INTRODUCTION

Top Abstract Introduction Materials and Methods Results and Discussion References

The gram-negative bacterium Actinobacillus actinomycetemcomitans is a capnophilic coccobacillus. The organism has been associated with several forms of periodontal disease such as localized juvenile periodontitis and rapidly progressive periodontitis, as well as with soft tissue abscesses and endocarditis (58). In a previous study 39 isolates of this periodontal pathogen had been screened for the presence of indigenous plasmids in an effort to evaluate the role(s) of such genetic elements in oral bacteria (32). Three plasmids, pVT736-1 (2 kb), pVT736-2 (>30 kb), and pVT745 (25 kb), were identified in two strains, suggesting that the occurrence of plasmids in A. actinomycetemcomitans was rare. The ultimate goal was to determine the biological properties of these plasmids, to assess their potential contribution to the pathogenicity of A. actinomycetemcomitans, and to evaluate their usefulness as tools in recombinant DNA technology. Previous work has focused mainly on the characterization of pVT736-1, one of the first rolling circle replicating (RCR) plasmids isolated from gram-negative bacteria (17, 18). It was shown that pVT736-1 was cryptic, that it was not related to RCR plasmids found in gram-positive bacteria, and that it encoded a new type of partitioning system (20).

Preliminary characterization suggested that there was no obvious phenotype associated with pVT745 (41). Its size was a strong indication that the plasmid replicated by a theta mechanism rather than by a rolling circle mode. Although pVT745 had been isolated from one strain of A. actinomycetemcomitans (VT745) only, it was demonstrated by Southern hybridization that this plasmid shared sequence homologies with chromosomal DNA from numerous A. actinomycetemcomitans isolates (39, 40). However, plasmid DNA did not hybridize with the genome of the strain from which it was isolated. It was suggested that the plasmid might have integrated into the chromosome of these strains. This was of particular interest since there is no evidence for the occurrence or frequency of natural genetic exchange among gram-negative bacteria found in the oral cavity. Integration of a plasmid into the A. actinomycetemcomitans chromosome may promote the transfer of chromosomal genes during conjugation or suggest the presence of one or more insertion elements or transposons. The goal of the current study was to obtain and analyze the nucleotide sequence of pVT745 in an effort to characterize plasmid-encoded functions. This would facilitate a determination of which genes pVT745 and the different A. actinomycetemcomitans chromosomes were sharing and whether such genes and/or their products could contribute to the organism's virulence. Sequence analysis of pVT745 revealed the presence of a cluster of genes encoding products homologous to proteins identified in type IV secretion systems (for recent review, see reference 11). Such transport systems are widespread and highly versatile since they can export protein and/or DNA/protein complexes. Their presence on plasmids from gram-negative organisms has been associated with conjugative transfer. Intra- and interspecies conjugative transfer of pVT745 derivatives was demonstrated.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results and Discussion References

Bacterial strains, plasmids, and growth conditions. The strains and plasmids used in this study are listed in Table 1. The A. actinomycetemcomitans recipient strain ATCC 29522Rif was isolated as a spontaneous rifampin-resistant mutant of strain ATCC 29522. A. actinomycetemcomitans was grown in TSBYE (3% Trypticase soy broth, 0.6% yeast extract) at 37°C in 10% CO₂. Escherichia coli JM109 was grown in YT medium (37). Where appropriate, antimicrobial agents were used at the following concentrations: ampicillin, 50 µg/ml; kanamycin, 50 µg/ml (except for strain ATCC 700685 with 100 µg/ml); rifampin, 100 µg/ml; and spectinomycin, 100 µg/ml.

DNA preparations and recombinant DNA techniques. Plasmid DNA was isolated from A. actinomycetemcomitans and E. coli as described previously (20). DNA templates used in sequencing were purified by CsCl buoyant density centrifugation (47) or by use of the Wizard Plus Midiprep kit (Promega). Restriction endonucleases and T4 DNA ligase were used in accordance with the manufacturer's instructions. Standard recombinant DNA techniques were performed as described by Sambrook et al. (47). DNA-DNA hybridization conditions and transformation by electroporation were as described previously (20).

