Design, synthesis, and anticancer activity of phosphonic acid diphosphate derivative of adenine-containing butenolide and its water-soluble derivatives of paclitaxel with high antitumor activity

Article abstract of DOI:10.1016/S0968-0896(03)00524-8

Synthesis of adenine derivative of triphosphono-γ -(Z)-ethylidene-2,3-dimethoxybutenolide 4 was accomplished by treatment of phosphonate 3 with 5-phosphoribosyl 1-pyrophosphate in the presence of 5-phosphoribosyl 1-pyrophosphate synthetase. It was found that triphosphonate 4 functions as an irreversible stoichiometric inactivator of the Escherichia coli ribonucleoside diphosphate reductase (RDPR). Triphosphonate 4 exhibited potent inhibitory activity against murine leukemias (L1210 and P388), breast carcinoma (MCF7), and human T-lymphoblasts (Molt4/C8 and CEM/0) cell lines. Paclitaxel ester derivatives of adenine-containing triphosphono-γ -(Z)-ethylidene-2,3-dimethoxybutenolide 8-10 were also synthesized. Like triphosphonate 4, compound 8 exhibited inhibitory property toward RDPR. It also induced microtubule assembly similar to paclitaxel (5). The structure of the chlorodiester linker in 8 was found to account for this dual property. After treatment of MCF7 cells with compounds 4, 5, and 8, fluorescence microscope examination demonstrated the presence of nucleus shrinkage or segmentation. Bifunctional prodrug 8 exhibited higher lipophilicity than 4 and higher water-solubility than 5. Pro-dual-drug 8 exhibited more pronounced anticancer activity relative to that of the triphosphonate 4 and paclitaxel (5). In contrast, compound 9, resulting from the linkage of triphosphonate 4 and paclitaxel (5) through a diester unit, was only found to function as a highly water-soluble prodrug for paclitaxel (5). It induced microtubule assembly in vitro, but did not show inhibitory property toward RDPR. On the other hand, compound 10, an aggregate of triphosphonate 4 and paclitaxel (5), neither functioned as an inhibitor of RDPR nor exhibited microtubule assembly stimulating activity in vitro.

Full text of DOI:10.1016/S0968-0896(03)00524-8

Bioorganic & Medicinal Chemistry 11 (2003) 4303–4313
Design, Synthesis, and Anticancer Activity of Phosphonic Acid
Diphosphate Derivative of Adenine-Containing Butenolide and Its
Water-soluble Derivatives of Paclitaxel with High Antitumor
Activity
Ali A. Moosavi-Movahedi,a,* Shahram Hakimelahi,b Jamshid Chamani,a
Ghadam Ali Khodarahmi,c Farshid Hassanzadeh,c Fen-Tair Luo,d Tai Wei Ly,d
Kak-Shan Shia,e Chi-Feng Yen,e Moti L. Jain,f Ramasamy Kulatheeswaran,d
Cuihua Xue,d Manijeh Pasdarb and Gholam Hossein Hakimelahia,d,e,
*
aInstitute of Biochemistry-Biophysics, Tehran University, Tehran, Iran
bDepartment of Cell Biology, Faculty of Medicine, University of Alberta, Edmonton, Alberta, Canada, T6G 2H7
cDepartment of Medicinal Chemistry, Faculty of Pharmacy, Isfahan University of Medical Sciences, Isfahan, Iran
dInstitute of Chemistry, Academia Sinica, Taipei, Taiwan 115, ROC
eTaiGen Biotechnology, 138 Shin Ming Rd., Taipei, Taiwan 114, ROC
fDepartment of Chemistry and Biochemistry, University of Arkansas, Fayetteville, AR 72701, USA
Received 29 May 2003; revised 25 June 2003; accepted 29 July 2003
Abstract—Synthesis of adenine derivative of triphosphono-g-(Z)-ethylidene-2,3-dimethoxybutenolide 4 was accomplished by
treatment of phosphonate 3 with 5-phosphoribosyl 1-pyrophosphate in the presence of 5-phosphoribosyl 1-pyrophosphate synthe-
tase. It was found that triphosphonate 4 functions as an irreversible stoichiometric inactivator of the Escherichia coli ribonucleoside
diphosphate reductase (RDPR). Triphosphonate 4 exhibited potent inhibitory activity against murine leukemias (L1210 and P388),
breast carcinoma (MCF7), and human T-lymphoblasts (Molt4/C8 and CEM/0) cell lines. Paclitaxel ester derivatives of adenine-
containing triphosphono-g-(Z)-ethylidene-2,3-dimethoxybutenolide 8–10 were also synthesized. Like triphosphonate 4, compound
8 exhibited inhibitory property toward RDPR. It also induced microtubule assembly similar to paclitaxel (5). The structure of the
chlorodiester linker in 8 was found to account for this dual property. After treatment of MCF7 cells with compounds 4, 5, and 8,
fluorescence microscope examination demonstrated the presence of nucleus shrinkage or segmentation. Bifunctional prodrug 8
exhibited higher lipophilicity than 4 and higher water-solubility than 5. Pro-dual-drug 8 exhibited more pronounced anticancer
activity relative to that of the triphosphonate 4 and paclitaxel (5). In contrast, compound 9, resulting from the linkage of triphos-
phonate 4 and paclitaxel (5) through a diester unit, was only found to function as a highly water-soluble prodrug for paclitaxel (5).
It induced microtubule assembly in vitro, but did not show inhibitory property toward RDPR. On the other hand, compound 10,
an aggregate of triphosphonate 4 and paclitaxel (5), neither functioned as an inhibitor of RDPR nor exhibited microtubule
assembly stimulating activity in vitro.
# 2003 Elsevier Ltd. All rights reserved.
Introduction
manifold is ribonucleoside diphosphate reductase
(RDPR) (EC 1.17.4.1).2 RDPR catalyzes the reduction
of four purine and pyrimidine ribonucleotides by direct
substitution of the 20-OH group of the ribonucleotides
with hydrogen.3,4 The physiological reducing species are
the mercapto groups of thioredoxin, which are regener-
ated from the oxidized form by NADPH-dependent
thioredoxin reductase.5,6 Inhibition of reductases by
certain ribonucleotide analogues obstructs the route for
the replication of the genetic materials necessary for
Mechanism-based inhibition of enzymes crucial to
metabolic pathways involved in cell division is an
attractive concept for the rational design of chemo-
therapeutic agents.1 Such an enzyme in the nucleic acid
*Corresponding authors. Tel.: +886-2-2790-1861; fax: +886-2-2796-
3606; e-mail: hosein@chem.sinica.edu.tw
0968-0896/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0968-0896(03)00524-8

4304
A. A. Moosavi-Movahedi et al. / Bioorg. Med. Chem. 11 (2003) 4303–4313
cancer cell division.7 However, a well-known problem in
anticancer therapy through inhibition of RDPR is the
bioavailability of this type of phosphorylated deriva-
tives. Nucleoside-50-diphosphate or triphosphate analo-
gues do not easily penetrate cellular membranes due to
their low lipophilicity.8
cellular delivery of triphosphonate 4 from a highly
lipophilic prodrug (i.e., 8) is expected to enhance its
intracellular inhibitory property toward RDPR, which
may result in profound activity against malignant tumor
cell lines.
Attempts to increase lipophilicity of RDPR inhibitors
are heretofore undescribed. The induction of high
water-solubility to paclitaxel ester derivative 8 by its
triphosphonate moiety, encouraged us to use compound
4 in particular as RDPR inhibitor. As such, two com-
pounds, triphosphonate 4 and paclitaxel (5), with differ-
ent modes of action are linked to each other. The adducts
8, 9, and 10 overcome the physiochemical disadvantages
of both parent compounds. Depending on the spacer,
both parent compounds, only one compound, or none of
them are released under physiological conditions.
On the other hand, paclitaxel (Taxol1, 5) is a highly
lipophilic anticancer agent for a variety of tumors (i.e.,
L1210 and P388 leukemias, MX-1 mammary tumor,
B16 melanoma, and CX-1 colon tumor xenografts),9
due to its unique ability to inhibit cell division as well as
other interphase processes by stabilizing microtubules.9
In the presence of paclitaxel, microtubules resist depoly-
merization, thus interfering with the G2 and M phases
of the cell cycle.9d A major problem encountered in the
use of paclitaxel for cancer chemotherapy is one of low
water-solubility, which makes its formulation for injec-
tion difficult.10 To date, a great deal of research has
been devoted towards the modification of paclitaxel in
order to create a more water-soluble and, therefore,
more easily formulated and delivered drug.
Results
Chemical and biological reactions in the synthesis of
adenine-containing triphosphono-ꢀ-(Z)-ethylidene-2,3-di-
methoxybutenolide 4 (Scheme 1)
Prodrug strategy may provide a utilitarian solution to
the biodistribution of potential drugs having either low
lipophilicity or low water-solubility. This strategy con-
sists of transient modification of the physiochemical
properties of a given compound through chemical deri-
vatization. Such temporary chemical modification is
usually designed to alter bioavailability while the inher-
ent pharmacological properties of the parent drug
remain intact.10 Prodrugs are designed, once the barrier
to delivery has been circumvented, to be converted into
the active drugs, in vivo, either by an enzymatic
mechanism10 or simple hydrolysis8,11 initiated under
physiological pH conditions.
6-Chloropurine-containing dimethylphosphono-g-(Z)-
ethylidene-2,3-dimethoxybutenolide 112 was reacted
with NH4OH in refluxing CH3CN to produce the cor-
responding adenine monoammonium phosphonate salt
2 as a mixture of two diastereoisomers in 70% yield.
Sequential treatment of 2 with Me3SiBr/DMF and H2O
gave phosphonic acid 3 in 45% yield over two steps.
Adenine-containing
phosphono-g-(Z)-ethylidene-2,3-
dimethoxybutenolide 3 and adenosine 50-monophos-
phate (AMP), as a reference compound, were individu-
ally incubated with PRPP and PRPP synthetase during
which time, the reaction proceeded linearly. The assays
were terminated after 6 h by the addition of MeOH and
the crude products purified on a DEAE Sephadex col-
umn with ammonium carbonate to afford diphosphate 4
(triphosphate equivalent, 47% yield) and ATP (>95%
yield), respectively.13
Herein we report the synthesis of adenine-containing
triphosphono-g-(Z)-ethylidene-2,3-dimethoxy-
butenolide 4, which was found to be an alternate sub-
strate for RDPR and exhibited notable anticancer
activity. Polar molecule 4 possesses high water-solubi-
lity, but very low lipophilicity. Consequently, the intra-
The structure of 4 was confirmed by its cleavage with
alkaline phosphatase first to a diphosphate analogue then
to monophosphate 3 (>90% yield). On the other hand,
compound 4 was found not to be a substrate for snake
venom or spleen phosphodiesterase. It remained intact
even after 24 h incubation at physiological conditions.
Kinetic parameters for PRPP synthetase with phospho-
nate 3
Using an established procedure,13 the kinetic para-
meters for PRPP synthetase were determined. The sub-
strate affinity of phosphonate 3 (Km=6.5 mM) for the
enzyme was found to be 27-times less than that of AMP
(Km=0.24 mM). The Vmax for conversion of monopho-
sphonate 3 (Vmax=0.12 mmol/unit/h) to triphosphonate
4 is 120 times lower than that for the conversion of
AMP (Vmax=14.5 mmol/unit/h) to ATP. Inhibitory
effect of phosphonate 3 (Ki=46.5 mM) toward phos-
phorylation of AMP was also evaluated.13
Scheme 1. Preparation of adenine-containing triphosphono-g-(Z)-
ethylidene-2,3-dimethoxybutenolide 4.

A. A. Moosavi-Movahedi et al. / Bioorg. Med. Chem. 11 (2003) 4303–4313
4305
Synthesis of paclitaxel ester derivatives of adenine-
containing triphosphono-ꢀ-(Z)-ethylidene-2,3-dimethoxy-
butenolide 8 and 9 (Scheme 2)
Lipophilicity and solubility tests
Lipophilicity and water-solubility were determined by
distribution between 1-octanol and water according to
the methods reported by Baker et al.14 (Table 1). Pacli-
taxel derivatives of adenine-containing triphos-
phonobutenolide 8, 9, and 10 were observed to exhibit
higher water-solubility than that exhibited by paclitaxel
(5). Prodrugs 8 and 9, as well as compound 10, also
exhibited higher lipophilicity as compared to adenine-
containing triphosphonobutenolide 4.
Acylation of paclitaxel (5) with dichloroacetyl chloride
in the presence of Et3N in CH2Cl2 at À15 ꢀC gave 20-
dichloroacetyl taxol (6) (95% yield). Alkylation of tri-
phosphonate 4 with compound 6 (DMF, 25 ꢀC) then
afforded prodrug 8 (45% yield). Likewise, treatment of
paclitaxel (5) with bromoacetyl bromide in the presence
of Et3N in CH2Cl2 at À15 ꢀC gave 20-bromoacetyl taxol
(7) (90% yield), which was subsequently alkylated with
triphosphonate 4 to give prodrug 9 (44% yield).
In general, the solubility of impure polar molecules (i.e.,
4) in organic solvents (i.e., 1-octanol) is higher than that
of their pure forms. Accordingly, we observed 27 times
increment in the lipophilicity of 4 when an equimolar
mixture of 4+5 was tested (see Table 1). However, this
combination did not improve the water-solubility of
paclitaxel (5), as expected.
Synthesis of adenine-containing paclitaxel 20-triphos-
phono-ꢀ-(Z)-ethylidene-2,3-dimethoxybutenolide
(Scheme 3)
10
Condensation of triphosphonate 4 with paclitaxel (5) in
DMF using diisopropylcarbodiimide (DIPC) as con-
densing agent gave phosphate ester 10 in 35% yield.
Hydrolysis of paclitaxel ester derivatives 8–10 in deuter-
ated water, phosphate buffer solution, and human plasma
(Scheme 4)
Paclitaxel-containing triphosphonate 8 was completely
hydrolysed to paclitaxel (5), triphosphonate 4, and
glyoxylic acid in a mixture of DMSO-d6 and distilled
D2O (pD 5.8, 1:1 mL/mL), a mixture of DMSO-d6 and
phosphate buffer solution (pH 7.2, 1:1 mL/mL), and in
human plasma at 37 ꢀC after 17.0, 2.0, and 0.40 h,
respectively. Analysis of the hydrolysis reaction in
DMSO-d6/D2O after 7.0 h by 1H NMR showed the
presence of triphosphonate ester 11, suggesting that it is
the intermediate in the formation of triphosphonate 4.
On the other hand, hydrolysis of paclitaxel-containing
triphosphonate 9 in DMSO-d6/D2O (43 h), DMSO-d6/
phosphate buffer solution (4.5 h), and in human plasma
(0.65 h) produced paclitaxel (5) and phosphoester deri-
vative 12. In this case, triphosphonate 4 or glyoxylic
acid was not detected. However, paclitaxel-containing
butenolide 10 was found to be intact in water or in
buffer solution at 37 ꢀC even after 55 h; yet in human
plasma, it was hydrolyzed completely to paclitaxel (5)
and triphosphonate 4 within 15 h.
Scheme 2. Synthesis of paclitaxel ester derivatives of adenine-contain-
ing triphosphono-g-(Z)-ethylidene-2,3-dimethoxybutenolide 8 and 9.
Table 1. Solubility and lipophilicity of triphosphonate 4, paclitaxel
(5), an equimolar mixture of 4+5, prodrugs 8 and 9, as well as the
triphosphonate ester 10
Compd
Solubility in
water (mM) 1-octanol (mM)
Solubility in
log P (1-octanol/
water)a
Triphosphonate 4
43Â106
58.6
48.9
59.0Â106
58.2Â106
56Â103
58Â103
69Â103
À5.9
6.0
Paclitaxel (5)
4+5b
4+5c
8
412Â105
1327
À4.5
56.7
6.0
0.3
29Â103
26Â103
14Â103
9
10
0.3
0.7
aPartition coefficients were calculated as P=[substrate]1-octanol/[sub-
strate]H O.
2
bSolubility in water and lipophilicity of 4 in an equimolar mixture of
4+5.
cSolubility in water and lipophilicity of 5 in an equimolar mixture of
4+5.
Scheme 3. Synthesis of adenine-containing paclitaxel 20-triphosphono-
g-(Z)-ethylidene-2,3-dimethoxybutenolide 10.

4306
A. A. Moosavi-Movahedi et al. / Bioorg. Med. Chem. 11 (2003) 4303–4313
ester derivative 10 remained intact. The conversions of 8
and 9 to paclitaxel (5) may account for their activity on
microtubule assembly.
Anticancer activity in vitro
Inhibition of the proliferation of murine leukemias
(L1210 and P388), breast carcinoma (MCF7), human T-
lymphocytes (Molt4/C8 and CEM/0), human embryo-
nic lung cell (HEL), and normal fibroblast (Hef522) by
50% (IC50)19 in the presence of compounds 2–10, 12,
glyoxylic acid, 9-(b-d–arabinofuranosyl)cytosine (ara-
C), equimolar mixtures of 4+6, 5+11, and
4+5+glyoxylic acid were carried out in vitro. Further-
more, compounds 4, 5, and 8–10 were tested for inhibi-
Scheme 4. Hydrolysis of the prodrugs 8 and 9 to their respective
parent drugs.
tion of the proliferation of B16 melanoma cells.18
A
mixture of DMSO/D2O (8:2 v/v) was used to solve the
compounds and their mixtures. The results are listed in
Table 2.
Inhibition of Escherichia coli RDPR
Using methodology previously described by Stubbe et
al.,15 compounds 4, 8, 9, 10, and 12 were evaluated for
inhibitory property against E. coli RDPR (EC 1.17.4.1).
The remaining enzyme activity was assayed as described
by Steeper and Steuart.16 The E. coli RDPR (1.35 mM)
was significantly inactivated upon incubation with tri-
phosphonate 4 or prodrug 8 by concentration-depen-
dent process. While inactivation of the enzyme achieved
immediately by 4, its inactivation by paclitaxel-contain-
ing triphosphonate 8 occurred after ca. 2 h. This is
consistent with the rate of hydrolysis of prodrug 8 to
parent drug 4 in the reaction media. The maximal inac-
tivation at 10, 25, 50, and 100 mM concentrations used
was 100% of the original enzyme activity. On the other
hand, at concentrations as low as 0.1 and 1.0 mM of 4,
instantaneous inactivation (t1/2 <10 s) of 10–15% and
69–78% of the enzyme was observed, respectively. No
change in the remaining enzyme activity was detected
over 10 s to 2 h. Such rapid inactivation did not allow
the determination of time-dependent inactivatory prop-
erty of triphosphonate 4 or its transport form 8 towards
RDPR. Chromatography of the inactivated enzyme on
a Sephadex G-50 column, resulted in no recovery of the
RDPR activity. In addition, dithiothreitol (DTT) did
not protect RDPR from inactivation by 4 or 8. Thus,
the inactivation may be due to an irreversible stoichio-
metric complexation of 4 with the active site of the
enzyme.17 On the other hand, compounds 9, 10, and 12
were found not to be inhibitors of RDPR.
Effect of adenine-containing triphosphonobutenolide 4,
paclitaxel (5), and bifunctional prodrug 8 on microtubule
bundle formation in MCF7 cells
The effects of compounds 4, 5, and 8 at their IC50 (Table 2)
on microtubule cycloskeletone were examined by
immunofluorescence microscopy using tubulin anti-
bodies (see Experimental). Triphosphonate 4 at 0.34 mM
did not exhibit differential microtubule assembly as
compared with the untreated MCF7 cells after 6.5 h.
Under the same conditions, however, both paclitaxel (5)
(2.71Â10À3 mM) and prodrug 8 (0.20Â10À3 mM) showed
similar ability to induce the formation of microtubule
bundles as well as different distribution of microtubules
in apoptotic MCF7 cells (Fig. 1).
Apoptotic morphology in MCF7 cells caused by adenine-
containing triphosphonobutenolide 4, paclitaxel (5), or
bifunctional prodrug 8
MCF7 cells were treated with higher concentrations
than IC50 of compounds 4 (4ÂIC50), 5 (2ÂIC50), and 8
(2ÂIC50). Chromatin staining of the cells revealed
nucleus shrinkage or fragmentation after 8.0 h (Fig. 2).
This morphological change is associated with cell death
occurring by a process called apoptosis.
Effect of paclitaxel ester derivatives of adenine-containing
triphosphono-ꢀ-(Z)-ethylidene-2,3-dimethoxybutenolide
8–10 on microtubule assembly in vitro
Discussion
An unusual potential pathway for the intracellular con-
version of monophosphonate 3 to its biologically active
anabolite 4 is demonstrated. The rate of phosphoryla-
tion of 3 to 4 by PRPP synthetase is, however, 120-times
less than that of the AMP. Thus, phosphonate 3 did not
exhibit high activity against malignant tumor cell lines.
Triphosphonate 4 showed significant inhibitory prop-
erty towards RDPR. Consequently, it exhibited notable
anticancer activity (see Table 2).
The ability of paclitaxel (5), prodrugs 8 and 9, as well as
20-paclitaxel phosphoester 10 to induce microtubule
assembly in vitro at a concentration of 7.0 mM were
examined according to an established procedure.18 The
percent of tubulin which had polymerized was deter-
mined to be 70% for paclitaxel (5), 64% for 8, and 34%
for 9. 20-Paclitaxel phosphoester 10 failed to exhibit
microtubule assembly stimulating activity in vitro.
HPLC analysis, after 3.5 h, demonstrated that 20-pacli-
taxel esters 8 and 9 were degraded, respectively, to 97
and 42% paclitaxel (5), while the paclitaxel phospho-
Because lower dosage of two drugs in combination has
shown better efficacy than individual drug applied
alone,20a the combined modality may be beneficial in

A. A. Moosavi-Movahedi et al. / Bioorg. Med. Chem. 11 (2003) 4303–4313
4307
Table 2. Inhibitory effects, IC50 (mM), of butenolides 2 and 3, triphosphonate 4, paclitaxel (5), 20-dichloroacetyl taxol (6), 20-bromoacetyl taxol (7),
prodrugs 8 and 9, equimolar mixtures of 4+6, 5+11, and 4+5+glyoxylic acid as well as the triphosphonate esters 10 and 12, glyoxylic acid, and
ara-C in DMSO/D2O (8:2 v/v) on the growth of malignant tumor cell lines, human embryonic lung cell (HEL), and normal fibroblast (Hef522)a
Compd
L1210
P388
MCF7
Molt4/C8
CEM/0
B16 melanoma
HEL
Hef522
ara-C
2
3
4
Paclitaxel (5)
6
0.17Æ0.00
85.49Æ2.71
11.42Æ1.15
0.62Æ0.01
4.50Æ10À3
4.36Æ10À3
5.47Æ10À3
0.54Æ10À3
2.90Æ10À3
2.01Æ10À3
2.34Æ10À3
3.89Æ10À3
5.70Æ0.11
45.63Æ1.38
69.86Æ1.94
>120
0.14Æ0.01
49.87Æ1.98
7.10Æ0.87
0.58Æ0.02
5.64Æ10À3
5.29Æ10À3
6.83Æ10À3
0.62Æ10À3
2.31Æ10À3
1.98Æ10À3
2.03Æ10À3
4.16Æ10À3
4.28Æ0.07
20.32Æ2.06
77.67Æ3.41
>120
1.03Æ0.02
>120
0.65Æ0.00
92.65Æ3.07
9.50Æ1.92
0.75Æ0.04
6.71Æ10À3
6.80Æ10À3
7.21Æ10À3
0.49Æ10À3
4.65Æ10À3
3.98Æ10À3
4.17Æ10À3
0.78Æ0.10
>120
—
—
1.25Æ0.13
74.18Æ2.45
45.76Æ1.11
35.87Æ0.98
0.73Æ0.06
0.58Æ0.04
0.47Æ0.01
0.86Æ0.02
—
1.98Æ0.05
80.46Æ2.96
62.17Æ1.87
53.14Æ1.26
1.89Æ0.09
2.05Æ0.07
1.72Æ0.08
2.45Æ0.03
—
13.90Æ1.01
0.34Æ0.01
2.71Æ10À3
2.34Æ10À3
3.95Æ10À3
0.20Æ10À3
1.77Æ10À3
1.02Æ10À3
1.56Æ10À3
3.01Æ10À3
7.99Æ0.86
68.74Æ1.70
80.09Æ2.14
>120
14.60Æ1.12
0.81Æ0.13
7.03Æ10À3
6.31Æ10À3
5.96Æ10À3
0.81Æ10À3
4.06Æ10À3
3.67Æ10À3
3.99Æ10À3
6.70Æ10À3
6.51Æ0.14
68.75Æ2.55
97.84Æ3.09
>120
—
15.60Æ1.01
21.50Æ10À3
—
—
7
Prodrug (8)
4+6
5+11
4+5+glyoxylic
acid
Prodrug (9)
10
12
Glyoxylic acid
DMSO/D2O
(8:2 v/v)
3.20Æ10À3
—
—
—
—
—
—
—
5.92Æ10À3
5.83Æ0.09
>120
17.00Æ10À3
1.20Æ0.04
—
0.94Æ0.00
113Æ3.67
83.49Æ2.14
120Æ3.75
>120
1.21Æ0.06
120Æ2.35
97.54Æ3.14
>120
74.25Æ2.81
—
>120
>120
>120
aThe IC50 values were estimated from dose–response curves compiled from three independent experiments and represent the compound concen-
tration (mM) required to inhibit cell proliferation by 50%. The mean (ÆSD) for 5–9, 4+6, 5+11, and 4+5+glyoxylic acid was <0.04Â10À3 in all
cases, except in B16 melanoma where (ÆSD) was <0.87Â10À3
.
the treatment of patients who cannot tolerate high
dosages of the drugs. One approach to combination
therapy involves the linkage of two anticancer agents
targeting at different pathways associated with the cell
division. As such, the cancer cells are expected to
become more susceptible to a dual drug, which can be
judged by its lower IC50 relative to the parent drugs
(compare IC50 values of 4, 5, and 8 in Table 2).
ferent cellular targets, was predicted. Indeed, when the
key experiments for a pro-dual-drug concept were car-
ried out, it was found that equimolar mixtures of 4+6,
5+11, and 4+5+glyoxylic acid exhibited about two
times more anticancer activity than paclitaxel (5), but
less activity relative to that of prodrug 8 (see Table 2).
20-Paclitaxel esters with electron-withdrawing sub-
stituents in the a-position show remarkable rate
enhancements of hydrolysis.10 Similarly, from a hydro-
lytic perspective, the ease of hydrolysis of novel prodrug
8 to paclitaxel (5) and phosphoester intermediate 11 is
due to the presence of electron-withdrawing substituents
(i.e., Cl and OPO2O–) at the a-position of its 20-ester
unit. Likewise, the terminal phosphate in 11 is intrinsi-
cally more labile due to the chloro-substitution in a-
position—note that phosphate hydrolysis may occur
either by water attack at the phosphorus center, accel-
erated by an inductive or hyperconjugative a-chloro
effect, or by water attack at the C–Cl bond, followed by
rapid cleavage of the hemiacetal to afford triphos-
phonobutenolide 4 and glyoxylic acid. As such, prodrug
8 and triphosphonate 4 exhibited similar inhibitory
property toward RDPR.
The ability of a drug to penetrate the cell membrane and
exhibit biological activity can be correlated to its lipo-
philicity.8,20b Triphosphonate 4 possessed high water-
solubility, but very low lipophilicity (Table 1). Conse-
quently, it was conjugated with paclitaxel (5), a highly
lipophilic antitumor agent, via a chlorodiester unit. In
comparison with triphosphonate 4, the lipophilicity of
the resulting bifunctional prodrug 8 was about 1000
times. Furthermore, prodrug 8 was also found to exhibit
500 times more water solubility than paclitaxel (5).
Thus, in comparison with adenine-containing triphos-
phonobutenolide 4 and paclitaxel (5), paclitaxel 20-tri-
phosphonate ester 8 possessed superior bioavailability
and greater anticancer activity.
Like impure organic compounds, a mixture of polar
molecule 4 with paclitaxel (5) showed more lipophilicity
for adjuvant compound 4 in comparison with individual
pure triphosphonate 4 (see Table 1). As such, an incre-
ment in the cell penetration of 4 is expected. Triphos-
phonate 4 and its derivative 11 are alternate substrates
for RDPR, which is crucial to a metabolic pathway
involved in the cell division. On the other hand, pacli-
taxel (5) and its derivative 6 inhibit cancer cell division
by stabilizing microtubules. Thus, relative to anticancer
activity of individual compounds, a synergistic effect on
the activity of their mixtures, capable of attacking dif-
The concept of pro-dual drug was further corroborated
by studying microtubule bundle formation and apopto-
sis in MCF7 cells in the presence of prodrug 8 as well as
parent drugs 4 and 5. It is reported that introduction of
an acyl moiety at the 20-position of paclitaxel resulted in
the loss of ability to promote microtubule assembly, but
not cytotoxicity.18 The ability of bifunctional prodrug 8
to induce microtubule assembly in vitro or inside MCF7
cells (see Fig. 1) is due to its hydrolysis to paclitaxel (5).
As shown in Fig. 2, chromatin staining of MCF7 cells
incubated with 4 (1.36 mM), 5 (5.42Â10À3 mM), or 8

4308
A. A. Moosavi-Movahedi et al. / Bioorg. Med. Chem. 11 (2003) 4303–4313
(0.40Â10À3 mM) revealed different numbers of apoptotic
nuclei. Indeed, the ability of compounds 4, 5, and 8 to
induce apoptosis correlates well with their anticancer
activity (Table 2).
Unlike triphosphonate 4, its phosphoester derivative 12
did not inhibit RDPR; yet exhibited weak anticancer
activity (see Table 2). This may be due to the slow rate
of hydrolysis of 12 to 4 by the action of phospho-
esterases inside the malignant tumor cell lines. On the
basis of this observation, and also to realize the impor-
tance of the linker in the biological activity of prodrugs
8 and 9, we synthesized paclitaxel 20-triphosphono-g-
(Z)-ethylidene-2,3-dimethoxybutenolide 10, which was
found to be about 250 times more soluble in water than
paclitaxel (5). We hypothesized that, in the presence of
phosphodiesterases, paclitaxel (5) and triphosphonate 4
may be liberated inside the tumor cells as potential
drugs. Derivative 10, however, was found to be stable
toward hydrolysis in either buffer solution (pH 7.2) or
in water (pH 5.8). Consequently, it did not show activity
on microtubule assembly. The activity of compound 10
against the examined malignant tumor cell lines in vitro
was found to be about 1000 times less than that of
paclitaxel (5), and at least 10 times less than that of the
triphosphonate 4 (Table 2). Accordingly, the controlled
release of the drugs based on the enzymatic cleavage of
To prove the rationale behind the design of our new
bifunctional prodrug 8, we synthesized 20-paclitaxel
ester derivative 9 in which a chlorine atom is replaced
with a hydrogen atom. Ester 9, under physiological
conditions, was hydrolyzed to paclitaxel (5) and stable
phosphoester 12. Prodrug 9 exhibited similar anticancer
activity in vitro when compared to paclitaxel (5); yet it
did not exhibit inhibitory property against RDPR.
Unlike phosphoester intermediate 11, phosphoester 12
(see Scheme 4) was not hydrolyzed to triphosphonate 4
under physiological conditions. This result confirms the
importance of chlorine atom at the a-position of the
ester linkage in dual-action anticancer prodrug 8. It
should be noted that since the linking moiety employed
is an ester, substantial rate enhancement due to non-
specific esterases may also occur in vivo and contribute
to the effective use of the prodrugs.
Figure 1. Immunofluorescence staining of microtubules in control and drug-treated MCF7 cells. Microtubules were labeled with a mouse mono-
clonal anti-tubulin antibody and detected with an FITC-conjugated secondary anti-mouse IgG: (A) untreated cells; (B) cells treated with 2.94 mM
adenine-containing triphosphonobutenolide 4; (C) treatment with 2.71Â10À3 mM of paclitaxel (5); (D) treatment with 0.98Â10À3 mM of bifunctional
prodrug 8. Microtubule distributions of apoptotic cells are indicated by white arrows.

A. A. Moosavi-Movahedi et al. / Bioorg. Med. Chem. 11 (2003) 4303–4313
4309
Figure 2. Drug-induced apoptosis in MCF7 cells: (A) untreated cells; (B) cells treated with 11.76 mM adenine-containing triphosphonobutenolide 4;
(C) treatment with 5.42Â10À3 mM of paclitaxel (5); (D) treatment with 1.96Â10À3 mM of bifunctional prodrug 8. Apoptotic nuclei are indicated by
white arrows.
paclitaxel 20-triphosphonate 10 was seemed to not be an
effective approach.
They also showed much higher lipophilicity as com-
pared to adenine-containing triphosphonobutenolide 4.
By virtue of its linker, bifunctional prodrug 8 was easily
hydrolyzed to paclitaxel (5) and triphosphonate 4 under
physiological conditions. Therefore, the biological
activity of 8 stems from both the adenine-containing
triphosphonobutenolide (RDPR inhibitor) and pacli-
taxel (microtubule assembly activator) moieties. Under
identical conditions, the hydrolysis of prodrug 9 to
paclitaxel (5) and triphosphonate derivative 12 was
found to be much slower, with biological activity only
comparable to that of paclitaxel (5). Paclitaxel 20-tri-
phosphonate 10, however, was found to not act as a
significant prodrug.
Finally, chemical stability of paclitaxel derivatives is
critical to their formulation since partial degradation of
conjugates can lead to precipitation of the poorly water
soluble paclitaxel. Compounds 8–10 in water (3.0 mg/
mL) remain clear and precipitate free for >15 h. They
also appear to show reasonable solid-state stability on
storage at 25 ꢀC.
Conclusions
A novel adenine-containing triphosphono-g-(Z)-ethyl-
idene-2,3-dimethoxybutenolide 4 was synthesized by
chemical and biological strategies. Compound 4 is a
significant inhibitor for RDPR and exhibited notable
anticancer activitiy against murine leukemias (L1210
and P388), breast carcinoma (MCF7), and human T-
lymphoblasts (Molt4/C8 and CEM/0) cell lines. Tri-
phosphonate 4 was found not to be significantly active
against proliferation of B16 melanoma cells. On the
other hand, monophosphonate 3 exhibited weak activity
against malignant tumor cell lines. This lack of sig-
nificant activity is proposed to be a consequence of its
slow rate of phosphorylation by PRPP synthetase in
cellular system. Paclitaxel ester derivatives of adenine-
Experimental
General
For anhydrous reactions, glassware was dried overnight
in an oven at 120 ꢀC and cooled in a desiccator over
anhydrous CaSO4 or silica gel. Reagents including E.
coli alkaline phosphatase, NADPH, ATP, and Hoechst
33342 were purchased from Sigma (St. Louis, MO,
USA) or Fluka (Switzerland). Antibodies were obtained
from Boehringer Mannheim (Indianapolis, IN, USA).
Solvents, including dry ether and tetrahydrofuran
(THF), were obtained by distillation from the sodium
ketyl of benzophenone under nitrogen. Other solvents,
including chloroform, dichloromethane, ethyl acetate,
and hexanes were distilled over CaH2 under nitrogen.
Absolute methanol and ethanol were purchased from
Merck (Germany) and used as received. E. coli RDPR
was isolated as described.15b
containing
triphosphono-g-(Z)-ethylidene-2,3-dime-
thoxybutenolide 8 and 9 as well as adenine-containing
paclitaxel 20-triphosphono-g-(Z)-ethylidene-2,3-dime-
thoxybutenolide 10 were also synthesized. Paclitaxel
derivatives 8–10 were observed to show much higher
water solubility than that exhibited by paclitaxel (5).

4310
A. A. Moosavi-Movahedi et al. / Bioorg. Med. Chem. 11 (2003) 4303–4313
Melting points were obtained with a Buchi 510 melting
point apparatus. Infrared (IR) spectra were recorded on
a Beckman IR-8 spectrophotometer. The wavenumbers
reported are referenced to the 1601 cmÀ1 absorption of
polystyrene. NMR spectra were obtained on a Varian
XL-300 (300 MHz) Spectrometer. DMSO-d6, D2O, or
CDCl3 was used as solvent; Me4Si (d 0.00 ppm) was
used as an internal standard. All NMR chemical shifts
are reported as d values in parts per million (ppm) and
coupling constants (J) are given in hertz (Hz). The
splitting pattern abbreviations are as follows: s, singlet;
d, doublet; t, triplet; q, quartet; br, broad; m, unre-
solved multiplet due to the field strength of the instru-
ment; and dd, doublet of doublets. For 31P NMR,
acetone-d6 or D2O was used as solvent; PO(Me)3 (d 1.59
ppm) was used as an internal standard. UV–vis spec-
troscopy was carried out using an HP8452A diode array
spectrophotometer. Microanalyses were performed on a
Perkin-Elmer 240-B microanalyzer.
chromatography (resine XAD-4, H2O) to afford 3 (0.17
g, 0.45 mmol) in 45% yield. Rf=0.09 (EtOAc/
MeOH=4:1); mp 264 ꢀC (dec.); IR (KBr): n=3100–
0
0
3450 (OH, NH2), 3112 (C8 H), 3098 (C2 H), 3021 (C5H),
1776 (C¼O), 1689 (C¼C), 1215 (P¼O) cmÀ1; UV
(EtOH): lmax (lge)=210 (4.57), 259 nm (4.19); 1H NMR
(300 MHz, DMSO-d6/D2O): d=3.78 (s, 3H, C2OCH3),
4.19 (s, 3H, C3OCH3), 6.32 (dd, J=16.4, 32.5 Hz, 1H,
PCH), 7.19 (br d, J=16.9 Hz, 1H, ¼CH), 7.78, 8.12 (2
s, 2H, C2 H+C8 H); 13C NMR (300 MHz, DMSO-d6):
d=60.20 (2-OCH3),60.29 (3-OCH3), 84.08 (C6), 91.65
0
0
0
(C5), 124.00 (C4), 135.10 (C5 ), 137.41 (C2), 143.15 (C2 ),
0
0
0
0
149.65 (C4 ), 153.08 (C3), 157.21 (C6 ), 157.43 (C8 ),
166.27 (C¼O);31P NMR (300 MHz, acetone-d6):
d=14.67; MS (70 eV): m/z (%): 383 (70) [M, phospho-
rus cluster]+; elemental analysis calcd (%) for
C13H14N5O7P (383.2576): C, 40.74; H, 3.68; N, 18.27; P,
8.09; found: C, 40.68; H, 3.55; N, 18.29; P 8.13.
5-Phosphoribosyl 1-pyrophosphate synthetase in the pre-
Purification refers to gravity column chromatography
on a DEAE Sephadex or Merck Silica Gel 60 (particle
size 230–400 mesh). Analytical TLC was performed on
precoated plates purchased from Merck (Silica Gel 60
F254). Compounds were visualized by use of UV light.
paration
of
(Z)-4-[2-triphosphono-2-(adenine-9-yl)-
ethylidene]-2,3-dimethoxy-Dꢁ,ꢂ-butenolide (4) and ATP.
Substrate affinities of phosphonate 3 and AMP for
PRPP synthetase as well as the inhibitory effect of 3
against the enzyme were evaluated according to an
established procedure.13 The resulting triphosphonate 4
(47% yield) and ATP (>95% yield) were, individually,
purified by using a DEAE Sephadex A-25 column
(1.5Â15 cm) that had been equilibrated in ammonium
carbonate (20 mM). A linear gradient of ammonium
carbonate (400 mL) was used to elute 4. The purity of 4
was determined to be 99.6% by analytical HPLC
(anion-exchange Partisphere column eluted with a gra-
dient of 8.0 mM to 1.0 M ammonium phosphate, pH
5.47) monitored at 260 nm: mp >250 ꢀC (dec.); UV
(EtOH): lmax (lge)=214 (4.38), 260 nm (4.27); 1H NMR
(300 MHz, D2O): d=3.60 (s, 3H, C2OCH3), 4.25 (s, 3H,
C3OCH3), 6.82 (dd, J=17.2, 33.1 Hz, 1H, PCH), 7.19
(Z)-4-[2-Ammonium methylphosphono-2-(adenine-9-yl)-
ethylidene]-2,3-dimethoxy-Dꢁ,ꢂ-butenolide (distereoiso-
meric mixture, 2). To a solution of 1 (4.12 g, 9.60
mmol) in CH3CN (50 mL) was added concentrated
NH4OH solution (100 mL). The solution was refluxed
for 6 h. The solvents were evaporated, and the residue
was crystallized from EtOH to give 2 (2.69 g, 6.72
mmol) in 70% yield. Rf=0.14 (EtOAc/MeOH=4:1);
mp 211–214 ꢀC; IR (KBr): n=3210–3153 (NH2), 3065
0
0
(C8 H), 3054 (C2 H), 2945 (C5H), 1780 (C¼O), 1695
(C¼C), 1221 (P¼O) cmÀ1; UV (EtOH): lmax (lge)=212
1
(4.60), 260 nm (4.25); H NMR (300 MHz, DMSO-d6/
D2O): d=3.71 (d, J=11.5 Hz, 1.5H, P(OCH3)), 3.76 (d,
J=11.5 Hz, 1.5H, P(OCH3)), 3.80 (br s, 3H, C2OCH3),
4.24 (br s, 3H, C3OCH3), 6.14 (dd, J=16.0, 31.8 Hz,
0.5H, PCH), 6.23 (dd, J=16.2, 32.2 Hz, 0.5H, PCH),
6.96–7.23 (m, 1H, ¼CH), 7.65, 7.83, 8.32 8.41 (4 s, 2H,
(dd, J=17.3, 22.3 Hz, 1H, ¼CH), 7.99, 8.30 (2 s, 2H,
31
0
0
C2 H+C8 H); P NMR (300 MHz, D2O): d=17.86 (d,
J=28.5 Hz, a-P), À22.48 (dd, J=19.6, 28.5 Hz, b-P),
À7.97 (d, J=19.6 Hz, g-P).
C2 H+C8 H); 13C NMR (300 MHz, DMSO-d6):
d=51.87, 52.01 (POCH3), 59.98, 60.01 (2-OCH3), 60.20,
60.46 (3-OCH3), 82.16, 83.14 (C6), 98.79, 99.87 (C5),
Treatment of triphosphonate 4 with E. coli alkaline phos-
phatase. Triphosphonate 4 (24 mM) was treated with
alkaline phosphatase (60 IU/mL) and 2-amino-2-
methylpropanol (65 mM, pH 9.8) at 37 ꢀC. Over 12 h
period sequential conversion of triphosphonate 4 to
diphosphonate and then to monophosphonate 3 was
observed. After 15 h, chromatography was performed
on thin-layer PEI-cellulose in 0.8 M LiCl/0.8 M formic
acid 1:1 (v/v) (Rf=0.09 for triphosphonate 4, 0.26 for
diphosphonate analogue, and 0.64 for monophos-
phonate 3) to afford 3 in >90% yield.
0
0
0
126.13, 127.51 (C4), 132.98, 133.04 (C5 ), 141.11, 142.20
0
(C2), 144.62, 145.31 (C2 ), 147.98, 148.45 (C4 ), 149.79,
0
0
0
150.12 (C3), 155.86, 156.41 (C6 ), 156.95, 157.02 (C8 ),
167.89, 167.96 (C¼O);31P NMR (300 MHz, acetone-d6):
d=16.37, 16.45; MS (DCl, 200 eV): m/z (%): 415 (95)
[M+1, phosphorus cluster]+; elemental analysis calcd
(%) for C14H19N6O7P (414.3480): C, 40.58; H, 4.62; N,
20.28; P, 7.48; found: C, 40.47; H, 4.58; N, 20.30; P,
7.41.
20-Dichloroacetylpaclitaxel (6). To a solution of pacli-
taxel (5) (0.146 g, 0.170 mmol) and Et3N (0.018 g, 0.18
mmol) in dry CH2Cl2 (8.0 mL) at À15 ꢀC was added
dropwise dichloroacetyl chloride (0.027 g, 0.18 mmol).
The reaction mixture was stirred for 2.0 h. The solution
was then partitioned between CH2Cl2 (30 mL) and
water (40 mL). The CH2Cl2 solution was washed with
water (2Â50 mL); then it was dried over MgSO4 and
(Z)-4-[2-Dihydrogenphosphono-2-(adenine-9-yl)ethylidene]-
2,3-dimethoxy-Dꢁ,ꢂ-butenolide (3). To a solution of 2
(0.41 g, 1.0 mmol) in DMF (10 mL) was added Me3SiBr
(1.09 g, 7.02 mmol). After the solution was stirred at
25 ꢀC for 7 h, a mixture of MeOH and H2O [6:3 (v/v),
18 mL] was added, and the solvents were evaporated.
The crude residue was purified by use of column

A. A. Moosavi-Movahedi et al. / Bioorg. Med. Chem. 11 (2003) 4303–4313
4311
filtered. Evaporation under reduced pressure and puri-
fication of the residue by use of column chromatography
(CH2Cl2) afforded 6 (0.15 g, 0.16 mmol) in 95% yield.
Rf=0.29 (hexanes/EtOAc=3:1); mp 171–173 ꢀC (dec.); 1H
NMR (300 MHz, CDCl3): d=1.07 (s, 3H, H3C17), 1.24
(s, 3H, H3C16), 1.67 (s, 3H, H3C18), 1.84 (s, 3H, H3C19),
2.23 (s, 3H, OCOCH3), 2.52–2.79 (m, 2H, H2C6), 2.36
(s, 3H, OCOCH3), 2.31–2.34 (m, 2H, H2C14), 3.71 (d,
J=6.8 Hz, 1H, HC3), 4.19 (d, J=8.3 Hz, 1H, HC20),
4.28 (d, J=8.3 Hz, 1H, HC20), 4.36 (dd, J=31.3, 8.3
Hz, 1H, HC7), 4.50–4.86 (br, 2H, 2ÂOH), 4.94 (br d,
1H, HC5), 5.65 (d, J=7.0 Hz, 1H, HC2), 5.77 (d, J=2.4
8.27; Cl, 2.33; P, 6.11; found: C, 48.86; H, 4.81; N, 8.32;
Cl, 2.30; P, 6.07.
(Z)-4-[2-(Paclitaxel-20 -O-carbomethyl)triphosphono-2-
(adenine-9-yl)ethylidene]-2,3-dimethoxy-Dꢁ,ꢂ-butenolide
(9). Compound 9 (0.43 g, 0.29 mmol) was prepared in
44% yield from 7 (0.392 g, 0.666 mmol) and ammonium
triphosphonate 4 (0.408 g, 0.666 mmol) in DMF (18.0
mL) by the method used for the synthesis of 8. Rf=0.24
(EtOAc); mp >250 ꢀC (dec.); 1H NMR (300 MHz,
CDCl3/D2O): d=1.09 (br s, 3H, H3C17), 1.22 (br s, 3H,
H3C16), 1.80 (br s, 3H, H3C18), 1.84 (br s, 3H, H3C19),
2.25 (s, 3H, OCOCH3), 2.48–2.76 (m, 2H, H2C6), 2.44
(br s, 3H, OCOCH3), 2.30–2.52 (m, 2H, H2C14), 3.81 (br
d, 1H, HC3), 3.80 (s, 3H, C2OCH3), 4.26 (s, 3H,
C3OCH3), 4.33 (br d, 2H, H2C20), 4.50 (m, 1H, HC7),
5.10 (br d, 1H, HC5), 5.69 (br d, 1H, HC2), 5.79 (br s,
0
0
Hz, 1H, HC2 ), 5.95 (dd, J=8.8, 2.4 Hz, 1H, HC3 ),
6.21–6.28 (m, 3H, HCCl2+HC10+HC13), 6.87 (d,
J=8.8, 1H, NH), 7.23–7.90 (m, 15H, 3ÂC6H5); MS
(FAB): m/z (%): 965 (84%) [M+1, chlorine cluster]+;
elemental analysis calcd (%) for C49H51NO15Cl2
(964.8399): C, 61.00; H, 5.33; N, 1.45; Cl, 7.35; found:
C, 60.97; H, 5.41; N, 1.50; Cl, 7.34.
0
2H, OCH2CO), 5.84 (br d, 1H, HC2 ), 6.02 (br d, 1H,
0
HC3 ), 6.10–6.52 (m, 3H, HC10+HC13+PCH), 7.28–
8.13 (m, 16H, ¼CH+3ÂC6H5), 7.70, 8.12 (2 br s, 2H,
20-Bromoacetylpaclitaxel (7). Compound 7 (0.149 g,
0.153 mmol) was prepared in 90% yield from 5 (0.146 g,
0.170 mmol) and bromoacetyl bromide (0.036 g, 0.18
mmol) in the presence of Et3N (0.018 g, 0.18 mmol) in
dry CH2Cl2 (8.0 mL) by the method used for the synth-
esis of 6. Rf=0.32 (hexanes/EtOAc=3:1); mp 163–
C2 H+C8 H); elemental analysis calcd (%) for
C62H76N9O28P3 (1,488.2900): C, 50.04; H, 5.15; N,
8.47; P, 6.25; found: C, 50.18; H, 5.27; N, 8.52; P,
6.39.
0
0
(Z)-4-[2-(Paclitaxel-20-O-triphosphono)-2-(adenine-9-yl)-
ethylidene]-2,3-dimethoxy-Dꢁ,ꢂ-butenolide (10). Ammo-
nium triphosphonate 4 (0.30 g, 0.49 mmol) was
dissolved in acetic acid (3.0 mL). After 5 min, the sol-
vent was evaporated and the residue was dissolved in
DMF (7.0 mL). To this solution were added paclitaxel
(5) (0.42 g, 0.49 mmol), DIPC (0.064 g, 108 mL, 0.51
mmol) and DMAP (0.10 g, 0.78 mmol). The resulting
solution was stirred at 25 ꢀC for 24 h. The reaction
mixture was diluted with water (20 mL) and extracted
with EtOAc (40 mL). The EtOAc solution was washed
with water (30 50 mL); then it was dried over MgSO4
and filtered. Evaporation under reduced pressure and
purification of the residue by use of silica gel column
chromatography (EtOAc) afforded the corresponding
triacid of 10 (MS (FAB): m/z (%):1,380 (65) [M+1,
phosphorus cluster]+), which was suspended in ammo-
nium carbonate solution (20 mL, 150 mM) and stirred
for 1 h. Filtration afforded triammonium salt 10 (0.24 g,
0.17 mmol) in 35% yield. Rf=0.31 (EtOAc); mp 224–
227 ꢀC (dec.); 1H NMR (300 MHz, CDCl3/D2O):
d=1.12 (s, 3H, H3C17), 1.19 (br s, 3H, H3C16), 1.67 (br
s, 3H, H3C18), 1.81 (s, 3H, H3C19), 2.17 (br s, 3H,
OCOCH3), 2.50–2.80 (m, 2H, H2C6), 2.33 (s, 3H,
OCOCH3), 2.29–2.48 (m, 2H, H2C14), 3.81 (br d, 1H,
HC3), 3.78 (s, 3H, C2OCH3), 4.21 (s, 3H, C3OCH3), 4.21
(br d, 2H, H2C20), 4.39 (m, 1H, HC7), 4.75 (br d, 1H,
166 ꢀC (dec.); H NMR (300 MHz, CDCl3): d=1.08 (s,
1
3H, H3C17), 1.26 (s, 3H, H3C16), 1.69 (s, 3H, H3C18),
1.83 (s, 3H, H3C19), 2.21 (s, 3H, OCOCH3), 2.53–2.80
(m, 2H, H2C6), 2.38 (s, 3H, OCOCH3), 2.30–2.33 (m,
2H, H2C14), 3.70 (d, J=6.8 Hz, 1H, HC3), 4.17 (d,
J=8.5 Hz, 1H, HC20), 4.27 (d, J=8.5 Hz, 1H, HC20),
4.36 (dd, J=30.8, 8.2 Hz, 1H, HC7), 4.50–5.01 (br, 2H,
20 OH), 4.88 (br d, 1H, HC5), 5.59 (d, J=2.5 Hz, 1H,
0
HC2 ), 5.60 (s, 2H, CH2Br), 5.67 (d, J=7.1 Hz, 1H,
0
HC2), 5.97 (dd, J=8.8, 2.5 Hz, 1H, HC3 ), 6.26–6.29 (m,
2H, HC10+HC13), 6.85 (d, J=8.8, 1H, NH), 7.21–7.86
(m, 15H, 30 C6H5); MS (FAB): m/z (%): 975 (79) [M+1,
bromine cluster]+; elemental analysis calcd (%) for
C49H52NO15Br (974.8458): C, 60.37; H, 5.38; N, 1.44;
Br, 8.20; found: C, 60.21; H, 5.49; N, 1.37; Br, 8.18.
(Z)-4-[2-(Paclitaxel-20-O-carbo-ꢁ-chloromethyl)triphos-
phono-2-(adenine-9-yl)ethylidene]-2,3-dimethoxy-Dꢁ,ꢂ
-
butenolide (8). To a solution of 6 (0.322 g, 0.333
mmol)) in DMF (15.0 mL) at 25 ꢀC was added ammo-
nium triphosphonate 4 (0.204 g, 0.333 mmol). The mix-
ture was stirred for 16.0 h. The solvent was then
evaporated under reduced pressure and the residue was
crystallized from MeOH to give 8 (0.23 g, 0.15 mmol) in
45% yield. Rf=0.21 (EtOAc); mp >250 ꢀC (dec.); H
1
NMR (300 MHz, CDCl3/D2O): d=1.10 (br s, 3H,
H3C17), 1.26 (br s, 3H, H3C16), 1.82 (br s, 3H, H3C18),
1.80 (br s, 3H, H3C19), 2.22 (s, 3H, OCOCH3), 2.50–
2.83 (m, 2H, H2C6), 2.41 (br s, 3H, OCOCH3), 2.28–
2.41 (m, 2H, H2C14), 3.76 (br d, 1H, HC3), 3.76 (s, 3H,
C2OCH3), 4.20 (s, 3H, C3OCH3), 4.23 (br d, 2H,
H2C20), 4.45 (m, 1H, HC7), 5.07 (br d, 1H, HC5), 5.72
0
HC5), 5.58 (br t, 1H, HC2 ), 5.67 (br d, 1H, HC2), 5.93 (br
0
d, 1H, HC3 ), 6.18–6.54 (m, 3H, HC10+HC13+PCH),
7.18–8.09 (m, 16H, ¼CH+3ÂC6H5), 7.79, 8.36 (2 s,
0
0
2H, C2 H+C8 H); MS (FAB): m/z (%): 1,431 (48)
[M+1, phosphorus cluster]+; elemental analysis calcd
(%) for C60H74N9O26P3 (1,430.2499): C, 50.39; H, 5.22;
N, 8.81; P, 6.50; found: C, 50.27; H, 5.10; N, 8.97; P,
6.69.
0
(br d, 1H, HC2), 5.81 (br d, 1H, HC2 ), 5.99 (br d, 1H,
0
HC3 ), 6.11–6.53 (m, 4H, HCCl+HC10+HC13+PCH),
7.20–7.94 (m, 16H, ¼CH+30 C6H5), 7.69, 8.08 (2 br s,
0
0
2H, C2 H+C8 H); elemental analysis calcd (%) for
C62H75N9O28ClP3 (1,522.7820): C, 48.90; H, 4.96; N,
Lipophilicity and solubility tests. An individual solution
of 4, 5, an equimolar mixture of 4+5, and 8–10 (30.0

4312
A. A. Moosavi-Movahedi et al. / Bioorg. Med. Chem. 11 (2003) 4303–4313
mg) in 1-octanol (15.0 mL) was shaken with distilled
water (pH 5.8, 15.0 mL) for 2.0 h. The layers were
separated, and their concentrations were determined by
an UV spectrophotometer at 227 (for 5, 4+5, and 8–10)
or 260 (for 4, 4+5, and 8–10) nm.14,21 The solubility
and lipophilicity data obtained for compounds 8–10,
possessing a joint skeleton of 4 and 5, at UV absorbance of
paclitaxel (5) (227 nm) were found to be consistent (limit
error Æ13%) with those obtained for 8–10 at UV absor-
bance of adenine-containing triphosphonobutenolide 4
(260 nm). Reported results are the average of both
determinations for 8–10 in duplicate (Table 1).
of each tube was transferred into another vials, diluted
with distilled water (300 mL), and then extracted with
CHCl3 (300 mL) to remove paclitaxel (5), which was
found to be identical with an authentic sample. The
aqueous solutions were, respectively, analyzed by
HPLC, employing an anion-exchange Partisphere col-
umn, using a gradient of 8.0 mM to 1.0 M ammonium
phosphate as mobile phase. We observed that com-
pound 8 was hydrolyzed to 4, 5, and glyoxylic acid after
0.40 h. Compound 9 was hydrolyzed to 5 and 12 after
0.65 h, and compound 10 was completely converted to 4
and 5 in human plasma media after 15 h.
Hydrolysis of prodrugs 8, 9, and 20-paclitaxel phospho-
ester 10 in a mixture of DMSO-d6 and D2O (1:1 mL/
mL, pD 5.8) or a mixture of DMSO-d6 and phosphate
buffer (1:1 mL/mL, pD 7.2)-(1HNMR study). Com-
pounds 8, 9, and 10 (0.10 mmol) were individually dis-
solved in a mixture of DMSO-d6 (2.0 mL) and D2O (2
mL, pD 5.8) or a mixture of DMSO-d6 (2.0 mL) and
Evaluation of the effectiveness of triphosphonate 4, pacli-
taxel triphosphonate esters 8–10, and triphosphonate
ester 12 against RDPR. All compounds 4, 8–10, and 12
were evaluated for inhibitory property against E. coli
RDPR (EC 1.17.4.1),15 and the remaining enzyme
activity was determined.16 The following is a repre-
sentative procedure: E. coli RDPR was isolated as
described,15b and used by the method of Baker et al.15a
Briefly, HEPES (50.0 mM, pH 7.60), MgSO4 (15.0
mM), EDTA (1.0 mM), ATP (1.60 mM), NADPH (0.50
mM), thioredoxin (12.0 mM), thioredoxin reductase
(0.79 mM), R1 subunit (12.8 mM), and R2 subunit (12.8
mM) along with variable concentrations of the inhibitor
4 were incubated at 37 ꢀC for different times (20 s–20
min). During the reactions, the respective solutions
scanned repeatedly from 190 to 620 nm. The remaining
enzyme activity for the conversion of CDP to dCDP
was determined by the method of Steeper and Steuart.16
Control experiments were performed with no inhibitor
in the incubation mixture.
1
deuterated phosphate buffer (2.0 mL, pD 7.2). The H
NMR spectra at 25 ꢀC were taken immediately. The
mixtures were then incubated at 37 ꢀC for various peri-
ods of time. The spectra of 8 completely changed to that
of the eliminated compounds 4, 5, and glyoxylic acid
after 17.0 h (in DMSO-d6/D2O) and 2.0 h (in DMSO-d6/
deuterated phosphate buffer), respectively. On the other
hand, after 7.0 h in DMSO-d6/D2O, prodrug 8 was
hydrolyzed to paclitaxel (5) and intermediate 11. Each
mixture was extracted with CDCl3 (2Â3.0 mL) to
remove paclitaxel (5), which was found to be identical
with an authentic sample. Each aqueous solution was
frozen and lypholized to afford chloro-intermediate 11
or a mixture of triphosphonate 4 and glyoxylic acid as
evidenced by 1H NMR analysis. For 11: mp 130–132 ꢀC
(dec.); 1H NMR (300 MHz, DMSO-d6/D2O): d=3.63 (s,
3H, C2OCH3), 4.24 (s, 3H, C3OCH3), 6.01 (br s, 1H,
OCHCl), 6.78–6.99 (m, 1H, PCH), 7.10–7.31 (m, 1H,
Microtubule assembly experiment. The ability of pacli-
taxel (5), prodrugs 8 and 9, as well as 20-paclitaxel
phosphoester 10 to induce microtubule assembly in
vitro were examined.18 The following is a representative
procedure: To PEM buffer solution (0.10 M Pipes, pH
7.1, 1.0 mM EGTA, and 0.8 mM MgSO4) was added
tubulin (12.0 mM), bifunctional prodrug 8 (7.0 mM), and
GTP (7.0 mM). The increase in turbidity (microtubule
polymerization) was monitored at 30 ꢀC by the increase
in apparent absorbance at 350 nm using a temperature-
controlled spectrophotometer. To determine the
amount of assembled tubulin, the samples were cen-
trifuged (200,000g for 4.0 min) at 37 ꢀC. The pellet was
dissolved in cold PEM buffer and allowed to depoly-
merize on ice for 25 min. The protein concentration in
the supernatant after centrifugation at 4 ꢀC was deter-
mined by the method of Bradford.22
0
0
¼CH), 7.85, 8.26 (2 br s, 2H, C2 H+C8 H). Following
the same method, compound 9 was found to be con-
verted to 5 and 12 after 43 h (in DMSO-d6/D2O) and 4.5
h (in DMSO-d6/deuterated phosphate buffer), respec-
tively, while 10 was stable even after 55 h. For 12: mp
>250 ꢀC (dec.); UV (EtOH): lmax (lge)=216 (4.24), 260
1
nm (4.19); H NMR (300 MHz, D2O): d=3.65 (s, 3H,
C2OCH3), 4.30 (s, 3H, C3OCH3), 5.52 (s, 2H,
OCH2CO), 6.85 (dd, J=18.0, 32.3 Hz, 1H, PCH),
7.15–7.28 (m, 1H, ¼CH), 7.86, 8.24 (2 br s, 2H,
0
0
C2 H+C8 H).
Hydrolysis of paclitaxel triphosphonate esters 8–10 in
the presence of human plasma. Paclitaxel triphos-
phonate esters 8–10 were separately dissolved in CHCl3
at a concentration of 15 mg/mL. From each stock solu-
tion 40 mL was transferred, individually, to 1.5-mL
tubes and the solvent in each tube was evaporated by
bubbling nitrogen gas into the respective solution. A
few drops of human plasma then was added to each
tube, and the mixtures were incubated at 37 ꢀC for var-
ious periods of time. In all cases, 2-propanol (300 mL)
was added at the proper intervals to precipitate the
proteins, and the resulting solution was vortexed and
centrifuged for 4.0 min. Supernatant solution (250 mL)
Anticancer test procedure in vitro. Murine leukemias
(L1210 and P388), breast carcinoma (MCF7), human T-
lymphoblasts (Molt4/C8 and CEM/0), human embryo-
nic lung cell (HEL), and normal fibroblasts (Hef522)
were cultured in DMEM supplemented with 10% FBS,
2.0 mM glutamine, 100 U/mL penicillin, and 100 mg/mL
streptomycin in a humidified atmosphere with 5% CO2
at 37 ꢀC.19 Under this condition, the generation time for
L1210, P388, MCF7, Molt4/C8, CEM/0, HEL, and
Hef522 cells was about 13, 12, 17, 18, 21, 45, and 48 h,
respectively. Compounds 2–10, 12, glyoxylic acid,

A. A. Moosavi-Movahedi et al. / Bioorg. Med. Chem. 11 (2003) 4303–4313
4313
ara-C, and equimolar mixtures of 4+6, 5+11, and
References and Notes
4+5+glyoxylic acid in DMSO/D2O (8:2 v/v), at var-
ious concentrations, were added to the cells (200 cells/
mL) in their exponential phase of growth. The cell
numbers of the control cultures, as well as that of the
cultures supplemented with the test compounds, were
determined after 24, 48, and 72 h of growth. The IC50
values were estimated from dose–response curves com-
piled from three independent experiments and represent
the compound concentration (mM) required to inhibit
proliferation of the respective malignant tumor cell lines
by 50% after 72 h incubation (Table 2).
1. Silverman, R. B. Mechanism-Based Enzyme Inactivation:
Chemistry and Enzymology; CRC: Boca Raton, FL, 1988, p 3.
2. Thelander, L.; Reichard, P. Annu. Rev. Biochem. 1979, 48, 133.
3. Sando, G. N.; Hogenkamp, H. P. C. Biochemistry 1973, 12,
3316.
4. Berglund, O. J. Biol. Chem. 1972, 247, 7270.
5. Hakimelahi, G. H.; Nemer, M. J. Sci. Tech. (Shiraz) 1985,
10, 1.
6. Blakley, R. L. J. Biol. Chem. 1966, 241, 176.
7. Goldberg, I. H. Acc. Chem. Res. 1991, 24, 191.
8. Meier, C.; Lorey, M.; De Clercq, E.; Balzarini, J. J. Med.
Chem. 1998, 41, 1417, and references cited therein.
9. (a) Souto, A. A.; Acuna, A. U.; Andreu, J. M.; Barasoain,
I.; Abal, M.; Amat-Guerri, F. Angew. Chem., Int. Ed. Engl.
1995, 34, 2710, and references cited therein. (b) Schiff, P. B.;
Fant, J.; Horwitz, S. B. Nature 1979, 277, 665. (c) Wiernik,
P. H.; Schwartz, E. L.; Strauman, J. J.; Dutcher, J. P.; Lipton,
R. B.; Phase, E. Cancer Res. 1987, 47, 2486, and references
cited therein. (d) Schiff, P. B.; Horwitz, S. B. Proc. Natl. Acad.
Sci. U.S.A. 1980, 77, 1561.
10. Greenwald, R. B.; Gilbert, C. W.; Pendri, A.; Conover,
C. D.; Xia, J.; Martinez, A. J. Med. Chem. 1996, 39, 424, and
references cited therein.
11. Oliyai, R.; Stella, V. J. Annu. Rev. Pharmacol. Taxicol
1993, 32, 521.
12. Hakimelahi, G. H.; Mei, N.-W.; Moosavi-Movahedi,
A. A.; Davari, H.; Hakimelahi, S.; King, K.-Y.; Hwu, J. R.;
Wen, Y.-S. J. Med. Chem. 2001, 44, 1749.
13. Balzarini, J.; De Clercq, E. J. Biol. Chem. 1991, 266, 8686
and references cited therein..
B16 melanoma cell proliferation. Inhibition of prolifera-
tion of B16 melanoma cells by compounds 4, 5, and 8–
10 were studied according to the reported procedure by
Mathew et al.18 Results are illustrated in Table 2.
Observation of microtubule assembly in MCF7 by immu-
nofluorescence microscopy. Cells were grown on cover-
slips and remained untreated or were incubated with
adenine-containing triphosphonobutenolide
4 (0.34
3
mM), paclitaxel (5) (2.71Â10À mM), or bifunctional
prodrug 8 (0.20Â10À3 mM).23 After 6.5 h, coverslips
were air-dried and fixed with freshly prepared 3.7%
paraformaldehyde in phosphate-buffered saline for 10.0
min. Residual aldehyde was quenched with 0.10 M gly-
cine in PBS for 5.0 min. Cells were permeabilized for 2.0
min at room temperature with 0.5% Triton X-100 in
PBS and blocked for 14.0 h at 4.0 ꢀC with normal goat
serum in PBS containing 0.2% bovine serum albumin
and 50.0 mM NH4Cl (1:1000 mL/mL). Microtubule
staining was carried out by primary anti-E7 mouse
antibody in PBS containing 0.2% BSA (1:20 mL/mL) for
45.0 min. Then, it was treated with goat secondary anti-
mouse FITC Fab fragment in PBS containing 0.2%
BSA (1:100 mL/mL) for 20.0 min. The coverslips were
mounted on glass slides by elvanol containing 0.2%
paraphenylene diamine and viewed under a Zeiss fluor-
escence microscope.
14. Baker, D. C.; Haskell, T. H.; Putt, S. R. J. Med. Chem.
1978, 21, 1218.
15. (a) Baker, C. H.; Banzon, J.; Bollinger, J. M.; Stubbe, J.;
Samano, V.; Robins, M. J.; Lippert, B.; Jarvi, E.; Resvick, R.
J. Med. Chem. 1991, 34, 1879. (b) Salowe, S. P.; Stubbe, J. J.
Bacteriol. 1986, 165, 363.
16. Steeper, J. R.; Steuart, C. D. Anal. Biochem. 1970, 34, 123.
17. For stoichiometric mechanism-based inhibitors, see: Salowe,
S. P.; Ator, M. A.; Stubbe, J. Biochemistry 1987, 26, 3408.
18. Mathew, A. E.; Mejillano, M. R.; Nath, J. P.; Himes,
R. H.; Stella, V. J. J. Med. Chem. 1992, 35, 145, and references
cited therein.
19. (a) Raic
M.; Pavelic, K.; De Clercq, E.; Mintas, M. J. Med. Chem.
1999, 42, 2673. (b) Raic-Malic, S.; Svedruzic, D.; Gazivoda,
T.; Marunovic, A.; Hergold-Brundic, A.; Nagl, A.; Balzarini,
J.; De Clercq, E.; Mintas, M. J. Med. Chem. 2000, 43, 4806.
20. (a) Schinazi, R. F.; Peters, J.; Williams, C. C.; Chance, D.;
Nahmias, A. J. Antimicrob. Agents Chemother. 1982, 22, 499.
(b) Hakimelahi, G. H.; Ly, T. W.; Moosavi-Movahedi, A. A.;
Jain, M. L.; Zakerinia, M.; Davari, H.; Mei, H.-C.; Sambaiah,
T.; Moshfegh, A. A.; Hakimelahi, S. J. Med. Chem. 2001, 44,
3710, and references cited therein.
-Malic, S.; Hergold-Brundic, A.; Nagl, A.; Grdisa,
´ ´ ´
Observation of apoptotic morphology in MCF7 cells by
a fluorescence microscope. Cells were grown on cover-
slips and remained untreated or were incubated with
´
´
´
´
´
´
adenine-containing triphosphonobutenolide
4 (1.36
mM), paclitaxel (5) (5.42Â10À3 mM), or bifunctional
prodrug 8 (0.4Â10À3 mM) for 8.0 h. The coverslips were
prepared in the same manner as described above. Nuclei
were labeled with Hoechst 33342 in PBS (1.0 mg mLÀ1
)
for 15.0 min in the dark at room temperature.24 The
coverslips were mounted on glass slides and viewed
under a Zeiss fluorescence microscope.
21. Hakimelahi, G. H.; Moosavi-Movahedi, A. A.; Sadegi,
M. M.; Tsay, S.-C.; Hwu, J. R. J. Med. Chem. 1995, 38, 4648,
and references cited therein..
22. Bradford, M. M. Anal. Biochem. 1970, 72, 248.
23. Swindell, C. S.; Krauss, N. E.; Horwitz, S. B.; Ringel, I. J.
Med. Chem. 1991, 34, 1176, and references cited therein.
24. Hakimelahi, S.; Parker, H. R.; Gilchrist, A. J.; Barry, M.;
Li, Z.; Bleackley, R. C.; Pasdar, M. J. Biol. Chem. 2000, 275,
10905.
Acknowledgements
For financial support, we thank the Ministry of Science
and Higher Education of Iran, Academia Sinica, and
Tehran University Research Council.