Total synthesis of guanacastepene A: A route to enantiomeric control

Article abstract of DOI:10.1021/jo051470k

The goal of the total synthesis of guanacastepene A served as a focus to bring together several chemical inquiries. One involved the synthesis of fused 5,7-hydrazulenones (see structure 20). Another issue had to do with the mechanistic intermediates in re

Full text of DOI:10.1021/jo051470k

VOLUME 70, NUMBER 26
DECEMBER 23, 2005
©
Copyright 2005 by the American Chemical Society
Total Synthesis of Guanacastepene A:
A Route to Enantiomeric Control
†
‡
§
|
⊥
Mihirbaran Mandal, Heedong Yun, Gregory B. Dudley, Songnian Lin, Derek S. Tan, and
Samuel J. Danishefsky*,
†,‡
Laboratory for Bioorganic Chemistry, Sloan-Kettering Institute for Cancer Research, 1275 York Avenue,
Box 106, New York, New York 10021, Department of Chemistry, Columbia University, Havemeyer Hall,
3
000 Broadway, New York, New York 10027
s-danishefsky@mskcc.org
Received July 15, 2005
The goal of the total synthesis of guanacastepene A served as a focus to bring together several
chemical inquiries. One involved the synthesis of fused 5,7-hydrazulenones (see structure 20).
Another issue had to do with the mechanistic intermediates in reductive cyclizations (see 17 to 18
and 19). The total synthesis required a mastery of an intramolecular Knoevenagel condensation of
a â,γ-unsaturated ketone (see compound 41). Actually, cyclization was best accomplished when
the terminal double bond of 41 was first converted to an epoxide. Further issues related to the
5
stereochemistry at C and, rather surprisingly, the propensity for â-face acetoxylation at C13.
Crystallographic verification of the assigned â-stereochemistry at C13 is provided. Finally, a route
to optically active material is provided (see compound 20). A key element in this construction was
an enantioselective addition of isopropenyl cuprate to 2-methylcyclopentenone (see compound 99).
Introduction
compounds that display activity against organisms (cf.
Staphylococcus aureus and Enterococcus faecalis) that are
resistant to agents such as methicillin and vancomycin.
The structure of guanacastepene A, a diterpenoid which
registered positive in this screening context, was solved
by Clardy and colleagues. As is often the case with
complex isolates, the assignments of gross structural
connectivity and stereostructures were first attempted
using spectroscopic criteria, including NMR ( H and C)
and high-resolution mass spectroscopy as well as infrared
and ultraviolet measurements. In the end, it remained
for single-crystal X-ray diffraction to clarify the structure
of guanacastepene A (1, Figure 1).1
Guanacastepene A was first identified and isolated as
part of a program to evaluate Endophytic fungi as sources
of potential antibiotics. In particular, the focus of the
1
search was directed to the identification and isolation of
†
Laboratory for Bioorganic Chemistry, Sloan-Kettering Institute for
Cancer Research.
‡
Department of Chemistry, Columbia University.
Present address: Department of Chemistry and Biochemistry, The
1
13
§
Florida State University.
|
Present address: Department of Medicinal Chemistry, Merck
Research Laboratories.
⊥
Present address: Tri-Institutional Research Program and Molec-
ular Pharmacology & Chemistry Program, Memorial Sloan-Kettering
Cancer Center.
The combination of the novel structure and promising
screening genealogy of guanacastepene A converged to
prompt significant interest in the community of chemists
as to its total synthesis. The privileged target status of
(1) (a) Brady, S. F.; Singh, M. P.; Janso, J. E.; Clardy, J. J. Am.
Chem. Soc. 2000, 122, 2116. (b) Singh, M. P.; Janso, J. E.; Luckman,
S. W.; Brady, S. F.; Clardy, J.; Greenstein, M.; Maiese, W. M. J.
Antibiot. 2000, 53, 256. (c) Brady, S. F.; Bondi, S. M.; Clardy, J. J.
Am. Chem. Soc. 2001, 123, 9900.
1
0.1021/jo051470k CCC: $30.25 © 2005 American Chemical Society
Published on Web 09/20/2005
J. Org. Chem. 2005, 70, 10619-10637
10619

Mandal et al.
FIGURE 2. Robinson annulation.
An evaluation of the logic of the Robinson annulation
for the purposes of assembling guanacastepene A leads
to the formalized hypothetical progression 4 f 5 f 6 f
FIGURE 1. Structure of guanacastepene A.
at the chemistry level survived some otherwise adverse
1
1
(Scheme 1). A wealth of methods, protocols, and
information with respect to its likely emergence as a
valuable resistance-breaking antibiotic agent. Thus, gua-
nacastepene A exhibits hemolytic activity against human
red blood cells.1b Clearly, for the guanacastepene A
structure to survive as a platform for discovering new
antibiotics of value in treating infectious diseases, it
would be necessary to differentiate antibiotic from
reagents have been fashioned and coordinated to facili-
tate the core transformation of 2 f 3. Given the extensive
research that has been directed to cover countless
permutations of the central theme captured in this
8
expression, it is perhaps inevitable that the literature
contains seemingly redundant solutions to this problem.
However, closer consideration of the precise issues in a
particular synthetic approach may well oblige the chem-
ist to draw from a specialized subset of the overall
Robinson annulation methodology. Thus, the existence
of a range of protocols and underlying thought processes,
which can be grouped under the rubric of “Robinson
annulation,” adds much to its value.
2
hemolytic selectivity. Certainly, the best opportunity for
accomplishing this goal would lie in gaining access to a
range of congeners of guanacastepene A. A careful survey
of the biological properties of such a focused compendium
of guanacastepene-related compounds may provide sug-
gestions as to directions to be followed for decoupling
antibiotic and hemolytic prospects in this series. This
potentially serious complication at the translational level
notwithstanding, guanacastepene still maintains its
standing as a stimulating target for chemical total
synthesis. It was in such a chemistry-dominated context
that our laboratory pursued this goal. Below, we record
the realization of the goal of the total synthesis of
guanacastepene A.3 It will be seen that the synthesis
was accomplished in the setting of a series of compli-
cationssmany of them unexpected.
Even a cursory inspection of the guanacastepene case
at hand soon reveals the need for improvisation in that
the sequence to be employed (4 f 5) must, ultimately,
give rise to a cycloheptenone rather than to the usual
cyclohexenone (cf. 2 f 3). Moreover, in moving toward
1
, the cycloheptenone must emerge such that its substit-
uents at carbons 8 and 11 translate, in the end, to the
required C -C11 anti backbone relationship. Well in
advance of dealing with the C
,4
8
8
-C11 anti correlation, there
would be a need to provide for the syn C11-C12 connectiv-
Given the extensive record of involvement of our
laboratory in the synthesis of steroids and biogenetically
ity.
5
related polyisoprenoids, it was not surprising that we
Results/Discussion
sought to translate the guanacastepene synthesis prob-
lem to that of a variation of the synthetic logic describable
under the term “Robinson annulation,” which can be
encompassed in the expression 2 f 3 (Figure 2). While
the short-term goal of the Robinson annulation itself is
well understood by practitioners of synthesis, its concep-
tual reach is broad and powerful, encompassing a diverse
range of methodological platforms. 6
We set, as our first goal, the synthesis of compound 5
1
(R ) H) starting from 2-methylcyclopentenone 4. Our
thought to reach 5 envisioned C-alkylation of the kineti-
9
cally generated enolate 7, fashioned from 4 using
methodology demonstrated by Piers,10 with an agent, RX
(Scheme 2). Such a sequence would be expected to secure
the future C -C anti relationship. RX would be chosen
,7
11
12
so that, in due course, there would be generated a
â-ketophosphonate, 12. It was assumed that intramo-
lecular Horner-Wadsworth-Emmons11 (HWE) reaction
would lead to the homo-Robinson annulation product 5.
(
2) For an explanation of the selectivity of membrane-lysing peptide
antibiotics, see: Zasloff, M. Nature 2002, 415, 389.
3) (a) Dudley, G. B.; Danishefsky, S. J. Org. Lett. 2001, 3, 2399. (b)
(
Dudley, G. B.; Danishefsky, S. J.; Sukenick, G. Tetrahedron Lett. 2002,
4
4
3, 5605. (c) Mandal, M.; Danishefsky, S. J. Tetrahedron Lett. 2004,
5, 3827. (d) Dudley, G. B.; Tan, D. S.; Kim, G.; Tanski, J. M.;
12
We initially took recourse to 5-iodo-1-pentene (8) as
RX.3a
Danishefsky, S. J. Tetrahedron Lett. 2001, 42, 6789. (e) Mandal, M.;
Danishefsky, S. J. Tetrahedron Lett. 2004, 45, 3831. (f) Tan, D. S.;
Dudley, G. B.; Danishefsky, S. J. Angew. Chem., Int. Ed. 2002, 41,
In pursuing the synthesis as discussed below, we were
operating in the racemic series. As will be described, after
2
185. (g) Lin, S.; Dudley, G. B.; Tan, D. S.; Danishefsky, S. J. Angew.
Chem., Int. Ed. 2002, 41, 2188. (h) Yun, H.; Danishefsky, S. J.
Tetrahedron Lett. 2005, 46, 3879.
(7) Reviews: (a) Gawley, R. E. Synthesis 1976, 777. (b) Jung, M. E.
Tetrahedron 1976, 32, 3.
(
4) For other syntheses of structures which connected with late
compounds described herein, thus constituting formal syntheses, see:
a) Shi, B.; Hawryluk, N. A.; Snider, B. B. J. Org. Chem. 2003, 68,
(8) Cf. inter alia: (a) Stork, G.; Ganem, B. J. Am. Chem. Soc. 1973,
95, 6152. (b) Stork, G.; Jung, M. E. J. Am. Chem. Soc. 1974, 96, 3682.
(c) Scanio, C. J. V.; Starrett, R. M. J. Am. Chem. Soc. 1971, 93, 1539.
(d) Zoretic, P. A.; Golen, J. A.; Saltzman, M. D. J. Org. Chem. 1981,
46, 3554. (e) Heathcock, C. H.; Ellis, J. E.; McMurry, J. E.; Coppolino,
A. Tetrahedron Lett. 1971, 12, 4995.
(
1
030. (b) Boyer, F.-D.; Hanna, I.; Ricard, L. Org. Lett. 2004, 6, 1817.
For other stimulating work in the guanacastepene series, cf. inter
alia: (c) Nguyen, T. M.; Seifert, R. J.; Mowrey, D. R.; Lee, D. Org. Lett.
2
1
002, 4, 3959. (d) Du, X.; Chu, H. V.; Kwon, O. Org. Lett. 2003, 5,
923. (e) Shipe, W. D.; Sorensen, E. J. Org. Lett. 2002, 4, 2063. (f)
(9) Stork, G.; Hudrlik, P. F. J. Am. Chem. Soc. 1968, 90, 4464.
(10) Piers, E.; Renaud, J.; Rettig, S. J. Synthesis 1998, 590.
(11) Maryanoff, B. E.; Reitz, A. B. Chem. Rev. 1989, 89, 863.
(12) Iodo-1-pentene was prepared by treatment of 5-bromo-1-pentene
with 3.0 equiv of NaI in refluxing acetone for 6 h as described
previously: Padwa, A.; Kamigata, N. J. Am. Chem. Soc. 1977, 99, 1871.
Hughes, C. C.; Kennedy-Smith, J. J.; Trauner, D. Org. Lett. 2003, 5,
4
1
113. (g) Sarpong, R.; Su, J. T.; Stoltz, B. M. J. Am. Chem. Soc. 2003,
25, 13624.
(
5) Danishefsky, S. J. Tetrahedron 1997, 53, 8689.
(
6) Rapson, W. S.; Robinson, R. J. Chem. Soc. 1935, 1285.
10620 J. Org. Chem., Vol. 70, No. 26, 2005

Total Synthesis of Guanacastepene A
SCHEME 1. Synthetic Strategy toward Guanacastepene A
SCHEME 2. Synthesis of Horner-Wadsworth-Emmons Precursor 12a
a
Key: (a) MeLi, THF, 0 °C, 1 h, then 2.5 equiv of 8, HMPA, -78 °C to rt, 63-72%; (b) Chloramine-T, H2O, H2SO4, acetone, 50 °C, 30
min; (c) Jones’ reagent, rt, 30 min, 54-63% for two steps; (d) (i) NaI, acetone, 15 min; (ii) (EtO)3P, PhH, reflux, 5 h, 70-74%.
SCHEME 3. Unsuccessful Horner-Wadsworth-Emmons Cyclizationa
a
Key: (a) Cs2CO3, PhMe, reflux, 12 h, 83% (14).
SCHEME 4. Preparation of Vinyl Iodide 17a
the total synthesis of guanacastepene was accomplished,
we turned our attentions to charting a path that would
lead to enantiohomogeneous guanacastepene. Here, again,
we were successful.
Alkylation of enolate 7 with 8 generated 9 in ap-
proximately 65-70% yield. According to plan, 9 lent itself
to transformation to 12 via ketone 11, in turn fashioned
from 10 via Jones oxidation. Compound 10 arose follow-
ing oxidation of the terminal olefin linkage of 9 with
13
Chloramine T. Unfortunately, several attempts to reach
compound 5 via HWE cyclization of 12 were unsuccessful
a
Key: (a) PPh3, imid, I2, CH2Cl2, 1 h, 92%; (b) MeLi, THF, 0
C, 1 h, then 2.5 equiv of 16, HMPA, -78 °C to rt, 74-76%.
(
Scheme 3). While we did not follow up in detail on every
°
failed attempt to synthesize 5, the case of the reaction of
2 with cesium carbonate was particularly telling. The
1
annulation by direct aldolization. Rather, another pos-
sibility to accomplish the consequences, if not the hoped-
for direct intramolecular style, of the homo-Robinson
annulation presented itself. As will be seen, this method,
while solving the 4 f 5 transformation, was not without
its own complications. It started on a promising note.
Thus, alkylation of enolate 7 with diiodide 16, prepared
product was 14 rather than the anticipated 5. Presum-
ably, 14 arose from cyclized product 13, following the
general pathway shown.14
Apparently, in the case of 12, internal aldolization
using a 5-exo-initiated pathway is kinetically favored over
the 7-endo bond closure, notwithstanding the enhanced
acidity of the â-ketophosphonate network. Indeed, this
setback underscores a key issue of directionality that
would have to be solved if transformations of the genus
15
as shown from the known 15, gave rise to 17 in ca. 75%
yield (Scheme 4).3a
At this stage, our hope was to accomplish reductive
cyclization of 17 with concomitant formation of 18.
1
6,17
4
f 5 were to be realizable by direct aldolization as its
ring-closing step. We shall revisit this question after the
total synthesis of the racemate is described.
Anxious to focus on the guanacastepene problem, we
set aside our preference to realize the homo-Robinson
In the event, lithiation of 17 with n-butyllithium in
hexanes afforded a 1.6:1 mixture of 18:19 (Scheme 5).
For obvious reasons, we term the former as the product
of reductive cyclization, while 19 is referred to as the
(
13) For the hydroxy-chlorination of olefins with chloramine-T,
(15) Yenjai, C.; Isobe, M. Tetrahedron 1998, 54, 2509.
(16) For a seminal report on this general cyclization strategy, see:
Corey, E. J.; Kuwajima, I. J. Am. Chem. Soc. 1970, 92, 395.
(17) Piers, E.; Walker, S. D.; Armbrust, R. J. Chem. Soc., Perkin
Trans. 1 2000, 633.
see: Damin, B.; Garapon, J.; Sillion, B. Synthesis 1981, 362.
(14) Snider and Hawryluk also observed only formation of a 5,5-
fused ring system in their attempts to perform a direct aldol with a
similar substrate. See ref 4a.
J. Org. Chem, Vol. 70, No. 26, 2005 10621

