modification-dependent enzymes
-
Upload
vuongnguyet -
Category
Documents
-
view
213 -
download
0
Transcript of modification-dependent enzymes
SURVEY AND SUMMARY
The other face of restrictionmodification-dependent enzymesWil A M Loenen1 and Elisabeth A Raleigh2
1Leiden University Medical Center PO Box 9600 2300RC Leiden The Netherlands and 2New England BiolabsInc 240 County Road Ipswich MA 01938-2723 USA
Received April 15 2013 Revised July 2 2013 Accepted July 10 2013
ABSTRACT
The 1952 observation of host-induced non-heredi-tary variation in bacteriophages by SalvadorLuria and Mary Human led to the discovery in the1960s of modifying enzymes that glucosylatehydroxymethylcytosine in T-even phages and ofgenes encoding corresponding host activities thatrestrict non-glucosylated phage DNA rglA and rglB(restricts glucoseless phage) In the 1980rsquos appreci-ation of the biological scope of these activities wasdramatically expanded with the demonstration thatplant and animal DNA was also sensitive to restric-tion in cloning experiments The rgl genes wererenamed mcrA and mcrBC (modified cytosine re-striction) The new class of modification-dependentrestriction enzymes was named Type IV as distinctfrom the familiar modification-blocked Types IndashIII Athird Escherichia coli enzyme mrr (modified DNArejection and restriction) recognizes both methyl-cytosine and methyladenine In recent years theuniverse of modification-dependent enzymes hasexpanded greatly Technical advances allow use ofType IV enzymes to study epigenetic mechanisms inmammals and plants Type IV enzymes recognizemodified DNA with low sequence selectivity andhave emerged many times independently during evo-lution Here we review biochemical and structuraldata on these proteins the resurgent interest inType IV enzymes as tools for epigenetic researchand the evolutionary pressures on these systems
GENETICS BIOCHEMISTRY AND STRUCTURES
Historical sketch
Like conventional modification-blocked restriction modi-fication-dependent restriction originally was diagnosed
owing to its biological effects when interstrain DNAtransfer was unexpectedly inhibited At the start phageswere the investigatory vehicles moving betweenEscherichia coli K12 E coli B and E coli C or Shigelladysenteria Sh (12) Later reduced plasmid phage orchromosomal transfer was found when alien modificationpatterns were present (3ndash5) Incoming DNA neededthe endogenous (for E coli K12) modification ofAm6ACN(6)GTGC (MEcoKI the A opposite theunderlined T is also modified) and Gm6ATC (Dam)Cm5CWGG (Dcm) occasionally had effects (6)lsquoOutgoingrsquo DNA was better accepted in many taxawithout any of these (7ndash10)
Progress in cloning and sequencing of restrictionenzyme (REase) genes other nucleases methyltransferase(MTase) genes and motor proteins began to feed data intoefforts to classify sequences and abstract from them sig-natures predictive of particular functions eg (11ndash15)Such signatures often correlate with physical proteindomains These domains can be split off from theoriginal protein and added to another and will thenoperate (mostly) as they are supposed to This result isthe basis for protein tagging with reporters and epitopesby molecular biologists As we see from the structuralorganization of modification-dependent REases thisapparently is also the basis for a mix-and-match evolu-tionary process in real lifemdashgrab a DNA-bindingdomain here a nuclease domain there and yoursquove got asite-specific (sort of) nuclease Sometimes a dimerizationsurface or a regulatory domain is needed as well
Finally with the advent of massive genome sequencingbioinformatic analysis has become a hypothesis generatorso that well-chosen biological and enzymatic tests can(hopefully) allow quick creation of strains and enzymesfor further research (16)
What biological DNA modifications are there
Biological DNA modifications have been studied formany years and much is known about their distribution
To whom correspondence should be addressed Tel +31 85 878 0248 Email wamloenenlumcnlCorrespondence may also be address to Elisabeth A Raleigh Tel +1 978 380 7238 Fax +1 978 921 1350 Email Raleighnebcom
The authors wish it to be known that in their opinion the first two authors should be regarded as Joint First Authors
56ndash69 Nucleic Acids Research 2014 Vol 42 No 1 Published online 29 August 2013doi101093nargkt747
The Author(s) 2013 Published by Oxford University PressThis is an Open Access article distributed under the terms of the Creative Commons Attribution License (httpcreativecommonsorglicensesby30) whichpermits unrestricted reuse distribution and reproduction in any medium provided the original work is properly cited
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
and the enzymes involved (17ndash19) Well-known basemodifications are C-5-methylcytosine (m5C)N4-methylcytosine (m4C) and N6-methyladenine (m6A)(Figure 1) These are widely distributed in cellular organ-isms particularly prokaryotes Other base modificationshave long been known in bacteriophage prominently5-hydroxymethylcytosine (hm5C) and derivatives of itwith sugar residues attached (ghm5C) (Figure 1) and5-hydroxymethyluracil (hm5U) Unusual modificationsof adenine have also been studied in phage Mu [Mommodification (20)] Fairly recently as methods for detec-tion of low frequency modifications have improved someof these exotic base modifications have also beenrecognized in higher organisms [hm5C (2122)] andlower eukaryotes [hm5U and the sugar-derivatized Jbase (23)] Bioinformatic investigation of coding se-quences related to modification enzymes suggests thatadditional unrecognized base modifications may still bediscovered (2425)
It is not only bases in DNA that may be modifiedEnzymatic sulfur modification of the phosphodiesterbackbone of DNA (PT-DNA for phosphothioesterDNA Figure 1) has recently been discovered in prokary-otes (26ndash28) PT-modified DNA is widespread modifica-tion is found in local sequence contexts compatible withsequence-specific addition and the similarity relationshipsamong the dnd genes encoding the modification machineryare consistent with extensive horizontal transfer as isfound for conventional restriction-modification (R-M)systems (29) This opens still further vistas for researchon the nature and biological consequences of modificationand restrictionSome of these modifications play important other roles
in the life of the host cell besides restriction wars in rep-lication timing in prokaryotes and in transcription regula-tion in prokaryotes and eukaryotes [eg (30ndash33)] Thistopic will not be addressed here except to note that themodifying enzymes that have acquired regulatory effectsin bacteria are normally conserved within a clade unlikecognate-modifying enzymes that accompany R-Msystems which are sporadically distributed (3435)
Molecular action what they do
Diversity of modification dependenceModifications that protect against conventional REasesinclude m5C hm5C ghm5C m6A m4C and mostrecently PT DNA (with sulfur replacing a non-bridgingoxygen) Neither hm5C nor ghm5C are known to beadded site-specifically instead they are found as universalsubstitutions in phage DNA The inverse could also betrue for each protective modification in Figure 1 thereare enzymes that attack DNA only when the modificationis present (Tables 1 and 2) Many of the enzymes weredescribed only recently and are distinct from the classicalexamples Many of the other modifications found inphages (1819) might be the object of undiscovered TypeIV enzymes hm5U and its glucosylated derivative J baseand the Mu modification N6 (1-acetoamido) adeninewould be interesting substratesThose modification-dependent enzymes that are classi-
fied as Type IV in REBASE (50) have been segregated(Table 1) from those classified as Type IIM (Table 2)The distinction between Type IIM and Type IV appearsto reflect production of defined bands on a gel in thereported characterizations This distinction may be mis-leading as bands on a gel can result from substratechoice in some cases (see further later in the text) As noother fundamental property unites the Type IV enzymesor distinguishes Type IIM from Type IV these authorsadvocate adding Type IIM to the Type IV classFor the most part those functions acting on hm5C also
act on m5C though with varying efficiency EcoK Mrr forwhich only in vivo evidence is available may be an excep-tionmdashit does not interfere with growth of hm5C-contain-ing T-even phages However phage-encoded restrictioninhibitors may confound interpretation of negativeresults obtained in vivo (see later in the text lsquoPhage-hostarms racersquo)
Figure 1 DNA modifications recognized by Type IV enzymesEnzymatic DNA modifications in the major groove of double-stranded DNA are methylation at cytosine C5 or N4 or at adenosineN6 and glucosylation of a pre-existing 5-hydroxymethylcytosine Thebeta-glucosyl derivative is shown other configurations and other sugarsare known to be added by some phages hm5C is incorporated duringreplication after conversion of the dCTP pool to hmdCTPPhosphorothioate modification of the backbone is carried outpostsynthetically Other biological DNA modifications are knownOnly those shown to elicit action of characterized Type IV enzymesare shown here
Nucleic Acids Research 2014 Vol 42 No 1 57
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
Other Mrr-related enzymes from Bacillus anthracisStreptomyces coelicolor and Zymomonas mobilis(identified bioinformatically see later in the text) werealso tested for activity in vivo Transformation efficiencyis reduced when a plasmid is prepared from a modifyinghost compared with the same plasmid from a non-mod-ifying host this reduction is alleviated when the corres-ponding Mrr-related gene is disrupted The specificity ofthis test depends on how thorough the genetic investiga-tion was if Dam Dcm EcoKM+ DNA transformsbetter than fully modified DNA modification specificitycould be either m6A or m5C or both hence the questionmarks in the table
The four systems listed for S coelicolor 3A constitute aparticularly exemplary analysis of this kind (37) In thiscase all four candidate R-M systems were deleted indi-vidually and together so that the effect of each could betested and each system was established in the related non-restricting host Streptomyces lividans For ScoA3Mrr theeffect of removing modifiable sites from the test plasmidwas also examined (for MEcoKI)
Diversity of functional organizationUnlike the classic Type IIP enzymes such as EcoRI andBamHI in which catalytic residues are embedded withinsequence-recognition structural elements the modifica-tion-dependent enzymes known so far exhibit separationof DNA binding and cleavage into different domains onthe same protein or even into different polypeptide chains(Table 3 and 4) In this they resemble Type I Type IIS orType III enzymes modification-blocked enzymes that alsoseparate recognition and cleavage For those also
Table 1 DNA modificationsa that elicit cleavage by Type IV enzymes
Protein m5C hm5C ghm5C m4C m6A PT References
EcoKMcrA (+) (+) () NT () NT (336)ScoA3McrA + NT NT NT (+) + (3738) Sco4631EcoKMcrBC + + (+) NT (33940)BanUMcrB (+) (41)BanUMcrB3 (+) (41)EcoKMrr (+) () () () (+) NT (4243)BanUMrr (+) (+) (41)ScoA3Mrr () (+) (37) Sco4213ZmoMrr (+) (+) (44) ZMO1932 ZmrrSauUSI + + NT (45)SauNewI (+) (46) NWMN_2386SepRPMcrR (+) (46) SERP2052ScoA3I (+) (37) Sco2863PvuRts1I family + + + NT NT (4748)GmrSD + NT NT (49)ScoA3II+III (+) () (37) Sco3261-62
aModifications m5C 5-methylcytosine hm5C 5-hydroxymethylcytosine ghm5C glucosylated hydroxymethylcytosine m4C N4-methylcytosinem6A N6-methyladenine PT phosphorothioation of non-bridging oxygen in DNA linkages also called S-DNA+ at least 100-fold less activity on this substrate than on substrates with+entry() (+) based on in vivo restriction of phage infection or plasmid transformation with appropriate host mutant configurations in vitro cleavageresults have not been reported(+) either m5C or m6A is recognized these were not distinguished in the reported experiments m6A sites tested were not cleaved but few modified sequences were testedNT not testedWhere the name found in REBASE (and listed at the left) is not the same as that used in the cited report the genomic locus_ID is given in theReferences column or the name used in the publication
Table 2 DNA modificationsa that elicit cleavage by other
modification-dependent enzymes (Type IIM)
Protein m5C hm5C ghm5C m4C m6A PT References
DpnI + NT (50)MspJI family + + NT (5152)SgeI + (53)AoxI +BisI +BlsI +GlaI + (54)GluI +KroI +MalI +MteI +PcsI +
aModifications m5C 5-methylcytosine hm5C 5-hydroxy-methylcytosine ghm5C glucosylated hydroxymethylcytosinem4C N4-methylcytosine m6A N6-methyladenine PT phosphoro-thioation of non-bridging oxygen in DNA linkages also called S-DNA+ at least 100-fold less activity on this substrate than on substrateswith+entry() (+) based on in vivo restriction of phage infection or plasmidtransformation with appropriate host mutant configurations in vitrocleavage results have not been reported(+) either m5C or m6A is recognized these were not distinguished inthe reported experiments m6A sites tested were not cleaved but few modified sequences weretestedNT not testedWhere the name found in REBASE (and listed at the left) isnot the same as that used in the cited report the genomiclocus_ID is given in the References column or the name used in thepublication
58 Nucleic Acids Research 2014 Vol 42 No 1
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
multiple evolutionary events apparently have occurred toconnect nuclease domains to recognition moieties (81)
Nuclease domainsEnzymes that recognize modified DNA with minimalsequence selectivity have emerged at least six times asexemplified by the McrA McrBC SauUSI MrrPvuRts1I and GmrSD families These exemplars are dis-cussed in more detail later in the text In brief nucleasedomains have been attached covalently or (for McrC) viaproteinndashprotein interaction to domains with DNA bindingand regulatory functions
EcoKMcrA carries a C-terminal H-N-Hc nucleasedomain identified bioinformatically (5882) (Figure 2)This nuclease domain is also found in modification-blocked nucleases (81) The purified binding-competentprotein did not cleave under a variety of buffer conditionsand cofactor additions (55) ScoA3McrA is designatedlsquoMcrArsquo due to its possession of a similar nucleasedomain For this enzyme cleavage depends on Mn2+ orCo2+ (38) and occurs at a variable distance from PT-modified sites Modification-blocked H-N-H REases alsooften exhibit unusual metal ion requirements [eg (83)]
McrBC The required McrC component (3940) is thenuclease moiety (65) (Figure 3) Mutational analysis
confirms that it is a PD-(DE)XK nuclease (65) in agree-ment with bioinformatic classification (64) Cleavageresults when McrC associates with full-lengthMcrBGTP complex bound to DNA and GTP ishydrolyzed (72) LlaJI a modification-blocked restrictionactivity exhibits a similar organization (85) althoughcleavage could not be demonstrated in vitroThe classic modification-dependent enzyme DpnI also
carries a PD-(DE)XK motif (see further later in the text)Mrr EcoKMrr contains a variant of the PD-(DE)XK
motif (6869) with the Mrr-N (E coli K12) presumedDNA-binding domain MspJI (see further later in thetext) also carries a nuclease domain in this family Aswith McrA McrBC and SauUSI nuclease domain simi-larity does not in itself dictate modification preferenceproperties the single-chain R-M system LlaGI hasconserved motifs characteristic of the E coli Mrrprotein but this enzyme does not target methylatedDNA (86)SauUSI This is a modification-dependent enzyme with
a phosphodiesterase cleavage domain akin to one origin-ally identified in phospholipase-D (45) Mutation of anyof the four conserved catalytic residues abolishes in vitroactivity This cleavage domain is also found in stand-alone
Table 3 Characteristics of Type IV restriction enzymes
Protein Subunitsdomains
DNABinding
Endonucleasedomain
NTP hydrolysis Recognition sitea comment and references
EcoKMcrA ndash In vitro cleavage not reportedN-terminal DBD (YgtR)m5CGR bound (5556) hm5Cm5C
discrimination (57)C-terminal H-N-Hc Bioinformatic ID (58) required for damage to DNA
in vivo (57)ScoA3McrA ndash Some CmCWGG and some S-DNA (PT modified)
sites are cleaved (38)N-terminal DBD Not related to EcoMcrA (38)C-terminal H-N-Hc 37 identical to EcoMcrA (38)
EcoKMcrBC GTP Rm5C(N30-35j)-(N30-3000)-Rm5Cb
McrB-N DUF3578 McrB binds DNA (40) via its N-terminal domain(59) by extrahelical modified base (60)
McrB-C P-loop NTPase (616263)McrC PD-(DE)XK Bioinformatic ID (64) Required for cleavage (3965)
EcoKMrr ndash m6A or m5C sequence specificity ambiguous(424366)
N-terminal Mrr-N Presumed DNA binding (67)C-terminal Mrr-cat (DE)
(DEQ) KBioinformatic ID (686970)
SauUS1 ATP or dATP Sm5CNGS two copies required for cleavage (45)N-terminal PLDc-2 (45)Middle P-loop NTPase (45)C-terminal DUF3427 (45) DBD
PvuRts1I PD-(DE)XK ndash mC(N11-13N9-10j)G 2-base extensions (47)Bioinformatic ID (64)
EcoCTGmrSD UTPgtgtGTP CTP Cuts T-even DNA (49)GmrS Motifs suggested DUF262 (71)To be confirmedGmrD DUF1524 To be determined
aRecognition sites (represented 5030) are those determined in vitro by binding or cleavage experimentsbMcrBC cleavage results in a double-strand cut near one Rm5C site (727374) but requires cooperation of two sites (3940) or a translocation block(73) The sites may be on different daughters across a fork (75) These are separated by 30ndash3000 (397274) and may be on either strand (3976)disposition of opposing nicks is not tightly constrained (73) and minor cleavage clusters are found 40 50 and 60 nt from the m5C (74)Degeneracy abbreviations B=C or G or T D=A or G or T H=A or C or T K=G or T M=A or C N=A or C or G or T R=A or GS=C or G V=A or C or G W=A or T Y=C or TCleavage positions are listed as (N to top cut to bottom cutj) If no number is listed the position of cleavage is not determined Space betweennumbers (eg PvuRts1I N11-13N9-10) indicates the range of positions at which cleavage may occur
Nucleic Acids Research 2014 Vol 42 No 1 59
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
nucleases and modification-blocked REases (8788)Interestingly two of the PLDc nuclease activities havebeen shown to work by a transesterification reaction likethat used by topoisomerases and transposases (8789)PvuRts1I has an apparently unusual nuclease domain
[ie not yet identified by sources curated by theNCBI Conserved Domain Database (90)] However thisenzyme was included in a categorization of PD-(DE)XKfamilies (64) a tentatively identified divalent metalion binding site Block B (47) corresponds to Block Dof Bujnicki and Rychlewski (64) Cleavage requiresMg2+ ions
EcoCTGmrSD Functional organization is less clearbut several possible nuclease motifs were identified inGmrS (71) Cleavage buffer contained Ca2+ and Mg2+
ions and UTP
Sequence context recognition
Many of the modification-dependent enzymescharacterized so far have little sequence specificity incontrast to conventional modification-blocked REasesRelatively complete characterization of sequence prefer-ence and cleavage position has been carried out for TypeIV enzymes EcoKMcrBC SauUSI and PvuRts1I(Table 3) and for Type IIM DpnI and the MspJI family(Table 4) Progress has been made with binding recogni-tion for EcoKMcrA Cleavage conditions have beenachieved for Sco3AMcrA (Table 3) For all of these rec-ognition of surrounding sequence context is degeneratewith preference for a neighboring base and frequently arequirement for two sites with suitable separation DpnI isin some respects an exception see later in the text
The remaining nucleases in Table 4 are less wellcharacterized The recognition sites might form a relatedseries It will be interesting to learn more about the rela-tionships among these and how the requirement formultiple modified positions is specified eg for BlsI andPkrI
McrA binding domainsThe two lsquoMcrArsquo enzymes are not similar in theirN-termini with homology limited to the C-terminalnuclease domain For EcoKMcrA (Figure 2) there isgood genetic evidence that base recognition lies in theN-terminus Extensive mutagenesis using insertion of
Table 4 Characteristics of other modification-dependent enzymes (Type IIM)
Protein Subunitsdomains
DNABinding
EndonucleaseDomain
Recognitionsite
Comment and references
DpnI family G m6AjTC 13 characterized isoschizomersN-terminal PD(DE)XK (50777879)C-terminal wHTH R m6AjTC (78)
MspJI family mC with preferences Second copy stimulates cleavageMspJI 5 mCNNR(N913j) (51)
N-terminal SRA-like (80)C-terminal Mrr-cat (DE)(DEQ)XK (51)
FspEI Like MspJI C m5C(N1216j) (52)LpnPI Like MspJI C m5CDG(N1014j) (52)AspBHI Like MspJI YS m5CNS(N812j) (52)RlaI Like MspJI V m5CW (52)SgrTI Like MspJI C m5CDS(N1014j) (52)SgeI Like MspJI m5CNNR(N913j) 49 identical to MspJI (53)No familyassigned
Information from httpwwwsibenzymecomproductsm2_type
AoxI jRG m5CYBisI G m5CjNGCBlsI RYNjR Y At least two m5C requiredGlaI R m5CjGY (54)GluI GmCjNG m5CKroI Gj m5CCGGCMteI GmCG m5CjNGm5CGm5CPcsI m5CG(N5jN2)m5CGPkrI Gm5CNjG m5C At least 3 m5C required
Figure 2 McrA functional domains Domain function was inferred in-directly from genetic analysis by Anton amp Raleigh (2004) (57) Manymutations in the N-terminal domain spared some activity in one ormore of three functional tests (grey segments) while others were defi-cient in all activities (black segments) One mutation (asterix) was fullyactive on m5C-containing substrates but fully inactive in the hm5Cchallenge in vivo Most mutations in the C-terminus (pale greysegment) retained function in one test that was interpreted asmeasuring m5C binding ability A predicted structural model byBujnicki Radlinska and Rychlewski (2000) (58) for this C-terminalregion is compatible with these results
60 Nucleic Acids Research 2014 Vol 42 No 1
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
five-amino acid linkers and classification with three func-tional tests allowed assignment of DNA recognition to theN-terminal portion with the C-terminal H-N-H domainimplicated in cleavage Of particular note a mutationdiscriminated in vivo between hm5C and m5C was foundin the N-terminal domain (57) The mutant was able tofully restrict bacteriophage lambda modified by MHpaIIbut not at all phage T4 containing hm5C In vitromodification-dependent binding was achieved with thefull-length His-tagged protein (5556) yielding a putativerecognition site (YgtR)mCG This recognition site iscompatible with in vivo observations (391)
Presumably the N-terminus of Sco3A McrA also rec-ognizes the DNA Recognition of both m5C and thephosphorothioate (PT) moiety must be accommodatedin the final reaction As either modification is sufficientto elicit cleavage more than one domain could beinvolved Cleavage occurred near some but not allDcm-modified sites (Cm5CWGG) Both synthetic
PT-containing oligonucleotides and unmethylated PT-modified plasmid were also cleaved on both sides of asymmetrically modified site PT modification is thoughtto be sequence-specific (2629) but the details are notyet clear
Novel McrB binding domainDNA-binding resides in the McrB N-terminal domain(405960) whereas translocation and cleavage coordin-ation reside in the C-terminal AAA+ regulatory andtranslocation domain (61ndash637273) The complete trans-lation product of mcrB McrB-L binds DNA specifically(4072) via its N-terminal 161 amino acids (aa) (59) Thecrystal structure of this domain (McrB-N) in complex withDNA has recently been published (60)McrB-N uses a strategy first discovered for DNA-MTase
action (92) it flips the C base out of the DNA helix into abinding pocket for inspection The pocket is large enoughto accommodate C m5C hm5C or m4C but too small if a
Figure 3 McrBC Assembly Model Two proteins are expressed from mcrB in vivo Both the complete protein (McrB-L) and a small one missing theN-terminus (McrB-S top row) bind GTP forming high-order multimers detected by gel filtration (second row) When visualized by scanningtransmission electron microscopy these appear as heptameric rings with a central channel Rings of McrB-L in top views show projections thatmay correspond to the N-terminal DNA-binding domain (red segment) Both forms can then associate with McrC judged again by gel filtrationMcrB-LGTP can bind to its specific substrate (RmC) in the absence of McrC (third row) in its presence the substrate is cleaved (fourth row)GTP hydrolysis is required for cleavage (arrow) a supershifted binding complex forms in the presence of GTP-gamma-S but no cleavage occursTranslocation accompanies GTP hydrolysis double-stranded cleavage requires collaboration between two complexes or a translocationblock The path of the DNA in the figure is arbitrary as is the conformation of McrC Modified from BourniquelAA and BickleTAComplex restriction enzymes NTP-driven molecular motors Bourniquel and Bickle (84) with permission Copyright 2002 Elsevier MassonSAS All rights reserved
Nucleic Acids Research 2014 Vol 42 No 1 61
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
glucose moiety is attached Conserved residues Y64 andL68 were noted to make van der Waals contact with themethyl group of the flipped out m5C these contacts aremissing when the pocket contains CThe flipping action can be compared with but is distinct
from that of eukaryotic m5C-specific regulatory proteinsthat use the SET and Ring-finger-Associated (SRA)domain to read DNA modification state (Figure 4A)This domain is found in most eukaryotes in accessoryproteins (eg UHRF1NP95SUVH5) of the DNMT1maintenance MTase (93ndash95) Despite the similarstrategy the McrB-N domain is not homologous butdisplays a distinct protein fold (60) Binding is accom-plished from the minor groove and extraction of the Ccreates a 30 bend toward the major groove resembling aglycosylase in this respect (Figure 4A) The eukaryoticproteins form a crescent from which loops project towrap around the DNA with recognition mediatedthrough both major and minor grooves (94) For McrB-N the authors suggest that the purine preference in the 50
position might result from flexibility constraints or inter-action with a non-conserved aa that occupies space left bythe flipped base Substitutions of this aa (Y41A or Y41Q)compromised binding activity
Sequence specificity novel phenotype and structuralmodel of MrrIn 1987 Heitman and Model discovered Mrr when theyfound that transfer of various foreign m6A MTasesinduced an SOS response due to DNA damage (42)This response to the presence of an incompatible MTaseremains the principal evidence that the E coli K12 Mrrprotein cleaves DNA Related proteins discussed later inthe text (Type IIM) have been more tractable for in vitrowork No concise description of the Mrr recognitionsequence has been forthcoming although several studieshave examined the spectrum of incompatible MTases(43669697) Both adenine and cytosine MTases confersensitivityMrr is also responsible for DNA damage that does not
depend on methylation at all foreign or otherwise Highhydrostatic pressure (HP) induces the SOS DNA damageresponse and lethality (98) The response did not dependon the activity of the endogenous MTases of E coli K12but did depend on both the presence of wild-type mrr andthe integrity of the SOS signal generation pathwayPossibly HP elicits a non-enzymatic modification or astructural change in DNA helicity that is acted on byMrr This HP phenotype was used to characterize mrrmutants which were fitted into a computer-assistedmodel of the Mrr protein (67) An N-terminal DNA-binding winged helix was proposed with a C-terminalnuclease domain previously identified (69) The functionalimportance of several conserved residues was confirmedSeveral of the selected mutants with null phenotypes wereisolated in a region far from the active site or bindingsurface identified bioinformatically These could affectinteraction with a component of the HP response Thisintriguing collection of informative mutants will illumin-ate in vitro characterization
Type IIM binding domainsType IIM enzymes of two families are well-characterizedwith respect to cleavage (Table 4) Crystal structures forboth have recently appeared
DpnI winged-helix DNA recognitionUnusually for modification-dependent enzymes DpnIcleaves a four-base site (Gm6ATC) with high fidelity(7799) to leave blunt ends when both strands of the siteare methylated At low concentration the enzyme nicksthe modified strand of a hemimethylated site (100) Thebehavior of the enzyme with respect to modificationpatterns within the canonical GATC sitemdashmodificationof C or A one strand or bothmdashhas been thoroughlyexplored (50) However only recently has cleavage ofnon-canonical adenine-methylated sites been examinedSiwek and co-workers (78) found evidence for consider-able relaxation of specificity at the outer base This experi-ment used substrates modified by a highly non-specificadenine MTase extensive DpnI cleavage cloning of thefragments and sequencing of the borders
Structure determination in the presence of DNA andvalidation experiments (78) place this enzyme togetherwith the other modification-dependent enzymes in thattwo domains segregate the cleavage function fromsequence recognition It also separates DpnI from theothers in that the cleavage domain also possesses somemodification and sequence specificity The main recogni-tion is accomplished by a monomeric winged-helixdomain which binds in the major groove and recognizesthe modifications on both strands in the same event Thestructure does not reveal a cleavage-competent complexhowever because the cleavage domain is far from theDNA Filter-binding experiments validated the ability ofthe C-terminal domain to bind alone to do so moretightly to fully methylated than to hemimethylated oligo-nucleotides and to compete with the full-length enzymereducing cleavage by it Expression of the N-terminalcleavage domain alone (in low yield) allowed validationof its cleavage activity Surprisingly this cleavage wasitself dependent on modification state and sequence ofthe substrate Modeling based on the structure of theblunt-end-producing Type IIP enzyme PvuII allowed pre-diction that the cleavage domain approaches from theminor groove Complete understanding of double-stranded cleavage will depend on understanding thedynamic transformations that allow the cleavage domainto approach and act at the site
MspJI coupling of cleavage with DNA recognitionThe six members of the MspJI family use the Mrr-catversion of the PD-(DE)XK nuclease to cut at definedlocations to one side relative to the modified base(12 bases on the modified strand 16ndash17 on the otherFigure 5A) only one modified base is required fordouble-strand cleavage to occur (unlike McrBC) (5152)However these enzymes are stimulated by the presence ofa second site in cis or in trans Symmetrically modifiedsites (such as m5CpGm5CpG in mammalian DNA)yield prominent bands of defined size (32 bp) containinga mixed population of sequences each with a m5C in the
62 Nucleic Acids Research 2014 Vol 42 No 1
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
middle (Figure 5B) This behavior is recapitulated by thePvuRts1I group of enzymes (exemplified by AbaSDFI inFigure 5C) except that the distances are shorter andrecognition of modification state is less well understoodDuring the characterization of MspJI Dcm
(Cm5CWGG) sites were the first recognized substrateyielding a clear banding pattern (51) Cleavage of differ-ently modified plasmids and designed oligonucleotidesubstrates allowed a good assessment of both modificationand sequence specificity This family shows preference forparticular bases nearby similar to McrBC
MspJI DNA recognition is mediated by anSRA-like domainRecently the crystal structure of MspJI without DNA hasbeen resolved at 205 A (80) Search of the MolecularModeling Database at NCBI (101) using VAST (102)
Figure 4 McrB-N in comparison to other base-flipping proteins (A)SRA domains SUVH5 (3Q0C) and UHRF1 (2ZKD) use loops extend-ing from a crescent formed from two beta sheets to flip C or m5C fromundeformed B-form DNA into a pocket (top row) whereas McrB-N(3SSC bottom row) uses loops from one beta-sheet to distort the DNAand flip the base It resembles the human alkyladenine glycosylase(1BNK) (bottom row) in bending the DNA toward the majorgroove while flipping the base via the minor groove Figure 5 ofSukackaite et al (60) (B) The SRA-like hemi-methylated 5mC recog-nition domains A ribbon model of the N-terminal domain of theMspJI structure (4F0Q and 4F0P left) compared with the SRAdomain of URHF1 (PDB 3FDE right) The crescent shape formed
Figure 4 Continuedby interacting beta sheets and helices aB and aC are the conservedfeatures of the SRA domain highlighted here Loops on the concaveside of UHRF1 participate in flipping the base and similar loops pre-sumably do so for MspJI Two of these vary in length among familymembers and may play roles in sequence context specificity Figure 2aand b from Horton et al (80)
Figure 5 Schematic diagrams of cleavage positions for MspJI andAbaSDFI Cleavage of both strands is elicited by a singly modifiedsite for both MspJI (A) and AbaSDFI (C) Cleavage position is fixedrelative to the modified site but with a four-base 50 extension forMspJI and a two-base 30 extension for AbaSDFI When a site is sym-metrically-modified (as for CpG sites in mammalian DNA) a 32 base-pair fragment is excised from the DNA (B) (A) Figure 2a and (B)Figure 3a reprinted with permission from Cohen-Karni et al (52)(C) Figure 5a from Wang et al (47)
Nucleic Acids Research 2014 Vol 42 No 1 63
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
showed that the N-terminal domain was structurallysimilar to that of the eukaryotic SRA domain with acrescent-shaped beta-sheet structure from which loopsproject (see Figure 4B and discussion earlier in the textMcrB) This structural homology allowed modeling of theDNA-bound structure with a flipped m5C The enzyme inthe crystal is a tetramer in which two monomers form aback-to-back dimer via the C-terminal regions thatcomprise the endonuclease Two back-to-back dimersgenerate a tetrameric protein with two cleavage domainspositioned (as in the Type IIP enzyme HindIII used formodeling the C-terminal cleavage domain interaction withDNA) so that a 4-base 50 extension would be created oncleavage of modeled DNA Cleavage is most efficient atmolar ratios that allow all four SRA-like domains to beoccupiedmdashtoo much enzyme prevents cleavage fromoccurring
Tracking and dimerization
McrBC as translocaseBourniquel and Bickle (84) have reviewed much of theenzymology of McrBC which will be briefly summarizedhere The Raleigh Bickle and Pingoud laboratories havecontributed to the following consistent picture of thein vitro reaction EcoKMcrBC cleavage results in adouble-strand cut near one RmC site (72ndash74) butrequires cooperation of two sites (3940) or a translocationblock (73) The sites may be on different daughters acrossa fork (75) These are separated by 30ndash3000 bp (397274)and may be on either strand (3976) cleavage occurs30ndash35 bases from the modified base with oppositenicks not tightly constrained (73) and minor cleavageclusters are found 40 50 and 60 nt from the m5C(74) hm5C DNA elicits cleavage also (39) A ring struc-ture is formed by 5ndash7 molecules of McrB in the presence ofGTP (Figure 3) (103) this complex can bind to a recog-nition element in DNA In the presence of McrC trans-location of the complex occurs and cleavage ensues whentranslocation is blocked Collision of translocatingcomplexes a protein barrier or a topological barrier willelicit double-strand cleavage adjacent to one recognitionelement or the other The enzyme will cleave when recog-nition elements are on opposite sides of a forked structure(75) This would allow action in vivo to prevent entry of aMTase gene even with rare sitesStructurally the McrB protein is proposed to be a
member of the AAA+ protein family of NTPases (104)many of which form ring-shaped complexes and partici-pate in molecular machines lsquoSensorrsquo segments found inthese proteins have been shown in some cases to playroles in coupling NTPase activity to intersubunit commu-nication and movement (105) Two of three elements ofthe GTP-binding motif proposed by Dila et al (61) werevalidated by mutational analysis (6562) The thirdproposed motif element was identified as amino acidsNTAD by Dila et al Alignment of AAA+ NTPases in(104) found this aligned with the motif designatedSensor-1 in (105) An interesting result was that mutationshere unexpectedly appeared to abrogate interaction withMcrC instead of changing which NTP would be
productive (62) It may instead play a role in coordinatingGTP binding and hydrolysis with DNA binding inter-action with McrC and cleavage
Intracellularly the story becomes more complex as themcrB gene encodes two products of 51 and 33 kD McrB-L and McrB-S the latter one starting from an in-frameinternal translation start site (106) Both in vivo andin vitro McrB-S can interfere with the function ofMcrB-L at least in part by forming complexes withMcrC unable to bind DNA (107) Both species can formmultimeric rings in the presence of GTP (103) as is usualfor AAA+NTPases (104)
SauUSI requires two sites and ATP hydrolysisSauUSI was originally annotated as a putative helicasefrom Staphylococcus aureus sp A single polypeptide issufficient for activity both in vitro and also in vivo as aclone in E coli using modified phage as a challenge Theamino acid sequence contains a PLDc domain at theN-terminus This contains a phosphodiesterase motiforiginally identified in Phospholipase D (108) it wasvalidated by mutagenesis of four catalytic residue candi-dates In the middle ATPase and helicase motifs wereproposed to account for ATP dependence of cleavageactivity A Domain of Unknown Function was identifiedat the C-terminus (Pfam DUF3427) (108) and wasproposed to recognize the substrate (45)
The purified enzyme cleaves modified DNA containingm5C and hm5C but not m4C in the presence of ATP ordATP but not other nucleotides The negative result form4C is firm plasmids modified at the same site by an m5CMTase (Dcm) or an m4C MTase (MPspGI) were testedThe former (Cm5CWGG) was sensitive whereas the latterwas resistant Thus the sequence preference is likely tobe satisfied m6A is likely not a substrate but few m6A-containing sites were examined
Like McrBC SauUSI requires the presence of two sitesfor efficient cleavage Presumably the ATPase activityparticipates in monitoring the presence of two sites asfor other nucleotide-dependent REases includingMcrBC The mechanism of communication is unknownThe enzyme belongs to a family of highly similarorthologues found in other sequenced Staphylococci(Tables 1 and 2) and more distant homologues can befound in sequenced bacterial and archaeal genomes
EVOLUTIONARY PRESSURES ON R-M SYSTEMS
Evolution by selfish propagation
One way to understand the massive variety of restrictionsystems and their sporadic distribution is to locate theevolutionary drivers of enzyme diversification in theenzyme genes themselves as selfish elements Work fromthe Kobayashi laboratory has elaborated clear examplesof selfish behavior in some Type II enzyme systems(109110) in which the host becomes lsquoaddictedrsquo to theR-M system Once a cell has acquired an R-M pair lossof the genes results in death of that cellrsquos descendants asthe REase is frequently still present and able to act on thegenome following loss of methylation activity In this
64 Nucleic Acids Research 2014 Vol 42 No 1
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
perspective the role played by modification-dependentenzymes is host defense to exclude systems with lsquoforeignrsquoMTase patterns and prevent the cell from loading up withparasites The exclusion event is accompanied by the deathof the cell (111ndash113) Weak sequence specificity of Type IVenzymes could then result from the need to control entryof a wide variety of invading systems
The selfish aspect certainly plays a role in R-M popula-tion biology but cannot be the whole story Type II R-Msystems can still be lost by inactivation of the R gene firstMoreover Type I systems escape this scenario withcomplex control of cleavage activity the restrictionassembly includes a methylation assembly to begin withtherefore the R protein cannot act unless the MTase ispresent in addition failure of the methylation activity inan intact complex leads to abrogation of R activity some-times by action of the ClpXP protease specifically on theR protein (114ndash116)
Furthermore in population terms a cell that acquiredand became addicted to an R-M system should lose incompetition with a sibling that never received thesystem Two factors could counter this First acquisitioncould be accompanied by an increase in the total numberof copies of the R-M system in the population asproposed for invading transposable elements Thisoverreplication results in more copies of the systemcreated than are lost whether to suicide or to other select-ive disadvantage [see eg (117118)] R-M gene amplifica-tion within a cell has been reported experimentally (119)but spread in a population has not been demonstrated yetA second factor that could counter the disability of addic-tion is localization of competition In a structured envir-onment (colonies on a plate or biofilm on large or smallsurfaces) killing of segregants preserves limiting nutrientsfor lineages that retain the toxinantitoxin pair (120121)Much of the real world is structured so this is an import-ant condition
Evolution by phage-host arms race
A second perspective supposes that the modification-dependent Type IV enzymes arose from the competitivecoevolutionary interaction between phages and theirhosts This was first enunciated by Revel and Luria (2)and most recently by Black and coworkers (122) seealso (123) That is hosts used modification-blockedrestriction to defend against phage infection T-evenphages developed methods of substituting modified basesfor the ordinary ones hosts developed Type IV enzymes indefense phages added sugar or other modifications (19) tothwart Type IV enzymes hosts extended Type IV enzymesto accommodate these decorations finally phages de-veloped protein inhibitors specific for these enzymes aswell T4 phages deliver a protein inhibitor (IPI) alongwith the DNA on infection which allows growth in thepresence of EcoCTGmrSD The locus responsible for thisinhibitor is highly variable among relatives of T4 asgmrSD is in enteric bacteria (both in distribution and inaa sequence) When phage with different IP1 alleles weretested for protection from cloned EcoCTGmrSD and itshomolog EcoUTGmrSD specificity was evident one or
the other or both or neither of the two homologs wascounteracted in individual cases (122) This variability ofthe outcome supports the idea that phage-host interactiondrives at least some of these developmentsIn this perspective the weak sequence selectivity of the
Type IV systems might simply reflect the lack of endogen-ous targets for the enzyme As the host does not present anyhm5C and the phage is completely substituted selection forsequence-specificity is weak Selection would act to spareany co-residentMTases This differs fromType II enzymeswhere the MTase and REase must co-evolve to allow thehost to survive Each Type IV system is compatible withsome suite of Type I-IIIMTases (and thus the R-M systemsas a whole) Methylated or hydroxymethylated bases maynot be recognized at all (EcoCTGmrSD) or the systemmay require one specific base in addition to the modifiedone (McrA McrBC MspJI and PvuRts1I) MTases mod-ifying sites not including that base are then compatible asDcm (CmCWGG) is compatible with the McrA McrBCand Mrr systems in E coli K12Type IV systems that restrict methylated bases in a
weakly specified sequence context confer an additionaladvantage in competition with phages Many phagessuch as have not evolved the nucleotide-substitutionstrategy used by the T-even phages These phagesnormally carry the modification pattern of the mostrecent host if the last host expressed an MTase creatinga susceptible site the Type IV enzyme of the new host willdestroy the invader and limit the infection This may beaccompanied by the death of the individual infectedtherefore protection can be conferred on the sibling popu-lation (111)A further implication of this scenario considers the fate
of a population invaded by phage Phage survival ofrestriction occurs at biologically relevant frequencies(106-102) The survivors of restriction carry the particu-lar methylation pattern of the particular cell and thus areresistant to all restriction systems it might have carried(Types IndashIV) A bacterial population as a whole thenbenefits from mechanisms that diversify the suite ofR-M systems so that such surviving phage do not havefree access to the entire population The extreme variabil-ity of R-M system content in isolates of the same species iscompatible with this idea [see eg (124) REBASEGenomes httptoolsnebcomvinczegenomes] Suchvariability also limits and shapes interstrain genetransfer (115125126)The Raleigh laboratory has built on elegant genetic
work with Type I enzymes in the Murray laboratory(127ndash130) to investigate a locus designated thelsquoImmigration Control locusrsquo or ICR that exemplifiesvariable R-M content Alternative DNA segments con-taining R-M systems are located at the same definedlocation in most E coli chromosomes between the yjiSand yjiA genes The ICR in the non-restricting E coli Cstrain [used in the original definition of the R-M phenom-enon (131)] contains a remnant of a Type I enzyme R geneand is 13 kilobase shorter than the same region in E coliK12 (132) The mechanism of segment replacement is stillunknown The ICR would be an example of the lsquodefenseislandsrsquo analyzed by the Koonin group (133) lsquoDefense
Nucleic Acids Research 2014 Vol 42 No 1 65
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
islandsrsquo contain genes that can defend against phage orother invading DNA these exhibit bioinformaticproperties similar to lsquomobilome islandsrsquo containing mobil-ization genes (transposases for example) However themechanism of mobilization has not been identified forthe ICR
FINAL THOUGHTS
The extreme diversity of R-M systems that recognizeordinary DNA seems likely to be approached by the di-versity of Type IV restriction systems Type IV enzymesare hard to find as most detection methods depend ondevelopment of genetic systems for each taxon or on ser-endipity Those characterized so far mostly stem frominitial genetic investigation of limits on infection trans-formation or transduction Barriers encountered provideleads to the genes involved Bioinformatic analysis hashelped to identify relatives which may be more tractableto biochemical investigation than the example originallyfound This approach has pitfalls the gene encodingMspJI was first thought to code for an enzyme recognizingan unmodified site because it is immediately adjacent toan (inactive it is now thought) cytosine MTase geneProvidentially the first expression host was devoid of sen-sitive sites whereas the first test substrate contained some(51) A combination of biological experiments with bio-informatics and biochemistry will be needed to reveal thefull spectrum of Type IV enzymes that may lurk within thevast universe of unidentified ORFs in bacterial systemsOne might begin with those strains whose genomes carryfew Type II systems Bacillus or Corynebacterium asopposed to Helicobacter or Neisseria [see the Genomessection of REBASE (50)]The role of lsquodefense islandsrsquo and their relation to the
lsquomobilomersquo in bacterial population biology remains to bedetermined If a defense island is similar to a mobilomeisland there should be a mechanism of mobilizationnearby which would boost the contribution oflsquooverreplicationrsquo to the account of selections acting onR-M systems R-M systems of all types can be found onor adjacent to known mobilizing elements (134135) buthave not been shown to move experimentallyOn another note it may turn out that evolutionarily
there is a continuum between the apparently modifica-tion-dependent and modification-blocked paths Onerelative of McrBC predicted by bioinformatics analysesis LlaI a system that recognizes an unmodified sequenceand requires two MTases to support it (136) The enzymeBamHI prefers to cleave DNA with m6A in its GGATCCsite and mutants can be isolated that require this modifiedbase (137) Are there native systems similarly protected bymodification of one position in the recognition site butdependent on modification at a different one An interest-ing evolutionary series can be imagined
ACKNOWLEDGEMENTS
The authors thank Rich Roberts and David Dryden forencouragement and critical comment They are grateful to
Tom Bickle Virginius Siksnys Yu Zheng and XiadongCheng for permission to reproduce figures They alsothank Ivan Correa for help drawing base structures andTasha Jose for illustration assistance They also thank theanonymous reviewers of a previous manuscript for caretaken with the review process greatly improving thepresent document
FUNDING
New England Biolabs (EAR) Funding for open accesscharge New England Biolabs (to EAR)
Conflict of interest statement None declared
REFERENCES
1 LuriaSE and HumanML (1952) A nonhereditary host-inducedvariation of bacterial viruses J Bacteriol 64 557ndash559
2 RevelHR and LuriaSE (1970) DNA-glucosylation in T-evenphage genetic determination and role in phage-host interactionAnnu Rev Genet 4 177ndash192
3 RaleighEA and WilsonG (1986) Escherichia coli K-12 restrictsDNA containing 5-methylcytosine Proc Natl Acad Sci USA83 9070ndash9074
4 Noyer-WeidnerM DiazR and ReinersL (1986) Cytosine-specific DNA modification interferes with plasmid establishmentin Escherichia coli K12 involvement of rglB Mol Gen Genet205 469ndash475
5 BlumenthalRM GregorySA and CooperiderJS (1985)Cloning of a restriction-modification system from Proteus vulgarisand its use in analyzing a methylase-sensitive phenotype inEscherichia coli J Bacteriol 164 501ndash509
6 SohailA LiebM DarM and BhagwatAS (1990) A generequired for very short patch repair in Escherichia coli is adjacentto the DNA cytosine methylase gene J Bacteriol 1724214ndash4221
7 MacNeilDJ (1988) Characterization of a unique methyl-specificrestriction system in Streptomyces avermitilis J Bacteriol 1705607ndash5612
8 VertesAA InuiM KobayashiM KurusuY and YukawaH(1993) Presence of mrr- and mcr-like restriction systems incoryneform bacteria Res Microbiol 144 181ndash185
9 MacalusoA and MettusAM (1991) Efficient transformation ofBacillus thuringiensis requires nonmethylated plasmid DNAJ Bacteriol 173 1353ndash1356
10 HolmesML NuttallSD and Dyall-SmithML (1991)Construction and use of halobacterial shuttle vectors and furtherstudies on Haloferax DNA gyrase J Bacteriol 173 3807ndash3813
11 PosfaiJ BhagwatAS PosfaiG and RobertsRJ (1989)Predictive motifs derived from cytosine methyltransferases NucleicAcids Res 17 2421ndash2435
12 MaloneT BlumenthalRM and ChengX (1995) Structure-guided analysis reveals nine sequence motifs conserved amongDNA amino-methyltransferases and suggests a catalyticmechanism for these enzymes J Mol Biol 253 618ndash632
13 TiminskasA ButkusV and JanulaitisA (1995) Sequence motifcharacteristics for DNA [cytosine-N4] and DNA [adenine-N6]methyltransferases Classification of all DNA methyltransferasesGene 157 3ndash11
14 DickmanMJ InglestonSM SedelnikovaSE RaffertyJBLloydRG GrasbyJA and HornbyDP (2002) The RuvABCresolvasome Eur J Biochem 269 5492ndash5501
15 KovallRA and MatthewsBW (1998) Structural functionaland evolutionary relationships between lambda-exonuclease andthe type II restriction endonucleases Proc Natl Acad Sci USA95 7893ndash7897
16 RobertsRJ ChangYC HuZ RachlinJN AntonBPPokrzywaRM ChoiHP FallerLL GuleriaJ HousmanGet al (2010) COMBREX a project to accelerate the functional
66 Nucleic Acids Research 2014 Vol 42 No 1
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
annotation of prokaryotic genomes Nucleic Acids Res 39D11ndashD14
17 RaePM and SteeleRE (1978) Modified bases in the DNAs ofunicellular eukaryotes an examination of distributions andpossible roles with emphasis on hydroxymethyluracil indinoflagellates Biosystems 10 37ndash53
18 WarrenRAJ (1980) Modified bases in bacteriophage DNAsAnnu Rev Microbiol 34 137ndash158
19 Gommers-AmptJH and BorstP (1995) Hypermodified bases inDNA FASEB J 9 1034ndash1042
20 SwintonD HattmanS CrainPF ChengCS SmithDL andMcCloskeyJA (1983) Purification and characterization of theunusual deoxynucleoside alpha-N-(9-beta-D-20-deoxyribofuranosylpurin-6-yl)glycinamide specified by the phageMu modification function Proc Natl Acad Sci USA 807400ndash7404
21 TahilianiM KohKP ShenY PastorWA BandukwalaHBrudnoY AgarwalS IyerLM LiuDR AravindL et al(2009) Conversion of 5-methylcytosine to 5-hydroxymethylcytosinein mammalian DNA by MLL partner TET1 Science 324930ndash935
22 KriaucionisS and HeintzN (2009) The nuclear DNA base5-hydroxymethylcytosine is present in Purkinje neurons and thebrain Science 324 929ndash930
23 BorstP and SabatiniR (2008) Base J discovery biosynthesisand possible functions Annu RevMicrobiol 62 235ndash251
24 KaminskaKH and BujnickiJM (2008) Bacteriophage MuMom protein responsible for DNA modification is a new memberof the acyltransferase superfamily Cell Cycle 7 120ndash121
25 IyerLM TahilianiM RaoA and AravindL (2009) Predictionof novel families of enzymes involved in oxidative and othercomplex modifications of bases in nucleic acids Cell Cycle 81698ndash1710
26 ZhouX HeX LiangJ LiA XuT KieserT HelmannJDand DengZ (2005) A novel DNA modification by sulphur MolMicrobiol 57 1428ndash1438
27 OuHY HeX ShaoY TaiC RajakumarK and DengZ(2009) dndDB A Database Focused on Phosphorothioation ofthe DNA Backbone PloS One 4 e5132
28 XuT LiangJ ChenS WangL HeX YouD WangZLiA XuZ ZhouX et al (2009) DNA phosphorothioation inStreptomyces lividans mutational analysis of the dnd locusBMC Microbiol 9 41
29 WangL ChenS VerginKL GiovannoniSJ ChanSWDeMottMS TaghizadehK CorderoOX CutlerMTimberlakeS et al (2011) DNA phosphorothioation iswidespread and quantized in bacterial genomes Proc Natl AcadSci USA 108 2963ndash2968
30 Loslashbner-OlesenA SkovgaardO and MarinusMG (2005) Dammethylation coordinating cellular processes Curr OpinMicrobiol 8 154ndash160
31 LowDA and CasadesusJ (2008) Clocks and switches bacterialgene regulation by DNA adenine methylation Curr OpinMicrobiol 11 106ndash112
32 FreitagM and SelkerEU (2005) Controlling DNA methylationMany roads to one modification Curr Opin Genet Dev 15191ndash199
33 ProhaskaSJ StadlerPF and KrakauerDC (2010) Innovationin gene regulation The case of chromatin computation J TheorBiol 265 27ndash44
34 SeshasayeeASN SinghP and KrishnaS (2012) Context-dependent conservation of DNA methyltransferases in bacteriaNucleic Acids Res 40 7066ndash7073
35 AntonBP (2011) Dissertation Boston University BostonMassachusetts USA
36 RaleighEA TrimarchiR and RevelH (1989) Genetic andphysical mapping of the mcrA (rglA) and mcrB (rglB) loci ofEscherichia coli K-12 Genetics 122 279ndash296
37 Gonzalez-CeronG Miranda-OlivaresOJ and Servın-GonzalezL(2009) Characterization of the methyl-specific restriction system ofStreptomyces coelicolor A3(2) and of the role played by laterallyacquired nucleases FEMS Microbiol Lett 301 35ndash43
38 LiuG OuHY WangT LiL TanH ZhouXRajakumarK DengZ and HeX (2010) Cleavage of
phosphorothioated DNA and methylated DNA by the type IVrestriction endonuclease ScoMcrA PLoS Genet 6 e1001253
39 SutherlandE CoeL and RaleighEA (1992) McrBC amultisubunit GTP-dependent restriction endonuclease J MolBiol 225 327ndash348
40 KrugerT WildC and Noyer-WeidnerM (1995) McrB aprokaryotic protein specifically recognizing DNA containingmodified cytosine residues EMBO J 14 2661ndash2669
41 SitaramanR and LepplaSH (2012) Methylation-dependentDNA restriction in Bacillus anthracis Gene 494 44ndash50
42 HeitmanJ and ModelP (1987) Site-specific methylases inducethe SOS DNA repair response in Escherichia coli J Bacteriol169 3243ndash3250
43 Waite-ReesPA KeatingCJ MoranLS SlatkoBEHornstraLJ and BennerJS (1991) Characterisation andexpression of the Escherichia coli Mrr restriction systemJ Bacteriol 173 5207ndash5219
44 KerrAL JeonYJ SvensonCJ RogersPL and NeilanBA(2010) DNA restriction-modification systems in the ethanologenZymomonas mobilis ZM4 Appl MicrobiolBiotech 89 761ndash769
45 XuSY CorvagliaAR ChanSH ZhengY and LinderP(2011) A type IV modification-dependent restriction enzymeSauUSI from Staphylococcus aureus subsp aureus USA300Nucleic Acids Res 39 5597ndash5610
46 MonkIR ShahIM XuM TanMW and FosterTJ (2012)Transforming the untransformable application of directtransformation to manipulate genetically Staphylococcus aureusand Staphylococcus epidermidis mBio 3 e00277ndash00211
47 WangH GuanS QuimbyA Cohen-KarniD PradhanSWilsonG RobertsRJ ZhuZ and ZhengY (2011)Comparative characterization of the PvuRts1I family ofrestriction enzymes and their application in mapping genomic5-hydroxymethylcytosine Nucleic Acids Res 39 9294ndash9305
48 JanosiL YonemitsuH HongH and KajiA (1994) Molecularcloning and expression of a novel hydroxymethylcytosine-specificrestriction enzyme (PvuRts1I) modulated by glucosylation ofDNAJ Mol Biol 242 45ndash61
49 BairCL and BlackLW (2007) A type IV modificationdependent restriction nuclease that targets glucosylatedhydroxymethyl cytosine modified DNAs J Mol Biol 366768ndash778
50 RobertsRJ VinczeT PosfaiJ and MacelisD (2010)REBASEndasha database for DNA restriction and modificationenzymes genes and genomes Nucleic Acids Res 38 D234ndashD236
51 ZhengY Cohen-KarniD XuD ChinHG WilsonGPradhanS and RobertsRJ (2010) A unique family of Mrr-likemodification-dependent restriction endonucleases Nucleic AcidsRes 38 5527ndash5534
52 Cohen-KarniD XuD AponeL FomenkovA SunZDavisPJ KinneySRM Yamada-MabuchiM XuSYDavisT et al (2011) The MspJI family of modification-dependent restriction endonucleases for epigenetic studies ProcNatl Acad Sci USA 108 11040ndash11045
53 LubysA VitkuteJ LubieneJ and JanulaitisA (2011)Restriction enzymes and their applications US Patent ApplicationPublication 20110207139 A1 httpappftusptogovnetahtmlPTOsearch-boolhtml (27 August 2013 date last accessed)
54 TarasovaGV NayakshinaTN and DegtyarevSK (2008)Substrate specificity of new methyl-directed DNA endonucleaseGlaI BMC Mol Biol 9 7
55 MulliganEA and DunnJJ (2008) Cloning purification andinitial characterization of E coli McrA a putative5-methylcytosine-specific nuclease Protein ExprPurif 62 98ndash103
56 MulliganEA HatchwellE McCorkleSR and DunnJJ (2010)Differential binding of Escherichia coli McrA protein to DNAsequences that contain the dinucleotide m5CpG Nucleic AcidsRes 38 1997ndash2005
57 AntonBP and RaleighEA (2004) Transposon-mediated linkerinsertion scanning mutagenesis of the Escherichia coli McrAendonuclease J Bacteriol 186 5699ndash5707
58 BujnickiJM RadlinskaM and RychlewskiL (2000) Atomicmodel of the 5-methylcytosine-specific restriction enzyme McrAreveals an atypical zinc finger and structural similarity tobetabetaalphaMe endonucleases Mol Microbiol 37 1280ndash1281
Nucleic Acids Research 2014 Vol 42 No 1 67
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
59 GastFU BrinkmannT PieperU KrugerT NoyerWeidnerM and PingoudA (1997) The recognition of methylatedDNA by the GTP-dependent restriction endonuclease McrBCresides in the N-terminal domain of McrB Biol Chem 378975ndash982
60 SukackaiteR GrazulisS TamulaitisG and SiksnysV (2012)The recognition domain of the methyl-specific endonucleaseMcrBC flips out 5-methylcytosine Nucleic Acids Res 407552ndash7562
61 DilaD SutherlandE MoranL SlatkoB and RaleighEA(1990) Genetic and sequence organization of the mcrBC locus ofEscherichia coli K-12 J Bacteriol 172 4888ndash4900
62 PieperU BrinkmannT KrugerT Noyer WeidnerM andPingoudA (1997) Characterization of the interaction between therestriction endonuclease McrBC from E coli and its cofactorGTP J Mol Biol 272 190ndash199
63 PieperU SchweitzerT GrollDH GastFU and PingoudA(1999) The GTP-binding domain of McrB more than just avariation on a common theme J Mol Biol 292 547ndash556
64 BujnickiJM and RychlewskiL (2001) Grouping together highlydiverged PD-(DE)XK nucleases and identification of novelsuperfamily members using structure-guided alignment ofsequence profiles J Mol MicrobiolBiotech 3 69ndash72
65 PieperU and PingoudA (2002) A mutational analysis of thePD DEXK motif suggests that McrC harbors the catalyticcenter for DNA cleavage by the GTP-dependent restrictionenzyme McrBC from Escherichia coli Biochemistry 415236ndash5244
66 KelleherJE and RaleighEA (1991) A novel activity inEscherichia coli K-12 that directs restriction of DNA modified atCG dinucleotides J Bacteriol 173 5220ndash5223
67 OrlowskiJ MebrhatuMT MichielsCW BujnickiJM andAertsenA (2008) Mutational analysis and a structural model ofmethyl-directed restriction enzyme Mrr Biochem Biophys ResCommun 377 862ndash866
68 AravindL MakarovaKS and KooninEV (2000) SURVEYAND SUMMARY holliday junction resolvases and relatednucleases identification of new families phyletic distribution andevolutionary trajectories Nucleic Acids Res 28 3417ndash3432
69 BujnickiJM and RychlewskiL (2001) Identification of a PD-(DE)XK-like domain with a novel configuration of the endonucleaseactive site in the methyl-directed restriction enzyme Mrr and itshomologs Gene 267 183ndash191
70 SteczkiewiczK MuszewskaA KnizewskiL RychlewskiL andGinalskiK (2012) Sequence structure and functional diversity ofPD-(DE)XK phosphodiesterase superfamily Nucleic Acids Res40 7016ndash7045
71 BairCL RifatD and BlackLW (2007) Exclusion of glucosyl-hydroxymethylcytosine DNA containing bacteriophages isovercome by the injected protein inhibitor IPI J Mol Biol 366779ndash789
72 StewartFJ and RaleighEA (1998) Dependence of McrBCcleavage on distance between recognition elements Biol Chem379 611ndash616
73 PanneD RaleighEA and BickleTA (1999) The McrBCendonuclease translocates DNA in a reaction dependent on GTPhydrolysis J Mol Biol 290 49ndash60
74 PieperU GrollDH WunschS GastFU SpeckC MuckeNand PingoudA (2002) The GTP-dependent restriction enzymeMcrBC from Escherichia coli forms high-molecular masscomplexes with DNA and produces a cleavage patternwith a characteristic 10-base pair repeat Biochemistry 415245ndash5254
75 IshikawaK HandaN SearsL RaleighEA and KobayashiI(2011) Cleavage of a model DNA replication fork by amethyl-specific endonuclease Nucleic Acids Res 39 5489ndash5498
76 StewartFJ PanneD BickleTA and RaleighEA (2000)Methyl-specific DNA binding by McrBC a modification-dependent restriction enzyme J Mol Biol 298 611ndash622
77 LacksS and GreenbergB (1977) Complementary specificity ofrestriction endonucleases DpnI and DpnII with respect to DNAmethylation J Mol Biol 114 153ndash168
78 SiwekW CzapinskaH BochtlerM BujnickiJM andSkowronekK (2012) Crystal structure and mechanism
of action of the N6-methyladenine-dependent type IIMrestriction endonuclease RDpnI Nucleic Acids Res 407563ndash7572
79 LacksS and GreenbergB (1975) A deoxyribonuclease ofDiplococcus pneumoniae specific for methylated DNA J BiolChem 250 4060ndash4066
80 HortonJR MabuchiMY Cohen-KarniD ZhangXGriggsRM SamaranayakeM RobertsRJ ZhengY andChengX (2012) Structure and cleavage activity of the tetramericMspJI DNA modification-dependent restriction endonucleaseNucleic Acids Res 40 9763ndash9773
81 CymermanIA ObarskaA SkowronekKJ LubysA andBujnickiJM (2006) Identification of a new subfamily of HNHnucleases and experimental characterization of a representativemember HphI restriction endonuclease Proteins 65 867ndash876
82 GorbalenyaAE (1994) Self-splicing group I and group II intronsencode homologous (putative) DNA endonucleases of a newfamily Protein Sci 3 1117ndash1120
83 ChanSH OpitzL HigginsL OrsquoLoaneD and XuSY (2010)Cofactor requirement of HpyAV restriction endonuclease PloSOne 5 e9071
84 BourniquelAA and BickleTA (2002) Complex restrictionenzymes NTP-driven molecular motors Biochimie 84 1047ndash1059
85 OrsquoDriscollJ HeiterDF WilsonGG FitzgeraldGFRobertsR and van SinderenD (2006) A genetic dissection ofthe LlaJI restriction cassette reveals insights on a novelbacteriophage resistance system BMC Microbiol 6 40
86 SmithRM DiffinFM SaveryNJ JosephsenJ andSzczelkunMD (2009) DNA cleavage and methylation specificityof the single polypeptide restriction-modification enzyme LlaGINucleic Acids Res 37 7206ndash7218
87 InterthalH PouliotJJ and ChampouxJJ (2001) The tyrosyl-DNA phosphodiesterase Tdp1 is a member of the phospholipaseD superfamily Proc Natl Acad Sci USA 98 12009ndash12014
88 SasnauskasG HalfordSE and SiksnysV (2003) How the BfiIrestriction enzyme uses one active site to cut two DNA strandsProc Natl Acad Sci USA 100 6410ndash6415
89 SasnauskasG ConnollyBA HalfordSE and SiksnysV (2007)Site-specific DNA transesterification catalyzed by a restrictionenzyme Proc Natl Acad Sci USA 104 2115ndash2120
90 Marchler-BauerA ZhengC ChitsazF DerbyshireMKGeerLY GeerRC GonzalesNR GwadzM HurwitzDILanczyckiCJ et al (2012) CDD conserved domains andprotein three-dimensional structure Nucleic Acids Res 41D348ndashD352
91 KravetzAN ZakharovaMV BeletskayaIV SinevaEVDenjmuchametovMM PetrovSI GlatmanLI andSoloninAS (1993) The cleavage sites and localization of genesencoding the restriction endonucleases Eco1831I and EcoHIGene 129 153ndash154
92 KlimasauskasS KumarS RobertsRJ and ChengX (1994)HhaI methyltransferase flips its target base out of the DNA helixCell 76 357ndash369
93 HashimotoH VertinoPM and ChengX (2010) Molecularcoupling of DNA methylation and histone methylationEpigenomics 2 657ndash669
94 RajakumaraE LawJA SimanshuDK VoigtPJohnsonLM ReinbergD PatelDJ and JacobsenSE (2011)A dual flip-out mechanism for 5mC recognition by theArabidopsis SUVH5 SRA domain and its impact on DNAmethylation and H3K9 dimethylation in vivo Genes Dev 25137ndash152
95 SharifJ and KosekiH (2011) Recruitment of Dnmt1 roles of theSRA protein Np95 (Uhrfi) and other factors Prog Mol BiolTransl Sci 101 289ndash310
96 Tesfazgi MebrhatuM WywialE GhoshA MichielsCWLindnerAB TaddeiF BujnickiJM Van MelderenL andAertsenA (2011) Evidence for an evolutionary antagonismbetween Mrr and Type III modification systems Nucleic AcidsRes 39 5991ndash6001
97 GrantGN JesseeJ BloomFR and HanahanD (1990)Differential plasmid rescue from transgenic mouse DNAs intoEscherichia coli methylation-restriction mutants Proc Natl AcadSci USA 87 4645ndash4649
68 Nucleic Acids Research 2014 Vol 42 No 1
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
98 AertsenA and MichielsCW (2005) Mrr instigates the SOSresponse after high pressure stress in Escherichia coli MolMicrobiol 58 1381ndash1391
99 WeiH TherrienC BlanchardA GuanS and ZhuZ (2008)The Fidelity Index provides a systematic quantitation of staractivity of DNA restriction endonucleases Nucleic Acids Res36 e50
100 LuL PatelH and BisslerJJ (2002) Optimizing DpnIdigestion to detect replicated DNA Biotechniques 33 316ndash318
101 MadejT AddessKJ FongJH GeerLY GeerRCLanczyckiCJ LiuC LuS Marchler-BauerAPanchenkoAR et al (2011) MMDB 3D structures andmacromolecular interactions Nucleic Acids Res 40 D461ndashD464
102 GibratJF MadejT and BryantSH (1996) Surprisingsimilarities in structure comparison Curr Opin Struct Biol 6377ndash385
103 PanneD MullerSA WirtzS EngelA and BickleTA (2001)The McrBC restriction endonuclease assembles into a ringstructure in the presence of G nucleotides EMBO J 203210ndash3217
104 NeuwaldAF AravindL SpougeJL and KooninEV (1999)AAA+ A class of chaperone-like ATPases associated with theassembly operation and disassembly of protein complexesGenome Res 9 27ndash43
105 OguraT and WilkinsonAJ (2001) AAA+ superfamilyATPases common structurendashdiverse function Genes Cells 6575ndash597
106 BearyTP BraymerHD and AchbergerEC (1997) Evidenceof participation of McrB(S) in McrBC restriction in Escherichiacoli K-12 J Bacteriol 179 7768ndash7775
107 PanneD RaleighEA and BickleTA (1998) McrBs amodulator peptide for McrBC activity EMBO J 175477ndash5483
108 PuntaM CoggillPC EberhardtRY MistryJ TateJBoursnellC PangN ForslundK CericG ClementsJ et al(2011) The Pfam protein families database Nucleic Acids Res40 D290ndashD301
109 NaitoT KusanoK and KobayashiI (1995) Selfish behavior ofrestriction-modification systems Science 267 897ndash899
110 KobayashiI (2001) Behavior of restriction-modification systemsas selfish mobile elements and their impact on genome evolutionNucleic Acids Res 29 3742ndash3756
111 FukudaE KaminskaKH BujnickiJM and KobayashiI(2008) Cell death upon epigenetic genome methylation a novelfunction of methyl-specific deoxyribonucleases Genome Biol 9R163
112 IshikawaK FukudaE and KobayashiI (2010) Conflictstargeting epigenetic systems and their resolution by cell deathnovel concepts for methyl-specific and other restriction systemsDNA Res 17 325ndash342
113 FukuyoM SasakiA and KobayashiI (2012) Success of asuicidal defense strategy against infection in a structured habitatSci Rep 2 238
114 MakovetsS TitheradgeAJ and MurrayNE (1998) ClpXand ClpP are essential for the efficient acquisition of genesspecifying type IA and IB restriction systems Mol Microbiol28 25ndash35
115 MurrayNE (2002) Immigration control of DNA in bacteriaself versus non-self Microbiology 148 3ndash20
116 MakovetsS PowellLM TitheradgeAJB BlakelyGW andMurrayNE (2004) Is modification sufficient to protect abacterial chromosome from a resident restriction endonucleaseMol Microbiol 51 135ndash147
117 OrgelLE and CrickFHC (1980) Selfish DNA the ultimateparasite Nature 284 604ndash607
118 KidwellMG and LischDR (2001) Perspective transposableelements parasitic DNA and genome evolution Evolution 551ndash24
119 SadykovM AsamiY NikiH HandaN ItayaMTanokuraM and KobayashiI (2003) Multiplication of arestriction-modification gene complex Mol Microbiol 48417ndash427
120 ChaoL and LevinBR (1981) Structured habitats and theevolution of anticompetitor toxins in bacteria Proc Natl AcadSci USA 78 6324ndash6328
121 MagnusonRD (2007) Hypothetical functions of toxin-antitoxinsystems J Bacteriol 189 6089ndash6092
122 RifatD WrightNT VarneyKM WeberDJ and BlackLW(2008) Restriction endonuclease inhibitor IPI of bacteriophageT4 a novel structure for a dedicated target J Mol Biol 375720ndash734
123 LabrieSJ SamsonJE and MoineauS (2010) Bacteriophageresistance mechanisms Nat Rev Microbiol 8 317ndash327
124 LinLF PosfaiJ RobertsRJ and KongH (2001)Comparative genomics of the restriction-modification systemsin Helicobacter pylori Proc Natl Acad Sci USA 982740ndash2745
125 MilkmanR RaleighEA McKaneM CrydermanDBilodeauP and McWeenyK (1999) Molecular evolution of theEscherichia coli chromosome V Recombination patterns amongstrains of diverse origin Genetics 153 539ndash554
126 HumbertO DorerMS and SalamaNR (2011)Characterization of Helicobacter pylori factors that controltransformation frequency and integration length during inter-strain DNA recombination Mol Microbiol 79 387ndash401
127 BarcusVA TitheradgeAJB and MurrayNE (1995) Thediversity of alleles at the hsd locus in natural populations ofEscherichia coli Genetics 140 1187ndash1197
128 SharpPM KelleherJE DanielAS CowanGM andMurrayNE (1992) Roles of selection and recombination in theevolution of type I restriction-modification systems inenterobacteria Proc Natl Acad Sci USA 89 9836ndash9840
129 LoenenWAM DanielAS BraymerHD and MurrayNE(1987) Organization and sequence of the hsd genes ofEscherichia coli K-12 J Mol Biol 198 159ndash170
130 SuriB and BickleTA (1985) EcoA the first member of a newfamily of type I restriction modification systems Geneorganization and enzymatic activities J Mol Biol 186 77ndash85
131 BertaniG and WeigleJJ (1953) Host-controlled variation inbacterial viruses J Bacteriol 65 113ndash121
132 SibleyMH and RaleighEA (2004) Cassette-like variation ofrestriction enzyme genes in Escherichia coli C and relativesNucleic Acids Res 32 522ndash534
133 MakarovaKS WolfYI SnirS and KooninEV (2011)Defense islands in bacterial and archaeal genomes and predictionof novel defense systems J Bacteriol 193 6039ndash6056
134 FurutaY AbeK and KobayashiI (2010) Genome comparisonand context analysis reveals putative mobile forms of restriction-modification systems and related rearrangements Nucleic AcidsRes 38 2428ndash2443
135 FurutaY and KobayashiI (2012) Restriction-modificationsystems as moblie epigenetic elements In RobertsAP andMullanyP (eds) Bacterial Integrative Mobile Genetic ElementsLandes Bioscience pp 85ndash103
136 OrsquoSullivanDJ and KlaenhammerTR (1998) Control ofexpression of LlaI restriction in Lactococcus lactis MolMicrobiol 27 1009ndash1020
137 WhitakerRD DornerLF and SchildkrautI (1999) A mutantof BamHI restriction endonuclease which requires N6-methyladenine for cleavage J Mol Biol 285 1525ndash1536
Nucleic Acids Research 2014 Vol 42 No 1 69
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
and the enzymes involved (17ndash19) Well-known basemodifications are C-5-methylcytosine (m5C)N4-methylcytosine (m4C) and N6-methyladenine (m6A)(Figure 1) These are widely distributed in cellular organ-isms particularly prokaryotes Other base modificationshave long been known in bacteriophage prominently5-hydroxymethylcytosine (hm5C) and derivatives of itwith sugar residues attached (ghm5C) (Figure 1) and5-hydroxymethyluracil (hm5U) Unusual modificationsof adenine have also been studied in phage Mu [Mommodification (20)] Fairly recently as methods for detec-tion of low frequency modifications have improved someof these exotic base modifications have also beenrecognized in higher organisms [hm5C (2122)] andlower eukaryotes [hm5U and the sugar-derivatized Jbase (23)] Bioinformatic investigation of coding se-quences related to modification enzymes suggests thatadditional unrecognized base modifications may still bediscovered (2425)
It is not only bases in DNA that may be modifiedEnzymatic sulfur modification of the phosphodiesterbackbone of DNA (PT-DNA for phosphothioesterDNA Figure 1) has recently been discovered in prokary-otes (26ndash28) PT-modified DNA is widespread modifica-tion is found in local sequence contexts compatible withsequence-specific addition and the similarity relationshipsamong the dnd genes encoding the modification machineryare consistent with extensive horizontal transfer as isfound for conventional restriction-modification (R-M)systems (29) This opens still further vistas for researchon the nature and biological consequences of modificationand restrictionSome of these modifications play important other roles
in the life of the host cell besides restriction wars in rep-lication timing in prokaryotes and in transcription regula-tion in prokaryotes and eukaryotes [eg (30ndash33)] Thistopic will not be addressed here except to note that themodifying enzymes that have acquired regulatory effectsin bacteria are normally conserved within a clade unlikecognate-modifying enzymes that accompany R-Msystems which are sporadically distributed (3435)
Molecular action what they do
Diversity of modification dependenceModifications that protect against conventional REasesinclude m5C hm5C ghm5C m6A m4C and mostrecently PT DNA (with sulfur replacing a non-bridgingoxygen) Neither hm5C nor ghm5C are known to beadded site-specifically instead they are found as universalsubstitutions in phage DNA The inverse could also betrue for each protective modification in Figure 1 thereare enzymes that attack DNA only when the modificationis present (Tables 1 and 2) Many of the enzymes weredescribed only recently and are distinct from the classicalexamples Many of the other modifications found inphages (1819) might be the object of undiscovered TypeIV enzymes hm5U and its glucosylated derivative J baseand the Mu modification N6 (1-acetoamido) adeninewould be interesting substratesThose modification-dependent enzymes that are classi-
fied as Type IV in REBASE (50) have been segregated(Table 1) from those classified as Type IIM (Table 2)The distinction between Type IIM and Type IV appearsto reflect production of defined bands on a gel in thereported characterizations This distinction may be mis-leading as bands on a gel can result from substratechoice in some cases (see further later in the text) As noother fundamental property unites the Type IV enzymesor distinguishes Type IIM from Type IV these authorsadvocate adding Type IIM to the Type IV classFor the most part those functions acting on hm5C also
act on m5C though with varying efficiency EcoK Mrr forwhich only in vivo evidence is available may be an excep-tionmdashit does not interfere with growth of hm5C-contain-ing T-even phages However phage-encoded restrictioninhibitors may confound interpretation of negativeresults obtained in vivo (see later in the text lsquoPhage-hostarms racersquo)
Figure 1 DNA modifications recognized by Type IV enzymesEnzymatic DNA modifications in the major groove of double-stranded DNA are methylation at cytosine C5 or N4 or at adenosineN6 and glucosylation of a pre-existing 5-hydroxymethylcytosine Thebeta-glucosyl derivative is shown other configurations and other sugarsare known to be added by some phages hm5C is incorporated duringreplication after conversion of the dCTP pool to hmdCTPPhosphorothioate modification of the backbone is carried outpostsynthetically Other biological DNA modifications are knownOnly those shown to elicit action of characterized Type IV enzymesare shown here
Nucleic Acids Research 2014 Vol 42 No 1 57
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
Other Mrr-related enzymes from Bacillus anthracisStreptomyces coelicolor and Zymomonas mobilis(identified bioinformatically see later in the text) werealso tested for activity in vivo Transformation efficiencyis reduced when a plasmid is prepared from a modifyinghost compared with the same plasmid from a non-mod-ifying host this reduction is alleviated when the corres-ponding Mrr-related gene is disrupted The specificity ofthis test depends on how thorough the genetic investiga-tion was if Dam Dcm EcoKM+ DNA transformsbetter than fully modified DNA modification specificitycould be either m6A or m5C or both hence the questionmarks in the table
The four systems listed for S coelicolor 3A constitute aparticularly exemplary analysis of this kind (37) In thiscase all four candidate R-M systems were deleted indi-vidually and together so that the effect of each could betested and each system was established in the related non-restricting host Streptomyces lividans For ScoA3Mrr theeffect of removing modifiable sites from the test plasmidwas also examined (for MEcoKI)
Diversity of functional organizationUnlike the classic Type IIP enzymes such as EcoRI andBamHI in which catalytic residues are embedded withinsequence-recognition structural elements the modifica-tion-dependent enzymes known so far exhibit separationof DNA binding and cleavage into different domains onthe same protein or even into different polypeptide chains(Table 3 and 4) In this they resemble Type I Type IIS orType III enzymes modification-blocked enzymes that alsoseparate recognition and cleavage For those also
Table 1 DNA modificationsa that elicit cleavage by Type IV enzymes
Protein m5C hm5C ghm5C m4C m6A PT References
EcoKMcrA (+) (+) () NT () NT (336)ScoA3McrA + NT NT NT (+) + (3738) Sco4631EcoKMcrBC + + (+) NT (33940)BanUMcrB (+) (41)BanUMcrB3 (+) (41)EcoKMrr (+) () () () (+) NT (4243)BanUMrr (+) (+) (41)ScoA3Mrr () (+) (37) Sco4213ZmoMrr (+) (+) (44) ZMO1932 ZmrrSauUSI + + NT (45)SauNewI (+) (46) NWMN_2386SepRPMcrR (+) (46) SERP2052ScoA3I (+) (37) Sco2863PvuRts1I family + + + NT NT (4748)GmrSD + NT NT (49)ScoA3II+III (+) () (37) Sco3261-62
aModifications m5C 5-methylcytosine hm5C 5-hydroxymethylcytosine ghm5C glucosylated hydroxymethylcytosine m4C N4-methylcytosinem6A N6-methyladenine PT phosphorothioation of non-bridging oxygen in DNA linkages also called S-DNA+ at least 100-fold less activity on this substrate than on substrates with+entry() (+) based on in vivo restriction of phage infection or plasmid transformation with appropriate host mutant configurations in vitro cleavageresults have not been reported(+) either m5C or m6A is recognized these were not distinguished in the reported experiments m6A sites tested were not cleaved but few modified sequences were testedNT not testedWhere the name found in REBASE (and listed at the left) is not the same as that used in the cited report the genomic locus_ID is given in theReferences column or the name used in the publication
Table 2 DNA modificationsa that elicit cleavage by other
modification-dependent enzymes (Type IIM)
Protein m5C hm5C ghm5C m4C m6A PT References
DpnI + NT (50)MspJI family + + NT (5152)SgeI + (53)AoxI +BisI +BlsI +GlaI + (54)GluI +KroI +MalI +MteI +PcsI +
aModifications m5C 5-methylcytosine hm5C 5-hydroxy-methylcytosine ghm5C glucosylated hydroxymethylcytosinem4C N4-methylcytosine m6A N6-methyladenine PT phosphoro-thioation of non-bridging oxygen in DNA linkages also called S-DNA+ at least 100-fold less activity on this substrate than on substrateswith+entry() (+) based on in vivo restriction of phage infection or plasmidtransformation with appropriate host mutant configurations in vitrocleavage results have not been reported(+) either m5C or m6A is recognized these were not distinguished inthe reported experiments m6A sites tested were not cleaved but few modified sequences weretestedNT not testedWhere the name found in REBASE (and listed at the left) isnot the same as that used in the cited report the genomiclocus_ID is given in the References column or the name used in thepublication
58 Nucleic Acids Research 2014 Vol 42 No 1
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
multiple evolutionary events apparently have occurred toconnect nuclease domains to recognition moieties (81)
Nuclease domainsEnzymes that recognize modified DNA with minimalsequence selectivity have emerged at least six times asexemplified by the McrA McrBC SauUSI MrrPvuRts1I and GmrSD families These exemplars are dis-cussed in more detail later in the text In brief nucleasedomains have been attached covalently or (for McrC) viaproteinndashprotein interaction to domains with DNA bindingand regulatory functions
EcoKMcrA carries a C-terminal H-N-Hc nucleasedomain identified bioinformatically (5882) (Figure 2)This nuclease domain is also found in modification-blocked nucleases (81) The purified binding-competentprotein did not cleave under a variety of buffer conditionsand cofactor additions (55) ScoA3McrA is designatedlsquoMcrArsquo due to its possession of a similar nucleasedomain For this enzyme cleavage depends on Mn2+ orCo2+ (38) and occurs at a variable distance from PT-modified sites Modification-blocked H-N-H REases alsooften exhibit unusual metal ion requirements [eg (83)]
McrBC The required McrC component (3940) is thenuclease moiety (65) (Figure 3) Mutational analysis
confirms that it is a PD-(DE)XK nuclease (65) in agree-ment with bioinformatic classification (64) Cleavageresults when McrC associates with full-lengthMcrBGTP complex bound to DNA and GTP ishydrolyzed (72) LlaJI a modification-blocked restrictionactivity exhibits a similar organization (85) althoughcleavage could not be demonstrated in vitroThe classic modification-dependent enzyme DpnI also
carries a PD-(DE)XK motif (see further later in the text)Mrr EcoKMrr contains a variant of the PD-(DE)XK
motif (6869) with the Mrr-N (E coli K12) presumedDNA-binding domain MspJI (see further later in thetext) also carries a nuclease domain in this family Aswith McrA McrBC and SauUSI nuclease domain simi-larity does not in itself dictate modification preferenceproperties the single-chain R-M system LlaGI hasconserved motifs characteristic of the E coli Mrrprotein but this enzyme does not target methylatedDNA (86)SauUSI This is a modification-dependent enzyme with
a phosphodiesterase cleavage domain akin to one origin-ally identified in phospholipase-D (45) Mutation of anyof the four conserved catalytic residues abolishes in vitroactivity This cleavage domain is also found in stand-alone
Table 3 Characteristics of Type IV restriction enzymes
Protein Subunitsdomains
DNABinding
Endonucleasedomain
NTP hydrolysis Recognition sitea comment and references
EcoKMcrA ndash In vitro cleavage not reportedN-terminal DBD (YgtR)m5CGR bound (5556) hm5Cm5C
discrimination (57)C-terminal H-N-Hc Bioinformatic ID (58) required for damage to DNA
in vivo (57)ScoA3McrA ndash Some CmCWGG and some S-DNA (PT modified)
sites are cleaved (38)N-terminal DBD Not related to EcoMcrA (38)C-terminal H-N-Hc 37 identical to EcoMcrA (38)
EcoKMcrBC GTP Rm5C(N30-35j)-(N30-3000)-Rm5Cb
McrB-N DUF3578 McrB binds DNA (40) via its N-terminal domain(59) by extrahelical modified base (60)
McrB-C P-loop NTPase (616263)McrC PD-(DE)XK Bioinformatic ID (64) Required for cleavage (3965)
EcoKMrr ndash m6A or m5C sequence specificity ambiguous(424366)
N-terminal Mrr-N Presumed DNA binding (67)C-terminal Mrr-cat (DE)
(DEQ) KBioinformatic ID (686970)
SauUS1 ATP or dATP Sm5CNGS two copies required for cleavage (45)N-terminal PLDc-2 (45)Middle P-loop NTPase (45)C-terminal DUF3427 (45) DBD
PvuRts1I PD-(DE)XK ndash mC(N11-13N9-10j)G 2-base extensions (47)Bioinformatic ID (64)
EcoCTGmrSD UTPgtgtGTP CTP Cuts T-even DNA (49)GmrS Motifs suggested DUF262 (71)To be confirmedGmrD DUF1524 To be determined
aRecognition sites (represented 5030) are those determined in vitro by binding or cleavage experimentsbMcrBC cleavage results in a double-strand cut near one Rm5C site (727374) but requires cooperation of two sites (3940) or a translocation block(73) The sites may be on different daughters across a fork (75) These are separated by 30ndash3000 (397274) and may be on either strand (3976)disposition of opposing nicks is not tightly constrained (73) and minor cleavage clusters are found 40 50 and 60 nt from the m5C (74)Degeneracy abbreviations B=C or G or T D=A or G or T H=A or C or T K=G or T M=A or C N=A or C or G or T R=A or GS=C or G V=A or C or G W=A or T Y=C or TCleavage positions are listed as (N to top cut to bottom cutj) If no number is listed the position of cleavage is not determined Space betweennumbers (eg PvuRts1I N11-13N9-10) indicates the range of positions at which cleavage may occur
Nucleic Acids Research 2014 Vol 42 No 1 59
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
nucleases and modification-blocked REases (8788)Interestingly two of the PLDc nuclease activities havebeen shown to work by a transesterification reaction likethat used by topoisomerases and transposases (8789)PvuRts1I has an apparently unusual nuclease domain
[ie not yet identified by sources curated by theNCBI Conserved Domain Database (90)] However thisenzyme was included in a categorization of PD-(DE)XKfamilies (64) a tentatively identified divalent metalion binding site Block B (47) corresponds to Block Dof Bujnicki and Rychlewski (64) Cleavage requiresMg2+ ions
EcoCTGmrSD Functional organization is less clearbut several possible nuclease motifs were identified inGmrS (71) Cleavage buffer contained Ca2+ and Mg2+
ions and UTP
Sequence context recognition
Many of the modification-dependent enzymescharacterized so far have little sequence specificity incontrast to conventional modification-blocked REasesRelatively complete characterization of sequence prefer-ence and cleavage position has been carried out for TypeIV enzymes EcoKMcrBC SauUSI and PvuRts1I(Table 3) and for Type IIM DpnI and the MspJI family(Table 4) Progress has been made with binding recogni-tion for EcoKMcrA Cleavage conditions have beenachieved for Sco3AMcrA (Table 3) For all of these rec-ognition of surrounding sequence context is degeneratewith preference for a neighboring base and frequently arequirement for two sites with suitable separation DpnI isin some respects an exception see later in the text
The remaining nucleases in Table 4 are less wellcharacterized The recognition sites might form a relatedseries It will be interesting to learn more about the rela-tionships among these and how the requirement formultiple modified positions is specified eg for BlsI andPkrI
McrA binding domainsThe two lsquoMcrArsquo enzymes are not similar in theirN-termini with homology limited to the C-terminalnuclease domain For EcoKMcrA (Figure 2) there isgood genetic evidence that base recognition lies in theN-terminus Extensive mutagenesis using insertion of
Table 4 Characteristics of other modification-dependent enzymes (Type IIM)
Protein Subunitsdomains
DNABinding
EndonucleaseDomain
Recognitionsite
Comment and references
DpnI family G m6AjTC 13 characterized isoschizomersN-terminal PD(DE)XK (50777879)C-terminal wHTH R m6AjTC (78)
MspJI family mC with preferences Second copy stimulates cleavageMspJI 5 mCNNR(N913j) (51)
N-terminal SRA-like (80)C-terminal Mrr-cat (DE)(DEQ)XK (51)
FspEI Like MspJI C m5C(N1216j) (52)LpnPI Like MspJI C m5CDG(N1014j) (52)AspBHI Like MspJI YS m5CNS(N812j) (52)RlaI Like MspJI V m5CW (52)SgrTI Like MspJI C m5CDS(N1014j) (52)SgeI Like MspJI m5CNNR(N913j) 49 identical to MspJI (53)No familyassigned
Information from httpwwwsibenzymecomproductsm2_type
AoxI jRG m5CYBisI G m5CjNGCBlsI RYNjR Y At least two m5C requiredGlaI R m5CjGY (54)GluI GmCjNG m5CKroI Gj m5CCGGCMteI GmCG m5CjNGm5CGm5CPcsI m5CG(N5jN2)m5CGPkrI Gm5CNjG m5C At least 3 m5C required
Figure 2 McrA functional domains Domain function was inferred in-directly from genetic analysis by Anton amp Raleigh (2004) (57) Manymutations in the N-terminal domain spared some activity in one ormore of three functional tests (grey segments) while others were defi-cient in all activities (black segments) One mutation (asterix) was fullyactive on m5C-containing substrates but fully inactive in the hm5Cchallenge in vivo Most mutations in the C-terminus (pale greysegment) retained function in one test that was interpreted asmeasuring m5C binding ability A predicted structural model byBujnicki Radlinska and Rychlewski (2000) (58) for this C-terminalregion is compatible with these results
60 Nucleic Acids Research 2014 Vol 42 No 1
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
five-amino acid linkers and classification with three func-tional tests allowed assignment of DNA recognition to theN-terminal portion with the C-terminal H-N-H domainimplicated in cleavage Of particular note a mutationdiscriminated in vivo between hm5C and m5C was foundin the N-terminal domain (57) The mutant was able tofully restrict bacteriophage lambda modified by MHpaIIbut not at all phage T4 containing hm5C In vitromodification-dependent binding was achieved with thefull-length His-tagged protein (5556) yielding a putativerecognition site (YgtR)mCG This recognition site iscompatible with in vivo observations (391)
Presumably the N-terminus of Sco3A McrA also rec-ognizes the DNA Recognition of both m5C and thephosphorothioate (PT) moiety must be accommodatedin the final reaction As either modification is sufficientto elicit cleavage more than one domain could beinvolved Cleavage occurred near some but not allDcm-modified sites (Cm5CWGG) Both synthetic
PT-containing oligonucleotides and unmethylated PT-modified plasmid were also cleaved on both sides of asymmetrically modified site PT modification is thoughtto be sequence-specific (2629) but the details are notyet clear
Novel McrB binding domainDNA-binding resides in the McrB N-terminal domain(405960) whereas translocation and cleavage coordin-ation reside in the C-terminal AAA+ regulatory andtranslocation domain (61ndash637273) The complete trans-lation product of mcrB McrB-L binds DNA specifically(4072) via its N-terminal 161 amino acids (aa) (59) Thecrystal structure of this domain (McrB-N) in complex withDNA has recently been published (60)McrB-N uses a strategy first discovered for DNA-MTase
action (92) it flips the C base out of the DNA helix into abinding pocket for inspection The pocket is large enoughto accommodate C m5C hm5C or m4C but too small if a
Figure 3 McrBC Assembly Model Two proteins are expressed from mcrB in vivo Both the complete protein (McrB-L) and a small one missing theN-terminus (McrB-S top row) bind GTP forming high-order multimers detected by gel filtration (second row) When visualized by scanningtransmission electron microscopy these appear as heptameric rings with a central channel Rings of McrB-L in top views show projections thatmay correspond to the N-terminal DNA-binding domain (red segment) Both forms can then associate with McrC judged again by gel filtrationMcrB-LGTP can bind to its specific substrate (RmC) in the absence of McrC (third row) in its presence the substrate is cleaved (fourth row)GTP hydrolysis is required for cleavage (arrow) a supershifted binding complex forms in the presence of GTP-gamma-S but no cleavage occursTranslocation accompanies GTP hydrolysis double-stranded cleavage requires collaboration between two complexes or a translocationblock The path of the DNA in the figure is arbitrary as is the conformation of McrC Modified from BourniquelAA and BickleTAComplex restriction enzymes NTP-driven molecular motors Bourniquel and Bickle (84) with permission Copyright 2002 Elsevier MassonSAS All rights reserved
Nucleic Acids Research 2014 Vol 42 No 1 61
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
glucose moiety is attached Conserved residues Y64 andL68 were noted to make van der Waals contact with themethyl group of the flipped out m5C these contacts aremissing when the pocket contains CThe flipping action can be compared with but is distinct
from that of eukaryotic m5C-specific regulatory proteinsthat use the SET and Ring-finger-Associated (SRA)domain to read DNA modification state (Figure 4A)This domain is found in most eukaryotes in accessoryproteins (eg UHRF1NP95SUVH5) of the DNMT1maintenance MTase (93ndash95) Despite the similarstrategy the McrB-N domain is not homologous butdisplays a distinct protein fold (60) Binding is accom-plished from the minor groove and extraction of the Ccreates a 30 bend toward the major groove resembling aglycosylase in this respect (Figure 4A) The eukaryoticproteins form a crescent from which loops project towrap around the DNA with recognition mediatedthrough both major and minor grooves (94) For McrB-N the authors suggest that the purine preference in the 50
position might result from flexibility constraints or inter-action with a non-conserved aa that occupies space left bythe flipped base Substitutions of this aa (Y41A or Y41Q)compromised binding activity
Sequence specificity novel phenotype and structuralmodel of MrrIn 1987 Heitman and Model discovered Mrr when theyfound that transfer of various foreign m6A MTasesinduced an SOS response due to DNA damage (42)This response to the presence of an incompatible MTaseremains the principal evidence that the E coli K12 Mrrprotein cleaves DNA Related proteins discussed later inthe text (Type IIM) have been more tractable for in vitrowork No concise description of the Mrr recognitionsequence has been forthcoming although several studieshave examined the spectrum of incompatible MTases(43669697) Both adenine and cytosine MTases confersensitivityMrr is also responsible for DNA damage that does not
depend on methylation at all foreign or otherwise Highhydrostatic pressure (HP) induces the SOS DNA damageresponse and lethality (98) The response did not dependon the activity of the endogenous MTases of E coli K12but did depend on both the presence of wild-type mrr andthe integrity of the SOS signal generation pathwayPossibly HP elicits a non-enzymatic modification or astructural change in DNA helicity that is acted on byMrr This HP phenotype was used to characterize mrrmutants which were fitted into a computer-assistedmodel of the Mrr protein (67) An N-terminal DNA-binding winged helix was proposed with a C-terminalnuclease domain previously identified (69) The functionalimportance of several conserved residues was confirmedSeveral of the selected mutants with null phenotypes wereisolated in a region far from the active site or bindingsurface identified bioinformatically These could affectinteraction with a component of the HP response Thisintriguing collection of informative mutants will illumin-ate in vitro characterization
Type IIM binding domainsType IIM enzymes of two families are well-characterizedwith respect to cleavage (Table 4) Crystal structures forboth have recently appeared
DpnI winged-helix DNA recognitionUnusually for modification-dependent enzymes DpnIcleaves a four-base site (Gm6ATC) with high fidelity(7799) to leave blunt ends when both strands of the siteare methylated At low concentration the enzyme nicksthe modified strand of a hemimethylated site (100) Thebehavior of the enzyme with respect to modificationpatterns within the canonical GATC sitemdashmodificationof C or A one strand or bothmdashhas been thoroughlyexplored (50) However only recently has cleavage ofnon-canonical adenine-methylated sites been examinedSiwek and co-workers (78) found evidence for consider-able relaxation of specificity at the outer base This experi-ment used substrates modified by a highly non-specificadenine MTase extensive DpnI cleavage cloning of thefragments and sequencing of the borders
Structure determination in the presence of DNA andvalidation experiments (78) place this enzyme togetherwith the other modification-dependent enzymes in thattwo domains segregate the cleavage function fromsequence recognition It also separates DpnI from theothers in that the cleavage domain also possesses somemodification and sequence specificity The main recogni-tion is accomplished by a monomeric winged-helixdomain which binds in the major groove and recognizesthe modifications on both strands in the same event Thestructure does not reveal a cleavage-competent complexhowever because the cleavage domain is far from theDNA Filter-binding experiments validated the ability ofthe C-terminal domain to bind alone to do so moretightly to fully methylated than to hemimethylated oligo-nucleotides and to compete with the full-length enzymereducing cleavage by it Expression of the N-terminalcleavage domain alone (in low yield) allowed validationof its cleavage activity Surprisingly this cleavage wasitself dependent on modification state and sequence ofthe substrate Modeling based on the structure of theblunt-end-producing Type IIP enzyme PvuII allowed pre-diction that the cleavage domain approaches from theminor groove Complete understanding of double-stranded cleavage will depend on understanding thedynamic transformations that allow the cleavage domainto approach and act at the site
MspJI coupling of cleavage with DNA recognitionThe six members of the MspJI family use the Mrr-catversion of the PD-(DE)XK nuclease to cut at definedlocations to one side relative to the modified base(12 bases on the modified strand 16ndash17 on the otherFigure 5A) only one modified base is required fordouble-strand cleavage to occur (unlike McrBC) (5152)However these enzymes are stimulated by the presence ofa second site in cis or in trans Symmetrically modifiedsites (such as m5CpGm5CpG in mammalian DNA)yield prominent bands of defined size (32 bp) containinga mixed population of sequences each with a m5C in the
62 Nucleic Acids Research 2014 Vol 42 No 1
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
middle (Figure 5B) This behavior is recapitulated by thePvuRts1I group of enzymes (exemplified by AbaSDFI inFigure 5C) except that the distances are shorter andrecognition of modification state is less well understoodDuring the characterization of MspJI Dcm
(Cm5CWGG) sites were the first recognized substrateyielding a clear banding pattern (51) Cleavage of differ-ently modified plasmids and designed oligonucleotidesubstrates allowed a good assessment of both modificationand sequence specificity This family shows preference forparticular bases nearby similar to McrBC
MspJI DNA recognition is mediated by anSRA-like domainRecently the crystal structure of MspJI without DNA hasbeen resolved at 205 A (80) Search of the MolecularModeling Database at NCBI (101) using VAST (102)
Figure 4 McrB-N in comparison to other base-flipping proteins (A)SRA domains SUVH5 (3Q0C) and UHRF1 (2ZKD) use loops extend-ing from a crescent formed from two beta sheets to flip C or m5C fromundeformed B-form DNA into a pocket (top row) whereas McrB-N(3SSC bottom row) uses loops from one beta-sheet to distort the DNAand flip the base It resembles the human alkyladenine glycosylase(1BNK) (bottom row) in bending the DNA toward the majorgroove while flipping the base via the minor groove Figure 5 ofSukackaite et al (60) (B) The SRA-like hemi-methylated 5mC recog-nition domains A ribbon model of the N-terminal domain of theMspJI structure (4F0Q and 4F0P left) compared with the SRAdomain of URHF1 (PDB 3FDE right) The crescent shape formed
Figure 4 Continuedby interacting beta sheets and helices aB and aC are the conservedfeatures of the SRA domain highlighted here Loops on the concaveside of UHRF1 participate in flipping the base and similar loops pre-sumably do so for MspJI Two of these vary in length among familymembers and may play roles in sequence context specificity Figure 2aand b from Horton et al (80)
Figure 5 Schematic diagrams of cleavage positions for MspJI andAbaSDFI Cleavage of both strands is elicited by a singly modifiedsite for both MspJI (A) and AbaSDFI (C) Cleavage position is fixedrelative to the modified site but with a four-base 50 extension forMspJI and a two-base 30 extension for AbaSDFI When a site is sym-metrically-modified (as for CpG sites in mammalian DNA) a 32 base-pair fragment is excised from the DNA (B) (A) Figure 2a and (B)Figure 3a reprinted with permission from Cohen-Karni et al (52)(C) Figure 5a from Wang et al (47)
Nucleic Acids Research 2014 Vol 42 No 1 63
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
showed that the N-terminal domain was structurallysimilar to that of the eukaryotic SRA domain with acrescent-shaped beta-sheet structure from which loopsproject (see Figure 4B and discussion earlier in the textMcrB) This structural homology allowed modeling of theDNA-bound structure with a flipped m5C The enzyme inthe crystal is a tetramer in which two monomers form aback-to-back dimer via the C-terminal regions thatcomprise the endonuclease Two back-to-back dimersgenerate a tetrameric protein with two cleavage domainspositioned (as in the Type IIP enzyme HindIII used formodeling the C-terminal cleavage domain interaction withDNA) so that a 4-base 50 extension would be created oncleavage of modeled DNA Cleavage is most efficient atmolar ratios that allow all four SRA-like domains to beoccupiedmdashtoo much enzyme prevents cleavage fromoccurring
Tracking and dimerization
McrBC as translocaseBourniquel and Bickle (84) have reviewed much of theenzymology of McrBC which will be briefly summarizedhere The Raleigh Bickle and Pingoud laboratories havecontributed to the following consistent picture of thein vitro reaction EcoKMcrBC cleavage results in adouble-strand cut near one RmC site (72ndash74) butrequires cooperation of two sites (3940) or a translocationblock (73) The sites may be on different daughters acrossa fork (75) These are separated by 30ndash3000 bp (397274)and may be on either strand (3976) cleavage occurs30ndash35 bases from the modified base with oppositenicks not tightly constrained (73) and minor cleavageclusters are found 40 50 and 60 nt from the m5C(74) hm5C DNA elicits cleavage also (39) A ring struc-ture is formed by 5ndash7 molecules of McrB in the presence ofGTP (Figure 3) (103) this complex can bind to a recog-nition element in DNA In the presence of McrC trans-location of the complex occurs and cleavage ensues whentranslocation is blocked Collision of translocatingcomplexes a protein barrier or a topological barrier willelicit double-strand cleavage adjacent to one recognitionelement or the other The enzyme will cleave when recog-nition elements are on opposite sides of a forked structure(75) This would allow action in vivo to prevent entry of aMTase gene even with rare sitesStructurally the McrB protein is proposed to be a
member of the AAA+ protein family of NTPases (104)many of which form ring-shaped complexes and partici-pate in molecular machines lsquoSensorrsquo segments found inthese proteins have been shown in some cases to playroles in coupling NTPase activity to intersubunit commu-nication and movement (105) Two of three elements ofthe GTP-binding motif proposed by Dila et al (61) werevalidated by mutational analysis (6562) The thirdproposed motif element was identified as amino acidsNTAD by Dila et al Alignment of AAA+ NTPases in(104) found this aligned with the motif designatedSensor-1 in (105) An interesting result was that mutationshere unexpectedly appeared to abrogate interaction withMcrC instead of changing which NTP would be
productive (62) It may instead play a role in coordinatingGTP binding and hydrolysis with DNA binding inter-action with McrC and cleavage
Intracellularly the story becomes more complex as themcrB gene encodes two products of 51 and 33 kD McrB-L and McrB-S the latter one starting from an in-frameinternal translation start site (106) Both in vivo andin vitro McrB-S can interfere with the function ofMcrB-L at least in part by forming complexes withMcrC unable to bind DNA (107) Both species can formmultimeric rings in the presence of GTP (103) as is usualfor AAA+NTPases (104)
SauUSI requires two sites and ATP hydrolysisSauUSI was originally annotated as a putative helicasefrom Staphylococcus aureus sp A single polypeptide issufficient for activity both in vitro and also in vivo as aclone in E coli using modified phage as a challenge Theamino acid sequence contains a PLDc domain at theN-terminus This contains a phosphodiesterase motiforiginally identified in Phospholipase D (108) it wasvalidated by mutagenesis of four catalytic residue candi-dates In the middle ATPase and helicase motifs wereproposed to account for ATP dependence of cleavageactivity A Domain of Unknown Function was identifiedat the C-terminus (Pfam DUF3427) (108) and wasproposed to recognize the substrate (45)
The purified enzyme cleaves modified DNA containingm5C and hm5C but not m4C in the presence of ATP ordATP but not other nucleotides The negative result form4C is firm plasmids modified at the same site by an m5CMTase (Dcm) or an m4C MTase (MPspGI) were testedThe former (Cm5CWGG) was sensitive whereas the latterwas resistant Thus the sequence preference is likely tobe satisfied m6A is likely not a substrate but few m6A-containing sites were examined
Like McrBC SauUSI requires the presence of two sitesfor efficient cleavage Presumably the ATPase activityparticipates in monitoring the presence of two sites asfor other nucleotide-dependent REases includingMcrBC The mechanism of communication is unknownThe enzyme belongs to a family of highly similarorthologues found in other sequenced Staphylococci(Tables 1 and 2) and more distant homologues can befound in sequenced bacterial and archaeal genomes
EVOLUTIONARY PRESSURES ON R-M SYSTEMS
Evolution by selfish propagation
One way to understand the massive variety of restrictionsystems and their sporadic distribution is to locate theevolutionary drivers of enzyme diversification in theenzyme genes themselves as selfish elements Work fromthe Kobayashi laboratory has elaborated clear examplesof selfish behavior in some Type II enzyme systems(109110) in which the host becomes lsquoaddictedrsquo to theR-M system Once a cell has acquired an R-M pair lossof the genes results in death of that cellrsquos descendants asthe REase is frequently still present and able to act on thegenome following loss of methylation activity In this
64 Nucleic Acids Research 2014 Vol 42 No 1
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
perspective the role played by modification-dependentenzymes is host defense to exclude systems with lsquoforeignrsquoMTase patterns and prevent the cell from loading up withparasites The exclusion event is accompanied by the deathof the cell (111ndash113) Weak sequence specificity of Type IVenzymes could then result from the need to control entryof a wide variety of invading systems
The selfish aspect certainly plays a role in R-M popula-tion biology but cannot be the whole story Type II R-Msystems can still be lost by inactivation of the R gene firstMoreover Type I systems escape this scenario withcomplex control of cleavage activity the restrictionassembly includes a methylation assembly to begin withtherefore the R protein cannot act unless the MTase ispresent in addition failure of the methylation activity inan intact complex leads to abrogation of R activity some-times by action of the ClpXP protease specifically on theR protein (114ndash116)
Furthermore in population terms a cell that acquiredand became addicted to an R-M system should lose incompetition with a sibling that never received thesystem Two factors could counter this First acquisitioncould be accompanied by an increase in the total numberof copies of the R-M system in the population asproposed for invading transposable elements Thisoverreplication results in more copies of the systemcreated than are lost whether to suicide or to other select-ive disadvantage [see eg (117118)] R-M gene amplifica-tion within a cell has been reported experimentally (119)but spread in a population has not been demonstrated yetA second factor that could counter the disability of addic-tion is localization of competition In a structured envir-onment (colonies on a plate or biofilm on large or smallsurfaces) killing of segregants preserves limiting nutrientsfor lineages that retain the toxinantitoxin pair (120121)Much of the real world is structured so this is an import-ant condition
Evolution by phage-host arms race
A second perspective supposes that the modification-dependent Type IV enzymes arose from the competitivecoevolutionary interaction between phages and theirhosts This was first enunciated by Revel and Luria (2)and most recently by Black and coworkers (122) seealso (123) That is hosts used modification-blockedrestriction to defend against phage infection T-evenphages developed methods of substituting modified basesfor the ordinary ones hosts developed Type IV enzymes indefense phages added sugar or other modifications (19) tothwart Type IV enzymes hosts extended Type IV enzymesto accommodate these decorations finally phages de-veloped protein inhibitors specific for these enzymes aswell T4 phages deliver a protein inhibitor (IPI) alongwith the DNA on infection which allows growth in thepresence of EcoCTGmrSD The locus responsible for thisinhibitor is highly variable among relatives of T4 asgmrSD is in enteric bacteria (both in distribution and inaa sequence) When phage with different IP1 alleles weretested for protection from cloned EcoCTGmrSD and itshomolog EcoUTGmrSD specificity was evident one or
the other or both or neither of the two homologs wascounteracted in individual cases (122) This variability ofthe outcome supports the idea that phage-host interactiondrives at least some of these developmentsIn this perspective the weak sequence selectivity of the
Type IV systems might simply reflect the lack of endogen-ous targets for the enzyme As the host does not present anyhm5C and the phage is completely substituted selection forsequence-specificity is weak Selection would act to spareany co-residentMTases This differs fromType II enzymeswhere the MTase and REase must co-evolve to allow thehost to survive Each Type IV system is compatible withsome suite of Type I-IIIMTases (and thus the R-M systemsas a whole) Methylated or hydroxymethylated bases maynot be recognized at all (EcoCTGmrSD) or the systemmay require one specific base in addition to the modifiedone (McrA McrBC MspJI and PvuRts1I) MTases mod-ifying sites not including that base are then compatible asDcm (CmCWGG) is compatible with the McrA McrBCand Mrr systems in E coli K12Type IV systems that restrict methylated bases in a
weakly specified sequence context confer an additionaladvantage in competition with phages Many phagessuch as have not evolved the nucleotide-substitutionstrategy used by the T-even phages These phagesnormally carry the modification pattern of the mostrecent host if the last host expressed an MTase creatinga susceptible site the Type IV enzyme of the new host willdestroy the invader and limit the infection This may beaccompanied by the death of the individual infectedtherefore protection can be conferred on the sibling popu-lation (111)A further implication of this scenario considers the fate
of a population invaded by phage Phage survival ofrestriction occurs at biologically relevant frequencies(106-102) The survivors of restriction carry the particu-lar methylation pattern of the particular cell and thus areresistant to all restriction systems it might have carried(Types IndashIV) A bacterial population as a whole thenbenefits from mechanisms that diversify the suite ofR-M systems so that such surviving phage do not havefree access to the entire population The extreme variabil-ity of R-M system content in isolates of the same species iscompatible with this idea [see eg (124) REBASEGenomes httptoolsnebcomvinczegenomes] Suchvariability also limits and shapes interstrain genetransfer (115125126)The Raleigh laboratory has built on elegant genetic
work with Type I enzymes in the Murray laboratory(127ndash130) to investigate a locus designated thelsquoImmigration Control locusrsquo or ICR that exemplifiesvariable R-M content Alternative DNA segments con-taining R-M systems are located at the same definedlocation in most E coli chromosomes between the yjiSand yjiA genes The ICR in the non-restricting E coli Cstrain [used in the original definition of the R-M phenom-enon (131)] contains a remnant of a Type I enzyme R geneand is 13 kilobase shorter than the same region in E coliK12 (132) The mechanism of segment replacement is stillunknown The ICR would be an example of the lsquodefenseislandsrsquo analyzed by the Koonin group (133) lsquoDefense
Nucleic Acids Research 2014 Vol 42 No 1 65
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
islandsrsquo contain genes that can defend against phage orother invading DNA these exhibit bioinformaticproperties similar to lsquomobilome islandsrsquo containing mobil-ization genes (transposases for example) However themechanism of mobilization has not been identified forthe ICR
FINAL THOUGHTS
The extreme diversity of R-M systems that recognizeordinary DNA seems likely to be approached by the di-versity of Type IV restriction systems Type IV enzymesare hard to find as most detection methods depend ondevelopment of genetic systems for each taxon or on ser-endipity Those characterized so far mostly stem frominitial genetic investigation of limits on infection trans-formation or transduction Barriers encountered provideleads to the genes involved Bioinformatic analysis hashelped to identify relatives which may be more tractableto biochemical investigation than the example originallyfound This approach has pitfalls the gene encodingMspJI was first thought to code for an enzyme recognizingan unmodified site because it is immediately adjacent toan (inactive it is now thought) cytosine MTase geneProvidentially the first expression host was devoid of sen-sitive sites whereas the first test substrate contained some(51) A combination of biological experiments with bio-informatics and biochemistry will be needed to reveal thefull spectrum of Type IV enzymes that may lurk within thevast universe of unidentified ORFs in bacterial systemsOne might begin with those strains whose genomes carryfew Type II systems Bacillus or Corynebacterium asopposed to Helicobacter or Neisseria [see the Genomessection of REBASE (50)]The role of lsquodefense islandsrsquo and their relation to the
lsquomobilomersquo in bacterial population biology remains to bedetermined If a defense island is similar to a mobilomeisland there should be a mechanism of mobilizationnearby which would boost the contribution oflsquooverreplicationrsquo to the account of selections acting onR-M systems R-M systems of all types can be found onor adjacent to known mobilizing elements (134135) buthave not been shown to move experimentallyOn another note it may turn out that evolutionarily
there is a continuum between the apparently modifica-tion-dependent and modification-blocked paths Onerelative of McrBC predicted by bioinformatics analysesis LlaI a system that recognizes an unmodified sequenceand requires two MTases to support it (136) The enzymeBamHI prefers to cleave DNA with m6A in its GGATCCsite and mutants can be isolated that require this modifiedbase (137) Are there native systems similarly protected bymodification of one position in the recognition site butdependent on modification at a different one An interest-ing evolutionary series can be imagined
ACKNOWLEDGEMENTS
The authors thank Rich Roberts and David Dryden forencouragement and critical comment They are grateful to
Tom Bickle Virginius Siksnys Yu Zheng and XiadongCheng for permission to reproduce figures They alsothank Ivan Correa for help drawing base structures andTasha Jose for illustration assistance They also thank theanonymous reviewers of a previous manuscript for caretaken with the review process greatly improving thepresent document
FUNDING
New England Biolabs (EAR) Funding for open accesscharge New England Biolabs (to EAR)
Conflict of interest statement None declared
REFERENCES
1 LuriaSE and HumanML (1952) A nonhereditary host-inducedvariation of bacterial viruses J Bacteriol 64 557ndash559
2 RevelHR and LuriaSE (1970) DNA-glucosylation in T-evenphage genetic determination and role in phage-host interactionAnnu Rev Genet 4 177ndash192
3 RaleighEA and WilsonG (1986) Escherichia coli K-12 restrictsDNA containing 5-methylcytosine Proc Natl Acad Sci USA83 9070ndash9074
4 Noyer-WeidnerM DiazR and ReinersL (1986) Cytosine-specific DNA modification interferes with plasmid establishmentin Escherichia coli K12 involvement of rglB Mol Gen Genet205 469ndash475
5 BlumenthalRM GregorySA and CooperiderJS (1985)Cloning of a restriction-modification system from Proteus vulgarisand its use in analyzing a methylase-sensitive phenotype inEscherichia coli J Bacteriol 164 501ndash509
6 SohailA LiebM DarM and BhagwatAS (1990) A generequired for very short patch repair in Escherichia coli is adjacentto the DNA cytosine methylase gene J Bacteriol 1724214ndash4221
7 MacNeilDJ (1988) Characterization of a unique methyl-specificrestriction system in Streptomyces avermitilis J Bacteriol 1705607ndash5612
8 VertesAA InuiM KobayashiM KurusuY and YukawaH(1993) Presence of mrr- and mcr-like restriction systems incoryneform bacteria Res Microbiol 144 181ndash185
9 MacalusoA and MettusAM (1991) Efficient transformation ofBacillus thuringiensis requires nonmethylated plasmid DNAJ Bacteriol 173 1353ndash1356
10 HolmesML NuttallSD and Dyall-SmithML (1991)Construction and use of halobacterial shuttle vectors and furtherstudies on Haloferax DNA gyrase J Bacteriol 173 3807ndash3813
11 PosfaiJ BhagwatAS PosfaiG and RobertsRJ (1989)Predictive motifs derived from cytosine methyltransferases NucleicAcids Res 17 2421ndash2435
12 MaloneT BlumenthalRM and ChengX (1995) Structure-guided analysis reveals nine sequence motifs conserved amongDNA amino-methyltransferases and suggests a catalyticmechanism for these enzymes J Mol Biol 253 618ndash632
13 TiminskasA ButkusV and JanulaitisA (1995) Sequence motifcharacteristics for DNA [cytosine-N4] and DNA [adenine-N6]methyltransferases Classification of all DNA methyltransferasesGene 157 3ndash11
14 DickmanMJ InglestonSM SedelnikovaSE RaffertyJBLloydRG GrasbyJA and HornbyDP (2002) The RuvABCresolvasome Eur J Biochem 269 5492ndash5501
15 KovallRA and MatthewsBW (1998) Structural functionaland evolutionary relationships between lambda-exonuclease andthe type II restriction endonucleases Proc Natl Acad Sci USA95 7893ndash7897
16 RobertsRJ ChangYC HuZ RachlinJN AntonBPPokrzywaRM ChoiHP FallerLL GuleriaJ HousmanGet al (2010) COMBREX a project to accelerate the functional
66 Nucleic Acids Research 2014 Vol 42 No 1
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
annotation of prokaryotic genomes Nucleic Acids Res 39D11ndashD14
17 RaePM and SteeleRE (1978) Modified bases in the DNAs ofunicellular eukaryotes an examination of distributions andpossible roles with emphasis on hydroxymethyluracil indinoflagellates Biosystems 10 37ndash53
18 WarrenRAJ (1980) Modified bases in bacteriophage DNAsAnnu Rev Microbiol 34 137ndash158
19 Gommers-AmptJH and BorstP (1995) Hypermodified bases inDNA FASEB J 9 1034ndash1042
20 SwintonD HattmanS CrainPF ChengCS SmithDL andMcCloskeyJA (1983) Purification and characterization of theunusual deoxynucleoside alpha-N-(9-beta-D-20-deoxyribofuranosylpurin-6-yl)glycinamide specified by the phageMu modification function Proc Natl Acad Sci USA 807400ndash7404
21 TahilianiM KohKP ShenY PastorWA BandukwalaHBrudnoY AgarwalS IyerLM LiuDR AravindL et al(2009) Conversion of 5-methylcytosine to 5-hydroxymethylcytosinein mammalian DNA by MLL partner TET1 Science 324930ndash935
22 KriaucionisS and HeintzN (2009) The nuclear DNA base5-hydroxymethylcytosine is present in Purkinje neurons and thebrain Science 324 929ndash930
23 BorstP and SabatiniR (2008) Base J discovery biosynthesisand possible functions Annu RevMicrobiol 62 235ndash251
24 KaminskaKH and BujnickiJM (2008) Bacteriophage MuMom protein responsible for DNA modification is a new memberof the acyltransferase superfamily Cell Cycle 7 120ndash121
25 IyerLM TahilianiM RaoA and AravindL (2009) Predictionof novel families of enzymes involved in oxidative and othercomplex modifications of bases in nucleic acids Cell Cycle 81698ndash1710
26 ZhouX HeX LiangJ LiA XuT KieserT HelmannJDand DengZ (2005) A novel DNA modification by sulphur MolMicrobiol 57 1428ndash1438
27 OuHY HeX ShaoY TaiC RajakumarK and DengZ(2009) dndDB A Database Focused on Phosphorothioation ofthe DNA Backbone PloS One 4 e5132
28 XuT LiangJ ChenS WangL HeX YouD WangZLiA XuZ ZhouX et al (2009) DNA phosphorothioation inStreptomyces lividans mutational analysis of the dnd locusBMC Microbiol 9 41
29 WangL ChenS VerginKL GiovannoniSJ ChanSWDeMottMS TaghizadehK CorderoOX CutlerMTimberlakeS et al (2011) DNA phosphorothioation iswidespread and quantized in bacterial genomes Proc Natl AcadSci USA 108 2963ndash2968
30 Loslashbner-OlesenA SkovgaardO and MarinusMG (2005) Dammethylation coordinating cellular processes Curr OpinMicrobiol 8 154ndash160
31 LowDA and CasadesusJ (2008) Clocks and switches bacterialgene regulation by DNA adenine methylation Curr OpinMicrobiol 11 106ndash112
32 FreitagM and SelkerEU (2005) Controlling DNA methylationMany roads to one modification Curr Opin Genet Dev 15191ndash199
33 ProhaskaSJ StadlerPF and KrakauerDC (2010) Innovationin gene regulation The case of chromatin computation J TheorBiol 265 27ndash44
34 SeshasayeeASN SinghP and KrishnaS (2012) Context-dependent conservation of DNA methyltransferases in bacteriaNucleic Acids Res 40 7066ndash7073
35 AntonBP (2011) Dissertation Boston University BostonMassachusetts USA
36 RaleighEA TrimarchiR and RevelH (1989) Genetic andphysical mapping of the mcrA (rglA) and mcrB (rglB) loci ofEscherichia coli K-12 Genetics 122 279ndash296
37 Gonzalez-CeronG Miranda-OlivaresOJ and Servın-GonzalezL(2009) Characterization of the methyl-specific restriction system ofStreptomyces coelicolor A3(2) and of the role played by laterallyacquired nucleases FEMS Microbiol Lett 301 35ndash43
38 LiuG OuHY WangT LiL TanH ZhouXRajakumarK DengZ and HeX (2010) Cleavage of
phosphorothioated DNA and methylated DNA by the type IVrestriction endonuclease ScoMcrA PLoS Genet 6 e1001253
39 SutherlandE CoeL and RaleighEA (1992) McrBC amultisubunit GTP-dependent restriction endonuclease J MolBiol 225 327ndash348
40 KrugerT WildC and Noyer-WeidnerM (1995) McrB aprokaryotic protein specifically recognizing DNA containingmodified cytosine residues EMBO J 14 2661ndash2669
41 SitaramanR and LepplaSH (2012) Methylation-dependentDNA restriction in Bacillus anthracis Gene 494 44ndash50
42 HeitmanJ and ModelP (1987) Site-specific methylases inducethe SOS DNA repair response in Escherichia coli J Bacteriol169 3243ndash3250
43 Waite-ReesPA KeatingCJ MoranLS SlatkoBEHornstraLJ and BennerJS (1991) Characterisation andexpression of the Escherichia coli Mrr restriction systemJ Bacteriol 173 5207ndash5219
44 KerrAL JeonYJ SvensonCJ RogersPL and NeilanBA(2010) DNA restriction-modification systems in the ethanologenZymomonas mobilis ZM4 Appl MicrobiolBiotech 89 761ndash769
45 XuSY CorvagliaAR ChanSH ZhengY and LinderP(2011) A type IV modification-dependent restriction enzymeSauUSI from Staphylococcus aureus subsp aureus USA300Nucleic Acids Res 39 5597ndash5610
46 MonkIR ShahIM XuM TanMW and FosterTJ (2012)Transforming the untransformable application of directtransformation to manipulate genetically Staphylococcus aureusand Staphylococcus epidermidis mBio 3 e00277ndash00211
47 WangH GuanS QuimbyA Cohen-KarniD PradhanSWilsonG RobertsRJ ZhuZ and ZhengY (2011)Comparative characterization of the PvuRts1I family ofrestriction enzymes and their application in mapping genomic5-hydroxymethylcytosine Nucleic Acids Res 39 9294ndash9305
48 JanosiL YonemitsuH HongH and KajiA (1994) Molecularcloning and expression of a novel hydroxymethylcytosine-specificrestriction enzyme (PvuRts1I) modulated by glucosylation ofDNAJ Mol Biol 242 45ndash61
49 BairCL and BlackLW (2007) A type IV modificationdependent restriction nuclease that targets glucosylatedhydroxymethyl cytosine modified DNAs J Mol Biol 366768ndash778
50 RobertsRJ VinczeT PosfaiJ and MacelisD (2010)REBASEndasha database for DNA restriction and modificationenzymes genes and genomes Nucleic Acids Res 38 D234ndashD236
51 ZhengY Cohen-KarniD XuD ChinHG WilsonGPradhanS and RobertsRJ (2010) A unique family of Mrr-likemodification-dependent restriction endonucleases Nucleic AcidsRes 38 5527ndash5534
52 Cohen-KarniD XuD AponeL FomenkovA SunZDavisPJ KinneySRM Yamada-MabuchiM XuSYDavisT et al (2011) The MspJI family of modification-dependent restriction endonucleases for epigenetic studies ProcNatl Acad Sci USA 108 11040ndash11045
53 LubysA VitkuteJ LubieneJ and JanulaitisA (2011)Restriction enzymes and their applications US Patent ApplicationPublication 20110207139 A1 httpappftusptogovnetahtmlPTOsearch-boolhtml (27 August 2013 date last accessed)
54 TarasovaGV NayakshinaTN and DegtyarevSK (2008)Substrate specificity of new methyl-directed DNA endonucleaseGlaI BMC Mol Biol 9 7
55 MulliganEA and DunnJJ (2008) Cloning purification andinitial characterization of E coli McrA a putative5-methylcytosine-specific nuclease Protein ExprPurif 62 98ndash103
56 MulliganEA HatchwellE McCorkleSR and DunnJJ (2010)Differential binding of Escherichia coli McrA protein to DNAsequences that contain the dinucleotide m5CpG Nucleic AcidsRes 38 1997ndash2005
57 AntonBP and RaleighEA (2004) Transposon-mediated linkerinsertion scanning mutagenesis of the Escherichia coli McrAendonuclease J Bacteriol 186 5699ndash5707
58 BujnickiJM RadlinskaM and RychlewskiL (2000) Atomicmodel of the 5-methylcytosine-specific restriction enzyme McrAreveals an atypical zinc finger and structural similarity tobetabetaalphaMe endonucleases Mol Microbiol 37 1280ndash1281
Nucleic Acids Research 2014 Vol 42 No 1 67
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
59 GastFU BrinkmannT PieperU KrugerT NoyerWeidnerM and PingoudA (1997) The recognition of methylatedDNA by the GTP-dependent restriction endonuclease McrBCresides in the N-terminal domain of McrB Biol Chem 378975ndash982
60 SukackaiteR GrazulisS TamulaitisG and SiksnysV (2012)The recognition domain of the methyl-specific endonucleaseMcrBC flips out 5-methylcytosine Nucleic Acids Res 407552ndash7562
61 DilaD SutherlandE MoranL SlatkoB and RaleighEA(1990) Genetic and sequence organization of the mcrBC locus ofEscherichia coli K-12 J Bacteriol 172 4888ndash4900
62 PieperU BrinkmannT KrugerT Noyer WeidnerM andPingoudA (1997) Characterization of the interaction between therestriction endonuclease McrBC from E coli and its cofactorGTP J Mol Biol 272 190ndash199
63 PieperU SchweitzerT GrollDH GastFU and PingoudA(1999) The GTP-binding domain of McrB more than just avariation on a common theme J Mol Biol 292 547ndash556
64 BujnickiJM and RychlewskiL (2001) Grouping together highlydiverged PD-(DE)XK nucleases and identification of novelsuperfamily members using structure-guided alignment ofsequence profiles J Mol MicrobiolBiotech 3 69ndash72
65 PieperU and PingoudA (2002) A mutational analysis of thePD DEXK motif suggests that McrC harbors the catalyticcenter for DNA cleavage by the GTP-dependent restrictionenzyme McrBC from Escherichia coli Biochemistry 415236ndash5244
66 KelleherJE and RaleighEA (1991) A novel activity inEscherichia coli K-12 that directs restriction of DNA modified atCG dinucleotides J Bacteriol 173 5220ndash5223
67 OrlowskiJ MebrhatuMT MichielsCW BujnickiJM andAertsenA (2008) Mutational analysis and a structural model ofmethyl-directed restriction enzyme Mrr Biochem Biophys ResCommun 377 862ndash866
68 AravindL MakarovaKS and KooninEV (2000) SURVEYAND SUMMARY holliday junction resolvases and relatednucleases identification of new families phyletic distribution andevolutionary trajectories Nucleic Acids Res 28 3417ndash3432
69 BujnickiJM and RychlewskiL (2001) Identification of a PD-(DE)XK-like domain with a novel configuration of the endonucleaseactive site in the methyl-directed restriction enzyme Mrr and itshomologs Gene 267 183ndash191
70 SteczkiewiczK MuszewskaA KnizewskiL RychlewskiL andGinalskiK (2012) Sequence structure and functional diversity ofPD-(DE)XK phosphodiesterase superfamily Nucleic Acids Res40 7016ndash7045
71 BairCL RifatD and BlackLW (2007) Exclusion of glucosyl-hydroxymethylcytosine DNA containing bacteriophages isovercome by the injected protein inhibitor IPI J Mol Biol 366779ndash789
72 StewartFJ and RaleighEA (1998) Dependence of McrBCcleavage on distance between recognition elements Biol Chem379 611ndash616
73 PanneD RaleighEA and BickleTA (1999) The McrBCendonuclease translocates DNA in a reaction dependent on GTPhydrolysis J Mol Biol 290 49ndash60
74 PieperU GrollDH WunschS GastFU SpeckC MuckeNand PingoudA (2002) The GTP-dependent restriction enzymeMcrBC from Escherichia coli forms high-molecular masscomplexes with DNA and produces a cleavage patternwith a characteristic 10-base pair repeat Biochemistry 415245ndash5254
75 IshikawaK HandaN SearsL RaleighEA and KobayashiI(2011) Cleavage of a model DNA replication fork by amethyl-specific endonuclease Nucleic Acids Res 39 5489ndash5498
76 StewartFJ PanneD BickleTA and RaleighEA (2000)Methyl-specific DNA binding by McrBC a modification-dependent restriction enzyme J Mol Biol 298 611ndash622
77 LacksS and GreenbergB (1977) Complementary specificity ofrestriction endonucleases DpnI and DpnII with respect to DNAmethylation J Mol Biol 114 153ndash168
78 SiwekW CzapinskaH BochtlerM BujnickiJM andSkowronekK (2012) Crystal structure and mechanism
of action of the N6-methyladenine-dependent type IIMrestriction endonuclease RDpnI Nucleic Acids Res 407563ndash7572
79 LacksS and GreenbergB (1975) A deoxyribonuclease ofDiplococcus pneumoniae specific for methylated DNA J BiolChem 250 4060ndash4066
80 HortonJR MabuchiMY Cohen-KarniD ZhangXGriggsRM SamaranayakeM RobertsRJ ZhengY andChengX (2012) Structure and cleavage activity of the tetramericMspJI DNA modification-dependent restriction endonucleaseNucleic Acids Res 40 9763ndash9773
81 CymermanIA ObarskaA SkowronekKJ LubysA andBujnickiJM (2006) Identification of a new subfamily of HNHnucleases and experimental characterization of a representativemember HphI restriction endonuclease Proteins 65 867ndash876
82 GorbalenyaAE (1994) Self-splicing group I and group II intronsencode homologous (putative) DNA endonucleases of a newfamily Protein Sci 3 1117ndash1120
83 ChanSH OpitzL HigginsL OrsquoLoaneD and XuSY (2010)Cofactor requirement of HpyAV restriction endonuclease PloSOne 5 e9071
84 BourniquelAA and BickleTA (2002) Complex restrictionenzymes NTP-driven molecular motors Biochimie 84 1047ndash1059
85 OrsquoDriscollJ HeiterDF WilsonGG FitzgeraldGFRobertsR and van SinderenD (2006) A genetic dissection ofthe LlaJI restriction cassette reveals insights on a novelbacteriophage resistance system BMC Microbiol 6 40
86 SmithRM DiffinFM SaveryNJ JosephsenJ andSzczelkunMD (2009) DNA cleavage and methylation specificityof the single polypeptide restriction-modification enzyme LlaGINucleic Acids Res 37 7206ndash7218
87 InterthalH PouliotJJ and ChampouxJJ (2001) The tyrosyl-DNA phosphodiesterase Tdp1 is a member of the phospholipaseD superfamily Proc Natl Acad Sci USA 98 12009ndash12014
88 SasnauskasG HalfordSE and SiksnysV (2003) How the BfiIrestriction enzyme uses one active site to cut two DNA strandsProc Natl Acad Sci USA 100 6410ndash6415
89 SasnauskasG ConnollyBA HalfordSE and SiksnysV (2007)Site-specific DNA transesterification catalyzed by a restrictionenzyme Proc Natl Acad Sci USA 104 2115ndash2120
90 Marchler-BauerA ZhengC ChitsazF DerbyshireMKGeerLY GeerRC GonzalesNR GwadzM HurwitzDILanczyckiCJ et al (2012) CDD conserved domains andprotein three-dimensional structure Nucleic Acids Res 41D348ndashD352
91 KravetzAN ZakharovaMV BeletskayaIV SinevaEVDenjmuchametovMM PetrovSI GlatmanLI andSoloninAS (1993) The cleavage sites and localization of genesencoding the restriction endonucleases Eco1831I and EcoHIGene 129 153ndash154
92 KlimasauskasS KumarS RobertsRJ and ChengX (1994)HhaI methyltransferase flips its target base out of the DNA helixCell 76 357ndash369
93 HashimotoH VertinoPM and ChengX (2010) Molecularcoupling of DNA methylation and histone methylationEpigenomics 2 657ndash669
94 RajakumaraE LawJA SimanshuDK VoigtPJohnsonLM ReinbergD PatelDJ and JacobsenSE (2011)A dual flip-out mechanism for 5mC recognition by theArabidopsis SUVH5 SRA domain and its impact on DNAmethylation and H3K9 dimethylation in vivo Genes Dev 25137ndash152
95 SharifJ and KosekiH (2011) Recruitment of Dnmt1 roles of theSRA protein Np95 (Uhrfi) and other factors Prog Mol BiolTransl Sci 101 289ndash310
96 Tesfazgi MebrhatuM WywialE GhoshA MichielsCWLindnerAB TaddeiF BujnickiJM Van MelderenL andAertsenA (2011) Evidence for an evolutionary antagonismbetween Mrr and Type III modification systems Nucleic AcidsRes 39 5991ndash6001
97 GrantGN JesseeJ BloomFR and HanahanD (1990)Differential plasmid rescue from transgenic mouse DNAs intoEscherichia coli methylation-restriction mutants Proc Natl AcadSci USA 87 4645ndash4649
68 Nucleic Acids Research 2014 Vol 42 No 1
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
98 AertsenA and MichielsCW (2005) Mrr instigates the SOSresponse after high pressure stress in Escherichia coli MolMicrobiol 58 1381ndash1391
99 WeiH TherrienC BlanchardA GuanS and ZhuZ (2008)The Fidelity Index provides a systematic quantitation of staractivity of DNA restriction endonucleases Nucleic Acids Res36 e50
100 LuL PatelH and BisslerJJ (2002) Optimizing DpnIdigestion to detect replicated DNA Biotechniques 33 316ndash318
101 MadejT AddessKJ FongJH GeerLY GeerRCLanczyckiCJ LiuC LuS Marchler-BauerAPanchenkoAR et al (2011) MMDB 3D structures andmacromolecular interactions Nucleic Acids Res 40 D461ndashD464
102 GibratJF MadejT and BryantSH (1996) Surprisingsimilarities in structure comparison Curr Opin Struct Biol 6377ndash385
103 PanneD MullerSA WirtzS EngelA and BickleTA (2001)The McrBC restriction endonuclease assembles into a ringstructure in the presence of G nucleotides EMBO J 203210ndash3217
104 NeuwaldAF AravindL SpougeJL and KooninEV (1999)AAA+ A class of chaperone-like ATPases associated with theassembly operation and disassembly of protein complexesGenome Res 9 27ndash43
105 OguraT and WilkinsonAJ (2001) AAA+ superfamilyATPases common structurendashdiverse function Genes Cells 6575ndash597
106 BearyTP BraymerHD and AchbergerEC (1997) Evidenceof participation of McrB(S) in McrBC restriction in Escherichiacoli K-12 J Bacteriol 179 7768ndash7775
107 PanneD RaleighEA and BickleTA (1998) McrBs amodulator peptide for McrBC activity EMBO J 175477ndash5483
108 PuntaM CoggillPC EberhardtRY MistryJ TateJBoursnellC PangN ForslundK CericG ClementsJ et al(2011) The Pfam protein families database Nucleic Acids Res40 D290ndashD301
109 NaitoT KusanoK and KobayashiI (1995) Selfish behavior ofrestriction-modification systems Science 267 897ndash899
110 KobayashiI (2001) Behavior of restriction-modification systemsas selfish mobile elements and their impact on genome evolutionNucleic Acids Res 29 3742ndash3756
111 FukudaE KaminskaKH BujnickiJM and KobayashiI(2008) Cell death upon epigenetic genome methylation a novelfunction of methyl-specific deoxyribonucleases Genome Biol 9R163
112 IshikawaK FukudaE and KobayashiI (2010) Conflictstargeting epigenetic systems and their resolution by cell deathnovel concepts for methyl-specific and other restriction systemsDNA Res 17 325ndash342
113 FukuyoM SasakiA and KobayashiI (2012) Success of asuicidal defense strategy against infection in a structured habitatSci Rep 2 238
114 MakovetsS TitheradgeAJ and MurrayNE (1998) ClpXand ClpP are essential for the efficient acquisition of genesspecifying type IA and IB restriction systems Mol Microbiol28 25ndash35
115 MurrayNE (2002) Immigration control of DNA in bacteriaself versus non-self Microbiology 148 3ndash20
116 MakovetsS PowellLM TitheradgeAJB BlakelyGW andMurrayNE (2004) Is modification sufficient to protect abacterial chromosome from a resident restriction endonucleaseMol Microbiol 51 135ndash147
117 OrgelLE and CrickFHC (1980) Selfish DNA the ultimateparasite Nature 284 604ndash607
118 KidwellMG and LischDR (2001) Perspective transposableelements parasitic DNA and genome evolution Evolution 551ndash24
119 SadykovM AsamiY NikiH HandaN ItayaMTanokuraM and KobayashiI (2003) Multiplication of arestriction-modification gene complex Mol Microbiol 48417ndash427
120 ChaoL and LevinBR (1981) Structured habitats and theevolution of anticompetitor toxins in bacteria Proc Natl AcadSci USA 78 6324ndash6328
121 MagnusonRD (2007) Hypothetical functions of toxin-antitoxinsystems J Bacteriol 189 6089ndash6092
122 RifatD WrightNT VarneyKM WeberDJ and BlackLW(2008) Restriction endonuclease inhibitor IPI of bacteriophageT4 a novel structure for a dedicated target J Mol Biol 375720ndash734
123 LabrieSJ SamsonJE and MoineauS (2010) Bacteriophageresistance mechanisms Nat Rev Microbiol 8 317ndash327
124 LinLF PosfaiJ RobertsRJ and KongH (2001)Comparative genomics of the restriction-modification systemsin Helicobacter pylori Proc Natl Acad Sci USA 982740ndash2745
125 MilkmanR RaleighEA McKaneM CrydermanDBilodeauP and McWeenyK (1999) Molecular evolution of theEscherichia coli chromosome V Recombination patterns amongstrains of diverse origin Genetics 153 539ndash554
126 HumbertO DorerMS and SalamaNR (2011)Characterization of Helicobacter pylori factors that controltransformation frequency and integration length during inter-strain DNA recombination Mol Microbiol 79 387ndash401
127 BarcusVA TitheradgeAJB and MurrayNE (1995) Thediversity of alleles at the hsd locus in natural populations ofEscherichia coli Genetics 140 1187ndash1197
128 SharpPM KelleherJE DanielAS CowanGM andMurrayNE (1992) Roles of selection and recombination in theevolution of type I restriction-modification systems inenterobacteria Proc Natl Acad Sci USA 89 9836ndash9840
129 LoenenWAM DanielAS BraymerHD and MurrayNE(1987) Organization and sequence of the hsd genes ofEscherichia coli K-12 J Mol Biol 198 159ndash170
130 SuriB and BickleTA (1985) EcoA the first member of a newfamily of type I restriction modification systems Geneorganization and enzymatic activities J Mol Biol 186 77ndash85
131 BertaniG and WeigleJJ (1953) Host-controlled variation inbacterial viruses J Bacteriol 65 113ndash121
132 SibleyMH and RaleighEA (2004) Cassette-like variation ofrestriction enzyme genes in Escherichia coli C and relativesNucleic Acids Res 32 522ndash534
133 MakarovaKS WolfYI SnirS and KooninEV (2011)Defense islands in bacterial and archaeal genomes and predictionof novel defense systems J Bacteriol 193 6039ndash6056
134 FurutaY AbeK and KobayashiI (2010) Genome comparisonand context analysis reveals putative mobile forms of restriction-modification systems and related rearrangements Nucleic AcidsRes 38 2428ndash2443
135 FurutaY and KobayashiI (2012) Restriction-modificationsystems as moblie epigenetic elements In RobertsAP andMullanyP (eds) Bacterial Integrative Mobile Genetic ElementsLandes Bioscience pp 85ndash103
136 OrsquoSullivanDJ and KlaenhammerTR (1998) Control ofexpression of LlaI restriction in Lactococcus lactis MolMicrobiol 27 1009ndash1020
137 WhitakerRD DornerLF and SchildkrautI (1999) A mutantof BamHI restriction endonuclease which requires N6-methyladenine for cleavage J Mol Biol 285 1525ndash1536
Nucleic Acids Research 2014 Vol 42 No 1 69
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
Other Mrr-related enzymes from Bacillus anthracisStreptomyces coelicolor and Zymomonas mobilis(identified bioinformatically see later in the text) werealso tested for activity in vivo Transformation efficiencyis reduced when a plasmid is prepared from a modifyinghost compared with the same plasmid from a non-mod-ifying host this reduction is alleviated when the corres-ponding Mrr-related gene is disrupted The specificity ofthis test depends on how thorough the genetic investiga-tion was if Dam Dcm EcoKM+ DNA transformsbetter than fully modified DNA modification specificitycould be either m6A or m5C or both hence the questionmarks in the table
The four systems listed for S coelicolor 3A constitute aparticularly exemplary analysis of this kind (37) In thiscase all four candidate R-M systems were deleted indi-vidually and together so that the effect of each could betested and each system was established in the related non-restricting host Streptomyces lividans For ScoA3Mrr theeffect of removing modifiable sites from the test plasmidwas also examined (for MEcoKI)
Diversity of functional organizationUnlike the classic Type IIP enzymes such as EcoRI andBamHI in which catalytic residues are embedded withinsequence-recognition structural elements the modifica-tion-dependent enzymes known so far exhibit separationof DNA binding and cleavage into different domains onthe same protein or even into different polypeptide chains(Table 3 and 4) In this they resemble Type I Type IIS orType III enzymes modification-blocked enzymes that alsoseparate recognition and cleavage For those also
Table 1 DNA modificationsa that elicit cleavage by Type IV enzymes
Protein m5C hm5C ghm5C m4C m6A PT References
EcoKMcrA (+) (+) () NT () NT (336)ScoA3McrA + NT NT NT (+) + (3738) Sco4631EcoKMcrBC + + (+) NT (33940)BanUMcrB (+) (41)BanUMcrB3 (+) (41)EcoKMrr (+) () () () (+) NT (4243)BanUMrr (+) (+) (41)ScoA3Mrr () (+) (37) Sco4213ZmoMrr (+) (+) (44) ZMO1932 ZmrrSauUSI + + NT (45)SauNewI (+) (46) NWMN_2386SepRPMcrR (+) (46) SERP2052ScoA3I (+) (37) Sco2863PvuRts1I family + + + NT NT (4748)GmrSD + NT NT (49)ScoA3II+III (+) () (37) Sco3261-62
aModifications m5C 5-methylcytosine hm5C 5-hydroxymethylcytosine ghm5C glucosylated hydroxymethylcytosine m4C N4-methylcytosinem6A N6-methyladenine PT phosphorothioation of non-bridging oxygen in DNA linkages also called S-DNA+ at least 100-fold less activity on this substrate than on substrates with+entry() (+) based on in vivo restriction of phage infection or plasmid transformation with appropriate host mutant configurations in vitro cleavageresults have not been reported(+) either m5C or m6A is recognized these were not distinguished in the reported experiments m6A sites tested were not cleaved but few modified sequences were testedNT not testedWhere the name found in REBASE (and listed at the left) is not the same as that used in the cited report the genomic locus_ID is given in theReferences column or the name used in the publication
Table 2 DNA modificationsa that elicit cleavage by other
modification-dependent enzymes (Type IIM)
Protein m5C hm5C ghm5C m4C m6A PT References
DpnI + NT (50)MspJI family + + NT (5152)SgeI + (53)AoxI +BisI +BlsI +GlaI + (54)GluI +KroI +MalI +MteI +PcsI +
aModifications m5C 5-methylcytosine hm5C 5-hydroxy-methylcytosine ghm5C glucosylated hydroxymethylcytosinem4C N4-methylcytosine m6A N6-methyladenine PT phosphoro-thioation of non-bridging oxygen in DNA linkages also called S-DNA+ at least 100-fold less activity on this substrate than on substrateswith+entry() (+) based on in vivo restriction of phage infection or plasmidtransformation with appropriate host mutant configurations in vitrocleavage results have not been reported(+) either m5C or m6A is recognized these were not distinguished inthe reported experiments m6A sites tested were not cleaved but few modified sequences weretestedNT not testedWhere the name found in REBASE (and listed at the left) isnot the same as that used in the cited report the genomiclocus_ID is given in the References column or the name used in thepublication
58 Nucleic Acids Research 2014 Vol 42 No 1
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
multiple evolutionary events apparently have occurred toconnect nuclease domains to recognition moieties (81)
Nuclease domainsEnzymes that recognize modified DNA with minimalsequence selectivity have emerged at least six times asexemplified by the McrA McrBC SauUSI MrrPvuRts1I and GmrSD families These exemplars are dis-cussed in more detail later in the text In brief nucleasedomains have been attached covalently or (for McrC) viaproteinndashprotein interaction to domains with DNA bindingand regulatory functions
EcoKMcrA carries a C-terminal H-N-Hc nucleasedomain identified bioinformatically (5882) (Figure 2)This nuclease domain is also found in modification-blocked nucleases (81) The purified binding-competentprotein did not cleave under a variety of buffer conditionsand cofactor additions (55) ScoA3McrA is designatedlsquoMcrArsquo due to its possession of a similar nucleasedomain For this enzyme cleavage depends on Mn2+ orCo2+ (38) and occurs at a variable distance from PT-modified sites Modification-blocked H-N-H REases alsooften exhibit unusual metal ion requirements [eg (83)]
McrBC The required McrC component (3940) is thenuclease moiety (65) (Figure 3) Mutational analysis
confirms that it is a PD-(DE)XK nuclease (65) in agree-ment with bioinformatic classification (64) Cleavageresults when McrC associates with full-lengthMcrBGTP complex bound to DNA and GTP ishydrolyzed (72) LlaJI a modification-blocked restrictionactivity exhibits a similar organization (85) althoughcleavage could not be demonstrated in vitroThe classic modification-dependent enzyme DpnI also
carries a PD-(DE)XK motif (see further later in the text)Mrr EcoKMrr contains a variant of the PD-(DE)XK
motif (6869) with the Mrr-N (E coli K12) presumedDNA-binding domain MspJI (see further later in thetext) also carries a nuclease domain in this family Aswith McrA McrBC and SauUSI nuclease domain simi-larity does not in itself dictate modification preferenceproperties the single-chain R-M system LlaGI hasconserved motifs characteristic of the E coli Mrrprotein but this enzyme does not target methylatedDNA (86)SauUSI This is a modification-dependent enzyme with
a phosphodiesterase cleavage domain akin to one origin-ally identified in phospholipase-D (45) Mutation of anyof the four conserved catalytic residues abolishes in vitroactivity This cleavage domain is also found in stand-alone
Table 3 Characteristics of Type IV restriction enzymes
Protein Subunitsdomains
DNABinding
Endonucleasedomain
NTP hydrolysis Recognition sitea comment and references
EcoKMcrA ndash In vitro cleavage not reportedN-terminal DBD (YgtR)m5CGR bound (5556) hm5Cm5C
discrimination (57)C-terminal H-N-Hc Bioinformatic ID (58) required for damage to DNA
in vivo (57)ScoA3McrA ndash Some CmCWGG and some S-DNA (PT modified)
sites are cleaved (38)N-terminal DBD Not related to EcoMcrA (38)C-terminal H-N-Hc 37 identical to EcoMcrA (38)
EcoKMcrBC GTP Rm5C(N30-35j)-(N30-3000)-Rm5Cb
McrB-N DUF3578 McrB binds DNA (40) via its N-terminal domain(59) by extrahelical modified base (60)
McrB-C P-loop NTPase (616263)McrC PD-(DE)XK Bioinformatic ID (64) Required for cleavage (3965)
EcoKMrr ndash m6A or m5C sequence specificity ambiguous(424366)
N-terminal Mrr-N Presumed DNA binding (67)C-terminal Mrr-cat (DE)
(DEQ) KBioinformatic ID (686970)
SauUS1 ATP or dATP Sm5CNGS two copies required for cleavage (45)N-terminal PLDc-2 (45)Middle P-loop NTPase (45)C-terminal DUF3427 (45) DBD
PvuRts1I PD-(DE)XK ndash mC(N11-13N9-10j)G 2-base extensions (47)Bioinformatic ID (64)
EcoCTGmrSD UTPgtgtGTP CTP Cuts T-even DNA (49)GmrS Motifs suggested DUF262 (71)To be confirmedGmrD DUF1524 To be determined
aRecognition sites (represented 5030) are those determined in vitro by binding or cleavage experimentsbMcrBC cleavage results in a double-strand cut near one Rm5C site (727374) but requires cooperation of two sites (3940) or a translocation block(73) The sites may be on different daughters across a fork (75) These are separated by 30ndash3000 (397274) and may be on either strand (3976)disposition of opposing nicks is not tightly constrained (73) and minor cleavage clusters are found 40 50 and 60 nt from the m5C (74)Degeneracy abbreviations B=C or G or T D=A or G or T H=A or C or T K=G or T M=A or C N=A or C or G or T R=A or GS=C or G V=A or C or G W=A or T Y=C or TCleavage positions are listed as (N to top cut to bottom cutj) If no number is listed the position of cleavage is not determined Space betweennumbers (eg PvuRts1I N11-13N9-10) indicates the range of positions at which cleavage may occur
Nucleic Acids Research 2014 Vol 42 No 1 59
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
