Design and syntheses of putative bioactive taxanes

Article abstract of DOI:10.1016/S0968-0896(02)00066-4

Reduction of 5α-hydroxy-7β,9α,10β-triacetoxy-4(20), 11(12)-taxadien-13-one 1 with activated zinc in glacial acetic acid led to rearranged products, including compounds with double bonds at C3-C4, C10-C11 or with an epoxide at C11-C12. Molecular modeling studies suggested that addition of a side chain at C-20 or C-5 of the taxanes with a C3-C4 double bond might lead to bioactivity. Semi-syntheses and results of bioactivities are discussed.

Full text of DOI:10.1016/S0968-0896(02)00066-4

Bioorganic & Medicinal Chemistry 10 (2002) 2387–2395
Design and Syntheses of Putative Bioactive Taxanes
Anastasia Nikolakakis,a Jian Hui Wu,b Gerald Batist,b Francoise Sauriol,c Orval Mamerd
and Lolita O. Zamira,b,
*
a
´ ´
Human Health Research Center, INRS-Institut Armand-Frappier, Universite du Quebec, 531Boulevard des Prairies, Laval,
´
Quebec, Canada, H7V 1B7
bMcGill Center for Translational Research in Cancer, Sir Mortimer B. Davis-Jewish General Hospital, 3755 Coˆte Ste. Catherine Rd.,
´
´
Suite D-127, Montreal, Quebec, Canada, H3T 1E2
cDepartment of Chemistry, Queen’s University, Kingston, Ontario, Canada, K7L 3N6
´
Biomedical Mass Spectrometry Unit, McGill University, 1130 Pine Avenue west, Montreal, Quebec, Canada, H3A 1A3
d
Received 6 September 2001; accepted 12 December 2001
Abstract—Reduction of 5a-hydroxy-7b,9a,10b-triacetoxy-4(20), 11(12)-taxadien-13-one 1 with activated zinc in glacial acetic acid
led to rearranged products, including compounds with double bonds at C3–C4, C10–C11 or with an epoxide at C11–C12. Mole-
cular modeling studies suggested that addition of a side chain at C-20 or C-5 of the taxanes with a C3–C4 double bond might lead
to bioactivity. Semi-syntheses and results of bioactivities are discussed. # 2002 Elsevier Science Ltd. All rights reserved.
Introduction
In this paper, we wanted to test this hypothesis by
building different taxane core skeletons with side chains
situated elsewhere than on C-13. We report the
Taxus canadensis is a low trailing shrub very common in
Quebec. Its composition of taxanes was shown to differ
from other yews.1ꢀ11 It is the only species producing
9-dihydro-13-acetylbaccatin III as the most abundant
taxane.1,2,11 In addition, taxinine and taxinine E are
found in larger amounts than other taxanes.8 Previous
work in our laboratory suggested that a major con-
formational change of the core skeleton of a taxane can
lead to unusual structure–activity-reactivities.7,8,12 For
example, taxuspine D, a taxane isolated from the
Canadian yew needles8 as well as from the Japanese
yew,13,14 was found to promote the polymerization of
tubulin with a potency corresponding to half of the
activity of paclitaxel.13 This result was surprising since
taxuspine D lacked all the key features essential for
bioactivity: a C-13 side chain with a 20-OH, a benzoyl
group on C-2 and an oxetane on C4-C5.15ꢀ17 We
explained this result by molecular modeling studies.12
The C12–C13 double bond caused a substantial change
in the conformation of the core skeleton, and the C-5
cinnamoyl in taxuspine D is found to mimic part of the
C-13 side chain of paclitaxel.
rearrangement reactions of
a taxinine derivative:
5a-hydroxy-7b,9a,10b-triacetoxy-4(20), 11(12)-taxadien-
13-one 1 when treated with activated zinc in glacial
acetic acid. We obtained taxanes with C3–C4, or C10–
C11 double bonds as well as a C11–C12 epoxide.
Molecular modeling studies, semi-syntheses, and bioac-
tivities of the C3–C4 double bond taxanes with C-5 or
C-20 side chains are reported.
Results and Discussion
Rearrangement of 5ꢀ-hydroxy-7ꢁ, 9ꢀ, 10ꢁ-triacetoxy-
4(20), 11(12)-taxadien-13-one, 1 (Scheme 1)
Initially, we intended to prepare taxinine derivatives
with C12–C13 double bonds. We, therefore, used the
reaction which had been successfully reported for tax-
anes.18 We were surprised to obtain unusual taxinine
derivatives, none of which had a C12–C13 double bond.
Indeed, treatment of 119,20 with activated Zn in glacial
acetic acid gave three major peaks on the analytical
HPLC. Two pure compounds were isolated upon pur-
ification by semi-preparative HPLC and were identified
as compounds 2 (13%) and 3 (21%). 1H NMR spectro-
scopy of the third peak isolated proved to be a mixture
*Corresponding author. Tel.: +1-450-687-5010-4260; fax: (office) +1-
450-686-5301, (residence) +1-514-481-2797; e-mail: lolita.zamir@
inrs-iaf.uquebec.ca
0968-0896/02/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(02)00066-4

2388
A. Nikolakakis et al. / Bioorg. Med. Chem. 10 (2002) 2387–2395
of three compounds. These peaks could not be purified
either by HPLC or preparative TLC. Cinnamoylation
proved useful as we were able to purify the cinnamoy-
lated products. Upon purification by semi-preparative
HPLC, three compounds were isolated and identified as
compounds 4 (37%), 5 (19%), and 6 (38%). This reac-
tion was repeated several times and depending on the
length of the reaction, the ratio between compounds 2–6
could be changed. Upon treatment of 1 with Zn in gla-
cial acetic acid for a longer period of time (42 h com-
pared to 25 h) the amounts of 2 and 3 were decreased
and we were able to isolate 5 as the major compound
The downfield shift of H-5 at d 4.26 in 1 to d 5.43 in 6
confirmed the positioning of the cinnamoyl group at C-
5. Only two acetyls were observed and they are located
on C-7 and C-9, as confirmed by HMBC correlations of
H-7 (d 5.64) and H-9 (d 5.57) with carbonyl acetyls d
170.1 and d 170.1, respectively. The disappearance of
the H-10 doublet at d 6.32 in 1 and the appearance of
the aliphatic protons H-10a and H-10b in 6 at d 3.37
and d 2.53 is consistent with deacetylation at C-10.
When comparing the NMR data of 4 to 6 we observe
that the major differences occur in ring A where the
double bond between C-11 and C-12 in 6 has been
replaced by an epoxide in 4. Me-18 is correlated to two
relatively deshielded carbons, C-11 (d 65.2) and C-12 (d
59.4) and to a carbonyl, C-13 (d 210.5). The NOE
correlations between H-3, H-7 and Me-18 confirm the a
orientation of Me-18. The NMR spectral data as well as
the HMBC and NOESY correlations of taxanes 2, 4
and 6 are shown in Tables 1–3.
1
from the mixture. The H NMR spectra of compounds
2 and 3 were very similar and only differed in the signals
of H-9, H-10, and H-11. The chemical shift of H-9
changed from d 6.20 in 2 to d 5.90 in 3. The C-10 proton
doublet at d 5.46 in 2 was replaced by two aliphatic
proton signals at d 2.22 and d 1.72 in 3. The chemical
shift of H-11 changed from d 1.64 in 2 to d 1.46 in 3.
The upfield changes to these proton signals and the
appearance of the two aliphatic protons confirm the
deacetylation of 2 at C-10. This was also confirmed by
the high-resolution mass spectrometry of the potassium
adducts: m/e 533.2154 for 2 and m/e 475.2099 for 3, a
difference of 58 units confirming the loss of an acetyl
group. The downfield shift of C-3 at d 34.8 in 1 to d
140.6 in 2 or d 141.7 in 3 and the upfield shift of C-20 at
d 111.3 in 1 to d 64.4 in 2 and d 64.8 in 3 confirm the
shift of the double bond from C4–C20 to C4–C3. The
upfield shift of C-11 at d 151.2 in 1 to d 54.0 in 2 or d
48.3 in 3 as well as the upfield shift of C-12 at d 138.5 in
1 to d 44.0 in 2 or d 47.3 in 3 confirm the disappearance
of the C12–C13 double bond. The new H-12 proton at d
2.58 for 2 or d 2.55 for 3 had a strong NOE correlation
with H-7, thereby confirming its a-orientation, and
indicating that Me-18 is b. The new H-11 proton at d
1.64 for 2 or d 1.46 for 3 had a strong correlation with
Me-18, thereby confirming its b-orientation.
The chemistry of taxanes is unusual and we did not
predict these rearrangements. Once the structures were
confirmed, mechanisms could be postulated. For exam-
ple, the conversion of compound 1 to 2 involved attack
of OAcꢀ on C-20 followed by migration of the C4–C20
double bond to its more stable tetrasubstituted C3–C4.
The transfer of the C3 hydride to C-11 was aided by the
U-configuration of the taxane. The C11–C12 double
bond picked up H+ to give taxane 2. Taxane 3 was
obtained from additional reduction of compound 2.
Similarly, the formation of compound 5 can be easily
rationalized by a hydride attack on C12 of taxane 1
triggering migration of the C12–C11 double bond to
C11–C10 and elimination of OAcꢀ. Taxane 6 was
obtained by reduction of 1. We were puzzled by the
transformation of taxane 1 to 4. The existence of the
epoxide is unambiguously confirmed: two quaternary
carbons at C11-C12 from the 13C NMR and the extra
oxygen atom in the FAB-HR-MS. The only possibility
is the transfer of an oxygen either from the neighboring
acetyl group or from the acetic acid.
The characterization of 5, albeit without the cinnamoyl
group at C-5, was discussed in our previous publication.20
Scheme 1. (a) Zn, glacial acetic acid, 23 ꢁC, 25 h, taxane 2 13%, taxane 3 21%; (b) (i) Zn, glacial acetic acid, 23 ꢁC, 25 h; (ii) trans-cinnamic acid,
DCC, 4-DMAP, toluene, 90 ꢁC, 5 h, taxane 4 37%, taxane 5 19%, taxane 6 38%.

A. Nikolakakis et al. / Bioorg. Med. Chem. 10 (2002) 2387–2395
2389
Semi-syntheses of taxanes 12, 13, 15 (Schemes 2 and 3)
compounds 2 or 3 with the docetaxel side chain, it was
first necessary to protect the primary C-20 alcohol.
Acetylation of compounds 2 and 3 with acetyl chloride in
pyridine, afforded compounds 7 (59% yield) and 8 (36%
yield), respectively. Esterification of the free C5-OH was
The semi-syntheses of compounds 12 and 13 (Scheme 2)
were very similar and provided compounds having a
C-5 side chain. In order to esterify the C5-OH of
Table 1. 1H and 13C NMR for taxane 2 in CDCl3
Position
d 1H mult.a J (J in Hz)
d
13Cb
HMBC
NOESYc
1
2.05 o.m
2.80 dd (15.3, 6.5)
2.65 o.m
—
44.0
26.7
2as, 14as, 16s, 17s, 19s, (5s, 6bs)
1w, 2b/14as, 20as
2a
2b
3
4
5
6a
6b
7
8
9
10
11
12
13
14a
14b
15
16
17
18
19
20a
20b
OAc
OAc
OAc
1, 3, 4, 8
140.6
134.5
67.7
—
4.39 br. s
2.04 o.m
1.93 o.m
5.81 dd (12.0, 4.6)
—
6.20 d (11.0)
5.46 d (11.0)
1.64 d (8.0)
2.58 qu (7.0)
—
2.66 o.m
2.51 d (19.9)
—
0.91 s
1.48 s
1.27 d (6.4)
1.18 s
6am, 20bm
See 1
5w, 6aw, 7w
6bw, 10s, 12s
33.2
66.3
46.2
75.7
72.9
54.0
44.0
215.2
39.2
6, 9, 19, 169.0
3, 7, 8, 10, 19, 169.9
9, 15, 168.5
17s, 19m, 2b/14as
7s, 11s, 12s, 18s
10s, 16s, 18s
7s, 10m, 18s
1, 2, 13
1, 2, 13, 15
14as, 20as
36.5
35.0
25.8
15.0
22.6
64.4
1, 11, 15, Me-17
1, 11, 15, Me-16
11, 12, 13
3, 7, 8, 9
3, 4, 5
3, 4, 5
169.0 (C-7)
169.9 (C-9)
168.5 (C-10)
1m, 14am, 11m, 17s
1m, 9s, 16s
10s, 11s, 12s
1s, 9s, 2bw
4.52 d (11.6)
4.20 d (11.6)
2.06 s
2as, 14bm, 20bs
5s, 20as
21.1
20.5
20.9
2.02 s
1.97 s
aMult. multiplicity: s, singlet; d, doublet; qu, quintet; dd, doublet of doublet; br, broad; m, multiplet; o, overlapping. The precision of the coupling
constants is Æ0.5 Hz.
bThe 13C chemical shifts were extracted from the HSQC and HMBC experiments (for quaternary carbons) (Æ0.2 ppm).
cNOESY intensities are marked as strong (s), medium (m) or weak (w).
Scheme 2. (a) Acetyl chloride, pyridine, 4 ꢁC, 18 h, taxane 7 59%, taxane 8 36%; (b) 9, DCC, 4-DMAP, CH2Cl2–toluene, 75 ꢁC, 5 h, taxane 10 (C-10
R=OAc, side chain R=4-OMe-Ph) 29%, taxane 11 (C-10 R=H, side chain R=4-OMe-Ph) 28%; (c) p-toluenesulfonic acid, methanol, 23 ꢁC,
taxane 12 57%, taxane 13 58%.

2390
A. Nikolakakis et al. / Bioorg. Med. Chem. 10 (2002) 2387–2395
realized with 2-(4-OMe)phenyl-1,3-oxazolidine of
N-Boc phenylisoserine (9),21 dicyclohexylcarbodiimide
and 4-(dimethylamino)pyridine, in dichloromethane–
toluene at 75 ꢁC affording diastereomers 10 (29%) and
11 (28%), respectively. Deprotection of 10 and 11 with a
catalytic amount of p-toluenesulfonic acid in methanol
afforded compounds 12 (57%) and 13 (58%), respec-
tively. The major differences in the NMR spectral data
between compounds 2 and 12 occur in ring C. A fourth
acetyl group is observed in 12 and is located on C-20 as
Table 2. 1H and 13C NMR for taxane 4 in CDCl3
Position
d 1H mult.a J (J in Hz)
d
13Cb
HMBC
NOESYc
1
1.99 o.m
1.98/1.95 o.m
2.75 o.m
—
42.1
24.5
36.9
2a/2b
3
4
2w, 7s, 14bs, 16m, 18m (o.14a)
20as
5
6a
5.60 br.dd (4.0, 1.6)
2.07 o.m
74.8
33.6
6b
7
8
1.91 o.m
5.75 dd (11.3, 5.2)
—
5.66 dd (12.3, 4.5)
2.16 dd (14.9, 4.5)
2.09 o.dd (14.9, 12.3)
68.4
47.0
77.2
30.9
19, 170.4
8, 170.4
3w, 6aw, 18w
9
2/10s, 17s, 19m
10a
10b
11
12
13
14a
14b
15
16
17
—
—
—
65.2
59.4
210.5
41.2
2.72 dd (19.4, 8.5)
1.90 o.d (19.4)
—
2, 12, 13
See H-3
40.6
28.2
23.9
14.7
12.8
116.6
0.78 s
1.58 s
1.98 s
0.88 s
5.45 s
5.10 s
2.05 s
2.05 s
—
1, 11, 15, Me-17
1, 11, 15, Me-16
11, 12, 13
3, 7, 8, 9
5
3, 5
170.4 (C-7)
170.4 (C-9)
1m, 14am, 17s
1m, 9s, 10aw, 16s
18
19
6b/2s, 9m, 20bw
5s, 20bs
20a
20b
OAc
OAc
10(CO)
20 CH¼
30¼CH
40-Ph
o
2m, 20as
21.1
21.1
165.9
117.1
146.7
134.1
128.0
129.3
130.0
6.25 d (15.7)
7.62 d (15.7)
—
134.1, 165.9
Me-18 (overlap)
7.59 br.d (ꢂ7.3)
m/p
7.41 m
aMult. multiplicity: s, singlet; d, doublet; dd, doublet of doublet; br, broad; m, multiplet; o, overlapping. The precision of the coupling constants is
Æ0.5 Hz.
bThe 13C chemical shifts were extracted from the HSQC and HMBC experiments (for quaternary carbons) (Æ0.2 ppm).
cNOESY intensities are marked as strong (s), medium (m) or weak (w).
Scheme 3. (a) 9, DCC, 4-DMAP, CH2Cl2–toluene, 4 ꢁC, 18 h, taxane 14 (side chain R=4-OMe-Ph) 12%; (b) p-toluenesulfonic acid, methanol,
23 ꢁC, 18 h, taxane 15 58%.

A. Nikolakakis et al. / Bioorg. Med. Chem. 10 (2002) 2387–2395
2391
confirmed by HMBC correlations of H-20 (d 5.12) with
Molecular modeling of taxanes
the carbonyl acetyl at d 170.0. The downfield shift of
H-5 at d 4.39 in 2 to d 5.69 in 12 confirmed the posi-
tioning of the docetaxel group at C-5. Similar compar-
isons can be made between compounds 3 and 13. A
third acetyl group is observed in 13 and is located on C-
20 as confirmed by HMBC correlations of H-20 (d 5.14)
with the carbonyl acetyl at d 170.3. The downfield shift
of H-5 at d 4.39 in 3 to d 5.73 in 13 confirmed the posi-
tioning of the docetaxel group at C-5. The NMR spec-
tral data as well as the HMBC and NOESY correlations
of taxane 12, are shown in Table 4.
In an attempt to rank the priority of compounds to be
synthesized, we built a database of 28 virtual com-
pounds, including the known bioactive compounds
paclitaxel, docetaxel and taxuspine D. The database was
docked against the refined structure of b-tubulin12 using
the DOCK program.22 According to the intermolecular
energy dock score (Table 5), compounds 12 (ꢀ38.4 kcal/
mol) and 13 (ꢀ36.1 kcal/mol) were predicted to be
bioactive since their dock scores were similar to that of
docetaxel (ꢀ37.3 kcal/mol). In addition, compounds 15
(ꢀ33.5 kcal/mol) and a virtual analogue of compound 5
(Scheme 1) wherein the cinnamoyl group had been
replaced by the paclitaxel side chain, compound 16
(ꢀ33.0 kcal/mol) had a similar dock score to taxuspine
D (ꢀ33.8 kcal/mol) and, therefore, were predicted to be
bioactive. However, after compounds 12, 13, and 15
were synthesized, the tubulin assembly assay showed no
bioactivity at 10 mM concentrations.23 This prompted us
to investigate the reason for the false positives in the
docking study. It is well known that solvation free
energies are omitted in the DOCK scoring function
although they can be taken into account implicitely.22
We therefore calculated the solvation free energy using
the AM1-SM5.4PDA model.24 In order to speed up the
The semi-synthesis of compound 15 from 3 afforded a
compound with a C-20 side chain. We postulated that
the primary C-20 alcohol of 3 would be more reactive
than the secondary alcohol at C-5 especially at lower
temperatures. Direct esterification of 3 was realized with
2-(4-OMe)phenyl-1,3-oxazolidine of N-Boc phenyliso-
serine (9),21 dicyclohexylcarbodiimide and 4-(dimethy-
lamino)pyridine, in dichloromethane–toluene at 4 ꢁC
affording diastereomers 14 (12%). Deprotection of 14
with a catalytic amount of p-toluenesulfonic acid in
methanol afforded compound 15 (58%). The downfield
shift of H-20 at d 4.20 in 3 to d 4.80 in 15 confirmed the
positioning of the docetaxel group at C-20.
Table 3. 1H and 13C NMR for taxane 6 in CDCl3
Position
d 1H mult.a J (J in Hz)
d
13Cb
HMBC
NOESYc
1
2a
2b
3
4
5
6a
6b
7
8
2.23 br.t (5.7)
1.95 o.m
1.85 o.m
3.22 br.d (4.4)
—
5.43 br.t (ꢂ2.9)
2.07 o.m
1.79 o.m
5.64 dd (11.7, 5.5)
—
40.1
25.4
2aw, 2bm, 14as, 16s, 17s
See 14a
1w, 2aw, 19w (see 6b)
7m, 2bm
36.3
147.1
74.6
34.0
6aw, 6bw, 20as
5s, 6bs, 7s
5s, 6as, 7m, 19w (see 2b)
3s, 6as, 10s, 18s
69.7
48.0
76.5
33.1
6, 8, 19, 170.1
9
5.57 dd (11.7, 4.7)
3.37 t (12.3)
2.53 dd (12.7, 4.9)
—
7, 8, 10, 19, 170.1
9, 11, 12, 16
2as, 10bs, 17s, 19m
7m, 10bs, 18s
9m, 10as, 17m
10a
10b
11
12
13
14a
14b
15
16/17
17/16
18
157.2
135.1
200.0
39.7
—
—
2.91 dd (19.4, 7.0)
1.94 o.d (19.4)
—
1w, 14bs, 16w, 19w, 20bw
14aw
1/15, 2, 13
40.1
36.6
24.6
13.9
12.9
114.2
1.16 s
1.57 s
2.19 s
0.83 s
5.26 s
4.91 s
2.07 s
2.05 s
—
1/15, 11, Me
1/15, 11, Me
11, 12, 13
3, 7, 8, 9
3, 5
3, 5, 4 (weak)
170.1
170.1
1m, 17s
1m, 2am, 9s, 10bm, 16s
3w, 7s, 10as, 30m
9w
19
20a
20b
OAc
OAc
10 (CO)
20 CH¼
30¼CH
40 Ph
o
5s, 20bs
2bs, 20as
21.2
20.9
166.2
118.1
145.6
134.4
128.6
128.9
130.4
6.44 d (16.1)
7.63 d (16.1)
—
7.74 d (7.7)
7.43 m
7.38 m
134.4, 166.2
128.6, 118.1
Me-18
145.6, 130.4
m
p
aMult. multiplicity: s, singlet; d, doublet; t, triplet; dd, doublet of doublet; br, broad; m, multiplet; o, overlapping. The precision of the coupling
constants is Æ0.5 Hz.
bThe 13C chemical shifts were extracted from the HSQC and HMBC experiments (for quaternary carbons) (Æ0.2 ppm).
cNOESY intensities are marked as strong (s), medium (m) or weak (w).

2392
A. Nikolakakis et al. / Bioorg. Med. Chem. 10 (2002) 2387–2395
Table 4. 1H and 13C NMR for taxane 12 in CDCl3
Position
d 1H mult.a J (J in Hz)
d
13Cb
HMBC
NOESYc
1
2a
2b
3
4
5
6a/b
7
8
9
10
11
12
2.14 o.t (7.4)
2.99 dd (15.5, 7.2)
2.76 d (15.8)
—
44.2
27.5
2am, 14aw, 16m, 17s
1w, 2bs, 20as
2as, 9s, 17m, 19w
147.6
127.9
68.4
30.7
65.7
46.2
75.5
73.1
53.3
44.0
217.5
38.8
—
5.69 br. s
2.0 o.m
5.74 t (8.8)
—
6w
5s, 7w, 19s
6s, 10s, 12s
6.19 d (11.2)
5.46 br.d (11.2)
1.70 br.d (6.5)
2.29 qu (7.0)
—
2.66 br.dd (20.3, 7.3)
2.02 o.m
—
7, 8, 10, 169.9
2bs, 17s, 19m,
7s, 11s, 12w, 18w
10w, 16s, 18s
7s, 10m, 18s
13
14a
14b
15
16
17
1m, 14bs, 16s
14as, 20aw
36.2
34.9
26.1
17.2
22.8
61.7
0.92 s
1.48 s
1.31 br.d
1.18 s
1, 11, 15, Me-17
1, 11, 15, Me-16
1w, 11s, 17s
1w, 2bw, 9s
10m, 11s, 12w
6w, 9w
18
19
3, 7, 8, 9
3, 4, 5, 170.0
20a
20b
20
5.12 d (11.9)
4.01 br.d (11.9)
4.52 br
2aw, 14bw, 20bs
20as
74.4
56.9
30w
OH-20
30
3.81 br
5.09 o.br
30-Ph
o
7.43 d (7.5)
7.35 t (7.8)
7.26 o.t
126.5
128.5
127.3
m
p
40(NH)
80 t-Bu(Me)
OAc
OAc
OAc
OAc
5.81 br.d (8.9)
1.39 br
2.06 s
28.3
21.2
20.6
20.55
21.0
2.04 s
2.02 s
2.98 s
aMult. multiplicity: s, singlet; d, doublet; t, triplet; qu, quintet; dd, doublet of doublet; br, broad; m, multiplet; o, overlapping. The precision of the
coupling constants is Æ0.5 Hz.
bThe 13C chemical shifts were extracted from the HSQC and HMBC experiments (for quaternary carbons) (Æ0.2 ppm).
cNOESY intensities are marked as strong (s), medium (m) or weak (w).
process, only single point calculations were performed
on the docked conformations. Consequently, the change
of conformational energy on binding were omitted. The
results were reported in Table 5. Our results clearly
show that solvation free enegies are an important factor
in ligand binding. As indicated by the corrected dock
scores, compounds 12 (ꢀ7.4 kcal/mol), 13 (ꢀ8.1 kcal/
mol) and 15 (ꢀ7.1 kcal/mol) should be inactive since
their corrected dock scores were much lower than that
of taxuspine D (ꢀ12.3 kcal/mol). Interestingly, the
ranking of taxane 16 (ꢀ14.5 kcal/mol), remained above
that of taxuspine D after the correction of the solvation
free energy. This virtual compound has a double bond
at C10–C11 and a paclitaxel side chain at C-5. The
synthesis of this compound is in progress.
Conclusion
A multidisciplinary approach combining bioorganic
mechanisms, synthetic chemistry and molecular model-
ing led to interesting taxane structures with double
bonds at C3–C4, C10–C11 or with a C11–C12 epoxide.
Addition of side chains situated elsewhere than on C-13
produced very different conformations which may act
on different cell targets.
Table 5. Docks intermolecular energy score, the solvation free energy
(ÁGsolv) and the corrected dock score (kcal/mol)
Compounds
Dock score
ÁGsolv
Corrected dock score
Paclitaxel
ꢀ44.4
ꢀ37.3
ꢀ33.0
ꢀ33.8
ꢀ38.4
ꢀ36.1
ꢀ33.5
ꢀ26.1
ꢀ22.5
ꢀ18.5
ꢀ21.5
ꢀ31.0
ꢀ28.0
ꢀ26.4
ꢀ18.3
ꢀ14.8
ꢀ14.5
ꢀ12.3
ꢀ7.4
Docetaxel
Taxane 16a
Taxuspine D
Taxane 12
Taxane 13
Taxane 15
Experimental
ꢀ8.1
ꢀ7.1
Instrumentation
aA virtual analogue of compound 5 (Scheme 1) wherein the cinnamoyl
group had been replaced by the paclitaxel side chain.
The instrumentation used was as previously described.20,25

A. Nikolakakis et al. / Bioorg. Med. Chem. 10 (2002) 2387–2395
2393
Rearrangement of 5ꢀ-hydroxy-7ꢁ,9ꢀ,10ꢁ-triacetoxy-
4(20),11(12)-taxadien-13-one, 1. 5a-Hydroxy-7b,9a,10b-
triacetoxy-4(20),11(12)-taxadien-13-one 1 (69 mg, 0.15
mmol) was dissolved in glacial acetic acid (0.89 mL) and
treated with activated Zn18 (0.83 g, 12.7 mmol) at 23 ꢁC
for 25 h. The mixture was diluted with EtOAc and fil-
tered through Celite. Heptane was added and the mix-
ture was evaporated under reduced pressure. The
residue was purified by semi-preparative HPLC (25–
100% CH3CN in H2O, 50 min gradient at a flow rate of
3 mL/min) affording 2 (9 mg, 13%, tR=24.76 min), 3
(13 mg, 21%, tR=27.38 min) and a mixture [11 mg,
18% (based on the average molecular weight of the
three compounds before cinnamoylation), tR=37.89
min]. The mixture (4 mg, 0.010 mmol) was dissolved in
toluene (0.5 mL) and was treated with trans-cinnamic
acid (8 mg, 0.054 mmol), DCC (11 mg, 0.053 mmol) and
DMAP (6 mg, 0.049 mmol) at 90 ꢁC for 5 h. The solu-
tion was diluted with EtOAc, washed with a saturated
NaHCO3 solution and brine. The organic phase was
dried, filtered and evaporated. The residue was purified
by semi-preparative HPLC (25–100% CH3CN in H2O,
50 min gradient at a flow rate of 3 mL/min) affording 4
(2 mg, 37%, tR=52.00 min), 5 (1 mg, 19%, tR=53.50
min) and 6 (2 mg, 38%, tR=54.75 min). The NMR data
of 2 is shown in Table 1. FAB-HR-MS for 2:
C26H38O9K [M+K]+ requires: 533.2153; found:
533.2154; 1H NMR (500 MHz, CDCl3) for 3: d 5.90 (dd,
J=12.6, 3.4 Hz, 1H, H-9), 5.82 (dd, J=11.9, 4.7 Hz,
1H, H-7), 4.54 (d, J=11.6 Hz, 1H, H-20a), 4.39 (br.dd,
J=4.1, 2.0 Hz, 1H, H-5), 4.20 (d, J=11.6 Hz, 1H, H-
20b), 2.74 (dd, J=15.6, 6.6 Hz, 1H, H-2a), 2.65 (dd,
J=19.8, 8.3 Hz, 1H, H-14a), 2.58 (o.d, J=15.6 Hz, 1H,
H-2b), 2.55 (qu, J=6.8 Hz, 1H, H-12), 2.48 (d, J=19.5
Hz, 1H, H-14b), 2.22 (br.t, 1H, H-10a), 2.07 (s, 3H,
OAc), 2.04 (o.m, 1H, H-1), 2.04 (o.m, 1H, H-6a), 2.02
(s, 3H, OAc), 1.90 (ddd, J=13.3, 4.2, 1.6 Hz, 1H, H-
6b), 1.72 (ddd, J=14.5, 7.4, 4.3 Hz, 1H, H-10b), 1.46
(o.m, 1H, H-11), 1.39 (s, 3H, H-17), 1.13 (o.d, 3H, H-
18), 1.13 (o.s, 3H, H-19), 0.91 (s, 3H, H-16); 13C NMR
(125 MHz, CDCl3) for 3: d 217.2 (C-13), 141.7 (C-3),
134.1 (C-4), 77.7 (C-9), 68.1 (C-5), 66.4 (C-7), 64.8 (C-
20), 48.3 (C-11), 47.3 (C-12), 46.1 (C-8), 43.9 (C-1), 39.8
(C-14), 37.0 (C-15), 34.7 (C-16), 33.4 (C-6), 31.9 (C-10),
26.7 (C-2), 25.0 (C-17), 22.4 (C-19), 21.6 (OAc), 20.2
(OAc), 15.3 (C-18). FAB-HR-MS for 3: C24H36O7K
[M+K]+ requires: 475.2098; found: 475.2099. The
NMR data for 4 is shown in Table 2; FAB-HR-MS for
4: C33H40O8K [M+K]+ requires: 603.2360; found:
603.2359. The NMR data for 5 was identical to litera-
ture;20 FAB-HR-MS for 5: C33H40O7K [M+K]+
requires: 587.2411; found: 587.2410. The NMR data for
6 is shown in Table 3; FAB-HR-MS for 6: C33H40O7K
[M+K]+ requires: 587.2411; found: 587.2410.
(0.2 mL) was treated with the chiral acid 9 (4.8 mg,
0.012 mmol) in 0.1 mL CH2Cl2. A catalytic amount of
4-DMAP and DCC (2.5 mg, 0.012 mmol) were added
and the solution was stirred at 75 ꢁC for 5 h. The reac-
tion mixture was diluted with CH2Cl2, washed with
brine, dried, filtered and evaporated. The residue was
purified by semi-preparative HPLC (25–100% CH3CN
in H2O, 50 min gradient at a flow rate of 3 mL/min)
affording taxane 10 (2 mg, 29%, tR=50.01 min). The
protected taxinine ester 10 (2 mg, 0.002 mmol) in
MeOH (1 mL) was treated with a catalytic amount of p-
toluenesulfonic acid (PTSA) at 23 ꢁC for 3 h. The solu-
tion was diluted with EtOAc, washed with brine to
neutrality, dried, filtered and evaporated. The residue
was purified by semi-preparative HPLC (25–100%
CH3CN in H2O, 50 min gradient at a flow rate of 3 mL/
min) affording taxane 12 (1 mg, 57%, tR=44.16 min).
The NMR data for 12 is shown in Table 4. FAB-HR-
MS: C42H57N1O14K [M+K]+ requires: 838.3416;
found: 838.3412.
Taxane 13. Taxane 3 (10 mg, 0.023 mmol) dissolved in
pyridine (0.5 mL) was treated with acetyl chloride (11
mL, 0.16 mmol) at 4 ꢁC for 18 h. The reaction was
quenched with brine and extracted with dichloro-
methane. The organic phase was dried, filtered and
evaporated. The residue was purified by semi-pre-
parative HPLC (25–100% CH3CN in H2O, 50 min gra-
dient at a flow rate of 3 mL/min) affording taxane 8 (4
mg, 36%, tR=34.47 min). Taxinine 8 (4 mg, 0.008
mmol) in toluene (0.2 mL) was treated with the chiral
acid 9 (5.3 mg, 0.013 mmol) in 0.1 mL CH2Cl2. A cata-
lytic amount of 4-DMAP and DCC (2.8 mg, 0.014
mmol) were added and the solution was stirred at 75 ꢁC
for 5 h. The reaction mixture was diluted with CH2Cl2,
washed with brine, dried, filtered and evaporated. The
residue was purified by semi-preparative HPLC (25–
100% CH3CN in H2O, 70 min gradient at a flow rate of
3 mL/min) affording taxane 11 (2 mg, 28%, tR=67.22
min). The protected taxinine ester 11 (2 mg, 0.002
mmol) in MeOH (1 mL) was treated with a catalytic
amount of p-toluenesulfonic acid (PTSA) at 23 ꢁC for 18
h. The solution was diluted with EtOAc, washed with
brine to neutrality, dried, filtered and evaporated. The
residue was purified by semi-preparative HPLC (25–
100% CH3CN in H2O, 50 min gradient at a flow rate of
3 mL/min) affording taxane 13 (1 mg, 58%, tR=47.31
min). 1H NMR (500 MHz, CDCl3) d 7.44 (d, J=7.7 Hz,
2H, H-2,6 of OBz), 7.35 (t, J=7.7 Hz, 2H, H-3,5 of
OBz), 7.26 (o.t, 1H, H-4 of OBz), 5.87 (dd, J=15.5, 4.0
Hz, 1H, H-9), 5.83 (br.d, J=9.0 Hz, 1H, NH-40), 5.75
(o.dd, J=11.2, 5.2 Hz, 1H, H-7), 5.73 (o.br.s, 1H, H-5),
5.14 (o.m, 1H, H-30), 5.14 (o.d, J=11.9 Hz, 1H, H-20a),
4.51 (br.s, 1H, H-20), 4.02 (br.d, J=11.7 Hz, 1H, H-
20b), 3.75 (br.s, 1H, OH-20), 2.92 (dd, J=15.2, 6.9 Hz,
1H, H-2a), 2.69 (d, J=15.5 Hz, 1H, H-2b), 2.66 (o, 1H,
H-14a), 2.28 (qu, J=6.3 Hz, 1H, H-12), 2.23 (br.m, 1H,
H-10a), 2.14 (o.br.t, 1H, H-1), 2.10 (o.m, 1H, H-6a),
2.08 (s, 3H, OAc), 2.04 (s, 3H, OAc), 2.03 (s, 3H, OAc),
1.97 (o.m, 1H, H-14b), 1.77 (ddd, J=14.8, 7.3, 4.3 Hz,
1H, H-10b), 1.53 (o.m, 1H, H-11), 1.40 (o.s, 3H, H-17),
1.40 (o.s, 9H, t-Bu), 1.18 (br.d, J=6.4 Hz, 3H, H-18),
1.13 (s, 3H, H-19), 0.92 (s, 3H, H-16); 13C NMR
Taxane 12. Taxane 2 (14 mg, 0.028 mmol) dissolved in
pyridine (0.5 mL) was treated with acetyl chloride (10
mL, 0.14 mmol) at 4 ꢁC for 18 h. The reaction was
quenched with brine and extracted with dichloro-
methane. The organic phase was dried, filtered and
evaporated. The residue was purified by flash chroma-
tography (CH2Cl2/acetone, 8:2) to give compound 7 (9
mg, 59%). Taxinine 7 (4 mg, 0.008 mmol) in toluene

2394
A. Nikolakakis et al. / Bioorg. Med. Chem. 10 (2002) 2387–2395
(125 MHz, CDCl3) d 217.2 (C-13), 170.3 (CO–OAc),
169.9 (CO–OAc), 169.2 (CO–OAc), 148.5 (C-3), 128.5
(C-3,5 of OBz), 127.3 (C-4 of OBz), 127.2 (C-4), 126.6
(C-2,6 of OBz), 77.4 (C-9), 74.4 (C-20), 68.7 (C-5), 65.6
(C-7), 61.8 (C-20), 56.7 (C-30), 47.5 (C-11), 47.0 (C-12),
46.0 (C-8), 43.8 (C-1), 39.3 (C-14), 36.6 (C-15), 34.4 (C-
16), 31.9 (C-10), 30.6 (C-6), 28.1 (t-Bu), 27.3 (C-2), 24.7
(C-17), 22.4 (C-19), 21.4 (OAc), 21.0 (OAc), 20.6 (OAc),
17.2 (C-18). FAB-HR-MS: C40H55N1O12K [M+K]+
requires: 780.3361; found: 780.3364.
ations saved at each cycle of pruning was set to 30 and
the number of anchor positions to explore was set at
300. The solvation of free energy was calculated by
performing single-point calculations on the docked
conformations using the AMI-SM5. 4PDA model as
implemented in the AMSOL program.24 The correlated
dock scores were obtained by subtracting the solvation
free energy from the dock scores.
Acknowledgements
Taxane 15. Taxinine 3 (9 mg, 0.021 mmol) in toluene
(0.4 mL) was treated with the chiral acid 9 (13 mg, 0.033
mmol) in 0.2 mL CH2Cl2. 4-DMAP (2 mg, 0.016 mmol)
and DCC (6.8 mg, 0.033 mmol) were added and the
solution was stirred at 4 ꢁC for 18 h. The reaction mix-
ture was diluted with CH2Cl2, washed with brine, dried,
filtered and evaporated. The residue was purified by
semi-preparative HPLC (25–100% CH3CN in H2O, 70
min gradient at a flow rate of 3 mL/min) affording tax-
ane 14 (2 mg, 12%, tR=52.21 min). The protected taxi-
nine ester 14 (2 mg, 0.002 mmol) in MeOH (1 mL) was
treated with a catalytic amount of p-toluenesulfonic
acid (PTSA) at 23ꢁC for 18 h. The solution was diluted
with EtOAc, washed with brine to neutrality, dried, fil-
tered and evaporated. The residue was purified by semi-
preparative HPLC (25–100% CH3CN in H2O, 50 min
gradient at a flow rate of 3 mL/min) affording taxane 15
(1 mg, 58%, tR=41.68 min). 1H NMR (500 MHz,
CDCl3) d 7.43 (o.d, 2H, H-2,6 of OBz), 7.36 (o.t, 2H,
H-3,5 of OBz), 7.28 (o.t, 1H, H-4 of OBz), 5.98 (o.m,
1H, H-7), 5.94 (o.m, 1H, H-9), 5.77 (o.m, 1H, NH-40),
5.20 (o.d, 1H, H-30), 4.84 (br.d, J=12.4 Hz, 1H, H-20a),
4.80 (br.d, J=12.4 Hz, 1H, H-20b), 4.53 (br.s, 1H, H-
20)), 4.18 (br.s, 1H, H-5), 2.68 (o.d, 1H, H-2b), 2.68
(o.m, 1H, H-14a), 2.61 (o.m, 1H, H-12), 2.22 (br.o.m,
1H, H-10a), 2.10 (o.m, 1H, H-14b), 2.07 (s, 3H, OAc),
2.05 (o.m, 1H, H-1), 2.05 (o.m, 1H, H-6a), 2.04 (dd,
J=15.4, 6.0 Hz, 1H, H-2a), 2.03 (s, 3H, OAc), 1.75
(o.m, 1H, H-10b), 1.48 (o.m, 1H, H-11), 1.41 (o.s, 9H,
t-Bu), 1.40 (o.s, 3H, H-17), 1.19 (s, 3H, H-19), 1.15 (o.d,
3H, H-18), 0.89 (s, 3H, H-16); 13C NMR (125 MHz,
CDCl3) d 216.9 (C-13), 170.1 (CO–OAc), 169.6 (CO–
OAc), 144.8 (C-3), 128.3 (C-3,5 of OBz), 127.6 (C-4),
127.3 (C-4 of OBz), 126.5 (C-2,6 of OBz), 77.4 (C-9),
74.2 (C-20), 67.1 (C-5), 67.1 (C-20), 65.9 (C-7), 56.0 (C-
30), 47.8 (C-11), 47.3 (C-12), 45.7 (C-8), 44.1 (C-1), 36.7
(C-15), 34.6 (C-16), 33.7 (C-6), 28.2 (t-Bu), 27.3 (C-2),
24.9 (C-17), 22.0 (C-18), 15.1 (C-19). FAB-HR-MS:
C38H53N1O11K [M+K]+ requires: 738.3256; found:
738.3256.
