A drug targeting motif for glycosidase inhibitors: An iminosugar-boronate shows unexpectedly selective β-galactosidase inhibition

Article abstract of DOI:10.1016/S0040-4039(02)02196-2

Boronic acids were tethered to iminosugars in compounds such as 8 and 13 in order to increase their affinity for cell surfaces where glycoprotein processing enzymes are operative. Surprisingly, this modification diminished α-mannosidase inhibition while increasing β-galactosidase inhibitory activity (8: K_i=2.0×10^-4 M versus E. coli β-galactosidase). The presence of a boronate in 8 and 13 has a profound impact on the specificity of this inhibition.

Full text of DOI:10.1016/S0040-4039(02)02196-2

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 43 (2002) 8905–8908
A drug targeting motif for glycosidase inhibitors:
an iminosugar–boronate shows unexpectedly selective
b-galactosidase inhibition
Leland L. Johnson, Jr.a,† and Todd A. Houstona,b,
*
aDepartment of Chemistry, Virginia Commonwealth University, Richmond, VA 23284-2006, USA
bSchool of Science, Griffith University, Nathan, QLD 4111, Australia
Received 31 July 2002; accepted 1 October 2002
Abstract—Boronic acids were tethered to iminosugars in compounds such as 8 and 13 in order to increase their affinity for cell
surfaces where glycoprotein processing enzymes are operative. Surprisingly, this modification diminished a-mannosidase inhibition
while increasing b-galactosidase inhibitory activity (8: Ki=2.0×10−4 M versus E. coli b-galactosidase). The presence of a boronate
in 8 and 13 has a profound impact on the specificity of this inhibition. © 2002 Elsevier Science Ltd. All rights reserved.
The effect of glycosidase enzymes on the structure of
cell-surface carbohydrates translates into important
biological consequences for processes such as viral
infection1 and tumor metastasis.2 This has attracted
research to identify novel inhibitors of glycosidases as
candidates for antitumor,3,4 antiviral5–8 and antidiabetic
therapies.9 There are a variety of known glycosidase
inhibitors, many of which are naturally-occurring alka-
loids that mimic the carbohydrate substrate.10–13 In
many cases, these iminosugars are believed to become
protonated in the active site and thus serve as mimics of
the electron-deficient transition state that must be
involved in glycoside hydrolysis.14 While some
inhibitors are highly selective for a given glycosidase,
the abundance of closely related enzymes within the
body makes selective inhibition a challenge. Here, we
report that attachment of a boronic acid to 1 converts
of enzyme action for the glycoprotein trimming
enzymes Mannosidase I and II (i.e. the boronate will
mask the mannose substrate and present the inhibitor
to the enzyme).
We are currently synthesizing both pyrrolidine (e.g.
1–4) and piperidine based ring systems of the general
structures shown in Fig. 1 to test this hypothesis. The
first chemistry that we sought to develop was a means
of installing boronic acids onto previously constructed
iminosugar systems. The target compounds were envis-
aged to arise from N-allyl amines through hydrobora-
tion with excess borane followed by conversion to the
boronic acid. Accordingly, the dimesylate 5 was synthe-
sized by the procedure described by Fleet15 and a
the a-mannosidase inhibitor to
a b-galactosidase
inhibitor.
We are interested in targeting drugs to specific cell-sur-
face carbohydrate structures as a means of increasing
their selectivity and efficacy. Boronic acid derivatives of
iminosugars would allow for reversible binding of gly-
coprotein carbohydrates based on the known affinity of
boronates for cis-vicinal diols in mannose and galac-
tose. This would target the inhibitor directly to the site
* Corresponding author. Tel.: +61-7-3875-7217; fax: +61-7-3875-6572;
e-mail: t.houston@mailbox.gu.edu.au
† Present address: Adenosine Therapeutics, LLC Charlottesville, VA,
USA.
Figure 1. Generalized target structures and known glycosi-
dase inhibitors.
0040-4039/02/$ - see front matter © 2002 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(02)02196-2

8906
L. L. Johnson, Jr., T. A. Houston / Tetrahedron Letters 43 (2002) 8905–8908
double displacement with allylamine provided 6 in
excellent overall yield from mannose. This compound
was converted to both the boronated target 8 and
non-boronated compound 9 through two separate steps
shown in Scheme 1. Use of a 20-fold excess of borane
ensures the formation of a monoboronate ester upon
methanol quench. The unreacted borane is converted to
trimethylborate that can be removed in vacuo. Com-
pound 6 was also converted to N-allyl-1,4-dideoxy-1,4-
imino-D-talitol by reaction with methanolic HCl and
the olefin was reduced through catalytic hydrogenation
to yield 9 (94% for two steps).
The phenyl derivatives 1215 and 13 were synthesized
according to Scheme 2. The known benzylamine 1015
was again synthesized from 5 by double displacement
of the dimesylate. Acetonide deprotection of 10 fur-
nished compound 12 while benzyl deprotection pro-
vided the secondary amine 11. This allowed the
phenylboronate to be installed through reductive ami-
nation to yield the final analog 13 (58% for three
steps).16 As with compounds 8 and 9, the aromatic
iminosugars 12 and 13 differ only in the substitution of
a hydrogen for a boronic acid, thus offering control
compounds to study the effect of boronate substitution
on enzyme inhibition.
Scheme 2. Synthesis of inhibitors 12 and 13.
very little inhibition (ca. 5–10%) of a-mannosidase
(Jack bean) and actually gave a slight increase in a-
galactosidase (green coffee bean) activity at 1.0 mM
concentration. Remarkably, both compounds were
much more effective at inhibiting b-galactosidase (E.
coli ) than their non-boronated counterparts 9 and 12.
An electrophilic boron species alone, in the form of
boric acid, is not sufficient to inhibit the enzyme.
Each compound was isolated as its hydrochloride salt
and was tested for inhibition of three common glycosi-
dases (Table 1). The boronic acids 8 and 13 displayed
Fleet has shown that the 1,4-iminotalitol compound 1
lacking N-alkyl substitution is a specific and competi-
tive inhibitor of human liver a-mannosidase.15 A closely
related class of compounds based on the 5-OH epimer
2 shows alternating a-mannosidase or a-fucosidase
inhibition depending on the nature of the N-alkyl
group.17 The structurally reduced lyxitol compound 3 is
also epimeric to the talitol stereochemistry at the
bridgehead. It is a potent inhibitor of a-galactosidase
while inhibiting b-galactosidase (A. niger) only at three-
fold higher concentration.18 Huber and Gaunt discov-
ered that -ribose inhibits b-galactosidase from E. coli
L
more effectively than its enantiomer.19 The talitol com-
pounds here match the stereochemical disposition of
L
-ribofuranose and might bind in a similar manner such
that it mimics distorted transitions of
ose found on the reaction coordinate.
D-galactopyran-
In the present case, it is not clear exactly how the
addition of a boronic acid imparts selective inhibitory
activity in this series. It is possible that the boron is
binding to an active site nucleophile in the aglycone
cavity since boronates make up a powerful class of
enzyme
inhibitors,
particularly
versus
serine
proteases.20–22 Inspection of the three-dimensional
structure of the E. coli b-galactosidase reveals that there
are residues outside of the catalytic carboxylates that
might act as nucleophiles toward the boron center,
most importantly Met-502.23 The sulfur of this highly-
conserved amino acid has been alkylated with an
Scheme 1. Synthesis of inhibitors 8 and 9.

L. L. Johnson, Jr., T. A. Houston / Tetrahedron Letters 43 (2002) 8905–8908
Table 1. Glycosidase inhibition at 1.0 mM inhibitor concentration
8907
Compds
E. coli b-galactosidase
Coffee bean a-galactosidase
Jack bean a-mannosidase
8
9
75–86
13
(10)
NI
B5
10
12
13
NI
45–55
NI
B5
(B5)
NT
NI
10
NT
Boric acid
Values are expressed as a percentage of activity lost (gained) in the presence of inhibitor. Results are from three independent determinations where
values fall within 5% range unless a range is noted. UV assays were performed at the enzyme’s optimal pH with requisite p-nitrophenylglycoside
substrates (NI, no inhibition; NT, not tested).
anomeric diazomethane derivative of galactose,24 and is
believed to participate as a nucleophile during carbohy-
drate hydrolysis in the absence of Glu-461.25 The elec-
trophilic site of the C-galactosyl affinity label used by
Sinnott24 is roughly similar to the position of the
boronate in 8 and 13. Alternatively, the boronate might
form a cyclic ester with the C-5/C-6 vicinal diol to create
the active species.26 Nonetheless, the specificity change
that the boronic acid imparts on 1, a selective mannosi-
dase inhibitor (Ki=1.2×10−4 M versus human liver
lysosomal a-mannosidase), is dramatic (8: Ki=2.0×10−4
M versus b-galactosidase) and should prove an effective
way to enhance other, more potent b-galactosidase
inhibitors.
9. (a) Joubert, P. H.; Bam, W. J.; Manyane, N. Eur. J. Clin.
Pharmacol. 1986, 30, 253. For general reviews, see: (b)
Winchester, B.; Fleet, G. W. J. J. Carbohydr. Chem. 2000,
19, 471; (c) Hausler, H.; Kawakami, R. P.; Mlaker, E.;
Severn, W. B., Wrodnigg, T. M.; Stutz, A. E. J. Carbo-
hydr. Chem. 2000, 19, 435; (d) Dwek, R. A.; Butters, T.
D.; Platt, F. M.; Zitzmann, N. Nature Rev. Drug Disc.
2002, 1, 65.
10. Fuhrmann, U.; Bause, E.; Ploegh, H. Biochim. Biophys.
Acta 1985, 825, 95.
11. Ganem, B. Acc. Chem. Res. 1996, 29, 340.
12. Sears, P.; Wong, C.-H. Chem. Commun. 1998, 1161.
13. Bols, M.; Lillelund, V. H.; Jensen, H. H.; Liang, X.
Chem. Rev. 2002, 102, 515.
14. (a) Dale, M. P.; Ensley, H. E.; Kern, K.; Sastry, K. A.
R.; Byers, L. D. Biochemistry 1985, 24, 3530; (b) Noten-
boom, V.; Williams, S. J.; Hoos, R.; Withers, S. G.; Rose,
D. R. Biochemistry 2000, 39, 11553; (c) See also: Jeong,
J.-H.; Murray, B. W.; Takayama, S.; Wong, C.-H. J. Am.
Chem. Soc. 1996, 118, 4227.
Acknowledgements
Financial support from the Mary Kapp Fund (VCU),
Jeffress Memorial Trust, and Horsley Cancer Fund
(Virginia Academy of Sciences) is gratefully acknowl-
edged. We appreciate productive discussions with Dr.
Evelyn Wolff (Insmed, Inc.) regarding the enzyme assays.
15. Fleet, G. W. J.; Son, J. C.; Green, D. S. C.; di Bello, I.
C.; Winchester, B. Tetrahedron 1988, 44, 2649.
16. Gray, C. W., Jr.; Houston, T. A. J. Org. Chem. 2002, 67,
5426.
17. Al Daher, S.; Fleet, G.; Namgoong, S. K.; Winchester, B.
References
Biochem. J. 1989, 258, 613.
18. Fleet, G. W. J.; Nicholas, S. J.; Smith, P. W.; Evans, S.
V.; Fellows, L. E.; Nash, R. J. Tetrahedron Lett. 1985, 26,
3127.
1. Feizi, T.; Larkin, M. Glycobiology 1990, 1, 17.
2. Hakomori, S. Adv. Cancer Res. 1989, 52, 257.
3. Humphries, M. J.; Matsumoto, K.; White, S. L.; Olden,
K. Proc. Natl. Acad. Sci. USA 1986, 83, 1752.
4. Pili, R.; Chang, J.; Partis, R. A.; Mueller, R. A.; Chrest,
F. J.; Passaniti, A. Cancer Res. 1995, 55, 2920.
5. Gruters, R. A.; Neefjes, J. J.; Tersmette, M.; de Goede,
R. E. Y.; Tulp, A.; Huisman, H. G.; Miedema, F.;
Ploegh, H. L. Nature (London) 1987, 330, 74.
6. Walker, B. D.; Kowalski, M.; Goh, W. C.; Kozarsky, K.;
Krieger, M.; Rosen, C.; Rohrschneider, L.; Haseltine, W.
A.; Sodroski, J. Proc. Natl. Acad. Sci. USA 1987, 84,
8120.
19. Huber, R. E.; Gaunt, M. T. Arch. Biochem. Biophys.
1983, 220, 263.
20. (a) Torssell, K. Ark. Kemi. 1957, 10, 529; (b) Antonov, V.
K.; Ivaniva, T. V.; Berezin, I. V.; Martinek, K. FEBS
Lett. 1970, 7, 23.
21. Phillip, M.; Bender, M. L. Proc. Natl. Acad. Sci. USA
1971, 68, 478.
22. Kettner, C. A.; Shenvi, A. B. J. Biol. Chem. 1984, 259,
15106.
23. Jacobson, R. H.; Zhang, X.-J.; DuBose, R. F.; Matthews,
B. W. Nature (London) 1994, 369, 761.
7. Karpas, A.; Fleet, G. W. J.; Dwek, R. A.; Petursson, S.;
Namgoong, S. K.; Ramsden, N. G.; Jacob, G. S.;
Rademacher, T. W. Proc. Natl. Acad. Sci. USA 1988, 85,
9229.
8. von Itzstein, M.; Wu, W.-Y.; Kok, G. B.; Pegg, M. S.;
Dyason, J. C.; Jin, B.; Phan, T. V.; Smythe, M. L.;
White, H. F.; Oliver, S. W.; Colman, P. M.; Varghese, J.
N.; Ryan, D. M.; Woods, J. M.; Bethell, R. C.; Hotham,
V. J.; Cameron, J. M.; Penn, C. R. Nature (London) 1993,
363, 418.
24. Sinnott, M. L. FEBS Lett. 1973, 94, 1.
25. Sinnott, M. L. Chem. Rev. 1990, 90, 1171.
26. The 11B NMR spectrum of 8 in D2O is simple at pD 2.0,
but becomes complex above pD 7.0 indicative of free
amine-catalyzed esterification of the boronate. The
hydrochloride salts 8, 9, 12 and 13 were collected by
filtration and deemed of sufficient purity (>95% by 1H
NMR) for use in the initial enzyme assays. Compound 8
was purified by precipitation with HCl in Et2O for deter-
mining its Ki value versus b-galactosidase. N-(Propyl-3-

