Synthesis of Dimeric ADP-Ribose and Its Structure with Human Poly(ADP-ribose) Glycohydrolase

Article abstract of DOI:10.1021/ja512528p

Poly(ADP-ribosyl)ation is a common post-translational modification that mediates a wide variety of cellular processes including DNA damage repair, chromatin regulation, transcription, and apoptosis. The difficulty associated with accessing poly(ADP-ribose) (PAR) in a homogeneous form has been an impediment to understanding the interactions of PAR with poly(ADP-ribose) glycohydrolase (PARG) and other binding proteins. Here we describe the chemical synthesis of the ADP-ribose dimer, and we use this compound to obtain the first human PARG substrate-enzyme cocrystal structure. Chemical synthesis of PAR is an attractive alternative to traditional enzymatic synthesis and fractionation, allowing access to products such as dimeric ADP-ribose, which has been detected but never isolated from natural sources. Additionally, we describe the synthesis of an alkynylated dimer and demonstrate that this compound can be used to synthesize PAR probes including biotin and fluorophore-labeled compounds. The fluorescently labeled ADP-ribose dimer was then utilized in a general fluorescence polarization-based PAR-protein binding assay. Finally, we use intermediates of our synthesis to access various PAR fragments, and evaluation of these compounds as substrates for PARG reveals the minimal features for substrate recognition and enzymatic cleavage. Homogeneous PAR oligomers and unnatural variants produced from chemical synthesis will allow for further detailed structural and biochemical studies on the interaction of PAR with its many protein binding partners. (Chemical Equation Presented).

Full text of DOI:10.1021/ja512528p

Article
pubs.acs.org/JACS
Synthesis of Dimeric ADP-Ribose and Its Structure with Human
Poly(ADP-ribose) Glycohydrolase
Michael J. Lambrecht,†,§ Matthew Brichacek,†,§,∥ Eva Barkauskaite,‡ Antonio Ariza,‡ Ivan Ahel,‡
and Paul J. Hergenrother*,†
†Department of Chemistry, Roger Adams Laboratory, University of Illinois, 600 South Mathews, Urbana, Illinois 61801, United States
‡Sir William Dunn School of Pathology, University of Oxford, South Parks Road, Oxford OX1 3RE, United Kingdom
S
* Supporting Information
ABSTRACT: Poly(ADP-ribosyl)ation is a common post-transla-
tional modification that mediates a wide variety of cellular
processes including DNA damage repair, chromatin regulation,
transcription, and apoptosis. The difficulty associated with
accessing poly(ADP-ribose) (PAR) in a homogeneous form has
been an impediment to understanding the interactions of PAR with
poly(ADP-ribose) glycohydrolase (PARG) and other binding
proteins. Here we describe the chemical synthesis of the ADP-
ribose dimer, and we use this compound to obtain the first human
PARG substrate-enzyme cocrystal structure. Chemical synthesis of
PAR is an attractive alternative to traditional enzymatic synthesis
and fractionation, allowing access to products such as dimeric ADP-
ribose, which has been detected but never isolated from natural
sources. Additionally, we describe the synthesis of an alkynylated dimer and demonstrate that this compound can be used to
synthesize PAR probes including biotin and fluorophore-labeled compounds. The fluorescently labeled ADP-ribose dimer was
then utilized in a general fluorescence polarization-based PAR−protein binding assay. Finally, we use intermediates of our
synthesis to access various PAR fragments, and evaluation of these compounds as substrates for PARG reveals the minimal
features for substrate recognition and enzymatic cleavage. Homogeneous PAR oligomers and unnatural variants produced from
chemical synthesis will allow for further detailed structural and biochemical studies on the interaction of PAR with its many
protein binding partners.
chain polymers (>50 units),10,11 this distribution is not
representative of PAR produced in whole cells and in vivo
INTRODUCTION
■
Poly(ADP-ribosyl)ation is a critically important post-transla-
where shorter oligomers (2−20 units) make up a greater
tional modification affecting a wide variety of cellular
percentage.10,12−14 The ADP-ribose dimer (1, Figure 1b)
processes.1,2 In cells, poly(ADP-ribose) (PAR) is produced
represents the simplest form of PAR that still retains the
structural features of the polymer (multiple pyrophosphates
by PAR polymerases (PARPs) from β-NAD+ in response to
DNA damage. PAR recruits a number of proteins responsible
for DNA repair, and is ultimately processed to ADP-ribose by
PAR glycohydrolase (PARG) (Figure 1a). The list of proteins
with PAR-binding activity is ever-growing; PAR has been
and an α-ribosyl linkage between ADP-ribose units), and is
produced in only small quantities in vitro.15 Though it has been
described as an “ideal ligand for x-ray co-crystal structure
determination”16 and detected by a variety of analytical
experimentally shown to bind proteins through at least four
methods including MS analysis17 there are no reports where
distinct protein binding domains, and is believed to interact
with >500 proteins.3−7 The details of PAR−protein interactions
on a molecular level, however, are poorly understood, primarily
due to the difficulty associated with obtaining PAR in
this compound has been isolated from natural sources or
enzymatically or chemically synthesized.
