Chain-terminating and clickable NAD+ analogues for labeling the target proteins of ADP-ribosyltransferases

Article abstract of DOI:10.1002/anie.201404431

ADP-ribosyltransferases (ARTs) use NAD⁺ as a substrate and play important roles in numerous biological processes, such as the DNA damage response and cell cycle regulation, by transferring multiple ADP-ribose units onto target proteins to form poly(ADP-ribose) (PAR) chains of variable sizes. Efforts to identify direct targets of PARylation, as well as the specific ADP-ribose acceptor sites, must all tackle the complexity of PAR. Herein, we report new NAD⁺ analogues that are efficiently processed by wild-type ARTs and lead to chain termination owing to a lack of the required hydroxy group, thereby significantly reducing the complexity of the protein modification. Due to the presence of an alkyne group, these NAD⁺ analogues allow subsequent manipulations by click chemistry for labeling with dyes or affinity markers. This study provides insight into the substrate scope of ARTs and might pave the way for the further developments of chemical tools for investigating PAR metabolism.

Full text of DOI:10.1002/anie.201404431

Angewandte
Chemie
DOI: 10.1002/anie.201404431
ADP-Ribosylation
Chain-Terminating and Clickable NAD+ Analogues for Labeling the
Target Proteins of ADP-Ribosyltransferases**
Yan Wang, Daniel Rçsner, Magdalena Grzywa, and Andreas Marx*
Abstract: ADP-ribosyltransferases (ARTs) use NAD+ as
a substrate and play important roles in numerous biological
processes, such as the DNA damage response and cell cycle
regulation, by transferring multiple ADP-ribose units onto
target proteins to form poly(ADP-ribose) (PAR) chains of
variable sizes. Efforts to identify direct targets of PARylation,
as well as the specific ADP-ribose acceptor sites, must all tackle
the complexity of PAR. Herein, we report new NAD+
analogues that are efficiently processed by wild-type ARTs
and lead to chain termination owing to a lack of the required
hydroxy group, thereby significantly reducing the complexity
of the protein modification. Due to the presence of an alkyne
group, these NAD+ analogues allow subsequent manipulations
by click chemistry for labeling with dyes or affinity markers.
This study provides insight into the substrate scope of ARTs
Figure 1. The structures of poly(ADP-ribose) (A), and NAD+ and the
and might pave the way for the further developments of
chemical tools for investigating PAR metabolism.
NAD+ analogues developed and employed in this study (B).
B
ased on its diverse functions, nicotinamide adenine dinu-
cleotide (NAD+) is well known to participate in a variety of
Successive transfer reactions onto the protein–mono-
cellular processes, such as redox metabolism, signaling path- (ADP-ribosyl) adduct, and subsequently onto the nascent
ways, and posttranslational modifications. In 1963, Chambon chain of several covalently linked ADP-ribosyl units, are the
et al. reported for the first time the formation of a nucleic acid basis for the formation of PAR, which consists of up to 200
like polymer derived from NAD+, namely poly(ADP- ADP-ribose units coupled through a 2’-O-a-d-ribofuranosyl-
ribose)[1] (PAR; Figure 1A). ADP-ribosylation is a reversible adenosine diphosphate backbone.[10,11]
posttranslational modification of proteins and catalyzed by
Efforts to identify direct targets of ARTs, as well as the
a
family of enzymes termed ADP-ribosyltransferases specific ADP-ribose acceptor sites, must all tackle the
(ARTs),[2] with ADP-ribosyltransferase diphtheria toxin- complexity of PAR, which complicates subsequent analysis.
like 1 (ARTD1, formerly known as PARP-1) as the best Current approaches harness mutant ARTs that catalyze only
described member. ARTs play important roles in a wide mono(ADP-ribos)ylation;[12] make use of protein-based or
range of biological processes, including DNA repair, main- chemical approaches for enrichment or to cleave PAR chains
tenance of genomic stability,[3–5] and transcriptional regula- from the substrate, thereby resulting in labeling;[13] or use
tion.[6,7] In the human genome, 22 different genes that contain chemically modified NAD+ analogues.[12b,14]
an ADP-ribosyltransferase catalytic domain have been iden-
Herein, we report new NAD+ analogues that are effi-
tified.[8] ADP-ribosylation comprises the transfer of single or ciently incorporated by wild-type ARTs and a) lead to PAR
multiple ADP-ribose moieties from NAD+ to specific amino chain termination owing to a lack of the required hydroxy
acid residues on a target protein, thereby leading to mono- group and b) bear an affinity tag that will allow subsequent
(ADP-ribos)ylation or poly(ADP-ribos)ylation. In this pro- manipulations such as labeling and sample enrichment (Fig-
cess, covalent transfer onto glutamic acid, aspartic acid, or ure 1B). Since little is known about the substrate scope of
lysine residues of target proteins is described.[9]
ARTs,[14,15] we developed four NAD+ analogues (1–4, Fig-
ure 1B) with systematic deoxygenation of the hydroxy groups
of adenosine and an alkyne group on the nucleobase. These
modifications enabled us to investigate the influence of the
respective OH group on ART (i.e., ARTD1) activity.
The synthesis of the analogues 1–4 was conducted by first
synthesizing the modified adenosine cores 5–8, which bear an
iodine atom at position 2 (Scheme 1 and Schemes S1,S2 in the
Supporting Information). The alkyne function was then
introduced with the Sonogashira reaction[16] to yield 9–12,
which were subsequently converted into the monophosphates
[*] Y. Wang, D. Rçsner, M. Grzywa, Prof. Dr. A. Marx
Fachbereich Chemie, Universitꢀt Konstanz
Universitꢀtsstrasse 10, 78457 Konstanz (Germany)
E-mail: andreas.marx@uni-konstanz.de
Homepage: http://www.uni-konstanz.de/chemie/~agmarx/
[**] We acknowledge partial funding by the DFG within SPP 1623 and
continuous discussions with A. Bꢁrkle, Konstanz.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201404431.
Angew. Chem. Int. Ed. 2014, 53, 8159 –8162
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
8159

.
Angewandte
Communications
Scheme 1. a) 1) CuI, trimethylsilylacetylene, Pd(PPh3)2Cl2, Et3N, DMF,
RT, 16 h; 2) NH3/MeOH, RT, 1.5 h, 45% for 9 in 2 steps, 80% for 10
in 2 steps, 81% for 11 in 2 steps, 82% for 12 in 2 steps. b) proton
sponge, POCl3, trimethyl phosphate, 08C, 5 h, 40% for 13, 47% for 14,
44% for 15, 61% for 16. c) b-NAM, CDI, triethylamine, DMF, RT. 96 h,
20% for 1, 23% for 2, 25% for 3 and 4. b-NAM=b-Nicotinamide
monophosphate, CDI=N,N’-Carbonyldiimidazole.
Figure 2. Modification of histone H1.2 and ARTD1 with the NAD+
analogues. The top panel shows the sulfo-Cy5 fluorescence and the
bottom panel shows the same gel stained with Coomassie Blue. Lanes
1 and 1’: negative control (ADP-ribosylation without oligonucleotide
duplex; for further experiments see Figure S2); lanes 2 and 2’: positive
control (ADP-ribosylation with natural NAD+); lanes 3 and 3’: ADP-
ribosylation with 1; lanes 4 and 4’: with 2; lanes 5 and 5’: with 3; lanes
6 and 6’: with 4. The total concentration of all NAD+ analogues was
1 mm.
13–16 through phosphorylation[17] (Scheme 1). The respective
monophosphates were subsequently converted into the
NAD+ analogues 1–4 by reaction with activated b-nicotin-
amide-monophosphate (Scheme 1).[18]
With NAD+ analogues 1–4 in hand, we tested them in
in vitro ADP-ribosylation assays with ARTD1, the best
studied ARTD family member. During ADP-ribosylation,
ARTD1 serves as its own acceptor[12] in a process known as
auto(ADP-ribos)ylation, as well as modifying acceptor pro-
teins in a trans(ADP-ribos)ylation process. We thus inves-
tigated the substrate scope of ARTD1 with regard to accept-
ing 1–4 in auto(ADP-ribos)ylation (Figure S5 in the Support-
ing Information) and trans(ADP-ribos)ylation (Figure 2)
reactions. In the trans(ADP-ribos)ylation assay, ARTD1
and histone H1.2, which is the main acceptor of ADP-
ribose[19] and is modified by ARTD1 and ARTD3 in vitro,[20]
were incubated with each of the NAD+ analogues in the
presence of an octameric oligonucleotide duplex that acti-
vates ARTD1.[21] Reaction with natural NAD+ served as
a positive control and for the negative control, the reaction
was performed in the absence of the activating oligonucleo-
tide duplex. After the enzymatic reaction, click chemistry was
performed to conjugate a fluorescent dye, sulfo-Cy5-azide
(Figure S1), to the alkyne-modified ADP-ribose units derived
from the incorporation of the modified NAD+ analogues.
After removing unreacted sulfo-Cy5-azide, the reaction
mixtures were analyzed by SDS-PAGE and the fluorescent
labeling was detected and compared with the gel stained with
Coomassie Blue.
undergo extensive poly(ADP-ribos)ylation. Differences in
the composition of the attached PAR led to variable migra-
tion propensity for the H1.2- and ARTD1-derived products
(Figure 2, lanes 3 and 3’) in PAGE analysis. Interestingly, we
found similar performance in the detection of PAR by using
1 and click chemistry, and conventional Western blots with an
antibody (Figure S3). For 2, we found that the absence of the
3’’-OH group influences the formation of PAR (Figure 2, lane
4). Although this analogue was clearly incorporated (since the
product could be stained by click reaction with a dye), the
automodification product pattern was altered compared to
the one obtained when 1 was used. This result suggests an
unexpected participation of the 3’’-OH group in ARTD1
catalysis. In contrast to the results obtained with 1 and 2, the
use of 3 and 4 prevents the formation of long PAR chains;
these analogues thus act as chain terminators, as shown by the
distinct fluorescent bands of labeled H1.2 and ARTD1
observed with 3 and 4 (Figure 2, lanes 5 and 6). Interestingly,
when the 3’’-OH group was provided with 3, PAR assembly
could not be rescued and a similar modification pattern to
that seen with the dideoxy analogue 4 (Figure 2, lanes 5, 6)
was observed. These findings confirm that the 2’’-OH group is
essential for PAR assembly by ARTD1.
As shown in Figure 2, all NAD+ analogues are accepted
for the trans(ADP-ribos)ylation of H1.2 by ARTD1
(Figure 2, lanes 3–6), with 1 and 2 showing the strongest
fluorescent signals owing to the formation of longer PAR
chains and the incorporation of multiple alkyne functional-
ities (Figure 2, lanes 3 and 4). With 1, both, ARTD1 and H1.2
Encouraged by these results, we investigated whether our
approach enables biotin labeling of the (ADP-ribos)ylated
products. Indeed, we were able to covalently connect biotin to
the reaction products that were obtained through the use of
the modified NAD+ analogues 1–4 (see Figure S4). Interest-
8160 www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 8159 –8162

Angewandte
Chemie
ingly, similar results were obtained when auto(ADP-ribos)-
ylation of ARTD1 was investigated (see Figure S5).
adenosine, we explored the substrate scope of ARTD1 in
terms of its dependence of the presence of a 2’’- and 3’’-OH
group in the assembly of poly(ADP-ribose). Whereas 1 led to
the formation of long PAR, 3 and 4 act as chain terminators
when applied in auto- or trans(ADP-ribos)ylation reactions.
This confirms the need for the 2’’-OH for PAR formation and
shows that the absence of this group cannot be rescued by the
presence of a 3’’-OH. Furthermore it was found that the 3’’-
OH group participates in the ADP-ribosylation of ARTD1
since the employment of 3 leads to less efficient PAR
formation as compared to the reactions where an analogue
with both 2’’- and 3’’-OH groups (1) was employed. We
further found that the modifications carried by 4 turn this
component into an efficient chain terminator of PAR syn-
thesis that competes with natural NAD+ in ADP-ribosylation
reactions. This study provides insight into the substrate scope
of ARTD1 and its catalytic mechanism for auto- and trans-
(ADP-ribos)ylation. Furthermore, the herein described
NAD+ analogues 1–4 add to the currently limited toolbox
of NAD+ analogues for studying ADP-ribosylation processes.
The chain-terminating analogues should facilitate the iden-
tification and analysis of protein targets of ADP-ribosylation
in proteomics approaches, since first results of labeling
histone H1.2 in the context of a cell lysate are very promising
(see Figure S7), and should thus spur progress towards the
development of new chemical tools to elucidate the metab-
olism of PAR.
To determine whether the NAD+ analogues are compet-
itive substrates for ARTD1-catalyzed ADP-ribosylation, we
tested 4 in a reaction competing with natural NAD+. For this
experiment, the enzymatic reaction was carried out as
described above, but with different ratios of natural NAD+
to 4. The results clearly show that 4 is incorporated by
ARTD1 in the presence of natural NAD+ (Figure 3). Even
Received: April 17, 2014
Published online: July 2, 2014
Keywords: ADP-ribosylation · ARTD1 · NAD+ ·
poly(ADP-ribosy)lation · posttranslational modifications
Figure 3. Labeling of ARTD1 and H1.2 with 4. The top panel shows the
sulfo-Cy5 fluorescence and the bottom panel shows the same gel
stained with Coomassie Blue. Lanes 1 and 1’: negative control (trans-
(ADP-ribos)ylation without oligonucleotide duplex); lanes 2 and 2’:
positive control (trans(ADP-ribos)ylation with 100% natural NAD+);
lanes 3 and 3’: trans(ADP-ribos)ylation with 75% natural NAD+ and
25% 4; lanes 4 and 4’: with 50% natural NAD+ and 50% 4; lanes 5
and 5’: with 25% natural NAD+ and 75% 4; lanes 6 and 6’: 100% 4.
The total concentration of all NAD+ analogues was 1 mm.
.
[1] P. Chambon, J. D. Weill, P. Mandel, Biochem. Biophys. Res.
Commun. 1963, 11, 39.
[2] B. A. Gibson, W. L. Kraus, Nat. Rev. Mol. Cell Biol. 2012, 13,
411.
[3] H. L. Ko, E. C. Ren, Biomolecules 2012, 2, 524.
[4] M. Christmann, M. T. Tomicic, W. P. Roos, B. Kaina, Toxicology
2003, 193, 3.
with only 25% of 4, a fluorescent signal corresponding to the
incorporation of the alkyne functionality was detected after
the click reaction with sulfo-Cy5-azide (Figure 3, lanes 3–6).
Interestingly, ADP-ribosylation of H1.2 by processing with 4
appears to take place at all employed concentrations of 4. By
contrast, ARTD1 ADP-ribosylation was only observed when
4 was exclusively employed. This indicates that 4 is more
efficiently used by ARTD1 for the trans(ADP-ribos)ylation
reaction of histone H1.2 than for auto(ADP-ribos)ylation
(Figure 3, lanes 3–6). We found that analogues 1–3 are also
incorporated by ARTD1 in the presence of natural NAD+
(see Figure S6).
[5] A. Bꢀrkle, BioEssays 2001, 23, 795.
[6] P. O. Hassa, M. O. Hottiger, Cell. Mol. Life Sci. 2002, 59, 1534.
[7] P. O. Hassa, C. Buerki, C. Lombardi, R. Imhof, M. O. Hottiger, J.
Biol. Chem. 2003, 278, 45145.
[8] M. O. Hottiger, P. O. Hassa, B. Lꢀscher, H. Schꢀler, F. Koch-
Nolte, Trends Biochem. Sci. 2010, 35, 208.
[9] D. Cervantes-Laurean, D. E. Minte, E. L. Jacobson, M. K.
Jacobson, Biochemistry 1993, 32, 1528.
[10] A. Bꢀrkle, FEBS J. 2005, 272, 4576.
[11] a) V. Schreiber, F. Dantzer, J. C. Ame, G. de Murcia, Nat. Rev.
Mol. Cell Biol. 2006, 7, 517; b) S. Messner, M. O. Hottiger,
Trends Cell Biol. 2011, 21, 534.
[12] a) Z. Tao, P. Gao, H.-w. Liu, J. Am. Chem. Soc. 2009, 131, 14258;
b) I. Carter-OꢁConnell, H. Jin, R. K. Morgan, L. L. David, M. S.
Cohen, J. Am. Chem. Soc. 2014, 136, 5201.
[13] a) J. D. Chapman, J.-P. Gagnꢂ, G. G. Poirier, D. R. Goodlett, J.
Proteome Res. 2013, 12, 1868; b) Y. j. Zhang, J. q. Wang, M.
Ding, Y. h. Yu, Nat. Methods 2013, 10, 981; c) S. Jungmichel, F.
Rosenthal, M. Altmeyer, J. Lukas, M. O. Hottiger, M. L. Nielsen,
Mol. Cell 2013, 52, 272.
In Summary, we report the design and syntheses of four
novel NAD+ analogues, which were modified at the purine
ring and in the number of hydroxy groups on the adenosine.
All of the modified NAD+ analogues (1–4) were applied in
enzymatic studies and found to be efficient substrates for
ARTD1 and were used to label ADP-ribosylated ARTD1 and
H1.2. By the systematic modification of the hydroxy groups of
Angew. Chem. Int. Ed. 2014, 53, 8159 –8162
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
8161

.
Angewandte
Communications
[14] H. Jiang, J. H. Kim, K. M. Frizzell, W. L. Kraus, H. Lin, J. Am.
Chem. Soc. 2010, 132, 9363.
[18] C. Moreau, G. K. Wagner, K. Weber, A. H. Guse, B. V. L. Potter,
J. Med. Chem. 2006, 49, 5162.
[19] A. Huletsky, G. de Murcia, S. Muller, M. Hengartner, L.
Mꢂnard, D. Lamarre, G. G. Poirier, J. Biol. Chem. 1989, 264,
8878.
[20] a) H. Okazaki, C. Niedergang, P. Mandel, Biochimie 1980, 62,
147; b) S. L. Rulten, A. E. Fisher, I. Robert, M. C. Zuma, M.
Rouleau, L. Ju, G. Poirier, B. Reina-San-Martin, K. W. Calde-
cott, Mol. Cell 2011, 41, 33.
[15] a) H. Mendoza-Alvarez, R. Alvarez-Gonzalez, Biochemistry
1999, 38, 3948; b) H. Mendoza-Alvarez, R. Alvarez-Gonzalez,
J. Mol. Biol. 2004, 336, 105; c) R. Alvarez-Gonzalez, J. Biol.
Chem. 1988, 263, 17690; d) R. Alvarez-Gonzalez, J. Moss, C.
Niedergang, F. R. Althaus, Biochemistry 1988, 27, 5378.
[16] K. Sonogashira, Y. Tohda, N. Hagihara, Tetrahedron Lett. 1975,
16, 4467.
[21] a) J. Fahrer, R. Kranaster, M. Altmeyer, A. Marx, A. Bꢀrkle,
Nucleic Acids Res. 2007, 35, e143; b) O. Popp, S. Veith, J. Fahrer,
V. A. Bohr, A. Bꢀrkle, A. Mangerich, ACS Chem. Biol. 2013, 8,
179.
[17] a) T. Kovꢃcs, L. ꢄtvçs, Tetrahedron Lett. 1988, 29, 4525; b) M.
Yoshikawa, T. Kato, T. Takenishi, Tetrahedron Lett. 1967, 8,
5065.
8162
www.angewandte.org
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 8159 –8162