Rationally designed squaryldiamides - A novel class of sugar-nucleotide mimics?

Article abstract of DOI:10.1039/c004165c

Sugar-nucleotides such as GDP-mannose, GDP-fucose and UDP-glucose are important biomolecules with a central role in carbohydrate and glycoconjugate biosynthesis, metabolism and cell signalling. Analogues and mimics of naturally occurring sugar-nucleotides are sought after as chemical tools and inhibitor candidates for sugar-nucleotide-dependent enzymes including glycosyltransferases. Many sugar-nucleotides bind to their target glycosyltransferases via coordination of the diphosphate group to a divalent metal cofactor in the active site. The identification of uncharged, chemically stable surrogates for the diphosphate group, with the ability to coordinate to a divalent metal, is therefore an important design criteria for the development of sugar-nucleotide mimics. Here, we describe the rational design and synthesis of a novel class of sugar-nucleotide mimics based on a squaryldiamide scaffold, an uncharged phosphate isostere. We demonstrate by comprehensive NMR titration experiments that the new sugar-nucleotide mimics coordinate efficiently to Mg²⁺, and provide results from biological studies with a therapeutically relevant mannosyltransferase from Trypanosoma brucei. Our findings suggest that squaryldiamides are a promising template for the development of sugar-nucleotide mimics, and illustrate the considerable potential of the squarylamide group as a fragment for inhibitor design. The Royal Society of Chemistry 2010.

Full text of DOI:10.1039/c004165c

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/obc | Organic & Biomolecular Chemistry
Rationally designed squaryldiamides – a novel class of sugar-nucleotide
mimics?†
Sven Niewiadomski,a,c Zeenat Beebeejaun,a Helen Denton,b Terry K. Smith,b Richard J. Morrisc and
Gerd K. Wagner*a
Received 16th March 2010, Accepted 12th May 2010
First published as an Advance Article on the web 7th June 2010
DOI: 10.1039/c004165c
Sugar-nucleotides such as GDP-mannose, GDP-fucose and UDP-glucose are important biomolecules
with a central role in carbohydrate and glycoconjugate biosynthesis, metabolism and cell signalling.
Analogues and mimics of naturally occurring sugar-nucleotides are sought after as chemical tools and
inhibitor candidates for sugar-nucleotide-dependent enzymes including glycosyltransferases. Many
sugar-nucleotides bind to their target glycosyltransferases via coordination of the diphosphate group to
a divalent metal cofactor in the active site. The identification of uncharged, chemically stable surrogates
for the diphosphate group, with the ability to coordinate to a divalent metal, is therefore an important
design criteria for the development of sugar-nucleotide mimics. Here, we describe the rational design
and synthesis of a novel class of sugar-nucleotide mimics based on a squaryldiamide scaffold, an
uncharged phosphate isostere. We demonstrate by comprehensive NMR titration experiments that the
new sugar-nucleotide mimics coordinate efficiently to Mg2+, and provide results from biological studies
with a therapeutically relevant mannosyltransferase from Trypanosoma brucei. Our findings suggest
that squaryldiamides are a promising template for the development of sugar-nucleotide mimics, and
illustrate the considerable potential of the squarylamide group as a fragment for inhibitor design.
2,3
transferases
and other glycoprocessing enzymes. These en-
Introduction
zyme classes have been identified as promising drug targets in
various therapeutic areas, including infection,4,5 inflammation,6,7
and cancer.7 Individual sugar-nucleotides also act as agonists
at the human P2Y14 receptor which has an important role in
immune function.8 Structural analogues and mimics of natu-
rally occuring sugar-nucleotide are therefore of significant in-
terest as candidate molecules for enzyme inhibitors, receptor
ligands and chemical tools in medicinal chemistry and chemical
biology.
Structurally, sugar-nucleotides (exemplified by GDP-mannose,
Fig. 1) are composed of a sugar, or sugar derivative, and a
nucleotide fragment. In mammalian cells, nine different sugar-
nucleotides have been identified, but this number is much larger in
other organisms.1 In many cases the diphosphate linkage is essen-
tial for the interaction of the sugar-nucleotide with its target pro-
tein and thus for its biological activity. However, the diphosphate
Sugar-nucleotides1 such as GDP-mannose (Fig. 1), GDP-fucose
and UDP-glucose are important biomolecules with a central
role in carbohydrate and glycoconjugate biosynthesis, metabolism
and cell signalling.1 In both prokaryotes and eukaryotes, sugar-
nucleotides serve as donor substrates for Leloir-type glycosyl-
aSchool of Pharmacy, University of East Anglia, Norwich, UK NR4 7TJ.
E-mail: g.wagner@uea.ac.uk; Fax: +44(0)1603 592003; Tel: +44 (0)1603
593889
bCentre for Biomolecular Sciences, The North Haugh, The University, St.
Andrews, KY16 9ST, Scotland
cDepartment of Computational & Systems Biology, John Innes Centre,
Norwich, UK NR4 7UH
† Electronic supplementary information (ESI) available: 1H-NMR spectra
1
for compounds 3, 4, 5a-g and 6a-g, H-NMR titration curves for 5c and
5g, and 1H-NMR spectra for the titration of 5g with MgCl2. See DOI:
10.1039/c004165c
Fig. 1 Molecular structures of GDP-mannose (left) and of a representative squarylamide targeted in this study. Both the diphosphate group in
GDP-mannose and the squaric acid moiety in the target squaryldiamides can coordinate to a divalent metal.
3488 | Org. Biomol. Chem., 2010, 8, 3488–3499
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
group also contributes to the limited stability and low membrane
permeability of sugar-nucleotides, unfavourable properties which
complicate their use as biological tools and therapeutic agents.9
In order to develop sugar-nucleotide mimics with improved
biological activity, chemical stability and membrane permeability,
many efforts have therefore been directed at the identification
of uncharged, chemically stable surrogates for the diphosphate
linkage in sugar-nucleotides.10,11 Functional groups that have
been examined as potential diphosphate bioisosteres include
phosphonates,12–17 malonates,18 malonodiamides,19 tartrates,18 vic-
inal alcohols,20,21 monosaccharides,18,22 amides,23 sulfonamides,23
and sulfamates.24,25 However, despite these considerable efforts
no generally applicable diphosphate bioisostere has so far been
identified.
Leloir-type glycosyltransferases catalyse the transfer of a sugar
(derivative) from the sugar-nucleotide donor to a suitable acceptor
substrate.2,3 In many cases, the diphosphate group of the donor
sugar-nucleotide is critically involved in orienting the donor
substrate in the correct, bioactive conformation, prior to the
transfer reaction.2,3 Glycosyltransferases of the GT-A family,
one of two major glycosyltransferase fold families, require the
presence of a divalent metal cation (typically Mn2+ or Mg2+) in the
active site for full catalytic activity.2,3 The catalytic metal mediates
sugar-nucleotide binding at the enzyme through simultaneous
interaction with the aspartate residue(s) of the so called DXD
motif characteristic for glycosyltransferases of the GT-A family
and the diphosphate group of the sugar-nucleotide (Fig. 1).2,3
Because of the importance of the sugar-nucleotide–metal in-
teraction for ligand binding in GT-A glycosyltransferases, the
capacity to coordinate to a divalent metal cation is an important
design criterion in the development of sugar-nucleotide mimics
as potential glycosyltransferase inhibitors. Interestingly, while for
many of the previously published sugar-nucleotide mimics such a
capacity for coordination to a divalent metal has been postulated,
this has rarely been proven experimentally.
Recently, the squaric acid group has been reported as
an uncharged mimic of the phosphate group in mono- and
oligonucleotides.26,27 However, the use of the squaric acid motif
as a mimic for the diphosphate linkage in sugar-nucleotides has
not previously been explored. To investigate this possibility, we
have rationally designed a novel class of sugar-nucleotide mimics
based on a squaric acid core as potential glycosyltransferase
inhibitors (Fig. 1). Herein, we describe the design and synthesis
of a series of novel squaryldiamides as potential mimics of GDP-
sugars such as GDP-mannose, the donor substrate of Leloir-type
mannosyltransferases. Importantly, we have also carried out NMR
studies to determine experimentally the capacity of these novel
squaryldiamides to coordinate to Mg2+, an important prerequisite
for binding at GT-A glycosyltransferases. To assess their biological
activity, we have evaluated these novel GDP-sugar mimics as
potential inhibitors of the GDP-mannose-dependent mannosyl-
transferase dolichol-phosphate mannose synthase (DPMS) from
Trypanosoma brucei, a validated target for the development of
novel anti-trypanosomal agents.28,29 The squarylamide derivatives
described herein are, to the best of our knowledge, the first
examples for GDP-sugar mimics based on the squaric acid
motif, and the main findings from this study may also have
implications for the design of inhibitors for other metal-dependent
glycosyltransferases.
Target design
The design of the GDP-sugar mimics targeted in this study was
informed by the following considerations (Fig. 1): Because of
its important contribution to binding, we decided to retain the
guanosine part of GDP-mannose. The diphosphate linkage was
to be replaced with a squaryldiamide group, which we expected
to coordinate to the catalytic metal in the DPMS active site.
Furthermore, we speculated that a suitable second substituent
at the cyclobutene-1,2-dione core, e.g. a benzylamine substituent
with a polar group at the ortho position (cf. 5c, Fig. 1), might
contribute to the interaction between inhibitor and metal through
the formation of a chelate complex.
Before embarking on the synthesis of the target squaryl-
diamides, we wanted to assess whether the GDP-mannose bind-
ing site of a Leloir mannosyltransferase could accommodate
these potential donor substrate mimics. We therefore carried
out molecular docking studies with the target squaryldiamides
and a representative mannosyltransferase, the fungal enzyme
Kre2p.30 Kre2p is one of only two GDP-mannose-dependent
mannosyltransferases whose structures had been solved at the
outset of this project.30,31 Like T. brucei DPMS, for which no crystal
structural is presently available, Kre2p requires a Mn2+ cofactor for
catalytic activity.30 The structure of Kre2p in complex with GDP,
released as a by-product during the mannosyltransferase reaction,
shows the nucleoside diphosphate bound to the donor binding
site mostly by van der Waals and stacking interactions as well as
key hydrogen bonds between and Asp161 and N1 and N2 of the
guanine base.30 In the binary Kre2p/GDP complex, Mn2+ forms
important interactions with the a and b phosphates of GDP and
Oe2 of Glu247.30 Kre2p belongs to the GT-A fold family of glyco-
syltransferases. Structural comparison with other GT-A enzymes
suggests that Glu247, Pro248 and Asp249 correspond to the so-
called DXD motif, which is required in GT-A glycosyltransferases
for the coordination of the divalent metal cofactor.30 However, in
the case of Kre2p, only Glu247, the first residue of the putative
DXD motif, interacts directly with the Mn2+ ion.
Molecular docking studies with squaryldiamides 5a–g indicated
that these novel GDP-sugar mimics can bind to the Kre2p sugar-
nucleotide binding site in a similar orientation as GDP. Docking
results for the representative squarylamide 5c show the guanine
base buried deep in the donor binding site and forming a hydrogen
bond with Asp161 (Fig. 2). Both the squaryldiamide core and the
2-nitrobenzyl substituent of 5c are in close proximity to the Mn2+
ion, which interacts with Glu247 in a similar fashion as in the
Kre2p/GDP complex. Squaryldiamide 5c coordinates to the metal
˚
cofactor via the oxygen atoms of the nitro group (distances: 1.6 A
˚
and 1.7 A) and the two squaryldiamide nitrogen atoms (distances:
˚
˚
3.9 A and 4.1 A), forming a chelate complex. We concluded from
these molecular docking results that the target squaryldiamides
can indeed adopt a similar geometry as GDP/GDP-mannose and
coordinate efficiently to the divalent metal co-factor in the donor
binding site of a mannosyltransferase. We therefore set out to test
these hypotheses experimentally
Chemistry
Asymmetrical squaryldiamides can be prepared by sequential re-
action of a squaric acid dialkylester with two different amines.26,27
This journal is
The Royal Society of Chemistry 2010
Org. Biomol. Chem., 2010, 8, 3488–3499 | 3489
©

View Article Online
Fig. 3 General synthetic strategy for target squaryldiamides (G: guanine;
R: ethyl or methyl; R¢: various substituents).
5¢-azidoguanosine 2, and the PPh3-mediated reduction of azide 2
to amine 3.
In the next step, strategy A (Fig. 3) required the preparation
of squarylmonoamide 4 from aminoguanosine 3 and a suitable
dialkylsquarate. Initially, we were concerned that this reaction
might be complicated by the formation of side products (e.g.
symmetrical or mixed squaryldiamides), potentially necessitating
the protection of the amino group in position 2 of the nucleobase.
Much to our delight we found that at room temperature in DMF,
the base-catalyzed reaction of diethylsquarate with 3 proceeds
exclusively at the aliphatic amino group in position 5¢. Despite a
prolonged reaction time of 92 h, no diamides were detected, and
no reaction was observed with the amino group in position 2. In
the 1H NMR spectrum of 4, the broad, exchangeable signal at d =
6.47 ppm, corresponding to the free amino group, confirms that
no reaction had occurred between this group and diethylsquarate.
This lack of reactivity of the amino group in position 2 is perhaps
not surprising, given the electron-withdrawing effect of the purine
ring system, which may contribute to the reduced nucleophilicity
of the amino group.
Fig. 2 Ligand-enzyme interactions of 5c in domain B of Kre2p as
predicted by molecular docking. Key: H-bonds – green, hydrophobic
contacts – red, ligand bonds – purple, non-ligand bonds – brown. Picture
created with LigPlot 47
Initially, we planned to install the aminoguanosine fragment,
which is present in all analogues, first, and the benzylamino
substituent, which is variable, second (Fig. 3, Strategy A). This
approach is advantageous in that it allows structural diversifi-
cation at the last step of the synthesis. Alternatively, the two
different substituents can also be introduced in the reverse order
(Fig. 3, Strategy B). For both synthetic pathways, 5¢-deoxy-5¢-
aminoguanosine 3 was required as a key synthetic building block.
3 was prepared in three steps from guanosine in 45% overall yield
via an adaptation of Dean’s procedure (Scheme 1).32 This reaction
sequence involved the synthesis of 5¢-deoxy-5¢-iodoguanosine 1
from guanosine via the selective substitution of the primary
alcohol with an iodo group, the conversion of 1 into 5¢-deoxy-
1
Interestingly, the H NMR spectrum of 4 showed two sets of
signals for the squarylamide proton (two triplets at, respectively,
d = 8.74 and 8.89 ppm), for both protons in position 5¢ of the
ribose, and for the protons of the ethyl group. The two sets
of signals can be attributed to the presence of 4 in two stable
rotameric forms, syn and anti (Fig. 4). In the syn conformation,
the squarylamide proton is facing towards the adjacent carbonyl
group, and in the anti conformation away from it. The existence
Scheme 1 Synthesis of key intermediates 3 and 4. Reagents and conditions: (i) I2, PPh3, imidazole, DMF, rt, 4 h, 65%; (ii) NaN3, DMF, 90 ◦C, 22 h,
76%; (iii) (a) PPh3, pyridine, rt, 5h; (b) NH3(aq), rt, 39 h, 92%; (iv) diethylsquarate, DIPEA, DMF, rt, 92 h, 80%.
3490 | Org. Biomol. Chem., 2010, 8, 3488–3499
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
Scheme 2 Synthesis of target squaryldiamides 5. Reagents and conditions: (i) R-NH2, DIPEA, DMF, rt, 1–7 days; (ii) R-NH2, DIPEA, EtOH or MeOH,
rt, 2–16 h; (iii) 3, DIPEA, DMF, rt, 5–7 days; (iv) aq. NH3 (35%), MeOH, 50 ◦C, 2 days. Substituents R: (a) n-butyl; (b) benzyl; (c) 2-NO2 benzyl; (d)
3-NO2 benzyl; (e) 4-NO2 benzyl; (f) carboxypropyl; (g) 2-CO2H benzyl.
were prepared in mostly excellent yields and with comparatively
short reaction times from diethylsquarate and the respective
primary amine (Scheme 2). In the subsequent step, the target
squaryldiamides 5c–g were obtained in good to excellent yields
from monoamides 6c–g and aminoguanosine 3. While these
reactions proceeded not significantly faster than the reaction of
4 with a primary amine, they yielded cleaner reaction products,
thus facilitating purification. Finally, squarylmonoamide 6c was
also reacted with aqueous ammonia to give the corrsponding
Fig. 4 Different rotamers of squarylamide 4.
squaryldiamide 7, which was required as a model compound for
the NMR titration experiments that were carried out next.
of stable rotamers has previously been described for other
squarylmonoamides.33,34 The interconversion of these rotamers
can be slow due to the delocalisation of the nitrogen lone pair
electrons with the cyclobutene-1,2-dione system and, as a result,
the restricted rotation about the squarylamide carbon-nitrogen
bond. Individual rotamers can also be stabilised by intramolecular
hydrogen bonding.33,34 The 1H NMR spectrum of 4 suggests that
in solution, this squarylamide exists as a 1 : 1 mixture of the syn
and anti rotamer. Only one set of signals is observed, however, for
the ribose protons 1¢–4¢ of 4, which indicates a similar orientation
and conformation of the ribose ring in both rotamers.
In order to obtain the target squaryldiamides 5, squaryl-
monoamide 4 was reacted with different amines (Scheme 2).
All amines required for the synthesis of the target squaryl-
diamides were commercially available, with the exception of 2-
aminomethylbenzoic acid, which was synthesised in two steps
according to Sun’s procedure.35 The reaction with squaryl-
monoamide 4 proceeded only sluggishly (1–2 days at room
temperature) with both benzylamine and n-butylamine (Scheme 2,
5a and 5b), and even more so with 2-nitrobenzylamine (5c; 7 days
at room temperature, followed by 4 days at 60 ◦C). The extremely
prolonged reaction time and the low isolated yield for 5c prompted
us to explore an alternative synthetic route for the preparation of
further analogues.
Coordination to Mg2+
In a series of squaryldiamide oligonucleotide mimics, a pro-
nounced downfield shift had previously been observed for the
1H NMR signals of the squarylamide protons upon addition
of MgCl2.26 This downfield shift had been attributed to the
coordination of the divalent metal ion to the squarylamide group
and a subsequent increase in acidity of the amide protons.26
We reasoned that this effect could be used as a diagnostic
marker to assess experimentally the predicted ability of the novel
squarylamides 5 to coordinate to divalent metal ions. To test
this hypothesis, we carried out comprehensive 1H NMR titration
experiments with three representative target compounds 5b, 5c
and 5g and MgCl2 in DMF-d7.
In the case of all three squaryldiamides 5b, 5c and 5g, upon
addition of MgCl2 a significant downfield shift was observed for
the signals corresponding to the two squarylamide protons, as
well as for those corresponding to the exchangeable NH and NH2
protons on the guanine base (Fig. 5 and ESI†). The representative
titration curve for 5b is shown in Fig. 5, with the chemical shift
values plotted against equivalents of MgCl2. Initially, the chemical
shift changes are very pronounced for all four signals, before a
plateau is reached at approximately two equivalents of MgCl2.
This suggests a 1 : 2 stoichiometry for the 5b–Mg2+ complex, with
one equivalent of Mg2+ coordinating to the squaryldiamide group,
To our satisfaction, we found that reversing the last two
reaction steps of our general synthesis offered a more practical
approach (Fig. 3, Strategy B). Thus, squarylmonoamides 6a–g
This journal is
The Royal Society of Chemistry 2010
Org. Biomol. Chem., 2010, 8, 3488–3499 | 3491
©

View Article Online
Fig. 5 NMR titration experiment with squarylamide 5b.
and a second equivalent in complex with the pyrimidine fragment
of the guanine base. The curves for the two squarylamide proton
signals show a particularly steep initial gradient, up to the addition
of approximately one equivalent of MgCl2, while the gradient for
the curves of the guanine NH and NH2 signals is more shallow.
The steeper curve for the squaryldiamide protons may suggest that
coordination of the first equivalent of Mg2+ occurs preferentially at
this position. This interpretation is supported by the equivalence
points determined for each titration. Extrapolation of the steep
initial and the shallow terminal sections of each curve results in
intersections, corresponding to the respective equivalence point,
at one equivalent of MgCl2 for the squaryldiamide protons, and
at two equivalents of MgCl2 for the guanine protons.
Fig. 6 NMR titration experiment with (a) squaryldiamide 7, and (b)
squarylmonoamide 6c.
To better understand the coordination behaviour of the squary-
lamide group, we carried out further titration experiments with
squarylmonoamide 6c and model squaryldiamide 7 (Fig. 6). We
anticipated that the formal removal of the guanosine fragment
in these two analogues would simplify the analysis of the titration
results. Also, information about the precise site at the cyclobutene-
1,2-dione core that is involved in coordination of the metal ion may
be obtained from direct comparison of 6c and 7. Previously, Sato
et al. have suggested for a model thymidine-based squaryldiamide
that Mg2+ coordinates to the oxygen rather than the nitrogen atoms
of the squaryldiamide group.26 In this case, 6c and 7 would be
expected to display very similar metal coordination behaviour.
Although the initial gradient is steeper, the titration curves for
the squarylamide protons in 7 correspond to those observed for
the squarylamide signals in 5b, suggesting a similar coordination
behaviour at the squarylamide group (Fig. 6a). In contrast to the
results for 7, titration of squarylmonoamide 6c with up to 10
equivalents of MgCl2 did not produce a stable endpoint for the
downfield shift of the amide proton signal, and no equivalence
point could be determined for this compound (Fig. 6b; the two
curves are for the amide proton of the two rotamers). The different
titration results for squaryldiamide 7 and monoamide 6c suggest
that efficient coordination of Mg2+ to the cyclobutene-1,2-dione
core requires the presence of two amide nitrogens, indicating that
the amide nitrogens rather than the two oxo groups are directly
involved in formation of the complex with the metal. Monoamide
6c may still be surrounded by several equivalents of MgCl2, as
evidenced by the downfield shift of the squarylamide proton
signals, but direct coordination to the metal appears to be relatively
weak. Importantly, coordination of the metal at the squarylamide
nitrogens also allows the ligand to adopt an orientation favourable
for binding to protein, as predicted by the results from molecular
docking (Fig. 2).
In summary, the results from the NMR titration experiments
with the novel squarylamides 5b, 5c, 5g, 6c and 7 suggested to
us that the diamides, if perhaps not the monoamides, coordinate
readily to a divalent metal such as Mg2+. With this feature, the
squaryldiamides meet, as expected, an important prerequisite for
effective diphosphate mimicry. Therefore, we next investigated if
the capacity for the coordination of divalent metals would translate
into biological activity towards a representative GDP-mannose-
dependent mannosyltransferase.
Biological results
The trypanosomal parasite Trypanosoma brucei is the causative
agent of African sleeping sickness (African trypanosomiasis),
which puts more than 60 million people at risk in sub-Saharan
Africa and led to 48,000 deaths in 2002 alone.36 Current treatments
for African trypanosomiasis are limited and problematic due to
toxicity, expense, administration routes and resistance, thus new
therapeutic approaches are urgently required. T. brucei has a cell-
surface protective coat of variant surface glycoproteins (VSG)
which the parasite uses to evade the immune system of the host.
The biosynthesis of glycosylphosphatidylinositol (GPI) anchors,
which link the VSG coat to the parasite membrane, has been
validated genetically and chemically as a therapeutic target.28,37,38
3492 | Org. Biomol. Chem., 2010, 8, 3488–3499
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
Table 1 Biological activity of squarylamides and reference compounds
IC50 value of 32 mM. The significant difference in activity between
guanosine on the one hand, and the guanine nucleotides and GDP-
mannose on the other illustrates the importance of the phosphate
groups for ligand binding, presumably through coordination of
the active site metal. Somewhat disappointingly, however, most
squarylamides showed only low to moderate inhibition of DPMS
activity at a concentration of 1 mM. Thus, the experimentally
demonstrated capacity of analogues 5b, 5c and 5g to coordinate
to Mg2+ did translate into DPMS inhibitory activity in vitro,
but to a lesser degree than expected. The most potent DPMS
inhibitor in this series is the 2-carboxybenzylamino derivative
5g, which also coordinates effectively to Mg2+. Interestingly, in
the nitrobenzyl series (5c–e), those analogues with the nitro
group in the meta (5d) or para position (5e) were slightly more
potent than 2-nitrobenzylamino derivative 5c. This suggests that
the contribution of a polar substituent in the ortho position to
stabilisation of the ligand-metal complex may be less relevant, at
least in vitro, than predicted by the Mg2+ titration experiments.
Analogues lacking the guanosine fragment (6b, 6c) are inactive,
which was expected, as this motif is known to be important for
biological activity.
We also evaluated squaryldiamide 5c for its inhibitory activity
against T.brucei GDP-Man pyrophosphorylase (GDP-Man PP).
GDP-Man catalyses the formation of GDP-Man from GTP and
mannose-1-phosphate and has been validated as a therapeutic
target.38 Interestingly, 5c acted as an inhibitor of GDP-Man PP
with an IC50 of 0.7 mM, suggesting that this squaryldiamide may
be able to mimic not only GDP-mannose, but also GTP binding.
Selected squarylamides were also tested for trypanocidal activity
against the bloodstream form of T. brucei in the Alamar Blue
cytotoxicity assay.40 Several analogues in this series showed activity
against live trypanosomes with ED50 values of, or below, 1 mM
(Table 1). The strongest effect in vivo was observed for benzylamino
derivative 5b, while 2-carboxybenzylamino derivative 5g, which
showed relatively potent DPMS inhibition, had comparatively
modest trypanocidal activity. The limited trypanocidal activity
of 5g may be due to its unfavourable physicochemical properties,
which may compromise its penetration through the cell membrane
of the parasite.
Residual DPMS
Compound
control
R
activity (%)a
ED50/mMb
nac
100 3.0
1.5 0.2
na
GDP-mannosed na
nde
GTPd
na
na
na
na
27.3 3.8
8.5 1.8f
78.2 2.3
97.4 4.5
87.1 0.2
81.3 0.3
78.0 2.0
84.1 5.2
69.6 2.1
65.1 3.0
nd
nd
nd
nd
GDPd
GMPd
guanosine
4
ethoxy
1.1 0.2
1.0 0.1
0.725 0.2
>1.5
low solubility
low solubility
0.985 0.2
1.0 0.2
>1.5
5a
5b
5c
5d
5e
5f
5g
6b
6c
n-butylamino
benzylamino
2-nitrobenzylamino
3-nitrobenzylamino
4-nitrobenzylamino
carboxypropylamino 91.9 5.6
2-carboxybenzylamino 64.0 1.0
benzylamino
2-nitrobenzylamino
97.9 0.7
104.4 6.3
>1.5
a In the presence of 1 mM inhibitor, unless stated otherwise; values are
means SD of 3 separate determinations counted twice. b Values are
means SD of 4 separate determinations. c Not applicable. d At 0.1 mM
concentration. e Not determined. f IC50 = 32 4 mM.
The mannosyltransferase dolichol-phosphate mannose synthase
(DPMS) is a Mg2+-dependent enzyme which catalyses the for-
mation of dolichol-phosphate-mannose (Dol-P-Man) from GDP-
mannose and dolichol-phosphate. In all eukaryotes, Dol-P-Man
is the essential mannose donor for all three mannosyltransferases
involved in GPI anchor biosynthesis, and for the last four
mannosyltransferases involved in N-glycosylation. Because of its
central role in GPI biosynthesis and N-glycosylation of VSG,
DPMS represents a promising new target for the development
of novel anti-trypanosomal agents.39
In order to assess the DPMS-inhibitory activity of the novel
squaryldiamides, full-length T. brucei DPMS was recombi-
nantly expressed in E. coli, and DPMS-containing membranes
were used to catalyse the formation of didehydrofarnesol-
phosphate[3H]mannose ([3H]-DFPM) from GDP-[3H]mannose
and didehydrofarnesol-phosphate in the presence or absence of
potential inhibitor. The residual DPMS activity in the presence of
inhibitor was determined from the amount of [3H]-DFPM formed,
relative to the amount of [3H]-DFPM formed in the non-inhibited
reaction (Table 1). For direct comparison, inhibition experiments
were also carried out with non-labelled GDP-mannose, GDP,
which is a by-product of the DPMS reaction and known feedback
inhibitor of DPMS,29 GTP, GMP and guanosine. Under these
conditions, GDP-mannose, GDP and, to a lesser degree, GTP
and GMP, but not guanosine, significantly reduced the DPMS-
catalysed formation of [3H]-DFPM. For GDP, we determined an
Discussion and conclusion
In light of their experimentally demonstrated ability to coordinate
to Mg2, the relatively limited DPMS inhibitory activity of the
novel squarylamides was surprising. It can be speculated that
this may be due to the unfavourable conformational preferences
of our novel squarylamides. 2D NMR spectra from NOESY
and ROESY experiments with 4 show interactions between the
squarylamide proton and protons in position 5¢, 4¢ and 3¢ of
the ribose (Fig. 7). The close proximity of the amide proton
and these protons suggests that both rotamers adopt a “bicyclic”
conformation, possibly stabilised by an internal hydrogen bond
between the amide proton and the 3¢-hydroxyl group (Fig. 4).
An important difference between both rotamers is the orientation
of the guanosine part relative to the cyclobutene-1,2-dione ring.
Sugar-nucleotides frequently adopt an extended conformation,
both in solution41 and bound to protein.42 It appears that in the
case of our novel squarylamides, only the anti rotamer can act as
an effective mimic of this extended sugar-nucleotide conformation.
This journal is
The Royal Society of Chemistry 2010
Org. Biomol. Chem., 2010, 8, 3488–3499 | 3493
©

View Article Online
2-Amino-9-((2R,3R,4S,5S)-3,4-dihydroxy-5-(iodomethyl)-tetra-
hydrofuran-2-yl)-1H-purin-6(9H)-one (1).32 To a suspension of
guanosine (10.02 g, 35.38 mmol), triphenylphosphine (30.62 g,
116.7 mmol) and imidazole (15.9 g, 243.3 mmol) in anhydrous
DMF (135 mL), iodine (28.47 g, 112.2 mmol) was added in
portions, over a period of five minutes, while stirring vigorously.
The orange solution was cooled to room temperature and stirred
for 4 h. The reaction mixture was poured into a stirred mixture
of DCM (1300 mL) and water (400 mL). A white precipitate
formed at the interface of the two layers. The precipitate was
filtered off, washed with DCM (100 mL) and dried at room
temperature overnight to give 1 as a white powder in 65% yield
(9.02 g, 22.9 mmol): Rf 0.37 (DCM–MeOH 7 : 3); mp 178 ◦C
(decomposition); dH (400 MHz, DMSO-d6) 3.41 (dd, 1H, J =
10.3 Hz, 10.6 Hz, H-5¢), 3.55 (dd, 1H, J = 10.3 Hz, 10.6 Hz, H-5¢),
3.93 (m, 1H, J = 5.9 Hz, 6.2 Hz, H-4¢), 4.06 (br, 1H, H-3¢), 4.63
(br, 1H, H-2¢), 5.40 (br, 1H, 3¢-OH), 5.56 (br, 1H, 2¢-OH), 5.70
(d, 1H, J = 6.2 Hz, H-1¢), 6.51 (s, 2H, 2-NH2), 7.92 (s, 1H, H-8),
10.68 (s, 1H, H-1).
Fig. 7 NOE interactions observed for both rotamers of squaryl-
monoamide 4.
At the same time, the conversion of the complete population of
a given squarylamide into the favourable anti rotamer may be
associated with a prohibitively high energy barrier, resulting from
restricted rotation about the squarylamide bond.
In summary, we have developed a series of squaryldiamides
as novel and rationally designed sugar-nucleotide mimics. A key
design criteria was the capacity of the squaric acid fragment to
coordinate to divalent metal ions. Importantly, we could demon-
strate experimentally that the new squaryldiamides coordinate
readily to Mg2+, thus mimicking an essential property of the
diphosphate linkage in sugar-nucleotides. The squaryldiamides
described herein therefore represent one of the first examples for
a sugar-nucleotide mimic whose capacity for metal coordination
has been experimentally demonstrated. Several analogues from
this series possess some biological activity against recombinant
DPMS, and the most potent squaryldiamide 5b was effective
against the bloodstream form of T. brucei with an ED50 of
700 mM. Conformational studies suggest that the generally only
modest bioactivity of these squarylamides may be due to their un-
favourable conformational preferences. Taken together, our results
suggest that squaryldiamides are a promising template for further
development of sugar-nucleotide mimics and novel glycosyltrans-
ferase inhibitors. Our findings illustrate the considerable potential
of the squarylamide group as a fragment for inhibitor design.
Future studies in this area will need to give due consideration to
conformational aspects of this exciting structural class.
2-Amino-9-((2R,3R,4S,5S)-3,4-dihydroxy-5-(azidomethyl)-tetra-
hydrofuran-2-yl)-1H-purin-6(9H)-one (2). 32 A solution of 1
(7.45 g, 18.9 mmol) and sodium azide (1.70 g, 26.1 mmol) was
stirred in anhydrous DMF (50 mL) at 90 ◦C for 22 h. The solvent
was removed in vacuo. The solid residue was triturated with water
(100 mL), filtered off, washed with water (2 ¥ 50 mL), EtOH
(25 mL) and diethylether (25 mL), and dried at room temperature
to give 2 as a dark yellow powder in 76% yield (4.45 g, 14.4 mmol):
Rf 0.35 (DCM–MeOH 7 : 3); mp 241 ◦C (decomposition); dH
(400 MHz, DMSO-d6) 3.51 (dd, 1H, J = 12.8 Hz, 13.2 Hz, H-5¢),
3.67 (dd, 1H, J = 12.8 Hz, 13.2 Hz, H-5¢), 3.98 (q, 1H, J =
3.3 Hz, H-4¢), 4.06 (t, 1H, J = 3.7 Hz, 4.0 Hz, H-3¢), 4.58 (t, 1H,
J = 5.1 Hz, H-2¢), 5.35 (br, 1H, 3¢-OH), 5.57 (br, 1H, 2¢-OH), 5.72
(d, 1H, J = 5.5 Hz, H-1¢), 6.55 (s, 2H, 2-NH2), 7.91 (s, 1H, H-8),
10.72 (br, 1H, H-1).
2-Amino-9-((2R,3R,4S,5S)-3,4-dihydroxy-5-(aminomethyl)-
tetrahydrofuran-2-yl)-1H-purin-6(9H)-one (3). 32 Triphenylphos-
phine (1.71 g, 6.52 mmol) was added to a solution of 2 (1.00 g,
3.24 mmol) in anhydrous pyridine (20 ml). The reaction was
stirred for 5 h at room temperature. Water (20 mL) and aqueous
ammonia (35%, 6.5 mL) were added, and the reaction was stirred
for another 39 h. The solvents were removed in vacuo. The residue
was triturated with EtOAc (85 mL) for 15 min, filtered off, and
washed consecutively with EtOAc, cold EtOAc–MeOH (1 : 1) and
water. The product was dried at 105 ◦C for two days to give 3
Experimental section
General
All reagents were obtained commercially and used without
further purification unless stated otherwise. TLC was carried
out on commercially available TLC aluminium sheets and spots
were visualised under UV at 254 nm. For preparative column
chromatography silica 60 (particle size 0.063–0.2 mm) was used.
Solvents are stated in portions of volume. 1H NMR and 13C NMR
spectra were recorded at, respectively, 399.918 MHz (1H) and
75.456 MHz (13C) on a Varian VXR 400 S spectrometer. Chemical
shifts d are measured from the deuterium lock peak of the solvent
and are recorded in ppm. Coupling constants J are given in Hz
and signal multiplicities are reported as follows: s = singlet, d =
duplet, dd = duplet of duplets, t = triplet, q = quartet, m =
multiplet and br = broad. Melting points were measured on a
capillary melting point apparatus and are not corrected. Mass
spectra were recorded at the EPSRC National Mass Spectrometry
Service Centre, University of Wales, Swansea.
◦
as a white powder in 92% yield (0.84 g, 2.98 mmol): mp 216 C
(decomposition); dH (400 MHz, DMSO-d6) 2.70 (dd, 1H, J =
5.1 Hz, 13.6 Hz, H-5¢), 2.78 (dd, 1H, J = 5.1 Hz, 13.6 Hz, H-5¢),
3.35 (br, 2H, 5¢-NH2), 3.78 (m, 1H, H-4¢), 4.07 (m, 1H, H-3¢), 4.43
(dd, 1H, J = 3.3 Hz, 5.5 Hz, H-2¢), 5.02 (br, 1H, 3¢-OH), 5.38 (br,
1H, 2¢-OH), 5.66 (d, 1H, J = 5.9 Hz, H-1¢), 6.50 (s, 2H, 2-NH2),
7.94 (s, 1H, H-8).
3-(((2R,3S,4R,5R)-5-(2-Amino-6-oxo-1H-purin-9(6H)-yl)-3,4-
dihydroxytetrahydrofuran-2-yl)methylamino)-4-ethoxycyclobut-3-
ene-1,2-dione (4). To a solution of 3 (401 mg, 1.42 mmol)
in anhydrous DMF (35 mL), DIPEA (125 mL, 92.0 mg,
0.71 mmol) and 3,4-diethoxycyclobut-3-ene-1,2-dione (215 mL,
248 mg, 1.46 mmol) were added. The reaction was stirred for
3494 | Org. Biomol. Chem., 2010, 8, 3488–3499
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
92 h at room temperature. The solvent was removed in vacuo.
The residue was sonicated with acetone for five minutes, filtered
off, washed with acetone and dried at 105 ◦C overnight to give
4 as a light yellow powder in 80% yield (464◦mg, 1.14 mmol): Rf
0.62 (DCM–MeOH–water 7 : 7 : 1); mp 196 C (decomposition);
dH (400 MHz, DMSO-d6) 1.29 (t, 1.5H, J = 7.0 Hz, CH3), 1.36
(t, 1.5H, J = 7.0 Hz, Et: CH3), 3.46–3.65 (m, 1H, H-5¢), 3.65–3.89
(m, 1H, H-5¢), 3.93 (br, 1H, H-4¢), 4.11 (br, 1H, H-3¢), 4.45 (br, 1H,
H-2¢), 4.61 (m, 2H, Et: CH2), 5.28 (br, 1H, 3¢-OH), 5.53 (br, 1H,
2¢-OH), 5.68 (d, 1H, J = 4.8 Hz, H-1¢), 6.47 (d, 2H, J = 9.9 Hz,
2-NH2), 7.83 (s, 1H, H-8), 8.74 (t, 0.5H, J = 5.1 Hz, 5¢-NH), 8.89
(t, 0.5H, J = 4.4 Hz, 5¢-NH), 10.67 (s, 1H, NH-1); dC (75 MHz,
DMSO-d6) 15.6 (Et: CH3), 46.2 (C-5¢), 69.3 (Et: CH2), 72.7 (C-3¢),
73.0 (C-2¢), 83.2 (C-4¢), 87.2 (C-1¢), 117.3 (C-5), 136.5 (C-8), 151.5
(C-4), 153.8 (C-2), 157.1 (C-6), 174.6 (Sq: C-3), 177.4 (Sq: C-4),
182.7 (Sq: C-2), 190.1 (Sq: C-1). m/z (ESI) 406.1237 (C16H18N6O7
requires 406.1237).
7.0 Hz, Ph: H-2, 6), 7.43 (t, 1H, J = 6.2 Hz, Sq-NH), 7.80 (br, 1H,
Sq-NH), 7.88 (s, 1H, H-8), 10.64 (br, 1H, NH-1); dC (75 MHz,
DMSO-d6) 45.5 (C-5¢), 46.7 (Bn: CH2), 70.8 (C-3¢), 72.6 (C-2¢),
83.4 (C-4¢), 86.6 (C-1¢), 117.0 (C-5), 127.4 (Ph: C-4), 127.5 (Ph:
C-2, 6), 128.7 (Ph: C-3, 5), 136.0 (Ph: C-1), 138.9 (C-8), 151.4
(C-4), 153.8 (C-2), 156.8 (C-6), 167.7 (Sq: C-4), 167.8 (Sq: C-3),
182.8 (Sq: C-1), 189.1 (Sq: C-2). m/z (ESI) 467.1547 (C21H21N7O6
requires 467.1546).
3-(((2R,3S,4R,5R)-5-(2-Amino-6-oxo-1H-purin-9(6H)-yl)-3,
4-dihydroxytetrahydrofuran-2-yl)methylamino)-4-(2-nitrobenzyl-
amino)cyclobut-3-ene-1,2-dione (5c). To a solution of 4 (15.0 mg,
36.9 mmol) in anhydrous DMF (2 mL) were added first DIPEA
(10.9 mL, 8.1 mg, 62.6 mmol), and then 2-nitrobenzylamine
hydrochloride (8.3 mg, 44.0 mmol). The reaction mixture was
stirred for seven days at room temperature, and then for four days
at 60 ◦C. The solvent was removed in vacuo, and the residue was
sonicated with acetone for five◦minutes. The precipitate was filtered
off, washed and dried at 105 C overnight to give 5c as a brown
powder in 13% yield (2.5 mg, 4.9 mmol): Rf 0.58 (DCM–MeOH–
water 7 : 7 : 1); mp 240 ◦C (decomposition); dH (400 MHz, DMSO-
d6) 3.71 (m, 1H, H-5¢), 3.94 (m, 1H, H-4¢), 4.05 (br, 1H, H-3¢), 4.51
(br, 1H, H-2¢), 4.97 (d, 2H, J = 4.9 Hz, Bn: CH2), 5.32 (br, 1H,
3¢-OH), 5.54 (br, 1H, 2¢-OH), 5.71 (d, 1H, J = 5.9 Hz, H-1¢), 6.49
(br, 2H, 2-NH2), 7.62 (m, 2H, Ph: H-4, 5), 7.71 (br, 1H, Sq-NH),
7.81 (d, 1H, J = 7.5 Hz, Ph: H-3), 7.89 (s, 1H, H-8), 7.91 (br, 1H,
Sq-NH), 8.10 (d, 1H, J = 7.7 Hz, Ph: H-6), 10.66 (br, 1H, NH-1);
dC (75 MHz, DMSO-d6) 44.3 (C-5¢), 45.6 (Bn: CH2), 70.8 (C-3¢),
72.7 (C-2¢), 83.5 (C-4¢), 86.8 (C-1¢), 118.3 (C-5), 123.0 (Ph: C-2),
125.1 (Ph: C-6), 131.0 (Ph: C-5), 133.8 (Ph: C-3), 134.7 (Ph: C-4),
136.1 (C-8), 147.7 (Ph: C-1), 151.4 (C-4), 153.6 (C-2), 156.8 (C-6),
162.6 (Sq: C-4), 167.4 (Sq: C-3), 176.7 (Sq: C-1), 182.9 (Sq: C-2).
m/z (ESI) 512.1403 (C21H20N8O8 requires 512.1404).
Preparation of squaryldiamides 5a–c from squarylmonoamide 4
3-(((2R,3S,4R,5R)-5-(2-Amino-6-oxo-1H-purin-9(6H)-yl)-3,4-
dihydroxytetrahydrofuran-2-yl)methylamino)-4-(butylamino)cy-
clobut-3-ene-1,2-dione (5a). To a solution of 4 (50.0 mg,
123 mmol) in anhydrous DMF (1 mL) were added first DIPEA
(11.0 mL, 8.2 mg, 63.1 mmol), and then n-butylamine (13.5 mL,
10.0 mg, 137 mmol). The reaction mixture was stirred at room
temperature for two days. The solvent was removed in vacuo,
and the residue was sonicated with acetone for five minutes.
The precipitate was filtered off, washed and dried at 105 ◦C
overnight to give 5a as a pale yellow powder in 79% yield (42.1 mg,
97.1 mmol): Rf 0.62 (DCM–MeOH–water 7 : 7 : 1); mp 238 ◦C
(decomposition); dH (400 MHz, DMSO-d6) 0.87 (t, 3H, J = 5.2 Hz,
Bu: CH3), 1.29 (m, 2H, Bu: C3H2), 1.47 (m, 2H, Bu: C2H2), 3.44
(m, 1H, H-5¢), 3.49 (br, 2H, Bu: C1H2), 3.70 (m, 1H, H-5¢), 3.95
(br, 1H, H-4¢), 4.05 (br, 1H, H-3¢), 4.52 (br, 1H, H-2¢), 5.32 (br, 1H,
3¢-OH), 5.54 (br, 1H, 2¢-OH), 5.72 (d, 1H, J = 4.6 Hz, H-1¢), 6.54
(br, 2H, 2-NH2), 7.43 (br, 2H, Sq: NH), 7.89 (s, 1H, H-8), 10.74
(br, 1H, NH-1); dC (75 MHz, DMSO-d6) 13.5 (Bu: CH3), 19.1 (Bu:
C3H2), 32.8 (Bu: C2H2), 42.9 (Bu: C1H2), 45.5 (C-5¢), 70.8 (C-3¢),
72.7 (C-2¢), 83.6 (C-4¢), 86.7 (C-1¢), 116.9 (C-5), 136.1 (C-8), 151.4
(C-4), 153.6 (C-2), 156.9 (C-6), 167.9 (Sq: C-3), 168.2 (Sq: C-4),
175.9 (Sq: C-1), 182.5 (Sq: C-2). m/z (ESI) 433.1709 (C18H23N7O6
requires 433.1710).
Preparation of squaryldiamides 5c–g from squarylmonoamides
6c–g
3-(((2R,3S,4R,5R)-5-(2-Amino-6-oxo-1H-purin-9(6H)-yl)-3,
4-dihydroxytetrahydrofuran-2-yl)methylamino)-4-(2-nitrobenzyl-
amino)cyclobut-3-ene-1,2-dione (5c). To a solution of 3 (10.1 mg,
35.8 mmol) in anhydrous DMF (2 mL) were added first DIPEA
(6.3 mL, 4.7 mg, 36.2 mmol), and then 6c (9.9 mg, 35.8 mmol).
The reaction was stirred for five days at room temperature. The
solvent was removed in vacuo, and the residue was purified by
column chromatography. First, the column was eluted with DCM–
MeOH (9 : 1) to remove unreacted 6c, followed by elution with
DCM–MeOH–water (7 : 7 : 1). Product-containing fractions were
combined and the solvents were evaporated to give 5c as a yellow
powder in 92% yield (16.8 mg, 32.8 mmol).
3-(((2R,3S,4R,5R)-5-(2-Amino-6-oxo-1H-purin-9(6H)-yl)-3,4-
dihydroxytetrahydrofuran-2-yl)methylamino)-4-(benzylamino)cy-
clobut-3-ene-1,2-dione (5b). To a solution of 4 (47.2 mg,
116 mmol) in anhydrous DMF (1.25 mL) were added first DIPEA
(10.4 mL, 7.7 mg, 60 mmol), and then benzylamine (12.7 mL,
12.5 mg; 120 mmol). The reaction mixture was stirred overnight
at room temperature. The solvent was removed in vacuo and the
residue was sonicated with dichloromethane–hexane (2 : 1, 10 mL)
for five minutes. The residue was filtered off, washed and dried at
105 ◦C overnight to give 5b as a light brown powder in 90% yield
(49.0 mg, 105 mmol): Rf 0.57 (DCM–MeOH–water 7 : 7 : 1); mp
3-(((2R,3S,4R,5R)-5-(2-Amino-6-oxo-1H-purin-9(6H)-yl)-3,
4-dihydroxytetrahydrofuran-2-yl)methylamino)-4-(3-nitrobenzyl-
amino)cyclobut-3-ene-1,2-dione (5d). To a solution of 3 (50.0 mg,
177 mmol) in anhydrous DMF (5 mL) were added first DIPEA
(15.4 mL, 11.4 mg, 88.4 mmol), and then 6d (49.0 mg, 177 mmol).
The reaction was stirred for five days at room temperature. The
solvent was removed in vacuo, and the residue was sonicated with
acetone for five minutes. The precipitate was filtered off, washed
with acetone and dried at 105 ◦C to give 5d as a light yellow powder
in 89% yield (80.9 mg, 158 mmol): Rf 0.49 (DCM–MeOH–water
◦
211 C (decomposition); dH (400 MHz, DMSO-d6) 3.71 (m, 1H,
H-5¢), 3.94 (m, 1H, H-4¢), 4.05 (br, 1H, H-3¢), 4.51 (m, 1H, H-2¢),
4.70 (br, 2H, Bn: CH2), 5.31 (br, 1H, 3¢-OH), 5.53 (d, 1H, J =
5.5 Hz, 2¢-OH), 5.71 (d, 1H, J = 6.2 Hz, H-1¢), 6.50 (br, 2H, 2-
NH2), 7.30 (t, 3H, J = 6.2 Hz, Ph: H-3, 4, 5), 7.37 (d, 2H, J =
This journal is
The Royal Society of Chemistry 2010
Org. Biomol. Chem., 2010, 8, 3488–3499 | 3495
©

View Article Online
7 : 7 : 1); mp 214 ◦C; dH (400 MHz, DMSO-d6) 3.73 (m, 1H, H-5¢),
3.96 (m, 1H, H-4¢), 4.06 (d, 1H, J = 4.3 Hz, H-3¢), 4.52 (m, 1H,
H-2¢), 4.84 (br, 2H, Bn: CH2), 5.31 (br, 1H, 3¢-OH), 5.54 (br, 1H,
2¢-OH), 5.71 (d, 1H, J = 5.5 Hz, H-1¢), 6.51 (br, 2H, 2-NH2), 7.57
(m, 1H, Sq: NH), 7.68 (br, 1H, Ph: H-4), 7.89 (s, 1H, H-8), 8.16 (s,
1H, Ph: H-2), 8.19 (br, 1H, Ph: H-6), 7.40-8.40 (1H, Sq: NH), 9.80
(br, 1H, NH-1); dC (75 MHz, DMSO-d6) 45.7 (C-5¢), 46.0 (Bn:
CH2), 71.0 (C-3¢), 72.9 (C-2¢), 83.7 (C-4¢), 87.0 (C-1¢), 117.1 (C-5),
122.3 (Ph: C-2), 122.7 (Ph: C-6), 130.6 (Ph: C-5), 134.5 (Ph: C-3),
136.4 (C-8), 141.5 (Ph: C-4), 148.1 (Ph: C-1), 151.6 (C-4), 153.8
(C-2), 157.1 (C-6), 167.6 (Sq: C-4), 182.6 (Sq: C-3), 182.9 (Sq:
C-1), 190.2 (Sq: C-2). m/z (ESI) 512.1398 (C21H20N8O8 requires
512.1404).
27.5 mg, 212 mmol) and 6g (39.0 mg, 142 mmol) as described for
the preparation of 5f. Thus, 5g was obtained as a white powder
in 99% yield (72.1 mg, 141 mmol): Rf 0.58 (DCM–MeOH–water
7 : 7 : 1); mp 267 ◦C (decomposition); dH (400 MHz, DMSO-d6)
3.67 (m, 1H, H-5¢), 3.93 (m, 1H, H-4¢), 3.98 (m, 1H, H-5¢), 4.06
(br, 1H, H-3¢), 4.50 (br, 1H, H-2¢), 5.01 (br, 2H, Bn: CH2), 5.40
(br, 1H, 3¢-OH), 5.56 (br, 1H, 2¢-OH), 5.69 (d, 1H, J = 6.0 Hz,
H-1¢), 6.75 (br, 2H, 2-NH2), 7.28 (br, 0.5H, Sq-NH), 7.35 (m, 3H,
Ph: H-3, 4, 5), 7.83 (d, 1H, J = 5.8 Hz, Ph: H-6), 7.86 (s, 1H,
H-8), 8.40 (br, 1.5H, Sq-NH), 9.85 (br, 1H, CO2H), 11.00 (br, 1H,
NH-1); dC (75 MHz, DMSO-d6) 45.6 (C-5¢), 46.5 (Bn: CH2), 71.0
(C-3¢), 72.9 (C-2¢), 83.8 (C-4¢), 86.9 (C-1¢), 117.1 (C-5), 123.0 (Ph:
C-1), 127.6 (Ph: C-4), 129.8 (Ph: C-5), 130.0 (Ph: C-3), 130.6 (Ph:
C-6), 136.5 (C-8), 138.1 (Ph: C-2), 151.9 (C-4), 154.1 (C-2), 157.5
(C-6), 168.2 (Sq: C-1), 182.6 (Sq: C-2), 182.8 (CO2H), 184.3 (Sq:
C-4), 189.7 (Sq: C-3). m/z (ESI) 511.1445 (C22H21N7O8 requires
511.1452).
3-(((2R,3S,4R,5R)-5-(2-Amino-6-oxo-1H-purin-9(6H)-yl)-3,
4-dihydroxytetrahydrofuran-2-yl)methylamino)-4-(4-nitrobenzyl-
amino)cyclobut-3-ene-1,2-dione (5e). The title compound was
prepared from 3 (50.0 mg, 177 mmol), DIPEA (15.4 mL, 11.4 mg,
88.4 mmol) and 6e (49.0 mg, 177 mmol) as described for the
preparation of 5d. Thus, 5e was obtained as a light yellow powder
in 74% yield (66.9 mg, 131 mmol): Rf 0.50 (DCM–MeOH–water
7 : 7 : 1); mp 210 ◦C (decomposition); dH (400 MHz, DMSO-d6)
3.73 (m, 1H, H-5¢), 3.96 (m, 1H, H-4¢), 4.07 (d, 1H, J = 4.3 Hz,
H-3¢), 4.52 (m, 1H, H-2¢), 4.84 (br, 2H, Bn: CH2), 5.32 (br, 1H,
3¢-OH), 5.55 (d, 1H, J = 4.0 Hz, 2¢-OH), 5.72 (d, 1H, J = 4.8 Hz,
H-1¢), 6.51 (br, 2H, 2-NH2), 7.58 (br, 2H, Ph: H-3, 5), 7.89 (s, 1H,
H-8), 8.24 (br, 2H, Ph: H-2, 6), 7.40-8.40 (2H, 2 ¥ Sq: NH), 9.94
(br, 1H, NH-1); dC (75 MHz, DMSO-d6) 45.8 (C-5¢), 46.3 (Bn:
CH2), 71.0 (C-3¢), 73.0 (C-2¢), 83.7 (C-4¢), 87.0 (C-1¢), 117.1 (C-5),
124.2 (Ph: C-2, 6), 128.8 (Ph: C-3, 5), 136.4 (C-8), 138.2 (Ph: C-4),
147.1 (Ph: C-1), 151.7 (C-4), 153.9 (C-2), 157.1 (C-6), 162.4 (Sq:
C-4), 167.7 (Sq: C-3), 182.6 (Sq: C-1), 184.1 (Sq: C-2). m/z (ESI)
512.1397 (C21H20N8O8 requires 512.1404).
3-(Butylamino)-4-ethoxycyclobut-3-ene-1,2-dione
(6a). n-
Butylamine (31 mL, 22.9 mg, 0.31 mmol) and DIPEA (30 ml,
22,3 mg, 0.17 mmol) were dissolved in EtOH (1 mL). To the
solution, 3,4-diethoxycyclobut-3-ene-1,2-dione (50 mL, 57.5 mg,
0.34 mmol) was added, and the mixture was stirred at room
temperature until TLC indicated completion of the reaction (ca.
2h). The solvents were removed in vacuo and the residue was
purified by column chromatography (DCM–MeOH 97.5 : 2.5)
to give 6a as a yellow solid in quantitative yield (61.9 mg,
0.31 mmol): Rf 0.17 (DCM, MeOH 975 : 25); mp 42 ◦C; dH
(400 MHz, DMSO-d6) 0.87 (t, J = 7.2 Hz, 3H, Bu: CH3), 1.30
(m, 2H, Bu: C3H2), 1.36 (t, J = 6.9 Hz, 3H, Et: CH3), 1.49 (m,
2H, Bu:C2H2), 3.29 (t, J = 6.5 Hz, 1H, Bu: C1H2), 3.48 (t, J =
7.0 Hz, 1H, Bu: C1H2), 4.65 (q, J = 6.9 Hz, 2H, Et: CH2), 8.59
(br, 0.5H, NH), 8.80 (br, 0.5H, NH); dC (75 MHz, DMSO-d6)
13.6 (Bu: CH3), 15.8 (Et: CH3), 19.1 (Bu: C3H2), 32.6 (Bu: C2H2),
43.5 (Bu: C1H2), 69.0 (Et: CH2), 171.5 (Sq: C-4), 175.6 (Sq: C-3),
182.2 (Sq: C-1), 189.8 (Sq: C-2). m/z (ESI) 197.1052 (C10H15NO3
requires 197.1052).
4-(2-(((2R,3S,4R,5R)-5-(2-Amino-6-oxo-1H-purin-9(6H)-yl)-3,
4-dihydroxytetrahydrofuran-2-yl)methylamino)-3,4-dioxocyclobut-
1-enylamino)butanoic acid (5f). To a solution of 3 (40.0 mg,
142 mmol) in anhydrous DMF (4 mL) were added first DIPEA
(12.4 mL, 9.2 mg, 71.2 mmol), and then 6f (43.2 mg, 190 mmol).
The reaction was stirred for seven days at room temperature. The
solvent was removed in vacuo, and the residue was sonicated with
acetone for five minutes. The precipitate was filtered off, washed
with acetone and dried at 105 ◦C to give 5f as a white powder
in 97% yield (63.8 mg, 138 mmol): Rf 0.33 (DCM–MeOH,/water
7 : 7 : 1); mp 242 ◦C (decomposition); dH (400 MHz, DMSO-d6)
1.73 (br, 2H, Bu: C3H2), 2.21 (br, 2H, Bu: C2H2), 3.49 (m, 2H, Bu:
C4H2), 3.72 (m, 1H, H-5¢), 3.97 (br, 1H, H-4¢), 4.07 (br, 1H, H-3¢),
4.52 (m, 1H, H-2¢), 5.30-5.60 (br, 2H, 2¢-OH, 3¢-OH), 5.71 (d, 1H,
J = 5.1 Hz, H-1¢), 6.57 (br, 2H, 2-NH2), 7.62 (br, 1H, Sq: NH),
7.89 (s, 1H, H-8), 10.45 (br, 1H, NH-1); dC (75 MHz, DMSO-d6)
26.6 (Bu: C-3), 35.0 (Bu: C-2), 43.1 (Bu: C-4), 45.7 (C-5¢), 71.0
(C-3¢), 72.9 (C-2¢), 83.8 (C-4¢), 87.1 (C-1¢), 117.2 (C-5), 136.7 (C-
8), 149.4 (C-4), 154.0 (C-2), 157.3 (C-6), 168.0 (Sq: C-1), 178.3
(CO2H), 182.7 (Sq: C-2), 183.6 (Sq: C-4), 188.0 (Sq: C-3). m/z
(ESI) 463.1450 (C18H21N7O8 requires 463.1452).
3-(Benzylamino)-4-ethoxycyclobut-3-ene-1,2-dione
(6b). n-
Benzylamine (37 mL, 36.3 mg, 0.34 mmol) was added at 0 ◦C
to a solution of 3,4-diethoxycyclobut-3-ene-1,2-dione (50 mL,
57.5 mg, 0.34 mmol) in EtOH (2 mL). The reaction was stirred
at room temperature overnight. The solvents were removed in
vacuo and the residue was purified by column chromatography
(DCM–MeOH 95 : 5) to give 6b as a colourless, viscous liquid in
83% yield (64.9 mg, 0.28 mmol): Rf 0.85 (DCM–MeOH 9 : 1); dH
(400 MHz, CDCl3) 1.42 (t, 3H, J = 7.0 Hz, 6.6 Hz, Et: CH3), 4.58
(d, 1.5H, J = 5.9 Hz, Bn: CH2), 4.66 (br, 0.5H, Bn: CH2), 4.75
(q, 2H, J = 7.0 Hz, 6.6 Hz, Et: CH2), 7.30 (m, 5H, Ph), 9.08 (br,
0.5H, NH), 9.29 (br, 0.5H, NH); dC (75 MHz, CDCl3) 15.6 (Et:
CH3), 48.4 (Bn: CH2), 69.7 (Et: CH2), 127.7 (Ph: C-4), 128.1 (Ph:
C-2, 6), 128.9 (Ph: C-3, 5), 137.1 (Ph: C-1), 172.5 (Sq: C-3), 177.7
(Sq: C-4), 183.0 (Sq: C-1), 189.6 (Sq: C-2). m/z (ESI) 231.0895
(C13H13NO3 requires 231.0895).
3-(2-Nitrobenzylamino)-4-ethoxycyclobut-3-ene-1,2-dione (6c).
2-Nitrobenzylamine hydrochloride (102.2 mg, 0.54 mmol) and
DIPEA (140 mL, 103.9 mg, 0.80 mmol) were dissolved in EtOH
(5 mL). To the solution, 3,4-diethoxycyclobut-3-ene-1,2-dione
(80 mL, 92.0 mg, 0.54 mmol) was added, and the reaction was
2-(4-(2-(((2R,3S,4R,5R)-5-(2-Amino-6-oxo-1H-purin-9(6H)-
yl)-3,4-dihydroxytetrahydrofuran-2-yl)methylamino)-3,4-dioxocy-
clobut-1-enylamino)methyl)benzoic acid (5g). The title com-
pound was prepared from 3 (40.0 mg, 142 mmol), DIPEA (37 mL,
3496 | Org. Biomol. Chem., 2010, 8, 3488–3499
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
stirred at room temperature until TLC indicated completion (ca. 2
h). The solvents were removed in vacuo and the residue was purified
by column chromatography (DCM–MeOH 97.5 : 2.5) to give 6c as
a pale yellow solid in 97% yield (145.4 mg, 0.53 mmol): Rf 0.28
(DCM–MeOH 975 : 25); mp 164 ◦C; dH (400 MHz, DMSO-d6)
1.23 (t, 1.5H, J = 6.6 Hz, Et: CH3), 1.38 (t, 1.5H, J = 7.0 Hz,
Et: CH3), 4.55 (q, 1H, J = 7.0 Hz, Et: CH2), 4.67 (q, 1H, J =
6.6 Hz, Et: CH2), 4.82 (d, 1H, J = 5.1 Hz, Bn: CH2), 5.03 (d, 1H,
J = 5.1 Hz, Bn: CH2), 7.60 (m, 2H, J = 7.7 Hz, Ph: H-4, 5), 7.79
(t, 1H, J = 7.7 Hz, 1.1 Hz, Ph: H-3), 8.09 (t, 1H, J = 5.9 Hz,
1.8 Hz, Ph: H-6), 9.07 (t, 0.5H, J = 5.1 Hz, NH), 9.32 (t, 0.5H,
J = 5.1 Hz, NH); dC (75 MHz, DMSO-d6) 15.7 (Et: CH3), 44.4
(Bn: CH2), 69.4 (Et: CH2), 125.2 (Ph: C-6), 129.4 (Ph: C-5), 130.1
(Ph: C-3), 133.2 (Ph: C-2), 134.2 (Ph: C-4), 147.9 (Ph: C-1), 173.1
(Sq: C-3), 177.6 (Sq: C-4), 183.0 (Sq: C-2), 189.1 (Sq: C-1). m/z
(ESI) 276.0748 (C13H12N2O5 requires 276.0746).
temperature overnight. The solvents were removed in vacuo and
the residue was purified by column chromatography (DCM–
MeOH 7 : 3) to give 6f as a colourless, viscous liquid in 97% yield
(106.9 mg, 0.47 mmol): Rf 0.43 (DCM–MeOH 7 : 3); dH (400 MHz,
DMSO-d6) 1.38 (m, 3H, Et: CH3), 1.75 (m, 2H, Bu: C3H2), 2.33
(m, 2H, Bu: C2H2), 3.49 (br, 2H, Bu: C4H2), 4.66 (m, 2H, Et: CH2),
8.67 (br, 0.4H, NH), 8.84 (br, 0.6H, NH); dC (75 MHz, DMSO-d6)
15.6 (Et: CH3), 25.4 (Bu: C3H2), 30.9 (Bu: C2H2), 43.1 (Bu: C4H2),
68.7 (Et: CH2), 172.4 (Sq: C-2), 174.5 (CO2H), 176.6 (Sq: C-1),
182.1 (Sq: C-3), 189.3 (Sq: C-4). m/z (ESI) 227.0793 (C10H13NO5
requires 227.0794).
2-((2-Ethoxy-3,4-dioxocyclobut-1-enylamino)methyl)benzoic
acid (6g). 2-Aminomethylbenzoic acid (100.0 mg, 0.66 mmol)
and DIPEA (170 mL, 126.1 mg, 0.98 mmol) were dissolved in
MeOH (20 mL). To the solution, 3,4-diethoxycyclobut-3-ene-
1,2-dione (98 mL, 112.7 mg, 0.66 mmol) was added, and the
reaction was stirred at room temperature overnight. The solvents
were removed in vacuo and the residue was purified by column
chromatography (DCM–MeOH 7 : 3) to give 6g as a pale orange
solid in quantitative yield (181.9 mg, 0.66 mmol): Rf 0.57 (DCM–
MeOH 7 : 3); mp 233 ◦C (decomposition); dH (400 MHz, DMSO-
d6) 1.36 (m, 3H, Et: CH3), 4.63 (m, 2H, Et: CH2), 4.70 (br, 1H,
Bn: CH2), 4.89 (br, 1H, Bn: CH2), 7.15-7.45 (m, 4H, Ph: H-3, 4,
5, NH), 7.82 (d, J = 7.3 Hz, 1H, Ph: H-6), 9.64 (br, 1H, CO2H);
dC (75 MHz, DMSO-d6) 15.9 (Et: CH3), 47.2 (Bn: CH2), 69.4 (Et:
CH2), 127.7 (Ph: C-5), 128.4 (Ph: C-1), 128.6 (Ph: C-3), 130.0
(Ph: C-4), 130.6 (Ph: C-6), 137.2 (Ph: C-2), 172.4 (Sq: C-2), 176.7
(CO2H), 177.4 (Sq: C-1), 183.6 (Sq: C-3), 189.3 (Sq: C-4).
3-(3-Nitrobenzylamino)-4-ethoxycyclobut-3-ene-1,2-dione (6d).
3-Nitrobenzylamine hydrochloride (100.0 mg, 0.53 mmol) and
DIPEA (140 mL, 103.9 mg, 0.80 mmol) were dissolved in EtOH
(5 mL). To the solution, 3,4-diethoxycyclobut-3-ene-1,2-dione
(78.5 mL, 90.3 mg, 0.53 mmol) was added, and the reaction was
stirred at room temperature overnight. The solvents were removed
in vacuo and the residue was purified by column chromatography
(DCM–MeOH 95 : 5) to give 6d as a pale yellow solid in 65% yield
(95.5 mg, 0.35 mmol): Rf 0.39 (DCM–MeOH 95 : 5); mp 113 ◦C;
d
H (400 MHz, DMSO-d6) 1.35 (m, 3H, Et: CH3), 4.64 (m, 2H, Et:
CH2), 4.81 (d, 2H, J = 4.8 Hz, Bn: CH2), 7.68 (q, 1H, J = 7.7 Hz,
Ph: H-5), 7.77 (d, 1H, J = 7.0 Hz, Ph: H-4), 8.17 (d, 1H, J =
7.7 Hz, Ph: H-6), 8.19 (s, 1H, Ph: H-2), 9.15 (br, 0.5H, NH), 9.36
(br, 0.5H, NH); dC (75 MHz, DMSO-d6) 15.5 (Et: CH3), 46.4 (Bn:
CH2), 69.0 (Et: CH2), 122.4 (Ph: C-2), 122.6 (Ph: C-6), 130.4 (Ph:
C-5), 134.4 (Ph: C-4), 140.5 (Ph: C-3), 148.1 (Ph: C-1), 173.2 (Sq:
C-3), 177.5 (Sq: C-4), 183.1 (Sq: C-2), 189.5 (Sq: C-1). m/z (ESI)
276.0748 (C13H12N2O5 requires 276.0746).
3-(2-Nitrobenzylamino)-4-aminocyclobut-3-ene-1,2-dione (7).
To a solution of 6c (30.0 mg, 109 mmol) in MeOH (5 ml) was added
aqueous NH3 (35%, 1mL). The reaction was stirred for two days
at 50 ◦C, by which time TLC indicated complete conversion.
The solvent was removed and the residue was recrystallised from
methanol to give 7 as a yellow powder in 34% yield (9.0 mg,
36.4 mmol): Rf value: 0.25 (DCM–MeOH 9 : 1); mp 284 ◦C
(decomposition); dH (400 MHz, DMSO-d6) 4.98 (d, 2H, J =
6.2 Hz, CH2), 7.20-8.05 (br, 3H, NH)7.62 (q, 2H, J = 7.3 Hz,
8.1 Hz, H-3, 5), 7.80 (t, 1H, J = 1.1 Hz, 7.3 Hz, 7.7 Hz, H-4), 8.12
(d, 1H, J = 1.1 Hz, 8.1 Hz, H-6); dC (75 MHz, DMSO-d6) 44.5
(CH2), 125.4 (C-3), 129.5 (C-4), 131.1 (C-1), 134.1 (C-6), 134.9
(C-5), 147.9 (C-2), 163.8 (Sq: C-3), 172.2 (Sq: C-4), 182.9 (Sq:
C-1), 183.3 (Sq: C-2). m/z (ESI) 247.0592 (C21H20N8O8 requires
247.0593).
3-(4-Nitrobenzylamino)-4-ethoxycyclobut-3-ene-1,2-dione (6e).
4-Nitrobenzylamine hydrochloride (100.0 mg, 0.53 mmol) and
DIPEA (140 mL, 103.9 mg, 0.80 mmol) were dissolved in EtOH
(5 mL). To the solution, 3,4-diethoxycyclobut-3-ene-1,2-dione
(78.5 mL, 90.3 mg, 0.53 mmol) was added, and the reaction was
stirred at room temperature overnight. The solvents were removed
in vacuo and the residue was purified by column chromatography
(DCM–MeOH 95 : 5) to give 6e as a pale yellow solid in 94% yield
(137.5 mg, 0.50 mmol): Rf 0.36 (DCM–MeOH 95 : 5); mp 168 ◦C;
dH (400 MHz, DMSO-d6) 1.29 (t, 1.5H, J = 6.2 Hz, Et: CH3),
1.37 (t, 1.5H, J = 6.2 Hz, Et: CH3), 4.65 (m, 2H, Et: CH2), 4.81
(d, 2H, J = 5.1 Hz, Bn: CH2), 7.58 (d, 2H, J = 7.7 Hz, Ph: H-3,
5), 8.23 (d, 2H, J = 8.4 Hz, Ph: H-2, 6), 9.16 (t, 0.5H, J = 5.1 Hz,
NH), 9.37 (t, 0.5H, J = 5.1 Hz, NH); dC (75 MHz, DMSO-d6)
15.5 (Et: CH3), 44.6 (Bn: CH2), 69.0 (Et: CH2), 123.9 (Ph: C-2,
6), 128.7 (Ph: C-3, 5), 146.3 (Ph: C-4), 147.0 (Ph: C-1), 172.4 (Sq:
C-3), 177.8 (Sq: C-4), 181.6 (Sq: C-2), 189.7 (Sq: C-1). m/z (ESI)
276.0748 (C13H12N2O5 requires 276.0746).
2-((1,3-Dioxoisoindolin-2-yl)methyl)benzoic acid. 43 Phthalide
(10.00 g, 74.6 mmol) and potassium phthalimide (15.00 g,
81.0 mmol) were suspended in anhydrous DMF (50 mL). The
reaction was heated to reflux overnight. After cooling to room
temperature, aqueous acetic acid (35%, 80 mL) was added and
the reaction was stirred for 30 min. A precipitate formed which
was filtered off, washed with water (3 ¥ 20 mL) and EtOH (2 ¥
20 mL). The precipitate was triturated with boiling aqueous EtOH
(60%, 60mL) for 30 min. Hot filtration and washing with EtOH
(2 ¥ 10 mL) gave ca. 15 g of crude product. This material was
recrystallised, in two portions, from glacial acetic acid (300 mL)
in the presence of Norit charcoal to give the title compound as a
white, crystalline powder in 58% yield (12.15 g, 43.2 mmol): Rf 0.59
4-(2-Ethoxy-3,4-dioxocyclobut-1-enylamino)butanoic acid (6f).
4-Aminobutanoic acid (50.0 mg, 0.49 mmol) and DIPEA (42 mL,
31.2 mg, 0.24 mmol) were dissolved in MeOH (7 mL). To the
solution, 3,4-diethoxycyclobut-3-ene-1,2-dione (72 mL, 82.8 mg,
0.49 mmol) was added, and the reaction was stirred at room
◦
(DCM–MeOH 7 : 3); mp 267 C; dH (400 MHz, DMSO-d6) 5.36
This journal is
The Royal Society of Chemistry 2010
Org. Biomol. Chem., 2010, 8, 3488–3499 | 3497
©

View Article Online
(s, 2H, CH2), 7.19 (d, 1H, J = 8.1 Hz, Bn: H-3), 7.36 (m, 1H, J =
7.3 Hz, 1.1 Hz, Bn: H-4), 7.47 (m, 1H, J = 7.7 Hz, 7.3 Hz, 1.1 Hz,
Bn: H-5), 7.77 (m, 2H, H-3, H-6), 7.89 (m, 2H, H-4, H-5), 8.08
(dd, 1H, J = 7.7 Hz, 1.1 Hz, Bn: H-6); dC (75 MHz, DMSO-d6)
123.4 (Phth: C-2, 5), 126.3 (Ph: C-3), 127.3 (Ph: C-5), 129.2 (Ph:
C-1), 130.9 (Ph: C-6), 131.9 (Ph: C-4), 132.6 (Phth: C-1, 6), 134.7
dockings that included Mn2+. This choice was validated by docking
GDP back into the crystal structure. For the calcium parameters
this redocking resulted in large cluster (93 out of 100) of states that
˚
fell within 0.1 A rmsd of the crystal structure coordinates. Ligand
geometries were optimised semi-empirically with GAMESS using
the AUSTIN MODEL 1 method.45,46 All dockings were performed
˚
=
(Phth: C-3, 4), 137.7 (Ph: C-2), 168.2 (2 ¥ C O), 168.5 (CO2H).
on a 60 ¥ 50 ¥ 46 grid of 0.375 A spacing centred on the known
binding site. Two rounds of docking were carried out for each
compound using AutoDock’s Lamarkian genetic algorithm. A
population size of 250 was chosen for the first round and a member
from the lowest energy cluster was then fed into the next round as
the input structure and optimised within a population of 200.
35
2-Aminomethylbenzoic acid.
2-((1,3-Dioxoisoindolin-2-
yl)methyl)benzoic acid (1.00 g, 3.56 mmol) was suspended in a
mixture of anhydrous DMF (7.5 mL) and EtOH (15 mL). The
mixture was heated to 75 ◦C until a clear solution was obtained.
To the solution, hydrazine hydrate (64%, 0.5 mL, 7.4 mmol) was
added, and the reaction was stirred overnight, at which stage
a precipitate had formed. The reaction mixture was cooled for
30 min in an ice bath to complete precipitation. TLC indicated
that different from the description in the original protocol35 the
precipitate, and not the filtrate, contained the desired product. The
precipitate was filtered off and washed with cold EtOH. The filter
cake was dissolved in a small volume of MeOH and loaded onto
silica. The silica column was washed with DCM–MeOH (9 : 1)
to remove phthalhydrazide, and then eluted with DCM–MeOH–
water (7 : 7 : 1). Product-containing fractions were combined and
the solvents were evaporated in vacuo to give the title compound as
a white powder in 44% yield (239 mg, 1.58 mmol): Rf 0.42 (DCM–
MeOH–water 7 : 7 : 1); mp 211 ◦C; dH (400 MHz, DMSO) 3.95 (s,
2H, CH2), 7.27-7.35 (m, 3H, H-3, 4, 5), 7.77 (d, 1H, J = 7.3 Hz, H-
6), 9.27 (br, 3H, NH2, CO2H); dC (75 MHz, DMSO-d6) 44.0 (CH2),
130.1 (C-5), 130.6 (C-3), 131.6 (C-4), 132.3 (C-1, 6), 141.2 (C-2),
169.6 (CO2H); m/z (ESI) 151.0632 (C8H9NO2 requires 151.0633).
Recombinant DPMS Assay. Full experimental details will
be published elsewhere.29 Briefly, E. coli membranes containing
recombinantly expressed, full-length (membrane-bound) T. brucei
DPMS (70 ng ml-1 total protein) was added to a reaction
mix (60 ml) of Na-Hepes (50 mM, pH 7.4), KCl (25 mM),
MgCl2 (5 mM), MnCl2 (5 mM), didehydrofarnesol-phosphate
(1 mg), GDP-[3H]Man (0.25 mCi/1 mM), and potential competing
substrate/inhibitor (1 mM unless otherwise stated) all in the
presence of 0.1% n-octyl-glucopyranoside and incubated at 30 ◦C
for 10min. The reaction was quenched and lipids extracted by
the addition of CHCl3–MeOH (1 : 1) making a final ratio of
10 : 10 : 3 (CHCl3–MeOH–H2O). The resulting supernatant after
centrifugation was dried and partitioned between butan-1-ol
and water, a portion butan-1-ol was counted as a measure of
didehydrofarnesol-phosphate-[3H]mannose product.
Cytotoxicity studies. The Alamar Blue viability test40 was used
to establish ED50 values against cultured bloodstream T. brucei
(strain 427) for all test compounds.
NMR titration experiments. 1H NMR titration experiments
with MgCl2 were carried out with compounds 5b, 5c, 5g, 6c and 7
under the following conditions. Test compounds (ca. 5–6 mmoles)
were dissolved, in separate NMR tubes, in DMF-d7 (ca. 450 mg).
To each tube, a solution of MgCl2 in DMF-d7 (0.35 mmol mg-1)
was added in steps corresponding to a total of 0.00 (“blank”),
0.25–2.00 (steps of 0.25), 2.5–4.0 (steps of 0.5), 5, 6, 8 and 10
Acknowledgements
We thank the Norwich Research Park for a studentship, and the
UEA and the EPSRC (First Grant EP/D059186/1; to GKW)
for financial support. This work was supported in part by a
Wellcome Trust Senior Research Fellowship (067441; to TKS). We
are grateful to Mr Colin MacDonald for assistance with the NMR
experiments, and to the EPSRC National Mass Spectrometry
Service Centre, Swansea, for the recording of mass spectra.
1
equivalents of MgCl2. After each addition of MgCl2, H NMR
spectra were recorded on a Varian VXR 400 S spectrometer at
room temperature, until no significant change of the chemical
shifts for the NH and NH2 protons was observed. For data
analysis, differences in chemical shift for the NH and NH2 protons,
relative to the “blank” sample, were plotted over the equivalents
of MgCl2, as shown in Fig. 5.
References
Molecular docking. The molecular docking experiments were
carried out with AutoDock 3.0.5.44 The X-ray crystal structure of
Kre2p was downloaded from the Protein Data Bank (PDB code
1s4o) and prepared for docking using the AutoDockTools. 1s4o
contains the coordinates of yeast alpha1,2-mannosyltransferase in
complex with GDP and Mn2+,30 which is thought to be the in vivo
ion co-factor. Docking parameters were not available for Mn2+
in versions of AutoDock prior to 4.0, so a number of available
metal ion parameters were evaluated by removing GDP and Mn2+
from the PDB file and comparing the docked metal ion position
to that of the original Mn2+ coordinates. These experiments were
performed on chain B using default parameters on a 40 ¥ 30 ¥
40 grid centered on the Mn2+ position. The calcium parameters
resulted in the best match so these were used in subsequent
1 G. K. Wagner, T. Pesnot and R. A. Field, A survey of chemical methods
for sugar-nucleotide synthesis, Nat. Prod. Rep., 2009, 26, 1172–1194.
2 L. L. Lairson, B. Henrissat, G. J. Davies and S. G. Withers, Gly-
cosyltransferases: structures, functions and mechanisms, Annu. Rev.
Biochem., 2008, 77, 521–555.
3 B. Schuman, J. A. Alfaro and S. V. Evans, Glycosyltransferase structure
and function, Top. Curr. Chem., 2007, 272, 217–257.
4 S. Ha, B. Gross and S. E. Walker, coli MurG: a paradigm for a
superfamily of glycosyltransferases, Curr. Drug Targets: Infect. Disord.,
2001, 1, 201–213.
5 S. Berg, D. Kaur, M. Jackson and P. J. Brennan, The glycosyl-
transferases of Mycobacterium tuberculosis - roles in the synthesis
of arabinogalactan, lipoarabinomannan, and other glycoconjugates,
Glycobiology, 2007, 17, 35R–56R.
6 A. J. Schottelius, A. Hamann and K. Asadullah, Role of fucosyltrans-
ferases in leukocyte trafficking: major impact for cutaneous immunity,
Trends Immunol., 2003, 24, 101–104.
3498 | Org. Biomol. Chem., 2010, 8, 3488–3499
This journal is
The Royal Society of Chemistry 2010
©

View Article Online
7 D. H. Dube and C. R. Bertozzi, Glycans in cancer and inflammation:
potential for therapeutics and diagnostics, Nat. Rev. Drug Discovery,
2005, 4, 477–488.
8 H. Ko, I. Fricks, A. A. Ivanov, T. K. Harden and K. A. Jacobson,
Structure–Activity Relationship of Uridine 5¢-Diphosphoglucose Ana-
logues as Agonists of the Human P2Y14 Receptor, J. Med. Chem., 2007,
50, 2030–2039.
9 A. C. Nottbohm and P. J. Hergenrother, Cyclic Compounds as
Phosphate Mimics, in Wiley Encyclopedia of Chemical Biology,
ed. Tadhg P. Begley, Wiley, New York, 2009, Vol. 3, pp 661–
67610.1002/9780470048672.wecb641.
10 X. Qian, M. M. Palcic, Glycosyltransferase Inhibitors, in Carbohy-
drates in Chemistry and Biology, ed. B. Ernst, G. Hart and P. Sinay¨;
Wiley-VCH, Weinheim, 2000, Part II, Vol. 3; pp 293–312.
11 P. Compain and O. R. Martin, Carbohydrate mimetics-based gly-
cosyltransferase inhibitors, Bioorg. Med. Chem., 2001, 9, 3077–
3092.
12 M. M. Vaghefi, R. J. Bernacki, N. K. Dalley, B. E. Wilson and
Robins, R. K., Synthesis of glycopyranosylphosphonate analogs of
certain natural nucleoside diphosphate sugars as potential inhibitors
of glycosyltransferases, J. Med. Chem., 1987, 30, 1383–1391.
13 M. M. Vaghefi, R. J. Bernacki, W. J. Hennen and R. K. Robins,
Synthesis of certain nucleoside methylenediphosphonate sugars as
potential inhibitors of glycosyltransferases, J. Med. Chem., 1987, 30,
1391–1399.
14 R. R. Schmidt and K. Frische, A new galactosyl transferase inhibitor,
Bioorg. Med. Chem. Lett., 1993, 3, 1747–1750.
15 A. J. Norris, J. P. Whitelegge, M. J. Strouse, K. F. Faull and T. Toyokuni,
Inhibition kinetics of carba- and C-fucosyl analogues of GDP-fucose
against fucosyltransferase V: implication for the reaction mechanism,
Bioorg. Med. Chem. Lett., 2004, 14, 571–573.
16 S. Vidal, I. Bruye`re, A. Malleron, C. Auge´ and J.-P. Praly, Non-
isosteric C-glycosyl analogues of natural nucleotide diphosphate sugars
as glycosyltransferase inhibitors, Bioorg. Med. Chem., 2006, 14, 7293–
7301.
17 R. Chang, T.-T. Vo and N. S. Finney, Synthesis of the C1-phosphonate
analog of UDP-GlcNAc, Carbohydr. Res., 2006, 341, 1998–
2004.
18 R. Wang, D. H. Steensma, Y. Takaoka, J. W. Yun, T. Kajimoto and
C.-H. Wong, A search for pyrophosphate mimics for the development of
substrates and inhibitors of glycosyltransferases, Bioorg. Med. Chem.,
1997, 5, 661–672.
19 B. M. Heskamp, G. H. Veeneman, G. A. Van DerMarel, C. A. A. van
Broeckel and J. H. van Boom, Design and synthesis of a trisubstrate
analogue for a(1→3)fucosyltransferase: A potential inhibitor, Tetrahe-
dron, 1995, 51, 8397–8406.
20 Y. J. Kim, M. Ichikawa and Y. Ichikawa, A rationally designed inhibitor
of -1,3-galactosyltransferase, J. Am. Chem. Soc., 1999, 121, 5829–
5830.
21 J. Grugier, J. Xie, I. Duarte and J.-M. Valery, Synthesis of 2-
(N-acetylamino)-2-deoxy-C-glucopyranosyl nucleosides as potential
inhibitors of chitin synthases, J. Org. Chem., 2000, 65, 979–984.
22 L. Ballell, R. J. Young and R. A. Field, Synthesis and evaluation of
mimetics of UDP and UDP-a-d-galactose, dTDP and dTDP-a-D-
glucose with monosaccharides replacing the key pyrophosphate unit,
Org. Biomol. Chem., 2005, 3, 1109–1115.
23 K. A. Winans and C. R. Bertozzi, An inhibitor of the human UDP-
GlcNAc 4-epimerase identified from a uridine-based library: A strategy
to inhibit O-linked glycosylation, Chem. Biol., 2002, 9, 113–129.
24 G. Gil-Fernandez, S. Perez, P. Vilas, C. Perez, F. G. De las Heras and A.
Garcia Gancedo, Antiviral activity of uridine 5¢-diphosphate glucose
analogs against some enveloped viruses in cell culture, Antiviral Res.,
1987, 8, 299–310.
25 E. Battaglia, A. Elass, R. D. Drake, P. Paul, S. Treat Magdalou, J. S.
Fournle-Gicleux, G. Siest, G. Vergoten, R. Lester and A. Radominska,
Characterization of a new class of inhibitors of the recombinant human
glucuronosyltransferase, UGT1*6, Biochim. Biophys. Acta, Gen. Subj.,
1995, 1243, 9–14.
27 K. Seio, T. Miyashita, K. Sato and M. Sekine, Synthesis and properties
of new nucleotide analogues possessing squaramide moieties as new
phosphate isosteres, Eur. J. Org. Chem., 2005, 5163–5170.
28 T. K. Smith, A. Crossman, J. S. Brimacombe and M. A. J. Ferguson,
Chemical validation of GPI biosynthesis as a drug target against
African sleeping sickness, EMBO J., 2004, 23, 4701–4708.
29 J. Nunes, H. Denton, R. T. Schwarz, A. Collier, G. K. Wagner, T. K.
Smith, manuscript in preparation.
30 Y. D. Lobsanov, P. A. Romero, B. Sleno, B. Yu, P. Yip, A. Her-
scovics and P. L. Howell, Structure of Kre2p/Mnt1p: a yeast a1,2-
mannosyltransferase involved in mannoprotein biosynthesis, J. Biol.
Chem., 2004, 279, 17921–17931.
31 J. Flint, E. Taylor, M. Yang, D. N. Bolam, L. E. Tailford, C. Martinez-
Fleites, E. J. Dodson, B. G. Davis, H. J. Gilbert and G. J. Davies, Struc-
tural dissection and high-throughput screening of mannosylglycerate
synthase, Nat. Struct. Mol. Biol., 2005, 12, 608–614.
32 D. K. Dean, An improved synthesis of 5¢-amino-5¢-deoxyguanosine,
Synth. Commun., 2002, 32, 1517–1521.
33 S. Tomas, R. Prohens, M. Vega, M. C. Rotger, P. M. Deya, P. Ballester
and A. Costa, Squaramido-based receptors: Design, synthesis, and
application to the recognition of tetraalkylammonium compounds,
J. Org. Chem., 1996, 61, 9394–9401.
34 M. C. Rotger, M. N. Pina, A. Frontera, G. Martorell, P. Ballester, P. M.
Deya and A. Costa, Conformational preferences and self-template
macrocyclization of squaramide-based foldable modules, J. Org. Chem.,
2004, 69, 2302–2308.
35 J. H. Sun and W. F. Daneker, An efficient synthesis of 2-(((9-fluorenyl-
methoxycarbonyl)amino)methyl)benzoic acid, Synth. Commun., 1998,
28, 4525–4530.
36 R. Pink, A. Hudson, M.-A. Mourie`s and M. Bendig, Opportunities and
Challenges in Antiparasitic Drug Discovery, Nat. Rev. Drug Discovery,
2005, 4, 727–740.
37 K. Nagamune, T. Nozaki, Y. Maeda, K. Ohishi, T. Fukuma, T. Hara,
R. T. Schwarz, C. Sutterlin, R. Brun, H. Riezman and T. Kinoshita,
Critical roles of glycosylphosphatidylinositol for Trypanosoma brucei,
Proc. Natl. Acad. Sci. U. S. A., 2000, 97, 10336–10341.
38 H. Denton, S. Fyffe and T. K Smith, GDP-Man pyrophosphorylase
is an essential in bloodstream form Trypanosoma brucei, Biochem. J.,
2010, 425, 603–614.
39 T. K. Smith, B. L. Young, H. Denton, D. L. Hughes and G. K. Wagner,
First small molecular inhibitors of T. brucei dolichol phosphate
mannose synthase (DPMS), a validated drug target in African sleeping
sickness, Bioorg. Med. Chem. Lett., 2009, 19, 1749–1752.
40 J. Mikus and D. Steverding, A simple colorimetric method to screen
drug cytotoxicity against Leishmania using the dye Alamar Blue,
Parasitol. Int., 2000, 48, 265–269.
41 Y. Sugawara and H. Iwasaki, Structure of disodium uridine diphospho-
glucose dihydrate, C15H22N2O17P2 ¥ 2 Na+ ¥ 2 H2O, and refinement
2-
of dipotassium glucose 1-phosphate dihydrate, C6H11O9P2- ¥ 2 K+
¥
2 H2O (monoclinic form), Acta Crystallogr., Sect. C: Cryst. Struct.
Commun., 1984, C40, 389–393.
42 M. A. Macnaughtan, M. Kamar, G. Alvarez-Manilla, A. Venot, J.
Glushka, J. M. Pierce and J. H. Prestegard, NMR structural charac-
terization of substrates bound to N-acetylglucosaminyltransferase V,
J. Mol. Biol., 2007, 366, 1266–1281.
43 J. Bornstein, P. E. Drummond and S. F. Bedell, a-Phthalimido-ortho-
toluic acid, Org. Synth., 1963, Coll. Vol. 4, 810–811.
44 G. M. Morris, D. S. Goodsell, R. S. Halliday, R. Huey, W. E. Hart, R. K.
Belew and A. J. Olson, Automated docking using a Lamarckian genetic
algorithm and an empirical binding free energy function, J. Comput.
Chem., 1998, 19, 1639–1662.
45 M. J. S. Dewar, E. G. Zoebisch, E. F. Healy and J. J. P. Stewart, The
development and use of quantum-mechanical molecular models. Part
76. AM1 – a new general purpose quantum mechanical molecular
model, J. Am. Chem. Soc., 1985, 107, 3902–3909.
46 M. W. Schmidt, K. K. Baldridge, J. A. Boatz, S. T. Elbert, M. S.
Gordon, J. H. Jensen, S. Koseki, N. Matsunaga, K. A. Nguyen, S. J.
Su, T. L. Windus, M. Dupuis and J. A. Montgomery, General Atomic
and Molecular Electronic-Structure System, J. Comput. Chem., 1993,
14, 1347–1363.
26 K. Sato, K. Seio and M. Sekine, Squaryl group as a new mimic
of phosphate group in modified oligodeoxynucleotides: Synthesis
and properties of new oligodeoxynucleotide analogues containing an
internucleotidic squaryldiamide linkage, J. Am. Chem. Soc., 2002, 124,
12715–12724.
47 A. C. Wallace, R. A. Laskowski and J. M. Thornton, Ligplot – a
program to generate schematic diagrams of protein ligand interactions,
Protein Eng., Des. Sel., 1995, 8, 127–134.
This journal is
The Royal Society of Chemistry 2010
Org. Biomol. Chem., 2010, 8, 3488–3499 | 3499
©