Large scale purification and characterization of TraI endonuclease encoded by sex factor plasmid R100.

The TraI protein encoded by plasmid R100 was purified in a large scale by monitoring the strand- and site-specific nicking activity at the origin of transfer, oriT. The N-terminal amino acid sequence of the purified protein was identical to that deduced from the DNA sequence of an open reading frame encoding TraI. The TraI protein is a DNA helicase which is highly processive and unwinds DNA in the 5′ to 3′ direction. The Stokes radius and the sedimentation coefficient for the TraI protein in 200 mM NaCl indicate that the protein is a rod-shaped monomer, whose native molecular weight is 186,000. Chemical cross-linking analysis revealed that there exist more dimers of TraI under the low salt conditions, under which both nicking and unwinding reactions catalyzed by TraI are the most efficient, indicating that the TraI protein is functionally active in a dimer form. TraI hardly introduced a nick into the linearized plasmid DNA and only slightly into the relaxed closed circular DNA, indicating that TraI requires superhelical structure of substrate DNA for the nicking reaction. Deletion analysis in the oriT region revealed that a particular region of 54 base pairs containing oriT is required for the nicking reaction.

The TraI protein encoded by plasmid R100 was purified in a large scale by monitoring the strand-and sitespecific nicking activity at the origin of transfer, oriT. The N-terminal amino acid sequence of the purified protein was identical to that deduced from the DNA sequence of an open reading frame encoding TraI. The TraI protein is a DNA helicase which is highly processive and unwinds DNA in the 5 to 3 direction. The Stokes radius and the sedimentation coefficient for the TraI protein in 200 mM NaCl indicate that the protein is a rod-shaped monomer, whose native molecular weight is 186,000. Chemical cross-linking analysis revealed that there exist more dimers of TraI under the low salt conditions, under which both nicking and unwinding reactions catalyzed by TraI are the most efficient, indicating that the TraI protein is functionally active in a dimer form. TraI hardly introduced a nick into the linearized plasmid DNA and only slightly into the relaxed closed circular DNA, indicating that TraI requires superhelical structure of substrate DNA for the nicking reaction. Deletion analysis in the oriT region revealed that a particular region of 54 base pairs containing oriT is required for the nicking reaction.
Conjugation is the process in which DNA is transferred from one bacterial cell harboring a sex factor plasmid, such as F or R100, to another by cell-to-cell contact (for the most recent review, see Frost et al. (1994)). One of the initial events in DNA transfer is the strand-and site-specific nicking at the origin of transfer, oriT, by the plasmid-specified endonuclease that is the traI gene product whose molecular mass is estimated to be about 192 kDa from the nucleotide sequence (Yoshioka et al., 1990;Bradshaw et al., 1990). The TraI protein encoded by plasmid F or R100 has been purified by monitoring its ATPase activity or by its molecular size, respectively. The TraI proteins purified introduce the strand-and site-specific nick, such that it is covalently linked with the 5Ј end of the nick (Inamoto et al., 1991(Inamoto et al., , 1994Reygers et al., 1991;Matson and Morton, 1991). The TraI proteins have been known to be DNA helicase (Abdel-Monem and Hoffman- Berling, 1976;Inamoto et al., 1994), which is supposed to unwind the duplex DNA from the nick introduced in the plasmid to provide the single-stranded DNA (Abdel-Monem et al., 1983), which is known to be transferred to the recipient cell (Ohki and Tomizawa, 1968;Rupp and Ihler, 1968).
In this paper, we purified the TraI protein encoded by R100 by monitoring the site-and strand-specific endonuclease activity and extensively analyzed the nicking and unwinding reactions catalyzed by TraI. We show that a dimer form of the TraI protein is active for both nicking and unwinding reactions, and that the superhelical plasmid DNA molecules with a specific region of 54 bp 1 are required for the nicking reaction.
The Klenow fragment of DNA polymerase I and T4 polynucleotide kinase were obtained from Takara. Restriction endonucleases, BamHI and HindIII (New England Biolabs), were used. Single-stranded DNAbinding protein (SSB) was obtained from Promega. RNase A and proteinase K were obtained from Sigma and Boehringer Mannheim, respectively. They were used as recommended by their suppliers.
DNA Preparation-Covalently closed circular plasmid DNAs used for in vitro nicking experiments were isolated from large-scale cultures, as described (Ohtsubo et al., 1978). An alkaline lysis method (Yoshioka et al., 1987;Maniatis et al., 1982) was used to prepare a small amount of plasmid DNA from large numbers of cell cultures. Single-stranded DNA of pUC119 or its derivative was prepared as described previously (Vieira and Messing, 1987).
Oligonucleotides-Oligonucleotides listed in Table I were synthesized using Applied Biosystems 392 DNA/RNA Synthesizer and purified by reversed-phase chromatography or by elution from a denaturing polyacrylamide gel after electrophoresis.
Nicking Reaction-The standard nicking reaction was performed as follows: 2 l of the covalently closed circular pHF11 DNA (37 fmol) in a solution (10 mM Tris-HCl (pH 8.0), 1 mM EDTA) was mixed with 3 l of 5 ϫ buffer N (150 mM Tris-HCl (pH 7.4), 25 mM MgCl 2 , 5 mM dithiothreitol, 50 g of bovine serum albumin/ml) and 8 l of Milli-Q water. After preincubation for 5 min at 37°C, the reaction was initiated by adding 2 l of the TraI protein in buffer A containing 200 mM NaCl and 20% glycerol. The mixture was incubated at 37°C for 30 min, and the reaction was stopped by adding 5 l of the stop solution (2% SDS, 33 mM EDTA). After addition of 1 l of proteinase K (20 mg/ml), the mixture was further incubated at 30°C for 30 min. DNA in the mixture was then extracted with phenol/chloroform and precipitated with ethanol. After digestion of the sample DNA with BamHI, the 5Ј ends of the DNA fragments were labeled with exchange reaction using T4 polynucleotide kinase and [␥-32 P]ATP (see Fig. 1) using the Megalabel TM kit (Takara). The 32 P-labeled fragments were denatured and electrophoresed in a 6% polyacrylamide sequencing gel containing 7 M urea. The radioactivity of bands of the fragments was determined using Bio-Image Analyzer (BAS1000 or BAS200, Fuji Photo Film). The nicking efficiency was estimated from the ratio of the nicked molecules (143 nt) to the half of total DNA molecules (236 and 143 nt) (see Fig. 1).
Purification of the TraI Protein-The TraI protein was purified monitoring the endonuclease activity assayed by the nicking reaction (see above) as follows. Cells (88 g) of E. coli CSR603 harboring pYY35-1, which overproduces the TraI protein, were lysed as described previously (Inamoto et al., 1994) and centrifuged at 39,000 ϫ g for 1 h to obtain the supernatant (Fraction I). Then, 0.3 g/ml of ammonium sulfate was added to Fraction I with stirring. The precipitate was collected by centrifugation at 27,000 ϫ g for 15 min, resuspended in 24 ml of buffer A (50 mM Tris-HCl (pH 7.5), 1 mM EDTA, 1 mM dithiothreitol, and 20% glycerol) and dialyzed twice against the same buffer (1 liter) for 6 h. The dialysate (Fraction II) was applied to a phosphocellulose P11 column (2.5 cm ϫ 7.6 cm; Whatman) equilibrated with buffer A, and the column was washed with 3 column volumes of the same buffer. Proteins were eluted with a 400 ml of 0 -1.0 M NaCl linear gradient in buffer A. The fractions (0.19 -0.31 M NaCl) with the TraI endonuclease activity were pooled and dialyzed against buffer A containing 50 mM NaCl. The dialysate (Fraction III) was then applied to a MonoQ (HR 5/5; Pharmacia) column equilibrated with buffer A containing 50 mM NaCl. The column was washed with 2 column volumes of the same buffer, and the proteins were eluted with a 20 ml of 50 -400 mM NaCl linear gradient in buffer A. The fractions (155-164 mM NaCl) with the TraI specific endonuclease activity were pooled. The dialysate (Fraction IV) was applied to a Superdex 200pg (HiLoad 16/60; Pharmacia) column equilibrated with buffer A containing 200 mM NaCl, and the column was developed with the same buffer. The peak fraction with the TraI endonuclease activity, which corresponded to the K av of 0.172, was pooled and dialyzed against buffer A containing 50 mM NaCl. The pooled solution (Fraction V) was applied to a HiTrap Heparin (10 ml; Pharmacia) column equilibrated with buffer A containing 50 mM NaCl. The column was washed with 5 column volumes of the same buffer, and the proteins were eluted with a 50-ml 50 mM to 1.0 M NaCl linear gradient in buffer A. The fractions (0.40 -0.53 M NaCl) with the TraI endonuclease activity were pooled and dialyzed against buffer A containing 50 mM or 200 mM NaCl. Total 12.8 mg of the protein was obtained and stored at Ϫ80°C.
