Evaluation of the role of three candidate human kinases in the conversion of the hepatitis C virus inhibitor 2′-C-methyl-cytidine to its 5′-monophosphate metabolite

Article abstract of DOI:10.1016/j.antiviral.2009.10.020

Nucleoside analogs are effective inhibitors of the hepatitis C virus (HCV) in the clinical setting. One such molecule, 2′-C-methyl-cytidine (2′-MeC), entered clinical development as NM283, a valine ester prodrug form of 2′-MeC possessing improved oral bioavailability. To be active against HCV, 2′-MeC must be converted to 2′-MeC triphosphate which inhibits NS5B, the HCV RNA-dependent RNA polymerase. Conversion of 2′-MeC to 2′-MeC monophosphate is the first step in 2′-MeC triphosphate production and is thought to be the rate-limiting step. Here we investigate which of three possible enzymes, deoxycytidine kinase (dCK), uridine-cytidine kinase 1 (UCK1), or uridine-cytidine kinase 2 (UCK2), mediate this first phosphorylation step. Purified recombinant enzymes UCK2 and dCK, but not UCK1, could phosphorylate 2′-MeC in vitro. However, siRNA knockdown experiments in three human cell lines (HeLa, Huh7 and HepG2) defined UCK2 and not dCK as the key kinase for the formation of 2′-MeC monophosphate in cultured human cells. These results underscore the importance of confirming enzymatic kinase data with appropriate cell-based assays. Finally, we present data suggesting that inefficient phosphorylation by UCK2 likely limits the antiviral activity of 2′-MeC against HCV. This paves the way for the use of a nucleotide prodrug approach to overcome this limitation.

Full text of DOI:10.1016/j.antiviral.2009.10.020

Antiviral Research 85 (2010) 470–481
Contents lists available at ScienceDirect
Antiviral Research
journal homepage: www.elsevier.com/locate/antiviral
Evaluation of the role of three candidate human kinases in the conversion
of the hepatitis C virus inhibitor 2ꢀ-C-methyl-cytidine to its
5ꢀ-monophosphate metabolite
Nina L. Golitsina∗, Francis T. Danehy Jr., Ross Fellows1, Erika Cretton-Scott2, David N. Standring
Idenix Pharmaceuticals, Inc., 60 Hampshire Street, Cambridge, MA 02139, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
Nucleoside analogs are effective inhibitors of the hepatitis C virus (HCV) in the clinical setting. One
such molecule, 2ꢀ-C-methyl-cytidine (2ꢀ-MeC), entered clinical development as NM283, a valine ester
prodrug form of 2ꢀ-MeC possessing improved oral bioavailability. To be active against HCV, 2ꢀ-MeC must
be converted to 2ꢀ-MeC triphosphate which inhibits NS5B, the HCV RNA-dependent RNA polymerase.
Conversion of 2ꢀ-MeC to 2ꢀ-MeC monophosphate is the first step in 2ꢀ-MeC triphosphate production and is
thought to be the rate-limiting step. Here we investigate which of three possible enzymes, deoxycytidine
kinase (dCK), uridine–cytidine kinase 1 (UCK1), or uridine–cytidine kinase 2 (UCK2), mediate this first
phosphorylation step. Purified recombinant enzymes UCK2 and dCK, but not UCK1, could phosphorylate
2ꢀ-MeC in vitro. However, siRNA knockdown experiments in three human cell lines (HeLa, Huh7 and
HepG2) defined UCK2 and not dCK as the key kinase for the formation of 2ꢀ-MeC monophosphate in
cultured human cells. These results underscore the importance of confirming enzymatic kinase data
with appropriate cell-based assays. Finally, we present data suggesting that inefficient phosphorylation
by UCK2 likely limits the antiviral activity of 2ꢀ-MeC against HCV. This paves the way for the use of a
nucleotide prodrug approach to overcome this limitation.
Received 10 March 2009
Received in revised form 20 October 2009
Accepted 21 October 2009
Keywords:
Antiviral drugs
siRNA
Nucleoside/nucleotide derivatives
Metabolism and activation
Uridine–cytidine kinase
Deoxycytidine kinase
© 2009 Elsevier B.V. All rights reserved.
1. Introduction
The HCV RNA-dependent RNA polymerase, NS5B, functions in
viral replication and is a highly pursued target. Drugs targeting
Chronic hepatitis C virus (HCV) infection is thought to affect
approximately 170 million individuals worldwide, and it can cause
cirrhosis of the liver and hepatocellular carcinoma. The standard
therapy for treating chronic HCV is a combination of pegylated
interferon (IFN) and ribavirin; however, this therapy is not effec-
tive in all patients and is associated with significant side effects.
Less than 50% of patients infected with HCV genotype 1, the most
common HCV genotype in the United States and Europe, benefit
from IFN/ribavirin therapy (Pawlotsky et al., 2007). In recent years
a substantial effort has been made to develop new inhibitors for
HCV. These molecules target various steps of the viral life cycle,
including viral entry, HCV RNA translation, post-translational pro-
cessing, HCV replication, and virus assembly or release (Pawlotsky
et al., 2007).
NS5B are divided into two classes: non-nucleoside analogs, which
bind to allosteric sites on the enzyme and inhibit RNA synthesis ini-
tiation and/or elongation (Liu et al., 2006; Cooper et al., 2007), and
nucleoside analogs, which bind to the enzyme’s active site and typ-
ically cause termination of viral RNA synthesis (De Francesco and
Carfí, 2007). Prodrug forms of several NS5B nucleoside inhibitors
have advanced into HCV clinical trials. The underlying nucleoside
analogs include 2ꢀ-C-methyl-cytidine or 2ꢀ-MeC (Standring et al.,
2003), 2ꢀ-fluoro, 2ꢀ-methyl-cytidine or 2ꢀ-F-2ꢀ-MeC (Ali et al., 2008;
Reddy et al., 2007; Lalezari et al., 2008) and 4ꢀ-azido cytidine or
4ꢀ-azidoC (Nelson et al., 2008). However, the utility of drugs tar-
geting NS5B has been somewhat limited by various factors such as
toxicity, the need for high doses of drug and rapid development of
resistance or lack of efficacy (Meanwell and Koszalka, 2008). While
HCV nucleosides appear to have a favorable resistance profile (Le
Pogam et al., 2008), the cytosine analogs currently in clinical trials
must be given at high doses to elicit potent antiviral effects in the
clinic.
∗
Corresponding author. Tel.: +1 617 995 9800; fax: +1 617 995 9801.
E-mail address: golistina.nina@idenix.com (N.L. Golitsina).
Present address: Centocor Discovery Research, 3210 Merryfield Row, San Diego,
In this study, we sought to further investigate factors influ-
encing the efficacy of the cytidine nucleoside analog 2ꢀ-MeC.
NM283 is a prodrug form of 2ꢀ-MeC that was shown to inhibit
HCV replication in HCV-infected chimpanzees (Standring et al.,
1
CA, USA.
2
Present address: McWhorter School of Pharmacy, Samford University, 800
Lakeshore Drive, Birmingham, AL, USA.
0166-3542/$ – see front matter © 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.antiviral.2009.10.020

