Stellatolides, a new cyclodepsipeptide family from the sponge ecionemia acervus: Isolation, solid-phase total synthesis, and full structural assignment of stellatolide a

Article abstract of DOI:10.1021/ja502744a

The marine environment is a rich source of metabolites with potential therapeutic properties and applications for humans. Here we describe the first isolation, solid-phase total synthesis, and full structural assignment of a new class of cyclodepsipeptides from the Madagascan sponge Ecionemia acervus that shows in vitro cytotoxic activities at submicromolar concentrations. Seven structures belonging to a new family of compounds, given the general name stellatolides, were characterized. The sequence and stereochemistry of all the amino acids in these molecules were established by a combination of spectroscopic analysis, chemical degradation, and derivatization studies. Furthermore, the complete structure of stellatolide A was confirmed by an efficient solid-phase method for the first total synthesis and the full structural assignment of this molecule, including the asymmetric synthesis of the unique B-hydroxy acid moiety (Z)-3-hydroxy-6,8-dimethylnon-4-enoic acid.

Full text of DOI:10.1021/ja502744a

Article
pubs.acs.org/JACS
Stellatolides, a New Cyclodepsipeptide Family from the Sponge
Ecionemia acervus: Isolation, Solid-Phase Total Synthesis, and
Full Structural Assignment of Stellatolide A
María Jesus Martín,† Raquel Rodríguez-Acebes,† Yesica García-Ramos,‡,§ Valentín Martínez,†
́
́
Carmen Murcia,† Isabel Digon,† Isabel Marco,†,▽ Marta Pelay-Gimeno,‡,§ Rogelio Fernandez,†
́
́
Fernando Reyes,†,▲ Andres M. Francesch,† Simon Munt,† Judit Tulla-Puche,*,‡,§
́
Fernando Albericio,‡,§,∥,⊥ and Carmen Cuevas*,†
†Research and Development, PharmaMar, S. A., Avenida de los Reyes 1, 28770 Colmenar Viejo, Madrid, Spain
‡Institute for Research in Biomedicine, Baldiri Reixac 10, 08028 Barcelona, Spain
§Biomedical Research Networking Centre on Bioengineering, Biomaterials and Nanomedicine, Barcelona Science Park,
Baldiri Reixac 10, 08028 Barcelona, Spain
∥Department of Organic Chemistry, University of Barcelona, Martí i Franques
̀
1-11, 08028 Barcelona, Spain
⊥School of Chemistry & Physics, University of Kwazulu-Natal, Durban 4001, South Africa
S
* Supporting Information
ABSTRACT: The marine environment is a rich source of
metabolites with potential therapeutic properties and applica-
tions for humans. Here we describe the first isolation, solid-
phase total synthesis, and full structural assignment of a new
class of cyclodepsipeptides from the Madagascan sponge
Ecionemia acervus that shows in vitro cytotoxic activities at
submicromolar concentrations. Seven structures belonging to a
new family of compounds, given the general name stellatolides,
were characterized. The sequence and stereochemistry of all
the amino acids in these molecules were established by a
combination of spectroscopic analysis, chemical degradation, and derivatization studies. Furthermore, the complete structure of
stellatolide A was confirmed by an efficient solid-phase method for the first total synthesis and the full structural assignment of this
molecule, including the asymmetric synthesis of the unique β-hydroxy acid moiety (Z)-3-hydroxy-6,8-dimethylnon-4-enoic acid.
INTRODUCTION
■
Cyclodepsipeptides from marine organisms have emerged
as a key class of bioactive compounds.1 A range of biological
activities, including cytotoxic, antiviral, and/or antifungal
properties, have been observed for cyclodepsipeptides isolated
from various marine sponges. These compounds include
callipeltins A,2 B−C,3 D−E,4 F−I,5 and J−M,6 papuamides
A−D7 and E−F,8 neamphamides A9 and B,10 theopapuamides
A11 and B−D,12 mirabamides A−D13 and E−H,14 homo-
phymines A,15 B−E and A1−E1,16 and pipecolidepsins A−C.17
From a chemical point of view, these cyclodepsipeptides show
intriguing structures and their synthesis poses a challenge.
During the course of our ongoing screening program to
Figure 1. Chemical structure and selected HMBC correlations for
identify new antitumor agents from marine organisms, extracts
stellatolide A (1).
of the Madagascan sponge Ecionemia acervus were found to
display cytotoxic activity against human tumor cell lines.
unique structural features compared to other marine cyclo-
Bioassay-guided fractionation of the aqueous extract of a
depsipeptides. The 22-membered macrocycle of stellatolide A is
specimen of this organism yielded a novel family of cyclic
depsiundecapeptides named stellatolides as bioactive compo-
nents. Of these, stellatolide A (1) (Figure 1) was the most
Received: March 18, 2014
abundant peptide isolated and was found to have a number of
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ja502744a | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
a
1
Table 1. H and 13C NMR Data for Stellatolide A
δC,
ppm
δH, ppm
(mult, J in Hz)
δC,
ppm
δH, ppm
(mult, J in Hz)
δC,
ppm
δH, ppm
(mult, J in Hz)
δC,
ppm
δH, ppm
(mult, J in Hz)
b
b
b
b
position
position
position
position
N-MeAla
171.6
diMeGln
173.9
3
40.5 1.67 (m), 1.25 (m)
26.1 1.67 (m)
3
72.5 4.61 (d, 4.2)
176.5
1
1
4
4
2
51.8 5.44 (q, 7.1)
13.2 1.29 (d, 7.1)
29.5 2.69 (s)
β-MeOTyr
2
59.0 4.11 (d, 9.2)
37.0 2.46 (m)
17.2 1.31 (d, 6.9)
44.9 2.71 (m)
14.0 1.33 (d, 7.2)
182.1
5
23.6 0.96 (d, 6.6)
21.4 0.91 (d, 6.0)
NH
NH2
3
3
5-Me
7.07 (br s), 6.80
c
(br s)
NMe
3-Me
4
NH
1
7.21 (d, 8.4)
HDMN
174.8
Gly
1
2
1
2
3
4
169.9
4-Me
5
172.2
45.3 2.59 (dd, 13.8,
9.4), 2.37 (dd,
13.7, 3.9)
53.7 4.93 (d, 9.5)
84.3 4.55 (d, 9.5)
129.4
NH
NH2
7.84 (br s), 7.10
2
44.1 3.96 (d, 17.1), 3.51
(d, 17.1)
3
4
5
66.5 4.88 (m)
c
(br s)
5/5′
6/6′
7
131.4 7.18 (d, 8.6)
116.0 6.80 (d, 8.6)
158.7
Thr(NH2)
NH
9.08 (m)
130.9 5.37 (dd, 10.8,
9.0)
1
171.2
2
56.4 4.49 (m)
49.1 3.97 (m)
16.8 1.46 (d, 6.9)
Ser(Me)
139.4 5.22 (dd, 10.8,
10.8)
MeO
NH
56.9 3.11 (s)
8.15 (d, 10.2)
N-MeGln
3
1
2
3
172.7
6
31.3 2.64 (m)
22.0 0.94 (d, 6.6)
48.1 1.15 (m)
4
55.2 4.51 (m)
6-Me
7
NH
NH2
71.5 3.80 (dd, 10.0, 7.1),
3.77 (dd, 10.0,
7.1)
1
170.7
2
55.6 4.79 (dd, 9.5, 6.6)
25.0 1.64 (m), 1.27 (m)
32.0 1.68 (m)
Thr2
3
1
2
3
4
173.3
MeO
NH
59.5 3.40 (s)
7.83 (d, 8.9)
8
26.9 1.57 (m)
4
60.0 4.35 (d, 2.4)
67.4 4.48 (m)
20.4 1.21 (d, 6.4)
8-Me
9
23.5 0.90 (d, 6.5)
23.0 0.90 (d, 6.4)
5
177.7
Thr1
NH2
7.78 (br s), 7.34
(br s)
1
172.8
2
57.3 5.21 (d, 3.1)
71.6 5.60 (qd, 6.3, 3.1)
14.8 1.19 (d, 6.3)
8.92 (d, 10.3)
NMe
30.3 2.93 (s)
NH
8.49 (d, 6.2)
β-HOAsn
3
Leu
4
1
2
174.0
1
2
171.9
NH
49.2 4.73 (dd, 10.8, 3.5)
59.1 4.75 (d, 4.2)
a
b
NMR data were taken in CD3OD at 500 and 125 MHz. NH and NH2 chemical shifts, multiplicities, and correlations were determined through
c
experiments in CD3OH. Assignments of NH2 are interchangeable.
made up of seven subunits and is connected through an amide
bond to a linear side chain comprising a complex tetrapeptide
terminated by a β-hydroxy acid. Four unusual nonproteinogenic
α-amino acid residues are present in the molecule, namely,
β-methoxytyrosine, 3,4-dimethylglutamine, 2,3-diaminobutanoic
acid, and 2,4-diamino-3-hydroxy-4-oxobutanoic acid (β-hydroxy-
asparagine). Its structure also contains a unique β-hydroxy acid
moiety that blocks the N-terminus, while the C-terminus is
lactonized via a D-allo-Thr residue. Here we report the isolation
and structural elucidation of stellatolides. In addition, as part
of our efforts to improve the availability of stellatolide A and
explore its potential therapeutic applications, we describe an
efficient solid-phase method for the first total synthesis of this
compound in multigram scale. This last achievement allowed
full stereochemical assignment, as well as the development
of an asymmetric synthesis of the unique β-hydroxy acid moiety,
(Z)-3-hydroxy-6,8-dimethylnon-4-enoic acid.
2.93 and 2.69 ppm, and several other methyl groups between
1
δH 1.46 and 0.90 ppm. Extensive analysis of the H and 13C
NMR data of 1, including correlation spectroscopy (COSY),
total correlation spectroscopy (TOCSY), heteronuclear single
quantum coherence (HSQC), and heteronuclear multiple bond
correlation (HMBC) spectra, revealed the presence of the
common amino acids Leu, Gly, and two Thr residues, together
with the following seven nonproteinogenic amino acid
moieties: N-methylalanine (N-MeAla), β-methoxytyrosine (β-
MeOTyr), N-methylglutamine (N-MeGln), O-methoxyserine
[Ser(Me)], 3,4-dimethylglutamine (diMeGln), 2,3-diaminobu-
tanoic acid [Thr(NH2)], and β-hydroxyasparagine (β-HOAsn).
The presence of the amino-linked (Z)-3-hydroxy-6,8-dimethyl-
non-4-enoic acid (HDMN) was also inferred from the
NMR spectra. Thus, a contiguous spin system comprising an
oxymethine signal (δH 4.88 ppm, δC 66.5 ppm), a double bond
at δH 5.37 and 5.22 ppm (δC 130.9 and 139.4 ppm), two
methylene groups (δH 2.59/2.37 ppm, δC 45.3 and δH 1.15 ppm,
δC 48.1 ppm), two methines at δH 2.64 (δC 31.3 ppm) and
1.57 ppm (δC 26.9 ppm), and three methyl groups at δH 0.94,
0.90, and 0.90 ppm (δC 22.0, 23.5, and 23.0 ppm, respectively)
was deduced. The sequence of this residue from C2 to C9 was
established by COSY and HMBC correlations. The C4−C5
olefin was assigned a Z geometry on the basis of a coupling
constant of 10.8 Hz between H4 and H5. Correlations in the
HMBC spectrum between the methylene at 2.59/2.37 ppm and
the carbon resonance at 174.8 ppm indicated the presence of
the carbonyl group at C1 of this residue.
RESULTS AND DISCUSSION
■
Isolation and Structural Elucidation of Stellatolide A.
The molecular formula of 1 was established as C66H107N15O22
by positive-ion high-resolution electrospray ionization time-of-
flight mass spectrometry [(+)-HR-ESI-TOF-MS], where the
[M + H]+ molecular ion was observed at m/z 1462.7764 (calcd
for C66H108N15O22, 1462.7787). The peptidic nature of this
molecule was evident from the presence of numerous carbonyl
carbons (δC 174.8−169.9 ppm) in the 13C NMR spectrum
1
obtained in CD3OD (see Table 1). In addition, the H NMR
spectrum showed signals corresponding to two methoxy groups
at δH 3.40 and 3.11 ppm, signals for two N-methyl groups at δH
These units were sequenced by use of a combination of
HMBC and rotating spectroscopy (ROESY) data. Long-range
B
dx.doi.org/10.1021/ja502744a | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 2. Chemical structures of stellatolides B−G (2−7).
correlations observed between α-protons and NH protons or
N-Me protons and carbonyl carbons of adjacent amino acids
(Figure 1) allowed us to establish the sequence as N-MeAla−
β-MeOTyr−N-MeGln−Leu−Gly−Ser(Me)−Thr1−diMeGln−
Thr(NH2)−Thr2−β-HOAsn. The presence of an ester bond
between the carbonyl group of N-MeAla and the hydroxyl
group of the Thr1 residue was suggested by the downfield
chemical shift of the H3 oxymethine proton (δH 5.60 ppm) of
this last residue7 and confirmed by the presence of an HMBC
correlation between this proton and the N-MeAla carbonyl
carbon at δC 171.6 ppm. In addition, the acylation of β-HOAsn
by HDMN was evidenced by HMBC correlation between the
H2 signal of β-HOAsn at δH 4.75 ppm and the carbonyl group
of HDMN at δC 174.8 ppm.
1, whereas the C4−C5 olefin was assigned a Z geometry on the
basis of a coupling constant of 10.9 Hz between H4 and H5.
