The discovery of BMS-275183: An orally efficacious novel taxane

Article abstract of DOI:10.1016/S0968-0896(03)00495-4

The evolution of 2, a C-4-methylcarbonate analogue of paclitaxel with minimal oral bioavailability and oral efficacy, into its C-3′ -t-butyl-3′-N-t-butyloxycarbonyl analogue (15i), a novel taxane with oral efficacy in preclinical models that is comparable to iv administered paclitaxel, is described.

Full text of DOI:10.1016/S0968-0896(03)00495-4

Bioorganic & Medicinal Chemistry 11 (2003) 4315–4323
The Discovery of BMS-275183: An Orally Efficacious Novel
Taxane
Harold Mastalerz,a,* Donald Cook,a Craig R. Fairchild,b Steven Hansel,c
Walter Johnson,a John F. Kadow,a Byron H. Long,b William C. Rose,b James Tarrant,a
Mu-Jen Wu,a May Quifen Xue,a Guifen Zhang,a Mary Zoecklerc and Dolatrai M. Vyasa
aDiscovery Chemistry, Bristol-Myers Squibb Pharmaceutical Research Institute, 5 Research Parkway, PO Box5100,
Wallingford, CT 06492-7660, USA
bOncology Drug Discovery, Bristol-Myers Squibb Pharmaceutical Research Institute, PO Box4000,
Princeton, NJ 08543-4000, USA
cDiscovery Metabolism and Pharmacokinetics, Bristol-Myers Squibb Pharmaceutical, Research Institute,
5 Research Parkway, PO Box5100, Wallingford, CT 06492-7660, USA
Received 12 March 2003; revised 16 July 2003; accepted 25 July 2003
Abstract—The evolution of 2, a C-4-methylcarbonate analogue of paclitaxel with minimal oral bioavailability and oral efficacy, into
its C-30-t-butyl-30-N-t-butyloxycarbonyl analogue (15i), a novel taxane with oral efficacy in preclinical models that is comparable to
iv administered paclitaxel, is described.
# 2003 Elsevier Ltd. All rights reserved.
Introduction
efflux mediated by P-gycloprotein (P-gp).8 In spite of
these limitations, recent efforts to identify an orally
active taxane have been encouraging. A water-soluble
prodrug of a paclitaxel analogue has shown oral efficacy
in preclinical in vivo models.5 Presumably, the improved
solubility provides higher drug/prodrug concentrations
in the gastrointestinal tract and thereby improves pas-
sive absorption. Co-administration of paclitaxel with P-
gp blockers9 such as cyclosporin A have been found to
significantly increase the oral bioavailability of pacli-
taxel. Analogues of 14b-hydroxy-10-deacetylbaccatin
show increased efficacy against cancer cell lines that
overexpress P-gp and have been reported to exhibit
good oral bioavailability and in vivo efficacy.10 It was
suggested that these analogues show better oral absorp-
tion because they are no longer good substrates for P-gp
mediated efflux in the gastrointestinal tract.
Paclitaxel (1), the diterpenoid natural product isolated
from Taxus brevifolia in 1971 by Wani and Wall1 is the
active constituent of the antitumor drug Taxol1. It is a
schedule dependent drug whose benefits are obtained
through prolonged tumor exposure times.2 Recently,
clinical utility has been demonstrated using repetitive,
once-weekly, administrations of moderate (less than
maximally tolerated) doses of paclitaxel.3 An oral tax-
ane would be ideal for such protracted regimens, pro-
viding
a
compliant and cost-effective way of
accomplishing an extended duration of exposure.
Unfortunately, the oral bioavailability (%F) of pacli-
taxel is very low4 and it does not show oral efficacy in
preclinical models.5 This is not surprising when the
molecule is viewed in terms of features that are con-
sidered predictive of ADME properties.6,7 Its molecular
weight of 854 gm/mole is high, and it carries 15 het-
eroatoms that can function as H-bond acceptors. Pacli-
taxel possesses poor aqueous solubility and is subject to
Our approach to identifying an orally active taxane was
largely empirical. It took advantage of our extensive
*Corresponding author. Tel.:+1-203-677-7773; fax:+1-203-677-7702;
e-mail: harold.mastalerz@bms.com
0968-0896/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0968-0896(03)00495-4

4316
H. Mastalerz et al. / Bioorg. Med. Chem. 11 (2003) 4315–4323
research experience in the taxane area and the
availability of a large inventory of novel analogues that
were generated during previous programs to identify
second generation taxanes for clinical development.
Representative members of that analogue inventory
were screened in a mouse oral exposure assay11 and this
provided us with promising leads for an oral analogue
program as well as insights into structural features that
affect the oral absorption of taxane analogues. One such
insight was that more than one structural modification
to paclitaxel would probably be required to achieve
an acceptable oral bioavailability (%F>20). A pro-
mising lead that emerged was the C-4-methyl car-
bonate analogue of paclitaxel (2), a compound which
had already demonstrated excellent in vivo efficacy
when administered iv and is currently in clinical
trials.12 It was selected as a starting point in our
quest for an orally bioavailable analogue with an in
vivo oral efficacy profile that is at least equivalent to iv
dosed paclitaxel in preclinical models. This paper13
describes the highlights of that work, an effort that led
to the discovery of BMS-275183 (15i), an orally bioa-
vailable and efficacious novel taxane that is currently in
Phase I clinical trials.
Analogue Synthesis
Compounds 15a–15h are analogues of 2 that have differ-
ent N-acyl groups at the C-30 position. Their
preparation (Scheme 1) entailed, N-acylation of the chiral
azetidinone (11)14 with the appropriate acylating agent,
base-promoted reaction15 of the resulting N-acylated aze-
tidinones, 12a to 12h, with the C-4-carbamate inter-
mediate 13 that is derived from 10-desacetyl baccatin
III,12b and the final removal of the silyl protecting groups.
Preparation of the C-30-t-butyl and C-30-i-propyl analo-
gues, 15i and 15j, used the racemic azetidinones, 7 and 8,
that were generated by the Staudinger reaction of the
imines derived from trimethylacetaldehyde (5) and iso-
butyraldehyde (6) with benzyloxyacetyl chloride and
acetoxyacetyl chloride respectively. Oxidative removal
of the N-p-methoxyphenyl groups of 7 and 8 was fol-
lowed by replacement of the respective benzyl and ace-
tyl groups with the triethylsilyl group to give 9 and 10.
After conversion to their N-Boc derivatives, excess
racemic 12i and 12j was allowed to react with 13 under
basic conditions. Removal of the silyl protecting groups
from the resulting mixtures of C-20, C-30 diastereomers
followed by chromatography afforded the major iso-
Scheme 1. (a) p-Anisidine, DCM then benzyloxacetylchloride or acetoxyacetylchloride, TEA, DCM, ꢀ78 ꢁC to rt; (b) Ce(NH4)2(NO3)6, aq CH3CN,
then H2, Pd/C, EtOH or K2CO3 in MeOH, then TESCI, imidazole (2 equiv), DMF; (c) R2COCl or (Boc)2O, DIPEA (1.2 equiv), DMAP (0.2 equiv),
DCM, 0 ꢁC to rt; (d) 13, LiHMDS (1.3 equiv), THF; (e) (HF)3TEA (4 equiv), THF, rt; (f) TFA, DCM; (g) R2COCl (1.5 equiv), aq NaHCO3, DCM.

H. Mastalerz et al. / Bioorg. Med. Chem. 11 (2003) 4315–4323
4317
mers, 15i and 15j, that possessed the natural configur-
ation of substituents at the C-20 and C-30 positions. The
observed diastereoselectivity is well known for the base-
promoted reaction of racemic azetidinones with the C-13-
hydroxy group of baccatin derivatives.14a,16 Compounds
15k to 15n are analogues of 15i that have different C-30-N-
acyl groups. They were prepared by first removing the Boc
group from 15i by treatment with TFA and then N-acy-
lating the resulting C-30-amino group using Schotten–
Baumann conditions. The latter chemistry is analogous to
that which had been used by Georg and coworkers17 to
efficiently generate C-30-N-acyl analogues of paclitaxel.
M109 Madison murine lung carcinoma model (Table 1),
it was found to be more than 2-fold less active than iv
administered paclitaxel. Since the program objective
was to identify an orally bioavailable taxane that
showed comparable or superior efficacy to iv dosed
paclitaxel, modifications of the C-30-phenyl and/or the
C-30-N-acyl groups of 2 were explored in an effort to
enhance oral bioavailability and efficacy. A systematic
screening protocol and decision tier was used to deter-
mine which analogues would be advanced. All com-
pounds were initially evaluated in
a
tubulin
polymerization assay and their in vitro cytotoxicity
against HCT 116, a human colon tumor cell line that is
sensitive to paclitaxel, and HCT 116/VM26, a human
colon cell line that is resistant to paclitaxel due to a
MDR phenotype overexpressing gp 170, were deter-
mined. They were screened for oral exposure in the
mouse and those compounds that showed higher drug
plasma levels than 2 were evaluated for oral efficacy in
the M109 murine lung tumor model. The compounds
that showed a log cell kill (LCK) ratio of between 0.8
and 1.2 for the orally dosed analogue versus iv dosed
paclitaxel were considered comparable in efficacy to iv
dosed paclitaxel in this model and were advanced to the
next evaluation tier. Here, their cytotoxicity against
A2780, a paclitaxel sensitive human ovarian cancer
cell line, and an A2780 cell line that is resistant to
paclitaxel due to mutated b-tubulin, were determined
together with their oral bioavailability in the rat and
oral in vivo efficacy in an A2780, human ovarian
carcinoma, tumor xenograft model. The choice of
candidates for final preclinical evaluation was based
on their performance in this evaluation tier.
Results and Discussion
One of several interesting leads that emerged form our
oral exposure screen was the C-4-methyl carbonate
analogue 2. It is as potent against the HCT-116, human
colon carcinoma, cell line as paclitaxel and slightly more
potent in a tubulin polymerization assay (Table 1).
More importantly, it exhibits significant efficacy advan-
tages in preclinical in vivo models.12 Because of its
excellent in vivo efficacy, it was considered a prime
candidate for optimization in the oral taxane program.
It is about 40 times more soluble than paclitaxel18 and
exhibits about 50-fold higher drug plasma levels than
paclitaxel in the mouse oral exposure screen. Initial
studies with amorphous samples of 2 suggested that it
might possess a significant oral bioavailability advan-
tage over paclitaxel but definitive studies with crystalline
material demonstrated that it possessed minimal (4%)
oral bioavailability in the rat. It lacked oral efficacy in
mice verses sc implanted murine neoplasms. For exam-
ple, when it was administered orally at its maximum
tolerated dose (MTD) to mice in the paclitaxel-sensitive
A group of C-30-N-alkoxycarbonyl analogues was initi-
ally surveyed. Previous SAR studies suggested that the
Table 1. In vitro and preliminary in vivo profile of potential orally-active taxane analogues
R1
R2
Tubulina
HCT-116b (nM)
R/S ratioc
Mouse oral absorption (ng/mL)d
PO/SC M109e
1
2
Phenyl
Phenyl
Phenyl
Phenyl
Phenyl
Phenyl
Phenyl
Phenyl
Phenyl
Phenyl
t-Butyl
i-Propyl
t-Butyl
t-Butyl
t-Butyl
t-Butyl
Phenyl
Phenyl
t-Butoxy
1.0
0.4
nd
4.0
2.3
0.33
2.9
114
68
22
9.9
24
65
13
24
76
87
3.9
7.1
14
84
3935
15,626
5718
6641
0.02 (160)
0.4 (32)
0.6 (15)f
0.3 (70)
0.9 (50)
0.1 (13)
0.7 (25)
1.3 (50)
nd
1.0 (160)
0.8 (25)
0.9 (12)
0.4 (15)
1.0 (160)
0.6 (100)
1.1 (60)
15a
15b
15c
15d
15e
15f
15g
15h
15i
15j
15k
15l
15n
15m
i-Butoxy
0.8
0.9
1.0
1.0
0.9
0.4
0.7
0.4
0.9
0.9
0.9
0.6
0.8
n-Butoxy
i-Propoxy
Cyclohexyloxy
Neopentyl
Cyclohexyl
Cyclobutyl
t-Butoxy
1.3
0.62
0.79
3.6
0.40
0.85
1.8
0.53
2.6
14
9479
10,950
7877
1759
5443
7636
7205
14,656
8087
16,853
18,223
t-Butoxy
i-Propoxy
Neopentyl
i-Butyl
84
6
>17
6.1
7.0
Cyclobutyl
aConcentration of test compound to give a change of 0.01 OD/h expressed as a ratio to the concentration of paclitaxel.
bIC50 for HCT-116, human colon tumor, cell line sensitive to paclitaxel.
cRatio of IC50 for HCT-116 multi-drug resistant (MDR) cell line/IC50 for HCT-116 sensitive cell line. The resistant HCT-116 cell line was estab-
lished by exposing cells to VM-26 and establishing a cell line resistant to these drugs.
dMouse plasma concentration of the analogue 1 h after oral dosing at 100 mg/kg in 10:10:80 Cremophor EL/ethanol/water.
eMadison Murine Lung Carcinoma, M109, tumors implanted sc. The analogue was given orally every day for 5 days beginning 4 days after
implantation and paclitaxel given iv on the same schedule. Values presented are a ratio of the log cell kill (LCK) for the analogue (optimum dose in
mg/kg/administration)/LCK for iv dosed paclitaxel. Log cell kill is defined as LCK=(TꢀC/TVDT)/3.32. TꢀC refers to the difference in days for the
tumor to reach target size for the treated versus the control group. TVDT is the tumor volume doubling time. All compounds were administered as
solutions in Cremophor EL/ethanol/water mixtures except 15h for which a Tween 80/water vehicle was used.
fThis result was confirmed when this study was repeated.

4318
H. Mastalerz et al. / Bioorg. Med. Chem. 11 (2003) 4315–4323
30-N-t-butoxycarbonyl analogue 15a might have
improved oral bioavailability and, like docetaxel (3),
which also has a C-30-N-t-butyloxycarbonyl group,
exhibit more potent cytotoxicity than paclitaxel 1.2
Indeed, 15a showed an impressive increase in oral
exposure in the mouse screen as well as subsequent oral
bioavailaibility studies in rats. It was 7-fold more cyto-
toxic than 2 against the paclitaxel-sensitive HCT-116
cell line but continued to show reduced potency against
the paclitaxel-resistant cell line (R/S ratio of 22) indi-
cating that the 30-N-t-butoxycarbonyl group does not
significantly reduce susceptibility to MDR based resis-
tance. Unfortunately, 15a did not show oral efficacy
comparable to iv administered paclitaxel in the M109
tumor model and therefore a number of additional 30-
N-alkoxycarbonyl analogues, 15b–15e, were prepared to
see if it would be possible to modulate the cytotoxicity
and thereby achieve improved oral efficacy. These ana-
logues exhibited drug plasma levels that were lower than
15a but still higher than 2 in the oral exposure screen
and their cytotoxicity against the HCT-116 cell line was
2–9-fold less than 15a. Of these compounds, only 15f,
an analogue that was comparable to paclitaxel in cyto-
toxicity against the HCT-116 cell line, showed oral effi-
cacy comparable to iv dosed paclitaxel in the M109
model. It, together with 15a and 15e, was advanced to the
next evaluation tier. Even thought the latter two com-
pounds did not show oral efficacy comparable to iv pacli-
taxel in the M109 model, it was felt that they deserved
evaluation in other in vivo tumor models because exhib-
ited such high oral exposure in the mouse screen.
methylcarbonate (15a) with small aliphatic groups
reduces susceptibility to P-gp mediated efflux and this is
in agreement with data that has been reported by Ojima
and coworkers for C-30-alkyl analogues derived from
14b-hydroxy-10-deacetyl baccatin.10a There the C-30-
alkyl analogues also performed better against cell lines
expressing MDR than their C-30-phenyl counterparts.
In any event, both 15i and 15j performed comparably to
15a in the M109 model in spite of their weaker cytotoxi-
city and lower oral exposure. The C-30-N-i-propoxy-
carbonyl analogue (15k) of 15i was prepared and it
showed a two-fold increase in oral exposure in the
mouse. Although, it was comparable in cytotoxic
against the HCT-116 cell line, it showed poor oral effi-
cacy in the M109 model. Lastly some aliphatic C-30-N-
acyl analogues of 15i were prepared. The C-30-N-neo-
pentanoyl analogue (15l) is isosteric to 15i and shows
reduced cytotoxicity but similar oral exposure. It was
found to be comparable to iv administered paclitaxel in
the M109 model but this required a six-fold higher dose
than for 15i. The C-30-N-isobutanoyl (15n) and C-30-N-
cyclobutanoyl (15m) analogues both showed higher oral
exposure but only the latter showed efficacy comparable
to iv dosed paclitaxel in the M109 model.
