Extended sugar-assisted glycopeptide ligations: Development, scope, and Applications

Article abstract of DOI:10.1021/ja073653p

Recently, we reported the development of sugar-assisted ligation (SAL), a novel peptide ligation method for the synthesis of glycopeptides. After screening a large number of glycoprotein sequences in a glycoprotein database, it became evident that a large proportion (approximately 53%) of O-glycosylation sites contain amino acid residues that will not undergo SAL reactions. To overcome these inherent limitations and broaden the scope of the method we report here the development of an extended SAL method. Glycopeptides containing up to six amino acid extensions N-terminal to the glycosylated residue were shown to facilitate ligation reactions with peptide thioesters, and these products were isolated in good yields. Kinetic analysis was used to show that as glycopeptides were extended by further amino acid residues, ligation reactions became slower. This finding was rationalized by molecular dynamics simulations using AMBER9. These studies suggested a general trend whereby the proximal distance between the reactive sites of the thioester intermediate (the N-terminal amine and the carbonyl carbon of the thioester) increased as glycopeptides were extended, thus slowing down the ligation rate. Each of the extended SAL methods showed broad tolerance to a number of different amino acid combinations at the ligation junction. Re-evaluation of the glycoprotein database suggested that 95% of the O-linked glycosylation sites can now be utilized to facilitate SAL or extended SAL reactions. As such, this method represents an extremely valuable tool for the synthesis of naturally occurring glycopeptides and glycoproteins. To demonstrate the applicability of the method, extended SAL was successfully implemented in the synthesis of the starting unit of the cancer-associated MUC1 glycoprotein.

Full text of DOI:10.1021/ja073653p

Published on Web 10/13/2007
Extended Sugar-Assisted Glycopeptide Ligations:
Development, Scope, and Applications
Richard J. Payne,† Simon Ficht,† Sishi Tang,‡ Ashraf Brik,†,# Yu-Ying Yang,†
David A. Case,‡ and Chi-Huey Wong*,†,§
Contribution from the Departments of Chemistry and Molecular Biology, The Scripps Research
Institute, 10550 North Torrey Pines Road, La Jolla, California 92037, and The Genomics
Research Center, Academia Sinica, 128 Section 2, Academia Road,
Nankang, Taipei 115, Taiwan
Received May 21, 2007; E-mail: wong@scripps.edu
Abstract: Recently, we reported the development of sugar-assisted ligation (SAL), a novel peptide ligation
method for the synthesis of glycopeptides. After screening a large number of glycoprotein sequences in a
glycoprotein database, it became evident that a large proportion (approximately 53%) of O-glycosylation
sites contain amino acid residues that will not undergo SAL reactions. To overcome these inherent
limitations and broaden the scope of the method we report here the development of an extended SAL
method. Glycopeptides containing up to six amino acid extensions N-terminal to the glycosylated residue
were shown to facilitate ligation reactions with peptide thioesters, and these products were isolated in
good yields. Kinetic analysis was used to show that as glycopeptides were extended by further amino acid
residues, ligation reactions became slower. This finding was rationalized by molecular dynamics simulations
using AMBER9. These studies suggested a general trend whereby the proximal distance between the
reactive sites of the thioester intermediate (the N-terminal amine and the carbonyl carbon of the thioester)
increased as glycopeptides were extended, thus slowing down the ligation rate. Each of the extended SAL
methods showed broad tolerance to a number of different amino acid combinations at the ligation junction.
Re-evaluation of the glycoprotein database suggested that 95% of the O-linked glycosylation sites can
now be utilized to facilitate SAL or extended SAL reactions. As such, this method represents an extremely
valuable tool for the synthesis of naturally occurring glycopeptides and glycoproteins. To demonstrate the
applicability of the method, extended SAL was successfully implemented in the synthesis of the starting
unit of the cancer-associated MUC1 glycoprotein.
Introduction
the protein part, the glycosylation pattern of a given glycoprotein
is not under the control of a coding template, but is dictated by
Protein glycosylation is a ubiquitous post-translational modi-
fication which introduces enormous structural diversity to
proteins.1 Estimations suggest that more than fifty percent of
all human proteins are glycosylated.2 The glycan modification
is of paramount importance for a variety of biological recogni-
tion events such as cell adhesion, cell differentiation, and cell
growth.3,4 Aberrant glycosylation of proteins often modifies
intracellular recognition and has been associated with a number
of serious illnesses including autoimmune diseases, infectious
diseases, and cancer.5,6 To understand the role of glycosylation
at a molecular level, it is imperative to have access to
homogeneous glycopeptides and glycoproteins. In contrast to
the relative activities of a number of glycosyltransferase
enzymes. The resulting glycoproteins are usually produced as
inseparable mixtures of glycoforms, and as such, biological
expression systems cannot be used for the production of
homogeneous glycoproteins. It is currently accepted that chemi-
cal and chemoenzymatic intervention can be used to solve the
availability problem. Peptide ligation methods have emerged
as powerful tools in this respect.7-10
One such method, native chemical ligation (NCL), has proven
to be very useful for peptide chemistry, where it has been
implemented in the synthesis of hundreds of proteins to
date.11 More recently, NCL has gained the spotlight as an
effective method for the preparation of glycopeptides and
glycoproteins.11-17 The obvious limitation of this method is the
† Department of Chemistry, The Scripps Research Institute.
‡ Department of Molecular Biology, The Scripps Research Institute.
§ The Genomics Research Center.
# Current address: Department of Chemistry, Ben Gurion University,
Beer Sheva 84105, Israel.
(7) Davis, B. G. Chem. ReV. 2002, 102, 579-601.
(1) Spiro, R. G. Glycobiology 2002, 12, 43-56.
(8) Pratt, M. R.; Bertozzi, C. R. Chem. Soc. ReV. 2005, 34, 58-68.
