Deletion of the Oxetane Ring in Docetaxel Analogues: Synthesis and Biological Evaluation

Article abstract of DOI:10.1021/ol036043c

(Equation presented) Two new docetaxel analogues have been prepared starting from 10-deacetylbaccatin III. Both derivatives lack the oxetane D-ring but possess the 4α-acetoxy group, which is important for biological activity. The influence of a more or le

Full text of DOI:10.1021/ol036043c

ORGANIC
LETTERS
2003
Vol. 5, No. 26
5031-5034
Deletion of the Oxetane Ring in
Docetaxel Analogues: Synthesis and
Biological Evaluation
Vaishali Deka, Joe1lle Dubois,* Sylviane Thoret, Franc¸oise Gue´ritte, and
Daniel Gue´nard
Institut de Chimie des Substances Naturelles, CNRS,
91198 Gif sur YVette Cedex, France
joelle.dubois@icsn.cnrs-gif.fr
Received October 17, 2003
ABSTRACT
Two new docetaxel analogues have been prepared starting from 10-deacetylbaccatin III. Both derivatives lack the oxetane D-ring but possess
the 4r-acetoxy group, which is important for biological activity. The influence of a more or less constrained C-ring was evaluated by adding,
or not adding, a double bond in this ring. Both compounds were found to be equally less active than docetaxel in biological assays.
Paclitaxel, a complex diterpenoid 1a, and its semisynthetic
analogue docetaxel 1b are the most important anticancer
drugs developed in the past decade for the treatment of
ovarian, breast, and nonsmall cell lung cancers.1 The unique
mechanism of action of paclitaxel and its densely function-
alized, stereochemically rich framework has elicited a great
deal of interest in structure-activity relationship studies for
the development of more potent analogues. These studies
have shown that the C13 ester side chain and the ester groups
at C2 and C4 are essential for biological activity.2
D-ring in the interaction with microtubules and cytotoxicity.
Several synthesized D-ring analogues3 have been found to
show some interaction with microtubules, though generally
to a lesser extent than paclitaxel or docetaxel, while the
D-seco analogues were weakly cytotoxic4 or totally inactive.5
(2) For general reviews on taxoid chemistry and structure-activity
relationship studies, see: (a) Taxol: Science and Applications; Suffness,
M., Ed.; CRC Press: Boca Raton, FL, 1995. (b) Taxane Anticancer
Agents: Basic Science and Current Status; Georg, G. I., Chen, T. T., Ojima,
I., Vyas, D. M., Eds.; ACS Symposium Series 583; American Chemical
Society: Washington, DC, 1995. (c) The Chemistry and Pharmacology of
Taxol and its DeriVatiVes; Farina, V., Ed.; Elsevier: Amsterdam, 1995.
(d) Kingston, D. G. I. J. Nat. Prod 2000, 63, 726-734. (e) Kingston, D. G.
I.; Yuan, H.; Jagtap, P. J.; Samala, L. The Chemistry of Taxol and Related
Taxoids. In Progress in the Chemistry of Organic Natural Products; Herz,
W., Kirby, G. W., Moore, R. E., Steglich, W., Tamm, Ch., Eds.; Springer-
Verlag: New York, 2002; Vol. 84.
(3) (a) Marder-Karsenti, R.; Dubois, J.; Bricard, L.; Gue´nard, D.; Gue´ritte-
Voegelein, F. J. Org. Chem. 1997, 62, 6631-6637. (b) Gunatilaka, A. A.
L.; Ramdayal, F. D.; Sarragiotto, M.; Kingston, D. G. I.; Sackett, D. L.;
Hamel, E.; J. Org. Chem. 1999, 64, 2694-2703. (c) Merckle´, L.; Dubois,
J.; Place, E.; Thoret, S.; Gue´ritte, F.; Gue´nard, D.; Poupat, C.;Ahond, A.;
Potier, P. J. Org. Chem 2001, 66, 5058-5065. (d) Dubois, J.; Thoret, S.;
Gue´ritte, F.; Gue´nard, D. Tetrahedron Lett. 2000, 41, 3331-3334.
(4) (a) Beusker, P. H.; Veldhuis, H.; Van den Bossche, B. A. C.;
Scheeren, H. W. Eur. J. Org. Chem. 2001, 1761-1768. (b) Beusker, P. H.;
Veldhuis, H.; Brinkhorst, J.; Hetterscheid, D. G. H.; Feichter, N.; Bugaut,
A.; Scheeren, H. W. Eur. J. Org. Chem. 2003, 689-705.
In the past 10 years, significant efforts have been made
to establish the actual contribution of the unique oxetane
(1) (a) Paclitaxel in Cancer Treatment; McGuire, W. P., Rowinsky, E.
K., Eds.; Marcel Dekker: New York, 1995. Vol. 8, pp 1-349.
10.1021/ol036043c CCC: $25.00 © 2003 American Chemical Society
Published on Web 11/22/2003