If short PAR oligomers such as the ADP-ribose dimer (1)
could be accessed through a modular chemical route, this
homogeneous and pure form. For use in vitro, PAR can be
enzymatically produced by PARP from β-NAD+ to afford
would afford these valuable compounds in purities and
quantities that are not attainable by enzymatic production.
polymers ranging from 2 to 200 units in length (Figure 1a) and
Additionally, chemical synthesis could allow for facile
subsequently fractionated down to narrow distributions by
anion exchange chromatography.8,9 While important advances
in this process have been made, enzymatic synthesis and
modification of PAR oligomers, providing access to unnatural
compounds such as the propargyl ADP-ribose dimer (2).
fractionation remains challenging, yielding small quantities after
repeated ion-exchange purification.10,11 Additionally, while in
Received: December 9, 2014
vitro enzymatic syntheses of PAR commonly produces long
© XXXX American Chemical Society
A
DOI: 10.1021/ja512528p
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
phate unit. Methods to construct pyrophosphates are often
low-yielding and involve long reaction times.22 To overcome
these challenges, we envisioned accessing the ADP-ribose
dimer (1) and propargyl ADP-ribose dimer (2) from
pyrophosphate-forming couplings of protected ribose phos-
phates 3 and 4, respectively, phosphorylated disaccharide 5, and
protected adenosine monophosphate 6 (Figure 1b). Dis-
accharide 5 would be formed from an α-selective glycosylation
of a protected adenosine by an orthogonally protected ribose.
RESULTS
■
Chemical Synthesis and Enzymatic Evaluation of an
ADP-Ribose Dimer. Attention was first turned toward
selective glycosylation of adenosine (Scheme 1). To control
selectivity, N-benzoyladenosine (7) was silylated in a two step
sequence to afford 5′,3′-silyl isomer 8 in ∼15:1 selectivity over
the 5′,2′-silylated isomer (Supporting Information Figure 1).
Next, D-ribose (9) was efficiently converted to β-glycosyl
fluoride 10 in a six step sequence involving introduction of the
small and nonparticipating allyl ether protecting groups for the
C-2 and C-3 alcohols of ribose, the 2-naphthylmethyl ether
group for protection of the C-5 alcohol, and conversion of the
C-1 alcohol to the glycosyl fluoride23 to allow for activation
under mild conditions (Supporting Information Figure 2).
Disaccharide 11 was synthesized by activation of glycosyl donor
10 with tin(II) chloride and silver(I) salts in the presence of
glycosyl acceptor 8 (Scheme 1). While traditional silver
perchlorate/tin(II) chloride23 promoted systems provided
some product, it was found that replacement of silver
perchlorate with silver hexafluorophosphate (AgPF6) led to
dramatically increased yields of 11. While the use of silver
hexafluorophosphate has been reported for activation of
thioglycosides,24 we are unaware of reports of its use for the
activation of glycosyl fluorides. Only the desired α-anomer of
disaccharide 11 was isolated, consistent with the α-stereo-
chemical selectivity of other ribose glycosylations, which has
been investigated in detail for similar systems.25,26 The
glycosylation was found to be exceptionally sensitive to the
steric bulk of the glycosyl donor (data not shown), and the
protecting groups chosen for compound 10 allowed for
efficient glycosylation while still providing enough orthogon-
ality to allow for further elaboration of 11.
Figure 1. (a) Synthesis and degradation of poly(ADP-Ribose), an
important cellular signaling molecule, and (b) ADP-ribose dimer (1)
and propargyl ADP-ribose dimer (2) and retrosynthetic disconnec-
tions.
Compounds such as 2 could be used in combination with click
chemistry to append various biochemical probes enabling
further interrogation of PAR biochemistry.