nucleases and modification-blocked REases (8788)Interestingly two of the PLDc nuclease activities havebeen shown to work by a transesterification reaction likethat used by topoisomerases and transposases (8789)PvuRts1I has an apparently unusual nuclease domain
[ie not yet identified by sources curated by theNCBI Conserved Domain Database (90)] However thisenzyme was included in a categorization of PD-(DE)XKfamilies (64) a tentatively identified divalent metalion binding site Block B (47) corresponds to Block Dof Bujnicki and Rychlewski (64) Cleavage requiresMg2+ ions
EcoCTGmrSD Functional organization is less clearbut several possible nuclease motifs were identified inGmrS (71) Cleavage buffer contained Ca2+ and Mg2+
ions and UTP
Sequence context recognition
Many of the modification-dependent enzymescharacterized so far have little sequence specificity incontrast to conventional modification-blocked REasesRelatively complete characterization of sequence prefer-ence and cleavage position has been carried out for TypeIV enzymes EcoKMcrBC SauUSI and PvuRts1I(Table 3) and for Type IIM DpnI and the MspJI family(Table 4) Progress has been made with binding recogni-tion for EcoKMcrA Cleavage conditions have beenachieved for Sco3AMcrA (Table 3) For all of these rec-ognition of surrounding sequence context is degeneratewith preference for a neighboring base and frequently arequirement for two sites with suitable separation DpnI isin some respects an exception see later in the text
The remaining nucleases in Table 4 are less wellcharacterized The recognition sites might form a relatedseries It will be interesting to learn more about the rela-tionships among these and how the requirement formultiple modified positions is specified eg for BlsI andPkrI
McrA binding domainsThe two lsquoMcrArsquo enzymes are not similar in theirN-termini with homology limited to the C-terminalnuclease domain For EcoKMcrA (Figure 2) there isgood genetic evidence that base recognition lies in theN-terminus Extensive mutagenesis using insertion of
Table 4 Characteristics of other modification-dependent enzymes (Type IIM)
Protein Subunitsdomains
DNABinding
EndonucleaseDomain
Recognitionsite
Comment and references
DpnI family G m6AjTC 13 characterized isoschizomersN-terminal PD(DE)XK (50777879)C-terminal wHTH R m6AjTC (78)
MspJI family mC with preferences Second copy stimulates cleavageMspJI 5 mCNNR(N913j) (51)
N-terminal SRA-like (80)C-terminal Mrr-cat (DE)(DEQ)XK (51)
FspEI Like MspJI C m5C(N1216j) (52)LpnPI Like MspJI C m5CDG(N1014j) (52)AspBHI Like MspJI YS m5CNS(N812j) (52)RlaI Like MspJI V m5CW (52)SgrTI Like MspJI C m5CDS(N1014j) (52)SgeI Like MspJI m5CNNR(N913j) 49 identical to MspJI (53)No familyassigned
Information from httpwwwsibenzymecomproductsm2_type
AoxI jRG m5CYBisI G m5CjNGCBlsI RYNjR Y At least two m5C requiredGlaI R m5CjGY (54)GluI GmCjNG m5CKroI Gj m5CCGGCMteI GmCG m5CjNGm5CGm5CPcsI m5CG(N5jN2)m5CGPkrI Gm5CNjG m5C At least 3 m5C required
Figure 2 McrA functional domains Domain function was inferred in-directly from genetic analysis by Anton amp Raleigh (2004) (57) Manymutations in the N-terminal domain spared some activity in one ormore of three functional tests (grey segments) while others were defi-cient in all activities (black segments) One mutation (asterix) was fullyactive on m5C-containing substrates but fully inactive in the hm5Cchallenge in vivo Most mutations in the C-terminus (pale greysegment) retained function in one test that was interpreted asmeasuring m5C binding ability A predicted structural model byBujnicki Radlinska and Rychlewski (2000) (58) for this C-terminalregion is compatible with these results
60 Nucleic Acids Research 2014 Vol 42 No 1
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
five-amino acid linkers and classification with three func-tional tests allowed assignment of DNA recognition to theN-terminal portion with the C-terminal H-N-H domainimplicated in cleavage Of particular note a mutationdiscriminated in vivo between hm5C and m5C was foundin the N-terminal domain (57) The mutant was able tofully restrict bacteriophage lambda modified by MHpaIIbut not at all phage T4 containing hm5C In vitromodification-dependent binding was achieved with thefull-length His-tagged protein (5556) yielding a putativerecognition site (YgtR)mCG This recognition site iscompatible with in vivo observations (391)
Presumably the N-terminus of Sco3A McrA also rec-ognizes the DNA Recognition of both m5C and thephosphorothioate (PT) moiety must be accommodatedin the final reaction As either modification is sufficientto elicit cleavage more than one domain could beinvolved Cleavage occurred near some but not allDcm-modified sites (Cm5CWGG) Both synthetic
PT-containing oligonucleotides and unmethylated PT-modified plasmid were also cleaved on both sides of asymmetrically modified site PT modification is thoughtto be sequence-specific (2629) but the details are notyet clear
Novel McrB binding domainDNA-binding resides in the McrB N-terminal domain(405960) whereas translocation and cleavage coordin-ation reside in the C-terminal AAA+ regulatory andtranslocation domain (61ndash637273) The complete trans-lation product of mcrB McrB-L binds DNA specifically(4072) via its N-terminal 161 amino acids (aa) (59) Thecrystal structure of this domain (McrB-N) in complex withDNA has recently been published (60)McrB-N uses a strategy first discovered for DNA-MTase
action (92) it flips the C base out of the DNA helix into abinding pocket for inspection The pocket is large enoughto accommodate C m5C hm5C or m4C but too small if a
Figure 3 McrBC Assembly Model Two proteins are expressed from mcrB in vivo Both the complete protein (McrB-L) and a small one missing theN-terminus (McrB-S top row) bind GTP forming high-order multimers detected by gel filtration (second row) When visualized by scanningtransmission electron microscopy these appear as heptameric rings with a central channel Rings of McrB-L in top views show projections thatmay correspond to the N-terminal DNA-binding domain (red segment) Both forms can then associate with McrC judged again by gel filtrationMcrB-LGTP can bind to its specific substrate (RmC) in the absence of McrC (third row) in its presence the substrate is cleaved (fourth row)GTP hydrolysis is required for cleavage (arrow) a supershifted binding complex forms in the presence of GTP-gamma-S but no cleavage occursTranslocation accompanies GTP hydrolysis double-stranded cleavage requires collaboration between two complexes or a translocationblock The path of the DNA in the figure is arbitrary as is the conformation of McrC Modified from BourniquelAA and BickleTAComplex restriction enzymes NTP-driven molecular motors Bourniquel and Bickle (84) with permission Copyright 2002 Elsevier MassonSAS All rights reserved
Nucleic Acids Research 2014 Vol 42 No 1 61
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
glucose moiety is attached Conserved residues Y64 andL68 were noted to make van der Waals contact with themethyl group of the flipped out m5C these contacts aremissing when the pocket contains CThe flipping action can be compared with but is distinct
from that of eukaryotic m5C-specific regulatory proteinsthat use the SET and Ring-finger-Associated (SRA)domain to read DNA modification state (Figure 4A)This domain is found in most eukaryotes in accessoryproteins (eg UHRF1NP95SUVH5) of the DNMT1maintenance MTase (93ndash95) Despite the similarstrategy the McrB-N domain is not homologous butdisplays a distinct protein fold (60) Binding is accom-plished from the minor groove and extraction of the Ccreates a 30 bend toward the major groove resembling aglycosylase in this respect (Figure 4A) The eukaryoticproteins form a crescent from which loops project towrap around the DNA with recognition mediatedthrough both major and minor grooves (94) For McrB-N the authors suggest that the purine preference in the 50
position might result from flexibility constraints or inter-action with a non-conserved aa that occupies space left bythe flipped base Substitutions of this aa (Y41A or Y41Q)compromised binding activity
Sequence specificity novel phenotype and structuralmodel of MrrIn 1987 Heitman and Model discovered Mrr when theyfound that transfer of various foreign m6A MTasesinduced an SOS response due to DNA damage (42)This response to the presence of an incompatible MTaseremains the principal evidence that the E coli K12 Mrrprotein cleaves DNA Related proteins discussed later inthe text (Type IIM) have been more tractable for in vitrowork No concise description of the Mrr recognitionsequence has been forthcoming although several studieshave examined the spectrum of incompatible MTases(43669697) Both adenine and cytosine MTases confersensitivityMrr is also responsible for DNA damage that does not
depend on methylation at all foreign or otherwise Highhydrostatic pressure (HP) induces the SOS DNA damageresponse and lethality (98) The response did not dependon the activity of the endogenous MTases of E coli K12but did depend on both the presence of wild-type mrr andthe integrity of the SOS signal generation pathwayPossibly HP elicits a non-enzymatic modification or astructural change in DNA helicity that is acted on byMrr This HP phenotype was used to characterize mrrmutants which were fitted into a computer-assistedmodel of the Mrr protein (67) An N-terminal DNA-binding winged helix was proposed with a C-terminalnuclease domain previously identified (69) The functionalimportance of several conserved residues was confirmedSeveral of the selected mutants with null phenotypes wereisolated in a region far from the active site or bindingsurface identified bioinformatically These could affectinteraction with a component of the HP response Thisintriguing collection of informative mutants will illumin-ate in vitro characterization
Type IIM binding domainsType IIM enzymes of two families are well-characterizedwith respect to cleavage (Table 4) Crystal structures forboth have recently appeared
DpnI winged-helix DNA recognitionUnusually for modification-dependent enzymes DpnIcleaves a four-base site (Gm6ATC) with high fidelity(7799) to leave blunt ends when both strands of the siteare methylated At low concentration the enzyme nicksthe modified strand of a hemimethylated site (100) Thebehavior of the enzyme with respect to modificationpatterns within the canonical GATC sitemdashmodificationof C or A one strand or bothmdashhas been thoroughlyexplored (50) However only recently has cleavage ofnon-canonical adenine-methylated sites been examinedSiwek and co-workers (78) found evidence for consider-able relaxation of specificity at the outer base This experi-ment used substrates modified by a highly non-specificadenine MTase extensive DpnI cleavage cloning of thefragments and sequencing of the borders
Structure determination in the presence of DNA andvalidation experiments (78) place this enzyme togetherwith the other modification-dependent enzymes in thattwo domains segregate the cleavage function fromsequence recognition It also separates DpnI from theothers in that the cleavage domain also possesses somemodification and sequence specificity The main recogni-tion is accomplished by a monomeric winged-helixdomain which binds in the major groove and recognizesthe modifications on both strands in the same event Thestructure does not reveal a cleavage-competent complexhowever because the cleavage domain is far from theDNA Filter-binding experiments validated the ability ofthe C-terminal domain to bind alone to do so moretightly to fully methylated than to hemimethylated oligo-nucleotides and to compete with the full-length enzymereducing cleavage by it Expression of the N-terminalcleavage domain alone (in low yield) allowed validationof its cleavage activity Surprisingly this cleavage wasitself dependent on modification state and sequence ofthe substrate Modeling based on the structure of theblunt-end-producing Type IIP enzyme PvuII allowed pre-diction that the cleavage domain approaches from theminor groove Complete understanding of double-stranded cleavage will depend on understanding thedynamic transformations that allow the cleavage domainto approach and act at the site
MspJI coupling of cleavage with DNA recognitionThe six members of the MspJI family use the Mrr-catversion of the PD-(DE)XK nuclease to cut at definedlocations to one side relative to the modified base(12 bases on the modified strand 16ndash17 on the otherFigure 5A) only one modified base is required fordouble-strand cleavage to occur (unlike McrBC) (5152)However these enzymes are stimulated by the presence ofa second site in cis or in trans Symmetrically modifiedsites (such as m5CpGm5CpG in mammalian DNA)yield prominent bands of defined size (32 bp) containinga mixed population of sequences each with a m5C in the
62 Nucleic Acids Research 2014 Vol 42 No 1
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
middle (Figure 5B) This behavior is recapitulated by thePvuRts1I group of enzymes (exemplified by AbaSDFI inFigure 5C) except that the distances are shorter andrecognition of modification state is less well understoodDuring the characterization of MspJI Dcm
(Cm5CWGG) sites were the first recognized substrateyielding a clear banding pattern (51) Cleavage of differ-ently modified plasmids and designed oligonucleotidesubstrates allowed a good assessment of both modificationand sequence specificity This family shows preference forparticular bases nearby similar to McrBC
MspJI DNA recognition is mediated by anSRA-like domainRecently the crystal structure of MspJI without DNA hasbeen resolved at 205 A (80) Search of the MolecularModeling Database at NCBI (101) using VAST (102)
Figure 4 McrB-N in comparison to other base-flipping proteins (A)SRA domains SUVH5 (3Q0C) and UHRF1 (2ZKD) use loops extend-ing from a crescent formed from two beta sheets to flip C or m5C fromundeformed B-form DNA into a pocket (top row) whereas McrB-N(3SSC bottom row) uses loops from one beta-sheet to distort the DNAand flip the base It resembles the human alkyladenine glycosylase(1BNK) (bottom row) in bending the DNA toward the majorgroove while flipping the base via the minor groove Figure 5 ofSukackaite et al (60) (B) The SRA-like hemi-methylated 5mC recog-nition domains A ribbon model of the N-terminal domain of theMspJI structure (4F0Q and 4F0P left) compared with the SRAdomain of URHF1 (PDB 3FDE right) The crescent shape formed
Figure 4 Continuedby interacting beta sheets and helices aB and aC are the conservedfeatures of the SRA domain highlighted here Loops on the concaveside of UHRF1 participate in flipping the base and similar loops pre-sumably do so for MspJI Two of these vary in length among familymembers and may play roles in sequence context specificity Figure 2aand b from Horton et al (80)
Figure 5 Schematic diagrams of cleavage positions for MspJI andAbaSDFI Cleavage of both strands is elicited by a singly modifiedsite for both MspJI (A) and AbaSDFI (C) Cleavage position is fixedrelative to the modified site but with a four-base 50 extension forMspJI and a two-base 30 extension for AbaSDFI When a site is sym-metrically-modified (as for CpG sites in mammalian DNA) a 32 base-pair fragment is excised from the DNA (B) (A) Figure 2a and (B)Figure 3a reprinted with permission from Cohen-Karni et al (52)(C) Figure 5a from Wang et al (47)
Nucleic Acids Research 2014 Vol 42 No 1 63
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
showed that the N-terminal domain was structurallysimilar to that of the eukaryotic SRA domain with acrescent-shaped beta-sheet structure from which loopsproject (see Figure 4B and discussion earlier in the textMcrB) This structural homology allowed modeling of theDNA-bound structure with a flipped m5C The enzyme inthe crystal is a tetramer in which two monomers form aback-to-back dimer via the C-terminal regions thatcomprise the endonuclease Two back-to-back dimersgenerate a tetrameric protein with two cleavage domainspositioned (as in the Type IIP enzyme HindIII used formodeling the C-terminal cleavage domain interaction withDNA) so that a 4-base 50 extension would be created oncleavage of modeled DNA Cleavage is most efficient atmolar ratios that allow all four SRA-like domains to beoccupiedmdashtoo much enzyme prevents cleavage fromoccurring
Tracking and dimerization
McrBC as translocaseBourniquel and Bickle (84) have reviewed much of theenzymology of McrBC which will be briefly summarizedhere The Raleigh Bickle and Pingoud laboratories havecontributed to the following consistent picture of thein vitro reaction EcoKMcrBC cleavage results in adouble-strand cut near one RmC site (72ndash74) butrequires cooperation of two sites (3940) or a translocationblock (73) The sites may be on different daughters acrossa fork (75) These are separated by 30ndash3000 bp (397274)and may be on either strand (3976) cleavage occurs30ndash35 bases from the modified base with oppositenicks not tightly constrained (73) and minor cleavageclusters are found 40 50 and 60 nt from the m5C(74) hm5C DNA elicits cleavage also (39) A ring struc-ture is formed by 5ndash7 molecules of McrB in the presence ofGTP (Figure 3) (103) this complex can bind to a recog-nition element in DNA In the presence of McrC trans-location of the complex occurs and cleavage ensues whentranslocation is blocked Collision of translocatingcomplexes a protein barrier or a topological barrier willelicit double-strand cleavage adjacent to one recognitionelement or the other The enzyme will cleave when recog-nition elements are on opposite sides of a forked structure(75) This would allow action in vivo to prevent entry of aMTase gene even with rare sitesStructurally the McrB protein is proposed to be a
member of the AAA+ protein family of NTPases (104)many of which form ring-shaped complexes and partici-pate in molecular machines lsquoSensorrsquo segments found inthese proteins have been shown in some cases to playroles in coupling NTPase activity to intersubunit commu-nication and movement (105) Two of three elements ofthe GTP-binding motif proposed by Dila et al (61) werevalidated by mutational analysis (6562) The thirdproposed motif element was identified as amino acidsNTAD by Dila et al Alignment of AAA+ NTPases in(104) found this aligned with the motif designatedSensor-1 in (105) An interesting result was that mutationshere unexpectedly appeared to abrogate interaction withMcrC instead of changing which NTP would be
productive (62) It may instead play a role in coordinatingGTP binding and hydrolysis with DNA binding inter-action with McrC and cleavage
Intracellularly the story becomes more complex as themcrB gene encodes two products of 51 and 33 kD McrB-L and McrB-S the latter one starting from an in-frameinternal translation start site (106) Both in vivo andin vitro McrB-S can interfere with the function ofMcrB-L at least in part by forming complexes withMcrC unable to bind DNA (107) Both species can formmultimeric rings in the presence of GTP (103) as is usualfor AAA+NTPases (104)
SauUSI requires two sites and ATP hydrolysisSauUSI was originally annotated as a putative helicasefrom Staphylococcus aureus sp A single polypeptide issufficient for activity both in vitro and also in vivo as aclone in E coli using modified phage as a challenge Theamino acid sequence contains a PLDc domain at theN-terminus This contains a phosphodiesterase motiforiginally identified in Phospholipase D (108) it wasvalidated by mutagenesis of four catalytic residue candi-dates In the middle ATPase and helicase motifs wereproposed to account for ATP dependence of cleavageactivity A Domain of Unknown Function was identifiedat the C-terminus (Pfam DUF3427) (108) and wasproposed to recognize the substrate (45)
The purified enzyme cleaves modified DNA containingm5C and hm5C but not m4C in the presence of ATP ordATP but not other nucleotides The negative result form4C is firm plasmids modified at the same site by an m5CMTase (Dcm) or an m4C MTase (MPspGI) were testedThe former (Cm5CWGG) was sensitive whereas the latterwas resistant Thus the sequence preference is likely tobe satisfied m6A is likely not a substrate but few m6A-containing sites were examined
Like McrBC SauUSI requires the presence of two sitesfor efficient cleavage Presumably the ATPase activityparticipates in monitoring the presence of two sites asfor other nucleotide-dependent REases includingMcrBC The mechanism of communication is unknownThe enzyme belongs to a family of highly similarorthologues found in other sequenced Staphylococci(Tables 1 and 2) and more distant homologues can befound in sequenced bacterial and archaeal genomes
EVOLUTIONARY PRESSURES ON R-M SYSTEMS
Evolution by selfish propagation
One way to understand the massive variety of restrictionsystems and their sporadic distribution is to locate theevolutionary drivers of enzyme diversification in theenzyme genes themselves as selfish elements Work fromthe Kobayashi laboratory has elaborated clear examplesof selfish behavior in some Type II enzyme systems(109110) in which the host becomes lsquoaddictedrsquo to theR-M system Once a cell has acquired an R-M pair lossof the genes results in death of that cellrsquos descendants asthe REase is frequently still present and able to act on thegenome following loss of methylation activity In this
64 Nucleic Acids Research 2014 Vol 42 No 1
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
perspective the role played by modification-dependentenzymes is host defense to exclude systems with lsquoforeignrsquoMTase patterns and prevent the cell from loading up withparasites The exclusion event is accompanied by the deathof the cell (111ndash113) Weak sequence specificity of Type IVenzymes could then result from the need to control entryof a wide variety of invading systems
The selfish aspect certainly plays a role in R-M popula-tion biology but cannot be the whole story Type II R-Msystems can still be lost by inactivation of the R gene firstMoreover Type I systems escape this scenario withcomplex control of cleavage activity the restrictionassembly includes a methylation assembly to begin withtherefore the R protein cannot act unless the MTase ispresent in addition failure of the methylation activity inan intact complex leads to abrogation of R activity some-times by action of the ClpXP protease specifically on theR protein (114ndash116)
Furthermore in population terms a cell that acquiredand became addicted to an R-M system should lose incompetition with a sibling that never received thesystem Two factors could counter this First acquisitioncould be accompanied by an increase in the total numberof copies of the R-M system in the population asproposed for invading transposable elements Thisoverreplication results in more copies of the systemcreated than are lost whether to suicide or to other select-ive disadvantage [see eg (117118)] R-M gene amplifica-tion within a cell has been reported experimentally (119)but spread in a population has not been demonstrated yetA second factor that could counter the disability of addic-tion is localization of competition In a structured envir-onment (colonies on a plate or biofilm on large or smallsurfaces) killing of segregants preserves limiting nutrientsfor lineages that retain the toxinantitoxin pair (120121)Much of the real world is structured so this is an import-ant condition
Evolution by phage-host arms race
A second perspective supposes that the modification-dependent Type IV enzymes arose from the competitivecoevolutionary interaction between phages and theirhosts This was first enunciated by Revel and Luria (2)and most recently by Black and coworkers (122) seealso (123) That is hosts used modification-blockedrestriction to defend against phage infection T-evenphages developed methods of substituting modified basesfor the ordinary ones hosts developed Type IV enzymes indefense phages added sugar or other modifications (19) tothwart Type IV enzymes hosts extended Type IV enzymesto accommodate these decorations finally phages de-veloped protein inhibitors specific for these enzymes aswell T4 phages deliver a protein inhibitor (IPI) alongwith the DNA on infection which allows growth in thepresence of EcoCTGmrSD The locus responsible for thisinhibitor is highly variable among relatives of T4 asgmrSD is in enteric bacteria (both in distribution and inaa sequence) When phage with different IP1 alleles weretested for protection from cloned EcoCTGmrSD and itshomolog EcoUTGmrSD specificity was evident one or
the other or both or neither of the two homologs wascounteracted in individual cases (122) This variability ofthe outcome supports the idea that phage-host interactiondrives at least some of these developmentsIn this perspective the weak sequence selectivity of the
Type IV systems might simply reflect the lack of endogen-ous targets for the enzyme As the host does not present anyhm5C and the phage is completely substituted selection forsequence-specificity is weak Selection would act to spareany co-residentMTases This differs fromType II enzymeswhere the MTase and REase must co-evolve to allow thehost to survive Each Type IV system is compatible withsome suite of Type I-IIIMTases (and thus the R-M systemsas a whole) Methylated or hydroxymethylated bases maynot be recognized at all (EcoCTGmrSD) or the systemmay require one specific base in addition to the modifiedone (McrA McrBC MspJI and PvuRts1I) MTases mod-ifying sites not including that base are then compatible asDcm (CmCWGG) is compatible with the McrA McrBCand Mrr systems in E coli K12Type IV systems that restrict methylated bases in a
weakly specified sequence context confer an additionaladvantage in competition with phages Many phagessuch as have not evolved the nucleotide-substitutionstrategy used by the T-even phages These phagesnormally carry the modification pattern of the mostrecent host if the last host expressed an MTase creatinga susceptible site the Type IV enzyme of the new host willdestroy the invader and limit the infection This may beaccompanied by the death of the individual infectedtherefore protection can be conferred on the sibling popu-lation (111)A further implication of this scenario considers the fate
of a population invaded by phage Phage survival ofrestriction occurs at biologically relevant frequencies(106-102) The survivors of restriction carry the particu-lar methylation pattern of the particular cell and thus areresistant to all restriction systems it might have carried(Types IndashIV) A bacterial population as a whole thenbenefits from mechanisms that diversify the suite ofR-M systems so that such surviving phage do not havefree access to the entire population The extreme variabil-ity of R-M system content in isolates of the same species iscompatible with this idea [see eg (124) REBASEGenomes httptoolsnebcomvinczegenomes] Suchvariability also limits and shapes interstrain genetransfer (115125126)The Raleigh laboratory has built on elegant genetic
work with Type I enzymes in the Murray laboratory(127ndash130) to investigate a locus designated thelsquoImmigration Control locusrsquo or ICR that exemplifiesvariable R-M content Alternative DNA segments con-taining R-M systems are located at the same definedlocation in most E coli chromosomes between the yjiSand yjiA genes The ICR in the non-restricting E coli Cstrain [used in the original definition of the R-M phenom-enon (131)] contains a remnant of a Type I enzyme R geneand is 13 kilobase shorter than the same region in E coliK12 (132) The mechanism of segment replacement is stillunknown The ICR would be an example of the lsquodefenseislandsrsquo analyzed by the Koonin group (133) lsquoDefense
Nucleic Acids Research 2014 Vol 42 No 1 65
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
islandsrsquo contain genes that can defend against phage orother invading DNA these exhibit bioinformaticproperties similar to lsquomobilome islandsrsquo containing mobil-ization genes (transposases for example) However themechanism of mobilization has not been identified forthe ICR
FINAL THOUGHTS
The extreme diversity of R-M systems that recognizeordinary DNA seems likely to be approached by the di-versity of Type IV restriction systems Type IV enzymesare hard to find as most detection methods depend ondevelopment of genetic systems for each taxon or on ser-endipity Those characterized so far mostly stem frominitial genetic investigation of limits on infection trans-formation or transduction Barriers encountered provideleads to the genes involved Bioinformatic analysis hashelped to identify relatives which may be more tractableto biochemical investigation than the example originallyfound This approach has pitfalls the gene encodingMspJI was first thought to code for an enzyme recognizingan unmodified site because it is immediately adjacent toan (inactive it is now thought) cytosine MTase geneProvidentially the first expression host was devoid of sen-sitive sites whereas the first test substrate contained some(51) A combination of biological experiments with bio-informatics and biochemistry will be needed to reveal thefull spectrum of Type IV enzymes that may lurk within thevast universe of unidentified ORFs in bacterial systemsOne might begin with those strains whose genomes carryfew Type II systems Bacillus or Corynebacterium asopposed to Helicobacter or Neisseria [see the Genomessection of REBASE (50)]The role of lsquodefense islandsrsquo and their relation to the
lsquomobilomersquo in bacterial population biology remains to bedetermined If a defense island is similar to a mobilomeisland there should be a mechanism of mobilizationnearby which would boost the contribution oflsquooverreplicationrsquo to the account of selections acting onR-M systems R-M systems of all types can be found onor adjacent to known mobilizing elements (134135) buthave not been shown to move experimentallyOn another note it may turn out that evolutionarily
there is a continuum between the apparently modifica-tion-dependent and modification-blocked paths Onerelative of McrBC predicted by bioinformatics analysesis LlaI a system that recognizes an unmodified sequenceand requires two MTases to support it (136) The enzymeBamHI prefers to cleave DNA with m6A in its GGATCCsite and mutants can be isolated that require this modifiedbase (137) Are there native systems similarly protected bymodification of one position in the recognition site butdependent on modification at a different one An interest-ing evolutionary series can be imagined
ACKNOWLEDGEMENTS
The authors thank Rich Roberts and David Dryden forencouragement and critical comment They are grateful to
Tom Bickle Virginius Siksnys Yu Zheng and XiadongCheng for permission to reproduce figures They alsothank Ivan Correa for help drawing base structures andTasha Jose for illustration assistance They also thank theanonymous reviewers of a previous manuscript for caretaken with the review process greatly improving thepresent document
FUNDING
New England Biolabs (EAR) Funding for open accesscharge New England Biolabs (to EAR)
Conflict of interest statement None declared
REFERENCES
1 LuriaSE and HumanML (1952) A nonhereditary host-inducedvariation of bacterial viruses J Bacteriol 64 557ndash559
2 RevelHR and LuriaSE (1970) DNA-glucosylation in T-evenphage genetic determination and role in phage-host interactionAnnu Rev Genet 4 177ndash192
3 RaleighEA and WilsonG (1986) Escherichia coli K-12 restrictsDNA containing 5-methylcytosine Proc Natl Acad Sci USA83 9070ndash9074
4 Noyer-WeidnerM DiazR and ReinersL (1986) Cytosine-specific DNA modification interferes with plasmid establishmentin Escherichia coli K12 involvement of rglB Mol Gen Genet205 469ndash475
5 BlumenthalRM GregorySA and CooperiderJS (1985)Cloning of a restriction-modification system from Proteus vulgarisand its use in analyzing a methylase-sensitive phenotype inEscherichia coli J Bacteriol 164 501ndash509
6 SohailA LiebM DarM and BhagwatAS (1990) A generequired for very short patch repair in Escherichia coli is adjacentto the DNA cytosine methylase gene J Bacteriol 1724214ndash4221
7 MacNeilDJ (1988) Characterization of a unique methyl-specificrestriction system in Streptomyces avermitilis J Bacteriol 1705607ndash5612
8 VertesAA InuiM KobayashiM KurusuY and YukawaH(1993) Presence of mrr- and mcr-like restriction systems incoryneform bacteria Res Microbiol 144 181ndash185
9 MacalusoA and MettusAM (1991) Efficient transformation ofBacillus thuringiensis requires nonmethylated plasmid DNAJ Bacteriol 173 1353ndash1356
10 HolmesML NuttallSD and Dyall-SmithML (1991)Construction and use of halobacterial shuttle vectors and furtherstudies on Haloferax DNA gyrase J Bacteriol 173 3807ndash3813
11 PosfaiJ BhagwatAS PosfaiG and RobertsRJ (1989)Predictive motifs derived from cytosine methyltransferases NucleicAcids Res 17 2421ndash2435
12 MaloneT BlumenthalRM and ChengX (1995) Structure-guided analysis reveals nine sequence motifs conserved amongDNA amino-methyltransferases and suggests a catalyticmechanism for these enzymes J Mol Biol 253 618ndash632
13 TiminskasA ButkusV and JanulaitisA (1995) Sequence motifcharacteristics for DNA [cytosine-N4] and DNA [adenine-N6]methyltransferases Classification of all DNA methyltransferasesGene 157 3ndash11
14 DickmanMJ InglestonSM SedelnikovaSE RaffertyJBLloydRG GrasbyJA and HornbyDP (2002) The RuvABCresolvasome Eur J Biochem 269 5492ndash5501
15 KovallRA and MatthewsBW (1998) Structural functionaland evolutionary relationships between lambda-exonuclease andthe type II restriction endonucleases Proc Natl Acad Sci USA95 7893ndash7897
16 RobertsRJ ChangYC HuZ RachlinJN AntonBPPokrzywaRM ChoiHP FallerLL GuleriaJ HousmanGet al (2010) COMBREX a project to accelerate the functional
66 Nucleic Acids Research 2014 Vol 42 No 1
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
annotation of prokaryotic genomes Nucleic Acids Res 39D11ndashD14
17 RaePM and SteeleRE (1978) Modified bases in the DNAs ofunicellular eukaryotes an examination of distributions andpossible roles with emphasis on hydroxymethyluracil indinoflagellates Biosystems 10 37ndash53
18 WarrenRAJ (1980) Modified bases in bacteriophage DNAsAnnu Rev Microbiol 34 137ndash158
19 Gommers-AmptJH and BorstP (1995) Hypermodified bases inDNA FASEB J 9 1034ndash1042
20 SwintonD HattmanS CrainPF ChengCS SmithDL andMcCloskeyJA (1983) Purification and characterization of theunusual deoxynucleoside alpha-N-(9-beta-D-20-deoxyribofuranosylpurin-6-yl)glycinamide specified by the phageMu modification function Proc Natl Acad Sci USA 807400ndash7404
21 TahilianiM KohKP ShenY PastorWA BandukwalaHBrudnoY AgarwalS IyerLM LiuDR AravindL et al(2009) Conversion of 5-methylcytosine to 5-hydroxymethylcytosinein mammalian DNA by MLL partner TET1 Science 324930ndash935
22 KriaucionisS and HeintzN (2009) The nuclear DNA base5-hydroxymethylcytosine is present in Purkinje neurons and thebrain Science 324 929ndash930
23 BorstP and SabatiniR (2008) Base J discovery biosynthesisand possible functions Annu RevMicrobiol 62 235ndash251
24 KaminskaKH and BujnickiJM (2008) Bacteriophage MuMom protein responsible for DNA modification is a new memberof the acyltransferase superfamily Cell Cycle 7 120ndash121
25 IyerLM TahilianiM RaoA and AravindL (2009) Predictionof novel families of enzymes involved in oxidative and othercomplex modifications of bases in nucleic acids Cell Cycle 81698ndash1710
26 ZhouX HeX LiangJ LiA XuT KieserT HelmannJDand DengZ (2005) A novel DNA modification by sulphur MolMicrobiol 57 1428ndash1438
27 OuHY HeX ShaoY TaiC RajakumarK and DengZ(2009) dndDB A Database Focused on Phosphorothioation ofthe DNA Backbone PloS One 4 e5132
28 XuT LiangJ ChenS WangL HeX YouD WangZLiA XuZ ZhouX et al (2009) DNA phosphorothioation inStreptomyces lividans mutational analysis of the dnd locusBMC Microbiol 9 41
29 WangL ChenS VerginKL GiovannoniSJ ChanSWDeMottMS TaghizadehK CorderoOX CutlerMTimberlakeS et al (2011) DNA phosphorothioation iswidespread and quantized in bacterial genomes Proc Natl AcadSci USA 108 2963ndash2968
30 Loslashbner-OlesenA SkovgaardO and MarinusMG (2005) Dammethylation coordinating cellular processes Curr OpinMicrobiol 8 154ndash160
31 LowDA and CasadesusJ (2008) Clocks and switches bacterialgene regulation by DNA adenine methylation Curr OpinMicrobiol 11 106ndash112
32 FreitagM and SelkerEU (2005) Controlling DNA methylationMany roads to one modification Curr Opin Genet Dev 15191ndash199
33 ProhaskaSJ StadlerPF and KrakauerDC (2010) Innovationin gene regulation The case of chromatin computation J TheorBiol 265 27ndash44
34 SeshasayeeASN SinghP and KrishnaS (2012) Context-dependent conservation of DNA methyltransferases in bacteriaNucleic Acids Res 40 7066ndash7073
35 AntonBP (2011) Dissertation Boston University BostonMassachusetts USA
36 RaleighEA TrimarchiR and RevelH (1989) Genetic andphysical mapping of the mcrA (rglA) and mcrB (rglB) loci ofEscherichia coli K-12 Genetics 122 279ndash296
37 Gonzalez-CeronG Miranda-OlivaresOJ and Servın-GonzalezL(2009) Characterization of the methyl-specific restriction system ofStreptomyces coelicolor A3(2) and of the role played by laterallyacquired nucleases FEMS Microbiol Lett 301 35ndash43
38 LiuG OuHY WangT LiL TanH ZhouXRajakumarK DengZ and HeX (2010) Cleavage of
phosphorothioated DNA and methylated DNA by the type IVrestriction endonuclease ScoMcrA PLoS Genet 6 e1001253
39 SutherlandE CoeL and RaleighEA (1992) McrBC amultisubunit GTP-dependent restriction endonuclease J MolBiol 225 327ndash348
40 KrugerT WildC and Noyer-WeidnerM (1995) McrB aprokaryotic protein specifically recognizing DNA containingmodified cytosine residues EMBO J 14 2661ndash2669
41 SitaramanR and LepplaSH (2012) Methylation-dependentDNA restriction in Bacillus anthracis Gene 494 44ndash50
42 HeitmanJ and ModelP (1987) Site-specific methylases inducethe SOS DNA repair response in Escherichia coli J Bacteriol169 3243ndash3250
43 Waite-ReesPA KeatingCJ MoranLS SlatkoBEHornstraLJ and BennerJS (1991) Characterisation andexpression of the Escherichia coli Mrr restriction systemJ Bacteriol 173 5207ndash5219
44 KerrAL JeonYJ SvensonCJ RogersPL and NeilanBA(2010) DNA restriction-modification systems in the ethanologenZymomonas mobilis ZM4 Appl MicrobiolBiotech 89 761ndash769
45 XuSY CorvagliaAR ChanSH ZhengY and LinderP(2011) A type IV modification-dependent restriction enzymeSauUSI from Staphylococcus aureus subsp aureus USA300Nucleic Acids Res 39 5597ndash5610
46 MonkIR ShahIM XuM TanMW and FosterTJ (2012)Transforming the untransformable application of directtransformation to manipulate genetically Staphylococcus aureusand Staphylococcus epidermidis mBio 3 e00277ndash00211
47 WangH GuanS QuimbyA Cohen-KarniD PradhanSWilsonG RobertsRJ ZhuZ and ZhengY (2011)Comparative characterization of the PvuRts1I family ofrestriction enzymes and their application in mapping genomic5-hydroxymethylcytosine Nucleic Acids Res 39 9294ndash9305
48 JanosiL YonemitsuH HongH and KajiA (1994) Molecularcloning and expression of a novel hydroxymethylcytosine-specificrestriction enzyme (PvuRts1I) modulated by glucosylation ofDNAJ Mol Biol 242 45ndash61
49 BairCL and BlackLW (2007) A type IV modificationdependent restriction nuclease that targets glucosylatedhydroxymethyl cytosine modified DNAs J Mol Biol 366768ndash778
50 RobertsRJ VinczeT PosfaiJ and MacelisD (2010)REBASEndasha database for DNA restriction and modificationenzymes genes and genomes Nucleic Acids Res 38 D234ndashD236
51 ZhengY Cohen-KarniD XuD ChinHG WilsonGPradhanS and RobertsRJ (2010) A unique family of Mrr-likemodification-dependent restriction endonucleases Nucleic AcidsRes 38 5527ndash5534
52 Cohen-KarniD XuD AponeL FomenkovA SunZDavisPJ KinneySRM Yamada-MabuchiM XuSYDavisT et al (2011) The MspJI family of modification-dependent restriction endonucleases for epigenetic studies ProcNatl Acad Sci USA 108 11040ndash11045
53 LubysA VitkuteJ LubieneJ and JanulaitisA (2011)Restriction enzymes and their applications US Patent ApplicationPublication 20110207139 A1 httpappftusptogovnetahtmlPTOsearch-boolhtml (27 August 2013 date last accessed)
54 TarasovaGV NayakshinaTN and DegtyarevSK (2008)Substrate specificity of new methyl-directed DNA endonucleaseGlaI BMC Mol Biol 9 7
55 MulliganEA and DunnJJ (2008) Cloning purification andinitial characterization of E coli McrA a putative5-methylcytosine-specific nuclease Protein ExprPurif 62 98ndash103
56 MulliganEA HatchwellE McCorkleSR and DunnJJ (2010)Differential binding of Escherichia coli McrA protein to DNAsequences that contain the dinucleotide m5CpG Nucleic AcidsRes 38 1997ndash2005
57 AntonBP and RaleighEA (2004) Transposon-mediated linkerinsertion scanning mutagenesis of the Escherichia coli McrAendonuclease J Bacteriol 186 5699ndash5707
58 BujnickiJM RadlinskaM and RychlewskiL (2000) Atomicmodel of the 5-methylcytosine-specific restriction enzyme McrAreveals an atypical zinc finger and structural similarity tobetabetaalphaMe endonucleases Mol Microbiol 37 1280ndash1281
Nucleic Acids Research 2014 Vol 42 No 1 67
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
59 GastFU BrinkmannT PieperU KrugerT NoyerWeidnerM and PingoudA (1997) The recognition of methylatedDNA by the GTP-dependent restriction endonuclease McrBCresides in the N-terminal domain of McrB Biol Chem 378975ndash982
60 SukackaiteR GrazulisS TamulaitisG and SiksnysV (2012)The recognition domain of the methyl-specific endonucleaseMcrBC flips out 5-methylcytosine Nucleic Acids Res 407552ndash7562
61 DilaD SutherlandE MoranL SlatkoB and RaleighEA(1990) Genetic and sequence organization of the mcrBC locus ofEscherichia coli K-12 J Bacteriol 172 4888ndash4900
62 PieperU BrinkmannT KrugerT Noyer WeidnerM andPingoudA (1997) Characterization of the interaction between therestriction endonuclease McrBC from E coli and its cofactorGTP J Mol Biol 272 190ndash199
63 PieperU SchweitzerT GrollDH GastFU and PingoudA(1999) The GTP-binding domain of McrB more than just avariation on a common theme J Mol Biol 292 547ndash556
64 BujnickiJM and RychlewskiL (2001) Grouping together highlydiverged PD-(DE)XK nucleases and identification of novelsuperfamily members using structure-guided alignment ofsequence profiles J Mol MicrobiolBiotech 3 69ndash72
65 PieperU and PingoudA (2002) A mutational analysis of thePD DEXK motif suggests that McrC harbors the catalyticcenter for DNA cleavage by the GTP-dependent restrictionenzyme McrBC from Escherichia coli Biochemistry 415236ndash5244
66 KelleherJE and RaleighEA (1991) A novel activity inEscherichia coli K-12 that directs restriction of DNA modified atCG dinucleotides J Bacteriol 173 5220ndash5223
67 OrlowskiJ MebrhatuMT MichielsCW BujnickiJM andAertsenA (2008) Mutational analysis and a structural model ofmethyl-directed restriction enzyme Mrr Biochem Biophys ResCommun 377 862ndash866
68 AravindL MakarovaKS and KooninEV (2000) SURVEYAND SUMMARY holliday junction resolvases and relatednucleases identification of new families phyletic distribution andevolutionary trajectories Nucleic Acids Res 28 3417ndash3432
69 BujnickiJM and RychlewskiL (2001) Identification of a PD-(DE)XK-like domain with a novel configuration of the endonucleaseactive site in the methyl-directed restriction enzyme Mrr and itshomologs Gene 267 183ndash191
70 SteczkiewiczK MuszewskaA KnizewskiL RychlewskiL andGinalskiK (2012) Sequence structure and functional diversity ofPD-(DE)XK phosphodiesterase superfamily Nucleic Acids Res40 7016ndash7045
71 BairCL RifatD and BlackLW (2007) Exclusion of glucosyl-hydroxymethylcytosine DNA containing bacteriophages isovercome by the injected protein inhibitor IPI J Mol Biol 366779ndash789
72 StewartFJ and RaleighEA (1998) Dependence of McrBCcleavage on distance between recognition elements Biol Chem379 611ndash616
73 PanneD RaleighEA and BickleTA (1999) The McrBCendonuclease translocates DNA in a reaction dependent on GTPhydrolysis J Mol Biol 290 49ndash60
74 PieperU GrollDH WunschS GastFU SpeckC MuckeNand PingoudA (2002) The GTP-dependent restriction enzymeMcrBC from Escherichia coli forms high-molecular masscomplexes with DNA and produces a cleavage patternwith a characteristic 10-base pair repeat Biochemistry 415245ndash5254
75 IshikawaK HandaN SearsL RaleighEA and KobayashiI(2011) Cleavage of a model DNA replication fork by amethyl-specific endonuclease Nucleic Acids Res 39 5489ndash5498
76 StewartFJ PanneD BickleTA and RaleighEA (2000)Methyl-specific DNA binding by McrBC a modification-dependent restriction enzyme J Mol Biol 298 611ndash622
77 LacksS and GreenbergB (1977) Complementary specificity ofrestriction endonucleases DpnI and DpnII with respect to DNAmethylation J Mol Biol 114 153ndash168
78 SiwekW CzapinskaH BochtlerM BujnickiJM andSkowronekK (2012) Crystal structure and mechanism
of action of the N6-methyladenine-dependent type IIMrestriction endonuclease RDpnI Nucleic Acids Res 407563ndash7572
79 LacksS and GreenbergB (1975) A deoxyribonuclease ofDiplococcus pneumoniae specific for methylated DNA J BiolChem 250 4060ndash4066
80 HortonJR MabuchiMY Cohen-KarniD ZhangXGriggsRM SamaranayakeM RobertsRJ ZhengY andChengX (2012) Structure and cleavage activity of the tetramericMspJI DNA modification-dependent restriction endonucleaseNucleic Acids Res 40 9763ndash9773
81 CymermanIA ObarskaA SkowronekKJ LubysA andBujnickiJM (2006) Identification of a new subfamily of HNHnucleases and experimental characterization of a representativemember HphI restriction endonuclease Proteins 65 867ndash876
82 GorbalenyaAE (1994) Self-splicing group I and group II intronsencode homologous (putative) DNA endonucleases of a newfamily Protein Sci 3 1117ndash1120
83 ChanSH OpitzL HigginsL OrsquoLoaneD and XuSY (2010)Cofactor requirement of HpyAV restriction endonuclease PloSOne 5 e9071
84 BourniquelAA and BickleTA (2002) Complex restrictionenzymes NTP-driven molecular motors Biochimie 84 1047ndash1059
85 OrsquoDriscollJ HeiterDF WilsonGG FitzgeraldGFRobertsR and van SinderenD (2006) A genetic dissection ofthe LlaJI restriction cassette reveals insights on a novelbacteriophage resistance system BMC Microbiol 6 40
86 SmithRM DiffinFM SaveryNJ JosephsenJ andSzczelkunMD (2009) DNA cleavage and methylation specificityof the single polypeptide restriction-modification enzyme LlaGINucleic Acids Res 37 7206ndash7218
87 InterthalH PouliotJJ and ChampouxJJ (2001) The tyrosyl-DNA phosphodiesterase Tdp1 is a member of the phospholipaseD superfamily Proc Natl Acad Sci USA 98 12009ndash12014
88 SasnauskasG HalfordSE and SiksnysV (2003) How the BfiIrestriction enzyme uses one active site to cut two DNA strandsProc Natl Acad Sci USA 100 6410ndash6415
89 SasnauskasG ConnollyBA HalfordSE and SiksnysV (2007)Site-specific DNA transesterification catalyzed by a restrictionenzyme Proc Natl Acad Sci USA 104 2115ndash2120
90 Marchler-BauerA ZhengC ChitsazF DerbyshireMKGeerLY GeerRC GonzalesNR GwadzM HurwitzDILanczyckiCJ et al (2012) CDD conserved domains andprotein three-dimensional structure Nucleic Acids Res 41D348ndashD352
91 KravetzAN ZakharovaMV BeletskayaIV SinevaEVDenjmuchametovMM PetrovSI GlatmanLI andSoloninAS (1993) The cleavage sites and localization of genesencoding the restriction endonucleases Eco1831I and EcoHIGene 129 153ndash154
92 KlimasauskasS KumarS RobertsRJ and ChengX (1994)HhaI methyltransferase flips its target base out of the DNA helixCell 76 357ndash369
93 HashimotoH VertinoPM and ChengX (2010) Molecularcoupling of DNA methylation and histone methylationEpigenomics 2 657ndash669
94 RajakumaraE LawJA SimanshuDK VoigtPJohnsonLM ReinbergD PatelDJ and JacobsenSE (2011)A dual flip-out mechanism for 5mC recognition by theArabidopsis SUVH5 SRA domain and its impact on DNAmethylation and H3K9 dimethylation in vivo Genes Dev 25137ndash152
95 SharifJ and KosekiH (2011) Recruitment of Dnmt1 roles of theSRA protein Np95 (Uhrfi) and other factors Prog Mol BiolTransl Sci 101 289ndash310
96 Tesfazgi MebrhatuM WywialE GhoshA MichielsCWLindnerAB TaddeiF BujnickiJM Van MelderenL andAertsenA (2011) Evidence for an evolutionary antagonismbetween Mrr and Type III modification systems Nucleic AcidsRes 39 5991ndash6001
97 GrantGN JesseeJ BloomFR and HanahanD (1990)Differential plasmid rescue from transgenic mouse DNAs intoEscherichia coli methylation-restriction mutants Proc Natl AcadSci USA 87 4645ndash4649
68 Nucleic Acids Research 2014 Vol 42 No 1
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
98 AertsenA and MichielsCW (2005) Mrr instigates the SOSresponse after high pressure stress in Escherichia coli MolMicrobiol 58 1381ndash1391
99 WeiH TherrienC BlanchardA GuanS and ZhuZ (2008)The Fidelity Index provides a systematic quantitation of staractivity of DNA restriction endonucleases Nucleic Acids Res36 e50
100 LuL PatelH and BisslerJJ (2002) Optimizing DpnIdigestion to detect replicated DNA Biotechniques 33 316ndash318
101 MadejT AddessKJ FongJH GeerLY GeerRCLanczyckiCJ LiuC LuS Marchler-BauerAPanchenkoAR et al (2011) MMDB 3D structures andmacromolecular interactions Nucleic Acids Res 40 D461ndashD464
102 GibratJF MadejT and BryantSH (1996) Surprisingsimilarities in structure comparison Curr Opin Struct Biol 6377ndash385
103 PanneD MullerSA WirtzS EngelA and BickleTA (2001)The McrBC restriction endonuclease assembles into a ringstructure in the presence of G nucleotides EMBO J 203210ndash3217
104 NeuwaldAF AravindL SpougeJL and KooninEV (1999)AAA+ A class of chaperone-like ATPases associated with theassembly operation and disassembly of protein complexesGenome Res 9 27ndash43
105 OguraT and WilkinsonAJ (2001) AAA+ superfamilyATPases common structurendashdiverse function Genes Cells 6575ndash597
106 BearyTP BraymerHD and AchbergerEC (1997) Evidenceof participation of McrB(S) in McrBC restriction in Escherichiacoli K-12 J Bacteriol 179 7768ndash7775
107 PanneD RaleighEA and BickleTA (1998) McrBs amodulator peptide for McrBC activity EMBO J 175477ndash5483
108 PuntaM CoggillPC EberhardtRY MistryJ TateJBoursnellC PangN ForslundK CericG ClementsJ et al(2011) The Pfam protein families database Nucleic Acids Res40 D290ndashD301
109 NaitoT KusanoK and KobayashiI (1995) Selfish behavior ofrestriction-modification systems Science 267 897ndash899
110 KobayashiI (2001) Behavior of restriction-modification systemsas selfish mobile elements and their impact on genome evolutionNucleic Acids Res 29 3742ndash3756
111 FukudaE KaminskaKH BujnickiJM and KobayashiI(2008) Cell death upon epigenetic genome methylation a novelfunction of methyl-specific deoxyribonucleases Genome Biol 9R163
112 IshikawaK FukudaE and KobayashiI (2010) Conflictstargeting epigenetic systems and their resolution by cell deathnovel concepts for methyl-specific and other restriction systemsDNA Res 17 325ndash342
113 FukuyoM SasakiA and KobayashiI (2012) Success of asuicidal defense strategy against infection in a structured habitatSci Rep 2 238
114 MakovetsS TitheradgeAJ and MurrayNE (1998) ClpXand ClpP are essential for the efficient acquisition of genesspecifying type IA and IB restriction systems Mol Microbiol28 25ndash35
115 MurrayNE (2002) Immigration control of DNA in bacteriaself versus non-self Microbiology 148 3ndash20
116 MakovetsS PowellLM TitheradgeAJB BlakelyGW andMurrayNE (2004) Is modification sufficient to protect abacterial chromosome from a resident restriction endonucleaseMol Microbiol 51 135ndash147
117 OrgelLE and CrickFHC (1980) Selfish DNA the ultimateparasite Nature 284 604ndash607
118 KidwellMG and LischDR (2001) Perspective transposableelements parasitic DNA and genome evolution Evolution 551ndash24
119 SadykovM AsamiY NikiH HandaN ItayaMTanokuraM and KobayashiI (2003) Multiplication of arestriction-modification gene complex Mol Microbiol 48417ndash427
120 ChaoL and LevinBR (1981) Structured habitats and theevolution of anticompetitor toxins in bacteria Proc Natl AcadSci USA 78 6324ndash6328
121 MagnusonRD (2007) Hypothetical functions of toxin-antitoxinsystems J Bacteriol 189 6089ndash6092
122 RifatD WrightNT VarneyKM WeberDJ and BlackLW(2008) Restriction endonuclease inhibitor IPI of bacteriophageT4 a novel structure for a dedicated target J Mol Biol 375720ndash734
123 LabrieSJ SamsonJE and MoineauS (2010) Bacteriophageresistance mechanisms Nat Rev Microbiol 8 317ndash327
124 LinLF PosfaiJ RobertsRJ and KongH (2001)Comparative genomics of the restriction-modification systemsin Helicobacter pylori Proc Natl Acad Sci USA 982740ndash2745
125 MilkmanR RaleighEA McKaneM CrydermanDBilodeauP and McWeenyK (1999) Molecular evolution of theEscherichia coli chromosome V Recombination patterns amongstrains of diverse origin Genetics 153 539ndash554
126 HumbertO DorerMS and SalamaNR (2011)Characterization of Helicobacter pylori factors that controltransformation frequency and integration length during inter-strain DNA recombination Mol Microbiol 79 387ndash401
127 BarcusVA TitheradgeAJB and MurrayNE (1995) Thediversity of alleles at the hsd locus in natural populations ofEscherichia coli Genetics 140 1187ndash1197
128 SharpPM KelleherJE DanielAS CowanGM andMurrayNE (1992) Roles of selection and recombination in theevolution of type I restriction-modification systems inenterobacteria Proc Natl Acad Sci USA 89 9836ndash9840
129 LoenenWAM DanielAS BraymerHD and MurrayNE(1987) Organization and sequence of the hsd genes ofEscherichia coli K-12 J Mol Biol 198 159ndash170
130 SuriB and BickleTA (1985) EcoA the first member of a newfamily of type I restriction modification systems Geneorganization and enzymatic activities J Mol Biol 186 77ndash85
131 BertaniG and WeigleJJ (1953) Host-controlled variation inbacterial viruses J Bacteriol 65 113ndash121
132 SibleyMH and RaleighEA (2004) Cassette-like variation ofrestriction enzyme genes in Escherichia coli C and relativesNucleic Acids Res 32 522ndash534
133 MakarovaKS WolfYI SnirS and KooninEV (2011)Defense islands in bacterial and archaeal genomes and predictionof novel defense systems J Bacteriol 193 6039ndash6056
134 FurutaY AbeK and KobayashiI (2010) Genome comparisonand context analysis reveals putative mobile forms of restriction-modification systems and related rearrangements Nucleic AcidsRes 38 2428ndash2443
135 FurutaY and KobayashiI (2012) Restriction-modificationsystems as moblie epigenetic elements In RobertsAP andMullanyP (eds) Bacterial Integrative Mobile Genetic ElementsLandes Bioscience pp 85ndash103
136 OrsquoSullivanDJ and KlaenhammerTR (1998) Control ofexpression of LlaI restriction in Lactococcus lactis MolMicrobiol 27 1009ndash1020
137 WhitakerRD DornerLF and SchildkrautI (1999) A mutantof BamHI restriction endonuclease which requires N6-methyladenine for cleavage J Mol Biol 285 1525ndash1536
Nucleic Acids Research 2014 Vol 42 No 1 69
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
multiple evolutionary events apparently have occurred toconnect nuclease domains to recognition moieties (81)
Nuclease domainsEnzymes that recognize modified DNA with minimalsequence selectivity have emerged at least six times asexemplified by the McrA McrBC SauUSI MrrPvuRts1I and GmrSD families These exemplars are dis-cussed in more detail later in the text In brief nucleasedomains have been attached covalently or (for McrC) viaproteinndashprotein interaction to domains with DNA bindingand regulatory functions
EcoKMcrA carries a C-terminal H-N-Hc nucleasedomain identified bioinformatically (5882) (Figure 2)This nuclease domain is also found in modification-blocked nucleases (81) The purified binding-competentprotein did not cleave under a variety of buffer conditionsand cofactor additions (55) ScoA3McrA is designatedlsquoMcrArsquo due to its possession of a similar nucleasedomain For this enzyme cleavage depends on Mn2+ orCo2+ (38) and occurs at a variable distance from PT-modified sites Modification-blocked H-N-H REases alsooften exhibit unusual metal ion requirements [eg (83)]
McrBC The required McrC component (3940) is thenuclease moiety (65) (Figure 3) Mutational analysis
confirms that it is a PD-(DE)XK nuclease (65) in agree-ment with bioinformatic classification (64) Cleavageresults when McrC associates with full-lengthMcrBGTP complex bound to DNA and GTP ishydrolyzed (72) LlaJI a modification-blocked restrictionactivity exhibits a similar organization (85) althoughcleavage could not be demonstrated in vitroThe classic modification-dependent enzyme DpnI also
carries a PD-(DE)XK motif (see further later in the text)Mrr EcoKMrr contains a variant of the PD-(DE)XK
motif (6869) with the Mrr-N (E coli K12) presumedDNA-binding domain MspJI (see further later in thetext) also carries a nuclease domain in this family Aswith McrA McrBC and SauUSI nuclease domain simi-larity does not in itself dictate modification preferenceproperties the single-chain R-M system LlaGI hasconserved motifs characteristic of the E coli Mrrprotein but this enzyme does not target methylatedDNA (86)SauUSI This is a modification-dependent enzyme with
a phosphodiesterase cleavage domain akin to one origin-ally identified in phospholipase-D (45) Mutation of anyof the four conserved catalytic residues abolishes in vitroactivity This cleavage domain is also found in stand-alone
Table 3 Characteristics of Type IV restriction enzymes
Protein Subunitsdomains
DNABinding
Endonucleasedomain
NTP hydrolysis Recognition sitea comment and references
EcoKMcrA ndash In vitro cleavage not reportedN-terminal DBD (YgtR)m5CGR bound (5556) hm5Cm5C
discrimination (57)C-terminal H-N-Hc Bioinformatic ID (58) required for damage to DNA
in vivo (57)ScoA3McrA ndash Some CmCWGG and some S-DNA (PT modified)
sites are cleaved (38)N-terminal DBD Not related to EcoMcrA (38)C-terminal H-N-Hc 37 identical to EcoMcrA (38)
EcoKMcrBC GTP Rm5C(N30-35j)-(N30-3000)-Rm5Cb
McrB-N DUF3578 McrB binds DNA (40) via its N-terminal domain(59) by extrahelical modified base (60)
McrB-C P-loop NTPase (616263)McrC PD-(DE)XK Bioinformatic ID (64) Required for cleavage (3965)
EcoKMrr ndash m6A or m5C sequence specificity ambiguous(424366)
N-terminal Mrr-N Presumed DNA binding (67)C-terminal Mrr-cat (DE)
(DEQ) KBioinformatic ID (686970)
SauUS1 ATP or dATP Sm5CNGS two copies required for cleavage (45)N-terminal PLDc-2 (45)Middle P-loop NTPase (45)C-terminal DUF3427 (45) DBD
PvuRts1I PD-(DE)XK ndash mC(N11-13N9-10j)G 2-base extensions (47)Bioinformatic ID (64)
EcoCTGmrSD UTPgtgtGTP CTP Cuts T-even DNA (49)GmrS Motifs suggested DUF262 (71)To be confirmedGmrD DUF1524 To be determined
aRecognition sites (represented 5030) are those determined in vitro by binding or cleavage experimentsbMcrBC cleavage results in a double-strand cut near one Rm5C site (727374) but requires cooperation of two sites (3940) or a translocation block(73) The sites may be on different daughters across a fork (75) These are separated by 30ndash3000 (397274) and may be on either strand (3976)disposition of opposing nicks is not tightly constrained (73) and minor cleavage clusters are found 40 50 and 60 nt from the m5C (74)Degeneracy abbreviations B=C or G or T D=A or G or T H=A or C or T K=G or T M=A or C N=A or C or G or T R=A or GS=C or G V=A or C or G W=A or T Y=C or TCleavage positions are listed as (N to top cut to bottom cutj) If no number is listed the position of cleavage is not determined Space betweennumbers (eg PvuRts1I N11-13N9-10) indicates the range of positions at which cleavage may occur
Nucleic Acids Research 2014 Vol 42 No 1 59
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
nucleases and modification-blocked REases (8788)Interestingly two of the PLDc nuclease activities havebeen shown to work by a transesterification reaction likethat used by topoisomerases and transposases (8789)PvuRts1I has an apparently unusual nuclease domain
[ie not yet identified by sources curated by theNCBI Conserved Domain Database (90)] However thisenzyme was included in a categorization of PD-(DE)XKfamilies (64) a tentatively identified divalent metalion binding site Block B (47) corresponds to Block Dof Bujnicki and Rychlewski (64) Cleavage requiresMg2+ ions
EcoCTGmrSD Functional organization is less clearbut several possible nuclease motifs were identified inGmrS (71) Cleavage buffer contained Ca2+ and Mg2+
ions and UTP
Sequence context recognition
Many of the modification-dependent enzymescharacterized so far have little sequence specificity incontrast to conventional modification-blocked REasesRelatively complete characterization of sequence prefer-ence and cleavage position has been carried out for TypeIV enzymes EcoKMcrBC SauUSI and PvuRts1I(Table 3) and for Type IIM DpnI and the MspJI family(Table 4) Progress has been made with binding recogni-tion for EcoKMcrA Cleavage conditions have beenachieved for Sco3AMcrA (Table 3) For all of these rec-ognition of surrounding sequence context is degeneratewith preference for a neighboring base and frequently arequirement for two sites with suitable separation DpnI isin some respects an exception see later in the text
The remaining nucleases in Table 4 are less wellcharacterized The recognition sites might form a relatedseries It will be interesting to learn more about the rela-tionships among these and how the requirement formultiple modified positions is specified eg for BlsI andPkrI
McrA binding domainsThe two lsquoMcrArsquo enzymes are not similar in theirN-termini with homology limited to the C-terminalnuclease domain For EcoKMcrA (Figure 2) there isgood genetic evidence that base recognition lies in theN-terminus Extensive mutagenesis using insertion of
Table 4 Characteristics of other modification-dependent enzymes (Type IIM)
Protein Subunitsdomains
DNABinding
EndonucleaseDomain
Recognitionsite
Comment and references
DpnI family G m6AjTC 13 characterized isoschizomersN-terminal PD(DE)XK (50777879)C-terminal wHTH R m6AjTC (78)
MspJI family mC with preferences Second copy stimulates cleavageMspJI 5 mCNNR(N913j) (51)
N-terminal SRA-like (80)C-terminal Mrr-cat (DE)(DEQ)XK (51)
FspEI Like MspJI C m5C(N1216j) (52)LpnPI Like MspJI C m5CDG(N1014j) (52)AspBHI Like MspJI YS m5CNS(N812j) (52)RlaI Like MspJI V m5CW (52)SgrTI Like MspJI C m5CDS(N1014j) (52)SgeI Like MspJI m5CNNR(N913j) 49 identical to MspJI (53)No familyassigned
Information from httpwwwsibenzymecomproductsm2_type
AoxI jRG m5CYBisI G m5CjNGCBlsI RYNjR Y At least two m5C requiredGlaI R m5CjGY (54)GluI GmCjNG m5CKroI Gj m5CCGGCMteI GmCG m5CjNGm5CGm5CPcsI m5CG(N5jN2)m5CGPkrI Gm5CNjG m5C At least 3 m5C required
Figure 2 McrA functional domains Domain function was inferred in-directly from genetic analysis by Anton amp Raleigh (2004) (57) Manymutations in the N-terminal domain spared some activity in one ormore of three functional tests (grey segments) while others were defi-cient in all activities (black segments) One mutation (asterix) was fullyactive on m5C-containing substrates but fully inactive in the hm5Cchallenge in vivo Most mutations in the C-terminus (pale greysegment) retained function in one test that was interpreted asmeasuring m5C binding ability A predicted structural model byBujnicki Radlinska and Rychlewski (2000) (58) for this C-terminalregion is compatible with these results
60 Nucleic Acids Research 2014 Vol 42 No 1
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
five-amino acid linkers and classification with three func-tional tests allowed assignment of DNA recognition to theN-terminal portion with the C-terminal H-N-H domainimplicated in cleavage Of particular note a mutationdiscriminated in vivo between hm5C and m5C was foundin the N-terminal domain (57) The mutant was able tofully restrict bacteriophage lambda modified by MHpaIIbut not at all phage T4 containing hm5C In vitromodification-dependent binding was achieved with thefull-length His-tagged protein (5556) yielding a putativerecognition site (YgtR)mCG This recognition site iscompatible with in vivo observations (391)
Presumably the N-terminus of Sco3A McrA also rec-ognizes the DNA Recognition of both m5C and thephosphorothioate (PT) moiety must be accommodatedin the final reaction As either modification is sufficientto elicit cleavage more than one domain could beinvolved Cleavage occurred near some but not allDcm-modified sites (Cm5CWGG) Both synthetic
PT-containing oligonucleotides and unmethylated PT-modified plasmid were also cleaved on both sides of asymmetrically modified site PT modification is thoughtto be sequence-specific (2629) but the details are notyet clear
Novel McrB binding domainDNA-binding resides in the McrB N-terminal domain(405960) whereas translocation and cleavage coordin-ation reside in the C-terminal AAA+ regulatory andtranslocation domain (61ndash637273) The complete trans-lation product of mcrB McrB-L binds DNA specifically(4072) via its N-terminal 161 amino acids (aa) (59) Thecrystal structure of this domain (McrB-N) in complex withDNA has recently been published (60)McrB-N uses a strategy first discovered for DNA-MTase
action (92) it flips the C base out of the DNA helix into abinding pocket for inspection The pocket is large enoughto accommodate C m5C hm5C or m4C but too small if a
Figure 3 McrBC Assembly Model Two proteins are expressed from mcrB in vivo Both the complete protein (McrB-L) and a small one missing theN-terminus (McrB-S top row) bind GTP forming high-order multimers detected by gel filtration (second row) When visualized by scanningtransmission electron microscopy these appear as heptameric rings with a central channel Rings of McrB-L in top views show projections thatmay correspond to the N-terminal DNA-binding domain (red segment) Both forms can then associate with McrC judged again by gel filtrationMcrB-LGTP can bind to its specific substrate (RmC) in the absence of McrC (third row) in its presence the substrate is cleaved (fourth row)GTP hydrolysis is required for cleavage (arrow) a supershifted binding complex forms in the presence of GTP-gamma-S but no cleavage occursTranslocation accompanies GTP hydrolysis double-stranded cleavage requires collaboration between two complexes or a translocationblock The path of the DNA in the figure is arbitrary as is the conformation of McrC Modified from BourniquelAA and BickleTAComplex restriction enzymes NTP-driven molecular motors Bourniquel and Bickle (84) with permission Copyright 2002 Elsevier MassonSAS All rights reserved
Nucleic Acids Research 2014 Vol 42 No 1 61
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
glucose moiety is attached Conserved residues Y64 andL68 were noted to make van der Waals contact with themethyl group of the flipped out m5C these contacts aremissing when the pocket contains CThe flipping action can be compared with but is distinct
from that of eukaryotic m5C-specific regulatory proteinsthat use the SET and Ring-finger-Associated (SRA)domain to read DNA modification state (Figure 4A)This domain is found in most eukaryotes in accessoryproteins (eg UHRF1NP95SUVH5) of the DNMT1maintenance MTase (93ndash95) Despite the similarstrategy the McrB-N domain is not homologous butdisplays a distinct protein fold (60) Binding is accom-plished from the minor groove and extraction of the Ccreates a 30 bend toward the major groove resembling aglycosylase in this respect (Figure 4A) The eukaryoticproteins form a crescent from which loops project towrap around the DNA with recognition mediatedthrough both major and minor grooves (94) For McrB-N the authors suggest that the purine preference in the 50
position might result from flexibility constraints or inter-action with a non-conserved aa that occupies space left bythe flipped base Substitutions of this aa (Y41A or Y41Q)compromised binding activity
Sequence specificity novel phenotype and structuralmodel of MrrIn 1987 Heitman and Model discovered Mrr when theyfound that transfer of various foreign m6A MTasesinduced an SOS response due to DNA damage (42)This response to the presence of an incompatible MTaseremains the principal evidence that the E coli K12 Mrrprotein cleaves DNA Related proteins discussed later inthe text (Type IIM) have been more tractable for in vitrowork No concise description of the Mrr recognitionsequence has been forthcoming although several studieshave examined the spectrum of incompatible MTases(43669697) Both adenine and cytosine MTases confersensitivityMrr is also responsible for DNA damage that does not
depend on methylation at all foreign or otherwise Highhydrostatic pressure (HP) induces the SOS DNA damageresponse and lethality (98) The response did not dependon the activity of the endogenous MTases of E coli K12but did depend on both the presence of wild-type mrr andthe integrity of the SOS signal generation pathwayPossibly HP elicits a non-enzymatic modification or astructural change in DNA helicity that is acted on byMrr This HP phenotype was used to characterize mrrmutants which were fitted into a computer-assistedmodel of the Mrr protein (67) An N-terminal DNA-binding winged helix was proposed with a C-terminalnuclease domain previously identified (69) The functionalimportance of several conserved residues was confirmedSeveral of the selected mutants with null phenotypes wereisolated in a region far from the active site or bindingsurface identified bioinformatically These could affectinteraction with a component of the HP response Thisintriguing collection of informative mutants will illumin-ate in vitro characterization
Type IIM binding domainsType IIM enzymes of two families are well-characterizedwith respect to cleavage (Table 4) Crystal structures forboth have recently appeared
DpnI winged-helix DNA recognitionUnusually for modification-dependent enzymes DpnIcleaves a four-base site (Gm6ATC) with high fidelity(7799) to leave blunt ends when both strands of the siteare methylated At low concentration the enzyme nicksthe modified strand of a hemimethylated site (100) Thebehavior of the enzyme with respect to modificationpatterns within the canonical GATC sitemdashmodificationof C or A one strand or bothmdashhas been thoroughlyexplored (50) However only recently has cleavage ofnon-canonical adenine-methylated sites been examinedSiwek and co-workers (78) found evidence for consider-able relaxation of specificity at the outer base This experi-ment used substrates modified by a highly non-specificadenine MTase extensive DpnI cleavage cloning of thefragments and sequencing of the borders
Structure determination in the presence of DNA andvalidation experiments (78) place this enzyme togetherwith the other modification-dependent enzymes in thattwo domains segregate the cleavage function fromsequence recognition It also separates DpnI from theothers in that the cleavage domain also possesses somemodification and sequence specificity The main recogni-tion is accomplished by a monomeric winged-helixdomain which binds in the major groove and recognizesthe modifications on both strands in the same event Thestructure does not reveal a cleavage-competent complexhowever because the cleavage domain is far from theDNA Filter-binding experiments validated the ability ofthe C-terminal domain to bind alone to do so moretightly to fully methylated than to hemimethylated oligo-nucleotides and to compete with the full-length enzymereducing cleavage by it Expression of the N-terminalcleavage domain alone (in low yield) allowed validationof its cleavage activity Surprisingly this cleavage wasitself dependent on modification state and sequence ofthe substrate Modeling based on the structure of theblunt-end-producing Type IIP enzyme PvuII allowed pre-diction that the cleavage domain approaches from theminor groove Complete understanding of double-stranded cleavage will depend on understanding thedynamic transformations that allow the cleavage domainto approach and act at the site
MspJI coupling of cleavage with DNA recognitionThe six members of the MspJI family use the Mrr-catversion of the PD-(DE)XK nuclease to cut at definedlocations to one side relative to the modified base(12 bases on the modified strand 16ndash17 on the otherFigure 5A) only one modified base is required fordouble-strand cleavage to occur (unlike McrBC) (5152)However these enzymes are stimulated by the presence ofa second site in cis or in trans Symmetrically modifiedsites (such as m5CpGm5CpG in mammalian DNA)yield prominent bands of defined size (32 bp) containinga mixed population of sequences each with a m5C in the
62 Nucleic Acids Research 2014 Vol 42 No 1
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
middle (Figure 5B) This behavior is recapitulated by thePvuRts1I group of enzymes (exemplified by AbaSDFI inFigure 5C) except that the distances are shorter andrecognition of modification state is less well understoodDuring the characterization of MspJI Dcm
(Cm5CWGG) sites were the first recognized substrateyielding a clear banding pattern (51) Cleavage of differ-ently modified plasmids and designed oligonucleotidesubstrates allowed a good assessment of both modificationand sequence specificity This family shows preference forparticular bases nearby similar to McrBC
MspJI DNA recognition is mediated by anSRA-like domainRecently the crystal structure of MspJI without DNA hasbeen resolved at 205 A (80) Search of the MolecularModeling Database at NCBI (101) using VAST (102)
Figure 4 McrB-N in comparison to other base-flipping proteins (A)SRA domains SUVH5 (3Q0C) and UHRF1 (2ZKD) use loops extend-ing from a crescent formed from two beta sheets to flip C or m5C fromundeformed B-form DNA into a pocket (top row) whereas McrB-N(3SSC bottom row) uses loops from one beta-sheet to distort the DNAand flip the base It resembles the human alkyladenine glycosylase(1BNK) (bottom row) in bending the DNA toward the majorgroove while flipping the base via the minor groove Figure 5 ofSukackaite et al (60) (B) The SRA-like hemi-methylated 5mC recog-nition domains A ribbon model of the N-terminal domain of theMspJI structure (4F0Q and 4F0P left) compared with the SRAdomain of URHF1 (PDB 3FDE right) The crescent shape formed
Figure 4 Continuedby interacting beta sheets and helices aB and aC are the conservedfeatures of the SRA domain highlighted here Loops on the concaveside of UHRF1 participate in flipping the base and similar loops pre-sumably do so for MspJI Two of these vary in length among familymembers and may play roles in sequence context specificity Figure 2aand b from Horton et al (80)
Figure 5 Schematic diagrams of cleavage positions for MspJI andAbaSDFI Cleavage of both strands is elicited by a singly modifiedsite for both MspJI (A) and AbaSDFI (C) Cleavage position is fixedrelative to the modified site but with a four-base 50 extension forMspJI and a two-base 30 extension for AbaSDFI When a site is sym-metrically-modified (as for CpG sites in mammalian DNA) a 32 base-pair fragment is excised from the DNA (B) (A) Figure 2a and (B)Figure 3a reprinted with permission from Cohen-Karni et al (52)(C) Figure 5a from Wang et al (47)
Nucleic Acids Research 2014 Vol 42 No 1 63
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
