Unnatural enantiomers of 5-azacytidine analogues: Syntheses and enzymatic properties

Article abstract of DOI:10.1016/S0223-5234(00)01184-3

Although 2'-deoxy-β-D-5-azacytidine (Decitabine) and β-D-5-azacytidine display potent antileukemic properties, their therapeutic use is hampered by their sensitivity to nucleophiles and to deamination catalysed by cytidine deaminase. As shown earlier [Shafiee M., Griffon J.-F., Gosselin G., Cambi A., Vincenzetti S., Vita A., Erikson S., Imbach J.-L., Maury G., Biochem. Pharmacol. 56 (1998) 1237-1242], β-L-enantiomers of cytidine derivatives are resistant to cytidine deaminase. We thus synthesized several 5-azacytosine β-L-nucleoside analogues to evaluate their enzymatic and biological properties. 2'-Deoxy-β-L-5-azacytidine (L-Decitabine), β-L-5-azacytidine, 1-(β-L-xylo-furanosyl)5-azacytosine, and 1-(2-deoxy-β-L-threo-pentofuranosyl)5-azacytosine were stereospecifically prepared starting from L-ribose and L-xylose. D- and L-enantiomers of 2'-deoxy-β-5-azacytidine were weak substrates of human recombinant deoxycytidine kinase (dCK) compared to β-D-deoxycytidine, whereas both enantiomers of β-5-azacytidine or the L-xylo-analogues were not substrates of the enzyme. As expected, none of the presently reported derivatives of β-L-5-azacytidine was a substrate of human recombinant cytidine deaminase (CDA). The prepared compounds were tested for their activity against HIV and HBV and they did not show any significant activity or cytotoxicity. In the case of L-Decitabine, this suggests that the enantioselectivities of concerned enzymes other than dCK and CDA might not be favourable. (C) 2000 Editions scientifiques et medicales Elsevier SAS.

Full text of DOI:10.1016/S0223-5234(00)01184-3

Eur. J. Med. Chem. 35 (2000) 1011–1019
´
© 2000 Editions scientifiques et me´dicales Elsevier SAS. All rights reserved
PII: S0223-5234(00)01184-3/SCO
Short communication
Unnatural enantiomers of 5-azacytidine analogues: Syntheses and enzymatic
properties
Gilles Gauberta, Christophe Mathe´a, Jean-Louis Imbacha, Staffan Erikssonb, Silvia Vincenzettic,
Daniela Salvatoric, Alberto Vitac, Georges Maurya*
aUMR 5625 du CNRS, De´partement de Chimie, Universite´ Montpellier II, Place Bataillon, 34095 Montpellier Cedex 5, France
bSwedish University of Agricultural Sciences, Department of Veterinary Medical Chemistry, the Biomedical Centre, Box 575,
S-75123 Uppsala, Sweden
cDipartimento di Scienze Veterinarie, Universita` di Camerino, 62024 Matelica, Italy
Received 17 April 2000; revised 21 June 2000; accepted 13 July 2000
Abstract – Although 2%-deoxy-b-
D
-5-azacytidine (Decitabine) and b-D-5-azacytidine display potent antileukemic properties, their
therapeutic use is hampered by their sensitivity to nucleophiles and to deamination catalysed by cytidine deaminase. As shown
earlier [Shafiee M., Griffon J.-F., Gosselin G., Cambi A., Vincenzetti S., Vita A., Erikson S., Imbach J.-L., Maury G., Biochem.
Pharmacol. 56 (1998) 1237–1242], b-
several 5-azacytosine b- -nucleoside analogues to evaluate their enzymatic and biological properties. 2%-Deoxy-b-
-Decitabine), b- -5-azacytidine, 1-(b- -xylo-furanosyl)5-azacytosine, and 1-(2-deoxy-b- -threo-pentofuranosyl)5-azacytosine were
stereospecifically prepared starting from -ribose and -xylose. - and -enantiomers of 2%-deoxy-b-5-azacytidine were weak
substrates of human recombinant deoxycytidine kinase (dCK) compared to b- -deoxycytidine, whereas both enantiomers of
b-5-azacytidine or the -xylo-analogues were not substrates of the enzyme. As expected, none of the presently reported derivatives
of b- -5-azacytidine was a substrate of human recombinant cytidine deaminase (CDA). The prepared compounds were tested for
their activity against HIV and HBV and they did not show any significant activity or cytotoxicity. In the case of -Decitabine, this
L
-enantiomers of cytidine derivatives are resistant to cytidine deaminase. We thus synthesized
L
L
-5-azacytidine
(L
L
L
L
L
L
D
L
D
L
L
L
´
suggests that the enantioselectivities of concerned enzymes other than dCK and CDA might not be favourable. © 2000 Editions
scientifiques et me´dicales Elsevier SAS
b-L-5-azacytidine / L-decitabine / synthesis / deoxycytidine kinase / cytidine deaminase
1. Introduction
5]. After triphosphorylation and incorporation into
cellular DNA, 1 or 2 forms a covalent adduct with the
DNA methylating enzyme DNA-(cytosine-5)methyl-
transferase which is therefore inhibited [6, 7]. This
may induce cell differentiation [8] or may activate
silent genes, particularly tumor suppressor genes [9,
10].
b-
D
-5-Azacytidine (b-
D
-5-azaC), 1, and 2%-deoxy-b-
D
-5-azacytidine (b-
D
-5-azadC), 2, are important anti-
leukemic agents used in clinical treatment [1, 2]. The
latter compound (Decitabine) is a potent antineoplas-
tic compound against human myeloid leukemic cells,
more active than b-
[3]. It has been shown that both b-
b- -5-azadC interfere with the process of DNA
D
-arabinofuranosylcytosine (araC)
The use of b-
D
-5-azaC or b- -5-azadC suffers how-
D
D
-5-azaC and
ever from several drawbacks. The inclusion of an
extra nitrogen atom into the cytosine nucleus in-
creases its chemical sensitivity with respect to nucleo-
philes and accounts for the instability of the com-
pounds in aqueous solution [11]. Decitabine has thus
to be administered as a freshly reconstituted solution
every 4 to 6 h [12]. Furthermore, several mechanisms
of drug resistance operate [10]. A deficiency in deoxy-
D
methylation and that the antileukemic activity of
these compounds originates from their ability to in-
duce the synthesis of hypomethylated DNA in cells [4,
* Correspondence and reprints:
E-mail address: glmaury@univ-montp2.fr (G. Maury).

1012
G. Gaubert et al. / European Journal of Medicinal Chemistry 35 (2000) 1011–1019
limiting step of nucleophilic deprotection, the benzoy-
lated sugar was reduced with sodium borohydride in
HMPA and then reoxidized with bistrimethylsilyltri-
fluoroacetamide to afford the nucleoside analogues in
improved yields [24, 25]. 2%,3%-Dideoxy-b-
tidine was prepared from b- -5-azacytosine using the
Barton–McCombie method [26] whereas the -enan-
tiomer was obtained by coupling the 5%-protected 1-
-2,3-dideoxyribose and the silylated
D
-5-azacy-
D
L
acetoxy-b-
L
heterocycle [16, 27].
Most previous syntheses of 5-azacytidine analogues
yield both a- and b-anomers often difficult to sepa-
rate. For this reason, we used regio- and stereospecific
methods to prepare the
from -ribose or -xylose. 1-O-Acetyl-2,3,5-tri-O-
benzoyl-b- -ribofuranose [28] was coupled with sily-
L
-enantiomers 3–6 starting
L
L
L
Figure 1. b-
D
- and b-
L
-5-azacytidine analogs studied as sub-
lated 5-azacytidine in the presence of trimethyl-
silyltriflate. According to Baker’s rule [29], the 2%-ben-
zoyl group controls the stereochemistry of the substi-
tution and only the b-anomeric derivative 7 was
obtained. Deprotection of 7 catalysed by sodium
strates of human dCK and human CDA.
cytidine kinase (dCK) or the inhibition of this enzyme
by an excess of the feedback inhibitor dCTP hampers
the phosphorylation of the drug. Another major
cause of resistance is induced by the ubiquitous cellu-
lar enzyme cytidine deaminase (CDA) since deamina-
tion of 1 or 2 results in total loss of activity [10, 13,
14].
methanolate in methanol afforded b- -ribofuranosyl-
L
5-azacytosine, 3, in good yield. The 3%- and 5%-posi-
tions of 3 were then protected using dichlorotetra-
isopropyldisiloxane. A Barton–McCombie elimination
[30] of the 2%-hydroxyl group of the 3%,5%-diprotected
compound followed by deprotection gave 2%-deoxy-
We have previously shown that a number of cy-
tidine analogues having the unnatural L stereochem-
istry are both substrates of human deoxycytidine
kinase and resistant to human cytidine deaminase
b-
bine), 4, in 42% overall yield from 3 (figure 2).
1,2-Di-O-acetyl-3,5-di-O-benzoyl- -xylo-furanose
[31] was similarly condensed with silylated 5-azacy-
tosine giving exclusively the b-anomeric derivative 9.
Deprotection with sodium methanolate afforded
L
-erythro-pentofuranosyl-5-azacytosine
( -Decita-
L
L
[15]. We therefore stereospecifically prepared the
L
-
enantiomers of 1 and 2 and other analogues (figure 1),
and we studied their enzymatic properties with respect
to dCK and CDA in the hope of getting phosphoryla-
tion, a lack of deamination of these compounds and
possibly an efficient incorporation into DNA. One
b- -xylofuranosyl-5-azacytosine, 5, in good yield.
L
Deacetylation of 9 followed by a Barton–McCombie
elimination of the 2%-hydroxyl group and debenzoyla-
tion yielded 2%-deoxy-b-
cytosine, 6 (figure 2).
L-threo-pentofuranosyl-5-aza-
b-
L
-5-azacytidine analogue, namely 2%,3%-dideoxy-b- -
L
5-azacytidine, was previously prepared and shown to
display anti-HBV activity but no cytotoxicity [16].
3. Biological results and discussion
The enzymatic properties of b- -5-azacytidine
L
2. Chemistry
derivatives 3–6 were determined with respect to hu-
man recombinant deoxycytidine kinase, to human
recombinant cytidine deaminase, and compared to the
The previous syntheses of 5-azacytidine analogues
are based either on the construction of the s-triazine
ring from a conveniently substituted sugar or on the
coupling of the silylated heterocycle and an halogeno-
properties of the corresponding b- -enantiomers
D
when available.
sugar. Thus, b-
D
-5-azaC and b-
D
-5-azadC have been
Deoxycytidine kinase is a key enzyme in the trans-
formation of nucleoside analogues to the correspond-
obtained using both methods [17–23]. To avoid the

G. Gaubert et al. / European Journal of Medicinal Chemistry 35 (2000) 1011–1019
1013
Figure 2. Syntheses of the b-L-5-azacytidine analogs 3–6.
ing 5%-triphosphate derivatives, whereas cytidine
deaminase often deactivates cytidine derivatives with
antiviral properties. Several studies of their action on
the
D
-enantiomers 1 and 2 have been previously re-
ported, and evaluations of the substrate character of
b- -5-azaC with respect to dCK have shown that it is
D

1014
G. Gaubert et al. / European Journal of Medicinal Chemistry 35 (2000) 1011–1019
either low or non-existent. Thus, kinetics in the pres-
ence of dCK from calf thymus gave a Km of 1200 mM
and a substrate efficiency of only 0.4 % compared to
b-
cells has been reported to have no affinity for
b- -5-azaC [33]. In contrast, b- -5-azaC appears
to be a substrate of uridine-cytidine kinase from vari-
ous sources [34, 35]. Concerning b- -5-azadC
D-dC [32]. Human dCK extracted from leukemic
D
D
D
(Decitabine), all existing studies indicate that this
compound is an average substrate of human or mam-
mal dCK, with Km ranging from 29 to 71 mM depend-
ing on the origin of the enzyme and on the conditions
[36, 37]. More recently, human dCK has been re-
ported to display with Decitabine 20 % of its activity
compared to b-
Most of the preceding kinetic studies of the enzy-
-5-azaC or -dC used
radioactivity to analyse the reaction medium. To take
directly into account the chemical decomposition of
D
-dC [38].
Figure 3. Lineweaver–Burk plots obtained in the phosphory-
lation of 2%-deoxy-b- -5-azacytidine and b- -5-azacytidine
catalysed by human dCK.
D
L
matic phosphorylation of b-
D
chemical properties regarding cytidine, adenosine or
guanosine derivatives [15, 39]. In contrast, no phos-
phorylation was observed in the case of both enan-
tiomers of b-5-azacytidine, 1 and 3. Concerning
D
- or
L
-5-azacytidine derivatives due to hydrolytic
opening of the s-triazine ring [11, 24], we used HPLC
to follow the kinetics of phosphorylation of 1–6 in
the presence of human dCK. Under our conditions,
substrate reversible decomposition was held below
2 % for the duration of the kinetics. With 5mM ATP
b- -5-azacytidine, our findings are in agreement with
D
previous results [35] and they further indicate that
human dCK has similar properties with both enan-
as phosphate donor, only 2%-deoxy-b-
D
-5-azacytidine,
tiomers of b-5-azaC or of b-5-azadC. The b- -xylo-
L
2, and 2%-deoxy-b- -5-azacytidine, 4, were phosphory-
L
or 2%-deoxy-xylo-derivatives 5 and 6 did not display
any substrate properties and, based on the consider-
ably relaxed enantioselectivity of human dCK, it is
likely that their antipodes are also mediocre
substrates.
lated (figure 3, table I). The efficiencies of phosphory-
lation were similar for the two enantiomers and they
were both largely below the efficiency of 2%-deoxy-b-
D
-cytidine phosphorylation (table I). The value of the
-5-azacytidine is similar
Km constant for 2%-deoxy-b-
D
Numerous studies have evaluated the sensitivity of
to the constants previously determined [36, 37]. The
weak enantioselectivity displayed by human dCK in
this case (EL/D=0.66) is consistent with its stereo-
b-
D
-5-azacytidine and 2%-deoxy-b- -5-azacytidine to
D
deamination catalysed by cytidine deaminase. The
source of the purified enzyme has a considerable
Table I. Kinetic parameters corresponding to the phosphorylation of 2%-deoxy-D- or L-5-azacytidine and the other prepared
analogues catalysed by human deoxycytidine kinasea.
Compound
Km (mM)
Vm (mmol/min per mg)
Vm/Km (relat)
2%-Deoxy-b-
2%-Deoxy-b-
2%-Deoxy-b-
D
-5-azacytidine, 2
-5-azacytidine, 4
-cytidine
94919
1794
0.1990.02
0.075
0.05
1
L
0.02490.0006
D
991.6
0.2490.01
b
b
b-
b-
b-
D-5-Azacytidine, 1
b
b
b
b
b
b
L
-5-Azacytidine, 2
-Xylofuranosyl-5-azacytidine, 5
L
2%-Deoxy-b-
L
-xylofuranosyl-5-azacytidine, 6
a The reaction medium contained 5 mM ATP, 5 mM MgCl2, the substrate (50–500 mM) and the enzyme (30 mg/mL) at 37 °C.
b No formation of monophosphate after incubation of the substrate and dCK (500 mg/mL) for 1 h at 37 °C.

G. Gaubert et al. / European Journal of Medicinal Chemistry 35 (2000) 1011–1019
1015
influence on the Km values of the two compounds [13,
40–42] which may partly result from their instability.
Using a low temperature (25 °C) and short kinetic
durations allowed us to limit the decomposition of the
substrates as observed from HPLC analysis. Only the
4.1.1. 1-(i- -Ribofuranosyl)5-azacytosine 3
L
A suspension of 5-azacytosine (2g, 17.9 mmol) in
hexamethyldisilazane (132 mL) and a catalytic amount
of ammonium sulfate was refluxed for 18 h under dry
argon. After evaporation, the oily residue was dissolved
in anhydrous 1,2-dichloroethane (125 mL). To this solu-
D
-enantiomers 1 and 2 were substrates of human
CDA (Km values 225 and 690 mM, respectively)
whereas no deamination occurred for the -enan-
tion were added 1-O-acetyl-2,3,5-tri-O-benzoyl-b-L-
L
ribofuranose (6.02 g, 12 mmol) in 1,2-dichloroethane (62
mL) and trimethylsilyltriflate (4.33 mL, 23.8 mmol). The
mixture was left standing at room temperature under
stirring for 5h, then diluted in chloroform (100 mL) and
neutralized with a saturated solution of sodium hydro-
genocarbonate. The organic layer was separated,
washed twice with 150 mL of water, dried and evapo-
rated. The residue was purified by silica gel column
tiomers 3–6, under the same conditions and in the
presence of increased concentrations of enzyme.
The prepared
L-nucleoside analogues 3–6 were
tested as inhibitors of HIV replication in CEM-SS
and MT-4 cell systems or as inhibitors of HBV repli-
cation in HepG2 cells following standard protocols
[43]. None of the compounds displayed any signifi-
cant antiviral or cytotoxic effect. Several causes could
account for the lack of activity. Our study only shows
chromatography. 1-(2,3,5-Tri-O-benzoyl-b-L-ribofura-
nosyl)5-azacytosine, 7, was eluted with a 0–2% gradient
of methanol in dichloromethane (4.65 g, 70%). To a
stirred suspension of this compound (4.65 g, 8.36 mmol)
in anhydrous methanol (1.8 mL), a solution of sodium
methanolate in methanol (0.2 M, 4.2 mL, 0.84 mmol)
was slowly added under anhydrous conditions. The
reaction mixture was stirred for 4h at room temperature
that
L
-Decitabine may be monophosphorylated in
cells and is resistant to enzymatic deamination. The
substrate properties and enantioselectivities of cellular
nucleotide kinases, viral DNA polymerases or other
concerned enzymes with respect to the reported com-
pounds have not been investigated and could be unfa-
vourable, even if HIV- or HBV DNA polymerases
have been shown to accept triphosphates of some
then left standing at 4 °C. 1-(b-L-Ribofuranosyl)5-azacy-
tosine crystallized as a white solid which was filtered,
washed with water and recrystallized in methanol with a
minimum of water (1.55 g, 76%): mp 233–235 °C; UV
L-nucleoside analogues as substrates or inhibitors
[44–47].
1
(ethanol 95) umax 244 nm, m 6500; H-NMR (DMSO-d6)
l 8.57 (s, 1H, H-6), 7.52 (bs, 2H, NH2), 6.63 (d, 1H,
H-1%, J1%-2%=4.9 Hz), 5.43 (bs, 1H, OH-2%), 5.17 (s, 1H,
OH-5%), 5.06 (bs, 1H, OH-3%), 4.1–3.9 (m, 2H, H-2% and
H-3%), 4.82 (m, 1H, H-4%), 3.7–3.5 (m, 2H, H-5%; H-5%%);
MS (FAB\0, G/T): 245 [M+H]+, 113 [BH2]+,
(FABB0, G/T): 243 [M−H]−, 111 [B]−; [a]D20= −30
(c 1.02, DMSO) (Lit. for the enantiomer: [h]2D0= +22.4
(c 1, H2O) [18]). Anal. C8H12N4O5, 1/5 H2O (C, H, N).
4. Experimental protocols
4.1. Chemistry
Melting points were determined in capillary tubes on
a Gallenkamp MFB-595-010 M apparatus and are un-
corrected. UV spectra were recorded on an Uvikon 810
(Kontron) spectrophotometer. 1H-NMR spectra were
obtained with a Bru¨ker AC 250 spectrometer. Chemical
shifts (l) are reported in ppm. Deuterium exchange and
decoupling experiments were performed to confirm pro-
ton assignments. FAB mass spectra were recorded in the
positive-ion or negative-ion mode on a JEOL DX 300
mass spectrometer: the matrix was a mixture (50/50, v/v)
of glycerol and thioglycerol (G/T). Specific rotations
were measured on a Perkin-Elmer model 241 spectropo-
larimeter. Elemental analyses were carried out by the
Service Central de Microanalyse du CNRS and were
within90.4 % of the calculated values, unless otherwise
4.1.2. 1-[3,5-O-(1,1,3,3-tetraisopropyldisiloxane-
1,3-diyl)-i-
1,3-Dichloro-1,1,3,3-tetra-isopropyldisiloxane
mL, 2.26 mmol) was slowly added to a suspension of
1-(b- -ribofuranosyl)5-azacytosine (0.50g, 2.05 mmol) in
L-ribofuranosyl]5-azacytosine
(0.65
L
anhydrous pyridine (20 mL) under argon. After stirring
for 1 h at room temperature, the mixture was diluted
with ethyl acetate (60 mL) and a saturated aqueous
solution of sodium hydrogenocarbonate (20 mL). The
organic layer was washed with water (20 mL), dried and
evaporated. The residue was chromatographed on silica
stated.
L
-Ribose and
L
-xylose were purchased from
gel (eluent:
0 to 4 % gradient of methanol in
Interchim-France.
dichloromethane). The product was obtained as a white

1016
G. Gaubert et al. / European Journal of Medicinal Chemistry 35 (2000) 1011–1019
powder (0.84 g, 84 %): mp 160–165 °C; UV (ethanol 95)
umax 242, m 5500; H-NMR (DMSO-d6) l 8.31 (s, 1H,
1-(2-deoxy-b-L-erythro-pentofuranosyl)5-azacytosine (86
1
mg, 84%) was recrystallized in methanol with a mini-
mum of water: mp 192–195 (Litt. for the enantiomer:
191–193 °C [18]); UV (ethanol 95): umax 243 nm m 6600
(Lit. for the enantiomer: umax 244 nm, m 6900 [20]);
1H-NMR (DMSO-d6) l 8.49 (s, 1H, H-6), 7.51 (bs, 2H,
NH2), 6.0 (m, 1H, H-1%), 5.22 (bs, 1H, OH-3%), 5.06 (bs,
1H, OH-5%), 4.21 (bs, 1H, H-3%), 3.8–3.5 (m, 3H, H-4%,
H-5%, H-5%%), 4.3–2.1 (m, 2H, H-2%, H-2%%); MS (FAB\
A, G/T) 217 [M+H]+, 113 [BH2]+, (FABB0, G/T):
215 [M−H]−, 107 [S]−; [a]2D0= −44 (c 1.02, DMSO).
Anal. C8H12N4O4, 1/10 H2O (C, H, N).
H-6), 7.56 (bd, 2H, NH2), 5.79 (bs, 1H, OH-2%), 5.53 (bs,
1H, H-1%), 4.3–4.1 (m, 2H, H-2%, H-3%), 4.1–3.8 (m, 3H,
H-4%, H-5%, H-5%%), 1.1–1.0 (m, 28H, iPr); MS (FAB\0,
G/T): 973 [2M+H]+, 487 [M+H]+, 113 [BH2]+,
(FABB0, G/T): 971 [2M−H]−, 485 [M−H]−, 111
[B]−; [a]2D0= +27 (c 1.02, DMSO).
4.1.3. 1-[2-Deoxy-3,5-O-(1,1,3,3-tetra-isopropyl-
disiloxane-1,3-diyl)-i-L-erythro-pentofuranosyl]5-
azacytosine 8
Phenoxy (thiocarbonyl) chloride (0.13 mL, 1.57
mmol) and 4-dimethylaminopyridine (0.38 g, 3.15
4.1.5. 1-(2-O-Acetyl-3,5-di-O-benzoyl-i-
L-
mmol) were added to a solution of 1-(3,5-O-TiPS-b-L-
xylofuranosyl)5-azacytosine 9
ribofuranosyl)5-azacytosine (0.51 g, 1.05 mmol) in anhy-
drous acetonitrile (32 mL). The suspension was stirred
for 4 h at room temperature under argon. After evapo-
ration, the residue was dissolved in dichloromethane (10
mL), washed with water (2×10 mL), with hydrochloric
acid 0.5 N (10 mL), and finally with water (2×10 mL).
The residue obtained after solvent evaporation was co-
evaporated with anhydrous dioxane then dissolved in
this solvent (8.5 mL). Tris(trimethylsilyl)silane (0.4 mL,
2.62 mmol) and a,a%-azoisobutyronitrile (0.05 g, 0.31
mmol) were added to the resulting solution which was
then heated to 100 °C for 2 h. After cooling, the evapo-
ration of the solvent gave a residue which was chro-
matographed on silica gel (eluent: 0 to 2 % gradient of
methanol in dichloromethane). The product 8 was ob-
tained as a white solid: mp 150–153 °C; UV (ethanol 95)
5-Azacytosine (3.65 g, 32.6 mmole) was suspended in
hexamethyldisilazane in the presence of a catalytic
amount of ammonium sulfate. The mixture was refluxed
for 18 h under dry argon. After evaporation, the resid-
ual oil was dissolved in anhydrous 1,2-dichloroethane
(240 mL). This solution was added to a solution of
1,2-di-O-acetyl-3,5-di-O-benzoyl-L-xylofuranose (9.5 g,
21.6 mmol) [31] and trimethylsilyltriflate (7.88 mL, 43.4
mmol) in 100 mL of 1,2-dichloroethane. After 5 h at
room temperature under stirring, the reaction mixture
was diluted with chloroform (200 mL) and neutralized
with an aqueous solution of sodium hydrogenocarbon-
ate. The organic layer was washed with water (2×300
mL), dried and evaporated. The crude residue was
purified by silica gel column chromatography and the
product was eluted with a 0–5% gradient of methanol in
dichloromethane. The final product obtained as a foam
was crystallized in absolute ethanol as white crystals: mp
116–118 °C; UV (ethanol 95) umax 228 nm, m 25800;
1H-NMR (DMSO-d6) l 8.45 (s, 1H, H-6), 7.9–7.4 (m,
12H, 2PhCO and NH2), 5.83 (d, 1H, H-1%, J1%-2%=2.5
Hz), 5.74 (d, 1H, H-3%, J3%-4%=4.2 Hz), 5.54 (t, 1H, H-2%,
J1%-2%=J2%-3%=2.1 Hz), 4.8–4.6 (m, 3H, H-4%, H-5%, H-5%%);
MS (FAB\0, G/T): 495 [M+H]+, 383 [S]+, 105
[PhCO]+ (FABB0, G/T): 493 [M−H]−, 111 [B]−, 121
[PhCO2]−. [a]2D0 = −95 (c 0.98, DMSO).
1
umax 242, m 5700; H-NMR (DMSO-d6) l 8.12 (s, 1H,
H-6), 7.41 (bd, 2H, NH2), 5.77 (dd, 1H, H-1%, J1%-2%=7.5
Hz, J1%-2%%=2.5 Hz), 4.48 (dd, 1H, H-3%, J3%-2%=5.8 Hz,
J3%-2%%=5.7 Hz), 3.9–3.7 (m, 2H, H-5%, H-5%%), 3.6–3.5 (m,
1H, H-4%), 2.4–2.2 (m, 2H, H-2%, H-2%%), 1–0.7 (m, 28H,
iPr); MS (FAB\0, G/T): 941 [2M+H]+, 471 [M+
H]+, 113 [BH2]+, (FABB0, G/T): 469 [M−H]−, 111
[B]−; [a]2D0= +81 (c 1.10, DMSO).
4.1.4. 1-(2-Deoxy-i-L-erythro-pentofuranosyl)5-
azacytosine 4
A solution of tetrabutylammonium fluoride (1M/
THF, 0.91 mL, 0.91 mmol) was slowly added to a
4.1.6. 1-(i-
To a stirred suspension of 1-(2-O-acetyl-3,5-di-O-ben-
zoyl-b- -xylofuranosyl)5-azacytosine (0.3g, 0.61mmol)
L
-xylofuranosyl)5-azacytosine 5
solution of 1-(2-deoxy-3,5-O-TiPS-b-
L
-erythro-pento-
L
furanosyl)5-azacytosine (0.21 g, 0.46 mmol) in anhy-
drous tetrahydrofuran (8.2 mL) under stirring. The
mixture was left 4h at room temperature under stirring,
then it was evaporated to dryness. The oily residue was
added to a minimum of methanol. The precipitate of
in dry methanol (1.8 mL), a methanolic solution of
sodium methanolate (0.02 M, 0.3 mL, 0.06 mmol) was
slowly added at room temperature and under argon.
After 4h of stirring, the mixture was left overnight at
room temperature, and 1-(b-L-xylofuranosyl)5-azacy-

G. Gaubert et al. / European Journal of Medicinal Chemistry 35 (2000) 1011–1019
1017
tosine was filtered and washed with methanol. The solid
obtained as a white powder was recrystallized in
methanol with a minimum of water (114 mg, 77%): mp
110–111 °C; UV (ethanol 95) umax 244 nm, m 4990;
1H-NMR (DMSO-d6) l 8.18 (s, 1H, H-6), 7.39 (bs, 2H,
NH2), 5.75 (d, 1H, OH-2%, J2%-OH=3.7 Hz), 5.50 (s, 1H,
H-1%), 5.31 (d, 1H, OH-3%, J3%-OH=2.8 Hz), 4.74 (t, 1H,
OH-5%, J5%-OH=J5%%-OH=5.6 Hz), 4.11 (m, 1H, H-4%),
3.92 (d, 1H, H-2%, J2%-OH=2.9 Hz), 3.85 (bs, 1H, H,
H-3%), 3.7–3.6 (m, 2H, H-5%, H-5%%); MS (FAB\0, G/T):
solved in dichloromethane (50 mL), and the solution
was washed with water (2×50 mL), with cold hydro-
chloric acid (0.5 N, 50 mL) and again with water (2×50
mL). Drying and evaporation gave the 2%-O-phenyloxy-
carbonyl intermediate which was coevaporated with an-
hydrous dioxane then dissolved in the same solvent
(22.7 mL). To this solution, tris(trimethylsilyl)silane
(1.05 mL, 7.03 mmol) and a,a’-azoisobutyronitrile (0.14
g, 0.84 mmol) were added, and the mixture was heated
at 100 °C for 2h. After cooling and evaporation, the
residue was chromatographed on silica gel using a 0–2%
gradient of methanol in dichloromethane. 1-(2-Deoxy-
245 [M+H]+, (FABB0, G/T): 243 [M−H]−: [a]2D0
+15 (c 1.02, DMSO). Anal. C8H12N4O5 (C, H, N).
=
3,5-di-O-benzoyl-b-L-threo-pentofuranosyl)5-azacyto-
sine (0.91 g, 75 %) was obtained as a white solid which
was recrystallized in absolute ethanol: mp 170–173 °C;
UV (ethanol 95) umax 228 nm, m 25700; 1H-NMR
(DMSO-d6) l 8.51 (s, 1H, H-6), 7.9–7.4 (m, 12H,
2PhCO and NH2), 6.15 (q, 1H, H-1%, J1%-2%=7.4 Hz,
J1%-2%%=1.7 Hz), 5.68 (t, 1H, H-3%, J2%-3%=J3%-4%=5.2 Hz),
4.8–4.6 (m, 3H, H-4%, H-5%, H-5%%), 3.0–2.9 (m, 1H,
H-2%), 2.4 (ld, 1H, H-2%%, J2%-2%%=15.4 Hz); MS (FAB\0,
G/T): 873 [2M+H]+, 437 [M+H], 325 [S]+, 113
[BH2]+, 105 [PhCO]+, (FABB0, G/T): 435 [M−H]−,
121 [PhCO2]−; [a]2D0= −138 (c 1.05, DMSO).
4.1.7. 1-(3,5-Di-O-benzoyl-i-L-xylofuranosyl)5-
azacytosine
To a solution of 1-(2-O-acetyl-3,5-di-O-benzoyl-b-L-
xylofuranosyl)5-azacytosine (5 g, 10.1 mmol) in a mix-
ture of acetic acid (20 mL) and pyridine (85 mL),
hydrazine hydrate 80 % (6.2 mL, 0.10 mol) was added at
room temperature under stirring. After 3h, the reaction
was stopped by adding 50 mL of acetone and stirring
was continued for 45 min. After partial evaporation, the
mixture was poured in water (200 mL) and ethyl acetate
was added (200 mL). The organic layer was washed with
a saturated aqueous solution of sodium hydrogenocar-
bonate (2×200 mL) and with water (2×200 mL). After
drying and evaporation, the residue was chro-
matographed on a silica gel column (eluent: 0 to 5 %
gradient of methanol in dichloromethane) giving the
product as a solid (2.06 g, 45 %) which was recrystal-
lized in ethanol with a minimum of dichloromethane:
mp 110–111 °C; UV (ethanol 95) umax 227 nm, m 25300;
1H-NMR (DMSO-d6) l 8.61 (s, 1H, H-6), 7.9–7.4 (m,
12H, 2PhCO and NH2), 6.38 (bs, 1H, OH-2%), 5.64 (s,
1H, H-1%), 5.37 (d, 1H, H-3%, J2%-3%=3.5 Hz), 4.8–4.6 (m,
3H, H-4%, H-5%, H-5%%), 4.37 (s, 1H, H-2%); MS (FAB\0,
G/T): 905 [2M+H]+, 341 [S]+, 113 [BH2]+, 105
[PhCO]+, (FABBO, G/T): 903 [2−H]−, 111 [B]−, 121
[PhCO2]−; [a]2D0= −123 (c 1.35, DMSO).
4.1.9. 1-(2-Deoxy-i-
azacytosine 6
L-threo-pentofuranosyl)5-
A solution of sodium methanolate in methanol (0.2
M, 0.23 mL, 0.05 mmol) was added to a suspension of
1-(2-deoxy-3,5-di-O-benzoyl-b-L-threo-pentofuranosyl)-
5-azacytosine (225 mg, 0.52 mmol) in anhydrous
methanol (3.9 mL) under argon. After 2h at room
temperature, 1-(2-deoxy-b-L-threo-pentofuranosyl)5-aza-
cytosine precipitated as a white solid which was filtered,
washed with methanol, and recrystallized in methanol
with a minimum of water (92 mg, 77 %): mp 180–
1
182 °C; UV (ethanol 95) umax 243 nm, m 5000 H-NMR
(DMSO-d6) l 8.35 (s, 1H, H-6), 7.40 (bs, 2H, NH2), 5.92
(d, 1H, H-1%, J1%-2%=6.7 Hz), 5.19 (bs, 1H, OH-5%), 4.75
(bs, 1H, OH-3%), 3.9–3.6 (m, 3H, H-4%, H-5%, H-5%%), 2.4
(m, 1H, H-2%), 1.95 (d, 1H, H-2%%, J2%-2%%=14.4 Hz); MS
(FAB\0, G/T) 229 [M+H]+, (FABB0, G/T) 227
[M−H]−; [a]2D0= +15 (c 1.02, DMSO). Anal.
C8H12N4O4 (C, H, N).
