Sugar-assisted ligation of N-linked glycopeptides with broad sequence tolerance at the ligation junction

Article abstract of DOI:10.1021/ja065601q

A novel method for the synthesis of N-linked glycopeptides using the sugar-assisted ligation strategy from cysteine free peptides is presented. The ligation junction tolerates a variety of amino acids, favoring less hindered amino acids and those with sid

Full text of DOI:10.1021/ja065601q

Published on Web 11/01/2006
Sugar-Assisted Ligation of N-Linked Glycopeptides with
Broad Sequence Tolerance at the Ligation Junction
Ashraf Brik,*,† Simon Ficht, Yu-Ying Yang, Clay S. Bennett, and Chi-Huey Wong*
Contribution from the Department of Chemistry and the Skaggs Institute for Chemical Biology,
The Scripps Research Institute, 10550 North Torrey Pines Road, La Jolla, California 92037
Received August 7, 2006; E-mail: abrik@bgu.ac.il; wong@scripps.edu
Abstract: A novel method for the synthesis of N-linked glycopeptides using the sugar-assisted ligation
strategy from cysteine free peptides is presented. The ligation junction tolerates a variety of amino acids,
favoring less hindered amino acids and those with side chains that could serve as a general base in the
ligation pathway. Since our approach allows the ligation of difficult junctions, the method could be applied
to the synthesis of large peptides by enzymatic removal of the sugar moiety. Alternatively, more complex
glycopeptides can be synthesized using glycosyltransferases. Together, this sequence of reactions should
be amenable to the synthesis of glycopeptides and glycoproteins and their deglycosylated products.
Introduction
also been investigated and successfully applied to the synthesis
of large peptides.6
The use of auxiliaries to assist chemical ligation in the context
of glycopeptide and glycoprotein synthesis has recently become
a major goal for several research groups. Macmillan and co-
workers investigated the applicability of 4,5,6-trimethoxy-2-
mercaptobenzyl and 1-(2,4-dimethoxyphenyl)-2-mercaptoethyl
auxiliaries for a cysteine-free NCL for glycopeptide synthesis.7
The Danishefsky group successfully used the 4,5,6-trimethoxy-
2-mercaptobenzyl auxiliary6 for the construction of complex
glycopeptides.8 Recently this group reported a similar approach
that takes advantage of an OfS followed by SfN migration
with the attachment of both coupling partner to a benzylic
auxiliary.9 These elegant examples have demonstrated the power
of auxiliaries in a cysteine-free glycopeptide ligation. In these
studies, however, only ligation at Gly-Gly and Gly-AA (AA
represents unhindered amino acids) junctions delivered the
desired products, limiting their scope in the synthesis of proteins
and glycoproteins. As a result, the development of auxiliaries
that can assist the ligation of cysteine-free glycopeptides with
tolerance to the ligation junction would be of great benefit to
the field of glycopeptide and glycoprotein synthesis.
The chemical synthesis of homogeneous glycopeptides and
glycoproteins from readily available materials has gained
considerable attention in organic synthesis.1 This field of
research is mainly driven by the fact that glycoproteins are
naturally expressed as heterogeneous mixtures of glycoforms,
rendering the study of the discrete effects of glycosylation in
protein function a difficult task.2 Chemical synthesis of glyco-
peptides and glycoproteins based on native chemical ligation
(NCL) represents one of the useful methods for the synthesis
of homogeneous glycopeptide structures.3 However, the restric-
tion imposed by the requirement for a cysteine residue at the
ligation site poses a significant limitation to this methodology.4
To overcome this limitation, several removable thiol-based
auxiliaries were introduced to fulfill the function of the cysteine
side chain in the ligation reaction. Kent and co-workers were
the first to extend the applicability of NCL by using a temporary
ethanethiol handle attached to the N-terminal peptide.5 Subse-
quent to this initial study, peptides with NR-linked auxiliaries
based on 1-phenyl-2-mercaptoethyl and 2-mercaptobenzyl have
We recently described a new approach for the synthesis of
â-O-linked glycopeptides lacking a cysteine residue.10 This
method that is based on sugar-assisted ligation (SAL) uses a
peptide thioester and a glycopeptide in which N-acetylglu-
cosamine (GlcNAc) â-linked to the serine side chain is modified
† Present address: Department of Chemistry, Ben Gurion University,
Beer Sheva 84105, Israel.
(1) For selected reviews: (a) Liu, L.; Bennett, C. S.; Wong, C.-H. Chem.
Commun. 2006, 1, 21-33. (b) Pratt, M. R.; Bertozzi, C. R. Chem. Soc.
ReV. 2005, 34, 58-68. (c) Davis, B. G. Chem. ReV. 2002, 102, 579-602.
(d) Seitz, O. ChemBioChem 2000, 1, 214-246.
(2) Helenius, A.; Aebi M. Science 2001, 291, 2364-2369.
(3) (a) Hackenberger, C. P. R.; Friel, C. T.; Radford, S. E.; Imperiali, B. J.
Am. Chem. Soc. 2005, 127, 12882-12889. (b) T. J. Tolbert, D. Franke, C.
H. Wong, Bioorg. Med. Chem. 2005, 13, 909-915. (c) Macmillan: D.;
Bertozzi, C. R. Angew. Chem., Int. Ed. 2004, 43, 1355-1359. (d) Warren,
J. D.; Miller, J. S.; Keding, S. J.; Danishefsky, S. J. J. Am. Chem. Soc.
2004, 126, 6576-6578. (e) Shin, Y.; Winans, K. A.; Backes, B. J.; Kent,
S. B. H.; Ellman, J. A.; Bertozzi, C. R. J. Am. Chem. Soc. 1999, 121,
11684-11689.
(6) (a) Offer, J.; Boddy, C.; Dawson, P. E. J. Am. Chem. Soc. 2002, 124, 4642-
4646. (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.
(7) Macmillan: D.; Anderson, D. W. Org. Lett. 2004, 6, 4659-4662.
(8) (a) Wu, B.; Chen, J.; Warren, J. D.; Chen, G.; Hua, Z.; Danishefsky, S. J.
Angew. Chem., Int. Ed. 2006, 45, 1-11. (b) Chen, J.; Warren, J. D.; Wu,
B.; Chen, G.; Wan, Q.; Danishefsky, S. J. Tetrahedron Lett. 2006, 47,
1969-1972.
(9) Chen, G.; Warren, J. D.; Chen, J.; Wu, B.; Wan, Q.; Danishefsky, S. J. J.
Am. Chem. Soc. 2006, 128, 7460-7462.
(4) Dawson, P. E.; Kent, S. B. H. Annu. ReV. Biochem. 2000, 69, 923-960.
(5) Cane, L. E.; Bark, S. J.; Kent, S. B. H. J. Am. Chem. Soc. 1996, 118,
5891-5896.
(10) Brik, A.; Ying, Y. Y.; Ficht, S.; Wong, C.-H. J. Am. Chem. Soc. 2006,
128, 5626-5627.
9
15026
J. AM. CHEM. SOC. 2006, 128, 15026-15033
10.1021/ja065601q CCC: $33.50 © 2006 American Chemical Society

Sugar-Assisted Ligation of N-Linked Glycopeptides
A R T I C L E S
Scheme 1. General Strategy of the Sugar-Assisted Chemical
at the C-2 acetamide with a thiol handle to mimic the cysteine
function. Upon completion of the ligation reaction, the thiol
handle can be reduced with H2/metal to the acetamide moiety,
furnishing the unmodified glycopeptides.11 Here we report that
this new method can be applied to the efficient synthesis of
N-linked glycopeptides. Despite the differences between the N-
and O-linked glycopeptides, the ligation rates were similar.
Importantly, the ligation chemistry tolerates a variety of amino
acids at the N-terminal glycopeptide, favoring less hindered
amino acids and those with side chains that can serve as a
general base in the ligation pathway. Moreover, the modified
glycopeptide can be transformed to the corresponding peptide
by glycosidases or to a more complex glycopeptide using
glycosyltransferases.
Synthesis of N-Linked Glycopeptides (R1, R3 ) Amino Acid Side
Chains, R2 ) Ph/(CH2)CONH2)
Results and Discussion
In addition to the naturally occurring O-linked glycopeptides,
N-linked glycosylation, which is a more prevalent form of
glycosylation, is found in a wide range of organisms ranging
from archaea to mammals and other eukaryotes.12 Due to the
important role of glycoproteins in many biological systems, their
preparation has been the goal of many laboratories for some
time.13 Encouraged by our previous results with SAL,10 we
sought to extend this strategy to the synthesis of N-linked
glycopeptides, despite the structural variations compared to the
O-linked case (Scheme 1). The differences could arise from the
change in the ring size of the SfN acyl transfer intermediate,
which is a 15-membered ring in the N-linked case versus a 14-
membered ring in the O-linked case. More importantly, the
reduced degree of rotational freedom around the glycan-amide
bond in the N-linked case can affect the proximity of the
nucleophilic amine to the thioester intermediate, thus hampering
the SfN acyl transfer. On the other hand, N-linked glycopep-
tides bear the N-acetyl-glucosamine unit as the first glycan
moiety attached to the peptide side chain. This ubiquitous
connectivity allows us to modify the C-2 acetamide with the
necessary thiol handle to assist the ligation.
Building Blocks and Glycopeptides Synthesis. In our
previous report, we relied on solution-phase synthesis to prepare
the thiol-modified O-linked glycopeptides. In order to make our
approach applicable for the synthesis of large glycopeptides and
ultimately for glycoproteins, we developed a solid-phase
synthesis of the N-linked glycopeptide precursor. We first
examined whether the necessary thiol-modified â-N-(GlcNAc)-
Asn building block 6 can be synthesized efficiently and used
in solid-phase peptide synthesis (SPPS). We started from the
known glycosyl azide 1 (Scheme 2A),14 which was reduced with
propanedithiol15 to give glycosyl amine 2 with a complete
retention of the stereochemistry. The â-glycosylamine 2 was
then coupled to the side chain of a protected aspartic acid by
employing HBTU/DIEA coupling conditions to furnish com-
pound 3.16 Following deprotection of the Troc protecting
group,17 the 2-amino functionality of compound 4 was coupled
to S-trityl-2-mercaptoacetic acid18 using HBTU/DIEA coupling
conditions, giving the fully protected compound 5. Finally, the
19
acid functionality was unmasked by employing Pd(PPh3)4 to
provide the desired building block 6 in gram quantities.
Glycopeptide synthesis bearing the thiol-modified â-N-
(GlcNAc) was carried out on Fmoc-Rink amide resin employing
routine PyBOP/NMM coupling conditions16 (Scheme 2B). The
coupling of the â-N-(GlcNAc)-Asn 6 to the N-terminal proline,
which is known to be a difficult coupling, was quantitatively
achieved as indicated by UV absorption at 302 nm of the Fmoc/
piperidine adduct. Upon completion of coupling of the amino
acid 6, the resin was split into five portions in which four
different amino acids (Gly, Ala, Val, Asp) were coupled to the
free N-terminal glycopeptide (Scheme 2B). Glycopeptide 7a,
without further elongation, was also prepared to evaluate the
effect of amino acid extension on the ligation reaction. Fol-
lowing hydrazine/methanol mixture (1:6) mediated acetate
removal, the glycopeptide-resin was exposed to the TFA/TIS/
TA/H2O (85:5:5:5) mixture to release the deprotected glyco-
(11) Yan, L. Z.; Dawson, P. E. J. Am. Chem. Soc. 2001, 123, 526-533.
(12) (a) Lechner, J.; Wieland, F. Annu. ReV. Biochem. 1989, 58, 173-194. (b)
Kornfeld, R.; Kornfeld, S. Annu. ReV. Biochem. 1985, 54, 631-664.
(13) (a) Hackenberger, C. P. R.; Friel, C. T.; Radford, S. E.; Imperiali, B. J.
Am. Chem. Soc. 2005, 127, 12882-12889. (b) Miller, J. S.; Dudkin, V.
Y.; Lyon, G. J.; Muir, T. W.; Danishefsky, S. J. Angew. Chem., Int. Ed.
2003, 42, 431-434. (c) Pratt, M. R.; Bertozzi, C. R. J. Am. Chem. Soc.
