ABSTRACT
The α-proteobacterial strain Sphingobium japonicum UT26 utilizes a highly chlorinated pesticide, γ-hexachlorocyclohexane (γ-HCH), as a sole source of carbon and energy, and haloalkane dehalogenase LinB catalyzes the second step of γ-HCH degradation in UT26. Functional complementation of a linB mutant of UT26, UT26DB, was performed by the exogenous plasmid isolation technique using HCH-contaminated soil, leading to our successful identification of a plasmid, pLB1, carrying the linB gene. Complete sequencing analysis of pLB1, with a size of 65,998 bp, revealed that it carries (i) 50 totally annotated coding sequences, (ii) an IS6100 composite transposon containing two copies of linB, and (iii) potential genes for replication, maintenance, and conjugative transfer with low levels of similarity to other homologues. A minireplicon assay demonstrated that a 2-kb region containing the predicted repA gene and its upstream region of pLB1 functions as an autonomously replicating unit in UT26. Furthermore, pLB1 was conjugally transferred from UT26DB to other α-proteobacterial strains but not to any of the β- or γ-proteobacterial strains examined to date. These results suggest that this exogenously isolated novel plasmid contributes to the dissemination of at least some genes for γ-HCH degradation in the natural environment. To the best of our knowledge, this is the first detailed report of a plasmid involved in γ-HCH degradation.
γ-Hexachlorocyclohexane (γ-HCH) (also called γ-BHC and lindane) is a halogenated organic insecticide which was once used worldwide but has since been prohibited in most countries due to its toxicity and long persistence in upland soil. Although γ-HCH is a representative man-made xenobiotic, some γ-HCH-degrading bacterial strains have been isolated and characterized (30, 32, 35, 42, 43, 47). Genes and enzymes for γ-HCH degradation were first well characterized for an α-proteobacterial strain, Sphingobium japonicum UT26, which converts γ-HCH to β-ketoadipate through the action of six enzymes: LinA (dehydrochlorinase), LinB (halidohydrolase), LinC (dehydrogenase), LinD (reductive dechlorinase), LinE (ring cleavage dioxygenase), and LinF (reductase) (14, 40). The linA-to-linF genes in UT26 are dispersed on the three large circular replicons: the linA, linB, and linC genes on the 3.6-Mb chromosome I; the linF gene on the 670-kb chromosome II; and the linDE operon with its regulatory gene (linR) on a 185-kb plasmid, pCHQ1 (39). Nearly identical lin genes have also been identified in other HCH-degrading bacterial strains, such as Sphingobium indicum B90 (31) and B90A (12) from India and Sphingobium francense Sp+ from France (7); most of the lin genes in these strains are closely associated with an insertion sequence, IS6100 (7, 32). pCHQ1 is conjugally transferable from UT26 to another Sphingomonas paucimobilis strain (39), and a recent report showed that the linA and linB genes in other strains are also located on plasmids (7). These observations indicate that lin genes must be spread by mobile genetic elements (MGEs).
The recent determination of various bacterial genome sequences has revealed that the horizontal transfer of various phenotypic genes has played a significant role in the evolution of bacteria and in their adaptation to environmental changes (17, 45). There is no doubt that various MGEs have contributed greatly to horizontal gene transfer (44, 46, 48, 61), but the direct detection of such events in natural bacterial communities is rare. Exogenous isolation of MGEs, designated “exogenous plasmid isolation” when such MGEs are plasmids, has been developed to capture transferable plasmids directly from the natural microbial community by using bacterial conjugation systems (56). This technique involves conjugal mating of a suitable recipient with a natural microbial community and subsequent selection of transconjugants that have acquired a genetic marker. This approach has been used for the isolation of plasmids involved in resistance to mercury and in the degradation of 2,4-dichlorophenoxyacetic acid and naphthalene (3, 4, 54, 58, 60). Considering the fact that most bacterial cells in the environment are not easily culturable by conventional techniques (1), the exogenous plasmid isolation technique will provide novel insights into horizontal gene transfer in the natural environment.
In the present study, the exogenous plasmid isolation technique was applied to isolate a gene for LinB activity from HCH-contaminated soil. Considering the fact that LinB is a key enzyme in the degradation not only of γ-HCH but also of β-HCH (41), the linB gene is one of the most important genes in sites contaminated by a technical mixture of HCH (t-HCH) that consists of α, β, γ, and δ isomers. The present report provides a snapshot of the dynamism of the linB gene in the natural environment.
MATERIALS AND METHODS
Bacterial strains, plasmids, and media.The bacterial strains and plasmids used in the present study are listed in Table 1. Escherichia coli cells were grown at 37°C in Luria-Bertani (LB) broth (34), and Sinorhizobium meliloti and Mesorhizobium loti cells at 30°C in TY medium (6). Other strains were grown in 1/3LB broth (14) or W minimal medium (20) supplemented with γ-HCH as a carbon source at 30°C. The solid media were prepared by the addition of 1.5% agar. Antibiotics were used at final concentrations of 50 μg/ml for ampicillin and phosphomycin (Pho), 25 μg/ml for kanamycin (Km), 10 μg/ml for gentamicin (Gm), 15 μg/ml for tetracycline (Tc), 200 μg/ml for trimethoprim, and 1,000 μg/ml for streptomycin (Sm).
Bacterial strains and plasmids used in this study
Methods of DNA manipulation.Established methods were used for the preparation of plasmid DNA, its digestion with restriction endonucleases, ligation, agarose gel electrophoresis, and the transformation of E. coli cells (34). Large plasmids from α-proteobacterial strains were extracted using the method described by Kado and Liu (22), and the transformation of bacterial cells by electroporation was performed as described previously (27). Nucleotide sequencing was performed with an ABI PRISM model 310 sequencer (Applied Biosystems, Foster City, CA). Southern blot analysis was carried out using the conventional protocols and a digoxigenin system (Roche Diagnostics, Mannheim, Germany). For the preparation of the linB probe, a linB gene fragment was amplified by PCR with UT26 total DNA as a template and the primers Bgl-linB-Xho.R (5′-GGGCTCGAGGATTATGCTGGGCGCAATC-3′) and linB.F (5′-TAAGGAGGAATATCGATGAGCCTC-3′). For the repA probe of pLB1, we used plasmid pLB1Tc (see below) as a template and the primers Hin_pLB1_65868.F (5′-CCCAAGCTTGTGCCACCGAAGTGAGC-3′) and pLB1_894_Bam.R (5′-CCCGGATCCGAACTTCTTCCGTCAACG-3′).
Microbiological methodology for a soil sample.The clay soil sample used in the present study was obtained from a field which had been contaminated with HCH isomers in Miyagi Prefecture, Japan. To determine the most probable number (MPN) of γ-HCH-degrading cells in the soil, 1-ml portions of serially diluted soil bacterial suspensions were inoculated into 9 ml of 1/10 W [containing the following per liter: KH2PO4, 170 mg; Na2HPO4, 980 mg; (NH4)2SO4, 100 mg; MgSO4, 48.7 mg; FeSO4, 0.52 mg; MgO, 10.75 mg; CaCO3, 2.0 mg; ZnSO4, 0.81 mg; CuSO4, 0.16 mg; CoSO4, 0.15 mg; and H3BO3, 0.06 mg) containing 10 ppm of γ-HCH and were incubated at 30°C for 2 weeks with shaking. An equal volume of ethyl acetate was mixed with each suspension to extract the solution that was used for the analysis of γ-HCH degradation with a Shimadzu GC-17A gas chromatograph equipped with an electron capture 63Ni detector (Shimadzu, Kyoto, Japan) as described previously (41).
Exogenous plasmid isolation was performed using a modified version of the procedure described by Top et al. (60). The bacterial fraction in 10 g of soil sample was suspended in phosphate-buffered saline (PBS) and filtered through 7-μm-pore-size filter paper (Advantec MFS, Inc., Tokyo, Japan) to remove nonbacterial materials. The filtered fraction was centrifuged at 10,000 × g for 20 min, and the cell pellet was resuspended in 0.5 ml of PBS buffer and mixed with 1 ml of an overnight culture of the recipient (UT26DB) cells. This mating mixture was concentrated by centrifugation at 10,000 × g for 2 min and spotted onto 1/3LB agar plates supplemented with cycloheximide (100 μg/ml). After overnight incubation at 30°C, cells were suspended in 1 ml of PBS, diluted, and spread on W agar plates supplemented with γ-HCH and Km.
Sequencing and computer analysis of pLB1.A 1.9-kb PstI fragment of pLB1 (bp 20285 to 22160) was cloned into the PstI site of an R6K-based Tc-resistant (Tcr) plasmid, pJP5608. The resulting plasmid, pRM26, was conjugally transferred from E. coli S17-1λpir to UT26DBT1 (see below), and selection was made for resistance to Tc. One such transconjugant carried the pLB1 derivative (pLB1Tc) in which pRM26 was integrated into pLB1 by single-crossover-mediated homologous recombination between the common PstI fragment shared by the two plasmids. To avoid contamination of endogenous plasmids in UT26DB, pLB1Tc was conjugally transferred from UT26DB to E. coli S17-1λpir and purified by a standard alkaline lysis method (34). Shotgun sequencing of pLB1Tc was performed, the DNA fragments not covered by shotgun sequencing were directly amplified by PCR, and the nucleotide sequences of the resulting PCR products were determined.
Sequence assembly and calculation of G+C contents were performed using Genetyx 13 software (SDC, Tokyo, Japan). Annotation was performed using the GenBank CDS translations/PDB/SwissProt/PIR/PRF protein databases and the DDBJ/EMBL/GenBank DNA databases with the BLAST program (http://www.ncbi.nlm.nih.gov/BLAST/ ). The conserved domains were searched using the InterProScan (http://www.ebi.ac.uk/InterProScan/ ) and the Pfam (http://www.sanger.ac.uk/Software/Pfam/search.shtml ) programs. The putative amino acid sequences were used to generate a neighbor-joining tree by running the CLUSTALW program (http://www.ddbj.nig.ac.jp/search/clustalw-j.html ).
Replication analysis of pLB1.The 2-kb DNA fragment containing the repA gene and its upstream region (bp 64908 to 894) was amplified by PCR using pLB1Tc as a template and the primers pLB1_64908.F (5′-TAAAAGCTTTCATCGCTTTCTCC-3′) and pLB1_894_Bam.R (see above). The amplicon was cloned into the multiple cloning site of a narrow-host-range vector, pK18mob (Kmr), to obtain pK18OR as a minireplicon. Insertion of the 3.3-kb BamHI-DraI fragment of pLB1 (bp 36846 to 40144) (see Fig. 2 and Table 2) into pK18mob generated pK18BD3, which was used as a negative control. A broad-host-range vector, pBBR1MCS-2 (Kmr), was used as a positive control. pK18OR, pK18BD3, and pBBR1MCS-2 (150 ng each) were introduced into the UT26 cells by electroporation to obtain Kmr transformants.
Genes, sites, and specific fragments on pLB1
Filter matings between pure cultures.The donor and recipient cells grown overnight were harvested by centrifugation, washed with TY broth, and resuspended in fresh TY broth. They were then mixed and subsequently spotted on a sterile 0.45-μm-pore-size cellulose acetate filter (Advantec) placed on a TY agar plate. After incubation at 30°C for 13 h, the cells on the filter were suspended in TY broth, diluted, and plated on selective agar plates. S. meliloti 1021, M. loti MAFF303099, and E. coli HB101 were selected with Sm, Pho, and Sm, respectively. The chromosomes of other recipient strains (Pseudomonas putida KT2440, S. paucimobilis IAM12578, Burkholderia multivorans ATCC 17616, and Acidovorax sp. strain KKS102) were marked with the TnMod-OGm-derived Gmr gene. The introduction of this gene into the genomes of the parental strains was carried out by electroporation-mediated mutagenesis using pTnMod-OGm (11).
