Composition of the ternary protein complex of the red cell membrane cytoskeleton

The red cell membrane skeletal network is constructed from actin, spectrin and protein 4.1 in a molar ratio of actin subunits/spectrin heterodimer/protein 4.1 of 2:1:1. This represents saturation of the actin filaments, since incubation with extraneous spectrin and protein 4.1 leads to no binding of additional spectrin, either to the inner surface of ghost membranes or to lipid-free membrane cytoskeletons. Partial extraction of spectrin from the membrane is accompanied by release of actin under all conditions. Regardless of the proportion of spectrin extracted, the molar ratio of spectrin dimers/actin subunits is constant at 1:2. This is not the result of release or cooperative breakdown of whole lattice junctions from the network, for the number of actin filaments, judged by capacity to nucleate polymerisation of added G-actin, remains unchanged even when as much as 60% of the total spectrin has been lost. A similar 1:2:1 stoichiometry characterises the complex formed when G-actin is allowed to polymerise in the presence of varying amounts of spectrin and protein 4.1. When this complex is treated with the depolymerising agent, 1 M guanidine hydrochloride, it breaks down into smaller units of the same stoichiometry. After cross-linking these can be recovered from a gel-filtration column. Complexes prepared starting from G-actin appear to be much more stable than those formed when spectrin and protein 4.1 are bound to F-actin.     

Spectrin promotes the association of F-actin with the cytoplasmic surface of the human erythrocyte membrane

We studied the binding of actin to the erythrocyte membrane by a novel application of falling ball viscometry. Our approach is based on the notion that if membranes have multiple binding sites for F-actin they will be able to cross-link and increase the viscosity of actin. Spectrin- and actin-depleted inside-out vesicles reconstituted with purified spectrin dimer or tetramer induce large increases in the viscosity of actin. Comparable concentrations of spectrin alone, inside-out vesicles alone, inside-out vesicles plus heat-denatured spectrin dimmer or tetramer induce large increases in the viscosity of actin. Comparable concentrations of spectrin alone, inside-out vesicles alone, inside-out plus heat denatured spectrin, ghosts, or ghosts plus spectrin have no effect on the viscosity of actin. Centrifugation experiments show that the amount of actin bound to the inside-out vesicles is enhanced in the presence of spectrin. The interactions detected by low-shear viscometry reflect actin interaction with membrane- bound spectrin because (a) prior removal of band 4.1 and ankyrin (band 2.1, the high- affinity membrane attachment site for spectrin) reduces both spectrin binding to the inside-out vesicles and their capacity to stimulate increase in viscosity of actin in the presence of spectrin + actin are inhibited by the addition of the water-soluble 72,000- dalton fragment of ankyrin, which is known to inhibit spectrin reassociation to the membrane. The increases in viscosity of actin induced by inside-out vesicles reconstituted with purified spectrin dimer or tetramer are not observed when samples are incubated at 0 degrees C. This temperature dependence may be related to the temperature-dependent associations we observe in solution studies with purified proteins: addition of ankyrin inhibits actin cross-linking by spectrin tetramer plus band 4.1 at 0 degrees C, and enhances it at 32 degrees C. We conclude (a) that falling ball viscometry can be used to assay actin binding to membranes and (b) that spectrin is involved in attaching actin filaments or oligomers to the cytoplasmic surface of the erythrocyte membrane.     

Spectrin-dependent and -independent association of F-actin with the erythrocyte membrane

Binding of F-actin to spectrin-actin-depleted erythrocyte membrane inside-out vesicles was measured using [3H]F-actin. F-actin binding to vesicles at 25 degrees C was stimulated 5-10 fold by addition of spectrin dimers or tetramers to vesicles. Spectrin tetramer was twice as effective as dimer in stimulating actin binding, but neither tetramer nor dimer stimulated binding at 4 degrees C. The addition of purified erythrocyte membrane protein band 4.1 to spectrin- reconstituted vesicles doubled their actin-binding capacity. Trypsinization of unreconstituted vesicles that contain < 10% of the spectrin but nearly all of the band 4.1, relative to ghosts, decreased their F-actin-binding capacity by 70%. Whereas little or none of the residual spectrin was affected by trypsinization, band 4.1 was significantly degraded. Our results show that spectrin can anchor actin filaments to the cytoplasmic surface of erythrocyte membranes and suggest that band 4.1 may be importantly involved in the association.     

Separation of the lipid bilayer from the membrane skeleton during discocyte-echinocyte transformation of human erythrocyte ghosts

The membrane skeleton, a protein lattice at the internal side of the red cell membrane, is principally composed of spectrin, actin and proteins 4.1 and 4.9. We have examined negatively stained red cell ghosts and demonstrated, on an ultrastructural level, a separation of the lipid bilayer from the membrane skeleton during echinocytic transformation. The electron micrographs of discoidal red cell ghosts suspended in hypotonic buffer revealed a filamentous reticulum that uniformly laminated the entire submembrane region. transformation of the discoidal ghosts into echinocytic form, as induced by incubation in isotonic buffer, resulted in a disruption of skeletal continuity underlying the surface contour of the membrane spicule. The submembrane reticulum extended into the base and the neck of the spiny processes of the crenated ghosts but was absent at the tip of these projections. In addition, membrane vesicles without a submembrane reticulum were detected either attached to the tips of the spicules or released into the supernatant from the echinocytic ghosts. Protein analysis revealed that the released vesicles were enriched in bands 3, 4.1 and 7 and contained very little of the membrane skeletal proteins, spectrin and actin. The data indicate that during echinocyte formation, parts of the lipid bilayer physically separate from the membrane skeleton, leading to a formation of skeleton-poor lipid vesicles.     

Pyruvate kinase-deficiency anemia: membrane approach

Low ionic strength extraction (37 degrees C, 30 min) of ghosts from PK-deficient erythrocytes provided crude spectrin extract. No significant differences in the extract composition compared to normal donors were observed. The reticulocyte-dependent spectrin extractability was found among the subjects with PK-deficiency anemia. Likewise ATP-depletion affects spectrin extractability and also leads to the adsorption of cytoplasmic protein MW 50,000 to the reticulocyte membrane. The measurement of membrane fluidity using the fluorescence probe DPH did not reveal significant alterations in the moiety of integral membrane constituents.     

Study of actin filament ends in the human red cell membrane

There is conflicting evidence concerning the state of the actin protofilaments in the membrane cytoskeleton of the human red cell. To resolve this uncertainty, we have analysed their characteristics with respect to nucleation of G-actin polymerization. The effects of cytochalasin E on the rate of elongation of the protofilaments have been measured in a medium containing 0.1 M-sodium chloride and 5 mM-magnesium chloride, using pyrene-labelled G-actin. At an initial monomer concentration far above the critical concentration for the negative ("pointed") end of F-actin, high concentrations of cytochalasin reduce the elongation rate of free F-actin by about 70%. The residual rate is presumed to correspond to the elongation rate at the negative ends. By contrast, the elongation rate on red cell ghosts or cytoskeletons falls to zero, allowing for the background of self-nucleated polymerization of the G-actin. The critical concentration of the actin in the red cell membrane has been measured after elongation of the filaments by added pyrenyl-G-actin in the same solvent. It was found to be 0.07 microM, compared with 0.11 microM under the same conditions for actin alone. This is consistent with prediction for the case of blocked negative ends on the red cell actin. The rate of elongation of actin filaments, free and in the red cell membrane cytoskeleton, has been measured as a function of the concentration of an added actin-capping protein, plasma gelsolin, with a high affinity for the positive ends. The elongation rate falls linearly with increasing gelsolin concentration until it approaches a minimum when the gelsolin has bound to all positive filament ends. The elongation rate at this point corresponds to the activity of the negative ends, and its ratio to the unperturbed polymerization rate (in the absence of capping proteins) is indistinguishable from zero in the case of ghosts, but about 1 : 4 in the case of F-actin. When ATP is replaced in the system by ADP, so that the critical concentrations at the two filament ends are equalized, the difference is equally well-marked: for F-actin, the rate at the equivalence point is about 40% of that in the absence of capping protein, whereas for ghosts the nucleated polymerization rate at the equivalence point is again zero, indicating that under these conditions the negative ends contribute little or not at all to the rate of elongation.(ABSTRACT TRUNCATED AT 400 WORDS)     

Influence of the surface potential on the Michaelis constant of membrane-bound enzymes: effect of membrane solubilization

We have purified from a membrane fraction of bovine brain a calmodulin-binding protein (calspectin) that shares a number of properties with erythrocyte spectrin: It has a heterodimeric structure with M_r 240 000 and 235 000 and binds to (dimeric form) or crosslinks (tetrameric form) F-actin. We show that calspectin (tetramer) is capable of inducing the polymerization of G-actin to actin filaments by increasing nucleation under conditions where actin alone polymerizes at a much slower rate. Thus, brain calspectin behaves in the same manner as erythrocyte spectrin, supporting the idea that, in conjunction with actin oligomers it comprises the cytoskeletal meshwork underlying the cytoplasmic surface of the nerve cell.     

Protein 4.1 is involved in a structural thermotropic transition of the red blood cell membrane detected by a spin-labeled stearic acid

Proteins involved in a structural transition in red blood cell membranes detected at 8 +/- 1.5 degrees C by a stearic acid spin-label have been investigated. Calcium loading of red blood cells with ionophore A23187 caused the disappearance of the 8 degrees C transition. Protein 4.1 appears to be the most susceptible protein to Ca2+ treatment. Antibodies specific for spectrin, band 3 (43K cytoplasmic domain), and protein 4.1 have been utilized as specific probes to modify membrane thermotropic properties. The 8 degrees C transition was eliminated by anti-4.1 protein antibodies but was not modified by the other antibodies. To further characterize the protein(s) involved in the transition, ghosts were subjected to sequential extraction of skeletal proteins. The extraction of band 6, spectrin, and actin did not modify the 8 degrees C transition. In contrast, high-salt extraction (1 M KCl) of spectrin-actin-depleted vesicles, a procedure that extracts proteins 2.1 and 4.1, was able to eliminate the 8 degrees C transition. Rebinding of purified protein 4.1 to the high salt extracted vesicles restored the 8 degrees C transition. These results indicate the involvement of protein 4.1 in the transition and suggest a functional membrane association of this protein. The binding of protein 4.1 to the membrane seems to contribute significantly to the thermotropic properties of red blood cells.     

Intramembrane particle aggregation in erythrocyte ghosts. II. The influence of spectrin aggregation.

Physicochemical properties of mixtures of spectrin and actin extracted from human erythrocyte ghosts have been correlated with ultrastructural changes observed in freeze-fractured erythrocyte membranes. (1) Extracted mixtures of spectrin and actin have a very low solubility (less than 30 mug/ml) near their isoelectric point, pH 4.8. These mixtures are also precipitated by low concentrations of Ca2+, Mg2+, polylysine or basic proteins. (2) All conditions which precipitate extracts of spectrin and actin also induce aggregation of the intramembrane particles in spectrin-depleted erythrocyte ghosts. Precipitation of the residual spectrin molecules into small patches on the cytoplasmic surface of the ghost membrane is thought to be the cause of particle aggregations, implying an association between the spectrin molecules and the intramembrane particles. (3) When fresh ghosts are exposed to conditions which precipitate extracts of spectrin and actin, only limited particle aggregation occurs. Instead, the contraction of the intact spectrin meshwork induced by the precipitation conditions compresses the lipid bilayer of the membrane, causing it to bleb off particle-free, protein-free vesicles. (4) The absence of protein in these lipid vesicles implies that all the proteins of the erythrocyte membrane are immobilized by association with either the spectrin meshwork or the intramembrane particles.     

Changes in membrane properties of rat red blood cells induced by cadmium accumulating in the membrane fraction

When rat red blood cells were incubated in a cadmium (Cd)-free medium following 1-h pretreatment with 0.5 mM CdCl2, incorporated Cd was retained in the cell during 14-h incubation and progressively accumulated in the membrane fraction, especially in the cytoskeleton fraction. In parallel to this accumulation, red cell filterability decreased and echinocytic cells increased, although intracellular ATP was maintained at the control level. The echinocytic shape was maintained in ghosts and cytoskeletons prepared from the Cd-loaded cells. In addition, the association of bands 2.1, 3, 4.2, and 4.5 with cytoskeletons increased and dissociation of cytoskeletal networks at low ionic strength decreased as the incubation time increased. Pretreatment of red blood cells with Cd also induced a release of small vesicles. These vesicles contained hemoglobin but were depleted of spectrin and actin, showing a phospholipid composition similar to that of red cell ghosts. These results suggest that the organization of cell membranes, especially cytoskeletal networks, is altered by Cd accumulation in the cytoskeleton fraction, which results in acceleration of age-related changes of red blood cells such as shape change and decreased filterability.     

Topology of the anion exchange protein AE1: the controversial sidedness of lysine 743

The topology of the band 3 (AE1) polypeptide of the erythrocyte membrane is not fully established despite extensive study. Residues near lysine 743 (K743) have been reported to be extracellular in some studies and cytoplasmic in others. In the work presented here, we have attempted to establish the sidedness of K743 using in situ proteolysis. Trypsin, papain, and proteinase K do not cleave band 3 at or near K743 in intact red cells, even under conditions that cause cleavage on the C-terminal side of the glycosylation site (N642) in extracellular loop 4. In contrast, trypsin sealed inside red cell ghosts cleaves at K743, as does trypsin treatment of inside-out vesicles (IOVs). The transport inhibitor 4,4'-diisothiocyanatodihydrostilbene-2,2'-disulfonate (H(2)DIDS), acting from the extracellular side, blocks trypsin cleavage at K743 in unsealed membranes by inducing a protease-resistant conformation. H(2)DIDS added to IOVs does not prevent cleavage at K743; therefore, trypsin cleavage at K743 in IOVs is not a consequence of cleavage of right-side-out or leaky vesicles. Finally, microsomes were prepared from HEK293 cells expressing the membrane domain of AE1 lacking the normal glycosylation site. This polypeptide does not traffic to the surface membrane; trypsin treatment of microsomes containing this polypeptide produces the 20 kDa fragment, providing further evidence that K743 is exposed at the cytoplasmic surface. Therefore, the actions of trypsin on intact cells, resealed ghosts, unsealed ghosts, inside-out vesicles, and microsomes from HEK293 cells all indicate that K743 is cytoplasmic and not extracellular.     

Actin in erythrocyte ghosts and its association with spectrin. Evidence for a nonfilamentous form of these two molecules in situ

Actin was isolated from erythrocyte ghosts. It is identical to muscle actin in its molecular weight, net charge, ability to polymerize into filaments with the double helical morphology, and its decoration with heavy meromyosin (HMM). when erythrocyte ghosts are incubated in 0.1 mM EDTA, actin and spectrin are solubilized. Spectrin has a larger molecular weight than muscle myosin. When salt is added to the EDTA extract, a branching filamentous polymer is formed. However, when muscle actin and the EDTA extract are mixed together in the presence of salt, the viscosity achieved is less than the viscosity of the solution if spectrin is omitted. Thus, spectrin seems to inhibit the polymerization of actin. If the actin is already polymerized, the addition of spectrin increases the viscosity of the solution, presumably by cross-linking the actin filaments. The addition of HMM of trypsin to erythrocyte ghosts results in filament formation in situ. These agents apparently act by detaching erythrocyte actin from spectrin, thereby allowing the polmerization of one or both proteins to occur. Since filaments are not present in untreated erythrocyte ghosts, we conclude that erythrocyte actin and spectrin associate to form an anastomosing network beneath the erythrocyte membrane. This network presumably functions in restricting the lateral movement of membrane-penetrating particles.     

Interactions of Plasmodium falciparum erythrocyte membrane protein 3 with the red blood cell membrane skeleton

Plasmodium falciparum parasites express and traffick numerous proteins into the red blood cell (RBC), where some associate specifically with the membrane skeleton. Importantly, these interactions underlie the major alterations to the modified structural and functional properties of the parasite-infected RBC. P. falciparum Erythrocyte Membrane Protein 3 (PfEMP3) is one such parasite protein that is found in association with the membrane skeleton. Using recombinant PfEMP3 proteins in vitro, we have identified the region of PfEMP3 that binds to the RBC membrane skeleton, specifically to spectrin and actin. Kinetic studies revealed that residues 38-97 of PfEMP3 bound to purified spectrin with moderately high affinity (K(D(kin))=8.5 x 10(-8) M). Subsequent deletion mapping analysis further defined the binding domain to a 14-residue sequence (IFEIRLKRSLAQVL; K(D(kin))=3.8 x 10(-7) M). Interestingly, this same domain also bound to F-actin in a specific and saturable manner. These interactions are of physiological relevance as evidenced by the binding of this region to the membrane skeleton of inside-out RBCs and when introduced into resealed RBCs. Identification of a 14-residue region of PfEMP3 that binds to both spectrin and actin provides insight into the potential function of PfEMP3 in P. falciparum-infected RBCs.     

An examination of the soluble oligomeric complexes extracted from the red cell membrane and their relation to the membrane cytoskeleton

A part of the spectrin extracted from red cell membranes at low ionic strength occurs in the form of a high-molecular weight oligomeric complex with actin and proteins 4.1 and 4.9. When the extraction is performed at 35 degrees, the spectrin is present in this complex as the dimer, all higher forms being dissociated. We have been unable to establish any correlation between the fraction of the spectrin thus complexed and the metabolic state of the cell. At least a large part of the complex appears to be a defined monodisperse species, sedimenting at 31S. The actin is present as short protofilaments. The average number of spectrin molecules associated with each molecule of complex has been studied by cytochalasin binding and electron microscopy. The complexes present the appearance in the electron microscope of spiders, in which the legs are spectrin dimers, attached to a globular element, containing by inference, actin and proteins 4.1 and 4.9; they are active in nucleating the polymerization of G-actin. The complexes are extremely stable, being resistant to dissociation under the conditions of the deoxyribonuclease assay, even after treatment with trypsin to degrade the actin-associated proteins. It is suggested that the complexes represent intact junctions of the membrane cytoskeletal network. Relevant structural features of the network are revealed by electron microscopy. The results lead to inferences concerning the mechanism of dissociation of the network from the membrane.     

Spectrin and protein 4.1 as an actin filament capping complex

Spectrin and protein 4.1, when added to G- or F-actin, cause the formation of short filaments, as judged by the appearance of powerful nucleating activity for G-actin polymerisation. F-Actin filaments are rapidly fragmented under physiological solvent conditions. The effect of cytochalasin E on the polymerisation reaction and the extent of reduction in the critical monomer concentration of actin when spectrin and 4.1 are added suggest that these proteins form a capping system for the more slowly growing, or 'pointed' ends of actin filaments. The interaction is not affected by calcium or by 4.9, the remaining constituent of the purified red cell membrane cytoskeleton.     

Influence of calmodulin on the human red cell membrane skeleton

The calcium receptor calmodulin interacts with components of the human red cell membrane skeleton as well as with the membrane. Under physiological salt conditions, calmodulin has a calcium-dependent affinity for spectrin, one of the major components of the membrane skeleton. It is apparent from our results that calmodulin inhibits the ability of erythrocyte spectrin (when preincubated with filamentous actin) to create nucleation centers and thereby to seed actin polymerization. The gelation of filamentous actin induced by spectrin tetramers is also inhibited by calmodulin. The inhibition is calcium dependent and decreases with increasing pH, similar to the binding of calmodulin to spectrin. Direct binding studies using aqueous two-phase partition indicate that calmodulin interferes with the binding of actin to spectrin. Even in the presence of protein 4.1, which is believed to stabilize the ternary complex, calmodulin has an inhibitory effect. Since calmodulin also inhibits the corresponding activities of brain spectrin (fodrin), it appears likely that calmodulin may modulate the organization of cytoskeletons containing actin and spectrin or spectrin analogues.     

Evidence for the asymmetrical binding of p-chloromercuriphenyl sulphonate to the human erythrocyte nucleoside transporter

Nucleosides cross the human erythrocyte membrane by a facilitated-diffusion process which is selectively inhibited by nanomolar concentrations of nitrobenzylthioinosine (NBMPR). The chemical asymmetry of the transporter was investigated by studying the effects of p-chloromercuriphenyl sulphonate (PCMBS) on uridine transport and high-affinity NBMPR binding in inside-out and right-side-out membrane vesicles, unsealed erythrocyte ghosts and intact cells. PCMBS was an effective inhibitor of the transporter (50% inhibition at 30 microM), but only when the organomercurial had access to the cytoplasmic membrane surface. PCMBS inhibition of NBMPR binding to ghosts was reversed by incubation with dithiothreitol. Both uridine and NBMPR were able to protect the transporter against PCMBS inhibition.     

Plasmodium falciparum erythrocyte membrane protein 3 (PfEMP3) destabilizes erythrocyte membrane skeleton

Plasmodium falciparum erythrocyte membrane protein 3 (PfEMP3) is a parasite-derived protein that appears on the cytoplasmic surface of the host cell membrane in the later stages of the parasite's development where it associates with membrane skeleton. We have recently demonstrated that a 60-residue fragment (FIa1, residues 38-97) of PfEMP3 bound to spectrin. Here we show that this polypeptide binds specifically to a site near the C terminus of alpha-spectrin at the point that spectrin attaches to actin and protein 4.1R in forming the junctions of the membrane skeletal network. We further show that this polypeptide disrupts formation of the ternary spectrin-actin-4.1R complex in solution. Importantly, when incorporated into the cell, the PfEMP3 fragment causes extensive reduction in shear resistance of the cell. We conjecture that the loss of mechanical cohesion of the membrane may facilitate the exit of the mature merozoites from the cell.     

Ultrastructure of the human erythrocyte cytoskeleton and its attachment to the membrane

We attached paraformaldehyde-fixed human erythrocyte ghosts to coated coverslips and sheared them to expose the cytoskeleton. Quick-freeze, deep-etch, rotary-replication, or tannic acid/osmium fixation and plastic embedding revealed the cytoskeleton as a dense network of intersecting straight filaments. Previous negative stain studies on spread skeletons found 5-6 spectrin tetramers intersecting at each actin oligomer, with an estimated 250 such intersections/microns 2 of membrane. In contrast, we found 3-4 filaments at each intersection and approximately 400 intersections/microns 2 of membrane. Immunogold labeling verified that the filaments were spectrin, but their lengths (29-37 nm) were approximately one-third that of extended spectrin dimers. The length and diameter of the filaments were sufficient to accommodate spectrin dimers, but not spectrin tetramers. Our results suggest that, in situ, spectrin dimers may associate as hexamers and octamers, rather than tetramers. We present several explanations that can reconcile our observations on intact cytoskeletons with previous reports on spread material. Extracting sheared ghosts with solutions of low ionic strength removed the cytoskeleton to reveal projections from the cytoplasmic surface of the membrane. These projections contained band 3, as shown by immunogold labeling, and they aggregated to a similar extent as intramembrane particles (IMP) when the cytoskeleton was removed, suggesting a direct relationship between these structures. Quantification indicated a stoichiometry of 2 IMP for each cytoplasmic projection. Cytoplasmic projections presumably contain other proteins besides band 3 since further treatment with high ionic strength solutions extracts peripheral proteins and reduces the diameter of projections by approximately 3 nm.     

Band 3 tyrosine kinase. Association with the human erythrocyte membrane

Band 3, the anion transport protein of the human erythrocyte membrane, is known to be phosphorylated in ghosts at tyrosine 8. The band 3 tyrosine kinase is now shown to be associated with the Triton X-100 insoluble membrane skeleton but not with spectrin or actin. The kinase was reversibly dissociated from membranes and skeletons at elevated ionic strength (50% at mu = 0.15). The binding capacity of the membranes exceeded their native complement of the kinase by at least 60-fold. Prior removal of all peripheral proteins from the cytoplasmic surface of inside-out vesicles did not diminish the rebinding of the kinase, whereas prior removal of band 3 and other accessory proteins from skeletons abolished the rebinding of the kinase. An excess of glyceraldehyde-3-P dehydrogenase, which binds to band 3 in the region of the phosphate acceptor tyrosine 8, both inhibited the phosphorylation of band 3 and released the kinase into solution. Soluble 40/45-kDa chymotryptic fragments from the cytoplasmic pole of band 3 were phosphorylated at least as well as membranous band 3 and caused the release of the kinase from Triton-extracted skeletons. Membrane skeletons lacked most of the membrane band 3, but retained most of the kinase. Nevertheless, the band 3 population solubilized by Triton X-100 from prelabeled ghosts was as well phosphorylated as the population of band 3 retained by the skeletons. Furthermore, the fraction of band 3 not associated with the skeletons following Triton X-100 extraction was a good substrate for the solubilized kinase. We conclude that this tyrosine kinase is reversibly bound to the membrane through electrostatic interactions with the polyacidic sequence surrounding the phosphate accepting tyrosine 8 on band 3. The kinase appears to be preferentially linked to those band 3 molecules associated with the membrane skeleton, but it impartially phosphorylates band 3 species free in the bilayer as well as band 3 fragments in solution. The resemblance of its plasma membrane binding behavior to that of tyrosine kinases of certain viruses causing oncogenic transformation is discussed.