2003, 125, 6149-6159. (d) Hojo, H.; Watabe, J.; Nakahara, Y.; Ito, Y.;
Nabeshima, K.; Toole, B. P. Tetrahedron Lett. 2001, 42, 3001-3004. (e)
Tolbert, T. J.; Wong, C. H.; J. Am. Chem. Soc. 2000, 122, 5421-5428. (f)
Mizuno, M.; Haneda, K.; Iguchi, R.; Muramoto, I.; Kawakami, T.; Aimoto,
S.; Yamamoto, K.; Inazu, T.; J. Am. Chem. Soc. 1999, 121, 284-290. (g)
Guo, Z. W.; Nakahara, Y.; Ogawa, T. Angew. Chem., Int. Ed. 1997, 36,
1464-1466.
(16) Albericio, F.; Carpino, L. A. Methods Enzymol. 1997, 289, 104-126.
(17) Woodward, R. B.; Heusler, K.; Gosteli, J.; Naegeli, P.; Oppolzer, W.;
Ramage, R.; Rangamathan S.; Vorbru¨ggen, H. J. Am. Chem. Soc. 1966,
88, 852-853.
(14) Matsubara, K.; Mukaiama, T. Chem. Lett. 1994, 15, 247-250.
(15) Bayley, H.; Standring, D. N.; Knowles, J. R. Tetrahedron Lett. 1978, 39,
3633-3634.
(18) Kupihar, Z.; Schmel, Z.; Kele, Z.; Penke, B.; Kovacs, L. Bioorg. Med.
Chem. 2001, 9, 1241-1247.
(19) Ciommer, M.; Kunz, H. Synlett 1991, 8, 593-595.
9
J. AM. CHEM. SOC. VOL. 128, NO. 46, 2006 15027

A R T I C L E S
Brik et al.
Scheme 2. (A) Synthesis of the Building Block 6a and (B) SPPS Synthesis of Glycopeptides on Fmoc-Rink Amide Polystyrene Resin
a Reagents and conditions: (a) propanedithiol, DIEA, MeOH, 4 h, rt, (b) Fmoc-(Asp)-OAll, HBTU, DIEA, DMF, 12 h, 90%; (c) Zn, AcOH, rt, 12 h,
85%; (d) TrtSCH2COOH, HBTU, DIEA, DMF, 4 h, rt, 78%; (e) Pd(PPh3)4, NMA, THF, 1 h, rt, 90%.
Table 1. Scope of the Sugar-Assisted Ligation
peptide from the solid support. The crude mixture of each
synthesis was purified by preparative HPLC to generate peptides
7a-7e in 55-65% yield (See Supporting Information). To
compare the ligation rates of the N- and the O-linked glyco-
peptides, we used the same thioester peptides bearing Gly, Ala,
His, and Val at the C-terminus.10
Glycopeptide Ligation. Having the model peptides in hand,
we turned our focus on the ligation study. The ligations of the
ligand junction
-AA1- -AA2-AA1-
mass (Da)
obsdb
unprotected synthetic peptides were performed by employing
conditions similar to NCL (6 M guanidine·HCl, pH 7.5-8.5).20
Thus, peptides were dissolved at a final concentration of 10-
12 mM (1:1.2 molar ratio of peptide thioester to glycopeptide)
followed by the addition of 2% thiophenol.21 The ligation
reactions were performed at 37 °C and were vortexed periodi-
cally to equilibrate the thiol additive. The progress of each
ligation reaction was monitored using analytical HPLC and
LCMS.
The rates of the various ligation reactions reported in Table
1 provided several insights into our strategy. As shown in Table
1, the observed ligation rate is dependent on the C-terminal
amino acid of the peptide thioester. These results are in
agreement with previous observation in the NCL.20 The presence
of a sterically hindered amino acid at the C-terminal thioester
peptide leads to a slow reaction relative to the unhindered model.
Indeed, in the case of the valine peptide thioester ligation with
the glycopeptide 7b (Table 1, entry 10), we observed the
thioester intermediate as the major product by LCMS, in addition
to a minor peak that corresponds to the desired product. The
use of glycopeptide 7a without the extension of an additional
amino acid adjacent to the asparagine leads to a very slow and
incomplete ligation (Table 1, entry 1), similar to the O-linked
glycopeptide case. This observation may be to do with the
decrease in the ring size of the SfN acyl transfer intermediate
entry
-AA2-
t (h)a
calcd
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Gly
Gly
Gly
Gly
Gly
Ala
Ala
Ala
Ala
Val
His
His
Gly
Gly
His
-
Gly
>48
∼12
∼24
∼48
∼12
∼40
>48
-
1373.5 ( 0.2
1430.6 ( 0.2
1444.5 ( 0.2
1472.5 ( 0.2
1488.5 ( 0.2
1444.5 ( 0.2
1486.2 ( 0.2
1486.2 ( 0.2
1504.2 ( 0.2
1472.4 ( 0.2
1511.1 ( 0.2
1569.2 ( 0.2
1489.1 ( 0.2
1511.2 ( 0.2
1591.3 ( 0.2
1373.6
1430.6
1444.6
1472.7
1488.7
1444.6
1458.7
1486.3c
1503.6
1472.2c
1511.6
1569.7
1488.6
1511.6
1591.7
Gly
Ala
Val
Asp
Gly
Ala
Val
Asp
Gly
Gly
Asp
Asn
His
His
Gly-Gly
Gly-Ala
Gly-Val
Gly-Asp
Ala-Gly
Ala-Ala
Ala-Val
Ala-Asp
Val-Gly
His-Gly
His-Asp
Gly-Asn
Gly-His
His-His
∼36
-
∼12
∼12
∼36
∼6
∼8
a t represents the time at which the thioester peptide was consumed, with
the major peak being the desired product (Figure 1). b Characterized by
LCMS. c After 48 h, the major product was the ligation thioester intermedi-
ate with less than 5% of product.
from a 15-membered ring to a 12-membered ring in the shorter
case. Moreover, the steric hindrance induced by the modification
of the side chain with the sugar moiety could contribute to the
sluggish rate.
