Enantioselective synthesis of kedarcidin chromophore aglycon in differentially protected form.

Full text of DOI:10.1002/1521-3773(20020315)41:6<1062::AID-ANIE1062>3.0.CO;2-8

COMMUNICATIONS
of 1. This route targeted the C1 C9 fusion bond for
disconnection and featured the transannular addition of a
vinyllithium intermediate to an internal diacetylene group.[9]
In theory, this approach was well suited for the synthesis of 1
itself, for it would allow for the assembly of the ansa-bridged
macrocycle prior to the reactive core functionality, although
the feasibility of the proposed anionic transannular addition
reaction, within the context of a functionally more complex
macrolactone structure, was uncertain. Here, we demonstrate
that the transannular cyclization strategy is effective within an
ansa-bridged macrolactone substrate and implement it suc-
cessfully in the first enantioselective synthesis of kedarcidin
chromophore aglycon, in differentially protected form (struc-
ture 2; MOM ˆ methoxymethyl, TIPS ˆ triisopropylsilyl,
TES ˆ triethylsilyl).
Enantioselective Synthesis of Kedarcidin
Chromophore Aglycon in Differentially
Protected Form**
Andrew G. Myers,* Philip C. Hogan,
Alexander R. Hurd, and Steven D. Goldberg
Kedarcidin is a highly reactive and complex chromoprotein
enediyne natural product with potent antiproliferative and
antibiotic activities.[1] It is a member of the subset of enediyne
[3]
antibiotics that includes neocarzinostatin,[2] C-1027, macro-
momycin,[4] actinoxanthin,[5] and maduropeptin.[6] Each of
these agents is composed of protein and small-molecule
(chromophore) components, which form a 1:1 complex. The
chromoprotein agents are thus distinguished from the
nonproteinaceous (10-membered) cyclic enediyne agents
(calicheamicin g1, esperamicin, and dynemicin A),[7] and
further by the fact that their chromophoric components,
when separated from the binding protein, are exceedingly
unstable. Kedarcidin chromophore (1) has extremely limited
iPrO
OMOM
O
CH3O
CH3O
HN
O
N
Cl
O
OTIPS
iPrO
OH
O
O
CH3O
N(CH3)2
OH
CH3O
HN
H3C
O
TESO
O
O
N
2
Cl
O
O
1
O
9
Our retrosynthetic analysis is outlined in Scheme 1. Final-
stage operations for the synthesis of 2 were envisioned to
involve C10 hydroxy group directed epoxidation of the
O
O
OH
H3C
HO
O
1
À
C8 C9 olefin of the ansa-bridged bicyclo[7.3.0]dodecadiene-
CH3
diyne 3, selective protection of the C10 hydroxy group, and
[9, 10]
À
dehydration to form the C4 C5 olefin
steps successfully
executed within our earlier, albeit simpler model system. The
substrate for these operations, compound 3, was imagined to
arise by cyclization of the bromide 4, induced by lithium
halogen exchange. Although potentially complicated by
deprotonation reactions (see below), as well as competing
carbonyl addition reactions, we felt that the proposed
lithium halogen exchange/transannular addition held suffi-
cient merit to pursue the development of a synthetic route to
4. Further, three s bonds within the precursor 4 were targeted
stability in solution and has been shown to undergo sponta-
neous cycloaromatization in the presence of sodium borohy-
dride.[1d] These issues of reactivity, coupled with structural
features such as the chloropyridine-containing ansa-bridge,
the highly unusual epoxy bicyclo[7.3.0]dodecadienediyne
core, and the appended naphthoic acid amide, kedarose, and
mycarose substituents, contribute to make the synthesis of 1 a
formidable challenge. In addition to our own efforts to
develop a synthesis of 1, the Hirama group has reported the
synthesis of core models,[8a] an ansa-bridged macrolactone
model,[8b] and subsequent glycosylation studies of this macro-
lactone model structure.[8c]
À
for disconnection: the internal diacetylene linkage (C7 C8),
À
À
the lactone C O bond, and the bromoenyne linkage (C1 C2).
These could reasonably be formed in any order (six different
sequences); two were investigated here. Both sequences
converged upon the intermediate 5 in a proposed Glaser
reaction[11] to form the last of the three s bonds targeted and
thus to yield the macrobicycle 4. The paths then diverged in
the ordering of the remaining two s bond forming steps, one
involving macrocyclization by lactonization, and the other by
In a new strategy for the synthesis of the reactive core
structures of the chromoprotein enediyne agents, we recently
described the synthesis of a model compound that comprises
the epoxy bicyclo[7.3.0]dodecadienediyne core functionality
[*] Prof. A. G. Myers, P. C. Hogan, A. R. Hurd, S. D. Goldberg
Department of Chemistry and Chemical Biology
Harvard University
À
Pd-catalyzed C C bond formation to generate the macro-
lactone. Both sequences were initiated by the same three
starting materials (6 8) and incorporated an acetonide
protective group within the diacetylene intermediate 7 to
position the hydroxymethyl and terminal alkyne substituents
in an orientation favorable for macrocyclization. Disconnec-
tion of the pyridyl ether bond within the intermediate 6
Cambridge, Massachusetts 02138 (USA)
Fax : (‡1)617-495-4976
E-mail: myers@chemistry.harvard.edu
[**] Financial support from the National Institutes of Health is gratefully
acknowledged. We thank Dr. Richard Staples and Mr. Andrew Haidle
for the X-ray analysis of compound 20.
1062
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
1433-7851/02/4106-1062 $ 17.50+.50/0
Angew. Chem. Int. Ed. 2002, 41, No. 6

COMMUNICATIONS
OMOM
O
OH
iPrO
OH
O
Cl
N
Cl
N
a–c
OCH3
points of
disconnection
CH3O
HO
CH3O
HN
MOMO
12
11
O
N
d, e
Cl
O
OTIPS
4
OH
2
O
NH2
O
5
NHBoc
Cl
N
Cl
N
f, g
10
OCH3
9
8
3
HO
HO
O
OCH3
MOMO
13
10
iPrO
OMOM
O
iPrO
OMOM
O
Scheme 2. Synthesis of b-azatyrosine residue 10. a) MOMCl (1.15 equiv),
iPr2NEt (3.0 equiv), CH2Cl2, 238C, 81%; b) oxalyl chloride (2.0 equiv),
DMSO (4.0 equiv); Et3N (9.0 equiv), CH2Cl2, À78 !08C, 92%; c) (S)-(À)-
Carreira ligand (1.1 mol%), Ti(OiPr)4 (0.5 mol%), 3,5-di-tert-butylsalicylic
acid (1 mol%), methyl trimethylsilyl ketene acetal (2.0 equiv), 90%, 94%
ee; d) PPh3 (1.2 equiv), diethyl azodicarboxylate (1.2 equiv), hydrazoic acid
in toluene (1.0n, 1.2 equiv); e) PPh3 (1.2 equiv), H2O (10 equiv), 238C,
77%; f) 12n HCl, methanol, 238C; g) di-tert-butyl dicarbonate (1.5 equiv),
Et3N (2.2 equiv), methanol, 238C, 88% (two steps).
CH3O
CH3O
CH3O
TBSO
HN
H
CH3O
HN
H
Cl
N
Cl
N
O
O
O
O
O
O
2
1
OTIPS
OTIPS
OH
Br
OH
Br
H
TBSO
8
7
5
4
H
The fragment 7 was prepared from the known intermediate
14[17] by a five-step sequence, as shown in Scheme 3. Cleavage
of the acetonide protective group and selective protection of
NHBoc
NHBoc
Cl
N
Cl
N
O
O
OCH3
H
O
OCH3
HO
O
HO
10
O
H3C CH3
Br
TBSO
O
CH3
CH3
Br
TBS
iPrO
O
S
O
O
O
HO
OPiv
O
Br
6
a, b
Br
TMS
TMS
TES
OH
H
OH
7
9
OMOM
O
14
15
TBS
H
CH3O
CH3O
OH
8
c
Scheme 1. Retrosynthetic analysis of 2. TBS ˆ tert-butyldimethylsilyl,
Boc ˆ tert-butoxycarbonyl.
HO
PivO
O
O
d, e
H
TMS
CH3
CH3
CH3
CH3
produced the two key fragments 9 and 10. This was considered
to be an advantageous feature of the approach, for it
established the rather difficult pyridyl ether bond early in
the sequence, rather than later, where stability issues would
severely limit the options available for that transformation.
Each of the four components (7 10) defined by this sequence
of disconnections was prepared in multi-gram quantities. The
naphthoic acid 8 was synthesized by protection of an
intermediate from a prior route.[12] The cyclic sulfate 9 was
also prepared by a prior route with d-erythronolactone as
starting material.[9] The remaining components, 7 and 10, were
prepared in optically pure form, as follows.
The crystalline b-azatyrosine residue 10 was synthesized on
a large scale by the highly practical and convenient Carreira
Singer aldol methodology[13] (>30 g in one reaction,
11[14] !12, Scheme 2), adapted here for the synthesis of a b-
amino acid by a subsequent Mitsunobu azidation reaction,[15]
followed by Staudinger reduction (12 !13).[16] The enantio-
selectivity of the asymmetric acetate aldol reaction was
reproducibly high (92 94% ee); one recrystallization of the
solid product provided optically pure material, in 90% yield.
Hydrolysis of the methoxymethyl ether 13 and Boc protection
of the amino group then afforded the desired b-azatyrosine
fragment 10 for subsequent coupling.
O
O
7
16
TES
H
Scheme 3. Synthesis of alcohol 7. a) 3n HCl, MeOH, 238C 88%;
b) pivaloyl chloride (1.13 equiv), pyridine, 238C, 87%; c) p-TsOH
(0.2 equiv), 2,2-dimethoxypropane (10 equiv), DMF, 708C, 93%;
d) LiHMDS (1.5 equiv), THF, À788C; TESCl (1.5 equiv), À78 !238C
92%; e) K2CO3 (2 equiv), MeOH, À108C, 18 h, 84%. TMS ˆ trimethyl-
silyl, HMDS ˆ bis(trimethylsilyl)amide.
the primary hydroxy group of the resulting triol by pivaloyl-
ation provided the diol 15. Reintroduction of an acetonide
protective group, now to mask the secondary and tertiary
hydroxy groups, was accomplished by using 2,2-dimethoxy-
propane and catalytic p-toluenesulfonic acid in DMF at 708C
to afford 16. C-Deprotonation of 16 with lithium bis(trime-
thylsilyl)amide and trapping of the resultant acetylide anion
with chlorotriethylsilane gave the corresponding differentially
silylated dialkyne. Selective cleavage of the pivaloate and
trimethylsilylalkyne groups then occurred upon exposure to
potassium carbonate in methanol at À108C to produce 7 in
55% yield (over 5 steps). This sequence served to transfer the
acetonide group of 14 to the contrathermodynamic position it
occupies in 7, as well as to transfer the location of the
Angew. Chem. Int. Ed. 2002, 41, No. 6
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
1433-7851/02/4106-1063 $ 17.50+.50/0
1063

COMMUNICATIONS
acetylenic (silyl) protective group, necessary to allow the
In an alternative and preferred sequence, ester bond
formation between 6 and 7 was accomplished first, and
macrocyclization was achieved in a subsequent Pd-mediated
process, after a particularly useful and efficient in situ C-
stannylation reaction (lower pathway, Scheme 4). Hydrolysis
of the methyl ester 6 and activation of the resultant carboxylic
acid allowed for smooth coupling with the primary hydroxy
group of 7 to afford ester 18 in 87% yield (2 steps). Initially,
we sought to close the macrocyclic ring by direct Pd-mediated
À
desired sequence of C C bond-forming reactions.
In the first fragment-assembly step the b-azatyrosine
residue 10 was coupled with the cyclic sulfate 9 to form, after
sulfate hydrolysis and hydroxy protection, the pyridyl ether 6
(Scheme 4).[9] Thus, addition of 9 to the potassium phenolate
derived from 10 led to smooth and selective cleavage of the
allylic sulfate ester bond (regioselectivity 9:1). Sulfate ester
hydrolysis[18] and subsequent protection of the resultant
secondary alcohol with tert-butyldimethylsilyl chloride[19] then
afforded 6. An apparent kinetic effect operated in the
silylation reaction which served to further enrich the mixture
in the desired allylic pyridyl ether regioisomer (6, selectivity
ꢀ20:1). The overall coupling sequence was quite efficient
(73% from 9) and routinely provided 6 in amounts of 5 15 g.
The pyridyl ether 6 served as a common point of departure
for two different sequences to the macrolactone 19
(Scheme 4). The first employed a more traditional macro-
lactonization reaction. For this sequence, it proved necessary
to transform the methyl ester 6 into the corresponding
trifluoroethyl ester prior to Pd-mediated coupling (see upper
pathway). Sonogashira coupling of the resultant trifluoroethyl
ester with the terminal alkyne of 7 then proceeded with
selective replacement of the trans-bromide to furnish the
bromoolefin 17 in 50 60% yield.[20] Bis-coupling in this
reaction was a problem that, despite many efforts, was never
completely suppressed. Saponification of the trifluoroethyl
ester 17 (the corresponding methyl ester underwent compet-
itive dehydrobromination) with lithium hydroperoxide clean-
ly afforded the corresponding hydroxy acid, which was readily
cyclized by using modified Yamaguchi conditions[21] at high
dilution (ca. 0.5 mm) to form 19 in 70% yield (efforts to
cyclize 17 directly were not successful).
À
C C bond formation within 18 (Castro Stephens Sonoga-
shira coupling). Although the desired macrolactone (19)
could be obtained in this way, yields were low (<40%) and
difficult to reproduce. Considering it likely that metalation of
the hindered terminal alkyne of 18 was problematic in the
coupling reaction, we investigated ways to activate this group
prior to cyclization. Stannylation with the reagent diethyl-
aminotrimethylstannane (caution! toxic), introduced by Lap-
pert et al.[22] and used by Stille[23] in his initial coupling studies,
proved extraordinarily successful within the multifunctional
substrate 18. Significantly, we found that the reaction could be
modified by the use of toluene as solvent (as opposed to the
use of neat diethylaminotrimethylstannane), and was rapid at
ambient temperature with a moderate excess of reagent
(3.0 equiv, ꢁ0.1m). Concentration of the reaction solution
removed volatile by-products as well as excess reagent.
À
Neither the alkynyl C H nor the amido proton resonance
was detected in H NMR analysis of the residue. Sequential
1
addition of ethyl ether, [Pd2(dba)3], and tri-2-furylphospha-
ne[23c] to the concentrated stannylation product then led to
smooth formation of the macrolactone 19 within 12 h at 238C
(68% yield from 18). A noteworthy feature of this cyclization
reaction was the fact that high dilution conditions were not
necessary to obtain good yields of 19, making this the method
of choice for the produc-
tion of 19. We have found
that the procedure de-
scribed for the in situ
BocHN
Cl
N
CO2CH2CF3
OH
stannylation of 18 is also
effective by using other,
structurally diverse ter-
O
c–e
f, g
O
TBSO
NHBoc
CH3
CH3
O
Br
Cl
N
minal alkynes as sub-
strates, and that it is tol-
erant of a range of func-
tional groups (ester,
amide, halide, and epox-
ide functional groups,
among others). We rec-
ommend it as a conven-
ient alternative to lithia-
tion/stannyl halide trap-
ping procedures. The
products of both cou-
pling macrocyclization
sequences (19, Scheme 4)
were shown to be indis-
9
+
NHBoc
O
TBS
O
O
Cl
N
17
TES
a, b
O
TBSO
O
O
OCH3
CH3
CH3
O
Br
TBSO
6
10
NHBoc
TBS
Br
Br
TES
Cl
N
19
TBS
O
O
O
H
i, j
c, h
TBSO
O
O
CH3
CH3
18
Br
Br
H
TBS
TES
Scheme 4. Synthesis of macrolactone 19. a) 10 (1.2 equiv), KHMDS (1.1 equiv), CH3CN, À408C; 9 (1.0 equiv), CH3CN,
À408C, 88%; b) TBSCl (1.5 equiv), Im (6.0 equiv), DMF, 238C, 83%; c) LiOH (2.0 equiv), 30% aqueous H2O2/THF
(1:3), 238C, 20 h; Na2SO3; d) trifluoroethanol (5.0 equiv), EDC ¥ HCl (1.5 equiv), DMAP (0.1 equiv), CH2Cl2, 238C, 86%
(2 steps); e) 8, (1.22 equiv), [Pd(PPh3)4] (0.1 equiv), CuI (0.32 equiv), Et3N (2.2 equiv), tert-butyl methyl ether, 238C,
61%; f) LiOH (2.0 equiv), 30% aqueous H2O2/THF (1:3), 238C, 30 min; g) 2,4,6-Trichlorobenzoyl chloride (20 equiv),
DMAP (10 equiv), Et3N (20 equiv), toluene, 508C, 70% (two steps); h) 7 (1.05 equiv), EDC ¥ HCl (2.4 equiv), DMAP
(0.1 equiv), CH2Cl2, 238C, 87% (2 steps); i) Et2NSn(CH3)3 (3.0 equiv), toluene, 238C; j) [Pd2(dba)3] (0.1 equiv), tri-2-
furylphosphane (0.8 equiv), Et2O, 23 8C, 68% (2 steps). Im ˆ imidazole, dba ˆ dibenzylideneacetone, EDC ˆ 1-[3-
(dimethylamino)propyl]-3-ethylcarbodiimide, DMAP ˆ 4-dimethylaminopyridine.
1
tinguishable by H NMR
analysis.[24] Parenthetical-
ly, we note that the ace-
tonide protective group
1064
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
1433-7851/02/4106-1064 $ 17.50+.50/0
Angew. Chem. Int. Ed. 2002, 41, No. 6

COMMUNICATIONS
within 17 and 18 was critical to the success of both macro-
cyclization reactions; presumably it serves as a preorganizing
element that favors ring closure. Substrates with alternative
noncyclic protective groups cyclized more slowly, and in lower
yield.
With ready access to gram-quantities of the macrolactone
intermediate 19, efforts turned toward the introduction of the
naphthoamide residue, and protective group manipulations
necessary to allow for Glaser cyclization to form the trans-
annular cyclization precursor 4 (Scheme 5). Toward this end,
Concentration afforded the trifluoroacetate salt, which was
treated with a solution of the naphthoic acid 8, previously
activated for coupling (EDC¥HCl, HOBT), to form the
corresponding naphthoic acid amide derivative in 81% yield
for the two-step sequence. Selective silyl protection of the C13
secondary hydroxy group then served to fully differentiate
each hydroxy group in the molecule (!5, 81%).[24]
At this point a second macrocyclization reaction was
achieved by employing carefully optimized, modified Eglin-
ton conditions[26] to afford the macrobicyclic product 4
(Scheme 5).[24] Thus, slow addition of a solution of the
macrolactone 5 in pyridine/THF (2:1) to a prewarmed
(458C) suspension of copper(ii) acetate and copper(i) iodide
in pyridine/THF (2:1) produced the transannular cyclization
substrate 4 in 63% yield.[24] This product proved to be
exceedingly unstable when concentrated and, in this regard,
was the most unstable intermediate in the entire synthetic
sequence.[27] Although 4 could be stored as a dilute solution in
toluene (À208C), typically, this was not done; rather, 4 was
subjected to direct metalation transannular cyclization
(Scheme 6).
NHBoc
Cl
N
O
O
O
O
a
RO
10
Br
19
CH3
CH3
O
H
H
20 R = H
b
21 R = TBS
c–e
a, b
3
4
iPrO
OMOM
O
CH3O
CH3O
Cl
iPrO
OMOM
O
HN
H
N
CH3O
CH3O
HN
D (≈ 60%)
O
O
O
O
N
OTIPS
OH
Cl
O
OTIPS
OH
c, d, b
TBSO
4
Br
H
O
5
H
f
HO
(> 90%)
D
iPrO
OMOM
O
[D2]3
CH3O
Scheme 6. Transannular cyclization and deuterium labeling. a) LiHMDS
(3.5 equiv), THF, toluene, À968C, tBuLi (1.7m, 5.0 equiv); HOAc/THF
(1:10), 45 60%; b) Et3N ¥ 3HF (75 equiv), CH3CN, 238C, 75 %; c) CD3OD,
238C; d) LiHMDS (3.5 equiv), THF, toluene, À968C, tBuLi (1.7m,
5.0 equiv); [D4]acetic acid/THF (1:3).
CH3O
Cl
HN
H
N
O
O
O
OTIPS
OH
Br
TBSO
Lithium halogen exchange within the intermediate 4 was
potentially complicated by at least three competing proton-
transfer events (from the secondary amide, the tertiary
hydroxy group, and the a-position of the lactone). In the
optimized procedure for transannular cyclization (Scheme 6),
these competing processes were mitigated by prior treatment
of 4 with a solution of LiHMDS in THF. After 5 min, a
solution of tert-butyllithium in pentane was added, followed
immediately (<3 s) by a quenching solution of acetic acid in
THF. After cleavage of the secondary TBS ether with
triethylamine trihydrofluoride and purification by flash col-
umn chromatography, the transannular cyclization product 3
was obtained, reproducibly, in 38 45% yield.[28] Some insight
into the mechanistic details of this process was gained by
conducting a deuterium labeling experiment (Scheme 6).
Treatment of 4 with [D4]methanol followed by concentration
from toluene (to remove methanol), dilution with THF, and
sequential addition at À968C of solutions of LiHMDS
4
Scheme 5. Glaser cyclization to prepare the transannular cyclization
precursor 4. a) TBAF (4.5 equiv), o-nitrophenol (5.0 equiv), THF, 238C,
98%; b) TBSCl (2.0 equiv), Im (6.0 equiv), DMF, 238C, 97%; c) 33%
TFA/CH2Cl2, H2O (2.0 equiv), 238C; d) 7 (1.3 equiv), EDC ¥ HCl
(3.0 equiv), HOBT (1.0 equiv), Et3N, CH2Cl2, 238C, 12 h, 81% (2 steps);
e) TIPSOTf (5.0 equiv), Et3N (15.0 equiv), CH2Cl2, À658C, 81%;
f) Cu(OAc)2 (60.0 equiv), CuI (15.0 equiv), pyridine, THF, 458C, 65%.
TBAF ˆ tetrabutylammonium fluoride, HOBTˆ 1-hydroxybenzotriazole.
the macrolactone 19 was globally desilylated by using o-
nitrophenol-buffered tetrabutylammonium fluoride in THF
(98% yield)[9] to provide the crystalline derivative 20, whose
structure was secured by single crystal X-ray analysis.[24] After
reintroduction of the TBS ether that had been cleaved in the
prior step (!21),[19, 24] the tert-butyl carbamate and acetonide
groups were removed selectively by exposure of 21 to a
solution of 33% trifluoroacetic acid in dichloromethane.
Angew. Chem. Int. Ed. 2002, 41, No. 6
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
1433-7851/02/4106-1065 $ 17.50+.50/0
1065

COMMUNICATIONS
(3.5 equiv in THF), tert-butyllithium (5.0 equiv in pentanes),
and [D4]acetic acid/THF (1:3) gave rise to the isotopically
labeled transannular cyclization product [D2]-3, after cleavage
allylic alcohol estimated to be 758 in 3 and 608 in the model
structure). It proved to be nontrivial to distinguish 22 from the
desired epoxide 24 spectroscopically; definitive assignment of
22 was achieved in two ways. First, dehydration of the C4
tertiary hydroxy group (Martin sulfurane,[29] benzene-1,4-
cyclohexadiene (1:1)) provided the cycloaromatized elimina-
tion product 23;[28] its enediyne precursor was not observable.
1H NMR analysis of 23 was consistent only with the assigned
structure, and not the cycloaromatized product that would
have arisen from 2. A more definitive structural assignment
was made by epoxidation/triethylsilylation of [8-D]-3.
1H NMR analysis of the resultant epoxide ([8-D]-22) showed
no vinyl proton resonance which conclusively established that
1
of the TBS ether, as before. H NMR analysis showed that
deuterium incorporation at C8 was virtually quantitative.
Interestingly, stereoselective incorporation of deuterium at
the a-position of the lactone had also occurred, albeit to a
lesser degree (ꢁ60%, stereochemistry assigned tentatively as
shown on the basis of 1H,1H NMR coupling constants). From
this result, it is reasonable to conclude that transannular
cyclization of 4 had proceeded through a polyanionic inter-
mediate, perhaps even a tetraanion.
All that remained to complete the synthesis of 2 was the
sequence of final stage operations initially outlined (see
above; Scheme 2). However, at this point we encountered a
serious problem when V-catalyzed hydroxy group directed
epoxidation of 3 was found to provide the C1 C12 epoxide
22[24] (after selective triethylsilylation of the C10 hydroxy
group), and not the desired C8 C9 epoxide (Scheme 7). This
stood in contrast to our earlier findings in a model system
lacking the ansa bridge, and was felt to reflect a less optimal
dihedral angle for the V-catalyzed process (dihedral angles of
À
epoxidation had occurred at C1 C12 of 3. A wide range of
alternative epoxidizing reagents was screened; however, most
of these led to decomposition of the substrate. Notably,
Jacobsen×s [MnIII(R,R)-(salen)] catalyst,[30] Yamamoto×s
asymmetric vanadium catalyst,[31] and dibutyltin oxide-tert-
butylhydroperoxide[32] each formed the undesired epoxide 22
as the only isolable product.
Further investigation revealed that the desired epoxide
24[24] was formed in preference to 22 when more hindered
hydroperoxides were used in the V-catalyzed process. Both
triphenylmethyl hydroperoxide[33] and 1,1-diphenylethyl hy-
droperoxide[34] gave the desired epoxide (24), with only a
trace of 22 (Scheme 8). The use of 1,1-diphenylethyl hydro-
peroxide led to an increased rate of reaction, and an improved
iPrO
OMOM
O
CH3O
CH3O
HN
O
N
Cl
O
O
OMOM
O
OTIPS
OH
iPrO
a
3
O
CH3O
CH3O
HN
TESO
O
22
N
Cl
O
OTIPS
OH
b
a
b
3
2
O
iPrO
OMOM
O
O
TESO
CH3O
24
CH3O
HN
Scheme 8. Epoxidation of
3 to afford the desired epoxide 24 and
O
subsequent dehydration to form 2. a) [VO(acac)2] (0.2 equiv),
CH3CPh2OOH (1.3 equiv), benzene, 238C; TESCl (50 equiv), Im
(100 equiv), CH2Cl2, 238C, 32%; b) Martin sulfurane (10 equiv), benzene,
238C, 83%.
N
Cl
O
O
OTIPS
OH
O
TESO
yield (32% vs. 20%, after in situ silylation), and was therefore
preferred.[35] With the desired epoxide 24 in hand, completion
of the synthesis of 2 was achieved by dehydration in the
presence of the Martin sulfurane. Synthetic 2 provided
spectral data in complete accord with the assigned struc-
ture.[24] In addition, cycloaromatization of 2[25] in the presence
1,4-cyclohexadiene produced the expected aromatic prod-
uct[28] which further confirms the assignment of 2.
iPrO
OMOM
O
CH3O
CH3O
HN
O
N
OTIPS
Cl
O
O
O
In conclusion, we have developed a convergent, enantiose-
lective synthetic route to kedarcidin chromophore aglycon in
differentially protected form (2). The route is 25 steps in the
longest linear sequence, with an average yield of 82% per step
(overall yield 1%).
TESO
23
Scheme 7. Epoxidation of3 to form the undesired epoxide 22; dehydration
and cycloaromatization. a) [VO(acac)2] (0.2 equiv), tBuOOH (20 equiv),
benzene, 238C; TESCl (50 equiv), Im (100 equiv), CH2Cl2, 238C; b) Martin
sulfurane (10 equiv), benzene-1,4-cyclohexadiene (1:1), 238C, 10% (four
steps). acac ˆ acetyl acetonate.
Received: December 12, 2001 [Z18370]
1066
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
1433-7851/02/4106-1066 $ 17.50+.50/0
Angew. Chem. Int. Ed. 2002, 41, No. 6

COMMUNICATIONS
[1] Isolation: a) K. S. Lam, G. A. Hesler, D. R. Gustavson, A. R. Cross-
well, J. M. Veitch, S. Forenza, K. Tomita, Antibiotics 1991, 44, 47 2;
b) S. J. Hofstead, J. A. Matson, A. R. Malacko, H. Marquardt, Anti-
biotics 1992, 45, 1250; characterization and structure proposal: c) J. E.
Leet, D. R. Schroeder, S. J. Hofstead, J. Golik, K. L. Colson, S. Huang,
S. E. Klohr, T. W. Doyle, J. A. Matson, J. Am. Chem. Soc. 1992, 114,
7946; d) J. E. Leet, D. R. Schroeder, D. R. Langley, K. L. Colson, S.
Huang, S. E. Klohr, M. S. Lee, J. Golik, S. J. Hofstead, T. W. Doyle,
J. A. Matson, J. Am. Chem. Soc. 1993, 115, 8432; structure revision:
e) S. Kawata, S. Ashizawa, M. Hirama, J. Am. Chem. Soc. 1997, 119,
12012.
[2] Neocarzinostatin:The Past, Present, and Future of an Anticancer Drug
(Eds.: H. Maeda, K. Edo, N. Ishida), Springer, Tokyo, 1997.
[3] K.-i. Yoshida, Y. Minami, R. Azuma, M. Saeki, T. Otani, Tetrahedron
Lett. 1993, 34, 2637.
[4] H. Chimura, M. Ishizuka, M. Hamada, S. Hori, K. Kimura, J. Inawaga,
T. Takeuchi, H. Umezawa, J. Antibiot. 1968, 21, 44.
[5] A. S. Khokhlov, B. Z. Cherches, P. D. Reshetov, G. M. Smirnova, I. B.
Sorokina, T. A. Prokoptzeva, T. A. Koloditskaya, V. V. Smirnov, J.
Antibiot. 1969, 22, 541.
[6] D. R. Schroeder, K. L. Colson, S. E. Klohr, N. Zein, D. R. Langley,
M. S. Lee, J. A. Matson, T. W. Doyle, J. Am. Chem. Soc. 1994, 116,
9351.
[7] Review: K. C. Nicolaou, W.-M. Dai, Angew. Chem. 1991, 103, 1453;
Angew. Chem. Int. Ed. Engl. 1991, 30, 1387.
[8] a) S. Kawata, F. Yoshimura, J. Irie, H. Ehara, M. Hirama, Synlett 1997,
250; b) F. Yoshimura, S. Kawata, M. Hirama, Tetrahedron Lett. 1999,
40, 8281; c) M. J. Lear, F. Yoshimura, M. Hirama, Angew. Chem. 2001,
113, 97 2;Angew. Chem. Int. Ed. 2001, 40, 946.
[18] B. M. Kim, K. B. Sharpless, Tetrahedron Lett. 1989, 30, 655.
[19] E. J. Corey, A. Venkateswarlu, J. Am. Chem. Soc. 1972, 94, 6190.
[20] Examples of stereoselective Pd-mediated coupling reactions of 1,1-
dibromoolefins, Z-selective: a) J. M. Nuss, R. A. Rennels, B. H.
Levine, J. Am. Chem. Soc. 1993, 115, 6991; b) S. Torii, H. Okumoto,
T. Tadokoro, A. Nishimura, M. A. Rashid, Tetrahedron Lett. 1993, 34,
2139; E-selective: c) W. R. Roush, K. J. Moriarty, B. B. Brown,
Tetrahedron Lett. 1990, 31, 6509; d) W. Shen, L. Wang, J. Org. Chem.
1999, 64, 8873; e) J. Uenishi, K. Matsui, Tetrahedron Lett. 2001, 42,
4353; see also: ref. [9]; for E-selective couplings of 1,1-dichloroolefins
with organozinc and organomagnesium reagents, see: f) A. Minato, K.
Suzuki, K. Tamao, J. Am. Chem. Soc. 1987, 109, 1257.
[21] J. Inanaga, K. Hirata, H. Saeki, T. Katsuki, M. Yamaguchi, Bull.
Chem. Soc. Jpn. 1979, 52, 1989. Our conditions differ from those
previously reported in that the macrolactonization is carried out by
slow addition of a solution of the hydroxy acid to a suspension of
trichlorobenzoyl chloride, triethylamine, and DMAP in toluene at
508C.
[22] K. Jones, M. F. Lappert, J. Chem. Soc. 1965, 1944.
[23] a) J. K. Stille, Angew. Chem. 1986, 98, 504; Angew. Chem. Int. Ed.
Engl. 1986, 25, 508; b) J. K. Stille, J. H. Simpson, J. Am. Chem. Soc.
1987, 109, 2138; c) V. Farina, Pure Appl. Chem. 1996, 68, 7 3.
[24] This compound was observed to be a mixture of two equilibrating
conformational isomers (atropisomers) in solution. A detailed anal-
ysis of the thermodynamics and kinetics of this process is presented
elsewhere.[25]
[25] A. G. Myers, A. R. Hurd, P. C. Hogan, J. Am. Chem. Soc., submitted.
[26] G. Eglinton, A. R. Galbraith, J. Chem. Soc. 1959, 889.
[27] Compounds 2, 3, 4, and 22 were not stable in neat form, but could be
stored as dilute solutions in toluene at À208C. Yields were obtained
using an internal standard (1H NMR).
[9] A. G. Myers, S. D. Goldberg, Angew. Chem. 2000, 112, 2844; Angew.
Chem. Int. Ed. 2000, 39, 2732.
[10] Reactive (Z)-enediyne synthesis by dehydration: a) R. R. Jones, R. G.
Bergman, J. Am. Chem. Soc. 1972, 94, 660; b) T. Doi, T. Takahashi, J.
Org. Chem. 1991, 56, 3465; c) K. Iida, M. Hirama, J. Am. Chem. Soc.
1995, 117, 8875; see also: ref. [9].
[11] a) C. Glaser, Ber. Dtsch. Chem. Ges. 1869, 2, 422; for a recent review,
see: b) P. Siemsen, R. C. Livingston, F. Diederich, Angew. Chem. 2000,
112, 2740; Angew. Chem. Int. Ed. 2000, 39, 2632.
[12] A. G. Myers, Y. Horiguchi, Tetrahedron Lett. 1997, 38, 4363.
[13] a) R. A. Singer, E. M. Carreira, W. Lee, J. Am. Chem. Soc 1994, 116,
8837; review: b) E. M. Carreira, R. A. Singer, Drug Discovery Today
1996, 1, 145.
[28] This compound exists as a single atropisomeric structure in solution. A
detailed analysis of the thermodynamics and kinetics of this process is
presented elsewhere.[25]
[29] J. C. Martin, R. J. Arhart, J. Am. Chem. Soc. 1971, 93, 4327.
[30] a) W. Zhang, J. L. Loebach, S. R. Wilson, E. N. Jacbosen, J. Am.
Chem. Soc. 1990, 112, 2801; review: b) E. N. Jacobsen in Catalytic
Asymmetric Synthesis (Ed.: I. Ojima), VCH, New York, 1993,
chap. 4.2.
[31] Y. Hoshino, H. Yamamoto, J. Am. Chem. Soc. 2000, 122, 10452.
[32] S. Kanemoto, T. Nonaka, K. Oshima, K. Utimoto, H. Nozaki,
Tetrahedron Lett. 1986, 27, 3387.
[14] D. G. Wishka, D. R. Graber, E. P. Seest, L. A. Dolak, F. Han, W. Watt,
J. Morris, J. Org. Chem. 1998, 63, 7851.
[33] Prepared by the method of D. E. Bissing, C. A. Matuszak, W. E.
McEwen, J. Am. Chem. Soc. 1964, 86, 3824.
[15] For reviews, see: a) O. Mitsunobu, Synthesis 1981, 1; b) D. L. Hughes,
Org. React. 1992, 42, 335.
[34] Prepared by the method of W. H. Richardson, V. F. Hodge, J. Org.
Chem. 1970, 35, 4012.
[16] For a convenient in situ protocol, see: E. Fabiano, B. T. Golding,
M. M. Sadeghi, Synthesis 1987, 190.
[35] Use of cumene hydroperoxide led to a mixture of 22 and 24 in which
the undesired epoxide 22 is favored.
[17] K. Iida, M. Hirama, J. Am. Chem. Soc. 1994, 116, 10310.
Angew. Chem. Int. Ed. 2002, 41, No. 6
¹ WILEY-VCH Verlag GmbH, 69451 Weinheim, Germany, 2002
1433-7851/02/4106-1067 $ 17.50+.50/0
1067