Synthesis and biological evaluation of novel macrocyclic paclitaxel analogues

Article abstract of DOI:10.1021/ol016124d

matrix presented This work describes the synthesis of two novel macrocyclic taxoid constructs by ring-closing olefin metathesis (RCM) and their biological evaluation. Computational studies examine conformational profiles of 1 and 2 for their fit to the β-

Full text of DOI:10.1021/ol016124d

ORGANIC
LETTERS
2001
Vol. 3, No. 16
2461-2464
Synthesis and Biological Evaluation of
Novel Macrocyclic Paclitaxel Analogues
Belhu B. Metaferia,† Jeannine Hoch,† Thomas E. Glass,† Susan L. Bane,‡
,§
Sabarni K. Chatterjee,‡ James P. Snyder,* Ami Lakdawala,§ Ben Cornett,§ and
,†
David G. I. Kingston*
Departments of Chemistry, Virginia Polytechnic Institute and State UniVersity,
Blacksburg, Virginia 24061, State UniVersity of New York at Binghamton,
Binghamton, New York 13902, and Emory UniVersity, Atlanta, Georgia 30322
dkingston@Vt.edu
Received May 14, 2001
ABSTRACT
This work describes the synthesis of two novel macrocyclic taxoid constructs by ring-closing olefin metathesis (RCM) and their biological
evaluation. Computational studies examine conformational profiles of 1 and 2 for their fit to the â-tubulin binding site determined by electron
crystallography. The results support the hypothesis that paclitaxel binds to microtubules in a “T” conformation.
The important diterpenoid natural product paclitaxel (Taxol)
is currently the world’s best-selling anticancer drug, with
sales of over $1.5 billion in 2000.1 It operates, in part at
least, by binding to microtubules and stabilizing them to
disassembly,2 and several recent studies have attempted to
delineate the binding conformation of paclitaxel on as-
sembled microtubules. Three models have been proposed on
the basis of the published electron crystallographic coordi-
nates of polymerized Râ-tubulin.3 A study based on REDOR
NMR and fluorescence spectroscopy by two of our groups
led to the proposal of a hydrophobically collapsed conforma-
tion,4 as did the modeling by Rao et al. based on photoaffinity
labeling.5 Giannakakou et al.6 conceived epothilone binding
models from the single-crystal X-ray structure of taxotere7
docked into â-tubulin.3 A more recent study of the binding
of paclitaxel to tubulin using the crystallographic density in
conjunction with an analysis of paclitaxel conformations
indicates that the T-shaped conformer, deprived of hydro-
phobic collapse, is the most likely bioactive form.8
A fruitful approach to the delineation of the conformation
of paclitaxel in its bound conformation on tubulin has been
through the synthesis of analogues with built-in conforma-
tional restrictions. A number of experiments in this direction
have been published, including the synthesis of analogues
with conformationally restricted side chains9 and studies of
analogues with bridges linking the C3′ and C2 phenyls with
† Virginia Polytechnic Institute and State University.
‡ SUNY, Binghamton.
(5) Rao, S.; He, L.; Chakravarty, S.; Ojima, I.; Orr, G. A.; Horwitz, S.
B. J. Biol. Chem. 1999, 274, 37990-37994.
§ Emory University.
(1) Thayer, A. M. Chem. Eng. News 2000, 78 (45, November 6), 20-
21.
(6) Giannakakou, P.; Gussio, R.; Nogales, E.; Downing, K. H.; Zaha-
revitz, D.; Bollbuck, B.; Poy, G.; Sackett, D.; Nicolaou, K. C.; Fojo, T.
Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 2904-2909.
(2) Schiff, P. B.; Fant, J.; Horwitz, S. B. Nature 1979, 277, 665-667.
(3) Nogales, E.; Wolf, S. G.; Downing, K. H. Nature 1998, 391, 199-
202.
(4) Li, Y.; Poliks, B.; Cegelski, L.; Poliks, M.; Gryczynski, Z.; Piszczek,
G.; Jagtap, P. G.; Studelska, D. R.; Kingston, D. G. I.; Schaefer, J.; Bane,
S. Biochemistry 2000, 39, 281-291.
(7) Gueritte-Voegelein, F.; Guenard, D.; Mangatal, L.; Potier, P.;
Guilhem, J.; Cesariao, M.; Pascard, C. Acta Crystallogr. 1990, C46, 781-
784.
(8) Snyder, J. P.; Nettles, J. H.; Cornett, B.; Downing, K. H.; Nogales,
E. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 5312-5316.
10.1021/ol016124d CCC: $20.00 © 2001 American Chemical Society
Published on Web 07/11/2001

both shorter10 and longer11 linkers. Only the latter of the two
bridged approaches yielded analogues having tubulin as-
sembly activity, and the activity of the best analogues was
less than that of paclitaxel.
Scheme 1a
In an attempt to model the T-paclitaxel conformation, we
have designed and synthesized the two analogues 1 and 2
linked between the C3′-phenyl group and the C4 position.
The T-paclitaxel conformation sustains nearly equal 9-10
Å distances between the C13 side chain terminal phenyl rings
and the C2-benzoyl phenyl group. Short, conformationally
directing linkers between these centers are out of the
question. On the other hand, the C4-acetate methyl lies very
close to one edge of the C3′ phenyl moiety in this conformer,
suggesting a possibility for conformational control. A variety
of positional isomers and C3′-C4 linkers were modeled as
T-Taxol mimics. In the end, proof of principle and synthetic
ease were integrated in structures 1 and 2.
a Reagents and conditions: (a) p-MeOC6H4NH2, MgSO4, CH2Cl2
(100%); (b) CH3COOCH2COCl, Et3N, -78 °C to room tempera-
ture, 12 h (85%); (c) lipase PS Amano, phosphate buffer, pH )
7.2, CH3CN, 24 h (98%); (d) 1 M, KOH, THF, 0 °C (quantitative);
(e) TIPSCl, imidazole, DMF (94%); (f) CAN, CH3CN, -5 °C
(92%); (g) PhCOCl, Et3N, DMAP, CH2Cl2 (90%).
The starting material in Scheme 2, the C4 hydroxy baccatin
derivative 7, was prepared from 10-deacetylbaccatin III (10-
DAB) using a reported protocol.16 Initial attempts to acylate
the hindered C-4 hydroxy group using the DCC/DMAP
method gave low yields, but the use of acid chloride and
LiHMDS in THF at 0 °C gave the desired C-4 modified
baccatin in 78% yield. A minor product acylated at the C-4
position but with a loss of the 1-dimethylsilyl group was
recovered and used for the subsequent step. Global depro-
tection of the silyl groups using HF/pyridine followed by a
selective C-10 acetylation with 0.1 mol % of CeCl3 and acetic
anhydride in THF gave a 94% yield of the desired 10-acetyl
derivative,17 and selective reprotection of the C7 hydroxy
as the triethylsilyl ether afforded 8 in good yield.
While the latter were not our optimal designs, they
nonetheless permitted a straightforward approach to targets
by means of the elegant ring-closing metathesis (RCM)
methodology in the crucial macrocyclization step. This
synthetic strategy was used effectively by Ojima and his
collaborators in preparing a large number of paclitaxel
analogues with C3′-C2 linkers,11,12 and it has the advantage
of tolerating a wide variety of functional groups.13 The
syntheses, bioactivities, and modeling of these compounds
are described below.
The synthesis started with the preparation of the â-lactam
6 as shown in Scheme 1. The starting material 3 was reacted
with p-anisidine to form p-methoxyphenyl (PMP) imine and
then converted to a racemic â-lactam by a Staudinger [2 +
2] cyclocondensation between a ketene generated from
acetoxyacetyl chloride and the PMP imine.14 The racemic
â-lactam was then subjected to an enzymatic resolution using
lipase; the yield was 98%, based on the desired enantiomer
4.15 Functional group manipulations then gave the target
lactam 6 through intermediate 5.
