Purification and Properties of the Endonuclease Specific for Apurinic Sites of Bacillus stearothermophilus *

An endonuclease specific for apurinic sites when doublestranded DNA is used as substrate has been isolated from the thermophilic bacterium, Bacillus stearothermophilus; it is a monomeric protein of about 28,000 daltons, without action on normal DNA strands or on alkylated sites. The enzyme is quite thermoresistant in the presence of other proteins, has an optimal temperature of MY, needs monovalent cations for optimal activity, is insensitive to EDTA, and is inhibited by divalent cations; it has no associated exonuclease activity. These latter properties are closer to those of Escherichia coli thermoresistant endonuclease IV, which is also insensitive to EDTA and has no exonuclease activity, and very different from those of the main endonuclease for apurinic sites of the same bacterium. The B. stearothermophilus enzyme is more resistant to urea and detergents than the main E. coli endonuclease for apurinic sites and has a higher content of hydrophobic amino acids.

An endonuclease specific for apurinic sites when doublestranded DNA is used as substrate has been isolated from the thermophilic bacterium, Bacillus stearothermophilus; it is a monomeric protein of about 28,000 daltons, without action on normal DNA strands or on alkylated sites. The enzyme is quite thermoresistant in the presence of other proteins, has an optimal temperature of MY, needs monovalent cations for optimal activity, is insensitive to EDTA, and is inhibited by divalent cations; it has no associated exonuclease activity. These latter properties are closer to those of Escherichia coli thermoresistant endonuclease IV, which is also insensitive to EDTA and has no exonuclease activity, and very different from those of the main endonuclease for apurinic sites of the same bacterium.
The B. stearothermophilus enzyme is more resistant to urea and detergents than the main E. coli endonuclease for apurinic sites and has a higher content of hydrophobic amino acids.
DNA spontaneously loses purines (1) and pyrimidines (2). The rate of base loss is considerably increased by treatment with chemicals like alkylating agents or by exposure to ionizing radiation. Endonucleases specific for apurinic sites in DNA have been found in Escherichia coli (3)(4)(5), in animals (6,7), and in plants (8).
E. coli possesses two endonucleases for apurinic sites. The main enzyme, which is responsible for 90% of the cell activity, has been completely purified by Verly and Rassart (9); it is thermolabile, is inhibited by EDTA, needs magnesium ions to be active, and might be the same enzyme as exonuclease III (10, 11). The accessory enzyme, which is responsible for 10% of the cell activity, has been called endonuclease IV (5); it resists heating at 45", is not inhibited by EDTA, and is devoid of exonuclease activity. Depurinated DNA has been repaired in vitro with three enzymes: the main E. coli endonuclease specific for apurinic sites, DNA polymerase I and the four deoxyribonucleosides triphosphates, ligase and its coenzyme (12). Gossard and * This work was supported by grants from the Medical Research Council of Canada and the Fonds de la Recherche Fondamentale Collective of Belgium. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "Wuertisemerzt" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
$ Recipient of studentships from the National Research Council and the Phenix Foundation. Verly (11) gave the details of the repair molecular mechanism. The spontaneous loss of DNA bases must be very high in thermophilic bacteria at the temperature at which they usually live so that we looked for an endonuclease hydrolyzing a phosphoester bond near apurinic sites in Bacillus stearothermophilus. The enzyme was found and purified. It is thermoresistant when protected by other proteins; the presence of a high percentage of hydrophobic amino acids, likely grouped in a central core, might explain this property. The B. stearothermophilus enzyme behaves more like endonuclease IV than like the main endonuclease for apurinic sites of E. coli (endonuclease VI).

MATERIALS AND METHODS
Media and Buffers-Sargeant's medium (13) consists of the following: 20 g of Bacto-tryptone (Difco), 10 g of yeast extract (Difco), 0.32 g of citric acid, 1.3 g of K,SO,, 3  alkylated, or alkylated-depurinated, and incubated at 37"; aliquots were taken after 0 to 120 min to measure the acidsoluble radioactivity. Fig. 2 shows that the extract had some action on untreated DNA, more on alkylated DNA, but that its action was far greater when alkylated sites were replaced by apurinic sites (alkylated-depurinated DNA The Escherichia coli results are from Verly and Rassart (9). Because no data were available for cysteine, methionine, and tryptophan, the molar percentages do not take account of the possible presence of these amino acids in the proteins. The polarity index is calculated by summing the polar amino acids and half the total of those of the intermediate class (23) Controls without enzyme were carried out in the same way. At the end of the incubation, 600 ~1 of 0.15 M NaCl, 0.015 M EDTA, pH 7.0, were added and the solution was dialyzed at 4" against the same buffer. Each sample was then split in two parts; one was denatured with NaOH and the other with formamide before sedimentation on neutral sucrose gradients.
The average number of breaks/strand was estimated from the sedimentation profile. crude extract activity on alkylated-depurinated DNA was found to be at 7.5. Among different buffers at pH 7.5 that were tested, 0.05 M Hepes was found the best for the enzyme activity.
Aliquots of crude extract diluted loo-fold with Buffer I containing 2% bovine serum albumin (20 ~1) were incubated with 20 ~1 of the alkylated-depurinated [3H]DNA solution for 10 min. The optimal temperature for the enzyme was found to be 60". Below this temperature, the logarithm of the reaction velocity plotted against l/T (T = absolute temperature in K) yields a straight line from which an activation energy of 21,000 cal/mol can be calculated for the enzymecatalyzed reaction. Physical Properties of Purified Enzyme -Two aliquots of Preparation VII (200 ~~11, containing 4 pg of protein, were submitted to polyacrylamide gel electrophoresis (see "Materials and Methods"). One of the gels was stained with Coomassie blue; the other was cut in 2-mm slices which were ground in 250 ~1 of Hepes buffer containing 4% bovine serum albumin and, after centrifugation, the enzyme activity of the supernatant was measured on alkylated-depurinated [3H]DNA. There was a single protein band corresponding to the enzyme activity (Fig. 3).
In another experiment, the solution of Preparation VII, which contained 1% sodium dodecyl sulfate, and three proteins of known molecular weights (ovalbumin, chymotrypsinogen, ribonuclease A) dissolved in the same buffer, were heated at 100" for 2 min and submitted to gel electrophoresis.
After staining with Coomassie blue, the migration coefficients were calculated. There was a linear relationship between the logarithm of the molecular weights of the standard proteins and their migration coefficients; Preparation VII gave a single band the migration coefficient of which corresponded to a molecular weight of 28,000.
The molecular weight of the endonuclease was also determined by Sephadex G-75 filtration.
Preparation VI was dialyzed against Buffer III and 0.1 M NaCl, and a 2-ml aliquot was filtered through the calibrated Sephadex G-75 column (see "Materials and Methods"). Reference to the calibration curve given by the standard proteins indicated a molecular weight around 27,000.
Preparation VII (180 pg of protein) was dialyzed against water, lyophilized, hydrolyzed in HCl, and analyzed for amino acid content on an automatic JEOL JLC-ASH apparatus. The molar percentages were calculated from the specific absorption determined experimentally with pure amino acids (Table II). Substrate Specificity of Purified Enzyme -Preparation VI (16 enzyme units) was incubated with 2 pg of T7 [3H]DNA, either untreated, alkylated, or depurinated.
Each sample was then split in two parts; one was denatured with NaOH and the other with formamide before sedimentation on sucrose gradients. Table III indicates that the untreated DNA contained no alkali-labile sites (= apurinic sites), the alkylated  (A); sodium dodecyl sulfate @)I or urea (C) was incubated at 37" for 30 min before measuring the acid-soluble radioactivity. After correction for controls without enzyme, the results are expressed as acid-soluble fractions. 0, B. stearothermophilus enzyme (Preparation VI); 0, E. coli enzyme (Preparation V of Verly and Rassart (9)).
DNA contained, per strand, an average of 5 of them (for 350 alkylated sites; see "Materials and Methods"), and the depurinated DNA an average of 26. Comparison of the formamide results, with and without enzyme, indicates that the alkalilabile sites have been hydrolyzed by the enzyme. On the other hand, the results were the same after a treatment with NaOH whether the enzyme was present or not. This clearly indicates that the enzyme was active only on alkali-labile sites; it thus had no action on normal strands or at alkylated sites ( i .e no activity of the N-glycosidase or of the endonuclease type).
The action of Preparation VI on heat-denatured DNA was also investigated; there was none on the normal strands, and the enzyme action on depurinated DNA was considerably decreased by the denaturation of the substrate. Preparation VI did not release acid-soluble material from sonicated DNA or DNA activated with pancreatic deoxyribonuclease (EC 3.1.4.5) (Sigma Chemical Co.); the preparation was thus devoid of an activity comparable to that of exonuclease III in Escherichia coli. Factors Affecting Catalytic Properties ofPurified Enzyme -Using acetate/barbital buffers, the optimal pH was found to be 7.5 for Preparation VI; the activity curve relative to pH was quite similar to that obtained with the crude extract. The optimal temperature, in the presence of 25% glycerol, was 45" and the enzyme was more labile after purification than in the crude extract. The rate of denaturation of Preparation VI in Hepes buffer, 25% glycerol, was studied at different temperatures; it followed first order kinetics with half-lives of 20 min at 40", 15 min at 50", 11 min at 60", and 8 min at 70". The addition of 0.4% bovine serum albumin protected the enzyme which could then be heated at 60" for 120 min without loss of activity.
The enzyme activity was influenced by monovalent cations. It passed through a maximum for 0.125 M KC1 (440% of the value without KCl) and 0.05 M NaCl (140% of the value without NaCl). Addition of EDTA to Preparation VI up to 10 mM did not inhibit the enzyme activity. On the other hand, Table IV shows that the divalent cations decreased the endonuclease activity.
We compared the action of several denaturation agents on the activity of the B. stearothermophilus enzyme (Preparation VI) and of the main E. coli endonuclease for apurinic sites (Preparation V of Verly and Rassart (9)). Up to 2% Triton X-100 had no effect on either enzyme, but Fig. 4 shows that the E. coli endonuclease for apurinic sites was much more sensitive to sodium deoxycholate, sodium dodecyl sulfate, or urea than the B. stearothermophilus enzyme.

DISCUSSION
An endonuclease which hydrolyzes DNA containing apurinic sites has been purified from Bacillus stearothermophilus cells. The final product appeared as a single protein band in a polyacrylamide gel electrophoresis performed in the presence of sodium dodecyl sulfate. The molecular weights of the native enzyme measured on Sephadex G-75 and of the sodium dodecyl sulfate-denatured enzyme determined by gel electrophoresis were nearly the same; the enzyme thus appears to be a monomeric protein of about 28,000 daltons.
With double-stranded DNA, the pure B. stearothermophilus enzyme is strictly specific for apurinic sites. This was shown using T7 phage DNA labeled with tritium and, after denaturation, the sucrose gradient centrifugation technique. Table  III shows that the enzyme introduced no break in untreated DNA, and that the action on alkylated DNA was restricted to the alkali-labile sites. Because NaOH does not produce breaks near alkylated sites (221, the conclusion is that the enzyme has no action on alkylated sites, either of the endonuclease type (see results after denaturation with formamide) or of the N-glycosidase type (see results after NaOH denaturation). The pure enzyme did not degrade sonicated DNA or DNA nicked with pancreatic deoxyribonuclease; it is thus without an activity similar to that of Escherichia cob exonuclease HI. The activity of the endonuclease of B. stearothermophilus on apurinic sites is considerably decreased when the DNA is denatured.
Because of its thermoresistance, absence of inhibition by EDTA, and absence of associated exonuclease activity, the endonuclease specific for apurinic sites of B. stearothermophi-Zus resembles more closely the endonuclease IV of E. cob (5) than the main endonuclease specific for apurinic sites found in this bacterium by Verly and Paquette (3,4) Fig. 4 shows that the E. coli enzyme is more readily denatured by urea and detergents. An amino acid analysis revealed a higher percentage of hydrophobic amino acids in the B. stearothermophilus endonuclease (Table II); the polarity index, calculated according to Vanderkooi and Capaldi (23), is 54.1% for the E. coli enzyme and only 44.2% for that of the thermophilic bacterium. Possibly a more important hydrophobic core might be responsible for the higher resistance to denaturation of the B. stearothermophilus endonuclease for apurinic sites.