Total synthesis of azumamides A and E

Article abstract of DOI:10.1002/anie.200602033

A zoom in on azumamides (2): An efficient synthetic route for two of the newly discovered marine natural products, namely azumamides A and E, has been developed (see picture). The present synthesis was followed by determination of the stereochemical structure of the azumamides. (Chemical Equation Presented)

Full text of DOI:10.1002/anie.200602033

Angewandte
Chemie
ment of new anticancer agents.[2] Recently, Fusetani and co-
workers isolated azumamides A–E (1–5; Table 1) from the
marine sponge Mycale izuensis;[3] they represent five unusual
cyclic peptides that show potent inhibitory activity on histone
deacetylase (HDAC) enzymes (IC50: 0.045–1.3 mm).
Table 1: The azumamides A–E (1–5).
Azumamide
R1
R2
R3
A (1)
B (2)
C (3)
D (4)
E (5)
NH2
NH2
OH
NH2
OH
H
Me
Me
Me
H
OH
OH
H
H
Me
Natural Product Synthesis
Histone acetylation plays a fundamental role in gene
expression, interfering with transcription-based processes and
the regulation of the cell cycle.[4] The modulation of histone-
modifying enzymes is recognized as a critical factor for the
epigenetic control of oncogene transcription and the activa-
tion of tumor repressors.[4b] These combined effects can force
cancer cells toward differentiation and growth arrest, render-
ing HDAC inhibitors as unique leads for the development of
specific, noncytotoxic antitumor antibiotics.
DOI: 10.1002/anie.200602033
Total Synthesis of Azumamides A and E**
Irene Izzo,* Nakia Maulucci, Giuseppe Bifulco, and
Francesco De Riccardis*
Azumamides belong to the rare class of HDAC inhibitors
typified by a cyclic tetrapeptide core.[4b] Structurally, they
include four nonribosomal amino acid residues, three of
which are a-amino acids belonging to the d series (d-Phe, d-
Tyr, d-Ala, d-Val), while the fourth one is a unique b-amino
acid assigned as (Z)-(2S,3R)-3-amino-2-methyl-5-nonene-
dioic acid, 9-amide (Amnaa) in azumamides A (1), B (2),
and D (4), and as (Z)-(2S,3R)-3-amino-2-methyl-5-nonen-
dioic acid (Amnda) in azumamides C (3) and E (5).[3]
In memory of Luigi Gomez-Paloma and Guido Sodano
In the last years, marine invertebrates have yielded some of
the most interesting and biologically important secondary
metabolites.[1] Among these fascinating structures, cyclic
peptides demonstrate impressive antiproliferative properties
and considerable promise as lead structures for the develop-
The configurations at C2 and C3 of the Amnaa residue in
azumamide A (1) were determined through derivatization
with (+)-MTPA-Cl of the free a-methyl-b-amino acid, once
isolated and after minor modifications.[3] Comparison of the
1H NMR spectrum of the MTPA amide with those recorded
for the stereochemically defined 3-amino-2-methylhexanoic
acid models indicated a 2S,3R pattern. The same configura-
tions were proposed for the remaining azumamides (2–5) on
the basis of their similar chemical shifts and coupling constant
values observed in their NMR spectra.[3]
The novel molecular features and the interesting bioac-
tivity of azumamides prompted us to embark on an integrated
project consisting of 1) the elaboration of a potentially
stereodivergent route for the synthesis of the cyclopeptide
architecture of the azumamides and 2) the search of the
solution structure for azumamide E (5). The first objective
served to unequivocally establish the stereochemistry of the
Amnaa and Amnda residues, while the second goal was a
prerequisite for the inhibitor–enzyme docking studies.
[*] Dr. I. Izzo, Dr. N. Maulucci, Prof. F. De Riccardis
Department of Chemistry
University of Salerno
Via Ponte Don Melillo, 84084 Fisciano/Salerno (Italy)
Fax: (+39)89-969-603
E-mail: iizzo@unisa.it
dericca@unisa.it
Prof. G. Bifulco
Department of Pharmaceutical Sciences
University of Salerno
Via Ponte Don Melillo, 84084 Fisciano/Salerno (Italy)
[**] Financial support from the MIUR (“Sintesi di Sostanze Naturali e di
loro analoghi sintetici con attività antitumorale”) and the Università
degli Studi di Salerno is acknowledged. Both the institutions are
also gratefully acknowledged. We thank Prof. Nobuhiro Fusetani
and Dr. Yoichi Nakao (University of Tokyo) and A. Ganesan
(University of Southampton) for valuable discussions, and Prof.
Agostino Casapullo and Dr. Patrizia Iannece for the mass spectra.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2006, 45, 7557 –7560
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
7557

Communications
Initially, we investigated the possible synthesis of azuma-
and regioselective silylation of the diol 14 produced the
requested (2R,3S)-1-[(tert-butyldiphenylsilyl)oxy]-5-benzy-
loxy-2-methyl-3-pentanol (15).
mide A (1). Our convergent approach is shown in Scheme 1
and featured the b-amino acid 7 and the all-d tripeptide 8 as
key synthons. The synthesis of the a-methyl-b-amino-w-
Three-step stereospecific amination at C3 and Boc-
protection furnished the orthogonally triprotected aminodiol
17 in good overall yield. Pd-mediated hydrogenation and
sequential Swern oxidation afforded the (3R,4S)-3-[(tert-
butyloxycarbonyl)amino]-5-[(tert-butyldiphenylsilyl)oxy]-4-
methylpentanal (19) in 93% yield over two steps. With the
aldehyde in hand, we were ready to complete the synthesis of
the key intermediate (Z)-(2S,3R)-3-[(tert-butyloxycarbony-
l)amino]-2-methyl-5-nonenedioic
Scheme 3).
acid,
9-amide
(25,
Scheme 1. Retrosynthetic plan for azumamide A. PG=protecting
group.
carbamoyl acid 7 relied on the stereochemically flexible
Brownꢀs crotylboration reaction[5] and started from the
commercially
Scheme 2).[6] The homoallylic acetate[7] 11 was obtained
through highly diastereo- and enantioselective (d.r.
available
3-benzyloxypropanol
(9,
a
> 99%; 98% ee[8]) crotylation performed with the chiral
reagent (+)-(E)-crotyl-Ipc2-borane on the readily available 3-
benzyloxypropanal (10). The subsequent reductive ozonoly-
sis[9] furnished variable ratios of the regioisomeric acetates 12
and 13. K2CO3-mediated hydrolysis of the mixture of acetates
Scheme 3. Reagents and conditions: a) KHMDS, THF, ꢀ788C!RT,
76%; b) LiOH, H2O/THF, 708C, 91%; c) DPPA, NH4Cl, Et3N, DMF,
79%; d) HF/py, py, 60%; e) TEMPO, phospate buffer, NaClO2, NaClO,
CH3CN, 37%. KHMDS=potassium 1,1,1,3,3,3-hexamethyldisilazane;
DPPA=diphenylphosphorylazide; TEMPO=2,2,6,6-tetramethyl-1-
piperidinyloxy, free radical.
We thus performed a highly stereoselective Wittig olefi-
nation[10] by reacting the diprotected aldehyde 19 with the
ylide formed from the phosphonium salt 20.[11] The Z-alkene
21 was obtained in good yield and with no detectable
accompanying E isomer (1H NMR spectroscopy). Lithium
hydroxide mediated ester hydrolysis and concurrent DPPA[12]
carboxyl group activation gave, in the presence of ammonium
chloride, the amide 23 in 72% yield over two steps. Finally,
HF-induced desilylation and a disappointingly low-yielding
oxidation of the primary alcohol to the carboxylic acid[13]
afforded the desired N-protected acid 25 in a modest yield of
22% for two steps.
With fragment 25 in hand, we were ready for the synthesis
of the linear tetrapeptide 6. Its preparation relied on solution-
phase peptide synthesis (Scheme 4) and gave key zwitterion 6,
which was ready for the macrolactamization reaction required
to produce azumamide A (1). Although the linear zwitterion
6 contained the Amnaa residue, which was expected to
facilitate the ring closure,[14,15] the macrolactamization reac-
tion performed with the most commonly used reagents for the
cyclization of homodetic peptides (DPPA,[12] FDPP,[16] and
EDC/HOBt;[15] see Supporting Information)[17] did not give
Scheme 2. Reagents and conditions: a) oxalyl chloride, DMSO, Et3N,
CH2Cl2, 98%; b) i) (+)-Ipc2BOMe, (E)-2-butene, tBuOK, nBuLi,
BF3·Et2O, THF, ꢀ788C; ii) Ac2O, py, CH2Cl2, 80%; c) O3, CH2Cl2, then
PPh3; NaBH4, EtOH; d) K2CO3, MeOH, 61% over three steps;
e) TPSCl, DMAP, py, CH2Cl2, 90%; f) MsCl, Et3N, THF; g) NaN3, DMF,
608C, 89% over two steps; h) H2, Pt2O, EtOAc; i) Boc2O, Et3N, CH2Cl2,
92% over two steps; j) H2, Pd/C, EtOH; k) oxalyl chloride, DMSO,
Et3N, CH2Cl2, 93% over two steps. Bn=benzyl; DMSO=dimethyl
sulfoxide; Ipc=isopinocampheyl; py=pyridine; TPS=tert-butyldiphe-
nylsilyl; DMAP=4-dimethylaminopyridine; Ms=mesyl; DMF=N,N-
dimethylformamide; Boc=tert-butyloxycarbonyl.
7558
www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 7557 –7560

Angewandte
Chemie
protected form of the Amnda residue, and proceed toward
the elaboration of azumamide E (5).
As expected, better results were obtained both in the
desilylation and in the oxidation[13] steps in the absence of
carboxamide. We isolated the protected b-amino acid 31 in a
satisfying overall yield of 75%. The C9,N-diprotected Amnda
(31) was then coupled with the all-d tripeptide 28 to give the
linear tetrapeptide 32. Treatment of 32 with a mixture of
trifluoroacetic acid/dichloromethane (1:1) led to quantitative
deprotection of its carboxylate and amino groups to give 33.
This time, we were delighted to find that the FDPP-mediated
macroclactamization reaction performed on 33 produced the
C9-protected cyclotetrapeptide 34 in 37% yield.[18] The ethyl
ester present in 34 was easily hydrolyzed in the presence of
LiOH, and the target azumamide E (5) was thus isolated in
75% yield.
Azumamide A (1) was subsequently obtained in 54%
yield through DPPA-induced amidation of azumamide E (5)
in the presence of ammonium chloride and a non-nucleophilic
base. The analytical data for 1 and 5 ([a]D values, electrospray
Scheme 4. Reagents and conditions: a) N-Fmoc-d-Ala, EDC, HOBt,
DIPEA, CH2Cl2; b) Et2NH, CH3CN, 93% over two steps; c) N-Fmoc-d-
Val, EDC, HOBt, DIPEA, CH2Cl2; d) Et2NH, CH3CN, 89% over two
steps; e) 25, EDC, HOBt, DIPEA, CH2Cl2, 61%; f) TFA/CH2Cl2 (1:1),
quant. Fmoc=9-fluorenylmethoxycarbonyl; EDC=N-(3-dimethylamino-
propyl)-N’-ethylcarbodiimide; HOBt=1-hydroxybenzotriazole;
DIPEA=diisopropylethylamine; TFA=trifluoroacetic acid.
1
mass spectrometry and H and 13C NMR spectroscopy data)
measured on the synthetic samples were almost identical to
those reported by Fusetani and co-workers[3] for the natural
products.[19] A conformational analysis aimed to obtain the
experimental structure for azumamide E (5) was carried out
by means of restrained molecular mechanics (MM) and
molecular dynamics (MD) calculations (10 ns, see Supporting
Information). The solution-state structure determined from
NMR spectroscopy (Figure 1) was obtained by using 17
the expected azumamide A (1). Considering that the number
of variables associated with the cyclization step implies a “test
it and see” approach, we decided to steer away from the
synthetically bothersome carbamoyl function present in the b-
amino acid residue, postponing its introduction to a later stage
of the synthesis. Thus, we chose to transform the ethyl ester
21, an intermediate in the synthesis of 25, into the corre-
sponding
(Z)-(2S,3R)-3-[(tert-butyloxycarbonyl)amino]-2-
methyl-5-nonenedioic acid, 9-ethyl ester (31, Scheme 5), a
Figure 1. NMR solution-state conformation of azumamide E (5)
obtained by restrained MD calculations using ROESY data
(tmix =100 ms, 600 MHz, [D6]DMSO, 300 K). O mid-gray, N dark gray,
C pale gray; H atoms omitted.
spatial restraints previously collected in the ROESY spectra
([D6]DMSO, 600 MHz, tmix: 50, 100, and 150 ms); a careful
analysis of its backbone did not reveal the presence of any
identifiable turns or defined secondary structure. The pre-
sentation of this structure may be considered as an introduc-
tion to structural studies of drug–receptor interactions that
will follow in due course.
Scheme 5. Reagents and conditions: a) HF/py, py, quant.; b) TEMPO,
phospate buffer, NaClO2, NaClO, CH3CN, 75%; c) 28, EDC, HOBt,
DIPEA, CH2Cl2, 86%; d) TFA/CH2Cl2 (1:1), quant.; e) FDPP, DIPEA,
DMF, 3 days, 37%; f) LiOH, H2O/THF, 08C, 75%; g) DPPA, NH4Cl,
Et3N, DMF, 54%; FDPP=pentafluorophenyl diphenylphosphinate.
In summary we have developed a successful protocol for
the solution-phase synthesis of azumamides A (1) and E (5),
Angew. Chem. Int. Ed. 2006, 45, 7557 –7560
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
7559

Communications
FDPP (inducing the formation of an activated pentafluoro-
phenyl ester) are preferred to carbodiimide-based coupling
reagents, whose urea group is difficult to remove (the only
exception is the water soluble carbodiimide, EDC), and to
“uronium” activators (for the concurrent end-capping through
tetramethylguanidylation of the amino group).
establishing unequivocally the configurations of the azuma-
mides, and presented a three-dimensional NMR solution-
state structure of azumamide E (5). Modification of the core
structure of the azumamides and evaluation of their effects on
the three HDAC enzymes is currently underway.
[18] Using DPPA and BOP-Cl [bis(2-oxo-3-oxazolidinyl)phosphinic
chloride (see: C. van der Auwera, M. J. O. Anteunis, Int. J. Pept.
Protein Res. 1987, 29, 574 – 588), we observed the formation of
trace amounts (ESI-MS analysis) of the cyclotetrapeptide 34
(see Supporting Information).
[19] The small variations in the values in the NMR spectra observed
for synthetic azumamide E (5) were attributed to its the
sensitivity to the apparent pH value of the CD3OD solutions
(see Supporting Information).
Received: May 22, 2006
Published online: September 8, 2006
Keywords: inhibitors · macrolactamization · natural products ·
.
peptides · total synthesis
[1] For recent reviews, see: J. W. Blunt, B. R. Copp, M. H. G. Munro,
P. T. Northcote, M. R. Prinsep, Nat. Prod. Rep. 2006, 23, 26 – 78,
and references therein.
[2] For recent reviews, see: Y. Hamada, T. Shiori, Chem. Rev. 2005,
105, 4441 – 4482, and references therein.
[3] Y. Nakao, S. Yoshida, S. Matsunaga, N. Shindoh, Y. Terada, K.
Nagai, J. K. Yamashita, A. Ganesan, R. W. M. van Soest, N.
Fusetani, Angew. Chem., DOI: 10.1002/ange.200602047; Angew.
Chem. Int. Ed., DOI: 10.1002/anie.200602047.
[4] a) S. Minucci, P. G. Pelicci, Nat. Rev. Cancer 2006, 6, 38 – 51;
b) M. Biel, V. Washolowski, A. Giannis, Angew. Chem. 2005,
117, 3248 – 3280; Angew. Chem. Int. Ed. 2005, 44, 3186 – 3216;
c) T. A. Miller, D. J. Witter, S. Belvedere, J. Med. Chem. 2003, 46,
5097 – 5116; note that the HDAC inhibition profiles observed for
azumamides E and C contradict the assumption that carboxylic
acids at the “zinc binding” domain give poor enzymatic activity.
See Ref. [4b].
[5] H. C. Brown, K. S. Bhat, J. Am. Chem. Soc. 1986, 108, 5919 –
5923; for recent reviews, see: M. Lei, M. P. Sibi, Tetrahedron
2002, 58, 7991 – 8035, and references therein.
[6] Experimental procedures and spectroscopic data can be found in
the Supporting Information.
[7] The acetylation was necessary for purification purposes.
[8] The enantiomeric excess of compound 11 was evaluated by
Mosher ester analysis of the alcohol resulting from acetate
hydrolysis (see Supporting Information).
[9] A. B. Benowitz, S. Fidanze, P. L. C. Small, Y. Kishi, J. Am. Chem.
Soc. 2001, 123, 5128 – 5129.
[10] D. R. Williams, K. Shamim, J. P. Reddy, G. S. Amato, S. M. Shaw,
Org. Lett. 2003, 5, 3361 – 3364.
[11] Compound 20 was easily prepared from commercially available
ethyl 4-bromobutyrate.
[12] K. Uwai, K. Ohashi, Y. Takaya, T. Ohta, T. Tadano, K. Kisara, K.
Shibusawa, R. Sakakibara, Y. Oshima, J. Med. Chem. 2000, 43,
4508 – 4515.
[13] M. Zhao, J. Li, E. Mano, Z. Song, D. M. Tschaen, E. J. J.
Grabowski, P. J. Reider, J. Org. Chem. 1999, 64, 2564 – 2566; the
oxidation reaction on 24 proved to be capriciously batch-
dependant. It is reported therein that the presence of carbon–
carbon double bonds can be troublesome. Two-step procedures
(from the alcohol to the aldehyde and from the aldehyde to the
carboxylic acid) gave yields of less than 30%. Fewer problems
were observed for the oxidation of 30 (Scheme 5).
[14] M. P. Glenn, M. J. Kelso, J. T. A. Tyndall, D. P. Fairlie, J. Am.
Chem. Soc. 2003, 125, 640 – 641.
[15] For recent reviews on peptide cyclizations, see: a) J. W. Davies, J.
Pept. Sci. 2003, 9, 471 – 501; b) P. Li, P. P. Roller, J. Xu, Curr. Org.
Chem. 2002, 6, 411 – 440, and references therein.
[16] a) S. Chen, J. Xu, Tetrahedron Lett. 1991, 32, 6711 – 6714; b) J.
Dudash, Jr., J. Jiang, S. C. Mayer, M. M. Joullie, Synth.
Commun. 1993, 23, 349 – 356.
[17] In the macrolactamization reaction, the reagents DPPA (the
successor of the conventional “azide” coupling methods) and
7560 www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 7557 –7560