INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Bacterial conjugation is a major mechanism for the transfer of chromosomal or plasmid DNA between prokaryotic cells. As a result, dissemination of antibiotic resistance determinants can occur, significantly influencing the virulence of various pathogens such as Escherichia coli, agrobacteria, enterococci, staphylococci, streptococci, streptomycetes, and the Bacteroides fragilis group (2, 4, 10, 22, 33, 34). The initial reactions during conjugal DNA transfer require multiple complex protein-DNA and protein-protein interactions that result in the formation of a relaxosome at the cis-acting origin of transfer (oriT). Relaxosomes are stable nucleoprotein complexes resulting in site- and strand-specific nicking at a site (the nic site) in the oriT region (24). The relaxase, or DNA strand transferase, cleaves at this nic site and attaches covalently to the 5' terminus. DNA is transmitted unidirectionally as a single strand into a recipient cell with 5'-to-3' polarity in a rolling-circle manner, followed by complementary-strand synthesis in the new host (for detailed reviews of conjugative transfer involving R and F factors, see references 25 and 44).

Much attention has been focused on the biochemical processes involved in the initiation of DNA transfer of the IncP plasmid RP4. The generation of the single DNA strand that is transferred requires formation of the initiation complex, or DNA relaxosome, in a cascade-like fashion. The first step in relaxosome formation is binding of an RP4 protein, TraJ, to a 19-bp inverted repeat within the oriT region and a 10-bp palindrome, srj (24, 43). In the second step, TraI, another RP4 protein, binds TraJ by both protein-protein interactions and the recognition of a 6-bp sequence, sri, in the nic region. Subsequently, site- and strand-specific cleavage of a unique phosphodiester bond at the nic site of the transfer origin occurs, followed by covalent attachment of TraI to the 5' terminus of the DNA strand involved in transfer. The binding and nicking events are thought to be independent processes (16, 26). TraI is also able to induce a second cleavage reaction proposed to terminate the rolling-circle model of transfer DNA replication (27). Following strand transfer to a new host, TraI catalyzes the recombination of two single-stranded DNA molecules at the RP4 nic site (27). A third protein, TraH, stabilizes the TraJ and TraI nucleoprotein complex by specific protein-protein interactions and does not bind DNA itself (24, 45). A fourth protein, TraK, interacts with oriT by binding DNA as a tetramer over a range of approximately 180 nucleotides, downstream of nic (46). Binding of TraK increases the formation of relaxation complexes in vitro and in vivo, possibly by influencing DNA topology to expose the nic site for more efficient cleavage by TraI (43).

Transferable antibiotic resistance plasmids of the B. fragilis group have also been described. They include the clindamycin-resistant plasmids pBFTM10 (12), pBF4 (40), and pBI136 (31) and the metronidazole-resistant plasmids pIP417, pIP419, and pIP421 (11, 37, 38). pBFTM10 is a 14.9-kb plasmid that is transferable within B. fragilis. When pBFTM10 is fused to the E. coli replicon pDG5 to form pGAT400, it can also transfer from B. fragilis to E. coli. Further, pGAT400 is also mobilizable within E. coli when coresident with the IncP plasmid R751, but it requires an intact pBFTM10 transfer region (12). Characterization of the pBFTM10 transfer region has demonstrated that only two genes, btgA and btgB, are necessary for transfer within B. fragilis and mobilization by R751 in E. coli. DNA sequence analysis of btgA revealed a helix-turn-helix DNA binding motif, suggesting that it was a DNA binding protein (12). In addition, the identification of three sets of inverted repeats (IRI, IRII, and IRIII) and a putative nic site in the region upstream of btgA suggested that this was the pBFTM10 oriT (12, 39).

In this paper, we report the cloning of the pBFTM10 oriT and the expression and purification of the BtgA protein, and we demonstrate specific binding of BtgA to its cognate oriT as determined by mobility shift assays and footprinting analyses.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Bacterial strains, plasmids, and media. The bacterial strains and plasmids used in this study are listed in Table 1. Media, antibiotic concentrations, and E. coli growth conditions were as previously described (1, 20). Bacto Agar and tryptone were obtained from Difco Co. (Detroit, Mich.). For strains requiring antibiotic selection, cells were grown at 37°C overnight in Luria broth (LB) supplemented with 200 µg of ampicillin per ml. E. coli XL1-Blue was used for routine cloning procedures (29). The putative oriT region was cloned into the BamHI site of pGEM-7Zf(+) (Promega, Madison, Wis.), to give pGEM-7ZT, and into the BamHI site of pACYC184, to form pJAYC1. E. coli BL21(DE3) was used as the host for btgA cloning and protein expression. A 646-bp fragment containing btgA was cloned into the BamHI site of the expression vector pET-19b (Novagen, Madison, Wis.) to give pDHBA.

Plasmid mobilization experiments. All mating experiments were performed as previously described (12, 23), in triplicate, using E. coli HB101 cells containing R751 as donors and E. coli DW1030 cells as recipients (Table 1). Bacterial cultures were grown overnight at 37°C in LB with the appropriate antibiotics. Overnight cultures were diluted 10-fold in LB and grown to early log phase. In a sterile microcentrifuge tube, 0.15 ml of donor cells was mixed with 1.35 ml of DW1030 cells and then pelleted by centrifugation at 10,000 × g for 1 min. After the supernatant was discarded, cell pellets were resuspended in 100 µl of 1× MPBS (8 mM Na₂HPO₄, 2 mM NaH₂PO₄, 145 mM NaCl [pH 6.9]). Resuspended cells were then spotted onto sterile 25-mm, 0.45-µm-pore-size Gelman GN-6 cellulose nitrate filters (Nalge Co., Rochester, N.Y.) on Luria agar plates and incubated at 37°C for 180 min. Transconjugants were enumerated by plating on selective media, and the mobilization frequencies of plasmids were normalized to the number of R751 transconjugants from the same experiment. The mobilization frequency of R751 was calculated as the mean ± standard error of the number of R751 transconjugants divided by the number of viable R751 donor cells.

DNA manipulations and enzymes. General DNA techniques were performed as described by Ausubel et al. (1). Purification of DNA for cloning was accomplished by using Geneclean II (Bio 101 Inc., Vista, Calif.) as recommended by the manufacturer. Competent E. coli BL21(DE3) cells for transformation were prepared by the calcium chloride method (3). Restriction endonucleases, T4 polynucleotide kinase, and pGEM-7Zf(+) were obtained from Promega, MunI was obtained from Boehringer Mannheim (Indianapolis, Ind.), and DNase I was obtained from Worthington Co. (Freehold, N.J.).

PCR techniques. The GeneAmp core reagent kit (Perkin-Elmer Cetus Co., Norwalk, Conn.) was used according to the manufacturer's specifications for all PCR amplifications, which were performed on a Perkin-Elmer Cetus GeneAmp PCR System 2400 thermal cycler with plasmid pGAT400 (12) as the template. Two primers, arot5B (5'-GGTGTAGTGGGATCCAGGTTCTTTCTTAGTGCC-3') and arot3B (5'-CGCGGATCCGACAGAAGTGGTTGTTTC-3'), were used to introduce BamHI restriction endonuclease sites at both ends of the oriT region cloned into the pGEM-7Zf(+) (Fig. 1) to give pJA-7ZT. Primer arot5B is complementary to nucleotides 334 to 351 upstream of the inverted repeats, and primer arot3B is complementary to nucleotides 621 through 638 immediately downstream of the inverted repeats (12). Similarly, primers arba5N (5'-CGCGGATCCTATGGATAAAGAAACAACAACC-3') and arba3B (5'-CGATGGATCCTACCGCCTCCCTGTATCTTAC-3') were used to introduce a BamHI restriction endonuclease site at the 5' and 3' ends of the B. fragilis btgA gene for cloning in pET-19b/BamHI to give pDHBA. All PCR products were electrophoresed through 1.3% agarose gels. The appropriate bands (646 bp for btgA and 304 bp for the oriT region) were excised and purified by using Geneclean II (Bio 101) prior to restriction endonuclease treatment and cloning.

View larger version (8K): [in this window] [in a new window]   FIG. 1.   Schematic representation of the pBFTM10 transfer region. Base pair numbers correspond to those used by Hecht et al. (12). Open rectangles represent the positions of the DNA oligonucleotides used for PCR amplification of the oriT region and the btgA gene: a, arba5N; b, arba3B; c, arot5B; and d, arot3B. The expanded view of the oriT region illustrates the positions of IRI, IRII, and IRIII relative to the start codon (ATG) of the btgA gene.

DNA sequence determination. Sequencing was performed by the Sanger dideoxy-chain termination method, using a Sequenase version 2.0 reagent kit as specified by the manufacturer (Amersham Co., Arlington Heights, Ill.).

Protein expression and purification. The BtgA protein was purified by an affinity binding procedure using His-Tag resin as specified by the manufacturer (Novagen). In brief, an overnight culture of E. coli BL21(DE3) carrying pDHBA was diluted 10-fold in fresh LB with ampicillin at a final concentration of 200 µg/ml and incubated at 37°C. Isopropyl--D-thiogalactopyranoside (IPTG) was then added to a final concentration of 1.0 mM, and cells were grown at 37°C with aeration for another 2 h before harvesting. Harvested cells were kept on ice and lysed by sonic disruption in a binding buffer containing 5 mM imidazole, 0.5 M NaCl, and 20 mM Tris-HCl (pH 7.9). Cell debris was removed by centrifugation at 4°C at 10,000 × g and cell extracts were incubated with His-Tag resin for 30 min at 4°C with gentle rotation.

Protein sequencing. An overnight culture of E. coli BL21(DE3) carrying pDHBA was grown under the conditions described above for protein expression. After IPTG induction, final harvesting, and resuspension in water, 100 µl of whole cells was mixed with 100 µl of 2× SDS-PAGE sample buffer (15) and boiled for 5 min. An aliquot was then applied directly to an SDS-10% polyacrylamide gel and electrophoresed at 100 V for 1 h. Proteins from the gel were transferred to a ProBlott membrane according to the manufacturer's instructions. After Ponceau-S (Sigma) protein staining, the BtgA protein band was excised and sequenced. The first 24 amino-terminal amino acids were sequenced with an ABI 477A protein sequencer (Applied Biosystems Co., Foster, Calif.) with an on-line ABI 120A phenylthiohydantoin analyzer. This analysis confirmed the correct fusion between the vector and BtgA protein.

DNA labeling and DNA mobility shift assay. The 304-bp oriT fragment was obtained by excision from pGEM-7ZT via BamHI restriction or by PCR. The resulting 304-bp oriT fragment was isolated by electrophoresis through a 1.3% agarose gel, purified by using Geneclean II, and labeled for footprinting assays using T4 DNA polynucleotide kinase as described by Sambrook et al. (29), with some modifications. When BamHI digestion was used, a dephosphorylation step was included. The labeling mix contained the 5' terminus of dephosphorylated DNA (1 to 50 pmol), 10× bacteriophage T4 polynucleotide kinase buffer (5 µl), [-³²P]ATP (3,000 Ci/mmol; 10 mCi/ml; 5 µl; Amersham Co., Arlington Heights, Ill.), bacteriophage T4 polynucleotide kinase (10 U), and water to 50 µl. If a higher efficiency of labeling was required, the reaction was performed in a volume of 20 µl, using 10 µCi of [-³²P]ATP (3,000 Ci/mmol; 10 mCi/ml). After incubation of this mixture for 30 min at 37°C, the enzyme was heat inactivated at 95°C for 5 min (29). Labeled fragments were then subjected to restriction endonuclease cleavage (with MunI, to obtain a 271-bp double-stranded DNA fragment specifically labeled on the upper strand, and with DdeI, to obtain a 295-bp double-stranded DNA fragment specifically labeled on the lower strand) overnight at 37°C. Labeled fragments were purified by electrophoresis through a 5% acrylamide-0.5× Tris-borate-EDTA gel. The purified DNA band was visualized by autoradiography, excised, eluted overnight with water, and used in DNase I footprinting assays.

DNase I footprinting. DNase I footprinting analysis was performed as described by Galas and Schmitz (8). Briefly, 10 ng of the oriT fragment (specifically labeled on the upper or lower strand) was incubated for 30 min at 4°C with various concentrations of purified BtgA protein in 1× binding buffer used for the gel shift assay described above, in a total reaction mixture volume of 40 µl. An equal volume of the mixture (40 µl) containing 2 mM CaCl₂ and 15 mM MgCl₂ was added, and the cleavage reaction was carried out in a total volume of 80 µl by adding DNase I to 2 ng. After 1 min of incubation at room temperature, DNase I was inactivated by adding an equal volume (80 µl) of stop solution containing 10 mM EDTA, 0.5% SDS, 100 mM NaCl, and 75 mg of tRNA per ml. Following phenol extraction and ethanol precipitation, samples were washed with 70% ethanol, dissolved in formamide dye solution, electrophoresed on an 8% polyacrylamide sequencing gel (28), and visualized by autoradiography. To resolve longer nucleotide sequences, increased electrophoresis time and Long Ranger acrylamide (FMC Bioproducts, Rockland, Maine) were used.

Digital imaging and data presentation. Autoradiographs and Coomassie blue-stained gels were scanned with a ScanMaker III (Microtek, Palm Bay, Fla.) or Hewlett-Packard ScanJet 4C (Hewlett-Packard, Louisville, Ky.) flatbed scanner at high resolution (1,000 dots/inch), using the software application Adobe Photoshop version 3.0.5 (Adobe Systems Inc., Buffalo, N.Y.). Scanned images were then transferred to the application Microsoft Powerpoint 4.0, text and labels were added, and images were printed at high resolution (1,440 dots/inch) on an Epson Stylus Color 1520 printer (Epson America Inc., Torrance, Calif.). Hewlett-Packard satin-gloss heavy-weight photographic paper was used for all prints.

Nucleotide sequence accession number. The GenBank accession number for the pBFTM10 mobilization region is M77806.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Cloning the oriT region of pBFTM10. Previous DNA sequence analysis of the pBFTM10 mobilization region in pGAT400 revealed three sets of inverted repeats and a possible nic site, based on consensus sequence immediately upstream of btgA (12). In addition, preliminary DNA relaxosome isolation studies also demonstrated that single-stranded DNA nicking occurred in or near this region (unpublished data). To determine if this area served as a functional oriT and a likely binding region for btgA, a 304-bp fragment encompassing the three inverted repeats and the putative nic site was initially amplified by PCR using plasmid pGAT400 as a template. Figure 1 illustrates the transfer region of pBFTM10 and the locations of the PCR primers (see Materials and Methods). Following cloning into the pGEM-7Zf(+) BamHI site, DNA sequence analysis confirmed that the amplified oriT region was properly cloned without any PCR-induced errors. The 304-bp oriT fragment was then excised, cloned again into the unique BamHI site of pACYC184 to form pJAYC1, and used in bacterial conjugation experiments.

Demonstration of oriT function by bacterial conjugation. Previously, pGAT400 was shown to mobilize between E. coli mating pairs when coresident with R751. pGAT400 contains btgA and btgB, both necessary for its own transfer, while R751 is presumed to provide the machinery necessary for transfer between E. coli cells (12, 17, 19, 30). To determine if the cloned putative oriT region was functional, pJAYC1 was transformed into E. coli HB101 containing pGAT400 and R751, or R751 alone, and used as a donor in bacterial conjugation experiments. HB101 cells containing either pGAT400, R751, pJAYC1, or pACYC184 were used as negative controls.

2

2

2

Cloning and expression of BtgA. The predicted amino acid sequence of BtgA revealed a helix-turn-helix DNA binding motif (amino acids 118 to 137) (12). This protein, therefore, seemed a likely candidate for binding to the oriT region. To test this possibility, btgA was cloned into the pET-19b expression vector to give pDHBA and overexpressed in E. coli BL21(DE3). The positions of primers arba5N and arba3B used to amplify btgA by PCR are illustrated in Fig. 1. DNA sequencing of the entire 630-bp insert and part of the pET-19b promoter sequence confirmed that the cloned gene contained no sequence errors. Following IPTG induction of pDHBA, an approximately 27-kDa band was visualized (Fig. 2). The predicted mass of BtgA, as determined from the DNA sequence, was 23.2 kDa, or 26.1 kDa with the leader sequence from the pET-19b expression vector included. To confirm that this band represented expressed BtgA, the first 24 amino acids of the amino terminus of the 26.1-kDa protein band were sequenced, revealing the correct predicted fusion protein.

View larger version (95K): [in this window] [in a new window]   FIG. 2.   Overproduction and purification of BtgA. Lane 1, Promega mid-range protein molecular mass marker (masses in kilodaltons are shown at the left); lane 2, proteins that did not bind to the His-Tag binding resin following the addition of crude extract; lane 3, BtgA protein eluted from the His-Tag column following multistep dialysis (see Materials and Methods).

Gel mobility shift assay. The BtgA-oriT complex was analyzed in an electrophoretic mobility shift assay to demonstrate noncovalent interaction of BtgA with the oriT (1, 8). Purified BtgA protein ranging from 10 to 300 ng and labeled 304-bp pBFTM10 oriT fragment were incubated with a binding buffer and electrophoresed in 5% polyacrylamide gels at 4°C (Fig. 3; see Materials and Methods). Increasing concentrations of BtgA (10, 50, and 100 ng [Fig. 3, lanes 2 to 4, respectively]) demonstrated a concentration-dependent manner of binding with oriT. At 300 ng (lane 5), BtgA demonstrated the maximum shift of oriT DNA (position A). The presence of poly(dI-dC) at 1 and 3 µg failed to disrupt the interaction between the oriT region and BtgA (lanes 6 and 7, respectively). To demonstrate the specificity of binding, specific competitor pJA-7ZT (containing pBFTM10 oriT) plasmid DNA was added in a 50-fold excess, resulting in partial inhibition of BtgA binding, while 100- and 300-fold excess competitor completely inhibited the binding reaction (lanes 8 to 10). Of note, storage of BtgA at 80°C resulted in some loss of binding activity after a few days, likely the result of precipitation.

View larger version (42K): [in this window] [in a new window]   FIG. 3.   Binding of purified BtgA protein to the pBFTM10 oriT. Ten nanograms of 304-bp BamHI pBFTM10 oriT fragment labeled at the 5' terminus with [-32P]ATP was incubated with purified BtgA (see Materials and Methods). Lane 1, labeled oriT fragment without addition of protein; lanes 2 to 5, labeled oriT fragment with addition of 10, 50, 100, and 300 ng of purified BtgA protein; lanes 6 and 7, oriT fragment incubated with 300 ng of BtgA in the presence of 1 and 3 µg of poly(dI-dC), respectively; lanes 8 to 10, labeled oriT fragment incubated with 300 ng of BtgA protein plus 50-, 100-, and 300-fold excess cold competitor oriT-containing plasmid pGEM-7ZT, respectively.

DNase I footprinting. DNase I protection experiments were used to localize the precise binding sites of BtgA on the pBFTM10 oriT. The oriT fragment was specifically labeled on the upper or lower strand with [-³²P]ATP and then incubated in the presence of increasing amounts of purified BtgA. These DNA-protein complexes were subjected to partial DNase I digestion to define the binding site(s) of BtgA. The resulting products were separated on 8% polyacrylamide sequencing gels and visualized by autoradiography (Fig. 4). Three distinct regions of protection were visualized on the lower strand. Two regions, consisting of 36 nucleotides (462 to 498 [Fig. 4A]) and 35 nucleotides (501 to 535 [Fig. 4A]), demonstrated the strongest protection after the addition of 0.5 µg of BtgA, while a third region, containing three identifiable nucleotides (536 to 558 [Fig. 4A]), demonstrated weaker protection at higher protein concentrations. In addition, 3 nucleotides, 498 to 500, showed weaker protection and were located between the strongly protected regions. The first protected region overlaps the right half of IRI (which contains a consensus sequence for the nick site) and the left half of IRII (Fig. 4 and 5), while the second protected region overlaps the right half of IRII and the left half of IRIII. The third, weaker region of protection overlaps 10 nucleotides of the right half of IRIII. Bands at 458 to 461 and 498 to 500 demonstrated possible minimal protection. No regions of protection were found in the remaining sequence. We detected no regions of protection on the upper strand that would correspond to the protected zones on the lower strand (Fig. 4B).

View larger version (75K): [in this window] [in a new window]   FIG. 4.   DNase I footprint analysis of the BtgA-pBFTM10 oriT protein complex. The specifically labeled upper and lower strands of oriT were incubated with increasing amounts of purified BtgA protein in binding reactions. The BtgA-oriT complexes were subjected to limited cleavage by DNase I and electrophoresed on 8% polyacrylamide sequencing gels. (A) Cleavage pattern of the lower strand of oriT. Quantities of purified BtgA in reaction mixtures: lane C, none; lanes 1 to 5, 100 ng, 300 ng, 500 ng, 1 µg, and 3 µg, respectively. Locations of the inverted repeats are shown by arrows. L and R designate the left and right arms, respectively, of the repeats. For IRI, only the right arm of the repeat is shown since the region corresponding to the left arm is undetectable under the conditions of the footprint. The locations of DNase I protection regions are indicated at the right by brackets. Asterisks represent the region of weak protection (see Results and Discussion). Nucleotide numbering at the left is from the labeled end and corresponds to the published sequence (12). The inset shows the DNase I digestion pattern up to the top of the gel, indicating that no regions of protection were seen, and also demonstrates that the extents of DNase I digestion were similar for all reactions. (B) Cleavage pattern of the upper strand of oriT. Quantities of BtgA in reaction mixtures: lanes 1 to 6, 100 ng, 300 ng, 500 ng, 1 µg, 2 µg, and 3 µg, respectively; lane C, none.

View larger version (41K): [in this window] [in a new window]   FIG. 5.   Double-stranded nucleotide sequence of the oriT region of pBFTM10. The dark gray boxed regions on the lower strand indicate regions of BtgA binding, as determined by DNase I protection assays. The light gray box indicates the region of weak binding. The location of a putative nic site based on the consensus sequence (12, 39) is shown, as are the left (L) and right (R) arms of the three inverted repeats and the start codon for btgA.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

The conjugative mechanisms required for transfer of plasmids in Bacteroides spp. are not well understood but are presumed to include initiation processes, similar to those of RP4 or the F plasmid (6, 9, 27, 44). Thus, it is presumed that relaxosomes are formed prior to transfer of a single strand of DNA. For RP4, this requires an oriT region, two DNA binding proteins, a stabilizing protein, and a nickase (16, 43). pBFTM10 has only two transfer genes, both of which are necessary for DNA transfer.

To further our understanding of the processes required for mobilization of plasmid DNA in Bacteroides spp., we first determined the location of the oriT region by testing for mobilization in E. coli. A 304-bp fragment that contained the three sets of inverted repeats identified previously as likely targets for protein binding was chosen. The putative promoter of btgA is immediately downstream of IRIII and was also included in the cloned fragment. Mobilization of the oriT region in E. coli was demonstrated only when both btgA and btgB were provided in trans, although R751 must also be coresident. It is presumed that the role of R751 is to provide a transfer apparatus for mobilization of Bacteroides transfer factors, including pGAT400, pB8-51, pLV22a, and the conjugal transposon Tn4399, in E. coli (12, 13, 21, 25, 32). However, it was previously unknown if cotransfer of R751 with these transfer factors was required. Normally R751 transfers at levels 1.4 orders of magnitude greater than those for the coresident transfer factor and has thus been found in transconjugants that received the mobilized transfer factor. Indeed, in these studies, R751 was efficiently self-transferred at levels 1.4 orders of magnitude greater than those for the mobilized pGAT400 or pJAYC1 (12, 30). As expected, we observed that 100% of pGAT400 transconjugants also contained R751. However, only 85% of pJAYC1 transconjugants also contained R751. Thus, it appears that cotransfer of R751 is not necessary for pJAYC1 mobilization, although it is not known why a difference between R751 cotransfer with pGAT400 and pJAYC1 was seen. This finding implies that R751 indeed may simply provide a transfer apparatus utilized by pBFTM10 and other transfer factors.

