Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Recently, a multifunctional deoxyribonucleoside kinase (dNK) was purified from Drosophila melanogaster (Dm) embryonic (S-2) cell cultures and its kinetic properties were characterized (Munch-Peterson et al., 1998a,b). The Dm-dNK showed a broad nucleoside substrate specificity with varying efficiency. The K_m values for dThd, dCyd, dAdo, and dGuo were 0.9, 1.0, 109, and 654 µM, respectively, and the V_max values for the different deoxynucleosides varied only by <2-fold (Munch-Peterson et al., 1998a,b). The recognition of the four different nucleosides by one and the same enzyme is unprecedented and differs markedly from the situation in mammalian cells. Mammalian cytosolic thymidine kinase (TK) is only able to phosphorylate dThd (and dUrd) as the natural substrate (Kit, 1976), whereas mitochondrial TK-2 can phosphorylate both dThd (and dUrd) and dCyd, albeit with a preference for dThd (Kit, 1976; Johansson and Karlsson, 1997). The mammalian dCyd kinase does not recognize dThd as a substrate, but in addition to dCyd, it can also convert dAdo and dGuo to their 5'-monophosphate derivatives (Johansson and Karlsson, 1995). Finally, mitochondrial dGuo kinase recognizes both dGuo and dAdo as a substrate (Johansson and Karlsson, 1996). The cDNA cloning of Dm-dNK revealed a sequence similarity to mammalian dCK, dGK, and TK-2 as well as to the herpes simplex virus (HSV) TK (Johansson et al., 1999). Munch-Petersen and colleagues revealed that the substrate specificity, size, and other characteristics of Dm-dNK is more related to TK-2 than to any of the other mammalian dNKs (Munch-Petersen et al., 1998a,b) and that the highest level of sequence similarity of Dm-dNK cDNA was with TK-2 (Johansson et al., 1999).

TK-2 shares several kinetic properties with herpetic TKs (including the pronounced substrate specificity for both dCyd and dThd) (Gentry et al., 1983), and tumor cells transfected with the gene encoding for the HSV TK becomes exquisitely sensitive to the cytostatic activity of a number of antiherpetic drugs (Balzarini et al., 1985, 1989). The sequence of Dm-dNK also showed that the enzyme is closely related to HSV TKs. Therefore, we decided to evaluate a variety of antiherpetic nucleoside analogues for their affinity for Dm-dNK and wanted to investigate whether Dm-dNK might be a useful target enzyme for the metabolic activation of antiherpetic compounds resulting in a potential cytostatic activity against insect cells. We found that the potent antiherpetic compound (E)-5-(2-bromovinyl)-2'-deoxyuridine (BVDU) (De Clercq et al., 1979) was not only an excellent substrate for Dm-dNK, but that it also seriously retarded or arrested the insect cell cycle progression and insect cell DNA synthesis, resulting in a pronounced selective cytostatic activity against the insect cells compared with mammalian cells. Our data revealed that Dm-dNK is a new, attractive target for the development of insecticidal drugs that are selectively cytostatic to insect, but not mammalian, cells.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Cells. Embryonic cells from the insects D. melanogaster S-2 and S. frugiperda (Sf-9) were cultivated in plastic 75 cm² culture bottles (Iwaki, International Medical, Brussels, Belgium) at 26°C. S-2 cells were seeded in insect cell medium consisting of Schneider's D. melanogaster powder medium (Serva, Boehringer Ingelheim, Heidelberg, Germany) (24.06 g, CaCl₂ 0.6 g, NaHCO₃ 0.4 g in 100 ml aqua distillata), adjusted to pH 6.5 with NaOH, and containing 10% inactivated (56°C, 30 min) fetal calf serum and 0.15% penicillin/streptomycin solution (final solution). Sf-9 cells were cultured in TC 100 basal medium (Life Technologies, Paisley, Scotland) containing 10% heat-inactivated fetal calf serum, 240 nM ZnSO₄.7 H₂O, 16 µM AlK(SO₄)₂.12 H₂O, L-glutamine (1 mg/ml), and dehydrated tryptose phosphate broth (Life Technologies) at 2.6 g/liter. The cells were subcultured twice weekly by scraping from the bottom of the cell culture flasks and were then resuspended in the insect cell culture medium at an initial density of ~250,000 cells/ml.

