Adenosine analogs bearing phosphate isosteres as human MDO1 ligands

Article abstract of DOI:10.1016/j.bmc.2018.02.006

The human O-acetyl-ADP-ribose deacetylase MDO1 is a mono-ADP-ribosylhydrolase involved in the reversal of post-translational modifications. Until now MDO1 has been poorly characterized, partly since no ligand is known besides adenosine nucleotides. Here, we synthesized thirteen compounds retaining the adenosine moiety and bearing bioisosteric replacements of the phosphate at the ribose 5′-oxygen. These compounds are composed of either a squaryldiamide or an amide group as the bioisosteric replacement and/or as a linker. To these groups a variety of substituents were attached such as phenyl, benzyl, pyridyl, carboxyl, hydroxy and tetrazolyl. Biochemical evaluation showed that two compounds, one from both series, inhibited ADP-ribosyl hydrolysis mediated by MDO1 in high concentrations.

Full text of DOI:10.1016/j.bmc.2018.02.006

Accepted Manuscript
Adenosine analogs bearing phosphate isosteres as human MDO1 ligands
Yuezhou Zhang, Mikael Jumppanen, Mirko M. Maksimainen, Samuli Auno,
Zulfa Awol, Léo Ghemtio, Harikanth Venkannagari, Lari Lehtiö, Jari Yli-
Kauhaluoma, Henri Xhaard, Gustav Boije af Gennäs
PII:
S0968-0896(17)32035-7
https://doi.org/10.1016/j.bmc.2018.02.006
BMC 14192
DOI:
Reference:
Bioorganic & Medicinal Chemistry
To appear in:
Received Date:
Revised Date:
Accepted Date:
16 October 2017
1 February 2018
5 February 2018
Please cite this article as: Zhang, Y., Jumppanen, M., Maksimainen, M.M., Auno, S., Awol, Z., Ghemtio, L.,
Venkannagari, H., Lehtiö, L., Yli-Kauhaluoma, J., Xhaard, H., Gennäs, G.B.a., Adenosine analogs bearing
phosphate isosteres as human MDO1 ligands, Bioorganic & Medicinal Chemistry (2018), doi: https://doi.org/
10.1016/j.bmc.2018.02.006
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Adenosine analogs bearing phosphate
isosteres as human MDO1 ligands
Leave this area blank for abstract info.
Yuezhou Zhang a,†, Mikael Jumppanen a,†, Mirko M. Maksimainen b, Samuli Auno a, Zulfa Awol a, Léo Ghemtio
c, Harikanth Venkannagari b, Lari Lehtiö b, Jari Yli-Kauhaluoma a, Henri Xhaard a and Gustav Boije af Gennäs a,
∗
a Drug Research Program, Division of Pharmaceutical Chemistry and Technology, Faculty of Pharmacy, FI-
00014 University of Helsinki, Finland
b Faculty of Biochemistry and Molecular Medicine, Biocenter Oulu, FI-90014 University of Oulu, Finland
c Drug Research Program, Division of Pharmaceutical Biosciences, Faculty of Pharmacy, FI-00014 University
of Helsinki, Finland

Bioorganic & Medicinal Chemistry
journal homepage: www.els evier.com
Adenosine analogs bearing phosphate isosteres as human MDO1 ligands
Yuezhou Zhang a,†, Mikael Jumppanen a,†, Mirko M. Maksimainen b, Samuli Auno a, Zulfa Awol a, Léo
Ghemtio c, Harikanth Venkannagari b, Lari Lehtiö b, Jari Yli-Kauhaluoma a, Henri Xhaard a and Gustav
Boije af Gennäs a, ∗
a Drug Research Program, Division of Pharmaceutical Chemistry and Technology, Faculty of Pharmacy, FI-00014
University of Helsinki, Finland
b Faculty of Biochemistry and Molecular Medicine, Biocenter Oulu, FI-90014 University of Oulu, Finland
c Drug Research Program, Division of Pharmaceutical Biosciences, Faculty of Pharmacy, FI-00014 University of
Helsinki, Finland
ARTICLE INFO
ABSTRACT
Article history:
Received
Received in revised form
Accepted
The human O-acetyl-ADP-ribose deacetylase MDO1 is a mono-ADP-ribosylhydrolase
involved in the reversal of post-translational modifications. Until now MDO1 has been poorly
characterized, partly since no ligand is known besides adenosine nucleotides. Here, we
synthesized thirteen compounds retaining the adenosine moiety and bearing bioisosteric
replacements of the phosphate at the ribose 5'-oxygen. These compounds are composed of
either a squaryldiamide or an amide group as the bioisosteric replacement and/or as a linker.
To these groups a variety of substituents were attached such as phenyl, benzyl, pyridyl,
carboxyl, hydroxy and tetrazolyl. Biochemical evaluation showed that two compounds, one
from both series, inhibited ADP-ribosyl hydrolysis mediated by MDO1 in high concentrations.
Available online
Keywords:
adenosine analogs;
adenosine diphosphate ribose;
mono-ADP-ribosylhydrolase;
human macrodomain-containing protein 1;
MDO1;
2009 Elsevier Ltd. All rights reserved
.
macrodomain;
phosphate isosteres
1. Introduction
The crystal structure of human MDO1 has been solved in the
apo state (PDB code: 2X47).6 MDO1 is closest to MDO2 (PDB
code: 4IQY)7 with a sequence identity of 57% and a root mean
square deviation of 1.61 Å over 220 Cα atoms. MDO1 and
MOD2 share essentially the same binding site for ADP-ribose,
Macrodomain proteins contain at least one copy of an α/β
globular domain of about 25 kDa that is evolutionarily conserved
among eukaryotes, prokaryotes and viruses.1
Most of
with two conservative changes only, Val271MDO1 to Ile189MDO2
;
macrodomain proteins studied to date have been shown to
function as ADP-ribose-binding modules.2 Humans have eleven
macrodomain proteins according to the Pfam 26.0 database:
macrodomain-containing proteins 1-3 (MDO1-3), five members
of the poly(ADP-ribose) polymerase superfamily, and three H2A
histone family members.3
The function(s) of human MDO1 (as well as its homologs
MDO2 and MDO3) are poorly known. MDO1 is localized in the
mitochondria, suggesting a contribution on ATP generation,
signaling, cellular differentiation and cellular growth.4 MDO1
has been shown to function as a mono-ADP-ribosylhydrolase and
a deacetylase of O-acetyl- ADP-ribose.5, 6 Indirectly, MDO1 may
thus be implicated in gene silencing.7 Human MDO1 is
associated with multiple cancers, such as ovarian carcinoma,8
breast carcinoma,9 acute leukemia10 and prostate cancer.11 This
makes the human MDO1 protein a promising new target for
potential therapeutic applications.12 As only a few synthetic
ligands have been reported, the endogenous ligand ADP-ribose
should serve as a starting point for ligand-based design.
Phe272MDO1 to Tyr190MDO2. Modeling the bound mode of ADP-
ribose in the apo MDO1 nonetheless requires a change in the
side-chain conformation of Phe306MDO1 to make it similar to
Phe224MDO2. Affinities towards ADP-ribose are comparable for
these proteins with a KD of 0.9 µM for MDO1 and a KD of 0.15
µM for MDO2.4, 13
The first objective of this study is to explore modifications to
the phosphate-ribose part of ADP-ribose. Phosphate-containing
compounds typically show poor bioavailability, are not
physiologically stable, show low permeability, and their
syntheses may be challenging.14, 15 The second objective is to
explore the binding pocket for a location that is able to
accommodate large and flat fluorescent groups such as fused
tricyclic aromatic rings. Such compounds, exemplified by
fluorescein, would be useful fluorescent probes in cell-based
studies of MDO1.
We report the design and synthesis of two series of adenosine
derivatives targeting the ADP-ribose binding pocket of MDO1.

Implementing an isosteric replacement strategy, adenosine was
coupled to unsymmetrically substituted squaryldiamides and
various amide linkers. Squaryldiamides possess two sp2 oxygen
atoms that carry a double partial negative charge through their
delocalized electrons, and have been reported to act as good
mimics of the phosphate moiety.16, 17 An amide group acts also as
a promising phosphate replacement, as exemplified by structural
isosteres found in the Protein Data Bank.18 Amides have also
been shown to be successful mimics of phosphodiester
internucleoside linkages in RNA19 and they are synthetically
feasible.
protected and unsymmetrically 3,4-substituted cyclobut-3-ene-
1,2-diones 13-20 that were subsequently deprotected in the
presence of aqueous trifluoroacetic acid (TFA) to give a set of
eight
different
3-(5’-amino-5’-deoxyadenosine)-4-
(alkylamino)cyclobut-3-ene-1,2-diones (21-28) in 45-90% yields.
The synthesis of the amide series was started by coupling
amine 3 with N-hydroxysuccinimide (OSu)-activated ester and
Fmoc- and tert-butyl-protected amino acids (Fmoc-Ser(tBu)-OSu
and Fmoc-Asp(OtBu)-OSu, respectively) to give compounds 29
and 30 in 70 and 75% yields, respectively (Scheme 2).
Deprotection of the acetonide- and the tert-butyl groups of
compounds 29 and 30 was performed in the presence of TFA in
H2O/CH2Cl2 to give the Fmoc-protected amino acid derivatives
31 and 33 in 70 and 90% yields, respectively. The serine
derivative 34 was obtained in 27% yield after deprotection of the
2. Results
2.1. Ligand-based design
Fmoc
group
with
diethylamine
(DEA)
in
N,N-
dimethylformamide (DMF).
In the first series of derivatives, we designed eight
asymmetrically substituted squaryldiamides with different
lengths, shapes and polarities (Table 1). We chose four benzylic
(compounds 21, 24, 26 and 28), aliphatic (compound 23), aryl
(compounds 22 and 25), and heteroaromatic (compound 27)
substituents. In the second series of five derivatives, adenosine
analogs with an amide linker were designed (compounds 31, 33-
34, 38 and 39, Table 2). To investigate the replacement of the
distal phosphate of ADP-ribose with a hydroxy or carboxyl
group, three of these compounds were designed as derivatives of
L-serine (compounds 31 and 34) and L-aspartic acid (compound
33). The carboxyl group of aspartic acid is similar to the
phosphate group especially in terms of geometry and pKa (pKa of
3-5 for carboxylic acid vs. 1.54 and 6.31 for the pKa1 and pKa2 of
phosphate).20 Compounds 31 and 33 are Fmoc-protected ones
tested to explore the ability of the binding pocket to
accommodate large and flat fluorescent groups. This could be
employed in the development of fused tricyclic fluorescent
probes for future screening studies of MDO1. Two compounds
contained a tetrazolyl group (compounds 38 and 39) and were
designed since tetrazole is a classical phosphate bioisosteric
replacement21 with a pKa around 4.90.22 The compound 39 with
the 1H proton of the tetrazole moiety substituted by a phenyl
group was designed to study the effect of an acidic proton of the
1H-tetrazole moiety for an enzymatic activity.
In the syntheses of the two tetrazole analogs 38 and 39, amine
3 was first reacted with either 1H-tetrazole-5-acetic acid or [(1-
phenyl-1H-tetrazol-5-yl)thio]acetic acid using N,N,N′,N′-
tetramethyl-O-(1H-benzotriazol-1-yl)uronium
hexafluorophosphate (HBTU) as a coupling reagent to give the
amides 36 and 37 in 80 and 85% yields, respectively (Scheme 3).
Subsequently, the acetonide-protecting groups of amides 36 and
37 were removed with TFA/H2O/CH2Cl2 (5:1:4) to give the
products 38 and 39 in 83 and 77% yields, respectively.
2.3. Biological activities
We first attempted to measure quantitatively the binding
affinities of the set of compounds, together with those of
benchmark nucleotides (ADP-ribose, ADP, AMP), using
isothermal titration calorimetry. We also tried differential
scanning fluorimetry, where the stabilizing effect of compounds
on temperature-dependent denaturation is measured by the
unspecific binding of a hydrophobic reporter dye to the protein
core.24 Molecules that are better stabilizers are thought to be
tighter binders. Unfortunately, as
a result of these two
experiments, no detection could be made in ITC and only very
small ∆Tm could be obtained even for the ADP and AMP controls
(less that one °C unit). For the endogenous substrate ADP-ribose
to the human MDO1, the KD measured using isothermal titration
calorimetry is in the order of 0.9 µM25 and the associated ∆Tm
measured in this study was about 5 °C. Thus, the results suggest
weaker affinities than for the endogenous substrate ADP-ribose,
in the low to high µM if not mM. Overall, these findings are not
surprising: it is unlikely that an optimal fit of the phosphate
replacement in the binding pocket, associated with an affinity as
good as that of ADP-ribose, will be found on a first try for our
compounds.
We therefore resorted to use a biochemical assay, where
presence of mono-biotin-ADP-ribosylated SRPK2 is followed
qualitatively using a western blot (Figure 1). In this setup the
negative control is the mono-biotin-ADP-ribosylated SRPK2
alone, whereas the positive control is composed of SPRK2 and
MDO1 without an inhibitor. Inhibited MDO1 cannot efficiently
hydrolyze the modification from SRPK2 resulting in stronger
signal than the positive control. The analysis revealed that 28 and
31, and to a lesser extent 27, inhibit the catalytic activity of
MDO1. Compound 28 is more efficient inhibitor than 31 based
on the visual comparison of their relative intensities.
2.2. Synthesis
The compounds of the first series, the squaryldiamides 21-28,
were synthesized in three reaction steps starting from diethyl
squarate (4) and two different amines (Scheme 1). First, the key
building block, acetonide-protected 5’-amino-5’-deoxyadenosine
(3) was synthesized. We treated 2',3'-O-isopropylideneadenosine
(1) first with diphenylphosphoryl azide (DPPA) and 1,8-
diazabicyclo[5.4.0]undec-7-ene (DBU), and subsequently with
sodium azide and 1,4,7,10,13-pentaoxacyclopentadecane (15-
crown-5) in 1,4-dioxane under reflux to yield the acetonide-
protected 5'-azido-5'-deoxyadenosine (2) in 86% yield. Next, the
azido intermediate (2) was converted to the desired key building
block (3) by using a modified Staudinger reaction in the presence
of polymer-bound triphenylphosphine.20 The acetonide-protected
5'-amino-5'-deoxyadenosine (3) was obtained in 70% yield.
We started the synthesis of target compounds by treating
diethyl squarate (4) with various aliphatic, aromatic and benzylic
amines in ethanol to give eight differently substituted alkoxy
amino squarates (5-12) in 25-92% yields. Only mono-substituted
products were formed in this reaction, which was consistent with
the literature.23 In a similar fashion, the acetonide-protected 5´-
amino-5´-deoxyadenosine (3) was reacted with alkoxy amino
squarates 5-12. This transformation produced the acetonide-
However, we could not confirm the results obtained from
western blot using the Alphascreen-based activity assay for
mono-ADP-ribosyl hydrolyzing macrodomains.26 The assay
utilizes the same principle as the western blot and measures
hydrolysis of ADP-ribosyl group from SRPK2. Dose-response

measurements did not show any significant inhibition for
compounds 28 and 31 against MDO1 or MDO2 (Supplementary
Material S1) even at the highest compound concentration 6.6
mM. This indicates that the ADP-ribose analogs lacking the
Scheme 1. Synthesis of the squaryldiamide derivatives.a
distal ribose only show low potencies against macrodomains
similar to ADP and AMP.
a Reagents and conditions: i) DPPA, DBU, NaN3, 15-crown-5, 1,4-dioxane, rt to 110 °C, overnight, 86%; ii) triphenylphosphine (polymer-bound), THF, H2O, rt,
4 h, 70%; iii) amine, EtOH, 0 °C to rt, 2 h, 25-92%; iv) DIPEA, EtOH, rt, 5 h to overnight, 37-94%; v) TFA/H2O (4:1), rt, 5 h, 45-90%.
Scheme 2. Synthesis of the amino acid derivatives.a
a
Reagents and conditions: i) Fmoc-Ser(tBu)-OSu or Fmoc-Asp(OtBu)-OSu, MeOH, rt, 2 h, 70-75%; ii) TFA/H2O/CH2Cl2 (4-5:1:1-4), rt, 5-24 h, 70-90%; iii)
DEA, DMF, rt, 5 h, 27%.
Scheme 3. Synthesis of the tetrazole derivatives.a
a Reagents and conditions: i) 1H-tetrazole-5-acetic acid or [(1-phenyl-1H-tetrazol-5-yl)thio]acetic acid, HBTU, DIPEA, DMF rt, 4 h, 80-85%; ii)
TFA/H2O/CH2Cl2 (5:1:4), rt, 4-5 h, 77-83%.
Figure 1. Western blot analysis of the inhibition efficiency for MDO1 in
the hydrolysis of mono-biotin-ADP-ribosylation from SRPK2 (A)
compounds 21-28 (B) compounds 31-34, and 38-39. Compounds 21, 26-28
and 33-39 were tested at 1 mM concentrations, and compounds 22-25 and 31
were tested at 0.8 mM concentrations. Compound 35 is not described in this
manuscript.27

As part of the design process, all the compounds were
prospectively docked to the three-dimensional structure of
MDO1. We hypothesized that the crystal structure of MDO2 can
be
used
Table 1. Inhibition of ADP-ribosyl hydrolysis and docking
scores for the compounds of the squaryldiamide series prepared
as shown Scheme 1.a
Table 2. Inhibition of ADP-ribosyl hydrolysis and docking
scores for the compounds of the amino acid series (Schemes 2
and 3).a
Docking
Compound
R
-
Inhibition
score
Docking
Compound
R
Inhibition
score
-10.376
-
-
AMP
ADP
31
33
-10.043
++
-
-
-11.029
-11.902
-10.678
-9.973
ADP-
ribose
-
-14.566
-10.203
-8.972
+++
34
38
-
-
21
22
-
-
39
-9.894
-
23
24
25
26
27
-10.017
-9.918
-
a
Readout: -, no reading; ++, moderate reading. Compounds were tested at
0.8-1 mM concentrations. Docking scores were calculated using the XP
scoring function (Schrödinger Ltd) The lower the value, the more favorably
the compound is predicted to bind.
-
to derive the binding mode of ADP-ribose to MDO1. This is
especially supported by the near-identical binding sites. Since the
binding site of the apo structure is occluded, the conformation of
the side chains Val181 and Phe306 had to be modified to
accommodate nucleotide ligands, and this was done by
reconstructing a homology model of MDO1 based on MDO2
(Supplementary Material S2). Redocking ADP-ribose to MDO1
led to a very similar binding mode as in MDO2 (Figure 2A and
Figure 2B) but only after a distance restraint had forced the
adenine ring onto its binding site. The docking procedure
validated nonetheless that in theory no large clashes occurred at
the binding site and that hydrogen-bonding groups of the protein
were geometrically suitably positioned, i.e. that the compounds
were potentially inhibitors of MDO1. ADP-ribose obtained
clearly the best score with a value of -14.566), while the docking
scores of the second most active compounds (ADP and 28) were
also among the best. We next focused on the analysis of the most
active compounds, 28 and 31 (Figure 2C and 2D).
-10.171
-10.692
-10.082
-11.239
-
-
+
28
+++
a
Readout: -, no reading; +, weak reading; +++, high reading. Compounds
were tested at 0.8-1 mM concentrations. Docking scores were calculated
using the XP scoring function (Schrödinger Ltd) The lower the value, the
more favorably the compound is predicted to bind.
For compounds 28 and 31, the adenine ring and the ribose
were constrained to occupy a site equivalent to that predicted for
ADP-ribose (Figure 2). The adenine ring is sandwiched between
the planar phenyl group of Phe306 (π-π interaction) and the
isopropyl side-chain of Val183. The amino group of adenine (N6)
is donating a hydrogen bond to the Asp160 carboxyl group. In
2.4. Docking simulations of ADP-ribose, compound 28 and
compound 31.

addition, the N1 nitrogen from the Watson-Crick edge of adenine
accepts a hydrogen bond from the main-chain nitrogen of Ile161.
The hydrogen of the ribose O3' hydroxy group donates a
hydrogen bond to the side-chain of Thr269 in many complexes.
We next examined which groups are docked in place of the
phosphate groups of ADP-ribose. In MDO2, the negative charge
Figure 2. The complex of human MDO2 with (A) ADP-ribose (PDB file: 4IQY) and predicted binding modes of modeled human MDO1 in complex with
(B) ADP-ribose, (C) compound 28, and (D) compound 31. Carbon, oxygen, and nitrogen atoms are shown in black, red, and blue, respectively. The carbon
atoms of binding site residues are shown in green. Putative hydrogen bonds are depicted as dashed lines in yellow.
prominent interaction is stacking on Phe272. Explaining the
selectivity of 28 over its analogs 21, 24, 26 and 27 is highly
speculative, but it may be that the methyl groups at the para
position create additional hydrophobic interactions.
of the proximal phosphate is neutralized by the positively
charged macrodipole at the N-terminus of helix 8. This
macrodipole is generated by the sum of the dipolar contributions
from the helical backbone. In addition, on the top of this helix,
the main chain NH of Val101 in MDO2 donates a hydrogen bond
to an oxygen atom of the proximal phosphate (Figure 2A); in
3. Conclusions
MDO1 the NH of the equivalent Val183 interacts with this time
In conclusion, we designed and synthesized novel ADP-ribose
the distal ADP-ribose phosphate, probably reflecting docking
analogs targeted to MDO1 by replacing the phosphates with
uncertainties (Figure 2B). For neither 28 nor 31 the amide
bioisosteres and a set of aromatic substituents in place of the
distal ribose. As a result, two inhibiting compounds were found,
nitrogen bonds with Gly182 and Val183 is satisfied in the docked
poses. This may reflect docking uncertainties, or water molecules
providing early indication about the SAR of MDO1 compounds.
that do play important role in molecular recognition and may
The results suggested that both a squaryldiamide and an amide
bridge these hydrogen-bonding groups, although they are not
linker were accepted by the binding site. A 4-methyl substituent
in the phenyl group of 28 provided additional potency perhaps
through interaction with an aromatic phenylalanine in the binding
included in the docking simulation. On the other hand, it is
possible that a next generation of compounds could be designed
that contain a hydrogen-bond acceptor pointing towards the main
site and could be used as a starting point for design of more
chain NH groups of Val183 and Gly182. Instead, for both 28 and
potent compounds. The next generation of derivatives could also
have an additional hydrogen-bond acceptor pointing towards the
main chain NH of Val183 and Gly182 to gain activity. Based on
compound 31 the ADP-ribose binding groove may accommodate
31 new interactions were created with Ser268 and Thr269. In
particular, an oxygen atom of the linker was found at close
distance of the Val171 main-chain nitrogen, indicating the
possibility of a hydrogen bond.
a bulky fluorophore group which could be a way for generation
of tool compounds for studies on biological roles of the
macrodomains. As the distal ribose is a key feature of the ADP-
ribose, proper interactions need to be designed at this region of
the binding pocket in order to achieve more potent inhibitors.
The third site of interest is that of the distal ribose. The ribose
especially benefits from interaction with Asp160 in MDO1 and
its equivalent Asp78 in MDO2. An equivalent interaction is not
possible for 28 and 31 that are devoid of oxygen or nitrogen
atoms in this region of the binding site. In contrast, an especially

4. Experimental
140.0, 120.4, 114.9, 90.8, 85.8, 84.2, 82.2, 52.5, 27.3, 25.5;
HRMS-ESI m/z calcd for C13H17N8O3 333.1418 [M+H]+, found
333.1424. 1H NMR spectral data is consistent with those
previously reported.28
4.1. Materials and general procedures
All reagents were commercially available. They were acquired
from Fluka (Buchs, Switzerland) and Sigma-Aldrich
(Schnelldorf, Germany) and were used without further
purification. All reactions in anhydrous solvents were conducted
in oven-dried glassware under argon. Thin-layer chromatography
(TLC) was performed using Silica Gel 60 F254 (Merck) and Silica
Gel 60 NH2 F254s aluminum sheets (Merck), visualized by UV
illumination and stained with ninhydrin in EtOH (1.5% w/v).
Column chromatography was performed with an automated high
performance flash chromatography Biotage Sp4-system or with a
Isolera Spektra One-system (Uppsala, Sweden) using a 0.1-mm
path length flow cell UV-detector/recorder module (fixed
wavelength 254 nm) for the Sp4-system or a variable UV-VIS
(200-800 nm) photodiode array detector for the Isolera Spektra
4.2.2. 2′,3′-O-Isopropylidene-5'-amino-5'-
deoxyadenosine (3)
Polymer-bound triphenylphosphine (loading: 1.4 mmol/g, 4.29 g,
6.02 mmol) was suspended in anhydrous THF (60 mL). The
mixture was left to stand for 5 min, then a solution of the azide 2
(1.00 g, 3.01 mmol) in anhydrous THF (2 mL) was added. The
suspension was stirred at rt for 4 h. Water (0.540 mL, 30.1 mmol)
was added to the reaction mixture and it was stirred at rt
overnight. The resin was washed with THF, the combined
solutions were dried with anhydrous Na2SO4, filtrated, and
evaporated. The crude product was purified by column
1
One-system, and the indicated mobile phase. The H and 13C
→ 4:1) to give
chromatography on silica gel (CHCl3/MeOH 10:1
the compound 3 as a white solid (645 mg, 2.11 mmol, 70%
yield). Rf = 0.25 (CHCl3 /MeOH, 1:4). 1H NMR (300 MHz,
CDCl3) δppm 8.34 (s, 1H), 7.91 (s, 1H), 6.03 (d, J = 3.1 Hz, 1H),
5.68 (br s, 2H), 5.46 (dd, J = 6.5, 3.1 Hz, 1H), 5.04 (dd, J = 6.4,
3.4 Hz, 1H), 4.27 (dd, J = 9.5, 4.3 Hz, 1H), 3.05 (dd, J = 13.4,
4.6 Hz, 1H), 2.98 (dd, J = 13.4, 5.8 Hz, 1H), 1.91 (br s, 2H), 1.62
(s, 3H), 1.39 (s, 3H); 13C NMR (75.4 MHz, CDCl3) δppm 155.8,
153.4, 150.3, 140.2, 120.7, 114.8, 91.0, 87.5, 83.9, 82.1, 44.0,
27.5, 25.6; HRMS-ESI m/z calcd for C13H19N6O3 307.1513
[M+H]+, found 307.1516.
NMR spectra were recorded on a Varian Mercury-VX 300
(Agilent Technologies, Santa Clara, California, USA) or Bruker
Ascent 400 (Bruker Corporation, Billerica, Massachusetts, USA)
spectrometer as solutions in CDCl3, DMSO-d6 and CD3OD.
Chemical shifts (δ) are reported as parts per million (ppm)
relative to the solvent peak (CDCl3 7.26 and 77.16 ppm, DMSO-
d6 2.50 and 39.52 ppm, CD3OD 3.31 and 49.00 ppm).
Multiplicities of peaks are represented by s (singlet), d (doublet),
t (triplet), q (quartet), quintet (qn), and m (multiplet). HPLC–MS
analyses were performed to determine purity of all tested
compounds. Purity of all tested compounds was > 95%. Mass
spectra were measured on a Bruker Daltonics Esquire-HPLC
spectrometer, with XTerra MS RP18 column (4.6 mm × 30 mm,
2.5 µm). High resolution mass spectra (HRMS) were measured
4.2.3. 3-Ethoxy-4-[(4-
fluorobenzyl)amino]cyclobut-3-ene-1,2-dione (8)
on
a
Waters Synapt G2 (Waters Corporation, Milford,
4-Fluorobenzylamine (156 µL, 171 mg, 1.36 mmol) was added at
0 °C to a solution of 3,4-diethoxycyclobut-3-ene-1,2-dione (0.20
mL, 2300 mg, 1.36 mmol) in EtOH (10 mL). The reaction
mixture was stirred at rt for 2 h. The solvents were removed in
vacuo and the residue was purified by column chromatography
on silica gel (n-heptane/EtOAc 1:1) to give compound 8 as a
white solid (189 mg, 0.76 mmol, 56% yield). Rf = 0.46. 1H NMR
(400 MHz, CDCl3) δppm 7.33-7.27 (m, 2H), 7.11–6.91 (m, 2H),
5.72 (br s, 1H), 4.77 (br s, 3H), 4.57 (br s, 1H), 1.45 (t, J = 7.1
Hz, 3H). 13C NMR (101 MHz, CDCl3) δppm 189.5, 183.1, 177.9,
172.3, 162.7 (d, J = 247.3 Hz), 132.9, 129.7, 116.07 (d, J = 21.6
Hz), 70.1, 48.1, 16.0; HRMS-ESI m/z calcd C13H13FNO3
250.0879 [M+H]+, found 250.0880.
Massachusetts, USA) and reported for the molecular ions
[M+H]+.
4.2. Synthesis procedures
Procedures for the synthesis and analysis of the remaining
compounds are available in the Supplementary Material.
4.2.1. 2′,3′-O-Isopropylidene-5'-azido-5'-
deoxyadenosine (2)
To a solution of 2´,3´-O-isopropylideneadenosine (1.00 g, 3.25
mmol) in dry 1,4-dioxane (10 mL), DPPA (1.40 g, 1.48 mL, 6.5
mmol) and DBU (1.34 g, 1.50 mL, 9.8 mmol) were added
dropwise. The reaction mixture was stirred at rt for 47 h. Sodium
azide (1.06 g, 16.3 mmol) and 15-crown-5 (0.527 g, 0.064 mL,
0.326 mmol) were added to the reaction mixture. The resulting
mixture was irradiated in a microwave reactor at 110 °C for 4 h
(1 bar, 600 rpm) and cooled to rt. Filtration of the reaction
mixture and evaporation of the filtrate gave a residue that was
redissolved in CHCl3 (90 mL). The organic phase was washed
with water (3 × 15 mL) and brine (3 × 15 mL). The organic phase
was dried with anhydrous Na2SO4, filtrated and evaporated to
give a brown oil. The crude product was purified by column
chromatography on silica gel (EtOAc/EtOH/n-heptane 3:1:4)
yielding the azide 2 as an amorphous yellow solid (0.928 g, 86%
yield) with some 15-crown-5 as an impurity. The product was
used in the next step without further purification. Rf = 0.33
4.2.4. 3-Ethoxy-4-[(3-nitrophenyl)amino]cyclobut-
3-ene-1,2-dione (9)
3-Nitroaniline (85 mg, 0.68 mmol) was added at 0 °C to a
solution of 3,4-diethoxycyclobut-3-ene-1,2-dione (0.10 mL, 120
mg, 0.68 mmol) in EtOH (3 mL). The reaction mixture was
stirred at rt for 5 h. The solvents were removed in vacuo and the
residue was purified by column chromatography on silica gel (n-
hexane/EtOAc 1:1) to give compound 9 as an orange solid (126
1
mg, 0.48 mmol, 71% yield). Rf = 0.29 (CHCl3/MeOH, 10:1). H
NMR (300 MHz, DMSO-d6) δppm 11.06 (br s, 1H), 8.32 (t, J =
2.2 Hz, 1H), 7.92 (ddd, J = 8.1, 2.2, 0.9 Hz, 1H), 7.78 (ddd, J =
8.2, 2.3, 1.0 Hz, 1H), 7.64 (t, J = 8.2 Hz, 1H), 4.80 (q, J = 7.1 Hz,
2H), 1.44 (t, J = 7.1 Hz, 3H); 13C NMR (75.4 MHz, DMSO-d6)
δppm 187.4, 184.2, 178.9, 169.3, 148.3, 139.4, 130.5, 125.3, 118.0,
113.6, 69.9, 15.5; HRMS-ESI m/z calcd C12H11N2O5 263.0662
[M+H]+, found 263.0668.
1
(CHCl3/ EtOH, 20:1). H NMR (300 MHz, CDCl3) δppm 8.32 (s,
1H), 7.90 (s, 1H), 6.23 (br s, 2H), 6.10 (d, J = 2.2 Hz, 1H), 5.45
(dd, J = 6.3, 2.2 Hz, 1H), 5.04 (dd, J = 6.3, 3.3 Hz, 1H), 4.41–
4.31 (m, 1H), 3.55 (dd, J = 5.7, 2.9 Hz, 2H), 1.60 (s, 3H), 1.37 (s,
3H). 13C NMR (75.4 MHz, CDCl3) δppm 156.0, 153.3, 149.3,

4.2.5. 3-Ethoxy-4-[(4-
methoxybenzyl)amino]cyclobut-3-ene-1,2-dione
(10)
solvent was removed in vacuo, and the residue was purified by
column chromatography on silica gel (CHCl3/MeOH 10:1) to
give compound 18 as a white solid (150 mg, 0.30 mmol, 91%
yield). Rf = 0.35 (CHCl3/MeOH, 10:1). 1H NMR (300 MHz,
DMSO-d6) δppm 8.30 (s, 1H), 8.17 (s, 1H), 7.65 (br s, 1H), 7.31
(br s, 2H), 7.22 (d, J = 8.5 Hz, 2H), 6.90 (d, J = 8.6 Hz, 2H), 6.19
(d, J = 2.7 Hz, 1H), 5.42 (dd, J = 6.3, 2.6 Hz, 1H), 5.01 (dd, J =
6.3, 3.5 Hz, 1H), 4.60 (d, J = 6.0 Hz, 2H), 4.30–4.20 (m, 1H),
4.01–3.81 (m, 1H), 3.73 (s, 3H), 1.54 (s, 3H), 1.32 (s, 3H); 13C
NMR (75.4 MHz, DMSO-d6) δppm 183.0, 182.6, 168.1, 167.8,
159.0, 156.3, 153.2, 149.0, 140.3, 130.9, 129.2, 119.3, 114.3,
89.0, 83.4, 81.4, 70.1, 55.4, 46.6, 45.4, 27.3, 25.5; HRMS-ESI
m/z calcd C25H28N7O6 522.2096 [M+H]+, found 522.2101.
4-Methoxybenzylamine (178 µL, 1.36 mmol) was added at 0 °C
to a solution of 3,4-diethoxycyclobut-3-ene-1,2-dione (0.20 mL,
240 mg, 1.36 mmol) in EtOH (5 mL). The reaction mixture was
stirred at rt for 2 h. The solvents were removed in vacuo, and the
residue was purified by column chromatography on silica gel (n-
hexane/EtOAc 1:1) to give compound 10 as a white solid (291
mg, 1.17 mmol, 86% yield). Rf = 0.41 (n-hexane/EtOAc, 1:1). 1H
NMR (300 MHz, CHCl3) δppm 7.24–7.02 (m, 3H), 6.91–6.81 (m,
2H), 4.75 (br s, 2H), 4.52 (br s, 2H), 3.78 (s, 3H), 1.42 (t, J = 7.5
Hz, 1H); HRMS-ESI m/z calcd for C14H16NO4 262.1001 [M+H]+,
found 262.1079.
4.2.9. 3-(2´,3´-O-Isopropylidene-5´-amino-5´-
deoxyadenosine)-4-[(4-
methylbenzyl)amino]cyclobut-3-ene-1,2-dione (20)
4.2.6. 3-[4-Fluorobenzyl)amino]-4-(2´,3´-O-
isopropylidene-5´-amino-5´-
deoxyadenosine)cyclobut-3-ene-1,2-dione (16)
To a solution of 3 (0.10 g, 0.33 mmol) in EtOH (3 mL) was
added DIPEA (59 µL, 0.33 mmol) and 12 (0.080 mg, 0.33
mmol). The reaction mixture was stirred at rt for 5 h and
evaporated in vacuo. Purification of the crude product by column
To a solution of compound 8 (179 mg, 0.72 mmol) in EtOH (10
mL) was added DIPEA (0.74 mL, 0.79 mmol) and 3 (241 mg,
0.79 mmol). The reaction mixture was stirred at rt for 6 h. The
solvent was removed in vacuo, and the residue was purified by
column chromatography on silica gel (CHCl3/MeOH 10:1) to
give 16 as a white solid (306 mg, 0.6 mmol, 84% yield). Rf =
→ 4:1) gave
chromatography on silica gel (CHCl3/MeOH 10:1
compound 20 as a white solid (140 mg, 0.28 mmol, 85%). Rf =
1
0.29 (CHCl3/MeOH, 10:1). H NMR (300 MHz, DMSO-d6) δppm
1
8.31 (s, 1H), 8.17 (s, 1H), 7.64 (br s, 1H), 7.37 (br s, 3H), 7.25–
7.08 (m, 4H), 6.19 (s, 1H), 5.49–5.35 (m, 1H), 5.01 (dd, J = 6.2,
3.4 Hz, 1H), 4.74–4.50 (m, 2H), 4.32–4.17 (m, 1H), 3.90 (br s,
1H), 3.80–3.63 (m, 1H), 2.27 (s, 3H), 1.54 (s, 3H), 1.32 (s, 3H);
13C NMR (75.4 MHz, DMSO-d6) δppm 182.7, 182.5, 167.5, 156.2,
152.9, 148.8, 139.9, 136.6, 135.8, 129.2, 127.5, 119.2, 113.7,
88.7, 85.2, 83.1, 81.2, 46.6, 45.2, 27.1, 25.3, 20.7; HRMS-ESI
m/z calcd C25H28N7O5 506.2146 [M+H]+, found 506.2152.
0.73 (CHCl3/MeOH, 10:1). H NMR (400 MHz, DMSO-d6) δppm
8.33 (s, 1H), 8.14 (s, 1H), 7.89 (br s, 1H), 7.34 (s, 4H), 7.17 (t, J
= 8.6 Hz, 2H), 6.19 (d, J = 2.8 Hz, 1H), 5.43 (dd, J = 6.6, 2.8 Hz,
1H), 5.03 (dd, J = 6.4, 3.4 Hz, 1H), 4.66 (d, J = 6.6 Hz, 2H),
4.37–4.17 (m, 1H), 4.00–3.70 (m, 2H), 1.54 (s, 3H), 1.32 (s, 3H).
13C NMR (101 MHz, DMSO-d6) δppm 182.7, 182.5, 167.9,
167.6, 161.4 (d, J = 243 Hz), 156.1, 152.9, 148.8, 139.8,
135.2, 129.6 (d, J = 8.4 Hz), 119.1, 115.4 (d, J = 21.2 Hz),
113.7, 88.7, 85.0, 83.1, 81.2, 46.0, 45.1, 27.0, 25.3; HRMS-
ESI m/z calcd C24H25FN7O5 510.1901 [M+H]+, found 510.1901.
4.2.10. 3-(5´-Amino-5´-deoxyadenosine)-4-[(4-
fluorobenzyl)amino]cyclobut-3-ene-1,2-dione (24)
4.2.7. 3-(2´,3´-O-Isopropylidene-5´-amino-5´-
deoxyadenosine)-4-[(3-
nitrophenyl)amino]cyclobut-3-ene-1, 2-dione (17)
A suspension of 16 (0.281 g, 0.55 mmol) in TFA/water (4:1, 10
mL) was stirred at rt for 6 h. The solvent was evaporated in
vacuo, and the residue was co-evaporated with toluene to remove
any residual TFA and H2O. The remaining residue was purified
by column chromatography on silica gel (CHCl3/MeOH 4:1) to
give compound 24 as a light green solid (194 mg, 0.41 mmol,
75% yield). Rf = 0.57 (CHCl3/MeOH, 10:1). 1H NMR (400 MHz,
DMSO-d6) δppm 8.33 (s, 1H), 8.15 (s, 1H), 8.10 (br s, 1H), 7.80
(br s, 1H), 7.41–7.31 (m, 2H), 7.29 (s, 2H), 7.21-7.10 (m, 2H),
5.91 (d, J = 5.9 Hz, 1H), 5.57 (d, J = 6.1 Hz, 1H), 5.38 (d, J = 4.9
Hz, 1H), 4.73-4.59 (m, , 3H), 4.20-4.12 (m, 1H), 4.06–3.97 (m,
1H), 3.94 (br s, 1H), 3.85–3.72 (m, 1H); 13C NMR (101 MHz,
DMSO-d6) δppm 182.6, 182.5, 168.0, 167.5, 161.5 (d, J = 243.0
Hz), 158.0 (q, J = 30.9 Hz) 156.1, 152.7, 149.4, 139.8, 135.2,
129.8 (d, J = 8.2 Hz), 119.2, 117.3 (q, J = 300.0 Hz), 115.4 (d, J
= 21.4 Hz) 87.4, 83.6, 72.8, 70.8, 46.0, 45.6; HRMS-ESI m/z
calcd C21H21FN7O5 470.1588 [M+H]+, found 470.1589. TFA
To a solution of compound 9 (64 mg, 0.31 mmol) in EtOH (3
mL) was added DIPEA (0.060 mL, 0.33 mmol) and 3 (0.10 g,
0.33 mmol). The reaction mixture was stirred at rt for 5 h. The
solvent was removed in vacuo, and the residue was purified by
column chromatography on silica gel (CHCl3/MeOH 10:1) to
give compound 17 as a yellow solid (155 mg, 0.30 mmol, 90%).
1
Rf = 0.46 (CHCl3/MeOH, 10:1). H NMR (300 MHz, DMSO-d6)
δppm 9.95 (br s, 1H), 8.35 (s, 2H), 8.15 (s, 1H), 7.82 (ddd, J = 8.1,
2.2, 0.9 Hz, 1H), 7.70 (d, J = 8.1 Hz, 1H), 7.58 (t, J = 8.1 Hz,
1H), 7.30 (s, 2H), 6.24 (d, J = 2.6 Hz, 1H), 5.47 (dd, J = 6.3, 2.7
Hz, 1H), 5.09 (dd, J = 6.3, 3.7 Hz, 1H), 4.33 (dt, J = 8.0, 4.2 Hz,
1H), 4.04 (dd, J = 14.0, 4.8 Hz, 1H), 3.87 (dd, J = 13.9, 7.3 Hz,
1H), 1.56 (s, 3H), 1.34 (s, 3H); 13C NMR (75.4 MHz, DMSO-d6)
δppm 184.3, 180.5, 169.5, 163.0, 156.1, 152.8, 148.7, 148.6,
140.3, 140.0, 130.6, 123.9, 119.2, 116.6, 113.8, 112.3, 88.6,
84.9, 83.1, 81.0, 45.4, 27.0, 25.3; HRMS-ESI m/z calcd for
C23H23N8O7 523.1684 [M+H]+, found 523.1689.
4.2.11. 3-(5´-Ammonium-5´-deoxyadenosine)-4-
[(3-nitrophenyl)amino]cyclobut-3-ene-1,2-dione
trifluoroacetate (25)
4.2.8. 3-(2´,3´-O-Isopropylidene-5´-amino-5´-
deoxyadenosine)-4-[4-
methoxybenzyl)amino]cyclobut-3-ene-1,2-dione
(18)
A suspension of 17 (88 mg, 0.16 mmol) in TFA/water (4:1, 3
mL) was stirred at rt for 5 h. The solvent was evaporated in
vacuo, and the residue was co-evaporated with toluene to remove
any residual TFA and H2O. The remaining residue was purified
by column chromatography on silica gel (CHCl3/MeOH 4:1) to
give compound 25 as a yellow solid (63 mg, 0.13 mmol, 66%
yield). Rf = 0.48 (CHCl3/MeOH, 4:1). 1H NMR (300 MHz,
To a solution of compound 10 (64 mg, 0.31 mmol) in EtOH (3
mL) was added DIPEA (0.060 mL, 0.33 mmol) and 3 (0.10 g,
0.33 mmol). The reaction mixture was stirred at rt for 5 h. The

DMSO-d6) δppm 10.44 (br zs, 1H), 8.43 (br s, 1H), 8.35 (br s, 1H),
8.22 (br s, 1H), 8.13 (s, 1H), 7.86–7.72 (m, 2H), 7.59 (t, J = 8.1
Hz, 1H), 7.24 (br s, 2H), 5.94 (d, J = 5.6 Hz, 1H), 5.56 (d, J = 5.9
Hz, 1H), 5.38 (d, J = 5.1 Hz, 1H), 4.71 (q, J = 5.5 Hz, 1H), 4.32–
4.21 (m, 1H), 4.14–3.99 (m, 2H), 3.97–3.83 (m, 1H); 13C NMR
(75.4 MHz, DMSO-d6) δppm 184.4, 180.4, 169.7, 163.0, 157.9 (q,
J = 30.9 Hz), 156.1, 152.7, 149.4, 148.6, 140.6, 139.9, 130.7,
123.9, 119.2, 117.3 (q, J = 300.2 Hz), 116.6, 112.3, 87.6, 83.3,
72.8, 70.8, 45.9; HRMS-ESI m/z calcd C20H19N8O7 483.1377
[M+H]+, found 483.1378.
1H), 7.82 (s, 1H), 7.72 (d, J = 7.6 Hz, 2H), 7.53 (t, J = 6.7 Hz,
2H), 7.36 (t, J = 7.6 Hz, 2H), 7.30–7.21 (m, 2H), 5.93–5.86 (m,
3H), 5.79 (d, J = 7.3 Hz, 1H), 5.35 (dd, J = 6.2, 4.1 Hz, 1H), 4.89
(dd, J = 6.2, 2.5 Hz, 1H), 4.44 (s, 1H), 4.37 (d, J = 7.0 Hz, 2H),
4.15 (t, J = 7.1 Hz, 1H), 3.92 (s, 1H), 3.74 (dd, J = 8.9, 3.9 Hz,
1H), 3.52 (d, J = 7.5 Hz, 1H), 3.48 (s, 2H), 1.61 (s, 3H), 1.35 (s,
3H), 1.10 (s, 9H); 13C NMR (75.4 MHz, CDCl3) δppm 171.2,
156.0, 153.5, 149.2, 144.1, 143.9, 141.4, 140.3, 127.8, 127.8,
127.2, 127.1, 125.2, 121.0, 120.1, 114.8, 92.0, 84.2, 83.2, 81.9,
74.0, 67.3, 62.3, 55.5, 47.3, 41.6, 27.7, 27.5, 25.6.; HRMS-ESI
m/z calcd C35H42N7O7 672.3146 [M+H]+, found 672.3146.
4.2.12. 3-(5´-Amino-5´-deoxyadenosine)-4-[(4-
methoxybenzyl)amino]cyclobut-3-ene-1,2-dione
(26)
4.2.15. tert-Butyl (S)-3-[[[(9H-fluoren-9-
yl)methoxy]carbonyl]amino]-4-[[[(3aS,4S,6S,6aS)-
6-(6-amino-9H-purin-9-yl)-2, 2-
A suspension of 18 (0.10 g, 0.20 mmol) in TFA/water (4:1, 3
mL) was stirred at rt for 5 h. The solvent was evaporated in
vacuo, and the residue was co-evaporated with toluene to remove
any residual TFA and H2O. The remaining residue was purified
by column chromatography on silica gel (CHCl3/MeOH 4:1) to
give compound 26 as a white solid (0.080 g, 0.17 mmol, 85%
yield). Rf = 0.38 (CHCl3/MeOH, 4:1). 1H NMR (300 MHz,
DMSO-d6) δppm 8.31 (s, 1H), 8.15 (s, 1H), 7.86 (br s, 1H), 7.57
(br s, 1H), 7.27 (s, 2H), 7.22 (d, J = 8.3 Hz, 2H), 6.89 (d, J = 8.3
Hz, 2H), 5.91 (d, J = 5.9 Hz, 1H), 5.52 (d, J = 5.9 Hz, 1H), 5.33
(d, J = 4.2 Hz, 1H), 4.73–4.64 (m, 1H), 4.61 (d, J = 5.9 Hz, 2H),
4.19–4.12 (m, 1H), 4.04–3.97 (m, 1H), 3.98–3.88 (m, 1H), 3.83–
3.74 (m, 1H), 3.72 (s, 3H); 13C NMR (75.4 MHz, DMSO-d6) δppm
182.6, 182.4, 167.9, 167.4, 158.6, 156.1, 152.7, 149.4, 139.8,
130.8, 128.9, 119.2, 114.0, 87.4, 83.6, 72.8, 70.8, 55.1, 46.2,
45.5; HRMS-ESI m/z calcd C22H24N7O6 482.1788 [M+H]+, found
482.1785.
dimethyltetrahydrofuro[3,4-d][1,3]dioxol-4-
yl]methyl]amino]-4-oxobutanoate (30)
To a solution of 3 (84 mg, 0.27 mmol) in anhydrous MeOH (10
mL) was added Fmoc-L-Asp(OtBu)-OSu (140 mg, 0.27 mmol).
The reaction mixture was stirred at rt for 2 h. The solvents were
removed in reduced pressure and the residue was purified by
column chromatography on silica gel (CHCl3/MeOH 10:1) to
give compound 30 as a white solid (160 mg, 0.22 mmol, 75%
yield). The product was used in the next step without further
purification. Rf = 0.50 (CHCl3/MeOH, 10:1). 1H NMR (300
MHz, CDCl3) δppm 8.56 (s, 1H), 8.43 (br s, 1H), 7.81 (s, 1H), 7.72
(d, J = 7.6 Hz, 2H), 7.54 (d, J = 7.4 Hz, 2H), 7.36 (t, J = 7.7 Hz,
2H), 7.31–7.19 (m, 2H), 6.11 (br s, 2H), 5.98 (d, J = 8.7 Hz, 1H),
5.83 (d, J = 4.5 Hz, 1H), 5.35 (t, J = 5.4 Hz, 1H), 4.89 (dd, J =
6.4, 2.3 Hz, 1H), 4.63 (br s, 1H), 4.45 (s, 1H), 4.38 (d, J = 7.3
Hz, 2H), 4.17 (t, J = 7.1 Hz, 1H), 4.03 (br s, 1H), 3.39 (d, J =
14.7 Hz, 1H), 2.90–2.73 (m, 2H), 2.10 (br s, 1H), 1.60 (s, 3H),
1.40 (s, 9H), 1.35 (s, 3H); 13C NMR (75.4 MHz, CDCl3) δppm
171.2, 170.3, 156.1, 156.0, 153.6, 149.1, 143.9, 143.8, 141.4,
141.4, 140.3, 127.8, 127.2, 125.2, 120.9, 120.1, 114.7, 92.3, 83.8,
82.8, 81.8, 81.7, 67.3, 51.9, 47.3, 41.5, 38.3, 28.2, 27.6, 25.5;
HRMS-ESI m/z calcd C36H42N7O8 700.3095 [M+H]+, found
700.3093.
4.2.13. 3-(5´-Amino-5´-deoxyadenosine)-4-[(4-
methylbenzyl)amino]cyclobut-3-ene-1,2-dione (28)
A suspension of 20 (0.10 g, 0.20 mmol) in TFA/water (4:1, 3 mL)
was stirred at rt for 5 h. The solvent was evaporated in vacuo, and
the residue was co-evaporated with toluene to remove any
residual TFA and H2O. The remaining residue was purified by
column chromatography (CHCl3/MeOH 4:1, KP-NH column) to
give compound 28 as solid (80 mg, 0.17 mmol, 85% yield). Rf =
0.18 (CH2Cl2/MeOH, 4:1, NH2 TLC). 1H NMR (300 MHz,
DMSO-d6) δppm 8.32 (s, 1H), 8.16 (s, 1H), 7.74 (br s, 1H), 7.45
(br s, 1H), 7.33 (s, 2H), 7.16 (s, 4H), 5.92 (d, J = 5.9 Hz, 1H),
5.57 (br s, 1H), 5.39 (br s, 1H), 4.75–4.54 (m, 3H), 4.22–4.10 (m,
1H), 4.08–3.90 (m, 2H), 3.88–3.69 (m, 1H), 2.26 (s, 3H); 13C
NMR (75.4 MHz, DMSO-d6) δppm 182.7, 182.6, 167.8, 167.5,
156.1, 152.8, 149.4, 139.9, 136.6, 135.8, 129.2, 127.5, 119.3,
87.5, 83.7, 72.8, 70.9, 46.6, 45.6, 20.7; HRMS-ESI m/z calcd
C22H24N7O5 466.1839 [M+H]+, found 466.1840.
4.2.16. (9H-Fluoren-9-yl)methyl [(S)-1-
[[[(2S,3S,4S, 5S)-5-(6-amino-9H-purin-9-yl)-3,4-
dihydroxytetra-hydrofuran-2-yl]methyl]amino]-3-
hydroxy-1-oxopropan-2-yl]carbamate (31)
A solution of 29 (94 mg, 0.14 mmol) in TFA/H2O/CH2Cl2 (4:1:1,
5 mL) was stirred at rt for 24 h. The solvent was evaporated in
vacuo, and the residue was co-evaporated with toluene to remove
residual TFA and H2O. The crude product was purified by
column chromatography on silica gel (CH2Cl2/MeOH 0→15%)
to give 31 as a white solid (56 mg, 0.10 mmol, 70% yield). Rf =
1
0.21 (CHCl3/ MeOH, 10:1). H NMR (400 MHz, DMSO-d6) δppm
8.44 (t, J = 5.9 Hz, 1H), 8.35 (s, 1H), 8.28 (s, 1H), 8.25 (d, J =
3.1 Hz, 1H), 7.88 (d, J = 7.5 Hz, 2H), 7.73 (t, J = 7.4 Hz, 2H),
7.50–7.26 (m, 6H), 5.84 (d, J = 6.6 Hz, 1H), 5.33 (br d, J = 55.9
Hz, 1H), 4.89 (br s, 1H), 4.67 (t, J = 5.8 Hz, 1H), 4.38–3.97 (m,
6H), 3.69–3.48 (m, 4H), 3.42–3.31 (m, 1H); 13C NMR (101
MHz, DMSO-d6) δppm 170.5, 156.0, 151.1, 149.0, 143.9, 143.8,
140.9, 140.7, 127.6, 127.1, 125.3, 120.1, 119.4, 87.8, 83.8, 72.8,
71.1, 65.8, 61.7, 57.3, 46.6, 41.1; HRMS-ESI m/z calcd for
C28H30N7O7 576.2207 [M+H]+, found 576.2205.
4.2.14. (9H-Fluoren-9-yl)methyl [(S)-1-
[[[(3aS,4S,6S,6aS)-6-(6-amino-9H-purin-9-yl)-2,2-
dimethyltetra-hydrofuro[3,4-d][1,3]dioxol-4-
yl]methyl]amino]-3-(tert-butoxy)-1-oxopropan-2-
yl]carbamate (29)
To a solution of 3 (0.10 g, 0.33 mmol) in anhydrous MeOH (10
mL) was added Fmoc-L-Ser(tBu)-OSu (160 mg, 0.33 mmol). The
reaction mixture was stirred at rt for 2.5 h under argon
atmosphere. The solvents were removed in vacuo, and the residue
was purified by column chromatography on silica gel
(EtOAc/EtOH/n-heptane 3:1:4) to give compound 29 as a white
solid (154 mg, 0.23 mmol, 70% yield). Rf = 0.65 (CHCl3/MeOH,
4.2.17. (S)-3-[[[(9H-Fluoren-9-
yl)methoxy]carbonyl]amino]-4-[[[(2S,3S,4S, 5S)-5-
(6-amino-9H-purin-9-yl)-3,4-
1
10:1). H NMR (300 MHz, CDCl3) δppm 8.49 (s, 1H), 8.11 (s,

dihydroxytetrahydrofuran-2-yl]methyl]amino]-4-
oxobutanoic acid (33)
The cell suspension was thawed, 10 µg/mL DNAase, 0.1 mM
Pefabloc, and 0.25 mg lysozyme was added followed by
incubation on ice for 5 min. The cells were disrupted by
sonication and the lysate was cleared by centrifugation (35000×g,
30 min, 4 °C). The supernatant was filtered using a 0.45-µM
filter syringe before it was loaded to two 1-mL HisTrap HP
columns (GE Healthcare, UK), which were pre-charged with Ni2+
and pre-equilibrated with binding buffer (20 mM HEPES pH 7.5,
500 mM NaCl, 10% glycerol, 10 mM imidazole, and 0.5 mM
TCEP). The bound protein was washed with 5 column volumes
(10 mL) of buffer supplemented with 25 mM imidazole and
eluted with buffer containing 350 mM imidazole. The eluted
protein fraction was concentrated and diluted in size exclusion
buffer (20 mM HEPES, pH 7.5, 250 mM NaCl, 5% glycerol, and
1 mM TCEP) to a final volume of 10 mL to bring the imidazole
concentration down to 50 mM. TEV protease was added (1:30
molar ratio) and the sample was incubated overnight at 4 °C. The
cleaved protein was run again through a 1-mL HisTrap HP
column (GE Healthcare, UK). The flow-through was collected
and concentrated to perform size exclusion chromatography
using Superdex S-75 column (GE Healthcare, UK). The collected
fractions were pooled, concentrated and divided into small
aliquots which were flash frozen in liquid nitrogen and stored at -
80 °C. Protein purity was analyzed with an SDS-PAGE
(Supplementary Material S3).
Compounds were tested as potential MDO1 inhibitors using a
biochemical assay. A mono-biotin-ADP-ribosylated SRPK2 was
used as a substrate for the hydrolysis as it can be robustly ADP-
ribosylated by ARTD10.29 A mixture of 2 µM substrate, 2 µM
MDO1 and 0.8-1 mM compound was incubated in 25 mM
HEPES pH 7.5, 100 mM NaCl for 16 hours. The reactions were
stopped by adding 2× Laemmli buffer (Bio-Rad, USA) and
incubating at 95 °C for 5 min. The samples were analyzed by
SDS-PAGE and transferred onto a nitrocellulose membrane
(Whatman, UK), which was blocked using 1% casein in 1× TBS
buffer (Bio-Rad, USA). Finally, the samples were visualized with
1:15000 streptavidin conjugated horseradish peroxidase
A mixture of 30 (120 mg, 0.17 mmol) in TFA/H2O/CH2Cl2
(5:1:4, 5 mL) was stirred at rt for 5 h. The solvent was removed
in vacuo and the resulting residue was co-evaporated with
toluene to remove any residual TFA and H2O. The crude product
was purified by column chromatography on silica gel
(MeOH/CHCl3 4:1) to give compound 33 as a white solid (110
mg, 0.19 mmol, 90% yield). The product was used in the next
step without further purification. Rf = 0.63 (MeOH/CHCl3, 4:1).
1H NMR (300 MHz, DMSO-d6) δppm 12.27 (br s, 1H), 8.38 (app t,
1Hz), 8.32 (s, 1H), 8.25 (s, 1H), 7.88 (d, J = 7.5 Hz, 2H), 7.70 (d,
J = 7.4 Hz, 2H), 7.41 (t, J = 7.4 Hz, 2H), 7.36–7.25 (m, 4H), 5.83
(d, J = 6.5 Hz, 1H), 5.38 (br s, 1H), 5.21 (br s, 1H), 4.73–4.65
(m, 1H), 4.46–4.37 (m, 1H), 4.30–4.15 (m, 4H), 4.04–3.97 (m,
2H), 3.58–3.26 (m, 2H), 2.67 (dd, J = 16.4, 5.1 Hz, 1H), 2.56 (d,
J = 8.9 Hz, 1H).; 13C NMR (75.4 MHz, DMSO-d6) δppm 171.6,
171.1, 155.8, 154.1, 150.0, 148.8, 143.8, 143.7, 141.2, 140.6,
127.6, 127.0, 125.3, 120.0, 119.3, 87.9, 83.7, 72.9, 71.1, 65.8,
51.5, 46.6, 41.2, 36.4; HRMS-ESI m/z calcd for C29H30N7O8
604.2156 [M+H]+, found 604.2164.
4.2.18. (S)-2-Amino-N-[[(2S,3S,4S,5S)-5-(6-amino-
9H-purin-9-yl)-3,4-dihydroxytetrahydrofuran-2-
yl]-methyl]-3-hydroxypropanamide (34)
A mixture of 31 (0.037 g, 0.064 mmol) and DEA (7.5 µL,
0.07 mmol) in DMF (0.50 mL) was stirred at rt for 5 h. The
solvent was removed in vacuo to give a residual syrup that was
purified by column chromatography (CHCl3/MeOH 0→15%,
KP-NH column) to give compound 34 as white crystals (6.0 mg,
0.017 mmol, 27 % yield). 1H NMR (400 MHz, CD3OD) δppm 8.28
(s, 1H), 8.25 (s, 1H), 5.94 (d, J = 5.9 Hz, 1H), 4.80 (t, J = 5.7 Hz,
1H), 4.23 (dd, J = 5.5, 3.7 Hz, 1H), 4.20–4.15 (m, 1H), 3.79 (dd,
J = 14.2, 5.2 Hz, 1H), 3.74–3.63 (m, 2H), 3.50 (dd, J = 14.2, 4.2
Hz, 1H), 3.45 (t, J = 5.6 Hz, 1H); 13C NMR (101 MHz, CD3OD)
δppm 175.9, 157.5, 153.9, 150.5, 142.0, 120.9, 90.6, 85.1, 74.7,
72.7, 65.5, 58.0, 42.2; HRMS-ESI m/z calcd for C13H20N7O5
354.1526 [M+H]+, found 354.1526.
(PerkinElmer, USA) using
a chemiluminescent substrate
(WesternBrightTM ECL, Advansta Corporation, USA).
Dose-response measurements were performed using the
previously reported Alphascreen-based activity assay for mono-
ADP-ribosyl hydrolyzing macrodomains.26 Compounds were
measured in quadruplicates from 6.6 mM to 0.66 µM using half-
logarithmic dilutions. After transfer of compounds to the assay
plate, MDO1 and MDO2 (final concentration 800 nM for both)
were added to the plate followed by the addition of a mono-ADP-
ribosylated SRPK2 (final concentration 50 nM). The plate was
incubated for 1 h 20 min at room temperature, followed by the
addition of the acceptor and donor bead mixture (final
concentration 5 µg/ml) and an incubation of 3 h at room
temperature in dark. The luminescence was read with a Tecan
Infinite M1000 Pro plate reader using the AlphaScreen detection
module with 600 ms excitation and 300 ms integration time.
4.3. Biochemical assays
The expression construct for human MDO1 (residues 58-325)
with an N-terminal 6×His tag and a thioredoxin fusion tag
followed by a TEV cleavage site (pNH-TrxT) was a generous gift
from the Structural Genomics Consortium, Oxford. DNA for
MDO2 was a kind gift from Dr. Ahola (University of Helsinki,
Finland). MDO2 (residues 7-243) was cloned to pNIC-Bsa4
vector also containing a TEV cleavage site. MDO1 and MDO2
were expressed by transforming the plasmids into E. coli
Rosetta2 (DE3) cells. The purification protocol was similar for
both enzymes. Glycerol stock was used to inoculate small 5-mL
pre cultures grown overnight in Luria broth (LB) media which
were then used to inoculate two large 1-L cultures containing
Terrific broth (TB) auto induction media with trace elements
(Formedium, UK). The media was additionally supplemented
with 8 g/L glycerol. The cultures were allowed to grow up to
OD600 of 1.5 at 37 °C with shaking after which the temperature
was reduced to 18 °C and the cultures were allowed to grow
overnight (14-16 h). The cells were then collected by
centrifugation (5500×g, 25 min, 4 °C), resuspended in lysis
buffer (50 mM HEPES, pH 7.5, 500 mM NaCl, 10% glycerol, 10
mM imidazole, and 0.5 mM TCEP; 1.5 mL/g of cells) and stored
at -20 °C.
4.4. Molecular modeling
Preliminary dockings by us and others (see Supplementary
Material S4) has shown that the MDO1 structure in its apo form
is not able to accommodate ligands. We thus relied on a three-
dimensional structure where the binding site had been made to
mimic the close homolog MDO2 bound to ADP-ribose. For this
purpose, the human MDO1 protein was remodeled using the
online webserver SWISS-MODEL.30 The best model was built
based on human MOD2 structure (PDB file: 4IQY) with 56.33%
sequence identity. Before docking, the modeled human MDO1
protein structure (PDB code: 2X47) was handled with protein

8.
Tian L, Wu Z, Zhao Y, et al. Differential induction of LRP16 by
preparation wizard (Schrödinger Ltd). The protein was assigned
atom type, side chain protonation states, and hydrogen network
using PROPKA at pH 7.0. The OPLS3 force field was applied to
minimize the protein. Glide receptor grids were generated by
defining a 10 Å box localized at the centroid of ADP-ribose
defined using the superimposed ADP-ribose-MDO2 protein
complex. H-bond constraint to Asp160 was applied during grids
generation, which will guarantee adenine stacked by Phe306 and
Val183.
All ligands listed in table 1 and 2 were drawn in 2D Sketcher,
then converted into 3D. The possible ionization states of ligand
were generated at pH 7.0+⁄-2.0 by LigPrep Suite.
The molecular docking experiments were carried out with
Maestro suite Glide and the Glide XP protocol.31 Default
parameters were used. Van der Waals scale factor was 1.0, and
partial charge cutoff was 0.25. For each ligand, the maximum 10
docking poses were output. The one with lowest energy was
suggested as most likely binding pose.
liganded and unliganded estrogen receptor α in SKOV3 ovarian carcinoma
cells. J Endocrinol. 2009;202(1):167-177.
9.
prognostic value of leukemia‐related protein 16 expression in invasive
ductal breast carcinoma. Cancer science. 2010;101(10):2262-2268.
10.
monocytic leukemia cell line with t (11; 21)(q13; q22). European journal of
haematology. 2007;79(1):25-31.
11.
LRP16 is an essential cofactor of androgen receptor. Endocrine-related
cancer. 2009;16(1):139-153.
12.
Zhao P, Lu Y, Han W. Clinicopathological significance and
Imagama S, Abe A, Suzuki M, et al. LRP16 is fused to RUNX1 in
Yang J, Zhao Y, Wu Z, et al. The single-macro domain protein
Vyas S, Chang P. New PARP targets for cancer therapy. Nature
Reviews Cancer. 2014;14(7):502-509.
13.
is an ADP‐ribose binding module. The EMBO journal. 2005;24(11):1911-
1920.
Karras GI, Kustatscher G, Buhecha HR, et al. The macro domain
14.
Yao Z-J, King CR, Cao T, et al. Potent inhibition of Grb2 SH2
domain binding by non-phosphate-containing ligands. J Med Chem.
1999;42(1):25-35.
15.
Wagner GK, Pesnot T, Field RA. A survey of chemical methods
for sugar-nucleotide synthesis. Natural product reports. 2009;26(9):1172-
1194.
16.
Acknowledgments
Seio K, Miyashita T, Sato K, Sekine M. Synthesis and properties
of new nucleotide analogues possessing squaramide moieties as new
phosphate isosters. Eur J Org Chem. 2005;2005(24):5163-5170.
This study was supported by grants from the Academy of
Finland (Project 257685 for GBG, 287063 and 294085 for LL
and MM), China Scholarship Council (Grant no. 2009629110),
University of Helsinki, Biocenter Oulu and Sigrid Jusélius
Foundation. Y.Z. and H.V. thank the Doctoral Program in
Informational and Structural Biology for organizing graduate
studies. We thank Miikka Olin for his help with the HPLC. Y.Z.
thank Mark Johnson for access to computational resources. The
Drug Discovery and Chemical Biology (DDCB) Biocenter
Finland network and the IT-Center for Science (CSC, Espoo,
Finland) are thanked for providing computational resources.
17.
synthesis and applications. Chem Soc Rev. 2011;40(5):2330-2346.
18. Zhang Y, Borrel A, Ghemtio L, et al. Structural Isosteres of
Storer RI, Aciro C, Jones LH. Squaramides: physical properties,
Phosphate Groups in the Protein Data Bank. Journal of Chemical Information
and Modeling. 2017;57(3):499-516.
19.
Mutisya D, Selvam C, Lunstad BD, et al. Amides are excellent
mimics of phosphate internucleoside linkages and are well tolerated in short
interfering RNAs. Nucleic Acids Res. 2014;42(10):6542-6551.
20.
Elliott TS, Slowey A, Ye Y, Conway SJ. The use of phosphate
bioisosteres in medicinal chemistry and chemical biology. MedChemComm.
2012;3(7):735-751.
21.
Lassalas P, Gay B, Lasfargeas C, et al. Structure property
relationships of carboxylic acid isosteres. J Med Chem. 2016;59(7):3183-
3203.
22.
Supplementary Material
Matthew A. Tetrazolone as an acid bioisostere: application to
marketed drugs containing a carboxylic acid. Organic & biomolecular
chemistry. 2016;14(39):9343-9347.
Supplementary data associated with this article can be found,
in the online version, at http://dx.doi.org/10.1016/j.bmc.201X.
XX.XXX. These data include the synthesis procedures of
compounds 5-7, 11-15, 19, 21-23, 27, 36-39, and the NMR
spectra of compounds as well as the abbreviations. Figures:
Figure S1, Dose-response measurements; Figure S2, three-
dimensional fold and structural differences between MDO2 and
MDO1; Figure S3, purification of recombinant MDO1; Figure
S4, docking to apo MDO1.
23.
Niewiadomski S, Beebeejaun Z, Denton H, Smith TK, Morris RJ,
Wagner GK. Rationally designed squaryldiamides–a novel class of sugar-
nucleotide mimics? Organic & biomolecular chemistry. 2010;8(15):3488-
3499.
24.
Niesen FH, Berglund H, Vedadi M. The use of differential
scanning fluorimetry to detect ligand interactions that promote protein
stability. Nature protocols. 2007;2(9):2212-2221.
25.
Neuvonen M, Ahola T. Differential activities of cellular and viral
macro domain proteins in binding of ADP-ribose metabolites. J Mol Biol.
2009;385(1):212-225.
26.
Haikarainen T, Maksimainen MM, Obaji E, Lehtiö L.
Development of an Inhibitor Screening Assay for Mono-ADP-Ribosyl
Hydrolyzing Macrodomains Using AlphaScreen Technology. SLAS
DISCOVERY: Advancing Life Sciences R&D. 2017:1-9.
References and notes
1.
Till S, Ladurner AG. Sensing NAD metabolites through macro
27.
After the western blot analysis of the compounds had been
domains. Front Biosci. 2009;14:3246-3258.
2.
functions, and their potential therapeutic implications. Mutation
Research/Reviews in Mutation Research. 2011;727(3):86-103.
3.
containing proteins: regulating new intracellular functions of mono (ADP-
ribosyl) ation. Nature Reviews Molecular Cell Biology. 2013;14(7):443-451.
4.
performed, we recognized that compound 35 (the Fmoc-deprotected
derivative of compound 33) contained small amount of impurities. Purity
determinations showed impure compound even after extensive purifications
of the compound, and therefore, we made a late stage decision to remove it
from the manuscript.
Han W, Li X, Fu X. The macro domain protein family: structure,
Feijs KL, Forst AH, Verheugd P, Lüscher B. Macrodomain-
28.
Ikeuchi H, Meyer ME, Ding Y, Hiratake J, Richards NG. A
critical electrostatic interaction mediates inhibitor recognition by human
asparagine synthetase. Biorg Med Chem. 2009;17(18):6641-6650.
29.
Chemical Probe Rescues Cells from Mono-ADP-Ribosyltransferase
ARTD10/PARP10-Induced Apoptosis and Sensitizes Cancer Cells to DNA
Damage. Cell chemical biology. 2016;23(10):1251-1260.
Malet H, Coutard B, Jamal S, et al. The crystal structures of
Chikungunya and Venezuelan equine encephalitis virus nsP3 macro domains
define a conserved adenosine binding pocket. Journal of virology.
2009;83(13):6534-6545.
Venkannagari H, Verheugd P, Koivunen J, et al. Small-Molecule
5.
Rosenthal F, Feijs KL, Frugier E, et al. Macrodomain-containing
proteins are new mono-ADP-ribosylhydrolases. Nature structural &
molecular biology. 2013;20(4):502-507.
6.
macrodomain proteins as novel O-acetyl-ADP-ribose deacetylases. J Biol
Chem. 2011;286(15):13261-13271.
30.
Biasini M, Bienert S, Waterhouse A, et al. SWISS-MODEL:
modelling protein tertiary and quaternary structure using evolutionary
Chen D, Vollmar M, Rossi MN, et al. Identification of
information. Nucleic Acids Res. 2014;42(W1):W252-W258.
31.
Friesner RA, Murphy RB, Repasky MP, et al. Extra precision
glide: Docking and scoring incorporating a model of hydrophobic enclosure
7.
Jankevicius G, Hassler M, Golia B, et al. A family of
for protein− ligand complexes. J Med Chem. 2006;49(21):6177-6196.
macrodomain proteins reverses cellular mono-ADP-ribosylation. Nature
structural & molecular biology. 2013;20(4):508-514.