Scheme 2a
a Reagents and conditions: (a) LiHMDS, THF, 0 °C, ClCOCH2-
CH2CHdCH2, (78%); (b) HF/py, THF, (91%); (c) 0.1 mol % of
CeCl3, acetic anhydride, THF, (94%); (d) TESCl, imidazole, DMF,
92%
(9) Barboni, L.; Lambertucci, C.; Appendino, G.; Vander Velde, D. G.;
Himes, R. H.; Bombardelli, E.; Wang, M.; Snyder, J. P. J. Med. Chem.
2001, 44, 1576-1587.
(10) Boge, T. C.; Wu, Z.-J.; Himes, R. H.; Vander Velde, D. G.; Georg,
G. I. Bioorg. Med. Chem. Lett. 1999, 9, 3047-3052.
(11) Ojima, I.; Chakravarty, S.; Tadashi, I.; Lin, S.; He, L.; Horwitz, S.
B.; Kuduk, S. B.; Danishefsky, S. J. Proc. Natl. Acad. Sci. U.S.A. 1999,
96, 4256-4261.
(12) Ojima, I.; Lin, S.; Inoue, T.; Miller, M. L.; Borella, C. P.; Geng,
X.; Walsh, J. J. J. Am. Chem. Soc. 2000, 122, 5343-5353.
(13) For a recent review on RCM, see: Furstner, A. Angew. Chem., Int.
Ed. 2000, 39, 3012-3043.
(14) Staudinger, H. Justus Liebigs Ann. Chem. 1907, 356, 51.
(15) Brieva, R.; Crich, J. Z.; Sih, C. J. J. Org. Chem. 1993, 58, 1068-
1075.
The coupling reaction of C4-modified baccatin derivative
8 and â-lactam 6 using the Holton-Ojima-Georg proto-
col18-20 gave the precursor taxoid-ω,ω′-diene 9 (92%), setting
(16) Chen, S.-H.; Kadow, J. F.; Farina, V. J. Org. Chem. 1994, 59, 6156-
6158.
(17) Holton, R. A.; Zhang, Z.; Clarke, P. A.; Nadizadeh, H.; Procter, D.
J. Tetrahedron Lett. 1998, 39, 2883-2886.
2462
Org. Lett., Vol. 3, No. 16, 2001

Scheme 3a
Table 1. Bioactivity Results for Compounds 1 and 2
cytotoxicity to
tubulin assembly
A2780 mammalian
cells (IC50, mg/mL)a
compound
activity (I50, mM)
paclitaxel
1
2
0.27 ( 0.05
2.54 ( 0.53
8.1 ( 1.0
0.136 ( 0.08
1.14 ( 0.03
1.07 ( 0.05
a The strain of A2780 used in these experiments consistently gave IC50
values in the range 0.06-0.15 µg/mL for paclitaxel. These values are higher
than those of other workers and are presumably due to variations in the
strain used. The same strain was used for paclitaxel and for compounds 1
and 2.
than those observed by him, but the tubulin assembly
activities are not as high as those of the best compounds he
obtained.12
a Reagents and conditions: (a) ((Cy3)P)2Cl2RudCHPh, CH2Cl2,
20 h, (72%); (b) HF/py THF 12 h, (78%).
In an effort to understand the reduced activities of 1 and
2 relative to that of paclitaxel, we performed Monte Carlo
conformational searches on the structures with several force
fields tailored to small molecules (MM2*, MM3*, and
MMFF (GBSA/H2O))21 and a 5 kcal/mol energy window.
For 1 and 2, the combined set of 500-1200 unique and fully
optimized conformers was found to contain torsional isomers
with excellent superposition of both diterpenoid core and
C2, C4, and C13 side chains on the same moieties of
T-paclitaxel. If the latter is the bioactive conformation at
â-tubulin, and if 1 and 2 can adopt this conformation, why
are the hybrid structures less efficacious than paclitaxel? One
possibility is that the T-forms are simply too high in energy.
