Interaction of the recA Protein of Escherichia coli with Adenosine 5 ’-0-( 3-Thiotriphosphate ) *

Incubation of the recA protein of Escherichia coli with the ATP analog adenosine 5’-0-(3-thiotriphosphate) (ATP(yS)) in the presence of DNA produces an irreversible inhibition of ATPase activity, although in the presence of ATP, ATP(@) shows an initial competitive inhibition. ATP(yS) is not appreciably hydrolyzed by recA protein and the inhibition of ATPase activity is due to the formation of stable complexes which contain equimolar amounts of ATP(yS) and recA protein. Formation of stable complexes requires DNA, which is also stably bound to recA protein in the presence of ATP(yS), at a ratio of 5 to 10 nucleotides/recA protein monomer. The DNA requirement is satisfied by either singleor double-stranded DNA, and in the latter case, the pH dependence is comparable to that observed for ATP hydrolysis. Binding of ATP(yS) is inhibited by other nucleoside diand triphosphates with efficiencies corresponding to their inhibitory effects on the ATPase activity of recA protein.

efficiently hydrolyzed by the recA protein. Here, we extend these observations and show that although inhibition is competitive, it is also essentially irreversible due to the formation of recA protein-ATP(yS) complexes that neither hydrolyze the ATP(@) nor dissociate during the course of most recA protein-dependent reactions. Furthermore, we find that the complexes contain 1 ATP(yS) molecule/recA protein monomer and require DNA for their formation.

EXPERIMENTAL PROCEDURES
Material~-[~~S]ATP(yS) was a generous gift from Dr. F. Eckstein from New England Nuclear. Unlabeled ATP(yS) and GTP(yS) were (Max Planck Institut, Gottingen, Germany) and was also purchased from Boehringer Mannheim; UTP(yS) was also generously donated by Dr. Eckstein. In various preparations, between 60 and 90% of the radioactivity of [35S]ATP(yS) was in ATP(yS), as judged by polyethyleneimine cellulose chromatography (1) and Norit adsorption (see below). Unlabeled NTP(yS) preparations showed a similar variation in purity, depending on their age, with the major contaminant being the nucleoside diphosphate. NTP(yS) concentrations were determined using the extinction coefficient of the corresponding NTP. Other nucleoside tri-and diphosphates were obtained from P-L Biochemicals and Sigma; r3H]ATP was from Amersham; NasPOaS. 12H20 was from Ventron (Alfa); Norit from Baker; nitrocellulose filters (HAWP, 45 nm, 24 mm diameter) from Millipore Corp.; calf thymus DNA from Sigma; poly(dT) and (dT)12 from P-L Biochemicals; other DNA's were prepared as described previously (4, 10). recA protein was Fraction I1 (lo), purified through the phosphocellulose step, and was greater than 90% pure. Its concentration was determined from the absorbance at 280 nm using an extinction coefficient (an A~w of 1 equals 51 p~ recA protein) calculated from the amino acid composition (11). Other reagents were as described previously (10).
Assay for Hydrolysis of ATP-ATP hydrolysis was assayed by thin layer chromatography on polyethyleneimine cellulose plates as described previously (10). The standard reaction mixture contained 20 mM buffer, 10 mM MgC12, 1 mM dithiothreitol, and DNA, ATP, and recA protein as indicated.
Assay for Hydrolysis of ATP(yS). Formation of inorganic thiophosphate (P03S) by the hydrolysis of [35S]ATP(yS) was monitored by Norit adsorption. An aliquot of the reaction mixture was mixed with an equal volume of Na3P03S and 2% Sarkosyl, and then 5 volumes of 50 mM Na3P03S containing 10% (w/v) Norit was added. After mixing, the Norit was removed by centrifugation and radioactive P O S in the supernatant fluid was measured. The standard reaction mixture was the same as for the ATPase assay.
Filter-binding Assay-Following incubation of the standard reaction mixture (described above), an aliquot was filtered with suction through a nitrocellulose filter previously soaked in 20 mM Tris-HC1 (pH 8.0), 10 mM MgC12, 1 mM dithiothreitol, and 30 mM NaCl (B buffer). The fdter was washed once with 10 volumes of either B buffer (low salt wash) or B buffer containing 1 M NaCl (high salt wash) followed by a second wash with 10 volumes of B buffer. When all liquid had passed through, the filter was dried and the bound radioactivity determined.
Filtration was routinely carried out at 2 ml/min; slower rates did not increase retention. At least 100 pg of recA protein could be retained by the filters and no difference in efficiency of retention was found between filtration of 20-1.11 or 90-pl aliquots of a 27 pM solution   Fig. 1) with a KITP(yS) of approximately 0.6 p~ as contrasted with a K , for ATP of approximately 20 p~ (12). ATP($) also inhibited the DS DNA-dependent ATPase and the UTPase activities of recA protein (data not shown).

Irreversible Inhibition
ATP(@) inhibited the extent as well as the initial rate of ATP hydrolysis, suggesting that in the presence of ATP($) the enzyme was irreversibly altered. To test this possibility, recA protein was preincubated with varying concentrations of ATP(yS) in the presence of SS DNA, then ATP was added, and the ATP(yS) concentration was adjusted so that it remained constant and ATPase activity was measured. As shown in Fig. 2, preincubation with increasing concentrations of ATP(yS) led to a corresponding increase in inhibition of the ATPase. The degree of inhibition was directly proportional to the amount of ATP(@) present during preincubation. Thus, since the ATP(yS) concentration was constant during ATP hydrolysis, inhibition must have occurred during the preincubation. Approximately 70% of the ATPase activity could be inhibited under these conditions, although >99% of the ATPase activity was due to recA protein.' Maximum inhibition occurred when there were equimolar concentrations of ATP(yS) and recA protein. Thus, ATP(@) causes an irreversible inhibition of the ATPase activity of the recA protein.
The irreversible inhibition required DNA. As shown in Fig.  3, preincubation of recA protein with excess ATP(yS) in the presence of varying amounts of SS DNA led to a progressive inhibition of ATPase activity, dependent on the ratio of DNA nucleotides to recA protein. Maximum inhibition occurred at a ratio of about 9 nucleotides/recA protein monomer. When recA protein was preincubated with ATP( $3) in the presence of DS DNA at pH 6.2, inhibition of both DS and SS DNAdependent ATPase activities occurred, whereas at pH 8.0, there was no inhibition, consistent with the pH optimum for DS DNA-dependent ATP hydrolysis and DS DNA binding by the recA protein (8,10). Preincubation in the absence of DNA gave no inhibition of DNA-independent ATPase activity at either pH 6.2 or pH 7.5.
These findings indicate that the irreversible inhibition of ATPase activity by ATP(yS) has requirements similar to those for ATP hydrolysis, Le. a DNA cofactor and the appropriate pH in the presence of DS DNA. These observations suggest 3 possible mechanisms for the inhibition: (i) ATP( yS) is hydrolyzed, but the thiophosphate produced is tightly bound to the enzyme and dissociates very slowly; (ii) hydrolysis is initiated and a covalent recA protein-ATP( $3) or recA protein-POaS intermediate accumulates; or (iii) no bonds are broken, but a stable noncovalent recA protein-ATP(yS) complex is formed. The experiments described below favor the 3rd alternative.
Lack of Hydrolysis of ATP(yS) by recA Protein-Incuba- The residual ATPase activity may result from enzyme molecules which bind the ADP contaminating the ATP(-@) preparation and are thus protected from ATP(yS) inhibition. . .

ATP[YSl:
..  produced at a rate of approximately 1 mol of ATP(@) hydrolyzed/mol of recA protein/1500 min, the limit of detection and several orders of magnitude below the rate of ATP hydrolysis (data not shown). Treatment of the reaction with either SDS or Sarkosyl did not produce an increase in the amount of P03S formed. Thus, it is unlikely that PO& produced by hydrolysis of ATP(yS) and bound noncovalently to the enzyme, is responsible for inhibition of ATPase activity.
Detection of recA Protein-ATP(yS) Complexes-Following incubation of recA protein with [35S]ATP(yS), the 35S was converted to a form that was stably retained by nitrocellulose filters even after washing with 1 M NaCl ( Table I) Fig. 2 except that the ATP(yS) concentration was 26 p~, and the concentration of +X174 SS DNA was varied. Aliquots were diluted 20-fold into an assay mixture as described in Fig. 2 except that ATP(@) was omitted and the +X174 SS DNA concentration was 168 p~. Initial velocities were determined from a time course.

Filter-binding assay for recA protein-ATP($) complexes
The standard reaction (50 pl) contained 20 mM buffer (either Tris- conversion required recA protein and either SS or DS DNA, although DS DNA was effective only at a pH at which the DS DNA-dependent ATPase was active. Retention on nitrocellulose fiiters was due to binding of recA protein to the filter since alkali-treated fiiters that have lost the capacity to bind DNA (9) were equally effective in this assay. Thus, the 35S in ["S]ATP(yS) exists in a complex with recA protein that can be retained on nitrocellulose fiiters in the presence of 1 M NaC1. Furthermore, the conditions for formation of this complex parallel those required for irreversible inhibition of ATPase activity. Incubation of the complexes at 60 "C for 2 min prior to Titration resulted in loss of the bound 35S (Table 11), although recA protein was still efficiently retained by the fiiters under these conditions (data not shown). The treatment at 60 "C did not produce inorganic P03S as judged by Norit adsorption.

Interaction of recA Protein with ATP(yS) 8853
remarkable stability accounts for the irreversibility of the inhibition of ATP hydrolysis by ATP(yS).
Although the turnover number for SS DNA-dependent ATP hydrolysis a t 37 "C is about 10 ADP produced/min/ recA protein monomer, 1 to 2 min were required for complete binding of ATP(@) to recA protein (Fig. 5). At 30 "C, binding of ATP(yS) in the presence of SS DNA was slightly slower, while in the presence of DS DNA (pH 6.2), the rate was much reduced (Fig. 6). In the presence of 20 p~ recA protein and 30 PM ATP(yS), the DS DNA-dependent reaction showed kinetics similar to the reaction in the presence of SS DNA (data not shown). These results indicate that the high affinity of recA protein for ATP(yS) is not a consequence of a rapid rate of formation of recA protein-ATP(yS) complexes, but is rather a result of their very slow rate of dissociation and, in fact, the rate of formation of these complexes is slower than the rate of turnover of ATP during hydrolysis.
Incubation in the absence of DNA at pH 6.2 or 8.0 produced  ATP( yS) suggests a noncovalent association. Characteristics of the Reaction Leading to Tight Binding ofATP(yS) to recA Protein-Although sensitive to heat treatment, the recA protein-ATP(yS) complexes were otherwise extremely stable, being resistant to exhaustive washing with 1 M NaCl and exposure to 30 mM EDTA. As shown in Fig. 4, the complexes had a half-life of about 90 min a t 37 "C. Their

Requirements for tight binding of ATP(yS) to recA protein
The standard reaction (50 pl) with 20 mM Tris-HC1 (pH 8.0) contained 4.3 p~ ["S]ATP(yS), 2.6 p~ recA protein, and either 101 p~ +X174 SS DNA, 125 p~ (dT)12, or 108 p~ (dT)lmj. Incubation was for 20 min at 30 "C at which time 30 p1 was fitered (low salt wash) as described under "Experimental Procedures." In the complete reaction, 0.6 mol of ATP(yS) was bound/mol of recA protein. Interaction of recA Protein with ATP(,&$ no increase in ATP(yS) binding over a period of at least 120 min, consistent with the failure of ATP(yS) to inhibit irreversibly the DNA-independent ATPase activity of recA protein. Furthermore, no stable binding of ATP(yS) was observed in the presence of DS DNA at pH 8.0.
As shown in Table 111, the formation of stable recA protein-ATP(7S) complexes required Mg". As noted earlier, once formed they were resistant to EDTA. Formation of the stable complex was also sensitive to salt, although once formed, the final complexes were stable to 1 M NaC1. Unlike the ATPase reaction (4), formation of complexes was not inhibited by Nethylmaleimide. The polynucleotide requirement for complex formation showed the same specificity as for ATP hydrolysis (10); in particular, (dT)12 did not promote complex formation but (dT)Iwo did (Table 111), indicating a similar polynucleotide size requirement as for ATP hydrolysis. As shown in Fig. 7, tight binding of ATP(+) depended on the ratio of recA protein to DNA, and saturation occurred at about 6 nucleotides/recA protein monomer with either SS or DS DNA, a value similar to that observed for the irreversible inhibition of ATPase activity. At pH 8.0, DS DNA (29 nucleotides/recA protein monomer) in the presence of subsaturating amounts of S S DNA (either 2 or 5 nucleotides/recA protein monomer) caused no additional binding of ATP(+) over the SS DNA level, although under these conditions, binding of recA protein to DS DNA is stimulated by SS DNA (2,9).
Stoichiometry of ATP(yS) Binding to recA Protein-Titration of recA protein in the presence of excess ATP(@) (Fig.  8) indicated that about 1.3 mol of ATP(yS) were bound/mol of recA protein monomer at 37 "C. Eighty per cent of the labeled ATP(yS) was bound when recA protein was in excess over ATP(@). When the titration was performed at 30 "C, 1.0 mol of ATP(yS) was bound/mol of recA protein monomer (data not shown). Titration of ATP(yS) in the presence of a constant amount of recA protein (Fig. 9) gave a ratio of 1.6 mol of ATP(@) bound/mol of recA protein monomer, although in this case, only 60% of the labeled ATP(yS) was bound. In similar titrations, a t 30 "C, ratios of 1.3 and 0.7 were observed with 2.6 and 22 PM recA protein, respectively. Finally, when recA protein was preincubated with varying amounts of unlabeled ATP(@) in the presence of DNA and  then excess ["5S]ATP(yS) added, binding of the labeled ATP(@) was maximally inhibited when the ratio during preincubation was 21 mol of ATP(yS)/mol of recA protein monomer (Fig. 10). Under these conditions, maximal binding of labeled ATP(@) was only 0.5 mol/mol of recA protein monomer. These results indicate that approximately 1 molecule of ATP(@) is tightly bound/recA protein monomer. Furthermore, since this ratio was also observed for the irreversible inhibition of ATPase activity by ATP($) (Fig.  2), Interaction of recA Protein with ATP(yS) 8855 protein, and the indicated NTP or NDP. Incubation was for 20 min at 30 "C, at which time 30yl aliquots were filtered (low salt wash) as described under "Experimental Procedures." In the absence of added NTP or NDP, 0.5 mol of ATP(+) was bound/mol of recA protein monomer. This value represents 100% binding.