Molecular Tools for the Study of ADP-Ribosylation: A Unified and Versatile Method to Synthesise Native Mono-ADP-Ribosylated Peptides

Article abstract of DOI:10.1002/chem.202100337

ADP-ribosylation (ADPr), as a post-translational modification, plays a crucial role in DNA-repair, immunity and many other cellular and physiological processes. Serine is the main acceptor for ADPr in DNA damage response, whereas the physiological impact of less common ADPr-modifications of cysteine and threonine side chains is less clear. Generally, gaining molecular insights into ADPr recognition and turn-over is hampered by the availability of homogeneous, ADP-ribosylated material, such as mono-ADP-ribosylated (MARylated) peptides. Here, a new and efficient solid-phase strategy for the synthesis of Ser-, Thr- and Cys-MARylated peptides is described. ADP-ribosylated cysteine, apart from being a native post-translational modification in its own right, proved to be suitable as a stabile bioisostere for ADP-ribosylated serine making it a useful tool to further biochemical research on serine ADP-ribosylation. In addition, it was discovered that the Streptococcus pyogenes encoded protein, SpyMacroD, acts as a Cys-(ADP-ribosyl) hydrolase.

Full text of DOI:10.1002/chem.202100337

Full Paper
doi.org/10.1002/chem.202100337
Chemistry—A European Journal
Molecular Tools for the Study of ADP-Ribosylation:
A Unified and Versatile Method to Synthesise Native Mono-
ADP-Ribosylated Peptides
+
[a]
+ [b]
[a]
Jim Voorneveld , Johannes Gregor Matthias Rack , Luke van Gijlswijk,
[a]
[a]
[a]
[a]
Nico J. Meeuwenoord, Qiang Liu, Herman S. Overkleeft, Gijsbert A. van der Marel,
Ivan Ahel,* and Dmitri V. Filippov*
[b]
[a]
Abstract: ADP-ribosylation (ADPr), as a post-translational
modification, plays a crucial role in DNA-repair, immunity and
many other cellular and physiological processes. Serine is the
main acceptor for ADPr in DNA damage response, whereas
the physiological impact of less common ADPr-modifications
of cysteine and threonine side chains is less clear. Generally,
gaining molecular insights into ADPr recognition and turn-
over is hampered by the availability of homogeneous, ADP-
ribosylated material, such as mono-ADP-ribosylated (MARy-
lated) peptides. Here, a new and efficient solid-phase strategy
for the synthesis of Ser-, Thr- and Cys-MARylated peptides is
described. ADP-ribosylated cysteine, apart from being a
native post-translational modification in its own right, proved
to be suitable as a stabile bioisostere for ADP-ribosylated
serine making it a useful tool to further biochemical research
on serine ADP-ribosylation. In addition, it was discovered that
the Streptococcus pyogenes encoded protein, SpyMacroD, acts
as a Cys-(ADP-ribosyl) hydrolase.
[
13–15]
Introduction
complex)
and removed (by (ADP-ribosyl)hydrolase
3
[
16]
(
ARH3)). Cysteine ADPr has been shown to be synthesized by
[17]
ADP-ribosylation is a complex post-translational modification
PARP7 (TiPARP) and PARP8, but hydrolases for erasing this
particular PTM remain unidentified to date.
[7,18]
(PTM) that partakes in a wide variety of cellular and physio-
Herein we
logical biology, including DNA damage response (DDR) and
immune-related processes. ADP-ribosylation occurs by transfer
describe the development of a new strategy to synthesize ADP-
ribosylated peptides, modified on the side chains of serine (Ser),
threonine (Thr) and cysteine (Cys) residues. We reasoned that
+
of ADP-ribose from NAD onto one of multiple types of amino
[
1–4]
acids including ones featuring a carboxylic acid (Glu/Asp),
an
ADPr-Cys containing peptides, besides emulating natural ADP-
[5,6]
[7–11]
[6,7,18]
alcohol (Ser/Tyr),
a thiol (Cys) or a guanidine (Arg).
The
ribosylated proteins,
may also serve as chemically stable
ADPr modification of amino acid side chains of proteins is
catalysed by members of the (ADP-ribosyl)transferase (ART)
superfamily, including poly(ADP-ribose) polymerases (PARPs).
ADP-ribosylated Ser linkages, possibly less susceptible to ARH3-
catalyzed hydrolysis. Thr-ADPr is a less common PTM and until
now, no family member of mammalian PARP has been
identified as being able to transfer ADPr onto threonine.
Reports have emerged however that some bacterial enzymes
are able to ADP-ribosylate threonine residues, for example, on
human ubiquitin (Ub). This modification inhibits the function
of PolyUb on multiple levels, such as biosynthesis, Ub
recognition and the reversal of the modification, and plays a
crucial role in bacterial colonization.
[10]
In the past few years, many insights have been acquired
regarding the biological role of ADP-ribosylated serine (Ser-
[8]
[12]
ADPr), its chemical stability,
and the distinctly different
[19]
mechanisms by which Ser-ADPr is introduced (by PARP1:HPF1
+
[
a] J. Voorneveld, L. van Gijlswijk, N. J. Meeuwenoord, Dr. Q. Liu,
Prof. H. S. Overkleeft, Prof. G. A. van der Marel, Dr. D. V. Filippov
Leiden Institute of Chemistry, Leiden University
Department of Bioorganic Synthesis,
In 2016, we developed the first solid phase synthesis of
ADP-ribosylated peptides, in which a phosphoribosylated
building block was used for the introduction of ribosylated
Einsteinweg 55, 2333CC, Leiden (The Netherlands)
E-mail: filippov@chem.leidenuniv.nl
[20]
amino acids in a peptide sequence. The ensuing installation
of the pyrophosphate moiety forced us to use acid labile tBu-
protection of the phosphotriester to attain orthogonality. After
removal of the tBu-groups, the ADP moiety was introduced via
+
[
b] Dr. J. G. M. Rack, Prof. I. Ahel
Sir William Dunn School of Pathology, University of Oxford
South Parks Road Oxford OX1 3RE (United Kingdom)
E-mail: ivan.ahel@path.ox.ac.uk
+
III
V
[21]
[
] These authors contributed equally to this work.
our established P À P coupling method with an adenosine
Supporting information for this article is available on the WWW under
https://doi.org/10.1002/chem.202100337
phosphoramidite. Although a variety of sequences, MARylated
[20]
[22]
on Gln, Asn or Cit sites as well as Ser have been prepared
using this method, inherent drawbacks of the methodology
include the presence of a carboxamide at the C-terminus of the
synthetic MARylated peptide and the extensive amount of
©
2021 The Authors. Chemistry - A European Journal published by Wiley-
VCH GmbH. This is an open access article under the terms of the Creative
Commons Attribution License, which permits use, distribution and re-
production in any medium, provided the original work is properly cited.
Chem. Eur. J. 2021, 27, 1–8
1
© 2021 The Authors. Chemistry - A European Journal published by Wiley-VCH GmbH
��
These are not the final page numbers!

Full Paper
doi.org/10.1002/chem.202100337
Chemistry—A European Journal
[
24]
protective group manipulations that is needed for some
peptide sequences. For instance, when the ADPr-modification
site in the target peptide is flanked by Ser or Thr at the C-
terminal or N-terminal sequence, the side chain protecting trityl
which proved useful for orthogonal protection in ribosides
[25]
and pyranosides, enabling α-selective glycosylation and also
cleavage in the final stage of the immobilized MARylated
peptides, and II) the bulky TBDPS group which not only
enhances α-selectivity during glycosylation, but also provides
orthogonality for the introduction of the phosphotriester, the
first step in the installation of the ADP-moiety. For phospho-
triester introduction, reagent 20 with fluorenylmethyl (Fm)
(Trt) groups must be replaced by an acetyl to allow further
processing of the tBu-protected phosphotriester.
The here-presented new, generally applicable strategy
toward MARylated peptides overcomes these disadvantages. In
the design of our new method we decided to again employ
Fmoc-based solid-phase peptide synthesis (SPPS) that is com-
patible with most peptide synthesizers and also to our
procedure for the introduction of the pyrophosphate moiety.
[26]
groups and reagent 21 equipped with 2-methylsulfonylethyl
[27,28]
(Mse) groups
were chosen (Scheme 4), as both of these can
be cleaved in an orthogonal fashion by treatment with DBU.
With these changes of the solid phase procedure and the
protecting group pattern of the newly designed ribosylated
amino acid building blocks, adenosine amidite 6 was chosen to
complete the ADPr installation, after which global deprotection
by subsequent treatment with DBU, TBAF and TFA would
furnish the MARylated peptides. We have utilized this new SPPS
approach to assemble a selection of biologically relevant
MARylated peptides. With these, we investigated the influence
of the ADP-ribosylated amino acid on signal turn-over by a
variety of human and microbial (ADP-ribosyl)hydrolases. We
found that the Ser>Thr exchange slowed the hydrolysis
reaction by ARH3, whereas Ser>Cys replacement abolished the
reaction thus indicating that the Cys-MARylated peptide is a
useful stabilized bioisostere for the study of human proteins. As
well, we found that SpyMacroD, a macrodomain-type enzyme
®
We selected the TentaGel resin equipped with the highly acid
[23]
sensitive S AC linker (Scheme 1), because this would return
oligopeptides with a C-terminal carboxylic acid, thus to ADPr-
peptides more closely resembling natural sequences, such as
those that would emerge from proteolytic processing of ADP-
ribosylated proteins. The acid-sensitive Trt and 4-methyltrityl
(Mtt) groups were chosen for protection of the side chains of
Ser-, Thr- and Lys-residues. We decided to postpone the
introduction of the phosphotriester to the final stage of the
SPPS leading to the design of ribosylated amino acid building
blocks 1–3 (Scheme 1) for incorporation of the prospected ADPr
moiety in the sequence. The building blocks are compatible
with standard Fmoc-based SPPS and are endowed with the
following protecting groups on the ribosyl moiety: I) the 4-
methoxybenzyl (PMB) groups on the secondary hydroxyls,
Scheme 1. Synthetic strategy for the preparation of peptides MARylated on their Ser-, Thr- or Cys-residue. P.G.=protecting group for amino acid side chains:
Trt for serine/threonine/histidine, Mtt for lysine, 2-PhiPr for glutamic acid, bis-Alloc for arginine.
Chem. Eur. J. 2021, 27, 1–8
www.chemeurj.org
2
© 2021 The Authors. Chemistry - A European Journal published by Wiley-VCH GmbH
��
These are not the final page numbers!

Full Paper
doi.org/10.1002/chem.202100337
Chemistry—A European Journal
from Streptococcus pyogenes, efficiently removed the Cys
modification.
concentration and nature of the activator, the results of which
are listed in Table 1. Coupling of donor 9 with acceptor 13
using the same conditions as described before in a similar
[12]
reaction gave a low yield of the wanted ribofuranosylated
Results and Discussion
Fmoc-Ser 12 together with a significant amount of side-product
1
1, originating from the Chapman-type rearrangement,
[32,33]
The glycosylation procedure toward the ribofuranosylated
Fmoc-amino acids 1–3 was first optimized by testing two
ribosyl donors; the known N-(phenyl)trifluoroacetimidate donor
(entry 1).
Changing the activator to TfOH (entry 2) led to
acid-catalyzed cleavage of one or more PMB-protecting groups
and the PMB-cation was scavenged by the acceptor, resulting in
Fmoc-Ser(PMB)-OAll together with a complex mixture of ribose
derived products. In an attempt to reduce the loss of the PMB
group, TBSOTf was used as an activator (entry 3). It was
reasoned that the softer character of the TBS cation would
decrease the acidity of the glycosylation conditions thus
diminishing the cleavage of the PMB groups. Although the use
of TBSOTf significantly improved the yield of product 12, side
product 11 still occurred in a 12% yield. A two-fold increase in
the concentration of donor 9 unexpectedly enlarged the
formation of side product 11 to 48% (entry 4). Finally,
increasing the temperature of the reaction mixture to À 40°C
[29]
[30]
1
0
and its trichloroacetimidate congener 9 (Table 1, see
Supporting Information for preparation). The latter was chosen
for its synthetic accessibility as donor 10 requires the labour-
intensive synthesis of 2,2,2-trifluoro-N-phenyl-acetimidoyl
chloride as a reagent. On the other hand, trichloroacetimidate
donors can undergo a Chapman-like rearrangement to form the
[31,32]
unreactive glycosylamides under glycosylation conditions.
As a model reaction for the glycosylation, the condensation of
serine acceptor 13 (Scheme 2, see Supporting Information for
preparation) with ribosyl donors 9 and 10 was examined by
varying the reaction conditions in terms of temperature,
did improve the yield of the glycosylation to 60% (entry 5). For
donor 10, the same set of reaction conditions were tested
(entries 6–10). It is of interest to note that using TfOH as
Table 1. Optimization of the glycosylation conditions with donors 9 and
activator with donor 10, the amount of acid-catalyzed PMB
ether cleavage has been significantly reduced in comparison
with trichloroacetimidate counterpart 9. However, the results of
the glycosylations proved to be more reproducible with TBSOTf,
particularly for scaling up the reaction.
1
0 and acceptor 13. The concentration (C) is the concentration of the
donor in DCM as solvent. Activators were used in 0.1 equivalent relative to
the donor and the reactions were carried out at 0.2 mmol scale.
Having optimized the reaction conditions (entry 8, Table 1)
[
34,35]
we glycosylated appropriately protected acceptors 13 (Ser),
[34]
1
4 (Thr)
preparation) with ribosyl donor 10.
furnished the suitably protected, ribofuranosylated amino acids
2, 16 and 17 in high α-selectivity (no β-product was observed)
and 15 (Cys) (See Supporting Information for
[29]
This transformation
entry
donor
C (M)
T (°C)
activator
11 (%)
12 (%)
1
1
2
3
4
5
6
7
8
9
9
9
9
9
0.1
0.1
0.1
0.2
0.1
0.1
0.1
0.1
0.2
0.1
À 50
À 50
À 50
À 50
À 40
À 50
À 50
À 50
À 50
À 40
TMSOTf
TfOH
TBSOTf
TBSOTf
TBSOTf
TMSOTf
TfOH
TBSOTf
TBSOTf
TBSOTf
34
n.d.
12
48
12
–
–
–
–
–
23
0
and good yields (56–70%, Scheme 2). For protection of the C-
terminus, the allyl ester was chosen since it can be selectively
39
17
60
32
53
56
23
48
removed by treatment with catalytic Pd(PPh ) under neutral
9
3 4
[36]
10
10
10
10
10
conditions. Treatment of amino acids 12, 16 and 17 with
Pd(PPh ) and 1,3-dimethylbarbituric acid (DMBA) as the allyl
3
4
[37]
cation scavenger furnished the required building blocks 1, 2
and 3 in good yields in a few steps.
1
0
The final building block, needed for the assembly of
MARylated peptides is adenosine phosphoramidite
6
(
Scheme 3) that enables the introduction of the ADPr moiety
III
V
[38]
via the P À P procedure. Silylation of the hydroxyl functions
in adenosine with TBSÀ Cl was followed by protecting the
Scheme 3. Synthesis of adenosine amidite 6. Reagents and conditions: I)
Scheme 2. Synthesis of Fmoc-based SPPS building blocks 1–3. Reagents and
TBSÀ Cl, imidazole, DMF, 50°C. II) Boc
2 2
O, DMAP, THF, reflux. III) TFA, H O,
conditions: I) TBSOTf, acceptors 13, 14 or 15, DCM, À 50°C. II) Pd(PPh
3 4
) ,
THF, 0°C. IV) 2-cyanoethyl-N,N-diisopropylchlorophosphoramidite, DIPEA,
DMBA, DCM. All=allyl.
DCM, rt.
Chem. Eur. J. 2021, 27, 1–8
www.chemeurj.org
3
© 2021 The Authors. Chemistry - A European Journal published by Wiley-VCH GmbH
��
These are not the final page numbers!

Full Paper
doi.org/10.1002/chem.202100337
Chemistry—A European Journal
exocyclic amine with a Boc group. The ensuing acid mediated,
and Mse protecting groups were removed by treatment of the
resins with 10% DBU in DMF. Monitoring the reaction progress
for 20 min by LC-MS showed that both Fm protecting groups
were completely eliminated under these conditions, whereas
only one Mse-group had been removed, leading to the
phosphodiester. Although the crystalline Mse reagent 21 is
easier to handle, the Fm protecting group was chosen for
further synthetic studies for its more efficient deprotection.
Thus, the assembly of the MARylated peptide was continued
with the deprotection of 22-Fm. Condensation of the resulting
phosphate monoester with adenosine phosphoramidite 6 and
oxidation of the resultant P(III)-P(V) intermediate gave immobi-
lized peptide 23, containing a partially protected pyrophos-
phate moiety. The cyanoethyl group was then removed from
the pyrophosphate in 23 by 10% DBU in DMF, after which the
silyl ethers were deprotected with TBAF and the remaining
protective groups finally removed with concomitant cleavage of
the target MARylated peptide from the resin by treatment with
10% TFA solution in DCM containing 2.5% TIS as a scavenger.
Monitoring of the latter deprotection by LC-MS analysis
revealed that the trityl and PMB-protecting groups were split
off instantly while the more stable Boc protective group on the
exocyclic amine of adenosine needed at least 2 h for its
removal. Purification with RP-HPLC of the obtained crude
product led to the isolation of homogeneous, MARylated
peptide 24, derived from the N-terminus histone H2B in 3.6%
overall yield.
[39]
regioselective desilylation
of the primary alcohol led to
partially protected adenosine 18, which was phosphitylated to
give phosphoramidite 6.
With all the required building blocks in hand, the solid
phase assembly of MARylated peptide 24, derived from the N-
terminus of human histone H2B, was undertaken (Scheme 4).
Standard Fmoc solid phase peptide synthesis methodology was
employed using amino acid building blocks having highly acid
sensitive side chain protecting groups (Mtt for lysine, Trt for
serine). As depicted in Scheme 4, Tentagel® S AC resin
preloaded with glycine was elongated using the selected
protected amino acid building blocks, including ribofuranosy-
lated Fmoc-Ser-OH 1 to give immobilized peptide 19. The
ensuing cleavage of the TBDPS-protecting group was tested
À
using three different F sources: TEA·3HF, HF·pyridine and
TBAF. Both deprotections using TEA·3HF and HF·pyridine
needed reaction times of up to 16 h to fully remove the silyl
protecting group whereas employing a 1 M·TBAF solution in
THF ensured full deprotection in 30 min. The TBAF treatment
was not only superior with regard to cleavage time but also in
the quality of the product according to LC-MS analysis of the
peptides after desilylation, removal of all other protecting
groups and cleavage from the solid support. After desilylation
of the primary hydroxyl in the ribose moiety, two phosphor-
amidite reagents were investigated for obtaining the ribosyl-5-
phosphomonoester. Two phosphoramidites, both bearing base
sensitive protecting groups, were tested: reagent 20 bearing
fluorenylmethyl (Fm) groups and reagent 21 equipped with 2-
methylsulfonylethyl (Mse) groups. The phosphitylation of the
immobilized peptide, having a ribose with 5-OH by either the
reagent 20 or 21, followed by the CSO mediated oxidation of
the formed phosphite to the phosphotriester intermediate
provided the immobilized, fully protected phosphoribosyl
peptides 22-Fm and 22-Mse, respectively. To convert the
phosphotriesters in these into the phosphomonoester, the Fm
Following the above-described general procedure, we next
set out to assemble Ser-ADPr oligopeptides 25 and 26 (Table 2),
both sequences that are confirmed to exist in nature bearing
this PTM as established by LC-MS/MS analysis of biological
[40,41]
samples.
The synthesis of peptide 25 proceeded uneventful
where 25 was obtained in 11% after HPLC-purification. En route
to peptide 26, incorporation of Thr-7 proved to be difficult and
required us to repeat the solid-phase condensation step
employing Fmoc-Thr(Trt)-OH twice. Peptide 26 also contains an
Scheme 4. Synthesis of MARylated H2B peptide 24.
Chem. Eur. J. 2021, 27, 1–8
www.chemeurj.org
4
© 2021 The Authors. Chemistry - A European Journal published by Wiley-VCH GmbH
��
These are not the final page numbers!

Full Paper
doi.org/10.1002/chem.202100337
Chemistry—A European Journal
Table 2. Synthesis of MARylated peptides 24–30. The amino acids indicated by bold print are the modification sites.
Number
Sequence
Yield (%)
Notes
ADPr
24
25
26
27
28
29
30
Ac-PAKS APAPKKG-OH
3.6
11
n.a.
n.a.
ADPr
Ac-GKS GAALSKKG-OH
ADPr
Ac-GKS SGPTSLFAVTVAPPGARG-OH
6.1
9.5
0.94
4.1
1.9
Strenuous coupling of Thr-7
n.a.
n.a.
EDT added in cleavage cocktail
tBuOOH was used instead of CSO
ADPr
Ac-GKSSGPT SLF-OH
ADPr
Ac-KEST LHLVLRL-OH
ADPr
Ac-PAKC APAPKKG-OH
ADPr
biotin-PAKC APAPKKG-OH
arginine, for whose side chain no suitable acid labile protecting
group is available, which led us to employ bis-Alloc protection
alkylsulfide. Since such overoxidation has not been detected
during the synthesis of the similar Cys-ADPr peptide 29, it is
postulated that this unwanted reaction has occurred on the
sulfur of the biotin tag. Oxidized biotin occurs as both α- and β-
sulfoxide and while the β-form nearly completely ablates its
affinity towards avidin the α-biotin sulfoxide still has strong
[20,42]
for the guanidine function.
This necessitated adaptation of
the deprotection procedure at the final stage of the synthesis of
Ser-ADPr peptide 26 by subjecting the resin to Pd(PPh ) and
3
4
DMBA as a scavenger prior to treatment with the acidic
cleavage cocktail. In this way and after HPLC purification
peptide 26 was obtained in 6.1% overall yield.
Having effectively completed the Ser-ADPr peptides 24–26,
our attention was turned to the assembly of ADPr peptides
with Thr or Cys at the site of ADP-ribosylation. Thr-ADPr peptide
[44]
binding properties
and can be used without loss of
[45]
streptavidin pull-down efficiency. Still, the synthesis of ADPr-
peptide 30 was repeated, using a milder oxidizing agent than
CSO for the oxidation of P(III)À P(V) precursor of the pyrophos-
phate. Indeed, the application of 0.5 M tBuOOH for 30 min
proved effective in the chemoselective oxidation of the
phosphite species whilst leaving the biotin moiety intact.
2
7, which is sharing part of the sequence of peptide 26, has
been selected to help determine the exact site of modification
as its identification by MS/MS of proteomic mixtures is not
Having obtained MARylated peptides 24–30, we set out to
investigate the differences in enzymatic turn-over of these
modifications. As ARH3 is the only known hydrolase of ADP-
[43]
always conclusive. The aforementioned difficulties incorporat-
ing the Thr-7 residue in 26 were not encountered in coupling
ribosylated Thr-building block 2 to obtain peptide 27. Another
relevant Thr-ADPr peptide is 28, a fragment containing the
ADP-ribosylation site in human Ub that is modified at Thr-66 by
the bacterial effector protein CteC, as detected by LC-MS/MS
[16]
ribosylated serine residues,
we first tested its ability to
hydrolytically remove the ADPr moiety from these peptides
(Figure 1a). We found that ARH3, but not its catalytic mutants
D77 N or D78 N, is capable of hydrolysing the glycosidic linkage
in 24 (Ser) and 27 (Thr). The turnover of the latter proved not as
efficient as the former (Figure 1a and b), which might be a
result of steric clash of the additional methyl group of the
threonine side chain within the enzyme active site. Please note
that we employed 45 min incubation, where we have demon-
[19]
analysis in proteomics studies. This ADPr peptide includes the
amino acids Glu and His and successful construction of this
sequence would mean that also these amino acids, in
appropriately protected form, can be included in the synthetic
scheme. In the SPPS to 28, which was obtained in 0.94% after
HPLC purification, the building blocks Fmoc-His(Trt)-OH and
Fmoc-Glu(O-2-PhiPr)-OH were used as both protecting groups
are removable by the levels of TFA used in the cleavage
cocktail. As was mentioned before, peptides containing ADP-
ribosylated cysteine can be considered as isosteric to ADP-Ser
peptides with the ADPr moiety relatively more stabile towards
enzymatic hydrolysis. The SPPS of Cys-ADPr peptide 29, the Ser-
to-Cys analogue of 24, was performed using ribofuranosylated
Cys building block 3. After deprotection and cleavage of the
immobilized Cys-ADPr peptide 29 from the resin, Ac-PAKC(PMB)
APAPKKG-OH was detected, a side product originating from the
migration of the PMB cation to the thiol of the Cys side chain.
Addition of ethane dithiol (EDT), a more potent thiol-based
scavenger, to the cleavage cocktail suppressed this side
reaction, providing 29 in 4.1% yield and good homogeneity
[12]
strated earlier
that Ser-MARylation turn-over is complete
<20 min. In contrast, the ADPr-Cys interglycosidic linkage was
largely stable towards enzymatic hydrolysis under the con-
ditions applied. Since ADP-ribosylation of cysteine residues is a
modification found in human cells, we expanded our inves-
tigation to all known human hydrolases and confirmed that
only ARH3 could remove the modification from serine and by
extension threonine (Figure 1b). Surprisingly, none of the tested
hydrolases was able to remove the modification from peptide
26. This suggests that either the modification is irreversible in
human cells or is reversed by a thus far unidentified enzyme. To
test whether the modification could in principle be reversed,
we tested several evolutionary divergent hydrolases of the
macrodomain and (ADP-ribosyl)hydrolase family from various
lower organisms (Figure 1c). While none of the ARH-like
enzymes was able to hydrolyse this particular linkage, Strepto-
(see Supporting Information for experimental data such as LC-
[46]
MS trace). Lastly, to obtain a useful pull-down tag for biological
experiments, N-biotin-Cys-ADPr peptide 30 was assembled.
After completion of the synthesis of 30, LC-MS analysis of the
crude product revealed a main product with a mass 16 Dalton
higher than expected, presumably due to the oxidation of an
coccus pyogenes MacroD (SpyMacroD) efficiently hydrolysed
the ADP-ribosyl-cysteinyl glycosidic bond. Earlier structural
studies on SAV0325, the Staphylococcus aureus homologue of
2
+
SpyMacroD, showed a Zn -binding motif within the active site
and the authors suggested that this zinc centre participates in
Chem. Eur. J. 2021, 27, 1–8
www.chemeurj.org
5
© 2021 The Authors. Chemistry - A European Journal published by Wiley-VCH GmbH
��
These are not the final page numbers!

Full Paper
doi.org/10.1002/chem.202100337
Chemistry—A European Journal
which also led to cleavage from the resin. RP-HPLC purification
of the crude products finally furnished the targeted peptides,
MARylated at a predetermined Ser, Thr or Cys residue, in decent
yields, homogeneity and sufficient quantities which allows their
usage in biological experiments. Our methodology and the set
of peptides synthesized vary in functional side chains flanking
the ADPr site, demonstrating that a wide set of amino acids can
be incorporated in this way. As well, functionalization of the
peptide with a biotin tag is allowed, provided that oxidation of
the phosphite species is performed with tBuOOH rather than
CSO.
The availability of the Ser-ADPr, Thr-ADPr and Cys-ADPr
peptides allowed us for the first time to directly assess and
compare liability of these linkages towards enzymatic hydrol-
ysis. We found that the additional methyl group in Thr, as
compared to Ser, leads to a pronounced reduction in turn-over
by ARH3. This suggests that the additional methyl group
hinders optimal substrate arrangement within the active site
due to increased steric hindrance but no other hydrolase was
identified as being able to hydrolyse Thr-ADPr. In contrast, Ser-
to Cys-ADPr peptide 29, in which the glycosidic bond nature
differs from O to S, is almost completely resistant to ARH3-
mediated hydrolysis. Given that Cys-modifications, which occur
in human cells and are suggested to be involved in regulation
of hypoxia, immunity, coronavirus response and nuclear
Figure 1. Enzymatic hydrolysis of interglycosidic linkages in ADP-ribosylated
serine, threonine and cysteine containing peptides 24, 27 and 29,
respectively. (a–c) Measurements of hydrolase activity against the various
ADPr-peptide linkages was assessed by converting released ADPr into AMP
via NUDT5 and subsequently measured using the AMP-Glo™ assay
(
(
Promega). Samples are background corrected and normalised to ARH3 wt
a+b) or SpyMacroD (c). Data are represented as mean values � SD
measured in triplicates. (a) Hydrolysis of Ser-, Thr- and Cys-linked ADPr by
ARH3 wt and catalytic mutant D77 N. (b) Hydrolysis of Ser-, Thr- and Cys-
linked ADPr by human (ADP-ribosyl)hydrolases. (c) Hydrolysis of Cys-linked
ADPr by evolutionary diverged (ADP-ribosyl)hydrolases of the macrodomain
and ARH-like families. Abbreviations: S2-MacroD, SARS-CoV-2 nsp3 macro-
domain; HsGDAP2, Homo sapiens Ganglioside-induced differentiation-asso-
ciated protein 2 (GDAP2); TcPARG, Thermomonospora curvata PARG,
Streptomyces coelicolor (SCO) ARH-like hydrolases indicated by their
respective gene identifiers.
[6,18,48,49]
receptors,
cannot be reversed by any of the known
human hydrolases, it may be that the modification is either
irreversible or is reversed by an as yet unidentified hydrolase or
mechanism. Our discovery of SpyMacroD as a Cys-(ADP-ribosyl)
hydrolase shows that efficient and specific hydrolysis is
possible, suggesting that a human enzyme harbouring this
activity may exist as well.
[47]
substrate recognition and catalysis. Our results suggest that
2
+
an increased Lewis acidity, due to the presence of the Zn ion,
relative to other macrodomains as well as production of
cysteine, a favourable zinc-coordination ligand, are supporting
the reaction. This result clearly shows that hydrolases can Acknowledgements
readily evolve into efficient cysteine deADP-ribosylating en-
zymes and that such activity may exist in humans. Together,
our data provide new insights into the turnover of ADP-
ribosylated substrate and highlight the suitability of the
MARylated peptides 24–30 as tools for the study of ADP-
ribosylation.
Streptomyces ARH-like proteins were a kind gift from Andreja
Mikoč at the Ruđer Bošković Institute (Zagreb, Croatia). Work in
I.A.’s laboratory is supported by Wellcome Trust (101794,
210634); Biotechnology and Biological Sciences Research
Council (BB/R007195/1); and Cancer Research United Kingdom
(C35050/A22284).
Conclusion
Conflict of Interest
Herein we describe the development of a new and robust
synthetic strategy to obtain peptides modified with mono(ADP-
ribose) at a chosen serine, threonine or cysteine side chain. For
this purpose, unprecedented ribofuranosylated Ser-, Thr- and
Cys-building blocks 1, 2 and 3 were developed and used in
Fmoc-based SPPS to obtain immobilized peptides that were
decorated with an orthogonally protected ribosyl unit at the
prospected ADPr site. This ribosylated part is then functional-
ized with a phosphomonoester prior to on-resin construction of
the pyrophosphate, yielding immobilized and partially pro-
tected ADP-ribosylated peptides. The peptides were then
subjected to a deprotection sequence, the final acid step of
The authors declare no conflict of interest.
Keywords: ADP-ribosylation · (ADP-ribosyl) hydrolase · poly
(ADP-ribose) polymerases · post-translational modification ·
solid-phase peptide synthesis
[
[
1] N. Ogata, K. Ueda, H. Kagamiyama, O. Hayaishi, J. Biol. Chem. 1980, 255,
616–7620.
2] N. Ogata, K. Ueda, O. Hayaishi, J. Biol. Chem. 1980, 255, 7610–7615.
7
[3] Y. Zhang, J. Wang, M. Ding, Y. Yu, Nat. Methods 2013, 10, 981–984.
Chem. Eur. J. 2021, 27, 1–8
www.chemeurj.org
6
© 2021 The Authors. Chemistry - A European Journal published by Wiley-VCH GmbH
��
These are not the final page numbers!

Full Paper
doi.org/10.1002/chem.202100337
Chemistry—A European Journal
[
[
4] R. Bredehorst, K. Wielckens, A. Gartemann, H. Lengyel, K. Klapproth, H.
Hilz, Eur. J. Biochem. 1978, 92, 129–135.
5] O. Leidecker, J. J. Bonfiglio, T. Colby, Q. Zhang, I. Atanassov, R. Zaja, L.
Palazzo, A. Stockum, I. Ahel, I. Matic, Nat. Chem. Biol. 2016, 12, 998–
[26] Y. Watanabe, T. Nakamura, H. Mitsumoto, Tetrahedron Lett. 1997, 38,
7407–7410.
[27] H. E. Rydon, R. C. Thompson, Tetrahedron Lett. 1968, 9, 2525–2526.
[28] C. Claesen, G. I. Tesser, C. E. Dreef, J. E. Marugg, G. A. van der Marel, J. H.
van Boom, Tetrahedron Lett. 1984, 25, 1307–1310.
1
000.
[
6] S. Buch-Larsen, I. Hendriks, J. Lodge, M. Rykær, B. Furtwängler, E.
Shishkova, M. Westphall, J. Coon, M. Nielsen, Cell Rep. 2020, 32, 108176.
7] S. Vyas, I. Matic, L. Uchima, J. Rood, R. Zaja, R. T. Hay, I. Ahel, P. Chang,
Nat. Commun. 2014, 5, 4426.
[29] H. A. V. Kistemaker, G. J. Van Der Heden Van Noort, H. S. Overkleeft,
G. A. Van Der Marel, D. V. Filippov, Org. Lett. 2013, 15, 2306–2309.
[30] R. R. Schmidt, J. Michel, Angew. Chem. Int. Ed. 1980, 19, 731–732; Angew.
Chem. 1980, 92, 763–764.
[
[
[
8] M. J. Suskiewicz, L. Palazzo, R. Hughes, I. Ahel, FEBS J. 2020, 1–12.
9] B. Lüscher, M. Bütepage, L. Eckei, S. Krieg, P. Verheugd, B. H. Shilton,
Chem. Rev. 2018, 118, 1092–1136.
[31] A. W. Chapman, J. Chem. Soc. Perkin Trans. 1 1925, 127, 1992–1998.
[32] K. Larsen, C. E. Olsen, M. S. Motawia, Carbohydr. Res. 2008, 343, 383–387.
[33] A. W. Chapman, J. Chem. Soc. Perkin Trans. 1 1925, 127, 1992–1998.
[
[
[
34] M. Ciommer, H. Kunz, Synlett 1991, 593–595.
[
[
[
10] M. S. Cohen, P. Chang, Nat. Chem. Biol. 2018, 14, 236–243.
11] J. G. M. Rack, L. Palazzo, I. Ahel, Genes Dev. 2020, 34, 263–284.
12] J. Voorneveld, J. G. M. Rack, I. Ahel, H. S. Overkleeft, G. A. Van Der Marel,
D. V. Filippov, Org. Lett. 2018, 20, 4140–4143.
35] S. Rio, J. M. Beau, J. C. Jacquinet, Carbohydr. Res. 1991, 219, 71–90.
36] H. Kunz, H. Waldmann, Angew. Chem. Int. Ed. 1984, 23, 71–72; Angew.
Chem. 1984, 96, 49–50.
[
[
37] H. Kunz, J. März, Synlett 1991, 591–593.
38] H. Gold, P. Van Delft, N. Meeuwenoord, J. D. C. Codée, D. V. Filippov, G.
Eggink, H. S. Overkleeft, G. A. Van Der Marel, J. Org. Chem. 2008, 73,
[
[
[
13] I. Gibbs-Seymour, P. Fontana, J. G. M. Rack, I. Ahel, Mol. Cell 2016, 62,
4
32–442.
14] J. J. Bonfiglio, P. Fontana, Q. Zhang, T. Colby, I. Gibbs-Seymour, I.
Atanassov, E. Bartlett, R. Zaja, I. Ahel, I. Matic, Mol. Cell 2017, 0, 1–9.
15] M. J. Suskiewicz, F. Zobel, T. E. H. Ogden, P. Fontana, A. Ariza, J. C. Yang,
K. Zhu, L. Bracken, W. J. Hawthorne, D. Ahel, D. Neuhaus, I. Ahel, Nature
020, 579, 598–602.
16] P. Fontana, J. J. Bonfiglio, L. Palazzo, E. Bartlett, I. Matic, I. Ahel, Elife
017, 6, 1–20.
17] K. M. Rodriguez, S. C. Buch-Larsen, I. T. Kirby, I. Siordia, D. Hutin, M.
Rasmussen, D. M. Grant, L. L. David, J. Matthews, M. L. Nielsen, M. S.
Cohen, Elife 2021, 10, 1–35.
18] A. Gomez, C. Bindesbøll, S. V. Satheesh, G. Grimaldi, D. Hutin, L.
MacPherson, S. Ahmed, L. Tamblyn, T. Cho, H. I. Nebb, A. Moen, J. H.
Anonsen, D. M. Grant, J. Matthews, Biochem. J. 2018, 475, 3827–3846.
19] F. Yan, C. Huang, F. Yan, C. Huang, X. Wang, J. Tan, S. Cheng, M. Wan, Z.
Wang, Mol. Cell 2020, 1–12.
9
458–9460.
[
[
[
39] C. Atdjian, L. Iannazzo, E. Braud, M. Ethève-Quelquejeu, Eur. J. Org.
Chem. 2018, 2018, 4411–4425.
40] C. A. Vivelo, R. Wat, C. Agrawal, H. Y. Tee, A. K. L. Leung, Nucleic Acids
Res. 2017, 45, D204–D209.
41] V. Ayyappan, R. Wat, C. Barber, C. A. Vivelo, K. Gauch, P. Visanpattanasin,
G. Cook, C. Sazeides, A. K. L. Leung, Nucleic Acids Res. 2020, 1–5.
42] F. Eisele, J. Kuhlmann, H. Waldmann, Chem. Eur. J. 2002, 8, 3362–3376.
43] D. M. Leslie Pedrioli, M. Leutert, V. Bilan, K. Nowak, K. Gunasekera, E.
Ferrari, R. Imhof, L. Malmström, M. O. Hottiger, EMBO Rep. 2018, 19, 1–
2
[
[
2
[
[
[
1
1.
[
[
[
44] R. K. Garlick, R. W. Giese, Biochem. J. 1990, 268, 611–613.
45] X. Zhou, J. Shearer, S. E. Rokita, J. Am. Chem. Soc. 2000, 122, 9046–9047.
46] J. G. M. Rack, R. Morra, E. Barkauskaite, R. Kraehenbuehl, A. Ariza, Y. Qu,
M. Ortmayer, O. Leidecker, D. R. Cameron, I. Matic, A. Y. Peleg, D. Leys,
A. Traven, I. Ahel, Mol. Cell 2015, 59, 309–320.
[
[
20] H. A. V. Kistemaker, A. P. Nardozza, H. S. Overkleeft, G. A. van der Marel,
A. G. Ladurner, D. V. Filippov, Angew. Chem. Int. Ed. 2016, 55, 10634–
[
[
[
47] C. D. Appel, G. K. Feld, B. D. Wallace, R. S. Williams, Protein Sci. 2016, 25,
1
0638; Angew. Chem. 2016, 128, 10792–10796.
1
682–1691.
48] L. Zhang, J. Cao, L. Dong, H. Lin, Proc. Natl. Acad. Sci. USA 2020, 117,
3447–13456.
[
21] H. Gold, P. Van Delft, N. Meeuwenoord, J. D. C. Codée, D. V. Filippov, G.
Eggink, H. S. Overkleeft, G. A. Van Der Marel, J. Org. Chem. 2008, 73,
1
9
458–9460.
49] M. E. Grunewald, M. G. Shaban, S. R. Mackin, A. R. Fehr, S. Perlman, J.
[
[
[
[
22] J. Voorneveld, J. G. M. Rack, I. Ahel, H. S. Overkleeft, G. A. Van Der Marel,
Virol. 2020, 94, 1–17.
D. V. Filippov, Org. Lett. 2018, 20, 4140–4143.
23] E. Giralt, D. Andreu, Proceedings of the Twenty-First European Peptide
Symposium, ESCOM, Leiden, 1991, p. 131.
24] Y. Oikawa, T. Yoshioka, O. Yonemitsu, Tetrahedron Lett. 1982, 23, 885–
Manuscript received: January 28, 2021
Accepted manuscript online: March 26, 2021
Version of record online: ■■■, ■■■■
8
88.
25] B. Samuelsson, J. Chem. Soc. Perkin Trans. 1984, 2371–2374.
Chem. Eur. J. 2021, 27, 1–8
www.chemeurj.org
7
© 2021 The Authors. Chemistry - A European Journal published by Wiley-VCH GmbH
��
These are not the final page numbers!

FULL PAPER
J. Voorneveld, Dr. J. G. M. Rack, L.
van Gijlswijk, N. J. Meeuwenoord, Dr. Q.
Liu, Prof. H. S. Overkleeft, Prof. G. A.
van der Marel, Prof. I. Ahel*, Dr. D. V.
Filippov*
1
– 8
ADP-ribosylated peptides: A
Ser-, Thr- or Cys-residues, the latter of
which can act as a stable isostere for
Ser-ADPr as Cys-ADPr is not reversed
by the Ser-ADPr hydrolase ARH3. Fur-
thermore, in using these peptides, the
first hydrolase capable of hydrolysing
Cys-ADPr was identified.
Molecular Tools for the Study of
ADP-Ribosylation: A Unified and
Versatile Method to Synthesise
Native Mono-ADP-Ribosylated
Peptides
universal approach to synthesise
mono-ADP-ribosylated peptides is
presented. These peptides can
provide a valuable tool for studying
ADP-ribosylation. This is demon-
strated by the synthesis of a variety of
ADP-ribosylated peptides modified on