INTRODUCTION

Top Abstract Introduction Materials and Methods Results and Discussion References

Conjugative transposons are gene transfer elements that can move from the genomes of donor cells to those of recipients, sometimes across large phylogenetic distances. They are one of the major vectors responsible for the spread of antibiotic resistance among bacterial pathogens (for recent reviews, see references 32, 37, and 39). The most intensely studied of the conjugative transposons is Tn916. This 18.3-kb element was originally isolated from the chromosome of Enterococcus faecalis DS16, where it mediated tetracycline resistance via the tet(M) gene (11). The complete DNA sequence of this element has been obtained (10). Open reading frames have been identified that have the potential to encode polypeptides with sequence similarity to proteins known to be involved in conjugation (e.g., the antirestriction protein Ard of plasmid Collb-P9 and the MbeA mobilization protein of plasmid ColE1). A functional oriT site has also been located (19), and there is evidence that this site is involved in single-stranded DNA transfer to the recipient (35, 40).

Tn5397 is a 21-kb tetracycline resistance-encoding conjugative transposon originally found in Clostridium difficile (24, 25). Tn5397 was shown to be transferred by a conjugation-like process from C. difficile strain 630 to Bacillus subtilis strain CU2189 and back to C. difficile, and between C. difficile strains (24). Furthermore, Tn5397 has also been shown to be capable of transfer in a model oral biofilm community, indicating that the element is likely to be able to transfer in natural environments (33). Tn5397 is related to Tn916; the central regions that are involved in conjugation of these two elements are very similar (14, 24, 25). However, Tn5397 can be distinguished from Tn916 in at least two important characteristics: first, Tn5397 contains a group II intron inserted into a gene almost identical to orf14 from Tn916 (25), and second, the DNA sequences of the ends of Tn5397 are completely different from those of Tn916. Instead of having the int and xis genes that have been shown to be required for integration and excision of Tn916 (22, 30, 34, 43, 44), Tn5397 contains the gene tndX, which encodes a putative protein not related to Int or Xis (2, 3, 9, 20, 27, 28, 36) but belonging to the large resolvase subgroup of site-specific recombinases (46).

Members of the resolvase/invertase family, such as Tn3, resolvase, and Mu invertase Gin, show significant amino acid sequence similarity within their N-terminal regions (12, 16, 21, 41). They make staggered 2-bp cuts at a central dinucleotide within the crossover site, leaving recessed 5' ends. The hydroxyl group of a conserved serine residue located near the N terminus of these recombinases is the nucleophile responsible for the cleavage of the DNA backbone at the dinucleotide (17, 23, 31). The members of the large resolvase subgroup have the resolvase/invertase catalytic domain in their N-terminal regions but are much larger and more diverse in their C termini (8, 45). Their molecular masses are between 50.7 and 82 kDa, while those of the typical members of the resolvase/invertase family are only about 20 kDa. The large resolvases are found associated with a range of genetic elements; for example, CisA is required for excision of DNA in spore development (38), XisF is involved in heterocyst development (5), C31 integrase is required for integration and excision of bacteriophage genomes (45), and TnpX is required for excision of the Clostridium perfringens transposon Tn4451 (8). Analysis of the amino acid sequence of TndX from Tn5397 revealed that it contains a 61.5-kDa putative polypeptide and is most closely related to TnpX, with 37% identity and 61% similarity over the full length of the proteins (46).

Recently we have shown that an intact tndX gene is required for conjugative transposition of Tn5397 and for the production of a circular form of the element (46). This finding, together with DNA sequence analysis of the transposon-chromosome junctions of integrated Tn5397, target sites of Tn5397, and DNA sequence of the joint of the circular form, allowed us to propose the following model for insertion and excision of the element. Briefly, the element is excised from the donor genome, generating a circular form with a GA dinucleotide at the joint of the ends of the element. The original target site is completely regenerated after Tn5397 excision and contains a central GA dinucleotide. The element is then transferred to the recipient cell by conjugation. In the recipient, it recognizes a target sequence containing a GA dinucleotide and inserts, resulting in an integrated copy of Tn5397 flanked by GA dinucleotides. We hypothesized that TndX is responsible for catalyzing the excision and insertion reactions. To test this, the tndX gene was cloned and the protein was overexpressed in Escherichia coli. The functions of this protein were tested by conducting excision and insertion assays in E. coli using mini-transposons derived from Tn5397. The roles of the highly conserved residues at the N-terminal resolvase/invertase domain and the role of the C-terminal nonconserved region of this protein were investigated by mutagenesis studies. The target site choices of the minitransposon in E. coli were also analyzed.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results and Discussion References

Bacterial strains. Bacterial strains used are listed in Table 1. E. coli BLR(DE3) was used as a host for overexpression of wild-type and mutant proteins. E. coli XL-1 Blue was used as a host for transformations and plasmid preparations. DNA from C. difficile 630 was used as template for amplification of tndX. DNAs from C. difficile CD37 and B. subtilis CU2189 were used as templates for amplification of the Tn5397 targets.

Construction of plasmids. The plasmids used are listed in Table 1 and were constructed as follows. pET-TndX was constructed by cloning the tndX gene (46) into a pET19 vector (Novagen) via XhoI and BamHI sites. The tndX gene was amplified by PCR using P1 (5' GCCTCGAGTTGTTAAAACAGCAAGC 3') and P2 (5' GCGGATCCCTATCAATGAGACACTGC 3') as primers. This amplification does not include the tndX ribosome binding site, and the gene is under the control of the vector translational signals. Pairs of complementary primers were designed to generate single amino acid substitutions within TndX as follows: P3 (5' GTTGCATTATACTCTGCGCTTTCACGAGATGATG 3') and P4 (5' CATCATCTCGTGAAAGCGCAGAGTATAATGCAAC 3') for pET-TndX(R17-A), P5 (5' CATTATACTCTCGCCTTGCGCGAGATGATGGGTTG 3') and P6 (5' CAACCCATCATCTCGCGCAAGGCGAGAGTATAATG 3') for pET-TndX(S19-A), and P7 (5' GTCACGTCTAGGAGCGAACAATGCACTATTC 3') and P8 (5' GAATAGTGCATTGTTCGCTCCTAGACGTGAC 3') for pET-TndX(R93-A). The negative-strand primers were used together with P1 to amplify the left-hand part of the mutated TndX, while the positive-strand primers were used together with P2 to amplify the right-hand part. P1 and P2 were then used to join the left- and right-hand segments of the DNA coding for the proteins, using equal molar amounts of the corresponding left- and right-hand PCR products as templates. These products were cloned into pET19 as previously described for tndX. DNA sequencing verified that no changes, other than the desired mutation, were introduced by these manipulations (data not shown). pET-TndX(N261) was constructed by using primers P1 and P9 (5' GCGGGATCCCTAAGATACCATGTGTCCTAG 3') to amplify the region encoding the N-terminal 261 amino acids of TndX and then cloning the PCR product into pET19 as described above; again, the construct was verified by DNA sequencing. To make miniTn5397A, the catP gene was amplified from pJIR62 (1) using primers P10 (5' CCTCTGCTTGTTCAGTTTCCGGGAGTGCAGTCGAAGTGGGC 3') and P11 (5' CGTTCCCCACCCAATAGACCGGTCTTTGTACTAACCTGTGG 3'). The left and right ends of Tn5397 was amplified using P12 (5' GCGGGATCCGCATATTACGCATCTCATTA 3') and P13 (5' GCCCACTTCGACTGCACTCCCGGAAACTGAACAAGCAGAGG 3') as a pair and P14 (5' CCACAGGTTAGTACAAAGACCGGTCTATTGGGTGGGGAACG 3') and P15 (5' GCGGGATCCGAAAACTGCTTGGATTCAG 3') as a pair, respectively. Since primers P10 and P13 and primers P11 and P14 are complementary to each other, these three PCR products were joined by subsequent PCRs to produce mini-Tn5397A. pSU-mini-Tn5397A was constructed by cloning mini-Tn5397A into pSU39 (4) via the BamHI site. Plasmids pET-CDatt and pET-TndX-CDatt were constructed by cloning the Tn5397 target site from C. difficile CD37 (including about 200 bp on each side of the insertion point) into pET19 and pET-TndX respectively, via the BamHI site. For pET-BSatt and pET-TndX-BSatt, the Tn5397 target site found in B. subtilis BS2 was amplified by PCR using P22 (5' GCGGGATCCCTTCCCGCGCGAATATCG 3') and P21 (5' GCGGGATCCGAAAACGGATGGGAATACG 3') as primers and genomic DNA from B. subtilis CU2189 as template; this product (including about 200 bp on each side of the insertion point) was cloned into pET19 and pET-TndX, respectively, via the BamHI site. pSU-mini-Tn5397B was made by cloning the joint region of the circular form of Tn5397 (amplified by PCR using primers P16 [5' ACCGGCGGATCCACGTGTATCAAGCAGAGGGAATCGGTAAA 3'] and P17 [5' AAAGGCGGATCCACAACCAGCAGGAAAACA 3']) into pSU39 via the BamHI site.

Protein expression and purification. TndX and each of its derivatives were expressed in E. coli BLR(DE3) cells as a fusion to the C terminus of a 10-histidine tag and purified by using TALON resin (Clontech). The protein expression and purification procedures from the manufacturers (Novagen and Clontech) were followed with some modifications. Briefly, to check protein expression, the cell pellet from 1 ml of culture was completely dissolved in denaturing lysis buffer (20 mM Tris-HCl [pH 8.0], 100 mM NaCl, 8 M urea), purified with TALON resin, eluted with 100 mM EDTA, and analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE).

Excision and insertion assays. In the excision assay, pSU-mini-Tn5397A, together with pET19 or with pET-TndX, was introduced into E. coli BLR(DE3) cells by transformation. Protein expression was induced by adding isopropyl--D-thiogalactopyranoside (IPTG) to a final concentration of 1 mM to the cell culture (optical density at 550 nm = 0.5), and the mixture was left for about 16 h before the cells were harvested. The samples were analyzed by SDS-PAGE to confirm the expression of TndX. Plasmid DNA was prepared (Qiagen) from these cells, and PCRs were conducted to analyze the excision products with the plasmid DNA as templates. Primers P18 (ACGTGTATCAAGCAGAGGGAATCGGTAAA) and P19 (CCACTTGATATGAAAAATCAAATGGCTC) were used to detect the formation of the circular form of the transposon. Primers P12 and P15 were used to detect the regenerated target site. These PCR products were analyzed by agarose gel electrophoresis and by DNA sequencing using Big Dye mix (Applied Biosystems) and an ABI PRISM 310 genetic analyzer (Applied Biosystems). In the insertion assay, pSU-mini-Tn5397B, together with pET-CDatt or pET-TndX-CDatt, was introduced into E. coli BLR(DE3) cells and TndX expression was induced. PCRs were conducted to detect the insertion of mini-Tn5397B into the target site. Primers P19 and P15 were used for the left transposon end-host genome junction, and P12 and P18 were used for the right. These PCR products were analyzed by agarose gel electrophoresis and DNA sequencing.

	RESULTS AND DISCUSSION

Top Abstract Introduction Materials and Methods Results and Discussion References

Expression of TndX in E. coli. The tndX gene is 1,599 bp and encodes a putative protein of 61.5 kDa (46). To confirm the presence of this protein and to obtain sufficient quantities for functional studies, tndX was amplified by PCR from genomic DNA of a Tn5397-bearing C. difficile strain 630 (Table 1) and cloned into pET19 to make pET-TndX. DNA sequencing of this construct confirmed that tndX was in the correct reading frame and that no mutations had been created as a result of the amplification and cloning procedures (data not shown). TndX was then expressed in E. coli BLR(DE3) cells as a fusion to the C terminus of a 10-histidine tag. This protein was purified under denaturing conditions (see Materials and Methods) and analyzed by SDS-PAGE. A protein of about the expected size, 61.5 kDa, was detected from cells carrying pET-TndX but not from those carrying pET19 (Fig. 1).

View larger version (45K): [in this window] [in a new window]   FIG. 1.   Expression of TndX in E. coli. TndX was expressed in E. coli BLR(DE3) cells as a fusion to the C terminus of a 10-histidine tag. This protein was purified under denaturing conditions and analyzed by SDS-PAGE. Lanes: M, SDS-PAGE molecular mass markers (Sigma) (in kilodaltons); 1, protein from cells containing pET19; 2, protein from cells containing pET-TndX.

TndX can catalyze excision and insertion of mini-Tn5397s in E. coli. To determine whether the overexpressed TndX was functional, independent assays were designed to test the two crucial steps of transposition, excision and insertion, upon expression of TndX in E. coli. Since the 21-kb Tn5397 was too large to be manipulated easily, two minitransposons were constructed. Mini-Tn5397A represents the linear parental transposon, and mini-Tn5397B represents the circular form of the transposon. These were used in the excision and insertion assays, respectively (Fig. 2A and 3A). In mini-Tn5397A, the central region of Tn5397 was replaced by a chloramphenicol acetyltransferase gene (catP) (42) while each side retained a fragment containing about 250 bp of element end sequence plus about 200 bp of C. difficile flanking genomic sequence. The 1.7-kb mini-Tn5397A was cloned into pSU39 to make pSU-mini-Tn5397A (Fig. 2A). To test for excision of mini-Tn5397A, pSU-mini-Tn5397A, together with pET-TndX or pET19, was introduced into E. coli BLR(DE3) cells before TndX expression was induced. The production of TndX was confirmed by SDS-PAGE after induction (data not shown). Plasmid DNA was prepared from the cells and used as templates in PCRs to detect the presence of the circular form of the minitransposon and the empty target site. The circular form was detected by PCR using primers P18 and P19 to amplify the joint of the two ends of the element (Fig. 2A). A product of about the expected size, 215 bp, was obtained in cells containing pET-TndX but not in those containing pET19 (Fig. 2B, left, lanes 1 and 2). DNA sequence analysis of this PCR product showed that it contained the ends of the element joined by a GA dinucleotide (data not shown). This joint was identical to that of the circular form of Tn5397 in C. difficile (46). To detect the empty target site remaining after mini-Tn5397A excision, primers P12 and P15 were used to amplify the target site by PCR. A PCR product of about the expected size, 438 bp, which represented the regenerated target site, was obtained only when TndX was expressed. The other PCR product, of about 1.7 kb, representing mini-Tn5397A integrated into its target site, was obtained in samples with and without the presence of TndX (Fig. 2B, right, lanes 1 and 2). DNA-sequencing analysis of the smaller PCR product showed that it was a regenerated target site identical to that in the C. difficile genome. The above data show that TndX overexpressed in E. coli catalyzes the excision of mini-Tn5397A, formation of a circular molecule, and regeneration of the target site. This is the same as the behavior of Tn5397 on excision in C. difficile (46).

View larger version (44K): [in this window] [in a new window]   FIG. 2.   Excision reactions catalyzed by wild-type and mutated TndX. (A) Structure of mini-Tn5397A and diagram of the excision assay. Mini-Tn5397A used in this assay contains a chloramphenicol acetyltransferase gene (catP, dotted box in the figure) flanked by fragments containing about 250 bp of the Tn5397 end sequence (diamond boxes) plus about 200 bp of C. difficile genomic sequence (wavy-line boxes) on each side. This minitransposon was cloned into pSU39 (single black line) to make pSU-mini-Tn5397A (Table 1). The GA dinucleotides at the transposon-genome junctions, at the joint of the left and right ends in the circular form, and at the regenerated target site are also shown. Primers that are used in PCRs are shown, and their orientation is represented by an arrow. (B) Detection of the products generated by excision of the minitransposon using PCR and agarose gel electrophoresis. All the lanes contain PCR products derived from plasmid preparations from E. coli BLR(DE3). In the left-hand gel, primers P18 and P19 were used in PCRs to detect the joint of the circular form of the minitransposon. In the right-hand gel, primers P15 and P12 were used to detect the presence of the minitransposon or the empty target site remaining after excision. Lanes: M, 100-bp DNA marker (Promega); 1, plasmids pET19 and pSU-mini-Tn5397A; 2. plasmids pET-TndX and pSU-mini-Tn5397A; 3, plasmids pET-TndX-(R17-A) and pSU-mini-Tn5397A; 4, plasmids pET-TndX-(S19-A) and pSU-mini-Tn5397A; 5, plasmids pET-TndX-(R93-A) and pSU-mini-Tn5397A; 6, plasmids pET-TndX-(N261) and pSU-mini-Tn5397A.

View larger version (40K): [in this window] [in a new window]   FIG. 3.   Insertion reactions catalyzed by wild-type and mutated TndX. (A) Structure of mini-Tn5397B and diagram of the insertion assay. Mini-Tn5397B used in this assay contains the joint of the ends of Tn5397 cloned in pSU39. The target site of Tn5397 in C. difficile was cloned into pET19, pET-TndX, pET-TndX-(R17-A), pET-TndX(S19-A), pET-TndX(R93-A), and pET-TndX(N261) to generate pET-CDatt, pET-TndX-CDatt, pET-TndX(R17-A)-CDatt, pET(S19-A)-CDatt, pET-TndX(R93-A)-CDatt, and pET-TndX(N261)-CDatt, respectively. Mini-Tn5397B, together with one of pET-CDatt, pET-TndX-CDatt, pET-TndX(R17-A)-CDatt, pET(S19-A)-CDatt, pET-TndX(R93-A)-CDatt, and pET-TndX(N261)-CDatt, was introduced into E. coli BLR(DE3) cells. The target site is represented by boxes containing wavy lines, the ends of Tn5397 in the minitransposon are represented by diamond boxes, and the single black line represents vector sequences. The GA dinucleotides present at the center of the target site, at the joint of the circular form, and at the transposon end-target sequences junctions are also shown. Primers used in PCRs are shown, and their orientation is represented by an arrow. Upon expression of TndX, total plasmid DNA was prepared and used as template in the PCRs to detect the insertion of the minitransposon by amplifying the transposon-target junctions. (B) Detecting the insertion of the minitransposon by PCR and agarose gel electrophoresis. All the lanes contain PCR products derived from plasmid preparations from E. coli BLR(DE3). In the left-hand gel, primers P15 and P19 are used to generate the PCR products (detection of the left transposon-target junction), and in gel, the right-hand primers P18 and P12 are used to generate the PCR products (detection of the right transposon-target junction). Lanes: M, 100-bp DNA marker (Promega); 1, plasmids pET-CDatt and pSU-mini-Tn5397B; 2, plasmids pET-TndX-CDatt and pSU-mini-Tn5397B; 3, plasmids pET-TndX(R17-A)-CDatt and pSU-mini-Tn5397B; 4, plasmids pET-TndX(S19-A)-CDatt and pSU-mini-Tn5397B; 5, plasmids pET-TndX(R93-A)-CDatt and pSU-mini-Tn5397B; 6, plasmids pET-TndX(N261)-CDatt and pSU-mini-Tn5397B.

Point mutations of key amino acids and deletion of the C-terminal extension all abolish the activity of TndX. TndX is homologous to other members of the resolvase/invertase family at its N terminus but has an unique extended C terminus (46). To investigate the role of some of the highly conserved residues in the N terminus of TndX (13, 15, 26) on excision and insertion of Tn5397, arginine 17, serine 19, and arginine 93 were changed to alanine by site-directed mutagenesis. A C-terminal deletion which retained the N-terminal 261 amino acids containing the complete resolvase/invertase domain of TndX was also created to see whether this domain alone is functional. Each of the mutated genes was cloned into the expression vector pET-19 as described for TndX, and the mutations were confirmed by DNA sequence analysis. The expression of each of these proteins was induced and confirmed by SDS-PAGE (data not shown). Assays were then carried out to analyze the ability of each of the mutated TndX proteins to catalyze excision and insertion of mini-Tn5397A and mini-Tn5397B, respectively, in E. coli. Neither joined element ends nor empty target sites were detected by PCR when each of these four mutant TndX derivatives was used in the excision assay (Fig. 2B, lanes 3 to 6), indicating that the element was not excised, so that the two ends could not be joined to generate a circular minitransposon. For the insertion assay, the target site of Tn5397 in C. difficile was cloned into the plasmids containing the mutant tndX genes, i.e., into pET-TndX(R17-A), pET-TndX(S19-A), pET-TndX(R93-A), and pET-TndX(N261) to generate pET-TndX(R17-A)-CDatt, pET-TndX(S19-A)-CDatt, pET-TndX(R93-A)-CDatt and pET-TndX(N261)-CDatt, respectively. In this assay, no transposon-target sequence junctions were detected by PCR when each of the four mutated proteins were used (Fig. 3B, lanes 3 to 6), suggesting the transposon could not recognize its target and insert in the presence of each of these mutated proteins. These results showed that arginine 17, serine 19, and arginine 93 of TndX were essential for the resolvase/invertase domain of TndX to catalyze excision and insertion reactions. The related large resolvases TnpX and the C31 integrase were also subject to site-directed mutagenesis of amino acids in comparable sites; it was found that these mutations abolish the activity of these two proteins (8, 45). Taken together, these data provide good evidence that the conserved amino acids are absolutely required for integration; our data also indicate that they are required for excision.

Target site choices of mini-Tn5397B in E. coli. Tn5397 was shown to insert into one specific target sequence in C. difficile but could enter different targets in the B. subtilis genome (46). For all the targets analyzed, no consensus sequences could be deduced apart from the central GA dinucleotide core (46). To determine the preference of mini-Tn5397B insertion between the original targets and other GA-containing sequences nearby in E. coli, the Tn5397 targets identified from C. difficile 630 and B. subtilis BS2 were amplified by PCR from genomic DNA of C. difficile CD37 and B. subtilis CU2189. The 400-bp DNA fragments containing one of the target sites from C. difficile CD37 and from B. subtilis BS2 were cloned into pET-TndX to make pET-TndX-CDatt and pET-TndX-BSatt, respectively (Table 1). Either pET-TndX-CDatt or pET-TndX-BSatt, together with pSU-mini-Tn5397B, was transformed into E. coli BLR(DE3) cells. Expression of the tndX gene was induced by ITPG. It was anticipated that TndX would catalyze the insertion of mini-Tn5397B into the target site on pET-TndX-CDatt or pET-TndX-BSatt, resulting in fusion of the two plasmids. However, there could be a mixture of plasmids, i.e., pET-TndX-CDatt (or pET-TndX-BSatt), pSU-mini-Tn5397B, and fused plasmids (some of the latter may have fused at different sites, depending on which targets are recognized by TndX). This would make the subsequent PCR experiment difficult to interpret (see below). Therefore, plasmids were extracted from the cells after induction and the plasmid mixture was used to transform E. coli XL-1 Blue. Transformants were selected on Luria-Bertani plates containing ampicillin and kanamycin. Each colony should contain a single type of fused plasmid with mini-Tn5397B inserted in its target. Plasmids were extracted from single colonies and used as templates in PCRs. A pair of primers, one reading out from the element end, P19, and the other reading toward the cloned target site from the vector, P20, were used (Fig. 3A). A total of 40 transformants containing inserted mini-Tn5397B for each cloned target site were analyzed. All 40 samples for the C. difficile 630 target gave the same-sized PCR products, as did the other 40 samples for the B. subtilis BS2 target. Further analysis of these PCR products by DNA sequencing showed that they all represented the minitransposon inserted into the GA dinucleotide of the cloned C. difficile or B. subtilis target site (data not shown). These results indicate that only the original target is used in these experiments. To rule out the possibility that all the transformants we analyzed in these experiments were siblings, we repeated the experiment and obtained identical results. No other sites were chosen as targets, although there were another 22 GA dinucleotides located in the region between the P20 primer and the central GA dinucleotide of the cloned target site. This data showed that TndX, the only Tn5397-encoded protein, determines the target site specificity of the minitransposon in E. coli.

Concluding remarks. In summary, we have demonstrated that minitransposons containing the ends of Tn5397 can transpose in E. coli. TndX is the only Tn5397-encoded protein required for excision and insertion of these elements. This showed for the first time that a member of the large-resolvase subgroup of the resolvase/invertase family of site-specific recombinase can mediate the insertion and excision of a conjugative transposon. The simple requirements for the transposition of this element indicate that it is a promising candidate to be modified to create a useful vector for gram-positive and gram-negative bacteria.