A novel extension of the Stork-Jung vinylsilane Robinson annelation procedure for the introduction of the cyclohexene of guanacastepene

Article abstract of DOI:10.1016/S0040-4039(01)02026-3

A new annelation procedure has been developed in a model system to introduce the cyclohexene ring of guanacastepene. Cyclohexenone 16 is prepared by sequential methylation and alkylation with allylic iodide 15. One-pot epoxidation, protodesilylation and hydrolysis of the THP forms dione 19, which cyclizes with NaOMe to 24 in 57% overall yield. Reduction, Mitsunobu inversion, and selective oxidation of the primary alcohol forms model 29. Phenyldimethylsilyl allylic iodide 36 is easier to make and more stable than trimethylsilyl allylic iodide 6 used by Stork.

Full text of DOI:10.1016/S0040-4039(01)02026-3

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 42 (2001) 9123–9126
A novel extension of the Stork–Jung vinylsilane Robinson
annelation procedure for the introduction of the cyclohexene of
guanacastepene
Barry B. Snider* and Bo Shi
Department of Chemistry MS 015, Brandeis University, Waltham MA 02454-9110, USA
Received 10 October 2001; revised 24 October 2001; accepted 29 October 2001
Abstract—A new annelation procedure has been developed in a model system to introduce the cyclohexene ring of guanacaste-
pene. Cyclohexenone 16 is prepared by sequential methylation and alkylation with allylic iodide 15. One-pot epoxidation,
protodesilylation and hydrolysis of the THP forms dione 19, which cyclizes with NaOMe to 24 in 57% overall yield. Reduction,
Mitsunobu inversion, and selective oxidation of the primary alcohol forms model 29. Phenyldimethylsilyl allylic iodide 36 is easier
to make and more stable than trimethylsilyl allylic iodide 6 used by Stork. © 2001 Elsevier Science Ltd. All rights reserved.
The novel, diterpene antibiotic guanacastepene (1),
which was isolated by Clardy from an unidentified
fungus growing on the tree Daphnopsis americana,
shows excellent activity toward methicillin-resistant
Staphylococcus aureus and vancomycin-resistant Ente-
rococcus faecium.1 Further biological studies indicated
that 1 has moderate activity against Gram-positive
bacteria, poor activity against Gram-negative bacteria,
and hemolytic activity against human red blood cells.2
We recently reported a 12-step synthesis of 4 (Scheme
1), the functionalized hydroazulenone ring system of
guanacastepene.3 Danishefsky and Magnus have also
reported routes to functionalized hydroazulene ring
systems4a,5 and Danishefsky has described alkylation
studies of the hydroazulene core.4b
Elaboration of 4 to guanacastepene could be carried
out by introduction of a methyl group and an oxy-
genated side chain to give 3. An aldol reaction would
give 2, in which X is an aldehyde precursor. Functional
group modification in the five- and six-membered rings
would complete a synthesis of guanacastepene. At first
glance, the Robinson annelation seems well suited to
the preparation of 2 from 4. However, there are several
concerns about this approach. First, a Michael addition
to form 3 may not give the correct stereoisomer. Sec-
ond, Robinson annelations on cycloalkenones,
although known, are uncommon.6 Third, the Nazarov
reagent, CH2ꢀCHCOCH2CO2R, which would introduce
X=CO2R that can be converted to an aldehyde, can
only be used with 1,3-dicarbonyl compounds.
OHC
X
OH
O
We decided to examine the Stork–Jung Robinson
annelation procedure,7 alkylation with 1-iodo-3-
trimethylsilyl-2-butene (6) to introduce a 3-oxobutyl
group, since either the methyl or trimethylsilylbutenyl
group can be introduced first, increasing the likelihood
that the stereochemistry of the new introduced center in
3 can be controlled (Scheme 2). 4,4-Dimethyl-2-cyclo-
hexenone (5) was chosen as a readily available model
for hydroazulenone 4. Preparation of the lithium eno-
late of 5 with LDA and DMPU in THF at −78°C,
addition of 6, and slow warming to 25°C affords 78%
of 7. A second alkylation with LDA, DMPU and MeI
provides 84% of bis-alkylated cyclohexenone 9. Alkyla-
tion in the opposite order proceeds equally well giving
81% of 88 in the first step and 82% of 9 in the second.