As a target for chemical synthesis, compounds such as 1 and
2 specifically and PAR more generally present several
challenges. Construction of the α-glycosidic linkage between
the ribofuranose and the adenosine 2′-hydroxyl involves
creation of a 1,2-cis-furanoside, a major challenge of synthetic
carbohydrate chemistry.18 Additionally, nucleosides are tradi-
tionally challenging substrates for chemical glycosylation due to
the reactivity of the nitrogenous base.19,20 While two important
chemical syntheses of 2′-O-α-ribosylated adenosine exist,20,21
drawbacks of both approaches related to their orthogonality
and efficiency inspired us to devise our own alternative route. A
second challenge is the presence of the repeating pyrophos-
a
Scheme 1. Synthesis of Phosphorylated Disaccharide
a
(a) SnCl2 (2.2 equiv), AgPF6 (2.2 equiv), 4 Å MS, CH2Cl2, −78 to 4 °C, 12 h, 72%; (b) Pd(PPh3)4 (20 mol %), N,N-dimethylbarbituric acid (10
equiv), MeOH, 16 h, 99%; (c) TBSOTf (3 equiv), DMAP (0.2 equiv), EtNiPr2 (6 equiv), CH2Cl2, 0 °C to rt, 2 h, 75%; (d) CCl3CO2H (46 equiv),
THF/H2O (1/1), 0 °C, 5 h, 85%; (e) (i-Pr)2NP(OBn)2 (1.2 equiv), 4,5-dicyanoimidazole (1.2 equiv), CH2Cl2/CH3CN (4/1), rt, 2 h, then
tBuOOH (5 equiv), 0 °C to rt, 3 h, 91%; (f) DDQ (1.5 equiv), CH2Cl2/H2O (4/1), 0 °C, 6 h, 92%; (g) Cl(i-Pr)2NPO(CH2)2CN (16) (1.2 equiv),
EtNiPr2 (1.2 equiv), THF, −78 °C to rt, 3 h, 99%; (h) H2O (3 equiv), 4,5-dicyanoimidazole (1.2 equiv), CH3CN, rt, 2 h, 99%.
B
DOI: 10.1021/ja512528p
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
a
Scheme 2. Synthesis of the ADP-Ribose Dimer (1) and Processing by PARG
a
(a) N-Chlorosuccinimide (4 equiv), EtNiPr2 (4 equiv), CH3CN, rt, 20 min, then DBU (10 equiv), CH3CN, rt, 20 min, then C-18 chromatography
and cation exchange, 72%; (b) H2, Pd/C, Et3N (20 equiv), tBuOH/H2O, rt, 16 h, 92%; (c) 3, N,N-carbonyldiimidazole (10 equiv), Et3N (4 equiv),
pyridine, rt, 2 h, then 18, ZnCl2 (8 equiv), DMF, rt, 96 h, then EDTA (Et3NH+ form) and C-18 chromatography, 79%; (d) NH3/MeOH (7M), rt,
20 h, then Bu4NF (33 equiv), THF, rt, 3 h, then precipitation, C-18 ion pairing chromatography, and cation exchange, 77%. Inset: Processing of 100
μM 1 by 7.5 nM human PARG shown by HPLC. Compound 1 is converted to ADP-ribose in 2 h by WT PARG while the catalytically inactive
mutants (E755N and E756N) do not process 1 at 10-fold higher enzyme concentration at 4 h.
Disaccharide 11 was converted to a phosphorylated
disaccharide to be used for pyrophosphate formation. Toward
this end, compound 11 was deallylated with Pd(PPh3)4 to
afford diol 12 and then silylated with TBSOTf to afford 13 and
enable a facile global desilylation at the end of the synthesis.
The 5′-TBS group of 13 was then selectively desilylated in the
presence of the secondary TBS ethers using trichloroacetic acid
to afford primary alcohol 14. Phosphoramidite coupling and
oxidation provided phosphate 15 in good yield. Removal of the
2-naphthylmethyl group of 15 under oxidative conditions,
coupling with chlorophosphoramidite 16, and subsequent
hydrolysis afforded H-phosphonate 5.
The fragment coupling reactions were next accomplished
(Scheme 2). First, H-phosphonate 5 was coupled to protected
adenosine phosphate 6 (synthesized as described in Supporting
Information Figure 3) to afford pyrophosphate 17. This
reaction is an adaptation of the classic Atherton-Todd
reaction,27 a method commonly used for formation of
phosphates and phosphoramidates.28 In this case, the reaction
likely proceeds through oxidative chlorination of H-phospho-
nate 5 to afford a chlorophosphate which can then condense
with phosphate 6 and afford pyrophosphate diester 17 upon
deprotection of cyanoethyl group of the intermediate
pyrophosphate triester with DBU. Use of this coupling method
afforded the product in good yield (72%) and allowed for the
reaction to be completed quickly (<1 h). Use of this H-
phosphonate coupling allowed for easier and more scalable
synthesis of compound 17 than the more commonly used
phosphomorphilodate and phosphorimidazolide methods to
form pyrophosphate diesters.22 Hydrogenolysis of the benzyl
groups of pyrophosphate 17 afforded deprotected phosphate
18. Attempts to form the second pyrophosphate with the H-
phosphonate method described above were unsuccessful,
leading us to synthesize (Supporting Information Figure 4)
and activate compound 3 as the phosphorimidazolide.
Condensation of the phosphorimidazolide of 3 with phosphate
18 allowed for formation of the second pyrophosphate to afford
fully protected dimer 19. Global deprotection of compound 19
gave the ADP-ribose dimer (1) in good yield after HPLC
purification. To confirm the structure of dimer 1, full two-
dimensional (2D) NMR spectroscopy data was obtained for
compound 1 as well as the preceding intermediates (see the
Supporting Information).
The processing of the ADP-ribose dimer (1) by PARG was
next examined (Scheme 2, inset). Treatment of dimer 1 with
human PARG at a concentration of 7.5 nM allowed for
complete conversion of 1 to ADP-ribose in approximately 2 h
(Scheme 2 and Supporting Information Figure 5) as detected
by LC/MS. No processing of 1 was observed by the
catalytically inactive E755N and E756N human PARG
mutants,29 even at 10-fold higher concentrations of enzyme
and longer reaction times (4 h) (Scheme 2). Additionally,
dimer 1 was shown to be a competent substrate for PARG from
other organisms (Tetrahymena thermophilia, Bos taurus, and
Thermomonospora curvata30) as well as for ADP-ribosylhy-
drolase 3 (ARH3), another enzyme with PAR processing
activity31 (Supporting Information Figure 6).
Human PARG Co-Crystal Structure. The inability of
mutant PARG to process 1 afforded the opportunity to
determine the three-dimensional structure of the PARG mutant
with this novel substrate. While X-ray structures of human
PARG exist,29 there are no co-crystal structures of human
PARG with a PAR oligomer. Dimer 1 was co-crystallized with a
catalytically inactive mutant (E756N) human PARG. The
C
DOI: 10.1021/ja512528p
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 2. Surface representation of human PARG (E756N) with bound 1 in the active site. The 2Fo-Fc electron density corresponding to the
ordered region of dimer 1 is shown in a gray mesh at contour level of 1σ. (b) Overlay of protozoan (4L2H, pink) and human PARG (this work,
blue) structures bound to a PAR fragment (pink) and dimer 1 (this work, yellow), respectively. Please note that, due to the flexibility observed for
the human N-1 β-phosphate group, two different β-phosphate conformations can be assigned, and only one has been retained for clarity of the figure.
Also, the N-1 ribose moiety is not shown as this portion of the structure is disordered.
a
Scheme 3. Synthesis of ADP-Ribose Dimer Derivatives
a
(a) 4, N,N-carbonyldiimidazole (5 equiv), Et3N (2 equiv), CH3CN, rt, 2 h, then 18, 4,5-dicyanoimidazole (8 equiv), rt, 16 h, 25%; (b) NH3/MeOH
(7M), rt, 14 h, then Bu4NF, THF, rt, 5 h, then precipitation, C-18 ion pairing chromatography, and cation exchange, 36%; (c) Biotin-Peg3-N3 (2
equiv), CuSO4·5H2O (2 equiv), Cu wire, H2O, rt, 16 h; (d) SulfoCy3-N3, CuSO4·5H2O (0.6 equiv), Cu Wire, H2O, rt, 16 h.
E756N mutant is devoid of PARG enzymatic activity without
alteration in the shape of the substrate binding pocket,29
making it an excellent candidate for crystallization studies. The
complex reported here (Figure 2a) was solved to 1.9 Å
resolution and represents the first crystal structure of human
PARG in complex with a PAR oligomer of any length.
In comparison with the published structure of mutant T.
thermophila PARG in complex with PAR oligomers (PDB ID:
4L2H, oligomers obtained through enzymatic synthesis and
isolation),17 there are similarities but also important differences.
In particular, the interactions with the N ADP-ribose unit
(Figure 2b) are retained in both structures. However, the
enzyme−substrate complex reported herein allows the first
view of the ribose−ribose O-glycosidic linkage and the N-1
adenine in the human system, and it is here where differences
with the T. thermophila enzyme are observed. Specifically, due
to the absence of a residue homologous to Arg164, present in
the protozoan enzyme, and due to the protozoan Ser297 being
replaced by Ala796 in the human enzyme, the N-1 adenine
group in human PARG appears to be significantly more solvent
exposed than in its protozoan orthologue and thus makes fewer
contacts with the enzyme (Figure 2b). The N-1 adenine group
interacts with the amide nitrogen of the conserved Leu752 in
human PARG (Leu252 in protozoan PARG), while the β-
phosphate forms H-bonds with the conserved Ala750 (Asn250
in protozoan PARG) (Figure 2b). In the human PARG
structure the N-1 β-phosphate is positioned differently to its
counterpart in protozoan PARG and it forms H-bonds with the
side chain of the Arg663 (located in the human PARG
accessory domain17), which is conserved in a majority of
metazoan PARGs. Both the human and T. thermophila PARG−
PAR structures suggest that binding of branched PAR
D
DOI: 10.1021/ja512528p
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
fragments appears to be discouraged due to the steric hindrance
around the 2″-OH position of the N-1 ribose.
