Salicylaldehyde ester-induced chemoselective peptide ligations: Enabling generation of natural peptidic linkages at the serine/threonine sites

Article abstract of DOI:10.1021/ol1003109

A serine/threonine-based chemoselective ligation method is described. It uses an O-salicylaldehyde ester at the C-terminus, reacting with N-terminal serine or threonine to realize peptide ligations. The utility of the O-salicylaldehyde ester enables the rapid coupling reaction and the production of an N,O-benzylidene acetal intermediate, which is readily converted into natural peptidic linkages (Xaa-Ser/Thr) at the ligation site.

Full text of DOI:10.1021/ol1003109

ORGANIC
LETTERS
2010
Vol. 12, No. 8
1724-1727
Salicylaldehyde Ester-Induced
Chemoselective Peptide Ligations:
Enabling Generation of Natural Peptidic
Linkages at the Serine/Threonine Sites
Xuechen Li,* Hiu Yung Lam, Yinfeng Zhang, and Chun Kei Chan
Department of Chemistry, The UniVersity of Hong Kong, Pokfulam Road,
Hong Kong, China
xuechenl@hku.hk
Received February 6, 2010
ABSTRACT
A serine/threonine-based chemoselective ligation method is described. It uses an O-salicylaldehyde ester at the C-terminus, reacting with
N-terminal serine or threonine to realize peptide ligations. The utility of the O-salicylaldehyde ester enables the rapid coupling reaction and
the production of an N,O-benzylidene acetal intermediate, which is readily converted into natural peptidic linkages (Xaa-Ser/Thr) at the ligation
site.
In addition to recombinant DNA technology to provide
polypeptides/proteins, chemical synthesis has dramatically
contributed to the exploration of the relationship of protein
structure to function. Moreover, with the rapid emergence
of middle-sized peptides (between 20 and 100 amino acids)
as therapeutics,1 such as fuzeon, synthetic peptide chemistry
has attracted a lot of attention. Merrifield’s linear solid-phase
peptide synthesis (SPPS) has provided a general tool to
prepare polypeptides.2 However, to chemically synthesize a
large polypeptide using linear solid-phase peptide synthesis
is very costly and sometimes even impossible. Thus, the
development of a method to achieve a convergent synthesis
using smaller polypeptide fragments becomes critical in terms
of both minimizing the cost of production of peptide
therapeutics1,3 and realizing chemical synthesis of proteins.4
In contrast with conventional fragment coupling, chemose-
lective ligations using partially protected or unprotected
peptide fragments would be advantageous to achieve an
efficient convergent synthesis. Use of unprotected peptides
helps circumvent the difficulty inherent to classical peptide
coupling reactions due to limited solubility; thus, it will
increase coupling yields and lead to easy purification and
characterization. The key issue in the development of peptide
coupling methods using unprotected peptide fragments is the
availability of a chemoselective reaction to specifically and
unambiguously join peptides through the C-terminus of a
peptide and the N-terminus of a second peptide. Using a
chemoselective reaction to join two peptide segments through
formation of an unnatural (i.e., nonpeptide) backbone
structure at the ligation site has permitted the facile prepara-
tion of a wide range of backbone-modified artificial peptides
and proteins.5 However, there are limited methods available
by which to accomplish peptide ligation through formation
(1) (a) Mark, V. Chem. Eng. News 2005, 83, 17–24. (b) Bruckdorfer,
T.; Marder, O.; Albericio, F. Curr. Pharm. Biotech. 2004, 5, 29–43.
(2) Merrifield, R. B. J. Am. Chem. Soc. 1963, 85, 2149–2154.
(4) Nilsson, B. L.; Soellner, M. B.; Raines, R. T. Annu. ReV. Biophys.
Biomol. Struct. 2005, 34, 91–118.
(3) Fuzeon was chemically manufactured using a three-fragement
assembly strategy. (a) Bray, B. L. Nat. ReV. Drug DiscoVery 2003, 2, 587–
(5) For a comprehensive review of chemoselective ligation methods,
see: Hackenberger, C. P. R.; Schwarzer, D. Angew. Chem., Int. Ed. 2008,
47, 10030–10074.