To test this idea, we sought to define the conformational
profile of compound 1 in solution by means of a NAMFIS
conformer deconvolution treatment, a procedure that avoids
the capriciousness of raw molecular mechanics energies.22,23
Careful integration of 2-D ROESY NMR cross-peaks for 1
in CDCl3 led to 27 well-defined intramolecular distances
accompanied by J(H2′-H3′) ) 1.54 Hz. NAMFIS analysis,
combining the latter NMR quantities and the full 514
conformer data set, yielded eight conformations with pre-
dicted populations ranging from 2 to 35% (∆∆G ) 0-1.7
kcal/mol). These comprise two nonpolar forms (40% total)
and extended conformers (60% total). Among them is an
unambiguous 3-D mimic of the T-conformation (5%). The
latter fully optimized structure is clearly within reasonable
energetic limits for binding to tubulin. Unlike the projected
mole fractions of paclitaxel torsional minima in CDCl3,23
no polar forms appear in the current NAMFIS solution.
Further pursuing the source of reduced activity in 1 and
2, we docked the T-conformers derived by Monte Carlo
searching into the paclitaxel binding pocket of the refined
electron crystallographic structure3 of â-tubulin with the
the stage for the RCM reaction shown in Scheme 3. The
RCM of 9 was catalyzed by Grubbs’ catalyst in CH2Cl2 at
room temperature and gave the desired 21-membered macro-
cyclic analogue 10 in 72% yield; the Z isomer was formed
exclusively, as evidenced by NMR coupling constants.
Deprotection of 10 gave the final bridged paclitaxel analogue
1 in 78% yield.
A repeat of the above reaction sequence using â-lactam
11 in place of 6 gave the second bridged analogue, 2.
Bioassays on the cyclic products 1 and 2 were carried out
in a microtubule assembly assay and a cytotoxicity assay;
the results are shown in Table 1. In the cytotoxicity bioassays,
compounds 1 and 2 have comparable bioactivities, and these
are only about eight times less than that of paclitaxel. In the
tubulin assembly assay, both compounds are also less active
than paclitaxel, with compound 1 being about 10-fold less
active and compound 2 about 30-fold less active. These
results can be compared with those obtained by Ojima et
al.; the cytotoxicity values relative to paclitaxel are greater
(18) Holton, R. A.; Biediger, R. J.; Boatman, P. D. In Taxol®: Science
and Applications; Suffness, M., Ed.; CRC Press: Boca Raton, FL, 1995;
pp 97-121.
(19) Ojima, I.; Habus, I.; Zhao, M.; Zucco, M.; Park, Y. H.; Sun, C. M.;
Brigaud, T. Tetrahedron 1992, 48, 6985-7012.
(20) Georg, G. I.; Boge, T. C.; Cheruvallath, Z. S.; Clowers, J. S.;
Harriman, G. C. B.; Hepperle, M.; Park, H. In Taxol®: Science and
Applications; Suffness, M., Ed.; CRC Press: Boca Raton, FL, 1995; pp
317-375.
(21) Mohamadi, F.; Richards, N. G. J.; Guida, W. C.; Liscamp, R.;
Lipton, M.; Caufield, C.; Chang, G.; Hendrickson, T.; Still, W. C. J.
Comput. Chem. 1990, 11, 440-449; cf. http://www.schrodinger.com/
macromodel2.html.
(22) Nevins, N.; Cicero, D.; Snyder, J. P. J. Org. Chem. 1999, 64, 3979-
3986.
(23) Snyder, J. P.; Nevins, N.; Cicero, D. O.; Jansen, J. J. Am. Chem.
Soc. 2000, 122, 724-725.
