RESEARCH ◥ (PARG) (fig. S2D), or PARP1 inhibitors (fig. S1C) REPORT had no effect. PARP1-E988K behaved similarly to the wild type (fig. S1E and fig. S2E). Carba- nicotinamide adenine dinucleotide, a nonreactive AGING PARP1 substrate, abrogated the complex (fig. S2F). Thus, disruption of the PARP1-DBC1 complex by + NAD+ does not require NAD+ cleavage or a cova- A conserved NAD binding pocket lently attached ADP-ribose. In HEK 293T cells, FK866, an inhibitor of NAD+ biosynthesis (16), increased the PARP1- that regulates protein-protein DBC1 interaction (Fig. 1D), as did depletion of NAD+ by genotoxic stress (fig. S3A). Interventions interactions during aging that increased NAD+ abundance (5, 17)decreased the PARP1-DBC1 interaction (Fig. 1, E and F, and Jun Li,1 Michael S. Bonkowski,1 Sébastien Moniot,2 Dapeng Zhang,3* fig. S3B). The SIRT1-DBC1 interaction was slightly Basil P. Hubbard,1† Alvin J. Y. Ling,1 Luis A. Rajman,1 Bo Qin,4 Zhenkun Lou,4 diminished by FK866, possibly through the seques- Vera Gorbunova,5 L. Aravind,3 Clemens Steegborn,2 David A. Sinclair1,6‡ tration of DBC1 by PARP1 (fig. S3C). Together, these data indicate that NAD+ inhibits PARP1- DNA repair is essential for life, yet its efficiency declines with age for reasons that are DBC1 complex formation in cells. unclear. Numerous proteins possess Nudix homology domains (NHDs) that have no To better understand the mechanism by which known function. We show that NHDs are NAD+ (oxidized form of nicotinamide adenine NAD+ inhibits complex formation, we identified Downloaded from dinucleotide) binding domains that regulate protein-protein interactions. The binding the DBC1 domain necessary for interaction with of NAD+ to the NHD domain of DBC1 (deleted in breast cancer 1) prevents it from PARP1. A truncated version of DBC1 lacking an inhibiting PARP1 [poly(adenosine diphosphate–ribose) polymerase], a critical DNA repair NHD conserved region (DBC1D354-396)hadim- protein. As mice age and NAD+ concentrations decline, DBC1 is increasingly bound to paired PARP1 binding in cells (fig. S4, A to D), PARP1, causing DNA damage to accumulate, a process rapidly reversed by restoring the whereas a recombinant DBC1 mutant containing + + abundance of NAD . Thus, NAD directly regulates protein-protein interactions, the the NHD domain (DBC1-NHD, residues 239 to 553) http://science.sciencemag.org/ modulation of which may protect against cancer, radiation, and aging. bound to a truncated PARP1 lacking the catalytic domain (PARP1-DCAT, residues 1 to 654) (fig. S5, A andB)andwasnotabrogatedbyNAD+, indicating he oxidized form of nicotinamide adenine (10, 11), with a conserved domain similar to Nudix that residues outside the minimal NHD domain may dinucleotide (NAD+) is critical for redox re- hydrolases that hydrolyze nucleoside diphosphates be necessary for NAD+ to dissociate the complex. actions and as a substrate for signaling by but lacking catalytic activity due to the absence of DBC1 interacted with the BRCT [BRCA1 (breast thePARPs[poly(adenosinediphosphate– key catalytic residues (9, 12, 13). cancer 1) C-terminal] domain of PARP1 (fig. S6A) T 14 18 ribose) polymerases] and the sirtuins (SIRT1 DBC1 is known to inhibit SIRT1 ( ), so we ( ) but not with the PARP1 catalytic domain to SIRT7) in the regulation of DNA repair, energy tested whether DBC1 might also inhibit PARP1 or PARP2, which lacks a BRCT domain (fig. S6, metabolism, cell survival, and circadian rhythms, as a way to coregulate these two major NAD+- B to D). A BRCT-deficient PARP1 had higher activ- + 1–3 among other functions ( ). Raising NAD concen- responsive pathways. In human embryonic kidney ity than the wild type (fig. S6, E to F), together on November 22, 2020 trations or directly activating the sirtuins delays (HEK) 293T cells, a SIRT1-independent interaction indicating that PARP1-DBC1 is mediated by con- aging in yeast, flies, and mice (4–6). Increased betweenDBC1andPARP1wasdetected(Fig.1A tacts between the NHD and PARP1-BRCT. amounts of sirtuin and PARP1 activity are also and fig. S1, A and B). The PARP1 inhibitors PJ-34 We generated an atomic-resolution homology associated with improved health and longevity and 3-aminobenzamide (3-AB) had no effect on model for the human DBC1-NHD. Our model is in humans (4, 7). Thus, understanding how cells the interaction (fig. S1C), nor did overexpression based on five known crystal structures of Nudix modulate NAD+ and PARP1 activity may shed of the adenosine diphosphate (ADP)–ribose hy- domains from other proteins (Fig. 2A, fig. S7, light on new therapies for major diseases, such drolase MACROD1 (MACRO domain containing 1) and supplementary materials and methods). as diabetes and cancer, and possibly also the (fig. S1D) or the PARP1 catalytic mutant, PARP1- NAD+ had the best fit of all riboside nucleotides. aging process (3, 8). E988K (where E988K denotes Glu988→Lys988)(fig. Substitutions of the amino acids predicted to alter Whether NAD+ has a third role in cells as a S1E). Thus, PARP1-DBC1 binding is independent of NAD+ binding inhibited PARP1-DBC1 binding direct regulator of protein-protein interactions PARP1 catalytic activity. either slightly (K353A and P366A) or substan- is a matter of speculation (9). The protein DBC1 The PARP1-DBC1 complex was abrogated by tially (Q391A) (Fig. 2B and fig. S8, A and B). (deleted in breast cancer 1) is one of the most NAD+ in a concentration-dependent manner, Radio-labeled or biotin-labeled NAD+ directly abundant yet enigmatic proteins in mammals whereas the SIRT1-DBC1 interaction was unaf- bound to DBC1 (Fig. 2, C and D) and was com- fected within physiological ranges of NAD+ (15), peted off with unlabeled NAD+ (fig. S9, A to D). 1Department of Genetics, Paul F. Glenn Center for the except at 500 mM (Fig. 1B). This effect was sur- Partial deletion of the NHD (DBC1D354-396)orQ391A + Biology of Aging, Harvard Medical School, Boston, MA 02115, prisingly specific: 200 mM of nicotinamide mononu- (DBC1Q391A) reduced NAD binding, whereas N- USA. 2Department of Biochemistry, University of Bayreuth, cleotide (NMN), nicotinamide riboside (NR), terminal and C-terminal truncations did not (Fig. 3 95440 Bayreuth, Germany. National Center for adenosine, adenosine triphosphate, and ADP- 2, C to D, and fig. S9E). Mutation of C387, a Biotechnology Information, National Library of Medicine, National Institutes of Health, Bethesda, MD 20894, USA. ribose, or 500 mM of nicotinamide and its struc- residue close to Q391 on the same helix (Fig. 2A), + 4Division of Oncology Research, Department of Oncology, tural analog 3-AB (2 mM), had no effect on the also decreased NAD binding (Fig. 2D) and reduced Mayo Clinic, 200 1st Street SW, Rochester, MN 55905, USA. PARP1-DBC1 complex, and NADH (the reduced the responsiveness of the PARP1-DBC1 complex to 5 Division of Biology, 434 Hutchinson Hall, River Campus, form of NAD+) (200 mM) or adenine (200 mM) NAD+ (Fig. 2E). University of Rochester, Rochester, NY 14627, USA. + 6Department of Pharmacology, School of Medical Sciences, were less effective than NAD (Fig. 1C and fig. PARP1 activity was inhibited by DBC1 in vitro The University of New South Wales, Sydney, New South S2, A to C). (Fig.3A).Incells,knockingdownDBC1increased Wales 2052, Australia. DBC1 mutants lacking regions outside the both PAR [poly(ADP-ribose)] concentrations be- *Present address: Department of Biology, St. Louis University, NHD (DBC1 and DBC1 ) behaved sim- fore and after exposure to paraquat, H O ,or St.Louis,MO63103,USA.†Present address: Department of Pharma- 1-500 243-923 2 2 cology, University of Alberta, Edmonton, Alberta T6G 2H7, Canada. ilarly to full length DBC1 (Fig. 1C), whereas etoposide (Fig. 3B and fig. S10, A to D) and the ‡Corresponding author. Email: [email protected] MACROD1, poly(ADP-ribose) glycohydrolase abundance of mRNAs positively regulated by Li et al., Science 355,1312–1317 (2017) 24 March 2017 1of6 RESEARCH | REPORT PARP1 (Fig. 3C) (19, 20). Reintroduction of wild- treatment (fig. S11B), and increased both non- with the PARP1-DBC1 complex in the presence of type DBC1 reduced PARP1 activity and partially homologous end joining (NHEJ) and homologous NMN or DNA damage (fig. S13, A and B), though restored gene expression, whereas DBC1Q391A recombination (HR) pathways in a PARP1-dependent we cannot rule out other interactions. These had no effect (Fig. 3C and fig. S10, E and F). manner (Fig. 3F and fig. S11C). Similarly, NMN results are consistent with a model in which Reducing DBC1 lowered the abundance of the treatment reduced the number of gH2AX foci in binding of NAD+ to the DBC1-NHD regulates the phosphorylated form of histone H2AX (gH2AX) paraquat-treated primaryhumanfibroblasts(Fig. two major pathways of DNA repair. (Fig. 3D), reduced DNA fragmentation (Fig. 3E 3G and fig. S12). No other major DNA repair DNA repair declines with age (21), in concert and fig. S11A), increased cell survival after paraquat proteins appeared to change their interactions with lower PARP1 activity (7). Our data indicate Downloaded from http://science.sciencemag.org/ on November 22, 2020 Fig. 1. Regulation of the PARP1-DBC1 interaction by NAD+. (A) Endogenous (D to F) The PARP1-DBC1 interaction after treatment with (D) FK866 or (E) DBC1 and PARP1 interact. IgG, immunoglobulin G; IP, immunoprecipitation. NMN for 24 hours or (F) in cells overexpressing NMNAT1, an NAD+ salvage (B)NAD+ dissociates the PARP1-DBC1 interaction. (C)EffectsofNAD+ and pathway gene to raise NAD+. DMSO, dimethyl sulfoxide; GAPDH, glyceraldehyde- structurally related molecules on the PARP1-DBC1 interaction. Flag-DBC1 was 3-phosphate dehydrogenase. Error bars indicate mean ± SEM. (D and E) One- incubated with molecules (200 mM) for 1 hour and then probed for PARP1. way analysis of variance (ANOVA) and Sidak’s post-hoc correction. (F) Unpaired NMN, nicotinamide mononucleotide; NR, nicotinamide riboside; WT, wild type. two-tailed t test.