Lithocholic acid analogues, new and potent α-2,3-sialyltransferase inhibitors

Article abstract of DOI:10.1039/b514915k

A new type of noncompetitive α-2,3-sialyltransferase inhibitor has been synthesized; we report the discovery, preparation and inhibitory activity of sixteen lithocholic acid analogues. The Royal Society of Chemistry 2006.

Full text of DOI:10.1039/b514915k

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Lithocholic acid analogues, new and potent a-2,3-sialyltransferase
inhibitors{
Kai-Hsuan Chang,a Lenselot Lee,b Jessica Chena and Wen-Shan Li*a
Received (in Cambridge, UK) 20th October 2005, Accepted 8th December 2005
First published as an Advance Article on the web 9th January 2006
DOI: 10.1039/b514915k
glutathione S-transferase inhibitor,11 and lithocholic acid (3), a
A new type of noncompetitive a-2,3-sialyltransferase inhibitor
substrate of nuclear pregnane X receptor,12 as potent inhibitors of
has been synthesized; we report the discovery, preparation and
a-2,3-ST with IC50 values of 350 and 21 mM, respectively.
inhibitory activity of sixteen lithocholic acid analogues.
Synthesis of epiandrosterone derivatives 4–5 and 27–28 was
achieved as depicted in Scheme 1. Epiandrosterone 23 was
esterified with the required amino acids to yield the corresponding
esters 24–26, which were subsequently converted to the epian-
drosterone–amino acid conjugates 4 and 27–28 by alkaline
hydrolysis of the Fmoc and acidolytic cleavage of the remaining
tBu or Boc groups. Efficient formation of epiandrosterone–
phosphate conjugate 5 was obtained from 23 via a three-step
procedure (Scheme 1). The alcohol 23 was converted to the bis-
(p-nitrophenyl)epiandrosteronyl phosphate intermediate 29 by
carefully controlling the release of one equiv. p-nitrophenol
with 1,8-diazabicyclo[5.4.0]undec-7-ene (DBU), followed by
alkaline-mediated replacement of the p-nitrophenyl with
methyl. The final product 5 was accomplished by quantitative
silyldemethylation of the P,P-dimethyl groups of 30 with
Sialic acids, nine-carbon amino sugars bearing a negative charge
under physiological conditions, are cell-type specific markers that
are widely distributed throughout the organism.1 Sialyltransferases
(ST), a class of glycosyltransferases, catalyze the transfer of sialic
acids to the terminal positions of growing oligosaccharide chains
of glycoconjugates.2 Hypersialylation at the non-reducing end of
glycoproteins or glycolipids is a vital biological process, which
plays a role in various physiological and pathological events, such
as cell–cell adhesion, immune defense, tumor cell metastasis,
invasions, and inflammation.3 Therefore, the discovery of potent
inhibitors of ST has become a promising strategy for chemother-
apeutic treatment of cancer4 and can be an invaluable tool to study
ST-dependent physiology in immune responses.5 Lately, ST
inhibitors with a structural mimetic of transition-state analogues,
bisubstrate analogues, donor analogues, and acceptor analogues
have been developed based on cytidine monophosphate-
N-acetylneuraminic acid (CMP-Neu5Ac) or disaccharides.6
Although these compounds have been demonstrated to effectively
inhibit STs, the inhibitors display poor permeability across cell
membranes and their clinical applications for studying the biology
of glycoconjugates are relatively limited.7 To our knowledge, few
ST inhibitors with a cell-permeable property have been reported so
far.7–8 Here we describe the discovery and synthesis of lithocholic
acid analogues, which employ a steroidal moiety to improve their
uptake,9 and demonstrate noncompetitive inhibition of a-2,3-
sialyltransferase (a-2,3-ST) in the presence of CMP-Neu5Ac.
Soyasaponin I (1),8 a rigid pentacyclic system with trisaccharide
from soybean saponin, was a new type of ST inhibitor and showed
a significant inhibition (Ki = 2.1 mM) toward a-2,3-ST in vivo.
Compound 1, where the pentacyclic ring is similar to the main
skeleton of steroidal compounds (Fig. 1), exhibits important
properties relating not only to cell permeability but also to invasive/
metastatic behaviors of cancer cell lines.10 Attractive attention was
stimulated by this report and we reasoned that some steroid-related
compounds might have potential for inhibiting a-2,3-ST activity.
As a result of our initial screening efforts, we have identified
epiandrosterone succinyl ester (2), a potent Schistosoma japonicum
Fig. 1 Main skeleton of lithocholic acid 3 is depicted as being formed
from 1 by cleaving bonds a and b, followed by converting the D-ring into a
5-membered ring, and oxidizing the terminal alcohol into carboxylic acid.
aInstitute of Chemistry, Academia Sinica, Taipei 115, Taiwan.
E-mail: wenshan@chem.sinica.edu.tw; Fax: +886(2)2783 1237;
Tel: +886(2)2789 8662
bDepartment of Chemistry, National Taiwan Normal University, Taipei
116, Taiwan. Fax: +886(2)2932 4249; Tel: +886(2)2935 0749
{ Electronic supplementary information (ESI) available: a-2,3-ST assay,
general experimental procedures and compound characterization. See
DOI: 10.1039/b514915k
Scheme 1 (a) Protected amino acid, DCC, DMAP, DCM, r.t., 30 min;
(b) DBU, DCM, r.t., 30 min; (c) trifluoroacetic acid (TFA), 2% H2O or 4 N
HCl, MeOH, r.t., 1 h; (d) P(O)(p-nitrophenyl)3, DBU, DCM, r.t., 15 h; (e)
DBU, MeOH, DCM, r.t., 15 h; (f) bromotrimethylsilane, DCM, r.t., 30 min.
This journal is ß The Royal Society of Chemistry 2006
Chem. Commun., 2006, 629–631 | 629

View Article Online
preparation of various 1,4-disubstituted 1,2,3-triazoles 18, 21–22
and 36–37 was successful using the copper(I)-catalyzed ligation of
azide 31 and their respective alkynes according to the click
chemistry concept approach.17 Subsequent Boc deprotection or
saponification afforded the racemic derivatives 19 or 20.
The inhibition assay of a-2,3-ST was done as described by
Schmidt15 using CMP-Neu5Ac as a substrate and the modified
disaccharide16 with a 4,5-dimethoxy-2-nitrobenzyl group as a UV-
labelled acceptor (Scheme 3). To enhance the potency in compound
2, we initially prepared four derivatives 4–5 and 27–28 with alterna-
tions of the succinyl ester unit. However, efforts to improve the IC50
were unsuccessful, for example, replacement of the succinyl moiety
of 2 with aspartyl, phosphate, Tyr, or Lys either reduced inhibitory
activity (IC50 > 350 mM for 4–5) or gave a product displaying no
biological functions (27–28) toward a-2,3-ST (Table 1). In addition,
a minor influence on the inhibitory property (IC50 = 450 mM for 6)
of the a isomer of 2 has been observed, suggesting negligible
interaction in the recognition of epiandrosterone/androsterone
analogues with succinyl-stereochemistry by the enzyme.
Scheme 2 (a) MeOH, Amberlite IR120, 80 uC, 12 h; (b) LAH, THF,
0 uC A r.t., 10 min; (c) DPPA, DBU, THF, r.t., 10 h; (d)NaN3, TBAI, 15-
crown-5, 110 uC, 15 h; (e) HBTU, DIPEA, H-Gly-OBut?HCl or L-Asp,
N,N-dimethylformamide (DMF), r.t., 30 min; (f) TFA, DCM, r.t., 1 h; (g)
succinic anhydride, DMAP, pyridine, 80 uC, 15 h; (h) TFA, 2% H2O, r.t.,
1 h; (i) succinic anhydride, DMAP, pyridine, 60 uC, 15 h; (j) CrO3, acetic
acid, 100 uC, 30 min; (k) acetic anhydride, pyridine, r.t., 5 h; (l) t-BuOH,
DMAP, DCC, DCM, r.t., 30 min; (m) NaOCH3, MeOH, r.t., 2 h; (n)
Fmoc-L-Asp(OBut)-OH, DMAP, DCC, DCM, r.t., 30 min; (o) DBU,
DCM, r.t., 1 h; (p) Fmoc-D-Asp(OBut)-OH, DMAP, DCC, DCM, r.t.,
30 min; (q) alkyne, CuSO4?5H2O, sodium ascorbate, THF/H2O (1 : 1), r.t.,
8 h; (r) NaOH, EtOH/H2O (1 : 1), r.t., 5 h.