showed that the N-terminal domain was structurallysimilar to that of the eukaryotic SRA domain with acrescent-shaped beta-sheet structure from which loopsproject (see Figure 4B and discussion earlier in the textMcrB) This structural homology allowed modeling of theDNA-bound structure with a flipped m5C The enzyme inthe crystal is a tetramer in which two monomers form aback-to-back dimer via the C-terminal regions thatcomprise the endonuclease Two back-to-back dimersgenerate a tetrameric protein with two cleavage domainspositioned (as in the Type IIP enzyme HindIII used formodeling the C-terminal cleavage domain interaction withDNA) so that a 4-base 50 extension would be created oncleavage of modeled DNA Cleavage is most efficient atmolar ratios that allow all four SRA-like domains to beoccupiedmdashtoo much enzyme prevents cleavage fromoccurring
Tracking and dimerization
McrBC as translocaseBourniquel and Bickle (84) have reviewed much of theenzymology of McrBC which will be briefly summarizedhere The Raleigh Bickle and Pingoud laboratories havecontributed to the following consistent picture of thein vitro reaction EcoKMcrBC cleavage results in adouble-strand cut near one RmC site (72ndash74) butrequires cooperation of two sites (3940) or a translocationblock (73) The sites may be on different daughters acrossa fork (75) These are separated by 30ndash3000 bp (397274)and may be on either strand (3976) cleavage occurs30ndash35 bases from the modified base with oppositenicks not tightly constrained (73) and minor cleavageclusters are found 40 50 and 60 nt from the m5C(74) hm5C DNA elicits cleavage also (39) A ring struc-ture is formed by 5ndash7 molecules of McrB in the presence ofGTP (Figure 3) (103) this complex can bind to a recog-nition element in DNA In the presence of McrC trans-location of the complex occurs and cleavage ensues whentranslocation is blocked Collision of translocatingcomplexes a protein barrier or a topological barrier willelicit double-strand cleavage adjacent to one recognitionelement or the other The enzyme will cleave when recog-nition elements are on opposite sides of a forked structure(75) This would allow action in vivo to prevent entry of aMTase gene even with rare sitesStructurally the McrB protein is proposed to be a
member of the AAA+ protein family of NTPases (104)many of which form ring-shaped complexes and partici-pate in molecular machines lsquoSensorrsquo segments found inthese proteins have been shown in some cases to playroles in coupling NTPase activity to intersubunit commu-nication and movement (105) Two of three elements ofthe GTP-binding motif proposed by Dila et al (61) werevalidated by mutational analysis (6562) The thirdproposed motif element was identified as amino acidsNTAD by Dila et al Alignment of AAA+ NTPases in(104) found this aligned with the motif designatedSensor-1 in (105) An interesting result was that mutationshere unexpectedly appeared to abrogate interaction withMcrC instead of changing which NTP would be
productive (62) It may instead play a role in coordinatingGTP binding and hydrolysis with DNA binding inter-action with McrC and cleavage
Intracellularly the story becomes more complex as themcrB gene encodes two products of 51 and 33 kD McrB-L and McrB-S the latter one starting from an in-frameinternal translation start site (106) Both in vivo andin vitro McrB-S can interfere with the function ofMcrB-L at least in part by forming complexes withMcrC unable to bind DNA (107) Both species can formmultimeric rings in the presence of GTP (103) as is usualfor AAA+NTPases (104)
SauUSI requires two sites and ATP hydrolysisSauUSI was originally annotated as a putative helicasefrom Staphylococcus aureus sp A single polypeptide issufficient for activity both in vitro and also in vivo as aclone in E coli using modified phage as a challenge Theamino acid sequence contains a PLDc domain at theN-terminus This contains a phosphodiesterase motiforiginally identified in Phospholipase D (108) it wasvalidated by mutagenesis of four catalytic residue candi-dates In the middle ATPase and helicase motifs wereproposed to account for ATP dependence of cleavageactivity A Domain of Unknown Function was identifiedat the C-terminus (Pfam DUF3427) (108) and wasproposed to recognize the substrate (45)
The purified enzyme cleaves modified DNA containingm5C and hm5C but not m4C in the presence of ATP ordATP but not other nucleotides The negative result form4C is firm plasmids modified at the same site by an m5CMTase (Dcm) or an m4C MTase (MPspGI) were testedThe former (Cm5CWGG) was sensitive whereas the latterwas resistant Thus the sequence preference is likely tobe satisfied m6A is likely not a substrate but few m6A-containing sites were examined
Like McrBC SauUSI requires the presence of two sitesfor efficient cleavage Presumably the ATPase activityparticipates in monitoring the presence of two sites asfor other nucleotide-dependent REases includingMcrBC The mechanism of communication is unknownThe enzyme belongs to a family of highly similarorthologues found in other sequenced Staphylococci(Tables 1 and 2) and more distant homologues can befound in sequenced bacterial and archaeal genomes
EVOLUTIONARY PRESSURES ON R-M SYSTEMS
Evolution by selfish propagation
One way to understand the massive variety of restrictionsystems and their sporadic distribution is to locate theevolutionary drivers of enzyme diversification in theenzyme genes themselves as selfish elements Work fromthe Kobayashi laboratory has elaborated clear examplesof selfish behavior in some Type II enzyme systems(109110) in which the host becomes lsquoaddictedrsquo to theR-M system Once a cell has acquired an R-M pair lossof the genes results in death of that cellrsquos descendants asthe REase is frequently still present and able to act on thegenome following loss of methylation activity In this
64 Nucleic Acids Research 2014 Vol 42 No 1
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
perspective the role played by modification-dependentenzymes is host defense to exclude systems with lsquoforeignrsquoMTase patterns and prevent the cell from loading up withparasites The exclusion event is accompanied by the deathof the cell (111ndash113) Weak sequence specificity of Type IVenzymes could then result from the need to control entryof a wide variety of invading systems
The selfish aspect certainly plays a role in R-M popula-tion biology but cannot be the whole story Type II R-Msystems can still be lost by inactivation of the R gene firstMoreover Type I systems escape this scenario withcomplex control of cleavage activity the restrictionassembly includes a methylation assembly to begin withtherefore the R protein cannot act unless the MTase ispresent in addition failure of the methylation activity inan intact complex leads to abrogation of R activity some-times by action of the ClpXP protease specifically on theR protein (114ndash116)
Furthermore in population terms a cell that acquiredand became addicted to an R-M system should lose incompetition with a sibling that never received thesystem Two factors could counter this First acquisitioncould be accompanied by an increase in the total numberof copies of the R-M system in the population asproposed for invading transposable elements Thisoverreplication results in more copies of the systemcreated than are lost whether to suicide or to other select-ive disadvantage [see eg (117118)] R-M gene amplifica-tion within a cell has been reported experimentally (119)but spread in a population has not been demonstrated yetA second factor that could counter the disability of addic-tion is localization of competition In a structured envir-onment (colonies on a plate or biofilm on large or smallsurfaces) killing of segregants preserves limiting nutrientsfor lineages that retain the toxinantitoxin pair (120121)Much of the real world is structured so this is an import-ant condition
Evolution by phage-host arms race
A second perspective supposes that the modification-dependent Type IV enzymes arose from the competitivecoevolutionary interaction between phages and theirhosts This was first enunciated by Revel and Luria (2)and most recently by Black and coworkers (122) seealso (123) That is hosts used modification-blockedrestriction to defend against phage infection T-evenphages developed methods of substituting modified basesfor the ordinary ones hosts developed Type IV enzymes indefense phages added sugar or other modifications (19) tothwart Type IV enzymes hosts extended Type IV enzymesto accommodate these decorations finally phages de-veloped protein inhibitors specific for these enzymes aswell T4 phages deliver a protein inhibitor (IPI) alongwith the DNA on infection which allows growth in thepresence of EcoCTGmrSD The locus responsible for thisinhibitor is highly variable among relatives of T4 asgmrSD is in enteric bacteria (both in distribution and inaa sequence) When phage with different IP1 alleles weretested for protection from cloned EcoCTGmrSD and itshomolog EcoUTGmrSD specificity was evident one or
the other or both or neither of the two homologs wascounteracted in individual cases (122) This variability ofthe outcome supports the idea that phage-host interactiondrives at least some of these developmentsIn this perspective the weak sequence selectivity of the
Type IV systems might simply reflect the lack of endogen-ous targets for the enzyme As the host does not present anyhm5C and the phage is completely substituted selection forsequence-specificity is weak Selection would act to spareany co-residentMTases This differs fromType II enzymeswhere the MTase and REase must co-evolve to allow thehost to survive Each Type IV system is compatible withsome suite of Type I-IIIMTases (and thus the R-M systemsas a whole) Methylated or hydroxymethylated bases maynot be recognized at all (EcoCTGmrSD) or the systemmay require one specific base in addition to the modifiedone (McrA McrBC MspJI and PvuRts1I) MTases mod-ifying sites not including that base are then compatible asDcm (CmCWGG) is compatible with the McrA McrBCand Mrr systems in E coli K12Type IV systems that restrict methylated bases in a
weakly specified sequence context confer an additionaladvantage in competition with phages Many phagessuch as have not evolved the nucleotide-substitutionstrategy used by the T-even phages These phagesnormally carry the modification pattern of the mostrecent host if the last host expressed an MTase creatinga susceptible site the Type IV enzyme of the new host willdestroy the invader and limit the infection This may beaccompanied by the death of the individual infectedtherefore protection can be conferred on the sibling popu-lation (111)A further implication of this scenario considers the fate
of a population invaded by phage Phage survival ofrestriction occurs at biologically relevant frequencies(106-102) The survivors of restriction carry the particu-lar methylation pattern of the particular cell and thus areresistant to all restriction systems it might have carried(Types IndashIV) A bacterial population as a whole thenbenefits from mechanisms that diversify the suite ofR-M systems so that such surviving phage do not havefree access to the entire population The extreme variabil-ity of R-M system content in isolates of the same species iscompatible with this idea [see eg (124) REBASEGenomes httptoolsnebcomvinczegenomes] Suchvariability also limits and shapes interstrain genetransfer (115125126)The Raleigh laboratory has built on elegant genetic
work with Type I enzymes in the Murray laboratory(127ndash130) to investigate a locus designated thelsquoImmigration Control locusrsquo or ICR that exemplifiesvariable R-M content Alternative DNA segments con-taining R-M systems are located at the same definedlocation in most E coli chromosomes between the yjiSand yjiA genes The ICR in the non-restricting E coli Cstrain [used in the original definition of the R-M phenom-enon (131)] contains a remnant of a Type I enzyme R geneand is 13 kilobase shorter than the same region in E coliK12 (132) The mechanism of segment replacement is stillunknown The ICR would be an example of the lsquodefenseislandsrsquo analyzed by the Koonin group (133) lsquoDefense
Nucleic Acids Research 2014 Vol 42 No 1 65
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
islandsrsquo contain genes that can defend against phage orother invading DNA these exhibit bioinformaticproperties similar to lsquomobilome islandsrsquo containing mobil-ization genes (transposases for example) However themechanism of mobilization has not been identified forthe ICR
FINAL THOUGHTS
The extreme diversity of R-M systems that recognizeordinary DNA seems likely to be approached by the di-versity of Type IV restriction systems Type IV enzymesare hard to find as most detection methods depend ondevelopment of genetic systems for each taxon or on ser-endipity Those characterized so far mostly stem frominitial genetic investigation of limits on infection trans-formation or transduction Barriers encountered provideleads to the genes involved Bioinformatic analysis hashelped to identify relatives which may be more tractableto biochemical investigation than the example originallyfound This approach has pitfalls the gene encodingMspJI was first thought to code for an enzyme recognizingan unmodified site because it is immediately adjacent toan (inactive it is now thought) cytosine MTase geneProvidentially the first expression host was devoid of sen-sitive sites whereas the first test substrate contained some(51) A combination of biological experiments with bio-informatics and biochemistry will be needed to reveal thefull spectrum of Type IV enzymes that may lurk within thevast universe of unidentified ORFs in bacterial systemsOne might begin with those strains whose genomes carryfew Type II systems Bacillus or Corynebacterium asopposed to Helicobacter or Neisseria [see the Genomessection of REBASE (50)]The role of lsquodefense islandsrsquo and their relation to the
lsquomobilomersquo in bacterial population biology remains to bedetermined If a defense island is similar to a mobilomeisland there should be a mechanism of mobilizationnearby which would boost the contribution oflsquooverreplicationrsquo to the account of selections acting onR-M systems R-M systems of all types can be found onor adjacent to known mobilizing elements (134135) buthave not been shown to move experimentallyOn another note it may turn out that evolutionarily
there is a continuum between the apparently modifica-tion-dependent and modification-blocked paths Onerelative of McrBC predicted by bioinformatics analysesis LlaI a system that recognizes an unmodified sequenceand requires two MTases to support it (136) The enzymeBamHI prefers to cleave DNA with m6A in its GGATCCsite and mutants can be isolated that require this modifiedbase (137) Are there native systems similarly protected bymodification of one position in the recognition site butdependent on modification at a different one An interest-ing evolutionary series can be imagined
ACKNOWLEDGEMENTS
The authors thank Rich Roberts and David Dryden forencouragement and critical comment They are grateful to
Tom Bickle Virginius Siksnys Yu Zheng and XiadongCheng for permission to reproduce figures They alsothank Ivan Correa for help drawing base structures andTasha Jose for illustration assistance They also thank theanonymous reviewers of a previous manuscript for caretaken with the review process greatly improving thepresent document
FUNDING
New England Biolabs (EAR) Funding for open accesscharge New England Biolabs (to EAR)
Conflict of interest statement None declared
REFERENCES
1 LuriaSE and HumanML (1952) A nonhereditary host-inducedvariation of bacterial viruses J Bacteriol 64 557ndash559
2 RevelHR and LuriaSE (1970) DNA-glucosylation in T-evenphage genetic determination and role in phage-host interactionAnnu Rev Genet 4 177ndash192
3 RaleighEA and WilsonG (1986) Escherichia coli K-12 restrictsDNA containing 5-methylcytosine Proc Natl Acad Sci USA83 9070ndash9074
4 Noyer-WeidnerM DiazR and ReinersL (1986) Cytosine-specific DNA modification interferes with plasmid establishmentin Escherichia coli K12 involvement of rglB Mol Gen Genet205 469ndash475
5 BlumenthalRM GregorySA and CooperiderJS (1985)Cloning of a restriction-modification system from Proteus vulgarisand its use in analyzing a methylase-sensitive phenotype inEscherichia coli J Bacteriol 164 501ndash509
6 SohailA LiebM DarM and BhagwatAS (1990) A generequired for very short patch repair in Escherichia coli is adjacentto the DNA cytosine methylase gene J Bacteriol 1724214ndash4221
7 MacNeilDJ (1988) Characterization of a unique methyl-specificrestriction system in Streptomyces avermitilis J Bacteriol 1705607ndash5612
8 VertesAA InuiM KobayashiM KurusuY and YukawaH(1993) Presence of mrr- and mcr-like restriction systems incoryneform bacteria Res Microbiol 144 181ndash185
9 MacalusoA and MettusAM (1991) Efficient transformation ofBacillus thuringiensis requires nonmethylated plasmid DNAJ Bacteriol 173 1353ndash1356
10 HolmesML NuttallSD and Dyall-SmithML (1991)Construction and use of halobacterial shuttle vectors and furtherstudies on Haloferax DNA gyrase J Bacteriol 173 3807ndash3813
11 PosfaiJ BhagwatAS PosfaiG and RobertsRJ (1989)Predictive motifs derived from cytosine methyltransferases NucleicAcids Res 17 2421ndash2435
12 MaloneT BlumenthalRM and ChengX (1995) Structure-guided analysis reveals nine sequence motifs conserved amongDNA amino-methyltransferases and suggests a catalyticmechanism for these enzymes J Mol Biol 253 618ndash632
13 TiminskasA ButkusV and JanulaitisA (1995) Sequence motifcharacteristics for DNA [cytosine-N4] and DNA [adenine-N6]methyltransferases Classification of all DNA methyltransferasesGene 157 3ndash11
14 DickmanMJ InglestonSM SedelnikovaSE RaffertyJBLloydRG GrasbyJA and HornbyDP (2002) The RuvABCresolvasome Eur J Biochem 269 5492ndash5501
15 KovallRA and MatthewsBW (1998) Structural functionaland evolutionary relationships between lambda-exonuclease andthe type II restriction endonucleases Proc Natl Acad Sci USA95 7893ndash7897
16 RobertsRJ ChangYC HuZ RachlinJN AntonBPPokrzywaRM ChoiHP FallerLL GuleriaJ HousmanGet al (2010) COMBREX a project to accelerate the functional
66 Nucleic Acids Research 2014 Vol 42 No 1
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
annotation of prokaryotic genomes Nucleic Acids Res 39D11ndashD14
17 RaePM and SteeleRE (1978) Modified bases in the DNAs ofunicellular eukaryotes an examination of distributions andpossible roles with emphasis on hydroxymethyluracil indinoflagellates Biosystems 10 37ndash53
18 WarrenRAJ (1980) Modified bases in bacteriophage DNAsAnnu Rev Microbiol 34 137ndash158
19 Gommers-AmptJH and BorstP (1995) Hypermodified bases inDNA FASEB J 9 1034ndash1042
20 SwintonD HattmanS CrainPF ChengCS SmithDL andMcCloskeyJA (1983) Purification and characterization of theunusual deoxynucleoside alpha-N-(9-beta-D-20-deoxyribofuranosylpurin-6-yl)glycinamide specified by the phageMu modification function Proc Natl Acad Sci USA 807400ndash7404
21 TahilianiM KohKP ShenY PastorWA BandukwalaHBrudnoY AgarwalS IyerLM LiuDR AravindL et al(2009) Conversion of 5-methylcytosine to 5-hydroxymethylcytosinein mammalian DNA by MLL partner TET1 Science 324930ndash935
22 KriaucionisS and HeintzN (2009) The nuclear DNA base5-hydroxymethylcytosine is present in Purkinje neurons and thebrain Science 324 929ndash930
23 BorstP and SabatiniR (2008) Base J discovery biosynthesisand possible functions Annu RevMicrobiol 62 235ndash251
24 KaminskaKH and BujnickiJM (2008) Bacteriophage MuMom protein responsible for DNA modification is a new memberof the acyltransferase superfamily Cell Cycle 7 120ndash121
25 IyerLM TahilianiM RaoA and AravindL (2009) Predictionof novel families of enzymes involved in oxidative and othercomplex modifications of bases in nucleic acids Cell Cycle 81698ndash1710
26 ZhouX HeX LiangJ LiA XuT KieserT HelmannJDand DengZ (2005) A novel DNA modification by sulphur MolMicrobiol 57 1428ndash1438
27 OuHY HeX ShaoY TaiC RajakumarK and DengZ(2009) dndDB A Database Focused on Phosphorothioation ofthe DNA Backbone PloS One 4 e5132
28 XuT LiangJ ChenS WangL HeX YouD WangZLiA XuZ ZhouX et al (2009) DNA phosphorothioation inStreptomyces lividans mutational analysis of the dnd locusBMC Microbiol 9 41
29 WangL ChenS VerginKL GiovannoniSJ ChanSWDeMottMS TaghizadehK CorderoOX CutlerMTimberlakeS et al (2011) DNA phosphorothioation iswidespread and quantized in bacterial genomes Proc Natl AcadSci USA 108 2963ndash2968
30 Loslashbner-OlesenA SkovgaardO and MarinusMG (2005) Dammethylation coordinating cellular processes Curr OpinMicrobiol 8 154ndash160
31 LowDA and CasadesusJ (2008) Clocks and switches bacterialgene regulation by DNA adenine methylation Curr OpinMicrobiol 11 106ndash112
32 FreitagM and SelkerEU (2005) Controlling DNA methylationMany roads to one modification Curr Opin Genet Dev 15191ndash199
33 ProhaskaSJ StadlerPF and KrakauerDC (2010) Innovationin gene regulation The case of chromatin computation J TheorBiol 265 27ndash44
34 SeshasayeeASN SinghP and KrishnaS (2012) Context-dependent conservation of DNA methyltransferases in bacteriaNucleic Acids Res 40 7066ndash7073
35 AntonBP (2011) Dissertation Boston University BostonMassachusetts USA
36 RaleighEA TrimarchiR and RevelH (1989) Genetic andphysical mapping of the mcrA (rglA) and mcrB (rglB) loci ofEscherichia coli K-12 Genetics 122 279ndash296
37 Gonzalez-CeronG Miranda-OlivaresOJ and Servın-GonzalezL(2009) Characterization of the methyl-specific restriction system ofStreptomyces coelicolor A3(2) and of the role played by laterallyacquired nucleases FEMS Microbiol Lett 301 35ndash43
38 LiuG OuHY WangT LiL TanH ZhouXRajakumarK DengZ and HeX (2010) Cleavage of
phosphorothioated DNA and methylated DNA by the type IVrestriction endonuclease ScoMcrA PLoS Genet 6 e1001253
39 SutherlandE CoeL and RaleighEA (1992) McrBC amultisubunit GTP-dependent restriction endonuclease J MolBiol 225 327ndash348
40 KrugerT WildC and Noyer-WeidnerM (1995) McrB aprokaryotic protein specifically recognizing DNA containingmodified cytosine residues EMBO J 14 2661ndash2669
41 SitaramanR and LepplaSH (2012) Methylation-dependentDNA restriction in Bacillus anthracis Gene 494 44ndash50
42 HeitmanJ and ModelP (1987) Site-specific methylases inducethe SOS DNA repair response in Escherichia coli J Bacteriol169 3243ndash3250
43 Waite-ReesPA KeatingCJ MoranLS SlatkoBEHornstraLJ and BennerJS (1991) Characterisation andexpression of the Escherichia coli Mrr restriction systemJ Bacteriol 173 5207ndash5219
44 KerrAL JeonYJ SvensonCJ RogersPL and NeilanBA(2010) DNA restriction-modification systems in the ethanologenZymomonas mobilis ZM4 Appl MicrobiolBiotech 89 761ndash769
45 XuSY CorvagliaAR ChanSH ZhengY and LinderP(2011) A type IV modification-dependent restriction enzymeSauUSI from Staphylococcus aureus subsp aureus USA300Nucleic Acids Res 39 5597ndash5610
46 MonkIR ShahIM XuM TanMW and FosterTJ (2012)Transforming the untransformable application of directtransformation to manipulate genetically Staphylococcus aureusand Staphylococcus epidermidis mBio 3 e00277ndash00211
47 WangH GuanS QuimbyA Cohen-KarniD PradhanSWilsonG RobertsRJ ZhuZ and ZhengY (2011)Comparative characterization of the PvuRts1I family ofrestriction enzymes and their application in mapping genomic5-hydroxymethylcytosine Nucleic Acids Res 39 9294ndash9305
48 JanosiL YonemitsuH HongH and KajiA (1994) Molecularcloning and expression of a novel hydroxymethylcytosine-specificrestriction enzyme (PvuRts1I) modulated by glucosylation ofDNAJ Mol Biol 242 45ndash61
49 BairCL and BlackLW (2007) A type IV modificationdependent restriction nuclease that targets glucosylatedhydroxymethyl cytosine modified DNAs J Mol Biol 366768ndash778
50 RobertsRJ VinczeT PosfaiJ and MacelisD (2010)REBASEndasha database for DNA restriction and modificationenzymes genes and genomes Nucleic Acids Res 38 D234ndashD236
51 ZhengY Cohen-KarniD XuD ChinHG WilsonGPradhanS and RobertsRJ (2010) A unique family of Mrr-likemodification-dependent restriction endonucleases Nucleic AcidsRes 38 5527ndash5534
52 Cohen-KarniD XuD AponeL FomenkovA SunZDavisPJ KinneySRM Yamada-MabuchiM XuSYDavisT et al (2011) The MspJI family of modification-dependent restriction endonucleases for epigenetic studies ProcNatl Acad Sci USA 108 11040ndash11045
53 LubysA VitkuteJ LubieneJ and JanulaitisA (2011)Restriction enzymes and their applications US Patent ApplicationPublication 20110207139 A1 httpappftusptogovnetahtmlPTOsearch-boolhtml (27 August 2013 date last accessed)
54 TarasovaGV NayakshinaTN and DegtyarevSK (2008)Substrate specificity of new methyl-directed DNA endonucleaseGlaI BMC Mol Biol 9 7
55 MulliganEA and DunnJJ (2008) Cloning purification andinitial characterization of E coli McrA a putative5-methylcytosine-specific nuclease Protein ExprPurif 62 98ndash103
56 MulliganEA HatchwellE McCorkleSR and DunnJJ (2010)Differential binding of Escherichia coli McrA protein to DNAsequences that contain the dinucleotide m5CpG Nucleic AcidsRes 38 1997ndash2005
57 AntonBP and RaleighEA (2004) Transposon-mediated linkerinsertion scanning mutagenesis of the Escherichia coli McrAendonuclease J Bacteriol 186 5699ndash5707
58 BujnickiJM RadlinskaM and RychlewskiL (2000) Atomicmodel of the 5-methylcytosine-specific restriction enzyme McrAreveals an atypical zinc finger and structural similarity tobetabetaalphaMe endonucleases Mol Microbiol 37 1280ndash1281
Nucleic Acids Research 2014 Vol 42 No 1 67
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
59 GastFU BrinkmannT PieperU KrugerT NoyerWeidnerM and PingoudA (1997) The recognition of methylatedDNA by the GTP-dependent restriction endonuclease McrBCresides in the N-terminal domain of McrB Biol Chem 378975ndash982
60 SukackaiteR GrazulisS TamulaitisG and SiksnysV (2012)The recognition domain of the methyl-specific endonucleaseMcrBC flips out 5-methylcytosine Nucleic Acids Res 407552ndash7562
61 DilaD SutherlandE MoranL SlatkoB and RaleighEA(1990) Genetic and sequence organization of the mcrBC locus ofEscherichia coli K-12 J Bacteriol 172 4888ndash4900
62 PieperU BrinkmannT KrugerT Noyer WeidnerM andPingoudA (1997) Characterization of the interaction between therestriction endonuclease McrBC from E coli and its cofactorGTP J Mol Biol 272 190ndash199
63 PieperU SchweitzerT GrollDH GastFU and PingoudA(1999) The GTP-binding domain of McrB more than just avariation on a common theme J Mol Biol 292 547ndash556
64 BujnickiJM and RychlewskiL (2001) Grouping together highlydiverged PD-(DE)XK nucleases and identification of novelsuperfamily members using structure-guided alignment ofsequence profiles J Mol MicrobiolBiotech 3 69ndash72
65 PieperU and PingoudA (2002) A mutational analysis of thePD DEXK motif suggests that McrC harbors the catalyticcenter for DNA cleavage by the GTP-dependent restrictionenzyme McrBC from Escherichia coli Biochemistry 415236ndash5244
66 KelleherJE and RaleighEA (1991) A novel activity inEscherichia coli K-12 that directs restriction of DNA modified atCG dinucleotides J Bacteriol 173 5220ndash5223
67 OrlowskiJ MebrhatuMT MichielsCW BujnickiJM andAertsenA (2008) Mutational analysis and a structural model ofmethyl-directed restriction enzyme Mrr Biochem Biophys ResCommun 377 862ndash866
68 AravindL MakarovaKS and KooninEV (2000) SURVEYAND SUMMARY holliday junction resolvases and relatednucleases identification of new families phyletic distribution andevolutionary trajectories Nucleic Acids Res 28 3417ndash3432
69 BujnickiJM and RychlewskiL (2001) Identification of a PD-(DE)XK-like domain with a novel configuration of the endonucleaseactive site in the methyl-directed restriction enzyme Mrr and itshomologs Gene 267 183ndash191
70 SteczkiewiczK MuszewskaA KnizewskiL RychlewskiL andGinalskiK (2012) Sequence structure and functional diversity ofPD-(DE)XK phosphodiesterase superfamily Nucleic Acids Res40 7016ndash7045
71 BairCL RifatD and BlackLW (2007) Exclusion of glucosyl-hydroxymethylcytosine DNA containing bacteriophages isovercome by the injected protein inhibitor IPI J Mol Biol 366779ndash789
72 StewartFJ and RaleighEA (1998) Dependence of McrBCcleavage on distance between recognition elements Biol Chem379 611ndash616
73 PanneD RaleighEA and BickleTA (1999) The McrBCendonuclease translocates DNA in a reaction dependent on GTPhydrolysis J Mol Biol 290 49ndash60
74 PieperU GrollDH WunschS GastFU SpeckC MuckeNand PingoudA (2002) The GTP-dependent restriction enzymeMcrBC from Escherichia coli forms high-molecular masscomplexes with DNA and produces a cleavage patternwith a characteristic 10-base pair repeat Biochemistry 415245ndash5254
75 IshikawaK HandaN SearsL RaleighEA and KobayashiI(2011) Cleavage of a model DNA replication fork by amethyl-specific endonuclease Nucleic Acids Res 39 5489ndash5498
76 StewartFJ PanneD BickleTA and RaleighEA (2000)Methyl-specific DNA binding by McrBC a modification-dependent restriction enzyme J Mol Biol 298 611ndash622
77 LacksS and GreenbergB (1977) Complementary specificity ofrestriction endonucleases DpnI and DpnII with respect to DNAmethylation J Mol Biol 114 153ndash168
78 SiwekW CzapinskaH BochtlerM BujnickiJM andSkowronekK (2012) Crystal structure and mechanism
of action of the N6-methyladenine-dependent type IIMrestriction endonuclease RDpnI Nucleic Acids Res 407563ndash7572
79 LacksS and GreenbergB (1975) A deoxyribonuclease ofDiplococcus pneumoniae specific for methylated DNA J BiolChem 250 4060ndash4066
80 HortonJR MabuchiMY Cohen-KarniD ZhangXGriggsRM SamaranayakeM RobertsRJ ZhengY andChengX (2012) Structure and cleavage activity of the tetramericMspJI DNA modification-dependent restriction endonucleaseNucleic Acids Res 40 9763ndash9773
81 CymermanIA ObarskaA SkowronekKJ LubysA andBujnickiJM (2006) Identification of a new subfamily of HNHnucleases and experimental characterization of a representativemember HphI restriction endonuclease Proteins 65 867ndash876
82 GorbalenyaAE (1994) Self-splicing group I and group II intronsencode homologous (putative) DNA endonucleases of a newfamily Protein Sci 3 1117ndash1120
83 ChanSH OpitzL HigginsL OrsquoLoaneD and XuSY (2010)Cofactor requirement of HpyAV restriction endonuclease PloSOne 5 e9071
84 BourniquelAA and BickleTA (2002) Complex restrictionenzymes NTP-driven molecular motors Biochimie 84 1047ndash1059
85 OrsquoDriscollJ HeiterDF WilsonGG FitzgeraldGFRobertsR and van SinderenD (2006) A genetic dissection ofthe LlaJI restriction cassette reveals insights on a novelbacteriophage resistance system BMC Microbiol 6 40
86 SmithRM DiffinFM SaveryNJ JosephsenJ andSzczelkunMD (2009) DNA cleavage and methylation specificityof the single polypeptide restriction-modification enzyme LlaGINucleic Acids Res 37 7206ndash7218
87 InterthalH PouliotJJ and ChampouxJJ (2001) The tyrosyl-DNA phosphodiesterase Tdp1 is a member of the phospholipaseD superfamily Proc Natl Acad Sci USA 98 12009ndash12014
88 SasnauskasG HalfordSE and SiksnysV (2003) How the BfiIrestriction enzyme uses one active site to cut two DNA strandsProc Natl Acad Sci USA 100 6410ndash6415
89 SasnauskasG ConnollyBA HalfordSE and SiksnysV (2007)Site-specific DNA transesterification catalyzed by a restrictionenzyme Proc Natl Acad Sci USA 104 2115ndash2120
90 Marchler-BauerA ZhengC ChitsazF DerbyshireMKGeerLY GeerRC GonzalesNR GwadzM HurwitzDILanczyckiCJ et al (2012) CDD conserved domains andprotein three-dimensional structure Nucleic Acids Res 41D348ndashD352
91 KravetzAN ZakharovaMV BeletskayaIV SinevaEVDenjmuchametovMM PetrovSI GlatmanLI andSoloninAS (1993) The cleavage sites and localization of genesencoding the restriction endonucleases Eco1831I and EcoHIGene 129 153ndash154
92 KlimasauskasS KumarS RobertsRJ and ChengX (1994)HhaI methyltransferase flips its target base out of the DNA helixCell 76 357ndash369
93 HashimotoH VertinoPM and ChengX (2010) Molecularcoupling of DNA methylation and histone methylationEpigenomics 2 657ndash669
94 RajakumaraE LawJA SimanshuDK VoigtPJohnsonLM ReinbergD PatelDJ and JacobsenSE (2011)A dual flip-out mechanism for 5mC recognition by theArabidopsis SUVH5 SRA domain and its impact on DNAmethylation and H3K9 dimethylation in vivo Genes Dev 25137ndash152
95 SharifJ and KosekiH (2011) Recruitment of Dnmt1 roles of theSRA protein Np95 (Uhrfi) and other factors Prog Mol BiolTransl Sci 101 289ndash310
96 Tesfazgi MebrhatuM WywialE GhoshA MichielsCWLindnerAB TaddeiF BujnickiJM Van MelderenL andAertsenA (2011) Evidence for an evolutionary antagonismbetween Mrr and Type III modification systems Nucleic AcidsRes 39 5991ndash6001
97 GrantGN JesseeJ BloomFR and HanahanD (1990)Differential plasmid rescue from transgenic mouse DNAs intoEscherichia coli methylation-restriction mutants Proc Natl AcadSci USA 87 4645ndash4649
68 Nucleic Acids Research 2014 Vol 42 No 1
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
98 AertsenA and MichielsCW (2005) Mrr instigates the SOSresponse after high pressure stress in Escherichia coli MolMicrobiol 58 1381ndash1391
99 WeiH TherrienC BlanchardA GuanS and ZhuZ (2008)The Fidelity Index provides a systematic quantitation of staractivity of DNA restriction endonucleases Nucleic Acids Res36 e50
100 LuL PatelH and BisslerJJ (2002) Optimizing DpnIdigestion to detect replicated DNA Biotechniques 33 316ndash318
101 MadejT AddessKJ FongJH GeerLY GeerRCLanczyckiCJ LiuC LuS Marchler-BauerAPanchenkoAR et al (2011) MMDB 3D structures andmacromolecular interactions Nucleic Acids Res 40 D461ndashD464
102 GibratJF MadejT and BryantSH (1996) Surprisingsimilarities in structure comparison Curr Opin Struct Biol 6377ndash385
103 PanneD MullerSA WirtzS EngelA and BickleTA (2001)The McrBC restriction endonuclease assembles into a ringstructure in the presence of G nucleotides EMBO J 203210ndash3217
104 NeuwaldAF AravindL SpougeJL and KooninEV (1999)AAA+ A class of chaperone-like ATPases associated with theassembly operation and disassembly of protein complexesGenome Res 9 27ndash43
105 OguraT and WilkinsonAJ (2001) AAA+ superfamilyATPases common structurendashdiverse function Genes Cells 6575ndash597
106 BearyTP BraymerHD and AchbergerEC (1997) Evidenceof participation of McrB(S) in McrBC restriction in Escherichiacoli K-12 J Bacteriol 179 7768ndash7775
107 PanneD RaleighEA and BickleTA (1998) McrBs amodulator peptide for McrBC activity EMBO J 175477ndash5483
108 PuntaM CoggillPC EberhardtRY MistryJ TateJBoursnellC PangN ForslundK CericG ClementsJ et al(2011) The Pfam protein families database Nucleic Acids Res40 D290ndashD301
109 NaitoT KusanoK and KobayashiI (1995) Selfish behavior ofrestriction-modification systems Science 267 897ndash899
110 KobayashiI (2001) Behavior of restriction-modification systemsas selfish mobile elements and their impact on genome evolutionNucleic Acids Res 29 3742ndash3756
111 FukudaE KaminskaKH BujnickiJM and KobayashiI(2008) Cell death upon epigenetic genome methylation a novelfunction of methyl-specific deoxyribonucleases Genome Biol 9R163
112 IshikawaK FukudaE and KobayashiI (2010) Conflictstargeting epigenetic systems and their resolution by cell deathnovel concepts for methyl-specific and other restriction systemsDNA Res 17 325ndash342
113 FukuyoM SasakiA and KobayashiI (2012) Success of asuicidal defense strategy against infection in a structured habitatSci Rep 2 238
114 MakovetsS TitheradgeAJ and MurrayNE (1998) ClpXand ClpP are essential for the efficient acquisition of genesspecifying type IA and IB restriction systems Mol Microbiol28 25ndash35
115 MurrayNE (2002) Immigration control of DNA in bacteriaself versus non-self Microbiology 148 3ndash20
116 MakovetsS PowellLM TitheradgeAJB BlakelyGW andMurrayNE (2004) Is modification sufficient to protect abacterial chromosome from a resident restriction endonucleaseMol Microbiol 51 135ndash147
117 OrgelLE and CrickFHC (1980) Selfish DNA the ultimateparasite Nature 284 604ndash607
118 KidwellMG and LischDR (2001) Perspective transposableelements parasitic DNA and genome evolution Evolution 551ndash24
119 SadykovM AsamiY NikiH HandaN ItayaMTanokuraM and KobayashiI (2003) Multiplication of arestriction-modification gene complex Mol Microbiol 48417ndash427
120 ChaoL and LevinBR (1981) Structured habitats and theevolution of anticompetitor toxins in bacteria Proc Natl AcadSci USA 78 6324ndash6328
121 MagnusonRD (2007) Hypothetical functions of toxin-antitoxinsystems J Bacteriol 189 6089ndash6092
122 RifatD WrightNT VarneyKM WeberDJ and BlackLW(2008) Restriction endonuclease inhibitor IPI of bacteriophageT4 a novel structure for a dedicated target J Mol Biol 375720ndash734
123 LabrieSJ SamsonJE and MoineauS (2010) Bacteriophageresistance mechanisms Nat Rev Microbiol 8 317ndash327
124 LinLF PosfaiJ RobertsRJ and KongH (2001)Comparative genomics of the restriction-modification systemsin Helicobacter pylori Proc Natl Acad Sci USA 982740ndash2745
125 MilkmanR RaleighEA McKaneM CrydermanDBilodeauP and McWeenyK (1999) Molecular evolution of theEscherichia coli chromosome V Recombination patterns amongstrains of diverse origin Genetics 153 539ndash554
126 HumbertO DorerMS and SalamaNR (2011)Characterization of Helicobacter pylori factors that controltransformation frequency and integration length during inter-strain DNA recombination Mol Microbiol 79 387ndash401
127 BarcusVA TitheradgeAJB and MurrayNE (1995) Thediversity of alleles at the hsd locus in natural populations ofEscherichia coli Genetics 140 1187ndash1197
128 SharpPM KelleherJE DanielAS CowanGM andMurrayNE (1992) Roles of selection and recombination in theevolution of type I restriction-modification systems inenterobacteria Proc Natl Acad Sci USA 89 9836ndash9840
129 LoenenWAM DanielAS BraymerHD and MurrayNE(1987) Organization and sequence of the hsd genes ofEscherichia coli K-12 J Mol Biol 198 159ndash170
130 SuriB and BickleTA (1985) EcoA the first member of a newfamily of type I restriction modification systems Geneorganization and enzymatic activities J Mol Biol 186 77ndash85
131 BertaniG and WeigleJJ (1953) Host-controlled variation inbacterial viruses J Bacteriol 65 113ndash121
132 SibleyMH and RaleighEA (2004) Cassette-like variation ofrestriction enzyme genes in Escherichia coli C and relativesNucleic Acids Res 32 522ndash534
133 MakarovaKS WolfYI SnirS and KooninEV (2011)Defense islands in bacterial and archaeal genomes and predictionof novel defense systems J Bacteriol 193 6039ndash6056
134 FurutaY AbeK and KobayashiI (2010) Genome comparisonand context analysis reveals putative mobile forms of restriction-modification systems and related rearrangements Nucleic AcidsRes 38 2428ndash2443
135 FurutaY and KobayashiI (2012) Restriction-modificationsystems as moblie epigenetic elements In RobertsAP andMullanyP (eds) Bacterial Integrative Mobile Genetic ElementsLandes Bioscience pp 85ndash103
136 OrsquoSullivanDJ and KlaenhammerTR (1998) Control ofexpression of LlaI restriction in Lactococcus lactis MolMicrobiol 27 1009ndash1020
137 WhitakerRD DornerLF and SchildkrautI (1999) A mutantof BamHI restriction endonuclease which requires N6-methyladenine for cleavage J Mol Biol 285 1525ndash1536
Nucleic Acids Research 2014 Vol 42 No 1 69
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
nucleases and modification-blocked REases (8788)Interestingly two of the PLDc nuclease activities havebeen shown to work by a transesterification reaction likethat used by topoisomerases and transposases (8789)PvuRts1I has an apparently unusual nuclease domain
[ie not yet identified by sources curated by theNCBI Conserved Domain Database (90)] However thisenzyme was included in a categorization of PD-(DE)XKfamilies (64) a tentatively identified divalent metalion binding site Block B (47) corresponds to Block Dof Bujnicki and Rychlewski (64) Cleavage requiresMg2+ ions
EcoCTGmrSD Functional organization is less clearbut several possible nuclease motifs were identified inGmrS (71) Cleavage buffer contained Ca2+ and Mg2+
ions and UTP
Sequence context recognition
Many of the modification-dependent enzymescharacterized so far have little sequence specificity incontrast to conventional modification-blocked REasesRelatively complete characterization of sequence prefer-ence and cleavage position has been carried out for TypeIV enzymes EcoKMcrBC SauUSI and PvuRts1I(Table 3) and for Type IIM DpnI and the MspJI family(Table 4) Progress has been made with binding recogni-tion for EcoKMcrA Cleavage conditions have beenachieved for Sco3AMcrA (Table 3) For all of these rec-ognition of surrounding sequence context is degeneratewith preference for a neighboring base and frequently arequirement for two sites with suitable separation DpnI isin some respects an exception see later in the text
The remaining nucleases in Table 4 are less wellcharacterized The recognition sites might form a relatedseries It will be interesting to learn more about the rela-tionships among these and how the requirement formultiple modified positions is specified eg for BlsI andPkrI
McrA binding domainsThe two lsquoMcrArsquo enzymes are not similar in theirN-termini with homology limited to the C-terminalnuclease domain For EcoKMcrA (Figure 2) there isgood genetic evidence that base recognition lies in theN-terminus Extensive mutagenesis using insertion of
Table 4 Characteristics of other modification-dependent enzymes (Type IIM)
Protein Subunitsdomains
DNABinding
EndonucleaseDomain
Recognitionsite
Comment and references
DpnI family G m6AjTC 13 characterized isoschizomersN-terminal PD(DE)XK (50777879)C-terminal wHTH R m6AjTC (78)
MspJI family mC with preferences Second copy stimulates cleavageMspJI 5 mCNNR(N913j) (51)
N-terminal SRA-like (80)C-terminal Mrr-cat (DE)(DEQ)XK (51)
FspEI Like MspJI C m5C(N1216j) (52)LpnPI Like MspJI C m5CDG(N1014j) (52)AspBHI Like MspJI YS m5CNS(N812j) (52)RlaI Like MspJI V m5CW (52)SgrTI Like MspJI C m5CDS(N1014j) (52)SgeI Like MspJI m5CNNR(N913j) 49 identical to MspJI (53)No familyassigned
Information from httpwwwsibenzymecomproductsm2_type
AoxI jRG m5CYBisI G m5CjNGCBlsI RYNjR Y At least two m5C requiredGlaI R m5CjGY (54)GluI GmCjNG m5CKroI Gj m5CCGGCMteI GmCG m5CjNGm5CGm5CPcsI m5CG(N5jN2)m5CGPkrI Gm5CNjG m5C At least 3 m5C required
Figure 2 McrA functional domains Domain function was inferred in-directly from genetic analysis by Anton amp Raleigh (2004) (57) Manymutations in the N-terminal domain spared some activity in one ormore of three functional tests (grey segments) while others were defi-cient in all activities (black segments) One mutation (asterix) was fullyactive on m5C-containing substrates but fully inactive in the hm5Cchallenge in vivo Most mutations in the C-terminus (pale greysegment) retained function in one test that was interpreted asmeasuring m5C binding ability A predicted structural model byBujnicki Radlinska and Rychlewski (2000) (58) for this C-terminalregion is compatible with these results
60 Nucleic Acids Research 2014 Vol 42 No 1
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
five-amino acid linkers and classification with three func-tional tests allowed assignment of DNA recognition to theN-terminal portion with the C-terminal H-N-H domainimplicated in cleavage Of particular note a mutationdiscriminated in vivo between hm5C and m5C was foundin the N-terminal domain (57) The mutant was able tofully restrict bacteriophage lambda modified by MHpaIIbut not at all phage T4 containing hm5C In vitromodification-dependent binding was achieved with thefull-length His-tagged protein (5556) yielding a putativerecognition site (YgtR)mCG This recognition site iscompatible with in vivo observations (391)
Presumably the N-terminus of Sco3A McrA also rec-ognizes the DNA Recognition of both m5C and thephosphorothioate (PT) moiety must be accommodatedin the final reaction As either modification is sufficientto elicit cleavage more than one domain could beinvolved Cleavage occurred near some but not allDcm-modified sites (Cm5CWGG) Both synthetic
PT-containing oligonucleotides and unmethylated PT-modified plasmid were also cleaved on both sides of asymmetrically modified site PT modification is thoughtto be sequence-specific (2629) but the details are notyet clear
Novel McrB binding domainDNA-binding resides in the McrB N-terminal domain(405960) whereas translocation and cleavage coordin-ation reside in the C-terminal AAA+ regulatory andtranslocation domain (61ndash637273) The complete trans-lation product of mcrB McrB-L binds DNA specifically(4072) via its N-terminal 161 amino acids (aa) (59) Thecrystal structure of this domain (McrB-N) in complex withDNA has recently been published (60)McrB-N uses a strategy first discovered for DNA-MTase
action (92) it flips the C base out of the DNA helix into abinding pocket for inspection The pocket is large enoughto accommodate C m5C hm5C or m4C but too small if a
Figure 3 McrBC Assembly Model Two proteins are expressed from mcrB in vivo Both the complete protein (McrB-L) and a small one missing theN-terminus (McrB-S top row) bind GTP forming high-order multimers detected by gel filtration (second row) When visualized by scanningtransmission electron microscopy these appear as heptameric rings with a central channel Rings of McrB-L in top views show projections thatmay correspond to the N-terminal DNA-binding domain (red segment) Both forms can then associate with McrC judged again by gel filtrationMcrB-LGTP can bind to its specific substrate (RmC) in the absence of McrC (third row) in its presence the substrate is cleaved (fourth row)GTP hydrolysis is required for cleavage (arrow) a supershifted binding complex forms in the presence of GTP-gamma-S but no cleavage occursTranslocation accompanies GTP hydrolysis double-stranded cleavage requires collaboration between two complexes or a translocationblock The path of the DNA in the figure is arbitrary as is the conformation of McrC Modified from BourniquelAA and BickleTAComplex restriction enzymes NTP-driven molecular motors Bourniquel and Bickle (84) with permission Copyright 2002 Elsevier MassonSAS All rights reserved
Nucleic Acids Research 2014 Vol 42 No 1 61
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
glucose moiety is attached Conserved residues Y64 andL68 were noted to make van der Waals contact with themethyl group of the flipped out m5C these contacts aremissing when the pocket contains CThe flipping action can be compared with but is distinct
from that of eukaryotic m5C-specific regulatory proteinsthat use the SET and Ring-finger-Associated (SRA)domain to read DNA modification state (Figure 4A)This domain is found in most eukaryotes in accessoryproteins (eg UHRF1NP95SUVH5) of the DNMT1maintenance MTase (93ndash95) Despite the similarstrategy the McrB-N domain is not homologous butdisplays a distinct protein fold (60) Binding is accom-plished from the minor groove and extraction of the Ccreates a 30 bend toward the major groove resembling aglycosylase in this respect (Figure 4A) The eukaryoticproteins form a crescent from which loops project towrap around the DNA with recognition mediatedthrough both major and minor grooves (94) For McrB-N the authors suggest that the purine preference in the 50
position might result from flexibility constraints or inter-action with a non-conserved aa that occupies space left bythe flipped base Substitutions of this aa (Y41A or Y41Q)compromised binding activity
Sequence specificity novel phenotype and structuralmodel of MrrIn 1987 Heitman and Model discovered Mrr when theyfound that transfer of various foreign m6A MTasesinduced an SOS response due to DNA damage (42)This response to the presence of an incompatible MTaseremains the principal evidence that the E coli K12 Mrrprotein cleaves DNA Related proteins discussed later inthe text (Type IIM) have been more tractable for in vitrowork No concise description of the Mrr recognitionsequence has been forthcoming although several studieshave examined the spectrum of incompatible MTases(43669697) Both adenine and cytosine MTases confersensitivityMrr is also responsible for DNA damage that does not
depend on methylation at all foreign or otherwise Highhydrostatic pressure (HP) induces the SOS DNA damageresponse and lethality (98) The response did not dependon the activity of the endogenous MTases of E coli K12but did depend on both the presence of wild-type mrr andthe integrity of the SOS signal generation pathwayPossibly HP elicits a non-enzymatic modification or astructural change in DNA helicity that is acted on byMrr This HP phenotype was used to characterize mrrmutants which were fitted into a computer-assistedmodel of the Mrr protein (67) An N-terminal DNA-binding winged helix was proposed with a C-terminalnuclease domain previously identified (69) The functionalimportance of several conserved residues was confirmedSeveral of the selected mutants with null phenotypes wereisolated in a region far from the active site or bindingsurface identified bioinformatically These could affectinteraction with a component of the HP response Thisintriguing collection of informative mutants will illumin-ate in vitro characterization
Type IIM binding domainsType IIM enzymes of two families are well-characterizedwith respect to cleavage (Table 4) Crystal structures forboth have recently appeared
DpnI winged-helix DNA recognitionUnusually for modification-dependent enzymes DpnIcleaves a four-base site (Gm6ATC) with high fidelity(7799) to leave blunt ends when both strands of the siteare methylated At low concentration the enzyme nicksthe modified strand of a hemimethylated site (100) Thebehavior of the enzyme with respect to modificationpatterns within the canonical GATC sitemdashmodificationof C or A one strand or bothmdashhas been thoroughlyexplored (50) However only recently has cleavage ofnon-canonical adenine-methylated sites been examinedSiwek and co-workers (78) found evidence for consider-able relaxation of specificity at the outer base This experi-ment used substrates modified by a highly non-specificadenine MTase extensive DpnI cleavage cloning of thefragments and sequencing of the borders
Structure determination in the presence of DNA andvalidation experiments (78) place this enzyme togetherwith the other modification-dependent enzymes in thattwo domains segregate the cleavage function fromsequence recognition It also separates DpnI from theothers in that the cleavage domain also possesses somemodification and sequence specificity The main recogni-tion is accomplished by a monomeric winged-helixdomain which binds in the major groove and recognizesthe modifications on both strands in the same event Thestructure does not reveal a cleavage-competent complexhowever because the cleavage domain is far from theDNA Filter-binding experiments validated the ability ofthe C-terminal domain to bind alone to do so moretightly to fully methylated than to hemimethylated oligo-nucleotides and to compete with the full-length enzymereducing cleavage by it Expression of the N-terminalcleavage domain alone (in low yield) allowed validationof its cleavage activity Surprisingly this cleavage wasitself dependent on modification state and sequence ofthe substrate Modeling based on the structure of theblunt-end-producing Type IIP enzyme PvuII allowed pre-diction that the cleavage domain approaches from theminor groove Complete understanding of double-stranded cleavage will depend on understanding thedynamic transformations that allow the cleavage domainto approach and act at the site
MspJI coupling of cleavage with DNA recognitionThe six members of the MspJI family use the Mrr-catversion of the PD-(DE)XK nuclease to cut at definedlocations to one side relative to the modified base(12 bases on the modified strand 16ndash17 on the otherFigure 5A) only one modified base is required fordouble-strand cleavage to occur (unlike McrBC) (5152)However these enzymes are stimulated by the presence ofa second site in cis or in trans Symmetrically modifiedsites (such as m5CpGm5CpG in mammalian DNA)yield prominent bands of defined size (32 bp) containinga mixed population of sequences each with a m5C in the
62 Nucleic Acids Research 2014 Vol 42 No 1
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
middle (Figure 5B) This behavior is recapitulated by thePvuRts1I group of enzymes (exemplified by AbaSDFI inFigure 5C) except that the distances are shorter andrecognition of modification state is less well understoodDuring the characterization of MspJI Dcm
(Cm5CWGG) sites were the first recognized substrateyielding a clear banding pattern (51) Cleavage of differ-ently modified plasmids and designed oligonucleotidesubstrates allowed a good assessment of both modificationand sequence specificity This family shows preference forparticular bases nearby similar to McrBC
MspJI DNA recognition is mediated by anSRA-like domainRecently the crystal structure of MspJI without DNA hasbeen resolved at 205 A (80) Search of the MolecularModeling Database at NCBI (101) using VAST (102)
Figure 4 McrB-N in comparison to other base-flipping proteins (A)SRA domains SUVH5 (3Q0C) and UHRF1 (2ZKD) use loops extend-ing from a crescent formed from two beta sheets to flip C or m5C fromundeformed B-form DNA into a pocket (top row) whereas McrB-N(3SSC bottom row) uses loops from one beta-sheet to distort the DNAand flip the base It resembles the human alkyladenine glycosylase(1BNK) (bottom row) in bending the DNA toward the majorgroove while flipping the base via the minor groove Figure 5 ofSukackaite et al (60) (B) The SRA-like hemi-methylated 5mC recog-nition domains A ribbon model of the N-terminal domain of theMspJI structure (4F0Q and 4F0P left) compared with the SRAdomain of URHF1 (PDB 3FDE right) The crescent shape formed
Figure 4 Continuedby interacting beta sheets and helices aB and aC are the conservedfeatures of the SRA domain highlighted here Loops on the concaveside of UHRF1 participate in flipping the base and similar loops pre-sumably do so for MspJI Two of these vary in length among familymembers and may play roles in sequence context specificity Figure 2aand b from Horton et al (80)
Figure 5 Schematic diagrams of cleavage positions for MspJI andAbaSDFI Cleavage of both strands is elicited by a singly modifiedsite for both MspJI (A) and AbaSDFI (C) Cleavage position is fixedrelative to the modified site but with a four-base 50 extension forMspJI and a two-base 30 extension for AbaSDFI When a site is sym-metrically-modified (as for CpG sites in mammalian DNA) a 32 base-pair fragment is excised from the DNA (B) (A) Figure 2a and (B)Figure 3a reprinted with permission from Cohen-Karni et al (52)(C) Figure 5a from Wang et al (47)
Nucleic Acids Research 2014 Vol 42 No 1 63
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
showed that the N-terminal domain was structurallysimilar to that of the eukaryotic SRA domain with acrescent-shaped beta-sheet structure from which loopsproject (see Figure 4B and discussion earlier in the textMcrB) This structural homology allowed modeling of theDNA-bound structure with a flipped m5C The enzyme inthe crystal is a tetramer in which two monomers form aback-to-back dimer via the C-terminal regions thatcomprise the endonuclease Two back-to-back dimersgenerate a tetrameric protein with two cleavage domainspositioned (as in the Type IIP enzyme HindIII used formodeling the C-terminal cleavage domain interaction withDNA) so that a 4-base 50 extension would be created oncleavage of modeled DNA Cleavage is most efficient atmolar ratios that allow all four SRA-like domains to beoccupiedmdashtoo much enzyme prevents cleavage fromoccurring
Tracking and dimerization
McrBC as translocaseBourniquel and Bickle (84) have reviewed much of theenzymology of McrBC which will be briefly summarizedhere The Raleigh Bickle and Pingoud laboratories havecontributed to the following consistent picture of thein vitro reaction EcoKMcrBC cleavage results in adouble-strand cut near one RmC site (72ndash74) butrequires cooperation of two sites (3940) or a translocationblock (73) The sites may be on different daughters acrossa fork (75) These are separated by 30ndash3000 bp (397274)and may be on either strand (3976) cleavage occurs30ndash35 bases from the modified base with oppositenicks not tightly constrained (73) and minor cleavageclusters are found 40 50 and 60 nt from the m5C(74) hm5C DNA elicits cleavage also (39) A ring struc-ture is formed by 5ndash7 molecules of McrB in the presence ofGTP (Figure 3) (103) this complex can bind to a recog-nition element in DNA In the presence of McrC trans-location of the complex occurs and cleavage ensues whentranslocation is blocked Collision of translocatingcomplexes a protein barrier or a topological barrier willelicit double-strand cleavage adjacent to one recognitionelement or the other The enzyme will cleave when recog-nition elements are on opposite sides of a forked structure(75) This would allow action in vivo to prevent entry of aMTase gene even with rare sitesStructurally the McrB protein is proposed to be a
member of the AAA+ protein family of NTPases (104)many of which form ring-shaped complexes and partici-pate in molecular machines lsquoSensorrsquo segments found inthese proteins have been shown in some cases to playroles in coupling NTPase activity to intersubunit commu-nication and movement (105) Two of three elements ofthe GTP-binding motif proposed by Dila et al (61) werevalidated by mutational analysis (6562) The thirdproposed motif element was identified as amino acidsNTAD by Dila et al Alignment of AAA+ NTPases in(104) found this aligned with the motif designatedSensor-1 in (105) An interesting result was that mutationshere unexpectedly appeared to abrogate interaction withMcrC instead of changing which NTP would be
productive (62) It may instead play a role in coordinatingGTP binding and hydrolysis with DNA binding inter-action with McrC and cleavage
Intracellularly the story becomes more complex as themcrB gene encodes two products of 51 and 33 kD McrB-L and McrB-S the latter one starting from an in-frameinternal translation start site (106) Both in vivo andin vitro McrB-S can interfere with the function ofMcrB-L at least in part by forming complexes withMcrC unable to bind DNA (107) Both species can formmultimeric rings in the presence of GTP (103) as is usualfor AAA+NTPases (104)
SauUSI requires two sites and ATP hydrolysisSauUSI was originally annotated as a putative helicasefrom Staphylococcus aureus sp A single polypeptide issufficient for activity both in vitro and also in vivo as aclone in E coli using modified phage as a challenge Theamino acid sequence contains a PLDc domain at theN-terminus This contains a phosphodiesterase motiforiginally identified in Phospholipase D (108) it wasvalidated by mutagenesis of four catalytic residue candi-dates In the middle ATPase and helicase motifs wereproposed to account for ATP dependence of cleavageactivity A Domain of Unknown Function was identifiedat the C-terminus (Pfam DUF3427) (108) and wasproposed to recognize the substrate (45)
The purified enzyme cleaves modified DNA containingm5C and hm5C but not m4C in the presence of ATP ordATP but not other nucleotides The negative result form4C is firm plasmids modified at the same site by an m5CMTase (Dcm) or an m4C MTase (MPspGI) were testedThe former (Cm5CWGG) was sensitive whereas the latterwas resistant Thus the sequence preference is likely tobe satisfied m6A is likely not a substrate but few m6A-containing sites were examined
Like McrBC SauUSI requires the presence of two sitesfor efficient cleavage Presumably the ATPase activityparticipates in monitoring the presence of two sites asfor other nucleotide-dependent REases includingMcrBC The mechanism of communication is unknownThe enzyme belongs to a family of highly similarorthologues found in other sequenced Staphylococci(Tables 1 and 2) and more distant homologues can befound in sequenced bacterial and archaeal genomes
EVOLUTIONARY PRESSURES ON R-M SYSTEMS
Evolution by selfish propagation
One way to understand the massive variety of restrictionsystems and their sporadic distribution is to locate theevolutionary drivers of enzyme diversification in theenzyme genes themselves as selfish elements Work fromthe Kobayashi laboratory has elaborated clear examplesof selfish behavior in some Type II enzyme systems(109110) in which the host becomes lsquoaddictedrsquo to theR-M system Once a cell has acquired an R-M pair lossof the genes results in death of that cellrsquos descendants asthe REase is frequently still present and able to act on thegenome following loss of methylation activity In this
64 Nucleic Acids Research 2014 Vol 42 No 1
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
perspective the role played by modification-dependentenzymes is host defense to exclude systems with lsquoforeignrsquoMTase patterns and prevent the cell from loading up withparasites The exclusion event is accompanied by the deathof the cell (111ndash113) Weak sequence specificity of Type IVenzymes could then result from the need to control entryof a wide variety of invading systems
The selfish aspect certainly plays a role in R-M popula-tion biology but cannot be the whole story Type II R-Msystems can still be lost by inactivation of the R gene firstMoreover Type I systems escape this scenario withcomplex control of cleavage activity the restrictionassembly includes a methylation assembly to begin withtherefore the R protein cannot act unless the MTase ispresent in addition failure of the methylation activity inan intact complex leads to abrogation of R activity some-times by action of the ClpXP protease specifically on theR protein (114ndash116)
Furthermore in population terms a cell that acquiredand became addicted to an R-M system should lose incompetition with a sibling that never received thesystem Two factors could counter this First acquisitioncould be accompanied by an increase in the total numberof copies of the R-M system in the population asproposed for invading transposable elements Thisoverreplication results in more copies of the systemcreated than are lost whether to suicide or to other select-ive disadvantage [see eg (117118)] R-M gene amplifica-tion within a cell has been reported experimentally (119)but spread in a population has not been demonstrated yetA second factor that could counter the disability of addic-tion is localization of competition In a structured envir-onment (colonies on a plate or biofilm on large or smallsurfaces) killing of segregants preserves limiting nutrientsfor lineages that retain the toxinantitoxin pair (120121)Much of the real world is structured so this is an import-ant condition
Evolution by phage-host arms race
A second perspective supposes that the modification-dependent Type IV enzymes arose from the competitivecoevolutionary interaction between phages and theirhosts This was first enunciated by Revel and Luria (2)and most recently by Black and coworkers (122) seealso (123) That is hosts used modification-blockedrestriction to defend against phage infection T-evenphages developed methods of substituting modified basesfor the ordinary ones hosts developed Type IV enzymes indefense phages added sugar or other modifications (19) tothwart Type IV enzymes hosts extended Type IV enzymesto accommodate these decorations finally phages de-veloped protein inhibitors specific for these enzymes aswell T4 phages deliver a protein inhibitor (IPI) alongwith the DNA on infection which allows growth in thepresence of EcoCTGmrSD The locus responsible for thisinhibitor is highly variable among relatives of T4 asgmrSD is in enteric bacteria (both in distribution and inaa sequence) When phage with different IP1 alleles weretested for protection from cloned EcoCTGmrSD and itshomolog EcoUTGmrSD specificity was evident one or
the other or both or neither of the two homologs wascounteracted in individual cases (122) This variability ofthe outcome supports the idea that phage-host interactiondrives at least some of these developmentsIn this perspective the weak sequence selectivity of the
Type IV systems might simply reflect the lack of endogen-ous targets for the enzyme As the host does not present anyhm5C and the phage is completely substituted selection forsequence-specificity is weak Selection would act to spareany co-residentMTases This differs fromType II enzymeswhere the MTase and REase must co-evolve to allow thehost to survive Each Type IV system is compatible withsome suite of Type I-IIIMTases (and thus the R-M systemsas a whole) Methylated or hydroxymethylated bases maynot be recognized at all (EcoCTGmrSD) or the systemmay require one specific base in addition to the modifiedone (McrA McrBC MspJI and PvuRts1I) MTases mod-ifying sites not including that base are then compatible asDcm (CmCWGG) is compatible with the McrA McrBCand Mrr systems in E coli K12Type IV systems that restrict methylated bases in a
weakly specified sequence context confer an additionaladvantage in competition with phages Many phagessuch as have not evolved the nucleotide-substitutionstrategy used by the T-even phages These phagesnormally carry the modification pattern of the mostrecent host if the last host expressed an MTase creatinga susceptible site the Type IV enzyme of the new host willdestroy the invader and limit the infection This may beaccompanied by the death of the individual infectedtherefore protection can be conferred on the sibling popu-lation (111)A further implication of this scenario considers the fate
of a population invaded by phage Phage survival ofrestriction occurs at biologically relevant frequencies(106-102) The survivors of restriction carry the particu-lar methylation pattern of the particular cell and thus areresistant to all restriction systems it might have carried(Types IndashIV) A bacterial population as a whole thenbenefits from mechanisms that diversify the suite ofR-M systems so that such surviving phage do not havefree access to the entire population The extreme variabil-ity of R-M system content in isolates of the same species iscompatible with this idea [see eg (124) REBASEGenomes httptoolsnebcomvinczegenomes] Suchvariability also limits and shapes interstrain genetransfer (115125126)The Raleigh laboratory has built on elegant genetic
work with Type I enzymes in the Murray laboratory(127ndash130) to investigate a locus designated thelsquoImmigration Control locusrsquo or ICR that exemplifiesvariable R-M content Alternative DNA segments con-taining R-M systems are located at the same definedlocation in most E coli chromosomes between the yjiSand yjiA genes The ICR in the non-restricting E coli Cstrain [used in the original definition of the R-M phenom-enon (131)] contains a remnant of a Type I enzyme R geneand is 13 kilobase shorter than the same region in E coliK12 (132) The mechanism of segment replacement is stillunknown The ICR would be an example of the lsquodefenseislandsrsquo analyzed by the Koonin group (133) lsquoDefense
Nucleic Acids Research 2014 Vol 42 No 1 65
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
islandsrsquo contain genes that can defend against phage orother invading DNA these exhibit bioinformaticproperties similar to lsquomobilome islandsrsquo containing mobil-ization genes (transposases for example) However themechanism of mobilization has not been identified forthe ICR
FINAL THOUGHTS
The extreme diversity of R-M systems that recognizeordinary DNA seems likely to be approached by the di-versity of Type IV restriction systems Type IV enzymesare hard to find as most detection methods depend ondevelopment of genetic systems for each taxon or on ser-endipity Those characterized so far mostly stem frominitial genetic investigation of limits on infection trans-formation or transduction Barriers encountered provideleads to the genes involved Bioinformatic analysis hashelped to identify relatives which may be more tractableto biochemical investigation than the example originallyfound This approach has pitfalls the gene encodingMspJI was first thought to code for an enzyme recognizingan unmodified site because it is immediately adjacent toan (inactive it is now thought) cytosine MTase geneProvidentially the first expression host was devoid of sen-sitive sites whereas the first test substrate contained some(51) A combination of biological experiments with bio-informatics and biochemistry will be needed to reveal thefull spectrum of Type IV enzymes that may lurk within thevast universe of unidentified ORFs in bacterial systemsOne might begin with those strains whose genomes carryfew Type II systems Bacillus or Corynebacterium asopposed to Helicobacter or Neisseria [see the Genomessection of REBASE (50)]The role of lsquodefense islandsrsquo and their relation to the
lsquomobilomersquo in bacterial population biology remains to bedetermined If a defense island is similar to a mobilomeisland there should be a mechanism of mobilizationnearby which would boost the contribution oflsquooverreplicationrsquo to the account of selections acting onR-M systems R-M systems of all types can be found onor adjacent to known mobilizing elements (134135) buthave not been shown to move experimentallyOn another note it may turn out that evolutionarily
there is a continuum between the apparently modifica-tion-dependent and modification-blocked paths Onerelative of McrBC predicted by bioinformatics analysesis LlaI a system that recognizes an unmodified sequenceand requires two MTases to support it (136) The enzymeBamHI prefers to cleave DNA with m6A in its GGATCCsite and mutants can be isolated that require this modifiedbase (137) Are there native systems similarly protected bymodification of one position in the recognition site butdependent on modification at a different one An interest-ing evolutionary series can be imagined
ACKNOWLEDGEMENTS
The authors thank Rich Roberts and David Dryden forencouragement and critical comment They are grateful to
Tom Bickle Virginius Siksnys Yu Zheng and XiadongCheng for permission to reproduce figures They alsothank Ivan Correa for help drawing base structures andTasha Jose for illustration assistance They also thank theanonymous reviewers of a previous manuscript for caretaken with the review process greatly improving thepresent document
FUNDING
New England Biolabs (EAR) Funding for open accesscharge New England Biolabs (to EAR)
Conflict of interest statement None declared
REFERENCES
1 LuriaSE and HumanML (1952) A nonhereditary host-inducedvariation of bacterial viruses J Bacteriol 64 557ndash559
2 RevelHR and LuriaSE (1970) DNA-glucosylation in T-evenphage genetic determination and role in phage-host interactionAnnu Rev Genet 4 177ndash192
3 RaleighEA and WilsonG (1986) Escherichia coli K-12 restrictsDNA containing 5-methylcytosine Proc Natl Acad Sci USA83 9070ndash9074
4 Noyer-WeidnerM DiazR and ReinersL (1986) Cytosine-specific DNA modification interferes with plasmid establishmentin Escherichia coli K12 involvement of rglB Mol Gen Genet205 469ndash475
5 BlumenthalRM GregorySA and CooperiderJS (1985)Cloning of a restriction-modification system from Proteus vulgarisand its use in analyzing a methylase-sensitive phenotype inEscherichia coli J Bacteriol 164 501ndash509
6 SohailA LiebM DarM and BhagwatAS (1990) A generequired for very short patch repair in Escherichia coli is adjacentto the DNA cytosine methylase gene J Bacteriol 1724214ndash4221
7 MacNeilDJ (1988) Characterization of a unique methyl-specificrestriction system in Streptomyces avermitilis J Bacteriol 1705607ndash5612
8 VertesAA InuiM KobayashiM KurusuY and YukawaH(1993) Presence of mrr- and mcr-like restriction systems incoryneform bacteria Res Microbiol 144 181ndash185
9 MacalusoA and MettusAM (1991) Efficient transformation ofBacillus thuringiensis requires nonmethylated plasmid DNAJ Bacteriol 173 1353ndash1356
10 HolmesML NuttallSD and Dyall-SmithML (1991)Construction and use of halobacterial shuttle vectors and furtherstudies on Haloferax DNA gyrase J Bacteriol 173 3807ndash3813
11 PosfaiJ BhagwatAS PosfaiG and RobertsRJ (1989)Predictive motifs derived from cytosine methyltransferases NucleicAcids Res 17 2421ndash2435
12 MaloneT BlumenthalRM and ChengX (1995) Structure-guided analysis reveals nine sequence motifs conserved amongDNA amino-methyltransferases and suggests a catalyticmechanism for these enzymes J Mol Biol 253 618ndash632
13 TiminskasA ButkusV and JanulaitisA (1995) Sequence motifcharacteristics for DNA [cytosine-N4] and DNA [adenine-N6]methyltransferases Classification of all DNA methyltransferasesGene 157 3ndash11
14 DickmanMJ InglestonSM SedelnikovaSE RaffertyJBLloydRG GrasbyJA and HornbyDP (2002) The RuvABCresolvasome Eur J Biochem 269 5492ndash5501
15 KovallRA and MatthewsBW (1998) Structural functionaland evolutionary relationships between lambda-exonuclease andthe type II restriction endonucleases Proc Natl Acad Sci USA95 7893ndash7897
16 RobertsRJ ChangYC HuZ RachlinJN AntonBPPokrzywaRM ChoiHP FallerLL GuleriaJ HousmanGet al (2010) COMBREX a project to accelerate the functional
66 Nucleic Acids Research 2014 Vol 42 No 1
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
annotation of prokaryotic genomes Nucleic Acids Res 39D11ndashD14
17 RaePM and SteeleRE (1978) Modified bases in the DNAs ofunicellular eukaryotes an examination of distributions andpossible roles with emphasis on hydroxymethyluracil indinoflagellates Biosystems 10 37ndash53
18 WarrenRAJ (1980) Modified bases in bacteriophage DNAsAnnu Rev Microbiol 34 137ndash158
19 Gommers-AmptJH and BorstP (1995) Hypermodified bases inDNA FASEB J 9 1034ndash1042
20 SwintonD HattmanS CrainPF ChengCS SmithDL andMcCloskeyJA (1983) Purification and characterization of theunusual deoxynucleoside alpha-N-(9-beta-D-20-deoxyribofuranosylpurin-6-yl)glycinamide specified by the phageMu modification function Proc Natl Acad Sci USA 807400ndash7404
21 TahilianiM KohKP ShenY PastorWA BandukwalaHBrudnoY AgarwalS IyerLM LiuDR AravindL et al(2009) Conversion of 5-methylcytosine to 5-hydroxymethylcytosinein mammalian DNA by MLL partner TET1 Science 324930ndash935
22 KriaucionisS and HeintzN (2009) The nuclear DNA base5-hydroxymethylcytosine is present in Purkinje neurons and thebrain Science 324 929ndash930
23 BorstP and SabatiniR (2008) Base J discovery biosynthesisand possible functions Annu RevMicrobiol 62 235ndash251
24 KaminskaKH and BujnickiJM (2008) Bacteriophage MuMom protein responsible for DNA modification is a new memberof the acyltransferase superfamily Cell Cycle 7 120ndash121
25 IyerLM TahilianiM RaoA and AravindL (2009) Predictionof novel families of enzymes involved in oxidative and othercomplex modifications of bases in nucleic acids Cell Cycle 81698ndash1710
26 ZhouX HeX LiangJ LiA XuT KieserT HelmannJDand DengZ (2005) A novel DNA modification by sulphur MolMicrobiol 57 1428ndash1438
27 OuHY HeX ShaoY TaiC RajakumarK and DengZ(2009) dndDB A Database Focused on Phosphorothioation ofthe DNA Backbone PloS One 4 e5132
28 XuT LiangJ ChenS WangL HeX YouD WangZLiA XuZ ZhouX et al (2009) DNA phosphorothioation inStreptomyces lividans mutational analysis of the dnd locusBMC Microbiol 9 41
29 WangL ChenS VerginKL GiovannoniSJ ChanSWDeMottMS TaghizadehK CorderoOX CutlerMTimberlakeS et al (2011) DNA phosphorothioation iswidespread and quantized in bacterial genomes Proc Natl AcadSci USA 108 2963ndash2968
30 Loslashbner-OlesenA SkovgaardO and MarinusMG (2005) Dammethylation coordinating cellular processes Curr OpinMicrobiol 8 154ndash160
31 LowDA and CasadesusJ (2008) Clocks and switches bacterialgene regulation by DNA adenine methylation Curr OpinMicrobiol 11 106ndash112
32 FreitagM and SelkerEU (2005) Controlling DNA methylationMany roads to one modification Curr Opin Genet Dev 15191ndash199
33 ProhaskaSJ StadlerPF and KrakauerDC (2010) Innovationin gene regulation The case of chromatin computation J TheorBiol 265 27ndash44
34 SeshasayeeASN SinghP and KrishnaS (2012) Context-dependent conservation of DNA methyltransferases in bacteriaNucleic Acids Res 40 7066ndash7073
35 AntonBP (2011) Dissertation Boston University BostonMassachusetts USA
36 RaleighEA TrimarchiR and RevelH (1989) Genetic andphysical mapping of the mcrA (rglA) and mcrB (rglB) loci ofEscherichia coli K-12 Genetics 122 279ndash296
37 Gonzalez-CeronG Miranda-OlivaresOJ and Servın-GonzalezL(2009) Characterization of the methyl-specific restriction system ofStreptomyces coelicolor A3(2) and of the role played by laterallyacquired nucleases FEMS Microbiol Lett 301 35ndash43
38 LiuG OuHY WangT LiL TanH ZhouXRajakumarK DengZ and HeX (2010) Cleavage of
phosphorothioated DNA and methylated DNA by the type IVrestriction endonuclease ScoMcrA PLoS Genet 6 e1001253
39 SutherlandE CoeL and RaleighEA (1992) McrBC amultisubunit GTP-dependent restriction endonuclease J MolBiol 225 327ndash348
40 KrugerT WildC and Noyer-WeidnerM (1995) McrB aprokaryotic protein specifically recognizing DNA containingmodified cytosine residues EMBO J 14 2661ndash2669
41 SitaramanR and LepplaSH (2012) Methylation-dependentDNA restriction in Bacillus anthracis Gene 494 44ndash50
42 HeitmanJ and ModelP (1987) Site-specific methylases inducethe SOS DNA repair response in Escherichia coli J Bacteriol169 3243ndash3250
43 Waite-ReesPA KeatingCJ MoranLS SlatkoBEHornstraLJ and BennerJS (1991) Characterisation andexpression of the Escherichia coli Mrr restriction systemJ Bacteriol 173 5207ndash5219
44 KerrAL JeonYJ SvensonCJ RogersPL and NeilanBA(2010) DNA restriction-modification systems in the ethanologenZymomonas mobilis ZM4 Appl MicrobiolBiotech 89 761ndash769
45 XuSY CorvagliaAR ChanSH ZhengY and LinderP(2011) A type IV modification-dependent restriction enzymeSauUSI from Staphylococcus aureus subsp aureus USA300Nucleic Acids Res 39 5597ndash5610
46 MonkIR ShahIM XuM TanMW and FosterTJ (2012)Transforming the untransformable application of directtransformation to manipulate genetically Staphylococcus aureusand Staphylococcus epidermidis mBio 3 e00277ndash00211
47 WangH GuanS QuimbyA Cohen-KarniD PradhanSWilsonG RobertsRJ ZhuZ and ZhengY (2011)Comparative characterization of the PvuRts1I family ofrestriction enzymes and their application in mapping genomic5-hydroxymethylcytosine Nucleic Acids Res 39 9294ndash9305
48 JanosiL YonemitsuH HongH and KajiA (1994) Molecularcloning and expression of a novel hydroxymethylcytosine-specificrestriction enzyme (PvuRts1I) modulated by glucosylation ofDNAJ Mol Biol 242 45ndash61
49 BairCL and BlackLW (2007) A type IV modificationdependent restriction nuclease that targets glucosylatedhydroxymethyl cytosine modified DNAs J Mol Biol 366768ndash778
50 RobertsRJ VinczeT PosfaiJ and MacelisD (2010)REBASEndasha database for DNA restriction and modificationenzymes genes and genomes Nucleic Acids Res 38 D234ndashD236
51 ZhengY Cohen-KarniD XuD ChinHG WilsonGPradhanS and RobertsRJ (2010) A unique family of Mrr-likemodification-dependent restriction endonucleases Nucleic AcidsRes 38 5527ndash5534
52 Cohen-KarniD XuD AponeL FomenkovA SunZDavisPJ KinneySRM Yamada-MabuchiM XuSYDavisT et al (2011) The MspJI family of modification-dependent restriction endonucleases for epigenetic studies ProcNatl Acad Sci USA 108 11040ndash11045
53 LubysA VitkuteJ LubieneJ and JanulaitisA (2011)Restriction enzymes and their applications US Patent ApplicationPublication 20110207139 A1 httpappftusptogovnetahtmlPTOsearch-boolhtml (27 August 2013 date last accessed)
54 TarasovaGV NayakshinaTN and DegtyarevSK (2008)Substrate specificity of new methyl-directed DNA endonucleaseGlaI BMC Mol Biol 9 7
55 MulliganEA and DunnJJ (2008) Cloning purification andinitial characterization of E coli McrA a putative5-methylcytosine-specific nuclease Protein ExprPurif 62 98ndash103
56 MulliganEA HatchwellE McCorkleSR and DunnJJ (2010)Differential binding of Escherichia coli McrA protein to DNAsequences that contain the dinucleotide m5CpG Nucleic AcidsRes 38 1997ndash2005
57 AntonBP and RaleighEA (2004) Transposon-mediated linkerinsertion scanning mutagenesis of the Escherichia coli McrAendonuclease J Bacteriol 186 5699ndash5707
58 BujnickiJM RadlinskaM and RychlewskiL (2000) Atomicmodel of the 5-methylcytosine-specific restriction enzyme McrAreveals an atypical zinc finger and structural similarity tobetabetaalphaMe endonucleases Mol Microbiol 37 1280ndash1281
Nucleic Acids Research 2014 Vol 42 No 1 67
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
59 GastFU BrinkmannT PieperU KrugerT NoyerWeidnerM and PingoudA (1997) The recognition of methylatedDNA by the GTP-dependent restriction endonuclease McrBCresides in the N-terminal domain of McrB Biol Chem 378975ndash982
60 SukackaiteR GrazulisS TamulaitisG and SiksnysV (2012)The recognition domain of the methyl-specific endonucleaseMcrBC flips out 5-methylcytosine Nucleic Acids Res 407552ndash7562
61 DilaD SutherlandE MoranL SlatkoB and RaleighEA(1990) Genetic and sequence organization of the mcrBC locus ofEscherichia coli K-12 J Bacteriol 172 4888ndash4900
62 PieperU BrinkmannT KrugerT Noyer WeidnerM andPingoudA (1997) Characterization of the interaction between therestriction endonuclease McrBC from E coli and its cofactorGTP J Mol Biol 272 190ndash199
63 PieperU SchweitzerT GrollDH GastFU and PingoudA(1999) The GTP-binding domain of McrB more than just avariation on a common theme J Mol Biol 292 547ndash556
64 BujnickiJM and RychlewskiL (2001) Grouping together highlydiverged PD-(DE)XK nucleases and identification of novelsuperfamily members using structure-guided alignment ofsequence profiles J Mol MicrobiolBiotech 3 69ndash72
65 PieperU and PingoudA (2002) A mutational analysis of thePD DEXK motif suggests that McrC harbors the catalyticcenter for DNA cleavage by the GTP-dependent restrictionenzyme McrBC from Escherichia coli Biochemistry 415236ndash5244
66 KelleherJE and RaleighEA (1991) A novel activity inEscherichia coli K-12 that directs restriction of DNA modified atCG dinucleotides J Bacteriol 173 5220ndash5223
67 OrlowskiJ MebrhatuMT MichielsCW BujnickiJM andAertsenA (2008) Mutational analysis and a structural model ofmethyl-directed restriction enzyme Mrr Biochem Biophys ResCommun 377 862ndash866
68 AravindL MakarovaKS and KooninEV (2000) SURVEYAND SUMMARY holliday junction resolvases and relatednucleases identification of new families phyletic distribution andevolutionary trajectories Nucleic Acids Res 28 3417ndash3432
69 BujnickiJM and RychlewskiL (2001) Identification of a PD-(DE)XK-like domain with a novel configuration of the endonucleaseactive site in the methyl-directed restriction enzyme Mrr and itshomologs Gene 267 183ndash191
70 SteczkiewiczK MuszewskaA KnizewskiL RychlewskiL andGinalskiK (2012) Sequence structure and functional diversity ofPD-(DE)XK phosphodiesterase superfamily Nucleic Acids Res40 7016ndash7045
71 BairCL RifatD and BlackLW (2007) Exclusion of glucosyl-hydroxymethylcytosine DNA containing bacteriophages isovercome by the injected protein inhibitor IPI J Mol Biol 366779ndash789
72 StewartFJ and RaleighEA (1998) Dependence of McrBCcleavage on distance between recognition elements Biol Chem379 611ndash616
73 PanneD RaleighEA and BickleTA (1999) The McrBCendonuclease translocates DNA in a reaction dependent on GTPhydrolysis J Mol Biol 290 49ndash60
74 PieperU GrollDH WunschS GastFU SpeckC MuckeNand PingoudA (2002) The GTP-dependent restriction enzymeMcrBC from Escherichia coli forms high-molecular masscomplexes with DNA and produces a cleavage patternwith a characteristic 10-base pair repeat Biochemistry 415245ndash5254
75 IshikawaK HandaN SearsL RaleighEA and KobayashiI(2011) Cleavage of a model DNA replication fork by amethyl-specific endonuclease Nucleic Acids Res 39 5489ndash5498
76 StewartFJ PanneD BickleTA and RaleighEA (2000)Methyl-specific DNA binding by McrBC a modification-dependent restriction enzyme J Mol Biol 298 611ndash622
77 LacksS and GreenbergB (1977) Complementary specificity ofrestriction endonucleases DpnI and DpnII with respect to DNAmethylation J Mol Biol 114 153ndash168
78 SiwekW CzapinskaH BochtlerM BujnickiJM andSkowronekK (2012) Crystal structure and mechanism
of action of the N6-methyladenine-dependent type IIMrestriction endonuclease RDpnI Nucleic Acids Res 407563ndash7572
79 LacksS and GreenbergB (1975) A deoxyribonuclease ofDiplococcus pneumoniae specific for methylated DNA J BiolChem 250 4060ndash4066
80 HortonJR MabuchiMY Cohen-KarniD ZhangXGriggsRM SamaranayakeM RobertsRJ ZhengY andChengX (2012) Structure and cleavage activity of the tetramericMspJI DNA modification-dependent restriction endonucleaseNucleic Acids Res 40 9763ndash9773
81 CymermanIA ObarskaA SkowronekKJ LubysA andBujnickiJM (2006) Identification of a new subfamily of HNHnucleases and experimental characterization of a representativemember HphI restriction endonuclease Proteins 65 867ndash876
82 GorbalenyaAE (1994) Self-splicing group I and group II intronsencode homologous (putative) DNA endonucleases of a newfamily Protein Sci 3 1117ndash1120
83 ChanSH OpitzL HigginsL OrsquoLoaneD and XuSY (2010)Cofactor requirement of HpyAV restriction endonuclease PloSOne 5 e9071
84 BourniquelAA and BickleTA (2002) Complex restrictionenzymes NTP-driven molecular motors Biochimie 84 1047ndash1059
85 OrsquoDriscollJ HeiterDF WilsonGG FitzgeraldGFRobertsR and van SinderenD (2006) A genetic dissection ofthe LlaJI restriction cassette reveals insights on a novelbacteriophage resistance system BMC Microbiol 6 40
86 SmithRM DiffinFM SaveryNJ JosephsenJ andSzczelkunMD (2009) DNA cleavage and methylation specificityof the single polypeptide restriction-modification enzyme LlaGINucleic Acids Res 37 7206ndash7218
87 InterthalH PouliotJJ and ChampouxJJ (2001) The tyrosyl-DNA phosphodiesterase Tdp1 is a member of the phospholipaseD superfamily Proc Natl Acad Sci USA 98 12009ndash12014
88 SasnauskasG HalfordSE and SiksnysV (2003) How the BfiIrestriction enzyme uses one active site to cut two DNA strandsProc Natl Acad Sci USA 100 6410ndash6415
89 SasnauskasG ConnollyBA HalfordSE and SiksnysV (2007)Site-specific DNA transesterification catalyzed by a restrictionenzyme Proc Natl Acad Sci USA 104 2115ndash2120
90 Marchler-BauerA ZhengC ChitsazF DerbyshireMKGeerLY GeerRC GonzalesNR GwadzM HurwitzDILanczyckiCJ et al (2012) CDD conserved domains andprotein three-dimensional structure Nucleic Acids Res 41D348ndashD352
91 KravetzAN ZakharovaMV BeletskayaIV SinevaEVDenjmuchametovMM PetrovSI GlatmanLI andSoloninAS (1993) The cleavage sites and localization of genesencoding the restriction endonucleases Eco1831I and EcoHIGene 129 153ndash154
92 KlimasauskasS KumarS RobertsRJ and ChengX (1994)HhaI methyltransferase flips its target base out of the DNA helixCell 76 357ndash369
93 HashimotoH VertinoPM and ChengX (2010) Molecularcoupling of DNA methylation and histone methylationEpigenomics 2 657ndash669
94 RajakumaraE LawJA SimanshuDK VoigtPJohnsonLM ReinbergD PatelDJ and JacobsenSE (2011)A dual flip-out mechanism for 5mC recognition by theArabidopsis SUVH5 SRA domain and its impact on DNAmethylation and H3K9 dimethylation in vivo Genes Dev 25137ndash152
95 SharifJ and KosekiH (2011) Recruitment of Dnmt1 roles of theSRA protein Np95 (Uhrfi) and other factors Prog Mol BiolTransl Sci 101 289ndash310
96 Tesfazgi MebrhatuM WywialE GhoshA MichielsCWLindnerAB TaddeiF BujnickiJM Van MelderenL andAertsenA (2011) Evidence for an evolutionary antagonismbetween Mrr and Type III modification systems Nucleic AcidsRes 39 5991ndash6001
97 GrantGN JesseeJ BloomFR and HanahanD (1990)Differential plasmid rescue from transgenic mouse DNAs intoEscherichia coli methylation-restriction mutants Proc Natl AcadSci USA 87 4645ndash4649
68 Nucleic Acids Research 2014 Vol 42 No 1
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
98 AertsenA and MichielsCW (2005) Mrr instigates the SOSresponse after high pressure stress in Escherichia coli MolMicrobiol 58 1381ndash1391
99 WeiH TherrienC BlanchardA GuanS and ZhuZ (2008)The Fidelity Index provides a systematic quantitation of staractivity of DNA restriction endonucleases Nucleic Acids Res36 e50
100 LuL PatelH and BisslerJJ (2002) Optimizing DpnIdigestion to detect replicated DNA Biotechniques 33 316ndash318
101 MadejT AddessKJ FongJH GeerLY GeerRCLanczyckiCJ LiuC LuS Marchler-BauerAPanchenkoAR et al (2011) MMDB 3D structures andmacromolecular interactions Nucleic Acids Res 40 D461ndashD464
102 GibratJF MadejT and BryantSH (1996) Surprisingsimilarities in structure comparison Curr Opin Struct Biol 6377ndash385
103 PanneD MullerSA WirtzS EngelA and BickleTA (2001)The McrBC restriction endonuclease assembles into a ringstructure in the presence of G nucleotides EMBO J 203210ndash3217
104 NeuwaldAF AravindL SpougeJL and KooninEV (1999)AAA+ A class of chaperone-like ATPases associated with theassembly operation and disassembly of protein complexesGenome Res 9 27ndash43
105 OguraT and WilkinsonAJ (2001) AAA+ superfamilyATPases common structurendashdiverse function Genes Cells 6575ndash597
106 BearyTP BraymerHD and AchbergerEC (1997) Evidenceof participation of McrB(S) in McrBC restriction in Escherichiacoli K-12 J Bacteriol 179 7768ndash7775
107 PanneD RaleighEA and BickleTA (1998) McrBs amodulator peptide for McrBC activity EMBO J 175477ndash5483
108 PuntaM CoggillPC EberhardtRY MistryJ TateJBoursnellC PangN ForslundK CericG ClementsJ et al(2011) The Pfam protein families database Nucleic Acids Res40 D290ndashD301
109 NaitoT KusanoK and KobayashiI (1995) Selfish behavior ofrestriction-modification systems Science 267 897ndash899
110 KobayashiI (2001) Behavior of restriction-modification systemsas selfish mobile elements and their impact on genome evolutionNucleic Acids Res 29 3742ndash3756
111 FukudaE KaminskaKH BujnickiJM and KobayashiI(2008) Cell death upon epigenetic genome methylation a novelfunction of methyl-specific deoxyribonucleases Genome Biol 9R163
112 IshikawaK FukudaE and KobayashiI (2010) Conflictstargeting epigenetic systems and their resolution by cell deathnovel concepts for methyl-specific and other restriction systemsDNA Res 17 325ndash342
113 FukuyoM SasakiA and KobayashiI (2012) Success of asuicidal defense strategy against infection in a structured habitatSci Rep 2 238
114 MakovetsS TitheradgeAJ and MurrayNE (1998) ClpXand ClpP are essential for the efficient acquisition of genesspecifying type IA and IB restriction systems Mol Microbiol28 25ndash35
115 MurrayNE (2002) Immigration control of DNA in bacteriaself versus non-self Microbiology 148 3ndash20
116 MakovetsS PowellLM TitheradgeAJB BlakelyGW andMurrayNE (2004) Is modification sufficient to protect abacterial chromosome from a resident restriction endonucleaseMol Microbiol 51 135ndash147
117 OrgelLE and CrickFHC (1980) Selfish DNA the ultimateparasite Nature 284 604ndash607
118 KidwellMG and LischDR (2001) Perspective transposableelements parasitic DNA and genome evolution Evolution 551ndash24
119 SadykovM AsamiY NikiH HandaN ItayaMTanokuraM and KobayashiI (2003) Multiplication of arestriction-modification gene complex Mol Microbiol 48417ndash427
120 ChaoL and LevinBR (1981) Structured habitats and theevolution of anticompetitor toxins in bacteria Proc Natl AcadSci USA 78 6324ndash6328
121 MagnusonRD (2007) Hypothetical functions of toxin-antitoxinsystems J Bacteriol 189 6089ndash6092
122 RifatD WrightNT VarneyKM WeberDJ and BlackLW(2008) Restriction endonuclease inhibitor IPI of bacteriophageT4 a novel structure for a dedicated target J Mol Biol 375720ndash734
123 LabrieSJ SamsonJE and MoineauS (2010) Bacteriophageresistance mechanisms Nat Rev Microbiol 8 317ndash327
124 LinLF PosfaiJ RobertsRJ and KongH (2001)Comparative genomics of the restriction-modification systemsin Helicobacter pylori Proc Natl Acad Sci USA 982740ndash2745
125 MilkmanR RaleighEA McKaneM CrydermanDBilodeauP and McWeenyK (1999) Molecular evolution of theEscherichia coli chromosome V Recombination patterns amongstrains of diverse origin Genetics 153 539ndash554
126 HumbertO DorerMS and SalamaNR (2011)Characterization of Helicobacter pylori factors that controltransformation frequency and integration length during inter-strain DNA recombination Mol Microbiol 79 387ndash401
127 BarcusVA TitheradgeAJB and MurrayNE (1995) Thediversity of alleles at the hsd locus in natural populations ofEscherichia coli Genetics 140 1187ndash1197
128 SharpPM KelleherJE DanielAS CowanGM andMurrayNE (1992) Roles of selection and recombination in theevolution of type I restriction-modification systems inenterobacteria Proc Natl Acad Sci USA 89 9836ndash9840
129 LoenenWAM DanielAS BraymerHD and MurrayNE(1987) Organization and sequence of the hsd genes ofEscherichia coli K-12 J Mol Biol 198 159ndash170
130 SuriB and BickleTA (1985) EcoA the first member of a newfamily of type I restriction modification systems Geneorganization and enzymatic activities J Mol Biol 186 77ndash85
131 BertaniG and WeigleJJ (1953) Host-controlled variation inbacterial viruses J Bacteriol 65 113ndash121
132 SibleyMH and RaleighEA (2004) Cassette-like variation ofrestriction enzyme genes in Escherichia coli C and relativesNucleic Acids Res 32 522ndash534
133 MakarovaKS WolfYI SnirS and KooninEV (2011)Defense islands in bacterial and archaeal genomes and predictionof novel defense systems J Bacteriol 193 6039ndash6056
134 FurutaY AbeK and KobayashiI (2010) Genome comparisonand context analysis reveals putative mobile forms of restriction-modification systems and related rearrangements Nucleic AcidsRes 38 2428ndash2443
135 FurutaY and KobayashiI (2012) Restriction-modificationsystems as moblie epigenetic elements In RobertsAP andMullanyP (eds) Bacterial Integrative Mobile Genetic ElementsLandes Bioscience pp 85ndash103
136 OrsquoSullivanDJ and KlaenhammerTR (1998) Control ofexpression of LlaI restriction in Lactococcus lactis MolMicrobiol 27 1009ndash1020
137 WhitakerRD DornerLF and SchildkrautI (1999) A mutantof BamHI restriction endonuclease which requires N6-methyladenine for cleavage J Mol Biol 285 1525ndash1536
Nucleic Acids Research 2014 Vol 42 No 1 69
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
five-amino acid linkers and classification with three func-tional tests allowed assignment of DNA recognition to theN-terminal portion with the C-terminal H-N-H domainimplicated in cleavage Of particular note a mutationdiscriminated in vivo between hm5C and m5C was foundin the N-terminal domain (57) The mutant was able tofully restrict bacteriophage lambda modified by MHpaIIbut not at all phage T4 containing hm5C In vitromodification-dependent binding was achieved with thefull-length His-tagged protein (5556) yielding a putativerecognition site (YgtR)mCG This recognition site iscompatible with in vivo observations (391)
Presumably the N-terminus of Sco3A McrA also rec-ognizes the DNA Recognition of both m5C and thephosphorothioate (PT) moiety must be accommodatedin the final reaction As either modification is sufficientto elicit cleavage more than one domain could beinvolved Cleavage occurred near some but not allDcm-modified sites (Cm5CWGG) Both synthetic
PT-containing oligonucleotides and unmethylated PT-modified plasmid were also cleaved on both sides of asymmetrically modified site PT modification is thoughtto be sequence-specific (2629) but the details are notyet clear
Novel McrB binding domainDNA-binding resides in the McrB N-terminal domain(405960) whereas translocation and cleavage coordin-ation reside in the C-terminal AAA+ regulatory andtranslocation domain (61ndash637273) The complete trans-lation product of mcrB McrB-L binds DNA specifically(4072) via its N-terminal 161 amino acids (aa) (59) Thecrystal structure of this domain (McrB-N) in complex withDNA has recently been published (60)McrB-N uses a strategy first discovered for DNA-MTase
action (92) it flips the C base out of the DNA helix into abinding pocket for inspection The pocket is large enoughto accommodate C m5C hm5C or m4C but too small if a
Figure 3 McrBC Assembly Model Two proteins are expressed from mcrB in vivo Both the complete protein (McrB-L) and a small one missing theN-terminus (McrB-S top row) bind GTP forming high-order multimers detected by gel filtration (second row) When visualized by scanningtransmission electron microscopy these appear as heptameric rings with a central channel Rings of McrB-L in top views show projections thatmay correspond to the N-terminal DNA-binding domain (red segment) Both forms can then associate with McrC judged again by gel filtrationMcrB-LGTP can bind to its specific substrate (RmC) in the absence of McrC (third row) in its presence the substrate is cleaved (fourth row)GTP hydrolysis is required for cleavage (arrow) a supershifted binding complex forms in the presence of GTP-gamma-S but no cleavage occursTranslocation accompanies GTP hydrolysis double-stranded cleavage requires collaboration between two complexes or a translocationblock The path of the DNA in the figure is arbitrary as is the conformation of McrC Modified from BourniquelAA and BickleTAComplex restriction enzymes NTP-driven molecular motors Bourniquel and Bickle (84) with permission Copyright 2002 Elsevier MassonSAS All rights reserved
Nucleic Acids Research 2014 Vol 42 No 1 61
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
glucose moiety is attached Conserved residues Y64 andL68 were noted to make van der Waals contact with themethyl group of the flipped out m5C these contacts aremissing when the pocket contains CThe flipping action can be compared with but is distinct
from that of eukaryotic m5C-specific regulatory proteinsthat use the SET and Ring-finger-Associated (SRA)domain to read DNA modification state (Figure 4A)This domain is found in most eukaryotes in accessoryproteins (eg UHRF1NP95SUVH5) of the DNMT1maintenance MTase (93ndash95) Despite the similarstrategy the McrB-N domain is not homologous butdisplays a distinct protein fold (60) Binding is accom-plished from the minor groove and extraction of the Ccreates a 30 bend toward the major groove resembling aglycosylase in this respect (Figure 4A) The eukaryoticproteins form a crescent from which loops project towrap around the DNA with recognition mediatedthrough both major and minor grooves (94) For McrB-N the authors suggest that the purine preference in the 50
position might result from flexibility constraints or inter-action with a non-conserved aa that occupies space left bythe flipped base Substitutions of this aa (Y41A or Y41Q)compromised binding activity
Sequence specificity novel phenotype and structuralmodel of MrrIn 1987 Heitman and Model discovered Mrr when theyfound that transfer of various foreign m6A MTasesinduced an SOS response due to DNA damage (42)This response to the presence of an incompatible MTaseremains the principal evidence that the E coli K12 Mrrprotein cleaves DNA Related proteins discussed later inthe text (Type IIM) have been more tractable for in vitrowork No concise description of the Mrr recognitionsequence has been forthcoming although several studieshave examined the spectrum of incompatible MTases(43669697) Both adenine and cytosine MTases confersensitivityMrr is also responsible for DNA damage that does not
depend on methylation at all foreign or otherwise Highhydrostatic pressure (HP) induces the SOS DNA damageresponse and lethality (98) The response did not dependon the activity of the endogenous MTases of E coli K12but did depend on both the presence of wild-type mrr andthe integrity of the SOS signal generation pathwayPossibly HP elicits a non-enzymatic modification or astructural change in DNA helicity that is acted on byMrr This HP phenotype was used to characterize mrrmutants which were fitted into a computer-assistedmodel of the Mrr protein (67) An N-terminal DNA-binding winged helix was proposed with a C-terminalnuclease domain previously identified (69) The functionalimportance of several conserved residues was confirmedSeveral of the selected mutants with null phenotypes wereisolated in a region far from the active site or bindingsurface identified bioinformatically These could affectinteraction with a component of the HP response Thisintriguing collection of informative mutants will illumin-ate in vitro characterization
Type IIM binding domainsType IIM enzymes of two families are well-characterizedwith respect to cleavage (Table 4) Crystal structures forboth have recently appeared
DpnI winged-helix DNA recognitionUnusually for modification-dependent enzymes DpnIcleaves a four-base site (Gm6ATC) with high fidelity(7799) to leave blunt ends when both strands of the siteare methylated At low concentration the enzyme nicksthe modified strand of a hemimethylated site (100) Thebehavior of the enzyme with respect to modificationpatterns within the canonical GATC sitemdashmodificationof C or A one strand or bothmdashhas been thoroughlyexplored (50) However only recently has cleavage ofnon-canonical adenine-methylated sites been examinedSiwek and co-workers (78) found evidence for consider-able relaxation of specificity at the outer base This experi-ment used substrates modified by a highly non-specificadenine MTase extensive DpnI cleavage cloning of thefragments and sequencing of the borders
Structure determination in the presence of DNA andvalidation experiments (78) place this enzyme togetherwith the other modification-dependent enzymes in thattwo domains segregate the cleavage function fromsequence recognition It also separates DpnI from theothers in that the cleavage domain also possesses somemodification and sequence specificity The main recogni-tion is accomplished by a monomeric winged-helixdomain which binds in the major groove and recognizesthe modifications on both strands in the same event Thestructure does not reveal a cleavage-competent complexhowever because the cleavage domain is far from theDNA Filter-binding experiments validated the ability ofthe C-terminal domain to bind alone to do so moretightly to fully methylated than to hemimethylated oligo-nucleotides and to compete with the full-length enzymereducing cleavage by it Expression of the N-terminalcleavage domain alone (in low yield) allowed validationof its cleavage activity Surprisingly this cleavage wasitself dependent on modification state and sequence ofthe substrate Modeling based on the structure of theblunt-end-producing Type IIP enzyme PvuII allowed pre-diction that the cleavage domain approaches from theminor groove Complete understanding of double-stranded cleavage will depend on understanding thedynamic transformations that allow the cleavage domainto approach and act at the site
MspJI coupling of cleavage with DNA recognitionThe six members of the MspJI family use the Mrr-catversion of the PD-(DE)XK nuclease to cut at definedlocations to one side relative to the modified base(12 bases on the modified strand 16ndash17 on the otherFigure 5A) only one modified base is required fordouble-strand cleavage to occur (unlike McrBC) (5152)However these enzymes are stimulated by the presence ofa second site in cis or in trans Symmetrically modifiedsites (such as m5CpGm5CpG in mammalian DNA)yield prominent bands of defined size (32 bp) containinga mixed population of sequences each with a m5C in the
62 Nucleic Acids Research 2014 Vol 42 No 1
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
middle (Figure 5B) This behavior is recapitulated by thePvuRts1I group of enzymes (exemplified by AbaSDFI inFigure 5C) except that the distances are shorter andrecognition of modification state is less well understoodDuring the characterization of MspJI Dcm
(Cm5CWGG) sites were the first recognized substrateyielding a clear banding pattern (51) Cleavage of differ-ently modified plasmids and designed oligonucleotidesubstrates allowed a good assessment of both modificationand sequence specificity This family shows preference forparticular bases nearby similar to McrBC
MspJI DNA recognition is mediated by anSRA-like domainRecently the crystal structure of MspJI without DNA hasbeen resolved at 205 A (80) Search of the MolecularModeling Database at NCBI (101) using VAST (102)
Figure 4 McrB-N in comparison to other base-flipping proteins (A)SRA domains SUVH5 (3Q0C) and UHRF1 (2ZKD) use loops extend-ing from a crescent formed from two beta sheets to flip C or m5C fromundeformed B-form DNA into a pocket (top row) whereas McrB-N(3SSC bottom row) uses loops from one beta-sheet to distort the DNAand flip the base It resembles the human alkyladenine glycosylase(1BNK) (bottom row) in bending the DNA toward the majorgroove while flipping the base via the minor groove Figure 5 ofSukackaite et al (60) (B) The SRA-like hemi-methylated 5mC recog-nition domains A ribbon model of the N-terminal domain of theMspJI structure (4F0Q and 4F0P left) compared with the SRAdomain of URHF1 (PDB 3FDE right) The crescent shape formed
Figure 4 Continuedby interacting beta sheets and helices aB and aC are the conservedfeatures of the SRA domain highlighted here Loops on the concaveside of UHRF1 participate in flipping the base and similar loops pre-sumably do so for MspJI Two of these vary in length among familymembers and may play roles in sequence context specificity Figure 2aand b from Horton et al (80)
Figure 5 Schematic diagrams of cleavage positions for MspJI andAbaSDFI Cleavage of both strands is elicited by a singly modifiedsite for both MspJI (A) and AbaSDFI (C) Cleavage position is fixedrelative to the modified site but with a four-base 50 extension forMspJI and a two-base 30 extension for AbaSDFI When a site is sym-metrically-modified (as for CpG sites in mammalian DNA) a 32 base-pair fragment is excised from the DNA (B) (A) Figure 2a and (B)Figure 3a reprinted with permission from Cohen-Karni et al (52)(C) Figure 5a from Wang et al (47)
Nucleic Acids Research 2014 Vol 42 No 1 63
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
showed that the N-terminal domain was structurallysimilar to that of the eukaryotic SRA domain with acrescent-shaped beta-sheet structure from which loopsproject (see Figure 4B and discussion earlier in the textMcrB) This structural homology allowed modeling of theDNA-bound structure with a flipped m5C The enzyme inthe crystal is a tetramer in which two monomers form aback-to-back dimer via the C-terminal regions thatcomprise the endonuclease Two back-to-back dimersgenerate a tetrameric protein with two cleavage domainspositioned (as in the Type IIP enzyme HindIII used formodeling the C-terminal cleavage domain interaction withDNA) so that a 4-base 50 extension would be created oncleavage of modeled DNA Cleavage is most efficient atmolar ratios that allow all four SRA-like domains to beoccupiedmdashtoo much enzyme prevents cleavage fromoccurring
Tracking and dimerization
McrBC as translocaseBourniquel and Bickle (84) have reviewed much of theenzymology of McrBC which will be briefly summarizedhere The Raleigh Bickle and Pingoud laboratories havecontributed to the following consistent picture of thein vitro reaction EcoKMcrBC cleavage results in adouble-strand cut near one RmC site (72ndash74) butrequires cooperation of two sites (3940) or a translocationblock (73) The sites may be on different daughters acrossa fork (75) These are separated by 30ndash3000 bp (397274)and may be on either strand (3976) cleavage occurs30ndash35 bases from the modified base with oppositenicks not tightly constrained (73) and minor cleavageclusters are found 40 50 and 60 nt from the m5C(74) hm5C DNA elicits cleavage also (39) A ring struc-ture is formed by 5ndash7 molecules of McrB in the presence ofGTP (Figure 3) (103) this complex can bind to a recog-nition element in DNA In the presence of McrC trans-location of the complex occurs and cleavage ensues whentranslocation is blocked Collision of translocatingcomplexes a protein barrier or a topological barrier willelicit double-strand cleavage adjacent to one recognitionelement or the other The enzyme will cleave when recog-nition elements are on opposite sides of a forked structure(75) This would allow action in vivo to prevent entry of aMTase gene even with rare sitesStructurally the McrB protein is proposed to be a
member of the AAA+ protein family of NTPases (104)many of which form ring-shaped complexes and partici-pate in molecular machines lsquoSensorrsquo segments found inthese proteins have been shown in some cases to playroles in coupling NTPase activity to intersubunit commu-nication and movement (105) Two of three elements ofthe GTP-binding motif proposed by Dila et al (61) werevalidated by mutational analysis (6562) The thirdproposed motif element was identified as amino acidsNTAD by Dila et al Alignment of AAA+ NTPases in(104) found this aligned with the motif designatedSensor-1 in (105) An interesting result was that mutationshere unexpectedly appeared to abrogate interaction withMcrC instead of changing which NTP would be
productive (62) It may instead play a role in coordinatingGTP binding and hydrolysis with DNA binding inter-action with McrC and cleavage
Intracellularly the story becomes more complex as themcrB gene encodes two products of 51 and 33 kD McrB-L and McrB-S the latter one starting from an in-frameinternal translation start site (106) Both in vivo andin vitro McrB-S can interfere with the function ofMcrB-L at least in part by forming complexes withMcrC unable to bind DNA (107) Both species can formmultimeric rings in the presence of GTP (103) as is usualfor AAA+NTPases (104)
SauUSI requires two sites and ATP hydrolysisSauUSI was originally annotated as a putative helicasefrom Staphylococcus aureus sp A single polypeptide issufficient for activity both in vitro and also in vivo as aclone in E coli using modified phage as a challenge Theamino acid sequence contains a PLDc domain at theN-terminus This contains a phosphodiesterase motiforiginally identified in Phospholipase D (108) it wasvalidated by mutagenesis of four catalytic residue candi-dates In the middle ATPase and helicase motifs wereproposed to account for ATP dependence of cleavageactivity A Domain of Unknown Function was identifiedat the C-terminus (Pfam DUF3427) (108) and wasproposed to recognize the substrate (45)
The purified enzyme cleaves modified DNA containingm5C and hm5C but not m4C in the presence of ATP ordATP but not other nucleotides The negative result form4C is firm plasmids modified at the same site by an m5CMTase (Dcm) or an m4C MTase (MPspGI) were testedThe former (Cm5CWGG) was sensitive whereas the latterwas resistant Thus the sequence preference is likely tobe satisfied m6A is likely not a substrate but few m6A-containing sites were examined
Like McrBC SauUSI requires the presence of two sitesfor efficient cleavage Presumably the ATPase activityparticipates in monitoring the presence of two sites asfor other nucleotide-dependent REases includingMcrBC The mechanism of communication is unknownThe enzyme belongs to a family of highly similarorthologues found in other sequenced Staphylococci(Tables 1 and 2) and more distant homologues can befound in sequenced bacterial and archaeal genomes
EVOLUTIONARY PRESSURES ON R-M SYSTEMS
Evolution by selfish propagation
One way to understand the massive variety of restrictionsystems and their sporadic distribution is to locate theevolutionary drivers of enzyme diversification in theenzyme genes themselves as selfish elements Work fromthe Kobayashi laboratory has elaborated clear examplesof selfish behavior in some Type II enzyme systems(109110) in which the host becomes lsquoaddictedrsquo to theR-M system Once a cell has acquired an R-M pair lossof the genes results in death of that cellrsquos descendants asthe REase is frequently still present and able to act on thegenome following loss of methylation activity In this
64 Nucleic Acids Research 2014 Vol 42 No 1
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
perspective the role played by modification-dependentenzymes is host defense to exclude systems with lsquoforeignrsquoMTase patterns and prevent the cell from loading up withparasites The exclusion event is accompanied by the deathof the cell (111ndash113) Weak sequence specificity of Type IVenzymes could then result from the need to control entryof a wide variety of invading systems
The selfish aspect certainly plays a role in R-M popula-tion biology but cannot be the whole story Type II R-Msystems can still be lost by inactivation of the R gene firstMoreover Type I systems escape this scenario withcomplex control of cleavage activity the restrictionassembly includes a methylation assembly to begin withtherefore the R protein cannot act unless the MTase ispresent in addition failure of the methylation activity inan intact complex leads to abrogation of R activity some-times by action of the ClpXP protease specifically on theR protein (114ndash116)
Furthermore in population terms a cell that acquiredand became addicted to an R-M system should lose incompetition with a sibling that never received thesystem Two factors could counter this First acquisitioncould be accompanied by an increase in the total numberof copies of the R-M system in the population asproposed for invading transposable elements Thisoverreplication results in more copies of the systemcreated than are lost whether to suicide or to other select-ive disadvantage [see eg (117118)] R-M gene amplifica-tion within a cell has been reported experimentally (119)but spread in a population has not been demonstrated yetA second factor that could counter the disability of addic-tion is localization of competition In a structured envir-onment (colonies on a plate or biofilm on large or smallsurfaces) killing of segregants preserves limiting nutrientsfor lineages that retain the toxinantitoxin pair (120121)Much of the real world is structured so this is an import-ant condition
Evolution by phage-host arms race
A second perspective supposes that the modification-dependent Type IV enzymes arose from the competitivecoevolutionary interaction between phages and theirhosts This was first enunciated by Revel and Luria (2)and most recently by Black and coworkers (122) seealso (123) That is hosts used modification-blockedrestriction to defend against phage infection T-evenphages developed methods of substituting modified basesfor the ordinary ones hosts developed Type IV enzymes indefense phages added sugar or other modifications (19) tothwart Type IV enzymes hosts extended Type IV enzymesto accommodate these decorations finally phages de-veloped protein inhibitors specific for these enzymes aswell T4 phages deliver a protein inhibitor (IPI) alongwith the DNA on infection which allows growth in thepresence of EcoCTGmrSD The locus responsible for thisinhibitor is highly variable among relatives of T4 asgmrSD is in enteric bacteria (both in distribution and inaa sequence) When phage with different IP1 alleles weretested for protection from cloned EcoCTGmrSD and itshomolog EcoUTGmrSD specificity was evident one or
the other or both or neither of the two homologs wascounteracted in individual cases (122) This variability ofthe outcome supports the idea that phage-host interactiondrives at least some of these developmentsIn this perspective the weak sequence selectivity of the
Type IV systems might simply reflect the lack of endogen-ous targets for the enzyme As the host does not present anyhm5C and the phage is completely substituted selection forsequence-specificity is weak Selection would act to spareany co-residentMTases This differs fromType II enzymeswhere the MTase and REase must co-evolve to allow thehost to survive Each Type IV system is compatible withsome suite of Type I-IIIMTases (and thus the R-M systemsas a whole) Methylated or hydroxymethylated bases maynot be recognized at all (EcoCTGmrSD) or the systemmay require one specific base in addition to the modifiedone (McrA McrBC MspJI and PvuRts1I) MTases mod-ifying sites not including that base are then compatible asDcm (CmCWGG) is compatible with the McrA McrBCand Mrr systems in E coli K12Type IV systems that restrict methylated bases in a
weakly specified sequence context confer an additionaladvantage in competition with phages Many phagessuch as have not evolved the nucleotide-substitutionstrategy used by the T-even phages These phagesnormally carry the modification pattern of the mostrecent host if the last host expressed an MTase creatinga susceptible site the Type IV enzyme of the new host willdestroy the invader and limit the infection This may beaccompanied by the death of the individual infectedtherefore protection can be conferred on the sibling popu-lation (111)A further implication of this scenario considers the fate
of a population invaded by phage Phage survival ofrestriction occurs at biologically relevant frequencies(106-102) The survivors of restriction carry the particu-lar methylation pattern of the particular cell and thus areresistant to all restriction systems it might have carried(Types IndashIV) A bacterial population as a whole thenbenefits from mechanisms that diversify the suite ofR-M systems so that such surviving phage do not havefree access to the entire population The extreme variabil-ity of R-M system content in isolates of the same species iscompatible with this idea [see eg (124) REBASEGenomes httptoolsnebcomvinczegenomes] Suchvariability also limits and shapes interstrain genetransfer (115125126)The Raleigh laboratory has built on elegant genetic
work with Type I enzymes in the Murray laboratory(127ndash130) to investigate a locus designated thelsquoImmigration Control locusrsquo or ICR that exemplifiesvariable R-M content Alternative DNA segments con-taining R-M systems are located at the same definedlocation in most E coli chromosomes between the yjiSand yjiA genes The ICR in the non-restricting E coli Cstrain [used in the original definition of the R-M phenom-enon (131)] contains a remnant of a Type I enzyme R geneand is 13 kilobase shorter than the same region in E coliK12 (132) The mechanism of segment replacement is stillunknown The ICR would be an example of the lsquodefenseislandsrsquo analyzed by the Koonin group (133) lsquoDefense
Nucleic Acids Research 2014 Vol 42 No 1 65
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
islandsrsquo contain genes that can defend against phage orother invading DNA these exhibit bioinformaticproperties similar to lsquomobilome islandsrsquo containing mobil-ization genes (transposases for example) However themechanism of mobilization has not been identified forthe ICR
FINAL THOUGHTS
The extreme diversity of R-M systems that recognizeordinary DNA seems likely to be approached by the di-versity of Type IV restriction systems Type IV enzymesare hard to find as most detection methods depend ondevelopment of genetic systems for each taxon or on ser-endipity Those characterized so far mostly stem frominitial genetic investigation of limits on infection trans-formation or transduction Barriers encountered provideleads to the genes involved Bioinformatic analysis hashelped to identify relatives which may be more tractableto biochemical investigation than the example originallyfound This approach has pitfalls the gene encodingMspJI was first thought to code for an enzyme recognizingan unmodified site because it is immediately adjacent toan (inactive it is now thought) cytosine MTase geneProvidentially the first expression host was devoid of sen-sitive sites whereas the first test substrate contained some(51) A combination of biological experiments with bio-informatics and biochemistry will be needed to reveal thefull spectrum of Type IV enzymes that may lurk within thevast universe of unidentified ORFs in bacterial systemsOne might begin with those strains whose genomes carryfew Type II systems Bacillus or Corynebacterium asopposed to Helicobacter or Neisseria [see the Genomessection of REBASE (50)]The role of lsquodefense islandsrsquo and their relation to the
lsquomobilomersquo in bacterial population biology remains to bedetermined If a defense island is similar to a mobilomeisland there should be a mechanism of mobilizationnearby which would boost the contribution oflsquooverreplicationrsquo to the account of selections acting onR-M systems R-M systems of all types can be found onor adjacent to known mobilizing elements (134135) buthave not been shown to move experimentallyOn another note it may turn out that evolutionarily
there is a continuum between the apparently modifica-tion-dependent and modification-blocked paths Onerelative of McrBC predicted by bioinformatics analysesis LlaI a system that recognizes an unmodified sequenceand requires two MTases to support it (136) The enzymeBamHI prefers to cleave DNA with m6A in its GGATCCsite and mutants can be isolated that require this modifiedbase (137) Are there native systems similarly protected bymodification of one position in the recognition site butdependent on modification at a different one An interest-ing evolutionary series can be imagined
ACKNOWLEDGEMENTS
The authors thank Rich Roberts and David Dryden forencouragement and critical comment They are grateful to
Tom Bickle Virginius Siksnys Yu Zheng and XiadongCheng for permission to reproduce figures They alsothank Ivan Correa for help drawing base structures andTasha Jose for illustration assistance They also thank theanonymous reviewers of a previous manuscript for caretaken with the review process greatly improving thepresent document
FUNDING
New England Biolabs (EAR) Funding for open accesscharge New England Biolabs (to EAR)
Conflict of interest statement None declared
REFERENCES
1 LuriaSE and HumanML (1952) A nonhereditary host-inducedvariation of bacterial viruses J Bacteriol 64 557ndash559
2 RevelHR and LuriaSE (1970) DNA-glucosylation in T-evenphage genetic determination and role in phage-host interactionAnnu Rev Genet 4 177ndash192
3 RaleighEA and WilsonG (1986) Escherichia coli K-12 restrictsDNA containing 5-methylcytosine Proc Natl Acad Sci USA83 9070ndash9074
4 Noyer-WeidnerM DiazR and ReinersL (1986) Cytosine-specific DNA modification interferes with plasmid establishmentin Escherichia coli K12 involvement of rglB Mol Gen Genet205 469ndash475
5 BlumenthalRM GregorySA and CooperiderJS (1985)Cloning of a restriction-modification system from Proteus vulgarisand its use in analyzing a methylase-sensitive phenotype inEscherichia coli J Bacteriol 164 501ndash509
6 SohailA LiebM DarM and BhagwatAS (1990) A generequired for very short patch repair in Escherichia coli is adjacentto the DNA cytosine methylase gene J Bacteriol 1724214ndash4221
7 MacNeilDJ (1988) Characterization of a unique methyl-specificrestriction system in Streptomyces avermitilis J Bacteriol 1705607ndash5612
8 VertesAA InuiM KobayashiM KurusuY and YukawaH(1993) Presence of mrr- and mcr-like restriction systems incoryneform bacteria Res Microbiol 144 181ndash185
9 MacalusoA and MettusAM (1991) Efficient transformation ofBacillus thuringiensis requires nonmethylated plasmid DNAJ Bacteriol 173 1353ndash1356
10 HolmesML NuttallSD and Dyall-SmithML (1991)Construction and use of halobacterial shuttle vectors and furtherstudies on Haloferax DNA gyrase J Bacteriol 173 3807ndash3813
11 PosfaiJ BhagwatAS PosfaiG and RobertsRJ (1989)Predictive motifs derived from cytosine methyltransferases NucleicAcids Res 17 2421ndash2435
12 MaloneT BlumenthalRM and ChengX (1995) Structure-guided analysis reveals nine sequence motifs conserved amongDNA amino-methyltransferases and suggests a catalyticmechanism for these enzymes J Mol Biol 253 618ndash632
13 TiminskasA ButkusV and JanulaitisA (1995) Sequence motifcharacteristics for DNA [cytosine-N4] and DNA [adenine-N6]methyltransferases Classification of all DNA methyltransferasesGene 157 3ndash11
14 DickmanMJ InglestonSM SedelnikovaSE RaffertyJBLloydRG GrasbyJA and HornbyDP (2002) The RuvABCresolvasome Eur J Biochem 269 5492ndash5501
15 KovallRA and MatthewsBW (1998) Structural functionaland evolutionary relationships between lambda-exonuclease andthe type II restriction endonucleases Proc Natl Acad Sci USA95 7893ndash7897
16 RobertsRJ ChangYC HuZ RachlinJN AntonBPPokrzywaRM ChoiHP FallerLL GuleriaJ HousmanGet al (2010) COMBREX a project to accelerate the functional
66 Nucleic Acids Research 2014 Vol 42 No 1
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
annotation of prokaryotic genomes Nucleic Acids Res 39D11ndashD14
17 RaePM and SteeleRE (1978) Modified bases in the DNAs ofunicellular eukaryotes an examination of distributions andpossible roles with emphasis on hydroxymethyluracil indinoflagellates Biosystems 10 37ndash53
18 WarrenRAJ (1980) Modified bases in bacteriophage DNAsAnnu Rev Microbiol 34 137ndash158
19 Gommers-AmptJH and BorstP (1995) Hypermodified bases inDNA FASEB J 9 1034ndash1042
20 SwintonD HattmanS CrainPF ChengCS SmithDL andMcCloskeyJA (1983) Purification and characterization of theunusual deoxynucleoside alpha-N-(9-beta-D-20-deoxyribofuranosylpurin-6-yl)glycinamide specified by the phageMu modification function Proc Natl Acad Sci USA 807400ndash7404
21 TahilianiM KohKP ShenY PastorWA BandukwalaHBrudnoY AgarwalS IyerLM LiuDR AravindL et al(2009) Conversion of 5-methylcytosine to 5-hydroxymethylcytosinein mammalian DNA by MLL partner TET1 Science 324930ndash935
22 KriaucionisS and HeintzN (2009) The nuclear DNA base5-hydroxymethylcytosine is present in Purkinje neurons and thebrain Science 324 929ndash930
23 BorstP and SabatiniR (2008) Base J discovery biosynthesisand possible functions Annu RevMicrobiol 62 235ndash251
24 KaminskaKH and BujnickiJM (2008) Bacteriophage MuMom protein responsible for DNA modification is a new memberof the acyltransferase superfamily Cell Cycle 7 120ndash121
25 IyerLM TahilianiM RaoA and AravindL (2009) Predictionof novel families of enzymes involved in oxidative and othercomplex modifications of bases in nucleic acids Cell Cycle 81698ndash1710
26 ZhouX HeX LiangJ LiA XuT KieserT HelmannJDand DengZ (2005) A novel DNA modification by sulphur MolMicrobiol 57 1428ndash1438
27 OuHY HeX ShaoY TaiC RajakumarK and DengZ(2009) dndDB A Database Focused on Phosphorothioation ofthe DNA Backbone PloS One 4 e5132
28 XuT LiangJ ChenS WangL HeX YouD WangZLiA XuZ ZhouX et al (2009) DNA phosphorothioation inStreptomyces lividans mutational analysis of the dnd locusBMC Microbiol 9 41
29 WangL ChenS VerginKL GiovannoniSJ ChanSWDeMottMS TaghizadehK CorderoOX CutlerMTimberlakeS et al (2011) DNA phosphorothioation iswidespread and quantized in bacterial genomes Proc Natl AcadSci USA 108 2963ndash2968
30 Loslashbner-OlesenA SkovgaardO and MarinusMG (2005) Dammethylation coordinating cellular processes Curr OpinMicrobiol 8 154ndash160
31 LowDA and CasadesusJ (2008) Clocks and switches bacterialgene regulation by DNA adenine methylation Curr OpinMicrobiol 11 106ndash112
32 FreitagM and SelkerEU (2005) Controlling DNA methylationMany roads to one modification Curr Opin Genet Dev 15191ndash199
33 ProhaskaSJ StadlerPF and KrakauerDC (2010) Innovationin gene regulation The case of chromatin computation J TheorBiol 265 27ndash44
34 SeshasayeeASN SinghP and KrishnaS (2012) Context-dependent conservation of DNA methyltransferases in bacteriaNucleic Acids Res 40 7066ndash7073
35 AntonBP (2011) Dissertation Boston University BostonMassachusetts USA
36 RaleighEA TrimarchiR and RevelH (1989) Genetic andphysical mapping of the mcrA (rglA) and mcrB (rglB) loci ofEscherichia coli K-12 Genetics 122 279ndash296
37 Gonzalez-CeronG Miranda-OlivaresOJ and Servın-GonzalezL(2009) Characterization of the methyl-specific restriction system ofStreptomyces coelicolor A3(2) and of the role played by laterallyacquired nucleases FEMS Microbiol Lett 301 35ndash43
38 LiuG OuHY WangT LiL TanH ZhouXRajakumarK DengZ and HeX (2010) Cleavage of
phosphorothioated DNA and methylated DNA by the type IVrestriction endonuclease ScoMcrA PLoS Genet 6 e1001253
39 SutherlandE CoeL and RaleighEA (1992) McrBC amultisubunit GTP-dependent restriction endonuclease J MolBiol 225 327ndash348
40 KrugerT WildC and Noyer-WeidnerM (1995) McrB aprokaryotic protein specifically recognizing DNA containingmodified cytosine residues EMBO J 14 2661ndash2669
41 SitaramanR and LepplaSH (2012) Methylation-dependentDNA restriction in Bacillus anthracis Gene 494 44ndash50
42 HeitmanJ and ModelP (1987) Site-specific methylases inducethe SOS DNA repair response in Escherichia coli J Bacteriol169 3243ndash3250
43 Waite-ReesPA KeatingCJ MoranLS SlatkoBEHornstraLJ and BennerJS (1991) Characterisation andexpression of the Escherichia coli Mrr restriction systemJ Bacteriol 173 5207ndash5219
44 KerrAL JeonYJ SvensonCJ RogersPL and NeilanBA(2010) DNA restriction-modification systems in the ethanologenZymomonas mobilis ZM4 Appl MicrobiolBiotech 89 761ndash769
45 XuSY CorvagliaAR ChanSH ZhengY and LinderP(2011) A type IV modification-dependent restriction enzymeSauUSI from Staphylococcus aureus subsp aureus USA300Nucleic Acids Res 39 5597ndash5610
46 MonkIR ShahIM XuM TanMW and FosterTJ (2012)Transforming the untransformable application of directtransformation to manipulate genetically Staphylococcus aureusand Staphylococcus epidermidis mBio 3 e00277ndash00211
47 WangH GuanS QuimbyA Cohen-KarniD PradhanSWilsonG RobertsRJ ZhuZ and ZhengY (2011)Comparative characterization of the PvuRts1I family ofrestriction enzymes and their application in mapping genomic5-hydroxymethylcytosine Nucleic Acids Res 39 9294ndash9305
48 JanosiL YonemitsuH HongH and KajiA (1994) Molecularcloning and expression of a novel hydroxymethylcytosine-specificrestriction enzyme (PvuRts1I) modulated by glucosylation ofDNAJ Mol Biol 242 45ndash61
49 BairCL and BlackLW (2007) A type IV modificationdependent restriction nuclease that targets glucosylatedhydroxymethyl cytosine modified DNAs J Mol Biol 366768ndash778
50 RobertsRJ VinczeT PosfaiJ and MacelisD (2010)REBASEndasha database for DNA restriction and modificationenzymes genes and genomes Nucleic Acids Res 38 D234ndashD236
51 ZhengY Cohen-KarniD XuD ChinHG WilsonGPradhanS and RobertsRJ (2010) A unique family of Mrr-likemodification-dependent restriction endonucleases Nucleic AcidsRes 38 5527ndash5534
52 Cohen-KarniD XuD AponeL FomenkovA SunZDavisPJ KinneySRM Yamada-MabuchiM XuSYDavisT et al (2011) The MspJI family of modification-dependent restriction endonucleases for epigenetic studies ProcNatl Acad Sci USA 108 11040ndash11045
53 LubysA VitkuteJ LubieneJ and JanulaitisA (2011)Restriction enzymes and their applications US Patent ApplicationPublication 20110207139 A1 httpappftusptogovnetahtmlPTOsearch-boolhtml (27 August 2013 date last accessed)
54 TarasovaGV NayakshinaTN and DegtyarevSK (2008)Substrate specificity of new methyl-directed DNA endonucleaseGlaI BMC Mol Biol 9 7
55 MulliganEA and DunnJJ (2008) Cloning purification andinitial characterization of E coli McrA a putative5-methylcytosine-specific nuclease Protein ExprPurif 62 98ndash103
56 MulliganEA HatchwellE McCorkleSR and DunnJJ (2010)Differential binding of Escherichia coli McrA protein to DNAsequences that contain the dinucleotide m5CpG Nucleic AcidsRes 38 1997ndash2005
57 AntonBP and RaleighEA (2004) Transposon-mediated linkerinsertion scanning mutagenesis of the Escherichia coli McrAendonuclease J Bacteriol 186 5699ndash5707
58 BujnickiJM RadlinskaM and RychlewskiL (2000) Atomicmodel of the 5-methylcytosine-specific restriction enzyme McrAreveals an atypical zinc finger and structural similarity tobetabetaalphaMe endonucleases Mol Microbiol 37 1280ndash1281
Nucleic Acids Research 2014 Vol 42 No 1 67
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
59 GastFU BrinkmannT PieperU KrugerT NoyerWeidnerM and PingoudA (1997) The recognition of methylatedDNA by the GTP-dependent restriction endonuclease McrBCresides in the N-terminal domain of McrB Biol Chem 378975ndash982
60 SukackaiteR GrazulisS TamulaitisG and SiksnysV (2012)The recognition domain of the methyl-specific endonucleaseMcrBC flips out 5-methylcytosine Nucleic Acids Res 407552ndash7562
61 DilaD SutherlandE MoranL SlatkoB and RaleighEA(1990) Genetic and sequence organization of the mcrBC locus ofEscherichia coli K-12 J Bacteriol 172 4888ndash4900
62 PieperU BrinkmannT KrugerT Noyer WeidnerM andPingoudA (1997) Characterization of the interaction between therestriction endonuclease McrBC from E coli and its cofactorGTP J Mol Biol 272 190ndash199
63 PieperU SchweitzerT GrollDH GastFU and PingoudA(1999) The GTP-binding domain of McrB more than just avariation on a common theme J Mol Biol 292 547ndash556
64 BujnickiJM and RychlewskiL (2001) Grouping together highlydiverged PD-(DE)XK nucleases and identification of novelsuperfamily members using structure-guided alignment ofsequence profiles J Mol MicrobiolBiotech 3 69ndash72
65 PieperU and PingoudA (2002) A mutational analysis of thePD DEXK motif suggests that McrC harbors the catalyticcenter for DNA cleavage by the GTP-dependent restrictionenzyme McrBC from Escherichia coli Biochemistry 415236ndash5244
66 KelleherJE and RaleighEA (1991) A novel activity inEscherichia coli K-12 that directs restriction of DNA modified atCG dinucleotides J Bacteriol 173 5220ndash5223
67 OrlowskiJ MebrhatuMT MichielsCW BujnickiJM andAertsenA (2008) Mutational analysis and a structural model ofmethyl-directed restriction enzyme Mrr Biochem Biophys ResCommun 377 862ndash866
68 AravindL MakarovaKS and KooninEV (2000) SURVEYAND SUMMARY holliday junction resolvases and relatednucleases identification of new families phyletic distribution andevolutionary trajectories Nucleic Acids Res 28 3417ndash3432
69 BujnickiJM and RychlewskiL (2001) Identification of a PD-(DE)XK-like domain with a novel configuration of the endonucleaseactive site in the methyl-directed restriction enzyme Mrr and itshomologs Gene 267 183ndash191
70 SteczkiewiczK MuszewskaA KnizewskiL RychlewskiL andGinalskiK (2012) Sequence structure and functional diversity ofPD-(DE)XK phosphodiesterase superfamily Nucleic Acids Res40 7016ndash7045
71 BairCL RifatD and BlackLW (2007) Exclusion of glucosyl-hydroxymethylcytosine DNA containing bacteriophages isovercome by the injected protein inhibitor IPI J Mol Biol 366779ndash789
72 StewartFJ and RaleighEA (1998) Dependence of McrBCcleavage on distance between recognition elements Biol Chem379 611ndash616
73 PanneD RaleighEA and BickleTA (1999) The McrBCendonuclease translocates DNA in a reaction dependent on GTPhydrolysis J Mol Biol 290 49ndash60
74 PieperU GrollDH WunschS GastFU SpeckC MuckeNand PingoudA (2002) The GTP-dependent restriction enzymeMcrBC from Escherichia coli forms high-molecular masscomplexes with DNA and produces a cleavage patternwith a characteristic 10-base pair repeat Biochemistry 415245ndash5254
75 IshikawaK HandaN SearsL RaleighEA and KobayashiI(2011) Cleavage of a model DNA replication fork by amethyl-specific endonuclease Nucleic Acids Res 39 5489ndash5498
76 StewartFJ PanneD BickleTA and RaleighEA (2000)Methyl-specific DNA binding by McrBC a modification-dependent restriction enzyme J Mol Biol 298 611ndash622
77 LacksS and GreenbergB (1977) Complementary specificity ofrestriction endonucleases DpnI and DpnII with respect to DNAmethylation J Mol Biol 114 153ndash168
78 SiwekW CzapinskaH BochtlerM BujnickiJM andSkowronekK (2012) Crystal structure and mechanism
of action of the N6-methyladenine-dependent type IIMrestriction endonuclease RDpnI Nucleic Acids Res 407563ndash7572
79 LacksS and GreenbergB (1975) A deoxyribonuclease ofDiplococcus pneumoniae specific for methylated DNA J BiolChem 250 4060ndash4066
80 HortonJR MabuchiMY Cohen-KarniD ZhangXGriggsRM SamaranayakeM RobertsRJ ZhengY andChengX (2012) Structure and cleavage activity of the tetramericMspJI DNA modification-dependent restriction endonucleaseNucleic Acids Res 40 9763ndash9773
81 CymermanIA ObarskaA SkowronekKJ LubysA andBujnickiJM (2006) Identification of a new subfamily of HNHnucleases and experimental characterization of a representativemember HphI restriction endonuclease Proteins 65 867ndash876
82 GorbalenyaAE (1994) Self-splicing group I and group II intronsencode homologous (putative) DNA endonucleases of a newfamily Protein Sci 3 1117ndash1120
83 ChanSH OpitzL HigginsL OrsquoLoaneD and XuSY (2010)Cofactor requirement of HpyAV restriction endonuclease PloSOne 5 e9071
84 BourniquelAA and BickleTA (2002) Complex restrictionenzymes NTP-driven molecular motors Biochimie 84 1047ndash1059
85 OrsquoDriscollJ HeiterDF WilsonGG FitzgeraldGFRobertsR and van SinderenD (2006) A genetic dissection ofthe LlaJI restriction cassette reveals insights on a novelbacteriophage resistance system BMC Microbiol 6 40
86 SmithRM DiffinFM SaveryNJ JosephsenJ andSzczelkunMD (2009) DNA cleavage and methylation specificityof the single polypeptide restriction-modification enzyme LlaGINucleic Acids Res 37 7206ndash7218
87 InterthalH PouliotJJ and ChampouxJJ (2001) The tyrosyl-DNA phosphodiesterase Tdp1 is a member of the phospholipaseD superfamily Proc Natl Acad Sci USA 98 12009ndash12014
88 SasnauskasG HalfordSE and SiksnysV (2003) How the BfiIrestriction enzyme uses one active site to cut two DNA strandsProc Natl Acad Sci USA 100 6410ndash6415
89 SasnauskasG ConnollyBA HalfordSE and SiksnysV (2007)Site-specific DNA transesterification catalyzed by a restrictionenzyme Proc Natl Acad Sci USA 104 2115ndash2120
90 Marchler-BauerA ZhengC ChitsazF DerbyshireMKGeerLY GeerRC GonzalesNR GwadzM HurwitzDILanczyckiCJ et al (2012) CDD conserved domains andprotein three-dimensional structure Nucleic Acids Res 41D348ndashD352
91 KravetzAN ZakharovaMV BeletskayaIV SinevaEVDenjmuchametovMM PetrovSI GlatmanLI andSoloninAS (1993) The cleavage sites and localization of genesencoding the restriction endonucleases Eco1831I and EcoHIGene 129 153ndash154
92 KlimasauskasS KumarS RobertsRJ and ChengX (1994)HhaI methyltransferase flips its target base out of the DNA helixCell 76 357ndash369
93 HashimotoH VertinoPM and ChengX (2010) Molecularcoupling of DNA methylation and histone methylationEpigenomics 2 657ndash669
94 RajakumaraE LawJA SimanshuDK VoigtPJohnsonLM ReinbergD PatelDJ and JacobsenSE (2011)A dual flip-out mechanism for 5mC recognition by theArabidopsis SUVH5 SRA domain and its impact on DNAmethylation and H3K9 dimethylation in vivo Genes Dev 25137ndash152
95 SharifJ and KosekiH (2011) Recruitment of Dnmt1 roles of theSRA protein Np95 (Uhrfi) and other factors Prog Mol BiolTransl Sci 101 289ndash310
96 Tesfazgi MebrhatuM WywialE GhoshA MichielsCWLindnerAB TaddeiF BujnickiJM Van MelderenL andAertsenA (2011) Evidence for an evolutionary antagonismbetween Mrr and Type III modification systems Nucleic AcidsRes 39 5991ndash6001
97 GrantGN JesseeJ BloomFR and HanahanD (1990)Differential plasmid rescue from transgenic mouse DNAs intoEscherichia coli methylation-restriction mutants Proc Natl AcadSci USA 87 4645ndash4649
68 Nucleic Acids Research 2014 Vol 42 No 1
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
98 AertsenA and MichielsCW (2005) Mrr instigates the SOSresponse after high pressure stress in Escherichia coli MolMicrobiol 58 1381ndash1391
99 WeiH TherrienC BlanchardA GuanS and ZhuZ (2008)The Fidelity Index provides a systematic quantitation of staractivity of DNA restriction endonucleases Nucleic Acids Res36 e50
100 LuL PatelH and BisslerJJ (2002) Optimizing DpnIdigestion to detect replicated DNA Biotechniques 33 316ndash318
101 MadejT AddessKJ FongJH GeerLY GeerRCLanczyckiCJ LiuC LuS Marchler-BauerAPanchenkoAR et al (2011) MMDB 3D structures andmacromolecular interactions Nucleic Acids Res 40 D461ndashD464
102 GibratJF MadejT and BryantSH (1996) Surprisingsimilarities in structure comparison Curr Opin Struct Biol 6377ndash385
103 PanneD MullerSA WirtzS EngelA and BickleTA (2001)The McrBC restriction endonuclease assembles into a ringstructure in the presence of G nucleotides EMBO J 203210ndash3217
104 NeuwaldAF AravindL SpougeJL and KooninEV (1999)AAA+ A class of chaperone-like ATPases associated with theassembly operation and disassembly of protein complexesGenome Res 9 27ndash43
105 OguraT and WilkinsonAJ (2001) AAA+ superfamilyATPases common structurendashdiverse function Genes Cells 6575ndash597
106 BearyTP BraymerHD and AchbergerEC (1997) Evidenceof participation of McrB(S) in McrBC restriction in Escherichiacoli K-12 J Bacteriol 179 7768ndash7775
107 PanneD RaleighEA and BickleTA (1998) McrBs amodulator peptide for McrBC activity EMBO J 175477ndash5483
108 PuntaM CoggillPC EberhardtRY MistryJ TateJBoursnellC PangN ForslundK CericG ClementsJ et al(2011) The Pfam protein families database Nucleic Acids Res40 D290ndashD301
109 NaitoT KusanoK and KobayashiI (1995) Selfish behavior ofrestriction-modification systems Science 267 897ndash899
110 KobayashiI (2001) Behavior of restriction-modification systemsas selfish mobile elements and their impact on genome evolutionNucleic Acids Res 29 3742ndash3756
111 FukudaE KaminskaKH BujnickiJM and KobayashiI(2008) Cell death upon epigenetic genome methylation a novelfunction of methyl-specific deoxyribonucleases Genome Biol 9R163
112 IshikawaK FukudaE and KobayashiI (2010) Conflictstargeting epigenetic systems and their resolution by cell deathnovel concepts for methyl-specific and other restriction systemsDNA Res 17 325ndash342
113 FukuyoM SasakiA and KobayashiI (2012) Success of asuicidal defense strategy against infection in a structured habitatSci Rep 2 238
114 MakovetsS TitheradgeAJ and MurrayNE (1998) ClpXand ClpP are essential for the efficient acquisition of genesspecifying type IA and IB restriction systems Mol Microbiol28 25ndash35
115 MurrayNE (2002) Immigration control of DNA in bacteriaself versus non-self Microbiology 148 3ndash20
116 MakovetsS PowellLM TitheradgeAJB BlakelyGW andMurrayNE (2004) Is modification sufficient to protect abacterial chromosome from a resident restriction endonucleaseMol Microbiol 51 135ndash147
117 OrgelLE and CrickFHC (1980) Selfish DNA the ultimateparasite Nature 284 604ndash607
118 KidwellMG and LischDR (2001) Perspective transposableelements parasitic DNA and genome evolution Evolution 551ndash24
119 SadykovM AsamiY NikiH HandaN ItayaMTanokuraM and KobayashiI (2003) Multiplication of arestriction-modification gene complex Mol Microbiol 48417ndash427
120 ChaoL and LevinBR (1981) Structured habitats and theevolution of anticompetitor toxins in bacteria Proc Natl AcadSci USA 78 6324ndash6328
121 MagnusonRD (2007) Hypothetical functions of toxin-antitoxinsystems J Bacteriol 189 6089ndash6092
122 RifatD WrightNT VarneyKM WeberDJ and BlackLW(2008) Restriction endonuclease inhibitor IPI of bacteriophageT4 a novel structure for a dedicated target J Mol Biol 375720ndash734
123 LabrieSJ SamsonJE and MoineauS (2010) Bacteriophageresistance mechanisms Nat Rev Microbiol 8 317ndash327
124 LinLF PosfaiJ RobertsRJ and KongH (2001)Comparative genomics of the restriction-modification systemsin Helicobacter pylori Proc Natl Acad Sci USA 982740ndash2745
125 MilkmanR RaleighEA McKaneM CrydermanDBilodeauP and McWeenyK (1999) Molecular evolution of theEscherichia coli chromosome V Recombination patterns amongstrains of diverse origin Genetics 153 539ndash554
126 HumbertO DorerMS and SalamaNR (2011)Characterization of Helicobacter pylori factors that controltransformation frequency and integration length during inter-strain DNA recombination Mol Microbiol 79 387ndash401
127 BarcusVA TitheradgeAJB and MurrayNE (1995) Thediversity of alleles at the hsd locus in natural populations ofEscherichia coli Genetics 140 1187ndash1197
128 SharpPM KelleherJE DanielAS CowanGM andMurrayNE (1992) Roles of selection and recombination in theevolution of type I restriction-modification systems inenterobacteria Proc Natl Acad Sci USA 89 9836ndash9840
129 LoenenWAM DanielAS BraymerHD and MurrayNE(1987) Organization and sequence of the hsd genes ofEscherichia coli K-12 J Mol Biol 198 159ndash170
130 SuriB and BickleTA (1985) EcoA the first member of a newfamily of type I restriction modification systems Geneorganization and enzymatic activities J Mol Biol 186 77ndash85
131 BertaniG and WeigleJJ (1953) Host-controlled variation inbacterial viruses J Bacteriol 65 113ndash121
132 SibleyMH and RaleighEA (2004) Cassette-like variation ofrestriction enzyme genes in Escherichia coli C and relativesNucleic Acids Res 32 522ndash534
133 MakarovaKS WolfYI SnirS and KooninEV (2011)Defense islands in bacterial and archaeal genomes and predictionof novel defense systems J Bacteriol 193 6039ndash6056
134 FurutaY AbeK and KobayashiI (2010) Genome comparisonand context analysis reveals putative mobile forms of restriction-modification systems and related rearrangements Nucleic AcidsRes 38 2428ndash2443
135 FurutaY and KobayashiI (2012) Restriction-modificationsystems as moblie epigenetic elements In RobertsAP andMullanyP (eds) Bacterial Integrative Mobile Genetic ElementsLandes Bioscience pp 85ndash103
136 OrsquoSullivanDJ and KlaenhammerTR (1998) Control ofexpression of LlaI restriction in Lactococcus lactis MolMicrobiol 27 1009ndash1020
137 WhitakerRD DornerLF and SchildkrautI (1999) A mutantof BamHI restriction endonuclease which requires N6-methyladenine for cleavage J Mol Biol 285 1525ndash1536
Nucleic Acids Research 2014 Vol 42 No 1 69
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
glucose moiety is attached Conserved residues Y64 andL68 were noted to make van der Waals contact with themethyl group of the flipped out m5C these contacts aremissing when the pocket contains CThe flipping action can be compared with but is distinct
from that of eukaryotic m5C-specific regulatory proteinsthat use the SET and Ring-finger-Associated (SRA)domain to read DNA modification state (Figure 4A)This domain is found in most eukaryotes in accessoryproteins (eg UHRF1NP95SUVH5) of the DNMT1maintenance MTase (93ndash95) Despite the similarstrategy the McrB-N domain is not homologous butdisplays a distinct protein fold (60) Binding is accom-plished from the minor groove and extraction of the Ccreates a 30 bend toward the major groove resembling aglycosylase in this respect (Figure 4A) The eukaryoticproteins form a crescent from which loops project towrap around the DNA with recognition mediatedthrough both major and minor grooves (94) For McrB-N the authors suggest that the purine preference in the 50
position might result from flexibility constraints or inter-action with a non-conserved aa that occupies space left bythe flipped base Substitutions of this aa (Y41A or Y41Q)compromised binding activity
Sequence specificity novel phenotype and structuralmodel of MrrIn 1987 Heitman and Model discovered Mrr when theyfound that transfer of various foreign m6A MTasesinduced an SOS response due to DNA damage (42)This response to the presence of an incompatible MTaseremains the principal evidence that the E coli K12 Mrrprotein cleaves DNA Related proteins discussed later inthe text (Type IIM) have been more tractable for in vitrowork No concise description of the Mrr recognitionsequence has been forthcoming although several studieshave examined the spectrum of incompatible MTases(43669697) Both adenine and cytosine MTases confersensitivityMrr is also responsible for DNA damage that does not
depend on methylation at all foreign or otherwise Highhydrostatic pressure (HP) induces the SOS DNA damageresponse and lethality (98) The response did not dependon the activity of the endogenous MTases of E coli K12but did depend on both the presence of wild-type mrr andthe integrity of the SOS signal generation pathwayPossibly HP elicits a non-enzymatic modification or astructural change in DNA helicity that is acted on byMrr This HP phenotype was used to characterize mrrmutants which were fitted into a computer-assistedmodel of the Mrr protein (67) An N-terminal DNA-binding winged helix was proposed with a C-terminalnuclease domain previously identified (69) The functionalimportance of several conserved residues was confirmedSeveral of the selected mutants with null phenotypes wereisolated in a region far from the active site or bindingsurface identified bioinformatically These could affectinteraction with a component of the HP response Thisintriguing collection of informative mutants will illumin-ate in vitro characterization
Type IIM binding domainsType IIM enzymes of two families are well-characterizedwith respect to cleavage (Table 4) Crystal structures forboth have recently appeared
DpnI winged-helix DNA recognitionUnusually for modification-dependent enzymes DpnIcleaves a four-base site (Gm6ATC) with high fidelity(7799) to leave blunt ends when both strands of the siteare methylated At low concentration the enzyme nicksthe modified strand of a hemimethylated site (100) Thebehavior of the enzyme with respect to modificationpatterns within the canonical GATC sitemdashmodificationof C or A one strand or bothmdashhas been thoroughlyexplored (50) However only recently has cleavage ofnon-canonical adenine-methylated sites been examinedSiwek and co-workers (78) found evidence for consider-able relaxation of specificity at the outer base This experi-ment used substrates modified by a highly non-specificadenine MTase extensive DpnI cleavage cloning of thefragments and sequencing of the borders
Structure determination in the presence of DNA andvalidation experiments (78) place this enzyme togetherwith the other modification-dependent enzymes in thattwo domains segregate the cleavage function fromsequence recognition It also separates DpnI from theothers in that the cleavage domain also possesses somemodification and sequence specificity The main recogni-tion is accomplished by a monomeric winged-helixdomain which binds in the major groove and recognizesthe modifications on both strands in the same event Thestructure does not reveal a cleavage-competent complexhowever because the cleavage domain is far from theDNA Filter-binding experiments validated the ability ofthe C-terminal domain to bind alone to do so moretightly to fully methylated than to hemimethylated oligo-nucleotides and to compete with the full-length enzymereducing cleavage by it Expression of the N-terminalcleavage domain alone (in low yield) allowed validationof its cleavage activity Surprisingly this cleavage wasitself dependent on modification state and sequence ofthe substrate Modeling based on the structure of theblunt-end-producing Type IIP enzyme PvuII allowed pre-diction that the cleavage domain approaches from theminor groove Complete understanding of double-stranded cleavage will depend on understanding thedynamic transformations that allow the cleavage domainto approach and act at the site
