Genetic and Functional Analyses of the mob Operon on Conjugative Transposon CTn341 from Bacteroides spp.

Bacterial strains and growth.The relevant bacterial strains and plasmids used in this study are described in Table 1. For many of the studies described in this report, the pFD699 model system was used (1, 35). Plasmid pFD699 is a chimera of an E. coli plasmid and CTn341, and it is a fully functional conjugative transposon in Bacteroides strains but a stable plasmid in E. coli. The complete sequence of CTn341 was derived from work with pFD699 and has the GenBank accession number AY515263.1. Bacteroides strains were grown in BHIS medium (per liter: 37 g Bacto brain heart infusion broth, 1.0 g cysteine, 5 mg hemin, 2.0 g NaHCO₃) in an anaerobic chamber (32), and Escherichia coli strains were grown aerobically in L broth (1% tryptone, 0.5% NaCl, 0.5% yeast extract). The following antibiotics were used for selection in E. coli: ampicillin (Ap, 100 μg/ml), spectinomycin (Sp, 50 μg/ml), nalidixic acid (50 μg/ml), erythromycin (Em, 250 μg/ml), kanamycin (35 μg/ml), and chloramphenicol (25 μg/ml). The antibiotics used for selection of Bacteroides were rifampin (Rf, 20 μg/ml), tetracycline (Tc, 5 μg/ml), erythromycin (Em, 10 μg/ml), gentamicin (50 μg/ml), and fusidic acid (Fu, 25 μg/ml).

Bacterial matings.Standard filter-mating protocols were used to mobilize plasmids between strains (16). To mobilize plasmids from E. coli donors into Bacteroides recipients, the mating plates were incubated aerobically and the selection plates anaerobically. To mobilize plasmids from Bacteroides to E. coli (“mating-out” assays), the mating plates were incubated anaerobically and the selection plates aerobically. Multiple attempts were made to demonstrate activity of the CTn341 Mob region in E. coli-based systems, but all were unsuccessful. Thus, for many of the assays designed to measure the transfer frequency of mutant derivatives of CTn341, we used pFD699 (Ap^r in E. coli) mutants in mating-out assays which measured transfer to E. coli recipients. In the mating-out assays, Bacteroides donors were induced with 1 μg of tetracycline per ml and E. coli transconjugants were selected aerobically on medium containing Ap (1). The transfer frequency was calculated as the mean of the number of transconjugants per input donor cell in three independent experiments.

Plasmid construction and DNA manipulations.In order to test the mobilization activities of the CTn341 oriT and mob genes, a mobilization-deficient plasmid was constructed. This plasmid, designated pFD1112, was formed by deletion of the Bacteroides mob region and ermF gene from the vector pFD340 (36) and insertion of the ermA-ermF resistance cassette for selection in both E. coli and Bacteroides (15). PCR-amplified sequences from CTn341 were cloned into the PstI and BamHI sites of pFD1112 to assess transfer as described below.

The entire mob region and the mobAB genes were cloned into shuttle vector pFD351 for complementation experiments. PCR was used to amplify fragments (Fig. 1; also see Table S1 in the supplemental material) with restriction sites for PstI and BamHI, and these were then directionally cloned into pFD351. A plasmid containing a truncated derivative of MobC₃₄₁ [pMobCΔN] (Fig. 1) was constructed by directional cloning of the deletion fragment into the BamHI and SstI sites of the Bacteroides expression vector pFD340. The deletion fragment containing the mobC₃₄₁ ribosome binding site and lacking the bases for codons 2 to 145 was constructed by PCR using the primers described in Table S1 in the supplemental material [primers mobC)TM-N and mobC)TM-C].

• Open in new tab
• Download powerpoint

FIG. 1.

Genetic map of the mob operon and RT-PCR evidence showing linkage between genes. (A) Map showing mob genes and functional domains of the Mob proteins. The regions used for the construction of complementing plasmids are shown below the map. The functional domains of the Mob proteins are shown above the map. The red bar above MobB indicates the location of the relaxase/mobilization nuclease domain (Pfam accession number 03432), and the red bar above MobC shows the TraG/TraD superfamily domain (Pfam accession number 02534). The Walker boxes are shown by the inverted triangles above the red bar, and the five MobC transmembrane domains are indicated by black diamonds above the map. (B) Design and results of RT-PCR experiment, showing the locations of the primers used and the agarose gel of the amplified products. The cDNA (shown in red) was generated by reverse transcription of the mob mRNA using primers located toward the end of the message in the mobC gene. The arrowheads to the right of the gel indicate the sizes of the expected 621- and 400-bp products. Lane A, 1-kb-ladder molecular-size marker; B, cDNA-1 plus primer set mobA/B; C, cDNA-2 plus primer set mobA/B; D, pFD699 DNA-positive control with primer set mobA/B; E, cDNA-1 plus primer set mobB/C; F, pFD699 DNA-positive control with primer set mobB/C. All primers are listed in Table S1 in the supplemental material.

Site-directed mutagenesis (SDM) of the suspected oriT sequences was performed using a QuikChange II site-directed mutagenesis kit (Stratagene) and mutagenic primers (see Table S1 in the supplemental material for the SDM primers). The region to be mutated was cloned into pUC19, mutated, and then subcloned into pFD1112 for analysis. All mutations were verified by nucleotide sequence analysis.

RNA isolation and cDNA synthesis.Bacteroides strains were grown to mid-log phase (A₅₅₀ of 0.3 to 0.5) in BHIS with or without tetracycline. Chloramphenicol was added to a final concentration of 100 μg/ml, and total cellular RNA was extracted using the hot acid phenol method previously described (28). Fifty micrograms of RNA was precipitated with 100% ethanol, treated with Turbo DNA-free DNase (Ambion) according to the manufacturer's instructions, and then checked for contaminating DNA by PCR. The RNA concentration was calculated by the absorption at A₂₆₀/A₂₈₀. cDNA was synthesized with a Superscript first-strand synthesis system (Invitrogen, Carlsbad, CA) using either gene-specific primers for reverse transcription (RT)-PCR (Table S1 in the supplemental material lists the RT-PCR primers) or 200 ng/μl random hexamers for quantitative RT-PCR.

Quantitative PCR.Quantitative RT-PCR was performed essentially as described previously, using a Bio-Rad iCycler with the real-time PCR detection system (Bio-Rad, Hercules, CA) (41). The PCR products were approximately 125 bp in size (see the quantitative RT-PCR primers in Table S1 in the supplemental material) and were verified by agarose gel electrophoresis and by melting point analysis according to the Bio-Rad iCycler software. The reaction mixture contained 12.5 μl 2× iQ SYBR green Supermix, 1.5 μl of 5 μM forward primer, 1.5 μl of 5 μM reverse primer, 8.5 μl H2O, and 1 μl cDNA template (diluted 1/10) per well. All samples were run in triplicate, and RNA without reverse transcriptase was run as a control for DNA contamination. Relative expression values were calculated using the method of Pfaffl (26). Fold induction was determined by comparing the results for uninduced cells to those for cells induced with 1 μg/μl tetracycline, using the sigma-54 modulation protein gene as the reference gene. The sigma-54 modulation protein gene was selected because its expression is not changed in the presence of tetracycline and its expression level is similar to that of the genes being examined.

Northern hybridization.Northern blotting was performed as described previously (28), using RNA isolated from uninduced strains and from those induced with 1 μg/ml tetracycline. The gene-specific probes were generated by PCR amplification of strain BTSpFD699 genomic DNA (see Table S1 in the supplemental material for the primers used) and labeled with [³²P]dCTP by using the Prime-a-Gene labeling system (Promega, Madison, WI). The resulting blots were exposed to a phosphorimaging cassette and visualized using a Typhoon 9410 (Amersham, Piscataway, NJ).

The mob genes are essential for CTn341 transfer.In a previous study, the 18-kb locus that mediates conjugative transfer of CTn341 was described (1). Mutational analysis of the locus indicated that most of the genes tested were required for the transfer of both CTn341 and coresident plasmids or the mobilizable transposon Tn4555. However, null mutations in two genes, mobA₃₄₁ and mobB₃₄₁, resulted in the loss of CTn341 but not of the coresident elements. As shown in Fig. 1, these genes are part of an apparent mob operon that is divergently transcribed from the tra region, and database searches suggested that these genes code for proteins involved in DTR reactions. Multiple sequence alignments of the MobA₃₄₁ protein (GenBank accession number AAS83498) revealed that this protein is highly conserved in nearly all of the Bacteroidetes phylum, including all Bacteroides species, and in many Flavobacteria, Porphyromonas, and Prevotella species. In most cases, the MobA₃₄₁ homologs are adjacent to putative relaxase proteins, such as MobB₃₄₁. MobA₃₄₁ has similarity to a newly described family of nicking accessory proteins referred to as the ribbon-helix-helix family which are found in conjugative systems in diverse bacteria (42) (see Fig. S1 in the supplemental material). These proteins have a role in helping the relaxase gain access to the nick site during DNA-processing reactions.

MobB₃₄₁ (GenBank sequence accession number AAS83499) was also highly conserved in Bacteroides species, usually with multiple homologs per genome. These homologs generally ranged between 40% and 99% identity and were usually associated with a MobA₃₄₁-like partner. MobB₃₄₁ has similarity to relaxase proteins and contains several conserved regions (Fig. 1; also see Fig. S2 in the supplemental material) (23). To verify a role for MobB₃₄₁, we tested the mobB₃₄₁ mutant Jo36 (1) and two complemented strains for transfer in mating-out assays. The data from these experiments, shown in Table 2 (assays 1 to 4), indicate that mobB₃₄₁ is essential for CTn341 transfer and that this defect can be complemented in trans when mobB₃₄₁ is present on a plasmid.

The third mob gene, mobC₃₄₁, codes for a protein (MobC₃₄₁, GenBank sequence accession number AAS83500) with significant homology to the VirD4 family of coupling proteins which mediate interactions between the relaxosome and the secretion pore. Multiple sequence alignments showed that the conserved Walker A and B motifs were present, as were predicted transmembrane domains (Fig. 1; also see Fig. S3 in the supplemental material). Previously and in the experiments whose results are shown in Table 2, we found that null mutations in mobC₃₄₁ were still able to mediate the transfer of CTn341 (1). This was unexpected, since the coupling proteins are essential in other T4SS conjugation systems (10). We reasoned that the lack of a phenotype might be due to complementation of the mutated mobC₃₄₁ by one or more of the MobC homologs in the chromosome. Four MobC homologs with greater-than-70% amino acid identity to MobC₃₄₁ are present in each of the commonly used Bacteroides strains (see Table S2 in the supplemental material). In order to determine whether MobC₃₄₁ is important for transfer, an N-terminally truncated MobC₃₄₁ protein lacking the first four membrane-spanning regions was constructed by the deletion of amino acid codons 2 to 145 (Fig. 1). This mutated protein's gene, cloned into an expression shuttle vector, pMobCΔN, lacked the regions needed to insert into the membrane but maintained the relaxase interaction motif. It was predicted that the MobB₃₄₁ relaxase would bind to the mutated MobC₃₄₁ in the cytoplasm but, since it could not associate with the T4SS in the membrane, CTn341 would not transfer. When the plasmid was tested in a CTn341 strain that was wild type with respect to mobC₃₄₁, there was about a 10-fold decrease in transfer frequency relative to the transfer frequency in the control (Table 2). This compared to about a 100-fold decrease for the strain with a mobC₃₄₁ null mutation. When pMobCΔN was expressed in the CTn341 mobC₃₄₁ mutant, transfer was nearly abolished and approached the background level (Table 2). These results suggest that MobC₃₄₁ preferentially interacts with the CTn341 conjugation apparatus but that MobC homologs can compensate somewhat for a loss in functionality.

The mob operon is regulated by tetracycline induction.The genetic organization of the mob genes suggested that they would be transcribed as an operon. This linkage was first examined by RT-PCR, as shown in Fig. 1B. The cDNA for these reactions was synthesized from Bacteroides fragilis strain IB255, containing CTn341 and induced with tetracycline, using primers located in the mobC coding region or at the end of the gene (primers cDNA-1 and cDNA-2, in Table S1 in the supplemental material). Then, using this cDNA as the template, specific primers were used to amplify the intergenic regions between mobA and mobB and between mobB and mobC as shown in Fig. 1B (primer sets mobA/B and mobB/C, in Table S1 in the supplemental material). In all cases, the intergenic regions were amplified and the size of the amplicons corresponded to the predicted sizes of 400 or 621 bp. In no case were there amplification products in control reaction mixtures lacking reverse transcriptase (data not shown).

Northern hybridization analyses were then used to further examine mob regulation. The transfer of CTn341 is known to be controlled by tetracycline; thus, RNA from tetracycline-induced and noninduced cultures was compared in a Northern blot analysis in order to determine if the mob genes were controlled at the transcriptional level (Fig. 2 A). The results readily showed significant induction of mob-specific mRNA in the presence of tetracycline, but much of the RNA was degraded and it was difficult to visualize the 4-kb mob mRNA. In order to optimize the levels of the intact 4-kb message, the mob operon was cloned onto a plasmid (Fig. 1, pMob). When this plasmid was introduced into a strain containing CTn341, a more distinct 4-kb mRNA was observed in tetracycline-induced cultures (Fig. 2B).

• Open in new tab
• Download powerpoint

FIG. 2.

Northern blot analysis of mob operon region. Fifty micrograms of RNA was separated on formaldehyde agarose gels, transferred to nylon membranes, and hybridized to ³²P-labeled probes from the mob operon as indicated. (A) RNA samples from B. fragilis cells containing CTn341 were either induced (+Tc) or not induced (−Tc) and probed with a 4-kb probe containing the entire mob operon. (B) RNA sample from tetracycline-induced B. fragilis containing CTn341 plus a multicopy plasmid with the cloned mob operon. Lanes are labeled to indicate induction where appropriate, and the full-length 4-kb mRNAs are shown by the arrows to the right of the blots. The apparent bands (*) at about 1.5 and 2.5 kb are a commonly observed compression artifact caused by the 16S and 23S rRNA. These artifacts are exacerbated by the extensive RNA degradation.

To confirm tetracycline control of the operon, quantitative RT-PCR analyses were performed with primers for mobA₃₄₁ and mobB₃₄₁ and for traA as a positive control. Previous studies using transcriptional fusions showed an increase in traA transcription under tetracycline induction conditions (5). The results of the quantitative RT-PCR analyses showed more than a 100-fold-increase in the mRNA levels of mobA₃₄₁, mobB₃₄₁, and traA under tetracycline induction conditions (Fig. 3). Tetracycline induction of the tra genes is controlled by the two-component regulatory system RteA/RteB (45). Quantitative RT-PCR was performed using rteA and rteB mutants to determine if they played a role in mob regulation. The results were very similar for both mutants and showed that the majority of tetracycline induction was mediated by RteAB, although there was still some weak induction in the mutants (Fig. 3). Interestingly, traA was more tightly controlled by RteAB.

• Open in new tab
• Download powerpoint

FIG. 3.

Quantitative RT-PCR analysis of the mobA, mobB, and traA genes in rteA and rteB mutants. Overnight cultures of the wild-type (BTSpFD699), rteA (BTSJo68), and rteB (BTSJo69) strains were subcultured and grown to mid-log phase either with or without 1 μg/μl tetracycline (1). RNA was isolated from the cultures, and two independent sets of RNA were used for analysis, using primers for the mobA, mobB, or traA gene as indicated. Each PCR measurement was performed in triplicate. The sigma-54 modulation protein gene was used as the reference gene. The results are expressed as the fold induction relative to the results with no tetracycline. Error bars represent standard deviations.

The CTn341 oriT and nick site are located upstream from the mob operon.The transfer origin, oriT, of many conjugative elements is located adjacent to the transfer genes (14). It was hypothesized that the oriT of CTn341 would be located in the intergenic region between the tra genes and the mob operon. To test this, we inserted three DNA fragments that spanned the region from traA to mobC into a mobilization-deficient shuttle plasmid, pFD1112, and then measured its transfer in mating-out assays with E. coli DH10B as the recipient. There was no transfer of the pFD1112 control, but a construct that contained the region between traA and mobA₃₄₁ was able to mediate transfer in the mating-out assays (Fig. 4). This region was then further subcloned into a series of overlapping fragments, and transfer was observed in all but one of these constructs. It was determined that these fragments all shared a common 120 bp directly upstream of mobA₃₄₁. In order to confirm the oriT location, a 270-bp fragment that encompassed the 120-bp region was tested in pFD1112. The results showed that this plasmid mediated transfer at a frequency similar to the transfer frequencies of all other constructs (Fig. 4). In fact, the transfer frequencies for all constructs were comparable to the transfer frequency of the parent fragment, suggesting that no cis-acting regions were lost during subcloning.

• Open in new tab
• Download powerpoint

FIG. 4.

Identification of the CTn341 oriT region. Three fragments from the mob operon region (shown by the numbered lines over the map) were PCR amplified, inserted into the mobilization-deficient plasmid pFD1112, and tested for the ability to mediate transfer of the plasmid in mating-out assays, as described in the text. Transfer frequencies were determined and are shown to the right. Based on these results, subfragments (shown below the map) were amplified and tested in the mating-out assay, and the results showed that there was a common 120-bp region just upstream of mobA (see parallel dashed vertical lines). Transfer frequencies are the number of transconjugants per input donor cell (means of the results of triplicate experiments).

In order to identify the presumed oriT nick site, a combination of bioinformatics and site-directed mutagenesis was used. Previously, consensus nick sites were determined for several Bacteroides-transmissible elements, and these were found to cleave between guanine and cytosine residues (34, 43). Using a pattern-matching homology search, we examined the 120-bp region for similarity to the consensus sequence. The results revealed six potential homologs with GC cleavage sites, and of these, four were located in proximity to the mobA gene (Fig. 5). Site-directed mutagenesis was used to change the GC to AT at four of the sites. Three of the four mutated sites mediated normal levels of transfer, but mutation of homolog 2 resulted in a complete loss of transfer, implying that it is the CTn341 nick site. In addition, this site was adjacent to a large imperfect inverted repeat which has the potential to fold into a stable structure (Fig. 5; also see Fig. S4 in the supplemental material). These secondary structures are typical of the relaxase cleavage sites, and none of the other putative cleavage sites possessed significant secondary structure.

• Open in new tab
• Download powerpoint

FIG. 5.

Identification of the CTn341 nick site. (A) Location of the oriT region between mobA and the divergently transcribed traA. DNA sequence showing four of the putative nick sequences (underlined and in upper case) that were mutated. The number 2 nick site shows the putative GC cleavage nick site with the arrowhead under the sequence, and the imperfect inverted repeat sequence is shown by the dashed arrows over the sequence. (B) Alignment of Bacteroides nick sites and generation of consensus. The Tn5520 nick site, Tn4555 nick site, and nick site homologues in the CTn341 oriT region are shown. Site-directed mutagenesis of the GC of homologue 2 resulted in a loss of transfer, indicating that it is the nick site of CTn341. Mutations of the other sites did not result in loss of the transfer phenotype.