Compounds and Radiochemicals. [CH₃-³H]dThd (radiospecificity 72 Ci/mmol) was obtained from Amersham (Buckinghamshire, UK), and [6-³H]BVDU (radiospecificity 1.0 Ci/mmol) was obtained from Moravek Biochemicals (Brea, CA). Nonradiolabeled BVDU (Fig. 1) was prepared according to a method described previously (Jones et al., 1979). The 5'-monophosphate (BVDU-MP), 5'-diphosphate (BVDU-DP), and 5'-triphosphate (BVDU-TP) of BVDU were synthesized by P. Herdewijn (Rega Institute, Leuven, Belgium). Thymidine (dThd), thymidine-5'-monophosphate (dTMP), and dTTP were obtained from Sigma Chemical Co. (St. Louis, MO).

View larger version (11K): [in this window] [in a new window]   Fig. 1.   Structural formula of (E)-5-(2-bromovinyl)-2'-deoxyuridine (BVDU).

Expression and Purification of Recombinant Dm-dNK. Dm-dNK was expressed as a fusion protein to glutathione-S-transferase. The cDNA sequence encoding Dm-dNK was cloned in the pGEX-5X-1 plasmid (Pharmacia Biotech, Uppsala, Sweden). The protein was expressed in Escherichia coli and purified as described (Johansson et al., 1999). The purity of the recombinant protein was determined by SDS-polyacrylamide gel electrophoresis (Phast System, Pharmacia Biotech). The protein concentration was determined with the Bradford Protein Assay (Bio-Rad, Richmond, CA), and BSA was used as the concentration standard.

dNK and TK-2 Assays The standard reaction mixture contained 5 mM ATP, 5 mM MgCl₂.6 H₂O, 9 mM KF, 5 mM phosphonoenolpyruvate, 5 µg pyruvate kinase, 10 mM dithiothreitol, 2 µM (0.1 µCi) [CH₃-³H]dThd, 5 µl of the appropriate concentrations of BVDU, BVDU-MP, BVDU-TP, dTMP, or dTTP (i.e., 200, 40, 8, 1.6, 0.32, 0.06, and 0.012 µM), and 10 µl of a purified recombinant Dm-dNK enzyme preparation in a total volume of 40 µl of 0.05 M Tris/HCl, pH 8.0. The reaction mixture was incubated at 37°C for 15 min and the reaction terminated by the addition of 75 µl of ice-cold 0.05 M Tris/HCl buffer, pH 8.0. The reaction proceeded linearly at all substrate concentrations for an incubation period up to >30 min. Then the mixture was applied onto DE81 filter disks to bind phosphorylated [CH₃-³H]dThd, and the remaining nonreacted [CH₃-³H]dThd was removed from the filters by washing with 1 mM ammoniumformate, pH 8.2, ethanol, and ether. The filters were then assayed for radioactivity in a toluene-based scintillation cocktail. The IC₅₀ values of BVDU, BVDU-MP, BVDU-TP, dTMP, and dTTP for [CH₃-³H]dThd phosphorylation was defined as the compound concentration that inhibited the conversion of [CH₃-³H]dThd to its 5'-monophosphate by 50%.

Inhibition of Insect Cell Proliferation by BVDU Approximately 80,000 S-2 or Sf-9 cells were suspended in 800 µl of insect cell growth medium and added to the wells of 48-well plates in the presence of varying concentrations of BVDU. The cells were then allowed to proliferate at 26°C and, twice a day, the cell number was determined by use of a Coulter counter type ZM (Coulter Electronics, Harpenden Hertz, UK).

Incorporation of [6-³H]BVDU into Methanol-Soluble and Methanol-Insoluble Material. The metabolism of radiolabeled [6-³H]BVDU was monitored as follows: S-2 and Sf-9 cells were seeded at 0.5 to 2 × 10⁶ cells/ml in a 5-ml cell culture flask (25 cm²) and incubated with 1 µM [6-³H]BVDU (5 µCi/5 ml). At 24 h, cells were scraped from the culture flask, centrifuged at 4°C for 10 min at 800g, washed 3 times with cold insect cell culture medium (without serum), and extracted with 60% cold methanol. After centrifugation at 2000g, the supernatants were filtered and HPLC analysis was performed using a Partisphere SAX radial compression column (Wattman, Clifton, NJ). A linear gradient of 5 mM (NH₄)H₂PO₄, pH 5.0, to 300 mM (NH₄)H₂PO₄, pH 5.0, was used to separate BVDU and its metabolites. The different fractions of the eluate were assayed for radioactivity in a toluene-based scintillation cocktail. The radiolabeled 5'-mono-, 5'-di-, and 5'-triphosphorylated derivatives of BVDU were identified with the corresponding synthetically prepared BVDU derivatives, which coeluated on the HPLC chromatograms with the radiolabeled metabolites detected in the cell extracts.

Thymidylate Synthase (TS) Activity Measurements by Assaying the Tritium Release from [5-³H]dCyd in Intact S-2 and Sf-9 Cells in the Presence of BVDU. Activity of TS in the intact insect S-2 and Sf-9 cells was measured by estimation of tritium release from [5-³H]deoxyuridylate that had been formed in the intact cells from 20 µM [5-³H]dCyd, in the reaction catalyzed by TS. The procedure for the determination of tritium release has been described previously (Balzarini and De Clercq, 1984). Briefly, 6 × 10⁶ insect cells were collected by centrifugation at 200g for 8 min and resuspended in 1 ml fresh insect medium. Then 240 µl of this cell suspension (1.5 × 10⁶ cells) was added to 300 µl of medium (control) or an appropriate amount of inhibitor together with 30 µl of medium, containing nonradiolabeled substrate (final concentration, 20 µM dCyd). After a 30-min preincubation period at 37°C, 15 µl (15 µCi) of radiolabeled substrate [5-³H]dCyd was added, and at 0, 30, and 60 min, 100 µl of the reaction mixture was withdrawn and mixed with 250 µl of a cold suspension of carbon black (160 mg/ml) in 5% trichloroacetic acid. After centrifugation at 1100g for 10 min, 200-µl samples of the supernatant (containing tritiated water derived from the tritium release during the TS-catalyzed conversion of [5-³H]dUMP to dTMP) were analyzed for radioactivity.

Flow Cytometric Insect Cell Cycle Analysis. Exponentially growing S-2 and Sf-9 cells were exposed to BVDU at 4.0, 0.8, 0.16, and 0.032 µM. At 24, 48, and 72 h, the DNA of the cells was stained with propidium iodide using the CycleTEST PLUS DNA Reagent Kit (Becton Dickinson, Le pont de Claix, France). The DNA content of the stained insect S-2 and Sf-9 cell cultures was assessed by flow cytometry on a FACScan equipped with CellQuest software (Becton Dickinson). Cell debris and cell clumps were excluded from the analysis using conventional dot plot gating.

In Vivo Assay on Insecticidal Activity BVDU was examined for its growth-inhibitory and lethal activities against larvae of S. frugiperda in an artificial diet feeding assay. Various amounts of the drug were added to the solid and liquid ingredients of the agar-based diet to result in final drug concentrations of 100, 200, and 500 µM BVDU. Six-day-old larvae of S. frugiperda were placed individually on portions of the diet in plastic containers (diameter = 4 cm, height = 3 cm). Twenty larvae were used for each BVDU concentration and for the (drug-free) control. The weight of each larva was determined every other day until prepupation or premature death of the larvae. Mortality was registered during larval and pupal development, and the hatch of the adults was carefully observed. Additional details on the preparation of the diet and the implementation of the S. frugiperda bioassay for testing insecticidal compounds were described previously (Breuer and Schmidt, 1996).

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Affinity of BVDU for Dm-dNK and Mitochondrial TK-2 The inhibitory activity of BVDU against the recombinant purified Dm-dNK and TK-2 was evaluated in the presence of 2 µM [CH₃-³H]dThd as the natural substrate. BVDU dose dependently inhibited dThd phosphorylation by Dm-dNK and TK-2 with IC₅₀ values of 2.6 µM and 0.34 µM, respectively. When different concentrations of BVDU (2 µM and 1 µM) were evaluated in the presence of different [CH₃-³H]dThd concentrations between 5 µM and 0.4 µM, Lineweaver-Burk plots of the data revealed competitive inhibition of BVDU against [CH₃-³H]dThd with a K_i value of 0.66 µM and a K_i/K_m value of 0.44, indicating a pronounced affinity of the drug for the Dm-dNK (Fig. 2). Because the choice of Tris as the buffer system may lower the association constant for the Mg²⁺-ATP complex compared with other buffers, we also have replaced Tris by 50 mM HEPES, pH 8.0, and performed similar kinetic experiments in the presence of 1 and 5 mM both MgCl₂ and ATP. Virtually identical K_i/K_m values were obtained under these experimental conditions (K_i/K_m = 0.44 and 0.46, respectively). When the kinetics of BVDU were examined against mitochondrial TK-2, competitive inhibition with respect of dThd as the natural substrate was recorded (Fig. 2). The K_i/K_m ratio was 0.15, which is 3-fold lower than that found for Dm-dNK. Our data strongly suggest that BVDU may act as an alternative substrate for Dm-dNK (and TK-2). Indeed, when examined directly for substrate activity against Dm-dNK using [³H]BVDU as the radiolabeled substrate, a K_m value of 4.5 µM and a V_max value of 400 nmol/µg protein/h were found, compared with a K_m value of 1.3 µM and a V_max value of 62.5 nmol/µg protein/h for the natural substrate [CH₃-³H]dThd. Thus, the phosphorylating capacity (V_max/K_m) of Dm-dNK for BVDU was almost twice as high as for dThd. HPLC analysis of the reaction products revealed the formation of BVDU-MP from BVDU at approximately the same extent as the formation of dTMP from dThd (data not shown). The identity of the BVDU-MP formed was attested by verifying its retention time on the HPLC chromatograms with synthetically prepared BVDU-MP. No other phosphorylated BVDU derivatives (i.e., the 5'-diphosphate of BVDU) were detected in the HPLC chromatograms. Interestingly, Dm-dNK was subject to pronounced feedback inhibition by dTTP, BVDU-TP, and BVDU-MP, but not by dTMP. The IC₅₀ values of the phosphorylated dThd and BVDU metabolites ranged between 3.1 and 12 µM, respectively, in the presence of 2 µM [CH₃-³H]dThd (Table 1). It should be noted that the kinetic experiments for Dm-dNK were performed with the fusion dNK protein to glutathione S-transferase. However, we have shown previously that the affinity values for nucleoside substrates obtained for this recombinant enzyme preparation were comparable with those obtained from purified Dm-dNK derived from the D. melanogaster insect cells (Munch-Peterson et al., 1998a,b; Johansson et al., 1999).

View larger version (14K): [in this window] [in a new window]   Fig. 2.   Double-reciprocal (Lineweaver-Burk) plots for inhibition of Dm-dNK by BVDU at 2 (), 1 (), or 0 µM (), and for inhibition of mitochondrial TK-2 by BVDU at 0.8 (), 0.4 (), or 0 µM ().

Inhibition of the Growth of S-2 and Sf Cells and a Variety of Mammalian Cells in Cell Culture The S-2 and Sf-9 insect cells were seeded at ~80,000 cells per 1-cm² well of a 48-well microplate and grown in the presence or absence of different concentrations of BVDU. Nontreated control cells grew exponentially as a function of the incubation time and reached the stationary growth phase within 80 (S-2) and 104 (S9) h after seeding (Fig. 3). By that time, the insect cells proceeded through 2.5 to 3 complete cell cycles. At concentrations of 0.16 and 4 µM, BVDU virtually completely blocked S-2 and Sf-9 insect cell proliferation at all time points (up to 104 h), respectively. BVDU, administered at a concentration as low as 0.032 µM (for S-2 cells) and 0.16 µM (for Sf-9 cells) still inhibited cell proliferation. The IC₅₀ values, or inhibitory concentration of BVDU that inhibit insect cell proliferation by 50%, were 0.11 and 0.86 µM for S-2 and Sf-9 cells, respectively (Fig. 3). Thus, BVDU was markedly cytostatic against the insect cells, being at least 5-fold more potent an inhibitor of S-2 cell proliferation than of Sf-9 cell proliferation. In contrast, BVDU was only marginally cytostatic (IC₅₀  100 µM) to mammalian cells (including murine leukemia L1210 and breast carcinoma FM3A cells, human T-lymphocyte wild-type CEM/0, Molt/4F, and TK-deficient CEM/TKcells, human B-lymphoblast Raji and TK-deficient Raji/TK cells, human wild-type osteosarcoma and TK-deficient osteosarcoma cells, human embryonic lung cells, and human cervix carcinoma HeLa cells), attesting to the highly selective susceptibility of insect cells to the growth-inhibitory effects of BVDU. Thus, whereas the insect cell proliferation was inhibited at BVDU concentrations <1 µM, none of the human cell lines investigated were markedly inhibited in their growth by BVDU at 100 µM (Table 2).

