Mosquito has a single multisubstrate deoxyribonucleoside kinase characterized by unique substrate specificity

Abstract

In mammals four deoxyribonucleoside kinases, with a relatively restricted specificity, catalyze the phosphorylation of the four natural deoxyribonucleosides. When cultured mosquito cells, originating from the malaria vector Anopheles gambiae, were examined for deoxyribonucleoside kinase activities, only a single enzyme was isolated. Subsequently, the corresponding gene was cloned and over-expressed. While the mosquito kinase (Ag-dNK) phosphorylated all four natural deoxyribonucleosides, it displayed an unexpectedly higher relative efficiency for the phosphorylation of purine versus pyrimidine deoxyribonucleosides than the fruit fly multisubstrate deoxyribonucleoside kinase (EC 2.7.1.145). In addition, Ag-dNK could also phosphorylate some medically interesting nucleoside analogs, like stavudine (D4T), 2-chloro-deoxyadenosine (CdA) and 5-bromo-vinyl-deoxyuridine (BVDU). Although the biological significance of multisubstrate deoxyribonucleoside kinases and their diversity among insects remains unclear, the observed variation provides a whole range of applications, as species specific and highly selective targets for insecticides, they have a potential to be used in the enzymatic production of various (di-)(deoxy-)ribonucleoside monophosphates, and as suicide genes in gene therapy.

INTRODUCTION

Deoxyribonucleoside kinases catalyze the phosphorylation of deoxyribonucleosides (dN) to the corresponding deoxyribonucleoside monophosphates (dNMP). They are the key enzymes in the salvage of dN originating from food or intra-cellular breakdown of DNA. Subsequently dNMPs are phosphorylated into diphosphates and triphosphates, which are the precursors of DNA. In mammals there are four deoxyribonucleoside kinases with overlapping specificities, but none of them is able to phosphorylate all four natural deoxyribonucleosides (for review see 1). In bacteria thymidine specific kinases are widely spread, but also additional kinases able to phosphorylate (deoxy)adenosine (dAdo), (deoxy)cytidine (dCyd) and (deoxy)guanosine (dGuo) have for example been shown in the Lactobacilli (2) and Bacilli genera (3,4).

Recently, it was shown that in the insect Drosophila melanogaster there was only one multisubstrate deoxyribonucleoside kinase (dNK, EC 2.7.1.145) present with the capacity to phosphorylate all four natural deoxyribonucleosides [dAdo, dCyd, dGuo, dThd (thymidine)] with equally high turnover and with higher efficiency than the mammalian kinases (5–7). However, this enzyme preferentially phosphorylates pyrimidine substrates. Additionally, this D.melanogaster dNK enzyme, also called Dm-dNK, was able to phosphorylate a wide range of nucleoside analogs. The unique properties of Dm-dNK have made this enzyme an attractive candidate for a number of applications, as a suicide gene in cancer gene therapy (8,9), in the enzymatic synthesis of d(d)NTPs and their analogs (7,8), and as a target for the development of new, nucleoside analog based, insecticides (10).

An open and interesting question is the biological significance of the multisubstrate Dm-dNK in the fruit fly. To date we have described one other multisubstrate deoxyribonucleoside kinase from the insect Bombyx mori (Bm-dNK) (11). However, only a recombinant form of Bm-dNK has been studied and so far it is not known how many deoxyribonucleoside kinases are present in this insect. Bm-dNK differs significantly from Dm-dNK with regard to the kinetic patterns displayed (11), which suggests that kinetic parameters and substrate specificities might not be well conserved among deoxyribonucleoside kinases from different insect species.

In this project the deoxyribonucleoside salvage pathway in different mosquito cell lines from Anopheles gambiae was investigated. Mosquitos are vectors for serious human diseases, for example the mosquito species A.gambiae is the most important malaria vector in Africa (12). According to the World Health Organization (1999), more than 500 million humans are infected with malaria each year resulting in more than 2 million fatal cases. Vector control is one of the most important ways to fight malaria. Therefore, characterization and understanding of the mosquito deoxyribonucleoside metabolism might identify new targets for specific insecticides to control this malaria vector.

In this study, A.gambiae cell lines were shown to possess a high capacity for phosphorylation of purine deoxyribonucleosides. This was shown to be due to a multisubstrate deoxyribonucleoside kinase (Ag-dNK) with significantly different kinetic and substrate specificity parameters than Dm-dNK or Bm-dNK.