The function of one of the mobilization proteins, BtgA, was tested after cloning, expression, and purification in different assays with the pBFTM10 oriT. The presence of a helix-turn-helix motif suggested that BtgA may be a DNA binding protein, and we tested whether it could bind to the pBFTM10 oriT. Electrophoretic mobility shift assays demonstrated significant specific binding of BtgA to the 304-bp oriT-containing fragment. Poly(dI-dC) in concentrations of up to 3 µg did not affect the mobility shift results. The addition of unlabeled oriT fragment contained in plasmid pJAYC1 was an effective competition for this binding reaction, confirming its specificity (Fig. 3). However, intermediate mobility shifts were seen at protein concentrations ranging from 10 to 100 ng (Fig. 3). These findings could represent the oligomerization of BtgA either before or during binding. Alternatively, this could represent a concentration-dependent increase in binding to additional regions within the oriT fragment, as seen for TraK from RP4 (46), and may be supported by the results of DNase I footprinting experiments (see below). Assays to determine if oligomers are involved in binding should resolve this observation.

DNase I protection assays localized BtgA binding to three regions on the lower strand of the pBFTM10 oriT (Fig. 5). The most prominent zones of protection were visualized at nucleotides 462 to 497 and 501 to 535. Taken together, these two regions span 71 nucleotides that cover parts (or all) of the three inverted repeats within the oriT region. The first region of protection toward the 5' terminus of the oriT region includes the right half of IRI, which also contains the consensus nick site sequence -TTCCTCTTG/C- (39). Results of preliminary experiments using labeled double-stranded linear oriT DNA combined with BtgA support this observation (unpublished data). A weaker zone of protection was visualized on nucleotides 536 to 558, which was visualized only with higher concentrations of protein. No large regions of protection were visualized on the upper strand of the pBFTM10 oriT, although individual bases could be protected in more compressed areas but not seen. This apparent asymmetric binding of BtgA protein to oriT could be due to a conformational change in the DNA as a result of wrapping or bending during BtgA-oriT interaction. This is not unusual, having been previously observed for single-stranded DNA binding proteins involved in replication (protection over 70 nucleotides on one strand only [5]) and the E. coli Fpg zinc finger protein (tested by using high-resolution hydroxyl radical footprinting analysis [35]), as well as several other bacterial and eukaryotic DNA binding proteins (7, 14, 18, 36, 41, 42, 47).

The location of BtgA binding strongly supports its role in DNA processing functions necessary for transfer initiation. Binding to recognition sequences within the oriT region is necessary for efficient transfer in RP4 and other plasmids. For RP4, TraJ binds to the right arm of a 19-nucleotide inverted repeat recognizing a 10-bp sequence within 8 nucleotides of the downstream nick site (45). TraJ binding to supercoiled DNA is required for binding and nicking activity of the relaxase TraI, presumably by conformational change of the nick site. TraK is not required for RP4 transfer but increases the efficiency of relaxosome formation. It binds downstream of the nick site over an approximate 180-nucleotide region, which requires as many as 15 to 20 monomers of TraK to form the complex resulting in bending of DNA (46). The precise role of BtgA in DNA processing has not been defined, but it is required for transfer of pBFTM10 (12). The relatively large region of protection overlapping the inverted repeats, and possible nick site, raises the possibility that BtgA is multifunctional.