N.L. Golitsina et al. / Antiviral Research 85 (2010) 470–481
471
2003) and in humans (Poordad et al., 2007), but was subsequently
2.2. Cloning and expression
discontinued. 2ꢀ-MeC, like other nucleoside analog inhibitors of
NS5B, must be converted into a 5ꢀ-triphosphate derivative to be
active.
cDNA clones containing gene NM 031432 or gene NM 012474,
for UCK1 and UCK2, respectively, were obtained from OriGene
(Rockville, MD). Full-length human UCK1 (encoding a 277 amino
acid protein of 31 kDa) and UCK2 (encoding a 261 amino acid
protein of 29 kDa) were amplified by PCR with primers con-
taining NdeI and XhoI restriction enzyme sites. UCK1 primers
were (from 5ꢀ to 3ꢀ) ACATATGGCTTCGGCGGGAGGCG and CTCTC-
GAGTCAGTGGGGTCTGCTGCTGGA; UCK2 primers were (from 5ꢀ
to 3ꢀ) AACATATGGCCGGGGACAGCGAGCAG and CTCTCGAGTCAAT-
GCGGCCTGCTGCTGGA. The PCR-amplified fragments were cloned
at a 5ꢀ-NdeI and 3ꢀ-XhoI sites into the pET-15b vector (EMD Bio-
sciences, Madison, WI), that contains an N-terminal His-tag. A cDNA
clone for the dCK gene NM 000788.1 was obtained from Invitrogen
(Carlsbad, CA) and sub-cloned into the pDEST17 expression vector
with an N-terminal His-tag (Invitrogen, Carlsbad, CA). The expres-
sion plasmids were sequenced (Tufts University, Boston, MA) to
confirm the correct full-length gene sequences. Expression plas-
mids with UCK1 or UCK2 were transformed into the Escherichia
coli strain BL21 (DE3) (Invitrogen, Carlsbad, CA) and dCK into
E. coli BL21-AITM (Invitrogen, Carlsbad, CA) competent cells. Five
milliliters of overnight culture grown from a single transformed
colony were added to 500 mL of fresh LB medium supplemented
with 100 ␮g/mL carbenicillin and grown at 37 ◦C. Protein expres-
sion was induced at cell density OD600 = 0.8 (UCK1 and UCK2) or at
a cell density of OD600 = 0.5 or 0.6 (dCK) by addition of isopropyl-
beta-d-thiogalactopyranoside (IPTG) to a final concentration of
1 mM (UCK1 or UCK2) or arabinose to a final concentration of 0.02%
(dCK). The bacteria were cultured for 4 h at 22 ◦C (UCK1 and UCK2)
or 37 ◦C (dCK), then harvested by centrifugation at 5000 × g for
10 min at 4 ◦C. Bacteria pellets were stored at −80 ◦C until purifica-
tion.
The aim of this study was to identify the kinase or kinases
responsible for the first step of 2ꢀ-MeC phosphorylation, the
conversion of 2ꢀ-MeC to the 5ꢀ-monophosphate form of 2ꢀ-MeC
(2ꢀ-MeCMP). This first phosphorylation step is inefficient as sug-
gested by recent data (Cretton-Scott et al., 2008; Standring et
al., 2008). The resulting low levels of 2ꢀ-MeCMP may restrict the
subsequent intracellular production of 2ꢀ-MeC triphosphate (2ꢀ-
MeCTP), which in turn limits the antiviral potency of this agent
in humans.
The kinases most likely to contribute to the initial phospho-
rylation of 2ꢀ-MeC include two uridine–cytidine kinases (UCKs),
UCK1 and UCK2, that participate in the first and rate-limiting
step of the pyrimidine-nucleotide salvage pathway in mam-
malian cells (Anderson and Brockman, 1964; Van Rompay et al.,
2003), and deoxycytidine kinase (dCK), which also participates
in the pyrimidine salvage pathway, primarily by phosphorylating
deoxynucleotides (Van Rompay et al., 2003). In addition to their
natural substrates, UCK and dCK kinases have been shown to phos-
phorylate a variety of nucleoside analogs (Beauséjour et al., 2002;
Van Rompay et al., 2001). UCK1 and UCK2 have been shown to
phosphorylate a range of uridine and cytidine analogs but are not
thought to phosphorylate deoxyribonucleosides or purine ribonu-
cleosides. Despite being closely related at the amino acid sequence
level (∼72% similarity), the UCK enzymes differ in their substrate
specificities (Van Rompay et al., 2001). In addition to its natural
substrates deoxycytidine, deoxyadenosine, and deoxyguanosine,
dCK has been shown to phosphorylate a broad range of nucleo-
side analogs (Johansson et al., 1997). There is a controversy in the
literature around the enzyme responsible for the phosphorylation
of 2ꢀ-MeC. Klumpp et al. (2008) reported that 2ꢀ-MeC is a substrate
for UCK1, not for dCK and UCK2. In contrast Murakami et al. (2007)
showed that UCK1 does not phosphorylate 2ꢀ-MeC, but found that
dCK phosphorylates cytidine and 2ꢀ-MeC in vitro, albeit with low
efficiency.
2.3. Protein purification
Fifteen milliliters of lysis solution (20 mM Tris–HCl, pH 7.8 for
UCK1 or pH 7.4 for UCK2 and dCK, 500 mM NaCl, 40 mM imidazole,
1× BugBuster solution [Novagen, Madison, WI], 1× ethylenedi-
aminetetraacetic acid [EDTA] free complete protease inhibitor
[Roche Applied Sciences, Indianapolis, IL], 10% glycerol, 37.5 U/mL
benzonase, 2 mM dithiothreitol [DTT]) were added per gram of
wet cell paste. The resuspended cell paste was incubated with
10,000 U/mL of lysozyme solution (EMD Chemicals, Gibbstown, NJ)
for 20 min at room temperature (RT) with gentle shaking. The lysate
was clarified at 19,000 × g for 1 h at 4 ◦C.
In the present study, we used two different strategies to address
this issue and investigate the contribution of UCK1, UCK2, and
dCK to the in vitro phosphorylation of 2ꢀ-MeC. The first evaluated
the phosphorylation of 2ꢀ-MeC by purified human recombinant
kinases, the second examined the effect of depleting the expression
of each kinase in cell culture with a gene-specific siRNA. We found
that both UCK2 and dCK could phosphorylate 2ꢀ-MeC in vitro, but
in cell culture experiments, UCK2 was the main kinase responsible
for the phosphorylation of 2ꢀ-MeC.
All purifications were performed on an AKTA Purifier (GE
Healthcare, Piscataway, NJ). For UCK1 and UCK2, clarified lysate
was loaded on a 1 mL HisTrap HP column (GE Healthcare, Piscat-
away, NJ) equilibrated with 10 column volumes (CV) of buffer A
(20 mM Tris–HCl, pH 7.8 for UCK1 or pH 7.4 for UCK2, 500 mM NaCl,
10% glycerol, 2 mM DTT, 0.4 mM Pefabloc SC [Roche Applied Sci-
ences, Indianapolis, IL], 0.1% Triton X-100 [Calbiochem, San Diego,
CA]) substituted with 8% buffer B (same as buffer A without Triton
X-100 and with 500 mM imidazole at 0.5 mL/min). For dCK, lysate
was loaded on a 5 mL HisTrap HP column equilibrated with the
buffer A at pH 7.4 substituted with 1.0% buffer B at 2.0 mL/min.
Unbound proteins were removed by 20 CV of 8% Buffer B. Two
step gradient washes with 10% and 20% buffer B, 20 CV each,
removed non-specifically bound E. coli proteins. His-tagged pro-
teins were eluted by a linear gradient (20–100% buffer B, 20 CV).
One milliliter (UCK1 and UCK2) or 5 mL (dCK) fractions were col-
lected and analyzed by sodium dodecyl sulfate polyacrylamide
gel electrophoresis (SDS PAGE) followed by staining with InVision
His-Tag In-gel (Invitrogen, Carlsbad, CA) and GelCode Blue Stain
Reagent (Pierce Biotechnology, Rockford, IL). Fractions containing
2. Materials and methods
2.1. Nucleoside and nucleoside analogs
2ꢀ-C-methyl-cytidine (2ꢀ-MeC), ␤-l-2ꢀ-deoxycytidine (LdC), ␤-
d-2ꢀ-deoxy-2ꢀ-fluoro-2ꢀ-C-methyl-cytidine (2ꢀ-F-2ꢀ-MeC) and their
corresponding triphosphates were synthesized at Idenix Pharma-
ceuticals, Inc. The remaining ribonucleosides, deoxynucleosides
and nucleoside analogs, 2ꢀ,3ꢀ-dideoxycytidine (ddC), cytosine ␤-
d-arabinofuranoside (AraC), 5-iodouridine (5-IodoU), 6-azauridine
(6-AzaU), 5-fluorocytidine (5-FC) and 2-chloro-2ꢀ-deoxyadenosine
(CdA) were obtained from Sigma–Aldrich (St. Louis, MO) or
ICN Biomedicals (Costa Mesa, CA). [6-3H]-2ꢀ-dC (9 Ci/mmol),
[5-3H(N)]-C (24.8 Ci/mmol), [5,6-3H]-U (34.5 Ci/mmol), 2ꢀ-MeC
(11.7 Ci/mmol) and 2ꢀ-F-2ꢀ-MeC (8.5 Ci/mmol), were obtained from
Moravek Biochemicals (Brea, CA). The structures of these com-
pounds are shown in Fig. 1.

472
N.L. Golitsina et al. / Antiviral Research 85 (2010) 470–481
Fig. 1. Nucleoside and nucleoside analogs structures.
pure His-tagged kinase were combined and buffer exchanged on
PD10 columns (GE Healthcare, Piscataway, NJ) equilibrated at 4 ◦C
with 50 mM Tris pH 7.8, 100 mM NaCl, 20% glycerol, 10 mM DTT
and 0.4 mM pefabloc. After clarification for 20 min at 16,000 × g
at 4 ◦C, protein concentration was measured using Coomassie Plus
(Pierce Biotechnology, Rockford, IL). Bovine serum albumin (BSA)
(New England BioLabs, Beverly, MA), was added to a concentra-
tion of 0.5 mg/mL and proteins aliquots were stored at −80 ◦C. The
purified proteins were >96% pure as evaluated by SDS PAGE.
The presence of phosphatases in the enzyme preparations was
excluded by comparison of the enzyme activity with and without
phosphatase inhibitor NaF (5 mM). No changes in the activity were
found (data not shown).
2.4. Enzyme phosphorylation assay
Phosphorylation assays were performed using fresh aliquots
of recombinant UCK1, UCK2 or dCK proteins. The test nucleo-
side substrates were added at a final concentration of 1 mM in a
50 ␮L reaction mixture containing 50 mM Tris–HCl, pH 7.6, 5 mM
MgCl2, 4 mM unlabeled ATP, 1 ␮Ci of [␥-33P] ATP (3000 Ci/mmol,
Amersham Biosciences, Piscataway, NJ), 0.5 mg/mL albumin and
5 ng of enzyme. The reaction mixtures were incubated 30 min
at 37 ◦C, heated at 100 ◦C for 2 min and centrifuged for 5 min at
16,000 × g. Two microliters of the reaction mixtures were spotted
on poly(ethylenimine)-cellulose F chromatography sheets (EMD
Chemicals, Madison, WI), and the nucleoside monophosphate
products were separated by thin-layer chromatography (TLC) in a
buffer containing either 0.5 M LiCl or NH4OH/isobuturic acid/water
(1:66:33, v/v). TLC sheets were autoradiographed on a Storm
860 PhosphorImager (Amersham Biosciences, Piscataway, NJ) and
quantified using ImageQuant TL v2003.03 (Amersham Biosciences,
The presence of deaminases in the purified enzymes was
excluded by HPLC analysis of 3H-2ꢀ-MeC after 1 h incubation at
37 ◦C with dCK, UCK1 or UCK2 in the presence of 2 mM Mg-ATP.
HPLC analysis did not reveal the appearance of any 2ꢀ-MeU deriva-
tives of NM107.

N.L. Golitsina et al. / Antiviral Research 85 (2010) 470–481
473
Piscataway, NJ). Products were identified using monophosphate
2.7. Western blotting
reference standards.
Steady-state activity parameters for dCK, UCK1 and UCK2 were
measured using a radiochemical method (Ives et al., 1969; Ahmed,
1981) with ATP as a phosphate donor. The assays were performed
in a total reaction volume of 0.05 mL in 100 mM Tris–HCl, pH 7.5,
100 mM KCl, 2 mM MgCl2, 2 mM ATP, 5 mM DTT, 0.5 mg/mL BSA
and 1 ␮M [6-3H]-2ꢀ-dC (Ci/mmol), or [5-3H(N)]-C (24.8 Ci/mmol),
or [5,6-3H]-U (34.5 Ci/mmol), with or without 5 mM NaF. Mg-ATP
was increased to 5 mM when 3H-2ꢀ-MeC (8.7 Ci/mmol) and 3H-
2ꢀ-F-2ꢀ-MeC (8.5 Ci/mmol) were used. All 3H-labeled substrates
were obtained from Moravek Biochemicals (Brea, CA). The amount
of enzyme added to the reaction was adjusted so that no more
than 20% of the substrate was consumed. After incubation at
37 ◦C samples were heated at 100 ◦C for 2 min, placed in ice for
1 min and centrifuged for 5 min at 16,000 × g to remove dena-
tured protein. An aliquot of 25 ␮L from each sample were diluted
in 200 ␮L of water to decrease salt concentration in samples and
50 ␮L of diluted samples were spotted in triplicate on Whatman
DE-81 filters, dried, washed by 1 mM phosphate buffer pH 7.0 as
described by Ahmed (1981) and dried again. For loading control
10 ␮L of each sample was spotted on two filters and dried. DE-
81 discs were placed in scintillation vials, eluted with 1 mL 0.2 M
NaCl/0.1 M HCl solution for 1 h and were counted in 18 mL of scin-
tillation fluid. Steady-state parameters were determined by using
SigmaPlot 2004 Enzyme Kinetics software and Michaelis–Menten
equation.
Huh7, HepG2 or HeLa cells were washed twice with 1 mL cold
phosphate buffered saline (PBS). Two hundred–300 ␮L of ice-cold
Reportasol (Novogen, Madison, WI) lysis buffer was added directly
to each well. Lysates were collected after 10 min incubation at RT,
clarified 10 min at 16,000 × g, and total protein concentrations were
measured with the BCA assay kit (Pierce, Fisher Scientific, Rockford,
IL).
Ten micrograms of cell lysate was loaded per lane on a 4–12%
gradient NuPage gel (Invitrogen, Carlsbad, CA), run with MES
buffer according the manufactures protocol and transferred to a
polyvinylidene fluoride (PVDF) membrane (Invitrogen, Carlsbad,
CA). The PVDF membrane was stained for 5 min in Ponceau S solu-
tion (Mallinkrodt, Hazelwood, MO), de-stained with water until
background was gone, and photographed. The membrane was
washed in Tris-buffered saline with 0.05% Tween-20 (TBST) at RT
until all stain was removed, and placed in 5% milk/1× TBS block-
ing buffer for at least 1 h at RT. The membrane was then incubated
with primary antibody diluted in TBST/1% milk overnight at 4 ◦C.
The blot was washed three times in TBS or TBST and incubated
with either goat anti-rabbit-HRP antibody (BioRad, Hercules, CA) or
goat anti-mouse IgG-AP (BioRad, Hercules, CA) diluted in TBST and
1% milk for 1 h at RT. The blot was then washed again three times
in TBS or TBST. For horseradish peroxidase conjugated antibody,
the enhanced chemoluminescence Western blotting detection kit
(GE Healthcare, Piscataway, NJ) was used for detection. For alkaline
phosphatase (AP) conjugated antibodies, the AP conjugate sub-
strate kit (BioRad, Hercules, CA) was used for detection.
2.5. siRNA transfection
Twelve-well plates were seeded with 1 × 105 cells per well
(HeLa and Huh7 cells) or 2 × 105 cells per well (HepG2 cells) in 1 mL
of appropriate cell medium. When cells reached ≥50% confluence,
the medium was replaced by 800 ␮L antibiotic-free medium. Trans-
fection agent LipofectamineTM (LP2000, Invitrogen, Carlsbad, CA)
was diluted 100× in Opti-MEM medium (Invitrogen, Carlsbad, CA)
and incubated 20 min at RT. The siRNAs were diluted in Opti-MEM
medium to 1 pmol/␮L, and 100 ␮L was mixed with 100 ␮L diluted
LP2000. A designated non-targeting siRNA (D001210, Dharmacon,
Fisher Scientific, Rockford, IL) provided a control for non-specific
siRNA transfection effects on gene expression. For UCK1 silenc-
ing, siRNAs HSS130222, HSS130223 and HSS130224 (Invitrogen,
Carlsbad, CA) and siRNA ID 1197 (Ambion, Austin, TX) were used.
For UCK2 silencing, siRNAs A (ID 893), B (ID 894), and C (ID 895)
(Ambion, Austin, TX) were tested. For dCK silencing, the test siR-
NAs were ID 68, ID 69, and ID 70 (Ambion, Austin, TX). After 1 h
incubation at RT, 200 ␮L of the siRNA/transfection agent mixture
were added to each well. The total amount of test siRNA was
100 pmol/well. Plates were incubated at 37 ◦C/5% CO2. Each experi-
ment was done at least in triplicate. Protein levels were analyzed by
Western blot, as described below, after 24, 48 and 72 h treatment
with siRNA.
2.8. Determination of the dCK activity in cell lysates
This was performed according to Arner et al. (1992) on aliquots
of the lysates prepared as described above in Section 2.7. The pro-
tein concentration of the lysate was first adjusted to 1 ␮g/␮L, then
25 ␮L of cell lysate were added to 25 ␮L of assay mix consist-
ing of 0.2 mM CdA substrate, 4 mM ATP, 4 ␮Ci 33P-␥ATP, 4 mM
MgCl2, 200 mM NaCl, 20 mM NaF, 4 mM DTT in 100 mM Tris–HCl
pH 7.4. After 45 min at 37 ◦C, 2 ␮L of the reaction mixtures were
spotted on poly(ethylenimine)-cellulose F chromatography sheets
(EMD Chemicals, Madison, WI), and the nucleoside monophos-
phate products were separated and quantified as described in
Section 2.4. CdA monophosphate produced by incubating CdA with
purified recombinant dCK in the presence of Reportazol provided
a chromatography standard, while reactions run without the CdA
substrate provided background controls for the TLC. The CdA phos-
phorylation in lysates from cells treated by siRNA was evaluated
as a % of the CdA phosphorylation seen in lysates from untreated
control cells.
2.9. HPLC determination of phosphorylation
Some 48 h after addition of siRNA to cells, medium was removed
and drug treatment was initiated by adding 1 mL of culture medium
containing [3H]-2ꢀ-MeC (or other nucleoside) to a final concen-
tration of 10 ␮M and 1000 dpm/pmol. Drug stock solutions were
made up freshly in antibiotic-free assay medium. The final dimethyl
sulfoxide concentration was 0.5%. Plates were gently mixed by
rocking back and forth and incubated at 37 ◦C/5% CO2 for 24 h. After
removal of medium, cells were washed three times with cold PBS
and extracted with 600 ␮L 70% methanol (at −20 ◦C) for at least
24 h at −20 ◦C. The cell extracts then were harvested, centrifuged
to remove cellular debris and dried down under a gentle stream of
filtered air. The dried cell extracts were resuspended in 200 ␮L of
water and analyzed by reverse-phase with ion-pairing chromatog-
raphy using an Agilent model 1100 instrument (Agilent, Santa Clara,
2.6. Antibodies
UCK1 and UCK2 antisera were generated by Capralogics (Hard-
wick, MA) (Shimamoto et al., 2002). Briefly, peptides encoding
amino acids 259–277 of UCK1 or amino acids 243–261 of UCK2
were separately conjugated to keyhole limpet hemocyanin via a
cysteine and injected into New Zealand White rabbits (two rab-
bits each). Antibodies were purified by standard procedures. For
Western blotting, the anti-UCK1 antibody was diluted 1:1000,
the anti-UCK2 antibody was diluted 1:5000, and the anti-dCK
antibody was diluted 1:500. Glyceraldehyde-3-phosphate dehy-
drogenase (GAPDH) was detected using a mouse monoclonal
antibody (Ambion, Austin, TX) at 1 ␮g/mL.

474
N.L. Golitsina et al. / Antiviral Research 85 (2010) 470–481
CA). The mobile phase consisted of buffer A (25 mM potassium
phosphate, 5 mM tetrabutyl ammonium phosphate (TBAP), pH 6.0)
and buffer B (methanol). A multi-stage linear gradient from 0% to
30% buffer B was run from 0 to 60 min according to the follow-
ing schedule: 10 min – 0% buffer B; 20 min – 20% buffer B; 30 min
20% buffer B; 35 min – 30% buffer B. Radioactivity was analyzed by
online detection using a radiomatic 500TR Flow Scintillation Ana-
lyzer (Packard/Perkin Elmer Life and Analytical Sciences, Waltham,
MA). Each analyte was quantified based on the area of its respective
radioactive peak and identified based on its peak elution pattern
compared to the elution patterns of cytidine, uridine and their cor-
responding 5ꢀ-phosphate derivatives. Identity was then confirmed
by co-elution of the radiolabeled peaks with authentic reference
standards. Finally the identity of monophosphate products was
further established by digestion of cell extracts with 1–5 units
of alkaline phosphatase for 60 min at 37 ◦C, thereby releasing the
parent nucleoside (e.g. 2ꢀ-MeC) which was identified as described
above.
2.10. Design of siRNA testing strategy and experimental controls
The optimal concentrations of siRNA reagents and the use of
single versus multiple siRNAs were evaluated in exploratory exper-
iments. To ensure that all observed silencing was due to a specific
effect of the siRNA on its target gene, the negative controls used
in each experiment included a transfection reagent that lacked
siRNA as well as a commercially available non-targeting siRNA con-
trol. Moreover, each gene-specific siRNA reagent served as negative
controls for the other two kinase genes.
The effect of the siRNAs on the expression of the specific target
kinase was determined in transfected cells by measuring the disap-
pearance of UCK1, UCK2 or dCK proteins by Western blotting. New
antibodies specific for each enzyme were developed in conjunc-
tion with this study (for details see Section 2.6). These antibodies
allowed the specific identification and assessment of UCK1, UCK2
and dCK levels in lysates from each cell line. Equivalent loading of
lysates for each transfection experiment was verified by Western
blot staining of GAPDH with GAPDH-specific antibodies.
Fig. 2. SDS PAGE of purified recombinant dCK, UCK1 and UCK2 enzymes.
M = proteins molecular weight markers.
MeC nucleoside analogs in vitro. The structures of the various
nucleosides and nucleoside analogs tested in this report are shown
in Fig. 1. His-tagged human UCK1, UCK2 and dCK enzymes were
each expressed in E. coli and purified to greater than 96% homo-
geneity as shown in Fig. 2.
The steady-state catalytic parameters for recombinant UCK1,
UCK2 and dCK, tested in standard DE-81 filter assay with 3H-labeled
compounds as described in Section 2, are summarized in Table 1.
The purified enzymes had Km and specific activity values similar
to those published by other researchers (Van Rompay et al., 2001;
Roy et al., 2004; Usova et al., 2004; Murakami et al., 2007).
Purified His-tagged recombinant dCK efficiently phosphory-
lated dC (Table 1) with a Km of 1.5 0.5 ␮M, within the range of
values reported in the literature (0.13–9 ␮M) (Shafiee et al., 1998;
Sabini et al., 2003; Usova et al., 2004; Roy et al., 2004; Murakami
et al., 2007). Consistent with literature reports dCK phosphory-
lated a wide range of cytidine analogs: AraC, ddC and 5-FC (data
not shown). The nucleoside monophosphate products of in vitro
phosphorylation reactions were resolved by TLC and their identity
was verified using appropriate reference standards and evaluated
as described in Section 2. Both 2ꢀ-F-2ꢀ-MeC and 2ꢀ-MeC were phos-
phorylated by dCK although with low efficiency: 118 and 438 times
less than dC, respectively (Table 1).
Finally, the impact of each kinase on 2ꢀ-MeC or 2ꢀ-F-2ꢀ-MeC
phosphorylation was determined in conjunction with the silencing
experiments by HPLC/radiometric quantification of total phos-
phorylated nucleoside analog species (appearing in cell lysates
primarily as the respective triphosphates) recovered from control
cells versus silenced cells following exposure to [3H]-labeled drug.
The relevant mono-, di- and triphosphate species were identified
using synthetic reference standards (see Section 2.9).
3. Results
These data are generally consistent with the findings of
Murakami et al. (2007) suggesting that 2ꢀ-F-2ꢀ-MeC and 2ꢀ-MeC
are potential, albeit less efficient, substrates for the human dCK
enzyme. Our data do not appear to support the finding of Klumpp
et al. (2008).
3.1. Phosphorylation of 2ꢀ-MeC by recombinant enzymes in vitro
Two different groups have obtained conflicting results related
to enzymes that can phosphorylate 2ꢀ-MeC to 2ꢀ-MeCMP. Klumpp
et al. (2008) reported that dCK and UCK2 do not phosphorylate 2ꢀ-
MeC and that 2ꢀ-MeC is a substrate for UCK1. Murakami et al. (2007),
on the other hand, showed that UCK1 does not phosphorylate 2ꢀ-
MeC. Instead they found that 2ꢀ-MeC can be phosphorylated by dCK
although with low efficiency.
Our paper sought to resolve this controversy using two differ-
ent approaches: by testing 2ꢀ-MeC phosphorylation with purified
kinases and by evaluating the impact of silencing the expression of
the specific kinase on 2ꢀ-MeC phosphorylation in cell cultures using
siRNA technique.
Purified His-tagged recombinant UCK1 efficiently phosphory-
lated C and U (Table 1) with a Km of 194 127 ␮M and Km of
452 37 ␮M for C and U, respectively, consistent with the range
of values reported in the literature (131–291 ␮M and 225–407 ␮M
for C and U, respectively) (Van Rompay et al., 2001; Roy et al.,
2004; Murakami et al., 2007). Consistent with a prior report (Van
Rompay et al., 2001), purified recombinant UCK1 displayed a rela-
tively restricted substrate profile; it effectively phosphorylated its
natural substrates, uridine and cytidine (Table 1), as well as the
nucleoside analog 5-IodoU and was less active with 6-AzaU and
5-FC (data not shown). UCK1 did not phosphorylate 2ꢀ-MeC and 2ꢀ-
F-2ꢀ-MeC tested at up to 5 mM (Table 1) and LdC (data not shown)
in the standard assay. Indeed, no trace of 2ꢀ-MeC or 2ꢀ-F-2ꢀ-MeC
We first evaluated whether the purified recombinant UCK1,
UCK2, or dCK enzymes could phosphorylate 2ꢀ-MeC and 2ꢀ-F-2ꢀ-

N.L. Golitsina et al. / Antiviral Research 85 (2010) 470–481
475
Table 1
Steady-state kinetic parameters for dCK, UCK1 and UCK2 with ATP as a phosphate donor. Kcat was calculated using molecular mass of 33.16, 33.707 and 31.572 kDa for
recombinant dCK, UCK1 and UCK2 subunits, respectively. Values are reported as means standard deviations of four experiments for dC, C or U, eight experiments for 2ꢀ-MeC
and two experiments for 2ꢀ-F-2ꢀ-MeC.
Compound
Km (␮M)
Vmax (nmol/min/mg)
kcat (s−1
)
kcat/Km (s−1 M−1
)
Relative efficiency
Enzyme
dC
1.5 0.5
451 225
230 74
125 30
87 33
162 89
0.07
0.048
0.09
46296
107
391
1
2ꢀ-MeC
2ꢀ-F-2ꢀ-MeC
0.0023
0.0084
dCK
C
U
194 127
452 37
NA, 5 mM
NA, 5 mM
544 377
1270 159
NA, 5 mM
NA, 5 mM
0.31
0.71
1598
1578
UCK1
UCK2
2ꢀ-MeC
2ꢀ-F-2ꢀ-MeC
C
U
72 21
121 27
3131 1718
NA, 5 mM
1975 500
49030 12700
837 496
1.03
1.41
0.44
14306
22511
139.8
1
1.6
0.0098
2ꢀ-MeC
2ꢀ-F-2ꢀ-MeC
NA, 5 mM
monophosphates was seen even after prolonged incubation (data
not shown) indicating that UCK1 does not utilize 2ꢀ-MeC and 2ꢀ-F-
2ꢀ-MeC as a substrates. Our results are in a good agreement with
Murakami et al. (2007), and contradict those of Klumpp et al. (2008).
Our purified recombinant UCK2 efficiently phosphorylated C
and U (Table 1) with Km values of 72 21 ␮M and 121 27 ␮M for
C and U, respectively, close to the values reported by Van Rompay
et al. (2001) (86 and 50 ␮M for C and U, respectively) and exhib-
ited a much wider substrate spectrum than UCK1 in accord with
the findings of Van Rompay et al. (2001). Thus, UCK2 phosphory-
lated the analogs 5-HydroxyU, 5-FC and 6-AzaU (data not shown).
UCK2 showed no activity against 2ꢀ-F-2ꢀ-MeC (Table 1). Contrary to
Klumpp et al. we found that 2ꢀ-MeC is a weak substrate for UCK2
with a Km of 3.1 1.7 mM (Table 1). 2ꢀ-MeC was 100× less effective
as a substrate in terms of catalytic efficiency (kcat/Km) than C for
UCK2 compared to 400× less effective than dC for dCK (Table 1).
In summary, our in vitro biochemical data suggest that the likely
candidate enzymes for phosphorylation of 2ꢀ-MeC are UCK2 and
dCK, and effectively rule out the involvement of UCK1.
Each kinase gene was silenced without significantly affecting
the other two genes using gene-specific siRNA reagents. The effect
of the siRNAs on the expression of the specific target kinase was
determined in transfected cells by measuring the disappearance of
UCK1, UCK2 or dCK proteins by Western blotting.
Finally, the impact of each kinase on 2ꢀ-MeC or 2ꢀ-F-2ꢀ-MeC
phosphorylation was determined in conjunction with the silencing
experiments by HPLC/radiometric quantification of total phos-
phorylated nucleoside analog species (primarily the respective
triphosphates) recovered from control cells versus silenced cells
following exposure to [3H]-labeled drug. The relevant mono-, di-
and triphosphate species were identified using synthetic reference
standards (see Section 2).
In summary, the use of different siRNA reagents and antibodies,
together with the evaluation of phosphorylation of the nucleosides,
provided an extensive panel of controls to ensure the correct inter-
pretation of all results.
3.3. Effect of siRNAs on kinase expression
In exploratory experiments, conditions for the most efficient
siRNA delivery and gene silencing were tested and optimized in
each cell line by measuring the decreased expression of target
kinase enzyme by Western blotting. The conditions investigated
included cell density, the use of different transfection reagents, the
use of different siRNA sequences, the use of single versus combi-
nations of gene-specific siRNAs at different ratios, and finally the
time required to achieve a maximum reduction in kinase levels fol-
lowing transfection of the siRNA. Some of these data are shown in
Figs. 3 and 4.
As seen in Fig. 3A, immunostaining with a UCK2-specific anti-
body reveals that the amount of UCK2 decreased substantially
in HeLa cells in response to transfection with UCK2-specific siR-
NAs but not with a control (random) siRNA. Immunostaining
with a GAPDH antibody verified that equal amounts of lysates
were loaded on this gel, demonstrating that the decrease in UCK2
expression was gene specific. Testing of three different siRNAs
singly or in combination, as described in Section 2 (data not
shown), established that maximal UCK2 suppression was attained
with a 1:1 combination of two siRNAs. As shown in Fig. 3A,
the combination of siRNAs B and C, targeting exon 4 (B) and
exon 5 (C) of the UCK2 gene, appeared to be slightly superior
to the combination of siRNAs A and C and was thus selected
for the remaining experiments. Essentially maximal UCK2 pro-
tein depletion was achieved by 24 h post-transfection (Fig. 4A) and
maintained through 72 h post-transfection (data not shown). Since
all three kinases were effectively depleted 48 h post-transfection in
response to their respective siRNA reagents (see below), this time
frame was routinely used when comparison between the proteins
3.2. siRNA analysis of the kinases involved in the phosphorylation
of 2ꢀ-MeC in human cell lines
Since the relative contribution of UCK2 and dCK to phosphoryla-
tion of 2ꢀ-MeC in the cellular environment likely depends on factors
such as the expression level of these enzymes and the competition
between 2ꢀ-MeC and the natural substrates, we decided to further
evaluate the contribution of each of these enzymes in human cell
lines. To this end, we used a siRNA approach to specifically target
each of the UCK1, UCK2 and dCK kinases. The effect of depleting
each enzyme on the extent of phosphorylation of 2ꢀ-MeC in cells
was used to determine the role of each kinase in the formation of
phosphorylated 2ꢀ-MeC species.
The inclusion of appropriate controls was an important aspect of
the experimental design for the siRNA study (see Section 2). A com-
bination of positive and negative controls was included in every
experiment to rule out any misinterpretation of the experimental
results.
All findings were verified in three different human cell lines:
HeLa cells and two human hepatocyte-derived cell lines (HepG2
hepatocellular liver carcinoma and Huh7 hepatoma cells). Of note,
the Huh7 cell line is permissive for the replication of the HCV repli-
con and is the cell line typically used for the preclinical assessment
of HCV antiviral agents. We also note that we attempted to study
the kinases in primary human hepatocytes. These experiments did
not yield useful information since we were unable to detect UCK2
or dCK by Western blot or get siRNA approaches to work in these
cells, which are more difficult to work with than cell lines.

476
N.L. Golitsina et al. / Antiviral Research 85 (2010) 470–481
Fig. 3. The effect of incubation with different gene-specific siRNAs on the kinase levels in HeLa cells. (A) Western blot analysis performed with UCK2-specific (top panel) and
GAPDH-specific (bottom panel) antibodies. For description of reagents and siRNAs see Sections 2 and 3. (B) Western blot analysis performed with dCK-specific (top panel)
and GAPDH-specific (bottom panel) antibodies.
Fig. 4. Effects of incubation time on the depletion of UCK1, UCK2 and dCK in HeLa cells by each gene-specific siRNA. Western blot were stained with antibodies recognizing
(A) UCK1 and UCK2, (B) dCK, and (C) GAPDH.

N.L. Golitsina et al. / Antiviral Research 85 (2010) 470–481
477
Fig. 5. Silencing of UCK1 and UCK2 expression in HepG2 cells. Western blot immunostained for UCK1 (A), UCK2 (B) and GAPDH (C).
was needed (e.g. for the phosphorylation experiments described
below).
siRNA, HSS130223 (data not shown). In HeLa cells (Fig. 4A), this
siRNA effectively depleted UCK1 protein but not UCK2 protein, as
judged by staining with our anti-UCK1 antibody that recognizes
both proteins (which are closely related at the amino acid sequence
level). Immunostaining for GAPDH again verified the loading of
equal amounts of lysates on the gel (Fig. 4C). The time course for
UCK1 (Fig. 4A) showed that 48 rather than 24 h was required to
achieve effective silencing for this protein. Fig. 5 shows the results
obtained in HepG2 cells and compares the staining obtained with
the antibody raised against UCK1 protein (Panel A) versus the anti-
UCK2 antibody that is specific for UCK2 (Panel B). Both antibodies
confirmed that the UCK2-specific siRNA depleted UCK2 protein.
Immunostaining with UCK1 antibody demonstrated that HepG2
cells have higher amounts of UCK1 compared to UCK2 than HeLa
(Fig. 4A) and Huh7 cells (Fig. 6). The data further confirmed that
UCK2 knockdown did not significantly deplete UCK1. Conversely,
only the UCK1-specific siRNA depleted UCK1 protein levels. GAPDH
For dCK, the Western blot analysis revealed that the amount of
kinase protein in HeLa cells was not affected by the control siRNA
but decreased significantly in response to transfection with each
of the three individual test siRNAs, ID68, ID69 and ID70 (Fig. 3B).
Equal protein loading was again verified by GAPDH immunostain-
ing (Fig. 3B). The most efficient suppression was seen with 1:1
combinations of two siRNAs, 68+70 or 69+70, (Fig. 3B). The com-
bination of siRNAs 69+70, targeting exons 2–3 (69) and exon 7
(70) of the dCK gene, gave the most effective silencing and was
selected for further studies. Silencing of dCK was evident (Fig. 4B)
through 72 h post-transfection and there was good suppression of
dCK protein levels at 48 h post-transfection, the time point chosen
for subsequent experiments.
For UCK1 silencing, four siRNAs were evaluated singly and in
combination and the best results were obtained with a single
Fig. 6. Gene-specific knockdown of kinase proteins in Huh7 cells. Western blot analysis using siRNA and antibody reagents as indicated on the figure.

478
N.L. Golitsina et al. / Antiviral Research 85 (2010) 470–481
staining again verified the loading of equal amounts of lysates on
the gel (Fig. 5C).
The experiment shown in Fig. 6 clearly demonstrates that
depletion of each of the three kinase proteins could be achieved
independentof theothertwoin Huh7cellsin responsetothe appro-
priate siRNA reagent. Similar results in terms of the extent and
specificity of silencing were obtained in all three cell lines (data
not shown). Importantly, kinase depletion was only observed in
response to the correct gene-specific siRNA in all three cell lines
and was not seen in cells treated with transfection agent alone
or transfected with the non-targeting random siRNA as seen in
Figs. 3–6.
3.4. Phosphorylation of 2ꢀ-MeC in cells after specific siRNA
transfection
Next, we determined how reductions in the levels of UCK1,
UCK2, and dCK proteins affected the ability of human cell lines to
phosphorylate 2ꢀ-MeC and 2ꢀ-F-2ꢀ-MeC. Cells were transfected with
the appropriate siRNA and cultured for a further 48 h to provide
extensive depletion of the target kinase protein that was verified
by parallel Western blot experiments. The depleted cells were then
incubated with the appropriate radiolabeled nucleoside analog for
24 h. The resultant phosphorylated nucleoside products (primar-
ily in the form of triphosphate species as described further below)
were then separated by HPLC and quantified (see Section 2 for
further details).
The results of these experiments in the three cell lines studied
are summarized in Figs. 7 and 8.
Compared to control cells, depletion of UCK2 protein reduced 2ꢀ-
MeC phosphorylation by about 5-fold in HeLa (Fig. 7A) and Huh7
(Fig. 8A) cells, and 3-fold in HepG2 cells (Fig. 7B). In contrast, UCK2
depletion did not reduce the phosphorylation of 2ꢀ-F-2ꢀ-MeC in
these cell lines (Figs. 7C and 8B).
Fig. 7. Effects of siRNA treatment on the phosphorylation of 2ꢀ-MeC in HeLa (A) and
HepG2 (B) cells and 2ꢀ-F-2ꢀ-MeC in HepG2 cells (C). Data reflect means and standard
deviations obtained from five (A) or two (B, C) independent experiments.
Next the impact of UCK1 depletion was determined. For 2ꢀ-MeC,
UCK1 depletion had little or no effect in HeLa (Fig. 7A), HepG2
(Fig. 7B) or Huh7 cells (Fig. 8A). Similarly, UCK1 knockdown did not
have a significant effect on 2ꢀ-F-2ꢀ-MeC phosphorylation as seen in
Figs. 7C and 8B.
For dCK, the effect of depletion on 2ꢀ-MeC phosphorylation was
found to be minimal (<20% reduction) after 48 h in Huh7 (Fig. 8A)
and HepG2 cells (data not shown); HeLa cells are less relevant to
the antiviral activity of 2ꢀ-MeC and were not tested in this assay.
In contrast, 2ꢀ-F-2ꢀ-MeC phosphorylation was reduced by ∼75% in
dCK-silenced Huh7 cells (Fig. 8B). Although this accords well with
the approximately 80% reduction of dCK protein levels estimated by
Western blot analysis, we further validated this result by measuring
the enzymatic dCK activity present in a portion of the cell lysates
using an assay based on the phosphorylation of CdA, a substrate that
is specific for dCK. Compared to control lysates, the lysates from
the silenced cells contained about 5-fold less dCK activity (data not
shown).
Taken together, the preceding results provide strong support
for the idea that, at least in cell culture, dCK is the primary enzyme
driving the initial phosphorylation of 2ꢀ-F-2ꢀ-MeC (Murakami et al.,
2007), whereas UCK2 is the key enzyme behind 2ꢀ-MeC phospho-
rylation. Less clear from the data is whether any role is left for
dCK in the intracellular phosphorylation of 2ꢀ-MeC. Since silenc-
ing is typically less than 100% efficient it is likely that much or all of
the residual 2ꢀ-MeC phosphorylation (20–40%) seen in the different
cell types simply reflects the remaining levels of UCK2. However,
given the limitations of the silencing methodology coupled with the
inherent variability of these complex cell-based assays, we cannot
completely exclude the possibility that dCK may make a minor con-
tribution to the overall phosphorylation of 2ꢀ-MeC in the tested cell
lines.
Fig. 8. Effects of siRNA treatment on the phosphorylation of 2ꢀ-MeC (A) and 2ꢀ-F-
2ꢀ-MeC (B) in Huh7 cells. Data reflect means and standard deviations obtained from
six (A) or three (B) independent experiments.

N.L. Golitsina et al. / Antiviral Research 85 (2010) 470–481
479
expected, no changein 2ꢀ-MeCTP levels was seen in cells transfected
with the other two siRNA species.
No deamination of 2ꢀ-MeC or its phosphorylated metabolites to
the corresponding uridine derivatives (e.g. 2ꢀ-MeU or 2ꢀ-MeUMP)
was seen in any cell type in this study. In our experience it takes a
longer time to generate these uridine species. Thus, low and vari-
able amounts (less then 4.5%) of 2ꢀ-MeUTP were found only in
HepG2 cells after 48 h incubation with 3H-2ꢀ-MeC (data not shown).
4. Discussion
The main goal of this study was to identify the human kinase(s)
responsible for the conversion of 2ꢀ-MeC to its 5ꢀ-monophosphate
metabolite, 2ꢀ-MeCMP. This is the first phosphorylation step in
the pathway leading to the production of the 2ꢀ-MeC triphosphate
active species that is responsible for inhibiting the HCV polymerase.
Data presented in this paper are consistent with the idea that the
first phosphorylation step for 2ꢀ-MeC is a rate-limiting step that
may restrict the antiviral activity of this agent in man. This limi-
tation, which likely applies to all the cytidine analogs currently in
clinical development for the treatment of HCV, may be overcome to
some extent by the use of high doses of nucleoside analogs, leading
to increased triphosphate production in the liver. However, high
doses of nucleoside analogs may also result in an increase in the
off-target toxicities and side effects encountered with these agents.
Two groups have reported different kinases responsible for 2ꢀ-
MeC phosphorylation. Klumpp et al. (2008) found that dCK did not
phosphorylate 2ꢀ-MeC but UCK1 did. Murakami et al. (2007) found
that UCK1 did not phosphorylate 2ꢀ-MeC whereas dCK did, although
with low efficiency.
Our study attempted to resolve this discrepancy using two
different approaches: by testing 2ꢀ-MeC phosphorylation with puri-
fied kinases and by investigating the impact of silencing the specific
kinase expression (by siRNA technique) on phosphorylation of 2ꢀ-
MeC.
We first evaluated the ability of recombinant UCK1, UCK2,
and dCK to phosphorylate 2ꢀ-MeC and several control nucleoside
analogs in vitro. While recombinant UCK2 and dCK were both able
to phosphorylate 2ꢀ-MeC in vitro, albeit with much less efficiency
than their natural substrates, UCK1 did not phosphorylate 2ꢀ-MeC
at all, consistent with reports that this enzyme has a relatively nar-
row substrate specificity (Van Rompay et al., 2001; Murakami et
al., 2007). Our biochemical results showing that dCK, but not UCK1,
can phosphorylate 2ꢀ-MeC generally agree with those of Murakami
et al. (2007) and contradict Klumpp et al. (2008). More importantly,
however, we also showed that UCK2 can catalyze this reaction. In
contrast to 2ꢀ-MeC, the closely related cytidine analog 2ꢀ-F-2ꢀ-MeC
was a substrate for dCK but not for the two UCK enzymes. This
finding confirms a prior report (Murakami et al., 2007) that 2ꢀ-F-
2ꢀ-MeC is phosphorylated by dCK and suggests that 2ꢀ-F-2ꢀ-MeC
behaves more like a deoxynucleoside than a ribonucleoside due to
the presence of the 2ꢀ-fluoro group.
While our biochemical data suggested that both UCK2 and
dCK could potentially be responsible for converting 2ꢀ-MeC to 2ꢀ-
MeCMP, a series of carefully controlled silencing experiments using
gene-specific siRNAs pinpointed UCK2 as the key enzyme respon-
sible for the monophosphorylation of 2ꢀ-MeC, at least in the three
different human cell lines tested in this study. Thus, reduction
of UCK2 levels decreased 2ꢀ-MeC phosphorylation by 70–80% in
all three human cell lines tested. In contrast, dCK appeared to be
responsible for no more than 20% of 2ꢀ-MeC phosphorylation in
Huh7 and HepG2 cells: this low value lies within the expected
range of variability for a complex cell-based assay. Given this and
the fact that our UCK2 silencing experiments almost certainly did
not achieve a 100% protein depletion, we cannot conclude with
Fig. 9. HPLC analysis detected phosphorylated 2ꢀ-MeC species in control Huh7 cells
(A); and in cells 48 h after transfection with siRNAs specific for UCK1 (B), dCK (C), or
UCK2 (D).
3.5. The effect of silencing on the phosphorylation profile of
2ꢀ-MeC in cells
The HPLC chromatogram shown in Fig. 9A depicts the profile
of 2ꢀ-MeC species metabolized in control Huh7 cell treated with
10 ␮M radioactive 2ꢀ-MeC for 24 h (Fig. 9A). Also shown is the
impact of silencing with siRNAs specific for UCK1 (Fig. 9B), dCK
(Fig. 9C) or UCK2 (Fig. 9D). Similar results were obtained in HepG2
cells (data not shown).
As can be seen from the chromatographic profiles, exposure
of cells to radioactive 2ꢀ-MeC resulted mainly in the formation
of the corresponding triphosphate product. Only a low amount
of 2ꢀ-MeCDP and no 2ꢀ-MeCMP was found under these condi-
tions. The predominance of 2ꢀ-MeCTP product coupled with the
lack of accumulation of the MP and DP intermediates is consis-
tent with the idea that the 2ꢀ-MeCMP intermediate, once formed,
is rapidly and efficiently converted to 2ꢀ-MeCTP. Thus, the first 2ꢀ-
MeC phosphorylation step appears to be the rate-limiting step in
2ꢀ-MeCTP production, at least in the cell systems studied in this
report.
A significant decrease in the amount of 2ꢀ-MeCTP end prod-
uct was seen only in cells transected with UCK2-specific siRNA
(Fig. 9D), indicating that a decrease in UCK2-mediated production
2ꢀ-MeCMP leads directly to a decrease in the TP active species. As

480
N.L. Golitsina et al. / Antiviral Research 85 (2010) 470–481
certainty that dCK has any role in the in vitro phosphorylation of
2ꢀ-MeC in cells. However, we do not exclude a minor role for this
enzyme.
ular the use of monophosphorylated nucleotide prodrugs provides
a potential approach to bypass the first kinase step, thereby lead-
ing to increased levels of the triphosphate active species and hence
potentially increased antiviral activity in the clinic compared to a
corresponding nucleoside analog.
In other respects, the siRNA results correlated well with those
obtained with the recombinant kinases and further validated the
biochemical findings. Thus, siRNA silencing ruled out a role for the
UCK1 kinase in the phosphorylation of 2ꢀ-MeC in the three human
cells lines. In addition to confirming the work of Murakami et al.
(2007) by showing that phosphorylation of 2ꢀ-F-2ꢀ-MeC is mediated
by dCK, our silencing experiments further rule out the involvement
of UCK1 or UCK2 in this process. Importantly, these results clearly
demonstrate that dCK is functional in the test cell types, thereby
establishing that the phosphorylation of 2ꢀ-MeC by UCK2 rather
than dCK is not simply due to a lack of a functional dCK kinase in
these cells.
Acknowledgments
We thank Jesse Potash (ProSanos Corporation) and Valérie
Philippon (former employee of Idenix Pharmaceuticals) for assis-
tance with manuscript preparation. We also thank our colleagues
John Bilello, Ben Mayes and Joseph McCarville for helpful com-
ments.
The idea that
a
UCK enzyme is responsible for 2ꢀ-MeC
References
phosphorylation in cultured cells was first demonstrated by
substrate competition studies showing that the intracellu-
lar phosphorylation of 2ꢀ-MeC to 2ꢀ-MeC triphosphate was
reduced in a dose-dependent fashion in the presence of cyti-
dine (EC50 = 19.17 4.67 ␮M) or uridine (EC50 = 20.92 7.10 ␮M),
but not deoxycytidine (EC50 > 100 ␮M) (E. Cretton-Scott, data not
shown). The present work now extends these in vitro competi-
tion experiments by demonstrating that UCK2, and not UCK1, is
the enzyme responsible for the initial phosphorylation of 2ꢀ-MeC
in cell culture.
Our study demonstrates the importance of using cell-based
systems to confirm results obtained on the phosphorylation of
nucleoside analogs by purified enzyme. As noted above, dCK was
shown to be fully functional in the cultured test cells but appeared
to play at best a minor role or perhaps even no role in the phos-
phorylation of 2ꢀ-MeC in cells. While the basis for the differences
in results between the two biochemical and cell-based systems
remains unknown, a variety of factors, such as intracellular enzyme
levels or activity, competition with the natural substrates, the exis-
tence of feedback mechanisms, the presence of specific natural
products within the cell, or the possibility that specific kinases
may function in multi-protein enzymatic complexes may lead to
differences between the two systems.
Our findings define UCK2 as the key kinase involved in the for-
mation of 2ꢀ-MeC monophosphate from 2ꢀ-MeC in cultured human
cells. Since the HCV replicon-containing Huh7 cell line is widely
used in HCV drug discovery and has been successfully used to pre-
dict which nucleosides will be active in HCV patients, we felt it
was very important to identify kinases responsible for activation of
2ꢀ-MeC in this key cell line.
Although among normal human tissues UCK2 expression was
observed only in placenta (Van Rompay et al., 2001), it should be
noted that very little is known about the expression of UCK1, UCK2
or dCK in primary human hepatocyte cultures or in vivo in the con-
text of the uninfected or HCV-infected liver. It remains formally
possible that the phosphorylation pathways used in hepatocytes or
tissues are different from those defined in cultured cells. However,
phosphorylation studies have shown that cytidine analogs such as
2ꢀ-MeC, 2ꢀ-F-2ꢀ-MeC, and 4ꢀ-azidoC are phosphorylated to about the
same limited extent (around 20 pmol of phosphorylated nucleoside
species per million cells) in primary hepatocytes from rat, monkey
or human as they are in cultured human cell lines (Cretton-Scott et
al., 2008; Standring et al., 2008; Cretton-Scott, unpublished results).
This suggests that the same restrictive first phosphorylation steps
are operative in these different systems. Conducting more study
in human hepatocytes especially in HCV-infected hepatocytes, will
further clarify this question.
Ahmed, N.K., 1981. Determination of pyrimidine phosphoribosyltransferase and uri-
dine kinase activities by an assay with DEAE-paper discs. J. Biochem. Biophys.
Methods 4, 123–130.
Ali, S., Leveque, V., Le Pogam, S., Ma, H., Philipp, F., Inocencio, N., Smith, M., Alker,
A., Kang, H., Najera, I., Klumpp, K., Symons, J., Cammack, N., Jiang, W.-R., 2008.
Selected replicon variants with low-level in vitro resistance to the hepatitis
C virus NS5B polymerase inhibitor PSI-6130 lack cross-resistance with R1479.
Antimicrob. Agents Chemother. 52, 4356–4369.
Anderson, E.P., Brockman, R.W., 1964. Feedback inhibition of uridine kinase by cyti-
dine triphosphate and uridine triphosphate. Biochim. Biophys. Acta 91, 380–386.
Arner, E.S., Spasokoukotskaja, T., Eriksson, S., 1992. Selective assays for thymi-
dine kinase 1 and 2 and deoxycytidine kinase and their activities in extracts
from human cells and tissues. Biochem. Biophys. Res. Commun. 188, 712–
718.
Beauséjour, C.M., Gagnon, J., Primeau, M., Momparler, R.L., 2002. Cytotoxic activity of
2ꢀ,2ꢀ-difluorodeoxycytidine, 5-aza-2ꢀ-deoxycytidine and cytosine arabinoside in
cells transduced with deoxycytidine kinase gene. Biochem. Biophys. Res. Com-
mun. 293, 1478–1484.
Cooper, C., Lawitz, E.J., Ghali, P., Rodriguez-Torres, M., Anderson, F.H., Lee, S.S., Proulx,
L., 2007. Antiviral activity of the non-nucleoside polymerase inhibitor, VCH-
759, in chronic hepatitis C patients: results from a randomized, double-blind,
placebo-controlled, ascending multiple dose study. Hepatology 46 (Suppl. 1),
864A.
Cretton-Scott, E., Perigaud, C., Peyrottes, S., Licklider, L., Camire, M., Larsson, M., La
Colla, M., Hildebrand, E., Lallos, L., Bilello, J., McCarville, J., Seifer, M., Liuzzi, M.,
Pierra, C., Badaroux, E., Gosselin, G., Surleraux, D., Standring, D.N., 2008. In vitro
antiviral activity and pharmacology of IDX184, a novel and potent inhibitor of
HCV replication. J. Hepatol. 48, S220.
De Francesco, R., Carfí, A., 2007. Advances in the development of new therapeu-
tic agents targeting the NS3-4A serine protease or the NS5B RNA-dependent
RNA polymerase of the hepatitis C. Adv. Drug Deliv. Rev. 59, 1242–
1262.
Ives, D.H., Durham, J.P., Tucker, V.S., 1969. Rapid determination of nucleoside kinase
and nucleotidase activities with tritium-labeled substrates. Anal. Biochem. 28,
192–205.
Johansson, M., Brismar, S., Karlsson, A., 1997. Human deoxycytidine kinase is located
in the cell nucleus. Proc. Natl. Acad. Sci. 94, 11941–11945.
Klumpp, K., Kalayanov, G., Ma, H., Le Pogam, S., Leveque, V., Jiang, W.-R., Inocen-
cio, N., De Witte, A., Rajyaguru, S., Tai, E., Chanda, S., Irwin, M.R., Sund, C.,
Winqist, A., Maltseva, T., Eriksson, S., Usiva, E., Smith, M., Alker, A., Najera, I.,
Cammack, N., Martin, J.A., Johansson, N.G., Smith, D.B., 2008. 2ꢀ-Deoxy-4ꢀ-azido
nucleoside analogs are highly potent inhibitors of Hepatitis C virus replica-
tion despite the lack of 2ꢀ-alfa-hydroxyl groups. J. Biol. Chem. 283, 2167–
2175.
Le Pogam, S., Seshaadri, A., Kosaka, A., Hu, S., Kang, H., Symons, J., Klumpp, K.,
Cammack, N., Najera, I., 2008. A high barrier to resistance may contribute to
the robust antiviral effect demonstrated by R1626 in HCV genotype 1-infected
treatment-naive patients. Hepatology 46, 813A.
Lalezari, J., Gane, E., Rodriguez-Torres, M., DeJesus, E., Nelson, D., Everson, G., Jacob-
son, I., Reddy, R., Hill, G.Z., Beard, A., Symonds, W.T., Berrey, M.M., McHutchison,
J.G., 2008. Potent antiviral activity of the HCV nucleoside polymerase inhibitor
R7128 with PEG-IFN and ribavirin: interim results of R7128 500 mg BID for 28
days. J. Hepatol. 48, S29.
Liu, Y., Jiang, W.W., Pratt, J., Rockeway, T., Harris, K., Vasavanonda, S., Tripathi,
R., Pithawalla, R., Kati, W.M., 2006. Mechanistic study of HCV polymerase
inhibitors in individual steps of the polymerization reaction. Biochemistry 45,
11312–11323.
Meanwell, N.A., Koszalka, G.W., 2008. 2007: a difficult year for HCV drug develop-
ment. Curr. Opin. Investig. Drugs 9, 128–131.
Murakami, E., Bao, H., Ramesh, M., McBrayer, T.R., Whitaker, T., Micolochick
Steuer, H.M., Schinazi, R.F., Stuyver, L.J., Obikhod, A., Otto, M.J., Furman, P.A.,
2007. Mechanism of activation of ␤-d-2ꢀ-deoxy-2ꢀ-fluoro-2ꢀ-C-methylcytidine
and inhibition of hepatitis C virus NS5B RNA polymerase. Antimicrob. Agents
Chemother. 51, 503–509.
The preceding observations suggest the possibility of devising
strategies to improve the antiviral activity of the nucleoside which
are inefficiently phosphorylated in the initial kinase step. In partic-

N.L. Golitsina et al. / Antiviral Research 85 (2010) 470–481
481
Nelson, D., Pockros, P., Godofsky, E., Rodriguez-Torres, M., Everson, G., Fried, M.,
Ghalib, R., Harrison, S., Nyberg, L., Shiffman, M., Chan, A., Hill, G., 2008. High
end-of-treatment response (84%) after 4 weeks of R1626, peginterferon alfa-2A
(40KD) and ribavirin followed by a further 44 weeks of peginterferon alfa-2A
and ribavirin. J. Hepatol. 48, S371.
Sabini, E., Ort, S., Monnerjahn, C., Konrad, M., Lavie, A., 2003. Structure of human
dCK suggests strategies to improve anticancer and antiviral therapy. Nat. Struct.
Biol. 10, 513–519.
Shimamoto, Y., Koizumi, K., Okabe, H., Kazuno, H., Murakami, Y., Nakagawa, F.,
Matsuda, A., Sasaki, T., Fukushima, M., 2002. Sensitivity of human cancer cells
to the new anticancer ribo-nucleoside TAS-106 is correlated with expression of
uridine-cytidine kinase 2. Jpn. J. Cancer Res. 93, 825–833.
Standring, D.N., Lanford, R., Wright, T., Chung, R.T., Bichko, V., Cretton-Scott, E., Pan-
Zhou, X., Bergelson, S., Qu, L., Tausek, M., Bridges, E., Moussa, A., Storer, R., Pierra,
C., Benzaria, S., Gosselin, G., La Colla, P., Sommadossi, J.P., 2003. NM 283 has
potent antiviral activity against genotype 1 chronic hepatitis C virus (HCV-1)
infection in the chimpanzee. J. Hepatol. 38, 3.
Pawlotsky, J.M., Chevaliez, S., McHutchison, J.G., 2007. The hepatitis
C
virus
life cycle as
1979–1998.
a target for new antiviral therapies. Gastroenterology 132,
Poordad, F., Lawitz, E.J., Gitlin, N., Rodriguez-Torres, M., Box, T., Nguyen, T.,
Pietropaolo, K., Liu, W., Fielman, B., Mayers, D., 2007. Efficacy and safety of val-
opicitabine in combination with pegylated interferon-a (PEG-IFN) and ribavirin
(RBV) in patients with chronic hepatitis C. Hepatology 46, 866A.
Reddy, R., Rodriguez-Torres, M., Gane, E., Robson, R., Lalezari, J., Everson, G.T., DeJe-
sus, E., McHutchison, J.G., Vargas, H.E., Beard, A., Rodriguez, C.A., Hill, G.Z.,
Symonds, W., Berrey, M., 2007. Antiviral activity, pharmacokinetics, safety, and
tolerability of R7128, a novel nucleoside HCV polymerase inhibitor, following
multiple, ascending, oral doses in patients with HCV genotype 1 infection who
have failed prior therapy. Hepatology 46, 232A.
Roy, B., Verri, A., Lossani, A., Spadari, S., Focher, F., Aubertin, A.-M., Gosselin, G.,
Mathe, C., Perigaud, C., 2004. Enantioselectivity of ribonucleotide reductase:
a first study using stereoisomers of pyrimidine 2ꢀ-azido-2ꢀ-deoxynucleosides.
Biochem. Pharmacol. 68, 711–718.
Standring, D.N., Lanford, R., Cretton-Scott, E., Licklider, L., Larsson, M., Pierra, C.,
Gosselin, G., Perigaud, C., Surleraux, D., Mayes, B., Moussa, A., Selden, J., 2008.
Potent antiviral activity of second generation nucleoside inhibitors, IDX102 and
IDX184, in HCV-infected chimpanzees. J. Hepatol. 48, S30.
Usova, E., Maltseva, T., Foldesi, A., Chattopadhayaya, J., Eriksson, S., 2004. Human
deoxycytidine kinase as a deoxyribonucleoside phosphorylase. J. Mol. Biol. 344,
1347–1358.
Van Rompay, A.R., Norda, A., Lindén, K., Johansson, M., Karlsson, A., 2001. Phosphory-
lation of uridine and cytidine nucleoside analogs by two human uridine–cytidine
kinases. Mol. Pharmacol. 59, 1181–1186.
Shafiee, M., Griffon, J.-F., Gosselin, G., Cambi, A., Vincenzetti, S., Vita, A., Eriksson,
S., Imbache, J.-L., Maury, G., 1998. A comparison of the enantioselectivities of
human deoxycytidine kinase and human cytidine deaminase. Biochem. Phar-
macol. 56, 1237–1242.
Van Rompay, A.R., Johansson, M., Karlsson, A., 2003. Substrate specificity and
phosphorylation of antiviral and anticancer nucleoside analogues by human
deoxyribonucleoside kinases and ribonucleoside kinases. Pharmacol. Ther. 100,
119–139.