MATERIALS AND METHODS

Insect cell lines and culture

The A.gambiae cell lines 4a-2s4, Sua1.1 and L3-5-3 were kindly provided by Dr H. M. Müller (EMBL, Heidelberg). These cell lines had been established from newly hatched larvae via a homogenization technique (13). L3-5-3 is described by Vizioli et al. (14). Culturing of D.melanogaster S-2 cells is described in Munch-Petersen et al. (6).

The insect cells were grown in tissue culture flasks (75 cm² from Greiner GmbH) in Schneider’s medium obtained from Sigma with addition of 10% heat-inactivated fetal calf serum (GibcoBRL Life Technologies).

Preparation of cell extracts

Crude cellular extracts were prepared by harvesting ∼5 × 10⁸ cells of each cell line by centrifugation for 20 min at 2700 g and the pellet was resuspended in 0.5 ml buffer A (20 mM K₂HPO₄, pH 7.4, 15% glycerol, 1 mM EDTA, 1 mM DTT). The cells were then disrupted by sonification and centrifuged for 30 min at 12 000 g to separate insoluble debris.

Column materials

Sephadex G-25, DEAE Sepharose CL-B6 and phenyl– Sepharose High Performance were obtained from Pharmacia Biotech Inc. 3′-dTMP Sepharose gel-matrix and 5′-dTMP Sepharose gel-matrix had been prepared according to the procedures described previously (15,16).

Protein purification

Step I: preparation of crude extract. Approximately 8.7 × 10¹⁰ A.gambiae cells of cell line 4a-2s4 were harvested by centrifugation for 20 min at 2700 g and the pellet was resuspended in 27 ml buffer A. The cells were then disrupted with a French press followed by centrifugation for 40 min at 13 000 g and the crude extract was collected (Fraction I).

Step II: streptomycin/ammonium sulfate precipitation and G-25 chromatography. The nucleic acids were precipitated with streptomycin sulfate to the final concentration of 0.7%. The resulting supernatant was fractionated with ammonium sulfate in two steps (20 and 65%) as described by Bohman and Eriksson (17). All centrifugations were performed for 20 min at 12 000 g. The ammonium sulfate pellet was suspended in buffer B {20 mM Tris–HCl, pH 8.0, 5 mM MgCl₂, 5 mM NaF, 10% glycerol, 2 mM DTT, 0.5 mM 3-[(3-cholamidopropyl)-dimethylammonio]-1-propanesulfonate (CHAPS), 50 mM ε-aminocaproic acid}, divided into two portions of 45 ml and desalted by passing over a Sephadex G-25 column (bed volume, 50 × 255 mm) with buffer B. The peak fractions were collected and pooled (Fraction II).

Step III: DEAE-ion exchange chromatography. Fraction II was applied to DEAE-ion exchange chromatography. The column (bed volume, 25 × 100 mm) was equilibrated in buffer C (20 mM Tris–HCl, pH 8.0, 5 mM MgCl₂, 10% glycerol, 2 mM DTT, 0.5 mM CHAPS). Unbound material was washed away with buffer C. The bound proteins were eluted by a 0–0.25 M KCl gradient in buffer C (8 bed volumes) followed by 0.5 M KCl in buffer C.

Step IV: 5′-aminothymidine–Sepharose chromatography. Fractions from the main peak were divided into three portions of ∼6–8 ml. Then, 0.5 mM phenylmethylsulfonylfluoride and 5 mM benzamidine were added and the portions applied to a 5′-aminothymidine–Sepharose column with thymidine bound to the matrix through the 5′-OH group. The column (bed volume 10 × 13 mm) was equilibrated in buffer D (10 mM Tris–HCl, pH 8.0, 5 mM MgCl₂, 10% glycerol, 2 mM DTT, 0.5 mM CHAPS). Unbound material was washed out with buffer D (15 times bed volume) and buffer E (10 mM Tris–HCl, pH 7.5, 5 mM MgCl₂, 10% glycerol, 2 mM DTT, 0.5 mM CHAPS) (5 times bed volume). The bound protein was eluted with buffer E + 10 mM thymidine (5 times bed volume). The elution fractions containing thymidine kinase (TK) activity were pooled.

