Interaction of lectins with membrane receptors on erythrocyte surfaces.

The interactions of human genotype AO erythrocytes (red blood cells) (RBCs) with N-acetylgalactosamine-reactive lectins isolated from Helix pomatia (HPA) and from Dolichos biflorus (DBA) were studied. Binding curves obtained with the use of tritium-labeled lectins showed that the maximal numbers of lectin molecules capable of binding to human genotype AO RBCs were 3.8 X 10(5) and 2.7 X 10(5) molecules/RBC for HPA and DBA, respectively. The binding of one type of lectin may influence the binding of another type. HPA was found to inhibit the binding of DBA, but not vice versa. The binding of HPA was weakly inhibited by a beta-D-galactose-reactive lectin isolated from Ricinus communis (designated RCA1). Limulus polyphemus lectin (LPA), with specificity for N-acetylneuraminic acid, did not influence the binding of HPA but enhanced the binding of DBA. About 80% of LPA receptors (N-acetylneuraminic acid) were removed from RBC surfaces by neuraminidase treatment. Neuraminidase treatment of RBCs resulted in increases of binding of both HPA and DBA, but through different mechanisms. An equal number (7.6 X 10(5) of new HPA sites were generated on genotypes AO and OO RBCs by neuraminidase treatment, and these new sites accounted for the enhancement (AO cells) and appearance (OO cells) of hemagglutinability by HPA. Neuraminidase treatment did not generate new DBA sites, but increased the DBA affinity for the existing receptors; as a result, genotype AO cells increased their hemagglutinability by DBA, while OO cells remained unagglutinable. The use of RBCs of different genotypes in binding assays with 3H-labeled lectins of known specificities provides an experimental system for studying cell-cell recognition and association.

Interaction energies in lectin-induced erythrocyte aggregation.

Two N-acetylgalactosamine-reactive lectins, Helix pomatia (HPA) and Dolichos biflorus (DBA), were used to study the energies involved in cell-cell interactions through the specific binding of these lectins to their membrane receptors on genotype AO human erythrocytes (red blood cells) (RBCs). The energy required to dissociate a unit of aggregated membrane area (gamma d) of two RBCs bridged by lectin molecules was determined from the shear force needed to dissociate two-cell aggregates in a flow channel. When HPA were used as bridging molecules, gamma d (0.4 X 10(-4) to 3.8 X 10(-4) dyn/cm) was proportional to the density (D = 175 to 1,060 molecules/micron 2) of HPA molecules bound on the RBC membrane. A similar gamma d/D ratio was also obtained for DBA. These results indicate that the number of lectin molecules bound on the interface plays an important role in determining the energy required for cell-cell dissociation. The aggregation energy per unit membrane area (gamma a) in lectin-induced aggregates was calculated from the degree of encapsulation of a lectin-bound, heat-sphered human RBC by a normal discoid RBC. A minimum of approximately 1,800 HPA molecules/micron 2 on the spheres was required to form stable aggregates with the RBC. By using spheres having a surface HPA density of 1,830 to 2,540 molecules/micron 2, or 1.1-1.5 X 10(12) combining sites/cm2, the gamma a value for HPA-induced aggregation was found to be 2.2 X 10(-3) dyn/cm. This higher value of gamma a than gamma d has been explained on the basis of several differences in aggregation and disaggregation processes. The gamma a value for DBA-induced aggregation was not obtainable by the sphere encapsulation method because of the relative low D values. A comparison of the present results with the published value of the free energy change of 5 kcal/mol for the interactions of HPA and DBA with their ligands suggests that only a small fraction of the lectin molecules bound to RBC surface participate in the bridging of adjacent cells.

Micromanipulation of adhesion of a Jurkat cell to a planar bilayer membrane containing lymphocyte function-associated antigen 3 molecules.

Cell adhesion plays a fundamental role in the organization of cells in differentiated organs, cell motility, and immune response. A novel micromanipulation method is employed to quantify the direct contribution of surface adhesion receptors to the physical strength of cell adhesion. In this technique, a cell is brought into contact with a glass-supported planar membrane reconstituted with a known concentration of a given type of adhesion molecules. After a period of incubation (5-10 min), the cell is detached from the planar bilayer by pulling away the pipette holding the cell in the direction perpendicular to the glass-supported planar bilayer. In particular, we investigated the adhesion between a Jurkat cell expressing CD2 and a glass-supported planar bilayer containing either the glycosyl-phosphatidylinositol (GPI) or the transmembrane (TM) isoform of the counter-receptor lymphocyte function-associated antigen 3 (LFA-3) at a concentration of 1,000 molecules/microns 2. In response to the pipette force the Jurkat cells that adhered to the planar bilayer containing the GPI isoform of LFA-3 underwent extensive elongation. When the contact radius was reduced by approximately 50%, the cell then detached quickly from its substrate. The aspiration pressure required to detach a Jurkat cell from its substrate was comparable to that required to detach a cytotoxic T cell from its target cell. Jurkat cells that had been separated from the substrate again adhered strongly to the planar bilayer when brought to proximity by micromanipulation. In experiments using the planar bilayer containing the TM isoform of LFA-3, Jurkat cells detached with little resistance to micromanipulation and without changing their round shape.

Determination of junction avidity of cytolytic T cell and target cell.

A direct measurement of the avidity of the junction between a cytotoxic T lymphocyte and its target cell was achieved by using a biophysical approach. A micromanipulation technique was used to determine the force required to separate a cytotoxic T cell (human clone F1, with specificity for HLA-DRw6) from its specific target cell (JY: HLA-A2, -B7, -DR4, w6) prior to delivery of the lethal hit. The force required to separate the F1-JY pair is 1.5 X 10(4) dynes per square centimeter. This junction avidity for F1-JY pairs is 6 to 13 times greater than that for F1-F1 and JY-JY pairs; the F1-JY conjugate requires a stronger separating force and is more easily rejoined than the homologous cell pairs. This study provides an estimate of the avidity of cytotoxic T cells for their target cells and insights into the biophysical correlates of the molecular complexes formed in the interaction of cytotoxic T cells and their targets during the cytotoxic process.

Membrane protein phosphorylation during stomatocyte-echinocyte transformation of human erythrocytes.

The normal, discoid shape of red blood cells represents an equilibrium between two opposing factors, i.e., stomatocytic and echinocytic transformations. Most stomatocytic agents were found to be inhibitors of calmodulin, a regulator of the phosphorylation of membrane proteins. We determined whether red cell shape transformations could be caused by changes in phosphorylation of membrane proteins, specifically the cAMP-dependent phosphorylation of ankyrin and band 4.1. Red blood cells were incubated with 32P and 100 microM chlorpromazine (stomatocytic transformation) or 30 mM sodium salicylate (echinocytic transformation) for various time intervals. Ghost membrane proteins were examined by polyacrylamide gel electrophoresis and autoradiography. Spectrin (beta-chain), ankyrin, band 3, band 4.1 and 4.9 were phosphorylated. No change was found in the degree and pattern of phosphorylation after stomatocytic transformation. Salicylate caused a reversible inhibition of transmembranous phosphate transport in both directions. The results indicate that the stomatocytic transformation induced by chlorpromazine and the echinocytic transformation induced by salicylate do not involve a change in phosphorylation, but that the echinocytic transformation induced by salicylate is associated with an inhibition of transmembranous transport of phosphate. Studies with salicylate suggest that the phosphorylation sites of band 3 are found mainly on the endofacial side of the membrane.

Genomic organization of mouse and human erythrocyte tropomodulin genes encoding the pointed end capping protein for the actin filaments.

Erythrocyte tropomodulin (E-Tmod), a globular protein of 359 residues, is highly expressed in the erythrocyte, heart and skeletal muscle. By binding to the N-terminus of tropomyosin (TM) and actin, E-Tmod blocks the elongation and depolymerization of the actin filaments at the pointed end. In erythrocytes, the E-Tmod/TM complex contributes to the formation of the short actin protofilament, which in turn defines the geometry of the membrane skeleton. In juvenile mice, over-expression of E-Tmod is associated with dilated cardiomyopathy. We have previously cloned the human E-Tmod cDNA, identified its TM-binding region, and mapped its gene to chromosome 9q22. Through genomic library screening and PCR-based genomic walking we have now cloned the mouse E-Tmod gene, whose coding region spans approximately 60kb containing nine exons and eight introns. The human E-Tmod gene obtained by PCR has an identical exon-intron organization. In sanpodo, a Tmod homologue in Drosophila, the exon boundaries are also conserved except that exons 2-5 and 6-7 are 'fused' and alternative splicing of two additional 5' exons and the 3' exons may give rise to several sanpodo isoforms. In a Tmod-like gene of C. elegans, exons 2-3 are 'fused', boundaries of exons 1, 7, 8, and 9 are conserved and exon/intron junctions of exons 4, 5 and 6 are shifted by a few residues. Analyses of 15 Tmod members from six species show no insertions or deletions of residues in the region of exons 6 and 7. A 5' rapid amplification of cDNA ends reveals that mouse E-Tmod transcripts obtained from embryonic stem cells, skeletal muscle and heart, but not smooth muscle, contain an additional 86bp untranslated cDNA sequence further upstream from exon 1. Thus, alternative promoters may provide a possible mechanism for tissue-specific expression and regulation of E-Tmod. This study is the first to report the exon organization of E-Tmod genes, which allows their regulation, manipulation, and disease relevance to be further investigated.

Erythrocytes in Duchenne dystrophy: osmotic fragility and membrane deformability.

Duchenne erythrocytes showed increased osmotic fragility as compared to controls (p less than 0.01), but individual values overlapped with controls, and only half of the Duchenne erythrocyte values were abnormal. When the effect of the smaller mean corpuscular volume of Duchenne erythrocytes was taken into account, there was no significant difference from controls in membrane deformability, as determined by microsieving or flow channel measurements. The increased osmotic fragility suggests minor changes in erythrocyte membrane properties in Duchenne muscular dystrophy.

Energy balance in red cell interactions.

Experiments were performed to elucidate the balance of energies involved in the formation of red blood cell (RBC) aggregates and in their disaggregation. In order to achieve a mean stable rouleau formation, the aggregating energy provided by macromolecular binding to the cell membrane must overcome the disaggregation energy of electrostatic repulsion between RBC surfaces and the effects of mechanical shear stress. In a quiescent suspension the net aggregation energy is largely stored in the membrane as a change in strain energy. The alterations in strain energy cause the curvature of the end cells in rouleaux of normal RBCs in Dx 80 to change from concave to convex and back again to concave as [Dx 80] was increased from 1 to 4 to 6 g/dl; computation of net aggregation energy per unit area (gamma) from changes in membrane strain energy yielded values on the order of 10(3) ergs/cm2. The end cells of neuraminidase-treated RBCs remained convex with [Dx 80] above 2 g/dl, and gamma is probably on the order of 10(2) ergs/cm2. The variations in gamma with [Dx 80] and RBC surface charge are similar to variations in reflectometric aggregation index without shear ( RAI0 ), indicating that RAI0 reflects gamma. The difference in gamma between normal and neuraminidase-treated RBCs represents the electrostatic repulsive energy, the magnitude of which varied inversely with dextran molecular size and directly with [Dx]. Moderate shearing in the reflectometer enhanced RBC aggregation by promoting cell-cell encounter, but high shear stresses cause RBC disaggregation. The energy required to disaggregate a unit interacting area of normal RBCs in Dx 80 in a flow channel is on the order of 10(4) ergs/cm2, which is much lower than gamma. These results suggest that the release of the stored membrane strain energy during disaggregation aids in the separation process. The results show that the understanding of RBC aggregation requires the considerations of surface charge, properties of aggregating agents, and the rheology of the cell membrane.

Angiogenic growth factor mRNA responses in muscle to a single bout of exercise.

A major adaptation to exercise is new capillary formation in skeletal muscle. On the basis of angiogenesis in tumors and during development, several angiogenic growth factors may be involved, including vascular endothelial growth factor (VEGF), basic fibroblast growth factor (bFGF), and transforming growth factor-beta 1 (TGF-beta 1). In 9-wk-old female Wistar rats, mRNA expression for these three growth factors in gastrocnemius muscle was examined by quantitative Northern analysis after a single 1-h run at 15 or 20 m/min at 10 degrees incline in room air. A third group ran at 15 m/min in 12% O2, and resting control groups were included at inspired O2 fractions of 0.21 and 0.12. Exercise significantly increased mRNA levels two- to fourfold, which was evident over the first 4 h postexercise; by 8 and 24 h, mRNA levels returned to baseline. For all three factors, mRNA levels were significantly higher after exercise at 20 than at 15 m/min. Hypoxia at rest doubled VEGF and TGF-beta 1 message but had no effect on bFGF. Hypoxic exercise further raised VEGF mRNA levels but had no effect on the other factors. We suggest that VEGF, bFGF, and TGF-beta 1 may be involved in the angiogenic response to exercise and that reduced intracellular PO2 (as occurs during normoxic exercise) may be part of the stimulus to such growth factor production.

Immunodetection of membrane skeletal protein 4.2 in bovine and chicken eye lenses and erythrocytes.

PURPOSE: Protein 4.2 is a major erythrocyte membrane skeletal protein, playing an important role in maintaining the integrity and stability of the membrane. It is a transglutaminase-like molecule with no enzymatic cross-linking activity. Several protein 4.2-associated proteins (i.e. band 3, ankyrin, and protein 4.1) and transglutaminase activities have been detected in the lens. The purpose of this study is to find out if protein 4.2 is also expressed in lens fiber membranes. METHODS: Western blot analysis of cell membranes isolated from bovine and chicken lens fibers and erythrocytes, and immunocytochemistry of frozen sections of bovine and chicken lens fibers were carried out using two protein 4.2-specific antibodies. These two peptide antibodies have been used to identify two alternatively spliced protein 4.2 isoforms in human erythrocyte membranes: the short (P4.2S, or hP4.2(691)) and the long (P4.2L, or hP4.2(721)) isoforms. RESULTS: Western blot analysis using anti-P4.2(L) antibody demonstrated specific immunoreactive polypeptides in bovine and chicken lens fiber membranes and erythrocyte membranes, co-migrating with hP4.2(721). Immunofluorescence staining of bovine and chicken lenses, using anti-P4.2(L) antibody, revealed specific signals along the cell membranes of cortical fibers. The signals exhibited a unique, patchy pattern along the cortical fiber cell membranes in both cross-sectional and longitudinal views. In cross sections, the labeling of anti-P4.2(L) along the entire cell membranes gave an appearance of a hexagonal shape of fiber cells. CONCLUSIONS: Protein 4.2, or its analogs, is present in the lens fiber membranes. Its specific staining pattern in the lens fibers suggests that it participates in the architecture of the lens fiber cell membranes, and may play a role in the lens mechanics and pathology.

Agglutination-induced erythrocyte deformation by two blood group A-specific lectins: studies by light and electron microscopy.

The adhesive energy in lectin-induced agglutination can be assessed by the deformation of erythrocytes in aggregates. Helix pomatia (HPA) and Dolichos biflorus (DBA) specifically agglutinated blood group A erythrocytes and induced a change in curvatures of cells at the end of the aggregates. The curvatures changed from concavity to convexity with increasing lectin concentrations. HPA-induced aggregates achieved the theoretical maximal end cell curvature of 0.27 microns-1 at 2-3 micrograms/ml; DBA-induced aggregates approached 0.23 microns-1, requiring 400 micrograms/ml. At any given lectin concentration, HPA showed greater surface binding, caused higher cell curvatures, induced larger aggregate size, and had greater adhesive energies, as compared to DBA. HPA and DBA are globular proteins with a diameter of approximately 5.5 nm and approximately 6.1 nm, respectively, as revealed by the negative staining electron microscopy. The former induced uniform intercellular spacing (approximately 15 nm), whereas the latter induced both smooth (approximately 20 nm) and ruffled spacings. The uniform intercellular spacing was not a function of lectin concentrations. Microscopic studies of erythrocyte deformation provided morphological correlates of biophysical findings of differential adhesive energies induced by these two blood group A-specific lectins.

Red cell membrane elasticity as determined by flow channel technique.

The elasticity of red cell membrane was determined in a rectangular flow channel under controlled shear flow. The relation between shear stress and cell extension ratio (lambda) has been analyzed with the use of Evans' two-dimensional model. The deformed cell shapes observed experimentally agreed well with the model with lambda up to 1.4. The best correlation was found at lambda = 1.2. The analysis suggests a nonlinear extensional membrane modulus in the low stress range encountered in the flow channel. In terms of an appropriate strain parameter, the elastic modulus is shown to rise toward the level encountered in micropipette aspiration experiments. The implications of the present findings in modeling of cell mechanics and in cell hemolysis are discussed.

The dynamics of shear disaggregation of red blood cells in a flow channel.

Red blood cell (RBC) rouleaux were formed in a flow channel in the presence of 2 g/dl dextran (molecular weight 76,000). The partial separation of RBC rouleau doublets adhering to the floor of the flow channel in response to small oscillatory shear stresses was observed experimentally. Theoretical analyses on displacement and drag force were performed to determine whether the motion of the cell involves membrane rotation (i.e., rolling) or sliding. From the experimental data and the results of theoretical analyses, it is concluded that, under the conditions of the experiments, the RBCs in a doublet separate from each other by rolling, rather than sliding of the sheared cell.

Erythrocyte tropomodulin binds to the N-terminus of hTM5, a tropomyosin isoform encoded by the gamma-tropomyosin gene.

Tropomodulin is a 40.6-kDa protein that binds to one end of the rod-like tropomyosin and inhibits its cooperativity and binding to actin. In myofibrils, tropomodulin has been localized at or near the free end of thin filaments. Using recombinant and chimeric molecules in a solid-phase binding assay, we demonstrate that it is the N-terminus of tropomyosin that interacts with tropomodulin. Among several tropomyosin isoforms tested, hTM5 encoded by the human gamma-tropomyosin gene has the highest affinity toward human erythrocyte tropomodulin. Tropomodulin may, therefore, regulate the length and/or organization of actin filaments by differential binding to tropomyosin isoforms. hTM5 exists in the human erythrocyte membrane skeleton. In non-muscle cells, tropomodulin may block the head-to-tail association of tropomyosins and their interaction with actin at the pointed end of actin filaments by preferentially binding to TM5 at its N-terminus.

Gene assignment, expression, and homology of human tropomodulin.

Tropomodulin is a newly characterized pointed end capping protein for actin filaments. It binds specifically to the N terminus of tropomyosin and blocks the elongation and depolymerization of tropomyosin-coated actin filaments. A 1.9-kb human tropomodulin cDNA clone was used to map its gene by fluorescence in situ hybridization. The tropomodulin gene was assigned to human chromosome 9q22.2-q22.3, a region that is also known to contain several other genes and disease loci and is proximal to the loci for gelsolin and alpha-fodrin. The gene for tropomodulin is expressed in major human tissues at different levels in the following order: heart and skeletal muscle much greater than that in brain, lung, and pancreas, which is greater than that in placenta, liver, and kidney. Human tropomodulin and a 64-kDa autoantigen in Graves disease (1D) are related: tropomodulin has 42 and 41% identity with the Graves protein in the N-terminal (69 residue) and C-terminal (194 residue) regions, respectively. The insertion of several homologous repeats in the midsection of the Graves protein, together with the extension of a proline-rich C terminus, accounts for the differences in length between the Graves protein (572 residues) and tropomodulin (359 residues). The significant sequence identity indicates that these two genes are evolved from a common ancestral gene.

Dynamic changes in viscoelastic properties in cytotoxic T-lymphocyte-mediated killing.

The biophysical properties of cytotoxic T lymphocytes during the killing of their target cells was investigated by using a human cytotoxic T lymphocyte clone, F1, and the target cell, JY, for which it is specific. In single cytotoxic cell/target cell pairs after their conjugation there are changes in the viscoelastic properties of the target cell in association with the lethal hit delivery and post-binding cytolytic steps. On the basis of these changes in the target cell, the complex cytolytic event can be divided into stages: the viscoelastic coefficients exhibited an initial increase followed by a return to resting values; thereafter these coefficients decreased below control and then rose again prior to lysis. The eventual killing of the target cell involves bubbling and swelling of the nucleus, clustering of granules, damage to the cytoplasmic membrane, cell swelling, and lysis. The viscoelastic changes involved in target cell death suggest the loss of integrity of its cytoskeletal apparatus.

Molecular cloning and characterization of human fetal liver tropomodulin. A tropomyosin-binding protein.

Human erythrocyte tropomodulin is a novel tropomyosin regulatory protein that binds to the end of erythrocyte tropomyosin and blocks heat-to-tail association of tropomyosin along actin filaments. It has been proposed to play a role in modulating the association of tropomyosin with the spectrin-actin complex in the erythrocyte membrane skeleton. Immunoscreening of a human fetal liver cDNA expression library in lambda gt11, followed by 5'-end extension by polymerase chain reaction from the same library, yielded a composite cDNA sequence of 2665 base pairs (bp). It contains a 34-bp 5'-untranslated region, a 1.6-kilobase (kb) 3'-untranslated region, and a complete open reading frame of 1077 bp that encodes a protein of 359 amino acids with a calculated molecular mass of 40.6 kDa and a pI of 4.8. Authenticity of the tropomodulin cDNA was confirmed by a complete sequence match of 49 predicted amino acids with the sequences of three tryptic peptides of the erythrocyte tropomodulin. The sequence has no internal repeats and no significant homology with any known proteins. Secondary structure predictions indicate that tropomodulin may consist of a series of seven or eight short alpha-helical segments and fold into a somewhat compact shape. The tropomyosin binding activity has been mapped to an N-terminal region containing residues 39-138. Nine independent PCR clones, five from a human reticulocyte cDNA library and four from the fetal liver cDNA library, revealed identical N-terminal 103 amino acids, suggesting that the sequence reported here may also be of erythrocyte tropomodulin. Northern analysis of human reticulocyte RNA showed two hybridizing bands of 2.7 and 1.6 kb, indicating that the 2665-bp cDNA sequence reported here was that of the longer transcript.

Tropomodulin-binding site mapped to residues 7-14 at the N-terminal heptad repeats of tropomyosin isoform 5.

Tropomodulin is a globular protein that caps the pointed end of actin filaments by complexing with the N-terminus of a tropomyosin (TM) molecule. TM consists of coiled coils except for the N-terminus, which may be globular. Here we report that human TM isoform 5 (hTM5) lacking the N-terminal 18 residues lost its binding activity toward tropomodulin. We further characterized the tropomodulin-binding site by creating a series of deletion and missense mutations within this region, followed by a solid-phase binding assay. I7, V10, and I14, hydrophobic residues located at the a and d positions of N-terminal heptad repeats involving intertwine, are essential for tropomodulin binding. R12, a positively charged residue at the f position, is also involved in recognition. In contrast, A2R and G3Y mutations, each creating a bulky N-terminus, did not alter the binding. In addition, rat TM5b, which differs from hTM5 in residues 4-6, exhibits a similar binding affinity. The tropomodulin-binding site, therefore, is mapped to residues 7-14 at the beginning of the long heptad repeats. Column chromatography revealed that hTM5 mutants remained capable of dimerization. Results also suggest tropomodulin has a groove-type, rather than a cavity-type, binding site for hTM5. We also mapped the epitope of monoclonal antibody LC1 to residues 4-10 of hTM5 and showed the competition between mAb LC1 and tropomodulin in hTM5 binding. Since the N-terminal residues need to overlap with the C-terminus of TM in their head-to-tail association, this investigation elucidates the mechanisms by which the tropomodulin-hTM5 complex is formed and functions in regulating the actin filaments.

Tropomyosin isoform 5b is expressed in human erythrocytes: implications of tropomodulin-TM5 or tropomodulin-TM5b complexes in the protofilament and hexagonal organization of membrane skeletons.

The human erythrocyte membrane skeleton consists of hexagonal lattices with junctional complexes containing F-actin protofilaments of approximately 33-37 nm in length. We hypothesize that complexes formed by tropomodulin, a globular capping protein at the pointed end of actin filaments, and tropomyosin (TM), a rod-like molecule of approximately 33-35 nm, may contribute to the formation of protofilaments. We have previously cloned the human tropomodulin complementary DNA and identified human TM isoform 5 (hTM5), a product of the gamma-TM gene, as one of the major TM isoforms in erythrocytes. We now identify TM5b, a product of the alpha-TM gene, to be the second major TM isoform. TM5a, the alternatively spliced isoform of the alpha-TM gene, which differs by 1 exon and has a weaker actin-binding affinity, however, is not present. TM4, encoded by the delta-TM gene, is not present either. In sodium dodecyl sulfate-polyacrylamide gel electrophoresis, hTM5 comigrated with the slower TM major species in erythrocyte membranes, and hTM5b comigrated with the faster TM major species. TM5b, like TM5, binds strongly to tropomodulin, more so than other TM isoforms. The 2 major TM isoforms, therefore, share several common features: They have 248 residues, are approximately 33-35 nm long, and have high affinities toward F-actin and tropomodulin. These common features may be the key to the mechanism by which protofilaments are formed. Tropomodulin-TM5 or tropomodulin-TM5b complexes may stabilize F-actin in segments of approximately 33-37 nm during erythroid terminal differentiation and may, therefore, function as a molecular ruler. TM5 and TM5b further define the hexagonal geometry of the skeletal network and allow actin-regulatory functions of TMs to be modulated by tropomodulin. (Blood. 2000;95:1473-1480)