Isolation and physico-chemical properties of blood platelet alpha-actinin

Investigation of calmodulin-like and rod domain mutations suggests common molecular mechanism for α-actinin-1-linked congenital macrothrombocytopenia

Actinin-1 mutations cause dominantly inherited congenital macrothrombocytopenia (CMTP), with mutations in the actin-binding domain increasing actinin's affinity for F-actin. In this study, we examined nine CMTP-causing mutations in the calmodulin-like and rod domains of actinin-1. These mutations increase, to varying degrees, actinin's ability to bundle actin filaments in vitro. Mutations within the calmodulin-like domain decrease its thermal stability slightly but do not dramatically affect calcium binding, with mutant proteins retaining calcium-dependent regulation of filament bundling in vitro. The G764S and E769K mutations increase cytoskeletal association of actinin in cells, and all mutant proteins colocalize with F-actin in cultured HeLa cells. Thus, CMTP-causing actinin-1 mutations outside the actin-binding domain also increase actin association, suggesting a common molecular mechanism underlying actinin-1 related CMTP.

Human CLP36, a PDZ-domain and LIM-domain protein, binds to α-actinin-1 and associates with actin filaments and stress fibers in activated platelets and endothelial cells

A 38-kd protein that associates with F-actin structures in activated platelets and endothelial cells was purified, cloned, and characterized. The protein contains an N-terminal PDZ motif, a large intervening sequence, and a C-terminal LIM domain and was identified as the human homolog of rat CLP36. The study showed that CLP36 associates with actin filaments and stress fibers that are formed during shape change and spreading of platelets and during migration and contraction of endothelial cells. CLP36 binds to α-actinin-1 as shown by coimmunoprecipitation, pull-down experiments, yeast 2-hybrid analysis, and blot overlay assays and colocalizes with α-actinin-1 along endothelial actin stress fibers. In contrast to α-actinin-1, CLP36 was absent from focal adhesions in both activated platelets and endothelial cells. The N-terminal part of CLP36 containing the PDZ domain and the intervening region, but not the LIM domain, targeted enhanced green fluorescent protein fusion proteins to stress fibers in endothelial cells. Yeast 2-hybrid analysis demonstrated that the intervening sequence, but not the PDZ or the LIM domain of CLP36, binds to the spectrinlike repeats 2 and 3 of α-actinin-1. The study further shows that CLP36 binds to α-actinin in resting platelets and translocates as a CLP36/α-actinin complex to the newly formed actin cytoskeleton in activated platelets. The results indicate that CLP36 binds via α-actinin-1 to actin filaments and stress fibers in activated human platelets and endothelial cells. The study suggests that CLP36 may direct α-actinin-1 to specific actin structures and at this position might modulate the function of α-actinin-1.

Human CLP36, a PDZ-domain and LIM-domain protein, binds to α-actinin-1 and associates with actin filaments and stress fibers in activated platelets and endothelial cells

A 38-kd protein that associates with F-actin structures in activated platelets and endothelial cells was purified, cloned, and characterized. The protein contains an N-terminal PDZ motif, a large intervening sequence, and a C-terminal LIM domain and was identified as the human homolog of rat CLP36. The study showed that CLP36 associates with actin filaments and stress fibers that are formed during shape change and spreading of platelets and during migration and contraction of endothelial cells. CLP36 binds to α-actinin-1 as shown by coimmunoprecipitation, pull-down experiments, yeast 2-hybrid analysis, and blot overlay assays and colocalizes with α-actinin-1 along endothelial actin stress fibers. In contrast to α-actinin-1, CLP36 was absent from focal adhesions in both activated platelets and endothelial cells. The N-terminal part of CLP36 containing the PDZ domain and the intervening region, but not the LIM domain, targeted enhanced green fluorescent protein fusion proteins to stress fibers in endothelial cells. Yeast 2-hybrid analysis demonstrated that the intervening sequence, but not the PDZ or the LIM domain of CLP36, binds to the spectrinlike repeats 2 and 3 of α-actinin-1. The study further shows that CLP36 binds to α-actinin in resting platelets and translocates as a CLP36/α-actinin complex to the newly formed actin cytoskeleton in activated platelets. The results indicate that CLP36 binds via α-actinin-1 to actin filaments and stress fibers in activated human platelets and endothelial cells. The study suggests that CLP36 may direct α-actinin-1 to specific actin structures and at this position might modulate the function of α-actinin-1.

