Anabolism of amdoxovir: Phosphorylation of dioxolane guanosine and its 5′-phosphates by mammalian phosphotransferases

Article abstract of DOI:10.1016/j.bcp.2004.06.019

Amdoxovir [(-)-β-D-2,6-diaminopurine dioxolane, DAPD], the prodrug of dioxolane guanosine (DXG), is currently in Phase I/II clinical development for the treatment of HIV-1 infection. In this study, we examined the phosphorylation pathway of DXG using 15 purified enzymes from human (8), animal (6), and yeast (1) sources, including deoxyguanosine kinase (dGK), deoxycytidine kinase (dCK), high K_m 5′-nucleotidase (5′-NT), guanylate (GMP) kinase, nucleoside monophosphate (NMP) kinase, adenylate (AMP) kinase, nucleoside diphosphate (NDP) kinase, 3-phosphoglycerate (3-PG) kinase, creatine kinase, and pyruvate kinase. In addition, the metabolism of ¹⁴C-labeled DXG was studied in CEM cells. DXG was not phosphorylated by human dCK, and was a poor substrate for human dGK with a high K_m (7 mM). Human 5′-NT phosphorylated DXG with relatively high efficiency (4.2% of deoxyguanosine). DXG-MP was a substrate for porcine brain GMP kinase with a substrate specificity that was 1% of dGMP. DXG-DP was phosphorylated by all of the enzymes tested, including NDP kinase, 3-PG kinase, creatine kinase, and pyruvate kinase. The BB-isoform of human creatine kinase showed the highest relative substrate specificity (47% of dGDP) for DXG-DP. In CEM cells incubated with 5 μM DXG for 24 h, 0.015 pmole/10⁶ cells (～7.5 nM) of DXG-TP was detected as the primary metabolite. Our study demonstrated that 5′-nucleotidase, GMP kinase, creatine kinase, and NDP kinase could be responsible for the activation of DXG in vivo.

Full text of DOI:10.1016/j.bcp.2004.06.019

Biochemical Pharmacology 68 (2004) 1879–1888
www.elsevier.com/locate/biochempharm
Anabolism of amdoxovir: phosphorylation of dioxolane guanosine
and its 50-phosphates by mammalian phosphotransferases
a,
b
a
c
*
Joy Y. Feng , William B. Parker , Megan L. Krajewski , Dominique Deville-Bonne ,
Michel Veronc, Preethi Krishnand, Yung-Chi Chengd, Katyna Borroto-Esodaa
aGilead Sciences, 4 University Place, 4611 University Drive, Durham, NC 27707, US
bSouthern Research Institute, 2000 Ninth Avenue South, Birmingham, AL 35205, USA
c
´
´
´
Unite de Regulation enzymatique des Activites cellulaires, Institut Pasteur, CNRS URA 2185, 75724 Paris Cedex 15, France
dDepartment of Pharmacology, School of Medicine, Yale University, 333 Cedar Street, New Haven, CT 06520, USA
Received 16 March 2004; accepted 14 June 2004
Abstract
Amdoxovir [(À)-b-D-2,6-diaminopurine dioxolane, DAPD], the prodrug of dioxolane guanosine (DXG), is currently in Phase I/II
clinical development for the treatment of HIV-1 infection. In this study, we examined the phosphorylation pathway of DXG using 15
purified enzymes from human (8), animal (6), and yeast (1) sources, including deoxyguanosine kinase (dGK), deoxycytidine kinase
(dCK), high Km 50-nucleotidase (50-NT), guanylate (GMP) kinase, nucleoside monophosphate (NMP) kinase, adenylate (AMP) kinase,
nucleoside diphosphate (NDP) kinase, 3-phosphoglycerate (3-PG) kinase, creatine kinase, and pyruvate kinase. In addition, the
metabolism of 14C-labeled DXG was studied in CEM cells. DXG was not phosphorylated by human dCK, and was a poor substrate
for human dGK with a high Km (7 mM). Human 50-NT phosphorylated DXG with relatively high efficiency (4.2% of deoxyguanosine).
DXG–MP was a substrate for porcine brain GMP kinase with a substrate specificity that was 1% of dGMP. DXG–DP was phosphorylated
by all of the enzymes tested, including NDP kinase, 3-PG kinase, creatine kinase, and pyruvate kinase. The BB-isoform of human creatine
kinase showed the highest relative substrate specificity (47% of dGDP) for DXG–DP. In CEM cells incubated with 5 mM DXG for 24 h,
0.015 pmole/106 cells (~7.5 nM) of DXG–TP was detected as the primary metabolite. Our study demonstrated that 50-nucleotidase, GMP
kinase, creatine kinase, and NDP kinase could be responsible for the activation of DXG in vivo.
# 2004 Elsevier Inc. All rights reserved.
Keywords: Amdoxovir; Anabolism; DXG; Phosphorylation; Phosphotransferase
1. Introduction
sphate (DP), and finally 50-triphosphate (TP). Studies have
shown that an elevated intracellular concentration of
nucleotide 50-triphosphate analogs (ddNTP) correlated
with enhanced activity [1]. Amdoxovir [(À)-b-D-2,6-dia-
minopurine, DAPD], the prodrug of dioxolane guanosine
(DXG), is currently in Phase I/II clinical development for
the treatment of HIV-1 infection (Fig. 1). Both in vitro and
in vivo studies have demonstrated that the anti-HIVactivity
seen upon treatment with DAPD is almost entirely due to
the generation of DXG by the action of adenosine deami-
nase (ADA) on DAPD [2–4]. DXG is subsequently phos-
phorylated to its 50-triphosphate (DXG–TP), which is a
strong alternative substrate inhibitor of the HIV reverse
transcriptase (RT) [2]. Since DXG is a deoxyguanosine
(dGuo) analog, it has been assumed that DXG is activated
to DXG–TP by the same phosphorylation pathway used by
the natural nucleoside dGuo. In addition, several other
Nucleoside analogs play important roles in anti-viral and
anti-cancer therapy. Most of these compounds need to be
transformed into their active 50-triphosphate form by cel-
lular or viral kinases to exert their pharmacological activ-
ity. Through three consecutive steps, nucleoside analogs
are phosphorylated to 50-monophosphate (MP), 50-dipho-
Abbreviations: AMP kinase, Adenylate kinase; BDG, 9-benzyl-9-dea-
zaguanine; DAPD, [(À)-b-D-2,6-diaminopurine dioxolane]; d4T, 20,30-
didehydro-30-deoxythymidine; DXG, dioxolane guanosine; dCK, deoxycy-
tidine kinase; dGK, deoxyguanosine kinase; DP, 50-diphosphate; GMP,
guanylate; MP, 50-monophosphate; MPA, mycophenolic acid; NMP,
nucleoside monophospahte; NDP, nucleoside diphosphate kinase; 50-NT,
50-nucleotidase; 3-PG, 3-phosphoglycerate; TP, 50-triphosphate.
* Corresponding author. Tel.: +1 919 294 7647; fax: +1 919 493 5790.
E-mail address: joy.feng@gilead.com (J.Y. Feng).
0006-2952/$ – see front matter # 2004 Elsevier Inc. All rights reserved.
doi:10.1016/j.bcp.2004.06.019

1880
J.Y. Feng et al. / Biochemical Pharmacology 68 (2004) 1879–1888
(dFdG) [11]. dCK, a cytosolic enzyme constitutively
expressed during the cell cycle, phosphorylates dCyd,
dAdo, and dGuo with ATP and UTP as phosphate donors,
and is known to have very broad substrate specificity [10].
An ubiquitous enzyme, the high Km 50-NT is specific for
the reversible dephosphorylation reaction of 6-oxypurine
nucleoside monophosphate (IMP and GMP) [12,13]. With
IMP as the phosphate donor, it also phosphorylates dGuo,
dAdo, acyclovir, tiazofurin, didanosine (ddI), and (À)-
carbovir [13].
Fig. 1. Structures of DAPD and DXG.
enzymes were known to be involved in the phosphorylation
of dGuo analogs (À)carbovir [5,6], acyclovir [7], and
ganciclovir [8]. In this study, we examined the phosphor-
ylation pathway of DXG using 15 purified enzymes from
human (8), animal (6), and yeast (1) sources (Table 1). The
enzymes in this study were divided into three groups: (1)
enzymes that phosphorylate deoxynucleoside to dNMP,
including dGuo kinase (dGK), deoxycytidine (dCyd)
kinase (dCK), and high Km 50-nucleotidase (50-NT); (2)
enzymes that phosphorylate deoxynucleoside-MP (dNMP)
to dexoynucleoside-DP (dNDP), including guanylate
(GMP) kinase, bovine liver nucleoside monophosphate
(NMP) kinase, and adenylate (AMP) kinase, which is also
known as myokinase; and (3) enzymes that phosphorylate
dNDP to deoxynucleoside-TP (dNTP), including nucleo-
side diphosphate (NDP) kinase, 3-phosphoglycerate (3-
PG) kinase, creatine kinase, and pyruvate kinase. In addi-
tion, we evaluated the metabolism of 14C-labeled DXG in
CEM cells. Previously, we reported the metabolism of
[8-3H]-DXG [2], however, the levels of DXG–MP, DP,
and TP were found to be significantly lower in later studies
[9] (personal communication with Dr. Robert St. Claire III,
Gilead Sciences).
GMP kinase has been purified from many eukaryotic
sources including bovine retina, pig brain, Baker’s yeast,
rat liver, and human erythrocytes [14]. However, there is
limited information on the activity of the purified recom-
binant human enzyme [14,15]. In this study, we chose to
use the commercially available GMP kinase purified from
pig brain. The high sequence similarity between the pig
GMP kinase and human GMP kinase (90.3% identity,
93.8% homology) [14] suggests that the information
gained on phosphorylation of dGMP and DXG–MP by
pig GMP kinase is likely to be directly transferable to the
human enzyme. NMP kinase from bovine liver [16,17] has
been shown to phosphorylate UMP, CMP, GMP, AMP, and
acyclo-GMP (only to a limited extent) with ATP as the
phosphate donor [7]. In this study, we also tested the
phosphorylation of DXG–MP using commercially avail-
able AMP kinase purified form pig heart, which is geneti-
cally similar to the human enzyme [18].
NDP kinase phosphorylates all natural nucleoside
diphosphates into triphosphates using ATP as the major
phosphate donor. The NDP kinase activity is found to be
much higher than many of the other enzymes involved in
nucleotide metabolism [19]. Eight isoforms (nM23-H1 to
nM23-H8) have been isolated in humans. The cytosolic
enzymes nM23-H1 (type A) and nM23-H2 (type B) share
high homology and are the most extensively studied [20–
23]. In this study, we used recombinant human enzyme
(type A, NM23-H1) and commercially available NDP
dGK is a mitochondrial enzyme constitutively expressed
during the cell cycle with dGuo and 20-deoxyadenosine
(dAdo) as natural substrates [10]. It has been shown to
phosphorylate anti-cancer nucleoside analogs, including 2-
chloro-20-deoxyadenosine (CdA), 9-b-D-arabinofurano-
sylguanine (araG), and 20,20-difluorodeoxyguanosine
Table 1
Kinases used in the study of DXG phosphorylation pathway
Kinases
EC #
Sources
dGK
2.7.1.113
2.7.1.74
3.1.3.5
2.7.4.8
2.7.4.4
2.7.4.3
2.7.4.6
Recombinant human enzyme
Recombinant human enzyme
Recombinant human enzyme
Porcine brain (Sigma #G9385)
Bovine liver (ICN #155962)
Porcine heart (CalBiochem #475945)
Bovine liver (Sigma #2635)
dCK
High Km 50-NT
GMP kinase
NMP kinase
AMP kinase
NDP kinase
Recombinant human enzyme, type A
3-PG kinase
2.7.2.3
2.7.3.2
Baker’s yeast (Sigma #P7634)
Recombinant human enzyme
Creatine kinase
Rabbit muscle, type I (Sigma #C3755)
Recombinant human enzyme (MM) (Sigma #C9858)
Recombinant human enzyme (BB) (Sigma #C9983)
Recombinant human enzyme (MB) (Sigma #C0984)
Pyruvate kinase
2.7.1.40
Rabbit muscle, type II (Sigma #P1506)

J.Y. Feng et al. / Biochemical Pharmacology 68 (2004) 1879–1888
1881
kinase purified from bovine liver, which was shown to
2.3. Enzymes
phosphorylate acyclovir-DP [24]. 3-PG kinase is a glyco-
lytic enzyme that catalyzes transfer of a phosphate group
from 3-phospho-D-glycerol phosphate to ADP forming 3-
phospho-D-glycerate and ATP, and is 96% conserved in
mammals [25]. It has been reported that human cellular 3-
PG kinase phosphorylated ganciclovir-DP to its tripho-
sphate form [8]. In addition, acyclovir-DP was phosphory-
lated by the 3-PG kinase partially purified from human
erythrocytes [24]. Recently, Krishnan et al. suggested
human 3-PG kinase as being the key enzyme for the last
step of phosphorylation during the activation of L-nucleo-
side analogs [23]. Creatine kinase phosphorylates ADP to
ATP using creatine phosphate as the phosphate donor [26].
There are four types of creatine kinase subunits that are
expressed specifically depending on species, development
stage and tissue. The two cytosolic types are M-CK
(muscle) and B-CK (brain), which form dimers of three
isoenzymes: MM-CK, MB-CK, and BB-CK. The enzyme
has been shown to phosphorylate 20,30-dideoxycytidine 50-
DP (ddCDP) [23], cidofovir-MP [27], carbovir-DP [5,6],
adefovir-DP [28], and 20,30-didehydro-30-deoxythymidine
(d4T)-DP [23]. Pyruvate kinase, a key glycolytic enzyme,
catalyzes the transfer of a phosphate group from phos-
pho(enol)pyruvate to ADP. In this study, we used commer-
cially available rabbit muscle pyruvate kinase, which was
shown to phosphorylate a series of the nucleoside analog
diphosphates to a limited extent [6,23,24].
The sources of the enzymes are summarized in Table 1.
The recombinant human dGK and dCK were generously
provided by Eriksson and coworkers [31]. The clone for the
high Km 50-NT was kindly provided by Spychala et al. [12]
and purified in our laboratory based on published methods.
Human recombinant NDP kinase (form A, NM23-H1) and
human recombinant 3-PG kinase were purified as pre-
viously described [23,32]. The molecular weights used
to calculate the kcat values for dGK, dCK, and 50-NT is 58,
60, and 195 kDa, respectively. The molecular weight used
for the calculation of kcat is 22 kDa for GMP kinase,
17 kDa for bovine liver and human NDP kinase, 47 kDa
for Baker’s yeast 3-PG kinase, 46 kDa for human 3-PG
kinase, 80 kDa for all of the creatine kinases, and 59 kDa
for rabbit pyruvate kinase.