Org. Lett., Vol. 3, No. 16, 2001
2463

DOCK protocol.24 This approach has been remarkably
successful for understanding a range of taxoid-tubulin
interactions: analysis of the role of the oxetane ring for
tubulin binding,25 prediction of the activity of a cyclopropyl
analogue of oxetane,25,26 and rationalization of the SAR of
C2-benzoyl derivatives.27,28 The best rigid fit for 1 is pictured
in Figure 1. The bridged paclitaxel does not fit nearly as
result was obtained for 2. Examination of the floor of the
cleft reveals both a kink in the C4-C3′ tether and close
contact between Ala233 and Phe272 of the protein and the
propene fragment of the linker. In an attempt to introduce a
degree of induced fit, we have performed molecular dynam-
ics on the complex of Figure 1 under different conditions at
several temperatures for 20 ps. The Phe272-tether interac-
tion is not abated. Thus, while the present results are
compatible with the T-Taxol binding motif, they suggest that
a flexible and extended bridge can be deleterious to full
expression of taxoid activity.
In conclusion, we have prepared the C4OAc-C3′Ph
bridged analogues 1 and 2. Both compounds are active but
somewhat less so than paclitaxel in tubulin assembly and
cytotoxicity assays. While we had hoped that the intra-
molecular tethers would provide sufficient molecular rigidity
to mimic T-Taxol exclusively, NMR conformational analysis
suggests that compound 1 can adopt near-nonpolar, T- and
side-chain extended conformational types without experienc-
ing undue torsional strain. A recent study of paclitaxel
analogues constrained around C3′ within the C13 side chain
led to a similar result.9 These outcomes offer an important
caveat to NMR interpretations that lead to single conforma-
tional solutions of the averaged ROE and J data.11,22 In the
present work, complementary docking studies strongly
suggest that although T-shapes of 1 and 2 fill the â-tubulin
binding pocket, they unfortunately encounter steric resistance
at the bottom of the cleft. While the bioassay and docking
studies are compatible with a bioactive T-form, they likewise
illustrate that the ideal tether length and constitution has not
yet been achieved. Future work will explore this issue in
some detail.
Figure 1. T-Taxol (magenta) bound to â-tubulin as described in
ref 7. The best T-form of compound 1 (yellow) is seated higher in
the same pocket as a result of close contact between the propene
moiety of the tether and Phe272 of the protein (white) at the bottom
of the illustration.
snugly within the hydrophobic binding pocket as does
T-Taxol. However, since the present docking involves a rigid
body for both the ligand and the protein, it lacks any degree
of induced fit, thereby maximizing any differences. A similar
Acknowledgment. This work was supported by a re-
search grant from the National Institutes of Health (CA-
69571). Mass spectra were obtained by Mr. William Bebout
at Virginia Tech and by the Nebraska Center for Mass
Spectrometry.
(24) Ewing, T. J. A.; Kuntz, E. D. J. Comput. Chem. 1997, 18, 1175-
1189; cf. http://www.cmpharm.ucsf.edu/kuntz/kuntz.html.
(25) Wang, M.; Cornett, B.; Nettles, J.; Liotta, D. C.; Snyder, J. P. J.
Org. Chem. 2000, 65, 1059-1068.
(26) Dubois, J.; Thoret, S.; Gue´nard, D.; Gue´ritte, F. Tetrahedron Lett.
2000, 65, 1059-1068.
(27) Kingston, D. G. I.; Chaudhary, A. G.; Chordia, M. D.; Gharpure,
M.; Gunatilaka, A. A. L.; Higgs, P. I.; Rimoldi, J. M.; Samala, L.; Jagtap,
P. G.; Giannakakou, P.; Jiang, Y. Q.; Lin, C. M.; Hamel, E.; Long, B. H.;
Fairchild, C. R.; Johnston, K. A. J. Med. Chem. 1998, 41, 3715-3726.
(28) Wang, M.; Snyder, J. P. Unpublished.
Supporting Information Available: Experimental pro-
cedures and full characterization data for compounds 1, 2,
and 6-11. This material is available free of charge via the
Internet at http://pubs.acs.org.
OL016124D
2464
Org. Lett., Vol. 3, No. 16, 2001