Ready protease-catalyzed synthesis of carbohydrate-amino acid conjugates.

Article abstract of DOI:10.1039/b104137c

The protease-catalyzed synthesis of amino acid est-carbohydrate conjugates as glycopeptide analogues has been achieved in a highly regioselective and carbohydrate-specific manner using amino acid vinyl ester acyl donors and minimally or completely unprotected carbohydrate acyl acceptors, which together probed active sites of proteases to reveal yield efficiencies that are modulated by the carbohydrate C-2 substitutent, and that may be exploited to allow selective one-pot syntheses.

Full text of DOI:10.1039/b104137c

View Article Online / Journal Homepage / Table of Contents for this issue
Ready protease-catalyzed synthesis of carbohydrate–amino acid
conjugates†
Viviane Boyer,a Mincho Stanchev,b Antony J. Fairbanks*b and Benjamin G. Davis*a
a Department of Chemistry, Univeristy of Durham, South Road, Durham, UK DH1 3LE.
E-mail: Ben.Davis@durham.ac.uk
b Dyson-Perrins Laboratory, South Parks Road, Oxford, UK OX1 3QY
Received (in Cambridge, UK) 10th May 2001, Accepted 6th July 2001
First published as an Advance Article on the web 12th September 2001
The protease-catalyzed synthesis of amino acid ester–
carbohydrate conjugates as glycopeptide analogues has been
achieved in a highly regioselective and carbohydrate-specific
manner using amino acid vinyl ester acyl donors and
minimally or completely unprotected carbohydrate acyl
acceptors, which together probed active sites of proteases to
reveal yield efficiencies that are modulated by the carbohy-
Scheme 1 For reagents and conditions, see ESI.
drate C-2 substituent, and that may be exploited to allow
selective one-pot syntheses.
With these acyl donors in hand, we investigated their utility
in transesterification reactions with a representative range of
carbohydrate acyl acceptors 9a–20a (Scheme 2, Table 1). After
exploring a range of conditions, the use of SBL lyophilised from
phosphate buffer (pH 8.0) in anhydrous pyridine at 45 °C
proved optimal. Initial variation of parent carbohydrate in the
completely deprotected series 9a–12a revealed exclusive O-6
Many carbohydrate–peptide conjugates display a wide variety
of potent biological activities of potential therapeutic and
commercial value.1,2 For example, glycoproteins act as critical
cell surface communication markers,3 glycopeptide motifs such
as the Thomsen–Friedenreich (Tf) antigen are associated with
cancer cell lines4 and an oligomeric sequence of the glycopep-
tide motif (AAT[Galb(1,3)GalNAca])n displays unusual non-
colligative antifreeze properties.5 Access to well-defined carbo-
hydrate–peptide conjugates and their analogues to probe the
nature of these properties is essential. A large number of elegant
methods have been developed for the synthesis and assembly of
N- and O-linked glycopeptides2,6 but these methods may be
complicated by low glycosylation efficiencies and extensive
protection regimes to ensure regioselectivity. To avoid these
potential problems we have investigated the utility of enzyme-
catalyzed regioselective acylation of carbohydrates as a one-
step method. Despite the ready construction of ester–carbohy-
drate linkages, there have been remarkably few syntheses of
amino acid esters of carbohydrates.7 Furthermore, although the
utility of hydrolases as powerful catalysts for regioselective
acylation of carbohydrates is well demonstrated,8 their use in
the transfer of amino acids to carbohydrates is, with very few
exceptions,9 neglected. This is all the more surprising given that
several biofunctional molecules, such as enkephalin–carbohy-
drate conjugates that modulate fibroblast and melanoma
growth,10 are themselves a-amino esters of carbohydrates.
Moreover, carbohydrate–peptide conjugates connected by po-
tentially metabolisable, sacrificial linkages, such as esters, have
high potential utility as prodrugs in which the glycan moiety
affords both protection and specific transport properties.11 We
therefore set ourselves the goal of establishing a ready, short
route for the creation of such ester-linked glycopeptides.
Initially, we chose the serine protease subtilisin Bacillus
lentus (SBL, EC 3.4.21.14) as a powerful catalyst for ester
synthesis12 and the representative amino acids phenylalanine
1a, aspartic acid 2a and glutamic acid 3a. Amino acid
derivatives (Scheme 1)† were chosen to probe not only the
amino acid specificity of SBL but also its tolerance for a variety
of amine (Ac, Boc, Fmoc, Z) and ester (Bn) protecting groups.
Pd(OAc)2-mediated transesterification13 of 1–8b with vinyl
acetate (Scheme 1) allowed the preparation of the correspond-
ing Phe, Asp and Glu; a and side-chain vinyl esters 1–8c as
acyl donors that render transesterifications essentially
irreversible.14
regioselectivity but only low isolated yields of either
9b or
-galactose 10b 6-O-phenylalaninate esters.15 However, a
higher yield of the 6-O-phenylalaninate ester of -mannose 11b
D
-glucose
D
D
indicated an exciting preference based only on the ster-
eochemistry of the parent carbohydrate. This crucial depend-
ency on carbohydrate acceptor was yet more dramatically
confirmed by the complete absence of product from the
attempted esterification of N-acetylglucosamine 12a from
which only 12a and the product of acyl donor hydrolysis 1b
were recovered. Next the effect of anomeric substituent was
probed. Introduction of a methyl substituent at O-1 increased
yield only slightly in the case of
acceptors 13–15a. Moreover, the near identical yields of a- and
-gluco
D
-galactose and -glucose acyl
D
b-glucosides 13, 14b indicated that, at least in the
D
series, anomeric stereochemistry had little or no effect on
overall yield. Most notably, the apparent specificity preference
of SBL for -manno acyl acceptors was further confirmed by
D
the higher yield (76%) of ester 16b obtained here from a-
D
-
mannoside 16a.
Thioglycosides and selenoglycosides are important glycosyl
donors16 and we next investigated their esterification to provide
potential glycopeptide donors, in which the glycosyl unit might
be further extended, and as further probes of the effect of
anomeric substituent. Consistent with both their larger size and
the potential for aromatic aglycones in carbohydrate substrates
to interact with protein surfaces,17 more dramatic results were
obtained for the thioglycosides 17–20a. A trend in the
efficiencies of the formation of 6-O-phenylalaninate products in
the order
D
-manno >
D
-gluco >
D
-galacto > N-acetyl- -gluco
D
emerged. In addition, for the first time, reduced regioselectivity
was observed for
D
-thioglucoside 17a (3+2, 6-O 17b: 3-O
17c).18
† Electronic supplementary information (ESI) available: experimental
details. See http://www.rsc.org/suppdata/cc/b1/b104137c/
Scheme 2
1908
Chem. Commun., 2001, 1908–1909
This journal is © The Royal Society of Chemistry 2001
DOI: 10.1039/b104137c

View Article Online
Table 1 Carbohydrate–amino acid coupling reactions
highly regioselective protease-catalyzed transesterification
process. The yields for this selective carbohydrate–peptide
conjugation of 23–76%, compare well with overall yields of
< 34% for alternative routes employing protection–deprotec-
tion strategies.7 The glycopeptides formed are powerful build-
ing blocks that will allow sugar reducing end (e.g. 17–20b) or
peptide N-terminal (e.g. 21) extension. In addition, we have
probed the substrate specificity of the proteases SBL and TL-
CLEC in this reaction using the novel vinyl esters 1–8c and this
has indicated a strong preference for phenylalanine but
flexibility in the N-protection that may be used. Furthermore,
we have successfully exploited striking differences in the rate of
reaction of carbohydrate acyl acceptors in this system to
perform exclusively mannose over N-acetylglucosamine se-
lective one-pot acylations. We have recently reported greatly
broadened substrate amino acid ester specificities for glycosy-
lated variants of SBL22 and we are currently exploring
transesterifications catalyzed by these glyco-SBLs with 4–8c
and other donors the results of which will be reported in due
course.
Coupling
pair
Yield of
R1
R2
R3
R4
R5
6-O-acyl (%)c
9a–1ca
OH
OH
OH
OH
H
H
OH
H
H
H
H
OH
H
H
OH
H
OH
OH
OH
OH
OH
OH
H
OH
H
OH
H
OH
H
H
H
H
OH
H
H
OH
H
24 9b
10a–1ca
11a–1ca
12a–1ca
13a–1ca
14a–1ca
15a–1ca
16a–1ca
17a–1ca
18a–1ca
19a–1ca
20a–1ca
16a–1cb
16a–2ca
16a–2cd
16a–3cd
24 10b
49 11b
NHAc OH
—
a-OMe
b-OMe
b-OMe
a-OMe
b-SPh
b-SPh
a-SPh
b-SePh
a-OMe
a-OMe
a-OMe
a-OMe
OH
OH
OH
H
OH
OH
H
OH
OH
H
OH
OH
H
25 13b
28 14b
30 15b
76 16b
44 17b + 29 17ce
36 18b
OH
62 19b
NHAc OH
H
H
H
H
23 20b
H
H
H
H
OH
OH
OH
OH
48 16b
32 16c
63 16c + 17 16de
60 16e
H
a 2 mg ml21 of lyophilized (from pH 8.0, 0.1 M phosphate) SBL
preparation, 45 °C, anhydrous pyridine, 120 h. b 1 mg ml21 of TL-CLEC,
45 °C, 1+25 water–pyridine. c All yields are for isolated, purified, single
compounds. d As for footnote a but for 500 h. e Yield of 3-O-acyl.
We thank the BBSRC for generous funding, Genencor
International for SBL, and Altus for TL-CLEC. We thank the
EPSRC for access to the Mass Spectrometry Service at Swansea
and the Chemical Database Service at Daresbury.
Next we investigated the effect of varying the amino acid acyl
donor. Disappointingly, but consistent with the observed low
affinity of SBL for other amino acid esters,19 none of the
aspartate or glutamate acyl donors were accepted as substrates.
In all cases only vinyl esters 4–8c were recovered indicating an
absence of acyl–enzyme intermediate formation. This con-
trasted with the reactions of 1c from which only transesterifica-
tion or hydrolysis products were recovered. In order to further
assess the utility of 1, 4–8c as acyl donor probes, we also
screened their reactivity with CLEC-thermolysin (TL-CLEC)20
as a protease with a different substrate specificity profile, that
includes b-aspartate esters.21 However, TL-CLEC also failed to
accept 4–8c and again only 1c was accepted, allowing the
preparation of 16b from 16a in 48% yield.
Notes and references
1 B. G. Davis, J. Chem. Soc., Perkin Trans. 1, 1999, 3215.
2 C. M. Taylor, Tetrahedron, 1998, 54, 11 317.
3 A. Varki, Glycobiology, 1993, 3, 97; R. A. Dwek, Chem. Rev., 1996, 96,
683.
4 G. F. Springer, J. Mol. Med., 1997, 75, 594.
5 T. Tsuda and S.-I. Nishimura, J. Chem. Soc., Chem. Commun., 1996,
2779.
6 O. Seitz, Chem. Bio. Chem., 2000, 1, 215; M. Meldal and P. M. St
Hilaire, Curr. Opin. Chem. Biol., 1997, 1, 552.
7 R. J. Tennant-Eyles and A. J. Fairbanks, Tetrahedron: Asymmetry,
1999, 10, 391 and references therein.
8 N. B. Bashir, S. J. Phythian, A. J. Reason and S. M. Roberts, J. Chem.
Soc., Perkin Trans 1, 1995, 2203.
9 S. Riva, J. Chopineau, A. P. G. Kieboom and A. M. Klibanov, J. Am.
Chem. Soc., 1988, 110, 584; Y.-F. Wang, K. Yakovlevsky, B. Zhang
and A. L. Margolin, J. Org. Chem., 1997, 62, 3488.
10 S. Horvat, J. Horvat, L. Vargadefterdarovic, K. Pavelic, N. N. Chung
and P. W. Schiller, Intl. J. Pept. Prot. Res., 1993, 41, 399.
11 R. D. Egleton, S. A. Mitchell, J. D. Huber, J. Janders, D. Stropova, R.
Polt, H. I. Yamamura, V. J. Hruby and T. P. Davis, Brain Res., 2000,
881, 37.
12 M. Dickman, R. C. Lloyd and J. B. Jones, Tetrahedron: Asymmetry,
1998, 9, 4099; M. Dickman, R. C. Lloyd and J. B. Jones, Tetrahedron:
Asymmetry, 1998, 9, 551.
13 M. Lobell and M. P. Schneider, Synthesis, 1994, 375.
14 Y.-F. Wang, J. J. Lalonde, M. Momongan, D. E. Bergbreiter and C.-H.
Wong, J. Am. Chem. Soc., 1988, 110, 7200. N-Protected vinyl ester
amino acids have been used previously as non-enzymatic acyl donors
with amines for peptide synthesis: F. Weygand and W. Steglich, Angew.
Chem., 1961, 73, 757. In all cases, 1–8c showed no non-enyzmatic
reaction with 9–20a.
Next, the effect of N-protection in the acyl donor was
investigated using Boc- and Z-protected phenylalanine donors
2, 3c, respectively. For 2c much lower rates of reaction were
observed than for 1c and after a comparable period of time
lower yields (32%) for the esterification of 16a were obtained.
However, extended reaction times gratifyingly allowed the
preparation of 6-O-phenylalaninates 16c, e from 2, 3c in 63 and
60% yields, respectively. The utility of 16c, e as glycopeptide
building blocks was confirmed through their quantitative N-
deprotection to methyl 6-O-phenylalaninyl-a- -mannopyrano-
D
side 21, which may be extended at its N-terminus.
Finally, the valuable specificity information obtained in these
screens was exploited to allow selective one-pot couplings. We
were delighted to find that different carbohydrate acyl acceptors
successfully competed in one-pot reactions to allow carbohy-
drate-selective esterification. Thus, in 1+1 mixtures of 12a +
16a and 19a + 20a (Scheme 3) mannosides reacted over N-
acetylglucosaminides with 1c in SBL-catalyzed acylations to
yield mannoside esters 16, 19b exclusively. In both reactions no
trace of 12b or 20b, respectively, was detected during this
highly selective process.
15 Regiochemistry of O-X-esterification products confirmed by 1H, 13C
NMR e.g. 18a 1H NMR (CD3OD) d 3.62 (H-6), 3.66 (H-6A); 13C NMR
(CD3OD) d 62.6 (C-6) ? 18b 1H NMR (CD3OD) d 4.23 (H-6), 4.32 (H-
6A); 13C NMR (CD3OD) d 65.9 (C-6) and/or HMBC, HSQC; OH–H d6-
DMSO COSY; OH acylation.
16 B. G. Davis, J. Chem. Soc., Perkin Trans. 1, 2000, 2137.
17 S. J. Chung, S. Takayama and C.-H. Wong, Bioorg. Med. Chem. Lett.,
1998, 8, 3359.
In summary, we have described a ready method for the
construction of glycan–peptide conjugates by exploiting a
18 At the helpful suggestion of a referee, the question of whether 17c is a
direct or indirect, rearranged acylation product was investigated.
Compound 17b, under standard reaction conditions but in the absence of
donor 1c, did not yield 17c.
19 K. Khumtaveeporn, G. DeSantis and J. B. Jones, Tetrahedron:
Asymmetry, 1999, 10, 2563.
20 CLECs are cross-linked enzyme crystals.
21 M. Miyanaga, M. Ohmori, K. Imamura, T. Sakiyama and K. Nakanishi,
J. Biosci. Bioeng., 2000, 90, 43.
22 K. Matsumoto, B. G. Davis and J. B. Jones, Chem. Commun., 2001,
903.
Scheme 3
Chem. Commun., 2001, 1908–1909
1909