Hypoxanthine-DNA glycosylase from Escherichia coli. Partial purification and properties.

Hypoxanthine-DNA glycosylase from Escherichia coli was partially purified by ammonium sulfate fractionation and by chromatography on Sephacryl S-200, DEAE-cellulose, and phosphocellulose P-11 columns. Analysis of the enzymatic reaction products was carried out on a minicolumn of DEAE-cellulose and/or by paper chromatography, by following the release of the free base [3H]hypoxanthine from [3H]dIMP-containing phi X174 DNA. In native conditions, the enzyme has a molecular mass of 60 +/- 4 kDa, as determined by gel filtration on Sephadex G-150 and Sephacryl S-200 columns. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis analysis revealed a major polypeptide band of an apparent molecular mass of 56 kDa, and glycerol gradient centrifugation indicated a sedimentation coefficient of 4.0 S. Hypoxanthine-DNA glycosylase from E. coli has an obligatory requirement for Mg2+ and is totally inhibited in the presence of EDTA. Co2+ can only partially replace Mg2+. The enzyme is inhibited by hypoxanthine which at 4 mM causes 85% inhibition. The optimal pH range of the enzymatic activity is 5.5-7.8, and the apparent Km value is 2.5 x 10(-7) M.

, and hypoxanthine-DNA glycosylase (Hyp-DNA glycosylase)' (4,24,25). These DNA glycosylases are small monomeric enzymes of 19-30 kDa in molecular mass and have been reported not to require cofactors such as divalent metal cations for their enzymatic activity (7). In this paper, we describe the partial purification and characterization of Hyp-DNA glycosylase from Escherichia coli. Contrary to previous reports on this activity (24), we have found that the enzyme has a molecular mass of about 60 kDa. It has an obligatory requirement for M$+ and is inhibited in the presence of EDTA.
[3H]dITP was prepared by deamination of [3H]dATP with 3 M nitrous acid (pH 3.5) for 14 h at 37 "C. To analyze the deamination products, an aliquot of the reaction was loaded onto TLC MN300 cellulose with authentic markers (dATP, dITP) and run in saturated ammonium sulfate, 1 M sodium citrate, isopropyl alcohol (80:18:2 v/ v/v). Under these conditions, 95-98% of the [3H]dATP was converted to [3H]dITP. On completion of the reaction, the salt was removed from the nucleotide by gel filtration on a column of Sephadex G-10 (1 X 4 cm) in 10 mM Tris-HC1 (pH 7.6) and 1 mM EDTA.