50-NT was purified based on a published method with
modifications [12]. Fifty milliliter overnight growth of
Escherichia coli (50 mg/L kanamycin) was used to inno-
culate 2 L volumes of LB media. Cells were grown to a
density of OD600 = 0.6–0.7 and induced with a final
concentration of 1 mg/mL of IPTG for 5 h. Cells were
collected by centrifugation at 8000 rpm for 10 min on a
Sorvall RC 5B Centrifuge with a SLA-150 rotor. The cell
pellets were suspended in 10 volume of homogenization
buffer containing 20 mM Tris–HCl, pH 8.0, 20 mM KCl,
1 mM EGTA, 1 mM 2-mercaptolethanol, 0.1 mM phenyl-
methansulfonyl fluoride (PMSF), 10 mg/mL of pepstatin
and leupeptine, and 0.2 mg/mL of lysozyme. The cells
were lyzed by stirring at room temperature in the lysis
buffer for 30 min, followed by sonication using Branson
Sonifier 450 (100% constant, output 3, 3 min  1 min).
The crude lysate was centrifuged at 14,000 rpm for 20 min
using a Sorvall RC 5B Centrifuge. The supernatant was
removed and subjected to 50% ammonium sulfate pre-
cipitation. After 1 h of incubation at 4 8C, protein preci-
pitates were collected by centrifugation at 8000 rpm on a
Sorvall RC 5B Centrifuge. The protein pellets were sus-
pended in 13 mL Buffer A (20 mM TrisCl, pH 7.0, 5 mM
imidazole, 0.6 M NaCl, and 5 mM 2-mercaptolethanol)
and dialyzed in Buffer A overnight (2 Â 2 L). The dialyzed
protein was loaded onto a pre-equilibrated Ni-NTA Super-
folw resin (bed volume = 3 mL), and incubated for 30 min
at 4 8C on a shaker. The mixture of the resin and protein
solution was loaded on to a column (3 cm  13 cm). The
column was washed with Buffer A (~70 mL) until no
protein was detected from the wash (OD280 < 0.07 AU).
The protein was eluted with Buffer B (20 mM Tris, pH 7.0,
20 mM–1 M imidazole, 0.6 M NaCl, and 5 mM 2-mercap-
tolethanol). The step gradient of Buffer B was set at a
rate of 20 mM per step, 10 mL per step. Five milliliters
of fractions was collected, and fractions containing
50-NT as determined by SDS-PAGE analysis were
pooled, exchanged into Buffer C (100 mM imidazole,
pH 7.0, 1 M NaCl, 10 mM 2-mercaptolethanol), and then
2. Material and methods
2.1. Reagents
[8-14C]-Inosine, [8-14C]-deoxyguanosine, and [8-14C]-
DXG were obtained from Moravek Biochemicals. ATP
was from Amersham Pharmacia Biotech. Inosine, IMP,
dGuo, dGMP, dGDP, dGTP, ADP, phospho(enol)pyruvate,
creatine phosphate, NAD+, DL-glyceraldehyde-3-phos-
phate, glyceraldehyde-3-phosphate dehydrogenase, and
N-acetyl-L-cysteine were from Sigma. All other materials
and reagents were of the highest quality available.
2.2. Cell culture
CEM cells, obtained from the American Type Culture
Collection were grown in RPMI 1640 medium (Gibco-
BRL) containing 10% fetal bovine serum (Gibco-BRL),
10 U/mL penicillin, 10 mg/mL streptomycin, and 50 mg/
mL gentamycin [29]. Cells were routinely tested for the
presence of mycoplasma by measuring [3H] thymidine
uptake [30]. Cells were maintained in a humidified
atmosphere of 5% CO2 in air at 37 8C. All experiments
were performed with cells that were proliferating at max-
imal rates. Cell numbers were determined with a coulter
counter.

1882
J.Y. Feng et al. / Biochemical Pharmacology 68 (2004) 1879–1888
mixed with an equal volume of glycerol. The final protein
concentration was 1.2 mg/mL in 50% glycerol, 50 mM
imidazole, pH 7.0, 0.5 M NaCl, and 5 mM 2-mercapto-
lethanol. The protein concentration was determined using
the Bradford assay (Bio-Rad protein assay) with BSA
as a standard. The enzyme was stored at À20 8C, under
which condition the enzyme is stable for at least two
years.
All of the enzymatic phosphorylation of dGDP and
DXG–DP reactions were carried out in 50 mM Tris acetate
(pH 7.5), 5 mM MgCl2 at 37 8C. At 2, 4, 6, 8, and 10 min, a
portion of the reaction mixture was removed and quenched
with 50 mM EDTA (final concentration after mixing). In
cases where EDTA interfered with the chromotography,
3.5–3.8% (w/w) HCl was used to stop the reaction. The
reaction products were subjected to HPLC analysis as
described below. NDP kinase reaction mixture contained
5 mM NaF and 4 mM ATP for NDP kinase from bovine
liver or 1 mM ATP for recombinant human NDP kinase
(form A). 3-PG kinase phosphorylation of dGuo and
DXG–DP was measured using coupled reactions in a
mixture containing 10 mM sodium phosphate (pH 7.5),
4 mM NAD+, and 4 mM DL-glyceraldehyde [23]. Prior to
the addition of the enzyme, the phosphate donor of this
reaction, 1,3-biphosphoglycerate was generated in situ by
incubating the reaction mixture with 8 unit/mL glyceral-
dehyde-3-phosphate dehydrogenase at room temperature
for 20 min. For creatine kinase, the reaction mixtures con-
tained 5 mM NaF, 20 mM creatine phosphate, 0.1% (w/w)
BSA, and 10 mM N-acetyl-L-cysteine as an activator [34].
Pyruvate kinase reactions were studied in the presence of
100 mM KCl and 4 mM phospho(enol)pyruvate.
2.4. Enzymatic phosphorylation of nucleosides and
nucleotides
The phosphorylation of dGuo and DXG by dGK and
dCK was conducted in a direct phosphorylation assay by
monitoring the formation of product from 14C-labeled
nucleoside substrates using ATP as phosphate donor
[33]. A mixture containing 50 mM Tris–HCl (pH 7.6),
2.5 mM ATP, 5 mM MgCl2, 125 mM KCl, 15 mM NaF,
10 mM DTT, 0.5 mg/mL BSA, and 0.5 mM–20 mM of the
corresponding nucleoside substrate was pre-incubated at
37 8C. The reaction was initiated by the addition of the
enzyme, and at each time interval, 5 mL of the reaction
mixture was removed and spotted onto Whatman DE81
filter paper (A). For each reaction, 3 Â 5 mL of mixture was
spotted onto another piece of Whatman DE81 filter paper
(B) to determine total counts. After air-dried, filter paper
(A) was washed with 5 mM ammonium formate (Â3) and
ethanol (Â1) and dried. Both filter papers (A) and (B) were
transferred to scintillation vials and eluted with 0.5 mL
elution buffer (0.2 M KCl, 0.1 M HCl). After the mixture
was incubated at room temperature for 30 min on a shaker,
4.5 mL of scintillation fluid was added to each vial, and
briefly mixed by vortex. The amount of the radio-labeled
reaction product and substrate were quantitated by a liquid
scintillation counter.
2.5. HPLC and data analysis for enzyme assays
Prior to HPLC analysis, each collected quenched mix-
ture was passed through a YM-10 Microcon centrifugal
filter unit (Millipore) and analyzed within 48 h using
HPLC (HP 1100, Agilent Technologies). The samples were
stored at room temperature prior to analysis. No degrada-
tion or conversion of the reaction mixture was detected
during analysis. The strong-ion exchange HPLC was per-
formed on a BioBasic AX (4.5 mm  100 mm, 5 mm)
column (ThermoHypersil-Keystone). Elution of the
nucleotides was accomplished with a 15 min linear gra-
dient from 10 mM ammonium phosphate (pH 5.5, 5%
MeOH) to 300 mM ammonium phosphate (pH 5.5, 5%
MeOH) buffer followed by 9 min of 300 mM ammonium
phosphate (pH 5.5, 5% MeOH) at a flow rate of 1.5 mL/
min [35]. The nucleotide substrate and product were
detected at 253 nm and quantitated based on peak area.
The chosen enzyme concentration was based on an
established linear correlation between the enzyme concen-
tration and the rate of the reaction. In each experiment, at
least five time points were taken before 15% of the sub-
strate was consumed. The initial reaction velocities were
determined using linear regression. The observed reaction
rate (v) was fitted to the Michaelis–Menten equation as a
function of substrate concentration ([S])
The phosphorylation of dGuo, inosine, and DXG by high
Km 50-NT were carried out at 37 8C in a reaction mixture
containing 50 mM imidazole–HCl (pH 6.5), 20 mM
MgCl2, 0.5 mg/mL BSA, 5 mM IMP (as phosphate donor),
5 mM ATP (as activator), and 14C-labeled nucleoside in a
volume of 20 mL [6,12]. The reaction was initiated by
adding 50-NT, and at each time interval 2 mL aliquots were
spotted onto polyethylenimine (PEI) thin-layer chromato-
graphy plates pre-spotted with 10 nmol of unlabeled
nucleoside. Plates were developed in 50% MeOH. Pro-
ducts and substrate were quantitated by cutting out the UV
spots and counting the radioactivity by liquid scintillation
counter.
The phosphorylation of dGMP and DXG–MP by GMP
kinase, NMP kinase, and AMP kinase was initiated by
adding enzyme to a pre-incubated 37 8C mixture of Tris
acetate (pH 7.5), 100 mM KCl, 10 mM MgCl2, and 4 mM
ATP. At 2, 4, 6, 8, and 10 min, a portion of the reaction
mixture was removed and quenched with 3.5–3.8% (w/w)
HCl. The reaction products were subjected to HPLC
analysis as described below.
ðkcat½EŠ ½SŠÞ
ðKm þ0½SŠÞ
n ¼
Thekinetic parameterskcat andKm were determined using
non-linear regression curve fitting (KaleidaGraph, 3.5).

J.Y. Feng et al. / Biochemical Pharmacology 68 (2004) 1879–1888
1883
2.6. Measurement of DXG–MP, –DP, and –TP in
CEM cells
phosphotransferases. Kinetic constants, kcat and Km, were
measured using steady state analysis. The catalytic effi-
ciency for each substrate is expressed as kcat/Km. The
relative phosphorylation efficiency for DXG and its
nucleotides is expressed as a percentage of that of dGuo
and its corresponding nucleotides. The results are sum-
marized in Tables 2–4.
CEM cells (~500,000 cells/mL) were incubated with
5 mM of [8-14C]-DXG (45 mCi/mmol) for 6 and 24 h at
37 8C. The cells (45 mL) were collected by centrifugation,
and the pellets were washed twice with ice-cold Puck’s
saline G [36] and mixed with 200 mL of ice-cold 0.5 M
perchloric acid. The acid-insoluble material was removed
by centrifugation at 12,000 Â g for 5 min, and the super-
natant fluid was removed and neutralized with 1 M potas-
sium phosphate (pH 7.5) and 8 M KOH. KClO4 was
removed by centrifugation at 12,000 Â g for 5 min, and
180 mL of the supernatant fluid was injected onto a Partisil
SAX strong ion-exchange column (4.6 mm  250 mm,
10 mm, ThermoHypersil-Keystone). Elution of the nucleo-
tides was accomplished with a 30-min linear gradient from
5 mM NH4H2PO4 (pH 2.8) to 750 mM NH4H2PO4 (pH
3.7) buffer followed by 10 min of 750 mM NH4H2PO4 (pH
3.7) at a flow rate of 2 mL/min. The fractions were
collected at 1-min intervals, and radioactivity was deter-
mined with a scintillation counter.
The dGK, dCK, and high Km 50-NT were tested for their
ability to phosphorylate dGuo and DXG. All three of the
enzymes phosphorylate the natural nucleoside dGuo effi-
ciently. DXG did not serve as a substrate for dCK. DXG
was phosphorylated by dGK at a rate of 0.025 sÀ1, similar
to the 0.030 sÀ1 rate for dGuo, however, the high Km
(7 mM) value of DXG revealed that the binding affinity
of DXG to the enzyme was more than 4000-fold weaker
than that of dGuo. As a result, the substrate specificity
(kcat/Km) for DXG was only 0.02% of that of dGuo for this
dGK catalyzed reaction. Both dGuo and inosine were
shown to be good substrates for the high Km 50-NT, with
Km values of 1 and 3.7 mM, and kcat values of 260 and
320 sÀ1, respectively. Compared to dGuo, DXG was phos-
phorylated at a rate that was 4.3-fold slower than dGuo and
was bound 5.6-fold weaker to the enzyme than dGuo.
Overall, DXG was phosphorylated by 50-NT at a level that
was 4.2% of dGuo.
3. Results
Among the three enzymes tested for the phosphorylation
of DXG–MP, GMP kinase was the only enzyme that
phosphorylated dGMP and DXG–MP. The Km value for
In this study, we examined the phosphorylation of dGuo,
DXG, and their mono- and di-phosphates by a series of
Table 2
Phosphorylation of dGuo and DXG to dGMP and DXG–MP
Enzymes
dGuo
DXG
Percentage of dGuoc
Km (mM)
kcat (sÀ1
)
kcat/Km
(mMÀ1sÀ1
Km (mM)
kcat (sÀ1
)
kcat/Km
(mMÀ1sÀ1
)
)
dGKa
(1.7 ꢀ 0.1) Â 10À3
0.167 ꢀ 0.004
1.0 ꢀ 0.1
0.030 ꢀ 0.002
2.7 ꢀ 0.1
18
7.0 ꢀ 0.5
0.025 ꢀ 0.008
Not detectable
60 ꢀ 10
0.0036
0.02
–
dCKa
16
Not detectable
5.6 ꢀ 0.9
DXG
–
50-NT (high Km)b
260 ꢀ 10
260
11
4.2
Enzyme
Inosine
Percentage of inosined
Km (mM)
kcat (sÀ1
)
kcat/Km
(mMÀ1sÀ1
Km (mM)
kcat (sÀ1
)
kcat/Km
(mMÀ1sÀ1
)
)
50-NT (high Km)b
3.7 ꢀ 0.7
ATP (2.5 mM) as the phosphate donor.
1040 ꢀ 130
280
5.6 ꢀ 0.9
60 ꢀ 10
11
3.9
a
b
c
d
IMP (5 mM) as the phosphate donor.
