A Scintillation Proximity Assay for Poly(ADP-ribose) Polymerase

Analytical Biochemistry 282, 24 –28 (2000) doi:10.1006/abio.2000.4604, available online at http://www.idealibrary.com on

A Scintillation Proximity Assay for Poly(ADP-ribose) Polymerase Anissa Cheung and Jie Zhang 1 Guilford Pharmaceuticals Inc., 6611 Tributary Street, Baltimore, Maryland 21224

Received January 5, 2000

Poly(ADP-ribose) polymerase (PARP) is an abundant nuclear protein in most of the eukaryotic tissues. When activated by DNA damage, PARP synthesizes poly(ADP-ribose) from NAD. Conventional radioactive PARP enzyme assay requires the separation of the polymer product from the NAD substrate, a rate-limiting step that hampers large-scale chemical library screening to identify novel small-molecule PARP inhibitors. By using biotinylated NAD, we have developed a scintillation proximity assay (SPA) for PARP. We demonstrated that PARP can incorporate the biotinylated ADP-ribose units into the radioactive poly(ADP-ribose) polymer, which can directly bind and excite the streptavidin-conjugated scintillation beads. PARP-SPA can be readily adapted to a 96-well format for automatic high-throughput screening for PARP inhibitors. © 2000 Academic Press

Poly(ADP-ribose) polymerase (PARP) 2 [EC 2.4.2.30] is a 113-kDa protein that uses NAD (␤-nicotinamide adenine dinucleotide) as its substrate to synthesize poly(ADP-ribose), a branched polymer that can consist of over 200 ADP-ribose units (1). This major nuclear protein has also been called PARS, ADPRT, or pADPRT. It modifies nuclear proteins with poly(ADP-ribosyl)ation (2). Many protein acceptors of poly(ADP-ribose) are involved in maintaining DNA integrity. They include histones, topoisomerases, DNA and RNA polymerases, DNA ligases, and Ca 2⫹- and Mg 2⫹-dependent endonuclease. PARP itself serves as a major acceptor 1 To whom correspondence should be addressed. Fax: 410-6316804. E-mail: [email protected]. 2 Abbreviation used: biotinylated NAD, N 6 -[N-(N-biotinyl-⑀-aminocapronyl)-6-aminohexylcarbamoylmethyl]NAD; ECL, enhanced chemiluminescence; HRP, horseradish peroxidase; NAD, ␤-nicotinamide adenine dinucleotide; PARP, poly(ADP-ribose) polymerase; SPA, scintillation proximity assay.

24

through intermolecular auto-ADP-ribosylation. Although PARP protein is present at a high level in nuclei in most tissues, its activation is dependent on the DNA damage. Two zinc-finger domains in the Nterminus of PARP are responsible for binding to and detecting damaged DNA, which triggers poly(ADP-ribose) synthesis. The catalytic site resides in the Cterminal domain where polymerization and branching of ADP-ribose takes place. In the middle of the PARP, there is a glutamate-rich region that serves as receptors for auto-ADP-ribosylation. Quite recently two additional PARP genes have been identified, and they are distinct from the original PARP, which is now designated PARP1 (3– 6). The two new forms of PARP are only about half the molecular weight of PARP1. PARP2 and PARP3 display 40 and 31% sequence identity to the catalytic C-terminal domain of PARP1. Despite lack of the zinc-finger DNA binding domain and the automodification domain, PARP2 still displays auto-ADP-ribosylation activity in a DNA-dependent manner (6). The PARP2 DNA binding domain resides in the first 64 amino acids at the N-terminus. Unlike PARP1, PARP2 does not modify purified histones by ADP-ribosylation. Little is known at present about the substrates for PARP2 and 3. Two other proteins, which share homology with PARP catalytic domain, are Tankyrase, a protein involved in regulating telomere integrity, and vPARP, a component of the mammalian vault ribonucleoprotein complex (7, 8). PARP and its analogs synthesize poly(ADP-ribose) which typically contains of 200 ADP-ribose units, with an average of one branch every 25 units of ADP-ribose. The homopolymer consists of repetitive ADP-ribose units linked by ␣(1⬙32⬘)ribosyl–ribose glycosidic bonds and of branching residues with ␣(1⵮32⬙)ribosyl– glycosidic bond. The catalytic domain of PARP and its analogs share significant homology with mono-ADP-ribosyl transferases, enzymes that transfer one ADP-ribose 0003-2697/00 $35.00 Copyright © 2000 by Academic Press All rights of reproduction in any form reserved.

SCINTILLATION PROXIMITY ASSAY FOR POLYMERASE

25

FIG. 1. The principles of PARP-SPA. A mixture of NAD, biotin-NAD, and 3H-NAD is used by PARP to synthesize poly(ADP-ribose) that contains the labels of biotin and tritium. The polymer binds to the avidin–SPA beads through the association of biotin–ADP-ribose and the avidin. Thus, 3H-ADP-ribose units are brought to close vicinity of the beads and excites the scintillant. No separation of 3H-NAD and poly(ADP-ribose) is necessary.

from NAD to covalently modify protein substrates. Diphtheria toxin, a prototypical mono-ADP-ribosyl transferase, has been shown to be able to use biotinylated NAD to modify elongation factor 2 with ADPribosylation (9). The purpose of the current study is to test if biotinylated NAD can also be utilized for poly(ADP-ribose) synthesis. The radioactive polymer with biotin labeling is then suitable for detection by scintillation proximity assay (SPA) (Fig. 1) (10). In SPA, the ␤-particles emitted from the substrates dissipate in the aqueous solution, while the radioisotope decays from products excite the scintillant beads that bind to the products. SPA eliminates the requirement of separation between free ligand and complex, and thus greatly increases throughput. We report here the development of such a PARP-SPA, which can be easily adapted for automatic screening for PARP inhibitors. MATERIALS AND METHODS

Recombinant human PARP (660 u/mg), mouse monoclonal antibody (C2-10) against PARP, mouse monoclonal antibody against poly(ADP-ribose) polymer (4335-MC), and N 6 -[N-(N-biotinyl-⑀-aminocapronyl)6-aminohexylcarbamoylmethyl]NAD (biotinylated NAD) were from Trevigen (Gaithersburg, MD). The C2-10 antibody recognizes epitope within the C-terminal part of the DNA binding domain of PARP. The 4335-MC antibody is specific for poly(ADP-ribose) polymers

10 –50 units in length and does not recognize structurally related RNA, DNA, ADP-ribose monomer, NAD, or other nucleic acid monomers. Nicotinamide [adenine-2, 8- 3H]dinucleotide ( 3H-NAD) (5 Ci/mmol) and nicotinamide adenine[adenylate- 32P]dinucleotide ( 32P-NAD) (800 Ci/mmol) were from NEN. The streptavidin-conjugated SPA beads were from Amersham Pharmacia Biotech. Activated calf thymus DNA and other chemical agents were from Sigma Co. (St. Louis, MO). Biotinylated radioactive poly(ADP-ribose) was synthesized in a 0.1-ml reaction that consisted of 50 mM Tris–HCl (pH 8.0), 4 mM MgCl 2, 5 ␮g/ml activated DNA, 1.5 ␮M NAD, 3H-NAD (1.6 Ci/mmol), 15 u/ml PARP, and 125 nM biotinylated NAD. The reaction was terminated at various time points by adding 0.2 mg streptavidin-conjugated SPA beads. After 30-min incubation, the plates were counted on a TopCounter (Packard). Samples were resolved on 10% sodium dodecyl sulfate (SDS)–polyacrylamide gel according to standard method using the Novex electrophoresis system. 32Plabeled polymer was detected by autoradiograph after drying the gel on filter paper under vacuum at 70°C. To detect biotinylated polymer, proteins were electrophoretically transferred from SDS gels to nitrocellulose. The nitrocellulose filter was blotted with 0.5 ␮g/ml (HRP)–avidin (Sigma) in TBST buffer consisting of 50 mM Tris–HCl (pH 7.5), 200 mM NaCl, 0.2% (v/v)

26

CHEUNG AND ZHANG

anti-poly(ADP-ribose) antibody (1:1000), and after washing, the second incubation with the HRP-conjugated anti-mouse IgG (1:5000). RESULTS

FIG. 2. Detection of poly(ADP-ribose) by immunoblotting and autoradiography. Poly(ADP-ribose) was synthesized by using NAD alone (A, lanes 1 to 3) or biotinylated NAD (bioNAD) (A, lanes 4 and 5), or biotinylated NAD and 32P-NAD (B, lanes 7 and 8). After the sample was resolved on 10% SDS gel and transferred to nitrocellulose, samples in A were probed with antibody against PARP (lane 1), antibody against poly(ADP-ribose) (lanes 2 and 3), or avidin–HRP (lanes 4 and 5) and developed with ECL, while those in B were exposed to X-ray film. Inclusion of 1 mM benzamide (Bz), a PARP inhibitor, abolished the signals for enzyme activities (lanes 3, 5, and 8).

Tween 20, for 1 h, followed by 4 ⫻ 10 min wash with TBST buffer, and developed by the enhanced chemiluminescence (ECL) method according to the manufacturer’s protocol. The same procedure was followed for immunoblot detection of the polymer, except that the first incubation for the nitrocellulose filter was with

We first tested whether biotinylated NAD would interfere in poly(ADP-ribose) synthesis by PARP. In the radioisotope PARP assay, the presence of biotinylate NAD had no effect on the enzymatic activity as determined by incorporation of radioactivity into the poly(ADP-ribose) polymer (data not shown). Similarly, the inhibition profiles of typical PARP inhibitors such as benzamide and 3-aminobenzamide remained the same whether biotinylated NAD was included or not (data not shown). These results indicate that the biotin moiety does not interfere in the catalytic process of the PARP enzyme. We then tested whether biotinylated ADP-ribose becomes incorporated into the poly(ADP-ribose) polymer. Using streptavidin-conjugated horseradish peroxidase directly, we detected in the reaction product high-molecular-weight (⬎110 kDa) bands which were labeled with biotin (Fig. 2A). PARP is known to be the predominant acceptor of poly(ADP-ribose) through intermolecular auto-(ADP-ribosyl)ation and typically the modification retards mobility of PARP in protein gel. The multiple species detected were typical of proteins that are heterogeneously labeled with different lengths of poly(ADP-ribose) polymer. The same reaction products can be recognized by a monoclonal antibody against human PARP and a monoclonal antibody against poly(ADP-ribose) polymer (Fig. 2A). When 32P-NAD was added into the reaction mixture, autoradiography revealed a 32P-labeled product that migrated at the same regions as the bands detected by antibodies (Fig. 2B).

FIG. 3. Optimized conditions for PARP-SPA. (A) A time course for biotinylated poly(ADP-ribose) synthesis by PARP. (B) The biotin–NAD/ NAD ratio was adjusted to maximize the SPA signals.

SCINTILLATION PROXIMITY ASSAY FOR POLYMERASE

27

FIG. 4. Use of PARP-SPA for characterization of PARP inhibitors. Dose–response experiments were performed to determine the IC 50 values of 3-aminobenzamide (3-AB) (A) and benzamide (Bz) (B) on PARP inhibition.

Thin-layer chromatography analysis reveals no difference between phosphodiesterase-degraded products of 32 P-labeled poly(ADP-ribose) and that of 32P- and biotin-labeled polymer (data not shown). This suggests that PARP cannot distinguish between NAD and biotinylated NAD, and neither polymerization nor branching is influenced by biotin. Taken together, these results indicate that PARP can use biotinylated NAD to synthesize poly(ADP-ribose) and the incorporation of biotin in the product can be used as a measurement of the enzyme activity. Furthermore, the biotinylated poly(ADP-ribose) serves as a substrate for poly(ADP-ribose) glycohydrolase (A.C., Jia-He Li, and J.Z., manuscript in preparation). Next, we determined whether the biotinylated radioactive poly(ADP-ribose) polymer can bind and excite

the streptavidin-coated scintillation beads. Direct incubation of the PARP reaction mixture with the beads resulted in a scintillation signal which is dependent on the presence of biotinylated NAD as substrate (Fig. 3A). The signal is abolished when Mg 2⫹, which is required for PARP activity, was chelated by EDTA. The sensitivity of the PARP-SPA was compared directly against the conventional radioactivity PARP assay by subjecting the same reaction mixture to the two assays. Same level of signals were detected when the filterwash procedure was used to separate unincorporated radioactive NAD from the polymer (data not shown). We also varied the molar ratios between the biotinylated NAD and regular NAD and found that an optimal SPA signal was achieved around 1:10 ratio for biotinylated NAD/NAD (Fig. 3B).

28

CHEUNG AND ZHANG

We used the PARP-SPA to determine the inhibition profiles of prototypical PARP inhibitors. Both 3-aminobenzamide and benzamide were found to dose dependently inhibited PARP in the SPA assay with IC 50 values of 5 and 1.3 ␮M, respectively (Figs. 4A and 4B). The PARP-SPA results are consistent with the reported values of IC 50 in the range of 1 to 10 ␮M for benzamide, and from 5 to 20 ␮M for 3-aminobenzamide (11). Thus, PARP-SPA is validated as an alternative method for the conventional radioactive PARP assay. The sensitivity of PARP-SPA remains the same as that of the radioisotope assay. To keep the cost of largescale screening low, we routinely performed the primary screening at a concentration of NAD around the lower micromolar range, so that a minimal amount of biotinylated NAD, an expensive ingredient, was consumed. Any initial “hit” was to be followed by conventional enzymatic kinetic analysis to determine the K i and the mode of inhibition.

tomation greatly increases the throughput of compound screening. After directly comparing PARP-SPA against the radioactive PARP assay, we found no difference in terms of sensitivity and inhibition profile of selected PARP inhibitors. There is accumulating evidence supporting PARP inhibition as a novel mechanism to treat ischemia/ reperfusion and inflammation-related injuries (13–15). A number of PARP inhibitors have demonstrated remarkable efficacy in reducing tissue damage in animal models of cerebral ischemia, traumatic brain injury, myocardial ischemia, retinal ischemia, septic shock, hemorrhagic shock, MPTP toxicity, streptozotocin toxicity, and arthritic inflammation. The remarkable protective effects that PARP inhibitors afford in animal models of diseases hold strong potentials for further drug development to treat the aforementioned conditions. The PARP-SPA may expedite the lead discovery process for developing prospective PARP inhibitors to combat the devastating diseases.

DISCUSSION

ACKNOWLEDGMENT

A major finding of the study is that PARP can utilize biotinylated NAD to synthesize poly(ADP-ribose) polymer. The biotin label greatly simplifies monitoring the enzymatic activities of PARP and its homologs and may facilitate identification and characterization of acceptor proteins of poly(ADP-ribose) polymer. A number of PARP-related enzymes have recently been cloned. They include PARP2, PARP3, Tankyrase, and vPARP (3– 8). The expressed recombinant PARP homologs can produce bona fide poly(ADP-ribose) polymer. There is also evidence for the existence of other PARP homologs yet to be cloned (12). In the past, biotinylated NAD has been shown to serve as a substrate for mono(ADPribose) transferase, such as diphtheria toxin. Now we demonstrated that biotinylated NAD can also be used for poly(ADP-ribosyl)ation. Based on the sequence homology in the catalytic domains among PARP and its homologs, it is likely that biotinylated NAD can be used by PARP2, PARP3, Tankyrase, and vPARP for synthesis of poly(ADP-ribose). Such biotin-labeling methods offer a convenient means to determine the activities of PARP and its analogs. Furthermore, there is little known about the substrates for PARP2, PARP3, Tankyrase, and vPARP. The biotin tag makes it possible to use affinity purification to identify and characterize acceptor proteins for poly(ADP-ribose). The biotinylation of poly(ADP-ribose) enables us to develop an SPA for PARP. The procedure circumvents the requirement of separation between substrate and product, a rate-limiting step in large-scale screening for PARP inhibitors. By eliminating the washing step, the PARP-SPA reduces assay time. The 96-well or other miniature assay format can be readily adapted to automatic liquid sample-handling robotics and the au-

We thank S. Lautar for help in figure preparation.

REFERENCES 1. Jacobson, M., and Jacobson, E. (1999) Trends Biochem. Sci. 24, 415– 417. 2. Ueda, K., and Hayaishi, O. (1985) Annu. Rev. Biochem. 54, 73–100. 3. Babiychuk, E., Cottrill, P. B., Storozhenko, S., Fuangthong, M., Chen, Y., O’Farrell, M. K., Van Montagu, M., Inze, D., and Kushnir, S. (1998) Plant J. 15, 635– 645. 4. Berghammer, H., Ebner, M., Marksteiner, R., and Auer, B. (1999) FEBS Lett. 449, 259 –263. 5. Johansson, M. (1999) Genomics 57, 442– 445. 6. Ame, J. C., Rolli, V., Schreiber, V., Niedergang, C., Apiou, F., Decker, P., Muller, S., Hoger, T., Menissier-de Murcia, J., and de Murcia, G. (1999) J. Biol. Chem. 274, 17860 –17868. 7. Smith, S., Giriat, I., Schmitt, A., and de Lange, T. (1998) Science 282, 1484 –1487. 8. Kickhoefer, V. A., Siva, A. C., Kedersha, N. L., Inman, E. M., Ruland, C., Streuli, M., and Rome, L. H. (1999) J. Cell. Biol. 146, 917–928. 9. Zhang, J. (1997) in Methods in Enzymology (McCormick, D. B., Suttle, J. W., and Wagner, C., Eds.), Vol. 280, pp. 255–265. Academic Press, San Diego, CA. 10. Hart, H. (1983) U.S. Patent 4,382,074. 11. Banasik, M., Komura, H., Shimoyama, M., and Ueda, K. (1992) J. Biol. Chem. 267, 1569 –1575. 12. Shieh, W. M., Ame, J. C., Wilson, M. V., Wang, Z. Q., Koh, D. W., Jacobson, M. K., and Jacobson, E. L. (1998) J. Biol. Chem. 273, 30069 –30072. 13. Szabo, C., and Dawson, V. L. (1998) Trends Pharmacol. Sci. 19, 287–198. 14. Piepper, A. A., Verma, A., Zhang, J., and Snyder, S. H. (1999) Trends Pharmacol. Sci. 20, 171–181. 15. Zhang, J. (1999) in Emerging Drugs: The Prospect for Improved Medicines (Fitzgerald, J. D., Bowman, W. C., and Taylor, J. B., Eds.), Vol. 4, pp. 209 –221. Ashley Publication Ltd., London.