Nucleoside diphosphate kinase and the activation of antiviral phosphonate analogs of nucleotides: Binding mode and phosphorylation of tenofovir derivatives

Article abstract of DOI:10.1080/15257770903155899

Tenofovir is an acyclic phosphonate analog of deoxyadenylate used in AIDS and hepatitis B therapy. We find that tenofovir diphosphate, its active form, can be produced by human nucleoside diphosphate kinase (NDPK), but with low efficiency, and that creati

Full text of DOI:10.1080/15257770903155899

This article was downloaded by: [University of Memphis]
On: 08 August 2012, At: 06:38
Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered
office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK
Nucleosides, Nucleotides and Nucleic
Acids
Publication details, including instructions for authors and
subscription information:
http://www.tandfonline.com/loi/lncn20
Nucleoside Diphosphate Kinase and the
Activation of Antiviral Phosphonate
Analogs of Nucleotides: Binding Mode
and Phosphorylation of Tenofovir
Derivatives
a
b
c
Kerstin Koch , Yuxing Chen , Joy Y. Feng , Katyna Borroto-Esoda
c
d
a
e
,
Dominique Deville-Bonne , Joël Janin & Solange Moréra
a
Yeast Structural Genomics, IBBMC UMR 8619 CNRS, Université
Paris-Sud, Orsay, France
b
Hefei National Laboratory for Physical Sciences at Microscale, and
School of Life Sciences, University of Science and Technology of
China, Hefei, Anhui, China
c
Gilead Sciences Inc., Foster City, California, USA
d
Enzymologie Moléculaire, UR4 Université Pierre et Marie Curie,
Paris, France
e
Laboratoire d’Enzymologie et Biochimie Structurales (LEBS), UPR
3082 CNRS, Gif-sur-Yvette, France
Version of record first published: 27 Aug 2009
To cite this article: Kerstin Koch, Yuxing Chen, Joy Y. Feng, Katyna Borroto-Esoda, Dominique Deville-
Bonne, Joël Janin & Solange Moréra (2009): Nucleoside Diphosphate Kinase and the Activation
of Antiviral Phosphonate Analogs of Nucleotides: Binding Mode and Phosphorylation of Tenofovir
Derivatives, Nucleosides, Nucleotides and Nucleic Acids, 28:8, 776-792
To link to this article: http://dx.doi.org/10.1080/15257770903155899
PLEASE SCROLL DOWN FOR ARTICLE
Full terms and conditions of use: http://www.tandfonline.com/page/terms-and-conditions
This article may be used for research, teaching, and private study purposes. Any
substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing,
systematic supply, or distribution in any form to anyone is expressly forbidden.

The publisher does not give any warranty express or implied or make any representation
that the contents will be complete or accurate or up to date. The accuracy of any
instructions, formulae, and drug doses should be independently verified with primary
sources. The publisher shall not be liable for any loss, actions, claims, proceedings,
demand, or costs or damages whatsoever or howsoever caused arising directly or
indirectly in connection with or arising out of the use of this material.

Nucleosides, Nucleotides and Nucleic Acids, 28:776–792, 2009
Copyright ꢀC Taylor & Francis Group, LLC
ISSN: 1525-7770 print / 1532-2335 online
DOI: 10.1080/15257770903155899
NUCLEOSIDE DIPHOSPHATE KINASE AND THE ACTIVATION OF
ANTIVIRAL PHOSPHONATE ANALOGS OF NUCLEOTIDES: BINDING
MODE AND PHOSPHORYLATION OF TENOFOVIR DERIVATIVES
1
2
3
3
Kerstin Koch, Yuxing Chen, Joy Y. Feng, Katyna Borroto-Esoda,
4
1
5
Dominique Deville-Bonne, Jo e¨ l Janin, and Solange Mor e´ ra
1
Yeast Structural Genomics, IBBMC UMR 8619 CNRS, Universit e´ Paris-Sud, Orsay, France
Hefei National Laboratory for Physical Sciences at Microscale, and School of Life Sciences,
2
University of Science and Technology of China, Hefei, Anhui, China
3
Gilead Sciences Inc., Foster City, California, USA
Enzymologie Mol e´ culaire, UR4 Universit e´ Pierre et Marie Curie, Paris, France
Laboratoire d’Enzymologie et Biochimie Structurales (LEBS), UPR 3082 CNRS,
4
5
Gif-sur-Yvette, France
2
Tenofovir is an acyclic phosphonate analog of deoxyadenylate used in AIDS and hepatitis B
therapy. We find that tenofovir diphosphate, its active form, can be produced by human nucleoside
diphosphate kinase (NDPK), but with low efficiency, and that creatine kinase is significantly more
active. The 1.65 A˚ x-ray structure of NDPK in complex with tenofovir mono- and diphosphate
shows that the analogs bind at the same site as natural nucleotides, but in a different conformation,
and make only a subset of the Van der Waals and polar interactions made by natural substrates,
consistent with their comparatively low affinity for the enzyme.
Keywords Nucleoside reverse transcriptase inhibitor; nucleotide binding; nucleotide
conformation; phosphorylation
Received 15 April 2009; accepted 17 June 2009.
The authors thank Dr. J e´ r oˆ me Deval (AFMB, Marseille, France) for helpful discussions; Dr. Michael
D. Miller (Gilead Science, Foster City, CA, USA) for providing tenofovir derivatives and editing the
manuscript; Dr. Donna Shewach and William Parker (Gilead Sciences, Durham, NC, USA) for their
expertise and suggestions on HPLC analysis of nucleotide analogs; and Dr. Fran c¸ ois Delfaud (MEDIT
S.A., Palaiseau, France) for providing access to proprietary software in the frame of the POPS program
of the System@tic Paris-Region Cluster. This work was supported by Conseil G e´ n e´ ral de l’Essonne (KK
and JJ), and Agence Nationale pour la Recherche contre le SIDA (SGM and DDB).
Address correspondence to Solange Mor e´ ra, Laboratoire d’Enzymologie et Biochimie Structurales
(LEBS), UPR 3082 CNRS, 91198-Gif-sur-Yvette, France. E-mail: morera@lebs.cnrs-gif.fr
7
76

Tenofovir Activation by Nucleoside Diphosphate Kinase
. INTRODUCTION
777
1
Nucleoside reverse transcriptase inhibitors (NRTI) are nucleoside
ꢁ
analogs that lack a 3 -hydroxyl group. Compounds such as azidothymidine
(AZT) and dideoxynucleosides are potent inhibitors of HIV replication
and play an essential part in antiretroviral therapies. However, they are
only prodrugs that must be converted to the fully phosphorylated forms
[
1]
in order to be incorporated by reverse transcriptase.
The conversion
relies on cellular kinases, but the analogs are generally poor substrates
for these enzymes, with the consequence that the active forms remain at
low concentration relative to the natural deoxynucleoside triphosphates
(dNTP) substrates, which greatly affects their antiviral potency. Acyclic
nucleoside phosphonate analogs have become a major class of antiviral nu-
cleoside derivatives.[ Among them, tenofovir (also called PMPA: [R]-9-[2-
2]
(phosphonomethoxy)propyl]adenine), administrated clinically as tenofovir
disoproxil fumarate to enhance cellular availability, has proved to be well
tolerated and efficient in combination with other drugs in large scale HIV
studies, and it is also approved for treatment against hepatitis B virus.[4]
Remarkably, tenofovir retains potency against HIV strains carrying several
of the mutations that cause resistance to other NRTI, and resistance to
[3]
[
5,6]
tenofovir itself appears to be comparatively difficult to select;
be fully suppressed with an α−borano analog.
it may even
[
7]
Tenofovir is an analog of deoxyadenylate (dAMP) in which the adenine
base is linked to an acyclic methoxypropyl group instead of a deoxyri-
bose, and the phosphonate group replaces the α-phosphate (Figure 1).
Its diphosphate derivative binds HIV reverse transcriptase like the natural
substrate dATP,[ and is incorporated by the enzyme while base-pairing
with uracil or thymine, causing termination of the DNA chain. In the cell,
tenofovir behaves as an analog of dAMP, and the monophosphate as an
analog of dADP. The active diphosphate form is found in abundance in
dendritic and Langherans cells, the first cell types targeted by a primary
HIV infection.[ The activation process requires a kinase that can add
a γ -phosphate onto the β-phosphate of dADP. Nucleoside diphosphate
kinase (NDPK), a ubiquitous enzyme, efficiently performs that reaction on
all natural NDP and dNDP, but it is extremely poor at phosphorylating
8]
9]
FIGURE 1 Chemical formula of tenofovir.

7
78
K. Koch et al.
ꢁ
NRTI derivatives, because the 3 OH of the substrate sugar moiety plays a
major role in its catalytic mechanism.[
10,11]
As a result, the phosphorylation
4
5
of dideoxyADP and of AZT diphosphate by NDPK is 10 or 10 -time
slower than for dADP or TDP,[
12,13]
which explains the low level of the
triphosphate forms found in cells treated with these NRTI. We show here
that this is also true for tenofovir monophosphate, which NDPK can convert
to the diphosphate form, but with a low efficiency relative to a natural
substrate. Furthermore, we demonstrate that other phosphotransferases
including creatine kinase and pyruvate kinase, can produce tenofovir
diphosphate more efficiently, and may therefore contribute to the activation
process.
X-ray structures are available for a number of NDPKs including the two
major human isozymes NDPK-A and B, and for the complexes with several
[
14]
natural substrates and phosphorylated NRTI derivatives.
All of them
show the same protein fold, which assembles into a homohexamer, except
in some bacterial enzymes that are tetramers. The substrate nucleotide
binding site is highly conserved, and the ligands that have been tested in
crystallographic experiments all bind similarly with few exceptions. The
present study describes a complex of tenofovir mono- and diphosphate
with the enzyme from Dictyostelium discoideum (Dd-NDPK). It shows that the
analogs also bind at that site, but in a conformation not found with normal
substrates, and which may contribute to their inefficient processing by the
enzyme.
2. MATERIALS AND METHODS
2
.1. Enzyme Expression and Activity
Recombinant human NDP kinase A (NDPK-A/NM23-H1) and the
H122G mutant of Dictyostelium discoideum NDP kinase (Dd-NDPK) were
expressed in E. coli and purified to homogeneity as described.[
15,16]
Natural
nucleotides were from Sigma. Tenofovir and its mono- and diphosphate
derivatives were synthesized by Trilink Biotechnologies (San Diego, CA,
USA). Human BB, MB, and MM creatine kinases (C9983, C0984, and
C9858), and rabbit muscle pyruvate kinase type II (PI506) were purchased
from Sigma (St. Louis, MO, USA). Human 3-phosphoglycerate kinase was a
generous gift from Dr. Preethi Krishnan and Dr. Y.C. Cheng (Yale University,
New Haven, CT, USA).
The phosphorylation of tenofovir monophosphate by NDPK-A
(
40 µg/mL) was carried out in Tris acetate buffer pH 7.5, 5 mM MgCl ,
2
◦
at 37 C. The donor substrate was ATP (1 mM). Aliquots (100 µl) were
removed every 2 minutes and quenched with 5 µl 3.5–3.8% HCl (final pH
3
–4). The reaction products were subjected to HPLC analysis as previously
[
17]
described.
Since dADP and dATP could not be sufficiently separated

Tenofovir Activation by Nucleoside Diphosphate Kinase
779
on HPLC from the more abundant ADP and ATP, a boronate column
was used to remove ATP and ADP from the reaction mixture prior to
HPLC analysis.[
18]
More than 98% of dADP and dATP were recovered,
and more than 99% of ATP and ADP were removed from the reaction
mixture. All enzymatic reactions involving creatine kinase, pyruvate kinase,
[
17]
and 3-phosphoglycerate kinase were conducted as previously described.
2
.2. Crystallization and X-Ray Structure Determination
Crystals of the H122G mutant protein of Dd-NDPK at 8 mg/ml in the
presence of 10 mM tenofovir diphosphate were obtained in hanging drops
containing 12% (w/v) PEG 1000 and 100 mM Hepes pH 7.5 over pits
consisting of 26% PEG 1000 in the same buffer. Crystals were flash-frozen
in the mother liquor and 20% glycerol. Data collection was carried out
at 100 K on beamline ID14-H1 at the European Synchrotron Radiation
◦
◦
Facility (Grenoble, France); 270 of data were collected in 1 frames, with
˚
1
second per frame exposure. Diffraction intensities to 1.65 A resolution
were evaluated with the program MOSFLM
[
19]
and further processed using
the CCP4 (Collaborative Computational Project Number 4, 1994) program
suite. Data collection and processing statistics are given in Table 1.
The crystals contain a hexamer in their asymmetric unit and, as they
are isomorphous to previously solved structures of Dd-NDPK, preliminary
phases could be calculated using coordinates from PDB entry 1S5Z.[
20]
TABLE 1 Crystallographic data and refinement statistics
Space group
P21
◦
Unit-cell parameters ( A˚ )
a = 69.4, b = 104.6, c = 69.2, β = 118
Resolution ( A˚ )
20–1.65 (1.74–1.65)
424591 (61272)
104433 (15223)
7.5 (35.7)
99.8 (100)
7.6 (1.8)
No. of observed reflections
No. of unique reflections
Rsym (%)a
Completeness (%)
I/σ
b
Rcryst (%)
18.1
20.5
0.004
1.29
Rfree (%)c
rms bond deviation ( A˚ )
rms angle deviation ( )
Average B ( A˚ )
◦
2
Protein atoms
16.0
38.2
26.7
24.4
Tenofovir diphosphate
Tenofovir monophosphate
Solvent
Values for the highest resolution shell are in parentheses
a
Rsym = ꢀhklꢀi|Ii(hkl) -<I(hkl)>|/ꢀhklꢀIi (hkl), where Ii(hkl) is the i th
observed amplitude of reflection hkl and <I(hkl)> is the mean amplitude for
all observations i of reflection hkl.
b
Rcryst = ꢀ ||Fobs|—|Fcalc||/ꢀ |Fobs|
c
5
% of the data were set aside for free R-factor calculation.

7
80
K. Koch et al.
The resulting electron density map showed the presence of a ligand in
all six subunits. After minor modifications of the model, refinement was
performed using CNS,[ leading to the statistics reported in Table 1. The
atomic coordinates and structure factors have been deposited in the Protein
Data Bank (PDB, http://www.rcsb.org) as entry 3FKB.
21]
2.3. Binding Site Comparisons
We used the MED-SuMo software to compare ligand binding sites
in NDPK and other proteins in the Protein Data Bank. This software,
developed by MEDIT SA (Palaiseau, France) implements the SuMo
[
22,23]
algorithm
to identify sets of surface chemical features (SCF) on a
protein. A SCF is a chemical group with a character relevant to ligand
binding, such as aromaticity, hydrophobicity, electric charge, or H-bond
donor or acceptor capacity. SCFs closest to each other in space are assem-
bled into triangles and a graph is built by connecting adjacent triangles
that share at least one edge. A graph matching algorithm can then be
used to efficiently compare two protein surfaces or ligand binding sites,
and MED-SuMo generates three-dimensional superimpositions based on
the match. A score is calculated by giving each matched SCF a weight that
depends on its chemical character and its distance to neighbouring SCFs[22]
and by summing the weights.
3. RESULTS
3.1. Tenofovir Monophosphate as a Substrate of NDPK and Other
Kinases
Purified recombinant human NDPK-A was able to phosphorylate teno-
fovir monophosphate in vitro with ATP as the phosphate donor. The steady-
state kinetic parameters for that reaction are reported in Table 2, and
they can be compared to those obtained with dADP, a natural substrate,
using the same protein sample. The K values are similar, but the turnover
m
is much lower with the analog, and the catalytic efficiency measured by
k /K is three orders of magnitude less. Thus, tenofovir monophosphate
cat
m
is a substrate of NDPK-A, albeit a poor one.
When other kinases were tested, no phosphorylation of tenofovir
ꢁ
monophosphate was detected using human 3 -phosphoglycerate kinase, but
rabbit muscle pyruvate kinase displayed about the same efficiency as NDPK-
A in terms of k /K , and all three isozymes of human creatine kinase
cat
m
(BB, MB, and MM) proved to be orders of magnitude more active. Table 2
shows that the BB isozyme has a high kcat with both dADP and tenofovir
5
−1 −1
monophosphate, and a k /K ratio above 3 × 10 M s for the analog
cat
m

Tenofovir Activation by Nucleoside Diphosphate Kinase
781
TABLE 2 Steady-state kinetic parameters for the enzymatic phosphorylation of tenofovir
monophosphate
kcat (s 1
−
)
kcat/Km (M .s
−1 −1
)
Ratioa
Enzyme Acceptor substrate
Km (mM)
NDPK-Ab
dADP
5
4
4
0.18 ± 0.03
65 ± 3
3.6 × 10
Tenofovir monophosphate 0.29 ± 0.03
0.12 ± 0.01
410
880
150
Pyruvate kinasec
dADP
3.0 ± 0.4
11 ± 3
280 ± 20
6.8 ± 2.3
9.3 × 10
Tenofovir monophosphate
Phosphoglycerate kinase
dADP
Tenofovir monophosphate
Creatine kinasef
dADP
620
d
2.0 ± 0.3
103 ± 7
5.2 × 10
e
e
ND
ND
-
6
5
0.08 ± 0.01 1030 ± 30
1.2 ± 0.20 400 ± 30
13 × 10
Tenofovir monophosphate
3.3 × 10
39
a
Ratio of kcat/Km for dADP versus the analog.
The donor substrate is ATP (1 mM).
Rabbit muscle pyruvate kinase type II (Sigma PI506); the donor substrate is phos-
b
c
pho(enol)pyruvate (4 mM).
dHuman 3-phosphoglycerate kinase; the donor substrate is DL-glyceraldehyde 3-phosphate
(
4 mM).
e
No product was detected.
f
Human BB isozyme (Sigma C9983); the MM and MB isozymes yielded similar results; the
donor substrate is creatine phosphate (20 mM).
in spite of its higher K . Similar k /K ratios were obtained with the MB
m
cat
m
and BB isozymes with tenofovir monophosphate as substrate.
3
.2. NDPK Structure in the Crystalline Complex
The protein used for crystallization was the H122G mutant of Dic-
tyostelium discoideum NDPK. Dd-NDPK, the first NDPK to have a crystal
structure determined,[
24]
is easy to overexpress and crystallize, and has
over 60% sequence identity with human NDPK-A and NDPK-B, which are
the product of the Nme1/Nm23-H1 and Nme2/Nm23-H2 genes, and the two
major isozymes found in human cells. All three proteins are very close in
structure, especially at the substrate binding site, and they have similar
enzymatic characteristics.[
10]
The NDPK catalytic mechanism operates in
two steps: the γ −phosphate group of the donor substrate (usually ATP) is
transferred first onto a histidine (His 122 in Dd-NDPK, His 118 in NDPK-A),
and then from the phosphohistidine onto the acceptor nucleotide that
replaces the donor and occupies the same site. The H122G substitution
prevents the transfer altogether, which makes the mutant useful when
structural studies are done in the presence of ATP or another donor
substrate that would transfer part of its phosphate to the histidine, resulting
in a mixture of chemical species at the active site. The substitution has no

7
82
K. Koch et al.
observable structural effect, and the H122G protein efficiently phosphory-
lates externally supplied imidazole.[
16]
In the crystals of the complex, the protein structure is well defined by
˚
the 1.65 A resolution electron density for 150 of the 155 residues in the
polypeptide chains, except for N-terminal residues 1–5. These residues are
disordered like in other crystal structures of Dd-NDPK. The asymmetric
unit contains six subunits labeled A to F for convenience. Each provides
independent information, and differences that affect the bound ligand
are described below. Nevertheless, the subunits have essentially identi-
cal structures, with a RMSD (root-mean-square distance after least-square
˚
superimposition) of 0.3–0.6 A over all the 150 Cα. They are also very
similar to wildtype Dd-NDPK subunits in complex with natural nucleotides.
˚
For instance, the Cα RMSD is 0.4–0.7 A with subunits of the ADP-BeF3
complex (PDB entry 2BEF), where BeF (beryllium fluoride) sits in between
3
the β-phosphate of ADP and the imidazole group of His 122, mimicking the
γ -phosphate of ATP.[ Thus, the absence of the H122 side chain and the
presence of a nucleotide analog in the present structure have little effect on
the protein conformation.
25]
3.3. The Bound Ligand
Electron density indicates that the ligand binding site is occupied in all
six subunits. As shown in Figure 2 for subunit A, the density fits tenofovir
monophosphate but the expected position of the second phosphate is
empty. Thus, we built the ligand as the monophosphate in subunit A and
in four other subunits. Well-defined density also indicates the presence
+
+
of an Mg
ion that ligates oxygens of the phosphonate, of the first
phosphate (called below the β−phosphate by analogy to natural substrates),
and of four water molecules in the octahedral geometry that is usual
+
+
for Mg . Subunit D contains electron density that extends beyond the
β−phosphate, and in which either a γ −phosphate or inorganic phosphate
can be placed. We chose to build tenofovir diphosphate in that subunit, but
the presence of a mixture containing inorganic phosphate and tenofovir
mono- and diphosphate cannot be excluded. Thus, the electron density
seen at the six ligand binding sites implies that tenofovir diphosphate has
been partly hydrolyzed on the time scale of the crystallographic experiment,
a plausible event given that the H122G mutant has low, but detectable
ATPase activity.[
16]
Table 3 reports the polar interactions and Van der Waals contacts that
tenofovir monophosphate makes in subunit A, and the diphosphate in
subunit D. The protein residues involved are shown in Figure 2. Whereas the
base and the linker make only nonpolar contacts, one of the phosphonate
oxygens receives two H-bonds from Lys 16 and Asn 119, and the phosphate
groups make H-bonds or salt bridges with the side chains of His 59, Arg 92,

Tenofovir Activation by Nucleoside Diphosphate Kinase
783
FIGURE 2 Tenofovir monophosphate bound to Dd-NDPK. The 2Fo-Fc electron density for tenofovir
monophosphate in subunit A is contoured at 1.5 σ. The ligand and the side chains in contact with it are
++
drawn as sticks colored by atom type, and the Mg ion, as a yellow ball.
Thr 98, and Arg 109. These residues are the same that interact with the
[
26]
phosphates of ADP in wildtype Dd-NDPK.
Thus, the analog occupies
essentially the same site. Nevertheless, Figure 3 shows that it binds in a
quite different conformation. The adenine base is sandwiched between
the side chains of Phe 64 and Val 116 as in NDPK-bound ADP, but it is
oriented in the opposite way. The methoxypropyl linker is well defined in
the electron density and it follows a different path from the ribose. The Pα
˚
atoms of ADP and of the tenofovir phosphonate are 3.5 A apart, but the Pβ
+
+
˚
atoms are within 0.3 A of each other, and Mg is bound in the same way,
offering the possibility for a phosphate group on His 122 to be transferred
to the tenofovir β-phosphate. However, the phosphonate group of the
bound analog overlaps with the expected position of the phosphohistidine
phosphate, which is where BeF is located in the ADP-BeF complex and,
3
3
therefore, the mode of binding seen in the crystal cannot be productive.
Figure 4 compares tenofovir monophosphate in subunit A to the
residual diphosphate seen in subunit D, and to tenofovir diphosphate
in HIV reverse transcriptase (PDB entry 1T05), superimposing the

7
84
K. Koch et al.
TABLE 3 Interactions made by phosphorylated tenofovir derivatives
in Dd-NDPK
Liganda
Protein
Distanceb ( A˚ )
Adenine base
C6
C8
N9
Phe 64 CB
Leu 68 CD2
Val 116 CG1
Thr 98 CG2
His 59 NE2
Phe 64 CE1
Leu 68 CD2
Tyr 56 CE1
Lys 16 NZ
Asn 119 ND2
Arg 109 NH2
Arg 92 NH1
Thr 98 OG1
His 59 NE2
3.5
3.7
3.6
3.5
3.8
3.7
3.8
4.0
2.9
2.9
2.8
3.6
3.0
2.9
ꢁ
Methoxypropyl linker
C6
C8
ꢁ
ꢁ
ꢁ
O9
C9
Phosphonate
Phosphates
O2A
O1B
O3B
O3G
aTenofovir monophosphate in subunit A, γ -phosphate of tenofovir
diphosphate in subunit D.
bNonpolar interactions shorter than 4 A˚ or polar interactions shorter
than 3.6 A˚ .
adenine moiety of the three structures. The conformations of the linker
seen in NDPK and reverse transcriptase are remarkably similar, the phos-
˚
phonate and α-phosphate positions are within 2.5 A of each other, but
the diphosphate moiety of tenofovir diphosphate adopts two orthogonal
orientations relative to the acyclic moiety as the result of P-O bond
rotations.
3
.4. The Ligand Binding Site in NDPK Structures
The present NDPK crystal structure is one of many that have been
determined in the presence of a variety of ligands and yield a wealth of
information on the substrate binding site. At the end of 2008, 65 entries of
the Protein Data Bank described NDPKs from 18 different origins ranging
from human to viral. Thirty-three, including the present one, contain
natural nucleotides or analogs. Table 4 reports their entry codes, the
nature of the ligand and the origin of the protein. The sequence identity
relative to Dd-NDPK is in the range 40% to 66%, with the mammalian
proteins at the top of the range, and the viral sequence at the bottom.
All are hexameric except for the Myxococcus xanthus tetramer (1NLK,
1
NHK), although they display no cooperative binding and obey Michaelian
kinetics.
To analyze the ligand sites, the ADP site in the ADP-BeF3 complex
of Dd-NDPK (2BEF) was taken as a reference, and all other structures
were systematically compared to it using the MED-SuMo software. For that

Tenofovir Activation by Nucleoside Diphosphate Kinase
785
FIGURE 3 The conformation of NDPK-bound ADP and tenofovir diphosphate. ADP (gray sticks) in
the ADP-BeF3 complex with wildtype Dd-NDPK (PDB entry 2BEF) is superimposed with tenofovir
monophosphate (TNM, colored by atom type) in subunit A of the present structure. The beryllium
fluoride moiety is located between the β-phosphate and the active site His 122, mimicking the γ -
phosphate of ATP.[
respectively.
25]
++
The yellow and gray balls are Mg ions bound to TNM and ADP- BeF3
purpose, a site was defined as the set of SuMo objects located within 6
˚
A of any ligand atom. The reference ADP site comprises 86 such objects
that belong to 19 amino acid residues. Table 4 reports the scores of the
comparison to other sites, obtained by summing the weights of matching
SuMo objects, and scaling the reference to 100. The Dd-NDPK structures
that contain ADP achieve very similar scores in the range of 82 to 91.
When other nucleotides or analogs occupy the binding site, the scores
are significantly lower: 74–86, and two entries are far below: 1BUX, where
ꢁ
ꢁ
the ligand is 3 -phospho-adenosine-5 -phospho-sulfate, a potent inhibitor
of NDPK,
[
27]
and 1MN7 where three residues are substituted. The ADP
binding site of human NDPK-A (2HVD) scores 73, less than in Dd-NDPK
due to the loss of interactions made by His 59, changed to Leu 55
in the human protein. The histidine is present in Pyrococcus horikoshii
NDPK (2DYA), yielding a score of 84. NDPKs from other sources and

7
86
K. Koch et al.
FIGURE 4 Phosphorylated tenofovir derivatives bound to NDPK and HIV reverse transcriptase.
Tenofovir monophosphate (colored by atom type) in subunit A of Dd-NDPK is compared to the
diphosphate (blue sticks) present in subunit D and in HIV reverse transcriptase (red sticks; PDB entry
T05[ ) after superimposing the adenine moiety. Two orthogonal orientations are shown. The green
8]
1
++
and blue balls are Mg ions bound to the mono- and diphosphate, respectively.
complexes with other ligands than ADP have scores in the range of 53 to
74.
The ligand site of the present complex implicates only 54 of the 86 SuMo
objects in the reference ADP site, and its score is 69, below all the natural
substrates of Dd-NDPK. The deletion of the His 122 side chain removes
only one SuMo object and has a marginal effect on the score, as shown
by entry 1B4S, which bears the same H122G mutation. Thus, the score of
the tenofovir site reflects its unusual mode of binding. The lowest scores
ꢁ
overall are near 60 for wildtype Dd-NDPK binding 3 -phospho-adenosine-
ꢁ
5
-phospho-sulfate, a ligand that also binds in a different way than natural
[
27]
substrates,
and for mimivirus NDPK in complex with dGDP or TDP. The
viral enzyme is highly divergent from the others in terms of its sequence
and enzymatic properties.[
28]
Nevertheless, the SuMo scores reported in
Table 4 uniquely identify NDPK, whatever the ligand. We used MED-SuMo
to compare all the entries in the PDB to the ADP site in 2BEF. The highest
scores were 20 for the coenzyme site in the human glycerol-3-phosphate
dehydrogenase 1-like protein (1PLA) and 15 for the site occupied by a
coenzyme analog in E. coli thymidylate synthase (1DDU). All other entries
scored below 13.

Tenofovir Activation by Nucleoside Diphosphate Kinase
787
TABLE 4 Comparison of the ligand binding sites in NDPK x-ray structures
PDB
entry
Organism/
Mutation
Ligand
Seq. Id.a SSuMob
Dictyostelium discoideum
ꢁ
ꢁ
ꢁ
2
1
1
1
1
BEF
KDN
B4S
MN9
B99
ADP
ADP
ADP
RTP
FUP
Adenosine-5 -diphosphate BeF3
Adenosine-5 -diphosphate AlF3
Adenosine-5 -diphosphate
100
100
99
99
100
100
91
90
86
84
H122G
H122G
Ribavirin triphosphate
ꢁ
ꢁ
ꢁ
ꢁ
2 ,3 -dideoxy-3 -fluoro-uridine-5 -
diphosphate
ꢁ
1
1
NDP
F6T
ADP
Adenosine-5 -diphosphate
100
100
82
82
ꢁ
ꢁ
TBD 2 -deoxy-thymidine-5 -α-borano-
diphosphate
2 ,3 -dehydro-thymidine-5 -
triphosphate
ꢁ
ꢁ
ꢁ
1F3F
D4T
3AN
H122G
99
80
ꢁ
ꢁ
1
1
1
1
3
1
HIY
S5Z
LWX
NDC
FKBc
BUX
3 -amino-adenosine-5 -diphosphate
100
100
100
100
99
78
77
75
74
69
59
ꢁ
SON Adenosine-5 -phosphonoacetic acid
ꢁ
ꢁ
AZD
TYD
3 azidothymidine-5 -diphosphate
ꢁ
Thymidine-5 -diphosphate
TNM Tenofovir monophosphate
H122G
ꢁ
ꢁ
PPS
3 -phospho-adenosine-5 -phospho-
sulfate
100
ꢁ
ꢁ
1MN7
ABT
3 -azidothymidine-5 -α-borano-
H122G/N119S/F64W
98
38
triphosphate
Mammalian
ꢁ
ꢁ
ꢁ
ꢁ
2
2
1
1
1
HVD
HVE
NUE
ZS6
ADP
ADP
Adenosine-5 -diphosphate
Human A
Human A S120G
Human B
Human 3
Human A
66
61
64
63
64
73
69
69
58
57
Adenosine-5 -diphosphate
GDP Guanosine-5 -diphosphate
ADP
ADP
Adenosine-5 -diphosphate
ꢁ
UCN
Adenosine-5 -diphosphate
H118G/F60W
Bovine retina
1
BE4
PCG
DGI
ADP
Cyclic GMP
64
61
53
62
Plant
S59
Microbial
ꢁ
ꢁ
1
2 -deoxy-guanosine-5 -diphosphate
Arabidopsis thaliana
ꢁ
ꢁ
2DYA
2DXE
2DXD
Adenosine-5 -diphosphate
Pyrococcus horikoshii
Pyrococcus horikoshii
Pyrococcus horikoshii
56
56
56
84
74
70
GDP Guanosine-5 -diphosphate
ANP
Phosphoamino-phosphonic acid
adenylate ester
ꢁ
2AZ3
CDP
cytidine-5 -diphosphate
Halobact. salinarium
Pyrococcus horikoshii
57
56
69
66
2DXF
GNP Phosphoamino-phosphonic acid
guanylate ester
ꢁ
1
3
2
1
1
1
NLK
B6B
B8Q
WKL
NHK
WKK
ADP
DGI
TYD
ADP
Adenosine-5 -diphosphate
Myxococcus xanthus
Mimivirus
Mimivirus
Thermus thermophilus
Myxococcus xanthus
Thermus thermophilus
47
40
40
63
47
63
66
61
58
57
56
53
ꢁ
ꢁ
2 -deoxy-guanosine-5 -diphosphate
ꢁ
ꢁ
Thymidine-5 -diphosphate
Adenosine-5 -diphosphate
CMP cyclic AMP
ꢁ
GDP Guanosine-5 -diphosphate
a
Percentage of identity between aligned sequences.
SuMo score of the comparison to the ADP binding site in entry 2BEF, normalized to 100.
Present structure.
b
c

7
88
K. Koch et al.
4. DISCUSSION
4
.1. NDPK and the Phosphorylation of Tenofovir Monophosphate
NDPK is able to convert tenofovir monophosphate to the diphosphate
−
1 −1
in vitro, but the value, k /K = 410 M s (cited in Table 2), indicates
cat
m
that the phosphorylation is not very efficient. The phosphorylation of
AZT diphosphate by the same human NDPK-A yields a similar value:
−
1 −1 [15]
ꢁ
1
92 M s .
All NRTI derivatives that lack a 3 OH group are poor
substrates of NDPK, which reflects a peculiar feature of that kinase. Natural
ꢁ
nucleotides bind in a conformation that allows the 3 OH of the sugar to
donate an intramolecular H-bond to the oxygen bridging the β- and the
γ -phosphates. This oxygen is the leaving group when a NTP substrate
phosphorylates the active site histidine, and the attacking group when a
NDP substrate dephosphorylates it. The intramolecular H-bond contributes
[
11]
to catalysis by lowering the energy of the transition state of the reaction.
ꢁ
ꢁ
d4T (2 ,3 -dehydro-thymidine) diphosphate is a significantly better substrate
−
1 −1 [15]
than other NRTI derivatives (k /K = 2650 M s ),
and this corre-
cat
m
lates with the observation that, when d4T diphosphate binds, a CH . . . O
bond implicating the 3 CH group of the dehydro sugar ring replaces the
ꢁ
OH . . . O bond present in a normal substrate. Albeit weaker, this H-bond
[
29]
also contributes to stabilizing the transition state.
Although the Km values for tenofovir monophosphate and dADP re-
ported in Table 2 are similar, the affinity of NDPK must be significantly less
for the analog than the natural substrate. The analog is a poor substrate for
which the histidine dephosphorylation step is rate limiting and slow and,
therefore, its dissociation constant K should be close to the observed K
d
m
(
0.29 mM). This is 50 times the K of NDPK-A for ADP (5.8 µM), and 3
d
[
20]
times that for d4T diphosphate (0.1 mM),
but less than for AZT diphos-
[
12]
phate (K = 6 mM)
and for adenosine phosphonoacetate (K = 0.6
m
d
[
30]
mM).
AZT diphosphate and d4T diphosphate bind NDPK in the same
[
31,29]
conformation as natural substrates,
but tenofovir derivatives display
a mode of binding that has never been observed before. The orientation
of the adenine base in tenofovir monophosphate corresponds to the syn
[
32]
conformation seen in cyclic AMP bound to Myxococcus xanthus NDPK;
all other nucleotides or analogs have the base in the anti conformation
when bound to NDPK.[
20]
Yet, the reversal of the base orientation does
not by itself explain the low affinity of tenofovir monophosphate. The
base maintains the same nonpolar interactions with protein groups, and
NDPK is very tolerant there: It accepts purine or pyrimidine bases in
its natural substrates, and even a substituted triazole group in ribavirin
diphosphate.[
33]
On the other hand, the 3 OH of the sugar in a natural
substrate makes polar interactions with the side chains of Lys 16 and Asn
ꢁ
[
26]
119,
in addition to the intramolecular H-bond to the phosphate oxygen.

Tenofovir Activation by Nucleoside Diphosphate Kinase
789
ꢁ
AZT diphosphate, d4T diphosphate, and other analogs that lack a 3 OH,
yet bind in the usual conformation, also lose these interactions and show
a low affinity. Moreover, the substitution of Asn 115 of human NDPK-A,
equivalent to Asn 119 in Dd-NDPK, reduces the catalytic efficiency by an
order of magnitude for natural substrates, while it increases it by the same
ꢁ
[34]
ꢁ
factor for 3 -deoxy analogs.
Thus, the absence of a 3 OH in tenofovir and
the consequent loss of interactions with protein groups is a likely cause of
its low affinity.
4.2. NDPK-Bound Tenofovir Derivatives
The methoxypropyl linker group of tenofovir does not mimic a cyclic
sugar, but rather adopts a low-energy conformation that is also observed
in the very different environment of reverse transcriptase, placing the
phosphonate group in a different position than the α-phosphate of a normal
substrate. However, the position seen in the H122G mutant may not be
allowed in the wildtype protein, because the phosphonate moiety would
be too close to the active site histidine and clash with the phosphate that
should be transferred. There is room for the phosphonate to move from
the observed position closer to the α-phosphate of ADP, and this can be
achieved by rotations around single P-O bonds without displacing the β-
phosphate. During the course of the reaction, tenofovir monophosphate
may therefore adopt a conformation closer to that of natural substrate.
Nevertheless, the present x-ray structure shows that a substrate analog
does not necessarily imitate the natural ligand when it binds to a protein.
Standard molecular modeling methods tend to optimize the fit between
molecules in the same site. Presumably, they would fail to predict the
reversal of the base orientation in tenofovir and the path followed by the
linker from the base to the phosphonate.
4
.3. Alternative Pathways for Tenofovir Activation
The phosphorylated derivatives that are account for the antiviral activity
[
9]
of tenofovir are relatively abundant in human cells. Yet, the first phos-
[
35,36]
phorylation step is very poorly catalyzed by human adenylate kinases,
and the kinetic and structural data reported here make it unlikely that
NDPK is the major player in the second step. Still, the abundance of
these housekeeping enzymes counterbalances their low catalytic efficiency,
whereas phosphorylated acyclic phosphonate analogs are relatively resistant
[
37]
to hydrolysis by nucleotidases
and remarkably stable in the cell with
a half-life of days.[
38,39]
On the other hand, phosphotransferases other
than adenylate kinase and NDPK are able to phosphorylate nucleoside
mono- and diphosphates. Some have been shown to phosphorylate NRTI
derivatives relatively efficiently, for instance phosphoglycerate kinase in

7
90
K. Koch et al.
the particular case of the L-nucleoside analogs.[40,41] While this particular
enzyme is not active on tenofovir monophosphate, we find that pyruvate
kinase is as efficient as NDPK, and creatine kinase much more so. These
enzymes, and others that have yet to be tested, may contribute to the
activation of tenofovir in cells.
4
.3. Conclusion
A prominent outcome of the present study is that a protein site designed
to bind natural nucleotides can easily accommodate a ligand that lacks a
sugar ring and where a phosphonate replaces the α-phosphate. Nucleotide
binding (Gene Oncology ID:166) is one of the most common molecular
functions of proteins encountered in nature. The PDB utility MolecularFunc-
tionTree Search finds it in about one of seven proteins in the PDB, which
offers a wide field for the design of acyclic phosphonate analogs targeting
these proteins. Although their nucleotide binding sites may not all be as
permissive as in NDPK, many should accept such compounds as ligands,
leading to the identification of novel families of substrates, agonists, or
inhibitors for the target proteins.
REFERENCES
1
2
3
4
5
. Stein, D.S.; Moore, K.H.P. Phosphorylation of nucleoside analog antiretrovirals: a review for
clinicians. Pharmacol. Ther. 2001, 21, 11–34.
. De Clercq, E.; Hol y´ , A. Acyclic nucleoside phosphonates: a key class of antiviral drugs. Nat. Rev. Drug
Discov. 2005, 4, 928–940.
. Pozniak, A. Tenofovir: what have over 1 million years of patient experience taught us? Int J. Clin.
Pract. 2008, 62, 1285–1293.
. Delaney, W.E.; Borroto-Esoda, K. Therapy of chronic hepatitis B: trends and developments. Curr. Op.
Pharmacol. 2008, 8, 532–540.
. Srinivas, R.V.; Fridland, A. Antiviral activities of 9-R-2-phosphonomethoxypropyl adenine (PMPA)
and bis(isopropylmethylcarbonyl)PMPA against various drug-resistant human immunodeficiency
virus strains. Antimicrob. Agents. Chemother. 1998, 42, 1484–1487.
6
7
8
. Wainberg, M.A.; Miller, M.D.; Quan, Y.; Salomon, H.; Mulato, A.S.; Lamy, P.D.; Margot, N.A.; Anton,
K.E.; Cherrington, J.M. In vitro selection and characterization of HIV-1 with reduced susceptibility
to PMPA. Antivir Ther. 1999, 4, 87–94.
. Frangeul, A.; Barral, K.; Alvarez, K.; Canard, B. In vitro suppression of K56R reverse transcriptase-
ꢁ
mediated tenofovir and adefovir-5 -diphosphate resistance conferred by the boranophosphonate
derivatives. Antimicrob. Agents Chemother. 2007, 51, 3162–3167.
. Tuske, S.; Sarafianos, S.G.; Clark, A.D. Jr.; Ding, J.; Naeger, L.K.; White, K.L.; Miller, M.D.; Gibbs,
C.S.; Boyer, P.L.; Clark, P.; Wang, G.; Gaffney, B.L.; Jones, R.A.; Jerina, D.M.; Hughes, S.H.; Arnold,
E. Structures of HIV-1 RT-DNA complexes before and after incorporation of the anti-AIDS drug
tenofovir. Nat. Struct. Mol. Biol. 2004, 11, 469–474.
9. Balzarini, J.; Van Herrewege, Y.; Vanham, G. Metabolic activation of nucleoside and nucleotide
reverse transcriptase inhibitors in dendritic and Langerhans cells. AIDS 2002, 16, 2159–2163.
1
0. Lascu I.; Gonin, P. The catalytic mechanism of nucleoside diphosphate kinases. J. Bioenerg. Biomembr.
000, 32, 237–246.
2
11. Janin, J.; Deville-Bonne, D. Nucleoside-diphosphate kinase: structural and kinetic analysis of reaction
pathway and phosphohistidine intermediate. Methods Enzymol. 2002, 354, 118–134.

Tenofovir Activation by Nucleoside Diphosphate Kinase
791
1
2. Bourdais, J.; Biondi, R.; Sarfati, S.; Guerreiro, C.; Lascu, I.; Janin, J.; V e´ ron, M. Cellular phospho-
rylation of anti-HIV nucleosides. Role of nucleoside diphosphate kinase. J. Biol. Chem. 1996, 271,
7887–7890.
13. Schneider, B.; Xu, Y.W.; Sellam, O.; Sarfati, R.; Janin, J.; Veron, M.; Deville-Bonne, D. Pre-steady
state of reaction of nucleoside diphosphate kinase with anti-HIV nucleotides. J. Biol. Chem. 1998,
273, 11491–11497.
1
4. Janin, J.; Dumas, C.; Mor e´ ra, S.; Xu, Y.; Meyer, P.; Chiadmi, M.; Cherfils, J. Three-dimensional
structure of nucleoside diphosphate kinase. J. Bioenerg. Biomembr. 2000, 32, 215–225.
15. Schneider, B.; Biondi, R.; Sarfati, R.; Agou, F.; Guerreiro, C.; Deville-Bonne, D.; Veron, M. The
mechanism of phosphorylation of anti-HIV D4T by nucleoside diphosphate kinase. Mol. Pharmacol.
2000, 57, 948–53.
1
1
1
6. Admiraal, S.J.; Schneider, B.; Meyer, P.; Janin, J.; V e´ ron, M.; Deville-Bonne, D.; Herschlag, D.
Nucleophilic activation by positioning in phosphoryl transfer catalyzed by nucleoside diphosphate
kinase. Biochemistry 1999, 38, 4701–4711.
7. Feng, J.Y.; Parker, W.B.; Krajewski, M.L.; Deville-Bonne, D.; Veron, M.; Krishnan, P.; Cheng, Y-C.;
ꢁ
Borroto-Esoda, K. Anabolism of amdoxovir: phosphorylation of dioxolane guanosine and its 5 -
phosphates by mammalian phosphotransferases. Biochem. Pharmacol. 2004, 68, 1879–1888.
ꢁ
8. Shewach, D. Quantitation of deoxyribonucleoside 5 -triphosphates by a sequential boronate and
anion-exchange high-pressure liquid chromatographic procedure. Analytical Biochemistry 1992, 206,
178–182.
1
2
9. Leslie, A.G. Integration of macromolecular diffraction data. Acta Cryst. D 1999, 55, 1696–1702.
0. Chen, Y.; Gallois-Montbrun, S.; Schneider, B.; V e´ ron, M.; Mor e´ ra, S.; Deville-Bonne, D.; Janin,
J. Nucleotide binding to nucleoside diphosphate kinases: X-ray structure of human NDPK-A in
complex with ADP and comparison to protein kinases. J. Mol. Biol. 2003, 332, 915–926.
2
1. Brunger, A.T.; Adams, P.D.; Clore, G.M.; Gros, P.; Grosse-Kunstleve, R.W.; Jiang, J.S.; Kuszewski, J.;
Nilges, N.; Pannu, N.S.; Read, R.J.; Rice, L.M.; Simonson, T.; Warren, G.L. Crystallography & NMR
System (CNS): A new software suite for macromolecular structure determination. Acta Cryst. D 1998,
5
4, 905–921.
2. Jambon, M.; Imberty, A.; Del e´ age, G.; Geourjon, C. A new bioinformatic approach to detect common
D sites in protein structures. Proteins 2003, 52, 137–145.
3. Jambon, M.; Andrieu, O.; Combet, C.; Del e´ age, G.; Delfaud, F.; Geourjon, C. The SuMo server: 3D
search for protein functional sites.Bioinformatics. 2005, 21, 3929–3230.
4. Dumas, C.; Lascu, I.; Mor e´ ra, S.; Glaser, P.; Fourme, R.; Wallet, V.; Lacombe, M.L.; V e´ ron, M.; Janin,
J. X-ray structure of nucleoside diphosphate kinase. EMBO J. 1992, 11, 3203–3208.
5. Xu, Y.W.; Mor e´ ra, S.; Janin, J.; Cherfils, J. AlF3 mimics the transition state of protein phosphorylation
in the crystal structure of nucleoside diphosphate kinase and MgADP. Proc. Natl. Acad. Sci. USA 1997,
2
2
2
2
3
9
4, 3579–3583.
ꢁ
2
2
6. Mor e´ ra, S.; Lascu, I.; Dumas, C.; LeBras, G.; Briozzo, P.; V e´ ron, M.; Janin, J. Adenosine 5 -
diphosphate binding and the active site of nucleoside diphosphate kinase. Biochemistry 1994, 33,
459–467.
ꢁ
7. Schneider, B.; Xu, Y.W.; Janin, J.; V e´ ron, M.; Deville-Bonne, D. 3 -Phosphorylated nucleotides
are tight binding inhibitors of nucleoside diphosphate kinase activity. J. Biol. Chem. 1998, 273,
28773–28778.
28. Jeudy, S.; Lartigue, A.; Claverie, J.M.; Abergel, C. Dissecting the unique nucleotide specificity of
Mimivirus nucleoside diphosphate kinase. J. Virol. 2009, 83, 7142–7150.
2
9. Meyer, P.; Schneider, B.; Sarfati, S.; Deville-Bonne, D.; Guerreiro, C.; Boretto, J.; Janin, J.; V e´ ron, M.;
Canard, B. Structural basis for activation of alpha-boranophosphate nucleotide analogues targeting
drug-resistant reverse transcriptase. EMBO J. 2000, 19, 3520–3529.
3
3
3
0. Chen, Y.; Morera, S.; Pasti, C.; Angusti, A.; Solaroli, N.; V e´ ron, M.; Janin, J.; Manfredini, S.; Deville-
Bonne, D. Adenosine phosphonoacetic acid is slowly metabolized by NDP kinase. Med. Chem. 2005,
1, 529–536.
1. Xu, Y.; Sellam, O.; Mor e´ ra, S.; Sarfati, S.; Biondi, R.; V e´ ron, M.; Janin, J. X-ray analysis of azido-
thymidine diphosphate binding to nucleoside diphosphate kinase. Proc. Natl. Acad. Sci. USA 1997,
94, 7162–7165.
2. Strelkov, S.V.; Perisic, O.; Webb, P.A.; Williams, R.L. The 1.9 A crystal structure of a nucleoside
ꢁ
ꢁ
diphosphate kinase complex with adenosine 3 ,5 -cyclic monophosphate: evidence for competitive
inhibition. J. Mol. Biol. 1995, 249, 665–674.

7
92
K. Koch et al.
3
3. Gallois-Montbrun, S.; Chen, Y.; Dutartre, H.; Sophys, M.; Morera, S.; Guerreiro, C.; Schneider, B.;
Mulard, L.; Janin, J.; Veron, M.; Deville-Bonne, D.; Canard, B. Structural analysis of the activation of
ribavirin analogs by NDP kinase: comparison with other ribavirin targets. Mol. Pharmacol. 2003, 63,
538–546.
3
3
3
4. Gallois-Montbrun, S.; Schneider, B.; Chen, Y.; Giacomoni-Fernandes, V.; Mulard, L.; Morera, S.;
Janin, J.; Deville-Bonne, D.; Veron, M. Improving nucleoside diphosphate kinase for antiviral
nucleotide analogs activation. J. Biol. Chem. 2002, 277, 39953–39959.
5. Robbins, B.L.; Greenhaw, J.; Connelly, M.C.; Fridland, A. Metabolic pathways for the activation
of the antiviral agent 9-R-2-phosphonomethoxypropyl adenine (PMPA) in human lymphoid cells.
Antimicrob. Agents Chemother. 1995, 39, 2304–2308.
6. Topalis, D.; Alvarez, K.; Barral, K.; Munier-Lehmann, H.; Schneider, B.; V e´ ron, M.; Guerreiro, C.;
Mulard, L.; El-Amri, C.; Canard, B.; Deville-Bonne, D. Acyclic phosphonate nucleotides and human
adenylate kinases: impact of a borano group on alpha-P position. Nucleosides Nucleotides Nucleic Acids.
2008, 27, 319–331.
3
7. Rinaldo-Matthis, A.; Rampazzo, C.; Balzarini, J.; Reichard, P.; Bianchi, V.; Nordlund, P. Crystal
structures of mitochondrial deoxyribonucleotidase in complex with two specific inhibitors. Mol
Pharmacol. 2004, 65, 860–967.
3
8. Robbins, B.L.; Srinivas, R.V.; Bishofberger, N.; Fridland, A. Anti-human immunodeficiency virus
activity and cellular metabolism of a potential prodrug of the acyclic nucleoside phosphonate 9-R-2-
phosphonomethoxypropyl adenine (PMPA) bis(isopropylmethylcarbonyl)PMPA. Antimicrob. Agents
Chemother. 1998 42, 612–617.
3
9. Pruvost, A.; Negredo, E.; Benech, H.; Theodoro, F.; Puig, J.; Gran, E.; Garcia, E.; Molto, J.; Grassi, J.;
Clotet, B. Measurement of intracellular didanosine and tenofovir phosphorylated metabolites and
possible interactions of the two drugs in human immunodeficiency virus-infected cells. Antimicrob.
Agents Chemother. 2005, 49, 1907–1914.
4
0. Krishnan, P.; Gullen, E.A.; Lam, W.; Dutschman, G.; Grill, S.P.; Chen, Y.C. Novel role of 3-
phosphoglycerate kinase, a glycolytic enzyme, in the activation of L-nucleoside analogs, a new class
of anticancer and antiviral agents. J. Biol. Chem. 2003, 278, 36726–36732.
4
1. Gallois-Montbrun, S.; Faraj, A.; Seclaman, E.; Sommadossi, J.P.; Deville-Bonne, D.; V e´ ron, M. Broad
specificity of human phosphoglycerate kinase for antiviral nucleoside analogs. Biochem. Pharmacol.
2004, 68, 1749–1756.