Synthesis, conformational analysis, and cytotoxicity of conformationally constrained aplidine and tamandarin A analogues incorporating a spirolactam β-turn mimetic

Article abstract of DOI:10.1021/jm040788m

With the aim of studying the contribution of the β II turn conformation at the side chain of didemnins to the bioactive conformation responsible for their antitumoral activity, conformationally restricted analogues of aplidine and tamandarin A, where the

Full text of DOI:10.1021/jm040788m

5700
J. Med. Chem. 2004, 47, 5700-5712
Synthesis, Conformational Analysis, and Cytotoxicity of Conformationally
Constrained Aplidine and Tamandarin A Analogues Incorporating a
Spirolactam â-Turn Mimetic
Marta Gutie´rrez-Rodr´ıguez,§ Mercedes Mart´ın-Mart´ınez,§ M. Teresa Garc´ıa-Lo´pez,§ Rosario Herranz,*,§
Fe´lix Cuevas,‡ Concepcio´n Polanco,‡ Ignacio Rodr´ıguez-Campos,‡ Ignacio Manzanares,‡ Francisco Ca´rdenas,#
Miguel Feliz,# Paul Lloyd-Williams,*,† and Ernest Giralt†,|
Instituto de Quı´mica Me´dica (CSIC), Juan de la Cierva 3, E-28006 Madrid, Spain, Pharma Mar, S.A.,
Polı´gono Industrial La Mina, Avda. de los Reyes 1, E-28770 Colmenar Viejo, Madrid, Spain, Unidad de RMN de Alto Campo,
SCT, Universidad de Barcelona, Josep Samitier 1, E-08028 Barcelona, Spain, Departamento de Qu´ımica Orga´nica,
Universidad de Barcelona, Martı´ i Franque`s 1-11, E-08028 Barcelona, Spain, and IRBB-Parque Cient´ıfico de Barcelona,
Josep Samitier 1, E-08028 Barcelona, Spain
Received February 12, 2004
With the aim of studying the contribution of the â II turn conformation at the side chain of
didemnins to the bioactive conformation responsible for their antitumoral activity, conforma-
tionally restricted analogues of aplidine and tamandarin A, where the side chain dipeptide
Pro8-N-Me-D-Leu7 is replaced with the spirolactam â II turn mimetic (5R)-7-[(1R)-1-carbonyl-
3-methylbutyl]-6-oxo-1,7-diazaspiro[4.4]nonane, were prepared. Additionally, restricted ana-
logues, where the aplidine (pyruvyl9) or tamandarin A [(S)-Lac9] acyl groups are replaced with
the isobutyryl, Boc, and 2-methylacryloyl groups, were also prepared. These structural
modifications were detrimental to cytotoxic activity, leading to a decrease of 1-2 orders of
magnitude with respect to that exhibited by aplidine and tamandarin A. The conformational
analysis of one of these spirolactam aplidine analogues, by NMR and molecular modeling
methods, showed that the conformational restriction caused by the spirolactam does not produce
significant changes in the overall conformation of aplidine, apart from preferentially stabilizing
the trans rotamer at the pyruvyl9-spirolactam amide bond, whereas in aplidine both cis and
trans rotamers at the pyruvyl9-Pro8 amide bond are more or less equally stabilized. These
results seem to indicate a preference for the cis form at that amide bond in the bioactive
conformation of aplidine. The significant influence of this cis/trans isomerism upon the
cytotoxicity suggests a possible participation of a peptidylprolyl cis/trans isomerase in the
mechanism of action of aplidine.
Introduction
The didemnins are a family of macrocyclic depsipep-
tides, isolated from several tunicates, which exhibit a
wide variety of biological activities, including antitu-
mor,1,2 antiviral,2 and inmunosupresive properties.2,3
These depsipeptides share a common 23-membered
macrocycle made up of six subunits [(3S,4R,5S)-isosta-
tine (Ist1), (2S,4S)-3-oxo-4-hydroxy-2,5-dimethylhexano-
ic acid (Hip2), Leu3, Pro4, [N,O-(Me)2-Tyr5], and Thr6]
and differ in the side chain attached to the threonine
NH, through a N-Me-D-leucine residue. Didemnin B
(Figure 1, 1), isolated in 1981 from the Caribbean
tunicate Trididemnum solidum,4 and aplidine (2, de-
hydrodidemnin B), isolated in 1990 from the Mediter-
ranean tunicate Aplidium albicans,5 stand out because
of their potent antitumoral activity. Didemnin B was
the first marine natural product to enter human anti-
cancer trials, but these had to be discontinued during
Figure 1. Most relevant didemnins and tamandarin A.
phase II tests because of its neuromuscular and car-
diotoxicity.1,2 Aplidine showed higher antitumoral po-
tencies and lower toxicity in preclinical and phase I
clinical studies.1,2,6-9 This better therapeutic profile has
facilitated its recent entry into phase II clinical trials.
Although the mechanism of action of aplidine is not
completely elucidated, it is known that it exerts anti-
* To whom correspondence should be addressed. For R.H.: phone,
+34 91 5622900; fax, +34 91 5644853; e-mail, rosario@iqm.csic.es. For
P.L.-W.: phone, +34 93 4021260; fax, +34 93 3397878; e-mail,
lloyd@qo.ub.es.
§ Instituto de Qu´ımica Me´dica (CSIC).
‡ Pharma Mar, S.A.
# Unidad de RMN de Alto Campo, SCT, Universidad de Barcelona.
† Departamento de Qu´ımica Orga´nica, Universidad de Barcelona.
| IRBB-Parque Cient´ıfico de Barcelona.
10.1021/jm040788m CCC: $27.50 © 2004 American Chemical Society
Published on Web 09/30/2004

Aplidine and Tamandarin A Analogues
Journal of Medicinal Chemistry, 2004, Vol. 47, No. 23 5701
proliferative effects through the inhibition of protein
synthesis and ornithine decarboxylase activity10-12 and
by induction of arrest and blockade of the G1 and G2
cell cycle phases, respectively.13 Additionally, aplidine
inhibits vascular endothelial growth factor (VEGF)
secretion and decreases the expression of its receptor
VEGFR-1 in human leukemia cells MOLT-4.14 Also, in
leukemic cells, aplidine is an extremely potent and rapid
apoptotic inductor via mitochondrial-mediated apoptotic
signaling routes15 and through activation of several
intracellular signaling kinases.16
Structure-activity relationship studies on didemnins
have shown the importance of the integrity of the
macrocyclic skeleton.2,17-19 However, two analogues,
tamandarins A (Figure 1, 3) and B, have recently been
isolated from a Brazilian ascidian of the family Didem-
nidae,20 which were reported to be somewhat more
potent in vitro than didemnin B.20,21 These analogues
have a hydroxyisovaleric acid (Hiv2) unit in the macro-
cyclic core instead of the hydroxyisovalerylpropionic acid
(Hip2) unit of the didemnins. With regard to the side
chain, it is difficult to establish precise structure-
activity relationships from the described data, but it is
important for activity, since the macrocycle hydrochlo-
ride was completely inactive22 and simple N-acyl deriva-
tives at the threonine NH retained very little antitumor
activity.19 The (N-Me-D-Leu7) residue (both the N-
methyl group and the D configuration must be present)
is essential for antitumoral activity.19,23
Figure 2. Aplidine side chain preferred conformations.
despite the localized differences between these two
conformers, both present similar hydrogen-bonding pat-
terns, and their overall conformations are alike and
broadly similar to that of didemnin B.
The fact that biological activities of the didemnins are
strongly dependent on the structure of the linear side
chain, along with its preference for adopting a â II turn
conformation, has led to speculation that this confor-
mational feature may be a determining element of
biological activity. Thus, some authors have suggested
that the side chain does not directly participate in the
binding of the didemnins to their receptors but that it
may stabilize the bioactive conformation of the macro-
cycle through the formation of a â turn,29 while others
have suggested that the presence of this turn is not
essential for activity.2,30 With the aim of shedding light
on this issue and contributing to the study of the
bioactive conformation of the didemnins, as a first step
toward the design of aplidine peptidomimetics, we
synthesized aplidine and tamandarin A analogues 4a
and 5a (Figure 3), where the Pro8-N-Me-D-Leu7 dipep-
tide of the side chain was replaced by a [4.4]spirolactam
â II turn mimetic. This replacement fixes the dipeptide
conformation with an additional methylene bridge and
allows the essential side chain and configuration of the
(D-Leu7) residue at the i + 2 position of the â turn to be
retained. Furthermore, according to the described data
for this type of â turn mimetic, it would fix the dipeptide
backbone φ8, ψ8, φ7, and ψ7 torsion angles into values
(-50.9°, 128.7°, 91.1°, and -5.4°) very close to those of
an ideal type II â turn31 and also near those observed
in the crystal structure of didemnin B24 and those
calculated for aplidine.27 On the other hand, because
the replacement of the [(S)-Lac9] group of didemnin B
by short-chain aliphatic acyl groups did not significantly
affect its antitumor activity,18,30 we also prepared
analogues 4b,c and 5b-d, where the (pyruvyl9) moiety
(a) was replaced by the isobutyryl (b), Boc (c), and
2-methylacryloyl (d) groups. The cytotoxicities of these
didemnin analogues were determined in human lung
(A549) and colon (HT-29) carcinoma cell lines. Finally,
with the aim of establishing conformation-cytotoxicity
relationships, the effect of the local conformational
constraint on the global conformation was studied in the
aplidine analogue 4a, in comparison with aplidine and
didemnin B, by using NMR and molecular modeling
methods. This conformational analysis required the
parallel analysis of the spirolactam derivative 6a (Fig-
ure 3) as a simple model of the side chain dipeptide.
X-ray diffraction analysis of the didemnin B (1) crystal
structure showed that its macrocycle was twisted and
puckered with a shape that was described as a “bent
figure eight”.24 This conformation was stabilized by
three intramolecular hydrogen bonds: a strong hydro-
gen bond between the (Leu3) carbonyl and the (Ist1)
amino group, forcing the adoption of the figure eight
shape, which is stabilized by another hydrogen bond
between the (N-Me-D-Leu7) carbonyl and the (Leu3)
amino group, and the third bond from the lactic (Lac9)
carbonyl to the (Thr6) amino group. This third hydrogen
bond fixes the linear side chain folded back toward the
macrocycle into a â II turn conformation. Two previous
conformational studies on didemnin B, using computa-
tional methods and solution NMR in deuterated DMSO-
25
d6 and CDCl3,26 showed the presence of a preferred
conformation, which overall was quite similar to the
crystal structure previously determined by X-ray crys-
tallography. The conformational comparison of taman-
darin A (3) with didemnin B, using various NMR
methods, showed that in fact the conformations of both
congeners in solution are strikingly similar and have
the same hydrogen bond network.20 On the other hand,
recent conformational studies on aplidine (2), also in
comparison with didemnin B, have shown the presence
of two slowly exchanging preferred conformers due to
the restricted rotation about the amide bond between
the (pyruvyl9) and (Pro8) side chain residues.27,28 These
conformations correspond to the cis and trans isomers
at that amide bond, and both are stabilized by a
hydrogen bond between the (Thr6) NH and the pyru-
vylcarboxamide CdO in the trans isomer or the (pyru-
vyl9) keto group in the cis isomer (Figure 2). In CDCl3
the cis/trans ratio was (∼1:1),27 while in DMSO there
was a (6:4) preference for the cis conformer.28 However,
Results and Discussion
Synthesis. The preparation of the aplidine and
tamandarin analogues 4 and 5 required the coupling of
their respective macrocycles to the appropriate [4.4]-

5702 Journal of Medicinal Chemistry, 2004, Vol. 47, No. 23
Gutie´rrez-Rodrı´guez et al.
Scheme 1a
a Reagents: (a) CH3COCO2H, HATU, HOAt, NMM, CH2Cl2; (b)
CH3COCOCl, TEA, DMAP, CH2Cl2; (c) (CH3)2CHCOCl, TEA,
DMAP, CH2Cl2; (d) Boc2O, (CH3)4NOH‚H2O, CH3CN; (e) CH2dC-
(CH3)-COCl, TEA, DMAP, CH2Cl2; (f) H2, Pd(C), MeOH; (g)
MeNH2‚HCl, BOP, NMM, CH2Cl2; (h) 1 N NaOH, MeOH.
the trifluoroacetate 7 with di(tert-butyl) dicarbonate in
dry acetonitrile, using tetramethylammonium hydroxide
as base,35 led directly to the N-Boc-acylated and C-
deprotected pseudodipeptide 9c.
Figure 3. Spirolactam analogues of aplidine (4), tamandarin
A (5), and the dipeptide pyruvyl-Pro-N-Me-D-Leu-NHMe (6a).
The 1H and 13C NMR spectra of the N-pyruvyl
derivatives 8a and 9a and the N-Boc derivative 9c in
CDCl3 showed the presence of two conformations at the
carboxamide bond in a 2:1 ratio. As in the case of
aplidine,27 in the pyruvic acid derivative 8a these slowly
exchanging conformational isomers were also observed
by TLC and HPLC. The assignment of these conforma-
tions was based on a comparison of the chemical shifts
for the C2 and C5 carbons of both conformations in each
compound, as is usually applied for Xaa-proline amide
bond isomers.36-38 Thus, in the N-pyruvyl derivatives
8a and 9a the major conformer was assigned as trans
because its C2 signal appeared downfield (48.91 ppm for
8a and 49.30 ppm for 9a) from that of the minor isomer
(47.72 and 48.03 ppm, respectively) while its C5 signal
appeared upfield (66.82 and 69.11 ppm) to that of the
cis isomer (67.82 and 69.11 ppm). However, by applica-
tion of the same criteria, in the N-Boc derivative 9c the
major conformation was assigned as cis. The high
percentage of the cis form in this derivative (66%) is in
agreement with the known preference of urethane
N-protected amino acids to adopt this conformation, as
exemplified by Boc-proline.39 In the isobutyryl deriva-
tives 8b and 9b and the 2-methylacryloyl derivatives
8d and 9d, only one preferred conformation was ob-
served, which was assigned as trans for the last two
compounds, based on the NOE effects observed in their
1D NOESY spectra between the (acryloyl) CH2 protons
and the 2-H spirolactam protons. This experiment could
not be used for determining the conformation of the
isobutyryl-Pro amide bond in compounds 8b and 9b
because the signals corresponding to their 2-H protons
overlapped with those of their 8-H protons. However,
spirolactam pseudodipeptide derivatives 9a-d, which
were obtained from the spirolactam analogue of H-Pro-
N-Me-D-Leu-OBn 732 by acylation and subsequent C
deprotection, as shown in Scheme 1. Attempts to
introduce the pyruvyl group as described for the syn-
thesis of aplidine33 by coupling the trifluoroacetate 7
with pyruvic acid using DCC/HOBt as activating agents
were unsuccessful. This coupling required the use of
HOAt/HATU as activating reagents and DIEA as base,
or alternatively, it was also carried out using pyruvyl
chloride as acylating agent, which had been freshly
prepared from pyruvic acid and dichloromethyl methyl
ether.34 In both cases the yield of the N-pyruvyl deriva-
tive 8a was not higher than 50%. Catalytic hydro-
genolysis of this compound in the presence of 10% Pd(C)
led quantitatively to the corresponding C-deprotected
spiropseudodipeptide 9a. Activation of this acid with
BOP, followed by treatment with methylamine hydro-
chloride, led to the N-methylcarboxamide 6a. The
isoburyryl and 2-methylacryloyl derivatives 8b and 8d
were obtained by reaction of trifluoroacetate 7 with the
isobutyryl and 2-methylacryloyl chlorides, respectively,
in the presence of TEA and catalytic amounts of
4-(dimethylamino)pyridine (DMAP). The C deprotection
of the isobutyryl derivative 8b was also carried out by
catalytic hydrogenolysis. However, in the case of the
2-methylacryloyl derivative 8d, hydrogenation, even
using a Lindlar catalyst, led to the simultaneous reduc-
tion of the double bond and C deprotection, furnishing
9b. Consequently, in this case the C deprotection was
carried out by saponification to give 9d. The reaction of

Aplidine and Tamandarin A Analogues
Journal of Medicinal Chemistry, 2004, Vol. 47, No. 23 5703
Scheme 2a
Figure 4. NH stretch region FT-IR data for a 1.4 mM sample
of the spiropseudodipeptide 6a in CHCl3.
a Reagents: HATU, HOAt, NMM, CH2Cl2/DMF.
dipeptide 6a showed the presence of two preferred
conformations in variable ratios depending on the
solvent (1:20 in CDCl3, 1:8 in acetone-d6, and 1:5 in
DMSO-d6). A trans conformation at the pyruvyl amide
bond was assigned to the major conformer, based on the
chemical shift comparison of both isomers.36-38,41 Thus,
as shown in Table 5, the 2-H protons (which correspond
to the δ protons of Pro) of the minor conformer appeared
shielded (3.52 and 3.56 ppm) with respect to those of
the major isomer (3.68 and 3.82 ppm). Similarly, the
C2 signal of the minor conformer appeared shielded
(49.01 ppm) with respect to that of the major one (50.21
ppm), while C5 appeared more shielded in the major
conformer (69.85 with respect to 70.21 ppm). To deter-
mine the possible existence of an intramolecular hydro-
gen bond between the methyl amide NH and one of the
carbonyls, we studied the temperature dependence of
that NH chemical shift between 35 and 55 °C in CDCl3
and DMSO-d6. In both solvents the ∆δ/∆T value of the
major conformation was 4.4 ppb/K. This value did not
conclusively establish the existence of a hydrogen bond
and suggests that the NH is probably involved in an
equilibrium between a hydrogen-bonded and a non-
hydrogen-bonded state.45,46 However, the low-field chemi-
cal shift of this NH in CDCl3 (7.34 ppm), in addition to
its small downfield shift (∆δ ) 0.26 ppm) when CDCl3
was replaced by DMSO-d6, indicated a high degree of
intramolecular hydrogen bonding.45-47 Moreover, as
shown in Figure 4 in the FT-IR spectra of 6a at 1.4, 5,
and 14 mM concentrations in CHCl3, the NH stretching
vibrations appeared as a strong broad band at 3366
cm-1 corresponding to a hydrogen-bonded amide proton
and as a very weak shoulder above 3400 cm-1 corre-
sponding to a small proportion of non-hydrogen-bonded
amide protons. The shape and frequency of these NH
stretching bands were concentration-independent, in-
dicating the absence of aggregation and that the hy-
drogen bond was intramolecular.46,48-51 Taking into
account the preference for a trans conformation of the
pyruvyl amide bond, two of the three carbonyls could
participate as acceptors in the intramolecular hydrogen
bond: either the pyruvic carboxamide group, forming a
â-turn conformation, or the spirolactam carbonyl group,
forming a γ-turn conformation.
it was tentatively assigned as trans on the basis of the
fact that R-alkyl-substituted prolines are known to
induce the trans form of the Xaa-Pro bond (Xaa ) acyl,
aminoacyl) preferentially because of unfavorable steric
interactions between the alkyl groups and the R proton
of Xaa in the cis form.36,40 Thus, N-acetyl-2-methylpro-
line-N′-methylamide has been shown to exist only in the
trans conformation.36
As shown in Scheme 2, the HATU/HOAt promoted
coupling of the side chain pseudodipeptides 9a-d to the
macrocyclic depsipeptides 1033 and 11,21 freshly pre-
pared from their N-Boc-protected derivatives by treat-
ment with a 4.6 M solution of HCl in EtOAc, led to the
desired aplidine and tamandarin A analogues 4a-c and
5a-d, respectively. Unlike aplidine, the RP-HPLC and
NMR analysis of these compounds did not show the
presence of cis/trans isomers.
Conformational Analysis of the Spirolactam
Aplidine Analogue 4a. Determination of the crystal
structure of the aplidine analogue 4a by X-ray diffrac-
tion analysis was not possible because suitable crystals
could not be obtained. Unlike aplidine, the 1H NMR
spectra of compound 4a in CDCl3 or DMSO-d6 solution
showed the presence of only one preferred conformation.
In Xaa-Pro-containing peptides the assignment of the
cis or trans conformation at the peptide bond is based
either on the NOE effects observed between the Xaa
residue R-H and the Pro R-H in cis conformers, or δ-H
in trans isomers,38,41-44 or on the different anisotropic
effects of the Xaa amide CdO group on the Pro carbons
and protons at the R and δ positions in both isomers,
which affect their respective chemical shifts.36-38,41 None
of these criteria could be applied to the assignment of
cis/trans conformation at the (pyruvyl9)-spirolactam
amide bond in the aplidine analogue 4a. The first is due
to the absence of R-H in the pyruvyl9 residue, and the
second is because it requires the assignment of the
chemical shifts of both isomers for their comparison.
Therefore, as a tool for the assignment of the (pyru-
vyl9)-spirolactam amide bond conformation, prior to the
conformational analysis of 4a by NMR and molecular
modeling, we studied the conformational preference at
that amide bond in the spirolactam side chain model,
compound 6a, as well as its preference for â- or γ-turn
conformations.
The preference between the â- and γ-turn conforma-
tions was studied by molecular modeling calculations
carried out on the spiropseudodipeptide analogue 12
Conformational Analysis of the Spirolactam
Pseudodipeptide 6a. The NMR spectra of pseudo-

5704 Journal of Medicinal Chemistry, 2004, Vol. 47, No. 23
Gutie´rrez-Rodrı´guez et al.
Table 1. Relevant Topographic Parameters of the Low-Energy Conformers for the Spiropseudodipeptide 12 and Ideal â and γ Turns
angle (deg)
distances (Å)
∆E
(kcal/mol)
conformer
ω1
φ2
ψ2
φ3
ψ3
C1O-HN
C6O-HN
C2-Mea
(3-H)-Meb
12a
0
3.2
177.15
177.72
-52.3
-62.0
-60
133.7
121.9
120
101.4
-82.1
80
-37.6
61.5
0
2.18
5.46
3.20
2.14
5.92
8.90
<7
3.97
9.31
12b
â II turnc
γ turnd
-76
76
a Distance between pyruvyl C2 and the methyl amide carbon. b Distance between pyruvyl 3-H and Me protons. c Ideal â II turn. d Ideal
inverse γ turn.
(Table 1), where the D-Leu residue was replaced by
D-Ala to simplify the calculations. The conformational
space of this pseudodipeptide was explored by simulated
annealing molecular dynamics, followed by minimiza-
tion, employing the AMBER force field52,53 and using
the model-building facility implemented in Insight II
to construct the initial molecular models. The obtained
minima were compared, and the repeated ones were
eliminated. Unique minima were clustered into families
according to their heavy atom rms values and peptide
backbone torsion angle values and were analyzed for
the presence of â and γ turns. Additionally, because the
previous NMR study had shown the presence of cis and
trans rotamers for the pyruvyl-spirolactam amide bond,
the corresponding dihedral angle (ω1) was also analyzed.
Among the resulting structures, only 6% had a cis
conformation at that amide bond, and all of these were
at least 2 kcal mol-1 above the global minimum. The
low-energy conformers found within a 2 kcal mol-1
energy window from the global minima (88%), all of
them corresponding to trans isomers, were clustered
into six families. Optimization using the semiempirical
AM1 Hamiltonian,54 as implemented in the MOPAC
module of the Insight II program package, was carried
out. This led to two families of conformers whose most
significant topographic parameters are shown in Table
1. The calculated dihedral angles and distances for
conformers 12a (69% of the frames) are in good correla-
tion with those expected for a â II turn conformation.55-57
Thus, the distance between the pyruvyl C2 (which
corresponds to the CRi of the â turn) and the methyl
amide carbon (which corresponds to the CRi+3 of the â
turn) was less than 7 Å (5.92 Å), in accordance with one
of the most significant criteria for the definition of â
turns.46,56 In addition, the value of 2.18 Å for the
distance between the pyruvyl amide oxygen and the
methyl amide HN, less than than 2.5 Å, supports the
existence of a hydrogen bond in conformers 12a.46 With
respect to the dihedral angles, the difference between
the calculated value for ψ3 for these conformers and that
of an ideal â II turn (-37.64°) was higher than the
generally accepted window of (30°; however, differences
of up to (45° for this torsion angle have been also
reported.56,58 On the other hand, the calculated values
for the φ3 and ψ3 dihedral angles for 12b conformers
(31% of the frames), as well as the distance between
the spirolactam carbonyl oxygen and the methyl amide
HN (2.14 Å), correspond to those of a hydrogen-bonded
inverse γ turn centered at the D-Ala residue (i + 1
residue of the γ turn). The presence of a weak NOE
effect between pyruvyl 3-H and the methyl protons of
the methyl amide, observed in the CDCl3 1D NOESY
spectrum of the pseudodipeptide 6a, is compatible only
with a â turn conformation, since the distance calculated
for those protons in conformer 12b (9.31 Å) is too long
for that NOE to be observed. Therefore, the combined
results of the NMR, FT-IR, and molecular modeling
studies support the high preference for a â II turn
conformation in the spiropseudodipeptide 6a.
NMR Analysis of the Spirolactam Aplidine Ana-
logue 4a. As already mentioned, the NMR spectra of
the aplidine analogue 4a in CDCl3 and DMSO-d6
solution showed the presence of only one preferred
conformation whose proton and carbon resonances were
completely assigned on the basis of 2D COSY, TOCSY,
NOESY, HSQC, and HMBC experiments. This assign-
ment was confirmed with 2D HSQC-TOCSY spectra (1H
and 13C chemical shifts are included in Table 1 of the
Supporting Information). Although the pyruvyl-spiro-
lactam amide bond conformation could not be directly
established from these NMR data, it was assigned as
trans based on the high preference for a trans confor-
mation at this amide bond, determined in the model
spiropseudodipeptide 6a. Moreover, the 1H and 13C
NMR data of the spirolactam unit of 4a were more
similar to those of the trans isomer of 6a than to those
of the cis isomer.
1
A comparison of H and 13C chemical shifts of the
restricted aplidine analogue 4a in CDCl3 with those
described for didemnin B26 (1) and both cis and trans
conformations of aplidine27 (2) did not indicate signifi-
cant differences at the macrocycle residues. Concerning
the side chain residues, the most significant differences
were found at the (Thr6) NH and â-H and at the (Leu7)
R-H protons (Figure 5). In the first case, the differences
at the (Thr6) protons could be related to the higher
strength of the hydrogen bond between the (pyruvyl9)
carboxamide carbonyl group and the (Thr6) NH in 4a
than in aplidine or didemnin B (see below), while in the
case of the (Leu7) R-H the differences could be attributed
to the local modification that the spirolactam introduces
into this residue. In the overall comparison of chemical
shifts, the data of the restricted compound 4a were more
similar to those of trans-aplidine than to those of the
cis isomer.
As shown in Table 2, the temperature coefficients
determined in the study of temperature dependence of
amide proton chemical shifts for the restricted aplidine

Aplidine and Tamandarin A Analogues
Journal of Medicinal Chemistry, 2004, Vol. 47, No. 23 5705
Table 2. Temperature Dependence of the NH Chemical Shifts of Didemnin B, cis- and trans-Aplidine and 4a, Given as -∆δ/∆T
(ppb/K)
didemnin Ba
4a
cis-aplidinec
trans-aplidinec
a
b
d
e
NH
CDCl3
DMSO-d6
CDCl3
CDCl3
CDCl3
DMSO-d6
(Ist1)
1.70
2.20
4.25
0.5
1.8
4.3
2.00
2.20
3.51
2.00
2.20
4.80
0.44
0.37
0.89
0.70
1.22
3.36
(Leu3)
(Thr6)
a Reference 26. b Reference 25. c Reference 27. d Determined by measuring the chemical shifts between 25 and 57.5 °C with intervals
of 6.5 °C. e Determined by measuring the chemical shifts between 25 and 75 °C with intervals of 10 °C.
Figure 6. Proton-deuterium exchange kinetics for compound
4a in MeOH-d4.
Table 3. Observed Interresidue NOE Correlations for the
Aplidine Analogue 4aa
Figure 5. Comparison of 1H NMR chemical shifts of the side
chain residues: (shaded bar) didemnin B; (cross-hatched bar)
cis-aplidine; (dotted bar) trans-aplidine; (clear bar) aplidine
spirolactam analogue 4a.
analogue 4a in CDCl3 were significantly lower than
those described for didemnin B26 and both cis and trans
conformations of aplidine.27 The three low values ob-
tained for the temperature coefficients of 4a are char-
acteristic of strongly hydrogen-bonded amide pro-
tons.45,46 In particular, it is noteworthy that the value
obtained for the (Thr6) NH (0.89 ppb/K) is much lower
than the corresponding values described for didemnin
B (4.25 ppb/K)26 and cis- and trans-aplidine (3.51 and
4.8 ppb/K).27 As expected, the temperature coefficients
in DMSO-d6 were higher than in CDCl3 but within the
normal values for hydrogen-bonded NHs.45,51 Further-
a Observed in the NOESY spectra registered in CDCl3 at 500
more, the low solvent dependence of the NH chemical
MHz (mixing times: 500 and 750 ms).
shifts from CDCl3 to DMSO-d6 [0.24 ppm for (Ist1), 0.08
ppm for (Leu3), and 0.06 ppm for (Thr6)] indicated the
low accessibility of the solvent to these NHs.52 This low
accessibility was also deduced from the study of proton-
deuterium exchange in MeOH-d4. As shown in Figure
6, the (Ist1) and (Leu3) NHs did not exchange with
deuterium after being dissolved in MeOH-d4 for 1 h,
which indicated an internal orientation of these NHs,
making them completely inaccessible to the solvent,
while the slow exchange of (Thr6) NH suggests its
participation in a hydrogen bond but in slow equilibrium
with a nonbonded state.
significant nonsequential NOEs observed between the
macrocycle residues were also in good agreement with
those reported for both cis and trans conformations of
aplidine and for didemnin B in CDCl3, indicating that
the three-dimensional structure of the macrocycle is
similar in all cases. Thus, the cross-peaks of (Ist1) NH
with (Pro4) R-H, (Me2Tyr5) R-H, and N-Me protons
supported the participation of that NH in a hydrogen
bond, forcing the macrocycle into the figure of eight
shape, and the cross-peaks of (Leu3) NH with (Thr6) R-H
and (D-Leu7) R-H, as well as (Leu3) R-H with (Thr6) NH
and (D-Leu7) â-H, supported the presence of the (Leu3)
NH f (D-Leu7) CdO hydrogen bond and the folding
back of the side chain toward the macrocycle. This
folding was also indicated by the NOE correlations of
the spirolactam 8-H protons (corresponding to the N-Me
of aplidine Leu7) with the (Ist1) NH and 4-H protons.
As in the case of didemnin B25,26 and aplidine,27 the
sequential NOE correlations observed in the NOESY
spectra between the NH or NCH3 and the R proton of
the preceding residue and between (Leu3) R-H and (Pro4)
δ-H allowed the assignment of a trans conformation to
all peptide bonds in 4a (Table 3). Moreover, the most

5706 Journal of Medicinal Chemistry, 2004, Vol. 47, No. 23
Gutie´rrez-Rodrı´guez et al.
With respect to the side chain, the NOE pattern was
also similar to that of aplidine and in good agreement
with that expected for a â II turn conformation. Thus,
strong NOE correlations were observed between (Leu7)
R-H and (Thr6) NH, the spirolactam 8-H and (Thr6) NH,
and the spirolactam 4-H protons (which correspond to
Pro8 â-H of aplidine), and between the pyruvyl 3-H
protons and (Thr6) â-H and γ-H. The overall NOE
correlation pattern did not change from CDCl3 to
DMSO-d6, indicating the absence of significant confor-
mational variability with the solvent.
Molecular Mechanics/Dynamics Calculations for
Spirolactam Aplidine Analogue 4a. Owing to the
inherent lack of precision associated with the determi-
nation of interproton distances in molecules of the size
of aplidine by NMR techniques,59 in addition to the
limited number of interatomic distances per residue
accessible experimentally, molecular mechanics/dynam-
ics calculations were carried out in order to provide more
information on the three-dimensional structure of the
spirolactam aplidine analogue 4a and to examine its
conformational preferences. Calculations were per-
formed using the well-established CHARMm force
field60,61 at 300K for reasonably long periods of time (see
Experimental Section) and without distance restraints
in order to allow the molecule to more fully explore the
conformational space available to it. Three implicit
solvent descriptions (ꢀ ) 4.8, 45, and 80) and two explicit
solvent models (H2O and DMSO) were applied. Tables
of derived quantities (interatomic distances, torsion
angles, and hydrogen-bonding patterns) resulting from
these calculations are included in the Supporting In-
formation (see Tables 2-4 in the Supporting Informa-
tion).
Good agreement was observed between the inter-
atomic distances derived from the unrestrained molec-
ular dynamics trajectories in the CHARMm all-atom
representation and the NOE-derived distances from the
experimental NMR data. Nearly all of the atomic
distances that corresponded to the measured NOE
distances were within those values that would be
expected to produce an NOE (see Table 2 in the
Supporting Information).
Analysis of the most relevant torsion angles found in
the simulations indicates that the overall three-dimen-
sional structure of the spirolactam aplidine analogue
4a is similar to that of the trans isomer of aplidine.27
Furthermore, comparison with reported results for
didemnin B by NMR studies25 and by X-ray diffraction24
indicates that the three-dimensional structures of all
three molecules are similar, the macrocycle in each case
adopting a twisted figure eight conformation.
Figure 7. Superposition of the representative snapshots from
the trajectories of the DMSO explicit-solvent simulations for
the aplidine spirolactam analogue 4a (green) and trans-
aplidine (orange).
the molecule’s side chain, a hydrogen bond between
(Thr6) NH and (pyruvyl9) carboxamide CO, analogous
to that found in aplidine27 and to that found in didemnin
B between (Thr6) NH and (Lac9) CO,24 is found in all
simulations. Another hydrogen bond between (Thr6) NH
and (spirolactam) CO is also detected. These hydrogen
bonds stabilize â- and γ-type turns, respectively, in the
molecule’s side chain. This result is also compatible with
the low ∆δ/∆T value observed for the (Thr6) NH proton,
as shown in Table 2.
The similarity of the three-dimensional structure of
the spirolactam analogue of aplidine 4a to that of
aplidine and to that of didemnin B is illustrated by the
superposition of representative snapshots from the
trajectories of the DMSO explicit-solvent simulations of
both these molecules and with the reported X-ray crystal
structure of didemnin B as shown in Figures 7 and 8.
Only those atoms belonging to the macrocycle in each
case were aligned, minimizing the root-mean-square
differences. The rmsd values of 1.1228 and 0.7319 Å,
respectively, were obtained between the spirolactam
analogue 4a and aplidine on one hand and between 4a
and didemnin B on the other.
Biological Results and Discussion. The cytotox-
icity of the conformationally restricted spirolactam
didemnin analogues 4 and 5 in human lung carcinoma
(A549) and human colon carcinoma (HT-29) cell lines
was determined as previously described.62 For compara-
tive purposes, didemnin B (1), aplidine (2), tamandarin
A (3), and dehydrotamandarin A were also included in
the assay. As indicated in Table 4, all the new com-
pounds were less cytotoxic than the model compounds
and, similarly to other didemnin congeners, none of
them showed selectivity for either of the two cell lines
assayed.18 The conformation restriction introduced to
enforce the â-turn conformation in the linear side chain
of aplidine led to a 20- and 50-fold decrease in the
cytotoxicity of the spirolactam containing 4a against
HT-29 and A549 cells, respectively. An even more
marked decrease (200-fold) was found in the cytotoxicity
of 5a with respect to that of dehydrotamandarin A.
The hydrogen-bonding patterns detected in the simu-
lations show that the transannular hydrogen bond
between (Ist1) NH and (Leu3) CO is present in almost
all simulations, as is that between (Leu3) NH and
(MeLeu7) CO. This result is compatible with the low ∆δ/
1
∆T values observed by H NMR for the (Ist1) NH and
(Leu3) NH protons, respectively (see Table 2). These
hydrogen bonds, also present in aplidine and in didem-
nin B, are responsible for the above-mentioned figure
eight disposition of the macrocycle. In the present study
another transannular hydrogen bond between (Leu3)
NH and (Ist1) CO is also detected in all simulations. In

Aplidine and Tamandarin A Analogues
Journal of Medicinal Chemistry, 2004, Vol. 47, No. 23 5707
Table 4. Cytotoxicity of Conformationally Restricted Didemnin Analogues 4 and 5
IC50 (nM)
compd
R1
(S)-Lac
pyruvyl
(S)-Lac
pyruvyl
pyruvyl
isobutyryl
Boc
X
A549a
HT-29b
didemnin B (1)c
(S)-[COCH(CH3)CO]
(S)-[COCH(CH3)CO]
2
2
aplidine (2)
0.18
0.95
0.47
8.55
89.1
8.68
93.8
93.8
4.56
93.9
0.45
0.95
0.47
8.90
89.1
8.68
93.8
93.8
4.56
93.9
tamandarin A (3)
CO
dehydrotamandarin A
CO
4a
4b
4c
5a
5b
5c
5d
(S)-[COCH(CH3)CO]
(S)-[COCH(CH3)CO]
(S)-[COCH(CH3)CO]
pyruvyl
isobutyryl
Boc
CO
CO
CO
CO
2-Me-acryloyl
a A549 ) human lung carcinoma cells. b HT-29 ) human colon carcinoma cells. c Values reported in ref 18.
high preference for the trans rotamer at the Xaa9-Pro8
amide bond, the lower flexibility at the side chain in 4a
could account for its 4-fold lower cytotoxicity.
Although a thorough conformational analysis of com-
pounds 4b, 4c, and 5a-d has not been carried out, the
importance of the R1-Pro9 bond in a cis conformation
would explain the considerable differences in cytotox-
icity between the N-acyl derivative 4b or 5b and their
corresponding N-urethanyl analogue 4c or 5c in both
series. Since no evidence of a mixture of rotamers was
detected either in the model isobutyryl 8b and 9b or in
the final macrocycle derivative 4b and 5b, it is reason-
able to assume that these latter compounds, which are
among the least cytotoxic spirolactam derivatives, do
exist in a trans conformation, as discussed above. In
contrast, on the basis of the preference for the cis
rotamer in the model Boc derivative 9c and the litera-
ture data,39 we presume that the only rotamer detected
in the macrocyclic analogues 4c and 5c, among the most
cytotoxic spirolactam derivatives, is also in this confor-
mation. It is interesting to note that compound 5c,
which showed the highest cytotoxicity among all these
spirolactams, is less cytotoxic than tamandarin A and
dehydrotamandarin A by only 5- and 10-fold, respec-
tively. Finally, the decrease in cytotoxicity of the spiro-
lactam containing the 2-Me-acryloyl moiety 5d could
also be attributed to the adoption of the (2-Me-acryl-
oyl)9-Pro8 amide bond only in the trans conformation,
as demonstrated in the model compounds 8d and 9d.
Figure 8. Superposition of the representative snapshots from
the trajectories of the DMSO explicit-solvent simulations for
the aplidine spirolactam analogue 4a (green) and the X-ray
crystal structure of didemnin B (orange).
Taking into account the results of the conformational
analysis of the aplidine analogue 4a in comparison with
aplidine, the decrease in the cytotoxicity produced by
the spirolactam conformational constraint cannot be
related either to changes in the macrocycle conformation
or to an inefficient induction of the preferred â II turn
conformation in the side chain. As mentioned above,
apart from the lower flexibility of the side chain in the
restricted analogue 4a, the most significant difference
is that, unlike aplidine in which the cis and trans
rotamers around the pyruvyl9-Pro8 amide bond are
similarly stabilized,27,28 the (pyruvyl9)-spirolactam amide
bond essentially adopts the trans form in 4a. This fact
strongly suggests the preference for the pyruvyl9-Pro
amide cis isomer in the bioactive conformation of
aplidine. In view of the overall 3D structural similarity
between didemnin B and compound 4a, including the
The significant influence of the Xaa9-Pro8 amide
bond cis/trans isomerism upon the cytotoxic activity of
didemnins, herein commented, suggests the possible
participation of a peptidyl-prolyl cis/trans isomerase
in their mechanism of action. This participation has
been previously suggested to explain the immunosup-
pressive effects of didemnin B,3,63 although, unlike the
macrocyclic immunosuppressants cyclosporin A (CsA),
FK506, and rapamycin, a didemnin B specific immu-
nophilin has not been identified. The first step in the

5708 Journal of Medicinal Chemistry, 2004, Vol. 47, No. 23
Gutie´rrez-Rodrı´guez et al.
Table 5. Relevant Analytical and Spectroscopic Data of Spirolactam Pseudodipeptides
6a
pyruvyl
NHCH3
76
8a
pyruvyl
OBn
56
syrup
C23H30N2O5
8b
8d
9a
pyruvyl
H
95
amorphous
C16H24N2O5
13.14 (20:80)
+3.5
(c 1, MeOH)
9b
9c
9d
R1
isobutyryl
OBn
2-Me-acryloyl
OBn
Isobutyryl
H
Boc
H
64
2-Me-acryloyl
H
R2
yield (%)
mp (°C)
formulaa
74
87
50
syrup
95
85
syrup
68-69
C17H28N2O4
NDc
192-194
C18H30N2O5
13.41 (50:50)
-3
amorphous
C17H26N2O4
6.38 (25:75)
C17H27N3O4
C24H34N2O4
C24H32N2O4
6.38 (50:50)
NDc
tR (min) (A/B)b 3.60 (30:70)
5.87, 6.72 (50:50) 4.50 (50:50)
[R]20
+69.4
NDc
+9.6
-2
+18.8
D
(c 1, MeOH)
(c 1, MeOH)
(c 1.1, MeOH) (c 1, MeOH)
(c 1, MeOH)
ESI-MS (m/e) 338.3 [M + H]+ 415.4 [M + H]+
415.4 [M + H]+ 413.3 [M + H]+ 325.1 [M + H]+ 323.3 [M-H]+ 353.4 [M - H]+ 323.2 [M + H]+
6a
8a
9a
1H
majore
minorf
(cis)
majore minorf
(trans) (cis)
majore
(trans)
minorf
(cis)
NMRd (trans)
8b
8d
9b
9c
9d
2-H 3.68, 3.82 3.52, 3.56 3.56-3.78, 3.92 3.60-3.70
3.56-3.68
1.64-1.93
1.96-2.12
3.58-3.70, 3.86
1.64-1.82
1.86-2.18
3.61-3.75
1.90-2.10
1.90-2.10
3.30, 3.49
3.30
3.57-3.67
1.55-2.25
3.57-3.67
3.19-3.44
3-H
4-H
8-H
9-H
R-H
â-H
R1
1.86-2.20
1.86-2.20
3.37
1.61-2.39
1.61-2.39
1.61-2.10
1.61-2.10
1.70-1.90
1.70-1.90
3.10, 3.70
3.22, 3.56-3.78 3.14, 3.60-3.70 3.20, 3.56-3.68
3.46
3.35
1.86-2.20, 2.51
1.61-2.39, 2.77 1.61-2.10, 2.64 1.64-1.93, 2.78 1.86-2.18, 2.60-2.74 1.90-2.10, 2.49 1.70-1.90, 2.50 1.55-2.25, 2.56
4.85
4.50
4.85
1.61-2.39
2.44
2.27 1.09, 1.12g
4.67 4.85
4.80
4.80
4.73
4.71
4.39
1.64
1.36
4.78
1.62, 1.91
1.61-2.10
1.64-1.93
1.71-2.22
1.70
1.55-2.25
2.38
2.29
5.19h
2.39
2.30
1.09, 1,12g
5.18h
6a
8a
majore (trans)
9a
13C NMRi
majore (trans)
minorf (cis)
minorf (cis)
8b
8d
majore (trans)
minorf (cis)
9b
9c
9d
C2
C3
C4
C5
C8
C9
CR
Câ
R1
50.21
25.73
37.45
69.85
40.83
49.01
25.79
37.37
70.21
41.14
48.91
24.92
36.47
66.83
47.72
24.76
35.63
67.03
47.87
24.86
35.76
67.55
40.49
29.55
52.71
37.62
18.64g
50.19
24.22
37.09
66.81
41.11
29.97
53.24
37.70
115.91h
49.30
24.93
36.01
48.03
24.70
36.60
48.04
25.01
36.55
67.57
40.64
30.78
53.46
36.94
18.65g
48.19
22.59
36.87
67.29
40.59
30.45
54.32
38.18
28.54
50.51
24.61
37.43
67.49
40.94
30.13
53.72
37.82
116.27h
69.11
40.77
41.01
29.90
53.42
37.01
27.09
197.49
41.31
32.50
53.58
37.44
26.79
197.49
29.49
54.54
36.81
29.31
52.90
37.28
26.94
198.04
32.31
53.18
37.39
26.94
197.65
27.17
201.45
27.58
198.70
a Satisfactory analyses for C, H, N. b NovapacK C18 (3.9 mm × 150 mm, 4 µm), A ) CH3CN, B ) 0.05% TFA in H2O. c Not determined.
d Spectra registered at 300 MHz in CDCl3 except for 6a (acetone-d6) and 9c (DMSO-d6, 90 °C). Spectra assignment based on COSY and
HSQC spectra. e Major rotamer at the pyruvyl amide bond. f Minor rotamer at the pyruvyl amide bond. g δ for the R1 ) CH3. h δ for the
R1 ) CH2. i Spectra registered at 75 MHz in CDCl3 except for 6a (acetone-d6).
spectrometers, operating at 50 or 100 MHz. Elemental analy-
ses were obtained on a CH-O-RAPID apparatus. Analytical
RP HPLC was performed on Waters µBondapak C18 (3.9 mm
× 300 mm, 4 µm) or Novapak C18 (3.9 mm × 150 mm, 4 µm)
columns with a flow rate of 1 mL/min and using a tunable UV
detector set at 214 nm. Mixtures of CH3CN (solvent A) and
0.05% TFA in H2O (solvent B) were used as mobile phases.
Analytical and preparative HPLC of the final peptides was
performed on a ThermoQuest HyperPrep PEP 100 C18 (250
mm × 21 mm) column with a flow rate of 7 mL/min, using a
tunable UV detector set at 270 nm. Mixtures of CH3CN and
0.05% TFA in H2O were used as mobile phases in isocratic
mode. Optical rotations were measured on a Perkin-Elmer 141
polarimeter. ESI-MS experiments were performed, in positive
mode, on a Hewlett-Packard 1100 MSD apparatus.
Synthesis of (5R)-7-[(1R)-1-Benzyloxycarbonyl-3-meth-
ylbutyl]-6-oxo-1-pyruvyl-1,7-diazaspiro[4.4]nonane (8a).
Method A. Pyruvic acid (19.2 mg, 0.22 mmol), HATU (100
mg, 0.26 mmol), HOAt (35 mg, 0.26 mmol), and NMM (48 µL,
0.44 mmol) were successively added, under argon, to a 0 °C
cooled solution of freshly prepared trifluoroacetate 732 (101 mg,
0.22 mmol) in dry CH2Cl2 (5 mL). After 24 h of stirring at room
temperature, the reaction mixture was diluted with CH2Cl2
(5 mL), and the solution was successively washed with 10%
citric acid solution (5 mL), saturated NaHCO3 (5 mL), and
brine (5 mL), dried over Na2SO4, and evaporated to dryness.
The residue was purified by radial chromatography, using (1:
3) hexane/EtOAc as eluant to give the title pseudodipeptide
derivative 8a as a syrup (42 mg, 47%), whose significant
analytical and spectroscopic data are shown in Table 5.
intracellular action of cyclosporin A, FK-506, and ra-
pamycin is their binding to their specific peptidyl-prolyl
isomerases, cyclophilin A (CsA), and FK506-binding
proteins (FK506 and rapamycin).3,64,65 Besides their
fundamental role in protein folding,66 peptidyl-prolyl
isomerases are crucial in signal transduction, cell cycle
regulation, and receptor trafficking and assembly65-68
so that interference of didemnins with peptidyl-prolyl
isomerase activity might be related to their antitumoral
and immunosuppressive effects.
Experimental Section
General Synthetic Methods. All reagents were of com-
mercial quality. Solvents were dried and purified by standard
methods. Amino acid derivatives were obtained from Bachem
Feinchemikalien AG. Analytical TLC was performed on alu-
minum sheets coated with a 0.2 mm layer of silica gel 60 F254
Merck, and preparative TLC was performed on 20 cm × 20
cm glass plates coated with a 2 mm layer of silica gel PF254
,
,
Merck. Silica gel 60 (230-400 mesh), Merck, was used for flash
chromatography. Preparative radial chromatography was
performed on 20 cm diameter glass plates coated with a 2 mm
layer of silica gel PF254, Merck. Melting points were taken on
a micro-hot-stage apparatus and are uncorrected. 1H NMR
spectra were recorded with Varian Gemini-200, Varian INOVA-
300, or Varian INOVA 400 spectrometer, operating at 200, 300,
or 400 MHz, using TMS as reference. 13C NMR spectra were
recorded with Varian Gemini-200 or Varian INOVA 400

Aplidine and Tamandarin A Analogues
Journal of Medicinal Chemistry, 2004, Vol. 47, No. 23 5709
Method B. Triethylamine (280 µL, 1.98 mmol) and DMAP
(4 mg, 33 µmol) were added to a 0 °C cooled solution of freshly
prepared trifluoroacetate 732 (151 mg, 0.33 mmol) in dry CH2-
Cl2 (4 mL). This mixture was slowly added to a solution of
pyruvyl chloride, freshly prepared from pyruvic acid (115 mg,
1.31 mmol) and dichloromethyl methyl ether (188 mg, 1.57
mmol) as previously described,34 in CH2Cl2 (3 mL). The
resulting reaction mixture was stirred at room temperature
for 6 h and diluted with CH2Cl2 (10 mL). The solution was
successively washed with 10% citric acid solution (5 mL),
saturated NaHCO3 (5 mL), and brine (5 mL), dried over Na2-
SO4, and evaporated to dryness. The residue was purified as
in method A to yield the derivative 8a (77 mg, 56%).
General Procedure for the Synthesis of N-Acyl Spiro-
lactam Pseudodipeptides 8b and 8d. Triethylamine (283
µL, 2 mmol), DMAP (6.1 mg, 50 µmol), and the corresponding
acyl chloride [isobutyryl chloride (106 mg, 1 mmol) or 2-meth-
ylacryloyl chloride (103 mg, 1 mmol)] were successively added
to a 0 °C cooled solution of freshly prepared trifluoroacetate
732 (228 mg, 0.50 mmol) in dry CH2Cl2 (10 mL). After 24 h of
stirring at room temperature, the reaction mixture was diluted
with CH2Cl2 (10 mL). The solution was successively washed
with 10% citric acid solution (5 mL), saturated NaHCO3 (5
mL), and brine (5 mL), dried over Na2SO4, and evaporated to
dryness. The residue was purified by flash chromatography,
using a 20-100% gradient of EtOAc in hexane as eluants. The
most significant analytical and spectroscopic data of the title
compounds 8b and 8d are summarized in Table 5.
h. Afterward, the MeOH was evaporated, and the residue was
dissolved in H2O (10 mL). This aqueous solution was succes-
sively washed with CH2Cl2 (10 mL), acidified with 0.1 N HCl
to pH 2, and extracted with CH2Cl2 (3 × 20 mL). The combined
organic extracts were washed with brine (20 mL), dried over
Na2SO4, and evaporated to dryness to yield the title pseudo-
dipeptide 9d as a white amorphous solid (55 mg, 85%), whose
significant analytical and spectroscopic data are summarized
in Table 5.
General Procedure for the Synthesis of Spirolactam
Pseudodepsipeptides 4a-c and 5a-d. The corresponding
pseudodipeptides 9a-d (14 µmol), HATU (12.4 mg, 32 µmol),
HOAt (4.5 mg, 33 µmol), and NMM (3.3 µL, 30 µmol) were
successively added, under argon, to a -5 °C cooled solution of
the respective macrocyclic depsipeptide hydrochlorides 1033
and 1121 (freshly prepared from 13 µmol of their N-Boc-
protected derivatives by treatment with 50 µL of 4.6 M HCl
solution in EtOAc) in a (2:1) CH2Cl2/DMF mixture (150 µmol).
The resulting mixture was stirred, first for 2 h at -5 °C and
then for 14 h at room temperature. Afterward, the solvent was
evaporated, and the residue was treated with EtOAc (2 mL).
The precipitate was filtered off, and the solution was washed
successively with 10% KHSO4 (2 × 1 mL), H2O (1 mL), a
saturated solution of NaHCO3 (1 mL), and brine (1 mL), dried
over Na2SO4, and evaporated to dryness. The residue was
purified by preparative HPLC performed on a HyperPrep PEP
100 C18 (250 mm × 21 mm) column with a flow rate of 7 mL/
min, using a (85:15) CH3CN/0.05% TFA mixture as eluant.
General Procedure for the Hydrogenolysis of the
Benzyl Esters 8a and 8b. Synthesis of the C-Deprotected
Pseudodipeptides 9a and 9b. A sample of 10% Pd(C) (30
mg) was added to a solution of the corresponding benzyl ester
8a and 8b (0.3 mmol) in MeOH (30 mL), and the mixture was
hydrogenated at room temperature and 1 atm of H2 pressure
for 1 h. The catalyst was filtered off and washed with MeOH
(5 mL), and the solution was evaporated to dryness. The
residue was dissolved in H2O (3 mL), and the resulting solution
was lyophilized to give the C-deprotected pseudodipeptides 9a
and 19b as white solids, whose analytical and spectroscopic
data are summarized in Table 5.
Synthesis of (5R)-7-[(1R)-1-(N-Methyl)carbamoyl-3-meth-
ylbutyl]-6-oxo-1-pyruvyl-1,7-diazaspiro[4.4]nonane (6a).
Pseudodipeptide 9a (32 mg, 0.10 mmol), BOP (53 mg, 0.12
mmol), and NMM (22 µL, 0.20 mmol) were successively added,
under argon, to a -5 °C cooled solution of methylamine
hydrochloride (6 mg, 0.09 mmol) in dry CH2Cl2 (2 mL). After
2 h of stirring at -5 °C and 20 h at room temperature, the
reaction mixture was diluted with CH2Cl2 (2 mL), and the
solution was successively washed with 10% citric acid solution
(2 mL), saturated NaHCO3 (2 mL), and brine (2 mL), dried
over Na2SO4, and evaporated to dryness. The residue was
purified by radial chromatography, using a 30-50% EtOAc
gradient in hexane as eluant, to give the title pseudodipeptide
derivative 6a as a syrup (25 mg, 76%), whose significant
analytical and spectroscopic data are shown in Table 5.
Synthesis of (5R)-1-(tert-Butoxycarbonyl)-7-[(1R)-1-
carboxy-3-methylbutyl]-6-oxo-1,7-diazaspiro[4.4]no-
nane (9c). Tetramethylammonium hydroxide pentahydrate
(158 mg, 0.87 mmol) and di(tert-butyl) dicarbonate (144 mg,
0.66 mmol) were added to a solution of freshly prepared
trifluoroacetate 732 (150 mg, 0.44 mmol) in acetonitrile (5 mL).
After 48 h of stirring at room temperature, the solvent was
evaporated to dryness, and the residue was dissolved in CH2-
Cl2 (3 mL). The solution was extracted with H2O (5 mL), and
the extracts were lyophilized. The resulting residue was
purified by flash chromatography, using an 8-40% gradient
of MeOH in CH2Cl2 as eluant, to give the title compound 9c
as a white solid (100 mg, 64%), whose analytical and spectro-
scopic data are summarized in Table 5.
[(5R)-7-[(1R)-1-Carbonyl-3-methylbutyl]-6-oxo-1-pyru-
vyl-1,7-diazaespiro[4.4]nonane7-9]aplidine (4a). White
amorphous solid (9 mg, 63%). HPLC [HyperPrep PEP 100 C18
(A/B, 85:15), φ ) 7 mL/min] tR ) 14.4 min. 1H NMR (300 MHz,
CDCl3): δ 0.90-1.00 (m, 24 H), 1.05-1.40 (m, 12 H), 1.40-
2.25 (m, 16H), 2.27-2.41 (m, 1H), 2.42-2.70 (m, 3H), 2.54 (s,
3H), 2.92-2.98 (m, 1H), 3.12-3.38 (m, 4 H), 3.54-3.78 (m, 4
H), 3.79 (s, 3H), 4.01-4.12 (m, 2H), 4.20-4.26 (m, 2H), 4.57-
4.62 (m, 2H), 4.77-4.82 (m, 2H), 5.18 (d, 2H, J ) 3 Hz), 5.37-
5.42 (m, 1H), 6.84 (d, 2H, J ) 9 Hz), 7.07 (d, 2H, J ) 9 Hz),
7.20 (d, 1H, J ) 10 Hz), 7.85 (d, 1H, J ) 10 Hz), 7. 87 (d, 1H,
J ) 5 Hz). 13C NMR (75 MHz, CDCl3): δ 11.86, 14.97, 15.47,
16.78, 17.12, 18.84, 21.20, 21.27, 23.62, 24.14, 15.03, 25.15,
25.30, 27.35, 27.51, 28.19, 30.46, 31.52, 34.27, 35.95, 39.01,
40.07, 41.59, 47.25, 49.72, 53.08, 55.51, 55.81, 57.38, 58.21,
66.66, 68.19, 69.23, 70.74, 81.69, 85.15. 114.33, 130.13, 130.56,
158.85, 161.12, 168.49, 169.81, 169.91, 170.81, 171.38, 171.69,
172.60, 173.34, 197.64, 205.19. ESI-MS m/e Calcd for C58H87-
N7O15 1121.6. Found 1122.3 [M + H]+.
[(5R)-7-[(1R)-1-Carbonyl-3-methylbutyl]-1-isobutyryl-
6-oxo-1,7-diazaespiro[4.4]nonane7-9]aplidine (4b). White
amorphous solid (10 mg, 69%). HPLC [HyperPrep PEP 100
1
C18 (A/B, 85:15), φ ) 7 mL/min] tR ) 19.0 min. H NMR (300
MHz, CDCl3): δ 0.86-1.00 (m, 24H), 1.12 (d, 3H, J ) 7 Hz),
1.18 (d, 3H, J ) 7 Hz), 1.34 (t, 2H, J ) 7 Hz), 0.90-1.30 (m,
7H), 1.56-2.25 (m, 16 H), 2.30-2.80 (m, 3H), 2.55 (s, 3H),
2.95-3.06 (m, 1H), 3.15-3.25 (m, 3 H), 3.65 (dd, 1H), 3.52-
3.79 (m, 6H), 3.79 (s, 3H), 3.98-4.15 (m, 1H), 4.28 (dd, 1H,
J ) 7 and 10 Hz), 4.59 (m, 2H), 4.79-4.85 (m, 2H), 5.17 (d,
1H, J ) 4 Hz), 5.40-5.44 (m, 1H), 6.84 (d, 2H, J ) 8 Hz), 7.06
(d, 2H, J ) 9 Hz), 7.24 (d, 1H, J ) 11 Hz), 7.90 (d, 1H, J ) 9
Hz), 8.56 (d, 1H, J ) 5 Hz). 13C NMR (75 MHz, CDCl3): δ
11.50, 14.92, 15.22, 16.68, 16.95, 18.50, 18.59, 18.81, 20.94,
23.42, 23.80, 24.49, 24.68, 24.90, 25.11, 26.91, 27.98, 30.93,
31.30, 35.58, 33.88, 34.16, 35.85, 36.24, 38.64, 38.84, 39.71,
41.29, 47.01, 47.76, 49.42, 49.62, 52.66, 55.26, 55.73, 57.12,
58.21, 66.57, 67.40, 68.10, 70.79, 81.57, 114.07, 130.05, 130.31,
158.57, 168.12, 169.66, 170.08, 170.56, 171.13, 171.96, 172.37,
174.04, 175.41, 205.04. ESI-MS m/e Calcd for C59H91N7O14
1121.6. Found 1122.8 [M + H]+.
[(5R)-1-(tert-Butoxycarbonyl)-7-[(1R)-1-carbonyl-3-meth-
ylbutyl]-6-oxo-1,7-diazaespiro[4.4]nonane7-9]aplidine (4c).
White amorphous solid (11 mg, 73%). HPLC [HyperPrep PEP
Synthesis of (5R)-7-[(1R)-1-Carboxy-3-methylbutyl]-1-
(2-methylacryloyl)-6-oxo-1,7-diazaspiro[4.4]nonane (9d).
A sample of 0.1 N NaOH (2.2 mL, 0.22 mmol) was added to a
solution of the benzyl ester 8d (83 mg, 0.2 mmol) in MeOH (3
mL), and the mixture was stirred at room temperature for 1
1
100 C18 (A/B, 85:15), φ ) 7 mL/min] tR ) 30.0 min. H NMR
(300 MHz, CDCl3): δ 0.85-0.97 (m, 24 H), 1.19-1.34 (m, 18
H), 1.48 (s, 9H), 1.50-2.20 (m, 6 H), 2.32-2.36 (m, 1H), 2.54

5710 Journal of Medicinal Chemistry, 2004, Vol. 47, No. 23
Gutie´rrez-Rodrı´guez et al.
1
(s, 3H), 2.58-2.72 (m, 2H), 2.97-3.08 (m, 1H), 3.10-3.22 (m,
3H), 3.34 (dd, 1H, J ) 4 and 14 Hz), 3.46-3.79 (m, 6H), 3.79
(s, 3 H), 4.03-4.12 (m, 2H), 4.28 (dd, 1H, J ) 7 and 13.2 Hz),
4.57-4.63 (m, 2H), 4.79-4.88 (m, 2H), 5.15 (d, 1H, J ) 3 Hz),
5.21-5.23 (m, 1H), 6.84 (d, 2 H, J ) 8 Hz), 7.07 (d, 2H, J ) 9
Hz), 7.28 (d, 1H, J ) 11 Hz), 7.79 (d, 1H, J ) 7 Hz), 7.82 (d,
1H, J ) 10 Hz). 13C NMR (75 MHz, CDCl3): δ 11.7, 15.2, 15.3,
16.8, 17.2, 18.7, 18.8, 21.2, 23.6, 23.9, 24.1, 24.9, 25.1, 25.4,
27.0, 28.2, 28.7, 29.9, 31.6, 34.5, 33.8, 34.2, 36.2, 36.6, 38.76,
39.1, 39.8, 41.3, 47.3, 47.8, 49.6, 49.9, 52.7, 55.5, 56.3, 57.4,
58.2, 66.7, 66.9, 68.3, 70.5, 80.7, 81.9, 114.3, 130.2, 130.5, 168.1,
169.8, 170.2, 170.8, 171.4, 172.5, 175.0, 205.0. ESI-MS m/e
Calcd for C60H93N7O15 1151.7. Found 1152.4 [M + H]+.
DMSO-d6 were used for NMR experiments. H NMR spectra
were recorded on Varian INOVA 300 and 400 MHz spectrom-
eters, and the 13C NMR spectra were recorded on a Varian
INOVA 300 operating at 75 MHz. Degassed solutions of 12
mM aplidine spirolactam analogue 4a in CDCl3 and DMSO-
d6 were used for NMR experiments. In this case, the 1H NMR
spectra were recorded on a Varian INOVA 500 MHz spectrom-
eter, while the 13C NMR spectra were recorded on a Varian
Mercury 400 spectrometer operating at 100 MHz. The mixing
time for the 2D TOCSY experiments was 70 ms, while for the
NOESY spectra mixing times of 500 and 750 ms were used.
The data were collected in phase using the hypercomplex
method, and the spectral width used was 5141 Hz in the two
dimensions. Four accumulations and 256 increments were
carried out in the TOCSY period, and 16 accumulations and
128 increments were carried out in the NOESY experiments.
The spectra were processed on Sun Microsystems SunBlade
100 and Ultra5 computers, using Varian VNMR 6.1C software.
NOE buildup curves were generated from five NOESY experi-
ments in DMSO-d6 (mixing times of 100, 150, 200, 300, and
400 ms), using the distance between the (Leu3) NH and (Thr6)
R-H determined in the X-ray structure of didemnin B for
calibration purposes.
[Hiv2]-[(5R)-7-[(1R)-1-carbonyl-3-methylbutyl]-6-oxo-1-
pyruvyl-1,7-diazaespiro[4.4]nonane7-9]aplidine (5a). White
amorphous solid (10 mg, 72%). HPLC [HyperPrep PEP 100
1
C18 (A/B, 85:15), φ ) 7 mL/min] tR ) 13.6 min. H NMR (300
MHz, CDCl3): δ 0.80-1.10 (m, 24H), 1.11-1.80 (m, 12H), 1.81-
2.30 (m, 10H), 2.45 (m, 1H), 2.55 (s, 3H), 2.57 (s, 3H), 3.07-
3.43 (m, 6H), 3.52-3.77 (m, 6H), 3.78 (s, 3H), 3.91 (m, 1H),
4.03 (m, 1H), 4.29 (m, 1H), 4.63 (m, 1H), 4.72 (m, 1H), 4.87
(m, 1H), 5.03 (d, 1H, J ) 4 Hz), 5.45 (m, 1H), 6.84 (d, 2H, J )
8 Hz), 7.07 (d, 2H, J ) 8), 7.29 (d, 1H, J ) 9 Hz), 7.81 (d, 1H,
J ) 9 Hz), 7.87 (d, 1H, J ) 5 Hz). ESI-MS m/e Calcd for
C55H83N7O14 1065.6. Found 1066.4 [M + H]+.
FT-IR Study of the Spiropseudodipeptide 6a. FT-IR
spectra were recorded on a Perkin-Elmer-Spectrum One FT-
IR spectrometer with 4 cm-1 resolution, using a 0.5 mm NaCl
cell. The spectra were obtained with 64 scans, using 1.4, 5,
and 14 mM solutions in dry CHCl3 at room temperature.
Background spectra were recorded with the solvent. The
spectra were processed using the Perkin-Elmer Spectrum for
Windows V3.02 software.
[Hiv2]-[(5R)-7-[(1R)-1-carbonyl-3-methylbutyl]-1-isobu-
tyryl-6-oxo-1,7-diaza-espiro[4.4]nonane7-9]aplidine (5b).
White amorphous solid (10 mg, 70%). HPLC [HyperPrep PEP
1
100 C18 (A/B, 85:15), φ ) 7 mL/min] tR ) 16.9 min. H NMR
(300 MHz, CDCl3): δ 0.80-1.07 (m, 24H), 1.08-1.47 (m, 12H),
1.48-2.30 (m, 16H), 2.36 (m, 2H), 2.56 (s, 3H), 2.65 (m, 1H),
2.96 (m, 1H), 3.18 (m, 2H), 3.36 (m, 2H), 3.65 (m, 6H), 3.78 (s,
3H), 3.91 (m, 1H), 4.02 (m, 1H), 4.25 (m, 1H), 4.63 (m, 1H),
4.72 (m, 2H), 4.87 (m, 1H), 5.02 (d, 1H, J ) 5 Hz), 5.45 (m,
1H), 6.84 (d, 2H, J ) 9 Hz), 7.07 (d, 2H, J ) 9 Hz), 7.27 (d,
1H, J ) 5 Hz), 7.88 (d, 1H, J ) 10 Hz), 8.32 (d, 1H, J ) 5 Hz).
ESI-MS m/e Calcd for C56H87N7O13 1065.6. Found 1066.7
[M + H]+.
Molecular Modeling Calculations on the Spiropseudo-
dipeptide 12. All calculations were run on an SGI workstation
(Fuel, RP14000, 500 MB RAM) under an Irix 6.5 operating
system. The initial conformation of the spiropseudodipeptide
12 was built using the library of fragments available in the
molecular modeling program package Insight II (version
2000.1, Biosym Technologies, San Diego, CA). Molecular
mechanics calculations were carried out using the AMBER
force field as implemented in the DISCOVER module. They
were conducted under vacuum with a distant-dependent
dielectric constant (4r) and a cutoff of 14 Å. The structures
were heated to 1000 K, equilibrated during 10 ps, and slowly
cooled to 300 K stepwise. In each step the temperature was
lowered by 100 K and the system was allowed to remain at
the new temperature for 100 ps. After cooling to 300 K, the
obtained conformations were energy-refined using a conjugated
gradient algorithm with a final gradient of 0.001 kcal/mol as
the convergence criterion. The conformers were stored and
used to start a new simulation at high temperature. This
procedure produced samples of 300 energy-minimized confor-
mations, which were compared with each other to eliminate
those that were identical and those more than 2 kcal/mol above
the global minimum. The protocol was run three times,
employing different starting structures of the spiro derivative
12. Finally, the resulting minima were fully optimized, using
the semiempirical AM1 Hamiltonian54 as implemented in the
MOPAC module of the Insight II program package.
Molecular Mechanics/Dynamics Simulations on the
Aplidine Spirolactam Analogue 4a. Molecular mechanics/
dynamics calculations were performed on Silicon Graphics 02
(MIPS R5000 processor) and Octane (MIPS R12000 processor)
computers running IRIX 6.5, using the well-established
CHARMM program package69 (versions 29b1 and 24b2, re-
spectively) and the CHARMm 23.1 force field.60,62 The initial
structure of the trans-pyruvyl-Pro8 isomer for modeling studies
was produced by the requisite editing of the reported X-ray
crystal structure of didemnin B. Both the CHARMm all-atom
and united-atom representations of the molecule were mod-
eled, the latter having all polar hydrogens explicitly repre-
sented. Charges were assigned to the atoms using the method
described by Gasteiger and Marsili,70 and their suitability was
confirmed by comparison with ESP charges derived from AM1-
optimized small substructures for pyruvyl prolines, dimeth-
[Hiv2]-[(5R)-1-(tert-butoxycarbonyl)-7-[(1R)-1-carbonyl-
3-methylbutyl]-6-oxo-1,7-diazaespiro[4.4]nonane7-9]apli-
dine (5c). White amorphous solid (10 mg, 70%). HPLC
[HyperPrep PEP 100 C18 (A/B, 85:15), φ ) 7 mL/min] tR ) 28.1
min. 1H NMR (300 MHz, CDCl3): δ 0.80-1.07 (m, 24H), 1.08-
1.67 (m, 12H), 1.48 (s, 9H), 1.68-2.30 (m, 10H), 2.41 (m, 1H),
2.55 (s, 3H), 2.68 (m, 1H), 2.94 (m, 1H), 3.07-3.40 (m, 4H),
3.42-3.72 (m, 6H), 3.78 (s, 3H), 3.90 (m, 1H), 4.01 (m, 1H),
4.29 (m, 1H), 4.63 (m, 1H), 4.77 (m, 1H), 4.87 (m, 1H), 5.02 (d,
1H, J ) 5 Hz), 5.25 (m, 1H), 6.84 (d, 2H, J ) 8 Hz), 7.07 (d,
2H, J ) 8 Hz), 7.31 (d, 1H, J ) 10 Hz), 7.50 (d, 1H, J ) 6 Hz),
7.85 (d, 1H J ) 10 Hz). 13C NMR (75 MHz, CDCl3): δ 12.07,
14.25, 17.12, 17.92, 19.15, 21.11, 21.18, 23.80, 24.00, 24.06,
24.86, 25.08, 25.15, 27.58, 28.21, 28.82, 30.33, 31.18, 33.78,
34.28, 35.77, 36.65, 39.07, 39.45, 39.86, 46.91, 48.14, 48.47,
52.62, 55.49, 57.13, 58.42, 66.36, 66.80, 69.14, 70.84, 79.19,
80.55, 114.27, 130.29, 130.59, 154.12, 158.81, 168.25, 169.84,
170.72, 170.80, 170.90, 171.25, 174.89. ESI-MS m/e Calcd for
C57H89N7O14 1095.6. Found 1096.9 [M + H]+.
[Hiv2]-[(5R)-7-[(1R)-1-carboxy-3-methylbutyl]-1-(2-meth-
ylacryloyl)-6-oxo-1,7-diazaspiro[4.4]nonane7-9]aplidine
(5d). White amorphous solid (10 mg, 72%). HPLC [HyperPrep
1
PEP 100 C18 (A/B, 85:15), φ ) 7 mL/min] tR ) 16.4 min. H
NMR (300 MHz, CDCl3): δ 0.85-0.96 (m, 18H), 1.02-1.05 (m,
6H), 1.14-1.45 (m, 12H), 1.49-1.64 (m, 4H), 1.68-1.77 (m,
1H), 1.89-2.05 (m, 3H), 1.99 (s, 3H), 2.10-2.28 (m, 4H), 2.43
(dd, 1H, J ) 8 and 17 Hz), 2.57 (s, 3H), 2.60-2.68 (m, 1H),
2.97 (bs, 1H), 3.13-3.40 (m, 4H), 3.54-3.77 (m, 5H), 3.79 (s,
3H), 3.89-4.07 (m, 2H), 4.27 (m, 1H), 4.64 (m, 1H), 4.73 (m,
1H), 4.88 (m, 1H), 5.03 (d, 1H, J ) 4 Hz), 5.30 (d, 1H, J ) 20
Hz), 5.30-5.39 (m, 1H), 6.84 (d, 2H, J ) 8 Hz), 7.08 (d, 2H,
J ) 8 Hz), 7.29 (s, 1H), 7.88 (d, 1H, J ) 10 Hz), 8.23 (d, 1H,
J ) 7 Hz). ESI-MS m/e Calcd for C56H85N7O13 1063.6. Found
1065.3 [M + H]+.
Nuclear Magnetic Resonance Spectroscopy. Solutions
of 59 mM spiropseudodipeptide 6a in CDCl3, acetone-d6, and

Aplidine and Tamandarin A Analogues
Journal of Medicinal Chemistry, 2004, Vol. 47, No. 23 5711
(6) Raymond, E.; Paz-Ares, L.; Izquierdo, M.; Belanger, K.; Maroun,
J.; Bowman, A.; Anthoney, A.; Jodrell, D.; Armand, J.-P.; Cortes-
Funes, H.; Germa-Lluch, J.; Twelves, C.; Celli, N.; Guzman, C.;
Jimeno J. Phase I Trials with Aplidine, a New Marine Derived
Anticancer Compound. Eur. J. Can. 2001, 37 (Suppl. 6), S32.
(7) Raymond, E.; Ady-Vago, N.; Ribrag, V.; Faivre, S.; Lecot, F.;
Wright, T.; Lo´pez-La´zaro, L.; Guzma´n, C.; Jimeno, J.; Armand,
J. P. Phase I and Pharmacokinetic Study of Aplidine, a Marine
Derived Compound, Given as a 24h Infusion Every 2 Weeks in
Patients (pts) with Advanced Solid Tumors and Non-Hodgkin
Lymphoma (NHL). Ann. Oncol. 2000, 11 (Suppl. 4), 134.
(8) Izquierdo, M. A.; Bowman, A.; Mart´ınez, M.; Cicchella, B.; Lo´pez-
La´zaro, L.; Guzma´n, C.; Germa´, J.; Smyth, J. A Phase I Study
of Aplidine (APL), a Marine Derived Compound, Given as a 1h
Infusion Weekly × 3 in Advanced Solid Tumors and Non-
Hodgkin Lymphoma Patients (pts.). Ann. Oncol. 2000, 11 (Suppl.
4), 134.
(9) Maroun, J.; Belanger, K.; Seymour, L.; Soulieres, D.; Charpen-
tier, D.; Goel, R.; Stewart, D.; Tomiak, E.; Jimeno, J.; Matthews,
S. Phase I Study of Aplidine (APL) in a 1 Hour Daily Infusion
× 5 q 3 Weeks in Patients (pts) with Solid Tumors and Low and
Intermediate Grade Non Hodgkins Lymphomas: A National
Cancer Institute of CanadasClinical Trials Group (NCICsCTG)
Study. Ann. Oncol. 2000, 11 (Suppl. 4), 134.
(10) Urdiales, J. L.; Morata, P.; Nun˜ez De Castro, I.; Sa´nchez-
Jime´nez, F. Antiproliferative Effect of Dehydrodidemnin B
(DDB), a Depsipeptide Isolated from Mediterranean Tunicates.
Cancer Lett. 1996, 102, 31-37.
(11) Lobo, C.; Garc´ıa-Pozo, S. G.; Nun˜ez de Castro, I.; Alonso, F. J.
Effect of Dehydrodidemnin B on Human Colon Carcinoma Cell
Lines. Anticancer Res. 1997, 17, 333-336.
(12) Gomez-Fabre, P. M.; de Pedro, E.; Medina, M. A.; Nun˜ez de
Castro, I.; Marquez, J. Polyamine Contents of Human Breast
Cancer Cells Treated with the Cytotoxic Agents Chlorphe-
niramine and Dehydrodidemnin B. Cancer Lett. 1997, 113, 141-
144.
(13) Erba, E.; Bassano, L.; Di Liberti, G.; Muradore, I.; Chiorino, G.;
Ubezio, P.; Vignati, S.; Codegoni, A.; Desiderio, M. A.; Faircloth,
G.; Jimeno, J.; D’Incalci, M. Cell Cycle Phase Perturbations and
Apoptosis in Tumour Cells Induced by Aplidine. Br. J. Cancer
2002, 86, 1510-1517.
(14) Broggini, M.; Marchini, S. V.; Galliera, E.; Borsotti, P.; Tarabo-
letti, G.; Erba, E.; Sironi, M.; Jimeno, J.; Faircloth, G. T.;
Giavazzi, R.; D’Incalci, M. Aplidine, a New Anticancer Agent of
Marine Origin, Inhibits Vascular Endothelial Growth Factor
(VEGF) Secretion and Blocks VEGF-VEGFR-1 (flt-1) Autocrine
Loop in Human Leukemia Cells MOLT-4. Leukemia 2003, 17,
52-59.
(15) Gajate, C.; An, F.; Mollinedo, F. Rapid and Selective Apoptosis
in Human Leukemic Cells Induced by Aplidine through a Fas/
CD95- and Mitochondrial-Mediated Mechanism. Clin. Cancer
Res. 2003, 9, 1535-1545.
(16) Cuadrado, A.; Garc´ıa-Ferna´ndez, L. F.; Gonza´lez, L.; Suarez, Y.;
Losada, A.; Alcaide, V.; Mart´ınez, T.; Ferna´ndez-Sousa, J. M.;
Sa´nchez-Puelles, J. M.; Mun˜oz, A. Aplidin Induces Apoptosis in
Human Cancer Cells via Glutathione Depletion and Sustained
Activation of the Epidermal Growth Factor Receptor, Src, JNK,
and p38 MAPK. J. Biol. Chem. 2003, 278, 241-250.
(17) Sakai, R.; Stroh, J. G.; Sullins, D. W.; Rinehart, K. L. Seven
New Didemnins from the Marine Tunicate Trididemnum soli-
dum. J. Am. Chem. Soc. 1995, 117, 3734-3748.
(18) Sakai, R.; Rinehart, K. L.; Kishore, V.; Kundu, B.; Faircloth, G.;
Gloer, J. B.; Carney, J. R.; Namikoshi, M.; Sun, F.; Hughes, R.
G., Jr.; Garc´ıa-Gravalos, D.; de Quesada, T. G.; Wilson, G. R.;
Heid, R. M. Structure-Activity Relationships of the Didemnins.
J. Med. Chem. 1996, 39, 2819-2834.
(19) Schmidt, U.; Griesser, H.; Haas, G.; Kroner, M.; Riedl, B.;
Schumacher, A.; Sutoris, F.; Haupt, A.; Emling, F. Synthesis
and Cytostatic Activities of Didemnin Derivatives. J. Pept. Res.
1999, 54, 146-161.
(20) Vervoot, H.; Fenical, W.; Epifanio, R. A. Tamandarins A and B:
New Cytotoxic Depsipeptides from a Brazilian Ascidian of the
Family Didemnidae. J. Org. Chem. 2000, 65, 782-792.
(21) Liang, B.; Richard, D. J.; Portonovo, P. S.; Joullie´, M. M. Total
Syntheses and Biological Investigations of Tamandarins A and
B and Tamandarin A Analogs. J. Am. Chem. Soc. 2001, 123,
4469-4474.
(22) Mayer, S. C.; Carroll, P. J.; Joullie´, M. M. The Cyclic Depsipep-
tide Backbone of the Didemnins. Acta Crystallogr. 1995, C51,
1609-1614.
(23) Jouin, P.; Poncet, J.; Dufour, M.-N.; Aumelas, A.; Pantaloni, A.;
Cros, S.; Franc¸ois, G. Antineoplastic Activity of Didemnin
Congeners: Nordidemnin and Modified Chain Analogues. J.
Med. Chem. 1991, 34, 486-491.
yltyrosine, and N-methyl amides employing MOPAC6. The
Gaussian 94 program package using the Hartree-Fock 6-31G*
basis set was used to check the adequacy of the force field
parametrization for the pyruvamide function. After energy
minimization the initial structures served as the starting point
for NVE molecular mechanics/dynamics calculations at 300 K.
These were performed without the introduction of restraints,
incorporating both implicit (ꢀ ) 1, 4.8, 45, and 80) and explicit
solvent descriptions for H2O and DMSO. For the latter cubic
boxes of 30 Å side length and 1000 TIP3P water molecules71
and of 31 Å side length and 216 DMSO molecules, respec-
tively,72 periodic boundary conditions were applied. After
adequate heating and equilibration of the systems, evolution
times were 10 ns for the all-atom implicit solvent simulations,
40 ns for all united-atom implicit solvent simulations, 4 ns
for the DMSO united-atom explicit solvent simulation, and 2
ns for the united-atom H2O explicit solvent simulation. Struc-
tures were saved periodically from each trajectory for further
analyses. Trajectories were analyzed for bond distances,
torsion angles, and hydrogen bonds by extracting the relevant
information from the CHARMM output files.
Cytotoxicity Measurement. A colorimetric assay using
the sulforhodamine B (SRB) reaction was adapted for a
quantitative measurement of cell growth and viability, follow-
ing the technique described by Skehan et al.62 Cells (A549 and
HT-29) were seeded in 96-well microtiter plates at 5 × 103 cells
per well in aliquots of 195 µL of RPMI medium, and they were
allowed to attach to the plate surface by growing in a drug-
free medium for 18 h. Afterward, samples were added in
aliquots of 5 µL [dissolved in (3:7) DMSO/H2O]. After 48-72
h of exposure, cells were fixed by adding 50 µL of cold 50%
(w/v) trichloroacetic acid and incubating at 4 °C for 60 min.
Then the plates were washed with deionized H2O and dried.
A 100 µL solution of SRB (0.4% w/v in 1% acetic acid) was
added to each microtiter well, and these were incubated at
room temperature for 10 min. Unbound SRB was removed by
washing with 1% acetic acid, the plates were air-dried, and
the bound stain was solubilized with Tris buffer. Optical
densities were read on an automated spectrophotometer plate
reader at a single wavelength of 490 nm. Data analysis was
automatically generated by the high-throughput screening
laboratory information management system.
Acknowledgment. This work was supported by
CICYT (Grant SAF2000-0147), MCYT-FEDER (Grant
BIO2002-2301), Generalitat de Catalunya (Group Con-
solidat 1999SGR0042 and Centre de Refere`ncia en
Biotecnologia), and Pharma Mar, S.A.
Supporting Information Available: Tables of 1H and 13
C
NMR chemical shifts for the spirolactam aplidine analogue 4a,
proton-proton distance measurements from the NOE buildup
curves, calculated interatomic distances, torsion angles, and
hydrogen bonding patterns for 4a, and combustion analysis
data. This material is available free of charge via the Internet
at http://pubs.acs.org.
Note Added after ASAP Posting. This manuscript
was released ASAP on 9/30/2004 with an author missing
from the byline. The correct version was posted on
10/8/2004.
References
(1) Rinehart, K. L. Antitumor Compounds from Tunicates. Med. Res.
Rev. 2000, 20, 1-27.
(2) Vera, M. D.; Joullie, M. M. Natural Products as Probes of Cell
Biology: 20 Years of Didemnin Research. Med. Res. Rev. 2002,
22, 102-145.
(3) Caufield, C. E.; Musser, J. H. Chapter 21. Macrocyclic Immu-
nomodulators. Annu. Rep. Med. Chem. 1989, 25, 195-204.
(4) Rinehart, K. L., Jr.; Gloer, J. B.; Hughes, R. G., Jr.; Renis, H.
E.; McGovren, J. P.; Swynenberg, E. B.; Stringfellow, D. A.;
Kuentzel, S. L.; Li, L. H. Didemnins: Antiviral and Antitumor
Depsipeptides from a Caribbean Tunicate. Science 1981, 212,
933-935.
(24) Hossain, M. B.; van der Helm, D.; Antel, J.; Sheldrick, G. M.;
Sanduja, S. K.; Weinheimer, A. J. Crystal and Molecular
Structure of Didemnin B, an Antiviral and Cytotoxic Depsipep-
tide. Proc. Natl. Acad. Sci. U.S.A. 1988, 85, 4118-4122.
(5) Rinehart, K. L.; Lithgow-Bertelloni, A. M. Dehydrodidemnin B.
WO 9104985, 1991.

5712 Journal of Medicinal Chemistry, 2004, Vol. 47, No. 23
Gutie´rrez-Rodrı´guez et al.
(25) Kessler, H.; Will, M.; Antel, J.; Beck, H.; Sheldrick, G. M.
Conformational Analysis of Didemnins. A Multidisciplinary
Approach by Means of X-ray, NMR, Molecular Dynamics, and
Molecular Mechanics Techniques. Helv. Chim. Acta 1989, 72,
530-555.
(26) Searle, M. S.; Hall, J. G.; Kyratzis, I.; Wakelin, L. P. G. Didemnin
B. Conformation and Dynamics of an Antitumour and Antiviral
Depsipeptide Studied in Solution by 1H and 13C n.m.r. Spec-
troscopy. Int. J. Pept. Protein Res. 1989, 34, 445-454.
(27) Cardenas, F.; Thormann, M.; Feliz, M.; Caba, J. M.; Lloyd-
Williams, P.; Giralt, E. Conformational Analysis of Dehydrodi-
demnin B (Aplidine) by NMR Spectroscopy and Molecular
Mechanics/Dynamics Calculations. J. Org. Chem. 2001, 66,
4580-4584.
(47) Alonso, E.; Lo´pez-Ortiz, F.; del Pozo, C.; Peralta, E.; Mac´ıas, A.;
Gonza´lez, J. Spiro â-Lactams as â-Turn Mimetics. Design,
Synthesis, and NMR Conformational Analisis. J. Org. Chem.
2001, 66, 6333-6338.
(48) Boussard, G.; Marraud, M. â-Turns in Model Dipeptides. An
Infrared Quantitative Analysis with NMR Correlation. J. Am.
Chem. Soc. 1985, 107, 1825-1828.
(49) Gallo, E. A.; Gellman, S. H. Hydrogen-Bond-Mediated Folding
in Depsipeptide Models of â-Turns and R-Helical Turns. J. Am.
Chem. Soc. 1993, 115, 9774-9788.
(50) Jones, I. G.; Jones, W.; North, M. Conformational Analysis of
Peptides and Pseudopeptides Incorporating an endo-(2S,3R)-
Norborn-5-ene Residue as a Turn Inducer. J. Org. Chem. 1998,
63, 1505-1513.
(28) Cardenas, F.; Caba, J. M.; Feliz, M.; Lloyd-Williams, P.; Giralt,
E. Analysis of Conformational Equilibria in Aplidine Using
Selective Excitation 2D NMR Spectroscopy and Molecular
Mechanics/Dynamics Calculations. J. Org. Chem. 2003, 68,
9554-9562.
(29) Shen, G. G.; Zukoski, C. F.; Montgomery, D. W. A Specific
Binding Site in Nb2 Node Lynphoma Cells Mediates the Effects
of Didemnin B, an Immunosuppressive Cyclic Peptide. Int. J.
Immunopharmacol. 1992, 14, 63-73.
(30) Mayer, S. C.; Ramanjulu, M.; Vera, M. D.; Pfizenmayer, A.;
Joullie´, M. M. Synthesis of New Didemnin B Analogs for
Investigations of Structure/Biological Activity Relationships. J.
Org. Chem. 1994, 59, 5192-5205.
(31) Genin, J. M.; Ojala, W. H.; Gleason, W. B.; Johnson, R. L.
Synthesis and Crystal Structure of a Peptidomimetic Containing
the (R)-4.4-Spiro Lactam Type II â-Turn Mimic. J. Org. Chem.
1993, 58, 2334-2337.
(32) Rodr´ıguez, I.; Polanco, C.; Cuevas, F.; Me´ndez, P.; Cuevas, C.;
Gallego, P.; Munt, S.; Manzanares, I.; Garc´ıa-Lo´pez, M. T.;
Gutie´rrez, M.; Herranz, R. Synthetic Methods for Aplidine and
New Antitumoral Derivatives, Methods of Making and Using
them. EP 01945484.2-2404, 2001.
(33) Jou, G.; Gonza´lez, I.; Albericio, F.; Lloyd-Williams, P.; Giralt,
E. Total Synthesis of Dehidrodidemnin B. Use of Uronium and
Phosphonium Salt Coupling Reagents in Peptide Synthesis in
Solution. J. Org. Chem. 1997, 62, 354-366.
(34) Ottenheijm, H. C. J.; De Man, J. H. M. Syntheses of R-Keto Acid
Chlorides. Synthesis 1975, 163-164.
(35) Khalil, E. M.; Subasinghe, N. L.; Johnson, R. L. An Efficient
and High Yield Method for the N-tert-Butoxycarbonyl Protection
of Sterically Hindered Amino Acids. Tetrahedron Lett. 1996, 37,
3441-3444.
(36) Delaney, N. G.; Madison, V. Novel Conformational Distributions
of Methylproline Peptides. J. Am. Chem. Soc. 1982, 104, 6635-
6641.
(37) Beausoleil, E.; Lubell, W. D. Steric Effects on the Amide Isomer
Equilibrium of Prolyl Peptides. Synthesis and Conformational
Analysis of N-Acetyl-5-tert-butylproline N′-Methylamides. J. Am.
Chem. Soc. 1996, 118, 12902-12908.
(38) Lee, Y.-C.; Jackson, P. L.; Jablonsky, M. J.; Muccio, D. D.; Pfister,
R. R.; Haddox, J. L.; Sommers, C. I.; Anantharamaiah, G. M.;
Chaddha, M. NMR Conformational Analysis of cis and trans
Proline Isomers in the Neutrophil Chemoattractant, N-Acetyl-
Proline-Glycine-Proline. Biopolymers 2001, 58, 548-561.
(39) Chernovitz, A. C.; Freedman, T. B.; Nafie, L. A. Vibrational CD
Studies of the Solution Conformation of N-Urethanyl-L-Amino
Acid Derivatives. Biopolymers 1987, 26, 1879-1900.
(40) Dumy, P.; Keller, M.; Ryan, D. E.; Rohwedder, B.; Wo¨hr, T.;
Mutter, M. Pseudo-Prolines as a Molecular Hinge: Reversible
Induction of cis Amide Bonds into Peptide Backbones. J. Am.
Chem. Soc. 1997, 119, 918-925.
(41) Madison, V.; Kopple, K. D. Solvent-Dependent Conformational
Distributions of Some Dipeptides. J. Am. Chem. Soc. 1980, 102,
4855-4863.
(42) Montelione, G. T.; Hughes, P.; Clardy, J.; Scheraga, H. A.
Conformational Properties of 2,4-Methanoproline (2-Carboxy-
2,4-methanoproline) in Peptides: Determination of Preferred
Peptide Bond Conformation in Aqueous Solution by Proton
Overhauser Measurements. J. Am. Chem. Soc. 1986, 108, 6765-
6773.
(51) Luppi, G.; Lanci, D.; Trigari, V.; Garavelli, M.; Garelli, A.;
Tomasini, C. Development and Conformational Analysis of a
Pseudoproline-Containing Turn Mimic. J. Org. Chem. 2003, 68,
1982-1993.
(52) Weiner, S. J.; Kollman, P. A.; Case, D. A.; Singh, U. C.; Ghio,
C.; Alagona, G.; Profeta, S., Jr.; Weiner, P. K. A New Force Field
for Molecular Mechanical Simulation of Nucleic Acids and
Proteins. J. Am. Chem. Soc. 1984, 106, 765-784.
(53) Cornell, W. D.; Cieplak, P.; Bayly, C. I.; Gould, I. R.; Merz, K.
M., Jr.; Ferguson, D. M.; Spellmeyer, D. C.; Fox, T.; Caldwell,
J. W.; Kollman, P. A. A Second Generation Force Field for the
Simulation of Proteins, Nucleic Acids and Organic Molecules.
J. Am. Chem. Soc. 1995, 117, 5179-5197.
(54) Dewar, J. M.; Zoebisch, E. G.; Eamonn, F. H.; Stewart, J. J. P.
AM1: A New General Purpose Quantum Mechanical Molecular
Model. J. Am. Chem. Soc. 1985, 107, 3002-3909.
(55) Rose, G. D.; Gierasch, L. M.; Smith, J. A. Turns in Peptides and
Proteins. Adv. Protein Chem. 1985, 37, 1-109.
(56) Wilmot, C. N.; Thorton, J. M. â-Turns and Their Distortions a
Proposed New Nomenclature. Protein Eng. 1990, 3, 479-493.
(57) Chou, K.-C. Prediction of Tight Turns and Their Types in
Proteins. Anal. Biochem. 2000, 286, 1-16.
(58) Perczel, A.; Hollo´si, M.; Sa´ndor, P.; Fasman, G. D. The Evalu-
ation of Type I and Type II â-Turn Mixtures. Int. J. Pept. Protein
Res. 1993, 41, 223-236.
(59) Neuhaus, D.; Williamson, M. P. The Nuclear Overhauser Effect
in Structural and Conformational Analysis, 2nd ed.; John Wiley
and Sons: New York, 2000.
(60) Momany, F. A.; Rone, R. Validation of the General Purpose
QUANTA 3.2 CHARMm Force Field. J. Comput. Chem. 1992,
13, 888-900.
(61) Momany, F. A.; Rone, R.; Kunz, H.; Frey, R. F.; Newton, S. Q.;
Scha¨fer, L. Geometry Optimization, Energetics and Solvation
Studies on Four and Five-Membered Cyclic and Disulfure
Bridged Peptides, Using the Programs QUANTA 3.3 and
CHARMm 22. J. Mol. Struct.: THEOCHEM 1993, 286, 1-18.
(62) Skehan, P.; Storeng, R.; Scudiero, D.; Monks, A.; McMahon, J.;
Vistica, D.; Warren, J. T.; Bokesch, H.; Kenney, S.; Boyd, M. R.
New Colorimetric Cytotoxicity Assay for Anticancer-Drug Screen-
ing. J. Natl. Cancer Inst. 1990, 82, 1107-1112.
(63) Shen, G. K.; Zukoski, C. F.; Montgomery, D. W. A Specific
Binding Site in Nb2 Node Lymphoma Cells Mediates the Effects
of Didemnin B, an Immunosuppresive Cyclic Peptide. Int. J.
Immunopharmacol. 1992, 14, 63-73.
(64) Fischer, G. Peptidyl-Prolyl cis/trans Isomerases and Their
Effectors. Angew. Chem., Int. Ed. Engl. 1994, 33, 1415-1436.
(65) Schmid, F. X. Prolyl Isomerases. Adv. Protein Chem. 2002, 59,
243-282.
(66) Schiene-Fischer, C.; Yu, C. Receptor Accessory Folding Helper
Enzymes: The Functional Role of Peptidyl Prolyl cis/trans
Isomerases. FEBS Lett. 2001, 495, 1-6.
(67) Gothel, S. F.; Marahiel, M. A. Peptidyl-Prolyl cis-trans Isomeras-
es, a Superfamily of Ubiquitous Folding Catalyst. Cell. Mol. Life
Sci. 1999, 55, 423-36.
(68) Shaw, P. E. Peptidyl-Prolyl Isomerases:
A New Twist to
Transcription. EMBO Rep. 2002, 3, 521-526.
(69) Brooks, B. R.; Bruccoleri, R. E.; Olafson, B. D.; States, D. J.;
Swaminathan, S.; Karplus, M. A Program for Macromolecular
Energy, Minimization and Dynamics Calculations. J. Comput.
Chem. 1983, 4, 187-217.
(43) Magaard, V. W.; Sanchez, R. M.; Bean, J. W.; Moore, M. L. A
Convenient Synthesis of the Conformationally Constrained
Amino Acid 5,5-Dimehtylproline. Tetrahedron Lett. 1993, 34,
381-384.
(44) Halab, L.; Lubell, W. D. Effect of Sequence on Peptide Geometry
in 5-tert-Butylprolyl Type VI â-Turn Mimics. J. Am. Chem. Soc.
2002, 124, 2474-2484.
(45) Kessler, H. Conformation and Biological Activity of Cyclic
Peptides. Angew. Chem., Int. Ed. Engl. 1982, 21, 512-523.
(46) Belvisi, L.; Gennari, C.; Mielgo, A.; Potenza, D.; Scolastico, C.
Conformational Preferences of Peptides Containing Reverse-
Turn Mimetic Bicyclic Lactams: Inverse γ-Turns versus Type-
II′ â-TurnssInsights into â-Hairpin Stability. Eur. J. Org. Chem.
1999, 389-400.
(70) Gasteiger, J.; Marsili, M. Iterative Partial Equalization of Orbital
Electronegativity a Rapid Access to Atomic Charges. Tetrahe-
dron 1980, 36, 3219-3288.
(71) Jorgensen, W. L.; Chandrasekhar, J.; Madura, J. D.; Impey, R.
W.; Klein, M. L. Comparison of Simple Potential Functions for
Simulating Liquid Water. J. Chem. Phys. 1983, 79, 926-935.
(72) Liu, H.; Mueller-Plathe, F.; van Gunsteren, W. F. A Force Field
for Liquid Dimethyl Sulfoxide and Physical Properties of Liquid
Dimethyl Sulfoxide Calculated Using Molecular Dynamics Simu-
lation. J. Am. Chem. Soc. 1995, 117, 4363-4366.
JM040788M