Synthesis and biological evaluation of 2-acyl analogues of paclitaxel (Taxol)

Article abstract of DOI:10.1021/jm980229d

The anticancer drug paclitaxel (Taxol) has been converted to a large number of 2-debenzoyl-2-aroyl derivatives by three different methods. The bioactivities of the resulting analogues were determined in both tubulin polymerization and cytotoxicity assays, and several analogues with enhanced activity as compared with paclitaxel were discovered. Correlation of cytotoxicity in three cell lines with tubulin polymerization activity showed reasonable agreement. Among the cell lines examined, the closest correlation with antitubulin activity was observed with a human ovarian carcinoma cell line.

Full text of DOI:10.1021/jm980229d

J . Med. Chem. 1998, 41, 3715-3726
3715
Syn th esis a n d Biologica l Eva lu a tion of 2-Acyl An a logu es of P a clita xel (Ta xol)
David G. I. Kingston,* Ashok G. Chaudhary, Mahendra D. Chordia, Milind Gharpure, A. A. Leslie Gunatilaka,
Paul I. Higgs, J ohn M. Rimoldi, Lakshman Samala, and Prakash G. J agtap
Department of Chemistry, Virginia Polytechnic Institute and State University, Blacksburg, Virginia 24061-0212
Paraskevi Giannakakou
Medicine Branch, Division of Clinical Sciences, National Cancer Institute, National Institutes of Health,
Bethesda, Maryland 20892
Yuan Q. J iang, Chii M. Lin, and Ernest Hamel
Laboratory of Drug Discovery Research and Development, Developmental Therapeutics Program, Division of Cancer Treatment
and Diagnosis, National Cancer Institute, Frederick Cancer Research and Development Center, Frederick, Maryland 21702
Byron H. Long, Craig R. Fairchild, and Kathy A. J ohnston
Bristol-Myers Squibb Pharmaceutical Research Institute, P.O. Box 4000, Princeton, New J ersey 05843
Received April 13, 1998
The anticancer drug paclitaxel (Taxol) has been converted to a large number of 2-debenzoyl-
2-aroyl derivatives by three different methods. The bioactivities of the resulting analogues
were determined in both tubulin polymerization and cytotoxicity assays, and several analogues
with enhanced activity as compared with paclitaxel were discovered. Correlation of cytotoxicity
in three cell lines with tubulin polymerization activity showed reasonable agreement. Among
the cell lines examined, the closest correlation with antitubulin activity was observed with a
human ovarian carcinoma cell line.
In tr od u ction
conclusions that the C-13 ester side chain is essential
for activity, that modifications to the “northern hemi-
sphere” (C-7, C-9, and C-10) have modest but often
beneficial effects on its bioactivity, and that changes to
the “southern hemisphere” (C-4, C-2, and the oxetane
ring) can have larger effects on activity, usually negative
but occasionally positive.4 Thus opening of the oxetane
ring leads to loss of activity,5 as does deacetylation or
deacetoxylation at C-4.6 Replacement of the C-4 acetate
with other acyl groups, however, results in restoration
or even enhancement of activity.7
The C-2 benzoate has only recently been explored as
a target for SAR studies of paclitaxel. Following our
initial report of the enhanced activity of certain C-2
aroyl paclitaxels,8 reports confirming and extending this
result have appeared from other laboratories.9 The
importance of the C-2 substituent has been highlighted
by studies that connect the bioactivity of paclitaxel with
a “hydrophobic collapse” of the side chain with the C-2
benzoate and the C-4 acetate substituents.10 In this
paper we report further examples of C-2 aroyl-substi-
tuted paclitaxels, and we provide both tubulin assembly
and cytotoxicity data for all compounds.
Interest in the novel diterpenoid paclitaxel (Taxol) (1)
continues at a high level from chemical, biological, and
clinical viewpoints. First isolated in low yield in the
late 1960s from the bark of the western yew, Taxus
brevifolia,1 paclitaxel was not initially considered a
viable clinical candidate because of concerns about its
supply and its lack of aqueous solubility. Encouragingly
positive results in new in vivo models in the mid-1970s
spurred reconsideration of its potential, and paclitaxel
entered preclinical studies in the late 1970s. The
discovery of its novel mechanism of action, involving
enhancement of the polymerization of tubulin and
formation of stable microtubule polymers,2 greatly
increased interest in the compound. Clinical trials
ultimately demonstrated its clinical effectiveness, and
paclitaxel is now approved for treatment of ovarian and
breast cancers, with promise also for treatment of lung,
skin, and head and neck cancers.3
Ch em istr y
Studies on the preparation of 2-aroyl analogues of
paclitaxel were hampered initially by the difficulty of
removing the 2-benzoyl group selectively. Thus, treat-
ment of cephalomannine (2), a close analogue of pacli-
taxel, with methanolic base led only to cleavage of the
side chain or of the 10-acetate or both,11 while treatment
of 7-O-(triethylsilyl)hexahydrobaccatin III (3) under
A large number of structure-activity relationship
(SAR) studies of paclitaxel have been carried out,
particularly in recent years, and have led to the general
S0022-2623(98)00229-5 CCC: $15.00 © 1998 American Chemical Society
Published on Web 08/20/1998

3716 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 19
Kingston et al.
Sch em e 1a
similar conditions gave hydrolysis in the order 10-
acetate > 4-acetate > 2-benzoate.12 Debenzoylations of
7,13-di-O-(triethylsilyl)baccatin III and 7-O-(triethylsi-
lyl)-10-deacetylbaccatin III have been reported,12 but the
yields of 2-debenzoyl product were low, and the reac-
tions were unsuited to development of a general syn-
thesis.
In seeking to address this need for a general approach
to the preparation of 2-debenzoylpaclitaxel, we inves-
tigated the use of various nonaqueous reagents and
conditions. Although the initial discovery of the ef-
fectiveness of such conditions was empirical, their use
can be understood on the basis of recent NMR and
molecular modeling studies, which indicate that pacli-
taxel adopts a compact globular form in aqueous media
but a more open conformation in chloroform.10 It seems
probable that this more open conformation allows nu-
cleophilic attack to occur more readily at the C-2
benzoate. Three different routes were developed for the
C-2 debenzoylation of paclitaxel, and each will be
described briefly in turn.
always accompanied by formation of some of the ortho
ester derivative 5. Although the ortho ester 5 could be
converted to the desired tri(tert-butoxycarbonyl) deriva-
tive under mild conditions, we desired an alternative
procedure. Since the overreaction appeared to be due
to the forcing conditions necessary to acylate at the
7-position, we adopted the expedient of first preparing
2′-O-(tert-butoxycarbonyl)paclitaxel (6) and converting
it to its 7-O-(triethylsilyl) derivative 7 (Scheme 1).
Compound 7 could be converted in good yield into 2′-
O,3′-N-di(tert-butoxycarbonyl)-7-O-(triethylsilyl)pacli-
taxel (8).
Selective deacylation of simple N-tert-butoxycarbonyl-
N-acylimides normally occurs under basic conditions to
give N-deacyl-N-tert-butoxycarbonyl derivatives.13 How-
ever, treatment of 8 with various bases unexpectedly
failed to give any N-debenzoyl product, instead yielding
a mixture of the 2-debenzoyl product 9 and the rear-
ranged debenzoyl product 10.14 Under some conditions
the undesired tetrahydrofuran 10 was the major prod-
uct, but careful control of conditions enabled us to
prepare 9 in reasonable yield (76%) from 8.
Th e ter t-Bu toxyca r bon yl Rou te. Our interest in
the development of selective methods for N-debenzoy-
lation of paclitaxel or its congener cephalomannine led
us to react paclitaxel with di-tert-butoxy dicarbonate
under various conditions, in an attempt to form 2′-O,7-
O,3′-N-tri(tert-butoxycarbonyl)paclitaxel (4). Formation
of this compound required forcing conditions and was
The availability of 9 enabled us to investigate reacy-
lation conditions. The use of carboxylic acid in the
presence of dicyclohexylcarbodiimide (DCC) and 4-pyr-
rolidinopyridine (PP) proved to be the most effective
method. The conditions required careful control, since
under uncontrolled conditions rearrangement to 10

Synthesis and Evaluation of Taxol 2-Acyl Analogues
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 19 3717
Sch em e 2a
occurred. Under controlled conditions, however, the
protected paclitaxel 8 could be prepared by benzoylation
of 9. Deprotection of 8 with HCl/HCOOH yielded
paclitaxel (1), thus indicating that no untoward rear-
rangements had occurred. The yield in the deprotection
step was poor, however, and since attempts to improve
it proved unrewarding, we sought a better route.
Th e P h a se-Tr a n sfer Ca ta lysis Rou te. The second
route made use of phase-transfer catalysis, which allows
the use of water-soluble bases while maintaining the
substrate in a nonaqueous environment. As indicated
above, such conditions might be expected to facilitate
selective hydrolysis of the 2-benzoate by allowing the
molecule to “open up” in the absence of hydroxylic
solvents. Various solvents, bases, and phase-transfer
catalysts (PTCs) were investigated, and eventually the
combination of benzene/dichloromethane (8:2) as organic
solvent, tetrabutylammonium chloride as the PTC, and
2 N NaOH as the base was selected as providing the
best combination of selectivity and reactivity. Under
these conditions 2′,7-di-O-(triethylsilyl)paclitaxel (11a )
was an acceptable substrate, and it was converted to
2-debenzoyl-2′,7-di-O-(triethylsilyl)paclitaxel (12a ) in
reasonable yield (72%) if the reaction was stopped at
approximately 50% conversion (Scheme 2). If reaction
was pushed beyond 50% conversion, the rearranged
derivative 15a was formed in increasing amounts, and
the overall yield of the desired product declined.
yields from paclitaxel by this route ranged from 27% to
52%, depending primarily on the difficulty of the reacy-
lation step (12a f 13a ). The major byproduct in the
deacylation and reacylation steps was the rearranged
paclitaxel 15a or its acylation product 16a .
Th e Tr iton B Rou te. Although the phase-transfer
catalytic route provided acceptable yields of product, the
debenzoylation step was a difficult one requiring careful
monitoring of the reaction mixture and careful judge-
ment of when to stop the reaction. We thus sought an
improved method for this step. Reasoning that the use
of a bulky protecting group at the 2′-position would
reduce undesired hydrolysis of the C-13 ester side chain,
we elected to use a 2′-O-tert-butyldimethylsilyl protect-
ing group. The 7-position was then protected with an
O-triethylsilyl protecting group in the usual way. The
protected derivative 11b could be conveniently prepared
in 95% yield in a one-pot reaction. After trials of several
different sets of reagents and conditions, we found that
the lipophilic base Triton B (benzyltrimethylammonium
hydroxide in methanol) provided a convenient and high-
yield method to prepare the 2-debenzoyl analogue 12b.
Thus treatment of 11b with Triton B in dichlo-
romethane at -78 °C with subsequent warming to -10
°C gave the desired debenzoyl analogue 12b in 73%
yield, or in 81% yield based on unreacted starting
material. Although this procedure also requires careful
monitoring, it is somewhat simpler experimentally than
the phase-transfer method and is now our method of
choice for this reaction. Reacylation of 12b and depro-
tection of the product 13b to the paclitaxel analogues
14a a -eg proceeded smoothly as described above, and
this method, besides being simpler experimentally, gave
somewhat improved overall yields of 32-62%, again
depending primarily on the acid used. The major
Reacylation of 12a with various carboxylic acids in
the presence of DCC and PP gave protected 2-acylpa-
clitaxel analogues of general structure 13a . The yield
of isolated product was acceptable (60-90%, depending
on the acid), provided that a large excess of the acid
component was used. Deprotection with HCl/MeOH
gave the analogues 14a a -eg (see Table 1). Overall

3718 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 19
Ta ble 1. 2-Acylpaclitaxel Analogues Prepared
Kingston et al.
phenyl ring
phenyl ring
substituent
phenyl ring
substituent(s)
substituent(s) (14d f-d n ) or
2-substituent (14ea -eg)
compd
compd
compd
14a a
14a b
14a c
14a d
14ba
14bb
14bc
14bd
14be
14bf
14bg
14bh
14bi
14bj
14bk
14bl
14bm
2-OMe
2-F
2-Cl
14bn
14bo
14bp
14bq
14ca
14cb
14cc
14cd
14ce
14cf
14cg
14ch
14d a
14d b
14d c
14d d
14d e
3-Br
3-CF3
3-CN
3-NO2
4-OH
4-OMe
4-OCH2Ph
4-F
4-N3
4-Cl
4-CN
4-NO2
2,3-F
14d f
14d g
14d h
14d i
14d j
14d k
14d l
14d m
14d n
14ea
14eb
14ec
14ed
14ee
14ef
14eg
3,4,5-OMe
3,5-OMe
3,4-F
3,5-N3
3,4-Cl
3,5-F
3,4,5-F
3,5-Cl
3,5-NO2
2-thiophenecarbonyl
3-thiophenecarbonyl
2-furoyl
3-furoyl
phenoxyacetyl
acetyl
2-N3
3-OEt
3-OPr
3-OiPr
3-OCH2Ph
3-OH
3-OMe
3-SMe
3-OPh
3-N3
3-F
3-I
3-Cl
3-COMe
2,5-F
2,4,5-F
2,3-OMe
2,5-OMe
valeryl
Sch em e 3
Sch em e 5
Sch em e 4
Sch em e 6
byproduct in the deacylation and reacylation steps was
again the rearranged paclitaxel 15b or its acylation
product 16b.
cyclic carbonate 20, which is formed in about 60%
overall yield from 19 as a key intermediate (Scheme 4).
The synthesis is somewhat lengthy, requiring four steps
from 19 to 20 and six steps from 20 to a 2-acylpaclitaxel.
The combined yields of these latter steps range from
3% to 25%, giving an overall yield for the route of
2-15%, depending on the acyl group.
Discu ssion . It is interesting to compare the relative
advantages and disadvantages of the various routes that
have now been reported for the preparation of 2-acyl-
2-debenzoylpaclitaxels. Four approaches have been
proposed, involving either paclitaxel or baccatin III as
the starting material and proceeding either through a
cyclic carbonate intermediate or by direct acylation at
C-2. One of the first methods was developed by Holton
and proceeds from 10-deacetylbaccatin III to the cyclic
1,2-carbonate derivative 18 and thence to a 2-aroylbac-
catin III analogue (Scheme 3).9h Thus protection of 10-
deacetylbaccatin III as its 7,10-di-O-(triethylsilyl)-13-
O-(trimethylsilyl) derivative followed by debenzoylation
at C-2 by treatment with Red-Al gave the 2-debenzoyl
analogue 17. Reaction of 17 with phosgene and pyridine
gave the cyclic carbonate 18, and treatment of 18 with
an aryllithium reagent gave a protected 2-debenzoyl-2-
aroylbaccatin III analogue. Conversion of this analogue
to a corresponding paclitaxel analogue could then be
accomplished by appropriate manipulations to attach
the C-13 ester side chain.15
A pathway that combines the direct acylation chem-
istry of our method with a baccatin III starting material
was developed by Chen et al.9e (Scheme 5). This route
proceeds via 7,13-di-O-(triethylsilyl)baccatin III (21),
which is debenzoylated with Red-Al to the 2-debenzoyl
derivative 22. Reacylation of this intermediate followed
by side-chain attachment gives 2-acylpaclitaxel ana-
logues in overall yields of 15-27%, depending on the
acid used. This method is thus somewhat more direct
than Nicolaou’s approach, while retaining the advan-
tages associated with a baccatin III starting material.
A similar pathway from 10-deacetylbaccatin III was also
developed by Holton.9h A final pathway from a baccatin
III starting material was developed by Chen9e and
improved by Ojima.9i It involves the deacylation of the
13-oxobaccatin III derivative 23 (Scheme 6) and subse-
quent reacylation and reduction to give the baccatin III
analogue 24.
The cyclic carbonate approach of Nicolaou9d (Scheme
4) is similar in many ways to Holton’s approach. It
proceeds from 10-deacetylbaccatin III (19) through the
Finally, Georg’s method9b is very similar to the Triton
B route described here. It uses a different base and

Synthesis and Evaluation of Taxol 2-Acyl Analogues
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 19 3719
slightly different reacylation conditions, but it is very
comparable in terms of yield and convenience.
compounds is significant for the data presented in
columns 3 and 4 of Table 2.
Each of the methods described has its own advantages
and limitations. The methods of Holton, Nicolaou, and
Chen, starting as they do from baccatin III or 10-
deacetylbaccatin III, provide a route for the synthesis
of paclitaxel analogues modified both at C-2 and on the
side chain and are thus the most versatile methods in
principle. The methods involving aryllithium addition
to a cyclic carbonate are, however, limited to the
preparation of analogues carrying substituents that are
compatible with the aryllithium functionality and can-
not be used to prepare derivatives with nitro, cyano,
azido, carbonyl, and similar groups. The routes de-
scribed in this paper and by Georg are both efficient
methods for the conversion of paclitaxel to various
2-debenzoyl-2-acyl analogues and are thus the method
of choice for this conversion.
In summary, the chemistry described provides a
convenient access to a wide range of 2-acylpaclitaxel
analogues through a carefully optimized set of reactions.
The method is versatile, since it is not limited to those
acyl groups which can be formed via a stable carbanion,
and it has been used successfully for the preparation of
the 50 paclitaxel analogues listed in Tables 1.
In addition, most of the analogues were evaluated in
a screening assay with microtubule protein at 37 °C
without GTP (column 5, Table 2). The microtubule-
associated proteins in the preparation have an enhanc-
ing effect on tubulin polymerization similar to that of
the higher glutamate concentration,18 and the higher
reaction temperature can further enhance taxoid effects
on tubulin assembly.19
Most of the compounds in the series were also
evaluated in four human cancer cell lines. The Burkitt
cells were grown in suspension culture, and the cell
number was quantitated. The ovarian and colon car-
cinoma lines were grown in monolayer cultures, and cell
protein was quantitated.
In conjunction with the parental ovarian line,20
a
paclitaxel-resistant line that expresses an altered cel-
lular â-tubulin21 was evaluated. The mutant ovarian
line was 21-fold more resistant to paclitaxel than the
parental line. Note that the data with this line are
presented relative to the resistance level obtained with
paclitaxel (column 8, Table 2). Thus, full restoration
of relative sensitivity with a modified taxoid would be
indicated by a value of about 0.05, and this was not
observed with any compound.
Biologica l Eva lu a tion
Tu bu lin -Ba sed Assa ys. There is excellent overall
agreement between the more extensive evaluations with
purified tubulin and the screening evaluation with
microtubule protein. There were only five significant
discrepancies between the two assays, and in all cases
these were compounds that were found to be less active
than paclitaxel in the purified tubulin assay. In the
screening assay four of these compounds (14b j,b k ,
bq,d i) were indicated to be more active than paclitaxel
and one (14bo) to be nearly equivalent to the parent
drug. Surveying the data overall, there is satisfactory
agreement between the biochemical and cytological
results (discussed below in greater detail). In only a
few cases does the microtubule data appear to be more
predictive than the purified tubulin data on the effects
of these compounds on cell growth (compounds 14bj,bk ,
bo,d i). Therefore, since the purified tubulin data were
obtained with the entire series of agents and verified
by repeat assays, we will primarily use these data for
comparisons with the cell proliferation data and struc-
ture-activity analysis.
Com p a r a tive Effects on Cell Gr ow th . The Bur-
kitt cells appear to be significantly less sensitive to
paclitaxel than the two parental carcinoma lines, as the
former had an IC50 value of 30 nM and the latter IC50
values of about 2 nM. We have not determined whether
this is due to an intrinsic difference in sensitivities or
to an artifact (e.g., the different incubation times,
culture methods, or parameters measured). Overall, the
range of activity was much smaller with the Burkitt
cells than with the other lines. Nevertheless, there was
broad similarity in sensitivity to the C-2-modified
analogues in most cases. Several compounds (14bf,d b,
d e,d m ) were more active than paclitaxel in all three
cell lines, and others were more active in at least two
lines (14a d ,bi,bj,bl,bn ,d k ). Except for compound 14bj,
these analogues were more active with tubulin as well,
and 14bj was more active than paclitaxel in the micro-
We have previously reported8,9f,16 that 2-debenzoyl-
2-(m-azidobenzoyl)paclitaxel (14bi) was more potent
than paclitaxel in enhancing tubulin polymerization and
inhibiting cell growth of the murine P388 leukemia,
human HL-60 leukemia, and human Burkitt lymphoma
CA46 lines and of several lines in the National Cancer
Institute human cancer screen. These observations
with 14bi, as well as less extensive data with several
other C-2-modified compounds,8 caused us to prepare
the series of compounds described here. We wished to
determine the effects of different modifications in the
C-2 substituent on the activity of taxoids and hoped to
prepare compounds more active than 14bi. Table 2
summarizes our evaluation to date of most of the
compounds whose syntheses are described here; data
for compounds not discussed in the text can be found
in the complete tables appearing as Supporting Infor-
mation. For all comparisons, we have expressed ana-
logue activity relative to that of paclitaxel, so that values
less than 1.0 indicate activity greater than that of the
parent compound and values greater than 1.0 activity
less than that of the parent compound.
We specifically designed a relatively facile assay16
with purified tubulin that is selective for compounds
more active than paclitaxel toward tubulin, the drug’s
cellular target. We chose a reaction condition (0.4 M
glutamate, no Mg2+, room temperature incubation, no
GTP) where paclitaxel yielded a relatively high "IC50"
value (23 ( 2 µM)17 and 14bi a low one (4.7 ( 0.1 µM).
Compounds less active than paclitaxel were also evalu-
ated in 0.8 M glutamate, a modification that substan-
tially enhances the activity observed with paclitaxel16
(IC50 value, 4.5 ( 0.1 µM), while that of 14bi changes
little (IC50 value, 2.8 ( 0.06 µM). Compounds with IC50
values less than 40 µM in either assay were evaluated
several times to confirm their activity. Based on the
standard deviations obtained, a 20% difference between

3720 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 19
Kingston et al.
Ta ble 2. Bioactivity Data for 2-Acylpaclitaxels Relative to That for Paclitaxela
cytotoxicity with human cell lines
polym of tubulin in
phenyl ring
substituent (14a a -d n ) or
polym of
microtubule
rel resistance
of paclitaxel-
glutamateb (22 °C)
Burkitt
ovarian
colon
compd 2-O-substituent (14ea -eg)
0.4 M
0.8 M
1.2
proteinc (37 °C) lymphomad carcinomae resistant linef carcinomag
14a a
14a d
14ba
14bc
14bf
14bg
14bi
14bj
14bk
14bl
14bn
14bo
14bp
14bq
14ca
14cf
14d a
14d b
14d c
14d d
14d e
14d f
14d g
14d h
14d i
14d j
14d k
14d l
14d m
14d n
14ea
14eb
14ec
14ed
2-OMe
2-N3
3-OEt
3-OiPr
3-OMe
3-SMe
3-N3
3-F
3-I
3-Cl
3-Br
3-CF3
3-CN
3-NO2
4-OH
4-Cl
2,3-F
2,5-F
2,4,5-F
2,3-OMe
2,5-OMe
3,4,5-OMe
3,5-OMe
3,4-F
3,5-N3
3,4-Cl
3,5-F
>1.7
0.38
1.4
1.5
0.70
1.1
0.67
1.4
0.77
1.7
15
0.095
3.3
0.18
0.40
1.6
0.05
0.32
1.4
0.32
1.3
31
7.6
14
0.95
0.67
0.37
0.29
0.71
0.52
0.27
0.96
0.35
2.9
1.9
0.36
1.6
17
0.35
1.2
1.0
1.3
0.27
1.0
0.17
>1.7
0.22
>1.7
0.20
1.2
>8.9
>140
0.50
2.7
0.37
0.70
0.79
0.92
0.80
1.3
1.0
0.69
19
1.4
2.6
3.1
0.45
0.85
1.7
0.32
3.0
1.1
0.88
52
>1.7
0.30
0.24
>1.7
0.63
>1.7
>1.7
>1.7
>1.7
0.29
>1.7
>1.7
0.20
>1.7
>1.7
>1.7
>1.7
0.41
0.29
>1.7
0.91
>1.7
>1.7
>1.7
>1.7
>1.7
2.3
0.67
0.42
NMh
1.3
6.7
0.67
1.6
>8.9
>8.9
>8.9
1.0
3.0
1.3
0.76
NM
NM
NM
0.57
0.48
NM
0.76
NM
0.57
0.57
0.40
0.27
1.1
0.62
0.48
NM
NM
NM
NM
NM
0.80
0.34
1.5
0.55
0.44
1.9
59
0.14
110
5.4
2.7
1.6
0.77
0.36
3.4
0.51
3.1
4.5
4.5
52
18
0.67
1.0
>3.3
0.67
>3.3
1.0
0.33
1.0
0.67
3.1
3.1
>8.9
>270
0.50
32
>48
0.12
>8.9
4.7
3.1
>48
2.6
2.2
3.3
3.4
1.3
1.0
4.0
0.67
0.76
0.60
1.8
0.44
4.2
4.1
1.9
26
7.8
0.12
3.2
0.38
20
4.8
4.3
3,4,5-F
3,5-Cl
4.7
1.3
0.67
2.3
0.67
0.33
2.0
3,5-NO2
2-thiophenecarbonyl
3-thiophenecarbonyl
2-furoyl
3-furoyl
>8.9
2.9
2.0
5.1
3.1
13
12
1.0
a
Activity data are recorded relative to the activity of paclitaxel in the corresponding assay, so numbers less than unity correspond to
analogues that are more active than paclitaxel. The complete table with data for all compounds is included as Supporting Information.
b
The analogues were evaluated as enhancers of the assembly of purified bovine brain tubulin at room temperature without GTP at the
indicated concentrations of monosodium glutamate (taken from 2 M stock solution adjusted to pH 6.6 with HCl).16 Polymer was harvested
by centrifugation, and the IC50 was defined as the drug concentration reducing the protein concentration of the supernatant by 50% as
compared to control reaction mixtures without drug. With active compounds, the pellet becomes progressively larger as compound
concentration increases. The highest drug concentration used was 40 µM. Compounds yielding IC50 values were evaluated in at least
three independent experiments, and those with IC50 values over 40 µM were evaluated at least twice. The values obtained for paclitaxel
were 23 ( 2 (SD) µM in 0.4 M glutamate and 4.5 ( 0.1 µM in 0.8 M glutamate. c Twice cycled microtubule protein was prepared following
the procedure of Williams and Lee27 and stored in liquid nitrogen before use. Quantification of tubulin polymerization potency was
accomplished following a modified procedure of Swindell et al.28 These modifications, in part, result in the expression of tubulin
polymerization potency as an effective concentration for any given compound. For this method, different compound concentrations in
polymerization buffer (0.1 M MES, 1 mM EGTA, 0.5 mM MgCl2, pH 6.6) were added to microtubule protein in polymerization buffer at
37 °C in microcuvette wells of a Beckman (Beckman Instruments) model DU 7400 UV spectrophotometer. A final microtubule protein
concentration of 1.0 mg/mL and compound concentrations of 2.5, 5.0, and 10 µM were used. Initial slopes of absorbance change measured
every 10 s were calculated by the program accompanying the instrument after initial and final times of the linear region encompassing
at least three time points were manually defined. Under these conditions linear variances were generally <10-6, slopes ranged from 0.03
to 0.002 A unit/min, and maximum absorbance was 0.15 A unit. Effective concentration (EC0.01) is defined as the interpolated concentration
d
capable of inducing an initial slope of 0.01 A unit/min and is calculated using the formula: EC0.01 ) concentration/slope. Human Burkitt
lymphoma CA46 cells were grown in 5.0-mL suspension cultures and counted after 24 h of growth at 37 °C under 5% CO2.25 IC50 values,
defined as the drug concentration required to reduce the increase in cell number by 50%, were determined in a minimum of two independent
experiments. The average IC50 value for paclitaxel was 30 nM. The highest drug concentration evaluated was 100 nM. e The human
ovarian carcinoma line 1A9 was derived by subcloning line A2780,20 and cells were grown in 96-well titer plates for 96 h at 37 °C under
5% CO2 in 0.1 mL of medium. The IC50 value was the drug concentration required to inhibit by 50% the increase in cell protein.26 The
average IC50 value for paclitaxel was 2.1 nM. The highest drug concentration examined was 100 nM. f The paclitaxel-resistant line
1A9(PTX22) was derived from 1A9 by a one-step selection in the presence of paclitaxel and verapamil (added to minimize likelihood of
selecting a P-glycoprotein-expressing line). 1A9(PTX22) expresses an altered â-tubulin that confers resistance to paclitaxel.21 IC50 values
for each drug were obtained in tandem with IC50 values for the parental line 1A9. The average IC50 value for paclitaxel was 44 nM, so
that the relative resistance of 1A9(PTX22) to the parental line for paclitaxel was 21-fold. The values presented in the table are the
relative resistance for the analogue divided by the relative resistance for paclitaxel. NM means no IC50 value was obtained for the analogue
g
in the parental and/or resistant line. Cells of the human colon carcinoma line HCT116 were grown in 96-well titer plates for 48 h at 37
°C under 5% CO2 in 0.1 mL of medium. The IC50 value was the drug concentration required to inhibit by 50% the increase in cell protein.
h
The average IC50 values for paclitaxel in the experiments presented here were in the range 2.0-2.5 nM. NM, not meaningful, indicating
that no IC50 value was obtained (i.e., IC50 >100 nM) for the parent line, for the resistant line, or for both lines.
tubule protein assay. Compounds that showed the
greatest differential activities between the three lines
were the following: 14bf,bl,bp ,d k (ovarian line rela-
tively sensitive); 14a a ,bc,be,bg,bo,ca ,cf,d g,d h ,ea ,eb,
ec,ed (Burkitt line relatively sensitive); 14a d ,bk ,d a
(colon line relatively sensitive); 14d a ,d n (ovarian line

Synthesis and Evaluation of Taxol 2-Acyl Analogues
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 19 3721
against the cell growth data obtained with the parental
ovarian line (Figure 1) and the colon line (Figure 2). The
best correlation of the tubulin assembly activity was
with the ovarian line. Every compound, except two, that
was more active than paclitaxel with tubulin was also
more active as an inhibitor of ovarian cell growth, and
the exceptions were similar to paclitaxel in their cyto-
toxicity toward the ovarian cells. Conversely, among
compounds with greater antiproliferative activity than
paclitaxel, only one was less potent than paclitaxel in
enhancing tubulin assembly. The poorest correlation
of the tubulin assembly data was with inhibition of
growth of the Burkitt cells (data not shown).
As was noted above, no analogue was able to reverse
completely paclitaxel resistance based on an altered
â-tubulin in the ovarian line 1A9(PTX22).21 Neverthe-
less, there did appear to be significant partial reversal
with a number of analogues. Restoration of taxoid
sensitivity appeared to be greatest (analogue/paclitaxel
< 0.4) with compounds 14ba ,bc,bi,bk ,d j. Of these, only
14bi,d j were more active than paclitaxel with purified
tubulin, but, perhaps significantly, 14bk was among the
compounds that were more active than paclitaxel in the
microtubule protein screening assay.
F igu r e 1. Correlation of analogue effects on stimulation of
tubulin assembly and on inhibition of growth of human ovarian
carcinoma 1A9 cells. The data of Table 2, columns 3 and 4
(the values for compounds 14a b,ba ,bj were averaged) are
plotted against the data of column 7. The cross indicates the
position of paclitaxel. Activities relative to paclitaxel are
plotted in both dimensions. The cluster of symbols at the upper
right represents 11 analogues inactive in both assays at the
highest drug concentrations examined. The correlation coef-
ficient for these data is 0.903 (obtained with Origin Microcal
Version 4.1).
Str u ctu r e-Activity Rela tion sh ip s. This discus-
sion will focus on the effects of this series of analogues
in enhancing the assembly of purified tubulin. As noted
above, correlations between the tubulin data and the
cell growth data are strong, but not absolute. Note that,
for ease of comparison, the compounds have been
tabulated throughout, as far as possible, in order of
increasing Hammett σ constant,22 and in no assay was
there a trend of activity with changes in the electron-
withdrawing or electron-donating effect of a substituent.
As reported initially,8 and confirmed more extensively
here, with benzenoid analogues substituent position was
highly important. Among compounds with single para-
substituents (compounds 14ca -ch ), only 14cd with a
p-fluoro group yielded an IC50 value, and it was 6.2-
fold higher than the value obtained with paclitaxel.
Moreover, introduction of a para-substituent into an
otherwise hyperactive analogue reduces the apparent
interaction with tubulin (cf. 14d h with 14bj, 14d j with
14bl, 14d c with 14d b, 14d l with 14d k , and 14d f with
14d g), although in one case (14d j) the resulting com-
pound is still more active than paclitaxel. Since the
negligible activity of the para-substituted derivatives
(14ca -ch ) is independent of the electronic nature of the
substituent, it seems likely that there is a significant
steric effect reducing binding of para-substituted C-2
benzenoid analogues to the paclitaxel site on tubulin.
F igu r e 2. Correlation of analogue effects on stimulation of
tubulin assembly and on inhibition of growth of human colon
carcinoma HCT116 cells. The data of Table 2, columns 3 and
4 (the values for compounds 14a b,ba ,bj were averaged) are
plotted against the data of column 9. The cross indicates the
position of paclitaxel. Activities relative to paclitaxel are
plotted in both dimensions. The cluster of symbols at the upper
right represents 4 analogues inactive in the tubulin assay at
the highest drug concentration examined and with IC50 values
greater than 200 nM in the cell proliferation assay. The
correlation coefficient for these data is 0.815 (obtained with
Origin Microcal Version 4.1).
In view of the apparent steric origin for the lack of
activity of the para-substituted analogues, we reasoned
that the problem might be avoided with meta-substitu-
tion. Such analogues theoretically could rotate about
the bond between the aryl ring and the carbonyl group.
Consistent with this idea, four analogues with relatively
small meta-substituents (14bf, methoxy; 14bi, azido;
14bl, chloro; 14bp , cyano) were significantly more active
than paclitaxel as inducers of tubulin assembly. Other
factors, perhaps electronic in origin, are clearly involved
in the interaction of C-2-modified taxoids with tubulin,
however, since some analogues with relatively small
substituents had moderately (14ba , ethoxy; 14bg, thi-
relatively resistant); and 14ec (colon line relatively
resistant).
We also evaluated the tubulin assay for its ability to
predict a compound’s antiproliferative activity. Figures
1 and 2 present graphical representations of the data
of Table 2, with the purified tubulin polymerization data
(both the 0.4 and 0.8 M glutamate results) plotted

3722 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 19
Kingston et al.
omethyl; 14bj, fluoro; 14bq, nitro) or severely (14be,
hydroxy; 14bm , acetyl) reduced activity, and one com-
pound with a relatively bulky bromo group (14bn ) was
more active than paclitaxel. Thus, significant steric
trends are evident in the alkoxy series (methoxy >
ethoxy > propoxy) but not in the halide series (Br > Cl
> F > I). With the exception of the m-bromide 14bn ,
no substituent group with a molecular mass greater
than 42 gave an analogue with activity greater than
that of paclitaxel.
previously described compound 14bi (2-debenzoyl-2-m-
azidobenzoylpaclitaxel). The best correlation between
tubulin assembly effects measured in a biochemical
assay and antiproliferative effects on cells was observed
in an ovarian carcinoma line designated 1A9. None of
the analogues synthesized could completely overcome
paclitaxel resistance due to an altered â-tubulin, but
compound 14bi and a number of other analogues
partially overcame this type of resistance to paclitaxel.
Exp er im en ta l Section
Only a small number of ortho-substituted benzoyl
analogues were prepared (compounds 14a a -a d ), and
all had small substituents. These were less active than
the analogous meta-substituted compounds, but the
o-azido derivative (14a d ) was significantly more active
than paclitaxel; it was also more cytotoxic than pacli-
taxel toward the ovarian and colon carcinoma cell lines.
A number of analogues with multiple substituents on
the benzoyl ring were also prepared (14d a -d n ). As
noted above, a para-substituent always made a negative
impact on activity, and generally a major negative
impact. Other substitution patterns were more vari-
able. In four of five analogues, a di-meta-substituent
pattern led to loss of activity (14d g, methoxy; 14d i,
azido; 14d m , chloro; 14d n , nitro; 14d m , however, did
have nominally23 greater activity than paclitaxel). The
exception was the di-m-fluoro analogue 14d k , which
was more active than the mono-m-fluoro compound
14bj. With ortho-meta-disubstitution, the results
obtained depended on whether the substituents were
on the same or opposite sides of the benzene ring,
although only two examples were available. Thus, if
the substituents were vicinal, the ortho effect domi-
nated, with the disubstituted analogues less active than
the monosubstituted meta-compounds (cf. 14d a with
14bj; 14d d with 14bf). In contrast, when the substit-
uents were on opposite sides of the benzene ring, there
was no loss or even enhancement of the activity ob-
served with the monosubstituted meta-compounds (cf.
14d b with 14bj; 14d e with 14bf).
Finally, in our effort to obtain analogues more active
than paclitaxel, we prepared a number of analogues
with non-benzenoid substituents at position C-2. The
nonaromatic analogues, such as acetyl (14ef) and
valeryl (14eg), were significantly less active than pa-
clitaxel, and even the aromatic analogues such as the
furoyl analogues 14ec,ed were also significantly less
active than paclitaxel in all assays. Only the thiophene-
carbonyl analogues 14ea ,eb showed improved activity
and then only in the antiproliferative studies with
Burkitt cells. These results parallel those obtained by
Nicolaou et al.,9d who observed that the thiophenecar-
bonyl analogues were the only non-benzenoid 2-acyl
analogues that showed improved antiproliferative activ-
ity as compared with paclitaxel, and this enhanced
activity was only observed in a few cell lines.
Gen er a l Exp er im en ta l P r oced u r es. All chemicals other
than solvents were obtained from Aldrich Chemical Co. and
used without further purification. All anhydrous reactions
were performed under Ar. THF was dried over sodium/
benzophenone. All reactions were monitored by TLC (silica
gel, GF) and the plates examined with UV light and developed
with vanillin spray. 1H and 13C NMR spectra were obtained
in CDCl3 at 270 and 400 MHz for proton spectra and assigned
primarily by comparison of chemical shifts and coupling
constants with those of related compounds and by appropriate
2D NMR techniques. Coupling constants are reported in Hz.
13C NMR spectra were assigned using HETCOR and DEPT
spectra. 1H NMR spectra sometimes showed the presence of
traces of ethyl acetate, since paclitaxel and its derivatives
retain ethyl acetate very tightly, and it cannot be removed
completely even on prolonged treatment in vacuo at 38 °C.
Exact mass measurements were performed at the Nebraska
Center for Mass Spectrometry. All compounds were homoge-
neous by TLC and were >90% pure as judged by NMR. All
compounds showing activity greater than paclitaxel were also
analyzed by HPLC (C-18 column, H2O/MeOH, 30:70) and
repurified if necessary to 95% purity or better.
2′-O,7-O,3′-N-Tr i(ter t-bu toxyca r bon yl)p a clita xel (4). A
solution of 2′-O,7-O,3′-N-tri(tert-butoxycarbonyl)-1,2-(R-tert-
butoxybenzylideneacetal)paclitaxel (5) (35 mg, 0.027 mmol) in
CHCl3 (2 mL) was stirred at room temperature for 24 h. It
showed disappearance of starting compound with the forma-
tion of new polar compound on TLC. The reaction mixture
was concentrated and dried under reduced pressure to afford
2′-O,7-O,3′-N-tri(tert-butoxycarbonyl)paclitaxel (4) (30 mg,
94%). 1H NMR: δ 1.05 (s 9H), 1.10 (s 6H), 1.36 (s 9H), 1.38-
1.41 (m, 2H), 1.44 (s 9H), 1.52 (s 1H), 1.74 (s 3H), 1.84 (s 3H),
1.87-1.91 (m, 1H), 2.09 (s 3H), 2.39 (s 3H), 2.58-2.64 (m, 1H),
3.87 (d, J ) 7.02 Hz, 1H), 4.11 (d, J ) 7.17 Hz, 1H), 4.13 (d,
J ) 7.17 Hz, 1H), 4.27 (d, J ) 8.24 Hz, 1H), 4.94 (d, J ) 7.78
Hz, 1H), 5.31-5.36 (m, 1H), 5.60 (d, J ) 7.17 Hz, 1H), 5.84 (d,
J ) 11.14 Hz, 1H), 5.92 (d, J ) 11.29 Hz, 1H), 5.97 (t, J )
9.00 Hz, 1H), 6.46 (s 1H), 7.12-7.15 (dd, J ) 7.78 and 7.47
Hz, 1H), 7.33-7.41 (m, 4H), 7.47-7.56 (m, 4H), 7.62-7.69 (m,
4H), 8.07 (d, J ) 7.01 Hz, 2H). 13C NMR: δ 10.94, 14.18, 20.69,
21.29, 22.34, 26.22, 27.22, 27.51, 27.65, 33.25, 34.61, 42.96,
46.44, 56.19, 60.84, 71.05, 74.13, 74.29, 74.73, 75.02, 78.80,
80.30, 82.77, 83.37, 83.55, 83.99, 127.90, 128.11, 128.50,
128.56, 128.99, 129.28, 129.64, 130.12, 131.51, 132.45, 133.77,
134.84, 137.78, 141.50, 151.74, 152.11, 152.82, 166.88, 168.30,
169.24, 170.12, 173.22, 202.08. HRFABMS: m/z [M + Na]+
1176.4796 (C62H75NO20Na requires 1176.4780).
2′-O,7-O,3′-N-Tr i(ter t-b u t oxyca r b on yl)-1,2-(r-ter t-b u -
toxyben zylid en ea ceta l)p a clita xel (5). To a stirred solution
of paclitaxel (1) (30 mg, 0.035 mmol) in dry acetonitrile (1.5
mL) was added di-tert-butyl dicarbonate (76.3 mg, 0.35 mmol)
under Ar. After the mixture stirred for 5 min, DMAP (17 mg)
was added at room temperature. The reaction mixture was
stirred for 4 h at room temperature and then diluted with 10
mL of EtOAc. The organic layer was washed with water and
brine, dried over Na2SO4, and evaporated to furnish a pale-
yellow residue, which was purified on PTLC (Analtech, 1000
µm; hexane/EtOAc, 3:1) to give 2′-O,7-O,3′-N-tri(tert-butoxy-
carbonyl)-1,2-(R-tert-butoxybenzylideneacetal)paclitaxel (5) (38
mg, 86%). 1H NMR: δ 0.95 (s 9H), 1.00 (s 9H), 1.16 (s 3H),
1.17 (s 3H), 1.28 (s 9H), 1.45 (s 9H), 1.51-1.64 (m, 2H), 1.66
(s 1H), 1.74 (s 3H), 1.82 (s 3H), 1.96-2.03 (m, 1H), 2.07 (s 3H),
Con clu sion s
We have evaluated a number of biological activities
of 50 analogues of paclitaxel modified at position C-2.
Enhanced activity was observed only in compounds that
retain a benzoyl group, provided they have a relatively
small meta-substituent or an o-azido group. No signifi-
cant improvement in activity was observed over the

Synthesis and Evaluation of Taxol 2-Acyl Analogues
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 19 3723
2.23 (s 3H), 2.68-2.76 (m, 1H), 3.42 (d, J ) 5.04 Hz, 1H), 4.26
(d, J ) 5.03 Hz, 1H), 4.73 (d, J ) 8.55 Hz, 1H), 4.76 (d, J )
8.55 Hz, 1H), 5.07 (d, J ) 8.85 Hz, 1H), 5.32-5.36 (dd, J )
7.63 and 7.33 Hz, 1H), 5.73 (t, J ) 8.54 Hz, 1H), 5.81 (d, J )
11.14 Hz, 1H), 5.91 (d, J ) 10.99 Hz, 1H), 6.45 (s 1H), 6.74-
6.77 (dd, J ) 7.48 and 6.87 Hz, 1H), 6.99-7.03 (dd, J ) 7.63
and 7.78 Hz, 2H), 7.34-7.38 (m, 5H), 7.44-7.49 (m, 3H), 7.57
(d, J ) 7.02 Hz, 4H). 13C NMR: δ 11.14, 14.41, 20.68, 20.80,
22.15, 25.74, 27.18, 27.44, 27.64, 30.30, 33.95, 34.04, 41.55,
43.62, 57.85, 71.42, 74.05, 74.49, 74.50, 76.38, 79.66, 82.77,
83.24, 83.43, 84.31, 86.36, 118.49, 126.40, 127.75, 127.80,
128.08, 128.30, 128.33, 128.84, 129.28, 131.43, 131.60, 134.09,
137.79, 141.72, 142.85, 151.67, 152.20, 152.89, 168.41, 169.13,
170.00, 173.07, 203.39. HRFABMS: m/z [M + H]+ 1210.5586
(C66H84NO20 requires 1210.5580).
in 0.5 mL of CH3CN was added to a solution of 2′-O-(tert-
butoxycarbonyl)-7-O-(triethylsilyl)paclitaxel (7) (92.5 mg, 0.09
mmol) in 0.5 mL of dry acetonitrile under Ar. After the
mixture stirred for 5 min at room temperature, DMAP (8 mg)
was added. The reaction mixture was stirred for 3 h at room
temperature and diluted with EtOAc. Solvents were removed
on a rotary evaporator. The residue was dissolved in EtOAc
and the solution washed sequentially with cold dilute HCl, cold
0.05 N NaHCO3, H2O, and brine and dried over Na2SO4. The
solvent was evaporated to yield crude product, which was
passed through a small silica gel column to yield pure 2′,N-
di-O-(tert-butoxycarbonyl)-7-O-(triethylsilyl)paclitaxel (8) (89
mg, 88%), Rf (hexane/EtOAc, 1:1) 0.55. 1H NMR: δ 0.50-0.58
(m, 6H), 0.90 (t, J ) 7.78 Hz, 9H), 1.04 (s 9H), 1.09 (s 3H),
1.14 (s 3H), 1.35 (s 9H), 1.47-1.51 (m, 1H), 1.63 (s 3H), 1.73
(s 1H), 1.79 (s 3H), 1.83-1.93 (m, 3H), 2.12 (s 3H), 2.38 (s 3H),
2.46-2.54 (m, 1H), 3.73 (d, J ) 7.17 Hz, 1H), 4.09 (d, J ) 8.39
Hz, 1H), 4.25 (d, J ) 8.40 Hz, 1H), 4.42-4.46 (m, 1H), 4.92 (d,
J ) 8.24 Hz, 1H), 5.59 (d, J ) 7.17 Hz, 1H), 5.87 (d, J ) 11.29
Hz, 1H), 5.96-6.01 (m, 2H), 7.16 (t, J ) 7.47 Hz, 1H), 7.31-
7.40 (m, 5H), 7.47-7.54 (m, 3H), 7.61-7.67 (m, 4H), 8.06 (d,
J ) 7.18 Hz, 2H). 13C NMR: δ 5.26, 6.70, 10.03, 13.96, 20.79,
21.15, 22.43, 26.41, 27.21, 27.50, 34.57, 37.10, 43.07, 46.59,
58.28, 60.68, 71.14, 72.06, 74.34, 74.89, 74.93, 78.78, 80.56,
83.32, 83.54, 84.30, 127.89, 128.11, 128.46, 128.51, 128.98,
129.37, 129.59, 130.10, 131.51, 133.12, 133.68, 134.78, 137.73,
140.71, 151.75, 152.86, 166.93, 169.07, 169.10, 170.19, 173.16,
201.89. HRFABMS: m/z [M + H]+ 1168.5282 (C63H82NO18
requires 1168.5301).
2′-O-(ter t-Bu toxyca r bon yl)p a clita xel (6). Paclitaxel (1)
(85.3 mg, 0.1 mmol) and acetonitrile (2 mL, freshly dried and
distilled over CaH2) were added to a flame-dried 25-mL round-
bottom flask under Ar. To this solution was added 21.8 mg
(0.1 mmol) of di-tert-butyl dicarbonate in 2.0 mL of dry CH3-
CN under Ar. After the mixture stirred for 5 min, DMAP (5
mg) was added at 0 °C. The reaction mixture was stirred for
2 h at room temperature and worked up by diluting with
EtOAc. Solvents were removed on a rotary evaporator. The
pale-yellow residue was dissolved in EtOAc and washed
sequentially with dilute HCl and cold 0.05 N NaHCO3. The
organic solution was washed with brine, dried over Na2SO4,
and evaporated to give 2′-O-(tert-butoxycarbonyl)paclitaxel (6)
(95 mg, 99.6%), Rf (hexane/EtOAc, 1:1) 0.36. 1H NMR: δ 1.12
(s 3H), 1.23 (s 3H), 1.45 (s 9H), 1.67 (s 3H), 1.85 (s 1H), 1.90
(s 3H), 2.16-2.22 (m, 1H), 2.21 (s 3H), 2.35-2.41 (m, 2H), 2.44
(s 3H), 2.51-2.55 (m, 2H), 3.80 (d, J ) 7.02 Hz, 1H), 4.18 (d,
J ) 8.21 Hz, 1H), 4.30 (d, J ) 8.54 Hz, 1H), 4.42 (m, 1H), 4.46
(d, J ) 9.46 Hz, 1H), 5.39 (d, J ) 2.9 Hz, 1H), 5.68 (d, J )
7.17 Hz, 1H), 5.92-5.95 (dd, J ) 2.44 and 9.3 Hz, 1H), 6.25 (t,
J ) 8.85 Hz, 1H), 6.28 (s 1H), 6.98 (d, J ) 9.31 Hz, 1H), 7.40-
7.43 (m, 7H), 7.47-7.50 (m, 3H), 7.52-7.61 (m, 1H), 7.73 (d,
J ) 7.02 Hz, 2H), 8.11 (d, J ) 7.02 Hz, 2H). 13C NMR: δ 9.58,
14.81, 20.81, 22.15, 22.66, 26.78, 27.58, 35.49, 35.57, 43.17,
45.53, 52.84, 58.47, 71.93, 72.09, 75.08, 75.59, 75.86, 76.41,
79.06, 81.01, 84.14, 84.42, 126.65, 127.16, 128.37, 128.63,
128.69, 129.00, 129.16, 130.19, 131.95, 132.68, 133.59, 133.64,
136.97, 142.83, 152.33, 166.96, 167.14, 168.33, 169.84, 171.25,
203.81. HRFABMS: m/z [M + H]+ 954.3908 (C52H60NO16
requires 954.3912).
2′,N-Di-O-(ter t-bu toxyca r bon yl)-2-d eben zoyl-7-O-(tr i-
eth ylsilyl)p a clita xel (9). A 0.1 N LiOH solution (0.45 mL)
was added to a stirred solution of 2′,7-di-O-(tert-butoxycarbo-
nyl)-7-O-(triethylsilyl)paclitaxel (8) (45 mg, 0.038 mmol) in 4.5
mL of THF at 0 °C. The ice bath was removed, and the
solution was stirred for 2 h at room temperature. TLC showed
the presence of two new spots at lower Rf in addition to starting
material. The reaction mixture was diluted with ether, and
the solution was washed with brine. The brine layer was
washed with additional ether, and the combined organic layer
was dried over Na2SO4 and evaporated. The crude product
was purified on PTLC (Analtech, 500 µm; hexane/EtOAc, 1:1).
The slower moving band was scraped and extracted to give
2′,N-di-O-(tert-butoxycarbonyl)-2-debenzoyl-7-O-(triethylsilyl)-
paclitaxel (9) (15.3 mg, 38%; 76% yield based on unrecovered
starting material). The faster moving band corresponded to
the rearranged paclitaxel 10. The product 9 had Rf (hexane/
EtOAc, 2:1) 0.21. 1H NMR: δ 0.59 (q, 6H), 0.92 (t, 9H), 1.18
(s 3H), 1.25 (s 3H), 1.30 (s 9H), 1.48 (s 9H), 1.65 (s 3H), 2.15
(s 3H), 2.16 (s 3H), 2.30 (s 3H), 2.55 (m, 2H), 3.37 (d, J ) 6.8
Hz, 1H), 3.86 (bt, J ) 4.0 Hz, 1H), 4.4 (dd, J ) 6.4, 3.7 Hz,
1H), 4.58 (bs, 2H), 4.95 (d, J ) 10.4 Hz, 1H), 5.95 (d, J ) 10.9
Hz, 1H), 6.0 (m, 1H), 6.08 (d, J ) 10.9 Hz, 1H), 6.32 (s 1H),
7.25-7.8 (m, 15H). MS: m/z 1064 (M + H+, 100%).
2′,N-Di-O-(ter t-bu toxyca r bon yl)-7-O-(tr ieth ylsilyl)p a -
clita xel (8) fr om 2′,N-Di-O-(ter t-bu toxyca r bon yl)-2-d e-
ben zoyl-7-O-(tr ieth ylsilyl)p a clita xel (9). A sample of 2′,N-
di-O-(tert-butoxycarbonyl)-2-debenzoyl-7-O-(triethylsilyl)-
paclitaxel (9) (2 mg, 0.0018 mmol) was treated with benzoic
acid (4.59 mg, 0.0338 mmol), DCC (7.75 mg, 0.038 mmol), and
a catalytic amount of PP in dry toluene (10 µL) under Ar. The
mixture was stirred overnight at 50 °C, and the solvent was
removed on a rotary evaporator. The crude reaction mixture
was purified by PTLC (500-µm layer; EtOAc/hexane, 1:2) to
yield 2′,N-di-O-(tert-butoxycarbonyl)-7-O-(triethylsilyl)pacli-
taxel (8) (1.5 mg, 68%), identical with material prepared
directly from paclitaxel.
2′-O-(ter t-Bu toxyca r bon yl)-7-O-(tr ieth ylsilyl)p a clita x-
el (7). Imidazole (34 mg, 5 mmol) followed by triethylsilyl
chloride (83.9 µL, 0.5 mmol) were slowly added to a stirred
solution of 2′-O-(tert-butoxycarbonyl)paclitaxel (6) (95.3 mg,
0.1 mmol) in 2 mL of dry DMF at 0 °C under Ar. The reaction
mixture was stirred for 3 h and quenched with EtOAc. The
organic layer was washed several times with H2O and brine
and dried with Na2SO4. The solvent was evaporated to obtain
pure 2′-O-(tert-butoxycarbonyl)-7-O-(triethylsilyl)paclitaxel (7)
(94.9 mg, 89%), Rf (hexane/EtOAc, 1:1) 0.66. 1H NMR: δ,
0.52-0.60 (m, 6H), 0.92 (t, J ) 15.9 Hz, 9H), 1.17 (s 3H), 1.21
(s 3H), 1.46 (s 9H), 1.68 (s 3H), 1.76 (s 1H), 1.85-1.91 (m, 1H),
2.03 (s 3H), 2.12-2.19 (m, 1H), 2.16 (s 3H), 2.35-2.53 (m, 2H),
2.44 (s 3H), 3.81 (d, J ) 7.17 Hz, 1H), 4.18 (d, J ) 8.4 Hz,
1H), 4.31 (d, J ) 8.2 Hz, 1H), 4.44-4.49 (dd, J ) 6.7 and 6.4
Hz, 1H), 4.93 (d, J ) 9.5 Hz, 1H), 5.42 (d, J ) 9.5 Hz, 1H),
5.69 (d, J ) 6.7 Hz, 1H), 5.94-5.97 (dd, J ) 2.7 and 9.5 Hz,
1H), 6.24 (t, J ) 9.15 Hz, 1H), 6.45 (s 1H), 6.95 (d, J ) 9.16
Hz, 1H), 7.32-7.47 (m, 7H), 7.48-7.51 (m, 3H), 7.57-7.61 (m,
1H), 7.74 (d, J ) 7.17 Hz, 2H), 8.12 (d, J ) 7.17 Hz, 2H). 13C
NMR: δ 5.26, 6.73, 10.09, 14.17, 20.84, 21.34, 22.70, 26.56,
27.60, 35.34, 37.19, 43.27, 46.68, 52.78, 58.39, 71.90, 72.15,
74.94, 75.00, 75.77, 76.51, 78.80, 81.04, 84.08, 84.22, 126.61,
127.16, 128.27, 128.62, 128.65, 128.95, 129.23, 130.18, 131.90,
133.46, 133.58, 137.01, 140.47, 152.29, 167.03, 168.27, 169.16,
Con ver sion of 2′,N-Di-O-(ter t-b u t oxyca r b on yl)-7-O-
(tr ieth ylsilyl)p a clita xel (8) to P a clita xel (1). 2′,N-Di-O-
(tert-butoxycarbonyl)-7-O-(triethylsilyl)paclitaxel (8) (9.5 mg)
was mixed with 99% formic acid (Fluka, 0.15 mL) and stirred
for 30 min at room temperature. The formic acid was removed
on a vacuum pump. The reaction mixture was diluted with
EtOAc, and the solution was washed with 5% NaHCO3, H2O,
and brine and dried. Solvents were removed by evaporation.
169.79, 201.82. HRFABMS: m/z [M + H]+ 1068.4747 (C58H74
NO16Si requires 1068.4776).
-
2′,N-Di-O-(ter t-bu toxyca r bon yl)-7-O-(tr ieth ylsilyl)p a -
clita xel (8). Di-tert-butyl dicarbonate (377.6 mg, 20 mmol)

3724 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 19
Kingston et al.
Purification of the residue by PTLC (EtOAc/hexanes, 1:1)
yielded paclitaxel (1) (2 mg, 28%), identical with an authentic
sample.
8.3 Hz, 1H), 4.48 (dd, J ) 9.4, 6.6 Hz, 1H), 4.67 (d, J ) 2.1 Hz,
1H), 4.94 (bd, J ) 8.8 Hz, 1H), 5.69 (d, J ) 7.0 Hz, 1H), 5.74
(dd, J ) 9.0, 2.1 Hz, 1H), 6.26 (bt, 1H), 6.45 (s 1H), 7.10 (d, J
) 8.9 Hz, 1H), 7.30-7.60 (m, 11H), 7.74 (dd, J ) 8.5, 1.5 Hz,
2H), 8.13 (dd, J ) 8.5, 1.4 Hz, 2H). 13C NMR: δ -5.86, -5.20,
5.27, 6.73, 10.11, 14.24, 18.11, 20.85, 21.50, 23.10, 25.49, 26.54,
35.55, 37.21, 43.31, 46.64, 55.63, 58.39, 71.36, 72.19, 74.92,
74.95, 75.10, 76.55, 78.82, 81.17, 84.22, 126.40, 126.97, 127.92,
128.68, 128.70, 128.71, 129.19, 130.20, 131.76, 133.60, 133.66,
134.03, 138.26, 140.14, 166.88, 167.03, 169.28, 170.13, 171.38,
201.67. FABMS: m/z (rel int.) [M + H]+ 1104 (5), 705 (3),
422 (40), 354 (12), 105 (100). HRFABMS: m/z [M + Na -
H]+ 1104.4936 (C59H79NO14Si2Na requires 1104.4937).
2′,7-Di-O-(tr ieth ylsilyl)p a clita xel (11a ). Imidazole (136.0
mg, 2.0 mmol) and triethylsilyl chloride (0.30 mL, 2.0 mmol)
were sequentially added to a solution of paclitaxel (1) (170.6
mg, 0.2 mmol) in dry DMF (1.0 mL). The reaction mixture
was stirred at room temperature for 2 h and diluted with
EtOAc. The solution was washed successively with water (20
mL × 2) and brine. The organic layer was separated, dried
over Na2SO4, and evaporated under reduced pressure to yield
a transparent syrup. The crude material was chromato-
graphed over silica gel using EtOAc/hexane (1:3) as eluent to
obtain 2′,7-di-O-(triethylsilyl)paclitaxel (11a ) as a white solid
(204 mg, 94%). Mp: 122-124 °C. 1H NMR: δ 0.45 (q, J )
7.8 Hz, 6H), 0.58 (q, J ) 7.8 Hz, 6H), 0.81 (t, J ) 7.8 Hz, 9H),
0.90 (t, J ) 7.8 Hz, 9H), 1.20 (s 3H), 1.25 (s 3H), 1.70 (s 3H),
2.05 (bs, 3H), 2.16 (s 3H), 2.32 (m, 1H), 2.50 (m, 1H), 2.54 (s
3H), 3.83 (d, J ) 7.0 Hz, 1H), 4.22 (d, J ) 8.0 Hz, 1H), 4.32 (d,
J ) 8.0 Hz, 1H), 4.48 (dd, J ) 10.6, 6.6 Hz, 1H), 4.70 (d, J )
2.1 Hz, 1H), 4.94 (bd, J ) 8.8 Hz, 1H), 5.70 (m, 2H), 6.24 (bt,
1H), 6.45 (s 1H), 7.10 (d, J ) 8.9 Hz, 1H), 7.30-7.60 (m, 11H),
7.74 (dd, J ) 8.5, 1.1 Hz, 2H), 8.13 (dd, J ) 8.5, 1.4 Hz, 2H).
13C NMR: δ 4.34, 5.28, 6.49, 6.72, 10.10, 14.15, 20.85, 21.46,
23.05, 26.53, 35.55, 37.21, 43.31, 46.64, 55.68, 58.39, 71.41,
72.19, 74.40, 74.92, 74.96, 76.55, 76.68, 76.99, 77.31, 78.82,
81.15, 84.22, 126.45, 127.01, 127.91, 128.65, 128.71, 129.20,
130.19, 131.71, 133.58, 133.65, 134.05, 138.41, 140.16, 166.90,
167.28, 170.06, 171.55, 201.67. FABMS: m/z [M + Na]+ 1104
(100).
2′-O-(ter t-Bu tyld im eth ylsilyl)-2-d eben zoyl-7-O-(tr ieth -
ylsilyl)p a clita xel (12b). Benzyltrimethylammonium hydrox-
ide (100 µL, 40% w/w solution in methanol) was added to a
solution of 2′-O-(tert-butyldimethylsilyl)-7-O-(triethylsilyl)pa-
clitaxel (11b) (110.4 mg, 0.1 mmol) in anhydrous CH2Cl2 (5
mL) at -78 °C. The reaction mixture was stirred for 10 min.
The reaction flask was transferred to a -10 °C bath (diethylene
glycol, dry ice) and stirred for 10-15 min. During this time
the progress of the reaction was monitored by TLC, which
indicated the formation of a more polar compound. The
mixture was diluted with cold CH2Cl2 (-78 °C, 10 mL) and
quenched with 5 mL of 0.1 N HCl. The organic layer was
removed, washed successively with water, dilute NaHCO3
solution, and brine, and dried over Na2SO4. Concentration
under reduced pressure yielded crude residue, which was
purified by PTLC (silica gel, 1000 µm; EtOAc/hexanes, 2:3) to
yield 2′-O-(tert-butyldimethylsilyl)-2-debenzoyl-7-O-(triethyl-
silyl)paclitaxel (12b) (73.0 mg, 73%; 81% based on unrecovered
starting material) and starting compound (11.0 mg, 10%).
Mp: 133-135 °C. 1H NMR: δ -0.28 (s 3H), -0.04 (s 3H),
0.57 (q, J ) 7.8 Hz, 6H), 0.80 (s 9H), 0.92 (t, J ) 7.8 Hz, 9H),
1.07 (s 3H), 1.15 (s 3H), 1.60-1.62 (m, 1H), 1.62 (s 3H), 1.96
(bs, 3H), 2.11-2.13 (m, 1H), 2.14 (s 3H), 2.40 (m, 1H), 2.42 (s
3H), 2.51 (m, 1H), 2.64 (br s 1H), 2.79 (d, J ) 5.9 Hz, 1H),
3.46 (d, J ) 7.0 Hz, 1H), 3.92 (t, J ) 6.3 Hz, 1H), 4.41 (dd, J
) 9.4, 6.6 Hz, 1H), 4.60 (d, J ) 1.6 Hz, 1H), 4.63 (bs, 2H), 4.95
(bd, J ) 8.6 Hz, 1H), 5.67 (dd, J ) 9.2, 1.6 Hz, 1H), 6.24 (bt,
1H), 6.37 (s 1H), 7.05 (d, J ) 9.2 Hz, 1H), 7.30-7.54 (m, 8H),
7.74 (dd, J ) 8.3, 1.5 Hz, 2H). FABMS: m/z (rel int.) [M +
Na]+ 1000 (3), 400 (22), 354 (84), 105 (100). 13C NMR: δ -5.75,
-5.19, 5.24, 6.72, 10.25, 14.12, 18.14, 20.81, 21,42, 23.22, 25.52,
26.31, 35.40, 37.21, 43.03, 46.64, 55.33, 58.24, 71.97, 72.18,
73.85, 74.76, 75.10, 78.09, 78.21, 82.36, 83.93, 126.43, 126.89,
128.05, 128.71, 128.82, 131.94, 133.90, 134.12, 137.92, 139.55,
167.73, 169.23, 169.90, 171.38, 202.45. HRFABMS: m/z [M
+ Na]+ 1000.4632 (C59H75NO14Si2Na requires 1000.4624).
Gen er a l P r oced u r e for th e Rea ction of 2-Deben zoyl-
2′,7-d i-O-(tr ieth ylsilyl)p a clita xel (12a ) w ith Ca r boxylic
Acid s. An appropriate carboxylic acid (0.1 mmol) was dis-
solved in dry toluene (0.2 mL), and DCC (20.6 mg, 0.1 mmol)
and PP (1.0 mg, 0.006 mmol) were added to the solution. The
mixture was stirred at room temperature for 5 min, compound
12a (10.0 mg, 0.01 mmol) was added, and the reaction mixture
was stirred at room temperature until the 12a was consumed
(TLC analysis). The reaction mixture was diluted with EtOAc
(10 mL) and filtered through a pad of silica gel and Celite,
which was washed with additional EtOAc (10 mL). The EtOAc
filtrate was concentrated on a rotary evaporator. Purification
using preparative TLC (silica gel, 500 µm, hexane/EtOAc, 2:1)
yielded 2-acyl-2-debenzoyl-2′,7-di-O-(triethylsilyl)paclitaxels
13a (60-90% yield).
Reaction of 2′,7-Di-O-(tr ieth ylsilyl)paclitaxel (11a) with
Na OH/P TC: 2-Deben zoyl-2′,7-d i-O-(tr ieth ylsilyl)p a clita x-
el (12a ). 2′,7-Di-O-(triethylsilyl)paclitaxel (11a ) (216 mg, 0.20
mmol) was dissolved in benzene/CH2Cl2 (8:2, 20 mL). Tet-
rabutylammonium hydrogen sulfate (800 mg, 2.35 mmol) and
2 N NaOH (20 mL) were added sequentially. The mixture was
stirred at room temperature, and the reaction was monitored
by TLC. After 1 h TLC showed the presence of a new low Rf
spot (0.3) along with starting material (Rf 0.75). After 2 h the
reaction mixture was diluted with benzene (25 mL), and the
solution was washed sequentially with water (2 × 25 mL) and
brine. The organic layer was separated, dried over Na2SO4,
and evaporated. Purification by preparative TLC (silica gel,
1000 µm; hexane/EtOAc, 1:1) yielded 2-debenzoyl-2′,7-di-O-
(triethylsilyl)paclitaxel (12a ) (70.0 mg, 35.8%), 7-(triethylsilyl)-
baccatin (15.0 mg, 10%), and starting material (110 mg, 51%).
1H NMR of 12a : δ 0.48 (q, J ) 7.8 Hz, 6H), 0.58 (q, J ) 7.8
Hz, 6H), 0.81 (t, J ) 7.8 Hz, 9H), 0.90 (t, J ) 7.8 Hz, 9H), 1.05
(s 3H), 1.10 (s 3H), 1.51 (s 3H), 1.95 (bs, 3H), 2.15 (s 3H), 2.40
(s 3H), 2.50 (m, 1H), 3.25 (bs, 1H), 3.49 (d, J ) 6.8 Hz), 3.92
(t, J ) 5.8 Hz, 1H), 4.42 (dd, J ) 10.6, 6.7 Hz, 1H), 4.62 (bs,
3H), 4.98 (bd, J ) 8.8 Hz), 5.63 (bd, J ) 9.5 Hz, 1H), 6.20 (bt,
1H), 6.37 (s 1H), 7.08 (d, J ) 9.5 Hz, 1H), 7.15-7.50 (m, 8H),
7.78 (dd, J ) 8.5, 1.3 Hz, 2H).
2′-O-(ter t-Bu tyld im eth ylsilyl)-7-O-(tr ieth ylsilyl)p a cli-
ta xel (11b). Imidazole (107 mg, 1.58 mmol) and tert-bu-
tyldimethylsilyl chloride (238 mg, 1.58 mmol) were added to
a stirred solution of paclitaxel (270 mg, 0.316 mmol) in 2.5
mL of anhydrous DMF. The solution was heated at 60 °C for
2 h. The mixture was cooled to room temperature, and
imidazole (107 mg, 1.58 mmol) and triethylsilyl chloride (150
µL, 1.34 mmol) were added. After stirring at room tempera-
ture for 1 h the reaction mixture was diluted with EtOAc, and
the solution was washed successively with H2O and brine. The
organic layer was dried over Na2SO4, and evaporation of
solvents under reduced pressure yielded crude material.
Purification by column chromatography over silica gel (EtOAc/
hexanes, 1:2) yielded 2′-O-(tert-butyldimethylsilyl)-7-O-(tri-
ethylsilyl)paclitaxel (11b) (325 mg, 95%) as an amorphous
solid. 1H NMR: δ -0.20 (s 3H), -0.02 (s 3H), 0.62 (q, J ) 7.8
Hz, 6H), 0.79 (s 9H), 0.92 (t, J ) 7.8 Hz, 9H), 1.17 (s 3H), 1.21
(s 3H), 1.70 (s 3H), 1.88-1.92 (m, 1H), 2.02 (bs, 3H), 2.09-2.2
(m, 1H), 2.16 (s 3H), 2.40 (m, 1H), 2.55 (m, 1H), 2.58 (s 3H),
3.83 (d, J ) 7.0 Hz, 1H), 4.19 (d, J ) 8.3 Hz, 1H), 4.30 (d, J )
Gen er a l P r oced u r e for th e Rea ction of 2′-O-(ter t-
Bu tyld im eth ylsilyl)-2-d eben zoyl-7-O-(tr ieth ylsilyl)p a cli-
ta xel (12b) w ith Ca r boxylic Acid s. An appropriate car-
boxylic acid (0.1 mmol) in dry toluene (0.2 mL) was treated
with DCC (20.6 mg, 0.1 mmol) and PP (1.0 mg, 0.006 mmol).
The mixture was stirred at room temperature for 5 min.
Compound 12b (10.0 mg, 0.01 mmol) was added, and the
reaction mixture was either stirred at room temperature or
heated on an oil bath at 60 °C until the starting compound
was consumed (TLC analysis, generally 4-16 h). The reaction

Synthesis and Evaluation of Taxol 2-Acyl Analogues
J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 19 3725
mixture was diluted with EtOAc (10 mL) and filtered through
55131, and CA 69571) and from Bristol-Myers Squibb
Pharmaceutical Research Institute (all to D.G.I.K.). We
thank Dr. Kenneth Snader and the Natural Products
Branch of the National Cancer Institute for a supply of
partially purified extracts of T. brevifolia that were used
in the earlier stages of this work as sources of paclitaxel.
a
pad of silica gel and Celite, which was washed with
additional EtOAc (10 mL). The EtOAc filtrate was concen-
trated on a rotary evaporator. Purification using preparative
TLC (silica gel, 500 m; hexane/EtOAc, 2:1) yielded 2-acyl-2′-
O-(tert-butyldimethylsilyl)-7-O-(triethylsilyl)paclitaxels 13b (60-
90% yield). Example: reaction of 3,5-dichlorobenzoic acid (60.0
mg, 0.31 mmol, 10 equiv), DCC (65 mg, 0.31 mmol, 10 equiv),
and catalytic PP (3.0 mg) with 12b (30 mg, 0.03 mmol) in dry
toluene (0.4 mL) at 60 °C for 16 h, followed by workup and
isolation as described, yielded 2-(3,5-dichlorobenzoyl)-2-deben-
zoyl-2′-O-(tert-butyldimethylsilyl)-7-O-(triethylsilyl)paclitax-
el (27.0 mg, 76%).
Dep r otection of Silyl Gr ou p s w ith Meth a n olic HCl. A
2-acyl-2-debenzoyl-2′,7-di-O-(triethylsilyl)paclitaxel (13a ) or a
2-acyl-2′-O-(tert-butyldimethylsilyl)-7-O-(triethylsilyl)paclitax-
el (13b) (0.005 mmol) was dissolved in freshly prepared
methanolic HCl (5%, v/v) solution. The mixture was stirred
at room temperature for 30-60 min and diluted with EtOAc
(10 mL). The solution was washed successively with dilute
NaHCO3 solution, H2O, and brine. The organic layer was dried
over Na2SO4 and concentrated to give crude product. Purifica-
tion was carried out using preparative TLC to give homoge-
neous product in 70-90% yield. Example: reaction of 2-de-
benzoyl-2-(3′,5′-dichlorobenzoyl)-2′-O-(tert-butyldimethylsilyl)-
7-O-(triethylsilyl)paclitaxel (25.0 mg) with freshly prepared
methanolic HCl (10%, v/v) at room temperature for 35 min as
described gave 2-(3,5-dichlorobenzoyl)-2-debenzoylpaclitaxel
(14d m ) (17.2 mg, 86%). 1H NMR: δ, 1.13 (s 3H), 1.23 (s 3H),
1.67 (s 3H), 1.78 (s 3H), 1.89 (m, 1H), 2.28 (m, 2H), 2.23 (s
3H), 2.33 (s 3H), 2.45 (d, J ) 3.4 Hz, 1H), 2.57 (m, 1H), 3.66
(bs, 1H), 3.79 (d, J ) 7.1 Hz, 1H), 4.14 (d, J ) 8.2, 1H), 4.40
(d, J ) 8.2 Hz, 1H), 4.43 (bt, 1H), 4.75 (bs, 1H), 4.96 (d, J )
8.0 Hz, 1H), 5.58 (d, J ) 7.1 Hz, 1H), 5.73 (dd, J ) 9.2, 1.6 Hz,
1H), 6.17 (bt, 1H), 6.26 (s 1H), 6.96 (d, J ) 9.2 Hz, 1H), 7.30-
7.74 (m, 9H), 7.74 (d, J ) 8.3 Hz, 2H) 8.00 (d, J ) 1.7 Hz, 2H).
FABMS: m/z (rel int.) [M + H]+ 922. HRFABMS: m/z [M +
H]+ 922.2592 (C47H50NO14Cl2 requires 922.2608).
Su p p or tin g In for m a tion Ava ila ble: Tables of bioactiv-
ity, 1H NMR, and MS data (4 pages). Ordering information
is given on any current masthead page.
Refer en ces
(1) Wani, M. C.; Taylor, H. L.; Wall, M. E.; Coggon, P.; McPhail, A.
T. Plant Antitumor Agents. VI. The Isolation and Structure of
Taxol, a Novel Antileukemic and Antitumor Agent from Taxus
brevifolia. J . Am. Chem. Soc. 1971, 93, 2325-2327.
(2) (a) Schiff, P. B.; Fant, J .; Horwitz, S. B. Promotion of Microtubule
Assembly in Vitro by Taxol. Nature 1979, 277, 665-667. (b)
Manfredi, J . J .; Horwitz, S. B. Taxol: An Antimitotic Agent with
a New Mechanism of Action. Pharmacol. Ther. 1984, 25, 83-
125.
(3) (a) Arbuck, S. G.; Blaylock, B. A. Taxol: Clinical Results and
Current Issues in Development. In Taxol: Science and Applica-
tions; Suffness, M., Ed.; CRC Press: Boca Raton, FL, 1995; pp
379-415. (b) Holmes, F. A.; Kudelka, A. P.; Kavanagh, J . J .;
Huber, M. H.; Ajani, J . A.; Valero, V. Current Status of Clinical
Trials with Paclitaxel and Docetaxel. In 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, 1994; pp 31-57.
(c) Donehower, R. C.; Rowinsky, E. K. An Overview of Experience
with Taxol (Paclitaxel) in the U.S.A. Cancer. Treat. Rev. 1993,
19 (Suppl.), 63-78. (d) Rowinsky, E. K.; Donehower, R. C.
Paclitaxel (Taxol). N. Engl. J . Med. 1995, 1004-1014.
(4) (a) Chen, S.-H.; Farina, V. Paclitaxel Structure-Activity Rela-
tionships and Core Skeletal Rearrangements. In Taxane Anti-
cancer 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, 1994; pp
247-261. (b) Georg, G. I.; Boge, T. C.; Cheruvallath, Z. S.;
Clowers, J . S.; Harriman, G. C. B.; Hepperle, M.; Park, H. The
Medicinal Chemistry of Taxol. In Taxol: Science and Applica-
tions; Suffness, M., Ed.; CRC Press: Boca Raton, FL, 1995; pp
317-375. (c) Kingston, D. G. I. The Chemistry of Taxol. Phar-
macol. Ther. 1991, 52, 1-34. (d) Kingston, D. G. I. Taxol: the
Chemistry and Structure-Activity Relationships of a Novel
Anticancer Agent. Trends Biotechnol. 1994, 12, 222-227. (e)
Kingston, D. G. I. Recent Advances in the Chemistry and
Structure-Activity Relationships of Paclitaxel. In Taxane An-
ticancer 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, 1994; pp
203-216. (f) Kingston, D. G. I. History and Chemistry. In
Paclitaxel in Cancer Treatment; McGuire, W. P., Rowinsky, E.
K., Eds.; Marcel Dekker: New York, Basel, Hong Kong, 1995;
Vol. 8, pp 1-33. (g) Nicolaou, K. C.; Dai, W.-M.; Guy, R. K.
Chemistry and Biology of Taxol. Angew. Chem., Int. Ed. Engl.
1994, 33, 15-44.
(5) Samaranayake, G.; Magri, N. F.; J itrangsri, C.; Kingston, D. G.
I. Modified Taxols. 5. Reaction of Taxol with Electrophilic
Reagents and Preparation of a Rearranged Taxol Derivative with
Tubulin Assembly Activity. J . Org. Chem. 1991, 56, 5114-5119.
(6) Neidigh, K. A.; Gharpure, M. M.; Rimoldi, J . M.; Kingston, D.
G. I.; J iang, Y. Q.; Hamel, E. Synthesis and Biological Evaluation
of 4-Deacetylpaclitaxel. Tetrahedron Lett. 1994, 35, 6839-6842.
(b) Chordia, M. D.; Chaudhary, A. G.; Kingston, D. G. I.; J iang,
Y. Q.; Hamel, E. Synthesis and Biological Evaluation of 4-Deac-
etoxypaclitaxel. Tetrahedron Lett. 1994, 35, 6843-6846. (c)
Datta, A.; J ayasinghe, L. R.; Georg, G. I. 4-Deacetyltaxol and
10-Acetyl-4-deacetyltaxotere: Synthesis and Biological Evalu-
ation. J . Med. Chem. 1994, 37, 4258-4260.
(7) (a) Chen, S.-H.; Kadow, J . F.; Farina, V.; Fairchild, C. R.;
J ohnston, K. A. First Syntheses of Novel Paclitaxel (Taxol)
Analogues Modified at the C4-Position. J . Org. Chem. 1994,
59, 6156-6158. (b) Chen, S.-H.; Fairchild, C.; Long, B. H.
Synthesis and Biological Evaluation of Novel C-4 Aziridine
Bearing Paclitaxel (Taxol) Analogues. J . Med. Chem. 1995, 38,
2263-2267. (c) Chen, S.-H.; Wei, J .-M.; Long, B. H.; Fairchild,
C. A.; Carboni, J .; Mamber, S. W.; Rose, W. C.; J ohnston, K.;
Casazza, A. M.; Kadow, J . F.; Farina, V.; Vyas, D.; Doyle, T. W.
Novel C-4 Paclitaxel (Taxol) Analogues: Potent Antitumor
Agents. Bioorg. Med. Chem. Lett. 1995, 5, 2741-2746. (d) Georg,
G. I.; Ali, S. M.; Boge, T. C.; Data, A.; Falborg, L.; Himes, R. H.
Selective C-2 and C-4 Deacylation and Acylation of Taxol: The
First Synthesis of a C-4 Substituted Taxol Analogue. Tetrahe-
dron Lett. 1994, 35, 8931-8934.
Syn th esis of Com p ou n d s 14a a -eg. Compounds 14a a -
eg were prepared from paclitaxel by either the PTC method
or the Triton B method described above, using the appropriate
carboxylic acid in the acylation step. The final products were
purified by PTLC, and product characterization was carried
out by 1H NMR spectroscopy, confirmed by HRMS in most
cases. Structural assignment as the paclitaxel analogue 14
rather than an isopaclitaxel analogue such as 15 or a depro-
tected version was based on the signal for H-2 at about 5.6-
5.7 ppm; in the isopaclitaxel analogues this signal occurs at
about 4.0 ppm. All products were at least 95% pure as
1
indicated by their H NMR spectra. Since the chemical shifts
for most protons are essentially unchanged from those of
paclitaxel, only the chemical shifts of those protons (AcO-4,
H-2, H-3, and H2-20) which showed significant variations from
compound to compound, together with the chemical shifts of
the substituent protons, are listed. 1H NMR and, in most
cases, HRMS data for compounds 14a a -eg are summarized
in Tables 3 and 4 in the Supporting Information.
Biologica l Assa ys. The assay for drug enhancement of
tubulin polymerization was performed as described previ-
ously.16 Two series of experiments were performed, the first
with 0.4 M glutamate (all compounds) and the second with
0.8 M glutamate (compounds less active than paclitaxel in the
first experiment). Electrophoretically homogeneous bovine
brain tubulin was prepared as described previously.24 Human
Burkitt lymphoma CA46 cells (a gift of Dr. Patrick O’Connor,
National Cancer Institute) were grown in suspension culture
as described previously.25 The ovarian carcinoma lines were
as described previously.20,21 They were grown in monolayer
cultures in microtiter plates, and IC50 values were determined
as described by Skehan et al.26
Ack n ow led gm en t. We are grateful for the financial
support for this work, which was provided by grants
from the National Cancer Institute (CA 48974, CA

3726 J ournal of Medicinal Chemistry, 1998, Vol. 41, No. 19
Kingston et al.
(8) Chaudhary, A. G.; Gharpure, M. M.; Rimoldi, J . M.; Chordia,
M. D.; Gunatilaka, A. A. L.; Kingston, D. G. I.; Grover, S.; Lin,
C. M.; Hamel, E. Unexpectedly Facile Hydrolysis of the 2-Ben-
zoate Group of Taxol and Synthesis of Analogues with Increased
Activities. J . Am. Chem. Soc. 1994, 116, 4097-4098.
5919. (b) Ojima, I.; Habus, I.; Zhao, M.; Zucco, M.; Park, Y. H.;
Sun, C. M.; Brigaud, T. New and efficient approaches to the
semisynthesis of taxol and its C-13 side chain analogues by
means of â-lactam synthon method. Tetrahedron 1992, 48,
6985-7012. (c) Holton, R. A.; Biediger, R. J .; Boatman, P. D.
Semisynthesis of Taxol and Taxotere. In Taxol: Science and
Applications; Suffness, M., Ed.; CRC Press: Boca Raton, FL,
1995; pp 97-121.
(9) (a) Georg, G. I.; Harriman, G. C. B.; Ali, S. M.; Datta, A.;
Hepperle, M.; Himes, R. H. Synthesis of 2-O-Heteroaroyl
Taxanes: Evaluation of Microtubule Assembly Promotion and
Cytotoxicity. Bioorg. Med. Chem. Lett. 1995, 5, 115-118. (b)
Georg, G. I.; Ali, S. M.; Boge, T. C.; Datta, A.; Falborg, L.; Park,
H.; Mejillano, M.; Himes, R. H. Synthesis of Biologically Active
2-Benzoyl Paclitaxel Analogues. Bioorg. Med. Chem. Lett. 1995,
5, 259-264. (c) Nicolaou, K. C.; Couladouros, E. A.; Nantermet,
P. G.; Renaud, J .; Guy, R. K.; Wrasidlo, W. Synthesis of C-2
Taxol Analogues, Angew. Chem., Int. Ed. Engl. 1994, 33, 1581-
1583. (d) Nicolaou, K. C.; Renaud, J .; Nantermet, P. G.; Coula-
douros, E. A.; Guy, R. K.; Wrasidlo, W. Chemical Synthesis and
Biological Evaluation of C-2 Taxoids. J . Am. Chem. Soc. 1995,
117, 2409-2420. (e) Chen, S.; Farina, V.; Wei, J .; Long, B.;
Fairchild, C.; Mamber, S. W.; Kadow, J . F.; Vyas, D.; Doyle, T.
W. Structure-Activity Relationships of Taxol: Synthesis and
Biological Evaluation of C-2 Taxol Analogues. Bioorg. Med.
Chem. Lett. 1994, 4, 479-482. (f) Grover, S.; Rimoldi, J . M.;
Molinero, A. A.; Chaudhary, A. G.; Kingston, D. G. I.; Hamel,
E. Differential Effects of Paclitaxel (Taxol) Analogues Modified
at Positions C-2, C-7, and C-3′ on Tubulin Polymerization and
Polymer Stabilization: Identification of a Hyperactive Paclitaxel
Derivative. Biochemistry 1995, 34, 3927-3934. (g) Ojima, I.;
Duclos, O.; Zucco, M.; Lavelle, F. Synthesis and Structure-
Activity Relationships of New Antitumor Taxoids. Effects of
Cyclohexyl Substitution at the C-3′ and/or C-2 of Taxotere
(Docetaxel). J . Med. Chem. 1994, 37, 2602-2608. (h) Holton, R.
A.; Kim, S. Process for the Preparation of Baccatin III Analogues
Bearing New C-2 and C-4 Functional Groups. U.S. Patent 5,-
399,726, March 21, 1995. (i) Ojima, I.; Kuduk, S. D.; Pera, P.;
Veith, J . M.; Bernacki, R. J . Syntheses and Structure-Activity
Relationships of Nonaromatic Taxoids: Effects of Alkyl and
Alkenyl Ester Groups on Cytotoxicity. J . Med. Chem. 1997, 40,
279-285.
(16) Lin, C. M.; J iang, Y. Q.; Chaudhary, A. G.; Rimoldi, J . M.;
Kingston, D. G. I.; Hamel, E. A Convenient Tubulin-Based
Quantitative Assay for Paclitaxel (Taxol) Derivatives More
Effective in Inducing Assembly than the Parent Compound.
Cancer Chemother. Pharmacol. 1996, 38, 136-140.
(17) IC50 is enclosed in quotation marks at its first appearance
because taxoids actually enhance tubulin polymerization. In the
assay used here, the drug concentration required to reduce the
protein in the supernatant by 50% was determined, since the
drug-induced polymer is removed from the reaction mixture by
centrifugation.
(18) Hamel, E.; del Campo, A. A.; Lowe, M. C.; Lin, C. M. Interactions
of Taxol, Microtubule-Associated Proteins, and Guanine Nucle-
otides in Tubulin Polymerization. J . Biol. Chem. 1981, 256,
11887-11894.
(19) Rimoldi, J . M.; Kingston, D. G. I.; Chaudhary, A. K.; Sama-
ranayake, G.; Grover, S.; Hamel, E. Modified Taxols, 9. Synthesis
and Biological Evaluation of 7-Substituted Photoaffinity Ana-
logues of Taxol. J . Nat. Prod. 1993, 56, 1313-1330.
(20) Behrens, B. C.; Hamilton, T. C.; Masuda, H.; Grotzinger, K. R.;
Whang-Peng, J .; Louie, K. G.; Knutsen, T.; McKoy, W. M.;
Young, R. C.; Ozols, R. F. Characterization of a cis-Diaminedichlo-
roplatinum(II)-Resistant Human Ovarian Cancer Cell Line and
Its Use in Evaluation of Platinum Analogues. Cancer Res. 1987,
47, 414-418.
(21) Giannakakou, P.; Sackett, D. L.; Kang, Y.-K.; Zhan, Z.; Buters,
J . T. M.; Fojo, T.; Poruchynsky, M. S. Paclitaxel resistant human
ovarian cancer cells have mutant â-tubulins that exhibit im-
paired paclitaxel driven polymerization. J . Biol. Chem. 1997,
272, 17118-17125.
(22) Hansch, C.; Leo, A.; Unger, S. H.; Kim, K. H.; Nikaitani, D.;
Lien, E. J . “Aromatic” Substituent Constants for Structure-
Activity Correlations. J . Med. Chem. 1973, 16, 1207-1216.
(23) The reader is reminded that only differences in IC50 values
greater than 20% are considered significant.
(24) Hamel, E.; Lin, C. M. Separation of Active Tubulin and Micro-
tubule-Associated Proteins by Ultracentrifugation and Isolation
of a Component Causing the Formation of Microtubule Bundles.
Biochemistry 1984, 23, 4173-4184.
(10) (a) Vander Velde, D. G.; Georg, G. I.; Grunewald, G. L.; Gunn,
C. W.; Mitscher, L. A. “Hydrophobic Collapse” of Taxol and
Taxotere Solution Conformations in Mixtures of Water and
Organic Solvent. J . Am. Chem. Soc. 1993, 115, 11650-11651.
(b) Williams, H. J .; Scott, A. I.; Dieden, R. A.; Swindell, C. S.;
Chirlian, L. E.; Francl, M. M.; Heerding, J . M.; Krauss, N. E.
NMR and Molecular Modeling Study of Active and Inactive Taxol
Analogues in Aqueous and Nonaqueous Solution. Can. J . Chem.
1994, 72, 252-260. (c) Ojima, I.; Kuduk, S. D.; Chakravarty,
S.; Ourevitch, M.; Be´gue´, J .-P. A Novel Approach to the Study
of Solution Structures and Dynamic Behavior of Paclitaxel and
Docetaxel Using Fluorine-Containing Analogues as Probes. J .
Am. Chem. Soc. 1997, 119, 5519-5527.
(25) ter Haar, E.; Kowalski, R. J .; Hamel, E.; Lin, C. M.; Longley, R.
E.; Gunasekera, S. P.; Rosenkranz, H. S.; Day, B. W. Discoder-
molide, a Cytotoxic Marine Agent that Stabilizes Microtubules
More Potently than Taxol. Biochemistry 1996, 35, 243-250.
(26) Skehan, P.; Storeng, R.; Scudiero, D.; Monks, A.; McMahon, J .;
Vistica, D.; Warren, J . T.; Bokesch, H.; Kenney, S.; Boyd, M. R.
New Colorimetric Cytotoxicity Assay for Anticancer-Drug Screen-
ing. J . Natl. Cancer Inst. 1990, 82, 1107-1112.
(27) Williams, R. C., J r.; Lee, J . C. Preparation of tubulin from Brain.
Methods Enzymol. 1982, 85, 376-385.
(28) Swindell, C. S.; Krauss, N. E.; Horwitz, S. B.; Ringel, I.
Biologically Active Taxol Analogues with Deleted A-Ring Side
Chain Substituents and Variable C-2′ Configurations. J . Med.
Chem. 1991, 34, 1176-1184.
(11) Miller, R. W.; Powell, R. G.; Smith, C. R., J r.; Arnold, E.; Clardy,
J . Antileukemic Alkaloids from Taxus wallichiana Zucc. J . Org.
Chem. 1981, 46, 1469-1474.
(12) Samaranayake, G.; Neidigh, K. A.; Kingston, D. G. I. Modified
Taxols. 8. Deacylation and Reacylation of Baccatin III. J . Nat.
Prod. 1993, 56, 884-898.
(13) Flynn, D. L.; Zelle, R. E.; Grieco, P. A. A Mild Two-Step Method
for the Hydrolysis/Methanolysis of Secondary Amides and Lac-
tams. J . Org. Chem. 1983, 48, 2424-2426.
(14) Farina, V.; Huang, S. The Chemistry of Taxanes: Unexpected
Rearrangement of Baccatin III During Chemoselective Deben-
zoylation with Bu3SnOMe/LiCl. Tetrahedron Lett. 1992, 33,
3979-3982.
(15) (a) Denis, J .-N.; Greene, A. E.; Guenard, D.; Gueritte-Voegelein,
F.; Mangatal, L.; Potier, P. A Highly Efficient, Practical Ap-
proach, to Natural Taxol. J . Am. Chem. Soc. 1988, 110, 5917-
J M980229D