Purification and Properties of an Endodeoxyribonuclease from Nuclei of Bovine Small Intestinal Mucosa*

An endodeoxyribonuclease has been purified from nuclei of bovine small intestinal mucosa to a homoge- neous state by a procedure involving affinity chromatography on heparin-agarose. The endonuclease, which was found to be bound to chromatin, has a pH optimum of 5.4. It requires Mn2+ or Coz+ for activity and its maximum activity with Md+ is about 80% of that with Mn2+. Its activity is strongly inhibited by sulfhydryl- blocking agents, and by ethidium bromide. The enzyme does not attack RNA and is inhibited by it. Its isoelec- tric point is 8.5 & 0.1, and its molecular weight is 49,000 zk 3,000, determined by sucrose gradient sedimentation and gel filtration on Sephadex G-100. Polyacrylamide gel electrophoresis in the presence of sodium dcdecyl sulfate indicated that the enzyme is composed of two nonidentical subunits with molecular weights of 30,000 and 23,000. "he enzyme catalyzes the endonucleolytic cleavage of circular duplex ColEl DNA via single strand scissions from the initial stage of degradation. The average size of the limit products of native phage T7 or Cowl DNA is about 2,000 to 1,500 base pairs, estimated by neutral sucrose sedimentation or agarose gel electrophoresis. The enzyme degrades denatured DNA about 20 times faster than native DNA. The products contain 5'-phosphoryl

by polyacrylamide gel electrophoresis and calculated that they had multiples of a molecular weight of 200 base pairs.' Similar degradation products of DNA have been found in terminally differentiating fetal lens fiber cells (3). As a next step, we examined the endonuclease activity in the nuclei of the small intestinal mucosa responsible for breaking the nucleosome linker of chromatin in the initial stage of degradation of chromatin DNA.
This paper describes the purification and some properties of an endonuclease activity from the nuclei of bovine small intestinal mucosa. Bovine mucosa was used instead of rat tissue, because large quantities of mucosal scrapings were required to obtain sufficient enzyme from the nuclei for study. The endonuclease seems to bind to chromatin and appears to be a nicking enzyme.
The possible function of this endonuclease in uiuo, and its participation in DNA synthesis in bovine small intestinal mucosa are discussed.

EXPERIMENTAL PROCEDURES
Substrates 32P-labeled Escherichia coli DNA was prepared as described previously (5). 32P-labeled and 3H-labeled T7 DNAs were prepared by the methods of Friedman and Smith (6), and Richardson (7), respectively. "P-labeled E. coli ribosomal RNA was prepared according to the method of Littauer et al. (8). ColEl DNA (Form I) and phage h DNA digested with restriction endonuclease Hind111 were & t s of Dr.
Y. Sakaki, and phage fd DNA was a gift of Dr. T. Takeya. DNA concentrations were expressed as nucleotide residues. Ty3'P]ATP was the method of Spudich and Watt (11). E. coli RNA was obtained from Mann. Dextran sulfate (Mr = 15,000) was from Pharmacia and carrier ampholite (pH 7 to 10) from LKB. All other chemicals were of the highest grade available commercially and were used without further purification.

Methods
Preparation of Nuclei from Bovine Intestinal Mucosa-All operations were performed at between 0 and 4°C. Bovine small intestine was obtained at a local slaughterhouse and brought to the laboratory on ice. The intestine was chopped into 30-cm segments after removing the fat. The segments were cut open longitudinally and washed briefly with cold 0.5% NaC1,0.5% KC1 and then thoroughly with cold Krebs-Ringer phosphate, pH 7.4, containing 6% dextran to remove mucin and remaining food particles. The mucosa was then obtained by scraping the intestine with the edge of a microscope slide glass by the method of Clark and Porteous (12).
