Localization of the ankyrin-binding site on erythrocyte membrane protein, band 3

The predominant attachment site of the spectrin-based cytoskeleton to the erythrocyte membrane occurs via the interaction of ankyrin with the cytoplasmic domain of band 3 (cdb3). In order to further characterize this interaction, we have conducted experiments to localize the ankyrin-binding site on cdb3. Four monoclonal and three antipeptide polyclonal antibodies were raised against cdb3 and used in competition studies to identify regions of close association of cdb3 with ankyrin. Antibodies to regions of cdb3 near the cytoplasmic domain-membrane spanning domain junction had no effect on 125I-ankyrin binding. Likewise, an antibody to a highly conserved region between residues 142 and 154 did not inhibit ankyrin binding. However, antibodies at or near the cysteine 201-317 cluster and the proposed proline-rich hinge in the center of cdb3 were potent inhibitors of ankyrin association, as were antibodies to the acidic NH2 terminus. Additional evidence for interaction of ankyrin with the NH2-terminal region of cdb3 was obtained by demonstrating the ability of ankyrin to inhibit tyrosine phosphorylation of cdb3 at its NH2 terminus by a purified calf thymus tyrosine kinase. These studies reveal two regions of cdb3, distant in primary sequence, which interact with ankyrin. A specific conformation of cdb3 may be required to permit these regions to simultaneously associate with ankyrin and allow binding to occur.     

Species-specific epitopes exist on the cytoplasmic amino-terminal domain of erythrocyte band 3 protein.

Monoclonal antibodies (P3-9H, P3-1F, P3-2H, P3-4A, and P3-4C) to human erythrocyte band 3 were produced using human erythrocyte membranes as the immunogen. All epitopes defined by these antibodies were found on the amino-terminal cytoplasmic domain of erythrocyte band 3. The antibodies crossreacted variously with erythrocyte band 3 of primates (chimpanzee, orangutan, Rhesus monkey, Japanese monkey, spider monkey, and capuchin monkey) in enzyme-linked immunosorbent assay. P3-9H did not crossreact with erythrocyte band 3 of any primate examined; P3-1F crossreacted only with that of chimpanzee; P3-2H crossreacted with erythrocyte band 3 of chimpanzee, spider monkey, and capuchin monkey; and P3-4A and P3-4C crossreacted with erythrocyte band 3 of all primates examined. These results suggest that evolutional changes in primates are accumulated in the amino-terminal cytoplasmic domain of band 3 and that species-specific epitopes exist on this domain.     

Antigenic determinants of the cytoplasmic domain of band 3 from bovine erythrocyte membrane

Rabbit antibodies were prepared against the cytoplasmic 38K-Da fragment of bovine band 3 and the immunological cross-reactivity with human, murine, rat, and chicken band 3 was examined. The antibodies cross-reacted with human and rodent band 3, indicating that there is an antigenic determinant(s) common to primate and nonprimate species. However, the antibodies did not recognize chicken band 3. Antigenic sites on the 38K-Da fragment were determined via amino acid sequence and immunoblotting analyses of proteolytic peptides of the fragment. Positions of antigenic determinants which were assumed to be common to primate and nonprimate species were mapped to the areas of residues 127-160 and of residues 259-304 in the primary structure of human band 3. Another epitope(s), which is absent in human band 3, existed in a region having a bovine-specific amino acid sequence. In addition, comparison of sequence data from different species showed that a proposed hinge region and a tryptophan-rich region on the cytoplasmic domain of band 3 [P. S. Low et al. (1984) J. Biol. Chem. 259, 13,070-13,076; R. R. Kopito and H. F. Lodish (1985) Nature (London) 316, 234-238] are also conserved in the bovine case.     

Characterization of the reversible conformational equilibrium of the cytoplasmic domain of erythrocyte membrane band 3

The cytoplasmic domain of the erythrocyte membrane protein, band 3, contains binding sites for hemoglobin, several glycolytic enzymes, and ankyrin, the linkage to the cytoskeleton. In an earlier study, we found evidence which suggested that band 3 might undergo a native conformational change. We demonstrate here that the cytoplasmic domain of band 3 does exist in a reversible, pH-dependent conformational equilibrium among 3 native states. At physiological salt concentrations this equilibrium is characterized by apparent pKa values of 7.2 and 9.2; however, these apparent pKa values change if the domain's sulfhydryl groups are modified. A major component of the structural change appears to involve the pivoting of two subdomains of the cytoplasmic domain at a central hinge, as evidenced by both hydrodynamic and fluorescence energy transfer measurements. The probable site of this hinge is between residues 176 and 191, a region highly accessible to proteases and also rich in proline. These structural rearrangements also apparently extend to the cluster of tryptophan residues near the N terminus, since the domain's intrinsic fluorescence more than doubles between pH 6.5 and 9.5. No measurable change in band 3 secondary or quaternary structure could be detected during the conformational transitions. A structural model of the cytoplasmic domain of band 3 is presented to show the possible spatial relationships between the regions of conformational change and the sites of peripheral protein binding.     

Characterization of the reversible conformational equilibrium in the cytoplasmic domain of human erythrocyte membrane band 3

The cytoplasmic domain of erythrocyte membrane band 3 (cdb3) serves as a center of membrane organization, interacting with such proteins as ankyrin, protein 4.1, protein 4.2, hemoglobin, several glycolytic enzymes, and a tyrosine kinase, p72syk. cdb3 exists in a reversible, pH-dependent conformational equilibrium characterized by large changes in Stokes radius (11 A) and intrinsic fluorescence (2-fold). Based on the crystallographic structure of the cdb3 dimer, we hypothesized that the above conformational equilibrium might involve the movement of flanking peripheral protein binding domains away from a shared dimerization domain. To test this hypothesis, we have mutated both donor (W105L) and acceptor (D316A) residues of a prominent H bond that bridges the above two domains and have examined the effect on the resulting conformational equilibrium. Analysis of the intrinsic fluorescence, Stokes radius, thermal stability, urea stability, and segmental mobility of these mutants reveals that the above H bond is indeed present in the low pH conformation of cdb3 and broken in a higher pH conformation. The data further reveal that cdb3 exists in three native pH-dependent conformations and that rupture of the aforementioned H bond occurs only during conversion of the low pH conformation to the mid-pH conformation. Conversion of the mid-pH conformation to the high pH conformation would now appear to involve structural changes primarily in the peripheral protein binding domain. Because ankyrin associates avidly with the low pH conformation of cdb3, ankyrin occupancy should strongly influence this structural equilibrium and thereby affect band 3 and perhaps global membrane properties.     

The interaction of hemoglobin with the cytoplasmic domain of band 3 of the human erythrocyte membrane

Previous studies point to the acidic amino-terminal segment of band 3, the anion transport protein of the red cell, as the common binding site for hemoglobin and several of the glycolytic enzymes to the erythrocyte membrane. We now report on the interaction of hemoglobin with the synthetic peptide AcM-E-E-L-Q-D-D-Y-E-D-E, corresponding to the first 11 residues of band 3, and with the entire 43,000-Da cytoplasmic domain of the protein. In the presence of increasing concentrations of the peptide, the oxygen binding curve for hemoglobin is shifted progressively to the right, indicating that the peptide binds preferentially to deoxyhemoglobin. The dissociation constant for the deoxyhemoglobin-peptide complex at pH 7.2 in the presence of 100 mM NaCl is 0.31 mM. X-ray crystallographic studies were carried out to determine the exact mode of binding of the peptide to deoxyhemoglobin. The difference electron density map of the deoxyhemoglobin-peptide complex at 5 A resolution showed that the binding site extends deep (approximately 18 A) into the central cavity between the beta chains, along the dyad symmetry axis, and includes Arg 104 beta 1 and Arg 104 beta 2 as well as most of the basic residues within the 2,3-diphosphoglycerate binding site. The peptide appears to have an extended conformation with only 5 to 7 of the 11 residues in contact with hemoglobin. In agreement with the crystallographic studies, binding of the peptide to deoxyhemoglobin was blocked by cross-linking the beta chains at the entrance to the central cavity. Oxygen equilibrium studies showed that the isolated cytoplasmic fragment of band 3 also binds preferentially to deoxyhemoglobin. The binding of the 43,000-Da fragment to hemoglobin was inhibited in the cross-linked derivative indicating that the acidic amino-terminal residues in the intact cytoplasmic domain also bind within the central cavity of the hemoglobin tetramer.     

Solution structure of the cytoplasmic domain of erythrocyte membrane band 3 determined by site-directed spin labeling

The cytoplasmic domain of the anion exchange protein (cdb3) serves as a critical organizing center for protein-protein interactions that stabilize the erythrocyte membrane. The structure of the central core of cdb3, determined by X-ray crystallography from crystals grown at pH 4.8, revealed a compact dimer for residues 55-356 and unresolved N- and C-termini on each monomer [Zhang et al. (2000) Blood 96, 2925-2933]. Given that previous studies had suggested a highly asymmetric structure for cdb3 and that pH dependent structural transitions of cdb3 have been reported, the structure of cdb3 in solution at neutral pH was investigated via site-directed spin labeling in combination with conventional electron paramagnetic resonance (EPR) and double electron electron resonance (DEER) spectroscopies. These studies show that the structure of the central compact dimer (residues 55-356) is indistinguishable from the crystal structure determined at pH 4.8. N-Terminal residues 1-54 and C-terminal residues 357-379 are dynamically disordered and show no indications of stable secondary structure. These results establish a structural model for cdb3 in solution at neutral pH which represents an important next step in characterizing structural details of the protein-protein interactions that stabilize the erythrocyte membrane.     

Purification and properties of human erythrocyte band 4.2. Association with the cytoplasmic domain of band 3

We have purified the human erythrocyte membrane protein band 4.2 to greater than 85% homogeneity. The protein was extracted from spectrin-actin-depleted inside-out vesicles in a pH 11 medium and purified by gel filtration in the presence of 1 M KI. The purified protein was heterogeneous and had an average S20,w of 5.5 and an average Stokes radius of 82 A. By electron microscopy, the protein appeared heterogeneous in size and shape, having a diameter ranging from 80 to 150 A. The protein bound saturably to band 4.2-depleted red cell inside-out vesicles, and the binding exhibited a concave Scatchard plot. Binding was reduced greater than 90% by proteolytic digestion of membranes. Digestion studies suggested that there are two classes of binding sites for band 4.2 on the cytoplasmic aspect of red cell membranes, one of which is likely to be band 3. The purified 43-kDa cytoplasmic domain of band 3 competed for band 4.2 binding to red cell membranes and could completely abolish binding when added at a concentration of greater than 200 micrograms/ml. The purification of band 4.2 and the characterization of its association with red cell membranes should facilitate the discovery of the function of this major red cell membrane protein.     

Interaction of deoxyhemoglobin with the cytoplasmic domain of murine erythrocyte band 3

The partial pressure of oxygen constitutes an important factor in the regulation of human erythrocyte physiology, including control of cell volume, membrane structure, and glucose metabolism. Because band 3 is thought to be involved in all three processes and because binding of hemoglobin (Hb) to the cytoplasmic domain of band 3 (cdb3) is strongly oxygen-dependent, the possibility that the reversible association of deoxyhemoglobin (deoxyHb) with cdb3 might constitute an O(2)-dependent sensor that mediates O(2)-regulated changes in erythrocyte properties arises. While several lines of evidence support this hypothesis, a major opposing argument lies in the fact that the deoxyHb binding sequence on human cdb3 is not conserved. Moreover, no effect of O(2) pressure on Hb-band 3 interactions has ever been demonstrated in another species. To explore whether band 3-Hb interactions might be widely involved in O(2)-dependent regulation of erythrocyte physiology, we undertook characterization of the effect of O(2) on band 3-Hb interactions in the mouse. We report here that murine band 3 binds deoxyHb with significantly greater affinity than oxyHb, despite the lack of significant homology within the deoxyHb binding sequence. We further map the deoxyHb binding site on murine band 3 and show that deletion of the site eliminates deoxyHb binding. Finally, we identify mutations in murine cdb3 that either enhance or eliminate its affinity for murine deoxyHb. These data demonstrate that despite a lack of homology in the sequences of both murine band 3 and murine Hb, a strong oxygen-dependent association of the two proteins has been conserved.     

Localization of the pyridoxal phosphate binding site at the COOH-terminal region of erythrocyte band 3 protein

A human erythrocyte Band 3 peptide, affinity labeled with pyridoxal phosphate, was purified by a combination of gel permeation and reverse-phase high performance liquid chromatography. The amino acid sequence of the transmembrane peptide was determined by sequencing subfragments of the peptide obtained from lysyl endopeptidase and staphylococcal proteinase V8 digestions. When a peptide containing the COOH-terminal of human erythrocyte Band 3 was also purified and sequenced, the affinity-labeled peptide was found to be located close to the COOH-terminal of Band 3, where it could be aligned with amino acid residues 852-927 of a murine erythrocyte Band 3, deduced from a nucleotide sequence of a cDNA clone (Kopito, R. R., and Lodish, H. F. (1985) Nature 316, 234-238). The amino acid sequence of the COOH-terminal region was highly homologous to that of murine Band 3. As a result, the sequence of the COOH-terminal peptide of Band 3 was established as follows. (Formula: see text). The pyridoxal phosphate binding site was identified as Lys-18 which corresponded to Lys-869 of the deduced sequence. It appears that the COOH-terminal region of Band 3 constitutes at least a part of the active center for anion transport in human erythrocyte membranes.     

Two-dimensional structure of the membrane domain of human band 3, the anion transport protein of the erythrocyte membrane.

The membrane domain of human erythrocyte Band 3 protein (M(r) 52,000) was reconstituted with lipids into two-dimensional crystals in the form of sheets or tubes. Crystalline sheets were monolayers with six-fold symmetry (layer group p6, a = b = 170 A, gamma = 60 degrees), whereas the symmetry of the tubular crystals was p2 (a = 104 A, b = 63 A, gamma = 104 degrees). Electron image analysis of negatively stained specimens yielded projection maps of the protein at 20 A resolution. Maps derived from both crystal forms show that the membrane domain is a dimer of two monomers related by two-fold symmetry, with each monomer consisting of three subdomains. In the dimer, two subdomains of each monomer form a roughly rectangular core (40 x 50 A in projection), surrounding a central depression. The third subdomain of the monomer measures approximately 15 x 25 A in projection and appears to be connected to the other two by a flexible link. We propose that the central depression may represent the channel for anion transport while the third subdomain appears not to be directly involved in channel formation.     

Interactions between protein 4.1 and band 3. An alternative binding site for an element of the membrane skeleton

Protein 4.1 from human erythrocytes formed a complex with band 3 in inside-out erythrocyte membrane vesicles and with soluble peptides derived from the cytoplasmic domain of band 3. Protein 4.1 labeled metabolically with 32P bound saturably to vesicles depleted of endogenous protein 4.1. The soluble cytoplasmic domain of band 3 (43K) competitively displaced approximately 60% of bound 32P-protein 4.1 from reconstituted membrane vesicles. Pretreatment of vesicles with anti-43K similarly inhibited the rebinding of protein 4.1. In solution, 125I-43K formed a complex with protein 4.1 that saturated at 1:1 stoichiometry and migrated as a discrete band when analyzed by nondenaturing polyacrylamide gel electrophoresis. In rate-zonal sedimentation in isotonic salt solutions, protein 4.1 and 43K sedimented as a sharp peak at 4.4 S. In experiments aimed at exploring the role of the protein 4.1-band 3 interaction in the organization of the membrane skeleton, the effect of spectrin was investigated. Spectrin and protein 4.1 formed a complex which co-sedimented in sucrose gradients, but the addition of 43K to preformed spectrin-protein 4.1 complexes resulted in disruption of the complex and co-sedimentation of most of the protein 4.1 with 43K. These results suggest that protein 4.1 can associate with band 3 in the erythrocyte membrane and that this association may modulate the attachment of the membrane skeleton to the membrane.     

Associations of human erythrocyte band 4.2. Binding to ankyrin and to the cytoplasmic domain of band 3

We have examined the associations of purified red cell band 4.2 with red cell membrane and membrane skeletal proteins using in vitro binding assays. Band 4.2 bound to the purified cytoplasmic domain of band 3 with a Kd between 2 and 8 X 10(-7) M. Binding was saturable and slow, requiring 2-4 h to reach equilibrium. This finding confirms previous work suggesting that the principal membrane-binding site for band 4.2 lies within the 43-kDa cytoplasmic domain of band 3 (Korsgren, C., and Cohen, C. M. (1986) J. Biol. Chem. 261, 5536-5543). Band 4.2 also bound to purified ankyrin in solution with a Kd between 1 and 3.5 X 10(-7) M. As with the cytoplasmic domain of band 3, binding was saturable and required 4-5 h to reach equilibrium. Reconstitution with ankyrin of inside-out vesicles stripped of all peripheral proteins had no effect upon band 4.2 binding to membranes; similarly, reconstitution with band 4.2 had no effect upon ankyrin binding. This shows that ankyrin and band 4.2 bind to distinct loci within the 43-kDa band 3 cytoplasmic domain. Coincubation of ankyrin and band 4.2 in solution partially blocked the binding of both proteins to the membrane. Similarly, coincubation of bands 4.1 and 4.2 in solution partially blocked binding of both to membranes. In all cases, the data suggest the possibility that domains on each of these proteins responsible for low affinity membrane binding are principally affected. The data also provide evidence for an association of band 4.2 with band 4.1. Our results show that band 4.2 can form multiple associations with red cell membrane proteins and may therefore play an as yet unrecognized structural role on the membrane.     

Inhibition of erythrocyte membrane shape change by band 3 cytoplasmic fragment

The ATP-dependent transformation of crenated white human erythrocyte ghosts into smoothed disc and cup forms is inhibited by the soluble 40-45-kilodalton (kDa) cytoplasmic portion of the major transmembrane protein, band 3. The band 3 fragment was prepared by chymotryptic treatment of inverted vesicles stripped of peripheral proteins. When present at greater than or equal to 0.2 mg per mg membrane protein (ie, greater than or equal to 2 mol fragment per mol endogenous band 3), the fragment significantly reduced the rate of shape change but did not alter the proportion of membranes that were ultimately converted into smoothed forms (greater than 90%). The inhibitory activity of the fragment could not be attributed to contamination of the fragment preparation by actin or proteolytic enzymes. ATP-independent shape transformation was not inhibited. The band 3 fragment may compete with endogenous, intact band 3 for an association with the spectrin-actin network required for ATP-dependent smoothing of crenated membranes.     

Primary structure of the cytoplasmic domain of human erythrocyte protein band 3. Comparison with its sequence in the mouse

We report here the peptide profile of the human cytoplasmic domain of band 3 protein (CDB-3). The peptide alignment was designed allowing for maximal homology with the murine protein whose sequence was deduced from cDNA analysis by Kopito and Lodish (Kopito, R.R., Anderson, M. and Lodish, H.F. (1987) J. Biol. Chem. 262, 8035-8040). In the human protein, part of the amino acid sequence has been determined by Kaul et al. (Kaul, R.K., Murthy, P.S.N., Reddy, A.G., Steck. T.L. and Kohler, H. (1983) J. Biol. Chem. 258, 7981-7990). We have sequenced most of the fragment not described by these author. The homology with the murine protein is high (90%), except in a few peptides where it is only 50%. The actual miniaturization of the techniques allows for the determination of a clear peptide profile of human CDB-3 starting from 10 ml blood samples. Our characterization of the peptide profile of membrane proteins is the first step towards the identification of genetic mutations, which have to be looked for in hemolytic anemia when the presence of an abnormal membrane protein is suspected.     

Both ankyrin and band 4.1 are required to restrict the rotational mobility of band 3 in the human erythrocyte membrane.

A population of band 3 proteins in the human erythrocyte membrane is known to have restricted rotational mobility due to interaction with cytoskeletal proteins. We have further investigated the cause of this restriction by measuring the effects on band 3 rotational mobility of rebinding ankyrin and band 4.1 to ghosts stripped of these proteins as well as spectrin and actin. Rebinding either ankyrin or 4.1 alone has no detectable effect on band 3 mobility. Rebinding both these proteins together does, however, reimpose a restriction on band 3 rotation. The effect on band 3 rotational mobility of rebinding ankyrin and 4.1 are similar irrespective of whether or not band 4.2 is removed from the membrane. We suggest that ankyrin and 4.1 together promote the formation of slowly rotating clusters of band 3.     

Structure of the spectrin-actin binding site of erythrocyte protein 4.1

The complete primary structure of the functional site of erythrocyte protein 4.1 involved in spectrin-actin associations has been determined. The sequence of this domain, which contains 67 amino acids and has a molecular mass of 8045 daltons, has been obtained by NH2-terminal sequence analysis of an 8-kDa chymotryptic peptide, three endoproteinase lysine C-cleaved peptides and two peptides obtained by Staphylococcus aureus protease V8 cleavage. All peptides including the 8-kDa domain peptide were purified by reverse-phase high performance liquid chromatography. Antibodies against two different synthetic peptides of the 8-kDa domain are able to inhibit the association between protein 4.1, spectrin, and F-actin, corroborating that the 8-kDa domain is responsible for the formation of a ternary complex. A computer search of the 8-kDa sequence with the National Biomedical Research Foundation database did not detect any significant homologies to known sequences. Protein 4.1 is not related to any known proteins and may represent a new protein superfamily.     

Hydrodynamic examination of the dimeric cytoplasmic domain of the human erythrocyte anion transporter, band 3.

Solution studies of the cytoplasmic domain (molecular mass approximately 40kDa) of band 3, the anion exchanger from human erythrocyte membranes, previously suggested a dimeric molecule on the basis of the relative techniques of calibrated gel filtration and calibrated preparative ultracentrifugation. This dimeric behavior is firmly established on an absolute basis by a combination of calibrated gel chromatography and absolute ultracentrifugation techniques. Sedimentation velocity in the analytical ultracentrifuge combined with calibrated gel chromatography give a molecular mass M of (77 +/- 4) kDa, a value confirmed by low-speed sedimentation equilibrium. Velocity sedimentation in the analytical ultracentrifuge gave a single sedimenting species with an s o 20,w of (3.74 +/- 0.07)S. Sedimentation equilibrium analysis was also used to establish the strength of the binding via the dissociation constant Kd, with a value from direct fitting of the concentration distribution curves of (2.8 +/- 0.5) microM, confirmed by a value of approximately 3 microM obtained from fitting a plot of molecular weight Mw,app versus cell loading concentration. Hydrodynamic calculations based on the classical translational frictional ratio showed that the protein was highly asymmetric, with an axial ratio of approximately 10:1, consistent with observations from electron microscopy.     

Crystallization and preliminary X-ray analysis of the cytoplasmic domain of human erythrocyte band 3

A cytoplasmic domain of the human erythrocyte membrane protein band 3 (M_r = 42,500), residues 1–379, expressed in and purified from E. coli, has been crystallized by the method of vapor diffusion in sitting drops with subsequent streak-seeding at room temperature. Initial crystals were grown from solutions containing 65–68% saturated ammonium sulfate at pH 4.9 and 2 mg/ml protein. Subsequent streak-seeding into solutions of 50–53% ammonium sulfate at pH 4.9 and 7 mg/ml protein produced single crystals suitable fur X-ray analysis, which contained pure protein as revealed by gel electrophoresis. The crystals belong to the monoclinic space group C2 with cell dimensions of a = 178.8 Å, b = 90.5 Å, c = 122.1 Å, and β = 131.3° and diffract at least to 2.7 Å resolution (at 100 K). A self-rotation function shows the presence of approximate 222 local symmetry. © 1995 Wiley-Liss, Inc.