Substrate Preparation
The substrate 6x174 RF DNA (5 pg) was nick-translated in the presence of [3H]dITP according to Davids et al. (27). The reaction mixture (50 pl) contained 50 mM Tris-HC1 (pH 7.5), 10 mM MgSO,, 50 pg/ml BSA, unlabeled dATP, dCTP, and dTTP (20 p~ each), 2 p M [3H]dITP, 5 units of E. coli DNA-polymerase I, and 1 pl of DNase-I from a stock solution of 1 pg/ml. After incubation for 30 min at 14 'C, the reaction was stopped by the addition of 5 p1 of 0.2 M EDTA, followed by incubation for 5 min at 65 "C to inactivate the DNase-I. The reaction mixture was loaded onto a Sephadex G-50 column (0.7 X 20 cm), equilibrated with 10 mM Tris-HC1 (pH 7.5) and 1 mM EDTA. The nick-translated DNA was eluted in the void volume in 95% yield and had a specific activity of 230,000 cpm/pg. This substrate will hereafter be referred to as [3H]dIMP-4X174 DNA. The standard reaction mixture (50 pl) contained 10 mM sodium and a limited amount of enzyme. After 15 min at 39 'C, the reaction was terminated by adding 3 pl of hypoxanthine from a stock solution of 10 mg/ml, and the mixture was chilled in an ice-water bath.
One unit of hypoxanthine-DNA glycosylase activity is defined as the amount of enzyme that catalyzes the release of 1 pmol of hypoxanthine/min at 39 "C.
Analyses of the Enzymutic Reaction Products DEAE-cellulose Analysis-For routine analysis, the enzymatic mixture was loaded onto a column filled with 0.7 ml of DEAE-cellulose (DE-52 Whatman) equilibrated with 10 mM sodium acetate (pH 4.4). The enzymatically released hypoxanthine was eluted with 2.5 ml of 10 mM sodium acetate (pH 4.4) and counted in a Beckman scintillation counter.
[3H]dIMP-$X174 DNA and free [WIdIMP, a product of exonuclease activity (see below), remained bound to the column and could be released for control purposes by washing the column with 1 M sodium acetate (pH 4.4).
Paper C h r o~t~r~h y Analysis-The enzymatically released hypoxanthine was also analyzed by paper chromatography in a solvent system that allowed the separation of hypoxanthine, deoxyinosine, and dIMP from one another (24). The reaction mixture (50 pl) was chilled to 0 "C and 3 pl of 10 mgfml hypoxanthine, 5 pi of 10 mg/ml deoxyinosine, 5 pl of 10 mg/ml dIMP, 5 pl of 2 mg/ml calf thymus DNA, 7 pl of 5 M NaCl, and 220 pl of ethanol (stored at -20 "C) were added. After 1 h at -70 "C, the sample was centrifuged at 12,000 X g for 15 min. The supernatant was removed and concentrated under reduced pressure to about 30 rl, applied to Whatman No. 3MM paper, and chromatographed for 16-20 h in a solvent system containing the upper phase of a mixture of ethyl acetatefn-propyl alcohol/H20 ( 4 1 2 v/v/v). The markers were localized as ultraviolet absorbing materials after the paper had been dried. Strips containing individual samples were then cut transversely into 2-cm pieces, transferred into vials containing 5 ml of hydrophobic scintillation liquid, and counted. This procedure was also used to estimate the amount of exonucleases in the various fractions by following the appearance of [3H]dIMP in the reaction mixture. The identity of hypoxanthine, deoxyinosine, and dIMP in the paper chromatographic analysis was confirmed by high pressure liquid chromatography analysis on a C-19 column (Waters) using 5% methanol in 20 mM NH,H2P04 as solvent.
Reduction with NaBH4-One microgram of dIMP-$X174 DNA was treated with active enzyme preparation (Fraction VI, see below) under the standard reaction conditions. The DNA was isolated by ethanol precipitation, dissolved in 0.1 ml of 0.1 M sodium borate (pH 9.2), treated with 0.75 mg of NaBH, (IO i.1 from a stock solution of 75 mg/ml in 50 mM NaOH), incubated for 1 h a t 0 "C, and neutralized with dilute phosphoric acid. The reduced DNA was dialyzed against 0.15 M NaCl, 0.015 M sodium citrate and isolated by ethanol precipitation.
Treatment with Exonuclease IZZ and Agarose-Gel Ekctrophoresis-DNA samples, dissolved in 10 pl of 50 mM Tris-HC1, 20 m M CaClz (pH 7.5), were treated with 5 units of exonuclease 111 (New England Biolabs) for 10 min at 37 "C and analyzed by alkaline agarose-gel electrophoresis as described previously (28). For positive control, we used reduced acid-depurinated DNA to demonstrate that in the presence of Ca2+ the exonucleolytic activity of exonuclease 111 is inhibited, while the AP endonucleolytic activity of the enzyme on apurinic DNA (29) and on reduced apurinic DNA (30) is maintained.

Bacterial Growth
A wild-type E. coli strain was grown to late log phase (5-6 h) in Lbroth medium at 37 'C with aeration and then harvested by centrifugation at 4000 X g for 5 min. The cell paste was stored for several weeks at -20 "C without losing the Hyp-DNA glycosylase activity.

Purification Procedures
Crude Ertract-Eighty grams of frozen bacteria were thawed in 150 ml of 10 mM sodium potassium phosphate buffer (pH 7.3), containing 1 mM 2-mercaptoethanol and 10% glycerol. NaCl was then added to 0.1 M and lysozyme to 0.1 mg/ml final concentration. The mixture was incubated for 15 min at 37 "C and immediately cooled in an ice-water bath. The debris was removed by centrifugation at 13,000 rpm in a Sorvall SS-34 rotor for 30 min at 4 "C. The supernatant is designated Fraction I.
Streptomycin Sulfate-One part of 10% streptomycin sulfate in phosphate buffer (pH 7.3) was slowly added to nine parts of Fraction I with gentle stirring. After 1 h at 4 "C, the precipitate was removed by centrifugation as above. This step removed most of the nucleic acids present in the crude extract.
First Ammonium Sulfate Fractionation-Solid ammonium sulfate (enzyme grade) was slowly added, with constant stirring, to the s t r e p~m y c i n -t r e a~d Fraction I until 42% saturation was reached.
After 60 min at 4 "C the precipitate was collected by centrifugation as above. The precipitate was resuspended in enzyme buffer (10 mM Hepes/NaOH (pH 623, 20 mM NaC1, 1 mM 2-mercaptoethanol, and 10% glycerol (v/v)) and dialyzed for 5 h against two changes of the same buffer. The dialysate is designated Fraction 11.
Sephucryl 5-200 Chroma~ogra~hy-Fraction I1 was applied to a column (3 X 100 cm) of Sephacryl S-200 superfine (Pharmacia LKB Biotechnology Inc.) equilibrated with enzyme buffer. The column was eluted with enzyme buffer, and fractions of 6 ml were collected. Column eluates were monitored by UV absorbance at 280 nm, and fractions were assayed for enzymatic activity. Fractions 44-52 (see Fig. 3), which contained most of the activity, were combined and designated Fraction 111.
Second Ammonium Sulfate Fractio~tion-Fraction 111 was treated with ammonium sulfate (65% saturation) and centrifuged as described above. The precipitate was resuspended in enzyme buffer and dialyzed for 12 h against three changes of enzyme buffer. The dialysate is designated Fraction IV.
DEAE-cellulose Fractionation-Fraction IV was applied to a column (20 X 3 cm) containing DEAE-cellulose (Whatman DE-52) equilibrated in enzyme buffer. The column was washed with 90 ml of enzyme buffer and eluted by a gradient (600 ml) of 20-1000 mM NaCl in enzyme buffer. Fractions of 9 ml were collected and assayed for enzyme activity and protein concentration. Fractions 28-33 (see Fig.  4) contained the enzymatic activity and were combined to form Fraction V.
Phsphocellulose Chromatography-Fraction V was concentrated by treatment with polyethyleneglycol-8000 (Sigma) or by ultrafiltration in an Amicon unit with a Diaflo PM-10 membrane. It was then dialyzed against enzyme buffer and applied to a column (5 x 1 cm) filled with phosphocellulose P-11 (Whatman) equilibrated with the same buffer. The column was eluted with 25 mi of a linear gradient of IO-1,000 mM NaCl in enzyme buffer. Hyp-DNA glycosylase activity eluted at 400-500 mM NaCl. At this stage the enzyme was separated from exonuclease activities, as indicated by the absence of release of ['HIdIMP by fractions that released hypoxanthine.

S~S -P o l y a c~~m~e Gel ElectrophoresQ
For SDS-PAGE we employed the gel and buffer system of Laemmli (32). Gels were run at room temperature at 150 V and then stained for 30 min at room temperature with Coomassie Brilliant Blue R-250 (Sigma) in acetic acid/methanol/water (L4.54.5). The gels were then destained for 1-2 h at 45 "C in acetic acid/methanol/water (5:7.587). The destaining solution was changed four to five times.  The ethanol-soluble material was analyzed by descending paper chromatography as described under "Experimental Procedures." The UVabsorbing markers dIMP, deoxyinosine, and hypoxanthine are indicated. U , Fraction 11, ammonium sulfate precipitate of the crude protein extract; W, Fraction VI, second ammonium sulfate precipitation, followed by phosphocellulose chromatography.

Assay of Enzymatic Activity
Since the enzymatic activity of Hyp-DNA glycosylase requires the presence of M$+ ions (see below), a concomitant activity of exonucleases was expected. To distinguish between these activities, we analyzed the monomeric reaction products by paper chromatography under conditions that enabled the separation between hypoxanthine and dIMP, the products of Hyp-DNA glycosylase and exonuclease activities, respectively. Fig. 1 illustrates a typical paper chromatographic analysis. It can be seen that the ammonium sulfate precipitate of the crude protein extract (Fraction 11) contains an exonuclease activity that is responsible for the release of about 40% of the monomeric products as dIMP (solid circles). However, upon further purification (Fraction VI), the amount of exonuclease activity was reduced to less than lo%, while that of Hyp-DNA glycosylase activity increased, respectively (open circles). Fig. 1 also shows that deoxyinosine is not liberated from the substrate DNA. Therefore, hypoxanthine is the only monomeric product released from the DNA that is expected not to bind to DEAE-cellulose. We thus use the flow-through fraction of the DEAE-cellulose minicolumn (see "Experimental Procedures") for routine assays of Hyp-DNA glycosylase activity. Fig. 2 indicates that the release of free base is accompanied by the formation of an apurinic site in the DNA. Hypoxanthine-containing DNA that had been treated with an active enzyme preparation (Fraction VI) was partially hydrolyzed by alkali (Fig. 2, lane 3 ) , while reduction with NaBH, subsequent to enzyme treatment rendered the DNA insensitive to alkaline hydrolysis (Fig. 2, lane 4 ) . The same DNA preparation, on the other hand, was hydrolyzed by the AP endonuclease activity of exonuclease I11 (Fig. 2, lane 5 ) , under reaction conditions that inhibit the exonuclease activity of the enzyme of an untreated DNA substrate (Fig. 2, lane 2 ) but maintain its AP endonuclease activity on reduced apurinic DNA (29,30, and data not shown).

Purification Procedure
Several attempts to purify the Hyp-DNA glycosylase activity to homogeneity failed, mainly due to its scarcity and apparent instability. The main difficulty in purifying Hyp-DNA glycosylase was that in most fractionation steps, the enzymatic activity could not be separated from the bulk of the proteins in the extracts. The purification procedure that gave the most satisfactory results is summarized in Table I. It consisted of an essential streptomycin sulfate treatment to remove the nucleic acids, followed by ammonium sulfate  precipitation. Separation of most of the proteins was achieved by exclusion chromatography through Sephacryl S-200 (Fig.  3). About 60% of the proteins were eluted in the excluded volume, while the Hyp-DNA glycosylase activity was eluted in the included volume, slightly after a marker of BSA. A second ammonium sulfate fractionation was then employed in which about half of the remaining proteins that accompanied Hyp-DNA glycosylase were removed with almost no loss of the enzymatic activity. Further enrichment was achieved by anion exchange chromatography (Fig. 4), which separated 70% of the bulk proteins while retaining about 80% of the remaining activity. Following the final step of cation exchange chromatography, the enzymatic activity was almost free of accompanyingproteins (Fig. 5 ) . However, at this stage, the enzyme became labile as manifested by the low recovery yield of the enzymatic activity. For storage purposes, the eluate from the phosphocellulose column was concentrated 10-fold by ultrafiltration, and glycerol was added to 50%. No apparent loss of enzymatic activity could be detected after 6 months of storage at -20 "C.   Table I Fraction VI was also devoid of AP endonucleases, as shown by alkaline agarose-gel electrophoresis of NaBH4-reduced apurinic DNA that had been treated with enzyme preparations as described by Weinberger and Sperling (28). We observed that reduced apurinic DNA treated by Fraction VI was not degraded by the alkaline treatment, while treatment with the crude extract completely degraded the DNA (data not shown). The depletion of exonucleases from Fraction VI was demonstrated by paper chromatographic analysis (Fig. 1) which shows that the amount of dIMP released by Fraction VI is diminished, relative to Fraction 11, with a concomitant increase of the amount of hypoxanthine.
The protein composition of the fractions obtained during the different stages of purification was analyzed by SDS-PAGE (Fig. 6). Lanes b-f show the protein profile of Fractions I-V (see Table I). Lanes g and i contain duplicate samples of the phosphocellulose purification step (Fraction VI). Following this step, most of the bulk proteins were removed, leaving a band of a major protein of an apparent molecular mass of 56 kDa.

Native Molecular Weight Determination
Native molecular mass was determined by gel filtration and glycerol gradient centrifugation. Fig. 7 shows the elution profile of a sample from the DEAE-cellulose purification step (Fraction V) on a calibrated column of Sephadex G-150. Hyp-DNA glycosylase eluted with an apparent molecular value of 60 & 4 kDa as compared to elution volumes of the indicated molecular mass markers.
To estimate the sedimentation coefficient of Hyp-DNA glycosylase, a sample from Step VI (Table I) Table I) was sedimented along with three protein markers through a 5-ml linear glycerol gradient (20-40%) in a Beckman SW 50.1 rotor for 22 h at 48,000 rpm and 2 "C. Fractions were collected from the bottom of the tube and assayed for Hyp-DNA glycosylase activity. Lowerpanel, Hyp-DNA glycosylase activity across the gradient; upper panel, calibration curve using the following protein markers as indicated by the arrows: 1, beef liver catalase (11 S); 2, BSA (4.3 S); 3, lysozyme (2.11 S).

TABLE I1 Effect of divalent cations on hypoxanthine-DNA glycosylase activity
The enzymatic activity was assayed as described under "Experimental Procedures." Magnesium cations were included in the assay only where indicated. The 100% activity is defined as the amount of [3H]hypoxanthine released by 2 microunits of enzyme from purification step I11 (Table I) (Table I). tivity. The sedimentation coefficient of the enzyme was estimated to be 4.0 k 0.2 S (Fig. 8).

Z L
Enzymatic Properties of Hyp-DNA Glycosylase Substrate Specificity-To test the activity of Hyp-DNA glycosylase on single-stranded DNA, we used heat-or alkalidenatured [3H]dIMP-dX174 DNA as substrate. S1 nuclease analysis, described by Bryant and Lehman (34), confirmed the persistence of more than 95% of this substrate in a singlestranded form under the reaction conditions. We have found that Hyp-DNA glycosylase acts about twice as fast on singlestranded DNA as on double-stranded DNA. On the other hand, the enzyme does not act on monomeric derivatives of hypoxanthine as indicated by the inability of Hyp-DNA glycosylase to release hypoxanthine from dITP.
pH Optimum-The enzymatic activity was determined in the pH range of 4-10 in phosphate buffer at a constant ionic strength of 25 mM. The optimal activity of Hyp-DNA glycosylase is in the pH range of 5.5-7.8.
Effect of Divalent Cations-The enzymatic activity of Hyp-DNA glycosylase has an obligatory requirement for magnesium cations and is totally inhibited in the presence of 1 mM EDTA. Fig. 9 shows the enzymatic activity as a function of M$+ concentration. The minimal M$+ concentration required for optimal enzymatic activity is 4 mM. Table I1 summarizes the effect of other divalent cations. Mn2+, Zn2+, and Ca2+ did not replace M$+, while Co2+ could only partially replace the requirement for M$+.
Inhibition of Hyp-DNA Glycosylase by Hypoxanthine and Related Purine Deriuatiues-Hyp-DNA glycosylase is inhibited by the end product of the reaction, as has been found for uracil-DNA glycosylase (9). Hypoxanthine at 2 mM caused  (Table I). 50% inhibition of the enzymatic activity, whereas 4 mM caused 85% inhibition (Fig. 10). Other purine derivatives did not inhibit the enzymatic activity. Optimal Temperature of Hyp-DNA Glycosylnse Activity-The effect of temperature on the enzymatic activity was determined by employing the standard assay at the temperature range between 4 and 55 "C. The optimal temperature for the activity of Hyp-DNA glycosylase is about 39 "C. The observed release of hypoxanthine at temperatures higher than 40 "C is due to residual enzymatic activity, since incubation of the substrate at these temperatures in the absence of an enzyme did not cause the release of hypoxanthine.
Heat Stability-The effect of heat treatment on the enzymatic activity was tested on Fraction VI (phosphocellulose column). The half-life of the enzymatic activity at 55 "C is 2 min. The enzyme remained fully active for several days at 4 "C in enzyme buffer, and for several months at -20 "C in enzyme buffer containing 50% (v/v) glycerol.
K, Value-The initial rate of hypoxanthine release was determined by the standard assay using DNA concentrations corresponding to a range of 7-20 nM [3H]dIMP. The K, of dIMP was estimated to be 2.5 X M (see Fig. 11).

