PerspectiveStructural Biology

PARP-1 Activation—Bringing the Pieces Together

See allHide authors and affiliations

Science  11 May 2012:
Vol. 336, Issue 6082, pp. 678-679
DOI: 10.1126/science.1221870

The repair of DNA strand breaks is crucial for cell survival. In higher eukaryotes, cellular responses to DNA strand breaks are coordinated by a process called poly(ADP-ribosyl)ation (where ADP is adenosine diphosphate). This highly dynamic process begins with the hydrolysis of NAD+ (nicotinamide adenine dinucleotide) by poly(ADP-ribose) polymerases (PARPs), resulting in the polymerization of ADP-ribose moieties onto a substrate protein. The most active PARP in DNA damage recognition, PARP-1, is highly stimulated by single- and double-strand breaks (1). Because it is a clinically important molecular target in the treatment of several cancers, including breast and ovarian cancers (2), understanding how PARP-1 is activated after binding DNA strand breaks is an ongoing challenge. On page 728 of this issue, Langelier et al. (3) report key structural information about PARP-1 domain organization around a DNA double-strand break (DSB) that explains this activation.

DNA-dependent PARP activation has been characterized extensively at the biochemical level. The identification of zinc-finger (Zn) structures in the amino-terminal region of PARP-1 as DNA-binding modules helped to elucidate the molecular mechanisms underlying DNA-dependent PARP-1 activation (4). However, PARP-1 is a large multidomain protein, in which the amino-terminal DNA-binding domain is separated from the carboxyl-terminal catalytic domain (CAT) by several regions, including a central automodification domain (AD) and a WGR domain [defined by the conserved residues tryptophan (W), glycine (G), and arginine (R)] of unknown function (see the first figure). It has been unclear how signals from the DNA-interacting regions are relayed to the catalytic domain. Crystallographic studies have elucidated individual PARP-1 domains and molecular interactions with DNA or inhibitors, but a global mechanistic understanding of the domain rearrangements that underlie PARP-1 activation has been elusive.

Modular domain architecture of human PARP-1.

Approximate domain boundaries are indicated by residue numbers. ART, ADP-ribosyl transferase domain.

To address these issues, Langelier et al. constructed the minimal assembly of domains that supports DNA DSB-dependent PARP-1 activation, consisting of the first and third zinc fingers coupled to the WGR and catalytic domains (Zn1, Zn3, WGR-CAT), and used it as a basis for their structure-function study. Although lacking Zn2 and the protein interaction module BRCT [Breast cancer type 1 susceptibility protein (BRCA1) C-terminal region], this truncated but active PARP-1 structure is the nearest to the full-length protein crystallized to date.

Model for DNA double-strand break-dependent activation of PARP-1.

Langelier et al. show that upon DSB generation, PARP-1 engages DNA as a monomer and adopts a compact conformation centered on the WGR domain. In this conformation (inset, magnified view), the WGR domain makes interdomain contacts with Zn1 and Zn3 and contributes to the formation of the DNA-binding interface. This rearrangement triggers a molecular switch in the HD region of the CAT domain, alters the flexibility and dynamics of the ART domain, and thereby activates PARP-1. It also places the BRCT/AD close to the CAT, making it available for intramolecular modification.

On the basis of their structural and functional data, Langelier et al. propose a PARP-1 activation model that involves the formation of a compact PARP-1–DNA complex, in which the DSB-interacting zinc fingers and the CAT domain lie on opposite faces of the WGR (see the second figure). This conformation distorts the helical subdomain (HD) within the CAT, thereby inducing a change in conformational flexibility of the CAT and repositioning the AD near the active site. This rearrangement explains the propensity of PARP-1 for poly(ADP-ribosyl) ating itself rather than target protein substrates, but raises the question of how PARP-1 accommodates the broad range of its known protein substrates.

The tertiary model (see the second figure) also indicates that PARP-1 does not form a dimer at the DSB, consistent with a previous structural study in which PARP-1 interacted with damaged DNA as a monomer (5). Yet, PARP-1 dimerization has been a recurrent theme in the literature, and many have reported PARP-1 as a catalytic dimer (68). Self-association behavior of PARP-1 through BRCT domains has also been reported (9), but a more recent characterization of the solution structure of PARP-1 BRCT domain indicated that this domain does not engage in dimer formation (10). Although Langelier et al. show that PARP-1 can be fully automodified as a monomer, it may also function as a dimer in certain contexts; for example, inactive mutants can complement each other (11).

Although the poorly characterized WGR domain is necessary for PARP-1 activation (11), the crucial contribution of this domain to the formation of a DNA damage recognition interface along with Zn1 and Zn3, as shown by Langelier et al., is unexpected. This central role of the WGR domain in PARP-1 activation may explain the observation that PARP-1 automodification is mainly restricted to a small region of the AD (11). In the model proposed by Langelier et al., the PARP-1-DNA structure has limited flexibility, and the small region of the AD prone to poly(ADP-ribosyl)ation is found near the catalytic pocket (12). Thus, Langelier et al.'s results could lead to a redefinition of what is currently considered the AD of PARP-1.

Furthermore, one could hypothesize from the model of Langelier et al. that the WGR could mediate PARP activation through DNA-independent signals. It is of considerable interest to determine whether atypical substrates will only trigger subtle distortions to the compact structure model or induce larger domain rearrangements. The WGR domain deserves further characterization as a substrate-recognition module. A WGR-centered activation mechanism would provide a strong rationale for developing small-molecule inhibitors specific to PARP-1.

Langelier et al.'s study is an impressive advance in our understanding of PARP-1 activation by DSBs. It provides the basis for unraveling the mechanism of PARP-1 activation in the context of a full-length protein. Examining essential domains for the binding of single-strand breaks will be highly informative, especially as regards Zn2, which appears to be critical for the interaction of PARP-1 with single-strand breaks (13). Furthermore, the study paves the way for designing a new generation of PARP-1–specific inhibitors that could disrupt the domain-domain interactions required for forming an active PARP-1 conformation.

Although PARP-1 displays robust DSB-binding activity (14), its identification as an integral component in either of the two main DSB repair pathways, homologous recombination and nonhomologous end joining (NHEJ), has not been rigorously demonstrated. However, recent studies strongly support interplay between DNA–protein kinase (DNA–PK) and PARP-1 in the DNA damage response through the NHEJ pathway (15, 16). The mechanism by which PARP-1 and DNA–PK bind to DNA DSBs must be elucidated before the precise role of these proteins in DNA repair can be understood.


Navigate This Article