Interestingly, calculations by the Snyder group using a
minireceptor approach and a refined model of the paclitaxel-
â-tubulin binding pocket predicted that derivatives with
oxygenated substituents at C4 and an appropriately placed
double bond maintain paclitaxel-like activity.6
Scheme 2 a
In the present communication, we report our continuing
endeavor toward the synthesis of D-ring-modified analogues
of docetaxel. Our target was the synthesis of the simple
D-seco docetaxel analogue 2a lacking the oxetane ring. This
kind of compound has been predicted by the Snyder group’s
calculations to be well accommodated in the â-tubulin
binding pocket.6 To evaluate the influence of a more
constrained C-ring on the biological activity, the synthesis
of 5,6-dehydro derivative 2b has also been designed.
a Reagents and conditions: (a) Pd/C, MeOH, H2, rt (98%); (b)
Pb(OAc)4, CH3CN, 0 °C (98%); (c) NaBH4, MeOH, 0 °C (80%).
We then decided to carry out the oxidative cleavage before
removal of the bromine, as we hypothesized that the presence
of the C5-R bromine would sufficiently hinder the R face of
the molecule to facilitate hydride delivery from the â-face
during the reduction of the C4-ketone. Oxidative cleavage
of compound 3 afforded the desired ketone 7 in high yield,
and as expected, reduction of the C4-ketone with sodium
borohydride was less stereoselective, affording 30% of the
R-hydroxide 8 along with 50% of the â-isomer 9 (Scheme
3). The stereochemistry of the C4 hydroxyl group was
We chose to proceed via our key intermediate 3, prepared
from 10-deacetylbaccatin III and successfully used for the
synthesis of our D-modified docetaxel analogues.3a,c,d This
compound can be debrominated or dehydrobrominated and
the 1,2-diol transformed by oxidative cleavage and stereo-
selective reduction to the R-C4 alcohol. Subsequent acety-
lation and introduction of the C2 and C13 functionalities
could afford the desired compounds 2a and 2b (Scheme 1).
Scheme 3 a
a Reagents and conditions: (a) Pb(OAc)4, CH3CN, 0 °C (99%);
(b) NaBH4, MeOH, 0 °C, 8 (30%) and 9 (50%) or BH3‚Me2S, (R)-
2-methyl CBS-oxazaborolidine, CH2Cl2, -20 °C, 8 (63%) and 9
(22%).
Scheme 1
determined by NOE experiments on both compounds 8 and
9. Compound 8 showed interactions between H4/H2 and
H4/CH319, characteristic of the â-orientation of H4, whereas
compound 9 exhibited the same correlations (H4/H3 and
H4/H14R) as compound 6.
To increase the yield of the R-isomer 8, several other
reducing agents (L- and K-Selectride, lithium triethylboro-
hydride, lithium tri-tert-butoxyaluminohydride, lithium boro-
hydride) were tried but did not give the desired reduction
product, in many cases because of concomitant debromina-
tion. Modification of the reaction conditions (solvent, tem-
perature) did not really improve the R-isomer yield. Finally,
the best result was obtained with BH3.Me2S in the presence
of (R)-2-methyl CBS-oxazaborolidine catalyst,8 which af-
forded the desired R-isomer in 63% yield along with 22%
of the â-isomer.
After a number of experiments with several reducing
agents, debromination of 3 was achieved with Pd/C in MeOH
under an atmosphere of hydrogen. Oxidative cleavage of the
vicinal diol in 4 with Pb(OAc)4 gave the ketone 5, reduction
of which with sodium borohydride proceeded with complete
facial selectivity yielding the undesired â-hydroxide 6 as
previously reported4a,7 (Scheme 2). Mitsunobu reaction on
the â-hydroxide to obtain acetylation with inversion of
configuration proved to be unsuccessful.
(5) (a) Barboni, L.; Dutta, A.; Dutta, D.; Georg, G. I.; Vander Velde, D.
G.; Himes, R. H.; Wang, M.; Snyder, J. P. J. Org. Chem. 2001, 66, 3321-
3329. (b) Samaranayake, G.; Magri, N. F.; Jitrangsri, C.; Kingston, D. G.
I. J. Org. Chem. 1991, 56, 5114-5119.
(6) Wang, M.; Cornett, B.; Nettles, J.; Liotta, D. C.; Snyder, J. P. J.
Org. Chem. 2000, 65, 1059-1068.
(7) â-Configuration of the C4 hydroxyl group was proved by NOESY
experiments, which showed H4/H3 and H4/H14R correlations.
Acetylation of 8 with acetic anhydride/DMAP proved to
be facile, affording the acetate 10 in high yield. It should be
noticed that, using the same conditions, compound 9 has also
(8) Corey, E. J.; Bakshi, R. K.; Shibata, S. J. Am. Chem. Soc. 1987,
109, 5551-5553.
5032
Org. Lett., Vol. 5, No. 26, 2003

been easily acetylated, contrary to what was reported by
Beusker et al. on their compound deprived of any functional-
ity at C5.4a Therefore, the presence of the C5-R-Br alters
the conformation of the molecule to render the 4â-hydroxy
group sterically less hindered. Debromination of 10 by
hydrogen in the presence of Pd/C afforded 11a in good yield.
Dehydrobromination proved to be more difficult. A number
of bases (DBU, DBN, lithium or potassium carbonate, etc.)
and experimental conditions (solvent, temperaure) were tested
for this reaction. Compound 11b was obtained with the best
yield by heating 10 in DMF in the presence of potassium
acetate. However this reagent afforded also the 5-acetyl
derivative 12 as a side product. This intermolecular substitu-
tion of the 5-R-bromine has never been reported before, but
this may be due, in our case, to the absence of a methyl or
methylene group at position 4-â. Then, the C1-C2 carbonate
of 11a and 11b was readily opened by phenyllithium as
previously described3a (Scheme 4).
Scheme 5 a
a Reagents and conditions: NaBH4, THF/MeOH (6:1), rt, 14a
(30%) and 15a (47%); 14b (45%) and 15b (25%).
dehydro compounds shows that the double bond induces a
conformational change in the C-ring and probably a different
position of the 4-acetyl group that restrains the accessibility
of the hydride to the R face.
Esterification was then realized with the 2-(4-OMe)phenyl
1,3-oxazolidine of N-Boc-phenylisoserine 16, DCC, and
DMAP in toluene at room temperature, affording 17 in good
yield. Removal of all the protective groups in one step with
p-toluenesulfonic acid in methanol afforded the desired
compounds 2a and 2b (86 and 50%, respectively) (Scheme
6).
Scheme 4 a
Scheme 6 a
a Reagents and conditions: (a) Ac2O, pyridine, DMAP, CH2Cl2,
40 °C, (86%); (b) H2, Pd/C, MeOH, (88%); (c) KOAc, DMF, 80
°C, 11b (48%) and 12 (12%); (d) PheLi, THF, -78 °C, 13a (67%)
and 13b (61%).
Contrary to our earlier results,3a,c,d the reduction of the C13
ketone with NaBH4 was no longer stereoselective. The
saturated derivative 13a led to the C13-R and C13-â isomers
in 30 and 47% yields, respectively, whereas the unsaturated
analogue 13b gave rise to the same isomers in 45 and 25%
yields, respectively9 (Scheme 5). In that case, modification
of the reducing agent or addition of a chiral reagent could
not improve the stereoselectivity. In the literature, borohy-
dride reduction of the C13-ketone leading to the C13-â-
hydroxy group has only been reported for 4-deacetyl-7-
(triethylsilyl)-13-oxo-baccatin III, where the 4-hydroxy group
participates in a transannular delivery of the hydride.10 The
difference in the stereoselectivity for the saturated and
a Reagents and conditions: (a) 16, DCC, DMAP, toluene, rt 17a
(93%), 17b (88%); (b) PTSA, MeOH, rt, 2a (86%), 2b (50%).
These compounds were found to be less active than
docetaxel, 70 times less in the microtubule disassembly
assay11 and up to a 1000 times less cytotoxic against KB
cell lines.12 No difference was observed between the two
compounds, indicating that an additional constraint on the
C-ring is not sufficient to restore the activity. It is to be noted
that these derivatives are the first D-seco analogues retaining
an interaction, though comparatively weak, with micro-
tubules.
(9) The stereochemistry of each compounds was assigned by NOE
experiments.
(10) Hoemann, M. Z.; Vander Velde, D. G.; Aube´, J.; Georg, G. I. J.
Org. Chem. 1995, 60, 2918-2921.
(11) Lataste, H.; Se´nilh, V.; Wright, M.; Gue´nard, D.; Potier, P. Proc.
Natl. Acad. Sci. U.S.A. 1984, 81, 4090-4094.
(12) Da Silva, A. D.; Machado, A. S.; Tempeˆte, C.; Robert-Gero, M.
Eur. J. Med. Chem. 1994, 29, 149-152.
Org. Lett., Vol. 5, No. 26, 2003
5033

