Purification, characterization, cloning, and expression of the chicken liver ecto-ATP-diphosphohydrolase Aileen F. Knowles 1 , Agnes K. Nagy 2 , Randy S. Strobel 3 and Mae Wu-Weis 1 1 Department of Chemistry, San Diego State University, San Diego, CA, USA; 2 West Los Angeles Veterans Affairs Medical Center, Los Angeles, CA, USA; 3 Department of Natural Sciences, Metropolitan State University, St Paul, MN, USA We previously demonstrated that the major ecto-nucleoside triphosphate phosphohydrolase in the chicken liver membranes is an ecto-ATP-diphosphohydrolase (ecto- ATPDase) [Caldwell, C., Davis, M.D. & Knowles, A.F. (1999) Arch. Biochem. Biophys. 362, 46–58]. Enzymatic properties of the liver membrane ecto-ATPDase are similar to those of the chicken oviduct ecto-ATPDase that we have previously purified and cloned. Using antibody developed against the latter, we have purified the chicken liver ecto-ATPDase to homogeneity. The purified enzyme is a glycoprotein with a molecular mass of 85 kDa and a specific activity of  1000 UÆmg protein )1 . Although slightly larger than the 80-kDa oviduct enzyme, the two ecto-ATPDases are nearly identical with respect to their enzymatic properties and mass of the deglycosylated proteins. The primary sequence of the liver ecto-ATPDase deduced from its cDNA obtained by RT-PCR cloning also shows only minor differences from that of the oviduct ecto-ATPDase. Immunochemical staining demonstrates the distribution of the ecto-ATPDase in the bile canaliculi of the chicken liver. HeLa cells transfected with the chicken liver ecto-ATPDase cDNA express an ecto-nucleotidase activity with character- istics similar to the enzyme in its native membranes, most significant of these is stimulation of the ATPDase activity by detergents, which inhibits other members of the ecto- nucleoside triphosphate diphosphohydrolase (E-NTPDase) family. The stimulation of the expressed liver ecto-ATPDase by detergents indicates that this property is intrinsic to the enzyme protein, and cannot be attributed to the lipid environment of the native membranes. The molecular identification and expression of a liver ecto-ATPDase, reported here for the first time, will facilitate future investigations into the differences between structure and function of the different E-NTPDases, existence of liver ecto-ATPDase isoforms in different species, its alteration in pathogenic conditions, and its physiological function. Keywords: ecto-ATP-diphosphohydrolase; chicken liver; E-NTPDase; expression; immunoaffinity purification. E(cto)-ATPases (E-ATPases; also known as E-NTPDases) (EC 3.6.1.5) are ubiquitous cell surface glycoproteins that hydrolyze nucleoside triphosphates. Some will also hydro- lyze nucleoside diphosphates. Their physiological substrates are probably the ligands of purinergic receptors, e.g. extracellular ATP, ADP, and UTP [1]. They may also play a role in regulating substrate concentration of ecto-protein kinases [2]. A substantial literature on the characterization of the E-ATPases in intact cells and plasma membrane preparations has accumulated since the 1970s (reviewed in [3]). Because of their low abundance and the lability of some E-ATPases to detergents, only the E-ATPases of rabbit muscle transverse tubules [4], chicken gizzard [5], human placenta [6], and chicken oviduct [7] have been purified to homogeneity. On the other hand, the cDNA sequences of more than two dozen related E-ATPases and soluble E-type ATPases have been reported, establishing an E-ATPase gene family [8]. The cDNAs of membrane-bound E-ATPases encode proteins of  500 amino acids. The bulk of the E-ATPase protein is extracellular with two transmembranous domains near the N- and C-termini. Variable numbers of potential N-glycosylation sites and protein kinase consensus motifs occur in the sequences. More importantly, all contain five highly conserved apyrase consensus regions [9,10] and 10 conserved cysteine residues, the latter are probably involved in disulfide bond formation. The E-ATPases can be divided into two groups based on their substrate selectivity and inhibition by azide. The ecto-ATP-diphosphohydrolases (ecto-ATPDases or ecto-apyrases) hydrolyze NDPs as well as NTPs and are inhibited by high concentrations of azide, whereas the ecto- ATPases show little activity toward NDPs and are not inhibited by azide. The ecto-ATPDases are comprised of different isoforms. The majority of the ecto-ATPDases that have been cloned are closely related to CD39, a cell surface antigen that is expressed on activated lymphocytes [11,12]. CD39s from several species have 60–90% identity in their primary sequences [12–15]. Biochemical and Correspondence to A. F. Knowles, Department of Chemistry, San Diego State University, San Diego, CA 92182-1030, USA. Fax: + 1619 594 4634, Tel.: + 1619 594 2065, E-mail: aknowles@chemistry.sdsu.edu Abbreviations: DMEM, Dulbecco’s modified Eagle’s medium. Definitions: E-ATPases are a family of cell surface (ecto) ATPases that hydrolyze extracellular ATP; they are also known as the E-NTPDases. Ecto-ATP-diphosphohydrolase (ecto-ATPDases) and ecto-ATPases are two different subfamilies of the E-ATPases. E-type ATPases are ATPases that have similar enzymatic characteristics and sequence homology to the E-ATPases, however, they are not membrane proteins. (Received 17 December 2001, revised 18 March 2002, accepted 22 March 2002) Eur. J. Biochem. 269, 2373–2382 (2002) Ó FEBS 2002 doi:10.1046/j.1432-1033.2002.02898.x immunolocalization studies indicated that CD39s are vas- cular ecto-ATPDases [16–20]. Two other ecto-ATPDases, cloned from chicken oviduct [21] and human brain [22], can be distinguished from the CD39/ecto-ATPDases because of significant sequence divergence from the latter. Interest- ingly, these two ecto-ATPDases have only 44% identity in their primary sequences. This cannot be entirely accounted for by species differences because the chicken and human ecto-ATPases have 57% sequence identity. Thus, the true relationship of the chicken and human ecto-ATPDases remains to be established. We previously reported the immunoaffinity purification of the chicken oviduct ecto-ATPDase to homogeneity [7] and its molecular cloning [21]. An ecto-ATPDase similar to the chicken oviduct enzyme is present in the chicken liver membranes [23]. We report here the complete purification of the chicken liver ecto-ATPDase, its enzymatic properties and localization, cloning of the full-length cDNA and its expression. EXPERIMENTAL PROCEDURES Materials The production and characterization of monoclonal anti- bodies against the chicken oviduct ecto-ATPDase were described previously [7]. Of the six monoclonal antibodies generated, MC18 was most suitable for Western blot analysis because of its strong and specific binding to the chicken ecto-ATPDase [7,23]. MC22 was employed to prepare the immunoaffinity column using hydrazide acti- vated Affi-gel [7] and MC27 was used for immunolocaliza- tion. Goat anti-(mouse IgG) IgG conjugated to alkaline phosphatase was purchased from Promega. Chicken liver polyA + RNA, Marathon TM cDNA amplification kit and Advantage 2 PCR enzyme system were purchased from Clontech. Pfu Turbo DNA polymerase was purchased from Strategene. Dulbecco’s modified Eagle’s media (DMEM), OptiMEM, fetal bovine serum, Lipofectamine, and genet- icin were purchased from Life Technologies Inc. N-Glyco- sidase F and restriction enzymes were purchased from New England Biolabs. ATP, ADP, and all other biochemical reagents were purchased from Sigma Chemical Co. Oligo- nucleotides used as primers for PCR and sequencing were synthesized at the San Diego State University Microchemi- cal Core Facility. Purification of chicken liver ecto-ATPDase Liver from freshly killed chickens was purchased at a local poultry farm. Membranes were prepared by homogenizing 1 lb of livers in 500 mL of isolation buffer (50 m M Tris/ HCl, pH 7.4, 0.25 M sucrose, and 1 m M EGTA) in a Waring blender for 1 min. After filtering the homogenate through cheesecloth, the filtrate was homogenized again in a Dounce homogenizer and then centrifuged at 5000 r.p.m. in an SS34 rotor in a Sorvall centrifuge for 10 min. The supernatant was centrifuged again at 16 000 r.p.m. for 20 min to precipitate the membranes. The membrane pellet was washed three times by repeated resuspension in a buffer containing 50 m M Tris/HCl, pH 7.4 and 1 m M EGTA and centrifugation. To extract the ecto-ATPDase, the mem- branes were solubilized in 50 m M Tris/HCl, pH 7.4 containing 5% NP-40 at  2mgproteinÆmL )1 and stirring at 4 °C overnight. After centrifugation at 16 000 r.p.m. for 20 min, the supernatant was filtered through Whatman no. 1 filter paper. The filtrate containing the extracted membrane proteins were applied to a DEAE Biogel A column (2.5 · 43 cm) pre-equilibrated with 50 m M Tris/HCl, pH 7.4 containing 0.1% NP-40 (chromatography buffer). After washing the column to elute the unbound proteins, the column was developed with a NaCl gradient consisting of 375 mL of chromatography buffer and 375 mL of chromatography buffer containing 1 M NaCl. The ATPase activity was eluted as a broad peak while more than 90% of the solubilized proteins remained bound to the column. Total recovery of activity was nearly 100%, partly attributable to the activating effect of NP-40 of liver membrane ecto- ATPDase [23]. The proteins eluted from the DEAE Biogel A column were applied to a ConA–Sepharose 4B column (1.5 · 10 cm) pre-equilibrated with chromatography buffer. After washing off the unbound proteins, the chicken liver ecto-ATPDase and other bound glycoproteins were eluted with 200 mL of chromatography buffer containing 1% a-methylmannoside. Fractions containing ATPase activity were pooled and applied to a second DEAE Biogel A column (1.5 · 10 cm) as a means of concentrating the proteins. The bound enzyme was eluted in a small volume of chromatography buffer containing 1 M NaCl. Fractions containing ATPase activity were pooled and desalted on Sephadex G-25 column (1.5 · 48 cm). The desalted and partially purified chicken liver ecto- ATPDase fraction ( 20 mL) was added to 3 mL of MC22-hydrazide resin [7], and equilibrated overnight by rocking at 4 °C. The slurry was poured into a small column, and washed sequentially with 50 mL each of chromatogra- phy buffer, buffer containing 0.5 M NaCl, and buffer again. The enzyme was eluted with 50 m M glycine, pH 2.5 containing 0.1% NP-40. Fractions of 1 mL were collected into tubes containing 0.1 mL of 1 M Tris/HCl, pH 8.0 plus 0.1% NP-40. The emergence of an 85-kDa protein coincided with elution of the ATPase activity (not shown). ATPase assays During the purification procedure, ATPase activities of the chicken liver ecto-ATPDase were assayed in 0.5 mL reac- tion mixture containing 50 m M Tris/HCl, pH 7.4, 0.1% NP-40, 4 m M MgCl 2 and 4 m M ATP at 37 °C for 5–30 min. After terminating the reactions by the addition of 0.1 mL 10% trichloroacetic acid, the denatured proteins were removed by centrifugation. Aliquots of the supernatant were used for determination of P i released by the AAM reagent (10 m M ammonium molybdate/5N H 2 SO 4 /acetone, 1 : 1 : 2, v/v/v) [7]. Absorbance of the phosphomolybdate complex was read at 355 nm. For characterizing the enzymatic properties of the purified chicken liver ecto-ATPDase, enzyme assays were carried out in 0.25-mL reaction mixtures using the buffer systems and substrate concentrations indicated in the legends. Phosphate released was determined by the mala- chite green reagent as described previously [7]. ATP and ADP hydrolysis activities of intact COS or HeLa cells transfected by chicken liver ecto-ATPDase 2374 A. F. Knowles et al. (Eur. J. Biochem. 269) Ó FEBS 2002 cDNA were determined using either attached cells in six- well plates or cell suspension obtained after trypsinization. For the six-well plates, cells were washed twice with 1 mL of buffered isotonic solution (0.1 M NaCl, 0.01 M KCl, and 25 m M Tris/HCl, pH 7.4) after aspiration of the culture media. The cells were then overlaid with 1 mL buffered isotonic solution containing 5 m M MgCl 2 and 5 m M ATP or ADP. After incubation at 37 °C for 15–30 min, the reaction mixture was collected by Pasteur pipettes and added to 0.1 mL 10% trichloroacetic acid. Aliquots (0.1– 0.4 mL) of the solution were used for determination of phosphate released using the AAM reagent. Alternatively, cells grown in 10-cm plates were trypsinized, suspended in culture media, and collected by centrifugation. After washing with buffered isotonic solution, the cells were resuspended in the same solution at 1–3 mg proteinÆmL )1 . Aliquots of cell suspension (50–100 lg cell protein) were used for enzyme assays in 0.5 mL isotonic reaction mixture with substrates as described above. The reaction was carried out for 10–30 min at 37 °C and stopped by the addition of 0.1 mL 10% trichloroacetic acid. The suspension was centrifuged to remove denatured proteins. Aliquots of the supernatant solution were used for P i determination by the AAM reagent. RT-PCR cloning Chicken liver polyA-RNA (1 lg) was reverse transcribed using avian myoblastosis virus reverse transcriptase and oligo(dT) as the primer (Marathon TM cDNA amplification kit, Clontech). Double-stranded cDNA was prepared according to the manufacturer’s instruction and served as template in the PCR. Oligonucleotides corresponding to the 5¢ and 3¢ ends of the chicken oviduct cDNA (GenBank accession no. AF041355) were used: forward primer, 5¢-ATGGAGTATAAGGGGAAGGTTGTTGC-3¢,and reverse primer, 5¢-TTGGATTTCCAGAAACACTGGA-3¢. PCR was carried out in a 50-lL reaction mixture containing 0.5 lL (from a total of 10 lL) cDNA template, 0.5 l M primers, 0.1 m M dNTP, and 1.25 U Pfu Turbo DNA polymerase (Strategene). Thermal cycling on a PTC-200 Peltier thermal cycler (MJ Research, Waltham, MA, USA) began with 4.5 min at 94 °C followed by 35 cycles of 94 °C for 45 s, 55 °Cfor1minand72°C for 3 min and ending with 10 min at 72 °C. A PCR product of  1.5 kb that corresponded to the length of the coding region of the chicken ecto-ATPDase was obtained. An aliquot (2 lL) of the reaction mixture was used for further amplification with the same primers and thermal cycling conditions but with Advantage 2 Taq DNA polymerase (Clontech) for subse- quent TA cloning. The 1.5-kb PCR product was gel purified and ligated to pCR2.1 (Invitrogen) and an aliquot of the ligation mixture was used to transform INVaF Escherichia coli cells (Invitrogen). White colonies that grew on agar containing ampicillin were selected and the presence of 1.5-kb insert was verified by digestion with EcoRI. DNA from one recombinant plasmid was isolated and digested with EcoRI. The resultant fragment was ligated with the mammalian expression vector pcDNA3 (Invitrogen) linear- ized with EcoRI and treated with bovine pancreatic alkaline phosphatase. An aliquot of the ligation mixture was used to transform DH5a E. coli cells. Orientation of the insert in the recombinant pcDNA3 was determined with appropriate restriction enzymes. One clone (pcDNA3-CL8) with the correct orientation was propagated for preparation of DNA for sequencing and transfection. DNA sequencing was provided by the San Diego State University Microchemical Core Facility. Transient and stable transfection COS-7 and HeLa cells were grown in DMEM containing 10% fetal bovine serum, penicillin (100 UÆmL )1 )and streptomycin (100 lgÆmL )1 ). Cells were plated either in six-well plates or 10-cm plates and were used for transfection after reaching 50–70% confluence. In the six-well plates, the cells were washed twice with OptiMEM and then layered with 1 mL OptiMEM containing DNA (1 lg per well) and Lipofectamine (5 lL per well) which had been premixed and incubated according to the manufacturer’s instruction. After 5 h, 1 mL of DMEM containing 20% fetal bovine serum was added to the wells. Twenty-four hours after transfection, the medium was replaced by fresh DMEM/ 10% fetal bovine serum. ATPase and ADPase activities were determined 48–72 h after transfection. When transfec- tion was carried out in 10-cm plates, the cells were overlaid with 6.4 mL OptiMEM containing premixed DNA (5 lg) and Lipofectamine (30 lL). After 5 h, DMEM and serum were added to bring the volume of the media to 10 mL and a final serum concentration to 10%. After 24 h, the media were replaced by fresh DMEM containing 10% serum. After another 24–48 h, the cells were harvested by trypsi- nization. Enzyme activities were determined using 25–50 lL cell suspension (50–100 lgcellprotein). For stable transfection, HeLa cells were first transfected in 10-cm plates with the chicken liver cDNA (in pcDNA3). Two days after transfection, the cells were harvested and divided into two T-25 flasks. The cells were allowed to attach overnight and geneticin was added at 400 lgÆmL )1 . Medium was replaced every three days. The established geneticin-resistant clones were propagated for activity determination and characterization. Deglycosylation Purified chicken liver ecto-ATPDase (2 lg) was precipitated by the addition of 9 vol. of ice-cold acetone. After centrifugation for 30 min, the protein pellet was dissolved in 0.5% SDS/2% 2-mercaptoethanol and heated for 10 min at 100 °C. Phosphate buffer (pH 7.5) and NP-40 were added to a final concentration of 50 m M and 1%, respect- ively. The protein solution was incubated with 1000 U of N-glycosidase F at 37 °C for 16 h. Aliquots of the protein solution were used for Western blot analysis. Gel electrophoresis and Western blot analysis SDS/PAGE was carried out on a mini-gel apparatus (Bio- Rad) in slab gels of 7.5% acrylamide according to Laemmli [24]. The gel was stained with silver nitrate [7]. For Western blot analysis, protein samples were mixed with sample buffer containing SDS but without reducing agents since the epitope recognized by MC18 is destroyed by reduction of disulfide bonds. Separated proteins were transferred to poly(vinylidene difluoride) membranes. The membranes were first blocked with NaCl/Tris (0.5 M NaCl and 20 m M Ó FEBS 2002 Molecular identification of a liver ecto-ATPDase (Eur. J. Biochem. 269) 2375 Tris/HCl, pH 7.4) with 2% BSA and then incubated for 2 h with the monoclonal antibody MC18 diluted in NaCl/Tris/ 2% BSA. After washing four times with NaCl/Tris containing 0.1% Tween 20, the membranes were incubated in a solution containing goat anti-(mouse IgG) Ig conju- gated to alkaline phosphatase for 1.5 h. After three washes with NaCl/Tris containing 0.1% Tween-20 and a final wash with NaCl/Tris, the immunoreative bands were detected by treating the membranes with the alkaline phosphatase substrates (Bio-Rad). Immunolocalization Paraffin-embedded sections of chicken liver fixed in 10% buffered formalin were used. Immunoperoxidase localiza- tion of ecto-ATPDase in these sections were performed with purified monoclonal antibody, MC27, using a standard immunohistochemical protocol [25]. RESULTS Purification of chicken liver ecto-ATPDase Chicken liver ecto-ATPDase, an integral membrane pro- tein, was solubilized from liver membranes by NP-40 and purified by ion-exchange, lectin-affinity, and immunoaffinity chromatographic separations as described in Experimental procedures. The purification scheme is summarized in Table 1. The most effective purification was obtained by immunoaffinity chromatography with a  50-fold purifica- tion in one single step. The significant loss of total activity probably resulted from irreversible binding of the majority of the liver ecto-ATPDase to MC22. Total purification was 2000-fold. The final preparation had a specific ATPase activity of  1200 lmolÆmin )1 Æmg )1 protein, similar to that of the purified chicken oviduct ecto-ATPDase ( 800 lmolÆmin )1 Æmg )1 protein) [7]. The purified enzyme is stable indefinitely at 4 °C in buffer containing 0.1% NP-40, but suffers significant loss of activity upon freezing. The purified chicken liver ecto-ATPDase contains a single protein band (Fig. 1A, lane 2) with an apparent molecular mass of 85 kDa, which is slightly higher than the molecular mass of the purified chicken oviduct ecto- ATPDase, 80 kDa (Fig. 1A, lane 1). In Western blot analysis with MC 18, an 85-kDa protein was also detected in both the chicken liver membranes (Fig. 1B, lane 3) and the purified ecto-ATPDase preparation (Fig. 1B, lane 1). Because the epitope detected by MC18 was sensitive to disulfide reduction, the samples used for Western blot analysis were not treated with 2-mercaptoethanol. Under these circumstances, a higher molecular mass band of  180 kDa was often detected in the protein sample (Fig. 1B, lane 1). This result suggests that the enzyme is able to form dimers. When the chicken liver ecto-ATPDase was treated with N-glycosidase F which removes N-linked oligosaccharides, Western blot analysis with MC18 revealed the presence of two protein bands with molecular masses of  55 and  100 kDa (Fig. 1B, lane 2). The molecular masses of the deglycosylated chicken liver ecto-ATPDase monomer is the same as the deglycosylated chicken oviduct ecto-ATPDase that we previously reported [21]. Thus, the slightly higher molecular mass of the native liver ecto- ATPDase indicates that its glycosylation is more extensive than that of the oviduct enzyme. Table 1. Purification of ecto-ATPDase from chicken liver. ND, not determined. Fraction Volume (mL) Total protein (mg) Total activity (lmolÆmin )1 ) Specific activity (lmolÆmin )1 Æmg )1 ) Yield (%) Purification (fold) Liver membranes 525 10200 6608 0.638 100 1 DEAE-Biogel A 490 784 6419 8.18 99 13 ConA-Sepharose 205 ND 3280 ND 50 ND Sephadex G-25 28.7 112 2944 26.3 45 41 M22-hydrazide gel 6.6 0.017 21.3 1242 0.32 1947 Fig. 1. Molecular masses of native and deglycosylated purified chicken liver ecto-ATPDase. (A) Purified chicken liver oviduct ecto-ATPDase (0.3 lg) and liver ecto-ATPDase (0.1 lg) were dissolved in SDS gel sample buffer containing 2-mercaptoethanol and applied to a 7.5% polyacrylamide gel. After electrophoresis, the gel was silver stained. Lane 1, chicken oviduct ectoATPDase; lane 2, chicken liver ecto- ATPDase. (B) Western blot analysis of purified chicken liver ecto- ATPDase before and after deglycosylation. Purified chicken liver ecto-ATPDase (2 lg) was treated with N-glycosidase F as described in Experimental procedures. An aliquot of the protein solution was treated with SDS gel sample buffer without 2-mercaptoethanol. After electrophoresis on 7.5% polyacrylamide gel and transfer to poly(vinylidene difluoride) membrane, the membrane was probed with monoclonal antibody MC18. Lane 1, purified chicken liver ecto- ATPDase; lane 2, deglycosylated chicken liver ecto-ATPDase; lane 3, chicken liver plasma membranes. 2376 A. F. Knowles et al. (Eur. J. Biochem. 269) Ó FEBS 2002 Enzymatic characteristics of the purified chicken liver ecto-ATPDase Most E-ATPases have broad substrate specificity with respect to nucleoside triphosphates. Ecto-ATPDases and CD39 also hydrolyze nucleoside diphosphates. The K m values obtained for ATP of the purified chicken liver ecto- ATPDase is 0.51 m M and that for ADP is 5.3 m M in the presence of 5 m M MgCl 2 at pH 7.4. Lower K m values for ATP and ADP, 0.13 m M and 0.72 m M , respectively, were obtained at pH 6.4. Unlike some other E-ATPases that exhibit similar or higher ATP hydrolysis activity in the presence of Ca 2+ [26], the CaATPase activity of the purified chicken liver ecto-ATPDase is  30% of the MgATPase activity at a divalent ion-ATP concentration of 5 m M at pH 7.4. On the other hand, the CaADPase activity is  80% of the MgADPase activity at a divalent ion-ADP concentration of 5 m M (data not shown). The ATPase and ADPase activities of the purified chicken liver ecto-ATPDase were affected differently by pH. Figure 2 shows that the pH–activity curves of ATP and ADP hydrolysis do not coincide. While maximal ATP hydrolysis activity was obtained in the pH range of 7.5–8.5, the pH optima for ADP was lower at 6.0–6.5. Thus, the ADPase/ATPase ratios vary significantly at different pH values. For example, a higher ADPase/ATPase ratio was obtained at pH 6.4 ( 0.5) than at pH 8.0 ( 0.1). Because of the lower K m for ADP and higher ADPase activity observed at pH 6–6.5, ADPase activities were determined at pH 6.4 in several of the experiments reported below. We showed previously that, in contrast to the chicken smooth muscle ecto-ATPase, which is inactivated by most detergents, the chicken liver plasma membrane ecto- ATPDase activity is increased by Triton X-100 and NP-40 [23,27]. As described above, the enzyme was extracted from the membranes by 5% NP-40, and all solutions used in its purification contained 0.1% NP-40. Table 2 shows that the purified chicken liver ecto-ATPDase was not affected by ConA or suramin, the former activates while the latter inhibits the chicken smooth muscle ecto-ATPase [23,28]. On the other hand, it was inhibited by high concentrations of azide, an inhibitor of CD39 and most ecto-ATPDases [29–31], and high concentrations of fluoride, vanadate, and pyrophosphate, inhibitors of the purified chicken oviduct ecto-ATPDase [7,31]. Like the other ecto-ATPDases, the ADPase activity of the purified chicken liver ecto-ATPDase was more sensitive to azide inhibition than its ATPase activity at either pH 7.4 or pH 6.4 (Fig. 3). At pH 7.4, 5m M azide inhibited ADP hydrolysis by  70% whereas ATP hydrolysis was inhibited by only 10%. Inhibition of ADP and ATP hydrolysis by azide was greater when the pH of assay solutions was 6.4. Azide inhibition was diminished if Ca 2+ wasusedinplaceofMg 2+ (data not shown). Inhibition of the enzyme by fluoride and vanadate was also more pronounced with ADP as the substrate and was greater at lower pH values. In contrast to azide, fluoride, and vanadate, 5 m M pyrophosphate inhibited ATP hydro- lysis significantly ( 50%), and the extent of inhibition was insensitive to pH. This difference could be the result of different modes of inhibition by these compounds. Pyro- phosphate was previously shown to be a competitive inhibitor of the oviduct ADPase activity [32], whereas Fig. 2. Effect of pH on ATP and ADP hydrolysis by the purified chicken liver ecto-ATPDase. Enzyme assays were carried out in 0.25-mL reaction mixtures using a wide-range buffer system (piperazine dihydrochloride/glycylglycine/NaOH [31]), covering the pH range of 4.6–9.5. The assay solutions contained 50 m M buffer, 0.1% NP-40, 5m M MgATP (d)or10m M MgADP (m) with 0.2–0.4 lgchicken liver ecto-ATPDase. The reaction was carried out for 5 min at 37 °C. Table 2. Effect of modulators on ATP and ADP hydrolysis of the purified chicken liver ecto-ATPDase. Purified chicken liver ecto-ATPDase (0.2– 0.4 lg) was preincubated with the indicated concentrations of modulators in a 0.25-mL reaction mixture at 37 °C for 5 min before initiating the reaction by the addition of ATP or ADP. Buffers were 50 m M Tris/HCl for pH 7.4 and 50 m M Mops for pH 6.4. Reaction time was 5 min. Duplicate samples were run for each condition. Results presented were the average (± SD) of three experiments. Values are given as percent activity. Addition ATP hydrolysis ADP hydrolysis pH 7.4 pH 6.4 pH 7.4 pH 6.4 ConA (50 lgÆmL )1 ) 102 ± 2.9 92 ± 1.6 110 ± 13.3 91 ± 2.9 Suramin (0.1 m M ) 106 ± 7.9 103 ± 3.6 112 ± 1.6 105 ± 8.2 Azide (10 m M ) 79 ± 3.6 57 ± 2.6 15 ± 0.4 8 ± 0.8 Fluoride (10 m M ) 98 ± 2.9 88 ± 1.2 9 ± 2.2 3 ± 0.8 Vanadate (1 m M ) 93 ± 2.3 78 ± 1.1 46 ± 3.4 12 ± 2.1 Pyrophosphate (5 m M ) 49 ± 1.9 56 ± 3.3 3 ± 0.4 7 ± 2.5 Ó FEBS 2002 Molecular identification of a liver ecto-ATPDase (Eur. J. Biochem. 269) 2377 inhibition by azide was of the mixed and uncompetitive type [31]. The mechanism of inhibition of fluoride and vanadate has not been investigated, but they are unlikely to be competitive inhibitors. Immunolocalization of the chicken liver ecto-ATPDase When thin sections of chicken liver were stained with the chicken ecto-ATPDase monoclonal antibody, MC27, the protein could be seen to be distributed at the bile canaliculi (Fig. 4). This localization of the ecto-ATPDase agrees with previous finding of distribution of cell surface ATPase activity in rat liver as determined by cytochemical staining [33–36]. Besides oviduct and liver, the chicken ecto- ATPDase is also present in the apical membranes of the oxyntic-peptic cells [37]. The distribution of the ecto- ATPDase on these epithelial cells is distinctly different from theotherATPDaseintheE-ATPasefamily,theCD39s [13,17,19]. Molecular cloning of chicken liver ecto-ATPDase The results described above indicate that: (a) the enzymatic properties of the chicken liver ecto-ATPDase are similar to that of the oviduct ecto-ATPDase; (b) the chicken liver ecto- ATPDase binds strongly to the monoclonal antibodies of the oviduct ecto-ATPDase, as well as an antibody [21] developed against the N-terminus of the chicken oviduct ecto-ATPDase (data not shown); and (c) the deglycosylated chicken liver and oviduct ecto-ATPDases have the same molecular mass, i.e.  55 kDa [21]. It seems likely that the two enzymes may have similar primary sequences despite the different molecular masses of the native enzymes. We decided to obtain the cDNA of the chicken liver ecto- ATPDase using RT-PCR starting with chicken liver polyA + RNA. Under the appropriate PCR conditions described in Experimental procedures, a 1.5-kb PCR product was obtained, which was introduced into the cloning vector, pCR2.1, by TA cloning. Three separate clones were sequenced using T7 promotor, M13, and gene specific primers, all yielding the same nucleotide sequence (Fig. 5, GenBank accession no. AF426405). The deduced primary sequence of the chicken liver ecto-ATPDase (Fig. 5) is nearly identical to that of the oviduct ecto- ATPDase differing in seven amino acids out of 493 amino acids. It has the same two transmembranous domains, five apyrase conserved regions (ACRs), 10 conserved cysteine residues, and 12 potential N-glycosylation sites as the oviduct ecto-ATPDase [21]. Northern blot analysis using total chicken liver RNA revealed a major transcript of  2 kb (data not shown). Expression of the chicken liver ecto-ATPDase cDNA To verify that the cDNA we obtained encodes an ecto- ATPDase, we carried out transient transfection in both COS-7 and HeLa cells. Both host cells express low ectonucleotidase activities, i.e. 5–10 nmolÆmin )1 Æmg )1 pro- tein with either MgATP or MgADP as the substrates. Cells transfected with the chicken liver ecto-ATPDase cDNA displayed 10- to 30-fold higher ATP hydrolysis activities. Because the expression level was higher in HeLa cells, which also have a more rapid growth rate than COS cells, we chose HeLa cells for stable transfection in order to characterize the expressed enzyme with respect to azide inhibition and the effect of NP-40 on ATP and ADP hydrolysis. Table 3 shows that: (a) in the absence of NP-40, the ATP hydrolysis activities were similar at pH 7.4 and 6.4, whereas ADP hydrolysis activity at pH 6.4 was threefold greater than at pH 7.4; (b) NP-40 (0.1%) increased both ATP and ADP hydrolysis activities by threefold to fivefold at both pH values; (c) ADPase/ATPase ratios varied between 0.2 and 0.64 depending on the pH and the presence of NP-40; (d) Fig. 3. Inhibition of ATP and ADP hydrolysis activities of purified chicken liver ecto-ATPDase by azide. ATP and ADP hydrolysis reac- tions were carried out in 0.25-mL reaction mixtures containing 50 m M Tris/HCl, pH 7.4 or 50 m M Mops, pH 6.4 with 5 m M MgCl 2 ,5m M ATP or 5 m M (ADP) in the absence and presence of the indicated concentrations of sodium azide.MgATP hydrolysis at pH 7.4 (s)and 6.4 (d); MgADP hydrolysis at pH 7.4 (h)and6.4(j). Fig. 4. Immunolocalization of ecto-ATPDase in chicken liver. Thin sections of chicken liver were stained with monoclonal antibody MC27 as described in Experimental procedures. Stained bile canaliculi are indicated by arrowheads. Bar ¼ 10 lm. 2378 A. F. Knowles et al. (Eur. J. Biochem. 269) Ó FEBS 2002 both ATPase and ADPase activities were inhibited by 10 m M azide, but ADPase activity was more sensitive to azide inhibition than ATP hydrolysis and inhibition in general was greater at pH 6.4 than at pH 7.4. These characteristics are similar to those displayed by the enzyme in its native membranes [23] and the purified enzyme described above. Figure 6 shows that MC18 detected protein bands of molecular masses of 80–85 kDa in HeLa cells expressing the chicken ecto-ATPDase but not in HeLa cells transfected with the pcDNA3 vector alone. DISCUSSION The existence of a cell surface ATPase in the bile canaliculi of liver was first demonstrated by ATPase activity staining in the 1950s [33,34]. Interestingly, this activity was both increased in magnitude and altered in cellular location in rat hepatomas induced by carcinogens [33,35]. Early biochemi- cal characterization of an ATPase activity in rat liver microsomes showed that it had the unusual properties of hydrolyzing also UTP and UDP and that hydrolyses of these nucleotides were inhibited by 5 m M azide [38]. We found that the extent of azide inhibition of ATPase activity in different plasma membrane preparations correlated with the ability of the membranes to hydrolyze ADP, and concluded that azide inhibition is a characteristic of a membrane-bound ATP diphosphohydrolase (ATPDase) [30]. We also showed that this activity is abundant in liver Fig. 5. Nucleotide sequence of the chicken liver ecto-ATPDase and the deduced primary sequence. Nucleotide numbers are on the left side and amino-acid residue numbers are on the right side of the figure. The transmembranous domains of the protein at the N- and C-terminus are shaded. The five apyrase conserved regions (ACRs) are underlined and in bold. The 10 cysteine residues conserved in all E-ATPases are indicated by a bold C with shading. Asparagine residues involved in potential N-glycosylation are indicated by bold italic N. The amino- acid residues that differ between chicken liver and oviduct ecto- ATPDases are at number 278–280, 316, 357, 461, 462. The different amino-acid residues in the chicken oviduct ecto-ATPDase are shown in parentheses following the corresponding amino-acid residue in the liver ecto-ATPDase. Table 3. Effect of NP-40 and azide on the ecto-ATPDase activities of HeLa Cells stably transfected by chicken liver ecto-ATPDase cDNA. HeLa cells were stably transfected with the chicken liver ecto-ATPDase cDNA as described under Experimental procedures. Cells were grown in DMEM containing 10% fetal bovine serum and 400 lgÆmL )1 geneticin. For ectonucleotidase assays, cells collected after trypsinization were used. Enzyme assays were carried out in a 0.5-mL reaction mixture containing either 50 m M Tris/HCl (pH 7.4) or 50 m M Mops (pH 6.4) with 5 m M MgATP or 5m M MgADP and 1 m M ouabain (inhibitor of the Na + /K + -ATPase) and 0.1 m M sodium azide (inhibitor of the mitochondrial ATPase) with 25 lL of trypsinized cells (50–75 lg cell protein). Reaction was carried out for 30 min at 37 °C. Duplicate samples were run for each condition. Results presented were average (± SD) of three experiments. Values are given as lmol P i Æmin )1 Æmg protein )1 . Addition pH 7.4 pH 6.4 ATPase ADPase ATPase ADPase None 0.26 ± 0.04 0.05 ± 0.01 0.24 ± 0.05 0.15 ± 0.05 0.1% NP-40 0.84 ± 0.28 0.28 ± 0.04 0.75 ± 0.22 0.48 ± 0.17 10 m M azide 0.19 ± 0.04 0.01 ± 0.01 0.14 ± 0.03 0.03 ± 0.05 0.1% NP-40 + 10 m M azide 0.59 ± 0.16 0.10 ± 0.02 0.40 ± 0.11 0.06 ± 0.05 Fig. 6. Expression of the chicken liver ecto-ATPDase in transiently and stably transfected HeLa cells. HeLa cells transfected with vector alone or chicken liver ecto-ATPDase cDNA were frozen to lyse the cells. Cell lysates (25 lg protein) were treated by SDS gel sample buffer without reducing agent, and subjected to SDS/PAGE on 7.5% acrylamide gel. Separated proteins were transferred to a poly(vinylidene difluoride) membrane, which was probed with monoclonal antibody MC18. Lane 1, chicken liver plasma membrane; lane 2, lysates of HeLa cells transfected with pcDNA vector; lane 3, lysates of HeLa cells trans- fected with chicken liver ecto-ATPDase cDNA. Ó FEBS 2002 Molecular identification of a liver ecto-ATPDase (Eur. J. Biochem. 269) 2379 and several other tissues [30]. Subsequently, an ATPase possessing NTPase and NDPase activities was partially purified from rat liver [39], and was shown to be responsible for the E-ATPase activity of intact rat hepatocytes [40]. Thattheecto-ATPDaseisthemajorE-ATPaseinlivers was supported by our recent study of the E-ATPase activity in the chicken liver plasma membranes. We showed that the chicken liver E-ATPase differs from the chicken smooth muscle ecto-ATPase with respect to substrate utilization, divalent ion requirement, thermal stability, azide inhibition, and response to a panel of modulators [23]. On the other hand, its enzymatic properties are similar to the chicken oviduct ecto-ATPDase that we have previously purified and cloned [7,21]. Utilizing antibodies developed for the chicken oviduct ecto-ATPDase, we achieved the first complete purification of a liver ecto-ATPDase. The purified chicken liver ecto-ATPDase is an 85-kDa membrane glycoprotein. Although slightly larger than the chicken oviduct ecto-ATPDase, it is nearly identical to the chicken oviduct ecto-ATPDase with respect to enzymatic properties. The similar properties of the two chicken ecto- ATPDases and their cross-reactivities with the same antibodies suggest that they may have similar primary sequences in spite of their different molecular masses. This proved to be the case when we obtained the cDNA of the chicken liver ecto-ATPDase by RT-PCR cloning. The two chicken ecto-ATPDases differ in only seven amino acids out of 493 amino-acid residues, all located in the C-terminal half of the molecules that are less conserved in the E-ATPases [21]. There is an additional N-glycosylation site in the liver ecto-ATPDase. These substitutions appear to have negli- gible effect on the enzymatic properties of the liver ecto- ATPDase. Upon transfection of HeLa cells with the chicken liver ecto-ATPDase cDNA, expression of the cDNA was demonstrated both by Western blotting and enzyme activity. The expressed protein still binds the chicken ecto- ATPDase monoclonal antibody MC18 but there is some evidence of heterogeneity in glycosylation in HeLa cells (Fig. 6). The expressed enzyme was able to hydrolyze ATP and ADP, was inhibited by 10 m M azide, and was markedly stimulated by the detergent NP-40 (Table 3). The response to detergents constitutes the most striking difference between the chicken ecto-ATPDases (E-NTPDase III) and chicken ecto-ATPase (E-NDPDase II). The ecto- ATPases of chicken gizzard, brain, and transverse tubules are inactivated by most detergents, except digitonin [4,5,23,27,41,42]. In contrast, the chicken ecto-ATPDases are activated by the same detergents and are extracted from the membranes by high concentrations of NP-40 [7,23] (and this study). This study shows unambiguously that the activity of the chicken liver ecto-ATPDase is increased by NP-40 even when the protein is expressed in nonliver host cells, leading to the conclusion that this property is intrinsic to the chicken liver ecto-ATPDase protein and cannot be attributed to any specific lipid environment of the liver cell membranes. The differential effects of detergents on the chicken ecto- ATPase and ecto-ATPDase are most likely related to the different regulatory mechanisms of the two enzymes. Compounds that promote oligomer formation, such as Con A and chemical cross-linking agents increase the activity of chicken ecto-ATPase [5,23,43], whereas deter- gents and other amphiphilic molecules, which prevent oligomer formation of membrane proteins, decrease its activity [27,28]. In contrast, the chicken liver ecto-ATPDase activity is not affected by ConA [23] while the chicken stomach ecto-ATPDase is actually markedly inhibited by a chemical cross-linking agent [44]. We and others have proposed that ecto-ATPase requires oligomerization for high activity [23,28,43,45], but the chicken ecto-ATPDase does not [23]. Interestingly, the CD39s, although also ecto- ATPDases, are adversely affected by detergents [20,46]. Studies from Guidotti’s laboratory showed that while the full-length membrane-bound rat CD39/ecto-ATPDase was inhibited by detergents, recombinant forms that lack either or both transmembranous domains had lower activities and were not inhibited by detergents [46,47]. It has been proposed that these domains are involved in the formation of a tetramer of the enzyme [46]. Our finding that the chicken liver ecto-ATPDase, whether expressed in the native or host cell membrane, is activated by detergents while unaffected by ConA, indicates that its activity is not dependent on its oligomerization status. While it is likely that the enzyme is fully active as monomers, the possibility that the enzyme has a more detergent-resistant quaternary structure cannot be ruled out at present. In either case, it will be of interest to determine if this unusual characteristic can be attributed to its transmembranous domains that have different sequences than those of the chicken smooth muscle ecto-ATPase and rat CD39/ecto- ATPDase. However, activation by detergents may be a unique property of the chicken ecto-ATPDases because the ecto-ATPDase activity of the rat liver plasma membranes is inhibited by NP-40 (A. F. Knowles, unpublished data). Nevertheless, like the chicken ecto-ATPDases, the rat liver ecto-ATPDase is also inhibited by azide (A. F. Knowles & A. K. Nagy, unpublished data). In contrast, the ecto- ATPDase activity of porcine liver, shown also to be distributed in the bile canaliculi, was reported to be less sensitive to azide inhibition [48,49]. Furthermore, analysis of the N-terminal amino-acid sequence of porcine liver ATP- Dase revealed identity to the rat liver lysosomal ATPase [49] suggesting that it may not be a member of the E-ATPase family. The possibility that more than one enzyme protein contribute to the overall liver ATPDase activity in different species will require further investigation. The function of the liver ecto-ATPDase is not established. While its localization in the bile canaliculi is suggestive of its involvement in bile acid secretion, experimental evidence is tenuous [50]. The proposal that it may play a role in purine salvage is more likely because 5¢-nucleotidase, which hydrolyzes AMP to adenosine, is also present in the bile canaliculi [50]. The liver ecto-ATPDase may also be important in regulating signaling by extracellular nucleo- tides because they elicit a variety of responses in liver [51]. It is expected then that the protein may be colocalized with a purinergic receptor. The molecular identification of a liver ecto-ATPDase should facilitate future investigation of its physiological function. ACKNOWLEDGEMENTS This work was supported by the California Metabolic Research Foundation. We thank Dr Jacques Perrault for providing the HeLa cells, and Drs Charles Caldwell and Rita Lim for helpful discussions. 2380 A. F. Knowles et al. (Eur. J. Biochem. 269) Ó FEBS 2002 REFERENCES 1. Burnstock, G. (1997) The past, present and future of purine nucleotides as signaling molecules. Neuropharmacology 36, 1127– 1139. 2. Redegeld, F.A., Caldwell, C.C. & Sitkovsky, M.V. (1999) Ecto- protein kinases: ecto-domain phosphorylation as a novel target for pharmaceutical manipulation. Trends Pharmacol. Sci. 20, 453–459. 3. Plesner, L. (1995) Ecto-ATPases: identities and functions. Int. Rev. Cytol. 158, 141–214. 4. Treuheit, M.J., Vaghy, P.L. & Kirley, T.L. (1992) Mg 2+ -ATPase from rabbit skeletal muscle transverse tubules is 67-kilodalton glycoprotein. J. Biol. Chem. 267, 11777–11782. 5. Stout, J.G. & Kirley, T.L. (1994) Purification and characterization of the ecto-Mg 2+ -ATPase of chicken gizzard smooth muscle. J. Biochem. Biophys. Methods 29, 61–75. 6. Christoforidis, S., Papamarcaki, T., Galaris, D., Kellner, R. & Tsolas, O. (1995) Human plancental ATP diphosphohydrolase is highly N-glycosylated plasma membrane enzyme. Eur. J. Biochem. 234, 66–74. 7. Strobel, R.S., Nagy, A.K., Knowles, A.F., Buegel, J. & Rosenberg, M.O. (1996) Chicken oviductal ecto-ATP- diphosphohydrolase. Purification and characterization. J. Biol. Chem. 271, 16323–16331. 8. Zimmermann, H., Beaudoin, A.R., Bollen, M., Goding, J.W., Guidotti, G., Kirley, T.L., Robson, S.C. & Sano, K. (2000) Proposed nomenclature for two novel nucleotide hydrolyzing enzyme families expressed on the cell surface. In Ecto-Atpases and Related Ectonucleotidases (Vanduffel, L. & Lemmens, R., eds), pp. 1–8. Shaker Publishing B.V., Maastricht, the Netherlands. 9. Handa, M. & Guidotti, G. (1996) Purification and cloning of a soluble ATP-diphosphohydrolase (apyrase) from potato tubers (Solanum tuberrosum). Biochem. Biophys. Res. Commun. 218, 916–923. 10. Vasconcelos, E.G., Ferreira, S.T., de Carvalho, T.M.U., de Souza, W., Kettlun, A.M., Mancilla, M., Valuenzuela, M.A. & Verjovski- Almeida, S. (1996) Partial purification and immunohistochemical localization of ATP diphosphohydrolase from Schistosoma man- soni: Immunological cross reactivities with potato apyrase and Toxoplasma gondii. J. Biol. Chem. 271, 22139–22145. 11. Kansas, G.S., Wood, G.S. & Tedder, T.F. (1991) Expres- sion, distribution, and biochemistry of human CD39. Role in activation-associated homotypic adhesion of lymphocytes. J. Immunol. 146, 2235–2244. 12. Maliszewski, C.R., Delespesse, G.J.T., Schoenborn, M.A., Armitage, R.J., Fanslow, W.C., Nakajima, T., Baker, E., Suth- erland, G.R., Poindexter, K., Birks, C., Alpert, A., Friend, D., Gimpel, S.D. & Gayle, R.B. III (1994) The CD39 lymphoid cell activation antigen. Molecular cloning and structural character- ization. J. Immunol. 153, 3574–3583. 13. Wang,T.F.,Rosenberg,P.A.&Guidotti,G.(1997)Character- ization of brain ecto-apyrase: evidence for only one ecto-apyrase (CD39) gene. Mol. Brain Res. 47, 295–302. 14. Matsumoto, M., Sakurai, Y., Komubo, T., Yagi, H., Makita, K., Matsui, T., Titani, K., Fujimura, Y. & Narita, N. (1999) The cDNA cloning of human placental ecto-ADP diphospho- hydrolases I and II. FEBS Lett. 453, 335–340. 15. Lemmens, R., Vanduffel, L., Kittel, A., Beaudoin, A.R., Benrezzak, O. & Se ´ vigny, J. (2000) Distribution, cloning and characterization of porcine nucleoside triphosphate diphospho- hydrolase-1. Eur. J. Biochem. 267, 4106–4114. 16. Kaczmarek, E., Koziak, K., Se ´ vigny, J., Siegel, J.B., Anrather, J., Beaudoin, A.R., Bach, F.H. & Robson, S.C. (1996) Identification and characterization of CD39/vascular ATP diphosphohydrolase. J. Biol. Chem. 271, 33116–33122. 17. Marcus, A.J., Broekman, M.J., Drosopoulos, J.H.F., Islam, N., Alyonycheva, T.N., Safier, L.B., Hajjar, K.A., Posnett, D.N., Shoenborn, M.A., Schooley, K.A., Gayle, R.B. & Maliszewski, C.R. (1997) The endothelial cell ecto-ADPase responsible for inhibition of platelet function is CD39J. Clin. Invest. 99, 1351– 1360. 18. Se ´ vigny, J., Grondin, G., Gendron, F P., Roy, J. & Beaudoin, A.R. (1998) Demonstration and immunolocalization of ATP diphosphohydrolase in the pig digestive system. Am.J.Physiol. 275, G473–G482. 19. Braun, N., Se ´ vigny, J., Robson, S.C., Enjyoji, K., Guchelberger, O., Hammer, K., Di Virgilio, F. & Zimmermann, H. (2000) Assignment of ecto-nucleoside triphosphate diphosphohydrolase- 1/cd39 expression to microglia and vasculature of the brain. Eur. J. Neurosci. 12, 4357–4366. 20. Lemmens, R., Kupers, L., Se ´ vigny, J., Beaudoin, A.R., Grondin, G., Kittel, A., Waelkens, E. & Vanduffel, L. (2000) Purification, characterization, and localization of an ATP diphosphohydrolase in porcine kidney. Am. J. Physiol. 278, F978–F988. 21. Nagy, A.K., Knowles, A.F. & Nagami, G.T. (1998) Molecular cloning of the chicken oviduct ecto-ATP-diphosphohydrolase. J. Biol. Chem. 273, 16043–16049. 22. Smith, T.M. & Kirley, T.L. (1998) Cloning, sequencing, and expression of a human brain E-type ATPase: homology to both ecto-ATPases and ecto-apyrases. Biochim. Biophys. Acta 1386, 65–78. 23. Caldwell, C.C., Davis, M.D. & Knowles, A.F. (1999) Ectonu- cleotidases of avian gizzard smooth muscle and liver plasma membranes: a comparative study. Arch. Biochem. Biophys. 362, 46–58. 24. Laemmli, K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680–685. 25. Wordinger, R. (1987) Manual of Immunoperoxidase Techniques, 2nd edn. ASCP Press, Chicago. 26. Gao, J P. & Knowles, A.F. (1993) The epidermal growth factor/ cAMP-inducible ectoCa 2+ -ATPase of human hepatoma Li-7A cells is similar to rat liver ectoATPase/hepatocyte cell adhesion molecule (Cell-CAM 105). Arch. Biochem. Biophys. 303, 90–97. 27. Caldwell, C.C., Norman, V., Urbina, A., Jarvis, A., Quinonz, C., Stemm, M. & Dahms, A.S. (1997) Regulatory differences among avian ecto-ATPases. Biochem. Biophys. Res. Commun. 238, 728–732. 28. Caldwell, C.C., Hornyak, S.C., Pendleton, E., Campbell, D. & Knowles, A.F. (2001) Regulation of chicken gizzard ecto-ATPase activity by modulators that affect its oligomerization status. Arch. Biochem. Biophys. 387, 107–116. 29. LeBel, D., Poirer, G.G., Phaneuf, S., St-Jean, P., Laliberte, J.F. & Beaudoin, A.R. (1980) Characterization and purification of calcium-sensitive ATP diphosphohydrolase from pig pancreas. J. Biol. Chem. 255, 1227–1233. 30. Knowles, A.F., Isler, R.E. & Reece, J.F. (1983) The common occurrence of ATP diphosphohydrolase in mammalian plasma membranes. Biochim. Biophys. Acta 731, 88–96. 31. Knowles, A.F. & Nagy, A.K. (1999) Inhibition of an ecto-ATP- diphosphohydrolase by azide. Eur. J. Biochem. 262, 349–357. 32. Debruyne. I. & Stockx, J. (1978) Nucleoside triphosphatase from the hen’s egg white and vitelline membrane. Purification, proper- ties, and relation to similar enzymes from the oviduct. Enzyme 23, 361–372. 33. Novikoff, A.B. (1957) A transplantable rat liver tumor induced by 4-dimethylaminoazobenzene. Cancer Res. 17, 1010–1027. 34. Wachstein, M. & Meisel, E. (1959) Enzymatic histochemistry of ethionine induced liver cirrhosis and hepatoma. J. Histochem. Cytochem. 7, 189–201. 35. Karasaki, S. (1972) Subcellular localization of surface adenosine triphosphatase activity in preneoplastic liver parenchyma. Cancer Res. 32, 1703–1712. Ó FEBS 2002 Molecular identification of a liver ecto-ATPDase (Eur. J. Biochem. 269) 2381 36. Okada, T., Zinchuk, V.S., Kobayashi, T. & Seguchi, H. (1999) Dynamics of rat liver ecto-ATPase during development suggests its involvement in bile acid efflux. Cell Tissue Res. 298, 511–518. 37. Lewis-Carl,S.&Kirley,T.L.(1997)Immunolocalizationofthe ecto-ATPase and ecto-apyrase (ATPDase) in chicken gizzard and stomach: purification and N-terminal sequence of the stomach ecto-apyrase. J. Biol. Chem. 272, 23645–23652. 38. Ernster, L. & Jones, L.C. (1962) Study of the nucleoside tri- and diphosphatase activities of rat liver microsomes. J. Cell Biol. 15, 563–577. 39. Lin, S H. & Fain, J.N. (1984) Purification of (Ca 2+ -Mg 2+ )- ATPase from rat liver plasma membranes. J. Biol. Chem. 259, 3016–3020. 40. Lin, S H. & Russel, W.E. (1988) Two Ca 2+ -dependent ATPases in rat liver plasma membrane. The previously purified (Ca 2+ - Mg 2+ )-ATPase is not a Ca 2+ -pump but an ecto-ATPase. J. Biol. Chem. 263, 12253–12258. 41. Cunningham, H.B., Yazaki, P.J., Domingo, R., Oades, K.V., Bohlen, H., Sabbadini, R.A. & Dahms, A.S. (1993) The skeletal muscle transverse tubular Mg-ATPase: identity with Mg-ATPases of smooth muscle and brain. Arch. Biochem. Biophys. 202, 32–43. 42. Megias, A., Martinez-Senac, M.M., Delgado, J. & Saborido, A. (2001) Regulation of transverse tubule ecto-ATPase activity in chicken skeletal muscle. Biochem. J. 353, 521–529. 43. Stout, J.G. & Kirley, T.L. (1996) Control of cell membrane ecto- ATPase by oligomerization state: intermolecular cross-linking modulates ATPase activity. Biochemistry 35, 8289–8298. 44. Lewis-Carl, S.A., Smith, T.M. & Kirley, T.L. (1998) Cross-linking induces homodimer formation and inhibits enzymatic activity of chicken stomach ecto-apyrase. Biochem. Mol. Biol. Int. 44, 463–470. 45. Beeler, T.J., Gable, K. & Keffer, J. (1983) Characterization of the membrane bound Mg 2+ -ATPase of rat skeletal muscle. Biochim. Biophys. Acta 734, 221–234. 46. Wang,T.F.,Ou,Y.&Guidotti,G.(1998)Thetransmembrane domains of ectoapyrase (CD39) affect its enzymatic activity and quaternary structure. J. Biol. Chem. 273, 24814–24821. 47. Grinthal, A. & Guidotti, G. (2000) Substitution of His59 converts CD39 apyrase into an ADPase in a quaternary structure depen- dent manner. Biochemistry 39, 9–16. 48. Leclerc, M C., Grondin, G., Gendron, F.P., Se ´ vigny, J. & Beaudoin, A.R. (2000) Identification, characterization, and immunolocalization of a nucleoside triphosphate diphospho- hydrolase in pig liver. Arch. Biochem. Biophys. 377, 372–378. 49. Se ´ vigny, J., Robson, S., Waelkens, E., Csizmadia, E., Smith, R.N. & Lemmens, R. (2000) Identification and characterization of a novel hepatic canalicular ATP diphosphohydrolase. J. Biol. Chem. 275, 5640–5647. 50. Che, M., Gatmaitan, Z. & Arias, I.M. (1997) Ectonucleotidases, purine nucleoside transporter, and function of the bile canalicular plasma membrane of the hepatocyte. FASEB J. 11, 101–108. 51. Dubyak, G.R. & El-Motassim, C. (1993) Signal-transduction via P2-purinergic receptors for extracellular ATP and other nucleo- tides. Am. J. Physiol. 265, C577–C606. 2382 A. F. Knowles et al. (Eur. J. Biochem. 269) Ó FEBS 2002 . ADP hydrolysis by  70% whereas ATP hydrolysis was inhibited by only 10%. Inhibition of ADP and ATP hydrolysis by azide was greater when the pH of assay solutions. ATP hydrolysis activity in the presence of Ca 2+ [26], the CaATPase activity of the purified chicken liver ecto-ATPDase is  30% of the MgATPase activity at