Identification of the Structural and Functional Domains of the Large Serine Recombinase TnpX from Clostridium perfringens *

Members of the large serine resolvase family of site-specific recombinases are responsible for the movement of several mobile genetic elements; however, little is known regarding the structure or function of these proteins. TnpX is a serine recombinase that is responsible for the movement of the chloramphenicol resistance elements of the Tn 4451/3 family. We have shown that TnpX binds differentially to its transposon and target sites, suggesting that resolvase-like excision and insertion were two distinct processes. To analyze the structural and functional domains of TnpX and, more specifically, to define the domains involved in protein-DNA and pro-tein-protein interactions, we conducted limited proteolysis studies on the wild-type dimeric TnpX 1–707 protein and its functional truncation mutant, TnpX 1–597 . The results showed that TnpX was organized into three major domains: domain I (amino acids (aa) 1–170), which included the resolvase catalytic domain; domain II (aa 170–266); and domain III (aa 267–707), which contained the dimerization region and two separate regions involved in binding

Mobile genetic elements, such as conjugative plasmids, transposons, integrons, and genomic islands are important vehicles for the transmission of virulence and antibiotic resistance genes in many microorganisms including Gram-positive bacteria such as Clostridium sp. (1,2). Antibiotic resistance transposons identified in the clostridia include the integrative mobilizable elements Tn4451 and Tn4453a, which confer chloramphenicol resistance. Integration and excision of these elements is mediated by the large serine recombinase TnpX (3)(4)(5).
Members of the serine recombinase family of site-specific recombinases catalyze strand exchange by a non-replicative DNA breakage and repair mechanism that involves a 2-bp staggered break across all four DNA strands and the formation of covalent phosphoserine linkages between the DNA strands and recombinase subunits (6). The large resolvases represent a subgroup within the serine recombinase family (7). The members of this subgroup are significantly larger (50 -82 kDa) than most other serine recombinase proteins (20 kDa) and catalyze a wider range of reactions (7). Little is known regarding the domain organization of large resolvases, and to date, the function and the mechanism by which the recombination process occurs is unclear.
Unlike Tn3-like resolvase proteins that only catalyze excision reactions, large serine recombinases have the ability to catalyze both the excision and integration of specific mobile DNA elements such as bacteriophage genomes, genomic islands, and mobile integrative elements (7). Several phage-encoded serine integrases related to TnpX have been shown to catalyze the insertion of their respective genomes into a specific target site (attP/attB recombination). The efficient excision of these integrated phage genomes has been demonstrated or postulated to require a recombination directionality factor in addition to the phage integrase (8 -12). By contrast, TnpX does not require any other proteins for excisive recombination (13).
Typical small resolvase proteins consist of two domains: an N-terminal catalytic domain (residues 1-140) that also mediates dimerization and a C-terminal helix-turn-helix DNA binding domain (7,14,15). The large serine recombinases identified to date show a high level of N-terminal sequence similarity to the catalytic domain of small resolvases, especially over the first 100 amino acids (aa). 1 The critical serine residue within the catalytic domain is conserved across all of the serine recombinases (7), and mutation of the proposed catalytic serine in TnpX abolishes recombinase activity (16). The catalytic resolvase domain is followed by a region of sequence conservation found only between members of the large serine recombinase group (7). None of the large recombinases appear to have a helix-turn-helix motif (7).
Although the biological properties of the truncated TnpX 1-597 * This research was supported by grants from the Australian National Health and Medical Research Council and the Australian Research Council. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  protein are not identical to those of the wild-type protein, it is clear that the Cys-rich C-terminal 110 aa of the 707-aa TnpX protein ( Fig. 1) are not essential for biological function (17). The TnpX binding sites at the left and right ends of Tn4451 (attL and attR, respectively) and at the junction of the circular intermediate (attCI) have been identified, and deletion analysis has shown that the presence of the highly charged 493-597 aa region is essential for DNA binding and biological activity (17). The next region between aa 392 and 492 appears to be required for dimerization. An unusual feature of TnpX is that, although it binds with almost equivalent affinity to its excisional target sites, attL and attR, it binds much more strongly to attCI than it does to any of its eventual attT insertional target sites. These data provide evidence that different types of synapses are formed in the excision and insertion reactions (17).
TnpX also contains a region that contains a cysteine-rich motif with a consensus sequence, LX 5 CX 2 CGX 13-27 YXCX 2-21 C (aa 318 -360) (7), which could be a zinc finger-like domain, and a small leucine-and valine-rich region (aa 493-527), which could be a leucine zipper-like domain (Fig. 1). Both of these regions have a high probability of forming coiled-coil structures and hence may be involve in protein-protein interactions (17).
In this study, we analyzed the structural and functional domain organization of TnpX by limited proteolysis analysis. The resultant fragments were identified, their gene regions subsequently were cloned and overexpressed, and the purified polypeptides were tested for their ability to bind specific target DNA sequences and to dimerize. The results showed that TnpX was organized into three major domains and that the N-terminal resolvase domain was not required for dimerization. A strong intramolecular interaction between the N-and C-terminal regions was revealed, and a previously unknown DNA binding site was identified between residues 583 and 707. Finally, it was shown that a small polypeptide consisting of the aa 533-587 region was capable of binding specifically to the TnpX binding sites.

EXPERIMENTAL PROCEDURES
Cloning, Expression, and Purification of TnpX Constructs-The strains and plasmids used for this study are described in Table I. The  six separate TnpX regions (Table I) were amplified by PCR from a plasmid carrying the relevant TnpX sequence. The oligonucleotides used for each construct (Table II) included an NdeI restriction site that incorporated a start codon and a XhoI site at the 3Ј end prior to the stop codon. PCR products were digested with NdeI and XhoI and cloned into pET22b for eventual overexpression with a hexahistidine tag (His 6 tag) at the C-terminal end. Each plasmid generated (Table I) was used to transform Escherichia coli C43(DE3) (18) cells to ampicillin resistance. Cells were grown in 500 ml of 2YT medium (19) containing ampicillin (100 g/ml) at 37°C to a turbidity at 600 nm of 0.5-0.6. The cultures were then transferred to 30°C, and expression was induced with 0.5 mM isopropyl-␤-D-thiogalactoside for 4 h. Cells were harvested by centrifugation at 6500 ϫ g for 10 min, and cell pellets were stored at Ϫ70°C until use. The cells were resuspended in 30 ml of cold buffer A (50 mM Tris-HCl (pH 7.2), 0.5 M NaCl, 5 mM ␤-mercaptoethanol) containing 2 mM phenylmethylsulfonyl fluoride and disrupted in a precooled French press and then by sonication (twice for 30 s). After centrifugation at 10,000 ϫ g for 45 min, the supernatant fraction was loaded onto a 2-ml Talon column (Clontech) preequilibrated in buffer A. The column was washed with 100 ml of buffer A, and the proteins were eluted using an imidazole gradient (5-400 mM). TnpX proteins were eluted between 50 and 200 mM imidazole. The peak fractions were pooled and either loaded onto a Superdex 200 column (Amersham Biosciences) for further purification or dialyzed against a solution containing 50 mM Tris-HCl (pH 7.2), 0.35 M NaCl, 5 mM ␤-mercaptoethanol, and 50% glycerol and stored at Ϫ70°C until needed.
Size Exclusion Chromatography-Aliquots (1-2 mg) of purified preparations of TnpX derivatives were applied to a Superdex 200 16/30 gel    filtration column preequilibrated in 50 mM Tris-HCl (pH 7.2), 0.35 M NaCl, and 2 mM DTT. Elution was performed in the same buffer at 0.4 ml/min, in 1-ml fractions collected, and in 15-l samples examined by Coomassie Blue staining of 12% SDS-PAGE gels. The column was calibrated using standard proteins (gel filtration calibration kit, Bio-Rad). The apparent molecular size of each protein was determined by interpolation from a standard curve of log molecular size versus K av . Limited Proteolysis-Purified TnpX 1-707 and TnpX 1-597 (80 -130 g) were solubilized in 50 mM Tris-HCl (pH 7.2), 0.5 M NaCl, and 2 mM DTT and incubated with chymotrypsin or trypsin (Sigma) (0.025-0.07 g) for various times at 30°C. The reactions were stopped by the addition of 3ϫ SDS sample buffer, and the samples were boiled for 3 min. The proteolytic fragments were then separated by SDS-PAGE on 10 or 12% gels and either visualized by Coomassie Blue staining or immunostained using anti-His tag antibodies.
