Purification and characterization of an (ADP-ribose)n glycohydrolase from human erythrocytes

An (ADP-ribose)n glycohydrolase from human erythrocytes was purified approximately 13,000-fold and characterized. On sodium dodecyl sulfate/polyacrylamide gel the purified enzyme appeared homogeneous and had an estimated relative molecular mass (Mr) of 59,000. Amino acid analysis showed that the enzyme had a relatively high content of acidic amino acid residues and low content of basic amino acid residues. Isoelectrofocusing showed that the enzyme was an acidic protein with pI value of 5.9. The mode of hydrolysis of (ADP-ribose)n by this enzyme was exoglycosidic, yielding ADP-ribose as the final product. The Km value for (ADP-ribose)n (average chain length, n = 15) was 5.8 microM and the maximal velocity of its hydrolysis was 21 mumol.min-1.mg protein-1. The optimum pH for enzyme activity was 7.4 KCl was more inhibitory than NaCl. The enzyme activity was inhibited by ADP-ribose and cAMP but not the dibutyryl-derivative (Bt2-cAMP), cGMP or AMP. These physical and catalytic properties are similar to those of cytosolic (ADP-ribose)n glycohydrolase II, but not to those of nuclear (ADP-ribose)n glycohydrolase I purified from guinea pig liver [Tanuma, S., Kawashima, K. & Endo, H. (1986) J. Biol. Chem. 261, 965-969]. Thus, human erythrocytes contain (ADP-ribose)n glycohydrolase II. The kinetics of degradation of poly(ADP-ribose) bound to histone H1 by purified erythrocyte (ADP-ribose)n glycohydrolase was essentially the same as that of the corresponding free poly(ADP-ribose). In contrast, the glycohydrolase showed appreciable activity of free oligo(ADP-ribose), much less activity on the corresponding oligo(ADP-ribose) bound to histone H1. The enzyme had more activity on oligo(ADP-ribose) bound to mitochondrial and cytosolic free mRNA ribonucleoprotein particle (mRNP) proteins than on oligo(ADP-ribose) bound to histone H1. It did not degrade mono(ADP-ribosyl)-stimulatory guanine-nucleotide-binding protein (Gs) and -inhibitory guanine-nucleotide-binding protein (Gi) prepared with cholera and pertussis toxins, respectively. These results suggest that cytosolic (ADP-ribose)n glycohydrolase II may be involved in extranuclear de(ADP-ribosyl)n-ation, but not in membrane de-mono(ADP-ribosyl)ation.     

Isolation and purification of poly(ADP-ribose) glycohydrolase from pig thymus

Poly(ADP-ribose) glycohydrolase has been purified about 12 300-fold from pig thymus with a recovery of 8.5%. The specific activity of the purified enzyme is 13.8 mumol min -1 mg protein -1. The molecular weight was estimated to be 59 000 by gel filtration through Sephadex G-100 in a non-denaturing solvent. Analysis of the final preparation by sodium dodecyl sulphate gel electrophoresis reveals two protein bands of molecular weight, 61 500 and 67 500. The Km value for poly(ADP-ribose) is estimated to be 1.8 microM monomer units. The enzyme preparation is free from phosphodiesterase, NADase and ADP-ribosyltransferase activities. The purified enzyme is inhibited by cyclic AMP, ADP-ribose, naphthylamine, histones H1, H2A, H2B, H3, polylysine, polyarginine, polyornithine and protamine. The inhibition by histone is relieved by an equal mass of DNA. Single-stranded DNA, poly(A), poly(I) and polyvinyl sulphate were inhibitory, but double-stranded DNA was not inhibitory.     

Cellular regulation of ADP-ribosylation of proteins. IV. Conversion of poly(ADP-ribose) polymerase activity to NAD-glycohydrolase during retinoic acid-induced differentiation of HL60 cells.

