Proc. Nati. Acad. Sci. USAVol. 88, pp. 6072-6076, July 1991Genetics

Trans-acting transposase mutant from Tn5ALISON DELONG* AND MICHAEL SYVANENtt*Department of Biology, Yale University, New Haven, CT 06511; and tDepartment of Medical Microbiology and Immunology, School of Medicine, Universityof California, Davis, CA 95616

Communicated by Jonathan Beckwith, March 15, 1991 (received for review December 9, 1990)

ABSTRACT Transposition of TnS and of its componentinsertion sequence ISSOR is regulated through the action of twoproteins it encodes: a cis-acting transposase, Tnp, and atrans-acting inhibitor of transposition, Inh. The mechanism ofthe cis-acting Tnp and the relevance of inhibition to cis actionhave been addressed in the current study. A specific colonymorphology assay for transposition of Tn5 was shown to besensitive to Inh produced in trans and was used to screen formutants in Inh and/or Tnp with altered regulation. A domi-nant mutant in ISSOR that promotes transposition in trans wasisolated and characterized. The mutant (449F) carries a Leu-*Phe mutation at position 449 in Tnp. This mutation reduces thefrequency of TnS or I5SOR transposition in cis but allowsTnp-449F to act as efficiently in trans as it does in cis. Tnp-449Fis sensitive to inhibition and, furthermore; Inh449F is acompetent inhibitor in trans. These results show that Tnp-449Fis a trans-acting transposase, unlike wild-type Tnp, which iscis-acting.

Transposons have various regulatory mechanisms that limittheir movement. With the transposon Tn5, regulation occursprimarily at the level of the transposition reaction (1, 2). Tn5is a compound transposon, comprising two nearly identical,inversely oriented repeat sequences, ISSOR and IS50L, anda central or unique region containing antibiotic-resistancegenes (3, 4). ISSOR, which is required for transposition ofTnS, codes for two proteins, Tnp and Inh, which are 476 and421 amino acids in length, respectively (Fig. 1; ref. 6). Thefrequency of TnS transposition is tightly regulated at severallevels. First, Inh directly inhibits transposition; second, Tnpexhibits a 100-fold preference for IRS elements located in cis(1, 2). Furthermore, Tnp is produced in very low amounts,due to a weak promoter, a poor ribosome binding site, and aposttranscriptional block barring expression of messagesinitiated at promoters external to IS50 (6). The tnp promoteris active only when hemimethylated (7, 8), as is the inner IRSof IS50 (7).

Inh is a diffusible (trans-acting) inhibitor of transpositionthat acts to block the transposition reaction directly, ratherthan by affecting expression of the tnp gene. In wild-typecells carrying ISSOR or TnS, Inh is more abundant than Tnp.Because Inh and Tnp form a protein complex (9), it is thoughtthat Inh may inhibit Tnp by direct binding.

In general, the mechanism causing site-specific DNA-binding proteins to act preferentially near the gene fromwhich they were synthesized is not known. In the case ofTnpthe ability of Inh to act efficiently in trans, especially if Inhinactivates Tnp by direct binding, may be responsible for Tnpacting near its site of synthesis (23). According to this model,the regulatory function of Inh itself is responsible for thecis-acting phenotype of Tnp. Here we report the isolation ofa trans-acting mutant ofTn5 transposase. Our results indicatethat the preference of Tnp for action in cis is independent ofits sensitivity to Inh.

MATERIALS AND METHODSMedia. Antibiotics were used in TYE agar (10) at the

following concentrations: kanamycin, 30 ,ug/ml; tetracy-cline, 12 ,ug/ml; carbenicillin, 125 ,ug/ml; chloramphenicol,20-50 jig/ml.

Strains. The Escherichia coli strains, phages, and plasmidsare listed in Table 1. To construct A421-449F, the 449Fmutation was subcloned from Tn5-449F into A421 by usingthe central 4.6-kilobase Not I fragment of pTG20-449F.AADL87 and AADL87-449F were constructed by crossing thehis-flanked transposons on pADL87 and pADL87-449F (seebelow) into the his DNA carried on ANK1038.