8908
L. L. Johnson, Jr., T. A. Houston / Tetrahedron Letters 43 (2002) 8905–8908
boronate)-1,4-dideoxy-1,4-imino-
D
-talitol HCl (8): 1H
3.53 (2H, m), 3.73 (1H, dd, J=6.3, 12.9), 3.88 (1H, dd,
J=4.1, 6.9 Hz), 3.99 (1H, app. q, J=3.9, 8.1 Hz), 5.18
(2H, m), 5.53 (1H, m); 13C NMR (75 MHz, pD 3.0 D2O):
l 55.37, 62.12, 62.73, 69.61, 69.74, 71.64, 72.90, 126.45,
126.89; HRMS (ESI+) m/z 204.1236 [(M+H+) calcd for
NMR (500 MHz, pD 2.0 D2O): l 0.82 (2H, t), 1.81 (2H,
m), 3.36 (1H, m), 3.39 (1H, m), 3.49 (1H, m), 3.56 (1H,
m), 3.70 (1H, m), 3.72 (1H, m), 3.80 (1H, m), 3.96 (1H,
m), 4.25 (1H, m), 4.41 (1H, m); 13C NMR (125 MHz, pD
2.0 D2O): l 11.4, 20.18, 50.34, 56.34, 63.34, 68.34, 70.00,
73.00, 73.34; 11B NMR (160 MHz, pD 2.0 D2O): l 18.65
(boric acid); HRMS (ESI+) m/z 250.1447 [(M+H+) calcd
C9H18NO4 204.1258]. N-Propyl-1,4-dideoxy-1,4-imino-D-
1
talitol HCl (9): H NMR (300 MHz, pD 3.0 D2O): l 0.53
(3H, t), 1.33 (2H, m), 2.98 (3H, m), 3.15 (1H, t, J=6.4
Hz), 3.34 (3H, m), 3.55 (1H, m), 3.84 (1H, dd, J=4.1 and
5.8 Hz), 3.99 (1H, app. q, J=4.1, 8.8 Hz). Like 9, the
allyl talitol derivative does not show significant inhibition
of b-galactosidase at 1.0 mM concentration [<10%].
for C9H21BNO6 250.1462]. N-Allyl-1,4-dideoxy-1,4-imino-
1
D
-talitol HCl: H NMR (300 MHz, pD 3.0 D2O): l 3.02
(1H, dd, J=3.3, 12.9 Hz), 3.20 (2H, m), 3.30 (1H, dd,
J=4.8, 12.3 Hz), 3.39 (1H, AB d, J=3.3 and 11.7 Hz),