(9) Liu, L.; Bennett, C. S.; Wong, C. H. Chem. Commun. 2006, 21-33.
(10) Brik, A.; Ficht, S.; Wong, C. H. Curr. Opin. Chem. Biol. 2006, 10, 638-
644.
(2) Apweiler, R.; Hermjakob, H.; Sharon, N. Biochim. Biophys. Acta 1999,
1473, 4-8.
(3) Dwek, R. A. Chem. ReV. 1996, 96, 683-720.
(4) Varki, A. Glycobiology 1993, 3, 97-130.
(11) Dawson, P. E.; Kent, S. B. H. Annu. ReV. Biochem. 2000, 69, 923-960.
(12) Wieland, T.; Bokelmann, E.; Bauer, L.; Lang, H. U.; Lau, H. Justus Liebigs
Ann. Chem. 1953, 583, 129-149.
(5) Miller, L. H.; Good, M. F.; Milon, G. Science 1994, 264, 1878-1883.
(6) Dube, D. H.; Bertozzi, C. R. Nat. ReV. Drug DiscoVery 2005, 4, 477-488.
9
10.1021/ja073653p CCC: $37.00 © 2007 American Chemical Society
J. AM. CHEM. SOC. 2007, 129, 13527-13536
13527

A R T I C L E S
Payne et al.
Scheme 1. Proposed Mechanism of Sugar-Assisted Ligation (SAL)
be chemoselective with all naturally occurring amino acid side
chains.24 The thiol auxiliary could then be removed by des-
ulfurization to afford the natural glycan motif. Initial concerns
over the orthogonality of this reaction in the presence of other
thiol containing amino acid residues was recently addressed by
Kent and co-workers18 and by our laboratory.25 Both reports
describe the use of a cysteinyl acetamidomethyl (Acm) protect-
ing group strategy to facilitate selective desulfurization reactions.
This method can be adopted for desulfurization of SAL products,
as demonstrated recently in the total synthesis of the antibacterial
glycoprotein diptericin.25 This has opened the door for the
synthesis of more complex glycopeptides and potentially
glycoprotein targets by this technology.26
The continuing goal of our laboratory was to apply the
existing SAL method to the total synthesis of a native
glycoprotein. A detailed search of over two hundred O-linked
glycoproteins was conducted using the database O-GlycBase
v6.00.27 This screen indicated that a large number of glycopro-
teins bear sterically hindered amino acid residues N-terminal
to the sugar-carrying amino acid (i.e., R1 and R3 in Scheme 1).
As a result, SAL cannot be used in these cases. To determine
the extent of this finding, each putative ligation junction adjacent
to naturally occurring O-glycosylation sites (n ) 535, see
Supporting Information) was scanned. This was conducted by
assessing the nature of the required thioester (two amino acids
N-terminal to the glycosylated residue, i.e., R1 in Scheme 1),
since it is established that the steric nature of the thioester
component is the most influential factor in SAL (and NCL)
ligation efficiency.24 Amino acid thioesters known to be
extremely difficult to ligate include proline, valine, isoleucine,
leucine, and threonine.28 Other thioesters cannot be used without
side-chain protection. These include glutamic acid and aspartic
acid due to the rapid formation of mixed cyclic anhydrides with
the carboxylate side chains29 and lysine due to the formation
of cyclic lactams.30 Additionally, in contrast to NCL, the use
of cysteine thioesters could be problematic for SAL reactions
because of cyclic thioester formation and promoted hydrolysis.
Remarkably, 53% of all O-glycosylation sites studied contained
such amino acids (indicated in red in Figure 1). A further 22%
of glycosylation sites contained thioesters that would facilitate
ligations, however, would do so at a sluggish rate (indicated in
orange). Only 25% of the O-glycosylation sites studied contained
thioesters that are known to undergo SAL at reasonable rates
(indicated in green). It is evident from this informatics study
that only a small proportion of the glycoproteome can be feasibly
synthesized by the existing ligation methodology.
requirement for an N-terminal cysteine or alanine residue.18
Unfortunately, many glycoprotein targets do not bear these
amino acids at strategically useful positions in the protein
sequence, and as such these methods cannot be implemented.
This fact has led to the development of cysteine-free ligation
techniques, which incorporate a thiol containing auxiliary at the
N-terminus.19,20 This method has proven useful for the synthesis
of glycopeptides; however, the use of these auxiliaries appears
to be limited to ligation sites containing amino acid side chains
of low steric bulk.21,22
Our laboratory has recently reported the development of a
sugar-assisted ligation (SAL) method for the synthesis of O-
and N-linked glycopeptides.23,24 This method utilizes a glyco-
peptide in which the carbohydrate (N-acetyl glucosamine) is
derivatized with a mercaptoacetate auxiliary at the 2-position.
In the presence of a peptide thioester, and under suitable ligation
conditions, thioester exchange is followed by an S f N acyl
transfer affording a ligated product with a native peptide
backbone (Scheme 1). The reaction cascade showed high
sequence tolerance at the ligation junction and was shown to
(13) Dawson, P. E.; Muir, T. W.; Clarklewis, I.; Kent, S. B. H. Science 1994,
266, 776-779.
To circumvent this problem, we proposed a modification of
the existing method whereby SAL could be conducted on
glycopeptides that had been extended by the addition of extra
amino acids N-terminal to the glycosylated residue. If successful
in this endeavor, it would represent a flexible new method,
providing access to a significantly greater number of biologically
relevant glycopeptide and glycoprotein targets. The first step
(14) 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.
(15) Mezzato, S.; Schaffrath, M.; Unverzagt, C. Angew. Chem., Int. Ed. 2005,
44, 1650-1654.
(16) Marcaurelle, L. A.; Mizoue, L. S.; Wilken, J.; Oldham, L.; Kent, S. B. H.;
Handel, T. M.; Bertozzi, C. R. Chem. Eur. J. 2001, 7, 1129-1132.