In every step in these procedures, the relative amount of the TraI protein in each fraction was monitored by electrophoresis in SDS-8% polyacrylamide gels according to Laemmli (1970). The protein concentration was determined by the Bio-Rad protein assay using bovine serum albumin as standard.
Determination of Stokes Radius-Stokes radius was determined by gel filtration of the purified TraI protein using Superdex 200pg column as described above. Ovalbumin, albumin, aldolase, catalase, ferritin, and thyroglobulin (Pharmacia gel filtration calibration kit) were used as standards.
Sucrose Density Gradient Sedimentation-Sucrose gradient (5 ϳ 20% w/v) was made in buffer A in tubes measuring 13 ϫ 51 mm for a Beckman SW 50.1 rotor. The TraI protein purified was layered on to the gradient in one tube. The solution containing standard proteins (catalase, aldolase, and albumin) was layered on to the gradient in the other tube. Centrifugation was carried out at 20°C for 10 h at 50,000 rpm.
Five-drop fractions were collected, and the concentration of proteins was determined by SDS-polyacrylamide gel electrophoresis or by the Bio-Rad protein assay.
Chemical Cross-linking-Chemical cross-linking was carried out as described (Runyon et al., 1993) with the following modification: crosslinking of TraI (2.4 M monomer) was carried out at 24°C in buffer A or in the nicking reaction buffer containing 10 -275 mM NaCl after preincubation for 10 min at 24°C. Cross-linking was initiated by addition of 1% glutaraldehyde to 0.1% and quenched after 30 min by addition of 100 mM lysine chloride to 10 mM. Products were analyzed by SDS-5% polyacrylamide gel electrophoresis.
Helicase Assay-The substrate used for the helicase assay (see Fig. FIG. 1. Specific nicking by the TraI protein. A, schematic representation of the structure of the BamHI fragment (236 bp) with or without a nick. Asterisks indicate the 32 P-labeled 5Ј ends of DNA strands (thick lines). The fragment was obtained by digestion with BamHI of the pHF11 DNA after treatment of the DNA with the TraI protein under the conditions described under "Materials and Methods." Note that the 5Ј end of the nick is attached by the TraI protein (open circle) and thus cannot be labeled with 32 P. B, a 6% polyacrylamide sequencing gel showing single-stranded DNA fragments generated from the BamHI fragment after denaturation of the sample DNA with heat. Lane 1, the sample DNA treated with TraI; lane 2, the sample DNA not treated with TraI. Lanes marked with M indicate sequencing ladders used as size markers. Note in lane 1 that a fragment of 143 nt is generated due to nicking by TraI at oriT (see A).
a Small letters indicate transition mutations such that G, A, T, and C were converted to T, C, G, and A, respectively, in the original sequence in the region ϩ18 to ϩ1.
b Coordinates to P-mcs are of pUC119 (Vieira and Messing, 1987). Coordinates to the other primers are of R100 (see Fig. 8).
