Total synthesis of largazole-devolution of a novel synthetic strategy

Article abstract of DOI:10.1055/s-0029-1216931

The cyanobacterial isolate largazole embodies a compelling combination of a relatively simple, yet unique cyclic depsipeptide structure with remarkable levels of selective cytotoxicity against cancer cell lines versus nontransformed cells. The unique stru

Full text of DOI:10.1055/s-0029-1216931

PAPER
2873
Total Synthesis of Largazole – Devolution of a Novel Synthetic Strategy
Total
Synthesis
o
o
f
Largazole Wang, Craig J. Forsyth*
Department of Chemistry, The Ohio State University, 100 W 18th Ave, Columbus, OH 43210, USA
Fax +1(614)2921685; E-mail: forsyth@chemistry.ohio-state.edu
Received 25 June 2009
the full ketide-derived domain of 1, with an unprotected
tripeptide derivative 4, representing the polypeptide re-
gion of largazole, might be effected via the mediation of
an N-heterocyclic carbene (NHC). The expected amide
Abstract: The cyanobacterial isolate largazole embodies a compel-
ling combination of a relatively simple, yet unique cyclic depsipep-
tide structure with remarkable levels of selective cytotoxicity
against cancer cell lines versus nontransformed cells. The unique
structure of largazole inspired a strategically novel and aggressive product 2 would retain the epoxide oxygen as a stereode-
approach towards its expedient total synthesis. This involved an ini-
tial dissection into an epoxy aldehyde and an unprotected tetrapep-
fined b-hydroxy group that would be engaged in a final
macrolactonization with the residual carboxylic acid
tide, representing the polyketide and polypeptide domains of
(2→1).
largazole, respectively. These fragments were successfully joined
using NHC-mediated amidation, but subsequent assembly of the cy-
clic depsipeptide via lactonization was thwarted. A reordering of
key couplings led to a successful assembly of largazole.
O
S
6
10
N
NH
Key words: total synthesis, heterocycles, depsipeptide, N-hetero-
cyclic carbene, thiazoline
2
N
S
1
O
O
O
O
13
21
17
22
15
S
N
H
29
1
The marine natural product largazole (1) (Scheme 1) was
isolated from a cyanobacterium collected from the waters
of Key Largo, Florida, USA. The structure and prelimi-
nary biological activities of this novel cyclic depsipeptide
were reported in early 2008. The structure of largazole
embodies two distinct structural/biogenetic domains,
polyketide thioester- and polypeptide-derived units.
While the structure of 1 is unprecedented, it is the potent
and remarkably selective cytotoxic activities against hu-
man cancer cell lines that most dramatically distinguishes
largazole from clinically employed anticancer agents.1
This has spurred many total syntheses,2 as well as ana-
logues and mechanism of action studies in the past two
years. The widespread supposition that largazole’s prima-
ry mechanism of action involves inhibition of histone
deacetylases seems to be well supported.1,2a,d We were
also intrigued by the combination of the simple, yet novel
architectural features that manifested themselves in such
highly selective biological activity. Rather than targeting
one of several obvious synthetic approaches towards 1,
we were intrigued by the potential to assemble this re-
markable natural product via an unprecedented cyclodep-
sipeptide total synthesis strategy. Herein we disclose this
novel strategy and its necessary devolution to achieve a
novel total synthesis of 1.
O
HO
S
N
N
H
N
O
S
O
OH
O
S
N
H
2
O
HO
S
N
H
N
O
O
S
N
O
O
+
H2N
S
H
4
3
Scheme 1 Structure of largazole (1) and initial retrosynthetic plan
Bode and Rovis had demonstrated the utility of NHC-
catalyzed amidations with a-substituted aldehydes and
simple amines.3 We were intrigued to test whether this ap-
proach could be amenable to the generation of b-hydroxy
amides from unprotected amino acids, such as 4, without
the necessity to protect the carboxylic acid moiety.
Our initial total synthesis plan incorporated one of the
most obvious disconnections of opening the cyclic dep-
sipeptide at the lactone to generate seco-acid 2
(Scheme 1). Less generally obvious was the recognition
that amidation of the a,b-epoxy aldehyde 3, embodying
Implementation of this synthetic plan was initiated by the
synthesis of the epoxy aldehyde 3 (Scheme 2). This began
with the generation of the thioester 7 containing a tetra-
zolyl sulfone moiety under Mitsunobu conditions. Julia
SYNTHESIS 2009, No. 17, pp 2873–2880
x
x.
x
x
.
2
0
0
9
Advanced online publication: 14.08.2009
DOI: 10.1055/s-0029-1216931; Art ID: C02409SS
© Georg Thieme Verlag Stuttgart · New York

2874
B. Wang, C. J. Forsyth
PAPER
olefination using 7 and the sensitive a,b-epoxy aldehyde
8 gave a 3.3:1.0 mixture of alkenes (E/Z)-9, which were
separated at the stage of primary alcohols 10 (45% yield,
two steps). Oxidation then provided the unstable a,b-
epoxy aldehyde 3 (87%).
O
HO
OH
O
Ot-Bu
O
N3
11
BOP
N
H
N3
R
+
i-Pr2NEt
CH2Cl2
94%
O
Ot-Bu
O
Ph3P, DIAD
AcSH, THF
LiOH, MeOH
H2O, 92%, 2 steps
OH
13 R = OH
14 R = SAc
15 R = SH
N
N
n-C7H15COSH
H2N
n-C7H15
S
N
N
6
N
S
N
12
N
Ph3P, DIAD
THF, 0 °C
S
O
O
N
Ph
O
O
O
Ph
7
EDCI⋅HCl
CH2Cl2, 96%
5
64%
BocHN
N
S
OH
O
16
H
KHMDS, THF
–78 °C to r.t.
OTBS
O
O
Ot-Bu
8
O
N3
O
O
N
H
O
BocHN
N
S
OR
S
17
9 R = TBS
S
TBAF, THF, 45%
10 R = H, E/Z = 3.3:1
Ph3P, MeCN
microwave, 65 °C, 95%
DMSO, SO3⋅Py, i-Pr2NEt
CH2Cl2, 0 °C
3
O
OR2
O
Scheme 2 Facile synthesis of the largazole polyketide derivative 3
N
H
N
The complementary polypeptide derivative 4 was gener-
ated in an expedient fashion beginning with known azido
acid 114 and valine ester 12 (Scheme 3). The azide moiety
was chosen to allow implementation of a late-stage
Staudinger reduction/aza-Wittig process for the installa-
tion of largazole’s thiazoline moiety.5 This involved effi-
cient amide formation between 11 and 12 to yield primary
alcohol 13, and a subsequent two-step replacement of the
hydroxy group with a thiol (92% yield). Thioesterification
between 15 and the cysteine and glycine-derived thiazole
166 provided azido thioester 17 in excellent yield. Subse-
quent thiazoline formation was effected in a similarly ef-
ficient fashion simply by treatment of 17 with
triphenylphosphine in acetonitrile under microwave irra-
diation to afford the complete polypeptide-derived do-
main of 1 in the terminally diprotected form of 18.5
R1HN
N
S
S
18 R1 = Boc, R2 = t-Bu
4 R1 = R2 = H
TFA
CH2Cl2
Bn
i-Pr2NEt
3
N
19
Cl–
CH2Cl2 48%
S
O
HO
S
N
N
H
N
1
O
S
O
OH
O
S
N
H
2
Scheme 3 Assembly of the largazole seco-acid 2
In what was anticipated to be the final deprotection step in
the projected total synthesis of 1, both the terminal N-Boc
and tert-butyl ester protecting groups of 18 were simulta-
neously removed using trifluoroacetic acid to afford
amino acid 4 as its ammonium trifluoroacetate salt
(Scheme 3). The key joining of the two biogenetically dis-
tinct domains of largazole was then initiated by amide for-
mation between 4 and epoxy aldehyde 3. This relied upon
the use of stoichiometric amounts of the NHC generated
by sufficient amounts of Hünig’s base (i-Pr2NEt) to render
the diisopropylethylammonium carboxylate salt and liber-
ate the free amine of 4, and to deprotonate the thiazolium
salt 19, precursor to the NHC.3a This resulted in 48% iso-
lated yield of the anticipated b-hydroxy amide 2, the seco-
acid of largazole. Accompanying the formation of 2 were
substantial amounts of the corresponding dehydrated di-
ene (e.g., vii, Scheme 4). In spite of this complication, a
stoichiometric NHC-mediated amidation procedure using
a,b-epoxy aldehydes was extended to the use of an unpro-
tected amino acid to result in a b-hydroxy amide that is a
common structural element among cyclic depsipeptides.
The general mechanism for the NHC-mediated amidation
proposed by Bode involves an internal redox wherein the
NHC–aldehyde adduct i, derived from aldehyde 3 and in
situ generated NHC undergoes tautomerization to ii,
which represents an oxidation of the initial aldehyde car-
bon (Scheme 4). Opening of the epoxide utilizing electron
density resident on the nitrogen then establishes the b-
alkoxy group in iii. A pair of subsequent proton transfers
leads to the b-hydroxy carbonyl v via enolate iv. Interme-
diate v represents the activated acyl donor that is captured
by the free amino group of 4 to yield the amide 2 and re-
lease the NHC. Similar amidation attempts using catalytic
amounts of 19 failed to provide satisfactory results, due to
rapid competitive imine formation between the epoxy al-
dehyde and the amine. Thus, a sequenced combination of
Synthesis 2009, No. 17, 2873–2880 © Thieme Stuttgart · New York

PAPER
Total Synthesis of Largazole
2875
O
O
O
O
N
H
O
S
N
N
H
S
N
N
O
R
N
O
S
O
S
S
N
H
OH
O
O
N
O
N
H
O
O
R
N
H
4
2
O
vii
S
N
Bn
n-C7H15
S
H
NHC 19
N
3
H2N
4
S
OH
O
O
NHC 19
Bn
Bn
N
N
R
R
O
v
Bn
O
H
S
S
S
vi
N
R
i-Pr2NEt
i
OH
i-Pr2NEt⋅HCl
OH
O
OH
O
Bn
Bn
N
OH
Bn
O
N
R
R
N
R
iv
S
S
S
H
ii
O
O
Bn
N
R
iii
S
O
R = n-C7H15
S
Scheme 4 Proposed mechanism for the formation of amide 2 (Bode et al.)3a and diene vii under NHC mediation
aldehyde, 19, and Hünig’s base operationally preceded lactonization step to complete a total synthesis of 1. At the
the addition of the naked amino acid.
time that this synthetic strategy was conceived and imple-
mented, there were no empirical results to deter our ex-
pectation that largazole could be assembled via a
seemingly trivial final lactonization of 2. Such was not to
be the case, however. Exhaustive attempts were made to
lactonize 2 or its hydroxy epimer via Keck,7 Yamaguchi,8
Mukaiyama,9 and Mitsunobu10 protocols and variants
thereof,11 with no notable successes. Indeed, there is no
evidence among the now substantial body of published lit-
erature regarding any success in attempts to convert the
seco-acid of largazole directly into the lactone. Conforma-
tional ring strain in 1, steric blockage about the valine
carboxyl residue, and a propensity towards hydroxyl elim-
ination to generate a conjugated diene may all conspire to
deter lactonization of 2 and its hydroxy epimer under each
of the chemical methods that have been applied.
The addition of imidazole as reported by Rovis and Bode
did not provide any improvement in this system. The use
of a different NHC reagent3b was also examined, but
similar results were obtained.
