Depsipeptide methodology for solid-phase peptide synthesis: Circumventing side reactions and development of an automated technique via depsidipeptide units

Article abstract of DOI:10.1021/jo060914p

The depsipeptide technique is a recently developed method for peptide synthesis which is applicable to difficult sequences when the synthetic difficulty arises because of aggregation phenomena. In the present work, application of the depsipeptide method t

Full text of DOI:10.1021/jo060914p

Depsipeptide Methodology for Solid-Phase Peptide Synthesis:
Circumventing Side Reactions and Development of an Automated
Technique via Depsidipeptide Units†,‡
Irene Coin,§ Rudolf Do¨lling,| Eberhard Krause,§ Michael Bienert,§ Michael Beyermann,*,§
Calin Dan Sferdean, and Louis A. Carpino*,
Leibniz-Institut fuer Molekulare Pharmakologie and Biosyntan GmbH, Robert-Roessle-Str. 10,
13125 Berlin, Germany, and Chemistry Department, Lederle GRC Tower, Box 34510,
UniVersity of Massachusetts, Amherst, Massachusetts 01003-4510
beyermann@fmp-berlin.de; lcarpino@chemistry.umass.edu
ReceiVed May 3, 2006
The depsipeptide technique is a recently developed method for peptide synthesis which is applicable to
difficult sequences when the synthetic difficulty arises because of aggregation phenomena. In the present
work, application of the depsipeptide method to extremely difficult sequences has been demonstrated
and a serious side reaction involving diketopiperazine formation uncovered and subsequently avoided by
the appropriate use of the Bsmoc protecting group. Many other aspects of the technique have been
investigated, such as the stability of the depsi units during assembly and workup procedures, the
completeness of the O-acylation step, the occurrence of epimerization of the amino acid activated during
O-acylation, and the nature of side products formed. In addition, the method was modified so as to allow
for completely automated syntheses of long-chain depsipeptides without the need for any interruption by
manual esterification procedures. Finally, the synthesis efficiency of the new depsipeptide technique was
shown to be comparable to that of the well-known pseudoproline technique.
Introduction
purification of the resulting products to be extremely difficult.
Current strategies to prevent difficulties in synthesis aim at
interfering with peptide folding and association by introducing
specific reversible modifications at the peptide backbone, using
N-alkylation systems (e.g., Hmb-protected amino acids1), pseu-
doprolines,2 or most recently depsipeptide units.
A main obstacle to the chemical synthesis of long peptides
and proteins is the occurrence of aggregation phenomena, which
cause both coupling and deprotection reactions during the
assembly of peptide chains to be slow and incomplete and the
Insertion of a depsipeptide unit in a sequence interrupts
the regular pattern of amide bonds at the site of a Ser/Thr
unit, with the peptide chain being extended from that point on
via the â-hydroxyl function. Each depsipeptide unit pro-
vides an additional ionizable moiety, thereby increasing solubil-
ity and facilitating purification, as reported for the case of the
â(1-42) amyloid peptide.3,4 The conversion to the target amide
† Dedicated to the memory of Professor Bruce Merrifield, a warm human
being and a great scientist.
‡ Abbreviations used: Bn, benzyl; Boc, tert-butyloxycarbonyl; Bsmoc, 1,1-
dioxobenzo[b]thiophene-2-ylmethyloxycarbonyl; DCM, dichloromethane; DIC,
N,N′-diisopropylcarbodiimide; DIEA, N,N-diisopropylethylamine; DMAP, 4-(di-
methylamino)pyridine; DMF, N,N-dimethylformamide; EDC, N-ethyl-N′-(3-
dimethylaminopropyl)carbodiimide hydrochloride; Fm, 9-fluorenemethyl; Fmoc,
9-fluorenemethyloxycarbonyl; Hmb, 2-hydroxy-4-methoxybenzyl; NEt3, triethyl-
amine; N-HATU, 1-[bis(dimethylamino)methylene]-1H-1,2,3-triazolo[4,5-b]py-
ridinium hexafluorophosphate 3-oxide; N-HBTU, 1-[bis(dimethylamino)meth-
ylene]-1H-benzotriazolium hexafluorophosphate 3-oxide; NMI, N-methylimi-
dazole; NMP, N-methylpyrrolidone; O-Su, succinimido ester; TFA, trifluoroacetic
acid; Z, benzyloxycarbonyl.
(1) Hyde, C.; Johnson, T.; Owen, D.; Quibell, M.; Sheppard, R. C. Int.
J. Pept. Prot. Res. 1994, 43, 431-440.
(2) (a) Mutter, M.; Nefzi, A.; Sato, T.; Sun, X.; Wahl, F.; Wahl, T.
Peptide Res. 1995, 8, 145-153. (b) Wo¨hr, T.; Wahl, F.; Nefzi, A.;
Rohwedder, B.; Sato, T.; Sun, X.; Mutter, M. J. Am. Chem. Soc. 1996,
118, 9218-9227.
§ Leibniz-Institut fuer Molekulare Pharmakologie.
| Biosyntan GmbH.
University of Massachusetts.
10.1021/jo060914p CCC: $33.50 © 2006 American Chemical Society
Published on Web 07/07/2006
J. Org. Chem. 2006, 71, 6171-6177
6171

Coin et al.
SCHEME 1. Conversion of a Depsipeptide to the Amide
SCHEME 2. Structure of Depsi-(VT)10-NH2 (1a)
Form through a Base-Promoted O,N-Acyl Shift
formation10 uncovered and subsequently avoided by judicious
use of a new protecting group. Many other questions that have
arisen during studies of the application of the technique were
addressed, for example, questions concerning the stability of
the depsi units during assembly and workup procedures, the
completeness of O-acylation, the occurrence of epimerization
of the amino acid activated during the O-acylation step,11 and
the eventual role played by the choice of solid support.4 In addi-
tion, the method has been modified so as to allow for completely
automated syntheses of long-chain depsipeptides without the
need for any interruption by manual esterification procedures.
The synthesis efficiency of the new depsipeptide method was
compared to that of the pseudoproline technique. For this
purpose, most of the peptides obtained via depsipeptide ana-
logues were synthesized also via pseudoproline, and the purities
of the crude products respectively obtained were compared.
Stability of Depsipeptide Units toward the Standard
Conditions of Solid-Phase Peptide Synthesis. Our first aim
was to check the stability of the depsipeptide bonds toward
repeated treatment with the standard reagents involved in solid-
phase synthesis. We took as a model the homooligopeptide
(VT)10, a difficult sequence12 forming â-structures,13 and built
a depsi bond at each Thr site, thus obtaining a poly-depsipeptide
chain made up of a series of alternate ester-amide units
(Scheme 2).
peptide is smoothly achieved through an O,N-acyl shift,5 which
occurs quantitatively under mildly basic conditions over a short
period of time (Scheme 1).
The idea of synthesizing an O-acyl isomer that could be
subsequently converted to a target amide as a means of cir-
cumventing synthetic difficulties was first described in 1998
for the assembly of some highly hindered R-methylserine
analogues of leucine-enkephalin.6 More recently, the technique
was rediscovered almost simultaneously by three research
groups,3,4a,7a and the depsipeptide method (otherwise called the
O-acyl isopeptide method4) began to be generalized for the
synthesis of difficult sequences where the difficulty arises
because of aggregation phenomena. Relative to the pseudopro-
line technique2 for solving similar problems via the same two
amino acids (Ser, Thr), the depsipeptide methodology has the
advantage of providing an isomer of the native amide which
needs only to be isomerized5 under essentially physiological
conditions to the native amide species. This methodology lends
itself to the generation of highly soluble prodrugs8,9 for
medicinal purposes and to the study of peptide folding and
association for highly aggregating natural systems (so-called
switch-peptides7 or click peptides4b).
Figure 1 reports the MALDI-MS spectra of the crude
(VT)10-NH2 (1) obtained via standard automated synthesis and
(10) (a) For an attempt to detect diketopiperazine formation during the
solid-phase syntheses of some hydroxyethylamine peptide isosteres, see:
Tamamura, H.; Kato, T.; Otaka, A.; Fujii, N. Org. Biomol. Chem. 2003, 1,
2468-2473. Tamamura, H.; Hori, T.; Otaka, A.; Fujii, N. J. Chem. Soc.,
Perkin Trans. 1 2002, 577-580. (b) For the slight occurrence of this side
reaction in the solid-phase synthesis of some Aâ(1-42) isopeptide mutants,
see: Sohma, Y.; Chiyomori, Y.; Kimura, M.; Fukao, F.; Taniguchi, A.;
Hayashi, Y.; Kimura, T.; Kiso, Y. Bioorg. Med. Chem. 2005, 13, 6167-
6174.
(11) (a) For an examination of the loss of configuration and the extent
of coupling for different activating reagents in the acylation of phenyl lactic
acid benzyl ester by Fmoc-Ala-OH, see: Davies, J. S.; How, J.; Le Breton,
M. J. Chem. Soc., Perkin Trans. 2, 1995, 2335-2339. (b) Previously,
varying amounts of configurational loss have been reported during on-resin
O-acylation of both primary and secondary hydroxyl-containing systems:
Sohma, Y.; Sasaki, M.; Hayashi, Y.; Kimura, T.; Kiso, Y. Chem. Commun.
2004, 124-125; Hamada, Y.; Ohtake, J.; Sohma, Y.; Kimura, T.; Kiso, Y.
Bioorg. Med. Chem. 2002, 10, 4155-4167. Sohma, Y.; Hayashi, Y.;
Skwarezynsky, M.; Hamada, Y.; Sasaki, M.; Kimura, T.; Kiso, Y.
Biopolymers (Pept. Sci.), 2004, 76, 344-356. (c) For early observations of
the loss of configuration observed during the loading of urethane-protected
amino acids onto hydroxy resins via the DCC or DIC/DMAP (cat.) methods,
see: Handford, B. O.; Jones, J. H.; Young, G. T.; Johnson, T. F. N. J. Chem.
Soc. 1965, 6814-6827. Jones, J. H.; Young, G. T. J. Chem. Soc. C 1968,
53-61.
Results and Discussion
In the present work, application of the depsipeptide method
to an extremely difficult sequence has been demonstrated and
a serious side reaction involving diketopiperazine (DKP)
(3) Carpino, L. A.; Krause E., Sferdean C. D., Schu¨mann, M.; Fabian,
H.; Bienert, M.; Beyermann, M. Tetrahedron Lett. 2004, 45, 7519-7523.
(4) (a) Sohma, Y.; Hayashi, Y.; Kimura, M.; Chiyomori, Y.; Taniguchi,
A.; Sasaki, M.; Kimura, T.; Kiso, Y. J. Pept. Sci. 2005, 11, 441-451. (b)
Taniguchi, A.; Sohma, Y.; Kimura, M.; Okada, T.; Ikeda, K.; Hayashi, Y.;
Kimura, T.; Hirota, S.; Matsuzaki, K.; Kiso, Y. J. Am. Chem. Soc. 2006,
128, 696-697.
(5) (a) Bergmann, M.; Brand, E.; Weinmann, F. Hoppe-Seyler’s Z.
Physiol. Chem. 1923, 131, 1-17. (b) Wakamiya, T.; Tarumi, Y.; Shiba, T.
Bull. Chem. Soc. Jpn. 1974, 47, 2686-2689.
(6) Horikawa, M.; Shigeri, Y.; Yumoto, N.; Yoshikawa, S.; Nakajima,
T.; Ohfune, Y. Bioorg. Med. Chem. Lett. 1998, 8, 2027-2032.
(7) (a) Dos Santos, S.; Chandravarkar, A.; Mandal, B.; Mimna, R.; Murat,
K.; Sauce`de, L.; Tella, P.; Tuchscherer, G.; Mutter, M. J. Am. Chem. Soc.
2005, 127, 11888-11889. (b) Mutter, M.; Chandravarkar, A.; Boyat, C.;
Lopez, J.; Dos Santos, S.; Mandal, B.; Mimna, R.; Murat, K.; Patiny, L.;
Sauce`de, L.; Tuchscherer, G. Angew. Chem., Int. Ed. 2004, 43, 4172-
4178.
(12) Gausepohl, H.; Rapp, W.; Bayer, E.; Frank, R. W. In InnoVation
and PerspectiVes in Solid-Phase Synthesis: Peptides, Polypeptides and
Oligonucleotides; Second International Symposium, Epton, R., Ed.; Intercept
Limited Pub: Andover, 1992; pp 381-385.
(13) Altmann, K. H.; Flo¨rsheimer, A.; Mutter, M. Int. J. Pept. Prot. Res.
1986, 27, 314-319.
(8) Skwarczynski, M.; Sohma, Y.; Noguchi, M.; Kimura, M.; Hayashi,
Y.; Hamada, Y.; Kimura, T.; Kiso, Y. J. Med. Chem. 2005, 48, 2655-
2666.
(9) Oliyai, R.; Stella, V. J. Bioorg. Med. Chem. Lett. 1995, 5, 2735-
2740.
6172 J. Org. Chem., Vol. 71, No. 16, 2006

Depsipeptide Methodology for Solid-Phase Peptide Synthesis
FIGURE 1. MALDI-MS spectra of crude products. (A) (VT)10-NH2 (1). Mass differences between peaks identify a series of Val and Thr deletions.
(B) depsi-(VT)10-NH2 (1a). Calculated values m/z (average mass, the same in both cases): [M + H]+calc) 2020.4; [M + Na]+calc) 2043.4; [M +
K]+calc) 2059.4. [M + H - 17]+ ) 2003.4 (neutral loss of NH3 from a side chain amino group).
calc
FIGURE 2. [Asn15]FBP28-WW-amide (3). Native Asp15 was substi-
tuted by Asn in order to avoid byproducts due to aspartimide
formation.16 Replacing the native free acid form with an amide at the
C-terminal position results in sharper peaks during HPLC analysis.
FIGURE 3. HPLC profiles of the crude Y19-K37 segment: (A) standard
Colored circles highlight sites chosen for the introduction of depsi units.
synthesis (2); (B) synthesis via the depsi analogue after shift (2a); (C)
synthesis via pseudo-proline (2b). Key: *, target product 2, in A
of the crude depsi-(VT)10-NH2 (1a), manually assembled as
described in the Experimental Section. In both cases, a TentaGel
SRam resin was used as solid support and the final cleavage
was carried out with TFA in the presence of scavengers.
The signals detected in the MALDI-MS spectrum of depsi-
(VT)10-NH2 are all attributable to the target depsipeptide and
do not give any indication for the presence of byproducts, thus
confirming that ester bonds are in general stable both toward
piperidine and under TFA cleavage conditions.
Application of the Depsipeptide Methodology to the
Extremely Difficult WW Domain FBP28. As a suitable model
to study the depsipeptide technique, we chose an appropriate
analogue of the WW domain FBP28 (Figure 2), a small, 37-
residue peptide, exhibiting a triple-stranded, antiparallel â-sheet
structure,14 which is impossible to synthesize using standard
SPPS15 (see Figure 5A).
Even the shorter C-terminal segment Y19-K37 (2) proved to
be a difficult sequence, with a standard SPPS via a TentaGel
SRam resin (amino linker) giving less than 25% of the target
peptide (Figure 3A). By introducing a depsipeptide unit at ES
27/28, the synthetic difficulties were completely overcome, and
the crude product obtained (2a) was of excellent quality
according to HPLC (Figure 3B).
coeluted with des-N; #, N23-K37 (43%); +, des-Y,Y,N; O, des-Y,N.
less than 20% of the target depsipeptide. By MALDI-MS
analysis, among the side products we found sequences arising
from apparent cleavage at the depsi bonds: a group of segments
having Glu27 as the C-terminal unit, and a product involving
truncation at Thr18.
Detection of Diketopiperazine Formation, Circumvented
by Use of Bsmoc Chemistry. To investigate the origin of
truncation at Thr18, the assembly was monitored step by step
after introduction of the second depsi unit at residues KT 17/
18, by cleaving and analyzing a small amount of sample at each
step. DKP formation was found to occur during removal of the
Fmoc group from Gly16, leading to 20% (according to HPLC
trace) of material missing the first two amino acids, product 4:
As no evidence of DKP formation was observed after standard
deprotection of the Leu26 residue (next to the depsi unit ES
27/28), the occurrence of such intramolecular aminolysis is
(14) Macias, M. J.; Gervais, V.; Civera, C.; Oschkinat, H. Nat. Struct.
Biol. 2005, 7, 375-379.
(15) Tremmel, S.; Beyermann, M.; Oschkinat, H.; Bienert, M.; Naumann,
D.; Fabian, H. Angew. Chem., Int. Ed. 2005, 44, 4631-4635.
On the other hand, synthesis with the same resin of the
complete sequence [Asn15]FBP28-NH2 via the introduction of
two depsi units (3a), at positions ES 27/28 and KT 17/18, gave
J. Org. Chem, Vol. 71, No. 16, 2006 6173

Coin et al.
FIGURE 4. HPLC profiles of crude products. To avoid artifacts in
estimating the extent of DKP formation, due to possible degradation
of the samples during workup and/or storage,18 removal of the NR-
protection from Leu26 was immediately followed by acetylation under
acidic conditions with Ac2O 50% v/v in DMF (5a, 5b). Identification
of the products was assessed by LC-MS analysis: (A) product of
standard Fmoc removal, 5a; (B) product of Bsmoc removal, 5b; (C)
product of standard coupling after Bsmoc removal, 6. Key: *, Ac-lac-
TWEKPQELK-NH2 5a; #, Ac-LE-lac-TWEKPQELK-NH2 5b; O,
Fmoc-TLE-lac-TWEKPQELK-NH2 6.
FIGURE 5. HPLC profiles of the crude [Asn15]FBP28-NH2: (A)
standard synthesis (3); (B) synthesis via two depsi units (ES 27/28,
KT 17/18) after shift, peptide 3b; (C) synthesis via two pseudoproline
units (ES 27/28, KT 17/18), peptide 3d; (D) depsi analogue containing
four depsi units (depsi-3c); (E) shifted depsi analogue (3c); (F) synthesis
via three pseudo-proline units, (ES 27/28, KT 12/13 coupled via HATU,
4 h, VS 5/6), peptide 3e. Key: *, target product 3, in A coeluted with
byproducts; #, des-Thr3; O, depsi-3c.
clearly dependent on the nature of the two amino acids adjacent
to the depsi bond (at the N-terminal side) and probably also on
their position within the sequence. To solve the problem of
truncation at Thr18 we used for R-amino protection the Bsmoc17
group, which is removed under milder conditions than is the
Fmoc residue.
Upon consideration of short DKP-prone sequences suitable
for a preliminary study, it was found that substitution of Ser28
with L-lactic acid in the C-terminal segment of the WW domain
provoked Fmoc removal from Fmoc-LE-(+)lac-TWEKPQELK
to take place with a 50% yield of product missing the first two
N-terminal amino acids (5a, Figure 4A). The amount of DKP
formation was markedly reduced by using the Bsmoc group,
which was quantitatively removed from Leu26 using 2%
piperidine v/v in DMF for 3 min (5b, Figure 4B), and coupling
of the following Thr residue could be performed via the standard
procedure, leading to a product of high purity (6, Figure 4C).
Suppression of Sequences Truncated at Glu27: Role of
the Solid Support. The formation of sequences of the type
XxxnfGlu27 was circumvented by performing an acetylation
step in the presence of NMI before introduction of the first depsi
unit. This indicates that the stronger activation process employed
for O-acylation (via DIC/NMI, 4 h) can give rise to undesired
additional acylation of free amino groups remaining available
on the resin after repeated cycles of standard couplings (N-
HBTU, 20 min), thus leading to the birth of new chains having
as the C-terminus an amino acid belonging to a depsi unit.
When assembling on a preloaded TentaGelA resin a depsi
analogue of Aâ(1-42) containing a depsi unit at Gly-Ser
25/26,4 Kiso and co-workers found large amounts of undesired
Aâ(1-25) and Aâ(26-42) (20% and 30%, respectively, ac-
cording to the HPLC trace), whereas performing the same
synthesis on a chlorotrityl resin gave no Aâ(1-25) and only a
small amount (1.6%) of Aâ(26-42). To explain these differing
results based on the use of the two resins, it was suggested that
the TFA-cocktail used for the final cleavage promotes hydrolysis
of the ester bond, which should occur only when the depsi-
peptide is still bound to the resin. To avoid such ester bond
hydrolysis, the use of an extremely acid-labile resin such as
the chlorotrityl resin appeared to be necessary. In light of our
results an alternative explanation can be proposed, according
to which the byproduct Aâ(1-25) is generated by acylation of
hydroxyl functions still available on the TGA resin when Gly25
is coupled to the hydroxyl function of Ser26 under O-acylation
conditions. This cannot occur in the case of the chlorotrityl resin,
since even if some chloride functions remained uncapped after
the treatment with MeOH which usually follows the loading
step, they are anyway unreactive toward the acylating species
that are formed during O-acylation. The other byproduct,
segment Aâ(26-42), may arise via DKP formation.
Based on these considerations, we can propose an optimized
strategy for application of the depsipeptide technique, according
to which a capping step in the presence of NMI is performed
after linking of the first amino acid to the resin, and every second
residue following a depsi unit is introduced with Bsmoc NR-
protection. Following this procedure we synthesized some depsi
analogues of FBP28-WW, differing in the number and positions
of the depsi units introduced. We obtained a crude product of
high purity by introduction of four depsi units: ES 27/28, KT
17/18, KT 12/13, and VS 5/6. In the analysis of the crude
depsipeptide product obtained by this synthesis (depsi-3c, Figure
5D) no sequences ending at Glu27 were found and no indication
of DKP formation was detected. In addition the subsequent
simultaneous O,N-shift at the four ester units was completed
smoothly in aqueous bicarbonate within 1 h (3c, Figure 5E).
To evaluate the occurrence of epimerization of the amino
acid activated during O-acylation, after shifting from the four-
depsi analogue, crude peptide 3c was hydrolyzed, and the
derivatized amino acid mixture analyzed in a GC-MS system.
The quantity of D-enantiomers found was 0.13% for Glu, 1.21%
for Val, and 1.66% for Lys, the last value being the sum of two
contributions (depsi units KT 17/18 and 12/13).
With regard to the synthesis of the other depsi analogues
containing smaller numbers of depsi units the appearance of
products due to incomplete O-acylation is notable, whereas all
ester bonds of the 4-depsi-[Asn15]FBP28-NH2 system were
smoothly formed using a standard protocol, a fact which is
indicative of an important role for conformational effects. In
particular, synthesis of the depsi peptide containing depsi units
at ES 27/28, KT 17/18, and VS 5/6 gave as the main byproduct
6174 J. Org. Chem., Vol. 71, No. 16, 2006

Depsipeptide Methodology for Solid-Phase Peptide Synthesis
SCHEME 3. Method for the Synthesis of the Ala/Thr
Depsidipeptide Building Blocka
For the first model of an automated synthesis the 31-mer
C-terminal segment (12) of the small, hydrophobic, globular
protein Crambin-4621 was chosen.
The full 46-mer has been cited as a difficult sequence by
Kent and co-workers,21 who reported an excellent total synthesis
by a unique one-pot native chemical ligation technique. Syn-
thetic difficulties were also described for the 31-mer. From a
preliminary standard synthesis of the native 31-mer the UV
deblocking trace showed difficulties at positions 27-28 (Fig-
ure 6A inset), which were completely eliminated by introduc-
tion of a depsi unit at AT 27/28 (Figure 6B inset). The
depsipeptide CRLPGTPEALC-AT-YTGCIIIPGATCPGD was
assembled using a standard automated protocol and introducing
the depsi unit via the depsidipeptide block 11. Synthesis of the
double depsi analogue CRLPGTPEALC-AT-YTGCIIIPG-AT-
CPGDYAN gave a comparable result, without showing any
further improvement.
a Key: (a) Cs2CO3/MeOH, followed by C6H5CH2Br/DMF;19 (b) Fmoc-
Ala-OH/EDC/DMAP/DCM;20 (c) H2/Pd(C)/EtOH; (d) Z-Ala-OH/EDC/
DMAP/DCM;20 (e) Fmoc-OSu/NEt3/CH3CN/H2O. The longer pathway
avoids possible difficulties due to the incomplete orthogonality between
Fmoc and OBn systems toward hydrogenolysis. The overall yields are
comparable (∼50%).
Conclusions. In conclusion, this work has demonstrated the
general stability of depsipeptides during the normal operations
of solid-phase peptide synthesis. By study of the very difficult
FBP28-WW sequence a serious side reaction due to diketopip-
erazine formation was identified and suppressed by substitution
of Bsmoc for Fmoc amino acids at appropriate positions. By
prior synthesis of a particular depsidipeptide, completely
automated syntheses of difficult long chain sequences was
possible. Finally, the synthesis efficiency of the new depsipep-
tide technique was shown to be comparable to that of the well-
known pseudoproline technique.
S6(Ac)-K37 and incomplete esterification was also detected when
the second depsi unit was introduced at KT 12/13 instead of
KT 17/18. Replacing NMI by the stronger acylation catalyst
DMAP did not help to improve the yield, and even worse results
were found using structure disrupting solvents such as DMF,
NMP, and DMSO, probably because of inefficient activation
of the carboxylic group via the carbodiimide. In the routine
procedure, O-acylation was always followed by a capping step
(Ac2O/NMI), thus allowing one to distinguish the products of
incomplete acylation from those of DKP formation, for which
the alcoholic function on the side chain is not acetylated.
Comparison of the Depsipeptide and Pseudoproline Tech-
niques. As part of the present work a comparison was made
between the new depsipeptide technique and the well-known
pseudoproline technique of Mutter and co-workers.2 This was
done for most of the depsipeptides examined in the case of both
the FBP28 domains and the Crambin sequence which is
described below and in all cases both methods were shown to
be equally efficient. Figures 3 and 5 show the results obtained
for the synthesis of the WW domain (the results for the synthesis
of Crambin are reported in the Supporting Information). It may
be noted that during the synthesis of peptide 3e (Figure 5F)
coupling of the pseudoproline at KT 12/13 had to be performed
manually (via HATU/DIEA/NMP/2 × 2 h, following standard
acetylation), since the standard procedure led to incomplete
coupling of about 40%.
Fully Automated Solid-Phase Synthesis via the Use of
Depsidipeptide Units. The necessity of performing an O-acyla-
tion step represents a possible drawback of the depsipeptide
method in comparison with the pseudoproline technique, since
N-acylation is notoriously more efficient and more likely to be
applied to automated protocols. Therefore, we prepared suitable
Fmoc derivatives of depsidipeptides in order to study their use
in the simultaneous introduction of two adjacent amino acids.
Scheme 3 shows two alternative methods for the synthesis of
the Ala/Thr system. The optical purity of depsidipeptide 8 was
investigated both by comparison of its NMR spectrum with that
of its authentically synthesized diastereoisomer Boc-Thr(Z-D-
Ala)-OBn (8a) and via an HPLC-based derivatization technique
using the Yamada reagent (Z-Val-Aib-Gly-OH)25 (see the
Supporting Information). No significant loss of configuration
was detected.
Experimental Section
General Procedures for the Synthesis of (VT)10 Sequence and
[Asn15]FBP28-WW. Standard solid-phase peptide synthesis was
performed on a peptide synthesizer using a standard Fmoc/tBu
protocol (0.25 M). A TentaGel-S-Ram resin (capacity 0.26
mmol/g) was used as solid support. Each cycle comprised double
couplings, unless otherwise stated. Manual standard couplings of
amino acid were carried out at a concentration of 0.25 M in DMF
using N-HBTU/DIEA activation (1:2 equiv) without preactivation,
couplings via DIC/HOBt (1:1 equiv) were carried out in DCM,
with a 5 min preactivation period. Fmoc removal was achieved by
piperidine 20% v/v in DMF (2 × 5 min). Standard O-acylations
were carried out at a concentration of 0.25 M in dry DCM for a
(16) Do¨lling, R.; Beyermann, M.; Haenel, J.; Kernchen, F.; Krause, E.;
Franke, P.; Brudel, M.; Bienert, M. J. Chem. Soc., Chem. Commun. 1994,
853-854.
(17) (a) Carpino, L. A.; Philbin, M.; Ismail, M.; Truran, G. A.; Mansour,
E. M. E.; Iguchi, S.; Ionescu, D.; El-Faham, A.; Warrass, R.; Weiss, M. S.
J. Am. Chem. Soc. 1997, 119, 9915-9916. (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, 4324-4338.
(18) Kertscher, U.; Bienert, M.; Krause, E.; Sepetov, N. F.; Mehlis, B.
Int. J. Pept. Prot. Res. 1993, 41, 207-211.
(19) Mosher, C. W.; Goodman, L. J. Org. Chem. 1972, 37, 2928-2933.
(20) Jiang, W.; Wanner, J.; Lee, R. J.; Bounaud, P. Y.; Boger, D. L. J.
Am. Chem. Soc. 2003, 125, 1877-1887.
(21) Bang, D.; Chopra, N.; Kent, S. B. H. J. Am. Chem. Soc. 2004, 126,
1377-1383.
(22) Cohen, S. B.; Halcomb, R. L. J. Am. Chem. Soc. 2002, 124, 2534.
(23) 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.
(24) Han, Y.; Albericio, F.; Barany, G. J. Org. Chem. 1997, 62, 4307.
(25) Yamada, T.; Nonomura, S.; Fujiwara, H.; Miyazawa, T.; Kuwata,
S. J. Chromatogr. 1990, 515, 475.
J. Org. Chem, Vol. 71, No. 16, 2006 6175

Coin et al.
FIGURE 6. Syntheses of crambin (16-46): HPLC profiles of crude products and UV monitoring traces for deblocking steps (insets). The UV
monitoring data measure, for each coupling cycle, the absorbance of the solution containing the Fm-piperidine adduct released during the Fmoc
removal step, which is an indication both of the efficiency of the previous coupling cycle and of the efficiency of the deblocking step itself. For
aggregated systems, where deblocking is slow, the instrument is programmed to repeat the deblocking step until the process is essentially complete.
For the full-size monitoring traces, see the Supporting Information: (A) standard synthesis (12a); (B) automated synthesis of depsi-12b (depsidipeptide
AT at 27/28). Key: *, target peptide, 27% purity; #, target depsipeptide, 44% purity.
time of 2 × 2 h, using DIC/NMI activation (1:1 equiv) without
preactivation. Standard acetylation was carried out using a solution
of 20% Ac2O, 10% DIEA in DMF (15 min). Acetylations in the
presence of NMI were performed using Ac2O/NMI (1:1 equivalents)
at a concentration of 0.25 M in DCM (2 × 90 min). All peptides
and depsipeptides were cleaved from the resin with TFA over a
period of 1-3 h in the presence of scavengers (standard cleavage
solution: TFA 88%, water 5%, TIS 2%, phenol 5%) and precipi-
tated in diethyl ether. O,N-Shifting of the depsipeptide to the amide
form was achieved by treatment in aqueous buffer (NaHCO3 0.1
M, pH 8.5) for 1 h.
(VT)10-NH2 (1) was automatically assembled on a TentaGel-
SRam resin (capacity: 0.26 mmol/g), using a standard 0.25 M Fmoc
protocol, performing double coupling at each step. MALDI-MS
spectrum is reported in the text.
Depsi-(VT)10-NH2 (1a). Boc-Thr-OH was linked to a TentaGel-
SRam resin (capacity: 0.26 mmol/g) by standard coupling. The
esterification of the hydroxyl group with Fmoc-Val-OH was carried
out according to the general procedure. A capping step in the
presence of NMI followed. Fmoc was deblocked with 20%
piperidine in DMF and Boc-Thr-OH coupled. These steps were
repeated in the same way up to completion of the sequence.
Cleavage was performed in standard TFA solution for 3 h. MALDI-
MS spectrum is reported in the text.
[Asn15]FBP28-NH2 (3c, via Four Depsi Units: ES 27/28, KT
17/18, KT 12/13, VS 5/6). Fmoc-Lys(Boc)-OH was linked to a
TentaGel-SRam resin (capacity: 0.26 mmol/g) by standard cou-
pling, followed by acetylation in the presence of NMI. The sequence
was extended to Y19 with introduction of the first depsi unit at ES
27/28 as described for peptide 2a. Boc-Thr-OH was manually
coupled and O-acylated by Fmoc-Lys(Boc)-OH. Acetylation in the
presence of NMI followed. The Fmoc group was removed and
Bsmoc-Gly-OH coupled by the standard procedure. Bsmoc removal
was carried out with 2% piperidine v/v in DMF (3 × 1 min), the
peptide-resin was washed with DMF (1 min) and Fmoc-Asn(Trt)-
OH coupled in the standard way. Fmoc-Ala-OH and Boc-Thr-OH
were successively coupled by the standard procedure. O-Acylation
by Fmoc-Lys(Boc)-OH and acetylation in the presence of NMI
followed. The Fmoc group was removed and Bsmoc-Tyr(tBu)-OH
coupled by the standard procedure. Bsmoc removal was carried
out with 2% piperidine v/v in DMF (3 × 1 min), the peptide-resin
was washed with DMF (1 min), and Fmoc-Glu(OtBu)-OH was
coupled by the standard procedure. The sequence was extended up
to Glu7 by the standard protocol. Boc-Ser-OH was manually coupled
and O-acylated by Fmoc-Leu-OH. Acetylation in the presence of
NMI followed. The Fmoc group was removed and Bsmoc-Ala-
OH coupled by the standard procedure. Bsmoc removal was carried
out with 2% piperidine v/v in DMF (3 × 1 min), the peptide-resin
was washed with DMF (1 min), and the last amino acids
subsequently coupled via the standard procedure. Cleavage was
carried out with a standard TFA solution (3 h). The depsi peptide
obtained GATA-VS-EWTEY-KT-ANG-KT-YYYNNRTL-ES-TWEK-
PQELK-NH2 was shifted to the amide form in 0.1 M aqueous
bicarbonate (1 h). MALDI-MS: [M + H]+calc ) 4358.70 (average),
found 4358.61.
[Asn15]FBP28-NH2 (3b, via Two Depsi Units: ES 27/28, KT
17/18). Fmoc-Lys(Boc)-OH was linked to a TentaGel-SRam resin
(capacity: 0.26 mmol/g) by standard coupling, followed by
acetylation in the presence of NMI. The sequence was extended to
Y19 with introduction of the first depsi unit at ES 27/28 as described
for peptide 2a. Boc-Thr-OH was manually coupled and O-acylated
by Fmoc-Lys(Boc)-OH. Acetylation in the presence of NMI
followed. The Fmoc group was deblocked and Bsmoc-Gly-OH
coupled by the standard procedure. Bsmoc removal was carried
out with 2% piperidine v/v in DMF (3 × 1 min), the peptide-resin
was washed with DMF (1 min) and Fmoc-Asn(Trt)-OH coupled
in the standard way. The rest of the sequence was completed
automatically. Cleavage was carried out with standard TFA solution
(3 h). The depsi peptide obtained, GATAVSEWTEYKTANG-KT-
YYYNNRTL-ES-TWEKPQELK-NH2, was shifted to the native
Ac-LE-(+)lac-TWEKPQELK-NH2 (Peptide 5a, Standard
Procedure). The sequence TWEKPQELK-NH2 was automatically
assembled on a TentaGel-SRam resin (capacity: 0.26 mmol/g),
using a standard 0.25 M Fmoc protocol. Lactic acid was coupled
by the standard method and acylated by means of Fmoc-Glu(OtBu)-
OH, with a subsequent acetylation step in the presence of NMI.
Fmoc was removed in the standard way and Fmoc-Leu-OH coupled.
The Fmoc protected peptide was treated with 20% piperidine v/v
in DMF (2 × 5 min), washed with DMF (1 min), and immediately
acetylated using a solution of acetic anhydride 50% v/v in DMF
amide form in NaHCO3 0.1M (1h). MALDI-MS: [M + H]+
4358.70 (average), found 4358.42.
)
calc
6176 J. Org. Chem., Vol. 71, No. 16, 2006

Depsipeptide Methodology for Solid-Phase Peptide Synthesis
(1 h). Peptide 5a was cleaved from the resin under standard cleavage
12a, Crambin (16-46) (standard synthesis): ESI-MS [M +
2H]2+ ) 1571.82, found 1571.7, [M + 3H]3+ ) 1048.21,
conditions (1 h). LC-MS: retention time 18.0min, [M + H]+
calc
calc
calc
) 1513.766 (monoisotopic) for 5a, found 1513.87; retention time
found 1047.9 [M + 4H]4+ ) 786.4, found 784.6. See Figure 1
calc
16.2 min: [M + H]+ ) 1271.640 (monoisotopic) for Ac-lac-
in the Supporting Information (p S25).
12b, Crambin (16-46) (via depsi unit AT 27/28, after shift):
calc
TWEKPQELK-NH2, found 1271.73.
ESI-MS reduced form [M + 2H]2+ ) 1571.82, [M + 3H]3+
Synthesis of Depsidipeptide 8. N-(tert-Butyloxycarbonyl)-
threonine Benzyl Ester (7). Following the method of Halcomb
and Cohen,22 13 g (59.3 mmol) of N-(tert-butyloxycarbonyl)-
threonine was dissolved in 250 mL of methanol and treated with
11 g (33.7 mmol) of Cs2CO3. The methanol was removed with a
rotary evaporator and the resulting oil was treated with 100 mL of
ethyl ether and the ether evaporated. The residual white solid was
redissolved in 125 mL of dry DMF and 8 mL (67.4 mmol) of benzyl
bromide was added slowly. The reaction mixture was stirred at room
temperature and the solvent removed by distillation under high
vacuum. During distillation of solvent the temperature of the oil
bath was kept at about 50 °C. The reaction mixture was partitioned
between ethyl acetate and water, and the organic layer was collected,
washed with 10% sodium bicarbonate solution and brine, and dried
over MgSO4. The crude benzyl ester was purified by column
chromatography on silica gel using ethyl acetate/ hexane 3:2 as
eluant to give 14.8 g (81%) of 7 as a colorless oil which was used
as such in the next step: 1H NMR (400 MHz, CDCl3) δ 1.24 (d,
3), 1.44 (s, 9), 2.1 (s, 1), 4.3 (m, 2), 5.2 (q, 2), 5.32 (d, 1), 7.3-
calc
) 1048.21, oxidized form [M +ca2lcH]2+
) 1569.82, [M +
calc
3H]3+
) 1046.88, found [M + 2H]2+) 1569.0, [M +
calc
3H]3+)1046.6. See Figure 6 in the Supporting Information(p S34).
Determination of Loss of Configuration during the Synthesis
of Depsidipeptide 8 by O-Acylation of Boc-Thr-OBn with Z-Ala-
OH via an NMR Technique. A small sample of the crude product
8 described in the text was removed prior to purification by column
1
chromatography, dissolved in CDCl3, and the H NMR spectrum
recorded at 400 Hz (462 scans). Examination of the spectrum at δ
1.19 showed no significant amount of the LD-form to be present
(see pp S20-21 (Supporting Information) for the 1H NMR spectral
data).
Determination of Loss of Configuration during the Synthesis
of Depsidipeptide 8 via an HPLC-Based Derivatization Tech-
nique. A. Removal of Boc Protection To Give H-Thr(Z-Ala)-
OBn. A 0.5 g (0.971 mmol) sample of crude 8 was dissolved in
10 mL of DCM/TFA (1:1, v/v), and the mixture was allowed to
stand at room temperature for 2 h. The solvent was removed by
means of a rotary evaporator. Traces of TFA were removed by
repetitive treatment with DCM followed by evaporation of the
solvent. Finally, the oil was dried under high vacuum and the free
amino depsidipeptide used without further purification.
B. Derivatization of H-Thr(Z-Ala)-OBn. To an ice-cold
solution of 120 mg (0.24 mmol) of H-Thr(Z-Ala)-OBn, obtained
as described in A above, 120 mg (0.3 mmol) of the Yamada
derivatization reagent Z-Val-Aib-Gly-OH,25 and 0.156 mL (0.89
mmol) of DIEA in 1 mL of DMF was added in one portion 114
mg (0.3 mmole) of N-HATU and the mixture stirred at room
temperature for 2 h. From the reaction mixture a small sample (0.1
mL) was withdrawn and diluted with acetonitrile to 2 mL, and from
this solution 20 µL was injected onto an HPLC column. The same
procedure was applied for the preparation of the LL(D)-isomer from
H-Thr(Z-D-Ala)-OBn in order to determine the appropriate retention
times of the two depsipentapeptides. Examination of the HPLC trace
derived from crude 8 showed that no more than 1% of the LL(D)-
isomer could have been present. The identity of the LL(L)- and LL-
(D)-isomers was confirmed by collection of the appropriate HPLC
fractions and subjecting each isomer to MALDI-MS analysis: LL-
(L)- calcd [M + Na]+ 812.348, found 812.320; LL(D)- calcd [M +
Na]+ 812.348, found 811.969. See pp S38-S40 in the Supporting
Information. The synthesis of the authentic LL(D)--form was carried
out using a sample of 8a following the derivatization technique
given for 8.
7.39 (m, 5); IR (DCM) 1741, 1714 cm-1
.
O-[(N-Benzyloxycarbonyl)alanyl]-N-(tert-butyloxycarbonyl)-
threonine Benzyl Ester (8). To an ice-cold solution of 3 g (9.708
mmol) of 7, 4.5 g (20.18 mmol) of Z-Ala-OH, and 0.355 g (2.9
mmol) of DMAP in 100 mL of DCM was added in portions 5.6 g
(29 mmol) of EDC. The reaction mixture was stirred at 0 °C for 1
h and then at room temperature for 4 h. The reaction mixture was
diluted with 750 mL of ethyl acetate, washed with 5% citric acid,
10% sodium bicarbonate solution, and brine, and dried over MgSO4.
After removal of the solvent with a rotary evaporator, the crude
1
product was sampled by H NMR for chiral purity (see later) and
the main portion purified by column chromatography on silica gel
using as eluant ethyl acetate/hexane (at first 3:7 and increasing the
ethyl acetate component gradually to 1:1) to give 4.5 g (97%) of
the pure ester as an oil: 1H NMR (400 MHz, CDCl3) δ 1.27 (d,
3), 1.31 (d, 3), 1.46 (s, 9), 4.18 (m, 1), 4.48 (dd, 1), 5.07-5.2 (m,
6), 5.25 (d, 1) 5.45 (m, 1), 7.26-7.37 (m, 10); IR (DCM) 1719
cm-1. Anal. Calcd for C27H34N2O8: C, 63.02; H, 6.66; N, 5.44;
Found: C, 63.04; H, 6.68; N, 5.42.
General Procedures for the Synthesis of Crambin (16-46).
All of the syntheses were carried out on a peptide synthesizer using
an UV detector for monitoring the deblocking step and equipped
with a 125-µm measuring loop. The peptides were assembled onto
an Fmoc-Amide resin with a loading of 0.67 mmol/g via Fmoc
chemistry on a 20-µmol scale with 5 equiv of the appropriate Fmoc-
AA-OH and 10 equiv of DIEA using standard instrument files. The
protected amino acids were dissolved in 125 µL of dry NMP,
N-HATU was dissolved in DMF (0.38 M), and DIEA (1.6 M) was
dissolved in NMP. The final concentration of the amino acid was
0.2 M. The activator and the base were delivered to the amino acid
cartridge, the mixture was preactivated for 30 s, the resulting
mixture was delivered to the reaction vessel, and the mixture was
shaken for 30 min. The deblocking step was carried out with 30%
piperidine in NMP for 2 × 3.5 min. After the final deblocking
step, the resin was washed with DCM and ethyl ether and dried
under vacuum and the peptide removed from the resin by treatment
with TFA in the presence of Reagent R24 (TFA-thioanisole-1,2-
ethanedithiol-anisole, 90:5:3:2) for 1 h. The TFA was removed with
a rotary evaporator, the peptide precipitated with cold ethyl ether
and lyophilized from acetic acid/water. Syntheses of analogues of
Crambine via use of depsidipeptide or pseudoproline units were
carried out under the same conditions given above, introducing the
building block in place of the threonine cartridge. Crude peptides
were dissolved in acetonitrile/water and injected onto an HPLC
column for HPLC analysis.
Acknowledgment. This work was supported in Berlin by
the Deutsche Forschungsgemeinschaft (BE 1434/5-1) and in
Amherst by the National Institutes of Health (GM-09706) and
the National Science Foundation (CHE-0078971). The National
Science Foundation is also thanked for grants used to purchase
the high-field NMR spectrometers used in this research. We
thank Annerose Klose, Heidemarie Lerch, and Dagmar Krause
for excellent technical assistance at the FMP, Berlin, and Dr.
Charlie Dickinson in Amherst for help with the COSY analysis.
Supporting Information Available: Description of general
experimental techniques. Detailed procedure of synthesis of com-
pounds 2, 2a,b, 3, 3a,d,e, 4, 5b, 6, 8a, 10, 11, and 9. Spectral data
for all new compounds. Monitoring traces and HPLC data for
automated Crambin (16-46) syntheses. Detailed evidence dem-
onstrating the lack of loss of configuration during the synthesis of
the Thr/Ala depsidipeptide. This material is available free of charge
via the Internet at http://pubs.acs.org.
JO060914P
J. Org. Chem, Vol. 71, No. 16, 2006 6177