The g5R (D250) Gene of African Swine Fever Virus Encodes a Nudix Hydrolase That Preferentially Degrades Diphosphoinositol Polyphosphates

The African swine fever virus (ASFV) g5R gene encodes a protein containing a Nudix hydrolase motif which in terms of sequence appears most closely related to the mammalian diadenosine tetraphosphate (Ap 4 A) hydrolases. However, purified recombinant g5R protein (g5Rp) showed a much wider range of nucleotide substrate specificity compared to eukaryotic Ap 4 A hydrolases, having highest activity with GTP, followed by adenosine 5′-pentaphosphate (p 5 A) and dGTP. Diadenosine and diguanosine nucleotides were substrates, but the enzyme showed no activity with cap analogues such as 7mGp 3 A. In common with eukaryotic diadenosine hexaphosphate (Ap 6 A) hydrolases, which prefer higher-order polyphosphates as substrates, g5Rp also hydrolyzes the diphosphoinositol polyphosphates PP-InsP 5 and [PP] 2 -InsP 4 . A comparison of the kinetics of substrate utilization showed that the k cat /K m ratio for PP-InsP 5 is 60-fold higher than that for GTP, which allows classification of g5R as a novel diphosphoinositol polyphosphate phosphohydrolase (DIPP). Unlike mammalian DIPP, g5Rp appeared to preferentially remove the 5-β-phosphate from both PP-InsP 5 and [PP] 2 -InsP 4 . ASFV infection led to a reduction in the levels of PP-InsP 5 , ATP and GTP by ca. 50% at late times postinfection. The measured intracellular concentrations of these compounds were comparable to the respective K m values of g5Rp, suggesting that one or all of these may be substrates for g5Rp during ASFV infection. Transfection of ASFV-infected Vero cells with a plasmid encoding epitope-tagged g5Rp suggested localization of this protein in the rough endoplasmic reticulum. These results suggest a possible role for g5Rp in regulating a stage of viral morphogenesis involving diphosphoinositol polyphosphate-mediated membrane trafficking.

Genes potentially encoding Nudix hydrolases are also present in the genomes of some double-stranded animal DNA viruses. Vaccinia virus has two Nudix genes, D9 and D10, which share 21% sequence identity and 33% sequence similarity, and several other members of the Poxviridae have pairs of related genes. So far, a single Nudix gene has been found in the genomes of the iridoviruses Chilo Iridescent Virus and Lymphocystis Disease Virus and in the sole representative of the Asfarviridae, African Swine Fever Virus (ASFV). Recently, the essential vaccinia virus D10 gene was shown to inhibit cap-dependent translation, but not cap-independent translation, when overexpressed in mammalian cells, and a role in the metabolism of cap-related structures, which have similarities to dinucleoside polyphosphates, was suggested ( 24 ). In this study, we describe for the first time the cloning and characterization of a 30-kDa viral Nudix hydrolase, the ASFV g5R (also known as D250) protein (g5Rp) and show that it hydrolyzes a range of nucleoside and dinucleoside polyphosphates but has only low activity toward cap analogue nucleotides containing 7-methylguanine. However, like the DIPP subfamily, g5Rp exhibits the greatest hydrolytic efficiency with diphosphoinositol polyphosphates.

The Nudix hydrolases are members of a phylogenetically widespread enzyme family that hydrolyze predominantly the diphosphate (pyrophosphate) linkage in a variety of nucleoside triphosphates, dinucleoside polyphosphates, nucleotide sugars and nucleotide cofactors having the general structure of a nucleoside diphosphate linked to another moiety, X ( 1 , 15 ). They all possess the Nudix sequence signature motif Gx 5 Ex 5 [UA]xREx 2 EExGU (where U is an aliphatic hydrophobic amino acid), formerly known as the MutT motif ( 1 , 14 ). It is likely that these enzymes play the dual roles of removing potentially toxic nucleotides and nucleotide metabolites (e.g., oxidized nucleotides, ADP-ribose) from the cellular pools and of adjusting the levels of nucleotide cofactors and messengers in response to metabolic pressures and requirements ( 1 ). Gene sequences have been determined for a large number of Nudix hydrolases from archaea, eubacteria, animals, plants, and fungi, and many of the gene products have also been isolated and studied ( 4 , 10 , 16 , 27 ). Among the most intriguing of these are the Ap 6 A hydrolases (also known as the DIPP subfamily), which hydrolyze highly phosphorylated dinucleoside polyphosphates such as diadenosine 5′,5′″-P 1 ,P 5 -pentaphosphate (Ap 5 A) and diadenosine 5′,5′″-P 1 ,P 6 -hexaphosphate (Ap 6 A) but which appear to be most active with the structurally unrelated nonnucleotide substrates, diphosphoinositol pentakisphosphate (PP-InsP 5 ) and bisdiphosphoinositol tetrakisphosphate ([PP] 2 -InsP 4 ) ( 2 , 5 , 13 , 19 ).

The g5R gene and flanking upstream sequence containing the ASFV promoter region were amplified by PCR from clone LMw16 ( 9 ) containing Malawi LIL 20/1 isolate DNA, with primers incorporating upstream KpnI (GCGGGTACCTGAATATCTGTGAACACGGC) and downstream BamHI (GCGGGATCCATGCTTATATCGTAAATAGTTTTTAATAAA) sites. These primers amplify the coding region of g5R without the C-terminal stop codon. This fragment was cloned upstream and fused in frame with a sequence encoding the HA epitope tag in the pcDNA3 vector (Invitrogen) to give plasmid UPg5RHA. This plasmid was transfected into Vero cells which were infected with the ASFV BA71V isolate. At various times postinfection (7, 12, and 16 h) cells were fixed in 4% paraformaldehyde, permeabilized in 0.2% Triton X-100, and stained with rat anti-HA monoclonal antibody (1 in 600 dilution; Roche) and Alexa Fluor 568 goat anti-rat immunoglobulin G conjugate (1 in 800 dilution; Molecular Probes). Cells were visualized with a Leica TCS NT confocal microscope.

Vero cells were incubated in inositol-free Dulbecco modified Eagle medium containing 10% dialyzed fetal calf serum for 4 days and then for a further 3 days in 24-well plates in the same medium containing 200 μCi of [ 3 H]inositol/ml. Cells were infected with the ASFV BA71V isolate at a multiplicity of infection of 3 to 5 in the presence of [ 3 H]inositol for either 4 h or 20 h. Cells were harvested in 50 μl of 0.6 M perchloric acid and then 35 μl of 1 M K 2 CO 3 -5 mM EDTA was added. Precipitates were removed by centrifugation and supernatants diluted in 1 mM EDTA and analyzed by HPLC as previously described ( 20 ). GTP was determined by in-line monitoring of absorbance (A 260 ) of extracts during the HPLC analysis; standards were used to determine the elution time and resolution from other nucleotides. ATP was determined in samples of unfractionated extracts by using a luciferase-based ATP assay kit (Calbiochem-Novabiochem).

Unless otherwise stated, nucleotide hydrolysis by g5Rp was assayed and reaction products identified by high-performance ion-exchange chromatography as follows. Assay mixtures (100 μl) containing 50 mM 1,3-bis[Tris(hydroxymethyl)-methylamino]-propane (pH 9.5), 10 mM MgCl 2 , substrate (various concentrations), and 0.84 μg of g5Rp were incubated for 10 min at 37°C, and 90 μl was applied to a 1-ml Resource-Q column (Pharmacia) at 2 ml/min in 35 mM NH 4 HCO 3 (pH 9.6). The elution system was as follows: buffer A (H 2 O), buffer B (0.7 M NH 4 HCO 3 , pH 9.6), and gradient of 5 to 100% buffer B over 10 min. Peaks were identified with the aid of standards and quantified by area integration. Diphosphoinositol polyphosphate hydrolysis by g5Rp was assayed as described previously for DIPP ( 19 ) in assay mixtures (500 μl) containing 50 mM HEPES (pH 7.2), 50 mM KCl, 4 mM CHAPS {3-[(3-cholamidopropyl)-dimethylammonio]-1-propanesulfonate}, 0.05 mg of bovine serum albumin/ml, 2 mM MgSO 4 , and 1 mM Na 2 EDTA at 37°C. At appropriate times, aliquots were quenched, neutralized, and subsequently analyzed by high-pressure liquid chromatography (HPLC) as previously described ( 20 ).

E. coli strain BL21(DE3) was transformed with pET-g5R, and expression was induced for 3 h with 0.5 mM isopropyl-1-thio-β-d-galactopyranoside as previously described ( 3 ). The induced cells (2.5 g) were harvested, washed, and resuspended in 50 ml of breakage buffer (50 mM Tris-HCl, pH 8.0, 0.1 M NaCl). The cell suspension was sonicated, and the soluble cell lysate was recovered by centrifugation at 20,000 × g for 25 min. The supernatant was applied at 1 ml/min to a 15-by-50-mm column of Ni 2+ -nitrilotriacetic acid (NTA)-agarose (Sigma) equilibrated with 37.5 mM Tris-HCl (pH 8.0)-225 mM NaCl. After elution of the unbound protein, a 30-min gradient of 0 to 50 mM histidine in 50 mM Tris-HCl (pH 8.0)-300 mM NaCl was applied at 1 ml/min. Fractions (1 ml) were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), and those containing pure g5R protein (g5Rp) were pooled and concentrated by ultrafiltration (Amicon).

The g5R ORF was amplified from the plasmid pST8 by using the PCR and an upstream T7 20-mer primer (TAATACGACTCACTATAGGG) and a downstream 37-mer BamHI primer (CACCCAGGATCCCACCTAATGCTTATATCGTAAATAG). After amplification with Pfu DNA polymerase, the DNA was recovered by phenol-chloroform extraction and digested with NdeI and BamHI, and the gel-purified restriction fragment was ligated into the NdeI and BamHI sites of the pET15b expression vector (Novagen) to give the construct pET-g5R, in which the g5R ORF is located downstream of a His tag sequence. The plasmid was used to transform Escherichia coli XL1-Blue cells for propagation.

The pST8 plasmid containing the cloned ASFV g5R open reading frame (ORF) was kindly donated by S. R. F. Twigg ( 26 ). Pfu DNA polymerase was from Stratagene. All mononucleotides and diadenosine and diguanosine polyphosphates were from Sigma, except p 5 A, which was synthesized by using recombinant LysU lysyl-tRNA synthetase and tetrapolyphosphate ( 5 ). Ap 2 G, Ap 3 G, Ap 4 G, Ap 5 G, and Ap 6 G were kindly provided by H. Schlüter (Freie Universität Berlin). 7-Methylated cap analogues were synthesized as previously described ( 8 , 25 ). [ 3 H]inositol was from Amersham Pharmacia Biotech. PP-[ 3 H]InsP 5 and [PP] 2 -[ 3 H]InsP 4 were prepared from incomplete phosphorylation of [ 3 H]inositol hexakisphosphate (InsP 6 ; NEN) and purified as described previously ( 19 ). Unlabeled PP-InsP 5 was a kind gift from J. R. Falck (University of Texas Southwestern Medical Center, Dallas).

RESULTS AND DISCUSSION

Sequence alignment.

An initial BLAST search with the ASFV g5Rp sequence against the GenBank nonredundant database showed it to be most closely related to the mammalian Ap4A hydrolases, enzymes that cleave Ap4A (and Gp4G) asymmetrically to AMP (GMP) and ATP (GTP) but which also have lower activity toward other dinucleoside and nucleoside polyphosphates with four or more phosphate groups. In a gapped alignment, human Ap4A hydrolase shares 24% overall sequence identity and 33% similarity with g5Rp over the aligned residues. These values increase to 36 and 48%, respectively, over a 56-amino-acid residue section that spans the Nudix motif (Fig. ). Of particular note is the presence in g5Rp of the conserved tyrosine residue that is found 17 amino acids to the C-terminal side of the Nudix motif in all animal, plant, and prokaryotic asymmetrical Ap4A hydrolases (10). Thus, it was anticipated that g5Rp might hydrolyze dinucleoside polyphosphates.

Open in a separate window

Purification and properties of g5Rp.

The ASFV g5R ORF was cloned into the pET15b expression vector and expressed in E. coli. After incubation with IPTG (isopropyl-β-d-thiogalactopyranoside), a major protein of the expected size (30 kDa) was induced (Fig. ). The recombinant histidine-tagged protein was purified to near homogeneity by chromatography on Ni2+-NTA-agarose (Fig. ). When assayed with a wide range of known Nudix hydrolase substrates, it showed no activity with nucleotide sugars, NAD+, NADH, FAD, or coenzyme A but, as predicted, it hydrolyzed nucleoside and dinucleoside polyphosphates (Fig. ). However, unlike the typical eukaryotic Ap4A hydrolases, g5Rp showed a much wider range of substrate utilization, particularly with respect to nucleoside 5′-triphosphates. Thus, when assayed with a fixed concentration of substrate (200 μM), g5Rp showed the highest activity with GTP, followed by adenosine 5′-tetraphosphate (p4A), adenosine 5′-pentaphosphate (p5A), and dGTP. The diguanosine nucleotides Gp3G, Gp4G, and Gp5G were also substrates, giving somewhat higher activity than the corresponding diadenosine nucleotides Ap3A, Ap4A, and Ap5A. Heterodinucleotides such as Ap4G showed intermediate activity, although Ap2G was inactive as a substrate. Introduction of a 7-methyl group on one or both of the guanosine residues substantially reduced the activity toward the guanine-containing dinucleotides such that the classical cap analogue, 7mGp3A, was completely inactive as a substrate (Fig. ). Since the methylated cap analogues are unstable at a pH higher than 9.0 due to nucleophilic attack on the C-8 position of the imidazole ring (8, 25), assays with these compounds were carried out at pH 8.3 rather than the optimum of pH 9.5 (see below), and the results were adjusted by comparison with the activity toward Gp4G at pH 8.3.

Open in a separate window

Open in a separate window

The enzyme had a strict alkaline pH optimum of 9.5; the activity at pH 7.0 was only 5% of that at pH 9.5 (not shown). Several other Nudix hydrolases have markedly alkaline pH optima. A divalent cation was absolutely required with Mn2+ at 3 mM supporting 50% higher activity than Mg2+ at 10 to 15 mM (not shown).

Products of nucleotide hydrolysis.

Analysis of the products of hydrolysis with various substrates indicated that g5Rp can act both as a phosphohydrolase and as a pyrophosphatase. For example, ATP yielded predominantly ADP and Pi (Fig. ). The small amount of AMP observed appears to come directly from ATP rather than ADP since ADP itself was not a substrate (data not shown). GTP yielded GDP + Pi. Similarly, p4A and p5A gave ATP + Pi and p4A + Pi, respectively, as initial products, with the nucleotide product then acting as a secondary substrate (Fig. ). With Gp3G as substrate, the products were GDP and GMP (Fig. ). With Gp4G, the major product was GDP, indicating a predominantly symmetrical mode of attack, with lower amounts of GTP and GMP (Fig. ). The GDP could not have all arisen from GTP after initial asymmetrical hydrolysis since it was always in great excess over GMP. Furthermore, it could not have arisen from a contaminating, reversible guanine nucleotide-dependent guanylate kinase activity acting on GTP and GMP since incubation of GDP on its own yielded no GTP or GMP (not shown). The last observation also shows that GTP and GMP are primary products in addition to GDP, so the enzyme must be capable of both symmetrical and asymmetrical hydrolysis by attack at two different phosphate groups in the active site. The excess of GMP over GTP in the products is due to conversion of some of the GTP to GDP + Pi. Two modes of attack are also indicated for Gp5G hydrolysis by the presence of p4G and the magnitude of the GDP peak (Fig. ). This pattern of alternative hydrolytic attack is very similar to that found with the eukaryotic Ap6A hydrolases (DIPP), Nudix hydrolases that prefer the higher-order polyphosphates to Ap4A, Gp4G, etc., and for which p4A and p5A are also good substrates (5, 13, 19). Hydrolysis of Ap6A by the Saccharomyces cerevisiae Ddp1 protein in the presence of H218O revealed dual sites of attack within the polyphosphate chain (5). Since this flexibility within the active site also extends to a remarkable ability to act as highly efficient phosphohydrolases toward the diphosphoinositol polyphosphates PP-InsP5 and [PP]2-InsP4, these compounds were also tested as substrates for g5Rp.

Open in a separate window

Activity with diphosphoinositol polyphosphates.

Despite an overall low degree of sequence similarity between g5Rp and the DIPP Nudix hydrolases, g5Rp was found to metabolize PP-InsP5 and [PP]2-InsP4. The first-order rate constants for both substrates were similar: k−1 (for PP-InsP5) = 2.9 ± 0.8 mg−1 min−1 (n = 7) and k−1 (for [PP] 2-InsP4) = 3.1 ± 0.8 mg−1 min−1 (n = 10). Unlike mammalian DIPP, g5Rp was found to preferentially remove the same β-phosphate from both substrates. DIPP removes the 5-β-phosphate from PP-InsP5 but attacks predominantly the other β-phosphate (tentatively assigned as 6-β-phosphate) from [PP]2-InsP4 to yield 5-PP-InsP5, which is then further metabolized to InsP6 (23, 28). However, the intermediate product of [PP]2-InsP4 metabolism by g5Rp did not cochromatograph with an authentic 5-PP-InsP5 standard, suggesting that it is probably 6-PP-InsP5 (Fig. ). This was then further slowly metabolized to InsP6. As previously shown for mammalian and fungal members of the DIPP subfamily, 5-PP-InsP5 was quickly converted by g5Rp to InsP6 (not shown), and no further hydrolysis products of InsP6 were observed starting from either PP-InsP5 or [PP]2-InsP4. Thus, in contrast to DIPP, g5Rp appears to preferentially remove the 5-β-phosphate from both PP-InsP5 and [PP]2-InsP4. In this respect, it resembles the unrelated multiple inositol polyphosphate phosphatase (7, 23).

Open in a separate window

Kinetic parameters.

Km and kcat values were determined for several of the best nucleotide substrates and for PP-InsP5 (Table ). For the nucleotides, Km values were all within the range of 0.5 to 4 mM, while the kcat values varied from 0.35 to 0.86 s−1. Comparison of the kcat/Km ratios confirmed that GTP was the most efficient nucleotide substrate, being ca. 10-fold better than Ap4A. However, although the kcat value for PP-InsP5 is 5- to 10-fold lower than for the nucleotides (0.076 s−1), the substantially lower Km of 1.2 μM for PP-InsP5 results in a 60-fold-higher kcat/Km ratio for this substrate compared to GTP and so allows classification of g5Rp as a novel diphosphoinositol polyphosphate phosphohydrolase on the basis that PP-InsP5 gives the highest kcat/Km ratio. We were unable to generate more detailed kinetic data for [PP]2-InsP4 hydrolysis due to insufficient amounts of this material. This classification is also consistent with estimated levels of PP-InsP5 in vivo. Given that InsP6 has a typical concentration range of 15 to 100 μM (22) and that PP-InsP5 is ca. 3.5% of the level of InsP6 (Table ), then PP-InsP5 will be ca. 0.5 to 3.4 μM. The Km for PP-InsP5 is 1.2 μM. The high Km values measured for p4A, p5A, Ap4A, and Gp4G suggest that these are not significant substrates in vivo since these compounds are generally found only in the submicromolar to low micromolar range and the known specific p4A and Ap4A hydrolases have Km values for their substrates within this range (11, 12). On the other hand, the Km values for ATP and GTP are within the normal concentration ranges of these nucleotides, so it is possible that they, too, may serve as substrates for g5Rp in vivo.

TABLE 1.

SubstrateKmb (mM)kcatb (s−1)kcat/Km (103 M−1 s−1)ATP0.920.370.40p4A1.360.860.63p5A0.580.450.78Ap4A3.900.350.09GTP0.670.671.00Gp4G1.700.560.34PP-InsP50.0012 ± 0.0030.076 ± 0.00963.3 ± 9.2Open in a separate window

TABLE 2.

MetaboliteMean % ± SEM for:
4 h uninfected4 h + ASFV20 h uninfected20 h + ASFVInsP48 ± 17 ± 18 ± 16 ± 1InsP551 ± 149 ± 153 ± 154 ± 1InsP648 ± 149 ± 146 ± 145 ± 1PP-InsP51.6 ± 0.11.7 ± 0.11.6 ± 0.10.9 ± 0.1[PP]2-InsP40.06 ± 0.020.07 ± 0.010.04 ± 0.020.04 ± 0.01ATPb1008310048GTPb10010010045Open in a separate window

Substrate metabolism in ASFV-infected cells.

To determine which of the major g5Rp substrates might be affected significantly in vivo during viral infection, Vero cells were labeled with [3H]inositol for 3 days and then infected with ASFV. The intracellular levels of inositol polyphosphate metabolites and of ATP and GTP were then measured 4 and 20 h postinfection. ASFV infection led to reductions in the levels of ATP, GTP, and PP-InsP5 concomitant with the expression of g5Rp, reaching 50% in each case by 20 h postinfection (Table ). Thus, all three compounds measured may be substrates in vivo. The reduction in PP-InsP5 is unlikely to be simply a consequence of depletion of the ATP pool since previous experiments with DDT1 MF-2 Syrian hamster vas deferens cells have shown that a 30-min treatment with NaF, which causes a 70% decrease in cellular ATP, doubles the level of PP-InsP5, indicating that the concentration of PP-InsP5 does not merely mirror that of ATP per se (20). This supports a specific role for g5Rp in reducing the intracellular level of PP-InsP5.

Expression and subcellular localization of HA epitope-tagged g5Rp.

A fragment of the ASFV genome, containing the g5R gene and a 250-bp upstream promoter region, was amplified by PCR and cloned upstream and in frame with a sequence encoding the HA epitope tag. This plasmid was transfected into ASFV-infected Vero cells and at various times postinfection the cells were fixed, permeabilized, stained with rat anti-HA monoclonal antibody and goat anti-rat Alexa Fluor 568 conjugate, and then visualized by confocal microscopy. Expression of the HA epitope-tagged g5R protein was observed from ca. 7 h postinfection and showed a localization typical of rough endoplasmic reticulum (Fig. ).

Open in a separate window

Conclusions.

The biochemical activity of a virally encoded Nudix hydrolase, the ASFV g5R protein, has been determined for the first time. It is able to hydrolyze a range of guanine and adenine nucleoside and dinucleoside polyphosphates but exhibits highest substrate efficiency with PP-InsP5. Taken together, the facts that (i) the in vitro activity of g5Rp is greatest toward PP-InsP5, (ii) the Km for PP-InsP5 is close to the known concentration of this compound in vivo, and (iii) the level of PP-InsP5 is reduced after viral infection, all suggest that this may be a physiologically relevant substrate in vivo, although confirmation of this will require measurement of PP-InsP5 levels in cells infected with ASFV from which the g5R gene has been deleted or placed under the control of an inducible promoter. The functional significance of this reduction in PP-InsP5 remains unclear, although we can predict that PP-InsP5-dependent pathways in infected cells may be inhibited. In yeast cells which lack InsP6 kinase and are therefore deficient in PP-InsP5 synthesis, the vacuoles are abnormally small and fragmented, indicating that InsP6 and PP-InsP5 may interact with proteins involved in membrane dynamics (21). ASFV replicates in cytoplasmic factory areas which are rich in membranes. During morphogenesis the virus particle is formed by a process that involves wrapping of the nucleoprotein core in a double membrane layer derived from the endoplasmic reticulum (6, 18). Formation of virus factories and virus morphogenesis requires manipulation of components of the cellular secretory pathway by the virus, and possibly manipulation of the metabolism of PP-InsP5 is important for this process. Transient expression of an HA epitope-tagged g5R protein under control of its own promoter in ASFV-infected cells showed an apparent localization in the rough endoplasmic reticulum. The g5R protein contains neither an obvious signal peptide sequence nor an endoplasmic reticulum retention signal. It is therefore very unlikely that the protein is translocated into the endoplasmic reticulum lumen. Instead, it may be associated with endoplasmic reticulum-resident proteins on the cytosolic side. The reason for this localization of g5R is not clear, but it suggests a localized effect of the enzyme on the metabolism of its substrates.

These proposals for the function of the ASFV g5R protein contrast with what has been suggested for the vaccinia D10 Nudix hydrolase. Overexpression of the D10 gene severely inhibited viral protein synthesis, decreased the steady-state level of late viral mRNA, and blocked the formation of infectious virus (24). However, cap-independent reporter gene expression was not affected. Overexpression of D9 also affected gene expression but to a lesser extent. Based on the knowledge that some Nudix hydrolases degrade cap-like dinucleoside polyphosphates, it was suggested that the D10 protein might bind to or hydrolyze cap structures and so affect the stability or translatability of mRNAs (24). Although we have not measured these aspects of gene expression in our system, the fact that 7-methylation of guanine dinucleotides greatly reduces their ability to act as substrates for g5Rp implies that g5Rp is unlikely to be involved in the turnover of mRNA caps.

Examination of the sequences of the virally encoded Nudix hydrolases suggests that different family members may indeed have different activities. Only ASFV g5Rp and the closely related iridovirus proteins have the tyrosine residue downstream of the Nudix motif in the same position that is conserved in all Ap4A hydrolases (Fig. ) (10). This residue is known to stack with one of the two adenine rings in Ap4A (S. Bailey, J. B. Rafferty, and A. G. McLennan, unpublished data). Although a few individual members of the poxvirus D9 and D10 subfamilies have tyrosine or phenylalanine at this position, this is not consistent, as would be expected if all D9 and D10 members perform the same function for each poxvirus. Unless a different aromatic residue in the D9 and D10 proteins functionally replaces this conserved tyrosine (e.g., the residue 7 positions from the right hand end of the sequences depicted in Fig. ), then it is likely that the g5R, D9 and D10 proteins do have enzymatically distinct activities and, therefore, functions. Future comparative analysis of cells infected with poxviruses and ASFV either overexpressing or deleted for these genes, combined with an in vitro analysis of the D9 and D10 proteins, should clarify the situation and lead to a complete understanding of the role of Nudix hydrolases in the infectious cycles of these large DNA viruses.