DNA sequencing. Three previously described clones of pVT745, pKN1, pKN2, and pKN3, were used to determine the nucleotide sequence of the plasmid (Table 1) (39). Smaller fragments of these clones, labeled A1 through F, were obtained by restriction enzyme digestion at the sites shown in Fig. 1. These fragments were then subcloned into either pUC19 (high copy) (57) or pGB2 (low copy) (12). The ends of the inserts were sequenced with M13 standard and reverse primers for pUC clones or custom-made primers flanking the multiple cloning site of pGB2. Automated sequencing was performed with fluorescent terminators by cycle sequencing with an Applied Biosystems model 373 DNA sequencer (DNA facility at the University of Texas Health Science Center, San Antonio). Generated sequences were used to design synthetic primers. Sequences were read on a Beckman Gelmate and entered directly in the computer. Additional sequencing was performed with pVT745 using the Omnibase DNA Cycle Sequencing System (Promega) to resolve ambiguities, to close gaps, and to cross restriction endonuclease sites.

View larger version (22K): [in this window] [in a new window]   FIG. 1.   Physical map of pVT745. Only restriction sites relevant for subcloning fragments A1 to F are depicted. Several of these sites are present more than once on the plasmid. The unique ScaI site (underlined) was the target site for the insertion of a kanamycin gene (Fig. 2.). The location and transcriptional orientation of the magA and magB gene clusters implicated in conjugation and the putative origins of replication (oriV) and transfer (oriT) are shown.

Annotation. MacDNASIS from Hitachi Software was used to analyze and assemble sequences and to determine the presence of putative genes. Open reading frames (ORFs) encoding at least 50 amino acids and displaying a translational start codon, as well as a potential E. coli Shine-Dalgarno consensus sequence (49), were identified. Smaller ORFs were listed only if they or their putative products showed significant similarities to known genes and/or proteins or if they appeared to be transcriptionally linked to adjacent ORFs. Generally, in the absence of experimental data the start codon farthest upstream was used to annotate the ORF start site.

Construction of recombinant pVT745 derivatives. A selective marker, kan, was inserted into pVT745 at a unique ScaI site located at the 3' end of gene AA05 (23 bp from the translational stop codon) via allelic exchange by homologous recombination. The different steps involved in the construction of the pVT745-derivatives are outlined in Fig. 2A. All recombinant constructs were obtained in E. coli JM109. Vector pGB2 (12) containing fragment B (B1+B2) from pVT745 (Fig. 1) was digested with ScaI. A 1.5-kb ClaI-fragment carrying kan from plasmid pJH1 (31) was blunt ended with Klenow polymerase I (Gibco-BRL) and ligated into the single ScaI site. A segment harboring the pGB2-specific marker, spc, was then deleted from this construct by double digestion with SphI/HincII, treatment with Klenow polymerase I, and religation of the free ends of the replicon. This last construct, pGB2R, was then used to transform A. actinomycetemcomitans VT745 by electroporation. Since pGB2 does not replicate in this host, only transformants that had the kan gene integrated into the resident plasmid, pVT745, by homologous recombination were able to grow in the presence of kanamycin. The pGB2 construct allowed for a single or a double crossover event to occur. Plasmid DNA isolated from the transformants was analyzed by restriction enzyme digestion and Southern blot hybridizations using pGB2 and the kanamycin resistance gene as probes. Results from these experiments confirmed that both a single crossover event and a double crossover event had occurred. The resulting pVT745 derivatives pDMG20 (single crossover) and pDMG21 (double crossover) were subsequently used in conjugation and mobilization assays (Fig. 2B).

View larger version (18K): [in this window] [in a new window]   FIG. 2.   Construction of recombinant pVT745 derivatives via allelic exchange. (A) A fragment of pVT745 was cloned into pGB2 and a kanamycin resistance gene inserted into the unique ScaI site of this construct for allelic recombination with pVT745. A detailed description of the construction is provided in Material and Methods. (B) Restriction endonuclease maps of the two types of recombinants, pDMG20 and pDMG21. The physical maps of key restriction endonuclease sites are shown along with the size of EcoRI fragments that can be derived from each recombinant type at the site of recombination. DNA segments are not drawn to scale. kan, kanamycin resistance gene; spc, spectinomycin resistance gene.

Mating experiments. Conjugative matings were performed between A. actinomycetemcomitans strains using JP2::pDMG20 and JP2::pDMG21 as donor strains and ATCC 29522Rif, ATCC 29522::pDMG3, and ATCC 700685::pDMG3 as recipient cells. In addition, E. coli JM109 served as an interspecies recipient strain. Donors and recipients were grown to mid-exponential phase in broth cultures and mixed in a total volume of 1 ml at a ratio of 1:1 (the recipient was A. actinomycetemcomitans) and 10:1 (the recipient was E. coli). The latter ratio was different to compensate for the much faster growth rate of E. coli. The mixture was then centrifuged, and the cells were washed and spotted onto a TSBYE agar plate. After incubation for 4 h (E. coli recipients) or 6 h (A. actinomycetemcomitans recipients) at 37°C in 10% CO₂, the cells were scraped off the plate and resuspended in 1 ml of TSBYE. Aliquots of serial dilutions of the suspension were then spread onto TSBYE to determine the number of A. actinomycetemcomitans transconjugants and the number of donor cells and on YT plates to determine the number of E. coli transconjugants. All plates contained the appropriate antibiotics. Incubation of plates was in 10% CO₂, except for YT plates. Transfer frequencies were expressed as the number of transconjugants per donor cell. Selected transconjugants were examined for the presence of plasmid DNA.

Nucleotide sequence accession number. The complete sequence of pVT745 from A. actinomycetemcomitans VT745 has been deposited in the GenBank database and assigned accession number AF302424.

	RESULTS AND DISCUSSION

Top Abstract Introduction Materials and Methods Results and Discussion References

General description. The entire sequence of pVT745 was determined to be 25,420 bp. Annotation of the derived sequences identified 36 ORFs likely to represent functional translated genes. A total of 12 of these ORFs were transcribed in a clockwise orientation, while the remaining 24 ORFs were transcribed counterclockwise. The positions and transcriptional orientations of all ORFs are depicted in Fig. 3. Table 2 lists the putative functions, the characteristics, and the closest relatives for the predicted product of each ORF. Seventeen ORFs and their products had no detectable homologs in the databases. However, some of these genes could be associated with conjugation and partitioning due to the fact that they appeared to be transcriptionally linked to genes with known functions (Table 2). In fact, many of the ORFs identified on pVT745 either overlapped or were separated by only a few nucleotides, indicating that they may be part of operons. Such potential operons are represented by genes magA1 and magA2, ssb to magB14, AA16 to AA14, and AA02 to AA06.

View larger version (29K): [in this window] [in a new window]   FIG. 3.   Genetic organization of pVT745. Genes are represented by boxes. Open boxes indicate that the corresponding ORFs are transcribed clockwise; hatched boxes indicate that the ORFs are transcribed counterclockwise. Kilobase coordinates are shown, as are the positions of the two PstI and the single BamHI and ScaI sites. Genes of unknown function are labeled AA (for A. actinomycetemcomitans) followed by a number. Other gene designations are associated with potential functions of corresponding gene products as listed in Table 2.

