INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The traM and traY genes of IncF conjugation systems are essential for transfer proficiency (19, 24, 25, 35). Numerous studies have established a role for traM and traY in regulation of transfer gene expression (6, 7, 34, 35, 41). Thus the transfer-deficient phenotype of traY and traM mutant derivatives is certain to reflect disruption of positive gene regulation as a minimum and may further reflect the loss of additional functions.

The TraM and TraY proteins of different IncF plasmids exhibit sequence-specific DNA binding activity at oriT (1, 8, 18, 21, 23, 30, 38, 39), and they have been implicated in the initiation stage of conjugative DNA transfer. During initiation of conjugative plasmid transfer an enzyme known as relaxase cleaves a defined strand of oriT DNA at a specific position (nic) in a transesterification reaction. Cleaving activity at oriT is exhibited as part of a nucleoprotein complex called the relaxosome. This complex includes the relaxase and auxiliary proteins, which can be host or plasmid encoded (12). These accessory factors impart specificity and stability to the complex and enhance the efficiency of the cleavage reaction. In many cases specific interactions between the auxiliary proteins and oriT DNA promote localized melting of the duplex and facilitate access of relaxase to its recognition site (31-33, 37, 40, 42, 46-49).

For the IncF system, extensive biochemical analysis has been dedicated to characterizing the relaxase-catalyzed cleavage reaction of oriT substrates in vitro. These studies have shown that the TraY proteins of IncF plasmids and the Escherichia coli histone-like integration host factor (IHF) promote the association of the TraI relaxase with oriT DNA (16) and enhance the cleaving activity of TraI in vitro (17, 28). An early study by Everett and Willetts using plasmid F additionally demonstrated the involvement of traI and traY in nic cleavage in vivo (10). We have been analyzing the composition and performance of the relaxosome of IncF plasmid R1 in vivo. A study to explore a possible role for oriT DNA binding protein TraM in the R1 relaxosome revealed an unexpected requirement for TraM protein to observe TraI-catalyzed strand scission in IHF-proficient E. coli in the absence of other transfer proteins (20). Thus, the R1 TraY protein was dispensable for cleavage on recombinant oriT substrates in vivo. Also contrary to expectations based on the previous biochemical and genetic analyses, TraY was dispensable for DNA processing occurring on recombinant mobilizable plasmids during bacterial conjugation (20). Notably, mobilization of the R1 oriT did not occur in the absence of both traM and traY.

The present study addresses the contribution of the TraY protein to the activity of the R1 relaxosome in vivo. Additionally, mobilization of R1 oriT by a heterologous transfer system was used to provide evidence for distinct functions for the proteins TraM and TraY during conjugative DNA strand transfer.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains and media. The E. coli strains used in this study are K-12 derivatives. Cells were grown in 2× tryptone-yeast extract (TY) (26). Antibiotics were used to select for plasmid-carrying strains in the following final concentrations: for R1-16 and pOX38-Km, kanamycin at 40 µg ml¹; for pUC-, pMMB67-, and pMS119-based constructions, dihydroampicillin (epicillin) at 100 µg ml¹; for pBR322 derivatives, tetracycline at 15 µg ml¹; for pGZ119 constructions, chloramphenicol at 10 µg ml¹.

Enzymes and reagents. Restriction endonucleases, calf intestinal phosphatase, and T4 DNA ligase were purchased from Boehringer Mannheim. DyNAzyme was obtained from Finnzymes (Espoo, Finland), and radiochemicals were from NEN.

Preparation of a traY expression plasmid for in vivo analyses. The traY gene (positions 3683 to 3985; numbering according to reference 14) was amplified from E. coli MC1061 (R1-16) by PCR using primers UYE (5'-CCGGAATTCTGTGCAATCATG) and DYB (5'-GGGGATCCTCTGTTTAATATTG). These oligonucleotides contained nonhomologous EcoR1 and BamHI linkers to facilitate ligation of the PCR product to the pBluescript vector (Stratagene). DNA sequence analysis confirmed that the resulting recombinant DNA encoded the wild-type TraY protein. The traY gene fragment was excised using EcoR1 and BamHI and introduced to expression vector pGZ119EH (22) to create pGZYM1.

Preparation of the standard DNA template. Oligonucleotide 8* (5'-AATTGGATGTTAGCCATCTGCCTGAGCT-3') was complementary to primer 8 (20) and had additionally 4 nucleotides (nt) at the 5' and 3' ends compatible with the EcoRI/SacI-digested vector. Annealed oligonucleotides 8* and 8 were ligated to linear pBluescript DNA, and transformed XL1 (Stratagene) cells were selected with dihydroampicillin and X-Gal (5-bromo-4-chloro-3-indolyl--D-galactopyranoside). A positive clone, confirmed by DNA sequencing, was purified, linearized, and reisolated for use in the cleavage assay.

Runoff DNA synthesis assay. E. coli strains harboring plasmids were grown at 37°C to stationary phase and then diluted in fresh medium containing antibiotics and 0.1 mM IPTG (isopropyl--D-thiogalactopyranoside) as indicated in the legends to Fig. 1 and 2. Incubation at 37°C was continued with shaking for 2.5 h. Cells were collected by centrifugation after making adjustments for optical density. The medium was thoroughly removed. Cell pellets were kept at 0°C and used immediately. For the reaction mixture, bacteria were resuspended in 25 µl of ice-cold buffer containing 10 mM Tris-HCl (pH 8.8) at 25°C, 1.5 mM MgCl₂, 50 mM KCl, 0.1% Triton X-100, 100 ng of oligonucleotide, 100 µM deoxynucleoside triphosphates, 5 µCi of [-³²P]dATP (3,000 Ci/mmol), 1.0 ng of standard DNA template, and 2 U of DyNAzyme. Viable-cell counts were obtained at harvest. Primer 8 was used as described previously (20) except that the in vitro DNA synthesis was stopped by the addition of 8 µl of formamide loading dye. The reaction products were applied immediately to denaturing 7% polyacrylamide (19:1 polyacrylamide/bisacrylamide ratio) Tris-borate-EDTA gels and resolved with oriT- and primer 8-specific polynucleotide size markers as described previously (20). Following electrophoresis radioactive products were visualized by autoradiography.

Mobilization assays. E. coli J5 donor strains carrying a mobilizable test plasmid (20), transfer-proficient pOX38-Km (4), and pGZYM1 to provide R1 traY in trans were cultured overnight in 2× TY medium with antibiotic selection for all plasmids. Conjugation was carried out as described previously (20). Transconjugants carrying the conjugative plasmid were selected on MacConkey agar plates containing 40 µg of kanamycin/ml and 25 µg of streptomycin/ml. Transconjugants harboring the mobilized plasmid were selected using 100 µg of dihydroampicillin/ml and 25 µg of streptomycin/ml. Viable counts for donors were determined with chloramphenicol selection. The conjugation frequency was expressed as the number of Km^r Sm^r transconjugants per donor cell. The mobilization frequency was expressed as the number of Ep^r Sm^r transconjugants per donor cell.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

The R1 TraY protein exhibits DNA binding activity. The unexpected ability of the TraM protein to perform functions anticipated for the TraY protein in the R1 system raised the question of whether the TraY protein from plasmid R1 has DNA binding activity. To clarify this point, the TraY protein of R1 was purified as a fusion protein with glutathione S-transferase (GST) according to the procedure of Nelson and Matson (29). Electrophoretic mobility shift assays (EMSA) were performed with the purified GST-TraY protein and various DNA fragments from the R1 oriT (data not shown). The region between positions 1865 and 2457 of the R1 sequence (14) was amplified by PCR and isolated or cleaved into multiple subfragments with either NdeI or DraI and then isolated for use as DNA ligands. EMSA with combinations of overlapping subfragments demonstrated a specific interaction between the TraY protein and oriT DNA between nt 2100 and 2158, in good agreement with the position of sbyA in the similarly structured oriT regions of F and R100 (18, 30).

View larger version (42K): [in this window] [in a new window]   FIG. 1.   Runoff DNA synthesis measures intracellular TraI-catalyzed cleavage activity. (A) The tra genes on plasmids maintained in vivo are illustrated schematically. The 1.2-kb BglII-PstI fragment of R1 oriT-traM DNA contained in pBR322-based substrate plasmid pBR111 is shown (top). In the second plasmid, pHP2, a 6.1-kb AsnI fragment carrying the R1 traI gene is placed under Ptac control in expression vector pGZ119EH (bottom) (B) oriT primer 8 and its complementary sequence were annealed, and a single copy of the hybrid was introduced into an unrelated vector. A KpnI fragment from this clone that contained the primer sequence was isolated and added exogenously to the cells in the cleavage assay. Reaction mixtures thus contained three primed templates for in vitro DNA synthesis: cleaved (top) and uncleaved (middle) oriT plasmid DNA released from bacterial cells and the exogenously added linear template (bottom). (C) E. coli AG1 harboring pBR111 and pHP2 was harvested after overnight culture without IPTG (lane 1) or after subculture in fresh medium without (lanes 2 and 3) or with IPTG (lanes 4 to 6). In the cleavage assay equivalent cell masses (as indicated by optical densities at 600 nm) were present in all reaction mixtures in addition to 1 ng of purified KpnI fragment. Reaction products synthesized on the different templates can be readily distinguished according to size on denaturing polyacrylamide gels (nic and std). Dideoxynucleotide-terminated DNA sequence ladders (ddATP and ddCTP) generated on oriT DNA with primer 8 were used to determine polynucleotide chain length (lanes A and C).

View larger version (54K): [in this window] [in a new window]   FIG. 2.   TraY stimulates TraI-catalyzed cleavage at oriT in vivo. Overnight cultures of E. coli AG1 strains carrying plasmids R1-16 (lane 1), pBR111M0, pGNKtraI, and pGZ119EH (lanes 2 and 3), pBR111M0, pGNKtraI, and pGZYM1 (lanes 4 and 5), or pBR111, pGNKtraI, and pGZYM1 (lanes 6 and 7) to provide the indicated combinations of proteins were subcultured in fresh medium with antibiotics and IPTG. The numbers of viable cells present in the reaction mixtures resolved in lanes 1 to 7 were 2.0 × 106, 1.4 × 106, 2.8 × 106, 0.34 × 106, 0.7 × 106, 0.8 × 106, and 1.7 × 106 CFU, respectively.

TraY promotes the cleaving activity of relaxase. To evaluate the role of TraY in the relaxosome, pGZYM1 was constructed with the R1 traY gene under the control of the P_tac promoter in pGZ119EH (22). pGZYM1 or vector DNA was introduced into E. coli AG1 carrying additionally an oriT-traM cleavage substrate containing wild-type traM (pBR111) or null traM (pBR111M0) and pGNKtraI (20), where relaxase expression is also regulated by P_tac. The various strains, each carrying three plasmids to achieve the desired combination of proteins (Fig. 2), were grown in the presence of IPTG. Cells were harvested at equivalent optical densities, and increasing numbers were assayed. Reaction products terminated at nic were observed from the control AG1 (R1-16), which carries all tra genes (lane 1), and were observed when TraY (lanes 4 and 5) or TraY and TraM (lanes 6 and 7) were present in addition to relaxase but not when relaxase alone was expressed in the IHF proficient wild-type strain (lanes 2 and 3). Conversely, no intracellular cleavage was detected on pBR111 in the presence of relaxase, TraM, and TraY when the appropriate constructions were maintained in IHF-deficient strain K5302 himA Smal::TnKm^r (13) (data not shown). These results demonstrate that TraY is a component of the R1 relaxosome and that TraY and TraM can act independently of each other in a manner that is sufficient to stimulate TraI-catalyzed cleaving activity in vivo when E. coli IHF is also present.

Mobilization of R1 oriT DNA requires traM in the presence and absence of traY. To investigate whether TraM and TraY impart equivalent functions to the relaxosome during conjugative transfer, the contribution of each to mobilization of the R1 oriT was assessed. Taking advantage of the plasmid specificity of TraY and TraM among IncF plasmids (10, 25, 43) we demonstrated earlier that mobilization of the R1 oriT by F-plasmid proteins did not occur in the absence of both TraM and TraY of R1 (20). Very efficient transfer was observed with TraM alone, however. If the essential role of traM during mobilization is solely to facilitate DNA processing at oriT, then the cleavage data imply that TraY should perform that function equally well. To test this, E. coli J5 carrying self-transmissible F derivative pOX38-Km and an R1 oriT test plasmid was transformed with traY expression construction pGZYM1 or a vector. In accordance with previous results (20), when the mobilization substrate carried the 1.2-kb BglII-PstI oriT fragment including the wild-type traM sequence, pMM-Mwt, mobilization of that plasmid occurred as efficiently as self-transfer by pOX38 (Table 1, row 1 [from the top]). When the oriT plasmid carried the traM null allele, pMM-M0, a 1,000-fold diminution in mobilization frequency compared to self-transmission by pOX38-Km was observed (Table 1, row 3). The additional presence of the R1 traY gene in the donor strain did not increase the efficiency of mobilization of the R1 oriT (Table 1, compare rows 3 and 4). Overexpression of traY through incubation of the donor strains in 0.1 mM IPTG for 1 h prior to the initiation of conjugation did not affect the frequency of mobilization (data not shown). Thus, in contrast to the requirements observed for oriT cleavage in vivo, TraM remained essential for mobilization of the R1 oriT even in the presence of TraY.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The failure of traY to compensate for the absence of traM during transfer suggests two possible explanations. One is that that the requirements for relaxase to efficiently catalyze origin cleaving during bacterial conjugation are different from those in the absence of conjugation (the conditions applied in the cleavage assays shown in Fig. 1 and 2 and in reactions reconstituted in vitro with purified proteins). We find this hypothesis improbable, but formally it cannot be excluded.

A more likely explanation is that the cleaving reaction on the mobilization substrate was actually catalyzed by the complex of TraI, TraY, and IHF but that a successful transfer of this substrate was prevented at a later stage of the process. In that case, the crucial step would require the TraM protein, either as part of the relaxosome or independent of that complex. We attempted to test directly the first part of this hypothesis by assaying for cleavage of the mobilization plasmids catalyzed during conjugation. Cleavage could not be reproducibly observed on either pMM-Mwt or pMM-M0 (data not shown). A number of positive controls indicated that the assay may be impaired by the presence of both the highly homologous F and R1 origin regions in the reaction. Reiterated cycles of DNA synthesis in the cleavage assay result in an enhancement of the nic-specific product when the majority of molecules are cleaved in vivo. When this is not the case, it is expected that nic-specific products generated in early cycles would be lost in subsequent cycles if they annealed to uncleaved molecules of pOX38-Km or a mobilizable plasmid and were elongated beyond nic. The frequency of pOX38-Km transfer is usually reduced at least 10-fold when additional plasmids are carried by the host (Table 1). If a low frequency of intracellular cleavage accompanies the low transfer frequency, then we expect that strand scission at the nic sites of pOX38-Km and the mobilizable plasmids would not be reliably detected with this assay. Resolution of this question, therefore, will require a different approach.

The second aspect of this hypothesis, namely, that TraM additionally contributes an essential function to the transfer process at a step distinct from the initial strand cleavage reaction at oriT, is intriguing. This function for TraM would be required in addition to its essential role in gene regulation (reviewed in reference 44). The nature of this activity for TraM is unknown but may involve an interaction with TraG-like protein TraD (9). Current models propose that TraG-like proteins perform their essential role in conjugation by physically linking the DNA substrate destined for transfer to the transport machinery that delivers the DNA to the recipient cell. The interface between the protein and plasmid is apparently an interaction between TraG and one or more proteins of the relaxosome (2, 3, 5, 15, 27, 36). Specific contacts at this stage are thought to determine how efficiently a relaxosome gains access to the transport complex. The present report demonstrates that recombinant R1 oriT-traM null molecules are not transferred despite the presence of the cognate TraY protein. It is conceivable that these plasmids transfer poorly because TraM is not available to efficiently couple the relaxosome to TraD.

The contribution of TraY to the R1 relaxosome remains poorly defined. Biochemical analysis of the F-plasmid proteins (16, 28) and recent genetic data (11) suggest that one function of TraY is to impart stability to the complex. Further work is necessary for a detailed understanding of the activity and regulation of IncF relaxosomes.