4A) was prepared as follows. Two times molar excess of the 51-mer oligonucleotide (P-mcs) labeled with 32 P at the 5Ј-end was mixed with single-stranded DNA of pUC119 in TE buffer (10 mM Tris-HCl (pH 8.0), 1 mM EDTA), and the mixture was heated at 65°C for 5 min, cooled down slowly (ϳ1°C/min) to room temperature, and applied to QIAGEN tip20 (QIAGEN Inc.) to remove non-annealed oligonucleotides. The DNA sample was dissolved in TE buffer. The standard helicase reaction was performed as follows: 0.5 l of the substrate (50 fmol) in TE buffer was mixed with 2 l of 5 ϫ buffer N, 0.7 l of 20 mM ATP, and 5.8 l of Milli-Q water. After preincubation for 5 min at 30°C, the reaction was initiated by adding 1 l of the TraI protein in buffer A containing 200 mM NaCl and 20% glycerol. The mixture was incubated at 30°C for 10 or 5 min, and the reaction was stopped by adding 2 l of the stop solution (50 mM EDTA, 0.6% SDS, 40% glycerol, 0.12% bromphenol blue). The reaction mixture was electrophoresed on a nondenaturing 8% polyacrylamide gel in TAE buffer (Maniatis et al., 1982), and the radioactivity of bands was determined using Bio-Image Analyzer.
Determination of the Direction of DNA Unwinding Catalyzed by TraI-A 24-mer oligonucleotide (P52) labeled with 32 P at the 5Ј end was annealed with a 42-mer oligonucleotide (P55) and used as a substrate (␣ in Fig. 6A). Also, a 24-mer oligonucleotide (P53) labeled with 32 P at the 5Ј end was annealed with a 42-mer (P54) and used as the other substrate (␤ in Fig. 6A). Helicase assay was performed as described above, except that 1 l of substrate ␣ or ␤ (0.2 pmol) was used in each reaction, and products were analyzed by 10% polyacrylamide gel electrophoresis.

RESULTS
Purification of the TraI Protein-The TraI protein of plasmid R100 has been shown to introduce a nick at oriT on a specific strand of the plasmid (Inamoto et al., 1994). Plasmid pHF11 carries a 236-bp BamHI fragment containing oriT of R100 (see Fig. 1). TraI introduces the nick at oriT in the BamHI fragment in this plasmid, resulting in generation of the single-stranded DNA of 143 nt long when the plasmid DNA was subsequently digested with BamHI, labeled with 32 P at the 5Ј ends, and denatured (see Fig. 1). It should be noted that nicking results in generation of another fragment of 93 nt also derived from the specific strand of the 236-bp BamHI fragment, but this fragment could not be labeled with 32 P because of the presence of TraI covalently linked to its 5Ј end (see Fig. 1).
Using this method of detection of the site-and strand-specific endonuclease activity of the TraI protein, we tried to purify the TraI protein in a large scale by column chromatography from the crude lysate of cells harboring plasmid pYY35-1 (traI ϩ ) overproducing the TraI protein. The crude cell lysate contained the TraI protein in an amount greater than 20% of total protein as estimated by SDS-polyacrylamide gel electrophoresis, but did not show any nicking activity (Table II). However, a fraction obtained from phosphocellulose column showed the nicking activity (Table II), suggesting that the initial crude lysate contained either an inhibitor(s) or a nuclease(s) causing degradation of the substrate DNA. The TraI protein in the phosphocellulose column fraction could be purified to homogeneity through MonoQ FPLC, Superdex 200pg, and HiTrap Heparin columns (see Table II). The specific activity of the TraI protein increased 6.2 times through the purification (Table II).
In these purification steps, the specific activity of the fraction obtained using MonoQ column was 4.6 times greater than that of the fraction obtained from phosphocellulose column. However, the relative amount of TraI estimated by scanning SDSpolyacrylamide gels increased only 1.1 times (Table II). This suggests that the phosphocellulose fraction still contained an inhibitor(s) or a nuclease(s). In phosphocellulose, Superdex 200pg or MonoQ column chromatography, the nicking activity corresponded to the single peak of protein mass (see Fig. 2A for the result of Superdex 200pg column chromatography). In Hi-Trap Heparin column chromatography, however, the profile of the nicking activity showed two peaks, which corresponded to those in the profile of protein concentration (Fig. 2B). We assume that this is due to multimerization of the TraI protein, as will be discussed later.
We carried out amino acid sequencing analysis of the purified TraI protein and found that the N-terminal amino acid sequence was MLSFSVV⅐⅐⅐, which is identical to that deduced from the nucleotide sequence of an open reading frame for the traI gene (Yoshioka et al., 1990). This in turn indicates that the N terminus of the TraI protein is neither blocked nor processed.
