Molecular characterization of the in situ red cell membrane calcium pump by limited proteolysis

In inside-out red cell membrane vesicles active calcium transport and the formation of the enzyme-phosphate complex (EP) of the calcium pump were simultaneously investigated and the effects of a limited proteolytic digestion examined. In order to visualize the proteolyzed EP forms we have induced the formation of a maximum level EP from [gamma-32P]ATP in the presence of Ca2+ + La3+ and applied a good-resolution acidic discontinuous sodium dodecyl sulfate-polyacrylamide gel electrophoresis system. Proteolysis of inside-out vesicle membranes by trypsin, Pronase, papain, or chymotrypsin produces a calmodulin-like activation of the calcium pump, abolishes its calmodulin sensitivity, and decreases the original 140-kDa EP complex to a limit polypeptide of 80 kDa. Trypsin digestion produces another major intermediary fragment of 90 kDa, which is still a low-activity calmodulin-sensitive form of the pump. The red cell calcium pump is activated by trypsin both in the absence and presence of Ca2+ during digestion although the rate of activation and the appearance of the 80-kDa polypeptide are enhanced by Ca2+. If proteolytic digestion is carried out by chymotrypsin, a calmodulin-insensitive maximum activation of the calcium pump coincides with the formation of a 125-130-kDa EP-forming polypeptide. Chymotrypsin and carboxypeptidase A have synergistic effects on the formation of this latter high-activity species. Based on these data we suggest a probable molecular arrangement for the functional parts of the red cell membrane calcium pump.     

Uncoupling the red cell sodium pump by proteolysis

In situ proteolysis of Na,K-ATPase was studied using inside-out red cell membrane vesicles. Proteolysis of the enzyme in its "E1" conformation with either trypsin or chymotrypsin inactivated cation translocation more than ATP hydrolysis. This was evident both in the absence of intravesicular alkali cations when Na-ATPase was compared to ATP-dependent 22Na+ influx, and in the presence of K+ when Na+/K+ exchange was compared to (Na+ + K+)-activated ATPase. This differential loss in pump versus hydrolysis was observed also when the activities of only intact, non-leaky vesicles were compared and therefore reflects intramolecular uncoupling rather than nonspecific leakage. Although oligomycin and thimerosal, like trypsin and chymotrypsin, inhibit the enzyme's conformational step(s), neither effect uncoupling. It is concluded that specific cleavage(s) of Na,K-ATPase, at least as it exists in situ, alters the reaction sequence with respect to the normal ordered mechanism. Accordingly, cytoplasmic Na+ and extracellular K+ bind to the enzyme, stimulate phosphorylation (ATP + E1----E1P + ADP) and dephosphorylation (E2P----E2 + Pi), respectively, but each is then released to the same side from which it had bound; presumably release occurs prior to the conformational transitions of E1P to E2P and E2 to E1. This conclusion is supported by experiments showing that, ar micromolar ATP concentration, the hydrolytic activity (Na-ATPase) of the trypsinized but not the unmodified enzyme is stimulated by K+, consistent with earlier experiments (Hegyvary, C., and Post, R. L. (1971) J. Biol. Chem. 246, 5234-5240) showing that the K X E2 to K X E1 transition is slower than the E2 to E1 transition.     

Functional domains of the in situ red cell membrane calcium pump revealed by proteolysis and monoclonal antibodies. Possible sites for regulation by calpain and acidic lipids

The functional domains of the in situ red cell membrane calcium pump were mapped by a double labeling technique. In inside-out vesicles (IOVs) the calcium pump was phosphorylated by [gamma-32P]ATP, the proteins blotted onto nitrocellulose and tagged by monoclonal antibodies raised against the purified pump protein. After proteolytic treatment of the IOVs by trypsin, chymotrypsin, or calpain-I, the fragmentation pattern of the enzyme was followed on the double-labeled immunoblots. The changes in the kinetics of the pump were examined by parallel measurements of the active calcium uptake in IOVs. By analysis of the results of tryptic digestion, it was possible to show that the antibodies recognized three different domains of the pump: 1) a Mr = 10,000-15,000 fragment (not seen directly) which includes the calmodulin-binding domain, 2) a nonphosphorylated Mr = 35,000 tryptic fragment, and 3) a phosphorylated fragment of Mr = 76,000-81,000. Chymotrypsin or calpain-I digestion of the membranes produced one major, Mr = 125,000 fragment, which had lost antibody-binding region 1. Production of this fragment coincided with the loss of calmodulin dependence and with a calmodulin-like activation of IOV calcium uptake (high Vmax, cooperativity in calcium activation). The Mr = 125,000 fragment was further activated by acidic lipids producing high Vmax and low K 1/2 (Ca2+) with no cooperativity. Based on these data a kinetic model and a functional map of the plasma membrane calcium pump is suggested.     

Dependence of the red blood cell calcium pump on the membrane potential

(1) It is shown that the rate of calcium extrusion from intact human red cells is faster at a membrane potential of approximately +50 mV (inside) than at approximately -50 mV. (2) The positive potential applied was the chloride potential of KCl cells in a K-gluconate medium when the Ca2+ sensitive K+ channel was blocked by 0.3mM quinidine. The negative potential resulted from the high K+ permeability in Ca2+ loaded cells (the cells were loaded to a Ca2+ activity in the cell water of about 50 microM). (3) It is further demonstrated that the Ca2+ affinity of the pump ATPase is decreased both at the internal (high affinity) and external (low affinity) site by increasing the proton concentration. Acidification thus inhibits internally and stimulates externally. (4) An indirect effect of the membrane potential on the pump activity via the accompanying pH shifts on either side of the membrane could be ruled out by choosing Ca2+ concentrations which are fully activating at the internal Ca2+ binding site at pH 6.5 and not yet inhibitory at the external Ca2+ binding site at pH 8. (5) The result is compatible with the assumption that the human red cell Ca-pump is exchanging Ca2+ for protons, yet is electrogenic by virtue of a stoichiometry of 1H+:1Ca2+ for this exchange.     

Conformational changes in human red cell membrane proteins induced by sugar binding

Summary We have previously shown that the human red cell glucose transport protein and the anion exchange protein, band 3, are in close enough contact that information can be transmitted from the glucose transport protein to band 3. The present experiments were designed to show whether information could be transferred in the reverse direction, using changes in tryptophan fluorescence to report on the conformation of the glucose transport protein. To see whether tryptophan fluorescence changes could be attributed to the glucose transport protein, we based our experiments on procedures used by Helgerson and Carruthers [Helgerson, A.L., Carruthers, A., (1987)J. Biol. Chem. 262:5464–5475] to displace cytochalasin B (CB), the specificd-glucose transport inhibitor, from its binding site on the inside face of the glucose transport protein, and we showed that these procedures modified tryptophan fluorescence. Addition of 75mm maltose, a nontransportable disaccharide which also displaces CB, caused a timedependent biphasic enhancement of tryptophan fluorescence in fresh red cells, which was modulated by the specific anion exchange inhibitor, DBDS (4,4-dibenzamido-2,2-stilbene disulfonate). In a study of nine additional disaccharides, we found that both biphasic kinetics and DBDS effects depended upon specific disaccharide conformation, indicating that these two effects could be attributed to a site sensitive to sugar conformation. Long term (800 sec) experiments revealed that maltose binding (±DBDS) caused a sustained damped anharmonic oscillation extending over the entire 800 sec observation period. Mathematical analysis of the temperature dependence of these oscillations showed that 2 m DBDS increased the damping term activation energy, 9.5±2.8 kcal mol^–1 deg^–1, by a factor of four to 39.7±5.1 kcal mol^–1 deg^–1, providing strong support for the view that signalling between the glucose transport protein and band 3 goes in both directions.     

On the substrate specificity of the red cell calcium pump

ATP-dependent active calcium transport in inside-out human red cell membrane vesicles is stimulated by magnesium essentially parallel with an increase in MgATP concentration. At a constant, low (1 μM) calcium concentration, increasing ATP and magnesium increase the maximum calcium transport rate irrespective of the constant or decreasing concentrations of CaATP present. K_Ca for calcium pumping is practically unchanged at variable ATP and magnesium concentrations. Free magnesium above 1–2 mM inhibits active calcium transport, probably through a direct interaction with the transport enzyme. Based on the experimental findings reported we suggest that the true, physiological substrate of the red cell calcium pump is MgATP.     

The targeting of the plasma membrane calcium pump in the cell

The information on the structural determinants that control the cellular distribution of P-type pumps is very scarce. However, recent experiments on the membrane targeting of the plasma membrane Ca²⁺ pump (PMCA) have provided interesting leads on the problem: they will be discussed in this succinct review. A general introduction on the biochemical properties of the PMCA pump will preface the discussion of the specific findings on the role of three distinct regions of the molecule in the targeting process.     

Molecular properties of the red cell calcium pump: I. Effects of calmodulin,proteolytic digestion and drugs on the kinetics of active calcium uptake in inside-out red cell membrane vesicles

In calmodulin-stripped inside-out human red cell membrane vesicles /IOV/ ATP + Mg²⁺-dependent active calcium uptake is stimulated by the addition of calmodulin. Calmodulin increases the maximum calcium transport rate /V_max/, decreases K_Ca, and does not affect K_ATP of calcium uptake. The action of both membrane bound and external calmodulin is competitively inhibited by phenothiazines. Drugs reacting with SH groups of proteins reversibly inhibit calcium pumping by decreasing V_max and not affecting K_Ca and K_ATP. The relative magnitude of calmodulin stimulation of calcium transport is unaltered by SH reagents.Mild proteolytic digestion of IOVs stimulates active calcium uptake and mimics the effects of calmodulin on the kinetic parameters — that is converts the system to a “high calcium-affinity” state. Proteolysis eliminates calcium-dependent calmodulin binding to IOV membranes and any further stimulation of calcium uptake by calmodulin. Based on these results the presence of a calmodulin-binding regulatory subunit of the red cell calcium pump at the internal membrane surface is postulated.     

Molecular properties of the red cell calcium pump: II. Effects of calmodulin,proteolytic digestion and drugs on the calcium-induced membrane phosphorylation by ATP in inside-out red cell membrane vesicles

In inside-out human red cell membrane vesicles /IOV/, in the absence of Mg²⁺, the only calcium-induced labelling by γ³²P-ATP occurs in a 140–150 000 molecular weight protein fraction, representing the hydroxylamine-sensitive phosphorylated intermediate /EP/ of the calcium pump. In the presence of Mg²⁺ calcium-induced phosphorylation is accelerated but several other membrane proteins are also phosphorylated through protein kinase action forming hydroxylamine-insensitive bonds. Addition of calmodulin accelerates EP formation both in the absence and presence of Mg²⁺.Treatment of the membrane with SH-group reagents significantly reduces EP formation. Mild trypsin digestion of IOVs, stimulating active calcium transport, eliminates calmodulin action and decreases the steady-state level of EP. In trypsin-digested IOVs the molecular weight of the ³²P-labelled EP is shifted to lower values /110–120 000/ We suggest that trypsin digestion cleaves off a 20–40 000 molecular weight calmodulin-binding regulatory subunit of the calcium pump molecule.     

Reversal of the calcium pump in human red cells

Summary Human red cells containing low ATP and high P_i concentrations were suspended in media with and without 2mm Ca²⁺, and the incorporation of (³²P)P_i into ATP was measured. There was some incorporation whatever the medium, but in every experiment there was an extra incorporation when the cells were in the Ca²⁺-containing medium. This extra incorporation was abolished by the ionophore A23187, which collapses the Ca²⁺ concentration gradient across the membranes, or by LaCl₃, which blocks the Ca²⁺ pump. Starved and phosphate-loaded cells also show an uptake of Ca²⁺ which is not apparent in fresh cells. Results are consistent with the idea that Ca²⁺-dependent incorporation of P_i into ATP is catalyzed by the Ca²⁺ pump using energy derived from the Ca²⁺ concentration gradient.     

Effects of divalent metal ions on the calcium pump and membrane phosphorylation in human red cells

In inside-out red cell membrane vesicles ATP-dependent calcium transport is activated by the divalent metal ions Mg²⁺, Mn²⁺, Co²⁺, Ni²⁺ and Fe²⁺. This activation is based on the formation of Me²⁺-ATP complexes which can serve as energy-donor substrates for the calcium pump, and probably, satisfy the requirement for free Me²⁺ in this transport process. Higher Me²⁺ concentrations inhibit calcium transport with various efficiencies. Mn²⁺ directly competes with Ca²⁺ at the transport site, while other divalent metal ions investigated have no such effect. The formation of the hydroxylamine-sensitive phosphorylated intermediate (EP) of the red cell membrane calcium pump from [γ-³²P]ATP is induced by Ca²⁺ while rapid dephosphorylation requires the presence of Mg²⁺. At higher concentrations Mn²⁺ and Ni²⁺ inhibit predominantly the formation of EP, while Co²⁺ and Fe²⁺ block dephosphorylation. The possible sites and nature of the divalent metal interactions with the red cell calcium pump are discussed. Hydroxylamine-insensitive membrane phosphorylation in inside-out vesicles from [γ-³²P]ATP is significantly stimulated by Mn²⁺ and Co²⁺, as compared to that produced by Mg²⁺, Fe²⁺ and Ni²⁺. Part of this labelling is found in phospholipids, especially in phosphatidylinositol. The results presented for the metal dependency of protein and lipid phosphorylation in red cell membranes may help in the characterization of ATP consumptions directly related to the calcium pump and those involved in various regulatory processes.     

The maximal velocity and the calcium affinity of the red cell calcium pump may be regulated independently

The kinetics of active Ca2+ transport in inside-out red cell membrane vesicles and the Ca2+-ATPase activity of the purified Ca2+ pump were studied and the effects of calmodulin, acidic phospholipids, and controlled trypsinization were compared. In the presence of calmodulin the maximal rate and the apparent affinity of the pump for Ca2+ were greatly increased in both preparations. The lowest value of Km(Ca) was between 0.5 and 0.7 microM depending on the concentration of calmodulin and on the enzyme preparation. Positive cooperativity for Ca2+ activation with a Hill coefficient of 1.6-1.7 was observed in all cases. When acidic phospholipids (phosphatidylinositol 4-phosphate was routinely used) were added to the inside-out vesicles or to the purified enzyme, maximal transport rates equal to those obtained with calmodulin were measured but the Km(Ca) decreased to 0.25 microM and the positive cooperativity disappeared (the Hill coefficient approached 1). Highly active, calmodulin-independent proteolytic fragments of molecular mass of 81 and 76 kDa were produced with controlled trypsinization. When the trypsin treatment was directed to obtain primarily the 81-kDa fragment, the preparation showed characteristics similar to those of the intact Ca2+ pump in the presence of calmodulin; that is, the same Vmax was obtained, the Km(Ca2+) was 0.5-0.6 microM, and the Hill coefficient was about 1.6. Addition of phosphatidylinositol 4-phosphate or allowing further proteolysis to produce the 76-kDa fragment, shifted the Km(Ca) to 0.25 and reduced the Hill coefficient to 1, without changes in the maximal rate. Based on these results it is suggested that the maximal velocity and the Ca2+ affinity on the erythrocyte Ca2+ pump may be regulated independently and that independent polypeptide regions of the enzyme are involved in the regulations.     

Conformational changes of the in situ nuclear pore complex.

By bridging the double membrane separating the cell nucleus and cytoplasm, nuclear pore complexes (NPCs) are crucial pathways for the exchange of ions, proteins, and RNA between these two cellular compartments. A structure in the central lumen of the NPC, called the nuclear transport protein, central granule, or nuclear plug, appeared to gate diffusion of intermediate-sized molecules (10-40 kDa) across the nuclear membranes. Visualization of the NPC required drying and fixation of the specimen for electron and atomic force microscopy (AFM), a requirement that has raised doubts about the physiological relevance of the observation. Here we present AFM images of the outer nuclear membranes and NPCs of Xenopus laevis oocytes under more physiological conditions. Measured under a variety of Ca2+ depletion conditions, the central granule appeared to occupy and occlude the lumen of the pore in >80% of NPCs compared to <10% in controls. In a few instances images were obtained of the same NPCs as the solution was changed from control saline to store depletion conditions, and finally to store repletion conditions. We conclude that the central lumen of the nuclear pore complex undergoes a conformational change in response to depletion of nuclear cisternal Ca2+ levels.     

Kinetics of inhibition of the plasma membrane calcium pump by vanadate in intact human red cells.

The lack of specific inhibitors of the plasma membrane Ca2+ pump (PMCA) has made vanadate (VO3-), a non-specific inhibitor, an invaluable tool in the study of PMCA function. However, three important properties of vanadate as an inhibitor of the PMCA in intact cells, namely its speed of action in different experimental conditions, the reversibility of its inhibitory effects at different doses, and its dose-response, had never been characterized, despite extensive use. We report here the speed, reversibility and dose-response of PMCA inhibition by vanadate in intact human red cells. Near maximal inhibitory concentrations (1mM) in the red cell suspension blocked almost instantly the uphill Ca2+ extrusion by the PMCA, regardless of the intracellular Ca2+ concentration, cation composition of the external media, membrane potential or volume-stability of the cell. PMCA inhibition by vanadate, at concentrations of 10mM and 1mM, was not reversed by washing, resuspending, and incubating the cells for up to 2h in vanadate-free media. Vanadate inhibited PMCA-mediated Ca2+ efflux in intact red cells with a K1/2 of approximately 3 microM, a value similar to that described for the Ca2+-ATPase in isolated red cell membranes.     

The red cell membrane and its cytoskeleton.

Gel-filtration (Sephadex G-75) analysis of hepatic cytosol reveals both qualitative and quantitative sex differences in oestrogen-binding proteins. The elution profile of [³H]oestradiol-labelled cytosol shows four species of oestrogen-binding proteins (peaks I, II, IV and V) common to both sexes. The amount of [³H]oestradiol binding in peak I is equivalent in both males and females and corresponds quantitatively to the specific oestrogen receptor. The amount of binding in the remaining three peaks is greater in males than females. In addition, an oestrogen-binding protein (peak III) is present that is unique to male cytosol. Proteinase-inhibition studies demonstrate that the observed multiplicity of oestrogen-binding proteins is not an artefact of proteolytic breakdown. Sex differences in oestrogen-binding proteins are absent in immature male and female animals; the oestrogen-binding protein profile in immature rats resembles that of an adult female. Gonadectomy of adult animals does not affect the oestrogen-binding-protein profile. In contrast, neonatal (day 1) castration results in partial feminization of the characteristic oestrogen-binding protein profile seen in the adult male; the appearance of Peak III is suppressed and marked decreases in the amount of oestradiol binding occurs in the remaining peaks. Hypophysectomy of adult animals results in near abolishment of the observed sex differences; the male oestrogen-binding protein profile is partially feminized and the female profile is partially masculinized, as characterized by the appearance of [³H]oestradiol binding in the region of peak III and increased amounts of binding in peaks IV and V. The present studies demonstrate a multiplicity of oestrogen-binding proteins in liver cytosol and raise the possibility that the presence of some of these proteins may be imprinted at birth through the hypothalamic–pituitary axis, by a mechanism requiring neonatal androgen exposure.     

Irreversible effects of calcium ions on the plasma membrane calcium pump

The calcium pump of human red cells can be irreversibly activated by preincubation of the membranes in the presence of calcium ions, with a pattern reminiscent of that produced by controlled trypsin attack. With 1 mm Ca²⁺, the activity of the basal enzyme increases three to fourfold over 30 to 60 min, to levels about half those obtained in the presence of calmodulin. On the whole, the effect occurs slowly, with a very low Ca²⁺ affinity at 37°C and is unaffected by serine-protease inhibitors. The activation caused by 1 mm Ca²⁺ is little affected by leupeptin (a thiol-protease inhibitor) and that obtained at 10 m Ca²⁺ is not inhibited. Preincubations at 0°C also lead to activation, to a level up to half that seen at 37°C, and the effect is not affected by leupeptin or antipain. No activation is observed by preincubating soluble purified Ca,Mg-ATPase in Ca²⁺-containing solutions at 37°C. Instead, calcium ions protect the detergent-solubilized enzyme from thermal inactivation, the effect being half-maximal between 10 and 20 m Ca²⁺. We conclude that the activation of the membrane-bound Ca,Mg-ATPase by Ca²⁺ should result from an irreversible conformational change in the enzyme and not from attack by a membrane-bound protease, and that this change presumably arises from the release of inhibitory particles existing in the original membrane preparations.We thank The Wellcome Trust for a research grant, the Medical Research Council for an equipment grant and the Regional Transfusion Service (Sheffield) for bank blood supplies.     

Characteristics and regulation of active calcium transport in inside-out red cell membrane vesicles

Sealed, inside-out human red cell membrane vesicles, prepared by a modified method of Steck (Steck T.L. (1974) in Methods in Membrane Biology (Korn, E.D., ed.), Vol 2, pp. 245–281, Plenum Press, New York), accomplish an ATP and Mg²⁺-dependent uphill calcium uptake with a reproducible maximum rate of 12–15 nmol/mg vesicle protein per min under physiological conditions. This maximum rate is increased by about 60–70% in the presence of a heatstable cytoplasmic activator protein (calmodulin) obtained from red cells. Calcium efflux from inside-out vesicles is smaller than 0.01 nmol/mg vesicle protein per min at intravesicular calcium concentrations between 0.1 and 20.0 mM.In the presence of Mg²⁺, active calcium uptake is supported by ATP, ITP, or UTP, but not by ADP, AMP, or p-nitrophenyl phosphate. The optimum pH for the process is 7.4–7.6, and the activation energy is 19–20 kcal/mol, irrespective of the presence or absence of calmodulin. Calcium uptake in inside-out vesicles is unaffected by ouabain or oligomycin, but blocked by low concentrations of lanthanum, ruthenium red, quercetin and phloretin. K⁺ and Na⁺, when compared to choline⁺ or Li⁺, significantly increase active calcium uptake. This stimulation by K⁺ and Na⁺ is independent of that by calmodulin.Concentrated red cell cytoplasm activates calcium uptake at low soluble protein:membrane protein ratios, while a ‘deactivation’ of the transport occurs at high cytoplasm: membrane protein ratios. A heat-labile cytoplasmic protein fraction antagonizing calmodulin activation, can be separated by DEAE-Sephadex chromatography. Based on these findings the regulation of active calcium transport in human red cells is discussed.