Bovine Thymus Poly(Adenosine Diphosphate Ribose) Polymerase PHYSICAL PROPERTIES AND BINDING TO DNA*

Purified bovine thymus poly(adenosine diphosphate ribose) polymerase is a monomeric protein with a single polypeptide chain having a molecular weight of approximately 130,000, determined by sodium dodecyl sulfate-gel electrophoresis, analytical ultracentrifuga- tion, and gel filtration. A high frictional ratio (1.81) indicated that the molecule has an elongated shape, or a high solvation, or both. The enzyme is a basic protein (PI 9.8). and amino acid analysis showed a relatively high lysine content. The enzyme activity is dependent on double-stranded DNA and is solely correlated with single- or double-stranded breaks on the DNA. Filter binding assay tech- nique showed that the enzyme-activating efficiency of DNA correlated sufficiently with its enzyme-binding efficiency. Thus, a very high enzyme-activating efficiency of a DNA fraction (active DNA) which was separated from the crude enzyme fraction is mainly due to its high enzyme-binding efficiency. It was also shown that single-stranded DNA and heparin had a strong inhibitory effect on the binding of the enzyme to double-stranded DNA, whereas competitive inhibitors did not affect the binding. We interpret these results to indicate that the binding of the enzyme to double-stranded DNA is a prerequisite step to its catalytic activity and has a dual function: (a) to position the enzyme on specific binding sites such as single- or double-stranded breaks on the DNA, and (b) to induce an active conformation of the enzyme.

Poly(ADI'-ribose) polymerase of eukaryotic cell nuclei catalyzes polymerization of the ADP-ribose moiety of NAD' into a homopolymer of repeating ADP-ribose units, which is known to be bound to various nuclear proteins (Refs. 1-6; for a review, see Ref. 7).
Recently, the enzyme has been purified to near homogeneity from various tissues in several laboratories (8)(9)(10)(11)(12). T h e purified enzyme absolutely requires DNA for its activity, as previously reported by Yoshihara (13) and Yamada et nl. (14) with partially purified enzvme. In recent reports (8,15). we described that a DNA fraction (active DNA) which was separated at a purification stage of bovine thymus poly(AD1'ribose) polymerase has a high enzyme-activating ability, Although the precise mechanism of DNA which supports the enzvme activity is still unclear, treatment of isolated nuclei (16) o r permeable cells (17) with DNA-endonuclease was * This work was supported in part by a Grants-in-aid 301034 anti :tOIOHO for Cancer Research from the Ministry of Education. Science, a n t i Culture of ,Japan. 'I'he costs of publication of this article were tiefrayed i n part hy the payment o f page charges. This article must therefore be hereby marked "cr~/1,4rllsc,ment" i n accordance with 18 I:.S.C. Section 1734 solely t o indicate this fact.
$ ' I ' o whom correspondence should h e addressed.
reported to elicit a considerable increase in the enzyme activity, indicating that the enzyme is activated by DNA damage. This activation was also implicated in DNA repair (16)(17)(18)(19)(20).
In this report, a filter binding assay employed in the study of DNA-protein interaction (21) was used to examine the enzyme binding of active DNA and of various other DNAs including their DNase I-treated products. Physicochemical properties of the purified enzyme will also be presented here. NaNi. 1 mM glutathione, and 0.5 mM dithiothreitol. Buffer B contained 30 mM NaCI, 3 mM sodium citrate. 2 mM Tris-HC1 (pH 7.4). and 2 mM EDTA.

DNA Preparations
Calf thymus DNA (type 1, Lot 26C9560) was a product of Sigma. "Active DNA," a DNA fraction which was separated from the enzyme during the purification and has a verv high enzvme-activating effiof denatured DNA, calf thymus DNA was heated at 95°C for 10 min ciencv, was prepared as described previously (15). For the preparation and rapidly cooled. Covalently closed circular ColEl plasmid DNA was prepared from Is'. coli AS45 met thv (ColEl) according to the method of Tomizawa et al. (22), except that the culture was treated with chloramphenicol for 18 h. Thus obtained closed circular ColEl DNA preparation was still contaminated (approximately 10% of the total DNA) with E. coli chromosomal DNA which was approximately 4 S and had a high activation ability on poly(ADP-ribose) polymerase (15). In order to eliminate this DNA, we further purified it by a 5 to 25? sucrose gradient centrifugation in a buffer (0.1 M NaCI, 0.025 M Tris-HCI (pH 5.6 Table IV). The reaction was carried out at 25°C for 30 min and stopped by adding EDTA to 10 mM. After termination of these reactions, the mixture was phenolextracted once and dialyzed against Buffer B.
In order to check the extent of digestion of DNAs by endonuclease treatments, 0.4 pg of the endonuclease-digested ColEl DNA was run on 15% agarose gel in the presence of 0.5 pg/ml of ethidium bromide using a horizontal slab gel electrophoresis apparatus (23) at 140 V for about 2 h a t room temperature. The DNA bands were visualized by the fluorescense from DNA-bound ethidium bromide under ultraviolet irradiation (24) and photographed with a camera equipped with an orange filter (Kenko MC YA3) and Neopan 400 Fuji film. Density of DNA bands on the photographic film was scanned with Joyce, Loebl and Co. microdensitometer MK 3CS. According to the microdensitometric analysis, DNase I (2 ng/ml) treatment of closed circular Poly(ADP-ribose) polymerase was purified by the method described in a previous report (8). The purity of the enzyme, examined by SDS'-polyacrylamide gel electrophoresis followed by amido black staining, was approximately 97%. The specific activity of the purified enzyme was 160 units/mg of protein. Reaction mixtures for the assay of this enzyme contained 25 mM Tris-HCI (pH 8.0), 10 mM MgCL, 0.5 mM dithiothreitol, 10 p~ [adenine-2,8-"H]NAD' (18 cpm/pmol), varying amounts of DNA, and 0.2 pg of the purified enzyme in a total volume of 0.2 ml. When 3H-labeled DNA was used in the assay mixture, 10 p~ [adenine-2,8-'HH]NAD' (18 cpm/pmol) was replaced by 10 p~ [adenine-U-'4C]NAD' (11 cpm/pmol). Histones were not added in all reaction mixtures because the purified enzyme could synthesize poly(ADP-ribose) without histones (8,25,26). The mixture was incubated at 25°C for 10 min, and the reaction was terminated by the addition of 2 ml of ice-cold 10% trichloroacetic acid. The acidinsoluble material was collected on a glass fiber filter, and the radioactivity was counted by a liquid scintillation spectrometer as described previously (27). One unit of the enzyme activity was defined as being equivalent to 1 nrnol of ADP-ribose incorporated into acidinsoluble material/min under the described conditions. The apparent K , for DNA was obtained from the double reciprocal plots of enzyme activity uersus DNA concentration as previously described (8).

SDS-Polyacrylamide Gel Electrophoresis
The gel electrophoresis and staining of protein bands with amido black were performed according to the method of Hayashi and Ohba (28). Protein samples were prepared by heating at 100°C for 2 min in 1% SDS with 5% /3-mercaptoethanol. Electrophoresis was performed on 5.0 and 7.5% polyacrylamide gels containing 0.1% SDS at pH 7.2 in sodium phosphate buffer a t a constant current of 8 mA/gel for 5 to 7 h. Marker proteins for molecular weight determination (bovine serum albumin, E. coli RNA polymerase subunits, and soybean trypsin inhibitor) were obtained from Boehringer.

Ultracentrifugal Analysis
All sedimentation velocity and equilibrium experiments were carried out with a Hitachi model 282 analytical ultracentrifuge.
Sedimentation Velocity-The sedimentation coefficients were calculated from the movement of the maximum of the gradient curve with the scanning systems. Before the run, all samples were dialyzed I The abbreviation used is: SDS, sodium dodecyl sulfate. against 0.1 M sodium phosphate buffer (pH 7.4) and were centrifuged in a 12-mm single sector cell for the schlieren system or in a 12-mm double sector cell equipped with quartz windows for the photoelectric scanning apparatus a t 280 nm. All measurements were done a t 60,000 rpm in a RA-72-TC rotor at 20°C.
Sedimentation Equilibrium-Before the run, all samples were dialyzed against Buffer A containing 0.1 M NaCl and were centrifuged in a 12-mm double sector cell equipped with quartz windows with a 3-mm sample column height. The gradient curve was detected by the photoelectric scanning at 280 nm. The time required to reach equilibrium was reduced by using the overspeed technique of Richards et al.
(29). The following sequence of speeds was used: 1) the enzyme was centrifuged for 5 h a t 7,900 rpm, 2) the speed was decreased to 4,000 rpm for 15 min, 3) the speed was increased to 5,700 rpm and to equilibrium. The sample was assumed to be a t equilibrium when scans taken over the space of several hours showed no more variation than that expected from experimental error. At the end of run, the speed was increased to 40,000 rpm to deplete the meniscus, then measurements were made for base-line correction. All measurements were done in a RA-72-TC rotor at 4OC. The slope of a plot of the logarithm of the optical density at 280 nm as a function of radial distance squared was determined by the method of least squares on a programmable Wang 600 calculator. The partial specific volume (i;) of the enzyme was calculated from the amino acid composition (30). Solvent densities were measured by pyknometry. University, Osaka, Japan) using a Hitachi KLA-5 amino acid analyzer. The poly(ADPribose) polymerase samples were extensively dialyzed against 0.1 M KC1 and then deionized water and hydrolyzed under reduced pressure in 6 N HCl for 24, 48, and 72 h at IlOOC. Cysteine and half-cystine were determined as cysteic acid after performic acid oxidation according to the method of Schram et aL (32). Tryptophan was determined spectrophotometrically according to the method of Edelhoch (33).

Isoelectric Focusing
Isoelectric focusing of the enzyme in 10 to 30% glycerol gradient was performed with LKB 8144 Ampholine electrofocusing equipment. P-Mercaptoethanol (0.1%) was added to the solution to stabilize the enzyme activity. The concentration of carrier Ampholine (pH range, 7 to IO) was 1.8%. The enzyme (0.9 mg) was applied and focusing was performed at 4°C for 3 days. The recovery of the enzyme activity after the run was approximately 30%.

Binding Assay
Binding of poly(ADP-ribose) polymerase to DNA was measured by a glass fiber filter assay which was employed in the study of adenovirus DNA.terminal protein complex (21). Just prior to use, filters (Whatman GF/F, 21 mm in diameter) were rinsed with 1 ml of a binding buffer (25 mM Tris-HCI (pH 8.0), 10 mM MgCL, 0.5 mM dithiothreitol, and 0.1 M NaCI). For all binding experiments, varying amounts of purified poly(ADP-ribose) polymerase, as indicated in each experiment, and 0.4 pg (5100 cpm) of %-labeled ColEl plasmid DNA were mixed in a final volume of 0.2 ml of the binding buffer. The reaction was started by the addition of the enzyme and incubated at 25°C for 10 min, then 1.0 ml of the chilled binding buffer was added. The diluted sample was filtered at a flow rate Of 30 to 40 ml/ min. The reaction tube was rinsed twice with 1 ml of the binding buffer, and the resultant solutions were passed through the filter. Then, the filter was dried for 30 min a t 80°C and counted in a toluenebased scintillation fluid. In the absence of the enzyme, less than 3% of 'H-labeled ColEl DNA was retained by the filter.

Properties 6207 Competitive BindLng Assay
When increasing amounts of unlabeled DNA are added to a constant amount of labeled ColEl DNA and then the enzyme is added, the unlabeled DNA competes with the labeled DNA for the enzyme, resulting in a decrease in labeled DNA.enzyme complex. This experiment was done in the presence of 0.4 pg (0.095 pmol) of DNase Itreated ( I O ng/ml) ColEl ['HIDNA, 0.6 pg (4.5 pmol) of the enzyme, and varying amounts of unlabeled DNA, the concentration of which is indicated in Fig. 7. Other conditions of this binding assav are the same as described above.

Other Methods
RNA was estimated spectrophotometrically (34). Protein and DNA were estimated by the method of Lowry et al. (35) and Burton (36). respectively. DNA was also estimated by measuring its absorbance at 260 nm. The minimum subunit molecular weight of the enzyme was determined by SDS-polyacrylamide gel electrophoresis. The calibration curve with proteins of known molecular weights showed no apparent molecular weight dependence on the gel concentration and a molecular weight value of 130,000 was obtained. This value is in good agreement with that obtained from sedimentation equilibrium data, indicating the absence of oligomer under the chosen buffer condition. Although the salt composition of the buffer was different from the enzyme assay mixture, the different concentrations of NaCl and MgCI? did not affect the molecular weight or S value of the enzyme.' S Value Determination-Results from sedimentation velocity experiments are shown in Fig. 2. A symmetrical single peak was always observed, indicating a homogeneous enzyme solution (data not shown). The sedimentation coefficient was dependent on the enzyme concentration and the .s(L),l.,c value obtained by extrapolation was 4.94 S .

Physical Properties
Stokes Radius Determination-The Stokes radius of the enzvme was obtained by comparing the elution position of the enzyme on a column of Sephadex G-200 with those of proteins of known Stokes radii and was estimated to be 59 A.
From the Stokes radius and s~ll,,~ obtained by the above determinations, the molecular weight of the enzyme was calculated to be 127,000 using the following equation (31): where M , = molecular weight, a = Stokes radius, s = sedimentation coefficient, cT= partial specific volume, 7 = viscosity of medium, p = density of medium, and N = Avogadro's number.
From the ,s~~,.,, and molecular weight determined by equilibrium centrifugation, the frictional ratio ( f / f m l i , ) of the enzyme was calculated to be 1.81 using the following equation (37):  Table I. In addition to these amino acids, a small unidentified peak was found just before marker glucosamine on the short column used for the separation of basic amino acids. The lysine content was the highest in this protein and the isoelectric focusing data showed a single peak of the enzyme activity at pH 9.8, indicating the basic nature of the enzyme.
Physical properties of the enzyme described above are summarized in Table 11.
Enzyme Activation and Binding Properties of DNA As previously described (13), the enzyme is tightly bound to chromatin in cells and, when isolated and purified, it definitely requires DNA not as a substrate (8,26) but as an enzyme activator with an unknown mechanism. Recently, the activity of a partially purified enzyme is reported to be enhanced in response to DNA strand breaks (19). The activity

Physical and DNA-binding
Properties of our purified bovine thymus enzyme was also increased by DNase I or DNase I1 treatment of calf thymus DNA (38). To determine further the interaction between the enzyme and DNA, we analyzed the formation of the enzyme e DNA complex by its retention on glass fiber filters and correlated the complex formation with poly(ADP-ribose) synthesis. In order to avoid structural ambiguity in high molecular DNA of animal origin and inhibitory activity of single-stranded DNA (8), covalently closed circular molecules of ColEl plasmid DNA and its controlled DNase I-digested products were used as a standard DNA in the present experiment.
Enzyme Activating Ability of ColEI DNA-As shown in Fig. 3, closed circular ColEl DNA which has no singlestranded breaks (nicks) shows essentially no enzyme-activating ability. When this DNA is digested mildly with increasing amounts of DNase I, the enzyme-activating ability of the DNA increases (Fig. 3) and the initial part of the response is linear to the number of nicks on the DNA (Fig. 4).
Binding Characteristics-In the binding assay used in this experiment, the composition of the buffer was the same as that of the enzyme assay except for the absence of NAD' and the inclusion of 0.1 M NaCI, as described later. Addition of NAD' reduced the binding efficiency to some extent, probably  through the formation of the product. The maximum forrnation of the complex was attained within 1 min of incubation a t 25"C, and the complex was detected even when the reaction mixture was incubated a t 0°C.
As shown in Fig. 5, whereas free ,'H-labeled ColEl DNA is not retained on glass fiber fdter, the DNA-enzyme complex is. When increasing amounts of the enzyme are added to a solution of ColEl ["HIDNA, an increasing proportion of the DNA is retained on the filter. T h e efficiency of retention of the complex with nicked ColEl DNA (DNase I-treated ColE1 DNA) is higher than that with closed circular ColE1 DNA, but the higher efficiency decreased to the level which was attained by the complex with closed circular ColEl DNA when the enzyme was heat-inactivated (65"C, 15 min).
Effect of NaCl-The enzyme is bound to chromatin in c i c~o and can be easily separated from chromatin with a buffer containing 0.3 M NaCI. In the enzyme purification step, the enzyme bound to DNA-cellulose was dissociated by a buffer containing 0.5 to 1.0 M NaCl (8). If the binding of the enzyme to nicked DNA is a prerequisite for the enzyme to be catalytically active, the addition of NaCl to the nicked DNA.enzyme complex may eventually lead to a loss of the enzyme activity accompanying the dissociation of the complex. Fig. 6 shows that both the activity and the binding efficiency of the enzyme decrease approximately in a parallel manner when the concentration of NaCl is increased. It is also noted that the binding of the enzyme to closed circular ColEl DNA is cataIyticaIly abortive and is thus nonspecific. The specificity of the binding is clear at 0.1 M NaCI, where the binding efficiency of the enzyme to closed circular ColEl DNA is much less than that to nicked ColEl DNA, Thus, in routine binding assay, 0.1 M NaCl was always added to the binding mixture. These results support the notion that the specific binding of the enzyme to nicked DNA is a prerequisite for the enzyme-activating process.
Competiti1.e Binding Assay-If our binding assay could detect the formation of a catalytically active enzyme.II)NA complex, the enzyme-activating efficiencies of various doublestranded DNAs should be well correlated with their enzymebinding efficiencies since various double-stranded DNAs, including their DNase I-digested products, activate the enzyme with varying efficiencies.
The formation of complexes between the enzyme and var- ious DNAs was tested by a competitive binding assay. As shown in Fig. 7, when increasing amounts of unlabeled DNAs are added to 0.4 pg of DNase I-digested (10 ng/ml) ColE1 [?H]DNA and then 0.6 p g of the enzyme is added, the unlabeled DNA competes with the labeied DNA for the enzyme, NOCI 1. i 1

FIG. 6.
Effect of NaCl on the efficiency of binding of poly(ADP-ribose) polymerase to 3H-labeled ColEl DNA a n d of the enzyme activity. The binding was determined as described under "Materials and Methods" except that NaCl was added in the binding buffer as indicated and the amount of the enzyme was 0.4 p g (3 pmol). The assay of the enzyme was determined as described under "Materials and Methods," except that NaCI was added in the assay mixture as indicated and 0.1 pg of the "H-labeled ColEl DNA was used. 0, the efficiency of the binding with DNase 2-treated ( The results obtained also indicate that all unlabeled DNAs are not equally effective competitors. Relative amounts of various unlabeled DNAs necessary to reduce the fdter-bound counts to 50% of control counts obtained in the absence of the unlabeled DNA were used to estimate the efficiencies of the binding of the enzyme to the DNAs (Table 111). To a fairly good approximation, these amounts reflect the efficiency of DNA in the formation of a catalytically active enzyme. DNA complex because the enzyme-binding efficiencies of various double-stranded DNAs are reasonably correlated with their enzyme-activating efficiencies. Although active DNA shows high enzyme-binding and enzyme-activating efficiencies, similar high efficiencies are achieved by extensive treatments of ColEl DNA and calf thymus DNA with DNase I (100 ng/ml).
The activation of the enzyme can also be introduced by the addition of full length linear ColE1 DNA produed by EcoRI restriction endonuclease (Table 111, Line 6). This observation supports the fact observed fist by Benjamin and Gill (19) that some of the DNA ends activate the enzyme. Our previous kinetic studies (8) suggested that the strong inhibitory effect of denatured DNA on the enzyme activity may be due to its high affinity with the enzyme, resulting in easy formation of an abortive enzyme. denatured DNA complex. This prediction is confirmed directly in the competitive binding assay (Table  111) in which heat-denatured DNA is shown to be the strongest competitor among various DNAs used in this experiment.
Effect of Various Inhibitors- Table IV shows that heparin has a strong inhibitory effect on both the binding of the enzyme to nicked ColEl DNA and the enzyme activity. Similar inhibitory effects are seen in the presence of an excess amount of ribosomal RNA or of closed circular ColE1 DNA. The inhibitory effects of closed circular ColEl DNA are another indication of an abortive binding of the enzyme to this DNA.
Thymidine and nicotinamide inhibit the enzyme reaction competitively with NAD' (39,40). N-Ethylmaleimide (8) and theophylline (41,42) also inhibit the enzyme activity. As shown in Table IV, at the enzyme-inhibitory concentrations, these inhibitors show little effect on the binding efficiency. Active DNA "The efficiency was calculated from the data in Fig. 7 and is represented as the reciprocal value of the relative amount of unlabeled DNA to reduce the fiter-bound counts to 50% of control counts obtained in the absence of the unlabeled DNA.
The efficiency is represented as the relative value of l/apparent K,. The method for the determination of the apparent K,,, is described under "Materials and Methods."  The enzyme activity was assayed as described under "Materials and Methods" except that various inhibitors were added to the assay mixture as indicated and 0.1 pg of unlabeled DNase I-treated (10 ng/ ml) ColEl DNA was used.
'The binding efficiency was determined by the same method as the competitive binding assay described under "Materials and Methods" except that various inhibitors were added instead of unlabeled DNA. ' Dithiothreitol was excluded from the enzyme assay mixture and from the binding buffer. Table 11, poly(ADP-ribose) polymerase from bovine thymus has a molecular weight of approximately 130,000 and is composed of a single polypeptide. Precise structure of activated enzyme bound to DNA is, however, not clear. A high frictional ratio of 1.81 of the enzyme calculated from the molecular weight and the s&,,< suggest a marked degree of hydration of the protein, or pronounced nonspherical molecular asymmetry, or both. Our previously reported (8) overestimation of the molecular weight (150,000 to 160,000) determined by gel chromatography according to the method of Andrews (43) may be due to the hydrodynamic property of the molecule.

As shown in
Recently, poly(ADP-ribose) polymerase has been purified from various tissues by several investigators. The molecular weights of these purified enzymes, determined by SDS-polyacrylamide gel electrophoresis, are 120,000 for the calf thymus enzyme by Mandel et al. (9), 130,000 for the enzyme of Ehrlich ascites tumor cells by Kristensen and Holtlund (lo), 63,500 for the pig thymus enzyme by Tsopanakis et al. (ll), and 50,000 for the major band of the rat liver enzyme preparation by Okayama et al. (12). These results suggest the presence of two molecular species of the enzyme, although the presence of both types of the enzyme in one tissue has not been found yet. Recently, we found an enzyme of molecular weight of 130,000 determined by SDS-polyacrylamide gel electrophoresis also in bovine liver," indicating that this molecular species may be distributed widely in the various bovine tissues. Recently, Kristensen and Holtlund (10) suggested the presence of an oligomeric conformation of their purified enzyme from Ehrlich ascites tumor cells. With our purified enzyme, however, we could not detect such a conformation by several analytical ultracentrifugal studies under different buffer conditions.
Our estimated PI value of 9.8 is similar to that of 9.4 for the K. Yoshihara, unpublished results enzyme from Ehrlich ascites tumor cells (10) but is significantly higher than that of 6.5 for calf thymus enzyme by Mandel et al. (9). The low estimation by Mandel et al. (9) may be due to contaminating DNA in their enzyme preparation. Closed circular ColEl DNA shows essentially no enzymeactivating ability and mild DNase I digestion of DNA increases the ability in proportion to the number of nicks on the DNA (Figs. 3 and 4). From the results shown in Fig. 4, the specific activity of the enzyme, and its molecular weight (130,000), the mean number of the enzyme molecules activated by one nick was calculated to be 2.6, based on an assumption that all enzyme molecules were active. Therefore, when the presence of an inactive population in the enzyme molecules is taken into account, this value suggests that an enzyme molecule is activated by one nick.
Efficiencies of various DNAs to activate the enzyme (Kn2 for DNA) were very different, and parallel changes occurred in the enzyme-activating and the enzyme-binding efficiencies of these DNAs (Table 111). Therefore, it would appear that variation in enzyme-activating efficiencies of DNAs is mainly the result of their different enzyme-binding capacities. The result of comparison of these two efficiencies suggests that a preference of the enzyme for DNA strand breaks such as nicks or ends is the basis for the enzyme-activating and enzymebinding efficiency of DNA.
We previously reported (44) a higher enzyme-act.ivating efficiency for poly(d(A)) .poly(d(T)) than for poly(d(G)). poly(d(C)) as an indication of sequence preference in the enzyme activation. However, this notion seems to be incorrect because the number of nicks or ends per unit weight of the poly(d(G)).poly(d(C)) (about 13 S) is much lower than that of the poly(d(A))-poly(d(T)) (about 5 S ) , mainly owing to different preparative procedures. Therefore, the difference in enzyme-activating efficiencies of these DNAs is largely due to the difference in the number of nicks or ends. High enzymeactivating and enzyme-binding ability of active DNA may also be attributable to its small S~O , ,~ of 6 S and a high content of nicks! Considering the polyanionic nature of denatured DNA, heparin, RNA, and closed circular DNA, and a high PI value for the enzyme, a nonspecific ionic interaction between these inhibitors and the enzyme is probably the basis for their inhibitory activity on the binding of the enzyme to nicked DNA ( Fig. 6; Table IV). None of the other inhibitors, such as thymidine, nicotinamide, theophylline, and N-ethylmaleimide, affected appreciably the nicked DNA-enzyme interaction.
All of these results indicate that the DNA-binding domain of the enzyme may be separated from its catalytic site and interact preferentially with nicks or ends of double-stranded DNA, and that this specific interaction (binding) induces the conformational change of the enzyme to an active form which initiates the synthesis of poly(ADP-ribose).
During the preparation of this manuscript, Ito et al. (45) reported a purification and characterization of poly(ADP-ribose) synthetase from calf thymus. This enzyme is a single polypeptide having an appoximate molecular weight of 110,000.