PARG activity is required for cell death by parthanatos

Cell death by a non-apoptotic pathway termed parthanatos is induced by hyperactivation of the DNA damage sensor PARP1, which uses NAD+ as a substrate to catalyse the poly-ADP-ribosylation (PARylation) of proteins. Parthanatos has been implicated in several pathological conditions such as ischaemia-reperfusion injury and neurodegenerative processes, including Alzheime[r]s and Parkinso[n]s diseases, as well as rare genetic disorders characterized by PARP1 hyperactivation. However, the precise sequence of molecular events by which excessive PARP1 activity causes cell death is currently unclear. Here we show that, in addition to PARP1-dependent PARylation, the execution of parthanatos also requires the hydrolysis of poly-ADP-ribose (PAR) chains by the glycohydrolase PARG. While complete inhibition of PARG activity prevents parthanatos, low levels of residual PARG activity are sufficient to support cell death by this pathway. Due to a phenotypic discrepancy between CRISPR/Cas9-generated PARG KO cells and PARG inhibitor-treated cells, we uncovered a new PARG splice variant predicted to encode an isoform of 53 kDa, termed PARG53, and provide evidence for the incorrect annotation of the previously reported isoforms PARG55 and PARG60. PARP1 hyperactivation during parthanatos leads to rapid and profound depletion of NAD+ and ATP pools, but full PARG inhibition only prevents the depletion of ATP, indicating that the depletion of these core metabolites can be uncoupled, and that ATP depletion is more closely correlated with cell death. This work sheds light on the initial steps of parthanatos induction, demonstrating that PAR formation and PAR hydrolysis are both required for this type of cell death.