Although the use of an unprotected amino acid in the
NHC-mediated amidation process was demonstrated to be
synthetically useful, there are some complicating structur-
al features of epoxy aldehyde 3 that may contribute to the
moderate yield leading to 2. These include the presence of
the g,d-alkene and the thioester in 3. The former was like-
ly to have contributed to a net dehydration to generate an
a,b-g,d-diene. For example, enolate iv could partition be-
tween the desired b-hydroxy carbonyl v and the b-elimi-
nation product conjugated diene vi (Scheme 4). An
empirical set of observations is that NHC-mediated ami-
dations of 4 with thioester–aldehyde 3 were considerably
lower yielding than similar conjugations using a,b-epoxy-
g,d-alkenyl aldehydes bearing a terminal siloxy group at
C21. Thus, the presence of an electron-withdrawing C21-
thioester was demonstrated to be detrimental to achieving
better chemical yields of the desired amidation products.
Consequently, a modification of the original total synthe-
sis plan was necessitated. The revised strategy involved
closure of the largazole cyclic depsipeptide via lactamiza-
tion of an ester-linked polypeptide/polyketide intermedi-
ate (20, Scheme 5). In contrast to our original synthetic
plan, the valine ester would be formed prior to the
polyketide-derived amide. Thus, the two key domains of
1 were reinvented as the complementary synthetic cou-
pling partners of valine-derived acid 21 and fully elaborat-
ed polyketide domain alcohol 22. Each of these was
The successful implementation of the direct convergent
coupling of amino acid 4 with epoxy aldehyde 3 resulting
in the largazole seco-acid, b-hydroxy amide 2, left a final
Synthesis 2009, No. 17, 2873–2880 © Thieme Stuttgart · New York

2876
B. Wang, C. J. Forsyth
PAPER
designed to incorporate parallel protecting groups con- the equatorially disposed valine isopropyl group in xi. In-
taining fluoren-9-ylmethanol (FmOH) on the ketide acid terception of the two epimeric diketopiperaziridinium
and peptide amino termini to facilitate simultaneous ions by the secondary alcohol of 22 would lead to (2S)-23
deprotection and subsequent lactamization.
and (2R)-23 from x and xi, respectively.
O
OH
O
S
FmOH
19, i-Pr2NEt, CH2Cl2
O
S
3
n-C7H15
S
OFm
N
77%
N
22
NH
1
H2N
O
O
O
O
O
FmocCl, dioxane
NaHCO3 (aq)
OH
N
H
n-C7H15
S
OH
N
4
96%
O
20
FmocHN
N
S
S
21
S
O
S
S
N
O
N
S
NH
OH
2,4,6-Cl3PhCOCl
i-Pr2NEt, THF, 40 °C
N
N
NH
O
+
22
21
FmocHN
then DMAP, toluene
85%
2
O
FmocHN
21
O
O
O
+
n-C7H15
S
OFm
23 2S/2R = 1:1.5
O
OH
O
S
OFm
1) Et2NH,CH2Cl2
72%
22
2) HATU, HOAt,
i-Pr2NEt, MeCN
Scheme 5 Second-generation retrosynthetic dissection of 1
1
The requisite protected carboxylate 22 was readily ob-
tained from a,b-epoxy aldehyde 3 (Scheme 2) via NHC-
mediated esterification with fluoren-9-ylmethanol
(Scheme 6). The complementary polypeptide derived
amino-protected carboxylic acid 21 was generated from 4
(Scheme 3) by simple and efficient Fmoc derivatization
(Scheme 6). The union of 21 and 22 was achieved by
Yamaguchi esterification to provide ester 23, unexpected-
ly as a 1:1.5 mixture of 2S/2R-valine epimers, respective-
ly. Without separation, these epimers were subjected to
the final two steps of deprotection and macrolactamiza-
tion, which generated a mixture of largazole (1) and 2-epi-
largazole (24) in a 1:1.5 ratio, in 72% combined yield.
These final products were then separated chromatograph-
ically.
+
O
S
N
N
O
NH
2
S
O
O
O
S
N
H
2-epi-largazole (24)
Scheme 6 Total synthesis of largazole
In summary, a novel total synthesis strategy towards lar-
gazole (1) was designed to implement a late-stage NHC-
mediated amidation of an unprotected polypeptide deriv-
ative 4 and a fully functionalized epoxy aldehyde 3. In
practice, the resultant amide emerged in modest yield as
b-hydroxy amide 2, bearing a free carboxylic acid, the
seco-acid of largazole. All attempts to affect macrolacton-
ization from hydroxy acid 2 or its hydroxy epimer were
unsuccessful in our hands, as corroborated by the original
total synthesis of 1.2 Nonetheless, the direct formation of
elaborate b-hydroxy amides from epoxy aldehydes and
unprotected amino acids using stoichiometric NHC medi-
ation, as demonstrated here, may well be applicable to the
rapid synthesis of cyclic depsipeptides bearing the b-acy-
lated amide structural motif. Several of the naturally oc-
curring cyclic depsipeptides that are structurally and
We hypothesize that the epimerization observed at the va-
line a-stereogenic center leading to (2R)-23 may occur via
the diketopiperazine intermediate xi (Scheme 7) generat-
ed in the process of the Yamaguchi esterification of 21
and 22. Either the mixed anhydride viii derived from acid
21 and trichlorobenzoyl chloride, or its DMAP derivative
ix may be susceptible to diketopiperazine formation via
involvement of the thiazoline nitrogen immediately lead-
ing to x. This initially formed structure x places the steri-
cally demanding isopropyl group in an axial orientation
on the newly formed six-membered ring. In the presence
of basic DMAP, enolization–protonation may at least par-
tially equilibrate the valine a-stereogenic center to favor
Synthesis 2009, No. 17, 2873–2880 © Thieme Stuttgart · New York

PAPER
Total Synthesis of Largazole
2877
and Kieselgel silica gel 60 F254s HP-TLC plates visualized by fluo-
rescence upon 254 nm irradiation and/or staining with anisaldehyde
reagent [90% EtOH (450 mL), aq H2SO4 (25 mL), AcOH (15 mL),
anisaldehyde (25 mL)] and/or PMA reagent [phosphomolybdic acid
(25 g), Ce(SO4)2 (87.5 g), H2O (479 mL), H2SO4 (25 mL)]. 1H NMR
and 13C NMR spectra of samples prepared in CDCl3 were refer-
enced to the residual CHCl3 peaks [d = 7.27 (1H) and d = 77.0 (13C)];
400-MHz or 500-MHz Bruker instruments were used. Optical rota-
tions were obtained on a Perkin-Elmer 241 polarimeter at the sodi-
um D line (589 nm) using 3.5 i.d. ꢀ 100 mm cylindrical glass cells
and were reported in concentration (c, g/100 mL) at 23 °C. HRMS
was performed on a Bruker MicroTOF mass spectrometer.
mechanistically related to 1 embody this specific func-
tional group array.12 The total synthesis of 1 was achieved
by a reordering of ester and amidation events, but was
complicated by the aggressive synthetic strategy of pre-
installing the thiazoline moiety in the acyl donor, N-pro-
tected carboxylic acid 21. This led to an unavoidable par-
tial epimerization at C2.
O
HO
N
H
S
O
N
Epoxy Alcohol 10
DIAD (0.49 mL, 2.37 mmol) was added to Ph3P (622 mg, 2.37
mmol) in THF (14 mL) at 0 °C. After the mixture had stirred for 5
min, a soln of alcohol 513 (562 mg, 2.09 mmol) in THF (3 mL) was
added to it. After another 5 min, a soln of thiooctanoic S-acid 614
(380 mg, 2.37 mmol) in THF (3 mL) was added. The reaction mix-
ture was kept at 0 °C for 10 min before sat. aq NH4Cl (20 mL) was
added. The product was extracted with Et2O (3 ꢀ 10 mL), dried
(MgSO4), concentrated, and purified by chromatography (EtOAc–
hexanes, 9:1); this gave sulfone 7 as a white solid; yield: 550 mg
(64%).
21
N
S
2,4,6-Cl3C6H2COCl, i-Pr2NEt
THF, 40 °C
NHFmoc
Cl
Cl
O
O
N
H
S
Cl
N
O
O
N
viii
To a –78 °C soln of sulfone 7 (320 mg, 0.78 mmol) and epoxy alde-
hyde 8 (202 mg, 0.94 mmol) in THF (7 mL) was added a 0.5 M soln
of KHMDS in toluene (1.7 mL, 0.85 mmol). After being stirred at
–78 °C for 1 h, the reaction mixture was slowly warmed to r.t. over
2 h and stirred for another 3 h. Aq phosphate buffer (pH 7) was then
added. The mixture was extracted with Et2O (2 ꢀ 5 mL), and the
combined extracts were dried (MgSO4), filtered, concentrated, and
purified by column chromatography (silica gel, EtOAc–hexanes,
10:1); this gave a mixture of (E)-9 and (Z)-9 (E/Z = 3.3:1.0, by
1H NMR analysis).
DMAP
N
S
O
NHFmoc
N
N
H
S
O
N
ix
N
S
DMAP
A 1 M soln of TBAF in THF (0.4 mL, 0.4 mmol) was added to a
soln of this mixture of 9 in THF (3 mL) and EtOAc (0.1 mL) at 0 °C.
After 30 min, pH 7 aq phosphate buffer was added. The resultant
mixture was extracted with Et2O (3 ꢀ 3 mL), the combined extracts
were dried (Na2SO4) and filtered, and the filtrate was concentrated
and purified by column chromatography (EtOAc–hexanes, 4:1);
this provided (E)-10 and (Z)-10.
NHFmoc
O
O
DMAP
HN
HN
S
S
*
N
*
N
O
O
Yield [(E)-10]: 77 mg (35%); yield [(Z)-10]: 24 mg (10%).
N
S
N
S
x
xi
Alcohol (E)-10
[a]D22 +20.4 (c 0.79, CHCl3).
NHFmoc
NHFmoc
22
22
IR (neat): 3410, 2927, 2856, 1691, 1459, 1410, 1123, 1085, 1040
cm–1.
1
24
1H NMR (500 MHz, CDCl3): d = 5.93 (dt, J = 15.5, 7.0 Hz, 1 H),
5.31 (dd, J = 15.5, 8.0 Hz, 1 H), 3.97 (dd, J = 12.5, 2.0 Hz, 1 H),
3.70 (dd, J = 12.5, 4.0 Hz, 1 H), 3.40 (dd, J = 8.0, 2.0 Hz, 1 H), 3.09
(m, 1 H), 2.93 (t, J = 7.5 Hz, 2 H), 2.54 (t, J = 7.5 Hz, 2 H), 2.36 (dt,
J = 7.5, 6.5 Hz, 2 H), 1.82 (br s, 1 H), 1.66 (m, 2 H), 1.31–1.21 (m,
8 H), 0.88 (m, 3 H).
Scheme 7 C2-Epimerization
All air-sensitive reactions were carried out under argon in oven-
dried glassware using standard syringe, cannula, and septa tech-
niques. Unless otherwise noted, all reactions were carried out at r.t.
(20–25 °C). Molecular sieves (MS) were activated by heating at
120 °C for 24 h under vacuum. THF, CH2Cl2, and toluene were pu-
rified by the Innovative Technology, Inc. Pure-Solv solvent purifi-
cation system, and MeCN and i-Pr2NEt were distilled from CaH2
under N2. Microwave-assisted reactions were performed in sealed-
tube mode on a CEM discover S-class reactor. CDCl3 was neutral-
ized by anhyd K2CO3. Other reagents were used as received from
commercial resources. Flash column chromatography was per-
formed on Silicycle SilicaFlash P60. Analytical TLC was per-
formed on Silicycle glass-backed TLC extra-hard-layer 60 plates
13C NMR (125 MHz, CDCl3): d = 199.4, 134.5, 128.5, 61.2, 59.9,
55.4, 44.2, 32.4, 31.6, 29.7, 28.9, 27.9, 25.7, 22.6, 14.1.
HRMS (ESI): m/z calcd for C15H26O3S [M + Na]: 309.1500; found:
309.1491.
Azidothiazole 17
i-Pr2NEt (230 mL, 1.31 mmol) and BOP [benzotriazol-1-yloxy-
tris(dimethylamino)phosphonium hexafluorophosphate; 553 mg,
1.25 mmol] were added to a soln of acid 114 (173 mg, 1.19 mmol)
in CH2Cl2 (9 mL). The mixture was stirred for 2 min before a soln
of L-valine tert-butyl ester 12 (217 mg, 1.25 mmol) in CH2Cl2 (1
Synthesis 2009, No. 17, 2873–2880 © Thieme Stuttgart · New York

2878
B. Wang, C. J. Forsyth
PAPER
mL) was added. The reaction mixture was stirred for 12 h and then
concentrated by rotary evaporation. The residue was purified by
column chromatography (silica gel, EtOAc–hexanes, 80:20); this
gave 13 as a colorless oil; yield: 336 mg (94%).
13C NMR (125 MHz, CDCl3): d = 174.4, 170.5, 163.2, 148.7, 121.2,
85.1, 81.8, 57.4, 41.5, 31.2, 28.3, 28.0, 24.7, 18.9, 17.6.
HRMS (ESI): m/z calcd for C23H36N4O5S2 [M + Na]: 535.2025;
found: 535.2022.
DIAD (0.33 mL, 1.6 mmol) was added to a soln of Ph3P (423 mg,
1.6 mmol) in THF (3 mL) at 0 °C. After 10 min, a soln of 13 (162
mg, 538 mmol) in THF (2 mL) was added. After another 15 min,
thioacetic acid (115 mL, 1.6 mmol) was added. The reaction mixture
was slowly warmed to r.t. over 2 h. Then Et2O (5 mL) and pH 7 aq
phosphate buffer were added. The product was extracted with Et2O
(3 ꢀ 5 mL), and the mixture was dried (MgSO4), filtered, and con-
centrated. The residue was kept at 4 °C for 10 h, before Et2O (1 mL)
was added. The resulting white precipitate was collected by filtra-
tion and washed with Et2O. The filtrate was concentrated, and then
hexanes (2 mL) were added. The resultant white precipitate was col-
lected by filtration and washed with hexanes. The filtrate was con-
centrated and the residue was dissolved in MeOH (5 mL), and then
H2O (1 mL) and LiOH·H2O (40 mg, 1.0 mmol) were added. After
the mixture had stirred for 10 min, an aq pH 7 phosphate buffer was
added. The product was extracted with Et2O (3 ꢀ 5 mL), and the
mixture was dried (MgSO4), filtered, and concentrated. The residue
was purified by flash chromatography (EtOAc–hexanes, 1:9); this
gave 15 as a yellow oil; yield: 193 mg (ca. 92%), with a small
amount (ca. 8%) of an inseparable impurity.
Fmoc-Protected Amino Acid 21
Thiazoline 18 (18.3 mg, 35.7 mmol) was dissolved in CH2Cl2–TFA
(1:1, 2 mL) and stirred at r.t. for 2 h. The solvent was removed and
the residue was dried under high vacuum for 1 h. Sat. aq NaHCO3
(0.4 mL) was added followed by FmocCl (9.2 mg, 0.43 mmol) in
1,4-dioxane (0.4 mL). After vigorous stirring of the mixture for 1 h,
1 M aq KHSO4 was added until pH 2 was achieved. The mixture
was then extracted with EtOAc (4 ꢀ 1 mL), and the combined ex-
tracts were dried (Na2SO4), filtered, concentrated, and purified by
column chromatography (silica gel, CHCl3–MeOH, 1:0 to 20:1);
this gave 21; yield: 19.8 mg (96%).
[a]D20 –36.1 (c 0.90, MeOH).
IR (neat): 3385, 2967, 1726, 1661, 1520, 1450, 1252, 1194, 1143,
1041 cm–1.
1H NMR (500 MHz, CD3OD): d = 8.18 (s, 1 H), 8.02 (br t, J = 5.5
Hz, 1 H), 7.81 (d, J = 7.5 Hz, 2 H), 7.67 (d, J = 7.5 Hz, 2 H), 7.62
(d, J = 8.5 Hz, 2 H), 7.40 (t, J = 7.5 Hz, 2 H), 7.31 (d, J = 7.5 Hz, 2
H), 4.60 (s, 2 H), 4.45 (d, J = 6.5 Hz, 2 H), 4.38 (m, 1 H), 4.24 (t,
J = 6.5 Hz, 1 H), 3.78 (d, J = –11.5 Hz, 1 H), 3.39 (d, J = –11.5 Hz,
1 H), 2.11 (m, 1 H), 1.61 (s, 3 H), 0.92 (d, J = 7.0 Hz, 3 H), 0.88 (d,
J = 7.0 Hz, 3 H).
EDCI·HCl (55 mg, 290 mmol) and DMAP (6 mg, 50 mmol) were
added to a soln of 15 (80 mg, ca. 250 mmol) and 166 (50 mg, 193
mmol) in CH2Cl2 (2 mL). After stirring for 20 min, the mixture was
concentrated, and then diluted with Et2O (3 mL) and sat. aq NH4Cl
(3 mL). The product was extracted with Et2O (3 ꢀ 3 mL), dried
(MgSO4), concentrated, and purified by flash chromatography
(EtOAc–hexanes, 85:15); this gave 17 as a pale yellow oil; yield:
103 mg (96%).
13C NMR (125 MHz, CD3OD): d = 175.4, 175.4, 172.9, 172.9,
171.7, 163.8, 157.4, 148.3, 143.8, 141.3, 127.4, 126.8, 124.8, 122.3,
119.6, 84.7, 66.7, 57.4, 57.3, 48.5, 30.5, 23.6, 18.2, 16.7.
HRMS (ESI): m/z calcd for C29H30N4O5S2 [M + Na]: 601.1555;
found: 601.1562.
[a]D22 +46.3 (c 8.6, CHCl3).
Fluorenylmethyl b-Hydroxy Ester 22
IR (neat): 3360, 2974, 2932, 2120, 1723, 1681, 1514, 1368, 1275,
1161 cm–1.
SO3·py (97 mg, 0.61 mmol) was added to a soln of epoxy alcohol
(E)-10 (87 mg, 0.31 mmol) in CH2Cl2 (0.7 mL), DMSO (120 mL,
1.7 mmol), and i-Pr2NEt (311 mL, 1.8 mmol) at 0 °C. After 1 h, an
aq pH 7 phosphate buffer was added. The resultant mixture was ex-
tracted with Et2O (3 ꢀ 1 mL), dried (MgSO4), filtered, concentrated,
and purified quickly by column chromatography (silica gel,
EtOAc–hexanes–i-Pr2NEt, 95:5:0.5); this gave epoxy aldehyde 3;
yield: 75 mg (87%). To a suspension of the thiazolium salt 19 (49
mg, 205 mmol) in CH2Cl2 (0.2 mL) was added i-Pr2NEt (71 mL, 0.21
mmol). The mixture was agitated manually for 1 min, and then add-
ed to a soln of aldehyde 3 (58 mg, 21 mmol) and fluoren-9-ylmeth-
anol (159 mg, 810 mmol) in CH2Cl2 (0.5 mL). The mixture was
stirred for 30 min before sat. aq NH4Cl (1 mL) was added. The mix-
ture was extracted with Et2O (3 ꢀ 1 mL), dried (MgSO4), filtered,
concentrated, and purified by column chromatography (silica gel,
EtOAc–hexanes, 1:9); this gave 22; yield: 76 mg (77%).
1H NMR (500 MHz, CDCl3): d = 8.09 (s, 1 H), 7.03 (d, J = 9.0 Hz,
1 H), 5.36 (br s, 1 H), 4.64 (d, J = 5.5, 2 H), 4.38 (dd, J = 8.5, 4.5
Hz, 1 H), 3.63 (d, J = –14 Hz, 1 H), 3.58 (d, J = –13.5, 1 H), 2.21
(m, 1 H), 1.71 (s, 3 H), 1.48 (app s, 18 H), 0.97 (d, J = 7.0 Hz, 3 H),
0.95 (d, J = 7.0 Hz, 3 H).
13C NMR (125 MHz, CDCl3): d = 184.2, 170.4, 170.1, 151.9, 124.2,
82.1, 67.1, 57.8, 36.3, 31.3, 28.3, 28.0, 22.0, 18.9, 17.6.
HRMS (ESI): m/z calcd for C23H36N6O6S2 [M + Na]: 579.2035;
found: 579.2024.
Thiazoline 18
A mixture of azidothiazole 17 (86 mg, 0.16 mmol), Ph3P (203 mg,
770 mmol), and MeCN (20 mL) was placed in a capped 35-mL CEM
microwave tube. The tube was flushed with argon and irradiated in
the microwave synthesizer at 65 °C for 9 h. The mixture was then
concentrated and purified by column chromatography (silica gel,
EtOAc–hexanes, 80:20); this gave 18 as a colorless yellow oil;
yield: 75 mg (95%).
[a]D20 –3.49 (c 1.58, CHCl3).
IR (neat): 3461, 2924, 2853, 1735, 1689, 1450, 1272, 1168 cm–1.
1H NMR (500 MHz, CDCl3): d = 7.81 (d, J = 7.5 Hz, 2 H), 7.62 (d,
J = 7.5 Hz, 2 H), 7.42 (m, 2 H), 7.34 (m, 2 H), 5.71 (dt, J = 15.5, 6.5
Hz, 1 H), 5.54 (dd, J = 15.5, 6.5 Hz, 2 H), 4.47 (m, 3 H), 4.14 (t,
J = 6.0 Hz, 1 H), 2.92 (t, J = 7.5 Hz, 2 H), 2.61 (d, J = 1.5 Hz, 1 H),
2.59 (d, J = 9.0 Hz, 1 H), 2.55 (t, J = 7.5 Hz, 2 H), 2.31 (q, J = 7.0
Hz, 2 H), 1.67 (m, 2 H), 1.33–1.19 (m, 8 H), 0.89 (m, 3 H).
[a]D22 –33.4 (c 3.89, CHCl3).
IR (neat): 3382, 1722, 1674, 1606, 1514, 1368, 1277, 1252, 1163,
1029 cm–1.
1H NMR (500 MHz, CDCl3): d = 7.94 (s, 1 H), 7.17 (d, J = 9.0 Hz,
1 H), 5.59 (br s, 1 H), 4.61 (d, J = 4.5 Hz, 2 H), 4.36 (dd, J = 8.5,
4.5 Hz, 1 H), 3.75 (d, J = –11.5 Hz, 1 H), 3.31 (d, J = –11.5 Hz, 1
H), 2.11 (m, 1 H), 1.57 (s, 3 H), 1.44 (s, 9 H), 1.43 (s, 9 H), 0.85 (d,
J = 7.0 Hz, 3 H), 0.82 (d, J = 7.0 Hz, 3 H).
13C NMR (125 MHz, CDCl3): d = 199.5, 172.0, 143.7, 143.6, 141.3,
132.7, 129.7, 127.9, 127.2, 125.0, 120.1, 68.5, 66.5, 46.8, 44.2,
41.6, 32.2, 31.6, 29.7, 28.9, 28.2, 25.7, 22.6, 14.1.
HRMS (ESI): m/z calcd for C29H36O4S [M + Na]: 503.2232; found:
503.2237.
