Functional studies of active-site mutants from Drosophila
melanogaster deoxyribonucleoside kinase
Investigations of the putative catalytic glutamate–arginine pair and
of residues responsible for substrate specificity
Louise Egeblad-Welin
1,2,*
, Yonathan Sonntag
1,*
, Hans Eklund
3
and Birgitte Munch-Petersen
1
1 Department of Science, Systems and Models, Roskilde University, Denmark
2 Department of Molecular Biosciences, Swedish University of Agricultural Sciences, Uppsala Biomedical Centre, Sweden
3 Department of Molecular Biology, Swedish University of Agricultural Sciences, Uppsala Biomedical Centre, Sweden
Drosophila melanogaster deoxyribonucleoside kinase
(Dm-dNK) phosphorylates the four natural deoxyribo-
nucleosides, thymidine, deoxycytidine, deoxyadenosine
and deoxyguanosine, which is a crucial step in the bio-
synthesis of DNA precursors via the salvage pathway.
In addition, Dm-dNK phosphorylates a number of
important nucleoside analogue pro-drugs [1,2], making
it a potential candidate for use in suicide gene therapy.
Keywords
catalytic mechanism; deoxyribonucleoside
kinase; dTTP; enzyme kinetics; nucleoside
analogues
Correspondence
B. Munch-Petersen, Department of Science,
Systems and Models, Roskilde University,
Box 260, DK 4000 Roskilde, Denmark

to an increased distance between the catalytic carboxyl group and 5¢-OH of
deoxythymidine (dThd) or deoxycytidine (dCyd). Mutation of Q81 to N
and Y70 to W was carried out to investigate substrate binding. The muta-
tions primarily affected the K
m
values, whereas the k
cat
values were of the
same magnitude as for the wild-type. The Y70W mutation made the
enzyme lose activity towards purines and negative cooperativity towards
dThd and dCyd was observed. The Q81N mutation showed a 200- and
100-fold increase in K
m
, whereas k
cat
was decreased five- and twofold for
dThd and dCyd, respectively, supporting a role in substrate binding. These
observations give insight into the mechanisms of substrate binding and
catalysis, which is important for developing novel suicide genes and drugs
for use in gene therapy.
Abbreviations
ACV, 9-(2-hydroxyethoxymethyl)-guanine; AraA, 9-(b-
D-arabinofuranosyl)-adenine; AraC, 1-(b-D-arabinofuranosyl)-cytosine; AraT,
1-(b-
D-arabinofuranosyl)-thymine; BVDU, (E)-bromvinyl-2¢-deoxyuridine; CdA, 2-chloro-2¢-deoxyadenosine; dAdo, deoxyadenosine; dCK,
cytosolic deoxycytidine kinase; dCyd, deoxycytidine; dGK, deoxyguanosine kinase; dGuo, deoxyguanosine; Dm-dNK, Drosophila
melanogaster deoxyribonucleoside kinase; dThd, deoxythymidine; F-AraA, 2-flouro-9-(b-
D-arabinofuranosyl)-adenine; FdUrd, 5-flouro-2¢-
deoxyuridine; HSV1-TK, Herpes simplex virus Type 1 thymidine kinase; TK, thymidine kinase.
1542 FEBS Journal 274 (2007) 1542–1551 ª 2007 The Authors Journal compilation ª 2007 FEBS

been suggested that Dm-dNK may be a putative sui-
cide gene in gene therapy. In vivo experiments with
cancer cell lines showed increased sensitivity towards
nucleoside analogues [12] and a bystander effect was
observed [13,14].
The 3D structures of Dm-dNK, human dCK, human
dGK and HSV1-TK show a similar binding mode for
the substrates in the active site. Three key residues in
Dm-dNK, identified and proposed as being responsible
for substrate specificity [4], were mutagenized and the
mutant enzymes characterized for their ability to phos-
phorylate native deoxyribonucleosides and nucleoside
analogues [15]. These mutations of residues 84, 88 and
110 (Fig. 2) converted dNK substrate specificity from
predominantly pyrimidine into purine.
It has been suggested that the reaction mechanism
proposed for HSV1-TK [16] also applies to other deoxy-
ribonucleoside kinases [3]. It is believed that E52 acts
as a base in the deprotonation of 5¢-OH, while the
transition state is stabilized by the positively charged
R105 (Fig. 2). The pK
a
of E52 is probably influenced
by the proximity of R105 which is high enough to act
as a base in the initial catalysis step. In a structural
study of Dm-dNK in which the enzyme was cocrystal-
lized with both deoxythymidine (dThd) and dTTP sep-
arately, E52 formed a hydrogen bond with the 5¢-OH
group of dThd, whereas it was moved 6.5 A
˚

and Q81 (Fig. 2). These two amino acid residues are
conserved among Dm-dNK, TK2, dGK, dCK and
HSV1-TK (Fig. 1). Y70 which anchors the 3¢-OH of the
deoxyribose moiety of the nucleoside, together with
E172 (Fig. 2), was mutated to W. This mutation was
performed to see whether the larger side chain would
affect substrate specificity. Q81, which forms two hydro-
gen bonds with the base, was mutated to N in order to
see how the increased distance between the base and
substrate-binding amino acid affected the binding of
dThd and deoxycytidine (dCyd).
Results
We performed site-directed mutagenesis of four active
site residues of Dm-dNK. Residue E52 was mutated to
D, H and Q, residue Y70 to W, residue Q81 to N and
residue R105 to K and H. The kinetic properties of
the active site mutants are summarized in Table 1. All
mutants were characterized with dThd and dCyd,
Fig. 2. Binding of the substrate dThd at the active site of Dm-dNK
[17]. Hydrogen-bonding residues are shown. E52, Y70, Q81 and
R105 were mutated in this study. Residues V84, M88 and A110,
mutated in a previous study [15], are also included.
Active site mutation of deoxyribonucleoside kinase L. Egeblad-Welin et al.
1544 FEBS Journal 274 (2007) 1542–1551 ª 2007 The Authors Journal compilation ª 2007 FEBS
Y70W was further characterized with three nucleoside
analogues: 1-(b-d-arabinofuranosyl)-cytosine (AraC),
1-(b-d-arabinofuranosyl)-thymine (AraT) and (E)-brom-
vinyl-2¢-deoxyuridine (BVDU).
Mutants of catalytic residues: E52 and R105
It is evident from the kinetic results that mutation of

cat
values were determined using a calculated mass of 26 785 kDa. It is assumed that there is one active site per monomer.
Where cooperativity is observed, the Hill coefficient (n) is given. When k
cat
⁄ K
m
is compared with the wild-type, dThd and dCyd is set to
100%. Kinetic parameters were determined from three independent experiments, except where indicated by * or # , which were based on
one or two experiments, respectively. The results are given as mean ± SD. ND, not detected.
Enzyme Substrate K
m
or K
0.5
(lM) (n) V
max
(lmolÆmin
)1
Æmg
)1
) k
cat
(s
)1
) k
cat
⁄ K
m
(s
)1
ÆM

3.2 · 10
3
(< 1%)
dCyd 5.8 ± 2.0 0.00259 ± 0.0006 1.2 · 10
)3
2.0 · 10
2
(< 1%)
E52Q dThd ND ND – –
dCyd ND ND – –
Y70W dThd# 251 ± 86
(n ¼ 0.6 ± 0.07)
5.2 ± 0.9 2.3 9.2 · 10
3
(< 1%)
dCyd 246 ± 34
(n ¼ 0.76 ± 0.002)
15.2 ± 0.3 6.8 2.8 · 10
4
(< 1%)
AraC# 1441 ± 463 4.6 ± 0.6 2.1 1.4 · 10
3
(< 1%)
AraT# 357 ± 43 2.3 ± 0.1 1.0 2.9 · 10
3
(< 1%)
BVDU# 4.9 ± 0.4 0.82 ± 0.07 0.4 7.5 · 10
4
(< 1%)
Q81N dThd 231 ± 24 12.1 ± 0.5 5.4 2.3 · 10

cat
values decreased by approxi-
mately fivefold with dThd and twofold with dCyd. An
interesting feature for this mutant is that it gained neg-
ative cooperativity with dThd and dCyd. The Y70W
mutant was further characterized using the nucleoside
analogues, AraC, AraT and BVDU. K
m
for AraC was
increased 59-fold, but this was less pronounced than
the increase of 106-fold for dCyd. V
max
was fairly
unchanged compared with the wild-type. The K
m
value
for BVDU was increased approximately twofold com-
pared that for the wild-type, which indicated a minor
change in the binding of BVDU. By contrast, k
cat
was
decreased approximately 15-fold. With AraT, K
m
increased approximately sixfold and k
cat
decreased
approximately fivefold compared with the wild-type.
Thus, for Y70W, the changes in K
m
with these ana-

(see Table 2).
The most striking result for Y70W was that it
became an almost entirely pyrimidine-specific kinase,
because phosphorylation of dAdo and dGuo was
almost abolished, compared with the wild-type. The
pyrimidine nucleoside analogues AraT, AraC, FdUrd
and BVDU were also phosphorylated by Y70W. The
purine analogue CdA was phosphorylated but less effi-
ciently compared with the wild-type, in accordance
with the lowered activity with purines for Y70W.
Mutant Q81N was slightly less efficient towards the
nucleosides compared with the wild-type. In particular,
phosphorylation of dGuo was reduced markedly.
Structure of Dm-dNK-E52D in complex with dTTP
The structure of one of the mutants, E52D, was solved
using X-ray crystallography at a resolution of 2.5 A
˚
in
complex with the feedback inhibitor dTTP. The R-fac-
tor and R
free
were 23.2 and 24.2%, respectively
(Table 3). Most of the protein could be found in
the electron-density map, with the exception of resi-
dues 1–11 and 210–230. The loop connecting a9 and
b5 (residues 195–200) was flexible and poor density
was observed. The electron density for the mutant resi-
due and the ligand was well defined. Structural super-
positioning of Dm-dNK-E52D–dTTP to the wild-type
Dm-dNK–dTTP (PDB ID: 1OE0) was performed and

F-AraA (%) 19 ± 6 ND 4 (5.0)
CdA (%) 120 ± 11 11.5 ± 0.5 (14.7) 96 (120)
ACV (%) NI ND NI
FdUrd (%) 48 ± 0 37 ± 8 (47.2) 23 (28.9)
BVDU (%)
a
54 19.5 ± 8.5 (24.9)
a
[15].
Active site mutation of deoxyribonucleoside kinase L. Egeblad-Welin et al.
1546 FEBS Journal 274 (2007) 1542–1551 ª 2007 The Authors Journal compilation ª 2007 FEBS
Mg ion that coordinates the phosphates of dTTP
(Fig. 3).
Discussion
Mutation of residues in the active site was intended to:
(a) validate the proposed reaction mechanism [3] by
mutating the putative catalytic base (E52) and arginine
(R105), thought to stabilize the transition state and
holding E52 in position during catalysis; and (b)
investigate the amino acid residues involved in sub-
strate binding (Y70 and Q81). The steady-state kinetics
of Dm-dNK is compulsory ordered with formation of
a ternary complex [1,2]. Pre-steady-state measurements
indicate that either the catalytic step or a preceding
step is rate determining for the overall forward reac-
tion (R. Browne, G. Andersen, G. Le, B. Munch-
Petersen and C. Grubmeyer, unpublished results).
Therefore, and because ATP is saturating in our
experiments, when evaluating the impact of the muta-
tions from the kinetic data, a change in the K

k
cat
value, although the K
m
value was relatively
unchanged. With its imidazole ring, histidine can act
as both a proton donor and an acceptor in enzymatic
reactions, and it should therefore theoretically be able
to replace glutamate as a base, to some extent. How-
ever, this is not the case. One reason may be an altered
local conformation, because the normal hydrogen
bond network will be affected. Modelling mutation of
E52H into the structure of wild-type Dm-dNK with
dCyd bound (PDB ID: 1J90) reveals an increase in the
distance between the 5¢-OH group and histidine of
 1A
˚
relative to glutamate. This alone could explain
the reduced catalytic rate.
Table 3. Data collection and refinement statistics for the dNK-E52D
in complex with dTTP
Dm-dNK E52D
Space group P2
1
Cell dimensions (A
˚
)a¼ 33.5
b ¼ 119.5
c ¼ 68.9
b ¼ 92.42°

1OE0) and Dm -dNK-E52D (red) illustrating the binding of the feed-
back inhibitor dTTP and the position of Mg
2+
in the active site.
L. Egeblad-Welin et al. Active site mutation of deoxyribonucleoside kinase
FEBS Journal 274 (2007) 1542–1551 ª 2007 The Authors Journal compilation ª 2007 FEBS 1547
E52Q
To further support to the hypothesis of the role of E52
as a proton abstractor it was also mutated to its amide –
glutamine. This mutant did not show any detectable
activity with either dThd or dCyd. The two amino acids
take up roughly the same volume, the position of the
side chain can be expected to occupy roughly the same
position and glutamine can participate in hydrogen
bonding. The result that E52Q did not show any activity
must therefore be a consequence of the reactivity of the
functional group and further support the hypothesis
that E52 acts as an initiating base in the reaction.
R105 mutations
R105H
R105 is thought to stabilize the transition state and
hold E52 in the correct position to initiate the catalytic
reaction. The transition state of this type of kinases is
considered to be close to trigonal bipyramidal geo-
metry of the phosphate to be transferred. Arg105 as
well as a Mg ion and arginines of the Lid-region sta-
bilize the negative phosphates of this state. The most
striking effect of mutation of arginine to histidine is an
almost 2000-fold decrease in the k
cat

When Y70 was mutated to W the kinetic results
showed that the K
m
values with dThd and dCyd were
dramatically increased, whereas the k
cat
values were
only slightly decreased. Thus, the mutation primarily
affects binding of the substrate. A similar point muta-
tion was made in HSV1-TK, namely Y101 to F. In
this study, the K
m
value for HSV1-TK-Y101F was
increased 12.5-fold, whereas the k
cat
value was twofold
lower [18]. The structure of HSV1-TK-WT was deter-
mined in complex with (North)-methanocarba-thymidine,
as was the structure of HSV1-TK-Y101F. A structural
superposition showed that there were no significant
changes in the polypeptide chain, except that the
hydrogen bond from Y101–3¢-OH was lost [18]. Based
on our results and the information from the structures
of HSV1-TK we suggest that the network of hydrogen
bonds is disrupted, and this gives rise to an increase in
K
m
. Also, the polarity is changed, and the increase in
the size of the side chain may create steric hindrance
for the substrate, making the base moiety of the sub-

ged, and this may be the reason for the low k
cat
value.
Q81N
The kinetic data for Q81N show a dramatic increase
in the K
m
values, whereas k
cat
is decreased slightly.
These results suggest that the tight anchoring of the
base (dThd or dCyd) is lost, thereby resulting in
poorer binding and higher K
m
values.
Active site mutation of deoxyribonucleoside kinase L. Egeblad-Welin et al.
1548 FEBS Journal 274 (2007) 1542–1551 ª 2007 The Authors Journal compilation ª 2007 FEBS
The relative phosphorylation of dAdo and dGuo
also showed a significant decrease compared with the
wild-type, most likely because of poorer binding of the
purine substrates.
In a previous study of HSV1-TK, Q125 (equivalent to
Q81 in Dm-dNK) was point mutated to N. The struc-
tures of both wild-type HSV1-TK and HSV1-
TK-Q125N were solved in complex with dThd, and the
main difference between the two structures was that the
tight binding of dThd (wild-type) was replaced by a
single water-mediated hydrogen bond (Q125N). Their
kinetic data supported this, because the K
m

) and
[
32
P]ATP[cP] (3000 CiÆmmol
)1
) were purchased from Amer-
sham Biosciences ( Uppsala, Sweden). BVDU (14.3 CiÆmmol
)1
),
1-b-d-arabinofuranosyl thymine (2.89 CiÆmmol
)1
) and 1-b-d-
arabinofuranosyl cytosine (23.30 CiÆmmol
)1
) were from
Moravek Biochemicals Inc. (Brea, CA). Radiolabelled nucle-
osides were diluted with the nonradioactive compounds to
the appropriate concentrations. When present in the radiola-
belled deoxynucleosides, ethanol was evaporated before use.
Non-radioactive nucleosides were from Sigma. Materials for
cloning, PCR, DNA sequencing, assay and crystallization
were standard commercially available products.
Site-directed mutagenesis and expression
plasmid
Expression plasmid pGEX-2T-Dm-dNK has been described
previously [2]. All mutants were constructed using site-
directed mutagenesis on the plasmid pGEX-2T-Dm-dNK
with truncation for 20 terminal amino acids. The primers
used to create the point mutations, where the changed
nucleotides are in boldface and underlined, are as follows:

Sequence verification
Plasmids of the seven mutants were transformed into XL1-
Blue Supercompetent Cells. Plasmids were isolated and the
insert sequenced using the dye terminator method (ABI
PRISM 310), in order to verify that the point mutations
were introduced, and that no other mutations or frame-
shifts had occurred.
Expression and purification
The seven pGEX-2T Dm-dNKDC20 mutants were trans-
formed into E. coli BL21-competent cells. Recombinant
proteins were expressed and purified and thrombin was
cleaved as described previously [2].
All proteins were stored at )80 °C, and a cryoprotectant
solution was added to a final concentration of: 10% (v ⁄ v)
glycerol, 0.1% (v ⁄ v) Triton X-100, 5 mm MgCl
2
and 5 mm
dithiothreitol, with the exception of Dm-dNK E52D CD20;
it was stored in 30% glycerol.
The purity of the proteins was determined by
SDS ⁄ PAGE [20] and the protein concentrations were deter-
mined using Bradford reagent [21].
Enzyme assays
Deoxynucleoside kinase activities were determined by initial
velocity measurements based on four time samples (0, 4,
8 and 12 min) using the DE-81 filter paper assay with trit-
ium-labelled substrates as described previously [2].
The standard assay conditions were: 50 mm Tris ⁄ HCl
pH 7.5, 2.5 mm MgCl
2

sion analysis and the Michaelis–Menten equation v ¼
V
max
Æ[S] ⁄ (K
m
+[S]) or the Hill equation v ¼ V
max
Æ[S]
n
⁄
(K
n
0:5
+[S]
n
) as described previously [22]. All kinetic data
were analysed using sigma plot.
Crystallization
Crystals of a C-terminally truncated (D20) recombinant
Dm-dNK mutant E52D were grown using the vapour diffu-
sion method by hanging drop geometry. The crystallization
solution was: 0.12 m NaAc pH 7.0, 0.1 m Mes pH 6.5 and
18% (w ⁄ v) monomethyl polyethylene glycol 2000. The
enzyme solution consisted of 5 mgÆmL
)1
mutant enzyme in
a1· NaCl ⁄ P
i
buffer with 5 mm dTTP, 5 mm Mg
2+

statistics are shown in Table 3. The coordinates have been
deposited with the PDB ID: 2jcs.
Acknowlegdements
This work was supported by grants from the Swedish
Research Council (to HE), the Swedish Cancer Foun-
dation (to HE), the Danish Research Council (to
BM-P) and the NOVO Nordisk foundation (to BM-P)
References
1 Munch-Petersen B, Piskur J & Søndergaard L (1998)
Four deoxynucleoside kinase activities from Drosophila
melanogaster are contained within a single monomeric
enzyme, a new multifunctional deoxynucleoside kinase.
J Biol Chem 273, 3926–3931.
2 Munch-Petersen B, Knecht W, Lenz C, Søndergaard L
& Piskur J (2000) Functional expression of a multisub-
strate deoxynucleoside kinase from Drosophila melano-
gaster and its C-terminal deletion mutants. J Biol Chem
275, 6673–6679.
3 Eriksson S, Munch-Petersen B, Johansson K & Eklund
H (2002) Structure and function of cellular deoxyribo-
nucleoside kinases. Cell Mol Life Sci 59, 1327–1346.
4 Johansson K, Ramaswamy S, Ljungcrantz C, Knecht
W, Piskur J, Munch-Petersen B & Eklund H (2001)
Structural basis for the substrate specificities of cellular
deoxynucleoside kinases. Nat Struct Biol 8, 616–620.
5 Sabini E, Ort S, Monnerjahn C, Konrad M & Lavie A
(2003) Structure of human dCK suggest strategies to
improve anticancer and antiviral therapy. Nat Struct
Biol 10, 513–519.
6 Wild K, Bohner T, Aubry A, Folkers G & Schulz GE

(2007) Enhanced toxicity of purine nucleoside analogs
in cells expressing Drosophila melanogaster nucleoside
kinase mutants. Gene Ther 14, 86–92.
13 Zheng X, Johansson M & Karlsson A (2000) Retro-
viral transduction of cancer cell lines with the gene
encoding Drosophila melanogaster multisubstrate
deoxyribonucleoside kinase. J Biol Chem 275, 39125–
39129.
14 Zheng X, Johansson M & Karlsson A (2001) Nucleo-
side analog cytotoxicity and bystander cell killing of
cancer cells expressing Drosophila melanogaster deoxy-
ribonucleoside kinase in the nucleus or cytosol. Biochem
Biophys Res Commun 289, 229–233.
15 Knecht W, Sandrini MP, Johansson K, Eklund H,
Munch-Petersen B & Pis
ˇ
kur J (2002) A few amino acid
substitutions can convert deoxyribonucleoside kinase
specificity from pyrimidines to purines. EMBO J 21,
1873–1880.
16 Wild K, Bohner T, Folkers G & Schulz GE (1997) The
structure of thymidine kinase from herpes simplex type
1 in complex with substrates and a substrate analogue.
Protein Sci 6, 2097–2106.
17 Mikkelsen NE, Johansson K, Karlsson A, Knecht W,
Andersen G, Piskur J, Munch-Petersen B & Eklund H
(2003) Structural basis for feedback inhibition of the
deoxyribonucleoside salvage pathway: studies of the
Drosophila deoxyribonucleoside kinase. Biochemistry 44,
5706–5712.

copy search in molecular replacement. Acta Crystallogr
D Biol Crystallogr 56, 1622–1624.
26 Jones TA, Zou JY, Cowan SW & Kjeldgaard M (1991)
Improved methods for building protein models in elec-
tron density maps and the location of errors in these
models. Acta Crystallogr 47, 110–119.
27 Murshudov GN, Vagin AA & Dodson EJ (1997)
Refinement of macromolecular structures by maximum-
likelihood method. Acta Crystallogr D53, 240–255.
L. Egeblad-Welin et al. Active site mutation of deoxyribonucleoside kinase
FEBS Journal 274 (2007) 1542–1551 ª 2007 The Authors Journal compilation ª 2007 FEBS 1551