(17) Hackenberger, C. P. R.; Friel, C. T.; Radford, S. E.; Imperiali, B. J. Am.
Chem. Soc. 2005, 127, 12882-12889.
(18) Pentelute, B. L.; Kent, S. B. H. Org. Lett. 2007, 9, 687-690.
(19) Offer, J.; Boddy, C. N. C.; Dawson, P. E. J. Am. Chem. Soc. 2002, 124,
4642-4646.
(25) Yang, Y. Y.; Ficht, S.; Brik, A.; Wong, C. H. J. Am. Chem. Soc. 2007,
129, 7690-7701.
(20) 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.
(26) Hsieh-Wilson, L. C. Nature 2007, 445, 31-33.
(27) Gupta, R.; Birch, H.; Rapacki, K.; Brunak, S.; Hansen, J. E. Nucleic Acids
Res. 1999, 27, 370-372.
(21) Macmillan, D.; Anderson, D. W. Org. Lett. 2004, 6, 4659-4662.
(22) Wu, B.; Chen, J. H.; Warren, J. D.; Chen, G.; Hua, Z. H.; Danishefsky, S.
J. Angew. Chem., Int. Ed. 2006, 45, 4116-4125.
(28) Hackeng, T. M.; Griffin, J. H.; Dawson, P. E. Proc. Natl. Acad. Sci. U.S.A.
1999, 96, 10068-10073.
(23) Brik, A.; Yang, Y. Y.; Ficht, S.; Wong, C. H. J. Am. Chem. Soc. 2006,
128, 5626-5627.
(29) Villain, M.; Gaertner, H.; Botti, P. Eur. J. Org. Chem. 2003, 3267-3272.
(30) De Koning, M. C.; Filippov, D. V.; van der Marel, G. A.; van Boom, J.
H.; Overhand, M. Eur. J. Org. Chem. 2004, 850-857.
(24) Brik, A.; Ficht, S.; Yang, Y. Y.; Bennett, C. S.; Wong, C. H. J. Am. Chem.
Soc. 2006, 128, 15026-15033.
9
13528 J. AM. CHEM. SOC. VOL. 129, NO. 44, 2007

Extended Sugar-Assisted Ligations
A R T I C L E S
Figure 1. Histogram representing the number of O-glycosylation sites of O-GlycBase v6.00 that can be utilized for sugar-assisted ligation (SAL). The
assessment was based on the C-terminal thioester component that would be required (two amino acid residues N-terminal to the glycosyl amino acid, R1):
(red) residues that cannot be ligated by SAL; (orange) residues that can potentially be ligated by SAL but represent slow ligations; (green) residues that can
be ligated by SAL.
Scheme 2. Synthesis of Glycosyl Amino Acid Building Block 1
Containing a Thiol Auxiliary at the 2-Positiona
Figure 2. Proposed transition states of (a) SAL (14-membered ring) and
(b) extended SAL with six amino acid extensions N-terminal to the
glycosylated residue (29-membered ring); R ) amino acid side-chain
functionality.
of the extended SAL, the transthioesterification, would be
identical to that of SAL (Scheme 1). However, as a result of
adding extra amino acids to the N-terminus, the proposed S f
N acyl transfer would now proceed through 17-29-membered-
ring transition states for single extension to six amino acid
extensions, respectively, compared with a 14-membered-ring
a Conditions: (a) Zn dust, acetic acid, 88%; (b) S-trityl-2-mercaptoacetic
acid, HBTU, DIEA, DMF, 90%; (c) Pd(PPh3)4, NMA, THF, 95%. HBTU
) 2-(1H-benzotriazole-1-yl)-1,1,3,3-tetramethylaminium hexafluorophos-
transition state for SAL (Figure 2). Precedence for S f N acyl
phate, DIEA ) N, N-diisopropylethylamine, NMA ) N-methylaniline.
shifts proceeding through comparable ring sizes encouraged us
in pursuing this strategy.31,32
containing glycopeptides. The synthesis began from the known
â-linked glycosyl amino acid 2, synthesized from (D)-glu-
Results and Discussion
cosamine in five steps (Scheme 2).33-35 Subsequent introduction
Synthesis of Glycosyl Amino Acid Building Block. The
initial phase of the research involved the synthesis of the
glycosylated serine building block 1, which was essential for
solid-phase peptide synthesis (SPPS) of the desired auxiliary
of the thiol auxiliary was achieved in two steps by liberation of
the C-2 amine under reductive conditions, followed by coupling
to S-trityl-2-mercaptoacetic acid to give 3 in 79% yield.23 With
(33) Saha, U. K.; Schmidt, R. R. J. Chem. Soc., Perkin Trans. 1 1997, 1855-
1860.
(31) Sasaki, S.; Shionoya, M.; Koga, K. J. Am. Chem. Soc. 1985, 107, 3371-
3372.
(34) Dullenkopf, W.; CastroPalomino, J. C.; Manzoni, L.; Schmidt, R. R.
Carbohydr. Res. 1996, 296, 135-147.
(32) Gennari, C.; Molinari, F.; Piarulli, U.; Bartoletti, M. Tetrahedron 1990,
46, 7289-7300.
(35) Paulsen, H.; Helpap, B. Carbohydr. Res. 1991, 216, 289-313.
9
J. AM. CHEM. SOC. VOL. 129, NO. 44, 2007 13529

A R T I C L E S
Payne et al.