Step V: phenyl–Sepharose chromatography. A sample of 1.5 M ammonium sulfate was added to the pooled fractions and the fractions centrifuged for 10 min at 20 000 g. The supernatant was chromatographed on a phenyl–Sepharose column (bed volume 10 × 13 mm) equilibrated with buffer F (50 mM K₂HPO₄, pH 7.5, 5 mM MgCl₂, 10% glycerol, 2 mM DTT) + 1.5 M ammonium sulfate. Unbound material including thymidine was washed out with buffer F + 1.5 M ammonium sulfate (10 times bed volume). Elution of the protein was performed in two steps. First with buffer F (approximately 5 times bed volume) and then with buffer F + 1% Triton X-100 (5 times bed volume).

SDS–PAGE was done according to the procedure of Laemmli (18) and the proteins were visualized by silver staining. The protein concentration was determined according to Bradford with BSA as standard protein (19).

Enzyme assay

Nucleoside kinase activities were determined by initial velocity measurements based on four time samples by the DE-81 filter paper assay using tritium-labeled nucleoside substrates. The assay was done as described in Munch-Petersen et al. (6). The standard assay conditions were: 50 mM Tris–HCl, pH 8.0, 2.5 mM MgCl₂, 10 mM DTT, 0.5 mM CHAPS, 3 mg/ml BSA, 2.5 mM ATP and 5 µM radiolabeled substrate (unless indicated otherwise). Radioactive nucleoside analogs were obtained from Moravek or Amersham. One unit (U) of nucleoside kinase activity is defined as 1 µmol of the corresponding monophosphate formed per minute.

Analysis of kinetic data

Kinetic data were evaluated as described in Knecht et al. (20) by non-linear regression analysis using the Michaelis–Menten equation v = V_max × [S]/(K_m + [S]).

Cloning of a multisubstrate deoxyribonucleoside kinase from A.gambiae

When deposited sequences from A.gambiae (GenBank) were analyzed for homology to Dm-dNK a putative dNK with an open reading frame (ORF) of 741 bp could be predicted. Subsequently this ORF was amplified from A.gambiae cDNA. Total RNA from A.gambiae was isolated from ∼5 × 10⁶ cells from the 4a-2s4 cultured cell line, using the RNAqueous™ kit (Ambion, TX). cDNA was prepared using the Advantage™ RT-for-PCR kit (Clontech Laboratories Inc., CA). The primer used was 2MSAgdNK: 5′-GTATGTCCAATTCGAATGGTAATAATG-3′. Both kit procedures were according to the instructions provided by the manufacturers.

The ORF of the A.gambiae multisubstrate deoxyribonucleoside kinase was amplified by PCR using the primers 1MSAgdNK-B-1: 5′-CGCGGATCCATGCCTCCGATAGCGAGCGAAAAGTTAGGCGCC-3′ and 2MSAgdNK-E: 5′-CCGGAATTCTCAGAAGTCCGTCTTGGCTCGCTTCGC-3′ and the isolated cDNA as the template. The PCR fragment was subsequently cut by BamHI/EcoRI and ligated into pGEX-2T vector (Amersham-Pharmacia) that was also cut by BamHI/EcoRI. The resulting plasmid, P665, was named pGEX-2T-Ag-dNK. The ORF in pGEX-2T-Ag-dNK was verified by sequencing and the ORF was deposited in the GenBank™ database with the accession number AF488801.

Sequence alignments were performed using CLUSTAL W (21) and PSORTII (22) was used to predict sorting signals in sequences.

Expression and purification of recombinant A.gambiae dNK (rAg-dNK)

