5′-(2-Nitrophenylalkanoyl)-2′-deoxy-5-fluorouridines as potential prodrugs of FUDR for reductive activation

Article abstract of DOI:10.1016/S0968-0896(03)00426-7

Four 5′-(2-nitrophenylalkanoyl)-2′-deoxy-5-fluorouridines (1a-d) were designed and synthesized as potential prodrugs of FUDR for reductive activation. Two methyl groups were introduced α to the ester carbonyl to increase both the rate of cyclization activation and the stability of the conjugates towards serum esterases. Chemical reduction of the nitro group into an amino leads to cyclization and release of the active FUDR. Kinetic analysis of the cyclization activation process indicates that the two methyl groups α to the ester carbonyl restrict the rotational freedom of ground state molecule and promote the cyclization reaction. However, the two methyl groups also were found to render the conjugates as poor substrates of E. coli B nitroreductase. Conjugate 1c, without the two methyl groups, was reduced by E. coli B nitroreductase (t_1/2=8 h) to give two products, a N-hydroxyl lactam and the drug FUDR, suggesting that the enzymatic reduction and subsequent cyclization activation proceeded through the hydroxylamine intermediate. These results indicate that cyclization activation will occur once the nitro group is reduced either to an amino or to a hydroxylamino group. The fact that the amino intermediates cyclized easily to release the incorporated drug FUDR suggests the feasibility of using peptide-linked acyl 2-aminophenylalkanoic acid esters as potential prodrugs for proteolytic activation.

Full text of DOI:10.1016/S0968-0896(03)00426-7

Bioorganic & Medicinal Chemistry 11 (2003) 3889–3899
50-(2-Nitrophenylalkanoyl)-20-deoxy-5-fluorouridines as Potential
Prodrugs of FUDR for Reductive Activation§
Bin Liuy and Longqin Hu*
Department of Pharmaceutical Chemistry, Ernest Mario School of Pharmacy, Rutgers, The State University of New Jersey, 160
Frelinghuysen Road, Piscataway, NJ 08854, USA
Received 24 March 2003; revised 3 June 2003; accepted 22 June 2003
Abstract—Four 50-(2-nitrophenylalkanoyl)-20-deoxy-5-fluorouridines (1a–d) were designed and synthesized as potential prodrugs of
FUDR for reductive activation. Two methyl groups were introduced a to the ester carbonyl to increase both the rate of cyclization
activation and the stability of the conjugates towards serum esterases. Chemical reduction of the nitro group into an amino leads to
cyclization and release of the active FUDR. Kinetic analysis of the cyclization activation process indicates that the two methyl
groups a to the ester carbonyl restrict the rotational freedom of ground state molecule and promote the cyclization reaction.
However, the two methyl groups also were found to render the conjugates as poor substrates of E. coli B nitroreductase. Conjugate
1c, without the two methyl groups, was reduced by E. coli B nitroreductase (t1/2=8 h) to give two products, a N-hydroxyl lactam
and the drug FUDR, suggesting that the enzymatic reduction and subsequent cyclization activation proceeded through the
hydroxylamine intermediate. These results indicate that cyclization activation will occur once the nitro group is reduced either to an
amino or to a hydroxylamino group. The fact that the amino intermediates cyclized easily to release the incorporated drug FUDR
suggests the feasibility of using peptide-linked acyl 2-aminophenylalkanoic acid esters as potential prodrugs for proteolytic activation.
# 2003 Elsevier Ltd. All rights reserved.
Introduction
It is important to understand that bioreductions are
found in both oxygenated and hypoxic cells. Ideally, the
presence of oxygen would promote a redox cycle that
suppresses the net reduction of a hypoxia-selective pro-
drug to its cytotoxic species, whereas the reaction cas-
cade leads to the formation of the cytotoxic species in
the absence of sufficient oxygen tension.3
One of the characteristics of many solid tumors is the
insufficient supply of oxygen to the core tissue due to
poorly developed vasculature. These oxygen-deficient
tumor cells are refractory to radiation therapy and most
chemotherapeutic agents. As a result, they are capable
of proliferating and causing tumor re-growth after
treatments.1 On the other hand, hypoxic tumor cells are
known to have a greater capacity for reductive reactions
as compared to normal well-oxygenated cells. This
unique feature differentiates neoplastic tissues from the
normal tissues and provides an attractive target for
selective anticancer chemotherapy. Several bioreduc-
tively activated nitro compounds, quinines, and aro-
matic N-oxides are currently in clinical trials as
hypoxia-selective cytotoxins and could potentially be
developed into selective anticancer prodrugs.2
Activation of aromatic nitro compounds via bioreduc-
tion to form cytotoxic species has been the subject of
many investigations over the past 20 years.2,4 There
has been much interest in exploiting this transformation
to activate prodrugs in the hypoxic region of solid
tumors or by nitroreductases introduced site-specifically
into tumor cells.5 The nitro group has been shown
to undergo up to six-electron (6e) reduction by flavin
containing enzymes in cells. As shown in Figure 1,
the formation of a nitro anion radical is the early
process in this reduction. The hypoxia selectivity
arises from the fact that in the normoxic tissues, the
radical intermediate can react with molecular oxygen,
leading to the regeneration of the parent nitro
compound. Consequently, the reduction to generate
highly cytotoxic species will only occur in the hypoxic
tissue.3
§Taken in part from the Masters Thesis of Bin Liu, Rutgers, The State
University of New Jersey, 2000.
*Corresponding author. Tel.:+732-445-5291; fax:+732-445-6312;
e-mail: longhu@rci.rutgers.edu
yCurrent address: Chugai Pharma USA, LLC, 6275 Nancy Ridge
Drive, San Diego, CA 92121, USA.
0968-0896/03/$ - see front matter # 2003 Elsevier Ltd. All rights reserved.
doi:10.1016/S0968-0896(03)00426-7

3890
B. Liu, L. Hu / Bioorg. Med. Chem. 11 (2003) 3889–3899
Also, nitroreduction by nitroreductases is of current
interest because new methods such as the antibody-
directed enzyme prodrug therapy (ADEPT) and gene-
directed enzyme prodrug therapy (GDEPT) have been
developed to introduce a specific enzyme into tumors.
In ADEPT, a conjugate of a tumor specific antibody
and an enzyme is administered, and the antibody–
enzyme conjugates accumulate selectively on the surface
of the antigen-expressing tumor cells. After a sufficient
time interval for its clearance from the normal (non-
target) tissue and the localization of the antibody–
enzyme conjugates into the tumor tissue, the prodrug of
an anticancer agent is administered and activated speci-
fically at the tumor site. The active drug can be of low
molecular weight, which would allow rapid diffusion
and reach tumor regions not accessible to the antibody–
zation reaction and thereby result in the release of the
active drug. Four potential prodrugs of FUDR (1a–d)
have been designed based on the bioreductive activation
mechanism shown in Scheme 1. The free 50-hydroxyl
group of the deoxyribose sugar in FUDR (6) is required
for its cytotoxic activity. The hydroxyl group must be
phosphorylated by thymidine kinase to FdUMP and
other metabolites to affect DNA synthesis. In com-
pounds 1a–d, the introduction of an ester linkage masks
this hydroxyl group leading to deactivation of FUDR.
Scheme 1 illustrates the proposed mechanism of activa-
tion upon reduction for the FUDR prodrugs 1a–d
designed. The reduction of the aromatic nitro group in
1a–d could occur in hypoxic tumor tissues or by a
nitroreductase such as the one from E. coli. The pro-
ducts of these bioreductions are believed to be either the
corresponding hydroxylamines 2a–d or the correspond-
ing amines 3a–d. This conversion of an electron-with-
drawing nitro group into a nucleophilic hydroxylamine
or amine is expected to trigger a spontaneous intramo-
lecular cyclization. The hydroxylamino group in 2a–d or
the amino group in 3a–d would attack the electrophilic
ester carbonyl carbon, leading to the formation of the
corresponding N-hydroxyl lactam 4a–d or lactam 5a–d
and the release of the effector anticancer drug FUDR
(6).
enzyme conjugate, producing
a
bystander effect.6
GDEPT is a related approach to ADEPT, where the
DNA encoding for a prodrug-activating enzyme is
selectively expressed and translated within a tumor cell.
An aerobic nitroreductase from Escherichia coli B is
currently under evaluation for use in both ADEPT and
GDEPT. E. coli B nitroreductase can reduce certain
aromatic nitro groups into their corresponding hydrox-
ylamines in the presence of NADH or NADPH as a
cofactor. Prodrugs which are activated by this enzyme
fall into two classes.7 The first class is exemplified by
2,4-dinitrobenzamides (e.g., CB1954) and 2,4-dini-
trophenyl related nitrogen mustards (e.g., SN23862), for
which the nitro reduction has the potential to increase
the alkylating reactivity. The second class of substrates
includes 4-nitrobenzyloxycarbamyl derivatives of a
range of amine-bearing cytotoxins. The enzyme reduces
the nitro group into the corresponding hydroxylamine
in the molecule, which then undergoes fragmentation to
give the effector cytotoxin.8,9 In this paper, we report
the design, synthesis, and evaluation of nitro-containing
prodrugs of FUDR, where nitro reduction would lead
to a cyclization activation process and the release of the
active drug FUDR.
Different lengths of the side chain can result in the for-
mation of five- or six-membered lactam ring, which
could influence the rate of cyclization. The two methyl
groups attached to the a-position of the carbonyl were
expected to restrict the rotational freedom of the
ground-state conformation of the molecule, and there-
fore, place the amine or hydroxylamine in a more
favorable position with respect to the ester carbonyl, a
concept termed as ‘stereopopulation control’. In addi-
tion, the methyl substitution has been found to accel-
erate the rate of lactonization and lactamization, which
is known as the ‘Thorpe–Ingold’ effect.10 This has also
been used in other strategies based on intramolecular
cyclization reactions and has been reviewed.11
Results and Discussion
Design
Our hypothesis is to use the nitro group as an electronic
trigger, which upon reduction in hypoxic tissues or by a
nitro reductase can initiate an intramolecular cycli-
Scheme 1. Proposed mechanism of activation for the FUDR prodrugs
upon reduction.
Figure 1. Selectivity of nitroaromatics towards hypoxia.

B. Liu, L. Hu / Bioorg. Med. Chem. 11 (2003) 3889–3899
3891
Scheme 2. Synthesis of conjugates 1a and 1b: (a) FUDR, DEAD, PPh3, THF, rt 20 h, 83%; (b) SOCl2, MeOH, 0 ꢀC–rt, 100%; (c) MeI, NaH, 15-
crown-5, DMF, À4 to 0 ꢀC, 100%; (d) 1 N NaOH, MeOH, reflux, 6 h, 93%; (e) FUDR, DEAD, PPh3, dioxane, rt 18 h, 53%.
One concern of using prodrugs containing ester bonds is
the stability of such linkages in the presence of esterases
found in human serum. The two methyl substitutions
would provide some steric hindrance around the ester
bond to slow down such esterase-catalyzed hydrolysis
and the resulting prodrugs might be sufficiently stable
under physiological condition to be useful.
crown-5. Sodium hydroxide mediated hydrolysis con-
verted the dialkylated ester 9 to the corresponding car-
boxylic acid 10 in 93% yield for the three steps. Final
coupling of the acid 10 with FUDR under the Mis-
tunobu reaction conditions afforded the target com-
pound 1b in 53% yield. The overall yield for the
synthesis of conjugate 1b was 49%.
The synthesis of target compounds 1c and 1d was
accomplished as shown in Schemes 3 and 4, respec-
tively. Reduction of 2-nitrophenylacetic acid (7) using
borane in THF15 gave the corresponding alcohol 11,
which was converted to its tosylate and treated with
potassium cyanide in NMP at 90–100 ꢀC. The resulting
nitrile 12 was hydrolyzed using 50% sulfuric acid to give
the carboxylic acid 13, which was coupled to FUDR
under the same Mitsunobo conditions to afford the
desired conjugate 1c in 39% overall yield. As shown in
Scheme 4, treatment of acid 13 with thionyl chloride in
methanol afforded the methyl ester 14 in 92% yield,
which was dialkylated with methyl iodide and LDA at
À78 ꢀC. This step turned out to produce a very complex
product mixture, and the desired compound 15 was
obtained in only 9% yield after purification by flash
column chromatography. The low yield might be
attributed to the strong electron withdrawing nitro
group, which led to difficulty in selectively dialkylating
the a-position on the side chain. The methyl ester 15
Synthesis
50-(2-Nitrophenyl)acetyl-20-deoxy-5-fluorouridine 1a was
prepared directly from the commercially available 2-
nitrophenylacetic acid (7) and FUDR (Scheme 2). The
conjugation between the acid and the 50-hydroxyl group
of the deoxyribose in FUDR was accomplished using
diethyl azodicarboxylate (DEAD) and triphenylphos-
phine (PPh3) in 83% yield. The Mistunobu reaction
condition is very mild and highly regioselective for the
primary hydroxyl over the secondary hydroxyl group in
the deoxyribose sugar of FUDR.12,13 The structure of
1
conjugate 1a was confirmed by H NMR, COSY, and
HRMS.
The route towards the synthesis of compound 1b started
with the esterification of 2-nitrophenylacetic acid (7)
with thionyl chloride in methanol.14 The methyl ester 8
was further dialkylated using methyl iodide and sodium
hydride in the presence of catalytic amount of 15-
Scheme 3. Synthesis of conjugate 1c: (a) BH3/THF, rt, 90%; (b) TsCl, pyridine, rt, 3 h, 82%; (c) KCN, NMP, 90–100 ꢀC, 1.0 h, 86%; (d) 50%
H2SO4, 130–135 ꢀC, 4.5 h, 100%; (e) FUDR, DEAD, PPh3, THF, rt, 16 h, 61%.
Scheme 4. Synthesis of conjugate 1d: (a) SOCl2, MeOH, 0 ꢀC–rt, 1 h, 92%; (b) LDA, CH3I, THF, À78 ꢀC to rt, 9%; (c) LiOH, MeOH/H2O (3:1),
5 ꢀC, 15 h, 100%; (d) FUDR, DEAD, PPh3, THF, rt, 32 h, 53%.

3892
B. Liu, L. Hu / Bioorg. Med. Chem. 11 (2003) 3889–3899
was further hydrolyzed using lithium hydroxide in
methanol and water at 0 ꢀC to produce the acid 16 in
quantitative yield. Mistunobu reaction was again
employed for the conjugation of acid 16 with FUDR,
which afforded the desired compound 1d in 53% yield.
ribose sugar in FUDR contribute to the stability of
FUDR conjugates 1b and 1d.
Chemical reduction
To test the feasibility of releasing the active drug FUDR
from conjugates 1a–d, two mild chemical reductions
were carried out first on compound 1b. One was hydro-
genation using a hydrogen balloon and the other was
sodium borohydride-mediated reduction. Both reac-
tions were done in the presence of 10% Pd/C catalyst at
room temperature.16 These conditions should not affect
the other functional groups present in the molecule
while reducing the nitro group to the corresponding
amine. Our results indicated that the cyclization did
occur after reduction. The lactam 5b and FUDR were
the only products isolated under both reduction condi-
tions. The amino intermediate 3b, however, was not
isolated, suggesting that the cyclization step was very
fast.
Stability test
The stability of these potential prodrugs under physio-
logical conditions is an important issue, since the ester
linkage in these molecules might be susceptible to enzy-
matic hydrolysis by esterases present in human serum.
Therefore, the stability of 1a–d was tested by incubating
each sample in sodium phosphate buffer (100 mM, pH
7.4) and in human serum at 37 ꢀC. Acetonitrile was used
to help dissolve the compounds, and the concentration
of the organic solvent in the incubation mixture was
kept below 5%. Aliquots (25 mL) were taken at different
time intervals and analyzed by HPLC on a C-18
reversed-phase column using acetonitrile–water (0.1%
trifluoroacetic acid) as the mobile phase. In the case of
incubation in human serum, the active esterases were
quenched with 7% HClO4 solution (100 mL) immedi-
ately after the aliquots were withdrawn.13
Bioreduction in hypoxic tissues or reduction by nitro-
reductases would most likely stop at the hydroxylamine
intermediate. It is known that cyclization via the
hydroxylamine is faster than cyclization via the amine
intermediate.17,18 Thus, our chemical reduction condi-
tions are simple and practical tests of the cyclization
activation process.
All four compounds were found to be stable in the
phosphate buffer at 37 ꢀC over three days of incubation.
However, their stability in human serum at 37 ꢀC differs
from each other as shown in Figure 2. Compounds 1b
and 1d with the two methyl groups at the a-position to
the ester carbonyl were found to be stable in human
serum; there was no observable hydrolysis after 3 days
of incubation. On the other hand, compounds 1a and 1c
without the two methyl substitutions were subjected to
hydrolysis catalyzed by esterases with half-lives of 68.6
and 12.4 h, respectively. These results indicate that the
two methyl groups did protect the esters from enzymatic
hydrolysis. The hydrolysis of methyl 2-nitrophenylace-
tate was also tested under the same conditions and was
found to have a half-life of 19 h in human serum at
37 ꢀC, which was much shorter than the 68.6 h half life
for compound 1a. Clearly, the two methyl groups at the
a-position to the ester carbonyl as well as the deoxy-
Between the two methods of chemical reduction,
hydrogenation was found to be milder than the sodium
borohydride reduction; thus, hydrogenation was chosen
to test cyclization activation process of compounds 1a,
1c, and 1d. Reduction of compound 1d gave a similar
result as observed for compound 1b; no amino inter-
mediate 3d but cyclized lactam 5d and FUDR were iso-
lated. On the other hand, the amino intermediates 3a
and 3c were isolated after hydrogenation of compounds
1a and 1c, respectively. Incubation of the amino inter-
mediates 3a and 3c in phosphate buffer (100 mM, pH
7.4) at 37 ꢀC led to the cyclization and release of FUDR,
which was analyzed by HPLC and confirmed by spec-
troscopic techniques.
In order to compare the rate of cyclization and release
of FUDR from the conjugates 1a–d upon reduction, we
studied the kinetics of the cyclization reaction by mon-
itoring the changes in UV–vis absorbance. In the case of
compound 1a, for which the amino intermediate 3a was
isolated, the maximum absorbance wavelength of amine
3a, cyclized lactam 5a, and FUDR were determined to
be 271, 248, and 268 nm, respectively. The maximum
change in UV absorption before and after cyclization of
compound 3a was at 249 nm. For the kinetic assay, we
withdrew the filtrates directly after partial hydrogena-
tion (4 min) of the nitro compounds, and immediately
diluted them with phosphate buffer at 37 ꢀC. The chan-
ges in UV absorbance at 249 nm were followed at dif-
ferent time intervals. The cyclization for compound 1b
upon reduction was found to be too fast to calculate the
kinetic constant. The calculated half-lives for other
compounds 1a, 1c, and 1d upon reduction were 14, 4,
and 2 min, respectively (Fig. 3). These results are con-
Figure 2. Stability of compounds 1a (*), 1b (*), 1c (!), and 1d (!)
in human serum. Relative concentration against time: expressed by the
ratio of the peak area at certain time interval versus peak area at time
zero for each compound.

B. Liu, L. Hu / Bioorg. Med. Chem. 11 (2003) 3889–3899
3893
sistent with the role of two methyl groups in restricting
the rotational freedom of the ground-state conforma-
tion of the molecule and facilitating the intramolecular
cyclization process as outlined in Scheme 1.
As a result, the reduction of compound 1c by E. coli B
nitroreductase was tested again in 10 mM phosphate
buffer, at pH 7.0 and 25 ꢀC (Fig. 5). Under this condi-
tion, the enzyme reduced compound 1c steadily as
shown by HPLC analysis. The formation of FUDR was
confirmed by comparison with authentic sample, and
the cyclized product N-hydroxyl lactam 4c was con-
firmed by analytical LC–MS. The calculated half-life for
the reductive activation is 8 h. A control experiment was
carried out in parallel under the same conditions with-
out the presence of nitroreductase, and no changes were
observed within the same time interval.
Nitroreductase assay
Based on the standard conditions published in the lit-
erature,19 compound 1c was first tested using an incu-
bation mixture containing compound 1c (0.1 mM),
nitroreductase (7 mg/mL), and cofactor NADH (1 mM)
in phosphate buffer (100 mM, pH 7.0) at 37 ꢀC. We
found that the concentration of 1c stopped changing
after 1 h of incubation and only 20% of 1c was con-
sumed. It seemed as though the enzyme was losing
activity or the cofactor NADH was decomposing. It has
been reported in the literature that NADH was unstable
under certain assay conditions.20 When we added more
NADH to the assay mixture after the 1-h incubation,
the substrate was further reduced. This suggested to us
that the problem was indeed due to the instability of
NADH under this assay condition.
This assay condition was thus used to test compounds
1a, 1b, and 1d. We found that compound 1b was not a
substrate of the E. coli B nitroreductase, and no sig-
nificant changes could be observed after 3 days of incu-
bation. Compound 1a and 1d were also poor substrates
of the enzyme. Although the release of FUDR and the
formation of cyclized lactams 4a and 4d could be
detected by HPLC, the half-lives of both activations
were longer than 3 days. Apparently, the two methyl
groups in 1d and the shorter spacing between phenyl
and FUDR in 1a adversely affected the substrate bind-
ing and/or reduction by the nitroreductase enzyme.
Rover and coworkers did a detailed study on factors
that influence the stability of NADH that included pH,
temperature and buffer ions.21 They found that 13% of
NADH decomposed after 40 min of incubation in 100
mM phosphate buffer. Since most of the nitroreductase
assays were carried out in phosphate buffer, we decided
to do our own test on the effect of phosphate buffer
concentration on the stability of NADH. We incubated
a 0.1 mM NADH solution in two different concentra-
tions (100 and 10 mM) of phosphate buffer at pH 7.0
and 37 ꢀC and monitored the absorbance change by
UV–vis at 340 nm, which is the maximum absorbance
wavelength for NADH. The half-life for the decom-
position of NADH was found to be 4.5 h in 100 mM
phosphate buffer, and it increased dramatically to 66 h
in 10 mM phosphate buffer. Decreasing the incubation
temperature from 37 ꢀC to 25 ꢀC resulted in a further
increase in stability of NADH to a half-life of 138.6 h in
10 mM phosphate buffer (Fig. 4).
The slow activation of compounds 1a–d by E. coli B
nitroreductase indicates that these compounds won’t be
appropriate for E. coli B nitroreductase-mediated
enzyme prodrug therapy. However, the fact that chemi-
cal reduction of all four conjugates, especially 1b, to
their corresponding amines leads to fast cyclization and
release of the incorporated drug FUDR suggests the
possibility of employing proteolysis to activate a tri-
partite conjugate such as 17 with a peptide acyl attached
to the amino group. Cleavage of the peptide acyl group
in 17 by a protease would produce the same amine
intermediates as in the reduction of conjugates 1a–d.
These intermediates (3a–d) have been shown in this
paper to undergo a facile intramolecular cyclization
reaction leading to the release of the incorporated drug
FUDR.
Figure 3. Kinetic study on the cyclization processes: Measuring UV
absorbance at 249 nm of the filtrates from 4 min’s hydrogenation of
compounds 1a (*), 1b (*), 1c (!), and 1d (!) against time: expres-
sed in the ratio of absorbance at certain time interval versus absor-
bance at T=120 min when the absorbance for each compound has
reached a constant.
Figure 4. Stability of NADH under different incubating conditions.
Relative UV absorbance of NADH against time: expressed by the
ratio of the absorbance at certain time interval versus absorbance at
time zero for each incubating condition. The conditions used were 100
mM phosphate buffer at 37 ꢀC (*),10 mM phosphate buffer at 37 ꢀC
(*), and 10 mM phosphate buffer at 25 ꢀC (!).

3894
B. Liu, L. Hu / Bioorg. Med. Chem. 11 (2003) 3889–3899
of a tumor or tissue-specific protease and the resulting
tripartite peptide conjugates might be developed into
clinically useful prodrugs for proteolytic activation.
This possibility is currently being explored in our
laboratory.
Experimental
General methods
Solvents were either ACS reagent grade or HPLC grade.
Unless otherwise stated, all reactions were magnetically
stirred and monitored by thin-layer chromatography
(TLC) using 0.25 mm Whatman precoated silica gel
plates. TLC plates were visualized using either 7% (w/
w) ethanolic phosphomolybdic acid or 1% (w/w) aqu-
eous potassium permagnate containing 1% (w/w)
NaHCO3. Flash column chromatography was per-
formed using silica gel (Merck 230–400 mesh). Yields
refer to chromatographically and spectroscopically (1H
NMR) homogeneous materials, unless otherwise noted.
