The Saccharomyces cerevisiae YOR163w Gene Encodes a Diadenosine 5′,5‴-P 1,P 6-Hexaphosphate (Ap6A) Hydrolase Member of the MutT Motif (Nudix Hydrolase) Family*

The YOR163w open reading frame on chromosome XV of the Saccharomyces cerevisiae genome encodes a member of the MutT motif (nudix hydrolase) family of enzymes ofM r 21,443. By cloning and expressing this gene in Escherichia coli and S. cerevisiae, we have shown the product to be a (di)adenosine polyphosphate hydrolase with a previously undescribed substrate specificity. Diadenosine 5′,5‴-P 1,P 6-hexaphosphate is the preferred substrate, and hydrolysis in H2 18O shows that ADP and adenosine 5′-tetraphosphate are produced by attack at Pβ and AMP and adenosine 5′-pentaphosphate are produced by attack at Pα with a K m of 56 μmand k cat of 0.4 s−1. Diadenosine 5′,5‴-P 1,P 5-pentaphosphate, adenosine 5′-pentaphosphate, and adenosine 5′-tetraphosphate are also substrates, but not diadenosine 5′,5‴-P 1,P 4-tetraphosphate or other dinucleotides, mononucleotides, nucleotide sugars, or nucleotide alcohols. The enzyme, which was shown to be expressed in log phase yeast cells by immunoblotting, displays optimal activity at pH 6.9, 50 °C, and 4–10 mm Mg2+ (or 200 μm Mn2+). It has an absolute requirement for a reducing agent, such as dithiothreitol (1 mm), and is inhibited by Ca2+ with an IC50 of 3.3 mm and F− (noncompetitively) with aK i of 80 μm. Its function may be to eliminate potentially toxic dinucleoside polyphosphates during sporulation.

Recent interest in the diadenosine polyphosphates (Ap n A) 1 has focused on their possible roles as regulators of cell proliferation. Ap 3 A has been proposed as a component of the interferon-induced antiproliferative response in mammalian cells (1), whereas Ap 4 A, the intracellular level of which has long been known to be associated with proliferation (2,3), may be an antagonist of this pathway. A key factor in this response in humans is the fragile histidine triad (FHIT) protein, an Ap 3 A hydrolase that is absent or defective in many common cancers (4,5). Precisely how this protein and its substrate, Ap 3 A, contribute to antiproliferation is not clear, but there is little doubt that the regulation of the intracellular levels of specific diadenosine polyphosphates is of great importance.
Eukaryotic Ap 3 A hydrolases exhibit an approximately 10fold preference for Ap 3 A over Ap 4 A as substrates 2 (6). They are members of the histidine triad (HIT) family of proteins and possess the catalytic sequence motif HXHXHX in which the central histidine residue forms a covalent enzyme-AMP reaction intermediate (4). Animals and higher plants also possess an asymmetrically cleaving Ap 4 A hydrolase that prefers Ap 4 A but is also active toward higher homologues, such as Ap 5 A and Ap 6 A, but inactive toward Ap 3 A (6). This enzyme belongs to the MutT motif (or nucleoside diphosphate linked to x (nudix)) family of nucleotide hydrolases (7,8). Together, these two enzymes are probably crucial for regulating the Ap 3 A/Ap 4 A ratio.
Many lower eukaryotes appear unusual in possessing one or more Ap 4 A phosphorylases in place of Ap 4 A hydrolase. For example, Saccharomyces cerevisiae has two Ap 4 A phosphorylases, Apa1 and Apa2, in addition to an Ap 3 A hydrolase (6, 9 -11). Like the Ap 4 A hydrolases, the yeast phosphorylases can also degrade Ap 5 A but not Ap 3 A (6,11). The phosphorylases appear to be distantly related to the HIT proteins, having an HXHXQ motif in place of the HXHXH histidine triad (10,12). Although S. cerevisiae does not have an Ap 4 A hydrolase, genes for five potential MutT motif proteins can be discerned in the genomic sequence. Here, we report that one of these, YOR163w (GenBank™ accession no. Z75071) from chromosome XV, encodes an Ap 6 A hydrolase that is also active against Ap 5 A and the adenosine 5Ј-polyphosphates p 5 A and p 4 A, but not Ap 4 A or Ap 3 A. This is the first time that an enzyme with this substrate specificity has been described. A preliminary report of this work has appeared (13).

Materials
Ap 6 A was synthesized by carbodiimide condensation of ATP (14). p 5 A was synthesized using the recombinant LysU lysyl-tRNA synthetase and tetrapolyphosphate (15,16). The plasmid pXLys5 was a gift from P. Plateau, and the LysU protein was purified as described (16). All other nucleotides were from Sigma. The cosmid clone pUOA1258, which carries the complete YOR163w open reading frame from yeast chromosome XV, was a gift from B. Dujon. The vector pPGY1 was a gift from L. D. Barnes
ACACACCATGGGCAAAACCGCGGATAAT) and d(AGGAATGGATC-CATATGTTTGCGGTGGCT) were synthesized to provide an NcoI restriction site at the start of the amplified gene and a BamHI site at the end. After amplification with Pfu DNA polymerase, the DNA was recovered by phenol/chloroform extraction and digested with NcoI and BamHI, and the gel-purified restriction fragment was ligated into the NcoI and BamHI sites of the pET15b vector (Novagen), thus regenerating the ATG initiator in the NcoI site and eliminating the His tag sequence from the vector. The resulting plasmid, pET163W, was used to transform E. coli XL1-Blue cells for propagation.
Cloning in Yeast-The YOR163w gene was amplified as above using the primers d(GTGGGGGAATTCAAAATGGGCAAAACCGC) and d(GA-ATAGCTCGAGATGTTTGC GGTGGCTTG). These primers provided an EcoRI restriction site at the start of the amplified gene and a XhoI site at the end. After amplification, the recovered DNA was digested with EcoRI and XhoI, and the gel-purified restriction fragment was ligated into the EcoRI and XhoI sites of the yeast centromere vector, pPGY1. The resulting construct, pPGY163W, generated the ATG initiator downstream of GAL1p, a galactose-inducible promoter. The plasmid was used to transform E. coli XL1-Blue cells for propagation.
Protein Expression in E. coli and Purification-E. coli strain BL21(DE3) was transformed with pET163W. A single colony was picked from an LB agar plate containing 20 g/ml ampicillin and inoculated into 10 ml of LB medium containing 60 g/ml ampicillin. After overnight growth, the cells were transferred to 1 liter of LB medium containing 60 g/ml ampicillin and grown to an A 600 of 0.9. Isopropyl-1-thio-␤-D-galactopyranoside was added to 0.4 mM, and the cells were induced for 4 h. The induced cells (5.1 g) were harvested, washed, and resuspended in 50 ml of breakage buffer (50 mM Tris-HCl, pH 8.0, 2 mM EDTA, 0.1 M NaCl). The cell suspension was sonicated, and the inclusion bodies were recovered by centrifugation at 10,000 ϫ g for 20 min. After washing by resuspension in breakage buffer containing 2.5 M urea, the inclusion bodies were dispersed in 6 ml of 6 M guanidinium-HCl, 10 mM DTT, and the extract was centrifuged at 100,000 ϫ g for 1 h. The supernatant was applied in 1-ml aliquots to a Bio-Rad Hi-Pore RP-304 reversed phase column (250 ϫ 4.6 mm), and the protein was eluted with a nonlinear gradient from 0.15% (v/v) trifluoroacetic acid to 0.1% (v/v) trifluoroacetic acid, 80% (v/v) CH 3 CN. Homogeneous YOR163w gene product eluted at 50% (v/v) CH 3 CN.
Protein Expression in Yeast and Purification-S cerevisiae strain INVScI was transformed with pPGY163W. A single colony was picked from an SC-Ura (Synthetic Complete medium without uracil) agar plate and inoculated into 100 ml of SC-Ura medium supplemented with 5% glucose. After 36 h, the cells were harvested by centrifugation, resuspended in 1 liter of SC-Ura (5% glucose), and further grown for 24 h. The cells (4.27 g) were again recovered by centrifugation, resuspended in 1 liter of SC-Ura (2% galactose, 1% raffinose), and grown for 16 h to fully induce expression of YOR163w. The induced cells (8.1 g) were harvested, washed, and resuspended in 8 ml of breakage buffer (50 mM Tris acetate, pH 7.5, 0.3 M NaCl, 10 mM 2-mercaptoethanol, 1 mM phenylmethylsulfonyl fluoride, 5 M trans-epoxysuccinyl-L-leucylamido-(4-guanidino)butane, and 1 mM benzamidine). An equal volume of 0.5-mm glass beads was added, and the cells were disrupted by shaking at 1600 rpm for 10 min at 4°C in a Mikro-Dismembrator U (B. Braun Biotech). The homogenate was decanted, and a cytosolic extract was recovered by centrifugation at 100,000 ϫ g at 4°C for 1 h. Crude extract (13.5 ml) was applied at 0.5 ml/min to a 50 ϫ 15-mm column of Ni 2ϩ -nitrilotriacetic acid-agarose (Sigma) equilibrated with 50 mM Tris acetate, pH 7.5, 0.3 M NaCl, and 10 mM 2-mercaptoethanol. After eluting the unbound protein, a linear gradient of 0 -30 mM histidine in equilibration buffer was applied at 1 ml/min. Fractions (4 ml) were assayed for Ap 6 A hydrolytic activity, pooled, and dialyzed against 10 mM sodium phosphate, pH 6.8, 0.01 mM CaCl 2 . The dialysate (20 ml) was applied at 1 ml/min to a 100 ϫ 7.8-mm Bio-Gel HPHT column (Bio-Rad), and the protein eluted with a linear gradient from 10 mM sodium phosphate, pH 6.8, 0.01 mM CaCl 2 to 350 mM sodium phosphate, pH 6.8, 0.01 mM CaCl 2 . Homogeneous YOR163w gene product eluted at 180 mM sodium phosphate.
Enzyme Assays and Product Identification-Potential substrates were screened by measuring the P i released by co-incubation of substrate with YOR163w protein and either inorganic pyrophosphatase or alkaline phosphatase. The standard assay (200 l) with phosphomonoester substrates was incubated for 30 min at 37°C and contained 50 mM Bis-Tris Propane buffer, pH 6.9, 5 mM MgCl 2 , 1 mM DTT, 0.35 mM substrate, 1 g (1 milliunit) YOR163w protein and 0.5 g (100 milliunit) inorganic pyrophosphatase. Assays with phosphodiester substrates contained 1 g (2 units) alkaline phosphatase instead of the pyrophosphatase. The P i released in each case was measured colori-metrically (17). Chromatographic fractions were screened for activity with 100 M Ap 6 A using a rapid luminometric assay supplemented with 1 mM DTT (18). The reaction products were identified by high performance ion-exchange chromatography. Assay mixtures (100 l) containing 50 mM Tris-HCl, pH 7.5, 5 mM MgCl 2 , 1 mM DTT, 0.16 mM substrate and 1 g of YOR163w protein were incubated for up to 10 min at 37°C and applied to a 1-ml Resource-Q column (Amersham Pharmacia Biotech) at 2 ml/min in 35 mM NH 4 HCO 3 , pH 9.6. The elution system comprised Buffer A (H 2 O) and Buffer B (0.7 M NH 4 HCO 3 , pH 9.6, with a gradient of 5-100% Buffer B over 10 min. Peaks were identified with the aid of standards and quantified by area integration. Kinetic parameters for p 5 A and p 4 A were calculated by measuring the initial rate of product formation by high performance liquid chromatography as described above. Because some of the products of Ap 6 A and Ap 5 A breakdown are also substrates for the enzyme, parameters for the dinucleotides were calculated from the initial rate of adenosine production upon co-incubation of the substrates with YOR163w protein and alkaline phosphatase. Assays containing 50 mM Bis-Tris Propane buffer, pH 6.9, 5 mM MgCl 2 , 1 mM DTT, substrate (various concentrations), 0.2 g (0.2 milliunit) of YOR163w protein, and 0.4 g (0.8 unit) of alkaline phosphatase were incubated for 5-10 min at 37°C and then applied at 1 ml/min to a 250 ϫ 4.6-mm Phenomenex Jupiter C18 column in 4 mM potassium phosphate, pH 6.1, 8% (v/v) methanol. The adenosine peak was integrated.
Determination of Site of Nucleophilic Substitution by Mass Spectrometry-A reaction mixture containing 100 mM Bis-Tris Propane buffer, pH 6.9, 10 mM MgCl 2 , 2 mM DTT, and 6 g of YOR163w protein was prepared in a final volume of 50 l. After freezing at Ϫ70°C, 50 l of ice-cold 2.0 mM Ap 6 A was added, and the mixture was immediately returned to Ϫ70°C and then lyophilized. The reaction was initiated by reconstitution in 100 l of H 2 18 O, and Ap 6 A was hydrolyzed completely to yield the major reaction products by incubation at 37°C for 20 min. Products were separated by high performance anion-exchange chromatography on a 1-ml Resource-Q column as described above. Peaks, monitored by their absorbance at 259 nm, were collected manually and lyophilized, and the distribution of the 18 O between AMP, ADP, and p 4 A was determined by positive ion electrospray mass spectrometry. Masses were also determined for the products of a control assay reconstituted in H 2 16 O. Samples for electrospray mass spectrometry were reconstituted in 20 l of 50% (v/v) acetonitrile, 0.1% (v/v) formic acid and injected at 10 l/min in to a VG-Quattro quadrupole mass spectrometer (Micromass U.K., Ltd.). Analysis was performed at a source temperature of 60°C, capillary voltage of 3.1 kV, and cone voltage of 45 V and with the skimmer offset to 5 V. A scan time of 1 s was employed, and the mass ranges of m/z 330 -365 Da, m/z 420 -455 Da, m/z 570 -630 Da were scanned for AMP, ADP, and p 4 A, respectively.
Immunoblotting-Protein extracts were analyzed on a 90 ϫ 50 ϫ 0.75-mm 15% SDS-polyacrylamide gel. The gel was equilibrated immediately after electrophoresis in transfer buffer (10 mM CAPS-NaOH, pH 11.0, 10% (v/v) methanol) for 10 min before electrophoretic transfer of the separated polypeptides to a nitrocellulose membrane at 150 mA and 4°C for 2 h. The membrane was blocked for 2 h at room temperature with phosphate-buffered saline containing 3% fat-free powdered milk and 0.2% Tween-20 and then probed with a 1:5000 dilution of whole rabbit anti-YOR163w antiserum (raised by standard procedures) followed by a 1:5000 dilution of horseradish peroxidase-conjugated goat anti-rabbit second antibody. After washing, the membrane was developed by enhanced chemiluminescence using the Amersham Pharmacia Biotech ECL kit.
Other Methods-Protein concentrations were measured by the Coomassie Blue dye binding method (19), using a mixture containing equal weights of bovine serum albumin, conalbumin, cytochrome c, and myoglobin as a standard.

Cloning and Expression of the YOR163w Gene Product
Open reading frame YOR163w potentially encodes a 188amino acid protein with a molecular mass of 21,575 Da. The intronless gene was polymerase chain reaction-amplified from the cosmid clone pUOA1258 using 29-base forward and reverse primers that included NcoI and BamHI sites, respectively, and the polymerase chain reaction product was inserted into the pET-15b expression vector. E. coli strain BL21(DE3) was transfected with the recombinant plasmid, and exponentially growing cells were induced for up to 3 h with isopropyl-1-thio-␤-Dgalactopyranoside. SDS-polyacrylamide gel electrophoresis of cell extracts revealed the presence in both the soluble fraction (10 -20%) and inclusion bodies (80 -90%) of a major band migrating with an apparent molecular mass of 24 kDa, which increased with induction time and so was presumed to be the required product (Fig. 1). Inclusion bodies were then isolated after 4 h of induction and solubilized, and the protein was purified in a single step to homogeneity by reversed phase chromatography (Fig. 2). The N-terminal sequence of the purified recombinant protein was determined to be GKTADNHG-PVRS by Edman degradation, and its mass was measured as 21,443 Da by electrospray mass spectrometry. These correspond exactly to the predicted sequence and mass (21,443.6 Da) for the 187-amino acid polypeptide lacking the N-terminal methionine. These data confirm the accuracy of the cloning procedure and the identity of the protein.
In order to confirm that the observed properties of the enzyme were not due to an alternative folding of the protein after reversed phase chromatography, the enzyme was also overexpressed in a soluble form in a yeast host system and purified conventionally by chromatography on a nickel affinity resin and hydroxyapatite. The enzyme binds tightly to the nickel column even though it was not expressed with a histidine tag. Both procedures yielded enzyme with very similar properties. The data presented here were obtained with the bacterially expressed protein.

Properties of the Protein
Substrates-Almost all MutT motif proteins studied so far are nucleotide pyrophosphatases that hydrolyze compounds containing an NDP linked to another moiety, hence the alternative name of nudix hydrolase (8). A wide range of nucleotides was assayed to determine the substrate(s) of the YOR163w protein. Of these, only Ap 6 A, p 5 A, Ap 5 A, and p 4 A yielded significant activity (Table I). ATP was very slowly degraded to ADP ϩ P i , whereas no activity was detectable with the following compounds in the presence of Mg 2ϩ or Mn 2ϩ ions: Ap 4 A, Ap 3 A, Ap 2 A, NAD ϩ , NADH, NADP ϩ , NADPH, desamino-NAD ϩ , FAD, coenzyme A, (deoxy)nucleoside 5Ј-triphosphates (GTP, CTP, UTP, ITP, dATP, dGTP, dCTP, and TTP), nucleoside 5Ј-diphosphates (ADP, GDP, CDP, and UDP), nucleoside 5Ј-monophosphates (AMP, GMP, CMP, and UMP), nucleotide sugars (ADP-ribose, IDP-ribose, ADP-glucose, GDP-glucose, GDP-mannose, CDP-glucose, UDP-glucose, UDP-galactose, and UDP-N-acetylgalactosamine), or nucleotide alcohols (CDPglycerol, CDP-choline, and CDP-ethanolamine). Diguanosine polyphosphates were not tested as substrates. Table I shows the kinetic constants obtained with the active substrates. K m values were all within the range 30 -70 M and had apparent catalytic constants (k cat ) below 1 s Ϫ1 . According to the calculated catalytic efficiencies (k cat /K m ), Ap 6 A is an 8-fold better substrate than Ap 5 A, the overall preference being Ap 6 A Ͼ p 5 A Ͼ Ͼ p 4 A Ͼ Ap 5 A. Kinetic parameters for p 5 A were calculated using subsaturating substrate concentrations only because substrate inhibition by this compound was observed above 50 M. Given the preference for Ap 6 A, we propose that this protein should be described as a diadenosine 5Ј,5ٞ-P 1 ,P 6 -hexaphosphate hydrolase.
Reaction Requirements-With 160 M Ap 6 A as substrate, the enzyme displayed optimal activity at pH 6.9, 50°C, and 4 -10 mM Mg 2ϩ . Mn 2ϩ at 200 M also sustained optimal activity, but Ca 2ϩ inhibited with an IC 50 of 3.3 mM. F Ϫ was also inhibitory (noncompetitive), with a K i of 80 M. In this respect, the yeast Ap 6 A hydrolase is similar to but less sensitive than the plant   2. Purification of YOR163w gene product. YOR163w gene product was purified in a single step from solubilized inclusion bodies as described under "Experimental Procedures." -----, acetonitrile gradient. Inset, a sample of the pooled peak was analyzed by SDS-polyacrylamide gel electrophoresis in a 15% gel. Lane 1, protein standards as in Fig. 1; lane 2, 1 g of YOR163w gene product. and animal Ap 4 A hydrolases, which have K i values for F Ϫ in the ranges 2-3 and 20 -30 M, respectively (20). The enzyme had an absolute requirement for a reducing agent, such as DTT (optimal at 1 mM).
Reaction Products-The reaction products generated from each of the four active substrates were determined after various incubation times by ion-exchange high performance liquid chromatography (Table I and Fig. 3). From the kinetics of product formation, the following overall conclusions were drawn. Ap 6 A yielded mainly p 4 A ϩ ADP (76%) but also p 5 A ϩ AMP (24%). AMP must be a primary breakdown product because ADP is resistant to further hydrolysis, hence the enzyme displays two alternative modes of attack on the Ap 6 A substrate. p 5 A, either alone or as a product of Ap 6 A breakdown, generated almost exclusively p 4 A ϩ P i . The example high performance liquid chromatography profile in Fig. 3 shows only a small amount of p 5 A; however, assays using shorter incubation times clearly showed the generation of equimolar amounts of p 5 A and AMP before the p 5 A itself is degraded to p 4 A ϩ P i . ATP was also observed, most likely due to the breakdown of the primary p 4 A product (see below). Similarly, Ap 5 A yielded predominantly p 4 A ϩ AMP (96%), but with a small amount of ATP ϩ ADP (4%), whereas p 4 A broke down to ATP ϩ P i . The preferential generation of p 4 A from both Ap 6 A and Ap 5 A suggests a reaction mechanism similar to the plant and animal Ap 4 A hydrolases, which always generate ATP from Ap n A substrates (n Ն 4). A binding pocket on these enzymes accommodates a pppA moiety, with the fourth phosphorus distal to the A being subject to nucleophilic attack by water (21,22). By analogy, the yeast Ap 6 A hydrolase appears to preferentially accommodate a pp-ppA moiety in the substrate binding site.
Regarding the generation of alternative products, this could occur in one of three ways. Using Ap 6 A as an example, these are (i) exclusive attack of the nucleophile (presumed to be water) on P ␤ , with elimination of the P ␤ -O(P ␥ ) bond, yielding p 4 A ϩ ADP, and elimination of the P ␤ -O(P ␣ ) bond, yielding p 5 A ϩ AMP; (ii) attack on P ␥ and elimination of the P ␥ -O(P ␤ ) bond, yielding p 4 A ϩ ADP, and attack on P ␤ and elimination of the P ␤ -O(P ␣ ) bond, yielding p 5 A ϩ AMP; (iii) attack on P ␤ and elimination of the P ␤ -O(P ␥ ) bond, yielding p 4 A ϩ ADP, and attack on P ␣ and elimination of the P ␣ -O(P ␤ ) bond, yielding p 5 A ϩ AMP. These possibilities can be distinguished by carrying out the reaction in the presence of H 2 18 O and following the fate of the 18 O by mass spectrometry (23) (Fig. 4). When Ap 6 A was hydrolyzed in the presence of H 2 16 O, the AMP and ADP  products had masses of 348 and 428 Da, respectively (Fig. 5, A and C), whereas hydrolysis in the presence of H 2 18 O resulted in fully 18 O-labeled AMP and ADP, with masses of 350 and 430 Da, respectively (Fig. 5, B and D). In both cases, the p 4 A product was unlabeled, with a mass of 588 Da (Fig. 5, E and F), whereas rapid degradation of the p 5 A prevented an assessment of its labeling pattern. Because only mechanism iii leads to labeling of both AMP and ADP and lack of labeling of p 4 A (Fig.  4), this must be the normal mode of attack and is, therefore, identical to that previously established for the Artemia and lupin Ap 4 A hydrolases (22)(23)(24).
Presence in Yeast-A rabbit polyclonal antibody was raised against the recombinant YOR163w protein in order to confirm that this protein is normally expressed in yeast. Fig. 6 shows that exponentially growing yeast cells express a protein that co-migrates on SDS-polyacrylamide gel electrophoresis with YOR163w. The native protein also migrates with an anomalously high mass of 24 kDa. This phenomenon has also been observed with the 16.7-kDa human Ap 4 A hydrolase, which usually migrates as 19 -21 kDa (7). DISCUSSION The recovery of active YOR163w protein after reversed phase chromatography and its high temperature optimum of 50°C reflect the high stability of the MutT motif proteins as a family. This purification system has also been found to yield pure and fully active recombinant human Ap 4 A hydrolase. 3 Stability can probably be attributed to the mixed ␤-sheet structure shown to be at the core of the E. coli MutT protein itself and other highly stable proteins (25).
So far, Ap 6 A and Ap 5 A have only been described in the secretory granules of certain specialized mammalian cells. It is not known whether they are present in the cytosol of eukaryotes in general, including yeast. If so, and if they are synthesized by aminoacyl-tRNA synthetases in a reaction similar to the lower order diadenosine polyphosphates (26,27), then both p 5 A and p 4 A would be required as adenylate acceptors. Neither of these compounds exists at detectable levels in vegetative yeast cells; however, they are both synthesized and excreted during the latter stages of sporulation following ascospore formation, reaching 1.5 and 2% of the concentration of ATP, respectively (28). They are not produced by asporogenous a/a or ␣/␣ strains placed in sporulation medium, and so they have been proposed as signals marking the end of sporulation (28). They may be synthesized by acetyl-CoA synthetase, which is known to generate them in vitro (29). Their presence in yeast cells suggests that Ap 6 A and Ap 5 A might also be synthesized at low levels during sporulation. Because Ap 5 A is a potent inhibitor (active in the nanomolar range) of the essential enzyme adenylate kinase (30), one function of the Ap 6 A hydrolase may 3 J. L. Cartwright  be to eliminate these potentially toxic dinucleotides during sporulation. In this context, it is of interest to note that the region upstream of the YOR163w gene contains a single copy of the stress response element 5Ј-AGGGG, which is known to contribute to the response to nitrogen starvation (31). The possibility that YOR163w expression is regulated by cellular stress remains to be determined. Alternatively, the accumulation of p 5 A and p 4 A during this period may reflect a higher activity of the Ap 6 A hydrolase during vegetative growth, its function being the removal of the substrates for Ap 6 A and Ap 5 A synthesis. Such functions would be in keeping with the "housecleaning" role proposed for the nudix hydrolase family (8).
In generating AMP ϩ p 5 A from Ap 6 A, the reaction mechanism of the yeast Ap 6 A hydrolase is identical to that previously determined for the production of AMP ϩ ATP by the Artemia and lupin Ap 4 A hydrolases, namely nucleophilic attack of water on P ␣ and elimination of the P ␣ -O(P ␤ ) bond (22)(23)(24). Interestingly, the Artemia Ap 4 A hydrolase can be forced to switch attack to P ␤ when presented with substrates containing nonscissile P ␣ -C phosphonate linkages, such as diadenosine 5Ј,5ٞ-P 1 ,P 4 -(P 1 ,P 2monofluoromethylene-P 3 ,P 4 -monofluoromethylene) tetraphosphate (ApCHFppCHFpA) (21), or ␣-thiophosphates, such as (R p ,S p )-diadenosine 5Ј,5ٞ-P 1 ,P 4 -(P 1 ,P 4 -dithio)-tetraphosphate (Ap s ppp s A) (32). This so-called frameshift mode of attack was attributed to a flexibility in the binding of the polyphosphate substrate to the enzyme, with either P ␣ or P ␤ being positioned next to a fixed catalytic center, a situation more commonly encountered with polymeric substrates (21). This flexibility is also demonstrated by the yeast Ap 6 A hydrolase with the natural substrate Ap 6 A and, to a lesser extent, Ap 5 A, with alternative sets of products being generated in each case. Fig. 7A shows a partial sequence alignment of the YOR163w protein with other known eukaryotic dinucleoside polyphosphate hydrolases, including the YA9E protein from Schizosaccharomyces pombe, which shares 43% sequence identity with YOR163w. The gene encoding YA9E has recently been cloned and expressed, 4 and the protein has Ap n A hydrolase activity with Ap 6 A and Ap 5 A as the preferred substrates, but with some activity toward Ap 4 A, in contrast to YOR163w, which has no activity with this substrate. Several observations can be made.
First, YOR163w has an extra proline residue inserted in the MutT motif, the sequence common to all members of this protein family. This may in part explain the exclusive accommodation of the longer polyphosphate chains compared with the Ap 4 A hydrolases and the YA9E protein, which do not have this extra residue. Second, the hydrophobic patch in the fungal enzymes located just N-terminal to the MutT motif (YOR163w residues 47-50) is more similar to that in the animal Ap 4 A hydrolases than the plant enzyme sequences (Fig. 7A). However, further toward the N terminus (YOR163w residues 26 -  human homologs (B). A, the YOR163w sequence (GenBank TM accession number Z75071) was aligned with lupin (GenBank TM accession number U89841) and barley (GenBank TM accession number Z99996) Ap 4 A hydrolases, S. pombe YA9E protein (SwissProt accession number Q09790), and human (SwissProt accession number P50583), pig (SwissProt accession number P50584) and mouse (SwissProt accession number P56380) Ap 4 A hydrolases using the CLUSTAL W program. B, the YOR163w sequence (GenBank TM accession number U55021) was aligned with the human diphosphoinositol polyphosphate phosphohydrolase DIPP (GenBank TM accession number AF062529) and the related human sequence of unidentified function (GenBank TM accession number AA916467). Amino acid identities with YOR163w are shaded black, amino acid similarities are shaded gray, the MutT motif (GX 5 EX 7 REX 2 EEXG) is underlined, and the hydrophobic patch is overlined. Numbers to the right and left of the sequences represent the positions in the respective complete amino acid sequences. 40), both of the fungal proteins, especially YOR163w, share additional sequence similarity with the two plant Ap 4 A hydrolases. This similarity is absent from the animal hydrolases, which do not align with any significance in this region. The fungal and plant enzymes also share the enzymic property of hydrolyzing both nucleoside and dinucleoside polyphosphates: the lupin Ap 4 A hydrolase degrades p 4 A, whereas this compound is a potent inhibitor of the animal Ap 4 A hydrolases (6,33). Thus, there may be a closer evolutionary relationship between the plant and fungal Ap n A hydrolases. Another activity that may be related is the dinucleoside polyphosphate hydrolase purified from the green alga Scenedesmus obliquus, an organism that, like S. cerevisiae, has an Ap 4 A phosphorylase (34). This enzyme hydrolyzes Ap n A with the preference Ap 5 A Ͼ Ap 4 A Ͼ Ap 6 A. No significant similarity to the E. coli orf186 gene product, an enzyme that prefers Ap 3 A as substrate but that also hydrolyzes ADP-ribose and NADH (35), was detected outside the MutT motif.
With regard to possible mammalian orthologs of YOR163w, a 41-amino acid sequence of this protein encompassing the MutT motif shows 50% identity and 65% similarity with two closely related but distinct sequences that are represented by several human, mouse, and rat clones in the GenBank TM expressed sequence tag data base. The alignment with the two human sequences is shown separately from the other dinucleoside polyphosphate hydrolases in Fig. 7B for clarity. One of these, DIPP, has recently been shown to be a diphosphoinositol polyphosphate phosphohydrolase, the first MutT motif protein with activity toward non-nucleotide substrates (36). Like YOR163w, DIPP shows positional flexibility in its site of attack. It is believed to attack a different pyrophosphoryl group in its two substrates, diphosphoinositol pentakisphosphate and bis(diphosphoinositol) tetrakisphosphate. Given the broad substrate specificity of some MutT motif proteins and the similarity between these proteins, it will be of particular interest to determine whether YOR163w can also hydrolyze diphosphoinositol polyphosphates.
In conclusion, we have characterized a dinucleoside polyphosphate nudix hydrolase from S. cerevisiae with a novel substrate specificity. Given the current power of genetic manipulation in yeast, phenotypic analysis of YOR163w knockouts in appropriate genetic backgrounds (e.g. apa1 and apa2 mutants) should help determine the role(s) of this enzyme and the relative contribution of hydrolases and phosphorylases to Ap n A catabolism and function in those organisms in which both types of enzyme exist (34).