Actin—membrane interactions: Association of G-actin with the red cell membrane

Chemically tritiated actin from rabbit skeletal muscle was used to investigate the association of G-actin with the red cell membrane. The tritiated actin was shown to be identical to unmodified actin in its ability to polymerize and to activate heavy meromyosin ATPase. Using sealed and unsealed red cell ghosts we have shown that G-actin binds to the cytoplasmic but not the extracellular membrane surface of ghosts. Inside-out vesicles which have been stripped of endogenous actin and spectrin by low-ionic-strength incubation bind little G-actin. However, when a crude spectrin extract containing primarily spectrin, actin, and band 4.1 is added back to stripped vesicles, subsequent binding of G-actin can be increased up to 40-fold. Further, this crude spectrin extract can compete for and abolish G-actin binding to unsealed ghosts. Actin binding to ghosts increases linearly with added G-actin and requires the presence of magnesium. In addition, actin binding is inhibited by cytochalasin B and DNAase I. Negative staining reveals an abundance of actin filaments formed when G-actin is added to reconstituted inside-out vesicles but none when it is added to unreconstituted vesicles. These observations indicate that added G-actin binds to the red cell membrane via filament formation nucleated by some membrane component at the cytoplasmic surface.     

Active calcium ion uptake by inside-out and right side-out vesicles of red blood cell membranes

The relationship between active extrusion of Ca⁺⁺ from red cell ghosts and active uptake of Ca⁺⁺ by isolated red cell membrane fragments was investigated by studying the Ca⁺⁺ uptake activities of inside-out and right side-out vesicles. Preparations A and B which had mainly inside-out and right side-out vesicles, respectively, were isolated from red cell membranes and were compared with respect to Ca⁺⁺ adenosine triphosphatase (ATPase) and ATP-dependent Ca⁺⁺ uptake activities. Preparation A had nearly eight times more inside-out vesicles and took up eight times more ⁴⁵Ca in the presence of ATP compared to preparation B. Separation of the ⁴⁵Ca-labeled membrane vesicles by density gradient centrifugation showed that the ⁴⁵Ca label was localized to the inside-out vesicle fraction. In addition, the ⁴⁵Ca taken up in the presence of ATP was lost during a subsequent incubation in the absence of ATP. The rate of ⁴⁵Ca loss was not influenced by the presence of EGTA, but was slowed in the presence of La⁺⁸ (0.1 mM) in the efflux medium. The results presented here support the thesis that the active uptake of Ca⁺⁺ by red cell membrane fragments is due to the active transport of Ca⁺⁺ into inside-out vesicles.     

Membrane alterations in irreversibly sickled cells: Hemoglobin–membrane interaction

Irreversibly sickled cells (ISCs) are sickle erythrocytes which retain bipolar enlongated shapes despite reoxygenation and owe their biophysical abnormalities to acquired membrane alterations. Freeze-etched membranes both of ISCs produced in vitro and ISCs isolated in vivo reveal microbodies fixed to the internal (PS) surface which obscure spectrin filaments. Intramembranous particles (IMPs) on the intramembrane (PF) surface aggregate over regions of subsurface microbodies. Electron microscopy of diaminobenzidine-treated ISC ghosts show the microbodies to contain hemoglobin and/or hemoglobin derivatives. Scanning electron microscopy and freeze-etching demonstrate that membrane–hemoglobin S interaction in ISCs enhances the membrane loss by microspherulation. Membrane-bound hemoglobin is five times greater in in vivo ISCs than non-ISCs, and increases during ISC production, paralleling depletion of adenosine triphosphate. Polyacrylamide gel electrophoresis of ISC membranes shows the presence of high-molecular-weight heteropolymers in the pre–band 1 region, a decrease in band 4.1 and an increase in bands 7, 8, and globin. The role of cross-linked membrane protein polymers in the generation of ISCs is discussed and is synthesized in terms of a unified concept for the determinants of the genesis of ISCs.     

Thermal properties of young red blood cells are indicative of an age-dependent regulation of membrane-skeleton interaction

The effects of red blood cell (RBC) age on membrane thermal properties have been investigated by using a 16-nitroxide stearic acid spin probe. We detected in unfractionated and most dense cells (2% fraction of circulating cells) a thermal transition at 40 degrees C that in young cells (1% fraction) was lowered at 33-35 degrees C. Spectrin seems to be directly involved in the transition detected in both young and unfractionated cells, as showed by the disappearance of the breaks after low salt extraction of spectrin. A further indication for a role of spectrin in this transition comes from its characteristic thermal unfolding above 40 degrees C. However, young cells did not show changes either in the thermal unfolding of spectrin or in the distribution of spectrin dimer, tetramer, and high oligomeric forms. These data rule out that spectrin of young RBC is modified in its thermal properties and indicate that young cells may have a different spectrin-membrane interaction. Treatment of unfractionated ghosts with an antibody specific for a fragment of the 10K domain of protein 4.1, which is fully competent for the spectrin-actin binding, produced an evident lowering of the transition temperature. The same antibody did not affect the thermal transition of young ghosts. Our results suggest that spectrin-membrane interactions may be regulated during RBC lifespan.     

The covalent modification of spectrin in red cell membranes by the lipid peroxidation product 4-hydroxy-2-nonenal

Spectrin strengthens the red cell membrane through its direct association with membrane lipids and through protein-protein interactions. Spectrin loss reduces the membrane stability and results in various types of hereditary spherocytosis. However, less is known about acquired spectrin damage. Here, we showed that α- and β-spectrin in human red cells are the primary targets of the lipid peroxidation product 4-hydroxy-2-nonenal (HNE) by immunoblotting and mass spectrometry analyses. The level of HNE adducts in spectrin (particularly α-spectrin) and several other membrane proteins was increased following the HNE treatment of red cell membrane ghosts prepared in the absence of MgATP. In contrast, ghost preparation in the presence of MgATP reduced HNE adduct formation, with preferential β-spectrin modification and increased cross-linking of the HNE-modified spectrins. Exposure of intact red cells to HNE resulted in selective HNE-spectrin adduct formation with a similar preponderance of HNE-β-spectrin modifications. These findings indicate that HNE adduction occurs preferentially in spectrin at the interface between the skeletal proteins and lipid bilayer in red cells and suggest that HNE-spectrin adduct aggregation results in the extrusion of damaged spectrin and membrane lipids under physiological and disease conditions.     

Possible identity of a membrane-bound with a soluble cyclic AMP-independent erythorocyte protein kinase that phosphorylates spectrin

A soluble casein kinase isolated and purified to homogeneity from the human erythrocyte cytosol by phosphocellulose and Sephadex G-200 chromatographies is indistinguishable from the membrane-bound casein (spectrin_kinase according and site-specificity criteria. The soluble enzyme shows an M_r of about 30 000 by gel filtration and comigrates with the purified membrane spectrin kinase as a single polypeptide of 32 000 Da on sodium dodecyl sulfate polyacrylamide gels. The soluble kinase phosphorylates spectrin in situ in spectrin kinase-depleted ghosts and catalyzes the in vitro phosphorylation of partially dephosphorylated spectrin with saturation kinetics identical to those displayed by the membrane spectrin kinase. When component 2 of spectrin that has been phosphorylated with [γ-³²P]ATP by either the soluble or the membrane kinases was subjected to limited proteolysis, the same 21500 Da papain-generated phosphopeptide was found to have been produced by the two enzymes. The same 21 500 Da phosphopeptide was identified after papain digestion of spectrin isolated from intact cells that had been incubated with ³²P_i. However, this particular peptide was not labeled in spectrin that had been phosphorylated in vitro by the catalytic subunit of cyclic AMP-dependent protein kinase. Identical phosphopeptide patterns were obtained by gel filtration and two-dimensional peptide maps of trypsin-cleaved component 2 of spectrin that had been labeled in situ, in intact ghosts or in spectrin kinase-depleted ghosts supplemented with the soluble kinase. These findings indicate a possible identity of the soluble with the membrane-bound casein (spectrin) kinase.     

Cross-linkings between spectrin and band 3 in human erythrocyte membranes

A specific structural association between spectrin component 1 and band 3 in human erythrocyte membrane has been demonstrated by covalent cross-linkings, specific labeling, and the technique of two-dimensional gel electrophoresis. A complex of 330,000 daltons, representing 1 + 3, was produced in mildly oxidized membranes at physiologic pH and isotonic conditions but not at hypotonic conditions (< 10 mM KCl or NaCl). The yield of this complex decreased dramatically as the monovalent cation concentration decreased from 90 mM to 30 mM. The presence of Mg⁺⁺ or Ca⁺⁺ (2 mM) at low ionic strength promoted 1 + 3 cross-linking in an amount similar to that produced at isotonic conditions. The specific segment of band 3 involved in the cross-linking was also investigated by means of chymotrypsin digestion of band 3 in the intact red cells. The results showed the cross-links between spectrin component 1 and the 55,000-dalton fragment of band 3 at physiologic pH and isotonic conditions. This is consistent with the idea that band 3 is anchored on or contacted with the submembrane meshwork at the cytoplasmic membrane surface.     

Interaction between microtubule-associated protein tau and spectrin

Tau factor, one of the microtubule-associated proteins (MAPs), is shown here to bind to spectrin. Evidence for an interaction between these two proteins is provided by spectrin affinity chromatography of brain MAPs, gel overlay of electrophoresed MAPs with ¹²⁵I-labelled spectrin, incorporation of tau factor in human erythrocyte ghosts, and demonstration that tau inhibits the F-actin cross-linking activity of tetrameric spectrin.The wide distribution of both tau and spectrin-like proteins in eukaryotic cells in in favor of the possible biological significance of this interaction. The results suggest that tau could be one of the proteins involved in the concerted regulation of microtubule and actin networks in the membrane vicinity.     

Topology of amino-phospholipids in the red cell membrane

The red cell membrane has an asymmetric arrangement of phospholipids. The amino-phospholipids are localized primarily on the inner surface of the membrane and the choline phospholipids are localized to a large extent on the outer surface of the membrane. Evidence is presented based on the use of covalent chemical probes in sequence that the red cell membrane contains heterogeneous domains of PE and PS and that the transport systems for Pi and K⁺ are asymmetrically arranged. Certain amino groups of PE, PS, and/or protein localized on the outer membrane surface are involved in Pi transport and certain amino groups of PE, PS, and/or protein localized on the inner surface of the membrane are involved in K⁺ transport. Cross-linking studies with DFDNB show that the cross-linked PE-PE molecules are rich in plasmalogens. This suggests that clusters of plasmalogen forms of PE occur in the membrane. Both PE and PS are cross-linked to membrane protein. These PE and PS molecules contain 24–28% 16:0 and 18:0 fatty acids and 12% fatty aldehydes. PE and PS molecules are cross-linked to a spectrin-rich fraction. It is proposed that the binding of spectrin to membrane PE and PS may help anchor spectrin to the inner surface of the membrane and regulate shape changes in the cell. K⁺-valinomycin forms a complex with TNBS and converts it from a non-penetrating proble to a penetrating probe. Valinomycin enhances K⁺ leak and Pi leak in the red cells. SITS inhibits completely the valinomycin-induced Pi leak and inhibits partially the valinomycin induced K⁺ leak. Valinomycin and IAA have additive effects on Pi leak. Ouabin has no effect on basal or valino-mycin-induced Pi leak. These data suggest that Pi leak and K⁺ leak occur by separate transport systems. In summary, the amino-phospholipids in the red cell membrane are asymmetrically arranged; some occur in clusters and some are closely associated with membrane proteins. Amino-phospholipids also are believed to bind spectrin to the inner surface of the membrane and also may play a role in cation and anion leak.     

Long ultradian rhythms in red blood cells and ghost suspensions: Possible involvement of cell membrane

Summary Oscillations in glyceraldehyde-3-phosphate dehydrogenase (GAPD) and glucose-6-phosphate dehydrogenase (G6PD) activities were recorded in suspensions of intact human red blood cells (RBCs) exposed to various light regimens. The periods of these oscillations, defined as “long ultradian,” ranged between 13 and 18 h regardless of light regimen. The patterns of enzymatic activities were the same when assayed at each time point, in full hypotonic hemolysates, and membrane-free hemolysates. However, if hemolysates were prepared by sonication the activity pattern did not exhibit significant oscillations and the activity was higher than that recorded in hypotonic hemolysates. The observed rhythms may reflect a time-dependent attachment and detachment of enzyme molecules from cell membrane, suggesting that at the bound state the enzyme molecules are (temporarily) inactive. Oscillations with similar long ultradian periods were also observed in Ca⁺⁺ concentration of suspended RBCs and in the binding of Ca⁺⁺⁴⁵ to human RBC ghosts. Treatment of the RBCs with A₂C or Diamide before the preparation of the ghosts changed or distorted the rhythmic pattern of Ca⁺⁺⁴⁵ binding. These results point to the role of the membrane in processing the long ultradian oscillations. The relation between this type of oscillations to circadian rhythm is discussed.     

Red Blood Cell Shape and Fluctuations: Cytoskeleton Confinement and ATP Activity

We review recent theoretical work that analyzes experimental measurements of the shape and fluctuations of red blood cells. Particular emphasis is placed on the role of the cytoskeleton and cell elasticity and we contrast the situation of elastic cells with that of fluid-filled vesicles. In red blood cells (RBCs), the cytoskeleton consists of a two-dimensional network of spectrin proteins. Our analysis of the wave vector and frequency dependence of the fluctuation spectrum of RBCs indicates that the spectrin network acts as a confining potential that reduces the fluctuations of the lipid bilayer membrane. However, since the cytoskeleton is only sparsely connected to the bilayer, one cannot regard the composite cytoskeleton membrane as a polymerized object with a shear modulus. The sensitivity of RBC fluctuations and shapes to ATP concentration may reflect the transient defects induced in the cytoskeleton network by ATP.     

Photodynamic effects of protoporphyrin on human erythrocytes. Nature of the cross-linking of membrane proteins

Protoporphyrin-sensitized photooxidation in human red blood cell membranes leads to severe deterioration of membrane structure and function. The membrane damage is caused by direct oxidation of amino acid residues, with subsequent cross-linking of membrane proteins. The chemical nature of these cross-links was studied in model systems, isolated spectrin and red cell ghosts. Cysteine and methionine are not involved in the cross-linking reaction. Further it could be shown that dityrosine formation, the crucial mechanism in oxidative cross-linking of proteins by peroxidase-H2O2 treatment, plays no role in photodynamic cross-linking. Experimental evidence indicated that a secondary reaction between free amino groups and a photooxidation product of histidine, tyrosine or tryptophan is involved in photodynamic cross-linking. This was deduced from the reaction observed between compounds containing a free amino group and photooxidation products of these amino acids, both in model systems, isolated spectrin and erythrocyte ghosts. In accordance, succinylation of free amino groups of membrane proteins or addition of compounds with free amino groups protected against cross-linking. Quantitative data and consideration of the reaction mechanisms of photodynamic oxidation of amino acids make it highly probable that an oxidation product of histidine rather than of tyrosine or tryptophan is involved in the cross-linking reaction, via a nucleophilic addition by free amino groups.     

Potassium-activated phosphatase from human red blood cells

Summary The cell membrane K⁺-activated phosphatase activity was measured in reconstituted ghosts of human red cells having different ionic contents and incubated in solutions of varying ionic composition. When K⁺-free ghosts are suspended in K⁺-rich media, full activation of the phosphatase is obtained. Conversely, very little ouabainsensitive activity is detected in K⁺-rich ghosts suspended in K⁺-free media. These results, together with the fact that Na⁺ competitively inhibits the effects of K⁺ only when present externally, show that the K⁺ site of the membrane phosphatase is located at the outer surface of the cell membrane. The Mg⁺⁺ requirements for K⁺ activation of the membrane phosphatase are fulfilled by internal Mg⁺⁺. Addition of intracellular Na⁺ to ATP-containing ghosts raises the apparent affinity of the enzyme for K⁺, suggesting that the sites where ATP and Na⁺ produce this effect are located at the inner surface of the cell membrane. The asymmetrical features of the membrane phosphatase are those expected from the proposed role of this enzyme in the Na⁺–K⁺-ATPase system.The authors are established investigators of the Consejo Nacional de Investigaciones Científicas y Técnicas, Argentina.     

Possible identity of a membrane-bound with a soluble cyclic AMP-independent erythrocyte protein kinase that phosphorylates spectrin

A soluble casein kinase isolated and purified to homogeneity from the human erythrocyte cytosol by phosphocellulose and Sephadex G-200 chromatographies is indistinguishable from the membrane-bound casein (spectrin) kinase according to physical and site-specificity criteria. The soluble enzyme shows an Mr of about 30000 by gel filtration and comigrates with the purified membrane spectrin kinase as a single polypeptide of 32000 Da on sodium dodecyl sulfate polyacrylamide gels. The soluble kinase phosphorylates spectrin in situ in spectrin kinase-depleted ghosts and catalyzes the in vitro phosphorylation of partially dephosphorylated spectrin with saturation kinetics identical to those displayed by the membrane spectrin kinase. When component 2 of spectrin that had been phosphorylated with [gamma-32P]ATP by either the soluble or the membrane kinases was subjected to limited proteolysis, the same 21500 Da papain-generated phosphopeptide was found to have been produced by the two enzymes. The same 21500 Da phosphopeptide was identified after papain digestion of spectrin isolated from intact cells that had been incubated with 32Pi. However, this particular peptide was not labeled in spectrin that had been phosphorylated in vitro by the catalytic subunit of cyclic AMP-dependent protein kinase. Identical phosphopeptide patterns were obtained by gel filtration and two-dimensional peptide maps of trypsin-cleaved component 2 of spectrin that had been labeled in situ, in intact ghosts or in spectrin kinase-depleted ghosts supplemented with the soluble kinase. These findings indicate a possible identity of the soluble with the membrane-bound casein (spectrin) kinase.     

Chloride permeability in human red cells: Influence of membrane protein rearrangement resulting from ATP depletion and calcium accumulation

Summary As 15% of band 3 protein, the assumed chloride channel, is associated with spectrin, the major peripheral protein of a lattice located at the red cell membrane-cytosol interface, the present study was undertaken to evaluate whether a rearrangement of the lattice modifies the functional property of band 3 protein. Such a rearrangement was modulated by depletion of cell ATP and/or by accumulation of Ca²⁺ ions within the cell.ATP depletion induces an inhibition of the electroneutral one-for-one chloride exchanges. Neither the modification of red cell morphology due to ATP depletion (discocyte-echinocyte transformation) nor a direct effect of the decrease in internal ATP level can account for this inhibition. On the other hand, it seems reasonable to consider that inhibition is related to the changes in membrane protein organization (formation of heteropolymers) induced by the decrease in ATP level. But it does not appear that the degree of inhibition is modified when this altered assembly of membrane protein is stabilized by disulfide linkages.Accumulation of Ca²⁺ ions in the cell at a relatively low concentration (10m range) inhibits chloride exchange without apparent modification of the assembly of membrane proteins. This effect of calcium on chloride exchanges is speculatively denoted as a direct effect of calcium.Calcium loading of fresh red cells at higher concentrations (500 to 1000 m) obtained by use of the ionophore A23187 induces a very strong inhibition of chloride exchanges. In this case, inhibition can be reasonably accounted for by two simultaneous effects of calcium: a direct effect which explains half of the inhibition and an indirect effect due to the formation of membrane protein complexes stabilized by covalent crosslinkages (activation by Ca²⁺ ions of a transglutaminase).It is interesting to note that intracellular calcium, whatever the level, inhibits electroneutral exchanges of chloride but increases net chloride movements.     

The sulfhydryl groups of the 35,000-dalton C-terminal segment of band 3 are located in a 9000-dalton fragment produced by chymotrypsin treatment of red cell ghosts

Five sulfhydryl groups of band 3, the anion-transport protein of the red blood cell membrane, can be labeled byN-ethylmaleimide (NEM). Two of these are located in a 35,000-dalton, C-terminal segment produced by chymotrypsin treatment of cells. Extensive treatment of unsealed ghosts with chymotrypsin results in the disappearance of the 35,000-dalton segment, but its two NEM-binding sites are preserved in a 9000-dalton peptide. The latter must therefore be a proteolytic product of the larger segment. Labeling of sulfhydryl groups of band 3 by an impermeant analog of NEM occurs in inside-out, but not in right-side-out vesicles derived from red cell ghosts, supporting the conclusion that NEM-reactive sulfhydryl groups, including those in the 35,000- and 9000-dalton segments, are exposed at the cytoplasmic face of the membrane. These findings support the conclusion that the 35,000-dalton segment crosses the bilayer, and suggest that the 9000-dalton segment may be a membrane-crossing portion of the 35,000-dalton segment.     

Changes in phosoholipid susceptibility toward phospholipases induced by ATP depletion in avian and amphibian erythrocyte membranes.

About half of the sphingomyelin content of fresh and ATP-depleted chicken erythrocytes is hydrolysed by sphingomyelinase. Removal of spingomyelin exposes the rest of the membrane phospholipids to hydrolysis by phospholipase C only in ATP-depleted but not in fresh cells. Addition of both sphinogomyelinase and phospholipase C to ATP-depleted cells causes about 60-70 percent hydrolysis of the total phospholipids accompanied by extensive (90 percent) hemolysis. The phospholipids of toad erythrocytes are partially available to phospholipase C activity in fresh cells (17-25 percent hydrolysis) without prior sphingomyelinase treatment. However, in ATP-depleted toad cells phospholipase C hydrolyses 66 percent of phospholipids and causes extensive lysis. Treatment of either fresh or ATP-depleted toad erythrocytes by sphingomyelinase together with phospholipase C induces hydrolysis of most of the phospholipds with complete lysis. Restoration of ATP to ATP-depleted cells endows them with resistance to the attack of phospholipase C. The correlation between changes in ATP level and membrane organization as revealed by increased susceptibility toward phospholipases is discussed.     

Role of spectrin in cross bonding of the red cell membrane

Membrane cross bonding--an adhesion between opposing areas of the cytoplasmic face of the red cell membrane--was achieved by treating red cells with heat, diamide, N-ethymaleimide, urea, or by ATP depletion in conjunction with cell shrinking. Membrane cross bonding could be recognized by the shape of the cells upon swelling. Quantitated by the percentage of cross-bonded red cells the effectivity of the treatments decreased in the order given above. Cross bonding was hardly reversible by reducing the diamide-induced S-S bonds with dithioerythritol. The effect of heat and urea treatment as well as ATP depletion was partly reversible. Transmission electron micrographs of the cross-bonded region showed basically parallel membranes. The distance between the respective phospholipid bilayers varied between 40 and 120 nm from cell to cell. Hb-free ghosts prepared from diamide-treated red cells could also be cross bonded. The following conclusions are drawn: spectrin provides the molecular cross link in membrane cross bonding. Aggregation and enrichment of spectrin in the cross-bonded region are probably involved in membrane cross bonding.     

On the mechanism of ATP-induced shape changes in human erythrocyte membranes. I. The role of the spectrin complex

Human erythrocyte ghosts have been shown, by scanning electron microscopy, to undergo ATP-dependent shape changes. Under appropriate conditions the ghosts prepared from normal disk-shaped intact cells adopt a highly crenated shape, which in the presence of Mg-ATP at 37 degrees C is slowly converted to the disk shape and eventually to the cup shape. These changes are not observed with other nucleotides or with 5'-adenylyl imidodiphosphate. Anti-spectrin antibodies, incorporated along with the Mg-ATP into the ghosts in amounts less than equivalent to the spectrin, markedly accelerate the shape changes observed with the Mg-ATP alone. The Fab fragments of these antibodies, however, have no effect. The conclusion is that the structural effect produced by the ATP is promoted by the cross-linking of spectrin by its antibodies, and may therefore itself be some kind of polymerization or network formation involving the spectrin complex on the cytoplasmic face of the membrane. The factors that contribute to the shape of the ghost and of the intact erythrocyte are discussed in the light of these findings.     

Association of spectrin with its membrane attachment site restricts lateral mobility of human erythrocyte integral membrane proteins

Interactions between spectrin and the inner surface of the human erythrocyte membrane have been implicated in the control of lateral mobility of the integral membrane proteins. We report here that incubation of “leaky” erythrocytes with a water-soluble proteolytic fragment containing the membrane attachment site for spectrin achieves a selective and controlled dissociation of spectrin from the membrane, and increases the rate of lateral mobility of fluorescein isothiocyanate-labeled integral membrane proteins (> 70% of label in band 3 and PAS-1). Mobility of membrane proteins is measured as an increase in the percentage of uniformly fluorescent cells with time after fusion of fluorescent with nonfluorescent erythrocytes by Sendai virus. The cells are permeable to macromolecules since virus-fused erythrocytes lose most of their hemoglobin. The membrane attachment site for spectrin has been solubilized by limited proteolysis of inside-out erythrocyte vesicles and has been purified (V). Bennett, J Biol Chem 253:2292 (1978). This 72,000-dalton fragment binds to spectrin in solution, competitively inhibits association of ³²P-spectrin with inside-out vesicles with a K_i of 10^?7M, and causes rapid dissociation of ³²P-spectrin from vesicles. Both acid-treated 72,000-dalton fragment and the 45,000 dalton-cytoplasmic portion of band 3, which also was isolated from the proteolytic digest, have no effect on spectrin binding, release, or membrane protein mobility. The enhancement of membrane protein lateral mobility by the same polypeptide that inhibits binding of spectrin to inverted vesicles and displaces spectrin from these vesicles provides direct evidence that the interaction of spectrin with protein components in the membrane restricts the lateral mobility of integral membrane proteins in the erythrocyte.