4.1.8. 1-(2-Deoxy-3,5-di-O-benzoyl-i-
pentofuranosyl)5-azacytosine 10
To a solution of 1-(3,5-di-O-benzoyl-b-
L
-threo-
-xylofura-
L
nosyl)5-azacytosine (1.27 g, 2.81 mmol) in dry acetoni-
trile (86.5 mL), phenoxy(thiocarbonyl) chloride (0.58
mL, 4.21 mmol) and 4-dimethylaminopyridine (1.03 g,
8.43 mmol) were added under stirring. The mixture was
left under stirring for 2h at room temperature, then it
was evaporated under vacuum. The residue was dis-
4.2. Enzymatic studies
Recombinant human deoxycytidine kinase has been
prepared and purified using human dCK cDNA [48]
cloned into the pET19b vector [49]. The preparation

1018
G. Gaubert et al. / European Journal of Medicinal Chemistry 35 (2000) 1011–1019
contained a histidine tag sequence and was more than
90% pure. The presence of the histidine tag had no
effect on the substrate kinetics of the enzyme [49]. The
specific activity of purified dCK was 60 U/mg protein, 1
unit of enzymatic activity being defined as the amount
of enzyme catalysing the phosphorylation of 1 nmol of
In kinetics of deamination catalysed by CDA, the
reaction mixture consisted of Tris–HCl 0.1 M pH 7.5,
the substrate (100–400 mM) and the enzyme (4.2 mg/mL)
at 25 °C. To hold substrate decomposition below 5%,
the kinetic duration was limited to 5 min and it was
assumed that the rate measured after that time and
under these conditions was not very different from the
initial rate. Kinetics were followed by HPLC as in the
case of dCK but using instead an isocratic elution with
triethylammonium acetate 20 mM pH 6.6 at a flow rate
of 1 mL/min. The retention times in min of the sub-
strates were: 1 (3.15), 2 (5.25), 3 (3.10), 4 (5.30), 5 (3.15),
and 6 (4.00). The corresponding products of reversible
decomposition were eluted at 2.3 min, 6.4 min, 2.25 min,
6.4 min, 2.4 min, and 5.00 min, respectively. Under
these conditions, no deamination product of compounds
1 and 2 was detected using UV detection between 220
and 400 nm, and especially at 240 nm near the maxi-
mum of absorption of 1-methyl-5-azauracil [50]. For this
reason, the progress of reaction was measured solely
from the decrease in intensity of the peak due to the
reactant. This was possible with good accuracy because
of the precision and reproducibility of the automatic
injector of the Waters System Alliance 2690. A nu-
cleoside analogue was considered as non substrate if
there was no detectable reaction after incubation of the
compound in 400 mM concentration in the presence of
CDA (42 mg/mL) for 1 h at 25 °C.
b-D-2%-deoxycytidine/min at 37 °C. Recombinant human
cytidine deaminase was prepared and purified as re-
ported [40]. The specific activity of pure CDA was 62
U/mg protein, 1 unit being defined as the amount of
enzyme catalysing the deamination of 1 mmole of b-
cytidine per min at 37 °C. b- -dC, b- -5-azaC and
b- -5-azadC were obtained from Sigma.
In kinetic studies with dCK, the reaction medium
D-
D
D
D
contained 50 mM Tris–HCl pH 7.6, 5 mM ATP, 5 mM
MgCl2, 1 mM dithiothreitol, 15 mM KCl, a concentra-
tion of substrate in the range of 50 to 500 mM depending
on the experiment, and 30 mg/mL of dCK. The reaction
was run at 37 °C until about 10 % of the substrate had
been transformed, and experiment length did not exceed
30 min. The compounds were considered as non sub-
strates of the enzyme if there was no phosphorylation
after 1 h at 37 °C in the presence of dCK 500 mg/mL.
Analysis of the reaction medium was performed by
HPLC on a Hypersil ODS 3 m column using a Waters
System Alliance 2690 equipped with a diode array detec-
tor Waters 996. 10 min of isocratic elution using eluent
A [5 mM Pic A-Waters (ion pairing agent: tetrabutylam-
monium hydrogen sulfate and phosphoric acid)] fol-
lowed by a 30-min gradient from eluent A to eluent B (5
mM Pic A in 50% aqueous acetonitrile) allowed the
separation of all reactants and products using a flow
rate of 1 mL/min. The retention times in min of each
substrate and its 5%-monophosphate (when observed)
were, respectively: 1 or 3 (2.1), 2 or 4 (3.5, 11.2), 5 (3.1),
6 (4.1). The retention times of ATP and ADP were 24.6
min and 21 min, respectively. Identifications of products
resulted from their UV spectra which were determined
at several wavelengths to identify the eventual decompo-
sition products. Based on the known absorptions of the
decomposition products [11, 24], it was estimated that
less than 5 % of the substrate was decomposed at the
end of any kinetic study. For each concentration of
substrate, the kinetic curves were determined by at least
three measurements of substrate transformation in func-
tion of time. Using the GraFit program (Erithacus
Software, 1992), the initial rates were obtained and used
to determine the apparent Vm and Km parameters and
standard deviations according to the Lineweaver–Burk
method.
Acknowledgements
This work was financed by the Agence Nationale de
Recherches sur le SIDA (ANRS, France) and the Centre
National de la Recherche Scientifique (CNRS, France).
We thank Dr A.-M. Aubertin (Universite´ Louis Pasteur,
Strasbourg, France) and Professor J.-P. Sommadossi
(University of Alabama at Birmingham, AL, USA) for
the results of the biological evaluations.
References
[1] Sorm F., Vesely J., Neoplasma 15 (1968) 339–343.
[2] Momparler R.L., Momparler L.F., Samson J., Leuk. Res. 8
(1984) 1043–1049.
[3] Momparler R.L., Onetto-Pothier N., Momparler L.F., Leuk.
Res. 14 (1990) 755–760.
[4] Bouchard J., Momparler R.L., Mol. Pharmacol. 24 (1983)
109–114.

G. Gaubert et al. / European Journal of Medicinal Chemistry 35 (2000) 1011–1019
1019
[30] Barton D.H.R., McCombie S.W., J. Chem. Soc., Perkin Trans.
16 (1975) 1574–1585.
[5] Wilson V.L., Jones P.A., Momparler R.L., Cancer Res. 43
(1983) 3493–3496.
[31] Gosselin G., Bergogne M.-C., Imbach J.-L., J. Heterocyclic
Chem. 30 (1993) 1229–1233.
[6] Cheng X., Curr. Opin. Struct. Biol. 5 (1995) 4–10.
[7] Ahmad I., Rao D.N., Crit. Rev. Biochem. Mol. Biol. 31 (1996)
361–380.
[32] Krenitsky T.A., Tuttle J.V., Koszalka G.W., Chen I.S.,
Beacham III L.M., Rideout J.L., Elion G.B., J. Biol. Chem. 251
(1976) 4055–4061.
[8] Momparler R.L., Pharmac. Ther. 30 (1985) 287–299.
[9] Jones P.A., Cell 40 (1985) 485–486.
[33] Coleman C.N., Stoller R.G., Drake J.C., Chabner B.A., Blood
46 (1975) 791–803.
[10] Momparler R.L., Coˆte´ S., Eliopoulos N., Leukemia 11 (Suppl.
1) (1997) 1–6.
[34] Lee T., Karon M., Momparler R.L., Cancer Res. 34 (1974)
2482–2488.
[11] Lin K.-T., Momparler R.L., Rivard G.E., J. Pharm. Sci. 70
(1981) 1228–1232.
[35] Vesely J., Cihak A., Pharmac. Ther. 2 (1978) 813–840.
[36] Momparler R.L., FEBS Symp. 57 (1979) 33–41.
[37] Vesely J., Cihak A., Neoplasma 27 (1980) 121–127.
[12] Pinto A., Zagonel V., Leukemia, Monograph 1 (7 Suppl.)
(1993) 51–60.
[13] Chabot G.G., Bouchard J., Momparler R.L., Biochem. Phar-
macol. 32 (1983) 1327–1328.
[38] Eriksson S., Kierdaszuk B., Munch-Petersen B., Oberg B.,
Johansson N.G., Biochem. Biophys. Res. Commun. 176 (1991)
586–592.
[14] Laliberte´ J., Marquez V.E., Momparler R.L., Cancer
Chemother. Pharmacol. 30 (1992) 7–11.
[39] Gaubert G., Gosselin G., Boudou V., Imbach J.-L., Eriksson
S., Maury G., Biochimie 81 (1999) 1041–1047.
[15] Shafiee M., Griffon J.-F., Gosselin G., Cambi A., Vincenzetti
S., Vita A., Erikson S., Imbach J.-L., Maury G., Biochem.
Pharmacol. 56 (1998) 1237–1242.
[40] Vincenzetti S., Cambi A., Neuhard J., Garattini E., Vita A.,
Protein Expression Purif. 8 (1996) 247–253.
[16] Lin T.-S., Guo X., Luo M.-Z., Lui M.-C., Zhu Y.-L.,
Dutschman G.E., Pai S.B., Li M.-M., Cheng Y.-C., Nu-
cleosides Nucleotides 14 (1995) 619–625.
[41] Vita A., Cacciamani T., Natalini P., Ruggieri S., Magni G.,
Adv. Exp. Med. Biol. 253B (1989) 71–77.
[42] Cacciamani T., Vita A., Cristalli G., Vincenzetti S., Natalini P.,
Ruggieri S., Amici A., Magni G., Arch. Biochem. Biophys. 290
(1991) 285–292.
[17] Piskala A., Sorm F., Collect. Czech. Chem. Commun. 29 (1964)
2060–2076.
[18] Winkley M.W., Robins R.K., J. Org. Chem. 35 (1970) 491–
495.
[43] Meier C., Aubertin A.-M., De Monte M., Faraj A., Som-
madossi J.-P., Pe´rigaud C., Imbach J.-L., Gosselin G., Antiviral
Chem. Chemother. 9 (1999) 41–51.
[19] Niedballa U., Vorbruggen H., J. Org. Chem. 39 (1974) 3672–
3674.
[44] Faraj A., Agrofoglio L.A., Wakefield J.K., McPherson S.,
Morrow C.D., Gosselin G., Mathe´ C., Imbach J.-L., Schinazi
R.F., Sommadossi J.-P., Antimicrobial Agents Chemother. 38
(1994) 2300–2305.
[20] Piskala A., Sorm F., Nucl. Acid Chem. 76 (1978) 435–441.
[21] Pilml J., Sorm F., Collect. Czech. Chem. Commun. 29 (1964)
2576–2577.
[45] Maga G., Amacker M., Hu¨bscher U., Gosselin G., Imbach
J.-L., Mathe´ C., Faraj A., Sommadossi J.-P., Spadari S., Nu-
cleic Acids Res. 27 (1999) 972–978.
[22] Novotny L., Vachalkova A., Piskala A., Collect. Czech. Chem.
Commun. 59 (1994) 1691–1693.
[23] Hanna N.B., Masojidkova M., Fiedler P., Piskala A., Collect.
Czech. Chem. Commun. 63 (1998) 222–230.
[46] Davis M.G., Wilson J.E., Van Draanen N.A., Miller W.H.,
Freeman G.A., Daluge S.M., Boyd F.L., Aulabaugh A.E.,
Painter G.R., Boone L.W., Antivir. Res. 30 (1996) 133–145.
[24] Beisler J.A., J. Med. Chem. 21 (1978) 204–208.
[25] Beisler J.A., Abbasi M.M., Driscoll J.S., J. Med. Chem. 22
(1979) 1230–1234.
[47] Aguesse-Germon S., Liu S.-H., Chevallier M., Pichoud C.,
Jamard C., Borel C., Chu C.K., Tre´po C., Cheng Y.-C.,
Zoulim F., Antimicrobial Agents Chemother. 42 (1998) 369–
376.
[26] Kim C.-H., Marquez V.E., Broder S., Mitsuya H., Driscoll J.S.,
J. Med. Chem. 30 (1987) 862–866.
[27] Lin T.-S., Luo M.-Z., Liu M.-C., Tetrahedron 51 (1995) 1055–
1068.
[48] Chottinger E.G., Shewach D.S., Datta N.S., Ashcraft E., Grib-
bin D., Ginsburg D., Fox I.H., Mitchell B.S., Proc. Natl. Acad.
Sci. USA 88 (1991) 1531–1535.
[28] Recondo E.F., Rinderknecht H., Helv. Chim. Acta 42 (1959)
1171–1173.
[49] Usova E.V., Eriksson S., Eur. J. Biochem. 248 (1997) 762–766.
[29] Baker B.R., Schaub R.E., Kissman H.M., J. Am. Chem. Soc.
77 (1955) 5911–5915.
[50] Jonas J., Horak M., Piskala A., Gut J., Collect. Czech. Chem.
Commun. 27 (1962) 2754–2760.