Two enzymatic activities of the nuclear enzyme poly(ADP-ribose) polymerase or transferase (ADPRT, EC 2.4.2.30), a DNA-associating abundant nuclear protein with multiple molecular activities, have been determined in HL60 cells prior to and after their exposure to 1 microM retinoic acid, which results in the induction of differentiation to mature granulocytes in 4-5 days. The cellular concentration of immunoreactive ADPRT protein molecules in differentiated granulocytes remained unchanged compared to that in HL60 cells prior to retinoic acid addition (3.17 +/- 1.05 ng/10(5) cells), as did the apparent activity of poly(ADP-ribose) glycohydrolase of nuclei. On the other hand, the poly(ADP-ribose) synthesizing capacity of permeabilized cells or isolated nuclei decreased precipitously upon retinoic acid-induced differentiation, whereas the NAD glycohydrolase activity of nuclei significantly increased. The nuclear NAD glycohydrolase activity was identified as an ADPRT-catalyzed enzymatic activity by its unreactivity toward ethenoadenine NAD as a substrate added to nuclei or to purified ADPRT. During the decrease in in vitro poly(ADP-ribose) polymerase activity of nuclei following retinoic acid treatment, the quantity of endogenously poly(ADP-ribosylated) ADPRT significantly increased, as determined by chromatographic isolation of this modified protein by the boronate affinity technique, followed by gel electrophoresis and immunotransblot. When homogenous isolated ADPRT was first ADP-ribosylated in vitro, it lost its capacity to catalyze further polymer synthesis, whereas the NAD glycohydrolase function of the automodified enzyme was greatly augmented. Since results of in vivo and in vitro experiments coincide, it appears that in retinoic acid-induced differentiated cells (granulocytes) the autopoly(ADP-ribosylated) ADPRT performs a predominantly, if not exclusively, NAD glycohydrolase function.     

Purification and characterization of an apurinic/apyrimidinic endonuclease from HeLa cells

An endodeoxyribonuclease from HeLa cells acting on apurinic/apyrimidinic (AP) sites has been purified to apparent homogeneity as judged by sodium dodecyl sulfate (SDS)-polyacrylamide gel electrophoresis. The presence of Triton X-100 was necessary throughout the purification for stabilization and stimulation of activity. The endonuclease has an apparent native molecular weight of 32,000 determined by molecular sieving and an apparent subunit molecular weight of 41,000 as judged by its electrophoretic mobility in SDS-polyacrylamide gels. The activity has an absolute requirement for Mg2+ or Mn2+ and a broad pH optimum between 6.7 and 9.0 with maximal activity near pH 7.5. The enzyme has no detectable exonuclease activity, nor any endonuclease activity on untreated duplex or single-stranded DNA. It is inhibited by adenine, hypoxanthine, adenosine, AMP, ADP-ribose, and NAD+, but it is unaffected by caffeine, the pyrimidine bases, ADP, ATP, or NADH. The use of a variety of damaged DNA substrates provided no indication that the enzyme acts on other than AP sites. The enzyme appears to cleave AP DNA so as to leave deoxyribose-5-phosphate at the 5' terminus and a 3'-OH at the 3' terminus; it also removes deoxyribose-5-phosphate from AP DNA which has deoxyribose at the 3' terminus. Specific antibody has been produced in rabbits which interacts only with a 41,000-dalton protein present in the purified enzyme (presumably the enzyme itself), as well as with partially purified AP endonuclease fractions from human placenta and fibroblasts.     

Change in activity of nuclear poly(ADP-ribose) glycohydrolase during the HeLa S3 cell cycle

The change in activity of nuclear poly(ADP-ribose) glycohydrolase during the cell cycle of HeLa S3 cells was investigated. The poly(ADP-ribose) glycohydrolase activity was solubilized from HeLa S3 cell nuclei and chromosomes only by sonication at high ionic strength. The enzyme hydrolyzed poly(ADP-ribose) exoglycosidically, producing ADP-ribose. After release from mitosis, the activity of the solubilized nuclear poly(ADP-ribose) glycohydrolase per nucleus or per unit protein, assayed with [3H]poly(ADP-ribose) (average chain length, n = 15) as substrate, was lowest in the early G1 phase and highest in the late G1 phase. The specific activity in the late G1 phase was about two times that in the early G1 phase. The high activity remained constant during the S-G2-M phase. A similar change during the cell cycle was observed after release from hydroxyurea block. These results suggest that the activity of poly(ADP-ribose) glycohydrolase doubled during the G1 phase of the cell cycle of HeLa S3 cells.     

Measurement of poly(ADP-ribose) glycohydrolase activity by high resolution polyacrylamide gel electrophoresis: Specific inhibition by histones and nuclear matrix proteins

We have developed a novel enzyme assay that allows the simultaneous determination of noncovalent interactions of poly(ADP-ribose) with nuclear proteins as well as poly(ADP-ribose) glycohydrolase (PARG) activity by high resolution polyacrylamide gel electrophoresis. ADP-ribose chains between 2 and 70 residues in size were enzymatically synthesized with pure poly(ADP-ribose) polymerase (PARP) and were purified by affinity chromatography on a boronate resin following alkaline release from protein. This preparation of polymers of ADP-ribose was used as the enzyme substrate for purified PARG. We also obtained the nuclear matrix fraction from rat liver nuclei and measured the enzyme activity of purified PARG in the presence or absence of either histone proteins or nuclear matrix proteins. Both resulted in a marked inhibition of PARG activity as determined by the decrease in the formation of monomeric ADP-ribose. The inhibition of PARG was presumably due to the non-covalent interactions of these proteins with free ADP-ribose polymers. Thus, the presence of histone and nuclear matrix proteins should be taken into consideration when measuring PARG activity.     

Evidence for a novel metabolic pathway of (ADP-ribose)n: pyrophosphorolysis of ADP-ribose in HeLa S3 cell nuclei

ADP-ribose liberated from (ADP-ribose)n by the action of (ADP-ribose)n glycohydrolase was converted to ATP and ribose 5-phosphate (ribose 5-P) in the presence of pyrophosphate (PPi) in HeLa S3 cell nuclei. This reaction was reversible and dependent on the simultaneous presence of ADP-ribose, PPi, Mg2+, and nuclei. These results suggest the presence of a novel enzyme in the nuclei, designated as ADP-ribose pyrophosphorylase, which catalyzes the reaction shown in Equation 1. ADP-ribose + PPi in equilibrium ATP + Ribose 5-P (1) This reaction could represent a pathway for the biosynthesis of ATP from (ADP-ribose)n in eukaryotic cell nuclei.     

TcPARP: A DNA damage-dependent poly(ADP-ribose) polymerase from Trypanosoma cruzi

Poly(ADP-ribose) polymerase (PARP) is a nuclear enzyme present in most eukaryotes and has been involved in processes such as DNA repair and gene expression. The poly(ADP-ribose) polymer (PAR) is mainly catabolised by poly(ADP-ribose) glycohydrolase. Here, we describe the cloning and characterisation of a PARP from Trypanosoma cruzi (TcPARP). The recombinant enzyme (Mr=65) required DNA for catalytic activity and it was strongly enhanced by nicked DNA. Histones purified from T. cruzi increased TcPARP activity and the covalent attachment of [32P]ADP-ribose moieties to histones was demonstrated. TcPARP required no magnesium or any other metal ion cofactor for its activity. The enzyme was inhibited by 3-aminobenzamide, nicotinamide, theophylline and thymidine but not by menadione. We demonstrated an automodification reaction of TcPARP, and that the removal of attached PAR from this protein resulted in an increase of its activity. The enzyme was expressed in all parasite stages (amastigotes, epimastigotes and trypomastigotes). When T. cruzi epimastigotes were exposed to DNA-damaging agents such as hydrogen peroxide or beta-lapachone, PAR drastically increased in the nucleus, thus confirming PAR synthesis in vivo and suggesting a physiological role for PARP in trypanosomatid DNA repair signalling.     

Inhibitory effect of tannin on poly(ADP-ribose) glycohydrolase from human placenta

Tannin, a naturally occurring polyphenolic compound, was found to inhibit the activity of purified poly(ADP-ribose) glycohydrolase from human placenta. The inhibition was dose-dependent and half maximal with 2.8 micrograms/ml of tannin. The inhibitory effect of tannin was two and three orders of magnitude more than those of ADP-ribose and cAMP, respectively. Kinetic analysis revealed that the inhibition by tannin was competitive with respect to the substrate poly(ADP-ribose).     

Characterization of two forms of poly(ADP-ribose) glycohydrolase in guinea pig liver

A poly(ADP-ribose) glycohydrolase from guinea pig liver cytoplasm has been purified approximately 45,000-fold to apparent homogeneity. The cytoplasmic poly(ADP-ribose) glycohydrolase designated form II differed in several respects from the nuclear poly(ADP-ribose) glycohydrolase I (Mr = 75,500) previously purified from the same tissue (Tanuma et al., 1986a). The purified glycohydrolase II consists of a single polypeptide with Mr of 59,500 estimated by a sodium dodecyl sulfate-polyacrylamide gel. A native Mr of 57,000 was determined by gel permeation. Peptide analysis of partial proteolytic degradation of glycohydrolases II and I with Staphylococcus aureus V8 protease revealed that the two enzymes were structurally different. Amino acid analysis showed that glycohydrolase II had a relatively low proportion of basic amino acid residues as compared with glycohydrolase I. Glycohydrolase II and I were acidic proteins with isoelectric points of 6.2 and 6.6, respectively. The optimum pH for glycohydrolases II and I were around 7.4 and 7.0, respectively. The Km value for (ADP-ribose)n (average chain length n = 15) and the Vmax for glycohydrolase II were 4.8 microM and 18 mumol of ADP-ribose released from (ADP-ribose)n.min-1.(mg of protein)-1, respectively. The Km was about 2.5 times higher, and Vmax 2 times lower, than those observed with glycohydrolase I. Unlike glycohydrolase I, glycohydrolase II was inhibited by monovalent salts. ADP-ribose and cAMP inhibited glycohydrolase II more strongly than glycohydrolase I. These results suggest that eukaryotic cells contain two distinct forms of poly(ADP-ribose) glycohydrolase exhibiting differences in properties and subcellular localization.     

Properties of purified calf thymus poly(adenosine diphosphate ribose) polymerase. Comparison of the DNA-independent and the DNA-dependent enzyme.

The physicochemical properties of the purified calf thymus poly(ADP-ribose) polymerase were investigated. The enzyme purified to homogeneity was shown to contain about 10% DNA on a weight basis and its activity to be DNA independent. After removing this fragment of DNA, called the sDNA fraction, the enzyme becomes DNA dependent. The activity of this enzyme preparation was entirely dependent on, and completely restored by, added calf thymus DNA or sDNA. However, the calf thymus DNA concentration needed was a hundred times higher than that of sDNA. The properties of the two enzyme preparations, DNA independent and DNA dependent, were essentially the same. They both reacted against the specific antibody obtained with the DNA-independent poly(ADP-ribose) polymerase. The pH optimum was around 8; the activity was stimulated by Mg2+, Mn2+ and Ca2+, and inhibited by high ionic strength, p-chloromercuribenzoate, ADP-ribose, AMP and polylysine. Nicotinamide, thymidine and NADP were shown to be competitive inhibitors. The enzymatic activity was stimulated by histone H1 when the ratio of DNA to histone H1 was 2. Histones H2A, H2B, H3 and H4 had little effect on the DNA-independent enzyme activity, but were strongly inhibitory for the DNA-dependent enzyme. This inhibitory effect could be reversed by allowing the DNA-dependent enzyme to react with the sDNA fraction before adding histone subfractions. The apparent Km for NAD of the DNA-dependent poly(ADP-ribose) polymerase was shown to vary with the DNA concentration. It was minimum when the amount of sDNA was 10% of that of the enzyme. The ratio of the apparent Km for sDNA to the enzyme concentration was constant at any enzyme concentration. The minimum estimation of the number of base pairs of sDNA required for maximal activation of one enzyme molecule was 16. For calf thymus DNA, this estimation was of 640. These results suggest that the activation of the enzyme needs the formation of some complex between the protein and a specific part of the DNA. This complex was preserved in the DNA-independent enzyme preparation.     

Effect of DNA on poly (ADP-ribose) glycohydrolase and the degradation of histone H1-poly (ADP-ribose) complex from HeLa cell nuclei.

A poly(ADP-ribose)-H1 histone complex has been isolated from HeLa cell nuclei incubated with NAD. The rate of poly(ADP-ribose) glycohydrolase catalyzed hydrolysis of the polymer in the complex is only 1/9 that of free poly(ADP-ribose), indicating that the polymer is in a protected environment within the complex. Comparison of the rate of hydrolysis of free poly(ADP-ribose) in the presence or absence of H1 to that in the complex synthesized de novo indicates a specific mode of packaging of the complex. This is further indicated by the fact that alkaline dissociation of the complex followed by neutralization markedly exposes the associated poly(ADP-ribose) to the glycohydrolase. The complex also partially unfolds when it binds to DNA as evidenced by a 2-fold increase in the rate of glycolytic cleavage of poly(ADP-ribose). This effect of DNA is not due to a stimulation of the glycohydrolase per se since hydrolysis of free polymer by the enzyme is strongly inhibited by DNA, especially single-stranded DNA. Inhibition of glycohydrolase by DNA results from the binding of the enzyme to DNA and conditions which decrease this binding (increased ionic strength or addition of histone H1 which competes for DNA binding) relieve the DNA inhibition.     