View larger version (21K): [in this window] [in a new window]   Fig. 3.   Growth of S-2 and Sf-9 cells in the presence of 4.0, 0.8, 0.16, and 0.032 µM BVDU as a function of time.

Metabolism of [6-³H]BVDU in S-2 and Sf-9 Cell Cultures Both S-2 and Sf-9 cell cultures efficiently metabolized [6-³H]BVDU (Table 3). On exposure of the insect cell cultures to 1 µM BVDU for 24 h, the majority of BVDU-derived radiolabel was recovered in the methanol-insoluble fraction in S-2 cells but in the methanol-soluble fraction in Sf-9 cells. Whereas a total of 0.36 pmol radiolabel/10⁶ cells was recovered in the 60% methanol-soluble BVDU-MP + BVDU-DP + BVDU-TP fraction of S-2 cells, 30-fold higher amounts of BVDU metabolites (i.e., 11 pmol radiolabel/10⁶ cells) had accumulated into the methanol-soluble fraction of Sf-9 cells. In the Sf-9 cell cultures, a slightly lower amount of radiolabel was recovered in the 60% methanol-insoluble fraction (i.e., nucleic acids) than in S-2 cells (Table 3). The ratios for accumulation of mono- versus di- and triphosphates of BVDU markedly differed between both insect cell lines. Whereas 73% of the total phosphorylated BVDU metabolites represented BVDU-TP in S-2 cells, the predominant BVDU metabolite in Sf-9 cells was BVDU-MP (77%). Thus, S-2 cells showed a markedly better conversion of BVDU to BVDU-TP and subsequent incorporation into nucleic acid than did Sf-9 cells, which predominantly tend to accumulate BVDU-MP and allow less incorporation of BVDU into their nucleic acids.

Inhibition of Tritium Release from [5-³H]dCyd by dThd and BVDU in Intact S-2 and Sf-9 Cell Cultures as a Tool to Measure In Situ TS Activity. The inhibitory effect of different concentrations of dThd and BVDU on the tritium release from 20 µM [5-³H]dCyd was evaluated in intact S-2 and Sf-9 cells. Tritium is released from [5-³H]dCyd after formation of [5-³H]dCyd-derived [5-³H]dUMP and subsequent conversion of [5-³H]dUMP to unlabeled dTMP in the TS-catalyzed reaction. dThd inhibited tritium release from [5-³H]dCyd by 50% at ~5 µM and 49 µM in S-2 and Sf-9 cells, respectively. BVDU, however, effected a 50% inhibition of tritium release from [5-³H]dCyd at an IC₅₀ value of 0.6 and 1.0 µM in S-2 and Sf-9 cells, respectively (Fig. 4), thus at a 10- to 50-fold lower concentration than dThd.

View larger version (51K): [in this window] [in a new window]   Fig. 4.   Effect of BVDU on the release of tritium from [5-3H]dCyd in intact S-2 and Sf-9 cell cultures.

Effect of BVDU on Cell Cycle Progression in Drug-Exposed S-2 and Sf-9 Cell Cultures. Fig. 5 shows the cell cycle distribution of S-2 (A) and Sf-9 (B) cell cultures exposed to different BVDU concentrations for 24, 48, and 72 h. At the lowest BVDU concentration (i.e., 0.05 µM), the S-2 cells accumulated in the S phase of the cell cycle after 24 h and slowly progressed to mitosis (G2/M phase) at longer incubation times. At higher drug concentrations (i.e., 0.5 and 5 µM), the cells were unable to enter the S phase and were arrested at the G1/S boundary. In contrast, the cell cycle distribution of Sf-9 cells was not markedly affected at 0.05 µM BVDU. In the presence of 0.5 µM, a moderate accumulation of Sf-9 cells in the S phase was noted throughout the entire incubation period. Treatment with 5 µM BVDU caused pronounced synchronization of the Sf-9 cell cultures, leading to predominant accumulation of the cells at the G1/S boundary of the cell cycle at 24 h, in the S phase at 48 h, and at the G1/S phase at 72 h.

View larger version (31K): [in this window] [in a new window]   Fig. 5.   Cell cycle distribution of BVDU-treated S-2 (A) and Sf-9 (B) cell cultures. BVDU was administered to S-2 and Sf-9 cell cultures at a final concentration of 0, 0.05, 0.5, or 5 µM. After 24, 48, and 72 h of exposure, the DNA content of the drug-treated cells was analyzed by propidium iodide (PI) staining and flow cytometry. Cell debris and doublets were excluded from the DNA histograms using conventional dot plot gating.

Insecticidal Activity of BVDU BVDU was added to an artificial diet of S. frugiperda larvae at 500, 100, and 20 µM, and the weight of the larvae during growth was determined (Fig. 6A). Retardation of larval growth was most striking at the highest BVDU concentration (500 µM); after 9 days, the mean weight of the larvae was only 17% of that of control animals, and all larvae subsequently died. At 100 µM BVDU, a significant (25-30%) growth retardation of the larvae was observed, which resulted in ~20% larval death (prepupal mortality), but also in ~50% pupal mortality. Virtually all adults that survived the pupal period were seriously affected (moulting defects, deformed wings, etc.). Only ~1 of 20 larvae eventually developed to an intact moth. Although 20 µM BVDU did not result in a statistically significant growth retardation of the larvae, 5% of the larvae died in the prepupal stage. Those larvae that survived the prepupal state further developed to apparently normal adults (Fig. 6B).

View larger version (32K): [in this window] [in a new window]   Fig. 6.   A, growth curves of S. frugiperda larvae in the presence of 500, 100, and 20 µM BVDU. Data represent the mean weight of 20 larvae. B, prepupal (larval) mortality, pupal mortality, and affected adults (moths) in the presence of 500, 100, and 20 µM BVDU.

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

In D. melanogaster embryonic S-2 cells, only a single dNK (Dm-dNK) was found being able to phosphorylate all four deoxynucleosides (Munch-Petersen et al., 1998a,b). So far, this property is unique and clearly differs from the more specialized mammalian dNKs that show a more stringent substrate specificity. However, because HSV type 1, HSV-2, and varicella-zoster virus (VZV) also encode a dNK enzyme that efficiently recognizes both dThd and dCyd as a substrate and proved able to phosphorylate a variety of synthetic guanosine analogues (i.e., acyclovir, ganciclovir, penciclovir, etc.) (Balzarini et al., 1993 and references therein), antiherpetic compounds also might be recognized as substrates by Dm-dNK. From an initial search of a variety of antiherpetic compounds, BVDU emerged as the most potent inhibitor of dThd phosphorylation by Dm-dNK. BVDU is an excellent substrate for HSV-1 and VZV TK and is also recognized as an efficient substrate by mitochondrial TK-2 (Cheng et al., 1981b), but not by mammalian TK-1 (Cheng et al., 1981a; Balzarini et al., 1982) or by mammalian dCK (unpublished observations). As a matter of fact, BVDU is a highly selective antiherpetic agent because of its virtual lack of metabolism (and associated toxicity) in mammalian cells. Therefore, our findings that BVDU is a very efficient substrate for Dm-dNK may be of great interest in view of the development of novel Dm-dNK substrates as cytostatic agents with selectivity for insect cells. Indeed, BVDU was found to have a pronounced cytostatic activity against both the S-2 and Sf-9 cells. These cell lines are derived from two insect species that belong to two different families within the arthropod class of insects. The observations strongly suggest that insect dNK can be regarded as an attractive novel target for the development of potent and selective insecticides. This premise was attested further by our findings that BVDU caused a detrimental effect on the normal developmental process of S. frugiperda larvae to the adult stage.

We have demonstrated that in insect cells, BVDU is markedly converted (presumably by the multifunctional dNK) to its phosphorylated derivatives and subsequently incorporated into insect cell DNA. Moreover, the insect cells accumulated in the S phase of the cell cycle or were even prohibited to enter the S phase on BVDU exposure. It is thus far unclear which metabolic event prevented or retarded DNA synthesis caused in the insect cells. Because TS represents the sole de novo enzyme providing cells with thymine nucleotide building blocks for DNA synthesis and because BVDU-MP has been shown earlier to be a potent inhibitor of this enzyme (K_i/K_m ~0.5) (Balzarini et al., 1982), inhibition of TS by BVDU-MP may contribute to the cytostatic activity of BVDU against the insect cells.

We have shown earlier that tumor cells transfected with the HSV-1 or HSV-2 TK gene and subsequently exposed to BVDU are strongly inhibited in their proliferative capacity, primarily because of TS inhibition on intracellular conversion of BVDU to BVDU-MP by the HSV-1 TK or HSV-2 TK (Balzarini et al., 1987; Balzarini and De Clercq, 1989). We also have demonstrated previously that BVDU-MP is a potent inhibitor of partially purified TS (K_i/K_m = 0.5) (Balzarini et al., 1982). We also now found a TS-specific suppression in the intact insect cells by BVDU. The pronounced inhibition of tritium release from [5-³H]dCyd by BVDU cannot just be ascribed to an inhibitory effect of BVDU on [5-³H]dCyd conversion to [5-³H]dCMP by dNK before its additional deamination to [5-³H]dUMP and TS-catalyzed tritium release. Indeed, we found that dThd is much less inhibitory to tritium release from [5-³H]dCyd in the intact insect cells than is BVDU, whereas both dThd and BVDU have a comparable affinity for dNK. The TS inhibition we observed in the intact S-2 and Sf-9 insect cells exposed to BVDU may lower the endogenous dTTP pools, resulting in a transient or sustained decreased availability of dTTP for DNA synthesis. Lack of this and other dNTP substrates for DNA synthesis may in turn cause accumulation of the insect cells at the G1/S boundary or in the S phase of the cell cycle, as observed in our experiments. The inhibition of TS in intact BVDU-treated insect cells (as measured by the inhibition of tritium release from [5-³H]dCyd) occurred at drug concentrations that were (approximately) 1 order of magnitude higher than the inhibitory concentrations of BVDU for insect cell proliferation. This may point to yet additional mechanisms to explain the cytostatic action of BVDU. Indeed, inhibition of DNA polymerase by BVDU-TP and incorporation of BVDU into the TCA-insoluble cell fraction (presumably DNA) may also contribute to the accumulation of the drug-treated insect cells in the S phase of their cell cycle. Potent inhibition of DNA polymerase may even prevent the cells to enter the S phase. It should be noted that Sf-9 cells accumulated much more BVDU-MP (300-fold) and BVDU-TP (3-fold) than did S-2 cells, although Sf-9 cells showed lesser BVDU-derived radiolabel in their nucleic acids than did S-2 cells. This may partly explain the ~5-fold higher sensitivity of S-2 cells to the cytostatic activity of BVDU compared with Sf-9 cells. These observations may point to the importance of the conversion of BVDU to its triphosphate, and subsequent incorporation into DNA, in the eventual cytostatic activity.

Munch-Petersen and collaborators have reported that Dm-dNK purified from D. melanogaster is not endowed with thymidylate (dTMP) kinase activity (Munch-Petersen et al., 1998a,b), an observation that we could confirm with our recombinant purified Dm-dNK preparation. Therefore, it is intriguing that S-2 cells easily convert BVDU-MP into BVDU-TP, whereas in Sf-9 cells, this conversion seems to be limited (as attested by the accumulation of BVDU-MP in the methanol-soluble fraction). There may be two explanations for this phenomenon. Unlike Sf-9 cells, the S-2 cells may contain a slightly different dNK that has, like HSV-1 TK, and VZV TK, associated dTMP kinase activity, thus efficiently recognizing and converting BVDU-MP as substrate to BVDU-DP. Hereafter, BVDU-DP is converted further to BVDU-TP by nucleoside diphosphate kinase. Alternatively, Dm-dNK and Sf-dNK may have no associated dTMP kinase activity, as it is the case for HSV-2 TK and mitochondrial TK-2, but the S-2 cells, unlike the Sf-9 cells, may contain a cellular dTMP kinase that is able to efficiently convert BVDU-MP to BVDU-DP. The latter metabolite is then readily converted to BVDU-TP by the nucleoside diphosphate kinase. Cellular dTMP kinases can differ markedly to one another depending the nature of the cell line. For example, whereas the anti-HIV drug azidothymidine monophosphate (AZTMP) is easily converted to azidothymidine diphosphate (AZTDP) by dTMP kinase in murine L1210 cells (Balzarini et al., 1987), AZTMP is poorly converted to AZTDP by dTMP kinase of human cells (Furman et al., 1986) because of an unusual low V_max value of human dTMP kinase for AZTMP. Therefore, it would not be unlikely that the different processing of BVDU-MP to BVDU-TP in S-2 and Sf-9 cells may be attributable to kinetic differences at the level of dTMP kinase. We found that metabolism of [³H]dThd by S-2 and Sf-9 cells was very similar with regard to dTMP, dTDP, and dTTP formation (data not shown), suggesting that the kinetic properties of the dTMP kinase in both cell lines for dTMP are not strikingly different. Therefore, additional studies are required to elucidate the molecular basis of the different kinetic behavior of BVDU in intact S-2 and Sf-9 cells, and this may become an important issue with regard to the development of novel drugs with selective insecticidal potential.

Our observations that BVDU-MP and BVDU-TP act as efficient product or feed-back inhibitors of Dm-dNK imply that the conversion of BVDU to its 5'-mono- and 5'-triphosphates may become self-regulatory. On the other hand, the inhibition of both TS (by BVDU-MP) and Dm-dNK (by BVDU-MP and BVDU-TP) may efficiently prevent the formation of dTTP. This should cause a sustained shortage of dTTP for DNA synthesis and a concomitantly increased competitive advantage of BVDU-TP over dTTP during its incorporation into DNA.

BVDU is highly selective in its cytostatic action for the insect cells investigated. Whereas the Sf-9 and S-2 insect cells were inhibited at IC₅₀ values of 0.11 and 0.87 µM, respectively, the mammalian murine cells were inhibited at an IC₅₀ value of 23 to 29 µM, whereas human cell proliferation was inhibited at BVDU concentrations of 100 µM. Also, the TK-deficient cells that virtually lack cytosolic TK and only express mitochondrial TK-2 were poorly sensitive to the cytostatic action of BVDU. Presumably, BVDU cannot efficiently reach the mitochondrial TK-2 to become converted to its phosphorylated derivative or, alternatively, BVDU is phosphorylated in the mitochondria by TK-2, but the phosphorylated product(s) are not harmful for this cellular compartment. These data are in agreement with observations published previously that BVDU very marginally inhibits growth of normal lung fibroblasts at 450 µM (Machida et al., 1982), DNA synthesis of normal primary rabbit kidney cells or Vero cells at 300 µM (De Clercq and Descamps, 1981; Larsson and Öberg, 1981), the proliferation of bone marrow granulocyte-monocyte progenitor cells at 120 to 600 µM (Wingard et al., 1983), and various lymphocyte responses at 150 to 300 µM (Marmer et al., 1982; Wingard et al., 1983). In fact, BVDU has been used in the treatment of herpetic keratitis, herpetic gingivostomatitis, herpes labialis, herpetic encephalitis, and VZV infections (i.e., chickenpox, shingles) with immune-competent and immune-compromised patients. BVDU has been given as eye drops of 0.1 or 0.5%, as a cream of 1.3 or 5%, or orally or parentally (i.v.) at 7.5 to 15 mg/kg/day for 5 subsequent days. At the latter drug doses, no remarkable side effects were observed (De Clercq et al., 1980; Wildiers and De Clercq, 1984; Maudgal et al., 1985a,b). There also are no experimental indications of mutagenicity, carcinogenicity, or teratogenicity at drug doses that are antivirally active in humans (Marquardt et al., 1985; De Clercq, 1986; Jelinek and De Clercq, 1987). Although we found BVDU highly cytostatic against the D. melanogaster and Spodoptera frugiperda cells, there is at least one report in the literature (Mazzacano and Fallon, 1995) that shows that BVDU is not markedly cytostatic to mosquito cells, but becomes more inhibitory when the mosquito cells were transfected with the TK gene of HSV-1. All observations together point to a selective cytostatic activity of BVDU against several, but not all, types of insect cells.

In conclusion, we have shown that insect cells belonging to different families within the arthropod class of insects are highly sensitive to the cytostatic action of (antiherpes) drugs that are recognized as substrates for phosphorylation by the multifunctional insect dNK (IC₅₀  0.1 µM). Because of the similarities between Dm-dNK and herpetic TKs, it would be worth evaluating additional HSV TK-dependent antiherpetic drugs for their (substrate) affinity for Dm-dNK and their potential cytostatic/insecticidal properties. The insect multifunctional dNK may be regarded as a suitable target in the development of an entirely novel type of selective insecticides.