Total synthesis of (±)-maoecrystal v

Article abstract of DOI:10.1021/ja309905j

The total synthesis of racemic maoecrystal V has been accomplished. Key steps include an intramolecular Diels-Alder cyclization to rapidly construct the core system from simple starting materials and the creation of the A-C ring trans-fusion through intra

Full text of DOI:10.1021/ja309905j

Article
pubs.acs.org/JACS
Total Synthesis of ( )-Maoecrystal V
Feng Peng†,§ and Samuel J. Danishefsky*,†,‡
†Department of Chemistry, Columbia University, Havemeyer Hall, 3000 Broadway, New York, New York 10027, United States
‡Laboratory for Bioorganic Chemistry, Sloan-Kettering Institute for Cancer Research, 1275 York Avenue, New York, New York
10065, United States
S
* Supporting Information
ABSTRACT: The total synthesis of racemic maoecrystal V has
been accomplished. Key steps include an intramolecular Diels−
Alder cyclization to rapidly construct the core system from
simple starting materials and the creation of the A−C ring trans-
fusion through intramolecular delivery of a hydrogen to the
hindered β-face of the ring system.
possibility that research directed to maoecrystal V could bring
with it new compounds with therapeutically exploitable
anticancer activity.
INTRODUCTION
■
Upon learning of the existence of maoecrystal V and,
particularly, following contemplation of its compelling
structure, we were already drawn to the prospect of undertaking
a program directed to its total synthesis. While one could
hardly be very confident, a priori, as to the outcome of such a
venture, we were increasingly attracted to the challenge.
Problems of this level of complexity often carry with them
significant learning opportunities. In retrospect, maoecrystal V
more than fulfilled our expectations in this regard.
Our interest in maoecrystal V as a fascinating chemical entity
was further augmented by perceptions, however preliminary,
that a successful synthesis program might enable the discovery
of new antimitotic agents of value. Of late, the team of Han-
Dong, from the Kunming Institute of Botany, has been
particularly active in searching for biologically potent
diterpenoids from the Isodon plants of the Laiatae family.1
Indeed, maoecrystal V was first isolated in 1994 from Isodon
eriocalyx Hara, which is distributed in Yunnan province. The
source of vegetation from which maoecrystal V was obtained
had been used in herbal medicine for some time. Eventually, in
2004, a single crystal of maoecrystal V was obtained.2
Crystallographic analysis revealed its structure to be that
shown in Figure 1. Moreover, maoecrystal V was evaluated for
cytotoxicity against a variety of human tumor cell lines, and it
was found to be selectively active, for instance, against the He
La cell line, with an impressive IC50 value of approximately 20
ng/mL. While this lead has not, to our knowledge, been
developed further, even these preliminary results raise the
Returning to the chemistry-based incentives, we could hardly
help but notice a collection of familiar substructural motifs in
maoecrystal V. The complexity of the total synthesis challenge
posed by 1 arises from the highly compact architecture that
interlocks these otherwise unexceptional patterns. Particularly
conspicuous is the grouping of the fully substituted positions
(8, 9, and 10). These carbon centers provide the basis for
fusions between the C-ring d-lactone and the bicyclic AB
system. Moreover, carbons 5 and 10 define an AB junction that
is trans-fused. The β-displayed hydrogen atom at C5 necessarily
bears a 1,3-diaxial relationship with the C9−C15 bond. It
appeared that delivery of a hydrogen from an external source to
this highly hindered β-face would be difficult. Also of interest in
the A-ring is the arrangement of the γ-disposed gem-dimethyl
cyclohexenone functionality incorporating carbons 1−4. It
would be necessary to accommodate an α-disposed methyl
group at the epimerizable C16. Indeed, the only carbon centers
on the maoecrystal V backbone lacking functional groups and
stereogenicity are those at positions 11, 12, and 14.
We were not the first to be enticed into the maoecrystal V
total synthesis challenge. Well before we took up the problem, a
variety of fascinating approaches had been disclosed from other
laboratories.3 The extensive undertakings directed to 1 have
culminated in its total synthesis by Yang and associates.4 Below,
we present our recently completed total synthesis of
maoecrystal V as well an account of the planning and
experimentation that enabled the reaching of that goal. To
provide a proper setting, we need to briefly recapitulate our
initially proposed, albeit ultimately unsuccessful, route.
Important lessons, gleaned from the breakdown of our first
approach, served to inform the eventually successful foray.
Received: October 7, 2012
Figure 1. Structure of maoecrystal V (1).
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ja309905j | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
prospects for creating a carbon−carbon bond at C9, subsequent
to cycloaddition, did not seem promising.
RESULTS AND DISCUSSION
■
As mentioned before, there are several motifs that could serve
to facilitate a retrosynthetic analysis in the context of what we
have termed pattern recognition analysis (PRA).5 Among the
recognizable patterns that attracted our attention is the bridged
bicyclo [2.2.2] octane domain (i.e., DE rings). The possibility
of building this motif by a Diels−Alder (DA) reaction6 between
a usefully substituted cyclohexadiene and an appropriately
activated ethylene presented itself [see (i) + (ii) → (iii), Figure
2]. At this early stage of analysis, we leave unspecified the
Such concerns regarding the feasibility of progression, even
beyond a successful bimolecular DA step, prompted consid-
eration of an intramolecular Diels−Alder (IMDA) reaction7 to
assemble the DE motif. Upon further reflection, it seemed that
an IMDA-based strategy would provide a means for dealing
with the vicinal bis neopentyl motif represented by carbons 9
and 10. Indeed, we were struck by the thought that the C10
center of (i) might carry, as one of its substituents, the
cyclohexadiene moiety previously described as (ii). In this way,
a means of dealing with the worrisome issue of bis-quaternary
C9−C10 bond formation would already be in place. The
activated olefin, destined to serve as the dienophile, would be
connected to the quaternary center at C10 through some
version of a hydroxymethyl linker. Theoretically, C10 could
already be housed within a six-membered ring corresponding to
the A-ring of maoecrystal V. This line of conjecture is captured
in the generalized hypothetical IMDA substrate (iv). It is
interesting to note that (iv) arises, conceptually, from
amalgamation of the previously hypothesized (i) and (ii)
through a Cα−Cβ−O spacer, wherein Cα is actually part of a
six-membered moiety destined to emerge as the A-ring of
maoecrystal V.
Still at the planning stage, it is well to recognize an
interesting stereochemical question in the projected IMDA
reaction. There is a subtle issue of facial selectivity in the
envisioned cycloaddition step. Thus, to reach maoecrystal V by
the route discussed above, the IMDA reaction must occur in
the rotameric sense [(iv)-a], which connects the emerging
configurations at C9 and C13 (in a relative sense) to the resident
stereogenicity [see (iv)-a → (v)]. For instance, if the IMDA
cycloaddition occurred through the alternative cyclohexadiene
rotamer, (iv)-b, the product after elimination of HY would be
(vi). The conversion of the hypothetical (vi) to maoecrystal V
(1) would be quite problematic, if even doable. Such a
Figure 2. IMDA-based strategy toward maoecrystal.
nature of functional groups A and Y. Needless to say,
delineation of these functions must be sensitive to the need
for ensuring regio- and stereoselectivities in productive
directions. Pursuing this line of analysis, it seemed that the
bridgehead center marked with an asterisk corresponding to C9
of the future maoecrystal V must already carry the means for
connectivity to the eventually required rings, A, B, and C. The
Figure 3. First attempted IMDA route toward maoecrystal V. Key: (a) Pd(OAc)2, 2-di-t-butylphosphino-2′-methylbiphenyl, K3PO4, THF, 80 °C, 12
h, 91%; (b) TMSCHN , Hunig’s base, CH CN/MeOH = 9:1, 6 h, 100%; (c) Bu SnCH OMOM, BuLi, THF, −78 to −40 °C, 30 min, 0.5% HCl
̈
2
3
3
2
workup, 75%; (d) HCl/MeOH, 50 °C, 75%; (e) PivCl, Py, DCM, 12 h, 96%; (f) NaBH , CeCl , MeOH, 0 °C, 2 h, 93%; (g) MOMCl, Hunig’s base,
̈
4
3
DCM, 12 h, 95%; (h) DIBAL-H, −78 °C, DCM, 30 min, 95%; (i) KH, 18-crown-6, ICH2SnBu3, 0 °C, THF, 6 h, 90%; (j) n-BuLi, −78 to −20 °C,
THF, 6 h, 88%; (k) Li, NH3(l), t-BuOH/THF, −78 °C, 20 min, −33 °C, 40 min; (l) 1 N HCl, 0 °C, THF/MeOH (10/1), 8 h, two steps, 78%; (m)
acid chloride, Py, DCM, 0 °C, 72%; (n) TBSOTf, TEA, DCM, 0 °C, 15 h, 81%; (o) 180 °C, sealed tube, toluene, 12 h.
B
dx.doi.org/10.1021/ja309905j | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
retrofitting would involve interchanging the etheno and ethano
bridges. The former would have to emerge as the
unfunctionalized carbons 11 and 12, while the latter would
require transformation to the extensively functionalized C15 and
C16 span.
In the opening stages of the program, we hoped to address
several issues. The first dealt with the matter of synthesizing a
usefully functionalized version of the genus (iv). We would
then probe the stereochemistry of the cycloaddition step, (i.e.,
which Curtin−Hammett rotamer of (iv) would define the
cycloaddition event). We well recognized that, in principle, the
direction and scope of that type of stereoselectivity could be a
function of the particular A-ring groups selected. These groups
must also be selected with the goal of complementarity as to
regiodirectionality in the cycloaddition step. High-quality
diene−dienophile regio-complementarity tends to lower the
activation barrier for DA or IMDA reactions.
Original Synthetic Strategy toward Maoecrystal V. In
our opening foray, we decided, initially, to introduce the
cyclohexadiene ultimately for the IMDA reaction in the form of
a stable anisole derivative. As was previously described,8 the
future cyclohexadiene (E ring) was joined to the future A ring
by a palladium-mediated arylation9 of 2, using 3-bromoanisole
(Figure 3). Following regiospecific O-methylation, 3 was in
hand. Through standard and pleasing chemistry,10 3 was
converted to 4 and, thence, after several high yielding steps, to
7. This compound was transformed with high stereoselection to
8 via a Still−Wittig [2,3-sigmatropic] rearrangement.11 Thus, a
system bearing the ultimately required quaternary substituted
C9 was in hand.
We had already made the decision that the DA diene (ii)
would, in the event, be presented as a cross-conjugated silyl
enol ether. In this fashion, the eventual introduction of the
required functionalities at C15 and C16 would have been
enabled. For the moment, we were content to delay inclusion
of the methyl group at the eventual C16, confident that the
required alkylation could be arranged at a later stage. There
seemed to be required only the Birch reduction of 8 to generate
a precursor to a viable cyclohexenone, which would be a
candidate precursor to provide the required cross-conjugated
siloxydiene. In practice, Birch reduction12 of the anisole
function 8 was nicely accomplished. However, these conditions,
in our hands, also caused reduction of the exomethylene group
(see formation of 9) as a 3:1 mixture of epimers. In retrospect,
the tendency of Birch-type conditions to effect reduction of
proximal isolated unhindered olefins is not without prece-
dent.13
Thermolysis of compound 11 at 180 °C in a sealed tube gave
rise to 13. We were, of course, pleased to find that the IMDA
reaction had indeed occurred. However, detailed NMR analysis
showed convincingly that the stereochemistry of the [2.2.2]
bicyclo-octane domain in 13 was of the undesired sort. Thus,
13 is not eligible for conversion to maoecrystal V for reasons
discussed above, now further complicated by an extraneous
methyl group.
While ultimately unsuccessful, potentially important mes-
sages were drawn from this effort. The progression described
was seen to validate the general IMDA route adumbrated in
Figure 1. In so doing, it underscored the difficulties associated
with predicting the facial selectivity of the cycloaddition step.
Thus, the overall rotameric state of the cyclohexadiene, which
determines the outcome of the IMDA reaction, arises from
several possible component factors. First, there might be an
inherent preference, rooted in the effective “A-value” of the two
relevant edges (i.e., whether the sp2 or sp3 carbons prefer to
turn toward one or the other sectors of the incident
cyclohexane). In the case at hand, it seems that the product
obtained is arising from the rotamer in which the presumably
more sterically demanding (sp3) carbons of the cyclohexadiene
are facing the more hindered side of the cyclohexane ring
(bearing the C-methyl and geminal dimethyl substituents).
However, the operative rotameric form of the cyclohexadiene
also defines which of the trans-related activating groups of the
dienophile (in the case of 11 as a fumarate) is presented in an
endo fashion, and which emerges exo. In the case of 13, the
future lactonic ester group had emerged endo, whereas the
carbomethoxyl group presented exo. In the end, it is the
weighting of steric factors in the inherent rotameric
distribution, with endo:exo presentational preferences in the
transition state, which defines the total Curtin−Hammett14
outcome. The difficulty of predicting the facial outcome was of
particular concern because there is required a substantial
number of steps in setting up the IMDA substrate in which this
question is posed. Hence, the prospect of trying to solve the
face selectivity uncertainty by noninformed screening of various
possibilities was not attractive.
Modified IMDA Route to Maoecrystal V. It was in
struggling with this problematic scenario that a rather
interesting solution presented itself. Perhaps the six-membered
A-ring could be presented in an achiral setting. In such a
context, the only consequence of face selectivity in the IMDA
step would relate to the endo:exo directivity of the two
activating groups of the trans-diactivated dienophile. On closer
reflection, even this question is, at most, of only temporary
relevance, because it was our goal to install a double bond
between carbons 8 and 14, thus rendering the original endo:exo
question moot. Put differently, in real terms the operative
dienophile would be a propiolate equivalent, wherein the
endo:exo issue does not even arise.15 A plan evolved wherein a
3,3-disubstituted cyclohexadiene would serve as an achiral
version of the future A-ring of 1. As above, we would hope to
take recourse to a cross-conjugated siloxydiene as the diene
component in the projected IMDA reaction. In this way, the
enoxysilane functionality of the resultant IMDA product could
be exploited to guide the emergence of the required
functionality at the C15−C16 bridge. As will be seen, a critical
part of the new plan was the installation of the double bond
between carbons 8 and 14. To increase the likelihood of
success, we would attempt to employ an α,β-sulfonylacrylate
system as our putative dienophile. We anticipated that if the
For the case at hand, having come this far, we continued with
the original plan. It was well appreciated that the presence of
the methyl group at C5 would pose a major obstacle to
completion of the total synthesis goal. However, such a study
could still serve as a model to evaluate the IMDA program. We
thus took the diastereometric mixture 9 forward. We selected a
fumarate function as the activated dienophile in the projected
IMDA step, confident that at some stage, the extraneous
carbomethoxyl group could be eliminated with the formation of
a double bond between the future C8 and C14. Fortunately, after
the acylation step, the C5 isomers were separable. We moved
forward with the major isomer, wherein the secondary methyl
group is syn to the resident MOM function at C1. Happily, enol
silylation of the acylated substrate provided, as the major
product, the desired cross-conjugated siloxydiene (see 11).
This compound was to serve as our IMDA substrate.
C
dx.doi.org/10.1021/ja309905j | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
IMDA reaction did indeed occur, the resultant adduct would be
prone to simple β-elimination of phenylsulfinic acid to generate
the required 8,14-double bond.
The aggregate of these considerations prompted us to
formulate specific compound 14 as our proposed IMDA
substrate (Figure 4). As seen, the operative rotameric state of
junction). Below, we relate the experiments by which this
general plan was reduced to practice.
The campaign to implement these conceptions began on a
rather positive note.16 Thus, two readily available compounds
20 and 21 could be joined, as shown, to provide 22 (Figure
6).17 Clearly, the reaction involved generating the enolate of 19,
Figure 4. Revised IMDA route toward maoecrystal V.
the siloxyhexadiene determines only the endo:exo disposition of
the two activating groups (see structures 15 and 16). Following
β-elimination, the two permutations converge (see enantio-
meric structures 17 and ent-17, which anticipate cleavage of the
silyl enol ether). We note that at the stage of 17, the two
nonconjugated double bonds in the putative A-ring are
diastereotopic.
Figure 6. Synthesis of intermediate 28 en route to maoecrystal V. Key:
(a) LDA, −78 °C, THF, 40%; (b) DIBAL-H, −78 °C, DCM, 2 h; (c)
MnO2, room temperature, DCM, 40 min, two steps, 68%; (d) 24, Py,
0 °C, DCM, 30 min, 86%; (e) TBSOTf, TEA, DCM, −78 °C, 12 h,
91%; (f) toluene, sealed tube, 166 °C, 1 h, then TBAF, THF, 62%; (g)
H2O2, NaOH, MeOH, 0 °C, 95%; (h) MgI2, DCM, 45 °C; (i)
Bu3SnH, AIBN, toluene, reflux, two steps, 50%.
If all would have proceeded well to this point, the central
challenge of the synthesis would then lie in translating this
diastereotopic distinction to allow for chemoselectivity between
the two double bonds of the A-ring. At first glance, the
prospects for realizing differentiation over these olefinic
linkages based on inherent levels of steric hindrance were not
promising. Fortunately, an interesting solution to this problem
also presented itself. As was alluded to above, a C8−C14 double
bond, at the stage of 17, was central to our proposed solution.
This α,β-unsaturated ester double bond would hopefully be
chemically distinct from the isolated olefins of the putative A
ring joining C1 and C2 as well as C4 and C5. The thought was to
exploit the α,β-unsaturated carbonyl character at C8−C14
double bond to introduce, in some manner, a hydroxyl group
at C8 (Figure 5). This alcohol would, in turn, confer selective
activation to the proximal C4−C5 double bond, thereby
allowing for the installation of the missing A-ring functionality.
Eventually, the tetrahydrofuranoid C-ring would arise from
joining the newly introduced C8 hydroxyl to C4 ultimately in
the required stereochemical sense (i.e., with a trans AC
which underwent a 1,4 addition to the β-chloroenone function
of 21. Ejection of chloride from this intermediate afforded 22.
Although the yield of 22 was modest (40%), this easily
obtained compound contains all of the carbon−carbon bonds
necessary to elaborate the dienic component of the envisioned
IMDA cycloaddition reaction required for maoecrystal. To pave
the way for setting up the fully equipped IMDA substrate, the
ester and ketone functions were reduced, as shown, and the
allylic alcohol arising from the ketone reduction underwent
selective oxidation with manganese dioxide to afford the ketone
23 in reasonable yield. Fortunately, the primary (neopentyl!)
alcohol function of 23 underwent smooth acylation with acyl
chloride 2418 under the conditions shown, to provide, following
enol silylation, the required IMDA substrate 25.
Moreover, as was previously found (cf., 11→13, Figure 3),
25 underwent thermally induced IMDA reaction to provide a
cycloadduct, which was not, per se, characterized. Happily, on
treatment of this product with TBAF, there was obtained
(presumably by β-elimination of phenylsulfinate) the α,β-
unsaturated lactone bearing a ketone at a carbon bridge of the
[2.2.2] bicyclooctane matrix. It will be appreciated that, for a
given direction (α or β) of attack on the diene, the “molecular
decision” between the two cyclohexadiene rotamers determines
only the enantiomeric state of the product after elimination of
the phenylsulfinate.
We have thus far not dealt with the problem of synthesizing
the natural enantiomer of maoecrystal V, because in the first
instance we anticipated being fully occupied, even with the
challenge of reaching its racemate. However, it is apparent on
reflection that if racemate 26 could indeed be established as a
Figure 5. Proposed completion of maoecrystal V synthesis.
D
dx.doi.org/10.1021/ja309905j | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
competent precursor for maoecrystal V, the task of reaching the
enantiomerically pure version of the natural product could be
accomplished through catalytically induced enantiospecificity in
the IMDA cycloaddition of achiral 25. Precedents for
successfully achieving control of enantiomeric outcomes in
the Diels−Alder reaction provide some basis for hope in this
regard.19
Happily, chemoselective nucleophilic epoxidation of the
conjugated double bond in structure 26 was readily
accomplished, as shown, to provide epoxide 27. Moreover, as
hoped, the epoxide could be opened at its secondary carbon
with magnesium iodide to provide a vicinal iodohydrin.20
Reduction of the iodo function was achieved via its reaction
with tri n-butyltin hydride. At this stage, we had in hand the α-
hydroxylactone 28, wherein the required trans fusion of the
B:D rings had been accomplished. The position “claimed” to
this point was fully secured by a crystallographic verification of
structure 28.
candidate for advancement would be a system such as 31,
wherein the configuration at C5 has somehow been corrected
such that the A:B junction is trans. The problem we faced was
that, as discussed above, external delivery of a hydrogen to the
required β-face of C5 would be highly disfavored from a purely
steric hindrance perspective. This β-face corresponds to the
inside of a molecular cavity arising from the interlocking of the
ABCD matrix.
In considering this problem, it seemed that a possible
solution might involve delivery of a proton to the β-face of C5
by suprafacial rearrangement of an already β-disposed hydrogen
from within the molecule. This line of conjecture led us to
consider an epoxide of the type 33 (Figure 8). We could
We next turned our attention to building the required
tetrahydrofuranoid substructure (i.e., ring B). It was our
expectation that the hydroxyl group, in place at C8 in structure
28, could be exploited to differentiate the two diastereotopically
related double bonds present in the A-ring of 28. Happily, this
turned out to be the case. As shown in Figure 7, treatment of
Figure 8.
confidently predict that, if access could be gained to an exo-
glycal of the type 32, attack by an external source of OH
+
would occur from the outside α-face. For the moment, we leave
unspecified (see asterisks) the functionality elsewhere that
might be compatible with reaching 32 and converting it to
epoxide 33. The key argument was that it ought to be possible
to prompt the rearrangement of 33 to 34 by Lewis acid catalysis
wherein suprafacial “hydride” migration of the β-hydrogen from
C4 would deliver it to the desired (albeit highly hindered) β-
face at C5, thereby producing 34.
In a previous disclosure,16 operating in a model series, a
compound corresponding in broad concept to 33, wherein
carbon centers 1, 2, 3, 15, and 16 were all methylene groups,
was synthesized. Indeed, the postulated α-face epoxidation
occurred, as did the rearrangement to the C4 ketone, C5-β-H
product, corresponding to 34. Below, we describe how this
strategic design was reduced to practice in a series of
compounds bearing the chemical implements necessary to
reach maoecrystal V. For this purpose, we must return to
compound 30.
Completion of Synthesis of Maoecrystal V. Having
demonstrated the feasibility of the exo-glycal epoxide rearrange-
ment sequence in the context of a model system,16 we now
turned our attention to achieving this transformation in the
suitably functionalized maoecrystal V series. Because the C16
ketone would be incompatible with the proposed conditions,
we elected to temporarily protect this functionality as a MOM
ether. In the event, intermediate 30 was converted to
compound 35, bearing a C16 secondary alcohol. Unfortunately,
as shown in Figure 9, the NaBH4 reduction of ketone 30 was
not stereoselective, and 35 was isolated as a 1:1 mixture of C16
isomers. Although highly troubling from an aesthetic
perspective, the lack of selectivity in the reduction step did
not materially impact on the overall efficiency or yield of the
synthesis. It proved possible to carry the mixture of isomers
through the synthesis, until the point at which reoxidation of
the alcohol epimers served to eliminate the issue of C16
stereointegrity (see ketone 44). Thus, the epimeric C16
secondary alcohols were masked with MOM protecting groups
to furnish intermediate 36, which, upon stereospecific
Figure 7. Installation of the tetrahydrofuranoid B-ring. Key: (a) m-
CPBA, DCM, room temperature, 18 h, 72%; (b) p-TsOH·H2O, DCM,
room temperature, 12 h, 90%.
28 with m-CPBA indeed gave rise to, apparently, a single
epoxide (29). While, in principle, the peracid oxidant could
have attacked 28 from its β-face under direction from the
hydroxyl function, the trajectory for such an epoxidation would
have been quite hindered. However, the C8 hydroxyl group
provides sufficient activation, even for the α-face of the C4−C5
double bond to distinguish it from the noncompetitive C1−C2
olefin.21 When this compound was subjected to the action of p-
toluenesulfonic acid, as shown, the anticipated cyclization
reaction occurred, undoubtedly with inversion at C5, to
generate the requisite tetrahydrofuranoid B-ring. However, as
fully expected, the junction of the A- and B-rings in structure 30
was cis.
It was hoped that the hydroxyl group at C4 would eventually
become a ketone that would serve as a precursor for
introduction of the gem dimethyl group at this center.22
However, it would, of course, be necessary to correct the non-
natural relative stereochemistry at C5 before the two methyl
groups were introduced at C4. In other words, a means must be
found for the α-disposed hydrogen at C5 in compound 30 to
eventually occupy a β-configuration corresponding to the
natural configuration of maoecrystal V. A hypothetical
E
dx.doi.org/10.1021/ja309905j | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 9. A-Ring functionalization and epimerization of C5. Key: (a) Ac2O, Py, DCM; (b) NaBH4, DCM/EtOH, 85% yield over two steps; (c)
MOMCl, i-Pr2NEt; (d) K2CO3, MeOH, 90% yield over two steps; (e) m-CPBA, room temperature, 95%; (f) DMP, NaHCO3, DCM, 0 °C, 85%; (g)
Ac2O, Py, DCM, 90%; (h) PhSH, Et3N; then NaBH4, EtOH/DCM, 78%; (i) Raney-Ni; (j) MsCl, DMAP, 65% over two steps; (k) K2CO3/MeOH,
room temperature, 95%; (l) DMDO; then BF3·OEt2, 82%.
epoxidation (m-CPBA) of the 1,2-double bond afforded 37,
bearing exclusively the α-oriented C1−C2 epoxide. At this stage,
Dess−Martin oxidation of the secondary alcohol, followed by
attempted silica gel purification, afforded a γ-hydroxyenone,
which was readily converted to its acetate, 38. The next goal
was the installation of an olefin corresponding to an exo-glycal
at the C4−C5 position. Happily, this goal could be
accomplished in a surprisingly straightforward fashion. Thus,
conjugate addition of thiophenol to the double bond of 38,
followed by reduction of the resultant β-thiophenoxy ketone
with NaBH4, yielded 39. The latter was subjected to Raney-Ni
desulfurization, followed by dehydration and C1 acetate
deprotection, to afford the key substrate 40. We were pleased
to observe that, upon exposure of this exo-glycal to the
previously described conditions, compound 40 rapidly under-
went the hoped-for sequence of exo-glycal epoxidation/
rearrangement to deliver 41 as a single epimer at carbons 1
and 5, although with epimeric −OMOM groups at C16.
With the tetrahydrofuranoid B-ring now in place, the next
milestone for completion of the total synthesis would require
Figure 10. Completion of the synthesis of maoecrystal V (1). Key: (a)
installation of the γ-dimethyl-α,β-unsaturated ketone function-
Lombardo reagent, DCM, room temperature, 85%; (b) CH2I2, Zn/Ag,
ality on the A-ring. The sequence by which this functional motif
Et2O, 36 °C, 88%; (c) PCC, DCM, room temperature, 76%; (d) H2,
was installed is shown in Figure 10. Thus, the C4 ketone of
PtO2, AcOH, 40%; (e) Lombardo reagent, DCM, 0 °C, 80%; (f) p-
intermediate 41 was converted to a cyclopropane (43) through
TsOH·H2O, benzene, 76 °C, 85%; (g) LDA, TMSCl, THF, −78 °C,
a two-step sequence consisting of Lombardo olefination23 (see
90%; then Pd(TFA)2, CH3CN, 80%; (h) TFDO, CH2Cl2, −78 °C→0
compound 42) followed by Zn/Ag-mediated Simmons−Smith
inspired cyclopropanation.24 Interestingly, under these con-
ditions, the MOM protecting group underwent CH2 insertion
(see MOM→MOE, 43). Upon exposure to PCC conditions,
the MOE group was removed and the epimeric alcohols were
oxidized to deliver the bis-ketone intermediate, 44, as a single
entity. The cyclopropane ring was then reductively cleaved to
furnish intermediate 45, possessing the key C4 gem-dimethyl
motif.
We next turned our attention to the selective functionaliza-
tion of the C16 ketone. In the event, we were pleased to find
that substrate 45 readily underwent regioselective olefination,
again via the Lombardo reagent, to afford exo-olefin 46. At this
stage, we were able to realize acid-mediated isomerization of
the exo-methylene group of 46 to afford 47, bearing the
trisubstituted olefin.25 Subsequent dehydrogenation of ketone
47, to install the requisite α,β-unsaturation, initially proved to
be quite challenging, presumably due to the neopentyl
character of the ketone. Ultimately, under our optimized
conditions, 47 was treated with LDA and TMSCl to generate a
stable TMS enol ether. Upon exposure to Pd(TFA)2, the
hoped-for Saegusa oxidation proceeded in good yield to afford
°C, dr = 1:1, 90%; (i) BF3·OEt2, DCM, room temperature, 85%.
the critical intermediate 48. At this stage, all that remained was
the installation of the C15-oxo, C16 α-methyl functionality. We
hoped to accomplish this goal. In the event, epoxidation of 48
with TFDO26 afforded a chromatographically separable
diastereomeric mixture of epoxides 49 and 50 (92% yield, 1:1
dr). Although we had anticipated that epoxide formation would
predominantly occur from the less hindered side of the olefin,27
to deliver 49 as the major isomer, we speculate that the
observed lack of stereoselectivity may be attributed to subtle
differences in the electronic environment surrounding the two
faces of the olefin under attack.28 In any case, the epoxide
isomers were readily separated, and, in the final step of the
synthesis, olefin 49 was treated with BF3·OEt2 to deliver
maoecrystal V (1) as a single isomer in 85% isolated yield.29
1
The H and 13C NMR data were in full accord with those
reported for the natural product itself. The mission of
synthesizing the racemate of maoecrystal V had been
accomplished.
F
dx.doi.org/10.1021/ja309905j | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
2009, 11, 4774−4776. (c) Nicolaou, K. C.; Dong, L.; Deng, L. J.;
Talbot, A. C.; Chen, D. Y. K. Chem. Commun. 2010, 46, 70−72.
(d) Lazarski, K. E.; Hu, D. X.; Sterm, C. L.; Thomson, R. J. Org. Lett.
2010, 12, 3010−3013. (e) Singh, V.; Bhalerao, P.; Mobin, S. M.
Tetrahedron Lett. 2010, 51, 3337−3339. (f) Baitinger, I.; Mayer, P.;
Trauner, D. Org. Lett. 2010, 12, 5656−5659. (g) Gu, Z.; Zakarian, A.
Org. Lett. 2011, 13, 1080−1082. (h) Dong, L.; Deng, L.; Lim, Y. H.;
Leugn, G. Y. C.; Chen, D. Y.-K. Chem.-Eur. J. 2011, 17, 5778−5781.
For an elegant synthesis of maoecrystal Z, see: (i) Cha, J. Y.; Yeoman,
J. T. S.; Reisman, S. E. J. Am. Chem. Soc. 2011, 133, 14964−14967.
(4) Gong, J.; Lin, G.; Sun, W.; Li, C.; Yang, Z. J. Am. Chem. Soc.
2010, 132, 16745−16746.
CONCLUSION
■
In summary, the total synthesis of maoecrystal V, albeit as a
racemate, has been completed. Among the key steps was the
IMDA reaction of 28, which served to construct virtually the
entire core system from a precursor obtained in one step from
readily available starting materials. Key features of the synthesis
included solution of the facial selectivity problem in the IMDA
reaction by utilization of an achiral A-ring equivalent. Another
feature involved creating the required trans-fusion of the A- and
C-rings through intramolecular delivery of a hydrogen to the
otherwise hindered β-face of C5 (see epoxidation of
intermediate 40 and rearrangement of the oxirane to 41).
The same logic was used to control the methyl group at C16,
although in this instance the elegance of the scheme was
compromised by a surprising nonstereospecific epoxidation of
the trisubstituted olefin between carbons 15 and 16.
The general directions that can be anticipated for an
improved synthesis are clear. Ideally, one would be able to
achieve enantiomeric control in the critical IMDA reaction.
Moreover, it would be preferable if the trisubstituted olefin,
eventually between carbons 15 and 16, could be introduced at a
much earlier point. Finally, it could well be hoped to achieve
stereoselectivity in the epoxidation of this 15−16 double bond
by fine-tuning the electronic characteristics of the A-ring.28
Future studies along these general lines are envisioned.
However, even in advance of these very important upgrades,
soon to be pursued, we feel that much has already been learned
from “operation maoecrystal V”.
(5) Wilson, R. M.; Danishefsky, S. J. J. Org. Chem. 2007, 72, 4293−
4305.
(6) Roush, W. R. In Comprehensive Organic Synthesis; Trost, B. M.,
Ed.; Pergamon: Oxford, 1991; Vol. 5, pp 513−550.
(7) (a) Nicolaou, K. C.; Snyder, S. A.; Montagnon, T.;
Vassilikogiannakis, G. Angew. Chem., Int. Ed. 2002, 41, 1668−1698.
(b) Liao, C.; Peddinti, R. K. Acc. Chem. Res. 2002, 35, 856−866.
(c) Ihara, M. Chem. Pharm. Bull. 2006, 54, 765−774.
(8) Peng, F.; Yu, M.; Danishefsky, S. J. Tetrahedron Lett. 2009, 50,
6586−6587.
(9) Fox, J. M.; Huang, X.; Chieffi, A.; Buchwald, S. J. Am. Chem. Soc.
2000, 122, 1360−1370.
(10) Stork, G.; Danheiser, R. J. Org. Chem. 1973, 38, 1775.
(11) Still, C.; Mitra, A. J. Am. Chem. Soc. 1978, 100, 1927−1928.
(12) Rabideau, P. W. Tetrahedron 1989, 45, 1579−1603.
(13) (a) Whitesides, G. M.; Whmann, W. J. J. Org. Chem. 1970, 35,
3565−3567. (b) Williams, D. R.; Robinson, L. A.; Nevill, C. R.; Reddy,
J. P. Angew. Chem., Int. Ed. 2007, 46, 915−918.
(14) Seeman, J. I. Chem. Rev. 1983, 83, 84−134.
(15) In our hands, in related cases, attempted use of a propiolate
ester for the IMDA reaction was unsatisfactory. Insofar as reaction
occurred, it was followed by aromatization.
ASSOCIATED CONTENT
* Supporting Information
Experimental procedures, spectral and other characterization
data. This material is available free of charge via the Internet at
http://pubs.acs.org.
■
S
(16) Peng, F.; Danishefsky, S. J. Tetrahedron Lett. 2011, 52, 2104−
2106.
(17) For the application of Birch-type alkylation in natural product
synthesis: (a) Kondo, K.; Sodeaka, M.; Mori, M.; Shibasaki, M.
Tetrahedron Lett. 1993, 34, 4219−4223. (b) Schultz, A. G.; Holoboski,
M. A.; Smvth, M. S. J. Am. Chem. Soc. 1996, 118, 6210−6219.
(c) Schultz, A. G.; Wang, A. J. Am. Chem. Soc. 1998, 120, 8259−8260.
(18) (a) Strekowski, L.; Kong, S.; Battiste, M. A. J. Org. Chem. 1988,
53, 901−904. (b) Guan, Z. H.; Zuo, W.; Zhao, L. B.; Ren, Z. H.; Liang,
Y. M. Synthesis 2007, 1465−1470. (c) Danishefsky, S. J.; Harayama, T.;
Singh, R. K. J. Am. Chem. Soc. 1979, 101, 7008−7012. (d) Danishefsky,
S. J.; Morris, J.; Mullen, G.; Gammill, R. J. Am. Chem. Soc. 1982, 104,
7591−7599.
AUTHOR INFORMATION
Corresponding Author
s-danishefsky@ski.mskcc.org
■
Present Address
§Merck &Co., Inc., Rahway, New Jersey 07065, United States.
Notes
The authors declare no competing financial interest.
(19) (a) Hayashi, Y.; Rohde, J. J.; Corey, E. J. J. Am. Chem. Soc. 1996,
118, 5502−5503. (b) Corey, E. J.; Shibata, T.; Lee, T. W. J. Am. Chem.
Soc. 2002, 124, 3808−3809. (c) Ryu, D. H.; Corey, E. J. J. Am. Chem.
Soc. 2003, 125, 6388−6390. (d) Corey, E. J. Angew. Chem., Int. Ed.
2002, 41, 1650−1667.
ACKNOWLEDGMENTS
■
Support was provided by the NIH (GM102872 to S.J.D.). F.P.
thanks Eli Lilly and Co. for a graduate fellowship. We thank
Drs. J. Decatur and Y. Itagaki for NMR and mass spectrometric
assistance, respectively; Aaron Sattler and Wes Sattler (Parkin
group, Columbia University) for X-ray experiments (NSF,
CHE-0619638); Yanqing Pan (Now Gilead Alberta ULC) for
synthesizing important intermediates; and William F. Berkowitz
for helpful discussions. Special thanks to Rebecca Wilson for
valuable help in editing the manuscript.
(20) Otsubo, K.; Inanaga, J.; Yamaguchi, M. Tetrahedron Lett. 1987,
28, 4435−4436.
(21) Hoveyda, A. H.; Evans, D. A.; Fu, G. C. Chem. Rev. 1993, 93,
1307−1370.
(22) Oppolzer, W.; Godel, T. J. Am. Chem. Soc. 1978, 100, 2583−
2584.
(23) (a) Lombardo, L. Tetrahedron Lett. 1982, 23, 4293−4294.
(b) Lombardo, L. Organic Syntheses 1987, 65, 81−83.
(24) (a) Rousseau, G.; Conia, J. M. Tetrahedron Lett. 1981, 22, 649−
652. (b) Denis, J. M.; Girard, C.; Conia, J. M. Synthesis 1972, 549−
551.
REFERENCES
■
(1) Sun, H. D.; Xu, Y. L.; Jiang, B. Diterpenoids from Isodon Species;
Science Press: Beijing, China, October 2001.
(2) Li, S. H.; Niu, X. M.; Shen, Y. H.; Zhang, H. J.; Sun, H. D.; Li, M.
L.; Tian, Q. E.; Lu, Y.; Cao, P.; Zhang, Q. T. Org. Lett. 2004, 6, 4327−
4330.
(3) (a) Gong, J.; Lin, G.; Li, C. C.; Yang, Z. Org. Lett. 2009, 11,
4770−4773. (b) Krawczuk, P. J.; Schone, N.; Baran, P. S. Org. Lett.
(25) Birman, V. B.; Danishefsky, S. J. J. Am. Chem. Soc. 2002, 124,
2080−2081.
(26) (a) Mello, R.; Fiorentino, M.; Fusco, C.; Curci, R. J. Am. Chem.
Soc. 1989, 111, 6749−6757. (b) D’Accolti, L.; Dinoi, A.; Fusco, C.;
Russo, A.; Curci, R. J. Org. Chem. 2003, 68, 7806−7810.
G
dx.doi.org/10.1021/ja309905j | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
(27) Indeed, in a model system, lacking the α,β-unsaturated ketone
functionality, we did observe selective epoxidation from the less
hindered face of the olefin.
(28) Woodward first postulated that if the unhindered face of an
olefin is electron-depleted, then an electron-deficient reagent might be
less likely to pass over that face. Thus, in such circumstances,
electronic and steric factors play competing roles, leading to the
formation of diastereomeric product mixtures. See: footnote, page 6, in
Woodward, R. B.; Bader, F. E.; Bickel, H.; Frey, A. J.; Kierstead, R. W.
Tetrahedron 1958, 2, 1−57.
(29) (a) Nii, K.; Tagami, K.; Kijima, M.; Munakata, T.; Ooi, T.;
Kusumi, T. Bull. Chem. Soc. Jpn. 2008, 81, 562−573. (b) Jung, M. E.;
Allen, D. A. Org. Lett. 2008, 10, 2039−2041. (c) Falrweather, K. A.;
Mander, L. N. Org. Lett. 2006, 8, 3395−3398. (d) Zhou, W.-S. Huaxue
Xuebao 1985, 43, 237−242. (e) Sudha, R.; Narasimhan, K. M.;
Waraswathy, V. G.; Sankararaman, S. J. Org. Chem. 1996, 61, 1877−
1879. (f) Ireland, R. E.; Hengartner, U. J. Am. Chem. Soc. 1972, 94,
3652−3653.
H
dx.doi.org/10.1021/ja309905j | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX