Synthesis and biological activity of aromatic amino acid phosphoramidates of 5-fluoro-2'-deoxyuridine and 1-β- arabinofuranosylcytosine: Evidence of phosphoramidase activity

Article abstract of DOI:10.1021/jm9603680

The amino acid phosphoramidate diesters of FUdR (2) and Ara-C (6), 5- fluoro-2'-deoxy-5'-uridyl N-(1-carbomethoxy-2-phenylethyl)phosphoramidate (5a), 5-fiuoro-2'-deoxy-5'-uridyl N-(1-carbomethoxy-2- indolylethyl)phosphoramidate (5b), 1-β-arabinofuranosylc

Full text of DOI:10.1021/jm9603680

J . Med. Chem. 1996, 39, 4569-4575
4569
Syn th esis a n d Biologica l Activity of Ar om a tic Am in o Acid P h osp h or a m id a tes of
5-F lu or o-2′-d eoxyu r id in e a n d 1-â-Ar a bin ofu r a n osylcytosin e: Evid en ce of
P h osp h or a m id a se Activity
Timothy W. Abraham,† Thomas I. Kalman,‡ Edward J . McIntee,† and Carston R. Wagner*,†
Department of Medicinal Chemistry, University of Minnesota, Minneapolis, Minnesota 55455, and Department of Medicinal
Chemistry, State University of New York-Buffalo, Buffalo, New York 14260
Received May 21, 1996X
The amino acid phosphoramidate diesters of FUdR (2) and Ara-C (6), 5-fluoro-2′-deoxy-5′-uridyl
N-(1-carbomethoxy-2-phenylethyl)phosphoramidate (5a ), 5-fluoro-2′-deoxy-5′-uridyl N-(1-car-
bomethoxy-2-indolylethyl)phosphoramidate (5b), 1-â-arabinofuranosylcytosine 5′-N-(1-car-
bomethoxy-2-phenylethyl)phosphoramidate (8a ), and 1-â-arabinofuranosylcytosine 5′-N-(1-
carbomethoxy-2-indolylethyl)phosphoramidate (8b), were synthesized and tested for their
antitumor activity against L1210 mouse lymphocytic leukemia cells and CCRF-CEM human
T-cell lymphoblastic leukemia cells. Ara-C phosphoramidates 8a ,b were found to be inactive
at a concentration of 100 µM, while the FUdR conjugates 5a ,b exhibited IC50 values within a
range of 0.30-0.40 µM. Stability studies revealed that >99% of the phosphoramidates remained
intact after incubation for >2 days in 20% calf or 20% human serum. Intracellular thymidylate
synthase (TS) inhibition studies revealed that treatment of L1210 and CCRF-CEM cells with
5a or 5b resulted in significant inhibition of TS in intact and permeabilized cells, while
treatment of L929 TK- cells with these compounds did not result in inhibition of TS activity
in intact cells. However, permeabilization of L929 TK- cells enhanced the activity of 5a ,b
toward intracellular TS by 900- and 1500-fold, respectively. In addition, incubation of cell-
free extracts of CEM cells with radiolabeled 5b resulted in the rapid production of FUdR 5′-
monophosphate and a lag in the generation of FUdR. Consequently, it is proposed that the
metabolism of the phosphoramidate diesters of FUdR in proliferating tissue proceeds through
two separate enzymatic steps involving P-N bond cleavage by an unknown phosphoramidase
followed by P-O bond cleavage by phosphatases such as 5′-nucleotidase.
In tr od u ction
phosphoramidate diesters, the associated activity was
shown to result from extracellular conversion to the
parent nucleoside.6
Several purine and pyrimidine base and nucleoside
analogs are important weapons in the anticancer and
antiviral chemotherapeutic arsenal. The biological
activity of most of these analogs requires intracellular
metabolism to 5′-mononucleotides by kinase-mediated
phosphorylation. The development of drug resistance
due to decreased nucleotide kinase activity has limited
the effectiveness of these agents. In order to overcome
the problem of drug resistance, prodrug approaches
have been developed to deliver phosphorylated nucleo-
side analogs as neutral derivatives into the cell. These
prodrugs must be converted intracellularly to the cor-
responding nucleotides. Several attempts have been
made to deliver the 5′-monophosphate of 5-fluoro-2′-
deoxyuridine (FUdR) to tumor cells via phosphorami-
date derivatives.1-5 In all but one report, the activity
of FUdR phosphoramidate triesters, rather than phos-
phoramidate diesters, was investigated, presumably
because phosphoramidate diesters would be charged
and therefore less likely to diffuse across the cell
membrane. The lack of observable cytotoxicity associ-
ated with the phosphoramidate diester, 5-fluoro-2′-
deoxy-5′-uridyl N-(1-carboxy-3-methylethyl)phosphor-
amidate, supported this assumption.2 In addition,
although cyclic phosphoramidate derivatives of FUdR
are rapidly converted in cell culture to the corresponding
Recently, our laboratory has reported the chemical
synthesis and biological activity of a series of aromatic
amino acid phosphoramidate di- and triesters of 3′-
azido-3′-deoxythymidine (AZT).7,8 One of these deriva-
tives, AZT 5′-N-(1-carbomethoxy-2-indolylethyl)phos-
phoramidate, 1, was found to be 8-fold more active than
the parent nucleoside and at least 10-fold less toxic.
Surprisingly, when the diester derivatives were incu-
bated in fetal calf serum at pH 7.2, 37 °C, for 6 days,
no degradation to the corresponding 5′-monophosphate
or nucleoside was observable. Preliminary mechanistic
studies demonstrated that 1 is internalized by lympho-
cytes to the same extent as AZT.7 However, in contrast
to peripheral blood mononuclear cells (PBMCs) incu-
bated with AZT, little or no free nucleoside and nearly
4-fold more phosphorylated AZT were observed in
PBMCs incubated with 1.7
* To whom correspondence should be addressed.
† University of Minnesota.
‡ State University of New York-Buffalo.
X Abstract published in Advance ACS Abstracts, October 15, 1996.
S0022-2623(96)00368-8 CCC: $12.00 © 1996 American Chemical Society

4570 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 23
Abraham et al.
Sch em e 1
The behavior of AZT phosphoramidates in human
PBMCs led us to propose that aromatic amino acid
phosphoramidate diesters may be useful for the delivery
of phosphorylated FUdR and 1-â-arabinofuranosylcy-
tosine (Ara-C) to neoplastic tissues. In this report we
describe the sythesis of a series of phenylalanine and
tryptophan phosphoramidate diesters (5a ,b and 8a ,b)
of FUdR (2) and Ara-C (6), respectively, and the study
of their ability to inhibit the growth of murine and
human leukemia cell lines. In addition, mechanistic
studies conducted to determine the extra- and intra-
cellular behavior of these compounds are also described.
Once the FUdR 5′-monophosphate was obtained, the
methyl esters of phenylalanine and tryptophan were
coupled to the phosphorylated nucleoside in a straight-
forward manner by minor modification of the method
of Moffat and Khorana.9 Typically, FUdR 5′-monophos-
phate, the amino acid methyl ester, and DCC were
dissolved in tert-butyl alcohol and water. After refluxing
for 4 h, the 5′-amino acid phosphoramidates of FUdR,
5a ,b, were isolated in yields ranging from 63% to 72%,
respectively.
In contrast to FUdR, the 5′-monophosphate (7) of
Ara-C was prepared directly by treatment with phos-
phorus oxychloride in triethyl phosphate (Scheme
2).10,11,14 The coupling of 7 with the methyl esters of
phenylalanine or tryptophan was carried out as de-
scribed above for FUdR, affording the corresponding
phosphoramidates 8a ,b in yields of 63% and 58%,
respectively.
Ch em istr y
The synthetic protocol developed for constructing
phosphoramidates of FUdR was based on a procedure
by Moffatt and Khorana, in which they describe the
dicyclohexylcarbodiimide (DCC)-mediated coupling of
adenosine 5′-monophosphate to p-anisidine.9 Initial
attempts to synthesize FUdR 5′-monophosphate directly
from FUdR, employing phosphorus oxychloride in tri-
ethyl phosphate, also resulted in 3′-OH phosphoryla-
tion.10,11 Therefore, it was necessary to protect the 3′-
OH in order to effect selective phosphorylation at the
5′-OH (Scheme 1). The 5′-OH of FUdR was first
protected as the monomethoxytrityl derivative (MMTrCl
in pyridine) and the 3′-OH as the acetate (acetic
anhydride in pyridine).12,13 Deprotection of the 5′-O-
MMTr-3′-O-AcFUdR with 80% acetic acid yielded 3′-O-
AcFUdR. Phosphorylation of the 5′-OH was accom-
plished by treating 3′-O-AcFUdR dissolved in pyridine
with 2-cyanoethyl phosphate in the presence of DCC to
yield 3′-O-AcFUdR 5′-(2-cyanoethoxy)phosphate, 3.12,13
Removal of both the acetyl and cyanoethyl moieties was
accomplished by treatment with LiOH in methanol and
water. The resulting FUdR 5′-monophosphate, 4, was
subsequently purified by ion-exchange chromatography
(H+).
Before testing the biological activity of the phosphor-
amidate diesters, the stability of these compounds in
culture media and human serum was determined. The
rates of decomposition of the phosphoramidates 5a ,b
and 8a ,b in serum were determined by incubating each
compound at a concentration of 100 µM in 20% fetal calf
or 20% human serum, pH 7.2, at 37 °C followed by
analysis of the remaining phosphoramidates by reverse-
phase HPLC at 14, 24, 40, and 60 h.
The rate of decomposition for the four phosphorami-
dates was shown to be negligible over 2.5 days, ranging
from 4.0 × 10-10 to 0.2 × 10-10 mol/h in fetal calf serum
and from 2.3 × 10-10 to 0.2 × 10-10 mol/h in human
serum. Typically, >99% of the added phosphoramidate
remained intact after incubation in culture media or
human serum for 2.5 days. Consequently, unlike 5′-
phosphorylated nucleosides, the phosphoramidates are
not rapidly degraded by endogenous blood phosphohy-
drolases or phosphorylases.

Activity of FUdR and Ara-C Phosphoramidates
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 23 4571
Ta ble 1. In Vitro Cytotoxicity of Phosphoramidate Diesters
Ta ble 2. Inhibition of Cellular Thymidylate Synthase Activity
in Intact and Permeabilized Cells
compd
cell line
L1210
CCRF-CEM
CEM-TK-
L929 TK-
L1210
CCRF-CEM
L929 TK-
L1210
CCRF-CEM
CEM-TK-
L929 TK-
L1210
CCRF-CEM
CEM-TK-
L929 TK-
CCRF-CEM
CCRF-CEM
CCRF-CEM
CCRF-CEM
IC50, µM
IC50, µM
2
0.012 ( 0.006
0.002 ( 0.002
66.3 ( 12.0
7.7 ( 1.9
compd
cell line
intact
0.043 ( 0.027 59 ( 35
7.3 ( 3.6 11 ( 4.9
permeabilized
I/P
2
L1210
0.0007
0.66
0.40
L929 TK-
L1210
4
0.015 ( 0.001
0.0055 ( 0.003
6.8 ( 1.8
4
0.017 ( 0.001 0.042
L929 TK-
L1210
>10
21 ( 9
0.008 ( 0.004 >1250
5.0 ( 1
3.1 ( 0.71
1.1 ( 0.74
2.3 ( 0.87
0.89 ( 0.25
0.67 ( 0.55
5a
4.2
5a
0.41 ( 0.1
CCRF-CEM 190 ( 92
61.3
>909
5.2
30.3
>1492
0.32 ( 0.01
86.5 ( 6.0
L929 TK-
L1210
>1000
12.0 ( 8.5
5b
200 ( 14
CCRF-CEM 27 ( 9.9
5b
0.40 ( 0.02
0.32 ( 0.12
58.1 ( 15.0
240 ( 28
L929 TK-
>1000
Table 1, 5a ,b are 29000-43000-fold less cytotoxic
toward CEM-TK- cells relative to CEM cells, respec-
tively, and 17000-21000-fold less cytotoxic toward the
mouse fibroblast cell line L929 TK- when compared to
L1210 cells. Therefore, because thymidine kinase is
necessary for the activity of 5a ,b, these compounds must
ultimately be converted intracellularly to FUdR and not
FUdR 5′-monophosphate. However, the transient in-
tracellular generation of the nucleoside monophosphate,
which is quickly dephosphorylated to the parent nucleo-
side, cannot be ruled out by these experiments.
6
7
8a
8b
0.090 ( 0.035
0.100 ( 0.023
135 ( 6
195 ( 13
Sch em e 2
In contrast to the FUdR derivatives, the phosphor-
amidates of Ara-C, 8a ,b, were shown to be inactive,
exhibiting IC50 values greater than 100 µM, although
Ara-C displayed potent cytotoxicity. The difference in
cell growth inhibitory activity between the FUdR and
Ara-C phosphoramidates may be due either to prefer-
ential transport of aromatic amino acid phosphorami-
dates of uridine relative to cytosine or to the substrate
specificity of the metabolizing enzyme or enzymes.
Meta bolism Stu d ies. To determine the mechanism
of action of the two active FUdR phosphoramidates
5a ,b, we chose to directly measure in situ the inhibition
of thymidylate synthetase (TS) activity in intact cells
with a rapid and convenient tritium release assay (Table
2).16 In addition, the inhibitory effect of the FUdR
analogs on TS activity was assayed in permeabilized
cells from normal and TK- mutant strains in order to
assess the role of cellular uptake and phosphorylation
on their antitumor activity.
Determination of the IC50 values in intact cells
revealed that 5b was 1.8- and 7.0-fold more effective
than 5a at inhibiting the cellular TS activity of L1210
and CCRF-CEM cells, respectively. Neither compound
was capable of inhibiting TS activity in L929 TK- cells,
which is consistent with the inability of these com-
pounds to inhibit the growth of thymidine kinase
deficient cells and implies that they are processed to
FUdR. The IC50 values for 5a ,b were increased by at
least 50- and 80-fold in L929 TK- cells compared to
L1210 cells, respectively. A similar trend was also
observed when the IC50 values for L929 TK- cells are
compared to the IC50 values for CCRF-CEM cells, in
which a 40- and 5-fold increase in the IC50 values was
observed for 5a ,b, respectively.
To assess the significance of transport on the inhibi-
tory activity of 5a ,b on intracellular TS, the effect of
permeabilization with high molecular weight dextran
sulfate on the activity of TS in L1210, CEM, and L929
TK- cells was determined as previously described.17
Permeabilization resulted in an increase of 4.2- and 5.2-
fold in the TS inhibitory activity in L1210 cells of 5a ,b,
Resu lts a n d Discu ssion
Biologica l Activity. The inhibitory activities of 5a ,b
and 8a ,b on the murine leukemia cell line L1210 and
human leukemia cell line CCRF-CEM were determined
and are given in Table 1. In each case, the cells were
treated with the compounds for 48 h and the number
of remaining viable cells was determined with a trypan
blue dye exclusion assay.15 The IC50 values for the
FUdR-containing compounds 5a,b were similar in CCRF-
CEM and L1210 cells, ranging from 0.32 to 0.41 µM,
respectively. Little observable dependence on either the
indolyl or benzyl side chain was detected. Nevertheless,
5a ,b were found to be 160-fold less cytotoxic than FUdR
toward CCRF-CEM cells and 34-fold less cytotoxic
toward L1210 cells.
The ascertain whether the antitumor activity of 5a ,b
was due to their intracellular conversion to either the
corresponding 5′-monophosphate or parent nucleoside,
cytotoxicity studies were conducted with cell lines devoid
of thymidine kinase (TK-) activity. As can be seen in

4572 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 23
Abraham et al.
phoramidate hydrolase activity detected in proliferating
tissue may be of a different nature.20,21 Characteriza-
tion of the enzyme responsible for hydrolytic activity
should shed light on its cellular function. We conclude
that amino acid diester phosphoramidates of FUdR
warrant further study as potential therapeutic agents
and prototypes of nucleotide prodrugs.
Exp er im en ta l Section
Ma ter ia ls. NMR (1H and 31P) spectra were recorded on a
GE Omega-300 spectrometer. An external standard of 85%
H3PO4 was used for all 31P-NMR spectra. FAB mass spectra
were obtained on a VG 7070E-HF mass spectrometer. Ana-
lytical TLC was performed on Analtech silica gel GHLF (0.25
mm) plates. Column chromatography was performed with
grade 62, 60-200 mesh silica gel. Flash chromatography was
performed with grade 60, 230-400 mesh Merck silica gel.
Column chromatography of water soluble compounds on silica
gel was performed using the following solvent gradient:
CHCl3:MeOH:H2O (5:2:0.25), CHCl3:MeOH:H2O (5:3:0.5), and
finally CHCl3:MeOH:H2O (5:4:1). Reverse-phase MPLC was
performed at a pressure of 10-12 psi using an FMI lab pump
equipped with a flow meter and pulse dampener. A C-18
column (3/4 in. × 18 in., 230-400 mesh) was used with a
solvent flow rate of ∼1.5 mL/min. A solvent gradient of H2O:
CH3CN (95:5) to H2O:CH3CN (80:20) was used. Pyridine was
distilled from BaO and stored over 3 Å molecular sieves. All
other solvents were reagent grade and used as received.
2-Cyanoethyl phosphate was prepared from the barium salt
(Aldrich Chemical Co.) by ion-exchange on BioRad AG 50W-
X8 (H+) resin. Concentration under reduced pressure refers
to solvent removal on a Buchi rotary evaporator. High vacuum
refers to <10-2 psi attained with a DuoSeal mechanical pump.
[6-3H]-5-Fluoro-2′-deoxyuridine 5′-monophosphate diammo-
nium salt (16.8 Ci/mmol) was purchased from Moravek Bio-
chemicals, Brea, CA. CCRF-CEM cells are a human T-lym-
phoblastoid leukemia cell line that was obtained from American
Type Culture Collection, Rockville, MD. CEM-TK- cells have
less than 1% of wild-type thymidine kinase activity and were
obtained from the NIH AIDS Research and Reference Reagent
Program, Ogden BioServices Corp., Rockville, MD. Trypan
blue stain (0.4%) in saline (0.85%) was purchased from GIBCO,
Grand Island, NY.
5′-O-(Mon om et h oxyt r it yl)-5-flu or o-2′-d eoxyu r id in e.
FUdR (2; 500 mg, 2.03 mmol) was dissolved in dry pyridine
(10 mL) under nitrogen. Monomethoxytrityl chloride (0.9419
g, 3.05 mmol) was added and the reaction mixture stirred at
room temperature for 48 h. The reaction was quenched with
methanol (5 mL), and then the mixture was concentrated to
dryness under reduced pressure. Flash chromatography on
silica gel, using chloroform/methanol (97/3) as the solvent, gave
the product as a colorless, crystalline solid (1.05 g, 100% yield).
1H-NMR (CDCl3): δ 9.67 (1H, s, H3), 7.82 (1H, d, J H-F ) 6 Hz,
H6), 7.50-6.82 (14H, m, aromatic H’s), 6.30 (1H, t, H1′), 4.56
(1H, br s, H3′), 4.08 (1H, br d, H4′), 3.77 (3H, s, OCH3), 3.40
(2H, m, H5′), 2.78 (1H, br s, 3′-OH), 2.48 (1H, m, H2′), 2.27
(1H, m, H2′).
F igu r e 1. Production of FUdR 5′-monophosphate (9) and
FUdR (0) from cell-free extracts of CEM cells incubated with
radiolabeled 5b. The radiolabeled compound was HPLC
purified, and the DPM values were corrected for background
counts as described in the Experimental Section.
respectively. For CCRF-CEM cells, an increase of 61-
and 30-fold for 5a ,b was observed, respectively. Unex-
pectedly, permeabilization of L929 TK- cells resulted
in a 900- and 1500-fold increase in the TS inhibitory
activity observed for 5a ,b, respectively.
In order to directly observe the generation of FUdR
5′-monophosphate and FUdR by CEM cells, we syth-
3
esized 6-3H-labeled 5b. Upon incubation of 10 mM H-
labeled 5b with CEM cell extracts, a time dependent
generation of FUdR 5′-monophosphate and FUdR was
observed (Figure 1). The production of FUdR lagged
significantly behind the production of FUdR 5′-mono-
phosphate. After 2 h, however, the rate of production
of FUdR 5′-monophosphate (138 DPM/h) and FUdR (132
DPM/h) became nearly identical.
Mech a n istic Im p lica tion s. The results obtained
are consistent with metabolism of 5a ,b to the corre-
sponding 5′-monophosphate before conversion of FUdR.
Although the cytotoxic response of 5a ,b maybe partially
due to direct inhibition of thymidylate synthetase, this
possibility seems unlikely given that the growth of TK-
cells was not significantly inhibited. It is proposed that
intracellular metabolism of the phosphoramidate di-
esters of FUdR proceeds through two separate enzy-
matic steps: one involving P-N bond cleavage by an
unknown phosphoramidase followed by P-O bond
cleavage by phosphatases such as 5′-nucleotidase.
The possible role of a lymphocytic phosphoramidase
in the antiviral activity of amino acid phosphoramidate
triesters and diesters of AZT and d4T has been sug-
gested.7,18,19 Production of d4T 5′-monophosphate, pre-
sumably by an endogenous phosphoramidase, from
CEM and CEM-TK- cells incubated with radiolabeled
d4T 5′-N-phenyl-N-(1-carbomethoxy-2-methylethyl)-
phosphoramidate has been observed.19 Similar studies
have demonstrated that intact PBMCs and cell-free
extracts of PBMCs are capable of converting 3′-fluoro-
2′-deoxythymidine (FLT) 5′-N-(1-carbomethoxy-2-in-
dolylethyl)phosphoramidate to FLT 5′-monophosphate
(McIntee, E. J ., Remmel, R. P., Schinazi, R. F., Abra-
ham, T. W., and Wagner, C. R., manuscript in prepara-
tion). Although the isolation of ribonucleoside 5′-
phosphoramidase from rabbit and rat liver has been
reported, deoxyuridine and deoxythymidine 5′-N-amino
acid phosphoramidates were shown not to be substrates
for the liver-derived enzyme, suggesting that the phos-
3′-O-Acetyl-5-flu or o-2′-d eoxyu r id in e. 5′-O-MMTrFUdR
(800 mg, 1.54 mmol) was dissolved in dry pyridine (20 mL)
under nitrogen. Freshly distilled acetic anhydride (1.45 mL,
15.4 mmol) was added and the reaction mixture heated at 100
°C (boiling water bath) for 4 h. Then ∼50 g of chipped ice
was added and the reaction mixture stirred until all the ice
had melted. The resulting mixture was extracted with chlo-
roform (4 × 20 mL), and the combined extracts were dried over
anhydrous MgSO4. Concentration under reduced pressure
followed by removal of excess pyridine under high vacuum
gave 3′-O-acetyl-5′-O-(monomethoxytrityl)-5-fluoro-2′-deoxy-
uridine as a yellow oil.
The 3′-O-acetyl-5′-O-(monomethoxytrityl)-5-fluoro-2′-deoxy-
uridine was redissolved in 80% acetic acid (20 mL) and heated
at 100 °C for 20 min. The reaction mixture was then
evaporated to dryness under reduced pressure. Repeated
evaporations from ethanol to remove traces of water and acetic

Activity of FUdR and Ara-C Phosphoramidates
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 23 4573
acid gave a light brown solid. Trituration with diethyl ether
gave 3′-O-acetyl-5-fluoro-2′-deoxyuridine as an off-white pow-
dery solid (356 mg, 80% yield). 1H-NMR (CD3OD): δ 8.23 (1H,
d, J H-F ) 6 Hz, H6), 6.25 (1H, t, H1′), 5.28 (1H, br d, H3′), 4.07
(1H, m, H4′), 3.79 (2H, m, H5′), 2.31 (2H, m, H2′), 2.06 (3H, s,
CH3CO).
1-â-Ar a bin ofu r a n osylcytosin e 5′-Mon op h osp h a te (7).
Triethyl phosphate (20 mL) was placed in a dry flask and
cooled to 0 °C in an ice bath under nitrogen. Distilled
phosphorus oxychloride (0.82 mL, 8.83 mmol) was added, and
then Ara-C (1.00 g, 4.11 mmol) was added in one portion. The
reaction mixture was stirred at 0 °C for 22 h and then poured
into a mixture of diethyl ether and petroleum ether (1:1, 200
mL). The suspension was extracted with water (4 × 50 mL),
and the combined aqueous extracts were neutralized to pH 7
using concentrated ammonium hydroxide.
The aqueous phase was then applied onto an ion-exchange
column (BioRad AG1-X8, HCOO-, 40 g, 3 cm × 13 cm,
prepacked in water). The column was eluted with water (400
mL) and then with 0.5 N formic acid (400 mL). The latter
eluant was concentrated to dryness and treated with acetone.
The colorless solid obtained was filtered and washed with
acetone (803 mg, 60% yield). 1H-NMR (D2O): δ 8.15 (1H, d,
H6), 6.25 (2H, two overlapping d, H5, H1′), 4.50 (1H, t, H3′),
4.25-4.05 (4H, m, H4′, H5′, H2′). 31P-NMR (D2O): δ 1.25.
HPLC: tR 1.58 min.
3′-O-Acetyl-5-flu or o-2′-deoxyu r idin e 5′-(2-Cyan oeth oxy)-
p h osp h a te (3). To a mixture of 2-cyanoethyl phosphate (120
mg, 0.795 mmol) in dry pyridine (5 mL) under nitrogen was
added a solution of 3′-O-acetylFUdR (100 mg, 0.347 mmol) and
DCC (572 mg, 2.78 mmol) in dry pyridine (5 mL). The reaction
mixture was stirred at room temperature for 24 h and then
diluted with water (20 mL) and stirred for another 1 h. The
precipitate of 1,3-dicyclohexylurea was filtered off and the
filtrate evaporated to dryness. The residue was subjected to
column chromatography on silica gel with a CHCl3:MeOH:H2O
gradient (see Materials). The product was isolated as a
colorless solid (136 mg, 93% yield). 1H-NMR (D2O): δ 8.05
(1H, d, J H-F ) 6 Hz, H6), 6.30 (1H, t, H1′), 5.40 (1H, br d, H3′),
4.40 (1H, m, H4′), 4.15 (2H, t, OCH2CH2CN), 4.05 (2H, m, H5′),
2.85 (2H, t, CH2CN), 2.55 (1H, dd, H2′), 2.40 (1H, dd, H2′),
2.10 (3H, s, CH3CO). 31P-NMR (D2O): δ 0.63.
5-F lu or o-2′-d eoxyu r id in e 5′-Mon op h osp h a te (4). 3′-O-
Acetyl-5-fluoro-2′-deoxyuridine 5′-(2-cyanoethoxy)phosphate (3;
135 mg, 0.32 mmol) was dissolved in a mixture of methanol
(5 mL) and 0.5 N aqueous LiOH (10 mL) and stirred at room
temperature for 24 h. The reaction mixture was concentrated
to dryness, introduced onto an ion-exchange column (BioRad
AG 50W-X8, H+), and eluted with water. Fractions containing
the product were combined and extracted with diethyl ether
(3 × 15 mL). The aqueous phase was then lyophilized to give
the crude product as a colorless solid which was used in the
next step without further purification. 1H-NMR (D2O): δ 7.95
(1H, d, J H-F ) 6 Hz, H6), 6.20 (1H, t, H1′), 4.45 (1H, br m,
H3′), 4.10 (3H, m, H4′, H5′), 2.35 (1H, m, H2′), 2.25 (1H, m,
H2′). 31P-NMR (D2O): δ 0.78. HPLC: tR 1.48 min.
1-â-Ar a bin ofu r a n osylcytosin e 5′-N-(1-Ca r bom eth oxy-
2-p h en yleth yl)p h osp h or a m id a te (8a ). As described above
for 5a , 1-â-arabinofuranosylcytosine 5′-monophosphate (250
mg, 0.774 mmol) was coupled with phenylalanine methyl ester
(970 mg, 5.41 mmol) in the presence of DCC (797 mg, 3.87
mmol). The colorless solid that was obtained was subjected
to column chromatography on silica gel using a CHCl3:MeOH:
H2O gradient. The product was isolated as a colorless solid
(237 mg, 63% yield). 1H-NMR (D2O): δ 7.72 (1H, d, J 6-5
)
7.5 Hz, H6), 7.30-7.10 (5H, m, Phe), 6.14 (1H, d, J 1′-2′ ) 5
Hz, H1′), 5.93 (1H, d, J 5-6 ) 7.5 Hz, H5), 4.34 (1H, dd, H3′),
4.06 (1H, dd, H4′), 3.87 (2H, br m, H2′, CHCO2Me), 3.76 (2H,
br s, H5′), 3.56 (3H, s, CO2CH3), 2.90 (2H, br d, PheCH2). 31P-
NMR (D2O): δ 7.00. FABMS: [M + 1]+ 484.86, [M + gly]+
577.24. HPLC: tR 8.03 min.22
1-â-Ar a bin ofu r a n osylcytosin e 5′-N-(1-Ca r bom eth oxy-
2-in d olyleth yl)p h osp h or a m id a te (8b). As described above
for 5a , 1-â-arabinofuranosylcytosine 5′-monophosphate (250
mg, 0.774 mmol) was coupled with tryptophan methyl ester
(1.18 g, 5.42 mmol) in the presence of DCC (797 mg, 3.87
mmol). Column chromatography on silica gel followed by
reverse-phase MPLC gave the product as a colorless, crystal-
line solid (233 mg, 58% yield). 1H-NMR (D2O): δ 7.57 (1H, d,
J 6-5 ) 7.5 Hz, H6), 7.47 (1H, d, indole H4), 7.37 (1H, d, indole
H7), 7.11 (2H, overlapping s and t, indole H2, H6), 6.97 (1H,
t, indole H5), 5.98 (1H, d, H1′), 5.72 (1H, d, J 5-6 ) 7.5 Hz,
H5), 4.29 (1H, dd, H3′), 4.00 (1H, dd, H4′), 3.91 (1H, dd,
CHCO2Me), 3.78 (3H, br m, H5′, H2′), 3.54 (1H, s, CO2CH3),
3.07 (2H, br d, indole CH2). 31P-NMR (D2O): δ 7.05.
FABMS: [M + 1]+ 524.00. HPLC: tR 8.82 min.22
[6-3H]-5-F lu or o-2′-d eoxy-5′-u r id yl N-(1-Ca r bom eth oxy-
2-in d olyleth yl)p h osp h or a m id a te (5b). [6-3H]-5-Fluoro-2′-
deoxyuridine 5′-monophosphate diammonium salt (125 µCi,
16.8 Ci/mmol) was diluted with unlabeled FUdR 5′-monophos-
phate diammonium salt (61 mg) to a final specific activity of
0.74 mCi/mmol. Ion-exchange chromatography on BioRad AG
50W-X8 (1.5 cm × 15 cm, 100-200 mesh, H+) and lyophiliza-
tion gave a colorless solid. This was coupled to tryptophan
methyl ester (257 mg, 1.18 mmol) in the presence of DCC (174
mg, 0.85 mmol) as described above for 5b.
Purification was by column chromatography on silica gel
(CHCl3:MeOH:H2O gradient) followed by reverse-phase HPLC
purification on a 10 × 250 mm 10 µm Alltech Econosphere
reverse-phase C8 preparative column. The HPLC system
consisted of a Beckman System Gold 406 analog interface
module, System Gold 166 programmable detector module, and
two 110B solvent delivery modules, a Hewlett Packard 3393A
integrator, and a Rheodyne manual injector. The compound
was eluted by using a gradient of water (solvent A) and
acetonitrile (solvent B). The gradient ran at 4.0 mL/min
starting at 100% A and 0% B with a linear gradient shift to
50% A and 50% B over the first 8 min, where it ran
isocratically for 2 min. From 10 to 12 min there was a linear
gradient change back to 100% A. Column effluent was
monitored at a wavelength of 255 nm, and the product was
5-F lu or o-2′-d eoxy-5′-u r id yl
N-(1-Ca r b om et h oxy-2-
p h en yleth yl)p h osp h or a m id a te (5a ). To the flask contain-
ing 5-fluoro-2′-deoxy-5′-uridine 5′-monophosphate (4) were
added phenylalanine methyl ester (400 mg, 2.24 mmol), DCC
t
(329 mg, 1.60 mmol), BuOH (5 mL), and water (1 mL).
A
reflux condenser was attached to the flask and the reaction
mixture heated in a boiling water bath for 4 h. After cooling
to room temperature the solvents were removed under reduced
pressure. The residue was resuspended in water (15 mL) and
extracted with diethyl ether (4 × 15 mL). The aqueous phase
was then lyophilized. The colorless solid that was obtained
was subjected to column chromatography on silica gel using a
CHCl3:MeOH:H2O gradient. The product was isolated as a
colorless solid (99 mg, 63% for two steps). 1H-NMR (D2O): δ
7.92 (1H, d, J H-F ) 6 Hz, H6), 7.25 (5H, m, Phe), 6.20 (1H, t,
H1′), 4.38 (1H, m, H3′), 4.01 (1H, m, H4′), 3.90 (1H, dd, CHCO2-
Me), 3.71 (2H, m, H5′), 3.63 (3H, s, CO2CH3), 2.95 (1H, dd,
PheCH2), 2.87 (1H, dd, PheCH2), 2.30 (1H, m, H2′), 2.13 (1H,
m, H2′). 31P-NMR (D2O): δ 7.08. FABMS: [M + 1]+ 488.05.
HPLC: tR 5.59 min.22
5-F lu or o-2′-d eoxy-5′-u r id yl N-(1-Ca r bom eth oxy-2-in -
d olyleth yl)p h osp h or a m id a te (5b). 3′-O-Acetyl-5-fluoro-2′-
deoxyuridine 5′-(2-cyanoethoxy)phosphate (3; 121 mg, 0.29
mmol) was converted to 5-fluoro-2′-deoxyuridine 5′-monophos-
phate (4) as previously described and then coupled with
tryptophan methyl ester (440 mg, 2.03 mmol) in the presence
t
of DCC (299 mg, 1.45 mmol), BuOH (5 mL), and H2O (1 mL).
Column chromatography was carried out on silica gel using a
CHCl3:MeOH:H2O gradient, and the product was further
purified by reverse-phase MPLC. The pure product was
isolated as a colorless, crystalline solid (109 mg, 72% for two
steps). 1H-NMR (D2O): δ 7.56 (1H, d, J H-F ) 6 Hz, H6), 7.51
(1H, d, indole H4), 7.32 (1H, d, indole H7), 7.12 (1H, s, indole
H2), 7.09 (1H, t, indole H6), 7.00 (1H, t, indole H5), 6.00 (1H,
t, H1′), 4.22 (1H, m, H3′), 3.96 (2H, m, H4′, CHCO2Me), 3.69
(2H, m, H5′), 3.65 (3H, s, CO2CH3), 3.12 (1H, dd, indoleCH2),
2.97 (1H, dd, indoleCH2), 2.09 (1H, m, H2′), 1.76 (1H, m, H2′).
31P-NMR (D2O): δ 7.20. FABMS: [M + 1]+ 527.0. HPLC: tR
6.07.22

4574 J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 23
Abraham et al.
collected from 5.7 to 7.7 min, yielding a colorless solid (20.3
mg, 23% yield, 0.76 mCi/mmol).
bovine serum or 20% human serum in PBS (pH 7.4) at 37 °C
for 14, 24, 40, and 60 h in triplicate. Cold HPLC grade
methanol (288.8 µL) was added to precipitate proteins, and
3.75 µL of 10 mM d4T in methanol was added as an internal
standard. Samples were then centrifuged (13200g, 10 min, 4
°C), and a 20 µL aliquot from the methanolic solution was
subjected to HPLC analysis. Analysis of the remaining
phosphoramidate was performed on a 4.6 × 250 mm 5 µm
Spherisorb reverse-phase C8 column. The HPLC system was
as described above. A standard curve (2-200 µM) based on
area under the curve (AUC) was constructed for the com-
pounds with d4T (50 µM) used as the internal standard. The
compounds were eluted by using a gradient of 50 mM am-
monium acetate (solvent A) and acetonitrile (solvent B) and
monitored at 260 nm. For the FUdR compounds, the gradient
ran at 1.5 mL/min and changed linearly from 85% A to 75% A
over the first 10 min. From 10 to 15 min there was a linear
gradient change to 60% A; then from 15 to 20 min the gradient
was changed back to 85% A. For the Ara-C compounds, the
gradient ran at 1.5 mL/min and changed linearly from 90% A
to 80% A over the first 10 min. From 10 to 15 min there was
a linear gradient change to 65% A; then from 15 to 20 min
the gradient was changed back to 90% A. The concentration
of the remaining phosphoramidate was determined from the
standard curve. Decomposition rates were determined by
linear plots of the remaining concentration of the phosphor-
amidates relative to time.
Meta bolism of [6-3H]-5-F lu or o-2′-d eoxy-5′-u r id yl N-(1-
Ca r bom eth oxy-2-in d olyleth yl)p h osp h or a m id a te (5b) in
CEM Cell Lysa tes. CEM cells were separated from their
culture medium by centrifugation (430g, 5 min, room temper-
ature). The residue (about 100 µL, 1 × 108 cells) was
resuspended in 1 mL of buffer (20 mM Tris-HCl, 500 mM
NaCl, pH 7.5) and sonicated with a VirSonic 300 cell disrupter
with microtip adapter (4 × 4 s bursts on ice). Lysate from ∼8
× 106 cells was incubated with [6-3H]-5-fluoro-2′-deoxy-5′-
uridyl N-(1-carbomethoxy-2-indolylethyl)phosphoramidate (10
mM, 0.76 mCi/mmol) for indicated lengths of time at 37 °C.
Enzymatic reactions were halted by quickly freezing the
samples in dry ice and storing at -20 °C until assayed.
Methanol was added to precipitate proteins and ensure
termination of enzymatic reactions. Samples were then
centrifuged (13200g, 15 min, 4 °C), and a 20 µL aliquot from
the methanolic lysate was subjected to HPLC analysis on a
4.6 × 250 mm 5 µm Spherisorb reverse-phase C8 column that
was protected by a 4.6 × 7.5 mm All-Guard Econosphere RP-
C8 precolumn. The HPLC system was identical with the one
described above. Metabolites were eluted by using a gradient
of 50 mM ammonium acetate (solvent A) and acetonitrile
(solvent B). The gradient ran at 1.5 mL/min using 99% A and
1% B isocratically for 4 min. From 4 to 9 min there was a
linear gradient change from 99% A to 70% A. From 9 to 12
min there was a linear gradient change back to 99% A; 0.5
mL fractions were collected for the first 16 min, and each was
mixed with scintillation cocktail (5 mL) and then counted on
a Beckman LS-3801 liquid scintillation counter. Metabolite
identification was based on retention times of coinjected cold
synthesized standards of FUdR 5′-monophosphate, FUdR, and
5b. The corresponding retention times for FUdR 5′-mono-
phosphate, FUdR, and 5b when monitored at 260 nm were
1.8, 5.4, and 10.5 min, respectively. Values are reported in
DPMs of radioactivity corresponding to the respective retention
times and the background DPMs from a 10 mM control sample
subtracted out.
Ack n ow led gm en t. We thank Dr. Rory Remmel for
many helpful discussions and suggestions and Ms. M.
C. Hsiao for assistance with the cellular enzyme assays.
This study was partially supported by Grant CA61908
(C.R.W.), AI27251 (T.I.K.), and CA35212 (T.I.K.) from
the National Institutes of Health. E.J .M. was supported
by the National Institute of Health Pharmacological
Sciences Training Grant GM07994.
Refer en ces
Cytotoxicity Assa ys. The purity of the compounds was
evaluated by HPLC prior to biological testing. In brief,
separation and quantitation of nucleoside and nucleotide
impurities present in the compounds to be tested were
performed on a 4.6 × 250 mm 5 µm Spherisorb reverse-phase
C8 column. The HPLC system consisted of a Spectra-Physics
SP8800 ternary HPLC pump and SP4600 integrator, a Kratos
Spectraflow 757 absorbance detector, and a Rheodyne manual
injector. The compounds were eluted by using a gradient of
50 mM ammonium acetate (solvent A) and acetonitrile (solvent
B) and monitored at 255 nm. For the FUdR compounds 5a ,b,
the gradient ran at 1.5 mL/min and changed linearly from 85%
A to 72% A over the first 15 min. From 15 to 20 min there
was a linear gradient change back to 85% A. The retention
time for FUdR was 2.30 min. For the Ara-C compounds 8a ,b,
the gradient ran at 1.5 mL/min and changed linearly from 90%
A to 75% A over the first 15 min. From 15 to 20 min there
was a linear gradient change back to 90% A. The retention
time for Ara-C was 2.29 min. Relative amounts of impurities
were determined by comparing the peak area of the nucleoside
and nucleotide to the area of the compound being evaluated.
Purity is expressed as a percent of total area.
(1) Remy, D. C.; Sunthankar, A. V.; Heidelberger, C. Studies on
Fluorinated Pyrimidines. XIV. The Synthesis of Derivatives
of 5-Fluoro-2′-deoxyuridine 5′-Phosphate and Related Com-
pounds. J . Org. Chem. 1962, 27, 2491-2500.
(2) Mukherjee, K. L.; Heidelberger, C. Studies on Fluorinated
Pyrimidines. XV. Inhibition of the Incorporation of Formate-
C14 into DNA Thymine of Ehrlich Ascites Carcinoma Cells by
5-Fluoro-2′-deoxyuridine-5′-monophosphate and Related Com-
pounds. Cancer Res. 1962, 22, 815-822.
(3) Hunston, R. N.; J ehangir, M.; J ones, A. S.; Walker, R. T.
Synthesis of 2′-deoxy-5-fluoro-5′-O-1′′,3′′,2′′-oxazaphosphacyclo-
hexa-2′′-yluridine 2′′-oxide and related compounds. Tetrahedron
1980, 36, 2337-2340.
(4) Farquhar, D.; Kuttesch, N. J .; Wilkerson, M. G.; Winkler, T.
Synthesis and Biological Evaluation of Neutral Derivatives of
5-Fluoro-2′-Deoxyuridine 5′-Phosphate. J . Med. Chem. 1983, 26,
1153-1158.
(5) Fries, K. M.; J oswig, C.; Borch, R. F. Synthesis and Biological
Evaluation of 5-Fluoro-2′-Deoxyuridine Phosphoramidate Ana-
logs. J . Med. Chem. 1995, 38, 2672-2680.
(6) Kumar, A.; Coe, P. L.; J ones, A. S.; Walker, R. T.; Balzarini, J .;
De Clercq, E. Synthesis and biological evaluation of some cyclic
phosphoramidate nucleoside derivatives. J . Med. Chem. 1990,
33, 2368-2375.
(7) Wagner, C. R.; McIntee, E. J .; Schinazi, R. F.; Abraham, T. W.
Aromatic amino acid phosphoramidate di- and triesters of 3′-
azido-3′-deoxythymidine (AZT) are non-toxic inhibitors of HIV-1
replication. Bioorg. Med. Chem. Lett. 1995, 5, 1819-1824.
The compounds were evaluated for their growth inhibitory
activity with L1210, L929 TK-, CEM, and CEM-TK- cells in
culture. The CEM and CEM-TK- cells were maintained in
RPMI 1640 medium supplemented with 20% heat-inactivated
fetal bovine serum, penicillin (100 U/mL), streptomycin (100
µg/mL), and 0.5% human interleukin-2. The cells were
cultured with and without compound at 37 °C in a 5% CO2-
95% air environment for 48 h, at which time portions were
counted for cell proliferation and viability by the trypan blue
dye exclusion method.15
(8) Abraham, T. W.; Wagner, C. R.
A phosphoramidite-based
synthesis of phosphoramidate amino acid diesters of antiviral
nucleosides. Nucleosides Nucleotides 1994, 13, 1891-1903.
(9) Moffatt, J . G.; Khorana, H. G. Nucleoside Polyphosphates. X.
The Synthesis and some Reactions of Nucleoside-5′-phosphoro-
morpholidates and Related Compounds. Improved Methods for
the Preparation of Nucleoside-5′-Polyphosphates. J . Am. Chem.
Soc. 1961, 83, 649-658.
(10) Yoshikawa, M.; Kato, T.; Takenishi, T. Studies of phosphoryla-
tion. III. Selective phosphorylation of unprotected nucleosides.
Bull. Chem. Soc. J pn. 1969, 42, 3505-3508.
Deter m in a tion of Decom p osition Ra tes in Ser u m . The
stability of 5a ,b and 8a ,b was assessed in fetal bovine serum
and human serum. In brief, 100 µM of the compound was
incubated in a solution of either 20% heat-inactivated fetal
(11) Yoshikawa, M.; Kato, T.; Takenishi, T. A novel method for
phosphorylation of nucleosides to 5′-nucleotides. Tetrahedron
Lett. 1967, 50, 5065-5068.

Activity of FUdR and Ara-C Phosphoramidates
J ournal of Medicinal Chemistry, 1996, Vol. 39, No. 23 4575
(12) Tener, G. M. 2-Cyanoethyl phosphate and its use in the synthesis
of phosphate esters. J . Am. Chem. Soc. 1961, 83, 159-168.
(13) Thomas, H.; Montgomery, J . A. Complex Esters of Thioinosinic-
(5′) Acid. J . Med. Chem. 1962, 5, 24-32.
Novel Phosphoramidate Derivatives of 3′-Azido-3′-Deoxythymi-
dine (AZT) as Anti-HIV Compounds. Antiviral Chem. Chemo-
ther. 1990, 1, 107-113.
(19) McGuigan, C.; Cahard, D.; Sheeka, H. M.; De Clercq, E.;
Balzarini, J . Aryl Phosphoramidate Derivatives of d4T Have
Improved Anti-HIV Efficacy in Tissue Culture and May Act by
the Generation of a Novel Intracellular Metabolite. J . Med.
Chem. 1996, 39, 1748-1753.
(20) Shabarova, Z. A. Synthetic Nucleotide-peptides. Prog. Nucleic
Acid Res. Mol. Biol. 1970, 10, 145-182.
(21) Kuba, M.; Okizaki, T.; Ohmori, H.; Kumon, A. Nucleoside
monophosphoramidate hydrolase from rat liver: purification and
characterization. Int. J . Biochem. 1994, 26, 235-245.
(22) Although the purity of 5a ,b and 8a ,b was verified by NMR,
HPLC, and MS analyses, their hydroscopic nature prevented the
obtaining of satisfactory (<0.5% of theoretical) elemental analy-
ses. Compounds were therefore judged pure by observation of
a single peak using two different RP-HPLC methods.
(14) Hong, C. I.; Nechaev, A.; West, C. R. Nucleoside Conjugates as
Potential Antitumor Agents. 2. Synthesis and Biological Activ-
ity of 1-(â-D-Arabino-furanosyl)cytosine Conjugates of Predniso-
lone and Prednisone. J . Med. Chem. 1979, 22, 1428-1432.
(15) Davidsohn, I.; Nelson, D. A. The Blood. In Clinical Diagnosis
by Laboratory Methods, 14th ed.; Davidsohn, I., Henry, J . B.,
Eds.; W. B. Saunders Co.: Philadelphia, PA, 1969; pp 140-157.
(16) Yalowich, J . C.; Kalman, T. I. Rapid Determination of Thymidy-
late Synthetase Activity and its Inhibition in Intact L1210
Leukemia Cells in Vitro. Biochem. Pharmacol. 1985, 34, 2319-
2324.
(17) Kalman, T. I.; Marinelli, E. R.; Xu, B.; Reddy, A. R.; J ohnson,
F.; Grollman, A. P. Inhibition of Cellular Thymidylate Synthesis
by Cytotoxic Propenal Derivatives of Pyrimidine Bases and
Deoxynucleosides. Biochem. Pharmacol. 1991, 42, 431-437.
(18) McGuigan, C.; Devine, K. G.; O’Connor, T. J .; Galpin, S. A.;
J effries, D. J .; Kinchington, D. Synthesis and Evaluation of some
J M9603680