Genetic Environment and Stability of cfr in Methicillin-ResistantStaphylococcus aureus CM05

Jeffrey B. Locke,a Shahad Rahawi,a Jacqueline LaMarre,b Alexander S. Mankin,b and Karen Joy Shawa

Trius Therapeutics, Inc., San Diego, California, USA,a and Center for Pharmaceutical Biotechnology, University of Illinois at Chicago, Chicago, Illinois, USAb

The Cfr methyltransferase confers resistance to many 50S ribosomal subunit-targeted antibiotics, including linezolid (LZD), viamethylation of the 23S rRNA base A2503 in the peptidyl transferase center. Methicillin-resistant Staphylococcus aureus strainCM05 is the first clinical isolate documented to carry cfr. While cfr is typically plasmid borne, in CM05 it is located on the chro-mosome and is coexpressed with ermB as part of the mlr operon. Here we evaluated the chromosomal locus, association withmobile genetic elements, and stability of the cfr insertion region in CM05. The cfr-containing mlr operon is located within a15.5-kb plasmid-like insertion into 23S rRNA allele 4. The region surrounding the cfr gene has a high degree of sequence similar-ity to the broad-host-range toxin/antitoxin multidrug resistance plasmid pSM19035, including a second ermB gene downstreamof the mlr locus and istAS-istBS. Analysis of several individual CM05 colonies revealed two distinct populations for which LZDMICs were either 8 or 2 !g/ml. In the LZDs colonies (designated CM05"), a recombination event involving the two ermB geneshad occurred, resulting in the deletion of cfr and the 3= flanking region (cfr-istAS-istBS-ermB). The fitness advantage of CM05"over CM05 (though not likely due to the cfr deletion itself) results in the predominance of CM05" in the absence of selectivepressure. Minicircles resulting from the ermB recombination event and the novel association of cfr with the pSM19035 plasmidsystem support the potential for the continued dissemination of cfr.

Resistance to linezolid (LZD) was reported shortly after its in-troduction into the clinic and included both staphylococcal

and enterococcal strains carrying G2576T mutations in one ormore alleles of the domain V-encoding region of 23S rRNA genes(21, 57). These mutations perturb the conformation of the pepti-dyl transferase center, which is the target of oxazolidinones in the50S ribosomal subunit (30, 52, 59). More recently, LZD resistancehas been linked to additional mutations identified in the genesencoding 50S ribosomal proteins L3 and L4 (34, 58). Most prob-lematic has been the identification of cfr, a horizontally transfer-able gene encoding a ribosomal methyltransferase that confersresistance to multiple classes of antibacterial agents targeting thepeptidyl transferase center by posttranscriptional modification ofthe 23S rRNA (29). Recent studies have shown that Cfr methylatescarbon 8 of nucleotide A2503 (20), thereby conferring a PhLOPSA

phenotype characterized by decreased binding of phenicols, lin-cosamides, oxazolidinones, pleuromutilins, and streptogramin Aclass antibiotics (29, 37, 56), as well as of 16-membered ring mac-rolides (55), to Cfr-methylated ribosomes. Of further concern isthe low fitness cost associated with Cfr methylation (31) and thedocumentation of clinical cfr strains possessing co-occurring mu-tations in 23S rRNA (14) and ribosomal proteins L3 (36) and L4(5). The need to maintain potency against strains with cfr and/orother LZD resistance determinants has supported the develop-ment of expanded-spectrum oxazolidinones, such as TR-700,which retains MIC activity in the presence of Cfr methylation (33)and also has greater potency than LZD against strains possessingribosomal structural mutations (34–36).

The cfr gene was originally identified on the pSCSF1 plasmidfrom Staphylococcus sciuri, isolated from the nasal swab of aflorfenicol-treated calf with a respiratory infection (Bavaria, Ger-many, 1997) (28, 51). The gene was later found in several staphy-lococcal veterinary isolates from swine (Staphylococcus aureus,Staphylococcus lentus) and cattle (Staphylococcus simulans, S. au-reus) on both pSCFS3 and pSCFS1 (28). It was also detected

among florfenicol-resistant porcine-derived staphylococci fromDenmark on pSCSF6, isolated from Staphylococcus warneri and S.simulans, and in the chromosome of Staphylococcus hyicus isolates(26). A recent report from China identified a cfr-positive Bacillusisolate from swine feces (16). The first clinical isolate documentedto carry the cfr gene was a methicillin-resistant Staphylococcus au-reus (MRSA) strain from a patient in Medellin, Colombia (2005),who had received two doses of linezolid (56). In this isolate, des-ignated CM05, the presence of the cfr gene was demonstrated to beassociated with oxazolidinone resistance (56). Subsequently, ad-ditional LZD-resistant (LZDr) clinical staphylococci (S. aureusand coagulase-negative staphylococci [CoNS]) of diverse geo-graphical origins (Arizona, Ohio, Belgium, Italy, Spain, Mexico,and Ireland) have been shown to carry the cfr gene (5, 7, 14, 17, 19,23, 24, 39–43, 53). Approximately one-quarter of the cfr-positiveisolates described in these reports are S. aureus, and the remainderare CoNS, largely Staphylococcus epidermidis. Recently, the firstclinical cfr-positive enterococci (Enterococcus faecium and Entero-coccus faecalis) were characterized and were shown to possessG2576T 23S rRNA mutations in addition to cfr (14). The cfr geneis typically monocistronic, associated with transposons, and plas-mid borne, which presumably can result in ready exchange be-tween Gram-positive strains (27, 29).

In the initial CM05 cfr locus study, the sequence of a 5.8-kbfragment containing the cfr gene was determined, and Southernblotting was utilized to document that this gene was chromo-

Received 29 July 2011 Returned for modification 9 September 2011Accepted 17 October 2011

Published ahead of print 24 October 2011

Address correspondence to Karen J. Shaw, kshaw@triusrx.com.

Copyright © 2012, American Society for Microbiology. All Rights Reserved.

doi:10.1128/AAC.05420-11

332 aac.asm.org 0066-4804/12/$12.00 Antimicrobial Agents and Chemotherapy p. 332–340

on June 8, 2012 by UNIV OF ILLINO

IS AT CHICAGO

http://aac.asm.org/

Downloaded from

somally located (56), rather than being present on the 33.7-kbpCM05 plasmid (GenBank accession no. GQ900387). The cfr genein CM05 was located downstream of the ermB ribosomal methyl-transferase gene, and both genes were shown to be constitutivelycoexpressed in the mlr operon (55). However, the total size, se-quence, and chromosomal locus of the cfr insertion in the CM05chromosome were unknown, as was the stability of cfr in thisstrain. Here we identify the region surrounding the cfr gene, elu-cidate the chromosomal site of insertion, and evaluate the stabilityof the cfr locus in MRSA CM05.

(Portions of this work were presented at the 49th InterscienceConference on Antimicrobial Agents and Chemotherapy [36a],San Francisco, CA, 12 to 15 September 2009.)

MATERIALS AND METHODSBacterial strains and culture conditions. S. aureus strains CM05,CM05!, and ATCC 29213 were cultured aerobically at 37°C on cation-adjusted Mueller-Hinton II agar (MHA; Becton Dickinson, FranklinLakes, NJ) or in Mueller-Hinton II broth (MHB). CFU were enumeratedby serially diluting bacteria in phosphate-buffered saline (PBS) and plat-ing on MHA.

Antimicrobial agents. Solutions of TR-700 (tedizolid, formerlyknown as torezolid; Trius Therapeutics, Inc., San Diego, CA), LZD(ChemPacific Corp., Baltimore, MD), vancomycin (VAN; Sigma-AldrichCorp., St. Louis, MO), chloramphenicol (CHL; Sigma), erythromycin(ERY; Sigma), and tiamulin (TIA; Wako Pure Chemical Industries, Ltd.,Richmond, VA) were prepared fresh prior to use in MIC assays or selectivemedia.

MIC testing. MIC assays were performed via broth microdilution inaccordance with CLSI guidelines (15), and values were determined visu-ally through detection with alamarBlue (Invitrogen Corp., Carlsbad, CA)as described previously (3). MIC assays were repeated at least 3 times,independently, with identical results.

PCR and genome sequencing. Sequencing of the CM05 genome wasperformed by pyrosequencing at the University of Illinois at Urbana-Champaign (49). Single-primer PCR (25) was used to elucidate the chro-mosomal sequences upstream and downstream of cfr. Targeted PCR(Platinum PCR SuperMix High Fidelity; Invitrogen) and primer exten-sion sequencing (Retrogen, Inc., San Diego, CA) were used for elucidationof the chromosomal integration site of cfr and for the assembly of somecontigs resulting from problems with automated assembly spanning re-peated sequences. rRNA rrn alleles 1, 3, 4, and 5 were amplified by PCRusing previously described primers (44). Allele 2 was amplified with prim-ers rrn6_RF (reverse complement to rrn6_R [38, 46]) and rrn2_R (44),and allele 6 was amplified using 16S_F (Table 1) and rrn6_R (38, 46).Sequence data were assembled using Vector NTI Advance 11 software(Invitrogen) and were annotated based on homology to sequence datafrom NCBI BLAST analyses (2). A list of primers used to amplify andsequence the cfr insertion region is provided in Table 1.

CM05 population analysis of variation in LZD susceptibility. A mas-ter glycerol stock sample of CM05 was grown to an optical density at 600nm (OD600) of 0.30 in MHB and was then diluted to "6.5 # 102 CFU/ml,and 100-!l aliquots were spread on MHA plates containing different con-centrations of LZD (0.25, 0.5, 1, 2, 4, 8, 16, or 32 !g/ml) or TR-700 (0.25,0.5, 1, or 2 !g/ml) or no drug. Following overnight growth, CFU wereenumerated, and the plating efficiencies of CM05 on the two drugs werecompared.

Serial passage of CM05 under nonselective conditions. A clonal cfr-positive CM05 culture was grown overnight in MHB. The following day,the culture was diluted 1:1,000 in MHB to start the subsequent cycle.Passaging was repeated for 75 days, and the ratio of LZDs to LZDr colonieswas determined every 5th passage by streaking a sample onto MHA, pick-ing 96 colonies, placing them in a 96-well plate with MHB, and thenstamping them into a plate containing MHB plus 4 !g/ml LZD. Once

cfr-positive colonies were under the level of detection by plating, the total-population DNA was isolated from each subsequent passage and was PCRamplified for the presence of the cfr gene using cfr_F/cfr_R (Table 1).

Serial passage of a CM05–CM05" mixture in the presence of LZDand TR-700. The initial CM05 glycerol stock (mixture of CM05 andCM05!) was serially passaged on MHA plates with continuous antibioticgradients (6) containing LZD (0.0 to 8 !g/ml) or TR-700 (0.0 to 4 !g/ml)as described previously (35). A glycerol stock was made from each passagefor subsequent MIC and sequence analyses of individual colonies. Follow-

TABLE 1 Primers used to amplify and sequence the CM05 cfr region

Primer namea Sequence (5= to 3=)b

MW1986_F ATGAATAATGACAAACTTCAG16S_F TTTTATGGAGAGTTTGATCC23S_1201_F GCTGTGGATTGTCCTTTGGACrepS_F ATGAATATCCCTTTTGTTGTAGrepS_700_F CTAATGGTTTTAAGCGGTACrepS_1104_F GTCAGAATGGAAAGAAGATTrepS_R TCAGCCTGTATCATAGCTAAACcfrL_F ATGGATAAGTTTAAAACAAGcfrL_R TTAAGAATCAATTTCTTTTTGCermL_F ATGTTGGTATTCCAAATGCGermL_R TTATTTATCTGCGTAATCACermB_F ATGAACAAAAATATAAAATATTCermB_715_F CTATTATTTAACGGGAGGAAATAAermB_R TTATTTCCTCCCGTTAAATAATAGcfr_F ATGAATTTTAATAATAAAACAAAGcfr_131_R TATCCTCAAATCGACTAATTCcfr_870_F ACGATACAATCCTACAATAAGcfr_R CTATTGGCTATTTTGATAATTACCAistAS_-122_R GTGTTTCTAGAGACCTTCACistAS_F ATGACGGAAGAGATTGTGGistAS_378_F ACTGATTGACCACTTAGistAS_R TTATCCATGTTCAAGCACTCistBS_F ATGAAGGAGTGCTTGAACATGGistBS_110_R TCAATCGCAGGTGTCGTTTCistBS_381_F CAAGGTTTATTTCGCCAGTGistBS_732_F GCTTGAAAATAGCCATTCATAAistBS_R TTATGAATGGCTATTTTCAAGzeta_F ATGGCAAATATAGTCAATTTTACzeta_836_F CCAAAACACCTAAACTTCCAGzeta_R TTAAATACCTGGAAGTTTAGGepsilon_F ATGGCAGTTACGTATGAAAAAACepsilon_R TTAAGCCACTTTCTCTTTATTComega_F GTGATTGTGGGAAATTTAGGomega_R TTAAAGTTTGTCAGGTAAATAdelta_F ATGATACAATATTACTATdelta_R CTATTCTTTTTCGTTTTCTAATTGgamma_F ATGAGTACAGTTATTTTAGCTGgamma_85_R ATCCGTCTTTCTTATCACTCTGgamma_590_F TTTCACTAGGACGTGTTCAAACgamma_1200_F TTATGCTTTAGTCGTTAAAACGgamma_1500_F TGGTAAAACCATTGATGATGgamma_R TTATTTTGAGAATGAAATAAAATCbeta_F ATGGCTAAAATTGGTTATGCbeta_R TTAACTATCCCTCTTTCCCTTGalpha_F ATGAAAAAATTTATTTATCGAGalpha_R CTAAAATTCTCTCGGTTGC23S_1890_R CCACTCCTCTTAACCTTCCAGilvA_F ATGACAGTCAAAACAACAGTTTCa Numbers correspond to the number of base pairs within or upstream/downstream ofthe designated gene.b Start and stop codon sequences are underlined.

CM05 cfr Genetic Environment and Stability

January 2012 Volume 56 Number 1 aac.asm.org 333

on June 8, 2012 by UNIV OF ILLINO

IS AT CHICAGO

http://aac.asm.org/

Downloaded from

ing 30 serial passages, total-population MICs were determined with cellsrecovered from both TR-700 and LZD passage conditions.

Growth curves. Three individual CM05 and CM05! colonies werecultured on MHA plus 4 !g/ml LZD or on MHA alone, respectively. Thefollowing day, multiple colonies from each plate were resuspended inMHB, and levels were adjusted to an OD600 of 0.30. Cultures were thendiluted 1:100 into test tubes containing 3 ml MHB. OD600 values wererecorded every 2 h for 24 h. An unpaired, 2-tailed t test was used tocompare OD600 values for CM05 and CM05! at each time point.

Nucleotide sequence accession number. The cfr-containing15,511-bp insertion region sequence, as well as partial flanking 23S rRNAgene sequences, has been deposited in the NCBI database under GenBankaccession no. JN849634. The deposited sequence extends 5= and 3= fromthe partial repS and istAS genes, respectively, previously deposited forCM05 (GenBank accession no. EF450709) (56).

RESULTSThe 15.5-kb plasmid-like cfr insertion region disrupts 23S rRNAallele 4. Analysis of the CM05 pyrosequencing data did not iden-tify the insertion site of the previously characterized mlr locus, sosingle-primer PCR was used to sequence the flanking region ex-tending outward from the 5= and 3= ends of mlr (repS and istBS,respectively). The resulting sequencing data from the istBS flank-ing sequence identified genes consistent with those of the multi-drug resistance plasmid pSM19035 (ermB, zeta, epsilon, omega,delta, gamma, beta, alpha) (4). Reanalysis of the assembled whole-genome pyrosequencing data revealed a contig containing an al-pha gene sequence flanking a portion of a 23S rRNA allele. Todetermine which 23S allele was disrupted, each of the six rrn rRNAoperons was amplified using template DNA isolated from S. au-reus CM05 in addition to ATCC 29213, which was used as a ref-erence. PCR amplicons of the predicted sizes were obtained for allsix alleles in ATCC 29213 and for five of the six alleles in CM05(allele 4 [rrlD] did not generate a PCR band) (Fig. 1). To confirmthe integration of the cfr plasmid-like region in rrlD, inward-facing PCR primers binding the external flanking genes (i.e., thefirst non-rRNA or tRNA genes) MW1986 and ilvA (derived fromthe S. aureus MW2 genome sequence; GenBank accession no.NC_003923) were paired with outward-facing cfr plasmid regionprimers (MW1986_F/repS_F and ilvA_F/alpha_R, respectively)(Table 1). The resulting amplicons were sequenced, which con-firmed specific 15,511-bp integration in 23S rRNA allele 4 be-

tween nucleotides 1256 and 1257 of the 2,923-bp gene (flankingsequences shown as GTTGAAGCAT on the left side of Fig. 2A andGATCGTAAGG on the right side). A 233-bp deletion was de-tected between 5S rRNA allele 4 and ilvA, resulting in a truncatedilvA gene (data not shown).

The associated cfr insertion region has high similarity to thebroad-host-range plasmid pSM19035. BlastN analysis of the15.5-kb insert demonstrated that nearly 60% of the sequence isidentical to regions of pSM19035 (GenBank accession no.NC_006979), a low-copy-number plasmid with a broad hostrange in Gram-positive bacteria, including Bacillus subtilis, Strep-tococcus pyogenes, and S. aureus (4, 10–13, 18). This high degree ofsimilarity extends from CM05 insert nucleotides 1 to 2218, whichinclude the pSM19035 plasmid origin of replication oriR, the repSgene, involved in the initiation of plasmid replication (22), and atruncated version of the copy number control gene copS, all ofwhich are in the reverse orientation relative to the publishedCM05 cfr gene (Fig. 2A) (56).

As shown in a previous CM05 study (56), the mlr operon lies inthe opposite orientation to repS and encodes resistance to macro-lides, lincosamides, and streptogramin B class antibiotics (due toexpression of the ermB ribosomal methyltransferase modifying23S rRNA base A2058) and to phenicols, lincosamides, oxazolidi-nones, pleuromutilins, streptogramin A, and 16-membered ringmacrolides (due to expression of cfr) (55, 56). The previously de-scribed sequences downstream of cfr included half of the istASgene, which is part of the IS21-558 insertion sequence (56). Single-primer PCR was used to sequence the insertion region distal toistAS and to aid in the assembly of orphan contigs from the pyro-sequenced genome. As predicted from the sequence of IS21-558,istBS was located adjacent to a complete copy of istAS (Fig. 2A).IS21-558 is a novel element, first described in pSCFS6, consistingof the istAS transposase, implicated in replicon fusion, and theistBS helper protein (26). In CM05, this region begins at nucleo-tide 5703 with a 57-bp imperfect terminal inverted repeat (26) andends with a 10-bp truncation of this sequence at nucleotide 8710.The 2,711-bp istAB region (bp 5703 to 8414) in CM05 showed4-bp and 13-bp mismatches with pSCFS6 and pSCFS3, respec-tively (26, 28, 29). The presumptive 1,338-bp istAS gene encodes a445-amino-acid protein identical to IstAS found in pSCFS3.Overlapping istAS by the last 8 nucleotides, the 753-bp istBS geneencodes a 250-amino-acid protein identical to IstBS in pSCFS3with the exception of an R169G substitution. It has been proposedthat the full-length 46-kDa IstA protein promotes simple transpo-sition events via a “cut-and-paste” mechanism involving circular-ization of a single copy of IS21 to translocate DNA from the donorto the target site (50). This process is aided by the 30-kDa IstBhelper protein (50).

The cfr gene and istAB operon are flanked by two large1,546-bp Tn917-derived repeated elements that include the ermBgene as well as the leader sequence ermL (Fig. 2A). This secondcopy of the ermB gene is identical to the ermB gene found in themlr locus and differs slightly from the ermB gene found inpSM19035 from Streptococcus pyogenes (Ser10 in CM05 versusAsn10 in pSM19035) (10). Each of these large repeats is flanked bya 63-bp perfect inverted repeat that begins with the sequence 5=-GGGGTCCC-3=. Outside of this region, the proximal 1,546-bprepeat is also flanked by a larger inverted repeat sequence contain-ing the cfrL leader, which may have resulted from a transposition/recombination event. This large region of duplication is antici-

FIG 1 PCR analysis of rrn alleles in S. aureus CM05. All six rRNA rrn loci (5S,16S, and 23S rRNA genes with flanking sequences) were amplified by PCR forCM05 and ATCC 29213 (positive-control 6-rrn allele reference strain). CM05rrn4 was not amplified, due to the 15.5-kb cfr region insertion in the middle ofthe 23S rRNA gene.

Locke et al.

334 aac.asm.org Antimicrobial Agents and Chemotherapy

on June 8, 2012 by UNIV OF ILLINO

IS AT CHICAGO

http://aac.asm.org/

Downloaded from

pated to be unstable due to the possibilities of crossing over,resulting in deletion of the intervening region containing cfr,istAS, and istBS along with one of the ermB copies.

Near-perfect identity to pSM19035 resumes at nucleotide 8412through the end of the insert at nucleotide 15511. This region ofhomology begins 5= of the ermL leader sequence and ermB, fol-lowed by the plasmid addiction system operon, comprising zetatoxin, epsilon antitoxin, and the omega autorepressor/regulatoron the complementary strand (Fig. 2A). Short 7-bp repeat omegabinding site elements are observed immediately upstream of theomega gene and the adjacent delta gene (parA), suggesting auto-regulation of the addiction system operon as well as the delta gene.The delta gene encodes an ATPase required for chromosomal par-titioning (45). The end of the insert contains genes encoding thegamma topoisomerase, the "/six site-specific recombinase, andthe hypothetical alpha gene product, whose function is unknown(10). The CM05 insert ends with a truncated "/six recombinaserecognition site (Fig. 2 and 3). The immediately flanking regioncontains the 1,667-nucleotide remainder of the 23S rRNA gene,which was precisely split by the insertion of the 15,511-bp ele-ment.

It should be noted that the 15.5-kb insertion region does notshare significant sequence similarity with the previously charac-terized 33.7-kb plasmid present in CM05 (GenBank accession no.GQ900387). This suggests that the acquisition and subsequentintegration of the cfr-carrying pSM19035-like plasmid occurredindependently of the endogenous plasmid in this strain.

Insertion of the cfr region into the CM05 chromosome oc-curs at a #/six site I recognition site. The "/six recombinase isknown to have a specific recognition sequence (six site) con-sisting of a 90-bp element that contains two regions (sites I andII) (Fig. 3A) (1). Recombination (Fig. 3B) is catalyzed by the"-recombinase and requires two 90-bp "/six sites in either a director an inverted orientation (9, 48). Two of these sites are present inpSM19035, a 29-kb broad-host-range plasmid (1, 10). DNA se-quence analysis suggests that the configuration observed in CM05is likely due to insertion of a pSM19035-like plasmid into thegenome, with the recombination event occurring precisely withinthe site I recognition region (5=-GTATAC-3=), which has previ-ously been shown to be critical for strand exchange (Fig. 3A) (9).This recombination event would have resulted in the location of aportion of site I at one end of the CM05 insert, with the remainderof the recombination site located at the other end of the 15,511-bpinsert (Fig. 3C). The 23S rRNA insertion site shares little sequencehomology with the "/six recognition site aside from the nucleo-tides at the site of insertion (data not shown).

LZD plating discrepancies reveal a subpopulation of cfr-negative CM05 (CM05"). A culture derived from a master glyc-erol stock of CM05 was plated onto MHA plates containing dif-ferent concentrations of LZD (0.25 to 32 !g/ml) or TR-700 (0.25to 2 !g/ml), an expanded-spectrum oxazolidinone whose potencyon the MIC level has been shown previously to be unaffected byCfr methylation (33). Although the MIC of LZD was 8 !g/ml,$30% of the CFU plated grew on plates containing 2 or 4 !g/ml of

FIG 2 CM05 cfr plasmid insertion region and recombination/excision scheme. (A) The 15.5-kb cfr region is inserted between bases 1256 and 1257 in 23S rRNAallele 4 (rrn4 rRNA operon) in CM05. The insertion in allele 4 (flanked by MW1986 and ilvA genes) is annotated based on the 6-rrn-operon S. aureus MW2genomic sequence (GenBank accession no. NC_003923) and is shown in the reverse complement orientation. The previously characterized mlr and IS21–558regions (56) are flanked by regions with high similarity to the broad-host-range multidrug resistance plasmid pSM19035 (18). (B through D) The instability ofthe region due to recombination between two identical 1,546-bp regions (light green bars) containing ermB genes and leader sequences (ermL) (B) leads to theexcision of 5,987-bp minicircles (C) and the simultaneous generation of the LZDs strain CM05!, lacking the istAB genes, cfr, and one copy of ermB (D).

CM05 cfr Genetic Environment and Stability

January 2012 Volume 56 Number 1 aac.asm.org 335

on June 8, 2012 by UNIV OF ILLINO

IS AT CHICAGO

http://aac.asm.org/

Downloaded from

LZD (Table 2). In contrast, TR-700 produced growth inhibitionconsistent with a MIC of 0.5 !g/ml (Table 2).

The differences in plating efficiencies were explored further.Analysis of individual colonies from the glycerol stock plating re-vealed a mixture of LZDr and LZDs clones. To determine whetherthe loss of the cfr gene was responsible for the differences in phe-notype, 10 colonies were chosen for further study. MIC analysisrevealed two distinct populations for which the LZD MIC waseither 8 !g/ml or 2 !g/ml (Table 3). PCR amplification (cfr_F/cfr_R) generated the predicted "1-kb cfr PCR product from chro-mosomal DNA isolated from all 5 of the LZDr colonies; however,no product was obtained from the 5 CM05! LZDs colonies (datanot shown).

Strain CM05" arises from recombination between duplicateermB regions. Comparative analysis of cfr region insertion se-quences between individual colonies of strains CM05 and CM05!revealed evidence of a recombination event between the two1,546-bp repeated ermB (Tn917) elements. In strain CM05, theresulting precise excision of one of the two Tn917 elements, onecopy of ermB, and the intervening cfr, istAS, and istBS genes would

lead to the generation of 5,987-bp minicircles (Fig. 2B and C).PCR was performed on a mixed CM05–CM05! total-DNA sam-ple to amplify the minicircles using outward-facing primer sets foristAS and istBS. The PCR bands generated matched the predictedsizes, and follow-up sequencing confirmed that they were indeedspecific to the istAS-istBS-ermB-cfr minicircle sequence (data notshown) (Fig. 4). The primer sets additionally generated smallerbands of "2 kb, which were sequenced and confirmed to containistAS and istBS in circular form without intervening cfr or ermBgenes, suggesting that this region can be excised independentlyfrom the greater 15-kb insert to form 2,454-bp minicircles (Fig. 4).The bridging sequence in these minicircles (istBS downstream se-quence, 5=-GCGAAATACA-3=; istAS upstream sequence, 5=-TGTGTATATG-3=) occurs at the imperfect terminal inverted repeatspreviously identified in the nearly identical 2,446-bp IS21-558 in-sertion sequence from pSCFS3 (28).

Nonselective serial passage of CM05 results in a CM05" pop-ulation of cells possessing the ermB-cfr-istAS-istBS excision/re-combination event. To explore the origin and to attempt to reca-pitulate the recombination event responsible for generating theCM05! strain, nonselective serial passage of a clonal cfr-positiveCM05 culture was performed. As a proxy to measure the loss of cfr,the ratio of LZDr to LZDs colonies was monitored every 5th pas-sage in MHB through replica plating of colonies in MHB aloneand MHB plus 4 !g/ml LZD. The data in Fig. 5 show that thepercentage of LZDr cfr-positive colonies in the population de-creased as the serial passage number increased. By passage 30, noLZDr colonies were detected through plating ($1% of the totalcell population) (Fig. 5), and the culture was PCR negative for cfrafter 75 passages. LZDs colonies from P75 were confirmed to pos-sess the same deletion of cfr-istAS-istBS-ermB as was documentedin the CM05! strain in the initial mixed population (Fig. 2D).

CM05" possesses a fitness advantage over CM05. To investi-gate whether the shift of the CM05 culture to a CM05! culturethrough nonselective passage was due to a high-frequency recom-

FIG 3 Model of the insertion of the 15.5-kb cfr-containing sequence into the CM05 chromosome. (A) The "/six recombinase recognition site of plasmidpSM19035, with the essential AT residues highlighted (9). Sites I and II of the recognition site are highlighted in green and gray, respectively. (B) Heterologousrecombination event between the 23S rRNA gene and "/six recombinase site I of the cfr-carrying pSM19035-like plasmid. (C) The insertion of the 15.5-kbcfr-containing sequence into the 23S rRNA gene results in the location of a portion of site I at each end of the insert.

TABLE 2 Plating efficiency of a mixed CM05–CM05! glycerol stock ona medium containing LZD or TR-700

Drug concn (!g/ml)

CFU on a medium containing:

LZD TR-700

0 69 690.25 64 620.5 61 01 68 02 14 04 21 NDa

8 0 ND16 0 ND32 0 NDa ND, not determined.

Locke et al.

336 aac.asm.org Antimicrobial Agents and Chemotherapy

on June 8, 2012 by UNIV OF ILLINO

IS AT CHICAGO

http://aac.asm.org/

Downloaded from

bination event or to a fitness advantage of CM05! cells over CM05cells, we compared the growth rates of clonal CM05 and CM05!strains in MHB (Fig. 6). These two strains had similar log-phasegrowth rates and attained the same stationary-phase OD600; how-ever, CM05 exhibited a longer lag phase than CM05!. The fitnessadvantage of CM05! over CM05 likely accounts for the domi-nance of CM05! under nonselective passage.

Passage of a mixed CM05–CM05" population in the pres-ence of LZD selects for cfr maintenance, whereas passage in thepresence of TR-700 does not. To further investigate the fitnessadvantage of CM05! and explore the drug-related selection andmaintenance of cfr-positive strains, we performed serial passageexperiments with mixed populations of CM05 and CM05! usingin vitro passage in the presence of TR-700 or LZD. Because Cfrconfers resistance to LZD, the competitive advantage of CM05!was not seen during passage of a mixed CM05–CM05! culture onLZD gradient plates. Thus, the serial-passage population in LZDresulted in cfr-positive colonies as the predominant clonal type(Fig. 7). Serial passage in the presence of TR-700, however, re-

sulted in a homogeneous CM05! population following 14 serialpassages, demonstrating the competitive growth advantage ofCM05! even under potentially selective conditions. Previousstructure-activity relationship (SAR) studies have shown that thereduced steric hindrance of the A-ring C-5 hydroxymethyl of TR-700, in contrast to the bulkier hydroxymethyl group present onLZD (and other oxazolidinones such as radezolid), accounts forthe retention of potency against cfr-positive strains (33) and theconcomitant lack of selective pressure to retain cfr. Despite reportsof clinical cfr strains possessing multiple classes of 50S mutation-based resistance mechanisms (5, 14, 36), CM05 under selectionwith LZD for 30 passages did not generate any cells with enhancedLZD resistance above the starting 8-!g/ml level. Whether this is astrain-specific phenomenon or whether it is related to the inser-tion of the cfr region into the 23S rRNA gene is unknown. Thesame was true for TR-700; however, the 16-fold-lower mutationfrequency of S. aureus with this compound than with LZD couldexplain the lack of resistance selection, as observed previously withATCC 29213 (35).

DISCUSSIONAll cfr-positive clinical isolates characterized to date, with the ex-ception of CM05, have possessed a plasmid-borne copy of thegene. Here we have built on the earlier hypothesis that despite thepresent chromosomal locus of cfr in CM05, its origin is plasmidbased (56). DNA sequence analysis suggests that the complex ge-netic environment of the cfr gene in CM05 likely arose from aseries of transposition and recombination events involving a plas-mid with a high degree of sequence similarity to the broad-host-range plasmid pSM19035 (4). In addition, plasmids with toxin/

TABLE 3 Characteristics of strains CM05 and CM05!

Strain

Presence or absencea

of: MIC (!g/ml)b of:

cfr ermB LZD TR-700 ERY CHL TIA VAN

CM05 % %% 8 0.5 &256 64 256 1CM05! ' % 2 0.5 &256 8 0.5 1a %, 1 copy present; %%, 2 copies present; ', absent.b MICs were determined by broth microdilution according to the CLSI guidelines for linezolid (LZD), TR-700, erythromycin (ERY), chloramphenicol (CHL), tiamulin (TIA), andvancomycin (VAN).

FIG 4 PCR amplification of cfr-containing minicircles from CM05! follow-ing recombination between ermB genes. Outward-facing primer sets encom-passing portions of istAS and istBS generated the expected bands for the "6-kbermB-cfr-istAS-istBS minicircles (larger bands). Additionally, these primersspecifically amplified 2.54-kb minicircles containing only the istAS and istBSgenes of the IS21–558 insertional element (smaller bands).

FIG 5 Serial passage of CM05 under nonselective conditions. A clonal CM05cfr-positive culture was serially passaged in MHB 75 times (1:1,000 dilutions).Every 5th passage, the ratio of cfr-positive to cfr-negative colonies was deter-mined. At passage 30, LZDr CM05 cells were no longer detected in the sam-pling ($1% of the population). By passage 75, the culture was PCR negativefor the cfr gene, suggesting a complete conversion to CM05! LZDs cells.

CM05 cfr Genetic Environment and Stability

January 2012 Volume 56 Number 1 aac.asm.org 337

on June 8, 2012 by UNIV OF ILLINO

IS AT CHICAGO

http://aac.asm.org/

Downloaded from

antitoxin maintenance systems highly similar to that of pSM19035have been seen in Enterococcus faecium and Enterococcus faecalis(32, 54). The instability of cfr within this greater chromosomalinsertion locus in the absence of selective pressure selects for re-combination events whereby the gene is excised and may poten-tially be transmitted horizontally. Furthermore, the association ofthe cfr gene with pSM19035 suggests that other strains containingfunctional cfr-positive versions of the free plasmid may exist, pro-viding another means of horizontal spread of this resistancemechanism into many pathogenic species.

The remnants of a recombination event within S. aureus CM05disrupted a 23S rRNA gene and the pSM19035 "/six site and waspossibly mediated by the "/six recombinase protein. These se-quences are found at the chromosomal junctions of the 15.5-kbinsertion. The "/six recombinase catalyzes resolution or inversionbetween two tandem or inversely oriented six sites, respectively, orintegration if the two sites are on different DNA molecules (8).The 23S insertion site is only marginally reminiscent of the known"/six recombinase recognition site, suggesting that the recombi-nase may have low selectivity or that some other mechanism wasinvolved in the genomic insertion. Insertion of drug resistanceplasmids into rRNA genes has not been reported, and thus, furtherwork is needed to determine what selective pressure or transcrip-tional advantages may result in these integration events. The factthat the plasmid bearing the cfr gene has been inserted into thevery genetic material that encodes the target of its modificationwarrants further investigation.

Earlier reports of CM05 described the Tn917 transposon seg-ment, which contains the ermB promoter, the ermB leader, andthe ermB gene followed by the cfr leader and cfr gene, which areconstitutively expressed (55, 56). Between the ermB gene and cfrleader is a possible Tn3-like resolution site (5=-CAGCGAAATTTTTAAATCTAT-3=) (47) followed by a 38-bp Tn917 terminal in-verted repeat. These sites are both within a 63-bp repeat that isfound in an inverted orientation flanking both copies of the ermBgene. Resolution between the Tn917 repeats, at the potential ressite, or anywhere within the duplicated ermB region would resultin the deletion of the intervening sequence containing cfr, istAS,and istBS. PCR and sequencing document the presence ofminicircles containing the intervening sequence within the cell.

The redundancy of rrn alleles opens the possibility that the

copy number of mlr can be increased by gene conversion. Associ-ation of mlr with transposon and plasmid elements may furtherincrease the mobility of mlr and its potential for horizontal trans-fer. However, the spontaneous deletion of cfr and 3= flanking se-quences suggests that either the gene or, more likely, the associatedgenetic elements are deleterious in the absence of antibiotic selec-tion (31). A recent study reported that a strain expressing an activeCfr protein had a growth rate equivalent to that of an isogenicstrain expressing a nonfunctional Cys119Ala mutant Cfr protein,thus demonstrating that Cfr methylation imposes a negligible fit-ness cost (31), which is supported by the existence of clinical cfrstrains possessing co-occurring 23S rRNA (14), L3 (36), or L4 (5)mutations. The lower fitness of CM05 than of CM05! in MHBmay be related to the coexpression of cfr and the ermB methyl-transferase genes. Methylation of two key peptidyl transferase cen-ter bases has been demonstrated to have detrimental effects thatmethylation of either one on its own would not have (31). Alter-natively, the structural or functional role of any components of thesurrounding genetic material (i.e., istAS, istBS, etc.) or the pres-ence of two ermB methyltransferase genes could also have a neg-ative impact on fitness and should be the subject of further study.

SAR and modeling work have revealed that the steric clashbetween the A2503 carbon 8 methyl group with C-5 acetamidesubstituents on oxazolidinones, such as LZD and radezolid (RX-1741), leads to increases in the MIC (33). Consistent with thishypothesis, the MIC of TR-700, which possesses a smaller hy-droxymethyl C-5 group, was the same for strains CM05 andCM05!. This argument is further supported by in vitro serial pas-sage of a mixed CM05–CM05! population in the presence ofTR-700, which resulted in a full conversion to the cfr-negativeCM05! strain. Presumably, similar selective pressures would existin in vivo systems, and drugs such as TR-700, or others not fallingwithin the PhLOPSA resistance phenotype, would not select forthe maintenance of cfr.

CM05 was the first reported clinical isolate carrying the cfrmethyltransferase gene. Here we have further documented thechromosomal locus of cfr, describing its novel association with apSM19035-like plasmid and its novel integration into a 23S rRNAallele. Although cfr is not contained within a free plasmid inCM05, it still exists in a highly mobile/unstable region, and thepresence of recombination-generated minicircles suggests that cfr

FIG 6 Growth profiles of strains CM05 and CM05!. The growth rates ofCM05 and CM05! were compared in MHB. The average OD600 values of threeindependent cultures for each strain are shown for each time point, along withthe standard deviations. The OD600 values for the cultures of the two strainswere significantly different from 2 to 16 h of growth (!, P # 0.05).

FIG 7 MIC profiles of a mixed CM05–CM05! culture passaged on gradientplates containing TR-700 or LZD. MIC testing of total-cell populations recov-ered from the edges of visible growth inhibition zones on antibiotic gradientplates showed no change in the MIC of TR-700 or LZD over 30 serial passages.LZD passage resulted in the immediate dominance of the cfr-positive CM05population by passage 1, whereas selection with TR-700 resulted in the shift toa homogeneous cfr-negative CM05! population by passage 14.

Locke et al.

338 aac.asm.org Antimicrobial Agents and Chemotherapy

on June 8, 2012 by UNIV OF ILLINO

IS AT CHICAGO

http://aac.asm.org/

Downloaded from

may still be horizontally transferable from this strain. The abilityof cfr to confer resistance to multiple classes of peptidyl transferasecenter-targeted antibiotics underscores the need for diligence inthe selection of therapeutic strategies in areas with precedents forcfr strain outbreaks.

ACKNOWLEDGMENTSThe CM05 strain was a gift from John Quinn. We acknowledge AlvaroHernandez and Jyothi Thimmapuram of the University of Illinois W. M.Keck Center for Comparative and Functional Genomics for assistancewith the sequencing and assembly of the CM05 genome. Finally, we thankJeffrey Hsiao for technical assistance.

REFERENCES1. Alonso JC, Weise F, Rojo F. 1995. The Bacillus subtilis histone-like

protein Hbsu is required for DNA resolution and DNA inversion medi-ated by the beta recombinase of plasmid pSM19035. J. Biol. Chem. 270:2938 –2945.

2. Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ. 1990. Basic localalignment search tool. J. Mol. Biol. 215:403– 410.

3. Baker CN, Tenover FC. 1996. Evaluation of Alamar colorimetric brothmicrodilution susceptibility testing method for staphylococci and entero-cocci. J. Clin. Microbiol. 34:2654 –2659.

4. Behnke D, Golubkov VI, Malke H, Boitsov AS, Totolian AA. 1979.Restriction endonuclease analysis of group A streptococcal plasmids de-termining resistance to macrolides, lincosamides and streptogramin-Bantibiotics. FEMS Microbiol. Lett. 6:5–9.

5. Bonilla H, et al. 2010. Multicity outbreak of linezolid-resistant Staphylo-coccus epidermidis associated with clonal spread of a cfr-containing strain.Clin. Infect. Dis. 51:796 – 800.

6. Bryson V, Szybalski W. 1952. Microbial selection. Science 116:45–51.7. Campanile F, et al. 2009. Abstr. 19th Eur. Cong. Clin. Microbiol. Infect.

Dis., abstr P929.8. Canosa I, Lopez G, Rojo F, Boocock MR, Alonso JC. 2003. Synapsis and

strand exchange in the resolution and DNA inversion reactions catalysedby the beta recombinase. Nucleic Acids Res. 31:1038 –1044.

9. Canosa I, Rojo F, Alonso JC. 1996. Site-specific recombination by thebeta protein from the streptococcal plasmid pSM19035: minimal recom-bination sequences and crossing over site. Nucleic Acids Res. 24:2712–2717.

10. Ceglowski P, Alonso JC. 1994. Gene organization of the Streptococcuspyogenes plasmid pDB101: sequence analysis of the orf$-copS region. Gene145:33–39.

11. Ceglowski P, Boitsov A, Chai S, Alonso JC. 1993. Analysis of the stabi-lization system of pSM19035-derived plasmid pBT233 in Bacillus subtilis.Gene 136:1–12.

12. Ceglowski P, Boitsov A, Karamyan N, Chai S, Alonso JC. 1993. Char-acterization of the effectors required for stable inheritance of Streptococcuspyogenes pSM19035-derived plasmids in Bacillus subtilis. Mol. Gen. Genet.241:579 –585.

13. Ceglowski P, Lurz R, Alonso JC. 1993. Functional analysis of pSM19035-derived replicons in Bacillus subtilis. FEMS Microbiol. Lett. 109:145–150.

14. Cercenado E, Marin M, Insa R, Bouza E. 2010. Abstr. 20th Eur. Cong.Clin. Microbiol. Infect. Dis., abstr P979.

15. CLSI. 2006. Methods for dilution antimicrobial susceptibility tests forbacteria that grow aerobically; approved standard, CLSI documentM7–A7, 7th ed, vol 26, no. 2. CLSI, Wayne, PA.

16. Dai L, et al. 2010. First report of the multidrug resistance gene cfr and thephenicol resistance gene fexA in a Bacillus strain from swine feces. Anti-microb. Agents Chemother. 54:3953–3955.

17. de la Torre M, et al. 2008. Abstr. 48th Intersci. Conf. Antimicrob. AgentsChemother., abstr C2-1835a.

18. Dixon JM, Lipinski AE. 1972. Resistance of group A beta-hemolyticstreptococci to lincomycin and erythromycin. Antimicrob. Agents Che-mother. 1:333–339.

19. Fritsche TR, Castanheira M, Mendes RE, Jones RN, Deshpande LM.2008. Abstr. 108th Gen. Meet. Am. Soc. Microbiol., abstr A-047.

20. Giessing AM, et al. 2009. Identification of 8-methyladenosine as themodification catalyzed by the radical SAM methyltransferase Cfr that con-fers antibiotic resistance in bacteria. RNA 15:327–336.

21. Gonzales RD, et al. 2001. Infections due to vancomycin-resistant Entero-coccus faecium resistant to linezolid. Lancet 357:1179.

22. Hashiba H, Takiguchi R, Joho K, Aoyama K, Hirota T. 1993. Identifi-cation of the replication region of Streptococcus thermophilus no. 29 plas-mid pST1. Biosci. Biotechnol. Biochem. 57:1646 –1649.

23. Jones RN, et al. 2009. Zyvox Annual Appraisal of Potency and Spectrumprogram: linezolid surveillance program results for 2008. Diagn. Micro-biol. Infect. Dis. 65:404 – 413.

24. Jones RN, Ross JE, Castanheira M, Mendes RE. 2008. United Statesresistance surveillance results for linezolid (LEADER Program for 2007).Diagn. Microbiol. Infect. Dis. 62:416 – 426.

25. Karlyshev AV, Pallen MJ, Wren BW. 2000. Single-primer PCR proce-dure for rapid identification of transposon insertion sites. Biotechniques28:1078, 1080, 1082.

26. Kehrenberg C, Aarestrup FM, Schwarz S. 2007. IS21-558 insertion se-quences are involved in the mobility of the multiresistance gene cfr. Anti-microb. Agents Chemother. 51:483– 487.

27. Kehrenberg C, Cuny C, Strommenger B, Schwarz S, Witte W. 2009.Methicillin-resistant and -susceptible Staphylococcus aureus strains ofclonal lineages ST398 and ST9 from swine carry the multidrug resistancegene cfr. Antimicrob. Agents Chemother. 53:779 –781.

28. Kehrenberg C, Schwarz S. 2006. Distribution of florfenicol resistancegenes fexA and cfr among chloramphenicol-resistant Staphylococcus iso-lates. Antimicrob. Agents Chemother. 50:1156 –1163.

29. Kehrenberg C, Schwarz S, Jacobsen L, Hansen LH, Vester B. 2005. Anew mechanism for chloramphenicol, florfenicol and clindamycinresistance: methylation of 23S ribosomal RNA at A2503. Mol. Microbiol.57:1064 –1073.

30. Kloss P, Xiong L, Shinabarger DL, Mankin AS. 1999. Resistance muta-tions in 23S rRNA identify the site of action of the protein synthesis in-hibitor linezolid in the ribosomal peptidyl transferase center. J. Mol. Biol.294:93–101.

31. LaMarre JM, Locke JB, Shaw KJ, Mankin AS. 2011. Low fitness cost ofthe multidrug resistance gene cfr. Antimicrob. Agents Chemother. 55:3714 –3719.

32. Lim SK, Tanimoto K, Tomita H, Ike Y. 2006. Pheromone-responsiveconjugative vancomycin resistance plasmids in Enterococcus faecalis iso-lates from humans and chicken feces. Appl. Environ. Microbiol. 72:6544 – 6553.

33. Locke JB, et al. 2010. Structure-activity relationships of diverse oxazo-lidinones for linezolid-resistant Staphylococcus aureus strains possessingthe cfr methyltransferase gene or ribosomal mutations. Antimicrob.Agents Chemother. 54:5337–5343.

34. Locke JB, Hilgers M, Shaw KJ. 2009. Mutations in ribosomal protein L3are associated with oxazolidinone resistance in staphylococci of clinicalorigin. Antimicrob. Agents Chemother. 53:5275–5278.

35. Locke JB, Hilgers M, Shaw KJ. 2009. Novel ribosomal mutations inStaphylococcus aureus strains identified through selection with the oxazo-lidinones linezolid and torezolid (TR-700). Antimicrob. Agents Che-mother. 53:5265–5274.

36. Locke JB, et al. 2010. Elevated linezolid resistance in clinical cfr-positiveStaphylococcus aureus isolates is associated with co-occurring mutations inribosomal protein L3. Antimicrob. Agents Chemother. 54:5352–5355.

36a.Locke JB, Mankin AS, LaMarre J, Brown-Driver V, Shaw KJ. Abstr. 49thIntersci. Conf. Antimicrob. Agents Chemother., poster C1-1364b.

37. Long KS, Poehlsgaard J, Kehrenberg C, Schwarz S, Vester B. 2006. Thecfr rRNA methyltransferase confers resistance to phenicols, lincosamides,oxazolidinones, pleuromutilins, and streptogramin A antibiotics. Antimi-crob. Agents Chemother. 50:2500 –2505.

38. Meka VG, et al. 2004. Linezolid resistance in sequential Staphylococcusaureus isolates associated with a T2500A mutation in the 23S rRNA geneand loss of a single copy of rRNA. J. Infect. Dis. 190:311–317.

39. Mendes RE. 2009. Abstr. 19th Eur. Cong. Clin. Microbiol. Infect. Dis.,abstr P928.

40. Mendes RE, et al. 2010. First report of staphylococcal clinical isolates inMexico with linezolid resistance caused by cfr: evidence of in vivo cfr mo-bilization. J. Clin. Microbiol. 48:3041–3043.

41. Mendes RE, et al. 2008. First report of cfr-mediated resistance to linezolidin human staphylococcal clinical isolates recovered in the United States.Antimicrob. Agents Chemother. 52:2244 –2246.

42. Mendes RE, et al. 2010. Assessment of linezolid resistance mechanismsamong Staphylococcus epidermidis causing bacteraemia in Rome, Italy. J.Antimicrob. Chemother. 65:2329 –2335.

CM05 cfr Genetic Environment and Stability

January 2012 Volume 56 Number 1 aac.asm.org 339

on June 8, 2012 by UNIV OF ILLINO

IS AT CHICAGO

http://aac.asm.org/

Downloaded from

43. Morales G, et al. 2010. Resistance to linezolid is mediated by the cfr genein the first report of an outbreak of linezolid-resistant Staphylococcus au-reus. Clin. Infect. Dis. 50:821– 825.

44. Pillai SK, et al. 2002. Linezolid resistance in Staphylococcus aureus: char-acterization and stability of resistant phenotype. J. Infect. Dis. 186:1603–1607.

45. Pratto F, et al. 2008. Streptococcus pyogenes pSM19035 requires dynamicassembly of ATP-bound ParA and ParB on parS DNA during plasmidsegregation. Nucleic Acids Res. 36:3676 –3689.

46. Prunier AL, Malbruny B, Tande D, Picard B, Leclercq R. 2002. Clinicalisolates of Staphylococcus aureus with ribosomal mutations conferring re-sistance to macrolides. Antimicrob. Agents Chemother. 46:3054 –3056.

47. Rojo F, Alonso JC. 1995. The beta recombinase of plasmid pSM19035binds to two adjacent sites, making different contacts at each of them.Nucleic Acids Res. 23:3181–3188.

48. Rojo F, Alonso JC. 1994. A novel site-specific recombinase encoded bythe Streptococcus pyogenes plasmid pSM19035. J. Mol. Biol. 238:159 –172.

49. Ronaghi M, Uhlen M, Nyren P. 1998. A sequencing method based onreal-time pyrophosphate. Science 281:363–365.

50. Schmid S, Seitz T, Haas D. 1998. Cointegrase, a naturally occurring,truncated form of IS21 transposase, catalyzes replicon fusion rather thansimple insertion of IS21. J. Mol. Biol. 282:571–583.

51. Schwarz S, Werckenthin C, Kehrenberg C. 2000. Identification of aplasmid-borne chloramphenicol-florfenicol resistance gene in Staphylo-coccus sciuri. Antimicrob. Agents Chemother. 44:2530 –2533.

52. Shinabarger DL, et al. 1997. Mechanism of action of oxazolidinones:effects of linezolid and eperezolid on translation reactions. Antimicrob.Agents Chemother. 41:2132–2136.

53. Shore AC, et al. 2010. Identification and characterization of the multi-drug resistance gene cfr in a Panton-Valentine leukocidin-positive se-quence type 8 methicillin-resistant Staphylococcus aureus IVa (USA300)isolate. Antimicrob. Agents Chemother. 54:4978 – 4984.

54. Sletvold H, et al. 2007. Comparative DNA analysis of two vanA plasmidsfrom Enterococcus faecium strains isolated from poultry and a poultryfarmer in Norway. Antimicrob. Agents Chemother. 51:736 –739.

55. Smith LK, Mankin AS. 2008. Transcriptional and translational control ofthe mlr operon, which confers resistance to seven classes of protein syn-thesis inhibitors. Antimicrob. Agents Chemother. 52:1703–1712.

56. Toh SM, et al. 2007. Acquisition of a natural resistance gene renders aclinical strain of methicillin-resistant Staphylococcus aureus resistant to thesynthetic antibiotic linezolid. Mol. Microbiol. 64:1506 –1514.

57. Tsiodras S, et al. 2001. Linezolid resistance in a clinical isolate of Staph-ylococcus aureus. Lancet 38:207–208.

58. Wolter N, et al. 2005. Novel mechanism of resistance to oxazolidinones,macrolides, and chloramphenicol in ribosomal protein L4 of the pneumo-coccus. Antimicrob. Agents Chemother. 49:3554 –3557.

59. Xiong L, et al. 2000. Oxazolidinone resistance mutations in 23S rRNA ofEscherichia coli reveal the central region of domain V as the primary site ofdrug action. J. Bacteriol. 182:5325–5331.

Locke et al.

340 aac.asm.org Antimicrobial Agents and Chemotherapy

on June 8, 2012 by UNIV OF ILLINO

IS AT CHICAGO

http://aac.asm.org/

Downloaded from