Stellatolide C (3) is closely related to 1. The [M + H]+
at m/z 1480.7898 (calcd for C66H110N15O23, 1480.7899) in the
(+)-HR-ESI-TOF-MS mass spectrum, 18 mass units higher
than 1, together with 1H and 13C NMR data pointed to 3 being
1
the acyclic derivative of 1. The H and 13C NMR spectra were
very similar to those of 1, except for the signals assigned to H3
and C3 of the Thr1 residue. In 3 these signals shifted upfield
to δ 4.44 and 67.5 ppm (CD3OD) respectively, relative to the
same signal of 1 (δH 5.60 ppm and δC 71.6 ppm).
As deduced from (+)-HR-ESI-TOF-MS and NMR data,
stellatolide D (4) (m/z 1446.7846 [M + H]+; calcd for
C66H108N15O21, 1446.7844) differed from 1 in that β-HOAsn
1
By use of liquid chromatography/mass spectrometry
(LC-MS) and after derivatization with Marfey’s reagent, L-FDAA
[Nα-(4,5-dinitrophenyl-5-fluoro)-L-alaninamide],18 the absolute
configurations of the amino acid residues were determined by
chromatographic comparison of the acid hydrolysate of 1 (6 N
HCl, 110 °C, 15 h) with appropriate amino acid standards. We
assigned the L configuration for N-MeAla, Leu, and N-MeGln
residues and the D configuration for Ser(Me) and the two
D-allo-Thr residues. The configurations of the nonproteinogenic
amino acids diMeGln, Thr(NH2), and β-HOAsn were assigned
as (2S,3S,4R), (2S,3S), and (2R,3S), respectively, by synthetic
preparation of the corresponding standards.19 To define the
chirality of the β-MeOTyr residue, an aliquot of 1 was ozonized.
This residue was then transformed to the corresponding
β-MeOAsp derivative after work-up by hydrogen peroxide.20
Hydrolysis, analysis by Marfey’s method and comparison
with synthetic sample revealed (2R,3R) as the stereochemistry.
The remaining stereochemical uncertainty of 1 concerns the
configuration of the two chiral centers in the moiety of HDMN.
Structural Elucidation of Stellatolides B−G. Stellatolide
B (2) (Figure 2) has a molecular formula of C65H105N15O22,
according to the (+)-HR-ESI-TOF-MS analysis (m/z
1448.7653 [M + H]+; calcd for C65H106N15O22, 1448.7637).
was replaced by an Asn residue. In the H NMR spectrum
of 4, the signal assigned to the β-hydroxy group of β-HOAsn
(δH 4.61 ppm) was absent, and a new methylene group at δH
2.96/2.73 ppm in the 1H spectrum and δC 37.2 ppm in the 13
C
data was observed. The stereochemistry of the amino acid
residues was established by Marfey’s analysis, as described for 1,
and resulted to be the same as 1 for the similar amino acids.
An R configuration was established for the new Asn residue
after Marfey’s derivatization and comparison with both R and S
standards.
Stellatolide E (5) was concluded to have a molecular formula
of C65H105N15O22, as deduced from (+)-HR-ESI-TOF-MS
data (m/z 1448.7630 [M + H]+; calcd for C65H106N15O22,
1448.7637). NMR data indicated that 5 differs from 1 in the
absence of a methyl group at the 3,4-dimethylglutamine
residue. The analysis of COSY, TOCSY, and HSQC experi-
ments allowed us to infer that the missing methyl group was at
C4, and the compound has therefore a 3-methylglutamine
residue (also present in callipeltins J and K).6 The absolute
configuration of the amino acid residues was determined by
Marfey’s method and was found to be the same as in 1, except
for 3-methylglutamine, whose absolute configuration was not
determined due to the unavailability of appropriate standards.
The molecular formula C66H104N14O21 of stellatolide F (6),
deduced from (+)-HR-ESI-TOF-MS data (m/z 1429.7636
[M + H]+; calcd for C66H105N14O21, 1429.7579), indicated the
loss of one ammonia molecule from 4. The NMR spectra for 6
were consistent with those of 4, except for the presence of a
signal corresponding to a methyl group shifted downfield to δ
1.65 ppm and an olefinic signal present at δH 6.79 ppm, indicating
that the Thr(NH2) in 4 was replaced by didehydroaminobutyric
acid (Dhb) in 6. A cross-peak observed in the ROESY
experiment (in CD3CN−H2O 5:1) between the NH at δH
1
This molecular formula and comparison of H and 13C NMR
spectra with those of 1 revealed that the only difference
between the compounds was the absence of a methyl group at
C8 in the HDMN moiety of the latter. This observation was
supported by analysis of the 1D and 2D NMR data. HPLC
analysis of the acid hydrolysate derivatized with Marfey’s
reagent and comparison with appropriate amino acid samples
indicated that all the amino acid residues in 2 had the same
absolute configuration as in 1. The configurations at C3 and C6
of the (Z)-3-hydroxy-6-methylnon-4-enoic acid (HMN) were
assumed to be identical to those of the similar residue found in
C
dx.doi.org/10.1021/ja502744a | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
8.98 ppm and the methyl group at δH 1.65 ppm indicated a
Z geometry for the double bond of the latter residue.
Scheme 1. Hexapeptide Synthesis following (A) Strategy 1
and (B) Strategy 2
a
Similarly, and as deduced from (+)-HR-ESI-TOF-MS data
and NMR spectra, stellatolide G (7) (m/z 1445.7607 [M + H]+;
calcd for C66H105N14O22, 1445.7528) differs from 1 by the re-
placement of Thr(NH2) by a Dhb residue. The close resem-
blances of the 1H and 13C NMR chemical shifts of all the amino
acid residues of 6 and 7 with the corresponding values of 1
suggest that chiral centers in the two latter molecules have the
same absolute configuration as in 1.
Total Synthesis. In order to establish the configuration of 1
and to further explore its pharmaceutical development and
in vivo evaluation, we undertook the development of a solid-
phase synthetic strategy for its preparation. To the best of our
knowledge, stellatolide A is one of the few complex β-OMeTyr-
containing peptides whose synthesis has been attempted.21
These kinds of peptide are not stable in moderate concentra-
tions of acid, thus leading to the elimination of MeOH from
the β-MeOTyr residue to give didehydrotyrosine. In addition,
and as observed in other depsipeptides of related families,
careful attention needs to be paid to transacylation, nitrile
formation from Asn/Gln residues, and both pyro-diMeGln and
diketopiperazine (DKP) formation.17b
a
Conditions: (a) piperidine−DMF (1:4); (b) Fmoc-N-MeGln
anhydride, DIPEA; (c) H-D-Tyr(tBu)-OAllyl, HATU, DIPEA; (d)
Fmoc-AA-OH, HATU, DIPEA; (e) Fmoc-D-Tyr-OAIIyl, DIPEA,
CH2Cl2.
Solid-Phase Strategies. First our efforts focused on
determining the best methodology and devising the optimal
anchoring, and thus the cyclization point (Figure 3). The
In the second strategy, Fmoc-D-Tyr-OAllyl was loaded on a
2-chlorotrityl chloride (2-CTC) resin through the phenol, and
the peptide was elongated until it reached the same hexapeptide
as in the previous strategy. In both cases, Boc-D-Thr-OH was
used at the branching position, since the linear arm would not
be constructed at this point. In both strategies, after performing
the esterification with Alloc-N-MeAla-OH 1324 and removing
the allyloxycarbonyl (Alloc) and allyl groups, cyclization with
benzotriazol-1-yloxytri(pyrrolidino)phosphonium hexafluoro-
phosphate (PyBOP, 2 equiv), 1-hydroxy-7-azabenzotriazole
(HOAt, 2 equiv), and N,N-diisopropylethylamine (DIPEA,
4 equiv) in N,N-dimethylformamide (DMF) was allowed to
proceed for 12 h. In both strategies, the cyclic peptide, which
was cleaved from the corresponding resins with <5% TFA, was
obtained with acceptable purity.
Figure 3. Anchoring and cyclization points (dashed lines).
In order to explore strategy 3, Fmoc-Gly-OH was anchored
to 2-CTC resin, and the peptide was elongated by introducing
D-Ser as the following residue, which was used instead of
D-Ser(Me), and Boc-D-Thr, N-MeAla (through esterification),
D-Tyr, and N-MeGln. The linear peptide was obtained
as a main product but with considerable formation of DKP
(N-MeAla−D-Tyr),22 even when exposure to piperidine was
minimized (1 × 3 min) and coupling of the following amino
acid (N-MeGln) was performed under neutral conditions
[N,N′-diisopropylcarbodiimide (DIPCDI)/HOAt]. In view of
these results, the Alloc group, which is removed in the absence
of base, was chosen for protecting the amino function of the
D-Tyr residue.
Once elongation of the sequence and cyclization had been
addressed, we next studied formation of the central core: elonga-
tion of the exocyclic arm, paying special attention to diMeGln,
in conjunction with esterification using a model system with Tyr
and Ile in place of β-MeO−D-Tyr and diMeGln, respectively.
First, the pentapeptide was constructed as in strategy 1, for
which superior results had been obtained in construction of the
cycle. Then Fmoc-D-allo-Thr-OH was used in place of Boc-D-
Thr-OH to allow growth of the exocyclic arm. Nevertheless,
presence of N-MeGln and the modified Tyr in 1 allows
cyclization on solid phase through side-chain anchoring of these
two residues (strategies 1 and 2), whereas Gly as a C-terminus
entails performing the cyclization in solution (strategy 3). None
of these candidate points considers the esterification reaction as
the cyclization step because of the slower kinetics of the ester
formation, especially in a cyclization reaction, which implies a
major risk of racemization.
In the first strategy, following the work of Lipton and co-
workers on callipeltin B,21a we introduced Fmoc-N-MeGlu
anhydride (Fmoc = fluorenylmethyloxycarbonyl) on a Sieber
resin in a regioselective manner. This approach allows the
release of the Gln-containing peptide by mild trifluoroacetic
acid (TFA) cleavage. As the cyclization between N-terminal
Tyr and C-terminal Gln would be highly unfavorable because
of competing DKP (MeAla−β-OMe-D-Tyr)22 formation, H-D-
Tyr-OAllyl23 was coupled as a second residue to the α-carboxyl
of Gln. After release of the Fmoc group, the peptide was
elongated by standard C to N Fmoc chemistry until the
hexapeptide was reached (Scheme 1). HPLC analysis of an
aliquot after cleavage from the resin showed a purity of 88%.
D
dx.doi.org/10.1021/ja502744a | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
even if the cycle was first formed, when the Ile residue was de-
protected, a negative ninhydrin was obtained, showing complete
O to N transacylation. The only way it could be minimized
was by performing the esterification with N-MeAla after the
Fmoc-Ile-OH introduction. In this manner, and although ester
formation was more difficult due to steric hindrance of the
Fmoc group [10 equiv of N-MeAla, 5 equiv of DIPCDI, and
0.2 equiv of 4-dimethylaminopyridine (DMAP) were needed to
show completion], the transacylation reaction was minimized
and no racemization of N-MeAla was detected by Marfey’s test.
This strategy (first strategy in Figure 3) was finally tested with
the key amino acid Alloc-β-MeO-D-Tyr(Trt)-OH 8. Although
the elongation proceeded as planned, the main problem arose
at the cyclization step. It was first attempted with PyBOP
(2 equiv), HOAt (2 equiv) and DIPEA (4 equiv) in DMF.
However, after 2 h, a peak with the same tR as starting material
but a mass that did not correspond to either linear or cyclic
peptide (32 units less than the latter) appeared. Moreover, the
UV profile was altered drastically from a smooth absorbance at
274.4 nm to an intense peak at 307.7 nm. Numerous attempts
involving harsh conditions [(7-azabenzotriazol-1-yl)oxy]tris-
(pyrrolidino) phosphonium hexafluorophosphate (PyAOP)/
HOAt/DIPEA, Lipton’s conditions, PyAOP/HOAt/collidine;21a
O-(7-azabenzotriazol-1-yl)-1,1,3,3-tetramethyluronium hexa-
fluorophosphate (HATU)/HOAt/DIEA microwave irradiation,
heating] gave the same outcome. In contrast, milder conditions
(DIPCDI/HOAt, DIPCDI/OxymaPure, Pfp activation) gave
two main peaks, one identified as the activated peptide and the
other lacking 18 units and upon first glance appearing to be the
desired cyclic product. The UV profile, however, was completely
different. The use of distinct solvents [DMF, DMF−CH2Cl2
(1:1), tetrahydrofuran (THF)] did not change the outcome
of the reaction, and the reaction was driven to the byproduct
(32 units less) previously observed when the harsher conditions
were used.
instead of the slower cyclization reaction. Coupling studies
with a model tripeptide designed to simulate the actual peptide
supported this notion. Reactivity of the β-MeO-D-Tyr carboxyl
functionality was studied by testing coupling conditions to form
a standard amide bond with Fmoc-N-MeGlu anhydride coupled
to a Sieber resin and H-Gly-OMe as a third amino acid. A 97:3
ratio (product/side product) was obtained when the DIPCDI/
HOAt system was used. On the basis of all these previous
experiments, we concluded that strategy 3 was the method of
choice for the synthesis of 1.
Further fine-tuning was needed in order to suppress DKP
formation, already devised in the simplified model. In addition
to using Alloc protection for the β-MeO-D-Tyr residue, other
conditions were tested, such as the tandem reaction with Fmoc-
N-MeGln-OH. However, only prior elongation of the exocyclic
arm before cycle completion effectively eliminated DKP
formation. This observation can be explained by the presence
of the exocyclic arm, which could imply a hindered environment
that slows down reactions, and therefore certain selectivity
between straight and side reactions is accomplished.25
Synthesis of Building Blocks and (Z)-3-Hydroxy-6,8-
dimethylnon-4-enoic Acid. Before peptide assembly on solid
phase, six of the component building blocks (Figure 4) were
These side products were formed when β-MeO-D-Tyr was
activated, thus preventing reaction of the carboxylic acid
and completion of the cycle. Scheme 2 shows the proposed
Figure 4. Chemical structures of amino acid building blocks.
stereoselectively synthesized and appropriately protected by an
Fmoc strategy, by modification and/or optimization of protocols
described in the literature: (2R,3R)-Alloc-β-MeO-Tyr(Trt)-OH
8,26 (2S,3S,4R)-Fmoc-3,4-diMeGln-OH 9,27 (2S,3S)-Fmoc-
Thr(NHBoc)-OH 10,19b (2R,3S)-Fmoc-3-OH-Asn-OH 11,28
Fmoc-N-Me-L-Gln 12,29 and N-Alloc-N-Me-L-Ala 13.30
Scheme 2. Demethoxylation of β-MeO-Tyr
The stereoisomerism of the unprecedented HDMN moiety
could not be determined by spectroscopic analysis, chemical
degradation, or derivatization due to the lack of appropriate
standards. Our group developed an asymmetric synthesis to
enantioselectively prepare the four possible stereoisomers:
(3R,6R), (3S,6R), (3R,6S), and (3S,6S). The synthesis involved
the use of the chiral Evans’ oxazolidinone chemistry to form the
C2/C3 bond as the key step (Scheme 3).
The use of commercially available pseudoephedrine as a
chiral auxiliary in the presence of LiCl allows highly
diastereoselective alkylations with alkyl halides according to
Myers’ chemistry.31 (1R,2R)-(−)- and (1S,2S)-(+)-pseudo-
ephedrinepropionamides 14 and 15 were alkylated with 1-iodo-
2-methylpropane 16 to give (S,R,R)-17 and (R,S,S)-18, re-
spectively. The latter two compounds were then converted to
the corresponding primary alcohols with LiH2N·BH3, and
subsequent oxidation with pyridinium chlorochromate (PCC)
yielded aldehydes (S)-19 and (R)-20. A variation of the Horner−
Wadsworth−Emmons reaction over (S)-19 and (R)-20 that
mechanism for the demethoxylation reaction. With activated
carboxylate, an intermediate oxazolone is formed, which later
undergoes demethoxylation to give the demethoxylated oxazolone.
As the demethoxylation reaction could not be prevented or
minimized, it was not possible to obtain 1 by cyclizing at this
position (first and second strategies in Figure 3). We envisaged
that the solution would be to achieve the participation of the
activated carboxyl of β-MeO-D-Tyr in a regular amide bond
E
dx.doi.org/10.1021/ja502744a | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
coupling and HOBt to minimize dehydration reactions; (viii)
cleavage of the protected peptide from the resins; and (ix)
cyclization in solution.
Scheme 3. Synthesis of Four Diastereoisomers of
(Z)-3-Hydroxy-6,8-dimethylnon-4-enoic acid
a
Elongation of the peptide chain was accomplished by step-
wise coupling of the corresponding Fmoc-protected amino
acids; the amino acids were activated with HOBt or HOAt and
DIPCDI in DMF for amide formation. Progress of the amino
acid coupling was monitored by the ninhydrin test (Kaiser test),
except when couplings were performed over N-Me amino acids.
Fmoc removal before each coupling step was achieved by
treatment of the resin-bound peptide with a 20% solution of
piperidine in DMF. All coupling reactions and deprotections
were monitored by cleavage of a small aliquot of resin with 2%
TFA in CH2Cl2 and reverse-phase HPLC-MS analysis.
The synthesis began by limited anchoring of Fmoc-Gly-
OH (0.38 equiv) to 2-CTC resin in the presence of DIPEA
(3.3 equiv) 36. Peptide elongation was accomplished with
4.0 equiv of the Fmoc-protected amino acid, furnishing pentapep-
tide 37 with a high coupling yield. By using DIPCDI/HOBt, we
minimized dehydration of the amide group into a nitrile group
(only 7%). An initial plan to proceed to esterification at this point
was revised because of the low yield observed in further external
chain elongation. As a result, we introduced the residue (2S,3S)-
Fmoc-Thr(NHBoc) 10 before formation of the ester bond.
This was achieved by first removing Fmoc-diMeGln, which was
efficiently accomplished without intramolecular lactamization of
the diMeGln into a pyroglutamic derivative, with only one 3 min
deprotection cycle, followed by coupling with 10 in the presence
of HOAt.
At this stage and as mentioned before, ester bond formation
required the use of a protecting group for the α-amino function,
such as the Alloc group, which is orthogonal to the Fmoc, and
the acid-labile protecting groups present in 37. N-Alloc-N-Me-
L-Ala 13 (10 equiv) was quantitatively incorporated in 2 h
when DIPCDI (10 equiv) and DMAP (1.6 equiv) were used, to
obtain hexapeptide 38. Removal of the Alloc group was carried
out with Pd(PPh3)4 in the presence of PhSiH3, and (2R,3R)-
Alloc-β-MeO-D-Tyr(Trt)-OH 8 (5 equiv) was coupled in 3 h
with 5 equiv of DIPCDI and HOAt to afford heptapeptide 39
with neither racemization nor demethoxylation. Introduction
of β-MeO-D-Tyr before growth of the exocyclic arm allowed
minimization of didehydrotyrosine formation (Scheme 2). At
this point, elongation of the exocyclic arm to the last β-hydroxy
acid, instead of removal of the Alloc group of the Tyr derivative,
allowed us to create a hindered environment in which DKP
formation is minimized.25
The exocyclic arm was sequentially constructed by incorpora-
tion of the four hydroxy acids 32−35 with the less reactive
HOBt coupling additive to give nonapeptides 40−43. The Alloc
group was removed from each of them, as described before,
and Fmoc-N-Me-L-Gln 12 (4 equiv) and Fmoc-L-Leu (4 equiv)
were anchored with a mixture of 4 equiv of HOAt and HATU
in the presence of DIPEA (10 equiv) to obtain undecapeptides
44−47, without DKP and dehydration formation.
After removal of the Fmoc group with washes of piperidine−
DMF (1:4), the unstable linear peptides were cleaved from
the resin [TFA−triisopropylsilane (TIS)−CH2Cl2 (1.5:2:96.5),
10 × 1 min] and precipitated over hexane at −78 °C. Macro-
lactamization of the dried crude product was successfully
performed by use of 1-ethyl-3-(3-dimethylaminopropyl)-
carbodiimide (EDC)−HOAt (5:5 equiv) under high dilution
conditions in CH2Cl2 for 5 h at 0 °C to afford the protected
peptides. Subsequent treatment with TFA−TIS−CH2Cl2
a
Conditions: (a) LDA, LiCl, THF, from −78 to 25 °C, 2 h, 74%; (b)
BH3·NH3, LDA, THF, from 0 to 25 °C, 2 h, 95%; (c) PCC, CH2Cl2,
25 °C, 2 h, 100%; (d) (CF3CH2O)2P(O)CH2CO2Et, KHMDS,
18-crown-6, THF, −78 °C, 2 h, 55%; (e) DIBAL, THF, −78 °C, 4 h,
95%; (f) MnO2, CH2Cl2, 25 °C, 2 h, 90%; (g) TiCl4, DIPEA, NMP,
−78 °C, 1 h, 50%; (h) LiOH, H2O2, THF/H2O, 0 °C, 2 h, 99%.
employed a phosphonate with electron-withdrawing groups,
ethyl 2-[bis(2,2,2-trifluoroethoxy)phosphoryl]acetate, together
with strongly dissociating conditions [potassium bis-
(trimethylsilyl)amide (KHMDS) and 18-crown-6 in THF],
yielded exclusively Z-alkenes (S)-21 and (R)-22. Ethyl esters
were subjected to a sequence of reduction−oxidation with
diisobutylaluminum hydride (DIBAL-H) and manganese(IV)
oxide, respectively, to yield aldehydes (S)-23 and (R)-24.
Treatment of the two chiral aldehydes with (R)- and (S)-4-
benzyl-2-oxazolidinone 25 and 26 provided alcohols (S,S,S)-27,
(S,R,S)-28, (R,R,R)-29, and (R,S,R)-30 with 70% enantiomeric
excess (ee). The four Mosher esters were made to assign the
absolute configuration of the alcohols. In our attempts to
improve the enantioselectivity of preparation of the desired acid
in bulk, we used various oxazolidinones. The best performance
was achieved with (R)-3-acetyl-4-isopropyl-5,5-diphenyloxazoli-
din-2-one 31, TiCl4, and (−)-sparteine with 95% ee and 75%
chemical yield. Hydrolysis with LiOH produced the desired four
diastereoisomers of HDMN.
Solid-Phase Synthesis of Stellatolide A. With all the
building blocks in hand, we undertook the challenging synthesis
of 1. Our optimized approach for solid-phase synthesis involved
the following: (i) cyclization via Gly to Leu, which implied
cyclization in solution; (ii) given the lability of β-MeOTyr in
acidic conditions, the starting Gly was introduced on 2-CTC
resin; (iii) Fmoc and Alloc were both used as temporary Nα
protecting groups (the latter, which is removed under milder
conditions than the former, is smaller and shows better
coupling, thus minimizing side reactions); (iv) minimal side-
chain protection was used: only trityl (Trt) for the Tyr
derivative and N-tert-butoxycarbonyl (Boc) for the Thr(NH2)
incorporation,32 whereas D-allo-Thr, L-Thr, (2R,3S)-Fmoc-3-
OH-Asn 11, and the terminated β-hydroxy acid were
introduced without protection of the hydroxyl function; (v)
sequential incorporation of N-MeAla and β-MeO-D-Tyr to
prevent O to N transacylation and pyro-diMeGlu formation;
(vi) elongation of the exocyclic arm before conclusion of the
endocyclic sequence, in order to minimize DKP formation;
(vii) use of HOAt as coupling additive for the most demanding
F
dx.doi.org/10.1021/ja502744a | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
1
(10:2:88) for 45 min led to global deprotection, followed by
reverse-phase HPLC purification of peptides 48−51 by use of a
C18 X-Bridge column and water/MeOH as eluents.
Spectroscopic data of the four diastereoisomers 48−51 were
compared with those of natural 1. Disappointingly, they differed
significantly from the spectrum of the natural compound.
Furthermore, the four synthetic peptides showed a different
HPLC retention time and, surprisingly, no in vitro cytotoxic
activity. We suspected that the originally proposed structure for
1 was incorrect.
Careful examination of H NMR spectrum of the synthetic
products revealed that the (3S,6S) configuration is to be
expected for the acidic residue. All 1H resonances fell within the
corresponding natural product values, and good correlation
was also observed for the respective JH,H coupling constants
(Figure 5). Having concluded that the absolute configuration
of the two chiral centers of the acid moiety was (3S,6S), we
focused our attention on finding the misassignment.
After detailed comparisons of NMR spectra of the four
synthetic peptides with those of the natural product, we
observed that signal patterns were quite different for the external
D-allo-Thr. Furthermore, we reconsidered the Marfey’s analysis
and detected a negligible peak that we considered could
correspond to an L-Thr residue. For some unexplained reason,
L-Thr in the natural product underwent racemization (or even
chemical degradation) under the strong acidic conditions used
for peptide hydrolysis before Marfey’s derivatization. Further-
more, marine cyclodepsipeptides isolated from sponges with a
β-hydroxy-α-amino acid placed at the branching position usually
present a D-allo configuration, (2R,3R)-D-allo-Thr being the most
frequent residue in callipeltins, neamphamides, theopapuamides,
and pipecolidepsin C. In addition, D-allo-β-hydroxyleucine found
in papuamides and mirabamides, and the (2R,3R,4R)-2-amino-3-
hydroxy-4,5-dimethylhexanoic acid (D-allo-AHDMHA) residue
in homophymines and pipecolidepsins A and B also have a
2R,3R absolute configuration. These two observations strongly
suggested that 1 has the same 2R,3R configuration at the
branching position and thus has an L configuration at the second
Thr residue (Thr2).
Figure 5. Partial 1H NMR spectra of stellatolide A (1) and
diastereoisomers 48−51.
a
Scheme 4. Solid-Phase Synthesis of Stellatolide A
a
Conditions: (a) piperidine−DMF (1:4); (b) Fmoc-D-Ser(Me)-OH, HOBt, DIPCDI, DMF; (c) Fmoc-D-allo-Thr-OH, HOBt, DIPCDI, DMF;
(d) (2S,3S,4R)-Fmoc-diMeGln-OH 9, HOBt, DIPCDI, DMF; (e) (2S,3S)-Fmoc-Thr(NHBoc)-OH 10, HOBt, DIPEA, DMF; (f) N-Alloc-N-Me-L-
Ala-OH 13, DIPCDI, DMAP, CH2Cl2; (g) Pd(PPh3)4, PhSiH3, CH2Cl2; (h) (2R,3R)-Alloc-β-MeO-Tyr(Trt)-OH 8, DIPCDI, HOAt, DMF; (i)
Fmoc-D-Allo-Thr-OH or Fmoc-L-Thr-OH, HOBt, DIPCDI, DMF; (j) (2R,3S)-Fmoc-3-OH-Asn-OH 11, HOBt, DIPCDI, DMF; (k) (Z)-3-hydroxy-
6,8-dimethylnon-4-enoic acid 32−35, HOBt, DIPCDI, DMF; (l) Fmoc-N-Me-L-Gln-OH 12, HOAt, HATU, DIPEA, DMF; (m) Fmoc-L-Leu,
HOAt, HATU, DIPEA, DMF; (n) (i) TFA−TIS−CH2Cl2 (1.5:2:96.5), 10 × 1 min, (ii) −78 °C, hexane; (o) HOAt, EDC, CH2Cl2, DMF, 0 °C, 5 h;
(p) TFA−TIS−CH2Cl2 (10:2:88), 25 °C, 45 min.
G
dx.doi.org/10.1021/ja502744a | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Therefore, we performed the synthesis, coupling 39 with the
L-Thr residue instead of D-allo-Thr, and following the sequence
depicted in Scheme 4, coupling with (3S,6S,Z)-3-hydroxy-6,8-
dimethyl-4-nonenoic acid 32. The corresponding synthetic
compound and the natural 1 revealed a perfect overlap of the
NMR spectra, identical analytical HPLC behavior, and the same
in vitro activity (Table 2). This finding highlights the relevance
role of the absolute configuration of some chiral centers of the
exocyclic arm. In this regard, for instance, activity is lost when
L-Thr is replaced by D-allo-Thr, a finding that is valuable for future
research purposes. In addition, the development of an efficient
synthesis of stellatolide A on solid support has circumvented the
issue of limited availability and allowed us to initiate advanced in
vivo evaluation of this compound.
ASSOCIATED CONTENT
* Supporting Information
Table 2. Cytotoxic Activity Data of Stellatolides A−G
■
S
GI50, μM
Additional text, structures, and three tables with experimental
procedures and spectral and other characterization data. This
material is available free of charge via the Internet at http://
pubs.acs.org.
stellatolide
lung, A549
colon, HT-29
breast, MDA-MB-231
A (natural)
0.08
0.08
0.64
4.73
0.21
0.90
1.23
>6.41
0.43
0.38
0.21
0.26
A (synthetic)
B
C
D
E
1.15
0.70
>6.75
0.71
>6.75
0.31
AUTHOR INFORMATION
■
Corresponding Authors
judit.tulla@irbbarcelona.org
ccuevas@pharmamar.com
2.69
1.60
F
>6.48
>6.41
2.14
G
>6.41
Present Addresses
▲(F.R.) Fundacion
́
MEDINA, 18016 Granada, Spain.
of the chirality of the Thr amino acid in the exocyclic arm for
the biological activities exerted by members of the stellatolide
family.
▽(I.M.) Clarke, Modet & Company, 28001 Madrid, Spain.
Notes
The authors declare the following competing financial
interest(s): M.J.M., R.R.-A., V.M., C.M., I.D., R.F., A.M.F.,
S.M., and C.C. are employees and shareholders of PharmaMar.
I.M. and F.R. were employees of PharmaMar.
On the basis of these results, we firmly established that the
β-hydroxy acid has an (S,S) configuration and unequivocally
confirmed the presence of the L-Thr residue located in the
exocyclic arm, which was clearly difficult to determine by
conventional methods. The presence of this residue was then
validated by total synthesis.
Biological Evaluation. A bioassay was conducted in
parallel during fractionation of the sponge extracts: cell-killing
capacity was evaluated against a panel of three human tumor
cell lines, including lung (A-549), colon (HT-29), and breast
(MDA-MB-231). In vitro antiproliferative activity was assessed
by the colorimetric SRB (sulforhodamine B) method, per-
formed as described previously.33 The GI50 (micromolar), a con-
centration that causes 50% growth inhibition, values obtained for
stellatolides in this assay are shown in Table 2. It is interesting to
note that stellatolides F and G containing didehydroaminobu-
tyric acid (Dhb) lose much of their potency and the acyclic
stellatolide C is also less potent. The other stellatolides showed
significant in vitro antiproliferative activity with values of
0.08−2.7 μM.
ACKNOWLEDGMENTS
■
́
We thank Dr. L. F. Garcıa-Fernan
́
dez for cytotoxicity assays,
Dr. J. L. Carballo for taxonomic identification of the sponge,
́
and Dr. S. Gonzalez for recording NMR spectra. This study was
partially funded by the CICYT (CTQ2012-30930), the
Generalitat de Catalunya (2009SGR1024), and the Institute
for Research in Biomedicine (IRB Barcelona). R.R.-A. thanks
the MEC and ESF for a Torres-Quevedo grant.
REFERENCES
■
(1) (a) Pelay-Gimeno, M.; Tulla-Puche, J.; Albericio, F. Mar. Drugs
2013, 11, 1693−1717. (b) Stolze, S. C.; Kaiser, M. Molecules 2013, 18,
1337−1367. (c) Blunt, J. W.; Copp, B. R.; Keyzers, R. A.; Munro, M.
H. G.; Prinsep, M. R. Nat. Prod. Rep. 2013, 30, 237−323 and
references cited therein.
(2) Zampella, A.; D’Auria, M. V.; Gomez Paloma, L.; Casapullo, A.;
Minale, L.; Debitus, C.; Henin, Y. J. Am. Chem. Soc. 1996, 118, 6202−
6209.
(3) D’Auria, M. V.; Zampella, A.; Paloma, L. G.; Minale, L.; Debitus,
C.; Roussakis, C.; Le Bert, V. Tetrahedron 1996, 52, 9589−9596.
(4) Zampella, A.; Randazzo, A.; Borbone, N.; Luciani, S.; Trevisi, L.;
Debitus, C.; D’Auria, M. V. Tetrahedron Lett. 2002, 43, 6163−6166.
(5) Sepe, V.; D’Orsi, R.; Borbone, N.; D’Auria, M. V.; Bifulco, G.;
Monti, M. C.; Catania, A.; Zampella, A. Tetrahedron 2006, 62, 833−
840.
(6) D’Auria, M. V.; Sepe, V.; D’Orsi, R.; Bellotta, F.; Debitus, C.;
Zampella, A. Tetrahedron 2007, 63, 131−140.
(7) Ford, P. W.; Gustafson, K. R.; Mckee, T. C.; Shigematsu, N.;
Maurizi, L. K.; Pannell, L. K.; Williams, D. E.; Silva, E. D. D.; Lassota,
P.; Allen, T. M.; Soest, R. V.; Andersen, R. J.; Boyd, M. R. J. Am. Chem.
Soc. 1999, 121, 5899−5909.
CONCLUSION
■
In conclusion, here we isolated seven novel cytotoxic
depsipeptides, stellatolides A−G, which contain various non-
natural amino acids from extracts of the sponge Ecionemia
acervus. Their structures were established by spectroscopic
methods and corroborated by the first total synthesis of
stellatolide A. This synthetic process allowed us establish that
for these branched peptides, in addition to the choice of the
protecting groups, resins, and coupling reagents, the order of
sequence completion of the two arms is crucial for minimizing
side reactions. Furthermore, the total synthesis of stellatolide A
allowed us to unambiguously establish its absolute configuration,
which was initially erroneously misassigned by advanced
Marfey’s analysis for one of the Thr residues. The strategy
presented allows the synthesis of analogues of stellatolides for
drug discovery purposes, thus permitting the establishment of
the first structure−activity relationships of this promising
antitumor family of compounds. Our findings reveal the key
(8) Prasad, P.; Aalbersberg, W.; Feussner, K.-D.; van Wagoner, R. M.
Tetrahedron 2011, 67, 8529−8531.
(9) Oku, N.; Gustafson, K. R.; Cartner, L. K.; Wilson, J. A.;
Shigematsu, N.; Hess, S.; Pannell, L. K.; Boyd, M. R.; McMahon, J. B.
J. Nat. Prod. 2004, 67, 1407−1411.
H
dx.doi.org/10.1021/ja502744a | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
(10) Yamano, Y.; Arai, M.; Kobayashi, M. Bioorg. Med. Chem. Lett.
2012, 22, 4877−4881.
out the reduction in solution after the cyclization step. Several
conditions were tested, such as Staudinger reaction, SnCl2 treatment,
hydrogenation with Pd, and Lindlar catalysts and Zn, but all of them
failed since the quality of the crude product decreased. Finally, the
Thr(NH2) residue was found to be the best choice to facilitate the
synthesis of this complex natural depsipeptide.
(33) Skehan, P.; Storeng, R.; Scudiero, D.; Monks, A.; McMahon, J.;
Vistica, D.; Warren, J. T.; Bokesch, H.; Kenney, S.; Boyd, M. R. J. Natl.
Cancer Inst. 1990, 82, 1107−1112.
(11) Ratnayake, A. S.; Bugni, T. S.; Feng, X.; Harper, M. K.; Skalicky,
J. J.; Mohammed, K. A.; Andjelic, C. D.; Barrows, L. R.; Ireland, C. M.
J. Nat. Prod. 2006, 69, 1582−1586.
(12) Plaza, A.; Bifulco, G.; Keffer, J. L.; Lloyd, J. R.; Baker, H. L.;
Bewley, C. A. J. Org. Chem. 2009, 74, 504−512.
(13) Plaza, A.; Gustchina, E.; Baker, H. L.; Kelly, M.; Bewley, C. A. J.
Nat. Prod. 2007, 70, 1753−1760.
(14) Lu, Z.; Wagoner, R. M. V.; Harper, M. K.; Baker, H. L.; Hooper,
J. N. A.; Bewley, C. A.; Ireland, C. M. J. Nat. Prod. 2011, 74, 185−193.
(15) Zampella, A.; Sepe, V.; Luciano, P.; Bellotta, F.; Monti, M. C.;
́ ̂
D’Auria, M. V.; Jepsen, T.; Petek, S.; Adeline, M.-T.; Laprevote, O.;
Aubertin, A.-M.; Debitus, C.; Poupat, C.; Ahond, A. J. Org. Chem.
2008, 73, 5319−5327.
(16) Zampella, A.; Sepe, V.; Bellotta, F.; Luciano, P.; D’Auria, M. V.;
Cresteil, T.; Debitus, C.; Petek, S.; Poupat, C.; Ahond, A. Org. Biomol.
Chem. 2009, 7, 4037−4044.
́
(17) (a) Coello, L.; Fernandez, R.; Reyes, J. F.; Francesch, A.;
Cuevas, C. Int. Pat. Appl. WO 2010/070078 A1, 2010. (b) Pelay-
Gimeno, M.; García-Martín, Y.; Martin, M. J.; Spengler, J.; Molina-
Guijarro, J. M.; Munt, S.; Francesch, A. M.; Cuevas, C.; Tulla-Puche,
J.; Albericio, F. Nat. Commun. 2013, 4, No. 2352, DOI: 10.1038/
ncomms3352.
(18) Marfey, P. Carlsberg Res. Commun. 1984, 49, 591−596.
(19) (a) Liang, B.; Carroll, P. J.; Joullie, M. M. Org. Lett. 2000, 2,
4157−4160. (b) Schmidt, U.; Mundinger, K.; Riedl, B.; Haas, G.; Lau,
R. Synthesis 1992, 1201−1202. (c) Wagner, R.; Tilley, J. W.; Lovey, K.
Synthesis 1990, 785−786. (d) Mori, K.; Iwasawa, H. Tetrahedron 1980,
36, 87−90.
(20) Zampella, A.; D’Orsi, R.; Sepe, V.; Casapullo, A.; Monti, M. C.;
D’Auria, M. V. Org. Lett. 2005, 7, 3585−3588.
(21) (a) Lipton and co-workers have published the synthesis of
callipeltin B, a natural product with a similar cycle but terminated in
pyro-diMeGln, which faciliates the synthetic process:
Krishnamoorthy, R.; Vazquez-Serrano, L. D.; Turk, J. A.; Kowalski,
J. A.; Benson, A. G.; Breaux, N. T.; Lipton, M. A. J. Am. Chem. Soc.
2006, 128, 15392−15393. (b) Ma and co-workers have accomplished
an elegant synthesis of papuamide B carried out in solution that
comprises 27 linear steps and an overall yield of 7%: Xie, W.; Ding, D.;
Zi, W.; Li, G.; Ma, D. Angew. Chem. 2008, 120, 2886−2890.
(22) Sequences containing an N-alkylamino acid and a D-amino acid
are very prone for DKP formation.
(23) In order to establish the synthetic strategy, the simplified Tyr
derivative was used.
(24) Esterification with Fmoc-N-MeAla-OH was more difficult due
to steric hindrance of the Fmoc group.
(25) This could constitute another example of “conformationally
restricted mobility protection”: Tulla-Puche, J.; Marcucci, E.; Fermin,
M.; Bayo-Puxan, N.; Albericio, F. Chem.Eur. J. 2008, 14, 4475−
́
4478.
(26) Rama Rao, A. V.; Chakraborty, T. K.; Laxma Reddy, K.;
Srinivasa Rao, A. Tetrahedron Lett. 1994, 35, 5043−5046.
(27) (a) Acevedo, C. M.; Kogut, E. F.; Lipton, M. A. Tetrahedron
2001, 57, 6353−6359. (b) C
2005, 70, 6218−6221.
(28) (a) Mattingly, P. G.; Miller, M. J.; Cooper, R. D. G.; Daughtery,
B. W. J. Org. Chem. 1983, 48, 3556−3559. (b) Guzman-Martínez, A.;
VanNieuwenhze, M. S. Synlett 2007, 1513−1516.
(29) Calimsiz, S.; Morales Ramos, A. I.; Lipton, M. A. J. Org. Chem.
̧alimsiz, S.; Lipton, M. A. J. Org. Chem.
́
̧
2006, 71, 6351−6356.
(30) Steven, C. M.; Watanabe, R. J. Am. Chem. Soc. 1950, 72, 725−
727.
(31) Myers, A. G.; Yang, B. H.; Chen, H.; McKinstry, L.; Kopecky, D.
J.; Gleason, J. L. J. Am. Chem. Soc. 1997, 119, 6496−6511.
(32) First, Fmoc-Thr(N3)-OH was used as building block 10 because
of its easier elaboration. The reduction on solid phase was successful
with SnCl2, thiophenol, and DIPEA, but the protection of the new
amine was not achieved in high yields. So the alternative was to carry
I
dx.doi.org/10.1021/ja502744a | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX