Sugar-assisted ligation in glycoprotein synthesis

Article abstract of DOI:10.1021/ja0708971

Sugar-assisted ligation (SAL) presents an attractive strategy for the synthesis of glycopeptides, including the synthesis of cysteine-free β-O-linked and N-linked glycopeptides. Here we extended the utility of SAL for the synthesis of α-O-linked glycopeptides and glycoproteins. In order to explore SAL in the context of glycoprotein synthesis, we developed a new chemical synthetic route for the α-O-linked glycoprotein diptericin ε. In the first stage of our synthesis, diptericin segment Cys(Acm) ³⁷-Gly⁵² and segment Val⁵³-Phe⁸² were assembled by SAL through a Gly-Val ligation junction. Subsequently, after Acm deprotection, diptericin segment Cys³⁷-Phe⁸² was ligated to segment Asp¹-Asn³⁶ by means of native chemical ligation (NCL) to give the full sequence of diptericin ε. In the final synthetic step, hydrogenolysis was applied to remove the thiol handle from the sugar moiety with the concomitant conversion of mutated Cys³⁷ into the native alanine residue. In addition, we extended the applicability of SAL to the synthesis of glycopeptides containing cysteine residues by carrying out selective desulfurization of the sulfhydryl-modified sugar moiety in the presence of acetamidomethyl (Acm) protected cysteine residues. The results presented here demonstrated for the first time that SAL could be a general and useful tool in the chemical synthesis of glycoproteins.

Full text of DOI:10.1021/ja0708971

Published on Web 05/25/2007
Sugar-Assisted Ligation in Glycoprotein Synthesis
Yu-Ying Yang,† Simon Ficht,† Ashraf Brik,*,†,§ and Chi-Huey Wong*,†,‡
Contribution from the Department of Chemistry, The Scripps Research Institute, 10550 North
Torrey Pines Road, La Jolla, California 92037, and Genomic Research Center, Academia
Sinica, 128 Sec. 2, Academia Road, Nankang, Taipei 115, Taiwan
Received February 7, 2007; E-mail: abrik@bgu.ac.il; wong@scripps.edu
Abstract: Sugar-assisted ligation (SAL) presents an attractive strategy for the synthesis of glycopeptides,
including the synthesis of cysteine-free â-O-linked and N-linked glycopeptides. Here we extended the utility
of SAL for the synthesis of R-O-linked glycopeptides and glycoproteins. In order to explore SAL in the
context of glycoprotein synthesis, we developed a new chemical synthetic route for the R-O-linked
glycoprotein diptericin ꢀ. In the first stage of our synthesis, diptericin segment Cys(Acm)37-Gly52 and segment
Val53-Phe82 were assembled by SAL through a Gly-Val ligation junction. Subsequently, after Acm
deprotection, diptericin segment Cys37-Phe82 was ligated to segment Asp1-Asn36 by means of native chemical
ligation (NCL) to give the full sequence of diptericin ꢀ. In the final synthetic step, hydrogenolysis was applied
to remove the thiol handle from the sugar moiety with the concomitant conversion of mutated Cys37 into
the native alanine residue. In addition, we extended the applicability of SAL to the synthesis of glycopeptides
containing cysteine residues by carrying out selective desulfurization of the sulfhydryl-modified sugar moiety
in the presence of acetamidomethyl (Acm) protected cysteine residues. The results presented here
demonstrated for the first time that SAL could be a general and useful tool in the chemical synthesis of
glycoproteins.
the glycosylation pattern on the desired glycoprotein.3a,4 The
synthetic strategies for glycopeptides and glycoproteins generally
Introduction
Glycosylation represents the most prevalent post-translational
modification of proteins. More than 50% of proteins in humans
are glycosylated.1 However, efforts toward understanding the
effects of carbohydrates on their attached proteins are greatly
hampered by the difficulty in obtaining sufficient amounts of
pure single glycoform for studies. Unlike protein or DNA,
oligosaccharide assembly is not controlled by a defined tem-
plate.2 As a result, glycoproteins usually exist as heterogeneous
mixtures of glycoforms, which possess various glycosylation
patterns on the same protein backbone. Recent advances in the
development of glycoprotein synthesis, however, have created
opportunities for researchers to obtain homogeneous glycopro-
teins in the laboratory.3 For instance, Gerngross and co-workers
recently have genetically engineered yeast to produce human
therapeutic glycoproteins in single glycoforms.3c-d
involve several techniques, including carbohydrate synthesis,
peptide assembly, and enzymatic elaboration.3a,4a,5 To success-
fully synthesize the desired glycopeptides or glycoproteins, every
element in each technique has to be carefully coordinated. Native
chemical ligation (NCL) represents the most widespread and
powerful peptide ligation method for protein synthesis.5a,c,6 It
allows the assembly of two unprotected synthetic peptides
through a chemoselective reaction between peptide bearing
N-terminal cysteine and a peptide bearing C-terminal thioester.
This strategy was further strengthened by the development of
expressed protein ligation (EPL) to extend its utility to the
synthesis of large-sized proteins.5d,7 However, in both NCL and
EPL, the requirement of N-terminal cysteine for ligation imposes
an intrinsic limitation to their application. For instance, to take
(4) (a) Davis, B. G. Chem. ReV. 2002, 102, 579-602. (b) Guo, Z.; Shao, N.
Med. Res. ReV. 2005, 25, 655-678. (c) Pratt, M. R.; Bertozzi, C. R. Chem.
Soc. ReV. 2005, 34, 58-68. (d) Brik, A.; Ficht, S.; Wong, C.-H. Curr.
Opin. Chem. Biol. 2006, 10, 638-644.
In comparison to biological methods, synthetic approaches
provide researchers more flexibility and control in manipulating
(5) (a) Shin, Y.; Winans, K. A.; Backes, B. J.; Kent, S. B. H.; Ellman, J. A.;
Bertozzi, C. R. J. Am. Chem. Soc. 1999, 21, 11684-11689. (b) Koeller,
K. M.; Smith, M. E. B.; Wong, C.-H. J. Am. Chem. Soc. 2000, 122, 742-
743. (c) Mcmillan, D.; Bertozzi, C. R. Angew. Chem., Int. Ed. 2004, 43,
1355-1359. (d) Hackenberger, C. P. R.; Friel, C. T.; Radford, S. E.;
Imperiali, B. J. Am. Chem. Soc. 2005, 127, 12882-12889. (e) Li, B.; Song,
H.; Hauser, S.; Wang, L.-X. Org. Lett 2006, 8, 3081-3084.
(6) (a) Dawson, P. E.; Muir, T. W.; Clark-Lewis, I.; Kent, S. B. H. Science
1994, 266, 776-779. (b) Dawson, P. E.; Kent, S. B. H. Annu. ReV. Biochem.
2000, 69, 923-960. (c) Marcaurelle, L. A.; Mizoue, L. S.; Wilken, J.;
Oldhm, L.; Kent, S. B. H.; Hndel, T. M.; Bertozzi, C. R. Chem.sEur. J.
2001, 7, 1129-1132.
† The Scripps Research Institute.
‡ Academia Sinica.
§ Current address: Department of Chemistry, Ben Gurion University,
Beer Sheva 84105, Israel.
(1) Apweiler, R.; Hermjakob, H.; Sharon, N. Biochim. Biophys. Acta 1999,
1473, 4-8.
(2) (a) Dwek, R. A.; Butters, T. D.; Platt, F. M.; Zitzmann, N. Nat. ReV. Drug
DiscoVery 2002, 1, 65-75. (b) Spiro, R. G. Glycobiology 2002, 12, 43-
56. (c) Graffenried, C. L. D.; Bertozzi, C. R. Curr. Opin. Cell Biol. 2004,
16, 356-363. (d) Sinclair, A. M.; Elliott, S. J. Pharm. Sci. 2005, 94, 1626-
1635.
(3) (a) Sears, P. Science 2001, 291, 2344-2350. (b) Xu, R.; Hanson, S. R.;
Zhang, Z.; Yang, Y.-Y.; Schultz, P. G.; Wong, C.-H. J. Am. Chem. Soc.
2004, 126, 15654-15655. (c) Hamilton, S. R., et al. Science 2006, 313,
1441-1443. (d) Li, H., et al. Nat. Biotechnol. 2006, 24, 210-215.
(7) (a) Muir, T. W.; Sondhi, D.; Cole, P. A. Proc. Nat. Acad. Sci. U.S.A. 1998,
95, 6705-6710. (b) Muralidharan, V.; Muir, T. W. Nature Methods 2006,
3, 429-438. (c) Tolbert, T. J.; Franke, D.; Wong, C.-H. Bioorg. Med. Chem.
2005, 13, 909-915.
9
7690
J. AM. CHEM. SOC. 2007, 129, 7690-7701
10.1021/ja0708971 CCC: $37.00 © 2007 American Chemical Society

