Synthesis of an isostere of an O-linked glycopeptide

Article abstract of DOI:10.1021/ol0490779

(Chemical Equation Presented) A route for the synthesis of an electrophilic, carbocyclic galactose equivalent from D-galactose is described. The strategy utilizes ring-closing metathesis with Grubbs's second-generation catalyst as the key step. The galact

Full text of DOI:10.1021/ol0490779

ORGANIC
LETTERS
2004
Vol. 6, No. 19
3221-3224
Synthesis of an Isostere of an O-Linked
Glycopeptide
Lisa J. Whalen and Randall L. Halcomb*
Department of Chemistry and Biochemistry, UniVersity of Colorado,
Boulder, Colorado 80309-0215
halcomb@colorado.edu
Received May 19, 2004
ABSTRACT
A route for the synthesis of an electrophilic, carbocyclic galactose equivalent from D-galactose is described. The strategy utilizes ring-closing
metathesis with Grubbs’s second-generation catalyst as the key step. The galactose-derived electrophile reacted in an S 2 fashion with N-Boc-
N
cysteine methyl ester to provide an r-galactosylserine isostere. The method was extended to the synthesis of a glycopeptide isostere.
The attachment of carbohydrate chains to proteins is a
ubiquitous post- and cotranslational modification known as
glycosylation. Glycosylation is linked to several biological
processes such as protein transport, cell adhesion, and signal
transduction.1 The conformations of proteins can be altered
by the attachment of carbohydrate residues,2 which may be
important for protein folding. Aberrant glycosylation is
associated with several diseases, including cancerous and
inflammatory conditions.3 As a result of the role glycosy-
lation plays in these biologically relevant issues, there is a
considerable amount of interest in glycoproteins. This in turn
creates a requirement for glycoproteins with a defined
composition in order to correlate structure and function of
specific glycoprotein motifs. However, glycoproteins are
biosynthesized as complex mixtures of proteins with varying
degrees of glycosylation at specific sites, each component
of which is called a glycoform. The presence of mixtures of
glycoforms complicates isolation of homogeneous glyco-
proteins from natural sources. As a result, the development
of efficient methods for synthesizing homogeneous glyco-
proteins is a goal for organic chemists.
A common way in which oligosaccharides are attached
to proteins is through the hydroxyl groups of serine and
threonine residues, as in the R-N-acetylgalactosaminylserine
unit 1. Although solid-phase peptide synthesis (SPPS) can
be adapted to allow incorporation of glycosylated amino
acids,4 the synthesis of glycopeptides using SPPS is not
entirely routine due to size limitations and problems with
deglycosylation during peptide synthesis. However, one
method that has potential to become routine is the chemo-
selective ligation of a carbohydrate chain to a full-length
protein.
(1) (a) Kornfeld, R.; Kornfeld, S. Annu. ReV. Biochem. 1985, 54, 631.
(b) Rademacher, T. W.; Parekh, R. B.; Dwek, R. A. Annu. ReV. Biochem.
1988, 57, 785. (c) Varki, A. Glycobiology 1993, 3, 97. (d) Dwek, R. A.
Chem. ReV. 1996, 96, 683.
(2) (a) Rasmussen, J. R. Curr. Opin. Struct. Biol. 1992, 2, 682. (b)
Andreotti, A. H.; Kahne, D. J. Am. Chem. Soc. 1993, 115, 3352. (c) Liang,
R.; Andreotti, A. H.; Kahne, D. J. Am. Chem. Soc. 1995, 117, 10395. (d)
Imperiali, B.; Rickert, K. W. Proc. Natl. Acad. Sci. U.S.A. 1995, 92, 97.
(e) Imperiali, B.; O’Connor, S. E. J. Am. Chem. Soc. 1997, 119, 2295. (f)
Live, D. H.; Kumar, R. A.; Beebe, X.; Danishefsky, S. J. Proc. Natl. Acad.
Sci. U.S.A. 1996, 93, 12759. (g) Mehta, S.; Meldal, M.; Ferro, V.; Duus, J.
O.; Bock, K. J. Chem. Soc., Perkin Trans. 1 1997, 1365. (h) Simanek, E.
E.; Huang, D.-H.; Pasternack, L.; Machajewski, T. D.; Seitz, O.; Millar,
D.; Dyson, H. J.; Wong, C.-H. J. Am. Chem. Soc. 1998, 120, 11567.
(3) Hakomori, S.-I. AdV. Cancer Res. 1989, 52, 257.
Progress toward the development of chemoselective liga-
tion strategies for this purpose was reported by Bertozzi,5
Flitsch,6 Jones,7 Davis,8 and Schmidt.9 These methods all
10.1021/ol0490779 CCC: $27.50 © 2004 American Chemical Society
Published on Web 08/20/2004

use two functional groups that exclusively react with each
other under mild conditions to produce the neoglycoproteins.
Although some functional glycoproteins were synthesized
using some of these methods, many of the new linkages are
substantially different from those found in wild-type glyco-
proteins. To overcome this obstacle, we propose new
chemistry for the ligation of the carbohydrate and peptide
components using techniques that construct linkages more
similar to those found in wild-type glycoproteins (Scheme
1).
and other researchers.11 These methods do not provide access
to the cyclohexene scaffold of 3 with the correct substitution
pattern around the cyclohexene ring. Furthermore, methods
for the synthesis of highly substituted cyclohexenes such as
3 from non-carbohydrate starting materials often require the
use of microbial oxidation12 or produce mixtures of stereo-
isomers.13 As a result, the synthesis of isostere 2 began from
galactose-derived alcohol 6 (Scheme 2).14 Following a
Scheme 2. Synthesis of Galactose-Derived Cyclohexenols
9a/9b
Scheme 1. Synthesis Plan for R-Galactosylserine Isostere 5
literature procedure,15 oxidation under Swern conditions16
provided the ketone, which was then converted to vinyl
bromide 7 as a 6:1 Z:E mixture. Removal of the TBS ether
unveiled the alcohol, which underwent smooth conversion
to aldehyde 8 under Swern conditions. Upon treatment with
CrCl2 and NiCl2, intramolecular Nozaki-Hiyama-Kishi
cyclization17 occurred to provide cyclohexenols 9a and 9b
in a 6:1 9a:9b ratio. Cyclohexenols 9a and 9b were separated
by HPLC for characterization, and assignment of the relative
configuration of each was accomplished by 1H NMR
analysis.18
To surmount the difficulty in constructing an acetal, we
propose a chemoselective SN2 reaction between activated
cyclohexene 3 and a thiolate derived from cysteine. This
would generate unsaturated thioether 4, whose olefin would
be reduced to provide glycosyl amino acid isostere 5.10
Isostere 5 differs in two atoms from wild-type structure 1,
but these substitutions are not expected to alter the confor-
mation significantly. However, the stability of 5 to chemical
and enzymatic hydrolysis should be higher. We chose to
demonstrate the principle of this strategy with the synthesis
of R-galactosylserine isostere 2.
Early experiments showed that the steric bulk of the
multiple benzyl ether protecting groups affected the planned
olefin reduction, so a route was devised to replace them with
acetates (Scheme 3). Oxidation of 9a/9b with Dess-Martin
The conversion of carbohydrates to functionalized cyclo-
hexanes or cyclopentanes has been investigated by Ferrier
(4) Garg, H. G.; Jeanloz, R. W. AdV. Carbohydr. Chem. Biochem. 1985,
43, 135.
(5) (a) Rodriguez, E. C.; Winans, K. A.; King, D. S., Bertozzi, C. R. J.
Am. Chem. Soc. 1997, 119, 9905. (b) Rodriguez, E. C.; Marcaurelle, L. A.;
Bertozzi, C. R. J. Org. Chem. 1998, 63, 7134. (c) Marcaurelle, L. A.;
Bertozzi, C. R. J. Am. Chem. Soc. 2001, 123, 1587.
Scheme 3. First-Generation Synthesis of Cyclohexenol 12
(6) (a) MacMillan, D.; Bill, R. M.; Sage, K. A.; Fern, D.; Flitsch, S. L.
Chem. Biol. 2001, 8, 133. (b) MacMillan, D.; Daines, A. M.; Bayrhuber,
M.; Flitsch, S. L. Org. Lett. 2002, 4, 1467. (c) Davis, N. J.; Flitsch, S. L.
Tetrahedron Lett. 1991, 32 (46), 6793.
(7) (a) Davis, B. G.; Jones, J. B. Synlett 1999, 1495. (b) Davis, B. G.;
Lloyd, R. C.; Jones, J. B. Bioorg. Med. Chem. Lett. 2000, 8, 1527.
(8) (a) Gamblin, D. P.; Garnier, P.; van Kasteren, S.; Oldham, N. J.;
Fairbanks, A. J.; Davis, B. G. Angew. Chem., Int. Ed. 2004, 43, 828. (b)
Rendle, P. M.; Seger, A.; Rodrigues, J.; Oldham, N. J.; Bott, R. R.; Jones,
J. B.; Cowan, M. M.; Davis, B. G. J. Am. Chem. Soc. 2004, 126, 4750.
(9) (a) Zhu, X.; Pachamuthu, K.; Schmidt, R. R. J. Org. Chem. 2003,
68, 5641. (b) Zhu, X.; Pachamuthu, K.; Schmidt, R. R. Org. Lett. 2004, 6
(7), 1083.
(10) Reduction of the olefin to the saturated carbocycle followed by
mesylation and reaction with a thiolate nucleophile failed. The reaction of
a mesylate related to compound 3 (P ) Bn, X ) OMs) with N-acetylserine
methyl ester did not undergo the substitution reaction to produce an oxygen-
linked analogue.
periodinane19 produced perbenzylated cyclohexenone 10. The
benzyl ethers were removed using boron trichloride, and the
free hydroxyls were reprotected as acetates to yield per-
3222
Org. Lett., Vol. 6, No. 19, 2004

Scheme 4. Synthesis of Enone 11 Using Ring-Closing Metathesis
acetylated cyclohexenone 11. Reduction of the ketone under
to provide the cyclohexenols 19 (major isomer shown). The
mixture of diastereomers was oxidized to the enone 20 with
Dess-Martin periodinane. Removal of the PMB ethers with
trifluoroacetic acid followed by acetylation of the free
hydroxyls furnished enone 11.
With suitable amounts of 11 available, R-galactosylserine
isostere 2 was completed (Scheme 5). Conversion of 12 to
the allylic mesylate was accomplished cleanly with meth-
anesulfonyl chloride.
Luche conditions20 provided pseudoequatorial cyclohexenol
12 as a 30:1 ratio of diastereomers.
The first-generation synthesis of 12 possessed several low
yields, which led us to develop a second-generation synthesis
(Scheme 4) using a ring-closing metathesis inspired by the
work of Madsen21 and Lowary.22 Starting from p-tolyl 1-thio-
â-D-galactopyranoside 13,23 the free hydroxyls of 13 were
protected as p-methoxybenzyl ethers and the anomeric
position deprotected with N-bromosuccimide in wet acetone24
to provide pyranose 14. Sodium borohydride reduction of
14 furnished diol 15. Selective protection of the primary
alcohol of 15 as a TBS ether followed by Swern oxidation
of the secondary alcohol provided ketone 16. Wittig olefi-
nation of the ketone and deprotection of the TBS ether
occurred smoothly to give alcohol 17. After Swern oxidation
of 17 to the aldehyde, addition of vinylmagnesium bromide
yielded a 4:1 mixture of allylic alcohols 18 (major isomer
shown). Ring-closing metathesis of the diastereomeric
mixture of alcohols 18 was carried out with 10 mol %
Grubbs’s second-generation catalyst25 in refluxing toluene
Scheme 5. Synthesis of R-Galactosylserine Isostere 2
(11) (a) Ferrier, R. J.; Middleton, S. Chem. ReV. 1993, 93, 2779 and
references therein. (b) Martinez-Grau, A.; Marco-Contelles, J. Chem. Soc.
ReV. 1998, 27, 155 and references therein.
(12) Hudlicky, T.; Entwistle, D. A.; Pitzer, K. K.; Thorpe, A. J. Chem.
ReV. 1996, 96, 1195 and references therein.
(13) (a) Mehta, D.; Lakshminath, S. Tetrahedron Lett. 2000, 41, 3509.
(b) Ogawa, S.; Ara, M.; Kondoh, T.; Saitoh, M.; Masuda, R.; Toyokuni,
T.; Suami, T. Bull. Chem. Soc. Jpn. 1980, 53, 1121.
(14) Marshall, J. A.; Seletsky, B. M.; Luke, G. P. J. Org. Chem. 1994,
59, 3413.
(15) Lubineau, A.; Billault, I. J. Org. Chem. 1998, 63, 5668.
(16) Mancuso, A. J.; Huang, S.-L.; Swern, D. J. Org. Chem. 1978, 43,
2480.
(17) Fu¨rstner, A. Chem. ReV. 1999, 99, 991 and references therein.
(18) In 9a, Jvinyl proton,1 ) 2.5 Hz and J1,2 ) 7 Hz, suggesting a diaxial
relationship between H-1 and H-2. In 9b, both of these coupling constants
are 4 Hz, suggesting that there is an axial-equatorial relationship between
H-1 and H-2. In addition, NOE interactions between H-1 and H-3 were
observed for 9a and between H-1 and H-2 for 9b.
(19) Dess, D. B.; Martin, J. C. J. Org. Chem. 1983, 48, 4155.
(20) Luche, J.-L. J. Am. Chem. Soc. 1978, 100, 2226.
(21) Hyldtoft, L.; Madsen, R. J. Am. Chem. Soc. 2000, 122, 8444.
(22) (a) Callam, C. S.; Lowary, T. L. Org. Lett. 2000, 2, 167. (b) Callam,
C. S.; Lowary, T. L. J. Org. Chem. 2000, 66, 8961.
(23) Wilstermann, M.; Balogh, J.; Magnusson, G. J. Org. Chem. 1997,
62 (11), 3659.
Reaction of the crude mesylate with N-Boc cysteine methyl
ester occurred with complete inversion at the mesylate-
bearing stereocenter to provide axial allylic sulfide 21.
Reduction of 21 with 750 psi H2 in the presence of
Wilkinson’s catalyst gave isostere 2 as the major isomer.
Following purification by preparative thin-layer chromatog-
raphy, no other isomers could be detected. Under these
conditions, an equal amount of carbon-sulfur bond reduction
was observed.
The synthesis of R-galactosylserine isostere 2 demonstrated
the proof-of-principle for this technique. Extension to the
synthesis of a glycopeptide isostere broadened the scope of
the method and required a peptide possessing a free cysteine.
(24) Groneberg, R. D.; Miyazaki, T.; Stylianides, N. A.; Schulze, T. J.;
Stahl, W.; Schreiner, E. P.; Suzuki, T.; Iwabuchi, Y.; Smith, A. L.; Nicolaou,
K. C. J. Am. Chem. Soc. 1993, 115, 7593.
(25) Scholl, M.; Ding, S.; Lee, C. W.; Grubbs, R. H. Org. Lett. 1999, 1
(6), 953.
Org. Lett., Vol. 6, No. 19, 2004
3223

The tripeptide N-Boc-Gly-Cys(SH)-Phe-OEt was chosen as
a representative example (synthesis in Supporting Informa-
tion). Following mesylation of 12, reaction of N-Boc-Gly-
Cys(SH)-Phe-OEt with the crude mesylate occurred with
complete inversion of configuration at the mesylate-bearing
carbon to provide axial allylic thioether 22 (Scheme 6).
Table 1. 1H NMR Analysis of R-Galactosylserine Isostere 2,
Glycopeptide Isostere 23, and Wild-Type R-Galactosylserine
Scheme 6. Synthesis of Glycopeptide Isostere 23
Nozaki-Hiyama-Kishi cyclization, the second using a ring-
closing metathesis. An allylic pseudoequatorial alcohol at
C-1 of this carbocycle was converted to an electrophilic
mesylate that reacted with the free thiols of cysteine and a
cysteine-containing peptide in an SN2 fashion to yield allylic
thioethers. Hydrogenation of the trisubstituted alkenes under
high pressure in the presence of Wilkinson’s catalyst
provided glycosyl amino acid and glycopeptide isosteres.
Reduction of 22 with 500 psi H2 in the presence of
Wilkinson’s catalyst proceeded with an equal amount of
carbon-sulfur bond reduction. Flash chromatography of the
reaction mixture provided 23 as a single isomer.
1H NMR analysis of 2 provided strong evidence that the
Acknowledgment. This research was supported by an
NSF Career Award (CHE-9733765). R.L.H. also thanks the
Camille and Henry Dreyfus Foundation (Camille Dreyfus
Teacher-Scholar Award), Pfizer (Junior Faculty Award), and
Novartis (Young Investigator Award) for support. This work
was facilitated by a 500 MHz spectrometer that was
purchased partly with funds from an NSF Shared Instru-
mentation Grant (CHE-0131003). L.J.W. thanks John Scott
(Bristol-Myers Squibb) for helpful suggestions, Richard
Shoemaker (University of Colorado) for assistance in
performing NMR studies, and the National Science Founda-
tion, the American Chemical Society Division of Medicinal
Chemistry, and Bristol-Myers Squibb for predoctoral fel-
lowships.
4
cyclohexane ring possessed a C1-like chair conformation.
Using automated difference SPT-NMR spectroscopy,26 sev-
eral coupling constants of the cyclohexane ring protons of 2
were obtained for comparison with the acetylated R-galac-
tosylserine (GalSer, Table 1). For 23, the same data was
obtained using a double-quantum-filtered COSY experiment.
Comparison of J1,2, J2,3, and J3,4 for 2, 23, and GalSer
suggests that the conformations of the carbocycles in each
compound are similar. However, this is a comparison of only
three coupling constants. Furthermore, as 2, 23, and GalSer
still possess protecting groups, this does not accurately reflect
their conformations in aqueous solution of the corresponding
deacylated compounds, which may be different. The large
J5,6ax values for 2 and 23 confirmed the stereochemistry at
C-5 resulting from the olefin reductions.
Supporting Information Available: Experimental details
and full characterization of all new compounds; 1H and 13
C
In summary, two routes to a carbocyclic galactose
equivalent were developed, one using an intramolecular
NMR spectra of selected compounds. This material is
available free of charge via the Internet at http://pubs.acs.org.
(26) Takacs, J. M.; Chandramouli, S. V.; Shoemaker, R. S. Tetrahedron
Lett. 1994, 35 (49), 9161.
OL0490779
3224
Org. Lett., Vol. 6, No. 19, 2004