Structure of human ADP-ribosyl-acceptor hydrolase 3 boundto ADP-ribose reveals a conformational switch that enablesspecific substrate recognitionReceived for publication, April 19, 2018, and in revised form, May 30, 2018 Published, Papers in Press, June 15, 2018, DOI 10.1074/jbc.RA118.003586

Yasin Pourfarjam‡1, Jessica Ventura‡1, Igor Kurinov§, Ahra Cho‡, Joel Moss¶2, and X In-Kwon Kim‡3

From the ‡Department of Chemistry, University of Cincinnati, Cincinnati, Ohio 45221, §Cornell University, Department of Chemistryand Chemical Biology, Northeastern Collaborative Access Team Advanced Photon Source (NE-CAT APS), Argonne, Illinois 60439,and ¶Pulmonary Branch, NHLBI, National Institutes of Health, Bethesda, Maryland 20892

Edited by Patrick Sung

ADP-ribosyl-acceptor hydrolase 3 (ARH3) plays importantroles in regulation of poly(ADP-ribosyl)ation, a reversible post-translational modification, and in maintenance of genomicintegrity. ARH3 degrades poly(ADP-ribose) to protect cellsfrom poly(ADP-ribose)– dependent cell death, reverses serinemono(ADP-ribosyl)ation, and hydrolyzes O-acetyl-ADP-ri-bose, a product of Sirtuin-catalyzed histone deacetylation.ARH3 preferentially hydrolyzes O-linkages attached to the ano-meric C1� of ADP-ribose; however, how ARH3 specifically rec-ognizes and cleaves structurally diverse substrates remainsunknown. Here, structures of full-length human ARH3 boundto ADP-ribose and Mg2�, coupled with computational model-ing, reveal a dramatic conformational switch from closed toopen states that enables specific substrate recognition. The glu-tamate flap, which blocks substrate entrance to Mg2� in theunliganded closed state, is ejected from the active site when sub-strate is bound. This closed-to-open transition significantlywidens the substrate-binding channel and precisely positionsthe scissile 1�-O-linkage for cleavage while securing tightly 2�-and 3�-hydroxyls of ADP-ribose. Our collective data uncover anunprecedented structural plasticity of ARH3 that supports itsspecificity for the 1�-O-linkage in substrates and Mg2�-depen-dent catalysis.

Regulation of poly(ADP-ribosyl)ation (PARylation),4 a re-versible post-translational modification (PTM) of proteins,

is required for maintaining genomic integrity and cellularresponses to DNA damage (1, 2). The addition of poly(ADP-ribose) (PAR) to proteins by PAR polymerase 1 (PARP1; alsoknown as ARTD1) plays a pivotal role in the repair of DNAsingle- and double-strand breaks (3–6). However, excessivePARylation by PARP1 often interferes with protein functionand may activate parthanatos, a PAR-dependent cell death path-way, which involves the translocation of PAR to the cytoplasm,eventually triggering the PAR-dependent release of apopto-sis-inducing factor (AIF) from mitochondria (7, 8). AIF releasedfrom mitochondria then goes to the nucleus, leading to DNAcleavage and cell death. The cellular level of PARylation and PAR istherefore tightly controlled by PAR synthesis and turnover.

In mammals, two enzymes, ADP-ribosyl-acceptor hydrolase3 (ARH3; also known as ADPRHL2) and PAR glycohydrolase(PARG), function in tandem to reverse PARylation (9). Thesehydrolytic enzymes commonly cleave the �(1�-2�) O-glycosidiclinkages in PAR chains (Fig. 1a) (10, 11). PARG has both exo-and endoglycohydrolase activities, acting at terminal and inter-nal sites of PAR chains, respectively (9, 12–14). Thus, in addi-tion to free ADP-ribose (ADPR), PARG generates short chainsof PAR that serve as a potent cell death signal (9). In contrast,ARH3 appears to catalyze primarily exocytic cleavage of PAR,generating free ADPR (9). Consistent with this differencein biochemical activity, ARH3 protects cells from oxidativestress–induced parthanatos by lowering the cytoplasmic PARlevel (9). ARH3 knockout cells are healthy under unstressedconditions. However, following H2O2-induced DNA damage,these cells show accumulation of cellular PAR, in particular inthe cytoplasm where it interacts with mitochondria, leading toAIF cleavage and release, enhanced nuclear accumulation ofAIF, and overall increased activity of the parthanatos pathway(9). Taken together, these findings indicate that ARH3 is a keyenzyme that not only controls PAR content but also determinescell fate during the DNA damage response.

PARG is unable to reverse a protein-bound mono(ADP-ribo-syl)ation (MARylation), the last step in completely erasing poly-(ADP-ribosyl)ation (15). This MARylation mark, particularlyon acidic residues, is removed by macrodomain-containingproteins with MAR hydrolytic activity such as terminal ADP-

The authors declare that they have no conflicts of interest with the contentsof this article. The content is solely the responsibility of the authors anddoes not necessarily represent the official views of the National Institutesof Health.

This article contains Figs. S1–S6 and Table S1.The atomic coordinates and structure factors (codes 6D36 and 6D3A) have been

deposited in the Protein Data Bank (http://wwpdb.org/).1 Both authors contributed equally to this work.2 Supported by the Intramural Research Program, NHLBI, National Institutes

of Health.3 Supported by the University of Cincinnati startup fund. To whom correspon-

dence should be addressed: Dept. of Chemistry, University of Cincinnati,301 Clifton Ct., Cincinnati, OH 45221. Tel.: 513-556-1909; Fax: 513-556-9239; E-mail: in-kwon.kim@uc.edu.

4 The abbreviations used are: PARylation, poly(ADP-ribosyl)ation; ARH,ADP-ribosyl-acceptor hydrolase; PAR, poly(ADP-ribose); PARP, PARpolymerase; PARG, PAR glycohydrolase; PTM, post-translational modi-fication; MAR, mono(ADP-ribose); ADPR, ADP-ribose; AIF, apopto-sis-inducing factor; MARylation, mono(ADP-ribosyl)ation; FL, full-

length; r.m.s.d., root mean square deviation; ADP-HPD, adenosinediphosphate (hydroxymethyl)pyrrolidine-2�,3�-diol.

croARTICLE

12350 J. Biol. Chem. (2018) 293(32) 12350 –12359

Published in the U.S.A.

by guest on September 25, 2020

http://ww

w.jbc.org/

Dow

nloaded from

ribose protein glycohydrolase (TARG1), MacroD1, and Mac-roD2 (16 –18). It has been shown that serine-specific ADP-ri-bosylation, a PTM specifically generated by PARP1/HPF1 andPARP2/HPF1 complexes, is the major cellular PTM followingDNA damage (19 –21). Importantly, in addition to PAR degra-dation, ARH3 was identified as the cellular eraser of serineMARylation marks (20, 22). Furthermore, ARH3 can specifi-cally digest O-acetyl-ADPR, a product of Sirtuin-catalyzedNAD�-dependent histone deacetylase reactions, which regu-late diverse biological processes, including chromatin remod-eling (11, 23, 24). Therefore, ARH3 is the only known enzymethat can specifically hydrolyze PAR, MAR PTMs, and O-acetyl-ADPR. For all these substrates, ARH3 preferentially hydrolyzesthe scissile �-O-linkage attached to the anomeric C1� position ofADPR (24) (Fig. 1a). However, the structural basis for the specific-ity of ARH3 for the 1�-O-linkage is unknown. Collectively, ARH3 isa distinctive, multitasking enzyme that controls two biologicallyimportant NAD�-dependent cellular signaling pathways.

The structure of the unliganded human ARH3, lacking theN-terminal 16 residues (ARH3�N16), shows a compact all-�-helical fold with a binuclear Mg2� center, which constitutesan archetype of the ARH superfamily (25). This di-Mg2�–containing catalytic center (MgA and MgB) is not found inPARG and is consistent with the Mg2�-dependent ADP-ribo-syl-acceptor hydrolase activity of ARH3 (11, 23). Consistently,mutations of Mg2�-coordinating residues of ARH3 led to adrastic decrease in PAR- and MAR-acceptor– hydrolyzingactivities (22, 25). However, we lack a fundamental mechanisticunderstanding of how ARH3 specifically recognizes and cleavesstructurally diverse substrates in a Mg2�-dependent manner.

To better understand the molecular mechanism of ARH3activities, we determined high-resolution crystal structures offull-length human ARH3 bound to Mg2� and ADPR, a productand an effective inhibitor of ARH3 activities (23, 24). Coupledwith computational analysis of ARH3–PAR substrate interac-tions, these structures reveal that substrate binding drives alarge-scale conformational transition from an unligandedclosed state to an open incision-competent enzyme state. Glu41

from the “glutamate flap,” which blocks the entry of substrate inthe closed state, is completely ejected from the binuclear Mg2�

center. The concomitant restructuring of the active site uncov-ers the catalytically important MgB, widens the leaving group–binding channel, and precisely positions the scissile 1�-O-link-age for hydrolysis while tightly securing the 2�- and 3�-OHgroups of ADPR. Unique structural features and conforma-tional flexibility of ARH3 strongly support ARH3’s specificityfor the 1�-O-linkage in structurally diverse substrates andMg2�-dependent catalysis.

Results

Structure of human ARH3FL–ADP-ribose–Mg2� complex

The crystal structure of apo-ARH3�N16 explained the lack ofADPR binding (25) (Fig. S1). We reasoned that the full-lengthARH3 (ARH3FL) might provide opportunities to capture ARH3in an ADPR-bound active form. The purified ARH3FL is likelymetal-free given its comparable basal activity without the addi-tion of Mg2� and in the presence of EDTA (Fig. 1b). The addi-

tion of Mg2� markedly enhances ARH3FL-mediated PAR hy-drolysis (Fig. 1b), which is consistent with a previous report(11). ARH3FL shows maximal activity with Mg2� followed byMn2� and Ca2� (Fig. 1b). Unlike DraG, the closest functionalhomolog that has a strong preference for Mn2� over Mg2�

(25–27), ARH3 exhibits rather comparable activity with eitherMg2� or Mn2�. We crystallized ARH3FL in complex withADPR and Mg2� and determined its crystal structure at 1.7-Åresolution (Fig. 1d). Thus, the numbering of amino acid resi-dues in this report deviates by 16 from that used for the apo-ARH3�N16 structure (25), but it is consistent with those used byOka et al. (11) and Abplanalp et al. (22). ARH3FL binds to oneADPR as evidenced by a clear and strong difference electrondensity at the active site (Fig. 1c).

ARH3FL adopts a compact all-�-helical fold with a centraldeep ADPR-binding cleft, a signature of the ARH3 superfamily(Figs. 1 and 2). The ARH3–ADPR–Mg2� complex structureidentifies two unique structural elements of ARH3, the “ade-nine cap” and glutamate flap, which undergo a structural rear-rangement upon ADPR binding and strongly contribute to thespecific substrate recognition of ARH3 (Figs. 1d and 2). Thebinuclear Mg2� catalytic center lies at the heart of a long,J-shaped substrate-binding channel. This structural featureenables ARH3 to bind and align both ADPR and the leavinggroup, e.g. two ADPR units (n ADPR and the n � 1 ADPRleaving group) in PAR substrates, for specific cleavage (Fig. 1,a and d).

The N-terminal 13 amino acids in the N-terminal extensionare disordered, which is consistent with their dispensable rolein ARH3 function (25). However, unexpectedly, Arg18 in theN-terminal extension, previously replaced by alanine in theapo-ARH3�N16 structure (25), contributes to ARH3 folding.Arg18 makes helix-capping interactions with the main-chaincarbonyls of Ser161 and Leu162 in �7 to stabilize �7 (Fig. S2).Arg18 further stabilizes ARH3 folding by forming a hydrogenbond with the side-chain carbonyl of Gln361 of �19 and a vander Waals contact with the side chain of Phe23 of �1. However,given the fully functional PAR hydrolysis activity of ARH3�N16

(25) and a remarkably long distance between Arg18 and Glu41

(� 32 Å), it is unlikely that Arg18 contributes to the observedconformational changes.

Structural comparison of the unliganded and ADPR-boundforms of ARH3 reveals dramatic conformational changes in theGlu41-containing flap motif, which we named the glutamateflap (Glu41-flap) (r.m.s.d. of 2.5 Å, comparing 52 C� atoms inthe Glu41-flap and its flanking residues of the unligandedARH3�N16 and complex D of the ARH3–ADPR–Mg2� com-plex) (Fig. 1d). The Glu41-flap is composed of the end of �1 thatexists as a 310-helix in the unliganded ARH3, �2, and a flexibleloop connecting �1 and �2 (L1) (Figs. 1d and 2). This substrate-induced conformational transition fully exposes the bimetalliccatalytic center for substrate engagement and to allow efficienthydrolysis to occur as described below. Together, these findingsimply that ARH3 can exist in at least two states: a substrate-bound “open” state and an unliganded “closed” state.

Four ARH3–ADPR–Mg2� complexes were found in theasymmetric unit (Fig. S3). The r.m.s.d. difference between thestructures of four ARH3–ADPR–Mg2� complexes (complexes

Structure of human ARH3 bound to ADP-ribose and Mg2�

J. Biol. Chem. (2018) 293(32) 12350 –12359 12351

by guest on September 25, 2020

http://ww

w.jbc.org/

Dow

nloaded from

A–D) and the unliganded ARH3 is summarized in Fig. S3b.Complexes B and D have nearly identical structures with alarge-scale conformational switch in the Glu41-flap (Fig. S3a).Complex D shows the largest degree of conformational transi-tion, which defines a fully open state, and therefore was primar-ily used for structural analyses. By contrast, complexes A and Cshow an intermediate conformational step between the unli-ganded ARH3 and complex D. A part of L1 in the Glu41-flap isdisordered in complexes A and C, further supporting its struc-tural flexibility (Fig. S3a).

ARH3 specifically exposes 1�-OH of ADP-ribose

ADPR is located at the deep ADPR-binding cleft of ARH3. Allthree parts of ADPR (adenosine, diphosphate, and the distalribose�) make extensive contacts with ARH3 (Figs. 1d and 3).ADPR has a surface area of 694 Å2 of which �80% (555 Å2 onaverage in complexes A–D) is buried by direct contacts withARH3. Overall, this matrix of ARH3–ADPR–Mg2� interac-

tions specifically exposes 1�-OH, corresponding to the scissileO-linkage in substrates (Figs. 1a and 3c), toward the catalyticMgB, strongly supporting its specificity for the 1�-O-linkage forsubstrate cleavage.

The distal ribose� of ADPR lies adjacent to the binuclearMg2� center where a group of acidic and polar residues exten-sively contact Mg2� ions and ADPR. The interactions betweenthe ribose� and two Mg2� ions are asymmetrical with moreextensive contacts on MgB. 3�-OH of the ribose� is directlycoordinated by MgB, and it is additionally hydrogen-bondedwith the side chain of Asn151 (Fig. 3, b and c). A water molecule(�-aqua ligand) that bridges MgA and MgB simultaneouslyengages 2�-OH of the ribose� with an unusually short distance(2.2 Å). These Mg2�-mediated ARH3 interactions with 2�- and3�-OH appear to secure tightly the ribose� to facilitate efficientcleavage of substrates at the 1�-O-linkage. In support of thishypothesis, in contrast to 2�- and 3�-OH groups, 1�-OH ofADPR that corresponds to the 1�-O-linkage in substrates is sol-

a

90o

MgAMgB

N

C

ADPR

Adenine Cap

Glu41-Flap

leaving group-binding site

MgA

MgB

Glu41

α1

α2

α3

α5

α8

α9

α6

α13

α10

α4

α12

α14

α16

α17α18

α19

d

Cleavageα(1"-2’) O-glycosidicbond

1"2"

3"

ribose’’

ribose’

1’2’

3’

n ADPR

(n-1) ADPR

12

34

56 78

9 O

HO OH

N

NN

N

NH2

O PO

O–O P

O

O–O O

HO OH

O

O OH

N

NN

N

NH2

O PO

O–O P

O

O–O O

HO OH

(ADPR)n

O

HO OH

N

NN

N

NH2

O PO

O–O P

O

O–O O

HO OH

OO O

HO OH

N

NN

N

NH2

O PO

O–O P

O

O–O O

HO OH

O

NH3+

O–

O

Poly(ADP-ribose) O-acetyl-ADP-ribose α-ADP-ribosyl-serine

1"

2"3"

1"

2"3"

Cleavage1"-O-linkage

Cleavage1"-O-linkage

cb

MgB

MgA

ADPR

PARylatedPARP1C

PARP1C

ARH3

EDTAPARG

Mg2+

+– + + –– – – – +– – – –+– – – + – N

o A

RH

3M

g2+

Mn

2+C

a2+

ARH3

PARylatedPARP1C

PARP1C75

M (kDa)