Synthesis of Derivatives of the ADP-Ribose Dimer
and Development of a PAR Binding Assay. Using
modifications of our synthetic route, variants of 1 were
constructed to provide tools to interrogate PAR biochemistry.
Specifically, analogues amenable to functionalization via bio-
orthogonal click chemistry were constructed such that these
compounds could be functionalized with various probes.
Toward this end, alkyne ribose phosphate 4 was synthesized
(Supporting Information Figure 7), activated as the corre-
sponding phosphorimidazolide, and coupled to phosphate 18
(Scheme 3) to furnish protected propargyl ADP-ribose dimer
20. Global deprotection provided propargyl ADP-ribose dimer
2. This derivative contains a nonreducing sugar and the non-
natural β-glycosidic linkage at the N-1 ribose. To demonstrate
its utility as a generally taggable PAR derivative, compound 2
was treated with a biotin azide under azide−alkyne Husigen
cycloaddition conditions to afford biotinylated ADP-ribose
dimer 21. Similar chemistry was performed with compound 2
and sulfo-Cy3 azide to afford Cy3 ADP-ribose dimer 22, a
fluorescent PAR derivative.
Using fluorescent dimer 22, we sought to develop a general
fluorescence polarization-based PAR−protein binding assay.
Incubation of compound 22 with E756N and E755N human
PARG provided binding curves (Figure 3a) that were used to
calculate dissociation constants. Compound 22 binds to E756N
and E755N human PARG (Figure 3a) with KD values of 83
7
and 208 14 nM, respectively (see the Supporting Information
for calculation). Fluorescent dimer 22 does not bind proteins
devoid of PAR binding activity such as bovine serum albumin
(Figure 3a) even at high concentrations of protein (Supporting
Information Figure 8a). Binding of compound 22 to E756N
PARG can be disrupted by preincubation of the protein with
compound 1 or the PARG inhibitor ADP-HPD32−35 (Figure
3b). Interestingly, it was found that ADP-HPD has substantially
lower affinity for the E755N PARG mutant both relative to
compound 1 and relative to its affinity for the E756N mutant
(Figure 3b and Supporting Information Figure 8b). This data
could suggest that the E756N mutant more accurately mimics
the wild type enzyme and is a better candidate for
crystallography than E755N human PARG.
Synthesis and Evaluation of Fragments of the ADP-
Ribose Dimer. Another benefit of a modular synthesis of
ADP-ribose oligomers is the ability to divert key intermediates
to create simple fragments of PAR oligomers and use these
fragments to investigate PARG substrate specificity. Toward
this end, compound 12 was globally deprotected to reveal the
fully deprotected disaccharide 2320,21 (Scheme 4). Additionally,
the 2-naphthyl methyl group of compound 14 was cleaved
selectively and phosphorylated at both primary alcohols to give
bisphosphate 24 (Scheme 4, known as iso-ADP-ribose36) upon
global deprotection (for a detailed synthetic scheme, see
Supporting Information Figure 9). This represents the first
chemical synthesis of this widely used molecule,36−39 and we
provide the first detailed characterization for this compound
including full 2D NMR spectroscopic analysis (see the
Supporting Information). Finally, global deprotection of 18
furnished compound 25 (Scheme 4), a fragment retaining the
N ADP-ribose unit and N-1 AMP unit, while lacking the N-1
ribose moiety.
Figure 3. Analysis of PAR−protein binding by fluorescence polar-
ization. (a) Binding of E755N (black squares) and E756N (red circles)
PARGs but not bovine serum albumin (blue diamonds) to compound
22 (7.5 nM). (b) Compound 22 binding to E756N mutant PARG
(250 nM) can be blocked by pretreatment with unlabeled ligand
(compound 1) or the PARG inhibitor ADP-HPD. Error bars represent
standard deviation (n = 3)
process the truncated substrates 23 and 24 at 25 nM (as shown
in Figure 4), and even at enzyme concentrations of 100 nM no
processing was observed (data not shown). Building out from
the glycosidic linkage, PARG was able to process compound 25
to adenosine monophosphate and ADP-ribose, indicating that
PARG is tolerant to truncated substrates in the N-1 region of
the polymer. Finally, in addition to truncations in the N-1
portion of the polymer, unnatural modifications of the natural
substrate are tolerated as PARG was able to process the
propargyl ADP-ribose dimer (2) to ADP-ribose and propargyl
ADP-ribose (Figure 4). Additionally, the β-glycosidic linkage of
the propargyl group of compound 2 was not hydrolyzed. That
PARG was able to process compounds 1, 2 and 25 with similar
efficiency but not compounds 23 and 24 suggests that the N
ADP-ribose motif is essential for PAR recognition, binding, and
processing but PARG appears more tolerant to substitutions in
the N-1 region.
