The total synthesis of coleophomones B and C

Article abstract of DOI:10.1002/1521-3773(20020902)41:17<3276::AID-ANIE3276>3.0.CO;2-P

Pushing the frontiers of olefin metathesis: As the coleophomones B (1) and C (2) differ only in the configuration of the Δ^16,17 double bond, ring-closing metathesis was chosen as the method for their construction following an initially converge

Full text of DOI:10.1002/1521-3773(20020902)41:17<3276::AID-ANIE3276>3.0.CO;2-P

COMMUNICATIONS
[13] A closed transition state has been assumed in accord with the
precedented reactivity patterns of allylboronic esters. For a review of
the chemistry of allylboronic esters, see W. R. Roush In Methods. Org.
Chem. (Houben-Weyl), 4th ed. 1952 , Vol. E21b, 1995, pp. 1410 1486.
[14] See Supporting Information for experimental details.
[15] Preliminary results indicate that the allenylation reaction is also
effective with aldehydes, as heating a solution of the cyclic vinyl-
boronic acid 14 and a-triisopropylsilyloxy acetaldehyde in toluene at
608C provides the homologated trisubstituted allene (ca. 50% yield,
d.r. ca. 5:1:0:0).
[16] For the allene 28, d.r. ꢁ 3:1:0:0; the stereochemistry of the major
diastereomer was assigned by analogy to 25; transition state A
(Scheme 3c) was used as a predictive model, with the aldehyde
substituent placed in an equatorial position about the bicyclic
transition structure.
[17] a) J. A. Tallarico, K. D. Depew, H. E. Pelish, N. J. Westwood, C. W.
Lindsley, M. D. Shair, S. L. Schreiber, M. A. Foley, J. Comb. Chem.
2001, 3, 312 318; b) H. E. Blackwell, L. Pÿrez, R. A. Stavenger, J. A.
Tallerico, E. Cope Etough, M. A. Foley, S. L. Schreiber, Chem. Biol.
2001, 8, 1167 1182; c) P. A. Clemons, A. N. Koehler, B. K. Wagner,
T. G. Sprigings, D. R. Spring, R. W. King, S. L. Schreiber, Chem. Biol.
2001, 8, 1183 1195.
Me
Me
O
O
O
Me
Me
O
HO
1: coleophomone A
24
Me
Me
O
22
Me
15
17
Me
Me
19
11
O
5
O
O
O
9
Me
13
Me
23
Me
7
25
3
O
1
O
O
H
H
O
O
3: coleophomone C
2: coleophomone B
antifungal activity,[1] inhibition of human heart chymase,[1,3]
and antibiotic properties.[2] In addition, their unique molec-
ular architectures are laden with unusual and challenging
features: The strained and rigid framework of the coleopho-
mones possesses a sensitive tricarbonyl moiety tethered into
an 11-membered macrocycle whose strain is derived from the
incorporation of a fused aryl ring and a highly unsaturated
bridging six-membered carbocycle with the point of attach-
ment being a quarternary carbon atom. Herein we report the
total synthesis of coleophomones B (2) and C (3)[4] by a route
that pushes the frontiers of the olefin metathesis reaction as a
means to construct challenging molecular complexity.
[18] P. Clemons, G. C. Micalizio, S. L. Schreiber, unpublished results.
The Total Synthesis of Coleophomones B
and C**
K. C. Nicolaou,* Georgios Vassilikogiannakis, and
Tamsyn Montagnon
Coleophomone A (1) is related to coleophomone B (2) by
an aldol reaction that affords, from the latter, the unique
spirocycle seen in the former. In view of the reported[2]
interconversion of 1 and 2, our approach to the coleophomone
family initially focused on the total synthesis of 2 and 3. These
differ only in the configuration of the double bond (D16,17) that
is E in 2 and Z in 3. From all the possible retrosynthetic
disconnections we therefore chose the one dissecting this
double bond by the olefin metathesis reaction (Scheme 1).
This disconnection, which led to the general precursor I,
promised an exciting synthetic adventure because of the
unprecedented nature of the challenge, made even more acute
by the requirement to control the geometrical outcome of the
projected ring-closing metathesis to deliver either isomer at
will and the possible occurrence of atropisomers within the
targeted frameworks. Our second major retrosynthetic dis-
Dedicated to Professor Thomas J. Katz
on the occasion of his 65th birthday
In 1998, a patent[1] from a group at Shionogi disclosed the
structures and biological activity of three novel natural
products, designated I-1 (1), I-2 (2), and I-3 (3), which had
been isolated from a broth produced by the fungus Stachy-
bothys cylindrospora RF-5900. Two of these compounds later
re-emerged in the literature when a group at Merck discov-
ered them in extracts from the fermentation of Coleophoma
sp. fungi and named them coleophomones A (1) and B (2).[2]
The biological profile of the coleophomone family includes
[*] Prof. Dr. K. C. Nicolaou, Dr. G. Vassilikogiannakis, Dr. T. Montagnon
Department of Chemistry and The Skaggs Institute for Chemical
Biology
ꢀ
connection operated on precursor I at the C8 C9 bond to
The Scripps Research Institute
generate the key building blocks II and III. This left the
considerable hurdle of their union to be negotiated in the
synthetic direction.
10550 North Torrey Pines Road, La Jolla, CA 92037 (USA)
Fax : (þ 1)858-784-2469
and
Scheme 2 summarizes the construction of the requisite
aromatic building block 7. The preference for the cyano group
in 7 was based upon literature precedent regarding our
subsequent intentions to effect preferential C-acylation over
the usually more facile O-acylation in the coupling step,[5]
while the p-bromobenzoate group was chosen as a protecting
group to enhance the chances for crystalline derivatives to be
obtained for X-ray crystallographic analysis purposes. Thus,
beginning with commercially available 2,3-dimethyl phenol
(4) and following a five-step literature procedure,[6] we arrived
at hydroxy acetonide 5, whose conversion to benzaldehyde
Department of Chemistry and Biochemistry
University of California, San Diego
9500 Gilman Drive, La Jolla, CA 92093
[**] We thank Dr. T. Tsuri (Shionogi) for generous gifts of coleophomo-
nes A, B, and C and spectral data, and Dr. K. Wilson (Merck) for
sharing relevant information and spectral data with us. We thank Dr.
R. Chadha, Dr. D. H. Huang, and Dr. G. Siuzdak for X-ray
crystallography, NMR spectroscopic, and mass spectrometric assis-
tance, respectively. This work was financially supported by the
National Institutes of Health (USA), the Skaggs Institute for
Chemical Biology, and grants from Abbott Laboratories, ArrayBio-
pharma, Bayer, Boehringer Ingelheim, DuPont, Glaxo, Hoffmann-
LaRoche, Merck, Novartis, Pfizer, and Schering Plough.
3276
¹ 2002 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0044-8249/02/4117-3276 $ 20.00+.50/0
Angew. Chem. Int. Ed. 2002, 41, No. 17

COMMUNICATIONS
olefin metathesis
O
O
olefin metathesis
Me
O
Me
O
R
a) MeOH, conc. H2SO4
Me
Me
b) LiHMDS, RBr
O
MeO
d) 1 M HCl
O
Me
Me
O
O
O
Me
Me
Me
Me
9a, b
8
O
R
c) LDA,
HMPA,
O
O
O
H
H
prenyl-Br
O
O
O
Me
2: coleophomone B
3: coleophomone C
11a, b
O
R
R
Me
Me
Me
Me
R2
Me
O
R2
d) 1 M HCl
acylation
O
O
MeO
O
Me
Me
10a, b
R1
12a, b
(allyl)
Me
a: R =
b: R =
Me
O
Me (prenyl)
OR
I
R2
Me
Scheme 3. Synthesis of substituted 1,3-cyclohexadiones 11a, 11b, 12a, and
12b. a) conc. H2SO4 (cat.), MeOH, 658C, 12 h, 85%; b) LiHMDS
(1.05 equiv), THF, ꢀ788C, 1 h; then allyl- or prenyl-Br (1.1 equiv),
ꢀ78!08C, 3 h, 80 85%; c) LDA (1.1 equiv), THF, slow addition of a
solution of 11a or 11b in THF/HMPA (7/1), ꢀ788C, 1 h; then prenyl-Br
(2.0 equiv), ꢀ78!208C, 12 h, 89%; d) 1m HCl/THF (10/1), 258C, 14 h, 95
98%. LiHMDS ¼ lithium bis(trimethylsilyl)amide; LDA ¼ lithium diiso-
propylamide.
R2
alkylation
alkylation
O
O
O
R1
X
Me
O
OR
II
III
Scheme 1. Retrosynthesis of coleophomones B (2) and C (3).
alkylation of the LiHMDS-derived enolate with either allyl or
prenyl bromide afforded the anti alkylated products 9a and
9b, respectively (80 85% yield). The bisalkylated congeners
(10a and 10b) were prepared by a second alkylation that
proved more demanding than the first. These compounds
were obtained in good yield (89%) by using HMPA as an
additive and LDA as the base to generate the corresponding
enolate (ꢀ788C) followed by addition of the allylic bromide
(ꢀ788C) and slow warming of the reaction mixture to ambient
temperature. Finally, deprotection of the resulting vinylogous
methyl esters (9a, 9b, 10a, and 10b) by aqueous acid
hydrolysis furnished the coveted 1,3-cyclohexadiones (11a,
11b, 12a, and 12b) in 95 98% yield.
With all of the requisite building blocks in hand, the task
now confronting us was the development of appropriate
conditions to achieve the desired C-acylation reaction. Whilst
smooth coupling took place (see Scheme 4) when the
unsubstituted and monosubstituted 1,3-cyclohexadiones 8,
11a, and 11b were combined with acyl cyanide 7 in the
presence of Et3N at ambient temperature to furnish 13a
Me Me
OH
O
O
Me
Me
ref. [6]
OH
4
5
a) Et3N, 4-DMAP
p-BrC6H4(CO)Cl
b) p-TsOH•H2O
c) MnO2
Me
H
O
O
O
O
Me
d) K2CO3,Br
CN
e) Et2AlCN
f) PCC
p-BrC6H4(CO)O
p-BrC6H4(CO)O
7
6
Scheme 2. Construction of acyl cyanide 7. a) Et3N (1.5 equiv), p-
BrC6H4(CO)Cl (1.05 equiv), 4-DMAP (0.05 equiv), CH2Cl2, 258C,
20 min, 97%; b) p-TsOH¥H2O (0.3 equiv), THF/H2O (3/2), 758C, 10 h,
85%; c) MnO2 (10 equiv), EtOAc, 258C, 1.5 h, 87%; d) K2CO3 (2.0 equiv),
3-bromo-2-methylpropene (1.5 equiv), acetone, 568C, 2 h, 91%;
e) Et2AlCN (1.1 equiv), toluene, 0!258C, 1 h, 82%; f) PCC (4.0 equiv),
CH2Cl2, 408C, 12 h, 73%. 4-DMAP ¼ 4-dimethylaminopyridine; p-
TsOH ¼ p-toluenesulfonic acid; PCC ¼ pyridinium chlorochromate.
Me
Me
O
R1
R2
Me
O
O
O
O
O
derivative 6 entailed sequential p-bromobenzoylation, aceto-
nide removal, and MnO2 oxidation (72% yield over the three
steps). Aldehyde 6 was then smoothly alkylated with 3-
bromo-2-methylpropene in the presence of K2CO3 (91%
yield) and subsequently converted to the desired acyl cyanide
7 by a two-step sequence involving initial addition of Nagata©s
reagent followed by PCC oxidation of the resulting cyanohy-
drin, events which proceeded in 60% overall yield.
R1
R2
a) Et3N or
CN
+
b) Et3N,
4-DMAP
O
O
Me
p-BrC6H4(CO)O
p-BrC6H4(CO)O
7
8 : R1 = R2 = H
13a
13b
13c
13d
13e
11a: R1 = H, R2 = allyl
11b: R1 = H, R2 = prenyl
12a: R1 = allyl, R2 = prenyl
12b: R1 = R2 = prenyl
The desired 1,3-cyclohexadiones (11a, 11b, 12a, and 12b)
were synthesized starting with CS-symmetrical 5-methyl-1,3-
cyclohexadione (8) and following the divergent pathway
shown in Scheme 3. Thus, methylation of 8 followed by
Scheme 4. Coupling of acyl cyanide 7 with various 1,3-cyclohexadiones.
a) 8, 11a, or 11b (1.2 equiv), Et3N (2.0 equiv), THF, 258C, 6 12 h, 91 98%;
b) 12a, 12b (1.2 equiv), Et3N (2.0 equiv), 4-DMAP (1.0 equiv), THF, 258C,
72 96 h, 83 86%.
Angew. Chem. Int. Ed. 2002, 41, No. 17
¹ 2002 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0044-8249/02/4117-3277 $ 20.00+.50/0
3277

COMMUNICATIONS
(98%), 13b (94%), and 13c (91%), respectively,[7] the
disubstituted substrates 12a and 12b proved recalcitrant,
refusing to react even after prolonged reaction times under
similar conditions. However, upon modification of the above
protocol to include 4-DMAP, the coupled products 13d and
13e could be isolated after 72 96 h, in 83 86% yield
(Scheme 4).
The scene was now set for the exploration to begin into the
feasibility of the crucial olefin-metathesis-based step to forge
the challenging 11-membered ring;[8] we started with the
simplest substrate, compound 13b. While the early Grubbs©
ruthenium catalyst[9] failed to induce ring closure within this
substrate, our ambition in this regard was realized when
diolefin 13b was subjected to the newest Grubbs©[10] catalyst,
IV (20 mol%, Scheme 5) in dichloromethane at reflux,
conditions that afforded a single macrocyclic product (14,
Scheme 5A) in 60% yield. X-Ray crystallographic analysis[11]
of 14 (see ORTEP representation) revealed its complete
structure, including the Z configuration of the D16,17 double
bond and the arrangement of the enol within the tricarbonyl
subunit. Having established the viability of macrocycle
formation by olefin metathesis, we then proceeded to inves-
tigate the simplest bisalkylated substrate, 13d (Scheme 5B),
as a substrate for this ring-closing reaction. As we had feared,
when this compound was treated with catalyst IV (10 mol%),
it was rapidly converted to spiropentene 15 in 85% yield as
shown in Scheme 5B. The formation of 15 provoked us to
increase the substitution on the terminal olefin to the gem-
dimethyl level to relocate the metathesis initiation site to the
less substituted olefin appended to the aromatic portion of the
molecule. We, therefore, subjected substrate 13c (Sche-
me 5A) to the same catalyst (IV, 30 mol%) and remarkably
once again, observed the formation of macrocycle 14, albeit in
reduced yield (30%).
While these results served as a proof of principle for our
initial premise that ring-closing metathesis would be a power-
ful tool to forge the challenging macrocyle of the coleopho-
mones, the tricarbonyl substrates employed left much to be
desired in terms of stability and ease of manipulation. To
improve upon these two aspects of the sequence we moved to
protect the troublesome tricarbonyl moiety, only to discover
that the new substrates harbored additional advantages. Thus,
exposure of 13a to excess diazomethane in diethyl ether
resulted in the formation of the vinylogous methyl esters 16a,
which precipitated out from the reaction mixture in 48%
yield, and 16b, which could be isolated from the remaining
solution in 48% yield (Scheme 6). The latter compound (16b)
was alkylated with prenyl bromide, after enolate generation
with LiHMDS and HMPA, to afford alkylated product 17
(63%) as a 1:1 mixture of geometrical isomers (D8,9), each of
which existed as a 1:1 pair of atropisomers as revealed by 1H
Me
H
O
O
O
Me
Me
O
p-BrC6H4(CO)O
13a
a) CH2N2
Me
O
OMe
O
OMe
O
O
7
8
Me
Me
O
O
p-BrC6H4(CO)O
p-BrC6H4(CO)O
16a
16b
b) LiHMDS,
HMPA,
prenyl-Br
Me
16
Hb
Ha
H
Me
Me
Ha
17
Me
Hb
O
O
9
H
O
OMe
O
Me
c) cat. IV
8
Me
O
OMe
O
p-BrC6H4(CO)O
p-BrC6H4(CO)O
18 (exclusively E)
17
Scheme 5. A) Forging a Z-configured 11-membered ring by an olefin
metathesis reaction involving disubstituted and trisubstituted double
Scheme 6. Protection of the tricarbonyl moiety and reversal of the
geometrical outcome of the olefin metathesis reaction leading to an (E)
D16,17 containing 11-membered ring. a) Excess CH2N2, Et2O, 0 8C, 30 min,
48% of 16a plus 48% of 16b; b) LiHMDS (1.05 equiv), THF, ꢀ788C, 1 h;
then prenyl-Br (2.0 equiv), ꢀ78!08C, 5 h, 63%; c) cat. IV (0.2 equiv),
CH2Cl2, 408C, 20 h, 30% of 18 plus 35% recovered starting material 17.
bonds. B) Selective formation of
a spiropentene from triolefin 13d.
a) cat. IV (0.2 equiv), CH2Cl2, 408C, 5 h, 60% for 13b; cat. IV (0.3 equiv),
CH2Cl2, 408C, 18 h, 30% for 13c; b) cat. IV (0.1 equiv), CH2Cl2, 408C, 1 h,
85%. Cy ¼ cyclohexyl; Ph ¼ phenyl.
3278
¹ 2002 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0044-8249/02/4117-3278 $ 20.00+.50/0
Angew. Chem. Int. Ed. 2002, 41, No. 17

COMMUNICATIONS
NMR spectroscopy. When 17 was subjected to the action of
Grubbs© catalyst IV (20 mol%) in refluxing dichloromethane
for 20 h, a single macrocycle (18) was isolated, in 30% yield,
and proven to be of the E configuration at D8,9. Under these
conditions, a considerable amount (30%) of the starting
material (17) was recovered and, interestingly, found to be
enriched in the Z isomer (D8,9). The most tantalizing feature
elucidated at this juncture was, however, the fact that the
macrocyclic double bond (D16,17) bore exclusively the E con-
figuration (see NOE effects on structure 18, Scheme 6). We
had now unearthed pathways that led us to both diaster-
eoisomers of the coleophomone macrocyclic skeleton in a
specific manner.
With these results providing the foundation for the next
phase, we proceeded to exploit the full potential of our
intelligence gathering for the final solution of the coleopho-
mone problem. Thus, fully substituted precursor 13e (see
Scheme 7) was treated with diazomethane under the estab-
lished conditions and, following flash column chromatogra-
phy, methoxy derivatives 19a (32%) and 19b (48%) were
isolated along with a small quantity of the third possible
regioisomer (not shown). As expected, 19a was a single
compound, while 19b proved to be a mixture of geometrical
isomers around the D8,9 bond (ca. 1.3:1), each of which again
consisted of a pair of atropisomers (ca. 1:1 by 1H NMR
spectroscopy). Upon exposure of the two regioisomers 19a
and 19b, independently, to the catalyst IV (10 mol%) we
observed the most pleasing results. Both 19a and 19b
succumbed to regio- and stereospecific ring-closing meta-
thesis to furnish, in high yields, the corresponding macrocycles
20a (80%, exclusively Z-configured at D16,17, ca. 4:1 ratio of
atropisomers in CDCl3, single atropisomer in CD3CN) and
20b (86%, exclusively E-configured at D16,17, ca. 1:1 E:Z isom-
ers at D8,9). The structures of 20a and 20b were established on
the basis of nOe studies and, in the former case (20a),
confirmed by an X-ray crystallographic analysis[11] (see
ORTEP representation). Thus, with exquisite specificity, the
olefin metathesis reaction furnishes the two macrocycles
processing the opposite configuration at the D16,17 bond, each
corresponding to a different coleophomone framework.
Furthermore, and in contrast to the reaction of the mono-
prenylated compound 17 (see Scheme 5), both D8,9 diaster-
eoisomers of 19b reacted smoothly, presumably because the
second prenyl group forced the molecule to adopt a more
favorable and rigid conformation. The rigidity and, therefore,
reduced entropy[12] of 19a and 19b is also presumed to be
responsible for the apparent irreversibility of the ring-closing
metathesis as evidenced by the absence of any spirocyclo-
pentene products in the reaction mixture and the high yields
of the desired macrocycles (20a and 20b). Another fascinat-
ing feature of these ring-closing metatheses is their exquisite
diastereoselectivity, that is, only the prenyl group occupying
the cis position to the adjacent methyl group (C12) partic-
ipates in the reaction. This preference presumably results
from the fact that it places the remaining prenyl side chain
trans to the C12 methyl group, an arrangement that allows
both groups to occupy the lower-energy equatorial positions
on the cyclohexane ring. Finally, we were able to confirm our
earlier hypothesis that the unprotected tricarbonyl system was
detrimental to the olefin metathesis reaction by comparing
the poor performance (< 20% yield of cyclized product) of
13e (Scheme 7) under the same cyclization conditions with
the stellar showing of its protected derivatives 19a and 19b.
As a prelude to the final drive toward the coleophomones,
the introduction of the D11,12 bond was accomplished in both
series by a one-pot phenylselenide formation/oxidation/syn-
elimination sequence to afford 21a as a single compound and
21b as a mixture of geometrical isomers (61 and 53% overall
yield, respectively, Scheme 8). This transformation increased
the ratio of geometrical isomers (D8,9) in 21b in favor of the
E isomer (20b E:Z ca. 1:1!21b E:Z ca. 3:1), presumably
owing to the apparent steric encumbrance of the C11
methylene group caused by the direct positioning of the C24
methyl group above it in the Z isomer (molecular modeling
and nOe studies). The recovered starting material 20b (15%)
was enriched in the Z isomer. Exposure of 21a and 21b to
K2CO3 in methanol at ambient temperature led to global
deprotection and afforded 22a (90% yield) and 22b (96%
yield), respectively, both as single compounds (see Scheme 8
and Table 1 for spectroscopic data). At this juncture, the
Scheme 7. Stereospecific olefin metathesis reactions leading to (Z) and
E isomers of 11-membered rings en route to the coleophomones. a) Excess
CH2N2, Et2O, 0 8C, 1 h, 32% of 19a plus 48% of 19b; b) cat. IV (0.1 equiv),
CH2Cl2, 408C, 3 h, 80%; c) cat. IV (0.1 equiv), CH2Cl2, 408C, 3 h, 86%.
Angew. Chem. Int. Ed. 2002, 41, No. 17
¹ 2002 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0044-8249/02/4117-3279 $ 20.00+.50/0
3279

COMMUNICATIONS
Table 1. Selected data for compounds 22a and 22b.
Me
24
Me
Me
Me
Me
22a: yellow oil; Rf ¼ 0.44 (silica gel, hexane/EtOAc, 1/1); IR (film): n˜max
¼
3436, 2920, 1651, 1549, 1449, 1261, 1008, 914, 726 cmꢀ1; 1H NMR (600 MHz,
CDCl3): d ¼ 15.71 (s, 1H), 7.38 (t, J ¼ 7.7 Hz, 1H), 7.11 (d, J ¼ 7.7 Hz, 1H),
6.89 (d, J ¼ 7.7 Hz, 1H), 6.29 (s, 1H), 5.28 (dd, J ¼ 11.0, 6.6 Hz, 1H), 4.56 (d,
J ¼ 12.1 Hz, 1H), 4.51 (d, J ¼ 12.1 Hz, 1H), 4.50 (br dd, J ¼ 6.5, 5.5 Hz, 1H),
4.34 (d, J ¼ 8.7 Hz, 1H), 4.10 (d, J ¼ 8.7 Hz, 1H), 2.84 (dd, J ¼ 14.8, 6.0 Hz,
1H), 2.69 (dd, J ¼ 14.3, 11.0 Hz, 1H), 2.22 (m, 2H), 2.08 (s, 3H), 1.82 (s,
3H), 1.55 (s, 3H), 1.50 ppm (s, 3H); 13C NMR (150 MHz, CDCl3): d ¼
198.8, 181.0, 164.1, 156.4, 139.8, 136.0, 134.4, 131.9, 131.7, 131.4, 124.1, 123.6,
122.2, 117.8, 115.1, 113.1, 69.7, 64.2, 59.2, 39.0, 37.1, 25.7, 23.2, 20.1, 18.1 ppm;
HRMS (MALDI): calcd for C25H27O4 [MHꢀH2O]: 391.1904; found:
391.1905; MS (ESI): 409 [MHþ], 431 [MNaþ]
Me
O
O
O
O
O
Me
Me
11
OMe
MeO
O
p-BrC6H4(CO)O
p-BrC6H4(CO)O
20a
20b
b) LiHMDS,
a) LiHMDS,
PhSeCl; then
aq. H2O2
PhSeCl; then
aq. H2O2
Me
Me
O
16
16
8
Me
17
Me
17
Me
Me
O
O
O
22b: amorphous solid; Rf ¼ 0.32 (silica gel, hexane/EtOAc, 1/1); IR (film):
n˜max ¼ 3412, 2920, 1651, 1556, 1455, 1262, 997, 866, 736 cmꢀ1
;
1H NMR
13
13
Me
Me
8
9
9
12
11
12
11
(500 MHz, CDCl3): d ¼ 16.54 (s, 1H), 7.29 (t, J ¼ 8.2 Hz, 1H), 7.12 (d, J ¼
8.2 Hz, 1H), 7.09 (d, J ¼ 7.4 Hz, 1H), 6.38 (s, 1H), 5.41 (br d, J ¼ 12.0 Hz,
1H), 4.66 (d, J ¼ 12.3 Hz, 1H), 4.56 (d, J ¼ 11.4 Hz, 1H), 4.55 (br dd, J ¼ 7.4,
5.5 Hz, 1H), 4.49 (d, J ¼ 12.3 Hz, 1H), 4.31 (d, J ¼ 11.4 Hz, 1H), 2.99 (dd,
J ¼ 14.2, 6.8 Hz, 1H), 2.44 (t, J ¼ 12.8 Hz, 1H), 2.33 (br d, J ¼ 13.8 Hz, 1H),
2.25 (dd, J ¼ 14.6, 6.4 Hz, 1H), 2.11 (s, 3H), 1.58 (br s, 3H), 1.57 (br s, 3H),
1.33 ppm (s, 3H); 13C NMR (125 MHz, CDCl3): d ¼ 200.0, 198.6, 183.0,
164.9, 153.5, 139.6, 135.7, 134.9, 132.2, 129.6, 128.1, 123.3, 122.9, 118.3, 118.0,
114.2, 76.3, 63.7, 58.8, 38.3, 34.4, 25.7, 20.1, 19.7, 18.0 ppm; HRMS
(MALDI): calcd for C25H27O4 [MHꢀH2O]: 391.1904; found: 391.1898;
MS (ESI): 409 [MHþ], 431 [MNaþ]
OMe
O
MeO
O
p-BrC6H4(CO)O
p-BrC6H4(CO)O
21a
21b
c) K2CO3
d) K2CO3
Me
Me
Hb
Ha
Hb
Ha
Me
Me
Me
Me
H
Ha
Ha
Hb
H
H
Hb
O
O
O
O
H
O
Me
Me
O
O
O
H
H
HO
HO
22a
e) CrO3•2py
22b
f) MnO2
strained and synthetically challenging macrocyclic scaffolds
associated with these molecules via an olefin metathesis
reaction. This unprecedented accomplishment bodes well for
this venerable reaction that is becoming increasingly enabling
for organic synthesis through the discovery of new catalysts to
initiate it and the design of novel strategies to apply it.
3: coleophomone C
2: coleophomone B
Scheme 8. Completion of the total synthesis of coleophomones B (2) and
C (3). a) LiHMDS (1.3 equiv), THF, ꢀ78!ꢀ108C, 1 h; then PhSeCl
(1.4 equiv), ꢀ788C!208C, 30 min; then sat. NH4Cl, (excess), 31% aq.
H2O2 (excess), 258C, 1 h, 61%. b) LiHMDS (1.3 equiv), THF, ꢀ78!258C,
1 h; then PhSeCl (1.4 equiv), ꢀ60!258C, 30 min; then sat. NH4Cl,
(excess), 31% aq. H2O2 (excess), 258C, 1 h, 53%, plus 15% recovered
20b; c) K2CO3 (3.0 equiv), MeOH, 258C, 3 h; then H2O (excess), 24 h,
90%; d) K2CO3 (3.0 equiv), MeOH, 258C, 40 min; then H2O (excess),
30 min, 96%; e) py (12 equiv), CrO3 (6.0 equiv), CH2Cl2, 0!258C, 20 min;
then 22a, 258C, 2 h, 81%; f) MnO2 (20 equiv), Et2O, 36 8C, 8 h, 73 %. py ¼
pyridine.
Received: June 14, 2002 [Z19535]
[1] T. Kamigaichi, M. Nakajima, H. Tami, , Japanese patent 10101666
1998.
[2] K. E. Wilson, N. N. Tsou, Z. Guan, C. L. Ruby, F. Pelaez, J.
Gorrochategui, F. Vicente, H. R. Onishi, Tetrahedron Lett. 2000, 41,
8705 8709.
[3] H. Urata, A. Kinoshita, K. S. Misono, F. M. Bumpus, A. Husain, J.
Biol. Chem. 1990, 265, 22348 22357.
coalescence to a single isomer in each series allowed addi-
tional and unequivocal confirmation of our earlier distinction
between Z/E isomerism and atropisomerism by using nOe
studies (see structures 22a and 22b, Scheme 8). Oxidation of
22b with MnO2 in diethyl ether produced, cleanly[13] and in
73% yield, coleophomone B (2) whose spectroscopic data
matched those of the natural product.[1,2] Likewise, coleopho-
mone C (3) was accessed by oxidation of 22a, this time with
Collins© reagent (this isomer reacts sluggishly with MnO2) in
81% yield and, once again, the spectroscopic data of the
synthetic sample proved to be identical to those of an
authentic sample.[1] Despite the reported equilibrium[2] be-
tween 1 and 2 we choose to refrain from declaring a formal
total synthesis of coleophomone A (1) pending further
experimentation.
[4] For nomenclature uniformity, we coined the name coleophomone C
for Shionogi©s I-3 compound,[1] despite the fact that it was not reported
by the Merck group in their coleophoma sp. study.[2]
[5] a) A. P. Krapcho, J. Diamanti, C. Cayen, R. Bingham, Org. Synth.
Coll. Vol. V. 1973, 198; initial O-acylation followed by rearrangement
to the product of C-acylation with various additives is well prece-
dented. For selected examples see: b) T. Ukita, S. Nojima, M.
Matsumoto, J. Am. Chem. Soc. 1950, 72, 5143 5144; c) Y. Tanabe,
M. Miyakado, N. Ohno, H. Yoshioka, Chem. Lett. 1982, 1543 1546;
d) Y. Oikawa, K. Sugano, O. Yonemitsu, J. Org. Chem. 1978, 43, 2087
2088; e) A. A. Akhrem, F. A. Lakhvich, S. I. Budai, T. S. Khlebnicova,
I. I. Petrusevich, Synthesis 1978, 925 927; f) J. E. Oliver, K. R. Wilzer,
R. M. Waters, Synthesis 1990, 1117 1119; g) H. Tabuchi, T. Hama-
moto, A. Ichihara, Synlett 1993, 651 652.
[6] M. Suzuki, T. Sugiyama, M. Watanabe, T. Murayama, K. Yamashita,
Agric. Biol. Chem. 1987, 51, 1121 1127.
[7] a) I. F. Montes, U. Burger, Tetrahedron Lett. 1996, 37, 1007 1010;
b) Q. Tang, S. E. Sen, Tetrahedron Lett. 1998, 39, 2249 2252.
[8] For examples of the synthesis of 11-membered rings by using olefin
metathesis see: a) H. El Sukkari, J.-P. Gesson, B. Renoux, Tetrahedron
Lett. 1998, 39, 4043 4046; b) J. D. Winkler, J. M. Holland, J. Kasparec,
P. H. Axelsen, Tetrahedron 1999, 55, 8199 8214; c) T. R. Hoye, M. A.
Promo, Tetrahedron Lett. 1999, 40, 1429 1432; d) M. Arisawa, C.
In conclusion we have developed an expedient and stereo-
controlled total synthesis of coleophomones B (2) and C (3)
through an initially convergent route that diverges at a late
stage to allow stereospecific construction of the highly
3280
¹ 2002 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0044-8249/02/4117-3280 $ 20.00+.50/0
Angew. Chem. Int. Ed. 2002, 41, No. 17

COMMUNICATIONS
Kato, H. Kaneko, A. Nishida, M. Nakagawa, J. Chem. Soc. Perkin
Trans. 1 2000, 1873 1876; for insightful discussion see: e) A. F¸rstner,
K. Radkowski, C. Wirtz, R. Goddard, C. W. Lehmann, R. Mynott, J.
Am. Chem. Soc. 2002, 124, 7061 7069.
[9] a) P. Schwab, M. B. France, J. W. Ziller, R. H. Grubbs, Angew. Chem.
1995, 107, 2179 2181; Angew. Chem. Int. Ed. Engl. 1995, 34, 2039
2041; b) P. Schwab, R. H. Grubbs, J. W. Ziller, J. Am. Chem. Soc. 1996,
118, 100 110.
[10] M. Scholl, S. Ding, C. W. Lee, R. H. Grubbs, Org. Lett. 1999, 1, 953
956.
Scheme 1. A!ABC or A!ABCD ring-formation approach to substituted
carbazoles by rhodium-catalyzed alkyne cyclotrimerization.
[11] CCDC-187042 and CCDC-187043 contain the supplementary crys-
tallographic data for this paper. These data can be obtained free of
charge via www.ccdc.cam.ac.uk/conts/retrieving.html (or from the
Cambridge Crystallographic Data Centre, 12, Union Road, Cam-
bridge CB21EZ, UK; fax: (þ 44)1223-336-033; or deposit@ccdc.ca-
m.ac.uk).
[12] For insightful discussion into the effects of entropy in the olefin
metathesis reaction and examples see: a) A. F¸rstner, K. Langemann,
J. Org. Chem. 1996, 61, 3942 3943; b) A. F¸rstner, K. Langemann,
Synthesis 1997, 792 803; c) A. F¸rstner, G. Seidel, N. Kindler,
Tetrahedron 1999, 55, 8215 8230; d) C. Galli, L. Mandolini, Eur. J.
Org. Chem. 2000, 3117 3125.
The diynes 2 were prepared in three steps starting from
readily available 2-iodoanilines 1. Sonogashira reactions
between 1 and terminal alkynes 5 followed by N-tosylation
of the aniline functionality gave alkynes 6, which thereafter
underwent N-ethynylation with alkynyliodonium salts 7 to
give the diynes 2 in good overall yields (Scheme 2, path A).
[13] It should be noted that no flash column chromatography was required
in the last two steps because compounds 22a, 22b, 1, and 2 were
isolated after a standard base/acid workup as spectroscopically pure
compounds.
R1
+ Ph
I
-
c) R2
R1
(R1 = H)
R
OTf
a) H
b)
5
7
R2: SiMe3, Ph, H
NHTs
Path A
Path B
6
2
1
(R1 = H)
R1
R
H
d)
A Highly Efficient and Flexible Synthesis of
Substituted Carbazoles by Rhodium-Catalyzed
Inter- and Intramolecular Alkyne
Cyclotrimerizations**
I
R2
b)
c)
5
+
I
R2: SiMe3
Ph
-
N
Me3Si
OTf
Ts
7a
8
Scheme 2. a) 5 (1.3 equiv), 5 mol% [PdCl2(PPh3)2], 10 mol% CuI, NEt3,
DMF, room temperature, 60 98%; b) TsCl, pyridine, THF, 70 92%;
c) potassium hexamethyldisilazane (KHMDS), toluene, 08C, then addition
of 7 (1.4 equiv), room temperature, 52 93% for 6!2, 87 98% for 8; d) 5
(1.4 equiv), 5 mol% [PdCl2(PPh3)2], 10 mol% CuI, 10 mol% PPh3, NEt3,
DMF, 808C, 54 68%.
Bernhard Witulski* and Carole Alayrac
Syntheses of substituted carbazoles have attracted consid-
erable attention because carbazole alkaloids are a growing
class of natural products that exhibit a variety of biological
activities.[1,2] Moreover, annelated carbazoles structurally
related to the natural indolocarbazole staurosporine have
gained significance because of their protein kinase C (PKC)
and topoisomerase inhibitory activity.[3]
Although several syntheses of substituted carbazoles as
well as chemical modifications of the carbazole nucleus are
established,[1,4] the major challenge in carbazole chemistry
arises from the question of how to functionalize regioselec-
tively up to eight aromatic ring positions. We have designed a
strategy for the assembly of the carbazole nucleus by an
A!ABC or A!ABCD ring-formation approach that is
based on the reliability of the Sonogashira reaction,[5] our
previously reported synthesis of functionalized ynamides,[6]
and the efficiency of transition-metal-catalyzed alkyne cyclo-
trimerizations (Scheme 1).[7]
Alternatively, N-ethynylation reactions with 7a were carried
out with tosylated 1 to give the ynamides 8, which were then
converted to 2 through Sonogashira reactions (Scheme 2,
path B). While the conversion of 1 to 6 proceeded readily at
room temperature, cross-coupling reactions with 8 to give 2
required heating to 808C, probably because of an increase of
steric hindrance. Notably, the use of trimethylsilylacetylene in
the Sonogashira reaction, or the use of 7a for the N-
ethynylation reaction following either path A or B in
Scheme 2 allowed further manipulations of the diynes 2
through protective group strategies.
Crossed alkyne cyclotrimerizations between diynes 2 and
monoalkynes 3 promoted by Wilkinson©s catalyst[6b,8] pro-
ceeded under very mild conditions (3 5 mol% [RhCl(PPh3)3],
5 6 equivalents of 3, toluene as solvent) to give the carbazoles
4 in high yields (Scheme 1, Table 1). Notably, the additional
triple bond in 2e did not interfere with the formation of
carbazole 4e (80% yield, Table 1, entry 5). The outstanding
efficiency of the carbazole formation by crossed alkyne
cyclotrimerizations is presumably a consequence of confor-
mational restrictions present in the diynes 2. A conformation
in which both alkyne moieties are close to each other appears
likely and therefore favors the overall catalytic process.
[*] Priv.-Doz. Dr. B. Witulski, Dr. C. Alayrac
Fachbereich Chemie der Universit‰t Kaiserslautern
Erwin-Schrˆdinger-Strasse, 67663 Kaiserslautern (Germany)
Fax : (þ 49)631-205-3921
E-mail: witulski@rhrk.uni-kl.de
[**] This work was supported by the Deutsche Forschungsgemeinschaft
and the Fonds der Chemischen Industrie. C.A. is grateful to the
Alexander von Humboldt Foundation for a Research Fellowship.
Supporting information for this article is available on the WWW under
http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2002, 41, No. 17
¹ 2002 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
0044-8249/02/4117-3281 $ 20.00+.50/0
3281