All reagents were purchased at the highest commercial
quality and used without further purification.
Figure 5. Reduction of compound 1c by E. coli B nitroreductase:
Composition of a solution of 1c reductively activated by nitror-
eductase against time. Shown are the disappearance of 1c (*) and
appearance of 4c (*) and FUDR (!).
Infrared spectra were recorded with a Perkin-Elmer
model 1600 series FTIR spectrometer using polystyrene
as an external standard. Infrared absorbance is reported
1
in reciprocal centimeters (cmÀ1). All H and 13C NMR
spectra were recorded on a Varian Gemini 300 or
200 MHz spectrometer at ambient temperature and
calibrated using residual undeuterated solvents as the
internal reference. Chemical shifts (300 MHz for 1H and
In summary, four 50-(2-nitrophenylalkanoyl)-20-deoxy-
5-fluorouridines (1a–d) were designed and synthesized
as potential prodrugs of FUDR for reductive activa-
tion. Two methyl groups were introduced to increase
the stability of the conjugates towards serum esterases.
Chemical reduction of the nitro group to an amino
functionality leads to cyclization and concomitant
release of the active incorporated drug FUDR. Kinetic
analysis of the cyclization activation process indicates
that the two methyl groups a to the ester carbonyl act to
restrict the rotational freedom of the ground state
molecule and promote the cyclization reaction. How-
ever, the two methyl groups seem to adversely affect the
affinity and/or catalysis of the conjugates as substrates
of the E. coli B nitroreductase. Only conjugate 1c was
found to be a substrate of E. coli nitroreductase with a
half life of around 8 h. 1a and 1d were poor substrates
of the nitroreductase enzyme with half lives greater
than 3 days while 1b was not significantly affected by
the enzyme over a three-day incubation. Two products
were found from the enzymatic reduction of 1c, a N-
hydroxyl lactam and the drug FUDR, suggesting that
the enzymatic reduction and subsequent cyclization
activation proceed through the hydroxylamine inter-
mediate. These results do indicate that once the nitro
group is reduced either to the amine or the hydro-
xylamine, the cyclization activation is bound to occur.
Although the poor activity of these conjugates as sub-
strates of E. coli B nitroreductase limited their use as
reductively activated prodrugs in ADEPT or GDEPT
using the reductase, the amine conjugates 3a–d might be
linked to a peptide sequence known to be a good substrate
1
75 MHz for 13C or 200 MHz for H and 50 MHz for
13C) are reported in parts per million (d) relative to
1
CD3OD (d 3.3 for H and 49.0 for 13C). Coupling con-
stants (J values) are given in hertz (Hz). The following
abbreviations were used to explain the multiplicities:
s=singlet; d=doublet; t=triplet; q=quartet; p=quin-
tet; m=multiplet; br=broad. Mass spectral data were
obtained from the University of Kansas Mass Spectro-
metry Laboratory (Lawrence, KS, USA).
50-(2-Nitrophenyl)acetyl-20-deoxy-5-fluorouridine
(1a).
The commercially available 2-nitrophenylacetic acid
(177 mg, 0.97 mmol), FUDR (200 mg, 0.81 mmol) and
PPh3 (250 mg, 0.95 mmol) were dried under vacuum
overnight before 5 mL of anhydrous THF was intro-
duced under argon atmosphere. DEAD (0.15 mL, 0.97
mmol) was then added and the reaction mixture was
stirred at room temperature for 20 h. After removal of
the solvent in vacuo, the residue was purified by flash
column chromatography on silica gel eluted with ace-
tone/hexanes (1:5!1:2!1:1) to afford 277 mg of the
desired product 1a as a white foam solid (83%). 1H
NMR (300 MHz, CDCl3) d 8.16 (d, 1H, J=7.8 Hz),
7.72–7.42 (m, 4H), 6.18 (t, 1H, J=6.0 Hz), 4.40 (dd, 1H,
J=3.3, 12.3 Hz), 4.31 (m, 1H), 4.11 (m, 2H), 3.34 (s,
2H), 2.25 (m, 1H), 1.95 (m, 1H); IR (KBr) 3446.1 (br),
3200.0, 3087.2, 1712.8, 1702.6, 1523.1, 1476.9, 1405.1,
1348.7, 1261.5, 1215.4, 1070.9, 1046.1, 784.6, 717.9
cmÀ1; MS (FAB, m-NBA) m/z (rel intensity): 410.1
(MH+, 7.1), 280.1 (12.9), 164.0 (6.6), 154.0 (100), 136.0

B. Liu, L. Hu / Bioorg. Med. Chem. 11 (2003) 3889–3899
3895
(91.2); HRMS calcd for C17H17FN3O8 (MH+)
410.1000, found 410.1003.
cmÀ1; MS (EI) m/z (rel intensity): 164.3 (M+–COOH,
3.6), 163.0 (100).
2-Nitrophenylacetic acid methyl ester (8). To a solution
of 2-nitrophenylacetic acid (40 g, 0.22 mol) in 240 mL of
methanol was added thionyl chloride (36 mL) over 30
min while maintaining the temperature at 0–4 ꢀC. The
reaction mixture was stirred at 0 ꢀC for 1 h and then at
room temperature for 18 h. After removal of the solvent
in vacuo, the residue was partitioned between ethyl
acetate and water. The organic phase was dried over
anhydrous Na2SO4 and filtered. Removal of ethyl ace-
tate in vacuo gave 43 g of methyl ester 8 as a yellow oil
in quantitative yield. 1H NMR (200 MHz, CDCl3) d
8.09–7.33 (m, 4H), 4.01 (s, 2H), 3.68 (s, 3H); 13C NMR
(50 MHz, CDCl3) d 170.6, 148.9, 133.8, 133.5, 129.9,
128.8, 125.4, 52.4, 39. 7; IR (neat) 2953.8, 1738.5,
1613.7, 1579.8, 1523.1, 1430.8, 1408.3, 1348.7, 1215.4,
1169.2, 1000.0, 861.5, 784.6, 707.7 cmÀ1; MS (FAB, m-
NBA) m/z (rel intensity): 196.2 (MH+, 100), 164.1
(55.5).
50-(2-Methyl-2-(2-nitrophenyl)propionyl)-20-deoxy-5-fluoro-
uridine (1b). 2-Methyl-2-(2-nitrophenyl)propionic acid
10 (20 mg, 0.089 mmol), FUDR (20 mg, 0.081 mmol)
and PPh3 (25 mg, 0.095 mmol) were dried under
vacuum for 16 h and dissolved in 0.5 mL of anhydrous
dioxane, to which was introduced DEAD (20 mL, 0.13
mmol) at room temperature. The reaction mixture was
stirred for 18 h. After removal of the solvent in vacuo,
the residue was purified by flash column chromato-
graphy on silica gel eluted with acetonehexanes
(1:5!1:2!1:1) to afford 22 mg of the desired product
1
1b as a white foam solid (53%). H NMR (300 MHz,
CDCl3) d 9.09 (br s, 1H), 7.96 (d, 1H, J=7.8 Hz), 7.65–
7.37 (m, 4H), 6.18 (t, 1H, J=6.0 Hz), 4.47 (dd, 1H,
J=3.3, 12.3 Hz), 4.24 (m, 1H), 4.15 (m, 2H), 2.30 (m,
1H), 1.72 (s, 3H), 1.67 (s, 3H), 1.59 (m, 1H); IR (KBr)
3456.4 (br), 3200.0, 3087.2, 2984.6, 2943.6, 2830.7,
1712.8, 1523.1, 1471.8, 1394.9, 1353.8, 1261.5, 1153.8,
1092.3, 1046.1, 892.3, 851.3, 789.7, 743.6 cmÀ1; MS
(FAB, m-NBA) m/z (rel intensity): 438.2 (MH+, 5.4),
308.1 (13.5), 192.1 (5.7), 164.1 (10.3), 154.1 (100);
HRMS calcd for C19H21FN3O8 (MH+) 438.1313,
found 438.1321.
2-Methyl-2-(2-nitrophenyl)propionic acid methyl ester
(9). A solution of 2-nitrophenylacetic acid methyl ester
8 (19.5 g, 0.1 mol), methyl iodide (14.3 mL, 0.23 mol)
and 15-crown-5 (5.5 g, 0.025 mol) in 130 mL of DMF
was stirred at À4 to 0 ꢀC, to which was slowly added a
small amount of sodium hydride (60% in oil) until the
mixture suddenly turned blue. The remaining portion of
sodium hydride (total 9.2 g, 0.23 mol) was added with
stirring over 40 min while the temperature was main-
tained at 0–4 ꢀC. Standing overnight after addition, the
reaction mixture was diluted with ethyl acetate. The
organic phase was washed sequentially with 1 N HCl
solution, 1 N KHCO3 solution and brine, dried over
anhydrous Na2SO4 and filtered. Removal of ethyl ace-
tate in vacuo afforded 22 g of the desired product 9 as a
yellow oil in quantitative yield. 1H NMR (200 MHz,
CDCl3) d 7.93–7.38 (m, 4H), 3.66 (s, 3H), 1.68 (s, 6H);
13C NMR (50 MHz, CDCl3) d 175.8, 148.9, 139.4,
133.5, 128.2, 127.9, 125.7, 52.2, 46.5, 27.7; IR (neat)
2986.2, 2951.8, 1738.7, 1607.6, 1575.9, 1526.2, 1481.3,
1434.9, 1356.7, 1288.3, 1247.0, 1230.8, 1147.0, 1113.1,
1049.8, 988.4, 860.9, 785.6, 748.0, 709.5 cmÀ1; MS
(FAB, m-NBA) m/z (rel intensity): 224.2 (MH+, 100),
192.2 (81.9), 164.2 (67.5).
2-(2-Nitrophenyl)ethanol (11). 2-Nitrophenylacetic acid
(4.7 g, 26 mmol) was dissolved in 15 mL of anhydrous
THF, to which was added slowly a solution of 1 N BH3
in THF (52 mL, 52 mmol) under argon. After the addi-
tion, the reaction was allowed to proceed at room tem-
perature for 20 h. Then the mixture was cooled to 0 ꢀC,
quenched by water, partitioned between saturated
NaHCO3 solution and ethyl acetate. The organic phase
was washed with brine and dried over anhydrous
Na2SO4. After removal of the solvent in vacuo, the
residue was purified by flash column chromatography
on silica gel eluted with ethyl acetate/hexanes
(1:10!1:5!1:3) to afford 3.9 g of 11 as a yellow oil
(90%). 1H NMR (200 MHz, CDCl3) d 7.81–7.20 (m,
4H), 3.77 (t, 2H, J=6.6 Hz), 3.02 (t, 2H, J=6.6 Hz),
2.78 (br s, 1H); 13C NMR (50 MHz, CDCl3) d 149.8,
133.9, 133.2, 133.0, 127.7, 124.9, 62.6, 36.2; IR (neat)
3361.8 (br), 2943.6, 2880.6, 1609.8, 1523.4, 1346.7,
1041.9, 860.4, 785.8, 740.5 cmÀ1; MS (FAB, m-NBA)
m/z (rel intensity): 168.0 (MH+, 28.8), 153.9 (100).
2-Methyl-2-(2-nitrophenyl)propionic acid (10). A solu-
tion of 2-methyl-2-(2-nitrophenyl)propionic acid methyl
ester 9 (8 g, 36 mmol) in 108 mL of methanol was trea-
ted with 1 N NaOH solution (108 mL) and was refluxed
for 6 h. After removal of methanol in vacuo, the reac-
tion mixture was adjusted with 1 N HCl solution to pH
2 and extracted with ethyl acetate. The organic phase
was washed with brine and dried over anhydrous
Na2SO4. Removal of ethyl acetate in vacuo gave 5.5 g of
10 as a yellow solid (93%). Mp (ethyl acetate) 146–
3-(2-Nitrophenyl)propionitrile (12). A solution of 2-(2-
nitrophenyl)ethanol 11 (1.7 g, 10 mmol) in 8 mL of
pyridine was treated with p-toluenesulfonyl chloride (1.4
g, 11 mmol) at room temperature. After the reaction
mixture was stirred at room temperature for 3 h, it was
partitioned between 1 N HCl solution and ethyl ether.
The organic phase was washed with 1 N NaHCO3 solu-
tion and brine, dried over anhydrous MgSO4 and fil-
tered. After removal of the solvent in vacuo, the residue
was purified by flash column chromatography on silica
gel eluted with ethyl acetate/hexanes (1:20 !1:15!1:10)
to afford 2.62 g of the desired tosylate as a white solid
(82%). Mp (ethyl acetate/hexanes) 73.5–75 ꢀC; 1H
NMR (300 MHz, CDCl3) d 7.95–7.25 (m, 4H), 4.34 (t,
2H, J=7.2 Hz), 3.25 (t, 2H, J=7.2 Hz), 2.42 (s, 3H);
148 ꢀC; H NMR (200 MHz, CDCl3) d 11.91 (br s, 1H),
1
8.00–7.39 (m, 4H), 1.72 (s, 6H); 13C NMR (50 MHz,
CDCl3) d 182.2, 148.5, 138.9, 133.6, 128.3, 128.1, 125.9,
46.6, 27.3; IR (KBr) 2992.4 (br), 2665.6, 1708.0, 1609.6,
1577.2, 1526.6, 1476.4, 1343.7, 1308.0, 1271.7, 1239.1,
1172.0, 1155.5, 944.5, 867.7, 789.5, 741.9, 703.3, 637.5

3896
B. Liu, L. Hu / Bioorg. Med. Chem. 11 (2003) 3889–3899
13C NMR (75 MHz, CDCl3) d 149.2, 145.1, 133.6,
133.5, 132.8, 131.8, 130.0, 128.4, 127.9, 125.2, 69.6, 33.2,
21.8; IR (KBr) 3066.7, 2964.1, 1594.9, 1523.1, 1353.8,
1717.9, 1702.6, 1523.1, 1471.8, 1400.0, 1348.7, 1261.5,
1189.7, 1169.2, 1082.0, 1046.9, 784.6, 851.3, 784.6, 743.6
cmÀ1; MS (FAB, m-NBA) m/z (rel intensity): 446.1
(M+Na+, 100), 424.1 (MH+, 16); HRMS calcd for
C18H19FN3O8 (MH+) 424.1156, found 424.1176.
1174.4, 1092.3, 969.2, 902.6, 810.3, 774.4, 656.4 cmÀ1
;
MS (FAB, m-NBA) m/z (rel intensity): 322.1 (MH+,
5.0), 150.1 (100).
3-(2-Nitrophenyl)propionic acid methyl ester (14). A
solution of 3-(2-nitrophenyl)propionic acid 13 (140 mg,
0.72 mmol) in 2 mL of methanol was charged with
thionyl chloride (0.12 mL) at 0 ꢀC and the reaction
mixture was stirred at this temperature for 1 h. After
concentration in vacuo, the residue was dissolved in
ethyl acetate, washed with water and dried over anhy-
drous Na2SO4. After removal of the solvent in vacuo,
the product was purified by flash column chromato-
graphy on silica gel eluted with ethyl acetate/hexanes
(1:20!1:15!1:10) to give 138 mg of the methyl ester 14
To a solution of the tosylate (3.2 g, 10.0 mmol) in 30
mL of 1-methyl-2-pyrrolidinone was added potassium
cyanide (0.98 g, 15 mmol) under argon atmosphere. The
mixture was heated while stirring at 90–100 ꢀC in an oil
bath for 1.0 h. Then water was added and the reaction
mixture was extracted with ethyl acetate (3ꢁ50 mL).
The organic phase was washed with brine and dried
over anhydrous Na2SO4. After removal of the solvent in
vacuo, the residue was purified by flash column chro-
matography on silica gel eluted with ethyl acetate/hex-
anes (1:15!1:10!1:5) to give 1.5 g of the desired nitrile
12 as a dark solid (86%). Mp (ethyl acetate/hexanes)
1
as a yellow oil (92%). H NMR (200 MHz, CDCl3) d
7.90–7.27 (m, 4H), 3.61 (s, 3H), 3.16 (t, 2H, J=7.6 Hz),
2.67 (t, 2H, J=7.6 Hz); 13C NMR (50 MHz, CDCl3) d
173.0, 149.1, 135.8, 133.4, 132.4, 127.8, 125.1, 52.0, 34.8,
28.6; IR (neat) 2953.8, 2922.3, 2850.3, 1736.2, 1524.1,
1437.0, 1346.6, 1288.7, 1259.9, 1196.4, 1167.4, 858.3,
788.1, 746.7, 708.7 cmÀ1; MS (ESI) m/z (rel intensity):
232.02 (M+Na+, 38), 210.0 (MH+, 34), 177.95 (80),
149.96 (100).
41–43 ꢀC; H NMR (200 MHz, CDCl3) d 8.03–7.42 (m,
1
4H), 3.21 (t, 2H, J=7.0 Hz), 2.82 (t, 2H, J=7.0 Hz);
13C NMR (50 MHz, CDCl3) d 145.0, 134.1, 133.3,
132.8, 128.9, 125.5, 119.0, 29.6, 18.6; IR (KBr) 3087.2,
2943.6, 2251.3, 1611.7, 1577.8, 1522.1, 1454.2, 1424.4,
1344.1, 1263.2, 1078.1, 864.0, 791.2, 749.4, 700.3 cmÀ1
;
MS (EI) m/z (rel intensity): 176.0 (M+, 4.6), 136.2
(18.3), 103.1 (100).
3-(2-Nitrophenyl)-2,2-dimethylpropionic acid methyl ester
(15). To a solution of 2.5 N n-butyl lithium in THF
(5.8 mL, 14.5 mmol) was added freshly distilled diiso-
propylamine (2.0 mL, 14.5 mmol) dropwise at À50 ꢀC.
After reaction proceeded for 45 min, a solution of ester
14 (736.8 mg, 3.52 mmol) in 25 mL of THF was added
and the reaction mixture quickly turned into a dark
suspension. Then methyl iodide (4.39 mL, 70.4 mmol)
was added slowly into the reaction mixture. The reac-
tion proceeded at À78 ꢀC for 1 h and was gradually
warmed to room temperature. After being stirred for an
additional 2.5 h, the reaction was quenched by saturated
NH4Cl solution at 0 ꢀC and extracted with methyl t-
butyl ether. The organic phase was washed with brine
and dried over anhydrous MgSO4. After concentration
in vacuo, the residue was purified by flash column
chromatography on silica gel eluted with ethyl acetate/
hexanes (1:20!1:15!1:10) to give 73 mg of the methyl
ester 15 as a yellow oil (9%). 1H NMR (300 MHz,
CDCl3) d 7.83–7.25 (m, 4H), 3.64 (s, 3H), 3.32 (s, 3H),
1.17(s, 6H); IR (neat) 2978.6, 2953.8, 1731.4, 1528.8,
1474.1, 1453.8, 1354.5, 1279.7, 1194.5, 1146.8, 1122.4,
854.4, 786.1, 736.1 cmÀ1; MS (FAB, m-NBA) m/z (rel
intensity): 238.1 (MH+, 9.5), 206.1 (6.7), 178.1 (8.8),
154.0 (100); HRMS calcd for C12H16NO4 (MH+)
238.1079, found 238.1079.
3-(2-Nitrophenyl)propionic acid (13). A solution of 3-(2-
nitrophenyl)propionitrile 12 (1.0 g, 5.7 mmol) in aqu-
eous sulfuric acid solution (5 mL, 50%) was heated at
130–135 ꢀC for 4.5 h. The reaction mixture was diluted
with 10 mL of saturated (NH4)2SO4 solution and
extracted with ethyl ether (3ꢁ40 mL). The organic
phase was dried over anhydrous MgSO4 and condensed
in vacuo to afford 1.1 g of the desired acid 13 as a white
solid in quantitative yield. Mp (acetone/hexanes) 111–
113 ꢀC; H NMR (200 MHz, CDCl3) d 8.00–7.37 (m,
1
4H), 3.25 (t, 2H, J=7.4 Hz), 2.82 (t, 2H, J=7.4 Hz);
13C NMR (50 MHz, CDCl3) d 178.6, 135.4, 133.5,
132.3, 127.9, 125.1, 34.7, 28.2; IR (KBr) 2923.1 (br),
1697.4, 1517.9, 1435.9, 1338.5, 1307.7, 1266.7, 1220.5,
928.2, 856.4, 789.7, 723.1 cmÀ1; MS (FAB, m-NBA) m/z
(rel intensity): 196.1 (MH+, 18.2), 178.1 (21.1), 150.1
(9.2), 154.1 (100).
50 -(3-(2-nitrophenyl)propionyl)-20 -deoxy-5-fluorouridine
(1c). 3-(2-Nitrophenyl)propionic acid 13 (143 mg, 0.73
mmol), FUDR (150 mg, 0.61 mmol) and PPh3 (192 mg,
0.73 mmol) were dried under vacuum for 16 h and dis-
solved in 5.5 mL of anhydrous THF, to which was
introduced DEAD (115 mL, 0.74 mmol) at room tem-
perature. The reaction mixture was stirred for 16 h.
After concentration in vacuo, the residue was purified
by flash column chromatography on silica gel eluted
with acetone/hexanes (1:5!1:2!1:1) to afford 158 mg
of the desired conjugate 1c as a white foam solid (61%).
1H NMR (200 MHz, acetone-d6) d 7.98 (dd, 1H, J=8.0,
1.4 Hz), 7.86 (d, 1H, J=7.0 Hz), 7.69–7.47 (m, 3H), 6.29
(dt, 1H, J=1.5, 6.6 Hz), 4.48–4.28 (m, 3H), 4.12 (dd,
1H, J=3.5, 8.3 Hz), 3.24 (t, 2H, J=7.2 Hz), 2.86 (br s,
1H), 2.85 (t, 2H, J=7.5 Hz), 2.36–2.30 (m, 2H); IR
(KBr) 3446.1 (br), 3200.0, 3076.9, 2953.8, 2820.5,
3-(2-Nitrophenyl)-2,2-dimethylpropionic acid (16).
A
solution of 3-(2-nitrophenyl)-2,2-dimethylpropionic
acid methyl ester 15 (73 mg, 0.3 mmol) in 5 mL of
methanol/water (3:1) was charged with lithium hydrox-
ide (20 mg, 0.83 mmol) at 0 ꢀC. The reaction mixture
was stirred at this temperature for 15 h. After removal
of methanol in vacuo, the residue was partitioned
between methyl t-butyl ether and brine. The organic
phase was dried over anhydrous MgSO4 and filtered.

B. Liu, L. Hu / Bioorg. Med. Chem. 11 (2003) 3889–3899
3897
Removal of the solvent in vacuo gave 70 mg of the
1
of 10% Pd/C. Both HPLC and TLC were used to
monitor the progress of reduction. At the end of reac-
tion, the catalyst was removed by filtration, the residue
was washed with methanol (3ꢁ5 mL). The combined
organic phase was concentrated in vacuo, and the resi-
due was purified by flash column chromatography on
silica gel eluted with acetone–hexanes to afford the
cyclized lactam, FUDR, and/or the amine intermediate.
HPLC analysis was performed on a C-18 reversed-phase
column (150ꢁ4.6 mm), using first an isocratic elution of
2% CH3CN for 5 min followed by a gradient elution
from 2% CH3CN to 70% CH3CN over 15 min and a
final isocratic elution of 70% CH3CN for 5 min, at a
flow rate of 1 mL/min and detection wavelengths at 220
and 280 nm.
desired acid 16 as a yellow oil quantitatively. H NMR
(300 MHz, CDCl3) d 9.42 (br s, 1H), 7.86–7.37 (m, 4H),
3.38 (s, 2H), 1.21(s, 6H); IR (neat) 3418.2 (br), 2974.4,
2918.7, 1699.3, 1529.7, 1475.9, 1355.0, 1286.3, 1152.2,
854.1, 786.2, 730.6, 667.9 cmÀ1; MS (FAB, m-NBA) m/z
(rel intensity): 224.1 (MH+, 5.3), 206.1 (8.8), 178.1
(19.7), 154.0 (100).
50-(3-(2-nitrophenyl)-2,2-dimethylpropionyl)-20-deoxy-5-
fluorouridine (1d). 3-(2-Nitrophenyl)-2,2-dimethylpro-
pionic acid 16 (69 mg, 0.31 mmol), FUDR (75.8 mg,
0.308 mmol) and PPh3 (84.8 mg, 0.31 mmol) were dried
under vacuum for 16 h and dissolved in 5 mL of anhy-
drous THF, to which was introduced DEAD (0.051 mL,
0.33 mmol) at room temperature. The reaction mixture
was stirred for 32 h. After concentration in vacuo, the
residue was purified by flash column chromatography
on silica gel eluted with acetone/hexanes (1:5!1:2!1:1)
to afford 52 mg of the desired conjugate 1d as a white
foam solid (53% based on the recovery of 20 mg of the
starting material 16). 1H NMR (300 MHz, CDCl3) d
10.03 (br s, 1H), 7.84 (d, 1H, J=7.2 Hz), 7.60–7.27 (m,
4H), 6.22 (t, 1H, J=6.3 Hz), 4.32 (dd, 1H, J=4.2, 10.3
Hz), 4.22 (d, 1H, J=3.0 Hz), 4.16 (dd, 2H, J=3.3, 5.1
Hz), 3.55 (br s, 1H), 3.32 (q, 2H, J=13.5 Hz), 2.51 (m,
1H), 2.11 (m, 1H), 1.20 (s, 6H); IR (KBr) 3445.7 (br),
3195.4, 3087.7, 2980.3, 2828.5, 1694.0, 1681.8, 1526.3,
1472.6, 1403.8, 1354.9, 1264.6, 1195.8, 1093.7, 1051.3,
970.4, 859.4, 788.5, 738.3 cmÀ1; MS (FAB, m-NBA) m/z
(rel intensity): 452.1 (MH+, 6.2), 322.1 (15.4), 178.0
(6.7), 154.0 (100); HRMS calcd for C20H23FN3O8
(MH+) 452.1469, found 452.1474.
General procedure B: cyclization. A solution of the con-
jugates 1a–d (0.05 mmol) in 3 mL of methanol under-
went hydrogenation according to procedure A. If the
product isolated was the amine intermediate, the iso-
lated intermediate was then incubated in 100 mM
phosphate buffer (pH 7.4) at 37 ꢀC and monitored by
HPLC. At the end of reaction, the solution was extrac-
ted with ethyl acetate, and the organic phase was
washed with brine, dried over anhydrous Na2SO4 and
filtered. Removal of the solvent in vacuo afforded the
lactam. FUDR in the aqueous phase was detected by
HPLC and confirmed by comparison with an authentic
sample.
General procedure C: NaBH4 reduction. To a solution of
NaBH4 (20 mg) in 0.5 mL of water was added the con-
jugates 1a–d (0.05 mmol) and a suspension of 10% Pd/
C in 2 mL of methanol under nitrogen atmosphere.
Both HPLC and TLC were used to monitor the pro-
gress of reaction. At the end of reaction, the reaction
mixture was diluted with acetone, and the catalyst was
removed by filtration. The residue after concentration
was partitioned between CH2Cl2 and water, and both
organic and aqueous phases were analyzed by HPLC
(the same method as mentioned above). The organic
phase was concentrated in vacuo and subjected to flash
column chromatography to afford the cyclized lactam,
FUDR, and/or the amine intermediate. FUDR was
detected in the aqueous phase by HPLC.
Stability determination in phosphate buffer. Each of the
conjugates 1a–d (1.5 mg) was dissolved in 200 mL of
CH3CN and stored at 0 ꢀC prior to use. A 10 mL solu-
tion was withdrawn and quickly added to 190 mL of pre-
warmed 100 mM phosphate buffer (pH 7.4) at 37 ꢀC to
give a final concentration of 1 mM. The resulting solu-
tion was incubated at 37 ꢀC while aliquots (25 mL) of
sample were removed at different time intervals and
injected directly into the HPLC injection port. HPLC
analysis with C-18 reversed-phase column used either a
gradient mobile phase from 28% CH3CN to 52%
CH3CN over 10 min or isocratic elution with a mobile
phase of 40% CH3CN, at a flow rate of 1 mL/min and
detection wavelengths at 220 and 280 nm.
3,3-Dimethyl-1,3-dihydroindol-2-one (5b). Reduction of
2-methyl-2-(2-nitrophenyl)propionic acid-FUDR con-
jugate 1b according to procedure A or C afforded the
lactam 5b as a white solid. Mp 149.5–151 ꢀC; H NMR
1
Stability determination in human serum. A 10-mL stock
solution was quickly added to 190 mL mixture of pre-
warmed human serum and phosphate buffer (4:1) at
37 ꢀC to give a final concentration of 1 mM. The solu-
tion was incubated at 37 ꢀC. At various time intervals,
aliquots (25 mL) were withdrawn and quenched imme-
diately by aqueous HClO4 solution (0.1 mL, 7%) to
stop enzyme hydrolysis. After centrifugation, the super-
natant was analyzed by the same HPLC method men-
tioned above.
(200 MHz, CDCl3) d 9.42 (br s, 1H), 7.28–6.96 (m, 4H),
1.44(s, 6H); 13C NMR (50 MHz, CDCl3) d 184.8, 140.2,
136.5, 127.8, 122.7, 122.6, 110.2, 45.0, 24.5; IR (KBr)
3157.9, 3097.4, 2969.7, 1714.7, 1676.4, 1622.0, 1485.2,
1474.1, 1461.1, 1411.3, 1385.3, 1339.1, 1226.3, 1213.3,
1174.1, 1014.2, 945.3, 812.9, 787.2, 754.8, 739.1, 715.0,
658.5, 619.2 cmÀ1; MS (FAB, m-NBA) m/z (rel inten-
sity): 162.2 (MH+, 100), 161.2 (M+, 65); HRMS calcd
for C10H12NO (MH+) 162.0919, found 162.0898.
General procedure A: hydrogenation. A solution of the
conjugates 1a–d (0.05 mmol) in 3 mL of methanol
underwent atmospheric hydrogenation in the presence
50-(2-Aminophenyl)acetyl-20-deoxy-5-fluorouridine (3a).
Compound 3a was obtained as a white solid from
reduction of 2-nitrophenylacetic acid–FUDR conjugate

3898
B. Liu, L. Hu / Bioorg. Med. Chem. 11 (2003) 3889–3899
1a according to procedure A. Mp 78–80 ꢀC; H NMR
(300 MHz, acetone-d6) d 7.69 (d, 1H, J=6.3 Hz), 7.64–
6.57 (m, 4H), 6.22 (t, 1H, J=6.8 Hz), 4.4 (m, 2H), 4.24
(dd, 1H, J=12.3, 2.7 Hz), 4.11 (dd, 1H, J=7.3, 3.3 Hz),
3.65 (s, 2H), 2.18 (ddd, 1H, J=14, 5.7, 2.4 Hz), 1.94
(dd, 1H, J=14, 7.2 Hz); IR (KBr) 3415.4 (br), 3200.0,
3066.6, 2953.8, 2820.5, 1702.6, 1687.2, 1620.5, 1471.8,
1359.0, 1261.5, 1097.4, 1051.3, 882.0, 753.8 cmÀ1; MS
(FAB, m-NBA) m/z (rel intensity): 380.2 (MH+, 4.2),
154.1 (100); HRMS calcd for C17H19FN3O6 (MH+)
380.1258, found 380.1296.
using 80 mM of each in 100 mM phosphate buffer (pH
7.4). The lmax for FUDR is 268 nm. The lmax for 3a is
271 nm. The lmax for 5a is 248 nm. The maximum
change in UV spectra before and after cyclization were
found to be around 249 nm. Thus, subsequent kinetic
studies on the cyclization of conjugates 1a–d upon reduc-
tion were performed by measuring the absorbance change
at 249 nm. A solution of the conjugates 1a–d (10 mg) and
10% Pd/C (1 mg) in 4 mL of methanol underwent atmo-
spheric hydrogenation for 4 min at room temperature. 1
mL of the reaction mixture was withdrawn into a cuvette
and centrifuged. A 5-mL aliquot of the supernatant solu-
tion was diluted 200 times with pre-warmed phosphate
buffer, pH 7.4 at 37ꢀC. The changes of the UV absor-
bance at 249 nm were monitored. The data was fitted to
first-order decay kinetics (y=y0+aeÀkt) to calculate the
kinetic constant and half-life of the cyclization activation
process for each compound.
1
1,3-Dihydroindol-2-one (5a). Lactam 5a was obtained
from 2-aminophenylacetic acid–FUDR conjugate 3a
according to procedure B as a white solid. Mp 126–
127 C; H NMR (200 MHz, CDCl3) d 9.21 (br s, 1H),
7.17–6.81 (m, 4H), 3.47 (s, 2H); 13C NMR (50 MHz,
CDCl3) d 178.4, 142.8, 128.1, 125.5, 124.8, 122.5, 110.1,
36.5; IR (KBr) 3216.1, 3076.9, 3035.9, 1701.8, 1619.2,
1472.5, 1387.2, 1333.9, 1304.5, 1239.5, 1176.0, 749.8
cmÀ1; MS (EI) m/z (rel intensity): 133.0 (M+, 100),
105.1 (11.1).
ꢀ
1
Determination of NADH stability in the incubation solu-
tion. NADH (0.1 mM) was incubated in 100 mM phos-
phate buffer (pH 7.0) at 37 ꢀC, NADH (0.1 mM) in 10
mM phosphate buffer (pH 7.0) at 37 ꢀC, and NADH
(0.1 mM) in 10 mM phosphate buffer (pH 7.0) at 25 ꢀC
and the stability was monitored by UV–vis at 340 nm.
The data were fitted to first order decay kinetics
(y=y0+aeÀkt) to calculate the kinetic constants and
half-lives of decay.
50-(3-(2-Aminophenyl)propionyl)-20-deoxy-5-fluorouridine
(3c). Compound 3c was obtained from reduction of 3-
(2-nitrophenyl)propionic acid–FUDR conjugate 1c
according to procedure A as a white foam solid. 1H
NMR (300 MHz, acetone-d6) d 7.86 (d, 1H, J=7.0 Hz),
7.02–6.54 (m, 4H), 6.28 (dt, 1H, J=6.6, 1.6 Hz), 4.45–
4.24 (m, 3H), 4.11 (m, 1H), 2.92–2.69 (m, 4H), 2.33–2.25
(m, 2H); IR (neat) 3372.1 (br), 3210.3, 3087.6, 1713.7,
1498.4, 1455.9, 1359.1, 1265.2, 1199.6, 1091.2, 1051.3,
Reduction with E. coli B nitroreductase. The enzyme (1.5
mL of 1.18 mg/mL stock solution, 7 mg/mL) was added to
the solution of the conjugates 1a–d (2.5 mL of 10 mM
stock solution in DMSO, 0.1 mM) and NADH (2.5 mL of
100 mM stock solution in phosphate, 1 mM) in 10 mM
phosphate buffer (243.5 mL, pH 7.0) at 25 ꢀC. Aliquots
were withdrawn and analyzed by HPLC. The half life of
reduction by E. coli nitroreductase was calculated based
on the disappearance of the substrate. Control experi-
ment was carried out under identical conditions without
the enzyme. The products were confirmed by comparison
with authentic samples and analytical LC–MS.
751.8 cmÀ1
.
3,4-Dihydro-1H-quinolin-2-one (5c). Lactam 5c was
obtained from 3-(2-aminophenyl)propionic acid–
FUDR conjugate 3c according to procedure B as a
white solid. Mp 151–153 ꢀC; 1H NMR (200 MHz,
CDCl3) d 8.30 (br s, 1H), 7.15–6.70 (m, 4H), 2.91 (t, 2H,
J=7.5 Hz), 2.57 (t, 2H, J=7.5 Hz); 13C NMR (50 MHz,
CDCl3) d 171.9, 137.4, 128.2, 127.7, 123.9, 123.3, 115.5,
30.9, 25.6; IR (KBr) 3195.9, 3093.6, 2975.6, 2912.5,
1683.5, 1594.4, 1492.2, 1437.9, 1386.5, 1341.4, 1282.0,
1247.1, 1012.3, 1199.2, 1033.1, 816.2, 748.5, 682.0 cmÀ1
MS (FAB, m-NBA) m/z (rel intensity): 148.1 (MH+,
100.0), 120.0 (10.5).
;
Acknowledgements
We gratefully acknowledge the financial support of
grant SNJ-CCR 700-009 from the State of New Jersey
Commission on Cancer Research and grant DAMD 17-
98-1-8544 from the US Army Department of Defense
Prostate Cancer Research Program.
3,3-Dimethyl-3,4-dihydro-1H-quinolin-2-one
(5d).
Reduction of 3-(2-nitrophenyl)-2,2-dimethylpropionic
acid–FUDR conjugate 1d according to procedure A
afforded compound 5d as a white solid. Mp 123–125 ꢀC;
1H NMR (200 MHz, CDCl3) d 7.70 (br s, 1H), 7.11–6.65
(m, 4H), 2.74 (s, 2H), 1.41 (s, 6H); 13C NMR (50 MHz,
CDCl3) d 176.6, 137.0, 128.7, 127.5, 122.5, 123.1, 114.7,
40.5, 37.5, 24.6; IR (KBr) 3198.4, 3071.8, 2965.2,
2925.4, 1671.1, 1596.6, 1492.7, 1389.5, 1258.5, 1094.0,
863.8, 758.4, 670.3 cmÀ1; MS (FAB, m-NBA) m/z (rel
intensity): 176.0 (MH+, 100); HRMS calcd for
C11H14NO (MH+) 176.1075, found 176.1056.
References and Notes
1. Coleman, C. N. J. Natl. Cancer Inst. 1988, 80, 310.
2. Siim, B. G.; Denny, W. A.; Wilson, W. R. Oncol. Res. 1997,
9, 357.
3. Sinhababu, A. K.; Thakker, D. R. Adv. Drug Deliv. Rev.
1996, 19, 241.
4. Rauth, A. M.; Melo, T.; Misra, V. Int. J. Radiat. Oncol.
Biol. Phys. 1998, 42, 755.
5. Denny, W. A. Eur. J. Med. Chem. 2001, 36, 577.
6. Melton, R.; Connors, T.; Knox, R. J. S.T.P. Pharma. Sci.
1999, 9, 13.
Kinetic study of cyclization at 37 ꢀC. First, the UV spec-
tra of FUDR, 2-aminophenylacetic acid–FUDR con-
jugate 3a and 1,3-dihydroindol-2-one 5a were acquired

B. Liu, L. Hu / Bioorg. Med. Chem. 11 (2003) 3889–3899
3899
7. Denny, W. A.; Wilson, W. R.; Hay, M. P. Br. J. Cancer 74,
S32. 1996.
15. Yoon, N. M.; Pak, C. S.; Brown, H. C.; Krishnamurthy,
S.; Stocky, T. J. Org. Chem. 1973, 16, 2180.
8. Hay, M. P.; Sykes, B. M.; Denny, W. A.; Wilson, W. R. A.
Bioorg. Med. Chem. Lett. 1999, 9, 2237.
9. Hay, M. P.; Wilson, W. R.; Sykes, B. M.; Denny, W. A.
Bioorg. Med. Chem. Lett. 1999, 9, 3417.
10. Atwell, G. J.; Sykes, B. M.; O’Connor, C. J.; Denny, W. A.
J. Med. Chem. 1994, 37, 371.
16. Berry, J. M.; Watson, C. Y.; Whish, W. J. D.; Threadgill,
M. D. J. Chem. Soc., Perkin Trans. 1 1997, 1147.
17. Sykes, B. M.; Atwell, G. J.; Denny, W. A.; Mclennan, D. J.;
O’Connor, C. J. J. Chem. Soc., Perkin Trans. 2 1995, 337.
18. Sykes, B. M.; Atwell, G. J.; Denny, W. A.; O’Connor, C. J.
J. Phys. Org. Chem. 1995, 8, 587.
11. Shan, D. X.; Nicolaou, M. G.; Borchardt, R. T.; Wang,
B. H. J. Pharm. Sci. 1997, 86, 765.
12. Mitsunobu, O. Synthesis 1981, 1, 1.
13. Hu, L.; Liu, B.; Hacking, D. R. Bioorg. Med. Chem. Lett.
2000, 10, 797.
19. Bridgewater, J.; Minton, N. P.; Michael, N. P.; Knox, R.;
Springer, C.; Collins, M. Eur. J. Cancer 1995, 31, 2362.
20. Yildirim, O.; Akbulut, U.; Arine, E.; Sungur, S. Biomat-
erials 1994, 15, 587.
21. Rover, L.; Fernandes, J. C. B.; Neto, G. D.; Kubota,
L. T.; Katekawa, E.; Serrano, S. H. P. Anal. Biochem. 1998,
260, 50.
14. Golebiowski, A.; Gorins, G.; Johnson, C. R.; Kiciak, K.
Polish J. Chem. 1993, 67, 685.