Highly homologous filamin polypeptides have different distributions in avian slow and fast muscle fibers

The high molecular weight actin-binding protein filamin is located at the periphery of the Z disk in the fast adult chicken pectoral muscle (Gomer, R. H., and E. Lazarides, 1981, Cell, 23: 524-532). In contrast, we have found that in the slow anterior latissimus dorsi (ALD) muscle, filamin was additionally located throughout the l band as judged by immunofluorescence with affinity-purified antibodies on myofibrils and cryosections. The Z line proteins desmin and alpha-actinin, however, had the same distribution in ALD as they do in pectoral muscle. Quantitation of filamin and actin from the two muscle types showed that there was approximately 10 times as much filamin per actin in ALD myofibrils as in pectoral myofibrils. Filamin immunoprecipitated from ALD had an electrophoretic mobility in SDS polyacrylamide gels identical to that of pectoral myofibril filamin and slightly greater than that of chicken gizzard filamin. Two-dimensional peptide maps of filamin immunoprecipitated and labeled with 125I showed that ALD myofibril filamin was virtually identical to pectoral myofibril filamin and was distinct from chicken gizzard filamin.     

Chicken gizzard filamin, retina filamin and cgABP260 are respectively, smooth muscle-, non-muscle- and pan-muscle-type isoforms: distribution and localization in muscles

We determined the full cDNA sequences of chicken gizzard filamin and cgABP260 (chicken gizzard actin-binding protein 260). The primary and secondary structures predicted by these sequences were similar to those of chicken retina filamin and human filamins. Like mammals, chickens have 3 filamin isoforms. Comparison of their amino acid sequences indicated that gizzard filamin, retina filamin, and cgABP260 were the counterparts of human FLNa (filamin a), b, and c, respectively. Antibodies against the actin-binding domain (ABD) of these 3 filamin isoforms were raised in rabbits. Using immunoabsorption and affinity chromatography, we prepared the monospecific antibody against the ABD of each filamin. In immunoblotting, the antibody against the gizzard filamin ABD detected a single band in gizzard, but not in striated muscles or brain. In brain, only the antibody against the retina filamin ABD produced a strong single band. The antibody against the cgABP260 ABD detected a single peptide band in smooth, skeletal, and cardiac muscle. In immunofluorescence microscopy of muscular tissues using these antibodies, the antibody against the gizzard filamin ABD only stained smooth muscle cells, and the antibody against the retina filamin ABD strongly stained endothelial cells of blood vessels and weakly stained cells in connective tissue. The antibody against the cgABP260 ABD stained the Z-lines and myotendinous junctions of breast muscle, the Z-lines and intercalated disks of cardiac muscle, and dense plaques of smooth muscle. These findings indicate that chicken gizzard filamin, retina filamin, and cgABP260 are, respectively, smooth muscle-type, non-muscle-type, and pan-muscle-type filamin isoforms.     

Phosphorylation of filamin and other proteins in cultured fibroblasts.

Incubation of subcellular fractions of fibroblasts with [³²P]ATP demonstrated 10 phosphoproteins whose phosphorylation can be increased by cyclic AMP or cyclic AMP-dependent protein kinase. One of these phosphoproteins, MW 240,000, resembles the actin binding protein, filamin, and can be selectively precipitated by antibodies to chicken gizzard filamin. Furthermore chicken gizzard filamin can be phosphorylated by skeletal muscle protein kinase and cyclic AMP stimulates this reaction.     

Potentiation of actomyosin ATPase activity by filamin

It was found that thin filaments from chicken gizzard muscle activate skeletal muscle myosin Mg2+-ATPase to a greater extent than does the complex of chicken gizzard actin and tropomyosin. The protein factor responsible for this additional activation has been now identified as the high Mr actin binding protein, filamin.     

A 250K-molecular-weight actin-binding protein from actin-based gels formed in sea urchin egg cytoplasmic extract

The actin-based gel formed at 35 degrees C in the cytoplasmic extract from eggs of a sea urchin, Tripneustes gratilla, contains several high-molecular-weight proteins. Among them, the 250K-molecular-weight protein was isolated and characterized. This protein migrated slightly more slowly than filamin from chicken gizzard upon polyacrylamide gel electrophoresis in the presence of sodium dodecyl sulfate. It reacted only very weakly with antibodies against chicken gizzard filamin or against a high-molecular-weight actin-binding protein from Physarum plasmodia. It did not react with antibodies against chicken erythrocyte alpha-spectrin nor against the 220K protein from the same egg. A chemical crosslinking experiment revealed the presence of dimers in the purified 250K protein preparation. A rotary shadowed specimen of such a preparation showed wavy single-stranded molecules 120-170 nm long, having five to six globular domains, which may represent dimers. The appearance was different from that of spectrin or actin-binding protein from macrophage or chicken gizzard filamin. This protein increased the viscosity of F-actin solution. It bound to F-actin preferably at low KCl concentrations such as 20 mM. The binding ability was not influenced by pH between 6.0 and 7.5, although it was somewhat reduced above pH 8.0. The binding was insensitive to low Ca ion concentrations. Electron microscopy using the negative staining technique supported the idea that this protein crosslinks actin filaments. In addition, a second protein from egg gels, with a reported molecular weight of about 220K (Kane, R.E., J. Cell Biol. 66, 305-315 (1975)), comigrated with human erythrocyte alpha-spectrin on an SDS-gel and reacted with antibodies against chicken erythrocyte alpha-spectrin. This suggests that this protein is a sea urchin egg spectrin. The role of these proteins in the cytoskeleton formation in the sea urchin egg is discussed.     

A 130K protein from chicken gizzard: Its localization at the termini of microfilament bundles in cultured chicken cells

A protein with a molecular weight of 130,000 (130K protein) was extracted from chicken gizzard and purified to homogeneity by ammonium sulfate fractionation and ion-exchange chromatography. Antibodies prepared against the pure protein were used for its immunochemical characterization and immunofluorescent visualization in cultured chicken cells. Both peptide mapping and immunochemical analysis indicated that the 130K protein is not related either structurally or antigenically to other mechanochemical proteins, including α-actinin, actin, myosin, tropomyosin, filamin and tubulin. Immunofluorescent labeling of different cultured embryonic chicken cells (from skin, heart and gizzard) indicated that the label was predominantly organized in intracellular plaques at the bottom of the cells and in some areas of cell-cell contact. Immunoprecipitation of the 130K protein from biosynthetically ³⁵S-methionine-labeled cultured cells, using the pure antibodies and Staphylococcus aureus, resulted in the specific isolation of a single labeled electrophoretic band indistinguishable from the chicken gizzard 130K protein. The 130K protein-rich plaques were found, by interference-reflection microscopy, to coincide with cell substrate adhesion plaques. Double immunofluorescent labeling for the 130K protein and other cytoskeletal proteins (actin, α-actinin and tropomyosin) indicated that the 130K protein-rich areas are localized at the termini of stress fibers. α-Actinin was found in close association with the 130K protein, while tropomyosin was usually excluded from those areas.     

Switching of filamin polypeptides during myogenesis in vitro

During chicken skeletal myogenesis in vitro, the actin-binding protein filamin is present at first in association with actin filament bundles both in myoblasts and in myotubes early after fusion. Later in mature myotubes it is found in association with myofibril Z disks. These two associations of filamin are separated by a period of several days, during which the protein is absent from the cytoplasm of differentiating myotubes (Gomer, R., and E. Lazarides, 1981, Cell, 23:524-532). To characterize the two classes of filamin polypeptides we have compared, by two-dimensional peptide mapping, 125I-labeled filamin immunoprecipitated from myoblasts and fibroblasts to filamin immunoprecipitated from mature myotubes and adult skeletal myofibrils. Myoblast filamin is highly homologous to fibroblast and purified chicken gizzard filamins. Mature myotube and adult myofibril filamins are highly homologous but exhibit extensive peptide differences with respect to the other three classes of filamin. Comparison of peptide maps from immunoprecipitated 35S-methionine-labeled filamins also shows that fibroblast and myoblast filamins are highly homologous but show substantial peptide differences with respect to mature myotube filamin. Filamins from both mature myotubes and skeletal myofibrils exhibit a slightly higher electrophoretic mobility than gizzard, fibroblast, and myoblast filamins. Short pulse-labeling studies show that mature myotube filamin is synthesized as a lower molecular weight variant and is not derived from a higher molecular weight precursor. These results suggest that myoblast and mature myotube filamins are distinct gene products and that during skeletal myogenesis in vitro one class of filamin polypeptides is replaced by a new class of filamin polypeptides, and that the latter is maintained into adulthood.     

Identification of a high molecular weight actin-binding protein in skeletal muscle.

A large polypeptide having a molecular weight of 240,000 as determined by electrophoresis in the presence of sodium dodecyl sulfate has been identified in whole cell homogenates from chick skeletal muscle myoblasts and the rat myoblast L6 cell line. A similar polypeptide was identified in both thigh and breast chicken skeletal muscle, but the latter contained less of this protein per g of tissue. Antibodies made to gizzard filamin (an actin-binding protein having a molecular weight of 240,000) cross-reacted with the partially purified Mr = 240,000 protein from chicken skeletal muscle. With use of the indirect immunofluorescence technique, the filamin antibody localized in the Z-line region of chicken skeletal muscle myofibrils. These results indicate that skeletal muscle contains a filamin-like protein that may form an integral part of the myofibril structure.     

Purification of mammalian filamin. Similarity to high molecular weight actin-binding protein in macrophages, platelets, fibroblasts, and other tissues

We have purified the high molecular weight actin-binding protein, filamin from guinea pig vas deferens. We find this mammalian filamin is very similar to chicken gizzard filamin in subunit molecular weight, amnio acid composition, actin-binding properties, immunological cross-reactivity, and the ability to be phosphorylated by cyclic AMP-dependent protein kinase. Anti-filamin antibodies cross-react with a high molecular weight macrophage actin-binding protein, and with a high molecular weight protein in platelets and fibroblasts. Furthermore like filamin, these proteins are also phosphorylated and cyclic AMP stimulates their phosphorylation. Anti-filamin antibodies do not cross-react with the erythrocyte membrane protein spectrin or with high molecular weight proteins in brain extracts. We conclude that filamin from avian and mammalian smooth muscle are very similar proteins and furthermore that many, but not all, non-muscle cells contain a protein closely related to filamin.     

An abundant and novel protein of 22 kDa (SM22) is widely distributed in smooth muscles. Purification from bovine aorta.

Using a rabbit polyclonal-antibody preparation directed against the chicken gizzard protein, we demonstrated by immunoblotting the presence of the 22 kDa protein (SM22) in a variety of chicken smooth-muscle-containing organs, including uterus, intestine, gizzard, oesophagus and aorta. Protein SM22 was present in only trace amounts in brain, liver and heart, and could not be detected in chicken breast muscle. The antibody preparation did not cross-react with extracts of bovine aorta. However, the presence of SM22 as a major component in bovine aorta and pig carotid was demonstrated by its co-migration with the purified chicken gizzard protein on one- and two-dimensional polyacrylamide electrophoretic gels. Its molar abundance relative to actin was estimated to be 0.9:6.0 and 1.4:6.0 for bovine aorta and pig carotid respectively. Like the chicken gizzard protein, it separates on pH-gradient electrophoresis into at least three variants, alpha, beta and gamma, with similar apparent Mr. Purification of the aorta SM22 showed it to have a similar amino acid composition to the chicken gizzard protein. We conclude that SM22 is widely distributed and an abundant and unique protein component of smooth-muscle tissues of birds and mammals.     

Ca2+-activated, phospholipid-dependent protein kinase catalyzes the phosphorylation of actin-binding proteins

Chicken gizzard vinculin and filamin were found to be phosphorylated by Ca2+-activated, phospholipid-dependent protein kinase (protein kinase C). These two actin-binding proteins serve as substrates for protein kinase C specifically in the free form, whereas they are little phosphorylated by protein kinase C in the presence of F-actin. In contrast, alpha-actinin from chicken gizzard is less susceptible to phosphorylation by protein kinase C, either in the presence or in the absence of F-actin. In light of these data, the possibility that Ca2+ and phospholipid-dependent phosphorylation by protein kinase C may modulate the function of actin-binding proteins has to be considered.     

Smooth muscle cells of the chicken aortic arch differ from those in the gizzard and the femoral artery in the distribution of F-actin, alpha-actinin and filamin

The distribution of F-actin, alpha-actinin and filamin in smooth muscle cells of the chicken was examined by immunofluorescent and immunoelectron microscopy. Those from the gizzard, the femoral artery and the aortic arch were compared. F-Actin labeled by NBD-phallacidin was seen diffusely distributed in the sarcoplasm in the gizzard and the femoral artery, but in the aorta it was observed as streaks and spots, with unstained areas in between. Epon sections of the aortic arch showed that bundles of thin myofilaments run in various directions interspersed with areas mostly occupied by intermediate filaments. alpha-Actinin labelling occurred in dense plaques along the sarcolemma in all the muscles examined. While dense bodies in the sarcoplasm were common and labelled for alpha-actinin in the gizzard and the femoral artery, hardly any were seen in the aortic arch and little labelling for alpha-actinin was observed in the sarcoplasm. Filamin was concentrated along the periphery of dense bodies and plaques in the gizzard and the femoral artery, but it was seen diffusely in the sarcoplasm of the aortic muscle. After chemical skinning of the latter, filamin labelling persisted only in the F-actin bundles, and other areas became negative. The present results show that smooth muscle cells of the aortic arch contrast with those of the gizzard and even with those of the femoral artery in the distribution of F-actin, alpha-actinin and filamin. The mechanisms of contraction and/or stress maintenance in the aortic smooth muscle may be different from those in other smooth muscles.     

Ultrastructural localization of alpha-actinin and filamin in cultured cells with the immunogold staining (IGS) method

Monospecific antibodies to chicken gizzard actin, alpha-actinin, and filamin have been used to localize these proteins at the ultrastructural level: secondary cultures of 14-d-old chicken embryo lung epithelial cells and chicken heart fibroblasts were briefly lysed with either a 0.5% Triton X-100/0.25% glutaraldehyde mixture, or 0.1% Triton X-100, fixed with 0.5% glutaraldehyde, and further permeabilized with 0.5% Triton X-100, to allow penetration of the gold-conjugated antibodies. After immunogold staining (De Mey, J., M. Moeremans, G. Geuens, R. Nuydens, and M. De Brabander, 1981, Cell Biol. Int. Rep. 5:889-899), the cells were postfixed in glutaraldehyde-tannic acid and further processed for embedding and thin sectioning. This approach enabled us to document the distribution of alpha-actinin and filamin either on the delicate cortical networks of the cell periphery or in the densely bundled stress fibers and polygonal nets. By using antiactin immunogold staining as a control, we were able to demonstrate the applicability of the method to the microfilament system: the label was distributed homogeneously over all areas containing recognizable microfilaments, except within very thick stress fibers, where the marker did not penetrate completely. Although alpha-actinin specific staining was homogeneously localized along loosely-organized microfilaments, it was concentrated in the dense bodies of stress fibers. The antifilamin-specific staining showed a typically spotty or patchy pattern associated with the fine cortical networks and stress fibers. This pattern occurred along all actin filaments, including the dense bodies also marked by anti-alpha-actinin antibodies. The results confirm and extend the data from light microscopic investigations and provide more information on the structural basis of the microfilament system.     

Filamin concentration in cleavage furrow and midbody region: frequency of occurrence compared with that of alpha-actinin and myosin

Affinity-purified rabbit antibody to purified chicken gizzard filamin was used in indirect immunofluorescence to localize filamin in dividing chick embryo cells. The antibody was shown to bind only chick embryo cell filamin when whole cell extracts were analyzed by the sensitive sodium dodecyl sulfate-polyacrylamide gel electrophoresis overlay technique described by Adair et al. (1978, J. Cell Biol. 790:281-285). The results show that filamin is located in stress fibers and membrane ruffles during interphase. As cells prophase, the condensing chromosomes are surrounded by diffuse antifilamin staining. No stress fibers are apparent. During metaphase and anaphase, the staining is bright but diffuse. There is often peripheral membrane staining. Filamin is not concentrated in the spindle region but neither is it excluded from the spindle. During cytokinesis, filamin is found highly concentrated in the cleavage furrow in 16 out of 100 cells examined. This frequency of concentration in the furrow is comparable to that observed for alpha-actinin (14%). Myosin concentration in the furrow is more frequent; it is observed in 37% of the cells examined. Neither myosin, alpha-actinin, nor filamin is observed concentrated in the furrow 100% of the time. We conclude that the results are consistent with, but not sufficient to prove, the hypothesis that alpha-actinin and filamin are essential components of the mechanism of cytokinesis.     

Filamin, a new high-molecular-weight protein found in smooth muscle and nonmuscle cells. Purification and properties of chicken gizzard filamin.

Filamin, a major high-molecular-weight protein of chicken gizzard smooth muscle, was purified to homogeneity by salt extraction, ammonium sulfate precipitation, agarose gel filtration, and diethylaminoethylcellulose ion-exchange chromatography. Purified filamin is an asymmetric oligomer consisting of two large subunits of identical size (2 X 250 000 daltons) as indicated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis, chemical cross-linking, sedimentation analysis (s10, wo = 10S) and Stokes'radius estimation (a = 120 A), It has no intersubunit disulfide but appears from oxidation studies to have adjacent thiols near the subunit interface. Filamin contains no amino sugars, methylated lysine, methylated histidine, or hydroxyproline, nor does it exhibit myosin-like ATPase activities. Its amino acid composition and physical properties differ from those of gizzard myosin, for which a pruification procedure is described. Filamin and the protein spectrin of erythrocyte membranes have strikingly similar physical properties, but they are chemically distinct.     

Isolation of a 240-kilodalton actin-binding protein from Dictyostelium discoideum

A high molecular weight actin-binding protein with subunit mass of 240 kilodaltons has been purified from vegetative amoebae of Dictyostelium discoideum. Briefly, a cell extract was prepared by homogenizing vegetative amoebae in 5 mM EGTA, 5 mM 1,4-piperazineethanesulfonic acid, 1 mM dithiothreitol, 0.02% NaN3, pH 7.0, followed by ultracentrifugation at 114,000 X g for 1 h. The 240-kDa protein in this extract was separated from actin by chromatography on ATP-saturated DEAE-cellulose and further purified by chromatography on hydroxylapatite and Sephacryl S-300. The 240-kDa protein increases the low shear viscosity of F-actin. Covalent cross-linking with dimethyl suberimidate demonstrates that the 240-kDa protein can form dimers in high salt (500 mM NaCl). Hydrodynamic studies in high salt demonstrate the presence of an asymmetric dimer (Stokes' radius = 8.6 nm, sedimentation coefficient = 12 S, native molecular weight = 434,000, and frictional ratio = 1.7). Rotary shadowing demonstrates that the monomer is a flexible rod of approximately 70 nm in length that can associate end to end to form a dimer of approximately 140 nm in length. The 240-kDa protein cross-reacts with antibodies to chicken gizzard filamin. The properties of the 240-kDa protein suggest that it is a member of the filamin class of actin-associated proteins.     

Vascular smooth muscle caldesmon

Caldesmon, a major actin- and calmodulin-binding protein, has been identified in diverse bovine tissues, including smooth and striated muscles and various nonmuscle tissues, by denaturing polyacrylamide gel electrophoresis of tissue homogenates and immunoblotting using rabbit anti-chicken gizzard caldesmon. Caldesmon was purified from vascular smooth muscle (bovine aorta) by heat treatment of a tissue homogenate, ion-exchange chromatography, and affinity chromatography on a column of immobilized calmodulin. The isolated protein shared many properties in common with chicken gizzard caldesmon: immunological cross-reactivity, Ca2+-dependent interaction with calmodulin, Ca2+-independent interaction with F-actin, competition between actin and calmodulin for caldesmon binding only in the presence of Ca2+, and inhibition of the actin-activated Mg2+-ATPase activity of smooth muscle myosin without affecting the phosphorylation state of myosin. Maximal binding of aorta caldesmon to actin occurred at 1 mol of caldesmon: 9-10 mol of actin, and binding was unaffected by tropomyosin. Half-maximal inhibition of the actin-activated myosin Mg2+-ATPase occurred at approximately 1 mol of caldesmon: 12 mol of actin. This inhibition was also unaffected by tropomyosin. Caldesmon had no effect on the Mg2+-ATPase activity of smooth muscle myosin in the absence of actin. Bovine aorta and chicken gizzard caldesmons differed in several respects: Mr (149,000 for bovine aorta caldesmon and 141,000 for chicken gizzard caldesmon), extinction coefficient (E1%280nm = 19.5 and 5.0 for bovine aorta and chicken gizzard caldesmon, respectively), amino acid composition, and one-dimensional peptide maps obtained by limited chymotryptic and Staphylococcus aureus V8 protease digestion. In a competitive enzyme-linked immunosorbent assay, using anti-chicken gizzard caldesmon, a 174-fold molar excess of bovine aorta caldesmon relative to chicken gizzard caldesmon was required for half-maximal inhibition. These studies establish the widespread tissue and species distribution of caldesmon and indicate that vascular smooth muscle caldesmon exhibits physicochemical differences yet structural and functional similarities to caldesmon isolated from chicken gizzard.     

Brownian motion of inert tracer macromolecules in polymerized and spontaneously bundled mixtures of actin and filamin

By use of light microscopy and fluorescence photobleaching recovery, we have studied (a) structures that form in a system composed of copolymerized rabbit muscle actin and chicken gizzard filamin and (b) the Brownian motion of inert tracer macromolecules in this matrix. We have used as tracers size-fractionated fluorescein-labeled ficoll and submicron polystyrene latex particles. In F-actin solutions, the relative diffusion coefficient of the tracer was a decreasing function of both tracer size and actin concentration. Also, a percolation transition for latex particle mobility was found to follow a form suggested by Ogston (Ogston, A. G. 1958. Trans. Faraday Soc. 54:1754-1757) for random filament matrices. The inclusion of filamin before polymerization resulted in increased tracer mobility. Below a filamin dimer-to-actin monomer ratio of 1:140, no structural features were observed in the light microscope. At or above this ratio for all actin concentrations tested, a three-dimensional network of filament bundles was clearly discriminated. Latex particles were always excluded from the bundles. By use of a dialysis optical cell in which polymerization could be initiated with very little hydrodynamic stress, we found that filamin can spontaneously bundle F-actin. A simple physical picture explains how dynamics can affect the structural result of coassembly and provides a further hypothesis on the balance between random filament cross-linking and large-scale bundling. Control of this balance may be important in cytoplasmic motile events.     

Calcium-dependent interaction of actin filaments with actin binding protein in the presence of calmodulin and caldesmon

Caldesmon, calmodulin-, and actin-binding protein of chicken gizzard did not affect the process of polymerization of actin induced by 0.1 M KCl. Caldesmon binds to F-actin, thus inhibiting the gelation action of actin binding protein (ABP; filamin). Low shear viscosity and flow birefringence measurements revealed that in a system of calmodulin, caldesmon, ABP, and F-actin, gelation occurs in the presence of micromolar Ca2+ concentrations, but not in the absence of Ca2+. Electron microscopic observations showed the Ca2+-dependent formation of actin bundles in this system. These results were interpreted by the flip-flop mechanism: in the presence of Ca2+, a calmodulin-caldesmon complex is released from actin filaments on which ABP exerts its gelating action. On the other hand, in the absence of Ca2+, caldesmon remains bound to actin filaments, thus preventing the action of ABP.