N-Acetylhexosaminidase inhibitory properties of C-1 homologated GlcNAc- and GalNAc-thiazolines

Article abstract of DOI:10.1016/j.bmcl.2008.03.067

Several C-1 homologated GlcNAc- and GalNAc-thiazolines, as well as a related GalNAc-thiazole, have been prepared. The compounds are analogues of GlcNAc-thiazoline, a potent transition-state-mimicking inhibitor of retaining β-N-acetylglycosaminidases. Kinetic evaluation of these fused pyranose-heterocycles against the bacterial N-acetylhexosaminidase SpHex suggests active site steric restrictions around the substrate anomeric carbon.

Full text of DOI:10.1016/j.bmcl.2008.03.067

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry Letters 18 (2008) 2944–2947
N-Acetylhexosaminidase inhibitory properties of C-1 homologated
GlcNAc- and GalNAc-thiazolines
Benjamin Amorelli,a Chunhua Yang,a Brian Rempel,b
Stephen G. Withersb and Spencer Knappa,*
aDepartment of Chemistry & Chemical Biology, Rutgers, The State University of New Jersey,
610 Taylor Road, Piscataway, NJ 08854, USA
bDepartment of Chemistry, University of British Columbia, Vancouver, BC, Canada V6T 1Z1
Received 28 February 2008; accepted 24 March 2008
Available online 28 March 2008
Abstract—Several C-1 homologated GlcNAc- and GalNAc-thiazolines, as well as a related GalNAc-thiazole, have been prepared.
The compounds are analogues of GlcNAc-thiazoline, a potent transition-state-mimicking inhibitor of retaining b-N-acetylglycos-
aminidases. Kinetic evaluation of these fused pyranose-heterocycles against the bacterial N-acetylhexosaminidase SpHex suggests
active site steric restrictions around the substrate anomeric carbon.
Ó 2008 Elsevier Ltd. All rights reserved.
Glycosidases mediate a variety of important cellular
functions. As an example, retaining N-acetyl-b-hexos-
aminidases of families 18 and 20 are a class of glycosyl
hydrolases that are responsible for the removal of 2-
acetamido-2-deoxy-D-glucopyranosyl and -galactopyr-
anosyl residues from glycoconjugates, and that utilize
a mechanism featuring substrate-assisted catalysis.1 An
inherited abnormality in the human glycosidases Hex
A and Hex B results in the accumulation of certain
gangliosides, causing the neurodegenerative disorders
Tay-Sachs and Sandoff’s diseases, respectively.2 The
hydrolysis of the terminal GlcNAc/GalNAc residue
from the aglycon polysaccharide features an oxazolini-
um intermediate held noncovalently in the active site
(Fig. 1).
osaminidase A, respectively).3,4 A schematic of H-bond-
ing interactions of SpHex with 1a indicated by the
crystallographic analysis is shown in Figure 2. A hydro-
phobic pocket surrounds the thiazoline ring of 1a, pro-
tecting the iminium carbon from water. Within this
pocket, Tyr393 and Asp313 residues can position the
2-acetamido group of the substrate for nucleophilic par-
ticipation at the anomeric center and diffuse the positive
charge distributed into the oxazolinium ring as it forms.
The normal role of the general acid residue Glu314 (cir-
The remarkable binding properties of GlcNAc- and
GalNAc-thiazolines 1 and their resistance to enzymatic
hydrolysis has facilitated the determination of the co-
crystal structure of 1a intact and bound in the active site
of the N-acetylhexosaminidase from Streptomyces plica-
tus (SpHex; Ki for 1a = 20 lM and 270 nM vs SpHex
and human b-hexosaminidase A, respectively; Ki for
1b = 100 lM and 820 nM vs SpHex and human b-hex-
Keywords: Glycosidase; O-GlcNAcase; Streptomyces plicatus; Pyra-
nose conformation; Transition state mimic.
*
Figure 1. Mechanism for retaining N-acetylhexosaminidases that use
substrate assistance.
Corresponding author. Tel.: +1 732 445 2627; e-mail: spencer.knapp
@rutgers.edu
0960-894X/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2008.03.067