100150250

75

M (kDa)

100150250

Figure 1. Structure of full-length human ARH3 bound to ADP-ribose and Mg2�. a, structures of ARH3 substrates: a linear unbranched poly(ADP-ribose)(left), O-acetyl-ADP-ribose (middle), and �-ADP-ribosylserine (right). ARH3 cleaves the 1�-O-linkage in substrates. The exoglycohydrolase activity of ARH3cleaves the �(1�-2�) O-glycosidic bond between n and n � 1 ADP-ribose, releasing ADP-ribose as a product. b, left, Mg2� enhances the ADP-ribosyl-acceptorhydrolase activity of ARH3. The ARH3-mediated hydrolysis of PAR on PARylated PARP1C was monitored in the presence and absence of Mg2� using a gel-basedassay. PARG has stronger PAR turnover activity than ARH3 and was used as a positive control. Right, metal preference of ARH3. ARH3 prefers Mg2� for catalysisfollowed by Mn2� and Ca2�. c, difference electron density maps (Fo � Fc) for ADPR and Mg2� ions contoured at 3.0 � (blue, ADPR; orange, Mg2�). d, structureof the ARH3–ADPR–Mg2� complex revealing two unique and flexible structural elements, adenine cap (green) and glutamate flap (Glu41-flap) (red), thatundergo conformational changes and strongly contribute to specific substrate recognition. The 1�-OH of ADPR (blue circled), corresponding to the scissile1�-O-linkage in substrates, is exposed to solvent, consistent with ARH3 specificity for 1�-O-linkage for cleavage. A putative binding site for the leaving group ishighlighted with a black ellipse. The N and C termini of ARH3 are indicated by N and C, respectively.

Structure of human ARH3 bound to ADP-ribose and Mg2�

12352 J. Biol. Chem. (2018) 293(32) 12350 –12359

by guest on September 25, 2020

http://ww

w.jbc.org/

Dow

nloaded from

vent-accessible and exposed to the leaving group– binding site(Figs. 1d and 3c), consistent with cleavage at the C1� position(11, 20, 24). Another water molecule (W1) is axially liganded toMgB and makes a hydrogen bond with 1�-OH (Fig. 3c). This W1ligand is located proximal to the anomeric C1� and appears wellaligned for nucleophilic attack of C1� during catalysis. At the otherside of the binuclear metal center, the carboxyl group of Asp77 iscoordinated by MgA and further secures 2�-OH (Fig. 3c).

Despite the structural similarity between ARH3 superfamilymembers, the binding orientation of ADPR in ARH3 is remark-ably different from that in DraG (28) (r.m.s.d. of 2.6 Å, compar-ing all C� positions of complex D and the DraG–ADPR com-plex (Protein Data Bank code 2WOE)) (Figs. 3a and S4a). Thisdistinct mode of ADPR binding can be accounted for by aunique adenine-binding module of ARH3, the adenine cap,which consists of �6, �12, and a loop flanking the C terminus of�6 (L6). Notably, �6 and �12 are missing in DraG (Figs. 2 and3a). Two aromatic amino acids from the adenine cap, Phe143

and Tyr149, sandwich the adenine base through extensive �–�

stacking interactions (Fig. 3a). In line with this finding, theARH3-Y149A mutant showed a drastic decrease in PAR- andMAR-acceptor– hydrolyzing activities (22, 25). Leu235 from theflexible �12 in the adenine cap moved �2.7 Å toward the ade-nine ring relative to the unliganded ARH3, forming a van derWaals contact with N6 of the adenine base (Fig. 3, a and b). Thechemical selectivity of the adenine base is further enhanced byhydrogen bonds between the backbone carbonyl of Gly147 andN6 and between the backbone nitrogen of Tyr149 and N7 (Fig. 3,a and b). The diphosphate moiety interacts with the side chainsof Ser148 and His182 and the main-chain nitrogen of Gly119.Consistently, substitution of Ser148 and His182 with alanine ledto loss of ADP-ribosyl-acceptor hydrolase activities (22, 25).

A conformational switch of ARH3 enables specific substraterecognition

The most striking feature in the ARH3–ADPR–Mg2� com-plex is the dramatic conformational change of the Glu41-flap(Fig. 4). The end of �1, which is at the beginning of the Glu41-

....

..

.

α1 α2

α3 α4 α5

α6 α7 α8 α9

α10 α11 α12

α13 α14 η1 α15

α16 α17 η2 α18 α19

20 30 40 50 60

hARH3 G R G G GD G E A A A ALL V DL S A RSLS.... FR CL C SFY AHDTV T VLRHVQSLEPDPGTPGSERDraG G R G G GD G E L M L LAV L EI Q S TG.PSVHD AL AF A ATV FMTKG A QYGIHRKM.TG...GGWLR

70 80 90 100 110 120

hARH3 TDDT M AL SL AK D D FA K G L V M G A KTEALYY A AR VQ EAF E AHR QEY KDPDR.. YG VV..TVF DraG TDDT M AL SL AK D D FA K G A A I P V RLKPGQI E SL GR GTL V CEE LWL .....SR VD NTCRRGI

130 140 150 160 170 180

hARH3 E G GNG AMR LA A TH K L P AGI V FA LS L L NPKCRDVF ARAQFNGK SY G V S YSSVQD. QK R Q ASDraG E G GNG AMR LA A TH R M G LPA L WV AQ I Y H....... TTTAPYSE DA A C A TLGHPAD EP L R NH

190 200 210 220 230 240

hARH3 L VH G L R PY I A AL M E S GYNGA LQAL L Q ESSSEHFLKQLLGH ED EGDAQSVLDA LGMEER SSDraG L VH G L R PY L M LI A D P SDAAC TLGR H G RGMKACRE.....E NR VHQ......H ..FHFE KG

250 260 270 280 290 300

hARH3 T N A ESV AI F S RLKKIGELLDQASVTREEVVSELG GIA F P YC LRCMEPDPEIPSAFN LQRTDraG T S I DTM VL Y T Q......................S AY. V Q HY FVT...........D FKSC

310 320 330 340 350 360

hARH3 LI GGD DT AG AGA YG D P W E I Q SIS M AI M V Q E D V Y L T IAT Y Q ES Q SC GYE T LA SLHR FQK DraG LI GGD DT AG AGA YG D P W E I Q TVN L ML V I S D E L Q Q A TGA T D SG L KL MKV R RR VDAL ALA

Glu41 Flap

Adenine Cap

Adenine Cap

ADPR interacting residues MgB coordinating residuesMgA coordinating residues

L1 L2

L6

L2 wall (DraG)

Figure 2. Structure-based alignment of ARH3 and DraG with structural elements and residues that are important for function. The Glu41 residue of theGlu41-flap of ARH3 that undergoes a large conformational change upon ADPR binding is indicated by a red box. A part of L2 of DraG (L2 wall) completely blocksthe conformational change of �1 and restricts its activity to cleavage of mono(ADP-ribosyl)ated substrates. The end of �1 in ARH3, which exists as 310-helix inthe unliganded form and undergoes 310-to-� transition upon ADPR binding (Fig. 4a), is indicated by a blue bar. Two aromatic residues in DraG that stabilize theL2 wall are indicated by a red triangle. Asp97 in DraG that is essential for the cleavage of MARylated arginine is indicated by a blue triangle.

Structure of human ARH3 bound to ADP-ribose and Mg2�

J. Biol. Chem. (2018) 293(32) 12350 –12359 12353

by guest on September 25, 2020

http://ww

w.jbc.org/

Dow

nloaded from

flap, exists as a kinked 310-helix structure in the unligandedARH3 (Fig. 4a). ADPR binding drives a 310-to-� transition,leading to �27° of rotation from the helical axis of the 310-helix.This rotation induces a concomitant �8.5-Å displacement ofL1 away from the active site, which accompanies �24° of rota-tion in �2 (Fig. 4a). Consequently, ADPR binding results in an�4.5-Å movement of the carboxylate of Glu41 of the Glu41-flapaway from the catalytic MgB. The straightened conformation of�1 is stabilized by new intrahelical stacking interactions within�1 among residues Phe39, Tyr40, and His43 (Fig. 4a).

