Replication functions of new broad host range plasmids isolated from polluted soils

Research in Microbiology 154 (2003) 499–509 www.elsevier.com/locate/resmic

Replication functions of new broad host range plasmids isolated from polluted soils Marie-Eve Gstalder a,∗ , Michel Faelen a , Natacha Mine a , Eva M. Top b , Max Mergeay a,c , Martine Couturier a a Laboratoire de Génétique des Procaryotes, IBMM, Université Libre de Bruxelles, rue des Prof. Jeener et Brachet, 12, 6104 Gosselies, Belgium b Department of Biological Sciences, 353B Life Sciences building, University of Idaho, Moscow, ID 83844-3051, USA c Laboratoire de Microbiologie, Division Déchets radioactifs and assainissement, SCK/CEN, Boeretang, 200 2400-MOL, Belgium

Received 2 May 2003; accepted 12 June 2003 First published online 13 June 2003

Abstract The nucleotide sequencing of replicons isolated from three new broad host range plasmids, pMOL98, pEMT8, and pEMT3, originating from polluted soils, showed a typical organization of iteron replicons replicating by the theta mode. In the pMOL98 replicon, the origin region and the rep gene were identified in complementation experiments. Sequence comparisons showed that the regions bearing these features are highly identical to regions in pIP02T and pSB102 and that the Rep proteins (but not the origin regions) of these three plasmids show some identity to the Rep proteins of the IncW group of plasmids. This suggests that pMOL98, pIPO2T, and pSB102 constitute a new Inc/Rep family, distantly related to the IncW group, but having an incompatibility phenotype different from the IncW phenotype. The pEMT8 replicon displayed an orf whose conceptually translated product is related to the Rep proteins of four plasmids, pSD20, pSW500, pMLb, and pALC1, not yet classified into any known incompatibility group. The vegetative origins of these plasmids were not similar, suggesting that the five plasmids could belong to a new family with similar Rep proteins but different incompatibility phenotypes. The pEMT3 replicon is clearly related to IncP replicons (sequence similarities and incompatibility phenotype), although sequence comparisons revealed some divergence with respect to the two well-documented subgroups IncPα and IncPβ. This suggests that in these plasmids, despite the existence of a powerful system of centralized control over replication, maintenance, and transfer functions, plasticity and evolution of these functions are at work. Our analysis confirms the extreme genetic flexibility of plasmids and the absolute necessity of using multiple techniques (PCR, DNA sequencing, DNA chips, and databases) to analyze the role of broad host range plasmids in the capture, recombination and spread of genetic traits among bacteria.  2003 Éditions scientifiques et médicales Elsevier SAS. All rights reserved. Keywords: Broad host range plasmid; Replication; Incompatibility

1. Introduction The importance of horizontal transfer in the evolution and diversification of bacteria is now well established. Plasmids play a primordial role in these exchanges. Through their conjugative transfer, mobilization, or retro-mobilization, they enable bacteria to acquire a wide variety of new genetic traits. These traits can be captured by plasmids from the chromosome as a result of transposition events or via plasmid integration into and imprecise excision from the host * Corresponding author.

E-mail address: [email protected] (M.-E. Gstalder).

chromosome [1,51]. Plasmids provide their hosts with a large array of phenotypes: resistance to antibiotics [11] or heavy metals [2,14,38,46], the ability to degrade recalcitrant compounds and use them as carbon and energy sources [13,17,60], and other traits offering a selective advantage to the host cell and allowing it to colonize specific biotopes. Broad host range (BHR) plasmids can transfer (or be mobilized) to and stably maintain the genes they carry in taxonomically distant species [54]. As such, they represent powerful vectors of recombinant DNA, promoting its spread in the environment. To be relevant, evaluation of the risk associated with the use of genetically modified organisms will thus require full understanding of

0923-2508/$ – see front matter  2003 Éditions scientifiques et médicales Elsevier SAS. All rights reserved. doi:10.1016/S0923-2508(03)00143-8

500

M.-E. Gstalder et al. / Research in Microbiology 154 (2003) 499–509

the distribution, maintenance, recombination, and conjugation of BHR plasmids. Yet the prevalence and diversity of these plasmids in the environment has not been studied in depth. In order to isolate, characterize, and identify BHR plasmids, we need to define the notion of “broad host range”. Classifying plasmids became essential after the discovery of R-plasmids in clinical bacterial isolates [11,36]. Classification was originally based on the incompatibility property, i.e., the inability of certain plasmids (now known to share common genetic determinants involved in the control of replication and partition) to be propagated in the same cell line. Two incompatible plasmids are placed in the same incompatibility group. Plasmids from enteric bacteria and Pseudomonas (which both belong to the γ-Proteobacteria) presently fall into 30 and 14 incompatibility groups respectively. Plasmids belonging to four of these groups (IncP, W, N, and Q) all transfer to and maintain themselves in both enteric bacteria and Pseudomonas (at the least). In most cases, transfer to other species has not been tested. It is hard to determine the host range of a plasmid, as transfer can be tested experimentally only between hosts representing a minute fraction of the bacterial world. We therefore need criteria establishing the phylogenetic distance that a plasmid should be able to cover to be considered a BHR plasmid. Szpirer et al. [51] have proposed the following experimentally manageable criterion: plasmids that can transfer between at least two branches of the Proteobacteria should be viewed as BHR plasmids. The study of movements of BHR plasmids in nature requires a suitable identification test. Plasmids lack a single ubiquitous molecule, analogous to that of the 16S rRNA gene, for identification. They consist of a variable assortment of gene cassettes involved in maintenance, transfer, and phenotypic expression. The only genetic traits shared by all plasmids are the genes involved in replication and its control. These genetic traits are therefore the most suitable for identification. Detailed studies of the mechanism and regulation of replication of various plasmids isolated from clinical samples have led to designing specific molecular probes for incompatibility determinants (Rep probes). These “replicon” probes detect basic replicons and this so-called replicon typing, based on DNA–DNA hybridization [8], greatly facilitates plasmid identification and classification. New techniques such as biparental, triparental, and retrotransfer matings have allowed the isolation of new BHR plasmids, especially from non-clinical environments such as soil, marine sediments, and water [9,10,33,57,58]. This has led to the characterization of more plasmid-borne phenotypic traits and replication and transfer functions, emphasizing the mosaic nature of plasmids (reviewed in [39]). Götz et al. [21] have developed primers and probes for detecting transfer and replication genes of IncP, IncN, IncQ, and IncW BHR plasmids in manure slurries [21] or marine sediments [50]. A large proportion of the plas-

mids isolated from these environments bear replication and incompatibility regions unrelated to known groups, suggesting that plasmids from clinical isolates represent only a subset of plasmid diversity [49]. This justifies further evaluation of the diversity and evolution of plasmids. We have sequenced and analyzed the replication functions of three new plasmids: pMOL98, pEMT8, and pEMT3. The first was isolated from hydrocarbon-polluted soil by the triparental exogenous plasmid isolation method, which selects plasmids on the sole basis of their BHR property [57]. The other two were isolated by the biparental exogenous isolation method, from agricultural soil treated with 2,4-dichlorophenoxyacetic acid (2,4-D) for many years (pEMT3, [58]) or for 18 days (pEMT8, [59]) before plasmid isolation. This latter method selects for mobilizable or conjugative plasmids with a given selectable marker and does not necessarily require the plasmid to be BHR. Nevertheless, several plasmids selected for their degradation capacity have turned out to be BHR. This applies notably to pEMT3 and pEMT8, which were checked for their ability to invade strains of α-, β-, and γ-proteobacteria ([58], Gstalder, Thesis, 2001). First analyses showed that pMOL98 and pEMT8 did not hybridize with any of the available replicon-specific probes used in replicon typing experiments and that pEMT3 hybridized with a specific IncP probe only at low stringency. In addition, the latter plasmid’s host range appeared more restricted than that of reference IncP plasmids. For these reasons, we undertook to analyze the replication functions of these three low-copy-number BHR plasmids.

2. Materials and methods 2.1. Bacterial strains, plasmids, and culture conditions Bacteria and plasmids are listed in Table 1. Bacterial cultures were grown in LB (10 g Bactotryptone, 5 g yeast extract, and 5 g NaCl per liter). Unless otherwise specified, all bacterial strains were grown at 30 ◦ C with good aeration. When required, kanamycin, tetracycline, and/or ampicillin was/were added (final concentration: 50, 20, and 100 µg/ml respectively). For R. metallidurans cultures, 1000 µg/ml kanamycin was used. 2.2. Plasmid DNA isolation and analysis Plasmid DNA was isolated by the classical alkaline lysis method [6], using Quiagen kits as recommended by the manufacturer, or by the method described by Kado and Liu [27]. All enzymes were used according to the suppliers’ instructions. Standard protocols [43] were followed for ligation of restriction fragments and agarose gel electrophoresis.

M.-E. Gstalder et al. / Research in Microbiology 154 (2003) 499–509

501

Table 1 Bacterial strains and plasmids (natural or vector) Strains

Genotypes/chromosomic markers

Escherichia coli DH5lac E. coli S17-1 E. coli HMS50 Ralstonia metallidurans AE 110

deoR, thi1, relA1, supE44, endA1, gyrA96, recA1, hsdR17, (argF,lac)U169 NalR pro, thi1, supE44, endA1, hsdR17, recA, RP4-2-Tc::Muc+ Kan::Tn7 StrR , TmpR

Plasmids pMOL98

Description Isolated by triparental mating, cryptic, pES1 tagged by mini-Tn5-Km1

pEMT8

Isolated by biparental mating, containing the first gene (tfdA) of 2,4-D degradative pathway; tagged by mini-Tn5-Km1 Isolated by biparental mating, containing initially the 2,4-D degradative pathway genes tfd, but the degradative capacity has disapeared; tagged by mini-Tn5-Km1 IncPα, AmpR , TetR , KanS IncPβ, TmpR repColE1, AmpR , pLac::ccdB repP15A, KanR , Mob+ , pLac::ccdB reppBBR1, KanR , Mob+ , pLac::ccdB

pEMT3 RP4KanS R751 pKIL18/19 pKIL194T pBHRK19

thyA, thi, endA, polA1 NalR mutant of the plasmid free derivative of the metal-resistant wild type strain CH34

References [22] [47] Molineux [35] References [57], Wilmotte, personnal communication

2.3. Bacterial matings

3. Results

Donor and recipient strains were grown overnight in LB. The donor and recipient were mixed (100 µl of each culture) and incubated overnight on LB plates at the desired temperature. Cells recovered from the plates were resuspended in 1 ml 10 mM MgSO4 , and appropriate dilutions were spread on plates containing different selective media in order to titrate transconjugants and parental cells of both types [32].

3.1. Construction of DNA libraries and isolation of replication functions from plasmids pMOL98, pEMT8, and pEMT3

2.4. Replicon typing and DNA–DNA hybridization experiments Probes for replicon typing were isolated by digestion with appropriate restriction enzymes as described [8] and purified twice on agarose gels. DNA fragments were labeled with digoxygenin by means of the “Dig-High prime kit”. Agarose gels were blotted onto nylon membranes using a standard protocol [43], and the transferred plasmid DNA was hybridized with digoxygenin-labeled probes according to the manufacturer’s instructions. 2.5. DNA sequencing The nucleotide sequences of pMOL98, pEMT8, and pEMT3 were sequenced by gene walking, using the “ABI Prism 310 Genetic Analyzer Perkin–Elmer Applied Biosystem” with the “BigDyeTM Terminator Cycle Sequencing Ready Reaction kit V2.0 Abiprism”. Fasta and Blast programs from NCBI were used to search for homologous sequences [41], and the “promoter predictor” program to predict potential promoters. Clone Manager 4.0 software was used to detect potential open reading frames.

[59] [58] [62] [25] [4] [18] [18]

Different strategies were used to isolate the replication functions of pMOL98, pEMT8, and pEMT3. Their sensitivity to only a few restriction enzymes, the large size of their replicons, or dispersion of the different regions essential to replication (e.g., IncP) may explain the difficulty of isolating the entire replicon of large BHR plasmids. 3.1.1. pMOL98 library and mini-plasmids DNA of plasmid pMOL98 was digested with EcoRI. The EcoRI fragments were cloned in the pKIL18 vector [4] and their extremities were sequenced (pMOL98 library). Independently, the EcoRI restriction mix was used directly for ligation. The ligated DNA was introduced into Escherichia coli DH5lac by electroporation and transformants resistant to kanamycin were selected: these were expected to carry the pMOL98 kanamycin resistance gene associated with a functional replicon. The shortest kanamycinresistant construct found, named pMEG1, consisted of two EcoRI fragments of, respectively, 2.3 and 9.3 kb. Hybridization experiments showed that the 2.3-kb fragment carried the mini-Tn5Kn1 transposon. The 9.3-kb EcoRI fragment of pMEG1 was sequenced by gene walking. Independently, A. Wilmotte has constructed a mini-plasmid by ligating a partial Sau3A digest of pMOL98 to the kanamycin resistance GenBlock of pUC4K (unpublished results). Restriction analysis of this 17-kb mini-pMOL98 indicated that a 12.5-kb segment of the plasmid corresponded to the 9.3-kb EcoRI fragment found in pMEG1 preceded by the 0.6- and 2.6-kb EcoRI fragments of our

502

M.-E. Gstalder et al. / Research in Microbiology 154 (2003) 499–509

library. This enabled us to reconstitute the DNA sequence of a 10.2-kb segment of this 12.5-kb region, accessible in the Genbank database under accession No. AJ519490. Finally, pMEG1 was partially digested with EclXI and the restriction mixture immediately religated and used to transform DH5lac. A 5.2-kb autoreplicative kanamycinresistant construct, pMEG2, was obtained. Restriction analysis of this mini-plasmid revealed the presence of the 1.8kb mini-Tn5-Kn1 associated with a 3.4-kb fragment. The 3.4-kb pMEG2 sequence is accessible under accession No. AJ345055. 3.1.2. pEMT8 library DNA of pEMT8 was digested with SacI. The fragments obtained were cloned in the positive selection vector pKIL194T [18]. The ligated DNA was introduced by electroporation into E. coli strain S17-1. This strain contains an integrated RP4 that can mobilize pKIL194T. Transformants were pooled and the mixture was grown and mated with R. metallidurans AE110. Because the pKIL194T vector is unable to replicate in this strain, selection of KanR transconjugants should allow selective recovery of plasmid constructs containing a cloned functional BHR replicon. A 14.8-kb plasmid (pMF100) was isolated, consisting of four SacI fragments of, respectively, 2.3, 3.2, 4.8, and 4.5 kb, the latter being the pKIL194T vector (as revealed by hybridization and endonuclease mapping experiments, data not shown). pMF100 was further digested with BglII, and the restriction mixture was immediately religated. An 8.9-kb autoreplicative, kanamycin-resistant mini-plasmid, pMF101, was recovered after introduction into AE110 by electroporation. Restriction endonuclease mapping showed that pMF101 did not carry the vector’s replication function and contained a 2.6-kb SacI–BglII fragment from pKIL194T, a 1.7-kb SacI–SphI fragment corresponding to the mini-Tn5Kn1, and a 4.6-kb SphI–BglII fragment of pEMT8, which ought to provide the functional replicon. We sequenced the 4.6-kb SphI–BglII fragment by gene walking. The resulting sequence was deposited in the Genbank database under accession No. AJ416427. 3.1.3. pEMT3 library Several libraries of pEMT3 DNA were generated, using different restriction endonucleases. This enabled us to identify one fragment containing the replication origin and a gene coding for the replication initiation protein. An 8-kb plasmid, pMEG100, expressing kanamycin resistance and containing the origin region, was isolated from a Sau3AI partial digest in the positive selection vector pKIL194T cleaved with BamHI. It was used to transform strain S17-1 and recovered after selection for KanR transformants. Plasmid pMEG100 could be maintained only in strains containing an IncP plasmid, either integrated into the chromosome as in S17-1 or extrachromosomal as in DH5lac/RP4. Hybridization experiments showed that the kanamycin resistance determinant originated from

the mini-Tn5-Kn1 carried by pEMT3 and not from the vector (data not shown). A plasmid conferring incompatibility towards pEMT3 was then obtained by cloning a 2400-bp PstI–SphI fragment of pMEG100 into pKIL19. This fragment was sequenced. A second region essential to pEMT3 replication was isolated from a PstI-library of pEMT3 DNA cloned in pKIL194T. The resulting 17-kb plasmid pMEG101 complemented pMEG100 and allowed its maintenance even in strains lacking an IncP helper plasmid or the parental pEMT3. Plasmid pMEG101 was shortened by EcoRI and HpaI restriction, yielding a 7.3-kb plasmid conferring kanamycin resistance, pMEG101B. The sequences of the pEMT3 replication elements are available on database under accession numbers AJ414161 and AJ414162. 3.2. Analysis of the autonomously replicating region of pMOL98 3.2.1. Sequence analysis The pMOL98 replicating region displays significant sequence identity to the corresponding regions of the pIPO2T (84%) and pSB102 plasmids (57%) [45,53]. Using the “Clone Manager 4.0” software, we identified ten potential orf s, (orf 1 to orf 10) on the 10.2-kb sequence (Fig. 1). These orf s and their putative products are listed in Table 2. One open reading frame, orf 3, spans nt 1793 to nt 3352. Significant identity was found between the corresponding 519-aa polypeptide and the putative replication proteins of the pIPO2T (91.1% identity) and pSB102 (62.2%) plasmids. Fig. 2 shows the alignment of the three proteins. The proteins were identified by database comparisons and show some degree of identity to several IncW-plasmid Rep proteins [45,53]. Sequence comparisons revealed significant identity between pMOL98 Rep and the Rep proteins encoded by the following plasmids: E. coli IncW plasmid pSa [37] (38% identity), Rhodothermus marinus plasmid pRM21 (accession No. NP_044344) (35% identity), Xanthomonas campestris pv. vesicatoria plasmid pXV2 [63] (34% identity), Bifidobacterium asteroides plasmid pAP1 (accession No. CAA72313) (29% identity), Azospirillum brasilense plasmid p90 [61] (27% identity), Chlorobium limicola plasmid pCL1 (accession No. NP_052164) (26% identity), Corynebacterium striatum plasmid pTP10 [52] (24% identity), Rhodopseudomonas palustris pMG101 [24] (24% identity), and Sphingomonas aromaticivorans pNL1 [42] (24% identity). Thus, orf 3 may code for a replication protein. It was tentatively named repA. All these plasmids are related to IncW plasmids by sequence similarities between their replication initiation proteins, but none has been tested for incompatibility. Four other orf s (orf 4, 8, 9, and 10) were named according to similarities between their putative products and previously characterized IncP proteins: ParB, KfrA, Ssb, and KorA, respectively. The organization of these orf s is the same in

M.-E. Gstalder et al. / Research in Microbiology 154 (2003) 499–509

503

Fig. 1. Schematic diagram of the structural organization of pMOL98 replication functions as deduced from the nucleotide sequence of a 10.2-kb region (accession No. AJ519490). The genes with their names are represented by large arrows. The interesting restriction sites are indicated (only one of the EclXI sites is located). The DR1, DR2, and DR3 boxes (gray boxes) have sequences identical to repeat regions detected in the pIPO2T sequence. The origin region shown to be essential is represented by a box and is enlarged below. In the enlargement, repeats are marked by gray arrows, DnaA-boxes by black boxes, and the IHF-binding site by an open circle. Table 2 orf s of pMOL98 and identities with pIPO2T and pSB102 sequences orf s

Position

Genes

orf 1 orf 2 orf 3

277−630 605−1585 1793−3352

orf 1 orf 2 repA

117 326 519

3352−3816 4476−6002 6270−662 6690−6980 7083−8117 9680−10039 10043−10262

parB orf 5 orf 6 orf 7 kfrA ssb korA

154 508 130 96 344 119 74

orf 4 orf 5 orf 6 orf 7 orf 8 orf 9 orf 10*

Protein Length Size (aa) (kDa)

Comparison with pIPO2T orf s Amino acid identities (%)

Comparison with pSB102 orf s Amino acid identities (%)

13.1 36.9 57.1

orf 33 orf 32 repA

96.6 94 91.1

orf 68 orf 69 repA

61.1 62.2 62.2

16.9 55.4 14.6 10.5 36.8 13.2 –

parB orf 29 orf 28b orf 28a orf 27 ssb orf 25

95.4 96.4 97.7 99 87.5 98.3 95.9

– orf 4 orf 5 orf 6 – ssb korA

– 67.5 62.5 33.3 – 75.4 64.5

Remarks

Other putative 466 bp orf (cf. discussion)

Truncated by construction

* The sequence of a pMOL98 region isolated by A. Wilmotte (unpublished result) indicates that orf 10 must encode a 132-aa protein showing 97.7% and 70.1% identity to pIPO2T ORF10 and pSB102 KorA, respectively.

pMOL98, pIPO2T, and pSB102, except that parB and kfrA are absent in pSB102. A region with features expected of a vegetative origin was detected 5.2 kb downstream from the rep gene. It contains four 19-bp repeats with a consensus sequence ACGCTGAAASTGTCTTGCS. A 27-bp AT-rich region (86% AT) adjacent to a 29-bp GC-rich region (73% GC) was detected near the first iteron cluster. Three potential DnaA-boxes [44] and one potential IHF site [20] were found in that same region. Despite this organization, no similarity was found between sequenced IncW plasmid origins and the putative pMOL98 origin. On the other hand, nearly identical putative ori regions were found in pMOL98, pIPO2T, and pSB102 (Fig. 2).

3.2.2. Complementation experiments A 553-bp PvuI–EclXI fragment (nt 8489 to 9041 of the 10.2-kb sequence) containing the putative origin of replication was cloned in the pKIL19 vector. In parallel, a 3184-bp EcoRI–NotI fragment of pMEG1 was inserted into the pBHRK19 vector (pBHRK19::rep), so that the repA gene of pMOL98 was placed under the control of the pLac promoter. Both constructs were introduced into an E. coli PolA− strain. Plasmid pKIL19 is derived from ColE1 and is unable to replicate in a PolA− background (absence of PolA), but we observed that the pKIL19 construct with the 553-bp insert could be maintained in the PolA− background, provided the strain was synthesizing pMOL98 RepA from pBHRK19::rep. This demonstrates that the 553-bp region

504

M.-E. Gstalder et al. / Research in Microbiology 154 (2003) 499–509

Fig. 2. Alignment of the pMOL98 replicative origin with the putative oriV regions of pIPO2T and pSB102. The aligned sequences span nt 8534 to nt 8731 of the pMOL98 sequenced region (accession No. AJ519490), nt 6969 to nt 7152 of the pSB102 sequence (accession No. AJ304453) and nt 6602 to nt 6796 of the pIPO2T sequence (accession No. AJ297913). Iterons are underlined, DnaA boxes are indicated by bold italic letters, putative IHF binding sites are indicated by open boxes.

functions as a replication origin in the presence of the pMOL98 RepA protein. A search for a putative promoter region with the “promoter predictor” program identified potential −35 and −10 boxes between nt 198 and nt 226, upstream from the EcoRI site preceding the orf 1, orf 2, and repA genes. Similar features have been detected at equivalent positions in pIPO2T and pSB102. Complementation experiments similar to the experiment described above were performed with a pBHRK19::rep construct bearing a terminator cassette (omega cassette) at the EcoRI site, upstream from the rep gene. Since this plasmid did not allow replication of the pKIL19-derived plasmid containing the pMOL98 ori region, we conclude that no promoter is present in the pMOL98 sequence between the EcoRI site and the repA start codon (data not shown). Transcription of repA probably proceeds from the hypothetical promoter mentioned above. 3.3. Sequence analysis of pEMT8 replication functions Only one putative gene was detected on the 4.1-kb sequenced fragment of pMF101. It spans nt 2613 to 3563, starts with an ATG codon, and should code for a 317-aa protein. Subcloning experiments showed that this orf is essential to replication. The translated peptide shows some identity to replication proteins encoded by Ruegeria sp. PR1b plasmid pSD20 (42%) (accession No. AF416330), Paracoccus alcaliphilus plasmid pALC1 (40%) [3], Erwinia stewartii plasmid pSW500 (34%) [16], and Mesorhizobium loti plasmid pMLb (35%) [28]. These proteins show no homology with other known proteins. This pEMT8 orf was thus tentatively called rep. A second 450-bp sequence essential to replication was detected about 2 kb upstream from this rep gene. It contains seven 17-bp direct repeats (iterons R1 to R7) with the consensus sequence BTCGTAYAWCRSCGAYR, two DnaA-boxes, and a 27-bp AT-rich sequence followed by a 31-bp GC-rich sequence (Fig. 3). No similarity

Fig. 3. Organization of pEMT8 replicon and partial sequences. (A) Schematic diagram of the structural organization of the pEMT8 replication function as deduced from its nucleotide sequence. The figure shows the 4117-bp SacI–BglII fragment of pEMT8 (accession No. AJ416427). The gene with its name is represented by large arrows. Promoters are indicated by open boxes. Origin regions shown to be essential are underlined. Repeats are marked by gray arrows, DnaA-boxes by dark gray boxes, and the IHF-binding site by an open circle. (B) Nucleotides 751 to 1250: the origin region. Repeats are underlined by arrows, DnaA boxes are indicated by bold letters, GC- and AT-rich regions are indicated by a dotted frame.

was detected, however, between this sequence and the sequences around the origins of replication of the four other plasmids.

M.-E. Gstalder et al. / Research in Microbiology 154 (2003) 499–509

505

Upstream from the rep gene, two putative promoter regions were detected. We called them P1 and P2. Located, respectively, 1.2 kb and 0.35 kb from rep, they display typical sequences of −35 and −10 boxes. Three other iterons, R8, R10, and R9, flank the rep gene. Three 16-bp repeats, RI, RII, and RIII, significantly similar to iterons R1 to R10, were found just upstream from the tentative P2 promoter. RI and RIII are in opposite orientation compared to R1-10 and RII. RII and RIII might thus form a hairpin structure. 3.4. Analysis of the pEMT3 replicon and comparison with other IncP plasmids Analysis of the pMEG100 sequence revealed a 450-bp region 83% identical to the replication origin of IncPβ plasmid R751 and 54% identical to that of the IncPα plasmid RP4. The region contains eight 15-bp repeats with the consensus sequence TRACACTTRAGGSVS (organized in two clusters of respectively five and three repeats), four putative DnaA-boxes, and one potential IHF binding site directly followed by an AT-rich region. This closely resembles the pattern seen in the IncPα RP4 [40] and IncPβ R751 [56] plasmids. The 3.1-kb insert of pMEG101B was found to contain two orf s. The first is similar to the trfA genes of R751 (76% identity) and RP4 (65% identity), while the second is highly similar to the ssb genes of these same two plasmids (respectively, 91% and 80% similarity). No characteristic −10 and −35 boxes were detected upstream from these genes, but upstream from the ssb gene, at nucleotides 43 to 55, we detected a sequence (TTTAGCSGCTAAA) specifically recognized by KorB, the main effector of the global regulation of IncP-plasmid maintenance and transfer systems [26]. In addition, a ninth repeat homologous to the iterons located at the pEMT3 origin was identified downstream from the putative trfA gene (Fig. 4).

Fig. 4. Organization of the pEMT3 replicon and partial sequences. (A) Schematic diagram of the structural organisation of pEMT3 replication. The figure shows the 1898-bp fragment of pEMT3, containing ssb and trfA (accession No. AJ414161) and the 1017-bp fragment of pEMT3 containing oriV (accession No. AJ414162). The genes with their names are represented by large open arrows. Promoters are indicated by open boxes. Origin regions shown to be essential are underlined. Repeats are marked by gray arrows, DnaA-boxes by dark gray boxes, and the IHF binding site by an open circle. (B) Origin region (nucleotides 1973 to 2872 of sequence No. AJ414162): repeats are underlined by arrows, bold italic letters represent DnaA boxes, open boxes indicate putative IHF binding sites and GCand AT-rich regions are indicated by a dotted frame.

4. Discussion All three plasmids studied here have structural features and a general organization characteristic of “iteron plasmids” [12,15,23]. The analysis of these DNA sequences suggests that each of these replicons contains two regions essential to replication. The first, spanning about 500 bp in all three plasmids, contains clusters of repeat sequences, DnaA boxes, AT- and GC-rich regions, and potential IHF binding sites, and thus all the features expected in an origin of replication (double helix opening site). Repeated sequences at plasmid origins, i.e., iterons, are specific binding sites for Rep proteins. On the basis of this pattern, pMOL98, pEMT8, and pEMT3 thus appear as plasmids with a theta mode of replication regulated by iterons. The second region required for replication consists of an orf coding for a polypeptide

significantly similar to Rep replication initiation proteins and hence called rep. The pMOL98 autoreplicative region sequenced here is highly identical to the corresponding regions of pIPO2T and pSB102. Since the iterons of these three plasmids are nearly identical and since their putative RepA proteins are more than 60% identical, they may constitute a new Rep/Inc family. The putative Rep proteins of these three plasmids are also somewhat identical to the Rep proteins of the IncW group of plasmids, but the observed identities do not lie in the origin region. This suggests that all these plasmids share a common ancestor but have evolved into different incompatibility groups, one of which includes pMOL98, pIPO2T, and pSB102.

506

M.-E. Gstalder et al. / Research in Microbiology 154 (2003) 499–509

Fig. 5. Alignment of the pMOL98, pIPO2T, and pSB102 putative Rep proteins. The Clustal W program [55] was used to align the putative Rep proteins of the three plasmids. Amino acid identities and similarities are indicated, respectively, by a black or gray background.

The longest orf that might correspond to the rep gene is 1560 bp long and encodes a 519-aa protein. Sequence analysis of this presumptive rep orf revealed different possible sizes for the putative Rep. As shown in Fig. 5, identity between the Rep proteins of pIPO2T and pMOL98 concerns the 466 C-terminal amino acids of the pMOL98 protein. In addition, identity between these two Rep proteins and their analogue in pSB102 is confined to the proteins’ 308 C-terminal amino acids. Results obtained in this study with the pMEG2 plasmid indicate that a protein possessing only the 342 C-terminal amino acids of the 519-aa putative Rep protein is functional in E. coli. This observation might reflect that the size prediction for the pMOL98 Rep protein (519 aa) is an overestimate or that the size of this protein is variable and determined by the host carrying the plasmid, as observed for IncP plasmids [40]. The same might apply to the presumptive Rep proteins of pIPO2T (462 aa) and pSB102 (484 aa). In addition to the above-mentioned Rep similarities (at DNA and predicted protein levels) between pMOL98, pIPO2T and pSB102, we found a putative promoter upstream from the rep gene and located nearly at the same position in all three plasmids. The Rep protein of the mini-pEMT8 replicon is significantly identical to the Rep proteins of four plasmids that have not yet been classified in any known Rep/Inc group: pSD20, pSW500, pMLb, and pALC1. Yet the origin regions of these plasmids show no significant identity, suggesting that they may belong to different Inc groups evolved

from a common ancestor. In pEMT8, we detected additional repeat sequences similar to those of the origin and organized as inverted repeats, in or around the consensus −35 and −10 boxes of the putative rep gene promoter sequences. These inverted repeats might generate a secondary structure that could be stabilized by binding of a Rep protein dimer, this leading to autoregulation of rep gene transcription as observed for other replicons [5,7,19,29,34]. The existence of Rep autoregulation has not been investigated in pSD20, pMLb, or pALC1, but in pSW500 the rep promoter is located within the origin of replication [16]. Two iterons of pSW500 lie between the promoter and the rep start codon and a third one iteron is present between the −35 and −10 boxes of the promoter. Binding of RepA to this iteron represses expression or the rep gene. As predicted on the basis of replicon typing experiments, pEMT3 Rep is similar to IncP Rep proteins, and especially to those of IncPβ plasmids. Yet unlike previously described plasmids of this family, pEMT3 lacks, between the rep gene and the origin, a (usually well conserved) sequence containing five repeats distinct from the origin iterons and bearing PstI, SalI and XhoI restriction sites and a transcription terminator [48]. Instead, the pEMT3 replicon contains a ninth iteron just downstream from the rep gene, as usually seen in IncPα plasmids. pEMT3 might thus be a representative of a third class of IncP plasmids. In IncP plasmids, expression of the rep gene is subject to the general co-regulation of maintenance and transfer genes. The operators/promoters of these genes (TTTAGCSGCTAAA) are recognized by the

M.-E. Gstalder et al. / Research in Microbiology 154 (2003) 499–509

Kor A and/or Kor B protein(s) encoded by the central control operon [26,30,31]. As in other IncP plasmids, the ssb and rep genes of pEMT3 seem to form an operon preceded by a KorB binding site. The replication functions of pEMT3 might thus be regulated like those of other IncP plasmids. Although limited, our analysis further illustrates the extreme genetic flexibility of plasmids and raises general comments: (i) available probes are not sufficient for detecting plasmids that diverge to some extent from reference plasmids of known Inc/Rep groups (pMOL98, pIPO2T, and pSB102, not recognized by the RepW probe, might nevertheless be related to the RepW replicon family, according to their Rep sequences); (ii) the available Rep-probe collection needs to be expanded to include probes recognizing new replicon families (the pEMT8 replicon, resembling that of three other new plasmids, should be viewed as belonging to a new replicon group); (iii) divergences within a replicon group can escape notice when plasmid identification is based solely on replicon typing with Rep probes (pEMT3, classified as an IncP/RepP plasmid, appears to represent a third subgroup, distinct from IncPα and IncPβ). In conclusion, a prerequisite to understanding the role of BHR plasmids in the capture, recombination, and spread of genetic traits among bacteria is to develop a flexible approach. To reach this goal, it will be important to identify not only new plasmid backbones (replication/maintenance genes and transfer genes) but also more accessory genes involved in plasmid phenotypic expression. PCR technologies, DNA sequencing, DNA chips, and databases should provide the appropriate tools for assessing the enormous diversity of plasmids. They should help field microbiologists learn more about the behavior of plasmids in nature and their ecological impact.

Acknowledgements We are especially grateful to A. Vuye and L. Wacheul for their participation in sequencing of these plasmids, and to A. Wilmotte for providing DNAs and unpublished results. This work was supported by a grant from the European Union to E.T. and M.M. (BIOTECH BIO2-CT92-0491), by the Convention ULB-SCK/CEN, by the Actions de Recherche Concertée, by the European Science Foundation (Plasmid Network), an EU concerted action (MECBAD, BIO4980099), and the Van Buuren Foundation. M.C. is “Chercheur Qualifié” of the FNRS.

507

References [1] E. Ayres-Sia, D.M. Kuenher, D.H. Figurski, Mechanism of retrotransfer in conjugation: Prior transfer of conjugative plasmid is required, J. Bacteriol. 178 (1996) 1457–1464. [2] M.J. Bale, J.C. Fry, M.J. Day, Plasmid transfer between strains of Pseudomonas aeruginosa on membrane filters attached to river stones, J. Gen. Microbiol. 133 (1987) 3099–3107. [3] D. Bartosik, M. Witkowska, J. Baj, M. Wlodarczyk, Characterization and sequence analysis of the replicator region of the novel plasmid pALC1 from Paracoccus alcaliphilus, Plasmid 45 (2001) 222–226. [4] P. Bernard, P. Gabant, E.M. Bahassi, M. Couturier, Positive-selection vectors using the F plasmid ccdB killer gene, Gene 148 (1994) 71–74. [5] F. Bex, P. Pierard, A. Desmyter, P. Dreze, M. Colet, M. Couturier, Mini-F E protein: The carboxy-terminal end is essential for E gene repression and mini-F copy number control, J. Mol. Biol. 189 (1986) 293–303. [6] H.C. Birnboim, J. Doly, A rapid alkaline extraction procedure for screening recombinant plasmid DNA, Nucleic Acids Res. 7 (1979) 1513–1522. [7] D.K. Chattoraj, R.J. Mason, S.H. Wickner, Mini-P1 plasmid replication: The autoregulation-sequestration paradox, Cell 52 (1988) 551– 557. [8] M. Couturier, F. Bex, P.L. Bergquist, W.K. Maas, Identification and classification of bacterial plasmids, Microbiol. Rev. 52 (1988) 375– 395. [9] C. Dahlberg, M. Hermansson, Abundance of Tn3, Tn21 and Tn501 transposase (tnpA) sequences in bacterial community DNA from marine environment, Appl. Environ. Microbiol. 61 (1995) 3051–3056. [10] C. Dahlberg, C. Linberg, V.L. Torsvik, M. Hermansson, Conjugative plasmids isolated from bacteria in marine environments show various degree of homology to each other and are not closely related to wellcharacterised plasmids, Appl. Environ. Microbiol. 63 (1997) 4692– 4697. [11] N. Datta, V.M. Hughes, Plasmids of the same Inc groups in Enterobacteria before and after the medical use of antibiotics, Nature 306 (1983) 616–617. [12] G. del Solar, J.C. Alonso, M. Espinosa, R. Diaz-Orejas, Broad-hostrange plasmid replication: An open question, Mol. Microbiol. 21 (1996) 661–666. [13] H. De Rore, K. Demolder, K. De Wilde, E. Top, F. Houwen, W. Verstraete, A transfer of the catabolic plasmid RP4::Tn4371 to indigenous soil bacteria and its effects on respiration and biphenyl breakdown, FEMS Microbiol. Ecol. 15 (1994) 71–77. [14] A.K. Drönen, V. Torsvik, J. Goksoyr, E.M. Top, Effect of mercury addition on plasmid incidence and gene mobilizing capacity in bulk soil, FEMS Microbiol. Ecol. 27 (1998) 381–394. [15] M. Espinosa, S. Cohen, M. Couturier, G. del Solar, R. Diaz-Orejaz, R. Giraldo, L. Janniere, C. Miller, M. Osborn, C.M. Thomas, Plasmid replication and copy number, in: C.M. Thomas (Ed.), The Horizontal Gene Pool, Harwood Academic Publishers, Amsterdam, 2000, pp. 1– 48. [16] J.F. Fu, H.C. Chang, Y.M. Chen, Y.S. Chang, S.T. Liu, Characterization of the replicon of plasmid pSW500 of Erwinia stewartii, Mol. Gen. Genet. 250 (1996) 699–704. [17] R.R. Fulthorpe, R.C. Wyndham, Transfer and expression of the catabolic plasmid pBRC60 in wild bacterial recipients in a fresh water ecosystem, Appl. Environ. Microbiol. 57 (1991) 1546–1553. [18] P. Gabant, C.Y. Szpirer, M. Couturier, M. Faelen, Direct selection cloning vectors adapted to the genetic analysis of gram-negative bacteria and their plasmids, Gene 207 (1998) 87–92. [19] D. Garcia-de-Viedma, R. Giraldo, M.J. Ruiz-Echevarria, R. Lurz, R. Diaz-Orejas, Transcription of repA, the gene of the initiation protein of the Pseudomonas plasmid pPS10, is autoregulated by interactions of the RepA protein at a symmetrical operator, J. Mol. Biol. 247 (1995) 211–223.

508

M.-E. Gstalder et al. / Research in Microbiology 154 (2003) 499–509

[20] J.A. Goodrich, M.L. Schwartz, W.R. McClure, Searching for and predicting the activity of sites for DNA binding proteins: Compilation and analysis of the binding sites for Escherichia coli integration host factor (IHF), Nucleic Acids Res. 18 (1990) 4993–5000. [21] A. Götz, R. Pukall, E. Smit, E. Tietze, R. Prager, H. Tschape, J.D. van Elsas, K. Smalla, Detection and characterization of broadhost-range plasmids in environmental bacteria by PCR, Appl. Environ. Microbiol. 62 (1996) 2621–2628. [22] D. Hanahan, Studies on transformation of Escherichia coli with plasmids, J. Mol. Biol. 166 (1983) 557–580. [23] D.R. Helinski, A.E. Toukdarian, R.P. Novick, Replication control and other stable maintenance mechanisms of plasmids, in: F.C. Neidhardt, R. Curtiss, J.L. Ingraham, E.C.C. Lin, K.B. Low, B. Magasanik, et al. (Eds.), Escherichia coli and Salmonella, American Society of Microbiology, Washington, DC, 1996, pp. 2295–2324. [24] M. Inui, J.H. Roh, K. Zahn, H. Yukawa, Sequence analysis of the cryptic plasmid pMG101 from Rhodopseudomonas palustris and construction of stable cloning vectors, Appl. Environ. Microbiol. 66 (2000) 54–63. [25] R.S. Jobanputra, N. Datta, Trimethoprim R factors in enterobacteria from clinical specimens, J. Med. Microbiol. 7 (1974) 169–177. [26] G. Jagura-Burdzy, D.P. Macartney, M. Zatyka, L. Cunliffe, D. Cooke, C. Huggins, L. Westblade, F. Khanim, C.M. Thomas, Repression at a distance by the global regulator KorB of promiscuous IncP plasmids, Mol. Microbiol. 32, (1999) 519–532. [27] C.I. Kado, S.-T. Liu, Rapide procedure for detection and isolation of large small plasmids, J. Bacteriol. 145 (1981) 1365–1373. [28] T. Kaneko, Y. Nakamura, S. Sato, E. Asamizu, T. Kato, S. Sasamoto, A. Watanabe, K. Idesawa, A. Ishikawa, K. Kawashima, T. Kimura, Y. Kishida, C. Kiyokawa, M. Kohara, M. Matsumoto, A. Matsuno, Y. Mochizuki, S. Nakayama, N. Nakazaki, S. Shimpo, M. Sugimoto, C. Takeuchi, M. Yamada, S. Tabata, Complete genome structure of the nitrogen-fixing symbiotic bacterium Mesorhizobium loti, DNA Res. 7 (2000) 331–338. [29] W. Kelley, D. Bastia, Replication initiator protein of plasmid R6K autoregulates its own synthesis at the transcriptional step, Proc. Natl. Acad. Sci. USA 82 (1985) 2574–2578. [30] K. Kostelidou, G. Jagura-Burdzy, C.M. Thomas, Mutational analysis of the global regulator KorA of broad-host-range plasmid RK2, J. Mol. Biol. 281 (1998) 453–463. [31] K. Kostelidou, C.M. Thomas, The hierarchy of KorB binding at its 12 binding sites on the broad host range plasmid RK2 and modulation of this binding by IncC1 protein, J. Mol. Biol. 295 (2000) 411–422. [32] P. Lejeune, M. Mergeay, F. Van Gijsegem, M. Faelen, J. Gerits, A. Toussaint, Chromosome transfer and R-prime plasmid formation mediated by plasmid pULB113 (RP4::mini-Mu) in Alcaligenes eutrophus CH34 and Pseudomonas fluorecens 6.2, J. Bacteriol. 155 (1983) 1015–1026. [33] A.K. Lilley, M.J. Bailey, The acquisition of indigenous plasmids by genetically marked Pseudomonas population colonizing the sugar beet phytosphere in relation to local environmental conditions, Appl. Environ. Microbiol. 63 (1997) 1577–1583. [34] P. Linder, G. Churchward, G.X. Xia, Y.Y. Yu, L. Caro, An essential replication gene, repA, of plasmid pSC101 is autoregulated, J. Mol. Biol. 181 (1985) 383–393. [35] M. Mergeay, D. Nies, H.G. Schlegel, J. Gerits, P. Charles, F. van Gijsegem, Alcaligenes eutrophus CH34 is a facultative chemolithotroph with plasmid-bound resistance to heavy metals, J. Bacteriol. 162 (1985) 328–334. [36] R.P. Novick, Plasmid incompatibility, Microbiol. Rev. 51 (1987) 381– 395. [37] M.S. Okumura, C.I. Kado, The region essential for efficient autonomous replication of pSa in Escherichia coli, Mol. Gen. Genet. 235 (1992) 55–63. [38] M. Osborn, K.D. Bruce, P. Strike, D. Richie, Distribution, diversity and evolution of the bacterial mercury resistance (mer) operon, FEMS Microbiol. Rev. 19 (1997) 239–262.

[39] M. Osborn, S. Bron, S. Firth, S. Holsappel, A.S. Huddleston, R. Kiewiet, W.J. Meijer, J.F. Seegers, R. Skurray, P. Terpstra, C.M. Thomas, P.B. Thorsted, E. Tietze, S.L. Turner, Chapter 8: The evolution of bacterial plasmids, in: C.M. Thomas (Ed.), The Horizontal Gene Pool, Harwood Academic Publishers, Amsterdam, 2000, pp. 301–361. [40] W. Pansegrau, E. Lanka, P.T. Barth, D.H. Figurski, D.G. Guiney, D. Haas, D.R. Helinski, H. Schwab, V.A. Stanisich, C.M. Thomas, Complete nucleotide sequence of Birmingham IncPα plasmids: Compilation and comparative analysis, J. Mol. Biol. 239 (1994) 623–663. [41] W.R. Pearson, D.J. Lipman, Improved tools for biological sequence comparison, Proc. Natl. Acad. Sci. USA 85 (1988) 2444–2448. [42] M.F. Romine, L.C. Stillwell, K.-K. Wong, S.J. Thurston, E.C. Sisk, C. Sensen, T. Gaasterland, J.K. Fredrickson, J.D. Saffer, Complete sequence of a 184-kilobase catabolic plasmid from Sphingomonas aromaticivorans F199, J. Bacteriol. 181 (1999) 1585–1602. [43] J. Sambrook, E.F. Fritsch, T. Maniatis, Molecular Cloning, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, 2001. [44] C. Schaefer, W. Messer, DnaA protein/DNA interaction. Modulation of the recognition sequence, Mol. Gen. Genet. 226 (1991) 34–40. [45] S. Schneiker, M. Keller, M. Dröge, E. Lanka, A. Pülher, W. Selbitschka, The genetic organization and evolution of the broad host range mercury resistance plasmid pSB102 isolated from a microbial population residing in the rhizosphere of alfalfa, Nucleic Acids Res. 29 (2001) 5169–5181. [46] S. Silver, Bacterial resistances to toxic metal ions-a review, Gene 179 (1996) 9–19. [47] R. Simon, U. Priefer, A. Pühler, A broad host range mobilization system for in vivo genetic engineering: Transposon mutagenesis in gram negative bacteria, Biotechnology 1 (1983) 784–791. [48] C.A. Smith, C.M. Thomas, Comparison of the organisation of the genomes of phenotypically diverse plasmids of incompatibility group P: Members of the IncP beta subgroup are closely related, Mol. Gen. Genet. 206 (1987) 419–427. [49] P.A. Sobecky, T.J. Mincer, M.C. Chang, D.R. Helinski, Plasmids isolated from marine sediment microbial communities contain replication and incompatibility regions unrelated to those of known groups, Appl. Environ. Microbiol. 63 (1997) 888–895. [50] P.A. Sobecky, T.J. Mincer, M.C. Chang, A. Toukdarian, D.R. Helinski, Isolation of broad-host-range replicons from marine sediment bacteria, Appl. Environ. Microbiol. 64 (1998) 2822–2830. [51] C. Szpirer, E.M. Top, M. Couturier, M. Mergeay, Retrotransfer or gene capture: A feature of conjugative plasmids, with ecological and evolutionary significance, Microbiology 145 (1999) 3321–3329. [52] A. Tauch, S. Krieft, J. Kalinowski, A. Pühler, The 51 409-bp R-plasmid pTP10 from the multiresistant clinical isolate Corynebacterium striatum M82B is composed of DNA segments initially identified in soil bacteria and in plant, animal, and human pathogens, Mol. Gen. Genet. 263 (2000) 1–11. [53] A. Tauch, S. Schneiker, W. Selbitschka, A. Pülher, L.S. van Overbeek, K. Smalla, C.M. Thomas, M.J. Bailey, L.J. Forney, A. Weightman, P. Ceglowski, T. Pembroke, E. Tietze, G. Schroder, E. Lanka, J.D. van Elsas, The complete nucleotide sequence and environmental distribution of the cryptic, conjugative, broad-host-range plasmid pIPO2 isolated from bacteria of the wheat rhizosphere, Microbiology 148 (2002) 1637–1653. [54] C.M. Thomas, The Horizontal Gene Pool—Bacterial Plasmids and Gene Spread, Harwood Academic Publishers, Amsterdam, 2000. [55] J.D. Thompson, D.G. Higgins, T.J. Gibson, CLUSTAL W: Improving the sensitivity of progressive multiple sequence alignment through sequence weighting, positions-specific gap penalties and weight matrix choice, Nucleic Acids Res. 22 (1994) 4673–4680. [56] P.B. Thorsted, D.P. Macartney, P. Akhtar, A.S. Haines, N. Ali, P. Davidson, T. Stafford, M.J. Pocklington, W. Pansegrau, B.M. Wilkins, E. Lanka, C.M. Thomas, Complete sequence of the IncPβ plasmid R751: Implications for evolution and organisation of the IncP backbone, J. Mol. Biol. 282 (1998) 969–990.

M.-E. Gstalder et al. / Research in Microbiology 154 (2003) 499–509

[57] E.M. Top, I. De Smet, W. Verstraete, R. Dijkmans, M. Mergeay, Exogenous isolation of mobilizing plasmids from polluted soils and sludges, Appl. Environ. Microbiol. 60 (1994) 831–839. [58] E.M. Top, W.E. Holben, L.J. Forney, Characterization of diverse 2,4-dichlorophenoxyacetic acid-degradative plasmids isolated from soil by complementation, Appl. Environ. Microbiol. 61 (1995) 1691– 1698. [59] E.M. Top, O.V. Maltseva, L.J. Forney, Capture of a catabolic plasmid that encodes only 2,4-dichlorophenoxyacetic acid: alpha-ketoglutaric acid dioxygenase (TfdA) by genetic complementation, Appl. Environ. Microbiol. 62 (1996) 2470–2476. [60] E.M. Top, D.P. Van Elsas, S.N. De Smet, L.J. Forney, Enhancement of 2,4-dichlorophenoxyacetic acid (2,4-D) degradation in soil by

509

dissemination of catabolic plasmids, Antonie Van Leeuwenhoek 73 (1998) 87–94. [61] A. Vande Broek, Y. Okon, J. Vanderleyden, Isolation and sequence analysis of repA from the incurable 90 MDa plasmid of Azospirillum brasilense, DNA Seq. 11 (2000) 101–107. [62] F. Van Gijsegem, A. Toussaint, M. Casadaban, Mu as a genetic tool, in: N. Symonds, A. Toussaint, P. van de Putte, M.M. Howe (Eds.), Phage Mu, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, 1987, pp. 215–250. [63] L.-T. Wu, Y.-H. Tseng, Characterization of the IncW cryptic plasmid pXV2 from Xanthomonas campestris pv. vesicatoria, Plasmid 44 (2000) 163–172.