Solid-phase synthesis of a cyclodepsipeptide: Cotransin

Article abstract of DOI:10.1021/ol800855p

(Chemical Equation Presented) The first solid-phase synthesis of cotransin-a cyclic depsipeptide having high pharmacological potential-was achieved, by a proper choice of coupling reagents and use of either TBAF or DBU for Fmoc removal to suppress the otherwise dominating, sequence-derived diketopiperazine formation. Starting the assembly from C-terminal lactic acid allowed fast and epimerization-free cyclization in solution. Novel conditions for orthogonal use of the Fmoc/Bsmoc-protection system were discovered, and an unexpected nucleophilic behavior of DBU was observed.

Full text of DOI:10.1021/ol800855p

ORGANIC
LETTERS
2008
Vol. 10, No. 17
3857-3860
Solid-Phase Synthesis of a
Cyclodepsipeptide: Cotransin
Irene Coin,*,† Monika Beerbaum,‡ Peter Schmieder,‡ Michael Bienert,† and
Michael Beyermann†
Leibniz-Institute of Molecular Pharmacology (FMP), Robert-Ro¨ssle-Str.10, D-12135
Berlin, Germany
coin@fmp-berlin.de
Received April 14, 2008
ABSTRACT
The first solid-phase synthesis of cotransinsa cyclic depsipeptide having high pharmacological potentialswas achieved, by a proper choice
of coupling reagents and use of either TBAF or DBU for Fmoc removal to suppress the otherwise dominating, sequence-derived diketopiperazine
formation. Starting the assembly from C-terminal lactic acid allowed fast and epimerization-free cyclization in solution. Novel conditions for
orthogonal use of the Fmoc/Bsmoc-protection system were discovered, and an unexpected nucleophilic behavior of DBU was observed.
Cotransin is a cyclic heptadepsipeptide structurally derived
from the fungal product HUN-7293.1 Like HUN-7293,
cotransin is able to repress, selectively and reversibly, the
expression of a subset of secreted and membrane proteins
in mammalian cells, being particularly efficient in inhibiting
cell adhesion molecule expression.2 Due to the crucial role
played by membrane proteins and adhesion molecules in the
context of a variety of diseases, cotransin and analogous
peptides have a high therapeutic potential. Up to now, the
synthesis of these compounds has been carried out in
solution, by applying the strategy first developed by Boger
and co-workers,3 which requires the separate assembly of
two segments, their condensation, and subsequent macro-
cyclization at the MeLeu3-Leu4 position. Although even a
small library of analogues has been prepared by using this
laborious procedure,3b a synthesis method based on a solid-
phase strategy would provide a more efficient way to prepare
much more analogues in a short time, thus allowing for
extended structure-function relationship studies. To syn-
thesize cotransin via SPPS is not an easy task: the peptide
contains three N-methylamino acids, which are difficult to
be acylated, and a lactic acid residue, which requires the
formation of an ester bond. In addition, the conformational
effects induced by the presence of N-alkylated residues favor
† Department of Peptide Chemistry.
‡ Department of NMR-Supported Structural Biology.
(1) (a) Hommel, U.; Weber, H. P.; Oberer, L.; Naegeli, H. U.;
Oberhauser, B.; Foster, C. A. Febs Lett. 1996, 379, 69. (b) Itazaki, H.;
Fujiwara, T.; Sato, A.; Kawamura, Y.; Matsumoto, K. Chem. Abstr. 1995,
123, 110270.
(3) (a) Boger, D. L.; Keim, H.; Oberhauser, B.; Schreiner, E. P.; Foster,
C. A. J. Am. Chem. Soc. 1999, 121, 6197. (b) Chen, Y.; Bilban, M.; Foster,
C. A.; Boger, D. L. J. Am. Chem. Soc. 2002, 124, 5431.
(2) (a) Besemer, J.; Harant, H.; Wang, S.; Oberhauser, B.; Marquardt,
K.; Foster, C. A.; Schreiner, E. P.; de Vries, J. E.; Dascher-Nadel, C.;
Lindley, I. J. D. Nature 2005, 436, 290. (b) Garrison, J. L.; Kunkel, E. J.;
Hegde, R. S.; Taunton, J. Nature 2005, 436, 285.
(4) (a) Gisin, B. F.; Merrifield, R. B. J. Am. Chem. Soc. 1972, 94, 3102.
(b) Khosla, M. C.; Bumpus, F. M.; Smeby, R. R. J. Am. Chem. Soc. 1972,
94, 4721. (c) Pedroso, E.; Grandas, A.; Delasheras, X.; Eritja, R.; Giralt,
E. Tetrahedron Lett. 1986, 27, 743.
10.1021/ol800855p CCC: $40.75
Published on Web 07/24/2008
2008 American Chemical Society

the formation of diketopiperazine (DKP) during removal of
the temporary N-protection.4 In particular, the Leu6-MeAla7
sequence following N-terminally the ester bond gives an
extraordinary extent of DKP formation during deblocking
of Leu6, thus making cotransin particularly suitable for
searching for appropriate conditions to suppress DKP forma-
tion during stepwise synthesis of depsipeptides in general.5
With the prospect of having a method in hand for preparing
a lot of analogues, we have aimed for using the convenient
Fmoc chemistry instead of the more harmful Boc chemistry,
although the latter can be advantageous with respect to
suppression of DKP formation.6 For depsipeptides bearing
a single ester bond, DKP formation could in principle be
avoided if the ring is disconnected at the N-terminal position
following the ester bond.7 In the case of cotransin, such a
strategy would be not appropriate because in this case the
cyclization should be performed at an N-methylated residue
(Leu6 to MeAla7), thus giving reason to a slow reaction and
occurrence of epimerization.
We describe here a simple strategy to synthesize cotransin
using Fmoc chemistry on solid phase, which relies on the
suppression of DKP formation by the use of an appropriate
base for Fmoc removal and on the choice of a disconnection
position that enables fast and epimerization-free ring closure
in solution.
The stepwise assembly of the linear peptide starts with
linkage of nonprotected (-)-lactic acid onto ClTrt-Cl poly-
styrene resin in the presence of DIEA (see Scheme 1). The
activated and is therefore expected to provide ideal conditions
for the final cyclization.8 Next, the R-hydroxyl group is
esterified with Fmoc-MeAla-OH using carbodiimide (DIC)
activation in the presence of N-methylimidazole (NMI) as
catalyst.9 After standard Fmoc removal with piperidine (20%
in DMF, 2 × 5 min), quantitative coupling of Fmoc-Leu-
OH onto the N-methyl alanine residue was achieved using
HATU activation, which gave here the best result in
comparison to PyBOP activation and acid fluoride coupling.10
By applying the standard conditions for Fmoc removal to
deprotect Fmoc-Leu6-MeAla7-(-)-lac-P (P: polymer sup-
port), rapid DKP formation took place, leading to complete
loss of the Leu-MeAla-DKP from the lactic acid residue.
Even by reducing the time of piperidine treatment to the point
of incomplete deblocking, we still observed almost quantita-
tive DKP formation. It has been reported5b that DKP can be
efficiently suppressed by using instead of the Fmoc group
the more base-labile Bsmoc group,11 Bsmoc being depro-
tected under less basic conditions than Fmoc. Nevertheless,
no recovery of the target peptide was achieved even after
fast removal of the Bsmoc group from Bsmoc-Leu6-MeAla7-
(-)-lac-P with diluted piperidine (2% in DMF, 3 × 1 min).
Deprotection by piperidine seems to be slower than the
DKP formation, so that all of the target peptide is
consumed. Thus, a more rapid deprotection should be
advantageous for diminishing DKP formation. DBU is a
strong base (pKa ∼ 12), which is able to remove Fmoc
very quickly.12 Indeed, a short-time DBU treatment led
to a complete Fmoc removal from Fmoc-Leu6-MeAla7-
(-)-lac-P with a reduced extent of DKP formation. The
best result was obtained by performing a kind of “flash”
treatment, in which the peptide-resin was simply washed
once with a solution of DBU 10% in DMF (10 s), rapidly
neutralized with a HOBt solution (20 s), and immediately
acylated with the next amino acid (preactivated Fmoc-
MePhe-OH). By this, for the deblocking-coupling step
resulting in Fmoc-MePhe-Leu-MeAla-(-)-lac-P, a 30%
yield was obtained, according to the resin loading
estimated by the amount of dibenzofulvene (dbf) formed
during the next deblocking step.
Scheme 1.
Scheme for the Synthesis of Cotransina
(8) Benoiton, N. L. Biopolymers 1996, 40, 245.
aSteps: (a) H-(-)-lac-OH/DIEA/DCM, (b) esterification via DIC/NMI,
(c) Fmoc removal and coupling via HATU, (d) (i) TBAF 0.15 M in DMF,
2 × 1 min, (ii) MeOH, 15 s, (iii) DCM, 15 s, (iV) Fmoc-MePhe-OH/HBTU/
DIEA preactivated, DMF, 20 min; (d′) (i) DBU 10% in DMF, 10 s, (ii)
HOBt 0.2 M in DMF, 20 s, (iii) Fmoc-MePhe-OH/HBTU/DIEA preacti-
vated, DMF, 20 min, (e) Fmoc removal and coupling via HBTU, (f) acidic
deprotection, (g) via HATU/DIEA in DCM (0.5 mM). More details are
reported in the Supporting Information.
(9) For the use of NMI (N-methyl-imidazole) as acylation catalyst, see:
Connors, K. A.; Pandit, N. K. Anal. Chem. 1978, 50, 1542. O-Acylations
are efficiently performed also using DMAP [4-(dimethylamino)pyridine]:
Hofle, G.; Steglich, W.; Vorbruggen, H. Angew. Chem., Int. Ed. 1978, 17,
569.
(10) HATU (1-[bis(dimethylamino)methylene]-1H-1,2,3-triazolo-[4,5-
b]pyridinium hexafluorophosphate 3-oxide): Carpino, L. A. J. Am. Chem.
Soc. 1993, 115, 4397. PyBOP [(Benzotriazol-1-yloxy)tripyrrolidinophos-
phonium hexafluorophosphate]: Coste, J.; Le-Nguyen, D.; Castro, B.
Tetrahedron Lett. 1990, 31, 205. For use of amino acid fluorides see:
Wenschuh, H.; Beyermann, M.; Krause, E.; Brudel, M.; Winter, R.;
Schu¨mann, M.; Carpino, L. A.; Bienert, M. J. Org. Chem. 1994, 59, 3275.
(11) (a) Carpino, L. A.; Philbin, M.; Ismail, M.; Truran, G. A.; Mansour,
E. M. E.; Iguchi, S.; Ionescu, D.; El Faham, A.; Riemer, C.; Warrass, R.;
Weiss, M. S. J. Am. Chem. Soc. 1997, 119, 9915. (b) Carpino, L. A.; Ismail,
M.; Truran, G. A.; Mansour, E. M. E.; Iguchi, S.; Ionescu, D.; El Faham,
A.; Riemer, C.; Warrass, R. J. Org. Chem. 1999, 64, 4324.
choice of this residue as C-terminal is based on the
consideration that lactic acid cannot form oxazolone when
(5) (a) Hamada, Y.; Shioiri, T. Chem. ReV. 2005, 105, 4441. (b) Coin,
I.; Dolling, R.; Krause, E.; Bienert, M.; Beyermann, M.; Sferdean, C. D.;
Carpino, L. A. J. Org. Chem. 2006, 71, 6171.
(12) DBU: Diaza(1,3)bicyclo[5.4.0]undecane: (a) Chang, C. D.; Waki,
M.; Ahmad, M.; Meienhofer, J.; Lundell, E. O.; Haug, J. D. Int. J. Pept.
Protein Res. 1980, 15, 59. (b) Wade, J.; Bedford, J.; Sheppard, R.; Tregear,
G. Pept. Res. 1991, 4, 194.
(6) Gu, W. X.; Silverman, R. B. J. Org. Chem. 2003, 68, 8774.
(7) Lopez-Macia, A.; Jimenez, J. C.; Royo, M.; Giralt, E.; Albericio, F.
Tetrahedron Lett. 2000, 41, 9765.
3858
Org. Lett., Vol. 10, No. 17, 2008

An alternative route for reducing DKP formation was
found to be the use of TBAF.13 By deblocking Fmoc-Leu-
MeAla-(-)-lac-P with TBAF (0.15 M in DMF, 2 min),
although Fmoc removal was slower than in the case of DBU,
DKP formation was even more reduced (∼ 50%). Interest-
ingly, after the TBAF treatment and DMF washings, the
coupling of the next activated amino acid (Fmoc-MePhe-
OH) was incomplete. Significant amounts (∼ 40%) of the
nonacylated peptide H-Leu-MeAla-(-)-lac-P were found, as
determined by LC-MS of the crude cleaved from the peptide
resin. By performing a washing with methanol (15 s) after
deblocking,13 followed by reswelling of the peptide-resin in
DCM (15 s), coupling of the next amino acid occurred
quantitatively, the yield of the deblocking-coupling step
being ∼ 45% (determined by quantification of liberated dbf
during the following deblocking step).
Thus, by using TBAF for Fmoc removal at this sensitive
position and HATU activation for couplings onto N-methyl
amino acid residues, stepwise assembly of the linear cotransin
was smoothly accomplished, as described in Scheme 1.
Crude linear cotransin was obtained in very good quality
and showed a single peak in LC-MS analysis. As the sole
impurity, the lactic acid as the product of DKP formation
was coeluted with the injection peak. Lactic acid was easily
removed by a preparative HPLC run (in principle, it could
also be simply washed out), and linear cotransin was obtained
in a high yield of 34% (based on the loading at the level of
Fmoc-MeAla-(-)-lac-P).
As expected, peptide cyclization by activating the C-terminal
lactic acid proceeded smoothly using HATU/DIEA in DCM
and was complete within less than 30 min.14 The LC-MS
analysis of the cyclic product showed a single peak, having
the mass value (m/z) calculated for cotransin, with the only
impurity observed being the cyclodimer of cotransin, the content
of which was less than 5% when the cyclization was carried
out at low concentration (0.5 mM). NMR data (see Supporting
Information) of the product isolated via preparative HPLC
assessed its identity and did not show the presence of epimers.
Cotransin was obtained in a total yield of 23% based on the
loading of Fmoc-MeAla-(-)-lac-P and exhibited in a prelimi-
nary biological test the expected activity.15
The Fmoc and Bsmoc groups have quite different sensitiv-
ity toward bases, the first being deprotected through a
ꢀ-elimination reaction promoted by abstraction of the acidic
proton at the 9-position of the fluorene ring17 and the second
through a Michael-type addition process.11 First of all, we
tested the stability of the two groups toward different bases
possibly compatible with the presence of an activated species
(Table 1, SI): the acylation catalysts NMI and DMAP, TBAF,
and the tertiary amine DBU. Both amino protecting groups
were stable in the presence of NMI, whereas DMAP
(0.2-0.4 M) provided no sufficient selectivity in the removal.
At the concentrations tested, good selectivity was found for
TBAF18 (0.1-0.2 M) and diluted DBU (0.2-0.5%, corre-
sponding to 0.013-0.033 M) in DMF, with Fmoc being
easily removed and Bsmoc being substantially stable (see
Table 1, SI).
On the basis of these results, the use of TBAF and DBU
in combination with a NR-Bsmoc-protected activated amino
acid seemed to be conceivable for a one-pot approach.
However, we observed that TBAF causes fast hydrolysis of
activated species (amino acid fluorides, OPfp, OSu, OpNp
esters) due to the unavoidable presence of water in the
system,19 so that its use in this context had to be discarded.
By adding DBU (0.013-0.033 M) to Fmoc-peptides (0.04
M) in the presence of an activated NR-Bsmoc-protected
amino acid (0.2 M), we observed no Fmoc removal.
Surprisingly, this was found to be due to acylation of DBU
by the activated species (Scheme 2). The reaction is so fast
that the base is acylated before the abstraction of the
fluorenylic proton from Fmoc can occur. In fact, even by
treating an activated Fmoc-amino acid with up to 1 equiv of
DBU, almost no Fmoc removal was observed, but only
acylation of DBU took place.
Although DBU is traditionally considered a non-nucleo-
philic base,20 several examples of its nucleophilic properties
are reported in recent literature.21 Our finding about the rapid
acylation of DBU gives a rationale to a recent report of Han
and co-workers,22 who studied the catalytic activity of
bicyclic amidines in acylation reactions and detected no
catalysis operated by DBU. The acyl-imidinium formed by
DBU is in fact a low reactive species: it was stable several
Although the synthesis of cotransin following our protocol
enables the synthesis of quite pure crude cyclodepsipeptide,
the effective exploitation of the resin capacity is still limited
by the extent of DKP formation. An approach to further
reduce DKP formation using the Fmoc chemistry may result
from developing a simultaneous deprotection-coupling
procedure (tandem deprotection/coupling),16 by which the
removal of the Fmoc group from the N-terminus is carried
out in the presence of an activated amino acid, orthogonally
protected at the NR, which acylates in situ the free amino
terminus as soon as liberated.
(16) The first example of coupling in situ during deprotection of Z is
reported by: Shute, R. E.; Rich, D. H. J. Chem. Soc., Chem. Commun. 1987,
1155. Use of Alloc for tandem deprotection/coupling reaction: Thieriet, N.;
Alsina, J.; Giralt, E.; Guibe´, F.; Albericio, F. Tetrahedron Lett. 1997, 38,
7275. Thieriet, N.; Gomez-Martinez, P.; Guibe´, F. Tetrahedron Lett. 1999,
40, 2505. Zorn, C.; Gnad, F.; Salmen, S.; Herpin, T.; Reiser, O. Tetrahedron
Lett. 2001, 42, 7049.
(17) Carpino, L. A.; Han, G. Y. J. Org. Chem. 1972, 37, 3404.
(18) Fresh prepared solutions of TBAF·3H2O in DMF dried over
molecular sieves were used. Bsmoc was stable toward TBAF only when
using dried DMF, whereas by using commercial DMF as such unidentified
products of degradation of the Bsmoc protected substrates were formed.
(19) No improvement resulted from attempts to remove water from the
system, neither by prolonged storage of the TBAF solution on molecular
sieves nor by working in the presence of N,O-bis(trimethylsilyl)acetamide,
nor by using other fluoride sources like TASF [tris(dimethylamino)sulfonium
difluorotrimethylsilicate] or CsF (the latter used alone or in the presence of
18-crown-6 ether).
(13) TBAF: tetrabutylammonium fluoride: Ueki, M.; Amemiya, M.
Tetrahedron Lett. 1987, 28, 6617.
(20) Oediger, H.; Mo¨ller, F.; Eiter, K. Synthesis 1972, 591.
(14) Ehrlich, A.; Heyne, H. U.; Winter, R.; Beyermann, M.; Haber, H.;
Carpino, L. A.; Bienert, M. J. Org. Chem. 1996, 61, 8831.
(15) Cotransin-mediated decrease in Corticotropin Releasing Factor
Receptor expression was investigated in the research group of Ralf Schu¨lein
at the FMP, Berlin. Further experiments are in progress.
(21) Nucleophilic catalysis of DBU in Baylis-Hillman reactions:
Aggarwal, V. K.; Mereu, A. Chem. Commun. 1999, 2311. Aza-Michael
addition promoted by DBU: Yeom, C.-E.; Kim, M. J.; Kim, B. M.
Tetrahedron 2007, 63, 904. See also: Ghosh, N. Synlett 2004, 574.
(22) Birman, V. B.; Li, X.; Han, Z. Org. Lett. 2007, 9, 37.
Org. Lett., Vol. 10, No. 17, 2008
3859

case in which both groups are used as NR-protection, but
show on the other hand that bases tested are not suited to be
used for a tandem deprotection-coupling system.
Scheme 2
.
Reaction between Bsmoc-Val-Y (Y ) F, OPfp, OSu,
OpNp) and DBUa
In conclusion, we described here an efficient protocol to
synthesize the cycloheptadepsipeptide cotransin in high
purity, using Fmoc chemistry on the solid phase. By
substituting either DBU or TBAF for piperidine, DKP
formation during Fmoc removal at the second position
following the ester bond has been reduced to an extent that
makes the synthesis of cotransin achievable in good yields
using Fmoc chemistry, without the need for implementation
of other protecting groups and chemistries (like for instance
the use of the Alloc24 group).16 The choice of an appropriate
disconnection position has also been of particular importance.
In fact, when we tried to assemble cotransin using MeLeu3-
Leu4 as the disconnection point, cyclization using PyBOP
occurred slowly and gave rise to extensive epimerization (∼
20%). For this cyclization, reduced epimerization was
reported3 by the use of a low activation method (azide
activation25), but this requires a long reaction time (up to
several days) and gives mostly lower yields.26 Our results
show that activation of lactic acid at the C-terminus of linear
cotransin allows for a fast and epimerization-free cyclization,
a method that can be adopted as a guideline principle for
the cyclization of substrates containing R-hydroxy acids,27
in perfect accordance with the known advantages given by
the use of C-terminal depsipeptide units for segment
coupling.28
a
The product was isolated via HPLC and identified via LC-MS and
NMR (see also Supporting Information).
days in aqueous solutions, DMSO, and MeOH, and it did
not react with nucleophiles (H-Xaa-OR). In the presence of
pyridine, Bsmoc-Val-DBU+ in MeOH gave Bsmoc-Val-OMe
(∼ 30% yield in 3 h).23 At increased concentration of DBU,
its nucleophilic behavior becomes apparent by a Michael-
like attack at the dioxobenzothiophene system of Bsmoc, in
the same way which is described for the removal operated
by piperidine11 (Scheme 3, see also SI for discussion about
the structure of the resulting product).
Scheme 3.
Bsmoc Removal Operated by DBUa
Our protocol for the synthesis of cotransin may represent
a general strategy for the synthesis of cyclodepsipeptides.
Acknowledgment. We are indebted to the Deutsche
Forschungsgemeinschaft (Be 1434/5-2) for financial support.
Supporting Information Available: Materials and meth-
ods, experimental procedures, and NMR spectra. LC-MS
analysis of the final product. Table 1. Discussion about the
adduct formed from Bsmoc removal via DBU. This material
is available free of charge via the Internet at http://pubs.acs.org.
a
The adduct was isolated via HPLC and identified via LC-MS and
NMR.
OL800855P
Taken together, our results identify new conditions for the
selective removal of Fmoc in the presence of Bsmoc, for a
(24) Kunz, H.; Unverzagt, C. Angew. Chem., Int. Ed. 1984, 23, 436.
(25) Izumiya, N.; Kato, T.; Waki, M. Biopolymers 1981, 20, 1785.
(26) Schmidt, R.; Neubert, K. Int. J. Pept. Protein Res. 1991, 37, 502.
(27) Gisin, B. F.; Merrifield, R. B.; Tosteson, D. C. J. Am. Chem. Soc.
1969, 91, 2691.
(23) Catalytic activity of DBU is reported for cyanoacylation of ketones:
Zhang, W.; Shi, M. Org. Biomol. Chem. 2006, 4, 1671. And for esterification
of carboxylic acids with dimethyl carbonate: Shieh, W.-C.; Dell, S.; Repicˇ,
O. J. Org. Chem. 2002, 67, 2188. In the latter work, the formation of a
carbamate intermediate is demonstrated, which shows analogies with the
complex identified by us.
(28) (a) Coin, I.; Schmieder, P.; Bienert, M.; Beyermann, M. J. Pept.
Sci. 2008, 14, 299. (b) Yoshiya, T.; Sohma, Y.; Kimura, T.; Hayashia, Y.;
Kiso, Y. Tetrahedron Lett. 2006, 47, 7905.
3860
Org. Lett., Vol. 10, No. 17, 2008