Purification and Characterization of a DNA-dependent ATPase from Escherichia coZi*

A DNA-dependent ATPase has been isolated and purified from an Escherichia coli cell-free extract. The ATPase has the following characteristics: preferential dependence on single-stranded DNA, specificity for ATP hydrolysis, K, value of 1.4 x lo-’ M for ATP, and molecular weight of approximately 69,000. The ATPase can be shown to bind to single stranded DNA. The resemblance between this ATPase and that isolated from vaccinia cores is discussed. Several reports have indicated that ATP is required for in vitro DNA replication (l-4). In toluene-treated cells of Escherichia cob, none of the other ribonucleoside triphosphates could replace ATP (5, 6). The ATP requirement may be explained by the existence of an enzyme which metabolizes ATP and participates in DNA replication. This hypothetical could be a DNA-dependent ATPase. We, therefore, undertook a study of such enzymatic

However, because of substantial loss of the enzyme activity (one-third was recovered), the specific activity was not increased by this treatment.
The specific activity of the last phosphocellulose fraction was 380 times higher than that of the extract freed of DNA (Table  I) Table I) from AB2463 or AB1157 was chromatographed on the first phosphocellulose column as described under "Purification of DNA-dependent ATPase." The ATPase was assayed for 30 min using 2 ~1 of each fraction under standard conditions: recap (0) and recA+ (0). The concentration of NaCl was determined by measuring the ionic strength of several fractions from each column (conduct meter: Philips-PR 9501). A, recA-; and A, recA+. Centrifugation-Because of substantial loss of enzyme activity during filtration on Sephadex columns, glycerol gradient centrifugation (10 to 30%) of the enzyme was carried out in the presence of bovine serum albumin which protected the enzyme from inactivation (see below). Thus, the recovery of the activity was 80%' and the sedimentation coefficient was calculated as 5.2 S using two marker enzymes: alcohol dehydrogenase (yeast) and alkaline phosphatase (Escherichia coli). The same sedimentation coefficient was obtained for the Fraction III and Fraction IV.
10 min of reaction time at concentrations higher than 10m5 M ATP, the initial velocity of the reaction was proportional to the enzyme concentration (from 5 units to 20 units). Fig. 2 shows the variation of initial velocity determined at 10 min with varying ATP concentrations and a Lineweaver-Burk plot of the data. The K, value obtained from the latter figure was 1.4 x lo-' M.
This gave approximately the same estimate of the molecular weight as was determined by the Sephadex G-150 filtration (assuming the same type of configuration for the proteins tested).

Specificity of Nucleoside
Triphosphate-The specific requirement for ATP has been one of the striking characteristics of DNA replication in E. coli. Therefore, we determined whether our DNA-dependent ATPase could hydrolyze other ribonucleoside triphosphates. As is shown in Fig. 2 The enzyme fractions collected after glycerol gradient centrifugation of Fraction V (in the absence of bovine serum albumin) were concentrated by 5% trichloroacetic acid precipitation and were analyzed on sodium dodecyl sulfate-polyacrylamide gel in order to estimate molecular weight (see "Materials and Methods"). A single band of protein was obtained. A value of 74,000 was calculated for this protein by comparison with standards of known molecular weight (trypsin, glyceraldehyde phosphate dehydrogenase, pyruvate kinase, catalase, and bovine serum albumin).
The similar molecular weights obtained for the ATPase activity under native conditions indicate that the single band of protein found on the gel may represent the monomeric state of the ATPase. Assay Conditions-Our assay conditions for DNA-dependent ATPase were adapted from those used by Wickner et al. (19) for the ATPase found in the preparation of dnaB gene product. We have investigated the question of whether these conditions were optimum for our ATPase activity in the presence of heat-denatured DNA. Using the Fraction IV, we have observed that the pH (from 6.5 to 8.0) and different types of buffer (Tris, potassium phosphate, and morpholino propane sulfonic acid) in the above pH range had little influence on enzyme activity. However, KC1 slightly activated (by 30%) the enzyme at concentrations between 0.05 and 0.15 M. The addition of MgCl, was not required for enzyme activity. Rifampicin (20 rg/ml) did not affect the ATPase activity. Bovine serum albumin (500 kg/ml) protected enzyme activity; in its presence the ATP degradation proceeded linearly for 1 hour, while in its absence, the reaction slowed after 10 min at 30". Analysis-In order to determine the products of the reaction, we have carried out the reaction with the same amount of Fraction IV in the presence of either [U-'"C]ATP or [y-32P]ATP. In the first test, 3.7 nmol of ADP were produced (production of AMP, less than 0.01 nmol), while in the second reaction 3.5 nmol of Pi were obtained. In both cases, no acid-precipitable radioactivity was found. Therefore, we concluded that the reaction consisted of hydrolysis of ATP into ADP and Pi. DNA (nmde) FKdiiS   FIG. 3 (left). Effect of heat-denatured or activated calf thymus DNA on ATPase activity.
The activated DNA (0) or its denatured form obtained by heating (0) was added to the standard reaction mixture together with the Fraction IV. The reaction was carried out for 30 min at 30".
Concentration of ATP-In order to determine the apparent K, value of the ATPase with respect to ATP, we measured the velocity of the reaction at different concentrations of ATP. For   FIG. 4 (right).
Binding property of ATPase to Ml3 DNA. Fifty microliters of Fraction IV were incubated for 10 min at 30" in the presence of 0.55 nmol of SzP-labeled Ml3 DNA (5 x 10' cpm) and 100 ~1 of H,O. Immediately afterward, the mixture was ultracentrifuged on the glycerol gradient (see "Materials and Methods").