Extent of single-stranded DNA required for efficient TraI helicase activity in vitro.

The IncF plasmid protein TraI functions during bacterial conjugation as a site- and strand-specific DNA transesterase and a highly processive 5' to 3' DNA helicase. The N-terminal DNA transesterase domain of TraI localizes the protein to nic and cleaves this site within the plasmid transfer origin. In the cell the C-terminal DNA helicase domain of TraI is essential for driving the 5' to 3' unwinding of plasmid DNA from nic to provide the strand destined for transfer. In vitro, however, purified TraI protein cannot enter and unwind nicked plasmid DNA and instead requires a 5' tail of single-stranded DNA at the duplex junction. In this study we evaluate the extent of single-stranded DNA adjacent to the duplex that is required for efficient TraI-catalyzed DNA unwinding in vitro. A series of linear partial duplex DNA substrates containing a central stretch of single-stranded DNA of defined length was created and its structure verified. We found that substrates containing >or=27 nucleotides of single-stranded DNA 5' to the duplex were unwound efficiently by TraI, whereas substrates containing 20 or fewer nucleotides were not. These results imply that during conjugation localized unwinding of >20 nucleotides at nic is necessary to initiate unwinding of plasmid DNA strands.

The IncF plasmid protein TraI functions during bacterial conjugation as a site-and strand-specific DNA transesterase and a highly processive 5 to 3 DNA helicase. The N-terminal DNA transesterase domain of TraI localizes the protein to nic and cleaves this site within the plasmid transfer origin. In the cell the C-terminal DNA helicase domain of TraI is essential for driving the 5 to 3 unwinding of plasmid DNA from nic to provide the strand destined for transfer. In vitro, however, purified TraI protein cannot enter and unwind nicked plasmid DNA and instead requires a 5 tail of singlestranded DNA at the duplex junction. In this study we evaluate the extent of single-stranded DNA adjacent to the duplex that is required for efficient TraI-catalyzed DNA unwinding in vitro. A series of linear partial duplex DNA substrates containing a central stretch of singlestranded DNA of defined length was created and its structure verified. We found that substrates containing >27 nucleotides of single-stranded DNA 5 to the duplex were unwound efficiently by TraI, whereas substrates containing 20 or fewer nucleotides were not. These results imply that during conjugation localized unwinding of >20 nucleotides at nic is necessary to initiate unwinding of plasmid DNA strands.
Helicases are ubiquitous enzymes involved in nucleic acid metabolism (for review, see Refs. [1][2][3][4]. The TraI protein of IncF plasmids (first characterized as DNA helicase I of Escherichia coli (5)) is essential for the transmission of bacterial genes during conjugation (6,7). In that process a copy of a conjugative plasmid is transferred unidirectionally in single-stranded form from one bacterial cell to another (for reviews, see Refs. 8 and 9). The TraI protein of IncF plasmids contributes to conjugative transfer in several ways. It is a component of a nucleoprotein complex, the relaxosome, which assembles with site specificity at the plasmid origin of transfer (oriT). Relaxosomes initiate the series of DNA-processing reactions that prepare plasmid DNA for transfer to a recipient cell. They are common to all self-transmissible and mobilizable plasmids in different degrees of complexity (for reviews, see Refs. 8, 10, and 11). The simplest systems employ a plasmid-encoded DNA transesterase, or relaxase, protein that acts on a specific phosphodiester bond, nic, in the transfer origin. Cleavage at this site provides a point of origin for the directed 5Ј to 3Ј transmission of the plasmid genome to a recipient cell. Relaxase proteins are also active in relaxosomes containing auxiliary DNA-binding proteins of host or plasmid origin. IncF, IncW, IncP, and IncQ systems offer well studied examples (12)(13)(14)(15)(16)(17)(18)(19)(20)(21). Among IncF plasmids factors known to stimulate nic cleavage include E. coli integration host factor and plasmid proteins TraM and TraY (22)(23)(24)(25)(26)(27). In the case of IncW plasmid R388, TrwA protein performs this role (13,14). The IncF system and the functionally analogous IncW-IncN family of DNA-mobilizing systems are (thus far) unique in that they additionally specify a DNA helicase activity essential for conjugative DNA transmission. The DNA helicase activities are located in C-terminal domains of the bifunctional F TraI and R388 TrwC proteins (28 -33). For reasons that are poorly understood, efficient gene transfer requires physical linkage of the relaxase and helicase domains (7,34).