Sugar-Assisted Ligation in Glycoprotein Synthesis
A R T I C L E S
Scheme 1. Proposed Mechanism of the Sugar-Assisted Ligation for R-Linked Glycopeptide Synthesis
advantage of these ligation methods, in many cases, introducing
cysteine mutations into protein sequences is necessary, which
could lead to the formation of mismatched disulfide bonds
during protein folding.
Recently, our group introduced a new chemoselective ligation
method named sugar-assisted ligation (SAL) for assembling
glycopeptides through non-cysteinyl residues and successfully
demonstrated its utility in the synthesis of cysteine-free â-O-
linked and N-linked glycopeptides.10 SAL adopts similar
concepts presented in NCL and removable auxiliary-assisted
ligation. However, instead of anchoring the auxiliary onto the
N-terminus of the peptide, we take advantage of the sugar by
modifying its acetamido moiety at the C2 position with a
sulfhydryl group to enable its transthioesterification with peptide
thioester (Scheme 1). Notably, we believe the restricted
conformation of the sugar moiety plays an important role in
the S f N acyl transfer by bringing the NR-amino group of
glycopeptide and the electrophilic carbonyl group of the thioester
intermediate into close proximity to facilitate the intramolecular
rearrangement via 14- or 15-membered ring.11 After the ligation,
the sulfhydryl group can be removed by hydrogenolysis to
regenerate the unmodified sugar.8b,10 Alternatively, the sulfhy-
dryl-modified sugar can be further elaborated with glycosyl-
transferases to extend its glycan structure or alkylated with
labeling reagents such as fluorophores.10b One important feature
To tackle this problem, different strategies were developed
to expand the repertoire of ligation junctions of NCL into non-
cysteinyl residues.8 One of the most notable approaches is the
removable auxiliary-assisted ligation strategy.8c-d,9 In this
strategy, a removable thiol auxiliary mimicking the function of
N-terminal cysteine was anchored onto the N-terminus of a
peptide to promote the transthioesterification and the subsequent
intramolecular S f N acyl transfer. The auxiliary was then
removed after the ligation to regenerate the unmodified peptide
chain. Various elegant auxiliaries have been developed to enable
the ligation of peptides or glycopeptides through non-cysteinyl
junctions8c-d,9 Nevertheless, so far, most of these auxiliaries
demand at least one Gly residue at its ligation site.8c-d,9a-e
(8) (a) Beligere, G. S.; Dawson, P. E. J. Am. Chem. Soc. 1999, 121, 6332-
6333. (b) Yan, L. Z.; Dawson, P. E. J. Am. Chem. Soc. 2001, 123, 526-
533. (c) Bang, D.; Makhatadze, V. T.; Kossiakoff, A. A.; Kent, S. B. Angew.
Chem., Int. Ed. 2005, 44, 3852-3856. (d) Macmillan, D. Angew. Chem.,
Int. Ed. 2006, 45, 7668-7672.
(9) (a) Cane, L. E.; Bark, S. J.; Kent, S. B. H. J. Am. Chem. Soc. 1996, 118,
5891-5896. (b) Low, D. W.; Hill, M. G.; Carrasco, M. R.; Kent, S. B. H.;
Botti, P. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 6554-6559. (c) Offer, J.;
Boddy, C.; Dawson, P. E. J. Am. Chem. Soc. 2002, 124, 4642-4646. (d)
Macmillan, D.; Anderson, D. W. Org. Lett. 2004, 6, 4659-4662. (e) Wu,
B.; Chen, J.; Warren, J. D.; Chen, G.; Hua, Z.; Danishefsky, S. J. Angew.
Chem., Int. Ed. 2006, 45, 4116-4125. (f) Chen, J.; Chen, G.; Wu, B.;
Wan, Q.; Tan, Z.; Hua, Z.; Danishefsky, S. J. Tetrahedron Lett. 2006, 47,
8013-8016.
(10) (a) Brik, A.; Yang, Y.-Y.; Ficht, S.; Wong, C.-H. J. Am. Chem. Soc. 2006,
128, 5626-5627. (b) Brik, A.; Ficht, S.; Yang, Y.-Y.; Bennett, C. S.; Wong,
C.-H. J. Am. Chem. Soc. 2006, 128, 15026-15033.
(11) Molecular simulations of transition states have been performed and will
be presented in a following publication. The simulations support the
hypothesis that the restricted conformation of the sugar helps to preorganize
the GalNAc-linked thioester intermediate and N-terminal amine for
nucleophilic attack.
9
J. AM. CHEM. SOC. VOL. 129, NO. 24, 2007 7691

A R T I C L E S
Yang et al.
Scheme 2. Synthesis of the Building Blocks 4a and 4ba
a Reagents and Conditions: (a) Fmoc-Thr-OAll (R ) CH3 for 2a) or Fmoc-Ser-OAll (R ) H for 2b), TMSOTf, molecular sieves AW 300, CH2Cl2,
-70 °C, R/â ) 4/3 (2aR, 45% yield; 2bR, 45% yield); (b) Zn/AcOH; (c) TrtS-CH2COOH, HBTU, DIEA, DMF, 84% (two steps); (d) Pd(PPh3)4, NMA,
THF, 95% (Trt ) trityl, TMSOTf ) trimethylsilyl trifluoromethanesulfonate).
about SAL, when compared to removable auxiliaries, is that
the nucleophile that participates in the S f N rearrangement is
a primary amine. This less sterically hindered nucleophile is
one of the attributing factors that allow SAL to accept a broader
spectrum of amino acid residues at its ligation site. Preferences
are for those amino acids possessing less hindered side chains
(Gly and Ala) and those containing a side chain that can serve
as a general base (His and Asp) in the ligation pathway.10b
This broad sequence tolerance exhibited in the ligation sites
of SAL represents the most attractive advantage in this method
for chemical glycoprotein synthesis. Given most naturally
occurring glycoproteins bear multiple glycosylation sites, SAL
could be quite useful if these appending glycans of glycoproteins
are eligible to participate in glycopeptide ligation. To enable
SAL to be a general method for the preparation of homogeneous
glycoproteins, herein, we made several advancements of this
method, including extending its utility to the synthesis of R-O-
linked glycopeptides, demonstrating selective desulfurization to
enable this method for the synthesis of cysteine-containing
glycopeptides and, more importantly, demonstrating its utility
in the synthesis of the complete R-O-linked glycoprotein
diptericin ꢀ.
Diptericin ꢀ is an 82-residue, cysteine-free R-O-linked
antibacterial glycoprotein.12 It contains two glycosylation sites,
Thr10 and Thr54, in which each threonine carries an N-acetyl-
galactosamine (GalNAc) moiety (Scheme 3A).5a,12,13 Previously,
Bertozzi and co-workers have successfully applied NCL to
chemically synthesize diptericin ꢀ.5a The synthesis was ac-
complished through the assembly of two glycopeptide seg-
ments, in which the glycopeptide thioester was prepared by
Fmoc-based solid-phase peptide synthesis (SPPS) using the
Kenner linker. Notably, since diptericin ꢀ contains no cysteine
residue, the Bertozzi group mutated Gly25 to cysteine to allow
NCL to be amenable to diptericin synthesis. In this report, we
took advantage of the glycosylation site Thr54 of diptericin ꢀ to
apply SAL for the first time to the synthesis of a complete
glycoprotein.
mucin-type O-linked glycans, which are typically R-linked to
serine or threonine residues.14 Mucin-type O-linked glycans exist
abundantly in naturally occurring glycoproteins. To apply SAL
for the synthesis of this type of R-O-linked glycopeptides and
glycoproteins, first, we prepared the NR-Fmoc-Thr[Ac3-R-
GalNAc(SH)] 4a and NR-Fmoc-Ser[Ac3-R-GalNAc(SH)] 4b as
the building blocks for SPPS. As shown in Scheme 2, the
synthetic strategy for building block 4a is straightforward. The
synthesis was started with TMSOTf-mediated glycosylation of
compound 115 with Fmoc-Thr-OAll at -78 °C to afford
glycosylated product 2a in an R and â mixture (R/â ) 4/3).
Compound 2aR was separated from the mixture using column
chromatography to give the desired pure R-isomer in 45% yield.
Following purification, the azido moiety was reduced to amine
by treatment with Zn/AcOH.16 The amino group was then
coupled with S-trityl-2-mercaptoacetic acid17 under standard
HBTU/DIEA coupling conditions18 to give the compound 3a.
Subsequently, the allyl group was removed by treatment with
a catalytic amount of Pd(PPh3)4 in the presence of N-methyla-
niline19 to furnish the building block 4a. Building block 4b was
also prepared using a similar synthetic strategy (Scheme 2).
Synthesis of R-O-Linked Glycopeptides Using SAL. To
investigate the ligation efficiency of SAL in R-O-linked
glycopeptides, we synthesized four short glycopeptides 5a-b
and 6a-b as well as peptide thioester Cys(Acm)37-Gly52 for
model studies (Table 1). Peptides containing the sequence from
diptericin ꢀ were designed to explore the optimal ligation
conditions for diptericin ꢀ synthesis. Having the building blocks
4a and 4b in hand, we synthesized Ser[R-GalNAc(SH)]-
containing glycopeptides 5a-b and Thr[R-GalNAc(SH)]-
containing glycopeptides 6a-b by means of Fmoc-based
(12) Bulet, P.; Hegy, G.; Lambert, J.; Dorsselaer, A. V.; Hoffmann, J. A.; Hetru,
C. Biochemistry 1995, 34, 7394-7400.
(13) (a) Winans, K. A.; King, D. S.; Rao, V. R.; Bertozzi, C. R. Biochemistry
1999, 38, 11799-11710. (b) Cudic, M.; Bulet, P.; Hoffmann, R.; Craik,
D. J.; Otvos, L., Jr. Eur. J. Biochem. 1999, 266, 549-558.
(14) (a) Hanisch, F.-G. Biol. Chem. 2001, 382, 143-149. (b) Hang, H. C.;
Bertozzi, C. R. Bioorg. Med. Chem. 2005, 13, 5021-5034.
(15) Koeller, K. M.; Smith, M. E. B.; Wong, C.-H. Bioorg. Med. Chem. 2000,
8, 1017-1026.
(16) Schultz, M.; Kunz, H. Tetrahedron Lett. 1992, 33, 5319-5322.
(17) Kupihar, Z.; Schmel, Z.; Kele, Z.; Penke, B.; Kovacs, L. Bioorg. Med.
Chem. 2001, 9, 1241-1247.
(18) Albericio, F.; Carpino, L. A. Methods Enzymol. 1997, 289, 104-126.
(19) Ciommer, M.; Kunz, H. Synlett 1991, 8, 593-595.
Results and Discussion
Synthesis of Building Blocks for SAL. Diptericin ꢀ bears
two GalNAc moieties.12 GalNAc is the core structure of the
9
7692 J. AM. CHEM. SOC. VOL. 129, NO. 24, 2007