Those analogues that performed as well as iv adminis-
tered paclitaxel in the M109 tumor model were then
evaluated in A2780, human ovarian carcenoma, xeno-
graft models (Table 2). In vitro all of these compounds
except for 15l and 15m, two of the C-30-t-butyl analo-
gues, were as cytotoxic as paclitaxel. Against paclitaxel-
resistant A2780, a subline which is resistant to paclitaxel
due to altered tubulin, the C-30-N-t-butyloxycarbonyl
analogues carrying the C-30-t-butyl (15i) and the C-30-i-
propyl (15j) groups performed best, showing only a 4 to
6-fold drop in potency. In vivo, in the A2780 xenograft
model, the latter two compounds continued to perform
well and showed a LCK of greater than or equal to 3.
The 30-phenyl analogues, 15c, 15f and 15h, performed at
least as well although higher doses were required to
achieve this effect. Of these compounds, the C-30-t-butyl
analogue (15i) showed the highest oral bioavailability in
the rat. Its solubility in 10:10:80 Cremophor El/ethanol/
water was 16.5 mg/mL; just a little more than that
observed for 2 in this medium. Taken all together, 15i
showed good oral efficacy in two tumor models, good
oral bioavailability, and potency against cell lines resis-
tant to paclitaxel due to MDR and altered tubulin. It
was therefore moved forward for additional preclinical
evaluation20 and emerged from those studies as a clin-
ical candidate for an orally active taxane.
The effect of replacing the N-benzoyl group at C-30 in 2
with aliphatic N-acyl groups isosteric to the carbamates
was examined to see if this would attenuate potency and
improved efficacy (Table 1). The neopentanoyl (15f) and
cyclobutanoyl (15h), analogues were less cytotoxic than
15a against the HCT-116 cell line while the cyclohex-
anoyl analogue (15g) was comparable. Only the former
two compounds showed drug plasma levels that were
higher than 2 in the oral exposure screen and both were
found to be as active as iv dosed paclitaxel when dosed
orally in the M109 model.
Further exploration of the C-30-carbamates included
analogues where the phenyl group at the 30 position was
replaced by a t-butyl or i-propyl group. This choice was
based on a report that 4, the C-30-t-butyl analogue of
docetaxel, is about 5 times more water-soluble than
docetaxel (3) and performs similarly in tubulin poly-
merization and cytotoxicity assays.19 It was thought
that this modification in our C-4-methyl carbamate ser-
ies may further improve solubility and thereby increase
oral absorption. Although the drug plasma levels for the
C-30-t-butyl (15i) and C-30-i-propyl (15j) analogues were
only found to be about half that seen for the C-30-
phenyl analogue (15a) in the mouse screen, subsequent
studies showed respectable oral bioavailability for 15i in
three species. Both compounds were found to be less
cytotoxic against the HCT-116 cell line than 15a but
they did exhibit reduced cross resistance against the
paclitaxel-resistant HCT-116 cell line. This suggests that
replacement of the C-30-phenyl group of the C-30-
Experimental
General experimental methods
1H NMR spectra were recorded on a Bruker AC-300
spectrometer and chemical shifts reported in ppm
downfield from a TMS standard. 13C NMR spectra
were recorded on the same instrument at 75.5 MHz.
High resolution mass spectral analysis was performed
on a Kratos MS50RF spectrometer in the FABmode

H. Mastalerz et al. / Bioorg. Med. Chem. 11 (2003) 4315–4323
4319
Table 2. Oral bioavailability, in vitro potency and oral efficacy of taxane analogues in human ovarian carcinoma (A2780) xenograft models
Rat oral Fa (%)
Cytotoxicity versus A2780
Analogue/paclitaxel ratiob
Oral efficacy versus sc A2780
R/S ratioc
LCK (cures/treated)
Oral dosed
15a
15c
15e
15f
15h
15i
15j
15l
15m
44
18
10
30
nd
52
36
51
34
0.2
0.9
0.5
1.4
1.0
1.6
0.7
17
29
15
32
0.9 (0/8)
7.0 (1/8)
1.5 (0/8)
3.8 (0/8)
4.1 (0/8)
4.4 (0/8)
3.0 (1/8)
1.8 (0/8)
2.1 (0/8)
32
100
55
100
140
65
28
36
120
15
3.5
6.1
>7.2
11
10
6.7
aDetermined using an oral dose of 40 mg/kg and an iv dose of 10 mg/kg of amorphous samples of the analogue with 10:10:80 Cremophor EL/
ethanol/water as the vehicle.
bRatio of IC50 for the analogue/IC50 for paclitaxel versus A2780 cells.
cRatio of IC50 versus paclitaxel resistant A2780 cells/IC50 versus A2780 cells.
dOptimal or maximally tolerated dose in mg/kg/administration given q2dx5 (every 2nd day for five treatments) except for 15 h where the indicated
dose was an LD50
.
using m-nitrobenzyl alcohol as the matrix. Column
chromatography was performed using silica gel 60 (200–
400 mesh). Unless otherwise noted, materials were
obtained from commercial sources and used without
further purification. Detailed descriptions of the in
vitro assays and the in vivo tumor models have been
published.20
(6 mL) and left stirring at RT overnight. After neu-
tralization with saturated aq NaHCO3 solution, the
reaction was extracted with EtOAc. The organic
extracts were washed with brine and dried (Na2SO4).
Removal of the solvents followed by silica gel column
chromatography (gradient elution with mixtures of 20–
50% EtOAc in hexane) afforded 151 mg (71%) of the
1
title compound: H NMR (CDCl3): d 0.96–2.58 [32H,
4-Deacetyl-30 -N-debenzoyl-4-N-neopentylcarbonyl-O-
methoxycarbonyl-30-paclitaxel (15f). tert-Butylacetyl
chloride (1.17 mL, 1.2 equiv) was added to a solution of
(3R,4S)-4-phenyl-3-triethylsilyloxy-azetidin-2-one (11,
1.94 g, 7 mmol), DIPEA (1.46 mL, 1.2 equiv), and
DMAP (171 mg, 0.2 equiv) in dry DCM (40 mL) at
0 ꢁC. After 3 h, the reaction was diluted with DCM (100
mL) and washed successively with aq HCl (1.0 M, 40
mL), saturated aq NaHCO3 solution (40 mL) and brine
(40 mL). After drying with Na2SO4, the solvent was
removed and the residue was chromatorgraphed on
silica gel (step gradient elution with 0–5% ethyl acetate
in hexane) to (3R,4S)-1-neopentylcarbonyl-4-phenyl-3-
triethylsilyloxy-azetidin-2-one (10f, 2.51 g, 96%) as an
oil: H NMR1 (CDCl3): 0.18–0.62 (m, 15e), 0.87 (s, 9H),
2.43 (d, 1H, J=13.8 Hz), 2.62 (d, 1H, J=13.8 Hz), 4.90
(d, 1H, J=5.7 Hz), 4.95 (d, 1H, J=6 Hz), 7.05–7.17 (m,
5H).
including 0.96 (s, 9H), 1.14 (s, 3H), 1.24 (s, 3H), 1.66 (s,
3H), 1.84 (s, 3H), 2.23 (s, 3H)], 3.58 (br s, 1H), 3.77 (s,
3H), 3.80 (d, 1H, J=5.5 Hz), 4.19 (d, 1H, J=8.3 Hz),
4.33 (d, 1H, J=8.7 Hz), 4.36 (m, 1H), 4.65 (d, 1H,
J=2.0 Hz), 4.95 (d, 1H, J=8.5 Hz), 5.58 (dd, 1H,
J=2.3, 8.8 Hz), 5.69 (d, 1H, J=7.0 Hz), 6.11 (d, 1H,
J=8.9 Hz), 6.16 (m, 1H), 6.27 (s, 1H), 7.29–8.12 (m,
10H); 13C NMR (CDCl3) : d 9.7, 15.0, 21.0, 26.9, 29.8,
31.0, 35.5, 35.9, 43.2, 45.9, 50.6, 54.7, 56.0, 58.4, 72.1,
72.5, 73.1, 75.0, 75.7, 76.1, 79.0, 83.1, 84.2, 127.3, 128.3,
128.8, 129.0, 129.2, 130.3, 133.4, 133.8, 138.8, 142.4,
153.4, 167.1, 171.4, 171.5, 173.2, 196.2, 203.65; LRMS
(ESI) 864 [(M+H)+]; HRMS (ESI): calcd for
C46H57NO15+H: 864.3801, found: 864.3790.
30-N-t-Butoxycarbonyl-4-deacetyl-30 -N-debenzoyl-4-O-
methoxycarbonyl-paclitaxel (15a). Following the same
procedure with di-t-butyl-dicarbonate gave the title
compound: 1H NMR (CDCl3) d 1.03–2.59 [30H,
including 1.14 (s, 3H), 1.27 (s, 3H), 1.32 (s, 9H), 1.71 (s,
3H), 1.87 (s, 3H), 2.24 (s, 3H)], 3.40 (d, 1H, J=5.1 Hz),
3.85–3.81 (m, 4H), 4.15 (d, 1H, J=8.8 Hz), 4.34 (d, 1H,
J=8.8 Hz), 4.40 (m, 1H), 4.63 (d, 1H, J=3.9 Hz), 4,97
(d, 1H, J=7.7 Hz), 5.32 (m, 2H), 5.68 (d, 1H, J=6.9
Hz), 6.19 (m, 1H), 6.27 (s, 1H), 7.61–7.26 (m, 8H) 8.13–
8.10 (m, 2H); LRMS (ESI) 864 [(MꢀH)ꢀ].
4-Deacetyl-30-N-debenzoyl-30-N-isobutoxycarbonyl-4-O-
methoxycarbonyl-paclitaxel (15b). Following the same
procedure with isobutylchloroformate gave the title
compound: 1H NMR (CDCl3) d 0.80–2.61 [m, 28H
including: 1.16 (s, 3H), 1.26 (s, 3H), 1.70 (s, 3H), 1.88 (s,
3H), 2.26 (s, 3H)], 3.46 (br s, 1H), 3.75 (m, 2H), 3.98 (m,
4H), 4.20 (d, 1H, J=8.6 Hz), 4.36 (d, 1H, J=8.5 Hz),
4.40 (m, 1H), 4.68 (br s, 1H), 4.96 (d, 1H, J=7.8 Hz),
5.35 (d, 1H, J=9.2 Hz), 5.53 (d, 1H, J=9.4 Hz), 5.71
A solution of 10f (525 mg, 1.4 equiv) and 4-deacetyl-7-
[bisisopropyl(methoxy)]silyloxy-4-O-methoxycarbonyl-
baccatin (13, 523 mg, 0.700 mmole) in dry THF (15 mL)
was cooled to ꢀ50 ꢁC and a solution of LiHMDSA
(0.84 mL, 1.2 equiv, 1.0 M in THF) was added with
stirring. After 40 min, the reaction was allowed to warm
to 0 ꢁC. After 1.5 h, this was quenched with a saturated
aq solution of NH4Cl and extracted with EtOAc. The
organic phase was washed with a saturated aq solution of
NH4Cl, water, brine and dried (Na2SO4). Removal of the
solvents followed by silica gel column chromatography
(gradient elution with mixtures of 0–20% EtOAc in hex-
ane) afforded 30-N-neopentyloxycarbonyl-7-[bisisopropyl
(methoxy)]silyloxy-4-deacetyl-30-N-debenzoyl-4-O-meth-
oxycarbonyl-20-triethylsilyloxy-paclitaxel (14f, 274 mg,
54%). This was taken directly and treated with triethy-
lamine trihydrofluoride (0.161 mL, 4 equiv) in dry THF

4320
H. Mastalerz et al. / Bioorg. Med. Chem. 11 (2003) 4315–4323
(d, 1H, J=7.0 Hz), 6.24 (m, 1H), 6.29 (s, 1H), 7.31–8.12
(m, 10H); LRMS (ESI) 866 [(M+H)+].
with cyclobutanecarbonyl chloride gave the title com-
1
pound: H NMR (CDCl3) d 1.14–2.53 [m, 27H includ-
ing: 1.14 (s, 3H), 1.25 (s, 3H), 1.67 (s, 3H), 1.84 (s, 3H),
2.24 (s, 3H)], 3.01 (m, 1H), 3.56 (br s, 1H), 3.81 (s, 3H),
3.82 (m, 1H), 4.20 (d, 1H, J=8.4 Hz), 4.34 (d, 1H,
J=8.5 Hz), 4.37 (m, 1H), 4.68 (d, 1H, J=2.3 Hz), 4.96
(d, 1H, J=8.6 Hz), 5.58 (dd, 1H, J=2.4, 9.0 Hz), 5.70
(d, 1H, J=7.0 Hz), 6.16 (m, 2H), 6.27 (s, 1H), 7.29–8.14
(m, 10H); LRMS (ESI) 848 [(M+H)+].
N-n-Butoxycarbonyl-4-deacetyl-30-N-debenzoyl-30-4-O-
methoxycarbonyl-paclitaxel (15c). Following the same
procedure with n-butyl chloroformate gave title com-
1
pound: H NMR (CDCl3) d 0.83–2.59 [30H, including
1.14 (s, 3H), 1.24 (s, 3H), 1.67 (s, 3H), 1.86 (s, 3H), 2.24
(s, 3H)], 3.42 (d, 1H, J=4.1 Hz), 3.83 (s, 3H), 4.15 (d,
1H, J=8.6 Hz), 4.34 (d, 1H, J=8.6 Hz), 4.34 (m, 1H),
4.63 (brs, 1H), 4.97 (d, 1H, J=7.0 Hz), 5.41 (m, 2H),
5.69 (d, 1H, J=7.0 Hz), 6.27 (m, 1H), 6.27 (s, 1H),
7.62–7.29 (m, 10H) 8.11 (d, 2H, J=7.4 Hz). Anal. calcd
For C45H55NO16: C, 62.42; H, 6.40; N, 1.62. Found: C,
62.28; H, 6.45; N, 1.55.
(ꢂ)-cis-3-benzyloxy-4-tert-butyl-1-p-methoxyphenyl-aze-
tidinone (7). Trimethylacetaldehyde (5, 20.3 mL, 1.25
equiv) was added to a strirred suspension of p-anisidine
(18.4 g, 0.150 mole) and anhydrous Na2SO4 (150 g) in
anhydrous DCM (250 mL) at rt. After 2 h, this was fil-
tered and the solid was wash with additional anhydrous
DCM. The solvent was removed from the filtrate and
the crystalline residue was dissolved in anhydrous DCM
(750 mL) and placed under a nitrogen atmosphere. TEA
(48.0 mL, 2.3 equiv) was added and the reaction was
cooled to ꢀ78 ꢁC. Benzyloxyacetyl chloride (27.2 mL
1.15 equiv) was added dropwise and then the reaction
was allowed to warm to rt. After 24 h, this was washed
twice with aq HCl (0.5 M), sat. aqueous NaHCO3 solu-
tion, brine and dried (Na2SO4). The solvent was
removed and the residue was chromatographed on a
silica gel column (gradient elution with 20% DCM in
hexane containing 0–20% EtOAc) to afford 7 (46.9 g,
4-Deacetyl-30 -N-debenzoyl-30 -N-isopropoxycarbonyl-4-
O-methoxycarbonyl-paclitaxel (15d). Following the
same procedure with i-propyl chloroformate gave the
1
title compound: H NMR (CDCl3) d 1.03–2.58 [m, 26H
including: 1.24 (s, 3H), 1.66 (s, 3H), 1.84 (s, 3H), 2.28 (s,
3H)], 3.44 (br s, 1H), 3.82 (s, 3H), 3.84 (m, 1H), 4.17 (d,
1H, J=8.6 Hz), 4.33 (d, 1H, J=8.5 Hz), 4.38 (m, 1H),
4.64 (m, 1H), 4.77 (sept, 1H, J=6.2 Hz), 4.96 (d, 1H,
J=7.8 Hz), 5.38 (m, 2H), 5.69 (d, 1H, J=7.1 Hz), 6.20
(m, 1H), 6.37 (s, 1H), 7.28–8.32 (m, 10H); LRMS (ESI)
852 [(M+H)+].
30-N-Cyclohexyloxycarbonyl-4-deacetyl-30-N-debenzoyl-
4-O-methoxycarbonyl-paclitaxel (15e). A mixture of
cyclohexanol (12.68 mL, 120 mole) and pyridine (10.68
mL, 1.1 equiv) was added dropwise to a solution of tri-
phosgene (11.87, 0.33 equiv) in anhydrous diethyl ether
(100 mL) at ꢀ78 ꢁC. After 1 h, the reaction was
removed from the bath and after a further 1.5 h, aq HCl
(1.0 N, 110 mL) was added. The organic phase was
separated and washed with brine (60 mL) and dried
(Na2SO4). Removal of the solvent followed by distilla-
tion afforded cyclohexyl chloroformate (14.2 gm, bp 41
92%): H NMR (CDCl3) d 1.09 (s, 9H), 3.81 (s, 3H),
1
4.15 (d, 1H, J=5.5 Hz), 4.77 (d, 1H, J=11.9 Hz), 4.81
(d, 1H, J=5.5 Hz), 5.03 (d, 1H, J=11.9 Hz) 6.87–7.43
(m, 9 Hz); LRMS (ESI) 340 ([M+H]+).
(ꢂ)-cis-4-tert-Butyl-3-triethylsilyloxy-azetidin-2-one (9).
A solution of ceric ammonium nitrate (60.4 g, 3.6
equiv) in 900 mL of water was added over 1 h to a well
stirred solution of 7 (10.38 g, 30.6 mmole) in acetonitrile
(600 mL) in an ice bath. This was extracted twice with
EtOAc and the combined organic extracts were washed
twice with satd aqueous NaHCO3 solution, 20% aqu-
eous NaHSO3 solution, satd aqueous NaHCO3 solution
and brine. After being dried (Na2SO4), the solvents were
removed and the residue was chromatographed on a
silica gel column (gradient elution with 10 to 40%
EtOAc in hexane) to afford 5.64 g of crude (ꢂ)-cis-3-
to 42 ꢁC at 4 mm Hg): H NMR (CDCl3) d 1.08–1.80
1
(m, 10H), 4.65 (m, 1H). Following the above procedure
with cyclohexyl chloroformate gave the title compound:
1H NMR (CDCl3) d 1.15–2.60 [m, 31H including: 1.15
(s, 3H), 1.26 (s, 3H), 1.68 (s, 3H), 1.85 (s, 3H), 2.26 (s,
3H)], 3.48 (br s, 1H), 3.86 (m, 4H), 4.18 (d, 1H, J=8.6
Hz), 4.40 (m, 3H), 4.68. (br s, 1H), 4.99 (d, 1H, J=7.9
Hz), 5.36 (d, 1H, J=8.9 Hz), 5.49 (d, 1H, J=9.4 Hz),
5.70 (d, 1H, J=7.1 Hz), 6.23 (m, 1H), 6.29 (s, 1H),
7.30–8.16 (m, 10H); LRMS (ESI) 892 [(M+H)+].
benzyloxy-4-tert-butyl-azetidin-2-one:
1H
NMR
(CDCl3) d 1.04 (s, 9H), 3.51 (d, 1H, J=5.2 Hz), 4.71 (m,
2H), 4.96 (d, 1H, J=11.9 Hz) 6.10 (brs, 1H), 7.35 (m,
5H). A suspension of this material (5.54 g, 23.8 mmole)
and 2.5 g of 10% Pd on charcoal in absolute EtOH (100
mL) was hydrogenated (34 psi, Parr apparatus) for 23 h.
A further 2 g of the Pd catalyst was added and the
hydrogenation was continued for a further 17 h at 50
psi. Removal of the catalyst followed by the solvent left
30-N-Cyclohexylcarbony-4-deacetyl-30-N-debenzoyl-4-O-
methoxycarbonyl-paclitaxel (15g). Following the same
procedure with cyclohexanecarbonyl chloride gave the
title compound: 1H NMR (CDCl3+D2O) d 1.17–2.60 [m,
49H including: 1.17 (s, 3H), 1.27 (s, 3H), 1.69 (s, 3H), 1.95
(s, 3H), 2.26 (s, 3H)], 3.77 (s, 3H), 3.84 (d, 1H, J=7.0 Hz),
4.23 (d, 1H, J=8.5 Hz), 4.35 (d, 1H, J=8.0 Hz), 4.39 (m,
1H), 4.70 (d, 1H, J=2.1 Hz), 4.98 (d, 1H, J=7.6 Hz), 5.60
(m, 1H), 5.74 (d, 1H, J=7.0 Hz), 6.16 (m, 1H), 6.29 (s,
1H), 7.33–8.16 (m, 10H); LRMS (ESI) 876 [(M+H)+].
1
crude (ꢂ)-cis-4-(tert-butyl)-3-hydroxy-azetidin-2-one: H
NMR (CDCl3+1 drop DO) d 1.05 (s, 9H), 3.48 (d, 1H,
J=5.0 Hz), 4.98 (d, 1H, J=5.0 Hz). This was dissolved in
dry DMF (40 mL) and imidazole (3.24 g, 2 equiv) and
triethylsilyl chloride (4.0 mL, 1 equiv) were added. After
10 min, the reaction was partitioned between water and
a mixture of EtOAc and hexane (1:1). The organic
phase was washed with water (twice), brine and then
dried (Na2SO4). The solvents were removed and the
30-N-Cyclobutyl-4-deacetyl-30-N-debenzoyl-4-O-methoxy-
carbonyl-paclitaxel (15h). Following the same procedure

H. Mastalerz et al. / Bioorg. Med. Chem. 11 (2003) 4315–4323
4321
residue was chromatographed on a silica gel column
(gradient elution with 20–25% EtOAc in hexane) to give
9 (3.86 g, 49%): 1H NMR (CDCl3) d 0.70 (m, 6H), 0.98
(m, 18H), 3.39 (d, 1H, J=5.0 Hz), 4.88 (dd, 1H, J=2.1,
5.0 Hz), 6.08 (brs, 1H).
imidazole (3.39 g, 2 equiv) and triethylsilyl chloride
(4.19 mL, 1 equiv) were added. After 10 min, the reac-
tion was partitioned between water and a mixture of
EtOAc and hexane (1:1). The organic phase was washed
twice with water, brine and then dried (Na2SO4). The
solvents were removed and the residue was chromato-
graphed on a silica gel column (gradient elution with
(ꢂ)-cis-4-tert-Butyl-1-tert-butyloxycarbonyl-3-triethyl-
silyloxy-azetidin-2-one (12i). A solution of 9 (2.04 g,
7.92 mmole), diisopropylethylamine (1.66 mL, 1.2
equiv), di-tert-butyl dicarbonate (1.90 g, 1.1 equiv) and
p-dimethylaminopyridine (194 mg, 0.2 equiv) in dry
DCM (24 mL) was left stirring at rt for 3 h. The reac-
tion was diluted with DCM, washed with brine and
dried (Na2SO4). Removal of the solvent followed by
silica gel column chromatography (gradient elution with
0–20% EtOAc in hexane) gave 12i (2.71 gm, 96%) as an
1
25–35% EtOAc in hexane) to give 10 (4.63 g, 77%): H
NMR (CDCl3) d 0.65–1.03 (m, 21H), 1.93 (m, 1H), 3.29
(dd, 1H, J=4.8, 9.1 Hz), 4.87 (dd, 1H, J=2.8, 4.7 Hz),
6.05 (br s, 1H).
(ꢂ)-cis-1-tert-Butyloxycarbonyl-4-iso-propyl-3-triethylsi-
lyloxy-azetidin-2-one (12j). A solution of 10 (1.05 g, 4.32
mmole), diisopropylethyl amine (0.90 mL, 1.2 equiv),
di-tert-butyl dicarbonate (1.04 g, 1.1 equiv) and p-
dimethylaminopyridine (106 mg, 0.2 equiv) in dry DCM
(10 mL) was left stirring at rt for 30 min. The reaction
was diluted with DCM, washed with brine and dried
(Na2SO4). Removal of the solvent followed by silica gel
column chromatography (gradient elution with 10–20%
EtOAc in hexane) afforded 1.31 g (88%) of 12b as an
1
oil: H NMR (CDCl3) d 0.70 (m, 6H), 1.00 (m, 9H),
1.09 (s, 9H), 1.53 (s, 9H), 3.90 (d, 1H, J=6.5 Hz), 4.93
(d, 1H, J=6.5 Hz).
(ꢂ)-cis-3-Acetoxy-4-isopropyl-1-p-methoxybenzyl-azeti-
din-2-one (8). Isobutyraldehyde (6, 4.62 mL, 1.25 equiv)
was added to a strirred suspension of p-anisidine (5.00
g, 40.7 mmole) and anhydrous Na2SO4 (25 g) in anhy-
drous DCM (80 mL) at rt. After 1 h, this was filtered
and the solid was wash with additional anhydrous
DCM. The solvent was removed from the filtrate and
the residue was dissolved in anhydrous DCM (200 mL)
and placed under a nitrogen atmosphere. Triethylamine
(13.1 mL, 2.3 equiv) was added and the reaction was
cooled to ꢀ78 ꢁC. Acetoxyacetyl chloride (5.00 mL 1.15
equiv) was added dropwise and the reaction was
allowed to warm to rt. After 20 h, this was washed twice
with 0.5 M HCl, satd aqueous NaHCO3 solution, brine
and dried (Na2SO4), the solvent was removed and the
residue was chromatographed on a silica gel column
(gradient elution with 20–30% EtOAc in hexane) to
1
oil: H NMR (CDCl3) d 0.66–1.07 (m, 21H), 1.53 (s,
9H), 2.15 (m, 1H), 3.87 (t, 1H, J=6.4 Hz), 4.88 (d, 1H,
J=6.1 Hz); LRMS (ESI) 344 [(M+H)]+.
30-tert-Butyl-30 -N-tert-Butyloxycarbonyl-4-deacetyl-30 -
dephenyl-30 -N-debenzoyl-4-O-methoxycarbonyl-pacli-
taxel (15i). A solution of 12i (2.71 g, 5 equiv) and 4-
deacetyl-7-[bisisopropyl(methoxy)]silyloxy-4-methoxy-
carbonyl-baccatin (13, 1.13 g, 1.52 mmole) in dry THF
(100 mL) under nitrogen was cooled to ꢀ50 ꢁC and a
solution of LiHMDSA (1.97 mL, 1.3 equiv, 1.0 M in
THF) was added. After 5 min this was transferred to a
bath that was maintained at ꢀ35 to ꢀ30 ꢁC for 20 h and
then ꢀ25 ꢁC for 24 h. The reaction was quenched with
saturated aqueous NH4Cl solution and extracted with a
mixture of EtOAc and hexane (1:1). The organic
extracts were washed with brine and dried (Na2SO4).
The solvents were removed and the residue was chro-
matographed (radial chromatography; 6-mm silica gel
plate; gradient elution with 5–20% EtOAc in hexane) to
afford 14a (1.55 g) as a mixture of 20,30-diastereomers.
This was dissolved in dry THF (60 mL) and triethyla-
mine trihydrofluoride (0.92 mL 4 equiv) were added.
After 22 h at rt, the reaction was neutralized with satu-
rated aq NaHCO3 solution and then extracted with
EtOAc. The organic extracts were washed with brine,
dried (Na2SO4) and the solvents were removed. The
residue was chromatographed (radial chromatography;
2 mm silica gel plate; gradient elution with 10–50%
EtOAc in hexane) to afford (in order of elution): 210 mg
1
afford 8 as a solid (7.15 g, 63%): H NMR (CDCl3) d
0.99 (d, 3H, J=7.0 Hz),1.02 (d, 3H, J=7.0 Hz), 2.20 (s,
3H), 3.82 (s, 3H), 4.24 (t, 1H, J=5.6 Hz), 6.06 (d, 1H,
J=5.3 Hz), 6.88–7.38 (m, 4H).
(ꢂ)-cis-4-Isopropyl-3-triethylsilyloxy-azetidin-2-one (10).
A solution of ceric ammonium nitrate (51.3 g, 3.6
equiv) in 750 mL of water was added to a well stirred
solution of the azetidinone (7.20 g, 26.0 mmole) in ace-
tonitrile (500 mL) in an ice bath over 1 h. The reaction
was then extracted twice with EtOAc and the combined
organic extracts were washed twice with satd aqueous
NaHCO3 solution, 20% aqueous NaHSO3 solution,
satd aqueous NaHCO3 solution and brine. After being
dried (Na2SO4), the solvents were removed to leave 4.26
g of crude (ꢂ)-cis-3-acetoxy-4-iso-propyl-azetidin-2-
(18%)
of
20S,30R-30-tert-butyl-3-0N-tert-butyloxy-
one: H NMR (CDCl3) d 0.86 (d, 3H, J=6.6 Hz), 0.99
carbonyl-4-deacetyl-30 -dephenyl-30 -N-debenzoyl-4-O-
methoxycarbonyl-paclitaxel {1H NMR (CDCl3) d 1.04
(s, 9H), 1.13 (s, 3H), 1.20 (s, 3H), 1.37 (s, 9H), 1.65 (s,
1H), 1.66 (s, 3H), 1.84–1.93 (m, 2H), 2.17 (s, 3H), 2.25
(s, 3H), 2.55 (m, 3H), 3.00 (d, 1H, J=6.5 Hz), 3.74 (d,
1H, J=10.8 Hz), 3.79 (d, 1H, J=6.9 Hz), 3.92 (s, 3H),
4.16 (d, 1H, J=8.5 Hz), 4.33 (d, 1H, J=8.5 Hz), 4.42
(m, 1H), 4.54 (d, 1H, J=6.5 Hz) 4.87 (d, 1H, J=10.6
Hz), 5.01 (d, 1H, J=7.7 Hz), 5.68 (d, 1H, J=7.0 Hz),
5.76 (m, 1H), 6.32 (s, 1H), 7.44–8.05 (m, 5H);. LRMS
1
(d, 3H, J=6.6 Hz), 1.89 (m, 1H), 2.17 (s, 3H), 3.52 (dd,
1H, J=4.8, 9.0 Hz), 5.96 (dd, 1H, J=2.5, 4.6 Hz), 6.38
(br s, 1H), LRMS (negative ESI) 170 [(MꢀH)ꢀ]. A sus-
pension of this material (4.26 g, 24.9 mmole) and
K2CO3 (102 mg, 0.03 equiv) in MeOH (40 mL) was
stirred at rt for 1.5 h. The reaction was neutralized with
Amberlite IR-20. Filtration followed by removal of the
solvent left crude (ꢂ)-cis-3-hydroxy-4-iso-propyl-azeti-
din-2-one. This was dissolved in dry DMF (40 mL) and

4322
H. Mastalerz et al. / Bioorg. Med. Chem. 11 (2003) 4315–4323
(ESI) 846 [(M+H)+]} and the 20R,30S isomer 15i (668
mg, 56%): H NMR (CDCl3) d 1.07 (s, 9H), 1.14 (s,
100%], 886.57 [(M+NH4+ACN)+, 30%]; 826.48
[(Mꢀ1)ꢀ, 100%]. HRMS (ESIꢀ): calcd for
C43H56NO15: 826.3650; found: 826.3645.
1
3H), 1.24 (s, 3H), 1.33 (s, 9H), 1.66 (s, 4H), 2.23 (s, 3H),
2.38–2.59 (m, 4H), 3.11 (d, 1H, J=5.8 Hz), 3.77 (d, 1H,
J=11.1 Hz), 3.82 (d, 1H, J=7.0 Hz), 3.96 (s, 3H), 4.20
(d, 1H, J=8.6 Hz), 4.33 (d, 1H, J=8.6 Hz), 4.39 (m,
1H), 4.53 (d, 1H, J=5.4 Hz) 4.88 (d, 1H, J=10.6 Hz),
4.98 (d, 1H, J=7.9 Hz), 5.69 (d, 1H, J=7.1 Hz), 6.03
(m, 1H), 6.28 (s, 1H), 7.40–8.11 (m, 5H); 13C NMR
(75 MHz, CDCl3) d 9.8, 14.3, 15.0, 21.0, 21.2, 22.3, 26.9,
27.5, 28.4, 35.1, 35.5, 35.9, 43.4, 45.9, 55.9, 58.4, 60.1,
60.5, 70.4, 72.1, 73.7, 75.2, 75.8, 76.2, 79.2, 79.6, 83.1,
84.3, 128.8, 129.3, 130.4, 133.1, 133.8, 143.0, 153.5,
155.9, 167.1, 171.3, 171.5, 175.4, 203.8; LRMS (ESI)
846 [(M+H)+]; HRMS (ESI): calcd for C43H58NO16:
844.3755, found: 844.3780.
30-tert-Butyl-4-deacetyl-30-dephenyl-30-N-debenzoyl-30-N-
isopropyloxycarbonyl-4-O-methoxycarbonyl-paclitaxel
(15k). Following the above procedure with isopropyl-
chloroformate afforded 15k: 1H NMR (CDCl3,
300 MHz) d 1.08 (s, 9H), 1.18–1.11 (m, 6H), 1.25 (s,
6H), 1.67 (s, 3H), 1.91 (s, 3H), 1.83–1.93 (m, 3H), 2.04
(s, 1H), 2.25 (s, 3H), 2.38 (d, J=8.9 Hz, 2H), 2.50–2.60
(m, 2H), 3.23 (d, J=4.8 Hz, 1H), 3.81–3.87 (bm, 2H),
3.99 (s, 3H), 4.21 (d, J=8.6 Hz, 1H), 4.33–4.38 (bm,
2H), 4.57 (d, J=3.9 Hz, 1H), 4.76 (m, 1H), 5.01 (m,
2H), 5.70 (d, J=7.1 Hz, 1H), 6.06 (m, 1H), 6.29 (s, 1H),
7.44–7.49 (m, 2H), 7.55–7.62 (m, 1H), (8.09 (d, J=7.3
Hz, 2H); 13C NMR (CDCl3, 75 MHz) d 203.79, 175.02,
171.53, 167.16, 156.39, 153.41, 142.85, 133.90, 133.27,
130.36, 128.85, 84.27, 83.35, 79.29, 77.45, 76.19, 75.80,
75.14, 73.38, 72.18, 70.44, 68.55, 60.47, 58.51, 56.13,
45.99, 43.43, 35.97, 35.58, 35.32, 27.47, 27.02, 26.97,
22.08, 21.25, 21.06, 15.06, 14.41, 14.32, 9.85; LRMS
(ESI): 832.46 [(M+1)+, 100%]; 830.44 [(Mꢀ1)ꢀ,
100%]. HRMS (ESIꢀ): calcd for C42H56NO16:
830.3599; found: 830.3607.
30 -N-tert-Butyloxycarbonyl-4-deacetyl-30 -N-debenzoyl-
30-dephenyl-30-isopropyl-4-O-methoxycarbonyl-paclitaxel
(15j). Following the above procedure with 12j afforded
15j: 1H NMR (CDCl3+D2O) d 1.03 (d, 3H, J=6.7 Hz),
1.09 (d, 3H, J=6.7 Hz), 1.14 (s, 3H), 1.24 (s, 3H), 1.31
(s, 9H), 1.66 (m, 3H), 1.83–2.02 (m, 5H), 2.24 (s, 3H),
2.25–2.59 (m, 3H), 3.68 (dd, 1H, J=2.0, 9.2 Hz), 3.82
(d, 1H, J=6.9 Hz), 3.98 (s, 3H), 4.19 (d, 1H, J=8.6
Hz), 4.34 (d, 1H, J=8.6 Hz), 4.39 (m, 1H), 4.43 (d, 1H,
J=2.0 Hz) 4.82 (br s, 1H), 4.98 (d, 1H, J=7.8 Hz), 5.69
(d, 1H, J=7.0 Hz), 6.11 (m, 1H), 6.28 (s, 1H), 7.45–8.12
(m, 5H); LRMS (ESI) 832 [(M+H)+].
30-tert-Butyl-4-deacetyl-30-dephenyl-30-N-debenzoyl-30-N-
neopentylcarbonyl-4-O-methoxycarbonyl-paclitaxel (15l).
Following the above procedure with tert-butylacetyl
1
chloride gave 15l: H NMR (CDCl2+D2O) d 1.00–2.56
30-tert-Butyl-30-N-cyclobutanecarbonyl-4-deacetyl-30-de-
phenyl-30 -N-debenzoyl-4-O-methoxycarbonyl-paclitaxel
(15m). Trifluoroacetic acid (15 mL) was added to a
solution of 15i (2.30 g, 2.72 mmole) in dry DCM (15
mL) at 0 ꢁC. After 1.5 h, this was diluted with DCM
(100 mL) and poured into a 0 ꢁC solution of NaHCO3
in water (150 mL). The organic phase was separated and
the solvent was removed. The crude 30-tert-butyl-4-dea-
cetyl-30-dephenyl-30-N-debenzoyl-4-O-methoxycarbonyl-
paclitaxel (16) was dissolved in DCM (15 mL) and a
saturated solution of NaHCO3 (15 mL) was added.
Cyclobutane carbonyl chloride (0.46 mL, 1.5 equiv) was
added with vigorous stirring. After 20 min, this was
diluted with ethyl acetate and the organic phase was
separated and dried (Na2SO4). Purification by pre-
parative reverse phase chromatography (step gradient
elution with 20–60% acetonitrile in water) afforded 15m
[39H, including 1.00 (s, 9H), 1.11 (s, 9H), 1.16 (s, 3H),
1.26 (s, 3H),1.69 (s, 3H), 1.91 (s, 3H), 2.26 (s, 3H)],
3.83 (d, 1H, J=7.1 Hz), 3.98 (s, 3H), 4.17 (d, 1H,
J=10.1 Hz), 4.26 (d, 1H, J=8.8 Hz), 4.37 (m, 2H),
4.55 (s, 1H), 5.00 (d, 1H, J=7.5 Hz), 5.73 (m, 2H),
6.02 (m, 1H), 6.29 (s, 1H), 7.45–8.13 (m, 5H); LRMS
(ESI) 844 [(M+H)+].
30-tert-Butyl-4-deacetyl-30-dephenyl-30-N-debenzoyl-30-N-
isobutylcarbonyl-4-O-methoxycarbonyl-paclitaxel (15n).
Following the above procedure with isobutyl-
chloroformate gave 15n: 1H NMR (CDCl3, 300 MHz) d
0.91 (dd, J=2.5 Hz, J=6.3 Hz, 6H), 1.08 (s, 9H), 1.15
(s, 3H), 1.25 (s, 3H), 1.68 (s, 3H), 1.90 (s, 3H), 2.07–2.00
(m, 3H), 2.25 (s, 3H), 2.45–2.39 (bm, 3H), 2.60–2.52
(bm, 2H), 3.28 (d, J=5.2 Hz, 1H), 3.82 (d, J=7.0 Hz,
1H), 3.99 (s, 3H), 4.43–4.13 (bm, 4H), 4.56 (dd, J=0.87
Hz, J=5.2 Hz, 1H), 5.00 (dd, J=2.0 Hz, J=9.5 Hz,
1H), 5.73 (d, J=7.0 Hz, 1H), 5.79 (d, J=10.1 Hz, 1H),
6.01 (m, 1H), 6.29 (s, 1H), 7.50–7.45 (m, 2H), 7.63–7.57
(m, 1H), 8.09 (d, J=7.3 Hz, 2H); 13C NMR (CDCl3) d
9.90, 13.88, 15.05, 21.09, 21.28, 22.38, 22.56, 26.26,
27.00, 27.54, 35.09, 35.59, 36.08, 45.81, 46.27, 56.11,
58.15, 58.44, 70.12, 72.16, 73.69, 75.12, 75.81, 76.24,
77.44, 78.95, 83.30, 84.29, 128.84, 129.48, 130.36,
133.37, 133.85, 142.55, 153.40, 166.94, 171.59, 172.61,
175.20, 203.81; LRMS (ESI): 830 [(M+H)+].
1
(1.47 g, 65%): H NMR (CDCl3, 300 MHz) d 1.06 (s,
9H), 1.14 (s, 3H), 1.24 (s, 3H), 1.66 (s, 3H), 1.88 (s, 3H),
1.92–1.76 (bm, 3H), 2.19–2.03 (bm, 4H), 2.23 (s, 3H),
2.58–2.36 (bm, 4H), 2.97 (p, J=7.9 Hz, 1H), 3.30 (d,
J=5.1 Hz, 1H), 3.80 (d, J=7.0 Hz, 1H), 3.98 (s, 3H),
4.14 (d, J=10.2 Hz, 1H), 4.22 (d, J=8.4 Hz, 1H), 4.40–
4.32 (bm, 2H), 4.55 (dd, J=1.1 Hz, J=5.2 Hz, 1H), 4.98
(dd, J=2.0 Hz, J=9.5 Hz, 1H), 5.69 (m, 2H), 5.99 (dd,
J=7.8 Hz, J=9.0 Hz, 1H), 6.27 (s, 1H), 7.48–7.43 (m,
2H), 7.62–7.55 (m, 1H), 8.08 (d, J=7.1 Hz, 2H); 13C
NMR (CDCl3, 75 MHz) d 9.72, 14.95, 18.29, 20.95,
22.12, 25.44, 26.85, 27.37, 35.47, 35.94, 39.91, 43.34,
45.74, 51.07, 56.03, 57.73, 58.36, 70.17, 72.06, 73.39,
74.94, 75.68, 76.09, 78.92, 83.21, 84.13, 128.72, 129.81,
130.22, 133.20, 133.74, 142.54, 153.21, 166.89, 171.46,
174.93, 174.99, 203.70; LRMS (ESI): 828.51 [(M+1)+,
References and Notes
1. Wani, M. C.; Taylor, H. L.; Wall, M. E.; Coggon, P.;
McPhail, A. T. J. Am. Chem. Soc. 1971, 93, 2325.

H. Mastalerz et al. / Bioorg. Med. Chem. 11 (2003) 4315–4323
4323
2. Rose, W. C., Taxol: science and applications. Suffness, M.
(ed.), CRC Press, Boca Raton, FL, 1995, 209.
mg/mL 1 h after being dosed orally in the mouse. Using this
value as a cut-off point for compounds of interest, the mouse
oral exposure assay was employed as a quick means of asses-
sing the potential oral bioavailability of taxane analogues that
were in our inventory as well as those that were generated in
the subsequent optimization studies.
3. (a) Luck, H. J.; Scholz, U.; Kuhnle, H.; Dethling, J.
Oncology 1999, 13 (Suppl. 3), 17. (b) Greco, F. A.; Thomas,
M.; Hainsworth, J. D. Cancer J. Sci. Am. 1999, 5, 179. (c)
Seidman, A. D. Semin. Oncol. 1999, 26 (Suppl. 8), 14.
4. Eiseman, J. L.; Eddington, N. D.; Leslie, J.; MacAuley, C.;
Sentz, D. L.; Zuhowski, M.; Kujawa, J. M.; Young, D.;
Egorin, M. Cancer Chemother. Pharmacol. 1994, 34, 465.
5. Rose, W. C.; Lee, F. Y. F.; Golik, J.; Kadow, J. Cancer
Chemother. Pharmacol. 2000, 46, 246.
6. Lipinski, C. A.; Lombardo, F.; Dominy, B. W.; Feeney,
P. J. Adv. Drug Deliv. Rev. 1997, 23, 3.
7. Van de Waterbeemd, H.; Smith, D. A.; Beaumont, K.;
Walker, D. K. J. Med. Chem. 2001, 44, 1313.
8. Sparreboom, A.; van Asperen, J.; Mayer, U.; Schinkel,
A. H.; Smit, J. W.; Meijer, D. K. F.; Borst, P.; Noojijen, W. J.;
Beijnen, J. H.; van Tellingen, O. Proc. Natl. Acad. Sci. U.S.A.
1997, 94, 2031.
9. Malingre, M. M.; Beijnen, J. H.; Schellens, J. H. M. Inves-
tig. New Drugs 2001, 19, 155.
10. (a) Vredenburg, M. R.; Ojima, I.; Veith, J.; Pera, P.; Kee,
F. C.; Sharma, A.; Kanter, P.; Greco, W. R.; Bernacki, R. J. J.
Natl. Can. Inst. 2001, 93, 1234. (b) Polizzi, D.; Pratesi, G.;
Monestiroli, S.; Tortoreto, M.; Zunino, F.; Bombardelli, E.;
Riva, A.; Morazzoni, P.; Colombo, T.; D’Incalci, M.; Zuc-
chetti, M. Clin. Cancer Res. 2000, 6, 2070. (c) Nicoletti, M. I.;
Colombo, T.; Rossi, C.; Monardo, C.; Stura, S.; Zucchetti,
M.; Riva, A.; Morazzoni, P.; Donati, M. B.; Bombardelli, E.;
D’Incalci, M.; Giavazzi, R. Cancer Res. 2000, 60, 842. (d)
Ferlini, C.; Distefano, M.; Pignatelli, F.; Lin, S.; Riva, A.;
Bombardelli, E.; Mancuso, S.; Ojima, I.; Scambia, G. Br. J.
Cancer 2000, 83, 1762. (e) Ojima, I.; Kuduk, S. D.; Pera, P.;
Veith, J. M.; Bernacki, R. J. J. Med. Chem. 1997, 40, 279. (f)
Ojima, I.; Slater, J. C.; Kuduk, S. D.; Takeuchi, C. S.; Gimi,
R. H.; Sun, C.-M.; Park, Y. H.; Pera, P.; Veith, J. M.; Ber-
nacki, R. J. J. Med. Chem. 1997, 40, 267. (g) Ojima, I.; Slater,
J. C.; Michaud, E.; Kuduk, S. D.; Bounaud, P.-Y.; Vrignaud,
P.; Bissery, M.-C.; Veith, J. M.; Pera, P.; Bernacki, R. J. J.
Med. Chem. 1996, 39, 3889. (h) Ojima, I.; Duclos, O.; Kun-
duk, S. D.; Sun, C.-M.; Slater, J. C.; Lavelle, F.; Veith, J. M.;
Bernacki, R. J. Bioorg. Med. Chem. Lett. 1994, 4, 2631.
11. Hansel, S. B.; Zoeckler, M. E.; McLaughlin, L. E.; AAPS
PharmSci 2001, 3, abstract. Our oral taxane program began
with a pilot study where a selected group of taxane analogues
that had been modified at one to three sites were screened in a
mouse oral exposure assay and then their %F in rats was
determined. For the oral exposure assay, analogues were dis-
solved in 10:10:80 Cremophor EL/ethanol/water and admi-
nistered orally to mice at a dose of 100 mg/kg. The drug
plasma exposure levels were determined after 1 h by HPLC
analysis and the reported values were the average obtained for
three mice. It was found that most of the compounds with
a%F>10 in the rat exhibited drug plasma levels greater than 4
12. (a) Chen, S.-H.; Wei, J.-M.; Long, B. H.; Fairchild, C. A.;
Carboni, J.; Mamber, S. W.; Rose, W. C.; Johnson, K.;
Casazza, A. M.; Kadow, J. K.; Farina, V.; Vyas, D. M.;
Doyle, T. W. Bioorg. Med. Chem. Lett. 1995, 5, 2741. (b)
Kadow, J. F.; Alstadt, T.; Chen, S.-H.; Dextraze, P.; Du, K.;
Fairchild, C.; Golik, J.; Hansel, S.; Johnston, K. A.; Kramer,
R. A.; Langley, D. R.; Lee, F.; Long, B.; Mastalerz, H.;
Ouellet, C.; Perrone, R.; Rose, W. C.; Scola, P.; Schulze, G.;
Staab, A.; Xue, Q.M; Walker, M.; Wei, J.-M.; Wittman, M.;
Wright, J. J. K.; Zoeckler, M.; Vyas, D. M. ACS Symposium
Series 2001, 796 (anticancer agents), 43. (c) Rose, W. C.;
Fairchild, C.; Lee, F. Y. Cancer Chem. Pharm 2001, 47, 97. (d)
Chen, S.-H. Frontiers of Biotech. Pharm 2002, 3, 157. (e) Kant,
J., Process for the Preparation of a Paclitaxel C-4 Methyl
Carbonate Analogue Patent Application, WO 0110856, 2001.
13. Portions of this work have been published in: Kadow, J.
F.; Mastalerz, H., Xue, Q. M.; Hansel, S.; Zoeckler, M. E.;
Rose, W. C.; Tarrant, J.G. Synthesis of C-4 Carbonate Tax-
anes and Therapeutic Use as Antitumor Agents for the Treat-
ment of Human or Veterinary Disease Patent Application,
WO 0156565, 2001.
14. (a) Holton, R. A.; Biediger, R. J.; Boatman, P. D., Taxol:
Science and Applications. Suffness, M. (ed.), CRC Press, Boca
Raton, FL, 1995, 97. (b) Holton, R. A.; Chen, S. H.; Vittorio,
F., 10-Deacetoxytaxol Derivatives, Patent Application, WO
9306093, 1993.
15. Holton, R. A.; Kim, H.-B.; Samoza, C.; Liang, F.; Biedi-
ger, J.; Boatman, D.; Shindo, M.; Smith, C. C.; Kim, S.;
Nadizadeh, H.; Suzuki, Y.; Tao, C.; Vu, P.; Tang, S.; Zhang,
P.; Murthi, K. K.; Gentile, L. N.; Jyanwei, H. L. J. Am. Chem.
Soc. 1994, 116, 1599. Ojima, I.; Habus, I.; Mangzhu, Z.;
Zucco, M.; Park, Y. H.; Sun, C. H.; Brigaud, T. Tetrahedron
1992, 48, 6985.
16. Lin, S.; Geng, X.; Qu, C.; Tynebor, R.; Gallagher, D. J.;
Pollina, E.; Rutter, J.; Ojima, I. Chirality 2000, 12, 431. Ojima,
I.; Slater, J. C. Chirality 1997, 9, 487.
17. Georg, I. G.; Boge, T. C.; Cheruvallath, Z. S.; Harriman,
G. C. B.; Hepperle, M.; Park, H.; Himes, R. H. Bioorg. Med.
Chem. Lett 1994, 4, 335.
18. In 10:10:80 Cremophor EL/ethanol/water the solutility of
2 was 12.9 mg/mL versus 0.3 mg/mL for paclitaxel (1).
19. Ali, S. M.; Hoemann, M. Z.; Aube, J.; Georg, G. I.; Mit-
scher, L. A.; McCall, R.; Jayasinghe, L. R. J. Med. Chem. 1995,
38, 3821. Ali, S. M.; Hoemann, M. Z.; Aube, J.; Georg, G. I.;
Mitscher, L. A.; Jayasinghe, L. R. J. Med. Chem. 1997, 40, 236.
20. Portions of these studies have been published: Rose, W. C.;
Long, B. H.; Fairchild, C. R.; Lee, F.Y .F.; Kadow, J. F. Clin.
Cancer Res. 2001, 7, 2016.