Interestingly, SAL tolerates several different amino acids at
N-terminus of the glycopeptide. This can be seen from the
examples where glycopeptide 7a is extended with various amino
acids. Despite the changes in the ligation rates going from
glycine to alanine and to valine (Table 1, entries 2-4), these
reactions proceeded well. In all of these cases the product was
(20) Hackeng, T. M.; Griffin, J. H.; Dawson, P. E. Proc. Natl. Acad. Sci. U.S.A.
1999, 96, 10068-10073.
(21) Dawson, P. E.; Churchill, M.; Ghadiri, M. R.; Kent, S. B. H. J. Am. Chem.
Soc. 1997, 119, 4325-4329.
9
15028 J. AM. CHEM. SOC. VOL. 128, NO. 46, 2006

Sugar-Assisted Ligation of N-Linked Glycopeptides
A R T I C L E S
lacks the ability to serve as a general base. When we used this
glycopeptide in the ligation with the same peptide thioester, the
rate was 3-fold slower (Table 1, entry 13) and similar to the
rate observed with glycopeptide 7d bearing the valine extension
(Table 1, entry 4). Furthermore, when glycopeptide 7a was
extended with a histidine residue, an amino acid known to serve
as a general base in many biological systems, the rate was faster
than the case of the glycine-extended glycopeptide ligation
(Table 1, entries 14, 15), supporting our general base hypoth-
esis.22
In all the ligation reactions presented in Table 1, the thioester
intermediate eluted in the HPLC analysis slightly before the
rearranged product. In our previous report we were able to
identify the thioester intermediate for the O-linked glycopeptide
by the mass spectrometry fragmentation pattern;10 however, in
the N-linked glycopeptide we did not observe similar mass
fragmentation. In order to differentiate the intermediate from
the product, we treated the suspected intermediate with a 10-
fold excess of 2-mercaptoethansulfonate (MES) at pH 7.4 and
the reaction was followed by LCSM. For example, when the
intermediates from entries 8 and 10 (1486.2, 1472.4 Da,
respectively) were treated with MES, they were rapidly con-
verted to the corresponding R-MES thioester peptides (758.3
and 744.3 Da, respectively) and the glycopeptide starting
material (841.3 and 871.3 Da, respectively).
In NCL, the unrearranged species has never been observed,
and it is known that the SfN acyl transfer is extremely rapid
and never becomes rate-limiting.23 In our ligation strategy, in
every instance we observed the intermediate in the reaction
pathway. The level of its accumulation was clearly dependent
on the ligation site. For the very difficult ligation junctions, the
major product was the intermediate thioester (Table 1, entries
8, 10). For the faster ones (Table 1, entries 2, 12, 5, 11, 12, 14,
15), the intermediate was formed rapidly (within the first few
minutes) and was eventually converted to the product in a rate
that is dependent on the ligation junction. Our results agree with
the studies on other auxiliaries,5,6 where the rearrangement of
the thioester intermediate is often rate-limiting.
Figure 1. Representative analytical HPLC traces of ligation reactions and
mass spectrometry analysis of the products (MALDI-TOF/MS). In all the
examples, the thioester substrate was completely consumed. (A) Ligation
reaction at 14 h (Table 1, entry 2): peak a, thioester substrate hydrolysis;
peak b, ligation product with the expected mass of 1431.0 ( 1 Da. (B)
Ligation reaction at 24 h (Table 1, entry 3): peak a, thioester substrate
hydrolysis; peak b, ligation intermediate; peak c, ligation product with the
expected mass of 1445.0 ( 1 Da, (C) Ligation reaction at 48 h (Table 1,
entry 9): peak a, thioester substrate hydrolysis; peak b, ligation product
with the expected mass of 1503.3 ( 1 Da. (D) Ligation reaction at 14 h
(Table 1, entry 12): peak a, thioester substrate hydrolysis; peak b, ligation
product with the expected mass of 1568.9 ( 1 Da.
Our study shows that the ligation reaction is more tolerant
to the steric hindrance around the nucleophilic amine compared
to the thioester elctrophile, as seen by comparing the ligation
rates of entries 4 and 10 (Table 1). In both cases the ligation
junction is Val-Gly; however, in entry 4 the thioester peptide
contains Gly at the C-terminus, while the nucleophilic amine
bears the Val residue (t ∼ 40 h). On the other hand, in entry
10, where the thioester peptide bears Val at the C-terminus while
the nucleophilic amine is part of a Gly residue, only the thioester
intermediate was observed. Similar results were also obtained
with the Gly-Ala junctions (Table 1, entries 3 and 6).
observed by HPLC analysis as the major peak with a complete
consumption of the starting material (Figure 1A,B). Preparative
HPLC of the ligation mixture gave the desired glycopeptides
in 70% yield (Table 1, entries 2, 3). Notably, ligation at the
Ala-Ala junction was also observed and gave the product in
50% yield (Table 1, entry 7). It should be noticed that in our
strategy the nucleophilic amine is primary, whereas in the other
known auxiliaries it is secondary, rendering it less nucleophilic
toward the thioester.6,7 These differences could account for the
increased sequence tolerance at the ligation junction and should
be taken into consideration for the design of new ligation
auxiliaries.
To our pleasant surprise, we found that in the case of
glycopeptide 7e containing N-terminal aspartic acid (Table 1,
entry 5), despite being a hindered amino acid, the ligation rate
was similar to that for the glycopeptide bearing a glycine residue
(Table 1, entry 2). We hypothesized that the carboxylic acid
side chain could serve as a general base in the ligation pathway,
thus facilitating the SfN acyl transfer. To test this, we prepared
a similar glycopeptide 7f, in which the N-terminal aspartic acid
was replaced with the sterically equivalent asparagine, which
The peptides that are used in these ligations bear the side
chains of Asp, Asn, His, Tyr, Ser, Thr, and Arg, in addition to
the free hydroxyl groups presented on the glycan core. Our
results show that SAL tolerates a variety of functional groups,
(22) It has been reported by Tam and coworkers (Zhang, L.; Tam, J. P.
Tetrahedron Lett. 1997, 38, 3-6.) that the imidazole of an N-terminal
histidine-containing peptide assisted in the ligation with a peptide thioester
(obtained by activating the thiocaboxylic acid of the C-terminal peptide
by Ellman’s reagent). The authors proposed a covalently bonded Nim-acyl
intermediate, without rolling out a mechanism that involves anchimeric
assistance of the imidazole moiety.
(23) Johnson, E. C. B.; Kent, S. B. H. J. Am. Chem. Soc. 2006, 128, 6640-
6646.
9
J. AM. CHEM. SOC. VOL. 128, NO. 46, 2006 15029

A R T I C L E S
Brik et al.
in our laboratory. Alternatively, the thiol group could also serve
as a handle for modification with various alkylating agents such
as thiol-reactive dyes for fluorescence labeling (Scheme 3).
Following SAL, the sugar moiety can be cleaved enzymati-
cally or chemically from the asparagine side chain, affording
the peptide structure (Scheme 3). In this case the sugar serves
as an auxiliary to assist the ligation at difficult junctions that
are not accessible with the existing auxiliaries (Table 1, entries
4, 7, 9, 12, 15).5-8 Preliminary studies were conducted with
the ligation product from Table 1, entry 9, and its desulfurized
version (1503.6, 1471.3 Da). Both substrates were treated
separately with PNGase A, an amidase that cleaves between
the innermost GlcNAc and asparagine residues from N-linked
glycopeptides and glycoproteins.26 Our results show that in both
reactions the product was formed after 2 h with the mass that
corresponds to the nonglycosylated peptide product (1268.3 Da).
However, the reaction was not completed after 24 h, providing
the product in 30% yield, possibly due to the unnatural
glycopeptide sequence that was used in our study.26
Alternatively, the thiol-containing GlcNAc moiety or the
desulfurization product could serve as an acceptor for enzymatic
elaboration to generate a more complex glycan (Scheme 3).27
To this end, we tested the enzymatic transfer of galactose
directly into the ligation product using â-1,4-galactosyltrans-
ferase (GalT).28 Although the thiol-modified sugar is not a
natural substrate for the transferase,29 we observed a full
conversion to the disaccharide glycopeptide when the ligation
product (Table 1, entries 3, 12) was treated with GalT and UDP-
galactose (Figure 2B). The reaction mixture was treated with
tris(2-carboxyethyl)phosphine hydrochloride (TCEP) to reduce
disulfide bonds formed during the enzymatic reaction followed
by product isolation using preparative HPLC (81% yield).
Transferring the sugar onto the ligation product before the thiol
handle removal allows further elaboration of the glycan moiety
through chemical dervitization. Further investigations into the
use of SAL in the synthesis of glycoproteins and neoglycon-
jugates are currently underway in our laboratory.
Figure 2. (A) HPLC and MALDI-TOF analysis of the crude desulfurization
reaction (1 h). A full conversion of the starting material from entry 9 (Table
1) to the desired product with the mass of 1470.5 Da (calcd mass ) 1471.5
Da) was observed. (B) HPLC and MALDI-TOF analysis of the crude
enzymatic transfer reaction (40 h) of galactose to the glycopeptide ligation
product (Table 1, entry 12). The desired product with the observed mass of
1731.9 Da (calcd mass ) 1731.8 Da) was isolated in 80% yield.
suggesting that SAL is chemoselective reaction. To check for
the compatibility of the lysine side chains under the ligation
conditions, two thioester peptides bearing the lysine residue at
different positions were prepared (Ac-LYRKG-SR, Ac-LYKAG-
SR) and tested with glycopeptide 7e. In both cases, similar
results were obtained in terms of rate and yield to the non-
lysine-containing peptides (see Supporting Information). It has
been reported that a lysine residue at the C-terminal peptide
could lead to lactam formation;24 however, in our study the
lactam formation was not observed.
Conclusions
Postligation Modification. Previously, we have shown that
the thiol handle can be removed efficiently by using desulfu-
rization conditions (H2/Pd/Al2O3, 6.0 M guanidine containing
100 mM phosphate buffer, pH 5.8, 10 mM TCEP) to furnish
the unmodified O-linked glycopeptide (Scheme 3).10 The
desulfurization reaction was first applied by Dawson and co-
workers to extend NCL for cysteine free peptides11 and was
nicely used by Kent and co-workers for the synthesis of cysteine-
free ubiquitin.25 In this study, the desulfurization reaction was
also successfully applied to the N-linked glycopeptide, and
several thiol-modified glycopeptides were desulfurized in up
to 90% isolated yield (Figure 2A). Although the removal of
the thiol handle by applying the desulfurization reaction could
restrict the use of our method to cysteine free peptides, several
orthogonal protecting groups can in principle be used to protect
the cysteine side chain during the reduction. Work toward
performing the ligation reaction followed by desulfurization in
the presence of cysteine residues is currently under investigation
We have demonstrated that SAL can be applied to N-linked
glycopeptide ligation. Despite the changes in the atom con-
nectivity and the ring size of the SfN acyl transfer intermediate
between N- and O-linked glycopeptides, the ligation reactions
proceeded successfully and in comparable rates. Notably, the
ligation junction tolerates a variety of sequences favoring less
hindered amino acids and those with side chains that can serve
as a general base in the ligation pathway. Moreover, since our
approach afforded ligation at difficult junctions, this method
can be applied to the synthesis of large peptides by subsequent
enzymatic removal of the sugar moiety. Alternatively, by using
GalT we were able to introduce galactose directly to the ligation
product, furnishing the disaccharide glycopeptide. Together,
(26) Fan, J. Q.; Lee, Y. C. J. Biol. Chem. 1997, 272, 27058-27064.
(27) For, recent reviews discussing the use of enzymes to elaborate glycoproteins,
see: Gorgan, M. J.; Pratt, M. R.; Maracaurelle, L. A.; Bertozzi, C. A. Annu.
ReV. Biochem. 2002, 71, 593-634.
(28) (a) Schanbacher, F. L.; Ebner, K. E. J. Biol. Chem. 1970, 245, 5057-
5061. (b) Beyer, T. A.; Sadler, J. E.; Rearick, J. I.; Paulson, J. C.; Hill, R.
L. AdV. Enzymol. 1981, 52, 23-175.
(24) de Koning, M. C.; Filippov, D. V.; Overhand, M; van der Marel G. A.;
van Boom J. H. Eur. J. Org. Chem. 2004, 850-857.
(25) Bang, D.; Makhatadze, G. I.; Tereshko, V.; Kossiakoff, A. A.; Kent, S. B.
Angew. Chem., Int. Ed. 2005, 44, 3852 -3856.
(29) For a previous study on the effect of substitution at C-2 of the GlcNAc
acceptor on GalT activity, see: Berliner, L. J.; Davis, M. E.; Ebner, K. E.;
Beyer, T. A.; Bell, J. E. Mol. Cell. Biochem. 1984, 62, 37-42.
9
15030 J. AM. CHEM. SOC. VOL. 128, NO. 46, 2006

Sugar-Assisted Ligation of N-Linked Glycopeptides
A R T I C L E S
Scheme 3. Possible Modifications of the Ligation Product
74.55, 72.93, 72.84, 68.52, 62.35, 57.18, 20.77, 20.62, 20.60. HRMS
(ESI-TOF) calcd for C15H21Cl3N2O9 [M + H]+: 479.0385. Found:
479.0375.
these sequences of reactions hold a great potential in the
synthesis of glycopeptides and glycoproteins.
Experimental Procedures
Compound 3. Compound 2 (2.0 g, 4.0 mmol) was added to a
preactivated Fmoc-Asp-OAll (2.48 mg, 6.20 mmol) with HBTU (2.4
g, 6.2 mmol) and DIEA (2.0 mL, 12 mmol) in dry DMF (20 mL) at 0
°C. The reaction mixture was stirred at room temperature for 12 h and
then diluted with 80 mL of ethyl acetate and washed with water and
brine. The organic layer was dried over MgSO4 and concentrated for
flash column chromatography (CHCl3/MeOH 10:1) to give 3.2 g of 3
(90%). 1H NMR (CDCl3, 500 MHz): δ 7.74 (d, J ) 7.7 Hz, 2H), 7.60
(d, J ) 7.4 Hz, 2H), 7.23-7.39 (m, 5H), 6.21 (dd, J ) 8.8, 4.1 Hz,
1H), 5.87 (m, 1H), 5.19-5.31 (m, 4H), 5.08 (dd, J ) 9.6, 9.6 Hz,
1H,), 4.76 (d, J ) 12.1 Hz, 1H), 4.69 (d, J ) 12.4 Hz, 1H), 4.62-4.66
(m, 2H), 4.42 (dd, J ) 10.3, 7.0 Hz, 1H), 4.27-4.33 (m, 2H), 4.05 (d,
J ) 12.1 Hz, 1H), 3.80-3.90 (m, 2H), 2.97 (m, 1H), 2.81 (m, 1H),
2.01-2.03 (2s, 9H). 13C NMR (CDCl3, 500 MHz): δ 170.70, 170.66,
170.61, 170.44, 169.34, 165.55, 162.58, 156.03, 155.44, 143.68, 143.53,
141.02, 131.44, 127.51, 126.88, 125.05, 124.99, 119.78, 118.36, 95.24,
79.06, 74.35, 73.28, 72.55, 68.04, 67.11, 66.03, 54.93, 50.31, 46.86,
44.70, 37.47, 36.35, 31.32, 20.49, 20.43, 20.39. HRMS (ESI-TOF) calcd
for C37H40Cl3N3O14 [M + H]+: 856.1649. Found: 856.1645.
General Methods. All chemical reagents were purchased from
Aldrich or Acros and used without further purification. Amino acids
and coupling reagent were purchased from Novabiochem. Tetrahydro-
furan (THF) was distilled over sodium/benzophenone, and methylene
chloride (CH2Cl2) was distilled over calcium chloride. 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 citric ammonium molybdate. Flash chroma-
tography was performed on silica Sila-P flash silica gel (40-63 µm,
1
Silicycle, Quebec, Canada). H and 13C NMR spectra were recorded
on a Bruker AMX 500-MHz spectrometer. Coupling constants (J) are
1
reported in hertz. All H chemical shifts are reported in δ referenced
to solvent. LCMS (Agilent 1100 LC coupled to an Agilent 1100 single
quadrapole mass spectrometer, column is an Agilent SB C8 50 × 4.6
mm) and/or reverse-phase HPLC analysis were performed on a Vydac
C18 column using a linear gradient of buffer A in buffer B over 30
min (buffer A, 0.1% trifuoroaetic acid (TFA) in water; buffer B, 0.1%
TFA in acetonitrile).
Compound 2. Propanedithiol (24 mmol, 2.6 mL) was added to a
degassed mixture of compound 1 (2.0 g, 4.0 mmol) and Et3N (24.0
mmol, 3.4 mL) in methanol (20 mL). The reaction was allowed to stir
at room temperature for 6 h, at which time it was concentrated, resulting
in glycosylamine 2, which was placed under high vacuum for 8 h and
Compound 4. Compound 3 (3.2 g) was dissolved in 20 mL of acetic
acid followed by the addition of 3.0 g of zinc dust. The reaction mixture
was stirred at room temperature for 8 h. After filtration the solvent
was removed under reduced pressure. The residue was purified using
flash column chromatography [ethyl acetate (EA)/hexanes (Hex), 3:1]
to give 2.6 g of 4 (88%). 1H NMR (DMSO-d6, 500 MHz): δ 8.62 (d,
J ) 8.8 Hz, 1H), 7.88 (d, J ) 7.7 Hz, 2H), 7.81 (d, J ) 8.1 Hz, 1H),
7.69 (d, J ) 7.4 Hz, 2H), 7.31-7.42 (m, 4H), 5.87 (m, 1H), 5.29 (d,
J ) 17.2 Hz, 1H), 5.17 (d, 1H, 10.7), 4.93-5.01 (m, 2H), 4.77 (dd, J
) 9.9, 9.9 Hz, 1H), 4.49-4.60 (m, 3H), 4.21-4.35 (m, 3H), 4.15 (dd,
J ) 12.5, 4.0, 1H), 3.92 (d, J ) 11.7 Hz, 1H), 3.82-3.84 (m, 1H),
1
used directly for the next step. H NMR (CDCl3, 500 MHz): δ 5.11
(dd, J ) 9.6, 9.6 Hz, 1H,), 5.06 (dd, J ) 9.6, 9.6 Hz, 1H), 5.83 (d, J
) 11.8 Hz, 1H), 4.63 (d, J ) 12.1 Hz, 1H), 4.24 (dd, J ) 12.5, 5.2
Hz, 1H), 4.17 (d, J ) 9.2 Hz, 1H), 4.11 (dd, J ) 12.1, 1.9 Hz, 1H),
3.64-3.68 (m, 2H), 2.10 (s, 3H), 2.04 (s, 3H), 2.03 (s, 3H). 13C NMR
(CDCl3, 500 MHz): δ 171.11, 170.71, 169.42, 154.89, 95.36, 86.18,
9
J. AM. CHEM. SOC. VOL. 128, NO. 46, 2006 15031

A R T I C L E S
Brik et al.
2.68-2.75 (m, 1H), 2.61 (dd, J ) 15.4, 7.7, 1H), 1.98-1.95 (3s, 9H).
13C NMR (DMSO-d6, 500 MHz): δ 171.26, 170.03, 169.87, 169.41,
169.35, 156.82, 143.75, 143.73, 140.71, 140.69, 132.30, 127.64, 127.10,
125.23, 125.19, 120.11, 117.54, 80.70, 75.36, 72.08, 68.64, 65.80, 64.99,
61.99, 54.93, 50.42, 46.55, 38.23, 36.93, 20.70, 20.52, 20.43. HRMS
(ESI-TOF) calcd for C34H39N3O12 [M + H]+: 682.2606. Found:
682.2583.
Strategy: Fmoc Cleavage. After treatment with 10% piperidine/DMF
(2 × 5 min) the resin was washed (5× DMF, 5× DCM, 5× DMF).
Coupling. After preactivation of 4 equiv of protected amino acid
(final concentration 0.1 M in DMF) for 5 min using 4 equiv of PyBOP
and 8 equiv of NMM, the solution was added to the resin. After 30
min, the resin was washed with DMF (5×), DCM (5×), and DMF
(5×).
Compound 5. Compound 4 (2.0 g, 2.8 mmol) was added to
preactivated S-trityl-2-mercaptoactic acid (1.86, 5.6 mmol) with HBTU
(2.10 g, 5.6 mmol) and DIEA (2.0 mL, 11.2 mmoL) in dry DMF (20.0
mL). The reaction mixture was stirred at room temperature for 4 h,
diluted with 80 mL of ethyl acetate, and washed with water and brine.
The organic layer was dried over MgSO4 and concentrated for flash
Capping. Acetic anhydride/pyridine (1:9) was added to the resin.
After 5 min the resin was washed with DMF (5×), DCM (5×), and
DMF (5×).
Coupling of the Sugar-Containing Monomer. After preactivation
of 4 equiv of building block 6 (final concentration 0.1 M in DMF) for
5 min using 4 equiv of PyBOP and 8 equiv of NMM, the solution was
added to the resin. After 12 h, the resin was washed with DMF (5×),
DCM (5×), and DMF (5×).
1
column chromatography (EA/Hex 1:2) to give 2.0 g of 5 (89%). H
NMR (CDCl3, 500 MHz): δ 7.74 (d, J ) 7.4 Hz, 2H), 7.58 (d, J )
5.5 Hz, 2H), 7.18-7.39 (m, 22H), 6.03 (d, J ) 8.4 Hz, 1H), 5.77-
5.81 (m, 2H), 5.24 (d, J ) 17.3 Hz, 1H), 5.14 (d, J ) 10.3 Hz, 1H),
5.03 (dd, J ) 9.5, 9.5 Hz, 1H), 4.93 (dd, J ) 10.3, 10.3 Hz, 1H), 4.80
(dd, J ) 8.8, 8.8 Hz, 1H), 4.54-4.64 (m, 2H), 4.38-4.51 (m, 2H),
4.20-4.35 (m, 3H), 4.04 (d, J ) 12.1 Hz, 1H), 3.90-3.94 (m, 1H),
3.69 (d, J ) 8.1 Hz, 1H), 2.84-2.89 (m, 1H), 2.64-2.69 (m, 1H),
2.04 (s, 3H), 2.01 (s, 3H), 1.95 (s, 3H). 13C NMR (CDCl3, 500 MHz):
δ 171.51, 171.05, 170.75, 170.58, 169.21, 156.02, 143.82, 141.20,
141.17, 131.52, 129.34, 128.26, 127.62, 127.10, 126.99, 125.17, 125.10,
119.89, 118.33, 80.23, 73.39, 71.87, 68.00, 67.70, 6.17, 66.04, 61.54,
53.07, 50.34, 47.02, 37.47, 35.86, 20.65, 20.49. HRMS (ESI-TOF) calcd
for C55H55N3O13S [M + Na]+: 1020.3348. Found: 1020.3351.
Compound 6. Compound 5 (2.0 g, 2.0 mmol) was suspended in
THF (20 mL), and N-methylaniline (2.2 mL, 20 mmol) and (PPh3)4Pd
(231 mg, 0.2 mmol) were added subsequently. The reaction mixture
was stirred at room temperature for 1 h. After removing the solvent
under reduced pressure, the residue was subjected to column chroma-
tography (MeOH/CH3Cl 9:1) to give the product in 90% yield (1.8 g,
Cleavage. 1. Removal of the Acetyl Protecting Groups of the
Sugar. The resin was washed with MeOH (10×), treated with MeOH/
hydrazine 6:1 (2 × 2.5 h), and washed (10× MeOH, 5× DMF, 10×
DCM).
2. Cleavage from the Resin. A mixture of TFA, thioanisole,
triisopropylsilane, and water (17:1:1:1) was added. After 2 h, the resin
was washed with TFA (4 × 4 mL).
Workup. The combined solutions were concentrated in vacuo. The
residue was dissolved in water, purified by preparative HPLC, and
analyzed by MALDI-TOF/MS (matrix R-cyano-4-hydroxycinnamic
acid).