Tyrosine phosphorylation of alpha-actinin in activated platelets

The integrin alpha(IIb)beta(3) mediates tyrosine phosphorylation of a 105-kDa protein (pp105) in activated platelets. We have partially purified a 105-kDa tyrosine-phosphorylated protein from platelets stimulated with phorbol 12-myristate 13-acetate and obtained the sequence of an internal 12-mer peptide derived from this protein. The sequence was identical to human alpha-actinin sequences deposited in the Swiss Protein Database. alpha-Actinin, a 105-kDa protein in platelets, was subsequently purified from activated platelets by four sequential chromatographic steps. Fractions were analyzed by Western blotting and probed with alpha-actinin and anti-phosphotyrosine antibodies. The distribution of alpha-actinin and pp105 overlapped throughout the purification. Furthermore, in the course of this purification, a 105-kDa tyrosine-phosphorylated protein was only detected in fractions that contained alpha-actinin. The purified alpha-actinin protein was immunoprecipitated with antibodies to phosphotyrosine in the absence but not in the presence of phenyl phosphate. alpha-Actinin resolved by two-dimensional gel electrophoresis of activated platelet lysates was recognized by the antibodies to phosphotyrosine, whereas pretreatment of the platelets with bisindolylmaleimide, a protein kinase C inhibitor that prevents tyrosine phosphorylation of pp105, inhibited the reactivity of the antibodies to phosphotyrosine with alpha-actinin. Taken together, these data demonstrate that a fraction of alpha-actinin is tyrosine-phosphorylated in activated platelets.

Interactions between smooth muscle alpha-actinin and lipid bilayers

alpha-Actinin has been proposed to be the actin-plasma membrane linker. This assumption is based on the discovery of direct interaction of alpha-actinin with two specific lipids, diacylglycerol and palmitic acid [Burn, P. (1988) Trends Biochem. Sci. 13, 79-83]. In our study, the binding of alpha-actinin with vesicles containing negatively charged phospholipids was measured by the method of 90 degrees light-scattering. Our results show that alpha-actinin is able to bind membranes containing negatively charged phospholipids, but not to bind membranes composed of neutral lipids only. Diacylglycerol and palmitic acid, on the other hand, have little effect on the binding of alpha-actinin to lipid vesicles. Analysis of binding isotherms in terms of a membrane binding model gave apparent dissociation constants which varied between 0.2 and 3 microM over a range of 5-20 mol % negatively charged phospholipid. Comparing the kinetics of alpha-chymotrypsin digestion of alpha-actinin in solution to those of vesicle-bound alpha-actinin, it can be seen that the cleavage site at the junction between the C-terminal and the central rod domain of alpha-actinin and another cleavage site on the C-terminal domain can be most effectively protected by its membrane binding. Analysis of the amide I and II regions of Fourier-transform infrared spectra of alpha-actinin revealed that the association of alpha-actinin with negatively charged phospholipid vesicles resulted in some perturbation of the protein secondary structure. Monolayers containing negatively charged phospholipid were layered and incubated on the surface of a polymerization solution of actin and alpha-actinin, and observed with an electron microscope. The results show that the bundle structure of actin filaments can be formed if diacylglycerol and palmitic acid are present in lipid layers.

Cation effects on the conformations of muscle and non-muscle alpha-actinins

We examined the effects of changing KCl concentration on the secondary structures of alpha-actinins using circular dichroism (CD), 1,1'-bis(4-anilino) naphthalene-5,5'-disulfonic acid (bisANS) fluorescence and proteolysis experiments. Under near-physiological conditions, divalent cations also were added and changes in conformation were investigated. In 25 mM KH2PO4, pH 7.5, increasing KCl from 0 to 120 mM led to decreases in alpha-helix conformation for brain, platelet and heart alpha-actinins (40.5-30.2%, 65.5-37.8% and 37.5-27.8%, respectively). In buffered 120 mM KCl, 0.65 mM calcium produced small changes in the CD spectra of both brain and platelet alpha-actinin, but had no effect on heart alpha-actinin. bisANS fluorescence of all three alpha-actinins also showed significant changes in conformation with increasing KCl. However, in buffered 120 mM KCl increasing concentrations of Ca2+ or Mg2+ did not have significant effects on the bisANS fluorescence of any alpha-actinin. Digestion of brain, platelet and heart alpha-actinins with alpha-chymotrypsin showed an increase of proteolytic susceptibility in 120 mM KCl. These experiments also showed that increasing the concentration of Ca2+ or Mg2+ led to greater changes in digestion fragment patterns in the absence of KCl than in the presence of 120 mM KCl. The results suggest that alpha-actinins exist in different conformations depending on the ionic strength of the medium, which could explain the differences in calcium and F-actin binding results obtained from different alpha-actinins.

Cation binding to chicken gizzard alpha-actinin

Gizzard alpha-actinin binds 45Ca2+ as shown by the calcium overlay method. Flow dialysis measurements in 20 mM Hepes (pH 7.5) reveal 3.5 +/- 1.8 (S.D.) high affinity calcium binding sites per dimer, with Kd1 = 6.36 +/- 0.34 x 10(-6) M, and 87.3 +/- 7.2 sites with Kd2 = 1.66 +/- 0.44 x 10(-4) M. Chymotrypsin and thermolysin digestion yielded peptides of gizzard alpha-actinin which, if they included the putative EF-hands, bound 45Ca2+ in 10 mM imidazole-HCl (pH 7.4) or 60 mM KCl, 10 mM imidazole-HCl (pH 7.4). In addition, peptides which include a region of the molecule more than 27 kDa from the N-terminal also bind calcium. In contrast, when KCl in the binding buffer was increased to 120 mM, calcium binding was eliminated. Flow dialysis data revealed no high-affinity binding and 82.5 +/- 3.3 calcium binding sites with calculated affinities in the millimolar range. These are divalent cation binding sites, not Ca(2+)-specific sites, because they are eliminated by the addition of up to 5 mM Mg2+. Structural changes produced upon cation binding to alpha-actinin measured by circular dichroism, proteolysis and bisANS fluorescence are substantial when binding K+ with only small changes upon binding of Ca2+ or Mg2+ in the presence of 120 mM KCl. These results suggest that monovalent and divalent cations have different effects on different parts of the molecule with a complete elimination of 45Ca2+ binding to the EF-hands being produced by 120 mM KCl.

Distribution and function of organized concentrations of actin filaments in mammalian spermatogenic cells and Sertoli cells

Actin filaments are concentrated in specific regions of spermatogenic cells and Sertoli cells. In spermatogenic cells they occur in intercellular bridges and in the subacrosomal space. In Sertoli cells they are abundant in ectoplasmic specializations and in regions adjacent to tubulobulbar processes of spermatogenic cells. At all of these sites, the filaments are morphologically related to the plasma membrane and+or intercellular membranes, and, as in many other cell types, are arranged in either bundles or networks. In at least two of the locations just indicated (ectoplasmic specializations and intercellular bridges), elements of the ER are closely related to the actin filaments. In tubulobulbar complexes, ER is present but is more distantly related to the filaments. Elements of the ER, when present, may serve a regulatory function. The filaments in ectoplasmic specializations and in regions adjacent to tubulobulbar processes of spermatogenic cells are suspected to be involved with the mechanism by which intercellular junctions are established, maintained, and degraded. In intercellular bridges, actin filaments may serve to reinforce and perhaps regulate the size of the cytoplasmic connections between differentiating germ cells. Filaments in the subacrosomal space may serve as a linking network between the acrosome and nucleus and may also be involved in the capping process. Because of the possibility that the actin filaments discussed before may be related to specific membrane domains involved with intercellular or interorganelle attachment, and that changes in these membrane domains are prerequisite to processes such as sperm release, turnover of the blood-testis barrier, formation of the acrosome, and coordination of spermatogenic cell differentiation, an understanding of exactly how these actin filaments are related to elements in the membrane and how this interaction is controlled is fundamental to our understanding, and perhaps our manipulating, of male fertility. I suspect that working out the molecular organization of these actin filament-containing sites and determining how their organization is controlled will be the major focus of research in this field over the next few years.

Immunochemical analysis of alpha-actinin of nemaline myopathy after two-dimensional electrophoresis

We analyzed alpha-actinin from human skeletal muscle by immunoblotting after two-dimensional electrophoresis. A monoclonal antibody, S alpha 5-17, was established after immunization in Balb/c mouse with crude alpha-actinin fraction from human soleus muscle. Western blotting and indirect immunofluorescence microscopy revealed that the antibody reacted selectively with alpha-actinin from human skeletal muscle and stained in a manner equivalent to that of type 1, 2A, 2B and 2C myofibers and cardiac atrial and ventricular muscles. No reactivity was observed in the arterial smooth muscle layer or in the central and peripheral nervous systems. The antibody exhibited 2 spots with different isoelectric points in a range more basic than that of actin upon immunoblotting after two-dimensional gel electrophoresis, suggesting the presence of 2 variants of alpha-actinin in human skeletal muscle. Analysis of type 2B-deficient muscle with nemaline myopathy or central core disease revealed that type 2B myofibers contained the basic variant, while type 1 and 2A myofibers contained only the acidic variant. Immunoblots performed after two-dimensional gel electrophoresis of muscles with nemaline myopathy revealed alpha-actinin variants indistinguishable from those of control muscles.

Calcium-sensitive, lipid-binding cytoskeletal proteins of the human placental microvillar region

In this study we describe a group of Ca2+-sensitive proteins located in the microvillar region of the human placental syncytiotrophoblast. By following the distribution of proteins between the particulate and supernatant phases of detergent-solubilized microvilli in the presence of defined concentrations of free Ca2+, we demonstrate a class of proteins of subunit molecular weights 72,000, 69,000, 38,000, 36,000, and 32,000 that associate with both the cytoskeleton and lipid at high concentrations of free Ca2+. These proteins can be released from microvilli using EGTA-containing buffers. Although they do not bind to phenyl-Sepharose, they will bind to phospholipids immobilized on phenyl-Sepharose columns in a Ca2+-dependent manner and show a marked preference for phospholipids with negatively charged headgroups. The results provide evidence for a sequence of events which may occur within the microvillus as the localized concentration of intracellular free Ca2+ rises.

Universal calibration of gel permeation chromatography and determination of molecular shape in solution

Gel permeation chromatography (GPC) has become a routine technique for both biology and polymer chemistry. By comparison our theoretical perception of the separation principle of GPC is still immature and conflicting and so is the assessment of the analytical informational content of this method. In order to discriminate between the various parameters that might influence GPC and thus to decide among the numerous propositions of calibration, several odd biopolymers (tropomyosin, spectrin, DNA, tobacco mosaic virus, alpha-actinin, ovomucoid) were selected. They were characterized by analytical ultracentrifugation as well as quasielastic light scattering, and they were compared to globular proteins including icosahedral viruses (tomato bushy stunt virus, turnip yellow mosaic virus, Q beta, MS2) on several different HPLC column matrices. The results demonstrate that the universal calibration principle of GPC is the viscosity radius, i.e., the molecular volume times a shape function which is defined by the intrinsic viscosity. Alternate propositions such as molecular weight, second virial coefficient, diffusion coefficient (Stokes radius), radius of gyration, mean linear projected length, contour length, and related measures seem to be excluded on the basis of the evidence presented. These results help to focus the physical picture which seems to govern GPC. Finally it is demonstrated that GPC is a versatile and unique tool with which to characterize molecular shape and dynamics in solution.

Effect of tropomyosin on the interactions of actin with actin-binding proteins isolated from pig platelets

Several non-muscle tropomyosins have been reported to lack the ability to polymerize in a head-to-tail manner [Dabrowska, R. et al. (1983) J. Muscle Res. Cell Motil. 1, 83-92; Côté, G.P. (1983) Mol. Cell. Biochem. 57, 127-146]. Unlike rabbit skeletal muscle tropomyosin, these proteins could therefore not protect the F-actin microfilaments neither from disassembly or from cross-linking by the other actin-associating factors. However, we have provided evidence that, in vitro, pig platelet tropomyosin, although shorter in molecular length, exhibits the same properties as the muscle protein: it self-associates and forms a 1:6 complex with platelet filamentous actin under physiological conditions [Prulière et al. (1984) J. Muscle Res. Cell Motil. 6, 126]. In this paper, we examine the effects of several other actin-binding proteins on the microfilaments saturated with this non-muscle tropomyosin. Since contractile proteins often vary with the cell type and may require different conditions for their interactions, we have developed a procedure which allows the parallel purification of actin-binding protein (ABP), vinculin, alpha-actinin, gelsolin as well as actin and tropomyosin from the same batch of cells. Thus, using an homogeneous system, we show by viscometry, sedimentation and densitometry, and by electron microscopy, that pig platelet tropomyosin can protect the structure of the microfilaments from the action of the modulating factors to the same extent as rabbit skeletal muscle alpha-tropomyosin. Our data suggest that interaction of ABP, vinculin or alpha-actinin can occur only with the ends of the filaments when F-actin is saturated with tropomyosin, while cross-linking takes place by interactions with sites localized along the entire length of F-actin in the absence of tropomyosin. Moreover, the presence of tropomyosin on F-actin leads to the total inhibition of gelsolin severing activity, although it did not prevent the binding of gelsolin to the F-actin--tropomyosin complex. This suggests that pig platelet as well as skeletal muscle tropomyosins have the ability to increase the strength of the interaction between actin monomers within the filament. This also suggests that the binding sites of gelsolin along the filaments are not localized in the groove of the F-actin helix.

Cytoskeletons of ADP- and thrombin-stimulated blood platelets. Presence of a caldesmon-like protein, alpha-actinin and gelsolin at different steps of the stimulation

Comparative analyses of the cytoskeletons of resting and stimulated platelets point out the involvement of a 79 kDa polypeptide in the activation step and its increased incorporation during aggregation. It appears as a doublet and cross-reacts with an antibody to chicken gizzard caldesmon, whereas no 150 kDa immunoreactive form was detected. alpha-Actinin and gelsolin were detected only in the aggregation step.

Proteins of the human placental microvillar cytoskeleton. alpha-Actinin

The human placental syncytiotrophoblast microvilli are supported by an underlying cytoskeleton consisting mainly of actin microfilaments. The major proteins associated with the actin have Mr values of 105 000, 80 000 and 68 000. The 105 000-Mr protein is recognized by an antibody preparation raised to purified chicken gizzard alpha-actinin. Electron microscopy has shown that the human placental protein has dimensions similar to those reported for muscle alpha-actinin. About half of the placental microvillar alpha-actinin is released from the cytoskeleton in the presence of Ca2+. This effect occurs at concentrations of Ca2+ greater than 0.3 muM and has been used as the basis of a method for the purification of the placental alpha-actinin. This sensitivity to Ca2+ is not affected by trifluoperazine and is therefore likely to be a property of the alpha-actinin as such rather than being mediated via calmodulin.

Properties of two isoforms of human blood platelet alpha-actinin

The structural and functional properties of the aa (2 X 97 kDa) and cc (2 X 94 kDa) isoforms of platelet alpha-actinin have been compared. Structural differences between aa and cc are revealed by their peptide maps, obtained from limited proteolysis, and by their immunological cross-reactivity. Both isoforms stimulate the Mg ATPase activity of actomyosin, bind to F-actin (high-speed sedimentation) and cross-link or gel actin filaments (low-speed sedimentation and viscometry), in a calcium-dependent manner. The study of the interaction with F-actin indicates that the binding of 1 molecule of aa or cc alpha-actinin/9-11 actin monomers is sufficient to produce maximal gelation in the presence of EGTA. CaCl2 at 0.1 mM strongly inhibits the binding of aa to F-actin and weakly that of cc, while it inhibits similarly the cross-linking of either aa or cc. The cross-linking efficiency of cc is 9, 7, 1.7 and 1.3 times higher than that of aa at 4, 20, 30 and 37 degrees C, respectively. The bb form (2 X 96 kDa), which is a proteolytic product of aa [Y. Gache et al. (1984) Biochem. Biophys. Res. Commun. 124, 877-881], behaves roughly as aa, but the calcium sensitivity of its binding to F-actin is intermediate between that of aa and cc. These results suggest that the effect of Ca2+ concentration on the binding of platelet alpha-actinin to F-actin may be partly dissociated from the effect on the cross-linking.

Hydrolysis of cytoskeletal proteins by the Ca2+-dependent protease during platelet activation

During platelet stimulation, the cytosolic Ca2+ concentration increases to micromolar levels. One consequence of this increase is that the Ca2+-dependent protease within platelets is activated. Activation of the Ca2+-dependent protease results in hydrolysis of actin-binding protein and P235. Actin-binding protein and P235 can both affect the organization of actin, thus activation of the Ca2+-dependent protease may provide a regulatory mechanism by which stimulus-induced changes in the organization of actin filaments could be directed. Although both actin-binding protein and P235 affect actin polymerization, stimulus-induced actin polymerization occurs before hydrolysis of actin-binding protein or P235 can be detected, thus it seems unlikely that hydrolysis of these proteins affects actin polymerization. Actin-binding protein also cross-links actin filaments into networks, a function that is lost when it is hydrolyzed by the Ca2+-dependent protease. Thus, hydrolysis of actin-binding protein may result in disruption of the actin filament networks that form early during platelet activation and permit the reorganization of filaments into the bundles present at later stages of platelet activation.

Isolation and some properties of macrophage alpha-actinin: evidence that it is not an actin gelling protein

We have isolated an actin-binding protein from rabbit alveolar macrophages which by virtue of its physical properties we classify as a nonmuscle alpha-actinin. The protein consists of two subunits of Mr 103 000 and has a Stokes' radius of 7.26 nm and a sedimentation coefficient of 6.83 X 10(-13) s-1. Under the electron microscope, rotary-shadowed molecules appeared as short rods with an average length of 39.9 nm. We have examined the nature of the interaction of macrophage alpha-actinin with F-actin. The binding of radioiodinated macrophage alpha-actinin to F-actin is calcium sensitive. At a low concentration of free calcium (less than 10(-9) M), the binding affinity is 4.2 X 10(6) M-1 and is relatively unaffected by changes in temperature, while in the presence of 0.1 mM Ca2+, binding is reduced more than 5-fold. The stoichiometry of binding suggests that alpha-actinin binds all along the length of the actin filaments. The affinity of 45Ca2+ for macrophage alpha-actinin is 4 X 10(6) M-1 with a capacity of four calcium ions per molecule. Although macrophage alpha-actinin has calcium-inhibitable actin gelation activity at 7 degrees C, its effect on the apparent viscosity of F-actin decreases with increasing temperature, and at 37 degrees C, no gel point is observed. Therefore, at the temperature at which macrophages function in vivo, alpha-actinin probably does not promote the isotropic gelation of actin.

Polymorphism of alpha-actinin from human blood platelets. Homodimeric and heterodimeric forms

In purified solutions of alpha-actinin from human blood platelets, three polypeptides a, b and c, of approximately 100 kDa, were observed by polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulfate. They were identified as alpha-actinin subunits on the basis of their cross-reactivity with antibodies against skeletal muscle alpha-actinin and their interaction with F-actin. After electrophoresis in the absence of sodium dodecyl sulfate, six forms were observed: aa, ab, bb, ac, bc, cc. The similarity between the one-dimensional peptide maps of a and b and the absence of b in the presence of calcium-dependent protease inhibitors indicated that b is probably derived from the a subunit. The c subunit differs from the a subunit. The results provide evidence that there are actually only two platelet alpha-actinin subunits a and c which give rise to three isoforms: two homodimers aa and cc, and one heterodimer ac.