Reaction of 9 with m-CPBA at 0°C forms the silyl
O
RO
AcO
2
guanacastepene (1)
X
RO
RO
O
O
O
3
4
Scheme 1. Retrosynthetic analysis for guanacastepene (1).
* Corresponding author. Tel.: 781-736-2550; fax: 781-736-1516;
e-mail: snider@brandeis.edu
0040-4039/01/$ - see front matter © 2001 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(01)02026-3

9124
B. B. Snider, B. Shi / Tetrahedron Letters 42 (2001) 9123–9126
OTHP
4.4 eq. PhMe2SiLi
2.2 eq. CuCN,THF
OTHP
OTHP
HO
I
CH3
O
O
HO
6
SiMe3
SiMe3
13
SiMe2Ph
14 (76%)
5
7
1) MsCl, Et3N, 0 °C, 15 min
2) NaI, acetone, 45 oC, 15 min
LDA, DMPU, 78%
I
CH3I, LDA,
DMPU, 84%
CH3I, LDA,
DMPU, 81%
SiMe2Ph
15 (82%)
I
CH3
O
O
6
SiMe3
SiMe3
Scheme 3. Preparation of silyl allylic iodide 15.
LDA, DMPU, 82%
O
9
8
m-CPBA
25 °C, 3 h
Alkylation of 4,4,6-trimethyl-2-cyclohexenone (8) with
15 provides 90% of 16 (Scheme 4). Alkylation of 5 with
15 proceeds in 75% yield and methylation of that
product affords 87% of 16, indicating that the alkyla-
tions can be carried out in either order, improving the
likelihood of obtaining the desired stereoisomer from 4.
Epoxidation of 16 with m-CPBA in CH2Cl2 for 1 h at
0°C affords silyl epoxide 17, which is not cleaved by
m-chlorobenzoic acid at room temperature, indicating
that dimethylphenylsilyl epoxide 17 is more stable to
acid than the trimethylsilylepoxide formed from 9.
Addition of pyr·(HF)x to the reaction mixture and
stirring at 25°C for 10 min efficiently cleaves silyl
epoxide 17 to the ketone and partially hydrolyzes the
THP providing a mixture of 18 and 19. Addition of
TsOH and MeOH and stirring at 25°C for 30 min
completes the hydrolysis of the THP affording 19 in
79% yield from 16 in a one-pot reaction.
5 eq NaOMe
MeOH, 2 h
O
O
+
O
O
10 (86%)
11 (76%)
12a, α-CH3CO (2%)
12b, β-CH3CO (13%)
Scheme 2. Preparation of tetrahydronaphthalene 11.
epoxide, which is cleaved by m-chlorobenzoic acid on
stirring for 3 h at 25°C as described by Stork to give
86% of diketone 10.
The aldol reaction of 10 to give dienone 11 was prob-
lematic since varying amounts of Michael adducts 12
are obtained. The best selectivity for 11 is achieved by
using excess strong base (5 equiv. of 1 M NaOMe in
MeOH), which gives 76% of the desired dienone 11 and
15% of a 7:1 mixture of 12b and 12a. More 12 (30–40%)
is obtained with 1 equiv. of NaOMe and little 11 is
formed with KOH in dioxane or DME or with pyrro-
lidine. The Michael adducts are formed as a 7:1 equi-
librium mixture and are not converted to 11 with excess
NaOMe in MeOH at reflux for 24 h, suggesting that the
Michael reaction is not readily reversible. The aldol
reaction is faster than the Michael reaction, but is
probably reversible. Dehydration of the aldol product is
irreversible and probably proceeds faster in excess base,
providing a plausible explanation for the observation
that excess base improves the yield of 11. The forma-
tion of Michael adduct byproducts may be the reason
for the limited use of Robinson annelations to prepare
dienones from cycloalkenones.
O
O
LDA, DMPU, 15 (90%)
SiMe2Ph
OTHP
8
16
m-CPBA
O
O
O
pyr·(HF)x
OR
O
SiMe2Ph
OTHP
18, R = THP
19, R = H
TsOH
MeOH/CH2Cl2
17
(79% from 16)
Scheme 4. Preparation of hydroxy dione 19.
Treatment of 19 with 5 equiv. of NaOMe in MeOH for
4 h affords 64% of the desired hydroxymethyl tetra-
hydronaphthalenone 24, 26% of the analogous
methoxymethyl tetrahydronaphthalenone 22, and 10%
of Michael products analogous to 12 (Scheme 5). Aldol
reaction gives alkoxide 20, which can form enolate 21.
