Total synthesis of baccatin III and taxol

Article abstract of DOI:10.1021/ja952692a

An intramolecular Heck reaction (90 → 91) serves as the key step in the total synthesis of the titled compounds. The synthetic route is based on utilizing the Wieland-Miescher ketone (5) as a matrix to provide the C and D rings of the targets and to provide functionality implements for joining this sector to an A ring precursor (6). Catalytically induced enantiotopic control and early emplacement of the oxetane are other features of the route.

Full text of DOI:10.1021/ja952692a

J. Am. Chem. Soc. 1996, 118, 2843-2859
2843
Total Synthesis of Baccatin III and Taxol
Samuel J. Danishefsky,*,1a John J. Masters,1b Wendy B. Young,1c J. T. Link,1a,d
Lawrence B. Snyder,1e Thomas V. Magee,1f David K. Jung,1g
Richard C. A. Isaacs,1h William G. Bornmann, Cheryl A. Alaimo,1i
Craig A. Coburn,1h and Martin J. Di Grandi1i
Contribution from the Laboratory for Bio-Organic Chemistry, Sloan-Kettering Institute for
Cancer Research, Memorial Sloan-Kettering Cancer Center, 1275 York AVenue,
New York, New York 10021
ReceiVed August 8, 1995X
Abstract: An intramolecular Heck reaction (90 f 91) serves as the key step in the total synthesis of the titled
compounds. The synthetic route is based on utilizing the Wieland-Miescher ketone (5) as a matrix to provide the
C and D rings of the targets and to provide functionality implements for joining this sector to an A ring precursor
(6). Catalytically induced enantiotopic control and early emplacement of the oxetane are other features of the route.
Background
In 1964, Wall and co-workers discovered that extracts from
the bark of Taxus breVifolia exhibited significant cytotoxicity
against KB cells. From these extracts was isolated a particularly
active principle which was termed taxol. It was deduced, again
by Wall and co-workers, that taxol corresponds to structure 32
(see Figure 1). Taxol (3) fared well in early preclinical
cytotoxicity challenges against a variety of cell lines, and in
time, emerged as a candidate compound for clinical evaluation.
Another step in the ascension of taxol (3) arose from a
disclosure of Horwitz and colleagues wherein the drug was
described to be a promoter of microtubule assembly.3 The
identification of such a mechanism of action for taxol (and
putative analogs) at the cellular level would facilitate biochemi-
cal and pharmacological investigation to go along with clinically
oriented investigations.
Relentlessly, taxol (3) and its analog taxotere (4)4 advanced
from the status of esoteric research curiosities to compounds
of serious clinical value. At the present time, taxol (3) has been
approved for use against metastatic ovarian and breast cancers
and is undergoing evaluation against a variety of other indica-
tions. The status of taxotere as an active drug is not as yet as
advanced as is that of taxol.5
Figure 1. Structures of baccatin III (1), 10-deacetylbaccatin III (2),
taxol (3), and taxotere (4).
tochemical route to taxol itself is, even at this writing,
complicated. Thus, in most taxol-containing plants, the fully
functional drug tends to be localized in nonrenewable domains.
On the other hand, baccatin III (1) and 10-deacetylbaccatin III
(2), themselves bereft of useful biological function, constitute
rather more accessible raw materials. They are isolable in
quantity from renewable regions of a variety of plants. Intense
efforts to achieve the semisynthesis of taxol (3) from 1 or 2
have been successful in ameliorating the drug availability
problem. Particularly critical in this regard were major semi-
synthesis contributions from Holton, Ojima, and Greene.6
Interest in the total synthesis of taxol (3) (via a total synthesis
of baccatin III (1)) was triggered by several factors. The novel
confluence of functionality of the tetracyclic ring system of 1
constitutes a challenge which must be accommodated in any
successful effort. It seemed likely that solutions of this total
synthesis problem would bring with them collateral advances
in the theory and practice of organic chemistry. Needless to
say, the favorable clinical findings pertinent to taxol (since given
the trademark name of Paclitaxel) added to the general interest
in the total synthesis goal. During the era where the availability
of taxol was feared to be a seriously limiting factor in its
oncological applications, total synthesis was seen by a hearty
few as a possible source of the drug.
Early on, it appeared that clinical applications of taxol would
be hindered by a lack of availability of the drug. The phy-
X Abstract published in AdVance ACS Abstracts, February 15, 1996.
(1) Current addresses: (a) Department of Chemistry, Columbia Univer-
sity, Havemeyer Hall, New York, NY 10027. (b) Eli Lilly and Co.,
Indianapolis, IN 46285. (c) Arris Pharmaceuticals, 385 Oyster Pt. Blvd.,
Suite 3, South San Francisco, CA 94080. (d) A portion of this work was
submitted by J. T. Link as part of his Ph.D. dissertation, Columbia
University, August 1995. Current addresses: (e) Bristol Myers Squibb, S.
Research Pkwy, Wallingford, CT 06492-7660. (f) Pfizer Central Research,
Eastern Pt. Rd., Groton, CT 06340. (g) Glaxo Pharmaceuticals, Five Moore
Drive, P.O. Box 13358, Research Triangle Park, NC 27709. (h) Merck &
Co., Sumney Town Pike, West Point, PA 19486. (i) Schering Plough
Research, 2015 Gallop Hill Rd., Kenilworth, NJ 07033. (j) Wyeth Ayerst,
401 N. Middleton Rd., Pearl River, NY 10965.
Consideration of the chemical complexity of baccatin III (1),
which in suitably protected form (at C7) would be the likely
synthetic intermediate en route to taxol (3), should have
(2) Wani, M. C.; Taylor, H. L.; Wall, M. E.; Coggon, P.; McPhail, A.
T. J. Am. Chem. Soc. 1971, 93, 2325.
(3) Horwitz, S. B.; Fant, J.; Schiff, P. B. Nature 1979, 277, 665.
(4) Mangatal, L.; Adeline, M. T.; Guenard, D.; Gueritte-Voegelein, F.;
Potier, P. Tetrahedron 1989, 45, 4177.
(5) For reviews, see: (a) Georg, G. I.; Chen, T. T.; Ojima, I.; Vyas, D.
M. Taxane Anticancer Agents; American Cancer Society: San Diego, 1995.
(b) Kingston, D. G. I.; Molinero, A. A.; Rimoldo, J. M. Progress in the
Chemistry of Organic Natural Products 61; Springer-Verlag: New York,
1993.
(6) (a) Denis, J. N.; Greene, A. E.; Guenard, D.; Gueritte-Voegelein, F.;
Mangatal, L.; Potier, P. J. Am. Chem. Soc. 1988, 110, 5917. (b) Holton, R.
A. Eur. Pat. Appl. EP400,971, 1990; Chem. Abstr. 1990, 114, 164568q. (c)
Holton, R. A. Workshop on Taxol and Taxus, 1991. (d) Ojima, I.; Habus,
I.; Zhao, M.; Georg, G. I.; Jayasinghe, L. R. J. Org. Chem. 1991, 56, 1681.
(e) Ojima, I.; Habus, I.; Zhao, M.; Zucco, M.; Park, Y. H.; Sun, C. M.;
Brigaud, T. Tetrahedron 1992, 48, 6985. (f) Ojima, I.; Sun, C. M.; Zucco,
M.; Park, Y. M.; Duclos, O.; Kuduk, S. Tetrahedron Lett. 1993, 34, 4149.
0002-7863/96/1518-2843$12.00/0 © 1996 American Chemical Society

2844 J. Am. Chem. Soc., Vol. 118, No. 12, 1996
Danishefsky et al.
engendered considerable skepticism and even disbelief that total
synthesis would supplant natural sources as a route to the drug.
More plausible, though as yet unrealized in practice, is the
prospect that mastery of the synthesis of baccatin III (1) will
bring with it new nuclei which, upon suitable conjugation with
biologically critical side chains, might provide medically
promising variants of taxol.
Scheme 1
Indeed much interesting chemistry has been developed
following synthetic explorations pertinent to taxol.7 Many
fascinating constructions and proposed constructions have been
adumbrated and disclosed. However, upon critical examination,
very few have carried with them realistic prospects for maturing
into comprehensive total synthesesspractical or otherwise. The
baccatin III (1) total synthesis problem has served as a stark
but useful reminder of the distinction between “regional
contributions” as opposed to programs which take account of
the full scope of a complex problem in synthesis.
The first total molecular solutions to the taxol problem, again
via 7-protected baccatin III, were provided in virtually concur-
rent disclosures by the research groups of Nicolaou8 and Holton.9
Given the difficulty of the challenge, and the multiple op-
portunities for failure, these two successes were clearly impor-
tant events in the field. In this paper, we document the total
synthesis of baccatin III and thence taxol which were achieved
in our laboratory.10
principle, in place. Thus, in 5, the future angular (C19) methyl
group and the oxygenated C7 of baccatin III (1) are readily
discerned as is the 6-membered C ring. A wealth of precedent
from terpenoid chemistry could be drawn upon to allow for
fashioning the necessary stereochemistry at C3 and C7 (baccatin
III numbering) from carbons 4a and 1 of 5. Furthermore, if
one imagines an eminently plausible deconjugation of the
conjugated enone double bond, originally present in compound
5, functionality emerges at C4 of baccatin III (1) (corresponding
to carbon 4 of the Wieland-Miescher ketone). Such a ketone
at C4 might be further exploitable to introduce a required oxygen
at C3 (Wieland-Miescher ketone numbering) corresponding to
C5 of baccatin III (1).
Left unspecified for the moment is the timing for introduction
of oxygenation at C9 of the baccatin (corresponding to C8 of
the Wieland-Miescher ketone). One could easily imagine using
the matrix of the bicyclic system of compound 5 to achieve
orderly functionalization at C8. However, there was concern,
as the synthesis began to unfold, that early inclusion of an
oxygen-based functional group at this center could lead to
intermediates whose additional complexity could prove to be
unmanageable.
We now consider the A ring substructure 6. The 1,3-diketone
system corresponds, overall, to functionality sites at C1 and C11
of baccatin III (1) and contains the required C-methyl pattern
(see Scheme 1). Two possibilities were entertained for bringing
together fragments derived from 5 and 6 in a useful way. For
this purpose we particularly concentrate on C9 and C10 of
baccatin III (1). Clearly, C9 is to be derived from building block
5. No possibility for bonding 6 to a fragment derived from
Wieland-Miescher ketone, in which C8 of 5 were dissipated,
was ever given serious credence. Such an excision would place
C8a (cf. 5) at great risk from a stereochemical standpoint and
would open up major hazards inherent in reconstruction. In
short, C9 of baccatin III (1) must be deriVed from C8 of
compound 5.
Two formulations for deriving C10 of the baccatin goal
presented themselves (Scheme 2). In one plan, an incision into
the matrix derived from 5 would be made between C7 and C8
as well as between C6 and C5 (see cleavage mode A). In this
view, a one-carbon fragment, destined to become C10 of baccatin
III (1), would be appended to 6 (at its carbon slated to emerge
at C11) and a bond established between C10 and C9 (baccatin
numbering). Alternatively, both C10 and C9 would be derived
from 5. The fragment to be presented for coupling with 6 could
have arisen from excision of C6 of the original Wieland-
Miescher ketone construct, with emergence of suitably dif-
ferentiated functionality fragments presented at its erstwhile C7
and C5 (see cleavage mode B). We note that compelling
Synthetic Planning
Our planning and thinking about a total synthesis of taxol
(3) (which in reality meant a total synthesis of baccatin III (1)
or a 7-protected derivative thereof) was guided from the
beginning by a central perception from which we never deviated.
It seemed that a suitable confluence of the Wieland-Miescher
ketone11ab (5) with the known and easily prepared (though not
commercially available) trimethylcyclohexane-1,3-dione (6)11c
might eventually lead to baccatin III (Scheme 1). This
prospectus had several alluring features. The sum of the two
eminently accessible components contain more than an ample
number of carbons (11 + 9 ) 20) to reach the core ABC sector
(19 carbons) of 1. Of greater significance than this type of
carbon count was the recognition that the functionalities for
development of the components and suitable melding were, in
(7) For reviews, see: (a) Swindell, C. S. Studies in Natural Products
Chemistry, Vol. 12; Elsevier Science Publishers: New York, 1993; pp 179-
231. (b) Nicolaou, K. C.; Dai, W. M.; Guy, R. K. Angew. Chem., Int. Ed.
Engl. 1994, 33, 15.
(8) (a) Nicolaou, K. C.; Zang, Z.; Liu, J. J.; Ueno, H.; Nantermet, P. G.;
Guy, R. K.; Claiborne, C. F.; Renaud, J.; Couladouros, E. A.; Paulvannan,
K.; Sorensen, E. J. Nature 1994, 367, 630. (b) Nicolaou, K. C.; Nantermet,
P. G.; Ueno, H.; Guy, R. K.; Coulandouros, E. A.; Sorensen, E. J. J. Am.
Chem Soc. 1995, 117, 624. (c) Nicolaou, K. C.; Liu, J. J.; Yang, H.; Ueno,
H.; Sorensen, E. J.; Claiborne, C. F.; Guy, R. K.; Hwang, C. K.; Nakada,
M.; Nantermet, P. G. J. Am. Chem Soc. 1995, 117, 634. (d) Nicolaou, K.
C.; Yang, Z.; Liu, J. J.; Nantermet, P. G.; Claiborne, C. F.; Renaud, J.;
Guy, R. K.; Shibayama, K. J. Am. Chem Soc. 1995, 117, 645. (e) Nicolaou,
K. C.; Ueno, H.; Liu, J. J.; Nantermet, P. G.; Yang, Z.; Renaud, J.;
Paulvannan, K.; Chadha, R. J. Am. Chem. Soc. 1995, 117, 653.
(9) (a) Holton, R. A.; Somoza, C.; Kim, H. B.; Liang, F.; Biediger, R.
J.; Boatman, D.; Shindo, M.; Smith, C. C.; Kim, S.; Nadizadeh, H.; Suzuki,
Y.; Tao, C.; Vu, P.; Tang, S.; Zhang, P.; Murthi, K. K.; Gentile, L. S.; Liu,
J. H. J. Am. Chem. Soc. 1994, 116, 1597. (b) Holton, R. A.; Kim, H. B.;
Somoza, C.; Liang, F.; Biediger, R. J.; Boatman, D.; Shindo, M.; Smith,
C. C.; Kim, S.; Nadizadeh, H.; Suzuki, Y.; Tao, C.; Vu, P.; Tang, S.; Zhang,
P.; Murthi, K. K.; Gentile, L. S.; Liu, J. H. J. Am. Chem. Soc. 1994, 116,
1599.
(10) For a preliminary account of this work, see: Masters, J. J.; Link, J.
T.; Snyder, L. B.; Young, W. B. and Danishefsky, S. J. Angew. Chem.
1995, 34, 1723.
(11) (a) Wieland, P.; Miescher, K. HelV. Chim. Acta 1950, 33, 2215. (b)
Ramachandrin, S.; Newman, M. S. Organic Syntheses; Wiley: New York,
1973; Collect. Vol. V, p 486. (c) Hargreaves, J. H.; Hickmott, P. W.;
Hopkins, B. J. J. Chem. Soc. C 1968, 2599.

Total Synthesis of Baccatin III and Taxol
J. Am. Chem. Soc., Vol. 118, No. 12, 1996 2845
Scheme 2
Scheme 4
Scheme 5
Scheme 3
The exploratory studies were conducted on racemic material.
Actually, the synthesis of the appropriate C8a S-enantiomer is
no more demanding than is that of the racemate. However,
the racemate is commercially available, and we drew on this
convenience during the reconnaissance phases of, what was
likely to be, an arduous journey.14 After validation in the
racemic series, the chemistry was translated to the single
enantiomer regime arising from compound 5, C8a ) S.
We conclude the section on analysis by consideration of the
timing of oxetane installation. The oxetane is, by all indications,
rather important for biological activity.15 Maximization of the
prospects that the total synthesis exercise might produce
significantly truncated, but still active, versions of taxol (3)
constituted a strong argument for early emplacement of the
oxetane. Moreover, such a strategy would avoid the need for
elaborate protecting devices to contain the “access handles”
which would be needed for a late stage fashioning of the
oxetane.
We were mindful that the presence of the oxetane ring from
nearly the outset might restrict the scope of feasible operations
for the orchestration of the synthesis. However, we hoped that
successful management of the problems engendered by the
resident oxetane might, in the long run, reduce the number of
steps. Furthermore, we would learn much about the “chemical
personality” of the oxetane. This knowledge could be helpful
in developing an analog program. Though at several stages of
the expedition we were destined to face serious consequences
from this high-risk course, it would in time be shown to be
sound (Vide infra).
arguments (similar to those advanced above as to C8 of 5)
dictated that C5 could not be sacrificed to the degradation. Any
scheme to excise this carbon from a construct derived via 5
would again undermine the otherwise securable stereochemistry
at C4a (slated to become C3 of baccatin III (1)). We were
understandably apprehensive about confronting the serious
questions which would be raised by attempts to reattach C4a of
a degraded version of 5 to an A ring derived intermediate.
We favored schemes anticipating formation of the C10-C11
baccatin bond on grounds of synthetic freshness. Thus, a C9-
C10 closure had already been investigated in Kende’s ground-
breaking assembly of the skeleton.12 Furthermore, as described
by the Rochester workers, such a coupling carried with it its
own risks and serious complications. Left unspecified at this
planning stage, is the matter of the sequencing of the C1-C2
and C10-C11 attachments. One could imagine establishing the
C10-C11 bond in the intermolecular step, reserving formation
of the C1-C2 bond for cyclization (Scheme 3). Alternatively,
the reverse order could be considered, wherein the C1-C2 bond
would be formed first through intermolecular means, delaying
C10-C11 bond construction for the cyclization (Scheme 4).
In practice, both sequences were explored, and we shall return
to this matter as the synthesis unfolds. At this stage, however,
it is well to emphasize that the concept of the Wieland-
Miescher ketone matrix gained particular favor from the fact
that the enantiomerically pure compound could be obtained in
quantity by L-proline-induced aldolization of the prochiral trione
7 (Scheme 5).13 Thus all stereochemistry of baccatin III (1)
would flow from a single stereogenic center (C8a of the
Wieland-Miescher ketone, corresponding to C8 of the baccatin),
which was itself induced by catalytic means.
Results and Discussion
Compound 5 was reduced with sodium borohydride according
to known protocols,16 affording 8 and thence the deconjugated
ketal 9 (Scheme 6). The secondary alcohol of 9, targeted to
emerge at C7 of the baccatin, was protected as its tert-butyl
dimethylsilyl ether.
Treatment of 10 with diborane was followed by oxidation
with hydrogen peroxide. Oxidation of the alcohol, thus ob-
(14) The R version of 5 is equally accessible and could be used to
synthesize ent-baccatin III. We have, in fact, prepared late stage ent
intermediates by this method.
(15) The Chemistry and Pharmacology of Taxol; Farina, V., Ed.; Elsevier
Press: New York, 1995.
(12) Kende, A. S.; Johnson, S.; Sanfillipo, P.; Hodges, J. C.; Jungheim,
L. N. J. Am. Chem. Soc. 1986, 108, 3513.
(13) Harada, N.; Sugioka, T.; Hisashi, U; Kuriki, T. Synthesis 1990, 53
and references therein.
(16) Heathcock, C. H.; Ratcliffe, R. J. Am. Chem. Soc. 1971, 93, 1746.

2846 J. Am. Chem. Soc., Vol. 118, No. 12, 1996
Danishefsky et al.
Scheme 6a
Scheme 8a
a Reagents: (a) NaBH4, EtOH, 0 °C, 97%; (b) Ac2O, DMAP, pyr,
CH2Cl2, 0 °C, 99%; (c) (HOCH2)2, PhH, naphthalenesulfonic acid,
reflux, 70%; (d) NaOMe, MeOH, THF, 98%; (e) TBSOTf, 2,6-lutidine,
CH2Cl2, 0 °C, 97%.
Scheme 7a
a Reagents: (a) OsO4, NMO, acetone, H2O, 66%; (b) TMSCl, pyr,
CH2Cl2, -78 °C f room temperature; (c) Tf2O, -78 °C f room
temperature; (d) (HOCH2)2, 40 °C, 69%; (e) NaH, THF, 45 °C, 74%.
principle, to follow the lead of Berkowitz,20 toward this
objective. Indeed, such a transformation was accomplished in
our laboratory on the closely related hydrindenone-derived
system 18 en route to 19.21
a Reagents: (a) (i) BH3‚THF, THF, 0 °C f room temperature; (ii)
H2O2, NaOH, H2O; (iii) PDC, CH2Cl2, 0 °C f room temperature; (iv)
NaOMe, MeOH, 62%; (b) (i) KHMDS, THF, -78 °C; (ii) PhNTf2,
81%; (c) Pd(OAc)2, PPh3, CO, Hunig’s base, MeOH, DMF, 73%; (d)
DIBAL, hexanes, -78 °C, 99%; (e) Me3S+I-, KHMDS, THF, 0 °C,
99%; (f) Al(OiPr)3, PhMe, reflux, 99%.
However, in the case at hand, a new approach was feasible.
The primary alcohol of 16 was readily differentiated by silylation
with TMS chloride (see compound 20). The secondary alcohol
was then activated by triflation (see structure 21). Remarkably,
treatment of 21 with ethylene glycol at reflux gave rise to the
desired oxetane 22. We presume that this transformation entails
a hypervalent silicon ether species on C20, triggering displace-
ment of the triflate. Another product, 23, arising from a pinacol-
like rearrangement was also detected in this reaction.17
We note that the formation of the oxetane by inversion of C5
bearing a leaving group has subsequently been employed, though
with different experimental protocols, in both the Nicolaou8 and
Holton9 syntheses. During the preparation of our initial
disclosure of this strategy,17 there appeared a similar outcome
in a semisynthetic setting.22 However, the Potier chemistry was
conducted on a substrate which lacked functionality at C7. The
demonstration shown here for building the oxetane onto a ring
system bearing four differentiated alcohol centers had significant
implications for total synthesis and for organizing an analog
program.
Considerable thought and experimentation was devoted to
selecting the proper blocking group for the tertiary hydroxyl
center of compound 22. Much would be expected from this
blocking group, not the least demanding property being its
susceptibility to removal at a late stage of the synthesis. It was
also necessary to differentiate, in a secure way, the C4 site from
the other alcohol centers at C1, C7, and C10 of the emerging
baccatin III. On the basis of these considerations, and after
surveying several other possibilities (Vide infra), we installed a
benzyl protecting group on the tertiary alcohol. This transfor-
mation was accomplished in high yield on compound 22 giving
tained, with PDC followed by equilibration of the resultant crude
ketone mixture16 afforded homogeneous, trans-fused 11 (Scheme
7). Having secured the R configuration at C3 (baccatin
numbering), we directed our attention to the fashioning of the
oxetane linkage.
We set as our next interim goal compound the allylic alcohol
15. It was hoped that a triol derived from 15 might be
transformed into the desired oxetane. In practice, two protocols
were pursued for addition of the methylene group destined to
become C20 of the oxetane. In one, the ketone 11 was converted
to its vinyl triflate 12.17 The latter, upon palladium-mediated
carbonylation-methoxylation gave rise to R,â unsaturated ester
13 which, upon reduction, led to the desired allylic alcohol 15.
However, as the synthesis progressed and large quantities of
intermediates had to be generated, this protocol was less than
optimal. Rather, we took recourse to Corey’s sulfonium ylide
methodology.18 Conversion of ketone 11 to spiroepoxide 14
occurred in high yield under the conditions shown. Lewis acid-
induced epoxide opening following precedents gave rise to the
desired allylic alcohol 15.19
Many possibilities were explored to convert allylic alcohol
15 to the next goal system, oxetane 22.17 Treatment of the
allylic alcohol 15 with osmium tetroxide gave rise to triol 16
(Scheme 8). In light of the lineup of resident â-face functional-
ity, we were surprised that this reaction was not stereospecific.
Approximately 15% of 17 wherein osmylation had occurred
from the more hindered â face was also obtained. Fortunately,
even on large scale, the separation of the two diols was a
straightforward matter and we proceeded to advance triol 16
toward the desired oxetane. It would have been possible, in
(20) Berkowitz, W. F.; Amarasekara, A. S.; Perumattam, J. J. J. Org.
Chem. 1987, 52, 3745.
(21) Isaacs, R. C. A.; Di Grandi, M. J.; Danishefsky, S. J. J. Org. Chem.
1993, 58, 3938.
(17) Magee, T. V.; Bornmann, W. G.; Isaacs, R. C. A.; Danishefsky, S.
J. J. Org. Chem. 1992, 57, 3274. This exploratory effort was conducted in
the racemic series.
(18) Corey, E. J.; Chaykovsky, M. J. Am. Chem. Soc. 1965, 87, 1353.
(19) Waddell, T. G.; Ross, P. A. J. Org. Chem. 1987, 52, 4802.
(22) Ettouate, L.; Ahond, A.; Poupat, C.; Potier, P. Tetrahedron 1991,
47, 9823.

Total Synthesis of Baccatin III and Taxol
J. Am. Chem. Soc., Vol. 118, No. 12, 1996 2847
Scheme 9a
Scheme 12
a Reagents: (a) BnBr, NaH, TBAI, THF, 0 °C f room temperature,
98%; (b) TsOH, acetone, H2O, 70 °C, 84%.
Scheme 10
Scheme 13
Scheme 11
Prototype C1 nucleophiles were also developed early in our
investigations. Some were fashioned in systems in which the
future C10 was already preinstalled on the A ring (see discussion
above appropriate to Scheme 1). For instance, we demonstrated
the feasibility of a Shapiro reaction on trisylhydrazone 36
derived from ketone 35 (Scheme 13). The lithio derivative 37
was coupled to a variety of aromatic aldehydes and to the highly
functionalized aldehyde 47 (Vide infra) derived from another
fragmentation strategy discussed below. In this way, we gained
confidence that compound 6 could be converted to effective
nucleophiles corresponding to either C1 or C11 of baccatin III
(1). It was our hope to exploit these capabilities for bond
formation with fragmentation products derived from a suitable
version of 26.
After many of the options were explored, the one shown in
Scheme 14 was adopted owing to its amenability to large-scale
work. Taking advantage of selective enolization of the keto
group in systems of the type 26 toward C7 (Wieland-Miescher
ketone numbering) specific ketone 25 was converted to silyl
enol ether 38. The latter was hydroxylated via a modified
Rubottom type protocol,27a with 3,3-dimethyldioxirane, followed
by treatment with camphorsulfonic acid (see compound 39).
Ring fragmentation was accomplished with lead tetraacetate,
40. The carboxyl carbon emerged as a methyl ester (see
compound 40) since the fragmentation was performed in
methanol. The aldehyde function in 40 was converted to the
dimethyl acetal (see compound 41). Lithium aluminum hydride
reduction of the ester gave rise to carbinol 42, which was
converted in a Grieco protocol to the o-nitrophenyl selenide
43.27b Oxidation with hydrogen peroxide then provided alkene
44. Ozonolysis of 44 gave aldehyde 45 which was destined to
serve as a key intermediate en route to baccatin III.
rise to benzyl ether 24 (Scheme 9). It was further possible to
cleave the ketal linkage of compound 24 without damaging the
oxetane moiety and without cleaving the silyl ether protecting
group (see compound 25).
With ketone 25 in hand, a variety of efficient degradation
sequences were developed.23 Some of these transformations
were conducted before the benzyl blocking group was settled
upon (see general system 26, Scheme 10). A key feature which
opened up a range of degradative options was the highly
selective enolization of the C6 ketone in the direction of C7 (see
Scheme 1 for Wieland-Miescher ketone numbering). That
being the case, it was possible by well-precedented chemisty
to prepare enone type 27, as well as aldehyde-ester 29 (derived
from the enone dihydroxylation product 28).
Another type of degradation product available from enone
type 27 was the R-oxygenated aldehydo ester 31 (Scheme 11).
This compound was derived by Schreiber type directed ozo-
nolysis24 of allylic alcohol derivatives (30). The latter arose
from 27 by a Wharton type transposition.25
The use of compound 6 as a nucleophilic coupling partner
with fragments obtained from system type 26 was facilitated
by the selective chemistry which this â-diketone undergoes. For
instance, monoketalization of 6 gave rise to 32 which could be
converted to vinyl iodide 33 (Scheme 12). The latter could be
lithiated to give rise to a C11-based nucleophile 34, which was
coupled to aromatic aldehydes as previously documented.26 We
shall return to lithio system 34 in discussing attempts to reduce
Scheme 3 to practice.
(23) Di Grandi, M. J.; Coburn, C. A.; Isaacs, R. C. A.; Danishefsky, S.
J. J. Org. Chem. 1993, 58, 7728.
(24) Schreiber, S. L.; Claus, R. E.; Reagan, J. Tetrahedron Lett. 1982,
23, 3867.
The previously mentioned aldehyde 47 was prepared from
46 which was in turn prepared by a sequence identical to that
described for 44, starting with compound 26a (the 4-OTBS
(25) (a) Wharton, P. S.; Bohlen, D. H. J. Org. Chem. 1961, 26, 3615.
(b) Maas, D. D.; Blagg, M.; Wiemer, D. F. J. Org. Chem. 1984, 49, 853.
(26) Di Grandi, M. J.; Jung, D. K.; Krol, W. J.; Danishefsky, S. J. J.
Org. Chem. 1993, 58, 4989.
(27) (a) Rubottom, G. M.; Vazquez, M. A.; Pelegrina, D. R. Tetrahedron
Lett. 1972, 3375. (b) Grieco, P. A.; Gilman, S.; Nishizawa, M. J. Org. Chem.
1976, 41, 1485.

2848 J. Am. Chem. Soc., Vol. 118, No. 12, 1996
Danishefsky et al.
Scheme 14a
Scheme 15
a Reagents: (a) TMSOTf, Et3N, CH2Cl2, -78 °C; (b) 3,3-dimeth-
yldioxirane, CH2Cl2, 0 °C; CSA, acetone, room temperature, 89%; (c)
Pb(OAc)4, MeOH, PhH, 0 °C, 97%; (d) MeOH, CPTS, 70 °C, 97%;
(e) LiAlH4, THF, 0 °C, 100%; (f) o-NO2C6H4SeCN, PBu3, THF, room
temperature, 88%; (g) 30% H2O2, THF, room temperature, 90%; (h)
O3, CH2Cl2, -78 °C; PPh3, 79%.
Scheme 16
version of 26). Cleavage of the dimethyl acetal of 46 followed
by reduction with lithium aluminum hydride, protection of the
resultant carbinol as its OTBS ether, and ozonolysis of the vinyl
group afforded aldehyde 47.
aldehyde ensemble (cf. 51 or 52) from its precursor guarding
group added further difficulties to an already complicated
program. While various potential solutions to this problem
remain to be explored, our preferences came to be focused on
a plan which called for establishing the C1-C2 bond first.
Our first demonstration that it would be possible to establish
the C1-C2 bond by the coupling strategy which we envisioned,
arose from the reaction of vinyllithium agent 37 (described
above) with aldehyde 47 (Scheme 16). A 60% yield of 53 was
obtained as a single stereoisomer. While this was a key
advance, its context did not fit the formulation implicit in
Scheme 4. In particular, the presence of a one-carbon unit at
C11 (which we number as C10′) deviated from the plan calling
for a C10-C11 closure.
Actually, we would have been willing to depart from our
plan. However, serious difficulties soon surfaced. Attempts
to protect the secondary alcohol (C2 of seco baccatin system
53) in this congested environment were unsuccessful. Failure
to protect this alcohol undermined efforts to develop the
functional groups along the “northern rim”. Again, while other
possibilities remained to be investigated, the results of these
expeditions were sufficiently discouraging to oblige a return to
the basic Scheme 4 prospectus.
Before discussing our successful route to baccatin III (1) via
initial C1-C2 bond formation (Scheme 4), we summarize our
efforts in the direction of first establishing a C9-C10 bond and
postponing the C1-C2 merger for the cyclization step (Scheme
15). Actually we had preferred this plan. Reductive cyclization
of a keto aldehyde precursor (cf. 48) was of particular interest
because it offered the prospect of delivering the C1-C2
functionality, configured in its required sense (see 49).28 We
achieved the coupling of lithium reagent 34 with several relevant
aldehydes. For the most part, these studies were carried out
with cholestanone-derived aldehydes of the type 50. The ready
accessibility of such steroid-based substrates greatly facilitated
the reconnaissance phase of our investigations. The translat-
ability of steroid models to actual pre-taxol constructs was
excellent save for those initiatives in the latter series wherein
the oxetane ring proved to be unstable.
While coupling of 34 and aldehydes of the type 50 could be
achieved, we have not at this writing succeeded in traversing
all of the steps required to reach a keto aldehyde corresponding
to 51 in the steroid series, or 52 in the oxetane containing series.
The problems were encountered when trying to constrain
(through an acetal linkage) or to otherwise protect the oxygens
at C9 and C10 (future baccatin numbering). Even protection of
the secondary alcohol in seco system 52 (X ) H) was
problematic. Furthermore, the revealing of the C1-C2 keto
For this program to be implemented, a methodological
advance had to be achieved. The problem involved the status
of C11 as we exploited the C1 vinyl nucleophile. In the case of
37, this issue had been obviated by employment of a carbon-
carbon bond (C11-C10′). This bond served to stabilize C11
during lithiation. In the conduct of the actual synthesis,
however, we could not employ such “protection”.
(28) For two examples describing the reduction of this scheme to practice,
albeit in much simpler (ring C aromatic) settings, see: (a) Swindell, C. S.;
Chander, M. C.; Heerding, P. G.; Rahman, L. T.; Raman, V.; Venkataraman,
H. Tetrahedron Lett. 1993, 34, 7005. Swindell; (b) C. S.; Fan, W. and
Klimko, P. G. Tetrahedron Lett. 1994, 35, 4959.

Total Synthesis of Baccatin III and Taxol
J. Am. Chem. Soc., Vol. 118, No. 12, 1996 2849
Scheme 19a
Scheme 17
Scheme 18a
a Reagents: (a) tBuLi, THF -78 °C; (b) add 61, THF, -78 °C,
90%; (c) TBAF, THF, 0 °C, 100%; (d) VO(acac)2, tBuOOH, CH2Cl2,
0 °C, 91%; (e) H2, Pd/C, EtOH, 0 °C, 90%; (f) e, then HOAc 0 °C to
room temperature, 56% from epoxide; (g) Cl2CO, pyr, CH2Cl2, 0 °C,
100%.
a Reagents: (a) H2NNH2, Et3N, EtOH, 72%; (b) I2, DBN, THF, 52%;
(c) I2, DBN, THF, 89%; (d) TMSCN, cat. KCN, 18-crown-6, CH2Cl2,
89%.
degradation of cholestanone, would serve as a model for
aldehyde 45 (which was destined to be advanced to baccatin
III (1)). The sequence started with dione 6, which was converted
to its monohydrazone 62. The latter, on treatment with iodine
in a Barton reaction,30 gave rise to iododienone 64. This
substance presumably arose from iodine induced dehydroge-
nation of the expected monoene iodide intermediate 63.
Compound 64 could be converted to its cyanohydrin 65 in
racemic form. The complications associated with the coupling
of racemic intermediates with optically pure agents would not
be of consequence since the sp3 hybridization, introduced for
purposes of metalation, would be discharged during workup.
A key finding was that compound 65 could, in fact, be
lithiated to give rise to 66 (Scheme 19). The latter reacted with
aldehyde 61 to give, after decyanation, a single stereoisomer
formulated as 67. While the configuration at the secondary
hydroxyl group in this substance was not proven at this stage,
it was established through crystallographic analysis of subse-
quent reaction products (Vide infra). It was then found that
directed epoxidation of the C14-C1 olefin produced a single
epoxide. Furthermore, this epoxide could be cleaved by
palladium-induced hydrogenolysis to generate diol 68 or further
reduced to the saturated diol 69 by the addition of acetic acid.
Compound 69 could be converted to cyclic carbonate 70 under
the conditions shown. It was in this way that the problem,
captured in the hypothetical expression 60, was solved.
Two options for the status of C11 at the stage of cyclization
to C10 were considered (Scheme 17). In one case, this carbon
would have emerged as the C11-C12 vinyl iodide of 54.
Alternatively, it might be presented as a vinyl triflate as shown
in structure 55. Recourse to a carbon appendage at C11 would
complicate reaching either 54 or 55.
Much effort was expended generating 1,4-cyclohexadienyl-
lithium reagents of the type 56. These efforts were unsuccessful.
Since we could not store the implement eventually needed at
C11 of the hypothetical vinyllithium reagent 56, it would be
necessary to create this functionality after coupling of C1 and
C2. To eventually reach 54 or 55, a C11 ketone seemed to be
required. Thus, the exact status of the “pro-ketonic” C11 during
the vinyllithium formation had to be determined. The possibility
of generating an agent of the type 58 with the C11 ketone already
in place from a precursor of the type 57 was shown to be feasible
in one instance,26 but was subsequently abandoned for lack of
generality.
We also did not pursue solutions where C11 was carried as a
protected alcohol at the stage when the vinlylmetal nucleophile
(cf. 59) was generated. The disentanglement of the fate of such
a C11 alcohol from the other resident hydroxyls (cf. C1, C2, C4,
and C7) during the manipulations prior to cyclization would be
very complicated.
We shall return to ketone 70 and to its utility in various
cyclization possibilities shortly. However, first we describe the
application of these findings to the oxetane-bearing 45 (Scheme
20). Compound 45 and lithium reagent 66, again, gave rise to
a single carbinol 71 after deprotection of the C11 ketone. At
The dilemma which we now faced could be formulated by
the need to generate a synthetic equivalent of hypothetical
construct 60. A solution was discovered (Scheme 18).29 It was
decided that the aldehyde acetal 61, which was derived by the
(29) Masters, J. J.; Jung, D. K.; Danishefsky, S. J.; Snyder, L. B.; Park,
T. K.; Isaacs, R. C. A.; Alaimo, C. A.; Young, W. B. Angew. Chem. Int.
Ed. Engl. 1995, 34, 452.
(30) Barton, D. H. R.; Bashiardes, G.; Fourrey, J.-L. Tetrahedron 1988,
44, 147-162.

2850 J. Am. Chem. Soc., Vol. 118, No. 12, 1996
Danishefsky et al.
Scheme 20a
a Reagents: (a) (i) THF, -78 °C, 93%; (ii) TBAF, THF, -78 °C, 80%; (b) mCPBA, CH2Cl2, room temperature, 80%; (c) H2, Pd/C, -5 °C,
EtOH, 65%; (d) CDI, NaH, DMF, 81%; (e) L-Selectride, THF, -78 °C, 93%.
Scheme 21
Scheme 22
this juncture, it was necessary to develop mild protocols to
preserve the oxetane ring during the course of the transforma-
tions to be described. In various constructs and at various
stages, the hydoxyl group at C2 tended to displace the carbon-
oxygen link at C20 with formation of products of the type 72.
Directed epoxidation of 71 produced 73 which underwent
hydrogenation to produce diol 74, wherein the C1 and C2
hydroxyl groups could be protected as a cyclic carbonate (see
compound 75). With the cyclic carbonate in place, the ketone
76 could be obtained by conjugate reduction of 75.
The steroid derived ketone 70 was used to explore ring
closures which we hoped to employ in closing the C10-C11 bond
en route to baccatin III (1). Many permutations were investi-
gated. One setback was our inability to convert ketone type
77 (derived from 70) to the corresponding vinyl iodide (Scheme
21). Eventually, we were able to reach a vinyl iodide via a
circuitous route (Vide infra). Unfortunately, we were unable
to effect Barbier31 or Nozaki-Kishi32 type ring closures of 78
or 79 in the model steroid-derived series.
We then turned to the possibility of an intramolecular Heck
type olefination reaction33,34 to construct the C10 f C11 baccatin
bond. The feasibility of such a palladium-catalyzed internal
Heck reaction was first demonstrated in our laboratory with
aromatic C ring substrates (see transformations 80 f 81 and
82 f 83, Scheme 22).35
of the steroid-baccatin III hybrid, cyclization would involve
creation of a potentially serious C17-C19 methyl-methyl
abutment. Of course, just such a condition would have to be
met for such a reaction to be of value in the baccatin III (1)
directed enterprise.
To initiate the plan for the Heck reaction we needed to extend
the linker arm providing C9 and C10 of the baccatin target
(Scheme 23). This was accomplished as follows. The ketone
70 underwent conversion to vinyl triflate 84 under carefully
specified conditions. The dimethyl acetal was cleaved, and the
resultant aldehyde underwent Wittig olefination to produce 85.
The key advance in the program was realized when 85 under-
went intramolecular Heck closure. In this case, however, addi-
tion of stochiometric tetrakis(triphenylphosphine)palladium over
several hours was required and a 50% yield of 86 was realized.29
It seemed that the oxidative addition step to the hindered C11
vinyl triflate bond was not the limiting feature of the reaction.
Indeed, after several hours it was possible to isolate a palladium-
It was well appreciated that the successful Heck reaction of
substrates 80 and 82 could not be taken as secure precedents
for producing C-saturated derivatives. For instance, in the case
(31) For a review, see: Blomberg and Hartog. Synthesis 1977, 18-30.
(32) Kress, M. H.; Ruel, R.; Miller, W. H.; Kishi, Y. Tetrahedron Lett.
1993, 34, 5999-6003.
(33) Heck, R. F. Palladium Reagents in Organic Syntheses; Academic
Press: New York, 1985; p 179.
(34) For a review, see: Meijere, A.; Meyer, F. E. Angew. Chem., Int.
Ed. Engl. 1994, 33, 2379.
(35) (a) Masters, J. J.; Jung, D. K.; Bornmann, W. G.; Danishefsky, S.
J. Tetrahedron Lett. 1993, 34, 7253. (b) Young, W. B.; Masters, J. J.;
Danishefsky, S. J. J. Am. Chem. Soc. 1995, 117, 5228.

Total Synthesis of Baccatin III and Taxol
J. Am. Chem. Soc., Vol. 118, No. 12, 1996 2851
Scheme 23a
a Reagents: (a) KHMDS, PhNTf2, THF, -78 °C, 90%; (b) pTsOH, acetone, H2O, 75 °C, 80%; Ph3PCH3Br, KOtBu, THF, 0 °C, 95%; (c)
Pd(PPh3)4, K2CO3, CH3CN, 85 °C, 50%; (d) Pd(PPh3)4 (1.1 equiv), K2CO3, CH3CN, 85 °C, 2 h; I2, Et2O.
Scheme 24
a Reagents: (a) PhNTf2, KHMDS, THF, -78 °C, 98%; (b) PPTS, acetone, H2O, 96%; (c) Ph3PdCH2, THF, -78 °C f 0 °C, 77%; (d) Pd(PPh3)4,
K2CO3, CH3CN, 85 °C, 49%.
Scheme 25
substrates 78 and 79, discussed above.) The main difficulty in
the Heck reaction seemed to arise from the olefin insertion-
elimination phases. The prolonged reaction time may reflect
the strain of the bridgehead olefin and the close methyl-methyl
contact in the cyclized baccatin-steroid diene 86. A further
complication arising from precipitation of palladium black
during the reaction brought with it the requirement of stoichio-
metric amounts of the palladium mediator.
The important point, however, was that a pathway had been
discovered which was potentially relevant to synthesizing
baccatin III (1). Of course, it was not clear that corresponding
steps could be reduced to practice in the more functionalized
oxetane-containing series (Scheme 24). The previously dis-
cussed ketone 76 was converted to vinyl triflate 88. Hydrolysis
of the acetal afforded 89. The latter, upon Wittig olefination,
provided substrate 90. The intramolecular Heck reaction
proceeded in a fashion similar to our findings in the cholestanone-
deriVed series (cf. 85 f 86) and the tetracyclic 91 was in hand!
It had been hoped that the now extraneous carbon (C10′) of
the exocyclic methylene group attached to C10 would be easily
excised by oxidative cleavage (cf. 95). Unfortunately, the
concluding phase of the synthesis proved to be very difficult.
Harbingers of difficulties were already present from our model
studies. Several attempts to achieve selective cleavage of the
C10-C10′ double bond in Heck products 81, 83, and 86 were
unsuccessful. The matter was probed in detail in the case of
86 (Scheme 25). Both ozone and osmium tetroxide attacked
the tetrasubstituted double bond first. In the case of ozonolysis
(cf. 86) this was followed by cleavage of the exocyclic olefin
to produce compound 92. It seemed that oxidants were directed
to the tetrasubstituted double bond first. The only oxidizing
agent which selectively attacked the exocyclic double bond was
dimethyldioxirane. This reaction gave rise to compound 93.
containing product whose structure could not be specified in
detail. Treatment of this entity with iodine led to compound
87.36 (The latter was used to produce the potential cyclization
(36) Masters, J. J. Unpublished results.

2852 J. Am. Chem. Soc., Vol. 118, No. 12, 1996
Danishefsky et al.
Scheme 26a
a Reagents: (a) TBAF, THF, room temperature, 92%; (b) TESOTf, Et3N, CH2Cl2, -78 °C, 92%; (c) MCPBA, NaHCO3, CH2Cl2, room temperature,
45%; (d) H2, Pd/C, EtOH, room temperature, 82%; (e) Ac2O, DMAP, pyr, room temperature, 66%; (f) PhLi, THF, -78 °C, 93%; (g) OsO4, pyr,
105 °C; (h) Pb(OAc)4, PhH, MeOH, 0 °C, 61%; (i) SmI2, Ac2O, THF, -78 °C, 92%; (j) KOtBu, (PhSeO)2O, THF, -78 °C; KOtBu, THF, -78 °C,
81%; (k) Ac2O, DMAP, pyr, 76%.
Unfortunately, attempts to transform the epoxide of compound
93 to a system where C10′ could be excised were unsuccessful.
For instance, compound 93 was resistant to conversion to a
system of the type 94.
that it may be difficult to conduct debenzylation in the later
stages of the synthesis.
The issue of when the benzyl group would be discharged
caused much concern. Eventually, the courage was mustered
to attempt its removal at the stage of 97 by hydrogenolysis.
Remarkably, the exocyclic methylene (C10′) group, true to its
“tradition” of stability toward many chemical reactions, survived
this hydrogenolysis unscathed! The resultant C4 alcohol was
converted to its acetate, 99.
It is our sense that the unusual chemistry encountered during
explorations directed at oxidation of the exocyclic methyl group
of compound 86 can be accommodated in a common rationale.
Thus sp3 hybridization at C10 would cause a significant abutment
of its R group (if this group is not a proton) with the vinylic
C18 methyl group (baccatin numbering). The acute bond angles
of the epoxide, of course, lessen this impact. In these terms,
there would be significant resistance to the attack of any agents
at the exocyclic double bond which involve transition states
leading to higher membered rings with sp3-hybridized bonds
(cf. ozonolysis, osmium tetroxide, ruthenium tetroxide).
After significant improvisation, it eventually proved to be
possible to cleave the C10′ methylene group in the steroidal
hybrid, resulting in conversion of 86 to enone 95. An account
of these experiments was recently provided elsewhere.37 Here,
we return to compound 91 and describe its conversion to
baccatin III (1) (Scheme 26).
The first step in this concluding phase involved changing the
stabilizing arrangement at C7. The robust TBS group had
successfully protected that future alcohol function from various
chemical incursions. However, in model probes to be described
shortly, it was discovered that the removal of the TBS group
would be problematic in substrates which carry significant
functionality beyond that present in 91. Hence, the TBS group
was removed at virtually the last feasible stage, and the alcohol
at C7 was reprotected with a more labile TES group (see
conversion of 91 f 96).
In the next phase of the synthesis, the cyclic carbonate was
cleaved with phenyllithium to give the C2 benzoate (see
compound 100a)38 as precedented in the experiments of both
Nicolaou8,39 and Holton9 on closely related substrates. It was
at this stage that the exocyclic methylene group was to be
cleaved. The campaign started with the formation of putative
osmate ester 101 obtained by treatment with osmium tetroxide
under forcing conditions. This substance was treated with lead
tetraacetate, whereby ketone 102 was obtained in a 61% yield.
The now extraneous epoxide oxygen was removed through the
action of samarium iodide in the presence of acetic anhydride
(see compound 103).40 We were now obliged to face the
possible consequences of having kept C9 underfunctionalized
throughout the course of the synthesis. Fortunately, we could
draw upon the powerful chemistry of the Holton group to rectify
the problem.9 Thus, treatment of 103 with potassium tert-
butoxide and phenylseleninic anhydride led, after acetylation,
to the acetoxy ketone 105.
In our original conceptions about synthesizing taxol (3), we
had envisioned a late stage oxidation of C13. Indeed, in 1992
we reported the earliest experiments on the matter which
indicated that such an oxidation could work.41 However, a
closer and still more reassuring analogy was reported by
Nicolaou and co-workers.8 The La Jolla team had shown that
In this substrate, epoxidation of the tetrasubstituted double
bond to produce compound 97 could be accomplished, albeit
in only 45% yield. The complication here was the tendency
for additional oxidation at the exocyclic methylene to provide
an unwanted bis-epoxide. The next step involved adjustment
of the status of the masked alcohol at C4. Throughout the
journey from compound 24, we had benefited from the stability
of the benzyl protecting group. However, it was determined
that the benzyl group was now going to be vulnerable to some
of the required oxidizing agents. Model probes also suggested
(38) The stereochemistry of the epoxidation reaction was established
through a crystal structure on compound 100b. The latter is the C7-TBS,
C4-Bn version of compound 100a formed through analogous chemistry via
91.
(39) Nicolaou, K. C.; Renaud, J.; Nantermet, P. G.; Couladouros, E. A.;
Guy, R. K.; Wrasidlo, W. J. Am. Chem. Soc. 1995, 117, 2409. Nicolaou,
K. C.; Nantermet, P. G.; Ueno, H.; Guy, R. K. J. Chem. Soc., Chem.
Commun. 1994, 295.
(40) For a review on the uses of samarium diiodide, see: Molander, G.
A. Chem. ReV. 1992, 92, 29.
(41) Queneau, Y.; Krol, W. J.; Bornmann, W. G.; Danishefsky, S. J. J.
Org. Chem. 1992, 57, 4043.
(37) Young, W. B.; Link, J. T.; Masters, J. J.; Snyder, L. B.; Danishefsky,
S. J. Tetrahedron Lett. 1995, 36, 4963.

Total Synthesis of Baccatin III and Taxol
J. Am. Chem. Soc., Vol. 118, No. 12, 1996 2853
Scheme 27a
a Reagents: (a) PCC, NaOAc, PhH, reflux, 64%; (b) NaBH4, MeOH, 79%; (c) HF‚pyr, THF, 85%; (d) refs 8, 9, and 42.
the oxidation could be performed on the Very substrate in
question. Indeed, in formal terms, our expedition could have
been terminated by the synthesis of 105 since this compound
was described as an intermediate in the Nicolaou total synthesis.8
It was, however, our intention to complete a fully self-contained
independent synthesis of baccatin III (1) itself. For that purpose,
an allylic oxidation with PCC was carried out on our fully
synthetic material to give 106 (Scheme 27). This product, in
turn, was reduced with sodium borohydride to provide
7-OTES baccatin III (107). Compound 107, as well, had been
an intermediate in the Nicolaou total synthesis.8 For our
purposes, however, we removed the triethylsilyl group with
HF‚pyr. In so doing we completed our relay-free total syn-
thesis of baccatin III (1). The material, thus obtained, was
identical with a natural sample in all respects, including optical
rotation.
As indicated earlier, baccatin III (1) has been converted to
taxol (3) by numerous groups. Indeed, this transformation is
currently practiced on a commercial scale. Thus, there would
seem to be little need to execute such a conversion as part of
our own program. However, for sentimental reasons, we carried
out the formalities involved. For that purpose, we substantially
followed the regimen of Ojima and co-workers42 to reach
â-lactam 108a. This was converted to 108b which served to
acylate 7-TES-baccatin III (107 obtained from natural 10-
deacetylbaccatin III (2)). By following these previously de-
scribed protocols8,9,42 we completed a total synthesis of taxol
(3), identical with the natural product in all respects, including
optical rotation.
the synthesis to and through this step also has taught us much
about the stability, as well as the vulnerability, of this fascinating
structural element. Unfortunately, the procedure for cleaving
the extraneous C10′ exo-methylene group proved to be unexpect-
edly complicated. The setbacks sustained on this front eroded
the efficiency of the total synthesis.
Nonetheless, much has been learned about the chemical nature
of baccatin-like structures and about bicyclic and tricyclic
analogues en route to the baccatin skeleton. We are currently
developing an oxetane-containing analog synthesis program
which draws heavily from the chemistry used in the total
synthesis. Results of this effort will be disclosed shortly.
Experimental Section
General Procedures. 1H NMR and 13C NMR spectra were recorded
on a Bruker AMX-400. Infrared spectra were recorded on a Perkin-
Elmer 1600 Series FTIR. Optical rotations were measured on a JASCO
DIP-370 polarimeter. Mass spectra were obtained on a JOEL JMS-
DX-303 HF mass spectrometer. Analytical chromatography was
performed on E. Merck silical gel 60 F254 plates (0.25 mm). Flash
chromatography was performed on Mallinckrodt silica gel 60 (230-
400 mesh). Tetrahydrofuran (THF) was distilled from sodium metal/
benzophenone ketyl. Dichloromethane (CH2Cl2), acetonitrile, benzene
(PhH), triethylamine, and pyridine were distilled from calcium hydride.
N,N-Dimethylformamide (DMF) was purchased from Aldrich in sure
seal containers. All other commercially obtained reagents were used
as received.
Ketone 11. A solution of 1043 (30.9 g, 91.3 mmol) in 300 mL of
THF at 0 °C was treated dropwise with BH3 (1 M solution in THF,
93.0 mL, 93.0 mmol). After complete addition, the mixture was slowly
allowed to warm to room temperature and stir for 13 h. The mixture
was cooled to 0 °C and treated dropwise with H2O (7.1 mL), followed
by 3 M NaOH (50 mL) and 30% H2O2. After 1 h, the mixture was
warmed to room temperature, stirred for 3 h, and concentrated under
reduced pressure. The residue was dissolved ether (600 mL) and H2O
(300 mL). The aqueous layer was washed with ether (4 × 200 mL)
and the combined extracts were washed with saturated NaCl (2 × 200
Conclusions
The total synthesis of baccatin III (1) has been achieved.
From baccatin III (1), following known protocols, taxol (3) itself
was obtained. The most rewarding aspect of the synthesis was
the ability to start with Wieland-Miescher ketone, itself
available through catalytic asymmetric induction, and to install
all of the stereochemistry required to reach baccatin III in a
sequential fashion. Our synthesis, though arduous, involves no
relays, no resolutions, and no recourse to awkwardly available
antipodes of the “chiral pool”.
In these studies we have discovered a variety of efficient
degradative schemes by which differentiated appendages mounted
to optically pure CD fragments could be retrieved. An original
protocol for delivering a synthetic equivalent of 60 was
developed and found to be compatible with an oxetane-
containing CD fragment. From here, after coupling vinyllithium
reagent 66 to CD construct 45, and appropriate functional group
manipulations, the vinyl triflate 90 was prepared. The defining
bond construction in the synthesis was the Heck reaction to
produce 91. The ability to carry out this reaction in such a
complicated setting could not have been anticipated in advance.
The capacity to carry the oxetane from the earliest stages of
mL), dried (MgSO4), and concentrated under reduced pressure.
A
solution of the crude carbinol in 300 mL of CH2Cl2 at 0 °C was treated
with PDC (66.7 g, 177 mmol) and 4 Å sieves (66.5 g) in portions over
a period of 1 h. The mixture was allowed to stir at 0 °C for 0.5 h and
then warmed to room temperature and stirred for 10 h. The mixture
was diluted with ether (500 mL), treated with Celite (50 g), and filtered
through a pad of Celite. The filtrate was concentrated under reduced
pressure, affording the crude ketone as a brown oil. A solution of the
crude ketone in 250 mL of MeOH at room temperature was treated
with NaOMe (25 wt % in MeOH, 28 mL). After 10 h, the mixture
was concentrated under reduced pressure and the residue was dissolved
in ether (500 mL) and H2O (300 mL). The aqueous layer was washed
with ether (2 × 150 mL), and the combined extracts were washed with
H2O (200 mL) and saturated NaCl (200 mL), dried (MgSO4), filtered
through Celite/MgSO4, and concentrated under reduced pressure,
affording 30.0 g (93%) of the trans-fused ketone 11 as an orange oil:
1H NMR (400 MHz, CDCl3) δ 3.97-3.83 (m, 4H), 3.78 (dd, J ) 11.1,
5 Hz, 1H), 2.46 (dd, J ) 12.3, 3.7 Hz, 1H), 2.41-2.27 (m, 2H), 2.00-
1.58 (m, 7H), 1.42 (dt, J ) 13.4, 4.9 Hz, 1H), 0.88 (s, 9H), 0.78 (s,
(42) Cf.: Ojima, I.; Sun, C. M.; Zucco, M.; Park, Y. H.; Duclos, O.;
Kuduk, S. Tetrahedron Lett. 1993, 34, 4149. Ojima, I.; Habus, I.; Zucco,
M.; Park, Y. M.; Sun, C. M.; Brigaud, T. Tetrahedron 1992, 48, 6985.
(43) Enantiomerically pure 10 was generated via the same procedures
as described for racemic 10; see ref 17.

2854 J. Am. Chem. Soc., Vol. 118, No. 12, 1996
Danishefsky et al.
3H), 0.07 (s, 6H); 13C NMR (100 MHz, CDCl3) δ 210.4, 109.0, 77.3,
64.3, 64.2, 52.2, 46.7, 42.4, 38.8, 35.1, 30.7, 30.5, 29.6, 25.8, 18.0,
10.5, -4.1, -4.8; IR (film) 2952, 2857, 1716, 1110, 1093, 1051, 869,
to -78 °C, treated with trifluoromethanesulfonic anhydride (14.5 mL,
86.4 mmol), and then allowed to warm to room temperature. After
1.5 h, ethylene glycol (90.0 mL, 1.62 mol) was added and the mixture
was heated at reflux for 14.5 h. The mixture was then cooled and
poured into saturated NaCl (400 mL) and saturated NaHCO3 (400 mL).
The aqueous layer was washed with CH2Cl2 (3 × 250 mL), and the
combined extracts were washed with 1:1 H2O:saturated CuSO4 (400
mL), saturated NaHCO3 (250 mL), and saturated NaCl (250 mL), dried
(MgSO4), and concentrated under reduced pressure. The residue was
purified by flash chromatography (SiO2, 3:2 to 1:1 hexane:EtOAc),
837 cm-1; [R]20 11.1° (c ) 1.8, CHCl3).
D
Spiroepoxide 14. A solution Me3S+I- (25.9 g, 127 mmol) in 275
mL of THF at 0 °C was treated with KHMDS (23.1 g, 110 mmol) in
portions over a period of 0.5 h. After 1 h, a solution of 11 (30.0 g,
84.6 mmol) in 100 mL of THF was added via cannula over a period of
0.5 h. After 0.5 h, the mixture was treated with H2O (25 mL) and
concentrated under reduced pressure. The residue was dissolved in
ether (500 mL) and H2O (300 mL). The aqueous layer was washed
with ether (150 mL), and the combined extracts were washed with H2O
(150 mL) and saturated NaCl (150 mL), dried (MgSO4), and concen-
trated under reduced pressure, affording 31.1 g (99%) of 14 as a light
yellow oil: 1H NMR (400 MHz, CDCl3) δ 3.93-3.89 (dd, J ) 10.8,
4.4 Hz, 1H), 3.30 (m, 4H), 2.63 (d, J ) 4.4 Hz, 1H), 2.23 (d, J ) 4.4
Hz, 1H), 1.99-1.92 (m, 1H), 1.87-1.76 (m, 3H), 1.64-1.58 (m, 3H),
1.48 (t, J ) 13.2 Hz, 1H), 1.24-1.19 (m, 3H), 0.94 (s, 3H), 0.85 (s,
9H), 0.01 (s, 6H); 13C NMR (100 MHz, CDCl3) δ 109.6, 78.8, 64.3,
64.1, 58.4, 47.1, 41.0, 40.1, 35.3, 32.9, 30.5, 30.4, 28.6, 25.8, 18.0,
10.7, -4.0, -4.8; IR (neat) 2950, 2856, 1471, 1252, 1106, 872, 836
1
affording 5.5 g (66%) of 22 as a white solid: mp 152-153 °C; H
NMR (400 MHz, CDCl3) δ 4.78 (dd, J ) 9.1, 2.2 Hz, 1H), 4.46 (d, J
) 7.6 Hz, 1H), 4.26 (d, J ) 7.6 Hz, 1H), 3.96-3.89 (m, 4H), 3.44
(dd, J ) 10.6, 7.2 Hz, 1H), 2.40 (s, 1H), 2.27 (ddd, J ) 16.3, 9.2, 7.1
Hz, 1H), 1.88 (ddd, J ) 15.1, 10.7, 2.4 Hz, 1H), 1.83-1.72 (m, 2H),
1.68-1.55 (m, 5H), 1.21 (s, 3H), 0.87 (s, 9H), 0.02 (s, 6H); 13C NMR
(100 MHz, CDCl3) δ 108.9, 88.2, 77.1, 76.4, 73.9, 64.3, 64.2, 46.6,
37.5, 36.5, 30.9, 30.1, 25.8, 18.0, 9.5, -4.1, -4.8; IR (film) 3416,
1094, 859, 836, 774 cm-1; [R]20 19.0° (c ) 3.65, CHCl3).
D
Benzyloxetane 24. A solution of 22 (7.40 g, 19.2 mmol) in 100
mL of THF at 0 °C was treated with NaH (2.3 g, 96 mmol). After 0.5
h, the mixture was treated with benzyl bromide (3.0 mL, 25 mmol)
and TBAI (2.13 g, 5.76 mmol) and allowed to warm to room
temperature. After 4 h, the mixture was cooled to 0 °C, and saturated
NH4Cl was added until a homogeneous biphasic mixture resulted. The
mixture was poured into ether (700 mL) and saturated NaCl (300 mL).
The aqueous phase was washed with ether (2 × 200 mL), and the
combined extracts were washed with saturated Na2S2O3 (2 × 150 mL)
and saturated NaCl (150 mL), dried (MgSO4), and concentrated under
reduced pressure. The residue was purified by flash chromatography
(SiO2, 9:1 to 17:3 hexane:EtOAc), affording 8.90 g (98%) of 24 as a
cm-1; [R]20 -0.7° (c ) 2.5, CHCl3).
D
Allylic Alcohol 15. A solution of 14 (31.1 g, 84.5 mmol) and Al-
(iPrO)3 (50.0 g, 245 mmol) in 300 mL of toluene was heated at reflux
for 38 h. The mixture was cooled to room temperature, diluted with
ether (1100 mL), and treated with saturated Na+K+ tartrate (500 mL).
The layers were separated, and the aqueous layer was treated with 1
M NaOH (500 mL) and washed with ether (3 × 300 mL). The
combined organic extracts were washed with saturated NaCl (500 mL),
dried (MgSO4), filtered through Celite, and concentrated under reduced
pressure, affording 30.9 g (99%) of 15 as a light yellow solid: mp
1
1
72-73 °C; H NMR (400 MHz, CDCl3) δ 5.57 (m, 1H), 4.03-3.92
white solid: mp 110-111 °C; H NMR (400 MHz, CDCl3) δ 7.38-
(m, 6H), 3.53 (dd, J ) 9.8, 6.5 Hz, 1H), 2.49 (br d, 1H), 2.23 (br d,
1H), 2.02 (m, 1H), 1.90 (dt, J ) 12.9, 2.5 Hz, 1H), 1.83 (dq, J ) 13.0,
2.5 Hz, 1H), 1.75 (dd, J ) 13.7, 4.7 Hz, 1H), 1.67 (dq, J ) 13.8, 2.5
Hz, 1H), 1.56 (t, J ) 13.2 Hz, 1H), 1.21 (dt, J ) 12.8, 4.5 Hz, 1H),
0.87 (s, 9H), 0.79 (s, 3H), 0.02 (s, 6H); 13C NMR (100 MHz, CDCl3)
δ 137.8, 123.0, 109.4, 75.8, 65.0, 64.23, 64.17, 41.6, 37.4, 33.3, 33.2,
32.1, 30.8, 25.8, 18.0, 8.9, -4.1, -4.8; IR (film) 3416, 2951, 2856,
1472, 1360, 1250, 1106, 1072, 872, 836, 774 cm-1; [R]20D 25.4° (c )
1.4, CHCl3).
7.24 (m, 5H), 4.95 (dd, J ) 9.0, 1.6 Hz, 1H), 4.55 (d, J ) 11.6 Hz,
1H), 4.42-4.38 (m, 3H), 3.95-3.89 (m, 4H), 3.41 (dd, J ) 10.4, 7.4
Hz, 1H), 2.22-2.20 (m, 1H), 2.00-1.91 (m, 2H), 1.85-1.79 (m, 2H),
1.68-1.58 (m, 3H), 1.30-1.25 (m, 1H), 1.25 (s, 3H), 0.87 (s, 9H),
0.02 (s, 3H), 0.01 (s, 3H); 13C NMR (100 MHz, CDCl3) δ 138.3,
128.4, 127.6, 127.3, 109.0, 82.8, 79.0, 76.7, 74.6, 65.0, 64.3, 64.2, 41.3,
38.1, 37.5, 36.9, 31.1, 29.5, 25.8, 18.0, 9.9, -4.0, -4.8; IR (film)
2951, 2856, 1471, 1360, 1256, 1095, 837 cm-1; [R]20D 33.3° (c ) 6.3,
CHCl3).
Triol 16.44 A solution of 15 (30.9 g, 84.5 mmol) and NMO (19.8
g, 170 mmol) in 1300 mL of acetone and 170 mL of H2O at room
temperature was treated with OsO4 (5 wt % in iPrOH, 42.0 mL, 4.23
mmol). After 15 h, the mixture was treated with saturated NaHSO3
(400 mL) and stirred for 0.5 h. The mixture was then poured into
EtOAc (400 mL) and saturated NaCl (400 mL). The aqueous layer
was washed with EtOAc (3 × 300 mL), and the combined extracts
were concentrated under reduced pressure. The residue was dissolved
in EtOAc (750 mL) and washed with saturated NaCl (300 mL), dried
(MgSO4), and concentrated under reduced pressure. The residue was
purified by flash chromatography (SiO2, EtOAc), affording 15.3 g (45%)
Ketone 25. A solution of 24 (8.90 g, 18.8 mmol) in 950 mL of
15:1 acetone:H2O was treated with pTsOH (4.64 g, 24.4 mmol), and
the mixture was heated at reflux for 2.5 h. The mixture was cooled to
room temperature, treated with saturated NaHCO3, and concentrated
with a stream of nitrogen. The residue was dissolved in ether (800
mL) and poured into saturated NaHCO3 (300 mL). The aqueous phase
was washed with ether (150 mL) and CH2Cl2 (200 mL). The combined
extracts were washed saturated NaHCO3 (2 × 200 mL) and saturated
NaCl (150 mL), dried (MgSO4), and concentrated under reduced
pressure. The residue was purified by flash chromatography (SiO2,
3:1:1 hexane:CH2Cl2:EtOAc), affording 6.80 g (84%) of 25 as a white
solid: 1H NMR (400 MHz, CDCl3) δ 7.36-7.28 (m, 5H), 5.04 (dd, J
) 9.2, 2.1 Hz, 1H), 4.59 (d, J ) 11.6 Hz, 1H), 4.49 (d, J ) 11.7 Hz,
1H), 4.50-4.44 (m, 2H), 3.43 (dd, J ) 10.6, 7.2 Hz, 1H), 2.48 (ddd,
J ) 15.6, 14.0, 6.5 Hz, 1H), 2.39-2.38 (m, 1H), 2.34 (s, 1H), 2.32 (s,
1H), 2.30-2.24 (m, 1H), 2.15 (ddd, J ) 13.2, 6.3, 2.0 Hz, 1H), 2.03-
1.94 (m, 2H), 1.42 (s, 3H), 1.36 (dt, J ) 14.1, 5.0 Hz, 1H), 0.89 (s,
9H), 0.03 (s, 3H), 0.02 (s, 3H); 13C NMR (100 MHz, CDCl3) δ 210.4,
137.9, 128.5, 127.7, 127.2, 82.4, 79.1, 76.3, 73.2, 65.1, 45.1, 38.9, 37.9,
37.7, 37.6, 36.6, 25.8, 18.0, 10.0, -4.0, -4.9; IR (film) 2948, 1707,
1
of 16 as an off-white solid: mp 157-158 °C; H NMR (400 MHz,
CDCl3) δ 3.98-3.90 (m, 5H), 3.77 (dd, J ) 11.5, 5.5 Hz, 1H), 3.56
(m, 2H), 3.10 (s, 1H), 2.62 (d, J ) 2.4 Hz, 1H), 2.31-2.29 (m, 1H),
2.06 (dd, J ) 14.7, 2.7 Hz, 1H), 1.87 (m, 2H), 1.80-1.70 (m, 2H),
1.62 (m, 2H), 1.52 (t, J ) 13.2 Hz, 1H), 1.29 (td, J ) 12.6, 6.4 Hz,
1H), 0.87 (s, 9H), 0.82 (s, 3H), 0.06 (s, 3H), 0.04 (s, 3H); 13C NMR
(100 MHz, CDCl3) δ 109.3, 73.8, 73.7, 69.3, 64.2, 64.18, 62.0, 42.4,
38.8, 37.9, 34.8, 30.8, 30.1, 25.8, 18.0, 12.3, -4.1, -4.8; IR (film)
3441, 2950, 2856, 1256, 1102, 864, 835 cm-1; [R]20 32.3° (c ) 1.1,
D
1455, 1255, 1098, 940, 734 cm-1; [R]20 27.4° (c ) 6.3, CHCl3).
CHCl3).
D
Oxetane 22.45 A solution of 16 (8.70 g, 21.6 mmol) and pyridine
(19.2 mL, 238 mmol) in 540 mL of CH2Cl2 at -78 °C was treated
with chlorotrimethylsilane (2.8 mL, 22.0 mmol). The mixture was
allowed to warm to room temperature, and chlorotrimethylsilane (280
µL, 2.2 mmol) was added every hour until starting material was
consumed as determined by TLC. After 3 h, the mixture was cooled
r-Hydroxy Ketone 39. A solution of 27 (7.20 g, 16.7 mmol) and
triethylamine (23.3 mL, 167 mmol) in 450 mL of CH2Cl2 at -78 °C
was treated with trimethylsilyl trifluoromethanesulfonate (16.7 mL, 83.5
mmol). After 4 h, the mixture was treated with saturated NaHCO3
(25 mL) and allowed to warm for 10 min. The mixture was poured
into hexane (900 mL) and saturated NaHCO3 (250 mL). The aqueous
phase was washed with hexane (200 mL), and the combined extracts
were washed with saturated NaHCO3 (2 × 250 mL) and saturated NaCl
(150 mL), quickly dried (MgSO4), and concentrated under reduced
pressure. The crude enol ether was dissolved in CH2Cl2 (300 mL),
(44) Formation of side product 17 has been described previously; see
ref 17.
(45) Formation of side product 23 has been described previously; see
ref 17.

Total Synthesis of Baccatin III and Taxol
J. Am. Chem. Soc., Vol. 118, No. 12, 1996 2855
cooled to 0 °C, and rapidly treated with 3,3-dimethyldioxirane (0.1 M
solution in acetone, 180 mL, 18.0 mmol). After 10 min, the mixture
was concentrated under reduced pressure. The residue was dissolved
in acetone (350 mL) and treated with CSA (106 mg). After 0.5 h, the
mixture was concentrated under reduced pressure and the residue was
purified by flash chromatography (SiO2, 6.5:2:1.5 to 6:2:2 hexane:CH2-
Cl2: EtOAc), affording 6.60 g (89%) of 39 as a white solid: mp 115-
o-Nitrophenyl Selenide 43. A solution of crude 42 and (2-
nitrophenyl)selenyl cyanide (3.15 g, 13.9 mmol) in 130 mL of THF at
room temperature was treated with tributylphosphine (3.73 mL, 15.2
mmol). After 0.25 h, the mixture was concentrated under reduced
aspirator pressure in a well-ventilated hood. The residue was slurried
in 17:3 hexane:EtOAc and purified by flash chromatography (SiO2,
17:3 hexane:EtOAc), affording 7.20 g (88%) of 43 as a yellow foam:
1H NMR (400 MHz, CDCl3) δ 8.24 (d, J ) 8.0 Hz, 1H), 7.39-7.24
(m, 8H), 4.90 (dd, J ) 9.6, 4.4 Hz, 1H), 4.69 (d, J ) 4.3 Hz, 1H), 4.66
(s, 1H), 4.57-4.50 (m, 3H), 3.72 (dd, J ) 11.6, 5.7 Hz, 1H), 3.33 (s,
3 H), 3.32 (s, 3), 3.04 (dt, J ) 11.6, 4.6 Hz, 1H), 2.85-2.78 (m, 1H),
2.34-2.27 (m, 2H), 2.04-1.93 (m, 2H), 1.90 (d, J ) 3.5 Hz, 1H),
1.74-1.60 (m, 2H), 1.03 (s, 3H), 0.88 (s, 9H), 0.09 (s, 3H), 0.07 (s,
3H); 13C NMR (100 MHz, CDCl3) δ 146.8, 138.0, 133.7, 133.4, 129.0,
128.5, 127.7, 127.4, 126.5, 125.1, 101.5, 81.9, 76.3, 70.9, 64.6, 53.2,
51.5, 41.7, 40.4, 38.6, 37.1, 25.9, 25.6, 25.2, 18.1, 15.6, -3.4, -4.9;
1
116 °C; H NMR (400 MHz, CDCl3) δ 7.37-7.25 (m, 5H), 5.04 (dd,
J ) 9.0, 1.6 Hz, 1H), 4.59 (d, J ) 11.6 Hz, 1H), 4.51 (m, 1H), 4.42 (d,
J ) 7.9 Hz, 1H), 4.33 (dd, J ) 6.9, 12.4 Hz, 1H), 4.33 (dd, J ) 12.4,
6.9 Hz, 1H), 3.50-3.60 (br s, 1H), 3.44 (dd, J ) 10.3, 7.4 Hz, 1H),
2.58 (dd, J ) 10.3, 7.0 Hz, 1H), 2.52-2.46 (m, 2H), 2.27-2.26 (m,
1H), 2.01-1.96 (m, 2H), 1.51 (s, 3H), 1.20 (t, J ) 12.6 Hz, 1H), 0.88
(s, 9H), 0.03 (s, 6H); 13C NMR (100 MHz, CDCl3) δ 210.1, 137.7,
128.6, 127.8, 127.2, 82.3, 78.9, 76.1, 73.0, 72.5, 65.2, 48.7, 46.5, 39.4,
37.7, 34.6, 25.8, 17.9, 11.3, -4.0, -4.9; IR (film) 3446, 1719, 1471,
1251096, 838 cm-1; [R]20 37.2° (c ) 1.95, CHCl3).
IR (film) 2952, 2855, 1513, 1332, 1087, 984, 837, 730 cm-1; [R]20
D
D
2.7° (c ) 0.85, CHCl3).
Aldehyde 40. A solution of 39 (3.90 g, 8.72 mmol) in 170 mL of
1:1 C6H6:MeOH at 0 °C was treated with lead(IV) acetate (5.05 g,
11.4 mmol). After 10 min, the mixture was diluted with ether (600
mL), affording a yellow slurry which was then filtered through SiO2
(300 mL) using ether as eluent. The clear filtrate was concentrated
under reduced pressure. The residue was dissolved in ether (800 mL),
washed with saturated NaHCO3 (2 × 150 mL) and saturated NaCl (150
mL), dried (MgSO4), and concentrated under reduced pressure, affording
4.4 g (97%) of 40 as a clear oil: 1H NMR (400 MHz, CDCl3) δ 9.78
(s, 1H), 7.37-7.26 (m, 5H), 4.89 (dd, J ) 9.2, 3.5 Hz, 1H), 4.61 (d,
J ) 11.2 Hz, 1H), 4.54 (d, J ) 7.8 Hz, 1H), 4.53 (d, J ) 7.8 Hz, 1H),
4.37 (d, J ) 11.2 Hz, 1H), 3.99 (dd, J ) 11.2, 6.1 Hz, 1H), 3.48 (s,
3H), 2.80 (dd, J ) 9.9, 7.8 Hz, 1H), 2.62 (d, J ) 17.1 Hz, 1H), 2.41
(dd, J ) 14.8, 5.0 Hz, 1H), 2.36-2.21 (m, 3H), 1.96-1.94 (m, 1H),
1.22 (s, 3H), 0.87 (s, 9H), 0.04 (s, 3H), 0.03 (s, 3H); 13C NMR (100
MHz, CDCl3) δ 201.3, 173.1, 137.9, 128.4, 127.60, 127.55, 81.8, 80.5,
75.9, 71.8, 64.9, 51.8, 50.4, 41.9, 40.6, 36.9, 31.0, 25.8, 18.0, 14.4,
-3.8, -4.7.
Acetal 41. A solution of crude 40 in 1100 mL of MeOH was treated
with collidinium p-toluenesulfonate (910 mg, 3.10 mmol), and the
mixture was heated at reflux. After 10 h, the mixture was concentrated
under reduced pressure and the residue was dissolved in ether (600
mL). The resulting white slurry was then filtered through basic alumina
(150 mL) using ether as eluent. The filtrate was concentrated under
reduced pressure, affording 4.40 g (97%) of 41 as a clear oil: 1H NMR
(400 MHz, CDCl3) δ 7.37-7.23 (m, 5H), 4.83 (dd, J ) 9.2, 3.8 Hz,
1H), 4.66-4.62 (m, 2H), 4.58 (d, J ) 11.1 Hz, 1H), 4.51 (d, J ) 7.7
Hz, 1H), 4.31 (d, J ) 11.1 Hz, 1H), 3.78 (dd, J ) 10.9, 6.0 Hz, 1H),
3.43 (s, 3H), 3.33 (s, 3H), 3.32 (s, 3H), 2.58-2.53 (m, 2H), 2.28 (ddd,
J ) 9.3, 5.9, 14.9 Hz, 1H), 2.20 (dd, J ) 10.5, 13.8 Hz, 1H), 2.05-
1.95 (m, 2H), 1.50 (dd, J ) 14.8, 5.6 Hz, 1H), 1.07 (s, 3H), 0.89 (s,
9H), 0.09 (s, 3H), 0.06 (s, 3H); 13C NMR (100 MHz, CDCl3) δ 173.5,
138.2, 128.3, 127.49, 127.46, 101.6, 81.8, 81.1, 76.1, 71.2, 64.5, 52.5,
51.6, 39.9, 39.7, 38.8, 37.3, 31.1, 25.8, 18.1, 15.6, -3.7, -4.9; IR (film)
2952, 2856, 1737, 1255, 1.89, 991, 837 cm-1; [R]20D 18.4° (c ) 2.25,
CHCl3).
Olefin 44. A solution of 43 (7.20 g, 11.1 mmol) in 140 mL of
THF at room temperature was treated with 11.5 mL of 30% H2O2.
After 12 h, the mixture was poured into ether (800 mL) and 10% NaOH
(200 mL). The aqueous layer was washed with ether (250 mL), and
the combined extracts were washed with 10% NaOH (3 × 150 mL)
and saturated NaCl (200 mL) and dried (MgSO4). Concentration under
reduced pressure afforded 4.72 g (90%) of 44 as a light yellow oil: 1H
NMR (400 MHz, CDCl3) δ 7.38-7.24 (m, 5H), 6.00-5.91 (m, 1H),
5.18 (d, J ) 17.2 Hz, 1H), 5.15 (dd, J ) 17.3, 10.2 Hz, 1H), 4.90 (dd,
J ) 8.0, 3.4 Hz, 1H), 4.63 (d, J ) 7.4 Hz, 1H), 4.59 (t, J ) 5.3 Hz,
1H), 4.55 (d, J ) 11.8 Hz, 1H), 4.51 (d, J ) 11.6 Hz, 1H), 4.45 (d, J
) 7.4 Hz, 1H), 3.72 (dd, J ) 9.2, 5.4 Hz, 1H), 3.26 (s, 6H), 2.77 (d,
J ) 9.3 Hz, 1H), 2.28 (ddd, J ) 3.2, 9.4, 13.4 Hz, 1H), 2.00-1.93 (m,
1H), 1.71 (dd, J ) 14.4, 4.7 Hz, 1H), 1.61 (dd, J ) 14.3, 4.7 Hz, 1H),
1.12 (s, 3H), 0.90 (s, 9H), 0.78 (s, 6H); 13C NMR (100 MHz, CDCl3)
δ 138.5, 135.0, 128.4, 127.5, 127.2, 119.5, 102.1, 82.9, 79.7, 77.2, 75.9,
71.9, 65.0, 52.5, 52.3, 48.7, 39.8, 39.2, 36.2, 25.9, 18.1, 16.9, -3.7,
-4.8; IR (film) 2948, 2929, 2850, 1461, 1250, 1078, 985, 832, 774
cm-1; [R]20D 15.9° (c ) 3.1, CHCl3); FAB HRMS m/e calcd for (M +
K) C27H44K1O5Si1 515.2586, found 515.2595.
Aldehyde 45. A solution of 44 (3.40 g, 7.15 mmol) and 5.25 g of
solid NaHCO3 in 525 mL of CH2Cl2 at -78 °C was treated with ozone
until a light blue solution resulted. The excess ozone was removed by
passing a stream of nitrogen through the solution for 5 min, and the
mixture was then treated with PPh3 (1.87 g, 7.15 mmol) and allowed
to warm to room temperature. After 0.5 h, the mixture was concentrated
under reduced pressure and the residue was purified by flash chroma-
tography (SiO2, 9:1 to 4:1 hexane:EtOAc), affording 2.7 g (79%) of
45 as a clear oil: 1H NMR (400 MHz, CDCl3) δ 9.99 (d, J ) 2.6 Hz,
1H), 7.34-7.32 (m, 5H), 4.92 (dd, J ) 7.6, 2.4 Hz, 1H), 4.65 (d, J )
8.0 Hz, 1H), 4.56-4.49 (m, 3H), 4.44 (d, J ) 11.2 Hz, 1H), 3.72 (dd,
J ) 6.4, 4.8 Hz, 1H), 3.25 (s, 3H), 3.24 (s, 3H), 2.96 (d, J ) 2.8 Hz,
1H), 2.25-2.18 (m, 1H), 2.00-1.94 (m, 1H), 1.77 (dd, J ) 12.4, 4.8
Hz, 1H), 1.69 (dd, J ) 14.4, 6.0 Hz, 1H), 1.20 (s, 3H), 0.090 (s, 9H),
0.09 (s, 3H), 0.08 (s, 3H); 13C NMR (400 MHz, CDCl3) δ 203.9, 138.0,
128.4, 127.6, 127.4, 101.9, 83.1, 77.4, 77.2, 76.0, 71.7, 65.5, 56.7, 52.7,
52.4, 39.6, 39.3, 34.1, 25.8, 19.1, 18.1, -4.0, -4.9; IR (film) 2929,
Alcohol 42. A solution of 41 (6.60 g, 12.6 mmol) in 20 mL of
THF was transferred via cannula to a 0 °C solution of LiAlH4 (623
mg, 16.4 mmol) in 180 mL of THF. After 1 h, the mixture was treated
with EtOAc (20 mL) followed by saturated NH4Cl (2 mL). After 0.5
h, the mixture was diluted with EtOAc (500 mL), treated with SiO2
(100 mL), and filtered through Celite and MgSO4 using EtOAc as
eluent. The filtrate was concentrated under reduced pressure, affording
2.78 g (quant.) of 42 as a clear oil: 1H NMR (400 MHz, CDCl3) δ
7.38-7.28 (m, 5H), 4.90 (dd, J ) 9.6, 4.6 Hz, 1H), 4.65-4.64 (m,
2H), 4.57-4.49 (m, 3H), 3.72 (dd, J ) 11.8, 5.6 Hz, 1H), 3.56 (q, J )
5.9 Hz, 2H), 3.30 (s, 3H), 3.28 (s, 3H), 2.98 (t, J ) 6.4 Hz, 1H), 2.29
(ddd, J ) 5.7, 9.7, 14.8 Hz, 1H), 2.19 (d, J ) 9.4 Hz, 1H), 1.99-1.90
(m, 2H), 1.82-1.79 (m, 1H), 1.64 (dd, J ) 14.7, 6.7 Hz, 1H), 1.47-
2856, 1708, 1471, 1387, 1254, 1089, 837, 775, 697 cm-1; [R]20 9.4°
D
(CHCl3, c ) 0.51); FAB HRMS m/e calcd for (M + K) C26H42K1O6-
Si1 517.2388, found 517.2379.
Monohydrazone 62. A solution of 611 (34.5 g, 224 mmol) in 250
mL of EtOH was treated with triethylamine (93.3 mL, 669 mmol) and
hydrazine monohydrate (21.7 mL, 447 mmol). After 1.5 h, the reaction
mixture was concentrated under reduced pressure and the residue was
purified by flash chromatography (SiO2, ether), affording 27.0 g (72%)
of 62 as a colorless oil: 1H NMR (400 MHz, CDCl3) δ 5.05 (br s,
2H), 2.59 (ddd, J ) 9.3, 6.3, 3.0 Hz, 1H), 2.49 dd, J ) 11.2, 6.8 Hz,
1H), 2.42 (m, 1H), 2.05 (m, 1H), 1.62 (m, 1H), 1.29 (s, 3H), 1.25 (s,
3H), 1.10 (d, J ) 6.8 Hz, 3H); IR (film) 3385, 2973, 2933, 1708, 1635,
1459 cm-1; CI (NH3) MS m/e calcd for (M + H) C9H16N2O1 169, found
169.
1.46 (m, 1H), 1.05 (s, 3H), 0.88 (s, 9H), 0.08 (s, 3H), 0.07 (s, 3H); 13
C
NMR (100 MHz, CDCl3) δ 137.4, 128.6, 127.9, 127.6, 101.5, 82.0,
81.6, 76.2, 70.8, 64.9, 61.8, 52.5, 51.8, 40.2, 39.7, 38.1, 37.1, 29.1,
25.8, 18.1, 15.4, -3.5, -4.9; IR (film) 3460, 2953, 2856, 1463, 1362,
Vinyl Iodide 63. A solution of 62 (13.5 g, 80.4 mmol) and DBN
(98 mL, 793 mmol) in 2.7 L of ether was treated dropwise with a
solution of I2 (40.8 g, 161 mmol) in 1 L of ether over 40 min. After
1254, 1156, 1086, 984, 837, 775, 734 cm-1; [R]20 6.2° (c ) 2.7,
D
CHCl3).

2856 J. Am. Chem. Soc., Vol. 118, No. 12, 1996
Danishefsky et al.
0.75 h, the mixture was filtered through SiO2 using ether as eluent.
Concentration of the filtrate under reduced pressure and purification
of the residue by flash chromatography (SiO2, 9:1 to 4:1 hexane:EtOAc)
afforded 11.0 g (52%) of 63 as a colorless oil and 4.1 g (19%) of 64
as a yellow oil.
saturated NaHCO3 (200 mL), and the aqueous layer was washed with
ether (200 mL). The combined extracts were washed with saturated
NaHCO3 (200 mL) and saturated NaCl (200 mL), dried (Na2SO4), and
concentrated under reduced pressure. The residue was purified by flash
chromatography (SiO2, 4:1 to 3:1 to 2:1 petroleum ether:ether),
affording 1.21 g (80%) of 73 as a white foam: 1H NMR (400 MHz,
CDCl3) δ 7.35-7.24 (m, 5H), 6.27 (dd, J ) 3.1, 1.4 Hz, 1H), 5.39 (d,
J ) 9.0 Hz, 1H), 4.97 (d, J ) 11.4 Hz, 1H), 4.86 (app t, J ) 7.4 Hz,
1H), 4.77 (d, J ) 11.3 Hz, 1H), 4.68 (d, J ) 9.0 Hz, 1H), 4.64 (t, J )
5.0 Hz, 1H), 4.52 (s, 1H), 3.59 (dd, J ) 11.8, 4.3 Hz, 1H), 3.51 (d, J
) 4.5 Hz, 1H), 3.27 (s, 6H), 2.89 (s, 1H), 2.34-2.28 (m, 1H), 2.17-
2.10 (m, 2H), 1.76 (d, J ) 6.4 Hz, 1H), 1.62 (s, 3H), 1.44 (s, 3H),
1.26 (s, 3H), 1.15 (s, 3H), 0.92 (s, 9H), 0.11 (s, 6H); 13C NMR (100
MHz, CDCl3) δ 201.2, 138.5, 137.5, 135.5, 128.3, 127.9, 127.6, 102.2,
82.7, 81.7, 77.5, 77.2, 72.7, 67.4, 66.7, 64.9, 53.9, 52.7, 52.3, 48.0,
47.4, 41.8, 41.2, 35.9, 25.9, 24.1, 19.9, 18.1, 17.3, 16.4, -3.6, -4.6;
IR (film) 3493, 2952, 2856, 1683, 1471, 1384, 1254, 1082, 837, 775
cm-1; [R]20D -36.0° (c ) 5.8, CHCl3); FAB HRMS m/e calcd for (M
+ H) C35H55O8Si1 631.3666, found 631.3672.
Iodo Dienone 64. A solution of 63 (11 g, 41.6 mmol) in 500 mL
of THF at 0 °C was treated with DBN (103 mL, 825 mmol). Three
portions of I2 (53 g, 209 mmol total) were added over 30 min, and the
reaction mixture was allowed to warm to room temperature. After 8
h, the mixture was poured into ether (750 mL) and H2O (500 mL).
The organic layer was washed with H2O (2 × 500 mL), saturated
Na2S2O3 (2 × 300 mL), and saturated NaCl (400 mL), dried (Na2-
SO4), and concentrated under reduced pressure. Purification of the
residue by flash chromatography (SiO2, 4:1 hexane:EtOAc) afforded
9.8 g (89%) of 64 as a yellow oil: 1H NMR (400 MHz, CDCl3) δ 6.81
(d, J ) 6.7 Hz, 1H), 6.54 (dd, J ) 7.9, 0.2 Hz, 1H), 1.83 (s, 3H), 1.27
(s, 6H); 13C NMR (100 MHz, CDCl3) δ 202.7, 138.1, 132.7, 131.9,
120.9, 53.7, 28.5, 15.3; IR (film) 2971, 1655, 1560, 1456, 1374, 1361,
1033 cm-1; CI(NH3) MS m/e calcd for (M + H) C9H12O1I1 263, found
263.
Cyanohydrin 65. A solution of 64 (9.8 g, 37 mmol) in 150 mL of
CH2Cl2 at 0 °C was treated with TMSCN (6.0 mL, 45 mmol), KCN
(0.24 g, 3.7 mmol), and 18-crown-6 (1.0 g, 3.8 mmol). The reaction
was allowed to warm to room temperature and stir for 8 h. After
cooling to 0 °C, the reaction was quenched by the addition of saturated
NaHCO3 (200 mL). The mixture was poured into CH2Cl2 (250 mL)
and washed with saturated NaHCO3 (200 mL). The organic layer was
concentrated under reduced pressure in a fume hood. The residue was
purified by flash chromatography (SiO2, 19:1 hexane:ether), affording
12.0 g (89%) of 65 as a yellow oil: 1H NMR (400 MHz, CDCl3) δ
6.61 (d, J ) 6.0 Hz, 1H), 5.52 (ddd, J ) 6.0, 3.3, 1.6 Hz, 1H), 1.95 (s,
3H), 1.23 (s, 3H), 1.17 (s, 3H), 0.23 (s, 9H); IR (film) 2975, 1590,
1463, 1254, 1141, 1099, 844 cm-1; CI(NH3) MS m/e calcd for (M +
H2O) C13H22N1O1Si1I1 379, found 379.
Diol 74. A solution of 73 (936 mg, 1.49 mmol) and 10% Pd/C
(1.016 g) in EtOH (87 mL) at -5 °C was placed under an atmosphere
of hydrogen. After 1.25 h, the mixture was diluted with EtOAc, filtered
through Celite using EtOAc as eluent, and concentrated under reduced
pressure. The residue was purified by flash chromatography (SiO2,
3:1 to 7:3 to 3:2 petroleum ether:ether), affording 610 mg (65%) of 74
as a white foam: 1H NMR (400 MHz, CDCl3) δ 7.37-7.30 (m, 5H),
6.12 (br s, 1H), 5.48 (d, J ) 9.2 Hz, 1H), 5.05 (d, J ) 11.4 Hz, 1H),
4.96 (d, J ) 11.4 Hz, 1H), 4.85 (app t, J ) 7.4 Hz, 1H), 4.59 (d, J )
9.0 Hz, 1H), 4.52 (app t, J ) 5.4 Hz, 1H), 4.06 (d, J ) 5.6 Hz, 1H),
3.59 (dd, J ) 10.6, 4.4 Hz, 1H), 3.37 (s, 1H), 3.28 (s, 3H), 3.23 (s,
3H), 3.12 (br s, 1H), 2.58-2.44 (m, 2H), 2.38-2.32 (ddd, J ) 13.2,
8.4, 4.4 Hz, 1H), 2.22-2.17 (m, 1H), 2.12 (s, 1H), 1.69 (d, J ) 1.2
Hz, 3H), 1.63-1.60 (m, 1H), 1.56 (s, 1H), 1.20 (s, 3H), 1.16 (s, 3H),
1.06 (s, 3H), 0.90 (s, 9H), 0.11 (s, 3H), 0.10 (s, 3H); 13C NMR (100
MHz, CDCl3) δ 203.6, 137.3, 132.8, 128.8, 128.0, 127.8, 102.1, 83.6,
82.0, 78.6, 77.2, 71.8, 70.5, 65.7, 53.1, 52.1, 51.4, 41.4, 41.3, 36.1,
34.2, 25.9, 18.2, 16.7, 16.1, -3.5, -4.6; IR (film) 3438, 2947, 1464,
Carbinol 71. A solution of diene 65 (3.93 g, 10.9 mmol) in 130
mL of THF at -78 °C was treated dropwise with tBuLi (1.7 M solution
in pentane, 12.5 mL, 21.2 mmol). After 0.5 h, a solution of aldehyde
45 in 40 mL of THF was added dropwise to the mixture via cannula.
An additional 20 mL of THF was used to transfer any remaining
aldehyde. After 0.25 h, the mixture was treated with saturated NH4Cl
(50 mL), diluted with ether (200 mL), and allowed to warm to room
temperature. The mixture was poured into ether (400 mL) and saturated
NH4Cl (150 mL). The aqueous layer was washed with ether (200 mL),
and the combined extracts were washed with saturated NaCl (200 mL),
dried (Na2SO4), and concentrated under reduced pressure. The residue
was purified by flash chromatography (SiO2, 9:1 to 4:1 hexane:EtOAc),
affording 5.57 g of a mixture of diastereomers (white foam). This
mixture of diastereomers (3.04 g, 4.26 mmol) in 26 mL of THF at
-78 °C was then treated dropwise with tetrabutylammonium fluoride
(1.0 M solution in THF, 25.6 mL, 25.6 mmol). After 0.25 h, the
mixture was diluted with hexane (26 mL) and directly purified by flash
chromatography (SiO2, hexane to 9:1 to 4:1 hexane:EtOAc), affording
2.11 g (75% from 45) of 71 as a single diastereomer (white foam): 1H
NMR (400 MHz, CDCl3) δ 7.35-7.16 (m, 5H), 6.74 (d, J ) 6.6 Hz,
1H), 6.27 (d, J ) 6.6 Hz, 1H), 5.61 (d, J ) 8.1 Hz, 1H), 4.98 (d, J )
6.9 Hz, 1H), 4.90 (s, 1H), 4.81 (d, J ) 11.7 Hz, 1H), 4.74 (d, J ) 11.6
Hz, 1H), 4.68 (d, J ) 8.1 Hz, 1H), 4.53 (s, 1H), 4.43 (t, J ) 4.9 Hz,
1H), 3.75 (t, J ) 4.0 Hz, 1H), 3.26 (s, 3H), 3.18 (s, 3H), 2.38 (d, J )
2.4 Hz, 1H), 2.34-2.25 (m, 1H), 2.01 (d, J ) 15.8 Hz, 1H), 1.85 (s,
3H), 1.42-1.40 (m, 2H), 1.31 (s, 3H), 1.30 (s, 3H), 1.27 (s, 3H), 0.97
(s, 9H), 0.19 (s, 3H), 0.18 (s, 3H); 13C NMR (100 MHz, CDCl3) δ
206.2, 157.9, 138.5, 137.2, 130.6, 128.3, 127.3, 127.2, 117.7, 102.4,
82.2, 81.2, 77.2, 76.0, 74.7, 69.6, 64.5, 53.1, 52.6, 51.2, 50.8, 42.9,
40.4, 34.8, 29.6, 25.9, 21.7, 20.5, 18.3, 15.6, -4.4, -4.6; IR (film)
1383, 1255, 1078, 1046, 837 cm-1; [R]20 35.2° (c ) 0.27, CHCl3);
D
FAB HRMS m/e calcd for (M + H) C35H57O8Si1 633.3823, found
633.3835.
Carbonate 75. A solution of 74 (0.76 g, 1.20 mmol) and
carbonyldiimidazole (4.88 g, 30.1 mmol) in 40 mL of DMF was treated
with NaH (48.6 g, 1.92 mmol, 95%) in portions over 20 min. The
reaction was quenched by addition of saturated NH4Cl (200 mL), and
the mixture was poured into ether (500 mL). The organic layer was
washed with NaHCO3 (300 mL), H2O (300 mL), and saturated NaCl
(300 mL), dried (Na2SO4), and concentrated under reduced pressure.
The residue was purified by flash chromatography (SiO2, 4:1 hexane:
EtOAc), affording 0.64 g (81%) of 75 as a white solid: mp 249-250
1
°C; H NMR (400 MHz, CDCl3) δ 7.40-7.24 (m, 5H), 6.13 (br s,
1H), 5.21 (d, J ) 9.9 Hz, 1H), 4.95 (d, J ) 10.3 Hz, 1H), 4.88 (d, J
) 10.1 Hz, 1H), 4.87-4.78 (m, 3H), 4.49 (app t, J ) 5.3 Hz, 1H),
3.60 (dd, J ) 11.7, 4.4 Hz, 1H), 3.25 (s, 3H), 3.23 (s, 3H), 3.07-3.02
(m, 1H), 2.69-2.65 (m, 1H), 2.41-2.39 (m, 1H), 2.24-2.16 (m, 2H),
1.70 (s, 3H), 1.58-1.55 (m, 2H), 1.27 (s, 3H), 1.19 (s, 3H), 1.13 (s,
3H), 0.89 (s, 9H), 0.11 (s, 3H), 0.81 (s, 3H); 13C NMR (100 MHz,
CDCl3) δ 200.0, 154.0, 137.8, 137.4, 129.1, 128.4, 127.8, 102.1, 91.3,
83.4, 81.8, 78.8, 77.2, 76.8, 71.8, 65.8, 53.6, 53.0, 51.5, 46.4, 41.9,
41.4, 36.7, 29.8, 25.9, 18.1, 16.0, 14.9, -3.3, -4.6; IR (film) 2952,
1794, 1677, 1364, 1047, 838 cm-1; [R]20 15.3° (c ) 0.66, CHCl3);
D
FAB HRMS m/e calcd for (M + Na) C36H54Na1O9Si1 681.3434, found
681.3467.
Ketone 76. A solution of 75 (984 mg, 1.50 mmol) in 30 mL of
THF at -78 °C was treated dropwise with a solution of L-Selectride
(1.0 M solution in THF, 4.50 mL, 4.50 mmol). After 1.25 h, the
mixture was treated with saturated NH4Cl (5 mL), diluted with ether
(300 mL), and allowed to warm to room temperature. The mixture
was poured into ether (100 mL), and saturated NH4Cl (75 mL). The
aqueous phase was washed with ether (100 mL) and the combined
extracts were washed with saturated NH4Cl (75 mL) and saturated NaCl
(75 mL), dried (K2CO3), and concentrated under reduced pressure. The
residue was purified by flash chromatography (SiO2, 7:1.5:1.5 to 5:3:2
3378, 2929, 2856, 1654, 1471, 1376, 1256, 1050, 837 cm-1; [R]20
D
29.3° (c ) 0.7, CHCl3); FAB HRMS m/e calcd for (M+) C35H54O7Si1
614.3639, found 614.3623.
Epoxide 73. A solution of 71 (1.49 g, 2.43 mmoles) and NaHCO3
(2.97 g) in 69 mL of CH2Cl2 at room temperature was treated with
mCPBA (95%, 1.04 g, 6.05 mmol). After 11 h, the mixture was treated
with Me2S (890 µL, 12.1 mmol) and diluted with H2O (40 mL) and
ether (200 mL). The mixture was poured into ether (400 mL) and

Total Synthesis of Baccatin III and Taxol
J. Am. Chem. Soc., Vol. 118, No. 12, 1996 2857
hexane:CH2Cl2:ether), affording 920 mg (93%) of 76 as a white solid:
mp 272-273 °C dec; H NMR (400 MHz, CDCl3) δ 7.39-7.24 (m,
reduced pressure. The residue was purified by flash chromatography
(SiO2, 9:1 to 17:3 petroleum ether:ether), affording 771 mg (77%) of
90 as a white foam: 1H NMR (400 MHz, CDCl3) δ 7.43-7.30 (m,
5H), 5.66-5.56 (m, 1H), 5.22-5.20 (d, J ) 9.9 Hz, 1H), 5.10-5.00
(m, 2H), 4.96-4.80 (m, 4H), 4.70 (s, 1H), 3.64 (dd, J ) 11.6, 4.4 Hz,
1H), 2.50-2.37 (m, 3H), 2.24-2.15 (m, 1H), 2.08 (s, 1H), 1.87 (dd, J
) 16.0, 7.2 Hz, 1H), 1.80-1.72 (m, 1H), 1.64 (s, 3H), 1.59-1.41 (m,
1H), 1.24 (s, 3H), 1.19 (s, 3H), 1.14 (s, 3H), 0.88 (s, 9H), 0.09 (s, 3H),
0.02 (s, 3H); 13C NMR (100 MHz, CDCl3) δ 154.0, 144.2, 137.2, 131.9,
129.5, 128.4, 127.9, 126.6, 120.2, 119.4, 90.0, 83.0, 82.2, 78.1, 77.2,
69.4, 65.8, 44.9, 44.6, 42.2, 41.1, 36.6, 27.2, 25.8, 23.4, 23.3, 18.1,
17.4, 17.3, 15.9, -3.4, -4.7; IR (film) 2940, 1797 cm-1; [R]20D -22.7°
(c ) 1.1, CHCl3); FAB HRMS m/e calcd for (M + H) C36F3H52O9S1-
Si1 672.2584, found 672.2583.
1
5H), 5.18 (d, J ) 9.8 Hz, 1H), 4.97-4.81 (m, 4H), 4.75 (s, 1H), 4.52
(t, J ) 5.2 Hz, 1H), 3.58 (dd, J ) 12.4, 4.8 Hz, 1H), 3.28 (s, 3H),
3.27-3.25 (m, 1H), 3.25 (s, 3H), 2.58-2.39 (m, 2H), 2.34 (s, 1H),
2.22-2.16 (m, 1H), 1.97-1.86 (m, 1H), 1.67 (d, J ) 14.8, 5.2 Hz),
1.54 (d, J ) 14.8, 5.6 Hz, 1H), 1.28 (s, 3H), 1.24 (s, 3H), 1.10 (s, 3H),
0.95 (d, J ) 6.4 Hz, 3H), 0.90 (s, 9H), 0.11 (s, 3H), 0.09 (s, 3H); 13C
NMR (100 MHz, CDCl3) δ 211.3, 153.8, 137.4, 129.1, 128.3, 127.8,
102.3, 93.5, 83.1, 82.0, 78.9, 77.2, 72.2, 65.7, 53.9, 53.6, 53.0, 46.2,
41.9, 41.4, 39.1, 37.0, 29.0, 27.0, 25.9, 23.5, 18.2, 16.8, 14.7, 14.6,
-3.3, -4.6; IR (film) 2950, 1786, 1709, 1115, 1049, 840 cm-1; [R]20
D
8.1° (c ) 0.27, CHCl3); FAB HRMS m/e calcd for (M + Na) C36H56-
Na1O9 Si1 683.3591, found 683.3620.
Vinyl Triflate 88. A solution of 76 (920 mg, 1.39 mmol) and
N-phenyl bis(trifluoromethanesulfonimide) (1.25 g, 3.49 mmol) in 47
mL of THF at -78 °C was rapidly treated with KHMDS (0.5 M
solution in toluene, 3.34 mL, 1.67 mmol). After 10 min, the mixture
was treated with saturated NH4Cl (5 mL), diluted with ether (300 mL),
and allowed to warm to room temperature. The mixture was poured
into ether (100 mL) and saturated NH4Cl (75 mL). The aqueous phase
was washed with ether (100 mL), and the combined extracts were
washed with saturated NH4Cl (75 mL) and saturated NaCl (75 mL),
dried (K2CO3), and concentrated under reduced pressure. The residue
was purified by flash chromatography (SiO2, 9:1 to 17:3 petroleum
ether:ether), affording 1.08 g (98%) of 88 as a white foam contaminated
with <5% of N-phenyl trifluoromethanesulfonimide: 1H NMR (400
MHz, CDCl3) δ 7.42-7.31 (m, 5H), 5.22 (d, J ) 12.4 Hz, 1H), 4.94-
4.84 (m, 4H), 4.75 (s, 1H), 4.49 (app t, J ) 6.0 Hz, 1H), 3.61 (dd, J
) 11.6, 4.4 Hz, 1H), 3.22 (s, 3H), 3.21 (s, 3H), 2.45-2.33 (m, 3H),
2.25-2.18 (m, 1H), 1.75 (dd, J ) 16.8, 5.6 Hz, 1H), 1.64 (s, 3H),
1.66-1.47 (m, 3H), 1.26 (s, 3H), 1.25 (s, 3H), 1.13 (s, 3H), 0.09 (s,
9H), 0.11 (s, 3H), 0.09 (s, 3H); 13C NMR (100 MHz, CDCl3) δ 154.2,
144.4, 137.3, 129.4, 128.4, 127.8, 126.5, 101.8, 90.3, 82.9, 82.2, 78.7,
77.2, 71.5, 65.7, 53.2, 52.8, 45.8, 45.0, 41.5, 41.4, 36.8, 30.9, 27.4,
25.9, 23.8, 23.1, 18.1, 17.4, 17.3, 15.4, -3.3, -4.6; IR (film) 2947,
Diene 91. A solution of 90 (234 mg, 0.314 mmol), K2CO3 (130
mg, 0.942 mmol), and 4 Å sieves in 16 mL of CH3CN at 90 °C was
treated with Pd(PPh3)4 (400 mg, 0.345 mmol) in portions over a period
of 16.5 h so as to maintain an amber-colored solution. The resulting
black slurry was allowed to cool to room temperature, diluted with
ether (40 mL), and filtered through Celite using ether as eluent. The
orange filtrate was concentrated under reduced pressure, and the residue
was purified by flash chromatography (SiO2, 17:3 to 4:1 petroleum
ether:ether), affording 85.4 mg (46%) of 91 as a white foam: 1H NMR
(400 MHz, CDCl3) δ 7.38-7.23 (m, 5H), 5.19 (s, 1H), 5.04-4.96 (m,
2H), 4.84 (s, 2H), 4.73-4.61 (m, 3H), 3.57 (dd, J ) 7.2, 9.7 Hz, 1H),
2.81 (d, J ) 16.7 Hz, 1H), 2.62-2.33 (m, 3H), 2.22 (d, J ) 4.7 Hz,
1H), 2.14 (d, J ) 16.4 Hz, 1H), 1.97-1.90 (m, 1H), 1.82-1.74 (m,
1H), 1.69 (s, 3H), 1.69-1.57 (m, 1H), 1.32 (s, 3H), 1.30 (s, 3H), 1.17
(s, 3H), 0.94 (s, 9H), 0.07 (s, 3H), 0.07 (s, 3H); 13C NMR (100 MHz,
CDCl3) δ 154.1, 143.2, 139.4, 138.1, 134.8, 128.4, 127.6, 127.5, 117.8,
93.7, 82.4, 80.3, 79.7, 74.5, 72.0, 64.8, 44.9, 44.7, 44.5, 40.0, 38.0,
29.4, 26.0, 25.5, 23.2, 22.6, 22.4, 20.6, 18.1, 16.4, -3.0, -4.5; IR (film)
2954, 2738, 1803, 1472, 1455, 1207, 1008 cm-1; [R]20 6.6° (c ) 1,
D
CHCl3); FAB HRMS m/e calcd for (M + H) C35H51O6Si1 595.3463,
found 595.3455.
Alcohol 95. A solution of 91 (255.6 mg, 0.430 mmol) in 8.5 mL
of THF at room temperature was treated with tetrabutylammonium
fluoride (1.0 M solution in THF, 1.29 mL, 1.29 mmol). After 6 h, the
mixture was diluted with EtOAc (3 mL) and concentrated under reduced
pressure. The residue was dissolved in CH2Cl2 and 4:1 hexane:EtOAc
and purified by flash chromatography (SiO2, 7:3 to 3:2 hexane:EtOAc),
affording 189.1 mg (92%) of 95 as a clear oil: 1H NMR (400 MHz,
CDCl3) δ 7.36-7.22 (m, 5H), 5.23 (s, 1H), 5.09 (dd, J ) 7.1, 2.2 Hz,
1H), 4.98 (d, J ) 9.6 Hz, 1H), 4.83 (d, J ) 11.4 Hz, 1 H), 4.80 (d, J
) 11.4 Hz, 1 H), 4.71 (d, J ) 4.9 Hz, 1H), 4.68 (s, 1H), 4.58 (d, J )
9.6 Hz, 1H), 3.58-3.52 (m, 1H), 2.66 (d, J ) 16.3 Hz, 1H), 2.58-
2.43 (m, 2H), 2.37-2.32 (m, 1H), 2.22-2.16 (m, 2H), 2.02-1.92 (m,
2H), 1.81-1.73 (m, 1H), 1.68 (s, 3H), 1.65-1.58 (m, 1H), 1.33 (s,
3H), 1.30 (s, 3H), 1.16 (s, 3H); 13C NMR (100 MHz, CDCl3) δ 154.1,
142.6, 138.8, 137.9, 135.6, 128.5, 127.7, 127.5, 118.4, 93.8, 82.9, 80.1,
79.6, 77.2, 74.4, 72.1, 64.9, 45.9, 45.0, 43.9, 40.2, 36.2, 29.5, 25.7,
23.2, 22.8, 21.0, 16.4; IR (film) 3472, 2926, 1799, 1455, 1021, 1009
1797, 1400, 1211, 1137, 1077, 991, 853 cm-1; [R]20 -19.5° (c )
D
0.59, CHCl3); FAB HRMS m/e calcd for (M + Na) C37H55Na1O11S1-
Si1 815.3084, found 815.3125.
Aldehyde 89. A solution of 88 (1.39 mmol) and PPTS (1.75 g,
6.97 mmol) in 60 mL of 9:1 acetone:H2O was heated at 75 °C for 9 h.
The mixture was cooled to room temperature, treated with saturated
NaHCO3 (10 mL), and concentrated with a stream of nitrogen. The
residue was diluted with ether (200 mL) and poured into saturated
NaHCO3 (50 mL). The aqueous phase was washed with ether (2 ×
50 mL), and the combined extracts were washed with saturated NaHCO3
(2 × 50 mL) and saturated NaCl (75 mL), dried (K2CO3), and
concentrated under reduced pressure. The residue was purified by flash
chromatography (SiO2, 9:1 to 3:1 petroleum ether:ether), affording 998
mg (96%) of 89 as a white foam: 1H NMR (400 MHz, CDCl3) δ 9.5
(s, 1H), 7.50-7.29 (m, 5H), 5.15 (d, J ) 10.2 Hz, 1H), 4.95-4.89 (m,
3H), 4.82 (d, J ) 9.8 Hz, 1H), 4.62 (s, 1H), 3.96 (dd, J ) 12.6, 4.9
Hz, 1H), 2.87 (d, J ) 18.0 Hz, 1H), 2.69 (s, 1H), 2.52-2.44 (m, 1H),
2.35 (dd, J ) 13.6, 6.0 Hz, 1H), 2.24-2.14 (m, 2H), 1.80 (dd, J )
18.0, 6.4 Hz, 1H), 1.64 (s, 3H), 1.60-1.41 (m, 2H), 1.32 (s, 3H), 1.13
(s, 6H), 0.88 (s, 9H), 0.08 (s, 3H), 0.02 (s, 3H); 13C NMR (100 MHz,
CDCl3) δ 200.9, 153.8, 143.9, 136.9, 129.4, 128.4, 127.9, 127.0, 90.2,
82.6, 81.8, 77.9, 77.3, 77.2, 70.2, 65.7, 50.2, 44.9, 44.7, 43.1, 37.0,
27.2, 25.8, 23.3, 23.0, 18.0, 17.6, 17.3, 15.3, -3.7, -4.8; IR (film)
cm-1; [R]20 51.4° (c ) 1, CHCl3); FAB HRMS m/e calcd for (M +
D
K) C29H36K1O6 519.2149, found 519.2166.
Diene 96. A solution of 95 (189.1 mg, 0.394 mmol) and triethyl-
amine (0.270 mL, 1.97 mmol) in 16.0 mL of CH2Cl2 at -78 °C was
treated with triethylsilyl trifluoromethanesulfonate (0.098 mL, 0.433
mmol). After 0.5 h, the mixture was treated with saturated NaHCO3
(4 mL) and diluted with ether (30 mL), and the mixture was poured
into ether (20 mL) and saturated NaHCO3 (20 mL). The aqueous layer
was washed with ether (20 mL), and the combined extracts were washed
with saturated NaHCO3 (40 mL) and saturated NaCl (40 mL), dried
(Na2SO4), and concentrated under reduced pressure. The residue was
purified by flash chromatography (SiO2, 9:1 to 4:1 hexane:EtOAc),
affording 216.2 mg (92%) of 96 as a white foam: 1H NMR (400 MHz,
CDCl3) δ 7.36-7.24 (m, 5H), 5.16 (s, 1H), 4.99 (dd, J ) 9.2, 2.7 Hz,
1H), 4.93 (d, J ) 9.6 Hz, 1H), 4.82 (s, 2H), 4.69 (d, J ) 4.9 Hz, 1H),
4.64-4.60 (m, 2H), 3.56 (dd, J ) 10.2, 7.0 Hz, 1H), 2.77 (d, J ) 16.6
Hz, 1H), 2.52-2.34 (m, 3H), 2.18 (d, J ) 4.9 Hz, 1H), 2.12 (d, J )
16.5 Hz, 1H), 1.93-1.90 (m, 1H), 1.84-1.72 (m, 1H), 1.67 (s, 3H),
1.67-1.59 (m, 1H), 1.28 (s, 6H), 1.15 (s, 3H), 0.98-0.93 (t, J ) 7.9
Hz, 9H), 0.59 (q, J ) 8.0 Hz, 6H); 13C NMR (100 MHz, CDCl3) δ
2938, 1796, 1722, 1403, 1211, 1137, 1083, 995, 881, 852 cm-1; [R]20
D
-19.4° (c ) 0.64, CHCl3); FAB HRMS m/e calcd for (M + H)
C35F3H50O10S1Si1 747.2847, found 747.2857.
Cyclization Precursor 90. A solution of 89 (998 mg, 1.34 mmol)
in 45 mL of THF at -78 °C was treated dropwise with a solution of
Ph3P‚CH2 (0.10 M solution in THF, 13.5 mL, 1.35 mmol). After
complete addition, the mixture was stirred at -78 °C for 10 min and
at 0 °C for 15 min. The mixture was then treated with saturated NH4-
Cl (5 mL), diluted with ether (150 mL), and allowed to warm to room
temperature. The mixture was poured into ether (150 mL) and saturated
NH4Cl (75 mL). The aqueous phase was washed with ether (100 mL),
and the combined extracts were washed with saturated NH4Cl (75 mL)
and saturated NaCl (75 mL), dried (K2CO3), and concentrated under

2858 J. Am. Chem. Soc., Vol. 118, No. 12, 1996
Danishefsky et al.
154.1, 143.2, 139.4, 138.1, 134.9, 128.4, 127.6, 127.5, 127.4, 117.7,
93.7, 82.4, 80.3, 79.8, 74.6, 72.1, 64.8, 44.8, 44.4, 40.0, 37.8, 29.4,
25.5, 23.2, 22.4, 20.7, 16.2, 7.0, 5.5; IR (film) 2954, 2876, 1805, 1455,
1008 cm-1; FAB HRMS m/e calcd for (M + H) C35H51O6Si1 595.3455,
found 595.3473.
extracts were washed with saturated NH4Cl (10 mL) and saturated
NaCl (10 mL), dried (Na2SO4), and concentrated under reduced
pressure. The residue was purified by flash chromatography (SiO2,
4:1 to 3:1 to 7:3 petroleum ether:ether), affording 42.4 mg (93%) of
100a as a white foam: 1H NMR (400 MHz, CDCl3) δ 8.08 (app d, J
) 7.9 Hz, 2H), 7.59-7.56 (m, 1H), 7.50-7.43 (m, 2H), 5.88 (d, J )
5.4 Hz, 1H), 5.15 (s, 1H), 5.05-5.02 (m, 2H), 4.34 (d, J ) 8.2 Hz,
1H), 4.23 (d, J ) 8.3 Hz, 1H), 3.91 (t, J ) 8.4 Hz, 1H), 2.99 (d, J )
16.4 Hz, 1H), 2.59-2.51 (m, 3H), 2.27 (s, 3H), 2.25-2.08 (m, 2H),
1.84 (dd, J ) 9.2, 14.9 Hz, 1H), 1.73-1.69 (m, 1H), 1.58 (s, 1H),
1.53 (s, 3H), 1.46 (s, 3H), 1.38 (s, 3H), 1.31 (s, 3H), 0.99 (s, 3H), 0.95
(t, J ) 8.0 Hz, 9H), 0.60-0.50 (m, 6H); 13C NMR (100 MHz,
CDCl3) δ 169.6, 140.7, 133.5, 130.0, 129.4, 128.6, 117.8, 108.8, 84.1,
82.3, 78.4, 77.2, 73.2, 71.5, 69.8, 62.7, 44.2, 44.1, 43.4, 42.3, 37.6,
26.0, 25.7, 25.6, 22.0, 21.4, 21.3, 16.1, 7.0, 5.7; IR (film) 3521, 2954,
Epoxide 97. A solution of 96 (171.5 mg, 0.289 mmol) and NaHCO3
(343 mg) in 9.7 mL of CH2Cl2 at room temperature was treated with
mCPBA (95%, 89.7 mg, 0.520 mmol). After 10.25 h, the mixture was
treated with Me2S (30 µL) and H2O (10 mL). The mixture was then
diluted with ether (50 mL) and poured into saturated NaHCO3 (30 mL).
The aqueous layer was washed with ether (20 mL), and the combined
extracts were washed with saturated NaHCO3 (2 × 40 mL) and
saturated NaCl (40 mL), dried (Na2SO4), and concentrated under
reduced pressure. The residue was purified by flash chromatography
(SiO2, 17:3 to 4:1 to 3:1 petroleum ether:ether), affording 79.6 mg
(45%) of 97 as a white foam: 1H NMR (400 MHz, CDCl3) δ 7.36-
7.24 (m, 5H), 5.13 (m, 2H), 5.01 (m, 2H), 4.90-4.81 (m, 3H), 4.68
(d, J ) 9.7 Hz, 1H), 3.67 (dd, J ) 10.0, 7.6 Hz, 1H), 3.04 (d, J ) 16.9
Hz, 1H), 2.59-2.50 (m, 1H), 2.44-2.40 (m, 1H), 2.28 (d, J ) 16.8
Hz, 1H), 2.15 (d, J ) 4.4 Hz, 1H), 1.98-1.92 (m, 2H), 1.60-1.54 (m,
1H), 1.37 (s, 3H), 1.36 (s, 3H), 1.16 (s, 3H), 0.92 (s, 3H), 0.98-0.92
(m, 9H), 0.56 (m, 6H); 13C NMR (100 MHz, CDCl3) δ 153.9, 140.0,
137.5, 128.5, 127.9, 127.8, 119.5, 91.5, 82.4, 80.0, 79.3, 77.2, 74.1,
70.5, 69.4, 65.0, 62.6, 43.9, 43.5, 43.4, 40.2, 37.6, 25.1, 24.5, 23.0,
22.9, 20.5, 15.8, 7.0, 5.6; IR (film) 2955, 1807, 1456, 1076, 1010, 733
cm-1; FAB HRMS m/e calcd for (M + H) C35H51O7Si1 611.3404, found
611.3416.
2876, 1733, 1277, 1247, 1098 cm-1; [R]20 15.1° (c ) 1.0, CHCl3);
D
FAB HRMS m/e calcd for (M+) C36H52O8Si1 640.3431, found
640.3410.
Ketone 102. A solution of 100 (42.4 mg, 0.066 mmol) in 2.4 mL
of pyridine was treated with OsO4 (319.7 mg, 1.258 mmol) and heated
to 105 °C. After 24 h, the mixture was cooled, diluted with EtOAc
(10 mL), and treated with saturated NaHSO3 (10 mL). The aqueous
layer was washed with EtOAc (4 × 10 mL), and the combined extracts
were washed with saturated NaHSO3 (2 × 10 mL), H2O (10 mL), and
saturated NaCl (10 mL), dried (Na2SO4), filtered through SiO2 (1 mL),
and concentrated under reduced pressure. The residue was dissolved
in 1:1 C6H6:MeOH (8 mL), cooled to 0 °C, and treated with lead(IV)
acetate (47.4 mg, 0.265 mmol). After 5 min, the mixture was diluted
with ether (20 mL) and the resulting yellow slurry was filtered through
SiO2 (2 mL) using ether as eluent. The filtrate was concentrated under
reduced pressure, and the residue was purified by flash chromatography
(SiO2, 4:1 to 7:3 hexane:EtOAc), affording 26.1 mg (61%) of 102 as
a white solid: 1H NMR (400 MHz, CDCl3) δ 8.11-8.09 (app d, 2H),
7.62-7.59 (app t, 1H), 7.50-7.46 (app t, 2H), 6.07 (d, J ) 5.5 Hz,
1H), 5.00 (d, J ) 8.6 Hz, 1H), 4.38 (d, J ) 8.5 Hz, 1H), 4.26 (d, J )
8.4 Hz, 1H), 3.76 (t, J ) 8.4 Hz, 1H), 3.35 (d, J ) 15.5 Hz, 1H), 2.85
(d, J ) 15.4 Hz, 1H), 2.70 (d, J ) 5.4 Hz, 1H), 2.51-2.49 (m, 1H),
2.30 (s, 3H), 2.30-2.18 (m, 2H), 1.84-1.70 (m, 2H), 1.50 (s, 3H),
1.45 (s, 3H), 1.32 (s, 3H), 1.04 (s, 3H), 0.95 (t, J ) 8.0 Hz, 9H), 0.63-
0.57 (m, 6H); 13C NMR (100 MHz, CDCl3) δ 202.7, 169.8, 166.9,
133.8, 130.1, 129.1, 128.7, 84.0, 82.0, 77.9, 77.2, 76.5, 73.1, 72.1, 65.8,
62.3, 50.9, 44.8, 44.4, 40.1, 37.7, 26.3, 25.8, 25.6, 23.0, 22.0, 21.2,
16.0, 15.3, 14.2, 6.9, 5.8, 5.3; IR (film) 3515, 2955, 2876, 1731, 1695,
1272, 1247, 1098, 712 cm-1; FAB HRMS m/e calcd for (M + H)
C35H51O9Si1 643.3303, found 643.3301.
Alcohol 98. A solution of 97 (79.6 mg, 0.130 mmol) and Pd(OAc)2
(24.9 mg, 0.111 mmol) in 4.7 mL of EtOH was placed under an
atmosphere of hydrogen. After 10.0 h, the mixture was diluted with
EtOAc (25 mL), filtered through Celite, and concentrated under reduced
pressure. The residue was purified by flash chromatography (SiO2,
7:3 hexane:EtOAc), affording 55.6 mg (82%) of 98 as a clear oil: 1H
NMR (400 MHz, CDCl3) δ 5.15 (s, 1H), 5.05 (s, 1H), 4.86 (d, J ) 4.7
Hz, 1H), 4.78-4.69 (m, 2H), 4.34 (d, J ) 9.0 Hz, 1H), 3.69 (dd, J )
9.7, 7.7 Hz, 1H), 3.02 (d, J ) 16.9 Hz, 1H), 2.84-2.77 (m, 1H), 2.70
(s, 1H), 2.44-2.38 (m, 1H), 2.38 (d, J ) 16.7 Hz, 1 H), 2.29-2.14
(m, 2H), 2.10 (d, J ) 4.7 Hz, 1H), 2.02-1.87 (m, 2H), 1.69-1.58 (m,
1H), 1.40 (s, 3H), 1.30 (s, 3H), 1.22 (s, 3H), 1.20 (s, 3H), 0.93 (t, J )
7.9 Hz, 9H), 0.54 (q, J ) 8.0 Hz, 6H); 13C NMR (100 MHz, CDCl3)
δ 153.9, 140.0, 119.5, 91.6, 88.3, 81.0, 80.1, 77.2, 74.3, 70.5, 69.4,
62.6, 43.8, 43.5, 43.0, 40.2, 37.6, 25.4, 24.6, 23.4, 23.0, 20.5, 15.7,
7.0, 5.6; IR (film) 3507, 2955, 2876, 1792, 1458, 1235, 1011 cm-1
;
[R]20 -24.5° (c ) 1.0, CHCl3); FAB HRMS m/e calcd for (M + K)
D
C28H44K1O7Si1 559.2493, found 559.2502.
Enone 103. A solution of 102 (26.1 mg, 0.041 mmol) and acetic
anhydride (38 µL, 0.41 mmol) in 2.0 mL of THF at -78 °C was treated
with SmI2 (0.1 M solution in THF, 165 µL, 0.0165 mmol) until a green
solution persisted. After 10 min, the mixture was treated with saturated
NH4Cl (5 mL), allowed to warm, and diluted with EtOAc (10 mL)
and H2O (2 mL). The mixture was poured into EtOAc (10 mL) and
saturated NH4Cl (5 mL), and the layers were separated. The aqueous
layer was washed with EtOAc (10 mL), and the combined extracts
were washed with saturated NaHCO3 (10 mL) and saturated NaCl (10
mL), dried (Na2SO4), and concentrated under reduced pressure. The
residue was purified by flash chromatography (SiO2 [2 mL], 7:3 hexane:
EtOAc), affording 23.4 mg (92%) of 103 as a white solid: 1H NMR
(400 MHz, CDCl3) δ 8.11-8.09 (app d, 2H), 7.61-7.57 (app t, 1H),
7.50-7.45 (app t, 2H), 5.87 (d, J ) 5.8 Hz, 1H), 4.89 (d, J ) 8.3 Hz,
1H), 4.34 (d, J ) 8.3 Hz, 1H), 4.16 (d, J ) 8.3 Hz, 1H), 3.72 (t, J )
9.1 Hz, 1H), 3.14 (d, J ) 15.6 Hz, 1H), 3.08 (d, J ) 5.8 Hz, 1H),
2.65-2.58 (m, 2H), 2.57 (d, J ) 15.6 Hz, 1H), 2.28 (s, 3H), 2.28-
2.24 (m, 1H), 1.97-1.86 (m, 1H), 1.82-1.74 (m, 1H), 1.75 (s, 3H),
1.41 (s, 3H), 1.35 (s, 3H), 1.13 (s, 3H), 0.97 (q, J ) 8.0 Hz, 9H), 0.62
(t, J ) 8.0 Hz, 6H); 13C NMR (100 MHz, CDCl3) δ 202.2, 170.0,
167.0, 144.6, 138.2, 133.6, 130.1, 129.3, 128.6, 84.5, 82.4, 80.2, 77.2,
76.5, 73.4, 72.7, 50.4, 44.6, 44.2, 39.6, 38.0, 30.2, 26.7, 26.0, 22.8,
22.3, 20.0, 16.1, 7.0, 5.3; IR (film) 3502, 2956, 1731, 1677, 1456, 1273,
1244, 1098, 730, 711 cm-1; FAB HRMS m/e calcd for (M + H)
C35H51O8Si1 627.3353, found 627.3376.
Acetate 99. A solution of 98 (55.6 mg, 0.107 mmol) and DMAP
(14.4 mg, 0.118 mmol) in 0.43 mL of pyridine at room temperature
was treated with acetic anhydride (100 µL, 1.07 mmol). After 16 h,
the mixture was treated with saturated NaHCO3 (5 mL) and then
diluted with EtOAc (15 mL). The aqueous layer was washed with
EtOAc (10 mL), and the combined extracts were washed with saturated
NaHCO3 (2 × 10 mL), dried (Na2SO4), and concentrated under reduced
pressure. The residue was purified by flash chromatography (SiO2,
4:1 hexane:EtOAc), affording 40.1 mg (66%) of 99 as a white foam
and 4.9 mg (9%) of recovered 98: 1H NMR (400 MHz, CDCl3) δ
5.15 (s, 1H), 5.01 (s, 1H), 4.98 (d, J ) 8.8 Hz, 1H), 4.88 (d, J ) 5.2
Hz, 1H), 4.63 (d, J ) 8.8 Hz, 1H), 4.55 (d, J ) 8.8 Hz, 1H), 3.90 (t,
J ) 8.0 Hz, 1H), 3.10 (d, J ) 17.1 Hz, 1H), 2.60-2.52 (m, 1H), 2.46
(d, J ) 5.2 Hz, 1H), 2.30 (d, J ) 18.5 Hz, 1H), 2.30-2.08 (m, 2H),
2.14 (s, 3H), 1.87-1.60 (m, 4H), 1.42 (s, 3H), 1.39 (s, 3H), 1.25 (s,
3H), 1.04 (s, 3H), 0.94 (t, J ) 8.0 Hz, 9H), 0.57 (m, 6H); 13C NMR
(100 MHz, CDCl3) δ 170.1, 153.3, 139.8, 119.2, 108.7, 90.9, 84.1,
80.7, 79.6, 77.2, 76.3, 70.9, 69.0, 62.2, 44.9, 43.7, 41.0, 40.2, 38.0,
29.7, 25.0, 24.7, 23.8, 21.8, 21.3, 20.4, 16.3, 7.0, 5.6; IR (film) 2950,
1800, 1730, 1243, 1013, 969 cm-1; [R]20D -17.3° (c ) 0.95, CHCl3);
FAB HRMS m/e calcd for (M + K) C30H46K1O8Si1 601.2599, found
601.2623.
Benzoate 100a. A solution of 99 (40.1 mg, 0.071 mmol) in 6 mL
of THF at -78 °C was treated with PhLi (0.5 M solution in 4:1 THF:
Et2O, 0.57 mL, 0.285 mmol). After 2 min, the mixture was treated
with saturated NH4Cl (10 mL) and then diluted with ether (15 mL).
The aqueous layer was washed with ether (20 mL), and the combined
r-Hydroxy Ketone 104. A solution of 103 (23.4 mg, 0.037 mmol)
in 2.5 mL of THF at -78 °C was treated with tBuOK (0.24 M solution

Total Synthesis of Baccatin III and Taxol
J. Am. Chem. Soc., Vol. 118, No. 12, 1996 2859
in THF, 0.62 mL, 0.149 mmol), and the mixture was placed in a -20
°C bath. After 0.75 h, the mixture was briefly warmed to 0 °C and
then transferred via cannula to a 0 °C solution of (PhSeO)2O (107.5
mg, 0.298 mmol) in 2.5 mL of THF. After 0.75 h, the mixture was
diluted with EtOAc (20 mL) and poured into saturated NaHCO3 (20
mL). The organic layer was washed with saturated Na2S2O3 (15 mL)
and saturated NaHCO3 (15 mL), dried (Na2SO4), and concentrated under
reduced pressure. The residue was dissolved in 2.5 mL of THF, cooled
to -78 °C, and treated with tBuOK (0.24 M solution in THF, 0.92
mL, 0.149 mmol). After 0.5 h, the mixture was treated AcOH (0.8 M
solution in THF, 0.47 mL, 0.373 mmol) and allowed to stir for 10 min
at -78 °C and then allowed to warm for 10 min. The mixture was
diluted with EtOAc (40 mL) and poured into saturated NaHCO3 (20
mL). The organic layer was dried (Na2SO4) and concentrated under
reduced pressure, and the residue was purified by flash chromatography
(SiO2 [2 mL], 9:1 to 4:1 to 7:3 hexane:EtOAc), affording 19.3 mg (81%)
of 104 as a white solid: 1H NMR (400 MHz, CDCl3) δ 8.05 (app d,
2H), 7.58 (app t, 1H), 7.43 (app t, 2H), 5.56 (d, J ) 6.64 Hz, 1H),
5.18 (d, J ) 2.5 Hz, 1H), 4.95 (d, J ) 8.2 Hz, 1H), 4.37 (dd, J ) 6.8,
8.0 Hz, 1H), 4.30 (d, J ) 8.4 Hz, 1H), 4.20 (d, J ) 2.3 Hz, 1H), 4.15
(d, J ) 8.5 Hz, 1H), 3.83 (d, J ) 6.6 Hz, 1H), 2.70-2.65 (m, 1H),
2.49-2.45 (m, 1H), 2.30 (s, 3H), 2.28-2.20 (m, 1H), 1.96 (s, 3H),
1.92-1.77 (m, 2H), 1.70 (s, 3H), 1.53 (s, 1H), 1.50 (s, 1H), 1.30-
1.23 (m, 1H), 1.09 (s, 3H), 1.08 (s, 3H), 0.91(t, J ) 7.9 Hz), 0.60-
0.48 (m, 6H); IR (film) 3455, 2956, 1727, 1452, 1370, 1271, 1248,
1107, 984, 821, 718 cm-1; FAB HRMS m/e calcd for (M + H)
C35H51O9Si1 643.3303, found 643.3316.
mixture was flushed through a plug of silica gel. Flash chromatography
(SiO2 [2 mL], 9:1 PhH:ether) afforded 10.3 mg (64%) of 106 as a white
solid.
7-O-TES-baccatin III (107).46 A solution of 106 (10.3 g, 0.0147
mmol) in 2.0 mL of MeOH was treated with NaBH4 (10.3 mg, 0.272
mmol) every 30 min for 2 h. The reaction was quenched with NH4Cl
(4 mL) and stirred vigorously for 15 min. The mixture was diluted
with EtOAc (20 mL), washed with saturated NaCl (10 mL), dried over
Na2SO4, filtered, and concentrated under reduced pressure. The residue
was purified by flash chromatography (SiO2 [2 mL], 7:3 hex:EtOAc)
to provide 8.1 mg (79%) of 107 along with 1.1 mg (11%) of recovered
106.
Baccatin III (1). A solution of 107 (8.1 mg, 0.012 mmol) in 0.8
mL of THF was treated with 0.1 mL of HF‚pyr in a Teflon flask. After
2 h, the reaction was diluted with ether (20 mL) and extracted with
saturated NaHCO3 (2 × 10 mL), saturated CuSO4 (10 mL), and
saturated NaCl (10 mL). The organic layer was dried over Na2SO4,
filtered, and concentrated under reduced pressure. The residue was
purified by flash chromatography (SiO2 [2 mL], 7:3 hex:EtOAc) to
afford 5.7 mg (85%) of 1 as a white solid. This material was identical
1
to an authentic sample of baccatin III by TLC, H NMR, 13C NMR,
IR, and optical rotation.
Acknowledgments. We dedicate this work to Dr. Monroe
Wall of the Research Triangle Institute for his discovery of taxol
and many other fascinating Natural Products. This research was
supported by the National Institutes of Health (NIH, Grant No.
AI16943). A Damon Runyon-Walter Winchell Cancer Fund
Postdoctoral Fellowship to J.J.M., an American Cancer Society
Postdoctoral Fellowship to W.B.Y., and a National Science
Foundation Postdoctoral Fellowship to C.A.C. are gratefully
acknowledged. We thank Dr. G. Sukenick (Memorial Sloan-
Kettering Cancer Center) and Barbara Sporer and Vinka
Parmakovich (Columbia University) for mass spectral analyses.
We also thank Susan de Gala of Yale University for crystal-
lographic analysis.
13-Deoxy-7-O-TES-baccatin III (105).46 A solution of 104 (19.3
mg, 0.03 mmol) and DMAP (1.8 mg, 0.015 mmol) in 0.2 mL of
pyridine at room temperature was treated with acetic anhydride (28.3
mL, 0.3 mmol). After 16 h, the mixture was diluted with EtOAc (20
mL) and poured into saturated NaHCO3 (15 mL). The organic layer
was washed with saturated NaCl (15 mL), dried (Na2SO4), and
concentrated under reduced pressure. The residue was purified by flash
chromatography (SiO2 [2 mL], 4:1 hexane:EtOAc), affording 15.9 mg
(76%) of 105 as a white solid.
13-Oxo-7-O-TES-baccatin III (106).46 A solution of 105 (15.9 mg,
0.0232 mmol) in 3.5 mL of PhH was treated with NaOAc (57.1 mg,
0.696 mmol), celite (150 mg), and PCC (150 mg, 0.696 mmol). The
resulting mixture was heated to reflux for 1 h. After cooling to room
temperature and diluting with 4 mL of 9:1 PhH:ether, the crude reaction
JA952692A
(46) Compound reported by Nicolaou; see: Nicolaou, K. C.; et al. J.
Am. Chem. Soc. 1995, 117, 653-659.