Synthesis 2009, No. 17, 2873–2880 © Thieme Stuttgart · New York

PAPER
Total Synthesis of Largazole
2879
Fluorenylmethyl Ester 23
1H NMR (500 MHz, CDCl3): d = 7.68 (s, 1 H), 7.21 (d, J = 8.0 Hz,
1 H), 6.37 (dd, J = 7.5, 5.0, Hz, 1 H), 5.89 (dt, J = 15.5, 6.5, Hz, 1
H), 5.80 (app t, J = 9.0 Hz, 1 H), 5.43 (dd, J = 15.5, 8.5, Hz, 1 H),
5.08 (d, J = –17.0, 8.0 Hz, 1 H), 4.28 (d, J = –11.0 Hz, 1 H), 4.28
(dd, J = –17.0, 2.0 Hz, 1 H), 4.23 (dd, J = 16.5, 5.0 Hz, 1 H), 3.20
(d, J = –11.0 Hz, 1 H), 2.88 (dt, J = 7.5, 2.0, Hz, 2 H), 2.82 (dd,
J = –16.5, 10.5, Hz, 1 H), 2.58 (d, J = –16.5 Hz, 1 H), 2.53 (t,
J = 7.5 Hz, 2 H), 2.29 (br q, J = 7.0 Hz, 2 H), 2.13 (m, 1 H), 1.80 (s,
3 H), 1.64 (m, 2 H), 1.30–1.23 (m, 8 H), 0.97 (d, J = 7.0 Hz, 3 H),
0.89 (d, J = 6.5 Hz, 3 H), 0.88 (m, 3 H).
A soln of 2,4,6-trichlorobenzoyl chloride (1.9 mL, 12 mmol) and i-
Pr2NEt (2.2 mL, 12 mmol) in THF (120 mL) was added to acid 21 (7.2
mg, 12 mmol). The mixture was stirred at 40 °C for 5 h before being
concentrated under vacuum. A soln of ester 22 (12.0 mg, 24.9 mmol)
in toluene (120 mL) was added to the residue, followed by DMAP
(1.5 mg, 12 mmol). The mixture was stirred for 16 h before a sat. aq
soln of NH4Cl (0.5 mL) was added. The mixture was extracted with
EtOAc (3 ꢀ 0.5 mL), dried (Na2SO4), filtered, concentrated, and pu-
rified by column chromatography (silica gel, EtOAc–hexanes, 1:4
to 2:3); this yielded 23 as a mixture of C2-epimers (2R/2S = 1.5:1);
yield: 11.0 mg (85%).
13C NMR (125 MHz, CDCl3): d = 199.4, 173.6, 169.4, 167.7, 167.6,
162.2, 147.7, 134.6, 128.5, 124.2, 85.0, 72.5, 59.7, 44.2, 41.6, 40.8,
40.3, 32.2, 32.2, 31.6, 28.9, 28.9, 27.8, 26.6, 25.6, 22.6, 18.8, 18.0,
14.1.
HRMS (ESI): m/z calcd for C58H64N4O8S3 [M + Na]: 1063.3784;
found: 1063.3789.
HRMS (ESI): m/z calcd for C29H42N4O5S3 [M + Na]: 645.2210;
found: 645.2201.
Largazole (1) and 2-epi-Largazole (24)
Et2NH (0.25 mL) was added to a soln of 23 (7.0 mg, 6.7 mmol) in
CH2Cl2 (0.5 mL). After the mixture had stirred for 3 h, the solvent
was removed under vacuum and the residue was redissolved in
CH2Cl2 (0.5 mL) and i-Pr2NEt (0.1 mL). After stirring of the result-
ant soln for 10 min, the solvent was removed under vacuum and the
residue was azeotropically dried twice with toluene (2 ꢀ 2 mL). The
residue was dissolved in MeCN (7 mL), and HATU [O-(7-azaben-
Acknowledgment
We thank Profs. H. Luesch, J. Hong, and A. Phillips for their early
helpful discussions in support of this research. Financial support of
this research from The Ohio State University and the OBIC-funded
Department of Chemistry mass spectrometry facility are gratefully
acknowledged.
zotriazol-1-yl)-N,N,N¢,N¢-tetramethyluronium
hexafluorophos-
phate; 5.1 mg, 13 mmol], HOAt (1-hydroxy-7-azabenzotriazole; 1.8
mg, 13 mmol), and i-Pr2NEt (4.6 mL, 27 mmol) were added. The re-
action mixture was stirred for 12 h before being concentrated under
vacuum. The residue was purified by column chromatography
(EtOAc–hexanes–MeOH, 10:10:1) to give a mixture of 1 and 24;
yield: 3.0 mg (72%). This mixture was separated by preparative
TLC (EtOAc–hexanes–MeOH, 10:10:1) to give 1 [yield: 1.2 mg
(29%)] and 24 [yield: 1.8 mg (43%)].
References
(1) Taori, K.; Paul, V. J.; Luesch, H. J. Am. Chem. Soc. 2008,
130, 1806.
(2) (a) Ying, Y.; Taori, K.; Kim, H.; Hong, J.; Luesch, H. J. Am.
Chem. Soc. 2008, 130, 8455. (b) Nasveschuk, C. G.;
Ungermannova, D.; Liu, X.; Phillips, A. J. Org. Lett. 2008,
10, 3595. (c) Seiser, T.; Kamena, F.; Cramer, N. Angew.
Chem. Int. Ed. 2008, 47, 6483. (d) Bowers, A.; West, N.;
Taunton, J.; Schreiber, S. L.; Bradner, J. E.; Williams, R. M.
J. Am. Chem. Soc. 2008, 130, 11219. (e) Ghosh, A. K.;
Kulkarni, S. Org. Lett. 2008, 10, 3907. (f) Ren, Q.; Dai, L.;
Zhang, H.; Tan, W.; Xu, Z.; Ye, T. Synlett 2008, 15, 2379.
(g) Numajiri, Y.; Takahashi, T.; Takagi, M.; Shin-ya, K.;
Doi, T. Synlett 2008, 16, 2483.
(3) (a) Chow, K. Y.; Bode, J. W. J. Am. Chem. Soc. 2004, 126,
8126. (b) Bode, J. W.; Sohn, S. S. J. Am. Chem. Soc. 2007,
129, 13798. (c) Vora, H. U.; Rovis, T. J. Am. Chem. Soc.
2007, 129, 13796.
(4) (a) Smith, N. D.; Goodman, M. Org. Lett. 2003, 5, 1035.
(b) Avenoza, A.; Cativiela, C.; Corzana, F.; Peregrina, J. M.;
Sucunza, D.; Zurbanoa, M. M. Tetrahedron: Asymmetry
2001, 12, 949.
Largazole (1)
[a]D20 +21 (c 0.10, MeOH).
IR (neat): 3370, 3085, 2926, 2854, 1738, 1682, 1552, 1504, 1259,
1100, 1029 cm–1.
1H NMR (500 MHz, CDCl3): d = 7.76 (s, 1 H), 7.18 (d, J = 9.5 Hz,
1 H), 6.44 (dd, J = 9.5, 3.0 Hz, 1 H), 5.83 (dt, J = 15.5, 6.5 Hz, 1 H),
5.67 (app t, J = 7.0 Hz, 1 H), 5.53 (dd, J = 16.0, 7.0 Hz, 1 H), 5.30
(dd, J = –17.5, 9.5 Hz, 1 H), 4.62 (dd, J = 9.5, 3.5 Hz, 1 H), 4.29
(dd, J = –17.5, 3.0 Hz, 1 H), 4.06 (d, J = –11.5 Hz, 1 H), 3.29 (d,
J = –11.5 Hz, 1 H), 2.91 (t, J = 7.0 Hz, 2 H), 2.86 (dd, J = –16.5,
10.0 Hz, 1 H), 2.70 (dd, J = –16.0, 2.5 Hz, 1 H), 2.53 (t, J = 7.5 Hz,
2 H), 2.32 (br q, J = 7.5 Hz, 2 H), 2.11 (m, 1 H), 1.87 (s, 3 H), 1.65
(m, 2 H), 1.29–1.27 (m, 8 H), 0.88 (m, 3 H), 0.70 (d, J = 6.5 Hz, 3
H), 0.53 (d, J = 6.5 Hz, 3 H).
13C NMR (125 MHz, CDCl3): d = 199.4, 173.6, 169.4, 168.9, 167.9,
164.6, 147.5, 132.7, 128.4, 124.2, 84.5, 72.0, 57.8, 44.2, 43.4, 41.1,
40.5, 34.2, 32.3, 31.6, 28.9, 28.9, 27.9, 25.7, 24.2, 22.6, 18.9, 16.7,
14.1.
(5) (a) Chen, J.; Forsyth, C. J. Org. Lett. 2003, 5, 1281.
(b) Chen, J.; Forsyth, C. J. J. Am. Chem. Soc. 2003, 125,
8734.
(6) Moody, C. J.; Bagley, M. C. J. Chem. Soc., Perkin Trans. 1
1998, 601.
HRMS (ESI): m/z calcd for C29H42N4O5S3 [M + Na]: 645.2210;
found: 645.2201.
(7) Boden, E. P.; Keck, G. E. J. Org. Chem. 1985, 50, 2394.
(8) Inanaga, J.; Hirata, K.; Saeki, H.; Katsuki, T.; Yamaguchi,
M. Bull. Chem. Soc. Jpn. 1979, 52, 1989.
2-epi-Largazole (24)
[a]D20 +43 (c 0.16, CDCl3).
(9) Mukaiyama, T.; Usui, M.; Saigo, K. Chem. Lett. 1976, 49.
(10) Kurihara, T.; Nakajima, Y.; Mitsunobu, O. Tetrahedron
Lett. 1976, 2455.
IR (neat): 3342, 3076, 2925, 2857, 1738, 1682, 1552, 1503, 1259,
1107, 1034 cm–1.
(11) Shiina, I.; Kubota, M.; Ibuka, R. Tetrahedron Lett. 2002, 43,
7535.
Synthesis 2009, No. 17, 2873–2880 © Thieme Stuttgart · New York

2880
B. Wang, C. J. Forsyth
PAPER
(13) Kang, S. H.; Kang, S. Y.; Choi, H.; Kim, C. M.; Jun, H.;
Youn, J. Synthesis 2004, 1102.
(14) Thiooctanoic S-acid 6 was prepared from octanoyl chloride
according to Quinn’s method: Shin, H.; Quinn, D. M. Lipids
1993, 28, 73; the crude product was purified by Kugelrohr
distillation (Büchi Glass Oven B-585, 110 °C, 6–7 mbar).
(12) (a) Fujisawa Pharmaceutical Co., Ltd., Jpn. Kokai Tokkyo
Koho JP 03141296, 1991. (b) Ueda, H.; Nakajima, H.; Hori,
Y.; Fujita, T.; Nishimura, M.; Goto, T.; Okuhara, M.
J. Antibiot. 1994, 47, 301. (c) Shigematsu, N.; Ueda, H.;
Takase, S.; Tanaka, H. J. Antibiot. 1994, 47, 311. (d) Ueda,
H.; Manda, T.; Matsumoto, S.; Mukumoto, S.; Nishigaki, F.;
Kawamura, I.; Shimomura, K. J. Antibiot. 1994, 47, 315.
(e) Masuoka, Y.; Nagai, A.; Shin-ya, K.; Furihata, K.; Nagai,
K.; Suzuki, K.; Hayakawa, Y.; Seto, H. Tetrahedron Lett.
2001, 42, 41.
Synthesis 2009, No. 17, 2873–2880 © Thieme Stuttgart · New York