Nuclei were prepared from mucosal scrapings by a modification of the method of Henner et al. (13). Tissue (wet volume 400 ml) was suspended in 3 volumes of 0.3 M sucrose containing 10 nm Tris/HCl, pH 8.0, 5 nm magnesium acetate, 3 mM CaCL, 0.5 mM dithiothreitol, and 0.1% Triton X-100, and homogenized with five strokes of a Teflon pestle in a Potter homogenizer. The homogenate was centrifuged at 1,000 X g for 10 min. T h e pellet was washed twice by suspending it in 3 volumes of 0.3 M sucrose buffer, homogenizing it in a Waring blendor at low speed for 2 min, and centrifuging the homogenate as described above. The final pellet was suspended in 5 volumes of 2.25 M sucrose containing 10 nm Tris/HCl, pH 8.0, 5 mM magnesium acetate, and 0.5 mm dithiothreitol. The suspension was layered over an equal volume of 2.25 M sucrose buffer and centrifuged at 45,000 X g for 45 mi n. The resultant pellet was washed three times by suspension in 0.3 M sucrose containing 10 nm Tris/HCl, pH 8.0, and 2 mM MgCL and then stored at -30°C as a pellet.
Preparation of DNA-Polyacrylamide Gel-"P-labeled E . coli DNA incorporated into polyacrylamide gel was used as a substrate for assay of the endonuclease activity. The DNA-gel suspension was prepared by the method of Melgar and Goldthwait (14) with a slight modification as follows. Air was removed under vacuum from a mixture of 2 ml of acrylamide stock solution (10% acrylamide, 2.5% bisacrylamide), 1 ml of 32P-labeled E. coli DNA (30 nmol/ml of DNA in 0.02 M KCI), and 0.25 ml of riboflavin solution (40 mg/ml). When polymerization under fluorescent lighting was complete, the DNA-gel was homogenized in a Waring blendor at low speed for 1 min with 10 volumes of 0.05 M Tris/HCl, pH 7.2, 0.05 M KCl, and 0.2 m EDTA. The homogenate was then passed through a 20-1111 syringe twice. DNA-gel was washed five times each by suspension and centrifugation at 15,000 X g for 5 min, first with 10 volumes of the same buffer, then with 10 volumes of 0.05 M Tris/HCl, pH 7.2, containing 0.02 M KC1 and 0.2 m~ EDTA, and finally with 10 volumes of 0.01 M potassium phosphate buffer, pH 7.0, containing 0.01 M KC1. The DNA concentration of the final DNA-gel suspension was adjusted to 15 nmol/ml. The yield of radioactivity in the DNA-gel was 40 to 60% of that of the input DNA. When heat-denatured DNA was incorporated in the polyacrylamide gel, DNA was released slowly during storage, so the gels were washed with the final buffer at least once a week.
Assay for Endonuclease Actiuity-The reaction mixture (0.3 m l ) contained 50 mM acetate buffer, pH 5.4, 3 mM MnCI2, 30 pg of bovine serum albumin, 0.1 ml of DNA-gel suspension (about 25,000 cpm) containing 3.3 mM potassium phosphate and 3.3 nm KC1, and enzyme preparation. After incubation for 30 min at 37'C, the reaction was terminated by adding 0.7 ml of 0.02 M EDTA, pH 8.0, at 0°C. The tubes were centrifuged at 2,000 X g for 10 min, and 0.5 ml of the supernatant was removed, dried in a planchet, and counted in a gas flow counter. Reactions were always run in duplicate. In control tubes without enzyme, 1 to 2% of the input radioactivity was released during incubation. One unit of the enzyme was defined as the amount that released 1 nmol of DNA in 30 min at 37°C under these conditions. The release of 32P-labeled DNA from DNA-gel up to 40% of the input DNA was proportional both to the amount of enzyme added and to the time of incubation.
