Distinctive inhibition of O-GlcNAcase isoforms by an α-GlcNAc thiolsulfonate

Article abstract of DOI:10.1021/ja076038u

O-GlcNAcase (OGA) promotes O-GlcNAc removal, and thereby plays a key role in O-GlcNAc metabolism, a feature of a variety of vital cellular processes. Two splice transcripts of human OGA encode "long OGA", which contains a distinct N-terminal O-GlcNAcase d

Full text of DOI:10.1021/ja076038u

Published on Web 11/10/2007
Distinctive Inhibition of O-GlcNAcase Isoforms by an r-GlcNAc Thiolsulfonate
Eun J. Kim,† Benjamin Amorelli,‡ Mohannad Abdo,‡ Craig J. Thomas,§ Dona C. Love,†
Spencer Knapp,*,‡ and John A. Hanover*,†
Laboratory of Cell Biochemistry and Biology, NIDDK, National Institute of Health, Bethesda, Maryland 20892,
Department of Chemistry & Chemical Biology, Rutgers, The State UniVersity of New Jersey,
Piscataway, New Jersey 08854, and NIH Chemical Genomics Center, NHGRI, National Institute of Health,
Bethesda, Maryland 20892
Received August 21, 2007; E-mail: jah@helix.nih.gov
Dynamic modification of cytoplasmic and nuclear proteins by
O-linked 2-acetamido-2-deoxy-â-D-glucopyranosyl (O-GlcNAc)
residues is an important signaling mechanism that shares charac-
teristics with phosphorylation.1 “O-GlcNAc-ylation” functions in
diverse cellular processes including nutrient sensing,2 modifying
protein target activity,3 and controlling gene expression4 and protein
degradation.5 Unusual patterns of O-GlcNAc-ylation have been
linked to insulin resistance6 and to several neurodegenerative
diseases.7 O-GlcNAcase (OGA) promotes O-GlcNAc removal and
thus plays a key role in O-GlcNAc metabolism. Human OGA exists
as two splice variants:8 the long isoform (hlOGA) gene consists of
16 exons and encodes a protein that has a N-terminus domain
homologous to hyaluronidase and a C-terminal region considered
to be a histone acetyltransferase (HAT) domain;9 the short isoform
(hsOGA) is encoded by an alternatively spliced transcript consisting
of only 10 exons and part of intron 10 that contains an alternative
stop codon. Short OGA therefore lacks the C-terminal HAT domain
and has 15 amino acids at its C-terminus different from those in
long OGA. Interestingly, a single nucleotide polymorphism located
in the intron 10 of the gene for OGA has been associated with
type II diabetes in Mexican Americans.10 We have recently
demonstrated that the short OGA hydrolyzes the â-O-linked
glycosidic bond of GlcNAc from glycoproteins in Vitro.11 A number
of mechanistic studies of long OGA and crystallographic studies
of homologous bacterial enzymes have demonstrated that the
O-GlcNAcase active site resides in the N-terminal domain, and that
catalysis requires aspartic acids at 174, 175, and 177 as key catalytic
residues that enable a substrate-assisted mechanism.12-14 Given that
this domain is also found in short OGA, it is surprising that short
OGA is comparatively resistant to previously described potent
inhibitors of long OGA and lysosomal hexosaminidases including
PUGNAc and NAG-thiazoline (see Supporting Information, S-1).
We previously measured the kinetic parameters of both OGA
Figure 1. Schematic protein structures of long and short OGA (A) and
enzymes by using the fluorogenic FDGlcNAc substrate 1, and found
that they behave differently as O-GlcNAcase catalysts.11 We report
here the differential inhibition characteristics of a new inhibitor,
R-GlcNAc thiolsulfonate 2, toward the OGA isoforms and the
effects of pseudo-glycosylation of long OGA by 2 on the properties
of the enzyme.
Previously described hexosaminidase inhibitors are much less
potent against the short OGA isoform than the longer, HAT domain
containing enzyme (see S-1). However, we have found that
R-GlcNAc thiolsulfonate 2 (see Supporting Information for its
synthesis) is a more effective inhibitor of short OGA than previous
potent inhibitors of hexosaminidases (see S-2). Respective inhibition
chemical structures of FDGlcNAc (1) and GlcNAc R-thiosulfonate (2) and
Lineweaver-Burk analyses of short (B) and long OGA (C) inhibition in
the presence of increasing concentrations of 2.
by 2 of long OGA and lysosomal hexosaminidase is about 4-fold
and 9-fold less than that of short OGA, indicating some degree of
selectivity. A Lineweaver-Burk plot of the initial velocity of
FDGlcNAc hydrolysis by both OGA isoforms at varied concentra-
tions of 2 reveals the mode of enzyme inhibition (Figure 1). The
inhibitor, thiolsulfonate 2, acts as a purely competitive inhibitor of
short OGA, with Ki ) 10 µM, indicating that inhibitor (2) and
FDGlcNAc substrate (1) are mutually exclusive in the active site
of the enzyme. The presence of 2 changes both the apparent Km
and the apparent Vmax, indicating a mixed-type inhibition for long
OGA. Therefore, 2 can bind competitively with the substrate not
only at the N-terminal active site but also at another site apart from
the active site of OGA.
† Laboratory of Cell Biochemistry and Biology, NIDDK, National Institute of
Health.
‡ Rutgers, The State University of New Jersey.
§ NIH Chemical Genomics Center, NHGRI, National Institute of Health.
9
14854
J. AM. CHEM. SOC. 2007, 129, 14854-14855
10.1021/ja076038u CCC: $37.00 © 2007 American Chemical Society

C O M M U N I C A T I O N S
that the amount of effective enzyme decreases as the time of
preincubation of long OGA with 2 increases (S-6). Furthermore,
thiol-GlcNAc-modified long OGA is completely insensitive to
cleavage by caspase 3. Long OGA is known to be a substrate for
caspase 3, and its cleavage sites are speculated to reside in the
N-terminal and middle domains.18 Complete caspase 3 resistance
of modified long OGA suggests a dramatic change in the conforma-
tion or steric properties of the latter, rendering the cleavage sites
inaccessible (S-5B). It is presently unclear whether the modification
of long OGA with S-GlcNAc occurs while the inhibitor is bound
to the active site or occurs independent of the active site binding
event; however, our results suggest that a change in the conforma-
tion of enzyme upon remote modification drastically alters the active
site architecture, resulting in a complete blockage of substrate
access. The covalent attachment of an R-S-GlcNAc unit to long
OGA accounts for the irreversible and time-dependent irreversible
inhibition of catalysis.
How might the R-linked GlcNAc thiolsulfonate 2 competitively
inhibit an enzyme that specifically processes â-linked substrates?
1
One hypothesis is that 2 assumes the S3 twist boat conformation
of the bound substrate. In this conformation, the sulfonyl oxygens
might be in position to act as H-bond acceptors19 from the general
acid residue and/or other donor residues in the active site (see S-7
for illustration).
Figure 2. Western blot analyses of long OGA (left) and short OGA (right)
treated with (+) and without (-) 2. Blot was stained by Ponceu S. solution
(upper) and probed with anti-His (middle panels) and WGA-biotin/
Streptavidin (lower panels). DTT treatment of the modified long OGA
reversed its modification (DTT+).
Acknowledgment. We thank Dr. Peter McPhie (NIH) for
helpful discussions on the enzyme inhibition kinetics. We are
thankful to NIH (Grant AI055760) for partial financial support and
to Rutgers University for a Reid graduate fellowship to BA and
the Intramural Research Program of the National Institute of
Diabetes and Digestive and Kidney Diseases, NIH.
The calculated specific inhibition constant (Ki) and catalytic
inhibition constant (Ki′) of 2 with respect to long OGA are 35 and
280 µM, respectively. Thus, the mixed inhibition of long OGA
results from a combination of competitive inhibition with a Ki
similar to that observed for the short isoform and a second
nonspecific component. The distinctive kinetic behavior toward
O-GlcNAcase catalysis of the two OGA isoforms implies that the
local environment of the enzyme active site differs. This difference
likely results from the presence of the C-terminal HAT domain in
long OGA and its absence in short OGA.
Supporting Information Available: Preparation of 2 and details
of the enzyme inhibition assays. This material is available free of charge
via the Internet at http://pubs.acs.org.
References
(1) Love, D. C.; Hanover, J. A. Sci. STKE 2005, 312, 1-14.
(2) Hanover, J. A. FASEB J. 2001, 15, 1865-1876.
(3) Du, X. L.; Edelstein, D.; Dimmeler, S.; Ju, Q.; Sui, C.; Brownlee, M. J.
Clin. InVest. 2001, 108, 1341-1348.
When long OGA was incubated with a saturating solution of 2
at 37 °C, a modification of the enzyme occurred: Western blot
analysis using the O-GlcNAc-specific CTD110.6 antibody (data
not shown) and wheat germ agglutinin lectin (WGA) confirmed
this to be a covalent attachment of GlcNAc (Figure 2, left panel).
Comparable modification did not occur in short OGA (Figure 2,
right panel). Thiolsulfonate modification has been reported to occur
at cysteine residues through the formation of disulfide linkages.15,16
Glycosyl thiolsulfonates have been used by Davis and co-workers
as reagents for protein pseudo-glycosylation at cysteine.15,16 The
chemical feasibility of disulfide bond formation with 2 and cysteine
residues was modeled by conjugating 2 with an N,C-protected
cysteine (S-3). We also find that treatment of the modified long
OGA with dithiothreitol reverses the modification (Figure 2, left
panel), suggesting the presence of a reducible R-GlcNAc-protein
disulfide bond. Long OGA is known to have two free cysteines
(positions 166 and 87817) that are possible modification sites.
Indeed, alkylation of free sulfhydryls in cysteine residues of long
OGA with N-ethylmaleimide completely blocks modification by 2
(S-4). On the basis of the enzyme inhibition studies, the presumed
GlcNAc disulfide modification by 2 likely occurs in the C-terminal
HAT domain. This region features a free cysteine at 878 that is
absent in short OGA.
(4) Yang, X.; Zhang, F.; Kudlow, J. E. Cell 2002, 110, 69-80.
(5) Zhang, F.; Su, K.; Yang, X.; Bowe, D. B.; Paterson, A. J.; Kudlow, J. E.
Cell 2003, 115, 715-725.
(6) McClain, D. A.; Crook, E. D. Diabetes 1997, 45, 1003-1009.
(7) Griffith, F. I.; Schmitz, B. Biochem. Biophys. Res. Commun. 1995, 213,
424-431.
(8) Comtesse, N.; Maldener, E.; Meese, E. Biochem. Biophys. Res. Commun.
2001, 283, 634-640.
(9) Toleman, C.; Paterson, A. J.; Whisenhunt, T. R.; Kudlow, J. E. J. Biol.
Chem. 2004, 279, 53665-53673.
(10) Lehman, D. M.; Fu, D.-J.; Freeman, A. B.; Hunt, K. J.; Leach, R. J.;
Johnson-Pais, T.; Hamlington, J.; Dyer, T. D.; Arya, R.; Abboud, H.;
Go¨ring, H. H. H.; Duggirala, R.; Blangero, J.; Konrad, R. J.; Stern, M. P.
Diabetes 2005, 54, 1214-1221.
(11) Kim, E. J.; Kang, D. O.; Love, D. C.; Hanover, J. A. Carbohydr. Res.
2006, 341, 971-982.
(12) Toleman, C.; Paterson, A. J.; Kudlow, J. E. Biochim. Biophys. Acta 2006,
1760, 829-839.
(13) 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, 635-644.
(14) Rao, F. V.; Dorfmueller, H. C.; Villa, F.; Allwood, M.; Eggleston, I. M.;
van Aalten, D. M. F. EMBO J. 2006, 25, 1569-1578.
(15) Davis, B. G.; Maughan, M. A. T.; Green, M. P.; Ullman, A.; Jones, J. B.
Tetrahedron: Asymmetry 2000, 11, 245-262.
(16) Gamblin, D. P.; Garnier, P.; Ward, S. J.; Oldham, N. J.; Fairbanks, A. J.;
Davis, B. G. Org. Biomol. Chem. 2003, 1, 3642-3644.
(17) Toleman, C. A.; Paterson, A. J.; Kudlow, J. E. J. Biol. Chem. 2006, 281,
3918-3925.
(18) Wells, L.; Gao, Y.; Mahoney, J. A.; Voseller, K.; Chen, C.; Rosen, A.;
Hart, G. W. J. Biol. Chem. 2002, 277, 1755-1761.
(19) Sivapriya, K.; Hariharaputran, S.; Suhas, V. L.; Chandra, N.; Chan-
drasekaran, S. Bioorg. Med. Chem. 2007, 15, 5659-5665.JA076038U
Strikingly, modification of OGA with thiol-linked GlcNAc
abolishes almost all of its enzymatic activity (S-5A). We also note
JA076038U
9
J. AM. CHEM. SOC. VOL. 129, NO. 48, 2007 14855