Crystallization of the gastric H,K-ATPase

Structural analysis of 2D crystals of gastric H+,K+-ATPase in different states of the transport cycle

The H+,K+-ATPase uses ATP to pump protons across the gastric membrane. We used electron crystallography and limited trypsin proteolysis to study conformational changes in the H+,K+-ATPase. Well-ordered 2D crystals were obtained with detergent-solubilized H+,K+-ATPase at low pH in the absence of nucleotides, E1 state, and in the presence of fluoroaluminate and ADP, mimicking the E1PADP state. Projection maps obtained with frozen-hydrated two-dimensional crystals of the H+,K+-ATPase in these two states looked very similar, suggesting only small conformational changes during the transition from the E1 to the E1P x ADP state. This result differs from the X-ray crystal structures of the related ATPase SERCA, which revealed substantially different conformations in the E1 and E1P x ADP states. To further characterize the conformational changes in the H+,K+-ATPase during its transport cycle, we performed limited proteolysis with trypsin. All examined states of the H+,K+-ATPase, including the E1 and E1P x ADP states present in the 2D crystals,showed characteristic differences in the digestion patterns. While the results from the limited proteolysis experiments thus show that the H+,K+-ATPase adopts distinct conformations during different stages of the transport cycle, the projection maps indicate that the structural rearrangements in the H+,K+-ATPase are much smaller than those observed in the related SERCA ATPase.

Functional consequences of the oligomeric form of the membrane-bound gastric H,K-ATPase

Cross-linking and two-dimensional crystallization studies have suggested that the membrane-bound gastric H,K-ATPase might be a dimeric alpha,beta-heterodimer. Effects of an oligomeric structure on the characteristics of E(1), E(2), and phosphoenzyme conformations were examined by measuring binding stoichiometries of acid-stable phosphorylation (EP) from [gamma-(32)P]ATP or (32)P(i) or of binding of [gamma-(32)P]ATP and of a K(+)-competitive imidazonaphthyridine (INT) inhibitor to an enzyme preparation containing approximately 5 nmol of ATPase/mg of protein. At <10 microM MgATP, E(1)[ATP].Mg.(H(+)):E(2) is formed at a high-affinity site, and is then converted to E(1)P.Mg.(H(+)):E(2) and then to E(2)P.Mg:E(1) with luminal proton extrusion. Maximal acid-stable phosphorylation yielded 2.65 nmol/mg of protein. Luminal K(+)-dependent dephosphorylation returns this conformation to the E(1) form. At high MgATP concentrations (>0.1 mM), the oligomer forms E(2)P.Mg:E(1)[ATP].Mg.(H(+)). The sum of the levels of maximal EP formation and ATP binding was 5.3 nmol/mg. The maximal amount of [(3)H]INT bound was 2.6 nmol/mg in the presence of MgATP, Mg(2+), Mg-P(i), or Mg-vanadate with complete inhibition of activity. K(+) displaced INT only in nigericin-treated vesicles, and thus, INT binds to the luminal surface of the E(2) form. INT-bound enzyme also formed 2.6 nmol of EP/mg at high ATP concentrations by formation of E(2).Mg.(INT)(exo):E(1)[ATP].Mg.(H(+)) which is converted to E(2).Mg.(INT)(exo):E(1)P.Mg.(H(+))(cyto), but this E(1)P form was K(+)-insensitive. Binding of the inhibitor fixes half the oligomer in the E(2) form with full inhibition of activity, while the other half of the oligomer is able to form E(1)P only when the inhibitor is bound. It appears that the catalytic subunits of the oligomer during turnover in intact gastric vesicles are restricted to a reciprocal E(1):E(2) configuration.

Structure and function of the calcium pump

Active transport of cations is achieved by a large family of ATP-dependent ion pumps, known as P-type ATPases. Various members of this family have been targets of structural and functional investigations for over four decades. Recently, atomic structures have been determined for Ca2+-ATPase by X-ray crystallography, which not only reveal the architecture of these molecules but also offer the opportunity to understand the structural mechanisms by which the energy of ATP is coupled to calcium transport across the membrane. This energy coupling is accomplished by large-scale conformational changes. The transmembrane domain undergoes plastic deformations under the influence of calcium binding at the transport site. Cytoplasmic domains undergo dramatic rigid-body movements that deliver substrates to the catalytic site and that establish new domain interfaces. By comparing various structures and correlating functional data, we can now begin to associate the chemical changes constituting the reaction cycle with structural changes in these domains.

Three-dimensional structure of the porcine gastric H,K-ATPase from negatively stained crystals

A low-resolution three-dimensional model of membrane-bound H,K-ATPase from pig gastric mucosa has been reconstructed by electron microscopy and image processing of two-dimensional crystals in negative stain. The crystal formation is induced by magnesium and vanadate, which stabilize the E2 conformation of the enzyme. The unit cell, with a size of a = b = 123 A, gamma = 90 degrees, has tetragonal p4 symmetry. There are four separate alpha beta protomers within each unit cell. The high-contrast region is limited to the cytoplasmic part of the protein. The total volume of the observed asymmetric protein domain corresponds to a molecular mass of 80-90 kDa. It consists mainly of a large pear-shaped domain measuring 60 x 45 A2, with a height of 50 A as measured perpendicular to the membrane plane. A small stalk segment, 20 A in length, forms a connection to the transmembrane region.

Oligomeric regulation of gastric H+,K+-ATPase

The H+,K+-ATPase of intact gastric vesicles has two Km values for ATP hydrolysis, 7 and 80 microM. Irradiation of vesicles with ultraviolet light in the presence of 1 mM ATP resulted in K+-ATPase activity that shows only the low affinity ATP binding. The irradiation stimulated or inhibited proton uptake rate compared with control vesicles at high or low ATP concentrations, respectively. The relation between proton uptake rate and K+-ATPase activity at different ATP concentrations was linear with irradiated vesicles and nonlinear with control vesicles. These results indicate that hydrolysis at the high affinity ATP binding site regulates the energy-transport coupling in negative and positive manners at high and low ATP concentrations, respectively. The complete inhibition of K+-ATPase by a specific proton pump inhibitor E3810 (rabeprazole) (2-([4-(3-methoxypropoxy)-3-methylpyridin-2-yl]methylsulf i nyl)-1H-benzimidazole sodium salt) occurred when E3810 bound to half of the alpha-subunit of H+,K+-ATPase in unirradiated vesicles at both 200 and 10 microM ATP, whereas the complete inhibition of proton uptake occurred when E3810 bound to half or a quarter of the alpha-subunit at 200 or 10 microM ATP, respectively. These results suggest that dimeric interaction between the alpha-subunits is necessary for the enzyme activity at all ATP concentrations and that dimeric or tetrameric interaction is necessary for proton transport at high or low ATP concentrations, respectively.

The site of acid secretion in the mammalian parietal cell

Initiation of acid secretion in the gastric mucosa is accompanied by a morphological transformation in which the acid pump, the H+/K(+)-ATPase, translocates from a cytoplasmic vesicular location to the secretory surface lining the canaliculi. Associated with the morphological changes, activation of K+ and Cl- pathways are necessary to supply K+ to the extracytoplasmic face of the pump. Although the pump in the secretory membrane is known to secrete acid, it is not known whether activation of the KCl pathway occurs in the tubulovesicular membrane prior to the formation of the canaliculus, or when the pump is in the secretory membrane. The cellular site of activation of acid secretion in the rabbit gastric parietal cell was investigated using the covalent binding of [3H]omeprazole as a probe of acid secretion in rabbit gastric glands that were undergoing stimulation in vitro. This compound depends on an acidic environment for activation and covalent binding to the H+/K(+)-ATPase. Electron microscopic autoradiography showed that activation of the enzyme occurred only when it was present in the canalicular membrane and not when it was present in the cytoplasmic tubulovesicular membrane. Hence there is likely to be a physical separation of K+ and/or Cl- pathways from the ATPase in the resting cell, and stimulation of acid secretion is dependent on colocalization of these pathways in the canalicular membrane.

Two-dimensional crystals of membrane-bound gastric H,K-ATPase

Two-dimensional crystallization of membrane-bound H,K-ATPase (EC 3.6.1.36) in vesicle preparations from parietal cells of hog gastric mucosa was induced by an imidazole buffer containing Mg2+ and VO3- ions. A continuous reorganization of the protein molecules started within a few hours by the formation of linear arrays. At later stages confluent two-dimensional crystals were formed. Electron microscopy and image processing showed that these were of a single tetragonal type. The asymmetric unit consisted of one pear-shaped protein domain corresponding to a H,K-ATPase protomer. Through stain-deficient contact regions four adjacent protein units were connected forming a tetrameric structure.

Identification of H+/K(+)-ATPase alpha,beta-heterodimers

Glutaraldehyde treatment of the C12E8 solubilized H+/K(+)-ATPase crosslinks the catalytic subunit with an apparent molecular mass of 94 kDa in SDS polyacrylamide gels into two Coomassie stained particles migrating at approx. 147 and 173 kDa. The subunit composition of these particles was determined from the comparative distribution of FITC fluorescence, wheat germ agglutinin and anti-beta antibody reactivity in control and crosslinked preparations. FITC exclusively labelled the catalytic monomer of the native preparation and its fluorescence was initially distributed into two broad bands centered at approx. 147 and 173 kDa after crosslinking. These fluorescent bands coincided with the Coomassie stained particles. A glycoprotein(s) detected by wheat germ agglutinin reactivity was present in diffuse areas between 65 and 86 kDa and 95 to 134 kDa in the control preparation. This area was also labelled by the anti-beta antibodies. With crosslinking, the distribution of the wheat germ agglutinin reactive protein and anti-beta antibodies coincided with the crosslinked particles labelled by FITC. The presence of both the catalytic monomer and the beta subunit glycoprotein in the crosslinked particles indicated that these proteins were closely associated in the C12E8 solution. This suggests that the minimal structural particle of the H+/K(+)-ATPase is an alpha,beta-heterodimer.

Gastric H+,K(+)-ATPase as a therapeutic target in peptic ulcer disease

The presence of unbuffered acid appears to be an essential contributory factor in the pathogenesis of peptic ulcer disease. Treatment has concentrated therefore on the reduction of acidity, and the last decade has seen the widespread and effective use of H2 antagonists. They are, at low doses, more successful in improving the natural history of duodenal ulcer disease than of gastric or esophageal ulceration. The H2 receptor plays a central role in activation of parietal cell acid secretion, and antagonists at this receptor block most (but not all) of the acid secretion due to even gastrinergic or muscarinic (vagal) stimulation. In hypergastrinemic states such as Zollinger-Ellison syndrome, or where acid secretion has to be inhibited by more than 20% over a 24-hr period, such as for treatment of esophagitis, NSAID damage, or gastric ulcers, the dose and frequency of administration of the currently available antagonists must be increased to achieve reliable therapy. This has led to a search for an alternative target for acid inhibitory drugs, such as the gastric acid pump, the H+,K(+)-ATPase. This article focuses on the function of this ATPase and suggests that inhibition of this pump will provide a more efficacious means of reduction of acid secretion by the stomach, hence improving and simplifying therapy of acid related diseases.

Glutaraldehyde crosslinking analysis of the C12E8 solubilized H,K-ATPase

A soluble porcine H,K-ATPase preparation was obtained with the nonionic detergent, C12E8. ATP hydrolysis by the soluble H,K-ATPase was stimulated with respect to the native preparation at pH 6.1, while the K(+)-phosphatase activity was comparable to the native enzyme. The soluble enzyme demonstrated characteristic ligand-dependent effects on ATP hydrolysis, including ATP activation of K(+)-stimulated hydrolysis with a K0.5 of 28 +/- 4 microM ATP, and inhibition with an IC50 of 2.1 mM ATP. The activation and inhibition of ATP hydrolysis by K+ was also observed with a K0.5 for activation of 2.8 +/- 0.4 mM KCl at 2.0 mM ATP (pH 6.1) and inhibition with an IC50 of 135 mM KCl at 0.05 mM ATP. 2-Methyl-8-(phenylmethoxy)imidazo[1,2a]pyridine-3-acetonitrile (SCH 28080), a specific inhibitor of the native H,K-ATPase, competitively inhibited the K(+)-stimulated activity with a Ki of 0.035 microM. The soluble enzyme was stable with a t0.5 for ATPase activity of 6 h between 4 and 11 degrees C. The demonstration of these related ligand responses in the catalytic reactions of the soluble preparation indicates that it is an appropriate medium for investigation of the subunit associations of the functional H,K-ATPase. Subunit associations of the active soluble enzyme were assessed following treatment with the crosslinking reagent, glutaraldehyde. The distribution of crosslinked particles was independent of the soluble protein concentration in the crosslinking buffer within the protein range 0.3 to 2.0 mg/ml or the detergent to protein ratio varied from 1 to 15 (w/w). The crosslinked pattern was unaffected by the presence or absence of K during crosslinking or nucleotide concentration. These observations suggest that crosslinking occurs in associated subunits that do not undergo rapid associations dependent upon enzyme turnover. Phosphorylation of the soluble enzyme with 0.1 mM MgATP produced a phosphoprotein at 94 kDa. A phosphoprotein obtained after glutaraldehyde treatment exhibited identical electrophoretic mobility to the crosslinked particle identified by silver stain. Glutaraldehyde treatment of soluble protein fractions resolved on a linear 10-35% glycerol gradient revealed several smaller peptides partially resolved from the crosslinked pump particle, but no active fraction enriched in the monomeric H,K-ATPase. This data indicates that the functional porcine gastric H,K-ATPase is organized as a structural dimer.

Purification and characterization of the proton translocating plasma membrane ATPase of red beet storage tissue

Plasma membranes were prepared from red beet (Beta vulgaris L.) storage tissue by partition in an aqueous two-phase system. A highly active proton-translocating ATPase was purified from these membranes by lysophosphatidylcholine extraction and glycerol density gradient centrifugation. The ATPase activity was inhibited by vanadate or dicyclohexyl carbodiimide, but was insensitive to azide, nitrate and molybdate at concentrations which inhibit the F1ATPase, the tonoplast ATPase, and acid phosphatase. Inhibition by vanadate was consistent with a non-competitive mechanism, with Ki = 10 microM. The Km for Mg-ATP was about 1 mM, magnesium ions were required, and the activity was stimulated by KCl and by lysophosphatidylcholine. The optimal pH was 6.5. The molecular mass by gel filtration in the presence of 2 g/liter octyl glucoside was 600 kDa, while dodecyl sulfate gel electrophoresis gave a polypeptide molecular mass of 100 kDa. After blotting onto nitrocellulose, the purified enzyme did not bind concanavalin A, although a concanavalin A-binding peptide of the plasma membrane runs to nearly the same position on the gel and showed some tendency to co-purify with the ATPase. Phospholipid vesicles into which the purified ATPase had been incorporated by the freeze-thaw technique showed vanadate-sensitive, ATP-dependent proton uptake. When the ATPase was reconstituted into lipid membranes at high protein to lipid ratios and incubated with ATP, two-dimensionally crystalline arrays of protein molecules were formed.

Assembly of two-dimensional membrane crystals of Na,K-ATPase

The assembly of vanadate-induced two-dimensional membrane crystals of Na,K-ATPase was analyzed by electron microscopy and image processing. Electron micrographs of negatively stained linear arrays of protein molecules were recorded and processed by correlation averaging methods. The arrays were compared with fully developed p21 crystals of the enzyme. On the basis of similarity in protein form, symmetry, and packing arrangement it was concluded that the fully developed crystals are built of tightly packed ribbons. Assembly pathways for two-dimensional membrane crystals of Na,K-ATPase are proposed.