A library approach to the generation of bisubstrate analogue sulfotransferase inhibitors

Article abstract of DOI:10.1021/ol0162217

(Equation presented) A library of potential bisubstrate analogue inhibitors (1) targeting sulfotransferase enzymes was generated by the chemoselective ligation of the PAPS mimic 2 with a panel of 447 aldehydes. Preliminary screening has identified compounds that inhibit estrogen sulfotransferase (EST), an enzyme relevant to breast cancer.

Full text of DOI:10.1021/ol0162217

ORGANIC
LETTERS
2001
Vol. 3, No. 17
2657-2660
A Library Approach to the Generation of
Bisubstrate Analogue Sulfotransferase
Inhibitors
Joshua I. Armstrong,† Xue Ge,† Dawn E. Verdugo,† Katharine A. Winans,†
,†,‡,§
Julie A. Leary,† and Carolyn R. Bertozzi*
Center for New Directions in Organic Synthesis, Howard Hughes Medical Institute,
and Departments of Chemistry and Molecular and Cell Biology, UniVersity of
California, Berkeley, California 94720
bertozzi@cchem.berkeley.edu
Received June 1, 2001
ABSTRACT
A library of potential bisubstrate analogue inhibitors (1) targeting sulfotransferase enzymes was generated by the chemoselective ligation of
the PAPS mimic 2 with a panel of 447 aldehydes. Preliminary screening has identified compounds that inhibit estrogen sulfotransferase (EST),
an enzyme relevant to breast cancer.
Sulfated biomolecules mediate a wide array of normal and
pathological cellular events. For example, sulfated carbohy-
drates arbitrate leukocyte adhesion to inflamed endothelium,1
sulfated proteins modulate HIV-1 infectivity,2 and sulfation
also regulates the activity of steroid hormones in vivo.3 The
role of sulfated biomolecules in numerous diseases has
prompted the search for inhibitors targeting the enzymes that
install sulfate esters, the sulfotransferases.
small molecule6 acceptor (Figure 1). Over 25 human
carbohydrate and protein sulfotransferases have been identi-
fied, including many homologous isozymes with similar
substrate specificities and overlapping tissue distributions.
The strong connection between sulfotransferases and several
disease states has prompted interest in the discovery of potent
and selective sulfotransferase inhibitors both as tools to
elucidate the role of these enzymes in vivo and as potential
therapeutics.7
We have been exploring the inhibitory potential of PAPS-
competitive compounds derived from substances traditionally
used to target the ATP-binding pocket of protein kinases.8
Although we have found drug-like PAPS-competitive com-
Sulfotransferases catalyze the transfer of a sulfuryl group
from 3′-phosphoadenosine 5′-phosphosulfate (PAPS) to a
hydroxyl (or amino) group on a carbohydrate,4 protein,5 or
† Department of Chemistry, University of California.
‡ Center for New Directions in Organic Synthesis, Howard Hughes
Medical Institute. The Center for New Directions in Organic Synthesis is
supported by Bristol-Myers Squibb as sponsoring member.
§ Department of Molecular and Cell Biology, University of California.
(1) Rosen, S. D. Am. J. Pathol. 1999, 155, 1013.
(6) Falany, C. N. FASEB J. 1997, 11, 206.
(7) Armstrong, J. I.; Bertozzi, C. R. Curr. Opin. Drug Disc. DeV. 2000,
3, 502.
(2) Farzan, M.; Mirzabekov, T.; Kolchinsky, P.; Wyatt, R.; Cayabyab,
M.; Gerard, N. P.; Gerard, C.; Sodroski, J.; Choe, H. Cell 1999, 96, 667.
(3) Kotov, A.; Falany, J. L.; Wang, J.; Falany, C. N. J. Steroid Biochem.
Mol. Biol. 1999, 68, 137.
(4) Bowman, K. G.; Bertozzi, C. R. Chem. Biol. 1999, 6, R9.
(5) Kehoe, J. W.; Bertozzi, C. R. Chem. Biol. 2000, 7, R57.
(8) (a) Armstrong, J. I.; Portley, A. R.; Chang, Y.-T.; Nierengarten, D.
M.; Cook, B. N.; Bowman, K. G.; Bishop, A.; Gray, N. S.; Shokat, K. M.;
Schultz, P. G.; Bertozzi, C. R. Angew. Chem., Int. Ed. 2000, 39, 1303. (b)
Verdugo, D. E.; Cancilla, M. T.; Ge, X.; Gray, N. S.; Chang, Y.-T.; Schultz,
P. G.; Negishi, M.; Leary, J. A.; Bertozzi, C. R. J. Med. Chem. 2001, in
press.
10.1021/ol0162217 CCC: $20.00 © 2001 American Chemical Society
Published on Web 07/25/2001

