Synthesis of nα-(1-phenyl-2-mercaptoethyl) amino acids, new building blocks for ligation and cyclization at non-cysteine sites: Scope and limitations in peptide synthesis

Article abstract of DOI:10.1021/jo049471k

A new and convenient method for the synthesis and incorporation of N ^α-(1-phenyl-2-mercaptoethyl)-derivatized amino acids applicable to chemical ligation at non-cysteine sites is presented. N^α- Auxiliary derivatives of glycine and al

Full text of DOI:10.1021/jo049471k

Synthesis of Nr-(1-Phenyl-2-mercaptoethyl) Amino Acids, New
Building Blocks for Ligation and Cyclization at Non-Cysteine
Sites: Scope and Limitations in Peptide Synthesis
Sylvie Tchertchian,†,‡ Oliver Hartley,† and Paolo Botti†,‡,*
Department of Structural Biology and Bioinformatics, University of Geneva, 1 rue Michel Servet,
1211 Geneva 4, and Geneprot Inc., 2 Pre´-de-la-Fontaine, 1217 Meyrin/GE, Switzerland
paolo.botti@medecine.unige.ch
Received March 31, 2004
A new and convenient method for the synthesis and incorporation of NR-(1-phenyl-2-mercaptoethyl)-
derivatized amino acids applicable to chemical ligation at non-cysteine sites is presented.
NR-Auxiliary derivatives of glycine and alanine were easily prepared using reductive amination
approaches. Several strategies for the incorporation of these derivatives into peptide chains were
investigated: coupling without protection, with acid-labile protection, with base-labile protection,
and via a novel protection strategy using the thiazolidine derivative. All amino acid derivatives
were successfully coupled to various peptide resins, and with the exception of those incorporating
Boc-protected derivatives, all resins yielded the desired peptide fragments. However, the coupling
of the two alanine derivative diastereomers generated some epimerization. Finally, N-terminal
auxiliary glycine and alanine peptides were cyclized, and the corresponding native circular peptides
were obtained upon successful removal of the auxiliary.
Introduction
This approach exploits a chemoselective reaction be-
tween two unprotected fragments, a C-terminal thioester
and an N-terminal cysteine, in aqueous solution at
neutral pH. In the first step of the reaction transthio-
esterification occurs at pH 6-8 by thiol exchange between
the free thiol of the N-terminal cysteine and the thioester
moiety on the other molecule. The newly generated
thioester then undergoes an S to N acyl shift due to the
proximity of the amino group to the thioester functional-
ity, thus generating a native amide bond at the ligation
site. By the same principle, NCL has been extended to
the generation of cyclic peptides via an intramolecular
reaction.9
A drawback with standard NCL is the intrinsic restric-
tion imposed by the requirement for cysteine at the site
of peptide ligation/cyclization. In proteins, as well as in
many cyclic peptides, cysteine occurs with a relatively
low frequency; many protein sequences do not have
suitably disposed cysteines for the NCL strategy, and
some lack cysteine residues entirely. Similarly, many
cyclic peptides of high biological relevance do not contain
cysteine residues. Although several methodologies for the
The ability to ligate or cyclize peptide fragments at
non-cysteine sites has become critical in the postgenomics
era1 and in the current efforts to develop enhanced drug
candidates.2 A rapid and efficient way to access proteins
discovered by the modern paradigms of pharmaceutical
research, notably proteome analyses, would represent a
key component in the advance toward accelerated dis-
covery of potential new therapeutic agents and drug
targets.3 Similarly, medical research would benefit from
new methodology facilitating the production of cyclic
peptides, which are becoming increasingly important as
new antimicrobial agents,4 in cancer research5 and in
various fields of medicinal chemistry.6 The past decade
has provided extensive demonstration of the utility of
native chemical ligation (NCL)7 for the rapid preparation
of medium-sized proteins with a high level of homogene-
ity.8
* To whom correspondence should be addressed at the University
of Geneva. Phone: (+41) 22 379 55 23. Fax: (+41) 22 379 55 02.
† University of Geneva.
‡ Geneprot Inc.
(1) Wilken, J.; Kent S. B. H. Curr. Opin. Biotechnol. 1998, 9, 412-
426.
(6) (a) Leduc, A. M.; Trent, J. O.; Wittliff, J. L.; Bramlett, K. S.;
Briggs, S. L.; Chirgadze, N. Y.; Wang, Y.; Burris, T. P.; Spatola, A. F.
Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 11273-11278. (b) Chen, S.
H.; Rodriguez, M. Drugs Future 2003, 28, 441-463. (c) Friedler, A.;
Zakai, N.; Karni, O.; Broder, Y. C.; Baraz, L.; Kotler, M.; Loyter, A.;
Gilon, C. Biochemistry 1998, 37, 5616-5622. (d) Chorev M.; Rubini
E.; McKee R. L.; Gibbons S. W.; Goldman M. E.; Caufield M. P.;
Rosemblatt M. Biochemistry 1991, 30, 5968-5974.
(2) Hruby, V. J. Life Sci. 1982, 31, 189-199.
(3) Rose, K.; Bougueleret, L.; Baussant, T.; Boehm, G.; Botti, P.;
Colinge, J.; Cusin, I.; Gaertner, H.; Gleizes, A.; Heller, M.; Jimenez,
S.; Johnson, A.; Kussmann, M.; Menin, L.; Menzel, C.; Ranno, F.;
Rodruiguez-Tome´, P.; Rogers, J.; Saudrais, C.; Villain, M.; Wetmore,
D.; Bairoch, A.; Hochstrasser, D. Proteomics 2004, 4, 2125-2150.
(4) (a) Mangoni, M. L.; Papo, N.; Mignona, G.; Andrei, D.; Shai, Y.;
Barra, D.; Simmaco, M. Biochemistry 2003, 42, 14023-14035. (b) Lee,
D. L.; Hodges, R. S. Biopolymers 2003, 71, 28-48. (c) Tam, J. P.; Lu,
Y. A.; Yang, J. L.; Yu, Q. Development of Novel Antimicrobial Agents:
Emerging Strategies; Horizon Scientific Press: Wymondham, UK,
2001; pp 215-240.
(7) (a) Wieland, T.; Bokelmann, E.; Bauer, L.; Lang, H. U.; Lau, H.
Liebigs Ann. Chem. 1953, 583, 129-149. (b) Dawson; P. E.; Muir, T.
W.; Clark-Lewis, I.; Kent, S. B. H. Science 1994, 266, 776-779.
(8) Dawson P. E.; Kent S. B. H. Annu. Rev. Biochem. 2000, 69, 923-
960.
(5) Jimeno, J. M. Anti-Cancer Drugs 2002, 13, S15-S19.
(9) Zhang, L.; Tam, J. P. J. Am. Chem. Soc. 1997, 119, 2363-2370.
10.1021/jo049471k CCC: $27.50 © 2004 American Chemical Society
Published on Web 11/13/2004
9208
J. Org. Chem. 2004, 69, 9208-9214

New Building Blocks for Peptide Ligation/Cyclization
SCHEME 1. Nr-(1-Phenyl-2-mercaptoethyl) Amino
cyclization of peptides both on the solid phase and in
solution have been described,10 a technique enabling both
NCL and cyclization of unprotected peptides in aqueous
solution at an X-X site (where X ) any amino acid)
would greatly facilitate the chemical synthesis of new
protein targets and would represent a valuable tool for
both the discovery and the large-scale synthesis of cyclic
peptides of potential therapeutic interest.
Acid Derivatives for ECLa
In this context, we have recently presented a novel and
practical method that extends NCL to the joining of
peptide segments at the X-Gly site,11 by making use of a
removable NR-(1-phenyl-2-mercaptoethyl) auxiliary de-
signed to closely mimic a native cysteine during ligation.
ECL (extended chemical ligation) strategy is based on
the same mechanistic pathway as NCL, generating an
amide bond at the ligation/cyclization site. Preparation
of the NR-(1-phenyl-2-mercaptoethyl) peptide fragments
for ECL is achieved by the synthesis of the appropriate
benzylamine precursor and its subsequent reaction with
the fully protected N-terminal bromoacetylated peptide.
Final acid treatment only removes the auxiliary when
on the amide bond, thus allowing unambiguous confir-
mation that the product is the desired material.12
This strategy presents a few limitations, however.
Although it has been successfully applied to generate a
native glycine residue at the ligation/cyclization site,12,13
it would not be appropriate for the unambiguous prepa-
ration of NR-(1-phenyl-2-mercaptoethyl) peptide frag-
ments featuring a substituted residue at the N-terminus.
Indeed, the reaction on the solid support of the benzyl-
amine derivative of choice with a secondary 2-bromo-
acetyl residue precursor of the desired N-terminal amino
acid (for example, the 2-bromopropionyl template is a
precursor for alanine) could not reach completeness
depending on the residue and the size of the peptide, and
may generate side products via an elimination mecha-
nism.14 Furthermore, when on a thioester resin, an
almost stoichiometric ratio between the benzylamine
precursor and the bromine is required to minimize
amidation of the thioester moiety.13
On the basis of these observations and on our previous
results, we set out to design a general method for the
synthesis of NR-(1-phenyl-2-mercaptoethyl) amino acid
derivatives (NR(Aux)-AA-OH)15 to enable the direct in-
corporation of the appropriate auxiliary-derivatized amino
acid into either CR-carboxyester or CR-thioester peptide
fragments. In this paper we present the chemical syn-
theses of these derivatives, comparing several NR protec-
tion approaches, including use of acid- or base-labile
protecting groups, direct coupling with unprotected de-
rivatives, and thiazolidine formation as a novel protecting
moiety for the 1,2-aminothiol of the auxiliary. We also
present a preliminary evaluation of the potential of these
novel building blocks with direct use in SPPS, and the
a P and P′ ) peptides. R1 ) amino acid side chain, R2 ) thiol
protecting group, R3 ) H or N protecting group, R4 ) alkyl chain.
syntheses of model circular peptides with glycine and
alanine at cyclization sites.
Results and Discussion
To exploit the concept of preformed NR-auxiliary-
derivatized amino acids ready for coupling to a peptide
resin by standard methods (Scheme 1), we prepared
auxiliary-derivatized glycine and alanine. We chose ala-
nine as the most suitable substituted candidate, because
it is statistically one of the most abundant residues in
natural proteins and it bears the smallest side chain
among the substituted amino acids.
Apart from enabling the preparation of derivatives
carrying both the amino acid and the auxiliary moiety
necessary for ligation, our synthetic approach presents
a further advantage. On the basis of a reductive amina-
tion key step, it involves the conjugation of an optically
pure amino acid (alanine) with the auxiliary moiety,
generating a new stereogenic center. This gives access
to two resulting diastereomers, which can be separated,
thus allowing a comparative study of their reactivity
during the ligation/cyclization reactions.
Synthesis of Nr-Auxiliary-Derivatized Glycine
and Alanine. The derivatized amino acids NR(Aux)-Gly-
OH and NR(Aux)-Ala-OH were obtained by the concise,
facile, and general strategy described in Scheme 2. The
synthesis of the NR(Aux)-Gly-OH derivative 4 can be
accomplished either via its benzylamine precursor 3 or
via the general procedure required for the side-chain-
functionalized amino acids (Scheme 2). The benzylamine
precursor 3 was synthesized from 1-(4-methoxyphenyl)-
2-(4-methylbenzylsulfanyl)ethanone (1) through the oxime
intermediate 2 using a preparation slightly modified from
that previously reported.11 Next, reductive amination
with glyoxylic acid16 gave racemic NR(Aux)-Gly-OH 4 in
(10) Davies, J. S. J. Pept. Sci. 2003, 9 (8), 471-501.
(11) Botti, P.; Carrasco, M. R.; Kent, S. B. H. Tetrahedron Lett. 2001,
42, 1831-1833.
(12) Low, D. W.; Hill, M. G. M.; Carrasco, M. R.; Kent, S. B. H.;
Botti, P. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 6554-6559.
(13) Cardona, V. M. F.; Hartley O.; Botti, P. J. Pept. Res. 2003, 61,
152-157.
(14) Canne, L. E.; Bark, S. J.; Kent, S. B. H. J. Am. Chem. Soc.
1996, 118, 5891-5896.
(15) When our work was in its final revision before submission, a
related work exploiting a similar concept was published: Clive, D. L.
J.; Hisaindee, S.; Coltart, D. M. J. Org. Chem. 2003, 68, 9247-9254.
J. Org. Chem, Vol. 69, No. 26, 2004 9209

Tchertchian et al.
SCHEME 2. Synthesis of Derivatized Amino Acidsa
a Reagents and conditions: (a) p-MeC6H4CH2SH, DIEA, DMF, rt, 4 h; (b) H2NOMe, pyr, EtOH, reflux, 3 h; (c) BH3‚THF, THF, 3 h
reflux; (d) glyoxylic acid, NaBH3CN, MeOH, rt, 17 h; (e) H2NCH(R1)CO2Et, BF3‚OEt, toluene, reflux, 6-8 h, then NaBH4, MeOH, 0 °C,
2 h; (f) NaOH, dioxane/H2O, rt, 17 h. R ) p-MeC6H4CH2-.
COS-Resin using the standard TBTU coupling method17
(Scheme 3). After Boc removal using two 1 min 100% TFA
cycles, standard HF treatment was performed to cleave
the peptide from the resin. The resulting peptide showed
a mass value corresponding to the N-terminal alanine
peptide without the auxiliary group. Since it is known11
that the auxiliary group can be cleaved only when on an
amide, this result can be explained by the carbamate
nature of the protected N-Boc group, suggesting that the
TFA-induced auxiliary cleavage most probably takes
place before or concomitantly with Boc removal. To find
selective Boc cleavage conditions, Boc-NR(Aux)-Gly-OH
8 was treated with different percentages of TFA (25%,
5%, and 1%) in dichloromethane, at room temperature.
While in the first two cases both the auxiliary and the
Boc group were completely cleaved in 30 min, leaving
native glycine as the reaction product, we detected some
unreacted material when the amount of TFA was reduced
to 1%. Encouraged by the latter result, we investigated
the possibility of using a super-acid-labile protecting
group, which might allow its own selective cleavage in
very mild conditions, leaving the auxiliary intact. To this
end we tried to prepare the NR-Bpoc-protected version of
(Aux)-Gly-OH as reported.18 Unfortunately, all our efforts
to prepare such a derivative completely failed, even on
the glycine template, and we decided to explore different
strategies.
SCHEME 3 N-Boc Protection and Couplinga
a Reagents and conditions: (a) Boc2O, NEt3, EtOH/CH2Cl2, rt,
48 h; (b) NaOH, dioxane/H2O, rt, 24 h; (c) H2N-GSYRFG-COS-
Resin, 9b,TBTU, DIEA, DMF, rt, 1.5 h, TFA (2 × 1 min), then
HF, p-cresol, 0 °C, 1 h. R ) p-MeC6H4CH2-, R2 ) -CH2CH2CO-
Leu-OH.
75% yield from the benzylamine precursor. The general
procedure for the preparation of the NR(Aux)-Gly-OH 4
and NR(Aux)-Ala-OH 7a,b involves a reductive amination
reaction between ketone 1 and the amino acid ethyl ester
of choice. The diastereomers 6a and 6b from the reaction
with L-alanine ethyl ester were separated by flash
chromatography to give pure compounds 6a (less polar)
and 6b (more polar). Subsequent saponification of 5, 6a,
and 6b afforded amino acids 4, 7a, and 7b in, respec-
tively, 97%, 92%, and 91% yield.
Synthesis and Side Reaction of Boc-Nr(Aux)-AA-
OH. Our first choice for NR protection of these amino
acids was to use classical Boc protection (Scheme 3). The
Boc-NR(Aux)-Gly-OH derivative 8 was obtained by reac-
tion of 5 with Boc-anhydride followed by the hydrolysis
of the ester bond, producing the desired compound in 56%
yield.
Synthesis of Nr-(1-Phenyl-2-mercaptoethyl) Pep-
tide Segments. Due to the challenges of using NR-acid-
labile-protected auxiliary-derivatized amino acids, we
decided to investigate coupling using both NR-unprotected
and NR-base-labile-protected amino acids (Aux)-AA-OH.
Coupling Using Nr-Unprotected (Aux)-AA-OH.
The coupling of unprotected derivatives 4, 7a, and 7b
(Scheme 4) involved a very judicious excess of the
incoming derivative (1.1 equiv) and is based on the
hypothesis that the activated NR-unprotected (Aux)-AA-
OH should be much more likely to couple to the unhin-
dered primary amine on the solid support rather than to
oligomerize by self-reacting with the secondary and much
Our attempts to prepare the NR-Boc-protected alanine
derivative of 9b gave relatively poor results. Indeed,
under conditions identical with those used to synthesize
Boc-NR(Aux)-Gly-OH derivative 8, 9b was obtained only
in 18% yield. The diminished reactivity of the alanyl
template was certainly due to its increased steric hin-
drance. Nevertheless, having both the NR-Boc-protected
derivatives 8 and 9b, we decided to start testing them
by coupling to a peptide resin.
In a typical experiment, Boc-NR(Aux)-Ala-OH 9b was
coupled to a model peptide resin of sequence GSYRFG-
(17) (a) Hackeng, T. M.; Griffin, J. H.; Dawson, P. E. Proc. Natl.
Acad. Sci. U.S.A. 2001, 96, 10068-10073. (b) Alewood, P.; Jones, A.;
Alewood, D.; Kent, S. B. H. Int. J. Pept. Protein Res. 1992, 40, 180-
193.
(16) De Nicola, A.; Einhorn, C.; Einhorn, J.; Luche J. L. J. Chem.
Soc., Chem. Commun. 1994, 879-880.
(18) Feinberg, R. S.; Merrifield, R. B. Tetrahedron 1972, 28, 5685.
9210 J. Org. Chem., Vol. 69, No. 26, 2004

New Building Blocks for Peptide Ligation/Cyclization
SCHEME 4. Coupling of N-Unprotected
(Aux)-AA-OHa
Thiazolidine Protection Scheme. To take advan-
tage of the 1,2-aminothiol moiety of our auxiliary-
derivatized amino acids 4, 7a, and 7b, we set out to
exploit a different protection scheme, which could pos-
sibly be easy, efficient, and general. Thiazolidine deriva-
tives have been used to protect the â-aminothiol moiety
of cysteine during syntheses of peptides.20 Recently, the
thiazolidine cyclic structure has also been successfully
exploited as an N-terminal protection scheme for cysteine
in the synthesis of a fully native folded protein of over
18 kDa.21 In multifragment ligation strategies a key issue
is the reactivity of the middle segment, which bears both
N-terminal cysteine and C-terminal thioester reacting
sites. In such a case, the N-terminal cysteine must be
protected to avoid cyclization.9 Accordingly, in a three or
more fragment ligation, the 1,2-aminothiol of the N-
terminal (Aux)-middle segment must be protected.13 In
the thiazolidine protection scheme the 1,2-aminothiol of
the N-terminal cysteine is blocked via a heterocyclic ring,
which provides a convenient temporary protection of this
residue for peptide segments with a C-terminal thioester
during ligation. In this paper we have extended the use
of the thiazolidine protection to the 1,2-aminothiol of the
NR-(2-mercaptoethyl) amino acids. In fact, the thiazoli-
dine formation with the secondary nitrogen of the NR-
auxiliary-derivatized amino acids generates a convenient
temporary tertiary amine, suitable for its use directly in
SPPS. The thiazolidine derivatives of glycine (15) and
alanine (16a and 16b) were conveniently prepared from
their precursors 4, 7a, and 7b in, respectively, 83%, 74%,
and 76% yields, by a two-step procedure including
removal of the sulfur protecting group and subsequent
condensation with formaldehyde (Scheme 6). Unfortu-
nately, NOE experiments on the two thiazolidine dia-
stereomers 16a and 16b did not allow their stereochem-
istry determination.
To test the stability of our thiazolidine derivatives
under coupling conditions, compound 15 was coupled in
solution using standard TBTU activation to glycine
methyl ester (H2NCH2COOMe), generating uniquely the
desired product. Subsequently, the thiazolidine 15 was
coupled to a model peptide on a Rink amide PEGA resin
(Scheme 6). After cleavage of the peptide from the resin
using standard Fmoc conditions (95% TFA, 2.5% TIS,
2.5% H2O), analysis of the crude showed no traces of ring
opening, thus demonstrating the stability of the thiazo-
lidine derivative in strong acidic conditions, for example,
the TFA cocktail required for the cleavages of peptides
in Fmoc-like chemistries. Future work will address a
more exhaustive study on the thiazolidine derivatives to
test also their full compatibility with the Boc chemistry.
Finally, the free 1,2-aminothiol of the N-terminal
auxiliary-derivatized glycinyl peptide was released by
a Reagents and conditions: (a) H2N-GSYRFG-COS-Resin, TBTU,
DIEA, DMF, rt, 1.5 h; (c) HF, p-cresol, 0 °C, 1 h. R2 ) -CH2CH2CO-
Leu-OH.
SCHEME 5. N-Fmoc Protection and Couplinga
a Reagents and conditions: (a) FmocCl, DIEA, dioxane/H2O, rt,
24 h; (b) H2N-GSYRFG-COS-Resin, HOBt, DCC, DMF, rt, 1.5 h,
then DBU (1% v/v), DMF, rt; (c) HF, p-cresol, 0 °C, 1 h. R )
p-MeC6H4CH2-.
more hindered amine of another incoming molecule. All
the unprotected amino acids 4, 7a, and 7b were coupled
to a model peptide resin of sequence GSYRFG-COS-Resin
in 1.1 equiv excess using standard TBTU activation for
1.5 h at room temperature. In each case, reaction
completion was monitored by ninhydrin analysis. The
crude peptide obtained after HF cleavage showed no
evidence by HPLC analysis of uncoupled amino acid
derivatives or double incorporation of the amino acid. The
peptide 11 resulting from the coupling of racemic NR(Aux)-
Gly-OH 4 was obtained as a mixture of diastereomers.
The coupling of each diastereomer NR(Aux)-Ala-OH gave
peptide diastereomers 12a and 12b, which had distinct
retention times in HPLC.
Coupling Using Nr-Base-Labile-Protected (Aux)-
AA-OH. For the base-labile protection strategy (Scheme
5) to be compatible with both our strategies would lie in
the possibility to remove the protecting group even on a
CR-thioester peptide. Recently, Clippingdale et al.19 uti-
lized DBU as an Fmoc deprotecting agent for the syn-
thesis of CR-thioester peptides via Fmoc chemistry chain
assembly. Although such a methodology is applicable only
to the preparation of small-sized peptides, its use in a
single step presents practically no side reactions.
To prepare the Fmoc-protected amino acid derivatives,
we utilized the highly reactive FmocCl. Thus, the desired
derivatives Fmoc-NR(Aux)-Gly-OH 13 and Fmoc-NR(Aux)-
Ala-OH 14b were obtained in, respectively, 63% and 36%
yield. Compound 14b was then successfully coupled to a
peptide resin CR-thioester, and the Fmoc group was
removed using 1% DBU in anhydrous DMF. No ap-
preciable amount of resin was lost during the treatment.
The desired compound 12b was finally obtained after HF
cleavage of the peptide.
(20) (a) King, F. E.; Clarl-Lewis, J. W.; Smith, G. R.; Wade, R. J.
Chem. Soc. 1959, 2264. (b) Ratner, S.; Clarke, H. T. J. Am. Chem. Soc.
1937, 59, 200. (c) Sheehan, J. C.; Yang, D.-D. H. J. Am. Chem. Soc.
1958, 80, 1158. (d) Hiskey, R. G.; Tucker, W. P. J. Am. Chem. Soc.
1962, 84, 4789 (e) Kemp, D. S.; Carey, R. I. J. Org. Chem. 1989, 54,
3640. (f) Wo¨hr, T.; Rohwedder, B.; Wahl, F.; Mutter, M. J. Am. Chem.
Soc. 1994, 118, 9218-9224.
(21) (a) Villain, M.; Vizzavona, J.; Gaertner, H. Peptides: the Wave
of the Future, Proceedings of Seventeenth American Peptide Sympo-
sium, San Diego, CA, June 9-14, 2001; American Peptide Society: San
Diego, CA, 2001; pp 107-108. (b) Villain, M.; Vizzavona, J.; Gaertner,
H. Collected Papers of the Seventh Symposium of Innovation and
Perspectives in Solid Phase Synthesis, Southampton, England, 2002;
Mayflower Worldwide, Ltd., 2002: Kingswinford, UK; pp 39-42.
(19) Clippingdale, A. B.; Barrow, C. J.; Wade, J. D. J. Pept. Sci. 2000,
225-234.
J. Org. Chem, Vol. 69, No. 26, 2004 9211

Tchertchian et al.
TABLE 1. Comparison of Cyclization Results for
Peptide Diastereomers 12a and 12b
SCHEME 6. Synthesis of Thiazolidines, Coupling,
and Resin Cleavagea
peptide
thioester
cyclic peptide/
diastereomeric
ratio,a %
hydrolysis product ratio,a %
12a
12b
20a/21a, 70/30
20b/21b, 60/40
20a, 85/15
20b, 70/30
a Ratios determined after reaction completion (17 h, 37 °C).
a Reagents and conditions: (a) HF, anisole, 0 °C, 1 h; (b)
HCHO(aq), NaOH, EtOH/H2O, rt, 1 h; (c) H2N-GAQFâA-CO2-
Resin, TBTU, DIEA, DMF, rt, 1.5 h; (d) TFA/TIS/H2O (95/2.5/2.5),
rt, 1 h; (e) MeONH2‚HCl, phosphate buffer, pH 3.5. R )
p-MeC6H4CH2-.
SCHEME 7. Cyclization of Peptide Thioestersa
FIGURE 1. HPLC analysis of cyclization products: (a) 20a
before and after a second purification of starting material 12a;
(b) 20b diastereomeric mixture.
TFA/TIS treatment to remove the auxiliary group was
performed on desalted compounds 19, 20a, and 20b.
While the final product 22 from linear precursor 11
resulted quantitatively in a single peak in HPLC analy-
sis, unfortunately crude 23 still showed the presence of
two peaks with the same expected mass for the cyclized
peptide without the auxiliary group. As reference stan-
dards we synthesized on the oxime resin22 two cyclic
peptide diastereomers of sequence AGSYRFG with al-
ternatively D- and L-alanine. The comparison with these
D- and L-alanine cyclic peptide reference standards al-
lowed the determination of epimerization ratios gener-
ated for peptides 20a and 20b, respectively 15% and 30%
(Figure 1).
These results reflect the different reactivity induced
by the initial use of the two amino acid derivatives 7a
and 7b, in both the coupling (epimerization process) and
cyclization steps. Indeed, one diastereomer was less
prompt to epimerization and favored peptide cyclization
more than the other, thus validating our synthetic efforts
to synthesize and separate amino acids 7a and 7b.
According to their different NMR spectra and distinct
HPLC retention times, amino acid derivatives 7a and 7b
show no detectable racemization of their alanine moiety.
Accordingly, the observed epimerization could not come
from the starting amino acid 7a or 7b, but presumably
occurred during either the coupling or the cyclization
step. HPLC traces on a C8 analytical column of purified
peptides 12a and 12b (as well as 11) showed in each case
a single symmetrical peak, and no trace of epimerized
material was detected, regardless of the slope of the
gradient. When however a C18 column was employed,
both peptides 12a and 12b showed a non-symmetrical
shape of the peak (shoulder). Finally, 12a was repurified,
a Reagents and conditions: (a) CH3CN/phosphate-imidazole
buffer, pH 7.4, rt, 0.5 h, or 37 °C, 0.5 or 17 h; (b) TFA/TIS (95/5),
rt, 5 or 17 h.
treatment with O-methylhydroxylamine under acidic
conditions (identical to those reported for N-terminal
cysteine derivatives) directly in the postcleavage mix-
ture.20,21
Cyclization Studies Using Nr(Aux)-AA-OH De-
rivatives. The coupling of glycine and alanine deriva-
tives 4, 7a, and 7b to a model peptide resin of sequence
GSYRFG-COS-Resin produced after HF cleavage an
HPLC trace (as crude) on a C8 analytical column with a
major single peak corresponding to the desired mass. The
obtained peptides 11, 12a, and 12b were then cyclized
in a phosphate/imidazole buffer at pH 7.4, and the
reactions were monitored by HPLC analysis. At each time
point 2-mercaptoethanol (20% v/v) was added to the
sample before injection to unambiguously distinguish the
circular unrearranged thiolactone intermediate from the
cyclized peptide with an amide bond.13 While cyclization
of peptide 11 generated only compound 19 as an expected
diastereomeric mixture after 0.5 h at room temperature,
cyclization of 12a or 12b afforded a mixture of cyclic
compound 20a or 20b and linear product 21a or 21b
resulting from the hydrolysis of the corresponding start-
ing materials (Scheme 7, Table 1).
HPLC analysis of the crude peptides showed the
presence of two distinct peaks (in different percentages,
depending on the used starting material 12a or 12b) with
the same expected mass for cyclic derivative 20a or 20b,
suggesting an epimerization process (Table 1, Figure 1).
(22) Osapay, G.; Profit, A.; Taylor, J. W. Tetrahedron Lett. 1990,
31, 6121-6124.
9212 J. Org. Chem., Vol. 69, No. 26, 2004

New Building Blocks for Peptide Ligation/Cyclization
35.5, 21.0, 14.2. HRMS (CI): m/z calcd for C21H27NO3S (M +
and the purity of the fractions was checked on HPLC with
a C18 column. A new sample from the pooled fraction
was then cyclized, producing this time only one peak with
the expected mass for the cyclized material (Figure 1),
thus demonstrating that the racemization occurred dur-
ing the coupling.
To reduce the extent of the racemization, we investi-
gated the use of carbodiimide-mediated coupling (DCC)
with alternatively HOBt and HOOBt (3-hydroxy-3,4-
dihydro-4-oxo-1,2,3-benzotriazine) as activating agent.
Unfortunately, preliminary results showed no significant
reduction of epimerization. Reproducible results were
obtained when the Fmoc-alanine derivative 14b was
employed. Future work will make a systematic study of
different coupling reagents and conditions.
H) 374.1790, found 374.1796.
Starting from ketone 1 (2.5 g, 8.74 mmol) and L-alanine
ethyl ester (1.38 g, 11.8 mmol), compound 6 was isolated as a
50/50 mixture of diastereomers (2.51 g, 74% after flash
chromatography). The diastereomers 6a,b were separated by
flash chromatography on silica gel (ether/pentane, 1/2) to give
pure 2-[1-(4-methoxyphenyl)-2-(4-methylbenzylsulfanyl)ethyl-
amino]propionic acid ethyl esters 6a (less polar) and 6b (more
1
polar). The following are data for diastereomer 6a. H NMR
(CDCl3, 400 MHz): δ 7.20 (d, 2H, J ) 8.1 Hz), 7.17 (d, 2H, J )
8.8 Hz), 7.10 (d, 2H, J ) 7.8 Hz), 6.83 (d, 2H, J ) 8.6 Hz),
4.18 (q, 2H, J ) 7.0 Hz), 3.79 (s, 3H), 3.70 (s, 2H), 3.67 (dd,
1H, J ) 4.0, 9.6 Hz), 3.10 (q, 1H, J ) 7.1 Hz), 2.61 (dd, 1H,
J ) 4.1, 13.7 Hz), 2.47 (dd, 1H, J ) 9.8, 13.4 Hz), 2.31 (s, 3H),
1.28 (d, 3H, J ) 7.0 Hz), 1.25 (t, 3H, J ) 6.9 Hz). 13C NMR
(CDCl3, 100 MHz): δ 176.0, 159.0, 136.6, 134.8, 134.3, 129.1,
128.9, 128.3, 113.9, 60.5, 58.1, 55.2, 53.6, 39.3, 34.9, 21.1, 19.7,
14.3. HRMS (CI): m/z calcd for C22H30NO3S (M + H) 388.1947,
found 388.1948. The following are data for diastereomer 6b.
1H NMR (CDCl3, 400 MHz): δ 7.20-7.15 (m, 4H), 7.09 (d, 2H,
J ) 7.9 Hz), 6.83 (d, 2H, J ) 8.8 Hz), 4.02 (q, 2H, J ) 7.1 Hz),
3.78 (s, 3H), 3.66 (pt, 1H, J ) 6.8 Hz), 3.57 (s, 2H), 3.22 (q,
1H, J ) 6.8 Hz), 2.71-2.65 (m, 2H), 2.31 (s, 3H), 1.22 (d, 3H,
J ) 6.8 Hz), 1.18 (t, 3H, J ) 7.1 Hz). 13C NMR (CDCl3, 100
MHz): δ 175.1, 159.0, 136.6, 135.2, 134.0, 129.2, 128.8, 128.4,
113.8, 60.6, 59.4, 55.2, 54.2, 39.1, 36.4, 21.1, 18.2, 14.1. HRMS
(CI): m/z calcd for C22H30NO3S (M + H) 388.1947, found
388.1944.
Conclusion
Our work greatly expands the applicability of the
extended chemical ligation and peptide cyclization by
making available the preformed auxiliary-derivatized
alanine and glycine ready to be used in SPPS. These
residues have been easily synthesized by a general three-
step procedure that could be extended to other amino
acids, irrespective of the nature of their side chains. Such
derivatives in both their unprotected and protected forms
have been successfully coupled to various peptide resins
(CR-thioester and -carboxyester). The alanine derivatives
however have been found to generate some epimerization
during the coupling. Finally, purified peptides with
N-terminal auxiliary-derivatized glycine and alanine
have been successfully cyclized, generating, after the
cleavage of the auxiliary group, the desired cyclic pep-
tides. Furthermore, the thiazolidine approach should
permit the development of functional derivatives suitable
for both single- and multiple-step extended chemical
ligation. Additional studies on the coupling conditions as
well as further exploitation of the thiazolidine protection
scheme will be addressed in a future work.
General Procedure for Preparation of Compounds 4
and 7. To a stirred solution of amino ester 5, 6a, or 6b (1
mmol) in dioxane (2 mL) was added a 4 M NaOH solution (1.25
mL, 5 mmol). The resulting mixture was stirred for 17 h at
room temperature. Water was added and the aqueous layer
extracted with ethyl acetate. The aqueous layer was then
acidified with a 1 M HCl solution and extracted with ethyl
acetate. The combined organic layers were dried over MgSO4,
filtered, and concentrated, affording pure amino acid 4, 7a, or
7b.
Starting from amino ester 5 (1.12 g, 3 mmol), rac-[1-(4-meth-
oxyphenyl)-2-(4-methylbenzylsulfanyl)ethylamino]acetic acid
(4) was isolated as a white powder (1 g, 97%). 1H NMR
(CD3OD, 400 MHz): δ 7.30 (d, 2H, J ) 8.6 Hz), 7.17 (d, 2H,
J ) 8.1 Hz), 7.12 (d, 2H, J ) 7.8 Hz), 7.01 (d, 2H, J ) 8.6 Hz),
4.30 (t, 1H, J ) 7.6 Hz), 3.82 (s, 3H), 3.66 (s, 2H), 3.61 (d, 1H,
J ) 16.9 Hz), 3.55 (d, 1H, J ) 16.9 Hz), 3.09 (dd, 1H, J ) 7.1,
13.9 Hz), 2.94 (dd, 1H, J ) 7.8, 13.9 Hz), 2.30 (s, 3H). 13C NMR
(CD3OD, 100 MHz): δ 168.8, 162.5, 138.2, 135.8, 131.0, 130.3,
130.1, 126.1, 115.9, 62.4, 55.9, 46.5, 36.6, 34.5, 21.1. HRMS
(CI): m/z calcd for C19H24NO3S (M + H) 346.1477, found
346.1480.
Starting from amino ester 6a (387 mg, 1 mmol), 2-[1-(4-
methoxyphenyl)-2-(4-methylbenzylsulfanyl)ethylamino]pro-
pionic acid 7a was isolated as a white powder (331 mg, 92%).
1H NMR (CD3OD, 400 MHz): δ 7.24 (d, 2H, J ) 8.6 Hz), 7.19
(d, 2H, J ) 8.1 Hz), 7.13 (d, 2H, J ) 8.0 Hz), 4.34 (t, 1H, J )
7.3 Hz), 3.82 (s, 3H), 3.71-3.63 (m, 2H), 3.50 (q, 1H, J ) 7.1
Hz), 3.06 (dd, 1H, J ) 7.3, 14.2 Hz), 2.93 (dd, 1H, J ) 8.1,
14.2 Hz), 2.31 (s, 3H), 1.43 (d, 3H, J ) 7.3 Hz). 13C NMR
(CD3OD, 100 MHz): δ 172.0, 162.5, 138.2, 135.8, 131.1, 130.3,
130.1, 126.0, 115.9, 62.1, 55.9, 55.1, 36.5, 34.9, 21.1, 16.3.
HRMS (CI): m/z calcd for C20H26NO3S (M + H) 360.1633,
found 360.1627.
Starting from amino ester 6b (387 mg, 1 mmol), 2-[1-(4-
methoxyphenyl)-2-(4-methylbenzylsulfanyl)-ethylamino]-pro-
pionic acid 7b was isolated as a white powder (327 mg, 91%).
1H NMR (CD3OD, 400 MHz): δ 7.29 (d, 2H, J ) 8.9 Hz), 7.17-
7.12 (m, 4H), 7.00 (d, 2H, J ) 8.8 Hz), 4.09 (dd, 1H, J ) 5.9,
9.1 Hz), 3.82 (s, 3H), 3.75 (q, 1H, J ) 7.3 Hz), 3.63-3.56 (m,
2H), 3.15 (dd, 1H, J ) 5.8, 13.9 Hz), 3.00 (dd, 1H, J ) 9.3,
13.9 Hz), 2.32 (s, 3H), 1.47 (d, 3H, J ) 7.3 Hz). 13C NMR
(CD3OD, 100 MHz): δ 171.7, 162.4, 138.3, 136.3, 131.0, 130.4,
Experimental Section
General Two-Step Procedure for Preparation of Com-
pounds 5 and 6. A mixture of 1-(4-methoxyphenyl)-2-(4-
methylbenzylsulfanyl)ethanone (1) (4 mmol), aminoester (1.35
equiv, 5.4 mmol), and boron trifluoride etherate (12 µL) was
refluxed in toluene (13 mL) for 6-8 h under N2 using a Dean-
Stark apparatus. The reaction mixture was then cooled to room
temperature and the solvent evaporated. The crude residue
was dissolved in methanol (24 mL). To the resulting solution
was added at 0 °C sodium borohydride (4 mmol), and the
reaction mixture was stirred at 0 °C under N2 for 1.5-2 h.
Water was added and the aqueous layer extracted with ethyl
acetate. The combined organic layers were dried over MgSO4,
filtered, and concentrated. Flash chromatography on silica gel
(ether/pentane, 1/2 and then 1/1) afforded compound 5 or 6.
Starting from ketone 1 (2.88 g, 10 mmol) and glycine ethyl
ester (1.39 g, 13.5 mmol), rac-[1-(4-methoxyphenyl)-2-(4-meth-
ylbenzylsulfanyl)ethylamino]acetic acid ethyl ester (5) was
isolated as an oil (2.38 g, 64%). 1H NMR (CDCl3, 400 MHz): δ
7.20 (d, 2H, J ) 7.9 Hz), 7.18 (d, 2H, J ) 8.5 Hz), 7.10 (d, 2H,
J ) 7.8 Hz), 6.84 (d, 2H, J ) 8.6 Hz), 4.16 (q, 2H, J ) 7.3 Hz),
3.78 (s, 3H), 3.74-3.68 (m, 3H), 3.28 (d, 1H, J ) 17.4 Hz),
3.10 (d, 1H, J ) 17.4 Hz), 2.66 (dd, 1H, J ) 4.5, 13.6 Hz), 2.56
(dd, 1H, J ) 9.3, 13.6 Hz), 2.31 (s, 3H), 1.24 (t, 3H, J ) 7.1
Hz). 13C NMR (CDCl3, 100 MHz): δ 172.4, 159.1, 136.6, 135.0,
133.9, 129.1, 128.8, 128.4, 113.9, 60.6, 59.7, 55.2, 48.3, 39.3,
J. Org. Chem, Vol. 69, No. 26, 2004 9213

Tchertchian et al.
130.1, 126.6, 115.7, 61.9, 55.9, 55.4, 37.2, 34.58, 21.1, 20.8, 15.1.
HRMS (CI): m/z calcd for C20H26NO3S (M + H) 360.1633,
found 360.1630.
CR-thioester (0.10 mmol) in DMF in the presence of TBTU (0.11
mmol) and DIEA (0.20 mmol). Completion of coupling was
monitored by a ninhydrin test. In the case of protected amino
acid 14b, the coupling reaction was performed in the presence
of HOBt (0.22 mmol) and DCC (0.17 mmol) and was followed
by Fmoc removal by treatment with 1% v/v DBU solution in
DMF. The peptides were then deprotected and cleaved from
the resin by treatment with anhydrous HF for 1 h at 0 °C with
500 µL of p-cresol as scavenger. After cleavage, the peptides
were precipitated with ether, dissolved in an acetonitrile/water
(50/50) mixture, and lyophilized. All peptides 10, 11, 12a, and
12b obtained from coupling with 9b, 4, 7a, 7b, and 14b were
purified by reversed-phase HPLC, affording white powders
characterized by MALDI mass spectrometry. Amino acid 15
(0.11 mmol) was coupled for 1.5 h to peptide carboxyester resin
(0.1 mmol). Peptide 17 was then cleaved from the resin by
treatment with a TFA/TIS/H2O (95/2.5/2.5) mixture, precipi-
tated with ether, dissolved in an acetonitrile/water (50/50)
mixture, and lyophilized. Crude peptide 17 (2 mg, 2.71 × 10-6
mol) was dissolved in a 0.2 M Na2HPO4, 0.2 M MeONH2‚HCl
buffer, pH 3.5 (544 µL), and the reaction was stirred at room
temperature. After 17 h, HPLC analysis of the mixture
revealed complete conversion to peptide 18. MALDI TOF MS
for peptide 10: m/z calcd for C43H63N11O12S 957.44, found
958.50. MALDI TOF MS for peptide 11: m/z calcd for
C51H71N11O13S2 1109.47, found 1111.14. MALDI TOF MS for
peptide 12a: m/z calcd for C52H73N11O13S2 1123.48, found
1125.13. MALDI TOF MS for peptide 12b: m/z calcd for
C52H73N11O13S2 1123.48, found 1125.19. Q-TOF MS for peptide
17: m/z calcd for C34H45N7O9S 727.83, found 728.33. Q-TOF
MS for peptide 18: m/z calcd for C33H45N7O9S 715.82, found
716.35.
General Two-Step Procedure for Thiazolidine Syn-
thesis. Standard HF cleavage of the thiol protecting p-
methylbenzyl group of amino acid 4, 7a, or 7b (1 mmol) in the
presence of 500 µL of anisole as scavenger affords, after
elimination of HF, a residue which is dissolved in a CH3CN/
H2O mixture (50/50). After concentration, the resulting aque-
ous phase is extracted twice with ether to remove anisole.
Concentration of the aqueous phase affords a white solid
residue which is dissolved in a 2/1 water/ethanol mixture (4
mL). A 4 M NaOH solution (500 µL, 2 mmol) was then added,
followed by HCHO (36% aqueous, 77 µL, 1 mmol). The reaction
mixture was stirred at room temperature for 1 h and concen-
trated. The aqueous phase was then acidified with a 1 M HCl
solution and extracted with ethyl acetate. The combined
organic layers were dried over MgSO4, filtered, concentrated,
and purified by flash chromatography (ethyl acetate 100%),
affording pure thiazolidine 15, 16a, or 16b.
Starting from amino acid rac-4, (830 mg, 2.41 mmol), rac-
[4-(4-methoxyphenyl)thiazolidin-3-yl]acetic acid (15) was iso-
lated as a white powder (505 mg, 83%). 1H NMR (CD3OD, 400
MHz): δ 7.39 (d, 2H, J ) 8.6 Hz), 6.92 (d, 2H, J ) 8.6 Hz),
4.33-4.27 (m, 1H), 4.28 (d, 1H, J ) 8.3 Hz), 4.05 (d, 1H, J )
8.3 Hz), 3.79 (s, 3H), 3.38 (d, 1H, J ) 17.2 Hz), 3.26 (dd, 1H,
J ) 6.6, 10.6 Hz), 3.22 (d, 1H, J ) 17.2 Hz), 3.13 (dd, 1H, J )
7.8, 10.6 Hz). 13C NMR (CD3OD, 100 MHz): δ 172.9, 161.4,
130.6, 130.4, 115.2, 72.2, 57.3, 55.8, 53.6, 35.3. HRMS (CI):
m/z calcd for C12H16NO3S (M + H) 254.0851, found 254.0846.
Starting from amino acid 7a (150 mg, 0.42 mmol), 2-[4-(4-
methoxyphenyl)thiazolidin-3-yl]propionic acid 16a was isolated
as a white powder (83 mg, 74%). 1H NMR (CD3OD, 400 MHz):
δ 7.30 (d, 2H, J ) 8.6 Hz), 6.90 (d, 2H, J ) 8.6 Hz), 4.35 (dd,
1H, J ) 6.1, 8.1 Hz), 4.20-4.16 (m, 1H), 3.78 (s, 3H), 3.35-
3.27 (m, 1H partly masked by CD3OD), 3.16 (dd, 1H, J ) 6.1,
10.4 Hz), 2.95 (dd, 1H, J ) 8.6, 10.4 Hz), 1.34 (d, 3H, J ) 7.3
Hz). 13C NMR (CD3OD, 100 MHz): δ 175.6, 161.1, 132.4, 130.2,
115.2, 69.3, 57.2, 55.7, 50.7, 38.2, 17.2. HRMS (CI): m/z calcd
for C13H18NO3S (M + H) 268.1008, found 268.1008
Acknowledgment. We thank the Swiss National
Science Foundation for financial support. We also thank
Irene Rossito-Borlat and Brigitte Dufour for technical
assistance and the NMR team of the Organic Chemistry
Department at the University of Geneva for the NMR
spectra. We thank Dr. Matteo Villain and Dr. Hubert
Gaertner for useful discussions. We finally thank Dr.
John Hill of Kent Mass Spectroscopy for his advice.
Starting from amino acid 7b (150 mg, 0.42 mmol), 2-[4-(4-
methoxyphenyl)thiazolidin-3-yl]propionic acid 16b was iso-
1
lated as a white powder (85 mg, 76%). H NMR (CD3OD, 400
Supporting Information Available: Experimental pro-
cedures for compounds 1, 2, 3, 4, 8, 13, 14b, 19, 20a, 20b,
21a, 21b, 22, and 23, NMR spectra of compounds 1, 2, 3, 4, 5,
6a, 6b, 7a, 7b, 8, 13, 14b, 15, 16a, and 16b, mass spectra of
peptides 10, 11, 12a, 12b, 17, 18, 19, 20a, 20b, 21a, 21b, and
23b, HPLC chromatograms of pure products 4, 7a, 7b, 15, 16a,
and 16b, and crude peptide 17, and cyclization reaction of
peptide 12a and crude peptide 23 (PDF). This material is
available free of charge via the Internet at http://pubs.acs.org.
MHz): δ 7.38 (d, 2H, J ) 8.6 Hz), 6.89 (d, 2H, J ) 8.8 Hz),
4.35 (t, 1H, J ) 6.1 Hz), 4.12 (d, 1H, J ) 8.8 Hz), 4.06 (d, 1H,
J ) 8.8 Hz), 3.77 (s, 3H), 3.42 (q, 1H, J ) 7.1 Hz), 3.22 (dd,
1H, J ) 6.3, 10.6 Hz), 3.07 (dd, 1H, J ) 5.8, 10.6 Hz), 1.27 (d,
3H, J ) 7.1 Hz). 13C NMR (CD3OD, 100 MHz): δ 176.7, 161.0,
132.9, 129.9, 115.0, 70.2, 59.1, 55.7, 51.8, 36.2, 13.3.
General Procedure for the Coupling of Derivatized
Amino Acids to Peptide Cr-Thioester or -Carboxyester
Resins and Subsequent HF Cleavage. Amino acid 4, 7a,
7b, 9b, or 14b (0.11 mmol) was coupled for 1.5 h to peptide
JO049471K
9214 J. Org. Chem., Vol. 69, No. 26, 2004