Evidence for a direct, nucleotide-sensitive interaction between actin and liver cell membranes

Developmental expression of the alpha-skeletal actin gene

Background

Actin is a cytoskeletal protein which exerts a broad range of functions in almost all eukaryotic cells. In higher vertebrates, six primary actin isoforms can be distinguished: alpha-skeletal, alpha-cardiac, alpha-smooth muscle, gamma-smooth muscle, beta-cytoplasmic and gamma-cytoplasmic isoactin. Expression of these actin isoforms during vertebrate development is highly regulated in a temporal and tissue-specific manner, but the mechanisms and the specific differences are currently not well understood. All members of the actin multigene family are highly conserved, suggesting that there is a high selective pressure on these proteins.

Results

We present here a model for the evolution of the genomic organization of alpha-skeletal actin and by molecular modeling, illustrate the structural differences of actin proteins of different phyla. We further describe and compare alpha-skeletal actin expression in two developmental stages of five vertebrate species (mouse, chicken, snake, salamander and fish). Our findings confirm that alpha-skeletal actin is expressed in skeletal muscle and in the heart of all five species. In addition, we identify many novel non-muscular expression domains including several in the central nervous system.

Conclusion

Our results show that the high sequence homology of alpha-skeletal actins is reflected by similarities of their 3 dimensional protein structures, as well as by conserved gene expression patterns during vertebrate development. Nonetheless, we find here important differences in 3D structures, in gene architectures and identify novel expression domains for this structural and functional important gene.

Acceleration of actin polymerization and rapid microfilament reorganization in cultured hepatocytes by cyclochlorotin, a hepatotoxic cyclic peptide

Cyclochlorotin (= chloropeptide, CC) is a hepatotoxic mycotoxin of Penicillium islandicum Sopp. The effect of CC on actin polymerization was examined by the measurement of changes in fluorescence intensity using pyrene-labeled actin and high shear viscosity. In the presence of CC, the time course of actin polymerization was accelerated in a dose dependent manner (2.5 ng/ml-2.5 microg/ml), without affecting the final level of viscosity. CC exerted a strong stabilizing effect on actin, enabling it to maintain its filamentous form in the presence of members of actin-binding proteins, including those of the gelsolin family prepared from hepatocytes. Microscopic observation revealed that in cultured hepatocytes, 1.0 microg/ml of CC induced bleb formation and changes in the microfilament. These observations indicated that after contact of the hepatocyte with CC, the following events were probable. The toxin passed through the cell membrane by a transport system and immediately reacted with the actin-actin binding proteins underlying the lipid bilayer. Bleb formation and hepatotoxicity were thus induced.

Comparative effects of hepatocyte growth factor and epidermal growth factor on motility, morphology, mitogenesis, and signal transduction of primary rat hepatocytes

Hepatocyte growth factor (HGF) and epidermal growth factor (EGF) are major hepatocyte mitogens, but HGF, also known as scatter factor (SF), has also been shown as a potent motogen for epithelial and endothelial cells. The mechanisms by which HGF is a stronger motogen compared to other mitogens are not understood. Here we report a comparative study of the effect of the two growth factors on cultured primary rat hepatocytes regarding their differential effects on morphology, mitogenicity, and motility as well as the phosphorylation of cytoskeletal-associated proteins. Using three different motility assays, both HGF and EGF increased the motility of hepatocytes, but HGF consistently elicited a significantly greater motility response than EGF. Additionally, HGF induced a more flattened, highly spread morphology compared to EGF. To examine if HGF and EGF phosphorylated different cytoskeletal elements as signal transduction targets in view of the observed variation in morphology and motility, primary cultures of 32P-loaded rat hepatocytes were stimulated by either HGF or EGF for up to 60 min. Both mitogens rapidly stimulated four isoforms of MAP kinase with similar kinetics and also rapidly facilitated the phosphorylation of cytoskeletal-associated F-actin. Two cytoskeletal-associated proteins, however, were observed to undergo rapid phosphorylation by HGF and not EGF during the time points described. One protein of 28 kDa was observed to become phosphorylated fivefold over controls, while the EGF-stimulated cells showed only a slight increase in the phosphorylation of this protein. Another protein with an apparent mwt of 42 kDa was phosphorylated 20-fold at 1 min and remained phosphorylated over 50-fold over control up to the 60 min time point. This protein was observed to become phosphorylated by EGF only after 10 min, and to a lesser extent (20-fold). Taken together, the data suggest that HGF and EGF stimulate divergent as well as redundant signal transduction pathways in the hepatocyte cytoskeleton, and this may result in unique HGF- or EGF-specific motility, morphology, and mitogenicity in hepatocytes.

Yeast mitochondria contain ATP-sensitive, reversible actin-binding activity

Sedimentation assays were used to demonstrate and characterize binding of isolated yeast mitochondria to phalloidin-stabilized yeast F-actin. These actin-mitochondrial interactions are ATP sensitive, saturable, reversible, and do not depend upon mitochondrial membrane potential. Protease digestion of mitochondrial outer membrane proteins or saturation of myosin-binding sites on F-actin with the S1 subfragment of skeletal myosin block binding. These observations indicate that a protein (or proteins) on the mitochondrial surface mediates ATP-sensitive, reversible binding of mitochondria to the lateral surface of microfilaments. Actin copurifies with mitochondria during subcellular fractionation and is released from the organelle upon treatment with ATP. Thus, actin-mitochondrial interactions resembling those observed in vitro may also exist in intact yeast cells. Finally, a yeast mutant bearing a temperature-sensitive mutation in the actin-encoding ACT1 gene (act1-3) displays temperature-dependent defects in transfer of mitochondria from mother cells to newly developed buds during yeast cell mitosis.

The inositol 1,4,5-trisphosphate receptor is localized on specialized sub-regions of the endoplasmic reticulum in rat liver

Inositol 1,4,5-trisphosphate (InsP3) is involved in the mobilization of Ca2+ from intracellular non-mitochondrial stores. In rat liver, it has been shown that the InsP3-binding site co-purifies with the plasma membrane. This suggests that in the liver the InsP3 receptor (InsP3R) associates with plasma membrane. We studied the subcellular distribution of the liver InsP3R by measuring the maximal binding capacity of [3H]InsP3 and using antibodies against the 14 C-terminal residues of the type 1 InsP3R. The antibodies recognized a large amount of an InsP3R protein of 260 kDa in a membrane fraction which is also enriched with [3H]InsP3-binding sites and with markers of the basal, the lateral and the bile-canalicular membrane and the plasma-membrane Ca2+ pump (PMCA). The fractions enriched in markers of the endoplasmic reticulum (ER) and the Ca2+ pump of the ER (SERCA2b) contained low levels of InsP3 receptors. The immunofluorescent labelling of cultured hepatocytes with anti-InsP3R antibodies indicated that the receptor is concentrated in the perinuclear area and in some regions near the plasma membrane. The fraction enriched with InsP3R is also contaminated with markers of the ER and with SERCA2b. It was exposed to alkaline medium (pH 10.5) to extract endogenous actin and membrane-associated proteins before being subfractionated by Percoll-gradient centrifugation. The alkaline treatment allowed partial separation of the markers of the ER from the markers of the plasma membrane. The InsP3R was recovered in the heavy subfraction, which was also enriched with markers for the ER and with the SERCA2b and contained low levels of markers of the plasma membrane. These data indicate that the InsP3R is neither localized on the plasma membrane itself nor homogeneously distributed on the ER membrane. This supports the view that part of the receptor is localized on a specialized sub-region of the ER which interacts with the plasma membrane.

Molecular cloning and expression of the chicken smooth muscle gamma-actin mRNA

We have investigated the expression of chicken smooth muscle gamma-actin mRNA by isolation and characterization of cDNAs representing this actin isoform and utilizing the cDNA to probe RNA from adult and developing cells. Nucleotide sequence elucidated from an apparent full length smooth muscle gamma-actin cDNA revealed that it contained 94 bp of 5' non-translated sequence, an open reading frame of 1131 bp, and 97 bp of 3' non-translated sequence. Within the 376 amino acid sequence deduced from the chicken cDNA were diagnostic amino acids at the NH2- and COOH-terminal regions which provided unequivocal identification of the gamma-enteric smooth muscle actin isoform. In addition, the chicken gamma-enteric actin deduced from our cDNA clones was found to differ from the sequence reported in earlier protein studies [J. Vandekerckhove and K. Weber, FEBS Lett. 102:219, 1979] by containing a proline rather than a glutamine at position 359 of the protein, indicating that the avian gamma-enteric actin isoform is identical to its mammalian counterpart. Comparison of the 5' and 3' non-translated sequence determined from the chicken cDNA to that elucidated for rat, mouse, and human showed that there is not a high degree of cross-species sequence conservation outside of the coding regions among these mRNAs. Northern hybridization analyses demonstrated that the gamma-enteric actin mRNA is expressed in adult aorta and oviduct tissues but not in adult skeletal muscle, cardiac muscle, liver, brain, and spleen tissues. The gamma-enteric actin mRNA was first observed in measurable quantities in gizzard tissue from 4-5 day embryos and increased in content in developing smooth muscle cells through 16-17 embryonic days. Following this initial increase during embryonic development, the gamma-enteric actin mRNA exhibits a decline in content until approximately 7 days posthatching, after which there is an increase in content to maximal levels found in adult gizzard tissue. In general, the developmental appearance of the gamma-enteric mRNA parallels that observed for this protein in previous studies indicating that the developmental expression of smooth muscle gamma-actin is regulated, in part, by an increased content of mRNA in chicken visceral smooth muscle cells during myogenesis.

Diacylglycerol-stimulated formation of actin nucleation sites at plasma membranes

Diacylglycerols, which are generated during phospholipase-catalyzed hydrolysis of phospholipids, stimulated actin polymerization in the presence of highly purified plasma membranes from the cellular slime mold Dictyostelium discoideum. The increased rate of actin polymerization apparently resulted from de novo formation of actin nucleation sites rather than uncapping of existing filament ends, because the membranes lacked detectable endogenous actin. The increased actin nucleation was mediated by a peripheral membrane component other than protein kinase C, the classical target of diacylglycerol action. These results indicate that diacylglycerols increase actin nucleation at plasma membranes and suggest a mechanism whereby signal transduction pathways may control cytoskeletal assembly.

A comprehensive analysis of the developmental and tissue-specific expression of the isoactin multigene family in the rat

The present study represents the first comprehensive analysis of isoactin gene expression in the developing rat. Our results clearly demonstrate that the developmental and tissue-specific expression of the actin multigene family is a highly integrated and complex process involving a variety of regulatory paradigms. The distinct temporal patterns of expression reported in this study indicate that there are three key phases in the regulation of expression of the actin multigene family during development. These include early embryonic development, late fetal development, and early postnatal development. The specific spatial patterns of expression observed in this study demonstrate that the expression of the actin multigene family is much more permissive than previously reported. This permissive expression includes a wide range of "ectopic" expression of the striated muscle isoactins as well as an extended expression of the alpha-smooth muscle isoactin. These findings expand our current understanding of the expression of the actin multigene family in development and provide a fundamental basis for future studies directed at investigating these processes.

Binding of actin to liver cell membranes: the state of membrane-bound actin

Previous work has shown that actin binds specifically and saturably to liver membranes stripped of endogenous actin (Tranter, M. P., S. P. Sugrue, and M. A. Schwartz. 1989. J. Cell Biol. 109:2833-2840). Scatchard plots of equilibrium binding data were linear, indicating that binding is not cooperative, as would be expected for F- or G-actin. To determine the state of membrane-bound actin, we have analyzed the binding of F- and G-actin to liver cell membranes. G-actin in low salt depolymerization buffer and EF-actin, a derivative that polymerizes very poorly in solution, bind to liver cell membranes as well as untreated actin in polymerization buffer. Phalloidin-stabilized F-actin binds, but to a lesser extent. The binding of F- and G-actins are mutually competitive and are inhibited by ATP, suggesting that both forms of actin bind to the same sites. For untreated actin in polymerization buffer, the time course of binding is biphasic, with an initial rapid component which is followed by a plateau phase, then a second, slower component. The binding kinetics of pure F-actin and pure G-actin are both monophasic and match the fast and slower components, respectively, of untreated actin. In the reconstituted system, membrane-bound actin does not stain with rhodamine-phalloidin, nor are actin filaments detected by EM. Distinct regions of amorphous material, however, are visible, which stain with an anti-actin antibody. The exact nature of this material has yet to be determined. A model of actin binding is presented.

Molecular links between the cytoskeleton and membranes

This review covers recent advances in non-erythroid spectrin re-distributions during development, structural motifs recently discovered in ankyrin, band 4.2, band 4.1, ezrin, talin, and myosin I, and our present understanding of actin-membrane interactions at focal adhesions and in liver, platelet, and Dictyostelium discoideum plasma membranes.