Sugar-Assisted Ligation in Glycoprotein Synthesis
A R T I C L E S
Table 1. Sugar-Assisted Ligation of R-Linked Glycopeptides
product mass (Da)
a
ligation
junction
t1
thioesterc
(equiv)
/
2
entry
-X1-
-X2-
(h)
reaction medium
obsdb
calcd
1
2
3
4
5
Gly
Val
Gly
Val
Val
Ser
Ser
Thr
Thr
Thr
Gly-Gly
Gly-Val
Gly-Gly
Gly-Val
Gly-Val
4
11
4
11
14
1.2
1.2
1.2
1.2
1.2
0.2 M phosphate buffer
0.2 M phosphate buffer
0.2 M phosphate buffer
0.2 M phosphate bufferd
0.2 M phosphate buffer +
2% PhSH
2677.4f
2720.4g
2691.5h
2733.6i
2733.6j
2677.4
2719.5
2691.4
2733.5
2733.5
6
7
8
Val
Val
Val
Thr
Thr
Thr
Gly-Val
Gly-Val
Gly-Val
14
9
7
1.2
1.5
2.0
0.1 M phosphate buffere
0.2 M phosphate bufferd
0.2 M phosphate bufferd
2733.6k
2733.6l
2733.6m
2733.5
2733.5
2733.5
b
c
a t1/2 indicates the required reaction time for the reactant glycopeptide to reach 50% consumption. Characterized by LC/MS. The concentration of the
reactant glycopeptide in each reaction is 6 mM. dReaction pH is around 7.2-7.4. eReaction pH is around 7.0. fIsolated yield ) 78%. gIsolated yield ) 56%.
hIsolated yield ) 73%. Isolated yield ) 59%. Isolated yield ) 45%. Isolated yield ) 47%. Isolated yield ) 63%. mIsolated yield ) 63%.
i
j
k
l
SPPS10b,20 on Rink-amide resin. Peptide thioester Cys(Acm)37-
Gly52 was prepared by Boc-based SPPS10a,21 on MBHA resin.
Herein, two ligation junctions Gly-Gly and Gly-Val were
examined, in which Gly-Val represents the ligation junction
found in diptericin ꢀ (Scheme 3A). As the results show in Table
1, R-O-linked glycopeptides were successfully synthesized by
SAL with the resulting ligation efficiencies similar to those of
â-O-linked10a and N-linked glycopeptides10b (Table 1, entries
1-4).22 GalNAc(SH)-Thr-containing glycopeptides exhibited
ligation rates similar to those of GalNAc(SH)-Ser-containing
glycopeptides, despite the additional steric factor contributed
by the methyl group of threonine (Table 1, entries 1-4). The
similarity of the ligation rates among various glycopeptides
indicates that the feature of the glycosidic linkage as well as
the glycan structure may have a negligible effect on the ligation
efficiency. Notably, in this report, thiophenol was excluded from
the ligation media in most of the experiments. As our earlier
studies have shown that the rate-determining step of SAL is S
f N acyl transfer,10b this suggests that thiophenol may
deteriorate the ligation efficiency, especially in cases exhibiting
a slower ligation rate. As shown in entry 5 (Table 1), the ligation
rate at the Gly-Val junction decreases when thiophenol was
present in the reaction solution (entry 4 vs 5). On the other
hand, to investigate the optimal ligation conditions for the Gly-
Val junction, a series of ligation studies were carried out using
glycopeptide 6b and peptide thioester Cys(Acm)37-Gly.52 As
shown in Table 1, the ligation efficiency at the Gly-Val junction
is affected by the reaction pH (Table 1, entries 4 and 6). An
ideal reaction pH for SAL is around 7.2 to 7.4, as measured by
a microelectrode.23 In this range of pH, SAL reaches its highest
efficiency while giving a minimum amount of hydrolyzed
thioester. As reactant concentrations were high, the best ligation
efficiency was observed at high phosphate buffer concentration,
to maintain the pH through the reaction (Table 1, entries 4 and
6). Moreover, the ligation efficiency at the Gly-Val junction
can be increased by adding more equivalents of peptide thioester
(Table 1, entries 4, 7, and 8).
Synthetic Strategy of Diptericin. Our synthetic strategy for
diptericin ꢀ is depicted in Scheme 3. In our synthetic design,
the whole protein sequence was dissected into three segments,
which were assembled sequentially from C- to N-terminus by
SAL and then NCL (Scheme 3). The residues Gly52-Val53 next
to the glycosylation site Thr54 represent an ideal ligation site to
demonstrate the general utility of SAL. The sequence of the
C-terminal segment Cys(Acm)37-Phe82 and middle segment Cys-
(Acm)37-Gly52 covers nearly all the standard L-amino acids,
except Met and Glu. In addition, the relatively slow Gly-Val
junction is difficult enough to represent a typical challenge in
glycoprotein synthesis by SAL. Performing SAL gave us the
46-residue glycopeptide Cys(Acm)37-Phe.82 After removing
Acm, Cys37-Phe82 was coupled to N-terminal segment Asp1-
Asn36 by NCL to generate the full-sized diptericin ꢀ. By
incorporating NCL into our synthetic strategy, the potential
difficulty associated with the synthesis of a 52-residue glyco-
peptide thioester was bypassed. To take advantage of NCL, Ala37
was temporarily mutated to cysteine. In the final synthetic step,
both the cysteine residue and the thiol handle on the sugar were
desulfurized8b,10 to complete the synthesis of diptericin ꢀ.
Synthesis of Diptericin Segment Cys(Acm)37-Phe82 by
SAL. The C-terminal segment of diptericin, the glycopeptide
Val53-Phe,82 was prepared by Fmoc-SPPS20 on 2-ClTrt resin
using the building block 4a. For the synthesis of the middle
segment, the peptide thioester Cys(Acm)37-Gly,52 the acetami-
domethyl (Acm) group was used to protect N-terminal cysteine
(20) Wellings, D. A.; Atherton, E. Methods Enzymol. 1997, 289, 44-67.
(21) (a) Alewood, P.; Alewood, D.; Miranda, L.; Love, S.; Meutermans, W.;
Wilson, D. Methods Enzymol. 1997, 289, 14-29. (b) Stewart, J. M. Methods
Enzymol. 1997, 289, 29-44.
(22) For instance: t1/2 (h) for the Gly-Gly junction for which the ligation
efficiency is less susceptible to varying the reaction conditions: 5 h for
â-O-linked glycopeptides; 4 h for both N-linked and R-O-linked glyco-
peptides. On the other hand, t1/2(h) for the Gly-Val junction for which the
ligation efficiency is more susceptible to varying the reaction conditions:
18 h for N-linked glycopeptides (6 M Gn‚HCl, 0.1 M PBS, 2% PhSH,
glycopeptide (1.2 equiv), peptide thioester (1 equiv)); 14 h for R-O-linked
glycopeptides (6 M Gn‚HCl, 0.1 M PBS, glycopeptide (1 equiv), peptide
thioester (1.2 equiv)).
(23) The initial pH of the buffer used for SAL is 8.5. After addition of the
peptides, the pH drops due to the residual TFA present in HPLC purified
peptides. As a consequence, in some cases, triethanolamine was added to
adjust the reaction pH to the ideal range 7.2-7.4.
9
J. AM. CHEM. SOC. VOL. 129, NO. 24, 2007 7693

A R T I C L E S
Yang et al.
Scheme 3. (A) The Sequence of Diptericin and Ligation Sites. (B) The Synthetic Strategy of Diptericin
from self-ligation and intramolecular cyclization. The coupling
of these two segments by SAL was carried out in 6 M Gn‚HCl,
200 mM phosphate, pH ) 8.5, where the concentration of
glycopeptide Val53-Phe82 and the peptide thioester Cys(Acm)37-
Gly52 was 6 mM and 12 mM, respectively. Due to the change
of pH (the measured pH was 6.4) after mixing both peptides,
the reaction pH was adjusted to 7.2 by carefully adding
triethanolamine (∼0.2 µL per 100 µL reaction volume).23 The
ligation was carried out at 37 °C and monitored by LC/MS.
After 48 h, LC/MS analysis showed that over 70% of the
glycopeptide Val53-Phe82 had been consumed,24 and there was
only a trace amount of peptide thioester remaining. The ligated
product was purified by reversed-phase HPLC and lyophilized
to afford the glycopeptide Cys(Acm)37-Phe82 in 28% isolated
yield. To increase the product yield, we added an extra 0.5 equiv
of peptide thioester into the reaction mixture after the ligation
had proceeded for 24 h. After 48 h, the reaction was quenched
and subjected to HPLC purification. Using this method, the
reaction was able to yield 36% of isolated product (Figure 1A).
The glycopeptide Cys(Acm)37-Phe82 was characterized by
electrospray mass spectrometry (the reconstructed mass was
5408 Da; the calculated average mass was 5409 Da).
replaced Val53 with either Gly or His. Glycopeptide Gly53-Phe82
and glycopeptide His53-Phe82 were then ligated to the peptide
thioester Cys(Acm)37-Gly52 (1.5 equiv), respectively. These
ligations were monitored by LC/MS. Interestingly, in the ligation
of glycopeptide His53-Phe82 with peptide thioester, the reaction
proceeded smoothly to completion after 24 h even at pH 6.4
(Figure 1B). Given the pKa of the imidazole side chain of
histidine is around 6.5, this experimental result further supports
our previous hypothesis on His and Asp, in which the side chain
of these residues can serve as a general base in the ligation
pathway.9b On the other hand, the ligation of glycopeptide Gly53-
Phe82 with peptide thioester also proceeded to completion in
24 h, when the reaction pH was adjusted to 7.2. Together, these
ligation results show that SAL is chemoselective and amenable
for the synthesis of large sized glycopeptides.
Synthesis of the N-Terminal Diptericin Segment Glyco-
peptide Thioester Asp1-Asn36. Synthesis of glycopeptide
thioesters in the context of glycoprotein synthesis remains a
challenge. While Boc-based SPPS has been intensively used
for the synthesis of peptide thioesters, its synthetic conditions
are too harsh to be compatible with the presence of carbohy-
drates. As a result, many efforts have been aimed at adapting
Fmoc-basedSPPStothesynthesisofglycopeptidethioesters.5a,5d,9e,25
One of the central problems with the Fmoc strategy is that the
basic conditions used for removing acetate esters commonly
used as protecting groups for sugars are not compatible with
the presence of the thioester moiety. Consequently, the acetyl
groups cannot be deprotected prior to the ligation step, which
could complicate the synthesis or lead to a loss of ligated product
To explore the effect of the N-terminal amino acid residue
of the glycopeptide Val53-Phe82 on its ligation efficiency, we
(24) The glycopeptide was consumed by two reaction paths: reacting with the
peptide thioester Cys(Acm)37-Gly52 and by generating a degraded fragment.
The mass of this degraded fragment (observed mass ) 3901 Da)
corresponds to a loss of the first two N-terminal residues of glycopeptide
Val53-Phe82. The same fragment mass was also observed in the cases of
glycopeptide His53-Phe82 and glycopeptide Gly53-Phe82 by LC/MS. The first
observation of this degraded fragment was in the crude glycopeptide cleaved
from the resin. In the beginning of the ligation, the amount of this degraded
fragment is less than 5%. However, its amount increased during the course
of the reaction. A similar degradation phenomenon was not observed for
the short glycopeptides (Table 1).
(25) (a) Warren, J. D.; Miller, J. S.; Keding, S. J.; Danishefsky, S. J. J. Am.
Chem. Soc. 2004, 126, 6576-6578. (b) Mezzato, S.; Schaffrath, M.;
Unverzagt, C.; Angew. Chem., Int. Ed. 2005, 44, 1650-1654.
9
7694 J. AM. CHEM. SOC. VOL. 129, NO. 24, 2007

Sugar-Assisted Ligation in Glycoprotein Synthesis
A R T I C L E S
Figure 1. (A) Analytical HPLC chromatograph of SAL of glycopeptide Val53-Phe82 and peptide thioester Cys(Acm)37-Gly52 at 48 h. Peak a, hydrolyzed
peptide thioester; Peak b, the remaining glycopeptide Val53-Phe82 and its degraded fragment (the ratio is around 2:1);24 Peak c, the ligated product. HPLC
gradient: 5-30% of CH3CN (+0.1% TFA) over 30 min at a flow rate of 1.5 mL/min. (B) ESI-MS spectrum of the glycopeptide Cys(Acm)37-Phe82 (the
reconstructed mass was 5408; the calculated average mass was 5409 Da). (C) Analytical HPLC chromatograph of SAL of glycopeptide His53-Phe82 and
peptide thioester Cys(Acm)37-Gly52 at the time point of 24 h. Peak a, the hydrolyzed thioester; peak b, the ligated product and the degraded fragment from
glycopeptide His53-Phe82 (the ratio is around 9:1).24 HPLC gradient: 5-50% of CH3CN (+0.1% TFA) over 30 min at a flow rate of 1.5 mL/min. (D)
ESI-MS spectrum of the ligated product of glycopeptide His53-Phe82 and peptide thioester Cys(Acm)37-Gly52 (the reconstructed mass was 5446 Da; the
calculated average mass was 5447 Da).
in the O-acetate removal step.5a,d,25b To facilitate the synthesis
of fully unprotected glycopeptide thioesters Asp1-Asn36, here
we introduced the acid-labile p-methoxybenzyl (PMB) group
for the protection of the hydroxyl groups of GalNAc. In addition,
since the C-terminal amino acid residue of the glycopeptide
thioester is asparagine,26 we were allowed to operate the whole
synthesis of glycopeptide thioester Asp1-Asn36 on solid support
by using the side-chain anchoring strategy.27
The preparation of building block 10 was started with the
reaction of compound 715 with p-thiocresol in the presence of
BF3‚OEt2, which afforded the corresponding thioglycoside in
85% isolated yield (Scheme 4).28 Following this step, the acetyl
groups on the sugar were replaced with PMB groups to give
compound 8. The glycosylation of compound 8 with Fmoc-
Thr-OAllyl was carried out at -15 °C to give compound 9 in
R and â mixture (R/â ) 3/2). The desired R isomer was then
purified by column chromatography to afford compound 9R in
45% yield. The azido group of compound 9R was converted to
the acetamido group by reductive acetylation using thioacetic
acid in pyridine.29 Subsequently, the allyl group was removed
by a catalytic amount of Pd(PPh3)4 in the presence of N-
methylaniline17 to furnish the compound 10.
For the synthesis of glycopeptide thioester Asp1-Asn36, we
employed the side-chain anchoring strategy by following the
procedure reported by Wang et al. (Scheme 4).27 The unpro-
tected side-chain of Fmoc-Asp-OAllyl was first anchored onto
Rink-amide resin; afterward, the N-terminal amine was de-
blocked for glycopeptide synthesis. Fmoc-based SPPS was
conducted by using standard HBTU/DIEA coupling condi-
tions,15,30 and the building block 10 was attached to the peptide
as the 10th residue. Once the full sequence of the glycopeptide
was completed, the C-terminal carboxylic acid was unmasked
and transformed into CR-thioester by treatment with 3-mercap-
topropionic acid in the presence of DIC/HOBt and DIEA in
DMF/DCM (4/1).27 Following this thioesterification step, the
glycopeptide thioester was cleaved from the resin with con-
comitant full deprotection by TFA/triethylsilane/thioanisole/H2O
(85/5/5/5). The crude product was then purified by preparative
reversed-phase HPLC and lyophilized to afford the glycopeptide
(26) It caught our attention that Asp or Glu on the C-terminus of the peptide
thioester has the potential to form an intramolecular anhydride during NCL,
which could lead to the generation of a branched ligation product. To avoid
(29) Takahashi, S.; Kuzuhara, H.; Nakajima, M. Tetrahedron 2001, 57, 6915-
6926.
this problem, the original Asp36 of diptericin was mutated to Asn36
.
(30) Since Asp-Gly sequence is prone to aspartimide formation during peptide
synthesis, we mutated Asp29 to an Asn residue. On the other hand, our
earlier efforts toward solving this problem by using the commercially
available Fmoc-Asp(OtBu)-(Hmb)Gly-OH in SPPS failed due to the
difficulty of complete removal of the Hmb group in the final global
deprotection step.
Alternatively, an orthogonal protecting group could be used to avoid this
side reaction (Villain, M.; Gaertner, H.; Botti, P. Eur. J. Org. Chem. 2003,
3267-3272).
(27) Wang, P.; Miranda, L. P. Int. J. Pept. Res. Therap. 2005, 11, 117-123.
(28) Ye, X.-S.; Wong, C.-H. J. Org. Chem. 2000, 65, 2410-2431.
9
J. AM. CHEM. SOC. VOL. 129, NO. 24, 2007 7695

A R T I C L E S
Yang et al.
Scheme 4 . (A) Synthesis of the Building Block 10. (B) Synthesis of Glycopeptide Thioester Asp1-Asn36 Using the Side-Chain Anchoring
Strategya
a Reagents and Conditions: (a) BF3‚OEt2, TolSH, CH2Cl2, 0 °C then rt, 85%; (b) NaOMe, MeOH, pH ∼11; (c) PMBCl, NaH, DMF, 0 °C then rt, 68%
(three steps); (d) Fmoc-Thr-OAll, NIS, cat. TfOH, molecular sieves AW 300, CH2Cl2, -15 °C, 83%, R/â ) 3/2; (e) AcSH, pyridine, 86%; (f) Pd(PPh3)4,
NMA, THF, 95%. TolSH ) p-thiocresol, NIS ) N-iodosuccinimide.
thioester Asp1-Asn36 in 9% isolated yield. This compound was
characterized by electrospray mass spectrometry (the recon-
structed mass was 3975 Da; the calculated average mass was
3975 Da).
Acm Removal, NCL, and Desulfurization. Having the
glycopeptide segments in hand, we turned our focus to genera-
tion of the full sequence of diptericin ꢀ by means of NCL. The
N-terminal cysteine of glycopeptide Cys(Acm)37-Phe82 was first
deblocked by treatment with Hg(OAc)2/AcOH (pH ) 4.0)31 to
give the glycopeptide Cys37-Phe82 in 83% isolated yield (see
Supporting Information). The glycopeptide Cys37-Phe82 (3 mM)
and glycopeptide thioester Asp1-Asn36 (4.5 mM) were then
ligated in 6.0 M Gn‚HCl, 0.2 M phosphate buffer, pH ) 7.9
containing 2% (v/v) PhSH and 2% (v/v) BnSH at 37 °C. After
16 h, the ligation was nearly complete and the ligated product
appeared as the major peak by HPLC. After reversed-phased
HPLC purification and lyophilization, the pure glycopeptide
Asp1-Phe82 was obtained in 47% yield (see Supporting Informa-
tion). Finally, the ligated product was subjected to hydrogenoly-
sis under the conditions of H2/(Pd/Al2O3), to reduce the thiol
handle attached to GalNAc, as well as the cysteine residue.8b
After HPLC purification and lyophilization, the full-sized
Figure 2. (A) HPLC analysis of the crude product from the desulfurization
step, which was eluted with a gradient of 5-30% of CH3CN (+0.1% TFA)
over 30 min at a flow rate of 1.5 mL/min. (B) ESI-MS spectrum of the
diptericin was obtained in 54% isolated yield (254 µg). The
final product was characterized by electrospray mass spectrom-
etry (the reconstructed mass was 9113 Da; the calculated average
mass was 9114 Da) (Figure 2).
pure product of the desulfurization step (the reconstructed mass is 9113
Da).
shown that the thiol handle of the sugar moiety can be easily
removed under H2 in the presence of 10-20 mg of Pd/Al2O3
per 1 mg of glycopeptide. Herein, the hydrogenolysis of
compound 11 was carried out in the same condition previously
reported,10 6M Gn‚HCl, 0.1 M phosphate buffer, pH 5.8
containing 10 mM TCEP; however, in order to obtain desulfu-
rization selectivity, we performed this reaction by varying the
amount of Pd/Al2O3 added. As a consequence, the reaction
containing 1 mg of glycopeptide was reacted with 10, 15, and
Selective Desulfurization. To allow SAL to be amenable
for the synthesis of glycopeptides or glycoproteins containing
cysteine residues, the sulfhydryl-modified sugar moiety has to
be desulfurized8b selectively in the hydrogenolysis. To inves-
tigate desulfurization selectivity, we synthesized a model
glycopeptide 11 bearing one Cys(Acm) residue and a sulfhydryl-
modified GalNAc R-linked to a Ser residue. Previously, we have
(31) Ingale, S.; Buskas, T.; Boons, G.-J. Org. Lett. 2006, 8, 5785-5788.
9
7696 J. AM. CHEM. SOC. VOL. 129, NO. 24, 2007

Sugar-Assisted Ligation in Glycoprotein Synthesis
A R T I C L E S
Figure 3. HPLC analysis of the selective desulfurization of glycopeptide 11 and MALDI-TOF/MS spectrum of glycopeptide 12 (desired product).
20 mg of Pd/Al2O3, respectively. The progress of the desulfu-
rization in these reactions was carefully monitored by LC/MS.
After 2.5 h, the desired product 12, in which the mass
corresponds to a loss of 32 Da from compound 11, was observed
as the major peak on the HPLC chromatograph (Figure 3). The
analytical data from these three reactions showed that the use
of 15 mg of Pd/Al2O3 gave the best selectivity and afforded
the product in 77% isolated yield. The reaction using 20 mg of
Pd/Al2O3 completed within 1 h; however, it gave poor desulfu-
rization selectivity [12/13, ∼3:1]. To examine the molecular
structure of the selectively desulfurized product 12, we removed
its Acm group and then subjected the product to NCL with the
peptide thioester Cys(Acm)37-Phe82 (see Supporting Informa-
tion). After 4 h, the clean ligated product as shown in the
analytical HPLC chromatograph suggested the selective des-
ulfurization would allow the SAL product to be sequentially
ligated to another peptide or glycopeptide. In addition, we also
tested the desulfurization selectivity in the glycopeptide contain-
ing two Cys(Acm) residues and a 46-residue glycopeptide Cys-
(Acm)37-Phe82 (see Supporting Information). Similar desulfu-
rization selectivity was obtained in these model studies;
however, we noticed that the desulfurization efficiency in long
glycopeptide decreased, which in turn affected its desulfurization
selectivity (see Supporting Information).32
potentially useful approach for the synthesis of glycoproteins,
as we demonstrated by the synthesis of diptericin ꢀ. Our
synthetic scheme combined the advantages given by SAL, NCL,
and desulfurization. Here, SAL aided us to prepare the 46-
resdiue C-terminal glycopeptide, which after Acm removal, was
subsequently ligated to the 36-residue N-terminal segment by
NCL to achieve the full-sized diptericin ꢀ. In the final step, the
desulfurization reaction was used to remove the thiol handle
from the sugar moiety and to convert the cysteine into the native
alanine residue. Moreover, we have demonstrated the sulfhydryl-
modified sugar moiety can be selectively desulfurized in the
presence of Cys(Acm), which expands the utility of SAL into
the synthesis of cysteine-containing glycopeptides or glycopro-
teins. Based on these results, we believe that SAL, in conjunction
with other ligation methods, should facilitate the synthesis of
homogeneous glycoproteins, which will help to elucidate the
effects of specific glycans on protein structure and function.33
Experimental Procedures
General Methods and Materials. Tetrahydrofuran (THF) was
distilled over sodium/benzophenone, and methylene chloride (CH2Cl2)
was distilled over calcium chloride. Reagents of commercial quality
were purchased and used without further purification. Glycosylation
experiments were performed using molecular sieves (AW 300), which
were flame-dried under high vacuum right before the reaction.
Analytical thin-layer chromatography (TLC) was performed using silica
gel 60 F254 glass plates, and compound spots were visualized by UV
light (254 nm) and by staining with acidic ceric ammonium molybdate.
Flash chromatography was performed on silica gel 60 Geduran (35-
75 µm, EM Science). 1H, 13C NMR spectra were recorded on a Bruker
AMX-500 MHz spectrometer. Coupling constants (J) are reported in
hertz (Hz), and chemical shifts are reported in parts per million (ppm)
relative to tetramethylsilane (TMS, 0.0 ppm). Fmoc-amino acids, Boc-
amino acids, HBTU, PyBOP, preloaded 2-ClTrt resins, MBHA resin
LL (100-200 mesh)‚HCl and Rink-amide resin were purchased from
Novabiochem. Boc-His(3-Bom)-OH was purchase from Bachem. HOBt,
N,N-diisopropylethylamine (DIPEA), N,N-dimethylformamide (Biotech
grade), HPLC grade acetonitrile, and trifluoroacetic acid were purchased
from Sigma-Aldrich and Fisher.
Conclusion
We have shown that SAL can be extended to the synthesis
of R-O-linked glycopeptides, in which the ligation rates are
similar to those of â-O-linked and N-linked glycopeptides,
making it a general method for various glycopeptide syntheses.
For the first time, we have demonstrated that SAL could be a
(32) While we were preparing this manuscript, Kent’s group reported selective
desulfurization of cysteine in peptides containing Cys(Acm) residues
(Pentelute, B. L.; Kent, S. B. H. Org. Lett. 2007, 9, 687-690). In their
report, Raney nickel was used as the catalyst. In our studies, we found that
the thiol handle of the sugar moiety reacts much more rapidly towards
desulfurization than the cysteine residue (data not shown), when Pd/Al2O3
was used as the catalyst. This indicates that even better desulfurization
selectivity might be achieved if different catalysts, like Raney nickel, or
other conditions were used to selectively desulfurize the thiol handle in
the presence of Cys(Acm) residues.
(33) Hsieh-Wilson, L. C. Nature 2007, 445, 31-33.
9
J. AM. CHEM. SOC. VOL. 129, NO. 24, 2007 7697

A R T I C L E S
Yang et al.
1
Mass Spectrometry. MALDI-TOF/MS (Applied Biosystems DE)
and ESI-TOF (IonSpec Ultima FTMS) mass spectra were obtained from
the routine analyses service at the Scripps Center of Mass Spectrometry.
LC/MS mass spectrometric analyses were performed on the Agilent
1100 MSD LC/MS system which utilizes the electrospray ionization
method.
mmol, 94%). H NMR (CDCl3, 500 MHz): δ 7.76 (d, 2H, J ) 7.4
Hz), 7.64 (d, 2H, J ) 5.9 Hz), 7.38-7.41 (m, 2H), 7.30-7.35 (m,
2H), 5.90-5.98 (m, 1H), 5.77 (d, 1H, J ) 9.2 Hz), 5.34-5.38 (m,
2H), 5.27 (d, 1H, J ) 10.7 Hz), 4.97 (d, 1H, J ) 3.0 Hz), 4.88-4.92
(m, 1H), 4.65-4.74 (m, 2H), 4.41-4.49 (m, 3H), 4.23-4.33 (m, 3H),
4.08-4.10 (m, 2H), 3.15 (dd, 1H, J ) 10.7, 3.0 Hz), 2.11 (s, 3H),
2.04 (s, 3H), 2.03 (s, 3H), 1.35 (d, 3H, J ) 6.2 Hz); 13C NMR (CDCl3,
125 MHz): δ 170.5, 170.3, 170.2, 170.1, 156.5, 143.8, 143.7, 141.2,
131.3, 127.6, 127.0, 125.1, 125.1, 119.9, 119.9, 119.3, 101.8, 77.3,
71.8, 67.4, 67.3, 67.2, 66.4, 62.1, 58.6, 49.7, 47.0, 20.7, 20.6, 20.5,
18.1. HRMS (ESI-TOF) calcd for C34H40N2O12 [M + H]+: 669.2654.
Found: 669.2641. The amine (650 mg, 0.973 mmol) was dissolved in
DMF (15 mL). TrtS-CH2COOH (651 mg, 1.95 mmol), HBTU (740
mg, 1.95 mmol), and DIPEA (678 µL, 3.89 mmol) were added to the
amine subsequently. The reaction mixture was stirred at room temper-
ature for 2 h. Once the reaction was done, the reaction solution was
concentrated under vacuum. The concentrated residue was then purified
by flash column chromatography (20-40% EtOAc in hexane) to give
Solid-Phase Peptide Synthesis. Solid-phase peptide syntheses were
carried out manually with syringes (Torviq) which were equipped with
Teflon filters. Preparative, semipreparative, and analytical reversed-
phase HPLC purifications were performed on a Hitachi D-7000 HPLC
system. All C18 HPLC columns were purchased from Grace Vydac.
The flow rates used for HPLC were 1.5 mL/min (analytical), 4.0 mL/
min (semipreparative), and 8.0 mL/min (preparative). Fmoc-SPPS:
Fmoc CleaVage: After treatment with 20% piperidine in DMF (2 ×
10 min), the resin was washed with DMF (4×), CH2Cl2 (4×), and DMF
(4×). Coupling: After preactivation of Fmoc-amino acid (5 equiv) for
5 min using HBTU (4 equiv, final concentration 0.5 M in DMF), HOBt
(2 equiv), and DIPEA (10 equiv), the solution was added to the resin.
After 1 h, the resin was washed with DMF (4×), CH2Cl2 (2×), and
DMF (2×). Glycosylated amino acid was coupled to the resin for 24
h. Double couplings were performed for some bulky amino acids. Acetyl
Group RemoVal: The resin was first washed with MeOH (10×) and
then treated with hydrazine/MeOH (1/6) (3 × 2 h). The resulting resin
was washed with MeOH (10×), DMF (5×), and CH2Cl2 (5×). Boc-
SPPS: Boc CleaVage: After treatment with 5% m-cresol/TFA (2 × 4
min), the resin was washed with DCM (4×) and with DMF (4×).
Coupling: After preactivation of Boc-amino acid (5 equiv) for 5 min
using PyBOP (4 equiv, final concentration 0.5 M in DMF) and
N-methylmorpholine (8 equiv), the solution was added to the resin.
After 45 min, the resin was washed with DMF (3×) and DCM (4×).
CleaVage: HF with 10% (v/v) anisole was used as cleavage solution.
Once the HF was removed, cold ethyl ether was added to the crude
reaction product and stirred for 10 min. The precipitate was collected
and redissolved in H2O/CH3CN (9/1) for HPLC purification.
1
pure compound 3a (877.7 mg, 0.892 mmol, 91%). H NMR (CDCl3,
500 MHz): δ 8.00 (brs, 1H), 7.77 (d, 2H, J ) 7.3 Hz), 7.63 (d, 2H, J
) 7.3 Hz), 7.16-7.47 (m, 19H), 5.95-6.01 (m, 1H), 5.72-5.82 (m,
2H), 5.39 (brs, 1H), 5.20-5.25 (m, 2H), 5.03-5.08 (m, 1H), 4.86-
4.89 (m, 1H), 4.37-4.50 (m, 6H), 4.19-4.28 (m, 3H), 4.04-4.13 (m,
2H), 2.14 (s, 3H), 2.03 (s, 3H), 1.95 (s, 3H), 1.28 (d, 3H, J ) 6.2 Hz);
13C NMR (CDCl3, 125 MHz): δ 171.2, 170.7, 170.6, 170.4, 169.2,
163.0, 156.9, 144.5, 144.2, 144.1, 141.7, 131.4, 129.9, 128.4, 128.1,
127.5, 127.3, 125.5, 125.4, 120.5, 120.4, 120.1, 99.8, 68.6, 67.7, 67.6,
67.6, 66.6, 62.5, 58.8, 48.2, 47.6, 39.0, 37.1, 36.9, 31.8, 21.2, 21.1,
21.0, 18.5. HRMS (ESI-TOF) calcd for C55H56N2O13S [M+Na]+:
1007.3395. Found: 1007.3390.
Compound 4a. Compound 3a (750 mg, 0.762 mmol) was dissolved
in THF (20 mL). Pd (PPh3)4 (88.0 mg, 0.076 mmol) and N-
methylaniline (827 µL, 7.62 mmol) were added, and then the reaction
mixture was stirred at room temperature for 45 min. Once the reaction
was done, the solution was concentrated under vacuum. The crude
product was purified by flash column chromatography (50% EtOAc
in hexane then 10% MeOH in CH2Cl2) to give pure compound 4a (690
Compound 2aR. Compound 1 (5.0 g, 10.5 mmol) and Fmoc-Thr-
OAllyl (4.83 g, 12.7 mmol) were mixed and dried under high vacuum
overnight prior to the reaction. The mixture was then dissolved in
anhydrous CH2Cl2 (30 mL) under N2. 3.0 g of fresh flame-dried
molecule sieves were added, and the mixture was stirred at room
temperature for 1 h. The reaction was then cooled down to -78 °C,
and TMSOTf (203.9 µL, 1.066 mmol) was added slowly into the
reaction solution. The mixture was kept at -78 °C and stirred for 45
min. Once the reaction was done, it was quenched by adding Et3N,
and the temperature was allowed to rise to room temperature. Molecule
sieves were removed by filtration through Celite, and the filtrate was
concentrated under vacuum. The crude product was purified by flash
column chromatography (20-50% EtOAc in hexane) to give the pure
1
mg, 0.731mmol, 95%). H NMR (MeOD, 500 MHz): δ 7.75 (d, 2H,
J ) 7.7 Hz), 7.57-7.63 (m, 3H), 7.47-7.50 (m, 1H), 7.31-7.33 (m,
6H), 7.12-7.26 (m, 11H), 5.38 (brs, 1H), 5.09 (dd, 1H, J ) 11.4, 2.9
Hz), 4.96 (brs, 1H), 4.24-4.45 (m, 6H), 4.13-4.16 (m, 2H), 4.03-
4.09 (m, 3H), 2.99 (d, 1H, J ) 13.2 Hz), 2.91 (d, 1H, J ) 13.2 Hz),
2.10 (s, 3H), 1.96 (s, 3H), 1.91 (s, 3H), 1.19 (d, 3H, J ) 7.0 Hz); 13C
NMR (MeOD, 125 MHz): δ 172.2, 172.1, 171.9, 171.9, 158.9, 145.7,
145.5, 145.1, 142.7, 142.7, 133.8, 133.8, 133.2, 133.1, 130.8, 130.1,
130.0, 129.1, 128.9, 128.3, 128.1, 126.2, 126.1, 121.1, 121.1, 100.3,
77.6, 69.7, 68.9, 68.3, 68.2, 67.8, 63.4, 61.6, 60.3, 37.6, 21.0, 21.0,
20.7, 20.7, 19.2, 14.6. HRMS (ESI-TOF) calcd for C52H52N2O13S [M
+ Na]+: 967.3082. Found: 967.3059.
Compound 2bR. Compound 1 (1.33 g, 2.81 mmol) and Fmoc-Ser-
OAllyl (858.8 mg, 2.34 mmol) were mixed together and dried under
high vacuum overnight prior to the reaction. The mixture was then
dissolved in anhydrous CH2Cl2 (15 mL) under N2. To this reaction
mixture, 1.5 g of fresh flame-dried molecule sieves were added;
afterward, it was stirred at room temperature for 1 h. The reaction
mixture was then cooled down to -78 °C, followed by slowly adding
TMSOTf (54.3 µL, 0.281 mmol) into the reaction solution. The reaction
was kept at -78 °C and stirred for 45 min. Once the reaction was
done, it was quenched by adding Et3N and the temperature was allowed
to rise to room temperature. Molecule sieves were then filtered off
through Celite, and the filtrate was concentrated under vacuum. The
crude product was purified by flash column chromatography (20-50%
EtOAc in hexane) to give pure R product (872 mg, 1.28 mmol, 45%).
1H NMR (CDCl3, 500 MHz): δ 7.76 (d, 2H, J ) 7.7 Hz), 7.62 (dd,
2H, J ) 7.0, 3.7 Hz), 7.40 (dd, 2H, J ) 7.7, 7.7 Hz), 7.32 (td, 2H, J
) 7.7, 3.3 Hz), 5.94 (m, 2H), 5.44 (brs, 1H), 5.28-5.38 (m, 3H), 4.94
1
R product (3.31 g, 4.77 mmol, 45.4%). H NMR (CDCl3, 500 MHz):
δ 7.76 (d, 2H, J ) 7.4 Hz), 7.62-7.64 (m, 2H), 7.36-7.41 (m, 2H),
7.30-7.34 (m, 2H), 5.90-5.99 (m, 1H), 5.68 (d, 1H, J ) 9.6 Hz),
5.46 (brs, 1H), 5.26-5.39 (m, 3H), 5.05 (d, 1H, J ) 3.3 Hz), 4.69 (d,
1H, J ) 5.2 Hz), 4.41-4.48 (m, 3H), 4.36 (dd, 1H, J ) 10.3, 7.3 Hz),
4.26-4.29 (m, 2H), 4.09 (d, 1H, J ) 6.6 Hz), 3.67 (dd, 1H, J ) 11.4,
3.7 Hz), 2.14 (s, 3H), 2.07 (s, 3H), 2.03 (s, 3H), 1.36 (d, 3H, J ) 6.2
Hz); 13C NMR (CDCl3, 125 MHz): δ 170.2, 169.9, 169.7, 156.7, 143.7,
143.6, 141.2, 141.1, 131.2, 127.6, 127.0, 125.2, 125.1, 119.9, 119.8,
119.2, 99.3, 76.9, 68.1, 67.4, 67.3, 66.9, 66.4, 61.7, 58.7, 57.6, 47.0,
20.6, 20.5, 20.4, 18.4. HRMS (ESI-TOF) calcd for C34H38N4O12 [M +
H]+: 695.2559. Found: 695.2553.
Compound 3a. Compound 2aR (2.68 g, 3.86 mmol) was dissolved
in AcOH (25 mL), followed by the addition of zinc powder (15.0 g).
The reaction mixture was stirred for 4 h at room temperature. Once
the reaction was done, zinc was filtered off through Celite and washed
with CH2Cl2. The filtrate was concentrated under vacuum. The
concentrated residue was then purified by flash column chromatography
(20% hexane in EtOAc) to give the pure amine product (2.40 g, 3.62
9
7698 J. AM. CHEM. SOC. VOL. 129, NO. 24, 2007

Sugar-Assisted Ligation in Glycoprotein Synthesis
A R T I C L E S
(d, 1H, J ) 3.3 Hz), 4.69-4.71 (m, 2H), 4.59-4.60 (m, 1H), 4.41 (d,
2H, J ) 7.0 Hz), 4.25 (t, 1H, J ) 7.3 Hz), 4.19 (t, 1H, J ) 6.6 Hz),
4.13 (dd, 1H, J ) 10.6, 2.9 Hz), 4.00-.4.04 (m, 2H), 3.63 (dd, 1H, J )
11.0, 3.3 Hz), 2.14 (s, 3H), 2.06 (s, 3H), 1.97 (s, 3H). 13C NMR (CDCl3,
125 MHz): δ 170.3, 169.8, 169.6, 169.1, 155.7, 143.7, 141.2, 131.2,
127.7, 127.6, 127.0, 125.0, 119.9, 119.2, 99.2, 69.7, 67.8, 67.4, 67.2,
67.1, 66.5, 61.6, 57.3, 54.4, 47.0, 20.5, 20.5, 20.4. HRMS (ESI-TOF)
calcd for C33H36N4O12 [M + Na]+: 703.2222. Found: 703.2208.
was then dissolved in 60 mL of dry DMF at 0 °C and fully protected
with PMB groups by adding PMBCl (20.5 mL, 151 mmol) and NaH
(3.62 g, 151 mmol). The reaction mixture was stirred at room
temperature for 2 h and quenched with MeOH. After removing MeOH,
it was concentrated under vacuum. Then the product was extracted with
EtOAc and washed with H2O. The organic layers were collected and
washed again with brine. After concentrating the organic layer under
vacuum, the crude residue was purified by flash column chromatography
(10-40% EtOAc in hexane) to give the pure PMB protected thiogly-
coside 8R and 8â, respectively [15.5 g, 23.2 mmol (R + â), 92%]. 1H
NMR (CDCl3, 500 MHz): δ 7.37 (d, 2H, J ) 8.1 Hz), 7.33 (d, 2H, J
) 8.8 Hz), 7.19 (d, 2H, J ) 8.5 Hz), 7.18 (d, 2H, J ) 8.8 Hz), 7.02
(d, 2H, J ) 7.8 Hz), 6.90 (d, 2H, J ) 8.4 Hz), 6.86 (d, 2H, J ) 8.5
Hz), 6.81 (d, 2H, J ) 8.5 Hz), 5.50 (d, 1H, J ) 5.5 Hz), 4.80 (d, 1H,
J ) 11.0 Hz), 4.63-4.67 (m, 2H), 4.32-4.47 (m, 5H), 3.98 (brs, 1H),
3.80 (s, 3H), 3.79 (s, 3H), 3.77 (s, 3H), 3.74 (dd, 1H, J ) 10.6, 2.6
Hz), 3.55 (dd, 1H, J ) 9.2, 7.0 Hz), 3.47 (dd, 1H, J ) 9.2, 5.9 Hz),
2.29 (s, 3H); 13C NMR (CDCl3, 125 MHz): δ 159.4, 159.3, 159.2,
137.6, 132.7, 130.4, 129.9, 129.8, 129.7, 129.6, 129.5, 129.4, 129.3,
113.9, 113.7, 113.6, 87.9, 78.7, 74.4, 73.1, 73.0, 72.0, 70.4, 68.3, 60.3,
55.3, 55.2, 55.1, 21.0. HRMS (ESI-TOF) calcd for C37H41N3O7S [M
+ Na]+: 694.2557. Found: 694.2556
Compound 3b. The amine (822 mg, 1.26 mmol) [HRMS (ESI-TOF)
calcd for C33H38N2O12 [M + H]+: 655.2497. Found: 655.2499] was
dissolved in DMF (10 mL). TrtS-CH2COOH (840 mg, 2.51 mmol),
HBTU (953 mg, 2.51 mmol), and DIPEA (875 µL, 5.03 mmol) were
added to the amine subsequently. The reaction was stirred at room
temperature for 2 h. Once the reaction was done, the solution was
concentrated under vacuum. The concentrated residue was then purified
by flash column chromatography (20-40% EtOAc in hexane) to give
1
pure compound 3b (1.03 g, 1.06 mmol, 84%). H NMR (CDCl3, 500
MHz): δ 7.76 (d, 2H, J ) 7.3 Hz), 7.60 (d, 2H, J ) 5.2 Hz), 7.36-
7.41 (m, 8H), 7.25-7.32 (m, 8H), 7.19-7.22 (m, 3H), 5.90-5.96 (m,
2H), 5.78-5.86 (m, 1H), 5.35 (d, 1H, J ) 2.9 Hz), 5.27 (d, 1H, J )
17.3 Hz), 5.23 (d, 1H, J ) 10.3 Hz), 4.98 (dd, 1H, J ) 11.4, 2.6 Hz),
4.71 (brs, 1H), 4.41-4.56 (m, 6H), 4.22 (t, 1H, J ) 6.6 Hz), 4.06-
4.11 (m, 2H), 3.98-4.01 (m, 2H), 3.85-3.87 (m, 1H), 2.13 (s, 3H),
1.96 (s, 3H), 1.93 (s, 3H); 13C NMR (CDCl3, 125 MHz): δ 170.4, 170.2,
170.0, 169.2, 168.5, 155.6, 143.8, 143.6, 141.2, 131.0, 129.3, 128.0,
127.6, 127.0, 124.9, 119.9, 119.3, 98.8, 69.8, 68.0, 67.6, 67.2, 67.0,
66.2, 61.8, 54.3, 47.6, 46.9, 38.5, 36.3, 20.6, 20.5, 20.4. HRMS (ESI-
TOF) calcd for C54H54N2O13S [M + Na]+: 993.3239. Found: 993.3240.
1
Compound 8â. H NMR (CDCl3, 500 MHz): δ 7.43 (d, 2H, J )
7.7 Hz), 7.27 (d, 2H, J ) 8.4 Hz), 7.18 (d, 2H, J ) 8.4 Hz), 7.13 (d,
2H, J ) 8.4 Hz), 6.96 (d, 2H, J ) 7.7 Hz), 6.80-6.87 (m, 6H), 4.75
(d, 1H, J ) 11.0 Hz), 4.60 (d, 1H, J ) 11.0 Hz), 4.54 (d, 1H, J ) 11.3
Hz), 4.43 (d, 1H, J ) 11.0 Hz), 4.40 (d, 1H, J ) 11.3 Hz), 4.33 (d,
1H, J ) 11.8 Hz), 4.31 (d, 1H, J ) 10.3 Hz), 3.86 (d, 1H, J ) 1.8
Hz), 3.73-3.78 (m, 11H), 3.57 (d, 1H, J ) 6.6 Hz), 3.49 (dd, 1H, J )
6.3, 6.3 Hz), 3.34 (dd, 1H, J ) 9.5, 2.2 Hz), 2.26 (s, 3H); 13C NMR
(CDCl3, 125 MHz): δ 159.2, 159.1, 158.8, 137.7, 133.0, 130.5, 129.7,
129.4, 129.3, 129.2, 129.1, 128.0, 113.7, 113.6, 113.3, 86.3, 82.0, 77.2,
73.8, 73.0, 71.7, 71.5, 68.0, 61.3, 55.0, 20.9. HRMS (ESI-TOF) calcd
for C37H41N3O7S [M + Na]+: 694.2557. Found: 694.2556.
Compound 4b. Compound 3b (420 mg, 0.433 mmol) was dissolved
in THF (10 mL), followed by adding Pd(PPh3)4 (50.0 mg, 0.043 mmol)
and N-methylaniline (470 µL, 4.33 mmol). The reaction mixture was
stirred at room temperature for 45 min. Once the reaction was done,
the reaction solution was concentrated under vacuum. The crude product
was purified by flash column chromatography (50% EtOAc in hexane
and then 10% MeOH in CH2Cl2) to give pure compound 4b (390 mg,
Fmoc-Thr-OAllyl. Cesium carbonate (2.28 g, 7.0 mmol) was added
to a suspension of Fmoc-Thr-OH (4.75 g, 13.9 mmol) in dry MeOH
(40 mL) under N2(g). The reaction mixture was stirred at room
temperature for 2 h and then evaporated to dryness under vacuum. The
concentrated residue was further dried under high vacuum for another
2 h. Afterward, the dried mixture was redissolved in dry DMF (40
mL) under N2 and added with allylbromide (1.45 mL, 16.7 mmol).
The reaction mixture was stirred at room temperature for 8 h. The white
precipitate was filtered off through Celite and washed with CH2Cl2.
The filtrate was concentrated under vacuum and then was purified by
flash column chromatography (25-50% EtOAc in hexane) to give a
1
0.420 mmol, 96%). H NMR (MeOD, 500 MHz): δ 7.76-7.78 (m,
2H), 7.60-7.65 (m, 2H), 7.16-7.41 (m, 19H), 5.36 (dd, 1H, J ) 2.6
Hz), 5.11-5.14 (m, 1H), 4.82 (brs, 1H), 4.35 (dd, 1H, J ) 11.4, 3.3
Hz), 4.27-4.32 (m, 3H), 4.22 (t, 1H, J ) 6.6 Hz), 4.17 (t, 1H, J ) 6.6
Hz), 4.05 (dd, 1H, J ) 11.0, 6.3 Hz), 3.99 (dd, 1H, J ) 11.0, 7.0 Hz),
3.91-3.93 (m, 1H), 3.82 (dd, 1H, J ) 10.3, 5.1 Hz), 2.82-2.93 (m,
2H), 2.13 (s, 3H), 1.92 (s, 3H), 1.89 (s, 3H); 13C NMR (MeOD, 125
MHz): δ 172.3, 172.2, 171.9, 171.8, 158.4, 146.0, 145.7, 145.4, 145.3,
142.7, 133.9, 133.2, 133.1, 130.8, 130.1, 130.0, 129.9, 129.2, 129.1,
128.9, 128.8, 128.3, 128.1, 128.0, 126.3, 121.1, 99.8, 70.4, 69.8, 68.8,
68.4, 68.2, 68.1, 63.0, 57.2, 37.5, 20.9, 20.7. HRMS (ESI-TOF) calcd
for C51H50N2O13S [M + Na]+: 953.2926. Found: 953.2892.
Compound 8R. Compound 7 (10.28 g, 27.55 mmol) and p-thiocresol
(5.1 g, 41.3 mmol) were dissolved in CH2Cl2 (40 mL) under N2(g). The
reaction solution was cooled down to 0 °C in ice bath. To this solution,
BF3‚OEt2 (6.92 mL, 55.1 mmol) was added slowly; afterward, the
reaction temperature was allowed to gradually increase to room
temperature. After 8 h, the reaction was quenched by slow addition of
saturated NaHCO3 and diluted with CH2Cl2. The product was washed
with saturated NaHCO3 and extracted with CH2Cl2. The organic layers
were collected and concentrated under vacuum. The concentrated
residue was then purified by flash column chromatography (20-40%
EtOAc in hexane) to give the pure product (R and â mixtures, 10.3 g,
23.6 mmol, 85%). The purified thioglycoside was dissolved in MeOH.
NaOMe (25 wt % in MeOH) was added dropwise to the reaction
solution until the pH reached around 11-12. The reaction mixture was
stirred at room temperature for 1 h and neutralized with acidic resin
(Dowex 50WX2-200(H)). The resin was then filtered off and washed
with CH2Cl2. The filtrate was concentrated under vacuum. The
concentrated residue was then purified by flash column chromatography
(5-10% MeOH in CH2Cl2). The pure thioglycoside (7.83 g, 25.2 mmol)
1
pure white powder (4.64 g, 12.2 mmol, 87%). H NMR (CDCl3, 500
MHz): δ 7.72 (d, 2H, J ) 7.4 Hz), 7.59 (d, 1H, J ) 5.9 Hz), 7.58 (d,
1H, J ) 5.9 Hz), 7.36 (d, 1H, J ) 7.3 Hz), 7.35 (d, 1H, J ) 7.7 Hz),
7.27 (d, 1H, J ) 7.3 Hz), 7.25 (d, 1H, J ) 7.4 Hz), 5.92 (d, 1H, J )
9.2 Hz), 5.82-5.89 (m, 1H), 5.29 (d, 1H, J ) 17.2 Hz), 5.20 (d, 1H,
J ) 10.3 Hz), 4.62-4.64 (m, 2H), 4.36-4.41 (m, 4H), 4.20 (t, 1H, J
) 7.0 Hz), 1.23 (d, 3H, J ) 5.9 Hz); 13C NMR (CDCl3, 125 MHz): δ
170.8, 156.7, 143.6, 143.5, 141.1, 131.2, 127.5, 126.9, 125.0, 119.8,
118.7, 67.7, 67.0, 66.0, 59.1, 46.9, 19.7. HRMS (ESI-TOF) calcd for
C22H23NO5 [M+Na]+: 404.1468. Found: 404.1468.
Compound 9R. Compound 8 (917 mg, 1.366 mmol) and Fmoc-
Thr-OAllyl (624.7 mg, 1.639 mmol) were mixed and dried under high
vacuum overnight prior to the reaction. The reaction mixture was
dissolved in dry CH2Cl2 under N2. The fresh flame-dried molecular
sieves (AW 300) (3 g) were added to the reaction mixture and stirred
for about 2 h; afterward, the reaction was cooled to -20 °C, followed
by addition of N-iodosuccinimide (338.1 mg, 1.503 mmol). To this
mixture, freshly prepared TfOH (0.05 equiv) was slowly added at
-20 °C. The reaction was stirred at -15 °C for 1 h. Once the reaction
was done, it was quenched by adding saturated Na2S2O3(aq) and saturated
NaHCO3(aq). The molecular sieves were filtered off through Celite and
9
J. AM. CHEM. SOC. VOL. 129, NO. 24, 2007 7699

