Polyadenosine Diphosphoribose Glycohydrolase
Created by Morgan Flynn
Polyadenosine diphosphoribose glycohydrolase (PDB accession 3SIG; EC 3.2.1.143, pI 9.90), a 29717.66 Da protein derived from Thermomonospora curvata, is the only enzyme that catalyzes the intracellular hydrolysis of poly(ADP-ribose) into free monomeric ADP-ribose in bacteria and mammals (3,9). The poly(ADP-ribose) glycohydrolase, or PARG, catalytic domain is a distant member of the ubiquitous ADP-ribose-binding macro domain family of proteins involved in the regulation of genome stability (5, 15). In the case of PARG, the conserved macro domain also functions as a ligand binding pocket with a high-affinity for Poly(ADP-ribose) (5). The ability of the fold to fulfill both catalytic and binding roles may explain the wide distribution of the macro domain throughout enzymes and across species.
Poly(ADP-ribose), or PAR, is a linear or branched chain of negatively charged, repeating ADP-ribose units joined via a unique glycosidic ribose-ribose linkage, with a branching frequency of up to 3% (Fig. 2) (1, 17). Due to its nucleotide component, PAR is also known as the “third type of nucleic acid”, although its biological role remains more mysterious than that of RNA and DNA (6, 18). Post-translational modification of proteins by poly(ADP-ribosyl)ation is vital to the preservation of genomic stability due to its contributions to DNA damage detection and repair, chromatin modification and transcription. Polymerization of ADP-ribose is also implicated in cell death pathways insulator function and mitotic apparatus function, though exact mechanisms for some actions are unknown to date. Poly(ADP-ribosyl)ation is essential for many physiological and pathophysiological outcomes, including carcinogenesis, aging, inflammation and neuronal function (4, 6).
Poly(ADP-ribosyl)ation is carried out by Poly(ADP-ribose) polymerases (PARPs). PARP-1, the most studied of the PARP enzymes, is a DNA nick-sensing enzyme that activates when bound to DNA breaks. It is a key player in DNA repair and the arrest of untimely cell death. PARP-1 is also implicated in the recruitment and synchronization of DNA repair enzymes and it modulates base excision repair (BER) capacity (20). It is well established that DNA base excision repair relies upon poly(ADP-ribosyl)ation. Poly(ADP-ribose) serves as a source of ATP to drive BER.
Activation of PARP-1 is one of the first markers of DNA damage. Upon recognition of and binding to a DNA strand break an enzyme-driven shuttling of PARP-1 begins, causing opening of the chromatin. Using nicotinamide adenine dinucleotide (NAD+) as an ADP-ribose donor, PARP-1 cleaves NAD+ into nicotinamide and ADP-ribose and then polymerizes the ADP-ribose onto acceptor proteins and itself (Fig. 1, 2)(4). The highly negative charge of covalently bound PAR dramatically affects the functionality of target proteins (4). Histones are highly poly(ADP-ribosyl)ated proteins. The sizeable negative charge conferred to histones that have experienced poly(ADP-ribosyl)ation creates an electrostatic repulsion between the histones and DNA. This process supports chromatin remodeling, DNA repair, and transcriptional regulation (4). Several transcription and DNA replication factors and signaling molecules are known to undergo poly(ADP-ribosyl)ation by PARP-1.
The hydrolysis of polymerized ADP-ribose is equally vital to genomic stability. The intracellular processes influenced by PARP-1 are also affected by the action of PARG (Fig. 2). The open chromatin allows PARG to enter the nucleus, where it moves to the site of PARP activity and interacts with DNA repair factor XRCC1, which activates PARP-1. As the process continues PARG becomes overabundant and due to excessive PARG, the PAR concentration decreases and thus the chromatin is able to adopt its original structure once again (20).
Poly(ADP-ribosyl)ation is transient (t1/2 < 1 minute) but very extensive in vivo, which points to a concerted activation of PARP and PARG. Polymer chains can comprise anywhere from 2 to 200 units on protein acceptors (4, 18). PAR metabolism relies on rapid turnover of PAR, which is solely catalyzed by PAR glycohydrolase (PARG) in eukaryotic cells. PARG is thought to reverse the action of PARP-1 by lowering transcription and promoting chromatin condensation as previously described (5, 19). The dynamic nature of cellular PAR metabolism rivals the enzymology of other reversible modifications, such as phosphorylation and acetylation. PARG-mediated breakdown of PAR yields a dramatic increase in free nuclear ADP-ribose, yet no protein modules are known that specifically recognize the ADP-ribose nucleotide (5).
Protein ADP-(ribosyl)ation has been studied for more than 40 years, but while much is understood about PAR and PAR polymerases, not much is known about the structure of PARG and the role it plays in cellular PAR metabolism (5). The first PARG gene was cloned in 1997, but there remains relatively scarce data about the PARG enzyme. Its structure was unknown until the crystal structure of PARG, derived from the bacterium Thermomonospora curvata, was presented by D. Slade, et.al., on September 4, 2011 (1). The derivation of PARG from T.curvata challenges the previously held assumption that there is no poly(ADP-ribosyl)ation in bacteria. To determine its structure, crystallized PARG, with and without ADP-ribose (PDB accession 3SIH) and with the PARG inhibitor,adenosine diphosphate hydroxymethyl pyrrolidinediol (PDB accession 3SII), was isolated from T. curvata and solved to 1.5 Angstroms (1). The crystallized structure of PARG in complex with ADP-ribose is of particular interest as we attempt to improve our understanding of how PARG activity controls reversible protein poly(ADP-ribosyl)ation and how defects in this regulation may be linked to human disease.
PARG exists as a monomer composed of 43% helices (primarily alpha-helices with 2 segments of 3/10 helix) and 19% beta sheets with the remainder organized into random coils and turns(1). It contains an ADP-ribose-binding macro domain fold with an amino-terminal extension (1). It is the macro domain that characterizes the unique intracellular function of PARG. ARH3 proteins show functional similarity to PARG, capably cleaving PAR as well, but this is where the similarity ends. ARH3 is a member of the ARH/DraG protein family and it consists of an all-alpha-helical fold with an active site containing two Mg2+ ions (14). The PARG active site lacks metal ions altogether.
Comparable tertiary structures to PARG do exist and are found in an Escherichia Coli putative YmdB protein (PDB accession 1SPV; Z score16.2; E value 2.7,19% identity) and an Archaeoglobulus fulgidus hypothetical protein AF1521 (PDB accession 2BFQ; Z score 15.6; E value 0.003, 16% identity) (1, 9). These proteins, like PARG, are members of the far reaching family of macro domain proteins. In fact, AF1521 and YmdB also share primary structural similarities with T. curvata PARG, as indicated by the results of Protein BLAST queries. AF1521 is a compact single domain structure and is the best studied of all the macro domain proteins. It displays mixed α/β structure and contains a single seven-stranded mixed β-sheet with strand order 1276354 and five α-helices. Within the fold of the Af1521 macro domain is a phosphate-binding loop homologous to those of the P-loop family of nucleotide hydrolases (13), including PARG. The affinity of ADP-ribose for the Af1521 macro domain is also high however, unlike PARG, AF1521 is capable of binding both monomeric ADP-ribose and ADP-ribose polymers (16). The simplest PAR chain capable of binding with PARG is an ADP-ribose dimer.
The ADP-ribose-binding cavity of PARG is flanked on one side by a diphosphate-binding loop that is highly conserved between PARG and other macro domain structures. A stretch of amino acids corresponding to a central portion of the PARG-specific signature sequence, GGg-X6–8-vQEE, sits opposite the diphosphate binding loop (12). Also specific to glycosidase enzymes such as PARG are 2 sequential acidic residues essential for activity that surround the glycosidic bond to be hydrolyzed. Glu 114 and Glu 115 serve this purpose for T. curvata PARG. Structural alignment dictates that the PARG-specific loop inserts into the macro domain fold to accommodate the Glu 115 side chain that projects into the PARG active site. The specificity of the PARGdiphosphate-loop allows only PARG and no other macro domain protein to hydrolyze PAR (1). The highly conserved PARG-specific signature sequence and macro domain suggest that PARG functions similarly across species and further cements the future of PARG as a protein of interest for the treatment of human disease.
The ligand-PARG complex structures reveal that ADP-α-ribose occupies a comparable position in many ADP-ribose macro domain family proteins (5). A minor rearrangement occurs for Val 226 and Phe 227 upon ADP-ribose binding. The Phe 227 side chain and the ribose" moiety of ADP-ribose are in close proximity, thereby ensuring intimacy between theribose” O4 and one of the phosphate groups within the active site (1). The ribose” moiety also forms hydrogen bonds with Glu 114 and Glu 115 via the 2’ -OH group and 1’ -OH group respectively (1). Few direct contacts exist between the adenosine moiety and the active site, with the exception of water-mediated polar interactions. The adenosine ribose’ moiety is pointedly less accessible than the ribose” moiety. Substituents including other ADP-ribose molecules are prevented from binding with the ribose’ 2’-OH group. Such an addition would require a dramatic and undesirable repositioning of the PARG carboxy-terminal α-helix causing exposure of the hydrophobic core of the macro domain.
The PARG active site, being limited by strict steric constraints, allows for an α(1”→2’) O-glycosidic linkage with an additional ADP-ribose group, providing further insight into PAR binding mechanisms. Recent models suggest that the (n-1) adenine moiety is partially enclosed by Ala 110, Ala 112, and Val 226 while the (n-1) ribose’ is close to Ser 98 and Gly 104. The lack of obvious binding sites for the (n-1) phosphate and ribose” moieties combined with the marked conformational freedom of the (n-1) ribose” moiety with respect to the terminal ADP-ribose and its associated (n-1) ribose suggest that bacterial PARG does not specifically bind additional PAR elements (1).
D. Slade, et. al. proposed a mechanism for PARG catalysis using structural discoveries from the T. curvata PARG complex with ADP-ribose and the corresponding PAR-PARG model (1). The key ribose–ribose O-glycosidic linkage must lie in direct hydrogen-bonding contact with Glu 115. Binding of the PAR terminus establishes this while constraining the conformation of the terminal ribose′′. Formation of a putative oxocarbenium intermediate is supported by the protonation of the (n − 1) PAR ribose′ 2′-OH leaving group via Glu 115, and by the stabilization of the positively charged oxocarbenium through close proximity with the terminal diphosphate group (1). The oxocarbenium intermediate is attacked by an ideally positioned water molecule and activated through concomitant deprotonation by Glu 115. This is followed by the release of a single ADP-β-ribose′′ and (n−1) PAR.
Both mutagenesis (PARG proteins E 114A and E 115A) and binding studies support this newly proposed mechanism. Although simultaneous mutation of both catalytic residues, Glu 114 and Glu 115, renders the enzyme inactive without disrupting the overall PARG fold, the isothermal titration calorimetry (ITC) ADP-ribose binding studies demonstrate that while mutation of Glu 115 into an alanine has no effect on binding, Glu 114-Ala mutation results in an approximate tenfold decrease in binding affinity (1). This supports the theory that Glu 114 is involved only in substrate binding, while Glu 115 may play a role in acid-base catalysis. The enzyme is also deactivated by the mutation of Phe 227 necessary to position the terminal ribose”, further evidence that the active cannot bind additional PAR elements. Likewise, mutations of Ser 98 and Val 226, implicated in binding the (n-1) ADP-ribose, greatly reduce PARG activity in T. curvata. Mutations of the corresponding catalytic residues in PARG from Homo sapiens have a similar effect on enzyme activity, suggesting a universal catalytic mechanism for mammalian and bacterial PARG (1).
The PARG inhibitor, Adenosine diphosphate (hydroxymethy1)pyrrolidinediol, or ADP-HPD, is an amino analog of ADP-ribose that acts as a highly potent, noncompetitive, and specific inhibitor of PARG, binding to Tyr 76 in the PARG active site (Fig. 3). It has no affect on the activity of PARP even at high concentrations. The H-bonding pattern between E 114 and E 115 and ADP-HPD is altered relative to the complex with ADP-ribose and V226 moves away from the active site when binding with ADP-HPD occurs (1). As a result, the active site is made unavailable for binding with any other ligand.
PARP-1 inhibition is currently an active and highly competitive area of investigation and it is much farther along than the study of PARG action and inhibition. The relatively new concept of incorporating control over PAR metabolism into the management of disease is an attractive avenue to explore. Comprehending the complex roles of PARP and PARG in the cell death process is tantamount to understanding how PARP and PARG inhibition can protect tissues from or enhance cytotoxicity depending on the nature and severity of DNA damage.
In a normal eukaryotic cell, PARP arrests cell death by facilitating DNA repair through the production of PAR from NAD+. Over-activation of the PAR pathway increases nicotinamide concentration and decreases cellular NAD+ and ATP, which leads to cell death. PARP-mediated cell death tends to be necrotic, a less controlled and less desirable mechanism than apoptosis, which poses danger for bystander cells. DNA damage, the most important factor in the regulation of PAR metabolism, can stimulate the catalytic activity of PARP-1 by about 500-fold (16). Recent work indicates that excessive creation of PAR by PARP-1 in response to high levels of DNA damage also stimulates the release of apoptosis-inducing factor (AIF) (10), further contributing to cell death. This PARP-mediated cellular suicide mechanism has been implicated in the pathomechanism of stroke, myocardial ischemia, diabetes, diabetes-associated cardiovascular dysfunction, shock, traumatic central nervous system injury, arthritis, colitis, allergic encephalomyelitis, and various other forms of inflammation (4). Blocking poly(ADP-ribose) metabolism by inactivating PARP has repeatedly been shown to reduce ischemia injury(16,17). While PARP-1 disruption blocks necrotic cell death, apoptosis may still occur, making PARP-1 inhibitors invaluable preservers of tissue.
In 2003, Xi-Chun Lu, et. al. investigated whether disrupting PAR metabolism via PARG inhibition could achieve similar protection to that given when PARP-1 is disrupted (17). They presented the first evidence that PARG inhibitors can ameliorate ischemic brain damage in vivo, supporting the role of PARG as a new therapeutic target for treating ischemia injury (17). Since then, the importance of PAR catabolism to the normal physiological state has been demonstrated repeatedly in the laboratory. Mice homozygous for an engineered PARG gene mutation (PargΔ4/Δ4), resulting in a complete depletion of all PARG isoforms (i.e., Parg null), die at embryonic day 3.5 (E3.5)(7). Embryos having the gene for PargΔ4/Δ4 and embryonic trophoblast stem cells accumulate high levels of PAR and undergo increased cell death by apoptosis as a result of NAD+ depletion associated with excessive accumulation of PAR (7). Likewise, Drosophila melanogaster mutants containing dysfunctional PARG exhibit increased lethality in the larval stages at normal developmental temperatures (8). These and other studies have shown that PARG inactivation offers protection against tissue injuries and cell death in animal models of cerebral ischemia, traumatic brain injuries (TBI), Parkinson’s disease, myocardial ischemia, type I diabetes, hemorrhagic shock, septic shock, inflammatory bowel disease, intestinal ischemia and other inflammatory responses (17). PAR catabolism is accelerated under genotoxic stress conditions related to such injuries and disease processes. This phenomenon is largely attributable to the enzymatic activity of PARG and is fueling current interest in PARG as it relates to preservation of the normal physiological state.
The same qualities that imbue PARG with tissue-protective faculties during ischemic insult make it an attractive enzyme to research oncologists. Cancer diagnoses are steadily rising at a frightening rate. As per the World Health Organization (WHO) data and statistics, 15 million new cancer patients will be diagnosed in the year 2020. Much effort has been put in by researchers and scientists to find drugs helpful in the treatment of cancers—the latest being the PARP and the PARG inhibitors (20). The intimate involvement of PARP and PARG in programmed and necrotic cell death makes them ideal candidates for use in the treatment and inhibition of cancers. Many currently employed chemotherapeutic agents can be less devastating to healthy tissues when used in combination with PARP and PARG inhibitors and the disruption of PAR metabolism alone in targeted cells may prove beneficial in the treatment of cancer.
New insights into PARG structure and catalytic activity are greatly improving overall understanding of the role PARG plays in the reversible Poly(ADP-ribosyl)ation of proteins and may answer many questions relating PARG activity and human disease. Researchers and pharmaceutical companies are eager to continue both experimental and clinical trials with PARP and PARG inhibitors. Since both PARP and PARG inhibitors now have well delineated sensitivity and specificity new insight into their medical potential is forthcoming. Their introduction in clinical trials is new hope for cancer patients. Perhaps the development of small, cell permeable PARG inhibitors allowing artificial interference with PAR metabolism, to manipulate the physiology of health and disease, is on the horizon (1).