Lithocholic acid (3), which potentially mimics a pentacyclic ring
of 1 (Fig. 1), showed a promising inhibition constant (IC50
=
21 mM), indicating an acceptable pharmacophore. Among the 16
synthetic analogues (Table 1), the terminal alcohol 7 and its
derivative 8 displayed a 5–17-fold decrease over 3, suggesting that
a carboxylic acid group is important for promoting affinity. This
was further confirmed by utilizing the method of peptide coupling
to extend the terminal carboxylic acid; the inhibitory properties of
compounds 9 and 10 can be restored completely compared to
those of 7–8. To determine the importance of the 3-hydroxyl
position of 3, the inhibition of a-2,3-ST activity by compound 11,
which has a ketone moiety in place of a hydroxyl group, was
evaluated. Compound 11 was found to have an IC50 of 139 mM,
representing 7-fold lower potency than 3. Furthermore, com-
pounds 12 and 13 exhibited a 2-fold potency increase over 3,
indicating that construction of the 3-hydroxyl portion of 3 is
tunable to the interaction of binding affinity. When the terminal
carboxylic acid of 12 was varied, the analogues 14–16 appeared to
reach a low micromolar affinity plateau (Table 1). Surprisingly,
replacement of the D-Asp with L-Asp, as in 17, gave a 2-fold
improvement in potency over 13. In addition, compound 3 was
further optimized by replacing the carboxylic acid with a 1,4-
disubstituted 1,2,3-triazole ligand using the click chemistry
approach. Compounds 20 and 21 showed an 8–12-fold potency
increase over 19, suggesting that adding the positively charged
amino group could abolish affinity. The primary alcohol 18 was at
least 10-fold less potent than 21 and 20-fold less potent than 22.
bromotrimethylsilane.13 Coupling of succinic anhydride to andros-
terone in the presence of 4-(dimethylamino)pyridine (DMAP) and
pyridine produced the desired isomer 6.
In addition, Scheme 2 illustrates the synthesis of lithocholic acid
analogues 7–22. Compound 3 was treated with Amberlite IR120 in
anhydrous methanol, and reduction of the resulting methyl
ester gave the alcohol 7. The hydroxyl group of 7 was converted
to a diphenylphosphate group of 8 in the presence of diphenyl
phosphorylazide (DPPA) and DBU. Condensation of
3
and the corresponding protected amino acids, H-Gly-OBut and
L-H-Asp(OBut)-OBut, using 2-(1H-benzotriazol-1-yl)-1,1,3,3-tetra-
methyluronium hexafluorophosphate (HBTU)/diisopropylethyl-
amine (DIPEA) as coupling agent furnished the protected
conjugates 32–33. Compounds 9–10 were obtained after removal
of the remaining tBu groups. Analogues 15–16 were prepared by
the esterification of 32–33 and succinic anhydride in a manner
similar to 6. Oxidation of 3 with CrO3 in acidic conditions produced
11 in 70% yield. Compound 12 was synthesized in a manner similar
to that described for 6, 15 and 16 by esterification of 3 with succinic
anhydride. In order to prepare analogues 13–14 and 17, the
hydroxyl group of 3 was transformed to an acetyl group, and
subsequent esterification gave 34. Intermediate 35, derived from 34
by deacylation, was converted to the desired product 13 by
dicyclohexylcarbodiimide (DCC)-promoted coupling followed by
deprotection of the Fmoc, Boc and tBu groups. Esterification of the
secondary alcohol 35 gave compound 14. The D-form amino acid–
lithocholic acid conjugate 17 was synthesized in a manner similar to
that described for 13. The 1,2,3-triazoles 18–22 required first the
synthesis of the terminal azide 31 (Scheme 2). The conversion of
diphenylphosphate 8 to the corresponding azide 31 was achieved in
the presence of excess sodium azide (NaN3), a catalytic amount of
tetrabutylammonium iodide (TBAI), and 15-crown-5.14 Reliable
Scheme 3 a-2,3-ST catalyzed sialylation of acceptor.
630 | Chem. Commun., 2006, 629–631
This journal is ß The Royal Society of Chemistry 2006

View Article Online
Table 1 a-2,3-Sialyltransferasea inhibitors
IC50
(mM)b Compound
IC50
(mM)b
Compound
#
4
#
480
14
13
5
6
400
450
15
16
22
18
7
351
17
6
8
9
. 100
18 . 100
Fig. 2 Lineweaver–Burk plot of the results of rat a-2,3-ST inhibition
assay for the synthetic inhibitor 22.
25
19
20
83
7
inhibitory activities among these synthetic derivatives. The results
demonstrated here should pave the way for the design of other ST
inhibitors.
10
16
We are grateful for funding from the National Science Council
(NSC 94-2113-M-001-017) and Academia Sinica (program project).
11
12
139
12
21
22
10
5
Notes and references
1 C. Traving and R. Schauer, Cell. Mol. Life Sci., 1998, 54, 1330;
M. Schwarzkopf, K.-P. Knobeloch, E. Rohde, S. Hinderlich,
N. Wiechens, L. Lucka, I. Horak, W. Reutter and R. Horstkorte,
Proc. Natl. Acad. Sci. USA, 2002, 99, 5267.
2 A. Harduin-Lepers, V. Vallejo-Ruiz, M.-A. Krzewinski-Recchi,
B. Samyn-Petit, S. Julien and P. Delannoy, Biochimie, 2001, 83, 727.
3 E. C. Seales, G. A. Jurado, B. A. Brunson, J. K. Wakefield, A. R. Frost
and S. L. Bellis, Cancer Res., 2005, 65, 4645; R. Kannagi, M. Izawa,
T. Koike, K. Miyazaki and N. Kimura, Cancer Sci., 2004, 95, 377;
L. K. Mahal, N. W. Charter, K. Angata, M. Fukuda, D. E. Koshland
and C. R. Bertozzi, Science, 2001, 294, 380.