A R T I C L E S
Yang et al.
washed with CH2Cl2. The filtrate was extracted with CH2Cl2 and washed
with saturated Na2S2O3(aq) and saturated NaHCO3(aq). After drying over
Na2SO4, the filtrate was evaporated to dryness under vacuum. The
concentrated residue was purified by flash column chromatography
(25-50% EtOAc in hexane) to give R product (506.7 mg, 0.546 mmol,
40%). 1H NMR (CDCl3, 600 MHz): δ 7.76 (d, 2H, J ) 7.9 Hz), 7.64
(d, 1H, J ) 7.4 Hz), 7.63 (d, 1H, J ) 7.4 Hz), 7.39 (d, 1H, J ) 7.4
Hz), 7.38 (d, 1H, J ) 7.4 Hz), 7.33 (d, 2H, J ) 8.3 Hz), 7.31 (d, 1H,
J ) 7.4 Hz), 7.30 (d, 1H, J ) 7.4 Hz), 7.20 (d, 2H, J ) 8.3 Hz), 7.17
(d, 2H, J ) 8.3 Hz), 6.90 (d, 2H, J ) 8.3 Hz), 6.86 (d, 2H, J ) 8.3
Hz), 6.81 (d, 2H, J ) 8.3 Hz), 5.89-5.96 (m, 1H), 5.74 (d, 1H, J )
9.2 Hz), 5.35 (d, 1H, J ) 17.1 Hz), 5.25 (d, 1H, J ) 10.1 Hz), 4.90 (d,
1H, J ) 3.5 Hz), 4.78 (d, 1H, J ) 10.9 Hz), 4.66-4.68 (m, 2H), 4.61
(d, 1H, J ) 11.0 Hz), 4.39-4.49 (m, 5H), 4.31-4.36 (m, 2H), 4.28
(dd, 1H, J ) 7.0, 7.0 Hz), 4.01 (brs, 1H), 3.95 (dd, 1H, J ) 6.6, 6.6
Hz), 3.88 (dd, 1H, J ) 10.5, 2.2 Hz), 3.79 (s, 3H), 3.78 (s, 3H), 3.77
(s, 3H), 3.53 (dd, 1H, J ) 9.2, 7.5 Hz), 3.48 (dd, 1H, J ) 8.8, 5.7 Hz),
1.30 (d, 3H, J ) 5.5 Hz); 13C NMR (CDCl3, 150 MHz): δ 169.9,
163.1, 159.4, 159.3, 159.2, 156.7, 143.9, 143.6, 141.2, 131.3, 130.3,
129.8, 129.7, 129.5, 129.4, 129.3, 127.6, 127.1, 127.0, 125.2, 125.1,
119.9, 119.3, 113.8, 113.7, 113.6, 99.5, 76.0, 74.7, 73.4, 73.1, 71.8,
69.9, 68.5, 67.3, 66.5, 59.5, 58.8, 55.1, 47.0, 18.7. HRMS (ESI-TOF)
calcd for C52H56N4O12 [M + H]+: 929.3967. Found: 929.3954.
18.4. HRMS (ESI-TOF) calcd for C51H56N2O13 [M + Na]+: 927.3674.
Found: 927.3680.
Solid-Phase Synthesis of Glycopeptide Val53-Phe82. The synthesis
of the C-terminal glycopeptide segment (53-82) was started with H2N-
Phe-2-ClTrt-resin and carried out by using Fmoc-based strategy. For
coupling the glycosylated amino acid onto the resin, we used 2 equiv
of building block 4a and left the reaction to shake for 24 h. Once the
SPPS was completed, the acetyl groups on the sugar were removed on
the solid support by treating with hydrazine/MeOH (1/6) for 6 h.
Following this O-acetate removal step, the resin was washed with
MeOH, DMF, and CH2Cl2. The resulting glycopeptide was deprotected
and cleaved from the resin by treating with TFA/H2O/Et3SiH/thioanisole
(17/1/1/1) for 50 min at room temperature. The crude glycopeptide
solution was evaporated to remove all the cleavage cocktail solution,
and then the dried residue was redissolved in H2O/CH3CN (1/1) for
HPLC purification. The purified glycopeptide was lyophilized to give
pure glycopeptide (18% yield). ESI-MS: 3528 Da (the peptide mass
was reconstructed from the experimental mass-to-charge (m/z) ratios
from all of the observed protonation states of the peptide).
Synthesis of Thioester Peptide Cys(Acm)37-Gly52. The synthesis
of the thioester peptide was begun with the preloading of 3-(tritylthio)
propanoic acid onto MBHA resin LL. The detailed preloading procedure
can be referred to in the Supporting Information for J. Am. Chem. Soc.
2006, 128, 5626-5627.
Solid-phase peptide synthesis was carried out by using the Boc
strategy. After the full sequence was completed on the solid support,
we removed the N-terminal Boc group to minimize side reactions during
HF cleavage. The resin was then dried under high vacuum for 4 h prior
to the final cleavage by HF with 10% (v/v) of anisole. The crude
thioester peptide was purified by HPLC to give the pure product in
43% yield. ESI-MS: 1982 Da (the peptide mass was reconstructed
from the experimental mass-to-charge (m/z) ratios from all of the
observed protonation states of the peptide).
Compound 10. Compound 9R (200 mg, 0.215 mmol) was dissolved
in pyridine (1.2 mL) at 0 °C, followed by addition of AcSH (1.2 mL).
After stirring for 1 min, the reaction temperature was raised to room
temperature and the reaction mixture was continued to stir for 4 h.
Once the reaction was done, the mixture was evaporated to dryness
under vacuum. The concentrated residue was purified by flash column
chromatography (20%-50% EtOAc in hexane and then 30% hexane
in EtOAc) to give pure product for the next allyl removal step (174.7
1
mg, 0.185 mmol, 86%). H NMR (CDCl3, 600 MHz): δ 7.77 (d, 2H,
J ) 7.4 Hz), 7.63 (d, 1H, J ) 6.5 Hz), 7.61 (d, 1H, J ) 6.1 Hz), 7.40
(d, 1H, J ) 7.9 Hz), 7.38 (d, 1H, J ) 7.9 Hz), 7.33 (d, 1H, J ) 7.9
Hz), 7.30 (d, 1H, J ) 7.4 Hz), 7.20-7.25 (m, 8H), 6.80-6.87 (m,
4H), 5.81 (m, 1H), 5.23-5.32 (m, 4H), 4.88 (d, 1H, J ) 11.0 Hz),
4.80 (d, 1H, J ) 3.5 Hz), 4.55-4.64 (m, 4H), 4.43-4.53 (m, 4H),
4.33-4.39 (m, 2H), 4.27 (dd, 1H, J ) 6.6, 6.6 Hz), 4.16 (m, 1H), 3.94
(brs, 1H), 3.87 (dd, 1H, J ) 6.6, 6.6 Hz), 3.79 (s, 3H), 3.77 (s, 3H),
3.73 (s, 3H), 3.47-3.55 (m, 2H), 1.94 (s, 3H), 1.22 (d, 3H, J ) 6.5
Hz); 13C NMR (CDCl3, 150 MHz): δ 170.5, 169.8, 162.9, 159.2, 159.1,
158.9, 156.2, 143.7, 143.4, 141.2, 141.1, 130.8, 130.5, 129.9, 129.8,
129.7, 129.6, 129.2, 129.1, 127.6, 127.0, 124.8, 124.7, 119.9, 119.8,
119.6, 113.7, 113.6, 113.4, 76.4, 76.3, 73.8, 73.1, 72.0, 71.0, 70.2, 68.5,
66.7, 66.0, 58.5, 55.1, 48.8, 47.1, 23.3, 18.3. HRMS (ESI-TOF) calcd
for C54H60N2O13 [M + Na]+: 967.3987. Found: 967.3984. In the allyl
removal step, the product from the previous step (720 mg, 0.762 mmol)
was dissolved in THF (10 mL), followed by addition of Pd(Ph3P)4 (88.1
mg, 0.076 mmol) and N-methylaniline (827 µL, 7.62 mmol) subse-
quently. The reaction mixture was stirred at room temperature for 45
min. Once the reaction was done, the reaction solution was evaporated
to dryness under vacuum. The concentrated residue was then purified
by flash column chromatography (50% EtOAc in hexane and then 10%
MeOH in CH2Cl2) to give the pure compound 10 (655 mg, 0.724 mmol,
Synthesis of Glycopeptide Thioester Asp1-Asn36. The synthesis
was based on the procedure reported by Wang et al. in Int. J. Pept.
Res. Therap. 2005, 11, 117-123. The synthesis started with Rink amide
resin. After removing the Fmoc group by treatment with 20% piperidine
(in DMF) for 20 min (×2), the resin was loaded with Fmoc-Asp-OAllyl
(4 equiv) by mixing together with the coupling solution containing
HBTU (4 equiv) and DIEA (10 equiv) in DMF for 2 h. Afterward, the
resin was washed with DMF (4 mL × 4), followed by acetylation of
free amines on the resin with Ac2O/DIEA/DMF (1:2:17) for 20 min.
Afterward, the resin was washed with DMF (4 mL × 4), DCM (4 mL
× 4), and DMF (4 mL × 4). Fmoc-SPPS was performed using HBTU/
DIEA coupling conditions. To couple the glycosylated amino acid, 2
equiv of building block 16 was used in the synthesis, and the coupling
time was elongated to 1 day. Once the full sequence of the glycopeptide
was completed on the resin, the allyl group was removed by treating
with Pd(PPh3)4 (25 mg/ 0.1 mmol resin) and PhSiH3 (10 equiv) in DCM
for 30 min (×2). After washing with DCM (4 mL × 4), DMF (4 mL
× 4), and then DCM (4 mL × 4), we transformed the free acid to
thioester by mixing the resin with the reaction solution containing ethyl
3-mercaptopropionate (24 equiv), DIEA (37.5 equiv), anhydrous HOBt
(30 equiv), DIC (30 equiv), and DCM/DMF (1/4) for 6 h (×2).
Afterward, the resin was washed with DCM (4 mL × 4), DMF (4 m
× 4), and then DCM (4 mL × 4). The glycopeptide was cleaved from
the resin with concomitant full deprotection by treatment with TFA/
H2O/thioanisole/TES (17/1/1/1). The crude product was subjected to
HPLC purification to give the pure product in 9% yield. ESI-MS: 3977
Da (the peptide mass was reconstructed from the experimental mass-
to-charge (m/z) ratios from all of the observed protonation states of
the peptide).
1
95%). H NMR (d6-DMSO, 600 MHz): δ 7. 87 (d, 2H, J ) 7.5 Hz),
7.71 (d, 2H, J ) 7.0 Hz), 7.39 (d, 1H, J ) 7.0 Hz), 7.37 (d, 1H, J )
7.0 Hz), 7.29 (t, 2H, J ) 7.0 Hz), 7.20-7.25 (m, 4H), 7.14 (d, 2H, J
) 8.3 Hz), 6.87-6.89 (m, 4H), 6.83 (d, 2H, J ) 8.3 Hz), 4.64 (dd,
2H, J ) 19.7, 11.0 Hz), 4.33-4.46 (m, 6H), 4.17-4.27 (m, 3H), 3.92-
3.96 (m, 3H), 3.71 (s, 9H), 3.66-3.70 (m, 2H), 3.42-3.50 (m, 2H),
1.90 (s, 3H), 1.09 (d, 3H, J ) 6.1 Hz); 13C NMR (d6-DMSO, 125
MHz): δ 172.8, 170.1, 158.7, 158.6, 158.6, 156.2, 143.9, 143.8, 140.8,
130.9, 130.9, 130.1, 129.3, 129.2, 128.8, 127.6, 127.0, 125.2, 125.2,
124.9, 120.1, 113.6, 113.5, 113.4, 98.7, 79.3, 79.0, 78.8, 77.2, 75.2,
73.5, 72.0, 71.0, 69.4, 68.8, 65.4, 59.3, 55.0, 54.9, 48.7, 46.8, 23.1,
General Procedure for SAL. The ligation of unprotected peptide
segments was performed as follows: Prior to the reaction, the ligation
solution (6 M Gn‚HCl, 0.2 M phosphate buffer, pH 8.5) was degassed
for 10 min. Subsequently, glycopeptide and thioester peptide were
9
7700 J. AM. CHEM. SOC. VOL. 129, NO. 24, 2007

