Striking ability of adenosine-2′(3′)-deoxy-3′(2′)-triphosphates and related analogues to replace ATP as phosphate donor for all four human, and the Drosophila Melanogaster, deoxyribonucleoside kinases

Article abstract of DOI:10.1081/NCN-120019510

In extension of an earlier report, six non-conventional analogues of ATP, three adenosine-2′-triphosphates (3′-deoxy, 3′-deoxy-3′-fluoro- and 3′-deoxy-3′-fluoroxylo-), and three adenosine-3′-triphosphates (2′-deoxy-, 2′-deoxy-2′-fluoro- and 2′-deoxy-2′-fluoroara-), were compared with ATP as potential phosphate donors for human deoxycytidine kinase (dCK), cytosolic thymidine kinase (TK1), mitochondrial TK2, deoxyguanosine kinase (dGK), and the deoxyribonucleoside kinase (dNK) from Drosophila melanogaster. With one group of enzymes, comprising TK1, TK2, dNK and dCK (with dAdo as acceptor), only 3′-deoxyadenosine-2′-triphosphate was an effective donor (5-60% that for ATP), and the other five analogues much less so, or inactive. With a second set, including dCK (dCyd, but not dAdo, as acceptor) and dGK (dGuo as acceptor), known to share high sequence similarity (≈45% sequence identity), all six analogues were good to excellent donors (13-119% that for ATP). With dCK and ATP], products were shown to be 5′-phosphates. With dCK, donor properties of the analogues were dependent on the nature of the acceptor, as with natural 5′-triphosphate donors. With dCK (dCyd as acceptor), K_m and V_max for the two 2′(3′)-deoxyadenosine-3′(2′)-triphosphates are similar to those for ATP. With dGK, K_m values are higher than for ATP, while V_max values are comparable. Kinetic studies further demonstrated Michaelis-Menten (non-cooperative) or cooperative kinetics, dependent on the enzyme employed and the nature of the donor. The physiological significance, if any, of the foregoing remains to be elucidated. The overall results are, on the other hand, highly relevant to studies on the modes of interaction of nucleoside kinases with donors and acceptors; and, in particular, to interpretations of the recently reported crystal structures of dGK with bound ATP, of dNK with bound dCyd, and associated modeling studies.

Full text of DOI:10.1081/NCN-120019510

This article was downloaded by: [Eindhoven Technical University]
On: 21 November 2014, At: 04:56
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
Striking Ability of Adenosine-2′(3′)-deoxy-3′(2′)-
triphosphates and Related Analogues to Replace ATP
as Phosphate Donor for All Four Human, and the
Drosophila Melanogaster, Deoxyribonucleoside Kinases
Krzysztof Krawiec a , Borys Kierdaszuk a f , Elena N. Kalinichenko b , Elena B. Rubinova b ,
Igor A. Mikhailopulo b , Staffan Eriksson c , Birgitte Munch-Petersen d & David Shugar a e g h
a Department of Biophysics, Institute of Experimental Physics , University of Warsaw ,
Warsaw, Poland
b Institute of Bioorganic Chemistry , National Academy of Sciences , Minsk, Belarus
c Department of Veterinary Medical Chemistry, Biomedical Centre , Swedish University of
Agricultural Sciences , Uppsala, Sweden
d Department of Life Sciences and Chemistry , University of Roskilde , Roskilde, Denmark
e Institute of Biochemistry and Biophysics , Polish Academy of Sciences , Warsaw, Poland
f Department of Biophysics , Institute of Experimental Physics, University of Warsaw , 93
Żwirkii Wigury St., Warsaw, PL-02-089, Poland
g Department of Biophysics , Institute of Experimental Physics, University of Warsaw , 93
Żwirkii Wigury St., Warsaw, PL-02-089
h Institute of Biochemistry and Biophysics, Polish Academy of Sciences , 5a Pawinśkiego St.,
Warsaw, PL-02-106, Poland
Published online: 24 Jun 2011.
To cite this article: Krzysztof Krawiec , Borys Kierdaszuk , Elena N. Kalinichenko , Elena B. Rubinova , Igor A. Mikhailopulo ,
Staffan Eriksson , Birgitte Munch-Petersen & David Shugar (2003) Striking Ability of Adenosine-2′(3′)-deoxy-3′(2′)-
triphosphates and Related Analogues to Replace ATP as Phosphate Donor for All Four Human, and the Drosophila Melanogaster,
Deoxyribonucleoside Kinases, Nucleosides, Nucleotides and Nucleic Acids, 22:2, 153-173, DOI: 10.1081/NCN-120019510
To link to this article: http://dx.doi.org/10.1081/NCN-120019510
PLEASE SCROLL DOWN FOR ARTICLE
Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained
in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the
Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and
are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and
should be independently verified with primary sources of information. Taylor and Francis shall not be liable for
any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever
or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of
the Content.
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. Terms & Conditions of access and use can be found at http://
www.tandfonline.com/page/terms-and-conditions

NUCLEOSIDES, NUCLEOTIDES & NUCLEIC ACIDS
Vol. 22, No. 2, pp. 153–173, 2003
Striking Ability of Adenosine-20(30)-deoxy-30(20)-
triphosphates and Related Analogues to Replace
ATP as Phosphate Donor for All Four Human,
and the Drosophila Melanogaster,
Deoxyribonucleoside Kinases
2
Krzysztof Krawiec,1 Borys Kierdaszuk,1, Elena N. Kalinichenko,
*
Elena B. Rubinova,2 Igor A. Mikhailopulo,2 Staffan Eriksson,3
Birgitte Munch-Petersen,4 and David Shugar1,5,
*
1Department of Biophysics, Institute of Experimental Physics,
University of Warsaw, Warsaw, Poland
2Institute of Bioorganic Chemistry, National Academy of Sciences,
Minsk, Belarus
3Department of Veterinary Medical Chemistry, Biomedical Centre,
Swedish University of Agricultural Sciences, Uppsala, Sweden
4Department of Life Sciences and Chemistry, University of Roskilde,
Roskilde, Denmark
5Institute of Biochemistry and Biophysics, Polish Academy of Sciences,
Warsaw, Poland
*Correspondence: David Shugar, Department of Biophysics, Institute of Experimental
_
Physics, University of Warsaw, 93 Zwirkii Wigury St., PL-02-089 Warsaw, and Institute of Bio-
chemistry and Biophysics, Polish Academy of Sciences, 5a Pawinꢀskiego St., PL-02-106 Warsaw,
Poland; Fax: þ(48)-39-12-16-23; E-mail: shugar@ibb.waw.pl; Borys Kierdaszuk, Department
_
of Biophysics, Institute of Experimental Physics, University of Warsaw, 93 Zwirkii Wigury
St., PL-02-089 Warsaw, Poland; Fax: þ(48)-22-5540001; E-mail: borys@biogeo.uw.edu.pl.
153
DOI: 10.1081/NCN-120019510
Copyright # 2003 by Marcel Dekker, Inc.
1525-7770 (Print); 1532-2335 (Online)
www.dekker.com

154
Krawiec et al.
ABSTRACT
In extension of an earlier report, six non-conventional analogues of ATP, three adeno-
sine-20-triphosphates (30-deoxy, 30-deoxy-30-fluoro- and 30-deoxy-30-fluoroxylo-),
and three adenosine-30-triphosphates (20-deoxy-, 20-deoxy-20-fluoro- and 20-deoxy-
20-fluoroara-), were compared with ATP as potential phosphate donors for human
deoxycytidine kinase (dCK), cytosolic thymidine kinase (TK1), mitochondrial TK2,
deoxyguanosine kinase (dGK), and the deoxyribonucleoside kinase (dNK) from Dro-
sophila melanogaster. With one group of enzymes, comprising TK1, TK2, dNK and
dCK (with dAdo as acceptor), only 30-deoxyadenosine-20-triphosphate was an effec-
tive donor (5–60% that for ATP), and the other five analogues much less so, or inactive.
With a second set, including dCK (dCyd, but not dAdo, as acceptor) and dGK (dGuo
as acceptor), knownto sharehigh sequence similarity (ꢀ 45% sequence identity), all six
analogues were good to excellent donors (13–119% that for ATP). With dCK and
ATP1, products were shown to be 50-phosphates. With dCK, donor properties of
the analogues were dependent on the nature of the acceptor, as with natural 50-tri-
phosphate donors. With dCK (dCyd as acceptor), Km and Vmax for the two 20(30)-
deoxyadenosine-30(20)-triphosphates are similar to those for ATP. With dGK, Km
values are higher than for ATP, while Vmax values are comparable. Kinetic studies
further demonstrated Michaelis-Menten (non-cooperative) or cooperative kinetics,
dependent on the enzyme employed and the nature of the donor. The physiological
significance, if any, of the foregoing remains to be elucidated. The overall results
are, on the other hand, highly relevant to studies on the modes of interaction of nucleo-
side kinases with donors and acceptors; and, in particular, to interpretations of the
recently reported crystal structures of dGK with bound ATP, of dNK with bound
dCyd, and associated modeling studies.
Key Words: Deoxyribonucleoside kinases; ATP analogues; Adenosine-20(3)0-
deoxy-30(20)-triphosphates; Donor-acceptor interdependence.
Abbreviations: ATP1, 30-deoxyadenosine-20-triphosphate; ATP2, 20-deoxyadeno-
sine-30-triphosphate; ATP3, 30-deoxy-30-fluoroadenosine-20-triphosphate (16);
ATP4, 20-deoxy-20-fluoroadenosine-30-triphosphate (14); ATP5, 9-(3-deoxy-3-
fluoro-b-D-xylofuranosyl)adenine-20-triphosphate (15); ATP6, 9-(2-deoxy-2-
fluoro-b-D-arabinofuranosyl)adenine-30-triphosphate (13); N, nucleoside; NTP,
nucleoside 50-triphosphate; dNK, deoxyribonucleoside kinase from Drosophila
melanogaster; dCK, deoxycytidine kinase; dGK, deoxyguanosine kinase; TK1,
cytosolic thymidine kinase (low-affinity form); TK2, mitochondrial thymidine
kinase; HMDS, hexamethyldisilazane; MeOH, methanol; TMS, tetramethylsilane;
TMS-Tfl, trimethylsilyl trifluoromethanesulfonate; DAST, (diethylamino)sulfur
trifluoride.
INTRODUCTION
Deoxyribonucleoside kinases, which play a key role in the salvage pathway for
synthesis of the deoxyribonucleotide precursors required for DNA synthesis, cata-
lyze the following reaction: 20-deoxyribonucleoside þ ATP ! 20-deoxyribonucleo-
side-50-phosphate þ ADP.

Phosphate Donors for Deoxyribonucleoside Kinases
155
Mammalian cells posses four deoxyribonucleoside kinases, viz. cytosolic deoxy-
cytidine kinase dCK (EC 2.7.1.74), cytosolic thymidine kinase TK1, mitochondrial
thymidine kinase TK2 (both EC 2.7.1.21), and mitochondrial deoxyguanosine kinase
dGK (EC 2.7.1.113).[1] Two of these, dCK and dGK, exhibit overlapping substrate
specificities towards dAdo and dGuo, whereas dCK and TK2 both phosphorylate
dCyd, and TK1 is highly specific for dThd and dUrd. Other organisms usually have
a different number of kinases, often with different and=or overlapping specificities,
e.g., the presumed unique deoxyribonucleoside kinase dNK (EC 2.7.1.145) from
Drosophila melanogaster, which is capable of phosphorylating all four natural deoxy-
ribonucleosides, albeit to highly varying degrees.[2]
The partially overlapping specificities of these enzymes for acceptor substrates
with different base and sugar moieties, including many nucleoside analogues which
must undergo phosphorylation for expression of antitumour or antiviral activ-
ities,[1,3–5] points to considerable flexibility in the active centers of these enzymes.
Furthermore, it has long been known that the ATP donor may frequently be
replaced by other 50-NTPs.[6] A striking example is the very marked preference of
mammalian deoxycytidine kinase (dCK), and of a bacterial (B. subtilis) deoxyguano-
sine kinase (dGK) for UTP relative to ATP.[7–10]
The foregoing led us to consider whether the location of the triphosphate group
of the donor is an absolute requirement for phosphate donor activity. In this context,
our attention was directed to a long-overlooked report of Kornberg and coworkers[11]
demonstrating that, although deoxyribonucleoside 30-triphosphates are not sub-
strates for E. coli DNA polymerase, their affinities for the enzyme (in the absence
of template) are comparable to those of the deoxyribonucleoside-50-triphosphates;
and, indeed, they compete with the latter for the 50-NTP binding sites. Subsequently
Sahyoun et al.[12] isolated from amphibian and mammalian cells a naturally occurring
specific inhibitor of adenylyl cyclase, identified as 20-deoxyadenosine-30-phosphate.
This was followed by numerous reports describing the inhibition of diverse enzymes
by nucleotides bearing 30-phosphate and 30-polyphosphate groups.[13–15]
We have previously shown that 20-deoxyadenosine-30-triphosphate (ATP1) and
30-deoxyadenosine-20-triphosphate (ATP2) are relatively good donors for highly puri-
fied mammalian dCK, and somewhat less effective with TK1 and TK2.[16] We have
now extended these studies to the mammalian deoxyguanosine kinase (dGK), and
the deoxyribonucleoside kinase (dNK) from Drosophila melanogaster. We have also
examined the potential donor properties of four additional 30(20)-triphosphates of
20(30)-deoxyadenosines with fluorine-substituted pentose moieties, viz. ATP3 (16),
ATP4 (14), ATP5 (15) and ATP6 (13), shown further below in Sch. 2.
EXPERIMENTAL PROCEDURES
Materials
[5-3H]-20deoxycytidine and [6-3H]-thymidine were from Amersham (UK), and
[2,8-3H]-20-deoxyadenosine and [8-3H]-20-deoxyguanosine were from Moravek
Biochemicals (USA). Non-labelled nucleosides, ATP, DTT, BSA, Tris buffer and
inorganic salts were from Sigma (USA).

156
Krawiec et al.
Enzymes
All deoxyribonucleoside kinases were recombinants, and procedures for cloning
and purification have been previously described, as follows: dCK,[17] dGK,[18,19]
TK1,[20] TK2[21] and dNK.[22] The activities of the individual enzymes with standard
substrates, determined with ATP as donor, are listed in footnote b to Table 4, below.
TK1 may exist in so-called low- (dimer) and high- (tetramer) thymidine-affinity
forms, the latter obtained by incubation of the concentrated enzyme with ATP.[23,24]
This study is limited to the low-affinity form of TK1, to allow testing ATP analogues
in the absence of ATP.
The preparation of dGK contains 0.1 mM ATP (added for purposes of stability).
Since this interferes with donor analogue assays, a simple procedure, using 1-mL
syringe columns of Sephadex G50, was used to remove ATP. Prior to kinetic
experiments, 100 mL aliquots of the enzyme were eluted through these columns by
centrifugation (800 ꢁ g, 4 min), repeated three times to lower background ATP
3-fold, an acceptable level.
Enzyme Assays
Activities of all kinases were followed at 37ꢂC by a radiochemical assay with the
use of 3H-labelled nucleosides, and DE-81 cellulose discs for product separation, essen-
tially as described by Ives et al.[25] Reactions were followed in 50 mM Tris-HCl buffer
pH 7.6, containing equimolar amounts of MgCl2 and phosphate donor (ATP or
ATP analogue) at 1 mM for activity measurements, or variable for kinetic measure-
3
ments, and 0.5 mg=mL BSA, 10 mM DTT, indicated concentrations of H-labeled
20-deoxynucleoside, and 0.75–200 ng of the appropriate kinase in a total volume of
50 mL. In each case, the level of enzyme was selected to obtain not more than 20%
(usually lower) substrate consumption during ꢀ30 min incubation, with constant
reaction rate. During the course of incubation, 10 mL samples were removed at 8,
16, 24 and 32 min and spotted on DE-81 discs. These were washed three times for
5 min with 5 mM ammonium formate, once in water, then dried and transferred to
scintillation vials, containing 0.3 mL 0.1 M HCl=0.2 M KCl. Following elution of
the product for 15 min, 2.5 mL liquid scintillator was added, and radioactivity
counted. The velocity of a reaction was obtained by least squares linear regression.
The activity of a given donor is expressed as the ratio of the velocity of the reaction
to that with ATP. Kinetic parameters were determined by the initial velocity method,
using non-linear regression fits of the Michaelis-Menten fn ¼ Vmax ꢁ [S ]=(Km þ [S ])g
or Hill fn ¼ Vmax ꢁ [S ]h=(Km þ [S ]h)g equations. For control (background) activity,
h
the reaction medium contained all components, except the phosphate donor, and
usually, with the exception of dGK (see above), showed no detectable activity.
Site(s) of Phosphorylation
Although deoxyribonucleoside kinases, with a 50-NTP as phosphate donor,
are known to phosphorylate only the primary 50-hydroxyl of a nucleoside accep-
tor, it was considered necessary to check whether this is also the case with our

Phosphate Donors for Deoxyribonucleoside Kinases
157
deoxynucleoside-30(20)-triphosphate donors. Hence products of phosphorylation
were treated at pH 7.5 with Russell’s Viper venom, a rich source of 50-nucleoti-
dase, verified as totally inactive against nucleoside-20(30)-phosphates, and dephos-
phorylation monitored by disappearance of radiolabeled product on DE-81
disks. An interesting, and fortunate, finding was that the venom 50-nucleotidase,
although inhibited by ATP, was not by ATP1.
Synthetic Procedures
General
UV spectra were recorded on a Specord M-400 (Carl-Zeiss, Germany). 1H NMR
spectra were run at 200.13 MHz at 23ꢂC on an AC-200 spectrometer equipped with
an Aspect 3000 data system (Bruker, Germany) with TMS as internal standard
(s ¼ singlet, d ¼ doublet, t ¼ triplet, m ¼ multiplet, br.s ¼ broad signal). Assignment
of proton resonances was confirmed, where possible, by selective homo-
nuclear decoupling. Procedures for TLC and HPLC are given in footnotes to
Table 2. Flash silica gel column chromatography of nucleosides was performed on
230–400 mesh silica gel (Merck, Germany). In condensation reactions, freshly
distilled trimethylsilyl trifluoromethanesulfonate (TMS-Tfl, Fluka, Switzerland)
was employed. Solutions of compounds in organic solvents were dried with
anhydrous sodium sulfate for 4 h. Unless otherwise indicated, reactions were carried
out at 20ꢂC.
Condensation of 1-O-acetyl-2-deoxy-2-fluoro-3,5-di-O-benzoyl-D-arabinofuranose
(2) with Silylated Adenine (1). A solution of arabinoside 2 (0.25 g, 0.62 mmol), tri-
methylsilyl derivative of adenine (1) [obtained from adenine (90 mg, 0.66 mmol) by
refluxing in HMDS (0.7 mL) in the presence of TMS-Cl (0.1 mL)] and TMS-Tfl
(0.3 mL, 1.8 mmol) in anh 1,2-dichloroethane (3 mL) was stirred at r.t. for 48 h,
poured into 5% aq. solution of NaHCO3 (10 mL) and extracted with CHCl3
(3 ꢁ 15 mL). The combined organic extracts were evaporated, the residue was treated
with MeOH (15 mL), saturated with dry ammonia gas at 0ꢂC, for 20 h and evapo-
rated. The residue was chromatographed [silica gel column, 5 ꢁ 40 cm; elution with
a linear MeOH gradient in CHCl3 (0 ! 10%, v=v; 2 ꢁ 300 mL)] to yield the N9-b-iso-
mer 3 (35 mg, 21%) and the N9-a-isomer 4 (18 mg, 11%; amorphous). Compounds 3
and 4 are identical in all respects (1H NMR and UV spectra, TLC and HPLC) to
those previously prepared.[26]
9-(2-Deoxy-2-fluoro-b-D-arabinofuranosyl)adenine (3). Compound 5 (1.12 g,
1.5 mmol)[3] was added to a solution of DAST (1.3 mL, 10.0 mmol) in a mixture of
anh. CH2Cl2 (20 mL) and pyridine (1.6 mL), the reaction mixture was heated at
50ꢂC for 3 h, poured into 5% aq. solution of NaHCO3 (20 mL) and extracted with
CHCl3 (3 ꢁ 25 mL). The combined organic extracts were evaporated, the residue
was treated with MeOH (15 mL), saturated with dry ammonia gas at 0ꢂC, for 20 h
and evaporated. The residue was dissolved in 80% aq. solution of CH3COOH

158
Krawiec et al.
(10 mL), the mixture was stirred for 1 h and evaporated. The residue was chromato-
graphed as described above to yield the N9-b-isomer 3 (40 mg, 10%).
Phosphorylation
The nucleosides 9–12 were converted to the corresponding triphosphates 13–16
(Sch. 2), as previously described for the preparation of ATP1 and ATP2 in Ref.[16]
with appropriate data given in Table 2.
,
RESULTS AND DISCUSSION
Syntheses
We have previously demonstrated that the coupling of trimetylsilylated N6-
benzoyladenine with 1,3,5-tri-O-benzoyl-2-deoxy-2-fluoro-b-D-ribofuranose in the
presence of an excess of TMS-Tfl in anhydrous acetonitrile under reflux led to a mix-
ture of isomeric benzoylated nucleosides. Debenzoylation of this mixture followed
by double silica gel column chromatography permitted isolation of pure nucleosides
9-(2-deoxy-2-fluoro-b-D-arabinofuranosyl)adenine (3; 14%), it’s a-anomer (14%),
and N7-a-arabinoside (25%).[26] We have now employed trimetylsilylated adenine
(1) in the reaction with 1-O-acetyl-3,5-di-O-benzoyl-2-deoxy-2-fluoro-b-D-ribofura-
nose (2) for the preparation of arabinoside 3 (for a recent review on nucleoside
synthesis, see[27]). The condensation of 1 with 2 in the presence of excess TMS-Tfl
in anhydrous dichloroethane at room temperature, followed by conventional
workup of the reaction mixture, deprotection of products and silica gel column chro-
matography, afforded arabinoside 3 and it’s a-anomer 4 in 21% and 11% yields,
respectively (Sch. 1).
We have also studied an alternative synthetic route to arabinoside 3, which
30
employs the readily accessible N6,O -dibenzoyl-50-O-(p-monomethoxy)trityladeno-
sine (5)[28] as starting compound. Treatment of 5 with DAST in CH2Cl2 in the
presence of pyridine at 50ꢂC for 3 h, followed by workup of the reaction mixture,
deprotection of products and silica gel column chromatography, gave arabinoside
3 in 10% yield (Sch. 1). The yield of the desired arabinoside 3 in this reaction was
not optimized. It was recently shown that the treatment of 9-(3-O-benzoyl-5-O-tri-
tyl-b-D-ribofuranosyl)-6-chloropurine with DAST under similar reaction conditions
resulted in the formation of the 20-deoxy-20-fluoro-arabinosyl derivative in 78%
yield.[29] Thus, the transformation of 5 into 3 deserves more detailed investigation
as a reasonable alternative to the convergent approach.
The synthesis of 20-deoxy-20-fluoroadenosine (6), 9-(3-deoxy-3-fluoro-b-D-xylo-
furanosyl) adenine (7) and 30-deoxy-30-fluoroadenosine (8) was previously described
in Ref.[28,30–32] Compounds 3, 6–8 were transformed into the respective derivatives
9–10 as reported previously [16,28 and references cited] (Sch. 2).
1
The H NMR and other physico-chemical data for compounds 11 and 12 have
been described in Ref.[28] Similar data for compounds 9 and 10 are listed in Tables 1
and 2.

Phosphate Donors for Deoxyribonucleoside Kinases
159
Scheme 1. (a) 1=2=TMS-Tf1 (1.1:1.0:3.0, mol), CH2ClCH2Cl, room temperature, 48 h; silica
gel column chromatography; (b) NH3=MeOH, room temperature, 48 h; silica gel column chro-
matography (combined yields: 3, 21%; 4, 11%); (c) 1-DAST=Py (1.0:2.0, mol), CH2Cl2, 50ꢂC,
3 h; 2-NH3OH=MeOH, room temperature, 20 h; 3-80% aq. CH3COOH, room temperature,
1 h (3, S10%).
The triphosphates 13 (ATP6), 14 (ATP4), 15 (ATP5), 16 (ATP3) were then pre-
pared as described previously,[16,28], as shown in Sch. 2. Relevant physico-chemical
data are presented in Table 2.
Conformational Analysis of the Furanose Rings
This was performed in order to reveal some possible correlations between the
stereochemistry of the furanose rings and the donor activities of ATP analogues

160
Krawiec et al.
Scheme 2. (a) 1-TMS-C1=Py; 2-BzCl; NH4OH; 3-MTr-Cl=Py (S65–75%); (b) 1-1M solution
2ꢃ
of PO(Im)3 in anh. pyridine, r.t., 4–6 h; 2-P2O7H2 x(Bu3HN1þ)2, DMF=Bu3N, r.t., 30 min;
3-0.5 M TEAB buffer, r.t., 3 h; 4-conc. NH4OH, r.t., 20 h; 5-80% aq. CH3COOH, r.t.,
1 h: 6-DEAE Sephadex [A-25, HCO3.1ꢃ, linear TEAB gradient (0.001–1 M, 2 ꢁ 500 mL)],
(S32–43%).

Phosphate Donors for Deoxyribonucleoside Kinases
161

162
Krawiec et al.
Table 2. Properties of 9-[2-deoxy-2-fluoro-5-O-(p-monomethoxy)trityl-b-D-arabinofurano-
syl]-N6-benzoyladenine (9) and its ribo-isomer 10 and triphosphates 13–16.
UV spectrum (MeOH)
lmax
HPLCb
Combined
yield (%)
TLCa
(Rf value)
Compound
e
Rt (min) Purity (%)
9
10
13
14
15
16
60
54
40
43
32
43
0.29
0.32
0.19
0.21
0.17
0.19
233.0; 280.0 33900; 27600
232.0; 281.0 41500; 23200
–
–
3.34
5.63
3.93
7.47
–
–
91
96
94
93
260.0
260.0
260.0
260.0
13600
14200
14000
13800
aTLC: aluminium sheets silica gel 60 F254 (Merck, Germany), solvent systems: 2-propanol=25%
aq. ammonia=water, 11 : 7 : 2 (vol).
bHPLC: Waters apparatus (Waters, USA); column Nova-Pac C-18 (3.9 ꢁ 300 mm)(Waters);
isocratic elution with 7% MeCN in 0.1M KH2PO4 (v=v), flow rate 0.7 mL=min (time of
analysis 15 min).
studied. The PSEUROT (version 6.3) program was employed for the conformational
analysis of the furanose rings of 9-(2-deoxy-2-fluoro-b-D-arabinofuranosyl)adenine
(3), 20-deoxy-20-fluoroadenosine (6), 9-(3-deoxy-3-fluoro-b-D-xylofuranosyl)adenine
(7) and 30-deoxy-30-fluoroadenosine (8). This program calculates the best fits of three
3
3
3J(H, H) and two J(H, F) experimental coupling constants (3JH1 ,H2 , JH2 ,H3 and
0
0
0
0
3
3
JH3 ,H4 , as well as JH1 ,F2 and JH3 ,F2 in the case 3 and 6, and JH2 ,F3 and JH4 ,F3
3
3
3
0
0
0
0
0
0
0
0
0
0
in the case of 7 and 8) to the five conformational parameters (P and cm for both the
Table 3. Pseudorotational parameters of 9-(2-deoxy-2fluoro-b-D-arabinofuranosyl)adenine
(3), 20-deoxy-20-fluoroadenosine (6), 9-(3-deoxy-3-fluoro-b-D-xylofuranosyl)-adenine (7), and
30-deoxy-30-fluoroadenosine (8) in D2O solutions.
N $ S
Compound
3a (JHH
PN
cm(N)
PS
cm(S)
rms
jDJmax
j
%S
equilibrium
3
)
ꢃ2.5
11.6
32.5
39.3
30.6
31.1
10.0b
40b
118.6
123.6
201.6
210.7
162b
40b
.132
.235
.000
.295
.000
.093
.000
.050
.21
.69
.00
.90
.00
.28
.00
.17
67
66
36
40
14
13
98
97
2T ! 1T0
(JHH & JHF
6c (JHH
(JHH & JHF
7d (JHH
(JHH & JHF
8d (JHH
(JHH & JHF
)
)
)
)
3T2 ! 1E
3
)
45b
40b
39b
4T 3E
4T 3T4
3
)
32.5
33.4
39b
3T4 2E
3T4 2E
2
)
160.4
158.7
35.1
35.0
3T2 ! E
2
3T2 ! E
aCoupling constants from Ref.[26]
bValues indicated were fixed during final calculations.
cCoupling constants from Ref.[30]
dCoupling constants from Ref.[37]
.
.
.

Phosphate Donors for Deoxyribonucleoside Kinases
163
N- and S-type conformers and the corresponding mole fractions).[33] Optimization of
pseudorotational parameters was performed essentially as described previously.[34]
The resulting optimised geometries of the N- and S-pseudorotamers are presented
in Table 3.
It was previously shown on the basis of 3J(H, H) analysis that the sugar moiety
of the b-arabinoside 3 is in a two-state N- and S-pseudorotational equilibrium in the
ratio 36 : 64[26]. Rather similar results were obtained in the present work (Table 3).
Note that, in the predominantly populated E (S-type) conformation, the F-C(20)-
1
C(10)-O(40) and F-C(20)-C(10)-N(9) fragments adopt the gauche orientation, whereas
the other possible gauche interactions [C(20)-F= C(30)-OH and C(30)-OH and the ring
oxygen] are not realised. Furthermore, in the less populated 3T2 (N-type) conforma-
tion, two gauche effects are operative [C(20)-F=C(30)-OH and C(20)-F=C(10)-N(9)]
along with the anomeric effect, which is more optimal in the N-type conforma-
tion.[35,36] Further support for the E $ T2 equilibrium results from the observed
3
1
3J(F, C(40)) coupling constant of 3.39 Hz,[26] which is an average value of the above
conformations: in the 1E conformation the C(40)-C(30)-C(20)-F torsion angle is about
3
95ꢂ, whereas in the T2 conformation the same torsion angle is about 165ꢂ.
In the case of riboside 6, the HO-C(30)-C(20)-F fragment adopts the gauche
orientation in both N- and S-type pentofuranose ring conformations. The gauche
interaction between C(20)-F and the ring oxygen, along with the anomeric effect,
drive the N $ S equilibrium toward the N-conformation. Note that the two
aforementioned effects slightly predominate over the gauche effect of HO-C(30)-
3
C(40)-O(40). It is noteworthy that J(F, C(40)) was found to be < 2.0 Hz, consistent
3
with a predominant 4T conformation (the C(40)-C(30)-C(20)-F torsion angle is
around 90ꢂ) as distinct from the less populated T4, where the same torsion angle
3
is around 165ꢂ.
There is very good concordance of results from the previous PSEUROT analysis
3
of J(H, H) of the furanose rings of the 30-fluoro substituted nucleosides 7 and 8[37]
3
3
and those herein presented, employing both J(H, H) and J(H, F) coupling con-
stants. The pentofuranose ring of the xyloside 7 assumes the most populated
3T4(N-type) conformation, mainly by virtue of the gauche effect of F-C(30)-C(40)-
O(40). Moreover, the anomeric effect drives the N $ S equilibrium toward the
N-conformation. In the case of riboside 8, like the isomeric riboside 6, the
HO-C(20)-C(30)-F fragment adopts the gauche orientation in both N- and S-type con-
formations. Therefore, the gauche interaction between C(30)-F and C(40)-O(40) bonds
2
is the predominant factor leading to the dominant population of the E (S-type)
conformation.
The values of the vicinal 3JH(1 ),C(8) ¼ 4.55 Hz and 3J H(1 ),C(4) < 2.0 Hz couplings
of the arabinoside 3 unambiguously point to the predominant anti-conformation
about the glycosidic bond.[26,38,39] By contrast, the isomeric riboside 6 reveals higher
conformational mobility about the glycosidic bond, reflected in identical values of
the aforementioned vicinal coupling constants, equal to 3.80 Hz. The predominant
anti-conformation of the adenine bases of the 30-fluorinated nucleosides 7 and
8 was previously deduced from the spin-lattice relaxation times of H(8) and
H(10).[37]
0
0
One may expect that introduction of a triphosphate group at the secondary
hydroxyl of nucleosides 3, 6–8 should not lead to essential changes of the N $ S

164
Krawiec et al.
conformational equilibrium of the pentofuranose rings. The cis-arrangement of the
fluorine atom and the phosphate group does not result either in effective steric repul-
sion, or in additional stabilisation of any conformation, by means of hydrogen bond-
ing. It was, however, shown that a 30-phosphate group in ribonucleosides drives the
N $ S pseudorotational equilibrium somewhat more to the S-type pseudo-
rotamers vs. the 30-hydroxyl, implying a stronger gauche effect of O(40)-C(40)-C(30)-
O-phosphate than of the O(40)-C(40)-C(30)-OH fragment.[16,38] Concurrently, the
enhanced ability of the former fragment to influence the N $ S pseudorotational
equilibria was shown to be lower in the case of purine compounds. In toto, there
are no evident reasons to expect noticeable changes in conformation on going from
nucleosides 3, 6–8 to the corresponding triphosphates 13–16.
Enzymatic Aspects
It should be noted that, with human dCK, it is now well documented that UTP
is a much more effective donor than ATP in vitro and, to a considerable extent,
in vivo.[8–10] But, since our 20(30) -triphosphates are analogues of ATP, the latter is
employed throughout as the reference donor.
The phosphate donor activities of all six ATP analogues, expressed as initial
rates of phosphorylation relative to that for ATP (taken as 100%), are shown in
Table 4. The major substrates of TK1 (dThd) and dGK (dGuo) were used as phos-
phate acceptors, while two substrates were used with dCK (dCyd and dAdo), TK2
(dThd and dCyd) and dNK (dCyd and dAdo). With dCK and TK2 the second sub-
strate (dAdo and dCyd, respectively) is also effectively phosphorylated, but differs
significantly in kinetic parameters from the major one. For dNK all four natural
20-deoxynucleosides are substrates,[2] leading to selection of one pyrimidine (dCyd)
and one purine (dAdo) acceptors. Footnote b to Table 4 lists the actual activities
of ATP with the different enzymes for the substrates employed, each at a concentra-
tion exceeding its Km value.
It should be noted that the nucleoside acceptor concentrations used here
(Table 4) are relatively high as compared with their physiological concentrations
(see[8] and references cited therein). Hence phosphorylating activities of the enzymes
may not reflect the situation under physiological conditions. Concentrations of sub-
strates used here were dictated by the need for substrate saturation, usually achieved
in a concentration range 10-fold higher than the appropriate Km values. Conse-
quently, initial rates measured here may be regarded as relatively good approxima-
tions of Vmax values, hence also of the turnover rates.
Site(s) of phosphorylation were first established by treatment of phosphorylated
products with 50-nucleotidase, as described under Experimental Procedures. With
dCK, and ATP1 as donor, phosphorylation of dCyd and dAdo led only to nucleo-
side-50-phosphates, other enzymes and phosphate donor analogues were not tested.
Quite striking is the fact that, with the enzyme dCK, and dCyd as acceptor, all
analogues exhibited donor activities comparable to that of ATP, with ATP1 as fully
active as ATP. However, with dAdo as acceptor, only ATP1 exhibited high activity,
60% that for ATP, the rates for all the others being below 10% that for ATP. Equally
striking is the finding that, with the enzyme dGK, ATP1 is even more active than

Phosphate Donors for Deoxyribonucleoside Kinases
165
Table 4. Phosphate donor activities of ATP analogues (all at 1 mM) towards deoxyribo-
nucleoside kinases,a expressed as initial rates relative to that of ATP (also at 1 mM, taken
as 100%b) and with different acceptors at concentrations indicated.
dCK
TK1
TK2
dNK
dGKc
% of donor
dCyd dCyd dAdo dThd dThd dCyd dCyd dAdo dGuo
N-conformer Donor 2 mM 25 mM 50 mM 30 mM 1 mM 50 mM 2.5 mM 100 mM 10 mM
ꢀ50
ꢅ50
ꢀ50
3
65
85
ATP
ATP1
ATP2
ATP3
ATP4
ATP5
ATP6
100
84
52
58
36
59
38
100
97
73
66
62
62
59
100
60
6
4
1.5
100
5.3
2.3
0
1
0
100
7
2.2
1.7
1.0
0.9
0.6
100
15
1.6
1.7
1.2
0.7
0.2
100
10
1
0.2
0.6
0
100
6
0.5
0
0.2
0
0
100
119
53
25
70
4
1
55
13
35
0
0
adCK (with dAdo as acceptor), TK1, TK2 and dNK, tentatively classified as group (a); dCK
(with dCyd as acceptor) and dGK (numbers in bold) as group (b), as described in the text.
bActivities of enzymes with ATP are as follows:
dCK=2 mM dCyd – 21 nmol=min=mg
dCK=25 mM dCyd – 27 nmol=min=mg
dCK=50 mM dAdo – 540 nmol=min=mg
TK2=1 mM dThd – 130 nmol=min=mg
TK2=50 mM dCyd – 520 nmol=min=mg
TK1=30 mM dThd – 330 nmol=min=mg
dNK=2.5 mM dCyd – 3.0 mmol=min=mg
dNK=100 mM dAdo – 2.70 mmol=min=mg
dGK=10 mM dGuo – 23 nmol=min=mg
c1% background activity (in absence of donor) due to contamination of dGK with ATP.
ATP, and that ATP2, ATP4 and ATP5 each exhibits activity at a level exceeding
50% that for ATP. Even ATP3 (25% activity) and ATP6 (13% activity) can be con-
sidered as reasonably good donors. It is consequently worth noting that, amongst all
five enzymes, dCK and dGK are known to exhibit highest sequence similarity, with
about 45% identity.[40]
With the other three enzymes, only ATP1 exhibits moderate donor activity,
5–15% that for ATP, whereas all the other analogues are only marginally (2%)
active, or not at all. Note that, with TK2, ATP1 is 2-fold more active vs. dCyd as
compared to dThd; and with dNK, it is more active vs. dCyd than dAdo. It should
be noted that TK2 and dNK, with similar donor specificities, also exhibit high (40%)
sequence identity.[40]
Overall, ATP1 is the most effective donor for all five kinases, whereas ATP2-
ATP6 together may range from excellent to neglible activity, depending on the
kinase employed, so that the enzymes (or, more precisely, enzyme activities) may
be tentatively classified in two groups: (a) TK1, TK2, dNK and dCK (but only with
dAdo as acceptor), with restricted donor specificity, are significantly active only with
ATP1; (b) dCK (with dCyd as acceptor) and dGK, with relaxed donor specificity,
which accept all six donor analogues at a level comparable with, or even superior

166
Krawiec et al.
to, that for ATP (cf. columns in bold and non-bold in Table 4). We revert to this
further, below.
The conformational parameters of our ATP analogues, considered to be similar
to those of the corresponding parent nucleosides (see last paragraph of Conforma-
tional analysis), exhibit marked variations, particularly as regards the N $ S equili-
bria, and the puckering modes of the furanose rings (Table 3). But, as may be seen
from Table 4, there is no apparent satisfactory correlation between these parameters
and donor activities. Presumably the enzyme constrains a bound donor to the
conformation required. It should, nonetheless, be noted that, in the case of dGK,
the N $ S equilibrium of the pentafuranose rings does appear to influence phosphate
donor activities. Furthermore, amongst the enzymes in group (a), the presence of an
electronegative fluorine at the pentose C(20) or C(30) does lead to a marked decrease
in donor activities of ATP3-ATP6 relative to their non-fluorinated analogues ATP1
and ATP2.
Kinetic parameters for two of the analogues, ATP1 and ATP2, as compared to
those for ATP, are presented in Table 5. It was previously shown that, with native
dCK, both ATP1 and ATP2 were competitive with respect to ATP.[16] It will be seen
that, with dCK, both Km and Vmax values are comparable for all three donors, with
25 mM dCyd as acceptor, in line with their rates of phosphorylation (see Table 4).
With TK1, Km values are also comparable, but Vmax values are much lower for
ATP1 and ATP2, in accord with their much poorer abilities as donors (Table 4).
Table 5. Kinetic parameters for the phosphate donors ATP1 and ATP2, and ATP for
comparison, with various deoxyribonucleoside kinases. Vmax is expressed as % reaction rate
with 1 mM ATP (see Table 4, footnote b), and h is the Hill coefficient. Acceptors employed,
and their concentrations, are listed for each kinase
Enzyme
dCK
TK1
TK2
dNK
dGK
Donor
(25 mM dCyd) (30 mM Thd) (1 mM Thd) (10 mM dCyd) (20 mM dGuo)
app
ATP Km
(mM) 7 ꢆ 1 (60 ꢆ 7)a 140b
2c
1.4d
100f
1.0d
20 ꢆ 3e
104
0.8
Vmax (%)
h
app
103 (106)a
0.7 (1.0)a
104f
1.6b
100f
1.0c
ATP1 Km
(mM) 11 ꢆ 1.5
116 ꢆ 37
15 ꢆ 4
24 ꢆ 1
no saturation
up to 4 mM
Vmax (%)
h
app
98.7
0.8–1.0
5.9
(1.0)g
7.1
1.0
18.6
1.3
0.5–0.6
ATP2 Km
(mM) 7.4 ꢆ 0.3
165 ꢆ 20
2.7
(1.0)g
36 ꢆ 13
2.2
0.7
28.5 ꢆ 3.5
1.6
0.8
130 ꢆ 67
67
0.75
Vmax (%)
h
73.5
1.0
aValues in brackets are for 50 mM dAdo as acceptor.
bData from.[24]
cData from.[57]
dData from.[2]
eA much lower value, 2.8 ꢆ 0.5 mM, was reported by Herrstro¨m Sjo¨berg et al.[58]
fCalculated using published Km values (see footnotes b-d) and reaction rates with 1 mM ATP.
gToo few data points to establish kinetic model, but kinetics appear hyperbolic.

Phosphate Donors for Deoxyribonucleoside Kinases
167
By contrast, with TK2 and dNK (sequence identity ꢀ 40%), Km values for both
ATP1 and ATP2 are much higher than for ATP.
Effects of Acceptor on Kinetics of Donor
Studies on nucleoside kinases have hitherto overlooked the possible influence of
a change of acceptor on the kinetics of the traditional ATP (or other 50-NTP) donor.
Our findings with nucleoside-20(30)-triphosphate donors clearly demonstrate that a
change of acceptor may appreciably affect the donor specificity, as shown for dCK
in Table 4, prompting us to examine the kinetic parameters for ATP with dCK, using
dAdo as acceptor in place of dCyd. This led to a change not only in kinetics, coop-
erative with dCyd (h ¼ 0.7), and non-cooperative with dAdo (h ꢀ 1), but also to a
marked increase in Km for ATP from 7 ꢆ 1 mM (with dCyd) to 60 ꢆ 7 mM (with
dAdo), while its Vmax=Km was still higher ( ꢀ 3-fold) with dAdo as acceptor than
with dCyd, due to the much higher turnover with dAdo (Table 5; Table 4, footnote
b). The ratio Vmax=Km for dCyd is 20-fold higher than for dAdo,[41] so that the
difference in Vmax=Km between the two acceptors at fixed (saturated) ATP concen-
tration is much higher. Nonetheless, this is the largest reported modification in
donor kinetics caused by a change of phosphate acceptor.
Furthermore, whereas activity of dATP with dCK (dCyd as acceptor), is
only ꢀ 60% that for ATP, and GTP, dGTP and dTTP are as effective as ATP, repla-
cement of dCyd by dAdo led to a marked decrease in activity of the donor, 2-3-fold
with dGTP and dTTP, and as much as 30-fold with dATP. The foregoing underlines
the broad substrate specificity with respect to donors for dCK (with dCyd as accep-
tor), but much less so with the group (a) enzymes (activities); and is particularly well
reflected in the case of dCK (with dAdo as acceptor), showing that a change of
acceptor may affect the range of specificity of the enzyme with respect to the donor.
Relevant to the foregoing are our earlier results on dependence of the relative
phosphate uptake form the ATP component of an ATP:UTP mixture, catalyzed
by dCK, on the phosphate acceptor,[8] demonstrating a remarkable preference for
UTP, relative to ATP, and applicable with both dCyd and dAdo as acceptors. It
would be interesting, in this context, to determine if 20(30)-deoxy-30(20)-triphosphates
of uridine behave like the adenosine analogues tested here, particularly as regards
dependence on phosphate acceptors.
Biological Aspects
The ability of deoxyadenosine -20- and 30-triphosphates to replace ATP as phos-
phate donors with some deoxyribonucleoside kinases clearly calls for a reexamination
of current concepts on the mode of binding, and the mechanism of action, of ATP as a
phosphate donor not only for nucleoside kinases, but also for other kinase systems,
e.g., protein kinases [see, e.g., 6,42]. In this context, it is worth noting that numerous
reports describe nucleoside-20(30)-phosphates and 30-polyphosphates as potent enzyme
inhibitors. A naturally occurring inhibitor of adenylyl cyclase, long ago isolated from
membranes of amphibian and mammalian cells, was identified as 20-deoxyadenosine-
30-phosphate.[12,43] Subsequently synthetic adenine nucleoside 30-polyphosphates were

168
Krawiec et al.
found to be even better inhibitors of such enzymes.[44] The two most potent inhibitors
of RNase A hitherto reported are the 20- and 30-phosphates of 50-diphosphoadeno-
sine.[14] The naturally occurring 30-phosphates of 50-phosphoadenosine (PAP) and
of adenosine-50-phosphosulfate (PAPS), both tight-binding inhibitors of nucleoside
diphosphate kinases, have been proposed as high-affinity drugs targeted to these
enzymes.[13] The natural occurrence and formation of nucleoside-30-phosphates,
resulting from enzymatic degradation of oligo and polynucleotides by tissue extracts,
has been reported by Bushfield et al.[45] We are, however, not aware of any reports
on the natural occurrence of nucleoside-20-phosphates, which might be anticipated
as products of catabolism of the interferon-induced 20-50 oligoadenylates.[46]
It has long been known that ATP may be replaced as a phosphate donor by other
50-NTPs in various kinase systems,[6,42] including viral TKs.[3,4] For example, UTP is
a far superior donor than ATP with human dCK,[8] and is 50-fold more effective than
ATP for the dGK from B. subtilis.[7] CTP is a good donor with vertebrate mitochon-
drial TK2,[1] and as efficient as ATP with dNK.[2,22] Probably the most striking exam-
ple is bovine mitochondrial dGK,[47] for which ATP is the best donor at its optimum
pH, 5.5, whereas at physiological pH its activity is drastically decreased, while CTP
and dCTP were as active as ATP, and UTP and dTTP twice as active.
The known broad specificity of bovine dGK for natural 50-NTP donors[47] is
reflected here by the broad specificity for 20-, 30- and 50-triphosphate donors of
human dGK (with dGuo as acceptor) and dCK (with dCyd as acceptor). With
TK1, TK2 (specific for pyrimidines) and dNK (preference for pyrimidines), only
ATP1 exhibited weak to moderate activity, and the other five analogues even lower
activity, or not at all (Table 4).
The foregoing also apparently correlates with the excellent donor properties of
UTP, with both human[19,48] and B. subtilis[7] dGK, and with human dCK ([8] and
references cited). By contrast, UTP is a poorer donor than ATP for TK1, TK2[8]
and dNK,[2] which are in group (a) in Table 4. One apparent exception is dCK,
which prefers UTP independently of acceptor,[8] but exhibits restricted donor activity
with dAdo (group (a)), and clearly suggests that donor and acceptor activities are
mutually interrelated.
Differences Between Highly Purified Natural and Cloned Enzymes
The five enzymes employed in this study were all cloned, and it is of interest,
in this context, to compare their activities with their natural counterparts previously
reported.[16] For native dCK, Km values, and Hill coefficients (in brackets), for ATP,
ATP1 and ATP2 (with dCyd as acceptor) are 47 ꢆ 3 mM[49] (h ¼ 0.7), 25 ꢆ 5 mM
(h ¼ 1) and 20 ꢆ 3 mM (h ¼ 1), respectively, as compared to 7 ꢆ 1 mM (h ¼ 0.7),
11.0 ꢆ 1.5 mM (h ¼ 0.8–1) and 7.4 ꢆ 0.3 mM (h ¼ 1) for the recombinant enzyme
(Table 5). For native TK2, the equivalent values for ATP and ATP2 (with dThd
as acceptor) were 10.0 ꢆ 1.5 mM[50] (h ¼ 0.4) and 50 ꢆ 6 mM (h ¼ 1), respectively, as
compared to 2 mM (h ¼ 1) and 36 mM (h ¼ 0.7) for the recombinant enzyme (Table
5). For TK1, with ATP2 as donor and dThd as acceptor, Km values and Hill coeffi-
cients were similar. Cloned dNK exhibited the same acceptor properties as the native
enzyme, but donor properties were not described.[22,51] For cloned dGK[18,52] no such

Phosphate Donors for Deoxyribonucleoside Kinases
169
results have been reported. One conceivable interpretation of the foregoing differ-
ences between native and cloned enzymes is the existence of post-translational
modifications of the former, e.g., it has been suggested that phosphorylation of
dCK[53–55] and TK1[56] may regulate their activity in situ.
Structural Aspects: Mode(s) of Binding of Donors and Acceptors
Crystal structures of the complex of dNK with dCyd, and of dGK with ATP
have been recently reported,[40] and the catalytic sites identified. In the dNK=dCyd
complex, dCyd is bound, as expected, in the nucleoside binding site. By contrast,
and unpredictably, ATP is bound by dGK (but in the absence of a nucleoside accep-
tor) in a direction opposite to the phosphate donor site, with the base at the deoxy-
ribonucleoside site, as expected for a bisubstrate analogue feedback inhibitor.
Furthermore, modeling of dCK, based on dGK (which shares high sequence similar-
ity to dCK, with ꢀ 45% identity), revealed no major differences in the side-chains
linking the substrate pocket, but with Ala100 in dCK replaced by Ser114 in dGK
˚
only 4 A from a purine substrate, in accord with the good activity of dCK for purine
substrates, but unable to account for the good substrate properties of dCyd (and
Cyd).
Our previous results,[16] and the present findings, add a totally new element to
the foregoing, which must be taken into account in studies of the modes of binding
of donors and acceptors by nucleoside kinases in general, particularly if we bear in
mind the present results demonstrating interdependence of binding of donors and
acceptors. We venture to predict that similar considerations probably apply to other
kinase systems.
It is, perhaps, relevant to underline the difficulties inherent in qualitative model-
ing of protein=ligand interactions from crystal structures alone, admirably illustrated
by the crystal structures of RNase A in complexes with its two most potent inhibi-
tors,[14] the 20- and 30-phosphates of 50-phosphoadenosine (referred to above), which
proved to be unpredictable, and totally at variance, with earlier crystal structures of
RNase=inhibitor complexes. It is now obvious that the crystal structures of dCK,
and=or dGK, complexed with one of our 20 (30)-triphosphate analogues, are essential
for further clarification of mode(s) of binding of donors and acceptors by the
enzymes embraced in the present investigation. Also required are modeling studies
which take into account the simultaneous binding of donors and acceptors, now
under way in our laboratories.
ACKNOWLEDGMENTS
The authors are grateful to Dr. G.G. Sivets and Dr. N.E. Poopeiko (Institute of
Bioorganic Chemistry, Minsk, Belarus) for gifts of 1-O-acetyl-3,5-di-O-benzoyl-2-
deoxy-2-fluoro-D-arabinofuranose (2) 20-deoxy-20-fluoroadenosine (6) and 30-deoxy-
30-fluoroadenosine (8). This investigation was supported by the Polish State
Committee for Scientific Research (KBN), which includes the individual project,
BST-718=BF, and international collaborative projects with Uppsala Biomedical

170
Krawiec et al.
Centre (Sweden) and Roskilde University (Denmark). S.E. is also indebted for
support from the Swedish Research Council.
REFERENCES
1. Arner, E.S.; Eriksson, S. Mammalian deoxyribonucleoside kinases. Pharmacol.
Ther. 1995, 67, 155–186.
ˇ
2. Munch-Petersen, B.; Piskur, J.; Sondergaard, L. Four deoxynucleoside kinase
activities from Drosophila melanogaster are contained within a single mono-
meric enzyme, a new multifunctional deoxynucleoside kinase. J. Biol. Chem.
1998, 273, 3926–3931.
3. Gentry, G.A. Viral thymidine kinases and their relatives. Pharmacol. Ther.
1992, 54, 319–355.
4. Shugar, D. Molecular Aspects of Chemotherapy; Shugar, D., Rode, W., Bor-
owski, E., Eds.; Springer-Verlag: Berlin and New York, 1992; 239–270.
5. De Clercq, E. Strategies in the design of antiviral drugs. Nature Reviews Drug
Discovery 2002, 1, 13–25.
6. Shugar, D. The NTP phosphate donor in kinase reactions: is ATP a mono-
polist? Acta Biochim. Pol. 1996, 43, 9–23.
7. Andersen, R.B.; Neuhard, J. Deoxynucleoside kinases encoded by the yaaG
and yaaF genes of Bacillus subtilis. Substrate specificity and kinetic analysis
of deoxyguanosine kinase with UTP as the preferred phosphate donor. J. Biol.
Chem. 2001, 276, 5518–5524.
8. Krawiec, K.; Kierdaszuk, B.; Eriksson, S.; Munch-Petersen, B.; Shugar, D.
Nucleoside triphosphate donors for nucleoside kinases: donor properties of
UTP with human deoxycytidine kinase. Biochem. Biophys. Res. Commun.
1995, 216, 42–48.
9. Shewach, D.S.; Reynolds, K.K.; Hertel, L. Nucleotide specificity of human
deoxycytidine kinase. Mol. Pharmacol. 1992, 42, 518–524.
10. White, J.C.; Capizzi, R.L. A critical role for uridine nucleotides in the regula-
tion of deoxycytidine kinase and the concentration dependence of 1-b-D-
arabinofuranosylcytosine phosphorylation in human leukemia cells. Cancer
Res. 1991, 51, 2559–2565.
11. Eglund, P.T.; Huberman, J.A.; Jovin, T.M.; Kornberg, A. Enzymatic synthesis
of deoxyrybonucleic acid. XXX. Binding of triphosphates to deoxyribonucleic
acid polymerase. J. Biol. Chem. 1969, 244, 3038–3044.
12. Sahyoun, N.; Schmitges, C.J.; Siegel, M.I.; Cuatrecasas, P. 20-Deoxyadenosine-
30-monophosphate: a naturally occurring inhibitor of adenylate cyclase in
amphibian and mammalian cells. Life Sci. 1976, 19, 1961–1970.
13. Schneider, B.; Xu, Y.W.; Janin, J.; Veron, M.; Deville-Bonne, D. 30-Phosphory-
lated nucleotides are tight binding inhibitors of nucleoside diphosphate kinase
activity. J. Biol. Chem. 1998, 273, 28,773–28,778.
14. Leonidas, D.D.; Shapiro, R.; Irons, L.I.; Russo, N.; Acharya, K.R. Crystal struc-
tures of ribonuclease A complexes with 50-diphosphoadenosine 30-phosphate and
0
0
˚
5 -diphosphoadenosine 2 -phosphate at 1.7 A resolution. Biochemistry 1997, 36,
5578–5588.

Phosphate Donors for Deoxyribonucleoside Kinases
171
15. Johnson, R.A.; Desaubry, L.; Bianchi, G.; Shoshani, I.; Lyons, E. Jr.; Taussig,
R.; Watson, P.A.; Cali, J.J.; Krupinski, J.; Pieroni, J.P.; Iyengar, R. Isozyme-
dependent sensitivity of adenylyl cyclases to P-site-mediated inhibition by
adenine nucleosides and nucleoside 30-polyphosphates. J. Biol. Chem. 1997,
272, 8962–8966.
16. Krawiec, K.; Kierdaszuk, B.; Kalinichenko, E.N.; Mikhailopulo, I.A.; Shugar,
D. Unusual nucleoside triphosphate donors for nucleoside kinases: 30-deoxy-
adenosine-20-triphosphate and 20-deoxyadenosine-30-triphosphate. Acta Biochim.
Polon 1998, 45, 87–94.
17. Chottiner, E.G.; Shewach, D.S.; Datta, N.S.; Ashcraft, E.; Gribbin, D.;
Ginsburg, D.; Fox, I.H.; Mitchell, B.S. Cloning and expression of human deoxy-
cytidine kinase cDNA. Proc. Natl. Acad. Sci. U.S.A. 1991, 88, 1531–1535.
18. Wang, L.; Hellman, U.; Eriksson, S. Cloning and expression of human mito-
chondrial deoxyguanosine kinase cDNA. FEBS Lett. 1996, 390, 39–43.
19. Johansson, M.; Karlsson, A. Cloning and expression of human deoxyguano-
sine kinase cDNA. Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 7258–7262.
20. Berenstein, D.; Christensen, J.F.; Kristensen, T.; Hofbauer, R.; Munch-
Petersen, B. Valine, not methionine, is amino acid 106 in human cytosolic
thymidine kinase (TK1). Impact on oligomerization, stability, and kinetic
properties. J. Biol. Chem. 2000, 275, 32187–32192.
21. Wang, L.; Munch-Peterson, B.; Herrstro¨m Sjo¨berg, A.; Hellman, U.; Bergman,
T.; Jornvall, H.; Eriksson, S. Human thymidine kinase 2: molecular cloning
and characterisation of the enzyme activity with antiviral and cytostatic nucleo-
side substrates. FEBS Lett. 1999, 443, 170–174.
ˇ
22. Munch-Petersen, B.; Knecht, W.; Lenz, C.; Sondergaard, L.; Piskur, J. Functional
expression of a multisubstrate deoxyribonucleoside kinase from Drosophila
melanogaster and its C-terminal deletion mutants. J. Biol. Chem. 2000, 275,
6673–6679.
23. Munch-Petersen, B.; Tyrsted, G.; Cloos, L. Reversible ATP-dependent transi-
tion between two forms of human cytosolic thymidine kinase with different
enzymatic properties. J. Biol. Chem. 1993, 268, 15621–15625.
24. Munch-Petersen, B.; Tyrsted, G.; Cloos, L.; Beck, R.A.; Eger, K. Different affi-
nity of the two forms of human cytosolic thymidine kinase towards pirimidine
analogs. Biochim. Biophys. Acta 1995, 1250, 158–162.
25. Ives, D.H.; Durham, J.P.; Tucker, V.S. Rapid determination of nucleoside
kinase and nucleotidase activities with tritium-labeled substrates. Anal. Bio-
chem. 1969, 28, 192–205.
26. Tennila¨, T.; Azhayeva, E.; Vepsa¨la¨nen, J.; Laatikainen, R.; Azhayev, A.; Mikhai-
lopulo, I.A. Oligonucleotides containing 9-(2-deoxy-2-fluro-b-D-arabinofurano-
syl)-adenine and -guanine: synthesis, hybridization and antisense properties.
Nucleosides, Nucleotides & Nucl. Acids 2000, 19, 1861–1884.
27. Vorbruggen, H.; Ruh-Pohlenz, C. Synthesis of nucleosides. Org. Reactions
¨
2000, 55, 1–630.
28. Kalinichenko, E.N.; Podkopaeva, T.L.; Poopeiko, N.E.; Kelve, M.; Saarma,
M.; Mikhailopulo, I.A.; Van Den Boogaart, J.E.; Altona, C. Role of the stereo-
chemistry of 30-fluoro-30-deoxy analogues of 2-5A in binding to and activation
of mouse RNase L. Rec. Trav. Chim. Pays-Bas 1995, 114, 43–50.

172
Krawiec et al.
29. Maruyama, T.; Takamatsu, S.; Kozai, S.; Satoh, Y.; Izawa, K. Synthesis of 9-
(2-deoxy-2-fluro-b-D-arabinofuranosyl)adenine bearing a selectively removable
protecting group. Chem. Pharm. Bull. 1999, 47, 966–970.
30. Mikhailopulo, I.A.; Sivets, G.G.; Poopeiko, N.E.; Khripach, N.B. Oxidation-
reduction sequence for the synthesis of peracylated fluorodeoxy pentofurano-
sides. Carbohyd. Res. 1995, 278, 71–89.
31. Zaitseva, G.V.; Sivets, G.G.; Kazimierczuk, Z.; Vilpo, J.A.; Mikhailopulo, I.A.
Convergent synthesis and cytostatic properties of 2-chloro-20-deoxy-20-fluoro-
adenosine and its N7-isomer. Bioorg. Med. Chem Lett. 1995, 5, 2999–3002.
32. Mikhailopulo, I.A.; Poopeiko, N.E.; Sivets, G.G.; Pricota, T.I.; Kvasyuk, E.I.;
Balzarini, J.; De Clercq, E. Synthesis and antiviral and cytostatic properties
of 30-deoxy-30-fluoro- and 20-azido-30-fluoro-20,30-dideoxy-D-ribofuranosides
of natural heterocyclic bases. J. Med. Chem. 1991, 34, 2195–2202.
33. Van Wijk, J., Haasnot, C.A.G.; de Leeuw, F.A.A.M.; Huckriede, B.D.; Westra
Hoekzema, A.J.A.; Altona, C. PSEUROT 6.3, Leiden Institute of Chemistry,
Leiden University, The Netherlands, 1999.
34. Seela, F.; Debelak, H.; Reuter, H.; Kastner, G.; Mikhailopulo, I.A. Different
conformation of 1-deazaadenosine and its 20-deoxyribonucleoside in crystals
and in solution. Tetrahedron 1999, 55, 1295–1308.
35. Thibaudeau, C.; Chattopadhyaya, J. Stereoelectronic Effects in Nucleosides and
Nucleotides and their Structural Implications; Uppsala University Press: Sweden,
1999.
36. Thibaudeau, C.; Plavec, J.; Chattopadhyaya, J. A new generalized Karplus-
type equation relating vicinal proton-fluorine coupling constants to H-C-C-F
torsion angles. J. Org. Chem. 1998, 63, 4967–4984.
37. Van Den Boogaart, J.E.; Kalinichenko, E.N.; Podkopaeva, T.L.; Mikhailo-
pulo, I.A.; Altona, C. Conformational analysis of 30-fluorinated A(20-50)A(20-
50)A fragments. Relation between conformation and biological activity. Eur.
J. Biochem. 1994, 221, 759–768.
38. Akhrem, A.A.; Mikhailopulo, I.A.; Abramov, A.F. 13C Nuclear magnetic reso-
nance spectroscopy of selected adenine nucleosides: structural correlation and
conformation about the glycosidic bond. Org. Magn. Res. 1979, 12, 247–253.
39. Davies, D.B. Conformations of nucleosides and nucleotides. Progr. Nucl.
Magn. Res. Spectr. 1978, 12, 135–225.
ˇ
40. Johansson, K.; Ramaswamy, S.; Ljungcrantz, C.; Knecht, W.; Piskur, J.; Munch-
Petersen, B.; Eriksson, S.; Eklund, H. Structural basis for substrate specificities
of cellular deoxyribonucleoside kinases. Nature Struct. Biol. 2001, 8, 616–620.
41. Eriksson, S.; Kierdaszuk, B.; Munch-Petersen, B.; Oberg, B.; Johansson, N.G.
Comparison of the substrate specificities of human thymidine kinase 1 and 2
and deoxycytidine kinase toward antiviral and cytostatic nucleoside analogs.
Biochem. Biophys. Res. Commun. 1991, 176, 586–592.
42. Shugar, D. Viral and host-cell protein kinases: enticing antiviral targets and
relevance of nucleoside, and viral thymidine, kinases. Pharmacol. Ther. 1999,
82, 315–335.
43. Sahyoun, N.; Schmitges, C.J.; Siegel, M.I.; Cuatrecasas, P. Inhibition of fat
cell membrane adenylate cyclase by 20-deoxyadenosine-30-monophosphate.
Life Sci. 1976, 19, 1971–1979.

Phosphate Donors for Deoxyribonucleoside Kinases
173
44. Desaubry, L.; Shoshani, I.; Johnson, R.A. Inhibition of adenylyl cyclase by a
family of newly synthesized adenine nucleoside 30-polyphosphates. J. Biol.
Chem. 1996, 271, 14,028–14,034.
45. Bushfield, M.; Shoshani, I.; Johnson, R.A. Tissue levels, source, and regulation
of 30-AMP: an intracellular inhibitor of adenylyl cyclases. Mol. Pharmacol.
1990, 38, 848–853.
46. Player, M.R.; Torrence, P.F. The 2-5A System: Modulation of Viral and Cel-
lular Processes Through Acceleration of RNA Degradation. Pharmacol. Ther.
1998, 78, 55–113.
47. Park, I.; Ives, D.H. Properties of a highly purified mitochondrial deoxyguano-
sine kinase. Arch. Biochem. Biophys. 1988, 266, 51–60.
48. Zhu, C.; Johansson, M.; Permert, J.; Karlsson, A. Phosphorylation of anti-
cancer nucleoside analogs by human mitochondrial deoxyguanosine kinase.
Biochem. Pharmacol. 1998, 56, 1035–1040.
49. In our earlier report [16] this value was incorrectly given as 15 mM, due to an
error in the Hill equation employed.
50. This value was previously reported incorrectly as 2.5 mM [16], due to an error in
the Hill equation employed.
ˇ
51. Knecht, W.; Munch-Petersen, B.; Piskur, J. Identification of residues involved
in the specificity and regulation of the highly efficient multisubstrate deoxy-
ribonucleoside kinase from Drosophila melanogaster. J. Mol. Biol. 2000, 301,
827–837.
52. Herrstro¨m Sjo¨berg, A.; Wang, L.; Eriksson, S. Substrate specificity of human
recombinant mitochondrial deoxyguanosine kinase with cytostatic and anti-
viral purine and pyrimidine analogs. Mol. Pharmacol. 1998, 53, 270–273.
53. Spasokoukotskaja, T.; Csapo, Z.; Sasvari-Szekely, M.; Virga, S.; Talianidis, I.;
Eriksson, S.; Staub, M. Effect of phosphorylation of deoxycytidine kinase
activity. Adv. Experimental Med. Biol. 2000, 486, 281–285.
54. Wang, L.M.; Kucera, G.L. Deoxycytidine kinase is phosphorylated in vitro by
protein kinase C alpha. Biochim. Biophys. Acta 1994, 1224, 161–167.
55. Csapo, Z.; Sasvari-Szkely, M.; Spasokoukotskaja, T.; Talianidis, I.; Eriksson,
S.; Staub, M. Activation of deoxycytidine kinase by inhibition of DNA syn-
thesis in human lymphocytes. Biochem. Pharmacol. 2001, 61, 191–197.
56. He, Q.; Skog, S.; Wu, C.; Johansson, A.; Tribukait, B. Existence of phosphory-
lated and dephosphorylated forms of cytosolic thymidine kinase (TK1).
Biochim. Biophys. Acta 1996, 1289, 25–30.
57. Munch-Petersen, B.; Cloos, L.; Tyrsted, G.; Eriksson, S. Diverging substrate
specificity of pure human thymidine kinases 1 and 2 against antiviral dideoxy-
nucleosides. J. Biol. Chem. 1991, 266, 9032–9038.
58. Herrstro¨m Sjo¨berg, A.; Wang, L.; Eriksson, S. Antiviral guanosine analogs as
substrates for deoxyguanosine kinase: implications for chemotherapy. Anti-
microb. Agents Chemother. 2001, 45, 739–742.
Received August 14, 2002
Accepted December 9, 2002