We thank the Natural Science and Engineering
Research Council of Canada, the Canadian Breast
Cancer Research Initiative grant and La Societe de
Recherche sur le Cancer for support via operating
grants to L.O.Z. Dr. Li Zhichao [Zhongling (Huizhou)
High Science and Technology Co. Ltd., PR China] and
Di-An Sun are gratefully acknowledged for a generous
supply gift of a sample of 2-deacetoxytaxinine J used in
the preparation of taxane 1.
References and Notes
1. Zamir, L. O.; Nedea, M. E.; Belair, S.; Sauriol, F.; Mamer,
O.; Jacqmain, E.; Jean, F. I.; Garneau, F.-X. Tetrahedron
Lett. 1992, 33, 5173.
2. Zamir, L. O.; Nedea, M. E.; Zhou, Z.-H.; Belair, S.; Caron,
G.; Sauriol, F.; Jacqmain, E.; Jean, F. I.; Garneau, F.-X.;
Mamer, O. Can. J. Chem. 1995, 73, 655.
3. Zamir, L. O.; Balachandran, S.; Zheng, Y. F.; Nedea,
M. E.; Caron, G.; Nikolakakis, A.; Vishwakarma, R. A.;
Sauriol, F.; Mamer, O. Tetrahedron 1997, 47, 15991.
4. Boulanger, Y. A.; Khiat, A.; Zhou, Z.-H.; Caron, G.;
Zamir, L. O. Tetrahedron 1996, 52, 8957.
5. Zamir, L. O.; Nedea, M. E.; Zhou, Z.-H.; Caron, G.;
Sauriol, F.; Mamer, O. Phytochemistry 1996, 41, 803.
6. Zamir, L. O.; Zhang, J.; Kutterer, K.; Sauriol, F.; Mamer,
O.; Khiat, A.; Boulanger, Y. Tetrahedron 1998, 54, 15845.
7. Zamir, L. O.; Zhang, J.; Wu, J.-H.; Sauriol, F.; Mamer, O.
J. Nat. Prod. 1999, 62, 1268.
8. Zamir, L. O.; Zhang, J.; Wu, J.-H.; Sauriol, F.; Mamer, O.
Tetrahedron 1999, 55, 14323.
9. Zhang, J.; Sauriol, F.; Mamer, O.; Zamir, L. O. Phyto-
chemistry 2000, 54, 221.
10. Zhang, J.; Sauriol, F.; Mamer, O.; You, X-L.; Alaoui-
Jamali, M. A.; Batist, G.; Zamir, L. O. J. Nat. Prod. 2001, 64,
450.
11. Gunawardana, G. P.; Premachandran, U.; Burres, N. S.;
Whittern, D.N.; Henry, R.; Spanton, S.; McAlpine, J. B. J.
Nat. Prod. 1992, 55, 1686.
12. Wu, J. H.; Zamir, L. O. Anti-Cancer Drug Des. 2000, 15,
73.
13. Kobayashi, J.; Hosoyama, H.; Shigemori, H.; Koiso, Y.;
Iwasaki, S. Experientia 1995, 51, 592.
Molecular modeling
14. Kobayashi, J.; Hosoyama, H.; Katsui, T.; Yoshida, N.;
Shigemori, H. Tetrahedron 1996, 52, 5391.
Based on the novel rearranged taxane skeletons 2, 3,
and 5, a series of novel side chains including those of
paclitaxel and docetaxel were virtually attached either at
C-5 or C-20 to generate a database of 28 virtual com-
pounds. Paclitaxel, docetaxel and taxuspine D were
added to the database as references. This database was
docked against the refined b-tubulin structure12 by using
the DOCK program.22 The docking procedure is the
same as our previous work26 except that the configur-
15. Gueritte-Voegelein, F.; Guenard, D.; Lavalle, F.; Le Goff,
M.-T.; Mangatal, L.; Potier, P. J. Med. Chem. 1991, 34, 992.
16. Georg, G. I.; Harriman, G. C. B., Vander Velde, D. G.;
Boge, T. C.; Cheruvallath, Z. S.; Datta, A.; Hepperle, M.;
Park, H.; Himes, R. H.; Jayasinghe, L. In Taxane Anticancer
Agents, ACS Symposium Series 583, Georg, G. I., Chen, T. T.,
Ojima, I., Vyas, D. M., (Eds.) American Chemical Society,
Washington DC, 1995, p 217.

A. Nikolakakis et al. / Bioorg. Med. Chem. 10 (2002) 2387–2395
2395
17. Kingston, D. G. I. In Taxane Anticancer Agents, ACS
Sumposium Series 583, Georg, G. I., Chen, T. T., Ojima, I.,
Vyas, D. M., (Eds.) American Chemical Society, Washington,
DC, 1995, p 203.
21. Didier, E.; Fouque, E.; Taillepied, I.; Commercon, A.
Tetrahedron Lett. 1994, 35, 2349.
22. Ewing, T. J. A.; Kuntz, I. D. J. Comput. Chem. 1997, 18,
1175.
18. Kelly, R. C.; Wicnienski, N. A.; Gebhard, I.; Qualls, S. J.;
Han, F.; Dobrowolski, P. J.; Nidy, E. G.; Johnson, R. A. J.
Am. Chem. Soc. 1996, 118, 919.
19. Morita, H.; Gonda, A.; Wei, L.; Yamamura, Y.; Waka-
bayashi, H.; Tayeka, K.; Itokawa, H. Phytochemistry 1998, 48,
857.
23. Swindell, C. S.; Krauss, N. E.; Horwitz, S. B.; Ringel, I. J.
Med. Chem. 1991, 34, 1176.
24. Hawkins, G. D.; Cramer, C. J.; Truhlar, D. G. J. Phys.
Chem. 1996, 100, 19824.
25. Nikolakakis, A.; Caron, G.; Cherestes, A. D.; Sauriol, F.;
Mamer, O.; Zamir, L. O. Bioorg. Med. Chem. 2000, 8, 1269.
26. Wu, J.H.; Batist, G.; Zamir, L.O. Anti-Cancer Drug Des.
2001, 16, 129.
20. Sun, D.-A.; Nikolakakis, A.; Sauriol, F.; Mamer, O.;
Zamir, L. O. Bioorg. Med. Chem. 2001, 9, 1985.