Novel O-glycosyl amino acid mimetics as building blocks for O-glycopeptides act as inhibitors of galactosidases

Article abstract of DOI:10.1016/j.bmcl.2003.10.011

As potential lead structures for a new class of glycosidase inhibitors the novel O-glycosyl amino acid mimetics 3'-O-[2,6-anhydro-D-glycero-L-gluco-heptitol-1-yl]-L-serine 3 and-L-threonine 4 were synthesized, employing regio- and stereoselective aziridine ring opening methodology. They proved to be stable in the presence of glycosidases and showed competitive inhibition of α-galactosidase from Aspergillus niger.

Full text of DOI:10.1016/j.bmcl.2003.10.011

Bioorganic & Medicinal Chemistry Letters 14 (2004) 73–75
Novel O-glycosyl amino acid mimetics as building blocks for
O-glycopeptides act as inhibitors of galactosidases
Lars Kroger, Dirk Henkensmeier, Andreas Schafer and Joachim Thiem*
Institute of Organic Chemistry, University of Hamburg, Martin-Luther-King-Platz 6, Hamburg, Germany
Received 11 August 2003; revised 8 October 2003; accepted 8 October 2003
Abstract—As potential lead structures for a new class of glycosidase inhibitors the novel O-glycosyl amino acid mimetics 30-O-[2,6-
anhydro-d-glycero-l-gluco-heptitol-1-yl]-l-serine 3 and-l-threonine 4 were synthesized, employing regio- and stereoselective azir-
idine ring opening methodology. They proved to be stable in the presence of glycosidases and showed competitive inhibition of a-
galactosidase from Aspergillus niger.
# 2003 Elsevier Ltd. All rights reserved.
The important role of carbohydrates in many biological
processes is evident, and tailored derivatives are dis-
cussed as promising leads for efficient pharmaceuticals.
However, the lability of the glycosidic bond towards
chemical and enzymatic degradation in vivo results in
low bioavailability of carbohydrate derivatives and pre-
vents their oral application.1 The synthesis of glyco
mimetics, representing the corresponding natural struc-
tures and offering higher metabolic stability, has already
received much attention, and further effort on their
synthesis are expected to be undertaken in the future.2 A
common approach is the application of C-glycosides,
which are no longer susceptible to cleavage by
glycosidases.3ꢀ6
it into a model peptide, inert against in vivo glycosidic
cleavage.13 In another approach connected with N-gly-
coproteins, we synthesized a N-glucoasparagine mimetic
(2), in which the asparagine has been shifted from the
anomeric center to position 2 of the carbohydrate.14 In
both these molecules the glycosidic acetal has changed
into a stable ether bond (Fig. 1).
A novel approach for the stereoselective synthesis of
b-alkoxy-a-amino acids was introduced by Nakajima
and Okawa via ring opening of chiral activated azir-
idine-2-carboxylates with alcohols in the presence of
Lewis acids.15,16 A regioselective attack by nucleophiles
such as alcohols, amines, carboxylic acids, thiols and
indols occurs only at position 3.17 Due to the moderate
nucleophilicity of alcohols, only aziridine ring openings
employing simple alcohols like methanol or ethanol are
reported to date using the alcohol as solvent as well.
Hence, we developed an optimized reaction protocol,
enabling us to utilize the secondary hydroxyl group of a
rather sterically demanding carbohydrate derivative as
nucleophile.18
Due to their therapeutic potential, the synthesis of gly-
copeptides and their mimetics represents a dynamically
developing field of research.7ꢀ9 Significant improve-
ments in activity, bioavailability, stability and solubility
of peptides were achieved by their glycosylation.10,11
The glycosidic linkage is susceptible to enzymatic
hydrolysis, but furthermore b-eliminations occur under
mild basic conditions in the case of O-glycosides of ser-
ine and threonine.12 Hence, glycopeptide mimetics can
be considered promising leads to novel pharmaceuticals.
Bednarski et al. synthesized the b-Gal-O-serine mimetic
1, stable against enzymatic hydrolysis, and incorporated
* Corresponding author. Tel.: +49-40-42838-4241; fax: +49-40-
42838-4325; e-mail: thiem@chemie.uni-hamburg.de
Figure 1. C-glycosidic b-Gal-O-Ser 1 and N-glucoasparagine 2 mimetic.
0960-894X/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2003.10.011

74
L. Kroger et al. / Bioorg. Med. Chem. Lett. 14 (2004) 73–75
¨
0
.
Scheme 1. Synthesis of 3 -O-[2,6-anhydro-d-glycero-l-gluco-heptitol-1-yl]-l-serine 3 and -l-threonine 4: (a) C3H3–Si(CH3)3, BF3 Et2O, MeCN,
.
48%; (b) O3, DCM; (c) NaBH4, THF, 62%; (d) BF3 Et2O, CHCl3, 52% (R=H), 55% (R=CH3); (e) H2, Pd/C, MeOH, 89% (R=H), 95%
(R=CH3); (f) (1) NaOMe, MeOH; (2) Amberlite IR-120(H +), 59% (R=H), 92% (R=CH3).
To further evaluate the potential of this method for the
synthesis of O-glycosyl amino acid mimetics, we pre-
pared the serine and threonine derivatives 3 and 4,
which can not be hydrolyzed. They have a C-glycosidic
linkage and a stable ether-bridge between the carbohy-
drate and the amino acid part, resulting in a linkage
elongated by one methylene group (Scheme 1).
The new C-glycosidic building block 5 was synthesized
by reacting 1,2,3,4,6-penta-O-acetyl-b-d-galactopyr-
anose (6) with propargyl trimethyl silane in the presence
of a Lewis acid. The expected19 a-linkage of the C-gly-
1
cosidic allene 7 was confirmed by H NMR, showing a
3J1,2-coupling of 4.1 Hz. Cleavage of the allene with
ozone gave the labile aldehyde 8, which was reduced
without isolation to the C-glycosidic methylene galactitol
5 employing sodium borohydride.20
Figure 2. Lineweaver–Burk plot. Hydrolysis of pNP-a-Gal by a-
galactosidase from A. niger with different concentrations of substrates
3 and 4 (A: substrate concentration 50 mmol; B: substrate concen-
tration 10 mmol). The common point of intersection indicates a com-
petitive inhibition.
Compound 5 was subjected to the Lewis acid catalyzed
ring opening of the activated aziridines 9 and 10, which
gave the glycosyl-amino acid building blocks 11 an 12 in
satisfactory yields of 52 and 55%, respectively.21
Hydrogenolytic deprotection furnished the inter-
mediates 13 and 14, which were used, after Fmoc-pro-
tection, as versatile building blocks for glycopeptide
solid phase synthesis of MUC1 mimetics. These results
will be published in due course. Zemplen deacetylation
of 13 and 14 gave the target compounds 322 and 4.23
attempted for these structures, which may be considered as
promising leads for a new class of glycosidase inhibitors.
To test the in vivo stability of the carbohydrate amino
acid mimetics 3 and 4, they were incubated with a-
galactosidase from Aspergillus niger24 and showed no
hydrolysis. As these mimetics have structures closely
related to the natural substrates of a-galactosidases,
they should interact with the active center without being
hydrolyzed by the enzyme. Thus, they might function as
competitive glycosidase inhibitors. This hypothesis was
tested by inhibition experiments of the a-galactosidase
catalyzed cleavage of para-nitrophenyl a-d-galactopyr-
anoside (Fig. 2). Comparison with the well known inhi-
bitor 1-deoxynojirimycin (Ki=18 mmol)25 showed
considerably lower results, but nevertheless a noticeable
competitive inhibition was observed for 3 (Ki=0.4
mmol) and 4 (Ki=0.3 mmol). No optimization was
References and notes
1. Wong, C.-H. J. Chin. Chem. Soc. 1999, 46, 271.
2. Chapleur, Y. Carboyhdrate Mimics; Wiley-VCH: Wein-
heim, 1998.
3. Levy, D. E.; Tang, C. The Chemistry of C-Glycosides;
Pergamon: Oxford, 1995.
4. Wittmann, V.; Kessler, H. Angew. Chem., Int. Ed. Engl.
1993, 105, 1138.
5. Westermann, B.; Walter, A.; Florke, U.; Altenbach, H.-J.
Org. Lett. 2001, 3, 1375.
6. Schafer, A.; Thiem, J. J. Org. Chem. 2000, 65, 24.
7. Dondoni, A.; Marra, A.; Massi, A. Tetrahedron 1998, 54,
2827.
8. Debenham, S. D.; Debenham, J. S.; Burk, M. J.; Toone,
E. J. J. Am. Chem. Soc. 1997, 119, 9897.

L. Kroger et al. / Bioorg. Med. Chem. Lett. 14 (2004) 73–75
¨
75
9. Lay, L.; Meldal, M.; Nicotra, F.; Panza, L.; Russo, G. J.
Chem. Soc., Chem. Commun. 1997, 1469.
10. Fisher, J. F.; Harrison, A. W.; Bundy, G. L.; Wilkinson,
K. F.; Bush, B. D.; Ruwart, M. J. J. Med. Chem. 1991, 34,
3140.
11. Bundy, G. L.; Pals, D. T.; Lawson, J. A.; Couch, S. J.;
Lipton, M. F.; Mauragis, M. A. J. Med. Chem. 1990, 33,
2276.
12. Marshall, R. D.; Neuberger, A. Adv. Carbohydr. Chem.
Biochem. 1970, 25, 407.
13. Bertozzi, C. R.; Hoeprich, P. D., Jr.; Bednarski, M. D. J.
Org. Chem. 1992, 57, 6092.
14. Schafer, A.; Klich, G.; Schreiber, M.; Paulsen, H.; Thiem,
J. Carbohydr. Res. 1998, 313, 107.
15. Nakajima, K.; Takai, F.; Tanaka, T.; Okawa, K. Bull.
Chem. Soc. Jpn. 1978, 51, 1577.
16. Larsson, U.; Carlson, R. Acta Chem. Scand. 1994, 48,
511.
17. Dauban, P.; Dubois, L.; Dau, M. E. T. H.; Dodd, R. J.
Org. Chem. 1995, 60, 2035.
component 5 (1 equiv) (2S)-benzyl-1-benzyloxycarbonyl-
aziridin-2-carboxylate 9 or (2S,3S)-benzyl-1-benzyloxy-
carbonyl-3-methylaziridin-2-carboxylate 10 (0.66 equiv),
respectively, were dissolved in dry chloroform. The solu-
tion was degassed and kept under high vacuum for 30
min. After flushing the flask with argon, the resulting
syrup was redissolved in the minimal amount of dry
chloroform and treated with a catalytic amount of 10%
.
BF3 Et2O in chloroform. The reaction was left at room
temperature for 16 h, treated with a catalytic amount of
.
BF3 Et2O again and left for further 16 h. After diluting
with chloroform, saturated sodium hydrogencarbonate
was added and the organic phase dried over magnesium
sulfate. Evaporation of the solvent followed by column
chromatography gave the desired products.
22. Compound 3: colorless solid; mp 134 ꢂC; [a]D20=+24.3 (c
1
0.5, H2O); H NMR (500 MHz, D2O): d=3.54 (dd, 1H,
3
2
0
0
H-7, J6,7=4.4 Hz, J7,7 =11.8 Hz), 3.58 (dd, 1H, H-7 ,
3
0
J6,7 =7.7 Hz), 3.58–3.64 (m, 2H, H-1, H-4), 3.74–3.83
(m, 6H, H-3, H-5, H-6, H-Ser-a, CH2-Ser-b), 3.87 (dd,
1H, H-10, J1,1 =10.2 Hz, J1 ,2=6.2 Hz), 4.10(ddd, 1H,
2
3
0
0
18. Henkensmeier, D.; Kroger, L.; Schafer A.; Thiem, J.,
manuscript in preparation.
3
H-2, J2,3=3.5 Hz) ppm.
23. Compound 4: colorless solid; mp 148 ꢂC; [a]D20=+21.4 (c
19. Deslongchamps, P. Stereoelectronic Effects in Organic
Chemistry; Pergamon: New York, 1983.
0.5, H2O); 1H NMR (400 MHz, D2O): d=1.18 (d, 1H,
20. Compound 5: colorless oil; [a]2D0=+31.3 (c 0.5, CH2Cl2);
CH3-Thr, Jb-Thr,CH3-Thr=6.5 Hz), 3.52 (d, 1H, H-Thr-a,
3
1H NMR (400 MHz, CDCl3): d=2.00, 2.01, 2.03, 2.05
3Ja-Thr, b-Thr=4.6 Hz), 3.54–3.70(m, 5H, C H2-1, H-4,
2
3
3
0
(4ꢁs, 12H, 4ꢁCH3), 3.60(dd, 1H, H-1, J1,1 =11.7 Hz,
CH2-7), 3.74 (ddd, 1H, H-6, J5,6=1.8 Hz, J6,7=4.6 Hz,
3
3J1,2=7.6 Hz), 3.77 (dd, 1H, H-10), 4.03 (dd, 1H, H-7,
3
0
J6,7 =6.5 Hz), 3.82 (dd, 1H, H-5, J4,5=3.4 Hz), 3.86
3
3
3
J6,7=4.1 Hz, 2J7,7 =11.7 Hz), 4.19–4.24 (m, 2H, H-2, H-
(dd, 1H, H-3, J2,3=6.2 Hz, J3,4=9.7 Hz), 3.93 (dd, 1H,
3
0
6), 4.35 (dd, 1H, H-70, J6,7 =8.1 Hz), 5.20(dd, 1H, H-3,
H-Thr-b), 4.06 (ddd, 1H, H-2, J1,2=3.9 Hz, J1 ,2=8.9
3
3
0
0
3J2,3=4.1 Hz, 3J3,4=8.1 Hz), 5.24 (dd, 1H, H-4, 3J4,5=3.1
Hz), 5.38 (dd, 1H, H-5, 3J5,6=3.6 Hz) ppm. Anal. calcd for
C15H22O10: C, 49.42, H, 6.12. Found: C, 49.10, H, 6.25.
21. General procedure for the aziridine ring opening: alcohol
Hz) ppm.
24. Bahl, O. P.; Agrawal, K. M. L. J. Biol. Chem. 1969, 244,
2970.
25. Dale, M. P. Biochemistry 1985, 24, 3530.