Alanine-scanning mutagenesis of the sixth transmembrane segment of gastric H+,K+-ATPase alpha-subunit.

The sixth transmembrane (M6) segment of the catalytic subunit plays an important role in the ion recognition and transport in the type II P-type ATPase families. In this study, we singly mutated all amino acid residues in the M6 segment of gastric H(+),K(+)-ATPase alpha-subunit with alanine, expressed the mutants in HEK-293 cells, and studied the effects of the mutation on the functions of H(+),K(+)-ATPase; overall K(+)-stimulated ATPase, phosphorylation, and dephosphorylation. Four mutants, L819A, D826A, I827A, and L833A, completely lost the K(+)-ATPase activity. Mutant L819A was phosphorylated but hardly dephosphorylated in the presence of K(+), whereas mutants D826A, I827A, and L833A were not phosphorylated from ATP. We found that almost all of these amino acid residues, which are important for the function, are located on the same side of the alpha-helix of the M6 segment. In addition, we found that amino acids involved in the phosphorylation are located exclusively in the cytoplasmic half of the M6 segment and those involved in the K(+)-dependent dephosphorylation are in the luminal half. Several mutants such as I821A, L823A, T825A, and P829A partly retained the K(+)-ATPase activity accompanying the decrease in the rate of phosphorylation.

M8, and M9 segments, only M5 and M9 have topogenic function, whereas the other three segments do not have this function in translation/insertion experiments in vitro (12). (ii) The M5/M6 hairpins are released from the membrane into the luminal side after trypsin digestion of Na ϩ ,K ϩ -ATPase and H ϩ ,K ϩ -ATPase in the absence of K ϩ (13,14). (iii) The M5/M6 hairpins are stabilized in the membrane in the presence of K ϩ (13,14). These results altogether suggest that the M5/M6 hairpins of the Na ϩ ,K ϩ -ATPase and H ϩ ,K ϩ -ATPase ␣-subunits are surrounded with other transmembrane segments and are intimately associated with K ϩ . It is proposed that these amphipathic characters of the M5/M6 hairpins are important for the transport of ions.
In gastric H ϩ ,K ϩ -ATPase, the M6 segment contains cysteine residues reported to be the major binding sites of the benzimidazole derivatives, which are irreversible inhibitors of gastric H ϩ ,K ϩ -ATPase (15,16). These cysteine residues are themselves not involved in the enzyme function (16). Rather, the covalent binding of the inhibitor with these cysteines interferes with the conformational changes of the ATPase. Therefore, it is likely that the M6 segment is close to the structural center important for the function of gastric H ϩ ,K ϩ -ATPase.
In the present study, we focused on the M6 segment of gastric H ϩ ,K ϩ -ATPase ␣-subunit and studied the role of the side chain of each amino acid residue from Cys 815 to Leu 833 in the M6 segment in the partial reactions of H ϩ ,K ϩ -ATPase to determine the affinity for cations. For this purpose, we performed alanine-scanning mutagenesis, which had been used previously to study the structure-function relationships of specific domain(s) of membrane proteins, including ion pumps and ion channels, especially for the binding site of drugs, ligands, and proteins (17)(18)(19)(20)(21).
cDNAs of ␣and ␤-Subunits of H ϩ ,K ϩ -ATPase-cDNAs of the ␣and ␤-subunits of H ϩ ,K ϩ -ATPase were prepared from rabbit gastric mucosa as described elsewhere (6). The ␣and ␤-subunit cDNAs were digested with EcoRI and XhoI. The obtained fragments were each ligated into the pcDNA3 vector treated with EcoRI and XhoI.
DNA Sequencing-DNA sequencing was performed by the dideoxy chain termination method using an Autocycle DNA sequencing kit and an ALFexpress DNA sequencer (Amersham Pharmacia Biotech).
Site-directed Mutagenesis-A plasmid containing cDNA for rabbit gastric H ϩ ,K ϩ -ATPase ␣-subunit was subjected to site-directed mutagenesis to create a unique restriction site of AvrII at the nucleotide position of 2436 (corresponding to Leu 813 ). The sequence of CCTGGG was mutated to CCTAGG without amino acid replacement. The following introduction of site-directed mutations in the M6 segment of the H ϩ ,K ϩ -ATPase ␣-subunit was carried out by sequential polymerase chain reaction (PCR) steps (7), in which appropriately mutated ␣-subunit cDNAs (segments between the AvrII site (nucleotide 2436) and the AatII site (nucleotide 2830)) were prepared. Two kinds of flanking sequence primers were prepared, one is the 5Ј-flanking sense primer, 5Ј-TGACAACCTGAAGAAGTCCATC-3Ј (nucleotide 2340 to 2361), and the other is the 3Ј-flanking antisense primer, 5Ј-CGATGCACACCTG-GAACACGAT-3Ј (nucleotide 2908 to 2929). In addition, sense and antisense synthetic oligonucleotides, each 21 bases long containing one or two mutated bases near the center, were designed (referred as the sense-mutating primer and the antisense-mutating primer). In the first PCR amplification step, the H ϩ ,K ϩ -ATPase ␣-subunit cDNA subcloned in pBluescript SK(Ϫ) was used as a DNA template (6). Two fragments were prepared in this step; one between the 5Ј-flanking sense primer and the antisense mutating primer, and the other between the sense mutating primer and the 3Ј-flanking antisense primer. Each amplified fragment was purified by gel electrophoresis, combined and incubated with the 5Ј-flanking sense primer and the 3Ј-flanking antisense primer in the second PCR amplification. The amplified fragment was purified by gel electrophoresis, subcloned in pCR-Script Amp SK(ϩ) (Stratagene), and sequenced. PCR was routinely carried out in the presence of 300 M of each dNTP, 6 M primers, 10 mM KCl, 20 mM Tris-HCl (pH 8.8), 10 mM (NH 4 ) 2 SO 4 , 2 mM MgSO 4 , 0.1% Triton X-100, 100 g/ml bovine serum albumin, and 2 units of Pfu DNA polymerase for 25 cycles. After sequencing, the amplified fragment in the second PCR was digested with AvrII and AatII, and ligated back into the relevant position of the wild-type construct of the ␣-subunit. Site-directed mutagenesis in Glu 822 and Asp 826 was carried out as described elsewhere by using the MutanK kit (6).
Cell Culture, Transfection, and Preparation of Membrane Fractions-Cell culture of HEK-293 was carried out as described previously (6,10). The ␣and ␤-subunit cDNA transfection was performed by the calcium phosphate method with 10 g of DNA per 10-cm dish. DNA was purified by cesium chloride density gradient centrifugation or with an EndoFree Plasmid Maxi kit (Qiagen, Tokyo). Cells were harvested 2 days after the DNA transfection. Membrane fractions of HEK cells were prepared as described previously (6). Briefly, cells in a 10-cm Petri dish were washed with phosphate-buffered saline and incubated in 2 ml of low ionic salt buffer (0.5 mM MgCl 2 and 10 mM Tris-HCl, pH 7.4) at 0°C for 10 min. After the addition of phenylmethylsulfonyl fluoride (1 mM) and aprotinin (0.09 unit/ml), the cells were homogenized in a Dounce homogenizer, and the homogenate was diluted with an equal volume of a solution comprising 500 mM sucrose and 10 mM Tris-HCl, pH 7.4. The homogenized suspension was centrifuged at 800 ϫ g for 10 min. The supernatant was centrifuged at 100,000 ϫ g for 90 min, and the pellet was suspended in a solution containing 250 mM sucrose and 5 mM Tris-HCl, pH 7.4.
SDS-Polyacrylamide Gel Electrophoresis and Immunoblot-SDSpolyacrylamide gel electrophoresis was carried out as described elsewhere (22). Membrane preparations (30 g of protein) were incubated in a sample buffer comprising 2% SDS, 2% ␤-mercaptoethanol, 10% glycerol, and 10 mM Tris-HCl, pH 6.8, at room temperature for 2 min and applied to the SDS-polyacrylamide gel. Immunoblot was carried out as described previously (6).
Quantification of Expressed H ϩ ,K ϩ -ATPase in the Membrane Fraction-A hog gastric vesicle preparation, which contains H ϩ ,K ϩ -ATPase, was prepared as described elsewhere (24). The membrane fractions of HEK cells in parallel with a series of diluted gastric vesicle preparations were run on the same SDS-polyacrylamide gel and blotted. The blots were scanned using an optical scanning image system. The content of H ϩ ,K ϩ -ATPase in the HEK membrane fraction was estimated from the standard curve obtained using the gastric vesicle preparation.
Assay of H ϩ ,K ϩ -ATPase Activity-H ϩ ,K ϩ -ATPase activity was measured in 1.2 ml of a reaction mixture comprising 50 g of membrane protein, 3 mM MgCl 2 , 160 M NADH, 0.8 mM phosphoenolpyruvate, 3 units/ml pyruvate kinase, 2.75 units/ml lactate dehydrogenase, 5 mM NaN 3 , 1 mM ouabain, 15 mM KCl, 40 mM Tris-HCl, pH 7.4, and 1 mM ATP. The decrease in the amount of NADH was measured at 37°C from the absorbance at 340 nm by a Beckman spectrophotometer as described elsewhere (25). The SCH28080-sensitive K ϩ -ATPase activity, defined as the H ϩ ,K ϩ -ATPase activity, was calculated as the difference between the K ϩ -ATPase activities in the presence and absence of 50 M SCH28080.
K ϩ -stimulated ATPase (K ϩ -ATPase) activity was measured in 1 ml of solution comprising 50 g of membrane protein, 3 mM MgCl 2 , 1 mM ATP, 5 mM NaN 3 , 2 mM ouabain, and 40 mM Tris-HCl, pH 7.4, in the presence and absence of 15 mM KCl. After incubation at 37°C for 30 min, the inorganic phosphate released was measured as described elsewhere (26). The K ϩ -ATPase activity was calculated as the difference between activities in the presence and absence of KCl.
Protein was measured using the BCA Protein Assay kit from Pierce (Rockford, IL) with bovine serum albumin as a standard.
Phosphorylation and Dephosphorylation-Fifty micrograms of membrane proteins was phosphorylated in 110 l of solution comprising 1 M ATP ([␥-32 P]ATP 4 ϫ 10 6 cpm), 2 mM MgCl 2 , 1 mM EGTA, 3 mM ouabain, and 40 mM Tris-HCl, pH 6.8, at 0°C for 15 s. When indicated, the phosphorylation reaction was performed at 10°C. The reaction was quenched by the addition of 590 l of ice-cold stop solution containing 10% trichloroacetic acid and 10 mM inorganic phosphate. The protein was collected at 13,000 ϫ g at 4°C for 3 min, and the pellet was washed with 500 l of ice-cold stop solution and 30% sucrose successively, solubilized in a sample buffer comprising 2% SDS, 2.5% dithiothreitol, 10% glycerol, and 50 mM Tris-HCl, pH 6.8, and subjected to the SDSpolyacrylamide gel electrophoresis under acidic conditions at pH 6.5 (10). The radioactivity associated with the H ϩ ,K ϩ -ATPase ␣-subunit separated on the gel was visualized and quantitated by digital autoradiography of the dried gel using Bio-Imaging Analyzer BAS2000 (Fuji Photo Film, Tokyo).
In the dephosphorylation experiments, the membrane protein phosphorylated above was incubated with various concentrations of KCl and 1 mM non-radioactive ATP at 0°C for 15 s. When indicated, it was incubated with 1 mM ADP and 1 mM non-radioactive ATP. The portions of the phosphorylated proteins were dephosphorylated in the presence of K ϩ or ADP. The reaction was stopped as described above, and the samples were run on the SDS-polyacrylamide gel. The radioactivity associated with the H ϩ ,K ϩ -ATPase ␣-subunit was visualized as described above. Fig. 1 shows a schematic model of the M5/M6 hairpin structure of rabbit gastric H ϩ ,K ϩ -ATPase ␣-subunit. It was shown that rabbit gastric H ϩ ,K ϩ -ATPase ␣-subunit spans the membrane between Lys 793 and Cys 815 (the M5 segment) and between Cys 815 and Lys 837 (the M6 segment), respectively, from the following experimental results: (i) Trypsin cleaved peptide bonds of hog gastric H ϩ ,K ϩ -ATPase ␣-subunit at Lys 791 and Lys 835 (which correspond to Lys 793 and Lys 837 in rabbit gastric H ϩ ,K ϩ -ATPase, respectively) in the cytoplasmic space ( Fig. 1) (27), and (ii) Cys 813 and Cys 822 of hog gastric H ϩ ,K ϩ -ATPase ␣-subunit (which correspond to Cys 815 and Cys 824 in rabbit gastric H ϩ ,K ϩ -ATPase, respectively) reacted with a cationic sulfenamide derived from omeprazole from the luminal space (15,16). It was also demonstrated that Leu 793 in sheep Na ϩ ,K ϩ -ATPase ␣ 1 -subunit (which corresponds to Leu 811 in rabbit gastric H ϩ ,K ϩ -ATPase) is exposed to the extracellular phase from the finding that the cysteine residue introduced in this position was labeled with a membrane-impermeable cysteine-directed reagent, N-biotinylaminoethyl methanethiosulfonate (28). Amino acid residues around Leu 793 (PLPLG) are well conserved between Na ϩ ,K ϩ -ATPase and gastric H ϩ ,K ϩ -ATPase ␣-subunits. Here, we replaced each of the amino acid residues between Cys 815 and Leu 833 , in or close to the M6 segment, of rabbit H ϩ ,K ϩ -ATPase ␣-subunit with alanine. These mutant ␣-subunit cDNAs were co-transfected with the gastric H ϩ ,K ϩ -ATPase ␤-subunit cDNA in HEK-293 cells. The expression levels of each mutant ␣-subunit in the membrane fractions were compared with that of the wild-type ␣-subunit in the Western blot detected with an anti-␣-subunit antibody. The expression levels ranged between 80% (C824A) and 125% (I818A) of that of the wild-type except for 31% for mutant L833A as shown in Table I. H ϩ ,K ϩ -ATPase Activity of the Mutants- Table I shows the K ϩ -ATPase and H ϩ ,K ϩ -ATPase activities of the M6 mutants. These two activities are comparable for each mutant, indicating that the K ϩ -ATPase activity of all the mutants is sensitive to 50 M SCH28080 and that this concentration is high enough to inhibit the K ϩ -ATPase activity of all the M6 mutants. Mutants C815A, I816A, T817A, F820A, C824A, T825A, F828A, and S832A retained 110, 76, 67, 65, 137, 62, 90, and 88% of the K ϩ -ATPase activity found in the wild-type, respectively. Mutants I818A, I821A, E822A, L823A, P829A, S830A, and V831A showed 31,33,15,47,34,25, and 30% of the K ϩ -ATPase activity found in the wild-type, respectively. Mutants L819A, D826A, I827A, and L833A almost lost the K ϩ -ATPase activity.

Construction of M6 Mutants of Gastric H
H ϩ ,K ϩ -ATPase consists of the ␣and ␤-subunits. The region involved in association with the ␤-subunit on the ␣-subunit is localized in and close to the fourth extracellular loop (between Arg 898 and Arg 922 ), which is very close to the M6 segment (29). Thus, it is not completely excluded that introducing a mutation in the M6 segment abolishes the assembly between the ␣and ␤-subunits, resulting in the loss of H ϩ ,K ϩ -ATPase activity. However, judged from the immunoprecipitation experiments, all the mutant ␣-subunits prepared in this study assembled with the ␤-subunit (data not shown), indicating that inactiva-tion of the K ϩ -ATPase activity in mutants L819A, D826A, I827A, and L833A is not due to the abolishment of the ␣/␤ assembly.
Phosphorylation Capacity of the Mutants-Membrane fractions of the cells expressing the wild-type enzyme and each alanine mutant were phosphorylated from [␥-32 P]ATP in the presence of 3 mM ouabain (to inhibit endogenous Na ϩ ,K ϩ -ATPase present in the membrane fraction) at 0°C for 15 s, separated on an SDS-polyacrylamide gel under weakly acidic conditions, and the patterns of the phosphorylated proteins were observed as reported previously (10) (Fig. 2). In the membrane fraction expressing the wild-type ␣and ␤-subunits, several radioactive bands were observed; doublet bands with molecular mass of about 100 kDa and a band with a molecular mass of 60 kDa ( Fig. 2A). The phosphorylation of the lower 100-kDa band was inhibited by 50 M SCH28080 ( Fig. 2A) or 1 mM sodium vanadate (data not shown). The phosphorylation of this band was not observed in the membrane fraction of cells expressing the ␤-subunit alone ( Fig. 2A) nor in mutant D387N, a phosphorylation site mutant (6), which cannot form an acylphosphate intermediate (data not shown). From these findings, it is concluded that the lower 100-kDa band represents the phosphorylated H ϩ ,K ϩ -ATPase ␣-subunit.
Mutants C815A, I816A, T817A, I818A, L819A, F820A, E822A, C824A, F828A, and S832A retained the ATP-dependent phosphorylation capacity. The phosphorylation was inhibited by 50 M SCH28080 (Fig. 2B). Mutants I821A, L823A, T825A, D826A, I827A, P829A, V831A, and L833A were hardly phosphorylated from ATP, because the phosphorylation levels of these mutants were much lower than that of the wild-type in the absence of SCH28080 (Fig. 2B). It should be noted that the phosphorylation levels of these mutants in the absence of SCH28080 were higher than those of the membrane fraction expressing the ␤-subunit alone or mutant D387N (data not shown) but comparable with the phosphorylation level of the wild-type in the presence of 50 M SCH28080 (Fig. 2B). Therefore, it is not completely excluded that these mutants retained very low capacity for phosphorylation at 0°C and that the sensitivity of the mutants to SCH28080 was lower than that of the wild-type. Mutant L819A retained the phosphorylation capacity, although this mutant lost the K ϩ -ATPase activity (Table I). Mutants I821A, L823A, T825A, P829A, and V831A apparently lost the phosphorylation capacity, although these FIG. 1. A schematic model of the M5/M6 hairpin structure of rabbit gastric H ؉ ,K ؉ -ATPase ␣-subunit. The sequence was obtained from Bamberg et al. (38). Sites cleaved with trypsin, between Lys 793 and Asn 794 and between Lys 837 and Ala 838 , are shown by the arrows (27). The binding sites of omeprazole, Cys 815 and Cys 824 , are shown by asterisks (15,16). mutants partly retained the K ϩ -ATPase activity (Table I). Later, phosphorylation of these mutants will be studied precisely.
Dephosphorylation Capacity of the Mutants-Next, we studied the dephosphorylation capacity of the wild-type H ϩ ,K ϩ -ATPase and the M6 mutants. Membrane fractions of the cells expressing the wild-type and the M6 mutants were phosphorylated from [␥-32 P]ATP followed by the incubation with KCl in the presence of 1 mM non-radioactive ATP at 0°C for 15 s. Even in the absence of KCl, the upper 100-kDa band present in HEK cells disappeared during incubation with cold ATP. Some portion of the phosphorylated ␣-subunit was dephosphorylated during incubation with cold ATP in the absence of KCl. The phosphorylated ␣-subunit of the wild-type H ϩ ,K ϩ -ATPase was dephosphorylated by the addition of K ϩ in a concentration-dependent manner. The half-maximally effective concentration of K ϩ for the dephosphorylation was about 0.1 mM (Fig. 3, A and  B). The phosphorylated ␣-subunits of mutants C815A and I818A were also dephosphorylated by the addition of K ϩ , although the sensitivity of these phosphorylated mutants for K ϩ were slightly lower than that of the wild-type enzyme (Fig. 3B). On the other hand, mutants L819A and E822A were dephosphorylated by only 20% and 50% in the presence of 50 mM KCl, respectively (Fig. 3B), indicating the presence of defect in the K ϩ -dependent dephosphorylation step, which resulted in complete or considerable loss of the overall H ϩ ,K ϩ -ATPase activity, respectively (Table I). The expression levels of mutant H ϩ ,K ϩ -ATPase ␣-subunits were estimated from the standard curve obtained using gastric vesicle preparation and expressed as percentages of the expression of the wild-type. All values of K ϩ -ATPase and H ϩ ,K ϩ -ATPase activities are given as means Ϯ S.E. for more than three membrane fractions. K ϩ -ATPase and H ϩ ,K ϩ -ATPase activities of the mutants were expressed as percentages of those of wild-type and normalized on the same expression levels.  2. A, phosphorylation of hog gastric vesicles, and the membrane fractions of the cells transfected with the wild-type ␣and ␤-subunit cDNAs, ␤-subunit cDNA alone. One microgram of gastric vesicles (GI) (lanes 1 and 2), and each 50 g of membrane fractions obtained from the cells co-transfected with the wild-type H ϩ ,K ϩ -ATPase ␣and ␤-subunit cDNAs (lanes 3 and 4) and from the cells transfected with the wild-type H ϩ ,K ϩ -ATPase ␤-subunit cDNA alone (lanes 5 and 6) were phosphorylated in 110 l of solution containing 1 M [␥-32 P]ATP (4 ϫ 10 6 cpm), 2 mM MgCl 2 , 1 mM EGTA, 3 mM ouabain, and 40 mM Tris-HCl (pH 6.8) at 0°C for 15 s. Prior to phosphorylation, the samples were preincubated with or without 50 M SCH28080 (shown as SCH 28080 ϩ or Ϫ) at 0°C for 30 min. The phosphorylation reaction was quenched by the addition of 590 l of ice-cold stop solution containing 10% trichloroacetic acid and 10 mM inorganic phosphate. The protein was collected at 13,000 ϫ g at 4°C for 3 min, washed with ice-cold stop solution and 30% sucrose, successively, and solubilized in a sample buffer for SDS-polyacrylamide gel electrophoresis. After gel electrophoresis at pH 6.5, the radioactivity was visualized by digital autoradiography of the dried gels using a Bio-Imaging Analyzer BAS2000. B, phosphorylation of the membrane fractions of the cells transfected with either the wild-type or the M6 mutant ␣-subunit cDNA in combination with the wild-type ␤-subunit cDNA. Fifty micrograms of membrane fractions of the cells co-transfected with the cDNA of the H ϩ ,K ϩ -ATPase ␤-subunit plus the cDNA of the wild-type H ϩ ,K ϩ -ATPase ␣-subunit or the indicated alanine mutant ␣-subunit were phosphorylated in the presence or absence of 50 M SCH28080 (shown as SCH 28080 ϩ or Ϫ). Bands representing the H ϩ ,K ϩ -ATPase ␣-subunit are shown by the arrow.
The phosphorylated ␣-subunit of the wild-type H ϩ ,K ϩ -ATPase was not dephosphorylated by the addition of ADP, indicating that the major phosphorylated intermediate is in the E 2 form (ADP-insensitive, K ϩ -sensitive form). Mutants L819A and E822A were not dephosphorylated by the addition of ADP, indicating that these mutants have no defect in the process of conversion from E 1 P to E 2 P (Fig. 4).
Substitutions of Leu 819 , Asp 826 , Ile 827 , and Leu 833 with Other Amino Acid Residues-Mutant L819A has a defect in the K ϩdependent dephosphorylation step, whereas mutants D826A, I827A, and L833A have defects in the phosphorylation steps. Here, we replaced Leu 819 with Gly, Val, or Met; Ile 827 with Val   FIG. 3. A and B, effects of K ϩ concentrations on the dephosphorylation reaction of the phosphorylated intermediates of the wild-type H ϩ ,K ϩ -ATPase and the M6 mutants. Fifty micrograms of each membrane protein were used. The wild-type H ϩ ,K ϩ -ATPase and mutants C815A, C818A, C819A, and E822A were phosphorylated from 1 M [␥-32 P]ATP at 0°C for 15 s. The phosphorylation was followed by the incubation with 1 mM non-radioactive ATP and indicated concentrations (0 -50 mM) of KCl at 0°C for 15 s. The reaction was quenched by the addition of the ice-cold stop solution. Precipitated proteins were separated on SDS-polyacrylamide gel, and the radioactivity associated with the H ϩ ,K ϩ -ATPase ␣-subunit was visualized by digital autoradiography (A) and expressed as the percentage of the control values measured in the absence of KCl (B): the wild-type H ϩ ,K ϩ -ATPase (q) and mutants C815A (), I818A (E), L819A (f), and E822A (‚). C, effects of K ϩ concentrations on the dephosphorylation reaction of the phosphorylated intermediates of the Leu 819 mutants. Fifty micrograms of membrane proteins; the wild-type H ϩ ,K ϩ -ATPase (q), and mutants L819G (E), L819A (f), L819V (), and L819M (᭛) were phosphorylated from [␥-32 P]ATP and dephosphorylated in the presence of indicated concentrations of KCl and 1 mM non-radioactive ATP. or Met; and Leu 833 with Gly, Val, or Met to study the role of the size of the side chain on the K ϩ -ATPase activity and its partial reactions. No K ϩ -ATPase activity was observed in mutant L819G, whereas mutants L819V and L819M retained 14% and 26% of the K ϩ -ATPase activity found in the wild-type enzyme, respectively (Table II). These mutants retained the phosphorylation capacity (Fig. 4). Mutants L819M and L819V were dephosphorylated by the addition of K ϩ , however, their sensitivity for K ϩ was 10 times and 100 times lower than that of the wild-type enzyme, respectively, whereas mutant L819G was dephosphorylated by only 30% in the presence of 50 mM KCl (Fig. 3C). These results suggest that the bulkiness or length of the side chain of the amino acid at this position (position 819) is critical for determining the affinity for K ϩ in the dephosphorylation step of the H ϩ ,K ϩ -ATPase.
Mutants I827V and I827M retained 43% and 49% of the K ϩ -ATPase activity found in the wild-type enzyme, respectively (Table II). These mutants also partially retained the phosphorylation capacity, 20% and 30% of the phosphorylation levels found in the wild-type enzyme, respectively. (Fig. 5). These results suggest that the bulkiness or length of the side chain of the amino acid at this position (position 827) is important in the phosphorylation step of the H ϩ ,K ϩ -ATPase.
Expression level of mutants L833G was 31% of that of the wild-type (Table II), which was comparable with that of mutant L833A. However, the expression levels of mutants L833V and L833M were significantly higher than that of mutant L833A. Mutants L833A and L833G lost K ϩ -ATPase activity and phosphorylation capacity, whereas mutants L833V and L833M retained them. These results suggest that the bulkiness or length of the side chain of the amino acid at this position (position 833) is important for the expression of the ␣-subunit as well as the phosphorylation step of the H ϩ ,K ϩ -ATPase. We also prepared a double mutant, E822D/D826E to study the interaction between Glu 822 and Asp 826 . E822D/D826E did not retain the K ϩ -ATPase activity and phosphorylation capacity, indicating that the roles of these negative side chains are not interchangeable.
Phosphorylation Capacity at Higher Temperature or after Longer Incubation-It is apparently contradictory that mutants I821A, L823A, T825A, P829A, and V831A considerably retained the H ϩ ,K ϩ -ATPase and K ϩ -ATPase activities, whereas they almost completely lost the ATP-dependent phosphorylation capacity as shown in Fig. 2. In the experiments shown in Table I and Fig. 2, the ATPase activities and phosphorylation capacity were assayed at 37°C and 0°C, respectively. It is expected that phosphorylation capacity may be observed in these mutants at higher temperatures. Hence, here we studied the phosphorylation capacity of these mutants at 10°C. Mutants I821A, L823A, and T825A were phosphorylated from ATP at 10°C, and their phosphorylation levels were 42%, 57%, and 65% of that of the wild-type, respectively (Fig. 6). Mutant P829A was partly phosphorylated from ATP. It is hard to observe whether mutants D826A, L827A, and V831A were phosphorylated from ATP at 10°C, because their phosphorylation levels were very low and not significantly different from that of the wild-type in the presence of 50 M SCH28080. However, it is noteworthy that the phosphorylation levels of these mutants were significantly higher than that of mutant D387N (as a negative control for phosphorylation), indicating that very low levels of phosphorylation were found in these mutants. These results altogether indicate that mutants I821A, L823A, T825A, and P829A partially retained phosphorylation capacity at 10°C. Unfortunately, we cannot estimate the phosphorylation levels at higher temperatures because of the high background phosphorylation.
It is expected that the mutations decrease the rate of phosphorylation resulting in the apparent loss of phosphorylation capacity at 0°C. Here, we studied the time course of phosphorylation between 15 and 60 s. Wild-type enzyme attained the steady state of phosphorylation within 15 s, whereas other mutants (I821A, L823A, T825A, D826A, L827A, and P829A) did not (Fig. 7). Mutants I821A, L823A, and T825A reached the steady state of phosphorylation within 30 s, and their phosphorylation levels were 40 -60% of that of the wild-type, indicating that the rate of phosphorylation of these mutants was decreased compared with the wild-type. However, a clear steady state of phosphorylation was not observed in mutants D826A and L827A because of their low phosphorylation levels. DISCUSSION In this study, we performed alanine-scanning mutagenesis of all the amino acid residues in the M6 segment of gastric H ϩ ,K ϩ -ATPase ␣-subunit, which is regarded as one of the major transmembrane segments involved in ion recognition and transport.
Single replacements of Leu 819 , Asp 826 , Ile 827 , or Leu 833 with alanine abolished the K ϩ -ATPase activity. In addition, Glu 822 is involved in determining the affinity for K ϩ in the K ϩ -ATPase activity (7, 8) and its dephosphorylation step as shown in the present study. The negative charge of this glutamic acid was proposed to inhibit the dephosphorylation step of H ϩ ,K ϩ -ATPase in the absence of K ϩ , and it was proposed that K ϩ stimulates the dephosphorylation by neutralizing the negative charge (30). It is very interesting that amino acid residues such as Leu 819 , Glu 822 , Asp 826 , and Leu 833 , which are important for the function of H ϩ ,K ϩ -ATPase, are predicted to be located on the same side of the putative ␣-helical structure of the M6 segment along the transmembrane direction (Fig. 8A). Ile 827 , which is also important for the function of H ϩ ,K ϩ -ATPase, is localized about 90°apart from these clustered important amino acid residues of Leu 819 , Glu 822 , Asp 826 , and Leu 833 on the helical wheel model. On the other hand, the K ϩ -ATPase activity was not suffered when the amino acid residues localized on the other side of the M6 segment such as Thr 817 , Phe 820 , Cys 824 , Phe 828 , and Cys 832 were replaced by alanine (Fig. 8A). Based on these findings and the following discussion, we pres-ent a hypothesis that Leu 819 , Glu 822 , Asp 826 , and Leu 833 face the putative ion channel-like structure (or putative ion-translocating pore), which consists of several transmembrane segments, M4, M5, and M6. The side chain of Ile 827 , which is close to Asp 826 , may support this structure. Similar helical wheel models were previously presented for the M4 segments of Na ϩ ,K ϩ -ATPase and H ϩ ,K ϩ -ATPase (31).
We also found that two groups of amino acid residues that are involved in different reaction steps of H ϩ ,K ϩ -ATPase show a unique distribution; that is, amino acid residues involved in the phosphorylation step such as Asp 826 , Ile 827 , and Leu 833 are located in the carboxyl-terminal half (cytoplasmic side) of the M6 segment, whereas those involved in the K ϩ -dependent dephosphorylation step such as Leu 819 and Glu 822 are located in the amino-terminal half (luminal side) of the M6 segment (Fig.  8B).
Phosphorylation of mutants at positions of Ile 821 , Leu 823 , Thr 825 , and Pro 829 was impaired at 0°C, but it was consider-TABLE II Expression of K ϩ -ATPase and H ϩ ,K ϩ -ATPase activities of M6 mutants The expression levels of mutant H Ϫ ,K ϩ -ATPase ␣-subunits were estimated from the standard curve obtained using gastric vesicle preparation and expressed as percentages of the expression of the wild-type. All values of K ϩ -ATPase and H ϩ ,K ϩ -ATPase activities are given as means Ϯ S.E. for more than three membrane fractions. K ϩ -ATPase and H Ϫ ,K Ϫ -ATPase activities of the mutants were expressed as percentages of those of wild-type and normalized on the same expression levels.  ably recovered at 10°C. It takes longer time for mutants I821A, L823A, T825A, and P829A to reach the steady state of phosphorylation than the wild-type. Therefore, the rate of phosphorylation of these mutants was significantly decreased by mutations. It is also noted that the steady-state phosphorylation levels of these mutants were lower compared with that of the wild-type.
Next, we discuss the role of bulkiness of the side chains at the position of amino acids 819, 826, 827, and 833 on the function of H ϩ ,K ϩ -ATPase. At the position of amino acid 819, mutations of Leu to Ala and Gly abolished the K ϩ -ATPase activity and K ϩ -dependent dephosphorylation. These mutants showed much lower or almost no sensitivity to K ϩ in the dephosphorylation step. However, a mutation to Met and Val partly retained the K ϩ -ATPase activity and K ϩ -dependent dephosphorylation (Table II and Fig. 3C). These mutants showed intermediate affinity for K ϩ between the wild-type and the alanine mutant. Therefore, the bulkiness of the side chain at this position is very important for the K ϩ -dependent dephosphorylation and determining the affinity for K ϩ , and small side chains cannot support the function.
At the position of amino acid 826, not only the Ala mutation but also the isosteric mutation to Asn and the isocharge mutation to Glu abolished the K ϩ -ATPase activity and phosphorylation (7). Therefore, both the present residue size of Asp and negative charge at this position are very important to retain the phosphorylation capacity. It is also noteworthy that the K ϩ -ATPase activity was not recovered when Glu 822 was replaced by Asp in combination with the mutation of Asp 826 to Glu (Table II). Therefore, there is no compensation of the size of the side chain between Glu 822 and Asp 826 . These two amino acids may play different roles in the binding (or coordination) of ions in the ion pocket as reported for the roles of Asn 796 and Asp 800 of sarcoplasmic Ca 2ϩ -ATPase (32,33). There are two Ca 2ϩ -binding pockets, the sites I and II. Asn 796 is involved in coordination of one Ca 2ϩ in the site II, whereas Asp 800 is involved in coordination of Ca 2ϩ both in the sites I and II.
At the position of amino acid 827, a mutation of Ile to Ala abolished the K ϩ -ATPase activity and phosphorylation, whereas mutations to Val and Met partly retained the K ϩ -ATPase activity and phosphorylation. Therefore, the bulkiness of the side chain at this position is important for the phosphorylation, and small side chains cannot support the function.
Leu 833 is localized at the end of the M6 segment (at the boundary between the M6 segment and cytoplasmic loop). Mutations of Leu to Gly and Ala significantly decreased the expression level, suggesting that this residue is involved in correct membrane insertion of the ␣-subunit. No K ϩ -ATPase activity and phosphorylation were observed in these mutants. On the other hand, the Val and Met mutants were expressed at the level of 60 -80% of the wild-type and partly retained the K ϩ -ATPase activity and phosphorylation. Therefore, the bulkiness of the side chain at this position is primarily important for the stable expression of the ␣-subunit and is also likely to be involved in the phosphorylation.
Rice and MacLennan (17) performed scanning mutagenesis of amino acid residues in the transmembrane segments, including the M6 of sarcoplasmic Ca 2ϩ -ATPase. They found a motif, (E/D)GLPA(T/V), in the M4 and M6 segments, which is partly conserved in other P-type ATPases. In rabbit gastric H ϩ ,K ϩ -ATPase, the sequence in the M6 segment corresponding to this motif is 826 DIFPSV 831 , which is not well conserved. The same group also highlighted Asn 796 , Asp 800 , and Gly 801 of sarcoplasmic Ca 2ϩ -ATPase as residues important for Ca 2ϩ transport (3,4,34), and Pro 803 as a residue important for determining the affinity for Ca 2ϩ (35). The corresponding amino acid residues in rabbit gastric H ϩ ,K ϩ -ATPase are Glu 822 , Asp 826 , Ile 827 , and Pro 829 , respectively. Although Glu 822 is not conserved between H ϩ ,K ϩ -ATPase and Ca 2ϩ -ATPase, this residue is also important for determining the affinity for K ϩ in H ϩ ,K ϩ -ATPase. Asp 826 and Ile 827 are also important for K ϩ -ATPase activity and its phosphorylation step. It should be noted, however, that the requirement for the bulkiness of the side chain at the position of amino acid 827 in gastric H ϩ ,K ϩ -ATPase is apparently different or opposite from that at the position of amino acid 801 in Ca 2ϩ -ATPase. Ca 2ϩ -ATPase requires small amino acid (Gly or Ala) at the position of 801 to retain its activity; the Ca 2ϩ -transport, Ca 2ϩ -ATPase, and ATP-dependent phosphorylation activities were abolished when Gly 801 was replaced by Val (34), whereas H ϩ ,K ϩ -ATPase requires bulky amino acids such as Ile, Met, and Val at the position of 827 to retain its activity; the K ϩ -ATPase and ATP-dependent phosphorylation activities were abolished when Ile 827 was replaced by Ala (Table I). It is also noteworthy that Leu 793 of sarcoplasmic Ca 2ϩ -ATPase was not pointed out as a residue important for Ca 2ϩ transport, whereas the corresponding residue, Leu 819 , of gas- FIG. 8. A helical wheel model and putative membrane topology of the M6 segment of the H ؉ ,K ؉ -ATPase. A, amino acid residues between Cys 815 and Leu 833 are plotted on the helical wheel. Leu 819 and Glu 822 are involved in K ϩ -dependent dephosphorylation (shown by red circles). Glu 826 , Ile 827 , and Leu 833 are involved in phosphorylation (shown by blue circles). Ile 821 , Leu 823 , Thr 825 , and Pro 829 are involved in the rate of phosphorylation (shown by green circles). B, putative membrane topology of the M6 segment. Amino acid residues involved in K ϩ -dependent dephosphorylation are concentrated in the luminal half, whereas those involved in phosphorylation are concentrated on the cytoplasmic half of the M6 segment. tric H ϩ ,K ϩ -ATPase is important for the function of H ϩ ,K ϩ -ATPase, especially determining the affinity for K ϩ in the dephosphorylation step. This difference in the role of Leu residue at the corresponding positions may reflect the difference in the reaction with K ϩ between H ϩ ,K ϩ -ATPase and Ca 2ϩ -ATPase, namely, gastric H ϩ ,K ϩ -ATPase transports K ϩ whereas Ca 2ϩ -ATPase does not.
Several mutations were also introduced into amino acid residues on the M6 segment of ouabain-resistant Na ϩ ,K ϩ -ATPase ␣-subunit, and the ability of the mutants to confer ouabain resistance to ouabain-sensitive HeLa cells were studied (5,36,37). Among them, the role of Asp 804 of Na ϩ ,K ϩ -ATPase is different from that of the corresponding amino acid of gastric H ϩ ,K ϩ -ATPase, Glu 822 . Asp 804 of Na ϩ ,K ϩ -ATPase was not replaceable with other amino acids (Leu, Asn, Glu). On the other hand, Glu 822 of gastric H ϩ ,K ϩ -ATPase was replaceable with Ala and Asp whereas mutants E822L and E822Q lost K ϩ -ATPase activity (7). In this study, we found that mutant E822A of gastric H ϩ ,K ϩ -ATPase showed a very low affinity for K ϩ in the dephosphorylation step, while keeping a normal phosphorylation activity (Figs. 3B and 2B). These differences of the role between Asp 804 of Na ϩ ,K ϩ -ATPase and Glu 822 of gastric H ϩ ,K ϩ -ATPase may reflect the differences in the specificity for transporting ions and stoichiometry between these two ATPases. Asp 808 of Na ϩ ,K ϩ -ATPase ␣-subunit was not replaced with Ala or Leu, whereas alanine mutants for Thr 807 , Ser 814 , and Tyr 817 were active. So far, there have been no reports for the roles of all the amino acid residues on the M6 segment on the Na ϩ ,K ϩ -ATPase function.
In conclusion, we identified five amino acid residues important for the H ϩ ,K ϩ -ATPase function in the M6 segment. Almost all of these amino acid residues are exclusively localized on the one side of the putative ␣-helical structure. Of these five amino acids, three located in the cytoplasmic half of the M6 segment are involved in the phosphorylation step whereas two located in the luminal half are involved in the dephosphorylation step. Furthermore, mutations of Ile 821 , Leu 823 , Thr 825 , and Pro 829 to alanine significantly decreased the rate of phosphorylation.