Percentage of dGuo is defined as (kcat/Km)DXG/(kcat/Km)dGuo  100%.
Percentage of inosine is defined as (kcat/Km)DXG/(kcat/Km)inosine  100%.
Table 3
Phosphorylation of dGMP and DXG–MP to dGDP and DXG–DP
Enzymes
dGMP
DXG–MP
Percentage of dGMPb
Km (mM)
kcat (sÀ1
)
kcat/Km
(mMÀ1sÀ1
Km (mM)
kcat (sÀ1
)
kcat/Km
(mMÀ1sÀ1
)
)
GMP kinasea
NMP kinasea
AMP kinasea
a
0.16 ꢀ 0.02
7.3 ꢀ 0.04
46
–
0.23 ꢀ 0.01
0.108 ꢀ 0.003
Not detectable
Not detectable
0.47
–
1.0
–
Not detectable
Not detectable
Not detectable
Not detectable
Not detectable
Not detectable
–
–
–
ATP (4 mM) as the phosphate donor.
b
Percentage of dGMP is defined as (kcat/Km)DXG–MP/ (kcat/Km)dGMP Â 100%.

1884
J.Y. Feng et al. / Biochemical Pharmacology 68 (2004) 1879–1888
Table 4
Phosphorylation of dGDP and DXG–DP to dGTP and DXG–TP
Enzymes
dGDP
DXG–DP
Km (mM)
Percentage of dGDPc
kcat (sÀ1
)
kcat/Km
(mMÀ1sÀ1
kcat (sÀ1
)
kcat/Km
(mMÀ1sÀ1
)
a
a
Km (mM)
)
NDP kinasea (bovine liver)
NDP kinaseb (human, form A)
3-PG kinase (Baker’s yeast)
3-PG kinase (human)
0.103 ꢀ 0.013
0.13 ꢀ 0.02
2.9 ꢀ 0.3
2.0 ꢀ 0.4
6.6 ꢀ 1.6
2.6 ꢀ 0.4
3.3 ꢀ 0.5
3.8 ꢀ 0.7
7.8 ꢀ 1.0
70 ꢀ 3
970 ꢀ 30
14500 ꢀ 300
1900 ꢀ 200
16 ꢀ 2
680
7460
5000
950
2.4
0.38 ꢀ 0.05
1.0 ꢀ 0.1
3.2 ꢀ 0.3
5.7 ꢀ 1.0
3.1 ꢀ 0.4
0.74 ꢀ 0.12
1.4 ꢀ 0.3
0.80 ꢀ 0.02
5.4 ꢀ 1.0
0.31 ꢀ 0.01
7.5 ꢀ 0.3
260 ꢀ 10
47 ꢀ 7
0.82
7.5
0.12
0.10
1.6
81
8.2
0.86
3.7
Creatine kinase (rabbit, type I)
Creatine kinase (human BB)
Creatine kinase (human MM)
Creatine kinase (human MB)
Pyruvate kinase (rabbit muscle)
a
0.27 ꢀ 0.02
5.9 ꢀ 0.5
0.13 ꢀ 0.01
2.0 ꢀ 0.2
6.4 ꢀ 0.4
0.088
8.0
45 ꢀ 3
17
47
9.4 ꢀ 0.6
114 ꢀ 1
2.8
0.093
2.5
3.3
30
8.3
210 ꢀ 20
27
1.2
4.4
ATP (4 mM) as the phosphate donor.
b
c
ATP (1 mM) as the phosphate donor.
Percentage of dGDP is defined as (kcat/Km)DXG–DP/(kcat/Km)dGDP Â 100%.
DXG–MP was 0.23 mM, similar to the 0.16 mM Km value
for dGMP. However, GMP kinase phosphorylated DXG–
MP at a rate 67-fold slower and 1% as efficiently as dGMP,
the natural substrate.
[8-14C]-DXG we used was >99% pure, it is not uncommon
to observe a significant formation of [8-14C]-GTP from a
trace of [8-14C]-guanine during studies of guanosine ana-
logs [37]. Once in the cell, [8-14C]-guanine is converted to
GMP by hypoxanthine-guanine phosphoribosyltransfer-
ase, and guanine is a far better substrate for this enzyme
than DXG is for the nucleoside kinases. As shown in
Table 5, no DXG metabolite was detected at 6 h in cells
treated with 5 mM DXG, and only very low concentrations
of DXG–MP and TP were detected at 24 h, at levels of
0.006 and 0.015 pmole/106 cells, respectively. The pre-
sence of 5 mM of mycophenolic acid (MPA) significantly
increased the DXG–MP, DP, and TP levels to 0.08, 0.23,
and 0.28 pmole/106 cells, respectively, after 24 h incuba-
tion with 5 mM DXG. DXG–TP levels were increased
approximately 19-fold higher than the non-MPA-treated
control (Table 5). Purine nucleoside phosphorylase (PNP)
is a key enzyme involving in the degradation of dGuo in
vivo. In order to verify that guanine was not generated from
DXG by PNP, we tested the effect of a potent inhibitor of
PNP, 9-benzyl-9-deazaguanine (BDG) on the levels of
GTP [38]. The levels of GTP in cells treated with 5 mM
DXG and 1 mg/mL BDG were similar to the cells treated
with 5 mM DXG only (data not shown), indicating that
DXG was not degraded to guanine by PNP. In addition,
BDG did not affect the metabolism of DXG to DXG–MP
The phosphorylation of dGDP and DXG–DP were
carried out by a series of enzymes including NDP kinase,
3-PG kinase, creatine kinase, and pyruvate kinase. DXG–
DP was phosphorylated at efficiencies of 0.12, 0.10, 1.6,
0.86, and 4.4% of dGDP by bovine NDP kinase, human
NDP kinase, Baker’s yeast 3-PG kinase, human 3-PG
kinase, and pyruvate kinase, respectively. For these five
enzymes, the Km values for DXG–DP were 0.7–7.7-fold
higher than that of dGDP, where the reaction rate for DXG–
DP was 33–226-fold slower than dGDP. Among the four
creatine kinases tested, human BB-isoform showed the
highest activity towards DXG–DP. Interestingly, even
though all of the creatine kinases phosphorylated DXG–
DP at a slower rate than dGDP, their apparent binding
affinity for DXG–DP was 2.1–4.8-fold higher than dGDP.
The metabolism of [8-14C]-DXG was studied in CEM
cells after 6 and 24 h of incubation with DXG at a
physiologically relevant concentration of 5 mM. The intra-
cellular levels of DXG–MP, DP, and TP were measured
using HPLC analysis. During incubation with 5 mM DXG,
a considerable level of GTP was detected due to a minor
contamination of [8-14C]-guanine. Even though the
Table 5
Anabolism of DXG in human CEM cells incubated with 5 mM DXG
Treatment of cells
Incubation time (h)
Intracellular concentration (pmole/106 cells)
DXG–MP
DXG–DP
DXG–TP
5 mM DXG
6
NDa
ND
ND
ND
0.015b
24
0.006
5 mM DXG + 5 mM MPA
5 mM DXG + 1 mg/mL BDG
6
0.08
0.08
0.24
0.23
0.24
0.28
24
6
ND
ND
ND
ND
24
0.006
0.012
a
ND: not detectable.
Converted to concentration ꢁ7.5 nM, based on cell volume of 2.1 ml/109 cells for human lymphoblasts [47].
b

J.Y. Feng et al. / Biochemical Pharmacology 68 (2004) 1879–1888
1885
Table 6
Phosphorylation pathways of nucleoside analogues that are currently used for anti-HIV infection
Compounds
Nucleoside to 50-MP
50-MP to 50-DP
50-DP to 50-TP
Emtricitabine
Lamivudine
Zalcitabine
Abacavir
dCK
UMP-CMP kinase
UMP-CMP kinase
UMP-CMP kinase?
Cytosolic deaminase and
GMP kinase
NDP kinase?a
3-PG kinase
dCK
dCK
Creatine kinase
Adenosine phosphotransferase
Creatine kinase/ NDP kinase/ pyruvatekinase/
3-PG kinase
Creatine kinase/ NDP kinaseb
NDP kinase?
Tenofovir
Not applicable
50-NT/adenylate synthetase/
AMP kinase
Didanosine
AMP kinase
adenylate lysase
Stavudine
Zidovudine
a
Thymidine kinase
Thymidine kinase
Thymidylate kinase?
Thymidylate kinase
Creatine kinase/ NDP kinase/ 3-PG kinase
NDP kinase?
‘‘?’’ indicated that the enzyme listed is based on speculation.
b
Data based on the studies of adefovir.
or DXG–TP (Table 5). These data suggested that PNP is
not a major player in the catabolism of DXG in vivo.
important analogs, including the dGuo analogs penciclovir
and ganciclovir, which were phosphorylated at levels of 10
and 1% of that of dGuo [13,40]. However, dGK cannot
phosphorylate 20,30-dideoxyadenosine (ddA), 20,30-dideox-
yguanosine (ddG) [13,41], or guanosine analogs such as
acyclovir, lobucavir, and 30-fluoro-20,30-dideoxyguanosine
(FLG) [41]. In this study, we showed that DXG was
phosphorylated by human dGK at a similar rate as dGuo.
However, the apparent binding affinity of DXG to dGK
was over 4000-fold weaker than dGuo. Although the
naturally occurring deoxynucleoside, dGuo, can be phos-
phorylated efficiently by human dCK, DXG does not serve
as a substrate for the enzyme. This observation is in partial
agreement with an earlier study by Kierdaszuk et al. [33]
where human dCK was completely inactive towards dGuo
analog 20,30-dideoxyguanosine (ddGuo).
4. Discussion
The nucleoside reverse transcriptase inhibitors (NRTI)
play critical roles in anti-HIV therapy. NRTIs use the cell’s
nucleoside phosphorylation pathways by sharing the same
enzymes, which also phosphorylate their corresponding
natural deoxynucleoside counterparts [1]. The phosphor-
ylation of currently clinically used NRTIs to their 50-MP
has been relatively well studied (Table 6) [1]. However, the
phosphorylation of NRTI 50-MP to 50-DP and then to 50-TP
is less known [38]. Moreover, it has been widely assumed
that the last step, catalyzed by NDP kinase, is shared by
most NRTIs, partially based on its broad substrate speci-
ficity for all four natural dNTPs [38]. However, recent
studies showed that 30-azido-30-deoxythymidine (AZT)-
DP [22], 20,30-dideoxyadenosine 50-diphosphate (ddADP)
[22], 20,30-dideoxythymidine 50-diphosphate (ddTDP), and
(À)-b-L-20,30-dideoxy-30-thiacytidine (3TC)-DP [38] are
all very poor [22] substrates of human NDP kinase. Even
d4T-DP, the best ddNDP substrate for the enzyme reported
so far, was phosphorylated at efficiency less than 0.09% of
that of dTDP [20]. Crystal structure studies suggested that
the 30-hydroxyl group, which is lacking in many NRTIs,
was critical for the activity of NDP kinase [22]. In addition,
the catalytic efficiency for the phosphorylation of the
unnatural L-configured ddNDP is much lower than their
D-enantiomers [39]. Since an elevated intracellular con-
centration of didexoynucleotide 50-triphosphate analogs
(ddNTP) correlated with enhanced activity [1], under-
standing of the individual phosphorylation pathway of
each nucleoside analog will help to predict and explain
drug potency and drug interactions, and to modify NRTI
structures to facilitate its activation.
Similar to previous reports [6,42–45], inosine was phos-
phorylated by 50-NT with a Km of 3.7 mM. However,
contrary to earlier studies, which showed that dGuo was
a poor substrate of 50-NT with a high Km (~13 mM) and a
very low Vmax [42,45], we found that dGuo was phos-
phorylated by 50-NT as efficiently as inosine, with a Km of
1 mM. These differences could be due to differences in
enzyme preparation. The catalytic efficiency for the phos-
phorylation of dGuo, inosine, and DXG by 50-NT is much
higher than the phosphorylation of dGuo and DXG by
dGK. Due to the ubiquitous presence of 50-NT in human
tissues and the relatively high substrate activity of DXG as
shown in this study, 50-NT is likely to play a key role in the
activation of DXG both in vitro and in vivo.
Among all of the enzymes tested for the phosphorylation
of DXG–DP, the human BB-isoform of creatine kinase
showed the highest relative substrate specificity (47% of
dGDP) for DXG–DP. In addition, 3-PG kinase and NDP
kinase had the highest catalytic efficiency for the phos-
phorylation of the natural dGDP, even though the relative
catalytic efficiency for DXG–DP phosphorylation by
human 3-PG kinase and NDP kinase is only 0.10 and
0.86% of that of dGDP, which were much less than many of
the enzymes tested in Table 4. Bourdais et al. [22] reported
that AZT-DP, ddADP, and ddTDP were very poor sub-
dGK is an enzyme with a broad substrate specificity,
which can phosphorylate all of the purine deoxyribonu-
cleosides (dGuo, dAdo, and dIno) [10]. Studies have shown
that dGK can phosphorylate many pharmacologically

1886
J.Y. Feng et al. / Biochemical Pharmacology 68 (2004) 1879–1888
Fig. 2. Phosphorylation pathway for DXG. The most important enzymes are shown in bold.
strates of NDP kinase (type B), mainly due to a significant
decrease in rate, instead of binding. Similarly, in this study,
there was a 230-fold decrease in the rate (kcat) and only a
3.7-fold decrease in the binding affinity (Km) for the
phosphorylation of DXG–DP by human NDP kinase,
compared to the phosphorylation of dGDP. It is worth
mentioning that even though DXG–DP is phosphorylated
by human NDP kinase at an efficiency that is only 0.1% of
dGDP, the inherent high activity of this enzyme warrants its
important role in the activation of DXG.
DXG–MP was enzymatically inefficient suggests that
the rate-limiting step in the DXG phosphorylation pathway
is the addition of the first phosphate. A different HPLC
profile was observed for cells treated with both DXG and
MPA and the level of metabolites followed the order of
DXG–TP = DXG–DP > DXG–MP. However, it should be
interpreted with caution since cell growth was inhibited by
MPA at the level tested.
Direct comparison of the catalytic efficiency (kcat/Km)
between 50-NT and GMP kinase indicates that the second
step of phosphorylation is 23-fold less efficient than the
first step, and therefore, the rate-limiting step in the
phosphorylation pathway is not the conversion of DXG
to DXG–MP. This apparent contradiction between the
enzymatic study and the cellular assays could be due to
the differences in the concentration of each enzyme and
possible involvement of other metabolic enzymes in cell.
Our enzymatic and cellular study demonstrated that high
Km 50-NT, GMP kinase, creatine kinase, and NDP kinase
are likely to play important roles in the phosphorylation
pathway of DXG (Fig. 2). Enzymes like dGK, 3-PG kinase,
and pyruvate kinase could also contribute to the activation
of DXG. This study also provides valuable information on
the phosphorylation pathway of dGuo analogs, which is
currently limited compared to the data available on the
phosphorylation pathway of other nucleoside analogs.
The cellular anabolism profile of a nucleoside analog
provides useful information in determining the rate-limit-
ing step(s) involved in its activation pathway. Previously,
we reported the intracellular level of DXG–TP to be 4.4
and 1.8 pmole/106 cells, respectively, in CEM and stimu-
lated PBMC cells after 24 h of incubation with 5 mM
[8-3H]-DXG [2]. However, later studies using a highly
sensitive Liquid Chromatography Mass Spectrometry
(LC-MS-MS) method indicated that the level of DXG–
TP was only 0.016 pmole/106 for stimulated PBMC cells
after 24 h incubation with 5 mM of DXG (personal com-
munication with Dr. Robert St. Claire III, Gilead
Sciences). In this study, we re-investigated the anabolism
of DXG using [8-14C]-DXG and found that the levels of
DXG metabolites were similar to the levels detected by
LC-MS-MS. The identities of the DXG metabolites were
further confirmed by their dramatic increase after the cells
were treated with MPA, a known stimulator of DXG
anabolism [9]. The high level of DXG metabolites
reported in our previous study is likely due to contaminant
of [8-3H]-guanine and inadequate HPLC resolution of
[8-3H]-DXG–TP and [8-3H]-GTP (anabolite of [8-3H]-
guanine). In this assay, we also detected radio-labeled
guanine nucleotides, but they were sufficiently separated
from the DXG-nucleotides. In the presence of 5 mM MPA,
the level of DXG–TP was increased 19-fold. This is
consistent with the enhancing effect of MPA on DXG
anti-viral activity and its stimulatory effect on DXG–TP
levels observed in PBMC [9,46]. MPA inhibits IMP dehy-
drogenase and results in increased intracellular IMP levels
and decreased GMP levels. Since IMP is the phosphate
donor in the phosphorylation of DXG by high Km 50-NT,
the observed enhancing effect of MPA further support the
hypothesis that high Km 50-NT is one of the key enzymes
for DXG activation. Furthermore, decreased GMP synth-
esis by MPA led to reduced dGTP levels and could
augment the activation and anti-viral activity of analogs
of dGuo [9]. The fact that DXG–TP was the primary
metabolite in cell culture and conversion of DXG to
Acknowledgements
We thank Dr. Staffan Eriksson for the generous gifts of
human recombinant dGK and dCK enzymes. We also
thank Dr. Josef Spychala and Dr. Beverly Mitchell for
providing the clone for high Km 50-NT. The authors are
especially grateful for the insightful discussion with Dr.
Josef Spychala, Dr. Wayne Miller, Dr. Robert Hart, and
Michelle Brosnan-Cook. DDB was recipient of a grant
from ANRS (Agence Nationale de Recherche contre le
Sida).
References
[1] Stein DS, Moore KHP. Phosphorylation of nucleoside analog
antiretrovirals: a review for clinicians. Pharmacol Ther 2001;21:
11–34.
[2] Furman PA, Jeffery J, Kiefer L, Feng JY, Anderson KS, Borroto-Esoda
K, et al. Mechanism of action of 1-b-D-2,6-diaminopurine dioxolane, a
prodrug of the human immnuodeficiency virus type 1 inhibitor 1-b-D-
dioxolane guanosine. Antimicro Agents Chemother 2001;45:158–65.

J.Y. Feng et al. / Biochemical Pharmacology 68 (2004) 1879–1888
1887
[3] Wang LH, Bigley JW, Brosnan-Cook M, Sista ND, Rousseau F. The
disposition of DAPD and it active metabolite DXG in therapy naive
and experienced HIV-infected subjects (Abstract 752). In: Proceedings
of the 8th Conference on Retroviruses and Opportunistic Infections. .
Chicago, IL, February 4–8, 2001; Foundation for Retrovirology and
Human Health; 2001. p. 272.
[22] Bourdais J, Biondi R, Sarfati S, Guerreiro C, lascu I, Janin J, et al.
Cellular phosphorylation of anti-HIV nucleosides. J Biol Chem
1996;271:7887–90.
[23] Krishnan P, Fu Q, Lam W, Liou J-Y, Dutschman G, Cheng Y-C.
Phosphorylation of pyrimidine deoxynucleoside analog diphosphates.
J Biol Chem 2002;277:5453–9.
[4] Mewshaw JP, Myrick FT, Wakefield DA, Hooper BJ, Harris JL,
McCreedy B, et al. Dioxolane guanosine, the active form of the
prodrug diaminopurine dioxolane, is a potent inhibitor of drug-resis-
tant HIV-1 isolates from patients failing standard nucleoside therapy. J
AIDS 2002;29:11–20.
[24] Miller WH, Miller LR. Phosphorylation of acyclovir diphosphate by
cellular enzymes. Biochem Pharmacol 1982;31:3879–84.
[25] Mori N, Singer-Sam J, Riggs AD. Evolutionary conservation of the
substrate-binding cleft of phosphoglycerate kinases. FEBS Lett
1986;204:313–7.
[5] Faletto MB, Miller WH, Garvey EP, St Clair MH, Daluge SM, Good
SS. Unique intracellular activation of the potent anti-human immu-
nodeficiency virus agent 1592U89. Antimicrob Agents Chemother
1997;41:1099–107.
[26] Wyss M, Kaddurah-Daouk R. Creatine and creatine metabolism.
Physiol Rev 2000;80:1107–213.
[27] Cihlar T, Chen MS. Identification of enzymes catalyzing two-step
phosphorylation of cidofovir and the effect of cytomegalovirus
infection on their activities in host cells. Mol Pharmacol 1996;50:
1502–10.
[6] Miller WH, Daluge SM, Garvey EP, Hopkins S, Reardon JE, Boyd FL,
et al. Phosphorylation of carbovir enantiomers by cellualr enzymes
determines the stereoselectivity of antiviral activity. J Biol Chem
1992;267:21220–4.
[28] Merta A, Votruba I, Jindrich J, Holy A, Cihlar T, Rosenberg I, et al.
Phosphorylation of 9-(2-phosphonomethoxyethyl)adenine and 9-(S)-
(3-hydroxy-2-phosphonomethoxypropyl)adenine by AMP(dAMP)
kinase from L1210 cells. Biochem Pharmacol 1992;44:2067–77.
[29] Parker WB, Shaddix SC, Vince R, Bennett LLJ. Lack of mitochondrial
toxicity in CEM cells treated with carbovir. Antiviral Res 1997;34:
131–6.
[7] Miller WH, Miller LR. Phosphorylation of acyclovir (acycloguano-
sine) monophosphate by GMP kinase. J Biol Chem 1980;255:7204–7.
[8] Matthews T, Boehme R. Antiviral activity and mechanism of action of
ganciclovir. Rev Infect Dis 1988;10:490–4.
[9] Myrick F, St Clair R, Hart R, Borroto-Esoda K. Effect of mycophenolic
acid on intracellular dioxolane guanosine triphosphate levels in PBMC
(Abstract #43). In: Proceedings of the 16th International Conference on
Antiviral Research. . Savannah, GA, April 27–May 1, 2003; 2003.
[10] Eriksson S, Wang L. The role of the cellular deoxynucleoside kinases
in activation of nucleoside analogs used in chemotherapy. In: Chu CK,
editor. Recent advances in nucleosides: chemistry and chemotherapy.
New York: Elsevier; 2002.
[30] Merkenschlager M, Kardamakis D, Rawle FC, Spurr N, Beverly PCL.
Rate of incorporation of radio-labelled nucleosides dose not necessa-
rily reflect the metabolite state of cells in culture: effects of latent
mycoplasma contamination. Immunology 1988;63:125–31.
[31] Wang J, Choudhury D, Chattopadhyaya J, Eriksson S. Stereoisomeric
selectivity of human deoxyribonucleoside kinases. Biochemistry
1999;38:16993–9.
[11] Zhu C, Johansson M, Permert J, Karlsson A. Phosphorylation of
anticancer nucleoside analogs by human mitochondrial deoxyguano-
sine kinase. Biochem Pharmacol 1998;56:1035–40.
[32] Schneider B, Xu YW, Sellam O, Sarfati R, Janin J, Veron M, et al. Pre-
steady state of reaction of nucleoside diphosphate kinase with anti-
HIV nucleotides. J Biol Chem 1998;273:11491–7.
[12] Spychala J, Chen V, Oka J, Mitchell BS. ATP and phosphate reci-
procally affect subunit association of human recombinant high Km
50-nucleotidase: role for the C-terminal polyglutamic acid tract in
subunit association and catalytic activity. Eur J Biochem 1999;259:
851–8.
[33] Kierdaszuk B, Bohman C, Ullman B, Eriksson S. Substrate specificity
of human deoxycytidine kinase toward antiviral 20,30-dideoxynucleo-
side analogs. Biochem Pharm 1992;43:197–206.
[34] Gerhardt W. Creatine kinase. Bergmeyer HU, Bergmeyer J, Grabl M,
editors. Methods of enzymatic analysis. Deerfield Beach, FL: Verlag
Chemie; 1983. p. 509.
[13] Arner ESJ. Eriksson S. Mammalian deoxyribonucleoside kinases.
Pharmacol Ther 1995;67:155–86.
[35] Miller HW, Beauchamp LM, Meade E, Reardon JE, Biron KK, Smith
AA, et al. Phosphorylation of ganciclovir phosphonate by cellular
GMP kinase determines the stereoselectivity of anti-human cytome-
galovirus activity. Nucleosides Nucleotides Nucleic Acids 2000;19:
341–56.
[14] Brady WA, Kokoris MS, Fitzgibbon M, Black ME. Cloning, char-
acterization, and modeling of mouse and human guanylate kinases. J
Biol Chem 1996;271:16734–40.
[15] Sekulic N, Shuvalova L, Spangenberg O, Konrad M, Lavie A. Strutural
characterization of the closed conformation of mouse guanylate
kinase. J Biol Chem 2002;277:30236–43.
[36] Puck TT, Cieciura SJ, Robinson A. Genetics of somatic mammalian
cells III long-term cultivation of euploid cells from human and animal
subjects. J Exp Med 1958;108:945–56.
[16] Strominger JL, Heppel LA, Maxwell ES. Nucleoside monophosphate
kinases: I. Transphophorylation between adenosine triphosphate and
nucleoside monophosphates. Biochim Biophys Acta 1959;32:412.
[17] Noda L. Nucleoside triphosphate-nucleoside monophosphokinases.
New York: Academic Press; 1962.
[37] Parker WB, Shaddix SC, Bowdon BJ, Rose LM, Vince R, Shannon
WM, et al. Metabolism of carbovir, a potent inhibitor of human
immunodeficiency virus type 1, and its effects on cellular metabolism.
Antimicrob Agents Chemother 1993;37:1004–9.
[18] Van Rompay AR, Johansson M, Karlsson A. Identification of a novel
human adenylate kinase: cDNA cloning, expression analysis, chromo-
some localization and characterization of the recombinant protein. Eur
J Biochem 1999;261:509–16.
[38] Parker WB, Allan PW, Niwas S, Montgomery JA, Bennett LLJ. Effect
of 9-benzyl-9-deazaguanine, a potent inhibitor of purine nucleoside
phosphorylase, on the cytotoxicity and metabolism of 6-thio-20-deoxy-
guanosine. Cancer Res 1994;54:1742–5.
[19] Henderson JF, Paterson ARP. Deoxyribonulceoside metabolism. In:
Nucleotide metabolism: an introduction. New York: Academic Press;
1973 (pp. 65–67, Chapter 14).
[39] Kreimeyer A, Schneider B, Sarfati R, Faraj A, Sommadossi JP, Veron
M, et al. NDP kinase reactivity towards 3TC nucleotides. Antiviral
Res 2001;2001:147–56.
[20] Schneider B, Biondi R, Sarfati R, Agou F, Guerreiro C, Deville-Bonne
D, et al. The mechanism of phosphorylation of anti-HIV D4T by
nucleoside diphosphate kinase. Mol Pharmacol 2000;57:948–53.
[21] Gallois-Montbrun S, Schneider B, Chen Y, Giacomoni-Fernandes V,
Mulard L, Morera S, et al. Improving nucleoside diphosphate kinase
for antiviral nucleotide analogs activation. J Biol Chem 2002;277:
39953–9.
[40] Sjonerg Herrstrom A, Wang L, Eriksson S. Substrate specificity of
human recombinant mitochondrial deoxyguanosine kinase with cyto-
solic and antiviral purine and pyrimidine analogs. Mol Pharmacol
1998;53:270–3.
[41] Sjonerg Herrstrom A, Wang L, Eriksson S. Antiviral guanosine
analogs as substrate for deoxyguanosine kinase: implication for
chemotherapy. Antimicrob Agents Chemother 2001;45.

1888
J.Y. Feng et al. / Biochemical Pharmacology 68 (2004) 1879–1888
[42] Keller PM, McKee SA, Fyfe JA. Cytoplasmic 50-nucleotidase cata-
lyzes acyclovir phosphorylation. J Biol Chem 1985;260:8664–7.
[43] Johnson MA, Fridland A. Phosphorylation of 20,30-dideoxyinosine by
cytosolic 50-nulceotidase of human lymphoid cells. Mol Pharmacol
1989;36:291–5.
transferase and nucleotidase activities. Arch Biochem Biophys 1994;
312:75–80.
[46] Borroto-Esoda K, Myrick F, Feng JF, Jeffrey JL, Furman PA.
In vitro combination of amdoxovir (DAPD) and the inosine mono-
phosphate dehydrogenase inhibitors mycophenolic acid and ribavirin
demonstrates potent activity against wild type and drug resistant
variants of HIV-1. Antimicrob Agents Chemother 2004, in press.
[47] Traut T. Physiological concentrations of purines and pyrimidines. Mol
Cell Biochem 1994;140:1–22.
[44] Saunders PP, Spindler CD, Tan MT, Alvarez E, Robins RK. Tiazofurin
is phosphorylated by three enzymes from Chinese hamster ovary cells.
Cancer Res 1990;50:5269–74.
[45] Pesi R, Turriani M, Allegrini S, Scolozzi C, Camici M, Ipata PL, et al.
The bifunctional cytosolic 50-nucleotidase: regulation of the phospho-