Role of negatively charged residues in the fifth and sixth transmembrane domains of the catalytic subunit of gastric H+,K+-ATPase.

The role of six negatively charged residues located in or around the fifth and sixth transmembrane domain of the catalytic subunit of gastric H+,K+-ATPase, which are conserved in P-type ATPases, was investigated by site-directed mutagenesis of each of these residues. The acid residues were converted into their corresponding acid amides. Sf9 cells were used as the expression system using a baculovirus with coding sequences for the alpha- and beta-subunits of H+,K+-ATPase behind two different promoters. Both subunits of all mutants were expressed like the wild type enzyme in intracellular membranes of Sf9 cells as indicated by Western blotting experiments, an enzyme-linked immunosorbent assay, and confocal laser scan microscopy studies. The mutants D824N, E834Q, E837Q, and D839N showed no 3-(cyanomethyl)-2-methyl-8(phenylmethoxy)-imidazo[1, 2a]pyridine (SCH 28080)-sensitive ATP dependent phosphorylation capacity. Mutants E795Q and E820Q formed a phosphorylated intermediate, which, like the wild type enzyme, was hydroxylamine-sensitive, indicating that an acylphosphate was formed. Formation of the phosphorylated intermediate from the E795Q mutant was similarly inhibited by K+ (I50 = 0.4 mM) and SCH 28080 (I50 = 10 nM) as the wild type enzyme, when the membranes were preincubated with these ligands before phosphorylation. The dephosphorylation reaction was K+-sensitive, whereas ADP had hardly any effect. Formation of the phosphorylated intermediate of mutant E820Q was much less sensitive toward K+ (I50 = 4.5 mM) and SCH 28080 (I50 = 1.7 microM) than the wild type enzyme. The dephosphorylation reaction of this intermediate was not stimulated by either K+ or ADP. In contrast to the wild type enzyme and mutant E795Q, mutant E820Q did not show any K+-stimulated ATPase activity. These findings indicate that residue Glu820 might be involved in K+ binding and transition to the E2 form of gastric H+,K+-ATPase.

Transport ATPases are able to convert the energy from ATP into active ion. transport. ATPases of the P-type class (1) form an acid-stable phosphorylated intermediate during the cata lytic cycle. This phosphorylated intermediate contains an aspartyl phosphate residue present in a conserved domain in the This work was supported by the Netherlands Foundation for Sci entific Research, Division of Medical Sciences Grant 902-22-086. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "adver tisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Phosphorylation of this residue and ion transport are cou pled in such a way that specific binding of the cation that has to be transported to the extracellular or intravesicular medium stimulates phosphorylation, whereas binding of the cation to be transported into the cytosol stimulates dephosphorylation. The latter process has only been demonstrated unequivocally in Na+,K+-ATPase and the gastric H+,K+-ATPase. The molecu lar mechanism of the coupling between the phosphorylation process on the one side and ion binding and transport on the other side is still far from being elucidated.
It is generally assumed that polar amino acid residues pres ent in the transmembrane domains might play a key role in transmembrane ion transport. In particular, negatively charged residues like those originating from aspartate and glutamate are likely candidates for such a role (3,4). The presence of four transmembrane segments in the N-terminal part of the catalytic subunits of these proteins is generally accepted. In these four transmembrane regions there is only one conserved negatively charged amino acid residue, which might be involved in transmembrane ion transport (3,(5)(6)(7)(8). In the C-terminal part of the catalytic subunit, however, the sec ondary structure is still disputed. Most authors assume the presence of six transmembrane segments in this area, but several models with only four transmembrane segments have been proposed too (9,10). In the last transmembrane segment there is a pair of negatively charged conserved amino acid residues, but mutational studies up to now give no indication for an important role in N a+,K+-ATPase (11,12).
Most negatively charged residues are present in the fifth and sixth transmembrane segments (see Fig. 1) that are assumed to be immediately C-terminal of the large intracellular loop. This region of P-type ATPases, however, is rather peculiar. Because of the relatively large number of negatively charged and other polar residues, the hydropathy index is rather low. Moreover, this region contains a number of proline residues, which gen erally give a break in an a-helix structure. In vitro translation studies with H+,K+-ATPase (13) did not show membrane in sertion properties for the fifth and sixth transmembrane seg ments. For sarcoplasmic and endoplasmic reticulum (SERCA)1-type Ca2+-ATPase, only a stop-transfer signal was found for the fifth but not for the sixth transmembrane seg ment. No signal anchor sequence was found for either of these transmembrane segments (14). Lutsenko et al. (15) recently showed that extensive tryptic digestion of N a'\ K +-ATPase led to membrane release of a water-soluble fragment (Gln™7-Arg829), which included the putative M5-M6 hairpin. Occlusion ofK +, however, prevented the release of this fragment from the membrane. These findings suggest that (i) the putative trans membrane segments M5 and M6 might be present in the mem brane in a form different from a classical «-helix and (ii) this part of the a-subunit plays a role in cation binding and transport.
The putative M5-M6 region in gastric H *\K" h-ATPase con tains at least three conserved negatively charged residues (Glu795, Glu820, and Asp824). Moreover, quite close to the C terminus of M6 there are three other negatively charged resi dues (Glu834, Glu837, and Asp839), which are also conserved in other ATPases. Although most models for P-type ATPases (16,17) do not consider these residues as being present in trans membrane segments, some models (see Fig. 1) do (18)(19)(20)(21). One of the reasons for the discrepancy between the models for H +,K+-ATPase and other mammalian P-type ATPases is the presence of two cysteine residues in the M5-M6 region, at least one of which is the target for extracellularly applied acidactivated omeprazole, an inhibitor of gastric acid secretion (19). Moreover, there is a cytosolically located tryptic digestion site at Lys791 (19,20), of which the corresponding amino acid resi due in most models is placed within M5. This results in a more C-terminally located position of the transmembrane segments M5 and M6 in H +,K+-ATPase as compared with the original models for Ca2+-ATPase and N a+,K+-ATPase (16,17).
Despite this uncertainty, a number of site-directed mutagen esis studies aimed at elucidating the function of these nega tively charged residues in the M5-M6 region in C a2+-ATP a se of both the SERCA (3,5,(22)(23)(24) and the plasma membrane type (25) as well as in N a+,K+-ATPase (26)(27)(28)(29) have been performed recently. From these studies, several candidate amino acids for a role in transmembrane cation transport have been proposed, but a consistent model has not yet been obtained.
With the gastric H 1 ,K+-ATPase (18,21,30), only one study with mutants has been published until now (31), since functional expression of this enzyme system has only recently been success fully carried out (31)(32)(33). We report here mutational studies in which six negatively charged amino acid residues within or close to the fifth and sixth transmembrane segments of the catalytic subunit have been converted into the corresponding acid amides. The study shows that the mutation E795Q has no effect, whereas the mutation D824N, E834Q, E837Q, or D839N prevents the formation of a phosphorylated intermediate. The mutation E820Q results in a phosphorylated intermediate with a markedly reduced sensitivity toward both K+ and the specific H +,K+-ATPase inhibitor 3-(cyanomethyl)-2-methyl-8{ phenyl me thoxy)imidazo[l,2a]pyridine (SCH 28080) in the phosphorylation reac tion. The hydrolysis of the phosphointermediate is not stimulated by K+, and no K^'-stimulated ATPase activity can be determined. This emphasizes the importance of these negatively charged res idues in the function of H '\ K +-ATPase.

EXPERIMENTAL PROCEDURES
Site-directed Mutagenesis -All DNA manipulations were done ac cording to standard molecular biology techniques described by Sambrook et al. Asn, Not only a change in the sequence of rat cDNA as desired was constructed, but also a recogniz able sequence of a restriction site was either introduced or deleted. After selection, digestion with Sphl, and purification of the mutant Sphl fragment from agarose gels, this fragment was religated into the dephosphorylated pUCÏ9BglII-HKttAiS, /îAI fragment. The mutated clones were digested with Bglll and Dral, and a 3.3-kilobase pair Bgltl fragment was isolated and purified from agarose gels. The baculovirus transfer vector pAcAS3, containing the code for the /3-gal actos i dase (36), was digested with Bam HI and dephosphorylated. The 3,3-kilobase pair Bglll fragment was ligated into this vector, and the pAcASa mutants were obtained.
Generation Production of Recombinant H ,K -ATPane -Sf9 cells were grown at 27 either in 100-ml spinner flask cultures or as monolayer cultures in nö-ern* culture flasks as described bv Klaassen et al. (32). For md production of'H ' ,K ' -ATPase, the cells were infected at a multiplicity of infection of ■ ) in the presence of V t ethanol (37) with the DLZßAS« or ßgßu imitated viruses and incubated for 3 days. Occasionally the mul tiplicity of infection was varied from 0.01 to 10, and the Sft) cells were incubated up to f> days.
Con focal Laxer Scan Microscopy -St'S cells were grown on sterile microscope coversi ides in complete growth medium and infected with a multiplicity of infection of 3 at 27 C C. After infection, cells were incu bated at 27 'C for 48 h in complete growth medium with additions as indicated. Cells were washed three times for 5 min with phosphatebuffered saline (PBS; pH 7.4) followed by fixation in 1% paraformalde hyde in 0,1 M phosphate buffer (pH 7.4) for 1 h at room temperature. Further processing was done by permeabilization at -2 0 °C in 100% methanol for 5 min. Next, the coverslides were dried at room temper ature, and aspedflc binding sites for antibodies were blocked by incu bation for 3 0 -6 0 min in PBS, 0.05*2 poly oxy ethyl ene sorbi tan monolaurate (Tween 20), 1% (w/v) gelatin, and 2% fetal calf serum with gentle rocking. After washing with PBS, 0.059< Tween 20 three times for 5 min each, the cells were incubated with the polyclonal antibody HKB (38) for 3 0 -6 0 min in PBS, 0,05'# Tween 20, 1% (w/v) gelatin, 2% ■ fetal bovine serum. Free antibodies were removed by washing the cells as above. Next, the cells were incubated with a fluorescently labeled secondary antibody (Dako, Glostrup, Denmark) for 1 h in PBS, 0.05% Tween 20, 1% (w/v) gelatin, and 2% fetal bovine serum. From this moment on, the coverslides were kept in the dark as much as possible, After washing as above, the cells were desalted by a short wash with water, dehydrated with 100% methanol, dried, and mounted in 10% (w/v) Mowiol (Hoechst, Amsterdam, The Netherlands), 2.5% (w/v) NaN.j, and 25% glycerol in 0,1 M Tris/Cl (pH 8,5). They were examined on a Bio-Rad MRC1000 confocal microscope. Images were averaged over eight scans.
Preparation of S ß Membranes -The SÍ9 cells were harvested by centrifugation at 2,000 x g for 5 min. After resuspension at 0 °C in 0.25 M sucrose, 2 mM EDTA and 25 mM Hepes/Tris (pH 7.0), the membranes were sonicated three times for 15 s at 60 W (Branson Power Company, Denbury, CT), After centrifugation for 30 min at 10,000 x g, the supernatant was recentrifugated for 60 min at 100,000 x g at 4 °C. The pelleted membranes were resuspended in the above mentioned buffer and stored at -2 0 °C.
Protein Determination -Protein was determined with the modified Lowry method described by Peterson (39) using bovine serum albumin as a standard.
Quantification of the Expression Level-T h e H \ K + -ATPase «-subunit content of the membrane fraction was determined by a quantita tive enzyme-linked immunosorbent assay (40), using the monoclonal antibody 5B6 (41, 42) and using purified pig H f,K*-ATPase as reference.
Western Blotting-Protein samples from the membrane fraction were solubilized in SDS-PAGE sample buffer and separated on SDS-gels containing 10% acrylamide according to Laemmli (43), For immunoblotting, the separated proteins were transferred to Immobilon polyvinylidenedifluoride membranes. The «-and /3-subunit of H ',K +-ATPase were detected as described earlier (32), with the polyclonal antibody HKB (38) recognizing the 565-585 region of the a-subunit of H +,K+-ATPase and the monoclonal antibody 2G11 (44) evoked against the ß-subunit. of H ' ,K ' -ATPase, respectively.
K ' -ATPase Activity Asvwy-The K'-activated ATPase activity was determined with a radiochemical method (45). For this purpose, 0.6-5 /^g of Sf9 membranes were added to 100 ¡±1 of medium, which contained 10 /xM 17-{"PIATP (specific activity 100-500 m C i-m m ol1), 1.0 niM MgCL, 0.2 mM EGTA, 0.1 mM EDTA, 0.1 mM ouabain, 1 m u NaN;i, 25 niM Tris-HCl ipH 7.0), and varying concentrations of KC1. After incu bation for 30 min at 37 °C, the reaction was stopped by adding 500 ¡A of ice-cold 10% (w/v) charcoal in 6% (w/v) trichloroacetic acid, and after 10 min at 0 °C, the mixture was centrifuged for 10 s (10,000 x g). To 0.2 ml of the clear supernatant, containing the liberated inorganic phosphate (;,2P¡), 3 ml of OptiFluor (Canberra Packard, Tilburg, The Netherlands) was added, and the mixture was analyzed by liquid scintillation analysis. In general, blanks were prepared by incubating in the absence of enzyme. The ATPase activity is presented as the percentage of the activity in the absence of added K 1 (5 juM), which is 70-150 nmol of P¡ liberated per mg per h. The latter activity is endogenously present in membranes of Sí9 cells.
ATP Phosphorylation Capacity-ATP phosphorylation was deter mined as described before (37 Déphosphorylation Studies-After ATP phosphorylation as described above, the reaction mixture was diluted from 60 to 200 ¡xi with nonra dioactive ATP (final concentration 1 mM) in order to prevent i-ephosphorylation with radioactive ATP and the ligands to be tested (5 raM ADP, 10 mM K^) (46). The mixture was further incubated for 5 or 10 s at 0 °C. Thereafter, the reaction was stopped as described above, and the residual phosphorylation level was determined.
Hydroxylamine Sensitivity of the Phosphorylated Intermediate -Af ter ATP phosphorylation, the acid-denaturated membranes present on the membrane filters were washed with 0.5 m imidazole-HCl (pH 7.5). After exposure of the filter to either ice-cold 0.5 M hydroxyl amineimidazole (pH 7.5) or 0.5 M imidazole-HCl (pH 7.5) for 10 min, the membranes were washed with 5% trichloroacetic acid in 0.1 m phos phoric acid and analyzed by liquid scintillation analysis.

RESULTS
Six baculoviruses were produced, each of which contains coding sequences for the ß-subunit as well as for a mutated a-subunit of gastric H ,_,K+-ATPase. In each of the mutants, one of the negatively charged carboxyl residues located in or around the fifth and sixth transmembrane segments of the «-subunit had been converted into an acid amide residue. These viruses were used to infect Sf9 cells. Fig, 2 shows that the «-subunit present in the membrane fractions of Sf9 cells infected with these mutated viruses has the same apparent molecular mass as the «-subunit of the enzyme of pig gastric mucosa (47). The antibody used to detect the a-subunit on the     The reduced sensitivity of m utant E820Q for SCH 28080 has another consequence. The background phosphorylation (Table  I, column 2 ) has been defined as the SCH 28080 (100 (jlm) -insensitive phosphorylation capacity. Since SCH 28080 even at 1 mM is not able to inhibit the formation of the phosphoryl ated intermediate of E820Q completely (see Fig. 7B), the back ground phosphorylation, attributed to the 140-kDa product, has been overestimated. This explains the significantly higher background phosphorylation level of mutant E820Q as compared with the wild type enzyme (Table I) and implicitly sug gests that the specific phosphorylation level of this mutant has been underestimated.
In the above mentioned studies, the effect of K+ on the specific phosphorylation capacity was investigated by preincu bation with this ion, thus preventing formation of a phospho rylated intermediate. We next prepared a phosphorylated in termediate in the absence of K+ and SCH 28080 and measured the residual amount of phosphorylated intermediate after in cubation for 5 and 10 s in the presence of either K+ or ADP. Fig.  8 shows that the phosphorylated intermediate obtained with the wild type virus is K+-sensitive as expected. There is hardly any effect of ADP on the dephosphorylation rate of this phos phorylated intermediate, in contrast to the situation with the pig enzyme, where a small effect has been observed (46). A similar behavior was found with the E795Q mutant. The E820Q mutant, however, was insensitive toward added K + up to 100 mM. The E820Q m utant also showed no sensitivity for ADP, suggesting that this mutation did not lead to a blockade of the E r P i£2-P conversion, which would have resulted in an ADP-sensitive phosphorylated intermediate. Surprisingly, the SCH 28080-insensitive phosphorylation, which in the SDS gel is responsible for the band of 140 kDa and is also present in uninfected cells, also decreased with time. The rate of dephos-  (5), and mutant E820Q (C) were phosphoryl ated at 0 °C with 0.1 i±m [y-^PlATP in the presence of 1 mM MgCl2 and 20 mM Tris-acetic acid (pH 6.0). After 10 s (t ~ 0) the incubation medium was diluted from 60 to 200 jllI with non radioactive ATP (final concentration 1 mM ) in 25 mM Tris-acetic acid (pH 6.0) in order to prevent re phosphorylation with radioactive ATP and incubated for either 5 or 10 s without further addition (squares) or with either 5 dim ADP (triangles) or 10 mM KC1 (circles). The residual phosphorylation level (E-P) was expressed as a percentage of the control and plotted as a function of time.
phorylation of this SCH 28080-insensitive phosphoprotein was increased by ADP but not by the presence of K+ (not shown).
In order to test whether the phosphorylated intermediate of the wild type enzyme and of the mutants E795Q and E820Q was an acylphosphate, the intermediate was treated with hydroxylamine, which converts the acylphosphate into a hydroxymate (51). Both the wild type enzyme and the mutants E795Q and E820Q showed similar hydroxylamine sensitivity (Pig. 9), indicating that also the K+-insensitive mutant E820Q had formed an acylphosphate as a phosphorylated intermediate. The figure also shows that the phosphorylated protein present in the membranes of uninfected Sf9 cells is at least in part an acylphosphate too. However, a further identification of the nature of this phosphorylated protein cannot be given. Since thapsigargin, vanadate, and ouabain have no effect on the level of this phosphorylated product (not shown) both SERCA-type Ca-ATPases and Na \ K 4'-ATPase can be excluded as candi dates for this 140-kDa protein.
Phosphorylation and dephosphorylation are key steps in the catalytic cycle of -ATPase. Fig. 10 shows the K+ depend ence of the overall ATPase activity of membranes of uninfected Sf9 cells and cells infected with viruses expressing the wild type enzyme and the mutants E795Q and E820Q. In this assay a relatively low ATP concentration (10 fxm) had to be used in order to obtain significant stimulation by K+. The wild type enzyme showed a biphasic activation curve with a maximum at 1 mM K +. This activation could be completely blocked by 100 ju,m SCH 28080 (not shown). A similar biphasic activation curve was found with the pig enzyme using a comparable low (5 versus 10 ¡iu) ATP concentration (45). The E820Q mutant showed, like the uninfected cells, no K+-dependent ATPase activity, Low K+ activated the ATPase activity of mutant E795Q, although the maximal level reached was less than that of the wild type enzyme, as is also the case for the phosphoryl ation capacity of this mutant (Fig. 6 ). At high K+, however, less inhibition was found with this mutant than with the wild type enzyme. The maximal ATPase activity with the wild type en zyme was only 60% above the background activity, whereas the steady-state phosphorylation level was 170% above the control level (Table I). The relatively minor increase in the ATPase assay is due to the high basal ATPase activity of the mem branes of SfD ceils. It is tempting to speculate that the high basal ATPase activity is related to the relatively high dephos phorylation rate of uninfected cells in the absence of K 1 (Fig. 9).

DISCUSSION
In the present study we converted six negatively charged amino acid residues present in the M5-M6 region of the a-subunit of gastric H +,K f -ATPase in their acid amide counterpart as a first approach to establish the importance of these gluta mate and aspartate residues for the function of this transport enzyme. The studies were carried out in the baculovirus sys tem, in which we were able to express the enzyme functionally, * by constructing viruses with the (mutated) a-subunit and the ß-subunit behind two different promoters (32,37). As a main functional parameter, we measured the presence of a phospho rylated intermediate both qualitatively by autoradiography of the 100-kDa phosphorylated intermediate and quantitatively by measuring the SCH 28080-sensitive ATP-phosphorylation level. We also determined the biosynthesis of immunoreactive a-subunit using a specific enzyme-linked immunosorbent assay (40). Although the amount of immunore active a-sub unit varied considerably from experiment to experiment, we have no indi cation that proteolytic breakdown of one of the mutants was enhanced. Moreover, we have no indication of routing problems with any of the mutants.
None of the mutants E834Q, E837Q, and D839N showed any ATP phosphorylation capacity, suggesting that each of these residues is essential for the enzyme to become phosphorylated. It might be that these residues are involved in H + binding, which is essential for ATP phosphorylation. Similar residues in other P-type ATPases have met less attention until now, due to the fact that in nearly all models these residues are located in the intracellular loop between M6 and M7 and not in trans membrane segments (see Fig. 1). Further studies on the precise location of these residues and mutational studies in other Ptype ATPases are needed to reach more definite conclusions.
The H +,K+-ATPase m utant D824N was also not active in terms of phosphorylation capacity by ATP. This Asp residue is completely conserved in N a+,K+-ATPase and C a2 -ATPases from both sarcoplasmic reticulum and plasma membrane. Mu tation of the similar Asp residue in either N a+,K+-ATPase (26,28) or plasma membrane Ca2+-ATPase (25) did not result in active enzyme either. In SERCAla Ca2+-ATPase, the similar mutant D800N did not show Ca2+ occlusion or Ca2+-induced phosphorylation by ATP (52). Thus, it is possible that this residue is involved in the binding of cations by all P-type ATPases, and thus amino acid substitutions affect phosphoryl ation from ATP.
Mutation of Glu795 into Gin results in formation of a phos-phorylated intermediate with apparently normal behavior to ward K+ and SCH 28080. The amount of this intermediate formed is smaller than that for the wild type enzyme, as might be expected from the measurement of immunoreactive «-sub unit. This might be due to the production of more inactive of-subunit. Mutational studies of the similar residue in SERCA Ca -ATPase and N a h,K+-ATPase suggest a more important role of this residue in these two P-type ATPases. The obtained results, however, markedly depend on the type of the amino acid residue chosen to replace the glutamate present in these two enzymes and possibly on the expression system used. Re placement of Glu779 into Asp or Leu in N a+,K*-ATPase ex pressed in HeLa cells did not result in active enzyme, whereas mutation of this Glu into Gin or Ala gave an active enzyme (27). Mutation of the same Glu residue into an Asp, using the baculovirus system, resulted in an active enzyme with only reduced cation affinity (29). Replacement of Glu779 by a Lys even re sulted in an increase in cation affinity in the latter system. Moreover, Glu779 in N a+,K+ -ATPase is the target for the car boxyl-specific reagent 4-(diazomethy 1)^7-(diethylamine)-coumarin (53,54), which inactivates the enzyme in a cation-protective way.
In SERCA Ca2+-ATPase, mutation of the corresponding res idue, Glu771, into Gin results in inhibition of Ca2+ transport (3) and Ca2+ occlusion (55). Phosphorylation from ATP at 2.5 mM Ca2+ and Ca2+-induced inhibition of phosphorylation by inor ganic phosphate still occurs (22). The dephosphorylation reac tion of the ADP-insensitive intermediate was blocked in this mutant. Mutation of this Glu residue by either Gly or Ala resulted in similar effects. Replacement of Glu by Lys resulted in a m utant in which Ca2 + had no effect on phosphorylation from either ATP or inorganic phosphate (56). Moreover the dephosphorylation step was not inhibited by Ca2+ in this mu tant. These experiments led Andersen to the suggestion that Glu771 might participate in countertransport of two protons/ Ca2+-ATPase cycle (56).
The most interesting m utant made in the present study is E820Q. This mutant yields a phosphorylated intermediate from ATP, but preincubation with either K+ or SCH 28080 had, in contrast to the wild type enzyme, hardly any effect on the steady-state ATP phosphorylation level. The hydrolysis of this phosphorylated intermediate was insensitive to both ADP and K+, and no K4 -activated ATPase activity could be detected in this mutant. In the pig enzyme, K+ lowers the steady-state phosphorylation level both by shifting the E 1 equilibrium to the right, which is assumed to occur through a cytosolic K+ binding site (45), and by increasing the rate of dephosphoryl ation, which occurs through an extracellular accessible K+ binding site. SCH 28080, a K+ antagonist, is assumed to sta bilize the E2 form of the enzyme, thus preventing formation of a phosphorylated intermediate (57,58). ADP stimulates hy drolysis of an P form of the enzyme (46). The K+ insensitiv ity of the phosphoenzyme of mutant E820Q is not due to inhi bition of the H^-P -> E2-P conversion, since ADP does not increase the hydrolysis rate of the phosphorylated intermedi ate. It is also not due to formation of an abnormal intermediate, since the hydroxylamine sensitivity of the phosphorylated in termediate indicates th at also in this m utant an acylphosphate had been formed. The finding that SCH 28080 does not prevent formation of a phosphorylated intermediate can be explained by assuming th a t the drug is no longer able to convert the mutated enzyme to the E.¿ form. A similar explanation can be given for the fact that vanadate does not completely preclude ATP phosphorylation. However, the mutation could also affect the binding of these drugs, The mutation of residue Glu820 thus affects an extracellularly accessible K+ binding site and possi-+ * bly also a cytosolically accessible K -site, Binding of extracellular K' to the pig enzyme results in a long range conformational change, which enhances the hydrol ysis rate of the E-P at Asp,JH *\ The results presented in this paper suggest that this process is no longer possible in mutant E820Q. This suggests that Glu820 is directly involved in K " binding. There are indications that the similar residue in other ATPases is also involved in cation binding. In addition, it is striking that in the K * -sensitive P-type ATPases this residue is an Asp or a Glu, whereas in the K* -insensitive Ca2 ' -ATPases this residue is an Asn. In Na ' ,K '-ATPase, mutation of this residue (AspM (U) into Asn or Glu resulted in an inactive enzyme as measured by the inability to confer ouabain resistance to ouabain-sensitive cells (26,28). In plasma membrane Ca2+-ATPase, mutation of this residue fAsn8'9) into alanine abol ished Ca2' * uptake and phosphorylation from ATP (25). In SERCAla Ca2 ' -ATPase, the N796A mutation resulted in the absence of ATP-dependent phosphorylation or Ca2 h occlusion, but the mutant still showed Ca2 ' -dependent inhibition of the phosphorylation from inorganic phosphate (3), This is ex plained by the assumption that for ATP-phosphorylation bind ing of two Ca2* ions is necessary, whereas binding of a single Ca24 ion can already inhibit phosphorylation from inorganic phosphate (52). All of these studies indicate that in all P-type ATPases the amino acids present on the site similar to Glu820 are involved in cation binding.
In summary, the present study emphasizes the importance of Glu820 for coupling between ATP phosphorylation and K transport in gastric H 1 ,K H -ATPase. Further studies with this enzyme are necessary to understand the structural basis for the specificity of and the kinetic differences between the vari ous P-type ATPases.