The difluoromethylene group as a replacement for the labile oxygen in steroid sulfates: A new approach to steroid sulfatase inhibitors

Article abstract of DOI:10.1016/j.bmcl.2003.09.089

Several estrone sulfate and estradiol sulfate analogues, in which the sulfate group was replaced with an α,α-difluoromethylenesulfonate group or an α,α-difluoromethylenetetrazole group, were examined as inhibitors of steroid sulfatase (STS). These compounds were 4.5-10.5 times more potent than their non-fluorinated analogues. Moreover, the presence of the fluorines changed the mode of inhibition from mixed to competitive. The inhibitor bearing the α,α-difluoromethylenetetrazole group exhibited an affinity for STS approaching that of the natural STS substrate, estrone sulfate. Possible reasons for the enhanced affinity of the fluorinated compounds compared to their non-fluorinated counterparts are discussed.

Full text of DOI:10.1016/j.bmcl.2003.09.089

Bioorganic & Medicinal Chemistry Letters 14 (2004) 151–155
The difluoromethylene group as a replacement for the labile
oxygen in steroid sulfates: a new approach to steroid
sulfatase inhibitors
Jennifer Lapierre, Vanessa Ahmed, Mei-Jin Chen, Mehdi Ispahany, J. Guy Guillemette
and Scott D. Taylor*
Department of Chemistry, University of Waterloo, Waterloo, Ontario, Canada N2L 3G1
Received 19 August 2003; accepted 29 September 2003
Abstract—Several estrone sulfate and estradiol sulfate analogues, in which the sulfate group was replaced with an a,a-difluor-
omethylenesulfonate group or an a,a-difluoromethylenetetrazole group, were examined as inhibitors of steroid sulfatase (STS).
These compounds were 4.5–10.5 times more potent than their non-fluorinated analogues. Moreover, the presence of the fluorines
changed the mode of inhibition from mixed to competitive. The inhibitor bearing the a,a-difluoromethylenetetrazole group exhib-
ited an affinity for STS approaching that of the natural STS substrate, estrone sulfate. Possible reasons for the enhanced affinity of
the fluorinated compounds compared to their non-fluorinated counterparts are discussed.
# 2003 Elsevier Ltd. All rights reserved.
Approximately one-third of breast tumors in post-
menopausal women require stimulation by estrogens for
optimal growth.1 Consequently, certain enzymes
responsible for the biosynthetic production of such
estrogens are being evaluated as targets for the treat-
ment of hormonally-dependent breast cancer. Until
fairly recently, two such enzymes, aromatase and 17b-
hydroxysteroid dehydrogenase, have been the main
focus of inhibitor/drug development. However, more
recently it has been proposed that another enzyme,
steroid sulfatase [STS, also known as estrone sulfatase
or arylsulfatase C (ASC)], may also be a potential target
for therapeutic intervention.2 STS catalyzes the hydro-
lytic desulfation of steroidal sulfates, such as estrone
sulfate (E1-S) or dehydroepiandrosterone sulfate
(DHEAS), to the corresponding steroids and inorganic
sulfate. Studies have suggested that these sulfated ster-
oids, by the reaction of STS, function as reservoirs for
the formation of estrogens in breast tumors.2 Over the
last several years, a number of reports have appeared
describing inhibitors of STS.3 In one report, Li and
co-workers examined a number of sulfonate analogues
as STS inhibitors.4a Among these compounds were sul-
fonates 1 and 2, in which the labile sulfate ester oxygen
of E1-S and estradiol sulfate (E2-S) was replaced with a
non-hydrolyzable methylene group. However, these
compounds were found to be poor inhibitors exhibiting
mixed inhibition with Ki values of 140 and 130 mM,
respectively, and have an affinity for STS that is con-
siderably poorer than that of substrates ES-1 and ES-2.
On the basis of these studies and on studies with other
sulfonate analogues,4b it was concluded that an oxygen
atom or an electronically similar link between the aryl
moiety and the sulfur atom is essential for high affinity
binding.4a
The difluoromethylenephosphonic acid group (DFMP),
R-CF2PO3À2, has been used extensively in the design of
inhibitors of enzymes that hydrolyze or bind phosphate
À2
esters (R-O-PO3 ).5À2In some instances, compounds
bearing the -CF2PO3 were dramatically better inhibi-
tors of phosphatases than theÀa2nalogous non-fluori-
nated analogues (R-CH2PO3 ).5f The increased
potency of the DFMP’s compared to their non-fluori-
nated analogues has been attributed to the lower pKa
values of the fluorinated phosphonic acids and/or H-
bonding interactions between the fluorines and residues
in the active sites.5aÀf In the case of STS, the low affinity
of 1 and 2, compared to ES-1 and ES-2, is probably not
a result of differences in pKa values of the sulfate and
*Corresponding author. Tel.: +1-519-888-4567x3325; fax: +1-519-
746-0435; e-mail: s5taylor@sciborg.uwaterloo.ca
0960-894X/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2003.09.089

152
J. Lapierre et al. /Bioorg. Med. Chem. Lett. 14 (2004) 151–155
sulfonate moieties since both should be completely
ionized at the pH under which the studies were per-
formed. The poor affinity of the sulfonates is probably a
result of loss of an interaction between the labile ester
oxygen of the sulfates and STS. Therefore, we decided to
examine the difluoromethylene moiety as a replacement
for the labile oxygen in anticipation that the fluorines
would partake in specific interactions with residues in the
active site, in a manner similar to that found with
difluoromethylenephosphonic acid inhibitors of certain
protein tyrosine phosphatases.5f To determine this, we
prepared two E1-S analogues, the difluoromethylene-
sulfonic acid 3 and the difluoromethylenetetrazole 4.6
Tetrazole 4 was prepared because it has been reported
that the tetrazole group could serve the role of the sul-
fate moiety at the CCK receptor in vitro in certain ana-
logues of CCK-8 sulfotyrosine-bearing peptides.7
Therefore, it was hypothesized that the tetrazole group
might also be a good sulfate surrogate in steroidal sul-
fates. Here we report the results of inhibition studies
with compounds 3 and 4, as well as with compound 5,
which is the 17-hydroxy analogue of 3, and with the
non-fluorinated compounds 1 and 6, with highly
purified STS.
lined in Scheme 1. The starting material, compound 7,
was described previously.6 The key step was the elec-
trophilic fluorination of sulfonate 9 using N-fluoro-
benzenesulfonimide.8 Compound 6 was prepared as
outlined in Scheme 2. The starting material, 11, was
prepared as previously described.6 Conversion of nitrile
12 to tetrazole 6 turned out to be somewhat challenging.
Standard conditions for converting nitriles to tetrazoles,
such as NaN3/NH4Cl/Á9 or Me3Si–N3/(Bu)2SnO/Á,10
resulted in little or no conversion. However, it was
found that by using the procedure of Bernstein and
Vacek,11 which involved heating a mixture of 12, NaN3
and Et3NHCl in N-methylpyrrolidinone at 150 ꢀC,
tetrazole 6 could be obtained in reasonable yield.
Human placental STS was purified to apparent homo-
geneity (one band on an SDS gel using silver staining)
using the procedure of Hernandez-Guzman et al.12
Inhibition studies were performed as follows. To the
wells of a 96-well plate was added an appropriate
amount of a stock solution of 1 or 3–6 in DMSO to
assay buffer which consisted of 0.33 M Tris–HCl, 0.1%
Triton, pH 7.4. To this was added an appropriate
amount of a stock solution of 4-methylumbelliferyl sul-
fate (4-MUS) in 0.33 M Tris–HCl, pH 7.4. These mix-
tures were incubated at 37 ꢀC for 2 min. The reaction
was initiated by adding 10 mL of purified STS in 20 mM
Tris–HCl, 0.1% Triton, pH 7.4. The final concentration
of STS was 0.005 mg/mL. The total volume was 100 mL,
the final concentration of DMSO was 10% and the final
concentration of Triton was 0.08%. The production of
4-methylumbelliferone was monitored over 10–15 min
using a Spectramax GeminiXS plate reader (excitation
359 nm, emission, 451 nm) at 37 ꢀC. Each reaction was
performed in triplicate. The concentration range over
which the inhibitors were tested ranged from approxi-
mately 0.5- to 3-fold Ki. Controls were performed in an
identical manner but did not contain STS. Initial rates
(v) were determined by taking the slopes of plots of the
change in relative fluorescence units with time. These
data were plotted as Lineweaver–Burk graphs and Ki
and aKi values were calculated from replots of the
slopes or intercepts of the Lineweaver–Burk graphs
according to the equations for mixed and competitive
inhibition.13
Compounds 3 and 4 were prepared as previously
reported.6 Compound 5, which bears a hydroxy group
at the 17-position in the D-ring, was prepared as out-
Li et al. have reported that compound 1 exhibits mixed
inhibition with purified STS and has a Ki value of
140 mM.4a,14 Similar results were reported for com-
Scheme 1.
Scheme 2.

J. Lapierre et al. /Bioorg. Med. Chem. Lett. 14 (2004) 151–155
153
pound 2. We also prepared compound 1 using the
methodology of Li et al. and examined it as an STS
inhibitor. Under our assay conditions, compound 1 also
exhibited mixed inhibition with a Ki of 600Æ74 mM and
an aKi of 1378Æ75 mM. The difference between our
result and that of Li et al. is most probably due to the
differences in assay conditions. STS activity in Li et al.’s
studies was determined radiometrically using E1-35S as
substrate at pH 7.0 in 100 mM Tris–acetate in the
absence of Triton and DMSO while our studies were
conducted using 4-MUS as substrate in 333 mM Tris–
HCl buffer at pH 7.4 with 0.08% Triton and 10%
DMSO. Interestingly, fluorosulfonate 3 did not exhibit
mixed inhibition but instead displayed only competitive
inhibition with a Ki of 57Æ6 mM (Fig. 1). Compound 5
was also a competitive inhibitor with a slightly lower Ki
value (37Æ5 mM). Tetrazole inhibitor 6 displayed mixed
inhibition with a Ki of 72.7Æ1.4 mM and an aKi of
213Æ38 mM. However, its fluoro analogue, compound
4, exhibited only competitive inhibition with a Ki of
16Æ3 mM (Fig. 2).
inhibit STS desulfation of 4-MUS with a Ki of 60 mM.15
Thus, it appears that compound 4 exhibits an affinity
for STS approaching that of the natural substrate. To
our knowledge, the only other report describing the use
of the tetrazole group as a sulfate mimetic was by Tilley
et al. who used the tetrazole group to mimic the sulfate
group in sulfotyrosine-bearing CCK-8 peptide analo-
gues.7 None of the tetrazole-bearing peptides exhibited
an affinity for the CCK receptor equal to that of the
analogous sulfated peptides.7 The superior affinity of
the tetrazole derivatives over the analogous sulfonates
in the present study is to our knowledge unprecedented
and is surprising when one considers that the two
groups are, from a structural point of view, quite dif-
ferent. Modeling studies will be necessary to determine
how the tetrazole group is interacting with STS.16
The second notable result is that replacing the bridging
methylene unit in 1 and 6 with a difluoromethylene
group results in a change in the mode of inhibition from
mixed inhibition to solely competitive inhibition. Li has
reported that estrone sulfate (ES-1) exhibits mixed-type
inhibition when using DHEA-35S as a substrate.4a Li
also noted that subtle changes in the sulfate group in
ES-1 can yield STS inhibitors that display inhibition
patterns that are different from ES-1.4a For example,
estrone phosphate and estrone-3-methylsulfonate are
purely competitive inhibitors.4a The CF2 group may
nullify the non-competitive binding mode as well as
increase the competitive binding mode such that the
non-competitive binding is relatively insignificant (or
both).
Several aspects of the above results are notable. First,
tetrazoles 4 and 6 are 3.5–8.3 times more potent than
their sulfonate analogues, 1 and 3, respectively. At pH
7.5 in Tris buffer and in the presence of Triton, E1-S has
been reported to have a Km of 21 mM.15 Under the same
conditions, E1-S has also been reported to competitively
The third, and perhaps the most significant result, is
that replacing the bridging methylene unit (in 1 and 6)
with a difluoromethylene group results in a 4.5–10.5-
fold increase in inhibitor potency. Since sulfonates 1
and 3, and tetrazoles 4 and 6, are ionized at physio-
logical pH,17 it stands to reason that enhanced affinity
of the fluorinated compounds 3 and 4 is probably not
due to the increased acidity of these compounds. It is
more likely that their enhanced affinity is due to inter-
actions of the fluorines with residues in the active site.
One possibility is that the fluorines are hydrogen bond-
ing with residues in the active site. Although fluorine is
capable of acting as an H-bond acceptor, the subject of
fluorine hydrogen bonds involving C–F is a subject
of much controversy.18aÀc Nevertheless, it appears that
such H-bonds can form in certain instances19 although
the optimal strength of such bonds is still unknown.
What residue(s) in STS might be involved in such an H-
bond with compounds 3 and 4? The mechanism of STS
has not yet been studied in detail. Most of the work that
has been done on the mechanism of aryl sulfatases
(AS’s) has been performed on arylsulfatase A (ASA).20
All AS’s known to date have a formylglycine residue in
the active site resulting from a post-translational modi-
fication of a serine or cysteine residue. The crystal
structure of human ASA shows that the formylglycine
exists as a hydrate which is stabilized by coordination to
a Mg+2 ion and H-bonds to Asp-281 and His-125. On
the basis of the STS crystal structure and other studies,
a mechanism has been proposed for ASA (Scheme
Figure 1. Inhibition of STS by sulfonate 3. Assays were conducted as
described in the text in the presence of 0 mM (&), 30 mM (x), 75 mM
(~), and 150 mM (^) compound 3.
Figure 2. Inhibition of STS by tetrazole 4. Assays were conducted as
described in the text in the presence of 0 mM (&), 10 mM (~), 25 mM
(x), and 50 mM (*) compound 4.

154
J. Lapierre et al. /Bioorg. Med. Chem. Lett. 14 (2004) 151–155
proteins that bind or hydrolyze phosphate esters for
many years, the analogous approach to sulfatase inhi-
bitor design has never been pursued. To our knowledge,
this is the first demonstration of the use of fluor-
omethylenesulfonic acids as enzyme inhibitors. In this
case, the fluorinated compounds were considerably bet-
ter inhibitors than their non-fluorinated analogues and
changed the mode of inhibition from mixed to compe-
titive, demonstrating the utility of this approach to sul-
fatase inhibitor design. The use of a tetrazole to mimic
the sulfate group is also worthy of note since the tetra-
zole-bearing compounds were superior inhibitors to the
sulfonate analogues. We expect that the approaches to
sulfatase inhibitor design outlined here will also be use-
ful for preparing inhibitors and probes of other proteins
that bind or hydrolyze sulfate esters.
Acknowledgements
Scheme 3.
We thank the Natural Sciences and Engineering
Research Council (NSERC) of Canada for supporting
this work through operating grants to S.D.T. and G.G.
and a postgraduate scholarship to J.L. We also thank
Debashis Ghosh for helpful discussions.
3).20,21 One of the hydroxyls of the formylglycine
hydrate attacks the sulfur atom of the substrate. It has
been hypothesized that the ester oxygen of the substrate
is involved in H-bonding with Lys-302 and His-229.
His-229 is a key residue, since it is believed to assist in
the cleavage of the S–O bond of the substrate by acting
as a general acid. The sulfate group is then eliminated
from the hydrate, resulting in formation of formyl gly-
cine which is then rehydrated.
References and notes
1. Henderson, I. C.; Cannelos, G. P. N. Eng. J. Med. 1980,
320, 17.
2. Pasqualini, J. R.; Chetrite, G. S. In Breast Cancer: Prog-
nosis, Treatment and Prevention; Pasqualini, J. R., Ed.;
Marcel Dekker: New York, 2002; p 187.
3. For a recent review see: Ahmed, S.; Owen, C. P.; James,
K.; Sampson, S.; Patel, C. K. Curr. Med. Chem. 2002, 9,
263.
4. (a) Li, P.-K.; Pillai, R.; Dibbelt, L. Steroids 1995, 60, 299.
(b) Dibbelt, L.; Li, P.-K.; Pillai, R.; Knuppen, R. J. Ster-
oid Biochem. Mol. Biol. 1994, 50, 261.
5. (a) Halazy, S.; Ehrhard, A.; Eggenspiller, A.; Berges-
Gross, V.; Danzin, C. Tetrahedron 1996, 51, 177. (b)
Blackburn, G. M.; Kent, D. E.; Kolkmann, F. J. Chem.
Soc., Perkin Trans. 1 1984, 1119. (c) Hebel, D.; Kirk,
K. L.; Kinjo, J.; Kovacs, T.; Lesiak, K.; Balzarini, J.; De
Clercq, E.; Torrence, P. F. Bioorg. Med. Chem. Lett. 1991,
1, 357. (d) Burke, T. R.; Kole, H. K.; Roller, P. P. Bio-
chem. Biophys. Res. Commun. 1994, 204, 129. (e) Chen, L.;
Wu, L.; Otaka, A.; Smyth, M. S.; Roller, P. R.; Burke,
T. R.; den Hertog, J.; Zhang, Z.-Y. Biochem. Biophys.
Res. Commun. 1995, 216, 976. (f) Burke, T. R., Jr.; Ye, B.;
Yan, X.; Wang, S.; Jia, Z.; Chen, L.; Zhang, Z.-Y.; Bar-
ford, D. Biochemistry 1996, 35, 15989.
In the crystal structure of STS, the formylglycine is
hydrated and the hydrate is sulfated.16 This is similar to
that found in the crystal structure of arylsulfatase B
(ASB).22 Nine out of 10 catalytically important residues
are identical in STS, ASA and ASB. The exception is
Gln-343 in STS as opposed to an Asn in ASA and ASB
and these residues function as ligands to the cation.
Superimposition of these nine a-carbon atom positions
by least squares minimization resulted in root mean
squared deviation of 0.39 A between STS and either
ASA or ASB.16 Thus, the mechanism for STS is most
likely very similar to that proposed for ASA. His-229 in
ASA, the residue that is believed to be involved in H-
bonding to the sulfate ester oxygen, corresponds to His-
290 in STS. Ghosh and co-workers have modeled E1-S
in the STS active site with its sulfate moiety super-
imposed with the crystallographically observed sulfate
of the sulfated formylglycine hydrate. The oxygen of the
labile S–O bond is within H-bonding distance of one of
the imidazole nitrogens of His-290. It has been pro-
posed that His-290 acts as a general acid in a manner
analogous to that proposed for His-229 in ASA.16 Thus,
it is possible that the fluorines in 3 and 4 enhance bind-
ing by H-bonding with His-290. Such H-bonding would
not be without precedent. The enhanced affinity of
difluoromethylenephosphonic acid inhibitors of protein
tyrosine phosphatase 1B, compared to their non-fluori-
nated analogues, has been attributed at least in part to
H-bonding with an N–H backbone proton.5f
6. Chen, M.-J.; Taylor, S. D. Tetrahedron Lett. 1999, 40,
4149.
7. Tilley, J. W.; Danho, W.; Lovey, K.; Wagner, R.; Swis-
tok, J.; Makofske, R.; Michaelewsky, J.; Triscari, J.; Nel-
son, D.; Weatherford, S. J. Med. Chem. 1991, 34, 1125.
8. Kotoris, C.; Chen, M.-J.; Taylor, S. D. J. Org. Chem.
1998, 63, 8052.
9. Finnegan, W. G.; Henry, R. A.; Lofquist, R. J. Am.
Chem. Soc. 1958, 80, 3908.
10. Wittenberg, S. J.; Donner, B. G. J. Org. Chem. 1993, 58,
4139.
11. Bernstein, P. R.; Vacek, E. P. Synthesis 1987, 1133.
Although a,a-difluoromethylenephosphonic acids have
been used as inhibitors and probes of enzymes and

J. Lapierre et al. /Bioorg. Med. Chem. Lett. 14 (2004) 151–155
155
12. Hernandez-Guzman, F. G.; Higashiyama, T.; Osawa, Y.;
Ghosh, D. J. Steroid. Biochem. Mol. Biol. 2001, 78, 441.
13. Segel, I. H. In Enzyme Kinetics; Wiley: New York, 1975;
p 346.
14. Two Ki values are obtained for compounds exhibiting
mixed inhibition. Only the lower Ki values were given in
ref 4a.
15. Daniel, W. L.; Chang, P. L. Enzyme 1990, 43, 212.
16. The structure of STS has recently been determined by X-
ray crystallography. See: Hernandez-Guzman, F. G.;
Higashiyama, T.; Pangborn, W.; Osawa, Y.; Ghosh, D. J.
Biol. Chem. 2003, 278, 22989. The coordinates for STS
were not available to the public at the time of this study.
17. Kotoris, C. C.; Chen, M.-J.; Taylor, S. D. Bioorg. Med.
Chem. Lett. 1998, 8, 3275.
18. (a) O’Hagan, D. O.; Rzepa, H. S. J. Chem. Soc., Chem.
Commun. 1997, 645. (b) Dunitz, J. D.; Taylor, R. Chem.
Eur. J. 1997, 3, 89. (c) Howard, J. A. K.; Hoy, V. J.;
O’Hagan, D. O.; Smith, G. T. Tetrahedron 1996, 52,
12613.
19. Barbarich, T. J.; Rithner, C. D.; Miller, S. M.; Anderson,
O. P.; Strauss, S. H. J. Am. Chem. Soc. 1999, 121, 4280.
20. von Bulow, R.; Schmidt, B.; Dierks, T.; von Figura, K.;
Uson, I. J. Mol. Biol. 2001, 305, 269 and references
therein.
21. For a more detailed description of the mechanism see ref
20.
22. Bond, S. C.; Clements, P. R.; Ashby, S. J.; Collyer, C. A.;
Harrop, S. J.; Hopwood, J. J.; Guss, J. M. Structure 1997,
5, 277.