Tautomeric modification of GIcNAc-thiazoline

Article abstract of DOI:10.1021/ol0706814

The potent OGIcNAcase (OGA) inhibitor GIcNAc-thiazoline has been modified by buffer- or acylation-induced imine-to-enamine conversion and then electrophile or radical addition (X_n = D₃, F, N ₃, OH, SMe, COCF₃, CF₃). Several functionalized GIcNAc-thiazolines show highly selective inhibition of OGA vs human hexosaminidase and thus have promise as tools for targeted investigations of OGA, an enzyme linked to diabetes and neurodegeneration. A new radical addition/fragmentation reaction of the N-(trifluoroacetyl)enamine has been discovered.

Full text of DOI:10.1021/ol0706814

ORGANIC
LETTERS
2007
Vol. 9, No. 12
2321-2324
Tautomeric Modification of
GlcNAc-Thiazoline
Spencer Knapp,*,† Mohannad Abdo,† Kehinde Ajayi,† Richard A. Huhn,†
Thomas J. Emge,† Eun Ju Kim,‡ and John A. Hanover*,‡
Department of Chemistry and Chemical Biology, Rutgers-The State UniVersity of
New Jersey, 610 Taylor Road, Piscataway, New Jersey 08854, and Laboratory of Cell
Biochemistry and Biology, National Institute of Diabetes and DigestiVe and Kidney
Diseases, National Institutes of Health, Bethesda, Maryland 20892
spencer.knapp@rutgers.edu
Received March 20, 2007
ABSTRACT
The potent O-GlcNAcase (OGA) inhibitor GlcNAc-thiazoline has been modified by buffer- or acylation-induced imine-to-enamine conversion
and then electrophile or radical addition (Xn D3, F, N3, OH, SMe, COCF3, CF3). Several functionalized GlcNAc-thiazolines show highly selective
)
inhibition of OGA vs human hexosaminidase and thus have promise as tools for targeted investigations of OGA, an enzyme linked to diabetes
and neurodegeneration. A new radical addition/fragmentation reaction of the N-(trifluoroacetyl)enamine has been discovered.
A wide variety of nuclear and cytoplasmic proteins are
modified on serine and threonine residues by the dynamic
addition and removal of â-O-GlcNAc units.1-3 These diverse
targets mediate important biological processes that may in
turn be regulated by the â-O-GlcNAc cycling. O-GlcNAc
addition and removal are catalyzed, respectively, by O-
GlcNac transferase (OGT) and O-GlcNAcase (OGA, Figure
1), the study of which has acquired urgency as the importance
of the O-GlcNAc modification to processes such as cellular
signaling and regulation4 and to disease states such as type
II diabetes,5 cancer,6 and Alzheimer’s7 has become clear. To
help sort out the mechanisms and effects of protein “O-
GlcNAc-ylation,” a significant effort has been directed
toward the development of inhibitors of OGA that do not
simultaneously inhibit the mechanistically related human
N-acetylhexosaminidases HexA and HexB. In comparison
† Rutgers-The State University.
‡ National Institutes of Health.
(1) Torres, C. R.; Hart, G. W. J. Biol. Chem. 1984, 259, 3308-3317.
(2) Wells, L.; Vosseller, K.; Hart, G. W. Science 2001, 291, 2376-2378.
(3) Zachara, N. E.; Hart, G. W. Chem. ReV. 2002, 102, 431-438.
(4) Slawson, C.; Housley, M. P.; Hart, G. W. J. Cell. Biochem. 2006,
97, 71-83.
(5) McClain, D. A.; Lubas, W. A.; Cooksey, R. C.; Hazel, M.; Parker,
G. J.; Love, D. C.; Hanover, J. A. Proc. Natl. Acad. Sci. U.S.A. 2002, 99,
10695-10699.
Figure 1. Hydrolysis of a serine/threonine linked â-O-GlcNAc
catalyzed by OGA, the OGA intermediate GlcNAc-oxazolinium
ion 1, and GlcNAc-thiazoline inhibitors 2-6.
(6) Shaw, P.; Freeman, J.; Bovey, R.; Iggo, R. Oncogene 1996, 12, 921-
930.
(7) Arnold, C. S.; Johnson, G. V. W.; Cole, R. N.; Dong, D. L.-Y.; Lee,
M.; Hart, G. W. J. Biol. Chem. 1996, 271, 28741-28744.
10.1021/ol0706814 CCC: $37.00
© 2007 American Chemical Society
Published on Web 05/18/2007

to the latter, OGA tolerates inhibitor groups larger than
acetamido -CH3, or its equivalent, in the acetamido binding
pocket and thus provides an avenue for potential optimization
of inhibitor selectivity, efficacy, solubility, transport, and
metabolic stability. GlcNAc-thiazoline8 (2, Figure 1) is a
The GlcNAc-thiazoline triacetate 8 is available on a
multigram scale by treatment of commercial glucosamine
pentaacetate 7 with P4S10 (Scheme 1).17 Whereas 2 binds in
Scheme 1. Tautomeric Deuteration of GlcNAc-Thiazoline
Figure 2. ORTEP view of one cation of the salt 8‚HO3SAr (Ar )
2,4-dinitrophenyl), showing the OS2 pyranose conformation.
nanomolar but nonselective inhibitor of OGA, HexA,9 and
HexB by virtue of its resemblence to the transition state
leading to the enzyme intermediate, oxazolinium ion 1.10 By
increasing the size of the thiazoline ring substituent from
methyl to ethyl, propyl, and isopropyl, Vocadlo and co-
workers were able to increase the selectivity for inhibition
of OGA (see 3-5).11,12 Analogous steric-based selectivity
improvements to other OGA inhibitors have also been
realized.13-15 On the other hand, inhibitors that possess
functionalized acetamido mimics could provide an expanded
array of options for biochemical and medicinal chemistry
studies.16 We have now prepared a new series of methyl-
modified GlcNAc-thiazolines 6 by exploiting the previously
unrecognized propensity of GlcNAc-thiazolines to undergo
buffer- and acylation-induced imine-to-enamine conversion.
enzyme active sites in an apparent pseudo-chair (4C1)
pyranose conformation,18,19 GlcNAc-thiazoline triacetates
such as 8 exist principally in a twist boat (OS2) in CDCl3
solution.20 In the solid state, the 2,4-dinitrobenzenesulfonic
acid salt of 8 also exhibits the S2 pyranose conformation
(Figure 2). The corresponding thiazoline triols (e.g., 2) are
closer to a boat (O,4B) in CD3OD solution.21
The methyl protons of 8 exchange with deuterium in
certain solvents in the presence of acid. As this reaction could
also be used to prepare tritiated 2, the deuteration was
optimized as follows. Treatment of 8 with 2.4 equiv of
pyridine, 1.2 equiv of triflic acid, and 100 equiv of D2O in
acetonitrile solution for 8 h at 23 °C and then extractive
workup gave the trideuterated GlcNAc-thiazoline 10. No
C-deuteration was detected in the absence of the buffer
components pyridine and triflic acid. Standard deacetylation
then led to the corresponding triol 11 without significant loss
of deuterium, according to integration of the methyl signal
in the 1H NMR spectrum. Alternatively, triol 2 was directly
trideuterated by treatment with the same buffer system, and
11 was separated from the buffer components by partitioning
between 1-butanol and saturated aqueous sodium bicarbonate
(93% yield, 95% D3). In polar solvents at acidic pH, 8
O
(8) Systematic name: (3aR,5R,6S,7R,7aR)-6,7-dihydroxy-5-hydroxy-
methyl-2-methyl-5,6,7,7a-tetrahydro-3aH-pyrano[3,2-d]thiazole.
(9) GlcNAc-thiazoline 2 has also been identified as a chemical chaperone
that may prevent misfolding of a HexA mutant associated with Tay-Sachs
disease. Tropak, M. B.; Reid, S. P.; Guiral, M.; Withers, S. G.; Mahuran,
D. J. Biol. Chem. 2004, 279, 13478-13487.
(10) 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. Knapp, S.; Vocadlo, D.; Gao, Z.; Kirk, B., Lou, J.;
Withers, S. G. J. Am. Chem. Soc. 1996, 118, 6804-6805.
(11) Macouley, M. S.; Whitworth, G. E.; Debowski, A. W.; Chen, D.;
Vocadlo, D. J. J. Biol. Chem. 2005, 280, 25313-25322.
(12) Stubbs, K. A.; Zhang, N.; Vocadlo, D. J. Org. Biomol. Chem. 2006,
4, 839-845.
(13) 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,
16484-16485.
(17) Knapp, S.; Kuhn, R. A.; Amorelli, B. Org. Synth. 2007, 84, 68-
76.
(18) Dennis, R. J; Taylor. E. J.; Macauley, M. S.; Stubbs, K. A.;
Turckenburg, J. P.; Hart, S. J.; Black, G. N.; Vocadlo, D. J.; Davies, G. J.
Nat. Struct. Mol. Biol. 2006, 13, 365-371.
(14) Kim, E. J.; Perreira, M.; Thomas, C. J.; Hanover, J. A. J. Am. Chem.
Soc. 2006, 128, 4234-4235.
(15) Additional recent OGA inhibitors: Shanmugasundaram, B.; De-
bowski, A. W.; Dennis, R. J.; Davies, G. J.; Vocadlo, D. J.; Vasella, A.
Chem. Commun. 2006, 4372-4374. Tolman, C.; Paterson, A. J.; Shin, R.;
Kudlow, J. E. Biochem. Biophys. Res. Commun. 2006, 340, 526-534. Lee,
T. N.; Alborn, W. E.; Knierman, M. D.; Konrad, R. J. Biochem. Biophys.
Res. Commun. 2006, 350, 1038-1043. Rao, F. V.; Dorfmueller, H. C.;
Villa, F.; Allwood, M.; Eggleston, I. M.; van Aalten, D. M. F. EMBO J.
2006, 25, 1569-1578. Perreira, M.; Kim, E. J.; Thomas, C. J.; Hanover, J.
A. Bioorg. Med. Chem. 2006, 14, 837-846.
(19) Lemieux, M. J.; Mark, B. L.; Cherney, M. M.; Withers, S. G.;
Mahuran, D. J.; James, M. N. G. J. Mol. Biol. 2006, 359, 913-929. Mark,
B. L.; Mahuran, D. J.; Cherney, M. M.; Zhao, D.; Knapp, S.; James, M. N.
G. J. Mol. Biol. 2003, 327, 1093-1109. Mark, B. L.; Vocadlo, D. J.; Knapp,
S.; Triggs-Raine, B. L.; Withers, S. G.; James, M. N. G. J. Biol. Chem.
2001, 276, 10330-10337.
(20) Vicinal proton coupling constants for 8: J1,2 ) 7, J2,3 ) 3, J3,4
)
1.5, and J4,5 ) 9 Hz. See also: Foces-Foces, C.; Cano, F. H.; Bernabe, M.;
Penades, S.; Martin-Lomas, M. Carbohydr. Res. 1984, 135, 1-11.
(21) Coupling constants for 2: J1,2 ) 7, J2,3 ) J3,4 ) 4, and J4,5 ) 9 Hz.
(16) Ritter, T. K.; Wong, C.-H. Tetrahedron Lett. 2001, 42, 615-618.
2322
Org. Lett., Vol. 9, No. 12, 2007

hydrolyzes to the acetamido mercaptan, and thioconjugates
can then be prepared by various S-alkylation and arylation
reactions.22 Hydrolysis of 10 led analogously to the tri-
deuteroacetamido mercaptan 12; however, ∼5% of the
deuterium was lost in the process. The transformations in
Scheme 1 are consistent with acid-promoted tautomerization
of 8 to give the enamine 9; reprotonation leads to sequential
replacement of all three methyl H’s.
by azido led to 18 and, following deacetylation, to 19.
Replacement of iodo with acetoxy and S-acetylthio was also
successful, and the resulting thiazolines 20 and 22 were
deacetylated (the latter in the presence of iodomethane) to
afford 21 and 23, respectively.
The ease of tautomerization of 8 suggested that an N-acyl-
enamine might also be accessible (Scheme 3).24 Reaction of
Would other electrophiles react with 9? Treatment of 8
with the same buffer but in the presence of 3.2 equiv of
N-bromosuccinimide gave the tribromide 14 (Figure 3). The
Scheme 3. Thiazoline Acylations
Figure 3. Tautomeric halogenation of GlcNAc-thiazoline.
dibromide 13 could be obtained (along with 14) by reducing
the amount of NBS to 2.2 equiv. The monobromide could
not be prepared selectively, evidently because the second
and third brominations are faster than the first. Fluorination,
however, could be effectively stopped after one substitu-
tion: exposure of 8 to buffer and 1.5 equiv of Selectfluor23
gave 15 in high yield. Standard deacetylation led to the fluoro
thiazoline triol 16.
8 with 1 equiv of TFAA indeed gave the enamine 24. In
methanol solution, 24 reverted to 8, and when treated with
methoxide, 24 gave 2. Upon acylation of 8 with 2.1 equiv
of TFAA, the C-acylated product 25 formed in good yield,
presumably through 24 as an intermediate. Deacetylation
gave keto triol 26, as confirmed by peracetylation to 27. Both
24 and 27 exist as pseudo-chair conformers in solution,
according to the vicinal proton J values.
During attempts to brominate 24, we discovered a new
trifluoromethylation reaction. Treatment of 24 with benzoyl
peroxide and a low power UV light source gave the 2,2,2-
trifluoroethyl thiazoline 28 (Scheme 4). This product
Iodination of 8 could also be stopped after a monosub-
stitution (Scheme 2). The product 17 proved to be unstable
Scheme 2. Tautomeric Iodination and Displacement Reactions
Scheme 4. Radical Trifluoromethylation
to storage but could be isolated, characterized, and subse-
quently treated with nucleophiles. Thus, substitution of iodo
shows the diagnostic five-bond coupling25 between the
thiazoline methylene H’s and the pyranose H-2. 19F NMR
(22) Knapp, S.; Myers, D. S. J. Org. Chem. 2001, 66, 3636-3638.
(23) Lal, G. S.; Pez, G. P.; Syvret, R. G. Chem. ReV. 1996, 96, 1737-
1755.
(24) Grobovsky, L. V.; Schmir, G. L. Tetrahedron 1971, 27, 1185-
1197.
(25) Knapp, S.; Myers, D. S. J. Org. Chem. 2002, 67, 2995-2999.
Org. Lett., Vol. 9, No. 12, 2007
2323

analysis (δ -63.9, t, J ) 10.3 Hz) supports this structure,
as do the spectra of the deacetylated product 29.
A radical chain mechanism (Scheme 5) accounts for the
formation of 28. The initiating radical can add to the CdC
than loss of CF3CO•, accounting for the formation of product
still bearing this group. Deacetylation of 30 gave triol 31,
but as a 1:4 mixture of isomers (respective H-1’s at 5.77
and 5.98 ppm). The change in isomeric composition upon
basic hydrolysis29 reflects thiazolidine ring opening to an
imine mercaptide intermediate, which recloses to give the
thermodynamic mixture of isomers of 31.
Scheme 5. Proposed Mechanism for Radical
Figure 4 shows the selective inhibition of human re-
combinant O-GlcNAcase by the new modified GlcNAc-
Trifluoromethylation
of 24, leading to a tertiary thiazolidine radical I, fragmenta-
tion of which would give the trifluoroacetyl radical and a
thiazoline product. The trifluoroacetyl radical likely frag-
ments further to provide carbon monoxide and the tri-
fluoromethyl radical.26 Addition of trifluoromethyl radical
to 24 again leads to a thiazolidinyl radical (II), and then the
chain is propagated by another fragmentation, giving 28 as
well as more trifluoromethyl radical.27
Figure 4. Inhibition by modified GlcNAc-thiazolines of O-
GlcNAcase in comparison to human placental â-hexosaminidase.
thiazolines, relative to their inhibition of human placental
â-hexosaminidase.30 While all seven new compounds show
somewhat reduced inhibition relative to the parent 2, the
azide 19 and the fluoride 16 exhibit excellent selectivity for
the O-GlcNAcase, and 19 retains nearly all of the inhibitory
ativity of 2. These two highly selective and potent GlcNAc-
thiazolines differ significantly from previously characterized
selective O-GlcNAcase inhibitors. The fluorine and azide
derivatives may prove useful for developing reagents for
imaging, labeling, and interfering with O-GlcNAc cycling
in living cells and tissues.
This mechanism is supported by the successful trapping
of the thiazolidine radical by methyl mercaptan (Scheme 6).
Scheme 6. Radical Addition of CH3SH
Acknowledgment. The authors are grateful to NIH
(AI055760) for financial support. K.A. thanks the GAANN
program for a graduate fellowship. The work at NIH was
supported by the Intramural Research Program of the
National Institute of Diabetes and Digestive and Kidney
Diseases.
Exposure of 24 to the radical initiating conditions, but in
the presence of excess mercaptan, led to the formation of
thioether adduct 30 as an inseparable 9:1 mixture of
stereoisomers (respective anomeric H’s at 5.94 and 6.21
ppm). Kinetic hydrogen atom abstraction likely occurs
preferentially from the less hindered â face28 and is faster
Supporting Information Available: Experimental details
and spectroscopic characterization for new compounds. This
material is available free of charge via the Internet at
http://pubs.acs.org.
(26) Maricq, M. M.; Szente, J. J.; Khitrov, G. A.; Dibble, T. S.; Francisco,
J. S. J. Phys. Chem. 1995, 99, 11875-11882.
OL0706814
(27) The reaction of s-collidine with triflic anhydride might be reinter-
preted in light of our observations. Binkley, R. W.; Ambrose, M. G. J.
Org. Chem. 1983, 48, 1777-1779.
(29) Bogna´r, R.; Gyo¨rgydea´k, Z.; Szila´gyi, L.; Sa´ndor, P.; Radics, L.
Liebigs Ann. Chem. 1979, 701-707.
(30) Assay procedure: Kim, E. J.; Kang, D. O.; Love, D. C.; Hanover,
J. A. Carbohydr. Res. 2006, 341, 971-982.
(28) Davies, D. I.; Parfitt, L. T.; Alden, C. K.; Claisse, J. A. J. Chem.
Soc. C 1969, 1585-1590.
2324
Org. Lett., Vol. 9, No. 12, 2007