Plasmid Constructions. pTG20 is pACYC184::TnS and isthe parent of all pACYC184::TnS and pACYC184::IS50(pADL53) plasmids used. pADL53-34am contains an Xba Ilinker at the Hpa I site at position 188 in ISSOR. In pADL53-561 the 400-bp Hpa I-Not I fragment from pRZ1037 (17) isinserted into the Hpa I-Not I backbone fragment ofpADL53.Derivatives of pADL53-449F carrying the 34am and 56Imutations were constructed using similar methods. PlasmidspADL87 and pADL88 are pBR333-based replicons. InpADL88, the chloramphenicol-resistance gene (cmlR) frompACYC184 is flanked by two 49-mers containing a 27-bpsequence that includes IRS from the outside end of Tn5. InpADL87, the Tn903 kanamycin-resistance gene (kanR) isflanked on one side by ISSOR-56I and on the other side by a61-mer containing the outside end of TnS. This transposon iscarried in the his region of pNK2242 (gift of N. Kleckner).For plasmid mutagenesis a recA strain with pTG20 wastreated with ethyl methanesulfonate, and plasmid DNA wasisolated (5) and transformed into SY1141.

Transposition and Inhibition Assays. Assays are describedin text and in table legends (see ref. 16 for more detaileddescriptions).

RESULTSA Lac Papillation Assay That Is Sensitive to Inhibition. A

Lac papillation assay that reflects Tn5 transposition frequen-cies in single E. coli colonies was used to screen directly formutations in IS50R affecting transposition regulation (17).The assay employs Tn5-620, (Fig. 1), a derivative of Tn5 thatcarries a wild-type copy of IS50R at its right end and 26 bpcomprising a functional outside end of TnS at its left end (16,18). The f3-galactosidase gene (lacZ) is fused at its ninthcodon to the left outside end, and the transposon contains atetracycline-resistance gene (tetR). Transcriptional and trans-lational start signals for lacZ are absent from TnS-620. Afraction of transpositions of TnS-620 into the E. coli chro-mosome produce gene fusions that code for active 3-galac-tosidase fusion proteins, conferring a Lac' phenotype on afew cells in a colony. On rich plates supplemented withlactose and the indicator 5-bromo-4-chloro-3-indolyl 3-D-galactopyranoside, these cells will grow to form blue papillae

Abbreviation: IRS, internal repeat sequence.tTo whom reprint requests should be addressed.

6072

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 28

, 202

1

Proc. Natl. Acad. Sci. USA 88 (1991) 6073

IS50L

kan Rh

'lacZ lac Y

IS5ORtnp

'A inhs h n1_-

4- IRS

tetR /

T 449 phe451 gIn

, t34 am

56 ile

FIG. 1. Structure of TnS and some of its derivatives. IRS is the 20-base-pair (bp) inverted internal repeat sequence that defines the ends ofTn5. P1P2 is the tandem promoter that directs transcription of tnp and inh, the overlapping structural genes for Tnp and Inh. n, h, and s showthe position of the NotI, HindIII, and SauI sites. The 34am mutation was made by inserting an Xba I linker into the 5' end of tnp to give anInh+ Tnp- derivative. Substitution of an isoleucine codon for the wild-type codon 56 of tnp (56I mutant; ref. 5) produces an Inh- Tnp-561derivative. The 449F mutant is from this study and 451Q is from ref. 6. In Tn5-620, most of ISSOL and the unique region has been deleted, thefragment shown is substituted for those sequences.

upon the lighter Lac- colony. The number of papillae percolony is thus a measure of the frequency of transposition ofTnS-620 (Fig. 2C). When a functional inh gene is placed intrans to the TnS-620 marker transposon, the frequency ofLacpapillation is visibly reduced (Fig. 2A). A plasmid carryingISSOR (pADL53, Fig. 2A) or an ISSOR derivative that makesonly Inh (pADL53-34am; data not shown) reduces the num-ber of papillae, while a plasmid carrying an ISSOR derivativethat makes only Tnp (pADL53-56I; data not shown) causesno reduction. Thus the papillation assay for transposition ofTnS-620 is sensitive to the suppression of transposition thatis mediated by a wild-type inh present in trans.

Isolation of an ISSOR Mutant with a Trans-Dominant Effecton Regulation. We used the Lac papillation assay to identifyTnS mutants with altered abilities to inhibit transposition intrans. In this screen, mutagenized pACYC184::Tn5 (pTG20)plasmid DNA was transformed into SY1141 (Alac recA/pRZ620), and the transformants were plated on papillationmedium and incubated at 42°C. The transformed colonies

Table 1. Strain list

Source/Strain Genotype ref.

E. coliRR1023 recA56/pOX38 11SY327 SY203 recA56 12SY1093 A(lac-pro)xII supE rpsL rpoB

gyrA malBA17SY1141 SY327/F'110/pRZ620 t

PhagesA421 b221 c1857 rex::Tn5 Oam29 Pam3 1A421-449F b221 cI857 rex::TnS-449F t

Oam29 Pam3ANK1038AADL87 t

PlasmidspACYC184 13pTG20 pACYC184::TnS 14pRZ620 ColE1::TnS-620 6pRZ1037 ColE1::IS50R-561 (Inh translation 15

initiation mutant)pADL53 pTG20 AIS50L tpADL53-34am pADL53::Xba I linker at Hpa I

site tpADL53-56I pADL53 carrying Inh translation

initiation mutation frompRZ1037 t

pADL87 tpADL88 tpTG20, pADL53, pADL53-34am, pADL53-561, pADL87, and

AADL87 all exist in wild-type and 449F allelic forms.*Laboratory collection.tThis work.tN. Kleckner, Harvard University.

were screened for increased numbers of papillae. Note thatTnS-620 was not mutagenized; TnS-620 is the reporter trans-poson employed for the purpose of visualizing the effects ofthe mutagenized pTG20 plasmids. Colonies exhibiting anexcess of papillae were purified and restreaked to confirmtheir phenotypes. Forty-two hyperpapillating colonies dis-played a partially defective inhibition phenotype. In theseisolates, papillation frequency was intermediate between thewild-type control (SY1141/pTG20) and the null control(SY1141/pACYC184). This is the phenotype expected formutations that reduce but do not abolish Inh function.However, for two mutants, the frequency of TnS-620 trans-position was higher in the presence of the mutant pTG20plasmids than in the presence of the vector (pACYC184)alone. The phenotype of one of these mutants (ISSOR-449F,which is carried on the plasmid pADL53-449F) is shown inFig. 2B. Rather than inhibit TnS-620 transposition, this mu-tant acts in trans to promote TnS-620 transposition, indicatingthat this mutant is dominant over the wild-type ISSOR carriedby pRZ620. The dominant effect of the TnS-449F mutant istemperature-sensitive: at 30°C the mutant displays wild-typeinhibition of papillation, rather than the dominant mutantphenotype (data not shown). The second mutant is an ISSORthat was converted to ISSOL (to be described elsewhere).A partial nucleotide sequence of the mutant ISSOR allele

was determined; a single C -* T transition at nucleotide 1437

in ISSOR was found that results in a Leu -- Phe replacementat amino acid position 449 in Tnp and was thus designatedTnS-449F. To show that the mutant phenotype was caused bythe identified sequence alteration, a 340-bp HindIII-Sau96Ifragment containing the 449F mutation (and no other se-quence alteration) was exchanged for the correspondingfragment of wild-type ISSOR. The phenotype of the resultingISSOR-containing plasmid was the same as the original ISSOR-449F mutant (data not shown). Thus, the dominant mutantphenotype of TnS-449F cosegregates with a single base-pairchange.

Identification of the ISSOR Gene Required for the Tn5-449FMutant Phenotype. To identify the ISSOR gene required forthe phenotypes of TnS-449F, tnp and inh were inactivatedindividually. Inactivation of the tnp open reading frame wasachieved through insertion of a synthetic octanucleotide intothe Hpa I site of IS50R (Fig. 1). The insertion introduces anamber codon at position 34 in the tnp coding sequence; thischange lies upstream from the translation initiation codon forInh. The 34am derivatives of ISSOR and ISSOR-449F aretherefore Tnp- and Inh+. Inactivation ofthe inh open readingframe was achieved by removing the translation initiationcodon for inh by substituting an isoleucine (at position 56 inTnp) for the initiator methionine in inh (15). ISSOR allelescarrying this mutation (here designated 561) are Tnp+ Inh-.Papillating colonies carrying the parental 449F mutant and its34am and 561 derivatives are shown in Fig. 2. Inactivation ofinh does not alter the 449F mutant phenotype (Fig. 2 E vs. B).However, inactivation of the tnp open reading frame in the

IRS

Tn 5-620 t

Genetics: DeLong and Syvanen

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 28

, 202

1

6074 Genetics: DeLong and Syvanen

A C

D E::4:m ::es' :.:::. h

FIG. 2. Papillation of TnS-620 and its inhibition by Inh and the 449F phenotype. Overnight cultures of SY1141 transformant strains carryingthe plasmids indicated were diluted in Luria-Bertani (LB) medium, spread on papillation medium, and incubated at 420C for 4 days. (A)SY1141/pADL53+ (Tnp+ Inh+). (B) SY1141/pADL53-449F (Tnp-449F Inh-449F). (C) SY1141/pACYC184 (Tnp- Inh-). (D) SY1141/pADL53-34am449F (Tnp- Inh-449F). (E) SY1141/pADL53-561449F (Tnp-56I449F Inh-).

ISSOR-449F mutant abolishes the mutant phenotype (Fig. 2Dvs. B) and reduces papillation as much as wild-type ISSORdoes. This result demonstrates that the tnp open readingframe is necessary for the dominant 449F mutant phenotype.It also shows that the 449F element codes for a functional Inhprotein that, in the absence of Tnp-449F, acts like thewild-type inhibitor.TnS449F Is Sllghtiy Defective in Transposition. The results

in Fig. 2 D and E show that the tnp product is responsible forthe phenotype of TnS-449F. Quantitative transposition as-says were employed to gain a better understanding of thenature ofTnp449F. To perform these assays, the 449F allelewas introduced into a A: :TnS (A421) transducing phage. WhenA421 infects kanamycin-sensitive cells, Tnp acts in cis topromote TnS transposition from the phage genome into thebacterial chromosome, generating kanamycin-resistantclones. Thus this assay measures the effects of the 449Fmutation in cis, while the papillation assay shows the effectsin trans. Transposition of Tn5 from A421 and A421-449F wasassayed at 300C, 370C, and 420C (Table 2). The 449F trans-poson is somewhat deficient in transposition at all tempera-tures, and the defect is more severe at high temperature(5-fold) than at low temperature (2- to 3-fold). This shows thatTnp-449F is not more active in transposition than wild-typeTnp and thus eliminates the possibility that the mutant

Table 2. Transposition proficiency of TnS-449FTransposition frequency x 106

Transposon 300C 370C 420CTnS 2.9 ± 0.7 2.0 ± 0.7 1.1 ± 0.3Tn5-449F 0.92 ± 0.18 0.52 ± 0.15 0.21 + 0.001The A infection assay (16) was used to determine the transposition

frequencies of TnS and TnS-449F. A421 or A421-449F was infectedinto SY327 cells at a multiplicity of infection of 1.0. Infected cellswere incubated 40 min at the temperature indicated, and kanamycin-resistant clones were selected at the same temperature. Transposi-tion frequency is expressed as kanamycin-resistant colonies arisingper infected cell. Assays were performed in triplicate.

phenotype of Tnp-449F is due to a much more active trans-posase.Tnp-449F Acts in Trans. In the quantitative assays de-

scribed above, we measured the activity of Tnp-449F by Ainfection assays. The kinetics of TnS transposition from aninfecting phage are qualitatively different from the kinetics oftransposition from a replicon already established in the cell.This difference results largely from the hypomethylated stateof A DNA and the activation of the tnp promoter by hypo-methylation (7, 8). These two factors cause a burst of Tnpsynthesis immediately following A::TnS infection; Inh syn-thesis lags behind (19). In contrast, in cells carrying anestablished TnS, such as TnS-620 in the papillation assay, Tnpand Inh are present at steady-state levels. To exclude thepossibility that Tnp-449F is defective only upon freshlyentering a cell, we also quantitatively assayed transpositionofTnS alleles that were fully established. In addition, we usedthe Inh- allele of ISSOR, IS50R-56I, to eliminate the effectsof Inh on Tnp activity.The abilities of Tnp-56I and Tnp-561449F to promote

transposition both in cis and in trans were compared directly.The two tnp alleles were inserted into a TnO derivativecarrying the kanamycin-resistance marker. Each transposonwas crossed onto phage A, and each recombinant phage waslysogenized into the chromosome of a recA strain carryingthe mini-F plasmid pOX38. In these cells, cis transpositioncatalyzed by a given Tnp results in the formation ofpOX38::TnS (i.e., transferable kanamycin resistance). Toallow simultaneous measurement of transposition promotedin trans by these alleles of tnp, a plasmid carrying two outsideIRS ends of TnS flanking a chloramphenicol-resistance(cmlR) determinant also was introduced into the cells carry-ing A: :Tn5 derivatives and pOX38. The cmlR transposon itselfis transposition-defective, but if Tnp is supplied in trans thiselement can transpose to give pOX38::cmlR (i.e., transfer-able chloramphenicol resistance).The results are shown in Table 3. Tnp-561 transposes in cis,

but transposition promoted in trans was undetectable; thepreference for action on the transposon present in cis com-

Proc. Nad. Acad Sci. USA 88 (1991)

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 28

, 202

1

Proc. Natl. Acad. Sci. USA 88 (1991) 6075

Table 3. The 449F mutation allows transposase to act in transfx 106

Transposase In cis* In transt fcis/ftransTnp-561 2.8 ± 1.9 <0.05 >55Tnp-561449F 0.27 ± 0.14 0.35 ± 0.3 0.77The recA56 donor strains carried pOX38 and pADL88. The

Tnp-561 strain carried AADL87, while the Tnp-561449F donor carriedAADL87-4. Donor strains were incubated with the recipient (SY1093)and dilutions were spread on selective media. The values shownrepresent the averages of five transposition frequencies (.) obtainedin two different experiments; the error is the standard deviation.*Transposition in cis is the movement of the kanR gene from eitherAADL87 or AADL87-4 onto pOX38 and is determined by dividingthe number of KanR SmR NalR exconjugants by the number of totalexconjugants.

tTransposition in trans is the movement of the cmIR gene ontopOX38 and is determined by dividing the number ofCmlR SmR NalRexconjugants by the total number of exconjugants and correctingthat frequency for the background frequency of pADL88 plasmidmobilization. This background frequency ofcmlR transfer (5.1 ± 3.4x 10-8) was determined by mating recA56/pOX38/pADL88 (pro-ducing no Tn5 transposase proteins) with SY1093 and was the sameas the frequency of cmlR transfer (4.7 ± 2.6 x 10-8) obtained withthe strain producing Tnp-561. In both the "no Tnp" control matingand the Tnp-561 mating (but not in the Tnp-56I449F mating), Cm1Rexconjugants also were CbR, indicating that cmlR transfer resultedfrom plasmid mobilization. The presence of an entire pADL88plasmid in the CmIR exconjugants from the Tnp-56I mating wasverified by Southern blot analysis, confirming that plasmid mobi-lization and not transposition had occurred.

pared with that in trans is >55-fold for Tnp-561. This showsthat Tnp-561 is similar to Tnp in its preference for cis action(1, 23). In contrast, despite the 10-fold defect of Tnp-561449Fin cis transposition, this transposase is quite proficient fortrans transposition (Table 3). In fact, the cis/trans ratio forTnp-561449F is nearly 1 (0.77, Table 3), indicating that the449F mutation greatly reduces Tnp's preference for endspresent in cis.The results in Table 3 suggest that the papillation pheno-

type ofthe original TnS-449F mutation is due to a trans-actingTnp-449F. The experiment described in Table 3, however,involved only Tnp. In contrast, in the original papillationscreen, cells carrying TnS-449F in trans to TnS-620 producedTnp, Inh, Tnp-449F, and Inh-449F. To test whether Tnp-449Fis also insensitive to Inh or Inh-449F, transposition of TnS-449F was assayed in the presence of the Inh+ Tnp- deriva-tives of ISSOR and ISSOR-449F (Table 4). The transpositionfrequencies of both TnS and TnS-449F were decreased >100-fold by the presence of either wild-type Inh or Inh-449F.Thus, TnS-449F is sensitive to inhibition and Inh-449F is acompetent inhibitor. This indicates that the original papilla-tion phenotype of TnS-449F is due to the trans-acting trans-posase alone.

Table 4. Inh-449F is inhibition-competentIS50 gene

Phage Plasmid products f x 106 Rel.A::TnS pACY184 - 21 1.0

pADL53-34am Inh+ 0.056 0.003pADL53-34am449F Inh-449F 0.051 0.002

A::Tn5-449F pACYC184 6.1 1.0pADL53-34am Inh+ 0.014 0.002pADL53-34am449F Inh-449F 0.032 0.005

Transposition frequencies (f) of TnS and Tn5-449F in SY327 cellscarrying the plasmids indicated were determined as described inTable 2, except that the infected cells were incubated and kanamy-cin-resistant clones were selected at 39.5°C. Relative transposition(Rel.) is the ratio of a given transposition frequency to the frequencyobtained with the same TnS allele in cells carrying pACYC184.

Table 5. Transposase inhibits Tn5 transpositionIS50 gene

Phage Plasmid product f x 106 Rel.

A::Tn5 pACYC184 4.0 1.0pADL53-56I Tnp-561 0.15 0.04pADL53-56I449F Tnp-561449F 0.8 0.2

A::Tn5-449F pACYC184 2.1 1.0pADL53-561 Tnp-561 0.02 0.01pADL53-561449F Tnp-561449F 1.1 0.5

Transposition frequencies (f) and relative transposition (Rel.) ofTnS and Tn5-449F in SY327 cells carrying the plasmids indicatedwere determined as described in Table 4.

Characterization of the Protein Products of ISSOR-449F. Toshow that the phenotype ofTn5-449F is due to a qualitativelydifferent Tnp product and not to differences in the amount ofTnp produced, we compared the protein products of ISSOR-449F and wild-type ISSOR by an immunoassay procedure (9).Overall, the wild-type and 449F mutant protein productswere not distinguishable from each other on the basis of theirtotal accumulation in cells carrying multiple copies of theappropriate ISSOR allele (data not shown).Tnp-561 Inhibits TnS Transposition in Trans. We also

examined the possibility that Tnp itself has inhibitor activity.To this end, the transposition proficiencies of TnS andTn5449F were assayed in the presence of the Inh- elementsIS50R-561 and ISSOR-561449F. As shown (Table 5) bothTnp-561 and Tnp-561449F have inhibitory activity of theirown. Furthermore, both Tn5 and Tn5-449F are sensitive tothis inhibition. There is, however, a significant quantitativedifference between the two Tnp alleles in that Tnp-561449F isa much weaker inhibitor than Tnp-561. This difference be-tween the two Tnp alleles was seen at 39.5°C (as in Table 5);at 30°C both Tnp alleles showed the same inhibitory activity(data not shown). Therefore the deficiency in inhibition of theTnp-449F allele is temperature-sensitive, as is its originalphenotype.

This experiment shows that Tnp-561, although transposi-tion-competent in cis, inhibits transposition in trans. The 25-to 100-fold inhibitory activity of Tnp-56I shown in Table 5compares to the 300- to 500-fold decrease caused by Inh alone(compare with Table 4). This "negative complementation"effect of Tnp-56I can explain why TnS-56I, which producesno Inh, does not transpose more frequently than wild-typeTnS (15). It seems likely that the wild-type Tnp allele alsopossesses inhibitory activity. This would explain why ISSORderivatives that overproduce Tnp and Inh cause transpositionof TnS to decline (20).

DISCUSSIONA mutant Tn5 transposase that stimulates transposition intrans has been isolated. The phenotype of ISSOR-449F isdominant, is temperature-sensitive, and segregates with asingle base-pair alteration. As assayed by papillation, themutant allele is inhibition-competent at 30°C but promotestransposition in trans at 42°C. Tnp-449F is necessary andsufficient for the dominant phenotype, while Inh-449F isinhibition-competent at all temperatures. Quantitative assaysshow that Tnp-449F and wild-type Tnp are equally sensitiveto Inh produced in trans. Unlike wild-type Tnp, however, the449F transposase acts as well in trans as in cis. The 449Flesion maps very near a second mutation, 451Q, that in-creases transposition frequency nearly 4-fold (17), suggestingthat this small carboxyl-terminal region is an importantdomain in Tnp.The ability of Tnp-449F to act in trans places clear con-

straints on the possible mechanisms proposed for the pref-erential cis activity of wild-type Tnp. Three previously con-

Genetics: DeLong and Syvanen

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 28

, 202

1

6076 Genetics: DeLong and Syvanen

sidered explanations become unlikely. (i) Because Tnp-449Fis still sensitive to Inh, cis action cannot be attributed simplyto the interaction between Tnp and Inh. (it) Since the 449Fmutation does not affect a dam methylation site, cis action isunlikely to arise from coordination of Tnp synthesis and IRSactivation. (iii) Intrinsic low transposase activity also cannotaccount for Tnp's cis-acting phenotype, since Tnp-449Fexhibits a partial deficiency in cis, but greatly enhancedactivity in trans.

Preferential cis action may reflect a propensity of Tnp toirreversibly bind DNA near its site of synthesis, coupled withan affinity for transposon ends that does not greatly exceedthe affinity for random DNA (21). Possibly, diffusion awayfrom the site of synthesis would result in nonspecific bindingof Tnp or, alternatively, the nascent polypeptide during Tnpsynthesis may bind DNA. Under this formulation the 449Fmutation could weaken Tnp binding to either nonspecificDNA or result in reversible binding to transposon ends. Thiswould explain why Tnp-449F is defective in cis but proficientin trans.

Recently isolated mutations in the transposase genes ofIS903 and ISIO appear to increase transposase activity intrans by increasing the transposase stability (11). We haveshown that the 449F mutation does not significantly alter theaccumulation of the Tnp polypeptide, though possibly themutation increases the functional stability. An increasedfunctional half-life might explain the somewhat paradoxicalbehavior of Tnp-449F in the various assays described here.While the mutant strongly enhanced transposition of a Tnp+reporter in the papillation assay, it also weakly inhibitedtransposition in the A infection assay (Table 5). Furthermore,inh and inh449F were present in the papillation experiments,while A infection assays show that Tnp-449F is sensitive toboth of these inh alleles. A relatively small increase in thefunctional half-life ofTnp could play a much larger role in thebehavior of TnS-449F in the papillation assay, which lastsseveral days, than in the A infection assay, which lasts onlya few hours.

Since Inh does not block Tnp synthesis, two primarymodels for Inh action can be considered (1, 2). Inh could actas a dominant negative allele ofTnp by forming transposition-defective oligomers with Tnp, or Inh could nonproductivelybind to the IRS elements and block Tnp access. A growingbody of evidence suggests that the first model is likely to becorrect. As we show here, tnp alleles exhibit negative com-plementation patterns suggesting the Tnp subunits can inter-act. Similarly, Biek and Roth (22) have obtained evidence ofintragenic complementation among inh alleles, supporting amodel in which Inh subunits interact. In addition, biochem-ical evidence indicates that an interaction of Tnp and Inh

mediates membrane localization of Inh (9). Therefore, itseems likely that both Tnp and Inh form oligomers, and thatInh blocks transposition by forming a transposition-defectiveoligomer with Tnp. In support of this model, native Tnp andInh proteins have been found to behave as polydisperseoligomers of 6 to 12 subunits in sucrose gradient sedimenta-tion experiments (T. Hanley and M.S., unpublished data).

We thank John Sedivy and Nancy Kleckner for help and usefuladvice. This work was supported by a grant from the NationalInstitutes of Health (GM28142).

1. Isberg, R. R., Lazaar, A. L. & Syvanen, M. (1982) Cell 30,883-892.

2. Johnson, R. C., Yin, J. C. P. & Reznikoff, W. S. (1982) Cell 30,873-882.

3. Auerswald, E.-A., Ludwig, G. & Schaller, H. (1980) ColdSpring Harbor Symp. Quant. Biol. 45, 107-114.

4. Mazodier, P., Cossart, P., Giraud, E. & Gasser, R. (1985)Nucleic Acids Res. 13, 195-205.

5. Maniatis, T., Fritsch, E. F. & Sambrook, J. (1982) MolecularCloning:A Laboratory Manual (Cold Spring Harbor Lab., ColdSpring Harbor, NY).

6. Krebs, M. P. & Reznikoff, W. S. (1986) J. Mol. Biol. 192,781-791.

7. Yin, J. C. P., Krebs, M. P. & Reznikoff, W. S. (1988) J. Mol.Biol. 199, 35-45.

8. McCommas, S. A. & Syvanen, M. (1988) J. Bacteriol. 170,889-894.

9. DeLong, A. & Syvanen, M. (1990) J. Bacteriol. 172, 5516-5518.10. Miller, J. H. (1972) Experiments in Molecular Genetics (Cold

Spring Harbor Lab., Cold Spring Harbor, NY).11. Derbyshire, K. M., Kramer, M. & Grindley, N. D. F. (1990)

Proc. Natl. Acad. Sci. USA 87, 4048-4052.12. Hopkins, J. D., Clements, M. B., Liang, T.-Y., Isberg, R. R.

& Syvanen, M. (1980) Proc. Natl. Acad. Sci. USA 77, 2814-2818.

13. Kieny, M. P., Lathe, R. & Lecocq, J. P. (1983) Gene 26, 91-99.14. Chang, A. C. Y. & Cohen, S. N. (1978) J. Bacteriol. 134,

1141-1156.15. Yin, J. C. P. & Reznikoff, W. S. (1988) J. Bacteriol. 170,

3008-3015.16. DeLong, A. (1989) Ph.D. thesis (Harvard Univ., Cambridge,

MA).17. Krebs, M. P. & Reznikoff, W. S. (1988) Gene 63, 277-285.18. Johnson, R. C. & Reznikoff, W. S. (1983) Nature (London)

304, 280-282.19. Rossetti, 0. L., Altman, R. & Young, R. (1984) Gene 32,91-98.20. Isberg, R. R. (1983) Ph.D. dissertation (Harvard Univ., Cam-

bridge, MA).21. Morisato, D., Way, J. C., Kim, H.-J. & Kleckner, N. (1983)

Cell 32, 799-807.22. Biek, D. P. & Roth, J. R. (1989) J. Bacteriol. 171, 2056-2066.23. Isberg, R. R. & Syvanen, M. (1981) J. Mol. Biol. 150, 15-32.

Proc. Natl. Acad. Sci. USA 88 (1991)

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 28

, 202

1