On the total synthesis and determination of the absolute configuration of rishirilide B: Exploitation of subtle effects to control the sense of cycloaddition of o-quinodimethides

Article abstract of DOI:10.1002/chem.200304931

The total synthesis of racemic rishirilide B has been accomplished. The synthesis serves to define the relative relationships of its stereogenic centers. Also, starting with readily available chiral pool, ent-rishirilide B was synthesized, thereby demonst

Full text of DOI:10.1002/chem.200304931

FULL PAPER
On the Total Synthesis and Determination of the Absolute Configuration
of Rishirilide B: Exploitation of Subtle Effects to Control the Sense of
Cycloaddition of o-Quinodimethides
Kana Yamamoto,[a] Martin F. Hentemann,[a, b] John G. Allen,[a, c] and
Samuel J. Danishefsky*[a, d]
Abstract: The total synthesis of racemic
rishirilide B has been accomplished. The
synthesis serves to define the relative
relationships of its stereogenic centers.
Also, starting with readily available
chiral pool, ent-rishirilide B was synthe-
sized, thereby demonstrating that natu-
ral configuration of rishirilide B. The
defining step in our total synthesis is the
facile cycloreversion of the bis(siloxy)-
benzocyclobutane and the intermolecu-
lar o-quinodimethide Diels Alder cy-
cloaddition. We believe that the tight
regiochemical guidance in this step aris-
es from a meshing of the electron-
donating effects of the symmetry-per-
turbing aromatic OTBS group of the o-
quinodimethide diene with the reactiv-
ity differential of the dienophile (ene-
dione), modulated by the hydroxyl
group at the a-position. The validity of
the hypothesis of hydroxy-directed acti-
vation of its vicinal ketone function in
the context of the enedione dienophile
warrants further study. This type of
activation may find broader applications
in distinguishing reactivity profiles of
key closely related functional groups in
organic substrates.
Keywords: cycloaddition
¥
fused-
ring systems ¥ quinodimethanes ¥
rishirilide ¥ total synthesis
From a structural standpoint, the rishirilides are of interest
in that they present merged aromatic and alicyclic domains.
While such arrangements need not be considered, per se,
intimidating from the standpoint of modern day synthesis,
they did prompt consideration of several issues. First, in the
light of the potential sensitivity of the resident functionality, it
seemed unlikely that the target systems would be assembled
from late-stage, classical ring-forming, electrophilic aromatic
substitution reactions. Hence, from the outset, one could
discern potentially major advantages in fashioning the aro-
matic and alicyclic sectors, concurrently, in highly advanced
forms. Moreover, the aromatic sectors (™naphtho∫ or ™ben-
zo∫) of both systems are desymmetrized by the presence of a
phenolic function. Provision would be necessary to give
appropriate guidance in fashioning the nonsymmetric naph-
thalene relative to the ™cyclohexane∫ domain of 1.
These considerations served to stimulate the thought that
intermolecular cycloaddition reactions of o-quinodimethides
might be of considerable value in enabling a convergent and
concise total synthesis of the rishirilides. Indeed, as we have
described elsewhere,[3] and will briefly summarize below, the
cycloaddition chemistry was first proposed with the rishir-
ilides well in mind. Happily, the key premises had been
realized in important model demonstrations.[3b] However,
application of the method to the rishirilide problem required,
in the end, significant improvisations beyond the proposed
Introduction
The history of the project to be related herein goes back to a
report describing the isolation of two new metabolites from
Streptomyces, OFR-1056. The structure of the more oxidized
A compound (2) was arrived at by a crystallographic deter-
mination. The formulation of the B compound, as being 1,
arose from spectroscopic data and from the presumption of
biogenetic connectivity (1 ! 2) between the two com-
pounds.[1, 2] This assumption was particularly relied upon to
suggest (but not prove) the stereochemistry of 1.
[a] Prof. Dr. S. J. Danishefsky, Dr. K. Yamamoto,
Dr. M. F. Hentemann, Dr. J. G. Allen
Laboratory for Bioorganic Chemistry
Sloan-Kettering Institute for Cancer Research
1275 York Avenue, New York, NY 10021 (USA)
Fax : (‡1)212-772-8691
E-mail: s-danishefsky@ski.mskcc.org
[b] Dr. M. F. Hentemann
Bayer Corporation West Haven, CT 06513 (USA)
[c] Dr. J. G. Allen
Eli Lilly and Company, Lilly Corporate Research Center
Indianapolis, IN 46285 (USA)
[d] Prof. Dr. S. J. Danishefsky
Department of Chemistry, Columbia University
Havemeyer Hall, New York, NY 10027 (USA)
3242
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
DOI: 10.1002/chem.200304931
Chem. Eur. J. 2003, 9, 3242 3252

3242 3252
cycloaddition reactions. The innovations required to reach
these targets may have broader consequences than the
successful attainment of the specific goal of providing a total
synthesis of rishirilide B (vide infra).
The rishirilides also evoked interest from a biological
perspective. Thus, compound 1 is a mm inhibitor of a2-
macroglobulin.[1] Since this complex protein, in turn, inhibits
the action of the fibrinolytic enzyme plasmin,[4] the rishirilides
could ultimately enhance plasmin-mediated fibrinolysis for
degradation of insoluble fibrin by suppression of plasmin
inhibition.[5] As such, it occurred to us that rishirilide
structures might be of value as functional ™small molecule∫
anti-clotting factors.[5]
Below, we document the first total synthesis of rishirilide B
(1). We also relate studies wherein target system 1 was
synthesized in an enantiomerically homogeneous form from
defined precursors. In this way, we are now able to formulate
the absolute configuration of rishirilide B. Parenthetically, we
enlarge on the concept of providing guidance to intermolec-
ular cycloadditions of o-quinodimethides through remote,
seemingly subtle, regiocontrol elements.
Scheme 1. Structures of rishirilides and the general synthetic strategy.
Results and Discussion
While our long-term goals included total syntheses of both 1
and 2, we started with the former. The hope was that perhaps
2 would be accessible by oxidative cyclization[6] (formally
hydroxylactonizaton) of 1 or of intermediates en route to 1.
Furthermore, compound 1, while chemically less interesting, is
actually more active against a2-macroglobulin than is 2.
In broad terms, the plan for addressing the total synthesis of
1 envisioned cycloaddition of o-quinodimethide 4, generated
from the fragmentation of a 1,2-trans-disubstituted benzocy-
clobutene (cf. 3), with a dienophile of the type 5.[7] At this
early conceptualization stage, we leave unspecified the nature
of the R-protecting groups in 3 and 4, the nature of the aryl-
desymmetrizing function W, the optimal g-substituents on the
dienophile (see 5) or the Y and Z group at the future C3 atom
of the rishirilide. At some point, to be determined on site, the
isoamyl group would be introduced at C4 by nucleophilic
alkylation of a keto function (Scheme 1).
Even casual inspection of this skeletal plan invites signifi-
cant questions at the level of intermediate 6 as to how the
groups, for instance, W and Z destined to emerge in rishirilide
B at C6 and C3, respectively, would be mutually coordinated.
There was scant knowledge about the way in which prefer-
ences exerted by resident functions might be exploited to
achieve tight control over the regiochemical sense of cyclo-
addition.
the presence of vicinal siloxy functions on the benzocyclobu-
tene would facilitate the cycloreversion step to generate the
reactive o-quinodimethide valence isomers 10 and 11, respec-
tively. We have already described our key discoveries
concerning this sort of cycloadditon.[3]
We found that benzocyclobutenes of the type 8 exhibit
remarkable reactivity with a range of dienophiles. Actually,
compounds in the parent series of this type (Wˆ H) are
thermochromic. The straw yellow color of these samples at
room temperature fades at lower temperature (ꢀ0 ! 208C).
NMR analysis, even of the colored form, fails to detect any
species corresponding to o-quinodimethides. Nonetheless, 8
reacts with a range of dienophiles at room temperature (cf.
maleic anhydride) and upward (cf. cyclohexenone; ca 1208C).
As had been described earlier, the chemistry of the generic
type of cycloadduct (cf. 14) has been developed in several
directions.[3a] The most widely studied modalities of progres-
sion from type 14 intermediates were 1) oxidation to naph-
thoquinone substructures and 2) twofold elimination to afford
naphthalenes (see 15 and 16, respectively, Scheme 2).
By contrast, unlike the situation which pertains with
unsubstituted system 8, the compounds bearing a peri
substituent on the benzo portion of the benzocyclobutene
(cf. 9, Wˆ OMe or OSiR3) are much less reactive. Interest-
ingly, compounds of the type 9 are colorless at room temper-
ature. Although the cycloaddition products produced, and
their apparent chemical behavior, are similar to the unsub-
stituted systems 8, the much higher temperature required for
reaction in the desymmetrizing peri-substituted cases (cf. 9) is
certainly disadvantageous. The retarding effect of the donat-
ing peri-oxy substituent may reflect attenuation of the key
cycloreversion reaction arising from abutments in the open o-
quinodimethide form with the ortho OR substituent. Alter-
However, before addressing such relatively subtle matters,
it was necessary to validate the central premise of the
feasibility of the proposed cycloadditions. While Diels Ald-
er-like reactions of o-quinodimethides with dienophiles are
well known, the usefulness of the reaction in the context of
syntheses has been confined to intramolecular cases.[8, 9]
Seeking substrates for intermolecular cases, we synthesized
benzocyclobutenes of the types 8 and 9. The thought was that
Chem. Eur. J. 2003, 9, 3242 3252
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3243

S. J. Danishefsky et al.
FULL PAPER
synthesis of idarubicinone and other goal structures.[3a] In this
report, however, we return to the initial rishirilide target.
With the background provided above, the central problem
in the rishirilide enterprise comes into sharper focus. First, to
accommodate the oxygen required eventually at C6, we would
necessarily be working with the peri-substituted and less
reactive class of benzocyclobutenes (cf. 9). Because of the
need to accommodate functionality at C2, C3, and C4,
particularly with the last two sites at quarternary carbon
atoms, it seemed unlikely that we could incorporate a
quinone-centered dienophile into our scheme. Rather, we
would have to make do with a less reactive dienophile than a
quinone (cf. 5, X ˆ H2), in the context of diene 11, for which
the diminished reactivity in the required sense was certainly
expected.
The issues, delineated above, invited solutions which
utilized more maximally activated dienophiles. In principle,
one might have supposed that the cycloaddition step could be
facilitated by recourse to Lewis acid catalysis, in the tradition
of the classical Diels Alder reaction. However, the case at
hand is novel in that a highly labile dienophile is being
generated in situ. We started with serious concerns about
whether Lewis acid catalysis would work well in such a setting.
While the matter was not surveyed exhaustively, early
attempts at catalysis in cycloreversion cycloaddition sequen-
ces were not rewarding.
Rather, we came to favor an alternative approach that took
in to account the functionality which must ulitmately be
required to reach rishrilide. As already alluded to above, it
was hoped that the isoamyl group could be introduced at a C4
ketone by ™nucleophilic alkylation.∫ We further entertained
the possibility that the C3 functionality, present at the point of
isoamylation, could direct the nucleophile to the required b-
face of C4. Extending this reasoning one step further, we
asked whether this future C3 hydroxy group might not play a
pivotal role in activating the cyclohexenone carbonyl center
and, in so doing, to enhance the directivity of the W function
implied in structure 11 (see arrows in Scheme 3). This line of
conjecture suggested evaluation of a dienophile of the type 19,
whereby a C3 hydroxyl group, vicinal to ketone, would
activate the cyclohexenone, either through hydrogen binding
or through targeted catalysis as in 20. The question is posed
implicitly in the proposed alignment, suggested by the
ensemble of 11 ‡ 19 (20), whereby we leave unresolved the
question of whether dienophilic enhancement would arise
from a local hydrogen bond or through a suitably interpo-
lated, metal-based catalyst (Scheme 4).
Scheme 2. Overview of o-quinodimethide Diels Alder chemistry.
natively, it may reflect incremental steric hindrance in the
cycloaddition step.
While the reactivity levels we had come to expect in the
unsubstituted series had been badly degraded, the regiochem-
ical patterns in the Diels Alder reactions of compound 9
were gratifyingly in the predicted direction. Thus, a peri-
methoxy or -siloxy group tends to direct the orientation in
cycloadditions with unsymmetrical dienophiles such that the
products arise from ™initial∫ bond formation at the meta-,
rather than the ortho-methide center.[2] Thus, the preferential
formation of intermediates 17 and 18 are illustrative of the
regiochemical preference governing the case of the peri-siloxy
function (Scheme 3).
Scheme 3. The directing effect of peri-substituent in o-quinodimethide
Diels Alder reaction.
As is common in synthesis directed toward provocative
natural products, methods initially devised to reach a partic-
ular structural type are apt to find application in other
pursuits, seemingly unrelated to the initial target that inspired
the initial study. So it was here. The concept of the oxygenated
o-quinodimethides, initially used for the rishirilide problem,
found excellent applications in a remarkably efficient total
Scheme 4. The hydroxy-directing effect in o-quinodimethide Diels Alder
reaction.
3244
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 3242 3252

Rishirilide B
3242 3252
We settled upon the specific dienophile 23 to launch the
program. This compound contains an ester group at C3, which
would correspond to the acid eventually required. The C2
methyl group would also be included, as would the critical C4
hydroxy group for purposes of activation (see discussion
above). Missing would be only be the means for implementing
the C1 keto group. Several possibilities (vide infra) were
entertained to deal with this problem if the basic underlying
assumptions of the plan were vindicated.
Indeed, compound 22, a possible precursor of 23 was
known, having been previously synthesized from the con-
densation of ethyl acetoacetate with crotonaldehyde.[10] The
conversion of 22 to 23 was accomplished by Rubottom-like
oxidation of its readily derived trimethylsilyl enol ether.[11]
Seemingly, only one stereoisomer was produced. At this stage
it was assumed, but not proven, that the a-methyl group had
shielded the a-face of the molecule, and that the hydroxyl-
ation product of 22 was indeed that formulated as 23.
In selecting our diene precursor for this opening study, we
initially favored a situation whereby the W blocking group
would be relatively stable, as would the two groups on the
cyclobutene moiety ultimately to be eliminated (see proto-
type step 6 ! 7 in Scheme 1). O-Methyl protection seemed
preferable to O-silyl derivatization of the oxygen atoms to be
eliminated in the formation of the naphthalene ring. Cleavage
of the silyl group in the context of Diels Alder adduct would
expose free hydroxyl groups, which could manifest other
forms of chemistry (oxidation, rearrangement) that might
compete with the bis-elimination scheme we required (see
6 ! 7). Accordingly, our sequence started with 2,3-dimethyl-
anisole (24), which was subjected to tetra side-chain bromi-
nation,[12] followed by reductive cyclization[13] with NaI to
afford 25. Subjection of the latter to the action of silver
tetrafluoroborate in methanol[14] provided 26, which was to
serve as the precursor for the o-quinodimethide system. In
this highly exploratory phase, we were willing to accept the
fact that the desired 26 was actually the minor product in the
synthesis relative to the predominant cis isomer (Scheme 5).
Scheme 6. Attempted Diels Alder reaction with 26 and dienophile 23.
a) 1608C, [D8]toluene 15 h; b) CSA, MeOH, 65%, 2 steps.
methanol units from a formal 1:1 adduct. At this stage, the
structure of this compound could not be asserted rigorously.
However, since only one product appeared to have been
produced, we assumed it to be the expected 27, arising from
the alignment contemplated in Scheme 4 (see Scheme 6). A
control experiment with TBS-protected 23 supported the
notion of the key role of the hydroxyl group in directing this
Diels Alder-like reaction. Benzocyclobutene 26 failed to
react with the TBS derivative of 23 (R ˆ TBS) (Scheme 6).
Assuming that the structure assignment is correct, we
addressed the next question, that is, the introduction of the
isoamyl group by nucleophilic alkylation of the C4 ketone.
Given the hindered nature of the ketone in 27, success could
not have been anticipated in advance. We had envisioned a
key role for the tertiary hydroxy group not only in directing an
isoamyl ogranometallic agent to the otherwise hindered
ketone, but also to its b-face, as required to reach rishirilide
B. In the event, reaction of the presumed 27 (i.e., with isoamyl
magnesium bromide (3.2 equiv)) afforded a 70% yield of a
crystalline product (m.p. 128 1298C), whose empirical for-
mula corresponds to the addition of C5H12 to 27. That this
structure indeed corresponds to 28 was rigorously established
by an X-ray determination.[3a]
It is well to reassert a scientific truism, that is, that the
attainment of a desired result following a series of experi-
ments does not necessarily establish the correctness of the
underlying rationale that led to the study. This appreciated, it
did seem that the strategic C3 alcohol function had seemingly
provided activation and directivity in both the critical cyclo-
addition nucleophilic alkylation steps en route to the bis-
protected 1-desoxyrishirilide B (28) (Scheme 7).
Scheme 7. Directed Grignard addition and unsuccessful oxidation at
benzylic position. a) Isoamylmagnesium bromide (1.08m), THF, 70%.
Scheme 5. Preparation of dienophile 23 and benzocylobutene 26. a) NaH,
EtOH; HCl(g), 36%; b) TMSCl, NaH, DMF; c) DMDO (0.066m), 42%,
2 steps; d) NBS, AIBN, hn; e) NaI, DMF, 56%, 2 steps; f) AgBF4, MeOH,
71% (cis:trans ˆ 2.5:1).
Clearly, the most efficient route to reach rishirilide B would
be to use the synthesis described above to reach an advanced
subgoal compound such as 28 and to go on from there to
™accomplish functionalization∫ at C1. While the blocking
groups of the starting materials en route to 28 had not been
selected with a view to enabling late stage global deprotec-
tion, we nevertheless used this C1 desoxy compound as a
model for exploring the proposed ™end game∫ functionaliza-
tions. Unfortunately, the results of these forays were negative.
In the event, reaction of trans-dimethoxybenzocyclobutene
26 and dienophile 23 occurred at 1608C over 15 hours
(Scheme 6). The crude cycloadduct was treated with cam-
phorsulfonic acid in methanol under reflux, providing a
compound (65% yield) corresponding to the loss of two
Chem. Eur. J. 2003, 9, 3242 3252
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3245

S. J. Danishefsky et al.
FULL PAPER
A variety of oxidation protocols, including the use of
bromosuccinimide and DDQ, as well as chromium- and
molybdenum-based oxidizing agents were unsuccessful. The
failures either took the form of no reaction, or extensive
decomposition leading to a multitude of uncharacterizable
products.
We then launched an exploratory investigation as to the
integratability of p-quinonemonoketals (cf. 30) with our o-
quinodimethide cycloaddition chemistry. In pursuing this
approach, we were mindful of the surprisingly poor perform-
ances of such quinoneketals in standard Diels Alder reac-
tions.[15] However, our hope was that systems of the type 9
would be sufficiently reactive to overcome this underempha-
sized and underappreciated problem. Also, we hoped to build
into the o-quinodimethide precursor (cf. 9), a protecting
group that would allow for liberation of the eventual phenol.
Two considerations guided our abandonment of the bis-
vicinal dimethoxy motif on the four-membered ring. First, as
mentioned earlier, our synthetic route to the required
dimethyl system gave the required trans compound only as
the minor product. We had in the interim developed an
acceptably stereoselective route to the vicinal trans-diol on
the four-membered ring (by means of the diketone, as shown
below, vide infra). Furthermore, we noted, in related model
studies, that bis-siloxy substituents on the cyclobutene are
more activating with respect to cycloaddtion chemistry than is
the corresponding trans vicinal-dimethoxy pattern. Since we
were inevitably working with the less reactive peri-substituted
benzocyclobutene series, it was particularly important to
mobilize all of the reactivity potential on the cyclobutene
moiety.
Scheme 8. Quinone-ketal Diels Alder strategy.
Scheme 9. Preparation of dienophile 36. Key: TSE ˆ 2-(Trimethylsilyl)-
ethyl. a) 1. TSEOH, Et2O, Na; 2. HCl(g), 34%; b) 1. NaH, TBSOTf, DMF;
2. DMDO (0.061m), 76%, 2 steps.
As we focused on a quality total synthesis of our goal
system 1, we needed to work out a much more selective route
to our trans-disubstituted benzocyclobutene. Earlier (see
Scheme 5), we had described a nonselective route to 26 via
25. We returned to 25, which was now treated with silver
acetate in aqueous acetic acid.[14] The crude product was
subjected to the action of sodium methoxide affording diol 37.
This compound was oxidized by a Swern protocol to produce
the benzocyclobutene 1,2-dione, 38.[2, 3b] The peri-methoxy
group was de-methylated through the action of 48% HBr, and
the resultant phenol was silylated, as shown (Scheme 10).
Indeed, it was found that use of the model benzocyclobu-
tene 29 with quinoneketal 30 did seem to provide a 1:1 adduct
formulated as 31. The thought was that the bis-elimination
product of an intermediate with useful functionality at ™pre
C1∫ might be subject to functionalization at carbons 2 and 3
(starting with conjugate addition at C2) and eventual depro-
tection at C1. Unfortunately, attempts to extend to the
cycloaddition step to encompass a cycloaddition of benzocy-
clobutene 32 with 30 were unsuccessful (Scheme 8). Once
again, we had experienced the retarding effects of a peri
substituent, this time in the context of a relatively unreactive
quinoneketal dienophile such as 30.
At this stage, we returned to the previously developed
hydroxyl-directing effect logic on the dienophile. Having
failed to accomplish late stage functionalization after the
tricyclic system (cf. 28 derived from 23) was in place, the hope
was to introduce the required handle to properly develop C1
before the cycloaddition. However, we also undertook to
build a dienophile from which the eventual acid attached to
C3 could in fact be liberated if all else were accomplished.
Similar considerations pertained to the C6 phenolic protect-
ing group.
Scheme 10. Efficient synthesis of benzocyclobutane 32. a) AgOAc, HOAc,
H2O, 93%; b) NaOMe, 55%, 2 steps; c) (COCl)2, DMSO, NEt3, 86%;
d) 48% HBr, 93%; e) TBSOTf, NEt3, 91%; f) NaBH4 then TBSOTf, NEt3,
83% (cis:trans ˆ 1:6).
With these considerations in mind, we initially targeted
compound 35 as a prospective precursor to reach a dienophile
of the required type. This compound was readily synthesized
from condensation of b-trimethylsilylethyl acetoacetate
(34)[16] with crotonaldehyde, by analogy with the chemistry
used to prepare 22 (Scheme 9).
Fortunately, careful reduction of the a-diketone with sodium
borohydride gave rise primarily to a trans-diol with high
stereoselection, presumably reflecting the stereochemical
guidance provided by the mono-reduced species. Finally,
bis-silylation of this diol afforded the silylated system 32. It
was envisioned that such compounds would serve as precur-
3246
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 3242 3252

Rishirilide B
3242 3252
sors of the trisilyloxy o-quionodimethide entering into cyclo-
addition.
role of hydroxyl-directing in this reaction. Thus Diels Alder
reaction with benzocyclobutene 32 and TBS-protected 41 did
occur, but afforded a near 1:1 mixture (Scheme 12).
Given our failure to install C1 functionality at the stage of
28, we focused on functionalizing 36 at an earlier stage, such
that the handle for the eventual C1 ketone would already be
present in the dieonophile. Fortunately, it proved possible to
achieve y-functionalization of 36 through the action of N-
bromosuccinimide (91% yield). Remarkably, neither the
bromide 39 nor its derived alcohol 40 (vide infra) underwent
cycloaddition with 32. Surprisingly, these compounds bearing
inductively activating groups proved to be less efficacious
dienophiles than the C1 unsubstituted counterpart (23).
While the reasons for the diminished reactivity exhibited by
39 and 40 are still far from clear, we sought to build a
dienophile that was sufficiently activated to undergo cyclo-
addition with the peri-substituted o-quinodimethide derived
from 32. The thought was to utilize the enedione 41. Access to
this compound was gained by means of solvolysis of bromide
39, as shown, providing diol 40, which upon Dess Martin
oxidation delivered our target dienophile 41 (Scheme 11).
Scheme 12. Final synthetic sequence to rishirilide B. Key: TSE ˆ 2-(tri-
methylsilyl)ethyl. a) toluene, 908C 12 h; b) CSA, pyiridine, MeOH; 72%
2 steps; c) isoamylmagnesium bromide (0.73m), THF; d) TAS-F, THF,
65%, 2 steps; e) CH2N2, Et2O, MeOH, 43% 45, 15% 46, 31% 1.
Scheme 11. Allylic oxidation of dienophile 36. Key: TSE ˆ 2-(trimethylsi-
lyl)ethyl; a) NBS, AIBN, CCl4, 91%; b) AgCO3, acetone, H2O, 63%;
c) Dess-Martin periodinane, CH2Cl2, 92%.
Following the precedent garnered in the des-oxy series (cf.
27 ! 28), the presumed 43 was subjected to the action of
excess isoamyl magnesium bromide. This experiment resulted
in mononucleophilic alkylation at C1, with the formation,
seemingly, of a single 1:1 addition product formulated as 44.
At this stage, we could take advantage of the removable
protecting groups we had installed at the phenolic and
carboxylic centers. Indeed, treatment of 44 with TAS-F
(TAS-F ˆ tris(dimethylamino)sulfur(trimethylsilyl)difluor-
ide) produced what we assumed to be (Æ)-rishirilde B (1;
Scheme 12).
Based solely on our own data, we were not in a position to
rigorously assert the stereochemistry at C2 C4 of our
synthetic end product. Indeed, there was some temporary
concern, since the 13C NMR spectrum obtained from our own
terminal product, presumed to be 1, did not fully correspond
to the previously published tabulated spectral data for the
natural product. No sample of naturally derived 1 was
available to allow for a direct material comparison, neither
was there a spectrum in chart form for superposition level
comparison.
Additionally temporary concern arose from another quar-
ter. Thus, treatment of fully synthetic racemic 1 with diazo-
methane in diethyl ether/methanol led to the formation of the
synthetic methyl ether/methyl ester, presumed to be 45.
Surprisingly, the high-field NMR spectrum (in chart readout)
of our fully synthetic compound was drastically different from
the spectrum (also in chart form) in studies directed to a total
synthesis of rishirilide, ostensibly, the same compound re-
ported by Hauser.[17] Barring a miss assignment, these spectra
should have been superimposable. Once again, no reference
Needless to say, these were grounds for concern as we
approached the critical cycloaddition on the basis of using
dienophile 40. It was hoped that the enedione network would
manifest greater reactivity than the failed putative dieno-
philes 39 and 40. However, even if this would be the case, a
significant issue of regiochemistry in distinguishing between
the directing effects of the two activating ketone groups would
be faced. Our governing rationale was centered around the C3
tertiary alcohol. In earlier studies (see Scheme 6) it appeared
that this hydroxyl group was activating the monoactivated
dienophile to the point where it underwent cycloaddition with
the peri-o-quinodimethide, derived from 26, in a regiospeci-
fied manner. In a similar way, it was hopedthat this hydroxyl
group would render its vicinal carbonyl group more ™direct-
ing∫ than the other (future C1) keto group, lacking hydroxyl-
centeredincremental activation . If this line of conjecture were
sustained, the preferred orientation would be that required to
reach rishirilide B.
A solution of compounds 32 and 41 in toluene was heated at
908 for 12 hours. NMR analysis indicated a relatively clean
product. While the spectrum did not rigorously establish the
structure of this crude to be 42, (stereochemistry undeter-
mined), this assignment was certainly not inconsistent, in any
way, with the data. Treatment of the presumed 42 with
camphorsulfonic acid in pyridine afforded a 72% yield of a
compound formulated as 43. Furthermore, once again a
control experiment with TBS derivative of 41 supported the
Chem. Eur. J. 2003, 9, 3242 3252
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3247

S. J. Danishefsky et al.
FULL PAPER
material was available for direct comparison with our
synthetic product.
carried through the five-step procedure described above to
furnish (‡)-41 (Scheme 13).
This state of confusion was much relieved, when we were
able to obtain a chart spectrum of that corresponding to
enantiopure compound prepared some time ago from the bis-
methylation of naturally occurring rishirilide B (1). The high-
field 1H NMR spectrum of our synthetic racemic 45 was
identical with the spectrum received from Fukuyama. Hence,
at this stage, we could be confident that our compound did
correspond to naturally occurring rishirilide B, and that the
claimed total synthesis product[17] is not rishirilide B methyl
ether/methyl ester (i.e., 45).
Having asserted this, it is important to take note that our
total synthesis, per se, to this stage did not rigorously establish
the actual relative configurational relationship in synthetic
rishirilide B to be that which we supposed. Formally, there still
remained the possibility that one or more errors had been
incorporated into the structure assignment of 1. Correspond-
ingly, there could have been one or more faults in our
assignments during the unfolding of our synthesis. Put differ-
ently, the fact that two compounds are identical does not
establish the correctness of their structure assignments.
As it turned out, these concerns could also be laid to rest.
Confidence was restored when our fully synthetic racemic 1
had been treated with diazomethane. In addition to the bis-
methylation obtained in 33% yield (and assigned as 45) as
discussed above, a mono-methylation product (i.e., 46), a
methyl ester of phenol (m.p. 92 948), was also obtained. A
crystallographic determination of this compound clearly
showed it to be 46 thus corroborating our assignments in full
(relative) stereochemical detail.[3a] Hence, the structure of
rishirilide B does indeed corresponds to 1, and rishirlide B
methyl ether/methyl ester is safely represented as 45. Re-
maining to be sorted out is the structure and stereochemistry
of the previously claimed total synthesis product,[17] which
cannot be 45.
With the gross structure and relative stereochemistry of
rishirilide B now secure, we addressed the question as to how
chemical synthesis might also serve to establish its absolute
configuration. This could be achieved by recourse to a starting
material of established absolute stereochemistry and proceed-
ing to reach the end target through steps that are stereo-
chemically unambiguous. Given availability of the natural
product, comparison between synthetic and naturally derived
material would be optimal. Even in the absence of an actual
reference sample, comparisons of the corresponding discern-
ing properties of synthetic and naturally derived material
could provide the answer. In the case at hand, there was no
naturally derived rishirilide available for direct comparison.
Fortunately, however, the measurement of the optical rotation
of natural rishirilide B has been recorded.[1]
We focused on synthesizing an enantiomerically defined
version of dienophile 41. In this effort we revisited with
modification, its synthesis in an enantio-defined context.
Thus, commercially available (R)-methylcyclohexanone (47),
derived in an unambiguous way from pulegone, was converted
to the known enone 48 by oxidation with 2-iodoxybenzoic
acid (IBX).[18] Subsequently, a'C-acylation with 2-(trimethyl-
silyl)ethoxycarboxy cyanide[19] furnished (À)-34, which was
Scheme 13. Preparation of enantiomerically pure enedione 41. Key:
TSE ˆ 2-(trimetylsilyl)ethyl. a) 2-Iodoxybenzoic acid (IBX), 658C, DMSO,
PhMe, 36%; b) iPr2NLi, 2-(trimethylsiyl)ethoxycarboxy cyanide, Et2O,
39%; c) 1. NaH, TBSOTf; 2. DMDO, acetone, 28%, 2 steps; d) 1. AIBN,
NBS, CCl4; 2. AgCO3, acetone, H2O, 49%, 2 steps; e) Dess-Martin period-
inane, CH2Cl2, 98%.
With dienophile (‡)-41 in hand, the completion of the
synthesis was accomplished analogously to that used in (Æ)-1.
Cycloadditon of 32 with (‡)-41 was followed by elimination,
nucleophilic isoamylation, and deprotection (Scheme 14).
Scheme 14. Completion of the synthesis of ent-rishirilide B. Key: TSE ˆ 2-
(trimetylsilyl)ethyl a) 1. 908C, PhMe; 2. ppts, MeOH, 65%, 2 steps; b) i-
pentylmagnesium bromide, THF, 73%; c) TAS-F, THF, 43%.
(À)-rishirilide B ([a]2D5 ˆ À13.6 (c ˆ 0.320 in EtOH)) was thus
produced in the laboratory. The literature reports pure
22
rishirilide B to have an optical rotation of [a] ˆ ‡12.8
DE
(c ˆ 0.488 in EtOH). Based on this enantiospecific synthesis,
and assuming the correctness of the literature and polaro-
graphic measurement, the natural enantiomer is that shown in
structure 1 (Scheme 1). Our fully synthetic optically pure
structure is thus (À)-1 corresponding to ent-rishirilide B.
Conclusion
The total synthesis of racemic rishirilide B has been accom-
plished. The synthesis serves to define rigorously, for the first
time, the relative relationships of its stereogenic centers. Also,
starting with (R)-3-methylcyclohexanone ent-rishirilide B
((À)-1) was synthesized, thereby demonstrating that natural
rishirilide B is as shown in structure (‡)-1 (Scheme 1).
The defining step in our total synthesis is the cycloreversion
of 32 and cycloaddition of its o-quinodimethide form with 41.
We believe that the tight regiochemical guidance in this step
arises from a meshing of the electron-donating effects of the
symmetry-perturbing aromatic OTBS group, with the reac-
tivity differential of the enedione modulated by the hydroxyl
group adjacent to the ketone at the future C1. We note that
3248
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 3242 3252

Rishirilide B
3242 3252
heating the mixture at reflux for 4 h, the solvent was evaporated and the
residue was purified by column chromatography (9:1 ! 4:1 ! 2:1 hexanes/
EtOAc) to give pure 27 as a colorless oil (0.41 g, 65%). IR (film): n ˆ 3472,
this hydroxy group, at the stage of compound 43, also directs
delivery of the isoamyl Grignard reagent to the required b-
face of the C4 ketone.
The validity of the hypothesis of hydroxy-directed activa-
tion of its vicinal ketone function in the context of the
enedione dienophile warrants further study. This type of
activation may find broader applications in distinguishing
reactivity profiles of key closely related functional groups in
organic substrates.
1
2933, 1744, 1682, 1624, 1463, 1240 cmÀ1; H NMR (400 MHz, CDCl3): d ˆ
9.05 (s, 1H), 8.60 (s, 1H), 7.47 (dd, J ˆ 7.92 Hz, 1H), 7.33 (d, J ˆ 8.30 Hz,
1H), 6.75 (d, J ˆ 7.65 Hz, 1H), 4.38 (s, 1H), 4.20 4.05 (m, 2H), 3.99 (s, 3H),
3.30 (dd, J ˆ 13.67, 16.92 Hz, 1H), 3.09 (dd, J ˆ 5.05, 16.89 Hz, 1H), 2.51 (dt,
J ˆ 5.51 Hz, 1H), 1.27 (d, J ˆ 6.68 Hz, 1H), 1.12 ppm (d, J ˆ 7.07 Hz, 3H);
13C NMR (125 MHz, CDCl3): d ˆ 195.73, 169.42, 157.15, 138.69, 137.36,
129.80, 127.52, 126.26, 125.25, 124.20, 119.22, 103.70, 81.56, 61.94, 55.54,
39.12, 34.30, 15.90, 14.06 ppm; HRMS (FAB ‡ ) calcd for C19H20O5:
‡
328.1311 [M ]; found: 328.1310.
Isoamyl magnesium bromide addition (28): The anthracene 27 (100 mg,
0.305 mmol) was dissolved in THF (3 mL) and isoamyl magnesium
bromide (1.08m Et2O, 0.91 mL, 0.91 mmol) was added dropwise over 2 h.
The mixture was partitioned between Et2O and sat. NaCl(aq), 1n HCl, and
sat. NaHCO3(aq), and the organic layer was dried (MgSO4) and evaporated.
The residue was purified by chromatog-
Experimental Section
General: All moisture- or air-sensitive reactions were performed under an
N2 or Ar atmosphere in oven-dried glassware. Unless otherwise noted,
extracts were dried with anhydrous MgSO4 and concentrated using a rotary
evaporator at aspirator pressure. All melting points were determined using
an Electrothermal 9100 capillary melting point apparatus and are uncor-
rected. Solvents used in moisture-sensitive reactions were dried using
standard methods. Reagents and starting materials were obtained from
commercial suppliers, and used without further purification unless other-
wise indicated. Flash chromatography was performed according to the
methods of Still, with silica gel (230 400 mesh) obtained from EM science
as the stationary phase. The term ™deactivated SiO2∫ refers to silica gel that
is pre-treated with 1% NEt3/hexanes for more than 30 min and washed
with hexanes prior to use. Purity of isolated compounds was assessed based
on examination of the 1H NMR spectroscopy and corresponding mass
spectral data. IR spectra were recorded on Perkin Elmer 1600 series FTIR
spectrometer from thin films on NaCl plates. 1H NMR spectra were
obtained with Bruker spectrometers at 400 or 500 MHz. 13C spectra were
obtained with Bruker instruments at 100 or 125 MHz. Spectra chemical
shifts were calibrated to the residual protio solvent resonance (CDCl3, 1H:
d ˆ 7.26 ppm; 13C: d ˆ 77.0 ppm). Low-resolution mass spectra were
determined with a PESciex Ap 130 spectrometer. High-resolution mass
spectra were determined by the University of California, Riverside MS
Facility, Riverside, CA 92521.
raphy (9:1 ! 4:1 hexanes/EtOAc) to
give the protected core 28 as a white
solid (85 mg, 70%). M.p. 128 1298C
(needles from Et2O-hexanes); IR
(film): n ˆ 3488, 2954, 2868, 1715, 1689,
1574, 1462, 1246 cmÀ1
;
1H NMR
(400 MHz, CDCl3): d ˆ 8.21 (s, 1H),
7.41 (s, 1H), 7.27 7.23 (m, 2H), 6.90
6.60 (m, 1H), 4.02 (s, 1H), 4.00 3.75
(m, 2H), 3.93 (s, 3H), 3.31 (dd, J ˆ 8.50, 17.91 Hz, 1H), 2.82 (dd, J ˆ 10.22,
17.89 Hz, 1H), 2.60 2.40 (m, 1H), 2.31 (brs, 1H), 2.08 (dt, J ˆ 3.80,
13.70 Hz, 1H), 1.65 1.50 (m, 1H), 1.30 1.20 (m, 2H), 1.02 (d, J ˆ 6.71 Hz,
3H), 0.95 0.76 (m, 1H), 0.86 (d, J ˆ 6.59 Hz, 3H), 0.74 (d, J ˆ 6.24 Hz,
3H), 0.70 0.65 ppm (m, 3H); 13C NMR (125 MHz, CDCl3): d ˆ 197.37,
173.47, 151.68, 138.32, 133.30, 129.80, 127.85, 126.57, 122.75, 120.12, 115.39,
83.85, 78.10, 65.13, 48.26, 35.71, 31.16, 28.29, 25.84, 22.74, 22.43, 18.41, 16.67,
‡
10.04, À1.91, À4.12, À4.39; MS: m/z: 423 [C24H32O5Na ].
trans-Bis(tert-butyldimethlsilyloxy)-3-methoxybenzocyclobutene:
Di-
methylanisole tetrabromide[12] (33 g, 60 mmol) and NaI (36 g, 239 mmol)
were heated together in DMF (240 mL) at 808C for 12 h.[13] The black
mixture was diluted with sat. NaCl(aq)
Cyclohexenone (23): NaH (35 mg, 60%, 1.47 mmol) was washed with
pentane and taken up in DMF (4 mL). 2-Carboxyethyl-3-methyl-cyclohex-
5-en-1-one[11] (0.22 g, 1.23 mmol) was dissolved in DMF (4 mL) and
transfered through a cannula at 08C into the NaH suspension. After
20 min the mixture was cooled to À788C, and TBSCl (0.22 g, 1.47 mmol)
was added in DMF (1 mL) through a
(250 mL) and extracted with pentane
(3 Â 500 mL). The pentane extracts
were dried (MgSO4) and evaporated,
and the residue was passed through a
column of silica gel (95:5 hexanes/
EtOAc) to give pure dibromide (8.5 g, 53%). 1H NMR (400 MHz, CDCl3):
d ˆ 7.38 (dd, J ˆ 7.97 Hz, 1H), 6.83 (d, J ˆ 8.46 Hz, 1H), 6.77 (d, J ˆ
7.42 Hz), 5.49 (s, 1H), 5.41 (s, 1H), 3.56 ppm (s, 3H).
cannula. After 20 min the mixture was
poured into pentane, washed with water
and NaCl(aq)
, dried (MgSO4), and
evaporated. The residue was taken up
in CH2Cl2 (10 mL) and cooled to
The dibromide (8.5 g, 32 mmol) was dissolved in acetic acid (158 mL) and
water (6.3 mL), and silver acetate (11.6 g, 70 mmol) was added. The
mixture was then heated (908C) 16 h in the dark[14] and then cooled in an
ice bath; sat. NaCl (aq) (100 mL) was added then added. The precipitate was
removed by filtration through Celite, and the filtrate was partitioned
between CH2Cl2 and sat. NaCl(aq), neutralized (NaHCO3), dried (MgSO4),
and evaporated. The residue was dissolved in MeOH (200 mL) and
NaOMe (25% in MeOH, 0.5 mL) was added at 08C.[14] After 5 h, the
mixture was partitioned between Et2O and 1n NH4Cl, the organic layer was
dried (MgSO4) and evaporated, and the residue was purified by recrystal-
lization (CH2Cl2/EtOAc/hexanes) to give cis-diol (2.9 g, 55%).
À788C and dimethyldioxirane (prepared according to the standard
procedure[20] from OXONETM, but without contact with the drying agent:
32 mL, 0.063m in acetone, 2 mmol) was added dropwise over 20 min. After
warming to room temperature over 4 h, most of the solvent was
evaporated, and the concentrate was partitioned between Et2O and
NaCl(aq). The organic layer was dried (MgSO4) and evaporated, and the
residue was purified by column chromatography (9:1 ! 4:1 hexanes/
EtOAc) to give pure 23 as a colorless oil (0.22 g, 78%): IR (film): nÄ ˆ
1
3484, 2982, 2934, 1726, 1679, 1461 cmÀ1; H NMR (400 MHz, CDCl3): d ˆ
7.11 7.00 (m, 1H), 6.13 (dd, J ˆ 2.92, 10.23 Hz, 1H), 4.26 4.15 (m, 2H),
4.14 (s, 1H), 2.65 2.57 (m, 1H), 2.50 2.40 (m, 1H), 2.36 2.28 (m, 1H),
1.25 (t, J ˆ 7.02 Hz, 3H), 1.14 ppm (d, J ˆ 6.73 Hz, 3H); 13C NMR
(125 MHz, CDCl3): d ˆ 195.68, 169.06, 152.91, 126.52, 80.92, 62.01, 38.79,
Methyl sulfoxide (5.5 mL, 77.4 mmol) was carefully added to oxalyl
chloride (19.4 mL, 38.7 mmol) at À788C in CH2Cl2 (30 mL). The cis-diol
(2.9 g, 17.6 mmol) was dried by azeotropic distillation with benzene and
cannulated into the oxalyl chloride solution in CH2Cl2 (60 mL), followed by
Et3N (24.5 mL, 176 mmol) added in three portions. The mixture was stirred
for 1 h and then washed with 1n HCl (3 Â 100 mL) and NaHCO3 (2 Â
100 mL), the organic extracts were dried (MgSO4) and evaporated, and the
residue recrystallized (EtOAc/hexanes) to give pure 3-methoxybenzocy-
clobutenedione (2.5 g, 86%). 1H NMR (400 MHz, CDCl3): d ˆ 7.69 (dd,
J ˆ 7.81 Hz, 1H), 7.58 (d, J ˆ 7.28 Hz, 1H), 7.13 (d, J ˆ 7.97 Hz), 4.24 (s, 1H),
5.41 (s, 1H), 3.56 ppm (s, 3H).
‡
33.03, 15.26, 14.06 ppm; MS: m/z: 221 [C10H14O4Na ].
Diels Alder reaction and aromatization (27): The enone 23 (0.38 g,
1.90 mmol) and benzocyclobutene 26 (0.37 g, 1.90 mmol) were dissolved in
[D8]toluene (1 mL) and heated at
1608C for 15 h in a sealed tube. The
solvent was evaporated, the residue was
dissolved in MeOH (10 mL), and CSA
(10 mg, 43 mmol) was added. After
Chem. Eur. J. 2003, 9, 3242 3252
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3249

S. J. Danishefsky et al.
FULL PAPER
3-Methoxybenzocyclobutenedione (200 mg, 1.23 mmol) was dissolved in
methanol (10 mL) and cooled to 08C, and NaBH4 (33 mg, 0.87 mmol) was
added. After 1 h, the solvent was removed at 08C, the residue was passed
through a plug of silica with cold EtOAc, and the filtrate was concentrated.
The residue was dissolved with CH2Cl2 (10 mL), the mixture cooled to
À788C, and TBSOTf (1.13 mL, 4.92 mmol) and Et3N (0.68 mL, 4.92 mmol)
were added. After 2 h, methanol (1 mL) was added. The organic solution
was washed with NH4Cl(aq) and NaCl(aq), dried (MgSO4), and evaporated.
Column chromatography (hexanes ! hexanes/EtOAc 95:5) gave the pure
trans compound as a colorless oil (338 mg, 70%). FTIR (film): n ˆ 2927,
2855, 1606, 1584, 1482 cmÀ1; 1H NMR (400 MHz, [D8]toluene): d ˆ 7.24 (dd,
J ˆ 7.70 Hz, 1H), 6.95 (d, J ˆ 7.17 Hz, 1H), 6.87 (d, J ˆ 8.25 Hz, 1H), 5.18 (s,
1H), 5.09 (s, 1H), 3.78 (s, 3H), 1.15 (s, 9H), 1.11 (s, 9H), 0.38 (s, 3H), 0.32 (s,
3H), 0.30 (s, 3H), 0.26 ppm (s, 3H); 13C NMR (125 MHz, CDCl3): d ˆ
155.6, 145.9, 131.0, 127.6, 115.2, 115.1, 79.3, 79.0, 56.9, 25.8, 18.0, À3.9, À4.5,
38.86, 33.08, 17.52, 15.31, À1.63 ppm; HRMS (DCI/NH3) calcd for
‡
C13H26NO4Si: 288.1631 [M ‡NH3]; found: 288.1638.
Hydroxy enone ((‡)-36): This compound was prepared in a similar manner
as (Æ)-36: [a]2D5 ˆ ‡22.74 (c ˆ 0.73, CHCl3); HRMS (DCI/NH3) calcd for
‡
C13H26NO4Si: 288.1631 [M ‡NH3]; found: 288.1634.
Hydroxy enedione (41): A solution of 36 (0.54 g, 2.0 mmol), NBS (0.39 g,
2.2 mmol), AIBN (16 mg, 0.1 mmol) in CCl4 (20 mL) was degassed with Ar
and then heated at reflux for 1 h. The mixture was cooled, pentane (50 mL)
was added and filtered through Celite
to remove succinimide. The filtrate was
evaporated and the reside was purified
by column chromatography (9:1 ! 4:1
hexanes/EtOAc) to give pure g-bro-
mide 39 as a white solid (0.64 g, 91%).
‡
IR (film): n ˆ 3475, 2953, 2898, 1741,
À4.7, À5.1 ppm; HRMS (NH3/CI): calcd for C27H43O5Si2: 394.2359 [M ];
1725, 1692, 1250, 837 cmÀ1 1H NMR
;
found: 394.2346.
(400 MHz, CDCl3): d ˆ 7.11 (dd, J ˆ
Crotonaldehyde condensation (35):[11] 2-(Trimethylsilyl)ethyl acetoace-
tate[16] (5 g, 24.8 mmol), 2-(trimethylsilyl)ethanol (7 mL, 49.6 mmol),
crotonaldehyde (2 mL, 24.8 mmol) were dissolved in Et2O (120 mL) and
cooled to 08C. Sodium (25 mg) was added and the mixture was stirred for
12 h. Hydrogen chloride was bubbled
2.35, 10.24 Hz, 1H), 6.02 (dd, J ˆ 2.10, 10.14 Hz, 1H), 5.00 (ddd, J ˆ 2.13,
9.42 Hz, 1H), 4.22 (s, 1H), 4.20 4.14 (m, 2H), 2.60 2.52 (m, 1H), 1.30 (d,
J ˆ 6.68 Hz, 3H), 1.0 0.91 (m, 2H), À0.02 (s, 9H); 13C NMR (125 MHz,
CDCl3): d ˆ 193.99, 168.44, 152.06, 125.04, 81.47, 65.19, 50.28, 48.33, 17.35,
13.48, À1.69; HRMS (DCI/NH3) calcd for C13H25NO4SiBr: 366.0736
‡
through the solution for 5 min at 08C
and left to stir for 4 h. The solvent was
evaporated and the residue was purified
[M ‡NH3]; found: 366.0721.
The bromide 39 (0.64 g, 1.82 mmol) was dissolved in acetone/water (8:1,
36 mL), AgCO3 (5 g, 18.2 mmol) was added, and the mixture stirred in the
dark for 24 h. The mixture was cooled to 08C, NaCl(aq) (20 mL) was added,
and the precipitate was removed by filtration through Celite. The filtrate
was concentrated to ca. 25 mL and partitioned between CH2Cl2 and
NaCl(aq). The organic layer was dried (MgSO4) and evaporated, and the
reside was purified by column chromatography (4:1 hexanes/EtOAc) to
give pure g-alcohol as a white solid (0.33 g, 63%). M.p. 107.4 107.88C
(needles from hexanes/EtOAc); IR (film): n ˆ 3484, 2952, 2906, 1701, 1682,
by column chromatography (9:1 ! 4:1
hexanes/EtOAc) to give pure 35 as a
colorless oil (2.1 g, 34%); IR (film): n ˆ
2955, 2898, 1738, 1680, 1249, 838 cmÀ1
;
1H NMR (400 MHz, CDCl3): d ˆ
7.00 6.90 (m, 1H), 6.03 (ddd, J ˆ 0.93, 2.57, 10.17 Hz, 1H), 4.25 (dd, J ˆ
8.17 Hz, 2H), 3.06 (d, J ˆ 11.65 Hz, 1H), 2.50 2.45 (m, 2H), 2.15 2.05 (m,
1H), 1.04 (d, J ˆ 9.96 Hz, 3H), 1.05 0.98 (m, 2H), 0.01 (s, 9H); 13C NMR
(125 MHz, CDCl3): d ˆ 194.57, 170.06, 149.68, 128.79, 63.36, 61.75, 33.07,
1254, 837 cmÀ1
;
1H NMR (400 MHz, CDCl3): d ˆ 7.12 (dd, J ˆ 4.91,
‡
32.79, 19.77, 17.28, À1.55 ppm; MS: m/z: 277 [C13H22O3SiNa ].
10.11 Hz, 1H), 6.13 (d, J ˆ 10.30 Hz, 1H), 4.40 (d, J ˆ 11.82 Hz, 1H),
4.30 4.18 (m, 2H), 4.16 (s, 1H), 2.37 (dq, J ˆ 7.03, 12.99 Hz, 1H), 1.24 (d,
J ˆ 7.08 Hz, 3H), 1.05 0.95 (m, 2H), 0.02 ppm (s, 9H); 13C NMR
(125 MHz, CDCl3): d ˆ 194.98, 171.67, 152.00, 124.94, 81.35, 66.15, 40.73,
17.22, 11.20, À1.62 ppm; HRMS (DCI/NH3) calcd for C13H26NO5Si:
6-(Trimethylsilylcarbetoxy)-5-(R)-methylcyclohexenone ((À)-35): This
compound was prepared based on the literature procedure.[21] nBuLi
(21 mL, 31 mmol) was added to a cold (08C) solution of iPr2NH (4.7 mL,
3.4 g, 34 mmol) in Et2O (110 mL), and the mixture was stirred for 10 min.
The resulting pale yellow solution was cooled to À788C, and enone 48
(5.2 g, 28 mmol) was added over 10 min, followed by 2-(trimethylsilyl)-
ethoxycarboxy cyanide (7.2 g, 80% purity, 34 mmol) over 30 min with a
syringe pump. After 30 min of stirring, water (200 mL) was added, and the
aqueous layer was extracted with Et2O (3 Â 80 mL), combined and
concentrated. The residue was purified by chromatography with silica gel
(1:19 ! 1:9 hexanes/EtOAc) to give pure (À)-35 as a clear oil (2.4 g,
9.5 mmol, 34%). [a]2D5 ˆ À59.71 (c ˆ 2.76 in CHCl3); MS: m/z: 277
‡
304.1580 [M ‡NH3]; found: 304.1584.
The alcohol (0.10 g, 0.35 mmol) was dissolved in CH2Cl2 (10 mL) at 08C,
Dess Martin reagent (0.44 g, 1.05 mmol) was added in three portions, and
the mixture stirred in the dark for 6 h. The mixture was partitioned between
CH2Cl2 and Na2S2O3/NaHCO3, the organic layer was dried (MgSO4) and
evaporated, and the reside was purified by column chromatography (4:1
hexanes/EtOAc) to give pure 29 as a white solid (91 mg, 92%). M.p. 108.5
108.88C (needles from hexanes/EtOAc); IR (film): n ˆ 3464, 2952, 2896,
1728, 1728, 1690, 1251, 839 cmÀ1; 1H NMR (400 MHz, CDCl3): d ˆ 6.96 (d,
J ˆ 10.33 Hz, 1H), 6.80 (d, J ˆ 10.35 Hz, 1H), 4.28 (s, 1H), 4.23 4.12 (m,
2H), 3.01 (q, J ˆ 6.51 Hz, 1H), 1.25 (d, J ˆ 6.49 Hz, 3H), 1.00 0.85 (m,
2H), À0.02 ppm (s, 9H); 13C NMR (125 MHz, CDCl3): d ˆ 195.12, 194.23,
167.84, 144.35, 136.08, 84.38, 66.26, 51.81, 17.34, 8.63, 6.12, À1.69 ppm;
‡
[C13H22O3SiNa ].
Hydroxy enone (36): NaH (0.43 g, 60%, 10.9 mmol) was washed with
pentane and taken up in DMF (20 mL). Enone 35 (2.30 g, 9.06 mmol) was
dissolved in DMF (10 mL) and transferred with a cannula at 08C into the
NaH suspension. After 20 min the mixture was cooled to À788C, and
TBSOTf (2.50 mL, 10.9 mmol) was
‡
HRMS (CDI/NH3) calcd for C13H24NO5Si: 302.1423 [M ‡NH3]; found:
302.1414.
added. After 20 min the mixture was
poured into pentane; then this mixture
Hydroxy enedione ((‡)-41): This compound was prepared in a similar
manner to the racemate of hydroxy enedione 41. M.p. 118 118.28C (white
flakes from hexanes/EtOAc); [a]2D5 ˆ ‡52.09 (c ˆ 0.86 in CHCl3); MS: m/z:
was washed with water and NaCl(aq)
,
dried (MgSO4), and evaporated. The
residue was taken up in acetone
(25 mL) and cooled to À788C and
306.9 [C13H24NO5Si‡Na ]. (‡)-g-bromide: [a]2D5 ˆ ‡175.7 (c ˆ 1.99 in
‡
‡
CHCl3); MS: m/z: 370.9 [C13H25NO4SiBr‡Na ]. (À)-g-alcohol: [a]D25
ˆ
‡
À93.40 (c ˆ 1.03 in CHCl3); MS: m/z: 309.1 [C13H26NO5Si‡Na ].
dimethyldioxirane (prepared according to the standard procedure,[20] but
without contact with drying agent: 148 mL, 0.61m in acetone, 90 mmol) was
added dropwise over 20 min. After warming to room temperature over 4 h,
the mixture was concentrated and the residue was partitioned between
Et2O and NaCl(aq), the organic layer was dried (MgSO4) and evaporated,
and the residue was purified by column chromatography (9:1 ! 4:1
hexanes/EtOAc) to give pure 36 as a colorless oil (1.86 g, 76%). IR (film):
cis/trans-Tris(tert-butyldimethylsilyloxy)benzocyclobutene (32): 3-Meth-
oxy benzocyclobutenedione 38 (1 g, 6.17 mmol) was heated at 110
1308C in 48% HBr for 7 h. The mixture was cooled, partitioned between
sat. NaCl(aq) and EtOAc, the organic layers were dried (MgSO4) and
evaporated, and the residue was purified by chromatography (4:1 hexanes/
EtOAc) to give pure phenol as a white
n ˆ 3484, 2955, 2899, 1735, 1684, 1250, 838 cmÀ1
;
1H NMR (400 MHz,
solid (0.92 g, 93%). 1H NMR
CDCl3): d ˆ 7.10 7.00 (m, 1H), 6.12 (ddd, J ˆ 2.16, 9.50 Hz, 1H), 4.23 4.15
(m, 2H), 4.14 (s, 1H), 2.65 2.55 (m, 1H), 2.50 2.40 (m, 1H), 2.36 2.28
(m, 1H), 1.13 (d, J ˆ 6.71 Hz, 3H), 1.0 0.94 (m, 2H), 0.00 ppm (s, 9H);
13C NMR (125 MHz, CDCl3): d ˆ 195.8, 169.25, 152.98, 126.56, 80.96, 64.57,
(400 MHz, CDCl3): d ˆ 7.65 (dd, J ˆ
7.66 Hz, 1H), 7.57 (d, J ˆ 7.35 Hz,
1H), 7.03 (d, J ˆ 7.76 Hz), 1.00 (s,
9H), 0.29 ppm (s, 6H).
3250
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 3242 3252

Rishirilide B
3242 3252
The phenol (0.78 g, 5.27 mmol) was suspended in CH2Cl2 (25 mL) at 08C,
and Et3N (0.95 mL, 6.85 mmol) and TBSCl (0.95 g, 6.32 mmol) were added.
After stirring for 30 min, the mixture was partitioned between NaCl(aq) and
CH2Cl2, the organic layer was dried (MgSO4) and evaporated, and the
residue was purified by column chromatography (95:5 hexanes/EtOAc) to
give pure (tert-butyldimethylsilyloxy)benzocyclobutenedione as a white
solid (1.23 g, 91%). IR (film): n ˆ 2927, 2858, 1798, 1767, 1309, 1126,
789 cmÀ1; 1H NMR (400 MHz, CDCl3): d ˆ 6.65 (dd, J ˆ 7.66 Hz, 1H), 7.57
(d, J ˆ 7.35 Hz, 1H), 7.03 (d, J ˆ 7.76 Hz, 1H), 1.00 (s, 9H), 0.28 ppm (s,
6H); 13C NMR (125 MHz, CDCl3): d ˆ 195.13, 190.82, 172.38, 163.89,
150.52, 138.20, 126.26, 114.69, 25.35, 18.25, À4.93 ppm; HRMS (DCI/NH3)
1H), 6.95 (d, J ˆ 7.12 Hz, 1H), 4.01 (s, 1H), 3.90 3.75 (m, 2H), 3.10 (q, J ˆ
6.78 Hz, 1H), 2.39 (s, 1H), 2.30 (ddd, J ˆ 3.36, 12.94 Hz, 1H), 1.68 (ddd, J ˆ
5.00, 13.62 Hz, 1H), 1.45 1.28 (m, 2H), 1.29 (d, J ˆ 6.74 Hz, 3H), 1.09 (s,
9H), 0.95 0.76 (m, 1H), 0.79 (d, J ˆ 6.35 Hz, 3H), 0.67 (d, J ˆ 6.44 Hz,
3H), 0.35 0.25 (m, 2H), 0.26, 0.24 (2s, 3H ea.), À0.23 ppm (s, 9H);
13C NMR (125 MHz, CDCl3): d ˆ 197.37, 173.47, 151.68, 138.32, 133.30,
129.80, 127.85, 126.57, 122.75, 120.12, 115.39, 83.85, 78.10, 65.13, 48.26, 35.71,
31.16, 28.29, 25.84, 22.74, 22.43, 18.41, 16.67, 10.04, À1.91, À4.12,
‡
À4.39 ppm; HRMS (FAB ‡ ) calcd for C32H50O6Si2Na: 609.3038 [M ‡Na];
found 609.3013.
Protected (À)-rishirilide B ((À)-44): This compound was prepared in a
‡
calcd for C14H18O3Si: 262.1025 [M ]; found 262.1034.
similar manner to (Æ)-44: [a]2D5 ˆ À5.22 (c ˆ 1.65 in CHCl3).
The silyloxydione (0.10 g, 0.38 mmol) was dissolved in MeOH (5 mL) at
08C and NaBH4 (10 mg) was added. After 1.5 h, acetone was added, and
the mixture was evaporated at 08C. The residue was filtered through a plug
of silica with EtOAc, and the filtrate was evaporated at 08C. The residue
was dissolved in CH2Cl2 (5 mL) and cooled to À788C, and Et3N (0.13 mL,
0.95 mmol) and TBSOTf (0.22 mL, 0.95 mmol) were added. After 1 h, the
mixture was evaporated, and the residue was purified by column
chromatography (hexanes) to give 32 as a mixture of trans and cis isomers
(colorless oil, 6:1 trans:cis, 0.15 g, 83%). The major trans isomer was
(Æ)-Rishirilide B (1): The 2-(trimethylsilyl)ethyl ester (22 mg, 38 mmol)
was dissolved in THF (1 mL), TAS-F (70.5 mg, 0.23 mmol) was added
under an atmosphere of nitrogen. After 30 min, the mixture was poured
onto a column of silica gel. Elution
(CH2Cl2 ! 9:1 CH2Cl2/MeOH ! 9:1
CH2Cl2/MeOH, 0.1% HOAc) and
evaporation gave a residue that was
applied in MeOH to a LiChroprep
C-18 column that was eluted (H2O !
characterized as follows: IR (film): n ˆ 2955, 2929, 2857, 1601, 1437 cmÀ1
;
3:1 H2O/MeOH ! 1:1 H2O/MeOH) to
1H NMR (400 MHz, CDCl3): d ˆ 7.18 (dd, J ˆ 7.81 Hz, 1H), 6.83 (d, J ˆ
7.19 Hz, 1H), 6.70 (d, J ˆ 8.18 Hz, 1H), 4.96 (s, 1H), 4.85 (s, 1H), 0.98 (s,
9H), 0.97 (s, 9H), 0.93 (s, 9H), 0.24 (s, 3H), 0.19 (s, 3H), 0.16 (s, 3H), 0.16
(s, 3H), 0.15 (s, 3H), 0.01 ppm (s, 3H); 13C NMR (125 MHz, CDCl3): d ˆ
169.39, 151.30, 146.10, 130.71, 119.84, 116.05, 79.37, 78.65, 25.94, 25.82,
25.71, 18.35, 18.17, 18.04, À2.93, À3.96, À4.04, À4.10, À4.55, À4.65 ppm;
give pure 1 as a beige solid (13 mg,
93%). IR (film): n ˆ 3399, 2954, 2868,
1677, 1625, 1571, 1451, 1277 cmÀ1
;
1H NMR (400 MHz, [D4]MeOH): d ˆ
8.41 (s, 1H), 8.38 (s, 1H), 7.43 (d, J ˆ
8.22 Hz, 1H), 7.28 (dd, J ˆ 7.93 Hz, 1H), 3.10 3.00 (s, 1H), 2.35 (dd, J ˆ
10.31 Hz, 1H), 1.73 (dd, J ˆ 8.96 Hz, 1H), 1.55 1.40 (m, 1H), 1.40 1.20
(m, 5H), 0.83 (d, J ˆ 6.54 Hz, 3H), 0.72 ppm (d, J ˆ 6.55 Hz, 3H); 13C NMR
(125 MHz, DMSO): d ˆ 197.68, 175.69, 152.97, 141.21, 132.34, 130.74,
126.06, 125.97, 125.21, 119.65, 119.36, 109.50, 82.83, 76.66, 48.04, 35.03,
‡
HRMS (FAB ‡ ) calcd for C26H50O3Si3 494.3099 [M ]; found 494.3050. The
minor cis-isomer was characterized by 1H NMR (400 MHz, CDCl3): d ˆ
5.44 (d, J ˆ 3.75 Hz, 1H), 5.29 ppm (d, J ˆ 3.83 Hz, 1H).
Diels Alder cycloadduct (43): The enedione 41 (50 mg, 83%, 85 mmol)
and benzocyclobutene 32 (20 mg, 70 mmol) were dissolved in [D8]toluene
(1 mL) and heated at 908C for 12 h. The solvent was evaporated, the
residue was dissolved in MeOH
‡
31.19, 28.02, 22.83, 22.54, 10.29 ppm; MS: m/z: 395 [M ‡Na].
(À)-Rishirilide B ((À)-1): This compound was prepared in a similar manner
to (Æ)-1: [a]2D5 ˆ À13.6 (c ˆ 0.32 in EtOH) (lit. [a]2D2 ˆ ‡12.8 (c ˆ 0.448 in
EtOH)).
(10 mL), and pyridine (6 mL, 70 mmol)
and CSA (16 mg, 70 mmol) were added.
(Æ)-Rishirilide B methyl ester (46) and (Æ)-dimethyl rishirilide B (45): (Æ)-
Rishirilide B (1) (13 mg, 35 mmol) was treated with diazomethane (solution
in Et2O) in MeOH (4 mL) for 15 min to give a mixture that was
evaporated; the residue was purified by chromatography on silica gel
(9:1 ! 4:1 ! 2:1 hexanes/EtOAc ! 9:1 CH2Cl2/MeOH, 0.1% HOAc) to
After heating the mixture at reflux 4 h,
the mixture was evaporated and the
reside was purified by column chroma-
tography (95:5 hexanes/EtOAc) to give
43 (9:1 mixture of regioisomers) as a
colorless oil (26 mg, 72%): IR (film): n ˆ 3480, 2954, 2859, 1747, 1703, 1458,
1255 cmÀ1; 1H NMR (400 MHz, CDCl3): d ˆ 9.03 (s, 1H), 8.60 (s, 1H), 7.65
(d, J ˆ 8.17 Hz, 1H), 7.55 (dd, J ˆ 8.12 Hz, 1H), 7.01 (d, J ˆ 7.49 Hz, 1H),
4.61 (s, 1H), 4.14 4.00 (m, 2H), 3.23 (q, J ˆ 6.48 Hz, 1H), 1.45 (d, J ˆ
6.51 Hz, 1H), 1.10 (s, 9H), 0.95 0.76 (m, 2H), 0.30 (s, 3H), À0.09 ppm (s,
3H); 13C NMR (125 MHz, CDCl3): d ˆ 193.55, 192.99, 168.68, 153.80,
137.21, 132.68, 130.75, 129.46, 127.71, 126.94, 125.82, 122.71, 115.67, 84.18,
65.84, 51.96, 25.78, 18.47, 17.13, 9.04, À1.72, À4.26, À4.35 ppm; HRMS
‡
(FAB ‡ ) calcd for C27H38O6NaSi: 537.2105 [M ‡Na]; found: 537.2092.
give 45 as a white solid (6 mg, 43%). 1H NMR (400 MHz, CDCl3): d ˆ 8.52
(s, 1H), 8.37 (s, 1H), 7.55 (d, J ˆ 8.25 Hz, 1H), 7.42 (dd, J ˆ 7.86 Hz, 1H),
6.90 (d, J ˆ 7.68 Hz, 1H), 4.21 (s, 1H), 4.01 (s, 3H), 3.34 (s, 3H), 3.11 (q, J ˆ
6.85 Hz), 2.50 (s, 1H), 2.30 (dt, J ˆ 3.84, 13.47 Hz, 1H), 1.75 1.60 (m, 1H),
1.45 1.30 (m, 2H), 1.27 (d, J ˆ 6.75 Hz, 3H), 1.25 1.20 (m, 1H), 0.95 0.80
(m, 2H), 0.79 (d, J ˆ 6.42 Hz, 3H), 0.69 ppm (d, J ˆ 6.40 Hz, 3H); MS: m/z:
Diels Alder cycloadduct ((‡)-43): This compound was prepared in a
similar manner as (Æ)-43 (as a 10:1 mixture of regioisomers): [a]2D5 ˆ ‡5.97
(c ˆ 2.68 in CHCl3).
Protected (Æ)-rishirilide B (44): The diketoanthracene 43 (107 mg,
0.208 mmol) was dissolved in THF (20 mL) and cooled to À788C. Isoamyl
magnesium bromide (0.73m Et2O,
‡
423 [C23H28O6Na ].
1.43 mL, 1.04 mmol) was added drop-
Further elution gave 46 as a brown residue (2 mg, 15%) that on standing in
MeOH for 17 d at À208C formed brown prisms. M.p. 94 968C; 1H NMR
(400 MHz, CDCl3): d ˆ 8.54 (s, 1H), 8.30 (s, 1H), 7.57 (d, J ˆ 8.23 Hz, 1H),
7.35 (dd, J ˆ 7.83 Hz, 1H), 6.94 (d, J ˆ 7.09 Hz, 1H), 4.22 (s, 1H), 3.36 (s,
3H), 3.12 (q, J ˆ 6.81 Hz), 2.54 (s, 1H), 2.31 (dt, J ˆ 3.25, 12.70 Hz, 1H),
1.75 1.60 (m, 1H), 1.45 1.30 (m, 2H), 1.26 (d, J ˆ 6.55 Hz, 3H), 1.25 1.20
(m, 1H), 0.95 0.80 (m, 2H), 0.79 (d, J ˆ 6.32 Hz, 3H), 0.68 ppm (d, J ˆ
wise over 15 min. After 1.5 h addi-
tional isoamyl magnesium bromide
(0.63 mL, 0.42 mmol) was added to
achieve completion as judged by thin-
layer chromatography (Rf 0.61, 5%
acetone in toluene). The mixture was
partitioned between Et2O and sat.
NaCl(aq), the organic layer was dried
‡
6.48 Hz, 3H); MS: m/z: 409 [C22H26O6Na ]. Further elution recovered the
natural product, 1 (4 mg, 31%).
(MgSO4) and evaporated, and the
residue was purified by chromatography (95:5 ! 9:1 hexanes/EtOAc) to
give the protected core as a colorless oil (85 mg, 70%). IR (film): n ˆ 3484,
2954, 1719, 1689, 1622, 1573, 1455, 1251 cmÀ1; 1H NMR (400 MHz, CDCl3):
d ˆ 8.51 (s, 1H), 8.36 (s, 1H), 7.58 (d, J ˆ 8.23 Hz, 1H), 7.35 (dd, J ˆ 7.78 Hz,
Trimethylsilylethylcyanoformate: This compound was prepared based on
the literature procedure.[19] KCN (1.82 g, 28 mmol) and Bu3NI (111 mg,
0.3 mmol) were added to a solution of trimethylsilylethylchloroformate[22]
in CH2Cl2 (70 mL) and water (70 mL) at RT, and the mixture was stirred
Chem. Eur. J. 2003, 9, 3242 3252
www.chemeurj.org
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3251

S. J. Danishefsky et al.
FULL PAPER
vigorously for 12 h. The two layers were
separated and the aqueous layer was
extracted with CH2Cl2 (3 Â 10 mL). The
organic layers were combined, dried,
[5] N. Aoki, Prog. Cardiovasc. Dis. 1979, 21 267.
[6] For recent review on oxidative cyclization, see; V. Singh, Acc. Chem.
Res. 1999, 32, 324.
[7] a) W. R. Dolibier, H. Koroniak, K. N. Houk, C. Sheu, Acc. Chem. Res.
1996, 29, 471; b) S. Niwayama, Y. Wang, K. N. Houk, Tetrahedron Lett.
1995, 36, 6201.
[8] For selected pioneering studies of intermolecular Diels Alder
reactions, see: a) M. P. Cava, A. A. Deana, K. Muth, 1959, 81, 6458;
b) Reference [13]; c) L. A. Errede, 1960, 83, 949.
[9] a) For reviews on applications of o-quinodimethides to natural
product synthesis, see: J. L. Charlton, M. M. Alauddin, Tetrahedron
1987, 43, 2873; J. L. Segura, N. MartÌn, Chem. Rev. 1999, 99, 3199;
b) For applications in steroid synthesis, see: P.-Y. Michellys, H.
Pellissier, M. Santelli. Org. Prep. Proced. Int. 1996, 28, 545; c) For
synthetic applications to other polycyclic natural products, see: T.
Kametani, Pure. Appl. Chem. 1978, 51, 747; W. Oppolzer, Synthesis
1978, 793; W. Oppolzer, Heterocycles 1980, 14, 1615; P. Magnus, T.
Gallagher, P. Brown, P. Pappalardo, Acc. Chem. Res. 1984, 17, 35.
[10] a) F. M. Hauser, S. A. Pogany, Synthesis 1980, 814; b) B.-D. Chong, Y.-
I. Ji, S.-S. Oh, J.-D. Yang, W. Baik, S. Koo, J. Org. Chem. 1997, 62, 9323.
[11] a) G. M. Rubottom, J. M. Gruber, R. K. Boeckman, Jr., M. Ramaiah,
J. B. Medwig, Tetrahedron Lett. 1978, 43, 4603; b) H. K. Chenault, S. J.
Danishefsky, J. Org. Chem. 1989, 54, 4249.
and concentrated to give trimethylsilylethylcyanoformate as a light yellow
oil, which was used without further purification (80% purity based on
1H NMR). 1H NMR (400 MHz, CDCl3) 4.42 (m, 2H), 1.14 (m, 2H),
0.07 ppm (s, 9H); 13C{1H} NMR (100 MHz, CDCl3) 144.4, 109.4, 68.15,
17.23, À1.57 ppm.
5-(R)-Methylcyclohexenone (47): This compound was prepared following
the literature procedure[18] as a yellow oil. IR (thin film): nÄ ˆ 1680,
1645 cmÀ1
;
1H NMR (400 MHz,
CDCl3): d ˆ 6.69 (ddd, J ˆ 2.6, 5.6,
10.1 Hz, 1H), 6.01 (brd, J ˆ 9.0 Hz,
1H), 2.48 (dd, J ˆ 3.3, 21.2 Hz), 2.42
(dt, J ˆ 4.4, 23.0 Hz, 1H), 2.22 (m, 1H),
2.12 (dd, J ˆ 11.8, 15.5 Hz, 1H), 2.03
(ddt, J ˆ 2.6, 9.7, 18.6 Hz, 1H), 1.07 ppm
(d, J ˆ 6.4 Hz, 3H); 13C{1H} NMR (100 MHz, CDCl3): d ˆ 200.1, 149.9,
129.8, 65.68, 46.05, 33.84, 30.16, 22.16, 21.01, 15.05, 13.88 ppm.
Acknowledgement
[12] J. C. Bill, D. S. Tarbell, O-Phthalaldehyde, Wiley, 1963.
[13] a) M. P. Cava, D. R. Napier, J. Am. Chem. Soc. 1957, 79, 1701; b) M. P.
Cava, J. F. Stucker, J. Am. Chem. Soc. 1957, 79, 1706.
This work was supported by the National Institutes of Health: AI 16943/
CA 28824 (SJD). Postdoctoral Fellowship Support is gratefully acknowl-
edged by KY (Helen Hay Whitney Foundation), MFH (5T32 CA 62948
05) and JGA (NIH-CA-80356). We thank Brian M. Bridgewater and
David G. Churchill for providing X-ray data. We thank Dr. George
Sukenick for mass spectral analyses.
√
[14] H. Nozaki, R. Noyori, N. Kozaki, Tetrahedron 1964, 20, 641.
[15] E. R. Jarvo, S. R. Boothroyd, M. A. Kerr, Synlett 1996, 898.
[16] Y. Ueda, G. Roberge, V. Vinet, Can. J. Chem. 1984, 62, 2936.
[17] F. M. Hauser, Y.-J. Xu, Org. Lett. 1999, 2, 335.
[18] K. C. Nicolaou, Y. -L. Zhong, P. S. Baran, J. Am. Chem. Soc. 2000, 122,
7596.
[19] Prepared based on literature procedure: Y. Nii, K. Okano, S.
Kobayashi, M. Ohno, Tetrahedron Lett. 1979, 27, 2517.
[20] W. Dam, Y.-Y. Chan, D. Cremer, J. Gauss, D. Scheutzow, M. Schindler,
J. Org. Chem. 1987, 52, 2800.
[1] H. Iwaki, Y. Nakayama, M. Takahashi, S. Uetsuki, M. Kido, Y.
Fukuyama, J. Antibiotics 1984, XXXVII, 1091.
[2] For
a preliminary account on this study see: J. G. Allen, S. J.
[21] S. G. Davies, G. Bhalay, Tetrahedron: Asymmetry 1996, 7, 1595.
[22] R. E. Shute, D. H. Rich, Synthesis 1987, 346.
Danishefsky, J. Am. Chem. Soc. 2001, 123, 351.
[3] a) J. G. Allen, M. F. Hentemann, S. J. Danishefsky, J. Am. Chem. Soc.
2000, 122, 571; b) M. F. Hentemann, J. G. Allen, S. J. Danishefsky,
Angew. Chem. 2000, 112, 2013; Angew. Chem. Int. Ed. 2000, 39, 1937.
[4] K. N. von Kaula, Synth. Fibrinolytic Thrombolytic Agents 1979, 53.
Received: March 10, 2003 [F4931]
3252
¹ 2003 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
Chem. Eur. J. 2003, 9, 3242 3252