A rationally designed inhibitor of α-1,3-galactosyltransferase

Full text of DOI:10.1021/ja9905391

J. Am. Chem. Soc. 1999, 121, 5829-5830
5829
A Rationally Designed Inhibitor of
r-1,3-Galactosyltransferase
Yong Jip Kim, Mie Ichikawa, and Yoshitaka Ichikawa*
Department of Pharmacology and Molecular Sciences
Johns Hopkins UniVersity School of Medicine
Baltimore, Maryland 21205
ReceiVed February 22, 1999
Because cell-surface glycoconjugates are known to play an
1
important role in a variety of cell-cell interactions, inhibitors
of the oligosaccharide-synthesizing enzymes, glycosyltransferases,
have been actively studied. A number of designed inhibitors have
been described: these are analogues of an acceptor molecule
2
3
(
oligosaccharide or peptide), a donor molecule (sugar nucle-
4
5
6
otide), and a transition-state mimic (mono- or bisubstrate).
Figure 1. Proposed reaction mechanism of R1,3-GalT and possible
interaction between the 1-N-iminosugar-based inhibitor and the enzyme.
Although a few successful examples of acceptor and bisubstrate
analogues have been reported, they are complex molecules and
are difficult to synthesize and unstable due to the incorporated
nucleoside diphosphate moiety.
tor-enzyme complex via a favorable electrostatic interaction
1
1
We have chosen R-1,3-galactosyltransferase (R1,3-GalT) as the
(Figure 1), as it does with â-galactosidase (â-Gal’ase). In this
paper we describe the rational design of an 1-N-iminosugar-based
UDP-galactose analogue 1, which was found to be a selective
potent inhibitor of R1,3-GalT but not of â1,4-GalT and shows a
significantly reduced inhibitory activity against â-Gal’ase because
of its incorporation of a neutral analogue of the UDP moiety.
In a preliminary screening of inhibitors of R1,3-GalT, we
confirmed that the iminosugar 2 alone could inhibit the R1,3-
7
target enzyme for our inhibitor design because of its biochemical
importance in xenotransplant rejection caused by the recipient’s
8
anti-Gal antibody reaction to the donor’s R-Gal epitope. Since
the R1,3-GalT reaction proceeds with overall retention of the
configuration of the galactose anomeric position via double
displacement, we assumed that a general base group (a carboxy-
late-containing amino acid residue) could be transiently involved
9
in forming a glycosyl-enzyme complex, as shown in Figure 1,
i
GalT reaction with a K of 70 µM but did not inhibit â1,4-GalT,
in a manner similar to that of a retaining glycosidase reaction.10
If this is the case, we hypothesized that a derivative of our newly
developed galactose-type 1-N-iminosugar 2 would be able to
efficiently inhibit the enzyme activity by forming a rigid inhibi-
whereas deoxygalactonojirimycin inhibited neither of the galac-
tosyltransferases effectively.12 Although we have identified the
1-N-iminosugar 2 as a novel inhibitor of R1,3-GalT, this imino-
sugar was also a very potent inhibitor of â-Gal’ase, with a K
i
of
4
nM.11 Our objective in the present study was to improve the
*
Corresponding author. E-mail: ichikawa@jhmi.edu.
potency of iminosugar 2 in inhibiting R1,3-GalT and to reduce
its inhibitory activity against â-Gal’ase by introducing various
groups on its nitrogen atom.
We have divided the UDP-galactose molecule into three
segments: a galactose residue (segment A), a pyrophosphate
moiety (segment B), and a uridine residue (segment C). Our
(
1) Varki, A. Glycobiology 1993, 3, 97-130.
(
2) (a) Hindsgaul, O.; Kaur, K. J.; Srivastava, G.; Blaszczyk-Thurin, M.;
Crawley, S. C.; Heerze, L. D.; Palcic, M. M. J. Biol. Chem. 1991, 266, 17858-
1
1
7862. (b) Lowary, T. L.; Hindsgaul, O. Carbohydr. Res. 1993, 249, 163-
95. (c) Kajihara, Y.; Kodama, H.; Wakabayashi, T.; Sato, K.; Hashimoto,
H. Carbohydr. Res. 1993, 247, 179-193.
(
3) (a) Bause, E. Biochem. J. 1983, 209, 323-330. (b) Hendrickson, T.
L.; Spencer, J. R.; Kato, M.; Imperiali, B. J. Am. Chem. Soc. 1996, 118, 7636-
637.
4) (a) Korytnyk, W.; Angelino, N.; Klohs, W.; Bernacki, R. Eur. J. Med.
7
(
Chem. 1980, 15, 77-84. (b) Camarasa, M. J.; Fernandez-Resa, P.; Garcia-
Lepez, M. T.; De las Heras, F. G.; Mendez-Castrillon, P. P.; Alarcon, B.;
Carrasco, L. J. Med. Chem. 1985, 28, 40-46. (c) Vaghefi, M. M.; Bernacki,
R. J.; Dalley, N. K.; Wilson, B. E.; Robins, R. K. J. Med. Chem. 1987, 30,
1
383-1391. (d) Cai, S.; Stroud, M. R.; Hakomori, S.; Toyokuni, T. J. Org.
Chem. 1992, 57, 6693-6696. (e) Endo, T.; Kajihara, Y.; Kodama, H.;
Hashimoto, H. Bioorg. Med. Chem. 1996, 4, 1939-1948. (f) Wang, R.;
Steensma, D. H.; Takaoka, Y.; Yun, J. W.; Kajimoto, T.; Wong, C.-H. Bioorg.
Med. Chem. 1997, 5, 661-672. (g) M u¨ ller, B.; Schaub, C.; Schmidt, R. R.
Angew. Chem., Int. Ed. 1998, 37, 2893-2897.
(
5) Platt, F. M.; Neises, G. R.; Reinkensmeier, G.; Townsend, M. L.; Perry,
V. H.; Proia, R. L.; Winchester, B.; Dwek, R. A.; Buttlers, T. D. Science
997, 276, 428-431.
6) (a) Inokuchi, J.; Radin, N. S. J. Lipid Res. 1987, 28, 565-571. (b)
Palcic, M. M.; Heerze, L. D.; Srivastava, O. P.; Hindsgaul, O. J. Biol. Chem.
1
(
1
989, 264, 17174-17181. (c) Qiao, L.; Murry, B. W.; Shimazaki, M.; Schultz,
J.; Wong, C.-H. J. Am. Chem. Soc. 1996, 118, 7653-7662. (d) Hashimoto,
H.; Endo, T.; Kajihara, Y.; J. Org. Chem. 1997, 62, 1914-1915.
strategy was to replace these three segments with a galactose-
type 1-N-iminosugar, a vicinal diol derived from L-tartaric acid,
and a 5′-thio-uridine, respectively, to avoid an unnecessary
negative charge on the molecule and to increase the stability of
the designed inhibitor.
(7) The Hindsgaul group has reported the first inhibitor of R1,3-GalT which
is an acceptor oligosaccharide analogue, see: Helland, A. C.; Hindsgaul, O.;
Palcic, M. M.; Stults, C. L. M.; Macher, B. A. Carbohydr. Res. 1995, 276,
9
1-98.
(
8) Galili, U. Immunol. Today 1993, 14, 480-482.
(
9) This type of “two-step double-displacement transfer mechanism” has
been proposed by Lowary and Hindsgaul for other R-glycosyltransferases,
see: ref 2b.
(11) Ichikawa, Y.; Igarashi, Y.; Ichikawa, M.; Suhara, Y. J. Am. Chem.
Soc. 1998, 120, 3007-3018.
(12) Takayama, S.; Chung, S. J.; Igarashi, Y.; Ichikawa, Y.; Sepp, A.;
Lechler, R. I.; Wu, J.; Hayashi, T.; Siuzdak, G.; Wong, C.-H. Bioorg. Med.
Chem., in press.
(10) (a) Koshland, D. E. Biol. ReV. 1953, 28, 416-436. (b) Sinnot, M. L.
Chem ReV. 1990, 90, 1171-1202. (c) White, A.; Tull, D.; Johns, K.; Withers,
S. G.; Rose, D. R. Nat. Struct. Biol. 1996, 3, 149-154.
1
0.1021/ja9905391 CCC: $18.00 © 1999 American Chemical Society
Published on Web 06/05/1999

5830 J. Am. Chem. Soc., Vol. 121, No. 24, 1999
Communications to the Editor
Table 1. Inhibitory Potency of the Compounds against R- and
â-Galactosyltransferases and R- and â-Galactosidases
whether the amino group could interact with the putative general
base in the enzyme’s active site, and the resulting UDP mimetic
was equivalent to 1 except for the fact that it lacked the galactose
framework. For this synthesis, the primary OH group of 12 was
sequentially converted to an amino group (Scheme 2). The
resulting amino compound 6 showed an improved inhibitory
potency, with an IC50 of 5 mM for R1,3-GalT. This result
suggested the presence of a general base (either an Asp or Glu
residue) in the active site, a situation which had previously been
inferred from the inhibition data obtained for the 1-N-iminosugar
inhibitory activity (IC50) against
compd
R-GalTasea
â-GalTaseb
â-Gal’asec
R-Gal’ased
e
MIf
NI
NI
NI
NI
NI
1
2
3
4
5
6
7.5 µM
17 µM
20 µM
200 µM
45 µM
50 µM
NI
g
h
15 µM
10 nM
45% at 5 mM
5 µM
7 µM
NI
i
70 µM
20% at 5 mM
5 mM
NI
NI
2
in our earlier study12 and others.7
a
Recombinant porcine R-1,3-galactosyltransferase from Calbiochem
Finally, we combined all three segments and synthesized a
b
(
San Diego, CA). Bovine â-1,4-galactosyltransferase from Sigma (St.
neutral UDP-galactose analogue 1 (Scheme 2). Oxidation of 12,
c
d
Louis, MO). â-Galactosidase (Aspergillus oryzae) from Sigma. R-
20
followed by reductive amination with the galactose-type 1-N-
e
f
Galactosidase (coffee beans) from Sigma.
K
is ) 4.4 ( 1.3 µM. No
iminosugar 2, afforded the conjugate 1 after acidic deprotection.
g
h
j
inhibition at 5 mM.
µM.
K
is ) 10 ( 1 µM.
K
i
) 4 nm. Kis ) 55 ( 14
This conjugate showed a potent inhibitory activity against R1,3-
GalT, with a K of 4.4 µM, and a diminished inhibitory potency
i
Scheme 1a
for â-Gal’ase (IC50 ) 17 µM). Since the â1,4-GalT reaction
involves inversion of the stereochemistry of the UDP-Gal, there
would be no glycosyl-enzyme complex formation with a nearby
carboxylate such as shown in Figure 1, and as we expected,
compounds 1-6 did not inhibit â1,4-GalT at concentrations up
to 5 mM (Table 1). We have also examined the inhibitory
potencies of compounds 1-6 against R-Gal’ase and found a
moderate increase in inhibitory activity (IC50 ) 20 µM for 1)
with increasing size of the substituent on the C-1 amino group
(
Table 1), a finding that is consistent with previous observations
11
concerning N-modification of those 1-N-iminosugars. An analy-
sis of the inhibition kinetics data21 indicated that, to our surprise,
neither 1 nor 2 showed a competitive mode of inhibition against
UDP-Gal, even though 1 is a UDP-Gal analogue. Similar
observations have been reported in inhibition studies of other
glycosyltransferases in which acceptor oligosaccharide analogues
were used as inhibitors.10 The type of inhibition we observed for
1 and 2 suggests that they interact not only with the free enzyme
a
Reagents and conditions: (a) MOMCl/iPr
Cl
phosphate buffer, pH 6.8; (d) 1 N HCl; (e) (i) TsCl/pyr, (ii) KSAc/DMF;
2
NEt/CH
, (ii) 2/BH
2
Cl
2 4
; (b) Bu NF/
THF; (c) (i) Dess-Martin periodinane/CH
2
2
3
‚pyr/MeOH-
(
f) (i) K
iii) Bu
/Lindlar catalyst/MeOH, (iii) CF
CH Cl , (ii) 2/BH ‚pyr/MeOH-phosphate buffer, pH 6.8, (iii) CF
2
CO
3
/MeOH, (ii) 2′,3′-O-isopropylidene-5′-O-tosyl-uridine/DMF,
CO H; (h) (i) TsCl/pyr, (ii) NaN /DMSO, (iii)
CO H; (i) Dess-Martin oxidation/
CO H.
(
4
NF/THF; (g) CF
3
2
3
H
2
3
2
21
2
2
3
3
2
but also with the enzyme-substrate complex.
In conclusion, we have demonstrated the rational design of an
inhibitor of R1,3-GalT by introducing an additional neutral UDP
analogue into the relatively simple motif of an 1-N-iminosugar.
The resulting conjugate showed an improvement in the desired
specific inhibitory potency against R1,3-GalT and suppressed the
unwanted inhibitory activity against the glycosidase. Although
we do not yet completely understand the structure-inhibitory
activity relationship determined by the stereochemistry of the diol
moiety of inhibitor 1, our approach may indicate a general strategy
for designing effective inhibitors of glycosyltransferases. We are
currently extending this design strategy to other 1-N-iminosugars
for potential inhibitor of glycosyltransferases.
When a simple alkyl group was introduced into 2 (segment
A), the inhibitory activity of the resulting N-butyl iminosugar 3
against R1,3-GalT was decreased by >500-fold (45% inhibition
at 5 mM, Table 1),13 and that for â-Gal’ase fell by >400-fold
(IC50 ) 5 µM). Incorporation of segment B, a triol group (a
pyrophosphate-like diol moiety), into 2 was examined with the
expectation that the diol moiety might fit into the enzyme binding
site to which the pyrophosphate moiety normally binds.
The synthesis was accomplished by reductive amination of 2
1
4
with the aldehyde derived from L-tartaric acid (Scheme 1).
A
15
TBDMS derivative of L-threitol 7 was converted to an MOM
derivative 9, which was oxidized16 and reductively aminated with
Acknowledgment. We appreciate the useful suggestions made by the
reviewers. Support from the American Cancer Society (JFA-515), the
National Institutes of Health (GM 52324), and the Mizutani Foundation
of Glycobiology are greatly appreciated. The NMR studies were
performed in the Biochemistry NMR Facility at Johns Hopkins University,
which was established by a grant from the National Institutes of Health
(GM 27512) and a Biomedical Shared Instrumentation Grant (1S10-
RR06262-0).
2
to afford 4. Compound 4, with the triol moiety as the pyro-
phosphate mimetics, was found to restore the inhibitory activity
against R1,3-GalT to some extent, with a K of 70 µM (Table 1).
i
Uridine (segment C) was not an inhibitor of R1,3-GalT;
therefore, we next prepared a neutral analogue of UDP 5 to see
whether this UDP mimetic could inhibit the enzyme activity. For
this conjugation we used a sulfur atom (sulfide) in order to confer
flexibility upon the linkage beween the diol moiety (segment B)
and the uridine residue. The threitol derivative 7 was converted
Supporting Information Available: Experimental procedures for the
synthesis of 1, 4, 5, and 6 and kinetic data (PDF). This material is available
free of charge via the Internet at http://pubs.acs.org.
17
to the thio derivative 11, which was then coupled with a 5′-O-
tosyl derivative of uridine18 to give 5. The neutral UDP mimetics
JA9905391
5
only weakly inhibited R1,3-GalT activity (20% inhibition at 5
19
mM). We next introduced an amino group into 5 to determine
(17) When a 5′-thio derivative of uridine was subjected to the coupling
reaction with a tosyl derivative of the threiol, only a small amount of the
coupling product was obtained, probably because of the competing intramo-
lecular cyclization of the thio moiety to the uracil group.
(
13) In the present study we used a recombinant porcine R-1,3-galacto-
syltransferase obtained from Calbiochem (San Diego, CA). The inhibition
assay was carried out as described in ref 12, except for the use of pH 7.0 and
LacNAc as the acceptor.
(18) (a) Levene, P. A.; Tipson, R. S. J. Biol. Chem. 1934, 106, 113-124.
(b) Horwitz, J. P.; Tomson, A. J.; Urbanski, J. A.; Chua, J. Org. Chem. 1962,
27, 3045-3048.
(
14) Synthesis of compounds 1, 4, 5, and 6 is described in Supporting
Information.
i
(19) UDP was found to inhibit R1,3-GalT with a K of 13 µM, see: ref 12.
(
15) Iida, H.; Yamazaki, N.; Kibayashi, C. J. Org. Chem. 1987, 52, 3337-
(20) Yoshida, T.; Lee, Y. C. Carbohydr. Res. 1994, 251, 175-186.
(21) Segel, I. H. Enzyme Kinetics; John Wiley & Sons: New York, 1973;
Chapter 4. See also Supporting Information.
3
342.
(16) Dess, D. B.; Martin, J. C. J. Org. Chem. 1983, 48, 4155-4156.