Study of human deoxycytidine kinase binding properties using intrinsic fluorescence or new fluorescent ligands

Article abstract of DOI:10.1016/S0223-5234(99)80092-0

A series of D- and L-enantiomers of cytidine or adenosine analogues and fluorescent N-methylanthraniloyl (MeNHBz) cytidine derivatives regiospecifically synthesized from cytidine or deoxycytidine were used to quantify the enantioselectivity of human deoxycytidine kinase (dCK) and to characterize its binding states. Both D- and L-enantiomers of these compounds induced significant bimodal quenchings of the intrinsic fluorescence of the enzyme. The ratios of dissociation constants Kd(D)/Kd(L) for the high affinity binding of non fluorescent cytidine derivatives were remarkably similar. β-D- and β-L-ATP gave monophasic titration curves and the enzyme displayed a preference for the natural enantiomer. This demonstrates the lack of enantioselectivity of dCK in the substrate binding steps of its mechanism. The results of other fluorescence experiments with MeNHBz-cytidine derivatives were consistent with an enzyme mechanism in which nucleotide binding precedes nucleoside binding.

Full text of DOI:10.1016/S0223-5234(99)80092-0

Eur. J. Med. Chem. 34 (1999) 423−431
© Elsevier, Paris
423
Short communication
Study of human deoxycytidine kinase binding properties using intrinsic
fluorescence or new fluorescent ligands
Manijeh Shafieea, Gilles Gosselina, Jean-Louis Imbacha, Gilles Divitab,
Staffan Erikssonc, Georges Maurya*
aLaboratoire de Chimie Bioorganique, UMR 5625 du CNRS, Département de chimie organique fine, Case courrier 006,
Université Montpellier II des Sciences et Techniques du Languedoc,Place Bataillon, 34095 Montpellier Cedex 5, France
bCentre de Recherche de Biochimie Macromoléculaire, UPR-1086 du CNRS, BP 5051, 34293 Montpellier Cedex 5, France
cDepartment of Veterinary Medical Chemistry, The Swedish University of Agricultural Sciences,
The Biomedical Centre, Box 575, SE-751 23 Uppsala, Sweden
(Received 15 July 1998; accepted 3 December 1998)
Abstract – A series of D- and L-enantiomers of cytidine or adenosine analogues and fluorescent N-methylanthraniloyl (MeNHBz) cytidine
derivatives regiospecifically synthesized from cytidine or deoxycytidine were used to quantify the enantioselectivity of human deoxycytidine
kinase (dCK) and to characterize its binding states. Both D- and L-enantiomers of these compounds induced significant bimodal quenchings
of the intrinsic fluorescence of the enzyme. The ratios of dissociation constants KdD/KdL for the high affinity binding of non fluorescent
cytidine derivatives were remarkably similar. â-D- and â-L-ATP gave monophasic titration curves and the enzyme displayed a preference for
the natural enantiomer. This demonstrates the lack of enantioselectivity of dCK in the substrate binding steps of its mechanism. The results
of other fluorescence experiments with MeNHBz-cytidine derivatives were consistent with an enzyme mechanism in which nucleotide binding
precedes nucleoside binding. © Elsevier, Paris
fluorescent cytidine derivatives / deoxycytidine kinase / enantioselectivity
1. Introduction
mechanism is more complex than previously be-
lieved [1]. Furthermore, deoxycytidine kinase has been
reported to have a markedly relaxed enantioselectivity
being able to catalyze the phosphorylation of both enan-
tiomers of a series of cytidine analogues [6]. In this
context, human deoxycytidine kinase (dCK) appears as
the key enzyme in the activation of a number of unnatural
L-nucleoside analogues into their 5≠-triphosphate deriva-
tives since the 5≠-monophosphorylation of nucleoside
analogues is generally considered as the most critical
step [7].
Deoxycytidine kinase (EC 2.7.1.74) is an important
enzyme in the salvage reactions of nucleoside synthesis.
It has a broad substrate specificity and catalyzes the
phosphorylation of all three natural deoxyribonucleo-
sides, â-D-dC, â-D-dA and â-D-dG by transferring the
γ-phosphoryl group of â-D-ATP or â-D-UTP [1]. As a
result, it may activate a large number of nucleoside
analogues and is therefore important in antiviral or
anticancer chemotherapies. An ordered Bi-Bi mechanism
has been postulated with either ATP [2] or the deoxy-
nucleoside [3] as the first substrate to bind to the enzyme.
However, the occurrence of bimodal kinetic curves,
negative cooperativity [4] and the apparent existence of
two binding sites or states [5] suggest that the enzymatic
We are currently studying the enantioselectivity of
recombinant dCK and we have previously shown that the
enzyme catalyzes the phosphorylation of, not only a
broad series of L-cytidine analogues [8], but also
L-adenosine analogues [9]. In the present work, we have
determined the affinity of dCK with respect to a series of
newly prepared fluorescent cytidine derivatives as well as
non fluorescent enantiomeric nucleoside or nucleotide
*Correspondence and reprints

424
Figure 1. N-Methyl anthraniloyl cytidine derivatives synthesized.
analogues using fluorescence spectroscopy, in order to
study the binding sites and the enantioselectivity of the
enzyme.
deoxycytidine with monomethoxytrityl chloride or
t-butyldimethylsilyl chloride, subsequent reaction with
N-methyl isatoic anhydride in the presence of sodium
hydride, and acidic deprotection. Under these conditions,
good yields of mono- or, in one instance, bis-MeNHBz
derivatives were obtained. The rapidly interconverting 2a
and 2b, although isolated by preparative HPLC, were not
stable enough to be used separately in binding experi-
ments and were used as an approximately 65:35 mixture.
All compounds 1–5 fluoresce strongly in water with a
maximum of emission at 440 nm when excited in the
320–350 nm range. These properties compare to that of
2≠(3≠)-MeNHBz-adenosine which presents a maximum
of emission at 428 nm in water upon excitation at
330–350 nm [10].
2. Chemistry
The N-methylanthraniloyl (MeNHBz) fluorophore has
often been used for the characterization of various
nucleoside or nucleotide binding proteins [10–12]. The
efficiency of this marker is probably due to its high
quantum yield and its modest bulk resulting in relatively
slight modifications of the binding efficiency to the
enzyme compared to the parent compounds [13]. The
introduction of the N-methylanthraniloyl group into a
nucleoside or a nucleotide molecule has been achieved in
basic aqueous medium through nucleophilic attack of
N-methyl isatoic anhydride by an alcoholate anion of the
substrate [10]. However, this reaction is restricted to
purine nucleoside or nucleotide labelling and yields only
mixtures of 2≠- or 3≠-O-N-methylanthraniloyl deriva-
tives in rapid equilibrium [12]. To characterize the bind-
ing sites and study the enantioselectivities of several
enzymes of cytidine or deoxycytidine metabolism [9], we
needed cytidine analogues regiospecifically substituted at
various positions on the molecule. The compounds 1–5
(figure 1) were prepared through selective protection of
the adequate hydroxyl and amino groups of cytidine or
3. Biological results and discussion
Prior to studying the binding of compounds 1–5 to
human deoxycytidine kinase, we determined the kinetic
properties of 2, 4 and 5 having a free 5≠-OH group and
thus being potential substrates towards dCK. Under our
conditions, only 2 and 4 were substrates of the enzyme
with Km constants comparable to that of â-D-dC and
â-D-riboC, respectively (figure 2, table I) [8]. In con-
trast, Vm constants were two orders of magnitude lower
than that of the reference compounds. The bis-MeNHBz

425
â-L-cytidine analogues. Assuming that there is a linear
relation between the fluorescence of free enzyme or
enzyme-ligand complex and enzyme or enzyme-ligand
concentration, respectively, dCK was titrated with the
nucleoside or nucleotide analogues in the presence of
MgCl2 (figure 3). We observed a decrease of intrinsic
fluorescence in all titrations without any shift of emission
wavelength maximum. The results were fitted to the
equations (1) or (2) (see experimental) and the probability
that the two fits were equally appropriate was determined
(table II). We found that the quenching of all cytidine
ligands 6–13 was biphasic irrespective of stereochemistry
and that the constants Kd1 and Kd2 corresponding to
enzyme-ligand complex dissociation were practically in-
dependent of the sugar structure. In all cases, the enantio-
selectivity of dCK based on the ratio of Kd constants
favoured only slightly the D-enantiomer in both phases of
the titration (KdD/KdL ratios were in the range of 0.5–0.6
and 0.3–0.5 for Kd1 and Kd2, respectively). The situation
was less straightforward for D- and L-dA but, whatever
the binding mode, dCK showed again a relaxed enantio-
selectivity (table II). The enzyme was also non enantio-
selective in the binding to â-D- or â-L-ATP. In contrast
with nucleoside analogues, in the case of â-ATP, the
results were best interpreted assuming that only one
binding site or enzymatic state was involved with a
preference for the natural enantiomer (KdD/KdL: 0.4)
(figure 3, table II).
Figure 2. Substrate properties of 3´-MeNHBz-riboC (●) and
2´(3´)-MeNHBz-dC (s) with respect to dCK. The medium
contained 50 mM Tris HCl pH 7.5, 5 mM ATP, 5 mM MgCl2,
1 mM dithiothreitol, 45 µg.ml-1 dCK.
derivative 5 was not a substrate. Thus, the affinity of the
MeNHBz-derivatives 1, 3 and 5 for dCK can be directly
assessed by steady state fluorescence studies of the
binding to the enzyme in the presence of MgCl2 and ATP.
The intrinsic fluorescence of human dCK is mostly due
to the presence of seven tryptophan residues per mono-
mer in the dimeric molecule. It has been previously used
to study the binding properties of â-D-dC, â-D-dA or
â-D-dU assuming the exclusive occurrence of static
quenching of fluorescence in the presence of these
compounds [5]. The enzyme fluorescence quenching us-
ing â-D-dC or â-D-dA was bimodal implying the exist-
ence of at least two binding sites or two conformational
states of the enzyme. We used this method to determine
the enantioselectivity of the ligand binding step in the
dCK mechanism with respect to a series of â-D- or
The prepared N-methylanthraniloyl cytidine deriva-
tives 1–4 all bound to dCK regardless of the fluorescent
label position in the molecule. The quenching of intrinsic
dCK fluorescence was most probably unimodal for 1 and
2 and bimodal for 3 and 4 (table II). Because of the
substitution at position 5≠, steady state titration experi-
ments were possible with 3 in the presence of saturating
concentrations of ATP and MgCl2. The presence of ATP
induced a reversal to unimodal quenching with a distinct
increase in the stability of the enzyme-ligand complex
Table I. Substrate properties of MeNHBz-cytidine derivatives with respect to dCK and as compared to unlabelled compounds. The reaction
medium contained 5 mM ATP, 5 mM MgCl2, the substrate (10 to 50 µM) and the enzyme (45 µg/mL for the MeNHBz compounds and
135 ng/mL for the unlabelled compounds).
Compound
Km
Vm
Vm/Km
(µM)
(µmol/min.mg)
ß-D-dC
ß-D-riboC
ß-D-2´(3´)-MeNHBz-riboC, 2
ß-D-3´-MeNHBz-dC, 4
ß-D-3´,N4-BisMeNHBz-dC, 5
9
0.24
0.21
0.002
0.001
a
1
60
63
12
a
0.13
0.001
0.003
–
aNo reaction observed.

426
Figure 3. Titration of 0.5 µM dCK with the enantiomers of A. â-dC or B. â-ATP using the intrinsic tryptophan fluorescence of the
enzyme.
(table II). This suggests that the ATP binding site of the
enzyme is indirectly involved in the binding of 3 to dCK
in the absence of ATP, so that the preliminary binding of
ATP-MgCl2 to the enzyme favours the subsequent bind-
ing of 3. This may be consistent with the enzyme
mechanism in which nucleotide binding precedes nucleo-
side binding [2].
