The role of poly(ADP-ribose) glycohydrolase (PARG) in cell death

Abstract

The post-translational modification of proteins with poly(ADP-ribose) (PAR) occurs as an early response to genotoxic insults. The nuclear poly(ADP-ribose) polymerase 1 (PARP-1) reacts to DNA nicks, activating its catalytical function and thereby producing large amounts of PAR. The substrate for this reaction is NAD+. Negatively charged linear PAR chains of up to 200 ADP-ribose monomers with maximal 3% branching are formed within minutes. Most of PAR is covalently bound on PARP-1 itself. Only 5-10% of PAR molecules are attached to other acceptor molecules. The turnover of PAR is tightly regulated. Its catabolic half-life ranges from seconds to hours depending on polymer size, complexity, protein association and the particular stress conditions to which cells are exposed. PAR glycohydrolase (PARG) degrades PAR via endo- and exoglycosidic activity. PARG produces primarily monomeric ADP-ribose albeit a few PAR polymers may arise. The coordinate action of PARP and PARG is required for proper cellular response to DNA damage and maintenance of genomic stability. Although the role of PARP-1 has been extensively studied, the biological function of PARG is not well established so far. Several studies with inhibitors and partial knockout models of PARG pointed to a possible role for PARG enzyme in the regulation of cell death or survival after alkylation- and oxidant-induced DNA damage. However, the lack of potent and cell permeable inhibitory molecules for PARG and the lethal phenotype of parg-/- mutants prevented the discovery of distinct functions in the cell death and survival network. Therefore we developed a RNA interference (RNAi) approach against PARG to reduce the amount of PARG protein in human and murine cells. Silencing of the parg gene resulted in a time-dependent reduction of parg mRNA, PARG protein and enzymatic activity which was maximal after 72 h of siRNA transfection. We found that as little as 10% of PARG protein is sufficient to ensure basic cellular functions as cell proliferation and DNA damage repair after sublethal doses of oxidative stress (H2O2). Cell survival following higher concentrations of H2O2 was increased in cells lacking PARG. In fact, the resident time of PAR molecules after oxidative stress was prolonged in cells silenced for PARG resulting in a protective phenotype. Dawson and collaborators demonstrated that PAR may act as a cell death signal upstream of apoptosis-inducing factor (AIF). In particular, PAR was found to mediate between DNA damage and translocation of AIF from mitochondria to the nucleus. On the basis of this information we used siRNA in human cells (HeLa) to selectively down-regulate the PAR metabolizing enzymes PARP-1, PARP-2 and PARG, either separately or in combination, to determine their individual contribution to modulate PAR as a cell death factor. Cell death mediated by the alkylating agent MNNG is characterized by the activation of PAR synthesis, moderate changes in ATP/ADP ratio of cells, activation of caspases 7 and 9 and finally AIF translocation from mitochondria to nucleus. We observed that only PARP-1, and not PARP-2 and PARG, are involved in this type of alkylation-induced cell death. PAR synthesized by PARP-2 as well as modified PAR molecules due to PARG reduction were unable to promote AIF translocation. We further investigated molecular signal cascades in oxidant-induced cell death in the presence or absence of PARG. As previously shown, PARG silenced cells are less sensitive to H2O2-induced cytotoxicity. The hypothesis was that specific PAR metabolites can act as a signalling factor in cell death, in particular monomeric ADP-ribose, that has been shown to bind to specific Ca2+ channels in vitro. We established a fluorescence spectroscopic determination of the influx of Ca2+ ions into the cytosol as an immediate response to an oxidative insult in mouse embryonic fibroblasts (MEFs) with a manipulated PAR metabolism. Using a set of PARP inhibitors, RNAi against PARP-1 and parp-1-/- cells, we investigated a PARP-dependent elevation of cytosolic Ca2+ after lethal doses of H2O2. Moreover, the translocation of AIF from mitochondria to nuclei was abolished when PARP function was eliminated leading to reduced cell death rates. Melastatin-like transient receptor potential channels 2 (TRPM2) were identified earlier required for Ca2+ gating from the extracellular space into the cytosol after oxidative insults. TRPM2 channels become activated by monomeric ADP-ribose, but not PAR and have been reported to act downstream of PARP-1. As free ADP-ribose is the final product of PARP/PARG cycle we tested the hypothesis, that PARG inhibition by RNAi could abolish the Ca2+ gating function of TRPM2. Indeed, we could demonstrate that cytososlic Ca2+ elevations after H2O2 were diminished when cells are silenced for the parg gene leading to a reduced cytotoxicity mediated by an impaired AIF translocation. Moreover, RNAi against TRPM2 as well as chemical inhibition of this channel showed a similar pattern in cytosolic Ca2+ fluxes after an oxidative insult. ADP-ribose loading of cells induces TRPM2 mediated Ca2+ fluxes in the absence of oxidative stress, suggesting that ADP-ribose is the key metabolite of the PARP/PARG cycle regulating TRPM2. We conclude that the interplay of PARP-1 and PARG controls a cell death pathway that operates between nucleus, cytoplasm, cell membrane and mitochondria and communicates via three types of chemical messengers: a nucleotide, a cation and proteins. Overall the results of this study contribute to the understanding of the specific involvements of PARP-1, PARP-2 and especially PARG in the response to DNA damage and the subsequent signal cascade that translates the duration of genotoxic insults into the cell fate of programmed cell death.