593. (b) Albericio, F. Curr. Opin. Chem. Biol. 2004, 8, 211–221
.
10.1021/ol1003109  2010 American Chemical Society
Published on Web 03/16/2010

of a natural peptide linkage at the ligation site. Clearly,
cysteine-based native chemical ligation (NCL),6 developed
by Kent and co-workers, is the first and the best method in
this regard, and it has become no doubt the most powerful
method in synthetic peptide chemistry. Cysteine-based NCL
features a thio capture between an N-terminal cysteine and
a C-terminal thioester, as a transthioesterification step which
is highly chemoselective, followed by a rapid 1,4 SfN acyl
transfer to afford a natural Xaa-Cys peptidic linkage (Xaa
represents any amino acid). Its efficiency, ease of operation,
and chemoselectivity (in the presence of any unprotected
amino acid) are very attractive to its users/practitioners; thus,
cysteine-based NCL has been widely used for chemically
synthesizing many proteins.7
ligation site.11 The method achieved great chemoselectivity,
but the ligation was extremely slow11 when threonine or
serine was used at the N-terminus, and more importantly,
the conversion from the pseudoproline structure into natural
peptidic bonds was not achieved.11 Therefore, although these
works have conceptually demonstrated the utility of weakly
activated alkylesters as acyl donors for chemoselective
peptide ligation at serine and threonine sites, they are unable
to provide a convergent synthesis of proteins or polypeptides
with the natural peptide sequence.
However, the rare presence of the cysteine residue (1.4%
content in proteins) has limited the utility of NCL (most
peptide pharmaceuticals rarely contain an internal cysteine1).
To address this issue, people have extensively searched for
alternative “native chemical ligation” methods at other amino
acid sites over the past decade.8 We are interested in the
development of ligations at serine9 and threonine sites. We
took note that the cysteine-NCL achieving the chemoselec-
tivity lies in that the N-terminal cysteine can differentiate
itself from other inner unprotected amino acid functional
groups with its bifunctionality: 1,2-mercaptoamine. On the
other hand, the capture-rearrangement chemical ligation
does not involve activating the carboxyl group; thus, it is
effective to overcome the racemization problem of the
conventional segment condensation method. Following the
same logic, it is conceivable that serine and threonine possess
great promise for achieving chemoselective ligation due to
their 1,2-hydroxyamine bifunctionality. We were attracted
by the imine-induced proximity acyl-transfer approach. Such
a strategy was originally introduced by Kemp10 and fully
developed by Tam to ligate a C-terminal glycolaldehyde
peptide with another peptide containing a Cys, Thr, or Ser
residue at the N-terminus to furnish a coupled product with
a pseudoproline structure (thiazolidine or oxazolidine) at the
Figure 1. Proposed Ser/Thr-based chemical ligation resulting in a
natural peptide bond at the ligation site.
We believe that, with further extensive investigations, a
Ser/Thr-based chemical ligation resulting in natural Xaa-Ser/
Thr linkage at the ligation site using such an imine
capture-rearrangement strategy can be realized. We con-
ceived of the idea of using a two-step strategy (Figure 1).
The first step involves the amine group of the N-terminal
serine or threonine reversibly reacting with the aldehyde
group of the C-terminus to form an imine, followed by the
cyclization from the R-hydroxyl group of the N-terminal
serine or threonine. Next, an OfN acyl transfer affords a
stable acetal intermediate. Other nucleophiles such as lysine
may also react with the aldehyde group, but the reaction is
reversible and unable to generate a stable product. The
second step is to remove the formed acetal group to afford
the natural peptide bond. In order to develop a practical
chemoselective peptide ligation based on the above strategy,
two questions must be addressed: (1) whether either the imine
capture step or the acyl transfer step can be accelerated (it
is not known which step is the rate-determining step) and
(2) whether the formed pseudoproline moiety12 can be readily
transformed into natural peptide linkages. The second
question is more challenging and critical because the removal
conditions should ideally be simple without any sophisticated
chemistry and compatible with other functionalities in
peptides or even oligosaccharides in glycopeptides.
(6) Dawson, P. E.; Muir, T. W.; Clark-Lewis, I.; Kent, S. B. H. Science
1994, 266, 776–779.
(7) (a) Kent, S. B. H. Chem. Soc. ReV. 2009, 38, 338–351. (b)
Muralidharan, V.; Muir, T. W. Nature Methods 2006, 3, 429–438. (c)
Cotton, G. J.; Muir, T. W. Chem. Biol. 1999, 6, R247–R256.
(8) For methods by post-modifying the cysteine-NCL product, see: (a)
Yan, L. Z.; Dawson, P. E. J. Am. Chem. Soc. 2001, 123, 526–533. (b) Wan,
Q.; Danishefsky, S. J. Angew. Chem., Int. Ed. 2007, 46, 9248–9252. (c)
Okamoto, R.; Kajihara, Y. Angew. Chem., Int. Ed. 2008, 47, 5402–5406.
(d) Canne, L. E.; Bark, S. J.; Kent, S. B. H. J. Am. Chem. Soc. 1996, 118,
5891–5896. (e) Offer, J.; Boddy, C. N. C.; Dawson, P. E. J. Am. Chem.
Soc. 2002, 124, 4642–4646. (f) Wu, B.; Chen, J.; Warren, J. D.; Chen, G.;
Hua, Z.; Danishefsky, S. J. Angew. Chem., Int. Ed. 2006, 45, 4116–4125.
(g) Crich, D.; Banerjee, A. J. Am. Chem. Soc. 2007, 129, 10064–10065.
(h) Haase, C.; Rohde, H.; Seitz, O. Angew. Chem., Int. Ed. 2008, 47, 6807–
6810. (i) Chen, J.; Wan, Q.; Yuan, Y.; Zhu, J.; Danishefsky, S. J. Angew.
Chem., Int. Ed. 2008, 47, 8521–8524. Other significant non-cysteine NCL-
derived chemoselective ligations include traceless Staudinger ligation and
decarboxylative amide ligation. (j) Nisson, B. L.; Kiessling, L. L.; Taines,
R. T. Org. Lett. 2000, 2, 1939–1941. (k) Tam, A.; Soellner, M. B.; Raines,
R. T. J. Am. Chem. Soc. 2007, 129, 11421–11430. (l) Bode, J. W.; Fox,
R. M.; Baucom, K. D. Angew. Chem., Int. Ed. 2006, 45, 1248–1252. For
a sugar-assisted ligation, see: (m) Brik, A.; Yang, Y. Y.; Ficht, S.; Wong,
C.-H. J. Am. Chem. Soc. 2006, 128, 5626–5627.
Among many candidates, we tentatively identified the
(thio)salicylaldehyde ester as a potential functional donor at
(11) (a) Liu, C.-F.; Tam, J. P. Proc. Natl. Acad. Sci. U.S.A. 1994, 91,
6584–6588. (b) Liu, C.-F.; Tam, J. P. J. Am. Chem. Soc. 1994, 116, 4149–
4153. (c) Tam, J. P.; Miao, Z. J. Am. Chem. Soc. 1999, 121, 9013–09022.
(12) Serine and threonine pseudoprolines in the form of ψMe,Me pro were
originally developed by Mutter and widely used in Fmoc SPPS: Haack, T.;
Mutter, M. Tetrahedron Lett. 1992, 33, 1589–1592.
(9) A “serine” ligation was reported deriving from post-modification of
a cystein-NCL product; see ref 7c.
(10) Kemp, D. S. Biopolymer 1981, 20, 1793–1804. (b) For a compre-
hensive review, see: Coltart, D. M. Tetrahedron 2000, 56, 3449–3491.
Org. Lett., Vol. 12, No. 8, 2010
1725

the C-terminus (Figure 1). We hypothesized that, when the
benzaldehyde ester is used, either the hydroxyl addition to
the imine or OfN acyl transfer would proceed faster.
