Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

ATP is released as a cotransmitter from cholinergic, adrenergic, and GABAergic neurons (Silinsky et al., 1998; Burnstock, 1999; Jo and Schlichter, 1999). Traditionally, the inactivation of neurotransmitter ATP in both the central and the peripheral nervous systems has been attributed to its breakdown by cell membrane-bound enzymes, classified as ecto-ATPases, ecto-apyrases, and ecto-5'-nucleotidases (Zimmermann, 1992; Plesner, 1995).

We have demonstrated, however, that neurogenic stimulation of the guinea pig vas deferens dramatically accelerates the degradation of exogenous ATP (Todorov et al., 1996). The difference between the rate of degradation of extracellular ATP by tissue preparations under resting conditions and that during nerve stimulation appears to be associated with a release of enzymes that break down ATP as well as ADP and AMP into ADO (Todorov et al., 1997). Inhibition of the propagation of neuronal action potentials with tetrodotoxin, suppression of adrenergic neurotransmission with guanethidine, or inhibition of exocytosis by omission of extracellular Ca²⁺ all prevented the release of nucleotidase activity, implying that sympathetic nerves are the source of the enzyme(s) (Todorov et al., 1997). Interestingly, the nucleotidase activity appears to be coreleased with neurotransmitter ATP and not with the sympathetic cotransmitter NE (Mihaylova-Todorova et al., 2001), suggesting that the proteins carrying the enzyme activity originate from a putative "ATP storage vesicle" rather than from a catecholamine storage vesicle.

We investigated the possibility that known ATPases, activated during the process of exocytosis, may be involved as releasable ATPases. The vacuolar H⁺-transporting ATPase, the Na⁺/K⁺-ATPase, the multidrug-resistance channel, and the cytosolic N-ethylmaleimide-sensitive fusion protein were rejected as possible candidates based on the evidence that their specific antagonists did not inhibit the releasable ATPase activity. We have found, however, that suramin and 6-N,N-diethyl-,-dibromomethylene-D-ATP (ARL67156) were potent inhibitors of the releasable ATPase activity (Todorov et al., 1997). Because both drugs have been shown to inhibit ecto-ATPases (Crack et al., 1995) we speculated that the releasable nucleotidases might share structural similarities with ecto-ATPases (Kennedy et al., 1997; Todorov et al., 1997).

Interestingly, none of the known members of the mammalian ecto-ATPase gene family, recently renamed ecto-nucleoside triphosphate diphosphohydrolases (ENTPDases) (EC 3.6.1.5) (Zimmermann and Braun, 1999), matches the neuronally released enzymes with regard to both substrate specificity and membrane localization. In fact, the ENTPDase members that possess ATPase activity such as ENTPDase1 (Maliszewski et al., 1994; Kaczmarek et al., 1996), ENTPDase2 (Kegel et al., 1997; Mateo et al., 1999; Vlajkovic et al., 1999), or ENTPDase3 (Smith and Kirley 1998) are membrane-bound proteins, whereas expression of the potentially soluble members such as the ENTPDase5 or ENTPDase6 (Chadwick and Frischauf, 1998) revealed that they poorly metabolize ATP (Mulero et al., 1999; Braun et al., 2000; Hicks-Berger et al., 2000). Moreover, none of the members of the ENTPDase family hydrolyze nucleotide monophosphates to nucleosides. Because soluble ENTPDases with ATPase activity have not as yet been identified, and the enzyme(s) released upon stimulation of the sympathetic nerves of the guinea pig vas deferens hydrolyzes AMP as well as ATP and ADP, other potential candidates for the releasable nucleotidases should be considered.

Members of the mammalian family of ecto-nucleotide pyrophosphatases/phosphodiesterases (ENPPases) break down nucleotide triphosphates, diphosphates, and dinucleotide-polyphosphates into monophosphates. Although ENPPases have a short membrane-spanning domain, the large extracellular domain containing the catalytic center could be cleaved from the membrane and released in soluble form (for review, see Bollen et al., 2000). However, ENPPases do not metabolize AMP to adenosine.

Alkaline- and tissue-nonspecific acid phosphatases are able to hydrolyze ATP, ADP, and AMP and exist in both glycosylphosphatidylinositol (GPI)-anchored and soluble forms (Ohkubo et al., 2000). Ecto-5'-nucleotidase (E.C. 3.1.3.5) or CD73 (Resta et al., 1993) is a 5'-monophosphoadenosine- or inosine-specific phosphatase (Zimmermann, 1992). Upon activation of phosphatidylinositol-specific phospholipase C, the GPI-anchored ecto-protein could be released in soluble form, while retaining its catalytic activity (Misumi et al., 1990; Lehto and Sharom, 1998).

Given the broad, mono-, di-, or triphosphonucleotide specificity of the releasable neuronal nucleotidases we explored the possibility that ENTPDases, ENPPases, phosphodiesterases, 5'-nucleotidases, or phosphatases may be responsible for the releasable nucleotidase activity. Our results suggest that the sympathetic nerves of the guinea pig vas deferens release either a soluble, heretofore unidentified enzyme that combines separate phosphohydrolase and 5'-monophosphate-diesterase catalytic activities or a mixture of separate enzyme entities, including a soluble ENTPDase-like ATPase and a soluble ecto-5'-nucleotidase-like AMPase.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Tissue Preparation. Male albino guinea pigs (350-400 g) were killed by decapitation. The vasa deferentia were removed, cleaned of connective tissue, and the lumen exposed by a section along the longitudinal axis. Three tissues, each from a different animal, were loaded in a superfusion chamber (inner volume of 200 µl; Brandel Inc., Gaithersburg, MD). Whatman (Maidstone, UK) 541 filters were cut to fit both ends of the chamber, which was then inserted vertically into a thermostatic block (36°C) with platinum "screen" electrodes at each end. The tissues were superfused from bottom to top (2 ml/min) with modified Krebs-HEPES buffer, pH 7.4, of the following composition: 140 mM NaCl, 5 mM KCl, 1 mM MgCl₂, 1.5 mM CaCl₂, 5 mM HEPES, and 11 mM glucose. The buffer was constantly bubbled with 100% O₂.

Nerve Stimulation and Sample Collection Protocols. The sympathetic nerves of the guinea pig vasa deferentia were stimulated for 30 s by electrical field stimulation (EFS) at 16 Hz, pulse duration of 0.2 ms, and supramaximal voltage. Three sessions of EFS were applied to the tissues at 30-min intervals. Samples of the Krebs-HEPES buffer superfusing the tissue preparations were collected for 30 s before and for 30 s during each of the stimulations in ice-cold test tubes containing the protease inhibitor leupeptin (1 µM). The samples were combined in two pools, one designated as prestimulation or P (collections before stimulation) and the other designated as S (collections during stimulation). If not used the same day, the pooled samples were frozen in liquid nitrogen and then stored at 86°C.

HPLC-Based Assay for Nucleotidase Activity. To study the properties of the releasable neuronal nucleotidases, the fluorescent 1,N⁶-etheno analogs of adenine nucleotides were used as substrates (Fig. 1). The rationale of the assay is based on the fact that the sequential dephosphorylation of 1,N⁶-etheno ATP (eATP) results in formation of 1,N⁶-etheno ADP (eADP) followed by formation of 1,N⁶-etheno AMP (eAMP), and finally formation of 1,N⁶-etheno adenosine (eADO). Each of the substrates and the resulting metabolites were quantified after separation by HPLC coupled with fluorescent detection. With this method the rate of hydrolysis of the substrate, as well as the type and rate of generation of the products were evaluated in a single chromatogram. eATP was used as initial substrate to study ATPase activity (Fig. 1A), whereas eADP was used to study ADPase activity (Fig. 1B), and eAMP was used to study AMPase activity (Fig. 1C)

View larger version (17K): [in this window] [in a new window]   Fig. 1.   Chromatographic image of etheno-purines before (Aa, Ba, and Ca) and after (Ab, Bb, and Cb) exposure to releasable nucleotidases. Structural formulae and representative chromatograms of eATP (Aa), eADP (Ba), and eAMP (Ca) in Krebs-HEPES buffer. The structural formula of eADO is shown as an inset in Cb. The etheno-modification, consisting of insertion of acyl group between nitrogen 1 and 6 of the adenine moiety, renders the molecule fluorescent but leaves unaltered the phosphate chain. Note that an invariable with time nonenzymatic dephosphorylation product (less than 5% for eATP, less than 8% for eADP, and less than 2% for eAMP) is always detected in the solutions. eATP incubated for 60 min at 37°C with superfusate S collected during nerve stimulation of the guinea pig isolated vas deferens was metabolized to eADP, eAMP, and eADO (Ab). Likewise, in presence of superfusate S, eADP was metabolized to eAMP and eADO (Bb), whereas eAMP was broken down to eADO (Cb). Metabolism of etheno-adenine nucleosides did not occur in presence of superfusate P, collected before the onset of the nerve stimulation of the guinea pig isolated vas deferens (data not shown).

HPLC-Based Assays for Adenine Nucleotides and ADO. ATP, ADP, AMP, and ADO were analyzed as described previously (Todorov et al., 1996). The etheno-adenine purines were separated on a gradient HPLC system equipped with a Resolve radial pack cartridge (8NV Ph 4 µm; 8 × 10 mm) (Waters, Milford, MA). The amount of each adenine purine was quantified using an RF 535 fluorescent monitor (Shimadzu, Columbia, MD) at an excitation wavelength of 230 nm and an emission wavelength of 420 nm. Buffer solutions consisted of 0.1 M phosphate (KH₂PO₄, pH 6.0) (buffer A) and 75% 0.1 M phosphate and 25% methanol (buffer B). The adenine nucleotides and ADO were separated using a gradient in which the concentration of buffer B was increased from 0 to 100% in 8 min according to Waters gradient profile 7. The HPLC equipment was controlled by, and data collected by, a Pentium II computer equipped with an LAC/E card and Millenium 2010 Chromatography Manager software (Waters). Identification of individual peaks in chromatograms was by comparison with the retention times of known etheno-adenine purine standards, and the concentration was determined by peak area per picomole relationship compared with standards. Standards were run with each set of samples.

Preparation of 1,N⁶-Etheno Derivatives of ATP, ADP, AMP, and ADO. Stock solutions of eATP, eADP, and eAMP were prepared by incubation of ATP, ADP, and AMP (1.5 × 10² mol/l) in citric phosphate buffer, pH 4, in the presence of 2-Cl-acetaldehyde for 40 min at 80°C. Furthermore, serial dilutions (3 µM-10 mM) of the stock solutions were prepared using deionized water (18 M) and stored at 20°C. The etheno-derivatives of adenine nucleotides are stable at 20°C for years and at room temperature for several days.

Hydrolysis of ATP, ADP, and AMP and Their Etheno-Derivatives by Commercially Available Enzymes. To test whether the modification of the molecules of ATP, ADP, or AMP by the etheno-group addition affects the enzymatic degradation of adenine nucleotides we have compared the rate of breakdown of native ATP and ADP with the rate of breakdown of their etheno-derivatives (eATP and eADP) by apyrase VI and VII purchased from Sigma-Aldrich (St. Louis, MO). The hydrolysis of AMP and that of eAMP by 5'-nucleotidase from Crotalus atrox venom (Sigma-Aldrich) was also evaluated. The etheno-analogs were metabolized by the commercially obtained enzymes with rates closely comparable with the rates for the nonmodified substrates (data not shown).

Time Course of Product Formation. eATP, eADP, or eAMP (1, 10, and 100 µM) was incubated with superfusate S or P for 0, 0.5, 1, 2, and 4 h at 37°C. At the end of the incubation periods the reactions were stopped with acidification with ice-cold citrate phosphate buffer, pH 4, and the samples analyzed for ATP, ADP, AMP, and ADO by HPLC.

Increase in Enzyme Activity with Protein Concentration. Samples of superfusate diluted 2-, 4-, or 8-fold with Krebs-HEPES buffer were tested for ATPase and AMPase activity. Concentration of enzyme(s) was achieved by reduction of the volume of the superfusate by filtration through Centricon centrifugal filters (Millipore Corporation, Bedford, MA) with membrane pore size cutoff of 30 kDa. Typically, a 2-ml sample was reduced to 40 µl. Enzyme activity was tested using 5, 10, and 20 µl of the concentrated sample, corresponding to theoretical 12.5-, 25-, and 50-fold concentration of proteins with molecular size higher than 30 kDa.

Kinetic Constants of the Releasable Nucleotidase Activity. Under the standard protocol used to study the enzyme activity of releasable nucleotidase, 5 µl of stock solution of eATP, eADP, or eAMP at a given concentration (3 µM-10 mM) was added to a mixture of 20 µl of superfusate collected during neurogenic stimulation of the guinea pig vas deferens (source of enzyme activity) and 25 µl of Krebs-HEPES buffer, pH 7.4. A similar sample was run using superfusate P. After incubation for 60 min at 37°C, the reaction was stopped by addition of 100 µl of ice-cold citrate phosphate buffer, pH 4. Aliquots (100 µl) were injected into the HPLC system for evaluation of the adenine nucleotides and adenosine present in the sample. The enzyme activity was estimated from the difference between the sum of products generated in the presence of superfusate collected during nerve stimulation (S) and the sum of products generated in the presence of superfusate collected under resting conditions (P). This difference (net product) was expressed in picomoles per minute per microliter of superfusate (pmol/min/µl). Normalization by microliters of superfusate was used to replace the conventionally used normalization by protein weight. Kinetic constants (K_m and V_max) were derived from Michaelis-Menten plots using the nonlinear regression analysis of GraphPad Prism software, version 3 (GraphPad Software, San Diego, CA).

Calcium and Magnesium Dependence of Releasable Nucleotidases. Nucleotidases were released in Krebs-HEPES buffer as described above. The collected samples were desalted by centrifugation through a Centricon filter, with a membrane pore cutoff of 30 kDa and then restored to their initial volume with Ca²⁺- and Mg²⁺-free Krebs-HEPES buffer. Because the initial concentration of Ca²⁺ and Mg²⁺ in the superfusing buffer was 1.5 and 1 mM, respectively, the concentration of the cations in the restored samples was estimated to be less than 0.075 and 0.05 mM, respectively. ATPase and AMPase activities were tested under these low Ca²⁺ and Mg²⁺ conditions as well as after supplementation with Ca²⁺ or Mg²⁺ to final concentrations of 1, 2, 3, and 10 mM. ATPase and AMPase activities were also tested in the presence of 5 mM EGTA at pH 7.4.

Chemicals. The following chemicals were purchased from Sigma-Aldrich: adenosine 5'-triphosphate (disodium salt); adenosine 5'-diphosphate (sodium salt); adenosine 5'-monophosphate (sodium salt); adenosine (hemisulfate salt); ,-methylene adenosine 5'-diphosphate (,-mADP); chemiacetal; citric acid; concanavalin A; EGTA; HEPES; levamisole; NaN₃; Na₂PO₄; phosphates inhibitor cocktail II; KH₂PO₄; apyrase VI and VII; 5'-nucleotidase from C. atrox venum; and 3-isobutyl-1-methylxanthine (IBMX), p-nitrophenyl 5'-thymidine monophosphate. Suramin hexasodium ([8-(3-benzamido-4-methylbenzamido) naphthalene-1,3,5-trisulphonic acid]) and ARL67156 were purchased from Sigma/RBI (Natick, MA). Methanol was purchased from B & J (Muskegon, MI). 2-Cl-acetaldehyde was prepared in our laboratory as described previously (Todorov et al., 1996). The Centricon centrifugal filter device (model 30; Amicon Bioseparations) was purchased from Millipore Corporation.

Statistical Analysis. Data were evaluated using GraphPad Prism software, version 3. Means were compared using standard t test, and p < 0.05 was considered to indicate a statistically significant difference.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

Substrate Specificity of Releasable Nucleotidases. eATP, eADP, and eAMP were all metabolized in the presence of superfusate from nerve-stimulated guinea pig vas deferens tissue preparations (S) (Fig. 1b) but remained unaffected by superfusate collected under resting conditions (P) (data not shown). The hydrolysis was time-dependent, sequential, and unidirectional from ATP to ADO. eADP was the first product formed from eATP, whereas eAMP and eADO appeared later in the time course of the reaction. When eADP was used as substrate, eAMP appeared first, and eADO was formed later (Fig. 1Bb). E-ADO was the first and only product during the time course of the metabolism of AMP (Fig. 1Cb). If sufficient time was allowed, eADO was the end product of the reaction regardless of whether eATP, eADP, or eAMP was used as initial substrate.

Increase in Enzyme Activity after Sample Concentration. The volume of samples of superfusate was reduced by filtration through Centricon filters with a membrane pore size cutoff of 30 kDa. The ATPase, ADPase, and AMPase activities were retained above the filter. There was no enzyme activity in the filtrate. In addition, we have observed an increase in specific activity (determined as activity per microliter of superfusate) consistent with concentration of enzyme(s) with size larger than 30 kDa. Dilution of the superfusate, on the other hand, led to a proportional decrease in specific enzyme activity.

Kinetic Constants of ATPase, ADPase, and AMPase Activity. The Michaelis-Menten plots of substrate concentration versus product velocity for all three activities and the Lineweaver-Burk reciprocal plots are shown in Fig. 2. The product velocity increased with the increase of the substrate concentrations (0.3-300 µM) according to a rectangular hyperbola. The K_m value for the ATPase activity, calculated from 28 separate experiments, was 33.6 ± 3.2 µM and the V_max was 0.204 ± 0.003 pmol/min/µl of superfusate (inserted table in Fig. 2). The ADPase activity had a K_m value of 21.0 ± 2.3 µM and V_max value of 0.111 ± 0.002 pmol/min/µl of superfusate (n = 4). The K_m value of the AMPase activity was 10.0 ± 1.1 µM, and the maximal velocity was 0.168 ± 0.003 pmol/min/µl of superfusate (n = 12). These data suggest that the AMPase operates at a slightly slower rate than the ATPase. The ADPase was the slowest of the three, having only one-half and two-thirds of the maximal velocities of the ATPase and the AMPase, respectively. The V_max/K_m ratio revealed that the AMPase had a three-fold higher efficiency than the ATPase or the ADPase. The ratio of the maximal velocities of the ATPase/ADPase was 2:1. ATP is therefore preferred 2-fold over ADP, which is consistent with the possibility that a single ENTPDase-like enzyme is responsible for the metabolism of both ATP and ADP (Plesner, 1995).

View larger version (23K): [in this window] [in a new window]   Fig. 2.   Michaelis-Menten plots and kinetic constants (Km and Vmax) for the neuronal ATPase, ADPase, and AMPase released from the guinea pig vas deferens during nerve stimulation. Increasing concentrations of e-ATP, e-ADP, or e-AMP were used as substrates for the ATPase, ADPase, and AMPase, respectively. The relationships between product velocity (net product) and substrate concentrations fit a rectangular hyperbola for each of the substrates (A). B, Lineweaver-Burk reciprocal plots of the ATPase (eATP used as substrate), ADPase (eADP used as substrate), and the AMPase (eAMP used as substrate) activities. The Km and Vmax values for each of the enzyme reactions are presented in the table. Note that the ratio of maximal velocities of ATPase/ADPase/AMPase approximates 4:2:3.

Calcium and Magnesium Dependence of Nucleotidase Activity. Previously, we have demonstrated that omission of Ca²⁺ from the superfusing solution or addition of cadmium (Cd²⁺) abolishes both the nerve stimulation-evoked release of neurotransmitters and the release of nucleotidase activity from the guinea pig vas deferens. These data have suggested that neuronal nucleotidases are released by a calcium-dependent mechanism (Todorov et al., 1997). To support the release of enzymes we had to maintain calcium in the superfusing solution. Magnesium, however, was not required for neuronal release of ATP, norepinephrine, or soluble enzymes. We therefore carried out release experiments using buffer containing calcium (1.5 mM) but not magnesium ions. The enzyme activity of these samples increased following a rectangular hyperbola when Mg²⁺ was increased from 0.6 to 4.8 mM. Maximal increase of the ADPase activity (12%), ATPase activity (8%), and AMPase activity (4%) was achieved when the Mg²⁺ concentration was increased to 2.4 mM. Further increase of Mg²⁺ concentration to 9.6 mM led to a decline in enzyme activity. Half-maximal activation of the ATPase, ADPase, and AMPase was achieved with 0.7, 1.3, and 0.85 mM Mg²⁺, respectively.

Pharmacological Characteristics of Releasable Neuronal Nucleotidases. The effects of ecto-ATPase inhibitors on the ATPase and AMPase activities are shown in Fig. 3. Suramin inhibited the releasable ATPase activity in a dose-dependent and noncompetitive manner (Fig. 3A) with a K_i value of 53 µM, derived from Dixon plot. On the other hand, suramin had no effect on the AMPase activity (Fig. 3B).

View larger version (30K): [in this window] [in a new window]   Fig. 3.   Inhibitory profiles of suramin (A and B) and ARL67156 (C and D) on releasable ATPase (A, C, and E) and AMPase (B, D, and F) activities. Guinea pig vasa deferentia were stimulated for 30 s with EFS (16 Hz, 0.2 s, 80 V) and samples of the superfusing solution were collected before (P) and during stimulation (S). Samples of superfusate (20 µl) in presence of increasing concentrations (0.3-1000 µM) of eATP in a total incubation volume of 50 µl were incubated at 37°C for 60 min. At the end of the incubation period the reaction was stopped with addition of citric phosphate buffer, pH 4, and the amount of metabolites formed (i.e., eADP, eAMP, and eADO) was analyzed using HPLC coupled with fluorescent detection. The releasable neuronal activity was estimated as the ATPase activity present in superfusate collected during EFS from which the resting activity (activity in sample P) has been subtracted. The velocity of the ATPase reaction was expressed as the amount of product [eADP + eAMP + eADO (in pmol)] formed per minute per microliter of superfusate. Releasable AMPase activity was tested in a similar way using eAMP as substrate and the velocity of the AMPase reaction was evaluated as the amount of eADO formed per minute per microliter of superfusate. Inhibitors were added to the reaction mixture 15 min before addition of substrate. In A, substrate (eATP) versus velocity plot of releasable ATPase in absence and in presence of varying concentrations of suramin (1-100 µM). Suramin inhibits ATPase activity in a noncompetitive manner. As shown in B, suramin has no effect on releasable AMPase activity. ARL67156 inhibited releasable ATPase activity in a noncompetitive manner (C). In E, slope versus inhibitor replots representing the changes of the apparent Km/Vmax ratio of the ATPase reaction as function of increasing ARL67156 concentrations demonstrate a Ki value of 55.8 µM. In D, complex inhibition of AMPase activity by ARL67156 consistent with binding of inhibitor to more than one enzyme binding sites in a noncompetitive way below 3 µM and in a competitive manner above 3 µM. In F, slope versus inhibitor concentration replots, demonstrating a Ki(1) value of 1.7 µM for the high-affinity binding site and a Ki(2) value of 75 µM for the low-affinity binding site of ARL67156 to the releasable AMPase.

View larger version (33K): [in this window] [in a new window]   Fig. 4.   Inhibitory profile of sodium azide, ,-mADP, and levamisole on releasable ATPase (A, C, and E) and AMPase (B, D, and F) activity. Superfusate was collected and enzyme reactions were carried out as described in Fig. 3. Sodium azide (5 and 10 mM) does not affect the releasable ATPase (A) or AMPase (B) activities. ,-mADP (0.1, 1, and 10 µM), a specific inhibitor of ecto-5'-nucleotidase inhibited competitively releasable AMPase activity (D), but had no effect on releasable ATPase activity (C). Levamisole (10 mM), a noncompetitive inhibitor of alkaline phosphatase, had no effect on releasable ATPase (E) or releasable AMPase (F).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

The results reported herein confirm our previous findings that the breakdown of extracellular ATP and consequently the level of extracellular ADO are regulated by soluble nucleotidases, released upon nerve stimulation of the sympathetic nerves of the guinea pig vas deferens (Todorov et al., 1996, 1997; Mihaylova-Todorova et al., 2001). Concentration of superfusate collected during nerve stimulation leads to an increase in nucleotidase activity and suggests that soluble proteins with a size greater than 30 kDa are involved.

In an attempt to more completely understand the nature of the releasable nucleotidases, we have examined the effects of several pharmacological agents that are known to inhibit ecto-phosphatases and ecto-phosphodiesterases. These families of adenine nucleotide-metabolizing enzymes exhibit broad substrate specificity. There is also evidence that ecto-phosphatases may be released from the cell membrane upon activation of endogenous phospholipases and cleavage of the anchoring GPI (Hooper, 1997). We have found, however, that the activity of releasable nucleotidases is not affected by either levamisole, a specific alkaline phosphatase antagonist, or by the phosphatase inhibitor cocktail II (Sigma-Aldrich), which is designed to block the actions of a wide range of phosphatases. These results suggest that alkaline-, acid-, neutral-, or protein-tyrosine phosphatases are not involved in the nerve stimulation-triggered metabolism of adenine nucleotides in the guinea pig vas deferens. Likewise, IBMX, a nonselective phosphodiesterase antagonist, failed to inhibit the ATPase and AMPase activities, thereby excluding the possibility that releasable nucleotidases may share catalytic properties with ecto-phosphodiesterases. Additionally, para-nitrophenyl thimidine monophosphate, a preferred substrate of ENPPases, had no influence on the eATP metabolism by releasable nucleotidases, suggesting that ENPPases do not contribute to the soluble ATPase activity.

Several lines of evidence presented in this study support our previous hypothesis that releasable nucleotidases may share catalytic properties with ecto-ATPases (Kennedy et al., 1997; Todorov et al., 1997; Westfall et al., 2000b). Our results show that like most ecto-ATPases the releasable nucleotidases depend on either Ca²⁺ or Mg²⁺ for activity. Removal of divalent cations by chelation or buffer exchange abolishes the ATPase, ADPase, and AMPase activities. Addition of Ca²⁺ or Mg²⁺, on the other hand, restored these activities. The effects of Ca²⁺ and Mg²⁺ were additive, suggesting that both cations use the same mechanism of activation. Mg²⁺ activated ADPase more than it activated ATPase or AMPase, suggesting that small differences in the metal ion coordination for ATP and ADP may exist. Recently, Chen and Guidotti (2001b) demonstrated that the metal-ATP enzyme complex of the ENDPase1 exists in a single form, whereas the metal-ADP-enzyme complex has two states, one corresponding to the intermediate complex formed during ATP hydrolysis and the second corresponding to the ADP binding as substrate for further hydrolysis.

Knowles and Nagy (1999) have shown a nucleotide-substrate-dependent inhibition effect of sodium azide on the chicken oviduct ecto-ATPDase. The effect of azide was prominent on the ADP hydrolysis, whereas the ATP hydrolysis was less influenced. Herein, we show that the soluble ATPase activity was not affected by sodium azide, but the rate of the ADPase activity was decreased. This is consistent with the hypothesis that the enzymes released upon nerve stimulation of the guinea pig vas deferens include an ATPDase.

The K_m value of the releasable ATPase (33 µM) determined in this work seems similar to the K_m values of ecto-ATPases purified from synaptosomes of the rat and mouse cortex (K_m = 39-53 µM; Nagy et al., 1986), but was lower than that of ATPase from rat striatal cholinergic synapse (K_m = 131 µM; James and Richardson, 1993). The expressed recombinant ENTPDases, however, show K_m values for ATP ranging from 75 to 400 µM that are in general higher than the ones found in tissues. ENTPDase1 and 3 use ADP as well as ATP as substrates, whereas ENTPDase2 is mostly an ATPase because it prefers ATP 10 times more than it prefers ADP. ENTDPase1 metabolizes ATP directly to AMP, and ADP does not appear as free product in the reaction. The guinea pig vas deferens neuronal nucleotidases generate ADP as a free product resulting from the ATP hydrolysis. ADP is detected in the reaction before it is hydrolyzed further to AMP. The velocity of ATP hydrolysis is 2-fold higher than that of ADP. The soluble ATPase released from the sympathetic nerves of the guinea pig vas deferens is therefore a triphosphate-diphosphohydrolase-like enzyme. Because ADP accumulates as a product of the metabolism of ATP, it seems that the releasable ATPase is more similar to ENTPDase3, which prefers ATP 3 to 4 times more than ADP (Smith and Kirley, 1998) than it is to ENTPDase1, which does not discriminate between ATP and ADP (Wang and Guidotti, 1996). On the other hand, recent work of Chen and Guidotti (2001a) demonstrated that the velocity of the ATP hydrolysis and the preference for ADP binding are functions of the oligomerization state of the ENTPDase monomers. Disruption of the tetrameric organization of the membrane-bound rat ENTPDase1 by either detergent solubilization or truncation of the N- and C-terminal segments of the recombinant protein decreased its preference for ADP 3.5 times and ADP appeared in solution during the metabolism of ATP. The K_m value for ATP was also decreased into the low micromolar range. It seems, therefore that the kinetics of the releasable ATPase from the guinea pig vas deferens reported in the current work approximates the kinetics of the C-terminal- and N-terminal-truncated rat ENTPDases. This opens the possibility that the releasable ATPases could in fact represent a proteolytic cleavage form of the membrane-bound ENTPDases.

The activity of ecto-ATPase in neuronal (Marti et al., 1996) and non-neuronal tissues (Bultmann et al., 1996) is inhibited by suramin. Suramin has also been shown to inhibit the ATPase activity of the Chinese hamster ovary cells transfected with the rat brain ecto-ATPase gene, but not that of Chinese hamster ovary cells transfected with the ecto-apyrase gene (Heine et al., 1999). Herein, we show that the neuronal ATPase activity released from the guinea pig vas deferens was inhibited by suramin in a noncompetitive manner and with a K_i value of 53 µM. At the same time, the breakdown of AMP remained unaffected.

Our results demonstrate that unlike suramin, which inhibits only the ATPase, ,-mADP and Con A inhibit only the metabolism of AMP and have no effect on the metabolism of ATP by releasable nucleotidases. These data suggest that the binding and hydrolysis of triphosphonucleotides is independent of the binding and catalytic hydrolysis of 5'-monophosphonucleotides. One possible explanation for these findings is that the sympathetic nerves of the guinea pig vas deferens release a separate enzyme that is pharmacologically similar to the members of the family of ecto-5'-nucleotidases. The apparent K_m value of the releasable AMPase activity reported herein (10 µM) is comparable with the K_m value of ecto-5'-nucleotidases isolated from electric organ of Torpedo (25 µM) and bovine cerebral cortex (46 µM) (Vogel et al., 1992). Ecto-5'-nucleotidase, a GPI-anchored protein, may become soluble upon activation of phospholipases. Previously, we have shown that the AMPase and the ATPase activities are coreleased with neurotransmitter ATP (Mihaylova-Todorova et al., 2001). It is not known at this time whether the events leading to exocytosis and release of neurotransmitters may concomitantly release or cleave ecto-enzymes from cell membranes.

ARL67156 inhibits the activity of ecto-ATPases expressed by blood (Crack et al., 1995) and smooth muscle cells (Westfall et al., 1996) as well as the ATPase activity of the releasable nucleotidases from guinea pig (Kennedy et al., 1997; Todorov et al., 1997; Westfall et al., 2000b) and rabbit vas deferens (Westfall et al., 2000a). Herein, we confirm our previous results that ARL67156 inhibits the ATPase activity and determined that the mode of inhibition is noncompetitive with a K_i value of 55.8 µM. In addition, we show that ARL67156 inhibits AMPase activity in a complex manner. Binding of ARL67156 to a high-affinity binding site (K_i = 1.7 µM) contributes to a noncompetitive inhibition of AMPase activity. At a second, low-affinity binding site (K_i = 75 µM), ARL67156 seems to inhibit AMPase activity in a competitive manner. It has been previously reported that ARL67156 inhibits purified human placental 5'-nucleotidase (Chayet et al., 1997) but the mode of inhibition was not tested. It will be interesting to determine whether inhibition of ecto-5'-nucleotidase activity is a common property of ARL67156 or whether this mode of inhibition is specific for the releasable AMPase. Based on the evidence provided by this study we favor the notion that at least two enzymes, an ATPDase and an AMPase that work cooperatively to break down extracellular ATP to ADO, are coreleased from the sympathetic nerves of the guinea pig vas deferens. Pharmacological analysis suggests that the ATPDase behaves as an ENTPDase (possibly C- and N-terminal truncated), whereas the AMPase closely resembles ecto-5'-nucleotidase. However, an alternative hypothesis that a single protein carries both ATPDase and AMPase activity cannot be rejected at this time. Enzymes showing functional characteristics of apyrases and amino acid sequence similarity with 5'-nucleotidases have been identified. For example, a soluble apyrase secreted from the salivary gland of adult female mosquito Aedes aegypti seems to share sequence similarity with vertebrate and bacterial 5'-nucleotidases (Champagne et al., 1995). It has to be noted, however, that this enzyme does not metabolize AMP. If the releasable nucleotidase described herein is a single enzyme then it should have at least two separate active centers that may be independently modulated by ecto-ATPase- and ecto-5'-nucleotidase inhibitors.