MspJI coupling of cleavage with DNA recognitionThe six members of the MspJI family use the Mrr-catversion of the PD-(DE)XK nuclease to cut at definedlocations to one side relative to the modified base(12 bases on the modified strand 16ndash17 on the otherFigure 5A) only one modified base is required fordouble-strand cleavage to occur (unlike McrBC) (5152)However these enzymes are stimulated by the presence ofa second site in cis or in trans Symmetrically modifiedsites (such as m5CpGm5CpG in mammalian DNA)yield prominent bands of defined size (32 bp) containinga mixed population of sequences each with a m5C in the
62 Nucleic Acids Research 2014 Vol 42 No 1
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
middle (Figure 5B) This behavior is recapitulated by thePvuRts1I group of enzymes (exemplified by AbaSDFI inFigure 5C) except that the distances are shorter andrecognition of modification state is less well understoodDuring the characterization of MspJI Dcm
(Cm5CWGG) sites were the first recognized substrateyielding a clear banding pattern (51) Cleavage of differ-ently modified plasmids and designed oligonucleotidesubstrates allowed a good assessment of both modificationand sequence specificity This family shows preference forparticular bases nearby similar to McrBC
MspJI DNA recognition is mediated by anSRA-like domainRecently the crystal structure of MspJI without DNA hasbeen resolved at 205 A (80) Search of the MolecularModeling Database at NCBI (101) using VAST (102)
Figure 4 McrB-N in comparison to other base-flipping proteins (A)SRA domains SUVH5 (3Q0C) and UHRF1 (2ZKD) use loops extend-ing from a crescent formed from two beta sheets to flip C or m5C fromundeformed B-form DNA into a pocket (top row) whereas McrB-N(3SSC bottom row) uses loops from one beta-sheet to distort the DNAand flip the base It resembles the human alkyladenine glycosylase(1BNK) (bottom row) in bending the DNA toward the majorgroove while flipping the base via the minor groove Figure 5 ofSukackaite et al (60) (B) The SRA-like hemi-methylated 5mC recog-nition domains A ribbon model of the N-terminal domain of theMspJI structure (4F0Q and 4F0P left) compared with the SRAdomain of URHF1 (PDB 3FDE right) The crescent shape formed
Figure 4 Continuedby interacting beta sheets and helices aB and aC are the conservedfeatures of the SRA domain highlighted here Loops on the concaveside of UHRF1 participate in flipping the base and similar loops pre-sumably do so for MspJI Two of these vary in length among familymembers and may play roles in sequence context specificity Figure 2aand b from Horton et al (80)
Figure 5 Schematic diagrams of cleavage positions for MspJI andAbaSDFI Cleavage of both strands is elicited by a singly modifiedsite for both MspJI (A) and AbaSDFI (C) Cleavage position is fixedrelative to the modified site but with a four-base 50 extension forMspJI and a two-base 30 extension for AbaSDFI When a site is sym-metrically-modified (as for CpG sites in mammalian DNA) a 32 base-pair fragment is excised from the DNA (B) (A) Figure 2a and (B)Figure 3a reprinted with permission from Cohen-Karni et al (52)(C) Figure 5a from Wang et al (47)
Nucleic Acids Research 2014 Vol 42 No 1 63
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
showed that the N-terminal domain was structurallysimilar to that of the eukaryotic SRA domain with acrescent-shaped beta-sheet structure from which loopsproject (see Figure 4B and discussion earlier in the textMcrB) This structural homology allowed modeling of theDNA-bound structure with a flipped m5C The enzyme inthe crystal is a tetramer in which two monomers form aback-to-back dimer via the C-terminal regions thatcomprise the endonuclease Two back-to-back dimersgenerate a tetrameric protein with two cleavage domainspositioned (as in the Type IIP enzyme HindIII used formodeling the C-terminal cleavage domain interaction withDNA) so that a 4-base 50 extension would be created oncleavage of modeled DNA Cleavage is most efficient atmolar ratios that allow all four SRA-like domains to beoccupiedmdashtoo much enzyme prevents cleavage fromoccurring
Tracking and dimerization
McrBC as translocaseBourniquel and Bickle (84) have reviewed much of theenzymology of McrBC which will be briefly summarizedhere The Raleigh Bickle and Pingoud laboratories havecontributed to the following consistent picture of thein vitro reaction EcoKMcrBC cleavage results in adouble-strand cut near one RmC site (72ndash74) butrequires cooperation of two sites (3940) or a translocationblock (73) The sites may be on different daughters acrossa fork (75) These are separated by 30ndash3000 bp (397274)and may be on either strand (3976) cleavage occurs30ndash35 bases from the modified base with oppositenicks not tightly constrained (73) and minor cleavageclusters are found 40 50 and 60 nt from the m5C(74) hm5C DNA elicits cleavage also (39) A ring struc-ture is formed by 5ndash7 molecules of McrB in the presence ofGTP (Figure 3) (103) this complex can bind to a recog-nition element in DNA In the presence of McrC trans-location of the complex occurs and cleavage ensues whentranslocation is blocked Collision of translocatingcomplexes a protein barrier or a topological barrier willelicit double-strand cleavage adjacent to one recognitionelement or the other The enzyme will cleave when recog-nition elements are on opposite sides of a forked structure(75) This would allow action in vivo to prevent entry of aMTase gene even with rare sitesStructurally the McrB protein is proposed to be a
member of the AAA+ protein family of NTPases (104)many of which form ring-shaped complexes and partici-pate in molecular machines lsquoSensorrsquo segments found inthese proteins have been shown in some cases to playroles in coupling NTPase activity to intersubunit commu-nication and movement (105) Two of three elements ofthe GTP-binding motif proposed by Dila et al (61) werevalidated by mutational analysis (6562) The thirdproposed motif element was identified as amino acidsNTAD by Dila et al Alignment of AAA+ NTPases in(104) found this aligned with the motif designatedSensor-1 in (105) An interesting result was that mutationshere unexpectedly appeared to abrogate interaction withMcrC instead of changing which NTP would be
productive (62) It may instead play a role in coordinatingGTP binding and hydrolysis with DNA binding inter-action with McrC and cleavage
Intracellularly the story becomes more complex as themcrB gene encodes two products of 51 and 33 kD McrB-L and McrB-S the latter one starting from an in-frameinternal translation start site (106) Both in vivo andin vitro McrB-S can interfere with the function ofMcrB-L at least in part by forming complexes withMcrC unable to bind DNA (107) Both species can formmultimeric rings in the presence of GTP (103) as is usualfor AAA+NTPases (104)
SauUSI requires two sites and ATP hydrolysisSauUSI was originally annotated as a putative helicasefrom Staphylococcus aureus sp A single polypeptide issufficient for activity both in vitro and also in vivo as aclone in E coli using modified phage as a challenge Theamino acid sequence contains a PLDc domain at theN-terminus This contains a phosphodiesterase motiforiginally identified in Phospholipase D (108) it wasvalidated by mutagenesis of four catalytic residue candi-dates In the middle ATPase and helicase motifs wereproposed to account for ATP dependence of cleavageactivity A Domain of Unknown Function was identifiedat the C-terminus (Pfam DUF3427) (108) and wasproposed to recognize the substrate (45)
The purified enzyme cleaves modified DNA containingm5C and hm5C but not m4C in the presence of ATP ordATP but not other nucleotides The negative result form4C is firm plasmids modified at the same site by an m5CMTase (Dcm) or an m4C MTase (MPspGI) were testedThe former (Cm5CWGG) was sensitive whereas the latterwas resistant Thus the sequence preference is likely tobe satisfied m6A is likely not a substrate but few m6A-containing sites were examined
Like McrBC SauUSI requires the presence of two sitesfor efficient cleavage Presumably the ATPase activityparticipates in monitoring the presence of two sites asfor other nucleotide-dependent REases includingMcrBC The mechanism of communication is unknownThe enzyme belongs to a family of highly similarorthologues found in other sequenced Staphylococci(Tables 1 and 2) and more distant homologues can befound in sequenced bacterial and archaeal genomes
EVOLUTIONARY PRESSURES ON R-M SYSTEMS
Evolution by selfish propagation
One way to understand the massive variety of restrictionsystems and their sporadic distribution is to locate theevolutionary drivers of enzyme diversification in theenzyme genes themselves as selfish elements Work fromthe Kobayashi laboratory has elaborated clear examplesof selfish behavior in some Type II enzyme systems(109110) in which the host becomes lsquoaddictedrsquo to theR-M system Once a cell has acquired an R-M pair lossof the genes results in death of that cellrsquos descendants asthe REase is frequently still present and able to act on thegenome following loss of methylation activity In this
64 Nucleic Acids Research 2014 Vol 42 No 1
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
perspective the role played by modification-dependentenzymes is host defense to exclude systems with lsquoforeignrsquoMTase patterns and prevent the cell from loading up withparasites The exclusion event is accompanied by the deathof the cell (111ndash113) Weak sequence specificity of Type IVenzymes could then result from the need to control entryof a wide variety of invading systems
The selfish aspect certainly plays a role in R-M popula-tion biology but cannot be the whole story Type II R-Msystems can still be lost by inactivation of the R gene firstMoreover Type I systems escape this scenario withcomplex control of cleavage activity the restrictionassembly includes a methylation assembly to begin withtherefore the R protein cannot act unless the MTase ispresent in addition failure of the methylation activity inan intact complex leads to abrogation of R activity some-times by action of the ClpXP protease specifically on theR protein (114ndash116)
Furthermore in population terms a cell that acquiredand became addicted to an R-M system should lose incompetition with a sibling that never received thesystem Two factors could counter this First acquisitioncould be accompanied by an increase in the total numberof copies of the R-M system in the population asproposed for invading transposable elements Thisoverreplication results in more copies of the systemcreated than are lost whether to suicide or to other select-ive disadvantage [see eg (117118)] R-M gene amplifica-tion within a cell has been reported experimentally (119)but spread in a population has not been demonstrated yetA second factor that could counter the disability of addic-tion is localization of competition In a structured envir-onment (colonies on a plate or biofilm on large or smallsurfaces) killing of segregants preserves limiting nutrientsfor lineages that retain the toxinantitoxin pair (120121)Much of the real world is structured so this is an import-ant condition
Evolution by phage-host arms race
A second perspective supposes that the modification-dependent Type IV enzymes arose from the competitivecoevolutionary interaction between phages and theirhosts This was first enunciated by Revel and Luria (2)and most recently by Black and coworkers (122) seealso (123) That is hosts used modification-blockedrestriction to defend against phage infection T-evenphages developed methods of substituting modified basesfor the ordinary ones hosts developed Type IV enzymes indefense phages added sugar or other modifications (19) tothwart Type IV enzymes hosts extended Type IV enzymesto accommodate these decorations finally phages de-veloped protein inhibitors specific for these enzymes aswell T4 phages deliver a protein inhibitor (IPI) alongwith the DNA on infection which allows growth in thepresence of EcoCTGmrSD The locus responsible for thisinhibitor is highly variable among relatives of T4 asgmrSD is in enteric bacteria (both in distribution and inaa sequence) When phage with different IP1 alleles weretested for protection from cloned EcoCTGmrSD and itshomolog EcoUTGmrSD specificity was evident one or
the other or both or neither of the two homologs wascounteracted in individual cases (122) This variability ofthe outcome supports the idea that phage-host interactiondrives at least some of these developmentsIn this perspective the weak sequence selectivity of the
Type IV systems might simply reflect the lack of endogen-ous targets for the enzyme As the host does not present anyhm5C and the phage is completely substituted selection forsequence-specificity is weak Selection would act to spareany co-residentMTases This differs fromType II enzymeswhere the MTase and REase must co-evolve to allow thehost to survive Each Type IV system is compatible withsome suite of Type I-IIIMTases (and thus the R-M systemsas a whole) Methylated or hydroxymethylated bases maynot be recognized at all (EcoCTGmrSD) or the systemmay require one specific base in addition to the modifiedone (McrA McrBC MspJI and PvuRts1I) MTases mod-ifying sites not including that base are then compatible asDcm (CmCWGG) is compatible with the McrA McrBCand Mrr systems in E coli K12Type IV systems that restrict methylated bases in a
weakly specified sequence context confer an additionaladvantage in competition with phages Many phagessuch as have not evolved the nucleotide-substitutionstrategy used by the T-even phages These phagesnormally carry the modification pattern of the mostrecent host if the last host expressed an MTase creatinga susceptible site the Type IV enzyme of the new host willdestroy the invader and limit the infection This may beaccompanied by the death of the individual infectedtherefore protection can be conferred on the sibling popu-lation (111)A further implication of this scenario considers the fate
of a population invaded by phage Phage survival ofrestriction occurs at biologically relevant frequencies(106-102) The survivors of restriction carry the particu-lar methylation pattern of the particular cell and thus areresistant to all restriction systems it might have carried(Types IndashIV) A bacterial population as a whole thenbenefits from mechanisms that diversify the suite ofR-M systems so that such surviving phage do not havefree access to the entire population The extreme variabil-ity of R-M system content in isolates of the same species iscompatible with this idea [see eg (124) REBASEGenomes httptoolsnebcomvinczegenomes] Suchvariability also limits and shapes interstrain genetransfer (115125126)The Raleigh laboratory has built on elegant genetic
work with Type I enzymes in the Murray laboratory(127ndash130) to investigate a locus designated thelsquoImmigration Control locusrsquo or ICR that exemplifiesvariable R-M content Alternative DNA segments con-taining R-M systems are located at the same definedlocation in most E coli chromosomes between the yjiSand yjiA genes The ICR in the non-restricting E coli Cstrain [used in the original definition of the R-M phenom-enon (131)] contains a remnant of a Type I enzyme R geneand is 13 kilobase shorter than the same region in E coliK12 (132) The mechanism of segment replacement is stillunknown The ICR would be an example of the lsquodefenseislandsrsquo analyzed by the Koonin group (133) lsquoDefense
Nucleic Acids Research 2014 Vol 42 No 1 65
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
islandsrsquo contain genes that can defend against phage orother invading DNA these exhibit bioinformaticproperties similar to lsquomobilome islandsrsquo containing mobil-ization genes (transposases for example) However themechanism of mobilization has not been identified forthe ICR
FINAL THOUGHTS
The extreme diversity of R-M systems that recognizeordinary DNA seems likely to be approached by the di-versity of Type IV restriction systems Type IV enzymesare hard to find as most detection methods depend ondevelopment of genetic systems for each taxon or on ser-endipity Those characterized so far mostly stem frominitial genetic investigation of limits on infection trans-formation or transduction Barriers encountered provideleads to the genes involved Bioinformatic analysis hashelped to identify relatives which may be more tractableto biochemical investigation than the example originallyfound This approach has pitfalls the gene encodingMspJI was first thought to code for an enzyme recognizingan unmodified site because it is immediately adjacent toan (inactive it is now thought) cytosine MTase geneProvidentially the first expression host was devoid of sen-sitive sites whereas the first test substrate contained some(51) A combination of biological experiments with bio-informatics and biochemistry will be needed to reveal thefull spectrum of Type IV enzymes that may lurk within thevast universe of unidentified ORFs in bacterial systemsOne might begin with those strains whose genomes carryfew Type II systems Bacillus or Corynebacterium asopposed to Helicobacter or Neisseria [see the Genomessection of REBASE (50)]The role of lsquodefense islandsrsquo and their relation to the
lsquomobilomersquo in bacterial population biology remains to bedetermined If a defense island is similar to a mobilomeisland there should be a mechanism of mobilizationnearby which would boost the contribution oflsquooverreplicationrsquo to the account of selections acting onR-M systems R-M systems of all types can be found onor adjacent to known mobilizing elements (134135) buthave not been shown to move experimentallyOn another note it may turn out that evolutionarily
there is a continuum between the apparently modifica-tion-dependent and modification-blocked paths Onerelative of McrBC predicted by bioinformatics analysesis LlaI a system that recognizes an unmodified sequenceand requires two MTases to support it (136) The enzymeBamHI prefers to cleave DNA with m6A in its GGATCCsite and mutants can be isolated that require this modifiedbase (137) Are there native systems similarly protected bymodification of one position in the recognition site butdependent on modification at a different one An interest-ing evolutionary series can be imagined
ACKNOWLEDGEMENTS
The authors thank Rich Roberts and David Dryden forencouragement and critical comment They are grateful to
Tom Bickle Virginius Siksnys Yu Zheng and XiadongCheng for permission to reproduce figures They alsothank Ivan Correa for help drawing base structures andTasha Jose for illustration assistance They also thank theanonymous reviewers of a previous manuscript for caretaken with the review process greatly improving thepresent document
FUNDING
New England Biolabs (EAR) Funding for open accesscharge New England Biolabs (to EAR)
Conflict of interest statement None declared
REFERENCES
1 LuriaSE and HumanML (1952) A nonhereditary host-inducedvariation of bacterial viruses J Bacteriol 64 557ndash559
2 RevelHR and LuriaSE (1970) DNA-glucosylation in T-evenphage genetic determination and role in phage-host interactionAnnu Rev Genet 4 177ndash192
3 RaleighEA and WilsonG (1986) Escherichia coli K-12 restrictsDNA containing 5-methylcytosine Proc Natl Acad Sci USA83 9070ndash9074
4 Noyer-WeidnerM DiazR and ReinersL (1986) Cytosine-specific DNA modification interferes with plasmid establishmentin Escherichia coli K12 involvement of rglB Mol Gen Genet205 469ndash475
5 BlumenthalRM GregorySA and CooperiderJS (1985)Cloning of a restriction-modification system from Proteus vulgarisand its use in analyzing a methylase-sensitive phenotype inEscherichia coli J Bacteriol 164 501ndash509
6 SohailA LiebM DarM and BhagwatAS (1990) A generequired for very short patch repair in Escherichia coli is adjacentto the DNA cytosine methylase gene J Bacteriol 1724214ndash4221
7 MacNeilDJ (1988) Characterization of a unique methyl-specificrestriction system in Streptomyces avermitilis J Bacteriol 1705607ndash5612
8 VertesAA InuiM KobayashiM KurusuY and YukawaH(1993) Presence of mrr- and mcr-like restriction systems incoryneform bacteria Res Microbiol 144 181ndash185
9 MacalusoA and MettusAM (1991) Efficient transformation ofBacillus thuringiensis requires nonmethylated plasmid DNAJ Bacteriol 173 1353ndash1356
10 HolmesML NuttallSD and Dyall-SmithML (1991)Construction and use of halobacterial shuttle vectors and furtherstudies on Haloferax DNA gyrase J Bacteriol 173 3807ndash3813
11 PosfaiJ BhagwatAS PosfaiG and RobertsRJ (1989)Predictive motifs derived from cytosine methyltransferases NucleicAcids Res 17 2421ndash2435
12 MaloneT BlumenthalRM and ChengX (1995) Structure-guided analysis reveals nine sequence motifs conserved amongDNA amino-methyltransferases and suggests a catalyticmechanism for these enzymes J Mol Biol 253 618ndash632
13 TiminskasA ButkusV and JanulaitisA (1995) Sequence motifcharacteristics for DNA [cytosine-N4] and DNA [adenine-N6]methyltransferases Classification of all DNA methyltransferasesGene 157 3ndash11
14 DickmanMJ InglestonSM SedelnikovaSE RaffertyJBLloydRG GrasbyJA and HornbyDP (2002) The RuvABCresolvasome Eur J Biochem 269 5492ndash5501
15 KovallRA and MatthewsBW (1998) Structural functionaland evolutionary relationships between lambda-exonuclease andthe type II restriction endonucleases Proc Natl Acad Sci USA95 7893ndash7897
16 RobertsRJ ChangYC HuZ RachlinJN AntonBPPokrzywaRM ChoiHP FallerLL GuleriaJ HousmanGet al (2010) COMBREX a project to accelerate the functional
66 Nucleic Acids Research 2014 Vol 42 No 1
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
annotation of prokaryotic genomes Nucleic Acids Res 39D11ndashD14
17 RaePM and SteeleRE (1978) Modified bases in the DNAs ofunicellular eukaryotes an examination of distributions andpossible roles with emphasis on hydroxymethyluracil indinoflagellates Biosystems 10 37ndash53
18 WarrenRAJ (1980) Modified bases in bacteriophage DNAsAnnu Rev Microbiol 34 137ndash158
19 Gommers-AmptJH and BorstP (1995) Hypermodified bases inDNA FASEB J 9 1034ndash1042
20 SwintonD HattmanS CrainPF ChengCS SmithDL andMcCloskeyJA (1983) Purification and characterization of theunusual deoxynucleoside alpha-N-(9-beta-D-20-deoxyribofuranosylpurin-6-yl)glycinamide specified by the phageMu modification function Proc Natl Acad Sci USA 807400ndash7404
21 TahilianiM KohKP ShenY PastorWA BandukwalaHBrudnoY AgarwalS IyerLM LiuDR AravindL et al(2009) Conversion of 5-methylcytosine to 5-hydroxymethylcytosinein mammalian DNA by MLL partner TET1 Science 324930ndash935
22 KriaucionisS and HeintzN (2009) The nuclear DNA base5-hydroxymethylcytosine is present in Purkinje neurons and thebrain Science 324 929ndash930
23 BorstP and SabatiniR (2008) Base J discovery biosynthesisand possible functions Annu RevMicrobiol 62 235ndash251
24 KaminskaKH and BujnickiJM (2008) Bacteriophage MuMom protein responsible for DNA modification is a new memberof the acyltransferase superfamily Cell Cycle 7 120ndash121
25 IyerLM TahilianiM RaoA and AravindL (2009) Predictionof novel families of enzymes involved in oxidative and othercomplex modifications of bases in nucleic acids Cell Cycle 81698ndash1710
26 ZhouX HeX LiangJ LiA XuT KieserT HelmannJDand DengZ (2005) A novel DNA modification by sulphur MolMicrobiol 57 1428ndash1438
27 OuHY HeX ShaoY TaiC RajakumarK and DengZ(2009) dndDB A Database Focused on Phosphorothioation ofthe DNA Backbone PloS One 4 e5132
28 XuT LiangJ ChenS WangL HeX YouD WangZLiA XuZ ZhouX et al (2009) DNA phosphorothioation inStreptomyces lividans mutational analysis of the dnd locusBMC Microbiol 9 41
29 WangL ChenS VerginKL GiovannoniSJ ChanSWDeMottMS TaghizadehK CorderoOX CutlerMTimberlakeS et al (2011) DNA phosphorothioation iswidespread and quantized in bacterial genomes Proc Natl AcadSci USA 108 2963ndash2968
30 Loslashbner-OlesenA SkovgaardO and MarinusMG (2005) Dammethylation coordinating cellular processes Curr OpinMicrobiol 8 154ndash160
31 LowDA and CasadesusJ (2008) Clocks and switches bacterialgene regulation by DNA adenine methylation Curr OpinMicrobiol 11 106ndash112
32 FreitagM and SelkerEU (2005) Controlling DNA methylationMany roads to one modification Curr Opin Genet Dev 15191ndash199
33 ProhaskaSJ StadlerPF and KrakauerDC (2010) Innovationin gene regulation The case of chromatin computation J TheorBiol 265 27ndash44
34 SeshasayeeASN SinghP and KrishnaS (2012) Context-dependent conservation of DNA methyltransferases in bacteriaNucleic Acids Res 40 7066ndash7073
35 AntonBP (2011) Dissertation Boston University BostonMassachusetts USA
36 RaleighEA TrimarchiR and RevelH (1989) Genetic andphysical mapping of the mcrA (rglA) and mcrB (rglB) loci ofEscherichia coli K-12 Genetics 122 279ndash296
37 Gonzalez-CeronG Miranda-OlivaresOJ and Servın-GonzalezL(2009) Characterization of the methyl-specific restriction system ofStreptomyces coelicolor A3(2) and of the role played by laterallyacquired nucleases FEMS Microbiol Lett 301 35ndash43
38 LiuG OuHY WangT LiL TanH ZhouXRajakumarK DengZ and HeX (2010) Cleavage of
phosphorothioated DNA and methylated DNA by the type IVrestriction endonuclease ScoMcrA PLoS Genet 6 e1001253
39 SutherlandE CoeL and RaleighEA (1992) McrBC amultisubunit GTP-dependent restriction endonuclease J MolBiol 225 327ndash348
40 KrugerT WildC and Noyer-WeidnerM (1995) McrB aprokaryotic protein specifically recognizing DNA containingmodified cytosine residues EMBO J 14 2661ndash2669
41 SitaramanR and LepplaSH (2012) Methylation-dependentDNA restriction in Bacillus anthracis Gene 494 44ndash50
42 HeitmanJ and ModelP (1987) Site-specific methylases inducethe SOS DNA repair response in Escherichia coli J Bacteriol169 3243ndash3250
43 Waite-ReesPA KeatingCJ MoranLS SlatkoBEHornstraLJ and BennerJS (1991) Characterisation andexpression of the Escherichia coli Mrr restriction systemJ Bacteriol 173 5207ndash5219
44 KerrAL JeonYJ SvensonCJ RogersPL and NeilanBA(2010) DNA restriction-modification systems in the ethanologenZymomonas mobilis ZM4 Appl MicrobiolBiotech 89 761ndash769
45 XuSY CorvagliaAR ChanSH ZhengY and LinderP(2011) A type IV modification-dependent restriction enzymeSauUSI from Staphylococcus aureus subsp aureus USA300Nucleic Acids Res 39 5597ndash5610
46 MonkIR ShahIM XuM TanMW and FosterTJ (2012)Transforming the untransformable application of directtransformation to manipulate genetically Staphylococcus aureusand Staphylococcus epidermidis mBio 3 e00277ndash00211
47 WangH GuanS QuimbyA Cohen-KarniD PradhanSWilsonG RobertsRJ ZhuZ and ZhengY (2011)Comparative characterization of the PvuRts1I family ofrestriction enzymes and their application in mapping genomic5-hydroxymethylcytosine Nucleic Acids Res 39 9294ndash9305
48 JanosiL YonemitsuH HongH and KajiA (1994) Molecularcloning and expression of a novel hydroxymethylcytosine-specificrestriction enzyme (PvuRts1I) modulated by glucosylation ofDNAJ Mol Biol 242 45ndash61
49 BairCL and BlackLW (2007) A type IV modificationdependent restriction nuclease that targets glucosylatedhydroxymethyl cytosine modified DNAs J Mol Biol 366768ndash778
50 RobertsRJ VinczeT PosfaiJ and MacelisD (2010)REBASEndasha database for DNA restriction and modificationenzymes genes and genomes Nucleic Acids Res 38 D234ndashD236
51 ZhengY Cohen-KarniD XuD ChinHG WilsonGPradhanS and RobertsRJ (2010) A unique family of Mrr-likemodification-dependent restriction endonucleases Nucleic AcidsRes 38 5527ndash5534
52 Cohen-KarniD XuD AponeL FomenkovA SunZDavisPJ KinneySRM Yamada-MabuchiM XuSYDavisT et al (2011) The MspJI family of modification-dependent restriction endonucleases for epigenetic studies ProcNatl Acad Sci USA 108 11040ndash11045
53 LubysA VitkuteJ LubieneJ and JanulaitisA (2011)Restriction enzymes and their applications US Patent ApplicationPublication 20110207139 A1 httpappftusptogovnetahtmlPTOsearch-boolhtml (27 August 2013 date last accessed)
54 TarasovaGV NayakshinaTN and DegtyarevSK (2008)Substrate specificity of new methyl-directed DNA endonucleaseGlaI BMC Mol Biol 9 7
55 MulliganEA and DunnJJ (2008) Cloning purification andinitial characterization of E coli McrA a putative5-methylcytosine-specific nuclease Protein ExprPurif 62 98ndash103
56 MulliganEA HatchwellE McCorkleSR and DunnJJ (2010)Differential binding of Escherichia coli McrA protein to DNAsequences that contain the dinucleotide m5CpG Nucleic AcidsRes 38 1997ndash2005
57 AntonBP and RaleighEA (2004) Transposon-mediated linkerinsertion scanning mutagenesis of the Escherichia coli McrAendonuclease J Bacteriol 186 5699ndash5707
58 BujnickiJM RadlinskaM and RychlewskiL (2000) Atomicmodel of the 5-methylcytosine-specific restriction enzyme McrAreveals an atypical zinc finger and structural similarity tobetabetaalphaMe endonucleases Mol Microbiol 37 1280ndash1281
Nucleic Acids Research 2014 Vol 42 No 1 67
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
59 GastFU BrinkmannT PieperU KrugerT NoyerWeidnerM and PingoudA (1997) The recognition of methylatedDNA by the GTP-dependent restriction endonuclease McrBCresides in the N-terminal domain of McrB Biol Chem 378975ndash982
60 SukackaiteR GrazulisS TamulaitisG and SiksnysV (2012)The recognition domain of the methyl-specific endonucleaseMcrBC flips out 5-methylcytosine Nucleic Acids Res 407552ndash7562
61 DilaD SutherlandE MoranL SlatkoB and RaleighEA(1990) Genetic and sequence organization of the mcrBC locus ofEscherichia coli K-12 J Bacteriol 172 4888ndash4900
62 PieperU BrinkmannT KrugerT Noyer WeidnerM andPingoudA (1997) Characterization of the interaction between therestriction endonuclease McrBC from E coli and its cofactorGTP J Mol Biol 272 190ndash199
63 PieperU SchweitzerT GrollDH GastFU and PingoudA(1999) The GTP-binding domain of McrB more than just avariation on a common theme J Mol Biol 292 547ndash556
64 BujnickiJM and RychlewskiL (2001) Grouping together highlydiverged PD-(DE)XK nucleases and identification of novelsuperfamily members using structure-guided alignment ofsequence profiles J Mol MicrobiolBiotech 3 69ndash72
65 PieperU and PingoudA (2002) A mutational analysis of thePD DEXK motif suggests that McrC harbors the catalyticcenter for DNA cleavage by the GTP-dependent restrictionenzyme McrBC from Escherichia coli Biochemistry 415236ndash5244
66 KelleherJE and RaleighEA (1991) A novel activity inEscherichia coli K-12 that directs restriction of DNA modified atCG dinucleotides J Bacteriol 173 5220ndash5223
67 OrlowskiJ MebrhatuMT MichielsCW BujnickiJM andAertsenA (2008) Mutational analysis and a structural model ofmethyl-directed restriction enzyme Mrr Biochem Biophys ResCommun 377 862ndash866
68 AravindL MakarovaKS and KooninEV (2000) SURVEYAND SUMMARY holliday junction resolvases and relatednucleases identification of new families phyletic distribution andevolutionary trajectories Nucleic Acids Res 28 3417ndash3432
69 BujnickiJM and RychlewskiL (2001) Identification of a PD-(DE)XK-like domain with a novel configuration of the endonucleaseactive site in the methyl-directed restriction enzyme Mrr and itshomologs Gene 267 183ndash191
70 SteczkiewiczK MuszewskaA KnizewskiL RychlewskiL andGinalskiK (2012) Sequence structure and functional diversity ofPD-(DE)XK phosphodiesterase superfamily Nucleic Acids Res40 7016ndash7045
71 BairCL RifatD and BlackLW (2007) Exclusion of glucosyl-hydroxymethylcytosine DNA containing bacteriophages isovercome by the injected protein inhibitor IPI J Mol Biol 366779ndash789
72 StewartFJ and RaleighEA (1998) Dependence of McrBCcleavage on distance between recognition elements Biol Chem379 611ndash616
73 PanneD RaleighEA and BickleTA (1999) The McrBCendonuclease translocates DNA in a reaction dependent on GTPhydrolysis J Mol Biol 290 49ndash60
74 PieperU GrollDH WunschS GastFU SpeckC MuckeNand PingoudA (2002) The GTP-dependent restriction enzymeMcrBC from Escherichia coli forms high-molecular masscomplexes with DNA and produces a cleavage patternwith a characteristic 10-base pair repeat Biochemistry 415245ndash5254
75 IshikawaK HandaN SearsL RaleighEA and KobayashiI(2011) Cleavage of a model DNA replication fork by amethyl-specific endonuclease Nucleic Acids Res 39 5489ndash5498
76 StewartFJ PanneD BickleTA and RaleighEA (2000)Methyl-specific DNA binding by McrBC a modification-dependent restriction enzyme J Mol Biol 298 611ndash622
77 LacksS and GreenbergB (1977) Complementary specificity ofrestriction endonucleases DpnI and DpnII with respect to DNAmethylation J Mol Biol 114 153ndash168
78 SiwekW CzapinskaH BochtlerM BujnickiJM andSkowronekK (2012) Crystal structure and mechanism
of action of the N6-methyladenine-dependent type IIMrestriction endonuclease RDpnI Nucleic Acids Res 407563ndash7572
79 LacksS and GreenbergB (1975) A deoxyribonuclease ofDiplococcus pneumoniae specific for methylated DNA J BiolChem 250 4060ndash4066
80 HortonJR MabuchiMY Cohen-KarniD ZhangXGriggsRM SamaranayakeM RobertsRJ ZhengY andChengX (2012) Structure and cleavage activity of the tetramericMspJI DNA modification-dependent restriction endonucleaseNucleic Acids Res 40 9763ndash9773
81 CymermanIA ObarskaA SkowronekKJ LubysA andBujnickiJM (2006) Identification of a new subfamily of HNHnucleases and experimental characterization of a representativemember HphI restriction endonuclease Proteins 65 867ndash876
82 GorbalenyaAE (1994) Self-splicing group I and group II intronsencode homologous (putative) DNA endonucleases of a newfamily Protein Sci 3 1117ndash1120
83 ChanSH OpitzL HigginsL OrsquoLoaneD and XuSY (2010)Cofactor requirement of HpyAV restriction endonuclease PloSOne 5 e9071
84 BourniquelAA and BickleTA (2002) Complex restrictionenzymes NTP-driven molecular motors Biochimie 84 1047ndash1059
85 OrsquoDriscollJ HeiterDF WilsonGG FitzgeraldGFRobertsR and van SinderenD (2006) A genetic dissection ofthe LlaJI restriction cassette reveals insights on a novelbacteriophage resistance system BMC Microbiol 6 40
86 SmithRM DiffinFM SaveryNJ JosephsenJ andSzczelkunMD (2009) DNA cleavage and methylation specificityof the single polypeptide restriction-modification enzyme LlaGINucleic Acids Res 37 7206ndash7218
87 InterthalH PouliotJJ and ChampouxJJ (2001) The tyrosyl-DNA phosphodiesterase Tdp1 is a member of the phospholipaseD superfamily Proc Natl Acad Sci USA 98 12009ndash12014
88 SasnauskasG HalfordSE and SiksnysV (2003) How the BfiIrestriction enzyme uses one active site to cut two DNA strandsProc Natl Acad Sci USA 100 6410ndash6415
89 SasnauskasG ConnollyBA HalfordSE and SiksnysV (2007)Site-specific DNA transesterification catalyzed by a restrictionenzyme Proc Natl Acad Sci USA 104 2115ndash2120
90 Marchler-BauerA ZhengC ChitsazF DerbyshireMKGeerLY GeerRC GonzalesNR GwadzM HurwitzDILanczyckiCJ et al (2012) CDD conserved domains andprotein three-dimensional structure Nucleic Acids Res 41D348ndashD352
91 KravetzAN ZakharovaMV BeletskayaIV SinevaEVDenjmuchametovMM PetrovSI GlatmanLI andSoloninAS (1993) The cleavage sites and localization of genesencoding the restriction endonucleases Eco1831I and EcoHIGene 129 153ndash154
92 KlimasauskasS KumarS RobertsRJ and ChengX (1994)HhaI methyltransferase flips its target base out of the DNA helixCell 76 357ndash369
93 HashimotoH VertinoPM and ChengX (2010) Molecularcoupling of DNA methylation and histone methylationEpigenomics 2 657ndash669
94 RajakumaraE LawJA SimanshuDK VoigtPJohnsonLM ReinbergD PatelDJ and JacobsenSE (2011)A dual flip-out mechanism for 5mC recognition by theArabidopsis SUVH5 SRA domain and its impact on DNAmethylation and H3K9 dimethylation in vivo Genes Dev 25137ndash152
95 SharifJ and KosekiH (2011) Recruitment of Dnmt1 roles of theSRA protein Np95 (Uhrfi) and other factors Prog Mol BiolTransl Sci 101 289ndash310
96 Tesfazgi MebrhatuM WywialE GhoshA MichielsCWLindnerAB TaddeiF BujnickiJM Van MelderenL andAertsenA (2011) Evidence for an evolutionary antagonismbetween Mrr and Type III modification systems Nucleic AcidsRes 39 5991ndash6001
97 GrantGN JesseeJ BloomFR and HanahanD (1990)Differential plasmid rescue from transgenic mouse DNAs intoEscherichia coli methylation-restriction mutants Proc Natl AcadSci USA 87 4645ndash4649
68 Nucleic Acids Research 2014 Vol 42 No 1
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
98 AertsenA and MichielsCW (2005) Mrr instigates the SOSresponse after high pressure stress in Escherichia coli MolMicrobiol 58 1381ndash1391
99 WeiH TherrienC BlanchardA GuanS and ZhuZ (2008)The Fidelity Index provides a systematic quantitation of staractivity of DNA restriction endonucleases Nucleic Acids Res36 e50
100 LuL PatelH and BisslerJJ (2002) Optimizing DpnIdigestion to detect replicated DNA Biotechniques 33 316ndash318
101 MadejT AddessKJ FongJH GeerLY GeerRCLanczyckiCJ LiuC LuS Marchler-BauerAPanchenkoAR et al (2011) MMDB 3D structures andmacromolecular interactions Nucleic Acids Res 40 D461ndashD464
102 GibratJF MadejT and BryantSH (1996) Surprisingsimilarities in structure comparison Curr Opin Struct Biol 6377ndash385
103 PanneD MullerSA WirtzS EngelA and BickleTA (2001)The McrBC restriction endonuclease assembles into a ringstructure in the presence of G nucleotides EMBO J 203210ndash3217
104 NeuwaldAF AravindL SpougeJL and KooninEV (1999)AAA+ A class of chaperone-like ATPases associated with theassembly operation and disassembly of protein complexesGenome Res 9 27ndash43
105 OguraT and WilkinsonAJ (2001) AAA+ superfamilyATPases common structurendashdiverse function Genes Cells 6575ndash597
106 BearyTP BraymerHD and AchbergerEC (1997) Evidenceof participation of McrB(S) in McrBC restriction in Escherichiacoli K-12 J Bacteriol 179 7768ndash7775
107 PanneD RaleighEA and BickleTA (1998) McrBs amodulator peptide for McrBC activity EMBO J 175477ndash5483
108 PuntaM CoggillPC EberhardtRY MistryJ TateJBoursnellC PangN ForslundK CericG ClementsJ et al(2011) The Pfam protein families database Nucleic Acids Res40 D290ndashD301
109 NaitoT KusanoK and KobayashiI (1995) Selfish behavior ofrestriction-modification systems Science 267 897ndash899
110 KobayashiI (2001) Behavior of restriction-modification systemsas selfish mobile elements and their impact on genome evolutionNucleic Acids Res 29 3742ndash3756
111 FukudaE KaminskaKH BujnickiJM and KobayashiI(2008) Cell death upon epigenetic genome methylation a novelfunction of methyl-specific deoxyribonucleases Genome Biol 9R163
112 IshikawaK FukudaE and KobayashiI (2010) Conflictstargeting epigenetic systems and their resolution by cell deathnovel concepts for methyl-specific and other restriction systemsDNA Res 17 325ndash342
113 FukuyoM SasakiA and KobayashiI (2012) Success of asuicidal defense strategy against infection in a structured habitatSci Rep 2 238
114 MakovetsS TitheradgeAJ and MurrayNE (1998) ClpXand ClpP are essential for the efficient acquisition of genesspecifying type IA and IB restriction systems Mol Microbiol28 25ndash35
115 MurrayNE (2002) Immigration control of DNA in bacteriaself versus non-self Microbiology 148 3ndash20
116 MakovetsS PowellLM TitheradgeAJB BlakelyGW andMurrayNE (2004) Is modification sufficient to protect abacterial chromosome from a resident restriction endonucleaseMol Microbiol 51 135ndash147
117 OrgelLE and CrickFHC (1980) Selfish DNA the ultimateparasite Nature 284 604ndash607
118 KidwellMG and LischDR (2001) Perspective transposableelements parasitic DNA and genome evolution Evolution 551ndash24
119 SadykovM AsamiY NikiH HandaN ItayaMTanokuraM and KobayashiI (2003) Multiplication of arestriction-modification gene complex Mol Microbiol 48417ndash427
120 ChaoL and LevinBR (1981) Structured habitats and theevolution of anticompetitor toxins in bacteria Proc Natl AcadSci USA 78 6324ndash6328
121 MagnusonRD (2007) Hypothetical functions of toxin-antitoxinsystems J Bacteriol 189 6089ndash6092
122 RifatD WrightNT VarneyKM WeberDJ and BlackLW(2008) Restriction endonuclease inhibitor IPI of bacteriophageT4 a novel structure for a dedicated target J Mol Biol 375720ndash734
123 LabrieSJ SamsonJE and MoineauS (2010) Bacteriophageresistance mechanisms Nat Rev Microbiol 8 317ndash327
124 LinLF PosfaiJ RobertsRJ and KongH (2001)Comparative genomics of the restriction-modification systemsin Helicobacter pylori Proc Natl Acad Sci USA 982740ndash2745
125 MilkmanR RaleighEA McKaneM CrydermanDBilodeauP and McWeenyK (1999) Molecular evolution of theEscherichia coli chromosome V Recombination patterns amongstrains of diverse origin Genetics 153 539ndash554
126 HumbertO DorerMS and SalamaNR (2011)Characterization of Helicobacter pylori factors that controltransformation frequency and integration length during inter-strain DNA recombination Mol Microbiol 79 387ndash401
127 BarcusVA TitheradgeAJB and MurrayNE (1995) Thediversity of alleles at the hsd locus in natural populations ofEscherichia coli Genetics 140 1187ndash1197
128 SharpPM KelleherJE DanielAS CowanGM andMurrayNE (1992) Roles of selection and recombination in theevolution of type I restriction-modification systems inenterobacteria Proc Natl Acad Sci USA 89 9836ndash9840
129 LoenenWAM DanielAS BraymerHD and MurrayNE(1987) Organization and sequence of the hsd genes ofEscherichia coli K-12 J Mol Biol 198 159ndash170
130 SuriB and BickleTA (1985) EcoA the first member of a newfamily of type I restriction modification systems Geneorganization and enzymatic activities J Mol Biol 186 77ndash85
131 BertaniG and WeigleJJ (1953) Host-controlled variation inbacterial viruses J Bacteriol 65 113ndash121
132 SibleyMH and RaleighEA (2004) Cassette-like variation ofrestriction enzyme genes in Escherichia coli C and relativesNucleic Acids Res 32 522ndash534
133 MakarovaKS WolfYI SnirS and KooninEV (2011)Defense islands in bacterial and archaeal genomes and predictionof novel defense systems J Bacteriol 193 6039ndash6056
134 FurutaY AbeK and KobayashiI (2010) Genome comparisonand context analysis reveals putative mobile forms of restriction-modification systems and related rearrangements Nucleic AcidsRes 38 2428ndash2443
135 FurutaY and KobayashiI (2012) Restriction-modificationsystems as moblie epigenetic elements In RobertsAP andMullanyP (eds) Bacterial Integrative Mobile Genetic ElementsLandes Bioscience pp 85ndash103
136 OrsquoSullivanDJ and KlaenhammerTR (1998) Control ofexpression of LlaI restriction in Lactococcus lactis MolMicrobiol 27 1009ndash1020
137 WhitakerRD DornerLF and SchildkrautI (1999) A mutantof BamHI restriction endonuclease which requires N6-methyladenine for cleavage J Mol Biol 285 1525ndash1536
Nucleic Acids Research 2014 Vol 42 No 1 69
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
middle (Figure 5B) This behavior is recapitulated by thePvuRts1I group of enzymes (exemplified by AbaSDFI inFigure 5C) except that the distances are shorter andrecognition of modification state is less well understoodDuring the characterization of MspJI Dcm
(Cm5CWGG) sites were the first recognized substrateyielding a clear banding pattern (51) Cleavage of differ-ently modified plasmids and designed oligonucleotidesubstrates allowed a good assessment of both modificationand sequence specificity This family shows preference forparticular bases nearby similar to McrBC
MspJI DNA recognition is mediated by anSRA-like domainRecently the crystal structure of MspJI without DNA hasbeen resolved at 205 A (80) Search of the MolecularModeling Database at NCBI (101) using VAST (102)
Figure 4 McrB-N in comparison to other base-flipping proteins (A)SRA domains SUVH5 (3Q0C) and UHRF1 (2ZKD) use loops extend-ing from a crescent formed from two beta sheets to flip C or m5C fromundeformed B-form DNA into a pocket (top row) whereas McrB-N(3SSC bottom row) uses loops from one beta-sheet to distort the DNAand flip the base It resembles the human alkyladenine glycosylase(1BNK) (bottom row) in bending the DNA toward the majorgroove while flipping the base via the minor groove Figure 5 ofSukackaite et al (60) (B) The SRA-like hemi-methylated 5mC recog-nition domains A ribbon model of the N-terminal domain of theMspJI structure (4F0Q and 4F0P left) compared with the SRAdomain of URHF1 (PDB 3FDE right) The crescent shape formed
Figure 4 Continuedby interacting beta sheets and helices aB and aC are the conservedfeatures of the SRA domain highlighted here Loops on the concaveside of UHRF1 participate in flipping the base and similar loops pre-sumably do so for MspJI Two of these vary in length among familymembers and may play roles in sequence context specificity Figure 2aand b from Horton et al (80)
Figure 5 Schematic diagrams of cleavage positions for MspJI andAbaSDFI Cleavage of both strands is elicited by a singly modifiedsite for both MspJI (A) and AbaSDFI (C) Cleavage position is fixedrelative to the modified site but with a four-base 50 extension forMspJI and a two-base 30 extension for AbaSDFI When a site is sym-metrically-modified (as for CpG sites in mammalian DNA) a 32 base-pair fragment is excised from the DNA (B) (A) Figure 2a and (B)Figure 3a reprinted with permission from Cohen-Karni et al (52)(C) Figure 5a from Wang et al (47)
Nucleic Acids Research 2014 Vol 42 No 1 63
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
showed that the N-terminal domain was structurallysimilar to that of the eukaryotic SRA domain with acrescent-shaped beta-sheet structure from which loopsproject (see Figure 4B and discussion earlier in the textMcrB) This structural homology allowed modeling of theDNA-bound structure with a flipped m5C The enzyme inthe crystal is a tetramer in which two monomers form aback-to-back dimer via the C-terminal regions thatcomprise the endonuclease Two back-to-back dimersgenerate a tetrameric protein with two cleavage domainspositioned (as in the Type IIP enzyme HindIII used formodeling the C-terminal cleavage domain interaction withDNA) so that a 4-base 50 extension would be created oncleavage of modeled DNA Cleavage is most efficient atmolar ratios that allow all four SRA-like domains to beoccupiedmdashtoo much enzyme prevents cleavage fromoccurring
Tracking and dimerization
McrBC as translocaseBourniquel and Bickle (84) have reviewed much of theenzymology of McrBC which will be briefly summarizedhere The Raleigh Bickle and Pingoud laboratories havecontributed to the following consistent picture of thein vitro reaction EcoKMcrBC cleavage results in adouble-strand cut near one RmC site (72ndash74) butrequires cooperation of two sites (3940) or a translocationblock (73) The sites may be on different daughters acrossa fork (75) These are separated by 30ndash3000 bp (397274)and may be on either strand (3976) cleavage occurs30ndash35 bases from the modified base with oppositenicks not tightly constrained (73) and minor cleavageclusters are found 40 50 and 60 nt from the m5C(74) hm5C DNA elicits cleavage also (39) A ring struc-ture is formed by 5ndash7 molecules of McrB in the presence ofGTP (Figure 3) (103) this complex can bind to a recog-nition element in DNA In the presence of McrC trans-location of the complex occurs and cleavage ensues whentranslocation is blocked Collision of translocatingcomplexes a protein barrier or a topological barrier willelicit double-strand cleavage adjacent to one recognitionelement or the other The enzyme will cleave when recog-nition elements are on opposite sides of a forked structure(75) This would allow action in vivo to prevent entry of aMTase gene even with rare sitesStructurally the McrB protein is proposed to be a
member of the AAA+ protein family of NTPases (104)many of which form ring-shaped complexes and partici-pate in molecular machines lsquoSensorrsquo segments found inthese proteins have been shown in some cases to playroles in coupling NTPase activity to intersubunit commu-nication and movement (105) Two of three elements ofthe GTP-binding motif proposed by Dila et al (61) werevalidated by mutational analysis (6562) The thirdproposed motif element was identified as amino acidsNTAD by Dila et al Alignment of AAA+ NTPases in(104) found this aligned with the motif designatedSensor-1 in (105) An interesting result was that mutationshere unexpectedly appeared to abrogate interaction withMcrC instead of changing which NTP would be
productive (62) It may instead play a role in coordinatingGTP binding and hydrolysis with DNA binding inter-action with McrC and cleavage
Intracellularly the story becomes more complex as themcrB gene encodes two products of 51 and 33 kD McrB-L and McrB-S the latter one starting from an in-frameinternal translation start site (106) Both in vivo andin vitro McrB-S can interfere with the function ofMcrB-L at least in part by forming complexes withMcrC unable to bind DNA (107) Both species can formmultimeric rings in the presence of GTP (103) as is usualfor AAA+NTPases (104)
SauUSI requires two sites and ATP hydrolysisSauUSI was originally annotated as a putative helicasefrom Staphylococcus aureus sp A single polypeptide issufficient for activity both in vitro and also in vivo as aclone in E coli using modified phage as a challenge Theamino acid sequence contains a PLDc domain at theN-terminus This contains a phosphodiesterase motiforiginally identified in Phospholipase D (108) it wasvalidated by mutagenesis of four catalytic residue candi-dates In the middle ATPase and helicase motifs wereproposed to account for ATP dependence of cleavageactivity A Domain of Unknown Function was identifiedat the C-terminus (Pfam DUF3427) (108) and wasproposed to recognize the substrate (45)
The purified enzyme cleaves modified DNA containingm5C and hm5C but not m4C in the presence of ATP ordATP but not other nucleotides The negative result form4C is firm plasmids modified at the same site by an m5CMTase (Dcm) or an m4C MTase (MPspGI) were testedThe former (Cm5CWGG) was sensitive whereas the latterwas resistant Thus the sequence preference is likely tobe satisfied m6A is likely not a substrate but few m6A-containing sites were examined
Like McrBC SauUSI requires the presence of two sitesfor efficient cleavage Presumably the ATPase activityparticipates in monitoring the presence of two sites asfor other nucleotide-dependent REases includingMcrBC The mechanism of communication is unknownThe enzyme belongs to a family of highly similarorthologues found in other sequenced Staphylococci(Tables 1 and 2) and more distant homologues can befound in sequenced bacterial and archaeal genomes
EVOLUTIONARY PRESSURES ON R-M SYSTEMS
Evolution by selfish propagation
One way to understand the massive variety of restrictionsystems and their sporadic distribution is to locate theevolutionary drivers of enzyme diversification in theenzyme genes themselves as selfish elements Work fromthe Kobayashi laboratory has elaborated clear examplesof selfish behavior in some Type II enzyme systems(109110) in which the host becomes lsquoaddictedrsquo to theR-M system Once a cell has acquired an R-M pair lossof the genes results in death of that cellrsquos descendants asthe REase is frequently still present and able to act on thegenome following loss of methylation activity In this
64 Nucleic Acids Research 2014 Vol 42 No 1
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
perspective the role played by modification-dependentenzymes is host defense to exclude systems with lsquoforeignrsquoMTase patterns and prevent the cell from loading up withparasites The exclusion event is accompanied by the deathof the cell (111ndash113) Weak sequence specificity of Type IVenzymes could then result from the need to control entryof a wide variety of invading systems
The selfish aspect certainly plays a role in R-M popula-tion biology but cannot be the whole story Type II R-Msystems can still be lost by inactivation of the R gene firstMoreover Type I systems escape this scenario withcomplex control of cleavage activity the restrictionassembly includes a methylation assembly to begin withtherefore the R protein cannot act unless the MTase ispresent in addition failure of the methylation activity inan intact complex leads to abrogation of R activity some-times by action of the ClpXP protease specifically on theR protein (114ndash116)
Furthermore in population terms a cell that acquiredand became addicted to an R-M system should lose incompetition with a sibling that never received thesystem Two factors could counter this First acquisitioncould be accompanied by an increase in the total numberof copies of the R-M system in the population asproposed for invading transposable elements Thisoverreplication results in more copies of the systemcreated than are lost whether to suicide or to other select-ive disadvantage [see eg (117118)] R-M gene amplifica-tion within a cell has been reported experimentally (119)but spread in a population has not been demonstrated yetA second factor that could counter the disability of addic-tion is localization of competition In a structured envir-onment (colonies on a plate or biofilm on large or smallsurfaces) killing of segregants preserves limiting nutrientsfor lineages that retain the toxinantitoxin pair (120121)Much of the real world is structured so this is an import-ant condition
Evolution by phage-host arms race
A second perspective supposes that the modification-dependent Type IV enzymes arose from the competitivecoevolutionary interaction between phages and theirhosts This was first enunciated by Revel and Luria (2)and most recently by Black and coworkers (122) seealso (123) That is hosts used modification-blockedrestriction to defend against phage infection T-evenphages developed methods of substituting modified basesfor the ordinary ones hosts developed Type IV enzymes indefense phages added sugar or other modifications (19) tothwart Type IV enzymes hosts extended Type IV enzymesto accommodate these decorations finally phages de-veloped protein inhibitors specific for these enzymes aswell T4 phages deliver a protein inhibitor (IPI) alongwith the DNA on infection which allows growth in thepresence of EcoCTGmrSD The locus responsible for thisinhibitor is highly variable among relatives of T4 asgmrSD is in enteric bacteria (both in distribution and inaa sequence) When phage with different IP1 alleles weretested for protection from cloned EcoCTGmrSD and itshomolog EcoUTGmrSD specificity was evident one or
the other or both or neither of the two homologs wascounteracted in individual cases (122) This variability ofthe outcome supports the idea that phage-host interactiondrives at least some of these developmentsIn this perspective the weak sequence selectivity of the
Type IV systems might simply reflect the lack of endogen-ous targets for the enzyme As the host does not present anyhm5C and the phage is completely substituted selection forsequence-specificity is weak Selection would act to spareany co-residentMTases This differs fromType II enzymeswhere the MTase and REase must co-evolve to allow thehost to survive Each Type IV system is compatible withsome suite of Type I-IIIMTases (and thus the R-M systemsas a whole) Methylated or hydroxymethylated bases maynot be recognized at all (EcoCTGmrSD) or the systemmay require one specific base in addition to the modifiedone (McrA McrBC MspJI and PvuRts1I) MTases mod-ifying sites not including that base are then compatible asDcm (CmCWGG) is compatible with the McrA McrBCand Mrr systems in E coli K12Type IV systems that restrict methylated bases in a
weakly specified sequence context confer an additionaladvantage in competition with phages Many phagessuch as have not evolved the nucleotide-substitutionstrategy used by the T-even phages These phagesnormally carry the modification pattern of the mostrecent host if the last host expressed an MTase creatinga susceptible site the Type IV enzyme of the new host willdestroy the invader and limit the infection This may beaccompanied by the death of the individual infectedtherefore protection can be conferred on the sibling popu-lation (111)A further implication of this scenario considers the fate
of a population invaded by phage Phage survival ofrestriction occurs at biologically relevant frequencies(106-102) The survivors of restriction carry the particu-lar methylation pattern of the particular cell and thus areresistant to all restriction systems it might have carried(Types IndashIV) A bacterial population as a whole thenbenefits from mechanisms that diversify the suite ofR-M systems so that such surviving phage do not havefree access to the entire population The extreme variabil-ity of R-M system content in isolates of the same species iscompatible with this idea [see eg (124) REBASEGenomes httptoolsnebcomvinczegenomes] Suchvariability also limits and shapes interstrain genetransfer (115125126)The Raleigh laboratory has built on elegant genetic
work with Type I enzymes in the Murray laboratory(127ndash130) to investigate a locus designated thelsquoImmigration Control locusrsquo or ICR that exemplifiesvariable R-M content Alternative DNA segments con-taining R-M systems are located at the same definedlocation in most E coli chromosomes between the yjiSand yjiA genes The ICR in the non-restricting E coli Cstrain [used in the original definition of the R-M phenom-enon (131)] contains a remnant of a Type I enzyme R geneand is 13 kilobase shorter than the same region in E coliK12 (132) The mechanism of segment replacement is stillunknown The ICR would be an example of the lsquodefenseislandsrsquo analyzed by the Koonin group (133) lsquoDefense
Nucleic Acids Research 2014 Vol 42 No 1 65
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
islandsrsquo contain genes that can defend against phage orother invading DNA these exhibit bioinformaticproperties similar to lsquomobilome islandsrsquo containing mobil-ization genes (transposases for example) However themechanism of mobilization has not been identified forthe ICR
FINAL THOUGHTS
The extreme diversity of R-M systems that recognizeordinary DNA seems likely to be approached by the di-versity of Type IV restriction systems Type IV enzymesare hard to find as most detection methods depend ondevelopment of genetic systems for each taxon or on ser-endipity Those characterized so far mostly stem frominitial genetic investigation of limits on infection trans-formation or transduction Barriers encountered provideleads to the genes involved Bioinformatic analysis hashelped to identify relatives which may be more tractableto biochemical investigation than the example originallyfound This approach has pitfalls the gene encodingMspJI was first thought to code for an enzyme recognizingan unmodified site because it is immediately adjacent toan (inactive it is now thought) cytosine MTase geneProvidentially the first expression host was devoid of sen-sitive sites whereas the first test substrate contained some(51) A combination of biological experiments with bio-informatics and biochemistry will be needed to reveal thefull spectrum of Type IV enzymes that may lurk within thevast universe of unidentified ORFs in bacterial systemsOne might begin with those strains whose genomes carryfew Type II systems Bacillus or Corynebacterium asopposed to Helicobacter or Neisseria [see the Genomessection of REBASE (50)]The role of lsquodefense islandsrsquo and their relation to the
lsquomobilomersquo in bacterial population biology remains to bedetermined If a defense island is similar to a mobilomeisland there should be a mechanism of mobilizationnearby which would boost the contribution oflsquooverreplicationrsquo to the account of selections acting onR-M systems R-M systems of all types can be found onor adjacent to known mobilizing elements (134135) buthave not been shown to move experimentallyOn another note it may turn out that evolutionarily
there is a continuum between the apparently modifica-tion-dependent and modification-blocked paths Onerelative of McrBC predicted by bioinformatics analysesis LlaI a system that recognizes an unmodified sequenceand requires two MTases to support it (136) The enzymeBamHI prefers to cleave DNA with m6A in its GGATCCsite and mutants can be isolated that require this modifiedbase (137) Are there native systems similarly protected bymodification of one position in the recognition site butdependent on modification at a different one An interest-ing evolutionary series can be imagined
ACKNOWLEDGEMENTS
The authors thank Rich Roberts and David Dryden forencouragement and critical comment They are grateful to
Tom Bickle Virginius Siksnys Yu Zheng and XiadongCheng for permission to reproduce figures They alsothank Ivan Correa for help drawing base structures andTasha Jose for illustration assistance They also thank theanonymous reviewers of a previous manuscript for caretaken with the review process greatly improving thepresent document
FUNDING
New England Biolabs (EAR) Funding for open accesscharge New England Biolabs (to EAR)
Conflict of interest statement None declared
REFERENCES
1 LuriaSE and HumanML (1952) A nonhereditary host-inducedvariation of bacterial viruses J Bacteriol 64 557ndash559
2 RevelHR and LuriaSE (1970) DNA-glucosylation in T-evenphage genetic determination and role in phage-host interactionAnnu Rev Genet 4 177ndash192
3 RaleighEA and WilsonG (1986) Escherichia coli K-12 restrictsDNA containing 5-methylcytosine Proc Natl Acad Sci USA83 9070ndash9074
4 Noyer-WeidnerM DiazR and ReinersL (1986) Cytosine-specific DNA modification interferes with plasmid establishmentin Escherichia coli K12 involvement of rglB Mol Gen Genet205 469ndash475
5 BlumenthalRM GregorySA and CooperiderJS (1985)Cloning of a restriction-modification system from Proteus vulgarisand its use in analyzing a methylase-sensitive phenotype inEscherichia coli J Bacteriol 164 501ndash509
6 SohailA LiebM DarM and BhagwatAS (1990) A generequired for very short patch repair in Escherichia coli is adjacentto the DNA cytosine methylase gene J Bacteriol 1724214ndash4221
7 MacNeilDJ (1988) Characterization of a unique methyl-specificrestriction system in Streptomyces avermitilis J Bacteriol 1705607ndash5612
8 VertesAA InuiM KobayashiM KurusuY and YukawaH(1993) Presence of mrr- and mcr-like restriction systems incoryneform bacteria Res Microbiol 144 181ndash185
9 MacalusoA and MettusAM (1991) Efficient transformation ofBacillus thuringiensis requires nonmethylated plasmid DNAJ Bacteriol 173 1353ndash1356
10 HolmesML NuttallSD and Dyall-SmithML (1991)Construction and use of halobacterial shuttle vectors and furtherstudies on Haloferax DNA gyrase J Bacteriol 173 3807ndash3813
11 PosfaiJ BhagwatAS PosfaiG and RobertsRJ (1989)Predictive motifs derived from cytosine methyltransferases NucleicAcids Res 17 2421ndash2435
12 MaloneT BlumenthalRM and ChengX (1995) Structure-guided analysis reveals nine sequence motifs conserved amongDNA amino-methyltransferases and suggests a catalyticmechanism for these enzymes J Mol Biol 253 618ndash632
13 TiminskasA ButkusV and JanulaitisA (1995) Sequence motifcharacteristics for DNA [cytosine-N4] and DNA [adenine-N6]methyltransferases Classification of all DNA methyltransferasesGene 157 3ndash11
14 DickmanMJ InglestonSM SedelnikovaSE RaffertyJBLloydRG GrasbyJA and HornbyDP (2002) The RuvABCresolvasome Eur J Biochem 269 5492ndash5501
15 KovallRA and MatthewsBW (1998) Structural functionaland evolutionary relationships between lambda-exonuclease andthe type II restriction endonucleases Proc Natl Acad Sci USA95 7893ndash7897
16 RobertsRJ ChangYC HuZ RachlinJN AntonBPPokrzywaRM ChoiHP FallerLL GuleriaJ HousmanGet al (2010) COMBREX a project to accelerate the functional
66 Nucleic Acids Research 2014 Vol 42 No 1
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
annotation of prokaryotic genomes Nucleic Acids Res 39D11ndashD14
17 RaePM and SteeleRE (1978) Modified bases in the DNAs ofunicellular eukaryotes an examination of distributions andpossible roles with emphasis on hydroxymethyluracil indinoflagellates Biosystems 10 37ndash53
18 WarrenRAJ (1980) Modified bases in bacteriophage DNAsAnnu Rev Microbiol 34 137ndash158
19 Gommers-AmptJH and BorstP (1995) Hypermodified bases inDNA FASEB J 9 1034ndash1042
20 SwintonD HattmanS CrainPF ChengCS SmithDL andMcCloskeyJA (1983) Purification and characterization of theunusual deoxynucleoside alpha-N-(9-beta-D-20-deoxyribofuranosylpurin-6-yl)glycinamide specified by the phageMu modification function Proc Natl Acad Sci USA 807400ndash7404
21 TahilianiM KohKP ShenY PastorWA BandukwalaHBrudnoY AgarwalS IyerLM LiuDR AravindL et al(2009) Conversion of 5-methylcytosine to 5-hydroxymethylcytosinein mammalian DNA by MLL partner TET1 Science 324930ndash935
22 KriaucionisS and HeintzN (2009) The nuclear DNA base5-hydroxymethylcytosine is present in Purkinje neurons and thebrain Science 324 929ndash930
23 BorstP and SabatiniR (2008) Base J discovery biosynthesisand possible functions Annu RevMicrobiol 62 235ndash251
24 KaminskaKH and BujnickiJM (2008) Bacteriophage MuMom protein responsible for DNA modification is a new memberof the acyltransferase superfamily Cell Cycle 7 120ndash121
25 IyerLM TahilianiM RaoA and AravindL (2009) Predictionof novel families of enzymes involved in oxidative and othercomplex modifications of bases in nucleic acids Cell Cycle 81698ndash1710
26 ZhouX HeX LiangJ LiA XuT KieserT HelmannJDand DengZ (2005) A novel DNA modification by sulphur MolMicrobiol 57 1428ndash1438
27 OuHY HeX ShaoY TaiC RajakumarK and DengZ(2009) dndDB A Database Focused on Phosphorothioation ofthe DNA Backbone PloS One 4 e5132
28 XuT LiangJ ChenS WangL HeX YouD WangZLiA XuZ ZhouX et al (2009) DNA phosphorothioation inStreptomyces lividans mutational analysis of the dnd locusBMC Microbiol 9 41
29 WangL ChenS VerginKL GiovannoniSJ ChanSWDeMottMS TaghizadehK CorderoOX CutlerMTimberlakeS et al (2011) DNA phosphorothioation iswidespread and quantized in bacterial genomes Proc Natl AcadSci USA 108 2963ndash2968
30 Loslashbner-OlesenA SkovgaardO and MarinusMG (2005) Dammethylation coordinating cellular processes Curr OpinMicrobiol 8 154ndash160
31 LowDA and CasadesusJ (2008) Clocks and switches bacterialgene regulation by DNA adenine methylation Curr OpinMicrobiol 11 106ndash112
32 FreitagM and SelkerEU (2005) Controlling DNA methylationMany roads to one modification Curr Opin Genet Dev 15191ndash199
33 ProhaskaSJ StadlerPF and KrakauerDC (2010) Innovationin gene regulation The case of chromatin computation J TheorBiol 265 27ndash44
34 SeshasayeeASN SinghP and KrishnaS (2012) Context-dependent conservation of DNA methyltransferases in bacteriaNucleic Acids Res 40 7066ndash7073
35 AntonBP (2011) Dissertation Boston University BostonMassachusetts USA
36 RaleighEA TrimarchiR and RevelH (1989) Genetic andphysical mapping of the mcrA (rglA) and mcrB (rglB) loci ofEscherichia coli K-12 Genetics 122 279ndash296
37 Gonzalez-CeronG Miranda-OlivaresOJ and Servın-GonzalezL(2009) Characterization of the methyl-specific restriction system ofStreptomyces coelicolor A3(2) and of the role played by laterallyacquired nucleases FEMS Microbiol Lett 301 35ndash43
38 LiuG OuHY WangT LiL TanH ZhouXRajakumarK DengZ and HeX (2010) Cleavage of
phosphorothioated DNA and methylated DNA by the type IVrestriction endonuclease ScoMcrA PLoS Genet 6 e1001253
39 SutherlandE CoeL and RaleighEA (1992) McrBC amultisubunit GTP-dependent restriction endonuclease J MolBiol 225 327ndash348
40 KrugerT WildC and Noyer-WeidnerM (1995) McrB aprokaryotic protein specifically recognizing DNA containingmodified cytosine residues EMBO J 14 2661ndash2669
41 SitaramanR and LepplaSH (2012) Methylation-dependentDNA restriction in Bacillus anthracis Gene 494 44ndash50
42 HeitmanJ and ModelP (1987) Site-specific methylases inducethe SOS DNA repair response in Escherichia coli J Bacteriol169 3243ndash3250
43 Waite-ReesPA KeatingCJ MoranLS SlatkoBEHornstraLJ and BennerJS (1991) Characterisation andexpression of the Escherichia coli Mrr restriction systemJ Bacteriol 173 5207ndash5219
44 KerrAL JeonYJ SvensonCJ RogersPL and NeilanBA(2010) DNA restriction-modification systems in the ethanologenZymomonas mobilis ZM4 Appl MicrobiolBiotech 89 761ndash769
45 XuSY CorvagliaAR ChanSH ZhengY and LinderP(2011) A type IV modification-dependent restriction enzymeSauUSI from Staphylococcus aureus subsp aureus USA300Nucleic Acids Res 39 5597ndash5610
46 MonkIR ShahIM XuM TanMW and FosterTJ (2012)Transforming the untransformable application of directtransformation to manipulate genetically Staphylococcus aureusand Staphylococcus epidermidis mBio 3 e00277ndash00211
47 WangH GuanS QuimbyA Cohen-KarniD PradhanSWilsonG RobertsRJ ZhuZ and ZhengY (2011)Comparative characterization of the PvuRts1I family ofrestriction enzymes and their application in mapping genomic5-hydroxymethylcytosine Nucleic Acids Res 39 9294ndash9305
48 JanosiL YonemitsuH HongH and KajiA (1994) Molecularcloning and expression of a novel hydroxymethylcytosine-specificrestriction enzyme (PvuRts1I) modulated by glucosylation ofDNAJ Mol Biol 242 45ndash61
49 BairCL and BlackLW (2007) A type IV modificationdependent restriction nuclease that targets glucosylatedhydroxymethyl cytosine modified DNAs J Mol Biol 366768ndash778
50 RobertsRJ VinczeT PosfaiJ and MacelisD (2010)REBASEndasha database for DNA restriction and modificationenzymes genes and genomes Nucleic Acids Res 38 D234ndashD236
51 ZhengY Cohen-KarniD XuD ChinHG WilsonGPradhanS and RobertsRJ (2010) A unique family of Mrr-likemodification-dependent restriction endonucleases Nucleic AcidsRes 38 5527ndash5534
52 Cohen-KarniD XuD AponeL FomenkovA SunZDavisPJ KinneySRM Yamada-MabuchiM XuSYDavisT et al (2011) The MspJI family of modification-dependent restriction endonucleases for epigenetic studies ProcNatl Acad Sci USA 108 11040ndash11045
53 LubysA VitkuteJ LubieneJ and JanulaitisA (2011)Restriction enzymes and their applications US Patent ApplicationPublication 20110207139 A1 httpappftusptogovnetahtmlPTOsearch-boolhtml (27 August 2013 date last accessed)
54 TarasovaGV NayakshinaTN and DegtyarevSK (2008)Substrate specificity of new methyl-directed DNA endonucleaseGlaI BMC Mol Biol 9 7
55 MulliganEA and DunnJJ (2008) Cloning purification andinitial characterization of E coli McrA a putative5-methylcytosine-specific nuclease Protein ExprPurif 62 98ndash103
56 MulliganEA HatchwellE McCorkleSR and DunnJJ (2010)Differential binding of Escherichia coli McrA protein to DNAsequences that contain the dinucleotide m5CpG Nucleic AcidsRes 38 1997ndash2005
57 AntonBP and RaleighEA (2004) Transposon-mediated linkerinsertion scanning mutagenesis of the Escherichia coli McrAendonuclease J Bacteriol 186 5699ndash5707
58 BujnickiJM RadlinskaM and RychlewskiL (2000) Atomicmodel of the 5-methylcytosine-specific restriction enzyme McrAreveals an atypical zinc finger and structural similarity tobetabetaalphaMe endonucleases Mol Microbiol 37 1280ndash1281
Nucleic Acids Research 2014 Vol 42 No 1 67
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
59 GastFU BrinkmannT PieperU KrugerT NoyerWeidnerM and PingoudA (1997) The recognition of methylatedDNA by the GTP-dependent restriction endonuclease McrBCresides in the N-terminal domain of McrB Biol Chem 378975ndash982
60 SukackaiteR GrazulisS TamulaitisG and SiksnysV (2012)The recognition domain of the methyl-specific endonucleaseMcrBC flips out 5-methylcytosine Nucleic Acids Res 407552ndash7562
61 DilaD SutherlandE MoranL SlatkoB and RaleighEA(1990) Genetic and sequence organization of the mcrBC locus ofEscherichia coli K-12 J Bacteriol 172 4888ndash4900
62 PieperU BrinkmannT KrugerT Noyer WeidnerM andPingoudA (1997) Characterization of the interaction between therestriction endonuclease McrBC from E coli and its cofactorGTP J Mol Biol 272 190ndash199
63 PieperU SchweitzerT GrollDH GastFU and PingoudA(1999) The GTP-binding domain of McrB more than just avariation on a common theme J Mol Biol 292 547ndash556
64 BujnickiJM and RychlewskiL (2001) Grouping together highlydiverged PD-(DE)XK nucleases and identification of novelsuperfamily members using structure-guided alignment ofsequence profiles J Mol MicrobiolBiotech 3 69ndash72
65 PieperU and PingoudA (2002) A mutational analysis of thePD DEXK motif suggests that McrC harbors the catalyticcenter for DNA cleavage by the GTP-dependent restrictionenzyme McrBC from Escherichia coli Biochemistry 415236ndash5244
66 KelleherJE and RaleighEA (1991) A novel activity inEscherichia coli K-12 that directs restriction of DNA modified atCG dinucleotides J Bacteriol 173 5220ndash5223
67 OrlowskiJ MebrhatuMT MichielsCW BujnickiJM andAertsenA (2008) Mutational analysis and a structural model ofmethyl-directed restriction enzyme Mrr Biochem Biophys ResCommun 377 862ndash866
68 AravindL MakarovaKS and KooninEV (2000) SURVEYAND SUMMARY holliday junction resolvases and relatednucleases identification of new families phyletic distribution andevolutionary trajectories Nucleic Acids Res 28 3417ndash3432
69 BujnickiJM and RychlewskiL (2001) Identification of a PD-(DE)XK-like domain with a novel configuration of the endonucleaseactive site in the methyl-directed restriction enzyme Mrr and itshomologs Gene 267 183ndash191
70 SteczkiewiczK MuszewskaA KnizewskiL RychlewskiL andGinalskiK (2012) Sequence structure and functional diversity ofPD-(DE)XK phosphodiesterase superfamily Nucleic Acids Res40 7016ndash7045
71 BairCL RifatD and BlackLW (2007) Exclusion of glucosyl-hydroxymethylcytosine DNA containing bacteriophages isovercome by the injected protein inhibitor IPI J Mol Biol 366779ndash789
72 StewartFJ and RaleighEA (1998) Dependence of McrBCcleavage on distance between recognition elements Biol Chem379 611ndash616
73 PanneD RaleighEA and BickleTA (1999) The McrBCendonuclease translocates DNA in a reaction dependent on GTPhydrolysis J Mol Biol 290 49ndash60
74 PieperU GrollDH WunschS GastFU SpeckC MuckeNand PingoudA (2002) The GTP-dependent restriction enzymeMcrBC from Escherichia coli forms high-molecular masscomplexes with DNA and produces a cleavage patternwith a characteristic 10-base pair repeat Biochemistry 415245ndash5254
75 IshikawaK HandaN SearsL RaleighEA and KobayashiI(2011) Cleavage of a model DNA replication fork by amethyl-specific endonuclease Nucleic Acids Res 39 5489ndash5498
76 StewartFJ PanneD BickleTA and RaleighEA (2000)Methyl-specific DNA binding by McrBC a modification-dependent restriction enzyme J Mol Biol 298 611ndash622
77 LacksS and GreenbergB (1977) Complementary specificity ofrestriction endonucleases DpnI and DpnII with respect to DNAmethylation J Mol Biol 114 153ndash168
78 SiwekW CzapinskaH BochtlerM BujnickiJM andSkowronekK (2012) Crystal structure and mechanism
of action of the N6-methyladenine-dependent type IIMrestriction endonuclease RDpnI Nucleic Acids Res 407563ndash7572
79 LacksS and GreenbergB (1975) A deoxyribonuclease ofDiplococcus pneumoniae specific for methylated DNA J BiolChem 250 4060ndash4066
80 HortonJR MabuchiMY Cohen-KarniD ZhangXGriggsRM SamaranayakeM RobertsRJ ZhengY andChengX (2012) Structure and cleavage activity of the tetramericMspJI DNA modification-dependent restriction endonucleaseNucleic Acids Res 40 9763ndash9773
81 CymermanIA ObarskaA SkowronekKJ LubysA andBujnickiJM (2006) Identification of a new subfamily of HNHnucleases and experimental characterization of a representativemember HphI restriction endonuclease Proteins 65 867ndash876
82 GorbalenyaAE (1994) Self-splicing group I and group II intronsencode homologous (putative) DNA endonucleases of a newfamily Protein Sci 3 1117ndash1120
83 ChanSH OpitzL HigginsL OrsquoLoaneD and XuSY (2010)Cofactor requirement of HpyAV restriction endonuclease PloSOne 5 e9071
84 BourniquelAA and BickleTA (2002) Complex restrictionenzymes NTP-driven molecular motors Biochimie 84 1047ndash1059
85 OrsquoDriscollJ HeiterDF WilsonGG FitzgeraldGFRobertsR and van SinderenD (2006) A genetic dissection ofthe LlaJI restriction cassette reveals insights on a novelbacteriophage resistance system BMC Microbiol 6 40
86 SmithRM DiffinFM SaveryNJ JosephsenJ andSzczelkunMD (2009) DNA cleavage and methylation specificityof the single polypeptide restriction-modification enzyme LlaGINucleic Acids Res 37 7206ndash7218
87 InterthalH PouliotJJ and ChampouxJJ (2001) The tyrosyl-DNA phosphodiesterase Tdp1 is a member of the phospholipaseD superfamily Proc Natl Acad Sci USA 98 12009ndash12014
88 SasnauskasG HalfordSE and SiksnysV (2003) How the BfiIrestriction enzyme uses one active site to cut two DNA strandsProc Natl Acad Sci USA 100 6410ndash6415
89 SasnauskasG ConnollyBA HalfordSE and SiksnysV (2007)Site-specific DNA transesterification catalyzed by a restrictionenzyme Proc Natl Acad Sci USA 104 2115ndash2120
90 Marchler-BauerA ZhengC ChitsazF DerbyshireMKGeerLY GeerRC GonzalesNR GwadzM HurwitzDILanczyckiCJ et al (2012) CDD conserved domains andprotein three-dimensional structure Nucleic Acids Res 41D348ndashD352
91 KravetzAN ZakharovaMV BeletskayaIV SinevaEVDenjmuchametovMM PetrovSI GlatmanLI andSoloninAS (1993) The cleavage sites and localization of genesencoding the restriction endonucleases Eco1831I and EcoHIGene 129 153ndash154
92 KlimasauskasS KumarS RobertsRJ and ChengX (1994)HhaI methyltransferase flips its target base out of the DNA helixCell 76 357ndash369
93 HashimotoH VertinoPM and ChengX (2010) Molecularcoupling of DNA methylation and histone methylationEpigenomics 2 657ndash669
94 RajakumaraE LawJA SimanshuDK VoigtPJohnsonLM ReinbergD PatelDJ and JacobsenSE (2011)A dual flip-out mechanism for 5mC recognition by theArabidopsis SUVH5 SRA domain and its impact on DNAmethylation and H3K9 dimethylation in vivo Genes Dev 25137ndash152
95 SharifJ and KosekiH (2011) Recruitment of Dnmt1 roles of theSRA protein Np95 (Uhrfi) and other factors Prog Mol BiolTransl Sci 101 289ndash310
96 Tesfazgi MebrhatuM WywialE GhoshA MichielsCWLindnerAB TaddeiF BujnickiJM Van MelderenL andAertsenA (2011) Evidence for an evolutionary antagonismbetween Mrr and Type III modification systems Nucleic AcidsRes 39 5991ndash6001
97 GrantGN JesseeJ BloomFR and HanahanD (1990)Differential plasmid rescue from transgenic mouse DNAs intoEscherichia coli methylation-restriction mutants Proc Natl AcadSci USA 87 4645ndash4649
68 Nucleic Acids Research 2014 Vol 42 No 1
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
98 AertsenA and MichielsCW (2005) Mrr instigates the SOSresponse after high pressure stress in Escherichia coli MolMicrobiol 58 1381ndash1391
99 WeiH TherrienC BlanchardA GuanS and ZhuZ (2008)The Fidelity Index provides a systematic quantitation of staractivity of DNA restriction endonucleases Nucleic Acids Res36 e50
100 LuL PatelH and BisslerJJ (2002) Optimizing DpnIdigestion to detect replicated DNA Biotechniques 33 316ndash318
101 MadejT AddessKJ FongJH GeerLY GeerRCLanczyckiCJ LiuC LuS Marchler-BauerAPanchenkoAR et al (2011) MMDB 3D structures andmacromolecular interactions Nucleic Acids Res 40 D461ndashD464
102 GibratJF MadejT and BryantSH (1996) Surprisingsimilarities in structure comparison Curr Opin Struct Biol 6377ndash385
103 PanneD MullerSA WirtzS EngelA and BickleTA (2001)The McrBC restriction endonuclease assembles into a ringstructure in the presence of G nucleotides EMBO J 203210ndash3217
104 NeuwaldAF AravindL SpougeJL and KooninEV (1999)AAA+ A class of chaperone-like ATPases associated with theassembly operation and disassembly of protein complexesGenome Res 9 27ndash43
105 OguraT and WilkinsonAJ (2001) AAA+ superfamilyATPases common structurendashdiverse function Genes Cells 6575ndash597
106 BearyTP BraymerHD and AchbergerEC (1997) Evidenceof participation of McrB(S) in McrBC restriction in Escherichiacoli K-12 J Bacteriol 179 7768ndash7775
107 PanneD RaleighEA and BickleTA (1998) McrBs amodulator peptide for McrBC activity EMBO J 175477ndash5483
108 PuntaM CoggillPC EberhardtRY MistryJ TateJBoursnellC PangN ForslundK CericG ClementsJ et al(2011) The Pfam protein families database Nucleic Acids Res40 D290ndashD301
109 NaitoT KusanoK and KobayashiI (1995) Selfish behavior ofrestriction-modification systems Science 267 897ndash899
110 KobayashiI (2001) Behavior of restriction-modification systemsas selfish mobile elements and their impact on genome evolutionNucleic Acids Res 29 3742ndash3756
111 FukudaE KaminskaKH BujnickiJM and KobayashiI(2008) Cell death upon epigenetic genome methylation a novelfunction of methyl-specific deoxyribonucleases Genome Biol 9R163
112 IshikawaK FukudaE and KobayashiI (2010) Conflictstargeting epigenetic systems and their resolution by cell deathnovel concepts for methyl-specific and other restriction systemsDNA Res 17 325ndash342
113 FukuyoM SasakiA and KobayashiI (2012) Success of asuicidal defense strategy against infection in a structured habitatSci Rep 2 238
114 MakovetsS TitheradgeAJ and MurrayNE (1998) ClpXand ClpP are essential for the efficient acquisition of genesspecifying type IA and IB restriction systems Mol Microbiol28 25ndash35
115 MurrayNE (2002) Immigration control of DNA in bacteriaself versus non-self Microbiology 148 3ndash20
116 MakovetsS PowellLM TitheradgeAJB BlakelyGW andMurrayNE (2004) Is modification sufficient to protect abacterial chromosome from a resident restriction endonucleaseMol Microbiol 51 135ndash147
117 OrgelLE and CrickFHC (1980) Selfish DNA the ultimateparasite Nature 284 604ndash607
118 KidwellMG and LischDR (2001) Perspective transposableelements parasitic DNA and genome evolution Evolution 551ndash24
119 SadykovM AsamiY NikiH HandaN ItayaMTanokuraM and KobayashiI (2003) Multiplication of arestriction-modification gene complex Mol Microbiol 48417ndash427
120 ChaoL and LevinBR (1981) Structured habitats and theevolution of anticompetitor toxins in bacteria Proc Natl AcadSci USA 78 6324ndash6328
121 MagnusonRD (2007) Hypothetical functions of toxin-antitoxinsystems J Bacteriol 189 6089ndash6092
122 RifatD WrightNT VarneyKM WeberDJ and BlackLW(2008) Restriction endonuclease inhibitor IPI of bacteriophageT4 a novel structure for a dedicated target J Mol Biol 375720ndash734
123 LabrieSJ SamsonJE and MoineauS (2010) Bacteriophageresistance mechanisms Nat Rev Microbiol 8 317ndash327
124 LinLF PosfaiJ RobertsRJ and KongH (2001)Comparative genomics of the restriction-modification systemsin Helicobacter pylori Proc Natl Acad Sci USA 982740ndash2745
125 MilkmanR RaleighEA McKaneM CrydermanDBilodeauP and McWeenyK (1999) Molecular evolution of theEscherichia coli chromosome V Recombination patterns amongstrains of diverse origin Genetics 153 539ndash554
126 HumbertO DorerMS and SalamaNR (2011)Characterization of Helicobacter pylori factors that controltransformation frequency and integration length during inter-strain DNA recombination Mol Microbiol 79 387ndash401
127 BarcusVA TitheradgeAJB and MurrayNE (1995) Thediversity of alleles at the hsd locus in natural populations ofEscherichia coli Genetics 140 1187ndash1197
128 SharpPM KelleherJE DanielAS CowanGM andMurrayNE (1992) Roles of selection and recombination in theevolution of type I restriction-modification systems inenterobacteria Proc Natl Acad Sci USA 89 9836ndash9840
129 LoenenWAM DanielAS BraymerHD and MurrayNE(1987) Organization and sequence of the hsd genes ofEscherichia coli K-12 J Mol Biol 198 159ndash170
130 SuriB and BickleTA (1985) EcoA the first member of a newfamily of type I restriction modification systems Geneorganization and enzymatic activities J Mol Biol 186 77ndash85
131 BertaniG and WeigleJJ (1953) Host-controlled variation inbacterial viruses J Bacteriol 65 113ndash121
132 SibleyMH and RaleighEA (2004) Cassette-like variation ofrestriction enzyme genes in Escherichia coli C and relativesNucleic Acids Res 32 522ndash534
133 MakarovaKS WolfYI SnirS and KooninEV (2011)Defense islands in bacterial and archaeal genomes and predictionof novel defense systems J Bacteriol 193 6039ndash6056
134 FurutaY AbeK and KobayashiI (2010) Genome comparisonand context analysis reveals putative mobile forms of restriction-modification systems and related rearrangements Nucleic AcidsRes 38 2428ndash2443
135 FurutaY and KobayashiI (2012) Restriction-modificationsystems as moblie epigenetic elements In RobertsAP andMullanyP (eds) Bacterial Integrative Mobile Genetic ElementsLandes Bioscience pp 85ndash103
136 OrsquoSullivanDJ and KlaenhammerTR (1998) Control ofexpression of LlaI restriction in Lactococcus lactis MolMicrobiol 27 1009ndash1020
137 WhitakerRD DornerLF and SchildkrautI (1999) A mutantof BamHI restriction endonuclease which requires N6-methyladenine for cleavage J Mol Biol 285 1525ndash1536
Nucleic Acids Research 2014 Vol 42 No 1 69
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
showed that the N-terminal domain was structurallysimilar to that of the eukaryotic SRA domain with acrescent-shaped beta-sheet structure from which loopsproject (see Figure 4B and discussion earlier in the textMcrB) This structural homology allowed modeling of theDNA-bound structure with a flipped m5C The enzyme inthe crystal is a tetramer in which two monomers form aback-to-back dimer via the C-terminal regions thatcomprise the endonuclease Two back-to-back dimersgenerate a tetrameric protein with two cleavage domainspositioned (as in the Type IIP enzyme HindIII used formodeling the C-terminal cleavage domain interaction withDNA) so that a 4-base 50 extension would be created oncleavage of modeled DNA Cleavage is most efficient atmolar ratios that allow all four SRA-like domains to beoccupiedmdashtoo much enzyme prevents cleavage fromoccurring
Tracking and dimerization
McrBC as translocaseBourniquel and Bickle (84) have reviewed much of theenzymology of McrBC which will be briefly summarizedhere The Raleigh Bickle and Pingoud laboratories havecontributed to the following consistent picture of thein vitro reaction EcoKMcrBC cleavage results in adouble-strand cut near one RmC site (72ndash74) butrequires cooperation of two sites (3940) or a translocationblock (73) The sites may be on different daughters acrossa fork (75) These are separated by 30ndash3000 bp (397274)and may be on either strand (3976) cleavage occurs30ndash35 bases from the modified base with oppositenicks not tightly constrained (73) and minor cleavageclusters are found 40 50 and 60 nt from the m5C(74) hm5C DNA elicits cleavage also (39) A ring struc-ture is formed by 5ndash7 molecules of McrB in the presence ofGTP (Figure 3) (103) this complex can bind to a recog-nition element in DNA In the presence of McrC trans-location of the complex occurs and cleavage ensues whentranslocation is blocked Collision of translocatingcomplexes a protein barrier or a topological barrier willelicit double-strand cleavage adjacent to one recognitionelement or the other The enzyme will cleave when recog-nition elements are on opposite sides of a forked structure(75) This would allow action in vivo to prevent entry of aMTase gene even with rare sitesStructurally the McrB protein is proposed to be a
member of the AAA+ protein family of NTPases (104)many of which form ring-shaped complexes and partici-pate in molecular machines lsquoSensorrsquo segments found inthese proteins have been shown in some cases to playroles in coupling NTPase activity to intersubunit commu-nication and movement (105) Two of three elements ofthe GTP-binding motif proposed by Dila et al (61) werevalidated by mutational analysis (6562) The thirdproposed motif element was identified as amino acidsNTAD by Dila et al Alignment of AAA+ NTPases in(104) found this aligned with the motif designatedSensor-1 in (105) An interesting result was that mutationshere unexpectedly appeared to abrogate interaction withMcrC instead of changing which NTP would be
productive (62) It may instead play a role in coordinatingGTP binding and hydrolysis with DNA binding inter-action with McrC and cleavage
Intracellularly the story becomes more complex as themcrB gene encodes two products of 51 and 33 kD McrB-L and McrB-S the latter one starting from an in-frameinternal translation start site (106) Both in vivo andin vitro McrB-S can interfere with the function ofMcrB-L at least in part by forming complexes withMcrC unable to bind DNA (107) Both species can formmultimeric rings in the presence of GTP (103) as is usualfor AAA+NTPases (104)
SauUSI requires two sites and ATP hydrolysisSauUSI was originally annotated as a putative helicasefrom Staphylococcus aureus sp A single polypeptide issufficient for activity both in vitro and also in vivo as aclone in E coli using modified phage as a challenge Theamino acid sequence contains a PLDc domain at theN-terminus This contains a phosphodiesterase motiforiginally identified in Phospholipase D (108) it wasvalidated by mutagenesis of four catalytic residue candi-dates In the middle ATPase and helicase motifs wereproposed to account for ATP dependence of cleavageactivity A Domain of Unknown Function was identifiedat the C-terminus (Pfam DUF3427) (108) and wasproposed to recognize the substrate (45)
The purified enzyme cleaves modified DNA containingm5C and hm5C but not m4C in the presence of ATP ordATP but not other nucleotides The negative result form4C is firm plasmids modified at the same site by an m5CMTase (Dcm) or an m4C MTase (MPspGI) were testedThe former (Cm5CWGG) was sensitive whereas the latterwas resistant Thus the sequence preference is likely tobe satisfied m6A is likely not a substrate but few m6A-containing sites were examined
Like McrBC SauUSI requires the presence of two sitesfor efficient cleavage Presumably the ATPase activityparticipates in monitoring the presence of two sites asfor other nucleotide-dependent REases includingMcrBC The mechanism of communication is unknownThe enzyme belongs to a family of highly similarorthologues found in other sequenced Staphylococci(Tables 1 and 2) and more distant homologues can befound in sequenced bacterial and archaeal genomes
EVOLUTIONARY PRESSURES ON R-M SYSTEMS
Evolution by selfish propagation
One way to understand the massive variety of restrictionsystems and their sporadic distribution is to locate theevolutionary drivers of enzyme diversification in theenzyme genes themselves as selfish elements Work fromthe Kobayashi laboratory has elaborated clear examplesof selfish behavior in some Type II enzyme systems(109110) in which the host becomes lsquoaddictedrsquo to theR-M system Once a cell has acquired an R-M pair lossof the genes results in death of that cellrsquos descendants asthe REase is frequently still present and able to act on thegenome following loss of methylation activity In this
64 Nucleic Acids Research 2014 Vol 42 No 1
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
perspective the role played by modification-dependentenzymes is host defense to exclude systems with lsquoforeignrsquoMTase patterns and prevent the cell from loading up withparasites The exclusion event is accompanied by the deathof the cell (111ndash113) Weak sequence specificity of Type IVenzymes could then result from the need to control entryof a wide variety of invading systems
The selfish aspect certainly plays a role in R-M popula-tion biology but cannot be the whole story Type II R-Msystems can still be lost by inactivation of the R gene firstMoreover Type I systems escape this scenario withcomplex control of cleavage activity the restrictionassembly includes a methylation assembly to begin withtherefore the R protein cannot act unless the MTase ispresent in addition failure of the methylation activity inan intact complex leads to abrogation of R activity some-times by action of the ClpXP protease specifically on theR protein (114ndash116)
Furthermore in population terms a cell that acquiredand became addicted to an R-M system should lose incompetition with a sibling that never received thesystem Two factors could counter this First acquisitioncould be accompanied by an increase in the total numberof copies of the R-M system in the population asproposed for invading transposable elements Thisoverreplication results in more copies of the systemcreated than are lost whether to suicide or to other select-ive disadvantage [see eg (117118)] R-M gene amplifica-tion within a cell has been reported experimentally (119)but spread in a population has not been demonstrated yetA second factor that could counter the disability of addic-tion is localization of competition In a structured envir-onment (colonies on a plate or biofilm on large or smallsurfaces) killing of segregants preserves limiting nutrientsfor lineages that retain the toxinantitoxin pair (120121)Much of the real world is structured so this is an import-ant condition
Evolution by phage-host arms race
A second perspective supposes that the modification-dependent Type IV enzymes arose from the competitivecoevolutionary interaction between phages and theirhosts This was first enunciated by Revel and Luria (2)and most recently by Black and coworkers (122) seealso (123) That is hosts used modification-blockedrestriction to defend against phage infection T-evenphages developed methods of substituting modified basesfor the ordinary ones hosts developed Type IV enzymes indefense phages added sugar or other modifications (19) tothwart Type IV enzymes hosts extended Type IV enzymesto accommodate these decorations finally phages de-veloped protein inhibitors specific for these enzymes aswell T4 phages deliver a protein inhibitor (IPI) alongwith the DNA on infection which allows growth in thepresence of EcoCTGmrSD The locus responsible for thisinhibitor is highly variable among relatives of T4 asgmrSD is in enteric bacteria (both in distribution and inaa sequence) When phage with different IP1 alleles weretested for protection from cloned EcoCTGmrSD and itshomolog EcoUTGmrSD specificity was evident one or
the other or both or neither of the two homologs wascounteracted in individual cases (122) This variability ofthe outcome supports the idea that phage-host interactiondrives at least some of these developmentsIn this perspective the weak sequence selectivity of the
Type IV systems might simply reflect the lack of endogen-ous targets for the enzyme As the host does not present anyhm5C and the phage is completely substituted selection forsequence-specificity is weak Selection would act to spareany co-residentMTases This differs fromType II enzymeswhere the MTase and REase must co-evolve to allow thehost to survive Each Type IV system is compatible withsome suite of Type I-IIIMTases (and thus the R-M systemsas a whole) Methylated or hydroxymethylated bases maynot be recognized at all (EcoCTGmrSD) or the systemmay require one specific base in addition to the modifiedone (McrA McrBC MspJI and PvuRts1I) MTases mod-ifying sites not including that base are then compatible asDcm (CmCWGG) is compatible with the McrA McrBCand Mrr systems in E coli K12Type IV systems that restrict methylated bases in a
weakly specified sequence context confer an additionaladvantage in competition with phages Many phagessuch as have not evolved the nucleotide-substitutionstrategy used by the T-even phages These phagesnormally carry the modification pattern of the mostrecent host if the last host expressed an MTase creatinga susceptible site the Type IV enzyme of the new host willdestroy the invader and limit the infection This may beaccompanied by the death of the individual infectedtherefore protection can be conferred on the sibling popu-lation (111)A further implication of this scenario considers the fate
of a population invaded by phage Phage survival ofrestriction occurs at biologically relevant frequencies(106-102) The survivors of restriction carry the particu-lar methylation pattern of the particular cell and thus areresistant to all restriction systems it might have carried(Types IndashIV) A bacterial population as a whole thenbenefits from mechanisms that diversify the suite ofR-M systems so that such surviving phage do not havefree access to the entire population The extreme variabil-ity of R-M system content in isolates of the same species iscompatible with this idea [see eg (124) REBASEGenomes httptoolsnebcomvinczegenomes] Suchvariability also limits and shapes interstrain genetransfer (115125126)The Raleigh laboratory has built on elegant genetic
work with Type I enzymes in the Murray laboratory(127ndash130) to investigate a locus designated thelsquoImmigration Control locusrsquo or ICR that exemplifiesvariable R-M content Alternative DNA segments con-taining R-M systems are located at the same definedlocation in most E coli chromosomes between the yjiSand yjiA genes The ICR in the non-restricting E coli Cstrain [used in the original definition of the R-M phenom-enon (131)] contains a remnant of a Type I enzyme R geneand is 13 kilobase shorter than the same region in E coliK12 (132) The mechanism of segment replacement is stillunknown The ICR would be an example of the lsquodefenseislandsrsquo analyzed by the Koonin group (133) lsquoDefense
Nucleic Acids Research 2014 Vol 42 No 1 65
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
islandsrsquo contain genes that can defend against phage orother invading DNA these exhibit bioinformaticproperties similar to lsquomobilome islandsrsquo containing mobil-ization genes (transposases for example) However themechanism of mobilization has not been identified forthe ICR
FINAL THOUGHTS
The extreme diversity of R-M systems that recognizeordinary DNA seems likely to be approached by the di-versity of Type IV restriction systems Type IV enzymesare hard to find as most detection methods depend ondevelopment of genetic systems for each taxon or on ser-endipity Those characterized so far mostly stem frominitial genetic investigation of limits on infection trans-formation or transduction Barriers encountered provideleads to the genes involved Bioinformatic analysis hashelped to identify relatives which may be more tractableto biochemical investigation than the example originallyfound This approach has pitfalls the gene encodingMspJI was first thought to code for an enzyme recognizingan unmodified site because it is immediately adjacent toan (inactive it is now thought) cytosine MTase geneProvidentially the first expression host was devoid of sen-sitive sites whereas the first test substrate contained some(51) A combination of biological experiments with bio-informatics and biochemistry will be needed to reveal thefull spectrum of Type IV enzymes that may lurk within thevast universe of unidentified ORFs in bacterial systemsOne might begin with those strains whose genomes carryfew Type II systems Bacillus or Corynebacterium asopposed to Helicobacter or Neisseria [see the Genomessection of REBASE (50)]The role of lsquodefense islandsrsquo and their relation to the
lsquomobilomersquo in bacterial population biology remains to bedetermined If a defense island is similar to a mobilomeisland there should be a mechanism of mobilizationnearby which would boost the contribution oflsquooverreplicationrsquo to the account of selections acting onR-M systems R-M systems of all types can be found onor adjacent to known mobilizing elements (134135) buthave not been shown to move experimentallyOn another note it may turn out that evolutionarily
there is a continuum between the apparently modifica-tion-dependent and modification-blocked paths Onerelative of McrBC predicted by bioinformatics analysesis LlaI a system that recognizes an unmodified sequenceand requires two MTases to support it (136) The enzymeBamHI prefers to cleave DNA with m6A in its GGATCCsite and mutants can be isolated that require this modifiedbase (137) Are there native systems similarly protected bymodification of one position in the recognition site butdependent on modification at a different one An interest-ing evolutionary series can be imagined
ACKNOWLEDGEMENTS
The authors thank Rich Roberts and David Dryden forencouragement and critical comment They are grateful to
Tom Bickle Virginius Siksnys Yu Zheng and XiadongCheng for permission to reproduce figures They alsothank Ivan Correa for help drawing base structures andTasha Jose for illustration assistance They also thank theanonymous reviewers of a previous manuscript for caretaken with the review process greatly improving thepresent document
FUNDING
New England Biolabs (EAR) Funding for open accesscharge New England Biolabs (to EAR)
Conflict of interest statement None declared
REFERENCES
1 LuriaSE and HumanML (1952) A nonhereditary host-inducedvariation of bacterial viruses J Bacteriol 64 557ndash559
2 RevelHR and LuriaSE (1970) DNA-glucosylation in T-evenphage genetic determination and role in phage-host interactionAnnu Rev Genet 4 177ndash192
3 RaleighEA and WilsonG (1986) Escherichia coli K-12 restrictsDNA containing 5-methylcytosine Proc Natl Acad Sci USA83 9070ndash9074
4 Noyer-WeidnerM DiazR and ReinersL (1986) Cytosine-specific DNA modification interferes with plasmid establishmentin Escherichia coli K12 involvement of rglB Mol Gen Genet205 469ndash475
5 BlumenthalRM GregorySA and CooperiderJS (1985)Cloning of a restriction-modification system from Proteus vulgarisand its use in analyzing a methylase-sensitive phenotype inEscherichia coli J Bacteriol 164 501ndash509
6 SohailA LiebM DarM and BhagwatAS (1990) A generequired for very short patch repair in Escherichia coli is adjacentto the DNA cytosine methylase gene J Bacteriol 1724214ndash4221
7 MacNeilDJ (1988) Characterization of a unique methyl-specificrestriction system in Streptomyces avermitilis J Bacteriol 1705607ndash5612
8 VertesAA InuiM KobayashiM KurusuY and YukawaH(1993) Presence of mrr- and mcr-like restriction systems incoryneform bacteria Res Microbiol 144 181ndash185
9 MacalusoA and MettusAM (1991) Efficient transformation ofBacillus thuringiensis requires nonmethylated plasmid DNAJ Bacteriol 173 1353ndash1356
10 HolmesML NuttallSD and Dyall-SmithML (1991)Construction and use of halobacterial shuttle vectors and furtherstudies on Haloferax DNA gyrase J Bacteriol 173 3807ndash3813
11 PosfaiJ BhagwatAS PosfaiG and RobertsRJ (1989)Predictive motifs derived from cytosine methyltransferases NucleicAcids Res 17 2421ndash2435
12 MaloneT BlumenthalRM and ChengX (1995) Structure-guided analysis reveals nine sequence motifs conserved amongDNA amino-methyltransferases and suggests a catalyticmechanism for these enzymes J Mol Biol 253 618ndash632
13 TiminskasA ButkusV and JanulaitisA (1995) Sequence motifcharacteristics for DNA [cytosine-N4] and DNA [adenine-N6]methyltransferases Classification of all DNA methyltransferasesGene 157 3ndash11
14 DickmanMJ InglestonSM SedelnikovaSE RaffertyJBLloydRG GrasbyJA and HornbyDP (2002) The RuvABCresolvasome Eur J Biochem 269 5492ndash5501
15 KovallRA and MatthewsBW (1998) Structural functionaland evolutionary relationships between lambda-exonuclease andthe type II restriction endonucleases Proc Natl Acad Sci USA95 7893ndash7897
16 RobertsRJ ChangYC HuZ RachlinJN AntonBPPokrzywaRM ChoiHP FallerLL GuleriaJ HousmanGet al (2010) COMBREX a project to accelerate the functional
66 Nucleic Acids Research 2014 Vol 42 No 1
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
annotation of prokaryotic genomes Nucleic Acids Res 39D11ndashD14
17 RaePM and SteeleRE (1978) Modified bases in the DNAs ofunicellular eukaryotes an examination of distributions andpossible roles with emphasis on hydroxymethyluracil indinoflagellates Biosystems 10 37ndash53
18 WarrenRAJ (1980) Modified bases in bacteriophage DNAsAnnu Rev Microbiol 34 137ndash158
19 Gommers-AmptJH and BorstP (1995) Hypermodified bases inDNA FASEB J 9 1034ndash1042
20 SwintonD HattmanS CrainPF ChengCS SmithDL andMcCloskeyJA (1983) Purification and characterization of theunusual deoxynucleoside alpha-N-(9-beta-D-20-deoxyribofuranosylpurin-6-yl)glycinamide specified by the phageMu modification function Proc Natl Acad Sci USA 807400ndash7404
21 TahilianiM KohKP ShenY PastorWA BandukwalaHBrudnoY AgarwalS IyerLM LiuDR AravindL et al(2009) Conversion of 5-methylcytosine to 5-hydroxymethylcytosinein mammalian DNA by MLL partner TET1 Science 324930ndash935
22 KriaucionisS and HeintzN (2009) The nuclear DNA base5-hydroxymethylcytosine is present in Purkinje neurons and thebrain Science 324 929ndash930
23 BorstP and SabatiniR (2008) Base J discovery biosynthesisand possible functions Annu RevMicrobiol 62 235ndash251
24 KaminskaKH and BujnickiJM (2008) Bacteriophage MuMom protein responsible for DNA modification is a new memberof the acyltransferase superfamily Cell Cycle 7 120ndash121
25 IyerLM TahilianiM RaoA and AravindL (2009) Predictionof novel families of enzymes involved in oxidative and othercomplex modifications of bases in nucleic acids Cell Cycle 81698ndash1710
26 ZhouX HeX LiangJ LiA XuT KieserT HelmannJDand DengZ (2005) A novel DNA modification by sulphur MolMicrobiol 57 1428ndash1438
27 OuHY HeX ShaoY TaiC RajakumarK and DengZ(2009) dndDB A Database Focused on Phosphorothioation ofthe DNA Backbone PloS One 4 e5132
28 XuT LiangJ ChenS WangL HeX YouD WangZLiA XuZ ZhouX et al (2009) DNA phosphorothioation inStreptomyces lividans mutational analysis of the dnd locusBMC Microbiol 9 41
29 WangL ChenS VerginKL GiovannoniSJ ChanSWDeMottMS TaghizadehK CorderoOX CutlerMTimberlakeS et al (2011) DNA phosphorothioation iswidespread and quantized in bacterial genomes Proc Natl AcadSci USA 108 2963ndash2968
30 Loslashbner-OlesenA SkovgaardO and MarinusMG (2005) Dammethylation coordinating cellular processes Curr OpinMicrobiol 8 154ndash160
31 LowDA and CasadesusJ (2008) Clocks and switches bacterialgene regulation by DNA adenine methylation Curr OpinMicrobiol 11 106ndash112
32 FreitagM and SelkerEU (2005) Controlling DNA methylationMany roads to one modification Curr Opin Genet Dev 15191ndash199
33 ProhaskaSJ StadlerPF and KrakauerDC (2010) Innovationin gene regulation The case of chromatin computation J TheorBiol 265 27ndash44
34 SeshasayeeASN SinghP and KrishnaS (2012) Context-dependent conservation of DNA methyltransferases in bacteriaNucleic Acids Res 40 7066ndash7073
35 AntonBP (2011) Dissertation Boston University BostonMassachusetts USA
36 RaleighEA TrimarchiR and RevelH (1989) Genetic andphysical mapping of the mcrA (rglA) and mcrB (rglB) loci ofEscherichia coli K-12 Genetics 122 279ndash296
37 Gonzalez-CeronG Miranda-OlivaresOJ and Servın-GonzalezL(2009) Characterization of the methyl-specific restriction system ofStreptomyces coelicolor A3(2) and of the role played by laterallyacquired nucleases FEMS Microbiol Lett 301 35ndash43
38 LiuG OuHY WangT LiL TanH ZhouXRajakumarK DengZ and HeX (2010) Cleavage of
phosphorothioated DNA and methylated DNA by the type IVrestriction endonuclease ScoMcrA PLoS Genet 6 e1001253
39 SutherlandE CoeL and RaleighEA (1992) McrBC amultisubunit GTP-dependent restriction endonuclease J MolBiol 225 327ndash348
40 KrugerT WildC and Noyer-WeidnerM (1995) McrB aprokaryotic protein specifically recognizing DNA containingmodified cytosine residues EMBO J 14 2661ndash2669
41 SitaramanR and LepplaSH (2012) Methylation-dependentDNA restriction in Bacillus anthracis Gene 494 44ndash50
42 HeitmanJ and ModelP (1987) Site-specific methylases inducethe SOS DNA repair response in Escherichia coli J Bacteriol169 3243ndash3250
43 Waite-ReesPA KeatingCJ MoranLS SlatkoBEHornstraLJ and BennerJS (1991) Characterisation andexpression of the Escherichia coli Mrr restriction systemJ Bacteriol 173 5207ndash5219
44 KerrAL JeonYJ SvensonCJ RogersPL and NeilanBA(2010) DNA restriction-modification systems in the ethanologenZymomonas mobilis ZM4 Appl MicrobiolBiotech 89 761ndash769
45 XuSY CorvagliaAR ChanSH ZhengY and LinderP(2011) A type IV modification-dependent restriction enzymeSauUSI from Staphylococcus aureus subsp aureus USA300Nucleic Acids Res 39 5597ndash5610
46 MonkIR ShahIM XuM TanMW and FosterTJ (2012)Transforming the untransformable application of directtransformation to manipulate genetically Staphylococcus aureusand Staphylococcus epidermidis mBio 3 e00277ndash00211
47 WangH GuanS QuimbyA Cohen-KarniD PradhanSWilsonG RobertsRJ ZhuZ and ZhengY (2011)Comparative characterization of the PvuRts1I family ofrestriction enzymes and their application in mapping genomic5-hydroxymethylcytosine Nucleic Acids Res 39 9294ndash9305
48 JanosiL YonemitsuH HongH and KajiA (1994) Molecularcloning and expression of a novel hydroxymethylcytosine-specificrestriction enzyme (PvuRts1I) modulated by glucosylation ofDNAJ Mol Biol 242 45ndash61
49 BairCL and BlackLW (2007) A type IV modificationdependent restriction nuclease that targets glucosylatedhydroxymethyl cytosine modified DNAs J Mol Biol 366768ndash778
50 RobertsRJ VinczeT PosfaiJ and MacelisD (2010)REBASEndasha database for DNA restriction and modificationenzymes genes and genomes Nucleic Acids Res 38 D234ndashD236
51 ZhengY Cohen-KarniD XuD ChinHG WilsonGPradhanS and RobertsRJ (2010) A unique family of Mrr-likemodification-dependent restriction endonucleases Nucleic AcidsRes 38 5527ndash5534
52 Cohen-KarniD XuD AponeL FomenkovA SunZDavisPJ KinneySRM Yamada-MabuchiM XuSYDavisT et al (2011) The MspJI family of modification-dependent restriction endonucleases for epigenetic studies ProcNatl Acad Sci USA 108 11040ndash11045
53 LubysA VitkuteJ LubieneJ and JanulaitisA (2011)Restriction enzymes and their applications US Patent ApplicationPublication 20110207139 A1 httpappftusptogovnetahtmlPTOsearch-boolhtml (27 August 2013 date last accessed)
54 TarasovaGV NayakshinaTN and DegtyarevSK (2008)Substrate specificity of new methyl-directed DNA endonucleaseGlaI BMC Mol Biol 9 7
55 MulliganEA and DunnJJ (2008) Cloning purification andinitial characterization of E coli McrA a putative5-methylcytosine-specific nuclease Protein ExprPurif 62 98ndash103
56 MulliganEA HatchwellE McCorkleSR and DunnJJ (2010)Differential binding of Escherichia coli McrA protein to DNAsequences that contain the dinucleotide m5CpG Nucleic AcidsRes 38 1997ndash2005
57 AntonBP and RaleighEA (2004) Transposon-mediated linkerinsertion scanning mutagenesis of the Escherichia coli McrAendonuclease J Bacteriol 186 5699ndash5707
58 BujnickiJM RadlinskaM and RychlewskiL (2000) Atomicmodel of the 5-methylcytosine-specific restriction enzyme McrAreveals an atypical zinc finger and structural similarity tobetabetaalphaMe endonucleases Mol Microbiol 37 1280ndash1281
Nucleic Acids Research 2014 Vol 42 No 1 67
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
59 GastFU BrinkmannT PieperU KrugerT NoyerWeidnerM and PingoudA (1997) The recognition of methylatedDNA by the GTP-dependent restriction endonuclease McrBCresides in the N-terminal domain of McrB Biol Chem 378975ndash982
60 SukackaiteR GrazulisS TamulaitisG and SiksnysV (2012)The recognition domain of the methyl-specific endonucleaseMcrBC flips out 5-methylcytosine Nucleic Acids Res 407552ndash7562
61 DilaD SutherlandE MoranL SlatkoB and RaleighEA(1990) Genetic and sequence organization of the mcrBC locus ofEscherichia coli K-12 J Bacteriol 172 4888ndash4900
62 PieperU BrinkmannT KrugerT Noyer WeidnerM andPingoudA (1997) Characterization of the interaction between therestriction endonuclease McrBC from E coli and its cofactorGTP J Mol Biol 272 190ndash199
63 PieperU SchweitzerT GrollDH GastFU and PingoudA(1999) The GTP-binding domain of McrB more than just avariation on a common theme J Mol Biol 292 547ndash556
64 BujnickiJM and RychlewskiL (2001) Grouping together highlydiverged PD-(DE)XK nucleases and identification of novelsuperfamily members using structure-guided alignment ofsequence profiles J Mol MicrobiolBiotech 3 69ndash72
65 PieperU and PingoudA (2002) A mutational analysis of thePD DEXK motif suggests that McrC harbors the catalyticcenter for DNA cleavage by the GTP-dependent restrictionenzyme McrBC from Escherichia coli Biochemistry 415236ndash5244
66 KelleherJE and RaleighEA (1991) A novel activity inEscherichia coli K-12 that directs restriction of DNA modified atCG dinucleotides J Bacteriol 173 5220ndash5223
67 OrlowskiJ MebrhatuMT MichielsCW BujnickiJM andAertsenA (2008) Mutational analysis and a structural model ofmethyl-directed restriction enzyme Mrr Biochem Biophys ResCommun 377 862ndash866
68 AravindL MakarovaKS and KooninEV (2000) SURVEYAND SUMMARY holliday junction resolvases and relatednucleases identification of new families phyletic distribution andevolutionary trajectories Nucleic Acids Res 28 3417ndash3432
69 BujnickiJM and RychlewskiL (2001) Identification of a PD-(DE)XK-like domain with a novel configuration of the endonucleaseactive site in the methyl-directed restriction enzyme Mrr and itshomologs Gene 267 183ndash191
70 SteczkiewiczK MuszewskaA KnizewskiL RychlewskiL andGinalskiK (2012) Sequence structure and functional diversity ofPD-(DE)XK phosphodiesterase superfamily Nucleic Acids Res40 7016ndash7045
71 BairCL RifatD and BlackLW (2007) Exclusion of glucosyl-hydroxymethylcytosine DNA containing bacteriophages isovercome by the injected protein inhibitor IPI J Mol Biol 366779ndash789
72 StewartFJ and RaleighEA (1998) Dependence of McrBCcleavage on distance between recognition elements Biol Chem379 611ndash616
73 PanneD RaleighEA and BickleTA (1999) The McrBCendonuclease translocates DNA in a reaction dependent on GTPhydrolysis J Mol Biol 290 49ndash60
74 PieperU GrollDH WunschS GastFU SpeckC MuckeNand PingoudA (2002) The GTP-dependent restriction enzymeMcrBC from Escherichia coli forms high-molecular masscomplexes with DNA and produces a cleavage patternwith a characteristic 10-base pair repeat Biochemistry 415245ndash5254
75 IshikawaK HandaN SearsL RaleighEA and KobayashiI(2011) Cleavage of a model DNA replication fork by amethyl-specific endonuclease Nucleic Acids Res 39 5489ndash5498
76 StewartFJ PanneD BickleTA and RaleighEA (2000)Methyl-specific DNA binding by McrBC a modification-dependent restriction enzyme J Mol Biol 298 611ndash622
77 LacksS and GreenbergB (1977) Complementary specificity ofrestriction endonucleases DpnI and DpnII with respect to DNAmethylation J Mol Biol 114 153ndash168
78 SiwekW CzapinskaH BochtlerM BujnickiJM andSkowronekK (2012) Crystal structure and mechanism
of action of the N6-methyladenine-dependent type IIMrestriction endonuclease RDpnI Nucleic Acids Res 407563ndash7572
79 LacksS and GreenbergB (1975) A deoxyribonuclease ofDiplococcus pneumoniae specific for methylated DNA J BiolChem 250 4060ndash4066
80 HortonJR MabuchiMY Cohen-KarniD ZhangXGriggsRM SamaranayakeM RobertsRJ ZhengY andChengX (2012) Structure and cleavage activity of the tetramericMspJI DNA modification-dependent restriction endonucleaseNucleic Acids Res 40 9763ndash9773
81 CymermanIA ObarskaA SkowronekKJ LubysA andBujnickiJM (2006) Identification of a new subfamily of HNHnucleases and experimental characterization of a representativemember HphI restriction endonuclease Proteins 65 867ndash876
82 GorbalenyaAE (1994) Self-splicing group I and group II intronsencode homologous (putative) DNA endonucleases of a newfamily Protein Sci 3 1117ndash1120
83 ChanSH OpitzL HigginsL OrsquoLoaneD and XuSY (2010)Cofactor requirement of HpyAV restriction endonuclease PloSOne 5 e9071
84 BourniquelAA and BickleTA (2002) Complex restrictionenzymes NTP-driven molecular motors Biochimie 84 1047ndash1059
85 OrsquoDriscollJ HeiterDF WilsonGG FitzgeraldGFRobertsR and van SinderenD (2006) A genetic dissection ofthe LlaJI restriction cassette reveals insights on a novelbacteriophage resistance system BMC Microbiol 6 40
86 SmithRM DiffinFM SaveryNJ JosephsenJ andSzczelkunMD (2009) DNA cleavage and methylation specificityof the single polypeptide restriction-modification enzyme LlaGINucleic Acids Res 37 7206ndash7218
87 InterthalH PouliotJJ and ChampouxJJ (2001) The tyrosyl-DNA phosphodiesterase Tdp1 is a member of the phospholipaseD superfamily Proc Natl Acad Sci USA 98 12009ndash12014
88 SasnauskasG HalfordSE and SiksnysV (2003) How the BfiIrestriction enzyme uses one active site to cut two DNA strandsProc Natl Acad Sci USA 100 6410ndash6415
89 SasnauskasG ConnollyBA HalfordSE and SiksnysV (2007)Site-specific DNA transesterification catalyzed by a restrictionenzyme Proc Natl Acad Sci USA 104 2115ndash2120
90 Marchler-BauerA ZhengC ChitsazF DerbyshireMKGeerLY GeerRC GonzalesNR GwadzM HurwitzDILanczyckiCJ et al (2012) CDD conserved domains andprotein three-dimensional structure Nucleic Acids Res 41D348ndashD352
91 KravetzAN ZakharovaMV BeletskayaIV SinevaEVDenjmuchametovMM PetrovSI GlatmanLI andSoloninAS (1993) The cleavage sites and localization of genesencoding the restriction endonucleases Eco1831I and EcoHIGene 129 153ndash154
92 KlimasauskasS KumarS RobertsRJ and ChengX (1994)HhaI methyltransferase flips its target base out of the DNA helixCell 76 357ndash369
93 HashimotoH VertinoPM and ChengX (2010) Molecularcoupling of DNA methylation and histone methylationEpigenomics 2 657ndash669
94 RajakumaraE LawJA SimanshuDK VoigtPJohnsonLM ReinbergD PatelDJ and JacobsenSE (2011)A dual flip-out mechanism for 5mC recognition by theArabidopsis SUVH5 SRA domain and its impact on DNAmethylation and H3K9 dimethylation in vivo Genes Dev 25137ndash152
95 SharifJ and KosekiH (2011) Recruitment of Dnmt1 roles of theSRA protein Np95 (Uhrfi) and other factors Prog Mol BiolTransl Sci 101 289ndash310
96 Tesfazgi MebrhatuM WywialE GhoshA MichielsCWLindnerAB TaddeiF BujnickiJM Van MelderenL andAertsenA (2011) Evidence for an evolutionary antagonismbetween Mrr and Type III modification systems Nucleic AcidsRes 39 5991ndash6001
97 GrantGN JesseeJ BloomFR and HanahanD (1990)Differential plasmid rescue from transgenic mouse DNAs intoEscherichia coli methylation-restriction mutants Proc Natl AcadSci USA 87 4645ndash4649
68 Nucleic Acids Research 2014 Vol 42 No 1
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
98 AertsenA and MichielsCW (2005) Mrr instigates the SOSresponse after high pressure stress in Escherichia coli MolMicrobiol 58 1381ndash1391
99 WeiH TherrienC BlanchardA GuanS and ZhuZ (2008)The Fidelity Index provides a systematic quantitation of staractivity of DNA restriction endonucleases Nucleic Acids Res36 e50
100 LuL PatelH and BisslerJJ (2002) Optimizing DpnIdigestion to detect replicated DNA Biotechniques 33 316ndash318
101 MadejT AddessKJ FongJH GeerLY GeerRCLanczyckiCJ LiuC LuS Marchler-BauerAPanchenkoAR et al (2011) MMDB 3D structures andmacromolecular interactions Nucleic Acids Res 40 D461ndashD464
102 GibratJF MadejT and BryantSH (1996) Surprisingsimilarities in structure comparison Curr Opin Struct Biol 6377ndash385
103 PanneD MullerSA WirtzS EngelA and BickleTA (2001)The McrBC restriction endonuclease assembles into a ringstructure in the presence of G nucleotides EMBO J 203210ndash3217
104 NeuwaldAF AravindL SpougeJL and KooninEV (1999)AAA+ A class of chaperone-like ATPases associated with theassembly operation and disassembly of protein complexesGenome Res 9 27ndash43
105 OguraT and WilkinsonAJ (2001) AAA+ superfamilyATPases common structurendashdiverse function Genes Cells 6575ndash597
106 BearyTP BraymerHD and AchbergerEC (1997) Evidenceof participation of McrB(S) in McrBC restriction in Escherichiacoli K-12 J Bacteriol 179 7768ndash7775
107 PanneD RaleighEA and BickleTA (1998) McrBs amodulator peptide for McrBC activity EMBO J 175477ndash5483
108 PuntaM CoggillPC EberhardtRY MistryJ TateJBoursnellC PangN ForslundK CericG ClementsJ et al(2011) The Pfam protein families database Nucleic Acids Res40 D290ndashD301
109 NaitoT KusanoK and KobayashiI (1995) Selfish behavior ofrestriction-modification systems Science 267 897ndash899
110 KobayashiI (2001) Behavior of restriction-modification systemsas selfish mobile elements and their impact on genome evolutionNucleic Acids Res 29 3742ndash3756
111 FukudaE KaminskaKH BujnickiJM and KobayashiI(2008) Cell death upon epigenetic genome methylation a novelfunction of methyl-specific deoxyribonucleases Genome Biol 9R163
112 IshikawaK FukudaE and KobayashiI (2010) Conflictstargeting epigenetic systems and their resolution by cell deathnovel concepts for methyl-specific and other restriction systemsDNA Res 17 325ndash342
113 FukuyoM SasakiA and KobayashiI (2012) Success of asuicidal defense strategy against infection in a structured habitatSci Rep 2 238
114 MakovetsS TitheradgeAJ and MurrayNE (1998) ClpXand ClpP are essential for the efficient acquisition of genesspecifying type IA and IB restriction systems Mol Microbiol28 25ndash35
115 MurrayNE (2002) Immigration control of DNA in bacteriaself versus non-self Microbiology 148 3ndash20
116 MakovetsS PowellLM TitheradgeAJB BlakelyGW andMurrayNE (2004) Is modification sufficient to protect abacterial chromosome from a resident restriction endonucleaseMol Microbiol 51 135ndash147
117 OrgelLE and CrickFHC (1980) Selfish DNA the ultimateparasite Nature 284 604ndash607
118 KidwellMG and LischDR (2001) Perspective transposableelements parasitic DNA and genome evolution Evolution 551ndash24
119 SadykovM AsamiY NikiH HandaN ItayaMTanokuraM and KobayashiI (2003) Multiplication of arestriction-modification gene complex Mol Microbiol 48417ndash427
120 ChaoL and LevinBR (1981) Structured habitats and theevolution of anticompetitor toxins in bacteria Proc Natl AcadSci USA 78 6324ndash6328
121 MagnusonRD (2007) Hypothetical functions of toxin-antitoxinsystems J Bacteriol 189 6089ndash6092
122 RifatD WrightNT VarneyKM WeberDJ and BlackLW(2008) Restriction endonuclease inhibitor IPI of bacteriophageT4 a novel structure for a dedicated target J Mol Biol 375720ndash734
123 LabrieSJ SamsonJE and MoineauS (2010) Bacteriophageresistance mechanisms Nat Rev Microbiol 8 317ndash327
124 LinLF PosfaiJ RobertsRJ and KongH (2001)Comparative genomics of the restriction-modification systemsin Helicobacter pylori Proc Natl Acad Sci USA 982740ndash2745
125 MilkmanR RaleighEA McKaneM CrydermanDBilodeauP and McWeenyK (1999) Molecular evolution of theEscherichia coli chromosome V Recombination patterns amongstrains of diverse origin Genetics 153 539ndash554
126 HumbertO DorerMS and SalamaNR (2011)Characterization of Helicobacter pylori factors that controltransformation frequency and integration length during inter-strain DNA recombination Mol Microbiol 79 387ndash401
127 BarcusVA TitheradgeAJB and MurrayNE (1995) Thediversity of alleles at the hsd locus in natural populations ofEscherichia coli Genetics 140 1187ndash1197
128 SharpPM KelleherJE DanielAS CowanGM andMurrayNE (1992) Roles of selection and recombination in theevolution of type I restriction-modification systems inenterobacteria Proc Natl Acad Sci USA 89 9836ndash9840
129 LoenenWAM DanielAS BraymerHD and MurrayNE(1987) Organization and sequence of the hsd genes ofEscherichia coli K-12 J Mol Biol 198 159ndash170
130 SuriB and BickleTA (1985) EcoA the first member of a newfamily of type I restriction modification systems Geneorganization and enzymatic activities J Mol Biol 186 77ndash85
131 BertaniG and WeigleJJ (1953) Host-controlled variation inbacterial viruses J Bacteriol 65 113ndash121
132 SibleyMH and RaleighEA (2004) Cassette-like variation ofrestriction enzyme genes in Escherichia coli C and relativesNucleic Acids Res 32 522ndash534
133 MakarovaKS WolfYI SnirS and KooninEV (2011)Defense islands in bacterial and archaeal genomes and predictionof novel defense systems J Bacteriol 193 6039ndash6056
134 FurutaY AbeK and KobayashiI (2010) Genome comparisonand context analysis reveals putative mobile forms of restriction-modification systems and related rearrangements Nucleic AcidsRes 38 2428ndash2443
135 FurutaY and KobayashiI (2012) Restriction-modificationsystems as moblie epigenetic elements In RobertsAP andMullanyP (eds) Bacterial Integrative Mobile Genetic ElementsLandes Bioscience pp 85ndash103
136 OrsquoSullivanDJ and KlaenhammerTR (1998) Control ofexpression of LlaI restriction in Lactococcus lactis MolMicrobiol 27 1009ndash1020
137 WhitakerRD DornerLF and SchildkrautI (1999) A mutantof BamHI restriction endonuclease which requires N6-methyladenine for cleavage J Mol Biol 285 1525ndash1536
Nucleic Acids Research 2014 Vol 42 No 1 69
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
perspective the role played by modification-dependentenzymes is host defense to exclude systems with lsquoforeignrsquoMTase patterns and prevent the cell from loading up withparasites The exclusion event is accompanied by the deathof the cell (111ndash113) Weak sequence specificity of Type IVenzymes could then result from the need to control entryof a wide variety of invading systems
The selfish aspect certainly plays a role in R-M popula-tion biology but cannot be the whole story Type II R-Msystems can still be lost by inactivation of the R gene firstMoreover Type I systems escape this scenario withcomplex control of cleavage activity the restrictionassembly includes a methylation assembly to begin withtherefore the R protein cannot act unless the MTase ispresent in addition failure of the methylation activity inan intact complex leads to abrogation of R activity some-times by action of the ClpXP protease specifically on theR protein (114ndash116)
Furthermore in population terms a cell that acquiredand became addicted to an R-M system should lose incompetition with a sibling that never received thesystem Two factors could counter this First acquisitioncould be accompanied by an increase in the total numberof copies of the R-M system in the population asproposed for invading transposable elements Thisoverreplication results in more copies of the systemcreated than are lost whether to suicide or to other select-ive disadvantage [see eg (117118)] R-M gene amplifica-tion within a cell has been reported experimentally (119)but spread in a population has not been demonstrated yetA second factor that could counter the disability of addic-tion is localization of competition In a structured envir-onment (colonies on a plate or biofilm on large or smallsurfaces) killing of segregants preserves limiting nutrientsfor lineages that retain the toxinantitoxin pair (120121)Much of the real world is structured so this is an import-ant condition
Evolution by phage-host arms race
A second perspective supposes that the modification-dependent Type IV enzymes arose from the competitivecoevolutionary interaction between phages and theirhosts This was first enunciated by Revel and Luria (2)and most recently by Black and coworkers (122) seealso (123) That is hosts used modification-blockedrestriction to defend against phage infection T-evenphages developed methods of substituting modified basesfor the ordinary ones hosts developed Type IV enzymes indefense phages added sugar or other modifications (19) tothwart Type IV enzymes hosts extended Type IV enzymesto accommodate these decorations finally phages de-veloped protein inhibitors specific for these enzymes aswell T4 phages deliver a protein inhibitor (IPI) alongwith the DNA on infection which allows growth in thepresence of EcoCTGmrSD The locus responsible for thisinhibitor is highly variable among relatives of T4 asgmrSD is in enteric bacteria (both in distribution and inaa sequence) When phage with different IP1 alleles weretested for protection from cloned EcoCTGmrSD and itshomolog EcoUTGmrSD specificity was evident one or
the other or both or neither of the two homologs wascounteracted in individual cases (122) This variability ofthe outcome supports the idea that phage-host interactiondrives at least some of these developmentsIn this perspective the weak sequence selectivity of the
Type IV systems might simply reflect the lack of endogen-ous targets for the enzyme As the host does not present anyhm5C and the phage is completely substituted selection forsequence-specificity is weak Selection would act to spareany co-residentMTases This differs fromType II enzymeswhere the MTase and REase must co-evolve to allow thehost to survive Each Type IV system is compatible withsome suite of Type I-IIIMTases (and thus the R-M systemsas a whole) Methylated or hydroxymethylated bases maynot be recognized at all (EcoCTGmrSD) or the systemmay require one specific base in addition to the modifiedone (McrA McrBC MspJI and PvuRts1I) MTases mod-ifying sites not including that base are then compatible asDcm (CmCWGG) is compatible with the McrA McrBCand Mrr systems in E coli K12Type IV systems that restrict methylated bases in a
weakly specified sequence context confer an additionaladvantage in competition with phages Many phagessuch as have not evolved the nucleotide-substitutionstrategy used by the T-even phages These phagesnormally carry the modification pattern of the mostrecent host if the last host expressed an MTase creatinga susceptible site the Type IV enzyme of the new host willdestroy the invader and limit the infection This may beaccompanied by the death of the individual infectedtherefore protection can be conferred on the sibling popu-lation (111)A further implication of this scenario considers the fate
of a population invaded by phage Phage survival ofrestriction occurs at biologically relevant frequencies(106-102) The survivors of restriction carry the particu-lar methylation pattern of the particular cell and thus areresistant to all restriction systems it might have carried(Types IndashIV) A bacterial population as a whole thenbenefits from mechanisms that diversify the suite ofR-M systems so that such surviving phage do not havefree access to the entire population The extreme variabil-ity of R-M system content in isolates of the same species iscompatible with this idea [see eg (124) REBASEGenomes httptoolsnebcomvinczegenomes] Suchvariability also limits and shapes interstrain genetransfer (115125126)The Raleigh laboratory has built on elegant genetic
work with Type I enzymes in the Murray laboratory(127ndash130) to investigate a locus designated thelsquoImmigration Control locusrsquo or ICR that exemplifiesvariable R-M content Alternative DNA segments con-taining R-M systems are located at the same definedlocation in most E coli chromosomes between the yjiSand yjiA genes The ICR in the non-restricting E coli Cstrain [used in the original definition of the R-M phenom-enon (131)] contains a remnant of a Type I enzyme R geneand is 13 kilobase shorter than the same region in E coliK12 (132) The mechanism of segment replacement is stillunknown The ICR would be an example of the lsquodefenseislandsrsquo analyzed by the Koonin group (133) lsquoDefense
Nucleic Acids Research 2014 Vol 42 No 1 65
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
islandsrsquo contain genes that can defend against phage orother invading DNA these exhibit bioinformaticproperties similar to lsquomobilome islandsrsquo containing mobil-ization genes (transposases for example) However themechanism of mobilization has not been identified forthe ICR
FINAL THOUGHTS
The extreme diversity of R-M systems that recognizeordinary DNA seems likely to be approached by the di-versity of Type IV restriction systems Type IV enzymesare hard to find as most detection methods depend ondevelopment of genetic systems for each taxon or on ser-endipity Those characterized so far mostly stem frominitial genetic investigation of limits on infection trans-formation or transduction Barriers encountered provideleads to the genes involved Bioinformatic analysis hashelped to identify relatives which may be more tractableto biochemical investigation than the example originallyfound This approach has pitfalls the gene encodingMspJI was first thought to code for an enzyme recognizingan unmodified site because it is immediately adjacent toan (inactive it is now thought) cytosine MTase geneProvidentially the first expression host was devoid of sen-sitive sites whereas the first test substrate contained some(51) A combination of biological experiments with bio-informatics and biochemistry will be needed to reveal thefull spectrum of Type IV enzymes that may lurk within thevast universe of unidentified ORFs in bacterial systemsOne might begin with those strains whose genomes carryfew Type II systems Bacillus or Corynebacterium asopposed to Helicobacter or Neisseria [see the Genomessection of REBASE (50)]The role of lsquodefense islandsrsquo and their relation to the
lsquomobilomersquo in bacterial population biology remains to bedetermined If a defense island is similar to a mobilomeisland there should be a mechanism of mobilizationnearby which would boost the contribution oflsquooverreplicationrsquo to the account of selections acting onR-M systems R-M systems of all types can be found onor adjacent to known mobilizing elements (134135) buthave not been shown to move experimentallyOn another note it may turn out that evolutionarily
there is a continuum between the apparently modifica-tion-dependent and modification-blocked paths Onerelative of McrBC predicted by bioinformatics analysesis LlaI a system that recognizes an unmodified sequenceand requires two MTases to support it (136) The enzymeBamHI prefers to cleave DNA with m6A in its GGATCCsite and mutants can be isolated that require this modifiedbase (137) Are there native systems similarly protected bymodification of one position in the recognition site butdependent on modification at a different one An interest-ing evolutionary series can be imagined
ACKNOWLEDGEMENTS
The authors thank Rich Roberts and David Dryden forencouragement and critical comment They are grateful to
Tom Bickle Virginius Siksnys Yu Zheng and XiadongCheng for permission to reproduce figures They alsothank Ivan Correa for help drawing base structures andTasha Jose for illustration assistance They also thank theanonymous reviewers of a previous manuscript for caretaken with the review process greatly improving thepresent document
FUNDING
New England Biolabs (EAR) Funding for open accesscharge New England Biolabs (to EAR)
Conflict of interest statement None declared
REFERENCES
1 LuriaSE and HumanML (1952) A nonhereditary host-inducedvariation of bacterial viruses J Bacteriol 64 557ndash559
2 RevelHR and LuriaSE (1970) DNA-glucosylation in T-evenphage genetic determination and role in phage-host interactionAnnu Rev Genet 4 177ndash192
3 RaleighEA and WilsonG (1986) Escherichia coli K-12 restrictsDNA containing 5-methylcytosine Proc Natl Acad Sci USA83 9070ndash9074
4 Noyer-WeidnerM DiazR and ReinersL (1986) Cytosine-specific DNA modification interferes with plasmid establishmentin Escherichia coli K12 involvement of rglB Mol Gen Genet205 469ndash475
5 BlumenthalRM GregorySA and CooperiderJS (1985)Cloning of a restriction-modification system from Proteus vulgarisand its use in analyzing a methylase-sensitive phenotype inEscherichia coli J Bacteriol 164 501ndash509
6 SohailA LiebM DarM and BhagwatAS (1990) A generequired for very short patch repair in Escherichia coli is adjacentto the DNA cytosine methylase gene J Bacteriol 1724214ndash4221
7 MacNeilDJ (1988) Characterization of a unique methyl-specificrestriction system in Streptomyces avermitilis J Bacteriol 1705607ndash5612
8 VertesAA InuiM KobayashiM KurusuY and YukawaH(1993) Presence of mrr- and mcr-like restriction systems incoryneform bacteria Res Microbiol 144 181ndash185
9 MacalusoA and MettusAM (1991) Efficient transformation ofBacillus thuringiensis requires nonmethylated plasmid DNAJ Bacteriol 173 1353ndash1356
10 HolmesML NuttallSD and Dyall-SmithML (1991)Construction and use of halobacterial shuttle vectors and furtherstudies on Haloferax DNA gyrase J Bacteriol 173 3807ndash3813
11 PosfaiJ BhagwatAS PosfaiG and RobertsRJ (1989)Predictive motifs derived from cytosine methyltransferases NucleicAcids Res 17 2421ndash2435
12 MaloneT BlumenthalRM and ChengX (1995) Structure-guided analysis reveals nine sequence motifs conserved amongDNA amino-methyltransferases and suggests a catalyticmechanism for these enzymes J Mol Biol 253 618ndash632
13 TiminskasA ButkusV and JanulaitisA (1995) Sequence motifcharacteristics for DNA [cytosine-N4] and DNA [adenine-N6]methyltransferases Classification of all DNA methyltransferasesGene 157 3ndash11
14 DickmanMJ InglestonSM SedelnikovaSE RaffertyJBLloydRG GrasbyJA and HornbyDP (2002) The RuvABCresolvasome Eur J Biochem 269 5492ndash5501
15 KovallRA and MatthewsBW (1998) Structural functionaland evolutionary relationships between lambda-exonuclease andthe type II restriction endonucleases Proc Natl Acad Sci USA95 7893ndash7897
16 RobertsRJ ChangYC HuZ RachlinJN AntonBPPokrzywaRM ChoiHP FallerLL GuleriaJ HousmanGet al (2010) COMBREX a project to accelerate the functional
66 Nucleic Acids Research 2014 Vol 42 No 1
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
annotation of prokaryotic genomes Nucleic Acids Res 39D11ndashD14
17 RaePM and SteeleRE (1978) Modified bases in the DNAs ofunicellular eukaryotes an examination of distributions andpossible roles with emphasis on hydroxymethyluracil indinoflagellates Biosystems 10 37ndash53
18 WarrenRAJ (1980) Modified bases in bacteriophage DNAsAnnu Rev Microbiol 34 137ndash158
19 Gommers-AmptJH and BorstP (1995) Hypermodified bases inDNA FASEB J 9 1034ndash1042
20 SwintonD HattmanS CrainPF ChengCS SmithDL andMcCloskeyJA (1983) Purification and characterization of theunusual deoxynucleoside alpha-N-(9-beta-D-20-deoxyribofuranosylpurin-6-yl)glycinamide specified by the phageMu modification function Proc Natl Acad Sci USA 807400ndash7404
21 TahilianiM KohKP ShenY PastorWA BandukwalaHBrudnoY AgarwalS IyerLM LiuDR AravindL et al(2009) Conversion of 5-methylcytosine to 5-hydroxymethylcytosinein mammalian DNA by MLL partner TET1 Science 324930ndash935
22 KriaucionisS and HeintzN (2009) The nuclear DNA base5-hydroxymethylcytosine is present in Purkinje neurons and thebrain Science 324 929ndash930
23 BorstP and SabatiniR (2008) Base J discovery biosynthesisand possible functions Annu RevMicrobiol 62 235ndash251
24 KaminskaKH and BujnickiJM (2008) Bacteriophage MuMom protein responsible for DNA modification is a new memberof the acyltransferase superfamily Cell Cycle 7 120ndash121
25 IyerLM TahilianiM RaoA and AravindL (2009) Predictionof novel families of enzymes involved in oxidative and othercomplex modifications of bases in nucleic acids Cell Cycle 81698ndash1710
26 ZhouX HeX LiangJ LiA XuT KieserT HelmannJDand DengZ (2005) A novel DNA modification by sulphur MolMicrobiol 57 1428ndash1438
27 OuHY HeX ShaoY TaiC RajakumarK and DengZ(2009) dndDB A Database Focused on Phosphorothioation ofthe DNA Backbone PloS One 4 e5132
28 XuT LiangJ ChenS WangL HeX YouD WangZLiA XuZ ZhouX et al (2009) DNA phosphorothioation inStreptomyces lividans mutational analysis of the dnd locusBMC Microbiol 9 41
29 WangL ChenS VerginKL GiovannoniSJ ChanSWDeMottMS TaghizadehK CorderoOX CutlerMTimberlakeS et al (2011) DNA phosphorothioation iswidespread and quantized in bacterial genomes Proc Natl AcadSci USA 108 2963ndash2968
30 Loslashbner-OlesenA SkovgaardO and MarinusMG (2005) Dammethylation coordinating cellular processes Curr OpinMicrobiol 8 154ndash160
31 LowDA and CasadesusJ (2008) Clocks and switches bacterialgene regulation by DNA adenine methylation Curr OpinMicrobiol 11 106ndash112
32 FreitagM and SelkerEU (2005) Controlling DNA methylationMany roads to one modification Curr Opin Genet Dev 15191ndash199
33 ProhaskaSJ StadlerPF and KrakauerDC (2010) Innovationin gene regulation The case of chromatin computation J TheorBiol 265 27ndash44
34 SeshasayeeASN SinghP and KrishnaS (2012) Context-dependent conservation of DNA methyltransferases in bacteriaNucleic Acids Res 40 7066ndash7073
35 AntonBP (2011) Dissertation Boston University BostonMassachusetts USA
36 RaleighEA TrimarchiR and RevelH (1989) Genetic andphysical mapping of the mcrA (rglA) and mcrB (rglB) loci ofEscherichia coli K-12 Genetics 122 279ndash296
37 Gonzalez-CeronG Miranda-OlivaresOJ and Servın-GonzalezL(2009) Characterization of the methyl-specific restriction system ofStreptomyces coelicolor A3(2) and of the role played by laterallyacquired nucleases FEMS Microbiol Lett 301 35ndash43
38 LiuG OuHY WangT LiL TanH ZhouXRajakumarK DengZ and HeX (2010) Cleavage of
phosphorothioated DNA and methylated DNA by the type IVrestriction endonuclease ScoMcrA PLoS Genet 6 e1001253
39 SutherlandE CoeL and RaleighEA (1992) McrBC amultisubunit GTP-dependent restriction endonuclease J MolBiol 225 327ndash348
40 KrugerT WildC and Noyer-WeidnerM (1995) McrB aprokaryotic protein specifically recognizing DNA containingmodified cytosine residues EMBO J 14 2661ndash2669
41 SitaramanR and LepplaSH (2012) Methylation-dependentDNA restriction in Bacillus anthracis Gene 494 44ndash50
42 HeitmanJ and ModelP (1987) Site-specific methylases inducethe SOS DNA repair response in Escherichia coli J Bacteriol169 3243ndash3250
43 Waite-ReesPA KeatingCJ MoranLS SlatkoBEHornstraLJ and BennerJS (1991) Characterisation andexpression of the Escherichia coli Mrr restriction systemJ Bacteriol 173 5207ndash5219
44 KerrAL JeonYJ SvensonCJ RogersPL and NeilanBA(2010) DNA restriction-modification systems in the ethanologenZymomonas mobilis ZM4 Appl MicrobiolBiotech 89 761ndash769
45 XuSY CorvagliaAR ChanSH ZhengY and LinderP(2011) A type IV modification-dependent restriction enzymeSauUSI from Staphylococcus aureus subsp aureus USA300Nucleic Acids Res 39 5597ndash5610
46 MonkIR ShahIM XuM TanMW and FosterTJ (2012)Transforming the untransformable application of directtransformation to manipulate genetically Staphylococcus aureusand Staphylococcus epidermidis mBio 3 e00277ndash00211
47 WangH GuanS QuimbyA Cohen-KarniD PradhanSWilsonG RobertsRJ ZhuZ and ZhengY (2011)Comparative characterization of the PvuRts1I family ofrestriction enzymes and their application in mapping genomic5-hydroxymethylcytosine Nucleic Acids Res 39 9294ndash9305
48 JanosiL YonemitsuH HongH and KajiA (1994) Molecularcloning and expression of a novel hydroxymethylcytosine-specificrestriction enzyme (PvuRts1I) modulated by glucosylation ofDNAJ Mol Biol 242 45ndash61
49 BairCL and BlackLW (2007) A type IV modificationdependent restriction nuclease that targets glucosylatedhydroxymethyl cytosine modified DNAs J Mol Biol 366768ndash778
50 RobertsRJ VinczeT PosfaiJ and MacelisD (2010)REBASEndasha database for DNA restriction and modificationenzymes genes and genomes Nucleic Acids Res 38 D234ndashD236
51 ZhengY Cohen-KarniD XuD ChinHG WilsonGPradhanS and RobertsRJ (2010) A unique family of Mrr-likemodification-dependent restriction endonucleases Nucleic AcidsRes 38 5527ndash5534
52 Cohen-KarniD XuD AponeL FomenkovA SunZDavisPJ KinneySRM Yamada-MabuchiM XuSYDavisT et al (2011) The MspJI family of modification-dependent restriction endonucleases for epigenetic studies ProcNatl Acad Sci USA 108 11040ndash11045
53 LubysA VitkuteJ LubieneJ and JanulaitisA (2011)Restriction enzymes and their applications US Patent ApplicationPublication 20110207139 A1 httpappftusptogovnetahtmlPTOsearch-boolhtml (27 August 2013 date last accessed)
54 TarasovaGV NayakshinaTN and DegtyarevSK (2008)Substrate specificity of new methyl-directed DNA endonucleaseGlaI BMC Mol Biol 9 7
55 MulliganEA and DunnJJ (2008) Cloning purification andinitial characterization of E coli McrA a putative5-methylcytosine-specific nuclease Protein ExprPurif 62 98ndash103
56 MulliganEA HatchwellE McCorkleSR and DunnJJ (2010)Differential binding of Escherichia coli McrA protein to DNAsequences that contain the dinucleotide m5CpG Nucleic AcidsRes 38 1997ndash2005
57 AntonBP and RaleighEA (2004) Transposon-mediated linkerinsertion scanning mutagenesis of the Escherichia coli McrAendonuclease J Bacteriol 186 5699ndash5707
58 BujnickiJM RadlinskaM and RychlewskiL (2000) Atomicmodel of the 5-methylcytosine-specific restriction enzyme McrAreveals an atypical zinc finger and structural similarity tobetabetaalphaMe endonucleases Mol Microbiol 37 1280ndash1281
Nucleic Acids Research 2014 Vol 42 No 1 67
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
59 GastFU BrinkmannT PieperU KrugerT NoyerWeidnerM and PingoudA (1997) The recognition of methylatedDNA by the GTP-dependent restriction endonuclease McrBCresides in the N-terminal domain of McrB Biol Chem 378975ndash982
60 SukackaiteR GrazulisS TamulaitisG and SiksnysV (2012)The recognition domain of the methyl-specific endonucleaseMcrBC flips out 5-methylcytosine Nucleic Acids Res 407552ndash7562
61 DilaD SutherlandE MoranL SlatkoB and RaleighEA(1990) Genetic and sequence organization of the mcrBC locus ofEscherichia coli K-12 J Bacteriol 172 4888ndash4900
62 PieperU BrinkmannT KrugerT Noyer WeidnerM andPingoudA (1997) Characterization of the interaction between therestriction endonuclease McrBC from E coli and its cofactorGTP J Mol Biol 272 190ndash199
63 PieperU SchweitzerT GrollDH GastFU and PingoudA(1999) The GTP-binding domain of McrB more than just avariation on a common theme J Mol Biol 292 547ndash556
64 BujnickiJM and RychlewskiL (2001) Grouping together highlydiverged PD-(DE)XK nucleases and identification of novelsuperfamily members using structure-guided alignment ofsequence profiles J Mol MicrobiolBiotech 3 69ndash72
65 PieperU and PingoudA (2002) A mutational analysis of thePD DEXK motif suggests that McrC harbors the catalyticcenter for DNA cleavage by the GTP-dependent restrictionenzyme McrBC from Escherichia coli Biochemistry 415236ndash5244
66 KelleherJE and RaleighEA (1991) A novel activity inEscherichia coli K-12 that directs restriction of DNA modified atCG dinucleotides J Bacteriol 173 5220ndash5223
67 OrlowskiJ MebrhatuMT MichielsCW BujnickiJM andAertsenA (2008) Mutational analysis and a structural model ofmethyl-directed restriction enzyme Mrr Biochem Biophys ResCommun 377 862ndash866
68 AravindL MakarovaKS and KooninEV (2000) SURVEYAND SUMMARY holliday junction resolvases and relatednucleases identification of new families phyletic distribution andevolutionary trajectories Nucleic Acids Res 28 3417ndash3432
69 BujnickiJM and RychlewskiL (2001) Identification of a PD-(DE)XK-like domain with a novel configuration of the endonucleaseactive site in the methyl-directed restriction enzyme Mrr and itshomologs Gene 267 183ndash191
70 SteczkiewiczK MuszewskaA KnizewskiL RychlewskiL andGinalskiK (2012) Sequence structure and functional diversity ofPD-(DE)XK phosphodiesterase superfamily Nucleic Acids Res40 7016ndash7045
71 BairCL RifatD and BlackLW (2007) Exclusion of glucosyl-hydroxymethylcytosine DNA containing bacteriophages isovercome by the injected protein inhibitor IPI J Mol Biol 366779ndash789
72 StewartFJ and RaleighEA (1998) Dependence of McrBCcleavage on distance between recognition elements Biol Chem379 611ndash616
73 PanneD RaleighEA and BickleTA (1999) The McrBCendonuclease translocates DNA in a reaction dependent on GTPhydrolysis J Mol Biol 290 49ndash60
74 PieperU GrollDH WunschS GastFU SpeckC MuckeNand PingoudA (2002) The GTP-dependent restriction enzymeMcrBC from Escherichia coli forms high-molecular masscomplexes with DNA and produces a cleavage patternwith a characteristic 10-base pair repeat Biochemistry 415245ndash5254
75 IshikawaK HandaN SearsL RaleighEA and KobayashiI(2011) Cleavage of a model DNA replication fork by amethyl-specific endonuclease Nucleic Acids Res 39 5489ndash5498
76 StewartFJ PanneD BickleTA and RaleighEA (2000)Methyl-specific DNA binding by McrBC a modification-dependent restriction enzyme J Mol Biol 298 611ndash622
77 LacksS and GreenbergB (1977) Complementary specificity ofrestriction endonucleases DpnI and DpnII with respect to DNAmethylation J Mol Biol 114 153ndash168
78 SiwekW CzapinskaH BochtlerM BujnickiJM andSkowronekK (2012) Crystal structure and mechanism
of action of the N6-methyladenine-dependent type IIMrestriction endonuclease RDpnI Nucleic Acids Res 407563ndash7572
79 LacksS and GreenbergB (1975) A deoxyribonuclease ofDiplococcus pneumoniae specific for methylated DNA J BiolChem 250 4060ndash4066
80 HortonJR MabuchiMY Cohen-KarniD ZhangXGriggsRM SamaranayakeM RobertsRJ ZhengY andChengX (2012) Structure and cleavage activity of the tetramericMspJI DNA modification-dependent restriction endonucleaseNucleic Acids Res 40 9763ndash9773
81 CymermanIA ObarskaA SkowronekKJ LubysA andBujnickiJM (2006) Identification of a new subfamily of HNHnucleases and experimental characterization of a representativemember HphI restriction endonuclease Proteins 65 867ndash876
82 GorbalenyaAE (1994) Self-splicing group I and group II intronsencode homologous (putative) DNA endonucleases of a newfamily Protein Sci 3 1117ndash1120
83 ChanSH OpitzL HigginsL OrsquoLoaneD and XuSY (2010)Cofactor requirement of HpyAV restriction endonuclease PloSOne 5 e9071
84 BourniquelAA and BickleTA (2002) Complex restrictionenzymes NTP-driven molecular motors Biochimie 84 1047ndash1059
85 OrsquoDriscollJ HeiterDF WilsonGG FitzgeraldGFRobertsR and van SinderenD (2006) A genetic dissection ofthe LlaJI restriction cassette reveals insights on a novelbacteriophage resistance system BMC Microbiol 6 40
86 SmithRM DiffinFM SaveryNJ JosephsenJ andSzczelkunMD (2009) DNA cleavage and methylation specificityof the single polypeptide restriction-modification enzyme LlaGINucleic Acids Res 37 7206ndash7218
87 InterthalH PouliotJJ and ChampouxJJ (2001) The tyrosyl-DNA phosphodiesterase Tdp1 is a member of the phospholipaseD superfamily Proc Natl Acad Sci USA 98 12009ndash12014
88 SasnauskasG HalfordSE and SiksnysV (2003) How the BfiIrestriction enzyme uses one active site to cut two DNA strandsProc Natl Acad Sci USA 100 6410ndash6415
89 SasnauskasG ConnollyBA HalfordSE and SiksnysV (2007)Site-specific DNA transesterification catalyzed by a restrictionenzyme Proc Natl Acad Sci USA 104 2115ndash2120
90 Marchler-BauerA ZhengC ChitsazF DerbyshireMKGeerLY GeerRC GonzalesNR GwadzM HurwitzDILanczyckiCJ et al (2012) CDD conserved domains andprotein three-dimensional structure Nucleic Acids Res 41D348ndashD352
91 KravetzAN ZakharovaMV BeletskayaIV SinevaEVDenjmuchametovMM PetrovSI GlatmanLI andSoloninAS (1993) The cleavage sites and localization of genesencoding the restriction endonucleases Eco1831I and EcoHIGene 129 153ndash154
92 KlimasauskasS KumarS RobertsRJ and ChengX (1994)HhaI methyltransferase flips its target base out of the DNA helixCell 76 357ndash369
93 HashimotoH VertinoPM and ChengX (2010) Molecularcoupling of DNA methylation and histone methylationEpigenomics 2 657ndash669
94 RajakumaraE LawJA SimanshuDK VoigtPJohnsonLM ReinbergD PatelDJ and JacobsenSE (2011)A dual flip-out mechanism for 5mC recognition by theArabidopsis SUVH5 SRA domain and its impact on DNAmethylation and H3K9 dimethylation in vivo Genes Dev 25137ndash152
95 SharifJ and KosekiH (2011) Recruitment of Dnmt1 roles of theSRA protein Np95 (Uhrfi) and other factors Prog Mol BiolTransl Sci 101 289ndash310
96 Tesfazgi MebrhatuM WywialE GhoshA MichielsCWLindnerAB TaddeiF BujnickiJM Van MelderenL andAertsenA (2011) Evidence for an evolutionary antagonismbetween Mrr and Type III modification systems Nucleic AcidsRes 39 5991ndash6001
97 GrantGN JesseeJ BloomFR and HanahanD (1990)Differential plasmid rescue from transgenic mouse DNAs intoEscherichia coli methylation-restriction mutants Proc Natl AcadSci USA 87 4645ndash4649
68 Nucleic Acids Research 2014 Vol 42 No 1
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
98 AertsenA and MichielsCW (2005) Mrr instigates the SOSresponse after high pressure stress in Escherichia coli MolMicrobiol 58 1381ndash1391
99 WeiH TherrienC BlanchardA GuanS and ZhuZ (2008)The Fidelity Index provides a systematic quantitation of staractivity of DNA restriction endonucleases Nucleic Acids Res36 e50
100 LuL PatelH and BisslerJJ (2002) Optimizing DpnIdigestion to detect replicated DNA Biotechniques 33 316ndash318
101 MadejT AddessKJ FongJH GeerLY GeerRCLanczyckiCJ LiuC LuS Marchler-BauerAPanchenkoAR et al (2011) MMDB 3D structures andmacromolecular interactions Nucleic Acids Res 40 D461ndashD464
102 GibratJF MadejT and BryantSH (1996) Surprisingsimilarities in structure comparison Curr Opin Struct Biol 6377ndash385
103 PanneD MullerSA WirtzS EngelA and BickleTA (2001)The McrBC restriction endonuclease assembles into a ringstructure in the presence of G nucleotides EMBO J 203210ndash3217
104 NeuwaldAF AravindL SpougeJL and KooninEV (1999)AAA+ A class of chaperone-like ATPases associated with theassembly operation and disassembly of protein complexesGenome Res 9 27ndash43
105 OguraT and WilkinsonAJ (2001) AAA+ superfamilyATPases common structurendashdiverse function Genes Cells 6575ndash597
106 BearyTP BraymerHD and AchbergerEC (1997) Evidenceof participation of McrB(S) in McrBC restriction in Escherichiacoli K-12 J Bacteriol 179 7768ndash7775
107 PanneD RaleighEA and BickleTA (1998) McrBs amodulator peptide for McrBC activity EMBO J 175477ndash5483
108 PuntaM CoggillPC EberhardtRY MistryJ TateJBoursnellC PangN ForslundK CericG ClementsJ et al(2011) The Pfam protein families database Nucleic Acids Res40 D290ndashD301
109 NaitoT KusanoK and KobayashiI (1995) Selfish behavior ofrestriction-modification systems Science 267 897ndash899
110 KobayashiI (2001) Behavior of restriction-modification systemsas selfish mobile elements and their impact on genome evolutionNucleic Acids Res 29 3742ndash3756
111 FukudaE KaminskaKH BujnickiJM and KobayashiI(2008) Cell death upon epigenetic genome methylation a novelfunction of methyl-specific deoxyribonucleases Genome Biol 9R163
112 IshikawaK FukudaE and KobayashiI (2010) Conflictstargeting epigenetic systems and their resolution by cell deathnovel concepts for methyl-specific and other restriction systemsDNA Res 17 325ndash342
113 FukuyoM SasakiA and KobayashiI (2012) Success of asuicidal defense strategy against infection in a structured habitatSci Rep 2 238
114 MakovetsS TitheradgeAJ and MurrayNE (1998) ClpXand ClpP are essential for the efficient acquisition of genesspecifying type IA and IB restriction systems Mol Microbiol28 25ndash35
115 MurrayNE (2002) Immigration control of DNA in bacteriaself versus non-self Microbiology 148 3ndash20
116 MakovetsS PowellLM TitheradgeAJB BlakelyGW andMurrayNE (2004) Is modification sufficient to protect abacterial chromosome from a resident restriction endonucleaseMol Microbiol 51 135ndash147
117 OrgelLE and CrickFHC (1980) Selfish DNA the ultimateparasite Nature 284 604ndash607
118 KidwellMG and LischDR (2001) Perspective transposableelements parasitic DNA and genome evolution Evolution 551ndash24
119 SadykovM AsamiY NikiH HandaN ItayaMTanokuraM and KobayashiI (2003) Multiplication of arestriction-modification gene complex Mol Microbiol 48417ndash427
120 ChaoL and LevinBR (1981) Structured habitats and theevolution of anticompetitor toxins in bacteria Proc Natl AcadSci USA 78 6324ndash6328
121 MagnusonRD (2007) Hypothetical functions of toxin-antitoxinsystems J Bacteriol 189 6089ndash6092
122 RifatD WrightNT VarneyKM WeberDJ and BlackLW(2008) Restriction endonuclease inhibitor IPI of bacteriophageT4 a novel structure for a dedicated target J Mol Biol 375720ndash734
123 LabrieSJ SamsonJE and MoineauS (2010) Bacteriophageresistance mechanisms Nat Rev Microbiol 8 317ndash327
124 LinLF PosfaiJ RobertsRJ and KongH (2001)Comparative genomics of the restriction-modification systemsin Helicobacter pylori Proc Natl Acad Sci USA 982740ndash2745
125 MilkmanR RaleighEA McKaneM CrydermanDBilodeauP and McWeenyK (1999) Molecular evolution of theEscherichia coli chromosome V Recombination patterns amongstrains of diverse origin Genetics 153 539ndash554
126 HumbertO DorerMS and SalamaNR (2011)Characterization of Helicobacter pylori factors that controltransformation frequency and integration length during inter-strain DNA recombination Mol Microbiol 79 387ndash401
127 BarcusVA TitheradgeAJB and MurrayNE (1995) Thediversity of alleles at the hsd locus in natural populations ofEscherichia coli Genetics 140 1187ndash1197
128 SharpPM KelleherJE DanielAS CowanGM andMurrayNE (1992) Roles of selection and recombination in theevolution of type I restriction-modification systems inenterobacteria Proc Natl Acad Sci USA 89 9836ndash9840
129 LoenenWAM DanielAS BraymerHD and MurrayNE(1987) Organization and sequence of the hsd genes ofEscherichia coli K-12 J Mol Biol 198 159ndash170
130 SuriB and BickleTA (1985) EcoA the first member of a newfamily of type I restriction modification systems Geneorganization and enzymatic activities J Mol Biol 186 77ndash85
131 BertaniG and WeigleJJ (1953) Host-controlled variation inbacterial viruses J Bacteriol 65 113ndash121
132 SibleyMH and RaleighEA (2004) Cassette-like variation ofrestriction enzyme genes in Escherichia coli C and relativesNucleic Acids Res 32 522ndash534
133 MakarovaKS WolfYI SnirS and KooninEV (2011)Defense islands in bacterial and archaeal genomes and predictionof novel defense systems J Bacteriol 193 6039ndash6056
134 FurutaY AbeK and KobayashiI (2010) Genome comparisonand context analysis reveals putative mobile forms of restriction-modification systems and related rearrangements Nucleic AcidsRes 38 2428ndash2443
135 FurutaY and KobayashiI (2012) Restriction-modificationsystems as moblie epigenetic elements In RobertsAP andMullanyP (eds) Bacterial Integrative Mobile Genetic ElementsLandes Bioscience pp 85ndash103
136 OrsquoSullivanDJ and KlaenhammerTR (1998) Control ofexpression of LlaI restriction in Lactococcus lactis MolMicrobiol 27 1009ndash1020
137 WhitakerRD DornerLF and SchildkrautI (1999) A mutantof BamHI restriction endonuclease which requires N6-methyladenine for cleavage J Mol Biol 285 1525ndash1536
Nucleic Acids Research 2014 Vol 42 No 1 69
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
islandsrsquo contain genes that can defend against phage orother invading DNA these exhibit bioinformaticproperties similar to lsquomobilome islandsrsquo containing mobil-ization genes (transposases for example) However themechanism of mobilization has not been identified forthe ICR
FINAL THOUGHTS
The extreme diversity of R-M systems that recognizeordinary DNA seems likely to be approached by the di-versity of Type IV restriction systems Type IV enzymesare hard to find as most detection methods depend ondevelopment of genetic systems for each taxon or on ser-endipity Those characterized so far mostly stem frominitial genetic investigation of limits on infection trans-formation or transduction Barriers encountered provideleads to the genes involved Bioinformatic analysis hashelped to identify relatives which may be more tractableto biochemical investigation than the example originallyfound This approach has pitfalls the gene encodingMspJI was first thought to code for an enzyme recognizingan unmodified site because it is immediately adjacent toan (inactive it is now thought) cytosine MTase geneProvidentially the first expression host was devoid of sen-sitive sites whereas the first test substrate contained some(51) A combination of biological experiments with bio-informatics and biochemistry will be needed to reveal thefull spectrum of Type IV enzymes that may lurk within thevast universe of unidentified ORFs in bacterial systemsOne might begin with those strains whose genomes carryfew Type II systems Bacillus or Corynebacterium asopposed to Helicobacter or Neisseria [see the Genomessection of REBASE (50)]The role of lsquodefense islandsrsquo and their relation to the
lsquomobilomersquo in bacterial population biology remains to bedetermined If a defense island is similar to a mobilomeisland there should be a mechanism of mobilizationnearby which would boost the contribution oflsquooverreplicationrsquo to the account of selections acting onR-M systems R-M systems of all types can be found onor adjacent to known mobilizing elements (134135) buthave not been shown to move experimentallyOn another note it may turn out that evolutionarily
there is a continuum between the apparently modifica-tion-dependent and modification-blocked paths Onerelative of McrBC predicted by bioinformatics analysesis LlaI a system that recognizes an unmodified sequenceand requires two MTases to support it (136) The enzymeBamHI prefers to cleave DNA with m6A in its GGATCCsite and mutants can be isolated that require this modifiedbase (137) Are there native systems similarly protected bymodification of one position in the recognition site butdependent on modification at a different one An interest-ing evolutionary series can be imagined
ACKNOWLEDGEMENTS
The authors thank Rich Roberts and David Dryden forencouragement and critical comment They are grateful to
Tom Bickle Virginius Siksnys Yu Zheng and XiadongCheng for permission to reproduce figures They alsothank Ivan Correa for help drawing base structures andTasha Jose for illustration assistance They also thank theanonymous reviewers of a previous manuscript for caretaken with the review process greatly improving thepresent document
FUNDING
New England Biolabs (EAR) Funding for open accesscharge New England Biolabs (to EAR)
Conflict of interest statement None declared
REFERENCES
1 LuriaSE and HumanML (1952) A nonhereditary host-inducedvariation of bacterial viruses J Bacteriol 64 557ndash559
2 RevelHR and LuriaSE (1970) DNA-glucosylation in T-evenphage genetic determination and role in phage-host interactionAnnu Rev Genet 4 177ndash192
3 RaleighEA and WilsonG (1986) Escherichia coli K-12 restrictsDNA containing 5-methylcytosine Proc Natl Acad Sci USA83 9070ndash9074
4 Noyer-WeidnerM DiazR and ReinersL (1986) Cytosine-specific DNA modification interferes with plasmid establishmentin Escherichia coli K12 involvement of rglB Mol Gen Genet205 469ndash475
5 BlumenthalRM GregorySA and CooperiderJS (1985)Cloning of a restriction-modification system from Proteus vulgarisand its use in analyzing a methylase-sensitive phenotype inEscherichia coli J Bacteriol 164 501ndash509
6 SohailA LiebM DarM and BhagwatAS (1990) A generequired for very short patch repair in Escherichia coli is adjacentto the DNA cytosine methylase gene J Bacteriol 1724214ndash4221
7 MacNeilDJ (1988) Characterization of a unique methyl-specificrestriction system in Streptomyces avermitilis J Bacteriol 1705607ndash5612
8 VertesAA InuiM KobayashiM KurusuY and YukawaH(1993) Presence of mrr- and mcr-like restriction systems incoryneform bacteria Res Microbiol 144 181ndash185
9 MacalusoA and MettusAM (1991) Efficient transformation ofBacillus thuringiensis requires nonmethylated plasmid DNAJ Bacteriol 173 1353ndash1356
10 HolmesML NuttallSD and Dyall-SmithML (1991)Construction and use of halobacterial shuttle vectors and furtherstudies on Haloferax DNA gyrase J Bacteriol 173 3807ndash3813
11 PosfaiJ BhagwatAS PosfaiG and RobertsRJ (1989)Predictive motifs derived from cytosine methyltransferases NucleicAcids Res 17 2421ndash2435
12 MaloneT BlumenthalRM and ChengX (1995) Structure-guided analysis reveals nine sequence motifs conserved amongDNA amino-methyltransferases and suggests a catalyticmechanism for these enzymes J Mol Biol 253 618ndash632
13 TiminskasA ButkusV and JanulaitisA (1995) Sequence motifcharacteristics for DNA [cytosine-N4] and DNA [adenine-N6]methyltransferases Classification of all DNA methyltransferasesGene 157 3ndash11
14 DickmanMJ InglestonSM SedelnikovaSE RaffertyJBLloydRG GrasbyJA and HornbyDP (2002) The RuvABCresolvasome Eur J Biochem 269 5492ndash5501
15 KovallRA and MatthewsBW (1998) Structural functionaland evolutionary relationships between lambda-exonuclease andthe type II restriction endonucleases Proc Natl Acad Sci USA95 7893ndash7897
16 RobertsRJ ChangYC HuZ RachlinJN AntonBPPokrzywaRM ChoiHP FallerLL GuleriaJ HousmanGet al (2010) COMBREX a project to accelerate the functional
66 Nucleic Acids Research 2014 Vol 42 No 1
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
annotation of prokaryotic genomes Nucleic Acids Res 39D11ndashD14
17 RaePM and SteeleRE (1978) Modified bases in the DNAs ofunicellular eukaryotes an examination of distributions andpossible roles with emphasis on hydroxymethyluracil indinoflagellates Biosystems 10 37ndash53
18 WarrenRAJ (1980) Modified bases in bacteriophage DNAsAnnu Rev Microbiol 34 137ndash158
19 Gommers-AmptJH and BorstP (1995) Hypermodified bases inDNA FASEB J 9 1034ndash1042
20 SwintonD HattmanS CrainPF ChengCS SmithDL andMcCloskeyJA (1983) Purification and characterization of theunusual deoxynucleoside alpha-N-(9-beta-D-20-deoxyribofuranosylpurin-6-yl)glycinamide specified by the phageMu modification function Proc Natl Acad Sci USA 807400ndash7404
21 TahilianiM KohKP ShenY PastorWA BandukwalaHBrudnoY AgarwalS IyerLM LiuDR AravindL et al(2009) Conversion of 5-methylcytosine to 5-hydroxymethylcytosinein mammalian DNA by MLL partner TET1 Science 324930ndash935
22 KriaucionisS and HeintzN (2009) The nuclear DNA base5-hydroxymethylcytosine is present in Purkinje neurons and thebrain Science 324 929ndash930
23 BorstP and SabatiniR (2008) Base J discovery biosynthesisand possible functions Annu RevMicrobiol 62 235ndash251
24 KaminskaKH and BujnickiJM (2008) Bacteriophage MuMom protein responsible for DNA modification is a new memberof the acyltransferase superfamily Cell Cycle 7 120ndash121
25 IyerLM TahilianiM RaoA and AravindL (2009) Predictionof novel families of enzymes involved in oxidative and othercomplex modifications of bases in nucleic acids Cell Cycle 81698ndash1710
26 ZhouX HeX LiangJ LiA XuT KieserT HelmannJDand DengZ (2005) A novel DNA modification by sulphur MolMicrobiol 57 1428ndash1438
27 OuHY HeX ShaoY TaiC RajakumarK and DengZ(2009) dndDB A Database Focused on Phosphorothioation ofthe DNA Backbone PloS One 4 e5132
28 XuT LiangJ ChenS WangL HeX YouD WangZLiA XuZ ZhouX et al (2009) DNA phosphorothioation inStreptomyces lividans mutational analysis of the dnd locusBMC Microbiol 9 41
29 WangL ChenS VerginKL GiovannoniSJ ChanSWDeMottMS TaghizadehK CorderoOX CutlerMTimberlakeS et al (2011) DNA phosphorothioation iswidespread and quantized in bacterial genomes Proc Natl AcadSci USA 108 2963ndash2968
30 Loslashbner-OlesenA SkovgaardO and MarinusMG (2005) Dammethylation coordinating cellular processes Curr OpinMicrobiol 8 154ndash160
31 LowDA and CasadesusJ (2008) Clocks and switches bacterialgene regulation by DNA adenine methylation Curr OpinMicrobiol 11 106ndash112
32 FreitagM and SelkerEU (2005) Controlling DNA methylationMany roads to one modification Curr Opin Genet Dev 15191ndash199
33 ProhaskaSJ StadlerPF and KrakauerDC (2010) Innovationin gene regulation The case of chromatin computation J TheorBiol 265 27ndash44
34 SeshasayeeASN SinghP and KrishnaS (2012) Context-dependent conservation of DNA methyltransferases in bacteriaNucleic Acids Res 40 7066ndash7073
35 AntonBP (2011) Dissertation Boston University BostonMassachusetts USA
36 RaleighEA TrimarchiR and RevelH (1989) Genetic andphysical mapping of the mcrA (rglA) and mcrB (rglB) loci ofEscherichia coli K-12 Genetics 122 279ndash296
37 Gonzalez-CeronG Miranda-OlivaresOJ and Servın-GonzalezL(2009) Characterization of the methyl-specific restriction system ofStreptomyces coelicolor A3(2) and of the role played by laterallyacquired nucleases FEMS Microbiol Lett 301 35ndash43
38 LiuG OuHY WangT LiL TanH ZhouXRajakumarK DengZ and HeX (2010) Cleavage of
phosphorothioated DNA and methylated DNA by the type IVrestriction endonuclease ScoMcrA PLoS Genet 6 e1001253
39 SutherlandE CoeL and RaleighEA (1992) McrBC amultisubunit GTP-dependent restriction endonuclease J MolBiol 225 327ndash348
40 KrugerT WildC and Noyer-WeidnerM (1995) McrB aprokaryotic protein specifically recognizing DNA containingmodified cytosine residues EMBO J 14 2661ndash2669
41 SitaramanR and LepplaSH (2012) Methylation-dependentDNA restriction in Bacillus anthracis Gene 494 44ndash50
42 HeitmanJ and ModelP (1987) Site-specific methylases inducethe SOS DNA repair response in Escherichia coli J Bacteriol169 3243ndash3250
43 Waite-ReesPA KeatingCJ MoranLS SlatkoBEHornstraLJ and BennerJS (1991) Characterisation andexpression of the Escherichia coli Mrr restriction systemJ Bacteriol 173 5207ndash5219
44 KerrAL JeonYJ SvensonCJ RogersPL and NeilanBA(2010) DNA restriction-modification systems in the ethanologenZymomonas mobilis ZM4 Appl MicrobiolBiotech 89 761ndash769
45 XuSY CorvagliaAR ChanSH ZhengY and LinderP(2011) A type IV modification-dependent restriction enzymeSauUSI from Staphylococcus aureus subsp aureus USA300Nucleic Acids Res 39 5597ndash5610
46 MonkIR ShahIM XuM TanMW and FosterTJ (2012)Transforming the untransformable application of directtransformation to manipulate genetically Staphylococcus aureusand Staphylococcus epidermidis mBio 3 e00277ndash00211
47 WangH GuanS QuimbyA Cohen-KarniD PradhanSWilsonG RobertsRJ ZhuZ and ZhengY (2011)Comparative characterization of the PvuRts1I family ofrestriction enzymes and their application in mapping genomic5-hydroxymethylcytosine Nucleic Acids Res 39 9294ndash9305
48 JanosiL YonemitsuH HongH and KajiA (1994) Molecularcloning and expression of a novel hydroxymethylcytosine-specificrestriction enzyme (PvuRts1I) modulated by glucosylation ofDNAJ Mol Biol 242 45ndash61
49 BairCL and BlackLW (2007) A type IV modificationdependent restriction nuclease that targets glucosylatedhydroxymethyl cytosine modified DNAs J Mol Biol 366768ndash778
50 RobertsRJ VinczeT PosfaiJ and MacelisD (2010)REBASEndasha database for DNA restriction and modificationenzymes genes and genomes Nucleic Acids Res 38 D234ndashD236
51 ZhengY Cohen-KarniD XuD ChinHG WilsonGPradhanS and RobertsRJ (2010) A unique family of Mrr-likemodification-dependent restriction endonucleases Nucleic AcidsRes 38 5527ndash5534
52 Cohen-KarniD XuD AponeL FomenkovA SunZDavisPJ KinneySRM Yamada-MabuchiM XuSYDavisT et al (2011) The MspJI family of modification-dependent restriction endonucleases for epigenetic studies ProcNatl Acad Sci USA 108 11040ndash11045
53 LubysA VitkuteJ LubieneJ and JanulaitisA (2011)Restriction enzymes and their applications US Patent ApplicationPublication 20110207139 A1 httpappftusptogovnetahtmlPTOsearch-boolhtml (27 August 2013 date last accessed)
54 TarasovaGV NayakshinaTN and DegtyarevSK (2008)Substrate specificity of new methyl-directed DNA endonucleaseGlaI BMC Mol Biol 9 7
55 MulliganEA and DunnJJ (2008) Cloning purification andinitial characterization of E coli McrA a putative5-methylcytosine-specific nuclease Protein ExprPurif 62 98ndash103
56 MulliganEA HatchwellE McCorkleSR and DunnJJ (2010)Differential binding of Escherichia coli McrA protein to DNAsequences that contain the dinucleotide m5CpG Nucleic AcidsRes 38 1997ndash2005
57 AntonBP and RaleighEA (2004) Transposon-mediated linkerinsertion scanning mutagenesis of the Escherichia coli McrAendonuclease J Bacteriol 186 5699ndash5707
58 BujnickiJM RadlinskaM and RychlewskiL (2000) Atomicmodel of the 5-methylcytosine-specific restriction enzyme McrAreveals an atypical zinc finger and structural similarity tobetabetaalphaMe endonucleases Mol Microbiol 37 1280ndash1281
Nucleic Acids Research 2014 Vol 42 No 1 67
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
59 GastFU BrinkmannT PieperU KrugerT NoyerWeidnerM and PingoudA (1997) The recognition of methylatedDNA by the GTP-dependent restriction endonuclease McrBCresides in the N-terminal domain of McrB Biol Chem 378975ndash982
60 SukackaiteR GrazulisS TamulaitisG and SiksnysV (2012)The recognition domain of the methyl-specific endonucleaseMcrBC flips out 5-methylcytosine Nucleic Acids Res 407552ndash7562
61 DilaD SutherlandE MoranL SlatkoB and RaleighEA(1990) Genetic and sequence organization of the mcrBC locus ofEscherichia coli K-12 J Bacteriol 172 4888ndash4900
62 PieperU BrinkmannT KrugerT Noyer WeidnerM andPingoudA (1997) Characterization of the interaction between therestriction endonuclease McrBC from E coli and its cofactorGTP J Mol Biol 272 190ndash199
63 PieperU SchweitzerT GrollDH GastFU and PingoudA(1999) The GTP-binding domain of McrB more than just avariation on a common theme J Mol Biol 292 547ndash556
64 BujnickiJM and RychlewskiL (2001) Grouping together highlydiverged PD-(DE)XK nucleases and identification of novelsuperfamily members using structure-guided alignment ofsequence profiles J Mol MicrobiolBiotech 3 69ndash72
65 PieperU and PingoudA (2002) A mutational analysis of thePD DEXK motif suggests that McrC harbors the catalyticcenter for DNA cleavage by the GTP-dependent restrictionenzyme McrBC from Escherichia coli Biochemistry 415236ndash5244
66 KelleherJE and RaleighEA (1991) A novel activity inEscherichia coli K-12 that directs restriction of DNA modified atCG dinucleotides J Bacteriol 173 5220ndash5223
67 OrlowskiJ MebrhatuMT MichielsCW BujnickiJM andAertsenA (2008) Mutational analysis and a structural model ofmethyl-directed restriction enzyme Mrr Biochem Biophys ResCommun 377 862ndash866
68 AravindL MakarovaKS and KooninEV (2000) SURVEYAND SUMMARY holliday junction resolvases and relatednucleases identification of new families phyletic distribution andevolutionary trajectories Nucleic Acids Res 28 3417ndash3432
69 BujnickiJM and RychlewskiL (2001) Identification of a PD-(DE)XK-like domain with a novel configuration of the endonucleaseactive site in the methyl-directed restriction enzyme Mrr and itshomologs Gene 267 183ndash191
70 SteczkiewiczK MuszewskaA KnizewskiL RychlewskiL andGinalskiK (2012) Sequence structure and functional diversity ofPD-(DE)XK phosphodiesterase superfamily Nucleic Acids Res40 7016ndash7045
71 BairCL RifatD and BlackLW (2007) Exclusion of glucosyl-hydroxymethylcytosine DNA containing bacteriophages isovercome by the injected protein inhibitor IPI J Mol Biol 366779ndash789
72 StewartFJ and RaleighEA (1998) Dependence of McrBCcleavage on distance between recognition elements Biol Chem379 611ndash616
73 PanneD RaleighEA and BickleTA (1999) The McrBCendonuclease translocates DNA in a reaction dependent on GTPhydrolysis J Mol Biol 290 49ndash60
74 PieperU GrollDH WunschS GastFU SpeckC MuckeNand PingoudA (2002) The GTP-dependent restriction enzymeMcrBC from Escherichia coli forms high-molecular masscomplexes with DNA and produces a cleavage patternwith a characteristic 10-base pair repeat Biochemistry 415245ndash5254
75 IshikawaK HandaN SearsL RaleighEA and KobayashiI(2011) Cleavage of a model DNA replication fork by amethyl-specific endonuclease Nucleic Acids Res 39 5489ndash5498
76 StewartFJ PanneD BickleTA and RaleighEA (2000)Methyl-specific DNA binding by McrBC a modification-dependent restriction enzyme J Mol Biol 298 611ndash622
77 LacksS and GreenbergB (1977) Complementary specificity ofrestriction endonucleases DpnI and DpnII with respect to DNAmethylation J Mol Biol 114 153ndash168
78 SiwekW CzapinskaH BochtlerM BujnickiJM andSkowronekK (2012) Crystal structure and mechanism
of action of the N6-methyladenine-dependent type IIMrestriction endonuclease RDpnI Nucleic Acids Res 407563ndash7572
79 LacksS and GreenbergB (1975) A deoxyribonuclease ofDiplococcus pneumoniae specific for methylated DNA J BiolChem 250 4060ndash4066
80 HortonJR MabuchiMY Cohen-KarniD ZhangXGriggsRM SamaranayakeM RobertsRJ ZhengY andChengX (2012) Structure and cleavage activity of the tetramericMspJI DNA modification-dependent restriction endonucleaseNucleic Acids Res 40 9763ndash9773
81 CymermanIA ObarskaA SkowronekKJ LubysA andBujnickiJM (2006) Identification of a new subfamily of HNHnucleases and experimental characterization of a representativemember HphI restriction endonuclease Proteins 65 867ndash876
82 GorbalenyaAE (1994) Self-splicing group I and group II intronsencode homologous (putative) DNA endonucleases of a newfamily Protein Sci 3 1117ndash1120
83 ChanSH OpitzL HigginsL OrsquoLoaneD and XuSY (2010)Cofactor requirement of HpyAV restriction endonuclease PloSOne 5 e9071
84 BourniquelAA and BickleTA (2002) Complex restrictionenzymes NTP-driven molecular motors Biochimie 84 1047ndash1059
85 OrsquoDriscollJ HeiterDF WilsonGG FitzgeraldGFRobertsR and van SinderenD (2006) A genetic dissection ofthe LlaJI restriction cassette reveals insights on a novelbacteriophage resistance system BMC Microbiol 6 40
86 SmithRM DiffinFM SaveryNJ JosephsenJ andSzczelkunMD (2009) DNA cleavage and methylation specificityof the single polypeptide restriction-modification enzyme LlaGINucleic Acids Res 37 7206ndash7218
87 InterthalH PouliotJJ and ChampouxJJ (2001) The tyrosyl-DNA phosphodiesterase Tdp1 is a member of the phospholipaseD superfamily Proc Natl Acad Sci USA 98 12009ndash12014
88 SasnauskasG HalfordSE and SiksnysV (2003) How the BfiIrestriction enzyme uses one active site to cut two DNA strandsProc Natl Acad Sci USA 100 6410ndash6415
89 SasnauskasG ConnollyBA HalfordSE and SiksnysV (2007)Site-specific DNA transesterification catalyzed by a restrictionenzyme Proc Natl Acad Sci USA 104 2115ndash2120
90 Marchler-BauerA ZhengC ChitsazF DerbyshireMKGeerLY GeerRC GonzalesNR GwadzM HurwitzDILanczyckiCJ et al (2012) CDD conserved domains andprotein three-dimensional structure Nucleic Acids Res 41D348ndashD352
91 KravetzAN ZakharovaMV BeletskayaIV SinevaEVDenjmuchametovMM PetrovSI GlatmanLI andSoloninAS (1993) The cleavage sites and localization of genesencoding the restriction endonucleases Eco1831I and EcoHIGene 129 153ndash154
92 KlimasauskasS KumarS RobertsRJ and ChengX (1994)HhaI methyltransferase flips its target base out of the DNA helixCell 76 357ndash369
93 HashimotoH VertinoPM and ChengX (2010) Molecularcoupling of DNA methylation and histone methylationEpigenomics 2 657ndash669
94 RajakumaraE LawJA SimanshuDK VoigtPJohnsonLM ReinbergD PatelDJ and JacobsenSE (2011)A dual flip-out mechanism for 5mC recognition by theArabidopsis SUVH5 SRA domain and its impact on DNAmethylation and H3K9 dimethylation in vivo Genes Dev 25137ndash152
95 SharifJ and KosekiH (2011) Recruitment of Dnmt1 roles of theSRA protein Np95 (Uhrfi) and other factors Prog Mol BiolTransl Sci 101 289ndash310
96 Tesfazgi MebrhatuM WywialE GhoshA MichielsCWLindnerAB TaddeiF BujnickiJM Van MelderenL andAertsenA (2011) Evidence for an evolutionary antagonismbetween Mrr and Type III modification systems Nucleic AcidsRes 39 5991ndash6001
97 GrantGN JesseeJ BloomFR and HanahanD (1990)Differential plasmid rescue from transgenic mouse DNAs intoEscherichia coli methylation-restriction mutants Proc Natl AcadSci USA 87 4645ndash4649
68 Nucleic Acids Research 2014 Vol 42 No 1
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
98 AertsenA and MichielsCW (2005) Mrr instigates the SOSresponse after high pressure stress in Escherichia coli MolMicrobiol 58 1381ndash1391
99 WeiH TherrienC BlanchardA GuanS and ZhuZ (2008)The Fidelity Index provides a systematic quantitation of staractivity of DNA restriction endonucleases Nucleic Acids Res36 e50
100 LuL PatelH and BisslerJJ (2002) Optimizing DpnIdigestion to detect replicated DNA Biotechniques 33 316ndash318
101 MadejT AddessKJ FongJH GeerLY GeerRCLanczyckiCJ LiuC LuS Marchler-BauerAPanchenkoAR et al (2011) MMDB 3D structures andmacromolecular interactions Nucleic Acids Res 40 D461ndashD464
102 GibratJF MadejT and BryantSH (1996) Surprisingsimilarities in structure comparison Curr Opin Struct Biol 6377ndash385
103 PanneD MullerSA WirtzS EngelA and BickleTA (2001)The McrBC restriction endonuclease assembles into a ringstructure in the presence of G nucleotides EMBO J 203210ndash3217
104 NeuwaldAF AravindL SpougeJL and KooninEV (1999)AAA+ A class of chaperone-like ATPases associated with theassembly operation and disassembly of protein complexesGenome Res 9 27ndash43
105 OguraT and WilkinsonAJ (2001) AAA+ superfamilyATPases common structurendashdiverse function Genes Cells 6575ndash597
106 BearyTP BraymerHD and AchbergerEC (1997) Evidenceof participation of McrB(S) in McrBC restriction in Escherichiacoli K-12 J Bacteriol 179 7768ndash7775
107 PanneD RaleighEA and BickleTA (1998) McrBs amodulator peptide for McrBC activity EMBO J 175477ndash5483
108 PuntaM CoggillPC EberhardtRY MistryJ TateJBoursnellC PangN ForslundK CericG ClementsJ et al(2011) The Pfam protein families database Nucleic Acids Res40 D290ndashD301
109 NaitoT KusanoK and KobayashiI (1995) Selfish behavior ofrestriction-modification systems Science 267 897ndash899
110 KobayashiI (2001) Behavior of restriction-modification systemsas selfish mobile elements and their impact on genome evolutionNucleic Acids Res 29 3742ndash3756
111 FukudaE KaminskaKH BujnickiJM and KobayashiI(2008) Cell death upon epigenetic genome methylation a novelfunction of methyl-specific deoxyribonucleases Genome Biol 9R163
112 IshikawaK FukudaE and KobayashiI (2010) Conflictstargeting epigenetic systems and their resolution by cell deathnovel concepts for methyl-specific and other restriction systemsDNA Res 17 325ndash342
113 FukuyoM SasakiA and KobayashiI (2012) Success of asuicidal defense strategy against infection in a structured habitatSci Rep 2 238
114 MakovetsS TitheradgeAJ and MurrayNE (1998) ClpXand ClpP are essential for the efficient acquisition of genesspecifying type IA and IB restriction systems Mol Microbiol28 25ndash35
115 MurrayNE (2002) Immigration control of DNA in bacteriaself versus non-self Microbiology 148 3ndash20
116 MakovetsS PowellLM TitheradgeAJB BlakelyGW andMurrayNE (2004) Is modification sufficient to protect abacterial chromosome from a resident restriction endonucleaseMol Microbiol 51 135ndash147
117 OrgelLE and CrickFHC (1980) Selfish DNA the ultimateparasite Nature 284 604ndash607
118 KidwellMG and LischDR (2001) Perspective transposableelements parasitic DNA and genome evolution Evolution 551ndash24
119 SadykovM AsamiY NikiH HandaN ItayaMTanokuraM and KobayashiI (2003) Multiplication of arestriction-modification gene complex Mol Microbiol 48417ndash427
120 ChaoL and LevinBR (1981) Structured habitats and theevolution of anticompetitor toxins in bacteria Proc Natl AcadSci USA 78 6324ndash6328
121 MagnusonRD (2007) Hypothetical functions of toxin-antitoxinsystems J Bacteriol 189 6089ndash6092
122 RifatD WrightNT VarneyKM WeberDJ and BlackLW(2008) Restriction endonuclease inhibitor IPI of bacteriophageT4 a novel structure for a dedicated target J Mol Biol 375720ndash734
123 LabrieSJ SamsonJE and MoineauS (2010) Bacteriophageresistance mechanisms Nat Rev Microbiol 8 317ndash327
124 LinLF PosfaiJ RobertsRJ and KongH (2001)Comparative genomics of the restriction-modification systemsin Helicobacter pylori Proc Natl Acad Sci USA 982740ndash2745
125 MilkmanR RaleighEA McKaneM CrydermanDBilodeauP and McWeenyK (1999) Molecular evolution of theEscherichia coli chromosome V Recombination patterns amongstrains of diverse origin Genetics 153 539ndash554
126 HumbertO DorerMS and SalamaNR (2011)Characterization of Helicobacter pylori factors that controltransformation frequency and integration length during inter-strain DNA recombination Mol Microbiol 79 387ndash401
127 BarcusVA TitheradgeAJB and MurrayNE (1995) Thediversity of alleles at the hsd locus in natural populations ofEscherichia coli Genetics 140 1187ndash1197
128 SharpPM KelleherJE DanielAS CowanGM andMurrayNE (1992) Roles of selection and recombination in theevolution of type I restriction-modification systems inenterobacteria Proc Natl Acad Sci USA 89 9836ndash9840
129 LoenenWAM DanielAS BraymerHD and MurrayNE(1987) Organization and sequence of the hsd genes ofEscherichia coli K-12 J Mol Biol 198 159ndash170
130 SuriB and BickleTA (1985) EcoA the first member of a newfamily of type I restriction modification systems Geneorganization and enzymatic activities J Mol Biol 186 77ndash85
131 BertaniG and WeigleJJ (1953) Host-controlled variation inbacterial viruses J Bacteriol 65 113ndash121
132 SibleyMH and RaleighEA (2004) Cassette-like variation ofrestriction enzyme genes in Escherichia coli C and relativesNucleic Acids Res 32 522ndash534
133 MakarovaKS WolfYI SnirS and KooninEV (2011)Defense islands in bacterial and archaeal genomes and predictionof novel defense systems J Bacteriol 193 6039ndash6056
134 FurutaY AbeK and KobayashiI (2010) Genome comparisonand context analysis reveals putative mobile forms of restriction-modification systems and related rearrangements Nucleic AcidsRes 38 2428ndash2443
135 FurutaY and KobayashiI (2012) Restriction-modificationsystems as moblie epigenetic elements In RobertsAP andMullanyP (eds) Bacterial Integrative Mobile Genetic ElementsLandes Bioscience pp 85ndash103
136 OrsquoSullivanDJ and KlaenhammerTR (1998) Control ofexpression of LlaI restriction in Lactococcus lactis MolMicrobiol 27 1009ndash1020
137 WhitakerRD DornerLF and SchildkrautI (1999) A mutantof BamHI restriction endonuclease which requires N6-methyladenine for cleavage J Mol Biol 285 1525ndash1536
Nucleic Acids Research 2014 Vol 42 No 1 69
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
annotation of prokaryotic genomes Nucleic Acids Res 39D11ndashD14
17 RaePM and SteeleRE (1978) Modified bases in the DNAs ofunicellular eukaryotes an examination of distributions andpossible roles with emphasis on hydroxymethyluracil indinoflagellates Biosystems 10 37ndash53
18 WarrenRAJ (1980) Modified bases in bacteriophage DNAsAnnu Rev Microbiol 34 137ndash158
19 Gommers-AmptJH and BorstP (1995) Hypermodified bases inDNA FASEB J 9 1034ndash1042
20 SwintonD HattmanS CrainPF ChengCS SmithDL andMcCloskeyJA (1983) Purification and characterization of theunusual deoxynucleoside alpha-N-(9-beta-D-20-deoxyribofuranosylpurin-6-yl)glycinamide specified by the phageMu modification function Proc Natl Acad Sci USA 807400ndash7404
21 TahilianiM KohKP ShenY PastorWA BandukwalaHBrudnoY AgarwalS IyerLM LiuDR AravindL et al(2009) Conversion of 5-methylcytosine to 5-hydroxymethylcytosinein mammalian DNA by MLL partner TET1 Science 324930ndash935
22 KriaucionisS and HeintzN (2009) The nuclear DNA base5-hydroxymethylcytosine is present in Purkinje neurons and thebrain Science 324 929ndash930
23 BorstP and SabatiniR (2008) Base J discovery biosynthesisand possible functions Annu RevMicrobiol 62 235ndash251
24 KaminskaKH and BujnickiJM (2008) Bacteriophage MuMom protein responsible for DNA modification is a new memberof the acyltransferase superfamily Cell Cycle 7 120ndash121
25 IyerLM TahilianiM RaoA and AravindL (2009) Predictionof novel families of enzymes involved in oxidative and othercomplex modifications of bases in nucleic acids Cell Cycle 81698ndash1710
26 ZhouX HeX LiangJ LiA XuT KieserT HelmannJDand DengZ (2005) A novel DNA modification by sulphur MolMicrobiol 57 1428ndash1438
27 OuHY HeX ShaoY TaiC RajakumarK and DengZ(2009) dndDB A Database Focused on Phosphorothioation ofthe DNA Backbone PloS One 4 e5132
28 XuT LiangJ ChenS WangL HeX YouD WangZLiA XuZ ZhouX et al (2009) DNA phosphorothioation inStreptomyces lividans mutational analysis of the dnd locusBMC Microbiol 9 41
29 WangL ChenS VerginKL GiovannoniSJ ChanSWDeMottMS TaghizadehK CorderoOX CutlerMTimberlakeS et al (2011) DNA phosphorothioation iswidespread and quantized in bacterial genomes Proc Natl AcadSci USA 108 2963ndash2968
30 Loslashbner-OlesenA SkovgaardO and MarinusMG (2005) Dammethylation coordinating cellular processes Curr OpinMicrobiol 8 154ndash160
31 LowDA and CasadesusJ (2008) Clocks and switches bacterialgene regulation by DNA adenine methylation Curr OpinMicrobiol 11 106ndash112
32 FreitagM and SelkerEU (2005) Controlling DNA methylationMany roads to one modification Curr Opin Genet Dev 15191ndash199
33 ProhaskaSJ StadlerPF and KrakauerDC (2010) Innovationin gene regulation The case of chromatin computation J TheorBiol 265 27ndash44
34 SeshasayeeASN SinghP and KrishnaS (2012) Context-dependent conservation of DNA methyltransferases in bacteriaNucleic Acids Res 40 7066ndash7073
35 AntonBP (2011) Dissertation Boston University BostonMassachusetts USA
36 RaleighEA TrimarchiR and RevelH (1989) Genetic andphysical mapping of the mcrA (rglA) and mcrB (rglB) loci ofEscherichia coli K-12 Genetics 122 279ndash296
37 Gonzalez-CeronG Miranda-OlivaresOJ and Servın-GonzalezL(2009) Characterization of the methyl-specific restriction system ofStreptomyces coelicolor A3(2) and of the role played by laterallyacquired nucleases FEMS Microbiol Lett 301 35ndash43
38 LiuG OuHY WangT LiL TanH ZhouXRajakumarK DengZ and HeX (2010) Cleavage of
phosphorothioated DNA and methylated DNA by the type IVrestriction endonuclease ScoMcrA PLoS Genet 6 e1001253
39 SutherlandE CoeL and RaleighEA (1992) McrBC amultisubunit GTP-dependent restriction endonuclease J MolBiol 225 327ndash348
40 KrugerT WildC and Noyer-WeidnerM (1995) McrB aprokaryotic protein specifically recognizing DNA containingmodified cytosine residues EMBO J 14 2661ndash2669
41 SitaramanR and LepplaSH (2012) Methylation-dependentDNA restriction in Bacillus anthracis Gene 494 44ndash50
42 HeitmanJ and ModelP (1987) Site-specific methylases inducethe SOS DNA repair response in Escherichia coli J Bacteriol169 3243ndash3250
43 Waite-ReesPA KeatingCJ MoranLS SlatkoBEHornstraLJ and BennerJS (1991) Characterisation andexpression of the Escherichia coli Mrr restriction systemJ Bacteriol 173 5207ndash5219
44 KerrAL JeonYJ SvensonCJ RogersPL and NeilanBA(2010) DNA restriction-modification systems in the ethanologenZymomonas mobilis ZM4 Appl MicrobiolBiotech 89 761ndash769
45 XuSY CorvagliaAR ChanSH ZhengY and LinderP(2011) A type IV modification-dependent restriction enzymeSauUSI from Staphylococcus aureus subsp aureus USA300Nucleic Acids Res 39 5597ndash5610
46 MonkIR ShahIM XuM TanMW and FosterTJ (2012)Transforming the untransformable application of directtransformation to manipulate genetically Staphylococcus aureusand Staphylococcus epidermidis mBio 3 e00277ndash00211
47 WangH GuanS QuimbyA Cohen-KarniD PradhanSWilsonG RobertsRJ ZhuZ and ZhengY (2011)Comparative characterization of the PvuRts1I family ofrestriction enzymes and their application in mapping genomic5-hydroxymethylcytosine Nucleic Acids Res 39 9294ndash9305
48 JanosiL YonemitsuH HongH and KajiA (1994) Molecularcloning and expression of a novel hydroxymethylcytosine-specificrestriction enzyme (PvuRts1I) modulated by glucosylation ofDNAJ Mol Biol 242 45ndash61
49 BairCL and BlackLW (2007) A type IV modificationdependent restriction nuclease that targets glucosylatedhydroxymethyl cytosine modified DNAs J Mol Biol 366768ndash778
50 RobertsRJ VinczeT PosfaiJ and MacelisD (2010)REBASEndasha database for DNA restriction and modificationenzymes genes and genomes Nucleic Acids Res 38 D234ndashD236
51 ZhengY Cohen-KarniD XuD ChinHG WilsonGPradhanS and RobertsRJ (2010) A unique family of Mrr-likemodification-dependent restriction endonucleases Nucleic AcidsRes 38 5527ndash5534
52 Cohen-KarniD XuD AponeL FomenkovA SunZDavisPJ KinneySRM Yamada-MabuchiM XuSYDavisT et al (2011) The MspJI family of modification-dependent restriction endonucleases for epigenetic studies ProcNatl Acad Sci USA 108 11040ndash11045
53 LubysA VitkuteJ LubieneJ and JanulaitisA (2011)Restriction enzymes and their applications US Patent ApplicationPublication 20110207139 A1 httpappftusptogovnetahtmlPTOsearch-boolhtml (27 August 2013 date last accessed)
54 TarasovaGV NayakshinaTN and DegtyarevSK (2008)Substrate specificity of new methyl-directed DNA endonucleaseGlaI BMC Mol Biol 9 7
55 MulliganEA and DunnJJ (2008) Cloning purification andinitial characterization of E coli McrA a putative5-methylcytosine-specific nuclease Protein ExprPurif 62 98ndash103
56 MulliganEA HatchwellE McCorkleSR and DunnJJ (2010)Differential binding of Escherichia coli McrA protein to DNAsequences that contain the dinucleotide m5CpG Nucleic AcidsRes 38 1997ndash2005
57 AntonBP and RaleighEA (2004) Transposon-mediated linkerinsertion scanning mutagenesis of the Escherichia coli McrAendonuclease J Bacteriol 186 5699ndash5707
58 BujnickiJM RadlinskaM and RychlewskiL (2000) Atomicmodel of the 5-methylcytosine-specific restriction enzyme McrAreveals an atypical zinc finger and structural similarity tobetabetaalphaMe endonucleases Mol Microbiol 37 1280ndash1281
Nucleic Acids Research 2014 Vol 42 No 1 67
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
59 GastFU BrinkmannT PieperU KrugerT NoyerWeidnerM and PingoudA (1997) The recognition of methylatedDNA by the GTP-dependent restriction endonuclease McrBCresides in the N-terminal domain of McrB Biol Chem 378975ndash982
60 SukackaiteR GrazulisS TamulaitisG and SiksnysV (2012)The recognition domain of the methyl-specific endonucleaseMcrBC flips out 5-methylcytosine Nucleic Acids Res 407552ndash7562
61 DilaD SutherlandE MoranL SlatkoB and RaleighEA(1990) Genetic and sequence organization of the mcrBC locus ofEscherichia coli K-12 J Bacteriol 172 4888ndash4900
62 PieperU BrinkmannT KrugerT Noyer WeidnerM andPingoudA (1997) Characterization of the interaction between therestriction endonuclease McrBC from E coli and its cofactorGTP J Mol Biol 272 190ndash199
63 PieperU SchweitzerT GrollDH GastFU and PingoudA(1999) The GTP-binding domain of McrB more than just avariation on a common theme J Mol Biol 292 547ndash556
64 BujnickiJM and RychlewskiL (2001) Grouping together highlydiverged PD-(DE)XK nucleases and identification of novelsuperfamily members using structure-guided alignment ofsequence profiles J Mol MicrobiolBiotech 3 69ndash72
65 PieperU and PingoudA (2002) A mutational analysis of thePD DEXK motif suggests that McrC harbors the catalyticcenter for DNA cleavage by the GTP-dependent restrictionenzyme McrBC from Escherichia coli Biochemistry 415236ndash5244
66 KelleherJE and RaleighEA (1991) A novel activity inEscherichia coli K-12 that directs restriction of DNA modified atCG dinucleotides J Bacteriol 173 5220ndash5223
67 OrlowskiJ MebrhatuMT MichielsCW BujnickiJM andAertsenA (2008) Mutational analysis and a structural model ofmethyl-directed restriction enzyme Mrr Biochem Biophys ResCommun 377 862ndash866
68 AravindL MakarovaKS and KooninEV (2000) SURVEYAND SUMMARY holliday junction resolvases and relatednucleases identification of new families phyletic distribution andevolutionary trajectories Nucleic Acids Res 28 3417ndash3432
69 BujnickiJM and RychlewskiL (2001) Identification of a PD-(DE)XK-like domain with a novel configuration of the endonucleaseactive site in the methyl-directed restriction enzyme Mrr and itshomologs Gene 267 183ndash191
70 SteczkiewiczK MuszewskaA KnizewskiL RychlewskiL andGinalskiK (2012) Sequence structure and functional diversity ofPD-(DE)XK phosphodiesterase superfamily Nucleic Acids Res40 7016ndash7045
71 BairCL RifatD and BlackLW (2007) Exclusion of glucosyl-hydroxymethylcytosine DNA containing bacteriophages isovercome by the injected protein inhibitor IPI J Mol Biol 366779ndash789
72 StewartFJ and RaleighEA (1998) Dependence of McrBCcleavage on distance between recognition elements Biol Chem379 611ndash616
73 PanneD RaleighEA and BickleTA (1999) The McrBCendonuclease translocates DNA in a reaction dependent on GTPhydrolysis J Mol Biol 290 49ndash60
74 PieperU GrollDH WunschS GastFU SpeckC MuckeNand PingoudA (2002) The GTP-dependent restriction enzymeMcrBC from Escherichia coli forms high-molecular masscomplexes with DNA and produces a cleavage patternwith a characteristic 10-base pair repeat Biochemistry 415245ndash5254
75 IshikawaK HandaN SearsL RaleighEA and KobayashiI(2011) Cleavage of a model DNA replication fork by amethyl-specific endonuclease Nucleic Acids Res 39 5489ndash5498
76 StewartFJ PanneD BickleTA and RaleighEA (2000)Methyl-specific DNA binding by McrBC a modification-dependent restriction enzyme J Mol Biol 298 611ndash622
77 LacksS and GreenbergB (1977) Complementary specificity ofrestriction endonucleases DpnI and DpnII with respect to DNAmethylation J Mol Biol 114 153ndash168
78 SiwekW CzapinskaH BochtlerM BujnickiJM andSkowronekK (2012) Crystal structure and mechanism
of action of the N6-methyladenine-dependent type IIMrestriction endonuclease RDpnI Nucleic Acids Res 407563ndash7572
79 LacksS and GreenbergB (1975) A deoxyribonuclease ofDiplococcus pneumoniae specific for methylated DNA J BiolChem 250 4060ndash4066
80 HortonJR MabuchiMY Cohen-KarniD ZhangXGriggsRM SamaranayakeM RobertsRJ ZhengY andChengX (2012) Structure and cleavage activity of the tetramericMspJI DNA modification-dependent restriction endonucleaseNucleic Acids Res 40 9763ndash9773
81 CymermanIA ObarskaA SkowronekKJ LubysA andBujnickiJM (2006) Identification of a new subfamily of HNHnucleases and experimental characterization of a representativemember HphI restriction endonuclease Proteins 65 867ndash876
82 GorbalenyaAE (1994) Self-splicing group I and group II intronsencode homologous (putative) DNA endonucleases of a newfamily Protein Sci 3 1117ndash1120
83 ChanSH OpitzL HigginsL OrsquoLoaneD and XuSY (2010)Cofactor requirement of HpyAV restriction endonuclease PloSOne 5 e9071
84 BourniquelAA and BickleTA (2002) Complex restrictionenzymes NTP-driven molecular motors Biochimie 84 1047ndash1059
85 OrsquoDriscollJ HeiterDF WilsonGG FitzgeraldGFRobertsR and van SinderenD (2006) A genetic dissection ofthe LlaJI restriction cassette reveals insights on a novelbacteriophage resistance system BMC Microbiol 6 40
86 SmithRM DiffinFM SaveryNJ JosephsenJ andSzczelkunMD (2009) DNA cleavage and methylation specificityof the single polypeptide restriction-modification enzyme LlaGINucleic Acids Res 37 7206ndash7218
87 InterthalH PouliotJJ and ChampouxJJ (2001) The tyrosyl-DNA phosphodiesterase Tdp1 is a member of the phospholipaseD superfamily Proc Natl Acad Sci USA 98 12009ndash12014
88 SasnauskasG HalfordSE and SiksnysV (2003) How the BfiIrestriction enzyme uses one active site to cut two DNA strandsProc Natl Acad Sci USA 100 6410ndash6415
89 SasnauskasG ConnollyBA HalfordSE and SiksnysV (2007)Site-specific DNA transesterification catalyzed by a restrictionenzyme Proc Natl Acad Sci USA 104 2115ndash2120
90 Marchler-BauerA ZhengC ChitsazF DerbyshireMKGeerLY GeerRC GonzalesNR GwadzM HurwitzDILanczyckiCJ et al (2012) CDD conserved domains andprotein three-dimensional structure Nucleic Acids Res 41D348ndashD352
91 KravetzAN ZakharovaMV BeletskayaIV SinevaEVDenjmuchametovMM PetrovSI GlatmanLI andSoloninAS (1993) The cleavage sites and localization of genesencoding the restriction endonucleases Eco1831I and EcoHIGene 129 153ndash154
92 KlimasauskasS KumarS RobertsRJ and ChengX (1994)HhaI methyltransferase flips its target base out of the DNA helixCell 76 357ndash369
93 HashimotoH VertinoPM and ChengX (2010) Molecularcoupling of DNA methylation and histone methylationEpigenomics 2 657ndash669
94 RajakumaraE LawJA SimanshuDK VoigtPJohnsonLM ReinbergD PatelDJ and JacobsenSE (2011)A dual flip-out mechanism for 5mC recognition by theArabidopsis SUVH5 SRA domain and its impact on DNAmethylation and H3K9 dimethylation in vivo Genes Dev 25137ndash152
95 SharifJ and KosekiH (2011) Recruitment of Dnmt1 roles of theSRA protein Np95 (Uhrfi) and other factors Prog Mol BiolTransl Sci 101 289ndash310
96 Tesfazgi MebrhatuM WywialE GhoshA MichielsCWLindnerAB TaddeiF BujnickiJM Van MelderenL andAertsenA (2011) Evidence for an evolutionary antagonismbetween Mrr and Type III modification systems Nucleic AcidsRes 39 5991ndash6001
97 GrantGN JesseeJ BloomFR and HanahanD (1990)Differential plasmid rescue from transgenic mouse DNAs intoEscherichia coli methylation-restriction mutants Proc Natl AcadSci USA 87 4645ndash4649
68 Nucleic Acids Research 2014 Vol 42 No 1
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
98 AertsenA and MichielsCW (2005) Mrr instigates the SOSresponse after high pressure stress in Escherichia coli MolMicrobiol 58 1381ndash1391
99 WeiH TherrienC BlanchardA GuanS and ZhuZ (2008)The Fidelity Index provides a systematic quantitation of staractivity of DNA restriction endonucleases Nucleic Acids Res36 e50
100 LuL PatelH and BisslerJJ (2002) Optimizing DpnIdigestion to detect replicated DNA Biotechniques 33 316ndash318
101 MadejT AddessKJ FongJH GeerLY GeerRCLanczyckiCJ LiuC LuS Marchler-BauerAPanchenkoAR et al (2011) MMDB 3D structures andmacromolecular interactions Nucleic Acids Res 40 D461ndashD464
102 GibratJF MadejT and BryantSH (1996) Surprisingsimilarities in structure comparison Curr Opin Struct Biol 6377ndash385
103 PanneD MullerSA WirtzS EngelA and BickleTA (2001)The McrBC restriction endonuclease assembles into a ringstructure in the presence of G nucleotides EMBO J 203210ndash3217
104 NeuwaldAF AravindL SpougeJL and KooninEV (1999)AAA+ A class of chaperone-like ATPases associated with theassembly operation and disassembly of protein complexesGenome Res 9 27ndash43
105 OguraT and WilkinsonAJ (2001) AAA+ superfamilyATPases common structurendashdiverse function Genes Cells 6575ndash597
106 BearyTP BraymerHD and AchbergerEC (1997) Evidenceof participation of McrB(S) in McrBC restriction in Escherichiacoli K-12 J Bacteriol 179 7768ndash7775
107 PanneD RaleighEA and BickleTA (1998) McrBs amodulator peptide for McrBC activity EMBO J 175477ndash5483
108 PuntaM CoggillPC EberhardtRY MistryJ TateJBoursnellC PangN ForslundK CericG ClementsJ et al(2011) The Pfam protein families database Nucleic Acids Res40 D290ndashD301
109 NaitoT KusanoK and KobayashiI (1995) Selfish behavior ofrestriction-modification systems Science 267 897ndash899
110 KobayashiI (2001) Behavior of restriction-modification systemsas selfish mobile elements and their impact on genome evolutionNucleic Acids Res 29 3742ndash3756
111 FukudaE KaminskaKH BujnickiJM and KobayashiI(2008) Cell death upon epigenetic genome methylation a novelfunction of methyl-specific deoxyribonucleases Genome Biol 9R163
112 IshikawaK FukudaE and KobayashiI (2010) Conflictstargeting epigenetic systems and their resolution by cell deathnovel concepts for methyl-specific and other restriction systemsDNA Res 17 325ndash342
113 FukuyoM SasakiA and KobayashiI (2012) Success of asuicidal defense strategy against infection in a structured habitatSci Rep 2 238
114 MakovetsS TitheradgeAJ and MurrayNE (1998) ClpXand ClpP are essential for the efficient acquisition of genesspecifying type IA and IB restriction systems Mol Microbiol28 25ndash35
115 MurrayNE (2002) Immigration control of DNA in bacteriaself versus non-self Microbiology 148 3ndash20
116 MakovetsS PowellLM TitheradgeAJB BlakelyGW andMurrayNE (2004) Is modification sufficient to protect abacterial chromosome from a resident restriction endonucleaseMol Microbiol 51 135ndash147
117 OrgelLE and CrickFHC (1980) Selfish DNA the ultimateparasite Nature 284 604ndash607
118 KidwellMG and LischDR (2001) Perspective transposableelements parasitic DNA and genome evolution Evolution 551ndash24
119 SadykovM AsamiY NikiH HandaN ItayaMTanokuraM and KobayashiI (2003) Multiplication of arestriction-modification gene complex Mol Microbiol 48417ndash427
120 ChaoL and LevinBR (1981) Structured habitats and theevolution of anticompetitor toxins in bacteria Proc Natl AcadSci USA 78 6324ndash6328
121 MagnusonRD (2007) Hypothetical functions of toxin-antitoxinsystems J Bacteriol 189 6089ndash6092
122 RifatD WrightNT VarneyKM WeberDJ and BlackLW(2008) Restriction endonuclease inhibitor IPI of bacteriophageT4 a novel structure for a dedicated target J Mol Biol 375720ndash734
123 LabrieSJ SamsonJE and MoineauS (2010) Bacteriophageresistance mechanisms Nat Rev Microbiol 8 317ndash327
124 LinLF PosfaiJ RobertsRJ and KongH (2001)Comparative genomics of the restriction-modification systemsin Helicobacter pylori Proc Natl Acad Sci USA 982740ndash2745
125 MilkmanR RaleighEA McKaneM CrydermanDBilodeauP and McWeenyK (1999) Molecular evolution of theEscherichia coli chromosome V Recombination patterns amongstrains of diverse origin Genetics 153 539ndash554
126 HumbertO DorerMS and SalamaNR (2011)Characterization of Helicobacter pylori factors that controltransformation frequency and integration length during inter-strain DNA recombination Mol Microbiol 79 387ndash401
127 BarcusVA TitheradgeAJB and MurrayNE (1995) Thediversity of alleles at the hsd locus in natural populations ofEscherichia coli Genetics 140 1187ndash1197
128 SharpPM KelleherJE DanielAS CowanGM andMurrayNE (1992) Roles of selection and recombination in theevolution of type I restriction-modification systems inenterobacteria Proc Natl Acad Sci USA 89 9836ndash9840
129 LoenenWAM DanielAS BraymerHD and MurrayNE(1987) Organization and sequence of the hsd genes ofEscherichia coli K-12 J Mol Biol 198 159ndash170
130 SuriB and BickleTA (1985) EcoA the first member of a newfamily of type I restriction modification systems Geneorganization and enzymatic activities J Mol Biol 186 77ndash85
131 BertaniG and WeigleJJ (1953) Host-controlled variation inbacterial viruses J Bacteriol 65 113ndash121
132 SibleyMH and RaleighEA (2004) Cassette-like variation ofrestriction enzyme genes in Escherichia coli C and relativesNucleic Acids Res 32 522ndash534
133 MakarovaKS WolfYI SnirS and KooninEV (2011)Defense islands in bacterial and archaeal genomes and predictionof novel defense systems J Bacteriol 193 6039ndash6056
134 FurutaY AbeK and KobayashiI (2010) Genome comparisonand context analysis reveals putative mobile forms of restriction-modification systems and related rearrangements Nucleic AcidsRes 38 2428ndash2443
135 FurutaY and KobayashiI (2012) Restriction-modificationsystems as moblie epigenetic elements In RobertsAP andMullanyP (eds) Bacterial Integrative Mobile Genetic ElementsLandes Bioscience pp 85ndash103
136 OrsquoSullivanDJ and KlaenhammerTR (1998) Control ofexpression of LlaI restriction in Lactococcus lactis MolMicrobiol 27 1009ndash1020
137 WhitakerRD DornerLF and SchildkrautI (1999) A mutantof BamHI restriction endonuclease which requires N6-methyladenine for cleavage J Mol Biol 285 1525ndash1536
Nucleic Acids Research 2014 Vol 42 No 1 69
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
59 GastFU BrinkmannT PieperU KrugerT NoyerWeidnerM and PingoudA (1997) The recognition of methylatedDNA by the GTP-dependent restriction endonuclease McrBCresides in the N-terminal domain of McrB Biol Chem 378975ndash982
60 SukackaiteR GrazulisS TamulaitisG and SiksnysV (2012)The recognition domain of the methyl-specific endonucleaseMcrBC flips out 5-methylcytosine Nucleic Acids Res 407552ndash7562
61 DilaD SutherlandE MoranL SlatkoB and RaleighEA(1990) Genetic and sequence organization of the mcrBC locus ofEscherichia coli K-12 J Bacteriol 172 4888ndash4900
62 PieperU BrinkmannT KrugerT Noyer WeidnerM andPingoudA (1997) Characterization of the interaction between therestriction endonuclease McrBC from E coli and its cofactorGTP J Mol Biol 272 190ndash199
63 PieperU SchweitzerT GrollDH GastFU and PingoudA(1999) The GTP-binding domain of McrB more than just avariation on a common theme J Mol Biol 292 547ndash556
64 BujnickiJM and RychlewskiL (2001) Grouping together highlydiverged PD-(DE)XK nucleases and identification of novelsuperfamily members using structure-guided alignment ofsequence profiles J Mol MicrobiolBiotech 3 69ndash72
65 PieperU and PingoudA (2002) A mutational analysis of thePD DEXK motif suggests that McrC harbors the catalyticcenter for DNA cleavage by the GTP-dependent restrictionenzyme McrBC from Escherichia coli Biochemistry 415236ndash5244
66 KelleherJE and RaleighEA (1991) A novel activity inEscherichia coli K-12 that directs restriction of DNA modified atCG dinucleotides J Bacteriol 173 5220ndash5223
67 OrlowskiJ MebrhatuMT MichielsCW BujnickiJM andAertsenA (2008) Mutational analysis and a structural model ofmethyl-directed restriction enzyme Mrr Biochem Biophys ResCommun 377 862ndash866
68 AravindL MakarovaKS and KooninEV (2000) SURVEYAND SUMMARY holliday junction resolvases and relatednucleases identification of new families phyletic distribution andevolutionary trajectories Nucleic Acids Res 28 3417ndash3432
69 BujnickiJM and RychlewskiL (2001) Identification of a PD-(DE)XK-like domain with a novel configuration of the endonucleaseactive site in the methyl-directed restriction enzyme Mrr and itshomologs Gene 267 183ndash191
70 SteczkiewiczK MuszewskaA KnizewskiL RychlewskiL andGinalskiK (2012) Sequence structure and functional diversity ofPD-(DE)XK phosphodiesterase superfamily Nucleic Acids Res40 7016ndash7045
71 BairCL RifatD and BlackLW (2007) Exclusion of glucosyl-hydroxymethylcytosine DNA containing bacteriophages isovercome by the injected protein inhibitor IPI J Mol Biol 366779ndash789
72 StewartFJ and RaleighEA (1998) Dependence of McrBCcleavage on distance between recognition elements Biol Chem379 611ndash616
73 PanneD RaleighEA and BickleTA (1999) The McrBCendonuclease translocates DNA in a reaction dependent on GTPhydrolysis J Mol Biol 290 49ndash60
74 PieperU GrollDH WunschS GastFU SpeckC MuckeNand PingoudA (2002) The GTP-dependent restriction enzymeMcrBC from Escherichia coli forms high-molecular masscomplexes with DNA and produces a cleavage patternwith a characteristic 10-base pair repeat Biochemistry 415245ndash5254
75 IshikawaK HandaN SearsL RaleighEA and KobayashiI(2011) Cleavage of a model DNA replication fork by amethyl-specific endonuclease Nucleic Acids Res 39 5489ndash5498
76 StewartFJ PanneD BickleTA and RaleighEA (2000)Methyl-specific DNA binding by McrBC a modification-dependent restriction enzyme J Mol Biol 298 611ndash622
77 LacksS and GreenbergB (1977) Complementary specificity ofrestriction endonucleases DpnI and DpnII with respect to DNAmethylation J Mol Biol 114 153ndash168
78 SiwekW CzapinskaH BochtlerM BujnickiJM andSkowronekK (2012) Crystal structure and mechanism
of action of the N6-methyladenine-dependent type IIMrestriction endonuclease RDpnI Nucleic Acids Res 407563ndash7572
79 LacksS and GreenbergB (1975) A deoxyribonuclease ofDiplococcus pneumoniae specific for methylated DNA J BiolChem 250 4060ndash4066
80 HortonJR MabuchiMY Cohen-KarniD ZhangXGriggsRM SamaranayakeM RobertsRJ ZhengY andChengX (2012) Structure and cleavage activity of the tetramericMspJI DNA modification-dependent restriction endonucleaseNucleic Acids Res 40 9763ndash9773
81 CymermanIA ObarskaA SkowronekKJ LubysA andBujnickiJM (2006) Identification of a new subfamily of HNHnucleases and experimental characterization of a representativemember HphI restriction endonuclease Proteins 65 867ndash876
82 GorbalenyaAE (1994) Self-splicing group I and group II intronsencode homologous (putative) DNA endonucleases of a newfamily Protein Sci 3 1117ndash1120
83 ChanSH OpitzL HigginsL OrsquoLoaneD and XuSY (2010)Cofactor requirement of HpyAV restriction endonuclease PloSOne 5 e9071
84 BourniquelAA and BickleTA (2002) Complex restrictionenzymes NTP-driven molecular motors Biochimie 84 1047ndash1059
85 OrsquoDriscollJ HeiterDF WilsonGG FitzgeraldGFRobertsR and van SinderenD (2006) A genetic dissection ofthe LlaJI restriction cassette reveals insights on a novelbacteriophage resistance system BMC Microbiol 6 40
86 SmithRM DiffinFM SaveryNJ JosephsenJ andSzczelkunMD (2009) DNA cleavage and methylation specificityof the single polypeptide restriction-modification enzyme LlaGINucleic Acids Res 37 7206ndash7218
87 InterthalH PouliotJJ and ChampouxJJ (2001) The tyrosyl-DNA phosphodiesterase Tdp1 is a member of the phospholipaseD superfamily Proc Natl Acad Sci USA 98 12009ndash12014
88 SasnauskasG HalfordSE and SiksnysV (2003) How the BfiIrestriction enzyme uses one active site to cut two DNA strandsProc Natl Acad Sci USA 100 6410ndash6415
89 SasnauskasG ConnollyBA HalfordSE and SiksnysV (2007)Site-specific DNA transesterification catalyzed by a restrictionenzyme Proc Natl Acad Sci USA 104 2115ndash2120
90 Marchler-BauerA ZhengC ChitsazF DerbyshireMKGeerLY GeerRC GonzalesNR GwadzM HurwitzDILanczyckiCJ et al (2012) CDD conserved domains andprotein three-dimensional structure Nucleic Acids Res 41D348ndashD352
91 KravetzAN ZakharovaMV BeletskayaIV SinevaEVDenjmuchametovMM PetrovSI GlatmanLI andSoloninAS (1993) The cleavage sites and localization of genesencoding the restriction endonucleases Eco1831I and EcoHIGene 129 153ndash154
92 KlimasauskasS KumarS RobertsRJ and ChengX (1994)HhaI methyltransferase flips its target base out of the DNA helixCell 76 357ndash369
93 HashimotoH VertinoPM and ChengX (2010) Molecularcoupling of DNA methylation and histone methylationEpigenomics 2 657ndash669
94 RajakumaraE LawJA SimanshuDK VoigtPJohnsonLM ReinbergD PatelDJ and JacobsenSE (2011)A dual flip-out mechanism for 5mC recognition by theArabidopsis SUVH5 SRA domain and its impact on DNAmethylation and H3K9 dimethylation in vivo Genes Dev 25137ndash152
95 SharifJ and KosekiH (2011) Recruitment of Dnmt1 roles of theSRA protein Np95 (Uhrfi) and other factors Prog Mol BiolTransl Sci 101 289ndash310
96 Tesfazgi MebrhatuM WywialE GhoshA MichielsCWLindnerAB TaddeiF BujnickiJM Van MelderenL andAertsenA (2011) Evidence for an evolutionary antagonismbetween Mrr and Type III modification systems Nucleic AcidsRes 39 5991ndash6001
97 GrantGN JesseeJ BloomFR and HanahanD (1990)Differential plasmid rescue from transgenic mouse DNAs intoEscherichia coli methylation-restriction mutants Proc Natl AcadSci USA 87 4645ndash4649
68 Nucleic Acids Research 2014 Vol 42 No 1
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
98 AertsenA and MichielsCW (2005) Mrr instigates the SOSresponse after high pressure stress in Escherichia coli MolMicrobiol 58 1381ndash1391
99 WeiH TherrienC BlanchardA GuanS and ZhuZ (2008)The Fidelity Index provides a systematic quantitation of staractivity of DNA restriction endonucleases Nucleic Acids Res36 e50
100 LuL PatelH and BisslerJJ (2002) Optimizing DpnIdigestion to detect replicated DNA Biotechniques 33 316ndash318
101 MadejT AddessKJ FongJH GeerLY GeerRCLanczyckiCJ LiuC LuS Marchler-BauerAPanchenkoAR et al (2011) MMDB 3D structures andmacromolecular interactions Nucleic Acids Res 40 D461ndashD464
102 GibratJF MadejT and BryantSH (1996) Surprisingsimilarities in structure comparison Curr Opin Struct Biol 6377ndash385
103 PanneD MullerSA WirtzS EngelA and BickleTA (2001)The McrBC restriction endonuclease assembles into a ringstructure in the presence of G nucleotides EMBO J 203210ndash3217
104 NeuwaldAF AravindL SpougeJL and KooninEV (1999)AAA+ A class of chaperone-like ATPases associated with theassembly operation and disassembly of protein complexesGenome Res 9 27ndash43
105 OguraT and WilkinsonAJ (2001) AAA+ superfamilyATPases common structurendashdiverse function Genes Cells 6575ndash597
106 BearyTP BraymerHD and AchbergerEC (1997) Evidenceof participation of McrB(S) in McrBC restriction in Escherichiacoli K-12 J Bacteriol 179 7768ndash7775
107 PanneD RaleighEA and BickleTA (1998) McrBs amodulator peptide for McrBC activity EMBO J 175477ndash5483
108 PuntaM CoggillPC EberhardtRY MistryJ TateJBoursnellC PangN ForslundK CericG ClementsJ et al(2011) The Pfam protein families database Nucleic Acids Res40 D290ndashD301
109 NaitoT KusanoK and KobayashiI (1995) Selfish behavior ofrestriction-modification systems Science 267 897ndash899
110 KobayashiI (2001) Behavior of restriction-modification systemsas selfish mobile elements and their impact on genome evolutionNucleic Acids Res 29 3742ndash3756
111 FukudaE KaminskaKH BujnickiJM and KobayashiI(2008) Cell death upon epigenetic genome methylation a novelfunction of methyl-specific deoxyribonucleases Genome Biol 9R163
112 IshikawaK FukudaE and KobayashiI (2010) Conflictstargeting epigenetic systems and their resolution by cell deathnovel concepts for methyl-specific and other restriction systemsDNA Res 17 325ndash342
113 FukuyoM SasakiA and KobayashiI (2012) Success of asuicidal defense strategy against infection in a structured habitatSci Rep 2 238
114 MakovetsS TitheradgeAJ and MurrayNE (1998) ClpXand ClpP are essential for the efficient acquisition of genesspecifying type IA and IB restriction systems Mol Microbiol28 25ndash35
115 MurrayNE (2002) Immigration control of DNA in bacteriaself versus non-self Microbiology 148 3ndash20
116 MakovetsS PowellLM TitheradgeAJB BlakelyGW andMurrayNE (2004) Is modification sufficient to protect abacterial chromosome from a resident restriction endonucleaseMol Microbiol 51 135ndash147
117 OrgelLE and CrickFHC (1980) Selfish DNA the ultimateparasite Nature 284 604ndash607
118 KidwellMG and LischDR (2001) Perspective transposableelements parasitic DNA and genome evolution Evolution 551ndash24
119 SadykovM AsamiY NikiH HandaN ItayaMTanokuraM and KobayashiI (2003) Multiplication of arestriction-modification gene complex Mol Microbiol 48417ndash427
120 ChaoL and LevinBR (1981) Structured habitats and theevolution of anticompetitor toxins in bacteria Proc Natl AcadSci USA 78 6324ndash6328
121 MagnusonRD (2007) Hypothetical functions of toxin-antitoxinsystems J Bacteriol 189 6089ndash6092
122 RifatD WrightNT VarneyKM WeberDJ and BlackLW(2008) Restriction endonuclease inhibitor IPI of bacteriophageT4 a novel structure for a dedicated target J Mol Biol 375720ndash734
123 LabrieSJ SamsonJE and MoineauS (2010) Bacteriophageresistance mechanisms Nat Rev Microbiol 8 317ndash327
124 LinLF PosfaiJ RobertsRJ and KongH (2001)Comparative genomics of the restriction-modification systemsin Helicobacter pylori Proc Natl Acad Sci USA 982740ndash2745
125 MilkmanR RaleighEA McKaneM CrydermanDBilodeauP and McWeenyK (1999) Molecular evolution of theEscherichia coli chromosome V Recombination patterns amongstrains of diverse origin Genetics 153 539ndash554
126 HumbertO DorerMS and SalamaNR (2011)Characterization of Helicobacter pylori factors that controltransformation frequency and integration length during inter-strain DNA recombination Mol Microbiol 79 387ndash401
127 BarcusVA TitheradgeAJB and MurrayNE (1995) Thediversity of alleles at the hsd locus in natural populations ofEscherichia coli Genetics 140 1187ndash1197
128 SharpPM KelleherJE DanielAS CowanGM andMurrayNE (1992) Roles of selection and recombination in theevolution of type I restriction-modification systems inenterobacteria Proc Natl Acad Sci USA 89 9836ndash9840
129 LoenenWAM DanielAS BraymerHD and MurrayNE(1987) Organization and sequence of the hsd genes ofEscherichia coli K-12 J Mol Biol 198 159ndash170
130 SuriB and BickleTA (1985) EcoA the first member of a newfamily of type I restriction modification systems Geneorganization and enzymatic activities J Mol Biol 186 77ndash85
131 BertaniG and WeigleJJ (1953) Host-controlled variation inbacterial viruses J Bacteriol 65 113ndash121
132 SibleyMH and RaleighEA (2004) Cassette-like variation ofrestriction enzyme genes in Escherichia coli C and relativesNucleic Acids Res 32 522ndash534
133 MakarovaKS WolfYI SnirS and KooninEV (2011)Defense islands in bacterial and archaeal genomes and predictionof novel defense systems J Bacteriol 193 6039ndash6056
134 FurutaY AbeK and KobayashiI (2010) Genome comparisonand context analysis reveals putative mobile forms of restriction-modification systems and related rearrangements Nucleic AcidsRes 38 2428ndash2443
135 FurutaY and KobayashiI (2012) Restriction-modificationsystems as moblie epigenetic elements In RobertsAP andMullanyP (eds) Bacterial Integrative Mobile Genetic ElementsLandes Bioscience pp 85ndash103
136 OrsquoSullivanDJ and KlaenhammerTR (1998) Control ofexpression of LlaI restriction in Lactococcus lactis MolMicrobiol 27 1009ndash1020
137 WhitakerRD DornerLF and SchildkrautI (1999) A mutantof BamHI restriction endonuclease which requires N6-methyladenine for cleavage J Mol Biol 285 1525ndash1536
Nucleic Acids Research 2014 Vol 42 No 1 69
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018
98 AertsenA and MichielsCW (2005) Mrr instigates the SOSresponse after high pressure stress in Escherichia coli MolMicrobiol 58 1381ndash1391
99 WeiH TherrienC BlanchardA GuanS and ZhuZ (2008)The Fidelity Index provides a systematic quantitation of staractivity of DNA restriction endonucleases Nucleic Acids Res36 e50
100 LuL PatelH and BisslerJJ (2002) Optimizing DpnIdigestion to detect replicated DNA Biotechniques 33 316ndash318
101 MadejT AddessKJ FongJH GeerLY GeerRCLanczyckiCJ LiuC LuS Marchler-BauerAPanchenkoAR et al (2011) MMDB 3D structures andmacromolecular interactions Nucleic Acids Res 40 D461ndashD464
102 GibratJF MadejT and BryantSH (1996) Surprisingsimilarities in structure comparison Curr Opin Struct Biol 6377ndash385
103 PanneD MullerSA WirtzS EngelA and BickleTA (2001)The McrBC restriction endonuclease assembles into a ringstructure in the presence of G nucleotides EMBO J 203210ndash3217
104 NeuwaldAF AravindL SpougeJL and KooninEV (1999)AAA+ A class of chaperone-like ATPases associated with theassembly operation and disassembly of protein complexesGenome Res 9 27ndash43
105 OguraT and WilkinsonAJ (2001) AAA+ superfamilyATPases common structurendashdiverse function Genes Cells 6575ndash597
106 BearyTP BraymerHD and AchbergerEC (1997) Evidenceof participation of McrB(S) in McrBC restriction in Escherichiacoli K-12 J Bacteriol 179 7768ndash7775
107 PanneD RaleighEA and BickleTA (1998) McrBs amodulator peptide for McrBC activity EMBO J 175477ndash5483
108 PuntaM CoggillPC EberhardtRY MistryJ TateJBoursnellC PangN ForslundK CericG ClementsJ et al(2011) The Pfam protein families database Nucleic Acids Res40 D290ndashD301
109 NaitoT KusanoK and KobayashiI (1995) Selfish behavior ofrestriction-modification systems Science 267 897ndash899
110 KobayashiI (2001) Behavior of restriction-modification systemsas selfish mobile elements and their impact on genome evolutionNucleic Acids Res 29 3742ndash3756
111 FukudaE KaminskaKH BujnickiJM and KobayashiI(2008) Cell death upon epigenetic genome methylation a novelfunction of methyl-specific deoxyribonucleases Genome Biol 9R163
112 IshikawaK FukudaE and KobayashiI (2010) Conflictstargeting epigenetic systems and their resolution by cell deathnovel concepts for methyl-specific and other restriction systemsDNA Res 17 325ndash342
113 FukuyoM SasakiA and KobayashiI (2012) Success of asuicidal defense strategy against infection in a structured habitatSci Rep 2 238
114 MakovetsS TitheradgeAJ and MurrayNE (1998) ClpXand ClpP are essential for the efficient acquisition of genesspecifying type IA and IB restriction systems Mol Microbiol28 25ndash35
115 MurrayNE (2002) Immigration control of DNA in bacteriaself versus non-self Microbiology 148 3ndash20
116 MakovetsS PowellLM TitheradgeAJB BlakelyGW andMurrayNE (2004) Is modification sufficient to protect abacterial chromosome from a resident restriction endonucleaseMol Microbiol 51 135ndash147
117 OrgelLE and CrickFHC (1980) Selfish DNA the ultimateparasite Nature 284 604ndash607
118 KidwellMG and LischDR (2001) Perspective transposableelements parasitic DNA and genome evolution Evolution 551ndash24
119 SadykovM AsamiY NikiH HandaN ItayaMTanokuraM and KobayashiI (2003) Multiplication of arestriction-modification gene complex Mol Microbiol 48417ndash427
120 ChaoL and LevinBR (1981) Structured habitats and theevolution of anticompetitor toxins in bacteria Proc Natl AcadSci USA 78 6324ndash6328
121 MagnusonRD (2007) Hypothetical functions of toxin-antitoxinsystems J Bacteriol 189 6089ndash6092
122 RifatD WrightNT VarneyKM WeberDJ and BlackLW(2008) Restriction endonuclease inhibitor IPI of bacteriophageT4 a novel structure for a dedicated target J Mol Biol 375720ndash734
123 LabrieSJ SamsonJE and MoineauS (2010) Bacteriophageresistance mechanisms Nat Rev Microbiol 8 317ndash327
124 LinLF PosfaiJ RobertsRJ and KongH (2001)Comparative genomics of the restriction-modification systemsin Helicobacter pylori Proc Natl Acad Sci USA 982740ndash2745
125 MilkmanR RaleighEA McKaneM CrydermanDBilodeauP and McWeenyK (1999) Molecular evolution of theEscherichia coli chromosome V Recombination patterns amongstrains of diverse origin Genetics 153 539ndash554
126 HumbertO DorerMS and SalamaNR (2011)Characterization of Helicobacter pylori factors that controltransformation frequency and integration length during inter-strain DNA recombination Mol Microbiol 79 387ndash401
127 BarcusVA TitheradgeAJB and MurrayNE (1995) Thediversity of alleles at the hsd locus in natural populations ofEscherichia coli Genetics 140 1187ndash1197
128 SharpPM KelleherJE DanielAS CowanGM andMurrayNE (1992) Roles of selection and recombination in theevolution of type I restriction-modification systems inenterobacteria Proc Natl Acad Sci USA 89 9836ndash9840
129 LoenenWAM DanielAS BraymerHD and MurrayNE(1987) Organization and sequence of the hsd genes ofEscherichia coli K-12 J Mol Biol 198 159ndash170
130 SuriB and BickleTA (1985) EcoA the first member of a newfamily of type I restriction modification systems Geneorganization and enzymatic activities J Mol Biol 186 77ndash85
131 BertaniG and WeigleJJ (1953) Host-controlled variation inbacterial viruses J Bacteriol 65 113ndash121
132 SibleyMH and RaleighEA (2004) Cassette-like variation ofrestriction enzyme genes in Escherichia coli C and relativesNucleic Acids Res 32 522ndash534
133 MakarovaKS WolfYI SnirS and KooninEV (2011)Defense islands in bacterial and archaeal genomes and predictionof novel defense systems J Bacteriol 193 6039ndash6056
134 FurutaY AbeK and KobayashiI (2010) Genome comparisonand context analysis reveals putative mobile forms of restriction-modification systems and related rearrangements Nucleic AcidsRes 38 2428ndash2443
135 FurutaY and KobayashiI (2012) Restriction-modificationsystems as moblie epigenetic elements In RobertsAP andMullanyP (eds) Bacterial Integrative Mobile Genetic ElementsLandes Bioscience pp 85ndash103
136 OrsquoSullivanDJ and KlaenhammerTR (1998) Control ofexpression of LlaI restriction in Lactococcus lactis MolMicrobiol 27 1009ndash1020
137 WhitakerRD DornerLF and SchildkrautI (1999) A mutantof BamHI restriction endonuclease which requires N6-methyladenine for cleavage J Mol Biol 285 1525ndash1536
Nucleic Acids Research 2014 Vol 42 No 1 69
Downloaded from httpsacademicoupcomnararticle-abstract421562434997by gueston 14 February 2018