When the amount of acid-soluble DNA was measured, the reaction was terminated by adding 20 pl of 0.2 M EDTA at 0°C and the mixture was centrifuged as above. The supernatant was then transferred to another tube and mixed with 0.2 ml of egg albumin (5 mg/ml) and 0.5 ml of 0.5 N perchloric acid. The mixture was stood for 10 min at 0°C and then centrifuged, and the total supernatant was transferred quantitatively to a planchet, neutralized by adding 1 drop of 2 N KOH, dried in an oven, and counted as above. Unless otherwise noted, single-stranded DNA-gel was used as a substrate.
Isoelectric Focusing-Isoelectric focusing was performed according to the LKB instruction manual (15) with a 110-ml column using pH 7 to 10 carrier ampholite and a 0 to 50% sucrose gradient containing 10% glycerol. Enzyme solution previously dialyzed against 1 liter of 1% glycine containing 10% glycerol was mixed with the dense gradient solution. Electrophoresis was carried out at 450 V for 48 h at 4°C.
Gel Electrophoresis-Polyacrylamide gel electrophoresis was carried out at 4°C on a 7.5% polyacrylamide gel (0.6 X 7 cm) at 2 mA/gel with 0.45 M p-alanine/acetate buffer, pH 4.5, for 1% h, as described by Reisfeld et al. (16). After electrophoresis, the gel was stained with 0.2% Coomassie brilliant blue in 7% acetic acid and destained fvst with 30% ethanol in 10% acetic acid and then with 5% methanol in 7.5% acetic acid. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis was performed using a 10% gel column (0.6 X 7 cm) containing 0.1% sodium dodecyl sulfate by the method of Laemmli and Favre (17). After electrophoresis for 2 h at 3 mA/gel at room temperature, the gel was stained as above.
ColEl DNA was applied to a 0.9% agarose disc or slab gel column containing 0.5 pg/ml of ethidium bromide and electrophoresis was carried out at 150 V for 2% h or at 50 V for 18 h at room temperature as described by Peacock and Dingman (18). DNA in the gel column was detected under fluorescent lighting and photographed.
Neutral and Alkaline Sucrose Density Gradient Centrifugations of 3ZP-labeled Phage T7DNA"Neutral and alkaline sucrose gradient sedimentations of DNA were performed in a 5 to 20% sucrose gradient containing 20 m~ Tris/HCl, pH 8.2,0.25 M NaCl, and 1 mM EDTA and in a 5 to 20% sucrose gradient containing 0.25 M NaOH and 1 mM EDTA, respectively. The gradients were centrifuged for 2% h or 4% h at 45,000 rpm at 4°C in a Hitachi RPS 50 rotor.
5'-Terminal Nucleotide Analysis-Phage fd DNA (12 nmol) or ColEl DNA (5 nmol) was incubated with 1 unit of endonuclease at 37°C for 60 min in 0.1 ml of reaction mixture containing 50 mM acetate buffer, pH 5.4, 3 nm MnC12, and 10 pg of bovine serum albumin. A control without endonuclease was run in parallel. The reaction was terminated by heating the mixture for 5 min at 100°C and then rapidly placing it in an ice bath. The solution was dialyzed against 10 nm Tris/HCl buffer, pH 7.5, containing 0.1 mM EDTA for 5 h and then incubated for 90 min at 45OC with 20 pg of alkaline phosphatase. The reaction was terminated by cooling the mixture to 0°C. Then the mixture was treated three times with an equal volume of saturated phenol, and twice with an equal volume of ether. The DNA solution was dialyzed against 1 mM Tris/HCl buffer, pH 7.5, containing 0.01 mM EDTA overnight. Then it was phosphorylated with 4 units of polynucleotide kinase and 20 nmol of [yR2P]ATP (5 X 10' cpm/ nmol) in 0.2 rnI of reaction mixture containing 20 mM Tris/HCl buffer, pH 8.0, 10 nm dithiothreitol, 10 mM MgC12 and 1 mM spermidine for 40 min at 37°C as described by Takeya et al. (19); additional polynucleotide kinase (4 units) was added and the incubation was continued for another 20 min. The reaction mixture was then treated f i s t with phenol and then with ether twice as described above. The solution was applied to a Sephadex G-75 column (1.2 X 32 cm) previously equilibrated with 10 nm Tris/HCl buffer, pH 7.5, containing 0.1 nm EDTA. The 5"32P-labeled polynucleotides were obtained in the exclusion volume and were lyophilized. 5"3ZP-labeled polynucleotides were dissolved in 20pl of distilled water and incubated with 5 pg of pancreatic DNase I in 0.1 ml of reaction mixture containing 20 mM Tris/HCl buffer, pH 7.6, and 10 nm MgC12 for 90 min at 37°C. Then a final concentration of 70 nm glycine/NaOH buffer, pH 9.2, was added and the solution was treated with 4 pg of snake venom phosphodiesterase for 60 min at 37°C. The resultant mononucleotide solution was applied to Whatman No. 3" paper together with authentic samples of the four mononucleotides. Ascending paper chromatography was performed in a solvent system of 0.1 M sodium phosphate buffer, pH 6.8/ammonium sulfate/n-propyl alcohol, 100/60/2. Spots of nucleotides were located under ultraviolet light, and cut out for counting in a liquid scintillation counter.
