Design, synthesis, and evaluation of a depsipeptide mimic of tendamistat

Article abstract of DOI:10.1021/jo9616062

The cyclic hexadepsipeptide framework of enniatin B was identified as a template matching the β-turn tripeptide of tendamistat. The modified analog 1 was synthesized as a tendamistat mimic and compared to the acyclic derivative 2 and the tripeptide Ac-Try-Arg-Tyr-OMe. These compounds were assembled from the dimeric esters 3-5. As an inhibitor of α-amylase, 1 is twice as potent as 2 and comparable to the tripeptide. NMR studies of 1 reveal four conformers in equilibrium in a 50:25:15:10 ratio; the ring conformation of the major component is similar to that of the enniatin B template, with the cis geometry of the α-hydroxyisovaleryl-N-methylvaline amide linkage; the other conformers differ in the position or presence of the cis amide linkage.

Full text of DOI:10.1021/jo9616062

J . Org. Chem. 1997, 62, 93-102
93
Design , Syn th esis, a n d Eva lu a tion of a Dep sip ep tid e Mim ic of
Ten d a m ista t
Andrea M. Sefler, Marisa C. Kozlowski, Tao Guo, and Paul A. Bartlett*
University of California, Department of Chemistry, Berkeley, California 94720-1460
Received August 19, 1996X
The cyclic hexadepsipeptide framework of enniatin B was identified as a template matching the
â-turn tripeptide of tendamistat. The modified analog 1 was synthesized as a tendamistat mimic
and compared to the acyclic derivative 2 and the tripeptide Ac-Try-Arg-Tyr-OMe. These compounds
were assembled from the dimeric esters 3-5. As an inhibitor of R-amylase, 1 is twice as potent as
2 and comparable to the tripeptide. NMR studies of 1 reveal four conformers in equilibrium in a
50:25:15:10 ratio; the ring conformation of the major component is similar to that of the enniatin
B template, with the cis geometry of the R-hydroxyisovaleryl-N-methylvaline amide linkage; the
other conformers differ in the position or presence of the cis amide linkage.
In tr od u ction
sequence 18Trp-19Arg-20Tyr plays in the interaction with
the enzyme active site.22,23 The crystal and NMR struc-
tures of tendamistat show these residues to be in a
â-Turns are common motifs in peptide-protein inter-
actions. Their convex structure presents the amino acid
side chains in a dense array, providing variability and,
thus, specificity in an efficient format. This secondary
structure consists of four residues in which the carbonyl
oxygen of the i residue is hydrogen bonded to the amide
hydrogen of the i + 3 residue.1-3 â-Turns are central to
many physiologically important interactions,4-6 which
has prompted the design of a number of rigid mimics and
strategies for inducing the conformation in an oligo-
peptide.7-11 One of the most straightforward methods
of inducing a â-turn is to incorporate the desired sequence
in a cyclic peptide,12 an approach that has been used suc-
cessfully in designing mimics of somatostatin,13,14 ten-
damistat,15,16 and other ligands.17-20
slightly distorted type I â-turn conformation.24,25
A
number of approaches to the design of small molecule
inhibitors of tendamistat have been described (Figure
1): incorporating the Trp-Arg-Tyr sequence in a cyclic
hexapeptide framework,15,16 complexing it with copper
ion,26 and embedding it in a rigid, polycyclic template.27
The previously described designs of tendamistat mim-
ics15,26,27 have used the program CAVEAT to identify
templates that match the backbone of the â-turn and that
can position the amino acid side chains of the 18Trp-
19Arg-20Tyr unit in the same orientation.28 A CAVEAT
search of the Cambridge Structural Database29 also
revealed that the backbone and side-chain atoms of
enniatin B (CSD code VHVIMH10) provide a very close
match to those of the tendamistat binding loop (see
Figure 1d);30 enniatin B is a cyclic hexadepsipeptide with
DLLLLL stereochemistry.31 The cyclic hexadepsipeptide is
an intriguing template for structure-based design for
several reasons. The macrocycle itself provides an en-
tropic advantage in binding if the molecule is constrained
in the correct conformation. The cyclic structure and
N-methylation convey protease resistance and oral bio-
availability, as exemplified by cyclosporin A32,33 and a
number of somatostatin and oxytocin mimics.13,34 Ad-
ditionally, the template is comprised of readily available
R-hydroxy acid and N-methylamino acid subunits.
Tendamistat, a potent microbial inhibitor of R-amy-
lase,21 has been used as a design target for â-turn tem-
plates because of the prominent role that the tripeptide
X Abstract published in Advance ACS Abstracts, December 15, 1996.
(1) Chou, P. Y.; Fasman, G. D. J . Mol. Biol. 1977, 115, 135.
(2) Wilmot, C. M.; Thornton, J . M. J . Mol. Biol. 1988, 203, 221.
(3) Richardson, J . S. Adv. Protein Chem. 1981, 34, 167.
(4) Smith, J . A.; Pease, L. G. CRC Crit. Rev. Biochem. 1980, 8, 315.
(5) Kabasch, W.; Sander, C. Biopolymers 1983, 22, 2577.
(6) Rose, G. D.; Gierasch, L. M.; Smith, J . A. Adv. Protein Chem.
1986, 37, 1.
(7) Kemp, D. S. Trends Biotech. 1990, 8, 249.
(8) Ball, J . B.; Alewood, P. F. J . Mol. Recog. 1990, 3, 55.
(9) Ho¨lzemann, G. Kontakte (Darmstadt) 1991, 1, 3.
(10) Hirschmann, R. Angew. Chem., Int. Ed. Engl. 1991, 30, 1278.
(11) Aubry, A.; Boussard, B.; Cung, M. T.; Marraud, M.; Vitoux, B.
J . Chim. Phys. 1988, 85, 345.
(21) Ve´rtesy, L.; Oeding, V.; Bender, R.; Zepf, K.; Nesemann, G. Eur.
J . Biochem. 1984, 141, 505.
(12) Marepalli, H. R.; Antohi, O.; Becker, J . M.; Naider, F. J . Am.
Chem. Soc. 1996, 118, 6531.
(13) Veber, D. F.; Freidinger, R. M.; Perlow, D. S.; Paleveda, W. J .;
Holly, F. W.; Strachan, R. G.; Nutt, R. F.; Arison, B. H.; Homnick, C.;
Randall, W. C.; Glitzer, M. S.; Saperstein, R.; Hirschmann, R. Nature
(London) 1981, 292, 55.
(22) Hofmann, O.; La´szlo´, V.; Braunitzer, G. Biol. Chem. Hoppe-
Seyler 1985, 366, 1161.
(23) Goto, A.; Matsui, Y.; Ohyama, K.; Arai, M.; Murao, S. Agric.
Biol. Chem. 1985, 49, 435.
(24) Pflugrath, J . W.; Wiegand, G.; Huber, R.; Ve´rtesy, L. J . Mol.
Biol. 1986, 189, 383.
(14) Kessler, H.; Bats, J . W.; Griesinger, C.; Koll, S.; Will, M.;
Wagner, K. J . Am. Chem. Soc. 1988, 110, 1033.
(15) Etzkorn, F. A.; Guo, T.; Lipton, M. A.; Goldberg, S. D.; Bartlett,
P. A. J . Am. Chem. Soc. 1994, 116, 10412.
(16) Matter, H.; Kessler, H. J . Am. Chem. Soc. 1995, 117, 3347.
(17) Bean, J . W.; Kopple, K. D.; Peishoff, C. E. J . Am. Chem. Soc.
1992, 114, 5328.
(18) Kopple, K. D.; Baures, P. W.; Bean, J . W.; D’Ambrosio, C. A.;
Hughes, J . L.; Peishoff, C. E.; Eggleston, D. S. J . Am. Chem. Soc. 1992,
114, 9615.
(19) McDermott, J .; Benoiton, N. Can. J . Chem. 1973, 51, 1915.
(20) Peishoff, C. E.; Ali, F. E.; Bean, J . W.; Calvo, R.; D’Ambrosio,
C. A.; Eggleston, D. S.; Hwang, S. M.; Kline, T. P.; Koster, P. F.;
Nichols, A.; Powers, D.; Romoff, T.; Samanen, J . M.; Stadel, J .; Vasko,
J . A.; Kopple, K. D. J . Med. Chem. 1992, 35, 3962.
(25) Kline, A. D.; Braun, W.; Wu¨thrich, K. J . J . Mol. Biol. 1988,
204, 675.
(26) Tian, Z.-Q.; Bartlett, P. A. J . Am. Chem. Soc. 1996, 118, 943.
(27) Kozlowski, M. C.; Bartlett, P. A. J . Org. Chem. 1996, 61, 7681.
(28) Lauri, G.; Bartlett, P. A. J . Comput. Aided Mol. Design 1994,
8, 51.
(29) Allen, F. H.; Kennard, O.; Taylor, R. Acc. Chem. Res. 1983, 16,
146.
(30) Bartlett, P. A.; Shea, G. T.; Telfer, S. J .; Waterman, S. In
Molecular Recognition: Chemical and Biological Problems; Roberts,
S. M., Ed.; Royal Society of Chemistry: London, 1989; p 182.
(31) Shishova, T. G.; Simonov, V. I. Kristallografia 1977, 22, 515.
(32) CyclosporinesNature of the Agent and its Immunologic Actions;
Kahan, B. D., Ed.; Grune and Stratton: New York, 1988.
(33) Ciclosporin; Borel, J . F., Ed.; Karger: Basel, 1986.
S0022-3263(96)01606-4 CCC: $14.00 © 1997 American Chemical Society

94 J . Org. Chem., Vol. 62, No. 1, 1997
Sefler et al.
F igu r e 1. Tendamistat â-turn24 (bare lines) superimposed on peptidomimetic templates (spherical atoms): (a) a cyclic
hexapeptide;15,16 (b) a tetrapeptide-copper complex;26 (c) a rigid tricycle;27 and (d) the cyclic hexadepsipeptide enniatin B.
Transformation of this structural lead into a potential
tendamistat mimic involves replacement of the isopropyl
side chains at the matching positions of enniatin B with
the indolylmethyl, guanidinopropyl, and hydroxybenzyl
groups of the tendamistat â-turn (Figure 2). In this
paper, we describe the synthesis of this derivative, 1, a
comparison of its binding affinity with other tendamistat
mimics and the acyclic analog 2, and determination of
its solution conformations by NMR.
quently, we formed the dimeric esters 3-5 at the outset
and assembled the linear hexamer without exposing the
hydroxyl groups (Scheme 2). Since the branched side-
chain subunits were expected to present the greatest
problem in the coupling reactions, the amides adjacent
to N-methylvaline and R-hydroxyisovalerate residues
were formed first; cyclization involved formation of the
amide between the hydroxy-Arg and N-Me-Tyr precur-
sors. Orthogonal protecting groups were employed to
allow selective deprotection of the side chain, amine, or
carboxylic acid functional groups. Moreover, we used the
ornithine side chain as an arginine precursor to avoid
problems that the reactive guanidino group could present
in the difficult coupling steps.
The forward synthesis is outlined in Schemes 3-5.
L-R-Hydroxyisovaleric acid, 6a , (Scheme 3) was prepared
from L-valine by diazotization35 and protected as the
2-(trimethylsilyl)ethyl (TMSE) ester, 6b.36 The yield was
low in this esterification step (ca. 40%) since dimerization
of the hydroxy acid was a significant side reaction.
Nevertheless, the TMSE protecting group was advanta-
Resu lts
Syn th esis. In principle, the cyclic hexadepsipeptide
1 could be synthesized from the hydroxy and amino acid
subunits by a number of strategies; however, the ease
with which tertiary amides with free hydroxyacyl groups
would lactonize argued against assembly of the linear
hexamer from the monomers in a linear stepwise se-
quence or by a convergent strategy in which the amide
linkages would be formed first (Scheme 1). Conse-
(34) Freidinger, R. M.; Williams, P. D.; Tung, R. D.; Bock, M. G.;
Pettibone, D. J .; Clineschmidt, B. V.; DiPardo, R. M.; Erb, J . M.;
Garsky, V. M.; Gould, N. P.; Kaufman, M. J .; Lundell, G. F.; Perlow,
D. S.; Whitter, W. L.; Veber, D. F. J . Med. Chem. 1990, 33, 1843.
(35) Li, W.; Ewing, W.; Harris, B.; J oullie, M. J . Am. Chem. Soc.
1990, 112, 7659.
(36) Sieber, P. Helv. Chim. Acta 1977, 60, 2711.

Synthesis of a Depsipeptide Mimic of Tendamistat
J . Org. Chem., Vol. 62, No. 1, 1997 95
minimized hydrolysis of the depsipeptide esters, which
was a significant side reaction with the standard TBAF/
THF solution. Also, because of the sensitive depsipeptide
esters, Boc deprotection (e.g., 3 f 3a ) was accomplished
with dilute TFA in methylene chloride, which gave a
cleaner product than the standard conditions with neat
TFA. After carboxyl deprotection of 3 and amine depro-
tection of 4, the dimers were coupled to form tetramer
7a with bromotris(pyrrolidino)phosphonium hexafluoro-
phosphate (PyBroP),38,39 a highly activated coupling
reagent that works well with hindered N-methyl sub-
strates. The TMSE ester was cleaved to prepare tet-
ramer 7b for the next coupling step.
The R-hydroxy analog of ornithine was formed by direct
diazotization of the diamino acid and protected as the
N-Cbz, TMSE ester 9c (Scheme 4).40,41 In preparing the
N-methyl-D-tryptophan derivative 10c, the indole nitro-
gen was first protected as the tert-butyldimethylsilyl
derivative to prevent alkylation at this position.42 Cou-
pling of these two components and subsequent protecting
group manipulations afforded the third dimeric interme-
diate, 5.
Although the reasons are not readily apparent, the
TBDMS group had to be removed prior to coupling the
dimeric and tetrameric fragments 5b and 7b; otherwise,
this reaction proceeded very poorly. HCl in ethyl acetate
was used for Boc-deprotection of 5 and a subsequent
intermediate (11b), since this reagent appeared to mini-
mize tert-butylation of the indole ring. PyBroP was used
both to couple tetramer 7b with dimer 5b and, after N-
and C-deprotection, to cyclize the linear hexamer 11c
(Scheme 5). Hydrogenolysis of the Cbz and benzyl
protecting groups and introduction of the guanidino side
chain then afforded the target compound 1.
The linear analog 2 was synthesized in a similar
fashion using the three dimeric fragments from prepara-
tion of the cyclic compound (Scheme S1 in the Supporting
Information). The coupling and deprotection steps were
altered to produce the linear hexadepsipeptide with the
residues mimicking the Trp-Arg-Tyr sequence at posi-
tions 3-5. After the final coupling step, the side-chain
protection was removed by hydrogenolysis and the guani-
dino group was introduced prior to removal of the
terminal protecting groups.
F igu r e 2. Enniatin B as scaffolding for peptidomimetic 1 and
acyclic comparison compound 2.
Sch em e 1
In h ibition . The inhibition constants for the cyclic and
linear mimic 1 and 2 were determined as described
previously (Table 1).27 However, due to the poor solubil-
ity of 1 in water, the assays were carried out in 5%
DMSO/H2O; Ki values of the control tripeptides WRY and
Ac-WRY-OMe were redetermined in this solvent to allow
comparison with previous results. Compared to water
alone, there is a 2-fold increase in the Ki values of these
tripeptides in 5% DMSO/water, which suggests that
hydrophobic effects make a significant contribution to the
binding energy.27
geous since it can be cleaved in the presence of the
depsipeptide esters. N-Methylation of N-Boc-valine and
N-Boc-tyrosine benzyl ether was accomplished as de-
scribed by McDermott and Benoiton.19 A variety of
methods were explored for the subsequent esterification
reactions to give 3 and 4. While the divalyl depsipep-
tide 4 was formed under standard coupling conditions
with EDC and DMAP in pyridine, the formation of 3
required the more active isopropenyl chloroformate re-
agent. For deprotection of the TMSE esters (e.g., 4 f
4a ), we found that silica-supported tetrabutylammonium
fluoride (TBAF),37 a nonhygroscopic source of fluoride ion,
The cyclic hexadepsipeptide 1 has only a 2-fold greater
affinity for R-amylase than the linear analog 2, and it is
weaker than the cyclic hexapeptides (e.g., cyclo[D-
(37) Clark, J . J . Chem. Soc., Chem. Commun. 1978, 789.
(38) Castro, B.; Dormoy, J . R. Tetrahedron Lett. 1973, 3243.
(39) Coste, J .; Frerot, E.; J ouin, P. Tetrahedron Lett. 1991, 32,
1967.
(40) Ohshiro, S.; Kuroda, K.; Fujita, T. Yakugaku Zasshi 1967, 87,
1184.
(41) Suzuki, E.; Inoues, S.; Goto, T. Chem. Pharm. Bull. 1968, 16,
933.
(42) Grieco, P.; Hon, Y.; Perez-Medrano, A. J . Am. Chem. Soc. 1988,
110, 1630.

96 J . Org. Chem., Vol. 62, No. 1, 1997
Sefler et al.
Sch em e 2
Sch em e 3
Sch em e 4
PFAWRY]15,16 or peptide-copper complexes26 reported
previously. To discern whether the cyclic compound is
constrained in a conformation that does not complement
the active site of R-amylase, the solution structure of
compound 1 was determined by NMR.
NMR Str u ctu r e of 1. Although compound 1 was
purified to homogeneity by reversed-phase HPLC, four
sets of peaks were observed in the 1H-NMR spectrum
reflecting a population composition of 50:25:15:10 at room
temperature. A NOESY spectrum was recorded at 50
°C to determine whether the extra peaks arose from
coeluting impurities or conformational heterogeneity of
1. The molecule showed positive nuclear Overhauser
enhancements under these conditions, yet a number of
the cross peaks were of the opposite sign. These cross
peaks arise from chemical exchange, which indicates that
the multiplicity of signals at room temperature reflect
different conformations of the macrocycle. The structural
studies were therefore carried out at 0 °C to ensure that
the molecule did not undergo conformational exchange
during the mixing time of the NOESY experiments. At
this temperature, exchange cross peaks were no longer
present, but only two sets of signals were observed
because signals from the two least populated conformers
were below the detection limits. Thus, only the two most
abundant conformers (referred to as “major” and “minor”)
were studied under these conditions.
For further studies, compound 1 was dissolved in 40%
CD3OD/D2O with a KH2PO4 buffer containing NaCl and
CaCl2, and the pH was adjusted to 7.0 with DCl to
simulate assay conditions. CD3OD was used instead of
DMSO-d6 as a cosolvent because satisfactory shimming
could not be obtained at 0 °C with the more viscous
DMSO solution. Because of the known ability of dep-
sipeptides to chelate metals,43 spectra were recorded in
the absence as well as presence of the sodium and
(43) Ovchinnikov, Y. A.; Ivanov, V. T.; Evstatov, A. V.; Mikhaleva,
I. I.; Bystrov, V. F.; Portnova, S. L.; Balashova, T. A.; Meshcheryakova,
E. N.; Tulchinsky, V. M. Int. J Pept. Protein Res. 1974, 6, 465.

Synthesis of a Depsipeptide Mimic of Tendamistat
J . Org. Chem., Vol. 62, No. 1, 1997 97
Sch em e 5
F igu r e 3. Inter-residue NOEs observed for major (a) and
minor (b) conformers of 1.
aromatic side-chain patterns. The Val and hVal reso-
nances were then differentiated on the basis of sequential
NOE’s to either the Trp or the Tyr residues (Figure 3).
All hydrogens for the major conformer and most of the
hydrogens for the minor conformer were assigned in this
manner.
Ta ble 1. In h ibitor s of r-Am yla sea
Ki (µM) in 5%
aqueous DMSO
inhibitor
Ki (µM) in water
tendamistatb
0.0002
2.4 ( 0.2c
520 ( 20d
100 ( 3d
14 ( 2b
GWR-D-Y‚Cu+2 complex
Trp-Arg-Tyr
870 ( 30
190 ( 10
30e
290 ( 10
140 ( 10
The presence of NOE cross peaks from hVal3 RH to
Trp4 N-Me and from Arg5 RH to Tyr6 N-Me for the
major conformer indicates that these amide bonds are
in the trans configuration; in turn, the lack of an hVal1
RH to Val2 N-Me NOE cross peak suggests that this
amide is cis (see Figure 3). For the less populated
component, the hVal3-Trp4 amide is trans (RH to N-Me
NOE observed), while the absence of a Arg5 RH to Tyr6
N-Me cross peak suggests that the latter amide is cis.
NOE’s between the Arg5 and Tyr6 side chains support
this assignment, since cis peptide bonds bring these side
chains closer together than trans linkages do. For the
minor conformer, the position of the hVal1 RH to Val2
N-Me NOE cross peak was obscured by the water sig-
nal so it could not be confirmed; however, molecular-
modeling studies suggest that low-energy conformations
for the hexadepsipeptide contain only one amide in the
cis configuration, in analogy to enniatin B itself.31 If the
cis amide is found at the hVal1-Val2 and Arg5-Tyr6
positions in the most abundant conformers, it is a
reasonable assumption that one of the least populated
components observed at room temperature contains the
cis linkage between hVal3 and Trp4 and that the other is
all trans.
Ac-Trp-Arg-Tyr-OMe
c[D-PFAWRY]
2
1
pH 7, 30 °C, using p-nitrophenyl maltotrioside as substrate.27
a
b From ref 21. c From ref 26. From ref 15. e Estimated from a two-
d
point determination.
calcium salts. No difference was observed, suggesting
that cation complexation is not a factor in determining
the conformations of the cyclic mimic 1 in 40% CD3OD/
D2O.
TOCSY and NOESY spectra of 1 were recorded at 0
°C. Interpretation of the TOCSY spectrum was compli-
cated by the three similar Val and hVal patterns and the
two benzylic AMX spin systems (see side chain labels in
Figure 2);44 only the resonances from the hArg side chain
could be identified uniquely from the TOCSY experiment.
The Trp and Tyr residues could be distinguished by NOE
cross peaks from the â-proton resonances to their unique
(44) We refer to the side chains by their amino acid equivalents,
using the prefix “h” to designate the R-hydroxy analog of the amino
acid.

98 J . Org. Chem., Vol. 62, No. 1, 1997
Sefler et al.
F igu r e 4. Superposition of low-energy structures of major (a) and minor (b) conformers of 1 that satisfy observed NOE constraints;
Trp4, hArg5, and Tyr6 side chains omitted for clarity.
F igu r e 5. Enniatin B from crystal structure31 (bare lines) superimposed with lowest energy structure of major conformer of 1
(spherical atoms); hArg5 side chain omitted for clarity.
The absence of NH protons made it difficult to obtain
torsional constraints to define backbone φ angles. How-
ever, Monte Carlo conformational searches of the four
combinations of amide configurations, followed by clus-
tering by the backbone atoms,28 resulted in only a few,
low energy conformations for each ring system; almost
all of the conformational heterogeneity arises from side-
chain torsions. The structures of the two major compo-
nents (hVal1-Val2 or Arg5-Tyr6 amide cis) were then
filtered against the relevant set of NOE constraints
(Figure 3) to give a set of closely related structures that
were both low energy as well as in accord with the
spectral information (Figure 4). The backbone of the
lowest energy structure of the primary component (cis
hVal1-Val2 amide) overlays reasonably well with the
crystal structure of enniatin B (rmsd ) 1.1 Å for all
backbone atoms, see Figure 5) because the cis amide bond
is located in the same position. However, the structure
for the cis Arg5-Tyr6 amide, as well as the modeled
structures for the other amide isomers, have significantly
different ring conformations, and it appears unlikely that
they could bind to R-amylase.
The underlying design of mimic 1 is therefore partially
validated, in that the structure of the template is largely
preserved in the major conformer. The relative orienta-
tion of the side chain CR-Câ bonds is similar to that in
tendamistat, although not as closely aligned as in the
cyclic peptide cyclo[D-PFAWRY].15 However, the multiple
conformations of the cyclic hexadepsipeptide backbone
detract from the binding affinity of 1. If the major
conformer of 1 is the only one able to bind to R-amylase,
correcting for the inactive components projects an inhibi-
tion constant for this form of 70 µM. This inhibitor is
somewhat weaker than the best cyclic and complexed
peptides (Table 1), but 3-fold better than the linear
control Ac-Trp-Arg-Tyr-OMe. Comparing the backbone
conformation of the major form of 1 with the binding loop
of tendamistat suggests why this compound is not a
better inhibitor of R-amylase. As shown in Figure 6, the
backbone turn of the depsipeptide is wider than the
tripeptide. The recently reported high-resolution struc-
ture of the tendamistat/R-amylase complex45 shows that
the tendamistat packs tightly into the active site of the
enzyme around the 18Trp-19Arg-20Tyr tripeptide. This
structure indicates that the backbone and bulky isopropyl
side chains at the far side of the depsipeptide ring may
clash sterically with the enzyme pocket.
Exp er im en ta l Section 46
2-(Tr im eth ylsilyl)eth yl (S)-2-Hydr oxy-3-m eth ylbu tan o-
a te, 6b. EDC (17.0 g, 84.3 mmol) and a catalytic amount of
DMAP (2.0 g, 16.41 mmol) were added to a solution of 8.8 g
(74.6 mmol) of (S)-2-hydroxy-3-methylbutanoic acid, 6a (syn-
thesized according to Li et al.35), and 53.5 mL (373 mmol) of
2-(trimethylsilyl)ethanol in 100 mL of pyridine at 0 °C under
Ar; the mixture was allowed to warm to 25 °C and stirred for
36 h. The pyridine was removed under reduced pressure, and
the residue was dissolved in 250 mL of ethyl acetate, washed
with 100 mL of 1 N HCl and 100 mL of H2O, and dried over
MgSO4. The solvent was removed with a rotary evaporator
at room temperature, and the excess 2-(trimethylsilyl)ethanol
was recovered by vacuum distillation. Purification by frac-
(45) Wiegand, G.; Epp, O.; Huber, R. J . Mol. Biol. 1995, 247, 99-
110.

Synthesis of a Depsipeptide Mimic of Tendamistat
J . Org. Chem., Vol. 62, No. 1, 1997 99
F igu r e 6. Tendamistat â-turn24 (bare lines) superimposed with lowest energy structure of major conformer of 1 (spherical atoms),
aligning by CR and Câ atoms of Trp-Arg-Tyr residues.
tional distillation under reduced pressure gave 6.52 g (40%)
Ester , 4. EDC (2.49 g, 13.0 mmol) and a catalytic amount of
DMAP (309 mg, 2.53 mmol) were added to a solution of N-Boc-
N-Me-L-valine (2.93 g, 12.65 mmol) and hydroxy ester 6b (2.5
g, 11.5 mmol) in 20 mL of pyridine at 0 °C under Ar. The
mixture was allowed to warm to 25 °C and stirred for 36 h.
After the pyridine was removed under reduced pressure, the
residue was dissolved in 200 mL of EtOAc and washed with
of ester 6b as a colorless liquid: bp 65 °C/100 mm Hg; [R]20
D
-6.6° (c ) 20 mg/mL, ethyl acetate); 1H NMR (400 MHz,
CDCl3) δ 4.23 (m, 2), 4.08 (d, 1, J ) 6.61), 2.27 (m, 1), 1.02 (m,
6), 0.06 (s, 11); 13C NMR (100 MHz) δ 169.4, 64.3, 64.1, 32.6,
18.7, 17.4, -1.6; IR (CH2Cl2) 1750, 1470, 1170, 1050, 993, 942,
867, 845 cm-1; MS (EI) m/ z 218 (M+), 191 (9.81), 174 (11.41),
157 (43.61), 145 (35.51), 117 (47.56), 101 (65.84), 93 (75.39);
HRMS (EI) m/ z calcd for C10H22O3Si 218.1347, found 218.1336.
(2S)-2-[[N-(ter t-Bu toxyca r bon yl)-N-m eth yl-O-ben zyl-L-
tyr osyl]oxy]-3-m eth ylbu ta n oic Acid 2-(Tr im eth ylsilyl)-
eth yl Ester , 3. DIEA (1.4 mL) and DMAP (178 mg, 1.45
mmol) were added to a solution of N-Boc-N-Me-O-Bn-L-tyrosine
(2.8 g, 7.27 mmol) and hydroxy ester 6b (1.75 g, 8.00 mmol)
in 7 mL of CH2Cl2. The mixture was cooled to -78 °C, and
freshly distilled isopropenyl chloroformate (875 µL, 8.00 mmol)
was added dropwise with stirring over 10 min. The mixture
was allowed to warm to 0 °C; after 1 h, EtOAc (100 mL) was
added and the solution was washed with 5% aqueous KHSO4
(2 × 50 mL), 5% aqueous NaHCO3 (2 × 50 mL), and brine (50
mL). The organic layer was worked up, and the crude product
was chromatographed (20% EtOAc/hexanes) to give 2.76 g
(65%) of a light yellow oil: [R]20D 35.2° (c ) 10 mg/mL, CHCl3);
1H NMR (400 MHz, CDCl3) δ 7.43-7.29 (m, 5), 7.15-7.08 (m,
2), 6.89 (d, 2, J ) 8.4), 5.03 (s, 2), 4.89-4.70 (m, 2), 4.27-4.21
(m, 2), 3.33-2.86 (m, 2), 2.4 and 2.5 (2s, 3), 2.24 (bs, 1), 1.38
and 1.35 (2s, 9), 1.08-0.92 (m, 8), 0.05 (s, 9); 13C NMR (100
MHz, CDCl3) δ 171.0, 169.4, 157.5, 154.9, 137.0, 129.8, 128.5,
127.8, 127.4, 114.8, 80.3, 79.8, 77.5, 76.7, 70.0, 63.6, 63.5, 59.8,
56.6, 56.4, 34.1, 34.0, 32.6, 29.9, 28.2, 28.0, 27.6, 19.0, 18.9,
18.8, 17.4, 17.2, 17.1, -1.6; IR (CHCl3) 2470, 1710, 1513, 1420,
1362, 1040, 930 cm-1; MS (EI) m/ z 609 (MH + Na+), 586
(MH+); HRMS (FAB) m/ z calcd for C32H48NO7Si (MH+)
586.3200, found 586.3183.
100 mL of 1 N HCl and 100 mL of H
2O, the solution was
worked up, and the crude product was chromatographed (10%
EtOAc/hexanes) to give 3.03 g (61%) of the ester 4 as a colorless
oil: [R]20D 33.6° (c ) 23.0 mg/mL, CHCl3); 1H NMR (400 MHz,
CDCl3) δ 4.80 (d, 1, J ) 4.6), 4.33-4.17 (m, 3), 2.85-2.76 (2
br s, 3), 2.45-2.18 (m, 2), 1.57 (bs, 9), 1.15-0.98 (m, 12), 0.06-
0.03 (m, 11); 13C NMR (100 MHz, CDCl3) δ 170.8, 169.9, 169.1,
77.9, 76.98, 76.4, 64.3, 63.9, 63.6, 32.5, 30.1, 28.4, 28.0, 20.6,
19.6, 18.7, 18.1, 17.8, 17.4, 17.2, -1.6; IR (CHCl3) 2470, 1740,
1520, 1425, 1390, 1370, 1040, 930 cm-1; MS (EI) 431 (M+),
293 (1.89), 260 (1.59), 157 (47.13), 145 (43.95), 130 (55.27), 101
(65.35); HRMS (FAB) m/ z calcd for
C
21H41O6NSi (MH+)
432.2786, found 432.2786.
(2S)-2-[[N-(ter t-Bu toxyca r bon yl)-N-m eth yl-O-ben zyl-L-
tyr osyl]oxy]-3-m eth ylbu ta n oic Acid , 3a . Didepsipeptide
3 (910 mg, 1.6 mmol) was dissolved in 50 mL of THF in a
round-bottomed flask under Ar. TBAF on SiO2 (4 g of 1.1
mmol fluoride/g) was added, and the suspension was stirred
at room temperature for 5 h. The yellow mixture was poured
onto 25 mL of saturated NH4Cl and extracted with EtOAc (3
× 25 mL). The combined organic layers were washed with
brine and dried over Na2SO4, and the solvent was evaporated.
The crude product was chromatographed (95:5:0.1 CH2Cl2/
MeOH/HOAc) to give 770 mg (100%) of acid 3a as a yellow
oil: 1H NMR (400 MHz, CDCl3) δ 7.43-7.30 (m, 5), 7.32-7.08
(m, 2), 6.90 (d, 2, J ) 8.51), 5.03 (s, 2), 4.71 and 4.58 (2m, 1),
3.25 (m, 1), 3.10-2.92 (m, 2), 2.76 and 2.67 (2s, 3), 1.72 (m, 1),
1.40 and 1.32 (2s, 9), 1.02-0.96 (m, 6); 13C NMR (100 MHz,
CDCl3) δ 157.5, 156.2, 155.1, 136.9, 129.9, 129.5, 128.5, 128.2,
127.8, 127.3, 127.3, 114.8, 114.7, 80.3, 80.2, 69.9, 61.4, 51.9,
34.4, 33.9, 32.4, 32.1, 28.2, 28.1, 25.0, 20.0, 14.1, 13.5; IR
(CHCl3) 2450, 1740, 1685, 1510, 1455, 1390, 1370, 1330, 1220,
1030, 930, 910 cm-1; MS (FAB) m/ z 486 (MH+); HRMS (FAB)
m/ z calcd for C27H36NO7 (MH+) 486.2492, found 486.2481.
(2S)-2-[(N-Meth yl-L-va lyl)oxy]-3-m eth ylbu ta n oic Acid
2-(Tr im eth ylsilyl)eth yl Ester , 4a . Didepsipeptide 4 (700
mg, 1.62 mmol) was dissolved in 5 mL of CH2Cl2 in a flame-
dried round-bottomed flask purged with Ar. Dry TFA (190
µL, 2.44 mmol) was added dropwise, and the solution was
stirred for 24 h at room temperature. The solvent was
evaporated, and the crude product was chromatographed (10%
EtOAc/hexane) to give 536 mg (100%) of the deprotected amine
4a as a yellow oil: 1H NMR (400 MHz, CDCl3) δ 4.75 (d, 1, J
) 4.54), 4.18 (m, 3), 2.19-2.15 (m, 2), 2.09 (s, 3), 0.99-0.91
(m, 12), 0.15-0.05 (m, 11); 13C NMR (100 MHz, CDCl3) δ 170.9,
170.0, 77.4, 63.6, 31.6, 30.0, 28.3, 22.7, 20.6, 18.7, 17.4, 17.3,
(2S )-2-[[N -(t er t -B u t o x y c a r b o n y l)-N -m e t h y l-L-v a l-
yl]oxy]-3-m et h ylb u t a n oic Acid 2-(Tr im et h ylsilyl)et h yl
(46) Gen er a l. Reagents and solvent were obtained from commercial
suppliers and were used as received, unless otherwise noted. Moisture-
or air-sensitive reactions were conducted under argon in distilled
solvents. The standard reaction workup involved drying the solution
of crude product over MgSO4, removing the solvent under reduced
pressure, and purifying the residue by flash chromatography47 with
the eluting solvent indicated. N-Boc-N-methyl-O-benzyl-L-tyrosine and
N-Boc-N-methyl-L-valine were prepared from the corresponding Boc-
protected amino acids with sodium hydride and methyl iodide in
analogy to the method of McDermott and Benoiton for the Cbz-
protected amino acids.19 Routine NMR spectra were obtained in CDCl3
unless otherwise noted. J values are given in Hz. The following
abbreviations are used: Boc ) N-(tert-butoxycarbonyl); DMAP ) N,N-
dimethyl-4-aminopyridine; EDC ) 1-[3-(dimethylamino)propyl]-3-eth-
ylcarbodiimide hydrochloride; TFA ) trifluoroacetic acid; TBAF )
tetrabutylammonium fluoride; DBU ) 1,8-diazabicyclo[5.4.0]undec-7-
ene; DIEA ) diisopropylethylamine.
(47) Still, W. C.; Kahn, M.; Mitra, A. J . J . Org. Chem. 1978, 43,
2923.

100 J . Org. Chem., Vol. 62, No. 1, 1997
Sefler et al.
14.1, -1.5, -1.8; IR (CHCl3) 3000, 2960, 2395, 1735, 1470,
1415, 1040, 925 cm-1; MS (FAB) m/ z 332 (MH+); HRMS (FAB)
m/ z calcd for C16H34NO4Si (MH+) 332.2252, found 332.2263.
(2S)-2-[[N-[(2S)-2-[[N-(ter t-Bu toxyca r bon yl)-N-m eth yl-
O-ben zyl-L-t yr osyl]oxy]-3-m et h ylb u t a n oyl]-N-m et h yl-L-
va lyl]oxy]-3-m et h ylb u t a n oic Acid 2-(Tr im et h ylsilyl)-
eth yl Ester , 7a . Compounds 3a (1.58 g, 3.3 mmol) and 4a
(809 mg, 2.2 mmol) were dissolved in 3 mL of freshly distilled
CH2Cl2, and PyBroP (1.54 g, 3.3 mmol) was added. The
mixture was placed under Ar and cooled to 0 °C, DIEA (1.54
mL, 8.8 mmol) was added dropwise with stirring, and the
mixture was stirred for 6 h as it slowly warmed to room
temperature. The solvent was evaporated, and the residue
was dissolved in 100 mL of EtOAc; this solution was washed
with 1 N HCl (2 × 100 mL), 1 N NaOH (2 × 100 mL), H2O (2
× 100 mL), and 100 mL of brine and worked up. The crude
product was chromatographed (30% EtOAc/hexane) to give
1.46 g (84%) of the tetramer 7a as a yellow oil: 1H NMR (400
MHz, CDCl3) δ 7.5-7.3 (m 5), 7.1 (m, 2), 6.9 (br d, 2), 5.03 (s,
2), 4.8 (m, 2), 4.2-4.3 (m, 4), 3.0-3.5 (m, 2), 2.6-3.0 (m, 6),
2.2-2.3 (m, 3), 2.1 (s, 9), 1.1-0.9 (m, 18), 0.05 (m, 11); 13C
NMR (100 MHz, CDCl3) δ 170.4, 169.2, 157.4, 154.4, 137.9,
130.3, 129.9, 128.5, 128.2, 127.8, 127.4, 114.8, 114.7, 80.4, 75.4,
70.0, 61.5, 33.8, 31.3, 29.8, 28.2, 27.9, 27.5, 20.7, 19.8, 19.5,
18.8, 17.7, 17.6, 17.0, -1.5, -1.8; IR (CHCl3) 2390, 1730, 1680,
1510, 1470, 1420, 1040, 925 cm-1; MS (FAB) m/ z 799 (MH+);
HRMS (FAB) m/ z calcd for C43H67N2O10Si (MH+) 799.4565,
found 799.4599.
(2S)-2-[[N-[(2S)-2-[[N-(ter t-Bu toxyca r bon yl)-N-m eth yl-
O-ben zyl-L-t yr osyl]oxy]-3-m et h ylb u t a n oyl]-N-m et h yl-L-
va lyl]oxy]-3-m eth ylbu ta n oic Acid , 7b. Cleavage of the
2-(trimethylsilyl)ethyl ester was carried out as described for
preparation of 3a ; the crude product was chromatographed (95:
5:0.1 CH2Cl2/MeOH/HOAc) to give a quantitative yield of the
tetrameric acid 7b: 1H NMR (400 MHz, CDCl3) δ 7.42-6.87
(m, 9), 5.05 (s, 2), 5.09-4.84 (m, 3), 4.69 (m, 1), 3.37-2.07 (m,
11), 1.47-0.83 (m, 27); 13C NMR (100 MHz, CDCl3) δ 173.4,
170.5, 157.5, 157.4, 137.0, 130.3, 129.9, 128.5, 128.2, 127.8,
127.3, 114.8, 114.7, 80.4, 75.7, 70.0, 61.5, 33.8, 31.3, 29.8, 28.2,
27.9, 27.5, 20.7, 19.8, 19.5, 18.8, 17.7, 17.6, 17.0; IR (CHCl3)
2395, 1735, 1675, 1505, 1480, 1390, 1365, 1020, 925 cm-1; MS
(FAB) m/ z 699 (MH+); HRMS (FAB) m/ z calcd for C38H55N2O10
(MH+) 699.3826, found 699.3887.
NMR (100 MHz, CD3CN) δ 176.0, 157.5, 138.5, 129.5, 128.8,
128.6, 70.6, 66.7, 41.2, 32.0, 26.4.
(S)-2-Hyd r oxy-5-[(ben zyloxyca r bon yl)a m in o]p en ta n o-
ic Acid 2-(Tr im eth ylsilyl)eth yl Ester , 9c. To a solution of
the acid 9b (5.1 g, 19 mmol) and 2-(trimethylsilyl)ethanol (13.6
mL, 5 equiv) in pyridine (25 mL, freshly distilled from CaH2)
at 0 °C were added EDC (4.4 g, 1.2 equiv) and DMAP (0.5 g,
0.2 equiv), and the resulting mixture was stirred at 0 °C to rt
for 50 h. Pyridine was evaporated under reduced pressure,
and the oily residue was dissolved in EtOAc (150 mL), washed
successively with saturated NH4Cl (125 mL), H2O (125 mL),
and saturated NaCl (100 mL), and worked up. The crude
product was chromatographed (0-40% EtOAc/hexanes gradi-
ent) to give 1.7 g (24%) of the ester 9c as a colorless oil: TLC
Rf 0.62 (1:1 EtOAc/hexanes); 1H NMR (500 MHz, CDCl3) δ 7.32
(m, 5), 5.06 (s, 2), 5.00 (s, 1), 4.24 (m, 2), 4.13 (s, 1), 3.21 (m,
2), 3.07 (s, 1), 1.82 (m, 1), 1.62 (m, 3), 1.00 (m, 2), 0.03 (s, 9);
13C NMR (125 MHz, CDCl3) δ 175.0, 156.4, 136.5, 128.4, 128.0,
70.0, 66.5, 64.1, 40.5, 31.2, 25.4, 17.3, -1.63; MS(FAB) m/ z
368.2 (MH+); HRMS(FAB) m/ z calcd for C18H30NO5Si (MH+)
368.1893, found 368.1887.
Nr-(ter t-Bu toxyca r bon yl)-NE-(ter t-bu tyld im eth ylsilyl)-
D-tr yp top h a n , 10b.42 A 1 M solution of LiHMDS (28.22 mL,
28.22 mmol) was added slowly to a solution of 1.53 g of TBSCl
(10.2 mmol) and 2.53 g of NR-Boc-D-tryptophan (8.3 mmol) in
42 mL of THF at -78 °C with stirring under Ar. After 24 h
at -78 °C, 25 mL of saturated aqueous NH4Cl was added, and
the mixture was warmed to 25 °C over 1 h and extracted with
EtOAc (3 × 50 mL). The combined organic layers were washed
with saturated NH4Cl (2 × 50 mL) and 50 mL of brine and
worked up. The crude product was chromatographed (90:10:
0.1 CH2Cl2/MeOH/HOAc) to give 3.04 g (88%) of the silyl
derivative 10b as a white solid: [R]20 -20.2° (c ) 21.3 mg/
D
mL, CHCl3) (lit.42 [R]20D-21.2°); mp 90-93 °C; H NMR (400
1
MHz, CDCl3) δ 7.58 (br s, 1), 7.44 (d, 1, J ) 7.7), 7.07 (m, 2),
7.01 (s, 1), 5.08 (br s, 1), 4.59 (br s, 1), 3.35 (m, 1), 3.23 (m, 1),
1.37 (br s, 9), 0.88 (s, 9), 0.55 (s, 6); 13C NMR (100 MHz, CDCl3)
δ 177.4, 155.5, 141.3, 131.0, 130.0, 121.5, 119.6, 118.7, 113.9,
112.3, 79.9, 76.7, 54.2, 28.3, 27.8, 27.7, 26.2, 25.6, 19.4; IR (CH2-
Cl2) 3450, 1720, 1460, 1170, 980, 850 cm-1; UV-vis (c ) 0.1
mM, MeOH) λmax 279.5 nm; MS (EI) m/ z 418 (M+), 344 (28.60),
244 (79.85), 188 (41.65), 130 (60.27); HRMS (EI) m/ z calcd
for C22H34N2O4Si 418.2288, found 418.2270.
(S)-2-Hyd r oxy-5-[(ben zyloxyca r bon yl)a m in o]p en ta n o-
ic Acid , 9b. L-Ornithine‚HCl (10 g, 60 mmol) was dissolved
in 10 mL of water and loaded on an Amberlite IRA-120 (H+)
column (80 mL of the resin), which was eluted first with water
(300 mL) and then with 3% NH4OH (400 mL). The solution
was lyophilized to give 9.2 g of an oil. The oil was dissolved
in 30 mL of water and cooled to 0 °C, and 10.5 mL of concd
H2SO4 (3 equiv) was added. A solution of 8.3 g of NaNO2 (2
equiv) in 30 mL of water was added dropwise over a period of
60 min. The resulting solution was left to stand as the
temperature rose from 0 °C to rt, and the reaction progress
was followed by 13C NMR. The reaction product, (S)-2-hy-
droxy-5-aminopentanoic acid, 9a ,40 was characterized by
NMR: 1H NMR (400 MHz, D2O) δ 3.87 (t, 1, J ) 6.2), 2.80 (q,
2, J ) 6.4), 1.4-1.6 (m, 4); 13C NMR (100 MHz, D2O) δ 183.7,
74.3, 42.2, 33.6, 25.8.
Nr-(ter t-Bu toxyca r bon yl)-NE-(ter t-bu tyld im eth ylsilyl)-
Nr-m eth yl-D-tr yp top h a n , 10c. Sodium hydride (0.27 g, 6.0
mmol) was added to a solution of 0.878 g of the silyl derivative
10b (2.1 mmol) and 1.05 mL of CH3I (17 mmol) in 8.0 mL of
THF at 0 °C under Ar, and the suspension was stirred and
allowed to warm to 25 °C. After 72 h, the mixture was
partitioned between 10 mL of EtOAc and 10 mL of water; the
aqueous layer was acidified to pH 2 with 3 N HCl and
extracted with EtOAc (4 × 25 mL). The combined organic
layers were washed with 25 mL of water, 5% aqueous Na2S2O3
(2 × 25 mL), 25 mL of water, and 25 mL of brine and worked
up. The crude product was chromatographed (95:5:0.1 CH2-
Cl2/MeOH/HOAc) to give 0.79 g (88% yield) of the N-methyl
derivative 10c as an off-white foam: [R]20 -39.55° (c ) 22.2
1
mg/mL, EtOAc); H NMR (400 MHz, acetone-d6) δ 7.64 (d, 1,
J ) 7.8), 7.52 (d, 1, J ) 7.8), 7.13-7.07 (m, 3), 4.92 (m, 1),
3.1-3.5 (m, 2), 2.65 and 2.8 (2s, 3), 1.36 and 1.10 (2s, 9), 0.92
(s, 9), 0.607 and 0.592 (2s, 6); 13C NMR (100 MHz, acetone-d6)
δ 173.2, 156.5, 155.6, 142.5, 142.3, 132.0, 131.8, 130.4, 129.8,
122.1, 120.2, 119.4, 114.7, 114.6, 79.8, 79.5, 60.1, 59.4, 31.4,
28.5, 28.1, 26.7, 26.6, 25.8, 25.2, 20.1, 20.0, -3.8, -3.9; IR (CH2-
Cl2) 3490, 1720, 1400, 1375, 1170, 792 cm-1; UV-vis (c ) 0.1
mM, MeOH) λmax 279.5 nm; MS (EI) m/z 432 (M+), 359 (14.44),
332 (14.97), 301 (22.41), 188 (36.38), 144 (71.23), 57 (99.39);
HRMS (EI) m/ z calcd for C23H36N2O4Si 432.2433, found
432.2453.
After 46 h, the reaction was complete, and excess NaNO2
was quenched by adding urea until the solution gave a
negative KI-starch test. The mixture was brought to pH 10
with 2 N NaOH and cooled in an ice bath; benzyloxycarbonyl
chloride (CbzCl, 10 mL, 1.1 equiv) and 2 N NaOH were added
alternately over a period of 2 h with vigorous stirring, keeping
the pH at about 10. The mixture was stirred at rt for another
2 h. The resulting suspension was washed with ether (3 ×
150 mL), and the clear aqueous layer was acidified with 6 N
HCl and extracted with EtOAc (4 × 50 mL), washed with
saturated NaCl (200 mL), and worked up to provide 9.2 g (68%)
of N-protected R-hydroxy acid 9b as a white solid after
chromatography on silica gel 60G (80% EtOAc/hexanes): mp
102-6 °C (lit.41 mp 107-108 °C); [R]D +1.5° (c 0.0098 g/mL,
MeOH); TLC: Rf 0.26 (CHCl3/MeOH/HOAc 9:1:0.5); 1H NMR
(400 MHz, CD3CN) δ 7.34 (m, 5), 5.70 (bs, 1), 5.03 (s, 2), 4.11
(S)-5-[(Ben zyloxyca r bon yl)a m in o]-2-[[Nr-(ter t-bu toxy-
ca r bon yl)-NE-(ter t-bu tyld im eth ylsilyl)-Nr-m eth yl-D-tr yp -
top h a n yl]oxy]p en ta n oic Acid 2-(Tr im eth ylsilyl)eth yl Es-
ter , 5a . To a solution of the hydroxy ester 9c (2.6 g, 7.1 mmol)
and the tryptophan derivative 10c (3.5 g, 8.1 mmol) in CH2-
Cl2 (25 mL) at 0 °C under Ar were added EDC (1.61 g, 1.2
equiv) and DMAP (180 mg, 0.2 equiv), and the mixture was
(t, 1, J ) 6.7), 3.11 (q, 2, J ) 6.4), 1.73 (m, 1), 1.56 (m, 3); 13
C

Synthesis of a Depsipeptide Mimic of Tendamistat
J . Org. Chem., Vol. 62, No. 1, 1997 101
stirred at 0 °C to rt for 24 h to give a clear yellow solution.
The solution was diluted with 20 mL of CH2Cl2, washed with
1 N HCl (2 × 20 mL), H2O (2 × 30 mL), and saturated NaCl
(30 mL), and worked up. The residue was chromatographed
(20% EtOAc/hexanes) to give 4.42 g (80%) of the dimeric ester
as a yellow foam: TLC Rf 0.78 (1:2 EtOAc/hexanes); 1H NMR
(400 MHz, CDCl3) δ 7.57 (d, 1, J ) 7.8), 7.45 (d, 1, J ) 7.8),
7.30 (m, 5), 7.12 (m, 2), 6.95 (s, 1), 5.21 (m, 1), 5.07 (s, 2), 4.99
(t, 1, J ) 5.4), 4.87 and 4.73 (s, 1), 4.22 (m, 2), 3.42 (m, 1),
3.18 (m, 3), 2.84 and 2.75 (s, 3), 1.85 (m, 2), 1.60 (m, 2), 1.38
and 1.15 (2s, 9), 1.01 (m, 2), 0.89 (s, 9), 0.56 (s, 3), 0.55 (s, 3),
0.04 (s, 9); 13C-NMR (100 MHz, CDCl3) δ 171.1, 169.5, 156.3,
141.4, 136.5, 130.5, 129.3, 128.8, 128.5, 128.0, 121.4, 119.5,
118.4, 113.8, 113.8, 113.2, 113.0, 79.9, 72.5, 66.6, 64.0, 60.3,
58.7, 57.9, 40.3, 31.2, 30.3, 28.3, 28.1, 27.9, 26.2, 25.6, 25.2,
24.3, 24.3, 21.0, 19.4, 17.4, 14.2, -1.6, -4.0; MS(FAB) m/ z
781.3 (M+); HRMS (FAB) m/ z calcd for C41H64N3O8Si2 (MH+)
782.4232, found 782.4221.
5-[(Ben zyloxyca r b on yl)a m in o]-(2S)-2-[[N-[(2S)-2-[[N-
[(2S)-2-[[N-(ter t-Bu t oxyca r b on yl)-N-m et h yl-O-ben zyl-L-
tyr osyl]oxy]-3-m eth ylbu ta n oyl]-N-m eth yl-L-va lyl]oxy]-3-
m e t h y l b u t a n o y l ]-N -m e t h y l -D -t r y p t o p h a n y l ]o x y ]-
p en ta n oic Acid 2-(Tr im eth ylsilyl)eth yl Ester , 11a . To a
solution of the ester 5a (1.3 g, 1.7 mmol) in THF (40 mL) under
Ar at rt was added TBAF on SiO2 (1.85 g, 1.1 mmol fluoride/
g, 5 equiv), and the mixture was stirred for 20 min. TLC (1:1
EtOAc/hexanes) showed that 5a (Rf 0.87) was cleanly con-
verted to a new compound (5, Rf 0.65). The reaction mixture
was poured onto saturated NH4Cl (100 mL) and extracted with
EtOAc (3 × 50 mL), and the organic layer was washed with
saturated NaCl (40 mL) and worked up. 1H NMR (400 MHz,
CDCl3) showed that only the TBDMS group had been removed.
This material was then dissolved in EtOAc saturated with
anhydrous HCl gas (15 mL) and left to stand at rt for 1 h.
Evaporation of the solvent afforded the N-deblocked ester 5b
as a yellow oil. To a solution of this material (ca. 1.7 mmol),
the tetradepsipeptide acid 7b (1.18 mg, 1 equiv), and PyBroP
(873 mg, 1 equiv) in CH2Cl2 (7.5 mL) at 0 °C under Ar was
added DIEA (326 µL, 1.1 equiv) dropwise, and the mixture
was allowed to stir as it warmed from 0 °C to rt for 20 h. The
reaction mixture was diluted with EtOAc (50 mL), washed
with 1 N HCl (3 × 50 mL), 1 N NaOH (3 × 50 mL), H2O (50
mL), and saturated NaCl (50 mL), and worked up. The residue
was chromatographed (40% EtOAc/hexanes) to give the coupled
product 11a as a white foam. The product was further purified
by normal phase MPLC on a Merck 40-63 µM silica column,
eluting with 40% EtOAc/hexanes at 20 mL/min and monitoring
the effluent by UV absorbance. The total yield of 11a was 554
mg (26%) of a yellow foam: TLC Rf 0.64 (1:1 EtOAc/hexanes);
1H NMR (300 MHz, CDCl3) δ 8.13 (s, 1), 7.62-7.02 (m, 16),
6.86 (d, 2), 5.73-5.52 (m, 2), 5.15-5.06 (m, 8), 5.05 (s, 2), 5.01
(s, 2), 4.71-4.64 (m, 1), 4.21 (m, 2), 3.75-3.71 (m, 1), 3.55-
3.45 (m, 1), 3.33-3.02 (m, 5), 3.01 (s, 3), 2.99 (s, 3), 2.69 and
2.58 (2s, 3), 2.21-2.14 (m, 1), 1.94-1.83 (m, 1), 1.65-1.55 (m,
2), 1.36 and 1.32 (2s, 9), 1.10-0.88 (m, 9), 0.80 (d, 3, J ) 6.24),
0.66 (bd, 3), 0.51 (d, 3, J ) 6.64), 0.03 (s, 9); MS(FAB) m/ z
1248.1 (MH+); HRMS (FAB) m/ z calcd for C68H94N5O15Si
(MH+) 1248.6516, found 1248.6524.
additional 5 min while the solution warmed to room temper-
ature. The solution was then concentrated with a rotary
evaporator without heating, and the residual solvent was
removed in vacuo for 30 min. The resultant clear glass was
lyophilized from benzene to yield compound 11c (103 mg,
100%) as a white powder: 1H NMR (400 MHz, CD3OD) δ 7.55
(d, 1, J ) 7.8), 7.26-7.40 (m, 12), 7.23 (d, 2, J ) 8.6), 7.10 (s,
1), 7.07 (t, 1, J ) 7.2), 6.99 (t, 1, J ) 7.1), 6.94 (d, 2, J ) 8.6),
5.55 (dd, 1, J ) 4.2, 11.4), 5.20 (d, 1, J ) 5.5), 5.06 (s, 2), 5.03
(s, 2), 5.02 (m, 2), 4.94 (m, 1), 3.37 (t, 1, J ) 5.9), 3.48 (dd, 1,
J ) 4.7, 15.3), 3.25 (m, 3), 3.17 (t, 2, J ) 6.8), 3.06 (s, 3), 3.03
(s, 3), 2.70 (s, 3), 2.17 (m, 2), 1.90 (m, 2), 1.64 (m, 3), 0.97 (d,
3, J ) 6.9), 0.95 (d, 3, J ) 6.7), 0.92 (d, 3, J ) 6.7), 0.77 (d, 3,
J ) 6.6), 0.63 (d, 3, J ) 6.9), 0.40 (d, 3, J ) 6.6); 13C NMR
(100 MHz, CD3OD) δ 172.9, 171.8, 171.6, 171.3, 171.1, 169.0,
160.0 (TFA), 158.8, 138.6, 138.4, 138.2, 131.9, 129.5, 129.5,
129.0, 128.9, 128.8, 128.5, 128.4, 128.4, 126.6, 124.5, 122.4,
119.8, 119.1, 116.6 (TFA), 112.4, 110.8, 78.5, 77.1, 74.2, 74.0,
70.9, 67.4, 63.4, 62.7, 59.0, 41.5, 35.4, 32.9, 32.7, 31.7, 30.9,
30.5, 29.5, 28.2, 26.4, 25.5, 20.1, 19.2, 19.2, 19.1, 17.7, 16.8;
MS (FAB) m/ z calcd for C58H74N5O13 (MH+) 1048.5283, found
1048.5290; 1070 (MNa+), 1048 (MH+), 958, 948, 914.
Cyclo[5-[(ben zyloxyca r bon yl)a m in o]-(2S)-2-[[N-[(2S)-
2-[[N-[(2S)-2-[(N-m eth yl-O-ben zyl-L-tyr osyl)oxy]-3-m eth -
ylbu ta n oyl]-N-m eth yl-L-va lyl]oxy]-3-m eth ylbu ta n oyl]-N-
m eth yl-D-tr yptoph an yl]oxy]pen tan oic acid], 1a. Precursor
11c (351 mg, 335 µmol) and PyBroP (315 mg, 670 µmol) were
dissolved in CH2Cl2 (300 mL), and the solution was cooled to
0 °C. DIEA (250 µL, 1.34 mmol) was added, and the ice bath
was allowed to warm to room temperature. After 4 days, the
solvent was removed under reduced pressure and the residue
was dissolved in EtOAc (100 mL) and washed with 1 N HCl
(5 × 50 mL), saturated NaHCO3 (2 × 50 mL), and brine (50
mL). The organic layer was worked up, and the residue was
chromatographed (90% ether:10% MeOH followed by 90%
CHCl3:10% MeOH (v/v)) to give 220 mg (64%) of the cyclic
product 1a as a white foam: 1H NMR (400 MHz, CDCl3) δ
8.12 (s, 1), 7.55 (d, 1, J ) 8.7), 7.26 (m, 11), 7.10 (m, 5), 6.84
(d, 2, J ) 8.5), 5.74 (m, 2), 5.34 (d, 1, J ) 6.3), 5.07 (m, 3), 4.97
(s, 2), 4.82 (m, 2), 3.87 (d, 1, J ) 10.8), 3.42 (m, 2), 3.31 (m, 1),
3.25 (s, 3), 3.10 (s, 3), 2.87 (m, 2), 2.78 (s, 3), 2.30 (m, 2), 2.04
(m, 2), 1.85 (m, 1), 1.68 (m, 2), 1.01 (m, 9), 0.88 (m, 3), 0.79 (d,
3, J ) 6.6), 0.62 (d, 3, J ) 6.7); 13C NMR (100 MHz, CDCl3) δ
170.9, 170.3, 169.4, 168.7, 167.8, 157.3, 156.4, 136.9, 136.6,
136.2, 130.0, 129.9, 129.2, 128.5, 128.1, 127.8, 127.4, 124.7,
127.0, 122.7, 122.0, 119.5, 118.2, 115.3, 114.6, 111.3, 109.8,
76.7, 75.5, 74.2, 70.0, 67.1, 66.5, 55.0, 40.3, 32.8, 31.9, 31.7,
31.1, 30.4, 30.0, 29.6, 29.3, 28.2, 27.1, 26.2, 26.1, 25.7, 25.2,
20.6, 19.7, 19.5, 19.3, 18.0, 17.9, 17.5, 17.3, 14.1; MS (FAB)
m/ z calcd for MNa+ C58H71N5O12Na 1052.4997, found
1052.4997; 1068 (MK+), 1052 (MNa+), 962, 918, 616, 537, 515,
444.
Cyclo[5-a m in o-(2S)-2-[[N-[(2S)-2-[[N-[(2S)-2-[(N-m eth -
yl-L-t yr osyl)oxy]-3-m et h ylb u t a n oyl]-N-m et h yl-L-va lyl]-
oxy]-3-m eth ylbu ta n oyl]-N-m eth yl-D-tr yp top h a n yl]oxy]-
p en ta n oic a cid ], 1b. Compound 1a (9.0 mg, 9.0 µmol) and
Pd(OH)2 (Degussa 20 wt % Pd, 11.8 mg, 22.0 µmol) were
suspended in MeOH (6 mL) and HOAc (63 µL, 111 µmol). After
being purged thoroughly with nitrogen, the mixture was
stirred vigorously under an H2 atmosphere for 40 min. The
reaction mixture was filtered through a 0.45-µm filter, the
filtrate was concentrated on a rotary evaporator, and the
remaining solvent was removed in vacuo to afford 6.1 mg (88%)
of compound 1b: 1H NMR (400 MHz, CD3OD) δ 7.52 (m, 1),
7.31 (m, 1), 7.07 (m, 5), 6.71 (m, 2), 5.70 (m, 1), 5.30 (d, 1, J )
8.6), 5.24 (m, 1), 5.06 (m, 1), 4.85 (m, 1), 3.96 (d, 0.7, J ) 10.8),
3.89 (d, 0.3, J ) 10.8), 3.32-3.38 (m, 2), 3.22 (s, 3), 3.06-3.17
(m, 1), 3.05 (s, 3), 2.90 (s, 3), 2.80 (m, 2), 2.06-2.33 (m, 2),
1.45-1.79 (m, 2), 1.28 (m, 3), 0.57-1.18 (m, 12), 0.46 (d, 2, J
) 6.8), 0.37 (d, 1, J ) 6.8); MS (FAB) m/ z calcd for C43H60N5O10
(MH+) 806.4340, found 806.4349; 844 (MK+), 828 (MNa+), 806
(MH+), 328, 283, 237.
5-[(Ben zyloxyca r b on yl)a m in o]-(2S)-2-[[N-[(2S)-2-[[N-
[(2S )-2-[(N -m e t h y l-O -b e n zy l-L-t y r o s y l)o x y ]-3-m e t h -
ylbu ta n oyl]-N-m eth yl-L-va lyl]oxy]-3-m eth ylbu ta n oyl]-N-
m eth yl-D-tr yp top h a n yl]oxy]p en ta n oic Acid , 11c. Com-
pound 11a (124 mg, 0.099 mmol) was dissolved in DMF (1 mL),
and TBAF (1.0 M in THF, 0.25 mL, 0.25 mmol) was added.
After 15 min, the mixture was diluted with EtOAc (80 mL),
and the organic layer was washed with 0.5 N HCl (3 × 40 mL),
H2O (40 mL), and saturated NaCl (40 mL) and worked up to
give 114 mg (100%) of the free acid 11b as a white foam. 1H-
NMR indicated the complete removal of the 2-(trimethylsilyl)-
ethyl ester, and the material was used without further
purification.
A solution of the free acid 11b in EtOAc (15 mL) was purged
with Ar via a glass pipet and cooled to -40 °C (dry ice/
acetonitrile bath), and dry HCl was bubbled vigorously through
it via a glass pipet for 10 min. The solution was then purged
with Ar for 15 min at -40 °C to remove excess HCl and for an
Cyclo[5-gu a n id in o-(2S)-2-[[N-[(2S)-2-[[N-[(2S)-2-[(N-
m e t h yl-L-t yr osyl)oxy]-3-m e t h ylb u t a n oyl]-N -m e t h yl-L-
va lyl]oxy]-3-m eth ylbu ta n oyl]-N-m eth yl-D-tr yp top h a n yl]-
oxy]p en ta n oic a cid ], 1. The deprotected intermediate 1a

102 J . Org. Chem., Vol. 62, No. 1, 1997
Sefler et al.
(7.4 mg, 9.2 µmol) and aminoiminomethanesulfonic acid48,49
(2.3 mg, 18.4 µmol) were dissolved in MeOH (6 mL), and Et3N
(6.4 µL, 46 µmol) was added. After being stirred for 7 h, the
solution was concentrated with a rotary evaporator and the
residue was dissolved in MeOH. The material was purified
over a preparative Vydac Proteins and Peptides (25 cm × 25
mm) 10 µm C18 reversed phase column with a solvent gradient
from 30% to 70% of 10% H2O/90% acetonitrile/0.1% TFA (v/v)
(solvent B) and 90% H2O/10% acetonitrile/0.1% TFA (v/v)
(solvent A); detection at 214 nm revealed that the product
eluted at ∼45% B:A. The appropriate fractions were pooled
and lyophilized to afford 5.7 mg (73%) of the target compound
1 as a white flocculent powder, which was insoluble in H2O,
partially soluble in 5-10% DMSO/H2O, and completely soluble
in 50% DMSO/H2O or DMSO: 1H NMR (400 MHz, CD3OD) δ
7.52 (m, 1), 7.30 (m, 1), 7.08 (m, 5), 6.74 (d, 0.5, J ) 8.5), 6.80
(d, 1.5, J ) 8.5), 5.70 (dd, 1, J ) 6.0, 10.8), 5.28 (d, 0.5, J )
8.7), 5.23 (d, 1.5, J ) 6.5), 4.94 (m, 2), 3.97 (d, 0.7, J ) 10.8),
3.89 (d, 0.3, J ) 10.8), 3.33 (m, 2), 3.26 (s, 3), 3.19 (m, 1), 3.05
(s, 3), 2.91 (s, 3), 2.83 (m, 2), 2.34 (m, 1), 2.08 (m, 1), 1.68 (m,
1), 1.60 (m, 1), 1.50 (m, 1), 1.28 (m, 2), 1.00 (m, 13), 0.67 (d, 2,
J ) 6.7), 0.46 (d, 2, J ) 6.9), 0.34 (d, 1, J ) 6.8); MS (FAB)
m/ z calcd for C44H62N7O10 (MH+) 848.4558, found 848.4556;
886 (MK+), 870 (MNa+), 848 (MH+).
NMR Gen er a l. Approximately 2 mg of cyclic hexamer 1
was dissolved in 300 µL of KD2PO4/NaOD (0.05 M/0.029 M)
in D2O, with 5 mM CaCl2, 50 mM NaCl, and 200 µL of CD3-
OD. The pH was adjusted to 6.98 with DCl. 1D 1H-NMR
spectra were recorded at 27 and 0 °C using presaturation to
suppress the water signal; 2D spectra were recorded at 0 °C.
Spectra were processed on a Bruker X-32 data station running
UxNMR. Time proportional phase incrementation (TPPI)50
was used to obtain phase-sensitive spectra. T1’s were mea-
sured using an inversion-recovery experiment.
TOCSY (500 MHz). Pulse sequence according to Bax et
al. with MLEV-17 spin-lock51 and DANTE presaturation52
during the relaxation delay, 512 experiments of 32 scans each,
relaxation delay of 1.5 s, mixing time of 50 ms, acquisition
time for one scan 0.184 s, size 2 K, spectral width in F2 and
F1 11.1 ppm, zero filling in F2 and F1 to 2 K, apodization in
both dimensions with squared sinebell shifted by π/2.
NOESY (300 MHz). Pulse sequence according to Boden-
hausen et al.,53 with presaturation during the relaxation delay
and the mixing time, 1.5 s presaturation during the relaxation
delay, 700 ms of presaturation during the mixing time,
acquisition time of 0.67 s, size 2 K, spectral width in F2 and
F1 11.1 ppm, apodization in both dimensions with squared
sinebell shifted by π/2. Interresidue NOEs observed are
depicted in Figure 3.
Molecu la r Mod elin g. All modeling was done using Mac-
roModel/Batchmin software version 4.0.54 Monte Carlo searches
(10,000 iterations) and minimizations were run without con-
straints. Four searches were performed with each of the three
amide bonds cis and all of them trans. Energy minimizations
were carried out with MacroModel’s TNCG method, using the
Amber* force field55 with GB/SA water solvation,56 and the
calculations were carried out until gradient convergence was
achieved. The structures were sorted into families with a
clustering algorithm57 by comparing the peptide backbone
atoms (CR, amide N or ester O, amide carbonyl carbon). The
families of the two major amide configurations were then
filtered against the NOE data (see Figure 3) by requiring
that two hydrogens showing an NOE cross peak be closer than
5.0 Å.
En zym ology. Due to the low solubility of compounds 1 and
2 in water, assays were performed in 5% DMSO/water. The
kinetic constants for hydrolysis of p-nitrophenyl R-D-malto-
trioside by porcine pancreatic R-amylase (PPA) were deter-
mined in water and in 5% DMSO/water as described previ-
ously.27 Km and Vmax are slightly higher in 5% DMSO/water
(Km ) 1.14 mM, Vmax ) 0.072 min-1) than in water (Km ) 1.05
mM, Vmax ) 0.067 min-1); however, Ki values were found to
differ significantly. As a result, Ki values for the control
compounds WRY (870 ( 30 µM) and AcN-WRY-OMe (190 (
10 µM) were redetermined in 5% DMSO/water. Dixon plots
for determination of the Ki values are provided in the Sup-
porting Information.
Ack n ow led gm en t. This work was supported by a
grant from the National Institutes of Health (GM30759);
A.M.S. received support from NIH Training Grant
GM08295 and M.C.K. from the Department of Educa-
tion and a Syntex fellowship. We also thank Prof. R.
Huber and Prof. W. Bode (Max Planck Institut fu¨r
Biochemie, Martinsried) for providing the coordinates
of tendamistat and the tendamistat-amylase complex
and Dr. Peter J ohann of Bo¨hringer-Mannheim for
providing a sample of p-nitrophenyl maltotrioside.
Su p p or tin g In for m a tion Ava ila ble: Experimental pro-
cedures for synthesis of the linear control 2; Dixon plots for
1
Ki determinations; H NMR spectra of compounds described
above (20 pages). This material is contained in libraries on
microfiche, immediately follows this article in the microfilm
version of the journal, and can be ordered from the ACS; see
any current masthead page for ordering information.
J O9616062
(48) Kim, K.; Lin, Y. T.; Mosher, H. S. Tetrahedron Lett. 1988, 29,
3183.
(49) Miller, A. E.; Bischoff, J . J . Synthesis 1986, 777.
(50) Rance, M.; Sorenson, W. W.; Bodenhausen, G.; Wagner, G.;
Ernst, R. R.; Wu¨thrich, K. Biochem. Biophys. Res. Commun. 1983, 117,
479.
(51) Bax, A.; Davis, D. G. J . Magn. Reson. 1985, 65, 355.
(52) Haasnoot, C. A. G. J . Magn. Reson. 1983, 53, 153.
(53) Bodenhausen, G.; Kogler, K.; Ernst, R. R. J . Magn. Reson. 1984,
58, 370.
(54) Mohamadi, F.; Richards, N. G. J .; Guida, W. C.; Liskamp, R.;
Caufield, C.; Chang, G.; Hendrickson, T.; Still, W. C. J . Comput. Chem.
1990, 11, 440.
(55) Ferguson, D. M.; Kollman, P. A. J . Comput. Chem. 1991, 12,
620.
(56) Still, W. C.; Tempczyk, A.; Hawly, R. C.; Hendrickson, T. J .
Am. Chem. Soc. 1990, 112, 6127.
(57) Lauri, G. A.; Bartlett, P. A. Unpublished results.