Mol Gen Genet (1983) 191:442-450 © Springer-Verlag 1983

Fine Structure of Transposition Genes of Tn2603 and Complementation of its tnpA and tnpR Mutations by Related Transposons Michiyasu Tanaka~ Tomoko Yamamoto, and Tetsuo Sawai Division of Microbial Chemistry, Faculty of Pharmaceutical Sciences, Chiba University, 1-33 Yayoi-cho, Chiba 260, Japan

Summary. The fine structure of the genes tnpA, tnpR and res of Tn2603 required for its own transposition, was determined. The order of the genes was tnpA-tnpR-res from the right end of the right hand side region in Tn2603, the tnpA and tnpR encoded gene products having molecular weights of 110,000 and 21,000, respectively. The 110,000 molecular weight polypeptides was absolutely required for replicon fusion as the first stage of transposition, and named transposase. On the other hand, the 21,000 molecular weight polypeptide was necessary for resolution of the cointegrate as the second stage of transposition, and named resolvase. We also examined the ability of various transposons, assumed to be closely related, to complement the tnpA and tnpR mutations of Tn2603. The results indicated that the mercury resistance transposon, Tn2613, and Tn501, can complement both genes, but TnAs and ~0 cannot at all. Tn501 had much less efficiency of complementation for tnpA than Tn2613. We have also discovered that the transposition frequency of transposons in the Tn2613 family systematically depend on their size of transposon.

Introduction It has been recognized that many genes on plasmids or chromosomes are located on transposable DNA units, namely transposons. They are for antibiotic resistance, heavy metal resistance, lactose utilization, toxin production and so on (for reviews see, Kleckner 1981; Kopecko 1980). A transposon is capable of sequential transposition from one replicon to another in a recA- background. Transposons can be divided into two classes on the basis of their structures and possible mechanisms of transposition (Galas and Chandler 1981). One of these classes has a structure such that accessory determinants like antibiotic resistance are flanked by two identical or similar insertion sequence (IS) elements at both ends in inverted or directed orientation, e.g., Tn5, Tn9, TnlO, Tn903, Tn1681, Tn2350 and so on (Berg etal. 1975; Gottesman and Rosner 1975; Kleckner et al. 1975; Nomura et al. 1978; So et al. 1979; Clerget et al. 1980). These transposons can be considered as being formed by a route in which the determinants have been bracketed by a pair of IS elements, suggesting that the essential function on the transposition is attributed to the IS element itself. Most transposons belonging to this

Offprint requests to. T. Yamamoto

class have been shown to transpose by a mechanism which does not form a cointegrate structure as an intermediate of transposition (Galas and Chandler 1981). On the other hand, another class of transposon has a pair of short inverted repeat (IR) sequences at both ends and encodes the functions necessary for transposition in the central region of the element, e.g., Tn3, y0, Tn501, Tn21, Tn2603 and so on (Heffron et al. 1977; Gill et al. 1979; Reed 1981; Kitts et al. 1982a; Schmitt et al. 1981; De La Cruz and Grinsted 1982; Tanaka et al. 1982). As one possible route by which these transposons are formed it has been postulated that a transposable element very like the minor transposon of Tnl721 flanks the appropriate genes, then deletion could result in their formation (Altenbuchner et al. 1981). Furthermore, the more complex transposons have been evolved through sequential alteration of structure by insertion, substitution or recombination (Tanaka et al. 1983). Transposons belonging to this class were shown to transpose via the formation of a cointegrate structure of two replicons joined by two copies of the transposon as an intermediate (Gill et al. 1978; Arthur and Sherratt 1979; Reed 1981; De La Cruz and Grinsted 1982; Grinsted et al. 1982; Tanaka et al. 1982). We previously reported the genetic structure and functions of Tn2603 which encodes the type II fl-lactamase (Ap 1 resistance), Sm/Sp resistance, Su resistance, and Hg resistance genes (Yamamoto et al. 1981a, b, 1982a; Tanaka et al. 1982). Tn2603 has a molecular size of 20 kb and a pair of short IRs at both ends. The resistance genes encoded by Tn2603 were located on the restriction endonuclease cleavage map, and the direction of transcription of each gene was determined (Yamamoto et al. 1981a, 1982a). We have also determined the genes necessary for transposition of Tn2603 (Tanaka et al. 1982). One of the genes, tnpA, is located at the right hand side of the internal portion of Tn2603 and encodes the | 10,000 molecular weight polypeptide that could be assigned to the transposase of Tn2603. The molecular weight of this polypeptide is identical with that of the transposase of Tn2602, which has a nearly identical structure to Tn3. Another gene encoding the polypeptide having a molecular weight of 21,000 was determined a.t the adjacent left part of tnpA (Tanaka et al. 1982). This 21,000 molecular weight polypeptide is predicted to be reAbbreviations: Ap, ampicillin; Sm, streptomycin; Sp, spectinomycin; Hg, mercury; Km, kanamycin; Tp, trimethoprim; Nx, nalidixic acid; Rif, rifampicin; Tc, tetracycline; Cm, chloramphenicol; kb, kilobase

443 Table 1. Plasmids used

Plasmids

Relevant markers a

Remarks/References

pMKI:: Tn2603# 1 pMK1 :: Tn2603del-170 pTY60 pMK1 :: Tn2601# 4 pMK1 ::Tn2602# 1 pMK1 ::Tnl# 2 pMK1 :: Tn3# 4 pSC179 (pSC101 :: Tn4) pMKt :: Tn21# 1 ColEl::Tn2608# 1 pMK1 ::Tn2613# 2 pUB781 (ColE1 :: Tn501) pBR322 pBR322 :: 7c~#11 pACYCt84 pACYC 184 : : Tn2613# 1 pACYC184 :: Tn2601 pACYC184:: Tn2602 pACYCI84:: 76# 1 pTY54

Ap, Sm/Sp, Su, Ap, Sm/Sp, Su, Ap, Sm/Sp, Su, Ap, Km Ap, Km Ap, Km Ap, Km Ap, Sm/Sp, Su, Sm/Sp, Su, Hg, Sm/Sp, Su, Hg Hg, Km Hg Ap, Tc Tc Cm, Tc Hg, Cm Ap, Cm, Tc Ap, Cm, Tc Cm Hg, Cm

pTY69

Cm

pTY70

Cm

pTY34

Ap, Cm, Tc b

F'lac R386 R388

lac +, tra + Tc, tra + Tp, Su, tra +

Yamamoto et al. (1981b) Tanaka'et al. (1982) This study Yamamoto et al. (1982b) Yamamoto et al. (1982b) Yamamoto et al. (1982b) Yamamoto et al. (1982b) Kopecko et al. (1976) This tudy, Kopecko et al. (1976) Tanaka et al. (1983) Tanaka et al. (1983) Bennett et al. (1978) Bolivar et al. (1977) This study, 7c~transposed into Ap resistance gene of pBR322 Chang et al. (1978) This study, Tn2613 transposed into Tc resistance gene of pACYC184 Our laboratory Our laboratory This study, 7~ transposed into Tc resistance gene of pACYC184 This study, met, res, tnpR, and tnpA genes of Tn501 were cloned into pACYCI84 with HindIII Tanaka et al. (1982), H2 fragment of pMKt :: Tn2603# 1 (tnpA +) was cloned into pACYC184 Tanaka et al. (1982), B2 fragment of pMK1 :: Tn2603# 1 (tnpA +, tnpR +, res +) was cloned into pACYC184 Yamamoto et al. (1981 a), H4 fragment of pMKI:: Tn2603# 1 was cloned into pACYC184 Our collection Dennison (1972) Datta and Hedges (1972)

Hg, Km Hg, Km Hg

Tc Km

Abbreviations are according to Novick et al. (1976) b Reduced level of resistance compared with pACYC184

a

solvase, involved in the resolution of the intermediate of transposition. In'the present study, we have determined that this 21,000 molecular weight polypeptide is involved in the resolution, resolvase, and that the internal resolution site, called res site, necessary for the resolution is located at the left part of the gene. On the other hand, our previous study also demonstrated that Tn2603 was evolved from an ancestral mercury transposon, Tn2613, and that some phenotypically different transposon has also originated from a c o m m o n ancestral transposon (Tanaka et al. 1983). In this paper, the functional relationship to tnpA and tnpR of various transposons belonging to this class has been described, in addition to the fine structure of the transposition machinery of Tn2603.

Materials and Methods

Bacterial Strains and Plasmids. The bacterial strains used are derivatives of Escherichia coli K12; C600, C600recA (recA and hal derivative of C600), AB2463 (recA str), ML1410 (met naI) and WA5023Riff (polA riJ) (Tanaka et al. 1982). Plasmids used are listed in Table 1. Determination o f Transposition Frequency onto R386. This method was previousyl described (Tanaka et al. 1982). Sm (12.5 ~tg/ml) or Hg (12.5 pg/ml) was used instead of Ap as selective drugs.

Determination o f Transposition Frequency onto R388. AB2463 harboring a donor plasmid and R388 was used as a donor strain for conjugal transfer, and ML1410 was used as a recipient. D o n o r and recipient strains were mixed in a ratio of 1:5, filtered by millipore filter, and then the filters were incubated on BHI agar plates at 37 ° C for 6 h. Transconjugants receiving R388:: Tn were selected with Sm (12.5 pg/ml), Yp (50 ~tg/ml) and Nx (50 lag/ml) on MuellerHinton agar or with Hg (12.5 lag/ml) instead of Sm. Transconjugants receiving R388 were selected with Tp (50 p~g/ml) and Nx (50 ~g/ml) on Mueller-Hinton agar. The transposition frequency is presented as the value of the number of R 3 8 8 : : T n + transconjugants divided by the number of R388 + transconjugants. Isolation o[ R388 : : Tn or a Co&tegrate Formed between Tn2603 or its Deletion Derivatives and the Acceptor R388 Plasmid. AB2463 harboring R388 and pTY 60 or its deletion derivatives in the presence or absence of a complementing plasmid was used as a donor for conjugal transfer. The strain was mated with C600recA, and transconjugants were selected for a transposon marker, with Tp (50 p.g/ml) and Nx (50 ~tg/ml) on Mueller-Hinton agar. Plasmid D N A was prepared from transconjugants, and its structure analyzed by digestion with restriction endonucleases. Isolation o f Plasmids Containing Various Transposons. Plasmids containing 76 from F'lac were constructed by the

444 method described by Guyer (1978). Plasmids containing other transposons have been described elsewhere (Yamamoto etal. 1981b, 1982b; Tanaka etal. 1982, 1983). pACYC184 :: Tn2601 and pACYC184:: Tn2602 were constructed by transposition of each transposon into pACYC184 from RGN14 and RGN823, respectively. pTY54 was constructed by cloning the HindIII fragment of 2bb::Tn501# 3 (our unpublished data) into pACYC184 and contains the tnpA, tnpR, res, and mer genes of Tn501. pSC179 (pSC101 :: Tn4) and pUB781 (ColE1 :: Tn501) were kindly provided by D.J. Kopecko and P. Bennett, respectively.

Tn2603

~s

~BBP I H

ori

II I II II

I H

.EESe 1 Ill II

II

SBg HH HBg

- sui sh" b[a

mer

HH u)o: e =. fnpA

Results

I pTY60 .

del-61 del-62 del-65 del-66 del-68 lkb

del-t?O H2

Other Methods. Preparation of plasmid DNA, digestion with restriction endonuclease, agarose gel electrophoresis, ligation with T4 DNA ligase and transformation were carried out by the methods previously described (Yamamoto et al. 1981 a).

P

I

• .

B2 .H4 E2

=

<

--

pTY61 pTY62 pTY65 pTY66 pTY68 pTY63 pTY69 pTY?O

= =

Fig. 1. Functional maps of pTY60 and its deletion derivatives. Abbreviations: E; EcoRI, B; BamHI, H; HindlII, S; SalI, P; PstI, Bg; BglII, ori; origin of replication, bla; fl-lactamase, str; streptomycin, sul; sulfonamide, mer; mercury, IR; inverted repeat. H2, B2, H4 and E2 represent the name of each fragment. ( - - ) ; deleted region

Determination of res and tnpR Genes in Tn2603 We have previously identified the gene and its product having a molecular weight of 110,000 requried for transposition of Tn2603 (Tanaka et al. 1982). The protein could be assumed to be transposase encoded by tnpA because of a molecular weight similar to the transposase of Tn3. Furthermore, we have identified a 21,000 molecular weight polypeptide encoded by the region of the left-hand side adjacent to the tnpA gene. Because of the locus and the molecular weight of the product, it was assumed to be similar to the resolvase encoded by the tnpR gene of Tn3 or 76 (Tanaka et al. 1982). To define the fine locations and the functions of these putative tnpA and tnpR genes of Tn2603, we constructed a series of deletions in Tn2603 which especially deleted the region encoding the 21,000 molecular weight polypeptide. Simultaneously, we attempted to define the res gene at which the resolvase acts and promotes the res-specific recombination. First, we tried to remove Tn5 from pMK1 ::Tn2603# 1 (Yamamoto et al. 1981a) to exclude the possibility of cointegrate fromation promoted by IS50 in Tn5. pMK1 ::Tn2603# 1 DNA was completely digested with PstI endonuclease, ligated with T4 DNA ligase. E. coli C600 was transformed by the ligated DNA, and transformants were selected with Su resitance. Since the Su resistance gene of Tn2603 contains one PstI site, the resultant transformants were assumed to harbor a plasmid in which two large PstI fragments of pMKl::Tn2603# 1 were jointed in the original orientation. The restirction map of the resultant plasmid, pTY60, lost almost the entire region of Tn5, the complete region of colicin E1 production and immunity, but retains all sequences of Tn2603. Then, a series of deletions was introduced into the internal region of Tn2603 with various restriction endonucleases as follows. pTY60 was completely digested with BamHI or HindIII, and ligated with T4 DNA ligase. After transformation of C600 with the ligated DNA, the transformants were selected with 12.5 pg/ml HgCI 2. Plasmid DNA was prepared from the resultant transformants, and the deleted regions were determined by restriction analysis. With BamHI, pTY61 was constructed which lost all of the internal BamHI fragments of Tn2603. With HindIII, two types of deleted plasmid were obtained. One of the types was pTY68, which

lost all of the internal HindlII fragments of Tn2603. Another type was pTY62, which retained a 1.6 kb HindIII fragment though losing the other HindlII fragments. We also constructed the deletion derivatives, pTY65 and pTY66, by partial digestion with EeoRI endonuclease. By digestion with PstI, pTY63 was constructed from pMK1 ::Tn2603del-170 which had previously been shown to delete the internal 1.2 kb EcoRI fragment encoding a putative tnpA (Tanaka et al. 1982). The deleted regions of the resultant plasmids are shown in Fig. I. To examine the effect on the normal transposition of deletions in Tn2603, their ability to transpose and the structures of transposition-products were determined. Each deletion plasmid and R388 coexisted in the reeA strain, AB2463, with or without a complementing plasmid, and were then mated with C600recA. Transconjugants were selected with Hg (6.3 lag/ml), Tp (50 lag/ml), and Nx (50 lag/ ml). Plasmid DNAs were prepared from transconjugants and their structures were analyzed by digestion with EcoRI endonuclease. The results are shown in Fig. 2 and Table 2. Each Hg resistance gene of pTY60 and pTY61 was proved to transpose normally, suggesting that BamHI fragments of Tn2603 do not affect its own transposition. On the other hand, pTY62, pTY65, pTY66, and pTY68 were shown to form cointegrate structure with R388. These structures contained two transposons in direct orientation at the junctions of two plasmids, suggesting the intermediate in transposition. pTY63 could not transpose onto R388 or form cointegrate structures in the presence of pACYC184. This agrees with our previous report that Tn2603del-170 was deficient in transposition. These results indicate that the function encoded in the region including the 1.2 kb EcoRI fragment can be assigned to the tnpA product, transposase, which mediates the formation of cointegrate structures of two replicons joined by two copies of transposon. Furthermore, the results suggest that pTY62, pTY65, pTY66, and pTY68 lacked the resolvase gene, tnpR, or/and resolution site, res, necessary for the resolution of the cointegrate. The cointegrate structure formed by pTY62 and pTY65 could not be resolved even in the coexistence of pTY70 which encodes a 21,000 molecular weight polypeptide assumed to be the

445

pACYCI84 ¢Cloning"~f H4 fragment E

B

"Clon~'~ggof E2 fragment 8

E

H~B

H

Fig. 2. 0.8% agarose gel electrophoresis of plasmid DNA digested with EcoRI endonuclease. Lane a: 2 phage DNA digested with HindIII endonuclase for molecular size standards (molecular sizes of the fragments are 23.7, 9.46, 6.67, 4.26, 2.25, 1.96 kb from top to bottom). Lane b; R388::Tn2603 de!-170# 1, that is a resolution product of R388::pTY63# 1 by pTY70. Lane c; pTY63. Lane d; R388::pTY63# 1, lane e; R388::pTY63#2 and lane f; R388::pTY63#3. R388::pTY63 were formed from pTY63 and R388 complemented by pTY69. Lane g; R388::Tn2603del170~ 11, lane h; R388::Tn2603del-170~ 12 and lane i; R388 :: Tn2603del-170~ 13. These three types of R388 :: Tn2603del170 were formed by transposition of Tn2603del-170 from pTY63 onto R388 by the complementation of pTY70. Lane j; R388

Table 2. Mode of transposition of Tn2603 and its deletion-deriva-

tives onto R388 Donor plasmid

pTY60 pTY61 pTY62 pTY62 pTY65 pTY65 pTY66 pTY68 pTY63 pTY63 pTY63

Complementing a plasmid

Structure of transposition product formed in recA

None None None pTY70 None pTY70 None None pACYC184 pTY69 pTY70

Normal Normal Cointegrate Cointegrate Cointegrate Cointegrate Cointegrate Cointegrate Not formed Cointegrate Normal

Resolution of cointegrate in presence of pTY70

No No No No No No Yes

a pTY69:pACYC184-H2 (tnpA +, tnpR-), pTY70: pACYC184B2 (tnpA +, tnpR +) tnpR product, resolvase. In the coexistence o f pTY69, pTY63 which has already been shown to lack tnpA, could form the cointegrate structure with R388, and the cointegrate could be resolved in the coexistence of pTY70, and therefore the Hg resistance gene of pTY63 could be trans-

E Resolution E

Fig. 3. Strategy for cloning and resolution of two res sites into pACYC184. Abbreviations: see legend to Fig. l, cat; chloraphenicol acetyltransferase, tet; tetracyline, res; resolution site, str/spc; streptomycin/spectinomycin. Notations: ( ~ ) ; H4 fragment of pMK 1 : : Tn2603# 1, (r'--1); E2 fragment of pMK 1 : : Tn2603# 1. The location of each fragment is shown in Fig. 1 posed normally onto R388 (Fig. 2 and Table 2). From these results, it can be considered that, pTY69 encodes transposase, pTY70 encodes transposase and resolvase, but contains the resolution site, res. Consequently, it has been shown that the 21,000 molecular weight polypeptide corresponds to the resolvase of Tn2603.

Cloning o f the res Site and Resolution by Resolvase To determine the precise location of the tnpR gene encoding resolvase and res site, we constructed a plasmid in which two res sites were cloned. The strategy for construction of such a plasmid is shown in Fig. 3. Firstly, the H4 fragment of pMK1 ::Tn2603# 1 encoding the type II fl-lactamase (bla) gene was cloned into pACYC184 and the orientation o f this fragment in the resultant plasmid, pTY34, was determined by restriction mapping. Second, pTY34 was partially digested with EcoRI endonuclease becasue pTY34 had two EeoRI sites, and the linearized pTY34 D N A was ligated with the E2 fragment of p M K l : : T n 2 6 0 3 # 1 D N A .

446 Table 3. Resolution of pTY99 complemented by each plasmid

Table 4. Complementation transposition of Tn2603del-170

Comp!ementing plasmid

SpS/Apra

Donor plasmid

Complementing plasmid a

pTY61 pTY65 pTY68 pMK1 :: Tn2613# 2 pUB781 pBR322 :: yo~ 11 pMKI ::TnAs b

120/120 0/90 0/9O 64/64 50/50 0/50 0/50

Relative transposition frequency

pMK1 : : Tn2603# 1 pMK1 :: Tn2603# 1 pMK1 :: Tn2603del-170 pMK 1 : : Tn2603del-170 pMK1 : : Tn2603del-170 pMK1 : :Tn2603del- 170 pMKI : :Tn2603del- 170 pMK 1 : :Tn2603del- 170 pMK 1 :: Tn2603del- 170 pMKI :: Tn2603del-170

pACYC184 pTY69 pACYC184 pTY69 pTY70 pACYC184 : : Tn2613# 1 pTY54 pACYC184: : 7 ~ 1 pACYC184: :Tn2601 pACYC184::Tn2602

1.00b 1.06 < 0.001 0.22 0.21 1.08 0.019 < 0.001 < 0.001 <0.001

a Sp: spectinomycin, Ap: ampicillin b pMKl::Tn2601#4, pMKl::Tn2602# 1, pMKI::Tnl# 2 and pMK1 :: Tn3# 4 (Yamamoto et al. 1982b) The ligated D N A was introduced into AB2463 by transformation, and tranformants were selected with Sp 12.5 lag/ml, and then their resistance patterns examined. Among these clones, Cm sensitive colonies were chosen, to obtain the clone in which the E2 fragment was inserted into the EcoRI site of the Cm resistance (cat) gene. Plasmid D N A was prepared from each clone, and the location and orientation of insertion of the E2 fragment confirmed. This plasmid was named pTY99 (Fig. 3). pTY99 has two res sites in direct orientation and can be stably maintained in the recA strain, AB2463, in the absence of the tnpR gene. If the resolvase encoded by the tnpR gene of Tn2603 is provided, it is expected to give an autonomous plasmid, pTY102, due to recombination between two res sites promoted by resolvase. Since pTY99 encodes two bla genes, one str/spc gene, one sul gene, and one tet gene, pTY102 is expected to encode one bla gene only. Therefore, the appearance of pTY102, that is the resolvase-dependent recombination between two res sites, can be determined by monitoring the clones sensitivity to Sp. E. coli AB2463 harboring pTY99 was transformed by various deletion derivatives from pTY60 which have been presumed to lack the res site but to retain tnpR, as indicated in Table 2. Transformants were selected with Hg (12.5 ~tg/ml) and Ap (25 ~tg/ ml), purified on the same agar, and their sensitivity to Sp examined. These results are shown in Table 3. All of the transformants in which pTY61 encoding the tnpR gene was introduced, exhibited sensitivity to Sp. In contrast, all of the transformants in which pTY65 or pTY68 encoding no tnpR gene was introduced, exhibited resistance to Sp. From these results, the precise location of the res site and tnpR gene of Tn2603 could be determined, as shown in Fig. 1. The structure of genes required for transposition of Tn2603 is consistent with that of Tn501, Tn1721, and Tn21 reported by Grinsted et al. (1982). By using this system, we next compared the functions of each resolvase of various transposons which can be considered to transpose via cointegrate structures. The transposons, Tn2613, Tn501, ~,6 and TnAs (Tn2601, Tn2602, Tnl, Tn3) were used. Each plasmid containing these transposons was introduced into AB2463/pTY99 by transformation, and transformants were selected with Km (25 ~tg/ml) for pMK1 ::Tn2603# 2 and pMK1 ::TnAs, Tc (25 ~tg/ml) for pBR322::76# 11, and colicin E1 for pUB781 (ColE1 :: Tn501). Transformants were examined for Sp sensitivity after purification on the selective agar. The results shown in Table 3 suggest that Tn2613 and Tn501 efficiently resolved pTY99, but 76 and TnA did not resolve at all. Since Tn2613 has been assumed to have a sequence in the region

a pTY69:pACYCI84-H2 (tnpA + derived from Tn2603). pTY70: pACYC184-B2 (tnpA + and tnpR + derived from Tn2603). pTY54:pACYC184 cloning the region which encodes tnpA and tnpR gene of Tn501 b Transposition frequency of Sm resistance gene was 1.06 × 10 -3 encoding the resolvase gene identical with Tn2603 (Tanaka et al. 1982), it is proper that the resolvase of Tn2603 and Tn2613 is interchangeable. On the other hand, Tn501 could also resolve at the res sites of Tn2603 though Tn501 has a different cleavage map from Tn2603 and Tn2613 (Tanaka et al. 1982). This result suggested that the function of the resolvase encoded by Tn501 was similar to that of Tn2603 and Tn2613 though with a minor sequence divergence. Complementation of a Transposase-Deficient Mutant of Tn2603 by Various Transposons We next compared the function of transposase encoded by the tnpA of various related transposons. Tn2603del-170 is a transposition-deficient derivative of Tn2603, due to deletion of the 1.2 kb-EcoRI fragment present in the righthand region of Tn2603 (Tanaka et al. 1982). The region including this fragment encodes an 110,000 molecular weight polypeptide, transposase, of which the function has been demonstrated above (Table2), and therefore, Tn2603del-170 is assigned to a transposase-defective mutant (tnpA-). To compare the functions of the transposases of various transposons with that of Tn2603, complementing plasmids containing each tnpA gene were constructed by transpoistion of cloning into pACYC184. Tn2613, 73, Tn2601 or Tn2602 was transposed into pACYC184, pTY54 is a plasmid in which the region encoding tnpA, tnpR, res, and Hg resistance gene of Tn501 was cloned into pACYC184. Each resultant plasmid for complementation and pMKl::Tn2603del-170 was introduced into E. coli AB2463 harboring R386 by transformation, and then used as a donor for mating with E. coli WA5023Rif r. The ability to transpose the Sm resistance gene to R386 was examined by mating assay as described previously (Tanaka et al. 1982). The monitoring of the transposition of the Sm resistance gene was because Tn2601 and Tn2602 mediate ampicillin resistance by elaboration with Type I fl-lactamase. In this experiment, all of the resultant transconjugants exhibited Km sensitivity suggesting that the mobilization of pMKl::Tn2603del-170 by R386 had not occurred. As shown in Table 4, the transposition-deficiency of Tn2603del-170 was complemented by the transposase of

447

Table 5. Transposition frequency of various transposons Donor plasmid

Size of transposon (kb)

pSC179 (pSCI01::Tn4)

pMKl::Tn2603#1 pMKt::Tn21# 1 ColEl::Tn2608# 1 pTY61 (Tn2603del-61) pMK1 : : Tn2613# 2 nd: not done a 3.10X10-3 3.62 X tO -4

22.5 20.0 19.0 13.5 11.2 7.2

10 -~

Relative transposition frequency onto R386

onto R388

Sm

Sm

Hg

0.64 nd 1.00" 1.00b 1.05 nd 1.98 nd nd nd nd 350

Hg

0.39 nd 1.00c 1.00a 1.09 nd 6.49 nd nd 35.0 nd 867

10:

C -1

10 0

r~

b 1.69×10-5 d 1.73 X 10 -5

C

.>_

IO

1.0

Tn2603, i.e., by pTY69 and pTY70 as was expected. The transposases of Tn2613 and Tn501 could complement also, but yr, Tn2601 and Tn2602 could not. Tn2613, which has been assigned to an evolutionary origin Of Tn2603 (Tanaka et al. 1983), storngly complemented, that is, fivehold higher than did pTY69 and pTY70 containing their own tnpA. This result suggests the possibility that Tn2613 encodes a transposase having a much higher efficiency of function than Tn2603, or that production of transposase by Tn2613 is much more than by Tn2603. On the other hand, Tn501 complemented at a lower level than did Tn2603 and Tn2613. Previously, we demonstrated by heteroduplex analysis that the regions encoding the tnpA gene of Tn2603 and Tn501 show different restriction maps but share a strong homology. Also, Tn21, which is closely related to Tn2603, has been demonstrated to be homologous with Tn501 in the tnpA gene (Rownd et al. 1980; De La Cruz and Grinsted 1982). On the basis of these observations, Tn2603 and Tn501 appear to encode similar but not identical transposases. We assume that this is the reason for the lower efficiency of complementation by Tn501. The results that yg, Tn2601 and Tn2602 could not complement Tn2603del-170 suggest that the transposases of these transposons had different specificities from Tn2603, though similar transposition mechanisms via cointegrate formation. Transposition Frequencies of Various Resistance Transposons Related to Tn2603 We previously reported that various complex resistance transposons have been evolved from an ancestral mercury transposon such as Tn2613 (Tanaka et al. 1983). Complementation analysis for transposition of Tn2603del-170 (Table 4) strongly supports this hypothesis. The analysis also demonstrated that the transposition efficiency was remarkably different among various related transposons, though they had exchangable tnpA genes. Especially, Tn2613 was assumed to have an extremely high transposability. To confirm this possibility, we compared the transposition frequencies of evolutionarily related transposons ranging in size between 7.2 kb and 22.5 kb. The structures and evolutionary relationships between Tn4, Tn2603, Tn21, Tn2608, Tn2613, and pTY61 containing Tn2603del-61 in which all BamHI fragments in Tn2603 were deleted, were previously described (Tanaka et al. 1983; Fig. 1). The transposition

[] e

A a 0.1

i 5

A A b c I

10

AA de t 15

A f

I 20kb

S i z e of T r a n s p o s o n

4. Relative transposition frequencies of resistance transposons related to Tn2603 onto R386 (n) and onto R388 (e). Transposition frequency is presented as a relative value compared with that of Tn2603. Abbreviations: a; pMKl::Tn2613# 2, b; pTY61 (Tn2603del-61), c; ColEl::Tn2608#l, d; pMKI::Tn21#I, e; pMK 1 :: Tn2603# l, f; pSC179 (pSC101 : :Tn4)

Fig.

frequencies onto R386 were measured as described in Materials and Methods, and results are shown for each size of transposon in Table 5 and Fig. 4. Transposition frequencies increased with decrease in size, as previously reported for ISl-promoted transposons (Chandler et al. 1982). However, the increase in transposition frequency was not proportional to decrease in size, that is, Tn2613 was capable of transposition at an extremely high efficiency compared with other transposons. This result supports the possibility obtained from complementation analysis of Tn2603del-170 by Tn2613. Tn2603del-61 which lacks all of the BamHI fragments in Tn2603 was capable of transposition at a 35-fold higher frequency compared with that of parental Tn2603, suggesting that the efficiency of transposition depends systematically on the size of the transposon. Discussion

In this paper, we have elucidated the fine genetic structure of the region necessary for normal transposition of Tn2603 (Fig. ]). The order of genes is res-tnpR-tnpA which span nearly 3 kb at the right-hand side of the elements. This structure if markedly homologous with that of Tn501, Tn1721 or Tn21 (Grinsted et al. 1982), but is different from that of y6 of TnAs (Tn3 etc.) which is tnpR-res-tnpA (Heffron et al. 1977; Reed 1981). In Tn3 and ~6, the res site at which the tnpR product, resolvase, acts is within the intercistronic region, and contains or is in contact with each promoter for transcription o f both tnpA and tnpR. There-

448 fore, the sequence-specific binding of resolvase to the res site leads to the subsidiary effect of repression of tnpR and tnpAgene expression (Gill etal. 1979; Reed 1981). It is interesting that the regulation system for expression of tnpA and tnpR in Tn2603 might be different from that in Tn3 and y8 because of the divergence of the genetic structure, but this is not elucidated at present. The transposase and resolvase of TnAs and 7~ were not able to function on Tn2603 (Table 3 and Table 4). The results suggested that TnAs and 76 are not only different in genetic structure but in the functions of transposase and resolvase. The transposase of Tn2613 which has been considered to be a direct ancester of Tn2603 (Tanaka et al. 1983), could complement Tn2603del-170 at an extremely high efficiency, while Tn501 could complement at low efficiency. This difference in efficiency of complementation suggests the evolutional distance between Tn2613 and Tn501 despite their similar phenotype. This result supports our previous hypothesis that Tn2603 has been evolved directly from the ancestral Tn2613, and that the essential functions for transposition would be retained (Tanaka et al. 1983). Grinsted etal. (1982) previously reported that the transposase of Tn501 could not complement Tn21. Since Tn21 is closely related to Tn2603 (Tanaka et al. 1983), this divergence in complementation of Tn501 on Tn2603 and Tn21 is strange. We can consider two possible explanations. One possibility is a difference in assay system of transpogition. We examined the transposition ability of Tn2603del-170 complemented by Tn501 onto R386, while mobilization frequency was monitored by transposition. Another possibility is that the IR sequence of Tn2603 is more homologous with Tn501 than with that of Tn21. The highly efficient function of the tnpA gene of Tn2613 was also confirmed by determination of the transposition frequency to compare the efficiencies of transposition among several transposons that have evolved from the ancestral Tn2613 (Table 5 and Fig. 4). The result also suggests that the efficiency of transposition of this group depends systematically on their size. This result agrees with the result concerning the ISl-flanked transposon reported by Chandler et al. (1982). However, the more drastic effect of their size appeared in transposition of the Tn2613 family than in the ISl-flanked tmnsposons. This phenomenon may be attributable to the specific properties of the Tn2613 family in the presence of size-limiting acceptance of the transposon by R388 and R386. We have demonstrated that the normal transposition of Tn2603 involves two stages. The first stage is the formation of a cointegrate structure of two, donor and target, replicons joined by two copies of the transposon, and is mediated by a transposase encoded by the tnpA gene. As discussed above, the Tn2603 transposase is not interehangable with that of Tn3 or of 7c~. The second stage is the resolution of the cointegrate structure into two replicons resulting in a copy of the transposon at its original site in the donor replicon and a new copy in the target replicon. Resolution of Tn2603-mediated cointegrates absolutely required resolvase and the res site at which the resolvase binds. The Tn2603 resolvase can be assigned to the 21,000 molecular weight polypeptide which has been identified in E. coli minicells (Tanaka et al. 1982). Lack of the tnpR or/and res genes resulted in accumulation of eointegrate structures which could be resolved by complementation in trans by the tnpR gene product, only when the res gene

existed in cis (Tables 2 and 3). We compared the functions of resolvases encoded by various transposons. Tn2613 and Tn501 are efficient at the res site of Tn2603, but those encoded by TnAs and 7~ are not. That is to say, Tn2613 and Tn501 could complement not only transposase but resolvase with Tn2603. TnAs and 7~ could not complement both proteins with Tn2603. The difference in functions of transposase and resolvase between both groups of transposon corresponds to the difference in their gene structures. Tn2613 complemented the transposase and the resolvase with Tn2603 with high efficiency. In contrast, Tn501 complemented well the resolvase with Tn2603 but only poorly the transposase. Therefore, Tn2603 is considered to be more closely related to Tn2613 than Tn501. This consideration is in agreement with the results of structural analysis of these transposons (Tanaka et al. 1983). Therefore, we concluded that the transposition functions were completely conserved throughout the evolution of transposons belonging to the Tn2613 family. Tn501 may be changed not only in structure but in the function of transposase compared with transposons of the Tn2613 family (Grinsted et al. 1982; Tanaka et al. 1983). TnAs and 7~ are considered to be more separated from the Tn2613 family than Tn501. This estimation is supported by sequence analysis of the IR of these transposons (Ohtsubo et al. 1978; Reed et al. 1979; Brown et al. 1980; De La Cruz et al. 1982; Hyde and Tu 1982). In evolution of transposons belonging to the Tn2613 family, the transposing ability had been successively reduced with increased size by insertion of additional DNA sequences (Tanaka et al. 1983). The resolvase of Tn501 has been reported to be induced by Hg 2+ (Kitts et al. 1982b). We examined whether the resolvase encoded by Tn2613 is induced by Hg 2÷. In Table 5, the transposition frequencies of Tn2613 onto R386 and R388 were measured in the absence of Hg 2÷. Transconjugants were picked up, and then examined for Km resistance. All transconjugants exhibited Km sensitivity, suggesting that the transconjugants were not due to plasmid mobilization. This result suggests that the level of resolvase production of Tn2613 is sufficient for resolution in the absence of Hg z+. When the transposition frequencies of Tn2613 onto R386 and R388 were measured in the presence of 6.3 gg/ml Hg 2+, the value was mostly identical with that in the absence of Hg 2+, as represented in Table 5. This suggests that the level of transposase production of Tn2613 was not affected by Hg 2+. Therefore, Tn2613 and Tn501 are considered to differ in the regulation of resolvase production. There may be also an evolutionary distance between both transposons. Since the resolvase production of Tn2603 was not also induced by Hg 2÷ (data are similar with Tn2613), it is probable that the Hg-noninducible production of resolvase is a common property among transposons in the Tn2613 family. Tn2603, Tn2613, and Tn501 encode transposase and resolvase that have a similar specificity of function. Therefore, if the Tn2613 family and Tn5Ol-type transposons were present in one replicon in direct orientations, a resolution would occur between each res site of both transposons. This event would generate a new hybrid transposon from both transposons. The hybrid transposon was described as the structure formed by 7~ and Tn3 (Reed 1981). We noticed that Tn2101 encoding Cb, Sm/Sp, Su and Hg resistance (Katsu et al. 1982; Katsu personal communication), has a restriction cleavage map identical to Tn2608 at the

449

left-hand portion including the Hg and Su resistance gene, but not similar on the right-hand side. On the basis o f this observation, we can estimate that Tn210l is a hybrid transposon formed between Tn2608-type and Tn5Ol-type transposons. Many multiple resistance transposons have been previously reported (Kopecko etal. 1976; Rubens et al. 1979a, b; Y a m a m o t o et al. 1981b, 1983; Medeiros et al. 1982; Katsu et al. 1982; Tanaka et al. 1983). These transposons are not identical in their phenotypes and genetic structures. However, some transposons are considered to be closely related and to have evolved from the same ancestral transposon (Tanaka et al. 1983). At present, the molecular mechanisms of evolution of th e multiple resistance transposons are not yet established, but we proposed a possibly evolutional strategy of transposons, Tn21, Tn2608, Tn4, and Tn2603, derived from Tn2613 (Tanaka et al. 1983). Grinsted et al. (1982) proposed the molecular model of formation o f multiple resistance transposon, such as Tn21 from the mercury resistance transposon. In addition to these models, we have postulated that one evoluationary driving force for multiple resistance transposons is formation of a hybrid transposon from two phenotypically different transposons mediated by the tnpR and res genes studied in this work. Also, we assume that homologous recombination by recA protein may involve the evolution of multiple resistance transposons. Recently, we observcd that the resolvase encoded by I R - R of Tn2610 encoding Cb, Sm/Sp, and Su resistance (Yamamaoto et al. 1983) could lead to resolution at res sites of Tn2603, but IR-L o f Tn2610 could not (unpublished data). This suggested that this resolvase may have the function of resolving a Tn2610-mediated cointegrate structure. However, this result suggests certain evolutional relationship between Tn2610 and Tn2603. Moreover, we have recently observed that the cloning of the res site of Tn2603 into pACYC184 activated the mobilization frequency of a cloned plasmid by a conjugative plasmid containing a Tn2613 family transposon (manuscript in preparation). However, this effect was only tenfold higher than that of pACYC184 cloning no res site. This result suggests that cointegrate formation between two plasmids mediated by resolvase and the res sites of Tn2603 can occur but this reaction has very low efficiency compared with that of resolution. Consequently, the resolvase of Tn2603 is considered to mediate two reactions, resolution and cointegration as a reverse reaction of resolution, resolution being the major role of resolvase rather than cointegration. We assume that this cointegrate formation may be concerned in plasmid mobilization and in cointegration of plasmid and other D N A rearrangements. References

Altenbucher, J, Choi C-L, Grinsted J, Schmitt R, Richmond MH (1981) The transposon Tn501 (Hg) and Tn1721 (Tc) are related. Genet Res 37:285-289 Arthur A, Sherratt D (1979) Dissection of the transposition process: a transposon-encoded site specific recombination system. Mol Gen Genet 175 : 267-274 Bennett PM, Grinsted J, Choi C-L, Richmond MH (1978) Characterization of Tn501, a transposon determining resistance to mercuric ions. Mol Gen Genet 159:101-106 Berg DE, Davis J, Allet B, Rochaix J-D (1975) Transposition of R factor genes to bacteriophage 2. Proc Natl Acad Sci USA 72: 3628-3632

Bolivar F, Rodriguez R, Greene D J, Betlach M, Heynecker HL, Boyer HW, Crosa J, Falkow S (1977) Construction and characterization of new cloning vehicles II, multiple purpose cloning system. Gene 2:95-113 Brown NL, Choi C-L, Grinsted J, Richmond MH (1980) Nucleotide sequence at the end of mercury transposon Tn501. Nucl Acids Res 8:1933-1945 Chandler M, Clerget M, Galas DJ (1982) The transposition frequency of ISl-flanked transposons is a function of their size. J Mol Biol 154:229-243 Chang ACY, Cohen SN (1978) Construction and characterization of amplifiable multicopy DNA cloning vehicles derived from the PI 5A cryptic miniplasmid. J Bacteriol 134:1141-1156 Clerget M, Chandler M, Caro L (1980) Isolation of an ISl-flanked kanamycin-resistance transposon from Rldrd-19. Mol Gen Genet 180:123-128 Datta N, Hedges RW (1972) Trimeth0prim resistance conferred by plasmids in Enterobacteriaceae. J Gen Microbiol 72: 349-356 De La Cruz F, Grinsted J (1982) Genetic and molecular characterization of Tn21, a multipler resistance transposon from RI00.1. J Bacteriol 151 : 222-228 Dennison S (1972) Naturally occurring R factor, derepressed for pillus synthesis, belonging to the same compatibility group as the sex factor F of Escherichia coli K-12. J Bacteriol 109:41~422 Galas DJ, Chandler M (1981) On the molecular mechanisms of transposition. Proc Natl Acad Sci USA 78:4858-4862 Gill R, Heffron F, Dougan G, Falkow S (1978) Analysis of sequences transposed by complementation of two classes of transposition-deficient mutant of Tn3. J Bacteriol 136:742-756 Gill R, Heffron F, Falkow S (1979) Identification of the protein encoded by the transposable element Tn3 which is required for its transposition. Nature (London) 282:797-801 Gottesman M, Rosener JL (1975) Acquisition of a determinant for chloramphenicol resistance by coliphage lambda. Proc Natl Acad Sci USA 72:5041-5045 Grinsted J, De La Cruz F, Altenbuchner J, Schmitt R (1982) Complementation of transposition of tnpA mutants of Tn3, Tn21, Tn50! and Tn1721. Plasmid 8:276-286 Guyer M (1978) The ~6 sequence of F is an insertion sequence. J Mol Biol 126:347-365 Heffron F, Bedinger P, Champoux J J, Falkow S (1977) Deletions affecting the transposition an antibiotic resistance gene. Proc Natl Acad Sci USA 74: 702-706 Hyde DR, Tu C-PD (1982) Insertion sites and the terminal nucleotide sequences of the Tn4 transposon. Nucl Acid Res I0:3981-3993 Katsu K, Inoue M, Mitsuhashi S (1982) Transposition of the carbencillin-hydrolyzing fl-lactamase gene. J Bacteriol 150: 483-489 Kitts PA, Lamond A, Sherratt DJ (1982a) Inter-replicon transposition of Tnl/3 occurs in two sequential genetically separable steps. Nature (London) 295:626-628 Kitts P, Symington L, Burke M, Reed R, Sherratt D (1982b) Transposon-specified site-specific recombination. Proc Natl Acad Sci USA 79:46-50 Kleckner N (1981) Transposable elements in prokaryotes. Ann Rev Genet 15: 341-404 Kleckner N, Chan RK, Tye B-K, Botstein D (1975) Mutagenesis by insertion of a drug-resistance element carrying an inverted repetition. J Mol Biol 97:561-575 Kopecko DJ (1980) Specialized genetic recombination systems in bacteria: their involvement in gene expression and evolution. In: Hahn F (ed) Progress in molecular and subcellular biology, vol 7. Springer, Berlin Heidelberg New York, p 135 Kopecko D J, Brevet J, Cohen SN (1976) Involvement of multiple translocating DNA segments and recombination hot spots in the structural evolution of bacterial plasmids. J Mol Biol 108:333-360 Medeiros AA, Hedges RW, Jacoby GA (1982) Spread of a "Pseu-

450 domonas-specific" fl-lactamase to plasmids of Enterobacteria. J Bacteriol 149:700-707 Nomura N, Yamagishi H, Oka A (1978) Isolation and characterization of transducing coliphage fd carrying a kanamycin-resistance gene. Gene 3:39-51 Novick RP, Clows RC, Cohen SN, Curtiss R, Datta N, Falkow S (1976) Uniform nomenclature for bacterial plasmids: a proposal. Bacteriol Rev 40:168-189 Ohtsubo H, Ohmori H, Ohtsubo E (1978) Nucleotide-sequence analysis of Tn3 (Ap): implications for insertion and deletion. Cold Spring Harbor Symp Quant Biol 43:1269-1277 Reed RR (1981) Resolution of cointegrates between transposons ~ and Tn3 defines the recombination site. Proc Natl Acad Sci USA 78 : 3428-3432 Reed RR, Young RA, Argentsinger S, Grindley NDF, Guyer M (1979) Transposition of the Escherichia coli insertion element ~ generates a five-base-pair repeat. Proc Natl Acad Sci USA 76: 4882-4886 Rownd RH, Easton AM, Barton CR, Womble DD, Mekell J, Sampathkumar P, Luckow VA (1980) Replication, incompatibility, and stability functions of R plasmid NR1. In: Alberts B (ed) ICN-UCLA Symposia on molecular and cellular biology, vol XIX. Academic Press, New York, p 311 Rubens CE, McNeill WF, Farrar WE (1979a) Transposable plasmid deoxyribonucleic acid sequence in Pseudomonas aeruginosa which mediates resistance to gentamicin and four other animicrobial agents. J Bacteriol 139:877-882 Rubens CE, McNeill WF, Farrar WE (1979b) Evolution of multiple-antibiotic-resistance plasmids mediated by transposable plasmid deoxyribonucleic acid sequences. J Bacteriol 140: 713--719 Schmitt R, Altenbuchner J, Grinsted J (1981) Complementation of transposition functions encoded by transposons Tn501 (Hg R) and Tn1721 (TetR). In: Levy SB, Clows RC, Loenic EL (eds) Molecular biology, pathogenicity, and ecology of bacterial plasmids. Plenum Publishing Corp, New York, p 359

So M, Heffron F, McCarthy BJ (1979) The E. coli gene encoding heat stable toxin is a bacterial transposon flanked by inverted repeats of IS/. Nature (London) 277:453-456 Tanaka M, Harafuji H, Yamamoto T (1982) A gene and its product required for transposition of resistance transposon Tn2603. J Bacteriol 151 : 723-728 Tanaka M, Yamamoto T, Sawai T (1983) Evolution of complex resistance transposons from an ancestral mercury transposon. J Bacteriol 153:1432-1438 Yamamoto T, Tanaka M, Baba R, Yamagishi S (1981 a) Physical and functional mapping of Tn2603, a transposon encoding ampicillin, streptomycin, sulfonamide, and mercury resistance. Mol Gen Genet 181:464-469 Yamamoto T, Tanaka M, Nohara C, Fukunaga Y, Yamagishi S (1981 b) Transposition of the oxaeillin-hydrolyzingpenicillinase gene. J Bacteriol 145:808-813 Yamamoto T, Tanaka M, Sawai T (1982a) Molecular properties of Tn2603, a transposon encoding ampicillin, streptomycin, sulfonamide and mercury resistance. In: Mitsuhashi S (ed) Drug resistance in bacteria. Japan Scientific Society Press, Tokyo, p 107 Yamamoto T, Yamagata S, Horii K, Yamagishi S (1982b) Comparison of transcription Of fl-lactamase genes specified by various ampicillin transposons. J Bacteriol 150:269-276 Yamamoto T, Watanabe M, Matsumoto K, Sawai T (1983) Tn2610, a transposon involved in the spread of the carbenicillin-hydrolyzingfl-lactomase gene. Mol Gen Genet 189:282-288

Communicated by F. Kaudewitz

Received April 5, 1983