Sugar-Assisted Ligation in Glycoprotein Synthesis
A R T I C L E S
dissolved in the ligation solution under Ar. The reaction was performed
at 37 °C and was vortexed periodically to equilibrate the reaction
mixture. Before HPLC or LC/MS analysis, TCEP (60 mM) or 10%
(v/v) of 2-mercaptoethanol was added to reduce any formed disulfide
bonds.
General Procedure for Desulfurization. Desulfurization was
performed at room temperature in 6 M Gn‚HCl, 0.2 M phosphate buffer,
pH 5.8 containing 15 mM TCEP. The buffer was degassed by bubbling
Ar through the solution for 10 min prior to use. Following the addition
of Pd/Al2O3 (15 times the weight of glycopeptide), the reaction mixture
was kept under H2 using a H2 balloon. The desulfurization reaction
was monitored by analytical HPLC chromatography and LC/MS. Once
the reaction was complete, Pd/Al2O3 was spun down by centrifuge,
and the supernatant was collected for HPLC purification.
and glycopeptide thioester segment Asp1-Asn36 (4.4 mM) was carried
out in a solution of 6 M Gn‚HCl, 200 mM phosphate buffer, pH 7.9
containing 2% (v/v) of thiophenol and 2% (v/v) benzylmercaptan at
37 °C. After 16 h, we observed that the ligation reaction was complete,
and the ligation product was confirmed by ESI-MS. After HPLC
purification, the pure product was obtained in 47% yield.
Acknowledgment. This work was supported by the NIH and
Skaggs Institute for Chemical Biology. S.F. is grateful for a
postdoctoral fellowship provided by the Deutsche Akademische
Austauschdienst (DAAD). We would like to thank Dr. Asad
Chavochi-Negad and Prof. M. Reza Ghadiri for assistance in
HF operation. We would also like to give our appreciation to
Dr. William Greenberg for editing the manuscript.
Removing Acm Group from Glycopeptide Segment Cys(Acm)37-
Phe82. Cys(Acm)37-Phe82 (3 mg) was dissolved in 1 mL of 10% AcOH
(pH 4.0) containing 30 equiv of Hg(OAc)2. The reaction mixture was
mixed well and left to sit at room temperature for 1 h. Once the reaction
was done, 120 equiv of DTT were added and the mixture was allowed
to react for 12 h to precipitate all Hg(OAc)2. The black precipitate was
spun down afterward, and the supernatant was collected for HPLC
purification.
Supporting Information Available: Complete ref 3a, com-
plete ref 3b, experimental procedures of glycopeptides synthesis,
analytical HPLC and mass data, and 1H and 13C NMR spectra.
This material is available free of charge via the Internet at
http://pubs.acs.org/JACS.
NCL of Glycopeptide Cys37-Phe82 and Glycopeptide Thioester
Asp1-Asn36. The ligation of glycopeptide segment Cys37-Phe82 (2.9 mM)
JA0708971
9
J. AM. CHEM. SOC. VOL. 129, NO. 24, 2007 7701