In a different approach to the determination of the
affinities of 1–5 for dCK, the fluorescence of the
N-methylanthraniloyl group was used to monitor the
titration of the ligand with the enzyme. Depending on the
ligand, a 30–100% increase in fluorescence intensity was
observed without any shift of the emission wavelength
maximum. Displacement experiments of 3 from its pre-
formed complex with dCK were performed by adding
saturating concentrations of a high affinity substrate of
the enzyme. Introducing 15 µM to 150 µM â-D-dC in a
first phase of the titration (concentration of added enzyme
below 0.2 µM) gave no change in fluorescence intensity,
whereas the addition of 150 µM â-D-dC in a second
Table II. Affinities of â-nucleoside and nucleotide derivatives to dCK: binding modes and values of dissociation constants (µM) determined
with the Grafit program. The medium contained 50 mM Tris HCl pH 7.5, 100 mM KCl, 1 mM dithiothreitol, 5 mM MgCl2, and 10 % glycerol.
Ligand
Biphasic binding (I)
Monophasic binding (II) Probability that II was
more appropriate than I
Kd1
Kd2
Kd
D-dC, 6
L-dC, 7
D-ddC, 8
L-ddC, 9
D-araC, 10
L-araC, 11
D-riboC, 12
L-riboC, 13
D-dA, 14
L-dA, 15
D-ATP, 16
L-ATP, 17
5’-MeNHBz-riboC, 1
2’(3’)-MeNHBz-riboC, 2
5’-MeNHBz-dC, 3
Idem + 5mM ATP
3’-MeNHBz-dC, 4
0.6 ± 0.2
1.3 ± 0.3
1.5 ± 0.4
2.4 ± 0.9
1.2 ± 0.7
2.4 ± 0.6
1.5 ± 0.8
2.7 ± 0.4
9.3 ± 3
2 ± 1.7
–
170 ± 30
300 ± 90
75 ± 20
270 ± 100
45 ± 10
140 ± 50
75 ± 40
220 ± 140
230 ± 40
95 ± 20
–
–
–
–
–
–
–
–
–
0.002
0.001
0.003
0.008
0.06
0.001
0.09
0.001
0.12
0.12
0.93
0.22
0.16
25 ± 4
65 ± 8
8.4 ± 1
21 ± 2
23 ± 4
24 ± 3
–
–
–
–
–
–
–
0.50
0.005
0.38
4.3 ± 1
–
4.3±0.7
220 ± 150
–
160±80
1.7 ± 0.1
–
0.08

427
phase of the titration (concentration of added enzyme
above 1 µM) resulted in only a 30–40% decrease in
fluorescence intensity. The titration of 3 by dCK in the
presence of 5 mM ATP increased about 2-fold the fluo-
rescence intensity compared to the titration in the absence
of ATP. This effect, already observed in the intrinsic
fluorescence data, is probably due to an improved binding
of the ligand to the protein in the presence of ATP as well
as a possible change of the binding site polarity. The
addition of â-D-dC in the final stage of the titration
decreased the fluorescence intensity to the same level as
in the absence of ATP. Fitting the experiment data to
equation (2) yielded two largely different Kd constants:
0.01 µM and 1.8 µM (figure 4). The value of the second
dissociation constant is comparable to Kd1 obtained for
the same compound using the intrinsic fluorescence of the
enzyme (table II). These results may be explained by the
occurrence of a strong binding of 3 at a non specific site
of dCK as well as a weaker binding at the substrate site.
Such a non specific binding could not be detected in
intrinsic fluorescence experiments with 1–5 because
ligand concentrations were much larger than enzyme
concentrations. Similar examples have been previously
reported in the literature in which the label on a fluores-
cent probe induces binding to an enzyme at a different
site than expected from its structure [14, 15]. Compara-
tively low fluorescence variations were observed in the
binding of 1 to dCK, and displacement experiments (as in
the case of 3) failed to give any significant fluorescence
change. In this case, the experimental data were best
correlated with equation (1), thus suggesting that the
binding was mainly non specific (Kd: 0.01 µM) (fig-
ure 4).
Finally, the results of our study of dCK affinity for
various ligands reflect the complexity of this enzyme,
already evident in kinetic studies of its mechanism [1].
Multiphasic variations of intrinsic or extrinsic fluores-
cence intensities in titrations are probably related to the
existence of several permanent or transient conforma-
tional states or binding sites of the enzyme depending on
substrate concentration. The justification for these mul-
tiple specific bindings of ligands as well as the non
specific binding in the case of 1 and 3 must await the
disclosure of the 3D-structure of the enzyme. Our results
further demonstrate that the lack of enantioselectivity in
the mechanism of human dCK primarily concerns the
binding steps of the substrates to the enzyme. They also
suggest that the binding of â-D-ATP to dCK precedes the
binding of nucleosidic substrates in the enzymatic mecha-
nism.
4. Experimental protocols
4.1. Chemistry
UV spectra were recorded on an Urikson 810 (KON-
1
TRON) spectrophotometer. H NMR spectra were ob-
tained with a Bruker AC 250 spectrometer. Deuterium
exchange and decoupling experiments were performed to
confirm proton 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).
Elemental analyses were carried out by the Service
Central de Microanalyse du CNRS and were within
± 0.4% of the calculated values, unless otherwise stated.
Chromatograms were obtained with a Waters System
Prep 4 000 equipped with an automatic injector WISP
715 ULTRA and a diode array detector 990. An SSQ
7 000 Finnegan MAT mass spectrometer was directly
connected to the chromatograph, and analyses were
performed either in the positive or in the negative mode
of electron spray ionization (ESI).
â-D-dC and â-D-ribo C were obtained from Sigma and
â-D-araC from Upjohn. â-D-ddC was a gift from J.
Balzarini (University of Leuven, Belgium). â-L-dC [16],
â-L-ddC [17], and â-L-dA [18] were synthesized as pre-
viously described. â-L-riboC, â-L-araC and â-L-ATP
were stereospecifically synthesized by multi-step reaction
sequences starting from commercially available L-sugars,
and the details of their synthesis will be reported else-
Figure 4. Titration of N-methyl anthraniloyl cytidine deri-
vatives 1 or 3 (0.1 µM) with human deoxycytidine kinase using
the fluorescence of the enzyme ligands.

428
where. N-Methyl isatoic anhydride has been purchased
from Aldrich.
reaction mixture 12 h at 100 °C. The product was purified
by chromatography on silica gel (eluent: 0–5% methanol
in dichloromethane) as a yellow solid (74%). MS (FAB >
0, G/T): 788 [M + H]+ . MS (FAB < 0, G/T):786 [M –
H]–, 514 [M – mMTr]–. 1H NMR (DMSO-d6) δ 7.65 (1H,
d, J = 7.6Hz, H-6), 7.2 (24H, m, H-arom.rings), 7.0 (1H,
m, NHCH3), 6.78 (4H, m, H-arom.rings), 6.2 (1H, d, J =
7.6Hz, H-5), 5.8 (1H, d, J = 5.3Hz, H-1≠), 5.7 (1H, s,
OH-2≠), 4.3 (1H, s, OH-3≠), 4.1 (1H, m, H-3≠), 3.75 (1H,
m, H-2≠), 3.7 (6H, s, OCH3), 3.65 (1H, m, H-4≠), 2.8
(2H, m, H-5≠ and H-5″). To a solution of â-D-5≠-O,
N4-bis(monomethoxytrityl)cytidine (5 g, 6.3 mmol, 1 eq)
in 50 mL of DMF, sodium hydride (305 mg, 7.6 mmol)
was added. The solution was then heated to 70 °C and
N-methylisatoic anhydride (909 mg, 7 mmol, 1.1 eq) was
added under argon. After 4 h, 50 mL of cold water were
added and the product extracted with chloroform. After
washing and drying, the residual oil was chromato-
graphed on silica gel (eluent: 0.1% methanol in chloro-
form) to give â-D-5≠-O, N4-bis(monomethoxytrityl)-
2≠(3≠)-O-(N-methylanthraniloyl)cytidine (4 g, 4.4 mmol,
70%). MS(FAB > 0, G/T): 921 [M + H]+, 833 [2M + H]+.
MS (FAB < 0, G/T): 919 [M – H]–.
4.1.1. â-D-5≠-O-(N-Methylanthraniloyl)cytidine 1
To a stirred solution of the hydrochloride of â-D-
2≠,3≠-O-(isopropylidene)cytidine (450 mg, 1.4 mmol,
1eq) in anhydrous DMF (6 mL), sodium hydride
(169 mg, 4.2 mmol) was added at room temperature
under argon. The mixture was heated at 70 °C and
N-methylisatoic anhydride was added in small portions.
After 40 h, 20 mL of water were added and the product
extracted with dichloromethane and ethyl acetate. The
residual oil was chromatographed on silica gel (eluent:
gradient of 0–5% ethyl acetate in chloroform) to give
â-D-2≠,3≠-O-(isopropylidene)-5-O-(N-
methylanthraniloyl)cytidine (308 mg, 0.74 mmol, 53%)
1
as an oil. H NMR (DMSO-d6, TMS) δ 7.70 (1H, m,
H-6a), 7.70 (1H, m, H-6), 7.68 (1H, m, H-3a), 7.45 (1H,
m, NHCH3), 7.42 (2H, s, NH2), 6.65 (1H, dd, J3a, 4a
=
J
4a–5a = 8.5Hz, H-4a), 6.5 (1H, t, J = 8.2 Hz, H-5a), 5.7,
(1H, d, J = 3.4 Hz, H-1≠), 5.62 (1H, d, J = 7.4Hz, H-5),
5.0 (1H, m, H-2≠), 4.85 (1H, m, H-3≠), 4.4 (2H, m, H-5≠
and H-5″), 4.3 (1H, m, H-4≠), 2.8 (1H, d, J = 5Hz,
NH-CH3), 1.4 (3H, s, CH3), 1.25 (3H, s, CH3). A solution
of â-D-2≠,3≠-O-(isopropylidene)-5≠-O-(N-methylanthra-
niloyl)cytidine (270 mg, 0.65 mmol) in 50% aqueous
formic acid (9 mL) was stirred for 48 h at room tempera-
ture. After evaporation of the solvent, addition of ethanol
and coevaporation, the residual product was chromato-
graphed on silica gel (eluent: gradient of 0–10% ethyl
acetate in dichloromethane) then purified by preparative
HPLC (eluent: gradient of 0–50% acetonitrile in water in
40 min, 1mL/min), retention time: 20 min. 1 (113 mg,
0.3 mmol, 46%) was obtained as a white solid: m.p.
133–136 °C. MS (FAB > 0, G/T): 377 [M + H]+, 753 [2M
+ H]+ . MS (FAB < 0, G/T): 375 [M – H]–. UV
A solution of â-D-5≠-O, N4-bis(monomethoxytrityl)-
2≠(3≠)-O-(N-methylanthraniloyl)cytidine
(3 g,
3.2 mmol) in 80% aqueous acetic acid (50 mL) was
stirred for 3 d at room temperature. After evaporation and
coevaporation with ethanol, the residue was chromato-
graphed on silica gel (eluent: gradient of 0–10% metha-
nol in dichloromethane). The residual solid was further
purified by preparative HPLC (eluent: gradient of 0–50%
acetonitrile in water, in 30 min) to give 2 as a solid
(50 mg, 0.32 mmol, 62%) directly characterized by MS
(retention times: 18.3 and 20.4 min for the 2≠-and the
3≠-isomer, respectively). M.p. 154–160 °C. UV (EtOH):
λmax 254 (e 15 100), 274 (e 8 300). MS (FAB > 0, G/T):
377 [M + H]+, 753 [2M + H]+ . MS (FAB < 0, G/T): 375
1
(EtOH-95): λmax 254 (e 14 800), 274 (e 8 100). H
1
[M – H]–, 751 [2M – H]–. H NMR (DMSO-d6): 2a
NMR (DMSO-d6) δ 7.7 (1H, dd, J5a, 6a = 8Hz, J4a, 6a
=
1.6 Hz, H-6a), 7.5 (1H, d, J = 7.4Hz, H-6), 7.45 (1H, m,
NHCH3), 7.38 (1H, m, H-3a), 7.1 (2H, d, NH2), 6.65 (1H,
dd, J4a, 5a = J3a, 4a = 8.5 Hz, H-4a), 6.5 (1H, t, J = 8.3Hz,
H-5a), 5.7 (1H, d, J = 3.4Hz, H-1≠), 5.59 (1H, d, J =
7.4Hz, H-5), 5.45 (1H, s, OH-2≠), 5.2 (1H, s, OH-3≠), 4.5
and 4.3 (2H, 2m, H-5≠ and H-5″), 4.1 (1H, m, H-4≠), 3.97
(1H, m, H-3≠), 3.94 (1H, m, H-3≠), 3.95 (1H, m, H-2≠),
2.7 (3H, d, J = 5Hz, NHCH3). Anal. C17H20N4O6, H2O
(C, H, N).