As previously described by Boge et al.,13 the absence of
the oxetane ring slightly modifies the overall conformation
of the taxane skeleton. Our compounds 2a and 2b display
changes similar to those of D-secopaclitaxel, especially for
the A-ring as shown by H14 chemical shifts and coupling
constants (Table 1).
biological activity. This small ring comprises three important
features: an oxygen atom, a constrained four-membered ring,
and a methylene group in a 1,3-diaxial interaction with the
19-methyl group. Deletion of the oxetane ring resulted in
suppression of all these features. The oxygen atom has been
shown not to be essential for tubulin interaction since 5(20)-
deoxydocetaxel is equipotent to paclitaxel in the microtubule
disassembly assay.3d,15 The constraint imposed by the small
ring actually influences the overall conformation of the
diterpene moiety as seen by NMR data in D-secopaclitaxel13
and in compounds 2. An additional double bond on the
C-ring is not sufficient to restore good affinity for micro-
tubules, suggesting that a constrained C-ring is not the
essential factor in the oxetane ring. The main difference
between docetaxel and compounds 2 lies in the C-ring
conformation and, therefore, in the orientation of the 4-ac-
etoxy group. This result suggests that the biological activity
is very sensitive to the position of this acetoxy group relative
to the C13-side chain. In compounds 2, the absence of a
methylene or a methyl group at position 4â may be critical
for the conformation of the C-ring. It is likely that the 1,3-
diaxial interaction between the 19-methyl group and a 4â-
bulky group would induce a conformation of the C-ring that
better fits the tubulin binding site. To check this hypothesis,
new D-seco compounds with a 20-methyl group and the 4R-
acetoxy moiety are currently being synthesized in our
laboratory.
Table 1. Comparison of Selected NMR Data and Coupling
Constants of Compounds 2a and 2b in CDCl3 with Docetaxel
and D-Secopaclitaxel
H14R H14â J 13,14R
J 13,14
â
J 2,3
docetaxel 1b
2.28
3.08
2.98
2.88
2.18
2.45
2.46
2.45
8 Hz
8 Hz
7 Hz
D-secopaclitaxel13
4.4 Hz 10.3 Hz 5.3 Hz
6 Hz
6 Hz
2a
2b
10 Hz
10 Hz
5.5 Hz
6.5 Hz
The presence of the 5,6-double bond in the C-ring in 2b
does not change the A-ring conformation but induces only
a modest conformational change, as seen with the H2,H3
coupling constant (Table 1). NOESY experiments have not
shown many differences in the conformational behavior in
solution except a much weaker NOE between H2′ and 4-OAc
than in docetaxel, suggesting a slightly different position for
the 4-acetoxy group that was confirmed by molecular
modeling studies.
Contrary to 4-deacetyldocetaxel,14 molecular modeling
studies and NOE experiments have not shown any significant
modification in the position of the side chain when the
oxetane D-ring is absent. The overall conformation of these
derivatives does not seem to be very different from that of
docetaxel.
Acknowledgment. The authors thank Dr. Alain Com-
merc¸on (Aventis-Pharma) for a gift of the docetaxel side
chain. Jean-Franc¸ois Gallard is acknowledged for performing
NOESY and ROESY experiments and Christiane Gaspard
for cytotoxicity evaluations.
In conclusion, contrary to what was predicted by Snyder’s
model,6 deletion of the oxetane ring is detrimental to
Supporting Information Available: Experimental pro-
cedures and full characterization data for compounds 2a, 2b,
4-15, and 17. This material is available free of charge via
the Internet at http://pubs.acs.org.
(13) Boge, T. C.; Hepperle, M.; Vander Velde D. G.; Gunn, C. W.;
Grunewald, G. L.; Georg, G. I. Bioorg. Med. Chem. Lett. 1999, 9, 3041-
3046.
(14) Gue´ritte-Voegelein, F.; Gue´nard, D.; Dubois, J.; Wahl; A.; Marder,
R.; Mu¨ller, R.; Lund, M.; Bricard, L.; Potier, P. 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, 1995; pp 189-202.
OL036043C
(15) However, 5(20)-deoxydocetaxel is about 100-fold less cytotoxic than
docetaxel against KB cell line (IC50(5(20)-deoxydocetaxel) ) 0.06 µM).
5034
Org. Lett., Vol. 5, No. 26, 2003