4 M. M. Fuster and J. D. Esko, Nat. Rev. Cancer, 2005, 57, 526.
5 M. A. Daniels, K. A. Hogquist and S. C. Jameson, Nat. Immunol., 2002,
3, 903.
13
12
a
Note that rat liver a-2,3-ST was obtained from CalBiochem at a
concentration of 1 mU/mL. The inhibition assay was carried out as
b
described in ref. 15, except for the use of the modified disaccharide16
as the acceptor.
6 D. Skropeta, R. Schwoerer, T. Haag and R. R. Schmidt, Glycoconjugate
J., 2004, 21, 205; K.-H. Chang, Y. S. Tao and W.-S. Li, Synlett, 2004,
37; L. J. Whalen, K. A. McEvoy and R. L. Halcomb, Bioorg. Med.
Chem. Lett., 2003, 13, 301; R. Schwoerer and R. R. Schmidt, J. Am.
Chem. Soc., 2002, 124, 1632; B. Muller, C. Schaub and R. R. Schmidt,
Angew. Chem., Int. Ed., 1998, 37, 2893.
These results are consistent with previous observations concerning
the value of the terminal carboxylic acid in binding. Analogue 22
was the most active compound with an IC50 of 5 mM. The
Lineweaver–Burk plot (Fig. 2) shows that compound 22 exhibits a
7 K.-Y. Lee, H. G. Kim, M. R. Hwang, J. I. Chae, J. M. Yang, Y. C. Lee,
Y. K. Choo, Y. I. Lee, S.-S. Lee and S.-I. Do, J. Biol. Chem., 2002, 277,
49341; A. K. Sarkar, T. A. Fritz, W. H. Taylor and J. D. Esko, Proc.
Natl. Acad. Sci. USA, 1995, 92, 3323.
8 T.-W. Lin, W.-W. Chang, C.-C. Chen and Y.-C. Tsai, Biochem. Biophys.
Res. Commun., 2005, 331, 953; C.-Y. Wu, C.-C. Hsu, S.-T. Chen and
Y.-C. Tsai, Biochem. Biophys. Res. Commun., 2001, 284, 466.
9 P. Yu, B. Liu and T. Kodadek, Nat. Biotechnol., 2005, 23, 746.
10 C. C. Hsu, T. W. Lin, W. W. Chang, C. Y. Wu, W. H. Lo, P. H. Wang
and Y. C. Tsai, Gynecol. Oncol., 2005, 96, 415.
noncompetitive inhibition (Ki
= 2.2 mM) toward cytidine
monophosphate N-acetylneuraminic acid (CMP-Neu5Ac).
However, we failed to identify the mode of inhibition between
22 and acceptor due to the solubility problem derived from the
modified disaccharide with a 4,5-dimethoxy-2-nitrobenzyl moiety.
It is possible that compound 22 binds to a-2,3-ST and alters the
structure of the active site, resulting in the decrease of enzyme
activity. Due to the potential of compounds 11, 13, 15–16, and 19–
22 as anticancer drugs, our collaborators recently evaluated their
cytotoxicity using in vitro assays against several tumor cell lines.
These results will be published elsewhere.
11 S.-C. Jao, J. Chen, K. Yang and W.-S. Li, Bioorg. Med. Chem., 2006,
14, 304.
12 S. J. Kim, E.-K. Bang, H. J. Kwon, J. S. Shim and B. H. Kim,
ChemBioChem, 2004, 5, 1517.
13 H. B. Wood and B. Ganem, Tetrahedron Lett., 1993, 34, 1403.
14 F. Liu and D. J. Austin, J. Org. Chem., 2001, 66, 8643.
15 C. Schaub, B. Mu¨ller and R. R. Schmidt, Glycoconjugate J., 1998, 15, 345.
16 S. B. Cohen and R. L. Halcomb, J. Org. Chem., 2000, 65, 6145.
17 V. V. Rostovtsev, L. G. Green, V. V. Fokin and K. B. Sharpless,
Angew. Chem., Int. Ed., 2002, 41, 2596.
Because of the physiological and pathological significance of ST,
there is strong demand for an efficient approach to searching for
cell-permeable inhibitors. This is the first report of a strategy for
designing effective a-2,3-ST inhibitors based on the structure of an
authentic inhibitor. Analogues 17, 20, and 22 display more potent
This journal is ß The Royal Society of Chemistry 2006
Chem. Commun., 2006, 629–631 | 631