Native chemical ligation at valine: A contribution to peptide and glycopeptide synthesis

Article abstract of DOI:10.1002/anie.200803523

(Chemical Equation Presented) A Val-uable link: The title transformation is achieved by a two-step ligation, radicalbased desulfurization strategy (see scheme; NCL=native chemical ligation). After S→N acyl transfer, in which the acyl acceptor is a γ-thiol valine derivative, and site-specific dethiolation, a valine residue appears at the site of ligation. This method accomplishes ligations at Thr-Val and Pro-Val sites, and allows successful ligation of glycopeptide fragments.

Full text of DOI:10.1002/anie.200803523

Angewandte
Chemie
DOI: 10.1002/anie.200803523
Synthetic Methods
Native Chemical Ligation at Valine: A Contribution to Peptide and
Glycopeptide Synthesis**
Jin Chen, Qian Wan, Yu Yuan, Jianglong Zhu, and Samuel J. Danishefsky*
Our laboratory has been pursuing the total synthesis of
naturally occurring glycoproteins bearing multiple oligosac-
charide domains. Specifically, efforts are well underway to
accomplish a de novo total synthesis of erythropoietin alpha
(EPO) in homogeneous form.[1] Although a variety of peptide
ligation strategies have been developed to facilitate the
merger of large, complex peptide and glycopeptide frag-
ments,[2–9] the need for highly efficient methodology continues
to motivate the chemical community to develop more power-
ful strategies. Our pursuit of the total synthesis of homoge-
neous EPO, as well as other biologically active glycopeptides,
has inspired new glycopeptide ligations.[10] To achieve our
most complex goals, we must learn how to overcome the
serious obstacles in joining glycopeptides in an iterative
fashion.
Native chemical ligation (NCL), developed by Kent and
co-workers,[3] constituted a fundamental advance, allowing
the joining of two substantial peptide domains. An additional
development provided by our group as well as others serves to
extend the NCL method to the assembly of peptides bearing
multiple sites of glycosylation.[10a,11] These methods currently
require the presence of a cysteine residue at the N-terminus of
the peptide coupling partner.
Application of the NCL method is often limited by the
paucity of cysteine residues in naturally occurring proteins
and glycoproteins. Several strategies have been investigated
to circumvent the need for a cysteine in the target at the
proposed ligation site. One approach involves appending an
auxiliary thiol group to the N-terminal amino acid, and
subsequent to ligation cleaving the auxiliary.[10b,c,4] This
approach suffers from certain limitations, as the reaction
may be inefficient at more hindered ligation sites and
difficulties can arise at the stage of auxiliary removal.
A second means by which to circumvent the need for a
resident cysteine residue involves the use of an amino acid
surrogate containing a thiol moiety. After ligation the
surrogate is converted into the desired amino acid. In this
context, ligation at a serine site has been achieved by post-
ligational conversion of cysteine into serine.[5] Similarly, a
formal methionine ligation has been accomplished by homo-
cysteine coupling, and subsequent post-ligational methyla-
tion.[6] Recently, Wong and co-workers reported a protocol
for cysteine-free thioester ligation,[2,7] which allows for
peptide assembly. Although lysine residues must be protected
and the reaction usually takes more than 48 hours even at
rather unhindered ligation sites, the method is quite promis-
ing.
In addition, two-step ligation/metal-based thiol reduction
protocols, which formally serve to accomplish ligation at N-
terminal alanine[8] and phenylalanine[9] residues, have been
developed. However, these methods may not be compatible
with a large range of functionalities, particularly sulfur-
containing groups which are frequently present in peptide
sequences. Alternatively, our group recently disclosed a mild
and highly versatile free-radical cysteine reduction protocol,
which tolerates all thiol-containing groups, as well as oligo-
saccharide domains.[10e]
Theoretically, native chemical ligation could be achieved
at any amino acid site, in the sense proposed by Yan and
Dawson,[8] in which a sulfhydryl group is temporarily installed
at a non-cysteine site. The implementation of this concept
requires that the key mechanistic steps (i.e. trans-thioester-
ification and then S!N acyl transfer) be operative. Post-
ligation desulfurization would then afford the desired peptide
or glycopeptide adduct possessing the natural amino acid
residue at the ligation site.
Herein, a strategy for implementing the logic of native
chemical ligation at valine residues is described. Valine is a
rather abundant amino acid, with approximately a 6.6%
frequency in nature, (compared to 1.7% for cysteine). There
are two potential valine surrogates, b-thiol-containing valine
(penicillamine) and a g-thiol valine. We envisioned that the g-
thiol valine, which contains a more reactive primary thiol
group, would serve as a more suitable precursor. As
illustrated in Scheme 1, an N-terminal thiol-modified valine
derivative, installed on peptide 2, would react with the
peptide 1 thioester by trans-thioesterification. The resultant
thioester-linked intermediate would then undergo a rapid
intramolecular acyl transfer, thus creating an amide bond.
Radical-based desulfurization would serve to remove the
thiol moiety and provide the desired peptide with valine at the
ligation site.
[*] Prof. S. J. Danishefsky
Department of Chemistry, Columbia University
3000 Broadway, New York, NY 10027 (USA)
Fax: (+1)212-772-8691
E-mail: s-danishefsky@mskcc.org
Dr. J. Chen, Dr. Q. Wan, Dr. Y. Yuan, Dr. J. Zhu, Prof. S. J. Danishefsky
Laboratory for Bioorganic Chemistry
Sloan-Kettering Institute for Cancer Research
1275 York Avenue, New York, NY 10065 (USA)
[**] Support for this research was provided by the National Institutes of
Health (CA28824). We thank Dr. George Sukenick, Hui Fang, and
Sylvi Rusli of the Sloan-Kettering Institute’s NMR core facility for
mass spectral and NMR spectroscopic analysis (SKI core grant no.:
CA02848). We would like to express our appreciation to Rebecca
Wilson for her assistance in preparing the manuscript.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.200803523.
Angew. Chem. Int. Ed. 2008, 47, 8521 –8524
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
8521

Communications
their results and our own (see below) for
the case of peptides bearing less hindered
C-terminal amino acids, penicillamine can
serve in fostering valine ligation. However,
our primary goal was that of developing a
logic which would encompass more gener-
ally applicable ligations including certain
b-branched C-terminal acyl donors. Our
successes in this regard are described
below.
Given the greater reactivity and dimin-
ished steric hindrance of a primary thiol
relative to a tertiary thiol, we felt that it
Scheme 1. Native chemical ligation at valine.
was worthwhile to develop a protocol for
the efficient synthesis of a modified valine
Penicillamine (2; Scheme 2), containing a b-thiol moiety,
is a commercially available valine surrogate. As shown in
Scheme 2, penicillamine was introduced to the N-terminus of
the peptide as the acyl acceptor, and the NCL was performed
under standard conditions. Because of the tertiary nature of
surrogate with a thiol group installed at its g position.
Drawing from the earlier work of Rapoport and Wolf[13] in
which a g-hydroxy valine was obtained from l-aspartic acid,
we modified the route thereby enabling the preparation of a
g-thiol valine derivative (11), albeit as two diastereomers.
Either epimer of 11 could, in principle, be used for our
purposes since the removal of the thiol group after the
ligation would eliminate the chirality at the b position. Thus,
as outlined in Scheme 3, the protected l-aspartic acid diester
6 was prepared from Fmoc-Asp-OtBu (see the Supporting
Information for details). The PhFl (9-(9-phenylfluorenyl))
group was used to block the a center, providing exclusively b-
alkylated compound 7. Regioselective reduction of 7 fur-
nished alcohols 8 and epi-8 as a mixture of diastereomers
(1:1), which were readily separated by chromatography. After
mesylation and treatment with the DBU salt of thioacetic
acid, acetylated thiol 9 was in hand. The latter was advanced
to the target compound, 11, in a straightforward manner, as
shown.
Having the required thiol-containing amino acids in hand,
we next sought to probe the versatility of our newly
developed protocol. A variety of substrates incorporating a
range of relevant functionalities was evaluated (Table 1). In
the cases of the less hindered C-terminal amino acids, such as
Gln (Table 1, entry 1) and Phe (Table 1, entry 2), the ligation
Scheme 2. Reagents and conditions: A. a) buffer at pH 6.5 (6.0m
Gn·HCl, 188.8 mm Na2HPO4), TCEP, RT, 9 h, 66% yield; b) TCEP,
VA-044, tBuSH, 378C, 3 h, 79% yield. B. c) buffer at pH 6.5 (6.0m
Gn·HCl, 188.8 mm Na2HPO4), TCEP, RT, 10 h, 4% yield based on LC-
MS trace. VA-044=2,2’-azobis-[2-(2-imidazolin-2-yl)propane] dihydro-
chloride. TCEP=tris(2-carboxyethyl)phosphine.
the thiol group, ligation proceeded rather slowly. Though the
ligation yield was reasonable in a case in which glutamine was
present at the C-terminus (Scheme 2A), when the more bulky
threonine residue was incorporated at the C-terminus the
reaction became prohibitively slow (Scheme 2B). During our
preparation of this communication, Seitz and co-workers
reported a ligation culminating in valine, using penicillamine
as the valine precursor.[12] They successfully achieved peptide
bond-formation when glycine, histidine, methionine, and
leucine were presented at the C-terminus. On the basis of
Scheme 3. Reagents and conditions: a) KHMDS, MeI, THF, ꢀ788C,
3 h, 97%; b) DIBAL-H, THF, ꢀ358C, 1 h, 83% (combined yield for 8
and epi-8); c) MsCl, Et3N, CH2Cl2, 08C, 1 h; d) AcSH, DBU, DMF, RT,
16 h, 73% in 2 steps; e) 1n NaOH, MeOH, 08C, 10 min; f) MMTS,
Et3N, CH2Cl2, RT, 30 min, 86% in 2 steps; g) HCl in EtOAc, RT, 82%.
KHMDS=potassium 1,1,1,3,3,3-hexamethyldisilazane; DIBAL-H=
diisobutylaluminum hydride; Ms=methylsulfonyl; Ac=acetyl;
DBU=1,8-diazabicyclo[5.4.0]undec-7-ene; MMTS=S-methyl methane-
thiosulfonate.
8522
www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 8521 –8524

Angewandte
Chemie
Table 1: NCL at valine by ligation and subsequent free-radical desulfurization.[a]
Entry Peptide 1
Peptide 2 Ligation
Product/Yield[b]/Time
Desulfurization
Product/Yield[b]/Time
1
14
2
3
4
5
14
14
15
14
[a] For detailed reaction conditions, please see the Supporting Information. [b] Yield of the isolated product. Fmoc=9-fluorenylmethoxycarbonyl.
proceeded very rapidly (1 h). By comparison, in the analo-
gous case of penicillamine as an acyl acceptor (2), ligation
with the same peptide (1) proceeded at a much slower rate
(9 h; Scheme 2A). As the C-terminus became sterically more
hindered, (Table 1, entries 3 and 4), the reaction rate dropped
correspondingly, but the ligations were still completed within
a reasonable time range and in good yield. Table 1, entry 3,
records an example of a non-cysteine driven NCL reaction in
which both peptides present b-branched amino acids (Thr and
Val) the ligation site. In contrast, the analogous reaction of
substrate 4 with penicillamine (2) displayed much lower
reactivity (Scheme 2B); only 4% product was formed in
10 hours. A range of C-terminal esters was investigated,
including the standard thiophenyl ester (Table 1, entries 3 and
4) as well as an ortho-thiophenolic ester, developed in our
laboratory to allow glycopeptide ligation (Table 1,
entry 1).[10a] We recently described a novel direct oxo-ester
ligation protocol, which circumvents the need for a thioester
intermediate, and allows ligation at sterically hindered C-
terminal sites.[14] In this context, we were pleased to find that
both the C-terminal p-cyanophenyl ester (Table 1, entry 2),
and the p-nitrophenyl ester (Table 1, entry 5) readily partici-
pated in our two-step valine ligation protocol. In addition, the
combination of a C-terminal p-nitrophenyl ester and an N-
terminal g-thiol valine furnished the ligation at a Pro-Val site
(Table 1, entry 5), which is extremely sterically demanding.
Additionally, substrates incorporating unprotected lysine
residues (Table 1, entry 2), as well as thiazolidine and
cysteine(Acm) moieties (Table 1, entry 1) were examined to
probe the limits of our system. In each of these cases, ligation
and subsequent reduction proceeded smoothly, providing the
desired peptide adducts in moderate to good overall yield.
Interestingly, it was found that although both g-thiol
valine epimers (i.e. 11 and epi-11) participate in the ligation
protocol, 11 is not markedly more reactive than epi-11, as
evidenced by a comparison of entries 3 and 4 in Table 1. Thus,
when 11 serves as the N-terminal amino acid, as in 14, ligation
proceeds in 4 hours to yield the product in 87% yield. An
analogous ligation using 15, which incorporated epi-11 as the
N-terminal amino acid, was somewhat slower, presumably
because of a 1,2-cis steric interaction between the b-methyl
group and the amide moiety during the S!N acyl transfer
step. That the difference in rates is relatively modest, suggests
that the rate-determining step in the coupling is the initial
intermolecular transthioesterification rather than the intra-
molecular S!N acyl transfer.
We next investigated the efficiency of the glycopeptide
coupling employing g-thiol valine 11 as a key partner. Thus,
glycopeptide 20, with an N-linked glycan, and a C-terminal
ortho-thiophenolic ester was prepared and coupled with
peptide 14 (Scheme 4). The ligation proceeded smoothly
and provided adduct 21’ in 1 hour. Subsequent thiol reduction
generated desired product 21 in nearly quantitative conver-
sion (Figure 1).
Angew. Chem. Int. Ed. 2008, 47, 8521 –8524
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
8523

Communications
This method represents an important extension
of existing NCL techniques, and should be
valuable in our ongoing programs to bring
synthesis to bear in domains which had previ-
ously been considered to be inaccessible to
chemistry-based interventions.[15]
Received: July 19, 2008
Revised: August 20, 2008
Published online: October 2, 2008
Keywords: desulfurization · glycopeptides ·
.
native chemical ligation · peptides · valine
[1] a) P. Lai, R. Everett, F. Wang, T. Arakawa, E.
Goldwasser, J. Biol. Chem. 1986, 261, 3116 –
3121; b) D. M. Ridley, F. Dawkins, E. Perlin, J.
Natl. Med. Assoc. 1994, 86, 129 – 135.
[2] a) J. Blake, Int. J. Pept. Protein Res. 1981, 17,
273 – 274; b) S. Aimoto, N. Mizoguchi, H. Hojo,
S. Yoshimura, Bull. Chem. Soc. Jpn. 1989, 62,
524 – 531; c) S. Aimoto, Biopolymers 1999, 51,
247 – 265.
[3] P. E. Dawson, T. W. Muir, I. Clark-Lewis, S. B. H.
Kent, Science 1994, 266, 776 – 779.
Scheme 4. Reagents and conditions: a) buffer at pH 6.5 (6.0m Gn·HCl, 188.8 mm
Na2HPO4), TCEP, RT, 30 min, 90% yield; b) TCEP, VA-044, tBuSH, 378C, 3 h, 89%
yield.
[4] D. W. Low, M. G. Hill, M. R. Carrasco, S. B. H.
Kent, Proc. Natl. Acad. Sci. USA 2001, 98, 6554 –
6559.
[5] R. Okamoto, Y. Kajihara, Angew. Chem. 2008,
120, 5482 – 5486; Angew. Chem. Int. Ed. 2008, 47,
5402 – 5406.
In summary, an efficient and broadly useful two-step
valine ligation protocol has been developed. The g-thiol
valine (11), containing a primary thiol group, serves as a
valine surrogate. The combination of the NCL method and
the metal-free reduction provides an efficient, two-step
protocol by which to accomplish a formal valine ligation.
[6] a) J. P. Tam, Q. Yu, Biopolymers 1998, 46, 319 – 327; b) K.
Pachamuthu, R. R. Schmidt, Synlett 2003, 659 – 662; c) A.
Saporito, D. Marasco, A. Chambery, P. Botti, C. Pedone, M.
Ruvo, Biopolymers 2006, 83, 508 – 518.
[7] R. J. Payne, S. Fichet, W. A. Greenberg, C.-H. Wong, Angew.
Chem. 2008, 120, 4483 – 4487; Angew. Chem. Int. Ed. 2008, 47,
4411 – 4415.
[8] L. Z. Yan, P. E. Dawson, J. Am. Chem. Soc. 2001, 123, 526 – 533.
[9] D. Crich, A. Banerjee, J. Am. Chem. Soc. 2007, 129, 10064 –
10065.
[10] a) J. D. Warren, J. S. Miller, S. J. Keding, S. J. Danishefsky, J. Am.
Chem. Soc. 2004, 126, 6576 – 6578; b) G. Chen, J. D. Warren, J.
Chen, B. Wu, Q. Wan, S. J. Danishefsky, J. Am. Chem. Soc. 2006,
128, 7460 – 7462; c) B. Wu, J. Chen, J. D. Warren, G. Chen, Z.
Hua, S. J. Danishefsky, Angew. Chem. 2006, 118, 4222 – 4231;
Angew. Chem. Int. Ed. 2006, 45, 4116 – 4125; d) G. Chen, Q.
Wan, Z. Tan, C. Kan, Z. Hua, K. Ranganathan, S. J. Danishefsky,
Angew. Chem. 2007, 119, 7527 – 7531; Angew. Chem. Int. Ed.
2007, 46, 7383 – 7387; e) Q. Wan, S. J. Danishefsky, Angew.
Chem. 2007, 119, 9408 – 9412; Angew. Chem. Int. Ed. 2007, 46,
9248 – 9252.
[11] For reviews on glycopeptide and glycoprotein synthesis, see
a) B. G. Davis, Chem. Rev. 2002, 102, 579 – 601; b) M. J. Grogan,
M. R. Pratt, L. A. Marcaurelle, C. R. Bertozzi, Annu. Rev.
Biochem. 2002, 71, 593 – 634; c) L. Liu, C. S. Bennett, C.-H.
Wong, Chem. Commun. 2006, 21 – 33.
[12] C. Haase, H. Rohde, O. Seitz, Angew. Chem. 2008, 120, 6912-
6915; Angew. Chem. Int. Ed. 2008, 47, 6807-6810.
[13] a) H. Rapoport, J. P. Wolf, J. Org. Chem. 1989, 54, 3164 – 3173;
b) P. Gmeiner, P. L. Feldman, M. Y. Chu-Moyer, H. J. Rapoport,
J. Org. Chem. 1990, 55, 3068 – 3074.
[14] Q. Wan, J. Chen, Y. Yuan, S. J. Danishefsky, 2008, submitted.
[15] R. M. Wilson, S. J. Danishefsky, Pure Appl. Chem. 2007, 79,
2189 – 2216.
Figure 1. Synthesis of glycopeptide 21. LC-MS trace of: A) Ligation
between glycopeptide 20 and peptide 14 after 30 min, 21’ is the ligation
product. B) 21’ after purification by HPLC methods; observed mass
[M+2H]2+ =1263.86, [M+3H]3+ =842.96. C) Desulfurization of 21’ after
3 h; 21 is the desulfurization product. D) 21 after purification by HPLC
methods; observed mass [M+2H]2+ =1248.02, [M+3H]3+ =832.23.
8524
www.angewandte.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2008, 47, 8521 –8524