INTRODUCTION

The prototypic orthopoxvirus, vaccinia virus, has been shown to encode the enzyme deoxyuridine 5-triphosphate nucleotidohydrolase (dUTPase, EC) (Broyles, 1993). This is perhaps not surprising considering that poxviruses are large DNA viruses that replicate solely in the cytoplasm of the host cell and have been shown to encode many of the enzymatic functions necessary for DNA precursor metabolism (reviewed in Moss, 1996; and Traktman, 1990). Also, poxviruses, unlike oncogenic viruses, do not induce cell proliferation and thus do not induce transcription of the host DNA replication machinery, much of which is regulated coordinately with the cell cycle. Due to its cytoplasmic site of replication and independence from host nuclear functions, vaccinia virus has been extensively studied as a model system and as a source of biological activities necessary for DNA metabolism and replication (Moss, 1996). In addition to dUTPase, other enzymes required for DNA precursor metabolism that have been shown to be virally encoded include thymidine kinase (Hruby and Ball, 1982), ribonucleotide reductase (Slabaugh et al., 1984), and thymidylate kinase (Hughes et al., 1991).

dUTPase, in removing dUTP from the dNTP pool and generating dUMP, is ostensibly involved in both maintaining the fidelity of DNA replication and in providing the precursor for the production of TMP by thymidylate synthetase (Kornberg and Baker, 1991). High levels of dUTP have been shown to be lethal to cells. DNA polymerase will readily utilize dUTP, resulting in the generation of highly uracil-substituted DNA upon replication (Warner and Duncan, 1978). Uracil-substituted DNA is acted upon by a cellular excision repair process initiated by uracil DNA glycosylase (Duncan, 1981), which can result in double-strand breaks and ultimately cell death (Tye et al., 1977). Consistent with vaccinia's independence from many host metabolic functions, a virally encoded uracil DNA glycosylase has been identified and shown to be essential (Stuart et al., 1993). Also, it has been suggested that it is the high levels of dUTP that mediate the process of thymineless death that results from treatment of mammalian cells with inhibitors of thymidylate synthetase such as N¹⁰-propargyl-5,8-dideazafolic acid (CB3717) or FdUrd. These inhibitors, in concert with low intracellular dUTPase levels, eventually result in high levels of dUMP incorporation and ultimately DNA fragmentation caused by an iterative excision repair process mediated by uracil DNA glycosylase (Tye et al., 1977; Curtin et al., 1991; Canman et al., 1993).

dUTPase has been shown to be essential in both Escherichia coli (El-Hajj et al., 1988) and yeast (Gadsden et al., 1993). dUTPase activity has been demonstrated in a number of other eukaryotic systems. These include Drosophila (Nation et al., 1989), higher plants (Pardo and Gutierrez, 1990; Pri-Hadash et al., 1992), human cells (Williams and Cheng, 1979), and within some members of the herpesvirus (Wohlrab and Francke, 1980) and retrovirus families (Elder et al., 1992; Threadgill et al., 1993). The native enzymes from all of these eukaryotic sources, while differing little in enzymatic characteristics, vary as to native structure and protomer molecular weight. It is of interest that the enzyme has been shown to be essential in a eukaryotic system in that, unlike E. coli, mammalian cells do not encode a dCTP deaminase and also are able to generate dUMP from dCMP via dCMP deaminase. Thus, the essentiality of the enzyme in precursor metabolism and for fidelity was less certain and could not be assumed from the prokaryotic model.

Reports from various systems have stimulated increased interest in the metabolic role of the enzyme. One is the determination that the enzyme is encoded by the non-primate lentiviruses and by types B and D retroviruses (Elder et al., 1992; Koppe et al., 1994). The role of dUTPase in the retroviral lifestyle is unclear, but reduced viral replication in nondividing or terminally differentiated cells does occur in equine infectious anemia virus (Threadgill et al., 1993; Steagall et al., 1995; Lichtenstein et al., 1995) and feline immunodeficiency virus (Wagaman et al., 1993; Lerner et al., 1995) upon dUTPase inactivation. Another is the demonstration in Drosophila (Giroir et al., 1987; Nation et al., 1989) and higher plants (Pardo and Gutierrez, 1990) that dUTPase is developmentally regulated.

We initially reported on the open reading frame (ORF)¹ F2L in vaccinia, identifying it as a sequence with significant homology to a retroviral protease-like gene (Slabaugh and Roseman, 1990). Subsequent additions of the sequence of dUTPase genes from E. coli and herpesvirus to the sequence data base demonstrated significant homology to the retroviral and vaccinia protease-like gene sequences (McGeoch, 1990). The protease-like sequences in these retroviruses have been shown to encode functional dUTPases (Elder et al., 1992). The F2L gene of vaccinia has also been shown to encode a functional dUTPase via biochemical analysis of the enzyme produced as a glutathione S-transferase fusion protein in E. coli (Broyles, 1993). In order to facilitate a detailed structural analysis of the vaccinia dUTPase we have cloned, expressed, and purified the enzyme to homogeneity. The enzyme is expressed at high levels and has a specific activity 36-fold higher than that of the fusion protein. Here we report the biochemical characterization of the dUTPase of vaccinia virus, show the native structure of the viral enzyme to be a trimer, and demonstrate that vaccinia dUTPase, while being catalytically specific for dUTP, is able to bind other nucleoside triphosphates at the active site.

MATERIALS AND METHODS

A 1.4-kilobase AccI DNA fragment containing the entire ORF F2L was cloned into the phagemid vector pIBI30 (International Biotechnologies, Inc.). An NdeI site (underlined below) was engineered at the initiating ATG start codon by site-specific mutagenesis (Kunkel et al., 1987) using the oligonucleotide 5-GTGAGTTAATATTGAACATAAAACTAA-3, thus generating the clone pF2/Nde. Following an NdeI-HindIII digest of the clone pF2/Nde, the fragment containing the F2 gene was cloned into the expression vector pT7-7 (Tabor and Richardson, 1985) that had been digested with NdeI-HindIII. The resulting clone, pT7-F2, was transformed into the bacterial strain BL21(DE3), a lysogen carrying the gene for T7 RNA polymerase under the control of the inducible p_L promoter (Studier et al., 1990). For expression, bacteria were propagated in Luria broth containing 100 µg/ml ampicillin. Upon reaching an A₅₉₀ of 0.8, cultures were induced with 0.4 m isopropyl --thiogalactopyranoside (IPTG) and incubated for 5 h at 30 °C.

All steps were performed on ice or at 4 °C. Typically, a 1-liter bacterial culture was harvested by centrifugation, and the cell pellet was resuspended in lysis buffer containing 50 m Tris-HCl (pH 7.5), 1 m EDTA, 2 m -mercaptoethanol, 20 m spermidine-HCl, 10% sucrose, and 0.2 m phenylmethylsulfonyl fluoride. Lysozyme was added to a final concentration of 0.2 mg/ml followed by incubation on ice for 30 min. Triton X-100 was then added to a final concentration of 0.1%, and the extracts were centrifuged for 20 min at 20,000 × g. Nucleic acids were removed by dropwise addition of 2% protamine sulfate (pH 7.0) to 0.2% followed by centrifugation for 20 min at 20,000 × g. The remaining supernatant was labeled the crude extract. Solid ammonium sulfate was then added to the crude extract to 60% saturation while stirring on ice. Following centrifugation at 20,000 × g for 20 min, the pellet was dissolved and then dialyzed in Buffer A (10 m MOPS (pH 7.5), 4 m MgCl₂, 2 m -mercaptoethanol, 10% glycerol, 0.2 m phenylmethylsulfonyl fluoride). The dialysate was applied to a Blue-Sepharose CL-6B column pre-equilibrated with Buffer A. After extensive washing with Buffer A, the column was eluted with a linear gradient of 0-1.0 NaCl in Buffer A. Fractions containing dUTPase were pooled, dialyzed in Buffer A, and applied to a CM-Sepharose column pre-equilibrated with Buffer A. The column was washed with 3-column volumes of Buffer A, and dUTPase was then eluted as a discrete peak with 0.1  NaCl in Buffer A. Fractions containing pure dUTPase were pooled and used for enzyme assays. The presence of dUTPase in column fractions was readily monitored on 15% polyacrylamide gels stained with Coomassie Brilliant Blue.

The relative amount of vaccinia dUTPase protein in crude bacterial extracts was determined by scanning laser densitometry of Coomassie-stained SDS-polyacrylamide gels using a Zenith model SL-504-XL instrument (Biomed Instrument Inc., Fullerton, CA).

Unless otherwise indicated, dUTPase activity was determined in a 20-µl reaction mixture containing 50 m MOPS (pH 7.5), 4 m MgCl₂, 2 m dithiothreitol, 0.2% bovine serum albumin, 50 µ [5-³H]dUTP (50 µCi/µmol), and enzyme. Initial velocity studies were done with dUTP in the range of 0.5 to 5 µ and 0.5 ng of enzyme. For competition studies 1 ng of enzyme and 100 µ 8-azido-ATP were preincubated on ice for 15 min prior to addition to reaction mixture containing between 0.5 and 10 µ dUTP. All assay mixtures were incubated at 25 °C for 5 min. Reactions were terminated by spotting the reaction mixture on a PEI-cellulose plate. The plate was developed in one dimension using a 0.5  LiCl, 0.1  (NH₄)₂SO₄ solvent system. Radioactivity was determined by one of two methods. For liquid scintillation, unlabeled nucleotide markers were visualized by UV light, appropriate regions were excised, eluted with 0.5  HCl, and counted. For analysis by an Automatic TLC Linear Analyzer (Berthold), plates were developed as above, dried, and scanned, and the results were analyzed using the computer program Chroma.

Protein concentration was determined by Bio-Rad protein assay using bovine serum albumin as the standard. One unit of enzyme activity is defined as the amount of pure dUTPase required to convert 1 µmol of dUTP to dUMP + PP_i/min at 25 °C.

Each irradiation mixture (50 µl, 4 °C) contained 25 m Bis-Tris at pH 7.5, 25 m NaCl, 2.4 µg of dUTPase, 2 or 5 m competing nucleotide, and either µ [-³²P]8-azido-ATP or 50 µ [-³²P]5-azido-dUTP. Each reaction was photolyzed for 1 min (302 nm lamp, 5 mW/cm²) on a parafilm surface, followed by the addition of 5 µl of 100 m ATP to reduce the background radioactivity on the glass fiber filter. To determine the amount of probe cross-linked to dUTPase, 32 µl of each reaction was spotted onto a 24-mm G4 glass fiber filter (Fisher) and immediately dipped into ice-cold 5% (w/v) trichloroacetic acid containing 1% (w/w) sodium pyrophosphate (trichloroacetic acid/PP_i). Each filter was then washed with 30 ml of the trichloroacetic acid/PP_i solution, batch washed with ethanol, dried, and counted by liquid scintillation counting. Control reactions were performed for each reaction to account for background radioactivity. The control reactions contained the exact amounts of probe and other reagents as the cross-linking reaction but was photolyzed before the addition of dUTPase.

Four aliquots of pure dUTPase having a UV absorbance of 3.78 (at 280 nm) were analyzed for amino acid composition at the University of Kentucky Macromolecular Structure and Analysis Facility on a Beckman 6300 High Performance Amino Acid Analyzer utilizing an o-phthalaldehyde detection method. The average of four determinations indicated that the dUTPase concentration in the original sample was 4.54 mg/ml. After correcting for the eight prolines and two cysteines per mol of dUTPase not detected by this method, the actual concentration was 4.84 mg/ml. This result indicates that E_mg/ml = 0.78.

Chromatography matrices were purchased from Pharmacia Biotech Inc. DNA restriction and modifying enzymes were obtained from Life Technologies, Inc. and U. S. Biochemical Corp. Deoxy[5-³H]uridine 5-triphosphate was purchased from Amersham Corp. (17 Ci/mmol). Deoxy[8-³H]adenosine 5-triphosphate (20.9 Ci/mmol) was purchased from DuPont NEN. Deoxy[5-³H]cytidine (18 Ci/mmol), thymidine (60 Ci/mmol), and guanosine (11 Ci/mmol) 5-triphosphates were all purchased from ICN. 5-Azido-dUTP was prepared as described previously (Evans and Haley, 1987). [-³²P]8-azido-ATP and [-³²P]5-azido-dUTP were prepared using published procedures (Evans and Haley, 1987; Glynn and Chappell, 1964). Nonradioactive nucleotides were obtained from Sigma Polygram CEL 300 PEI/UV₂₅₄ thin layer chromatography plates were obtained from VWR.

RESULTS

For overexpression of dUTPase, the vaccinia ORF F2L was cloned into the expression vector pT7-7 in order to generate the plasmid pT7-F2. Introduction of an NdeI site at the amino terminus of ORF F2L allowed for the expression of the protein with no additional sequence. Induction of BL21(DE3) cells transformed with pT7-F2 by the addition of IPTG resulted in the production of a 16.4-kDa peptide that is not present in uninduced extracts (Fig. 1, lanes 2 and 3). This is consistent with the predicted molecular weight of ORF F2L. This protein was detectable as early as 30 min after induction (data not shown) and constituted a major protein product by 2-h post-induction. After 5 h the F2 protein constituted approximately 15% of the total protein as determined by scanning laser densitometry of a Coomassie-stained SDS-polyacrylamide gel (Fig. 1, lane 3). Solubility of the highly expressed protein was significantly improved by growing the bacterial cultures at 30 °C.

To determine whether expression of dUTPase by the pT7-F2 construct produced a functional enzyme, dUTPase assays were performed on induced BL21(DE3) bacterial cell extracts of equal protein concentration transformed with either pT7-7 or pT7-F2 (Fig. 2). Enzyme assays containing 1 or 5 µg of the extracts could not give an accurate comparison of relative activity; after a 5-min incubation of the extract containing pT7-F2, the dUTP substrate was almost completely consumed with over 90% being converted to dUMP. A substantial amount of endogenous dUTPase activity can also be observed as extracts of the BL21 (DE3) host containing pT7-7 produced dUMP levels of approximately 17 and 40% using 1 or 5 µg, respectively. Relative enzyme activity was demonstrated by an assay using 0.3 µg of the protein extracts. The extract containing pT7-7 converted less than 2% of the dUTP to dUMP, whereas the pT7-F2 extract converted almost 40% of the substrate to product. This showed that the transformant containing pT7-F2 exhibited approximately 20-fold more dUTPase activity than that containing pT7-7. We concluded that the pT7-F2 construct produces a highly active dUTPase.

Recombinant vaccinia dUTPase was typically purified from 1-liter bacterial cultures as described under ``Materials and Methods.'' Fig. 1 shows the protein composition at each purification step. Five hours after induction with IPTG, bacteria were lysed enzymatically, the cell debris was removed by centrifugation, and the remaining supernatant was labeled the crude extract (lane 3). Nucleic acids were removed, and proteins were precipitated by addition of ammonium sulfate and then dialyzed (lane 4). Fractions containing dUTPase after Blue-Sepharose CL-6B column chromatography were pooled and dialyzed (lane 5). CM-Sepharose column chromatography yielded highly pure dUTPase (lane 6). Typically, this purification procedure yielded between 4 and 5 mg of recombinant dUTPase per liter of bacterial culture (Table I). Recently, we have improved the yield of pure dUTPase to consistently over 10 mg per liter by inclusion of 0.1% Nonidet P-40 to Buffer A during the dialysis steps (data not shown). The purified enzyme is highly stable as samples stored at 80 °C for over 1 year retain nearly original levels of activity.

Table I. Purification of vaccinia dUTPase Fraction ml Total protein Total units Specific activity Yield Purification mg 105 units/mg % -fold Crude extract 40.0 100 3.40 3,400 100 Ammonium sulfate 9.0 50 1.87 3,700 55 1.1 Blue Sepharose 43.5 6.2 1.53 24,800 45 7.3 CM Sepharose 16.7 4.2 1.51 36,000 44 10.6

Computer analysis of the amino acid sequence of ORF F2L predicted a molecular mass of 16.4 kDa. This was consistent with the apparent molecular weight of recombinant vaccinia dUTPase on SDS-polyacrylamide gels. Gel filtration chromatography was used to determine the molecular mass of the native enzyme. Purified vaccinia dUTPase eluted from a Superose 12 HR10/30 FPLC gel filtration column as a discrete peak that retained enzymatic activity (not shown). The retention time of 14.1 min, coincident with the retention time of the marker ovalbumin, corresponded to an approximate molecular mass of 45 kDa (Fig. 3). We conclude that the vaccinia-encoded dUTPase enzyme is a trimer.

Initial velocity studies were performed in which the concentration of dUTP was varied between 0.5 and 5 µ and dUTPase was present at 0.5 ng. In the concentration range used, dUTP demonstrated Michaelis-Menten-type substrate kinetics with an apparent K_m of 0.94 µ (Fig. 4).

The pH dependence of the enzyme was tested. Enzyme activity varied little over a pH range of between 6 and 10 (data not shown).

In order to assess the metal ion requirement of the enzyme, various chelators were added to the standard reaction mixture minus MgCl₂ (Table II). For this analysis, purified enzyme was dialyzed exhaustively against Buffer A minus MgCl₂. Enzyme activity was unchanged by this dialysis step, and the addition of MgCl₂ to the reaction did not significantly increase activity (not shown). Addition of 1.0 m EDTA, a chelator of Mg²⁺ and Ca²⁺, inhibited enzymatic activity to 5% of the control. Addition of the Ca²⁺ chelator EGTA and the Zn²⁺ chelator o-phenanthroline at a concentration of 1.0 m resulted in a reduction of enzyme activity to 66 and 63%, respectively. To assess whether enzymatic activity required a specific divalent cation, the ability of different divalent cations to restore enzymatic activity in the presence of EDTA was tested. Preincubation of the dialyzed enzyme with EDTA for 5 min, followed by addition of the indicated cations, showed that inhibition by EDTA was reversible by addition of the divalent cations Mg²⁺, Mn²⁺, or Zn²⁺ (Table III). Zn²⁺ and Mn²⁺ were equally effective at restoring enzyme activity to a level 90% of the control, whereas Mg²⁺ restored activity to 60% of the control. Therefore, although dUTPase can utilize any of these divalent cations as a prosthetic group, Mn²⁺ and Zn²⁺ were most effective.

Table II. Effect of various chelators on dUTPase activity Reaction conditions Relative activitya % 0.1 m EDTA 41 1.0 m EDTA 5 0.1 m EGTA 90 1.0 m EGTA 66 1.0 m o-phenanthroline 63 a  % relative activity is the percentage of dUTP converted to dUMP as compared with an assay where no chelator is added.

Table III. Effect of addition of various cations at 4 m after preincubation with 1.0 m EDTA Additive Relative activitya % None 5 Mg2+ 68 Mn2+ 90 Zn2+ 90 a  % relative activity is the percentage of dUTP converted to dUMP as compared with an assay where no cation or chelator is added. Standard reactions minus cation were preincubated with 1.0 m EDTA for 5 min. The indicated cation was then added to 4 m.

Substrate specificity of the enzyme was assessed by substituting the four deoxynucleoside triphosphates normally found in DNA for dUTP in the standard reaction mixture at a concentration of 0.1 m (50 µCi/µmol). The products of the reaction were analyzed by thin layer chromatography, and appropriate regions were excised for scintillation counting. Under these conditions, hydrolysis of 0.1% of the substrate is detectable. No dephosphorylation of the four substituted deoxynucleotides was observed (data not shown). This demonstrated that the sole substrate for hydrolysis by vaccinia dUTPase is dUTP. However, cross-linking protection studies done to assess the utility of using the photoaffinity analogues 5-azido-dUTP and 8-azido-ATP as cross-linking agents for active site mapping showed that cross-linking of -³²P-labeled analogues to dUTPase was significantly inhibited in the presence of nucleotide competitors (Table IV). Although the degree of protection by 2 m dUTP was slightly less than the protection afforded by the other dNTPs, this was probably due to the rapid hydrolysis of dUTP, thus lowering its effective concentration. An increase in the dUTP concentration to 5 m resulted in protection comparable with that obtained with the other dNTPs. This suggested that while dUTPase is catalytically specific for dUTP, binding of other nucleoside triphosphates at the active site may occur. To assess this more directly, a competition experiment was done. dUTPase enzyme was incubated with 100 µ 8-azido-ATP and then used in an initial velocity experiment (Fig. 5). The results demonstrate that 8-azido-ATP acts as a competitive inhibitor of dUTP with a K_i of approximately 173 µ.

Table IV. Protection from photolabeling by nucleotide competitors Percent protection refers to the percent decrease in dUTPase photolabeling in the presence of the competing nucleotide, compared with a control reaction without a nucleotide competitor. Each photolabeling reaction was initiated by the addition of dUTPase to a reaction mixture containing buffer, competing nucleotide, and photoprobe. Following a brief mixing of the reaction and a 10-s incubation, the reaction was photolyzed for 1 min. The incorporation of 32P was determined as described under ``Materials and Methods.'' Competitor % protection 5 µ 8-azido-ATP % protection 50 µ 5-azido-dUTP 2 m dUTP 81 65 5 m dUTP 91 86 2 m deoxyuridine 22 46 2 m dTTP 83 90 2 m dATP 88 93 2 m dGTP 91 92

DISCUSSION

Using a bacterial system we have overexpressed and purified the recombinant vaccinia virus dUTPase enzyme to homogeneity and shown it to be highly active. dUTPases have now been identified and characterized in a number of eukaryotic viruses including retroviruses, herpes, and poxviruses (Elder et al., 1992; Wohlrab and Francke, 1980; Broyles, 1993), and in a number of eukaryotic systems including higher plants (Pardo and Gutierrez, 1990; Pri-Hadash et al., 1992), rat (Hokari and Sakagishi, 1987), yeast (Gadsden et al., 1993), fungi (McIntosh et al., 1994), and humans (Williams and Cheng, 1979; Climie et al., 1994). Characterization of the enzymatic properties of the vaccinia dUTPase enzyme shows it to share many of the characteristics of its eukaryotic and prokaryotic homologues in that the vaccinia dUTPase is a highly specific metalloenzyme that functions within a broad pH range. We found that purified dUTPase that has been dialyzed exhaustively retains activity; however, activity is abolished by EDTA and can be restored by the addition of divalent cations. This suggests that a divalent cation is required for enzyme activity and that the prosthetic group is tightly bound by the enzyme. Using gel filtration we have determined the native structure of the vaccinia enzyme to be a trimer in composition. The enzyme shares this trimeric structure with dUTPases from rat (Hokari and Sakagishi, 1987), E. coli (Cedergren-Zeppezauer et al., 1992), and the human enzyme (McIntosh et al., 1992), with which it shares its highest amino acid identity (63%).

Broyles (1993) demonstrated that the F2L gene of vaccinia virus had dUTPase activity when expressed as a glutathione S-transferase fusion protein. Here, we confirm that finding using authentic vaccinia dUTPase purified from a bacterial expression system. However, we determined a K_m approximately 100-fold lower and a specific activity of the enzyme 36-fold higher than that determined by Broyles. We assume these differences are the direct result of characterizing a fusion protein as compared with authentic dUTPase.

Upon infection, poxviruses express a number of enzymes that function to provide nucleotide precursors for DNA replication (Traktman, 1990). It is noteworthy that a number of these enzymes have significant homology to their human homologues; thymidylate kinase and thymidine kinase have amino acid identities of 42 and 68%, respectively. We have previously shown that the vaccinia dUTPase, ORF F2L, is expressed early in infection, consistent with its role in DNA metabolism (Slabaugh and Roseman, 1990). Of the viral enzymes that are involved in nucleotide metabolism, the virally encoded ribonucleotide reductase (Slabaugh et al., 1984) is of particular relevance here as it will accept UDP as a precursor and would ultimately contribute to the intracellular dUTP pool. Therefore, it is advantageous for vaccinia virus to encode both a dUTPase to reduce the dUTP pool and a uracil DNA glycosylase that acts by excision repair to remove dUTP that has been misincorporated into replicating DNA. While it has been demonstrated that the viral DNA glycosylase is essential for viral replication (Stuart et al., 1993; Millns et al., 1994), the situation for dUTPase is less clear. Perkus et al. (1991) successfully deleted the dUTPase gene as part of a large, viable deletion that removed a total of 55 ORFs from vaccinia virus. Our attempts to disrupt the viral dUTPase gene in vivo using the selectable marker gpt have been unsuccessful.²

Whether the large deletion in vaccinia allows for a compensatory modulation of host functions similar to what is seen in the herpes simplex virus (HSV) system remains to be determined (Lirette and Caradonna, 1990). These authors showed that during HSV infection of HeLa cells, the cellular dUTPase was dephosphorylated, and its activity was reduced. However, during infection by an HSV dUTPase mutant, the cellular enzyme remained phosphorylated and active. They concluded that a productive infection by the HSV dUTPase mutant was made possible by the retention of cellular activity. In mature human T cells, it has been shown that expression and phosphorylation of dUTPase is dependent on the cell cycle (Strahler et al., 1993). We have determined that the vaccinia dUTPase is phosphorylated during infection of HeLa cells.³

Consistent with many studies of dUTPase, we have shown the enzyme to be highly specific for hydrolysis of dUTP. Using labeled dNTPs we detected no dephosphorylation of any other substrate but dUTP. However, we have performed cross-linking studies using the photoaffinity analogues 5-azido-dUTP and 8-azido-ATP to determine whether these probes could be used for active site mapping. Protection studies showed that [-³²P]5-azido-dUTP and [-³²P]8-azido-ATP can be specifically inhibited from cross-linking to the enzyme by the presence of nucleotide competitors. These results suggest that while dUTPase is catalytically active for only dUTP, other nucleotides can occupy the active site. In order to test this hypothesis we used 8-azido-ATP as a competitive inhibitor of enzyme activity. This showed that 8-azido-ATP acted as a competitive inhibitor of dUTP. This confirmed directly the ability of 8-azido-ATP to bind to the active site.

The development of agents that selectively inhibit dUTPase is of great interest as the enzyme is obviously an excellent target for the inhibition of cell proliferation. Recent work testing the ability of substrate analogues to inhibit the activity of dUTPase of E. coli, and to inhibit the proliferation of human cancer cells in culture, has shown some success (Zalud et al., 1995). Our demonstration here that 8-azido-ATP binds the active site, albeit weakly, may have implications not only for the design of enzyme inhibitors but also for the elucidation of the active site itself.