Properties of normal and mutant polypeptide fragments from the dimer self-association sites of human red cell spectrin.

We have examined the properties and interactions of expressed polypeptide fragments from the N-terminus of the alpha-chain and the C-terminus of the beta-chain of human erythroid spectrin. Each polypeptide comprises one complete structural repeating unit, together with the incomplete repeat that interacts with its partner when spectrin tetramers are formed. The shared repeat thus generated is made up of two helices from the C-terminal part of the beta-chain and one helix from the N-terminus of the alpha-chain. Three mutant beta-chain fragments with amino acid substitutions in the incomplete terminal repeat were also studied. The alpha- and beta-chain fragments were both substantially monomeric, as shown by sedimentation equilibrium. Circular dichroism analysis and thermal denaturation profiles revealed that the complete repeat present in each fragment had entered the stable tertiary fold. Unexpectedly, the conformational stability of the folded beta-chain repeat was found to be grossly perturbed by the mutations, all of them well beyond its C-terminal boundary; possible explanations for this phenomenon are considered. Sedimentation equilibrium showed that in equimolar mixtures the wildtype alpha- and beta-chain peptides formed a 1:1 complex. Mixing curves, observed by circular dichroism, revealed that association was accompanied by an increase in alpha-helicity. From continuous-variation profiles an association constant in the range 1-2 x 10(6) M-1 was inferred. The association was unaffected by the apparently unstructured anionic tail of 54 residues, found at the C-terminus of the spectrin beta-chain. Of the three mutations in the beta-chain fragment, one (an Ala-->Val replacement in the A helix segment of the incomplete repeat) had a relatively small effect on the association with the alpha-chain fragment, whereas Trp-->Arg mutations in the A and in the remote B helix segments were much more deleterious. These observations are consistent with the relative severities of the haemolytic conditions associated with the mutations.

Actomyosin motor in the merozoite of the malaria parasite, Plasmodium falciparum: implications for red cell invasion.

The genome of the malaria parasite, Plasmodium falciparum, contains a myosin gene sequence, which bears a close homology to one of the myosin genes found in another apicomplexan parasite, Toxoplasma gondii. A polyclonal antibody was generated against an expressed polypeptide of molecular mass 27,000, based on part of the deduced sequence of this myosin. The antibody reacted with the cognate antigen and with a component of the total parasite protein on immunoblots, but not with vertebrate striated or smooth muscle myosins. It did, however, recognise two components in the cellular protein of Toxoplasma gondii. The antibody was used to investigate stage-specificity of expression of the myosin (here designated Pf-myo1) in P. falciparum. The results showed that the protein is synthesised in mature schizonts and is present in merozoites, but vanishes after the parasite enters the red cell. Pf-myo1 was found to be largely, though not entirely, associated with the particulate parasite cell fraction and is thus presumably mainly membrane bound. It was not solubilised by media that would be expected to dissociate actomyosin or myosin filaments, or by non-ionic detergent. Immunofluorescence revealed that in the merozoite and mature schizont Pf-myo1 is predominantly located around the periphery of the cell. Immuno-gold electron microscopy also showed the presence of the myosin around almost the entire parasite periphery, and especially in the region surrounding the apical prominence. Labelling was concentrated under the plasma membrane but was not seen in the apical prominence itself. This suggests that Pf-myo1 is associated with the plasma membrane or with the outer membrane of the subplasmalemmal cisterna, which forms a lining to the plasma membrane, with a gap at the apical prominence. The results lead to a conjectural model of the invasion mechanism.

Actin polymerisation regulates integrin-mediated adhesion as well as rigidity of neutrophils.

Activation of adherent neutrophils causes them to convert from selectin-mediated rolling to integrin-mediated immobilisation and migration. Migration is known to depend on formation and redistribution of filamentous (F) actin, but although immobilisation in seconds parallels early cortical actin polymerisation, no link has been proven. We tested the effect of the actin-polymerising agent jasplakinolide (10 microM) on adhesive and mechanical properties of neutrophils. Pretreated cells were able to adhere and roll on immobilised platelets in a flow-based adhesion assay, but whereas untreated rolling cells became immobilised in seconds when chemotactic formyl peptide (fMLP, 0.1 microM) was superfused over them, the cells treated with jasplakinolide continued rolling. Pretreatment with jasplakinolide also blocked de novo expression of integrin CD11b and shape change which otherwise occurred in minutes after treatment with fMLP. Jasplakinolide directly caused actin polymerisation within neutrophils, evidenced by a marked increase in rigidity (resistance to aspiration into a 5 microm micropipette) and increase in association of actin with the Triton-insoluble cytoskeleton. These results indicate that rearrangement of the actin cytoskeleton regulates integrin-mediated adhesion of activated neutrophils, as well as their migration and mechanical properties.

Membrane proteins at the interface of erythrocytes fused by treatment with polyethylene glycol.

Fusion of human red cells through the action of polyethylene glycol gives rise to pairs or higher clusters with a common membrane envelope, in which a barrier at the position of the original interface can be seen in phase contrast. At early times this septum contains lipids, as judged by labelling with a fluorescent lipophile, and transmembrane protein; this was shown by the presence of the preponderant component, band 3, detected by a fluorescent label, covalently attached before fusion at an extracellular site, or by immunofluorescence with anti-band 3 antibody. Ankyrin, which is bound to band 3, is also observed in the septum. The lipid thereafter disappears from the interface, carrying most of the band 3 with it. A continuous membrane skeletal network, defined by the presence of spectrin (detected by immunofluorescent staining in epifluorescence and confocal microscopy) appears to persist for long periods, but in many cells interruptions develop in the septum. In other fused pairs, particularly at longer times, the interface is seen to have vanished completely. Protease inhibitors have no discernible effect on any of these observations. The results suggest a model for the events that follow fusion. Covalent cross-linking of membrane proteins beyond a critical level causes inhibition of fusion, suggesting that proteins, probably the membrane skeletal network, regulate the fusion process. The efficiency of fusion is strikingly dependent on the composition of the isotonic medium, being relatively high at an orthophosphate concentration of 5 mM and minimal at 20 mM.

Interaction of dystrophin fragments with model membranes.

The interaction with membrane lipids of recombinant fragments of human dystrophin, corresponding to a single structural repeating unit of the rod domain, was examined. Surface plasmon resonance, constant-pressure isotherms in a Langmuir surface film balance, and interfacial rheology were used to observe binding of the polypeptides and its effects on the properties of the lipid film. Modification of the monolayer properties was found to depend on the presence of phosphatidylserine in the lipid mixture and on the native tertiary fold of the polypeptide; thus a fragment with the minimum chain length required for folding (117 residues) or longer caused a contraction of the surface area at constant pressure, whereas fragments of 116 residues or less had no effect. The full extent of contraction was reached at a surface concentration of lipid corresponding to an average area of about 42 A2 per lipid molecule. A dystrophin fragment with the native, folded conformation induced a large increase in surface shear viscosity of the lipid film, whereas an unfolded fragment had no effect. Within a wide range of applied shear, the shear viscosity remained Newtonian. Binding of liposomes to immobilized dystrophin fragments could be observed by surface plasmon resonance and was again related to the conformational state of the polypeptide and the presence of phosphatidylserine in the liposomes. Our results render it likely that intact dystrophin interacts directly and strongly with the sarcolemmal lipid bilayer and grossly modifies its material properties.

Further analysis of the role of spectrin repeat motifs in alpha-actinin dimer formation.

Protein constructs consisting of repeats 1-4, repeats 1-3 and repeats 2-4 of the rod domain of chicken alpha-actinin were expressed as fusion proteins in Escherichia coli. Based on the evidence of circular dichroism spectra and cooperative thermal unfolding profiles both truncated rod fragments were judged to have assumed the native structural fold. The thermal stabilities were in both cases significantly lower than that of the intact rod (repeats 1-4). Analyses by sedimentation equilibrium and velocity provided further evidence to show that fragment 1-4 is entirely dimeric in the concentration range of these experiments, resembling therefore the rod domain isolated by proteolytic digestion of native alpha-actinin. Fragment 2-4, and probably also 1-3, show concentration-dependent association, with dissociation constants, estimated by sedimentation equilibrium, in the 1-10 microM range. Thus, in confirmation of earlier work, all four repeats are required to generate a maximally stable anti-parallel dimer (Kd approximately 10 pM), suggesting the presence of binding sites in all of them to allow for aligned pairing.

Physical properties of dystrophin rod domain.

We have prepared two fragments of the human dystrophin rod domain, each containing eight spectrin-like repeating units, by expression in Escherichia coli. The first corresponds to the central portion of the rod, the other to three repeats from the N-terminal end, fused to five repeats from the C-terminal end. The latter makes up the entire mutant rod, found in a patient with mild (Becker-type) muscular dystrophy. Both fragments were found to possess an ordered, stable structure, and had the form of short rod-like particles in the electron microscope. Molecular weight determinations by sedimentation equilibrium revealed that both polypeptides were monomeric in solution, suggesting that the dystrophin rod domain is incapable of forming an antiparallel homodimer. This supports the inference from sequence analyses [Winder et al., 1995: FEBS Lett. 369:27-33, 1996: Biochem. Soc. Trans. 24:2805] that the dystrophin rod domain lacks the arrangement of sites required for lateral self-association, and that dystrophin, unlike the other known proteins of the spectrin superfamily, may thus exist as a monomer.

Stability of the dystrophin rod domain fold: evidence for nested repeating units.

An examination of fragments of the human dystrophin rod domain, corresponding to a single structural repeating unit, showed that a critical chain length, defined with a precision of one residue at the C-terminal end, is required for formation of the native tertiary fold. We report here that extending the chain by six residues beyond this minimum results in a large increase in conformational stability. This is not related to a change in association state of the polypeptide. The results support the conjecture that successive repeating units in the rod domain of the spectrinlike proteins form a nested structure, in which the N-terminal part of the three-helix bundle of one repeat packs into the overlapping structure of the preceding repeat. This would be expected to affect functional characteristics related to flexibility of the dystrophin rod domain.

Contractile protein system in the asexual stages of the malaria parasite Plasmodium falciparum.

F-actin was detected in asexual-stage Plasmodium falciparum parasites by fluorescence microscopy of blood films stained with fluorescent phalloidin derivatives. F-actin was present at all stages of development and appeared diffusely distributed in trophic parasites, but merozoites stained strongly at the poles and peripheries. No filament bundles could be discerned. A similar distribution was obtained by immunofluorescence with 2 polyclonal anti-actin antibodies, one of which was directed against a peptide sequence present only in parasite actin (as inferred from the DNA sequence of the gene). A monoclonal anti-actin antibody stained very mature or rupturing schizonts but not immature parasites. Myosin was identified in immunoblots of parasite protein extracts by several monoclonal anti-skeletal muscle myosin antibodies, as well as by a polyclonal antiserum directed against a consensus conserved myosin sequence (IQ motif). The identity of the polypeptides recognised by these antibodies was confirmed by overlaying blots with biotinylated F-actin. The antiserum and one of the monoclonal antibodies were used in immunofluorescence studies and were found to stain all blood-stage parasites, with maximal intensity towards the poles of merozoites. Our results are consistent with the presence of an actomyosin motor system in the blood-stage malaria parasite.

Origins of the parasitophorous vacuole membrane of the malaria parasite: surface area of the parasitized red cell.

There is conflicting evidence on whether the parasitophorous vacuole membrane, in which the malaria parasite becomes encapsulated when it enters the red cell, represents a part of the host cell membrane or is derived, at least in part, from the parasite. We have measured the surface area of populations of red cells before and after invasion by up to four merozoites of the malaria parasite, Plasmodium falciparum. The dimensions of the merozoite are such that, if it enveloped itself entirely in host cell membrane during entry, the loss of surface area would amount to some 4 square microns 2 or 3% of the total for each parasite internalized. Our measurements show that within the 99% level of confidence any loss of surface area is less than 1 square micron 2 per parasite internalized. Area measurements on red cells that have been allowed to lose known proportions of their membrane by metabolically induced vesiculation reveal, moreover, that diminutions in surface area in the range of interest are readily detectable. Our observations on recently invaded (young ring-stage) parasites appear to exclude any significant change in surface area of the host cell following invasion. This implies that, if indeed there is internalization of host cell membrane lipid on invasion, as the best evidence shows, it is compensated by parasite-derived lipid, and conversely the parasitophorous vacuole membrane probably contains a contribution of parasite-derived material, presumably that seen to be discharged by the apical organelles, the rhoptries, at the time of invasion.

Association of structural repeats in the alpha-actinin rod domain. Alignment of inter-subunit interactions.

Fragments of the rod domain of chicken alpha-actinin, which comprises four spectrin-like repeat sequences, have been prepared by expression in Escherichia coli. Electron microscopy reveals that all products containing three or four complete repeats are rod-like. Self-association of fragments was detected by chemical cross-linking and analytical equilibrium sedimentation. The intact rod domain forms a stable dimmer, which does not dissociate measurably in the accessible concentration range. Elimination of either terminal repeat (repeat 1 or repeat 4) greatly diminishes the extent of dimerisation. The fragment comprising repeats 1-3 dimerises appreciably, with an association constant estimated from the sedimentation equilibrium distribution of approximately 5 x 10(5) M-1. The fragment made up of repeats 2-4 dimerises to a small extent, but also forms aggregates at high concentrations. The results are most easily reconciled with an aligned structure for the rod domain in solution, in which repeat 1 associates with repeat 4 of the partnering chain, and repeat 2 with repeat 3, rather than with a staggered structure, in which one of the terminal repeats does not participate in dimerisation. Possible explanations for the apparent difference observed between the alpha-actinin rod structure in solution and in two-dimensional crystalline arrays are examined.

Perturbation of red blood cell membrane rigidity by extracellular ligands.

It is known that binding of extracellular antibodies against the major sialoglycoprotein, glycophorin A, reduced the deformability of the red blood cell membrane. This has been taken to result from new or altered interactions between the glycophorin A and the membrane skeleton. We have shown by means of the micropipette aspiration technique that antibodies against the preponderant transmembrane protein, band 3, induce similar effects. A definite but much smaller reduction in elasticity of the membrane is engendered by univalent Fab fragments of the anti-band 3 antibodies. By examining cells genetically devoid of glycophorin A or containing a variant of this constituent, truncated at the inner membrane surface, we have shown that the anti-band 3 antibodies do not act through the band 3-associated glycophorin A. We examined the effect of anti-glycophorin A antibodies on homozygous Wr(a+b-) cells, in which an amino acid replacement in band 3 annihilates the Wright b (Wrb) epitope (comprising sequence elements of glycophorin A and band 3) and thus, by implication disrupts or perturbs the band 3-glycophorin A interaction; these cells show a much smaller response to an anti-glycophorin A antibody than do normal controls. We infer that in this case anti-glycophorin A antibodies exert their rigidifying effect through the associated band 3. Another anti-glycophorin A antibody, directed against an epitope remote from the membrane surface, however, increases the rigidity of both Wr(a+b-) and normal cells. This implies that not all antibodies act in the same manner in modifying the membrane mechanical properties. The effect exerted by anti-band 3 antibodies appears not to be transmitted through the band 3-ankyrin-spectrin pathway because the rigidifying effect of the intact antibody persists at alkaline pH, at which there is evidence that the ankyrin-band 3 link is largely dissociated. The large difference between the effects of saturating concentrations of the divalent and univalent anti-band 3 antibodies implies the existence of an overriding effect on rigidity, resulting from the bifunctionality of the intact antigen. Freeze-fracture electron microscopy shows that the anti-band 3 promotes the formation of small clusters of intra-membrane proteins. Extracellular ligands may in general act by promoting strong or transient interactions between integral membrane proteins, thereby impeding local distortion of the membrane skeletal network in response to shear.

Minimum folding unit of dystrophin rod domain.

Fragments of the rod domain of dystrophin, which consists of spectrin-like repeating sequences, have been prepared by expression in Escherichia coli. The phasing established earlier for the dystrophin rod, as well as for Drosophila spectrin and smooth muscle alpha-actinin, suggested a length of less than 113 residues for the dystrophin repeat that we have chosen. Fragments with a common N-terminus and lengths between 113 and 119 residues were prepared. The formation of the stable native tertiary fold could be recognized by resistance to proteolysis, the circular dichroism spectrum in the regions of both peptide and aromatic absorption bands and the resolution of the long-wavelength component in the tryptophan absorption spectrum. It was found that the critical length for folding was 117 residues: shortening the chain by 1 further residue resulted in loss of the capacity to form a defined tertiary structure. Residue 117 is a glutamine; replacement of this by a methionine residue did not impair the ability of the chain to enter the folded conformation, implying that it is the length of the C-terminal alpha-helix, rather than any specific side-chain interaction, that is critical in determining the stability of the native structure. The fragment of 119 residues forms a significantly more stable structure than that of 117. It appears that the minimum unit capable of forming the native fold extends some residues into the adjoining sequence repeat.

Interaction of protein 4.1 with the red cell membrane: effects of phosphorylation by protein kinase C.

Phosphorylation with endogenous protein kinase C causes the membrane skeletal protein, band 4.1, to lose its capacity to attach to one of two classes of high-affinity binding sites on the red cell membrane. These sites are the ones eliminated by proteolysis in situ of glycophorin C; the surviving type of site is located in a C-terminal peptide of the glycophorin C, retained on the membrane after proteolysis, which is also the site of attachment of p55. A synthetic peptide, comprising the 28 C-terminal residues of glycophorin C, also binds protein 4.1. Phosphorylation of the intact cells, stimulated by phorbol ester, approximately halves the retention of glycophorin C in the membrane cytoskeletons prepared from these cells and reduces the affinity of extracellular glycophorin C epitopes for their antibody.

Association state of human red blood cell band 3 and its interaction with ankyrin.

We have studied the association state of band 3, the anion channel and predominant transmembrane protein of the human red blood cell, and the anomalous stoichiometry and dynamics of its interaction with ankyrin, which acts as a link to the spectrin of the membrane skeletal network. Band 3 exists in benign nonionic detergent solutions as a dimer. Tetramer is formed irreversibly in the course of manipulations, particularly in ion-exchange chromatography. The dimer in solution binds ankyrin without self-associating. In ankyrin-free inside-out membrane vesicles and when incorporated into phosphatidylcholine liposomes, only some 10% to 15% of band 3 chains bind ankyrin at saturation. Moreover, in liposomes this was independent of protein:lipid ratio between 1:2 and 1:40. The bound fraction of band 3 remains with the detergent-extracted membrane cytoskeleton, but is released if the ankyrin has been cleaved with chymotrypsin before detergent treatment; thus, the attachment to the membrane cytoskeleton is entirely through ankyrin and not through other constituents such as protein 4.1. The ratio of band 3 to ankyrin in this complex implies that it consists of two chains of band 3 and one chain of ankyrin, at least after detergent extraction. The bound and free populations of band 3 exchange freely in the membrane. In the artificial liposome membrane binding of ankyrin to band 3 dimers cause association of the band 3 into higher aggregates, as seen in freeze-fracture electron microscopy. Successive manipulations of the red blood cell membrane, which are involved in the preparation of ghosts, of inside-out vesicles, and of inside-out vesicles stripped of peripheral proteins are accompanied by progressive aggregation of intramembrane particles, as judged by freeze-fracture electron microscopy. Thus the intramembrane particles are evidently stabilized in the intact cell by the peripheral protein network and the cytosolic milieu. Aggregation may be expected to limit the number of functional ankyrin binding sites. However, although extraneous ankyrin binds to the unoccupied binding site on the spectrin tetramers in intact ghost membranes, little or no ankyrin can bind to the unoccupied band 3 dimers in situ, perhaps by reason of occlusion of binding sites by the membrane skeletal network.

Concentrated Tris solutions for the preparation, depolymerization, and assay of actin: application to erythroid actin.

High concentrations of Tris are effective in dissociating actin-containing complexes, such as the red cell membrane cytoskeleton. A preparative procedure for red cell actin is based on the dissociation of the membrane skeletal complex in a buffer containing 1 M Tris hydrochloride, followed by gel filtration chromatography in the same medium. The actin is recovered as the monomer and is fully native, as judged by its critical concentration of polymerization, inhibition of DNase I, stimulation of myosin ATPase, and the appearance in the electron microscope of filaments, both bare and decorated with heavy meromyosin, and of magnesium ion-induced paracrystals. The Tris solution causes rapid depolymerization of F-actin with no denaturation, and the solution of monomeric actin in this medium is stable for many weeks in the cold; concentrated Tris is more reliable than guanidinium chloride for the depolymerization of F-actin in the estimation of total actin concentration by the DNase I inhibition assay.

Mechanism of regulation of malarial invasion by extraerythrocytic ligands.

Invasion of red cells by Plasmodium falciparum in vitro was inhibited by a range of extracellular ligands, none of which block the major receptors for merozoites. Most effective, in terms of dose response, were two monoclonal antibodies against the Wrb antigen on glycophorin A; wheat germ agglutinin which also binds to glycophorin, and an anti-band 3 monoclonal antibody, caused inhibition of invasion at higher levels of saturation, while concanavalin A, which binds to band 3, was without effect. All the ligands except concanavalin A, increased the rigidity of the host cell membrane. The anti-Wrb antibodies generated the highest dose response effect, but no correlation between invasion and shear elastic modulus of the membrane could be established. All ligands, with the exception of concanavalin A, caused a reduction in the translationally mobile fractions of band 3 and glycophorin, as revealed by fluorescence recovery after photobleaching (FRAP). Invasion diminished with loss of mobile band 3, engendered by bound wheat germ agglutinin or anti-band 3, falling precipitately when the mobile fraction fell below 40% of that in unperturbed membranes. Both anti-Wrb antibodies suppressed invasion completely at concentrations insufficient to affect significantly either membrane rigidity or intramembrane protein diffusion. A univalent anti-glycophorin A (Fab) fragment, the parent antibody of which was previously shown to inhibit invasion strongly, had only a modest effect on invasion and induced a correspondingly small change in the mobile fraction of band 3.(ABSTRACT TRUNCATED AT 250 WORDS)

Analysis of the phasing of four spectrin-like repeats in alpha-actinin.

Selected fragments of the central rod of chicken gizzard alpha-actinin were expressed as fusion proteins in Escherichia coli, with the aim of determining the positions in the sequence of the four successive spectrin-like repeats that make up this domain. The criteria for an independently folding unit were resistance to proteolysis and the high alpha helicity characteristic of the native protein. Sequences containing repeats 1-4, 2-4, 3-4 and 4 all generated stable fragments on digestion with trypsin and/or thermolysin and N-terminal sequencing gave the most probable starting position of each repeat. The sequences of all four inferred repeats and the sequences of the entire rod, were separately expressed and were shown to assume a stable, protease-resistant fold in solution. The repeat boundaries established in this way differed from those originally deduced from sequence alignments; the N-terminal boundaries of the repeats were 14-24 residues nearer the C-terminus than predicted. The ability to express individual repeats should facilitate identification of the binding sites for the cytoplasmic domains of beta 1 integrins and intercellular cell adhesion molecule-1 which have been localised to the rod domain of alpha-actinin.