Mandal et al.
SCHEME 5. Reductive Cyclization of 17a
SCHEME 6. Preparation of Dideuterated Vinyl
a
Iodide 17-d
2
a
Key: (a) Et3N, CD3OD, rt, 20 h, 94% of 17-d1; (b) DBU,
CD3OD, rt, 4 h, 91% of 17-d2.
reduction pathway, suppression of cyclopentanone proton
transfer should result in a corresponding attenuation of
formation of 19. Thus, we surmised that if the R-protons
were replaced with less “kinetically acidic” deuterium
atoms, the product distribution should shift in a direction
favoring cyclization over reduction.19 In practice, deu-
teration of starting material 17 was accomplished with
DBU in deuterated methanol (Scheme 6). Interestingly,
the use of a weaker base, such as triethylamine, gave
rise only to monodeuterated intermediate. We note that
deuteration had occurred with selectivity, as shown, at
a
Key: (a) 2.2 equiv of n-BuLi, THF, 0 °C, 30 min, 1.6:1 18:19;
b) PCC, powdered sieves, CH2Cl2, ca. 70%.
(
reduction product. Upon treatment with PCC, compound
1
8 undergoes a precedented oxidative rearrangement18
to give rise to enone 20.
In an effort to improve the yield of 18, we examined
the effects of concentration on the ratio of formation of
3b
the R-face (17-d
1
).
1
8 relative to 19. Indeed, we found that, upon dilution
0 1 2
Each substrate (17-d , 17-d , and 17-d ) was subjected
of the reaction mixture, the bias toward formation of the
desired intramolecular cyclization product, 18, could be
further encouraged. Under optimal conditions, the ratio
of 18:19 could be increased to 3.5:1. Since decreasing
concentration favors intramolecular reaction pathways,
these results suggest that the undesired reduction prod-
uct, 19, proceeds, at least partially, through an intermo-
lecular protonolysis pathway. However, as a practical
matter, we were already operating in a dilute setting and
we could not hope to solve our problem by adjustments
of the concentrations.
to standard cyclization conditions, and the ratio of
formation of cycloadduct (18) to reduction product (19)
was evaluated (Scheme 7). As shown below, although
incorporation of deuterium did not entirely eliminate the
formation of 19, it did significantly shift the product ratio
in favor of 18. These results implicate enolization of the
cyclopentanone as the major proton source in the reduc-
tion pathway. Interestingly, although deuterium isotopes
are frequently employed as valuable tools in probing
reaction mechanisms, they are only rarely considered as
a means to overcome problems in synthetic chemistry.
A propos of the problem of competitive noncyclizing
reduction, we sought to identify the source of the hy-
pothesized proton transfer. We were able to rule out
transfer from solvent by conducting the reaction in THF-
To probe the mechanism of the protonolysis still
further, we examined the origins of formation of 19-d
2
,
the noncyclized reduction product of the dideutero sub-
3c
strate (Scheme 8). Upon quenching the reaction mixture
with H O, we found that only 30% of the total uncyclized
product incorporated deuterium at the vinylic position
cf. 22:23). This interesting result seems to teach that in
d
8
. No deuterium incorporation in 19 and no change in
the ratio of products was observed. Furthermore, workup
of the reaction mixture with D O resulted in no exchange
2
2
(
of iodine for deuterium in 17, indicating that at the time
of the quenching of the reaction, there was no remaining
viable vinyllithium species. Postulating that the enoliz-
able cyclopentanone may actually serve as the proton
source, we treated the reaction mixture with acetic
anhydride for 30 min prior to workup. Indeed, following
this sequence, one could isolate significant but variable
amounts of enol acetate 21. This outcome indicates the
existence of an enolate intermediate in the reaction
mixture.
the dideutero case, the major proton source in the much
suppressed reduction pathway is no longer the enolizable
cyclopentanone. When the mixture was quenched with
2
D O, minimal change in the ratio of 22 to 23 was
observed. It was thus deemed unlikely that significant
amounts of vinyllithium species remained, preworkup.
Furthermore, the solvent was seen to be a highly improb-
able source of proton transfer, as conduct of the reaction
8
in THF-d did not result in deuterium incorporation at
the vinylic position. Accordingly, we are left with the
possibility that the “n-butyl” group may be a minor
contributing proton source in the quenching of the
vinyllithium intermediate. For instance, perhaps, iodobu-
tane, which is formed in the initial metal-halogen
(
19) For a general discussion of isotope effects, see: Lowry, T. H.;
Richardson, K. S. Isotope effects. In Mechanism and Theory in Organic
Chemistry, 3rd ed.; Harper Collins: New York, 1987; p 232. For related
applications to synthesis, see: (a) Laplaza, C. E.; Davis, W. M.;
Cummins, C. C. Organometallics 1995, 14, 577. (b) Schrock, R. R.;
Baumann, R.; Reid, S. M.; Goodman, J. T.; Stumpf, R.; Davis, W. M.
Organometallics 1999, 18, 3649. (c) Vedejs, E.; Little, J. J. Am. Chem.
Soc. 2002, 124, 748.
If, as the evidence suggests, the enolizable protons R
to the keto group are serving as proton sources in the
(18) For an early example of a similar oxidative transposition, see:
B u¨ chi, G.; Egger, B. J. Org. Chem. 1971, 36, 2021.
10622 J. Org. Chem., Vol. 70, No. 26, 2005

Total Synthesis of Guanacastepene A
SCHEME 7. Improving the Cyclization of 17 through a Kinetic Isotope Effecta
a
Key: (a) 5.0 equiv of n-BuLi (inverse addition), THF, 0 °C, 30 min.
SCHEME 8. Investigating the Mechanism of the Noncyclizing Reductiona
a
Key: (a) 5.0 equiv of n-BuLi (inverse addition), THF, 0 °C, 30 min.
SCHEME 9. Strategy for Formation of 25 through Dialkylation
SCHEME 10. Unsuccessful Dialkylation of 20a
exhange of nBuLi and 17, may undergo E2 elimination,
generating a source for protonolysis of the vinyllithium
species.
In summary, the main lesson of this detour is that, in
the nondeutero pathway described earlier, the main
proton source in the undesired reduction is the enolizable
cyclopentanone of 17. Replacing the enolizable protons
of 17 with less kinetically acidic deuterium isotopes
significantly suppresses the noncyclizing reduction path-
way. The minimal amount (<9%) of reduction product,
which is still observed in the dideutero case, is perhaps
attributable to an E2 elimination of the n-butyl iodide
species generated in the lithium-halogen exchange.
We return to the guanacastepene A total synthesis
problem. With compound 20 in hand, we directed our
attention to system 6. At this stage of our deliberations,
much design and experimentation would be necessary
before defining the optimal nature of the R function at
a
Key: (a) LiHMDS, allyl iodide; (b) LiHMDS (MeI or TMSCl).
C in a sequential fashion. To do this in a sensible way,
8
it would be necessary to know the likely outcome, at the
stereochemical level, of the transformation of 20 to 25
(Scheme 9). In particular, we wanted to gain an assess-
ment as to whether the last group introduced in the
alkylation event would enter syn or anti to the resident
methyl function at C . In principle, specific schemes
could be proposed to deal with either stereochemical
eventuality. While a priori it might be assumed that the
group introduced in the last alkylation event would enter
anti to the resident C11 methyl function, this issue had
to be studied, particularly as to the sorts of selectivity
margins to be anticipated.
11
8
C . As in the case of the building of the seven-membered
B ring, the paradigm of Robinson annulationsbroadly
construedsserved to help organize our thinking. In other
8
words, R at C must, at an appropriate stage undergo
cyclization (presumably by some variation of the aldol
condensation).
However, before addressing the all-critical specifics of
such a prospectus, it was necessary to establish the sort
of latitude open to us in terms of the interfacing of the
We initially sought to accomplish dialkylation of 20
through sequential installation of the allyl and methyl
substituents.3d Accordingly, 20 was deprotonated and
treated with allyl iodide to give rise to 26 (Scheme 10).
We were pleased to find that allylation had proceeded
with high stereoselectivity (although the relative config-
uration of the allyl group had not been determined at
this stage). Unfortunately, our attempts to install the
8
configurations at C and C11. Not surprisingly, our
earliest thoughts were to introduce the methyl group and
the future R group (albeit undefined in its specifics) at
J. Org. Chem, Vol. 70, No. 26, 2005 10623

Mandal et al.
SCHEME 11. Preparation of Silyl Enol Ether 28a
For comparative purposes, we sought to reverse the
order of introduction of the methyl and allyl groups at
the quaternary carbon center. It was assumed that
inverting the sequence of the two alkylation events would
lead to the C
8
epimer of 29. The synthesis of this target
commenced with 1,4-hydride addition to 27 (Scheme
27
1
3). This step was followed by addition of allyl iodide.
Importantly, the formation of 29 was not observed from
this reaction sequence. Rather, a new product, isomeric
with 29, was obtained. This compound was formulated
as 32. Thus, an important finding had been made. The
a
Key: (a) 1.5 equiv of LiHMDS, THF, -78 °C, 1 h, then 3.0
equiv of Eschenmoser’s salt, THF, -78 °C to rt, 20 min; (b)
m-CPBA, CH2Cl2/NaHCO3 (aq) (2:1), 60-70% over two steps; (c)
.0 equiv of vinyl-MgBr, 1.5 equiv of CuI, 4.5 equiv of HMPA, 4.5
equiv of TMSCl, THF, -78 °C, 20 min, 77%.
3
8
stereochemistry at C is indeed dictated by the order in
which the substituents are introduced, with the second
alkylation occurring anti to the C11 methyl group.
The relative stereochemistries of 29 and 32 were first
assigned through extensive NMR analysis (Figure 3). In
methyl group in a following step were hampered by the
difficulty of generating the R-enolate (i.e., C ) of 26.
8
Clearly, our original direct dialkylation strategy would
have to be reconsidered. Nonetheless, the success of the
first alkylation (cf. 20 f 26) did provide us with some
valuable encouragement as to the feasibility of achieving
1
13
summary, H and C resonances were assigned from
J-resolve, COSY, NOESY, DEPT, HMQC, and HMBC
data. In analogy to the modeling studies of guanacaste-
stereoselective alkylation at C
8
of a cycloheptenone
1a
pene by Clardy and co-workers, two possible gauche
system of the type 20.
butane-conformers about the C
identified using the MM2 force field. NOE interactions
involving C12-H, C17-H , and C19-H supported assign-
ment as the pseudo-chair conformation shown below. The
stereochemical configuration could then be assigned
based upon NOE interactions between C16-H and both
-H and C10-H . The NMR analysis of 32 was
complicated by overlapping H signals; however, the C
NOE interactions are consistent with the pseudo-chair
conformation and C conformation shown.
Although we had foreseen some level of selectivity for
introduction of the C methyl group anti to the resident
11 methyl substituent (cf. 28 f 29), we had not
9
-C10 bond of 29 were
A revised strategy was designed to circumvent the
problematic enolate regeneration step. We postulated
that an exo-methylene group, such as found in 27, could
be of value in a modified route (Scheme 11). To synthesize
3
3
C
8
2
7, we revisited chemistry that had been developed in
3
our own laboratory some years ago. Thus, cyclopentenone
C
9
R
R
2
0 was subjected to enolization, as shown, and the
1
7
-
20
presumed enolate was treated with Eschenmoser’s salt
H
2
to produce the dimethylaminomethylene Mannich base
8
2
1
product. Upon exposure of this material to m-CPBA,
elimination occurred readily to afford 27. Vinyl cuprate
addition to 27 in the presence of TMSCl22 provided our
target compound, 28, in which the enolate had been
trapped as a silyl enol ether.
8
C
anticipated apparently exclusive methylation from the
R-face. Needless to say, we were very pleased with this
favorable outcome, which would prove to be of great value
in the completion of the synthesis of guanacastepene A.
Though we have not pursued the generality of the
finding, this result may underscore the underappreciated
stereoselectivity margins that can be attained in the
alkylation of appropriately substituted cycloheptenone
systems.
The lithium enolate was then generated from 28 as
shown (Scheme 12). Exposure of this enolate to the action
of methyl iodide provided exclusively 29 (vide infra for
structural arguments in support of this stereochemical
assignment). Importantly, no other diastereomer was
observed following this reaction.23 At this point, the
ketonic function of 29 was protected as a dioxolane ketal
with concurrent olefin migration,24 and the terminal
olefin was taken through a hydroboration-oxidation
We return to compound 31. Our plan envisioned
28
utilization of a Turner-Fujimoto logic, wherein the
25
protocol to afford 30. Following concomitant ketal
carbonyl group of an intermediate enol lactone (cf. 33)
would be attacked with generalized nucleophile 34, with
deprotection and Jones oxidation,26 the target compound
3
1 was in hand. We note, parenthetically, that in the
3f
the expectation of reaching 35. Indeed, the desired enol
deprotection of the ketal precursor to 31, reconjugation
29
lactone 33 was prepared as shown (Scheme 14). We
of the enone had not occurred.
examined its reactions with methyl (34, Y ) H) and ethyl
(
34, Y ) Me) nucleophiles. However, in practice, we were
(
20) Schreiber, J.; Maag, H.; Hashimoto, N.; Eschenmoser, A. Angew.
unable to isolate or even identify any of the hoped-for 35
or its fully conjugated dienone isomer. Instead complex
mixtures of products were obtained. We note that, unlike
the situation in the Turner-Fujimoto reaction, the
carbonyl group in 33 is present in the context of a dienol
lactone. Thus, hypothetical attack by 34 leads to dieno-
late 36. By analogy to the textbook dehydro case, the
sequence 36 f 37 f 38 f 35 would be necessary. In the
case at hand, somewhere in this projected sequence, the
Chem., Int. Ed. Engl. 1971, 10, 330.
(
21) (a) Danishefsky, S. J.; Kitahara, T.; McKee, R.; Schuda, P. F.
J. Am. Chem. Soc. 1976, 98, 6715. (b) Piers, E.; Marais, P. C. J. Chem.
Soc., Chem. Commun. 1989, 122.
(
22) For a discussion of the role of TMSCl in cuprate additions, with
leading references, see: Frantz, D. E.; Singleton, D. A. J. Am. Chem.
Soc. 2000, 122, 3288.
(
23) For an example of another stereoselective cycloheptenone
alkylation, see: Pearson, A. J.; Bansal, H. S. Tetrahedron Lett. 1986,
7, 287.
24) (a) Stork, G.; Uyeo, S.; Wakamatsu, T.; Grieco, P.; Labovitz, J.
2
(
J. Am. Chem. Soc. 1971, 93, 4945. (b) (a) Daignault, R. A.; Eliel, E. L.
Organic Syntheses; Wiley: New York, 1973; Collect. Vol. V, p 303. (b)
De Leeuw, J. W.; De Waard, E. R.; Beetz, T.; Huisman, H. O. Recl.
Trav. Chim. Pays-Bas 1973, 92, 1047.
(27) Ojima, I.; Kogure, T. Tetrahedron Lett. 1972, 13, 5035.
(28) (a) Turner, R. B. J. Am. Chem. Soc. 1950, 72, 579. (b) Fujimoto,
G. I. J. Am. Chem. Soc. 1951, 73, 1856.
(29) Barkley, L. G.; Knowles, W. S.; Raffelson, H.; Thompson, Q. E.
J. Am. Chem. Soc. 1956, 78, 4111.
(
25) Brown, H. C.; Chen, J. C. J. Org. Chem. 1981, 46, 3978.
(
26) Bowden, K.; Heilbron, I. M.; Jones, E. R. H.; Weedon, B. C. L.
J. Chem. Soc. 1946, 39.
10624 J. Org. Chem., Vol. 70, No. 26, 2005

Total Synthesis of Guanacastepene A
SCHEME 12. Preparation of Acid 31a
a
Key: (a) 1.5 equiv of MeLi, THF, 0 °C, 10 min, then 5.0 equiv of HMPA, 5.0 equiv of MeI, -78 °C to rt, 15 min, 96%; (b) (CH2OH)2,
p-TsOH, PhH, reflux, 11 h, 88%; (c) 9-BBN, THF, rt, 1 h, then 3 M NaOH, 30% H2O2, rt, 3 h, 71%; (d) Jones’ reagent, acetone, 2 h, 77%.
SCHEME 13. Preparation of the Opposite
ester, 44, arising from formal retro-Claisen cleavage of
the â-ketoester ensemble in the side chain of 41. In
addition, in the absence of provisions to rigorously
exclude all possible contamination from oxygen, two other
products, 45 and 46, were identified, albeit in low yields
and with poor reproducibility. These two compounds were
separated and their NMR spectra studied. It was promis-
ing to find that each of these tricyclic core structures
exhibited conformational mobility, as judged by the
coalescence of otherwise complex spectra following heat-
ing. Similar conformational behavior had been described
for guanacastepene A.1a
We first focused on understanding the nature of the
retro-Claisen cleavage step, which clearly had occurred
in preference to the expected formation of compound 42.
Needless to say, such a retro-Claisen reaction on an
enolizable â-ketoester, particularly one bearing two
highly enolizable R-protons, had not been expected,
certainly under the relatively mild conditions (ethanol,
a
Diastereomer, 32
a
Key: (a) PhMe2SiH, Wilkinson’s catalyst ([Ph3P]3RhCl; 1%),
benzene, reflux, 20 min; (b) MeLi, THF, 0 °C, 10 min, then allyl
iodide, -78 °C to rt, 30 min, 69% over two steps.
analogy with the usual Turner-Fujimoto progression28
breaks down.
Of course, our original purpose in trying to realize the
sequence 33 f 36 f 35 was the hope that, in the end, Y
in 35 would correspond to a group (cf. inter alia CtN,
2
CO R, CHO) that would facilitate building the olefinic
hydroxyaldehyde functionality embracing carbons 15, 4,
and 5. With the breakdown of this initiative, alternatives
directed to the same subgoal were pursued. 3
6
0 °C) used in this experiment. In an orienting experi-
ment, we studied the possibility of executing a deliberate
retro-Claisen reaction on ketal 47 (Scheme 17). Interest-
ingly, even under conditions identical to those used on
f
In particular, we returned to compound 30. Oxidation
of the latter under controlled conditions afforded ketal
aldehyde 39 (Scheme 15).30 As was anticipatable by
precedent, the aldehyde function reacted smoothly with
4
1, compound 48 was not observed. Thus, the ethoxide-
mediated degradative excision of the carboxymethyl
group is dependent on the existence of a carbonyl group
31
2
ethyldiazoacetate in the presence of SnCl . In this way,
compound 40, equipped with a â-ketoester on the side
chain, was in hand. Finally, deprotection of the ketal
linkage led to 41, wherein the â,γ-unsaturated keto
arrangement in the future B,C sector had withstood
potential conjugation.32
With compound 41 in hand, it seemed that there was
now a clear path to be followed to reach the tricyclic core
structure of guanacastepene A. It was expected that an
intramolecular Knoevenagel condensation would com-
plete construction of the ring system (Scheme 16).33 We
recognized that such a reaction could produce either the
fully conjugated dienyl ring structure 43 or, conceivably,
the partially conjugated dienone, 42. In either case, we
could imagine programs that would be feasible for
introduction of the required oxygen at C14 and, thence,
3
at the future C of guanacastepene A.
We took this finding to suggest that the retrograde
Claisen step is driven by attack of the cycloheptanone
ketonic oxygen (either as its dienolate 49 or as its
ethoxide addition product, 50) on the â-ketone of the
â-ketoester side chain (Scheme 18). In either case, further
unraveling of the system leads to the observed major
product, 44.
Furthermore, it was of interest that no adventitious
oxidation products were noted in the attempted retro-
grade Claisen reaction. Conversely, in the early stages
of the study, we could not obtain Knoevenagel products
(
cf. 42 or 43) independent of oxidation.
We could envision two possible pathways that would
account for the products observed from the Knoevenagel-
like cyclization reaction. In one, adventitious oxidation
occurs on starting material 41. The oxidized intermedi-
ates thus produced would be quite prone to cyclization,
affording 45 and 46. Failure to find products 42 and 43
might suggest that cyclization in the absence of prior
oxidation does not occur. Alternatively, it is conceivable
that such a cyclization does occur, but under the condi-
tions of our experiment, adventitious oxygenation of the
product gives rise to the observed 45 and 46.
We tested this second possibility by conducting the
cyclization reaction in deuterated solvent, in an anaerobic
setting in a gastight NMR tube, involving repeated
freeze-thawing and evacuation (Scheme 19). Under
the acetoxy function at C13
.
It was, therefore, surprising to find that under a
variety of attempted conditions to achieve intramolecular
Knoevenagel condensation we could not obtain either
compound 42 or 43. The major product was actually the
(
30) (a) Dess, D. B.; Martin, J. C. J. Org. Chem. 1983, 48, 4155. (b)
Ireland, R. E.; Liu, L. J. Org. Chem. 1993, 58, 2899.
(
31) Holmquist, C. R.; Roskamp, E. J. J. Org. Chem. 1989, 54, 3258.
32) Hagiwara, H.; Uda, H. J. Chem. Soc., Chem. Commun. 1987,
(
1
351.
(33) We were encouraged by literature precedent involving cycliza-
tion of a similar â-keto ester side chain onto a cycloheptanone system
with an adjacent quaternary center: van der Baan, J. L.; Barnick, J.
W. F. K.; van Beek, G.; Bickelhaupt, F.; Spek, A. L. Tetrahedron 1992,
4
8, 2773.
J. Org. Chem, Vol. 70, No. 26, 2005 10625

Mandal et al.
FIGURE 3. Pseudo-chair and pseudo-twist-boat conformations of 29 and 32 with selected strong NOE interactions. Interactions
used to assign the C8 stereochemical configuration are shown in bold.
SCHEME 14. Unsuccessful Attempts at Turner-Fujimoto Cyclizationa
a
Key: (a) NaOAc, Ac2O, 140 °C, 80 min, 59%; (b) MeLi or EtMgBr.
SCHEME 15. Preparation of â-Keto Ester 41a
While we were gaining some confidence in our under-
standing of the variously competitive pathways operating
in our system, we had not yet realized a protocol by which
we could obtain ω-oxidized products of the type 45 or 46
in serviceable yields. Accordingly, we examined the
purposeful introduction of an oxygen atom into our
bicyclic substrate prior to attempted intramolecular
Knoevenagel condensation. Toward this end, we revisited
â-ketoester 41 and studied its reaction with m-CPBA
(
Scheme 20). In the event, despite the presence of a
potentially competing â-ketoester linkage in this mol-
ecule, epoxidation of the future C -C14 bond occurred to
a
Key: (a) Dess-Martin periodinane, CH2Cl2, rt, 2 h, 83%; (b)
1
ethyl diazoacetate (N2CHCOOEt), SnCl2, CH2Cl2, rt, 3.5 h; (c)
TsOH, H2O in acetone (5%), 70 °C, 90 min, 80% over two steps.
give rise to 53. The stereochemistry of the epoxide linkage
was formulated as R. This assignment was validated on
the basis of subsequent events. In particular, treatment
of this compound with sodium ethoxide in ethanol at
room temperature led to the formation of the 14-hydroxy-
lated bicyclic structure 54. Indeed, when 54 was sub-
jected to the action of sodium ethoxide/ethanol at 50 °C,
a 74% yield of the much desired compound 55 was
obtained. By the same token, treatment of compound 53
under the more forcing sodium ethoxide/ethanol, 50 °C
conditions also led to compound 55 in acceptable yield.
these conditions, the tetradeuterated tricyclic system 51
was detected by mass spectrometric analysis. Further-
more, when this compound was exposed to air, it gave
rise to an ꢀ-hydroxylated trideuterated compound, as-
signed as 52. Thus, under conditions of rigorous oxygen
exclusion, a non-oxidative adduct can apparently be
produced. A model for late-stage ω-oxidation had pre-
sented itself, though by happenstance.
10626 J. Org. Chem., Vol. 70, No. 26, 2005

Total Synthesis of Guanacastepene A
SCHEME 16. Attempted Intramolecular Knoevenagel Condensation of 41a
a
Key: (a) NaOEt, EtOH, 60 °C, 19 h.
SCHEME 17. Control Study: Attempted
Retrograde Claisen Condensation of 47
5
formulated as 58, in which inversion at C would be
necessary to reach the required (5â) guanacastepene
stereochemistry. The notion was that the principal
trajectory of hydride delivery by a hindered reducing
agent had occurred from the â-face, thereby leading to
the R-alcohol at C
supposition that hydride delivery would occur from the
â-face anti to the angular methyl at C . In conformational
analysis terms, the guanacastepene stereochemistry at
would correspond to formation of an equatorial alcohol
5
. Our formulation was based on the
8
At last, a serviceable route to the construction of the
tricyclic core structure had been realized. The key was
C
5
38
by means of axial delivery. With this bias influencing
our tentative assignments, the major compound was
subjected to Mitsunobu inversion.39 Compound 58 was
converted to its dibenzoate, 59, as shown. Hydrolysis of
the dibenzoate, in turn, produced the original minor
product, 57. Each diol, 57 and 58, could be converted to
its cyclic acetonide derivative, formulated as 60 and 61,
respectively. It was possible to sustain the cyclic ac-
etonide protecting group through the protocol which
accomplished desilylation at C14. The resultant product,
1
to oxidize (via epoxidation) the C -C14 olefin prior to
Knoevenagel cyclization.
Our next area of focus would be that of dealing with
the novel R,â-unsaturated aldehyde embracing carbons
3
, 4, and 15, as well as the allylic alcohol encompassing
carbons 3, 4, and 5. It seemed prudent to protect the C14
alcohol before generating hydroxy functionality at C or
15. Accordingly, compound 55 was subjected to the
5
C
action of triethylsilyl triflate under the conditions shown,
thereby affording 56 (Scheme 21).
30
when subjected to Dess-Martin oxidation, gave rise to
In planning concurrent reduction of both the ester and
keto functions, we were concerned that the ylidene
â-ketoester function would be subject to facile conjugate
5
the C14 ketone in each of the two epimeric C acetonides
(62 and 63).
At this point, we faced the issue of introduction of the
C13 acetoxy function. We first studied the problem in the
3
g
reduction. In actuality, this was not a significant
problem. Treatment of 56 with DIBAL-H led to smooth
reduction of the ester and ketonic functions with pres-
ervation of the diene system. A separable mixture of two
diol products was obtained, reflecting diastereomers at
major diol series, which we construed to have the 5-epi-
guanacastepene stereochemistry, in the context of its
acetonide (63). Subsequently, we undertook to study the
5
acetoxylation of the ketone in the minor (C -â) series (62),
3
4
C
5
(vide infra). The use of other reducing agents was
which now became readily available through the Mit-
sunobu protocols described above. The thought was that
as we came closer to reaching guanacastepene itself, it
would be possible to rigorously distinguish between the
two late-stage diastereomers, 57 and 58.
explored in an attempt to render the reaction more
stereospecific. However, recourse to other reducing agents,
36
such as LiAlH
4
, Li(OtBu)
3
AlH,35 9-BBN, and NaBH
4
/
3
7
CeCl
3
,
in fact led to reduced stereoselectivities.
We theorized that the minor reduction product corre-
sponds to structure 57, bearing the required â-configu-
5
ration at C . The major compound was correspondingly
Our initial approach to the installation of the C13
acetoxy function involved formation of a C13-C14 enol
derivative (cf. 64, Scheme 22).40 We anticipated that
epoxidation of the enol would occur from the R-face, anti
to the resident C12-isopropyl and C11-methyl functional-
ities. Two stereochemically divergent pathways can then
(34) For a discussion of the general preference for axial hydride
delivery to cyclohexenones, with leading references, see: Davis, A. P.
In Methods in Organic Chemistry (Houben-Weyl) Stereoselective Syn-
thesis, Workbench Edition E21; Helmchen, G., Hoffmann, R. W.,
Mulzer, J., Schaumann, E., Eds.; Georg Thieme Verlag: New York,
(38) Wu, Y.-D.; Houk, K N.; Trost, B. M. J. Am. Chem. Soc. 1987,
1
996; Vol. 7, 4034.
35) For a review of reductions using lithium trialkoxyaluminum
109, 5560.
(
(39) For a review of the Mitsunobu reaction, see: Hughes, D. L. Org.
hydrides, see: M a´ lek, J. Org. React. 1985, 34, 1.
React. 1992, 42, 335.
(
36) Krishnamurthy, S.; Brown, H. C. J. Org. Chem. 1977, 42, 1197.
37) Gemal, A. L.; Luche, J.-L. J. Am. Chem. Soc. 1981, 103, 5454.
(40) Brevet, J.-L.; Fournet, G.; Gore, J. Synth. Commun. 1996, 26,
(
4185.
J. Org. Chem, Vol. 70, No. 26, 2005 10627

Mandal et al.
SCHEME 18. Proposed Pathway of Formation of 44
SCHEME 19. Mechanistic Studies on the
We now digress briefly to explain our paradigm for
analysis of the stereochemistry of the C13-acetoxylated
compounds. In the precursor ketone, 63, the C13 geminal
protons are coupled to the C12 R-proton with distinct
coupling constants (13.2 and 7.6 Hz). Based on analogy
to guanacastepene A (J(13-HR,12-HR) ≈ 7 Hz), we could
a
Serendipitous Oxidation
tentatively assign the 13.2 Hz coupling constant to 13H -
â
1
R R R
2H and the 7.6 Hz coupling constant to 13H -12H .
This assignment was supported by observed NOE inter-
actions. In this way, we looked forward to assigning the
stereochemistry of the C13-acetoxylated products 69 and
a
Key: (a) NaOEt-d5, EtOH-d6, 60 °C, 17 h; (b) exposure to air,
rt, <4 h.
7
0 by evaluating the J(13-H,12-HR) coupling constants in
SCHEME 20. Synthesis of Oxidatively
Functionalized Gunacastane Skeleton 55a
the context discussed above.
In the event, ketone 63 was converted to epoxide 68
as shown (Scheme 23). The stereochemistry of the ep-
oxide was not determined at this stage. The epoxide was
then subjected to acidic conditions, which should favor
the solvolysis pathway, with retention of the epoxide
stereochemistry. As discussed above, we had expected
this pathway to yield R-acetoxy ketone (cf. 70), and were
thus surprised to find 69 as the major product. As
discussed above, the assignment of the stereochemistry
of 69 at C13 was based on an evaluation of the vicinal
1
3-H/12-H coupling constant (6.7 Hz), which correlates
to the coupling constant observed for the natural product
7.5 Hz), in the 13H -12H
series.1a
By contrast, when 68 was subjected to pyrolysis
(
R
R
conditions, expected to result in overall inversion of
stereochemistry, a ca. 1:1 mixture of 69 and 70 was
observed. Again, we turned to an analysis of the 13-H/
a
Key: (a) m-CPBA, CH2Cl2, 0 °C, 2 h, 89%; (b) NaOEt, EtOH,
rt, 30 min, 82%; (c) NaOEt, EtOH, 50 °C, 6 h, 74%; (d) NaOEt,
EtOH, 50 °C, 6 h, 80%.
12-H coupling constant to assign the C13 stereochemistry
for compound 70. As expected for a structure with a trans
relationship between 13-H and 12-H, the observed
be envisioned from epoxide 65. Under a solvolytic path-
way (path a), the C14-O epoxide bond is cleaved to afford
R-hydroxyketone 66, with retention of stereochemistry
at C13. By contrast, a thermolytic pathway (path b)
triggering strictly intramolecular bond reorganization
events would entail transfer of the intact acetoxy function
from C14 to C13, resulting in an overall inversion of
J
(13-Hâ,12-HR) was 12.8 Hz. The product obtained from the
solvolytic pathway led us to conclude that epoxidation
(
cf. 63 f 68) must have occurred from the seemingly
more hindered â-face of the enol acetate. However, the
surprising lack of product stereoselectivity resulting from
4
1
(41) For acid catalyzed and thermal rearrangements of enol ether
epoxides, cf. inter alia: (a) Zhu, Y.; Shu, L.; Tu, Y.; Shi, Y. J. Org.
Chem. 2001, 66, 1818. (b) Williamson, K. L.; Johnson, W. S. J. Am.
Chem. Soc. 1961, 83, 4563. (c) Soloway, A. H.; Considine, W. J.;
Fukushima, D. K.; Gallagher, T. F. J. Am. Chem. Soc. 1954, 76, 2941.
stereochemistry at C13
.
Given our initial expectation
that epoxidation would occur from the seemingly less
congested R-face, we anticipated that the thermolytic
pathway would be the one which would lead to the
guanacastepene (C13-â) stereochemistry from the pre-
sumed 65.
(
d) Heathcock, C. H.; Smith, S. C. J. Org. Chem. 1994, 59, 6828. (e)
Hern a´ ndez, R.; Vel a´ zquez, S. M.; Su a´ rez, E. J. Org. Chem. 1994, 59,
6395.
10628 J. Org. Chem., Vol. 70, No. 26, 2005

Total Synthesis of Guanacastepene A
SCHEME 21. Preparation of the Diastereomeric Keto-Acetonides 62 and 63a
a
Key: (a) Et3SiOTf, pyridine, CH2Cl2, 0 °C, 80-85%; (b) DIBAL-H, CH2Cl2, -78 to 0 °C (58/57 80:20); (c) 2,2-dimethoxypropane,
PPTS, CH2Cl2, 0 °C, 67% (from 56); (d) TBAF, THF, 0 °C, 91-98%, then Dess-Martin periodinane, pyridine, CH2Cl2, 90%; (e) 2,2-
dimethoxypropane, PPTS, CH2Cl2, 0 °C, 86% (from 56); (f) HF‚pyridine, pyridine, THF, then Dess-Martin periodinane, CH2Cl2, 77-
8
5%; (g) PPh3, benzoic acid, DIAD, THF, -78 °C to rt; (h) DIBAL-H, CH2Cl2, -78 to 0 °C.
SCHEME 22. Proposed Oxidation of C13 via Solvolytic and Thermolytic Pathways
the thermolytic pathway in any case merited further
scrutiny.
We sought to examine the origins of the nonhomoge-
These results serve to illuminate the course of the
thermolytic reaction pathway. Normally, one would
expect this reaction to produce acetoxyketone 70, with
inversion of stereochemistry from epoxide 68-â. In this
case, the results indicate the existence of a competing
neity of the thermolysis pathway. Upon subjection of 68
(
ratio of â- to R-epoxide ≈ 90:10) to thermolytic conditions
C
14-oxido heterolytic pathway, perhaps a consequence of
(150 °C for 18 min), a ca. 2:1 mixture of 70 and 69 was
3e
the allylic relationship of the epoxide to the C -C olefin.
observed (Scheme 24). In addition to these two acetox-
yketone diastereomers, a hydroxyketone intermediate,
1
2
(
42) The thermolysis study described above for the 5-R series has
7
2, was identified. This intermediate was very predomi-
been separately performed on the equivalent substrate in the 5-â series
(
see 62), which corresponds to the guanacastepene stereochemistry.
nantly of the 13-â configuration (ca. 40:1). Acetylation of
In studies with this substrate, there appears to be a preference for
formation of the 13-â acetoxy product (74) over the 13-R acetoxy product
(13-epi-74). The ratio of 74 to 13-epi-74 was approximately 2.3: 1.
7
2 and combination of this product with the acetoxyke-
4
2
tones afforded an overall 1:1 mixture of 70 and 69.
J. Org. Chem, Vol. 70, No. 26, 2005 10629

Mandal et al.
SCHEME 23. Studies on the Stereoselective Oxidation of 63a
a
Key: (a) Et3N, DMAP, AcCl, Ac2O, 100 °C, 90%; (b) DMDO/acetone, CH2Cl2, -78 to 0 °C; (c) p-TsOH, MeNO2, then Ac2O, pyridine,
DMAP, 60% from 63; (d) 150 °C (neat), then Ac2O, pyridine, DMAP, 65% from 63 (69/70 ca. 1:1).
SCHEME 24. Investigations on the Lack of Stereoselectivity in the Thermolysis of 68a
a
Key: (a) DMDO/acetone, CH2Cl2, 150 °C, then Me2S, 80%; (b) Ac2O, DMAP, pyridine, CH2Cl2, rt, 80%.
Thus, compound 68 could undergo heterolytic cleavage
to generate intermediate 71, which could be converted
to either acetoxyketone or hydroxyketone. In either event,
product generated under this pathway would be formed
with retention of the epoxide stereochemistry, thus
eroding the overall selectivity for the inversion product,
and the stereochemical nature of this late stage inter-
mediate was confirmed as corresponding to that of
guanacastepene A (Figure 4).44
We digress briefly to note that, although we had
inferred the stereochemistry of the intermediate epoxide
to be â, based on its conversion to the acetoxyketone 74,
an alternate mechanism has been proposed by Magnus
and Ollivier.45 According to the Magnus hypothesis,
epoxidation may occur from the R-face of the silyl enol
70.
Having thus identified the origin of the lack of stereo-
selectivity in the thermolysis route, we returned to the
task of installing the guanacastepene C13 â-acetoxy
functionality. Clearly, the solvolytic pathway described
above would be a viable means of accomplishing this goal.
However, given that the intermediate epoxide formed,
albeit surprisingly was actually of the â-configuration,
we opted to install the acetoxy group via the more
straightforward Rubottom oxidation protocol.43 Thus,
conversion of compound 62 to its silyl enol ether, followed
by exposure to DMDO, gave rise to hydroxyketone 73 (ca.
(43) (a) Rubottom, G. M.; Vazquez, M. A.; Pelegrina, D. R. Tetra-
hedron Lett. 1974, 4319. (b) Danishefsky, S. J.; Masters, J. J.; Young,
W. B.; Link, J. T.; Snyder, L. B.; Magee, T. V.; Jung, D. K.; Isaacs, R.
C. A.; Bornmann, W. G.; Alaimo, C. A.; Coburn, C. A.; DiGrandi, M. J.
J. Am. Chem. Soc. 1996, 118, 2843. (c) Rubottom, G. M.; Gruber, J.
M.; Boeckman, R. K., Jr.; Ramaiah, M.; Medwid, J. B. Tetrahedron
Lett. 1978, 4603.
(
44) The X-ray shows that, while the first carbon of the isopropyl
group is, indeed, â, the geminal dimethyls are R-disposed. Therefore,
in addition to confirming the stereochemical assignment described
above, the X-ray also rationalizes why attack of oxidizing agents occurs
primarily from the â-face of the molecule. We and others continue to
investigate this matter.
1
7
0:1 â:R), which was subsequently acetylated to afford
4 (Scheme 25). An X-ray structure was obtained for 73,
10630 J. Org. Chem., Vol. 70, No. 26, 2005

Total Synthesis of Guanacastepene A
SCHEME 25. Completion of the Synthesis of Guanacastepene Aa
a
Key: (a) Et3SiOTf, Et3N, CH2Cl2; (b) DMDO/acetone, CH2Cl2, -78 °C, then Me2S, 82-90% over two steps; (c) Ac2O, pyridine, DMAP,
CH2Cl2, 96%; (d) PPTS, MeOH, 70 °C; (e) PhI(OAc)2, TEMPO, CH2Cl2, 59-65% over two steps.
results, it is highly unlikely that an inversion pathway,
such as that suggested by Magnus, is responsible for the
observed C13 stereochemistry. The reader will note that
this mechanistic evaluation was performed in the C
series. As indicated above, our studies have demonstrated
that the nature of the remote C stereocenter does not
impact the stereochemistry of epoxidation at C13
5
-R
5
.
We return to the completion of the synthesis of gua-
nacastepene A. Hydrolysis of the acetonide of 74 followed
by TEMPO-catalyzed oxidation46 of the resultant primary
alcohol afforded guanacastepene A (1), whose spectral
1
13
data ( H NMR at 25 °C and -50 °C, C NMR at 25 °C
and -50 °C, IR, and mass spectra) were found to be in
complete accord with those obtained for the natural
material (Scheme 25).4
7,48
Due to the instability of gua-
FIGURE 4. X-ray structure of 73.
nacastepene, it would not be possible to obtain a sample
of the natural product. Given the crystallographic deter-
mination on compound 74, we could be very confident of
the correctness of our assignment.
SCHEME 26. Proposed Alternate Mechanism for
a
Formation of 73
Indeed, prior to obtaining verification from the crystal-
lographic determination, we had prepared and evaluated
the remaining three diastereomers of 1, with respect to
5
C and C13, to demonstrate that the guanacastepene
1
epimers can be differentiated by H NMR. Indeed, the
epimeric C13-acetoxy functionalities were shown to ex-
hibit markedly different coupling constants. In addition,
5
the epimeric C -hydroxy groups exhibited differential
effects on the peak shape of the C15 aldehyde. This
evidence is thus overwhelming, in the absence of a direct
comparison of samples, that the total synthesis of racemic
guanacastepene had, indeed, been accomplished.
We now sought to adapt our program to allow us to
gain access to fully synthetic optically active natural
product. In this context, we identified intermediate 20
from our racemic route as a key target compound for our
modified asymmetric synthesis. With enantioenriched 20
in hand, the route to optically active guanacastepene A
a
Key: (a) DMDO/acetone, CH2Cl2, -78 °C, then Me2S.
ether 75, affording 76, which could plausibly rearrange
to the C13-â-hydroxyketone, 73, as shown (Scheme 26).
To test the applicability of the proposed Magnus
(
46) De Mico, A.; Margarita, R.; Parlanti, L.; Vescovi, A.; Piancatelli,
G. J. Org. Chem. 1997, 62, 6974.
(47) An authentic sample is no longer available for comparison (J.
Clardy, personal communication). In fact, at the present time chemical
synthesis is the only source of this compound.
mechanism to our own case, we prepared the C13
-
dideuterated compound 79, which was then enolized to
afford deuterated silyl enol ether 80 (C13-D/C13-H ) 85:
1
6
(48) H NMR for synthetic guanacastepene at 25 °C (acetone-d , 400
MHz): δ 9.91 (br s, 1H), 7.45 (d, J ) 1.1 Hz, 1H), 5.48 (d, J ) 6.5 Hz,
1
1
H), 4.62 (m, 1H), 3.97 (br s, 1H, OH), 2.08 (s, 3H), 1.99 (m), 1.90 (m),
1
5) (Scheme 27). The latter was subjected to the Rubot-
.79 (m), 1.63 (m), 1.40 (m), 1.27 (s, 3H), 1.12 (d, J ) 6.6 Hz, 3H), 1.09
tom conditions described above, followed by acetoxylation.
The ratio of deuterated acetoxy compound 81 to nondeu-
terated 69 was again 85:15, thus confirming that signifi-
cant deprotonation at C13 had not occurred. Given these
(s, 3H), 0.93 (d, J ) 6.4 Hz, 3H).
1
H NMR for synthetic guanacastepene
6
at -50 °C (acetone-d , 500 MHz, key signals): major δ 9.96 (s), 7.42
(
s), 5.45 (d, J ) 5.6 Hz), 4.64 (m), 4.59 (d, J ) 5.4 Hz) (OH), 2.10 (s);
minor δ 9.72 (s), 7.49 (s), 5.53 (d, J ) 7.1 Hz), 4.52 (m), 4.48 (d, J )
4.1 Hz, OH), 2.11 (s). These data match the data obtained directly from
1
the H NMR spectra of the natural product. We are grateful to Prof.
Jon Clardy and Dr. Sean F. Brady for providing detailed NMR spectra
of natural guanacastepene A.
(45) Magnus, P.; Ollivier, C. Tetrahedron Lett. 2002, 43, 9605.
J. Org. Chem, Vol. 70, No. 26, 2005 10631

Mandal et al.
SCHEME 27. Evaluation of the Applicability of the Magnus Mechanism to Guanacastepene Aa
a
Key: (a) Et3N, CD3OD, rt, 20 h, 90%; (b) DBU, CD3OD, rt, 4 h, 95%; (c) Et3SiOTf, Et3N, CH2Cl2, 79%; (d) DMDO/acetone, CH2Cl2,
78 °C, then Me2S; (e) Ac2O, pyridine, DMAP, CH2Cl2, ca. 60%.
-
SCHEME 28. Proposed Homo-Robinson
Annulation Strategy
TABLE 1. Successful Ring Expansion of Robinson
Annulated Substrates
would converge with our established synthesis in the
racemic series. In this context, we revisited a general
strategy that had been considered early in the project:
that of using a homo-Robinson annulation to assemble a
structure such as 20. Our early efforts to assemble the
B,C-system through a Robinson annulation strategy had
been unsuccessful due to the preferential formation of a
five-membered ring in lieu of the requisite seven-
membered ring (see Scheme 3). We were hopeful that an
alternate approach might allow us to circumvent this
problem.
a
Key: (a) (i-Pr)2NH (2.5 equiv), n-BuLi (2.8 equiv), TMSCl (2.0
equiv), THF, -78 °C; (b) Et2Zn (1.5 equiv), CH2I2 (1.1 equiv), Et2O,
°C; (c) FeCl3 (2.5 equiv), 0 °C; (d) NaOAc, reflux.
0
and 8951 were prepared from the Hajos-Parrish and
52
53
Weiland-Meischer ketones, respectively. Known com-
pound 91 was prepared by a direct Robinson annulation-
annular decarboxylation protocol.54 Importantly, the
route from cyclohexenone (cf. 82) to cycloheptadienone
(
cf. 86) can be conducted without purification of any of
In a revised early phase strategy, we would draw from
the intermediates. The yields reported below are overall
yields from the four-step procedure.
49
the previously reported work of Saegusa, involving
oxidative cleavage of silyloxycyclopropanes to generate
ring-expanded R,â-unsaturated ketones. We proposed to
extend the Saegusa teaching to produce ring-expanded
cross-conjugated dienones. Thus, our proposed route
would commence with a standard Robinson annulation
to generate a fused cyclohexenone of the type 82 (Scheme
8). The latter would be converted to the cyclopropane
4 via the cross-conjugated silyloxydiene 83. On the basis
of both independent chemical reasoning and precedent,
Having established this homo-Robinson equivalent
protocol as a viable means to generate fused cyclohep-
tadienone structures, we returned to our goal of reaching
bicyclic compound 20. Our program commenced with silyl
enol ether 7 (Scheme 29). Michael addition of 7 to methyl
55
vinyl ketone, under Mukayama conditions, gave rise to
2
8
intermediate 93, which, following aldol cyclization, yielded
the fused cyclohexenone intermediate 94. This compound
was subjected to the four-step ring-expansion sequence
3
we anticipated that treatment of 84 with FeCl would
induce oxidative cleavage of the interior cyclopropane
bond, giving rise to chloroketone 85. Following â-elimina-
tion, the cross-conjugated dienone 86 would be in hand.
As recently described,3h we first demonstrated the
cyclopropanation route to homo-Robinson annulation
(
50) Caine, D.; Kotian, P. L.; McGuiness, M. D. J. Org. Chem. 1991,
5
6, 6307.
(51) Kim, M.; Kawada, K.; Watt, D. S. Synth. Commun. 1989, 19,
2017.
(
52) Micheli, R. A.; Hajos, Z. G.; Cohen, N.; Parrish, D. R.; Portland,
L. A.; Sciamanna, W.; Scott, M. A.; Wehrli, P. A. J. Org. Chem. 1975,
40, 675.
products in the context of racemates. Several examples
are summarized below (Table 1). Starting materials 8750
(53) Wieland, P.; Miescher, K. Helv. Chim. Acta 1950, 33, 2215.
(
54) Ozaki, Y.; Kubo, A.; Kim, S. W. Chem. Lett. 1993, 6, 993.
55) Duhamel, P.; Dujardin, G.; Hennequin, L.; Poirier, J. M. J.
(
(49) Ito, Y.; Fujii, S.; Saegusa, T. J. Org. Chem. 1976, 41, 2073.
Chem. Soc., Perkin Trans. 1 1992, 387.
10632 J. Org. Chem., Vol. 70, No. 26, 2005

Total Synthesis of Guanacastepene A
SCHEME 29. Application of Homo-Robinson Annulation Strategy to Guanacastepene Aa
a
Key: (a) MVK, AcOH, BF3‚Et2O, -20 °C, 97%; (b) NaOMe, 98%; (c) (i-Pr)2NH, n-BuLi, TMSCl, THF, -78 °C; (d) Et2Zn, CH2I2, Et2O,
0
°C; (e) FeCl3, 0 °C; (f) NaOAc, reflux, 40% yield for four steps; (g) Wilkinson’s catalyst, H2, 83%.
SCHEME 30. Formal Enantioselective Total Synthesis of Guanacastepene Aa
a
Key: (a) CuSCN, isopropenyllithium, (S)-2-methoxymethylpyrrolidine, 4 Å MS, -100 °C, 96%, 90% ee; (b) TBSOTf, TEA, 90%; (c)
MVK, AcOH, BF3‚Et2O, -20 °C, 90%; (d) H2, Pd/C, 92%; (e) NaOMe, 98%; (f) (i-Pr)2NH, n-BuLi, TMSCl, THF, -78 °C; (g) Et2Zn, CH2I2,
Et2O, 0 °C; (h) FeCl3, 0 °C; (i) NaOAc, reflux, 40% yield over four steps; (j) Wilkinson’s catalyst, H2, 83%
as described above, to afford ultimately cycloheptadi-
Conclusions
enone 98 in 40% overall yield from 94. Finally, selective
Wilkinson reduction56 of 98 was easily executed, giving
In summary, the main goals of the guanacastepene A
project have been met. A total synthesis of the racemate
has been accomplished. A route to enter the appropriate
enantiomeric series has been fashioned and documented
for use. The route to guanacastepene involved a multi-
faceted travelogue, which led us along several fascinating
excursions. First, we would cite the issue of reductive
cyclization relative to reduction (see compounds 18 and
rise to 20 in racemic form.
At this point, there remained only the adaptation of
this route to allow for the enantioselective preparation
of key intermediate 20. To accomplish this goal, we made
use of an enantioselective conjugate addition described
by Quinkert and colleagues.57 Thus, as described by the
German workers, addition of isopropenyl cuprate to
methyl cyclopentenone 4 afforded 99 in 90% ee (Scheme
1
9) and the way in which deuterated compounds were
used both to influence yield distributions and to teach
mechanistic pathways. The solution to the control of
3
0). It is only appropriate to note that the Quinkert
protocol, however impressive in its outcome, does appar-
ently require the use of substantial excesses of chiral
additive. Thus, from the perspective of convenience, there
is certainly room for new technology to be brought to bear
on the problem. Nonetheless, we moved forward. Thus,
silyl enol ether formation on 99, followed by Michael
addition to methylvinyl ketone, gave rise to 100. The
isopropenyl group of the latter was hydrogenated to
afford highly enantioenriched 93. When subjected to the
conditions described above for the racemic protocol,
compound 93 was smoothly converted to optically en-
riched 20. This novel reaction protocol constitutes the
basis of an enantioselective formal synthesis of guana-
castepene A. Following the routes documented above in
the racemic series, compound 20 could be converted to
enantioenriched or possibly enantiopure guanacastepene
A (depending on full purification through crystalliza-
tion), though we have not in practice done so at this
writing.
8
stereochemistry at C was critical to the success of the
program. The sorting out of the issues of how to conduct
intramolecular Knoevenagel condensation (see precursor
5
3) turned into an unexpected challenge. The solution
to the problem of stereoselective acetoxylation at C13 was
solved only after recognizing that attack of the C12-C13
enol was occurring from the â-face (see 62 f 73). Finally,
it was the lure of guanacastepene which led us to study
means for accomplishing the equivalent of the homo-
3h
Robinson annulation. With the benefit of a valuable
57
enabling contribution from Quinkert, we could exploit
our chemistry to enable the enantiocontrolled synthesis
of guanacastepene itself. Inevitably, some of our findings
provoke questions that have not been fully answered.
Indeed, the journey to guanacastepene had carried with
it many interesting teachings and stimulating questions
in chemistry, well worth continuing attention.
Experimental Section
8
â-(4′-Carbethoxy-3′-ketobutyl)-8R,11â-dimethyl-12â-
(
56) Bhattacharyya, S.; Karpha, T. K.; Mukherjee, D. Synth. Com-
mun. 1989, 19, 673.
57) Quinkert, G.; M u¨ ller, T.; K o¨ niger, A.; Schultheis, O.; Sicken-
berger, B.; D u¨ rner, G. Tetrahedron Lett. 1992, 33, 3469.
isopropyl-1,14-oxiranylhydroazulen-3-one 53. In a 25 mL
conical flask, the â,ú-diketo ester 41 (96 mg, 265 µmol, 1.0
(
equiv) was dissolved in 5 mL of CH
2
Cl
2
and cooled to 0 °C.
J. Org. Chem, Vol. 70, No. 26, 2005 10633

Mandal et al.
m-CPBA (57-86%, 420 mg, 1.39-2.09 mmol, 5.3-8.0 equiv)
Knoevenagel Cyclization of γ-Hydroxyenone 54 to
Tricyclic Alcohol 55. A solution of freshly prepared sodium
ethoxide (0.1 M in ethanol, 0.26 mL, 0.026 mmol, 1.1 equiv)
was added to a solution of γ-hydroxyenone 54 (9.1 mg, 0.024
mmol, 1.0 equiv) in 0.74 mL of ethanol in a resealable vial.
The vial was closed with a Teflon-lined cap and heated at 50
°C for 6 h. The reaction mixture was cooled to room temper-
ature, diluted with 10 mL of ether, and washed with two 3-mL
portions of water and then with 3 mL of saturated sodium
chloride solution. The combined aqueous layers were extracted
with 5 mL of ether, and the combined organic layers concen-
trated at reduced pressure, diluted with 10 mL of ether,
washed with 3 mL of saturated sodium chloride solution, dried
was added in three portions over 1 h. After an additional 2 h,
5
2 2 3
mL of 1 M Na S O was added and the reaction removed
from the bath and allowed to warm to rt. The mixture was
partitioned between 15 mL of Et O and 5 mL of satd NaHCO
),
2
3
.
The organic phase was washed 1× with brine, dried (MgSO
4
filtered, and evaporated to yield the crude product. Purification
by silica flash chromatography (4:1 hexanes/EtOAc, then 2:1
hexanes/EtOAc) yielded the epoxide 53 as a colorless oil (89
mg, 89%). TLC: R
hexanes/EtOAc). IR (film): 2945, 1745 (CdO), 1716 (CdO),
f f
0.31 (2:1 hexanes/EtOAc), R 0.50 (1:1
1
9
698 (CdO), 1646, 1473, 1367, 1317, 1238, 1174, 1084, 1033,
1
42. H NMR (500 MHz, CDCl
3
, 81:19 keto/enol form): (keto
), 3.47 (AB d, 1H, J )
), 3.42 (AB d, 1H, J ) 15.5 Hz, C4-H ), 3.24
app s, 1H, C14-H), 3.23 (obsc d, 1H, J ) 15.8 Hz, C2-H ),
), 2.49 (ddd, 1H,
), 2.32 (d, 1H, J ) 15.8 Hz, C2-
), 1.95-1.78 (obsc
), 1.84 (obsc m, 1H, C7-H ), 1.75 (ddd,
H, J ) 14.0, 10.0, 5.5 Hz, C7-H ), 1.51 (m, 1H, C18-H), 1.43
obsc m, 1H, C9-H ), 1.39 (ddd, 1H, J ) 13.9, 10.6, 0.9 Hz,
C13-H ), 1.28 (obsc t, 3H, J ) 7.1 Hz, C26-H ), 1.26 (obsc s,
H, C16-H ), 1.20 (td, 1H, J ) 10.0, 7.0 Hz, C12-H), 0.92 (d,
H, J ) 6.5 Hz, C19-H ), 0.91 (s, 3H, C17-H ), 0.85 (d, 3H,
), (enol form, diagnostic peaks) δ 12.11 (s,
form) δ 4.19 (q, 2H, J ) 7.1 Hz, C25-H
5.5 Hz, C4-H
2
4
over MgSO , filtered, and concentrated at reduced pressure.
1
a
b
Purification by flash column chromatography on 4 g of silica
gel (elution with 30% ethyl acetate/hexanes) provided 6.4 mg
(74%) of tricyclic alcohol 55. The tricyclic alcohol (55) synthe-
sized in this way has the exact NMR characteristics of
compound 46. In this way, it confirms the stereochemistry of
the C14-hydroxy to be R.
(
a
2
.58 (ddd, 1H, J ) 17.3, 10.1, 5.5 Hz, C6-H
a
J ) 17.3, 10.0, 5.5 Hz, C6-H
b
b a
H ), 2.12 (dd, 1H, J ) 14.2, 6.9 Hz, C13-H
m, 3H, C3-H
1
(
a
, C4-H
2
a
b
4-Carbethoxy-8R,11â-dimethyl-12â-isopropyl-14R-tri-
1
,2 3,4
b
ethylsilyloxy-(∆, ∆ )-tricycle 56. A solution of alcohol 55
b
3
(66 mg, 0.18 mmol, 1.0 equiv) and pyridine (89 µL, 1.1 mmol,
3
3
3
6.0 equiv) in 9 mL of CH Cl in a 25-mL round-bottomed flask
2
2
3
3
was stirred at 0 °C under argon, and triethylsilyl triflate (0.13
mL, 0.55 mmol, 3.0 equiv) was added rapidly dropwise. After
5 min, the reaction mixture was diluted with 27 mL of ether,
washed successively with 9-mL portions of water, 1.0 M
aqueous HCl, and a 1:1 mixture of saturated sodium bicarbon-
J ) 6.6 Hz, C20-H
3
1
3
1
H, C5-OH), 4.99 (s, 1H, C4-H). C NMR (100 MHz,
CDCl
3
): (keto form) δ 212.6 (C3), 202.7 (C5), 167.2 (C15), 66.85
(
(
C1), 62.88 (C14), 61.32 (C25), 51.07 (C12), 49.90 (C8), 49.22
C4), 44.68 (C2), 44.18 (C11), 38.38 (C6), 34.62 (C7), 32.98 (C9),
4
ate and saturated sodium chloride solutions, dried over MgSO ,
32.24 (C10), 32.05 (C13), 29.02 (C18), 23.13 (C19), 22.66 (C20),
0.00 (C16), 14.60 (C17), 14.08 (C26), (enol form, additional
filtered, and concentrated at reduced pressure. The crude
material was purified by flash column chromatography on 9 g
of silica gel (gradient elution with 10-30% ethyl acetate/
hexanes) to provide 74 mg (85%) of 56 as a colorless oil. TLC:
2
nonoverlapping peaks, partially assigned) δ 178.7 (C4/5), 172.8
C4/5), 66.92, 62.93, 59.92 (C25), 51.13, 50.10, 44.85, 38.07,
(
3
4
2.38, 30.20, 20.33, 14.68, 14.24. ESI-MS m/z (rel int): (pos)
f
R 0.34 (8:2 hexanes/EtOAc). IR (film): 2956.2, 2874.7, 1735.2,
+
-
1
01.2 ([M + Na] , 100); (neg) 413.0 ([M + Cl] , 100), 376.9
3
1674.3, 1456.5, 1366.6, 1226.6. H NMR (400 MHz, CDCl ): δ
-
+
(
[M - H] , 40). HRMS (EI): m/z 401.2311 (M + Na ), calcd
for C22 378.2406.
â-(4′-Carbethoxy-3′-ketobutyl)-8R,11â-dimethyl-12â-
6.02 (d, 1H, J ) 1.6 Hz), 4.45 (br s, 1H), 4.16-4.31 (m, 2H),
2.50-2.60 (m, 2H), 1.85-2.13 (m, 2H), 1.59-1.85 (m, 8H), 1.27
34 5
H O
(
t, 3H, J ) 7.1 Hz), 1.23 (s, 3H), 0.90-1.00 (m, 15H), 0.89 (d,
8
1
,2
3H, J ) 6.6 Hz), 0.56-0.63 (m, 6H). HRMS (EI): m/z 497.3056
isopropyl-14R-hydroxy-(∆ )-hydroazulen-3-one 54. In a
+
(
M + Na ), calcd for C28
46 4
H O Si 474.3165.
2
5 mL conical flask, the epoxide 53 (54.7 mg, 145 µmol, 1.0
equiv) was azeotroped 1× from PhH, and then NaOEt (0.1 N
8R,11â-Dimethyl-5R-hydroxy-4-hydroxymethyl-12â-iso-
0
1,2 3,4
in anhyd EtOH, prepared from Na , 1.6 mL, 159 µmol, 1.1
propyl-14R-triethylsilyloxy-(∆,
tion of keto-ester 56 (79 mg, 0.166 mmol, 1.0 equiv) in 16.5
mL of CH Cl in a 50 mL round-bottomed flask under argon
was immersed deeply in a -78 °C-cooling bath. A solution of
DIBAL-H (1.0 M in CH Cl , 1.6 mL, 1.6 mmol, 10 equiv) was
∆ )-tricycle 58. A solu-
equiv) was added via syringe and the solution was stirred at
rt. After 15 min, the reaction mixture was poured into satd
2
2
aq NH
extracts were washed 1× with H
MgSO
4
Cl and extracted 3× with EtOAc. The combined organic
2
O and 1× with brine, dried
2
2
(
4
), filtered, and evaporated to yield the crude product
added slowly down the side of the flask, and the resulting
mixture was stirred for 30 min at -78 °C and then for 30 min
at 0 °C. The reaction mixture was quenched (carefully at first)
with ethyl acetate (10 mL). The resulting solution was washed
with 1:1 mixture of 1.0 M aqueous HCl/saturated sodium
chloride solution (2 × 10 mL) and then with 1:1 mixture of
saturated sodium bicarbonate/saturated sodium chloride solu-
as a clear oil (53.4 mg). Purification by silica flash chroma-
tography (2:1 hexanes/EtOAc) yielded the γ-hydroxyhydroa-
f
zulenone 54 as a clear oil (44.6 mg, 82%). TLC: R 0.23 (1:1
hexanes/EtOAc). IR (film): 3456 (br, O-H), 2961, 1742 (Cd
O), 1716 (CdO), 1652 (enone CdO), 1150, 1450, 1368, 1316,
1
1
233, 1178, 1095, 1032, 874. H NMR (500 MHz, CDCl
3
, 92:8
keto/enol form): (keto form) δ 5.88 (d, 1H, J ) 1.3 Hz, C2-
H), 4.51 (br s, 1H, C14-H), 4.20 (q, 2H, J ) 7.1 Hz, C25-H ),
), 2.61 (app ddd, 2H, J ) 9.3, 6.3, 3.1
4
tion. The organic layer was dried over MgSO , filtered, and
concentrated at reduced pressure to afford a mixture, which
was ca. 4:1 mixture (R:â) of C5 diastereomers. The diastere-
omers were separated carefully by preparative TLC (25% ethyl
acetate in hexanes) to obtain 0.039 g of major R-epimer in 54%
yield and minor â-epimer in 8% yield (6 mg). R-Epimer of 58.
2
3
.47 (app s, 2H, C4-H
2
Hz, C6-H
2
), 2.02 (obsc m, 1H, C9-H
R
), 1.98 (obsc m, 2H, C10-
),
, C12-H), 1.64
app sxt, 1H, J ) 6.7 Hz, C18-H), 1.50 (dd, 1H, J ) 13.2, 7.0
Hz, C9-H ), 1.28 (t, 3H, J ) 7.1 Hz, C26-H ), 1.17 (s, 3H,
C16-H ), 1.00 (d, 3H, J ) 6.7 Hz, C19-H ), 0.95 (s, 3H, C17-
), 0.92 (d, 3H, J ) 6.6 Hz, C20-H ), (enol form, diagnostic
peaks) δ 12.13 (s, 1H, C5-OH), 5.04 (s, 1H, C4-H). C NMR
H
1
(
2
), 1.89 (obsc m, 1H, C13-H
a
), 1.87 (obsc m, 1H, C7-H
, C14-OH, C7-H
a
.83-1.75 (m, 4H, C13-H
b
b
TLC: R
2934.1, 2874.3, 1456.0, 1239.3, 1049.9, 1013.4, 987.6. H NMR
(400 MHz, CDCl
f
0.25 (3:1 hexanes/EtOAc). IR (film): 3282.2, 2953.3,
1
â
3
3
): δ 6.01 (s, 1H), 4.31-4.44 (m, 3H), 4.11 (br
3
3
d, 1H, J ) 9.4 Hz), 2.29-2.33 (b, 2H), 1.51-2.00 (m, 9H), 1.36-
1.49 (m, 2H), 1.21-1.30 (m, 1H), 1.08 (s, 3H), 0.94-0.99 (m,
H
3
3
1
3
1
6
4
2H), 0.87 (d, 3H, J ) 6.6 Hz), 0.79 (s, 3H), 0.56-0.60 (m,
(
100 MHz, CDCl
C1), 123.8 (C2), 73.2 (C14), 61.3 (C25), 53.9 (C12), 49.7 (C8),
9.2 (C4), 48.6 (C11), 38.6 (C6), 36.5 (C13), 34.3 (C10), 32.5
C7), 32.3 (C9), 28.2 (C18), 24.1 (C16), 23.7 (C19), 22.1 (C20),
3
): δ 208.7 (C3), 202.9 (C5), 167.3 (C15), 166.8
+
H). ESI-MS m/z (rel int): (pos) 457.3 ([M + Na] , 100); (neg)
(
-
69.0 ([M + Cl] , 55).
4
(
8R,11â-Dimethyl-5â-benzoyloxy-4-benzoyloxymethyl-
1
,2 3,4
2
0.9 (C17), 14.1 (C26). ESI-MS m/z (rel int): (pos) 401.4 ([M
12â-isopropyl-14R-triethylsilyloxy-(∆,
∆
)-tricycle 59. A
+
-
-
+
Na] , 100); (neg) 413.0 ([M + Cl] , 65), 377.0 ([M - H] ,
mixture of diol 58 (40 mg, 0.092 mmol, 1.0 equiv), Ph P (120
mg, 0.46 mmol, 5.0 equiv), and benzoic acid (56 mg, 0.46 mmol,
3
1
00).
10634 J. Org. Chem., Vol. 70, No. 26, 2005

Total Synthesis of Guanacastepene A
5
.0 equiv) in 9 mL of THF in a 25-mL round-bottomed flask
was cooled at -78 °C under argon. DIAD (90 µL, 0.46 mmol,
.0 equiv) was added dropwise and the resulting yellow
120.8, 99.5, 77.2, 74.4, 67.7, 60.5, 54.6, 46.1, 38.2, 38.0, 37.6,
35.8, 34.1, 27.9, 26.4, 25.3, 25.2, 23.9, 23.6, 22.5, 17.2, 6.9, 5.1.
+
5
ESI-MS m/z (rel int): (pos) 497.3 ([M + Na] , 100). HRMS
+
solution stirred overnight with gradual warming to room
temperature. After 15 h, the reaction mixture was diluted with
(EI): m/z 497.3449 (M + Na ), calcd for C29
H
50
O
3
Si 474.3529.
8
R,11â-Dimethyl-12â-isopropyl-5R,15-isopropylidene-
2
7 mL of ether, washed with 9 mL of half-saturated sodium
bicarbonate and 9 mL of saturated sodium chloride, dried with
MgSO , filtered, and concentrated under reduced pressure.
1,2 3,4
dioxy-14R-triethylsilyloxy-(∆,
∆ )-tricycle 61. Compound
6
1 was prepared from 58 in 91% yield following the similar
4
procedure that has been used to prepare 60. TLC: R
0.92
2.3: 1 hexanes/ethyl acetate). IR (film): 2954, 2912, 2874,
f
Partial purification was accomplished by flash column chro-
matography on silica gel (elution with 5% ethyl acetate/
(
1
1
455, 1413, 1377, 1223, 1198, 1089, 1060. H NMR (400 MHz,
hexanes) to 55 mg (92% yield) of 59. TLC: R
f
0.82 (2:1 hexanes/
CDCl
3
): δ 5.65 (app s, 1H), 4.32-4.22 (m, 3H), 4.06 (d, J ) 14
EtOAc). IR (film): 2966.0, 1718.6, 1452.5, 1314.2, 1269.7,
Hz, 1H), 1.79-1.60 (m, 5H), 1.56-1.46 (m, 4H), 1.43-1.35 (m,
1
1
175.6, 1109.6, 1067.9, 1025.8, 968.4, 709.7. H NMR (500
MHz, CDCl ): δ 8.03-7.96 (m, 4H), 7.56-7.50 (m, 2H), 7.42-
.35 (m, 4H), 6.14 (app s, 1H), 5.81 (m, 1H), 4.96 (d, J ) 12. 3
Hz, 1H), 4.79 (d, J ) 12.4 Hz, 1H), 4.45 (m, 1H), 2.14-1.98
m, 2H), 1.95-1.45 (m, 10H), 1.13 (s, 3H), 1.02-0.92 (m, 18H),
5
0
H), 1.34-1.29 (m, 4H), 0.98 (s, 3H), 0.91-0.0.84 (m, 12H),
.80 (s, d, J ) 7 Hz, 3H), 0.76 (s, 3H), 0.55-0.0.48 (m, 6H).
3
7
+
ESI-MS m/z 497.3 [M + Na] , calcd for C29
H
50
O
3
Si 474.3.
Si
+
HRMS (EI): m/z 497.3425 (M + Na ), calcd for C29
74.3529.
R,11â-Dimethyl-12â-isopropyl-5â,15-isopropylidene-
50 3
H O
(
4
+
0
.65-0.59 (m, 6H). HRMS (EI): m/z 665.3652 (M + Na ), calcd
8
for C40 642.3741.
H
54
O
5
1,2 3,4
dioxy-14-keto-(∆,
silyl ether 60 (31 mg, ca. 0.065 mmol) in 0.75 mL of THF in a
5-mL round-bottomed flask was cooled at 0 °C under argon,
∆ )-tricycle 62. A solution of triethyl-
8
R,11â-Dimethyl-5â-hydroxy-4-hydroxymethyl-12â-iso-
1
,2 3,4
propyl-14R-triethylsilyloxy-(∆,
tion of dibenzoate 59 (55 mg, 0.085 mmol) in 9 mL of CH
∆
)-tricycle 57. A solu-
Cl
1
2
2
and HF-pyridine (80 µL) was added. The mixture was stirred
at 0 °C for 5 min and warmed to room temperature. After being
stirred at room temperature for 1 h, the reaction mixture was
quenched with TMSOMe (1 mL) and stirred for 30 min,
evaporated under reduced pressure, and purified by prepara-
tive TLC (20% ethyl acetate in hexanes) to afford TES-
f
in a 25-mL round-bottomed flask was cooled at -78 °C under
argon, and a solution of DIBAL-H (1.0 M in CH Cl , 0.85 mL,
.85 mmol, 0 equiv) was added dropwise down the sides of the
2
2
0
flask. The resulting colorless solution was stirred for 30 min
at -78 °C and 30 min at 0 °C and then quenched with 27 mL
of ethyl acetate (carefully at first). The reaction mixture was
washed with two 9-mL portions of a 1:1 mixture of 1.0 M
aqueous HCl/saturated sodium chloride solution and one 9-mL
portion of a 1:1 mixture of saturated sodium bicarbonate/
saturated sodium chloride solutions. The organic layer was
deprotected alcohol in 85% yield. TLC: R 0.2 (4:1 hexanes/
1
3
EtOAc). H NMR (500 MHz, CDCl ): δ 5.87 (s, 1H), 4.31-
4.41 (m, 2H), 4.26 (d, 1H, J ) 14.8 Hz), 4.11 (d, 1H, J ) 15.4
Hz), 2.25 (br t, 1H, J ) 13.7 Hz), 1.44-1.89 (m, 13H), 1.38 (s,
3H), 1.22-1.34 (m, 2H), 0.98-1.05 (m, 6H), 0.91 (d, 3H, J )
1
3
dried with MgSO
4
, filtered, and concentrated at reduced
6.6 Hz), 0.88 (s, 3H). C NMR (100 MHz, CDCl ): δ 158.5,
3
pressure. Partial purification by flash column chromatography
134.4, 133.8, 122.0, 100.0, 75.0, 68.0, 60.7, 55.4, 47.0, 38.5, 37.8,
on 4 g of silica gel (gradient elution with 30-50% ethyl acetate/
36.9, 36.0, 34.8, 28.5, 26.9, 25.6, 25.4, 24.3, 24.0, 22.9, 17.4.
+
hexanes) provided 35 mg of a colorless oil 57 in 67% yield.
HRMS (EI): m/z 383.2570 (M + Na ), calcd for C H O
2
3
36
3
1
TLC:
CDCl
R
f
0.24 (2:1 hexanes/EtOAc). H NMR (400 MHz,
360.2664. A solution of Dess-Martin periodinane (72 mg, 0.17
mmol, 1.3 equiv) and pyridine (41 µL, 0.51 mmol, 3.8 equiv)
in 7 mL of CH Cl was prepared under argon. At the same
3
): δ 5.93 (s, 1H, C2-H), 4.37 (br m, 2H, C5-H, C14-H),
4
.33 (d, 1H, J ) 12.4 Hz, C15-H), 4.21 (br d, 1H, J ) 12.0 Hz,
C15-H), 2.43 (br s, 1H), 2.14 (br s, 1H), 1.91 (m, 1H), 1.83-
.70 (m, 6H), 1.64 (m, 2H), 1.60 (hept, 1H, J ) 6.8 Hz, C18-
H), 1.39 (br m, 2H), 1.00 (s, 3H, C16/17-H ), 0.97 (t, 9H, J )
), 0.95 (d, 3H, J ) 6.0 Hz, C19/20-H ), 0.89 (s,
), 0.88 (d, 3H, J ) 6.5 Hz, C19/20-H ), 0.60 (q,
): δ 157.3, 142.0,
2
2
time, a solution of TES-deprotected alcohol (48 mg, 0.13 mmol,
1.0 equiv) in 7 mL of CH Cl was prepared in a 50-mL round-
1
2
2
3
bottomed flask under argon. 5.2 mL of the Dess-Martin
solution was added dropwise to the alcohol solution, and after
10 min, the remaining 1.8 mL was added. After an additional
10 min, the reaction mixture was purified directly by flash
column chromatography on 10 g of silica gel (gradient elution
with 10-20% ethyl acetate/hexanes) to provide 43 mg (90%)
7
3
6
1
3
.9 Hz, C28-H
3
3
H, C16/17-H
3
1
3
3
2 3
H, C27-H ). C NMR (100 MHz, CDCl
29.5, 121.0, 77.2, 73.8, 72.7, 71.3, 68.7, 63.1, 4.5, 38.0, 37.1,
6.9, 36.6, 32.7, 28.2, 27.3, 25.5, 23.8, 22.4, 6.9, 5.0. ESI-MS
+
m/z (rel int): (pos) 457.3 ([M + Na] , 100); (neg) 469.3 ([M +
of the keto-acetonide 62 as a colorless oil. TLC: R 0.36 (4:1
f
-
+
Cl] , 100). HRMS (EI): m/z 457.2985 (M + Na ), calcd for
Si 434.3216
R,11â-Dimethyl-12â-isopropyl-5â,15-isopropylidene-
hexanes/EtOAc). IR (film): 2937, 2870, 1715, 1620, 1453, 1378,
1
C
26
H
46
O
3
1221, 1175, 1149, 1022, 903, 866. H NMR (400 MHz, CDCl ):
3
8
δ 6.81 (s, 1H), 4.32-4.39 (m, 2H), 4.07 (dd, 1H, J ) 17.8, 1.8
Hz), 2.49 (dd, 1H, J ) 18.1, 7.5 Hz), 2.37 (br t, 1H, J ) 12
Hz), 2.10-2.19 (m, 1H), 1.82-1.98 (m, 2H), 1.56-1.82 (m, 7H),
1
,2 3,4
dioxy-14R-triethylsilyloxy-(∆,
mL round-bottom flask, the 14-TES-5â,15-diol 57 (ca. 35 mg,
.08 mmol, 1.0 equiv) was dissolved in 12 mL of CH Cl and
cooled to 0 °C. PPTS (7.6 mg) and DMP (0.196 mL, 1.6 mmol,
0 equiv) were added in sequence, and the reaction mixture
was stirred for 30 min at 0 °C before quenching with K CO
∆ )-tricycle 60. In a 100
1
3
.43 (s, 3H), 1.38 (s, 3H), 1.04 (d, 3H, J ) 6.6 Hz), 1.01 (s,
0
2
2
13
H), 0.95 (s, 3H), 0.93 (d, 3H, J ) 6.7 Hz). C NMR (100 MHz,
): δ 205.5, 146.4, 138.8, 133.1, 129.5, 100.1, 67.6, 60.1,
1.4, 46.0, 41.2, 38.2, 36.9, 36.4, 33.8, 28.5, 26.9, 25.0, 24.5,
4.2, 23.9, 22.3, 17.9. ESI-MS m/z (rel int): (pos) 381.1 ([M +
CDCl
3
2
5
2
2
3
.
The resulting solution was stirred for 10 min, filtered, con-
centrated under reduced pressure, and purified by silica gel
column chromatography (hex/EtOAc ) 10/1 followed by 4/1)
+
-
-
Na] , 100), (neg) 393.0 ([M + Cl] , 33), 357.2 ([M - H] , 100).
8R,11â-Dimethyl-12â-isopropyl-5R,15-isopropylidene-
dioxy-14-keto-(∆, ∆ )-tricycle 63. A solution of triethyl-
1
,2 3,4
to yield the 14-TES-5,15-acetonide 60 (31 mg, 81% yield, as
1
white solids. TLC: R
MHz, CDCl
Hz, C14-H/C5-H), 4.27 (d, 1H, J ) 4.4 Hz, C5-H/14-H), 4.26
f
0.74 (4:1 hexanes/EtOAc). H NMR (400
silyl ether 61 (31 mg, ca. 0.065 mmol) in 0.75 mL of THF in a
15-mL round-bottomed flask was cooled at 0 °C under argon,
and HF-pyridine (80 µL) was added. The mixture was stirred
at 0 °C for 5 min and warmed to room temperature. After being
stirred at room temperature for 1 h, the reaction mixture was
quenched with TMSOMe (1 mL) and stirred for 30 min,
evaporated under reduced pressure, purified by preparative
TLC (20% ethyl acetate in hexanes) to afford TES-deprotected
3
): δ 5.70 (s, 1H, C2-H), 4.35 (dd, 2H, J ) 8.4, 6.4
(
d, 1H, J ) 15.2 Hz, C15-H), 4.11 (d, 1H, J ) 15.3 Hz, C15-H),
2
.23 (br m, 1H), 1.84 (m, 1H), 1.81-1.69 (m, 4H), 1.61 (hept,
H, J ) 6.8 Hz, C18-H), 1.54 (dd, 2H, J ) 10.1, 3.1 Hz), 1.44
1
(
s, 3H, C40-H
3
), 1.41 (m, 1H), 1.37 (s, 3H, C40-H
H), 1.00 (s, 3H, C16/17-H ), 0.97 (d, 3H, J ) 6.5 Hz, C19/20-
), 0.95 (t, 9H, J ) 7.9 Hz, C28-H ), 0.87 (d, 3H, J ) 6.5 Hz,
), 0.84 (s, 3H, C16/17-H ), 0.58 (q, 6H, J ) 7.9 Hz,
). C NMR (100 MHz, CDCl ): δ 157.3, 134.8, 132.1,
3
), 1.26 (m,
2
H
3
3
3
alcohol in 86% yield. TLC: R
f
0.17 (4:1 hexanes/EtOAc); IR
C19/20-H
3
1
3
(film): 3440 (br), 2940, 2869.0, 1450, 1377, 1223, 1198, 1087,
3
1
C27-H
2
3
1025, 864. H NMR (400 MHz, CDCl
3
): δ 5.82 (br-s, 1H), 4.39-
J. Org. Chem, Vol. 70, No. 26, 2005 10635

Mandal et al.
0.004 mmol, 0.2 equiv) were added, followed by Ac O (18 µL,
0.19 mmol, 10 equiv). The reaction mixture was stirred for 3
h, diluted with 5 mL of toluene, and concentrated under
reduced pressure. Purification by column chromatography on
ca. 1 g of silica gel (gradient elution with 5-10% ethyl acetate/
hexanes) provided 7.9 mg (97%) of acetate 74 as an oil. TLC:
4
.27 (m, 3H), 4.09 (d, J ) 14.4 Hz, 1H), 1.87-1.71 (m, 6H),
.61-1.40 (m, 8H), 1.36-1.32 (m, 4H), 1.01 (s, 3H), 0.94 (d, J
2
1
)
6.5 Hz, 3H), 0.86 (d, J ) 6.0 Hz, 3H), 0.82 (s, 3H). ESI-MS
+
m/z 383.1 [M + Na] , calcd for C23
H
36
O
3
360.2. HRMS (EI):
36 3
H O 360.2664. A
+
m/z 383.2561 (M + Na ), calcd for C23
solution of Dess-Martin periodinane (72 mg, 0.17 mmol, 1.3
equiv) and pyridine (41 µL, 0.51 mmol, 3.8 equiv) in 7 mL of
CH
solution of TES-deprotected alcohol (48 mg, 0.13 mmol, 1.0
equiv) in 7 mL of CH Cl was prepared in a 50-mL round-
bottomed flask under argon. A 5.2 mL portion of the Dess-
Martin solution was added dropwise to the alcohol solution,
and after 10 min, the remaining 1.8 mL was added. After an
additional 10 min, the reaction mixture was purified directly
by flash column chromatography on 10 g of silica gel (gradient
elution with 10-20% ethyl acetate/hexanes) to provide 43 mg
R
f
4.6 (5% ethyl acetate in CH
1750.8, 1721.3, 1613.5, 1371.5, 1216.3, 1089.8, 1018.7. H NMR
(500 MHz, CDCl ): δ 6.90 (s, 1H), 5.37 (d, 1H, J ) 6.5 Hz),
2 2
Cl ). IR (film): 2938.0, 2872.9,
1
2 2
Cl was prepared under argon. At the same time, a
3
2
2
4.29 (m. 2H), 4.03 (d, 1H, J ) 16.5 Hz), 2.21 (dt, 1H, J ) 12
Hz, J ) 2 Hz), 2.03 (s, 3H), 2.0-1.8 (m, 3H), 1.68-1.48 (m,
5H), 1.37 (s, 3H), 1.30-1.25 (m, 4H), 1.13 (s, 3H), 1.02 (d, 3H,
13
3
J ) 7 Hz), 0.863 (m, 6H). C NMR (100 MHz, CDCl ): δ 199.7,
169.5, 145.1, 140.6, 133.5, 133.0, 100.1, 74.3, 67.6, 60.1, 55.2,
45.6, 38.0, 36.8, 36.3, 34.0, 27.0, 25.0, 24.8, 24.6, 23.7, 23.3,
+
23.0, 20.9, 18.8. HRMS (EI) m/z 439.2462 (M + Na ), calcd
(
90%) of the keto-acetonide 63 as a colorless oil. TLC: R
f
0.36
36 5
for C25H O 416.2563.
(
4:1 hexanes/EtOAc). IR (film): 2956, 2870, 1716, 1626, 1450,
Enol Acetate 64. In a 5 mL vial containing keto-acetonide
3 (17 mg, 0.047 mmol) in 1.6 mL of acetic anhydride was
1
1
(
377, 1266, 1174, 1092. H NMR (400 MHz, CDCl
br-s, 1H), 4.50 (br-d, J ) 15 Hz, 1H), 4.30 (m, 1H), 4.10 (d, J
16 Hz, 1H), 2.41 (dd, J ) 18.5 Hz, J ) 7.5 Hz, 1H), 2.09 (dd,
3
): δ 6.67
6
added freshly distilled triethylamine (0.825 mL) and then
DMAP (8.7 mg). The resulting solution was cooled to 0 °C, and
acetyl chloride (165 µL) was added slowly via syringe. The
reaction mixture was stirred for 30 min with slowly warmed
to room temperature and then was heated at 100 °C. After 12
h, the reaction mixture was cooled to room temperature and
filtered through a bed of Celite, and the filtrate was concen-
trated to yield the crude product, which was purified by
preparative TLC (using 20% ethyl acetate in hexanes) to afford
)
J ) 18 Hz, J ) 13 Hz, 1H), 1.90 (m, 1H), 1.83-1.73 (m, 3H),
1
1
.72-1.66 (m, 1H), 1.61-1.51 (m, 2H), 1.48-1.40 (m, 2H),
.38-1.33 (m, 4H), 1.30 (s, 3H), 0.977-0.92 (m, 9H), 0.86 (d,
1
3
J ) 6.5 Hz, 3H). C NMR (100 MHz, CDCl
1
3
): δ 206.2, 147.2,
41.4, 132.8, 127.7, 99.8, 67.4, 60.7, 51.6, 46.9, 41.3, 39.3, 37.6,
7.0, 34.7, 28.7, 25.2, 25.0, 24.3, 24.0, 22.5, 19.5. HRMS (EI):
3
+
m/z 381.2408 (M + Na ), calcd for C23
34 3
H O 358.2508.
8
R,11â-Dimethyl-13â-hydroxy-12â-isopropyl-5â,15-iso-
64 in 89% yield. TLC: R
f
0.4 (8:2 ) hexanes\ethyl acetate).
1
,2 3,4
propylidenedioxy-14-keto-(∆,
mL pear-shaped flask, the keto-acetonide 62 (8.5 mg, 0.024
∆ )-tricycle 73. In a 10
IR (film): 2937.3, 2869.8, 1769.6, 1453.6, 1371.8, 1223.0,
1
1199.0, 1091.8. H NMR (500 MHz, CDCl ): δ 5.83 (d, J ) 2.0
3
mmol, 1.0 equiv) was dissolved in 1 mL of CH
Et N (33 µL, 0.24 mmol, 10 equiv) and then Et
.12 mmol, 5 equiv) were added via syringe. The reaction
2
Cl
3
2
under argon.
Hz, 1H), 5.47 (s, 1H), 4.36 (d, J ) 14.5 Hz, 1H), 4.30 (m, 1H),
4.09 (d, J ) 15.5 Hz, 1H), 2.13 (s, 3H), 1.87-1.48 (m, 7H),
1.42 (m, 1H), 1.39 (s, 3H), 1.32 (m, 4H), 0.98 (m, 6H), 0.927
3
SiOTf (27 µL,
0
1
3
mixture was stirred for 10 min and then diluted with 5 mL of
ethyl acetate, washed with 1 mL of saturated sodium bicar-
(d, J ) 7 Hz, 3H), 0.90 (d, J ) 6 Hz, 3H). C NMR (100 MHz,
3
CDCl ): δ 168.4, 150.5, 147.9, 134.0, 133.6, 122.6, 112.0, 99.6,
bonate solution, dried over Na
under reduced pressure to give an oil. This oil was azeotropi-
cally dried with toluene, dissolved in 2.4 mL of CH Cl in a
0-mL pear-shaped flask, and cooled at -78 °C under argon.
2 4
SO , filtered, and concentrated
67.8, 60.9, 59.5, 46.9, 40.3, 38.1, 37.6, 34.6, 27.6, 26.1, 25.8,
+
25.4, 24.3, 23.5, 22.8, 21.5. ESI-MS: m/z 423.2 ([M + Na] ,
36 4
calcd for C25H O 400.26.
2
2
1
C13â-Acetate 69. To a solution of enol acetate 64 (2 mg,
A solution of DMDO (0.52 mL, 0.036 mmol, 1.5 equiv) was
added dropwise down the sides of the flask. The resulting
0.0048 mmol) in 0.5 mL of CH Cl was added slowly a solution
2
2
of DMDO in acetone (0.07 M in acetone, 68 µL, 1 equiv) at -
50 °C, and the resulting solution was stirred for 40 min at -
50 °C and then 20 min at 0 °C. Dimethyl sulfide (1 µL) was
added, and the resulting mixture was warmed to room
temperature and evaporated to dryness. This crude epoxide
2
solution was stirred for 15 min at -78 °C, and then Me S (27
µL, 0.37 mmol, 15 equiv) was added to quench excess DMDO.
The quenched reaction solution was stirred for 5 min at -78
°C, and then the cooling bath was removed and after an
additional 15 min the solution was concentrated under reduced
pressure. Purification by flash column chromatography (gradi-
ent elution with 5-20% ethyl acetate/hexanes) provided 7.3
mg (82%) of hydroxy-ketone 73 as a colorless oil. In a separate
run, 13 mg of 73 was dissolved in 0.34 mL of ethyl acetate
was dissolved in MeNO (0.5 mL), and p-TSA (10 mol %) was
2
added. The resulting solution was stirred for 20 min at rt.
Saturated NaHCO (1 mL) was added and the reaction mixture
3
diluted with water and extracted with EtOAc. The organic
layers were dried with MgSO and concentrated under reduced
4
(
0.1 M solution) in a small (4-mL) vial, which was left
pressure to afford the crude product. To this crude product in
uncapped and placed in a larger vial containing an excess
amount of hexanes. The larger vial was sealed, and the
hydroxy-ketone (73) was allowed to crystallize by vapor-
diffusion over the course of 2 days to provide white needles
2 2
CH Cl (0.5 mL) were added a catalytic amount of DMAP, 80
µL of pyridine, and 40 µL of acetic anhydride. The resulting
solution was stirred for 3 h (or until the reaction was complete).
The reaction mixture was evaporated to dryness and purifica-
tion using preparative TLC (using 3% ethyl acetate/chloroform)
(
mp 184.6-186.4 °C). TLC: R
f
0.28 (4:1 hexanes/EtOAc). IR
(
film): 3422.0, 2938.8, 1710.1, 1614.2, 1455.2, 1379.5, 1223.4,
afforded 69 as the major isomer in 60% yield. TLC: R
f
0.23
1
1
3
088.7. H NMR (400 MHz, CDCl ): δ 6.97 (s, 1H), 4.32-4.40
(8:2) hexanes\ethyl acetate). IR (film): 2938.1, 1750.1, 1718.6,
1
(
m, 2H), 4.05-4.12 (m, 2H), 2.12-2.34 (m, 3H), 1.95-2.06 (m,
3
1613.3, 1372.1, 1216.2, 1093.7. H NMR (500 MHz, CDCl ): δ
1
1
1
H), 1.83-1.95 (m, 1H), 1.51-1.75 (m, 4H), 1.45 (s, 3H), 1.30-
6.8, 5.41 (d, 1H, J ) 6.5 Hz), 4.49 (d, 1H, J ) 16.5 Hz), 4.31
(m, 1H), 4.14 (d, 1H, J ) 16.5 Hz), 2.03 (s, 3H), 1.91 (m, 1H),
1.84-1.71 (m, 3H), 1.63-1.44 (m, 6H), 1.37 (s, 3H), 1.31 (s,
3H), 1.14 (s, 3H), 1.02 (d, 3H, J ) 7.5 Hz), 0.93 (s, 3H), 0.86
.43 (m, 1H), 1.36 (s, 3H, partially obscured), 1.22 (s, 3H),
.07-1.13 (m, 6H), 0.94 (s, 3H). 13C NMR (100 MHz, CDCl
3
):
δ 204.9,145.7, 140.6, 133.3, 117.8, 100.4, 74.7, 67.8, 60.3, 56.3,
+
4
2
3
5.9, 38.3, 36.9, 36.5, 34.4, 27.2, 25.2, 24.8, 24.6, 24.2, 23.7,
(d, 3H, J ) 7.0 Hz). ESI-MS: m/z 439.3 ([M + Na] , calcd for
+
3.4, 19.5. ESI-MS m/z (rel int): (pos) 397.2 ([M + Na] , 100),
25 36 5
C H O 416.25.
+
-
79.2 ([M + H] , 80); (neg) 409.2 ([M + Cl] , 40).
Preparation of C13â Acetate 69 and C13R Acetates 70
1
3â-Acetoxy-8R,11â-dimethyl-12â-isopropyl-5â,15-iso-
by Thermolysis of the Epoxy Acetate Derived from 64.
1
,2 3,4
propylidenedioxy-14-keto-(∆,
mL, pear-shaped flask, hydroxy ketone 73 (7.3 mg, 0.019 mmol,
.0 equiv) was dissolved in 0.9 mL of CH Cl under argon.
Pyridine (31 µL, 0.38 mmol, 20 equiv) and DMAP (0.5 mg,
∆
)-tricycle 74. In a 10-
To a solution of enol acetate 64 (7 mg, 0.017 mmol) in 0.5 mL
2 2
of CH Cl was added slowly a solution of DMDO in acetone
1
2
2
(0.07 M in acetone, 249 µL, 1 equiv) at - 50 °C, and the
resulting solution was stirred for 40 min at -50 °C and then
10636 J. Org. Chem., Vol. 70, No. 26, 2005

Total Synthesis of Guanacastepene A
0 min at 0 °C. Dimethyl sulfide (1 µL) was added, and the
1
2
6
1750, 1371, 1219, 1020. H NMR (400 MHz, acetone-d , 25
resulting mixture was warmed to room temperature and
evaporated to dryness. The crude epoxy acetate was heated
neat at 150 °C for 18 min. The reaction mixture was cooled to
°C): δ 9.91 (br s, 1H, C15-H), 7.45 (d, 1H, J ) 1.1 Hz, C2-
H), 5.48 (d, 1H, J ) 6.5 Hz, C13-H), 4.62 (m, 1H, C5-H),
3.97 (br s, 1H, OH), 2.08 (s, 3H, OAc), 1.99 (m), 1.90 (m), 1.79
room temperature and was dissolved in 1 mL of CH
2
Cl
2
.
(m), 1.63 (m), 1.40 (m), 1.27 (s, 3H, C16/17-H
3
), 1.12 (d, 3H, J
Pyridine (80 µL), a catalytic amount of DMAP, and acetic
anhydride (40 µL) were added, and the mixture was stirred
for 12 h. Evaporation provided crude product, which was
purified by preparative TLC (3% ethyl acetate in chloroform)
to afford the acetates in 1:1 de ratio. The C13-â acetate 69
was obtained in 31% yield (2.2 mg) and also C13-R acetate 70
) 6.6 Hz, C19/20-H ), 1.09 (s, 3H, C16/17-H ), 0.93 (d, 3H, J
3
3
1
) 6.4 Hz, C19/20-H ). H NMR (500 MHz, acetone-d , -50 °C,
3
6
key signals): major δ 9.96 (s, C15-H), 7.42 (s, C2-H), 5.45 (d,
J ) 5.6 Hz, C13-H), 4.64 (m, C5-H), 4.59 (d, J ) 5.4 Hz, OH),
2.10 (s, OAc); minor 9.72 (s, C15-H), 7.49 (s, C2-H), 5.53 (d, J
) 7.1 Hz, C13-H), 4.52 (m, C5-H), 4.48 (d, J ) 4.1 Hz, OH),
1
13
was obtained in 31% yield (2.2 mg). C13-R Acetate 70. H NMR
2.11 (s, OAc). C NMR (125 MHz, acetone-d , -50 °C, observed
6
(
3
500 MHz, CDCl ): δ 6.79 (br-s, 1H), 5.44 (d, 1H, J ) 12.0
signals): δ 201.1, 200.4, 193.3, 192.6, 170.4, 170.2, 161.6, 158.2,
149.34, 149.29, 140.1, 138.6, 133.3, 131.7, 74.6, 73.5, 62.5, 59.3,
59.2, 54.8, 50.4, 47.7, 46.4, 41.9, 38.7, 38.5, 35.5, 35.4, 27.7,
Hz), 4.48 (br-d, 1H, J ) 16.0 Hz), 4.30 (m, 1H), 4.08 (d, 1H, J
16.0 Hz), 2.10 (s, 3H), 1.93-1.72 (m, 6H), 1.62-1.44 (m, 3H),
.41-1.35 (m, 4H), 1.31 (s, 3H), 1.02 (s, 3H), 0.97 (d, J ) 7
Hz, 3H), 0.93 (s, 3H), 0.91 (d, J ) 7.5 Hz, 3H). ESI-MS: m/z
)
1
2
3
5.9, 25.6, 23.7, 23.5, 21.0. 20.2. ESI-MS m/z (rel int): (pos)
97.2 ([M + Na] , 100); (neg) 409.1 ([M + Cl] , 100).
+
-
+
4
36 5
39.3 ([M + Na] , calcd for C25H O 416.25.
Guanacastepene A (1). A 10-mL, round-bottomed flask
Acknowledgment. This work was supported by a
equipped with a reflux condenser and an argon inlet was
grant from the National Institutes of Health (CA-
28824). G.B.D. was an NIH Postdoctoral Fellow (1 F32
NS11150-01). H.Y. is grateful for a Predoctoral Fellow-
ship from Eli Lilly & Company. D.S.T. was a Damon
Runyon Cancer Research Foundation Postdoctoral Fel-
low (DRG-1641). S.L. was a US Army Breast Cancer
Research Program Postdoctoral Fellow (DAMD-17-99-
charged with a solution of acetonide 74 (7.9 mg, 0.019 mmol,
1
.0 equiv) and PPTS (0.95 mg, 0.0038 mmol, 0.2 equiv) in 1.9
mL of methanol under argon and heated at 70 °C for 30 min.
The reaction mixture was allowed to cool to room temperature
and then was concentrated under reduced pressure (at room
temperature). Purification by column chromatography on ca.
1
g of silica gel (elution with ethyl acetate) followed by
azeotropic drying with benzene at room temperature provided
diol, which was not stable to storage and used immediately in
the next step. The freshly prepared diol from the previous step
1
-9373). We thank Dr. George Sukenick, Sylvi Rusli,
Anna Dudkina, and Hui Fang (NMR Core Facility, CA-
0
2848) for mass spectral analysis and Dr. Louis Todaro
was dissolved in 1.6 mL of CH
flask under argon, and PhI(OAc)
equiv) was added. A solution of TEMPO (0.30 mg, 0.0019
mmol, 0.1 equiv) in 0.3 mL of CH Cl was added, and the
2
Cl
2
2
in a 10-mL, pear-shaped
(12.3 mg, 0.038 mmol, 2.0
(Hunter College) for X-ray crystallographic analysis.
2
2
Supporting Information Available: Experimental pro-
resulting solution was stirred for 3 h at room temperature and
concentrated under reduced pressure (at room temperature).
Purification by column chromatography on ca. 1 g of silica gel
1
13
cedures, H and C NMR spectral data, optical rotations,
HRMS, and additional information on key intermediates. This
material is available free of charge via the Internet at
http://pubs.acs.org.
(
gradient elution with 20-30% ethyl acetate/hexanes) provided
.2 mg (59%) of guanacastepene A (1) as an amorphous white
solid. TLC: R 0.73 (1:3 hexanes/EtOAc). IR (film): 3433 (br),
4
f
JO051470K
J. Org. Chem, Vol. 70, No. 26, 2005 10637