General Procedure for Chemical Ligation. The ligation of
unprotected glycopeptide segments was performed as follows: 0.1 M
phosphate buffer (pH 8.5) containing 6 M guanidine was degassed with
argon for 20 min before use. Peptide thioester and glycopeptide were
dissolved at a final concentration of 10 mM followed by the addtion
of 2% thiophenol. The ligation reaction was performed in a heating
block at 37 °C and was vortexed periodically to equilibrate the thiol
additive. Before analysis, TCEP (50 mM) was added to reduce any
disulfide bonds. The reaction was monitored using LCMS and/or
reverse-phase HPLC analysis performed on a Vydac C18 column using
a linear gradient (95-60%) of buffer A in buffer B over 30 min.
General Procedure for Desulfurization. Desulfurization reactions
were performed in a 0.1 M phosphate buffer containing 6.0 M guanidine
at pH 5.8 and 10 mM TCEP at room temperature. The buffer was
degassed by bubbling argon through for 20 min before each use. Pd/
Al2O3 was added (10-20 times the weight of peptide) and the reaction
was kept under hydrogen. The desulfurization reaction was monitored
by analytical HPLC as described above.
1
90%). H NMR (MeOD + CDCl3 (v/v ) 1/4), 500 MHz): δ 7.72-
7.77 (m, 2H), 7.55-7.64 (m, 3H), 7.50-7.55 (m, 1H), 7.34-7.41 (m,
3H), 7.24-7.31 (m, 7H), 7.18-7.23 (m, 5H), 7.11-7.15 (m, 2H), 5.15
(dd, J ) 10.3, 9.6 Hz, 2H), 5.10 (d, J ) 9.9 Hz, 1H), 4.96 (dd, J )
9.6, 9.6 Hz, 2H), 4.49 (m, 1H), 4.21-4.26 (m, 3H), 4.12 (dd, J ) 7.4,
7.4 Hz, 1H), 4.00-4.07 (m, 3H), 3.78-3.81 (m, 1H), 1.99 (s, 3H),
1.98 (s, 3H), 1.95 (s, 3H). 13C NMR (MeOD + CDCl3 (v/v ) 1/4),
500 MHz): δ 172.16, 171.65, 171.06, 145.10, 144.93, 144.77, 142.23,
142.21, 133.52, 133.49, 132.87, 132.79, 130.33, 129.76, 129.66, 129.57,
128.88, 128.56, 128.52, 127.94, 127.81, 127.79, 126.15, 126.08, 120.67,
79.28, 78.98, 74.32, 74.15, 69.59, 68.06, 67.96, 62.94, 53.61, 48.01,
38.89, 37.45, 20.75, 20.98. HRMS (ESI-TOF) calcd for C52H51N3O13S
[M + Na]+: 980.3035. Found: 980.3032.
Glycopeptide Synthesis. Solid-phase chemistry was carried out in
syringes equipped with Teflon filters, purchased from Torviq. If not
differently described, all reactions were carried out at room temperature.
Preparative HPLC was performed on a Hitachi (D-7000 HPLC system)
instrument using a preparative column (Grace Vydac “Protein &
Peptide” C18) and a flow rate of 9 mL/min. DMF was purchased from
Aldrich in biotech grade. Commercial reagents were used without
further purification. Resins, protected amino acids, and PyBOP were
purchased from Novabiochem.
Enzymatic Removal of the Glycan Moiety with N-Glycosidase
A. A total of 0.6 nmol of substrate (Table 1, entry 9, and its desulfurized
version, 1503.6 and 1471.3 Da) dissolved in 40 µL of pH 8.0, 50 mM
sodium phosphate buffer was mixed with N-glycosidase A (10 U,
purchased from New England Biolabs) and incubated at 37 °C. For
LCMS analysis, 2.5 µL of reaction solution was taken from the
Eppendorf tube and diluted into 35 µL of water.
Enzymatic Transfer. Glycopeptide product from entry 12, Table 1
(1.0 mg, 0.63 µmol) and UDP-Gal (0.6 mg, 1.0 µmol) were dissolved
in 110 µL of HEPES buffer (130 mM, pH 7.4 + 0.25% Triton X-100)
containing freshly prepared MnCl2 solution (0.95 µmol). â-1,4-GalT
(9.4 µL, 5 mU) and alkaline phosphatase (0.04 µL, 30 mU) were added,
and the reaction was shaken gently at ambient temperature for 40 h.
The reaction was then incubated with an excess of TCEP for 30 min
and purified by HPLC employing a gradient of 100:0 A:B to 10:90
A:B over 30 min at a flow rate of 1.5 mL min-1, where A ) 0.1%
TFA in H2O and B ) 0.1% TFA in MeCN. The retention time of the
desired product was 9.16 min. Lyophilization of the desired fractions
afforded pure disaccharide glycopeptide in 80% yield. Similar results
were obtained with the ligation product from entry 3, Table 1 (calcd
mass, 1605.7 Da; observed, 1606.4 Da).
Preloading of the Rink Amide Resin. The resin (0.69 mmol/g)
was washed (5× DCM, 5× DMF), followed by removal of the Fmoc
group by treating it with 10% piperidine/DMF (2 × 5 min) and another
washing step (5× DMF, 5× DCM, 5× DMF). For preactivation of the
first protected amino acid, 1 equiv of PyBOP and 2 equiv of NMM
were added to a solution of the building block (0.1 M) in DMF. After
5 min of preactivation, the mixture was added to the resin. After 2 h
the resin was washed (5× DMF, 5× DCM, 5× DMF), capped with
acetic anhydride/pyridine (1:9) (2 × 5 min), and washed (5× DMF,
5× DCM, 5× DMF).
Glycopeptide Solid-Phase Synthesis According to the Fmoc-
9
15032 J. AM. CHEM. SOC. VOL. 128, NO. 46, 2006

Sugar-Assisted Ligation of N-Linked Glycopeptides
A R T I C L E S
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). C.S.B. is grateful for the Ruth L.
Kirschstein NRSA postdoctoral fellowship provided by the NIH
(GM073500-02).
Supporting Information Available: Experimental procedures
and characterization of the products and other detailed results.
This material is available free of charge via the Internet at
http://pubs.acs.org.
JA065601Q
9
J. AM. CHEM. SOC. VOL. 128, NO. 46, 2006 15033