Conjugation-like ORFs. GenBank analysis revealed that the predicted proteins encoded within a DNA segment of 12 kb showed homology to the type IV family of secretion systems (10). Corresponding systems found on plasmids of gram-negative bacteria are associated with DNA transfer such as in conjugation events, whereas their presence in the chromosomes of bacterial pathogens such as Bordetella pertussis and Helicobacter pylori has been implicated in the export of proteinaceous virulence factors (for a review, see reference 11). BLASTX searches revealed that the pVT745-specific proteins showed high levels of similarity with both chromosomally encoded type IV transport proteins and traditional plasmid transfer systems (Table 2). Based on these homologies it was reasonable to assume that pVT745 was self-transmissible, and therefore genes implicated in conjugation were designated mag (mating-associated genes). As shown for other conjugative plasmids transfer functions seem to be clustered in two regions. The first region is responsible for DNA processing functions and, except for Legionella pneumophila (48), is not found in chromosomally encoded type IV secretion systems. The other region is most probably associated with mating pore formation, pilus assembly, and entry exclusion. To separate these two functions within the mag genes, the letter "A" was added to genes whose products showed homologies to proteins required for conjugal DNA processing, and the letter "B" was added to genes encoding products involved in mating aggregate formation and entry exclusion. As with genes found in homologous systems, either the ribosome-binding site or the initiation codon of most ORFs located in magA and magB overlapped with the end of the previous ORF, suggesting that these genes are transcriptionally linked. Most of the genes clustered in magB seem to be associated with the formation of a multicomponent pore or channel which spans both bacterial membranes and is used to transport DNA (11). Although, the location and organization of genes in magA and magB resembles those for other conjugative plasmids and type IV secretion systems, the location of magB12 (encoding the VirD4 homolog), magB13 (encoding a lipoprotein), and magB14 (encoding a TrbM homolog) at the end of magB is rather unusual. Gene trbM is only found in the IncP-specific transfer operon and its role in conjugation, if any, is unknown (45). Also, the presence of the additional lipoprotein encoded by magB13 has been reported for the type IV secretion system of Brucella suis only (43).

Conjugative transfer functions. The presence of ORFs homologous to genes implicated in conjugation suggested that pVT745 is able to mediate its own conjugative transfer and to support the mobilization of non-self-transmissible, coresident plasmids. Mobilizable plasmids, such as the RSF1010 derivative, pMMB67 (16), only carry genes necessary for DNA processing and an oriT site corresponding to the plasmid-encoded nickase. They lack the genes required for mating pore formation. Since pVT745 did not carry any known selective marker which would allow the study of conjugation transfer functions, a kanamycin resistance gene derived from pJH1 (31) was inserted into the plasmid by homologous recombination as described in Materials and Methods. The integration of the kan gene via single and double crossover events was verified by EcoRI restriction enzyme analysis and Southern blot hybridization (not shown). Two types of recombinants were generated which resulted in construct pDMG20 and construct pDMG21.

7

8

Replication and partition functions. The putative origin of replication was located downstream of magB within a ca. 0.63-kb noncoding region (Fig. 1). This region showed 88 to 91% identity to the origin of replication of Haemophilus ducreyi plasmid, pLS88 (4.8 kb) (14). Similar regions have been described for Pasteurella multocida plasmid pIG1 (5.4 kb) (56), and two plasmids, pYFC1 and pAB2, were isolated from Pasteurella haemolytica (9, 55). These data indicate that there is an evolutionary link between multiple plasmids in the HAP group. Replication of pLS88 seems to be independent of any plasmid-encoded protein (14). Nucleotide sequence analysis performed by Wright et al. (56) revealed that the replication regions contained two to three major inverted repeats (IR20, IR16, and IR38). Two of these repeats can be found in the putative oriV of pVT745. However, a region of approximately 0.5 kb present in all of the above-mentioned plasmids is missing on pVT745. This deletion might have an effect on the host range of pVT745. Whereas pAB2 and pIG1 replicated in numerous gram-negative organisms, including E. coli (55, 56), conjugation experiments with the pVT745 derivatives indicated that pVT745 could be transferred to E. coli but was unable to support its replication in the absence of a specific E. coli replicon (see above). This was confirmed by transformation of JM109 with pDMG20 and pDMG21 via electroporation. E. coli transformants could only be obtained with pDMG20 which carried the pGB2 replicon. However, as already observed in the mating experiments, the plasmid was structurally unstable. All transformants carried a 7.8-kb plasmid only, which was identical to pGB2R. Therefore, it can be concluded that the pVT745-specific oriV is not functional in E. coli and that genes in the conjugative transfer region cannot be maintained in JM109. The reason for the latter is unknown. However, it is of interest to note that plasmids pAB2 and pIG1 are mobilizable (55, 56) but not conjugative. The other plasmids, pLS88 and pYFC1, are neither conjugative nor mobilizable.

DNA sequences shared by pVT745 and A. actinomycetemcomitans genomes. It was shown previously that the genomes of numerous strains of A. actinomycetemcomitans contain regions with homology to pVT745 (39). Southern hybridization studies with three different pVT745-specific fragments allowed for the identification of five strain-dependent groups, A to E, based on hybridization patterns (39). Additional hybridization studies were performed with smaller fragments ranging in size from 1.3 to 7 kb (40). A comparison of the hybridizing fragments with the pVT745 sequence in hand showed that some of these regions of homology were associated with genes located in the magB cluster for groups A, D, and E. However, none of the strains seemed to contain a complete magB operon. In addition, none of the five groups exhibited similarities to any of the genes found in magA. Strains with hybridization patterns C and D showed a high degree of homology with a probe containing the pVT745-specific oriV. The other three groups also hybridized with this probe. However, the signal for strain 725, representing pattern E, was very weak (40). The last region of homology was associated with ORFs homologous to genes found in Yersinia (AA09 and AA10) and Neisseria (AA12) spp. (Table 2). Strain VT747, the only representative of group D, hybridized to DNA fragments carrying AA09 to AA12. Representatives of groups A and C showed some similarities to the DNA segment harboring AA10 to AA12. The presence of remnants of the pVT745-specific oriV and conjugative system suggests that this plasmid, or a related vector, once inserted into the chromosome of various A. actinomycetemcomitans strains with subsequent strain-specific loss of the majority of plasmid-encoded genes. However, the sequences shared could also have been inherited from different donor organisms at different times, instead of being the result of a single event that occurred in the distant past.

Conclusions. Plasmid pVT745 is a true composite with blocks of genes which seem to have been acquired from a variety of bacterial sources, such as Neisseria spp., Serratia spp., and the HAP family. We failed to identify genes with similarities to putative virulence factors, or antibiotic resistance genes. However, phenotypes other than conjugative transfer might be associated with one or more of the ORFs of unknown function. Significant sequence similarities were found at the protein level to bacteria belonging to the , , and  subgroups of proteobacteria. This was particularly apparent with proteins encoded in the magB cluster, which is of a rather chimeric nature. Predicted proteins were homologous to chromosomally encoded components of type IV secretion systems present in Brucella abortus, Brucella suis, Helicobacter pylori, and Bordetella pertussis, all of which are associated with virulence, and to the highly similar plasmid-encoded proteins of gram-negative bacteria which were shown to be involved in conjugative DNA transfer (11). There is no simple single mechanism to explain this diversity in the magB gene cluster. O'Callaghan et al. (43) have suggested that the different protein secretion systems each evolved independently from the DNA transfer system. However, this assumption is not readily supported by the data presented in this report, since magB contains sets of genes with homologies to both DNA and protein secretion systems from a variety of plasmids and organisms. Since the arrangement of genes is very similar to those in other type IV secretion systems, it seems unlikely that these genes originated from different sources and were then arranged in the pattern found on pVT745. Such an explanation would also rely on extensive gene exchange between A. actinomycetemcomitans and other bacteria, yet there is currently no evidence for this. In addition, the pVT745 secretion system did not exhibit any significant similarities to the other systems at DNA level. It is possible that the pVT745-specific transfer system separated from the others early on and evolved independently. If this assumption is correct, the homologs detected in the chromosomes of various A. actinomycetemcomitans strains would indeed have been the result of an insertion of all or part of pVT745. This would raise the question as to the driving force(s) that caused the incorporation of pVT745 into bacterial chromosomes. No known insertion element or phages, or remnants thereof, were detected on pVT745, and one can only speculate if such elements were once present and then lost.