Physicochemical Properties of TraI-The molecular mass of   Inamoto et al. (1994). e Data from Yoshioka et al. (1990). the TraI protein encoded by R100 has been estimated to be 192 kDa from the nucleotide sequence of the traI gene and 180 kDa by SDS-polyacrylamide gel electrophoresis of the protein purified previously by monitoring its molecular size (see Table III).
To determine the molecular weight of the purified TraI protein, we carried out gel filtration, which yielded a Stokes radius of 63.2 Å, and also sucrose gradient sedimentation analysis, which gave a sedimentation coefficient of 6.7 S in the buffer containing 200 mM NaCl. According to the method described by Siegel and Monty (1966), we calculated a native molecular mass for TraI to be 186,000 with a friction rate of 2.1 (Table  III). These results indicate that the TraI protein is rod-shaped and exists as monomers in the buffer containing 200 mM NaCl. We then carried out the sedimentation analysis of the purified TraI protein in the buffer containing 50 mM NaCl, and found that the sedimentation coefficient sifted to 8.9 S (Table  III). This suggests that TraI forms dimers under the low salt conditions. To verify this assumption, we carried out chemical cross-linking analysis and found that TraI monitored by SDSpolyacrylamide gel electrophoresis existed as dimers under the low salt conditions but as monomers under the high salt conditions (Fig. 3).
Unwinding Reaction by TraI-We have recently demonstrated that TraI encoded by R100 is a DNA helicase (Inamoto et al., 1994). We quantitatively analyzed the helicase activity of TraI purified as described above using the single-stranded pUC119 DNA annealed with the 51-mer oligonucleotide labeled with 32 P at its 5Ј end as substrate (see Fig. 4). The unwinding reaction by TraI was found to reach a plateau in several minutes (Fig. 5A, a). Eighty percent of the substrate added were unwound by addition of 2.2 TraI molecules per substrate DNA molecule (Fig. 5A, b). The optimal NaCl concentration was 20 mM, and the efficiency of the reaction was reduced to one-hundredth in 100 mM NaCl (Fig. 5A, c). These results combined with the data shown in Fig. 3 suggest that the TraI helicase is functionally active in a dimer form.
Lineweaver-Burk plotting of the data gave V max of 0.47 nM/ min and K m of 6.6 nM at 30°C, indicating efficient binding of TraI to the substrate DNA. The value of V max corresponds to an unwinding velocity of 290 bp/s, which means that the unwinding reaction is immediately completed once the reaction starts.
From these values, we calculated k cat (turnover) to be 0.12/min and k cat /K m to be 1.9 ϫ 10 7 M Ϫ1 min Ϫ1 . The values indicate that TraI does not rapidly turn over. These results show that the TraI helicase is highly processive.
We examined several parameters for the unwinding reaction. The unwinding reaction occurred most efficiently at 37°C. Optimal pH was 7.2. The unwinding reaction absolutely required Mg 2ϩ and ATP. The optimal concentration of Mg 2ϩ is 5 mM, and that of ATP is 4 mM as far as we measured the activity at the concentrations from 0.25 to 4 mM. We also examined the effect of the single-stranded DNA-binding protein, SSB, on the unwinding reaction. SSB promoted the unwinding reaction slightly, about 5%, at the concentration of 16 nM tetramer (3.2 times over substrate), but inhibited the reaction to a certain extent at the concentrations of over 120 nM (24 times over substrate); SSB inhibited the reaction by 31% at 400 nM (80 times over substrate) and by 47% at 1.6 M (320 times over substrate).
We also determined the direction of DNA unwinding by TraI using the substrate, ␣ or ␤, which is a 24-bp duplex with the 5Јor 3Ј-overhang of 18 nt, respectively (Fig. 6A). TraI unwound substrate ␣ with the 5Ј-overhang very efficiently, but unwound substrate ␤ with the 3Ј-overhang very poorly (Fig. 6B). This indicates that TraI unwinds DNA in the 5Ј to 3Ј direction. This agrees with the direction of unwinding by TraI (Helicase I) encoded by F (Lahue and Matson, 1988).
Nicking Reaction Catalyzed by TraI-The nicking reaction catalyzed by TraI reached a plateau essentially in a relatively short time (Fig. 5B, a). However, nicking did not occur beyond 55% of substrate molecules even upon long incubation of the reaction mixture and/or addition of much TraI protein (Fig. 5B,  b). The nicking reaction efficiently occurred under low salt, but not under high salt; nicking could hardly be detected at the concentration of NaCl over 100 mM (Fig. 5B, c). This indicates that the TraI nickase is active in a dimer form.
We then examined several parameters for the nicking reaction. The nicking reaction occurred most efficiently at 37°C. Optimal pH was 7.6. The optimal concentration of Mg 2ϩ was 10 mM. Unlike the unwinding reaction, the nicking reaction occurred in the absence of ATP. ATP rather inhibited the nicking activity, probably because ATP enhanced the helicase activity that leads to movement of TraI on DNA, resulting in inhibition of the nicking reaction. We also examined the effect of SSB on the nicking reaction. Interestingly, SSB inhibited the nicking reaction to 33% at 13 nM (5.4 times over substrate) and to 7% at 210 nM (87 times over substrate).
Superhelical Structure and DNA Sequence Required for the Nicking Reaction by TraI-To see whether TraI binds to a specific DNA sequence around oriT, we carried out gel-shift assay and filter-binding assay using the restriction fragments of plasmid DNA digested with several restriction enzymes. TraI was, however, found to bind not specifically to the frag-ment containing oriT (data not shown). Nicking did not occur when the plasmid DNA linearized with a restriction enzyme, such as HindIII which cleaves the DNA at a single site, was used as substrate (Fig. 7A). Nicking did not efficiently occur when the closed circular DNA was relaxed by treatment with topoisomerase I and used as substrate, while nicking occurred very efficiently when superhelical DNA was used as substrate (Fig. 7A). When the nicking reaction was carried out using the superhelical DNA in the presence of ethidium bromide, the nicking efficiency was decreased or increased in a manner dependent on the concentration of ethidium bromide (Fig. 7B). These results show that TraI requires superhelical DNA as substrate for the nicking reaction.
As described above, the sequence recognized by TraI could not be determined using the linear fragments of plasmid DNA. To define the region required for nicking by TraI, therefore, we constructed a set of plasmids with a deletion in the oriT region (see Fig. 8) and examined them to see whether TraI introduces FIG. 5. Unwinding reaction (A) and nicking reaction (B) by TraI. a, time course of the unwinding or nicking reaction. Concentrations of TraI and substrate used were 11 and 5 nM, respectively. b, stoichiometry of the unwinding or nicking reaction. c, effect of the NaCl concentration on the unwinding or nicking reaction. Concentrations of TraI and substrate used were 22 and 5 nM, respectively. The helicase and nicking activities were assayed by the conditions described under "Materials and Methods." FIG. 6. Direction of DNA unwinding by TraI. A, substrate DNA, ␣ and ␤, used to determine the direction of DNA unwinding by TraI. Asterisks indicate the 5Ј ends labeled with 32 P. B, a 10% polyacrylamide gel, showing the unwinding activity of the TraI protein. Substrate ␣ or ␤ was incubated with TraI, and the products were analyzed using the conditions described under "Materials and Methods." Note that substrate ␣ was unwound, whereas substrate ␤ was not unwound, demonstrating that TraI travels in the 5Ј to 3Ј direction on single-stranded portion of the substrate and unwinds the double-stranded DNA. a nick or not. As shown in Fig. 8, a particular region of 54 bp containing oriT was found to be essential for nicking. This indicates that this 54-bp region is recognized by TraI for the nicking at oriT.

DISCUSSION
In this paper, we have shown that the TraI protein of R100 is a rod-shaped molecule with molecular weight 186,000. This is consistent with the result obtained by electron microscopy that the protomer of TraI protein of plasmid F is rod-shaped (Abdel-Monem et al., 1977). We have also shown here that the TraI protein forms dimers under the low salt conditions and monomers under the high salt conditions, and that dimers of TraI are active in both nicking and unwinding reactions. Previous stoichiometric analysis has suggested that a multimer consisting of about 70 to 90 monomers of TraI of F shows the highest helicase activity (Kuhn et al., 1978;Benz and Muller, 1990), and that 10 molecules of TraI of F per DNA molecule show the highest helicase activity although substantial activity was detected in the reaction involving a 1:1 ratio of enzyme molecules to DNA molecules (Lahue and Matson, 1988). These results differ from ours obtained for TraI of R100. It is noted here, however, that several other helicases, except one, form either dimers or hexamers (Lohman, 1992). Rep helicase and helicase II (UvrD) have been shown to be dimerized upon binding to either single-stranded or duplex DNA (Wong et al., 1992;Runyon et al., 1993). This reminds us the result that in the purification procedure of TraI protein, chromatography using the HiTrap Heparin column showed two peaks, although chromatography using other columns showed a single peak (see Fig. 2). This may suggest that TraI was dimerized upon binding to heparin, a DNA analog.
We have shown here that the TraI protein introduces the nick into only a little over half of the substrate DNA even upon long incubation and/or addition of an excess amount of the TraI protein. It is possible that the nicking reaction requires some factors, which are not present in the reaction mixture. Such factors may include the traY gene product and the integration host factor IHF, which have been recently shown to stimulate the nicking reaction catalyzed by TraI (Inamoto et al., 1994) by binding to the sites (sbyA and ihfA) located immediately adjacent to the nicking site oriT . It might also be possible that TraI possesses the topoisomerase activity to convert superhelical DNA to relaxed closed circular DNA, which was shown in this paper to be poorly nicked by the TraI protein. However, the TraI protein has no topoisomerase activity, since the relaxed circular DNA molecules were not detected in an agarose gel containing chloroquine after electrophoresis of the sample treated with TraI (data not shown). We cannot, however, exclude the possibility that the nicked molecules with TraI covalently attached are still in a superhelical DNA form and can be readily converted to the superhelical DNA molecules without TraI, leading to an equilibrium state between two kinds of superhelical DNA molecules with and without TraI.
We have recently found that the TraI protein has the singlestranded DNA binding activity and cleaves the single-stranded DNA at oriT. 2 This DNA cleaving activity is inhibited by SSB. These findings lead us to assume that the oriT region in the double-stranded DNA molecules melts locally, providing a single-stranded DNA portion which is cleaved by TraI. This assumption may be supported by our present result that the nicking reaction is strongly inhibited by SSB.
The nicking reaction occurs most efficiently under the low salt conditions, but it hardly occurs at high salt (100 mM NaCl or more), as described in this paper. The cleavage reaction of the single-stranded DNA by TraI occurs most efficiently under the low salt conditions like the nicking reaction, but it still occurs even at the NaCl concentration 100 mM unlike the nicking reaction. 2 It is possible that the superhelical DNA, that is the substrate for the nicking reaction, cannot be melted locally under the high salt conditions, thus not providing any single-stranded DNA portion which is bound and cleaved by the TraI protein. Alternatively, dimers of TraI formed under the low salt conditions may form a DNA-protein complex at the partially melted region around oriT which leads to the cleavage reaction, but monomers formed under the high salt conditions may not form such a complex. FIG. 7. DNA substrate required for the nicking reaction. A, nicking at oriT on superhelical DNA (closed circles), relaxed closed circular DNA (open circles), or linearized DNA (closed squares) of plasmid pHF11. The nicking activity was assayed as described under "Materials and Methods." B, nicking of superhelical pHF11 DNA in the presence of ethidium bromide. The nicking reaction was initiated by adding 2 l of the TraI protein (477 fmol/l) after addition of ethidium bromide to the reaction mixture and preincubation for 12 min.
FIG. 8. The essential region for nicking by TraI. A critical portion of the oriT region is shown at the top. Numbers are nucleotide positions defining the 3Ј end of the nick as ϩ1, and the 5Ј end of the nick as Ϫ1. Inverted repeat sequences are indicated by pairs of arrows between two DNA strands. The IHF-binding site, ihfA , and the TraY-binding site, sbyA , are shown by hatched and open boxes, respectively. The pHF plasmids carry a DNA segment from the oriT region of R100 (thick lines). These DNA segments were synthesized by polymerase chain reaction using oligonucleotide primers with a BamHI site. The polymerase chain reaction-amplified fragments were digested with BamHI and cloned into the BamHI site of pUC18. The results of nicking by TraI are shown on the right side of the panel.