Poly(ADP-ribose) accessibility to poly(ADP-ribose) glycohydrolase activity on poly(ADP-ribosyl)ated nucleosomal proteins

Hydrolysis of protein-bound 32P-labelled poly(ADP-ribose) by poly(ADP-ribose) glycohydrolase shows that there is differential accessibility of poly(ADP-ribosyl)ated proteins in chromatin to poly(ADP-ribose) glycohydrolase. The rapid hydrolysis of hyper(ADP-ribosyl)ated forms of histone H1 indicates the absence of an H1 dimer complex of histone molecules. When the pattern of hydrolysis of poly(ADP-ribosyl)ated histones was analyzed it was found that poly(ADP-ribose) attached to histone H2B is more resistant than the polymer attached to histone H1 or H2A or protein A24. Polymer hydrolysis of the acceptors, which had been labelled at high substrate concentrations (greater than or equal to 10 microM), indicate that the only high molecular weight acceptor protein is poly(ADP-ribose) polymerase and that little processing of the enzyme occurs. Finally, electron microscopic evidence shows that hyper(ADP-ribosyl)ated poly(ADP-ribose) polymerase, which is dissociated from its DNA-enzyme complex, binds again to DNA after poly(ADP-ribose) glycohydrolase action.     

Effects of chemically defined tannins on poly (ADP-ribose) glycohydrolase activity.

Three classes of chemically defined tannins, gallotannins, ellagitannins and condensed tannins were examined for their inhibitory activities against purified poly (ADP-ribose) glycohydrolase. Ellagitannins showed higher inhibitory activities than gallotannins. In contrast, condensed tannins, which consist of an epicathechin gallate (ECG) oligomer without a glucose core were not appreciably inhibitory. Kinetic analysis revealed that the inhibition of ellagitannins was competitive with respect to the substrate poly(ADP-ribose), whereas gallotannins exhibited mixed-type inhibition. These results suggest that conjugation with glucose of hexahydroxy-diphenoyl (HHDP) group, which is a unique component of ellagitannins, potentiated the inhibitory activity, and that the structure of ellagitannins may have a functional domain which competes with poly(ADP-ribose) on the poly(ADP-ribose) glycohydrolase molecule.     

Purification and properties of poly(ADP-ribose) polymerase from pig-thymus nuclei.

The nuclear enzyme poly(ADP-ribose) polymerase has been purified about 9200-fold from pig thymus nuclei with a 46% yield. An aqueous organic solvent system was used for the isolation of the polymerase from nuclei and for its purification by chromatography at sub-zero temperatures. Electrophoretic analysis under both denaturing and non-denaturing conditions revealed a single protein band suggesting that the preparation was homogeneous and that the enzyme is composed of one polypeptide chain. The molecular weight estimated from sodium dodecyl sulphate-/polyacrylamide gel electrophoresis was 63 500 and from gel filtration through columns of Sephadex G-100, 58 000. The enzyme preparation was free from poly(ADP-ribose)-degrading enzymes and from DNA. The purified polymerase showed an absolute requirement for both DNA and histones. The maximal specific activity of the homogeneous preparation measured by the standardized assay, was 20.7 mu mol NAD+ incorporated x min-1 x mg-1 of protein at 37 degree C. Amino-terminal group analysis with dansyl chloride did not reveal a terminal amino acid suggesting that the amino-terminal group may be blocked. In the presence of histones, the Km for NAD+ was 23 micrometer.     

Lignin inhibits (ADP-ribose)n glycohydrolase activity

(ADP-ribose)n glycohydrolase activity was inhibited in vitro by lignin, a naturally occurring polymethoxyphenolic compound. However, coniferyl alcohol, which is a main component of lignin, was not inhibitory even at 100 micrograms/ml. Lignin caused competitive inhibition with respect to the substrate (ADP-ribose)n and its Ki value was 18 micrograms/ml. These results suggest that lignin with a polymerized structure has a functional domain that interacts with the (ADP-ribose)n glycohydrolase molecule at the same site as (ADP-ribose)n.     

