5-(trifluoromethyl)-β-L-2′-deoxyuridine, the L-enantiomer of trifluorothymidine: Stereospecific synthesis and antiherpetic evaluations

Article abstract of DOI:10.1016/S0968-0896(01)00062-1

As part of our ongoing work on β-L-nucleoside analogues as potential antiviral drugs, we have synthesized 5-(tri-fluoromethyl)-β-L-2′-deoxyuridine (L-TFT), the hitherto unknown L-enantiomer of trifluorothymidine (CF₃dUrd, TFT). We have also studied the effect of L-TFT on human and herpes simplex virus (HSV) type 1 and 2 thymidine kinases, and human thymidine phosphorylase, as well as its anti-HSV-1 and anti-HSV-2 activities in cell cultures. L-TFT has been found: (i) to inhibit HSV-1 TK with activity comparable to TFT, with no effect on human TK, (ii) to be phosphorylated by the viral enzyme with similar efficiency to TFT, (iii) to be resistant, in contrast to TFT, to hydrolysis by human thymidine phosphorylase. Unfortunately, when evaluated in cell cultures, L-TFT did not show any anti-HSV-1 and anti-HSV-2 activities. Copyright

Full text of DOI:10.1016/S0968-0896(01)00062-1

Bioorganic & Medicinal Chemistry 9 (2001) 1731–1738
5-(Trifluoromethyl)-ꢀ-L-20-deoxyuridine, the L-Enantiomer
of Trifluorothymidine: Stereospecific Synthesis and
Antiherpetic Evaluations
Raul Salvetti,a Arnaud Marchand,d Massimo Pregnolato,a Annalisa Verri,b
Silvio Spadari,b Federico Focher,b Martin Briant,c Jean-Pierre Sommadossi,c
d,e
Christophe Mathed,*and Gilles Gosselin
a
`
Dipartimento di Chimica Farmaceutica, Universita degli Studi, I-27100 Pavia, Italy
bIstituto di Genetica Biochimica ed Evoluzionistica, CNR, I-27100 Pavia, Italy
cNovirio-Pharmaceuticals, Inc., Cambridge, MA, USA
d
e
´
`
Laboratoire de Chimie Organique Biomoleculaire de Synthese, U.M.R. 5625 CNRS-UM II,
´
´ ´
Laboratoire Cooperatif Novirio-CNRS-Universite Montpellier II, 34095 Montpellier, France
Universite Montpellier II, 34095 Montpellier, France
Received 9 January 2001; revised 27 February 2001; accepted 2 March 2001
Abstract—As a part of our ongoing work on b-l-nucleoside analogues as potential antiviral drugs, we have synthesized 5-(tri-
fluoromethyl)-b-l-20-deoxyuridine (l-TFT), the hitherto unknown l-enantiomer of trifluorothymidine (CF3dUrd, TFT). We have
also studied the effect of l-TFT on human and herpes simplex virus (HSV) type 1 and 2 thymidine kinases, and human thymidine
phosphorylase, as well as its anti-HSV-1 and anti-HSV-2 activities in cell cultures. l-TFT has been found: (i) to inhibit HSV-1 TK
with activity comparable to TFT, with no effect on human TK, (ii) to be phosphorylated by the viral enzyme with similar efficiency
to TFT, (iii) to be resistant, in contrast to TFT, to hydrolysis by human thymidine phosphorylase. Unfortunately, when evaluated
in cell cultures, l-TFT did not show any anti-HSV-1 and anti-HSV-2 activities. # 2001 Elsevier Science Ltd. All rights reserved.
Introduction
rated into viral DNA.3 TFT, as well as its phosph-
orylated metabolites, can also inhibit viral DNA
biosynthesis by acting as competitive inhibitors of virus-
encoded thymidine kinase4 and HSV DNA polymerase2
with respect to thymidine and thymidine triphosphate.
TFT shows also a concomitant cytotoxicity in unin-
fected cells probably related to the inhibition of cellular
enzymes including human thymidine kinase, thymidyl-
ate synthase as well as DNA polymerases.5,6
Infections by herpes viruses are among the most com-
mon and easily transmitted human viral diseases.
Numerous distinct herpes viruses have been identified
and the treatment by therapeutic agents of diseases
caused by some of them, such as herpes simplex virus
(HSV) has produced clinical benefit. Among these
drugs, 5-(trifluoromethyl)-b-d-20-deoxyuridine [CF3dUrd,
trifluorothymidine (TFT)]1 has been approved by the
Food and Drug Administration since 1980 for use in the
topical treatment of primary keratoconjonctivitis and
epithelial keratitis.2 The mode of action of TFT is based
on its intracellular conversion to TFT monophosphate
by both cellular and virus-encoded thymidine kinase.
Further phosphorylation steps by nucleotide kinases
gives TFT triphosphate which is preferentially incorpo-
In the last years, l-nucleoside analogues, the mirror
images of the natural d-nucleosides, have drawn con-
siderable attention as potential antiviral drugs.7 The
possibility that l-nucleoside derivatives may be more
efficient than the corresponding d-enantiomers, owing
to their powerful antiviral activity and favorable toxi-
city profile, has been demonstrated for some of them.8
On the basis of these findings, we thought it was of
interest to synthesize the hitherto unknown l-enantio-
mer (5-CF3-b-l-dUrd 6, l-TFT) of TFT, in order to
study its effect on human and herpes simplex virus type
*Corresponding author. Tel.:+33-4-6714-4776; fax: +33-4-6704-
2029; e-mail: cmathe@univ-montp2.fr
0968-0896/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0968-0896(01)00062-1

1732
R. Salvetti et al. / Bioorg. Med. Chem. 9 (2001) 1731–1738
1 and 2 thymidine kinases, human thymidine phos-
phorylase (TK, HSV-1 TK, HSV-2 TK and TP, respec-
tively) as well as its anti-HSV-1 and HSV-2 activities in
cell cultures, comparatively to TFT.
affording the 20-deoxy-b-l-threo derivative 3 as a white
foam in 92% overall yield. Deprotection of 3 with solid
sodium methoxide in dry methanol provided the
unprotected 20-deoxynucleoside 4 as a crystalline solid
in 67% after purification. In order to prepare the target
compound 6, the 20-deoxy-b-l-threo-pentofurano-
nucleoside 4 was selectively converted into its 50-O-ben-
zoyl derivative 5. Inversion of configuration at C-30 was
achieved via Mitsunobu conditions14 by using benzoic
acid as incoming nucleophile. The reaction of 5 with
diethyl azodicarboxylate (DEAD), benzoic acid and tri-
phenylphosphine (PPh3) at 0 ꢀC, followed by debenzoy-
lation with methanolic ammonia provided, after
purification on silica gel column, the desired enantiomer
l-TFT 6 as a crystalline white solid in 23% overall
yield.
Results and Discussion
Chemistry
Several methodologies have been described in the lit-
erature for the preparation of 20-deoxynucleoside ana-
logues.9 For our purpose, we decided to synthesize in a
stereospecific manner the target molecule 6 by conden-
sing a suitable protected l-pentofuranose sugar with the
commercially available 5-(trifluoromethyl)uracil. In
accord with Baker’s rule,10,11 and owing to 2-O-acyl
participation during condensation, we selected as start-
ing sugar 1,2-di-O-acetyl-3,5-di-O-benzoyl-l-xylo-fur-
anose 1 (Scheme 1). This compound was readily
prepared from l-xylose following a synthetic pathway
previously described by Gosselin et al.12 The glycosyla-
tion reaction of silylated 5-(trifluoromethyl)uracil with 1
was carried out under Vorbruggen conditions, in anhy-
drous 1,2-dichloroethane using (trimethylsilyl) tri-
fluoromethane sulfonate (TMSOTf) as a catalyst.13
From the crude coupling product, regioselective 20-O-
deacetylation with hydrazine hydrate gave 1-(3,5-di-O-
benzoyl-b-l-xylo-furanosyl)-5-(trifluoromethyl)uracil 2
in 82% overall yield after silica gel column chromato-
graphy. Compound 2 was then reacted with O-phenyl
chloro(thio)formate (PhOC(S)Cl) and 4-dimethylamino-
pyridine (DMAP) in anhydrous 1,2-dichloroethane to
give the corresponding 20-O-[phenoxy(thiocarbonyl)]
intermediate, which was subsequently deoxygenated
with tris(trimethylsilyl)silane in dry dioxane in the
presence of a,a0-azoisobutyronitrile (AIBN), thus
Biological studies
L-TFT specifically inhibits herpes simplex virus thymi-
dine kinases. l-TFT and its d-counterpart (TFT) have
been tested against both herpes simplex virus (HSV-1
and HSV-2) and human thymidine kinases. Interest-
ingly, this study demonstrated that the l-enantiomer
inhibits HSV-1 TK with an IC50 value of 0.5 mM with
no effect against human TK at concentrations up to
200 mM, the highest concentration tested (Fig. 1A).
Such a result is particularly interesting if we compare it
with the values obtained for TFT, the clinically used
Viroptic1. In fact, TFT inhibits not only herpes virus
TK (HSV-1 TK, IC50=0.25 mM; HSV-2 TK,
IC50=2 mM) but also human TK (IC50=2 mM) (Fig.
1B). Table 1, where a comprehensive view of the effect
of both enantiomer on HSV-1 and HSV-2 TK’s is pre-
sented, clearly shows that TFT and l-TFT pre-
ferentially inhibit HSV-1 TK. l-TFT was further
studied in order to better understand its mechanism of
Scheme 1. Reagents and conditions: (a) (i) silylated 5-trifluoromethyluracil, TMSOTf, (CH2Cl)2, rt; (ii) H2NNH2. H2O, acetic acid, pyridine, rt;
(b) (i) PhOC(S)Cl, DMAP, (CH2Cl)2, rt; (ii) (Me3Si)3SiH, AIBN, dioxane, reflux; (c) NaOMe, MeOH, rt; (d) BzCl, pyridine, 0 ꢀC; (e) (i) DEAD,
benzoic acid, PPh3, THF, 0 ꢀC; (ii) MeOH–NH3, rt.

R. Salvetti et al. / Bioorg. Med. Chem. 9 (2001) 1731–1738
1733
action. Experiments in which the inhibitor was tested at
different substrate (thymidine) concentrations demon-
strated that l-TFT inhibits HSV-1 TK with a competi-
tive or partially mixed-type mechanism of action with
a Ki value of 0.5 mM (Fig. 2). The kinetic behavior
suggests a specific interaction of l-TFT with the active
site of the viral enzyme.
inhibitor, we incubated HSV-1 TK with l-TFT (or its
d-enantiomer, as control) in the presence of [g-32P]-
ATP. Resolution of the reaction products was per-
formed by HPLC as described in Experimental. Figure
3 shows the comparison of the chromatograms of the
products of reaction in which both enantiomers are tes-
ted as substrate. For each enantiomer a specific peak of
radioactivity, corresponding to the monophosphate
form, follows the elution of ATP. No peak of labeled
monophosphate is present when the reaction mixture is
immediately heated at 80 ꢀC, centrifuged and injected in
HPLC (data not shown). The slight differences in the
elution time of d- and l-TFTMP in the two elutions are
mirrored by differences in ATP elution, thus they do not
depend upon enantiomeric properties of the com-
pounds. This result demonstrates that, in vitro, HSV-1
TK is able to phosphorylate l-TFT as well as d-TFT.
L-TFT is phosphorylated by HSV-1 thymidine kinase.
To understand whether l-TFT is a non-substrate inhi-
bitor of viral TK or, like its d-counterpart, a substrate-
Under our chromatographic conditions, we can only
quantify the presence of ATP and nucleoside mono-
phosphate since the labeled nucleoside diphosphate
eventually formed by the thymidylate kinase activity
associated with the HSV-1 TK is not completely sepa-
rated from the peak of labeled ATP. Therefore, we have
no information whether l-TFT is further converted to
diphosphate by the viral thymidylate kinase activity.
L-TFT is resistant to phosphorolysis by human thymidine
phosphorylase. The antiviral activity of several nucleo-
side analogues is often limited by their intracellular
rapid degradation by thymidine phosphorylase (TP).
Thus, in an attempt to avoid this degradation and con-
sequently to improve the drug activity, modified
nucleosides have been synthesized; in particular, com-
pound in which the d-deoxyribose moiety has been
replaced by the l-enantiomer.7 In order to verify if
l-TFT, like l-thymidine and its derivatives,7 is resistant to
human TP degradation, we incubated both enantiomer
Figure 1. Effect of different concentrations of l-TFT (panel A)
or TFT (panel B) on HSV-1 (*) and human (*) thymidine kinase
activities.
Table 1. Effect of TFT and l-TFT nucleoside analogues on herpes
virus thymidine kinases
Compound
HSV-1 TK IC50 (mM)
HSV-2 TK IC50 (mM)
Figure 2. Lineweaver–Burk plot of the effect of l-TFT on the activity
of HSV-1 TK in the presence of increasing concentrations of the sub-
strate [3H]-dThd. l-TFT concentrations: 0 mM (*), 0.25 mM (*),
0.5 mM (&), 1 mM (&) and 2 mM (~).
TFT
l-TFT
0.25
0.5
2.0
14.5

1734
R. Salvetti et al. / Bioorg. Med. Chem. 9 (2001) 1731–1738
of TFT with recombinant human TP15 as described
in Experimental, the reaction being followed for 0, 10,
20 and 30 min. Interestingly, L-TFT, contrary to its
d-enantiomer, is completely resistant to degradation by
human TP (Fig. 4).
affected [3H]-dThd incorporation in HeLa 5a cells,
whereas l-enantiomer was inactive (Fig. 5). This fact
suggests that l-TFT is probably unable to cross the cell
membrane17 and, indirectly, that an enantioselective
active transport is responsible for the cellular uptake
of TFT. We assumed that, if present in the cytoplasm,
l-TFT would have competed with [3H]-dThd for the
enzymatic phosphorylation, thus reducing the amount
of [3H]-dThd incorporated into DNA during the DNA
synthesis.
Effect of L-TFT on [3H]-thymidine incorporation in
HeLa TKꢁ/HSV-1 TK+ cells. In order to verify the
ability of l-TFT to inhibit also in vivo thymidine
phosphorylation by HSV-1 TK, we have evaluated if
l-TFT was able to reduce the [3H]-dThd incorporation
into DNA in intact cells. For this study we used HeLa
cells TKꢁ/HSV-1 TK+ (HeLa 5a) in which the human
TK is replaced by HSV-1 TK and [3H]-dThd incor-
poration into DNA is strictly dependent upon viral
enzyme activity, as demonstrated by the use of specific
non-substrate16 and substrate inhibitors7 of viral TK.
Cells were incubated in the presence of [3H]-dThd and
increasing concentrations of l-enantiomer (or d-
enantiomer, as control). As expected, TFT strongly
Antiviral assays
The antiviral activities of L-TFT, comparatively to
TFT, against HSV-1 and HSV-2, were determined by
plaque reduction (PR) assays in human foreskin fibro-
blast (HFF) cells. The results are summarized in Table
2. From these data, it appears that l-TFT does not
demonstrate any antiherpetic activity in cell cultures, in
contrast to its d-counterpart (TFT).
Figure 4. Activity of human thymidine phosphorylase on TFT (*)
and L-TFT (*).
Figure 5. Effect of different concentrations of TFT (*) or L-TFT (*)
on DNA synthesis evaluated as [3H]-dThd incorporation in HeLa cells
TKꢁ/HSV-1 TK+. 100% corresponds to 3 pmol of acid precipitable
[3H]-dTMP after 20 min incubation at 37 ꢀC. Each point is the average
of three determinations.
Figure 3. Phosphorylation of TFT and l-TFT by HSV-1 TK. HPLC
elution profiles of the reaction products of TFT (panel A) and l-TFT
(panel B) after 60 min incubation with HSV-1 TK and [g32P]-ATP.

R. Salvetti et al. / Bioorg. Med. Chem. 9 (2001) 1731–1738
Table 2. Antiviral activities of TFT (Trifluorothymidine) and l-TFT
1735
400 MHz, 13C NMR spectra at 100 MHz and 19F NMR
at 235 MHz in (CD3)2SO at room temperature with a
Bruker DRX 400. Chemical shifts (d) are quoted in
parts per million (ppm) referenced to the residual sol-
vent peak (CD3)CD2HSO being set at d-H 2.49 and d-C
39.5 relative to tetramethylsilane (TMS). 19F chemical
shits are reported using trichlorofluoromethane as
external reference. Deuterium exchange and COSY
experiments were performed in order to confirm proton
assignments. Coupling constants, J, are reported in
Compound
Virus
(cells)a
EC50
(mg/mL)b
CC50
(mg/mL)c
SId
TFT
HSV-1
HSV-2
HSV-1
HSV-2
2.2
10.5
>50
>50
100
100
>50
>50
45.5
9.5
—
L-TFT
—
aHSV-1 and HSV-2 were propagated in human foreskin fibroblasts
(HFF) and virus titer determined by PR assays.
bEC50 (50% effective concentration) or compound concentration
required to reduce viral plaque formation by 50% compared to the
untreated control.
1
Hertz. 2-D H–13C heteronuclear COSY were recorded
for the attribution of 13C signals. FAB mass spectra
were recorded in the positive-ion or negative-ion mode
on a JEOL SX 102. The matrix was a glycerol and
thioglycerol mixture (G–T: 50:50, v/v). Specific rota-
tions were measured on a Perkin-Elmer Model 241
spectropolarimeter (path length 1 cm), and are given in
units of 10ꢁ1 cm2 gꢁ1. Elemental analyses were carried
out by the Service de Microanalyses du CNRS, Division
de Vernaison (France). Thin layer chromatography was
performed on precoated aluminium sheets of Silica Gel
60 F254 (Merck, Art. 5554), visualization of products
being accomplished by UV absorbance followed by
charring with 10% ethanolic sulphuric acid and heating.
Column chromatography was carried out on Silica Gel
60 (Merck, Art. 9385). Commercial reagents and sol-
vents analytical grade were used unless otherwise stated.
[3H-methyl]-thymidine (25 Ci/mmol) and [g32P]-ATP
(3000 Ci/mmol) were from Amersham (Arlington
Heights, IL). 5-(trifluoromethyl)-b-d-20-deoxyuridine
(TFT) (batch 45749000) was purchased from Instel
Chimios S.A. (France).
cCC50 (50% cytotoxic concentration) was defined as the compound
concentration required to reduce the cell number by 50%.
dSI, selectivity index (CC50/EC50).
Conclusion
In this work, we have synthesized in a stereospecific
manner 5-(trifluoromethyl)-b-l-20-deoxyuridine (l-TFT
6) the hitherto unknown l-enantiomer of TFT, in order
to study: i) its affinity for human, HSV-1 and HSV-2
thymidine kinases; (ii) its susceptibility to be phos-
phorylated by the aforementioned thymidine kinases
and to be hydrolyzed by human thymidine phos-
phorylase; (iii) its effect on [3H]-dThd incorporation in
HeLa cells TKꢁ/HSV-1 TK+. Our results have demon-
strated that l-TFT specifically inhibits HSV thymidine
kinases with no effect against human thymidine kinase.
The best inhibition was observed with HSV-1 TK with
an IC50 value comparable to the one obtained with
TFT. The mechanism of inhibition for l-TFT is purely
competitive. HPLC analysis of the reaction products
showed that HSV-1 TK was able to phosphorylate
l-TFT to the corresponding monophosphate with effi-
ciency comparable to TFT. In contrast to its d-counter-
part, l-TFT was fully resistant to degradation by
human thymidine phosphorylase and had no effect on
[3H]-dThd incorporation into DNA in intact cells. Anti-
HSV-1 and HSV-2 activities of l-TFT were subse-
quently evaluated in HFF cells using PR assays.
Unfortunately, the l-enantiomer of TFT did not show
any antiherpetic activity in cell cultures. This lack of
antiviral activity could be attributed to a deficiency of
cellular uptake, of the anabolic pathway from the start-
ing nucleoside to the 50-triphosphate form, or the lack
of interaction of the l-triphosphate with the viral DNA
polymerase. However that may be, the present findings
do not preclude the pursuit of studies on b-l-nucleoside
analogues as potential antiviral drugs. Experiments
related to these topics are currently in progress in our
laboratory.
1-(3,5-Di-O-benzoyl-ꢀ-L-xylo-furanosyl)-5-trifluoromethyl-
uracil 2. A mixture of 5-trifluoromethyluracil (6.3 g,
35 mmol), hexamethyldisilazane (200 mL) and a cataly-
tic amount of ammonium sulphate was refluxed under
argon for 18 h. The clear solution was concentrated to
dryness under reduced pressure. The oily residue was
dissolved in dry CH2Cl2 (170 mL) and a solution of 112
(17 g, 38.5 mmol) in CH2Cl2 (170 mL) and TMSOTf
(14 mL, 77.6 mmol) were added. The reaction was stir-
red for 1 h under argon and diluted with CHCl3
(250 mL). The solution was washed with saturated
NaHCO3 (2ꢂ400 mL), water (2ꢂ200 mL), dried over
anhydrous Na2SO4 and evaporated to dryness. The
resulting crude material was dissolved in a mixture of
pyridine and acetic acid (4:1, v/v, 250 mL), then hydra-
zine hydrate (98%, 5.2 mL, 105 mmol) was added. After
15 h stirring, acetone (100 mL) was added and stirring
was continued for 2 h. The solvents were removed and
the residue was dissolved in CH2Cl2 (300 mL). The
organic layer was washed with saturated NaHCO3
(2ꢂ200 mL) and water (100 mL), dried over anhydrous
Na2SO4 and evaporated to dryness. Silica gel column
chromatography of the residue using as eluent a step-
wise gradient of MeOH (0–2%) in CH2Cl2 afforded
compound 2 as a white foam (14.9 g, 82% overall yield
from 1): [a]2D0 – 83 (c 1.00 in DMSO); UV lmax (EtOH)
Experimental
Chemistry
Evaporation of solvents was carried out on a rotary
evaporator under reduced pressure. Melting points were
determined in open capillary tubes on a Gallenkamp
MFB-595-010 M apparatus and are uncorrected. UV
spectra were recorded on an Uvikon 931 (Kontron)
1
264 nm (e 10100), 231 nm (e 26200); H NMR (DMSO-
d6) d 4.52 (1H, d, J=3.9 Hz, 20-H), 4.65 (1H, m, 50-H),
4.87 (2H, m, 40-H and 500-H), 5.42 (1H, m, 30-H), 5.73
(1H, s, 10-H), 6.47 (1H, d, 20-OH), 7.4–7.9 (10H, m,
1
spectrophotometer. H NMR spectra were recorded at

1736
R. Salvetti et al. / Bioorg. Med. Chem. 9 (2001) 1731–1738
0
2ꢂC6H5), 8.20 (1H, s, 6-H), 11.92 (1H, s, 3-NH); 13C
NMR (DMSO-d6) d 62.7 (50-C), 78.1 (30-C), 78.4 (20-C),
80.6 (40-C), 93.0 (10-C), 103.3 (5-C, q, J=31.8 Hz), 123.5
(CF3, q, J=269.5 Hz), 129.5–134.7 (C-Arom), 141.6 (6-
C, q, J=5.9 Hz), 150.3 (2-C), 159.8 (4-C), 165.3 (CO),
166.4 (CO); 19F NMR (DMSO-d6) d ꢁ61.6 (CF3); m/z
(FAB>0) 521 (M+H)+, 341 (S)+, 181 (BH2)+; m/z
(FAB<0) 1039 (2MꢁH)ꢁ, 519 (M–H)ꢁ, 179 (B)ꢁ.
Anal. calcd for C24H19F3N2O8/0.5 H2O: C, 54.45; H,
3.81; N, 5.29; F, 10.77. Found: C, 54.46; H, 3.80; N,
5.57; F, 10.34.
(1H, d, J1 ꢁ2 =6.9 Hz, 1 -H), 8.51 (1H, s, 6-H), 11.88
(1H, s, 3-NH); 13C NMR (DMSO-d6) d 41.7 (20-C), 60.3
(50-C), 69.2 (30-C), 86.2 (10-C), 87.0 (40-C), 102.8 (5-C, q,
J=31.0 Hz), 123.7 (CF3, q, J=267.0 Hz), 143.9 (6-C, q,
J=6.0 Hz), 150.5 (2-C), 159.9 (4-C); 19F NMR (DMSO-
d6) d ꢁ62.3 (CF3); m/z (FAB>0) 889 (3M + H)+, 593
(2M + H)+, 297 (M+H)+, 181 (BH2)+, 117 (S)+; m/z
(FAB<0) 887 (3MꢁH)ꢁ, 591 (2MꢁH)ꢁ, 295 (MꢁH)ꢁ,
179 (B)ꢁ. Anal. calcd for C10H11F3N2O5: C, 40.54; H,
3.74; N, 9.46. Found: C, 40.63; H, 3.81; N, 9.51.
0
00
1-(5-O-Benzoyl-2-deoxy-ꢀ-L-threo-pentofuranosyl)-5-tri-
fluoromethyluracil 5. Benzoyl chloride (0.94 mL,
8.10 mmol) in dry pyridine (14 mL) was added dropwise
to a stirred solution of compound 4 (2.0 g, 6.75 mmol) in
dry pyridine (55 mL) at 0 ꢀC. After addition was com-
pleted, water (10 mL) was added and the solvents were
removed. The residue was dissolved in CH2Cl2 (50 mL)
and the organic layer was washed with saturated
NaHCO3 (2ꢂ50 mL), water (50 mL), dried over anhy-
drous Na2SO4 and evaporated under reduced pressure.
The residue was purified by silica gel column chroma-
tography using a stepwise gradient of MeOH (0–4%) in
CH2Cl2 to afford compound 5 as a white foam (2.47 g,
86%): [a]2D0 ꢁ62 (c 0.39 in DMSO); UV lmax (EtOH)
1-(3,5-Di-O-benzoyl-2-deoxy-ꢀ-L-threo-pentofuranosyl)-
5-trifluoromethyluracil 3. To a stirred solution of 2
(17.7 g, 34 mmol) in dry CH2Cl2 (200 mL) were added
successively DMAP (12.4 g, 102 mmol) and phenoxy
(thiocarbonyl) chloride (6.90 mL, 51 mmol). After 30
min, the solvent was removed under reduced pressure.
The residue was dissolved in CH2Cl2 (200 mL) and the
organic layer was washed with 0.5 N HCl (2ꢂ200 mL),
brine (300 mL), dried over anhydrous Na2SO4 and
evaporated to dryness. The resulting crude material was
dissolved in dry dioxane (200 mL) AIBN (2.24 g,
13.6 mmol) and tris(trimethylsilyl)silane (13.6 mL,
44.1 mmol) were added. The resultant solution was
heated under reflux for 30 min. After cooling to room
temperature, the solvent was removed under reduced
pressure. Chromatography of the residue on a silica gel
column using as eluent a stepwise gradient of MeOH
(0–2%) in CH2Cl2 afforded compound 3 as a white
foam (15.8 g, 92%): [a]2D0 ꢁ120 (c 1.03 in DMSO); UV
1
262 nm (e 10600), 229 (e 14400); H NMR (DMSO-d6)
0
00
0
00
2.07 (1H, d, J2 ꢁ2 =14.9 Hz, 2 -H), 2.62 (1H, m, 2 -H),
4.27 (1H, m, 40-H), 4.38 (1H, m, 30-H), 4.55 (2H, m, 50-
H and 500-H), 5.680 (1H, d, J=3.1 Hz, 30-OH), 6.03 (1H,
0
0
d, J1 ꢁ2 =7.0 Hz, 1 -H), 7.5–8.0 (5H, m, C6H5), 8.57 (1H,
s, 6-H), 11.88 (1H, s, 3-NH); 13C NMR (DMSO-d6) d
41.5 (20-C), 64.3 (50-C), 69.5 (30-C), 83.5 (40-C), 86.3 (10-
C), 103.2 (5-C, q, J=32.0 Hz), 123.6 (CF3, q,
J=267.0 Hz), 129.6–134.4 (C-Arom), 143.8 (6-C, q,
J=6.0 Hz), 150.6 (2-C), 159.9 (4-C), 166.5 (CO); 19F
NMR (DMSO-d6) d ꢁ62.4 (CF3); m/z (FAB>0) 401
(M+H)+, 221 (S)+, 181 (BH2)+; m/z (FAB<0) 399
(MꢁH)ꢁ, 179 (B)ꢁ. Anal. calcd for C17H15F3N2O6/0.5
H2O: C, 49.88; H, 3.94; N, 6.84; F, 13.92. Found: C,
50.17; H, 4.08; N, 7.01; F, 13.46.
1
lmax (EtOH) 263 nm (e 11900), 231 nm (e 33200); H
NMR (DMSO-d6) 2.50 (1H, m, 20-H), 2.92 (1H, m, 200-
H), 4.65 (2H, m, 40-H and 50-H), 4.80 (1H, m, 500-H),
5.69 (1H, m, 30-H), 6.12 (1H, dd, J1 ꢁ2 =1.7 Hz,
0
0
0
00
J1 ꢁ2 =7.6 Hz), 7.4–7.9 (10H, m, 2ꢂC6H5), 8.20 (1H, s,
6-H), 11.89 (1H, s, 3-NH); 13C NMR (DMSO-d6) d 39.2
(20-C), 62.9 (50-C), 74.0 (30-C), 81.6 (40-C), 86.5 (10-C),
103.4 (5-C, q, J=32.0 Hz), 123.5 (CF3, q, J=267.0 Hz),
129.5–134.6 (C-Arom), 141.6 (6-C, q, J=6.0 Hz), 150.3
(2-C), 159.9 (4-C), 165.5 (CO), 166.3 (CO); 19F NMR
(DMSO-d6)
d
ꢁ62.3 (CF3); m/z (FAB>0) 505
1-(2-Deoxy-ꢀ-L-erythro-pentofuranosyl)-5-trifluorome-
thyluracil 6. DEAD (0.94 mL, 6 mmol) was added to a
stirred solution of compound 5 (0.80 g, 2 mmol), PPh3
(1.57 g, 6 mmol) and benzoic acid (0.73 g, 6 mmol) in dry
tetrahydrofuran (THF) (40 mL) at 0 ꢀC under argon.
The resulting solution was stirred for 1 h at 0 ꢀC. Sol-
vent was removed and the residue was subjected to silica
gel column chromatography using as eluent a stepwise
gradient of MeOH (0–2%) in CH2Cl2. The appropriate
fractions were pooled and directly treated with metha-
nolic ammonia (previously saturated at ꢁ10 ꢀC and
tightly stoppered; 50 mL) for 18 h at room temperature.
Evaporation to dryness and column chromatography
on silica gel using a stepwise gradient of MeOH (0–7%)
in CH2Cl2 afforded compound 6 (136 mg, 23%) which
was crystallized from a chloroform/acetone mixture: mp
180–181 ꢀC (lit.,18 186–189 ꢀC for the d-enantiomer);
[a]2D0 ꢁ41 (c 0.97 in DMSO) {lit.,19 for the d-enantiomer,
[a]2D0 + 47.3 (c 1.00 in water)}; UV lmax (EtOH) 263 nm
(M+H)+, 325 (S)+, 181 (BH2)+; m/z (FAB<0) 1511
(3M – H)ꢁ, 1007 (2M – H)ꢁ, 503 (M – H)ꢁ, 179 (B)ꢁ.
Anal. calcd for C24H19F3N2O7: C, 57.14; H, 3.79; N,
5.55; F, 11.30. Found: C, 56.96; H, 3.85; N, 5.64; F,
11.05.
1-(2-Deoxy-ꢀ-L-threo-pentofuranosyl)-5-trifluoromethyl-
uracil 4. To a stirred solution of 3 (28.0 g, 55.5 mmol)
in dry MeOH (555 mL) was added solid sodium meth-
oxide (4.5 g, 83.3 mmol). The resulting solution was
stirred for 20 min, neutralized by addition of 1 N HCl
(83 mL) and evaporated to dryness. The residue was
subjected to a silica gel column chromatography, with a
stepwise gradient of MeOH (0–7%) in CH2Cl2 to afford
compound 4 (11 g, 67%) which was crystallized from a
chloroform/acetone mixture: mp 178–179 ꢀC; [a]D20 ꢁ3 (c
1
1.01 in DMSO); UV lmax (EtOH) 263 nm (e 9900); H
0
0
00
NMR (DMSO-d6) 2.0 (1H, d, J2 ꢁ2 =14.7 Hz, 2 -H),
2.52 (1H, m, 200-H), 3.69 (2H, m, 50-H and 500-H), 3.90
(1H, m, 40-H), 4.22 (1H, m, 30-H), 4.79 (1H, t,
J=5.5 Hz, 50-OH), 5.30 (1H, d, J=3.0 Hz, 30-OH), 6.03
(e 9700); H NMR (DMSO-d6) 2.20 (2H, m, 20-H an0d
1
200-H), 3.58 (1H, ddd, J5 ꢁ4 =2.9 Hz, J5 ꢁ5 =11.9 Hz, 5 -
0
0
0
00
H), 3.66 (1H, ddd, 500-H), 3.81 (1H, dd, J4 ꢁ3 =6.5 Hz,
0
0

R. Salvetti et al. / Bioorg. Med. Chem. 9 (2001) 1731–1738
1737
40-H), 4.43 (1H, m, 30-H), 5.21 (1H, t, J=4.4 Hz, 50-
OH), 5.26 (1H, d, J=4.4 Hz, 30-OH), 6.07 (1H, t,
J=6.0 Hz, 10-H), 8.72 (1H, s, 6-H), 11.83 (1H, s, 3-NH);
13C NMR (DMSO-d6) d 41.5 (20-C), 61.1 (50-C), 70.2
(30-C), 86.3 (10-C), 88.5 (40-C), 103.3 (5-C, q,
J=31.9 Hz), 123.9 (CF3, q, J=268.8 Hz), 143.1 (6-C, q,
J=6.1 Hz), 150.4 (2-C), 159.9 (4-C); 19F NMR (DMSO-
d6) d ꢁ61.3 (CF3); m/z (FAB>0) 297 (M+H)+, 181
(BH2)+, 117 (S)+; m/z (FAB<0) 295 (MꢁH)ꢁ, 179
(B)ꢁ. UV, NMR and MS spectra were superimposable
on those obtained from a commercially available TFT
sample. Anal. calcd for C10H11F3N2O5: C, 40.54; H,
3.74; N, 9.46; F, 19.24. Found: C, 40.52; H, 3.67; N,
9.54; F, 19.03.
90 ꢀC. The tubes were then centrifuged at 10,000 rpm
for 5 min, the supernatant (20 mL) was injected in HPLC
and analyzed following the gradient conditions described
above. The relative peak areas of nucleoside (thymidine,
TFT or l-TFT) and base (eventually produced) were
used to determine the enzymatic activity of TP.
Cells. The cells used in this study were HeLa 5a (HeLa
TKꢁ transformed to the TK+ phenotype with a func-
tional copy of the HSV-1 TK gene). The cells were
obtained from Professor G. Della Valle (University of
Bologna, Italy). The cells were maintained at 37 ꢀC in
Dulbecco’s modified essential medium (DMEM) con-
taining 10% foetal calf serum (FCS).
In vivo incorporation of [3H]-thymidine in HeLa cells.
HeLa 5a cells were grown in DMEM with foetal calf
serum at 37 ꢀC. Exponentially growing cells were tryp-
sinized, resuspended in DMEM without calf serum at
the concentration of 6ꢂ105/mL and 100 mL aliquots
were incubated for 30 min at 37 ꢀC in test tubes. Finally,
to each tube, [3H]-Thd (2.4 mCi) and different con-
centrations of the nucleoside analogue were added.
Incubation was then continued for 0, 10, 20 and 30 min.
At each time, 0.08 mL samples of culture were spotted
on 25 mm GF/C (Whatman) filters. The filters were
washed three times in 5% (v/v) TCA for 5-10 min and
twice in ethanol. They were then dried and the acid
insoluble radioactivity was estimated by scintillation
counting of the filters in 1 mL of scintillating mixture
Betamax (ICN-Biomedicals).
Biological studies
Thymidine kinase assays. HSV-1 and HSV-2 TK were
purified and assayed as previously described:16,20 briefly,
enzyme was incubated at 37 ꢀC for 30 min in a mixture
(25 mL) containing 30 mM HEPES [4-(2-hydroxyethyl)-
1-piperazineethanesulfonic acid] K+, pH 7.5, 6 mM
MgCl2, 6 mM ATP, 0.5 mM dithiothreitol (DTT), and
0.4 mM (HSV-1) or 0.6 mM (HSV-2) [3H-methyl]-dThd
(2200 cpm/pmol). Human cytosolic TK, purified from
HeLa cells,16 was assayed in the same reaction condi-
tions and 0.5 mM [3H-methyl]-dThd (2200 cpm/pmol).
The reaction was terminated by spotting 20 mL of the
incubation mixture on a 25-mm DEAE paper disk (DE-
81 paper, Whatman). The disk was washed with an
excess of 1 mM ammonium formate (pH 3.6) in order to
remove unconverted nucleoside, then with ethanol. [3H]-
TMP was estimated by scintillation counting in 1 mL of
BetamaxTM (ICN-Biomedicals).
Antiherpetic activities in cell cultures
Media and virus strains. The medium utilized was
Earle’s minimal essential medium (MEM) containing
Earle’s balanced salt solution supplemented with either
2 or 10% fetal bovine serum (FBS), 100 U of penicillin
per mL, 25 mg of gentamicin per mL and 2 mM l-gluta-
mine. HSV-1 (strain E-377) and HSV-2 (strain MS)
were used in the virus inhibition assays.
When [g-32P]-ATP was used in TK assay, enzyme was
incubated in the mixture described above containing
1 mM [g-32P]-ATP (400 cpm/pmol), 2 mM MgCl2 and
40 mM of TFT or l-TFT. After 1 h at 37 ꢀC, to each
sample 1.5 mL of EDTA 0.5 M were added to stop the
reaction. Samples were heated 5 min at 80 ꢀC and cen-
trifuged at 10,000 rpm 10 min. 20 mL of the supernatant
was injected in HPLC.
Cell cultures. The routine growth and passage of human
foreskin fibroblat (HFF) cells were performed using
MEM supplemented with 10% FBS.
The reverse-phase method employing the Bio-Rad
100 MAPS preparative system was used. Nucleosides
and nucleotides were resolved on a Superspher 100
RP18 column (Merck) at room temperature in the fol-
lowing conditions: injection volume, 20 mL; detection,
UV 260 nm; eluents, buffer A (20 mM KH2PO4, pH
5.6), buffer B (20 mM KH2PO4, pH 5.6, 60% metha-
nol). Gradient conditions: 0–5 min, 0% buffer B; 15
min, 50% buffer B; 16–28 min, 100% buffer B. Flow
rate: 0.5 mL/min. 56 fractions (fraction volume: 250 mL)
were collected and counted in a b-counter.
Plaque reduction assay for HSV-1 and HSV-2. 2 days
prior to use, HFF cells were plated into six-well plates
and incubated at 37 ꢀC with 5% CO2 and 90% humid-
ity. On the date of assay, the drug was made up in
MEM with 2% FBS and then serially diluted 1:5 in
MEM using six concentrations of drug. The initial
starting concentration was usually 100 mg/mL down to
0.03 mg/mL. The virus used for infection was diluted in
MEM containing 10% FBS to a desired concentration
which will give 20–30 plaques per well. The media was
then aspirated from the wells and 0.2 mL of virus added
to each well in triplicate with 0.2 mL of media being
added to drug toxicity wells. The plates are then incu-
bated for 1 h with shaking every 15 min. After the incu-
bation period, MEM containing the various drug
concentrations was added to appropriate wells in dupli-
cate and in a volume 2.0 mL. Pooled human globulin
Nucleoside degradation by human thymidine phosphoryl-
ase and HPLC analysis of the reaction products. A mix-
ture (25 mL) containing 0.1 mM thymidine or 0.1 mM of
TFT or l-TFT, 50 mM sodium arsenate, pH 6.2 and an
amount of recombinant thymidine phosphorylase
(TP)15 to give linear reaction rate was incubated at
37 ꢀC for 0, 10, 20 and 30 min and then heated 5 min at

1738
R. Salvetti et al. / Bioorg. Med. Chem. 9 (2001) 1731–1738
obtained from Baxter Health Care Corp. was diluted
1:500 and added to the media that the drug was diluted in
to prevent extra-cellular spread of HSV. The cultures
were then incubated for 3 days. At the end of the incuba-
tion period, 1.0 ml of 0.1% crystal violet in 20% metha-
nol was added to each well and incubated for 10 min. The
liquid was then aspirated off, monolayers were washed,
and plaques were enumerated using a stereomicroscope.
5. Fischer, P. H.; Prusoff, W. H. In Chemotherapyof Viral
Infections, Handbook of Experimental Pharmacology; Came,
P. E.; Caliguiri, L. A., Eds.; Springer-Verlag: Berlin, 1982,
p 95–116.
6. Shannon, W. M. In Antiviral Agents and Viral Diseases
of Man; Galasso, G. J., Ed.; Raven Press: New York, 1984,
p 55–121.
7. Spadari, S.; Maga, G.; Verri, A.; Focher, F. Exp. Opin.
Invest. Drugs 1998, 7, 1285.
8. Wang, P.; Hong, J. H.; Cooperwood, J. S.; Chu, C. K.
Antiviral Res. 1998, 40, 19.
Acknowledgements
9. Robins, M. J.; Wilson, J. S.; Hansske, F. J. Am. Chem. Soc.
1983, 105, 4059, and references therein.
10. Baker, B. R.; Joseph, J. P.; Schaub, R. E.; Williams, J. H.
J. Org. Chem. 1954, 19, 1786.
11. Baker, B. R.; Schaub, R. E.; Kissman, H. M. J. Am.
Chem. Soc. 1955, 77, 5911.
12. Gosselin, G.; Bergogne, M.-C.; Imbach, J.-L. J. Hetero-
cyclic Chem. 1993, 30, 1229.
13. Vorbruggen, A.; Krolikiewicz, K.; Bennua, B. Chem. Ber.
1981, 114, 1234.
14. Hughes, D. L. Org. React. 1992, 42, 335.
15. Focher, F.; Spadari, S.; Ubiali, D.; Pregnolato, M.; Zhi,
C.; Gambino, J.; Wright, G. J. Med. Chem. 2000, 43, 2601.
16. Focher, F.; Hildebrand, C.; Freese, S.; Ciarrocchi, G.;
Noonan, T.; Sangalli, S.; Brown, N. C.; Spadari, S.; Wright,
G. J. Med. Chem. 1988, 31, 1496.
These investigations were supported by Grants from
A.N.R.S., Agence Nationale de Recherche sur le SIDA
(France), CNR PF Biotecnologie and EC TMR-
ERBMRXCT 970125 (Italy) and Novirio Pharmaceu-
ticals, Inc., Cambridge, MA (USA). A.M. is particularly
grateful to the Ministere de l’Education Nationale, de la
Recherche et de la Technologie (France) for a doctoral
fellowship. A.V. was supported by a fellowship pro-
vided by Fondazione A. Buzzati-Traverso.
References
17. Upston, J. M.; Gero, A. M. Biochim. Biophys. Acta 1995,
1236, 249.
18. Heidelberger, C.; Parsons, D. G.; Remy, D. C. J. Med.
Chem. 1964, 7, 1.
19. Kawakami, H.; Ebata, T.; Koseki, K.; Matsushita, H.;
Naoi, Y.; Itoh, K.; Mizutani, N. Heterocycles 1990, 31, 569.
20. Focher, F.; Sangalli, S.; Ciarrocchi, G.; Pedraly-Noy, G.;
Sandoli, D.; Rebuzzini, A.; Wright, G.; Brown, N. C.;
Spadari, S. J. Chemother. 1989, 1, 1107.
1. Heidelberger, C.; Parsons, D.; Remy, D. C. J. Am. Chem.
Soc. 1962, 84, 3597.
2. Carmine, A. A.; Brogden, R. N.; Heel, R. C.; Speight,
T. M.; Avery, G. S. Drugs 1982, 23, 329.
3. Prusoff, W. H.; Zucker, M.; Mancini, W. R.; Otto, M. J.;
Lin, T.-S.; Lee, J.-J. Antiviral Res. 1985, S1, 1.
4. Wigdahl, B. L.; Parkhurst, J. R. Antimicrob. Agents
Chemother. 1978, 14, 470.