Escherichia coli KY895 (F^–, tdk-1, ilv) (11) was transformed by pGEX-2T-Ag-dNK (P665) using standard techniques. Transformed KY895 were grown to an A_600nm of 0.5–0.6 in LB/ampicillin (100 µg/ml) medium at 37°C and protein expression was induced by adding isopropyl-β-d-thiogalactoside to a final concentration of 100 µM. The cells were further grown for 4 h at 25°C and subsequently harvested by centrifugation. The cell pellet was subjected to sonication in the binding buffer A (20 mM NaPO₄ pH 7.3; 150 mM NaCl; 10% glycerol; and 0.1% Triton X-100) in presence of a protease inhibitor cocktail (Complete™ – EDTA free from Roche Dignostics), subjected to centrifugation at 10 000 g for 30 min, filtered and loaded onto the column. A 1 ml column (glutathione–Sepharose available from Pharmacia) was equilibrated in binding buffer A. After loading of the sample, the column was washed with 20 ml of binding buffer A. Subsequently the column was washed with 2.5 ml 10 mM ATP/MgCl₂ in (A) and incubated for 1 h at room temperature and then 30 min at 4°C. Afterwards the column was washed again with 5 ml of buffer A and 1 ml of thrombin (50 U/ml) solution was applied on the column. The column was gently shaken O/N at 4°C to efficiently cleave the rAg-dNK from the glutathione S-transferase (GST)-tag. The rAg-dNK was eluted from the column in buffer A.

RESULTS

Characterization of the cell lines

Three A.gambiae cell lines were grown to be tested for the deoxyribonucleoside kinase activity. The doubling times for the A.gambiae cell lines 4a-2s4, Sua1.1 and L3-5-3 were 32, 38 and 29 h, respectively, making L3-5-3 the fastest growing cell line. Crude extracts from the three mosquito cell lines and the D.melanogaster cell line S-2 were prepared and analyzed for their capacity to phosphorylate the four natural deoxyribonucleosides dAdo, dCyd, dGuo and dThd. The activities and the activities relative to the TK activity within each cell line are given in Table 1. The three A.gambiae lines had the capacity to phosphorylate all four natural deoxyribonucleosides. While the D.melanogaster cell line phosphorylated dThd and dCyd with an 8-fold higher rate than dAdo and dGuo, the mosquito cell lines phosphorylated dAdo and dGuo with an equal or higher activity than dThd and dCyd. In the cell line Sua1.1 an up to 3-fold higher enzyme activity for purine deoxyribonucleosides compared with pyrimidine deoxyribonucleosides was found. These results suggested that the organization and characteristics of enzymes involved in phosphorylation of deoxyribonucleosides differ between the two species.

Table 1. Extracts of the insect cell lines were analyzed for their capacity to phoshorylate the natural deoxyribonucleosides dThd, dCyd, dAdo and dGuo.

	Anopheles gambiae cell lines						Drosophila melanogaster
	4a-2s4		Sua1.1		L3-5-3		S-2
Substrate (100 µM)	mU/1010 cells	%	mU/1010 cells	%	mU/1010 cells	%	mU/1010 cells	%
dThd	3.8	100	2.6	100	3.2	100	6.8	100
dCyd	4.2	112	1.8	69	4.4	143	6.8	100
dAdo	4.4	114	5.6	216	5.8	185	0.8	12
dGuo	5.0	130	4.0	156	4.4	139	0.8	12

The results were normalized to 10¹⁰ cells. For easier comparison, the results are also presented as percent of the dThd phosphorylating capacity (100%) of each cell line.

Purification of a multisubstrate deoxyribonucleoside kinase from A.gambiae (Ag-dNK)

Even though L3-5-3 was the fastest growing wild-type line, the 4a-2s4 line was chosen for the large-scale cell production for protein purification because of its ability to attach only loosely to the culture flask surfaces resulting in a minimal loss of cells during harvesting. The purification started from ∼8.7 × 10¹⁰ A.gambiae cells and was a modified version of the procedure used to isolate Dm-dNK from D.melanogaster S-2 cells (6). A surprising discovery was made during Step III of the purification procedure, the DEAE-ion exchange chromatography (Fig. 1). The different deoxyribonucleoside kinases in mammalian cells can be separated by DEAE-ion exchange chromatography (23). However, in D.melanogaster S-2 cells only one activity peak for all four natural deoxyribonucleosides was found (6). In the cell line 4a-2s4 we found, after testing all fractions with the four natural deoxyribonucleosides, two additional minor peaks (named A and B) with dGuo and/or dAdo phosphorylating activity (Fig. 1a). Peak C was the activity peak covering all four deoxyribonucleosides tested and contained the fractions used for further purification of Ag-dNK. The peaks A and B, eluting from the DEAE-ion exchange chromatography at lower KCl concentration, were retested for their activity with ribonucleosides (Fig. 1b). Two prominent peaks of activity, with Ado and Guo phosphorylating activity (fraction 27) and Ado phosphorylating activity (fraction 32) were identified. The relative activities of these two fractions with other substrates were further investigated and are presented in Table 2. The activity profile of these two fractions suggests that they are composed of adenosine kinase-like enzyme(s), which also exhibit a partial purine and pyrimidine deoxyribonucleoside kinase activity.

Figure 1.

Elution profile from DEAE-ion exchange chromatography. Activities were measured with 10 (a) or 100 µM (b) (deoxy-)ribonucleoside substrate. Activity peaks were arbitrarily named A, B and C.

Table 2. Substrate specificities of DEAE chromatography fractions 27 and 32 (activity peak B).

Substrate (5 µM)	#27 (%)	#32 (%)
Ado	100	100
Guo	59 ± 2.5	3.5 ± 1.5
Cyd	n.d.	n.d.
dThd	n.d.	0.16 ± 0.03
dCyd	0.08	0.27 ± 0.11
dAdo	n.d.	0.26 ± 0.2
dGuo	1.4	0.42 ± 0.36
ddC	n.d.	n.d.
ddA	0.09 ± 0.08	0.024 ± 0.005
ddG	n.d.	n.d.
FddThd	n.d.	n.d.
FdUrd	n.d.	0.04 ± 0.03
AraT	n.d.	n.d.
AraC	n.d.	n.d.
AZT	n.d.	n.d.
D4T	n.d.	0.011 ± 0.009
CdA	0.14 ± 0.13	0.48 ± 0.08
ACV	n.d.	n.d.
BVDU	n.d.	0.27 ± 0.15

The activities are relative to the activity with Ado that was set to 100% for each fraction. Relative activities are means from three measurements ± SD. n.d., not detectable. AraC, 1-β-d-arabinofuranosylcytosine; AraT, 1-β-d-arabinofuranosylthymine; AZT, 3′-azido-2′,3′-dideoxythymidine, FdUrd, 5-fluorodeoxyuridine.

The fourth and fifth purification step revealed further differences from the Dm-dNK purification. In contrast to Dm-dNK, Ag-dNK did not bind to 3′-dTMP Sepharose, where thymidine is linked to the matrix through the 3′ group, but bound strongly to thymidine linked through the 5′-OH group (data not shown). Another difference was that the mosquito enzyme, in contrast to Dm-dNK, could not be recovered from the hydrophobic phenyl–Sepharose chromatography by elution with CHAPS, but had to be eluted by Triton X-100. The purification is summarized in Table 3, and Figure 2 shows purified Ag-dNK in comparison with recombinant Dm-dNK (7). Unfortunately, the high detergent and a low protein content in the Ag-dNK fractions, originating from the 5′ dTMP Sepharose and the phenyl–Sepharose chromatography, did not allow determination of the protein concentration by the Bradford method.

Table 3. Purification table for the purification of A.gambiae multisubstrate deoxyribonucleoside kinase (Ag-dNK) from 8.7 × 10¹⁰ cells of the A.gambiae 4a-2s4 cell line.

	Total protein (mg)	Total TK activity (mU)	Specific activity (mU/mg)	Purification (fold)
I. Crude extract	769	36.3	0.0472	1
II. Desalted ammonium sulfate fraction	489	27.9	0.0571	1.31
III. DEAE Sepharose (Peak C)	66	22.2	0.3364	7.13
IV. 5′-TMP Sepharose	a	17.1
V. Phenyl–Sepharose	a	14.4

TK activity was measured at 10 µM dThd.

^aFractions resulting from the 5′-dTMP Sepharose and the phenyl–Sepharose chromatography had a too high detergent and a too low protein content to allow determination of the protein concentration by the Bradford method.

Figure 2.

SDS–PAGE of purified A.gambiae multisubstrate deoxyribonucleoside kinase (Ag-dNK). The gel was silver stained. Lane 1, Ag-dNK (60 µl from fraction V); lane 2, 0.25 µg rDm-dNK; lane 3, 0.1 µg rDm-dNK.

Substrate specificity of Ag-dNK

Purified Ag-dNK phosphorylated all four natural deoxyribonucleosides as shown in Table 4 but characterized by the ratio, which is very different from Dm-dNK. The relative specificity of Ag-dNK with ribonucleosides and nucleoside analogs was also investigated and is presented in Table 4. Surprisingly and in contrast to Dm-dNK (6,7) we could not show phosphorylation of ribonucleosides or dideoxyribonucleosides by Ag-dNK. On the other hand, Ag-dNK was able to phosphorylate a few medically interesting nucleoside analogs including 2′,3′-didehydro-3′-deoxythymidine (D4T), 5-bromo-vinyl-deoxyuridine (BVDU) and 2-chloro-deoxyadenosine (CdA) with substantially higher activity than recombinant Dm-dNK (BVDU and CdA) (7) or native Dm-dNK (D4T) (6). This sustains the fact that the multisubstrate deoxyribonucleoside kinase from mosquitos differs significantly from other deoxyribonucleoside kinases described so far.

Table 4. Substrate specificities of native Ag-dNK and Dm-dNK.

Substrate (5 µM)	Ag-dNK (%)	Dm-dNK (%)a
dThd	100	100
dCyd	142 ± 30	100
dAdo	76 ± 20	5
dGuo	158 ± 59	1
Cyd	n.d.	–
Ado	n.d.	–
Guo	n.d.	–
ddC	n.d.	0.2
ddA	n.d.	–
ddG	n.d.	–
FddThd	n.d.	0.3
FdUrd	29 ± 3.9	110
AraT	6.1 ± 4.2	15
AraC	6.7 ± 1.8	10
AZT	n.d.	0.5
D4T	4.6 ± 3.7	0.5
CdA	311 ± 96	–
ACV	n.d.	<0.1
BVDU	192 ± 63	–

The activities are relative to the activity with dThd that was set to 100%. Relative activities are means from three to six measurements ± SD.

^aComparable data for native Dm-dNK are from Munch-Petersen et al. (6). n.d., not detectable; –, not reported.

Cloning and expression of recombinant Ag-dNK

As a result of the ongoing international efforts in sequencing the genome of A.gambiae (http://www.niaid.nih.gov/newsroom/releases/celera.htm), we were able to identify and subsequently subclone an ORF for Ag-dNK from the cell line 4a-2s4. The ORF encodes a protein of 246 amino acids and a calculated mass of 28.1 kDa. This fits reasonably well with the apparent molecular weight observed for native Ag-dNK, running just below recombinant Dm-dNK (29 kDa) in the SDS–PAGE shown in Figure 2. When the Ag-dNK amino acid sequence (Fig. 3) was examined by the signal prediction method for cellular transport signals (22) one patch of amino acids matching a potential nuclear import signal (PAKRAKT) was identified beginning at amino acid 238 and located at the very C-terminus of the enzyme like the nuclear import signal in Dm-dNK (7,9). In contrast, no nuclear import signal could be predicted in Bm-dNK (11). The alignment of Ag-dNK, Dm-dNK and Bm-dNK ORFs (Fig. 3) indicates a very high identity among the three insect kinases. Only the N- and C-termini are slightly divergent. Another difference between the three enzymes is the calculated isoelectric points with 5.99, 7.76 and 8.19 for Ag-dNK, Bm-dNK and Dm-dNK, respectively.

Figure 3.

Multiple alignment of the predicted amino acid sequences of the known insect deoxyribonucleoside kinases, Ag-dNK, Dm-dNK and Bm-dNK. Dm-dNK residues closest to both substrates, as determined within the crystal structure of rDm-dNKΔC20 (25), are marked with arrows. Above the sequence the phosphate binding P-loop and the LID region are marked (25). The putative nuclear import signals in Ag-dNK and Dm-dNK (7,9) are marked.

Kinetic properties of rAg-dNK

The full-length ORF of Ag-dNK was over-expressed as a fusion protein with GST. rAg-dNK could be purified to nearly homogeneity by gluthatione–Sepharose affinity chromatography followed by thrombin cleavage (data not shown). The relation between velocity and substrate concentration was determined in the same way as for recombinant Dm-dNK (7) and Bm-dNK (11) for the four natural deoxyribonucleosides and compared with the catalytic efficiencies of recombinant Dm-dNK and Bm-dNK. Phosphorylation of deoxyribonucleoside substrates by deoxyribonucleoside kinases is a bi-substrate reaction and for Dm-dNK it has been show to follow a compulsory ordered steady-state reaction mechanism with formation of a ternary complex with the phosphate donor and acceptor (6). Because the phosphate donor concentration (2.5 mM ATP) was well above the expected K_m, the obtained K_m values for deoxyribonucleoside substrates are expected to be close to the real K_m values. Similarly, the observed V_max is likely to be close to the real V_max because both substrates were available at saturating concentrations (Table 5). In contrast to rDm-dNK or rBm-dNK-8His, rAg-dNK only shows an ∼2-fold difference in the catalytic efficiency between the most preferred substrate dThd and the least efficiently converted substrate dAdo. Nevertheless, the catalytic efficiency for the conversion of dThd to TMP by rAg-dNK is ∼5000-fold lower when compared with the efficiency by which rDm-dNK catalyzes this reaction.

Table 5. Apparent kinetic parameters of recombinant A.gambiae multisubstrate deoxyribonucleoside kinase (rAg-dNK).

Substrate	rAg-dNK				rDm-dNKa	rBm-dNK-8Hisb
	Km	Vmax	kcat	kcat/Km	kcat/Km	kcat/K0.5
	(µM)	(mU/mg)	(s–1)	(M–1 s–1)	(M–1 s–1)	(M–1 s–1)
dThd	5.8	29	0.014	2414 {100}	1.2 × 107 {100}	1.2 × 105 {100}
dCyd	12.6	65	0.03	2380 {99}	7.2 × 106 {60}	1.8 × 105 {150}
dAdo	18.1	44	0.021	1160 {48}	1.9 × 105 {0.8}	4770 {4}
dGuo	9.5	41	0.019	2000 {83}	2.3 × 104 {0.19}	769 {0.6}

Best fit of the Michaelis–Menten equation to all data. The k_cat values were calculated using the equation V_max = k_cat × [E], where [E] = total enzyme concentration and is based on one active site/monomer. The phosphate donor (2.5 mM ATP) was present at saturating concentration.

^aData from Munch-Petersen et al. (7) and ^bKnecht et al. (11) are also presented for easy comparison. In brackets the relative catalytic efficiencies for rAg-dNK, rDm-dNK and rBm-dNK-8His are given (the catalytic efficiency with dThd was set to 100% for each enzyme).

DISCUSSION

Previously, two multisubstrate deoxyribonucleoside kinases from the fruit fly D.melanogaster and the silkworm Bombyx mori, closely related to the mammalian mitochondrial thymidine kinase (TK2) family (7,11), have been characterized (5–7,11,24). The two enzymes are capable of phosphorylating all four natural deoxyribonucleosides but they clearly prefer pyrimidine substrates. In this study another insect, a mosquito, which is an important disease vector for malaria, was examined for the presence of a multisubstrate deoxyribonucleoside kinase. A significant kinase activity was detected in all three cultivated mosquito cell lines examined (Table 1). The capacity of the mosquito cells to phosphorylate dThd and dCyd was similar to the capacity of the Drosophila S-2 cells. On the other hand, the mosquito capacity to phosphorylate dAdo and dGuo was significantly higher (Table 1). One of the cell lines, 4a-2s4, was selected for further characterization of the deoxyribonucleoside kinase activity. Three separate peaks of deoxyribonucleoside kinase activity were observed upon separation on DEAE-ion exchange chromatography (Fig. 1a). The major peak (peak C) corresponded to a multisubstrate deoxyribonucleoside kinase, similarly as observed in the case of Dm-dNK found previously in D.melanogaster (6). The smaller peaks (peaks A and B) were likely to correspond to ribonucleoside kinases with partial deoxyribonucleoside kinase activities (Table 2 and Fig. 1b). The multisubstrate deoxyribonucleoside kinase from A.gambiae, which we named Ag-dNK, was further purified (Table 3 and Fig. 2) and then the substrate specificity was determined on the purified native enzyme. This was substantially different from the previously determined characteristics of Dm-dNK or Bm-dNK (Table 4) (6,11). Ag-dNK clearly exhibited higher relative activities towards purine substrates than Dm-dNK (Table 4). An important example is also the superior ability of Ag-dNK to phosphorylate the nucleoside analog stavudine (D4T), which is not or only weakly phosphorylated by native Dm-dNK (6) and neither by TK1 nor TK2 (23). Also, a surprisingly high phosphorylation of BVDU (193%) and CdA (311%) in relation to dThd at 5 µM (Table 4) was found when compared with recombinant Dm-dNK, with 54 and 126% for BVDU and CdA, respectively, at 500 µM (7).

Several sequences coding for a putative mosquito kinase were obtained from the database and a putative ORF was deduced from them (Fig. 3). From the phylogentic point of view Dm-dNK and Ag-dNK are related closer to each other than to Bm-dNK (data not shown), reflecting well the phylogenetic relationship among the three insects. The Ag-dNK ORF was over-expressed and the corresponding recombinant enzyme rAg-dNK indeed possessed the kinase activities, similar to the native Ag-dNK (Table 5). While rDm-dNK phosphorylates all four natural deoxyribonucleosides with equally high turnover numbers, it clearly prefers pyrimidines as shown by the catalytic efficiencies in Table 5. Surprisingly, the catalytic efficiency for the conversion of dThd versus dAdo differs only ∼2-fold for rAg-dNK, while for rDm-dNK the difference in catalytic efficiency between dThd and dGuo was 525- and 250-fold between dCyd and dGuo for rBm-dNK-8His. Therefore, Ag-dNK deserves to be called multisubstrate kinase even more than Dm-dNK or Bm-dNK and has the closest efficiencies for all natural deoxyribonucleosides reported so far.

The 3D structure for Dm-dNK has been determined recently (25) and amino acids determining the specificity for pyrimidine deoxyribonucleosides versus purine deoxyribonucleosides have been determined (26). However, as can be seen in Figure 3, among the three enzymes there is only one difference (M118) within the amino acids, which are believed to be the closest to the substrate. A stereo view of dCyd bound to the active center of Dm-dNK is illustrated in Figure 4. Apparently, the shown amino acid residues are not responsible for the observed diversity in substrate specificity among the three enzymes. Residues outside the substrate binding pocket and/or conformational changes taking place during the reaction might play a major role in the specific properties of each insect multisubstrate kinase.

Figure 4.

Stereoview of the nucleoside binding (dCyd) in rDm-dNKΔC20 (25). The only differing amino acid between Dm-dNK, Ag-dNK and Bm-dNK in this region (I105 in Bm-dNK instead of M118 in Dm-dNK and Ag-dNK) is inserted in green. The stereoview was produced using the program Swiss-Pdb Viewer (http://us.expasy.org/spdbv/).

So far the biological significance of the observed differences among Dm-dNK, Bm-dNK and Ag-dNK cannot be explained. However, these variations could be useful for a number of applied aspects, like development of specific and selective insecticides and recruitment of enzymes with novel properties. Balzarini et al. (10) demonstrated the growth inhibiting effect of the nucleoside analog BVDU on D.melanogaster S-2 and Spodoptera frugiperda Sf-9 cells, promoting the idea to use nucleoside analogs as insecticides activated by an insect multisubstrate deoxyribonucleoside kinase. Encouraging results with BVDU have also been reported by Mazzacano and Fallon (27). Considering a high capacity of mosquito cells for purine deoxyribonucleoside phosphorylation, we would also suggest testing of dGuo/dAdo analogs as mosquito insecticides, and for example use CdA as a lead-structure. Mazzacano and Fallon (27) also showed Anopheles albopictus mosquito wild-type cells to be more suspectible to the nucleoside analog acyclic guanosine (ACV) than TK^– cells transformed with the TK from human Herpes simplex virus 1 (HSV 1), the only enzyme so far to accept ACV as a substrate. This suggested the existence of an enzyme in mosquitos with superior ability to phosphorylate ACV than HSV1 TK. In our study of A.gambiae cells, we could not show any phosphorylation of ACV. This might be due to the low concentration of substrates (5 µM) tested by us, or due to the existence of a different deoxyribonucleoside kinase in A.albopictus. In addition, novel insect enzymes, which can convert a spectrum of nucleoside analogs, such as ACV, that are poor substrates for mammalian deoxyribonucleoside kinases, could be valuable in cancer gene therapy as suicide genes, and could also provide improvements in the enzymatic phosphorylation and production of a range of (di-)(deoxy-) ribonucleoside analogs used for biotechnological applications like DNA sequencing (8,9).

DDBJ/EMBL/GenBank accession no. AF488801