Nucleotide sequence accession number.The complete sequence for the circular plasmid pLB1 has been deposited in DDBJ/EMBL/GenBank under accession number AB244976.
RESULTS
Exogenous isolation of a plasmid involved in γ-HCH degradation.In the present study, we used a clay soil sample from Miyagi Prefecture, Japan. The soil had been contaminated by t-HCH, and still contained low levels of HCH isomers (0.03 to 0.1 mg/liter, mainly α-HCH). The number of culturable bacterial cells on each 1/100 NB (30 mg of beef extract per liter and 50 mg of peptone per liter) agar plate was 1.4 × 107 CFU/g of soil. The number of indigenous γ-HCH-degrading bacterial cells was estimated to be at least 103 MPN/g of soil. The bacterial fraction collected from the soil was mated with a linB mutant of UT26 (UT26DB) (41), using the protocol described in Materials and Methods. We obtained one putative transconjugant, UT26DBT1, that was able to utilize γ-HCH as a sole carbon source on the W agar plate plus Km. The control experiments without UT26DB did not give rise to such transconjugants. We further confirmed that UT26DBT1 was indeed a transconjugant of UT26DB by the following three results: (i) the 16S rRNA gene of UT26DBT1 was completely identical to that of UT26; (ii) linB::Km-specific DNA fragments were amplified by PCR using the UT26DBT1 colony as a template; and (iii) three UT26-endogenous plasmids, pCHQ1 (185 kb), pUT1 (30 kb), and pUT2 (5 kb), were also observed in UT26DBT1 (Fig. 1). UT26DBT1 harbored one additional plasmid, designated pLB1. Southern blot analysis with the linB gene as a probe indicated that pLB1 carries the linB homologue(s) (Fig. 1).
Plasmids residing in UT26DB and its transconjugant UT26DBT1. Gel electrophoresis of plasmids (a) and Southern blot analysis with the linB gene as a probe (b) are shown. Lanes 1, UT26DB; lanes 2, UT26DBT1. The white arrowheads indicate endogenous plasmids pCHQ1 (185 kb), pUT1 (30 kb), and pUT2 (5 kb), respectively. The black arrowhead indicates pLB1.
Sequence analysis of pLB1.Our determination of the entire nucleotide sequence of pLB1Tc purified from E. coli S17-1λpir cells revealed that its parental plasmid, pLB1, is 65,998 bp in size. The average G+C content of pLB1 is 60.2%, and this score approximately corresponds to those of plasmids from other α-proteobacterial strains, such as pNL1 (62%) from Sphingomonas aromaticivorans F199, pSymA (60%) from S. meliloti 1021, and pMLa (59%) from M. loti MAFF303099. pLB1 was found to carry 50 coding sequences (CDSs) (Fig. 2 and Table 2), which represent 72.1% of the total plasmid DNA. The remaining 18,385-bp regions consist of partial gene fragments and intergenic regions. The putative products of 24 of 50 CDSs showed low similarity (less than 48% identity) to those deposited in public databases.
Circular map of pLB1. CDSs or gene remnants outside the circle are transcribed in the clockwise direction and those inside in the counterclockwise direction. The putative functions of genes are shown by the following colors: red, replication and stable inheritance; orange, conjugative transfer; blue, transposase and resolvase; magenta, regulation; yellow, unknown; and green, others. Gene remnants are shown in gray. Large and small circles indicate the putative oriV and oriT regions, respectively. The region of the IS6100-linB cluster is shown by the light green arrow. The positions of tnp in the IS6100-linB cluster are almost equal to those of IS6100 (Table 2).
(i) The IS6100-linB cluster.pLB1 carried two directly oriented copies of genes whose nucleotides were identical to linB from S. indicum B90A. To confirm the functionality of the two pLB1-specified linB genes, pRM14 and pRM15 were constructed such that each carried one of the two linB genes (bp 17148 to 18531 and 14965 to 16348, respectively) (Table 1). The introduction of these plasmids into UT26DB gave rise to transformants able to grow on a W agar plate containing γ-HCH as a carbon source, indicating that both linB genes on pLB1 are functional. The two linB genes and three complete and directly oriented copies of IS6100 were clustered in the order IS6100-linB-IS6100-linB-IS6100, and this cluster was preceded by two truncated versions of this IS element. This 6-kb region including the IS6100 remnants (bp 13187 to 19323) was designated the IS6100-linB cluster. Although the transposition of IS6100 usually generates 8-bp direct repeats of its target site (8), no obvious duplication of such a target site was observed around this cluster.
(ii) Genes for replication.The predicted pLB1 replication initiator protein, RepA, is unique because only three homologues with homology at a low level were found in databases: RepA of Agrobacterium radiobacter K84 plasmid pAgK84 (33% identity), Rep of Agrobacterium tumefaciens 1D1422 plasmid pTAR (31% identity), and a hypothetical protein of a Zymomonas mobilis ZM4 plasmid (27% identity). An InterProScan search revealed that the pLB1 RepA protein contains in its middle part a putative winged-helix motif that would promote the binding to DNA. Although the upstream region of the repA gene (bp 64921 to 65998) showed no similarity to other well-characterized replication origin (oriV) sequences, this region was relatively A+T rich (55.8% of G+C content) and was found to contain seven 9-bp direct repeats (5′-kCwAwCwsd-3′) and one inverted repeat (Fig. 3). The direct repeats may serve as iterons (9), which are the interaction sites of iteron-type replication initiator proteins. Additionally, two putative DnaA boxes (38) were also located in this region (bp 64945 to 64953 and 65134 to 65142).
Structure of the putative oriV region of pLB1. The 9-bp direct repeats (5′-kCwAwCwsd-3′), which may serve as iterons, are shown by solid arrows, and one inverted repeat is indicated by dashed arrows. Putative DnaA-binding sequences are boxed, and parts of repA and orf21 are shaded. Putative ribosome-binding sites and start codons of orf21 and repA are shown in boldface.
(iii) Genes for stable inheritance.The parA-parB2 cluster (bp 47479 to 49253) is considered to be involved in the partitioning of pLB1. The putative ParA protein shows 28% identity to an IncC-like protein of S. meliloti MBA19 plasmid pMBA19a (Table 2) and belongs to the Walker-type ATPases (63). A phylogenetic tree of partitioning ATPases, in which each clade is thought to coincide with the genetic organization of the par loci (15, 16), demonstrated that ParA of pLB1 belongs to a novel clade (Fig. 4). The putative ParB2 protein has a functional domain (IPR003115) required for its binding to a specific DNA sequence (partition site). Another putative partitioning gene, parB1 (bp 3079 to 4854), was found far away from the parA-parB2 cluster. Although ParB1 also had the predicted functional domain, its C-terminal region, which was postulated to be involved in dimerization, showed a very low level of similarity to that of ParB2.
Phylogenetic tree of putative Walker-type partition ATPases. Lengths of horizontal lines reflect relative evolutionary distances among the 33 Walker-type partition ATPases encoded by various plasmids or chromosomes. The GenBank accession numbers of respective proteins are shown in brackets. Groups of similar sequences are labeled. pLB1 is shown in boldface. The scale bar indicates 0.05 substitution per site.
Two CDSs, rsv and orf2, encode a putative resolvase (site-specific serine recombinase) and nuclease, respectively. Their predicted products show homology with ParA and ParB of plasmid RP4, respectively (49), and they are considered to be involved in multimer resolution (13, 21). (Note that ParA and ParB of RP4 are different from ParA, ParB1, and ParB2 of pLB1.) Therefore, the Rsv and Orf2 proteins might contribute to the stable maintenance of pLB1 in the host cells by the site-specific resolution and subsequent decatenation of the multimer form of pLB1. The putative products of orf14 and orf15 showed high similarity (82 to 85% identity) to antitoxin and toxin, respectively, of the classical proteic killer system for plasmid maintenance (26), and they may be involved in the genetic addiction system of pLB1.
(iv) Genes for conjugative transfer.Putative genes for conjugative DNA transfer, constituting the DNA transfer and replication (Dtr) system and mating pair formation (Mpf) system, were found in a 15-kb region of pLB1 (Fig. 2). The gene organization suggested that this region consists of two transcriptional units (mobC-virD2 and orf21-orf18) (Fig. 5). The putative origin of transfer (oriT; bp 49254 to 49443) was deduced on the basis of typical features of other previously well-characterized oriT sequence (65). This region, which had a relatively low G+C content (52.6%), contained inverted repeats (bp 49302 to 49327 and 49386 to 49425) and the putative nic site (5′-TATCCCGC-3′), which showed significant similarity to the RP4-type site (49).
Genetic organization of the putative transfer region of pLB1 and comparison with related systems. Genes encoding similar functions are displayed in the same color. Circles indicate the putative oriT regions of the respective plasmids. The numbers below the genes indicate the percent amino acid identity (>20%) to the corresponding gene product from pLB1. The lengths of TraO homologues are various, because only their C-terminal regions are conserved, which contain a putative functional domain for primase. The GenBank accession numbers of the respective nucleotide sequences are as follows: broad-host-range cryptic plasmid pIPO2, NC_003213; mercury resistance plasmid pSB102, NC_003122; A. tumefaciens strain C58 plasmid pTi, NC_003065; catabolic plasmid pWW0, NC_003350; and antibiotic resistance plasmid R388, BR000038.
We believe that the four putative gene products (MobC, VirD2, Orf18, and VirD4) play a role in the Dtr system. MobC contained at its C terminus the MobC domain (PF05713) that is believed to act as a molecular wedge for the relaxosome-induced melting of oriT DNA (66), and thus MobC may facilitate relaxosome function. VirD2 has been predicted to serve as a relaxase capable of site-specific nicking of plasmid DNA at oriT, because it contains each of the three defined motifs that are conserved in the relaxases of IncP-1 and Ti plasmids (50). As is the case with TraC of IncP-1 plasmids (49), the orf18 product is assumed to be transported to the recipient cells during the conjugative transfer of single-stranded plasmid DNA and is thought to subsequently initiate the synthesis of the complementary strand. The presence of the nucleotide-binding Walker motifs A and B in VirD4 allowed us to categorize this protein as being isofunctional with TrwB from an IncW plasmid, TraG from IncP-1 plasmids, and VirD4 from the Ti plasmid, all of which are necessary for conjugal DNA transfer.
The products of the 10 pLB1-specified genes (virB1 to virB6 and virB8 to virB11) putatively involved in the Mpf system showed 28% (VirB6) to 62% (VirB11) identity with the products of the virB gene clusters of symbiotic plasmids from rhizobial strains (Table 2). The level of identity was relatively low, but the functional domains of the putative products were conserved, suggesting that they are functional as the Mpf system. The gene order of the virB cluster on pLB1 was fundamentally the same as that in the Mpf gene clusters in various conjugative plasmids (Fig. 5) (55). The position of the Dtr-related gene cluster relative to that of the Mpf-related cluster on pLB1 was furthermore conserved in pIPO2 and pSB102, although several CDSs (such as traB, traP, and traQ) are located within the Mpf-related gene clusters on the latter two plasmids. This indicates that the putative transfer region of pLB1 is more compact than those of other plasmids. The VirB7 protein of Ti plasmid and its homologues in other Mpf systems of various plasmids encode outer-membrane-anchored lipoproteins to stabilize VirB9 and its homologues (18, 29). The virB7 homologues are flanked by virB6 and <