Figure 1. Sulfotransferases catalyze sulfuryl-group transfer from
3′-phosphoadenosine 5′-phosphosulfate (PAPS) to biomolecules.
pounds with moderate to high potency in vitro,8 we anticipate
difficulty in achieving selective inhibition of related sulfo-
transferase enzymes in vivo since the family of sulfotrans-
ferases shares a similar PAPS-binding pocket. To surmount
this problem, we focus here on the preparation of bisubstrate
analogues.
Bisubstrate analogues can provide specificity via key
interactions within both binding pockets of the enzyme.9 This
is a clear advantage in the case of sulfotransferases that share
high sequence similarity in the PAPS-binding region but
possess disparity in other regions including the acceptor-
binding site (Figure 2a).10 In addition to specificity, potency
is achieved from the entropic advantage of linking structures
that mimic each substrate (Figure 2b). Many of the traditional
design strategies for bisubstrate analogues that target group
transfer enzymes, such as kinases11 and glycosyltrans-
ferases,12 result in potent inhibitors in vitro. However, the
compounds are usually too hydrophilic or charged to cross
cell membranes and are therefore not useful for cellular
studies.
Figure 2. (a) Schematic of a sulfotransferase active site containing
both the conserved PAPS-binding pocket and an acceptor-binding
pocket with variable amino acid residues. (b) Generalized structure
of a sulfotransferase bisubstrate analogue inhibitor. (c) The first
phase of library design utilizes an aminooxy-functionalized PAPS
mimic to identify pharmacophores that target the acceptor-binding
pocket. (d) An aldehyde-derived “hit” from the first phase will be
used as a scaffold to identify PAPS substitutes.
inhibitors via the chemoselective coupling of aminooxy
nucleoside 2 with commercially available aldehydes (Scheme
1). The oxime-forming reaction tolerates numerous functional
groups, an attribute that has been exploited for the chemo-
selective ligation of pharmacophores in a library synthesis,13
attachment of sugars to peptides,14 and the modification of
cells.15 Once a hit is identified, the library strategy can be
reversed by using the aldehyde-derived acceptor-mimic as
a scaffold; derivatizing this compound with an aminooxy
linker would allow us to probe for more bioavailable
inhibitors targeting the PAPS-binding site (Figure 2d).
Ultimately, the two-step strategy may yield compounds with
Our strategy focused on the design of a library possessing
two elements: (1) a PAPS mimic that directs the compounds
to sulfotransferases and (2) a variable component that is
hydrophobic and drug-like to fill the acceptor-binding pocket
(Figure 2c). The target library, structure 1 (Scheme 1),
Scheme 1
(9) Radzicka, A.; Wolfenden, R. Methods Enzymol. 1995, 249, 284.
(10) Kakuta, Y.; Pedersen, L. G.; Pedersen, L. C.; Negishi, M. Trends
Biochem. Sci. 1998, 23, 129.
(11) (a) Traxler, P. M.; Wacker, O.; Bach, H. L.; Geissler, J. F.; Kump,
W.; Meyer, T.; Regenass, U.; Roesel, J. L.; Lydon, N. J. Med. Chem. 1991,
34, 2328. (b) Parang, K.; Till, J. H.; Ablooglu, A. J.; Kohanski, R. A.;
Hubbard, S. R.; Cole, P. A. Nat. Struct. Biol. 2001, V8, 37. (c) Yan, X.;
Lawrence, D. S.; Corbin, J. D.; Francis, S. H. J. Am. Chem. Soc. 1996,
118, 11684.
(12) Wong, C. H. Acc. Chem. Res. 1999, 32, 376.
(13) Maly, D. J.; Choong, I. C.; Ellman, J. A. Proc. Natl. Acad. Sci.
U.S.A. 2000, 97, 2419.
(14) Marcaurelle, L. A.; Bertozzi, C. R. Chem. Eur. J. 1999, 5, 1384.
(15) Lemieux, G. A.; Bertozzi, C. R. TIBTECH 1998, 16, 506.
contains an oxime linkage between the elements, allowing
the parallel synthesis of hundreds of potential bisubstrate
2658
Org. Lett., Vol. 3, No. 17, 2001

linked pharmacophores that simultaneously bind in both
substrate-binding pockets.
Library synthesis was performed in a 96-well plate format
using 25 mM 2 (DMSO containing 0.2% AcOH) and 20 mM
aldehyde (DMSO, 447 aldehydes total). The reactions were
incubated at room temperature in the dark for 48 h before
they were frozen for later use. Approximately 10% of the
library was analyzed by reversed-phase HPLC, and the major
peak was isolated and characterized by electrospray ioniza-
tion mass spectrometry. The reaction yields were typically
high (>90%) as determined by a reduction in peak intensity
for 2. However, more than one product peak was observed
in approximately 10% of the 40 reactions analyzed, preclud-
ing the determination of an accurate yield via this method.
The library was then screened for inhibitors of the cytosolic
enzyme, estrogen sulfotransferase (EST), using the previously
developed immobilized enzyme mass spectrometry (IEMS)
assay.21 Unusually high levels of sulfated estrogen are found
in breast cancer cells; thus EST inhibitors are of potential
therapeutic value.22 Briefly, EST was immobilized on agarose
beads via reductive amination and a mixture of the library
components was analyzed by mass spectrometry both before
and after exposure to the immobilized enzyme. A reduction
in peak intensity for a library component indicated binding
to the enzyme. Figure 3 shows the mass spectrometry data,
As the first step toward this goal, we have synthesized 2
and used it to prepare library 1 consisting of 447 unique
compounds. We chose the 2′-deoxyadenosine scaffold after
evaluating the X-ray crystallographic data for both EST16
and the sulfotransferase domain of heparin N-deacetylase/
N-sulfotransferase17 and finding no key interactions between
the enzyme active site residues and the 2′-hydroxyl group
of PAPS. In addition, the 2′-deoxy analogues are more
hydrophobic and more synthetically tractable, requiring fewer
protecting group manipulations to install the 5′-aminooxy
linker.
The new PAPS analogue 2 was prepared from the
azidonucleoside 318 (Scheme 2) by a Pd-mediated reduction
Scheme 2a
a Reagents and conditions: (a) 10% Pd/C, AcOH, MeOH (62%);
(b) FmocNHOCH2COOH (4), EEDQ, DMF (63%); (c) (i) (i-
Pr)2NP(ONPE)2, tetrazole, CH2Cl2; (ii) mCPBA (71%); (d) (i)
nPrNH2, CH3CN, 1 h; (ii) BSA, DBU (57%).
Figure 3. Results from the IEMS screen (performed in duplicate)
of the library (subset of 67 members) against EST. Compounds
displaying greater than a 55% decrease in ion abundance after
enzyme incubation were considered “hits.” Error bars indicate the
result of each replicate.
of the azido group to the amine followed by coupling with
2-[(N-fluorenemethyloxycarbonyl)aminooxy]acetic acid (4)19
to provide 5. Standard phosphoramidite chemistry was then
employed using di(p-nitrophenethyl (NPE)) diisopropylphos-
phoramidite20 followed by oxidation with mCPBA to provide
the phosphorylated nucleoside 6. Complete deprotection was
achieved via the stepwise addition of excess n-propylamine
to a solution of 6 (100 mM, CH3CN) followed by treatment
with bis(trimethylsilyl)acetamide (BSA) and 1,8-diazabicyclo-
[5.4.0]undec-7-ene (DBU). Purification of 1 via reversed-
phase HPLC eluting with a gradient of 0 to 20% CH3CN in
25 mM NH4OAc provided the ammonium salt of 2 in 57%
yield.
expressed as percent (%) decrease in peak intensity for 67
compounds. We have previously demonstrated a strong
correlation between a reduction in peak intensity and
inhibitory activity,8b and we were interested in establishing
this correlation for library 1. We therefore screened the 67
compounds for EST inhibitory activity using a previously
developed radiolabel-transfer TLC assay.8b We found that
all compounds displaying greater than 50% inhibition at 200
µM also shared greater than 55% ion abundance decrease
in the IEMS assay (Figure 4). Although several of the “hit”
(16) Kakuta, Y.; Pedersen, L. G.; Carter, C. W.; Negishi, M.; Pedersen,
L. C. Nat. Struct. Biol. 1997, 4, 904.
(17) Kakuta, Y.; Sueyoshi, T.; Negishi, M.; Pedersen, L. C. J. Biol. Chem.
1999, 274, 10673.
(18) Mag, M.; Engels, J. W. Nucleosides Nucleotides 1988, 7, 725.
(19) Trevisiol, E.; Defrancq, E.; Lhomme, J.; Laayoun, A.; Cros, P.
Tetrahedron 2000, 56, 6501.
(21) Cancilla, M. T.; Leavell, M. D.; Chow, J.; Leary, J. A. Proc. Natl.
Acad. Sci. U.S.A. 2000, 97, 12008.
(22) Qian, Y.; Deng, C.; Song, W. C. J. Pharm. Exp. Ther. 1998, 286,
555.
(20) Uhlmann, E.; Engels, J. W. Tetrahedron Lett. 1986, 27, 1023.
Org. Lett., Vol. 3, No. 17, 2001
2659

Figure 4. A comparison of the IEMS binding data (shaded bars) and radiolabel-transfer inhibition data (dark bars) reveals several EST
inhibitors. Compounds 8, 12, 18, and 20 were confirmed as hits. False positives (“hits” from the IEMS assay, but not from the radiolabel-
transfer assay) are expected since compounds may bind to the enzyme without inhibiting activity. Examples are compounds 7, 11, 13, 14,
and 16. Each assay was performed in duplicate; the error bars represent the result of each replicate. The R groups correspond to substituents
in compound 1.
compounds in the IEMS assay lacked inhibitory activity
(compounds 7, 11, 13, 14, 16), it is noteworthy that the IEMS
assay successfully identified all of the inhibitors revealed
by the radiolabel-transfer TLC assay. Specifically, com-
pounds 8, 18, and 20 demonstrated significant inhibitory
activity (>80% inhibition) against EST. Surprisingly, these
compounds vary widely in aldehyde structure. This implies
that the aldehyde structures bind to disparate regions of the
active site or that the structures contain similar binding motifs
that would have been difficult to rationally predict. Our
strategy therefore reveals a “fingerprint” of pharmacophores
that target the acceptor-binding pocket of EST. As we expand
our efforts and screen 1 against a panel of sulfotransferases,
we expect to generate “pharmacophore fingerprints” for each
enzyme, allowing us to identify unique structures that bind
with specificity to each acceptor-binding pocket.
binding pocket. This approach will hopefully yield cell
permeable, specific, and potent sulfotransferase inhibitors for
use in future in vivo studies.
Acknowledgment. We are grateful to Jonathan Ellman
and his colleagues for numerous helpful discussions regard-
ing library design and synthesis. We acknowledge generous
funding from the Camille and Henry Dreyfus Foundation
and Dupont. This research was supported by grants to C.R.B.
from the American Cancer Society (RPG9700501BE) and
the National Institutes of Health (GM59907-01). J.A.L.
acknowledges the National Institute of Health (GM47356)
for financial support.
Supporting Information Available: Experimental pro-
cedures and spectral data for all new compounds and the
structures of all aldehydes used to generate library 1. This
material is available free of charge via the Internet at
http://pubs.acs.org.
We are currently screening library 1 against a panel of
sulfotransferases to identify inhibitors for each enzyme and
determine the selectivity of the “hits”. With the most potent
inhibitors in hand, the strategy will be reversed, using the
aldehyde-derived moiety as an anchor to probe the PAPS-
OL0162217
2660
Org. Lett., Vol. 3, No. 17, 2001