DISCUSSION
■
We present here the first total chemical synthesis of the ADP-
ribose dimer 1. This represents the first time that this
compound has been obtained and characterized and represents
The ability of PARG to process these dimer fragments was
next examined (Figure 4). We found that PARG was unable to
E
DOI: 10.1021/ja512528p
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Scheme 4. Synthesis of Simplified PAR Fragments
an alternative to accessing PAR oligomers via enzymatic
synthesis and fractionation. Key to the synthesis of 1 has
been the development of a scalable α-selective glycosylation of
adenosine and the development of an efficient H-phosphonate
coupling method for pyrophosphate formation. In addition to
the chemistry presented here, the H-phosphonate coupling
method has found wide use in our laboratory as a fast, high-
yielding, and convenient method to form pyrophosphate
diesters for a wide variety of substrates. We have found that
reactions using this coupling method scale easier than
traditional phosphoimidazolide and phosphomorphilodate
chemistry as this method does not require the addition and
subsequent removal of metal additives such as ZnCl2.
The ability to access compound 1 enabled the first crystal
structure of human PARG bound to a cognate substrate.
Consistent with the previously reported17 T. thermophilia
structure, the N ADP-ribose unit appears to make the most
contacts with the protein. We found the N-1 adenine ring in
the human structure to be more solvent exposed than in the T.
thermophilia structure and found changes to the orientation of
the N-1 pyrophosphate moiety due to hydrogen bonds with
Arg663 not present in the T. thermophilia structure. Chemical
synthesis of PAR oligomers should facilitate crystallography
experiments with proteins other than PARG and should enable
understanding of PAR−protein interactions on a molecular
level.
The modular nature of our synthetic route allows
intermediates such as 12, 14, and 18 to be easily diverted to
the synthesis of PAR fragments. Experiments with PARG show
that the N ADP-ribose unit of PAR is essential for binding and
recognition by PARG, but that truncations or unnatural
modifications are tolerated in the N-1 unit (summarized in
Figure 4). This finding is consistent with the PARG−PAR
crystal structures where the N unit is bound most tightly
(Figure 2). The somewhat promiscuous nature of PARG
demonstrated here could enable the generation of alternative
PARG substrates and inhibitors inspired by ADP-ribose.
The synthetic route to the ADP-ribose dimer (1) served as a
blueprint for the synthesis of other short PAR oligomers,
Figure 4. LC/MS assay of enzymatic cleavage of PAR derivatives by T.
thermophilia PARG.
providing access to unnatural PAR variants including the
propargyl ADP-ribose dimer (2). The ability of compound 2 to
be used as a generally taggable PAR derivative will enable the
synthesis of a variety of biochemical probes through click
chemistry, as demonstrated by the synthesis of the biotinylated
ADP-ribose dimer (21) and the fluorescent ADP-ribose dimer
(22).
Fluorescent ADP-ribose dimer 22 should be a valuable tool
for PAR binding assays as well as for high-throughput screening
campaigns. The development of PARG inhibitors has been
hindered by a lack of convenient methods for high-throughput
screening, as traditionally PARG activity is monitored by a
radiometric TLC assay40 or immunochemical dot blot
assay.41,42 The approach described herein provides quantitative
binding data and can be performed rapidly in a high-throughput
format. Additionally, as fluorescence polarization does not
monitor enzymatic activity but rather directly measures binding
of compound 22 to various proteins, it can be applied to find
inhibitors of PAR−protein binding events for which there is not
an enzymatic readout.
F
DOI: 10.1021/ja512528p
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
(16) Oberoi, J.; Richards, M. W.; Crumpler, S.; Brown, N.; Blagg, J.;
Bayliss, R. J. Biol. Chem. 2010, 285, 39348−39358.
(17) Barkauskaite, E.; Brassington, A.; Tan, E. S.; Warwicker, J.;
Dunstan, M. S.; Banos, B.; Lafite, P.; Ahel, M.; Mitchison, T. J.; Ahel,
I.; Leys, D. Nat. Commun. 2013, 4, 1−8.
ASSOCIATED CONTENT
* Supporting Information
Figures detailing synthesis of compounds 3, 4, 6, 8, 10, and 24
as well as additional PARG processing and fluorescence
polarization data. Crystallographic data. Experimental proce-
dures and characterization of chemical products. This material
is available free of charge via the Internet at http://pubs.acs.org.
■
S
(18) Demchenko, A. V. Handbook of Chemical Glycosylation; Wiley
VCH: Weinheim, 2008.
(19) Knapp, S.; Gore, V. K. J. Org. Chem. 1996, 61, 6744−6747.
(20) Mikhailov, S. N.; Kulikova, I. V.; Nauwelaerts, K.; Herdewijn, P.
Tetrahedron 2008, 64, 2871−2876.
AUTHOR INFORMATION
Corresponding Author
*hergenro@uiuc.edu
■
(21) van der Heden van Noort, G. J.; Overkleeft, H. S.; van der
Marel, G. A.; Filippov, D. V. Org. Lett. 2011, 13, 2920−2923.
(22) Wagner, G. K.; Pesnot, T.; Field, R. A. Nat. Prod. Rep. 2009, 26,
1172−1194.
Present Address
∥M.B.: University of Maine, Department of Chemistry, 5706
Aubert Hall, Orono, Maine 04469.
(23) Mukaiyama, T.; Murai, Y.; Shoda, S. Chem. Lett. 1981, 10, 431−
432.
(24) Lear, M. J.; Yoshimura, F.; Hirama, M. Angew. Chem., Int. Ed.
2001, 40, 946−949.
(25) Larsen, C. H.; Ridgway, B. H.; Shaw, J. T.; Smith, D. M.;
Woerpel, K. A. J. Am. Chem. Soc. 2005, 127, 10879−10884.
(26) van Rijssel, E. R.; van Delft, P.; Lodder, G.; Overkleeft, H. S.;
Author Contributions
§M.J.L. and M.B. contributed equally.
Notes
The authors declare no competing financial interest.
́
van der Marel, G. A.; Filippov, D. V.; Codee, J. D. C. Angew. Chem., Int.
Ed. 2014, 53, 10381−10385.
ACKNOWLEDGMENTS
■
(27) Atherton, F. R.; Openshaw, H. T.; Todd, A. R. J. Chem. Soc.
1945, 660.
We thank the University of Illinois for supporting this work.
M.J.L. is a member of the National Institutes of Health
Chemistry−Biology Interface Training Grant (NRSA 1-T-32-
GM070421). We are grateful to B. Yestrepsky for synthesis of
ADP-HPD, B. Drown and D. Leys for advice, and J. Tucker for
materials. Work in the I.A. laboratory is supported by the
Wellcome Trust and the European Research Council.
(28) Le Corre, S. S.; Berchel, M.; Couthon-Gourves
̀
, H.; Haelters, J.-
P.; Jaffres, P.-A. Beilstein J. Org. Chem. 2014, 10, 1166−1196.
̀
(29) Tucker, J. A.; Bennett, N.; Brassington, C.; Durant, S. T.;
Hassall, G.; Holdgate, G.; McAlister, M.; Nissink, J. W. M.; Truman,
C.; Watson, M. PLoS One 2012, 7, e50889.
(30) Slade, D.; Dunstan, M. S.; Barkauskaite, E.; Weston, R.; Lafite,
P.; Dixon, N.; Ahel, M.; Leys, D.; Ahel, I. Nature 2011, 477, 616−620.
(31) Niere, M.; Mashimo, M.; Agledal, L.; Dolle, C.; Kasamatsu, A.;
̈
REFERENCES
■
364.
Kato, J.; Moss, J.; Ziegler, M. J. Biol. Chem. 2012, 287, 16088−16102.
(32) Slama, J. T.; Aboul-Ela, N.; Jacobson, M. K. J. Med. Chem. 1995,
38, 4332−4336.
(1) Malanga, M.; Althaus, F. R. Biochem. Cell Biol. 2005, 83, 354−
(2) Schreiber, V.; Dantzer, F.; Ame,
Mol. Cell Biol. 2006, 7, 517−528.
(3) Krietsch, J.; Rouleau, M.; Pic, E.; Ethier, C.; Dawson, T. M.;
Dawson, V. L.; Masson, J.-Y.; Poirier, G. G.; Gagne, J.-P. Mol. Aspects
Med. 2013, 34, 1066−1087.
(4) Gagne, J. P.; Isabelle, M.; Lo, K. S.; Bourassa, S.; Hendzel, M. J.;
Dawson, V. L.; Dawson, T. M.; Poirier, G. G. Nucleic Acids Res. 2008,
36, 6959−6976.
(5) Pleschke, J. M.; Kleczkowska, H. E.; Strohm, M.; Althaus, F. R. J.
Biol. Chem. 2000, 275, 40974−40980.
(6) Ahel, I.; Ahel, D.; Matsusaka, T.; Clark, A. J.; Pines, J.; Boulton, S.
J.; West, S. C. Nature 2008, 451, 81−85.
(7) Ahel, D.; Horejsi, Z.; Wiechens, N.; Polo, S. E.; Garcia-Wilson,
E.; Ahel, I.; Flynn, H.; Skehel, M.; West, S. C.; Jackson, S. P.; Owen-
Hughes, T.; Boulton, S. J. Science 2009, 325, 1240−1243.
(8) Andrabi, S. A. S.; Kim, N. S. N.; Yu, S.-W. S.; Wang, H. H.; Koh,
D. W. D.; Sasaki, M. M.; Klaus, J. A. J.; Otsuka, T. T.; Zhang, Z. Z.;
Koehler, R. C. R.; Hurn, P. D. P.; Poirier, G. G. G.; Dawson, V. L. V.;
Dawson, T. M. T. Proc. Natl. Acad. Sci. U. S. A. 2006, 103, 18308−
18313.
́
J.-C.; de Murcia, G. Nat. Rev.
(33) Slama, J. T.; Aboul-Ela, N.; Goli, D. M.; Cheesman, B. V.;
Simmons, A. M.; Jacobson, M. K. J. Med. Chem. 1995, 38, 389−393.
(34) Koh, D. W.; Coyle, D. L.; Mehta, N.; Ramsinghani, S.; Kim, H.;
Slama, J. T.; Jacobson, M. K. J. Med. Chem. 2003, 46, 4322−4332.
(35) Patel, C.; Koh, D. W.; Jacobson, M. K.; Oliveira, M. A. Biochem.
J. 2005, 388, 493−500.
́
́
(36) Wang, Z.; Michaud, G. A.; Cheng, Z.; Zhang, Y.; Hinds, T. R.;
Fan, E.; Cong, F.; Xu, W. Genes Dev. 2012, 26, 235−240.
(37) Wang, Z.; Gagne,
e86010.
́
J.-P.; Poirier, G. G.; Xu, W. PLoS One 2014, 9,
(38) Zhang, F.; Chen, Y.; Li, M.; Yu, X. Proc. Natl. Acad. Sci. U. S. A.
2014, 111, 7278−7283.
(39) DaRosa, P. A.; Wang, Z.; Jiang, X.; Pruneda, J. N.; Cong, F.;
Klevit, R. E.; Xu, W. Nature 2015, 517, 223−226.
(40) Men
673.
(41) Bachir Affar, El; Duriez, P. J.; Shah, R. G.; Sallmann, F. R.;
́
ard, L.; Poirier, G. G. Biochem. Cell Biol. 1987, 65, 668−
Bourassa, S.; Kupper, J.-H.; Burkle, A.; Poirier, G. G. Anal. Biochem.
̈
1998, 259, 280−283.
(42) Okita, N.; Ashizawa, D.; Ohta, R.; Abe, H.; Tanuma, S.-I.
Biochem. Biophys. Res. Commun. 2010, 392, 485−489.
(9) Fahrer, J.; Kranaster, R.; Altmeyer, M.; Marx, A.; Burkle, A.
Nucleic Acids Res. 2007, 35, e143−e143.
(10) Kiehlbauch, C. C. C.; Aboul-Ela, N. N.; Jacobson, E. L. E.;
Ringer, D. P. D.; Jacobson, M. K. M. Anal. Biochem. 1993, 208, 26−34.
(11) Tan, E. S.; Krukenberg, K. A.; Mitchison, T. J. Anal. Biochem.
2012, 428, 126−136.
NOTE ADDED IN PROOF
■
For a recent synthesis of O-methyl acetal protected PAR
oligomers see the work of Filippov, et al. (Angew. Chem. Int. Ed.
2015, DOI: 10.1002/anie.201412283).
(12) Malanga, M.; Bachmann, S.; Panzeter, P. L.; Zweifel, B.; Althaus,
F. R. Anal. Biochem. 1995, 228, 245−251.
(13) Kleczkowska, H. E.; Malanga, M.; Szumiel, I.; Althaus, F. R. Int.
J. Radiat. Biol. 2002, 78, 527−534.
(14) Malanga, M.; Althaus, F. R. J. Biol. Chem. 1994, 269, 17691−
17696.
(15) Popp, O.; Veith, S.; Fahrer, J.; Bohr, V. A.; Burkle, A.;
̈
Mangerich, A. ACS Chem. Biol. 2013, 8, 179−188.
G
DOI: 10.1021/ja512528p
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX