Structure-based design of nucleoside-derived analogues as sulfotransferase inhibitors

Article abstract of DOI:10.1039/c9ra07567d

Sulfotransferases (STs) catalyse the transfer of a sulfonyl group ('sulfation') from the enzyme co-factor 3′-phosphoadenosine 5′-phosphosulfate (PAPS) to a variety of biomolecules. Tyrosine sulfation of proteins and carbohydrate sulfation play a crucial role in many protein-protein interactions and cell signalling pathways in the extracellular matrix. This is catalysed by several membrane-bound STs, including tyrosylprotein sulfotransferase 1 (TPST1) and heparan sulfate 2-O-sulfotransferase (HS2ST1). Recently, involvement of these enzymes and their post-translational modifications in a growing number of disease areas has been reported, including inflammation, cancer and Alzheimer's disease. Despite their growing importance, the development of small molecules to probe the biological effect of TPST and carbohydrate ST inhibition remains in its infancy. We have used a structure-based approach and molecular docking to design a library of adenosine 3′,5′-diphosphate (PAP) and PAPS mimetics based upon 2′-deoxyadenosine and using 2′-deoxy-PAP as a benchmark. The use of allyl groups as masked methyl esters was exploited in the synthesis of PAP-mimetics, and click chemistry was employed for the divergent synthesis of a series of PAPS-mimetics. A suite of in vitro assays employing TPST1 and HS2ST, and a kinase counter screen, were used to evaluate inhibitory parameters and relative specificity for the STs.

Full text of DOI:10.1039/c9ra07567d

RSC Advances
View Article Online
View Journal | View Issue
PAPER
Structure-based design of nucleoside-derived
analogues as sulfotransferase inhibitors†
Cite this: RSC Adv., 2019, 9, 32165
a
Dominic P. Byrne,b Hollie Parsons,a Neil G. Berry,a
*
Neil M. Kershaw,
David G. Fernig,b Patrick A. Eyers
and Richard Cossticka
b
Sulfotransferases (STs) catalyse the transfer of a sulfonyl group (‘sulfation’) from the enzyme co-factor 30-
phosphoadenosine 50-phosphosulfate (PAPS) to a variety of biomolecules. Tyrosine sulfation of proteins
and carbohydrate sulfation play a crucial role in many protein–protein interactions and cell signalling
pathways in the extracellular matrix. This is catalysed by several membrane-bound STs, including
tyrosylprotein sulfotransferase 1 (TPST1) and heparan sulfate 2-O-sulfotransferase (HS2ST1). Recently,
involvement of these enzymes and their post-translational modifications in a growing number of disease
areas has been reported, including inflammation, cancer and Alzheimer's disease. Despite their growing
importance, the development of small molecules to probe the biological effect of TPST and
carbohydrate ST inhibition remains in its infancy. We have used a structure-based approach and
molecular docking to design a library of adenosine 30,50-diphosphate (PAP) and PAPS mimetics based
upon 20-deoxyadenosine and using 20-deoxy-PAP as a benchmark. The use of allyl groups as masked
methyl esters was exploited in the synthesis of PAP-mimetics, and click chemistry was employed for the
divergent synthesis of a series of PAPS-mimetics. A suite of in vitro assays employing TPST1 and HS2ST,
and a kinase counter screen, were used to evaluate inhibitory parameters and relative specificity for the STs.
Received 18th August 2019
Accepted 3rd October 2019
DOI: 10.1039/c9ra07567d
rsc.li/rsc-advances
STs, the membrane-associated STs exhibit a higher degree of
substrate selectivity.3,7
Introduction
Tyrosine sulfation is a relatively common post-translational
modication of proteins catalysed by tyrosylprotein sulfo-
transferases 1 and 2 (TPST1 and 2) (Fig. 1A). Tyrosine sulfation
was rst reported in bovine brinogen in 1954.8 It has since
been shown to play an important role in a wide range of bio-
logical processes.9 For example, it is implicated in the modu-
lation of extracellular protein–protein interactions such as
those involved in leukocyte adhesion10–12 and hemostasis.13–15 It
also plays a role in visual function,16 agonist binding to
hormone receptors17 and the proteolytic processing of bioactive
peptides.18 Sulfation of several tyrosine residues in the N-
terminal domain of the chemokine receptor CCR5 has been
shown to be crucial in mediating HIV entry into cells.19,20 In
addition, tyrosine sulfation of the Duffy antigen/receptor for
chemokines (DARC) is important for the malaria parasite Plas-
modium vivax binding and entry into erythrocytes.21
Heparan sulfates (HS) are highly sulfated GAGs that can exist
either conjugated to proteins or as free, oen extended, chain-
forming fragments.22 HS has been shown to bind to extracel-
lular growth factors, cytokines, enzymes and cell matrix
proteins, and also to act as co-receptors to generate signalling-
competent growth factor complexes.23,24 As a component of
the extracellular matrix, HS has roles in important processes
such as cell adhesion,25 viral invasion,26 and Alzheimer's
disease.27 HS 2-O-sulfotransferase (HS2ST1) transfers a sulfonyl
The post-translational covalent modication of biomolecules
through the addition of a sulfonyl group (SO3ꢀ) to a hydroxyl
or amino acceptor moiety is of importance in a very broad
range of biological processes.1,2 The ubiquitous donor of
activated sulfate is 30-phosphoadenosine 50-phosphosulfate
(PAPS). Sulfation (also termed sulfonation) is catalysed by
enzymes called sulfotransferases (STs) which are separated
into two general classes. Cytosolic STs (SULTs) sulfate both
small endogenous and exogenous compounds, including
hormones, neurotransmitters, and a variety of pharmaceutical
agents, and have critical roles in detoxication and excre-
tion.3,4 In contrast, the membrane-associated STs are located
in the Golgi apparatus and sulfate endogenous secreted
macromolecules such as proteins and polysaccharides,
including glycosaminoglycans (GAGs).5,6 The STs possess
a structurally-conserved PAPS-binding domain, but exhibit
distinct substrate binding sites, and in contrast with cytosolic
aDepartment of Chemistry, University of Liverpool, Liverpool L69 7ZD, UK. E-mail:
nkershaw@liverpool.ac.uk
bDepartment of Biochemistry, Institute of Integrative Biology, University of Liverpool,
Liverpool L69 7ZB, UK
† Electronic supplementary information (ESI) available: Details of molecular
docking, additional gures, recombinant protein production, assay details,
experimental procedures and characterisation of compounds. See DOI:
10.1039/c9ra07567d
This journal is © The Royal Society of Chemistry 2019
RSC Adv., 2019, 9, 32165–32173 | 32165

View Article Online
RSC Advances
Paper
Fig. 1 (A) Tyrosylprotein sulfotransferase catalysed sulfation of a tyrosine residue. (B) Heparan sulfate 2-O-sulfotransferase catalysed sulfation of
an iduronic acid (IdoA) subunit of heparan sulfate.
group to the 2-OH of iduronic acid (IdoA) and more rarely to represents an attractive therapeutic option in oncology. A range
glucuronic acid (GlcA) to give IdoA2S and GlcA2S, respectively of inhibitors have been discovered based upon natural prod-
(Fig. 1B).28 HS2ST-knockout mouse models have revealed ucts,37 polychlorinated phenols,38 and compounds designed to
a crucial role for 2-O-sulfated HS in kidney and eye develop- be bi-substrate mimics.39,40 More recently, irreversible clinical
ment, and for nervous system function.29 Additionally, 2-O-sul- steroid sulfatase inhibitors such as the orally active non-
fation of HS is essential for regulation of broblast growth steroidal aryl sulfamate irosustat (STX64) have entered clinical
factor (FGF)-2 signalling,30 and has been shown to play a role in trials.41 Related to sulfation, the reversible process of protein
prostate cancer cell proliferation in vitro.31
phosphorylation, which was also originally described in the
Given the importance that regulated sulfation plays in 1950s, is catalysed by >500 diverse members of the human
a plethora of biological processes, it is somewhat surprising protein kinase superfamily.42 In contrast to sulfation, the kinase
that modulation of the TPSTs and carbohydrate STs using small eld is approaching maturity in the biomedical context, with
molecules remains relatively understudied.32–36 In contrast, the over 50 kinase inhibitors approved in cancer and inammatory
SULTs have received signicant attention, mainly focussed on disease.43–47 Of interest, we recently described facile real-time
estrogen (steroid) sulfotransferase (SULT1E) through modula- biochemical assays, which were previously developed for ana-
tion of the substrate binding site. Many human tumours are lysing phosphorylation in the kinase eld, to evaluate tyrosine
hormone-dependent, playing a central role in their develop- and carbohydrate sulfation in vitro. Using this high-throughput
ment and survival. Estrogen sulfatase inhibition therefore enzymatic approach, we discovered that HS2ST and the TPSTs
Fig. 2 Crystallographic analysis of the sulfotransferase active site. (A) Nucleoside-binding domain of TPST1 complexed with PAP (PDB ID: 5WRI).
TPST1 is rendered as grey cartoon. Residues interacting with PAP are labelled and rendered as thin sticks (carbon – grey, nitrogen – blue, oxygen
– red). Crystallographic waters are rendered as slate spheres PAP is rendered as coloured sticks (carbon – green, nitrogen – blue, oxygen – red).
Black dotted lines indicate hydrogen bonds. (B) Nucleoside-binding domain of HS2ST complexed with PAP (PDB ID: 4NDZ). HS2ST is rendered as
grey cartoon. Residues interacting with PAP are labelled and rendered as thin sticks (carbon – grey, nitrogen – blue, oxygen – red). PAP is
rendered as coloured sticks (carbon – green, nitrogen – blue, oxygen – red). Black dotted lines indicate hydrogen bonds.
32166 | RSC Adv., 2019, 9, 32165–32173
This journal is © The Royal Society of Chemistry 2019

View Article Online
Paper
RSC Advances
Fig. 3 Summary of chemical structures. (A) PAP-mimetics (B) PAPS-mimetics.
were novel targets for several families of previously known
Results and discussion
Crystallographic analysis, compound selection and molecular
protein kinase inhibitor, suggesting that the ATP and PAPS
binding site of these classes of enzyme might share similar
druggable features.48,49
docking
The crystal structures of TPST1 and HS2ST (PDB IDs: 5WRI50 and
4NDZ51) bound to PAP and sulfate acceptor substrates were used
to facilitate the rational design of potential enzyme inhibitors.
Central to the nucleoside-binding domain of the STs are highly-
conserved structural features termed the 50-phosphosulfate-
binding (50-PSB) and 30-phosphate-binding (30-PB) loops (Fig. SI
1A and B†). These make a network of hydrogen bonds involving
the protein surface and the 30- and 50-phosphates of PAP. The
adenine ring is involved in a hydrophobic or p-stacking interac-
tion and a hydrogen bond to N-6 is frequently observed.
Comparison of the structures of TPST1 and HS2ST bound to
PAP reveals very similar hydrogen bonding interactions between
the 30-phosphate of PAP and the sidechains of a number of basic
residues (Fig. 2A and B). The 50-phosphate of PAP also forms
a number of analogous interactions in both structures.
However, in TPST1 a water molecule bridges the 40- and 50-
oxygens, whereas in HS2ST a backbone interaction with Ala-
1085 performs this function. The hydrophobic interaction
with adenine is made by Leu-84 in TPST1 and Ala-1085 in
HS2ST. N-6 makes a number of contacts in both structures. In
TPST1, N-1 makes a backbone interaction with Asn-295 and N-3
makes a sidechain interactions with Tyr-239. In contrast, N-1
and N-3 do not make any contacts in HS2ST but N-7 accepts
a hydrogen bond from the sidechain of Ser-1088. The 20-
hydroxyl of PAP forms a hydrogen with the 30-phosphate but
does not make any contacts to the protein in either structure.
In this report, we describe a structure-based approach to ST
inhibitor design, based on our understanding of co-factor and
substrate binding. Targeting the substrate binding site of the
SULTS has previously been exploited in the design of small
molecule inhibitors.37–40 However, this strategy would be
signicantly more challenging for the membrane-associated
STs as the binding interactions are protein–protein/protein–
polysaccharide in nature. We opted to target the PAPS binding
site with the intention of developing a general ST inhibitor
initially. Subsequent exploitation of differences in the substrate
binding sites could then conceivably be used to develop selec-
tivity once general inhibition had been achieved. It is note-
worthy that many protein kinase inhibitors exert their action
through competitive inhibition of the ATP (analogous to PAPS)
binding site. We utilised the published crystal structures of
TPST1 complexed to adenosine 30-50-diphosphate (PAP) and
substrate peptides50 and HS2ST1 bound to PAP and a hepta-
saccharide.51 Analysis of the PAP-binding site and subsequent
molecular docking were utilised to guide the design of a library
of PAP and PAPS mimetics based upon 20-deoxyadenosine.
Compounds were subsequently assessed for inhibitory activity
towards puried HS2ST1 and TPST1 in the presence of the
sulfate donor PAPS using our microuidic-shi assays. Finally,
a counter screen against the canonical ATP-dependent Ser/Thr
kinase protein kinase A (PKA) was performed to gauge relative
specicity for ST inhibition.
Scheme 1 Reagents and conditions: (a) TMSCl, pyridine, MMTrCl, 0 ꢁC to RT then aq. NH3; (b) NaHMDS, allyl bromide, TBAI, DMF, ꢀ20 ꢁC to RT;
(c) O3, NaOH, MeOH, DCM, ꢀ78 ꢁC, then AcOH; (d) NaHMDS, methyl bromoacetate, TBAI, DMF, ꢀ20 ꢁC to RT.
This journal is © The Royal Society of Chemistry 2019
RSC Adv., 2019, 9, 32165–32173 | 32167

View Article Online
RSC Advances
Paper
Given that the 20-hydroxyl does not appear to directly
contribute to binding we reasoned that it would be feasible to
remove it in our subsequent inhibitor design and thus focus on
derivatives of 20-deoxyadenosine. Although 20-deoxyribonucleo-
sides are known to preferentially adopt a different sugar pucker
conformation compared with ribose nucleosides, the inversion
barrier is small (ca. 20 kJ molꢀ1) and the conformers are in rapid
equilibrium in solution.52 20-Deoxyadenosine has also previ-
ously served as a starting point for SULT inhibitor design.40 This
strategy would have the additional benet of reducing the need
for the lengthy protecting group strategies required for selective
modication of ribonucleosides. 20-Deoxy-PAP represents the
closest analogue to PAP based on this scaffold and would serve
as a benchmark from which to compare our library of
compounds (Fig. 3A). Surprisingly, this compound has not
previously been assessed as a ST inhibitor.
A series of PAP-mimetics probing the binding requirements
at the 30- and 50-positions were proposed, with isosteric
replacement of the phosphates for less metabolically labile
substituents. Four functional groups were selected, with
differing pKa values such that a range of ionisation states at
a physiologically relevant pH could be assessed (1–4) (Fig. 3A).
Bis-methyl ester 1 and bis-amide 2 would be non-ionised,
whereas bis-hydroxamic acid 3 and bis-carboxylic acid 4
would be half- and fully-ionised, respectively. Compounds 5 and
6 were selected as a result of an unanticipated differentiation of
the 30- and 50-hydroxyls during the synthesis of 1–4, vide infra.
Analogues of PAPS containing a hydrolytically stable isostere
of the 50-phosphosulfate were also considered. Compounds 7–
14 were selected (Fig. 3B) bearing a triazole as a phosphate
bioisostere,53,54 and either a carboxylate or phosphate terminus
to mimic the sulfate of PAPS. The presence or absence of the 30-
phosphate would also be assessed (e.g. 7 vs. 8). Compounds
bearing a benzyl group on the 50-carboxylate or phosphate
terminus (6, 9–10, 13–14), were selected to investigate the
possibility of occupying both the nucleoside and substrate
binding sites. Whilst it was recognised that PAP and PAPS
mimetics would not achieve selective inhibition of one ST over
another, it could be anticipated that selectivity could be
subsequently designed into a potent, general ST inhibitor.
A molecular docking study was used to predict the mode
and relative strength of compound binding. A protocol was
rst developed that reproduced the binding pose of PAP with
TPST1 and HS2ST accurately (see ESI for details,† RMSD, 0.04
Synthesis
20-Deoxy-PAP was synthesised using a modied literature
procedure (see ESI for details†).55 It was envisaged that a suit-
ably protected derivative of bis-methyl ester 1 would act as
a point of late-stage diversication to access 1–4. The synthesis
of this key intermediate (7) is summarised in Scheme 1. Mon-
omethoxytrityl (MMTr) protection of 20-deoxyadenosine using
the transient protection protocol occurred in high yield to give
15.56,57 It had been expected that the required methylene methyl
ester functionality at the 30- and 50-positions could be intro-
duced though bis-alkylation of 15 using methyl bromoacetate.58
Unfortunately, under a range of conditions, only decomposition
of the starting material was observed. The reason for decom-
position is not clear, as the reaction was successful using t-butyl
bromoacetate.59 However, due to the relatively harsh conditions
required to remove the t-butyl groups, it did not prove possible
to advance this material through to 17. An alternative strategy
was devised using allyl groups as latent methyl esters.60 Double
allylation of 15 proceeded smoothly, and was followed by
a modied ozonolysis to provide key intermediate 17 in a 54%
yield (31% overall from 20-deoxyadenosine).61–63
Scheme 2 Reagents and conditions: (a) 80% aq. AcOH, RT; (b) (i) 6 M
NH3, MeOH, RT; (ii) 80% aq. AcOH, RT; (c) (i) 50% wt. aq. NH2OH,
MeOH, RT; (ii) 80% aq. AcOH, RT; (d) (i) 1 M aq. NaOH, MeOH, RT; (ii)
80% aq. AcOH, RT.
ꢀ
ꢀ
A and 0.13 A respectively). This protocol was then employed
to dock PAPS, 20-deoxy-PAP and 1–14. Phosphates were
docked in their monobasic form. Carboxylic acids were
docked as carboxylates. All were predicted to bind in an
analogous fashion to PAP in both TPST1 and HS2ST, and the
docking scores were also very similar (Table SI 1, Fig. SI 2A–
Q†). The 50-triazolecarboxylate/phosphate of 7–14 overlaid
well with the predicted binding pose for PAPS (Fig. SI 3†). The
benzyl groups of 6, 9–10, 13–14 were predicted to span the
region between the nucleoside and substrate binding sites
(Fig. SI 4†). This provided condence in the synthesis of 20-
deoxy-PAP and compounds 1–14.
Scheme 3 Reagents and conditions: (a) NaHMDS, allyl bromide, TBAI,
DMF, ꢀ45 ꢁC to RT; (b) O3, NaOH, MeOH, DCM, ꢀ78 ꢁC; (c) i-Pr2-
NP(OBn)2, 1H-tetrazole, MeCN, 0 ꢁC to RT then mCPBA; (d) (i) 10% Pd/
C, H2, 1 atm, Et3N, MeOH, RT, then Na+-Dowex®; (ii) 80% aq. AcOH,
RT; (e) (i) 80% aq. AcOH, RT; (ii) NaI, MeCN, 80 ꢁC.
32168 | RSC Adv., 2019, 9, 32165–32173
This journal is © The Royal Society of Chemistry 2019

View Article Online
Paper
RSC Advances
During the double allylation of 15, it was noticed that a small
amount of a mono-allylated product was also produced. Acety-
lation of the puried product conrmed that allylation had
occurred at the 30-position (Fig. SI 5†). No mono-allylation at the
50-position could be detected. Selective 20-O-alkylation of
unprotected ribonucleosides is an established process under
a number of conditions.65,66 However, protection of the 50-
hydroxyl is normally required for selective modication of the
30-hydroxyl of 20-deoxynucleosides.61 We speculate that this
selectivity is caused by steric hindrance of the 50-OH by the
bulky MMTr group as 15 is resistant to standard 50-halodeox-
ygenation conditions whereas the unprotected nucleoside
reacts smoothly. Further studies into this novel selectivity is
currently underway. Although unexpected, it was recognised
that this selectivity could potentially be exploited to differen-
tiate the 30- and 50-positions and thus generate a number of non-
Scheme 4 Reagents and conditions: (a) alkyne, 5 mol% CuSO4$5H2O, symmetric PAP mimetics. Aer signicant optimisation,
conditions were developed to reliably furnish the 30-allylated
product 18 in a 40% yield (75% based on recovered starting
material) (Scheme 3). Performing the addition of allyl bromide
above ꢀ45 ꢁC improved the conversion of starting material, but
10 mol% sodium ascorbate, t-BuOH/H2O (1 : 1), RT; removal of the
benzyl group of 9 under hydrogenative conditions occurred in
excellent yield to provide 7. Conversion of 9 to the 30-dibenzylphos-
phate followed by exhaustive hydrogenation provided 8 in a 66% yield
over the two steps. An alternative phosphate protecting group was
required for the synthesis of 10. Accordingly, formation of the bis- at the expense of the isolated yield of 18. This was then
cyanoethyl-protected 30-phosphate afforded 23 in a 75% yield.73 This
advanced to the orthogonally protected nucleoside 20 through
a two-step procedure. First, the alkene was oxidatively cleaved,
as previously described, and then the 50-hydroxyl was converted
was followed by selective deprotection using the method of Gaffney
and Reese74 to provide 10 in a 74% yield (Scheme 5).
to the dibenzylphosphate using standard phosphoramidite
conditions.67
With 17 in hand, completion of the synthesis of 1–4 was
achieved in short order. Accordingly, deprotection of the MMTr
group using aqueous AcOH furnished bis-methyl ester 1 in
quantitative yield.64 Treatment of 17 with 6 M methanolic
ammonia followed by MMTr deprotection gave bis-amide 2 in
an 85% yield over two steps. Conversion to bis-hydroxamic acid
Removal of benzyl groups under hydrogenative conditions68
followed by MMTr deprotection proceeded smoothly to give 5 in
a 91% yield. Alternatively, cleavage of the MMTr group followed
by selective removal one of the benzyl groups, by treatment with
sodium iodide in acetonitrile at elevated temperature,
completed the synthesis of 6 in good overall yield (59%).69
3 was achieved in a similar fashion, albeit in lower yield. Finally,
The synthesis of PAPS mimetics 7–14 commenced from 20-
basic ester hydrolysis of 17 and MMTr deprotection proceeded
deoxy-50-azidoadenosine (21)70 via a copper-catalysed alkyne
in high yield to give bis-acid 4 (Scheme 2).
Scheme 5 Reagents and conditions: (a) 10% Pd/C, H2, 1 atm, Et3N, MeOH, RT, then Na+-Dowex®; (b) (i) i-Pr2NP(OBn)2, 1H-tetrazole, MeCN,
0 ꢁC to RT then mCPBA; (ii) 10% Pd/C, H2, 1 atm, Et3N, MeOH, RT, then Na+-Dowex®; (c) i-Pr2NP(OCH2CH2CN)2, 1H-tetrazole, MeCN, 0 ꢁC to RT
then mCPBA; (d) TMG, TMSCl, MeCN, RT, then aq. NH3; (e) NaI, MeCN, 80 ꢁC; abbreviations: TMG ¼ 1,1,3,3-tetramethylguanidine.
This journal is © The Royal Society of Chemistry 2019
RSC Adv., 2019, 9, 32165–32173 | 32169

View Article Online
RSC Advances
Paper
acetylene cycloaddition (CuAAC), or “click” coupling, with the PAP and coenzyme A (CoA) as the positive controls (both are
appropriate alkyne partner.71 21 was synthesised in two steps competitive PAPS inhibitors), and the pan-kinase inhibitor
from 20-deoxyadenosine (See ESI for details†). CuAAC under staurosporine A as a negative control. Compounds were initially
standard conditions with benzyl propiolate gave triazole 9 in screened at 400 mM in the presence 20 nM HS2ST and 0.1 mM
excellent yield. Reaction of 21 with dibenzyl ethynylphospho- TPST1 respectively. Full dose–response curves were then ob-
nate72 also proceeded smoothly to provide 22 in good yield tained for the most active compounds. The results are sum-
(Scheme 4).
Hydrogenation of 22 gave 11 in excellent yield (89%). 22
marised in Fig. 4A–D.
20-Deoxy-PAP exhibited moderate inhibitory activity against
could also be converted to the 30-dibenzylphosphate which gave HS2ST (IC50 ¼ 12.7 mM ꢂ 1.2) and TPST1 (IC50 ¼ 3.6 mM ꢂ 1.2)
12 upon global deprotection in a 47% yield. Treatment of 22 compared with PAP (IC50 ¼ 2.0 mM and 1.5 mM respectively)
with sodium iodide in acetonitrile, as previously described, gave (Fig. 4A–D). This demonstrates that the 20-hydroxyl is not
13 in moderate yield. Finally, 30-phosphorylation of 13 and essential for binding. However, despite not appearing to make
deprotection gave 14 in a 56% yield over two steps (Scheme 5). any contacts with the protein, its absence does reduce potency
somewhat. This raises the possibility that the hydrogen bond
from the 20-hydroxyl to the 30-phosphate (Fig. 2A and B) may
function as a conformational lock,75 reducing the entropic loss
Determination of ST inhibitory activity and a simple kinase
counter-screen.
on binding to the protein and warrants further study.
The ability of 20-deoxy-PAP and 1–14 to inhibit HS2ST and
TPST1 were initially evaluated by in vitro enzyme assay, using
Fig. 4 Summary of enzymatic inhibition assays against HS2ST and TPST1. (A) Enzymatic analysis of HS2ST1 inhibition by a panel of PAP- and
PAPS-mimetic compounds. MBP-tagged HS2ST (20 nM) was incubated with PAPS (5 mM) in the presence of the appropriate nucleoside analogue
(400 mM). HS2ST sulfotransferase activity was assayed using a fluorescent hexasaccharide substrate (2 mM) and normalised to DMSO (4% v/v) or
buffer control. Data shown is mean and SD of 4 repeat experiments. (B) Full dose–response curves for selected compounds. HS2ST (20 nM) was
incubated with increasing concentration of the indicated compound in the presence of PAPS (5 mM) for 15 min at 20 ꢁC. HS2ST activity calculated
as previously described. Data from two independent experiments are combined. (C) Enzymatic analysis of TPST1 inhibition by a panel of PAP- and
PAPS-mimetic compounds. TPST1 (0.1 mM) was incubated with PAPS (5 mM) in the presence of the appropriate nucleoside analogue (400 mM) for
30 min at 20 ꢁC. TPST1 activity was measured using fluorescently-labelled CC4-tide (2 mM) and normalised to DMSO (4% v/v) or buffer control.
(D) Full dose–response curves for selected compounds. TPST1 (0.1 mM) was incubated with increasing concentrations of the indicated
compound in the presence of PAPS (5 mM) for 30 min at 20 ꢁC. TPST1 activity was measured using CC4-tide and normalised to DMSO or buffer
controls as previously described. The data shown is from duplicate experiments.
32170 | RSC Adv., 2019, 9, 32165–32173
This journal is © The Royal Society of Chemistry 2019

View Article Online
Paper
RSC Advances
Fig. 5 Immunological evaluation of TPST1 inhibition towards substrate and a Ser/Thr protein kinase counter-screen. (A) Immunoblots evaluating
TPST1 sulfotransferase activity in the presence of a panel of PAP- and PAPS-mimetic compounds. GST-CC4-tide (1 mg) was incubated in the
presence of TPST1 (0.2 mg), PAPS (5 mM), and a fixed concentration of the indicated compound (400 mM) for 15 min. After termination of the
reaction using SDS-PAGE sample buffer, tyrosine sulfation was visualised by immunoblotting using a monoclonal sulfotyrosine antibody (top
panel), with equal GST-CC4-tide and TPST1 loading confirmed using an antibody to detect 6xHis tagged proteins (bottom panel). (B) Enzymatic
inhibition of PKA catalytic activity by a panel of PAP- and PAPS-mimetic compounds. PKA kinase (1 nM) was incubated with ATP (5 mM) in the
presence of the appropriate nucleoside analogue (400 mM) for 30 min at 20 ꢁC. PKA activity was calculated in real-time using fluorescently-
labelled substrate peptide (2 mM) and normalised to DMSO (4% v/v) or buffer control.77 Data in B is mean and SD of 3 individual experiments.
Staurosporine is included at 40 mM as a generic inhibitor of kinase activity. For A, similar results were seen in two independent experiments.
None of the PAP-mimetics (1–6) showed any inhibitory recognises sulfotyrosine, using PAP as positive control (Fig. 5A).
activity against HS2ST (Fig. 4A). Carboxylic acid derivatives can Compounds were screened at 400 mM in the assay. In common
make electrostatic interaction with residues such as a lysine and with the TP real-time ST1 enzyme assay using a synthetic
arginine, which are oen present in phosphate-binding protein peptide, a broad range of inhibitory activities was observed. For
domains.76 However, the trigonal geometry of carboxylates example, 20-deoxy-PAP, and PAP-mimetics 5 and 6 showed
differs from the tetrahedral phosphate group. This may complete suppression of tyrosine sulfation at this concentra-
preclude efficient binding and alternative isosteres are currently tion. In contrast, bis-hydroxamic acid 3, bis-acid 4, and 8, 9, 12
being investigated. Several of the triazole-containing PAPS- and 13 showed partial suppression of tyrosine sulfation.
mimetics (8, 9, 12 and 13) were weak inhibitors of HS2ST with However, PAPS-mimetic 7 did not show any activity in this
IC50 values ranging from 100 to 300 mM (Fig. 4B). An identical assay.
structure–activity relationship (SAR) was observed for both the
Finally, a simple counter-screen was set up employing PKA.
triazole-carboxylate (7–10) and triazole-phosphate (11–14) This was carried out in order to assess relative specicity for ST
series. For compounds bearing a free carboxylate or phosphate inhibition, and using the generic kinase inhibitor staurosporine
(7, 8, 11 and 12), 30-phosphorylation was essential for activity A as a positive control and PAP and CoA as negative controls,
(e.g. 7 vs. 8). In contrast, the corresponding benzylated neither of which bind nor inhibit PKA. Compounds were
analogues (9, 10, 13 and 14) showed activity only in the absence screened at 400 mM in the presence 1 nM PKA. The results are
of 30-phosphorylation (e.g. 9 vs. 10).
summarised in Fig. 5B. Interestingly, 20-deoxy-PAP showed no
Similar results were obtained using TPST1, although activity against PKA, which demonstrates clearly that selective
a broader range of inhibitory activities, compared with HS2ST, ST inhibition can be achieved from the 20-deoxyadenosine
was observed (Fig. 4C). PAP-mimetics 5 and 6, and PAPS- scaffold. With the exception of 11, which was inactive in all of
mimetics 7–9, 12 and 13 were weak inhibitors with IC50 the ST assays, 1–14 all showed some inhibitory activity towards
ranging from 100 to 300 mM (Fig. 4D). Comparing 5 and 6, the PKA indicating a lack of specicity for STs over a generic protein
presence a benzyl group on the 50-phosphate led to a decrease in kinase.
activity. The SAR for PAPS-mimetics 7–14 was very similar to
Given that members of this rst generation set of
that observed against HS2ST although compound 7, which compounds are only modest inhibitors of TPST1 and HS2ST,
bears a free carboxylate and lacks a 30-phosphate, also showed and are derived from 20-deoxyadenosine, it is perhaps unsur-
activity against TPST1.
prising that this promiscuous inhibitory activity towards
In order to further assess TPST1 inhibition by 20-deoxy-PAP a protein kinase is observed. It is noteworthy that 20-deoxy-PAP
and compounds 1–14 using a complementary assay, immuno- retained much of the ST inhibitory activity compared with PAP
blotting of a TPST1 protein substrate (GST-CC4tide) was and also showed no PKA inhibition. Further work into relevant
implemented with a monoclonal antibody that specically structural features of PAP and 20-deoxy-PAP that engender ST-
This journal is © The Royal Society of Chemistry 2019
RSC Adv., 2019, 9, 32165–32173 | 32171

View Article Online
RSC Advances
Paper
selective inhibition is ongoing, and the PAP- and PAPS-
mimetics described in this study provide useful starting
points for future development.
4 M. Negishi, L. G. Pedersen, E. Petrotchenko, S. Shevtsov,
A. Gorokhov, Y. Kakuta and L. C. Pedersen, Arch. Biochem.
Biophys., 2001, 390, 149–157.
5 Y. S. Yang, C. C. Wang, B. H. Chen, Y. H. Hou, K. S. Hung and
Y. C. Mao, Molecules, 2015, 20, 2138–2164.
6 J. Liu, A. F. Moon, J. Sheng and L. C. Pedersen, Curr. Opin.
Struct. Biol., 2012, 22, 550–557.
7 Y. Kakuta, L. G. Pedersen, L. C. Pedersen and M. Negishi,
Trends Biochem. Sci., 1998, 4, 129–130.
8 F. R. Bettelheim, J. Am. Chem. Soc., 1954, 76, 2838–2839.
9 K. L. Moore, J. Biol. Chem., 2003, 278, 24243–24246.
Conclusions
Analysis of the nucleoside-binding domains of TPST1 and
HS2ST and molecular docking identied 20-deoxy-PAP as
a structurally simplied analogue PAP. This led to the design of
a library of potential inhibitors of the STs based on 20-deoxy-
adenosine. Divergent syntheses of a range PAP- and PAPS-
mimetics have been successfully developed. Evaluation of the 10 V. Ramachandran, M. U. Nollert, H. Qiu, W. J. Liu,
inhibitory activity of the synthetic targets against HS2ST and
TPST1 was carried out using our previously reported in vitro
R. D. Cummings, C. Zhu and R. P. McEver, Proc. Natl.
Acad. Sci. U. S. A., 1999, 96, 13771–13776.
enzyme assays and a complementary immunoblot technique. A 11 S. D. Rodgers, R. T. Camphausen and D. A. Hammer,
PKA counter-screen was also developed to assess specicity for Biophys. J., 2001, 81, 2001–2009.
ST inhibition. 20-Deoxy-PAP, which to our knowledge has not 12 J. H. Ippel, C. J. de Haas, A. Bunschoten, J. A. van Strijp,
previously been assessed as an ST inhibitor, was demonstrated
to have attenuated activity towards HS2ST and TPST1 compared
J. A. Kruijtzer, R. M. Liskamp and J. Kemmink, J. Biol.
Chem., 2009, 284, 12363–12372.
with PAP. This raises questions surrounding the intramolecular 13 G. L. Hortin, Blood, 1990, 76, 946–952.
hydrogen bond between the 20-hydroxyl and 30-phosphate of 14 A. Leyte, H. B. van Schijndel, C. Niehrs, W. B. Huttner,
PAP and the possibility that it may play a pre-organisational role
in binding. PAPS-mimetics 8, 9, 12 and 13 were identied in all
M. P. Verbeet, K. Mertens and J. A. van Mourik, J. Biol.
Chem., 1991, 266, 740–746.
ST assays and, whilst weak inhibitors, represent opportunities 15 D. A. Michnick, D. D. Pittman, R. J. Wise and R. J. Kaufman,
for further investigation. We have previously identied, low/ J. Biol. Chem., 1994, 269, 20095–20102.
sub-micromolar sulfotransferase inhibitors such as rottlerin 16 Y. Kanan, J. C. Siefert, M. Kinter and M. R. Al-Ubaidi, PLoS
and suramin (non nucleoside-derived).48,49 The next step will be
One, 2014, 9, e105409.
to obtain co-crystal structures of these and 8, 9, 12 and 13 with 17 S. Costagliola, V. Panneels, M. Bonomi, J. Koch, M. C Many,
TPST1 and HS2ST which we are currently undertaking using G. Smits and G. Vassart, EMBO J., 2002, 21, 504–513.
published protocols. This will allow further structure-based 18 J. R. Bundgaard, J. Vuust and J. F. Rehfeld, EMBO J., 1995, 14,
compound design to be used. Interestingly, a broader range of 3073–3079.
inhibitory activities was shown against TPST1 (which sulfates 19 M. Farzan, T. Mirzabekov, P. Kolchinsky, R. Wyatt,
proteins) relative to HS2ST (which sulfates glycans), which hints
at the possibility of achieving selective ST inhibition in the
M. Cayabyab, N. P. Gerard, C. Gerard, J. Sodroski and
H. Choe, Cell, 1999, 96, 667–676.
future, perhaps by additional targeting of the distinct substrate- 20 E. G. Cormier, M. Persuh, D. A. Thompson, S. W. Lin,
binding pockets in these enzymes.
T. P. Sakmar, W. C. Olson and T. Dragic, Proc. Natl. Acad.
Sci. U. S. A., 2000, 97, 5762–5767.
21 H. Choe, M. J. Moore, C. M. Owens, P. L. Wright, N. Vasilieva,
W. Li, A. P. Singh, R. Shakri, C. E. Chitnis and M. Farzan,
Mol. Microbiol., 2005, 55, 1413–1422.
Conflicts of interest
There are no conicts to declare.
22 J. R. Bishop, M. Schuksz and J. D. Esko, Nature, 2007, 446,
1030–1037.
23 A. Nagarajan, P. Malvi and N. Wajapeyee, Front. Endocrinol.,
2018, 9, 483.
Acknowledgements
We would like to thank the following for nancial support:
European Commission FET-OPEN grant [ArrestAD no. 737390] 24 S. Sarrazin, W. C. Lamanna and J. D. Esko, Cold Spring
(to DGF and PAE). BBSRC grant number BB/N021703/1 (to PAE).
NWC (to DGF).
Harbor Perspect. Biol., 2011, 3, a004952.
25 R. J. Weiss, J. D. Esko and Y. Tor, Org. Biomol. Chem., 2017,
15, 5656–5668.
26 S. Y. Kim, J. Zhao, X. Liu, K. Fraser, L. Lin, X. Zhang,
F. Zhang, J. S. Dordick and R. J. Linhardt, Biochemistry,
2017, 56, 1151–1162.
27 P. Chiodelli, A. Bugatti, C. Urbinati and M. Rusnati,
Molecules, 2015, 20, 6342–6388.
References
1 W. Leung, I. Backstrom and M. B. Bally, Oncotarget, 2016, 7,
55811–55827.
2 N. Gamage, A. Barnett, N. Hempel, R. G. Duggleby,
K. F. Windmill, J. L. Martin and M. E. McManus, Toxicol. 28 H. N. Bethea, D. Xu, J. Liu and L. C. Pedersen, Proc. Natl.
Sci., 2006, 90, 5–22. Acad. Sci. U. S. A., 2008, 105, 18724–18729.
3 E. Chapman, M. D. Best, S. R. Hanson and C. H. Wong, 29 C. L. Merry and V. A. Wilson, Biochim. Biophys. Acta, 2002,
Angew. Chem., Int. Ed., 2004, 43, 3526–3548.
1573, 319–327.
32172 | RSC Adv., 2019, 9, 32165–32173
This journal is © The Royal Society of Chemistry 2019

View Article Online
Paper
RSC Advances
30 S. Ashikari-Hada, H. Habuchi, N. Sugaya, T. Kobayashi and 53 H. Isobe, T. Fujino, N. Yamazaki, M. Guillot-Nieckowski and
K. Kimata, Glycobiology, 2009, 6, 644–654. E. Nakamura, Org. Lett., 2008, 10, 3729–3732.
31 B. W. Ferguson and S. Datta, Prostate Cancer, 2011, 2011, 54 A. S. Rowan, N. I. Nicely, N. Cochrane, W. A. Wlassoff,
a893208.
A. Claiborne and C. J. Hamilton, Org. Biomol. Chem., 2009,
7, 4029–4036.
55 B. L Hevroni, A. H. Sayer, E. Blum and B. Fischer, Inorg.
Chem., 2014, 53, 1594–1605.
32 F. Vargas, M. D. Tuong and J. C. Schwartz, J. Enzyme Inhib.,
1986, 1, 105–112.
33 J. W. Kehoe, D. J. Maly, D. E. Verdugo, J. I. Armstrong,
B. N. Cook, Y. B. Ouyang, K. L. Moore, J. A. Ellman and 56 G. S. Ti, B. L. Gaffney and R. A. Jones, J. Am. Chem. Soc., 1982,
C. R. Bertozzi, Bioorg. Med. Chem. Lett., 2002, 12, 329–332. 104, 1316–1319.
34 W. Zhou, Y. Wang, J. Xie and R. J. Geraghty, Biochem. 57 M. Suzuki, T. Sekido, S.-i. Matsuoka and K. Takagi,
Biophys. Res. Commun., 2017, 482, 1207–1212. Biomacromolecules, 2011, 12, 1449–1459.
35 T. F. Gesteira, V. J. Coulson-Thomas, A. Taunay-Rodrigues, 58 A. Tatibouet, A. C. Simao and P. Rollin, Lett. Org. Chem.,
V. Oliveira, B. E. Thacker, M. A. Juliano, R. Pasqualini, 2005, 2, 47–50.
W. Arap, I. L. Tersariol, H. B. Nader, J. D. Esko and 59 M. A. Potopnyk and S. Jarosz, Eur. J. Org. Chem., 2013, 23,
M. A. Pinhal, J. Biol. Chem., 2011, 286, 5338–5346. 5117–5126.
36 S. T. Cheung, M. S. Miller, R. Pacoma, J. Roland, J. Liu, 60 C. Neff, F. Bellot, J.-B. Waern, F. Lambert, J. Brandel,
A. M. Schumacher and L. C. Hsieh-Wilson, ACS Chem.
Biol., 2017, 12, 3126–3133.
G. Serratrice, F. Gaboriau and C. Policar, J. Inorg. Biochem.,
2012, 112, 59–67.
37 Y. Otake, A. L. Nolan, U. K. Walle and T. Walle, J. Steroid 61 J. Ju, D. H. Kim, L. Bi, Q. Meng, X. Bai, Z. Li, X. Li,
Biochem. Mol. Biol., 2000, 73, 265–270.
M. S. Marma, S. Shi, J. Wu, J. R. Edwards, A. Romu and
N. J. Turro, Proc. Natl. Acad. Sci. U. S. A., 2006, 103, 19635–
19640.
38 V. S. Parker, E. J. Squirewell, H. J. Lehmler, L. W. Robertson
and M. W. Duffel, Environ. Toxicol. Pharmacol., 2018, 58,
196–201.
39 J. I. Armstrong, D. E. Verdugo and C. R. Bertozzi, J. Org.
Chem., 2003, 68, 170–173.
40 J. I. Armstrong, X. Ge, D. E. Verdugo, K. A. Winans, J. A. Leary
and C. R. Bertozzi, Org. Lett., 2001, 3, 2657–2660.
41 L. W. Woo, D. Ganeshapillai, M. P. Thomas, O. B. Sutcliffe,
B. Malini, M. F. Mahon, A. Purohit and B. V. L. Potter,
ChemMedChem, 2011, 6, 2019–2034.
62 P. Sundararaman, E. C. Walker and C. Djerassii, Tetrahedron
Lett., 1978, 19, 1627–1628.
63 J. A. Marshall, A. W. Garofalo and R. C. Sedrani, Synlett,
1992, 8, 643–645.
64 N. Valiaeva, D. L. Wyles, R. T. Schooley, J. B. Hwu,
J. R. Beadle, M. N. Prichard and K. Y. Hostetler, Bioorg.
Med. Chem., 2011, 19, 4616–4625.
65 S. K. Jana, P. Leonard, S. A. Ingale and F. Seela, Org. Biomol.
Chem., 2016, 14, 4927–4942.
42 L. J. Wilson, A. Linley, D. E. Hammond, F. E. Hood,
J. M. Coulson, D. J. MacEwan, S. J. Ross, J. R. Slupsky, 66 V. Gopalakrishnan, K. Vaijayanti and K. N. Ganesh,
P. D. Smith, P. A. Eyers and I. A. Prior, Cancer Res., 2018,
78, 15–29.
43 L. H. Jones, Cell Chem. Biol., 2018, 25, 30–35.
44 A. C. Borisa and H. G. Bhatt, Eur. J. Med. Chem., 2017, 140, 1–
19.
Nucleosides Nucleotides, 1992, 11, 1263–1273.
67 W. R. Birmingham, C. A. Starbird, T. D. Panosian,
D. P. Nannemann, T. M. Iverson and B. O. Bachmann, Nat.
Chem. Biol., 2014, 10, 392–399.
¨
¨
68 G. Dufner, R. Schworer, B. Muller and R. R. Schmidt, Eur. J.
45 A. Tang, K. Gao, L. Chu, R. Zhang, J. Yang and J. Zheng,
Oncotarget, 2017, 8, 23937–23954.
46 T. Hunter, Curr. Opin. Cell Biol., 2009, 21, 140–146.
47 P. Traxler, Expert Opin. Ther. Targets, 2003, 7, 215–234.
Org. Chem., 2000, 20, 1467–1482.
69 F. Xue and C. T. Seto, Org. Lett., 2010, 12, 1936–1939.
70 M. Mattarella, L. Berstis, K. K. Baldridge and J. S. Siegel,
Bioconjugate Chem., 2014, 25, 115–128.
48 D. P. Byrne, Y. Li, P. Ngamlert, K. Ramakrishnan, C. E. Eyers, 71 A. S. Rowan, N. I. Nicely, N. Cochrane, W. A. Wlassoff,
C. Wells, D. H. Drewry, W. J. Zuercher, N. G. Berry,
D. G. Fernig and P. A. Eyers, Biochem. J., 2018, 475, 2417–
2433.
A. Claiborne and C. J. Hamilton, Org. Biomol. Chem., 2009,
7, 4029–4036.
72 T. E. McAllister and M. E. Webb, Org. Biomol. Chem., 2012,
10, 4043–4049.
49 D. P. Byrne, Y. Li, P. Ngamlert, K. Ramakrishnan, C. E. Eyers,
C. Wells, D. H. Drewry, W. J. Zuercher, N. G. Berry, 73 M. W. Powner and J. D. Sutherland, ChemBioChem, 2008, 9,
D. G. Fernig and P. A. Eyers, Biochem. J., 2018, 475, 2435–2455. 2386–2387.
50 S. Tanaka, T. Nishiyori, H. Kojo, R. Otsubo, M. Tsuruta, 74 P. R. J. Gaffney and C. B. Reese, J. Chem. Soc., Perkin Trans. 1,
K. Kurogi, M. C. Liu, M. Suiko, Y. Sakakibara and
Y. Kakuta, Sci. Rep., 2017, 7, 8776.
51 C. Liu, J. Sheng, J. M. Krahn, L. Perera, Y. Xu, P. H. Hsieh,
2001, 192–205.
75 A. H. Miah, R. C. Copley, D. O'Flynn, J. M. Percy and
P. A. Procopiou, Org. Biomol. Chem., 2014, 12, 1779–1792.
W. Dou, J. Liu and L. C. Pedersen, J. Biol. Chem., 2014, 289, 76 T. S. Elliott, A. Slowey, Y. Yeb and S. J. Conway,
13407–13418. MedChemComm, 2012, 3, 735–751.
52 G. M. Blackburn, M. J. Gait, D. Loakes and D. M. William, 77 D. P. Byrne, M. Vonderach, S. Ferries, P. J. Brownridge,
Nucleic Acids in Chemistry and Biology, Royal Society of
C. E. Eyers and P. A. Eyers, Biochem. J., 2016, 473, 3159–3175.
Chemistry, Cambridge, 2006.
This journal is © The Royal Society of Chemistry 2019
RSC Adv., 2019, 9, 32165–32173 | 32173