Enantioselective total synthesis of (-)-taxol

Article abstract of DOI:10.1021/ja9939439

Enantioselective total synthesis of taxol has been accomplished. Coupling reaction of the optically pure A-ring hydroxy aldehyde with the aromatic C-ring fragment followed by Lewis acid mediated eight-membered B- ring cyclization gave the desired ABC endo-tricarbocycle. The C-ring moiety of this product was reduced under Birch conditions to the cyclohexadiene derivative, which was oxygenated by singlet oxygen from the convex β-face to give the C4β,C7β-diol stereoselectively. For introduction of the C19- methyl, the cyclopropyl ketone was prepared via cyclopropanation of the C- ring allylic alcohol or conjugate addition of a cyano group to the C-ring enone. Reductive cleavage of the cyclopropane ring followed by isomerization of the resulting enol to the corresponding ketone gave the crucial synthetic intermediate containing the C19-methyl group. Regioselective transformation of three hydroxyl groups of this intermediate, conversion of the C4-carbonyl group to the allyl chloride, and introduction of the C10-oxygen functionality afforded a precursor for D-ring construction. Dihydroxylation of the allyl chloride moiety followed by basic treatment of the resulting diol gave a fully functionalized taxol skeleton. Functional group manipulation of this product including attachment of the C13 side chain provided (-)-taxol.

Full text of DOI:10.1021/ja9939439

J. Am. Chem. Soc. 2000, 122, 3811-3820
3811
Enantioselective Total Synthesis of (-)-Taxol
Hiroyuki Kusama,† Ryoma Hara,† Shigeru Kawahara,† Toshiyuki Nishimori,†
Hajime Kashima,† Nobuhito Nakamura,† Koichiro Morihira,† and Isao Kuwajima*,‡
Contribution from the Department of Chemistry, Tokyo Institute of Technology, O-okayama, Meguro-ku,
Tokyo 152-8551, Japan, and the Research Center for Biological Function, The Kitasato Institute,
Shirokane, Minato-ku, Tokyo 108-8642, Japan
ReceiVed NoVember 8, 1999
Abstract: Enantioselective total synthesis of taxol has been accomplished. Coupling reaction of the optically
pure A-ring hydroxy aldehyde with the aromatic C-ring fragment followed by Lewis acid mediated eight-
membered B-ring cyclization gave the desired ABC endo-tricarbocycle. The C-ring moiety of this product
was reduced under Birch conditions to the cyclohexadiene derivative, which was oxygenated by singlet oxygen
from the convex â-face to give the C4â,C7â-diol stereoselectively. For introduction of the C19-methyl, the
cyclopropyl ketone was prepared via cyclopropanation of the C-ring allylic alcohol or conjugate addition of
a cyano group to the C-ring enone. Reductive cleavage of the cyclopropane ring followed by isomerization of
the resulting enol to the corresponding ketone gave the crucial synthetic intermediate containing the C19-
methyl group. Regioselective transformation of three hydroxyl groups of this intermediate, conversion of the
C4-carbonyl group to the allyl chloride, and introduction of the C10-oxygen functionality afforded a precursor
for D-ring construction. Dihydroxylation of the allyl chloride moiety followed by basic treatment of the resulting
diol gave a fully functionalized taxol skeleton. Functional group manipulation of this product including
attachment of the C13 side chain provided (-)-taxol.
Introduction
The taxane diterpenes1 (Figure 1) isolated from several yew
trees have a common tricarbocyclic structure containing an sp2
carbon on their bridgehead C11-sites.2 Such a unique structural
feature as well as many oxygenated asymmetric centers has
made them extremely challenging synthetic targets. Among
various taxanes, taxol3 (1) isolated from extracts of the bark of
Taxus breVifolia was shown to act as a promoter of microtubule
assembly4 and has been well-documented to exhibit promising
activity against a number of human cancers.5 Further, some
members of its family such as taxinine and taxuspines were
also reported to exhibit multidrug resistance reversing activity.6
† Tokyo Institute of Technology.
‡ The Kitasato Institute.
(1) (a) Lythgoe, G. The Alkaloids; R. H. F., Ed.; Academic Press: New
York, 1968; Vol. 10, p 597. (b) Miller, R. W. J. Nat. Prod. 1980, 43, 425.
(c) Kingston, D. G. I.; Molinero, A. A.; Rimoldo, J. M. Prog. Chem. Org.
Nat. Prod. 1993, 61, 1.
(2) The taxane numbering system, as illustrated in Figure 1, is used
throughout.
(3) Wani, M. C.; Taylor, H. L.; Wall, M. E.; Coggon, P.; McPhail, A.
T. J. Am. Chem. Soc. 1971, 93, 2325.
Figure 1.
(4) Horwitz, S. B.; Fant, J.; Schiff, P. B. Nature 1979, 277, 665. Schiff,
P. B.; Horwitz, S. Proc. Natl. Acad. Sci. U.S.A. 1980, 77, 1561. Schiff, P.
B.; Manfredi, J. J.; Horwitz, S. B. Pharmacol. Ther. 1984, 25, 83.
(5) For reviews, see: (a) Taxol: Science and Applications; Suffness,
M., Ed.; CRC: Boca Raton, FL, 1995. (b) Taxane Anticancer Agents: Basic
Science and Current Status; Georg, G. I., Chen, T. T., Ojima, I., Vyas, D.
M., Eds.; ACS Symposium Series 583; American Chemical Society:
Washington, DC, 1995. (c) Runowicz, C. D.; Wiernik, P. H.; Einzig, A. I.;
Goldberg, G. L.; Horwitz, S. B. Cancer 1993, 71, 1591. (d) Rowinsky, E.
K.; Onetto, N.; Canetta, R. M.; Arbuck, S. G. Semin. Oncol. 1992, 19,
646.
(6) (a) Kobayashi, J.; Ogiwara, A.; Hosoyama, H.; Shigemori, H.;
Yoshida, N.; Sasaki, T.; Li, Y.; Iwasaki, S.; Naito, M.; Tsuruo, T.
Tetrahedron 1994, 50, 7401. (b) Morita, H.; Wei, L.; Gouda, A.; Takeya,
K.; Itokawa, H.; Fukaya, H.; Shigemori, H.; Kobayashi, J. Tetrahedron 1997,
53, 4621.
These features have prompted synthetic studies to lead to
successful total syntheses.7,8,10
The major problems in taxane synthesis seem to be focused
on the following two aspects: (1) construction of a highly
distorted ABC-tricarbocyclic structure, and (2) stereocontrol of
many asymmetric centers. To solve the first problem, we
developed a useful methodology which allowed construction
of the taxane ABC-tricarbocycle from AC-fragment A via
connection between the C9 and C10 atoms.9 Although this aldol-
like C-C bond formation is assumed to produce several
stereoisomers, the cyclized product was usually obtained as a
10.1021/ja9939439 CCC: $19.00 © 2000 American Chemical Society
Published on Web 04/05/2000

3812 J. Am. Chem. Soc., Vol. 122, No. 16, 2000
Kusama et al.
single endo conformer containing C9R,C10â substituents ir-
Scheme 1. Retrosynthetic Analysis
respective of the geometry of the dienol silyl ether moiety (eq
1).
We have previously described a synthesis of the relatively
simple taxusin, where the present methodology was efficiently
applicable for construction of the taxusin ABC-tricarbocycle.
Subsequent functional group manipulations completed the total
synthesis of (+)-taxusin.8b
Considering a much more complex and challenging structure
as well as important biological activities, total syntheses of
(7) Taxol synthesis: (a) Holton, R. A.; Somoza, C.; Kim, H.-B.; Liang,
F.; Biediger, R. J.; Boatman, P. 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. N.; Liu, J. H. J. Am. Chem. Soc. 1994, 116, 1597. Holton,
R. A.; Kim, H.-B.; Somoza, C.; Liang, F.; Biediger, R. J.; Boatman, P. 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. N.; Liu, J. H. J. Am.
Chem. Soc. 1994, 116, 1599. (b) Nicolaou, K. C.; Nantermet, P. G.; Ueno,
H.; Guy, R. K.; Couladouros, E. A.; Sorensen, E. J. J. Am. Chem. Soc.
1995, 117, 624. Nicolaou, K. C.; Liu, J.-J.; Yang, Z.; 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. 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. 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. (c) Danishefsky, S. J.; Masters, J. J.;
Young, W. B.; Link, J. T.; Snyder, L. B.; Magee, T. V.; Jung, D. K.; Isaacs,
R. C. A.; Bornmann, W. G.; Alaimo, C. A.; Coburn, C. A.; Di Grandi, M.
J. J. Am. Chem. Soc. 1996, 118, 2843. (d) Wender, P. A.; Badham, N. F.;
Conway, S. P.; Floreancig, P. E.; Glass, T. E.; Gra¨nicher, C.; Houze, J. B.;
Ja¨nichen, J.; Lee, D.; Marquess, D. G.; McGrane, P. L.; Meng, W.;
Mucciaro, T. P.; Mu¨hlebach, M.; Natchus, M. G.; Paulsen, H.; Rawlins, D.
B.; Satkofsky, J.; Shuker, A. J.; Sutton, J. C.; Taylor, R. E.; Tomooka, K.
J. Am. Chem. Soc. 1997, 119, 2755. Wender, P. A.; Badham, N. F.; Conway,
S. P.; Floreancig, P. E.; Glass, T. E.; Houze, J. B.; Krauss, N. E.; Lee, D.;
Marquess, D. G.; McGrane, P. L.; Meng, W.; Natchus, M. G.; Shuker, A.
J.; Sutton, J. C.; Taylor, R. E. J. Am. Chem. Soc. 1997, 119, 2757. (e)
Mukaiyama, T.; Shiina, I.; Iwadare, H.; Sakoh, H.; Tani, Y.; Hasegawa,
M.; Saitoh, K. Proc. Jpn. Acad. Ser. B 1997, 73, 95. Shiina, I.; Iwadare,
H.; Sakoh, H.; Hasegawa, M.; Tani, Y.; Mukaiyama, T. Chem. Lett. 1998,
1. Shiina, I.; Saitoh, K.; Fre´chard-Ortuno, I.; Mukaiyama, T. Chem. Lett.
1998, 3. Mukaiyama, T.; Shiina, I.; Iwadare, H.; Saitoh, M.; Nishimura,
T.; Ohkawa, N.; Sakoh, H.; Nishimura, K.; Tani, Y.; Hasegawa, M.;
Yamada, K.; Saitoh, K. Chem. Eur. J. 1999, 5, 121.
(8) Taxusin synthesis: (a) Holton, R. A.; Juo, R. R.; Kim, H.-B.;
Williams, A. D.; Harusawa, S.; Lowenthal, R. E.; Yogai, S. J. Am. Chem.
Soc. 1988, 110, 6558. (b) Hara, R.; Furukawa, T.; Horiguchi, Y.; Kuwajima,
I. J. Am. Chem. Soc. 1996, 118, 9186. Hara, R.; Furukawa, T.; Kashima,
H.; Kusama, H.; Horiguchi, Y.; Kuwajima, I. J. Am. Chem. Soc. 1999, 121,
3072. (c) Paquette, L. A.; Zhao, M. J. Am. Chem. Soc. 1998, 120, 5203.
Paquette, L. A.; Wang, H.-L.; Su, Z.; Zhao, M. J. Am. Chem. Soc. 1998,
120, 5213.
(9) (a) Horiguchi, Y.; Furukawa, T.; Kuwajima, I. J. Am. Chem. Soc.
1989, 111, 8277. (b) Furukawa, T.; Morihira, K.; Horiguchi, Y.; Kuwajima,
I. Tetrahedron 1992, 48, 6975. (c) Seto, M.; Morihira, K.; Katagiri, S.;
Furukawa, T.; Horiguchi, Y.; Kuwajima, I. Chem. Lett. 1993, 133. (d)
Morihira, K.; Seto, M.; Furukawa, T.; Horiguchi, Y.; Kuwajima, I.
Tetrahedron Lett. 1993, 34, 345.
natural taxol and its structural analogues are expected to bring
much greater advances both in synthetic organic chemistry and
in the clinical field. We explored an efficient pathway to the
more challenging (-)-taxol. This paper describes the full details
of our synthesis of (-)-taxol.10
Synthetic Plan
Our program for taxol synthesis was based on the initial
construction of endo tricarbocyclic intermediate I (Scheme 1)
with correct stereochemistry at C1 and C2, followed by
appropriate functional group manipulations on the B- and
C-rings. The structural features of I favor an introduction of
the C19-methyl from the convex face of the C-ring. Our choice
of a C-ring fragment was rather crucial at this stage, but among
several candidates, we preferred a cyclohexadiene derivative
since it was expected that the use of the diene C-ring fragment
III would permit an eventual installation of C4 and C7 oxygen
functionalities from the convex â-face of I. We also envisioned
that the present approach would lead to an enantioselective
synthesis of taxol by using the aldehyde II with a chiral center
corresponding to the C1 site. Thus, a chelation-controlled
coupling of the optically active R-hydroxy aldehyde11 with a
C-ring fragment would confirm the stereochemical outcome at
the C1 and C2 sites, and a subsequent B-ring cyclization would
lead to the diastereoselective formation of the key intermediate
I for taxol. On the basis of the retrosynthetic analysis shown in
Scheme 1, we executed an enantioselective total synthesis of
(-)-taxol.
Enantioselective Preparation of the A-Ring Fragment
The optically active A-ring fragment 8 was prepared as
follows (Schemes 2 and 3).11 Addition of lithiated propargyl
ether to propionaldehyde followed by Lindlar reduction and
(10) For the preliminary communication, see: Morihira, K.; Hara, R.;
Kawahara, S.; Nishimori, T.; Nakamura, N.; Kusama, H.; Kuwajima, I. J.
Am. Chem. Soc. 1998, 120, 12980.
(11) (a) Nakamura, T.; Waizumi, N.; Horiguchi, Y.; Kuwajima, I.
Tetrahedron Lett. 1994, 35, 7813. (b) Seto, M.; Morihira, K.; Horiguchi,
Y.; Kuwajima, I. J. Org. Chem. 1994, 59, 3165.

EnantioselectiVe Total Synthesis of (-)-Taxol
J. Am. Chem. Soc., Vol. 122, No. 16, 2000 3813
Scheme 2. Preparation of the A-Ring Fragment. 1a
basic media with good recovery was quite difficult due to a
facile formation of dimeric isomers; hence, the crude mixture
including a monomer and dimers was treated with N,N′-
dimethylethylenediamine in boiling benzene to convert 6 to its
aminal. Purification of the resultant aminal by silica gel column
chromatography was accompanied by removal of the protecting
group to give pure hydroxy aldehyde 6 as a monomer.
Recrystallization from benzene-hexane afforded the enantio-
merically pure R-hydroxy aldehyde (>98% ee). The optical
purity of 6 was determined by 1H NMR analysis of the MTPA
ester derived from the reduction of 6, followed by acylation of
the resulting alcohol. After protecting the aldehyde moiety as
an aminal, the pivaloyl group was changed to a TIPS group in
a one-pot operation, and elution through a silica gel column
again allowed removal of the aminal moiety to give aldehyde
7. According to a similar operation, Peterson olefination of 7
afforded dienol silyl ether 8 as a mixture of geometrical isomers
on the vinyl sulfide moiety (E:Z ) ca. 1.5:1).14 Although 8 could
be prepared from 6 without removal of the aminal moiety as in
the conversion of 6 to 7, the above procedure gave much higher
reproducibility.
a (a) THF, -78 to 0 °C, 3 h; (b) hexane, rt, 7 h; (c) CH2Cl2, -78 °C
to rt, 10 h; (d) THF, - 78 °C, 0.5 h; 0 °C, 0.5 h; (e) Et2O, 0 °C, 3 h;
CH2Cl2, rt, overnight; (f) MeOH, rt, 3 h, 52% for six steps; (g) CH2Cl2,
-78 to 0 °C, overnight, 92%.
Scheme 3. Preparation of the A-Ring Fragment. 2a
Construction of the ABC Tricarbocycle
(a) Use of Cyclohexadiene Derivative as a C-Ring Frag-
ment. As the first candidate for a C-ring fragment, 2-bromo-
1,3-cyclohexadiene derivative 12 was chosen and prepared as
shown in Scheme 4. Thus, introduction of a triethylsiloxy group
onto the 6-position of 2-bromocyclohexenone was performed
by oxidation of the corresponding enol silyl ether with m-CPBA.
Addition of PhSCH(Li)OMe15 to the carbonyl group of 9
occurred from the opposite side of the OTES group to give the
cis-diol derivative predominantly along with a small amount
Scheme 4. Preparation of the C-Ring Fragmenta
a (a) CH2Cl2, -78 °C to rt, 4 h, 71%; (b) K2OsO2(OH)4, K2Fe(CN)6,
K2CO3, DHQ-PHN, t-BuOH, 0 °C, 11 h, 98%; (c) benzene, reflux, 0.5
h, 98%; (d) (i) benzene, reflux, 0.5 h; (ii) THF, -23 °C, 6 h, 62%; (e)
(i) benzene, reflux, 0.5 h; (ii) THF, 0 °C, 3 h, 64%.
Swern oxidation gave enone 2. Conjugate addition of isobutyric
ester enolate to 2 took place smoothly to produce the keto ester
shown, which underwent Claisen-like cyclization on exposure
t
to BuOK. The resulting â-diketone was then converted to
pivaloate 3. Removal of the THP protecting group followed by
Swern oxidation afforded aldehyde 4 in ca. 50% overall yield
from propargylic ether. The above transformation could be
performed on a molar scale without purification of the inter-
mediates.
Conversion of 4 to enol silyl ether 5 was achieved as shown
in Scheme 3. Owing to the geminal methyl groups, enol ether
5 was obtained as the E-isomer exclusively. The geometry of
the double bond was determined by the observed NOE from
the gem-dimethyl group to the olefinic proton. Application of
Sharpless’s asymmetric dihydroxylation12 using DHQ-PHN as
a chiral ligand to 5 produced R-hydroxy aldehyde 6 in good
enantioselectivity (90% ee).13 However, isolation of 6 from the
a (a) (1) THF, -78 °C, 0.5 h; (2) CH2Cl2, 0 °C, 0.5 h, 96% in two
steps; (b) (1) THF, -78 °C, 2 h; (2) THF-MeOH, reflux, 4 h, cis-
isomer 54%, trans-isomer 7% from 9; (c) (1) benzene, reflux, 6 h; (2)
CH2Cl2, 0 °C to rt, 0.5 h, 64% from 10; (d) THF, rt, 2 days, 82%.
(13) Absolute stereochemistry was determined on the basis of the X-ray
crystallographic analysis of the derivative of 7. See also: Takenaka, Y.;
Kubo, S.; Ohashi, Y.; Nakamura, T.; Waizumi, N.; Horiguchi, Y.;
Kuwajima, I. Acta Crystallogr. 1994, C50, 1820.
(14) We initially planned benzyloxymethylenation of the carbonyl group
of 7 on the basis of the high efficacy realized in taxusin synthesis;8b however,
the reaction of 7 with benzyloxymethyllithium did not give the desired
adduct in acceptable yield.
(12) Kolb, H. C.; VanNieuwenhze, M. S.; Sharpless, K. B. Chem. ReV.
1994, 94, 2483.
(15) Mandai, T.; Hara, K.; Nakajima, M.; Otera, J. Tetrahedron Lett.
1983, 24, 4993.

3814 J. Am. Chem. Soc., Vol. 122, No. 16, 2000
Kusama et al.
of the corresponding trans-isomer. The O,S-acetal moiety was
changed to a dibenzyl acetal via dimethyl acetal 10. Conversion
of the vicinal diol moiety to the corresponding thionocarbonate
11, followed by heating with diazaphospholidine according to
Corey et al.,16 afforded cyclohexadiene 12. This compound
should be kept in the presence of 2,6-di-tert-butyl-4-methyl-
phenol (BHT) under a nitrogen atmosphere at low temperature;
otherwise 12 easily decomposes to form unidentified white
solids probably due to radical polymerization.
Scheme 6a
The chelation-controlled addition of the vinyllithium reagent
derived from 12 to 8 was performed in the presence of Mg(II)
ion to give the corresponding coupling product 13a as a single
isomer (Scheme 5). The stereochemistry at the C2 site was
determined on the basis of the observed NOEs of cyclized
1
product 15 at the later stage. The H NMR spectrum of the
resulting adduct was highly broadened. This suggested that 13a
consisted of atropisomers arising from the rotational barrier on
the C2-C3 bond, which were interconverting at room temper-
ature. This phenomenon was also found in a similar substrate
having an aromatic C-ring.17 On the contrary, connection of
the C1,2-diol moiety as the boronic ester gave 14 which showed
1
sharp signals in the H NMR spectrum. Preliminary studies
suggested that the boronate derivative consisted of the desired
(P)-isomer where an acetal moiety is closely situated to the
terminal of the dienol ether to favor the eight-membered B-ring
cyclization.17
Among several Lewis acid examined, TiCl2(OiPr)2 proved
to be the most efficient to induce the B-ring cyclization in this
case. The stereochemistry of cyclized product 15, namely the
conformation of the B-ring and configuration of the C9 and
C10 sites as well as the C2 hydroxyl group, was determined on
the basis of the observed NOEs shown below.18 For subsequent
removal of the boron protecting group, the use of pinacol with
DMAP efficiently produced the corresponding diol 16.
a (a) Benzene, reflux 1 h; (b) (1) CH2Cl2, -78 °C, 1 h; (2) CH2Cl2,
-23 °C, 1 h, 80% from 16; (c) (1) CH2Cl2, rt, 4 h; (2) benzene, reflux,
4 h, 85% from 18a; (d) EtOH, rt, 1 h; (e) CH2Cl2, 0 °C, 2.5 h, 87%.
Protection of the 1,2-diol moiety with a benzylidene group
produced R-isomer 17a exclusively, and reduction of the C13
keto group followed by silylation with TBSOTf afforded 18a
(Scheme 6). Singlet oxygen oxygenation of the 1,3-diene moiety
took place selectively from the â-face of the C-ring, and
subsequent treatment with Bu3SnH and a catalytic amount of
AIBN induced both the peroxide bond cleavage and removal
of the phenylthio group to yield 19a. Removal of the benzyl
group, followed by protection of the C7,9-diol, afforded 21a
as a mixture of two diastereomers (R:â ) ca. 1:1).
Scheme 5. Construction of the Tricarbocyclea
(b) Use of an Aromatic C-Ring Fragment. A drawback in
the use of the C-ring fragment 12 was lability of the 1,3-diene
moiety which appeared to undergo facile polymerization to form
a considerable amount of waxy substances under several reaction
conditions or during workup procedures. Although such po-
lymerization could be prevented by performing operations in
the presence of a small amount of BHT, development of a more
convenient route was desirable, especially for large-scale
operations. We conceived a new route by way of a C9-keto
tricarbocycle with a readily available aromatic C-ring. Thus,
2-bromobenzaldehyde dibenzylacetal 22 was used as a C-ring
fragment. Successive operations of coupling between the A-
and C-rings, protection of the 1,2-diol moiety, and SnCl4-
induced eight-membered B-ring cyclization gave the corre-
sponding tricarbocycle 23 after removal of the boron protecting
group. The stereochemistry of 23 was the same as that of 15
(Scheme 7).
From compound 23, a Birch reduction precursor was prepared
(Scheme 8). Reduction of the C13-carbonyl group followed by
protection of the C2- and C13-hydroxyl groups with TBSOTf
a (a) THF, -78 °C, 0.5 h (preparation of magnesium alkoxide), then
lithiated 12, -78 °C, 1 h, 68%; (b) benzene, rt, 0.5 h, 77%; (c) CH2Cl2,
-78 °C to 0 °C, 1.5 h; (d) benzene, rt, 59% from 14 in two steps.
(18) Although the cyclization precursor 14 was a mixture of geometrical
isomers on the vinyl sulfide moiety (E:Z ) ca. 1.5:1), E/Z isomerization
took place under the cyclization reaction conditions and the thermodynami-
cally most stable tricarbocycle 15 was obtained as a major product. For the
proposed mechanisms of the E/Z isomerization, see ref 9d.
(16) Corey, E. J.; Hopkins, P. B. Tetrahedron Lett. 1982, 23, 1979.
(17) Nakamura, T.; Waizumi, N.; Tsuruta, K.; Horiguchi, Y.; Kuwajima,
I. Synlett 1994, 584.

EnantioselectiVe Total Synthesis of (-)-Taxol
J. Am. Chem. Soc., Vol. 122, No. 16, 2000 3815
Scheme 7. Construction of the Tricarbocycle (Aromatic
Route)a
With the expectation that use of a sterically hindered alcohol
as a proton source would allow for preferential protonation on
the sterically less-demanding aromatic ring rather than at the
C9 site, we examined Birch reduction in the presence of various
highly substituted alcohols, and we were pleased to find that
the use of 2,2,4-trimethyl-3-isopropyl-3-pentanol led to a
satisfactory result. Thus, reduction of 25a (K/liquid NH3, THF,
-78 °C, 2 h) and successive in situ treatment with EtOH (rt, 2
h) gave an almost 1:1 mixture of conjugated diene 26 and benzyl
alcohol 27. Since the latter could be recovered as starting
material 25a via Swern oxidation in excellent yield, a total
conversion is considered acceptable for synthetic purposes.
Ketone 26 was converted to allylic alcohol 20a, which was
found to be identical to that obtained from the diene-type C-ring
fragment shown in Schemes 5 and 6 (Scheme 9).
a (a) THF, -78 °C, 0.5 h (preparation of magnesium alkoxide), then
lithiated 22, -78 °C, 1 h, 74%; (b) benzene, rt, 12 h, 90%; (c) (1)
CH2Cl2, -78 to -45 °C, 1.5 h; (2) benzene, rt, 40 h, 76% in two steps.
Scheme 8. Preparation of Birch Reduction Precursora
Scheme 9a
a (a) (1) CH2Cl2, -78 °C, 1.5 h; (2) CH2Cl2, -78 °C, 2 h, 82%
from 23; (b) benzene, reflux, 2 h, quantitative; (c) (1) EtOH, rt, 2 h,
91%; (2) CH2Cl2, -78 °C to rt, 0.5 h, quantitative.
a (a) THF, 0 °C, 1 h, 98%; (b) (1) benzene, reflux, 1 h, 89%; (2)
MeOH - THF, 0 °C, 1 h, quantitative; (c) CH2Cl2-MeOH, rt, 1 h; rt,
17 h, 74%.
produced 24. Removal of the phenylthio and benzyl groups
followed by Swern oxidation afforded the desired aryl ketone
25a.
We also investigated the Birch reduction of derivatives of
25a with various protecting groups on the C2-OH: reaction of
the di-tert-butylsilylene derivative 25b induced C2-O bond
cleavage predominantly (eq 3). Use of the C2-O-SEM substrate
25c gave a much more satisfactory result (eq 4), but we
encountered much difficulty in removing the SEM group at a
later stage; the SEM group was thus deemed to be unsuitable
for the present purpose.
Several attempts at Birch reduction of aryl ketone 25a
disclosed a rather unexpected result. Reduction in the presence
of tert-butyl alcohol as a proton source failed to reduce the
aromatic ring but produced the corresponding benzyl alcohol
27 exclusively (eq 2). This unusual phenomenon may be
attributed to the unique structural feature of substrate 25a which
allows 25a to exhibit behavior different from that of the usual
benzoyl group: semiempirical MO calculation of 25a using the
PM3 Hamiltonian19 revealed that the benzene ring is situated
almost perpendicular to the C9 carbonyl group.
Introduction of C19 Methyl Group
(a) Route via Cyclopropanation. For introduction of the
C19-methyl, we adopted a strategy similar to that of our taxusin
synthesis involving a hydroxyl group-directed cyclopropanation
on the ∆3,8-double bond, followed by ring cleavage of the
(19) SPARTAN (ver. 4.1.1) was used for PM3 calculations.

3816 J. Am. Chem. Soc., Vol. 122, No. 16, 2000
Kusama et al.
Scheme 10. Preparation of the Triol 20ba
cyclopropyl ketone. By using Et2Zn/ClCH2I20 or Et2Zn/ICH2I,
we attempted to transfer a methylene group on 21a, but the
∆3,8-double bond of 21a was found to be very unreactive, giving
back the starting materials irrespective of extensive examinations
(eq 5).
a (a) THF, -78 to 0 °C, 1.5 h; (b) (1) CH2Cl2, -78 °C, 1 h; (2)
CH2Cl2, -23 °C, 0.5 h, 62% from 16 in three steps; (c) (1) CH2Cl2, rt,
7 h; (2) benzene, reflux, 12 h, 80% in two steps; (d) EtOH, rt, 4.5 h;
(e) (1) THF, -78 °C, 2 h, 67%; (2) MeOH, 0 °C, 0.2 h, 79%; (f)
CH2Cl2, rt, 0.5 h, CH2Cl2-MeOH, rt, 0.5 h, 50%.
Scheme 11. Preparation of the Cyclopropyl Ketone 30aa
We assumed that the failure of 21a to undergo cyclopropa-
nation was due to the strong coordination of the C2 oxygen to
zinc on the C4-OH group where a methylene group on the zinc
has little chance to engage the ∆3,8-double bond. To circumvent
this problem, we explored di-tert-butylsilylene as a protecting
group to make the C2 oxygen much less basic. Thus, ketone
16 was treated successively with BuLi and tBu2Si(H)Cl at low
temperature. On warming, hydrogen evolved to give the
dioxasilapentane.21 A successive application of procedures
similar to those described in Scheme 6 afforded triol 20b, which
was also prepared via the aromatic C-ring route as shown in
Scheme 10.
Then the C7,9-diol moiety of 20b was protected with various
protecting groups and cyclopropanation of the ∆3,8-bond was
investigated. Although the substrates with a carbonate or an
acetonide group on the C7,9-diol did not react with the zinc
carbenoid reagent, the benzylidene-protected substrate was found
to undergo cyclopropanation (Scheme 11). Of the two diaster-
eomers of 21b (R:â ) ca. 1:4), the major â-isomer underwent
the expected methylenation on treatment with a zinc reagent
prepared by ZnEt2 and ClCH2I in toluene to produce 29, whereas
a (a) CH2Cl2, -23 °C, 12 h, 79%; (b) toluene, 0 °C, 0.5 h, 66%; (c)
CH2Cl2, rt, 12 h, 77%.
the methylene transfer could not be effected with the R-isomer.
NOE experiments of the R-isomer indicated that it contains the
boat conformation of the 1,3-dioxolane ring where the benzylic
hydrogen seems to prevent an approach of the reagent to the
double bond. Dess-Martin oxidation of 29 afforded the
corresponding cyclopropyl ketone 30a.
(20) Denmark, S. E.; Edwards, J. P. J. Org. Chem. 1991, 56, 6974.
(21) Tanino, K.; Shimizu, T.; Kuwahara, M.; Kuwajima, I. J. Org. Chem.
1998, 63, 2422.

EnantioselectiVe Total Synthesis of (-)-Taxol
J. Am. Chem. Soc., Vol. 122, No. 16, 2000 3817
During synthetic studies of taxusin, we discovered that the
C13-OH plays a critical role in converting the enol produced
by cyclopropane ring cleavage to the desired ketone. Conversion
of a ∆3,4-enol bearing a protected C13-hydroxyl to a C3R-
protonated ketone via intermolecular protonation is exceedingly
difficult because the protonation to C3 must occur from a narrow
space between the A- and C-rings. To overcome the difficulty
encountered in our prior synthesis of taxusin, we achieved this
enol/keto isomerization by exploiting the closely situated C13-
OH for protonation on C3 from the R-face.8b
Scheme 12. Reductive Cleavage of the Cyclopropyl Ketonea
On the basis of this observation, we initially removed the
C13- as well as C1,2-silicon protecting groups of 30a, and the
resulting triol 30b was used for the ring cleavage. Under the
influence of SmI2, the cyclopropane ring cleavage took place
readily, but the resulting enol was exceptionally unstable, and,
in the presence of air, the enol quickly decomposed to form a
complex mixture of unidentified products (eq 6).
a (a) (1) EtOH, rt, 0.5 h, 84%; (2) CH2Cl2, -45 °C, 2 h; (3) THF,
rt, 2.5 h, 95% in two steps; (b) (1) benzene, reflux, 1 h, 86%; (2) THF,
rt, 1 h, quantitative; (c) (1) THF, rt, 5 h, quantitative; (2) THF-cat.
BHT, rt, 3 h; (d) degassed MeOH-cat. BHT, rt, 2.5 days, 45% in two
steps (ca 50% of 31 was recovered); 65% by repeating this isomerization
procedure.
Then we decided to examine the enol/keto isomerization by
using the C1,2-diol-protected substrates. Several experiments
revealed that the choice of the protecting group on the C1,2-
diol and the C7,9-diol moieties was quite important to realize
the isomerization of the enol to the corresponding ketone, and
the following procedure was found to give a satisfactory
result: replacing the C7,9-protecting group of 30a with carbon-
ate, followed by removal of the di-tert-butylsilylene group with
TBAF, gave the diol. After protection of the C1,2-diol moiety
with a benzylidene group (single isomer), treatment with
methanolic K2CO3 yielded 30c. On exposure to SmI2-HMPA22
in methanol, 30c was smoothly converted to an enol stable
enough for manipulation under various reaction conditions.
Similar to the taxusin synthesis,8b the C13-protected enol did
not isomerize to the ketone at all, and removal of the TBS group
followed by basic treatment (NaOMe/MeOH) induced the
isomerization to give an almost 1:1 equilibrium mixture23 of
31 and the C3R-protonated ketone 32. Finally, 32 was obtained
in 65% yield by repeating twice this enol-keto isomerization
procedure. The stereochemistry of the C3 site was determined
by NOE experiments as well as by the coupling constant of the
C3 proton (doublet, J ) 5.6 Hz) (Scheme 12).
substituted 33, even with the use of any type of methylcopper
or manganese reagent (eq 7). A molecular modeling study
suggested that the C9-methoxy group might prevent the
conjugate addition to the C-ring enone moiety due to a
buttressing effect. Thus, we undertook an experiment using
acetonide-protecting substrate 34. Use of a combination of Me2-
CuLi/TMSCl24 induced the desired addition reaction to afford
methylated product 35 in good yield as the enol silyl ether (eq
8).
(b) Conjugate Addition Approach. Although starting from
the tricarbocycle 16 we could prepare the important synthetic
intetmediate 32 for taxol synthesis, the protecting groups of the
C1,2- and C7,9-diols had to be exchanged frequently to promote
the desired transformations as described in the above section.
To improve this situation, we reinvestigated a route including
a conjugate addition reaction for introduction of the C19-methyl
group. During our study of taxusin synthesis, we experienced
that introduction of a methyl group onto the C-ring enone via
conjugate addition could not be effected with C9,10-dimethoxy-
On the basis of these observations, we attempted to introduce
the angular methyl group via conjugate addition to enone 36,
which was readily available from triol 20a (Scheme 13). Various
(22) Batey, R. A.; Motherwell, W. B. Tetrahedron Lett. 1991, 32, 6211.
(23) Isomerization of the enol containing benzylidene groups on the C1,2-
and C7,9-diol moieties to the corresponding ketone did not proceed at all.
Thus, the free C7,9-hydroxyl groups were indispensable to obtain the desired
ketone.
(24) (a) Horiguchi, Y.; Matsuzawa, S.; Nakamura, E.; Kuwajima, I.
Tetrahedron Lett. 1986, 27, 4025. (b) Matsuzawa, S.; Horiguchi, Y.;
Nakamura, E.; Kuwajima, I. Tetrahedron 1989, 45, 349.

3818 J. Am. Chem. Soc., Vol. 122, No. 16, 2000
Scheme 13. Conjugate Addition Routea
Kusama et al.
Thus, starting from enantiomerically pure R-hydroxy aldehyde
8, the important synthetic intetmediate 32 could be prepared
by using the aromatic C-ring fragment 22 and conjugate addition
methodology without undergoing troublesome exchange of the
protecting groups on the C1,2- and C7,9-diol moieties.
Functional Group Manipulation for (-)-Taxol Synthesis
With the intermediate 32 in hand, the first task we faced was
how to discriminate three hydroxyl groups. The most desirable
process was to protect the C7- and the C13-OH with different
types of protecting groups and then to oxidize the C9-hydroxyl
group to the corresponding ketone. Since the C7-OH seemed
to be much less sterically demanding, selective protection of
the C7-hydroxyl group by silylation (TBSOTf/2,6-lutidine) was
initially examined, but silylation unexpectedly took place on
the C9-OH selectively. Second, oxidation of the C9-OH was
attempted, but Dess-Martin periodinane induced selective
oxidation of the C13 hydroxyl group. By taking account of these
features, we adopted the following strategy for protecting
intermediate 32. The C7,9-diol moiety was initially protected
as a boronic ester (PhB(OH)2/Py), and then the C13-OH was
silylated with TBSOTf. After removal of the boron protecting
group with hydrogen peroxide, selective oxidation of the C9-
OH could be achieved with the Dess-Martin periodinane. The
remaining C7-OH could then be protected as a 2-methoxy-2-
propyl ether (MOP),26 giving 42 in an appropriately differenti-
ated form.
a (a) (1) CH2Cl2, -23 °C, 1 h, 90%; (2) CH2Cl2, rt, 1 h, 95%; (b)
toluene, rt, 6 h, 74%; (c) THF, 0 °C, 0.2 h, 88%; (d) (1) CH2Cl2, -78
°C, 0.2 h, 78%; (2) THF, -78 °C, 0.3 h, 95%; (e) benzene-pyridine,
50 °C, 0.5 h, 88%; (f) acetone, rt, 0.5 h, quantitative.
Scheme 14a
efforts to direct the introduction of a methyl group or its
equivalent using copper reagents or stabilized carbanions were
fruitless, probably due to a much larger steric congestion around
the reaction site as compared with that of taxusin-type substrate
35. Nevertheless, a cyano group could be introduced quite
efficiently by treatment of 36 (R:â ) ca. 1:5) with Et2AlCN in
toluene to give stable enol 37 in good yield.25 Stereochemistries
of the benzylidene and p-methoxybenzylidene moieties were
determined by NOE experiments. The phenyl group of the
benzylidene moiety was assigned the R-configuration on the
basis of the NOE between the C2â-proton and the acetal methine
proton. As for the p-methoxybenzylidene moiety, the NOEs
between the acetal methine proton and the C7R as well as the
C10R protons revealed the â-configuration of the p-methox-
yphenyl group. Under the above reaction conditions, the
stereochemistry of the p-methoxybenzylidene moiety was
completely isomerized to the thermodynamically stable â-iso-
mer; however, isomerization of the benzylidene stereocenter also
occurred gradually to afford a small amount of the adduct
containing the â-benzylidene. Enol 37, the desired compound,
was converted to the TBS enol ether 38.
Next, the conversion of the cyano group of 38 to a methyl
group was attempted by using a Wolff-Kishner-type reduction
of the derived hydrazones or radical deoxygenation of the
alcohol derivatives such as a xanthate, but all attempts failed.
Furthermore, displacement of the hydroxyl group of 39 with a
bromine atom only gave cyclopropane derivative 40. Hence we
changed our plan to the efficient preparation of cyclopropyl
ketone 30c as a common synthetic intermediate of the diene
C-ring and the aromatic C-ring routes, and 40 was converted
to 30c by acid treatment quantitatively.
a (a) (1) CH2Cl2, rt, 0.25 h; (2) CH2Cl2, -45 °C, 12 h; (e) H2O-
AcOEt, rt, 1 h, 70% from 32 in three steps; (b) (1) CH2Cl2, rt, 1 h,
92%; (2) CH2Cl2, rt, 5 min, 97%; (c) THF, -78 °C, 1 h, 89%; (d)
Et2O, rt, 1 h, 91%; (e) (1) MeOH, rt, 10 h, 88%; (2) CH2Cl2, rt, 3 min,
89%.
(26) We initially chose a TES group as a protecting group on the C7-
OH. However, in the case of introduction of a C10 functionality, migration
of the TES group to a ∆9,10 enolate oxygen occurred to give an enol silyl
ether. Protection with Bn or BOM groups also failed to afford a
C7-epimerized alcohol under basic reaction conditions.
(25) Murai, A.; Tanimoto, N.; Sakamoto, N.; Masamune, T. J. Am. Chem.
Soc. 1988, 110, 1985.

EnantioselectiVe Total Synthesis of (-)-Taxol
J. Am. Chem. Soc., Vol. 122, No. 16, 2000 3819
Scheme 15. Dihydroxylation of the C-Ringa
The next problem was to introduce the methylene group on
the C4 site. We had already found Wittig reactions to be
inefficient in our taxusin synthesis; therefore we decided to
explore a more efficient methodology which may also serve to
introduce a C5 functionality for installation of the D-ring at a
later stage. To this end, we examined a selective conversion of
42 to 44 containing an appropriate allylsilane functionality via
enol triflate 43. Although 42 contains two ketone functions, the
difference in steric circumstances allowed us to selectively
generate a ∆4,5-enolate by treatment with KHMDS and quench-
ing with Tf2NPh to give 43 in good yield. Gratifyingly, the
reaction of 43 with the Grignard reagent shown in Scheme 14
in the presence of a catalytic amount of Pd(Ph3P)4 afforded 44
in good yield.27 Treatment of 44 with NCS in MeOH28 gave
the corresponding C5R-chloride 45.
For construction of the D-ring, introduction of a diol moiety
on the ∆4,20-double bond of 45 was next examined, but OsO4
oxidation unexpectedly took place at the ∆11,12-double bond on
the A-ring, yielding a C11,12-diol exclusively (eq 9).
a (a) (1) THF, -23 °C, 0.5 h, 80%; (2) CH2Cl2, rt, 0.5 h, 92%; (b)
toluene, reflux, 10 h, 68% at 92% conversion; (c) Et2O, rt, 12 h, 86%.
To increase steric congestion around the A-ring to effect
selective dihydroxylation on the exomethylene moiety, we
decided to introduce a C10 functional group initially. Func-
tionalization of the C10 site could be performed by generation
of the ∆9,10-enolate with LDA, followed by oxidation with
MoO5‚pyr‚HMPA (MoOPH),29 giving the C10R-alcohol selec-
tively. Although the stereochemical result was opposite from
the natural form, inversion of the C10R-oxygen functionality
to the thermodynamically more favored C10â configuration was
achieved by treatment with a base. Among the several attempts
examined, the following procedure gave the most satisfactory
result. Thus, acetylation of the C10-hydroxyl group (Ac2O/
DMAP) gave 46R which, on heating with DBN, induced clean
isomerization to afford the desired C10â-acetate 46â (Scheme
15).
Figure 2. Conformation of C10R- and C10â-Acetates.
which the R-face of the exomethylene was the least hindered
site for dihydroxylation (Figure 2).
On applying OsO4 oxidation to 46R and 46â, the former still
gave the corresponding C11,12-diol exclusively, whereas in the
reaction with the latter, dihydroxylation took place selectively
on the ∆4,20-site from the R-face to afford the desired 47 in
good yield.
Such a difference may be due to the unique structural features
of these tricarbocycles; NOE experiments as well as semiem-
pirical MO calculations19 of the C10R-acetate 46R indicated
that it contains the B- and C-rings as chair-chair and boat
conformations, respectively. In such a structure, both the upper
and lower sites of the ∆4,20-double bond suffer from severe steric
hindrance, and dihydroxylation should occur on the sterically
less demanding A-ring double bond. On the other hand, NMR
experiments on the C10â-acetate 46â revealed the conformer
containing the chair-boat B-ring and the chair form C-ring, in
Heating 47 with DBU in toluene cleanly induced the oxetane
ring closure to form the ABCD tetracyclic ring system 48
(Scheme 16). Then, we examined acetylation of the C4R-OH
of 48, but the reaction could not be effected under the usual
conditions (Ac2O or AcCl/Et3N-DMAP). Even on using a
combination of methyllithium and acetic anhydride, acetylation
took place very sluggishly to yield the desired product in only
ca. 30% yield. Considering that the phenyl group might prevent
the acetylation of the C4R-OH, a sequence of replacement of
the benzylidene group with the sterically less-demanding
carbonate was undertaken.
Since the MOP protecting group was also cleaved under the
reaction conditions for removal of the benzylidene group (H2/
Pd(OH)2), the MOP group was initially replaced by a TES group
and then the C1,2-diol protecting group was converted from
the benzylidene to the carbonate. The resulting carbonate cleanly
underwent acetylation at the C4R-OH to afford 50, and then
the carbonate group was converted to the C2-benzoyloxy group
by using known methodology,7a,b,30 giving the C7,13-protected
baccatin III. The TES group on the C7-OH was replaced by a
Troc group and the TBS group was removed by treatment with
(27) This cross-coupling reaction was highly sensitive toward the ligands
on palladium, and the use of triphenylphosphine was indispensable in
obtaining the coupling product.
(28) The use of methanol as solvent was essential to obtain the allyl
chloride in good yield. In the reaction of CH2Cl2, a vinylsilane was obtained
as a major product.
(29) Vedejs, E.; Engler, D. A.; Telschow, J. E. J. Org. Chem. 1978, 43,
188.

3820 J. Am. Chem. Soc., Vol. 122, No. 16, 2000
Kusama et al.
Scheme 16. Completion of the Total Synthesis of Taxola
Conclusion
The total synthesis of natural (-)-taxol has been achieved.
The most rewarding aspects are (1) an initial introduction of
an asymmetric center on C1 which has allowed for control of
all of the other asymmetric sites at later stages, (2) use of a
simple C-ring fragment such as bromobenzaldehyde, and (3) a
facile construction of the endo tricarbocycle, making it easy to
introduce requisite functionalities stereoselectively.
Introduction of the angular methyl presented a synthetic
problem. For cyclopropanation of the ∆3,8-double bond and the
following cleavage of the cyclopropyl ketone 30c, we had to
exchange the protecting groups often, making this synthetic route
rather lengthy. However, this problem could be overcome by
using the alternative conjugate addition approach.
From the unique ABC tricarbocyclic structures, several
interesting transformations were achieved: Birch reduction of
a highly distorted aryl ketone function, and isomerization of
the enol 31 to the ketone 32.
Total synthesis is always accompanied by the formation of
several artificial compounds, and we could produce several
substrates with taxane-like carbon skeletons; among these, new
useful drugs of stronger bioactivities might be produced.33
Acknowledgment. This work was financially supported by
the a Grant-in-Aid from the Ministry of Education, Science,
Sports, and Culture of the Japanese Government.
a (a) Toluene, reflux, 4 h, 86%; (b) (1) MeOH, rt, 0.5 h; (2) DMF,
rt, 5 h, 97% from 48 in two steps; (c) (1) EtOH, rt, 1 h, 97%; (2)
CH2Cl2, -78 to 0 °C, 1 h, 94%; (3) CH2Cl2, rt, 6 h, 66%; (d) (1) THF,
-78 °C, 83%; (2) THF, rt, 2 h, 88%; (3) CH2Cl2, rt, 3 h, 94%; (e) (1)
THF, rt, 2 days, 80%; (2) THF, -78 to 0 °C, 0.5 h, 77% at 90%
conversion; (f) AcOH-H2O, rt, 16 h, 84%.
Supporting Information Available: Spectroscopic data and
experimental procedures. This material is available free of
charge via the Internet at http://pubs.acs.org.
TASF.7a,31 Finally, acylation of the C13-OH by the â-lactam
method7a,32 followed by removal of the C7- and C2′-protecting
groups completed our enantioselective total synthesis of (-)-
taxol.
JA9939439
(32) Ojima, I.; Habus, I.; Zhao, M.; Zucco, M.; Park, Y. H.; Sun, C. M.;
Brigaud, T. Tetrahedron 1992, 48, 6985.
(33) Morihira, K.; Nishimori, T.; Kusama, H.; Horiguchi, Y.; Kuwajima,
I.; Tsuruo, T. Bioorg. Med. Chem. Lett. 1998, 8, 2973, 2977. Kusama, H.;
Morihira, K.; Nishimori, T.; Nakamura, T.; Kuwajiima, I. Tetrahedron Lett.
1999, 40, 4235.
(30) Wender, P. A.; Kogen, H.; Lee, H. Y.; Munger, J. D., Jr.; Wilhelm,
R. S.; Williams, P. D. J. Am. Chem. Soc. 1989, 111, 8957.
(31) Since the removal of the two silyl groups on 50 with TASF caused
partial epimerization at the C7 stereocenter, the TES group on the C7-OH
was exchanged to a Troc group.