Elimination of the hydroxyl group at the ring fusion
will give the desired product 24. Elimination of the side
chain hydroxyl group will give methylenecyclohexanone
23, which can undergo addition of methoxide and then
elimination of the ring fusion hydroxyl group to give
22.12 We thought that using a non-nucleophilic cosol-
vent might improve the selectivity for 24.
We now turned our attention to adapting the Stork–
Jung vinylsilane Robinson annelation to the introduc-
tion of a functionalized chain containing an aldehyde
precursor. The functionalized 3-silylallylic iodide can-
not be prepared analogously to 6 from trimethylsilyl-
propargyl alcohol. Fortunately, we found that addition
of phenyldimethylsilylcuprate9 to propargylic alcohol
1310 gives 76% of the desired phenyldimethylsilyl allylic
alcohol 14 and only 1–2% of the regioisomer (Scheme
3). CuCN-catalyzed addition of PhMe2SiLi·Et2Zn as
described by Oshima11 gives 94% of a difficultly separa-
ble 84:16 mixture of 14 and the regioisomer. Formation
of the mesylate of 14 with MsCl and Et3N and displace-
ment with NaI in acetone at 45°C provides 82% of the
desired functionalized dimethylphenylsilyl allylic iodide
15.
The best results are obtained by treating 19 with 5
equiv. of 0.2 M NaOMe in 4:1 benzene/MeOH, which
gives 72% of the desired product 24, 8% of methoxy
adduct 22, and 14% of Michael products. Treatment of
THP ether 18 with 5 equiv. of NaOMe in MeOH

B. B. Snider, B. Shi / Tetrahedron Letters 42 (2001) 9123–9126
9125
affords only 22 and Michael products, indicating that
then with 15 (68%) provides 32 (Scheme 7). Epoxida-
tion, protodesilylation with pyr·(HF)x and hydrolysis
with TsOH affords 70% of 33, which cyclizes with 5
equiv. NaOMe in 4:1 benzene/MeOH to give 51% of
hydroxy dienone 34.
OTHP is a better leaving group than OH.
OH
OH
5 eq NaOMe
4:1 C6H6/MeOH
25 °C, 2 h
NaO
OH
O
ONa
19
OTHP
SiMe2Ph
20
21
O
O
1) LDA, DMPU, MeI (71%)
2) LDA, DMPU, 15 (68%)
-NaOH
O
-NaOH
OH
OMe
O
NaOMe
-NaOH
32
H
O
1) m-CPBA
2) pyr·(HF)x
O
31
3) TsOH, MeOH, CH2Cl2
OH
OH
O
5 eq NaOMe
4:1 C6H6/MeOH
25 °C, 2 h
23
24 (72%)
22 (8%)
O
O
Scheme 5. Aldol reaction and dehydration to give 24.
To complete the model study, we needed to reduce the
ketone of 24 to the axial alcohol and oxidize the
primary alcohol to the aldehyde to introduce the func-
tionality present on the cyclohexene of guanacastepene
(1). Reduction of 24 with L-Selectride affords mainly
the equatorial cyclohexenol 25, while reduction with K-
or KS-Selectride13 gives a complex mixture (Scheme 6).
We therefore reduced 24 with LiAlH(O-t-Bu)3 to give a
98:2 mixture of 25 and 28 and used a Mitsunobu
reaction to invert the stereochemistry.14
34 (51%)
33 (70%)
Scheme 7. Synthesis of 34.
Regiospecific addition of the dimethylphenylsilyl anion
to propargylic alcohols not only provides easy access to
functionalized 3-silylallylic iodides, but is advantageous
for the parent system. Oshima’s procedure, CuCN-cata-
lyzed addition of Me2PhSiLi·Et2Zn to 2-butyn-1-ol
(35), provides 89% of the alcohol and 4% of the readily
separable regioisomer (Scheme 8). In this case, use of
the cuprate by Pancrazi’s procedure gives 74% of the
alcohol that is harder to purify. Mesylation and dis-
placement with NaI proceeds in 86% yield, completing
a three-step synthesis of the non-volatile, acid-stable
1-iodo-3-dimethylphenylsilyl-2-butene (36). In contrast,
five steps are required for the preparation of volatile,
acid-sensitive 1-iodo-3-trimethylsilyl-2-butene (6) from
2-propyn-1-ol. Preparation of 10 from 8 using 36 pro-
ceeds in 86% yield as compared to only 70% using 6.
OH
OBz
OBz
LiAlH(O-t-Bu)3
THF, 3 h
BzOH, Ph3P
DEAD, THF
-78→ 25 °C
OH
-78→ 25 °C
24
26, β-OBz
27, α-OBz
25
KOH, 45 °C, 2 h
MeOH,H2O
OH
OH
CHO
OH
+
CHO
OH
TEMPO
NCS
Bu4NCl
CH2Cl2/H2O
29
30
(7% from 24)
28, β-OH
25, α-OH
(56% from 24)
1) PhMe2SiLi, Et2Zn
I
CH3
HO
cat CuCN, THF (89%)
Scheme 6. Preparation of hydroxyaldehyde 29.
CH3
2) MsCl, Et3N, 0 °C
3) NaI, acetone, 45 °C
(86% two steps)
SiMe2Ph
36
35
Treatment of 25 with 4 equiv. of PPh3, DEAD and
BzOH affords an 88:12 mixture of the desired diben-
zoate 26 and stereoisomer 27, which is probably formed
by an SN1 reaction. The Mitsunobu reaction with p-
nitrobenzoic acid was less successful.15 Hydrolysis of
the mixture with KOH in 9:1 MeOH/H2O gives a
difficultly separable, acid-sensitive 88:12 mixture of 28
and 25 in 72% yield from 24. Selective oxidation of the
primary alcohol by Einhorn’s procedure16 with 2 equiv.
of TEMPO, 4 equiv. of NCS, and 1 equiv. of Bu4NCl
in CH2Cl2/H2O requires 16 h, but gives the desired
hydroxyaldehyde 29 in 56% overall yield from 24 along
with 7% of hydroxyaldehyde 30 formed from the 25 in
the mixture.17 The spectral data for 29 show the
expected similarity to those of guanacastepene.
Scheme 8. Synthesis of 36.
We decided to explore the utility of 36 with isophorone
(37) since reaction of the morpholine dienamine of
isophorone (42) with MVK has been reported to give
only 12% of tetrahydronaphthalenone 41.18 Alkylation
of 37 with 36 yields 86% of 38, which is epoxidized and
protodesilylated to give 88% of 39 (Scheme 9). Treat-
ment of 39 with 10 equiv. of NaOMe in 2:1 benzene/
MeOH affords 52% of a 3:1 mixture of Michael
adducts 40b and 40a, and 39% of the desired dienone
41. Although the data for 40b are identical to those
reported, the spectral data for 41 are similar to, but
clearly different from those previously reported.18 NOE
studies confirmed the structure of 39, suggesting that
This annelation sequence works equally well on 2-
cycloheptenone (31). Alkylation with MeI (71%) and

9126
B. B. Snider, B. Shi / Tetrahedron Letters 42 (2001) 9123–9126
References
O
O
LDA, DMPU, 36
SiMe2Ph
1. Brady, S. F.; Singh, M. P.; Janso, J. E.; Clardy, J. J. Am.
Chem. Soc. 2000, 122, 2116–2117.
2. Singh, M. P.; Janso, J. E.; Luckman, S. W.; Brady, S. F.;
Clardy, J.; Greenstein, M.; Maiese, W. M. J. Antibiot.
2000, 53, 256–261.
38
(86% based on
recovered 37)
1) m-CPBA
2) pyr·(HF)x
37
O
O
O
3. Snider, B. B.; Hawryluk, N. A. Org. Lett. 2001, 3,
569–572.
4. (a) Dudley, G. B.; Danishefsky, S. J. Org. Lett. 2001, 3,
2399–2401; (b) Dudley, G. B.; Tan, D. S.; Kim, G.;
Tanski, J. M.; Danishefsky, S. J. Tetrahedron Lett. 2001,
42, 6789–6791.
+
O
10 eq NaOMe
2:1 C6H6/MeOH
45 °C, 12 h
O
40a, α-CH3CO (13%)
40b, β-CH3CO (39%)
41 (39%)
39 (88%)
5. Magnus, P.; Waring, M. J.; Ollivier, C.; Lynch, V. Tetra-
hedron Lett. 2001, 42, 4947–4950, 6977.
Scheme 9.
6. (a) Boeckman, Jr., R. K.; Springer, D. M.; Alessi, T. R.
J. Am. Chem. Soc. 1989, 111, 8284–8286; (b) Jansen, B. J.
M.; Hendrikx, C. C. J.; Masalov, N.; Stork, G. A.;
Meulemans, T. M.; Macaev, F. Z.; de Groot, Ae. Tetra-
hedron 2000, 56, 2075–2094 and references cited therein.
7. (a) Stork, G.; Jung, M. E. J. Am. Chem. Soc. 1974, 96,
3682–3684; (b) Stork, G.; Jung, M. E.; Colvin, E.; Noel,
Y. J. Am. Chem. Soc. 1974, 96, 364–3686.
8. (a) Smith, H. A.; Huff, B. J. L.; Powers, W. J.; Caine, D.
J. Org. Chem. 1967, 32, 2851–2856; (b) Torri, J.; Azzaro,
M. Bull. Soc. Chim. Fr. Part 2 1978, 283–291.
9. Betzer, J.-F.; Pancrazi, A. Synlett 1998, 1129–1131.
10. Blanchette, M. A.; Malamas, M. S.; Nantz, M. H.;
Roberts, J. C.; Somfai, P.; Whritenour, D. C.; Masa-
mune, S.; Kageyama, M.; Tamura, T. J. Org. Chem.
1989, 54, 2817–2825.
we had prepared 41 and that the structure of the adduct
of 42 and MVK had been misassigned.
The literature data18 for the dienone prepared from
dienamine 42 are identical to those reported for
dienone 44, which has been prepared by oxidation of
b-ionone to give 43 and cyclization19 (Scheme 10).
Reaction of 42 with MVK could also give 43, which
will cyclize to 44, not 41.
O
O
O
MVK
tol
N
O
reflux
43
42
44 (12%)
11. Wakamatsu, K.; Nonaka, T.; Okuda, Y.; Tu¨ckmantel,
W.; Oshima, K.; Utimoto, K.; Nozaki, H. Tetrahedron
1986, 42, 4427–4436.
Scheme 10.
12. For a similar sequence, see: Kim, S.; Fuchs, P. L. J. Am.
In conclusion, we have extended the Stork–Jung
Robinson annelation procedure by using 15 with cyclo-
hexenone 5 and 2-cycloheptenone (31) to prepare
hydroxymethyl dienones 24 and 34. We have also devel-
oped a protocol to convert hydroxy ketone 24 to the
desired hydroxy aldehyde 29. This sequence should be
suitable for annelation of the guanacastepene cyclohex-
ene ring onto hydroazulene 4 to give 2 and further
elaboration to complete the synthesis of guanacaste-
pene (1). The regiospecific preparation of 3-silylallylic
halides 15 and 36 from propargylic alcohols should
significantly extend the scope of the Stork–Jung
Robinson annelation procedure.
Chem. Soc. 1993, 115, 5934–5940.
13. Wiebe, J. P.; Deline, C.; Buckingham, K. D.; Dave, V.;
Stothers, J. B. Steroids 1985, 45, 39–51.
14. Ho, T.-L.; Su, C.-Y. J. Org. Chem. 2000, 63, 3566–3568.
15. Hughes, D. L.; Reamer, R. A. J. Org. Chem. 1996, 61,
2967–2971.
16. Einhorn, J.; Einhorn, C.; Ratajczak, F.; Pierre, J.-L. J.
Org. Chem. 1996, 59, 7452–7454.
17. Oxidation of 28 with Ru(PPh3)3Cl2 in toluene under air
at 0°C provides 63% of 29, 15% of hydroxyketone 24,
and 5% of keto aldehyde. See: Tomioka, H.; Takai, K.;
Oshima, K.; Nazaki, H. Tetrahedron Lett. 1981, 22,
1605–1608.
18. (a) Hickmott, P. W.; Simpson, R. J. Chem. Res. (S) 1992,
302–303; (b) Hickmott, P. W.; Simpson, R. J. Chem. Res.
(M) 1992, 2447–2475.
Acknowledgements
´
19. Tutorskaya, O. O.; Elyashberg, M. E.; Yampol’skii, Yu.
Yu.; Karasev, Yu. Z.; Miropol’skaya, M. A.;
Samokhvalov, G. I. J. Org. Chem. USSR 1990, 26,
247–250; Chem. Abstr. 1990, 113, 59594b.
We thank the NIH (GM50151) for generous financial
support.