Mechanistically, the observed conformational switch of theGlu41-flap is required for specific substrate recognition. In theunliganded state, Glu41 of the Glu41-flap completely masks MgB

from access to substrate (Fig. 4c). This closed state is thereforeenzymatically inactive, and Glu41 must be ejected away fromMgB to allow substrate to enter the bimetallic catalytic center.In support of this hypothesis, the ARH3–ADPR–Mg2� com-plex shows a conformational switch in the Glu41-flap from a

closed state to an open state. This substantial active-siterestructuring unmasks MgB and significantly widens the leav-ing group– binding site (Figs. 1d and 4, b and c), which nowallows substrate entrance to the dimetallic catalytic center. Thisopen conformation of ARH3 provides an optimal alignmentbetween the scissile 1�-O-linkage and catalytic groups (Mg2�

and catalytic residues) and therefore constitutes an “incision-competent” enzyme state (Figs. 3d and 4b). Collectively, theGlu41-flap controls substrate access to the binuclear Mg2� cen-ter, and its closed-to-open conformational switch enables spe-cific binding and cleavage of substrates.

To gain further mechanistic insights into substrate hydroly-sis, we modeled di-ADPR, a substrate with the largest leavinggroup (Fig. 1a), to the active site of ARH3 in catalytic position(Fig. 4, b and c). In this ARH3– di-ADPR–Mg2� model, then � 1 ADPR leaving group snugly fits into the curved J-shapedsubstrate binding channel (Fig. 4b), whereas n ADPR occupiesthe identical site as that seen in the ARH3–ADPR–Mg2� com-

OOH

OH

OPOPON

N

NN

H2NO

HO OH

–OO–O

OOH

a

c

Adenine Cap(ARH3)

MgA MgB

ADPR(ARH3)

ADPR(DraG)

F143 Y149

L235

α6

α12

ADPR

1"-OH

2"-OH 3"-OH

D314

E41

D77

T76

D78

D316

T317

μ-aqualigand

water (W1)

MgAMgB

n ADPR

(n-1) ADPR

D314

E41

MgA

MgB

1"-scissileO-glycosidic

bond

d

N151

ARH3:ADPR:Mg2+ DraG

H-bondMg2+ Coordination

L235

N151Y149

F143 S148 H182

D314

D771"2"

3"

G147

12 3

4

56

789

b

Figure 3. ARH3 specifically exposes the scissile 1�-O-linkage in substrates for cleavage. a, structural superposition of ADPR-bound forms of ARH3 (wheat)and DraG (gray) reveals a distinctive ADPR-binding mode in ARH3. The adenine cap of ARH3 (green) grasps the adenine ring and is essential for ARH3 activities.Hydrogen bonds contributed by the main-chain atoms of the adenine cap to N6 and N7 of the adenine ring impart specificity. b, diagram showing interactionsbetween ADPR and ARH3. c, close-up of the binuclear Mg2� catalytic center and ADPR binding in ARH3. A matrix of Mg2�-mediated coordination andhydrogen-bonding interactions secure ADPR in the active site, which is consistent with Mg2�-dependent catalysis by ARH3. The 1�-OH of ADPR (blue circled),corresponding to the scissile 1�-O-linkage, is specifically exposed to solvent, strongly supporting ARH3 specificity for the 1�-O-linkage as a site for cleavage. d,close-up of the ARH3– di-ADPR–Mg2� complex model. The energy-minimized computational model for the ARH3– di-ADPR–Mg2� complex was generatedusing the ARH3–ADPR–Mg2� complex as a starting model (see “Experimental procedures”). In this model, 3�-OH of the n � 1 ADPR leaving group is directlycoordinated by MgB, replacing the W1 water ligand, and MgB tightly secures both n and n � 1 ADPR units for efficient cleavage to occur.

Structure of human ARH3 bound to ADP-ribose and Mg2�

12354 J. Biol. Chem. (2018) 293(32) 12350 –12359

by guest on September 25, 2020

http://ww

w.jbc.org/

Dow

nloaded from

plex. Notably, 3�-OH of the adenosine ribose� of n � 1 ADPR isdirectly coordinated to MgB, replacing the axially coordinatedwater ligand (W1), and it is positioned within hydrogen-bond-ing distance to Asp314 and Glu41 (Figs. 3d and 4b). These find-ings suggest that MgB is important for catalysis by securingADPR and the leaving group to promote the subsequent cleav-age at the C1� position.

Structural superposition of ADPR-bound forms of ARH3and DraG reveals that the observed 310-to-� transition of �1and the conformational flexibility of the Glu41-flap are uniquein ARH3. Analysis of secondary structures using Dictionary ofSecondary Structure of Proteins (DSSP) (29) indicates thatDraG lacks the flexible 310-helix structure at the end of �1 (Fig.2). Furthermore, a part of L2 of DraG (named the “L2 wall”)tightly caps �1 through an edge-stacking �–� interactionbetween the side chains of Trp50 from L2 and Phe29 from �1(Figs. 3a and S4). This structural arrangement of DraG effec-tively restricts the conformations of �1 and L1 (correspondingto the Glu41-flap in ARH3) to the closed state in which substrateaccess to the leaving group– binding site is prevented (Fig. 4a).The DraG mechanism instead involves a ring opening of theribose�, which positions the scissile 1�-N-linkage in close prox-imity to Asp97 of DraG for cleavage (28). This structural featurewould enforce DraG substrate specificity to MARylated argi-nine (28). In ARH3, Phe29 and Trp50 of DraG are replaced withalanine and serine, respectively, relieving the constraininginteraction. Consequently, the L2 of ARH3, equivalent to the L2wall of DraG, is disordered. Furthermore, Asp97 of DraG isreplaced with glycine in ARH3 (Fig. 2). Taken together, theconformational flexibility of the Glu41-flap, the lack of the L2wall, metal preference for Mg2�, and the lack of the catalyticresidue equivalent to DraG Asp97 in ARH3 (25) (Fig. 2), allsuggest a distinct catalytic mechanism for ARH3 and explaindifferent substrate specificity in ARH3.

Binuclear metal center of the ARH3–ADPR–Mg2� complex

In the ARH3–ADPR–Mg2� complex, two Mg2� ions (MgA

and MgB) have an octahedral coordination geometry and are

fully occupied with only 3.3-Å intermetal distance (3.8 Å in theunliganded ARH3 structure). They are bridged by the bidentatecarboxyl group of Asp316 as well as by the �-aqua ligand (Fig.3c). The bridging �-aqua ligand is nearly symmetrically coordi-nated by two Mg2� ions (1.9 and 1.8 Å to MgA and MgB, respec-tively) and fixes 2�-OH of the ribose� of ADPR (Fig. 3c).

In contrast to MgA, whose coordination ligands remain iden-tical (including Thr76, Asp77, Asp78, and Asp316), the MgB coor-dination setup is dynamically changed. In the unligandedARH3, Glu41 of the Glu41-flap is coordinated by MgB. UponPAR substrate binding, 3�-OH of the adenosine ribose� of then � 1 ADPR leaving group replaces Glu41 and directly coordi-nates MgB. Thus, in the substrate-bound state, MgB engagesboth ADPR and the leaving group to precisely expose the scis-sile 1�-O-linkage for cleavage (Figs. 3d and 4b). Asp314 andGlu41 further aid in the correct positioning of the n � 1 ADPRleaving group by interacting with 3�-OH of the adenosineribose�. Finally, in the ARH3–ADPR (product) complex, theleaving group departs, and a new water ligand (W1) is axiallycoordinated by MgB (Figs. 3c and 4b). This dynamic switchingof the metal-site makeup in each catalytic step of ARH3 is con-sistent with Mg2�-dependent ARH3 activity.

Asp314 is essential for the formation of binuclear metal center

The structure of the ARH3–ADPR–Mg2� complex showsthat the dimetallic catalytic center of ARH3 is tightly packedagainst ADPR and appears ideally optimized for cleavage of thescissile 1�-O-linkage in substrates. Therefore, a subtle change inthe active-site arrangement may result in a detrimental effecton ARH3 functions. Consistently, a conservative substitution ofAsp314 with glutamate led to the loss of enzymatic activity (22,25) (Fig. S5a). To gain further insights into this structure–function relationship, we determined the crystal structure ofARH3D314E bound to ADPR and Mg2� at 1.6-Å resolution(Figs. 5 and S5b). The overall structure and ADPR-bindingmode are nearly identical to those for the ARH3WT–ADPR–Mg2� complex (r.m.s.d. of 0.2 Å, comparing all C� positions ofcomplexes D of ARH3WT and ARH3D314E). Strikingly, however,

MgA

MgB

D314

D77

ADPR

α1α2

L1

F39 Y40

H43

~ 27°

~ 24°~ 8.5Å

E41

ba

Glu41-Flap

unliganded ARH3

Glu41-Flap

ARH3:ADPR:Mg2+

(n-1) ADP-ribose

n ADP-ribose E41

ARH3:ADPR:Mg2+

unliganded ARH3

c

MgB

Figure 4. A conformational change of ARH3 enables specific substrate recognition and cleavage. a, structural comparison of the unliganded (gray/black)and ADPR-bound forms (orange/red) of ARH3 reveals a conformational switch in the Glu41-flap. The end of �1 undergoes a 310-to-� transition upon ADPRbinding, which results in displacement of Glu41 away from MgB and significantly widens the leaving group– binding site of ARH3. b and c, ARH3 switchesconformations to enable specific poly(ADP-ribose) recognition. The structure of the ARH3– di-ADPR–Mg2� complex model is superimposed onto those of theARH3–ADPR–Mg2� complex and the apo-ARH3�N16, and di-ADPR is shown on the surfaces of ARH3. Comparison of solvent-accessible surfaces reveals an openplatform that can accommodate n � 1 ADPR in the ARH3–ADPR–Mg2� complex (b), whereas n � 1 ADPR binding is completely obstructed by the Glu41-flapin the unliganded ARH3 (c).

Structure of human ARH3 bound to ADP-ribose and Mg2�

J. Biol. Chem. (2018) 293(32) 12350 –12359 12355

by guest on September 25, 2020

http://ww

w.jbc.org/

Dow

nloaded from

MgB is completely missing in the ARH3D314E–ADPR–Mg2�

complex, and the side chain of Glu314 is flipped out from thebinuclear metal center, presumably due to a steric clash causedby the one-carbon–longer side chain in Glu314 (Fig. 5). Thedistance between Glu314 and the corresponding MgB in theARH3WT–ADPR–Mg2� complex is beyond the range of directcoordination (3.7 Å), explaining the loss of MgB and the loss ofenzymatic activity. Consistent with the important role of MgB

in securing ADPR (Fig. 3c), the overall B factors for atoms in theribose� of ADPR in the ARH3D314E–ADPR–Mg2� complex aresignificantly higher than those in the ARH3WT–ADPR–Mg2�

complex. Collectively, these findings indicate that Asp314 isrequired for the formation of the binuclear Mg2� center andsuggest a critical role of MgB for catalysis.

Discussion

Our structures of the ARH3–ADPR–Mg2� complex and acomputational model of the ARH3– di-ADPR–Mg2� complexreveal the previously unknown conformational plasticity of theGlu41-flap, which strongly supports ARH3 specificity for the1�-O-linkage for cleavage. Our studies also explain the pub-lished mutational analysis data for ARH3. These structuralfindings are consistent with Mg2�-dependent hydrolysis andprovide important clues to the catalytic mechanism of ARH3.First, the observed conformational switch in ARH3 is requiredto specifically recognize substrates. In the closed unligandedstate, Glu41 coordinates MgB and prevents substrates fromentering into the dimetallic catalytic center (Fig. 6). To engagesubstrates, the Glu41-flap is completely moved away from MgB,inducing substantial active-site rearrangement from a closed toan open state. This conformational transition now allows sub-strates to enter the active site of ARH3. Second, our structuresstrongly support the observation that ARH3 favors the 1�-O-linkage in substrates for cleavage (24). In the ARH3–ADPR–Mg2� complex, 2�- and 3�-OH of ADPR are secured by Mg2�

ions, the bridging �-aqua ligand, and active-site residues,whereas 1�-OH corresponding to the scissile 1�-O-linkage insubstrates is specifically exposed to the open leaving group–

binding site (Fig. 3c). Substrates containing 2�- or 3�-O-linkagetherefore would result in a serious steric clash with the tightlypacked binuclear metal center of ARH3, leading to inefficientsubstrate cleavage. Consistently, IC50 values for ARH3 inhibi-tion by 2�- or 3�-N-acetyl-ADPR, analogs of O-acetyl-ADPR,were significantly higher than that for ADPR (24). In ourARH3– di-ADPR–Mg2� model, the di-ADPR is slightly bentdue to the simultaneous engagement of 3�-OH of n ADPR and3�-OH of n � 1 ADPR by MgB (Figs. 3d and 6). This structuralfeature would likely further reinforce the correct positioning ofthe scissile 1�-O-linkage for efficient cleavage (Fig. 6). Third, thesubstrate-induced widening of the leaving group– binding siteenables ARH3 to specifically bind and cleave substrates with astructurally diverse leaving group (Figs. 1a and 4), supportingARH3’s broad substrate specificity.

The flexible Glu41-flap dynamically switches between closedand open conformations and plays important roles in substraterecognition (Fig. 6). In the unliganded ARH3, Glu41 of theGlu41-flap serves as the key residue that constitutes the closedenzyme state by masking MgB. In the PAR substrate– boundstate, Glu41 is released from MgB and instead interacts bothwith ADPR and the leaving group (Fig. 6), which presumablycontributes to the precise alignment of the scissile O-linkage forsubsequent cleavage and constitutes the open enzyme state.Therefore, we propose the Glu41-flap as the gate that controls sub-strate entrance to the active site. Given the structural flexibility andsolvent accessibility of the Glu41-flap, it is also possible that theGlu41-flap functions as a protein–protein interaction module. Pro-teins that specifically interact with the closed conformation of theGlu41-flap of ARH3 might conformationally lock ARH3 in an inac-tive state and thereby regulate ARH3 function.

A notable ARH3 inhibition by ADP-HPD (Fig. S6a) (30), ananalog of the oxocarbenium ion intermediate, raises the possi-bility that the ARH3 mechanism might involve an oxocarbe-nium intermediate in the distal ribose� of ADPR in a similar wayto PARG (15, 31), which is followed by water-mediated nucleo-philic attack at the anomeric C1� position. The observed 18Oincorporation to C1� during hydrolysis of O-acetyl-ADPR (24)is also consistent with this model. In glycosidase superfamilymembers, a catalytic acidic residue is typically found near thescissile bond, e.g. Glu756 in rat PARG (Glu752 in human PARG)(31, 32), to activate a water molecule for nucleophilic attack onthe anomeric carbon (33). In the ARH3–ADPR–Mg2� com-plex, Asp314 is located proximal to both 1�-OH (correspondingto the scissile 1�-O-linkage) and the axial water ligand (W1) ofMgB (Fig. 3c). It is plausible that Asp314 might function as ageneral acid or base to protonate the leaving group and thenactivate W1 for backside attack of the anomeric C1� in a man-ner similar to Glu756 in rat PARG (Fig. S6b) (31). In support ofthis catalytic role of Asp314, the ARH3-D314N mutant shows adramatic reduction in ARH3 activities (22, 25). Although morework is needed to determine the catalytic mechanism of ARH3,it is unlikely that ARH3 has a redox chemistry step for catalysisgiven its preference for Mg2� over Mn2� (Fig. 1b).

ARH3 is a unique multitasking enzyme that regulates cellularconcentrations of both PAR, either free or attached to proteins,and mono(ADP-ribosyl)ated substrates. Elevated PARylationlevels are often found in many human cancers, including

ADPR

2"-OHD314

E41

D77

T76 D78

D316

H-bondMg2+ Coordination

MgAMgB

ARH3-WTARH3-D314E

E3141"-OH

Figure 5. Asp314 is essential for the formation of the binuclear metalcenter. Structural comparison of the WT and catalytically inactive ARH3-D314E mutant reveals a side-chain flipping of Glu314, which leads to the lossof MgB. This suggests that Asp314 is essential for the formation of the binu-clear metal center in ARH3 and that MgB is required for catalysis.

Structure of human ARH3 bound to ADP-ribose and Mg2�

12356 J. Biol. Chem. (2018) 293(32) 12350 –12359

by guest on September 25, 2020

http://ww

w.jbc.org/

Dow

nloaded from

BRCA-deficient breast cancers (34, 35) and triple-negativebreast cancers (36, 37). ARH3�/� cells undergo enhanced PAR-dependent cell death upon genotoxic stresses but remainhealthy under unstressed conditions. We suggest that pharma-cological intervention in ARH3-dependent pathways could be asafe and efficient therapeutic strategy for cancers with up-reg-ulated PARylation by increasing the cytoplasmic PAR level andtriggering parthanatos, either alone or in combination withcurrent chemotherapeutic agents. Our structures reveal exten-sive adenine interactions with the adenine cap and a large con-formational switch of the Glu41-flap that are crucial for specificengagement of structurally diverse substrates. These uniquestructural elements of ARH3 can be exploited for the develop-ment of specific ARH3 inhibitors, which have potential thera-peutic applications, as well as the ability to advance our under-standing of the role of ARH3 in regulation of PARylation. Aconcern is that ARH3 also catalyzes hydrolysis of O-acetyl-ADPR, which may be involved in other cellular pathways suchas chromatin remodeling. This activity also relies on the ability ofARH3 to act at the C1� position of ADPR. ARH3’s contribution tothe O-acetyl-ADPR–dependent pathways, in contrast to otherproteins in the ARH-macrodomain superfamily, is not known.Thus, some ARH3 inhibitors may affect multiple signaling path-ways due to its multiple enzymatic activities and substrates.

Experimental procedures

Plasmids and protein purification

Human ARH3FL was cloned into a modified pET21b vectorwith an N-terminal His6 tag and a following cleavage site forPreScission protease (pET21b-His6-pps). The pET21b-His6-pps-ARH3FL plasmid was introduced into Escherichia coliRosetta 2 (DE3) cells, and ARH3FL was induced by adding 1 mM

isopropyl �-D-thiogalactoside overnight at 16 °C. ARH3FL waspurified by affinity capture on a nickel-nitrilotriacetic acid (GEHealthcare) column. After elution with imidazole (250 mM), theprotein was loaded onto a heparin column (GE Healthcare) and

eluted with a NaCl gradient (0.1–1 M). Fractions with ARH3FL

were pooled, and PreScission protease was added to cleave theN-terminal histidine tag. Finally, ARH3FL was loaded to a Sep-hacryl 200 (GE Healthcare) size-exclusion column using a buffercontaining 25 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1 mM DTT, and5% glycerol. The purified ARH3FL was concentrated to �30mg/ml and then stored at �80 °C. A gene encoding the ARH3-D314E mutant was synthesized (GeneUniversal, Inc.) and clonedinto pET-21b-pps vector. The ARH3-D314E mutant was purifiedusing the same protocol as the WT protein. The human PARGcatalytic domain (residues 448–976) was cloned in pET21b-His6-pps, expressed in E. coli Rosetta (DE3) cells, and purified by nickel-nitrilotriacetic acid–affinity chromatography, heparin chroma-tography, and Sephacryl 200 size-exclusion chromatographyas described previously (32). For preparation of PARylatedPARP1, the DNA-binding domain (residues 1–374) ofhuman PARP1 and the PARP1C catalytic domain (residues375–1014) were purified as described previously (6, 31).

Crystallization and data collection

The WT and the D314E mutant of ARH3 (10 mg/ml) werecocrystallized with 5 mM ADPR by hanging-drop vapor diffu-sion. ADPR binding is required for crystallization as the unli-ganded ARH3FL did not yield any crystals. The protein solutionwas mixed with an equal volume of well solution (22% PEG4000, 0.1 M sodium acetate, pH 4.5, and 0.1 M MgSO4) andincubated at 22 °C. Crystals were briefly equilibrated in a har-vesting solution (26% PEG 4000, 0.1 M sodium acetate, pH 4.5,0.1 M MgSO4, and 5 mM ADPR), transferred to a cryoprotectantsolution (26% PEG 4000, 0.1 M sodium acetate, pH 4.5, 0.1 M

MgSO4, 5 mM ADPR, and 15% glycerol), and then flash-cooledin liquid nitrogen for data collection.

X-ray diffraction data were collected at the NE-CAT 24ID-Ebeamline at the Advanced Photon Source. The ARH3WT–ADPR–Mg2� complex crystals (P1 symmetry; four ARH3–ADPR–Mg2� complexes per asymmetric unit) diffracted to

n ADP-ribose-binding site

leaving group-binding site

D314

E41

D77

Glu41-Flap

ScissileO-glycosidic bond

(1"-O-linkage)

Unliganded “closed” conformation(n-1) ADPR binding is obstructed

by the Glu41-Flap

Substrate-bound, incision-competent “open” conformation Product-bound “open” conformation

MgA

MgB

(n-1) ADP-riboseleaving group

D314

E41

D77 D314

E41

D77

Figure 6. ARH3 switches between closed and open conformations to specifically recognize and cleave substrates. ARH3 switches between a self-inhibited and an incision-competent conformation for specific substrate recognition and cleavage. In the absence of PAR substrates, the Glu41-flap caps MgB,constituting a closed, self-inhibited enzyme state. Substrate binding induces a conformational switch in the Glu41-flap, allowing substrate entrance to theactive site. This active-site rearrangement in ARH3 specifically exposes the scissile 1�-O-linkage to the catalytic Asp314 for cleavage, explaining ARH3 specificityfor the 1�-O-linkage in structurally diverse substrates.

Structure of human ARH3 bound to ADP-ribose and Mg2�

J. Biol. Chem. (2018) 293(32) 12350 –12359 12357

by guest on September 25, 2020

http://ww

w.jbc.org/

Dow

nloaded from

1.7-Å resolution, and ARH3D314E–ADPR–Mg2� complex crys-tals (P1 symmetry; four ARH3–ADPR–Mg2� complexes perasymmetric unit) diffracted to 1.6-Å resolution. X-ray data setswere collected with an Eiger 16M detector and processed usingHKL2000 (38) and SCALEPACK (38, 39). Data collection sta-tistics are shown in Table S1.

Structure determination

The full-length human ARH3WT–ADPR complex structurewas determined by molecular replacement using MolRep (40)in the CCP4 suite with the apo-ARH3�N16 structure (25) as asearch model. The asymmetric unit contains four ARH3 mole-cules, and they all show a strong difference electron density forthe bound ADPR at the active site, which was supported bypolder map calculations. The restraints for ADPR were gener-ated using Monomer Library Sketcher in the CCP4i suite (41).The model was manually rebuilt using Coot (42) and refinedwith PHENIX (43) to an Rfactor of 18.4% and an Rfree of 21.8%.The ARH3D314E–ADPR complex structure was determined bymolecular replacement using MolRep (40) with the ARH3WT–ADPR–Mg2� complex structure as a search model. TheARH3D314E–ADPR–Mg2� complex model was rebuilt usingCoot (42) and refined with PHENIX (43) to an Rfactor of 16.7%and an Rfree of 19.9%. Crystallographic data statistics are shownin Table S1. The Ramachandran plot shows that �98% of theresidues in both ARH3WT–ADPR–Mg2� and ARH3D314E–ADPR–Mg2� complexes are in the favored regions, and allthe others are in the allowed regions. No outlier residue wasobserved.

ARH3 activity assay

ARH3 activity was measured against PARylated PARP1using a method similar to that for PARG activity measurementas described previously (31). The C-terminal catalytic domainof PARP1 (PARP1C; residues 375–1014) containing the auto-modification domain was PARylated in the presence of theN-terminal DNA-binding domain of PARP1, a nicked DNA,and NAD� as described previously (6, 31). Briefly, PARylationof PARP1 was performed at 37 °C in a reaction buffer contain-ing 50 mM Tris-HCl, pH 7.5, 100 mM NaCl, and 2 mM DTT. ToPARylate PARP1, PARP1C (2 �M), DNA-binding domain (res-idues 1–374) (2 �M), and a nicked DNA (2 �M) were preincu-bated for 10 min on ice. 200 �M NAD� was then added to thereaction, and it was then incubated at 37 °C for 30 min. Afterdesalting with a buffer containing 50 mM Tris-HCl, pH 7.5, 100mM NaCl, and 5% glycerol, purified human ARH3 proteins(WT and D314E) or human PARG was treated with PARylatedPARP1 substrates in the presence or absence of 5 mM EDTA ordivalent metals (Mg2�, Mn2�, and Ca2�) and incubated for 60min at 37 °C. The level of modification of PARP1 was visualizedby Coomassie Blue staining of SDS-polyacrylamide gels.

Computational modeling of the ARH3– di-ADPR–Mg2�

complex

The di-ADP-ribose model was generated by covalently link-ing 2�-OH of the adenosine ribose� of n � 1 ADPR to the ano-meric C1� of the distal ribose� of n ADPR in the ARH3–ADPR–Mg2� complex using YASARA (44). The n ADPR molecule, two

Mg2� ions, and the bridging �-aqua ligand remained anchoredto the experimentally identified site in the ARH3–ADPR–Mg2� complex. Then n � 1 ADPR was docked to the putativen � 1 ADPR-binding site (leaving group– binding site) of theJ-shaped substrate-binding channel of ARH3. The ARH3– di-ADPR–Mg2� model was then subjected to energy minimiza-tion using the AMBER ff14SB protein force field in the Chimerapackage (45). All atoms in the protein and di-ADPR ligand wereallowed to move during energy minimization. All structuralfigures were prepared using PyMOL.

Author contributions—Y. P. validation; Y. P., J. V., I. K., and A. C.investigation; Y. P., J. V., and J. M. writing-original draft; J. V., A. C.,and J. M. formal analysis; I. K., J. M., and I.-K. K. writing-review andediting; I.-K. K. supervision.

Acknowledgments—This work used NE-CAT beamlines (NationalInstitutes of Health Grant P41 GM103403) and an Eiger 16M detec-tor on 24-ID-E beam line (National Institutes of Health GrantS10OD021527) at the APS (United States Department of EnergyGrant DE-AC02-06CH11357).

References1. Hakmé, A., Wong, H. K., Dantzer, F., and Schreiber, V. (2008) The ex-

panding field of poly(ADP-ribosyl)ation reactions. ’Protein modifications:beyond the usual suspects’ review series. EMBO Rep. 9, 1094 –1100CrossRef Medline

2. Gibson, B. A., and Kraus, W. L. (2012) New insights into the molecular andcellular functions of poly(ADP-ribose) and PARPs. Nat. Rev. Mol. CellBiol. 13, 411– 424 CrossRef Medline

3. Mégnin-Chanet, F., Bollet, M. A., and Hall, J. (2010) Targeting poly(ADP-ribose) polymerase activity for cancer therapy. Cell. Mol. Life Sci. 67,3649 –3662 CrossRef Medline

4. El-Khamisy, S. F., Masutani, M., Suzuki, H., and Caldecott, K. W. (2003) Arequirement for PARP-1 for the assembly or stability of XRCC1 nuclearfoci at sites of oxidative DNA damage. Nucleic Acids Res. 31, 5526 –5533CrossRef Medline

5. Beck, C., Robert, I., Reina-San-Martin, B., Schreiber, V., and Dantzer, F.(2014) Poly(ADP-ribose) polymerases in double-strand break repair: fo-cus on PARP1, PARP2 and PARP3. Exp. Cell Res. 329, 18 –25 CrossRefMedline

6. Kim, I. K., Stegeman, R. A., Brosey, C. A., and Ellenberger, T. (2015) Aquantitative assay reveals ligand specificity of the DNA scaffold repairprotein XRCC1 and efficient disassembly of complexes of XRCC1 and thepoly(ADP-ribose) polymerase 1 by poly(ADP-ribose) glycohydrolase.J. Biol. Chem. 290, 3775–3783 CrossRef Medline

7. Andrabi, S. A., Kim, N. S., Yu, S. W., Wang, H., Koh, D. W., Sasaki, M.,Klaus, J. A., Otsuka, T., Zhang, Z., Koehler, R. C., Hurn, P. D., Poirier, G. G.,Dawson, V. L., and Dawson, T. M. (2006) Poly(ADP-ribose) (PAR) poly-mer is a death signal. Proc. Natl. Acad. Sci. U.S.A. 103, 18308 –18313CrossRef Medline

8. Yu, S. W., Andrabi, S. A., Wang, H., Kim, N. S., Poirier, G. G., Dawson,T. M., and Dawson, V. L. (2006) Apoptosis-inducing factor mediates poly-(ADP-ribose) (PAR) polymer-induced cell death. Proc. Natl. Acad. Sci.U.S.A. 103, 18314 –18319 CrossRef Medline

9. Mashimo, M., Kato, J., and Moss, J. (2013) ADP-ribosyl-acceptor hydro-lase 3 regulates poly(ADP-ribose) degradation and cell death during oxi-dative stress. Proc. Natl. Acad. Sci. U.S.A. 110, 18964 –18969 CrossRefMedline

10. Miwa, M., Tanaka, M., Matsushima, T., and Sugimura, T. (1974) Purifica-tion and properties of glycohydrolase from calf thymus splitting ribose-ribose linkages of poly(adenosine diphosphate ribose). J. Biol. Chem. 249,3475–3482 Medline

Structure of human ARH3 bound to ADP-ribose and Mg2�

12358 J. Biol. Chem. (2018) 293(32) 12350 –12359

by guest on September 25, 2020

http://ww

w.jbc.org/

Dow

nloaded from

11. Oka, S., Kato, J., and Moss, J. (2006) Identification and characterization ofa mammalian 39-kDa poly(ADP-ribose) glycohydrolase. J. Biol. Chem.281, 705–713 CrossRef Medline

12. Davidovic, L., Vodenicharov, M., Affar, E. B., and Poirier, G. G. (2001)Importance of poly(ADP-ribose) glycohydrolase in the control of poly-(ADP-ribose) metabolism. Exp. Cell Res. 268, 7–13 CrossRef Medline

13. Braun, S. A., Panzeter, P. L., Collinge, M. A., and Althaus, F. R. (1994)Endoglycosidic cleavage of branched polymers by poly(ADP-ribose) gly-cohydrolase. Eur. J. Biochem. 220, 369 –375 CrossRef Medline

14. Hatakeyama, K., Nemoto, Y., Ueda, K., and Hayaishi, O. (1986) Purifica-tion and characterization of poly(ADP-ribose) glycohydrolase. Differentmodes of action on large and small poly(ADP-ribose). J. Biol. Chem. 261,14902–14911 Medline

15. Slade, D., Dunstan, M. S., Barkauskaite, E., Weston, R., Lafite, P., Dixon, N.,Ahel, M., Leys, D., and Ahel, I. (2011) The structure and catalytic mecha-nism of a poly(ADP-ribose) glycohydrolase. Nature 477, 616 – 620CrossRef Medline

16. Jankevicius, G., Hassler, M., Golia, B., Rybin, V., Zacharias, M., Timinszky,G., and Ladurner, A. G. (2013) A family of macrodomain proteins reversescellular mono-ADP-ribosylation. Nat. Struct. Mol. Biol. 20, 508 –514CrossRef Medline

17. Sharifi, R., Morra, R., Appel, C. D., Tallis, M., Chioza, B., Jankevicius, G.,Simpson, M. A., Matic, I., Ozkan, E., Golia, B., Schellenberg, M. J., Weston,R., Williams, J. G., Rossi, M. N., Galehdari, H., et al. (2013) Deficiency ofterminal ADP-ribose protein glycohydrolase TARG1/C6orf130 in neuro-degenerative disease. EMBO J. 32, 1225–1237 CrossRef Medline

18. Rosenthal, F., Feijs, K. L., Frugier, E., Bonalli, M., Forst, A. H., Imhof, R.,Winkler, H. C., Fischer, D., Caflisch, A., Hassa, P. O., Lüscher, B., andHottiger, M. O. (2013) Macrodomain-containing proteins are new mono-ADP-ribosylhydrolases. Nat. Struct. Mol. Biol. 20, 502–507 CrossRefMedline

19. Bonfiglio, J. J., Fontana, P., Zhang, Q., Colby, T., Gibbs-Seymour, I., Ata-nassov, I., Bartlett, E., Zaja, R., Ahel, I., and Matic, I. (2017) Serine ADP-ribosylation depends on HPF1. Mol. Cell 65, 932–940.e6 CrossRefMedline

20. Fontana, P., Bonfiglio, J. J., Palazzo, L., Bartlett, E., Matic, I., and Ahel, I.(2017) Serine ADP-ribosylation reversal by the hydrolase ARH3. Elife 6,e28533 CrossRef Medline

21. Palazzo, L., Leidecker, O., Prokhorova, E., Dauben, H., Matic, I., and Ahel,I. (2018) Serine is the major residue for ADP-ribosylation upon DNAdamage. Elife 7, e34334 CrossRef Medline

22. Abplanalp, J., Leutert, M., Frugier, E., Nowak, K., Feurer, R., Kato, J., Kis-temaker, H. V. A., Filippov, D. V., Moss, J., Caflisch, A., and Hottiger, M. O.(2017) Proteomic analyses identify ARH3 as a serine mono-ADP-ribosyl-hydrolase. Nat. Commun. 8, 2055 CrossRef Medline

23. Ono, T., Kasamatsu, A., Oka, S., and Moss, J. (2006) The 39-kDa poly-(ADP-ribose) glycohydrolase ARH3 hydrolyzes O-acetyl-ADP-ribose, aproduct of the Sir2 family of acetyl-histone deacetylases. Proc. Natl. Acad.Sci. U.S.A. 103, 16687–16691 CrossRef Medline

24. Kasamatsu, A., Nakao, M., Smith, B. C., Comstock, L. R., Ono, T., Kato, J.,Denu, J. M., and Moss, J. (2011) Hydrolysis of O-acetyl-ADP-ribose iso-mers by ADP-ribosylhydrolase 3. J. Biol. Chem. 286, 21110 –21117CrossRef Medline

25. Mueller-Dieckmann, C., Kernstock, S., Lisurek, M., von Kries, J. P., Haag,F., Weiss, M. S., and Koch-Nolte, F. (2006) The structure of human ADP-ribosylhydrolase 3 (ARH3) provides insights into the reversibility of pro-tein ADP-ribosylation. Proc. Natl. Acad. Sci. U.S.A. 103, 15026 –15031CrossRef Medline

26. Nordlund, S., and Noren, A. (1984) Dependence on divalent cations of theactivation of inactive Fe-protein of nitrogenase from Rhodospirillumrubrum. Biochim. Biophys. Acta 791, 21–27 CrossRef

27. Li, X. D., Huergo, L. F., Gasperina, A., Pedrosa, F. O., Merrick, M., andWinkler, F. K. (2009) Crystal structure of dinitrogenase reductase-activat-ing glycohydrolase (DraG) reveals conservation in the ADP-ribosylhydro-lase fold and specific features in the ADP-ribose-binding pocket. J. Mol.Biol. 390, 737–746 CrossRef Medline

28. Berthold, C. L., Wang, H., Nordlund, S., and Högbom, M. (2009) Mecha-nism of ADP-ribosylation removal revealed by the structure and ligandcomplexes of the dimanganese mono-ADP-ribosylhydrolase DraG. Proc.Natl. Acad. Sci. U.S.A. 106, 14247–14252 CrossRef Medline

29. Kabsch, W., and Sander, C. (1983) Dictionary of protein secondary struc-ture: pattern recognition of hydrogen-bonded and geometrical features.Biopolymers 22, 2577–2637 CrossRef Medline

30. Finch, K. E., Knezevic, C. E., Nottbohm, A. C., Partlow, K. C., and Hergen-rother, P. J. (2012) Selective small molecule inhibition of poly(ADP-ribose)glycohydrolase (PARG). ACS Chem. Biol. 7, 563–570 CrossRef Medline

31. Kim, I. K., Kiefer, J. R., Ho, C. M., Stegeman, R. A., Classen, S., Tainer, J. A.,and Ellenberger, T. (2012) Structure of mammalian poly(ADP-ribose) gly-cohydrolase reveals a flexible tyrosine clasp as a substrate-binding ele-ment. Nat. Struct. Mol. Biol. 19, 653– 656 CrossRef Medline

32. Tucker, J. A., Bennett, N., Brassington, C., Durant, S. T., Hassall, G.,Holdgate, G., McAlister, M., Nissink, J. W., Truman, C., and Watson, M.(2012) Structures of the human poly(ADP-ribose) glycohydrolase cata-lytic domain confirm catalytic mechanism and explain inhibition by ADP-HPD derivatives. PLoS One 7, e50889 CrossRef Medline

33. Vasella, A., Davies, G. J., and Böhm, M. (2002) Glycosidase mechanisms.Curr. Opin. Chem. Biol. 6, 619 – 629 CrossRef Medline

34. Gottipati, P., Vischioni, B., Schultz, N., Solomons, J., Bryant, H. E., Djure-inovic, T., Issaeva, N., Sleeth, K., Sharma, R. A., and Helleday, T. (2010)Poly(ADP-ribose) polymerase is hyperactivated in homologous recombi-nation-defective cells. Cancer Res. 70, 5389 –5398 CrossRef Medline

35. Bieche, I., Pennaneach, V., Driouch, K., Vacher, S., Zaremba, T., Susini, A.,Lidereau, R., and Hall, J. (2013) Variations in the mRNA expression ofpoly(ADP-ribose) polymerases, poly(ADP-ribose) glycohydrolase andADP-ribosylhydrolase 3 in breast tumors and impact on clinical outcome.Int. J. Cancer 133, 2791–2800 CrossRef Medline

36. Ossovskaya, V., Koo, I. C., Kaldjian, E. P., Alvares, C., and Sherman, B. M.(2010) Upregulation of poly(ADP-ribose) polymerase-1 (PARP1) in triple-negative breast cancer and other primary human tumor types. Genes Can-cer 1, 812– 821 CrossRef Medline

37. Rojo, F., García-Parra, J., Zazo, S., Tusquets, I., Ferrer-Lozano, J., Menen-dez, S., Eroles, P., Chamizo, C., Servitja, S., Ramírez-Merino, N., Lobo, F.,Bellosillo, B., Corominas, J. M., Yelamos, J., Serrano, S., et al. (2012) Nu-clear PARP-1 protein overexpression is associated with poor overall sur-vival in early breast cancer. Ann. Oncol. 23, 1156 –1164 CrossRef Medline

38. Otwinowski, Z., and Minor, W. (1997) Processing of X-ray diffraction datacollected in oscillation mode. Methods Enzymol. 276, 307–326 CrossRefMedline

39. Pflugrath, J. W. (1999) The finer things in X-ray diffraction data collection.Acta Crystallogr. D Biol. Crystallogr. 55, 1718 –1725 CrossRef Medline

40. Vagin, A., and Teplyakov, A. (2010) Molecular replacement with MOL-REP. Acta Crystallogr. D Biol. Crystallogr. 66, 22–25 CrossRef Medline

41. Winn, M. D., Ballard, C. C., Cowtan, K. D., Dodson, E. J., Emsley, P., Evans,P. R., Keegan, R. M., Krissinel, E. B., Leslie, A. G., McCoy, A., McNicholas,S. J., Murshudov, G. N., Pannu, N. S., Potterton, E. A., Powell, H. R., et al.(2011) Overview of the CCP4 suite and current developments. Acta Crys-tallogr. D Biol. Crystallogr. 67, 235–242 CrossRef Medline

42. Emsley, P., and Cowtan, K. (2004) Coot: model-building tools for mo-lecular graphics. Acta Crystallogr. D Biol. Crystallogr. 60, 2126 –2132CrossRef Medline

43. Afonine, P. V., Grosse-Kunstleve, R. W., Echols, N., Headd, J. J., Moriarty,N. W., Mustyakimov, M., Terwilliger, T. C., Urzhumtsev, A., Zwart, P. H.,and Adams, P. D. (2012) Towards automated crystallographic structurerefinement with phenix.refine. Acta Crystallogr. D Biol. Crystallogr. 68,352–367 CrossRef Medline

44. Krieger, E., Koraimann, G., and Vriend, G. (2002) Increasing the precisionof comparative models with YASARA NOVA—a self-parameterizingforce field. Proteins 47, 393– 402 CrossRef Medline

45. Pettersen, E. F., Goddard, T. D., Huang, C. C., Couch, G. S., Greenblatt,D. M., Meng, E. C., and Ferrin, T. E. (2004) UCSF Chimera—a visualiza-tion system for exploratory research and analysis. J. Comput. Chem. 25,1605–1612 CrossRef Medline

Structure of human ARH3 bound to ADP-ribose and Mg2�

J. Biol. Chem. (2018) 293(32) 12350 –12359 12359

by guest on September 25, 2020

http://ww

w.jbc.org/

Dow

nloaded from

KimYasin Pourfarjam, Jessica Ventura, Igor Kurinov, Ahra Cho, Joel Moss and In-Kwon

reveals a conformational switch that enables specific substrate recognitionStructure of human ADP-ribosyl-acceptor hydrolase 3 bound to ADP-ribose

doi: 10.1074/jbc.RA118.003586 originally published online June 15, 20182018, 293:12350-12359.J. Biol. Chem. 

  10.1074/jbc.RA118.003586Access the most updated version of this article at doi:

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

  http://www.jbc.org/content/293/32/12350.full.html#ref-list-1

This article cites 45 references, 13 of which can be accessed free at

by guest on September 25, 2020

http://ww

w.jbc.org/

Dow

nloaded from