Efficient Synthesis and Stereochemical Revision of Coibamide A

Article abstract of DOI:10.1021/jacs.5b09286

Coibamide A is a highly potent antiproliferative cyclodepsipeptide originally isolated from a Panamanian marine cyanobacterium. Herein we report an efficient solid-phase strategy for assembly of highly N-methylated cyclodepsipeptides, which is invaluable in generating coibamide A derivatives for structure-activity relationship studies. As a consequence of our synthetic studies, two stereochemical assignments of coibamide A were revised and the total synthesis of this natural compound was achieved for the first time.

Full text of DOI:10.1021/jacs.5b09286

Communication
pubs.acs.org/JACS
Efficient Synthesis and Stereochemical Revision of Coibamide A
Guiyang Yao,†,‡,§ Zhengyin Pan, Chunlei Wu, Wei Wang, Lijing Fang,* and Wu Su*
†,§
†
†
,†
,†
†Guangdong Key Laboratory of Nanomedicine, Institute of Biomedicine and Biotechnology, Shenzhen Institutes of Advanced
Technology, Chinese Academy of Sciences, Shenzhen, Guangdong, 518055, P. R. China
‡
Departent of Pharmaceutical Engineering, School of Chemistry and Chemical Engineering, Southeast University, Nanjing, 211189, P.
R. China
*
S Supporting Information
Coibamide A has been reported to be a side-to-tail cyclized
depsipeptide possessing a 22-membered macrocycle with one
hydroxy acid unit and ten L-amino acids where eight of the
residues are N-methylated. The sequence and stereochemistry of
the amino acids in this molecule were established by a
combination of spectroscopic analysis, chemical degradation,
and derivatization studies. The absolute configuration of
threonine was determined using computational models of the
two possible isomeric structures, constrained by ROESY
correlations. The synthesis of a macrocyclic peptide containing
multiple densely N-methylated amides are challenging as a
consequence of the low inherent reactivity of N-methylated
ABSTRACT: Coibamide A is a highly potent antiprolifer-
ative cyclodepsipeptide originally isolated from a Pan-
amanian marine cyanobacterium. Herein we report an
efficient solid-phase strategy for assembly of highly N-
methylated cyclodepsipeptides, which is invaluable in
generating coibamide A derivatives for structure−activity
relationship studies. As a consequence of our synthetic
studies, two stereochemical assignments of coibamide A
were revised and the total synthesis of this natural
compound was achieved for the first time.
7
amino acids. This problem is exacerbated by side reactions such
as 2,5-diketopiperizine (DKP) formation, epimerization during
the coupling steps, polymerization in the cyclization step, etc. In
yclodepsipeptides are cyclic heterodetic peptides consist-
C
ing of hydroxy and amino acids linked by ester and amide
1
bonds. Some of the cyclodepsipeptides isolated from both
2
014, He et al. reported the first total synthesis of the proposed
marine and terrestrial sources contain N-methylated amino acids.
structure of coibamide A using a [(4 + 1) + 3 + 3]-peptide
Such structural features may help to improve the receptor
8
fragment-coupling strategy in solution phase. But the analytical
2
selectivity, metabolic stability, and hydrophobicity. A range of
and biological activity data for the synthetic sample were
inconsistent with those reported for natural coibamide A. Based
on their results, further determination and reassignment of the
full structure of coibamide A would be needed. Compared with
the multistep and time-consuming synthesis in solution phase, a
straightforward solid-phase peptide synthesis (SPPS) route
would facilitate and speed up the synthesis of coibamide A.
Recently, Nabika et al. reported their synthetic efforts toward
coibamide A by using Fmoc-based SPPS followed by the cleavage
of the resulting linear peptide from the resin and its subsequent
biological activities, including anticancer, antiviral, antifungal,
and immunosuppressive properties, have been observed for N-
3
methylated cyclodepsipeptides. The remarkable biological
activities along with their complex molecular scaffolds make
4
them interesting targets for chemical synthesis. Coibamide A
(
Figure 1) is a highly N-methylated cyclodepsipeptides, which
9
macrolactonization. However, although numerous coupling
agents were evaluated, the macrolactonization between the
5
11
hydroxyl group of MeThr and C-terminal MeAla failed to
1
1
provide any of the desired product, and only the [D-MeAla ]-
epimer of 1a was formed in a low yield of 3.8%. In order to
validate the configuration of this natural cyclodepsipeptide and
to supply samples for further pharmaceutical development, we
set out to establish a flexible and efficient solid-phase method for
the total synthesis of coibamide A and its analogues.
Figure 1. Originally proposed structure of coibamide A (1a).
First our efforts were focused on determining the cyclization
point and optimizing the coupling methodology (Scheme 1).
Besides the unfavorable late-stage macrolactonization, a proper
cyclization position should also avoid those sterically encum-
bered by the N-methylated amide bonds. Considering the high
risk of competing DKP (MeAla-Tyr(Me)) formation during the
was isolated by McPhail and co-workers from a Panamanian
marine cyanobacterium in 2007. Coibamide A exhibited low
5
nanomolar inhibitory activities against the proliferation of a
number of cancer cell lines with a potentially new mechanism of
6
action. As a promising lead agent, further exploration of this
molecule in cancer drug discovery was however limited by the
lack of availability from natural sources.
Received: September 2, 2015
©
XXXX American Chemical Society
A
DOI: 10.1021/jacs.5b09286
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
Scheme 1. Retrosynthetic Analysis of 1a
reagent set of BTC/collidine was expected to effect the coupling
of both N-methylated amino acids and natural amino acids of
coibamide A.
Among the nine building blocks, Fmoc-N-Me-Ser(Me)-OH
and Me Val-HIV-OH are not commercially available and thus
2
were stereoselectively prepared in solution phase. Fmoc-N-Me-
Ser(Me)-OH was synthesized according to a previously reported
16
procedure. The synthesis of the Me Val-HIV-OH fragment is
2
shown in Scheme 2. Reductive methylation of L-valine [aqueous
Scheme 2. Synthesis of Me Val-HIV-OH (9a) and Me Val-D-
2
2
a
HIV-OH (9b)
a
Reagents and conditions: (a) H , Pd/C, aq. HCHO; (b) allyl
2
bromide, Cs CO , DMF; (c) EDCI, DMAP, CH Cl ; (d) PPh ,
2
3
2
2
3
DIAD, THF; (e) Pd(PPh ) , morpholine, THF.
3
4
17
formaldehyde, H , Pd(C)] produced Me Val-OH (6). Allyl
2
2
1
0
cyclization between N-terminal Tyr(Me) and C-terminal
ester 8 was prepared by selective O-allylation of (S)-(+)-2-
hydroxy-3-methylbutyric acid (7) with allyl bromide and
Cs CO . Me Val-HIV-OH (9a) was afforded by formation of
9
9
MeLeu , the macrocyclization site was chosen at the MeIIe -
Ala8 junction, affording precursor 2. Using orthogonally
protected MeThr would allow successive solid-phase assembly
of the main peptidyl chain of precursor 4 followed by
introduction of MeAla via the ester bond formation and
elongation of the side chain. Conveniently, Fmoc-N-methyl-O-
tert-butyl-L-threonine is commercially available. However, the
use of this tert-butyl (tBu) ether derivative would require
establishing an Fmoc-based solid-phase method that also has to
be compatible with trifluoroacetic acid (TFA) deprotection.
Thus, aryl hydrazide resin was selected because the aryl hydrazide
linker is stable in both strongly acidic and basic conditions during
Boc- and Fmoc-based solid-phase synthesis, and yet it can be
2
3
2
the sterically hindered ester bond between 6 and 8 using EDCI/
DMAP and subsequent removal of the allyl group with
[Pd(PPh ) ] and morpholine.
3
4
With all of the building blocks in hand, the assembly of the
precursor 13 on Fmoc-hydrazinobenzoyl AM resin could be
initiated (Scheme 3). After removal of the Fmoc protecting
group of the resin with 20% piperidine in DMF, the first amino
acid, MeIle, was activated with BTC and attached onto the resin.
A substitution level of 0.30 mmol/g was preferred to facilitate the
tertiary amide bond formation. To avoid truncated products,
acetic anhydride was then added to cap any of the remaining
amino groups on the resin. As expected, the Fmoc-based solid-
phase synthesis of the heptapeptide 11 proceeded smoothly
employing BTC/collidine as the coupling reagents for N-
methylated amino acids, MeSer(Me), MeThr(tBu), MeLeu, a
1
0
cleaved under very mild oxidative conditions. The reactivity of
the transient diazene species toward nucleophiles also allows
cyclative cleavage, which releases the cyclodepsipeptide directly
into solution without the need for an additional cyclization
1
1
reaction.
second MeSer(Me), and Me Val-HIV. All couplings were carried
2
To overcome the difficulties associated with the synthesis of
N-methylated amide bonds, the use of highly electrophilic
species such as acid chlorides and acid anhydrides has attracted
great attention. Triphosgene [bis(trichloromethyl) carbonate,
BTC] converts carboxylic acids into acid chlorides, which readily
acylate even sterically hindered and electronically deactivated
out at room temperature with a single-coupling cycle for 1 h or
less until a negative chloranil test showed the absence of free
secondary amine on the resin.
To commence the side chain synthesis, the tBu protecting
group on resin-bound 11 was successfully removed by treating
the resin twice with TFA/TIS (triisopropylsilane)/water cocktail
(95/2.5/2.5) for 2 and 5 min, respectively. HPLC analysis of a
small sample cleaved from the resin revealed that TFA-induced
1
2
amines. It has shown superiority over benzotriazole-based
reagents such as HATU or HOAt/DIC in the couplings of N-
methylated or aromatic amino acids in both Fmoc-based and
18
amidolysis of the N-methylated peptide could be suppressed
using a minimized reaction time during deprotection of the tert-
butyl ether. Esterification of the free secondary hydroxyl group of
4 with MeAla mediated by BTC/collidine was not efficient
compared to amidation. An acceptable yield was achieved when
symmetric anhydride formed by the reaction of Fmoc-N-Me-Ala-
OH and DCC in a ratio of 2:1 was used, allowing the coupling
reaction to be finished in 2.5 h with a single-coupling cycle.
Attachment of the second residue of the side chain, Fmoc-
Tyr(Me)-OH, onto the peptidyl resin was carried out using the
optimized protocol for BTC-mediated coupling of unmethylated
1
3
Boc-based SPPS. However, it is commonly regarded that BTC
does not work properly in the coupling of urethane-protected
unmethylated amino acids, because the NH proton of the
urethane function is susceptible to side reactions during strong
activation with the BTC reagent. In the previous study, we
demonstrated that increasing the molar ratio of BTC to the
amino acids significantly improved the activation efficiency of
14
unmethylated amino acids. Further research revealed that the
addition of 1 equiv of HOAt to the BTC activated intermediate
15
could dramatically accelerate the coupling rate. Therefore, the
B
DOI: 10.1021/jacs.5b09286
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
a
Scheme 3. Solid-phase Synthesis of Precursor 13
a
Reagents and conditions: (a) 20% piperdine/DMF; (b) Fmoc-N-Me-Ile-OH; BTC, collidine, DIEA, THF; then acetic anhydride, DIEA, DMF; (c)
Fmoc-N-Me-Ser(Me)-OH; BTC, collidine, DIEA, THF; (d) Fmoc-N-Me-Thr(tBu)-OH; BTC, collidine, DIEA, THF; (e) Fmoc-N-Me-Leu-OH;
BTC, collidine, DIEA, THF; (f) Me Val-HIV−OH (9a), BTC, collidine, DIEA, THF; (g) TFA/TIS/H O (95/2.5/2.5); (h) Fmoc-N-Me-Ala-OH;
2
2
DCC, DMAP, CH Cl , DMF; (i) Fmoc-Tyr(Me)-OH or Boc-Tyr(Me)-OH; BTC, collidine, THF, HOAt, DIEA, DMF; (j) Fmoc-N-Me-Leu-OH;
2
2
BTC, collidine, DIEA, THF; (k) Fmoc-Ala-OH, BTC, collidine, THF, HOAt, DIEA, DMF.
amino acids with HOAt as an additive. However, removal of the
Fmoc group on 12a with 20% piperidine in DMF resulted in
partial degradation of the linear peptide via DKP formation even
when exposure to piperidine was minimized (3 min). It was
envisioned that DKP formation could be avoided under acidic
deprotection conditions. Therefore, Boc-Tyr(Me)-OH was
attached to the resin as an alternative of Fmoc-Tyr(Me)-OH
to form 12b. Applying the aforementioned TFA-deprotection
procedure used for the tert-butyl ether, Boc group was removed
with no degraded linear peptide detected. The following two
residues, MeLeu and Ala, were incorporated into the side
peptidyl chain to give 13 by the coupling reactions mediated by
BTC/collidine and BTC/collidine/HOAt, respectively.
Cu(OAc) /Py/H O in CH Cl and semipreparative HPLC
2 2 2 2
purification, precursor 14 was obtained in 15% yield. The Fmoc
group of 14 was removed, and the cyclization was carried out in
the presence of EDCI/HOAt/DIEA in DCM. After purification,
the proposed structure of coibamide A (1a) was obtained in 56%
yield with spectroscopic data identical to those reported by He et
8
1
13
al. We also noted that H and C NMR spectral data of 1a did
5
not agree with those of natural coibamide A. The significant
difference in NMR spectral data once again brings into question
the original structural assignment of the natural product.
The synthetic strategy was further explored in the synthesis of
11
the [D-MeAla ]-epimer of 1a (1b), which was found to be
slightly less potent than the natural one, but still exhibited
nanomolar cytotoxicity toward a number of different cancer cell
lines. Starting from resin-bound alcohol 4, D-MeAla residue was
incorporated to the resin by esterification. Following the same
procedure for the side chain elongation, resin cleavage, and
cyclization, 1b was synthesized in an overall yield of 12%. Its
spectroscopic data were consistent with those reported by
After removal of the Fmoc group from 13, cyclative cleavage of
the cyclodepsipeptide from resin 2 was investigated under a
variety of reaction conditions. The transient diazene species was
formed by oxidation of the hydrazide with N-bromosuccinimide
10
(
NBS) or atmospheric O in the presence of Cu(II) salts.
2
However, the following intramolecular nucleophilic attack by
alanine amine did not occur possibly due to the steric hindrance
of the C-terminal MeIle residue. Direct aminolysis of resin-
bound peptide 2 under air at elevated temperatures (55 and 90
9
Nabika et al. Although X-ray structures of the synthetic samples
were not available, possible epimerization during the coupling
reactions and macrocyclization was excluded by synthesis of both
11
the proposed structure 1a and [D-MeAla ]-epimer 1b, which
1
9
°
C) did not provide any of the cyclized product, either.
were reported by two individual groups previously.
Consequently, the solid-phase macrocyclization approach had to
be adjusted and a solution-phase cyclization was conducted
Our next effort was focused on elucidation of the stereo-
chemical structure of natural coibamide A. A detailed comparison
1
(Scheme 4). After cleavage from resin 13 using a mixture of
of the H NMR spectrum of the natural product with that of
synthetic samples disclosed that 1b was more similar to natural
coibamide A than 1a. The H signals attributed to the seven
1
Scheme 4. Synthesis of the Proposed Structure of Coibamide
A (1a)
a
amino acid residues belonging to the macrocycle of 1b were
almost identical to those of natural coibamide A, while significant
differences were observed for the other four residues outside the
1
1
macrocycle. This suggested that the figuration of MeAla residue
in natural coibamide A should be revised from L to D, and at least
one exocyclic residue also needs to be revised. Among the four
4
3
exocyclic residues, misassignment of MeLeu and MeSer(Me)
was less possible because their counterparts also exist in the
macrocycle, the stereochemistry of which were believed to be
2
correct. The figuration of HIV was more suspicious in view of
1
1
the limited influence on H signals of the terminal Me Val .
2
2
11
Therefore, [D-HIV ]-[D-MeAla ]-diastereomer 1c was antici-
pated to be the real structure of natural coibamide A, thus
becoming our next target (Figure 2).
a
Reagents and conditions: (a) Cu(OAc) , Py, CH Cl /H O; (b)
2
2
2
2
Et NH, CH CN; (c) EDCI, HOAt, DIEA, CH Cl .
2
3
2
2
C
DOI: 10.1021/jacs.5b09286
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
Figure 2. Chemical structures of 1b and 1c.
To convert the stereochemistry of HIV, a Mitsunobu reaction
REFERENCES
■
of Me Val-OH (6) and alcohol 8 was carried out and Me Val-D-
2
2
(1) Davies, J. S. J. Pept. Sci. 2003, 9, 471−501.
2) (a) Chatterjee, J.; Gilon, C.; Hoffman, A.; Kessler, H. Acc. Chem.
HIV-OH (9b) was obtained after removal of the allyl group with
Pd(PPh ) ] and morpholine. Following the sequence depicted
(
[
Res. 2008, 41, 1331−42. (b) Chatterjee, J.; Laufer, B.; Kessler, H. Nat.
3
4
in Schemes 3 and 4, we then performed the synthesis of 1c, using
b and Fmoc-D-N-Me-Ala-OH instead of 9a and Fmoc-N-Me-
Ala-OH, respectively. The resulted synthetic compound and the
natural product revealed a perfect overlap of the NMR spectra
Protoc. 2012, 7, 432−44.
9
(3) Sivanathan, S.; Scherkenbeck, J. Molecules 2014, 19, 12368−12420.
(
4) For recent examples, see: (a) Pelay-Gimeno, M.; García-Ramos, 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,
352. (b) Murai, M.; Kaji, T.; Kuranaga, T.; Hamamoto, H.; Sekimizu,
K.; Inoue, M. Angew. Chem., Int. Ed. 2015, 54, 1556−60. (c) Martín, M.
J.; Rodríguez-Acebes, R.; García-Ramos, Y.; Martínez, V.; Murcia, C.;
́ ́
Digon, I.; Marco, I.; Pelay-Gimeno, M.; Fernandez, R.; Reyes, F.;
Francesch, A. M.; Munt, S.; Tulla-Puche, J.; Albericio, F.; Cuevas, C. J.
Am. Chem. Soc. 2014, 136, 6754−62.
(
see Supporting Information). Therefore, the absolute structure
2
of coibamide A should be revised from 1a to 1c. Furthermore, a
bioactivity assay was performed using the MTT method. As
expected, 1c displayed similar in vitro activity against MDA-MB-
31 cells (IC50 3.9 nM), corroborating the authenticity of our
structural revisions.
2
In summary, the total synthesis and stereochemical revision of
coibamide A were achieved for the first time by using an efficient
and robust solid-phase method we have developed. The
combined application of both the Boc- and Fmoc-based strategy
on aryl hydrazide resin allows for successive assembly of the main
and side peptidyl chains by solid-phase peptide synthesis. The
use of BTC to generate all of the amide bonds on solid support is
also featured. Establishment of the revised structure of coibamide
A paves the way for structure−activity relationship studies and
elucidation of the unknown mechanism of action, which are
currently underway in our laboratory.
(5) Medina, R. A.; Goeger, D. E.; Hills, P.; Mooberry, S. L.; Huang, N.;
Romero, L. I.; Ortega-Barría, E.; Gerwick, W. H.; McPhail, K. L. J. Am.
Chem. Soc. 2008, 130, 6324−5.
(
̈
6) Hau, A. M.; Greenwood, J. A.; Lohr, C. V.; Serrill, J. D.; Proteau, P.
J.; Ganley, I. G.; McPhail, K. L.; Ishmael, J. E. PLoS One 2013, 8, e65250.
(
(
7) Stolze, S. C.; Kaiser, M. Synthesis 2012, 44, 1755−77.
8) He, W.; Qiu, H. B.; Chen, Y. J.; Xi, J.; Yao, Z. J. Tetrahedron Lett.
2
(
014, 55, 6109−12.
9) Nabika, R.; Suyama, T. L.; Hau, A. M.; Misu, R.; Ohno, H.; Ishmael,
J. E.; McPhail, K. L.; Oishi, S.; Fujii, N. Bioorg. Med. Chem. Lett. 2015, 25,
302−6.
(10) Woo, Y.; Mitchell, A. R.; Camarero, J. A. Int. J. Pept. Res. Ther.
2
(
007, 13, 181−90.
11) (a) Rosenbaum, C.; Waldmann, H. Tetrahedron Lett. 2001, 42,
ASSOCIATED CONTENT
■
5
677−80. (b) Shigenaga, A.; Moss, J. A.; Ashley, F. T.; Kaufmann, G. F.;
*
S
Supporting Information
Janda, K. D. Synlett 2006, 4, 551−4.
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/jacs.5b09286.
(12) (a) Falb, E.; Yechezkel, T.; Salitra, Y.; Gilon, C. J. Pept. Res. 1999,
53, 507−17. (b) Eckert, H.; Forster, B. Angew. Chem., Int. Ed. Engl. 1987,
2
2
(
6, 894−5. (c) Fuse, S.; Mifune, Y.; Takahashi, T. Angew. Chem., Int. Ed.
014, 53, 851−5.
13) (a) Thern, B.; Rudolph, J.; Jung, G. Angew. Chem., Int. Ed. 2002,
Experimental procedures and analysis data (PDF)
4
4
(
1, 2307−9. (b) Thern, B.; Rudolph, J.; Jung, G. Tetrahedron Lett. 2002,
3, 5013−6.
14) Fang, L.; Wu, C.; Yu, Z.; Shang, P.; Cheng, Y.; Peng, Y.; Su, W.
AUTHOR INFORMATION
■
Corresponding Authors
lj.fang@siat.ac.cn
wu.su@siat.ac.cn
*
Eur. J. Org. Chem. 2014, 2014, 7572−6.
*
(15) Fang, L.; Yao, G.; Pan, Z.; Wu, C.; Wang, H. S.; Burley, G. A.; Su,
W. Org. Lett. 2015, 17, 158−61.
Author Contributions
G.Y. and Z.P. contributed equally.
§
(16) Freidinger, R. M.; Hinkle, J. S.; Perlow, D. S.; Arison, B. H. J. Org.
Chem. 1983, 48, 77−81.
Notes
(17) King, A. M.; De Ryck, M.; Kaminski, R.; Valade, A.; Stables, J. P.;
Kohn, H. J. Med. Chem. 2011, 54, 6432−42.
The authors declare no competing financial interest.
(
18) (a) Urban, J.; Vaisar, T.; Shen, R.; Lee, M. S. Int. J. Pept. Protein Res.
1
996, 47, 182−189. (b) Vaisar, T.; Urban, J. J. Mass Spectrom. 1998, 33,
ACKNOWLEDGMENTS
■
5
05−24.
This work was supported by the “Hundred Talents Program” of
Chinese Academy of Sciences, Shenzhen Sciences & Technology
Innovation Council (KQCX20130628112914285,
KQCX2015033117354154, and JCYJ20150316143416083)
and the National Natural Science Foundation of China (Grant
Nos. 21402232 and 21432003). The authors are grateful for the
assistance of Mass facility from Peking University Shenzhen
Graduate School.
(19) Fallows, A. J.; Singh, I.; Dondi, R.; Cullis, P. M.; Burley, G. A. Org.
Lett. 2014, 16, 4654−7.
D
DOI: 10.1021/jacs.5b09286
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX