Demonstration of a Novel Type of ATP-diphosphohydrolase (EC 3.6.1.5) in the Bovine Lung*

A novel type of ATP-diphosphohydrolase (ATPDase) is demonstrated in bovine lung. The enzyme has an optimum pH of 7.5 and catalyzes the hydrolysis of the B- and y-phosphate residues from diphospho- and triphosphonucleosides. It requires Ca2+ or Mg2+ and is insensitive to ouabain, an inhibitor of Na’/K+-ATPase, P’,P-di(adenosine 5’)-pentaphosphate, an inhibitor of adenylate kinase, and tetramisole, an inhibitor of alkaline phosphatase. In contrast, sodium azide (10 mM), a known inhibitor of ATPDases and mitochondrial ATPases, as well as mercuric chloride (10 p ~ ) and gossypol (2,2’-bis[8-formyl-1,6,7-trihydroxy-5-iso- propyl-3-methylnaphthalene]) (35 p ~ ) are powerful inhibitors of this enzyme. The same inhibition profile is obtained with ATP or ADP as substrate, thereby supporting the concept of a common catalytic site for these substrates. This

Several years ago, we described the first mammalian ATPDase in the pig pancreas (Type I) (20,21). Since then, mammalian ATPDases have been found in rat pancreas (22), mouse liver and brain, dog kidney, human tumors (23), human term placenta (24), and myometrium (25). An ATPDase was also localized in bovine aorta smooth muscles and endothelial cells (26)(27)(28). We have recently demonstrated that the properties of the aorta ATPDase (Type 11) are different from those of the pancreas ATPDase (Type I) (29)(30)(31). Heat denaturation curves, 6oCo y-irradiation-inactivation curves, and migration patterns on polyacrylamide gel electrophoresis under nondenaturing conditions are the parameters on which we reached our conclusion.
The importance of extracellular nucleotides in blood coagulation is well established (32)(33)(34). ADP stimulates platelet aggregation, ATP antagonizes the effects of ADP on platelets, and adenosine inhibits the action of platelet stimulants by their binding on Az receptors (35). The threshold of platelet activation by ADP is reached between 2 and 5 ~L M (36). Results from our laboratory showed that platelet aggregation induced by 2 ~L M ADP can be prevented or even reversed by adding a purified ATPDase fraction from bovine aorta smooth muscle cells (30).
Lung is a highly vascularized tissue where circulating platelets and microemboli become easily trapped in the capillary bed (37). Nevertheless, perfusion studies demonstrate that ATP is rapidly metabolized into adenosine on a single passage through the pulmonary capillary bed (38,39). Using an indicator dilution technique in rat perfused lungs, Smith and Ryan (40) observed that the mean transit times of ATP, AMP, and an intravascular marker (blue dextran) were identical, suggesting that the enzyme activities responsible for degrading ATP and AMP were confined to the vasculature.
Subsequently, a series of elegant experiments utilizing electron microscopic cytochemistry have demonstrated that enzymes responsible for the degradation of ATP and AMP in rat lung were localized to the luminal surface of the plasma membrane of capillary endothelial cells (40, 41).
Until recently, it was generally accepted that extracellular nucleotides would be degraded by three different enzymes: an ATPase, an ADPase, and a 5'-nucleotidase (39, 42-48). Although ADPase activities have been reported in the microvasculature of the lung (38)(39)(40)(41)49), it seems reasonable to propose that a similar endothelial ectonucleotidase system would be present in large and small blood vessels. Since ATPDases have been described in large vessels (26)(27)(28)(29)(30)(31), the purpose of this work was to verify the presence of an ATPDase 4699 in the lung and to compare its properties with those of the pancreas and bovine aorta ATPDases. Our results demonstrate the existence of an hitherto unclescribed type of ATPDase which we propose to identify as Type 111.
ATP-diphosphohydrolase Assays-Enzyme activity was routinely determined at 37 "C in the following incubation medium: 90 mM Tris imidazole (pH 7.5), 2 mM MgCl,, 1.4 mM CaCl,, 5.0 mM tetramisole, 2.0 mM substrate, and 10-50 pg of enzyme in a final volume of 0.5 ml. Reaction was stopped with 0.8 ml of ice-cold 10% (w/v) trichloroacetic acid and inorganic phosphorus was measured according to LeBel et al. (50). One unit of enzyme corresponds to the liberation of 1.0 pmol of phosphate/min/mg of protein a t 37 "C. and V,,,.x,.,p were estimated under the same conditions with [y-3ZP]ATP as the substrate. The liberated 32P was measured as described by LeBel et al. (20). Proteins were estimated by the technique of Bradford using bovine serum albumin as standard (51). A series of inhibitors were tested on ATPase and ADPase activities a t concentrations which were effective in other systems.
Enzyme Denaturation by Physical Methods-Heat denaturation curves were obtained by incubating the microsomal fraction at temperatures varying from 1 to 70 "C for 3 min. ATPase and ADPase activities were then measured as described above. Radiation inactivation of the microsomal fraction was carried out by exposing the enzyme to a y 6oCo source for various lengths of time at -78 "C. The molecular mass of the native form of the enzyme was estimated as previously described (52,53).

RESULTS
A microsomal fraction enriched in both ATPase and AD-Pase activities was isolated by differential centrifugation. Analysis of the different fractions reveals that these activities are concentrated in a microsomal fraction (pellet 11) that sediments at 22,000 x g for 90 min. Further purification is obtained by centrifugation of the fraction on a sucrose cushion (35%) ( Table I). In the latter fraction, the ATPase/ADPase ratio increases substantially, suggesting the existence of nucleotide phosphohydrolase activities in pellet 11 which are removed during fractionation. p H dependence for enzyme catalysis was first examined and as illustrated in Fig. 1, the optimum pH for ADP and ATP were similar at 7.25. Both ATPase and ADPase required Ca2+ or Mg2+. Maximal rates of catalysis were obtained with 15 and 10 mM Ca2+, and with 10 and 4 mM Mg2+, using ATP or ADP, respectively, as substrate (data not shown).
The enzyme preparation was purified by polyacrylamide gel electrophoresis under nondenaturing conditions along with the bovine aorta ATPDase preparation isolated according to C6t6 et al. (29). As shown in Fig. 2, a single band is responsible for the hydrolysis of ATP and ADP in the lung preparation, as for the bovine aorta ATPDase preparation, suggesting that the same enzyme might be hydrolyzing both substrates in bovine lungs. Moreover, the aorta enzyme moved at a locus different from that of the bovine lung preparation. This result supports the view that a distinct isoform of ATPDase is present in lungs.
The presence of an ATPDase activity in bovine lung was further assessed by testing substrate specificity. As shown in Table 11, it was found that the enzyme preparation hydrolyses triphosphonucleosides a t comparable rates and diphosphonucleosides a t slightly slower rates. Moreover, a combination of ATP and ADP at 2 mM concentration was not additive. And very low levels of hydrolysis were measured with nucleoside monophosphate, P-glycerophosphate, andp-nitrophenyl phosphate.
The biochemical properties of the lung ATPDase were further defined with various inhibitors (Table 111). Significant inhibitions were obtained with gossypol (35 FM), an antifertility agent known to reduce ATP content of spermatozoa (54), sodium azide (10 mM), mercuric chloride (10 pM), and tetramisole (5 mM), an alkaline phosphatase inhibitor (55), whereas no significant inhibition was observed with Ap5A (100 p~) , an inhibitor of adenylate kinase (56) and ouabain (3 mM), an inhibitor of Na+/K+-ATPase (57). The level of inhibition were comparable with either ATP or ADP as substrate, further supporting the concept that a common catalytic site is involved in the hydrolysis of these nucleotides. This possibility was confirmed by comparing the inactivation profiles of ADPase and ATPase activities after heat and 6oCo y-irradiation-inactivation. As shown in Fig. 3, superimposable curves of heat inactivation were obtained with ADP and ATP, with a temperature of 63 "C corresponding to 50% residual activity. Radiation-inactivation with 6 o C~ also produced identical curves with ADP and ATP as substrates (Fig.  4). These curves allowed an estimation of the molecular mass of the enzyme on its native form, as described by Kepner  than 5% substrate hydrolyzed). Km,npl, was estimated at 7 ? 2 p~ (mean f S.E.) and Vmn..nl,p at. 1.1 ? 0.3 pmol of Pi/min/ mg of protein (mean ? S.E.).

DISCUSSION
Our laboratory has been involved for many years in the st,udy of ATPDase. As shown by Dixon and Webb (58), t,his enzyme has the property of catalyzing the hydrolysis of y- similarities of denaturation curves by heat and 6oCo y-irradiations with ATP or ADP as substrate. In agreement with our conclusion, combinations of ADP and ATP did not result in any increase in the total phosphate released, as one would expect from the activities of two distinct enzymes. The fact that AMP, P-glycerophosphate, andp-nitrophenyl phosphate are not hydrolyzed to a significant extent by our enzyme preparation, joined to the fact that tetramisole is routinely added to our enzyme assay, rule out the possibility that an alkaline phosphatase is responsible for the hydrolysis of diphospho-and triphosphonucleosides. Ap6A, an effective inhibitor of adenylate kinase, does not alter the hydrolysis rate of ADP or ATP, and therefore rules out the possibility that phosphate is produced from ADP as a result of the conversion of this substrate to AMP and ATP. Moreover, the fact that our enzyme preparation can hydrolyze purine as well as pyrimidine diphospho-and triphosphonucleotides corroborates our conclusion.
Biochemical and electrophoretic properties of the lung ATPDase differ from those of the previously described ATPDases from pig pancreas (Type I) (20, 21) and smooth muscle cells of bovine aorta (Type 11) (29)(30)(31), and so constitute a novel type of ATPDase. The three main facts on which this conclusion was reached are different migration patterns after polyacrylamide gel electrophoresis, and different denaturation curves by heat and 6oCo y-irradiations (29). The latter technique allowed a gross estimation of the native molecular mass of these enzyme. The molecular mass of bovine lung ATPDase (70 f 3 kDa) is much lower than the values reported for pig pancreas (132 kDa) (29) or bovine aorta (189 kDa) (29).
Several ADPase and ATPase activities have been reported in the lung (38-41). For instance, Dawson et al. (49) described plasma-membrane ATPase and ADPase activities in rat lung microsomal fractions. These activities showed an optimum p H of 7.5 and had comparable sedimentation properties. However, the authors at the time believed that these activities were attributable to separate proteins.
Perfusion studies have demonstrated that ATPase, AD-Pase, and 5'-nucleotidase activities in lungs are confined to the vascular endothelial cells and that nucleotides do not enter the cells prior to their hydrolysis into adenosine (40). Since ATPDases described in large vessels (26-31) are involved in the control of platelet aggregation and blood coagulation, it seems reasonable to propose that the ATPDase isolated in bovine lung fullfills a similar task in pulmonary capillary bed. Our laboratory has recently reported ATPDase activity in brain capillaries.* Defective hydrolysis of extracellular nucleotides in blood vessels have been associated to various diseases. For instance, pulmonary thrombosis is particularly associated with diabetes (59-61). Perfusion studies performed on diabetic rat lungs demonstrate that ADPase and AMPase activities are significantly reduced, thus increasing risks of thrombosis (62). We are now investigating the localization of this ATPDase in bovine lung and the demonstration of its physiological role in such pathological conditions.