This bifunctional arrangement of DNA transesterase and helicase activities is shared by adeno-associated virus replication initiation (Rep) 1 proteins (35). As revealed by very recent structural data, the endonuclease domain of adeno-associated virus-5 Rep proteins (36) bear the closest structural and functional relatedness to the N-terminal DNA transesterase domains of the F TraI and R388 TrwC proteins (37). 2 Adenoassociated virus-2 Rep78 protein and an alternative splicing product, Rep68, bind the viral replicative form origin of replication, cleave the target DNA in a site and strand-specific manner, and mediate vectorial unwinding of the DNA duplex via an ATP-dependent helicase activity, thus initiating a strand displacement mechanism of viral DNA replication. The N-terminal 224 amino acids of Rep78/68 (39) are involved in interaction with recognition sequences (resolution binding sites) in the viral terminal repeats and in the viral integration site on human chromosome 19 (40). This protein cleaves with site and strand specificity the terminal resolution site, trs, near resolution binding sites. The helicase activity of Rep78/68 has 3Ј to 5Ј polarity and is believed to assist the initiation of adeno-associated virus DNA replication at several stages including modifying the terminal repeat hairpin before nicking at trs, conversion of the terminal repeat region from the linear to hairpin form, and finally, possibly as a replicative helicase during synthesis of the viral genome (41).
The conjugative helicases initiate duplex unwinding unidirectionally from the plasmid transfer origin after strand-specific cleavage at nic by their N-terminal relaxase domains. The molecular mechanisms of these early steps of plasmid strand transfer remain poorly understood. Genetic and biochemical approaches have characterized the role of auxiliary factors in assisting the nic-specific cleavage reaction as well as the contribution of origin DNA sequence and topology (23)(24)(25)(26)(27)(42)(43)(44). Nonetheless, the product of that reaction, the open circular form, is not directly accessible to the duplex unwinding activity of the physically linked DNA helicase domain. Thus, although recruitment of the conjugative helicase to oriT is achieved by site-specific recognition of the N-terminal relaxase domain and/or via interaction with additional oriT binding auxiliary proteins, the reaction as reconstituted thus far fails to effectively load the helicase at nic (45). In the cell, then, the relaxosome probably acquires a higher order structure in a staged initiation process that triggers helicase activation, similar perhaps to the orchestrated unwinding of origin DNA during replication initiation in prokaryotic and eukaryotic cells (46).
Purified TraI and TrwC proteins do not exhibit sequence specificity for binding and translocating (5Ј to 3Ј) on ssDNA (29,47,48). Reconstitution of the duplex-unwinding activity in vitro requires a region of ssDNA adjacent to the duplex (49). In early work Hoffmann-Berling and co-workers (47) estimated that the length of ssDNA required by TraI on the duplex 5Ј end lies within an approximate range of 12-200 nt. At the time partial duplexes were generated by exonucleolytic erosion of flush termini, and technical limitations prevented a narrower definition of this requirement for TraI. To gain insights to the extent of localized duplex unwinding that may occur to start plasmid DNA strand transfer, we have investigated the length requirement for ssDNA within a linear duplex necessary for efficient TraI-catalyzed DNA unwinding. We find that singlestranded regions shorter than 27 nt do not support efficient helicase activity.

EXPERIMENTAL PROCEDURES
Expression and Purification of R1 TraI-The expression construction pHP2 (50) carries the 5.3-kilobase pair traI gene of plasmid R1drd16 (GenBank TM accession number AY423546) under the control of the P tac promoter in vector pGZ119HE (51). Cultures (600 ml) of E. coli SCS1[pHP2] were grown at 37°C in 2ϫ TY (16 g/liter Bacto-Tryptone, 10 g/liter Bacto yeast extract, 5 g/liter NaCl) medium (52) supplemented with 0.1% (w/v) glucose, 25 mM MOPS (pH 8.0), 25 g/ml thiamine/HCl, and 10 g/ml chloramphenicol. At an A 600 of 0.5, isopropyl-1-thio-␣-Dgalactopyranoside was added to a concentration of 1 mM. Shaking of cells was continued at 30°C for 5 h. Cells were harvested by centrifugation and resuspended in a total volume of 100 ml of 20 mM spermidine, 200 mM NaCl, and 2 mM EDTA (pH 8.0) and frozen at Ϫ20°C. Frozen cells were thawed overnight at 4°C, and the suspension was adjusted to final concentrations of 50 mM Tris/HCl (pH 7.6), 30 mM NaCl, 5% (v/v) sucrose, 0.15% (v/v) Brij-58, 275 g/ml lysozyme, and 0.1 mM phenylmethylsulfonyl fluoride. After 1 h at 0°C the lysis mixture was centrifuged at 100,000 ϫ g for 90 min at 4°C. Solid ammonium sulfate was slowly added to the supernatant to 50% saturation (0.29 mg/ml), stirred for 30 min on ice, and then centrifuged 27,000 ϫ g for 30 min at 4°C. The precipitate was dissolved in a total volume of 40 ml in buffer A (20 mM Tris/HCl (pH 7.6), 10% glycerol (v/v), 0.1 mM EDTA, 1 mM DTT, and 50 mM NaCl). Fraction I was dialyzed against 3 changes of a 5-fold volume of buffer A for 3 h. The dialyzed fraction was loaded on a Pharmacia HiTrap TM heparin-Sepharose column. Proteins were eluted with a 0.05-1.2 M gradient of NaCl using buffer B (20 mM Tris/HCl (pH 7.6), 10% glycerol (v/v), 0.1 mM EDTA, 1 mM DTT, 1.2 M NaCl). TraI eluted at 415 mM NaCl (fraction II). Soluble ammonium sulfate was added to a final concentration of 1 M, and the mixture was applied to a phenyl-Sepharose column equilibrated with buffer C (20 mM Tris/HCl (pH 7.6), 10% glycerol (v/v), 0.1 mM EDTA, 1 mM DTT, and 1 M (NH 4 ) 2 SO 4 ). The protein was eluted using a decreasing gradient of 1 to 0 M (NH 4 ) 2 SO 4 in buffer D (20 mM Tris/HCl (pH 7.6), 10% glycerol (v/v), 0.1 mM EDTA, and 1 mM DTT). The TraI protein eluted at ϳ250 mM (NH 4 ) 2 SO 4 (fraction III). Peak fractions were pooled, dialyzed against buffer D, concentrated against polyethylene glycol 20000 supplemented with glycerol to a final concentration of 40% (v/v), and stored at Ϫ20°C (fraction IV). This fraction was greater than 90% pure TraI protein based on Coomassie-stained polyacrylamide-denaturing gels. Protein concentration was determined using the Bradford protein assay (Bio-Rad) with bovine serum albumin as standard.
Preparation of Helicase Substrates-Linear partial duplex substrates containing a single-stranded gap were prepared in several steps. Three dsDNA fragments (A, B, C) were generated in independent PCR reactions performed with 200 M each dNTP, 20 ng of pBluescript II SK DNA (Stratagene), 0.2 M each primer (Table I), and 1 unit of Dynazyme II polymerase (Finnzymes) in the buffer provided. To render one strand in each fragment sensitive to subsequent exonuclease attack, one primer in each pair was 5Ј-phosphorylated before amplification. The reverse primer used to amplify fragment A and the forward primers employed to create fragments B and C were phosphorylated nonradioactively by T4 polynucleotide kinase and reisolated using mini Quick Spin™ columns (Roche Applied Science). PCR products A and B were common to every substrate. Product C was unique for each because the site of annealing of the forward primer in each case was positioned a few nt further removed on the template pBluescript DNA relative to the reverse primer, which was constant for each preparation. Product C for all substrates was radiolabeled internally by adding [␣-32 P]dATP to the PCR reaction mixture. The conditions for amplification were 94°C for 3 min then 30 cycles at 94°C for 30 s, 60°C for 40 s, and 72°C for 1 min followed by one step at 72°C for 5 min. For the 30ss and 40ss substrates 55°C annealing temperature was required. The 5Ј-phosphorylated strand of each PCR fragment was then selectively degraded by 5 units of exonuclease III (New England Biolabs)/ 20-l PCR reaction mix at 37°C for 1 h. After nuclease treatment, the products were purified with Qiagen PCR purification kits. The yield and concentration of ssDNA were determined on 1.4% agarose Tris borate EDTA gels. To create the partial linear duplex DNA substrates, the three protected strands of fragments A, B, and C were combined at a molar ratio of 3:3:1 in 10 mM Tris/HCl (pH 8.5) and 200 mM NaCl and heat-denatured at 94°C for 5 min. Annealing was achieved in reiterated cycles of decreasing temperature (1°C over four 25-s increments per cycle) to a final temperature of 16°C. This hybridization mix was used directly in helicase assays or stored at Ϫ20°C.
Verification of Partial Duplex Substrate Structure-Typical nuclease   Helicase Assay-Standard reactions (20 l) were performed at 37°C for 20 min in a solution containing 40 mM Tris/HCl (pH 7.5), 4 mM MgCl 2 , 1 mM DTT, 10% glycerol, 50 g/ml bovine serum albumin, 1.8 mM ATP, ϳ3-3.5 ng of DNA (375 pM) substrates, and 0 -52 nM TraI. For kinetic studies carried out in the presence of 52 nM enzyme a scaled-up (14-fold) reaction mixture was assembled and warmed to 37°C, and the reaction was started by the addition of protein. Portions (20 l) were removed at 0.5, 1, 1.5, 2, 3, 4, 5, 7.5, 10, 12.5, 15, and 20 min, and the reactions were stopped by the addition of 0.2 volumes of loading buffer (50 mM EDTA, 50% glycerol, 1% SDS, and 0.1% bromphenol blue). The products were resolved on 1.4% agarose gels in Tris borate EDTA buffer at 7 V/cm for 2.5 h. Radiolabeled DNA was visualized by autoradiography of the dried gels. Data were quantified using ImageQuant software (Molecular Dynamics). The percent unwound helicase substrate was determined after background correction: % unwound ϭ signal intensity in displaced fragment divided by total substrate signal.

Expression, Purification, and ATPase Activity of R1 TraI
Protein-TraI was overexpressed in E. coli SCS1 and purified from crude cell extracts as described under "Experimental Procedures." The final fraction contained a greater than 90% homogeneous solution of TraI as judged by SDS-PAGE and Coomassie Blue staining (data not shown). The purified protein exhibited an apparent molecular mass of 180 kDa. Typically, 20 -25 mg of protein were obtained from 4.8 liters of cultured cells.
Extensive studies have characterized the NTPase activity of TraI from plasmid F (97% identity between predicted proteins) (53,54). In good agreement with these reports the ATPase activity of TraI from plasmid R1 was dependent on the presence of a ssDNA effector and activated by Mg 2ϩ . In the presence of 2 mM ATP maximal activity was observed at 2 mM MgCl 2 , and half-maximal values were obtained at 0.3 and 20 mM MgCl 2 (not shown). One unit of enzymatic activity is defined as the amount of enzyme needed to hydrolyze 1 nM ATP in 20 min at 30°C. The progress of the purification procedure was monitored by measuring the ATPase activity at the later stages, as summarized in Table II. A specific activity of 1970 kilounits/mg was determined for the final fraction containing TraI.
Preparation and Verification of the Helicase Substrate Structures-The TraI protein of plasmid F is known to require a region of ssDNA adjacent to the duplex to facilitate its DNA unwinding activity (47)(48)(49). To evaluate the requirement of this enzyme for a 5Ј ssDNA tail, we chose to create a series of linear partial duplex substrates containing a centrally located single-stranded gap of variable length. Duplex arms were generated by hybridizing two distinct non-overlapping singlestranded fragments to one common longer fragment of complementary ssDNA as illustrated in Fig. 1. To be able to control the gap size precisely, the position of the 5Ј end of the primer utilized to amplify fragment C was varied incrementally relative to the start of the left arm duplex (Fig. 1B). A series of partial duplex molecules were generated with this approach that differed only in the defined length of the single-stranded region present after hybridization.
The structure of the resulting products was verified based on nuclease sensitivity (Fig. 2). The oriT region of IncF(II) plasmid R1 is essentially devoid of commonly used restriction endonuclease cleavage sites. Utilization of this DNA for creating suitable helicase substrates would make verification of the gap size in each substrate prohibitively difficult. Therefore, pBluescript DNA was chosen to be able to exploit the dense arrangement of restriction endonuclease recognition sites in the multiple cloning site when these are positioned in and around the singlestranded gap of the partial duplex hybridization products. For each substrate preparation sensitivity of intermediate products and the product of their hybridization to restriction endonucleases and single strand-specific nuclease S1 was analyzed. The results for three of the seven different substrates used in this study are presented in detail. Briefly, fragments A, B, and C were first synthesized in independent PCR reactions using pBluescript DNA as template. Fragment C was labeled internally in this reaction in the presence of [␣-32 P]dATP. One strand of each was then selectively degraded 5Ј to 3Ј by exonuclease III. The remaining ssDNA molecules were pooled and hybridized to create a unique linear molecule of partial duplex character. The series of seven substrates contains an ssDNA region 67-12 nt long flanked by 2 dsDNA arms. The right arm contains radiolabeled DNA and varies in length in accordance with the gap size. TraI is expected to bind to the gap, translocate 5Ј to 3Ј on the single-stranded loading region, enter the downstream duplex, and displace the radiolabeled strand (Ն427 nt). Each substrate in the series is prepared from a slightly longer fragment C DNA with a concomitant reduction in gap size. B, sequence 643-711 of pBluescript II SK DNA including the multiple cloning site is shown. The map indicates the different primer positions (arrows) relative to the restriction endonuclease sites (above) used to verify each substrate structure. Precise limitation of the extent of single-stranded gap is achieved by incremental shifts in the position of the fragment C forward primer toward the end of the left arm duplex designated by the position of the reverse primer B. Numbers below the C primers indicate the distance (in nt) from primer B.

ssDNA Requirement for TraI Duplex Unwinding
In Fig. 2A the structure of the hybridization product expected to contain a single-stranded gap 67 nt in length is demonstrated. Importantly, for every substrate just one of the component ssDNA fragments was radiolabeled. Thus, not all DNA species are visible in the autoradiogram. The different DNA species present after -exonuclease III treatment of the pooled PCR products before their hybridization (lane 2) were also treated with several nucleases and resolved electrophoretically through non-denaturing agarose gels (lanes 3-6). The fastest and the slowest migrating labeled species were insensitive to restriction endonucleases (lanes 3-5) but were com-pletely degraded by S1 nuclease (lane 6). In contrast, the band migrating somewhat less than the 405-bp double-stranded standard DNA fragment (lane 13) was resistant to S1 nuclease (lane 6). Treatment with PstI did not alter the apparent migra-  2. Verification of the partial duplex substrate structure. The pattern of nuclease sensitivity used to confirm the partial duplex structure and the length of single-stranded gap present in the 67ss (A), 27ss (B), and 16ss substrates (C). Sensitivity of intermediate products (left) and the product of their hybridization (right) to restriction endonucleases and single strand-specific nuclease S1 is shown. Radiolabeled species include exonuclease-digested fragment C (left, ssDNA), residual double-stranded fragment C (dsDNA), and the lower strand of the right duplex arm of hybridized products (right). Sizes of the predicted products of nuclease treatment when the labeled DNA is present as a fully duplex intermediate product or partially duplex hybridized product are indicated in the upper or lower portions, respectively, of the insets (far left). For clarity, the unlabeled fragments expected, but not detectable by autoradiography, are indicated in light gray.
ssDNA Requirement for TraI Duplex Unwinding consistent with values expected for nuclease-treated doublestranded PCR product C (see inset, Fig. 2A) before one strand was removed with -exonuclease III. We conclude, therefore, that this band represents a residual double-stranded 427-bp fragment C resulting from incomplete exodegradation. The slowest migrating band was not present in all preparations. Taken together, the resistance to endonucleases, sensitivity to S1, and the observations that this band was eliminated by heat denaturation of the samples (not shown) or incubation with TraI helicase (Fig. 3A) before electrophoresis suggest that this product is in fact an artifact possibly caused by partial annealing of single-stranded products of the exonuclease reaction in the pooled population of intermediate DNA fragments. The fastest migrating band would, thus, appear to represent the desired single-stranded fragment C.
After heat denaturation and hybridization, the population of DNA products revealed two discernable bands (lane 7). The very faint band comigrating with dsDNA present in the nonhybridized population (lane 6) was insensitive to S1, ApaI, and PstI digestion (lanes 12, 8, 9, respectively), consistent with the preceding assignment of this band to fully duplex fragment C. The prominent species migrating between the 621-and 700-bp DNA standards (lanes 1 and 13) was reduced to a single species migrating somewhat less than the 405-bp standard band upon S1 treatment (lane 12). Incubation with NotI released a slightly more mobile band than that produced by S1 (lane 10). These results support the conclusion that the prominent labeled DNA in lane 7 is a partial duplex molecule with a central region of ssDNA flanked by two duplex arms. The single-stranded gap can be cleaved by S1 nuclease to release the right and left duplexes, but only the 427-bp labeled right arm is visible. Duplex character left of the central single-stranded gap was confirmed with PvuII. Two PvuII recognition sites are present in the fragment of pBluescript DNA common to all partial duplex substrates (fragment A). Treatment of the hybridization products would be expected to yield two radiolabeled fragments 448 and 170 bp long. Release of the 448-bp fragment requires that the site left of the single-stranded gap is present in doublestranded form. PvuII digestion of this preparation yielded two bands exhibiting similar mobility. Correlation of expected and existing gap size is provided by sensitivity to NotI and resistance to ApaI and PstI endonuclease digestion in accordance with the sequence of the forward primer used to generate fragment C (Fig. 1B).
The structure of intermediate products and hybridized products in the preparation of all seven partial duplex substrates described in this study was demonstrated using the same general analysis (not shown). The selection of enzymes was varied as the expected gap size was narrowed from 67 to 12 nt. Verification of the structure of the substrate with the shortest FIG. 3. TraI catalyzes 5 to 3 unwinding of the linear 67ss partial duplex. A, increasing amounts of TraI, as indicated above the lanes (nM), were incubated with the partial duplex substrate in the standard helicase reaction, and the products were resolved electrophoretically (right). Controls lacking cofactors are indicated (lanes [11][12][13]. Heat-denatured substrate was applied to lane 14. The position of migration of unwound product is indicated (*). The products of reaction mixtures containing exonuclease-treated PCR products with 0 (lane 2) and 52 nM TraI (lane 3) were also resolved. Labeled dsDNA fragments of known length (as shown in bp) were applied for reference (lane 1). B, the kinetics of unwinding from 67ss were determined in the presence of 52 nM TraI under standard helicase assay conditions. Portions of the reaction were removed at the times indicated above the lanes (min), and catalysis was terminated in ice-cold EDTA. Products of the reactions (lanes 2-14) and a heat-denatured sample of the substrate (‚, lane 15) were resolved electrophoretically on native gels. The position of migration of unwound product is indicated (*).

FIG. 4. Concentration dependence of the unwinding reaction.
The maximal helicase activity was determined using seven DNA substrates with defined single-stranded gaps. Each panel shows an autoradiogram (right) of a representative gel for substrates containing single-stranded regions ranging from 67 to 12 nt as illustrated (left). Standard helicase assays were performed with increasing amounts of purified TraI protein as indicated above the lanes (nM). Products of the reactions (lanes 1-10) and a heat-denatured sample of the substrates (‚, lane 11) were resolved on native Tris borate EDTA-agarose gels. Controls lacking cofactors are indicated (lanes 8 -10).
region of ssDNA relied on digestion with XhoI, which cuts the right duplex arm only in this case, combined with continued sensitivity to S1 nuclease. For completeness, the results for two additional substrates containing single-stranded regions 27 and 16 nt in length are shown in Fig. 2. The preparation of the 27-nt gapped duplex (Fig. 2B, lane 8) contains some residual double-stranded fragment C. Thus, PvuII digestion of the substrate (lane 12) yields in addition to those products predicted for the partially duplex DNA, an ϳ300-bp fragment diagnostic for the fully duplex species. Members of this series of partial duplex substrates are designated according to the confirmed length of single-stranded gap in nt, for example, 67ss.
The 67ss DNA Substrate Supports Helicase Activity-To ascertain whether the length of ssDNA present in 67ss, which is an intermediate value in the earlier estimation of 12-200 nt (47), is sufficient for efficient TraI helicase activity, the dependence of the reaction on protein concentration and the time course of fragment displacement from the 67ss substrate at 52 nM TraI were determined (Fig. 3). The lowest concentration of protein that yielded detectable unwinding (16%) after 20 min at 37°C (Fig. 3A) was 2.6 nM. This corresponds to a protein (monomer) to substrate ratio of 7:1. Given that a 4 -6-fold excess of unlabeled ssDNA was also present in each labeled substrate preparation, we estimate that the protein to DNA ratio for this reaction mixture approaches 1:1. Performance of the enzyme under these conditions was, thus, similar to reported activities of F TraI at similar protein to DNA ratios on partially duplex circular DNA substrates (48). The comparison is important because in that case TraI was exposed to thousands of nt of ssDNA for binding before displacing the annealed complementary fragment. Complete unwinding of 67ss was observed at protein concentrations greater than 26 nM, again in good agreement with the performance of a Ͼ10-fold excess of F TraI on heteroduplex single-stranded cirular substrates (48). The requirement for ATP and divalent cation cofactors was assessed by omission (Fig. 3A, lanes 11-13). As expected no helicase activity was detected in the absence of ATP. The ad-dition of MgCl 2 was not essential, as has been observed previously (48). At the maximum protein concentration (Fig. 3B) 55% of the fragment was displaced after 90 s. Near quantitative unwinding of the substrate was observed after 7.5 min at 37°C. We conclude that a single-stranded region 67 nt long 5Ј to the duplex is, therefore, sufficient for efficient TraI helicase activity.
Maximal Helicase Activity on Substrates with Shortened Single-stranded Regions-Given that the 67-nt gap was sufficient to support highly efficient DNA unwinding activity by TraI, additional substrates were prepared, and their structures were characterized, as described above (Fig. 2 and data not shown), which limited the extent of single-stranded region adjacent to the duplex in a stepwise manner (67, 40, 30, 27, 20, 16, and 12 nt). The relative efficiency of DNA unwinding activity catalyzed by TraI on the series of substrates was compared using the standard helicase reaction and increasing amounts of purified protein (Fig. 4). The data are presented graphically in Fig. 5. The efficiency of the reaction was dependent on protein concentration and the length of ssDNA present on the substrate. At the highest protein concentrations nearly quantitative unwinding of the 67ss and the 40ss substrates was observed. In comparison, the maximum efficiency of unwinding was reduced to 80 and 60% when the ssDNA was shortened to 30 or 27 nt, respectively. Very limited unwinding of the partial duplex DNAs containing a single-stranded region Յ20 nt was observed. These experiments were repeated (n Ն 5) using two independent preparations of each substrate (Fig. 5).
The kinetics of the unwinding reaction on the series of substrates at 52 nM TraI were also compared (Fig. 6). The rates of the overall reaction decreased as the length of ssDNA present in the substrates was limited. The extent of unwinding on substrates containing Յ40 nt of ssDNA did not continue to increase to 100% efficiency but reached a plateau after ϳ7.5 min. Although the assay employed here measures the combined efficiency of loading the helicase and duplex unwinding, we infer that the rate-limiting step in the overall reaction is the initiation of the unwinding process, which is the binding to the available ssDNA and entering the duplex, and not the rate of unwinding itself. We conclude that a single-stranded region at least 27 nt in length is sufficient to support efficient TraI- ssDNA Requirement for TraI Duplex Unwinding catalyzed unwinding of the adjacent duplex. In contrast, singlestranded regions shorter than 20 nt are not sufficient to stimulate efficient helicase activity. DISCUSSION The initiation pathway for TraI-catalyzed unwinding of duplex DNA in isolation or as part of the conjugative relaxosome has not been described. Recruitment of replicative helicases to replication origins typically requires the activity of sequencespecific DNA-binding proteins that confer site specificity to the helicase activity as well as induce localized distortions in the duplex to provide regions of ssDNA for the helicase to bind (46). The IncF-IncW family of DNA/mobilizing systems achieve site specificity in oriT DNA unwinding by physical linkage of the conjugative helicase to an N-terminal DNA relaxase domain (34,38,55,56), 2 which acts to cleave the transfer origin with nt precision, and presumably via interactions with additional components of the relaxosome. In contrast to studies of the initiation process of replication at prokaryotic origins of replication, however, reconstitution of origin unwinding and the initiation of conjugative DNA synthesis in vitro has not been achieved. Although some helicases initiate the unwinding reaction from a nicked substrate, this activity has not been observed in vitro for TraI (54). The observations that (i) the intermediate (open) product of the TraI-catalyzed cleavingjoining reaction at nic failed to support helicase entry to the open circular form as it has been reconstituted in vitro thus far (45) and (ii) the helicase activity when studied in isolation required a single-stranded tail 5Ј to the duplex (49) imply that localized duplex melting at nic is necessary to load the TraI helicase. In vitro reconstitution of helicase activity for enzymes that require single-stranded ends to enter the duplex is typically achieved when a short fragment or oligonucleotide targeted for displacement is annealed to a comparatively large single-stranded circular or linear phage DNA. Substrates of this type offer an excess of entry sites as the helicase can bind anywhere on the available ssDNA and translocate to the duplex region. Early studies of TraI by Hoffmann-Berling and co-workers (47) sought to limit the extent of ssDNA at the single strand/double strand junction, resulting in the estimation that ϳ200 nt of ssDNA tail were required for this enzyme (47). With the aim of gaining insight to the initiation pathway utilized by TraI as a conjugative helicase, the present study required preparation of a series of linear DNA substrates where the length of single-stranded region adjacent to the target duplex could be defined. The linear substrates used contained a continuous DNA lattice with two terminally flush duplex arms separated by a central single-stranded region of variable length. The effect of varying the extent of ssDNA present at single strand/double strand junctions on the efficiency of TraI-catalyzed unwinding was observed. The end product of the assay measured the combined efficiency of loading at the single-stranded region and subsequent unwinding of a ϳ450-nt fragment from the downstream target duplex. We found that decreasing the available single-stranded region stepwise from 67 to 40 nt did not curtail the efficiency of unwinding, but a further reduction to 30 nt resulted in a loss of 20% of the maximal activity under standard assay conditions. Limiting the ssDNA to 27 nt decreased the maximal activity to only 60% that observed on the 67ss and 40ss substrates. When the gap size was decreased further to 20, 16, or 12 nt, efficient duplex unwinding was no longer observed.
This assay cannot distinguish the effect of gap size on individual steps presumed to occur in the initiation process such as binding affinity, enzyme orientation, translocation on the ssDNA lattice, duplex entry, etc. In view of the fact that during conjugation loading of TraI helicase is both site-and strandspecific, we expect that the nature of the region flanking nic in oriT is an important effector of the initiation pathway. Dissection of these steps, therefore, will be carried out on sequencespecific substrates. The necessity of conducting the current study on different DNA is due to the virtual absence of restriction recognition sites in the IncF oriT. In the absence of suitable sites, confirmation of the actual substrate gap size would be quite difficult to prove. Moreover, all gel systems we have tested were unable to differentiate large fragments of fully duplex linear DNA from molecules containing 30 bp or less of melted duplex for a similar series of oriT-specific substrates (described below). Thus the rigorous evaluation of the extent of ssDNA present in the substrates that is crucial to the present analysis would be practically untenable.
The current findings have been integrated into the construction of oriT-specific heteroduplex DNA substrates where variable lengths of the sequence surrounding nic are present as open (noncomplementary) duplex and flanked by two doublestranded arms, one of which contains double-stranded oriT DNA in its native context relative to nic. We presently know that reconstitution of both the nic-cleaving and the DNA-unwinding activities of TraI on substrates of this type is possible. 3 The effect of the extent of duplex melting around nic on the efficiency of TraI-catalyzed nic cleavage and DNA unwinding is under investigation. These studies will further assess the effects of the additional presence of auxiliary DNA-binding proteins of the IncF relaxosome as well as E. coli SSB protein on the activities of TraI at its origin-specific initiation site.