Table 1. Isolated Yields of Extended Sugar-Assisted Ligation
(SAL) Reactions of Glycopeptides 4-8 and C-Terminal Glycine
Thioester 9a
Scheme 3. SPPS (Fmoc Strategy) of Extended Glycopeptides
4-8a
isolated
ligation
yield
ligation
ligation
method
glycopeptide extension
ligation
junction
half-life
t1
XX
/
2
SAL
Gly
4: GlyVal
5: GlyValLeu
6: GlyArgValLeu
7: GlySerArgValLeu
8: GlyAlaSerArgValLeu
Gly-Gly
Gly-Gly
Gly-Gly
Gly-Gly
Gly-Gly
Gly-Gly
91%
86%
70%
60%
38%b
49%
9 h
exSAL
dexSAL
texSAL
qexSAL
pexSAL
12 h
19 h
19 h
19 h
>24 h
a Conditions: 4:1 v/v NMP:6 M Gn‚HCl 1 M HEPES pH 8.5, 2% PhSH,
37 °C, 96 h. b The HPLC trace of this ligation suggests that the yield was
similar to the texSAL (see Figure 3). The reduced isolated yield is a result
of difficult product isolation by preparative HPLC.
Figure 3. Kinetics for ligation reactions of glycopeptides 4-8 and
C-terminal glycine peptide thioester 9 using SAL (red triangle), exSAL
(blue square), dexSAL (yellow diamond), texSAL (brown circle), qexSAL
(green circle), and pexSAL (black diamond).
a PyBOP ) benzotriazole-1-yl-oxy-tris-pyrrolidino-phosphonium hexaflu-
orophosphate; Pip ) piperidine; TFA ) trifluoroacetic acid; TIS )
triisopropylsilane; TA ) thioanisole; NMM ) N-methylmorpholine.
the desired thiol auxiliary in place, the final step involved
palladium-catalyzed removal of the allyl group to afford the
desired â-glycosyl amino acid building block 1 in 95% yield
(Scheme 2).
Solid-Phase Peptide Synthesis of Glycopeptides. Synthesis
of the N-terminal extended glycopeptides 4-8 was achieved
by SPPS following the Fmoc-strategy, starting from Fmoc-
protected Rink amide resin. To avoid unnecessary expenditure
of monomer 1, only 1 equiv was coupled and the reaction time
was increased to 6 h. The UV absorption of the Fmoc/piperidine
adduct at 302 nm suggested that these conditions provided
acceptable coupling yields. Amino acids possessing sterically
hindered side chains were incorporated N-terminal to the
glycosylated amino acid to demonstrate the applicability of the
method to challenging glycopeptide and glycoprotein sequences.
Specifically, we chose N-terminal amino acid sequences which
were not ligatable by the normal SAL methodology. For
example, glycopeptide 8 contained the sequence Ala-Ser-Arg-
Val-Leu before incorporation of an N-terminal glycine. After
the desired amino acid sequence was coupled, acetate groups
were first removed by hydrazinolysis and the resin
treated with trifluoroacetic acid/thioanisole/triisopropylsilane/
water (17:1:1:1) to release the fully unprotected oligomers
from the solid support. After purification by reverse phase
Figure 4. Structures of model thioester intermediates 10-16 for molecular
dynamics simulations.
HPLC, glycopeptides 4-8 were obtained in yields between 28
and 54%. Peptide thioesters bearing the amino acids glycine,
9
13530 J. AM. CHEM. SOC. VOL. 129, NO. 44, 2007

Extended Sugar-Assisted Ligations
A R T I C L E S
Figure 5. Distribution profile for the distance between the amine of the
N-terminus and the thioester carbonyl in the model thioester intermediates.
histidine, alanine, and tyrosine on the C-terminus were synthe-
sized using SPPS via the Boc strategy (see Supporting Informa-
tion).
Extended Sugar-Assisted Ligation Screen. Initial ligation
reactions of the extended glycopeptides and thioester 9 bearing
a C-terminal glycine residue were conducted under the standard
ligation conditions (6 M Gn‚HCl, 100 mM potassium dihydro-
gen phosphate, pH ) 8.5, 2% PhSH, 37 °C). In our hands, these
conditions caused significant quantities of hydrolyzed thioester,
leading to diminished ligation yields. To circumvent this
problem, a range of mixed solvent ligation conditions were
analyzed. The most effective solvent system proved to be a
mixture of N-methyl pyrrolidinone (NMP) and HEPES buffer
(4:1 v/v NMP/6 M Gn‚HCl, 1 M HEPES, pH 8.5, 2% PhSH,
37 °C) which allowed for facile ligation reactions, coupled with
minimal thioester hydrolysis.36 Gratifyingly, the reaction of all
extended glycopeptides (4-8) and thioester (9) under these
conditions gave ligated products which were isolated in good
yields when compared to the original SAL reaction (Table 1).
The extended SAL (exSAL) gave an isolated yield of 86%,
higher than the double-extended SAL (dexSAL) and triple-
extended SAL (texSAL) which reacted in 70% and 60% yields,
respectively. The quadruple-extended and penta-extended SAL
(qexSAL and pexSAL) reactions proceeded in lower yields (38%
and 49%, respectively).
Kinetic Studies. It is clear from the above study that there
is a significant reduction in ligation yield as amino acids are
added to the N-terminus of the glycopeptides. The next phase
of the research focused on the analysis of the ligation kinetics
in the hope that this would shed some light on the isolated yields
obtained. To this end, a direct rate comparison of the extended
SAL methods with the SAL method reported previously was
conducted.23 The rate of product formation was monitored by
HPLC every 2 h for the first 11 h, and an endpoint was
determined after 24 h (Figure 3). The half-life of the traditional
SAL reaction was 9 h, slightly faster than the exSAL reaction
which displayed a half-life of 12 h (Table 1). In agreement with
the ligation yields shown in Table 1, ligation rates became more
sluggish as additional amino acids were added to the N-terminus.
Interestingly, dexSAL, texSAL, and qexSAL reactions exhibited
similar rates of ligation with half-lives of 19 h (Table 1).
Addition of a sixth amino acid residue in pexSAL resulted in a
Figure 6. Snapshots of molecular dynamics simulations, showing the local
maxima distances (in Å) between the N-terminal amine and the thioester
carbonyl carbon for (a) 10, (b) 11, (c) 12, and (d) 13.
significant drop in ligation rate, whereby the reaction had only
proceeded in 42% after 24 h. These kinetic studies clearly
indicate that SAL is more facile than its extended counterparts;
however, these rates differ by less than an order of magnitude,
and as such, exSAL, dexSAL, texSAL, qexSAL, and pexSAL
represent synthetically useful methods.
Molecular Dynamics Simulations. To rationalize the ob-
served kinetics, molecular dynamics studies were conducted.
The simulations were used to calculate distances between the
two proposed reaction centers in the thioester intermediate, that
is, the N-terminal amine and the carbonyl carbon of the thioester
intermediate. On the assumption that the S f N acyl shift is
the rate-limiting step of the reaction (as was observed in the
kinetic studies), this information could then be used to gain an
understanding of the relative rates of reactions observed. To
this end, six glycopeptides were built using Maestro37 and
molecular dynamics simulations performed using AMBER 9.38
These included a control glycopeptide 10, with no extensions
N-terminal to the glycosylated serine residue and 11-16 (Figure
4), which contain the same extensions as glycopeptides 4-8
described in the ligation and kinetic studies above (Table 1 and
Figure 3). Equilibration of the system was necessary prior to
collection of the data. The density profiles of the glycopeptides
with 0-6 extensions show that all the systems are equilibrated
after 20 ps of constant pressure simulations (see Supporting
Information). Over the subsequent 10 ns of simulation, the
distance profile of all glycopeptides showed random fluctuations
of HN-CO distance and spontaneous transitions between long
and short HN-CO distances (see Supporting Information).
Figure 5 describes the distribution of HN-CO distances that
fall into the range of 3 to 17 Å. It is clear from these studies
that glycopeptide 11 with one glycine extension exhibits the
highest percentage of snapshots with the proximity distance
(37) Schro¨dinger Inc. http://www.schrodinger.com (accessed 08/02/2007).
(38) Case, D. A.; et al., Amber 9; University of California: San Francisco, CA,
2006.
(36) Ficht, S.; Payne, R. J.; Brik, A.; Wong, C. H. Angew. Chem., Int. Ed. 2007,
46, 5975-5979.
9
J. AM. CHEM. SOC. VOL. 129, NO. 44, 2007 13531

A R T I C L E S
Payne et al.
Table 2. Scope of the Extended Sugar-Assisted Ligation (exSAL)a
Table 4. texSAL, qexSAL, and pexSAL of Glycopeptides 6-8 with
C-Terminal Glycine and Alanine Peptide Thioestersa
ligation
junction
-AA2-AA1-
isolated
ligation
yield
isolated
glycopeptide
ligation
junction
ligation
yield
entry
thioester
glycopeptide
entry
thioester
XX
1
2
3
4
5
6
7
8
9
Gly
His
Ala
Gly
His
Ala
Gly
His
Ala
Gly
Gly
Gly
His
His
His
Asp
Asp
Asp
Gly-Gly
His-Gly
Ala-Gly
Gly-His
His-His
Ala-His
Gly-Asp
His-Asp
Ala-Asp
86%
70%
44%
77%
73%
77%
91%
91%
64%
1
2
3
4
5
6
Gly
Gly
Gly
Ala
Ala
Ala
GlyArg
Gly-Gly
Gly-Gly
Gly-Gly
Ala-Gly
Ala-Gly
Ala-Gly
60%
38%
49%
48%
31%
34%
GlySerArg
GlyAlaSerArg
GlyArg
GlySerArg
GlyAlaSerArg
a Conditions: 4:1 v/v NMP/6 M Gn‚HCl 1 M HEPES, pH 8.5, 2% PhSH,
37 °C, 96 h.
a Conditions: 4:1 v/v NMP:6 M Gn‚HCl 1 M HEPES pH 8.5, 2% PhSH,
37 °C, 96 h.
13 (see Supporting Information). For compounds 11-12 this
distance is 5.75-5.91 Å, presumed to be favorable for the S f
N acyl transfer. In contrast, control peptide 10 and extended
glycopeptide 13 exhibited maximal frequency at distances of
8.61 and 8.70 Å, respectively, suggesting a less favorable acyl
shift and hence a slower ligation rate, as was observed in the
kinetic studies (Figure 3).
Table 3. Scope of the Double Extended Sugar-Assisted Ligation
(dexSAL)a
From the distance profile obtained from the molecular
simulation studies of 10 we would still expect a ligation to take
place, albeit at a significantly slower rate (10 displays a
maximum population of 22.4% between 8 and 9 Å, see
Supporting Information). However, it has previously been
established that glycopeptides lacking an amino acid extension
N-terminal to the glycosylated amino acid are unable to ligate
using SAL.23 It is therefore presumed that ring strain in the
transition state is an additional factor that must be considered.
The transition state for glycopeptide 10 would be an 11-
membered ring, compared with a 14-membered ring for the
single extended glycopeptide 11, which is able to ligate rapidly
with a range of thioesters. The 11-membered ring transition state
is therefore presumed to possess significant ring strain, which
results in an inhibited ligation rate.
isolated
ligation junction
-AA2-AA1-
ligation
yield
entry
thioester
glycopeptide
1
2
3
4
5
6
7
8
9
Gly
His
Ala
Tyr
Gly
His
Ala
Tyr
Gly
His
Ala
Tyr
Gly
Gly
Gly
Gly
His
His
His
His
Asp
Asp
Asp
Asp
Gly-Gly
His-Gly
Ala-Gly
Tyr-Gly
Gly-His
His-His
Ala-His
Tyr-His
Gly-Asp
His-Asp
Ala-Asp
Tyr-Asp
70%
65%
53%
44%
48%
44%
28%
24%
76%
60%
49%
68%
10
11
12
Scope of exSAL. To study the effect of other amino acid
residues at the ligation junction in exSAL, the glycopeptide-
peptide thioester pairs shown in Table 2, entries 2-9 were
examined. Glycopeptides containing N-terminal histidine and
aspartic acid residues were synthesized by SPPS (Fmoc strategy)
in an analogous fashion to those described in Scheme 3 (see
Supporting Information). These were reacted with peptide
thioesters containing C-terminal glycine, histidine and alanine
residues using the mixed solvent system conditions. The ligation
between glycopeptide 4 and the peptide thioester 9 bearing a
C-terminal glycine gave an isolated yield of 86% (entry 1). The
peptide thioester containing a C-terminal histidine residue also
ligated in good yield (70%, entry 2), consistent with the
previously reported SAL method.23,24 As a result of changing
the C-terminal amino acid of the thioester to an alanine residue,
the ligation efficiency decreased (44%, entry 3) in accord with
the established reactivity of peptide thioesters in SAL and NCL
reactions.28 Remarkable ligation yields were achieved for
glycopeptides which bear an N-terminal aspartic acid or histidine
residue. The isolated ligation yields for these glycopeptides
exceeded those for the glycine glycopeptide (64-91% yield,
a Conditions: 4:1 v/v NMP/6 M Gn‚HCl 1 M HEPES, pH 8.5, 2% PhSH,
37 °C, 96 h.
between 4 and 6 Å. A significant percentage of the total
population (36%) exhibits HN-CO distances within 6 Å, which
suggests a favorable S f N acyl transfer reaction. This is indeed
the experimental observation, whereby the glycopeptide contain-
ing the identical extension to 11 underwent the most rapid
ligation (Figure 3). As the number of amino acids in the
extension increases from one to six, the percentage of snapshots
that satisfy N-C proximity decreases monotonously, with the
maximum distribution shifting to a longer distance range. When
the length of the glycopeptide is increased to include six amino
acid extensions in 16, the total population within 6 Å HN-CO
distance range is reduced to 6%. An anomaly exists for
glycopeptide 15 containing five extensions, which shows the
longest NH to CO distance out of all the model systems
examined.
To visualize the results of the molecular simulation studies,
Figure 6 shows the snapshot solutions where the NH to CO
distance showed maximal frequency in model compounds 10-
9
13532 J. AM. CHEM. SOC. VOL. 129, NO. 44, 2007

Extended Sugar-Assisted Ligations
A R T I C L E S
Scheme 4. Synthesis of Galactosamine-R-O-threonine Building
Blocks 21 and 22a
reactions with sterically challenging tyrosine peptide thioesters
clearly demonstrates the potential of this method.
Encouraged by the observed flexibility of the exSAL and
dexSAL methods to a variety of amino acids at the ligation
junction, we wondered if amino acids other than glycine could
also be ligated to glycopeptides 6, 7, and 8 Via texSAL, qexSAL,
and pexSAL, respectively. To this end, glycopeptides 6-8 were
reacted with a peptide thioester containing a C-terminal alanine
residue. Gratifyingly, these glycopeptides ligated in satisfactory
yields (48%, 31%, and 34% for texSAL, qexSAL, and pexSAL,
respectively) after 96 h reaction time. As such, although reaction
rates are significantly slower, these examples still represent
potentially useful ligation reactions (Table 4, entries 4-6).
Future research will aim to elucidate how many N-terminal
extensions can be tolerated by the SAL method before the
isolated yields become unsatisfactory.
Synthesis of MUC1 Repeat Starting Unit by exSAL. After
demonstrating the effectiveness of the extended SAL methodol-
ogy for the synthesis of model glycopeptides, we next turned
our attention to the synthesis of a glycopeptide of biological
significance, the MUC1 repeating unit. MUC1 is a heavily
O-glycosylated glycoprotein that is present at the interface
between epithelial cells and their extracellular matrix.39 The
extracellular domain consists of tandem repeating units com-
posed of twenty amino acid residues. In tumor cells, the
expression of MUC1 is drastically increased, which is coupled
with down regulation of certain glycosyltransferases and
concomitant overexpression of several sialyltransferases, result-
ing in prematurely sialylated glycans linked to the MUC1
protein.40 This aberrant glycosylation results in a significant
alteration in the conformation of the peptide backbone thus
leading to the exposure of distinct peptide epitopes, which
become accessible to the immune system.41 As such, the
incorporation of specific glycans to the MUC1 repeating unit
is considered as a promising target for the production of
immunostimulating antigens.42
a Conditions: (a) (i) CCl3CN, K2CO3, CH2Cl2, (ii) Fmoc-Thr-OAllyl,
TMSOTf, DCM, ether, -30 °C, 55% over two steps 2:1 R:â; (b) Zn, Ac2O,
AcOH, 96%; (c) (i) Zn, AcOH, (ii) TrtSCH2CO2H, HBTU, DIEA, DMF,
58% over two steps; (d) Pd(Ph3)4, NMA, THF, 76% for 21 (R ) Ac),
quantitative for 22 (R ) COCH2STrt).
entries 7-9). Surprisingly, even sterically challenging ligation
junctions such as His-His (73%, entry 5), Ala-His (77%, entry
6), and His-Asp (91%, entry 8) gave excellent yields.
Subsequent desulfurization of the ligated products described
in Table 2 furnished glycopeptides containing the native N-acetyl
functionality at the 2-position. This was achieved by hydrogena-
tion conditions using palladium on alumina.23,24 Products were
generally isolated (after HPLC purification) in quantitative yields
(see Supporting Information).
Scope of dexSAL. Encouraged by the implementation of
exSAL to a range of different ligation partners, we examined
the scope of a further extension by application of the dexSAL
method to a variety of amino acids at the ligation junction (Table
3). The glycopeptides were synthesized Via SPPS (Fmoc
strategy, see Supporting Information). These glycopeptides were
reacted with peptide thioesters bearing C-terminal glycine,
alanine, histidine, and tyrosine residues using the previously
described mixed solvent system. Reactions using glycine
glycopeptides afforded ligation products in good yields (44-
70%, entries 1-4). In contrast, ligations of glycopeptides
containing an N-terminal histidine residue proceeded in sig-
nificantly lower yields when compared to the exSAL method
described in Table 2 (24-48%, entries 5-8). This can be
rationalized by a slower rate of ligation in this reaction, validated
by the kinetic and molecular dynamics studies (Figures 3 and
5). Gly-His and His-His ligations were isolated in satisfactory
yields; however, incorporation of alanine and tyrosine peptide
thioesters resulted in a significant decline in ligation yield (28%
and 24%, respectively, entries 7 and 8). Aspartate extended
glycopeptides ligated in good yields (49-76%, entries 9-12)
for all cases; however, these were slightly lower than the exSAL
cases shown in Table 2. This is once again presumed to be due
to a more sluggish rate of ligation for these glycopeptides as
described in the kinetic studies. The ability to conduct dexSAL
As part of our long-term goal to synthesize repeating
oligomers of cancer-associated glycopeptides we chose to
embark on the synthesis of the C-terminal starting unit of the
MUC1 tandem repeat glycoprotein to demonstrate the ap-
plicability of exSAL to this system. In both the C-terminal
starting unit and MUC1 tandem repeat the amino acid adjacent
to the N-terminal glycan is the sterically encumbered valine
residue. This would make the ligation of the corresponding
C-terminal glycine thioester difficult by standard SAL methods.
Extension by a further amino acid would now lead to a
glycopeptide bearing an N-terminal glycine that could be reacted
with a C-terminal histidine peptide thioester. This ligation was
shown to be efficient in the model exSAL studies (Table 2,
entry 2, 70% isolated yield).
To prepare the proposed glycopeptide target, two galac-
tosamine building blocks, R-linked to Fmoc-threonine, were
first synthesized (Scheme 4). Formation of the trichloroace-
timidate and subsequent glycosylation from the known alcohol
(39) Dziadek, S.; Kunz, H. Synlett 2003, 1623-1626.
(40) Dziadek, S.; Brocke, C.; Kunz, H. Chem. Eur. J. 2004, 10, 4150-4162.
(41) Dziadek, S.; Griesinger, C.; Kunz, H.; Reinscheid, U. M. Chem. Eur. J.
2006, 12, 4981-4993.
(42) Dziadek, S.; Kowalczyk, D.; Kunz, H. Angew. Chem., Int. Ed. 2005, 44,
7624-7630.
9
J. AM. CHEM. SOC. VOL. 129, NO. 44, 2007 13533

A R T I C L E S
Payne et al.
Scheme 5. Synthesis of the MUC1 Starting Unit Diglycopeptide by SPPS Followed by exSAL and Desulfurization
1743 gave a 2:1 mixture of R/â isomers in 55% yield. It is
important to note that the use of a mixed solvent system in the
glycosylation reaction (1:1 v/v dichloromethane/diethyl ether)
gave higher selectivity for the R-anomer than that previously
reported.25 Reduction of the desired R-anomer 18 with zinc dust
in the presence of acetic anhydride and acetic acid gave glycosyl
amino acid 19 containing an N-acetyl group in 96% yield.
Alternatively, reduction of 18 with zinc dust and acetic acid,
followed by coupling to S-trityl-2-mercaptoacetic acid gave
glycosyl amino acid 20 containing the thiol auxiliary at the
2-position.25 Finally, the allyl groups of 19 and 20 were cleaved
using Pd(PPh3)4 and NMA to afford the desired glycosyl amino
acid building blocks 21 and 22 in 76% and quantitative yield,
respectively.
(43) Koeller, K. M.; Smith, M. E. B.; Wong, C. H. Bioorg. Med. Chem. 2000,
8, 1017-1026.
9
13534 J. AM. CHEM. SOC. VOL. 129, NO. 44, 2007

Extended Sugar-Assisted Ligations
A R T I C L E S
Figure 7. Histogram representing the O-glycosylation sites from O-GlycBase that could not be synthesized using sugar-assisted ligation (SAL). These sites
have been re-evaluated for applications of the extended SAL methods (exSAL, dexSAL, texSAL, qexSAL, or pexSAL, two to six amino acid residues
N-terminal to the glycosyl amino acid) Key: (red) residues that cannot be ligated by exSAL, dexSAL, texSAL, qexAL, or pexSAL; (orange) residues that
can potentially be ligated by exSAL, dexSAL, texSAL, qexAL, or pexSAL but represent slow ligations; (green) residues that can be ligated by exSAL,
dexSAL, texSAL, qexAL, or pexSAL.
Synthesis of the desired 15-mer diglycopeptide 23 was
achieved by SPPS via the Fmoc strategy (see Supporting
Information). The N-acetylgalactosamine building block 21 was
incorporated as the first residue, and glycosyl amino acid 22,
containing the thiol auxiliary at the 2-position, was coupled as
the third residue from the N-terminus of the glycopeptide
(Scheme 5). After acetate removal, side-chain deprotection, and
cleavage from the resin, the resulting diglycopeptide 24 was
isolated in 15% overall yield. Diglycopeptide 24 was ligated
with the corresponding peptide thioester (Ac-APPAH-S(CH2)2-
CONH2, 25, synthesized by the Boc strategy) under the mixed-
solvent conditions. This furnished the desired 20-mer glyco-
peptide 26 in 70% yield. Finally, desulfurization of the thiol
auxiliary gave the starting unit of the MUC1 repeat 27 in
quantitative yield (Scheme 5). The successful implementation
of the exSAL method for the synthesis of this MUC1 starting
unit will now serve as a platform to sequentially ligate amino
acid units of the MUC1 repeat in future research. It is hoped
that the resulting oligomers may serve as new immunogenic
constructs for the production of immunostimulating antigens
for the development of cancer vaccines.
methodology for ligation of glycopeptides extended by up to
six amino acid residues N-terminal to the glycosylated amino
acid should significantly increase the synthetic accessibility of
glycopeptides (and potentially glycoproteins) by this method.
To demonstrate the proposed generality and potential scope of
the extended SAL methods with respect to the synthesis of
native glycoproteins, we chose to re-examine all targets from
O-GlycBase v6.00 that could not be synthesized using the
traditional SAL methodology. These glycoproteins were re-
evaluated by scanning amino acid residues N-terminal to the
O-glycosylation sites (by up to six amino acids) until a feasible
ligation junction was found. As with the previous informatics
study, a feasible ligation junction was qualified by assessing
the nature of the thioester component of the proposed ligation.
The application of extended SAL disconnections to these
O-glycosylation sites is depicted in Figure 7.
It is clear from Figure 7 that the probability that a given
glycopeptide or glycoprotein target can be synthesized is greatly
increased by the application of exSAL, dexSAL, texSAL,
qexSAL, and pexSAL methodologies. Of the O-glycosylation
sites that could not be synthesized by SAL, 65% are now
accessible by these methods (indicated in green in Figure 7). A
further 26% of O-glycosylation sites represent slower ligations,
Revised Scope of SAL for the Assembly of Naturally
Occurring Glycoproteins. The ability to utilize the SAL
9
J. AM. CHEM. SOC. VOL. 129, NO. 44, 2007 13535

A R T I C L E S
Payne et al.
but remain synthetically accessible (indicated in orange in Figure
7). Notably, a combination of the results of this study and those
described in Figure 1 suggest that 95% of all O-glycosylation
sites reported in O-Glycbase can be utilized for either SAL or
extended SAL reactions. Of these O-glycosylation sites, 62%
represent rapid ligations, while 38% represent slower ligation
reactions.
or extended SAL technologies. As such, this method, either
alone, or in combination with other ligation strategies should
aid in the efficient syntheses of native glycopeptide and
glycoprotein targets that were not previously attainable.
Current efforts in our laboratory are focused on the application
of this method to ligate glycopeptides bearing pre-assembled
glycans, which can be installed using synthetic or enzymatic
methods.
Conclusion
Experimental Section
The inherent limitations in our previously developed SAL
method for the total synthesis of native glycoproteins inspired
the development of extended sugar-assisted ligation (SAL)
methods, namely exSAL, dexSAL, texSAL, qexSAL, and
pexSAL. Ligation products were isolated in high yields when
extended glycopeptides were reacted with peptide thioesters
bearing a C-terminal glycine residue by using a mixed-solvent
system, which served to protect the thioester from hydrolysis.
Kinetic studies indicated that ligation reactions proceeded at
progressively slower rates as further amino acid extensions were
incorporated into the glycopeptide. These observations were
rationalized by the use of molecular dynamics simulations which
suggested that the distances between the required reaction sites
(the N-terminus and the carbonyl carbon of the thioester) in
the proposed thioester intermediate also increased as glycopep-
tides were extended. As such, the rate is presumed to be a
function of the proximity of these two entities and hence the
ease at which the glycopeptide thioester intermediate can
rearrange to afford the ligation product. All five of the extended
SAL methods were shown to exhibit broad tolerance for a
variety of amino acid residues at the ligation junction, including
challenging residues such as alanine and tyrosine. In addition,
the auxiliaries of several ligation products were desulfurized
efficiently to afford glycopeptides displaying the native N-
acetylglucosamine moiety. As a preliminary application, exSAL
was effectively implemented in the synthesis of a naturally
occurring glycopeptide, the starting unit of the MUC1 repeat.
A detailed search of over five hundred O-glycosylation sites
using the glycoprotein database O-GlycBase v6.00 reveals that
many more sequences are accessible (when compared to SAL)
using exSAL, dexSAL, texSAL, qexSAL, or pexSAL. Specif-
ically, 95% of all O-glycosylation sites reported in O-Glycbase
v6.00 can potentially be constructed by applications of the SAL
General Procedure for Extended SAL Reactions. Glycopeptides
(1.5 equiv, approximately 3 µmol) were dissolved in 150 µL of
deoxygenated ligation buffer [4:1 v/v N-methyl-2-pyrrolidinone (NMP)/6
M guanidine hydrochloride, 1 M HEPES, pH ) 8.5]. This solution
was transferred to an eppendorf tube containing the thioester (ca. 2
µmol). Thiophenol (2% by volume, 3 µL) was added, and the reaction
was mixed gently. The ligation mixture was incubated at 37 °C with
gentle mixing every 12 h until the reaction was confirmed to be
complete by LC-MS. The ligation reactions were quenched by the
addition of TCEP solution (0.6 mL of a 10 mg/mL solution), and the
product was purified by HPLC.
General Procedure for the Desulfurization of the Thiol Auxiliary.
Ligated glycopeptides (1.0 mg) were dissolved in a degassed solution
of 6 M Gn‚HCl, 100 mM potassium dihydrogen phosphate, 10 mM
TCEP buffer, pH ) 5.8 (1.0 mL). Palladium on alumina (5 wt %, 16
mg) was added, and the reaction was stirred under an H2 atmosphere
for 1 h. The reaction was filtered, and the product was purified by
HPLC.
Acknowledgment. This work was supported by the NIH and
the Skaggs Institute for Chemical Biology. R.J.P. is grateful
for funding provided by the Lindemann Trust Fellowship. S.F.
is grateful for a postdoctoral fellowship provided by the
Deutsche Akademische Austauschdienst (DAAD). We would
also like to thank Dr. William Greenberg for proof reading this
manuscript.
Supporting Information Available: Complete ref 38; detailed
molecular dynamics simulation procedures and results, experi-
mental procedures, product characterization, and other detailed
results. This material is available free of charge via the Internet
at http://pubs.acs.org.
JA073653P
9
13536 J. AM. CHEM. SOC. VOL. 129, NO. 44, 2007