Effect of chemical modification of cytoplasmic activator protein on human red cell membrane Ca++ + Mg+ ATPase.

A highly purified cytoplasmic activator protein of human red cell membrane Ca⁺⁺ + Mg⁺⁺ ATPase was prepared by two step purification scheme utilizing Diethylaminoethyl cellulose (DE-52) and sephadex (G-100) column chromatography. This purified protein can elicit a maximum activation of membrane Ca⁺⁺ + Mg⁺⁺ ATPase at low calcium concentrations. The stimulatory effect of this protein can be rendered totally ineffective by chemical modification with N-bromosuccinimide. The results suggest a possible role of methionine oxidation in the regulation of the Ca⁺⁺ + Mg⁺⁺ ATPase activator activity.     

A protein activator of the plasma membrane Ca++-ATPase of heart sarcolemma

A detergent extract of dog or beef heart sarcolemmal vesicles was prepared and found to have a stimulatory effect on the Ca⁺⁺-ATPase of plasma membranes from human erythrocyte and cardiac sarcolemma. A procedure is described which enriches the activating fraction. The protein nature of the preparation is illustrated by its sensitivity to boiling and to the proteolytic enzyme(s) trypsin and chymotrypsin. SDS polyacrylamide gels indicate that the protein(s) involved have a molecular weight of 56 and 60 kDa. The sarcolemmal activator can stimulate the Ca⁺⁺-ATPase activity of the isolated enzyme more than 100% in the presence of saturating amounts of calmodulin. The activation is calcium dependent, being greatest at approximately 10µm Ca⁺⁺, free, but does not change theK _m for Ca⁺⁺. A possible physiological role for the activator is discussed.     

Calcium binding to the subunit c ofE. coli ATP-synthase and possible functional implications in energy coupling

The 8-kDa subunit c of theE. coli F₀ ATP-synthase proton channel was tested for Ca⁺⁺ binding activity using a⁴⁵Ca⁺⁺ ligand blot assay after transferring the protein from SDS-PAGE gels onto polyvinyl difluoride membranes. The purified subunit c binds⁴⁵Ca⁺⁺ strongly with Ca⁺⁺ binding properties very similar to those of the 8-kDa CF₀ subunit III of choloroplast thylakoid membranes. The N-terminal f-Met carbonyl group seems necessary for Ca⁺⁺ binding capacity, shown by loss of Ca⁺⁺ binding following removal of the formyl group by mild acid treatment. The dicyclohexylcarbodiimide-reactive Asp-61 is not involved in the Ca⁺⁺ binding, shown by Ca⁺⁺ binding being retained in twoE. coli mutants, Asp61Asn and Asp61Gly. The Ca⁺⁺ binding is pH dependent in both theE. coli and thylakoid 8-kDa proteins, being absent at pH 5.0 and rising to a maximum near pH 9.0. A treatment predicted to increase the Ca⁺⁺ binding affinity to its F₀ binding site (chlorpromazine photoaffinity attachment) caused an inhibition of ATP formation driven by a base-to-acid pH jump in whole cells. Inhibition was not observed when the Ca⁺⁺ chelator EGTA was present with the cells during the chlorpromazine photoaffinity treatment. An apparent Ca⁺⁺ binding constant on the site responsible for the UV plus chlorpromazine effect of near 80–100 nM was obtained using an EGTA-Ca⁺⁺ buffer system to control free Ca⁺⁺ concentration during the UV plus chlorpromazine treatment. The data are consistent with the notion that Ca⁺⁺ bound to the periplasimic side of theE. coli F₀ proton channel can block H⁺ entry into the channel. A similar effect occurs in thylakoid membranes, but the Ca⁺⁺ binding site is on the lumen side of the thylakoid, where Ca⁺⁺ binding can modulate acid-base jump ATP formation. The Ca⁺⁺ binding to the F₀ and CF₀ complexes is consistent with a pH-dependent gating mechanism for control of H⁺ ion flux across the opening of the H⁺ channel.This work was supported in part by grants from the Department of Energy and the U.S. Department of Agriculture.On leave from the Institute of Soil Science and Photosynthesis, Russian Academy of Science, Pushchino, Russia.     

Biochemical properties of isolated transverse tubular membranes

This review addresses the major biochemical and structural characteristics of isolated transverse tubule (T-tubule) membranes, including methods of isolation and morphology of purified membranes, evaluation of attendant membrane activities, including ion pumps and channels, and structural and compositional analyses of functionally relevant components. Particular emphasis is placed on the Mg²⁺-ATPase, its localization in the T-system, its unusual kinetic properties, its possible functions, and its potential regulation by diacylglycerol and other biologically-relevant lipids. Conclusions are drawn with respect to the biochemical markers characteristic of T-tubule membranes and the criteria to be applied in the assessment of isolated T-tubule membrane purity.Abbreviations CMC critical micelle concentration - Con-A concanavalin A - DHP dihydropyridine - E-C excitation-contraction - E-P phosphoenzyme - FITC fluorescein isothiocyanate - IgG immunoglobulin G - immuno-EM immunoelectron microscopic - IP₃ inositol 1,4,5-triphosphate - LPC lysophosphatidylcholine - NBD-C1 7-chloro-4-nitrobenz-2-oxa-1,3-diozole - NBD-F 7-fluoro-NBD - NEM N-ethylmaleimide - PL phospholipid - SDS-PAGE sodium dodecyl sulfate-polyacrylamide gel electrophoresis - SL sarcolemma - SR sarcoplasmic reticulum - STX saxitoxin - TTX tetrodotoxin - T-tubule transverse tubule - WGA wheatgerm agglutinin     

Technical considerations for assessing alterations in skeletal muscle sarcoplasmic reticulum Ca++-sequestration functionin vitro

A multiple measurement system for assessing sarcoplasmic reticulum (SR) Ca⁺⁺-ATPase activity and Ca⁺⁺-uptake was used to examine the effects of SR fractionation and quick freezing on rat white (WG) and red (RG) gastrocnemius muscle.In vitro measurements were performed on whole muscle homogenates (HOM) and crude microsomal fractions (CM) enriched in SR vesicles before and after quick freezing in liquid nitrogen. Isolation of the CM fraction resulted in protein yields of 0.96±0.1 and 0.99±0.1 mg/g in WG and RG, respectively. The percent Ca⁺⁺-ATPase recovery for CM compared to HOM was 14.5% (WG) and 10.1% (RG). SR Ca⁺⁺-activated Ca⁺⁺-ATPase activity was not affected by quick freezing of HOM or CM, but basal ATPase was reduced (P<0.05) in frozen HOM (5.12±0.18–3.98±0.20 mole/g tissue/min in WG and from 5.39±0.20–4.48±0.24 mole/g tissue/min in RG). Ca⁺⁺-uptake was measured at a range of physiological free [Ca⁺⁺] using the Ca⁺⁺ fluorescent dye Indo-1. Maximum Ca⁺⁺-uptake rates when corrected for initial [Ca⁺⁺]_f were not altered in HOM or CM by quick freezing but uptake between 300 and 400nM free Ca⁺⁺ was reduced (P<0.05) in quick frozen HOM (1.30±0.1–0.66±0.1 mole/g tissue/min in WG and 1.04±0.2–0.60±0.1 mole/g tissue/min in RG). Linear correlations between Ca⁺⁺-uptake and Ca⁺⁺-ATPase activity measured in the presence of the Ca⁺⁺ ionophore A23187 were r=+0.25, (P<0.05) and r=+0.74 (P<0.05) in HOM and CM preparations, respectively, and were not altered by freezing. The linear relationships between HOM and CM maximum Ca⁺⁺-uptake (r=+0.44, P<0.05) and between HOM and CM Ca⁺⁺-ATPase activity (r=+0.34, P<0.05) were also not altered by tissue freezing. These data suggest that alterations in maximal SR Ca⁺⁺-uptake function and maximal Ca⁺⁺-ATPase activity may be measured in both HOM and CM fractions following freezing and short term storage. (Mol Cell Biochem139, 41–52, 1994)     

Effects of morphine tolerance and dependence on Mg++ dependent ATPase activity of synaptic vesicles.

The effect of morphine on ATPase of synaptic plasma membranes (SPM) and synaptic vesicles isolated from the mouse brain was studied. The activity of synaptic vesicle Mg⁺⁺-dependent ATPase from mice rendered morphine tolerant and dependent by pellet implantation was 40% higher than that from placebo implanted mice. However, the activities of Mg⁺⁺-dependent ATPase and Na⁺, K⁺ activated ATPase of SPM of tolerant and nontolerant mice were not significantly different. The activity of synaptic vesicular Mg⁺⁺-dependet ATPase was dependent on the concentration of Mg⁺⁺ but not of Ca⁺⁺; maximum activity was obtained with 2 mM MgCl₂. On the other hand, Mg⁺⁺-dependent ATPase activity of SPM was dependent on both Mg⁺⁺ and Ca⁺⁺, activity being maximum using 2 mM MgCl₂ and 10^?5 M CaCl₂. It is suggested that this stimulation of ATPase activity may alter synaptic transmission and may thus be involved in some aspects of morphine tolerance and dependence.     

The effect of anti-ATPase antibodies upon the Ca++ transport of sarcoplasmic reticulum

Sheep or guinea pig antisera against the purified Ca⁺⁺ transport ATPase of sarcoplasmic reticulum inhibit Ca⁺⁺ transport due to a complement-dependent damage of the membrane, which causes massive leakage of Ca⁺⁺. The Ca⁺⁺-activated ATPase activity is only slightly affected even at ten times higher antibody concentration than that required for inhibition of Ca⁺⁺ transport. Antibodies prepared against the Ca⁺⁺ binding protein (C₁ protein) have no influence upon either ATPase activity or Ca⁺⁺ transport and ferritin-labeled anti-C₁ antibodies do not bind to microsomes.     

THE SARCOTUBULAR SYSTEM OF FROG SKELETAL MUSCLE : A Morphological and Biochemical Study

In the frog skeletal muscle cell a well defined and highly organized system of tubular elements is located in the sarcoplasm between the myofibrils. The sarcoplasmic component is called the sarcotubular system. By means of differential centrifugation it has been possible to isolate from the frog muscle homogenate a fraction composed of small vesicles, tubules, and particles. This fraction is without cytochrome oxidase activity, which is localized in the mitochondrial membranes. This indicates that the structural components of this fraction do not derive from the mitochondrial fragmentation, but probably from the sarcotubular system. This fraction, called sarcotubular fraction, has a Mg⁺⁺-stimulated ATPase activity which differs from that of muscle mitochondria in that it is 3 to 4 times higher on the protein basis as compared with the mitochondrial ATPase, and is inhibited by Ca⁺⁺ and by deoxycholate like the Kielley and Meyerhof ATPase. We therefore conclude that the "granules" of the Kielley and Meyerhof ATPase, which were shown to have a relaxing effect, are fragments of the sarcotubular system. The isolated sarcotubular fraction has a high RNA content and demonstrable activity in incorporating labeled amino acids, even in the absence of added supernatant.     

Effect of tissue specificity of brain soluble fractions on Na+, K+-ATPase activity

Previous evidence from this laboratory indicated that catecholamines and brain endogenous factors modulate Na⁺, K⁺-ATPase activity of the synaptosomal membranes. The filtration of a brain total soluble fraction through Sephadex G-50 permitted the separation of two fractions-peaks I and II-which stimulated and inhibited Na⁺, K⁺-ATPase, respectively (Rodríguez de Lores Arnaiz and Antonelli de Gomez de Lima, Neurochem. Res.11, 1986, 933). In order to study tissue specificity a rat kidney total soluble was fractionated in Sephadex G-50 and kidney peak I and II fractions were separated; as control, a total soluble fraction prepared from rat cerebral cortex was also processed. The UV absorbance profile of the kidney total soluble showed two zones and was similar to the profile of the brain total soluble. Synaptosomal membranes Na⁺, K⁺- and Mg²⁺-ATPases were stimulated 60–100% in the presence of kidney and cerebral cortex peak I; Na⁺, K⁺-ATPase was inhibited 35–65% by kidney peak II and 60–80% by brain peak II. Mg²⁺-ATPase activity was not modified by peak II fractions. ATPases activity of a kidney crude microsomal fraction was not modified by kidney peak I or brain peak II, and was slightly increased by kidney peak II or brain peak I. Kidney purified Na⁺, K⁺-ATPase was increased 16–20% by brain peak I and II fractions. These findings indicate that modulatory factors of ATPase activity are not exclusive to the brain. On the contrary, there might be tissue specificity with respect to the enzyme source.     

Effects of beta-bungarotoxin on calcium uptake by sarcoplasmic reticulum from rabbit skeletal muscle

A fluorescent chelate probe and a Millipore filtration technique have been used to study the effects of β-bungarotoxin (β-toxin) on passive and active Ca⁺⁺ uptake and ATPase in fragmented sarcoplasmic reticulum (SR) of rabbit skeletal muscle. β-Toxin at 3 × 10^?6 M did not affect ATPase activity. In the absence of ATP, β-Toxin increased the passive uptake of Ca⁺⁺; in the presence of ATP, active Ca⁺⁺ uptake was inhibited. The effect of β-toxin in SR can be detected at concentrations as low as 10^?9 M. The results suggest that β-toxin induces Ca⁺⁺ leakage in SR membranes.     

Separate effects of mercurial compounds on the ionophoric and hydrolytic functions of the (Ca+++Mg++)-ATPase of sarcoplasmic reticulum

Summary We have shown that a Ca⁺⁺-ionophore activity is present in the (Ca⁺⁺+Mg⁺⁺)-ATPase of rabbit skeletal muscle sarcoplasmic reticulum (A.E. Shamoo & D.H. MacLennan, 1974.Proc. Nat. Acad. Sci. USA 71:3522). Methylmercuric chloride inhibited the (Ca⁺⁺+Mg⁺⁺)-ATPase and Ca⁺⁺ transport, but had no effect on the activity of the Ca⁺⁺ ionophore. Mercuric chloride inhibited ATPase, transport and ionophore activity. The ATPase and transport functions were more sensitive to methylmercuric chloride than to mercuric chloride. The two functions were inhibited concomitantly by methylmercuric chloride but slightly lower concentrations of mercuric chloride were required to inhibit Ca⁺⁺ transport than were required to inhibit ATPase. Methylmercuric chloride and mercuric chloride probably inhibited ATPase and Ca⁺⁺ transport by blocking essential-SH groups. However, it appears that there are no essential-SH groups in the Ca⁺⁺ ionophore and that mercuric chloride inhibited the Ca⁺⁺ ionophore activity by competition with Ca⁺⁺ for the ionophoric site. Blockage of Ca⁺⁺ transport by mercuric chloride probably occurs both at sites of essential-SH groups and at sites of ionophoric activity. These data suggest the separate identity of the sites of ATP hydrolysis and of Ca⁺⁺ ionophoric activity.     

Coordinated, coenzyme Q reversible, 2,5-dibromothymoquionine inhibition of electron transport and ATPase in Escherichia coli.

2,6-dibromothymoquinone (DBMIB) and other coenzyme Q analogs partially inhibit electron transport and the membrane-bound Mg⁺⁺ stimulated ATPase of $\text{E. coli}$ membranes. The inhibitions by DBMIB are fully reversed by coenzyme Q₆, and other analogs show partial reversal by coenzyme Q₆. Electron transport reactions inhibited are NADH and lactate oxidase, NADH menadione reductase, lactate phenazinemethosulfate reductase and duroquinol oxidase. The concentrations of DBMIB required are similar for electron transport and ATPase inhibition and inhibitions are all increased by uncouplers. Electron transport and ATPase are not inhibited in a DBMIB insensitive mutant. Soluble ATPase extracted from the membranes does not show DBMIB inhibition under either high or low Mg⁺⁺ conditions. Lipophilic chelators show additional inhibition over DBMIB. It appears that coenzyme Q functions at three sites in $\text{E. coli}$ electron transport where ATPase activity is controlled. Coenzyme Q deficient mutants also show decreased electron transport and ATPase activity which is restored by coenzyme Q.     

Ultracytochemical study of Ca++-ATPase and K+-NPPase activities in retinal photoreceptors of the guinea pig

Summary Ca⁺⁺-ATPase activity was demonstrated histochemically at light- and electron-microscopic levels in inner and outer segments of retinal photoreceptor cells of the guinea pig with the use of a newly developed one-step lead-citrate method (Ando et al. 1981). The localization of ouabain-sensitive, K⁺-dependent p-nitrophenylphosphatase (K⁺-NPPase) activity, which represents the second dephosphorylative step of the Na⁺-K⁺-ATPase system, was studied by use of the one-step method newly adapted for ultracytochemistry (Mayahara et al. 1980). In retinal photoreceptor cells fixed for 15 min in 2% paraformaldehyde the electron-dense Ca⁺⁺-ATPase reaction product accumulated significantly on the inner membranes of the mitochondria but not on the plasmalemma or other cytoplasmic elements of the inner segments. The membranes of the outer segments remained unstained except the membrane arrays in close apposition to the retinal pigment epithelium. The cytochemical reaction was Ca⁺⁺- and substrate-dependent and showed sensitivity to oligomycin. When Mg⁺⁺-ions were used instead of Ca⁺⁺-ions, a distinct reaction was also found on mitochondrial inner membranes.In contrast to the localization of the Ca⁺⁺ -ATPase activity, the K⁺-NPPase activity was demonstrated only on the plasmalemma of the inner segments, but not on the mitochondria, other cytoplasmic elements or the outer segment membranes. This reaction was almost completely abolished by ouabain or by elimination of K⁺ from the incubation medium.Fellow of the Alexander von Humboldt Foundation, Bonn, Federal Republic of Germany     

Subcellular Distribution of Basal and Cation-sensitive ATPases in Roots of Phaseolus vulgaris

A smooth microsomal fraction isolated from homogenates of Pbaseolus vulgaris root tissue has been found to possesss a highly active basal ATPase (measured in the absence of added cations). The microsomal membranes also feature a cation-sensitive ATPase which responds to Mg²⁺, Na⁺ and K⁺, but in a manner that is highly variable with pH. In contrast, membrane fragments prepared by a technique designed to yield purified plasma membrane were capable of little or no hydrolysis of ATP either in the presence or absence of added cations. This suggests that the microsomal activity is a reflection of membrane-bound ATPase which has been derived from cytoplasmic membranes, possibly the tonoplast, rather than plasma membrane.     

Ca++- and Na+, K+-ATPase activities in the fungiform papilla of the tongue of Rana Esculenta (Anura Ranidae)

In the fungiform papilla of Rana esculenta (Anura Ranidae), the Ca⁺⁺-ATPase is mainly distributed on the basolateral membrane of the sensory area cells (i.e., neuroepithelial, supporting, and mucous cells). Apical membranes of all cells facing the surface present a slight enzymatic activity. Lateral wall cells have a strong Ca⁺⁺-ATPase activity on basolateral and apical membranes. Strong Na⁺, K⁺-ATPase activity occurs on the apical surface of neuroepithelial cells. Ca⁺⁺-ATPase activity is absent on the surface of endothelial cells of the capillaries located under the sensory area. These observations lead us to conclude that the sensory area of fungiform papilla is the selective way for calcium influx. Furthermore the absence of ATPase activity on the surface of the endothelial cells indicates that there is no functional barrier to calcium influx into capillary, and that calcium can be removed by vessels from the sensory area.     

The phospholipid requirement of the (Ca + Mg)-ATPase from human platelets

The phospholipid requirement of the (Ca²⁺ + Mg²⁺)-ATPase present in a membrane fraction from human platelets was studied using various purified phospholipases. Only those phospholipases, which hydrolyse the negatively charged phospholipids, inhibited the (Ca²⁺ + Mg²⁺)-ATPase activity. The ATPase activity could be restored by adding mixed micelles of Triton X-100 and phosphatidylserine or phosphatidylinositol. Micelles with phosphatidic acid, phosphatidylcholine, phosphatidylethanolamine or sphingomyelin could not be used for reconstitution and inhibited the activity of the native enzyme.     

Reconstitution and Characterization of a Calmodulin-Stimulated Ca-Pumping ATPase Purified from Brassica oleracea L

Purification and functional reconstitution of a calmodulin-stimulated Ca²⁺-ATPase from cauliflower (Brassica oleracea L.) is described. Activity was purified about 120-fold from a microsomal fraction using calmodulin-affinity chromatography. The purified fraction showed a polypeptide at 115 kD, which formed a phosphorylated intermediate in the presence of Ca²⁺, together with a few polypeptides with lower molecular masses that were not phosphorylated. The ATPase was reconstituted into liposomes by 3-([cholamidopropyl]-dimethylammonio-)1-propanesulfonate (CHAPS) dialysis. The proteoliposomes showed ATP-dependent Ca²⁺ uptake and ATPase activity, both of which were stimulated about 4-fold by calmodulin. Specific ATPase activity was about 5 μmol min⁻¹ (mg protein)⁻¹, and the Ca²⁺/ATP ratio was 0.1 to 0.5 when the ATPase was reconstituted with entrapped oxalate. The purified, reconstituted Ca²⁺-ATPase was inhibited by vanadate and erythrosin B, but not by cyclopiazonic acid and thapsigargin. Activity was supported by ATP (100%) and GTP (50%) and had a pH optimum of about 7.0. The effect of monovalent and divalent cations (including Ca²⁺) on activity is described. Assay of membranes purified by two-phase partitioning indicated that approximately 95% of the activity was associated with intracellular membranes, but only about 5% with plasma membranes. Sucrose gradient centrifugation suggests that the endoplasmic reticulum is the major cellular location of calmodulin-stimulated Ca²⁺-pumping ATPase in Brassica oleracea inflorescences.     

Effects of Germination on NA-K-stimulated Adenosine 5'-Triphosphatase and ATP-dependent Ion Transport of Isolated Membranes from Cotyledons

A purified membrane fraction featuring ATPase activity was isolated from cotyledon tissue of Phaseolus vulgaris at different stages of germination. The fraction is enriched in both basal and Na⁺-K⁺-stimulated ATPase and is relatively free of contamination by fragments of mitochondrial membrane and microsomes. The isolated membranes have been tentatively identified as partially purified plasma membrane.     

Beta-adrenergic receptor characteristics of postnatal rat myocardial cell preparations

Summary Primary mycolardial cell cultures and freshly isolated cardiac cells in suspension resprensent two isolated, whole cell models for investigating cellular transsarcolemmal⁴⁵Ca⁺⁺ exchange in response to a receptor-coupled stimulus. Studies were performed to characterize beta-adrenergic receptor binding, beta-adrenergic receptor mediated cellular calcium (⁴⁵Ca⁺⁺) exchange, and viability in purified primary myocardial cell cultures and freshly isolated cardiac cells in suspension obtained from 3-to 3-d-old Sprague-Dawley rats. In addition, beta-adrenergic receptor binding was characterized in whole-heart crude membrane preparations. All three preparations had saturable beta-adrenergic binding sites with the antagonist [¹²⁵I]iodopindolol ([¹²⁵I]IPIN). The suspensions had a significantly lower B _max (42±6 fmol/mg protein) than the membranes and cultures (77±8 and 95±10 fmol/mg protein, respectively). The K _D of the cultures (218±2.0 pM) was significantly higher than that for the suspensions (107 ±1.3 pM) and membranes (93±1.3 pM). Viability was significantly lower in the suspensions (57%) when compared to 94% viability in myocardial cell cultures after 3 h of incubation in Kreb's Henseleit buffer. Incubation of the cultures with 5.0×10⁻⁷ M isoproterenol resulted in a significant increase in⁴⁵Ca⁺⁺ exchange as early as 15 s. In contrast,⁴⁵Ca⁺⁺ exchange into the suspensions was not increased. Although both primary cell cultures and cardiac cells in suspension possess saturable beta-adrenergic receptors, only the monolayer cultures exhibited functional beta-adrenergic receptor-mediated⁴⁵Ca⁺⁺ exchange. Of the two intact cell models investigated, these data suggest that primary myocardial cell cultures are more suitable than cell suspensions for investigating beta-adrenergic receptor binding and functions in the postnatal rat heart. This research was supported by The University of Texas Research Institute, a grant from the Texas Advanced Research Technology Program awarded to S. W. Leslie and R. E. Wilcox, and contract 223-86-2109 from the Food and Drug Administration.     

Water-soluble proteins of the human red cell membrane

Summary Procedures were developed for preparation of red cell membranes almost free of hemoglobin but with minimal loss of membrane proteins. Two water-soluble protein fractions are described, each constituting about 25% of the ghost protein. The first is ionically bonded and can be solubilized in water rapidly at pH 7.0 and more slowly at higher ionic strength solutions, with a minimal rate at 20mm. This fraction contains four major components with molecular weights ranging from 30,000 to 48,000. The second fraction can only be solubilized at an appreciable rate if Ca⁺⁺ is absent and at higher pH (9.0). It is predominantly a single molecular weight component (150,000). It tends to aggregate at higher ionic strength and in the presence of Ca⁺⁺. Evidence is presented suggesting that the water-soluble proteins are present at the inner face of the membrane. The lipids remain in a water-insoluble residue that contains four major protein components ranging in molecular weight from 30,000 to 100,000. The latter is the predominant component. Only the residue contains the Na⁺–K⁺-activated ATPase, the cholinesterase, antigenic activity and most of the sialic acid and carbohydrate. The first water-soluble fraction contains a Mg⁺⁺-activated ATPase. The extraction of the water-soluble proteins is accompanied by anatomical changes resulting finally in the formation of small membranous vesicles.