ADP-Ribosylation of Highly Purified Rat Brain Mitochondria

Highly purified synaptic and nonsynaptic mitochondria were prepared from rat brain, and their ADP-ribosyl transferase and NAD glycohydrolase activities were investigated. Data show that there is no significant difference in ADP-ribosyl transferase activity between these two types of subcellular preparations. However, NAD glycohydrolase activity appeared to be much higher in nonsynaptic mitochondria. The specific activity of both enzymes was investigated in the presence of the inhibitor nicotinamide or its analogue 3-aminobenzamide or other adenine nucleotides, such as ATP or ADP-ribose. The inhibitory effect of nicotinamide or 3-aminobenzamide on ADP-ribosyl transferase appears rather weak compared with their effect on NAD glycohydrolase activity. However, ADP-ribose and ATP appeared more effective in inhibiting ADP-ribosyl transferase. Our results provide evidence for the existence of ADP-ribosyl transferase activity in rat brain mitochondria. When NAD glycohydrolase was inhibited totally by nicotinamide, the transfer of ADP-ribose from NAD to mitochondrial proteins still occurred. The chain length determinations show that the linkage of ADP-ribose to mitochondrial proteins is oligomeric.     

Mg2+-dependent/poly(ADP-ribose)-sensitive endonuclease

A novel endonuclease from adult hen liver nuclei has been purified to a homogeneous state through salt extraction, ammonium sulfate fractionation, gel filtration, acetone fractionation, and successive chromatography of 1) hydroxyapatite and DNA Sepharose and 2) hydroxyapatite and isoelectric focusing. The endonuclease has a pH optimum at 9.0 and requires Mg2+ for activity. The enzyme hydrolyzes more rapidly in the order of polynucleotide: denatured DNA = rRNA greater than poly(dA) = poly(dT) greater than poly(dC) = poly(dG) greater than native DNA. This endonuclease degrades denatured DNA about 20 times more rapidly than does the native DNA. The products contain 5'-phosphoryl and 3'-hydroxyl termini and all four deoxynucleotides are present while dGMP is predominant. The enzyme cleaves the circular duplex PM2 DNA, endonucleotically, via single strand scission. The isoelectric point is 10.2 +/- 0.2 and the molecular weight is 43,000 +/- 2,000, determined by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and gel filtration. Pyridoxal 5'-phosphate and 2,3-butanedione inhibit the catalytic activity, respectively. The inhibition of DNA binding activity was also seen with former, but not with the latter. Purified Mg2+-dependent alkaline endonuclease was used to investigate the nature of poly(ADP-ribose) inhibition of the enzyme. In contrast to the Ca2+/Mg2+-dependent endonuclease (Yoshihara, K., Tanigawa, Y., Burzio, L., and Koide, S. S. (1975) Proc. Natl. Acad. Sci. U. S. A. 72, 289-293), ADP-ribosylation of the endonuclease protein was not observed. When 100 ng of the poly(ADP-ribose) having four to five ADP-ribose units per molecule were added to the nuclease assay system (total volume of 0.2 ml) 14% inhibition was observed, and increase in the chain length increased the inhibition. When 100 ng of poly(ADP-ribose) consisting of 20 or more units of the ADP-ribose per mol were added, the inhibition was over 95%. The possible role of the poly(ADP-ribose)-sensitive endonuclease is discussed.     

Purification and characterization of poly(ADP-ribose) glycohydrolase. Different modes of action on large and small poly(ADP-ribose)

Poly(ADP-ribose) glycohydrolase was purified approximately 74,000-fold to apparent homogeneity from calf thymus with a yield of 3.2%. The enzyme was a monomeric protein of Mr = 59,000, as determined by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The action of glycohydrolase on poly(ADP-ribose) was exoglycosidic in the direction of adenosine terminus----ribose terminus; radioactive ADP-ribose monomers were immediately produced from evenly labeled poly(ADP-ribose), but not from the polymer labeled selectively at the ribose terminus. The enzymatic degradation of large poly(ADP-ribose) (greater than 20 ADP-ribose residues) proceeded in a biphasic as well as bimodal manner. In the early and rapid phase, the enzyme degraded part of large polymers successively, leaving the remainder completely intact, and accumulated ADP-ribose monomers and small polymers of the size less than half of original polymers, indicating that the enzyme action was processive up to a certain extent. In the late and 20-fold slower phase, by contrast, the enzyme degraded the accumulated small polymers gradually and evenly, i.e. in a nonprocessive manner. The Km for large polymers was approximately 100-fold lower than that for small polymers. Similar rates and processivities were observed with large and small polymers bound to various proteins. These results suggested that the glycohydrolase may regulate differentially the levels of large and small poly(ADP-ribose) in the cell.