However, as the penalty, the acyl transfer has to progress
through an unfavored 1,5-acyl transfer. More importantly,
the formed N,O-benzylidene acetal group is expected to be
readily removed under acidic conditions. Potentially, the
reactivity of the coupling step and removal step could be
tuned by introducing certain substituents on the phenyl ring
of the salicylaldehyde. Installation of photoreactive groups
onto the phenyl ring of the salicylaldehyde ester, such as a
nitro group, would likely enable the removal of the N,O-
benzylidene acetal under UV light.
intermediate derived from coupling between Fmoc-Ala-
salicylaldehyde ester and serine was observed in almost
quantitative yield. Impressively, the thiophenol ester is a
rather reactive acyl donor and has been used for direct
aminolysis, but it is completely inert under this condition
and does not compete with the serine reacting with the
salicylaldehyde ester.
We began by testing the hypothesis with the commercially
available salicylaldehyde. As a model study, we prepared
FmocAla 1 with a salicylaldehyde ester at the C-terminus
(Figure 2). To our delight, the coupling between 1 and serine
benzyl ester 2 proceeded smoothly in pyridine/acetic acid11
(1:1, mol/mol), resulting in 85% conversion in 30 min and
a quantitative conversion in 2 h to afford N,O-benzylidene
acetal intermediate 3.13 Furthermore, after evaporation of the
solvent (pyridine/acetic acid), the crude reaction mixture was
directly subjected to the next step. The removal of the acetal
group can be readily achieved under acidic conditions (TFA/
H2O/i-Pr3SiH, 5 min) with quantitative conversion to afford
the “natural” Ala-Ser dipeptide 4. Such conditions are
generally used in peptide synthesis and even glycoprotein
synthesis.14
Figure 3. Chemoselectivity studies.
With this success in hand, we next explored the scope of
the coupling reaction between N-terminal serine or threonine
and the C-terminal salicylaldehyde ester of some rather
hindered ꢀ-branched amino acid sites (e.g., Val, Pro, Ile, and
Thr). These ꢀ branched amino acids when used at the
C-terminus generally retard the peptide coupling step dra-
matically; thus, most known ligation methods require a
prolonged time for completion at these amino acid sites16
(i.e., 48 h for cysteine-NCL) or are limited to less hindered
amino acids at the C-terminal site. However, these amino
acids possessing a salicylaldehyde ester at the C-terminus
react with both serine and threonine derivatives surprisingly
rapidly, resulting in >70% conversion after 30 min and
completion within 5 h.17 Treatment with (TFA/H2O/i-Pr3SiH)
for 5 min gave rise to peptides with natural peptide bonds
at the ligation site (Table 1). Tripeptide Val-Thr-Ser-OBn
(entry 5) was isolated in diastereomerically pure form,
without detection of the epimerized diastereomer by LCMS
and NMR.18
Figure 2. Model study.
To demonstrate the feasibility of this ligation strategy in
peptide segment ligation using unprotected peptides, a two-
step sequential ligation was carried out between the O-
salicylaldehyde ester peptide 519 and unprotected peptide 6
possessing an N-terminal serine (Figure 4). With only one
purification, the desired peptide 7 with a natural peptide bond
(Thr-Ser) at the ligation site was obtained in 72% yield. It
is noteworthy that peptide 7 possesses a thiophenol ester at
the C-terminus. The thiophenol ester, being a rather reactive
acyl donor in chemical ligation, is completely inert and fully
The phenolic ester has been known to directly condense
with amines to afford amides.15 To rule out this direct
condensation possibility and evaluate the chemoselectivity,
two convincing competition experiments were carried out.
As outlined in Figure 3, we compared the reactivities between
an alanine and a serine derivative when coupling with
O-salicylaldehyde ester 1 and the reactivities between a
thiophenol ester and a salicyladehyde ester when coupling
with serine 2. In both cases, only the N,O-benzylidene acetal
(16) Hackeng, T. M.; Griffin, J. F.; Dawson, P. E. Proc. Natl. Acad.
Sci. U.S.A. 1996, 96, 10068–10073.
(13) The N,O-benzylidene acetal intermediate between BocNH-Ala-
salicylaldehyde ester and H2N-Ser-OMe was fully characterized and studied;
see the Supporting Information.
(17) NH2-Thr-Phe-OEt dipeptide was used here. Under the reaction
conditions, the formation of a diketopiperazine was not observed.
(18) The epimerization issue was also studied using Fmoc-Ala-Phe(L)-
salicylaldehyde ester and Fmoc-Ala-Phe(D)-salicylaldehyde ester; see the
Supporting Information.
(14) Yamamoto, N.; Tanabe, Y.; Okamoto, R.; Dawson, P. E.; Kajihara,
Y. J. Am. Chem. Soc. 2008, 130, 501–510.
(15) (a) Chen, G.; Wan, Q.; Tan, Z.; Kan, C.; Hua, Z.; Ranganathan,
K.; Danishefsky, S. J. Angew. Chem., Int. Ed. 2007, 46, 7383–7387. (b)
Yuan, Y.; Chen, J.; Wan, Q.; Tan, Z.; Chen, G.; Kan, C.; Danishefsky,
S. J. J. Am. Chem. Soc. 2009, 131, 5432–5437.
(19) Peptide salicylaldehyde esters could be synthesized using Fmoc
SPPS chemistry with the protocol developed by Dawson for the preparation
of peptide thiosesters: Blanco-Canosa, J. B.; Dawson, P. E. Angew. Chem.,
Int. Ed. 2008, 47, 6851–6855.
1726
Org. Lett., Vol. 12, No. 8, 2010

Table 1. Scope of the Reaction between Serine/Threonine and
Salicylaldehyde Estersa
Ab
(%)
Ac
(%)
R1
R2
Bd
A-S
V-S
P-S
%
1
2
3
4
5
6
7
8
Ala
Val
Pro
Ile
Val-Thr
Ala
Val
Pro
Ile
Ser-OBn
Ser-OBn
Ser-OBn
Ser-OBn
85
80
71
70
76
76
73
75
80
74
90
NRf
NR
100
100
100
100
100
100
100
100
100
100
95
98
99
95
97
99
96
98
98
92
86
I-S
Ser-OBn
VT-S
A-TF
V-TF
P-TF
I-TF
VT-TF
K-S
Thr-Phe-OEt
Thr-Phe-OEt
Thr-Phe-OEt
Thr-Phe-OEt
Thr-Phe-OEt
Ser-OMe
Figure 4. Serine ligation with a unprotected peptide possessing a
thiophenyl ester at the C terminus.
9
10
11
12
13
Val-Thr
Lys(TFA)e
Ala
ceuticals contain internal serine or threonine, such as forteo,
fuzeon, nesiritide, and corticoliberin; thus, this Ser/Thr-
induced chemoselective peptide ligation method should find
great applications in synthetic peptide chemistry and peptide
pharmaceutical manufacturing. Next, we will apply this
method to convergently synthesize the antiretroviral drug
fuzeon and glycopolypeptides.
Ala-OMe
Phe-OEt
NR
NR
Ala
a All reactions were performed at the concentration of around 0.05 M.
b The reaction conversion at 30 min. Five microliters of sample from the
reaction mixture was taken, diluted with acetonitrile/water, and directly
injected into the LCMS system; the conversion is calculated on the basis
of the consumption of the salicylaldehyde. c Conversion at 5 h. d The
deprotection step is performed between 5 and 10 min. e Derived from the
treatment of Fmoc-Lys(Boc)-O-salicylaldehyde ester with TFA. f NR: no
reaction.
Acknowledgment. We thank the Department of Chemistry
and Faculty of Science of The University of Hong Kong for
providing a startup fund and a Unversity Development Fund
for purchasing a LCMS. We thank Ms. Bonnie Yan at The
University of Hong Kong for the help with recording NMR
spectra.
orthogonal to this serine/threonine based chemical ligation.
This outcome adds high value to this ligation protocol
because the C-terminal thioester peptide is ready for further
elongation from the N terminus to the C terminus.
In summary, we reported herein a new method utilizing
Ser and Thr at the N-terminus to achieve a chemoselective
ligation resulting in natural Xaa-Ser and Xaa-Thr bonds
at the ligation site. Serine and threonine are well represented
(12.7%) in proteins, and many synthetic peptide pharma-
Supporting Information Available: Synthetic methods
and characterization. This material is available free of charge
via the Internet at http://pubs.acs.org.
OL1003109
Org. Lett., Vol. 12, No. 8, 2010
1727