Others-RNase activity was measured using "P-labeled E. coli RNA as a substrate by counting acid-soluble radioactivity. Micrococcal nuclease digestion of E. coli DNA was carried out as described by Webb and Felsenfeld (20). Molecular weights of proteins were measured by centrifugation on a 5 to 20% sucrose density gradient at 45,000 rpm for 18 h at 4°C in a Hitachi RPS 50 rotor. Fractions collected from the bottom of the gradient were assayed for endonuclease and hog spleen DNase I1 activities, respectively. Protein concentration was determined by measuring the absorbance at 280 nm, assuming Akmm = 1.00 for a solution of 1 mg of protein/ml.

Detection of Endonuclease Activity in Crude Extracts
Unless otherwise noted, centrifugation was carried out at 20, OOO X g for 10 min, and all operations were perfomed at 0-4°C. Crude extracts were eluted from nuclei stepwise with increasing NaCl concentrations as follows. The purified nuclear pellet was thawed in 3 volumes of Buffer A (20 mM Tris/ HCl, pH 7.5, 1 m~ EDTA, and 10 mM 2-mercaptoethanol) and disrupted by sonication in a Branson sonifier (model B-12) operated at maximum output three times for 1 min each. The disrupted nuclear suspension was stirred for 10 min, centrifuged, and the supernatant was saved. The pellet was resuspended in 3 volumes of Buffer A, homogenized in a Waring blendor at low speed for 3 min and at high speed for 5 min, stirred, and then centrifuged. The supernatants were combined and centrifuged at 105,000 X g for 1 h. The final 20,000 X g pellet was then suspended in 6 volumes of Buffer A containing 0.15 M NaCI, homogenized, stirred for 30 min, and centrifuged as above. These processes were performed successively in 6 volumes of Buffer A containing 0.3 M, 0.6 M, and 1.0 M NaCI. The resultant 105,000 X g supernatants were dialyzed extensively against Buffer B (20 m~ potassium phosphate buffer, pH 7.0, containing 0.1 mM EDTA, 5 m~ 2mercaptoethanol, 25 m~ NaCl, and 10% glycerol), and the slight precipitates formed were removed by centrifugation.
The endonuclease activity for degrading denatured DNAgel and the protein concentration of each fraction were measured (Fig. 1). The specific activity and yield of endonuclease activity were highest in the fractions extracted with 0.3 M to 0.6 M NaCI. No endonuclease activity which acted preferentially on native DNA-gel could be detected in any fraction, even when the assay was performed in the presence of Mg2' and Ca2' (21). This result shows that an endonuclease that acts preferentially on single-stranded DNA was present in nuclei in a chromatin-bound form.

Enzyme Purification
Preparation of Crude Extract-Nuclei (4 ml) were washed twice with 5 volumes of Buffer A and then extracted twice with 5 voiumes of Buffer A containing 0.6 M NaCl. The crude extract was dialyzed against 20 volumes of Buffer B, and the resultant precipitate was removed by centrifugation (Fraction I, 25 ml).

DEAE-cellulose Chromatography-Fraction I was applied
to a DEAE-cellulose column (1.2 X 6 cm) which had been equilibrated with Buffer B. The unadsorbed fractions were pooled (Fraction 11, 33 ml). Heparin-agarose Affinity Chromatography-Fraction I1 was applied to a heparin-agarose column (1.0 X 3 cm) previously equilibrated with Buffer B. The column was washed with 30 ml of the equilibrating buffer, and then material was eluted with 60 ml of a linear gradient of 0.025 M to 0.5 M NaCl in the same buffer. The flow rate was 3 ml/h and fractions of 3 ml were collected. The active fractions eluted with between 0.1 M and 0.25 M NaCl were pooled (Fraction 111, 18 ml) (Fig.  2).
Sephadex G-100 Gel Filtration-Fraction 111 was concentrated to 3.2 ml by dialysis against polyethylene glycol solution containing 20 m~ potassium phosphate buffer, pH 7.0, 5 mM 2-mercaptoethanol, 0.1 mM EDTA, 0.1 M NaC1, and 10% glycerol, and layered on a Sephadex G-100 column (5 X 1 0 0 cm) that had been equilibrated with the same buffer. It was then eluted with the same buffer at a flow rate of 20 ml/h. The active fractions were collected (Fraction IV, 60 ml), concentrated by dialysis against polyethylene glycol solution, and then dialyzed against 50 m~ potassium phosphate buffer, pH 7.0, containing 5 mM 2-mercaptoethanol, 0.1 m~ EDTA, 50 m~ NaCl, and 30% glycerol for 16 h. The results of a typical purification are summarized in Table I.
Although only 30-fold increase of specific activity was achieved, the starting material was extracted after washing the nuclei twice with Buffer A, which removed large quantities of protein. Thus the increase of specific activity is several hundredfold if calculated on the basis of that of the nuclear pellet. The total activity of the DEAE-cellulose fraction was always 1.5-to 2-fold greater than that of the crude extract.
This may be due to the existence of an inhibitor(s) for the endonuclease. The final preparation of enzyme lost 20% of its activity on storage for 1 month at -20°C. When NaCl was omitted from Buffer B, the enzyme activity was completely lost during the purification procedures. Unless otherwise noted, Fraction IV was used in all subsequent experiments.
Purity of the Enzyme-When Fraction IV was subjected to 7.5% polyacrylamide gel electrophoresis (pH 4.5), only one major band was evident (Fig. 3A), but the enzyme activity eluted from the gel was low. When an excess amount of  Fraction I11 was subjected to gel electrophoresis, the position of the enzyme activity corresponded with that of the major band of Fraction IV. Fraction IV had no detectable ribonuclease activity; when 5 nmol of "P-labeled E. coli rRNA (2,000 cpm/nmol) was incubated with 1% units of the enzyme for 60 min at 37"C, the formation of acid-soluble material was less than 0.5%. Properties of the Enzyme Molecular Weight-Th'e molecular weight of the purified enzyme was estimated by Sephadex G-100 gel filtration with marker proteins and by zone sedimentation in a 5 to 20% sucrose density gradient as described under "Methods." Values of 49,000 & 3,000 were obtained by these procedures assuming that the enzyme is a globular protein. Electrophoresis of the purified enzyme in the presence of 0.1% sodium dodecyl sulfate on 10% polyacrylamide gel showed that it was composed of two nonidentical polypeptide chains with molecular weights of about 30,000 and 23,000 (Fig. 3, B and C).
Isoelectric Point-The heparin-agarose fraction (Fraction 111) was used for determination of the isoelectric point. The sample (19 ml, 200 units) was dialyzed and subjected to electrofocusing as described under "Methods." After electrofocusing, fractions (2.0 m l ) were collected from the bottom of the column, and the endonuclease activity and pH of each were measured (Fig. 4). The isoelectric point of the enzyme was 8.5 & 0.1.
Requirements for Activity-The optimal pH was 5.4 in 50 mM sodium acetate buffer. In 50 mM sodium cacodylate buffer, pH 5.8, and 50 m potassium phosphate buffer, pH 6.0, the activities were 9 0 % and 30%, respectively, of the maximum itol stimulated the reaction about 20 to 40%. NaCl or potassium phosphate at 50 mM inhibited the enzyme activity about 80%. Ethidium bromide (1 pg/0.3 ml) completely inhibited the reaction, and E. coli RNA (10 pg/0.3 ml) also inhibited the activity 50%. Rabbit muscle G-actin, which is a specific inhibitor of pancreatic DNase I, had no effect. Bovine serum albumin (30 pg/0.3 ml) increased the activity about 2-fold. Comparison of the Activities on Native and Denatured DNA-Gels-The rates of hydrolysis of native and denatured DNA-gels were linearly proportional to the time of incubation with the enzyme (Fig. 5). The rate of degradation of denatured DNA-gel was about 20 times that of degradation of native DNA-gel. On incubation for 2 h, formation of acid-soluble nucleotides from denatured DNA-gel amounted to about 5% of the total, while none were released from native DNA-gel. When heat-denatured E. coli DNA (1.5 nmol) was incubated with excess endonuclease (3 units) for 12 h, about 30% of the input DNA became acid-soluble.
Mode of Action ofthe Enzyme-ColEl DNA (Form I) was digested with the enzyme and subjected to agarose slab gel electrophoresis as described under "Methods." The purified :I ; ; . the latter was labile during overnight dialysis against 1% glycine containing 10% glycerol at 4°C. Electrofocusing was carried out as described under "Methods." After focusing, fractions (2 m l ) were collected from the bottom of the column. The absorbance at 280 nm and the pH were determined and aliquots ( 6 0 pl) were assayed as described under "Methods," except that the incubation time was 150 min. 0, pH; M , enzyme activity.  (Form II), and then to a unit length of DNA (Form 111) (Fig. 6 ) . The above results indicate that the enzyme acts on double-stranded DNA endonucleolytically with single strand scissions. In order to elucidate whether the endonuclease acts on specific sites of DNA at the initial stage of attack, ColEl DNA was first digested with the enzyme, and then the digest was treated with Eco RI endonuclease. No new band was observed on agarose gel electrophoresis, indicating that the endonuclease did not act on ColEl DNA at specific sites (data not shown).
Size of Products of Hydrolysis of Native T7 DNA-The size of the products in limit digests of native T7 DNA was analyzed by neutral sucrose density gradient centrifugation (Fig. 7). The average molecular weight of the products was estimated as 1.35 X IO", which corresponded to about 2,000 base pairs (22). No further shift of the peak toward the top of the gradient was detected on increasing the incubation time. When the products of limit digestion were analyzed by alkaline DNA was digested with 1 unit of endonuclease under standard conditions. Aliquots were taken from the reaction mixture a t intervals and mixed with EDTA at a final concentration of 10 mM. The samples were centrifuged and analyzed as described under "Methods." The centrifugation profile was obtained from an aliquot taken after 60min incubation. "H-labeled T7 DNA was added as an internal marker. M , '"P; M , "H. sucrose gradient centrifugation, however, the peak fraction was shifted further to the top, and this material was calculated to contain several hundred nucleotide residues. When ColEl DNA was digested with excess endonuclease, the final product was calculated to be about 1,500 base pairs on agarose gel electrophoresis with a marker of phage X DNA digested with Hind111 endonuclease (data not shown). The above results indicate that limited numbers of single strand scissions were introduced into native T7 and ColEl DNA even after extensive digestion with the enzyme.
Digestion of E . coli DNA of Various Sizes with Endonuclease-To examine how the endonuclease acts on DNA of smaller size (less than 2,000 base pairs), we first digested E. coli DNA with various amounts of micrococcal nuclease. The size of the products was estimated by agarose gel electrophoresis and neutral sucrose density gradient sedimentation. Then the digested E. coli DNA of various sizes was further digested with the endonuclease and the final products were analyzed by alkaline sucrose density gradient sedimentation. When the size of E. coli DNA was about 20,000 bases, 16 nicks were introduced into the DNA (Fig. 8, a and b). On the other hand, when it was about 2,000 bases only one nick was introduced (Fig. 8, c and d).
Analysis of Acid-soluble Nucleotides of Denatured E. coli DNA-The size of the products of limit digestion of denatured E. coli DNA was analyzed by DEAE-Sephadex A-25 chromatography in the presence of 7 M urea (23). Only small amounts of mono-to pentanucleotides were produced and more than 90% of the acid-soluble nucleotides were bigger than 8 nucleotides (data not shown). h. Acid-soluble material was measured (about 20%) and then the incubation mixture was adjusted to pH 8.0, and divided in half. Onehalf was treated with 5 p g of alkaline phosphatase for 30 min at 45"C, and the other half was incubated in parallel without alkaline phosphatase. Aliquots of each mixture were digested with venom (4 p g ) or spleen (60 pg) phosphodiesterase under the respective optimal conditions. Aliquots taken at the indicated times were digested with 3 pg of alkaline phosphatase. The radioactivity of the supernatant was determined after Norit treatment. W, venom enzyme and alkaline phosphatase; M , venom enzyme, but not alkaline phosphatase; A-A, spleen enzyme and alkaline phosphatase; A-A, spleen enzyme, but not alkaline phosphatase.

TABLE 11
Identification of 5'-terminal nucleotides produced by endonuclease digestion Phage fd DNA (12 nmol) or ColEl DNA (5 nmol) was incubated with 1 unit of endonuclease for 60 min under standard reaction conditions. 5"Terminal nucleotides were identified as described under "Methods." A control without enzyme was run simultaneously and control values (about 10 to 30% of those of endonuclease-treated samples) were subtracted. Identification of the Terminal P of Digestion Products-Denatured E. coli DNA was subjected to limit digestion with the endonuclease, and then the reaction products were digested with calf spleen or snake venom phosphodiesterase (Fig. 9). The products were sensitive to the venom enzyme, but were not attacked by the spleen enzyme without pretreatment of alkaline phosphatase. These results showed that the reaction products were 5'-phosphoryl and 3"hydroxyl terminated nucleotides.

5"Deoxymononu
Identification of Nucleotides at the 5"Termini of Products-To confirm that the endonuclease has no base specificity in limit degradation of single-or double-stranded DNA, we treated phage fd DNA or ColEl DNA with the enzyme and labeled the 5'4ermini with [y-"PIATP using polynucleotide kinase after treatment with alkaline phosphatase. The products were then digested with pancreatic DNase I followed by snake venom phosphodiesterase, and the 5'-mononucleotides were identified by paper chromatography as described under "Methods." Results with both single-and doublestranded DNA showed that almost equal amounts of the four bases were present at the 5"termini (Table 11). Similar results were obtained with denatured E. coli DNA.

DISCUSSION
In this work we purified an endonuclease to homogeneity from nuclei of bovine s m d intestinal mucosa and studied its properties. The enzyme has several unique properties which distinguish it from other mammalian endonucleases reported previously, ( a ) On repeated washing, the enzyme activity remains associated with the nuclear pellet and is eluted best with 0.3 to 0.6 M NaC1. These results indicate that the enzyme is bound fairly tightly to chromatin, although not so tightly as non-histone proteins. Recently, two endonucleases associated with chromatin have been reported. One is Ca"+"g"-dependent DNase (21), which is eluted with 0.6 M NaCl from rat liver nuclei (24). This enzyme has not been characterized in detail, but it is strongly activated by Ca2' in the presence of Mg2+. In contrast, the bovine intestinal endonuclease described in this paper was inhibited by Ca2+ in the presence of Mn2+ or Mg2+. The other endonuclease reported was purified from HeLa S3 cells (25). It has a molecular weight of 22,000, a pH optimum of 7.2 and isoelectric point of 5.1 k 0.2. In contrast, the molecular weight, pH optimum, and isoelectric point of bovine intestinal endonuclease are 49,000 f 3,000,5.4 in 50 mM acetate buffer, and 8.5 f 0.1, respectively. (b) The results obtained by agarose gel electrophoresis of ColEl DNA clearly indicate that the enzyme makes single strand breaks without showing any base specificity at the initial stage of attack. The enzyme hydrolyzes denatured DNA about 20 times faster than native DNA. In this respect, the purified endonuclease has common characteristics with mammalian DNase V, which is located in nuclei and hydrolyzes denatured DNA about 5 times faster than native DNA (26). However, there are several differences in the properties of the purified endonuclease and DNase V. First, all the DNase Vs so far examined (26)(27)(28) were co-purified with DNA polymerase a, which is extracted from the cytosol or nucleokarysm fraction of cells. DNA polymerase a in rat small intestinal mucosa, although not characterized in detail, has been reported to be extracted from nuclei in the absence of NaCl (29). On the contrary, our purified enzyme is extracted from nuclei in the presence of 0.3 to 0.6 M NaCl. Second, calf thymus DNase V has a molecular weight of 53,000 f 2,500, and is composed of four identical subunits. Bovine intestinal endonuclease is, however, composed of two nonidentical subunits with molecular weights of 30,000 and 23,000. Third, the endonuclease copurified with DNA polymerase a from rat regenerating liver is not adsorbed to DNA-cellulose (28), while the endonuclease reported here is eluted with 0.15 M NaCl from heparin-agarose, which has many properties in common with DNA-cellulose (30). Thus, bovine intestinal endonuclease appears to be essentially a nicking endonuclease, and has some unique properties different from those of other mammalian endonucleases so far reported.
The size of limit digestion products of native T7 and ColEl DNA is about 2,000 to 1,500 nucleotides. The enzyme reported here cannot act on "'P-labeled E. coli DNA of smaller size (less than 1,000 bases). Why the endonuclease does not act on DNA of smaller size is still unknown, but there are several restriction endonucleases that recognize specific nucleotide sequences but are nonspecific in their cleavage (31,32), and in this respect, it should be further studied whether the purified endonuclease recognizes specific nucleotide sequences or not. Meanwhile, this property is interesting in connection with the physiological function of the endonuclease. The enzyme may well be associated with DNA synthesis, judging from its location in nuclei and its production of 3'-hydroxyl ends, which activate DNA polymerase a in uitro. It is reported that the size of Okazaki pieces in mammalian cells is about 100 nucleotides, while that of E. coli is 1,000 nucleotides (33). The distance between the two origins of replication in eukaryotic cells is about 10,OOO bases (34). If an endonuclease is needed for relieving the torque in unwinding duplex DNA before DNA polymerase initiates incorporation of nucleotides in mammalian cells, the enzyme should be an endonuclease such as that reported here which nicks native duplex DNA but does not act on DNA of smaller size (z.e. less than 100 nucleotides). The precise nature of the endonuclease, however, requires further characterization especially with respect to how the enzyme recognizes its action sites in the replicative form of chromatin.
Our purpose in isolating and characterizing this endonuclease from nuclei of bovine small intestinal mucosa was to identify the enzyme that breaks the nucleosome linker of chromatin during the initial stage of DNA degradation in situ.
We are now attempting to determine whether the endonuclease activity reported here is also present in the nuclei of rat small intestinal mucosa. It would be interesting to study how chromatin DNA is digested by the purified endonuclease in a reconstituted system in vitro.