B. Amorelli et al. / Bioorg. Med. Chem. Lett. 18 (2008) 2944–2947
2945
gesting that 1a is a true transition state mimic for
OGA and presumably also for the mechanistically re-
lated hexosaminidases.9
Unlike OGA, the hexosaminidases do not tolerate sub-
strate or inhibitor amide groups much larger than acet-
amido in the binding pocket.3,9,10 Whether groups at C-
1 (see methylated analogue 2) would be tolerated has not
yet been established. Crystallographic studies of the
piperidine carbohydrate analogue GalNAc-isofagomine
(Ki = 21 lM vs SpHex)11 bound in the active site of
SpHex show a key noncovalent interaction between
the equatorial piperidine N–H and the general acid res-
idue Glu314, an interaction absent3 in the case of the
SpHexÆ1a complex. A thiazoline related to 1, but which
might additionally possess a group capable of hydrogen
bonding or other interaction with Glu314, ought to be
an even more effective inhibitor. Therefore, we also set
out to enable this additional enzyme(Glu314)–inhibitor
interaction through the preparation of functionally elab-
orated thiazolines such as 3. The evaluation of 2 and 3
as a means to probe glycosidase tolerance of substitu-
tions (shown in boxes) for –H at the anomeric center
of parent inhibitors GlcNAc- and GalNac-thiazolines
1 is the subject of this report. The route to 2 was de-
scribed earlier,12 and an unusual fragmentation leading
to the fused GalNAc-thiazole 11 was also reported.13
Figure 2. Schematic showing interactions between 1a and the active
site of SpHex.
cled) is to assist binding and departure of the aglycon
oxygen from C-1. However, Glu314 does not interact
with 1a, as the latter lacks a hydrogen bond donor in
this position.
Human O-GlcNAcase (OGA) is a critical enzyme
for post-translational modification of intracellular
proteins.5 OGA is a family 84 glycoside hydrolase
responsible for the cleavage of GlcNAc from b-O-Glc-
NAc-modified serine and threonine residues. The afore-
mentioned polysaccharide glycosidases exhibit little
sequence similarity to OGA; however, GlcNAc-thiazo-
line 1a has also been shown to inhibit OGA
(Ki = 70 nM),4 implicating this enzyme as a hydrolase
with the 2-acetamido substrate-assisted catalytic mecha-
nism as well.6
Lactone 4
oxycarbonylmethylenetriphenylphosphorane,
(Scheme 1),13 treated with tert-but-
reacted
slowly under reflux, but at 140 °C in a sealed reaction
tube gave a mixture of thiazoline ester 5 and the chain
extended vinylogous carbonate 6 in 55% combined iso-
1
lated yield (respective ratio 2.6:1). The H NMR signal
for the thiazoline CH3 of 5 at 2.17 ppm (br s, long
range coupled to H-2)14 and the downfield thioaceta-
mide CH3 singlet (2.47 ppm) of 6 are indicative of their
structures. Deprotection of the mixture of 5 and 6 with
trifluoroacetic acid in dichloromethane gave 7 in 86%
yield (Scheme 1). Compounds 5 and 6 were both con-
verted to 7 in this reaction, as the amount of either
alone is insufficient to account for the isolated yield.
Some hydrolysis of 7 to the mercapto amide under
these conditions might be expected.14,15 However, with
exclusion of moisture, 7 was stable to the reaction con-
ditions, concentration without workup, and spectro-
scopic characterization in chloroform solution.
Installation of nitrogen was carried out by first convert-
ing 7 to the acyl azide. Attempts to convert 7 directly
by using diphenylphosphoryl azide (DPPA)16 were
unsuccessful. The conversion of 7 to the mixed anhy-
dride was followed by addition of aqueous sodium
azide. After concentration of the organic layer, the acyl
azide was verified by infrared spectroscopic analysis, as
this intermediate is not stable to chromatography on
silica gel. The product shows a strong stretch at
2135 cmꢀ1 due to the azide, but there was also partial
rearrangement to isocyanate 8 (2257 cmꢀ1). Heating
the crude azide for 2 h in toluene solution completed
the rearrangement to 8. The isocyanate was also unsta-
Past efforts towards glycosidase inhibitor design have
included the synthesis and evaluation of inhibitors that
superficially resemble posited electrostatic and confor-
mational changes within the transition state, including
a flattened substrate (4H3) pyranose ring and a much
debated glycosyl cation-like structure (sp2 hybridized
and partially positively charged at C-1). Free energy
correlation studies7 have recently demonstrated, how-
ever, that strong adventitious binding by many glycosi-
dase inhibitors does not in itself indicate transition state
mimicry.8 Vocadlo and coworkers have homologated
the acetamido group of 1a, and shown that this family
of inhibitors [sp3 hybridized at C-1 (pyranose number-
ing), 4C1 pyranose pseudo-chair conformation when
bound in the active site] displays excellent correlation
of free energy of activation calculated from enzymatic
substrate hydrolysis relative to inhibitor binding, sug-
1
ble to silica gel, although H and 13C NMR spectra of
the crude concentrated reaction mixture provide evi-
dence for an efficient and reliable conversion to 8.

2946
B. Amorelli et al. / Bioorg. Med. Chem. Lett. 18 (2008) 2944–2947
Figure 3. Unit cell of crystalline GalNAc-thiazoline 1b, showing the
4C1 pseudo-chair pyranose conformation.
NAc-thiazolines bind in hexosaminidase active sites as
pyranose pseudo-chairs,3,4,9 which conformations most
closely match the substrate-derived transition state (see
Fig. 1). Thus, more conformational change upon en-
zyme binding is required of 1a compared with 1b, which
is already in a chair.
Inhibition studies3 of the new C-1 homologated thiazo-
lines (2, 10, and 3) and thiazole 11 against SpHex reveal
greatly reduced affinities relative to the parent thiazo-
lines 1a and 1b (Chart 1). The relatively small C-1
methyl group of 2 reduces binding by a factor of 10 rel-
Scheme 1. Synthesis of elaborated GaINAc-thiazolines 10 and 3.
þ
ative to 1a, but larger groups (i.e., –CH2NH3 and
–CH2NHBoc) reduce the binding affinity by factors of
at least 100-fold. Thus, the kinetic studies on the amino-
methyl thiazoline 3 show no evidence of a binding inter-
action with the general acid Glu314. The decrease in
inhibition might be due to changes in inhibitor size,
shape, conformation, or, in the case of 3, a possibly
Trapping 8 with tert-BuOH to form the stable carba-
mate was extremely slow, giving 10 in only 28% yield
after 48 h.
Thiazoline 9 was deprotected with n-Bu4NF to carba-
mate triol 10. After treatment of 10 with TFA in dichlo-
romethane solution, aminomethyl thiazoline triol 3 was
isolated as the trifluoroacetate salt. TFA salt 3 is some-
what unstable in acidic methanol; therefore,3 was char-
1
acterized by H NMR in dry CD3CN: the diagnostic
ABq (3.42 and 3.57 ppm, J = 14 Hz each) for the
þ
–CH2NH3 unit, the broad singlet at 2.21 ppm for the
thiazoline CH3, and the electrospray mass spectrum
(m/z 249 MH+) confirm the structure.
The parent GalNAc-thiazoline 1b adopts a pseudo-chair
4C1 conformation in D2O solution according to 1H
NMR analysis, with J2,3 = 7.8 Hz.14 In the single crystal,
4
1b also exists as a C1 chair conformation, as shown in
Figure 3. Likewise, the homologated analogue N-Boc-
aminomethyl GalNAc-thiazoline 10 has a chair-like
pyranose conformation in CD3OD solution, according
1
to its H NMR spectrum (J2,3 = 8.4 Hz). In contrast,
the GlcNAc-thiazoline 1a exists primarily in
pseudo-boat conformation in CD3OD solution
a
Chart 1. Inhibition studies of the new C-1 homologated thiazolines (2,
and 10 and 3 and thiazole 11 against SpHeX.
(J2,3 = 4.0 Hz).10 Both GlcNAc- and (presumably) Gal-

B. Amorelli et al. / Bioorg. Med. Chem. Lett. 18 (2008) 2944–2947
2947
T. J.; Khan, A. R.; Cao, Z.; Robbins, P. W.; James, M. N.
G.; Triggs-Raine, B. L. J. Biol. Chem. 1998, 273, 19618;
(d) Drouillard, S.; Armand, S.; Davies, G. J.; Vorgias, C.
E.; Henrissat, B. Biochem. J. 1997, 328, 945.
unfavorable polar interaction. In particular, the portion
of the enzyme active site near the anomeric carbon and
the scissile bond, which must normally accommodate
the aglycon upon initial binding of the substrate, evi-
dently cannot accommodate bulky substituents at the
anomeric carbon when introduced in the context of
the GalNAc-thiazoline inhibitors. This is presumably a
consequence of the tightness of the enzyme binding
pocket surrounding the pseudo-equatorial C-1 b-substi-
tuent, which is H in the case of the best inhibitors 1a and
1b.
2. Gravel, R. A.; Clark, J. T. R.; Kaback, M. M.; Mahuran,
D.; Sandhoff, K.; Suzuki, K. In The Metabolic Basis of
Inherited Disease; Scriver, C. R., Beaudet, A. L., Sly, W.
S., Valle, D., Eds.; McGraw-Hill: New York, 1995; p
1807.
3. Mark, B. L.; Vocadlo, D. J.; Knapp, S.; Triggs-Raine, B.
L.; Withers, S. G.; James, M. N. G. J. Biol. Chem. 2001,
276, 10330.
4. (a) Macauley, M. S.; Whitworth, G. E.; Debowski, A. W.;
Chin, D.; Vocadlo, D. J. J. Biol. Chem. 2005, 280, 25313;
(b) Tropak, M. B.; Blanchard, J. E.; Withers, S. G.;
Brown, E. D.; Mahuran, D. Chem. Biol. 2007, 14, 153.
5. (a) Torres, C. R.; Hart, G. W. J. Biol. Chem. 1984, 259,
3308; (b) Wells, L.; Vosseller, K.; Hart, G. W. Science
2001, 291, 2376; (c) Hanover, J. A. FASEB J. 2001, 15,
1865; Bacterial OGA (d) Dorfmueller, H. C.; Borodkin, V.
S.; Schimpl, M.; Shepherd, S. M.; Shpiro, N. A.; van
Aalten, D. M. F. J. Am. Chem. Soc. 2006, 128, 16484.
6. (a) Macauley, M. S.; Whitworth, G. E.; Debowski, A. W.;
Chin, D.; Vocadlo, D. J. J. Biol. Chem. 2005, 280, 25313;
(b) Cetinbas, N.; Macauley, M. S.; Stubbs, K. A.;
Drapala, R.; Vocadlo, D. J. Biochemistry 2006, 45, 3835.
7. (a) Lienhard, G. E. Science 1973, 180, 149; (b) Wolfenden,
R. Nature 1969, 223, 704; (c) Mader, M. M.; Bartlett, P. A.
Chem. Rev. 1997, 97, 1281.
Acknowledgements
We thank NIH (AI055760) for partial financial support,
and Rutgers University for a Reid graduate fellowship
to B.A. The work at UBC was supported by the Cana-
dian Institutes for Health Research, the Natural Science
and Engineering Research Council, and the University
of British Columbia. We thank Dr. Brian Patrick and
David Vocadlo at UBC for determining the 3-dimen-
sional structure of 1b.
Supplementary data
8. Wicki, J.; Williams, S. J.; Withers, S. G. J. Am. Chem. Soc.
2007, 129, 4530.
Experimental procedures and copies of spectra for new
compounds. Crystallographic data (excluding structure
factors) for the structures in this paper have been depos-
ited with the Cambridge Crystallographic Data Centre
as Supplementary Publication No. CCDC 678775. Sup-
plementary data associated with this article can be
found, in the online version, at doi:10.1016/
j.bmcl.2008.03.067.
9. Whitworth, G. E.; Macauley, M. S.; Stubbs, K. A.;
Dennis, R. J.; Taylor, E. J.; Davies, G. J.; Greig, I. R.;
Vocadlo, D. J. J. Am. Chem. Soc. 2007, 129, 635.
10. Knapp, S.; Abdo, M.; Ajayi, K.; Huhn, R. A.; Emge, T. J.;
Kim, E. J.; Hanover, J. A. Org. Lett. 2007, 9, 2321.
11. Mark, B. L.; Vocadlo, D. J.; Zhao, D.; Knapp, S.;
Withers, S. G.; James, M. N. G. J. Biol. Chem. 2001, 276,
42131.
12. Knapp, S.; Yang, C.; Haimowitz, T. Tetrahedron Lett.
2002, 43, 7101.
References and notes
13. Knapp, S.; Amorelli, B.; Doss, G. A. Tetrahedron Lett.
2004, 45, 8507.
14. Knapp, S.; Myers, D. S. J. Org. Chem. 2002, 67, 2995.
15. Knapp, S.; Myers, D. S. J. Org. Chem. 2001, 66, 3636.
16. Shioiri, T.; Ninomiya, K.; Yamada, S. J. Am. Chem. Soc.
1972, 94, 6203.
1. (a) Tews, I.; Perrakis, A.; Oppenheim, A.; Dauter, Z.;
Wilson, K. S.; Vorgias, C. E. Nat. Struct. Biol. 1996, 3,
638; (b) Tews, I.; Terwisscha van Scheltinga, A. C.;
Perrakis, A.; Wilson, K. S.; Dijkstra, B. W. J. Am. Chem.
Soc. 1997, 119, 7954; (c) Mark, B. L.; Wasney, G. A.; Salo,