DISCUSSION
This study presented the partial purification of an enzymatic activity from E. coli extracts that releases hypoxanthine from dIMP-containing DNA. Such activity was initially discovered by Karran and Lindahl in E. coli (24) and in calf thymus (25) and has been termed hypoxanthine-DNA glycosylase. However, attempts to obtain the enzyme in a substantially purified form have been hampered by its scarcity and apparent instability. Functionally and mechanistically, the enzyme belongs to a group of enzymes that initiate the repair of DNA-containing altered bases by hydrolyzing the glycosylic bond between the altered base and the deoxyribose, leading to the formation of apurinic or apyrimidinic sites.
Hypoxanthine may occur in DNA in vivo, either by spontaneous deamination of adenine residues (25) or by misincorporation of dIMP residues from traces of dITP that might exist in the deoxynucleoside triphosphates pool. Although the latter possibility is less probable (25), the incorporation of dITP into DNA in vitro by E. coli DNA polymerase I (3,35) and by DNA polymerase a from HeLa cells (4) indicates that such incorporation might also occur in vivo. For our studies, we have employed DNA that was nick-translated in the presence of [3H]dITP instead of dGTP. This substrate corresponds to DNA damaged by misincorporation and contains a stable Watson-Crick I:C base pair, rather than an I:T pair that results from deamination and may exist in the less stable "Wobble" configuration.
Hypoxanthine-DNA glycosylase was partially purified by ammonium sulfate fractionation, gel filtration, and ion-exchange chromatography. The partially purified enzyme preparation has been shown to be free of contaminating deoxyribonucleases that accompanied the initial stages of purification. In nondenaturing conditions, the enzyme exhibited a molecular mass of 60 kDa as determined by gel filtration methods. This observation was substantiated further by estimating the sedimentation coefficient of the enzyme using glycerol gradient centrifugation. The results show that the enzymatic activity (4.0 S) migrated close to BSA (4.2 S), which has a molecular mass of 68 kDa. Analysis of the purified preparation under denaturing conditions by polyacrylamide gel electrophoresis in the presence of SDS revealed a major band of 56 kDa (Fig. 6, lanes g and j ) . The sedimentation and gel filtration data are consistent with the assignment of this band to Hyp-DNA glycosylase. However, the presence of minor quantities of multiple polypeptide species in the enzyme preparation does not exclude other possibilities, for example, the persistence of multiple subunits in the 60-kDa native enzyme. We have also shown that the enzyme has an obligatory requirement for M$+ and is totally inhibited in the presence of EDTA. Inhibition of the enzymatic activity could also be achieved by hypoxanthine, the end product of the enzymatic reaction, at a concentration of 5 mM.
The hypoxanthine-DNA glycosylase described here differs from the previously described enzyme (24,25) in three features. First, its molecular mass appears to be 56 kDa, whereas the molecular mass of the previously described activity is 30 kDa. Second, the enzymatic activity described here is M$+dependent, while the previously described one has no requirement for M$+ or other divalent cations as cofactors. Third, end product inhibition is observed for the presently described activity, whereas it did not occur previously. As far as the first two features are concerned, our hypoxanthine-DNA glycosylase also differs from other DNA glycosylases that have been collectively described as polypeptides of 20-30 kDa in molecular mass that do ,not require M$+ for their activity. However, Micrococcus luteus uracil-DNA glycosylase (10) and E. coli 3-methyladenine DNA glycosylase (13) are inhibited by the respective free base end product, while 3-methyladenine from calf thymus has been reported not to be inhibited by 3-methyladenine (36).
The apparent discrepancy with respect to the requirement