(65%): δ 7.9 (1H, dd, J5a, 6a = 8Hz, J4a, 6a = 1.6 Hz, H-
6a), 7.7 (1H, d, J = 7.4 Hz, H-6), 7.4 (1H, m, CH3NH),
7.4 (1H, m, H-3a), 7.2 (2H, d, NH2), 6.7 (1H, dd, J4a, 5a
≈ J3a, 4a = 8.5 Hz, H-4a), 6.6 (1H, m, H-5a), 5.9 (1H, d,
J = 5.9 Hz, H-1≠), 5.75 (1H, d, J = 7.6Hz, H-5), 5.73 (1H,
m, OH-2≠), 5.18 (1H, m, H-3≠), 4.35 (1H, m, H-2≠), 4.15
(1H, m, H-4≠), 3.65 (2H, m, H-5≠ and H-5≠≠), 2.8 (3H,
d, J = 3.8Hz, NHCH3). 2b (35%): δ 7.7 (2H, H-6 and
H-6a), 7.4 (2H, m, CH3NH and H-3a), 7.2 (2H, d, NH2),
6.7(1H, dd, J4a, 5a ≈ J3a, 4a = 8.5 Hz, H-4a), 6.6 (1H, m,
H-5a), 6.5 (1H, d, J = 5.1Hz, H-1≠), 5.75 (1H, d J =
7.7Hz, H-5), 5.18 (1H, m, H-3≠), 5.1 (1H, m, OH-3≠),
4.35 (1H, m, H-2≠), 3.9 (1H, m, H-4≠), 3.1 (3H, s,
NHCH3). Anal. C17H20N4O6,1/2 H2O (C, H, N).
4.1.2. â-D-2≠(3≠)-O-(N-Methylanthraniloyl)cytidine 2
â-D-5-O,N4-Bis(monomethoxytrityl)cytidine was pre-
pared from cytidine using the procedure previously de-
scribed for adenosine derivatives [19], after heating the

429
4.1.3.
cytidine 3
â-D-5≠-O-(N-Methylanthraniloyl)-2≠-deoxy-
oxycytidine (1.3 g, 1.44 mmol) in 30 mL of 80% aqueous
acetic acid was stirred for 48 h at 50 °C. After evapora-
tion and coevaporation of the reaction medium with
ethanol, the residue was chromatographed on silica gel
(eluent: gradient of 0–12% methanol in chloroform)
yielding 400 mg of 3 as a solid which was further purified
by HPLC (eluent: gradient of 0–50% acetonitrile in
water, in 40 min) as a white solid (350 mg, 1 mmol,
75%). M.p. 215–219 °C. MS (FAB > 0, GT): 361 [M +
H]+, 721[2M + H]+, 383[M + Na+ ]+, 249[M-MeNHBz +
H]+, 134[MeNHBz]+. MS (FAB < 0, G/T): 359[M – H]–,
719[2M – H]–. UV (EtOH 95): λmax 254 (e14 600),
â-D-5≠-O-t-Butyldimethylsilyl-2≠-deoxycytidine was
prepared in 95% yield from 2≠-deoxycytidine using a
published procedure [20]. From this compound (1.6 g,
4.7 mmol), â-D-5≠-O-t-butyldimethylsilyl-N4, 3≠-bis-
(monomethoxytrityl)-2≠-deoxycytidine was prepared in
61% yield using conditions already published [19]. MS
(FAB > 0, G/T): 886 [M + H]+, 614 [M – mMTr + H]+,
273[mMTr]+ . MS (FAB < 0, GT): 884[M – H]–, 612[M
– mMTr – H]–. 1H NMR (DMSO-d6): δ 7.68 (1H, d, J =
7.26 Hz, H-6), 7.5 (24H, m, H-arom.rings), 7.0 (1H, s,
CH3NH), 6.91 and 6.8 (2[2H], 2d, J = 8.9 and 6.3 Hz,
H-arom.rings), 6.3 (1H, d, J = 7.1 Hz, H-5), 6.2(1H, s,
H-1≠), 4.1 (1H, s, H-3≠), 3.9 (1H, s, H-4≠), 3.8 (6H, 2s,
2 [OCH3]), 3.49 and 3.2 (2H, 2m, H-5≠ and H-5″), 1.6
and 1.38 (2H, 2m, H-2≠ and H-2″), 0.8 (9H, s, C-tBu),
0.0(6H, s, Si-tBu). A solution of dry â-D-5≠-O-t-
butyldimethylsilyl-N4, 3≠-bis(monomethoxytrityl)-2≠-
deoxycytidine (2.3 g, 2.6 mmol) in 15 mL of THF was
treated with tetrabutylammonium fluoride (2 mL, 1 eq)
and stirred for 2 h at room temperature. Addition of
30 mL of cold water, extraction with chloroform, washing
with water, drying and evaporation gave a solid which
was chromatographed on silica gel (eluent: 0.2% ethyl
acetate, 0.2% triethylamine in chloroform) and purified as
a white solid (1.76 g, 2.3 mmol, 88%). MS (FAB > 0,
G/T): 772 [M + H]+, 544 [M – mMTr + 2Na+]+,
273[mMTr]+. MS (FAB < 0, G/T): 770 [M – H]–, 534 [M
– mMTr + Cl]–, 498 [M – mMTr]–. â-D-N4,3≠-
1
λmax 274 (e 8 100). H NMR (DMSO-d6) δ 7.75 (1H,
dd, J5a, 6a = 8Hz, J6a, 4a = 1.6 Hz, H-2a), 7.55 (1H, d, J =
7.4 Hz, H-6), 7.45 (1H, d, J = 5 Hz, CH3NH), 6.35 (1H,
t, J3a, 5a = 1.5 Hz, J3a, 4a = 8.5 Hz, H-5a), 7.2 (2H, d,
NH2), 6.75 (1H, d, J5a, 4a = J3a, 4a = 8.5 Hz, H-4a), 6.6
(1H, t, H-3a), 6.15 (1H, t, J = 7.5 Hz, H-1≠), 5.6 (1H, d,
J = 7.4 Hz, H-5), 5.4 (1H, s, OH-3≠), 4.4 (1H, m, H-5≠),
4.35 (1H, m, H-5″), 4.3 (1H, m, H-3≠), 4.05 (1H, m,
H-4≠), 2.8 (3H, d, J = 5 Hz, NHCH3), 2.2 (1H, m, H-2≠),
2.05 (1H, m, H-2″). Anal. C17H20N4O5 (C, H, N).
4.1.4.
cytidine 4
â-D-3≠-O-(N-Methylanthraniloyl)-2≠-deoxy-
â-D-5≠-O-Monomethoxytrityl-2≠-deoxycytidine was
prepared from 2≠-deoxycytidine using a previously re-
ported procedure for adenosine derivatives [19], after
heating the reaction mixture for 48 h at 40 °C. The
product was chromatographed on silica gel (eluent: gra-
dient of 0–12% methanol in chloroform) and obtained as
a solid (49%). MS (FAB > 0, G/T): 500 [M + H]+, 999
[2M + H]+, 1 498 [3M + H]+. MS (FAB < 0 G/T): 498 [M
– H]–, 534 [M + Cl]–, 997 [2M – H]–, 1 033 [2M + Cl]–.
1H NMR (DMSO-d6) δ 7.65 (1H, d, J = 7.6 Hz, H-6), 7.1
(14H, m, H-arom.rings), 7.1 (2H, m, NH2), 6.75 (2H, d,
J = 8.8 Hz, H-arom.rings), 5.98 (1H, t, J = 6.3, H-1≠),
5.55 (1H, d, J = 7.6 Hz, H-5), 5.22 (1H, d, J = 3.9 Hz,
OH-3≠), 4.1 (1H, m, H-3≠), 3.75 (1H, m, H-4≠), 3.6 (3H,
s,-O-CH3), 3.1 (2H, m, H-5≠ and H-5″), 2.1 and 1.95 (2H,
2m, H-2≠ and H-2″). To a solution of â-D-5≠-O-
monomethoxytrityl-2≠-deoxycytidine (1.9 g, 3.8 mmol, 1
eq) in 20 mL of DMF, sodium hydride (167 mg,
4.2 mmol) then N-methylisatoic anhydride (742 mg,
5.7 mmol, 1.5 eq) in 5 mL of DMF were added, and the
mixture was stirred for 36 h at room temperature. After
addition of 20 mL of cold water, extraction with 100 mL
of chloroform then with 50 mL of ethylacetate, the
product was chromatographed on silica gel (eluent: gra-
dient of 0–5% methanol in chloroform) to yield 1 g of a
mixture of two compounds. MS (FAB > 0, G/T): 634 [M
+ 2H]+, 833 [2M + H]+, 273[mMTr]+. MS (FAB < 0,
Bismonomethoxytrityl-2≠-deoxycytidine
(1.71 g,
2.2 mmol) thus obtained was dissolved in 20 mL of dry
DMF and treated with sodium hydride (133 mg,
3.3 mmol). A solution of N-methylisatoic anhydride
(742 mg, 5.7 mmol) in 5 mL of DMF was added and the
mixture was stirred for 10 h at 60 °C. Upon addition of
30 mL of cold water, â-D-5≠-O-(N-methylanthraniloyl)-
N4,3≠-bismonomethoxytrityl-2≠-deoxycytidine precipi-
tated as a white solid which was dried (1.4 g, 1. 55 mmol,
90%). MS (FAB > 0, G/T): 905 [M + H]+, 632 [M –
mMTr + H]+, 273[mMTr]+. MS (FAB < 0, G/T): 903 [M
– H]–, 633 [M – mMTr – H]–. H NMR (DMSO-d6):
1
δ 7.7 (1H, dd, J5a, 6a = 8Hz, J6a, 4a = 1.55 Hz, H-6a), 7.5
(27H, m, H-6, NH, H-arom.ring and H-5a), 7.1 and 6.85
(2[2H], 2d, J = 8.9 and 6.3 Hz, H-arom.rings), 6.7(1H, d,
J5a, 4a = J3a, 4a = 8.4 Hz, H-4a), 6.55 (1H, t, J = 8.3 Hz,
H-3a), 6.15 (1H, d, J = 7.4 Hz, H-5), 5.9 (1H, t, J = 6.9
Hz, H-1≠), 4.45 and 4.35 (2H, 2m, H-5≠ and H-5″), 4.25
(1H, s, H-3≠), 3.97 (1H, s, H-4≠), 3.6 (2[3H], 2s,
2[OCH3]), 2.8 (3H, d, J = 4.9 Hz, NHCH3), 1.45 (2H, m,
H-2≠ and H-2″).
A
solution of â-D-5≠-O-(N-
methylanthraniloyl)-N4,3≠-bismonomethoxytrityl-2≠-de-

430
G/T): 631 [M – H]–, 1 263 [2M – H]–, 667 [M + Cl]–,
764[M – 2H]2–.
The enzyme activity was measured by HPLC analysis
of the reaction medium which contained 50 mM Tris HCl
pH 7.5, 5 mM ATP, 1 mM dithiothreitol, 5 mM MgCl2,
and the enzyme (45 mg/mL). After 10 min of preincuba-
tion at 37 °C, the substrate (10–50 µM) was added.
Analyses of the reaction mixture were performed by
HPLC on a Hypersil ODS 3µ column with the following
eluent (A: Triethylammonium acetate 20 mM pH 6.6, B:
Triethylammonium acetate and 50% of acetonitrile):
10 min of 90% A + 10% B, 15 min from 90% A + 10% B
to 100% B, 10 min of 100% B (1 mL /min). The retention
times in min of each substrate and its 5≠-monophosphate
(when observed) were, respectively: 2 (23.4 and 25.3,
21.2 broad), 4 (27.4, 23.2), 5 (29.6). The products were
characterized by UV and by their mass spectrum which is
directly determined by connecting the HPLC chromato-
graph to the MS spectrometer (mode ESI < 0). The
method did not allow reliable determination of kinetic
parameters at low substrate concentration. For this rea-
son, initial concentrations of substrates were kept above
10 µM. In each case, the reaction was run until about 20%
of the substrate had been transformed. For each concen-
tration of substrate, the kinetic curves were determined by
at least three measurements of substrate transformation as
a function of time. The kinetic parameters were deter-
mined from initial rates with the Lineweaver-Burk
method and the Grafit program (Erithacus Software,
1992). The compounds 1–5 were weak or very weak
inhibitors of dCK since a large excess (100 µM) was
needed in all cases to detect the inhibition with respect to
20 µM â-D-2≠-deoxycytidine as substrate: 30% for 1 and
2, 55% for 3, 50% for 4 and 0% for 5, under the same
conditions as in substrate studies.
The mixture was dissolved in 30 mL of 80% aqueous
acetic acid and the solution was stirred for 3 d at room
temperature. After evaporation, then coevaporation with
ethanol, the residue was chromatographed on silica gel
(eluent: gradient of 0–12% methanol in chloroform) then
purified by HPLC (eluent: gradient of 0–50% acetonitrile
in water, for 40 min) which yielded 2 products: (i) 4
(300 mg, 0.83 mmol, 22%) as a white solid (m.p.
223–227 °C). HPLC/UV/MS: Retention time 26.2 min,
MS ESI > 0: 361 [M + H]+, 462 [M + NEt3 + H]+, 721
[2M + H]+. MS (FAB > 0, GT): 361 [M + H]+, 721 [2M
+ H]+, 743 [2M + Na+ ]+. MS (FAB < 0, G/T): 359 [M –
H]–. UV (EtOH 95: λmax 254 (e 15 600), λmax 274 (e
1
8 600). H NMR (DMSO-d6) δ 7.85 (2H, m, H-6 and
H-6a), 7.58 (1H, m, CH3NH), 7.4 (1H, t, J3a, 5a = 1.5Hz,
J3a, 4a = 8.5 Hz, H-3a), 7.19 (2H d, NH2), 6.7 (1H, d, J5a,
4a = 8.5 Hz, H-4a), 6.6 (1H, t, J5a, 4a = 8.5 Hz, J5a, 6a
=
8Hz, H-5a), 6.25 (1H, dd, J = 5.63 and 8.6Hz, H-1≠), 5.8
(1H, d, J = 7.4 Hz, H-5), 5.35 (1H, d, J = 5.9 Hz, H-3≠),
5.21 (1H, s, OH-5≠), 4.1 (1H, d, J = 1.9Hz, H-4≠), 3.7
(2H, s, H-5 and H-5″), 2.8 (3H, d, J = 5Hz, NHCH3), 2.4
and 2.3 (2H, 2m, H-2≠ and H-2″). Anal. C17H20N4O5: C,
56.66; H, 5.59; N, 15.55. Found: C, 56.26; H, 5.55; N,
14.99. (ii) 5 (130 mg, 0.26 mmol, 7%) as a white solid
(m.p. 226–229 °C). HPLC/UV/MS: Retention time
29.6 min. MS ESI > 0: 494 [M + H]+, 595 [M + NEt3 +
H]+, 988 [2M + 2H]2 +. MS (FAB > 0, GT): 494 [M + H]
+, 228[M – 2MeNHBz + H]+, 134[MeNHBz]+. MS (FAB
< 0, GT): 492 [M – H]–, 985 [2M – H]–. UV (EtOH 95):
1
λmax 257 (e 19 300), λmax 283 (e 15 200). H NMR
(DMSO-d6) δ 7.9 (1H, dd, J = 7.7 and 2.8 Hz, H-6, 7.78
(2H, m, 2 H-6a), 7.48 (7H, m, 2NH, H-3a and H
arom.ring), 6.7 (1H, d, J3a, 4a = J5a, 4a = 8.5 Hz, H-4a), 6.6
(1H, t, J5a, 6a = 7.5 Hz, H-3a), 6.28 (1H, t, J = 6.9 Hz,
H-1≠), 5.4 (1H, t, J = 6.7 and 7.8 Hz, H-3≠), 5.23 (1H, dd,
J = 7.7 and 2.8 Hz, H-5), 5.21 (1H, s, OH-5≠, 4.1 (1H, s,
H-4≠), 3.65 (2H, s, H-5 and H-5″), 3.3 (3H, s, NHCH3),
2.8 (3H, d, J = 5 Hz, NHCH3), 2.4 and 2.2 (2H, 2m, H-2≠
and H-2″). Anal. C25H27N5O6 (C, H, N).
4.3. Fluorescence measurements
The experiments were performed on a SLM-Smart
8 000 spectrofluorometer equipped with a 450 W lamp.
The tryptophan intrinsic fluorescence was excited at
290 nm to minimize the substrate inner-filter effect, and
the emission was measured at 332 nm with a spectral
band pass of 2 nm or 4 nm for excitation or emission,
respectively. Experiments were carried out at 25 °C in
50 mM Tris HCl pH 7.5, 100 mM KCl, 1 mM dithiothrei-
tol, 5 mM MgCl2 and 10% glycerol. In measurements of
intrinsic tryptophan fluorescence variations, the enzyme
(0.5 µM) was equilibrated for 2 h in the fluorescence
buffer then titrated with each ligand (0–600 µM). The
results were corrected for dilution and for internal
quenching as previously described [23]. Curve fitting was
performed with the program Grafit using the equations
4.2. Kinetic experiments
Recombinant human deoxycytidine kinase has been
prepared as previously published [21, 22] with a histidine
tag sequence. It was more than 90% pure and had very
similar properties as compared to the enzyme without a
tag sequence. The enzyme specific activity was 240 U/mg
protein, 1 unit of enzymatic activity being defined as the
amount of enzyme that catalyzes the phosphorylation of
1 nmol of â-D-2≠-deoxycytidine per min at 37 °C.

431
(1) or (2) which assume the existence of one or two
binding sites or enzyme states, respectively:
References
[1] Arner E.S.J., Eriksson S., Pharmacol. Ther. 67 (1995) 155–186.
[2] Kim M.Y., Ives D.H., Biochemistry 28 (1989) 9043–9047.
(1) F = Fmin–{(Eo + Lo + Kd) – [(Eo + Lo + Kd)2 –
4EoLo]0.5}(∆F)/2Eo
[3] Datta N.S., Shewach D.S., Mitchell B.S., Fox I.H., J. Biol. Chem.
264 (1989) 9359–9364.
where F is the observed relative fluorescence intensity,
Fmin is the fluorescence intensity at the beginning of the
titration, ∆F is the variation of fluorescence intensity
between the initial value and at saturating concentration
of substrate, Eo is the total enzyme concentration, Lo is
the total ligand concentration and Kd the dissociation
constant of the enzyme-ligand complex;
[4] Sarup J.C., Fridland A., Biochemistry 26 (1987) 590–597.
[5] Kierdaszuk B., Rigler R., Eriksson S., Biochemistry 32 (1993)
699–707.
[6] Furman P.A., Wilson J.E., Reardon J.E., Painter G.R., Antivir.
Chem. Chemother. 6 (1995) 345–355.
[7] Nair V., Jahnke T.S., Antimicrob. Agents Chemother. 39 (1995)
1017–1029.
[8] Shafiee M., Griffon J.F., Gosselin G., Cambi A., Vincenzetti S., Vita
A., Eriksson S., Imbach J.-L., Maury G., Biochem. Pharmacol. 56
(1998) 1237–1242.
(2) F = Fmin – ({(Eo + Lo + Kd1) – [(Eo + Lo + Kd1)2
– 4EoLo]0.5}(F1)/2Eo + [F2Lo/(Kd2 + Lo)])
[9] Pelicano H., Pierra C., Eriksson S., Gosselin G., Imbach J.-L.,
Maury G., J. Med. Chem. 40 (1997) 3969–3973.
where F1 and F2 are the fluorescence quenching values
induced by the first and the second binding respectively,
Kd1 and Kd2 are the corresponding dissociation con-
stants. The Grafit program was used to determine the
standard errors and to compare the curve fits correspond-
ing to the equations (1) or (2). All calculations were
performed assuming that the stoichiometry of binding is
one ligand per dimer of protein [5].
In experiments using the fluorescence of MeNHBz-
derivatives, a solution of a fixed concentration of fluo-
rescent ligand (0.1 µM) in the fluorescence buffer
(700 mL) was excited at 350 nm and the emission ob-
served at 440 nm with a spectral band pass of 4 or 8 nm
for excitation or emission, respectively. The MeNHBz-
nucleoside was equilibrated at 25 °C then titrated with the
enzyme up to a concentration of 1.5 µM or 5µΜ
depending on the concentration of the ligand, and the
increase of fluorescence intensity corrected for dilution.
In attempting to determine the dissociation constants by
this method, the equations (1) and (2) were used with the
following modifications: the minus sign after Fmin was
replaced by a plus sign in both equations, and F1 and F2
are the fluorescence increases induced by the first and the
second binding, respectively.
[10] Hiratsuka T., Biochim. Biophys. Acta 742 (1983) 496–508.
[11] Woodward S.K.A., Eccleston J.F., Geeves M.A., Biochemistry 30
(1991) 422–430.
[12] Rensland H., Lautwein A., Wittinghofer A., Goody R.S.,
Biochemistry 30 (1991) 11181–11185.
[13] Pelicano H., Maury G., Elalaoui A., Shafiee M., Imbach J.-L.,
Goody R.S., Divita G., Eur. J. Biochem. 248 (1997) 930–937.
[14] Delahunty M.D., Wilson S.H., Karpel R.L., J. Mol. Biol. 236 (1994)
469–479.
[15] Rao R., Al-Shawi M.K., Senior A.E., J. Biol. Chem. 263 (1988)
5569–5573.
[16] Holy A., Collect. Czech. Chem. Commun. 37 (1972) 4072–4087.
[17] Gosselin G., Mathe C., Bergogne M.C., Aubertin A.-M., Somma-
dossi J.-P., Schinazi R., Imbach J.-L., Nucleosides Nucleotides 14
(1995) 611–617.
[18] Urata H., Ogura E., Shinohara K., Ueda Y., Akagi M., Nucleic Acids
Res. 20 (1992) 3325–3332.
[19] Herdewijn P., Pauwels R., Baba M., Balzarini J., De Clercq E., J.
Med. Chem. 30 (1987) 2131–2137.
[20] Ogilvie K.K., Can. J. Chem. 51 (1973) 3799–3807.
[21] Chottinger E.G., Shewach D.S., Data N.S., Ashcraft E., Gribbin D.,
Ginsburg D., Fox I.H., Mitchell B.S., Proc. Natl. Acad. Sci. USA 88
(1991) 1531–1535.
[22] Usova E., Eriksson S., Eur. J. Biochem. 248 (1997) 762–766.
[23] Elalaoui A., Divita G., Maury G., Imbach J.-L., Goody R.S., Eur. J.
Biochem. 221 (1994) 839–846.