Class II broad-spectrum mercury resistance transposons in Gram-positive bacteria from natural environments

Res. Microbiol. 152 (2001) 503–514  2001 Éditions scientifiques et médicales Elsevier SAS. All rights reserved S0923-2508(01)01224-4/FLA

Class II broad-spectrum mercury resistance transposons in Gram-positive bacteria from natural environments Elena Bogdanovaa∗ , Leonid Minakhina, Irina Bassa , Alexander Volodina, Jon L. Hobmanb , Vadim Nikiforova a Institute of Molecular Genetics, Russian Academy of Sciences, Moscow 123182, Russia b School of BioSciences, University of Birmingham, Edgbaston, Birmingham, B15 2TT, UK

Received 10 July 2000; accepted 10 October 2000

Abstract – We have studied the mechanisms of the horizontal dissemination of a broad-spectrum mercury resistance determinant among Bacillus and related species. This mer determinant was first described in Bacillus cereus RC607 from Boston Harbor, USA, and was then found in various Bacillus and related species in Japan, Russia and England. We have shown that the mer determinant can either be located at the chromosome, or on a plasmid in the Bacillus species, and is carried by class II mercury resistance transposons: Tn5084 from B. cereus RC607 and B. cereus VKM684 (ATCC10702) and Tn5085 from Exiguobacterium sp. TC38-2b. Tn5085 is identical in nucleotide sequence to TnMERI1, the only other known mer transposon from Bacillus species, but it does not contain an intron like TnMERI1. Tn5085 is functionally active in Escherichia coli. Tn5083, which we have isolated from B. megaterium MK64-1, contains an RC607-like mer determinant, that has lost some mercury resistance genes and possesses a merA gene which is a novel sequence variant that has not been previously described. Tn5083 and Tn5084 are recombinants, and are comprised of fragments from several transposons including Tn5085, and a relative of a putative transposon from B. firmus (which contains similar genes to the cadmium resistance operon of Staphylococcus aureus), as well as others. The sequence data showed evidence for recombination both between transposition genes and between mer determinants.  2001 Éditions scientifiques et médicales Elsevier SAS class II broad-spectrum mercury resistance transposons / horizontal gene transfer

1. Introduction Mercury resistance genes (mer determinants) provide a suitable model system for studying gene transfer in environmental bacteria. Mercury is present in the environment both as a result of natural processes, and from anthropogenic sources [17, 36]. Resistance to mercury salts is found in many genera of bacteria isolated from the environment, and many mer determinants have been sequenced. Two main mer determinant types have been described: narrow-spectrum mer determinants confer resistance to inorganic mercury salts only. Broad-spectrum mer determinants confer resistance to organomercurials as well as inorganic mercury salts [7, 17, 29, 40, 42, 44].

∗ Correspondence and reprints.

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

Several studies have been devoted to describing the mechanisms of mer determinant horizontal transfer in Gram-negative bacteria. The mer determinants from these bacteria are often located on plasmids and on different transposons, many of which have been studied in detail [14, 17, 18, 24, 25, 27, 32, 33, 35, 43, 44]. The majority of mercury resistance transposons that have been studied are class II. They are characterized by the presence of 35–48-bp terminal inverted repeats (IR), transposase (tnpA) and resolvase (tnpR) genes, a res-internal resolution site, and genes other than those required for transposition [14]. The mechanisms of mer determinant horizontal transfer in Gram-positive bacteria are less well understood [17]. Few data are available about plasmidborne mer determinants in environmental Grampositive bacteria [3, 5, 7], and most of these determinants are usually believed to be located on the chromosome [16, 20, 28, 30, 41, 42]. Until now, only one mer transposon from Gram-positive bacteria has

504

E. Bogdanova et al. / Res. Microbiol. 152 (2001) 503–514

been described: TnMERI1 from Bacillus megaterium MB1, which is a class II transposon [20, 21]. The objective of this study was to ascertain genetic structures involved in the horizontal transfer of the best characterized and most widely distributed broad-spectrum mer determinant, initially described in Bacillus cereus RC607 isolated from Boston Harbor, USA [41, 42]. This mer determinant had also been identified in the 30 strains including B. subtilis, B. firmus and B. lentus and also in B. megaterium MB1 in Minamata Bay, Japan [21, 30] and B. cereus and unidentified Bacillus strains in England [16]. We had previously found this mer determinant in Exiguobacterium sp. TC38-2b from Carpathia (Ukraine), and in the sole mercury-resistant (HgR ) B. cereus VKM684 (ATCC10702) strain among 26 randomly selected mercury-sensitive (HgS ) B. cereus strains from the collection [6, 7]. This strain had originally been obtained by the ATCC collection from a nondocumented locale and habitat prior to 1948 [10]. We have also identified a fragment of the RC607 mer determinant in B. megaterium MK64-1 from the Kuril islands [7]. (The B. megaterium MK64-1 strain mer determinant will be named the RC607-like mer determinant below.) It has been reported that the RC607 mer determinant is localized only in the chromosome in all studied environmental Bacillus species in USA, Japan and England [16, 20, 28, 30, 41, 42] and B. cereus VKM684 [7]. However, we identified this mer determinant in a plasmid in the closely related genus, Exiguobacterium [7]. Therefore in this work we attempted to determine whether the RC607 mer determinant can be found as plasmid-localized in environmental Bacillus strains. In an earlier study we identified the terminal inverted repeats (IRs) of a class II transposon in the downstream sequence flanking the mer determinants in B. cereus RC607, Exiguobacterium sp. TC38-2b and B. megaterium MK64-1, and suggested the involvement of class II transposons in the dissemination of mer determinants in Gram-positive bacteria [7]. Indeed, in 1999 the RC607 mer determinant was found in the TnMERI1 transposon from B. megaterium MB1, the only known Bacillus mer transposon [20, 21]. The structure of its flanking repeats was not completely determined, nor has its functional activity been described. Therefore this work investigated whether the RC607 and RC607-like mer determinants are located on transposons in Exiguobacterium sp.

TC38-2b, B. cereus RC607, B. cereus VKM684 and in B. megaterium MK64-1. We also studied the structure and functional activity of first discovered mer transposons. 2. Materials and methods 2.1. Bacterial strains and plasmids

The mercury-resistant (HgR ) strain used in this study, Exiguobacterium sp. TC38-2b, is from our collection of low G+C content Gram-positive bacteria [7]. HgR B. cereus VKM684 (ATCC10702) has been previously described [6, 7, 10]. HgR B. cereus RC607 was obtained from Mahler in 1990. These strains were preserved for long-term storage by freeze drying. At the same time the stock culture was maintained at 4–8◦ C on Luria agar (LA) with 7 µg mL−1 of HgCl2 . In this work we isolated the HgS derivative of B. cereus RC607 strain (see below). In mating experiments, we used its spontaneous mutant, VKM607-14 (HgS RifR ), which was resistant to 50 µg mL−1 of rifampicin on LA. The Escherichia coli K12 strains used were: HB101 (thi-1 proA2 leu-2 recA13 hsdS20 rpsL20) [37], a RifR derivative of UB5201 (pro met recA56 gyrA) [13] and IF238 (prototroph, gyrA91) [11]. The conjugative plasmid pRP1.2 (TcR ) [8], a derivative of the broad-host-range plasmid RP1 [23, 34] as well as pACYC184KS (CmR TcS ) [24] were used for transposition assays. The plasmids pACYC184, pUC19, and pTZ19 were used in cloning experiments. The list of the main recombinant plasmids with the inserts from DNA from Exiguobacterium sp. TC38-2b, B. cereus RC607 and B. megaterium MK64-1 is shown in table I. 2.2. Media and growth conditions

Unless otherwise stated, cells were grown in Luria broth (LB) [37] or on LA plates. LA was supplemented with selective agents at the following concentrations (µg mL−1 ): Ap, ampicillin (100); Tc, tetracycline (10); Cm, chloramphenicol (20); Rif, rifampicin (40); HgCl2 (5–8). 2.3. Cloning and other DNA manipulations

Restriction enzyme digests, DNA ligations and other molecular biology methods were carried out

E. Bogdanova et al. / Res. Microbiol. 152 (2001) 503–514 Table I. The main recombinant plasmids.

Plasmid pKLH3.1a pKLH3.2a pKLH3.3a pKLH3.4 pKLH3.5 pKLH3.6 pKLH3.7 pKLH3.8 pKLH3.9

pYW33b p35Eb pYW33.1 pYW33.2 pYW33.3 pKLH6.1 pKLH6.2 pKLH6.3 pKLH6.4 pKLH304.1a

Description 12.0-kb HindIII fragment from pKLH3 in pTZ19 1.3-kb EcoRI (through BglII) fragment from pKLH3.1 in pUC19 2.1-kb BglII/EcoRI (through StuI) fragment from pKLH3.1 in pUC19 1.7-kb EcoRI/EcoRI fragment from pKLH3.1 in pTZ19 1.9-kb EcoRI/HindIII (through BglII) fragment from pKLH3.1 in pTZ19 3.0-kb BglII/BglII fragment from pKLH3 in pUC19 1.8-kb HindIII/PstI (through two EcoRI sites) fragment from pKLH3.6 in pTZ19 13.8-kb HindIII/PstI fragment (after pKLH3.1 and pKLH3.7 inserts ligation) in pTZ19 2.1-kb HindIII/XbaI (see figure 1: XbaI site locates between two NcoI sites) fragment from pKLH3.1 in pTZ19 12.8-kb HindIII fragment from RC607 in pUC9 3.7-kb EcoRI/HindIII fragment from pYW33 in pUC9 5.5-kb HindIII/BglII fragment from pYW33 in pTZ19 0.6-kb HindIII/EcoRI fragment from pYW33 in pTZ19 1.1-kb EcoRI/PstI (through SspI) fragment from pYW33 in pTZ19 2.6-kb HindIII/HindIII fragment from pKLH6 in pTZ19 1.5-kb EcoRI/EcoRI fragment from pKLH6 in pTZ19 5.5-kb HindIII/BglII fragment from pKLH6 in pTZ19 14.6-kb HindIII/PstI fragment (after ligation of the inserts from pYW33 and pKLH3.7) in pTZ19 8.8-kb HindIII/HindIII fragment from pKLH304 in pACYC184

a The inserts were shown in [7]. b Obtained from I. Mahler, Brandeis University, Waltham, MA, USA.

by standard procedures [37]. The recombinant clones (table I) were obtained by standard methods [37]. Some of these clones have been previously described [7]. Recombinant plasmid DNA was prepared from E. coli by a standard alkaline lysis procedure [4]. Southern blot hybridization and DNA sequencing were carried out as described [7]. Alignment of DNA sequences was performed using the VOSTORG pro-

505

gram [45] and BLAST subroutines [1] (National Center for Biotechnology Information, Bethesda, MD). Plasmid preparations from the RC607 strain were isolated by alkaline lysis [4] with modifications. We had earlier [7] found that for the isolation of plasmid DNA from environmental Bacillus strains it was necessary: 1) to grow cells in nutritionally poor medium; 2) to wash cells with the lysis buffer; 3) to lyse cells quickly at the lowest possible temperature and consequently to find individual lysis conditions for each strain, and 4) immediately after isolation of DNA to deproteinize it with phenol and SDS. Therefore, the RC607 strain was grown overnight in M9 medium containing 0.25 volumes of aminopeptide [5], 0.25% glucose and 1 µg mL−1 HgCl2 . Cell pellets were washed with 0.5 volumes of lysis buffer (50 mM glucose, 25 mM Tris.HCl, pH 8.0, 20 mM EDTA) and incubated in the same buffer containing 2 mg mL−1 of lysozyme for 10–12 min at 37◦ C, cooled to 0–4◦ C and NaOH/SDS added. After neutralization with Na acetate, the suspension was centrifuged for 15 min at 15 000 g at 0–4◦ C. RNA was precipitated by the addition of LiCl to 2.5 M, then DNA was precipitated by the addition of 0.6 volumes of propan-2ol. Resuspended DNA was deproteinized with phenol/chloroform with addition of 2% SDS according to standard procedures [37]. The size of the RC607 strain plasmids were determined as a sum of the molecular masses of their EcoRI fragments. The isolation of pKLH3 plasmid from Exiguobacterium sp. TC38-2b and total DNA isolation from the Bacillus strains were described in [7]. 2.4. Curing of the HgR plasmid (pKLH6) in B. cereus RC607

In the first experiment, the initial HgR RC607 strain was grown in LB without HgCl2 at 42◦ C for 18 h. The bacterial suspension was diluted with LB by 5 × 107 times and grown at 42◦ C for a further 24 h. In the second experiment, the strain was grown at 42.5◦ C for 72 h, diluting the culture with fresh LB twice, every 24 h. In the first experiment 10% of the colonies that grew were HgS , while 60% were HgS in the second experiment. 2.5. Mating conditions

We used the original B. cereus RC607 (HgR RifS ) as donor and the VKM607-14 (HgS RifR ) strain as

506

E. Bogdanova et al. / Res. Microbiol. 152 (2001) 503–514

recipient. Mating was carried out as described in [7], but cultures of donor and recipient were grown separately in LB containing 3 µg mL−1 HgCl2 or 5 µg mL−1 rifampicin, respectively, or on LA containing 5 µg HgCl2 or 50 µg mL−1 rifampicin. The pKLH6 (HgR plasmid from RC607) transfer frequency was less than 1 × 10−8 per recipient cell. 2.6. Transposition assays

Transposition was tested in mating experiments essentially as described by G. Kholodii et al. [24]. Donor strains for these experiments were constructed by introducing the plasmids pKLH3.8, pKLH6.4 and pKLH3.1 (all – HgR ApR ) (table I and figure 1) by transformation into HB101 harboring the conjugative plasmid pRP1.2 (TcR ). The resulting clones were propagated on LB agar containing HgCl2 plus Tc for 120–160 h, then grown overnight in LB broth in the presence of HgCl2 (3 µg mL−1 ) plus Tc (5 µg mL−1 ). These cultures were diluted 50-fold into fresh LB without antibiotic selection, grown for 3.5 h, mixed in tubes with a late logarithmic culture of a RifR recipient (UB5201, recA), and dropped on LB plates, which were then incubated for 24 h (all incubations described were at 30◦ C). Aliquots of the mixed culture from the plates were suspended in LB with 0.3 µg mL−1 HgCl2 , and plated on LB agar containing HgCl2 , (selecting for the transposon mercury resistance marker), plus Rif (selection against the donor). Another aliquot of the mixed growth suspension was plated onto LB agar containing Tc, (selecting for the conjugative plasmid pRP1.2), plus Rif. The ratio of the number of transconjugants on the HgCl2 plus Rif plates to the total number of TcR (pRP1.2 – carrying) RifR transconjugants gave the transposition frequency. Each transposition experiment was repeated twice. Genetic linkage of markers expected in recombinant plasmids and consequently cointegrate resolution was tested by mating HgR TcR RifR transconjugants from each cross with HB101 (pACYC184KS, CmR TcS ). All HgR TcR RifR transconjugants tested (150 colonies per experiment) were also ampicillinresistant and could transfer all three resistance markers (HgR TcR ApR ) to HB101 pACYC184KS (TcS CmR ).This result indicated that a cointegrate plasmid (pRP1.2::Tn5085::pKLH3.8) had been transferred. As a control, a similar mating was performed with the donor strain HB101, containing both pRP1.2

Figure 1. Schematic representation of the complete structure of Tn5085, Tn5084 and the transposon from B. cereus VKM684, and restriction digest maps of the regions flanking the transposons. The restriction maps of Tn5085 and Tn5084 transposition modules were constructed in the same way as in figure 5. The triangles mark IRs, rectangles near IRL mark the tnpA gene fragment [7]. res, tnpR and tnpA denote the res region, resolvase and transposase genes. The restriction maps of the mercury resistance modules were constructed by analysis of cloned fragments and sequence data (Ac: TC38-2b, X99457, Y08064, [7]; RC607, M22708 [42], Y08065 [7], AB036431, AF138877 [15]). The restriction map of the transposon from B. cereus VKM684 [7] and the flanking sequences of every transposon were constructed by Southern blot hybridization with 32 P-labeled cloned fragments as the probes shown in this figure and table I. Only experimentally determined restriction sites are shown for VKM684. The sequences of the Tn5085 insertion site in pKLH3 (the plasmid from Exiguobacterium sp. TC38-2b) and the Tn5084 insertion site in pKLH6 (the naturally occurring plasmid from B. cereus RC607) [this work] are shown under the Tn5085 and Tn5084 schemes. Duplications of the DNA target are underlined. Abbreviations: K, KpnI; S, Stu I; others – as in figure 5. Tn5085 scheme is darkened to show its region in the chimeric Tn5084–Tn5085 transposon.

(TcR ) and pKLH3.1 (HgR ApR ). pKLH3.1 contains a defective Tn5085 in which one-third of the tnpA gene has been deleted (figure 1). The frequency of HgR ApR TcR co-transfer into UB5201 (RifR recA) per total number of TcR transconjugants was 4 × 10−7 . The subsequent crossing of this ‘transconjugant’ with HB101 (pACYCKS, TcS CmR ) showed that the HgR ApR TcR co-transfer resulted from the mobilization of plasmid pKLH3.1 by pRP1.2. Cointegrate pRP1.2::Tn5085::pKLH3.8 (HgR ApR TcR ) in UB5201 (RifR ) was resolved after mating of this strain with the E. coli strain IF238 (which is a prototroph and has a complete recombination system) and selecting for TcR on minimal agar. The sec-

E. Bogdanova et al. / Res. Microbiol. 152 (2001) 503–514

507

ondary transconjugants HgR ApR TcR RifS , formed as a result of the cross, were twice purified to separate colonies on minimal agar with Tc and were tested for transposon and other donor plasmid markers (HgCl2 and Ap, respectively). The resulting transconjugants: pRP1.2::Tn5085 (HgR ApS TcR ) originated from cointegrate resolution. The transconjugants were also tested by analysis of their DNA. Plasmid DNA was isolated from 10– 100 mL of overnight cultures and analyzed initially without restriction in 0.6% agarose gel electrophoresis. Under these conditions, plasmid DNA of clones pKLH3.1, pKLH3.8 and pKLH6.4 could be easily distinguished by size from pRP1.2 and from cointegrates. 3. Results 3.1. The broad-spectrum mer determinant from B. cereus RC607 is plasmid-borne

Plasmid and total DNA preparations were isolated from RC607 as described in Materials and methods and compared with the clone pYW33 DNA. This clone was used as the standard and contained the broad-spectrum mer determinant originally isolated from RC607 total DNA within a 12.8-kb HindIIIHindIII fragment [42]. All DNAs were analyzed by restriction with HindIII and HindIII/EcoRI and blot-hybridization with the individual 32 P labeled DNA probes pYW33, pYW33.2, p35E (RC607) and pKLH3.2-3.5 (TC38-2b) (figure 1 and table I) and mixtures of probes. Figure 2A, lines 1–6, shows the result of the hybridization experiment with the mixture of pKLH3.2 and pKLH3.3 probes, containing mer determinant genes: merR1, ORF2, ORF3, ORF4 and merA (figure 3 legend contains these genes description). The figure 2A and all other hybridization experiments demonstrated that: 1) both the RC607 total DNA and the RC607 plasmid DNA appeared to contain the mer determinant; and 2) this mer determinant is identical to the cloned mer determinant in pYW33. These data indicated that the mer determinant in the RC607 strain may be plasmid-borne. This HgR plasmid was named pKLH6. The localization of the mer determinant in RC607 at a plasmid was confirmed by elimination of pKLH6 (see Materials and methods). Growth of the strain at elevated temperatures resulted in a proportion of the bacteria becoming HgS . DNA isolated from these

Figure 2. A. Southern blot DNA–DNA hybridization. DNA samples were electrophoresed in a 1% agarose gel, transferred to a HAWP membrane (Millipore) and probed with: lanes 1–8, the mixture of 32 P-labeled inserts from both pKLH3.2 and pKLH3.3; lanes 9– 16, a mixture of 32 P-labeled inserts from pKLH3.4, pKLH3.5 and pYW33.2. Lanes 1, 4, 9 and 12 contain total DNA extracted from RC607; 2, 5, 10 and 13, plasmid DNA extracted from RC607; 3, 6, 11 and 14, the pYW33 clone used as the standard; 7 and 15, plasmid DNA extracted from an HgS derivative of RC607; 8 and 16, total DNA extracted from the HgS RC607 strain. DNA preparations in lanes 1–3 and 9–11 were digested by HindIII. DNA preparations in lanes 4–8, 12–16 were digested by the mixture of EcoRI and HindIII. B. EcoRI restriction pattern of DNA fragments from plasmid preparations of the original HgR B. cereus RC607 strain and from its HgS derivative. Lane 1 contains as size marker a Pst I digest of lambda phage DNA; lane 2, HgR RC607 strain; lane 3, HgS RC607 strain.

bacteria (both by total and plasmid DNA preparations) did not hybridize with pYW33 and other labeled probes (figure 2A, lines 7, 8, 15 and 16). Simultaneously, plasmid DNA isolated from the HgS RC607 strains (from all 14 studied colonies) was

508

E. Bogdanova et al. / Res. Microbiol. 152 (2001) 503–514

missing restriction fragments of about 21 kb in total (figure 2B), which suggests that the size of pKLH6 is about 21 kb. (All nine studied colonies remaining as HgR had pKLH6.) The strain also harbors another plasmid (or plasmids) of 13.5 kb. We found no evidence in mating experiments to show that pKLH6 was self-transmissible. The results described above were obtained using the RC607 strain received from the USA about 10 years ago and maintained on plates in our laboratory during this time. We also isolated plasmid DNA from a resuscitated vial of freeze-dried RC607 and found that the plasmid profile from this strain did not differ from the ones described above (data not shown). Figure 2A, lines 9–14 shows the result of the hybridization experiment with the mixtures of inserts from pKLH3.4, pKLH3.5 and pYW33.2 as the probes, containing merB3 gene of the mer determinant and DNA fragments flanking the mer determinant from the both sides including distantly located pYW33.2 insert (figure 1 and table I). It demonstrated the similarity of mer determinant flanks in RC607 total DNA, plasmid DNA and pYW33 clone. All these DNA preparations appeared to contain an identical mer determinant as a part of a 12.8-kb identical HindIII-HindIII fragment. Detailed restriction analysis of the left flank of each DNA fragment present in the naturally occurring pKLH6 plasmid and in the pYW33 clone further confirmed the identity of their HindIII-HindIII fragments. The 5.5-kb HindIII-BglII fragments from pKLH6 and from pYW33 were subcloned to produce the clones pKLH6.3 and pYW33.1, respectively. Both clones were then digested with HindIII, NcoI, EcoRI, SspI and PstI (figure 1) which produced identically sized DNA fragments (data not shown). Thus, these data indicated that the pYW33 clone been apparently obtained from the pKLH6. 3.2. The RC607 and RC607-like mer determinants are found in three closely related class II transposons: Tn5085, Tn5084 and Tn5083

The mer determinant flanking sequences in Exiguobacterium sp. TC38-2b, B. cereus RC607 and B. megaterium MK64-1 were analyzed by restriction analysis, Southern blot hybridization and partial sequencing of the cloned fragments from their plasmids.

3.2.1. Tn5085 is virtually identical to TnMERI1 from B. megaterium MB1 from Japan but does not contain an intron

In an earlier study we identified the RC607 mer determinant in the conjugative plasmid pKLH3 from Exiguobacterium sp. TC38-2b, (from Carpathia, Ukraine) [7]. In that work we analyzed the subclones pKLH3.1 and pKLH3.6 containing the mer determinant and its flanking regions, derived from pKLH3 (figure 1). We found that they contained a complete set of transposition module elements: a res region with a putative resI site (figure 4A), and genes for resolvase and a complete transposase, which were all transcribed in the same direction (figure 5), and two nonidentical 38-bp inverted repeats at its ends (figure 6). Thus, the mer determinant was located in a transposon which was named Tn5085. The size of Tn5085 including the mercury resistance genes is 11.8 kb (figure 1). Tn5085 has the same gene organization as TnMERI1, the only other known mer transposon from B. megaterium MB1 from Japan [20]: restriction analysis revealed no differences in transposition module structure between Tn5085 and TnMERI1 (figure 5). Sequencing of approximately 70% of the transposition module upstream of merB3 and downstream of merB1 (figure 3) found 11 nucleotide substitutions over 3.4 kb of sequenced DNA (i.e. 0.3% difference) between Tn5085 and TnMERI1. However, unlike TnMERI1, Tn5085 contained no intron. The mercury resistance determinant (about 6.8 kb; its structure is shown in figure 3) has not been completely sequenced in either Tn5085 or TnMERI1, but their restriction maps are identical [20]. The sequenced regions of the merB3 gene [17, 21], are identical (Ac: Tn5085, X99457; TnMERI1, AB022308, AB027306 [21]). It is therefore likely, that Tn5085 and TnMERI1 are identical except for the presence of the intron in TnMERI1. 3.2.2. Tn5085 shows transposition activity in E. coli cells

We constructed a complete Tn5085 transposon by ligation of two DNA fragments, one from pKLH3.1 and the other from pKLH3.7, using pTZ19 as vector (table I, figure 1). The genetic data showed that transpositions of Tn5085 into the recipient plasmid, pRP1.2 in E. coli HB101 (recA) occurred at a frequency of 2 × 10−4 –1.4 × 10−3 and resulted in cointegrate plasmid pRP1.2::Tn5085::pKLH3.8 formation. These cointegrates were resolved in E. coli IF238

E. Bogdanova et al. / Res. Microbiol. 152 (2001) 503–514

509

Figure 3. Schematic representation of the mosaic structure of Tn5084 and Tn5083. The mercury resistance determinant (mer determinant) includes: the merA gene encoding the enzyme mercuric reductase. MerB 1, merB 2 and merB 3 (denoted B1, B2, B3 ), genes encoding different organomercurial lyase enzymes; merR1, merR2 (denoted R1, R2 ), genes encoding proteins regulating mercury resistance; ORF2, ORF3 and ORF4 (denoted ORFS ), genes encoding proteins whose functions are discussed [15, 17, 20, 21, 42]. The tnpR, tnpA and tnpA genes encode resolvase, transposase and a fragment of transposase [7]. IRR , IRL and res denote terminal repeats and putative res I site. Unfilled and unshaded genes or parts of them have nucleotide sequences virtually identical to the same regions of Tn5085. The sequences in the shaded and darkened regions differ from the same regions of Tn5085 by about 6–10% and about 40%, respectively. The sequences of the res regions and parts of the tnpR and tnpA genes are shown in figure 4A, B. Ac of the elements of the transposition modules are shown in figure 5, Ac of mercury resistance modules are: Tn5085, X99457, Y08064 [7]; Tn5084, M22708 [42], AB036431, AF138877 [15]; Tn5083, Y09907.

(which has a complete recombination system) probably owing to a general recA-dependent recombination system of the IF238 host bacterium. Electrophoresis of DNA preparations from these transconjugants confirmed the transposon insertion into pRP1.2 and cointegrate resolution (data not shown). Restriction analysis of the plasmid DNA from several pRP1.2::Tn5085 (HgR ApS TcR ) transconjugants (using restriction enzymes shown above the diagram of Tn5085 in figure 1, or with combinations of them) showed that all transconjugants mapped as a simple insertion of transposon DNA into the recipient plasmid, pRP1.2 (pRP1.2 is a pRP1 deletant that has lost approximately 10 kb of DNA from 0–10 region (G. Kholodii, unpublished)). We identified 8 different Tn5085 insertion sites in plasmid RP1.2. Of these, 5 were located to the 9.5-kb region between positions 30.3–39.8 and 3 insertions were located in a 4.5-kb region at positions 58.0–12.4 (using the coordinates of pRP1). Insertion of Tn5085 was effected in both orientations.

Figure 4. Alignments of two regions of transposons Tn5085, Tn5084 and Tn5083 : A) partial sequence (from the beginning) of the tnpR gene and the res region upstream of this gene across the putative res I site; B) the end of tnpR gene and start of tnpA. A dot indicates a base identical to that in the Tn5085 sequence; a gap is indicated by a dash. The site of the intron insertion into the identical sequence of TnMERI1 [20] is shown by a rectangle. Ac: Tn5085, Y17750; Tn5084, Y17748; Tn5083, Y18009; TnB.firm. (Tn from B. firmus ), M90749, M90750 [22].

3.2.3. Transposon Tn5084 differs from Tn5085 only in the res region and in the tnpR gene

In this work we have shown that the mer determinant in B. cereus RC607 (from Boston Harbor, USA) is located in pKLH6, the naturally occurring plasmid from this strain. Analysis of pKLH6 and clones pYW33 and pKLH6.1 derived from pKLH6 (figure 1 and table I) showed that in this plasmid the RC607 mer determinant was found in the mercury resistance transposon, which we called Tn5084 (figure 1). Restriction analysis and Southern blot hybridization showed Tn5084 to closely resemble Tn5085 (figure 5 and figure 1). Sequencing of about 60% of the trans-

510

E. Bogdanova et al. / Res. Microbiol. 152 (2001) 503–514

Figure 6. Alignment of the terminal IRs of Tn5085, Tn5084 and Tn5083 and comparison with IRs from class II transposons. Ac for the IRs: Tn5085, Y08064, Y17752; Tn5084, Y08065, AJ277277; Tn5083, Y09907; Tn4430 from B. thuringiensis, Tn3 from E. coli are from [2].

Figure 5. Restriction endonuclease maps of part of the mercury resistance determinant and the transposition module from Tn5085 (from Exiguobacterium sp. TC38-2b), Tn5084 (from B. cereus RC607) and Tn5083 (from B. megaterium MK64-1), compared with TnMERI1 (Ac: AB022308 and AB027306, [20, 21]). The restriction maps were constructed by analysis of the clones shown in figure 1 and pKLH304.1 from MK64-1 (table I and [7] and their subclones (E. Bogdanova, unpublished). Southern blot hybridization using pKLH3.7 clone (cloned fragment of pKLH3 from Exiguobacterium sp. TC38-2b), as a probe, was used in part for the construction of the restriction map of Tn5084. Abbreviations: E, EcoRI; A, AvaII; Dr, DraI; X, XbaI; B, Bgl II; H, HindIII; N, NcoI; P, Pst I. Horizontal arrows indicate the location and orientation of the genes or parts of them verified by DNA sequencing. Dashed lines indicate the approximate location of the genes. MerB3 denotes the organomercurial lyase gene (see figure 3). tnpR and tnpA denote the resolvase and transposase genes, respectively. The triangles show the inverted repeat, IRR . Open boxes indicate the sequenced regions of the transposons. Not all restriction sites are shown in these regions. Ac: Tn5085, X99457, Y17750-Y17752; Tn5084, M22708 [42], Y17748, Y17741, Y17749, AJ277277; Tn5083, Y09907, Y18009. The darkened triangle denotes the site of the intron insertion in TnMERI1.

position module, upstream of merB3 and downstream of merB1 (figure 3), showed only 0.2% nucleotide substitutions in Tn5084 compared with Tn5085, except within a 621-bp region (containing the entire tnpR gene and the res region) in which the transposons differed by 10% (figures 3 and 4). TnpRs encoded by Tn5084 and Tn5085 were 8% different and both were closely related to the resolvase of a putative transposon from B. firmus (figure 7).The terminal IRs from Tn5084 and Tn5085 were identical (figure 6). We had previously shown that all the sequenced genes (about 4.2 kb) of the mercury resistance determinants from Tn5084 and from Tn5085 are virtually identical [7]. The merB3 genes of both transposons are identical (Ac; Tn5084, M22708; Tn5085,

X99457). It seems probable, therefore, that the two transposons differed only in res and the resolvase gene region. (After our work was completed and our Tn5084 sequences were assigned confidential accession numbers (Y17741, Y17748, Y17749), Endo, Silver and Huang submitted their sequence of the part of this transposon to GeneBank (AB036431). These sequences are almost identical.) We were not able to reconstruct Tn5084 to test its functional activity in the same way as we had reconstructed Tn5085, i.e. by ligation of two large HindIII-HindIII fragments of Tn5084 from the inserts from pYW33 and pKL6.1 (figure 1). The ligation of these inserts formed the transposon with a deletion in its tnpA gene because there was an additional HindIII site in the Tn5084 tnpA gene (figure 5 and figure 1). Therefore, we constructed the chimeric transposon Tn5084-Tn5085 by ligation of the inserts from pYW33 (Tn5084) and pKLH3.7 (Tn5085) into pTZ19 to produce pKLH6.4 (figure 1). In experiments similar to those described above, the transposition frequency of the chimeric Tn5084-Tn5085 was about 1 × 10−3 . Like Tn5085, the chimeric transposon was transposed as a cointegrate plasmid (pRP1.2::Tn5084-Tn5085::pKLH6.4). Cointegrate resolution only occurred in E. coli IF238. The two studied sites of the Tn5084-Tn5085 integration to plasmid pRP1.2 were located near the integration sites of Tn5085. Tn5084-Tn5085 was inserted in both orientations. 3.2.4. Transposon Tn5084 is located on the chromosome of the B. cereus VKM684 (ATCC10702) strain

The B. cereus VKM684 chromosomally located mer determinant has previously been shown to be identical to the RC607-TC38-2b determinant [7]. In this study restriction digests of total VKM684 DNA

E. Bogdanova et al. / Res. Microbiol. 152 (2001) 503–514

511

3.2.5. The structure of transposon Tn5083 provides evidence for recombination, both between transposition and mercury resistance genes from different transposons

Figure 7. Neighbor-joining distance dendrogram of the TnpR amino acid sequences was constructed by PHYLO_WIN program [12]. Pairwise distances were computed by Poisson correction. The distance scale indicates the number of amino acid substitutions per site. Bootstrap percentages (500 replicates) are shown to the left of the node being considered. Ac for the TnpR sequences: Tn21 from E. coli (M10791), Tn1546 from Enterococcus faecium (Q06237), Tn5085 from Exiguobacterium TC38-2b (Y17750), Tn5084 from B. cereus RC607 (Y17748), Tn917 from Enterococcus faecalis (P06693), Tn5083 from B. megaterium MK64-1 (Y18009), Tn from B. firmus (B42707), Tn552 from S. aureus (P18358).

using EcoRI, HindIII, BglII, NcoI, PstI and AvaII, and Southern blot hybridization identified elements of a complete transposition module indistinguishable from Tn5084 (figure 1). In the region downstream of tnpA, at the end of the putative transposition module of the transposon from VKM684, we found an AvaII site located inside IRR , as in Tn5084 (figure 5). The homology of the transposition module from VKM684 with the labelled DNA probe from pKLH3.7 (from TC38-2b), was interrupted, as in the case of Tn5084, immediately beyond this AvaII site. At the left flank of the transposon, downstream of merB1, no NcoI or PstI restriction sites like those found in the DNA flanking Tn5084 in RC607, were found (figure 1). The inserts from pYW33.2 and pYW33.3, carrying the Tn5084 flanking DNA did not hybridize with VKM684 (figure 1). This suggests that a transposon, probably indistinguishable from Tn5084 from RC607 but having different flanking sequences, might be localized in the B. cereus VKM684 chromosome.

Earlier we had identified an RC607-like mer determinant in the pKLH304 plasmid from B. megaterium MK64-1, and we had cloned the mer determinant containing a fragment of this plasmid as the pKLH304.1 clone [7]. In the present study we analyzed this clone and found that the RC607-like mer determinant is located in the mercury resistance class II transposon which we have named Tn5083. We determined the nucleotide sequence upstream of the merB3 gene, from the DraI-site found in TnMERI1, Tn5085 and Tn5084 (figure 5). The alignment in figure 4A showed that the sequence of Tn5083 was identical in all four transposons up to the putative resI site. The homology ended abruptly at the center of the putative resI site. Further on, tnpR and tnpA were clearly evident in the Tn5083 sequence, and these genes were transcribed in the opposite direction to the transcription of the mer operon (figures 3 and 5). Thus, the Tn5083 mer determinant was located inside the transposition module whose structure was close to those of Tn5084-Tn5085 in organization, but considerably different in the DNA sequence. Over the sequenced region the Tn5083 transposition module differed from Tn5085 by 40% nucleotide substitutions. The highest similarity score obtained from BLAST analysis of the Tn5083 transposition module was with resolvase and transposase from a putative transposon from B. firmus [22], showing about 12% difference in nucleotide sequence. The mercury resistance determinant from Tn5083 did not contain merR2, merB2 or merB1 genes, and had, on an average, 3% nucleotide substitutions compared with the same region of the Tn5085Tn5084 mer determinant. These nucleotide changes were not randomly distributed throughout the length of the mer determinant (figure 3). The complete structure of Tn5083 and its functional activity remain unknown. 4. Discussion The published data and the results presented above show that a number of different class II mercury resistant transposons are found in Gram-positive bacteria from the environment: the broad-spectrum RC607 mer determinant can be located on closely related class II transposons: TnMERI1 [20], Tn5085 and

512

E. Bogdanova et al. / Res. Microbiol. 152 (2001) 503–514

Tn5084, and a RC607-like mer determinant – on Tn5083. Three sets of facts indicate the real participation of these transposons in the spread of the mer determinants in the environment. 1) Tn5084 can be localized both at the plasmid in the RC607 strain and at the chromosome in the VKM684 strain. 2) Five-bp direct repeats at the outer ends of the IRs (figure 1) are a strong indication of the true mobility of Tn5084 and Tn5085. 3) Data on recombinant (mosaic) transposons, shown in this paper, also point to the essential role of transposons in the transfer of mer operons. Tn5084 may be regarded as a recombinant transposon formed from Tn5085 and from an unknown transposon differing from Tn5085 by 10% nucleotide substitutions. Tn5084 may have been formed by two recombination acts: the first into homologous res subsites or the spacer (site-specific recombination), and the second near the end of the tnpR gene, as shown by a one-nucleotide deletion which destroys the translation frame and leads to TnpR elongation for 4 amino acids in Tn5084 (the recombination mechanism for this event is unknown) (figures 3 and 4B). Tn5083 is also a mosaic. Analysis of the sequence of its mercury resistance determinant (Ac:Y09907) shows that it includes regions from two different sequence types: merB3, merR1 and open reading frames (ORFs) with 0.2% nucleotide substitutions compared with the same regions of Tn5085, and a 692-bp region inside merA with 0.6% nucleotide substitutions compared with the same region of Tn5085. The fragments of merA, coding for the N- and C-terminal regions of mercuric reductase, have 6% and 8.5% nucleotide substitutions compared to Tn5085 and represent a novel sequence variant which has not been described previously (figure 3). It seems reasonable to assume that during the formation of Tn5083, Tn5085 lost its transposition module as a result of site specific recombination at homologous resI sites (figure 4A) and acquired a transposition module related to the B. firmus transposition module (the B. firmus transposon also contains similar genes to the cadmium resistance operon from S. aureus). The absence of several genes (merB1, merB2 and merR2, approximately 3 kb) at the end of the Tn5083 operon compared with the Tn5085-Tn5084 operons may indicate some additional features of Tn5083 transposon formation (figure 3).

Our results show that transposition gene exchange by recombination in the res region may be similarly widespread in the evolution of mercury resistance transposons of Gram-positive bacteria, as it is for transposons from Gram-negative bacteria [44]. The group of closely related transposons Tn5085, TnMERI1 and Tn5084 is globally distributed and can be found in diverse environmental habitats (soil in Carpathia (Ukraine), Minamata Bay (Japan) and Boston Harbor (USA)). The geographical spread of these transposons is also accompanied by horizontal transfer between bacterial genera (Exiguobacterium and Bacillus). One can presume that the global spread of these transposons is accounted for by the properties of both their transposition modules and the particular features of their mer determinants. We have shown for the first time in this work that Tn5085 and probably Tn5084, both from Gram-positive bacteria, are functional in cells from evolutionarily remote hosts, Gram-negative E. coli. Their mercury resistance determinant is able to express in E. coli [7, 15, 21, 42]. The presence of three different merB genes in this mer determinant ensures the resistance of bacteria to a wide range of organomercurials [15, 21] and probably favors the highly successful colonization of a wide set of bacteria (Bacillus, Exiguobacterium, and Clostridium) [31]. The cause of the inability of Tn5085 and Tn5084 to resolve co-integrates in E. coli remains unknown. It is possible that the resolution systems of these transposons are active only in their natural hosts, or they have defective resolution systems and use their natural hosts recombination system. The location of the mercury resistance genes on the HgR plasmid (pKLH6) in B. cereus RC607 was surprising, as all other studies reported the exclusively chromosomal localization of this mer determinant in Bacillus strains, including strain RC607. It is possible that during our work a transposition of Tn5084 from the chromosome to the plasmid took place to form pKLH6. However, the fact that the Tn5084-flanking DNA sequences in pYW33, pKLH6, and in the total DNA preparations are identical (see section 3.1) refutes this idea. The pKLH6 plasmid could be simultaneously found in the autonomous state and in the chromosomeintegrated state. But we have found that pKLH6 is a naturally temperature-sensitive plasmid and can be lost without selective pressure at 42◦ C (temperature

E. Bogdanova et al. / Res. Microbiol. 152 (2001) 503–514

sensitivity has been described, for example, for some plasmids from S. aureus and B. thuringiensis during their replication in B. subtilis cells [26, 38]). The high segregation frequency of HgS clones during the growth of the RC607 strain at the elevated temperature (see section 2.4) suggests that this hypothesis is also unlikely. It is also possible that pKLH6 was integrated into the chromosome of the initial RC607 strain and excised from it during storage in our laboratory. Integration of a plasmid into the chromosome, or excision from it, is known in Bacillus [9, 19, 39]. However, it more likely that the HgR plasmid could not be previously isolated using the existing DNA extraction methods, and this was only possible in this study by using a method specifically designed for plasmid extraction from environmental Bacillus strains ([7] and Materials and methods). This work has shown that the broad-spectrum RC607 mer determinant from Gram-positive bacteria may spread in the environment with the help of plasmids and transposons analogous to the dispersal of Gram-negative HgR determinants [32, 33, 44]. Acknowledgements We thank I. Mahler (Brandeis University, Waltham, MA) for the Bacillus sp. RC607 strain and the plasmids from this strain. We thank K. Brodolin for help in tree construction and G. Kholodii for guidance in transposition experiments and helpful discussion. We thank Nigel Brown, Sofia Mindlin and Steve Kidd for critical reading of this manuscript. We thank E.I. Molchanova for technical assistance. This research was supported by the Russian Foundation for Basic Research, grant 99-04-48559; the Foundation for Support of Leading Scientific Schools, grant 0015-97751; and the State Science and Technology program ‘Novel Methods of Bioengineering’. JLH was supported by a grant (6/B10333 to Professor N.L. Brown) from the Biotechnology and Biological Sciences Research Council. The EMBL accession numbers for the nucleotide sequences of the transposon regions reported in this paper are: Tn5083 (B. megaterium MK64-1, pKLH304), Y18009; Tn5084 (B. cereus RC607, pKLH6), AJ277277, Y17741, Y17748, Y17749; Tn5085 (Exiguobacterium sp. TC38-2b, pKLH3), Y17750, Y17751, Y17752.

513

References [1] Altschul S.F., Gish W., Miller W., Myers E.W., Lipman D.J., Basic local alignment search tool, J. Mol. Biol. 215 (1990) 403–410. [2] Arthur M., Molinas C., Depardieu T., Courvalin P., Characterization of Tn1546, a Tn3 -related transposon conferring glycopeptide resistance by synthesis of depsipeptide peptidoglycan precursors in Enterococcus faecium BM4147, J. Bacteriol. 175 (1993) 117–127. [3] Belliveau B.H., Trevors J.T., Mercury resistance determined by a self-transmissible plasmid in Bacillus cereus 5, Biol. Metals 3 (1990) 188–196. [4] Birnboin H.C., Doly J., A rapid alkaline extraction procedure for screening recombinant plasmid DNA, Nucl. Acids Res. 7 (1979) 1513–1523. [5] Bogdanova E.S., Mindlin S.Z., Occurrence of two structural types of mercury reductases among Gram-positive bacteria, FEMS Microbiol. Lett. 78 (1991) 277–280. [6] Bogdanova E.S., Mindlin S.Z., Pakrova E., Kocur M., Rouch D., Mercuric reductase in environmental Gram-positive bacteria sensitive to mercury, FEMS Microbiol. Lett. 97 (1992) 95–100. [7] Bogdanova E.S., Bass I.A., Minakhin L.S., Petrova M.A., Mindlin S.Z., Volodin A.A., Kalyaeva E.S., Tiedjie J.M., Hobman J.L., Brown N.L., Nikiforov V.G., Horizontal spread of mer operons among Gram-positive bacteria in natural envinronment, Microbiology 44 (1998) 609–620. [8] Danilevich V.N., Stepashin Yu.G., Volozhantsev N.N., Volkovoi K.I., Isolation and characterization of the deletion mutants of temperature-sensitive plasmid pEG1, Genetika (Moscow) 16 (1980) 1958–1966. [9] Conrad B., Bashkirov V.I., Hofemeister J., Imprecise excision of plasmid pE194 from the chromosomes of Bacillus subtilis pE194 insertion strains, J. Bacteriol. 174 (1992) 6997–7002. [10] Dornbush A.C., Pelcak E.J., Assay of aureomycin level in body fluid, Ann. NY Acad. Sci. 51 (1948) 218–220. [11] Felton J., Michaelis S., Wright A., Mutation in two unliked genes are required to produce asparagine auxotrophy in Escherichia coli , J. Bacteriol. 142 (1980) 221–228. [12] Galtier N., Gouy M., Gautier C., SEAVIEW and PHYLO_WIN: two graphic tools for sequence alignment and molecular phylogeny, CABIOS 12 (1996) 543–548. [13] Grinsted J., de la Cruz F., Altenbuchner J., Schmitt R., Complementation of transposition of tnpA mutants of Tn3, Tn21, Tn501, and Tn1721, Plasmid 8 (1982) 276–286. [14] Grinsted J., de la Cruz F., Schmitt R., The Tn21 subgroup of bacterial transposable elements, Plasmid 24 (1990) 163– 189. [15] Gupta A., Phung L.T., Chakravarty L., Silver S., Mercury resistance in Bacillus cereus RC607: transcriptional organization and two new open reading frames, J. Bacteriol. 181 (1999) 7080–7086. [16] Hart M.C., Elliott G.N., Osborn A.M., Ritchie D.A., Strike P., Diversity amongst Bacillus merA genes amplified from mercury resistant isolates and directly from mercury polluted soil, FEMS Microbiol. Ecology 27 (1988) 73–84. [17] Hobman J.L., Brown N.L., Bacterial Mercury Resistance Genes, in: Sigel H., Sigel A. (Eds.), Metal Ions In Biological

514

[18]

[19]

[20]

[21]

[22]

[23]

[24]

[25]

[26]

[27]

[28]

[29] [30]

[31]

E. Bogdanova et al. / Res. Microbiol. 152 (2001) 503–514 Systems, Vol. 34, Marcel Dekker Inc., New York, 1997, pp. 527–567. Hobman J.L., Kholodii G.Ya., Nikiforov V.G., Ritchie D.A., Strike P., Yurieva O.V., The sequence of the mer operon of pMER 327/419 and transposon ends of pMER 327/419, 330 and 05, Gene 146 (1994) 73–78. Hofemeister J., Israeli-Reches M., Dubnau D., Integration of plasmid pE194 at multiple sites on the Bacillus subtilis chromosome, Mol. Gen. Genet. 189 (1983) 58–68. Huang Ch.-Ch., Narita M., Yamagata T., Itoh Y., Endo G., Structure analysis of a class II transposon encoding the mercury resistance of the Gram-positive bacterium Bacillus megaterium MB1, a strain isolated from Minamata Bay, Japan. Gene 234 (1999) 361–369. Huang Ch.-Ch., Narita M., Yamagata T., Endo G., Identification of three mer B genes and characterization of a broadspectrum mercury resistance module encoded by a class II transposon of Bacillus megaterium strain MB1, Gene 239 (1999) 361–366. Ivey D.M., Guffanti A., Shen Z., Kudyan N., Krulwich T.A., The cadC gene product of alkaliphilic Bacillus firmus OF4 partially restores Na+ resistance to an Escherichia coli strain lacking an Na+ /H+ antiporter (NhaA), J. Bacteriol. 174 (1992) 4878–4884. Jacob A.E., Shapiro J.A., Yamamoto L., Smith D.L., Cohen S.N., Berg D., Plasmid studied in Escherichia coli and other enteric bacteria, in: Bukhari A.I., Shapiro J.A., Adhya S.L. (Eds.), DNA Insertion Elements, Plasmids, and Episomes, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York, 1977, pp. 607–638. Kholodii G.Ya., Mindlin S.Z., Bass I.A., Yurieva O.V., Minakhina S.V., Nikiforov V.G., Four genes, two ends, and a res region are involved in transposition of Tn5053 : a paradigm or a novel family of transposons carrying either a mer operon or an integron, Molecular Microbiology 17 (1995) 1189–1200. Kholodii G.Ya., Yurieva O.V., Gorlenko Zh.M., Mindlin S.Z., Bass I.A., Lomovskaya O.L., Kopteva A.V., Nikiforov V.G., Tn5041: a chimeric mercury resistance transposon closely related to the toluene degradative transposon Tn4651, Microbiology 143 (1997) 2549–2556. Lereclus D., Guo S., Sanchis V., Lecadet M.-M., Characterization of two Bacillus thuringiensis plasmids whose replication is thermosensitive in B. subtilis , FEMS Microbiol. Let. 49 (1988) 417–422. Liebert C.A., Wireman J., Smith T., Summers A.O., Phylogeny of mercury resistance (mer ) operons of Gramnegative bacteria isolated from the fecal flora of primates, Appl. Envir. Microb. 63 (1997) 1066–1076. Mahler I., Levinson H.S., Wang Y., Halvorson H.O., Cadmium and mercury-resistant Bacillus strains from a salt marsh and from Boston Harbor, Appl. Environ. Microbiol. 52 (1986) 1293–1298. Misra T.K., Bacterial resistances to inorganic mercury salts and organomercurials, Plasmid 25 (1992) 4–16. Nakamura K., Silver S., Molecular analysis of mercuryresistant Bacillus isolates from sediment of Minamata Bay, Japan. Appl. Environ. Microbiol. 60 (1994) 4596–4599. Narita M., Koizumi T., Huang C., Endo G., Broad-spectrum mercury resistance and its genetic characterization of

[32]

[33]

[34]

[35]

[36]

[37]

[38]

[39]

[40] [41]

[42]

[43]

[44]

[45]

anaerobic mercury-resistant bacterium, Clostridium butyricum Mersaru, isolated from sediment of Minamata Bay, Japan. DDBJ/EMBL/GenBank submission (1999) Ac: AB024961. Nikiforov V.G., Bass I.A., Bogdanova E.S., Gorlenko, Zh.M., Kalyaeva E.S., Kopteva A.V., Lomovskaya O.V., Minakhin L.S., Minakhina S.V., Mindlin S.Z., Petrova M.A., Kholodii G.Ya., Yurieva O.V., Distribution of mercury resistance transposons in environmental bacterial populations, Molekularnaya Biologiya 33 (1999) 55–62. Osborn A.M., Bruce K.D., Strike P., Ritchie D.A., Distribution, diversity and evolution of the bacterial mercury resistance (mer ) operon, FEMS Microbiol. Rev. 19 (1997) 239–262. Pansegrau W., Lanka E., Barth P.T., Figurski D.H., Guiney D.G., Haas D., Helinski D.R., Schwab H., Stanisich V.A., Thomas C.M., Complete nucleotide sequence of Birmingham IncP-alpha plasmids. Compilation and comparative analysis, J. Mol. Biol. 239 (1994) 623–663. Pearson A.J., Bruce K.D., Osborn A.M., Ritchie D.A., Strike P., Distribution of class II transposase and resolvase genes in soil bacteria and their association with mer genes, Appl. Environ. Microbiol. 62 (1996) 2961–2965. Robinson J.B., Tuovinen O.H., Mechanisms of microbial resistance and detoxification of mercury and organomercury compounds: physiological, biochemical and genetic analyses, Microbiol. Rev. 48 (1984) 95–124. Sambrook J., Fritsch E.F., Maniatis T., Molecular cloning: A Laboratory Manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York, 1989. Scheer-Abramowitz J., Gryczan T.J., Dubnau D., Origin and mode of replication of plasmids pE194 and pUB110, Plasmid 6 (1981) 67–77. Shishido K., Noguchi N., Kim C., Ando T., Isolation of a tetracycline-resistance plasmid excised from a chromosomal DNA sequence in Bacillus subtilis , Plasmid 10 (1983) 224–234. Silver S., Phung L.T., Bacterial heavy metal resistance: new surprises, Annu. Rev. Microbiol. 50 (1996) 753–789. Wang Y., Mahler I., Levinson H.S., Hallorson H.O., Cloning and expression in Escherichia coli of chromosomal mercury resistance genes from a Bacillus sp., J. Bacteriol. 69 (1987) 4848–4851. Wang Y., Moore M., Levinson H.S., Silver S., Walsh C., Mahler I., Nucleotide sequence of a chromosomal mercury resistance determinant from a Bacillus sp. with broadspectrum mercury resistance, J. Bacteriol. 171 (1989) 83– 92. Yurieva O., Nikiforov V., Catalytic center quest: Comparison of transposases belonging to the Tn3 family reveals an invariant triad of acidic amino acid residues, Biochem. Mol. Biol. International. 38 (1996) 15–20. Yurieva O.V., Kholodii G.Ya., Minakhin L.S., Gorlenko, Zh.M., Kalyaeva E.S., Mindlin S.Z., Nikiforov V.G., Intercontinental spread of promiscous mercury-resistance transposons in environmental bacteria, Mol. Microbiol. 24 (1997) 321–329. Zharkikh A.A., Rzhetsky A.Yu., Morosov P.S., Sitnikova T.L., Krushkal J.S., VOSTORG: a package of microcomputer programs for sequence analysis and construction on phylogenetic trees, Gene 101 (1991) 251–254.