Protein Sequence Determination-Purified TnpX 1-597 was digested with chymotrypsin and proteolytic fragments separated by 12% SDS-PAGE, electroblotted onto polyvinylidene difluoride membranes, and stained with Coomassie Blue. The desired bands were then excised, and the N-terminal amino acid sequence was determined by automated Edman degradation performed at the Protein Chemistry Facility, La Trobe University.
Circular Dichroism Spectroscopy-Circular dichroism measurements were performed on a Jasco 810 spectropolarimeter using a 0.1-cm path length cuvette at 25°C. TnpX derivatives were examined in the far-UV range between 200 and 250 nm in 50 mM Tris-HCl (pH 7.2), 0.35 M NaCl, and 2 mM DTT. Each spectrum represents the average of 5-10 scans processed for base-line subtraction and smoothing using software provided by the manufacturer.

RESULTS
TnpX Is Organized into Discrete Domains-Computer analysis of the primary structure of TnpX led to the suggestion that TnpX is a multidomain protein (17). To investigate the structural organization of TnpX, limited proteolytic digestion experiments using chymotrypsin were performed on purified fulllength TnpX 1-707 and a truncated version of the protein, TnpX 1-597 , which has been shown to be fully functional in vivo (17). The results (Fig. 2) revealed a clear digestion pattern with a relatively small number of chymotryptic fragments, indicating that many of the Phe, Tyr, Leu, and Trp residues of TnpX are not readily accessible to the protease. For each protein, digestion yielded three major products with molecular masses of 64, 52, and 46 kDa for TnpX 1-707 ( Fig. 2A) and 51, 39, and 33 kDa for TnpX 1-597 (Fig. 2B). The difference in masses of the fragments generated from TnpX 1-707 and TnpX 1-597 corre- sponded to the absence of the last 110 C-terminal amino acids in the latter protein. All of these fragments were shown to be C-terminal fragments, because they reacted with anti-His tag antibodies (Fig. 2, bottom panels).
To determine whether the different proteolytic fragments were generated by cleavage at sites that were also highly accessible to other proteases, trypsin was used to digest both TnpX 1-707 and TnpX 1-597 . Both proteins yielded digestion products very similar to those obtained with chymotrypsin (data not shown).
Identification of the N-termini of the Stable Proteolytic Fragments-Two proteolytic fragments generated from TnpX 1-597 , corresponding to the 39-and 33-kDa bands (Fig. 2B), were submitted for N-terminal sequencing so that the chymotrypsin digestion site could be mapped. The N-terminal sequences of these fragments were KLHKRK and GTHSNR, respectively, indicating that cleavage had occurred before Lys-267 and Gly-312. We could not generate sufficient amount of the 51-kDa product for sequencing, but based on the size of this TnpX fragment, we suggest that cleavage occurred between residues 165 and 170. Because the proteolytic pattern obtained with TnpX 1-707 was comparable to that obtained with TnpX 1-597 , it was concluded that the 46-kDa proteolytic fragment generated from TnpX 1-707 resulted from cleavage at the same site. Note that all of the chymotrypsin cleavage sites were located within putative random coil regions according to secondary structure predictions carried out using PSIPRED (22).
Based on these data, it appeared that TnpX was organized into three major domains: 1) an N-terminal domain (aa 1-165/ 170) that included the catalytic resolvase domain; 2) a middle domain from aa 165/170 to 266; and 3) the C-terminal domain (from aa 267), which could be further digested at residue 312. This last product was very stable and had a predicted ␣-helical structure. The middle domain contained two regions of unknown function that were conserved within the large serine recombinase family (7), namely, residues 177-203 (predicted secondary structure: ␤-strand-␣-helix) and residues 240 -263 (predicted secondary structure: ␣-helix-␤-strand).
The Stable Proteolytic Fragments TnpX 267-707 and TnpX 267-597 Are Correctly Folded and Are Still Capable of Dimerization-We have recently shown by gel filtration that TnpX 1-707 is a dimer in solution (17). Structural predictions indicated that there was a high probability that the C-terminal domain consisted of a significant amount of ␣-helical structures and multiple coiled-coil regions. As a consequence, we asked whether the C-terminal region of TnpX was involved in the dimerization process. Proteolytic fragments generated from TnpX 1-707 and TnpX 1-597 were assayed for their ability to multimerize using Far-Western analysis. The products of chymotrypsin digests were separated by SDS-PAGE and transferred to a nitrocellulose membrane. The proteins were renatured by incubation of the membrane under conditions that allowed the recovery of activity of the undigested protein. The membrane was then probed with N-terminal T7tagged TnpX 1-707 before incubation with anti-T7 antibodies. The results (Fig. 3) showed that TnpX 1-707 , TnpX 1-597 , and both TnpX 267-707 and TnpX 267-597 could bind T7-tagged TnpX 1-707 under these conditions. These results suggest that TnpX 267-707 and TnpX 267-597 are still capable of dimerization.
To confirm these conclusions, proteins corresponding to aa 267-707 and 267-597 were cloned into pET22b, overexpressed as C-terminal His 6 -tagged proteins, and purified as described previously (13). After metal chelation chromatography on a Talon resin, the proteins were over 95% pure and showed an apparent molecular mass of 53 and 40 kDa, respectively. An analysis of their CD spectra indicated that both TnpX 267-707 and TnpX 267-597 were folded in solution, because they showed the double minima at ϳ222 and 208 nm, which is characteristic of an ␣-helical conformation (Fig. 4A). These data are in agreement with secondary structure predictions from primary sequence analysis (22). Gel filtration experiments revealed that both TnpX 267-707 and TnpX 267-597 were dimers in solution (Fig. 4B).
To see whether a TnpX 312-597 fragment was still capable of dimerization, we carried out limited chymotryptic digestion on TnpX 267-597 and separated the reaction products on a Superdex 200 column. The TnpX 312-597 product was eluted with an ap-parent molecular mass of 95 kDa, in agreement with the hypothesis that the dimerization regions were located between aa 312 and 597 (data not shown).
TnpX 267-707 Lacks the Resolvase Domain but Still Binds DNA Specifically-Previous studies on the large resolvase SpoIVCA have suggested that the region located N-terminal to the cysteine-rich region (before residue 310) could be involved in DNA binding (23). By contrast, recent studies using different truncated versions of TnpX have shown that purified TnpX 1-356 and TnpX 1-492 have lost the capability to bind DNA (17). To identify and localize the DNA binding domain, chymotrypsin fragments generated from TnpX 1-707 and TnpX 1-597 were tested for DNA binding activity by Southwestern blotting using a specific 32 Plabeled attL fragment as a probe. TnpX 1-707 and each of the major C-terminal chymotryptic products, TnpX 170 -707 , TnpX 267-707 , and TnpX 312-707 , were still able to bind DNA under these conditions (Fig. 5). These results implied that a DNA binding domain was located in the C-terminal region of the protein as expected from previous studies that localized the major DNA binding domain to the aa 492-597 region (17). When a partial chymotryptic digest of TnpX 1-597 was examined in the same way, TnpX 1-597 and each of the C-terminal proteolytic fragments were also shown to bind DNA but with reduced affinity (data not shown). These data suggested that either TnpX 1-597 was incorrectly refolded after renaturation or that a DNA binding domain was also located between residues 598 and 707.
To confirm the data obtained by Southwestern analysis and to further define the DNA binding regions, DNA binding experiments were conducted on purified TnpX 267-707 , TnpX 267-597 , and TnpX 312-597 using TnpX 1-707 and TnpX 1-597 as positive controls. The specific binding of each of the purified proteins at similar concentrations was determined by gel mobility analysis using a radioactively labeled 360-bp attL fragment and a 405-bp attCI fragment. Only TnpX 1-707 , TnpX 1-597 , and TnpX 267-707 interacted with these fragments (Fig. 6, A and B, lanes 2-4). To confirm that TnpX 267-707 , which does not contain the catalytic resolvase domain, was binding the DNA specifically, competitive gel mobility shift assays were performed. The addition of an excess of unlabeled attL DNA completely abolished the binding of both TnpX 1-707 and TnpX 267-707 , whereas the addition of the same amount of nonspecific DNA (an unrelated pfoA gene fragment from Clostridium perfringens) had no effect (Fig. 6C). Finally, in the absence of the N-terminal resolvase domain, the removal of the last 110 aa had a dramatic effect on DNA binding since TnpX 267-597 and TnpX 312-597 only bound with very low affinity to attL and attCI (Fig. 6, A and B, lanes 5  and 6).
The Charged Domain (aa 533-583) and the Cysteine-rich Domain (aa 583-707) Are Both Involved in DNA Binding-These experiments and those of the previous study (17) provide evidence that the regions from aa 493 to 597 and aa 598 to 707 are both involved in DNA binding. The amino acid region of 493-597 contains two strings of oppositely charged amino acids, whereas the C-terminal cysteine-rich region (aa 598 -707) contains two putative FCS (consensus sequence CX 2 CX 9 -24 FCSX 2 CX 3 (F/Y)) zinc-coordinating domains (24), which are located between residues 591-616 and 639 -664.
To provide more direct evidence that the oppositely charged region and the FCS motifs were involved in DNA binding, fragments encoding aa 533-597, aa 533-707, and aa 583-707 were cloned into pET22b and the resultant proteins were overexpressed and purified as before. The level of expression for TnpX 583-707 was very low, leading to a poor yield of purified TnpX 583-707 protein. After Talon chromatography, TnpX 533-707 and TnpX 533-597 were over 95% pure and showed an apparent molecular mass of 20 and 15 kDa, respectively, as observed by Tris-Tricine SDS-PAGE. The CD spectra of these polypeptides indicated that they were both folded with TnpX 533-597 showing the double minimal characteristic of an ␣-helical conformation (data not shown). In addition, gel filtration experiments revealed that both TnpX 533-597 and TnpX 533-707 were monomers in solution (data not shown). Gel mobility shift assays were then conducted on purified TnpX 533-597 , TnpX 533-707 , and the small amount of TnpX 583-707 that was available. Each of these proteins was able to form a complex with attL DNA (Fig. 7A). The addition of excess, unlabeled attL DNA abolished the binding of TnpX 533-597 and reduced the binding of TnpX 533-707 (Fig.  7, B and C, lanes 4 and 5), whereas the addition of the same amount of nonspecific DNA had no affect (Fig. 7, B and C, lanes   FIG. 7. Gel mobility shift analysis of purified TnpX 533-597 , TnpX 533-707 , and TnpX 583-707 . A, a fixed amount (0.1 pmol) of a 360-bp 32 P-labeled attL fragment was incubated with varying amounts of purified TnpX 533-597 (72.5, 50.75, and 29 pmol), TnpX 533-707 (22, 14.5, and 5.5 pmol), and TnpX 583-707 (23.75 and 11.9 pmol). The resultant protein-DNA complexes were analyzed on a 6% polyacrylamide gel. B and C, competition gel mobility shift assays were performed with purified TnpX 533-597 (B) and purified TnpX 533-707 (C) Both proteins were incubated at the highest and lowest protein concentrations used above with 32 P-labeled attL (0.1 pmol), an excess (5.1 pmol) of the same unlabeled fragment, or an excess of unlabeled nonspecific DNA (pfo fragment, 5.1 pmol). The first lane in each series is the no protein control. 6 and 7). Based on these results, it is concluded that TnpX contains at least two DNA binding regions consisting of residues 533-597 and 598 -707. Only the amino acid region of 533-597 is essential for biological activity (17).
TnpX  and TnpX 267-597 Interact and Can Cooperate in DNA Binding-Since TnpX 1-597 but not TnpX 267-597 was capable of binding specifically to the DNA, we decided to analyze the ability of the N-terminal resolvase domain to restore DNA binding after intermolecular association. Therefore, a TnpX 1-266 derivative was cloned into pET22b, the resultant protein purified as before and shown by CD spectroscopy to be folded (data not shown).
To determine whether there was any interaction between the N-and C-terminal domains of TnpX, gel filtration experiments were conducted with purified TnpX 1-266 and TnpX 267-597 . The elution patterns for TnpX 1-266 , TnpX 267-597 , and a TnpX 1-266 -TnpX 267-597 mixture were monitored by analyzing each fraction by SDS-PAGE. Purified TnpX 1-266 primarily eluted with an apparent molecular of 23.3 kDa (Fig. 8), corresponding to a monomer. Purified TnpX 267-597 eluted as described previously with an apparent molecular mass of 95.5 kDa, corresponding to a dimer. When both fragments (with an excess of TnpX 267-597 ) were preincubated for 15 min at room temperature before loading onto the gel filtration column, TnpX 1-266 was displaced to a high molecular form and co-eluted with TnpX 267-597 (Fig. 7) with an apparent molecular mass of 132 kDa corresponding to the size of a TnpX 1-597 dimer binding to two TnpX 1-266 monomers. These data suggested that there were protein-protein interactions between these two parts of the TnpX protein.
To see whether the interaction between these two domains was sufficient to restore DNA binding, gel shift experiments were carried out. The individual TnpX 1-266 and TnpX 267-597 proteins were not able to form a complex with attL DNA (Fig.  9, lanes 3 and 4). However, when these polypeptides were mixed in equimolar amounts and added to attL DNA, we observed a protein-DNA complex that migrated to the same position as the protein-DNA complex obtained with TnpX 1-597 (Fig. 9, lane 5). In addition, the gel filtration-purified TnpX 1-266 -TnpX 267-597 protein complex (Fig. 8) also had the ability to bind DNA (Fig. 9, lane 6). These data provide good evidence that intramolecular interactions occur between the N-and C-terminal fragments of TnpX and that these interactions can restore the ability of truncated derivatives to form a protein complex that binds to the specific TnpX binding site. Although we could restore the full DNA binding activity by incubating TnpX 1-266 with TnpX 267-707 (Fig. 9, lane 8), no in vitro or in vivo recombination activity could be detected (data not shown), indicating that covalent linkage is required for biological activity. DISCUSSION TnpX is the only protein responsible for the precise excision of Tn4451 and the only transposon-encoded protein required for integration (13). TnpX-mediated recombination requires the binding of TnpX, presumably as a dimer, to each of its target sites followed by synapse formation to allow correct alignment of the GA-dinucleotides at the cleavage site, which is presumably mediated by TnpX-TnpX interactions. The final steps in recombination involve strand cleavage, strand trans- fer, and religation. In the excision reaction, TnpX binds with equal affinity to the attL and attR ends of the transposon and strand exchange leads to the formation of a non-replicating circular intermediate (5,17). By contrast, integration involves the binding of TnpX to the attCI site at the joint of the circular intermediate, significantly weaker binding to an insertional target site, attT, and subsequent strand exchange (17).
Prior to this study, little was known regarding the domain structure of TnpX. We have now shown that TnpX consists of three major domains (I, II, and III), each separated by exposed linker regions (Fig. 10). Each of the regions sensitive to proteolytic digestion was located within regions that, according to secondary structure predictions, consisted of random coils (17). These regions were located primarily in the N-terminal part of the protein (around residue 170, at residue 266, and at residue 312).
The N-terminal domain, domain I, was found to extend to a region around residue 170 and encompassed the catalytic resolvase region (aa 1-104), which is highly conserved among serine recombinases (7) and has been shown to be essential for TnpX function (5,16). Domain II extended from aa 170 to 266 and contained two regions that were conserved between members of the large serine recombinase group (7). Finally, domain III stretched from residues 267 to 707 and was found to be highly resistant to proteolytic digestion. This domain contained a weakly conserved zinc finger-like motif (from residues 318 to 360) and two FCS-zinc finger motifs.
Gel filtration of the chymotryptic TnpX fragments showed that the 51-kDa fragment (residues 170 -597) remained tightly associated with purified TnpX 267-597 and TnpX 312-597 . Furthermore, when the N-and C-terminal fragments were produced as the recombinant proteins TnpX 1-266 and TnpX 267-597 , they were found by gel filtration analysis to interact strongly (Fig.  8), providing evidence for intramolecular interactions between these TnpX domains.
Previous studies have shown that TnpX 1-597 and TnpX 1-492 exist as dimers, whereas TnpX 1-356 exists in two forms, both a dimer and monomer (17). In this study, it was shown that TnpX 1-266 behaved as a monomer, whereas TnpX 312-597 formed a dimer. Together these data suggest that the dimerization site lies at least partially within the amino acid region of 312-356. This region corresponds to a weak C4 zinc finger motif (residues 324 -360) that has a predicted secondary structure also consistent with this type of motif (2-3 ␤-strands and an ␣-helix). Zinc finger motifs from many proteins have been shown to mediate both homodimerization and heterodimerization depending on the protein (25). However, further mutagenesis studies are required to confirm the role of this region in TnpX dimerization Although gel mobility shift assays conducted on TnpX 1-266 and TnpX 267-597 demonstrated that neither protein was capable of DNA binding, the complex formed when they were preincubated together was shown to have the same DNA binding properties as native TnpX 1-597 (Fig. 9). These data indicated that the complex had the correct structural conformation in solution and that, not only were both domains intimately connected to each other by strong intramolecular interactions, this interaction was required for DNA binding (Fig. 10). However, despite the fact that the DNA binding properties were fully restored and the complex exhibited a similar pattern on gel filtration, no recombinase activity was evident either in vitro or in vivo. We conclude that, whereas specific DNA binding can tolerate a discontinuity in the peptide backbone, the catalytic activity of TnpX cannot.
Recent studies that utilized truncated derivatives of TnpX indicated the presence of an essential DNA binding region between aa 492 and 597 (17). We have now confirmed this conclusion by the cloning and purification of a fragment localized to this region and have shown that it was properly folded and was able to bind DNA specifically. Why was this polypeptide (TnpX 533-597 ) active in DNA binding while TnpX 267-597 had very poor DNA binding activity? It would seem that the formation of a dimer in the absence of the 1-266 region inhibits DNA binding, possibly because the putative DNA binding site is buried and is not accessible to the DNA. Conformation change brought about by intramolecular interaction between the N-and C-terminal domains appears to be required for DNA binding that is mediated by the functional dimeric form. This process may be required for specific synapse formation prior to recombination. It is clear that no such constraints are placed on the much smaller TnpX 533-597 protein, which can bind efficiently as a monomer.
The C-terminal 110 aa of TnpX are not essential for biological activity, although their deletion does yield a TnpX enzyme with altered biological activity (17). Unexpectedly, the region between residues 583 and 707 was shown to bind specifically to TnpX target DNA. This region contains two FCS motifs (residues 591-612 and 639 -660) that are known to be zinc binding domains in various chromatinic proteins and transcription factors (24). DNase I protection studies had previously indicated that the area of protection afforded by TnpX 1-707 is more extensive than that of TnpX 1-597 (17), which may be due to additional protein-DNA contacts mediated by the FCS motifs that are absent in TnpX 1-597 . The contribution of the 583-707 region to the function of TnpX is unclear. When the last 110 aa of TnpX are removed, recombination between attL and attR is somewhat reduced, whereas in vivo attT/attCI recombination is more efficient (17).
In summary, in this study, we have identified two regions that were important for DNA binding, neither of which contained the catalytic resolvase domain. The DNA binding region from residues 533 to 597 could bind DNA independently and was essential for the biological activity of TnpX. This region clearly represents the major DNA binding site. The second DNA binding region, residues 583-707, was not essential for biological activity. In addition, the dimerization domain of TnpX was localized to a region in close proximity to a putative C4 zinc finger domain. Finally, TnpX was shown to contain three major domains with intramolecular association between the N-and C-terminal domains required for TnpX dimers to bind specifically to the DNA.