Structural basis for allosteric PARP-1 retention on DNA breaks

DNA death grip Poly(ADP-ribose) polymerase–1 (PARP-1) binds to DNA breaks and recruits DNA repair components. Cancer-killing PARP-1 inhibitor (PARPi) compounds all block the same catalytic site but exhibit vastly different efficacy. Zandarashvili et al. investigated the molecular impact of PARPi binding to PARP-1 (see the Perspective by Slade and Eustermann). Different PARPi molecules perturb PARP-1 allostery in diverse manners: Some drive allostery to promote release of PARP-1 from DNA, and others drive allostery to promote retention. These insights help explain the different efficacies in the clinic and enable conversion of a pro-release, ineffective cancer-killing compound to a pro-retention, more effective PARPi. Science, this issue p. eaax6367; see also p. 30 Different poly(ADP-ribose) polymerase–1 (PARP-1) inhibitors used to treat cancer either trap or release PARP-1 at DNA break sites. INTRODUCTION Poly(ADP-ribose) polymerase–1 (PARP-1) is an abundant enzyme in the cell nucleus that regulates genome repair by binding to DNA damage sites and creating the poly(ADP-ribose) posttranslational modification. PARP-1 hyperactivity leads to cell stress or death associated with many prominent diseases (e.g., cardiovascular disease and several common neurodegenerative disorders). PARP-1 has notably emerged as an effective clinical target for a growing list of cancers. Clinical PARP-1 inhibitor (PARPi) compounds all bind at the same location at the catalytic center of the enzyme to block the binding of substrate nicotinamide adenine dinucleotide (NAD+) and prevent poly(ADP-ribose) production, yet they exhibit vastly different outcomes in tumor cell killing and efficacy in the clinic—a paradox that has confounded the development of PARPi. The resolution of this paradox likely lies in the realization that the most effective PARPi compounds trap PARP-1 at the site of a DNA break, generating a lesion that becomes cytotoxic, especially in tumor cells with deficiencies in the repair of DNA strand breaks. RATIONALE The molecular roots of PARP-1 trapping on DNA remain poorly understood. We focused on the retention of PARP-1 on damaged DNA, examining a panel of PARPi that included those currently approved for clinical use. Solution biophysical approaches, especially hydrogen/deuterium exchange mass spectrometry (HXMS), combined with x-ray structures and a battery of biochemical assays, were used to interrogate the molecular impact of PARPi binding to PARP-1 engaged on sites of DNA damage. Structure-guided modification of PARPi through medicinal chemistry was combined with chromatin fractionation to monitor trapped PARP-1 and with cell survival assays to assess PARPi efficacy, so as to probe the molecular underpinnings of the variable outcomes between clinical PARPi. RESULTS HXMS experiments revealed that a critical allosteric regulatory domain of PARP-1, the helical domain (HD), is affected in distinct ways depending on the particular PARPi engaged in the NAD+-binding site adjacent to the HD. Certain PARPi destabilized specific HD regions, some had no effect on the HD, and others actually stabilized regions of the HD. PARPi that destabilized the HD increased PARP-1 affinity for DNA and retained PARP-1 on DNA breaks. Conversely, PARPi that stabilized the HD decreased PARP-1 affinity for DNA breaks. PARPi molecules were thus classified into three types: type I, allosteric pro-retention on DNA; type II, non-allosteric; and type III, allosteric pro-release from DNA. X-ray structure analysis identified PARPi contacts with the HD structural element helix αF, which was established to be the discriminating factor between the types of PARPi. We found that type I PARPi contact helix αF to initiate an allosteric chain reaction that travels ~40 Å through the multidomain PARP-1 molecule and culminates in increased DNA binding affinity. Structure-guided mutagenesis of helix αF disrupted PARPi contacts and abrogated the allosteric effects of a type I inhibitor, transforming it into a non-allosteric type II inhibitor. Other mutations that disrupted PARP-1 allostery, including one identified in a de novo PARPi-resistant patient with ovarian cancer, also prevented type I PARPi from retaining PARP-1 on a DNA break. Type III PARPi influenced PARP-1 allostery in a manner that reduced DNA binding and favored DNA release. Structure-inspired modification of a pro-release (type III) inhibitor converted it to a pro-retention (type I) inhibitor that conferred potent PARP-1 trapping within the cellular context and increased its ability to kill cancer cells. CONCLUSION Our findings establish the impact of clinical PARPi on PARP-1 allostery and demonstrate that allostery plays a critical role in cellular PARP-1 trapping and can increase potency toward cancer cell killing. The results illuminate the molecular basis for the fine-tuning of PARPi to achieve allosteric effects and to influence PARP-1 retention on DNA damage and trapping on chromatin in cells. In contrast to cancer, other diseases would seem to benefit from PARP-1 inhibition but not cell death (e.g., cardiovascular disease, neurodegenerative diseases, and inflammation). Our studies provide the molecular understanding and the appropriate toolset to create and evaluate tunable PARPi for clinical applications where PARP-1 trapping and associated cytotoxicity are either desirable or undesirable in specific patients. PARPi impact on PARP-1 allostery. PARP-1 (tan) uses multiple domains to detect DNA breaks, and DNA damage detection is allosterically coupled to poly(ADP-ribose) production. PARPi bind to the catalytic domain to inhibit PARP-1 activity. Type I PARPi influence PARP-1 allostery and retain PARP-1 on DNA (left, UKTT15 in green), whereas type III PARPi perturb PARP-1 allostery and release PARP-1 from DNA (right, veliparib in red). Type II PARPi do not influence PARP-1 allostery. The success of poly(ADP-ribose) polymerase–1 (PARP-1) inhibitors (PARPi) to treat cancer relates to their ability to trap PARP-1 at the site of a DNA break. Although different forms of PARPi all target the catalytic center of the enzyme, they have variable abilities to trap PARP-1. We found that several structurally distinct PARPi drive PARP-1 allostery to promote release from a DNA break. Other inhibitors drive allostery to retain PARP-1 on a DNA break. Further, we generated a new PARPi compound, converting an allosteric pro-release compound to a pro-retention compound and increasing its ability to kill cancer cells. These developments are pertinent to clinical applications where PARP-1 trapping is either desirable or undesirable.