Actin and associated proteins at the cytoskeleton-plasma membrane interface stabilize the membrane bilayer, control cell shape, and delimit specialized membrane domains. To identify membrane proteins that bind directly to F-actin, we have developed a blot overlay assay with 125I-labeled F-actin. In the soil amoebae, Dictyostelium discoideum, the major proteins reactive in this assay are p30a, a 34-kD peripheral membrane protein that is concentrated in filopodia and at sites of cell-cell adhesion, and ponticulin, a 17-kD transmembrane glycoprotein required for efficient chemotaxis and for control of pseudopod dynamics. Proteins with apparent molecular masses of approximately 34- and approximately 17-kD also are observed on F-actin blot overlays of many mammalian cell lines. However, in mammalian cells, the most prominent F-actin binding proteins in this assay exhibit apparent molecular masses of 78-, 80-, 81-, approximately 120-, and 205-kD. Bovine neutrophils contain the 78-, 81-, and 205-kD proteins, all of which co-isolate with a plasma membrane-enriched fraction. We have previously identified the 78-, 80-, and 81-kD proteins as moesin, radixin, and ezrin, respectively. These proteins, which are members of the protein 4.1 superfamily, colocalize with actin in cell surface extensions and have been implicated in the protrusion of microvilli, filopodia, and membrane ruffles. The 205-kD protein (p205) appears to be absent from current databases, and its characteristics are still under investigation. We here report that the 120-kD protein is drebrin, a submembranous actin-binding protein originally identified as a developmentally regulated brain protein. Thus, it appears that F-actin blot overlays provide an efficient assay for simultaneous monitoring of a subset of F-actin binding proteins, including p30a, ponticulin, moesin, radixin, ezrin, p205, and drebrin.
A wide range of molecules promote nerve growth, and these include cell adhesion molecules (CAMs), NCAM, N-cadherin, and the L1 glycoprotein are CAMs that are normally found on both the advancing growth cone and also on cellular substrates, and in general operate via a homophilic binding mechanism. In recent years it has become clear that nerve growth stimulated by these CAMs does not rely on the adhesion function of these molecules, but instead requires that the CAMs activate second messenger cascades in neurons. A large body of evidence supports the hypothesis that homophilic binding of the CAM in the substrate to the CAM in the neuron leads to activation of the neuronal FGF receptor, possibly via a direct interaction in cis between the CAM and the FGF receptor. The consequential activation of PLC gamma is both necessary and sufficient to account for the neurite outgrowth response stimulated by the above three CAMs. Based on the above model, we reasoned that soluble CAMs might also be able to stimulate neurite outgrowth and that such agents might be developed as potential therapeutic agents for stimulating nerve regeneration. To this end we have made soluble chimeric molecules consisting of the extracellular domain of NCAM or L1 fused to the Fc region of human IgG 1. We have found that these molecules can stimulate neurite outgrowth from rat and mouse cerebellar granule cells cultured on a variety of tissue culture substrates and that they do so by activating the FGF receptor signal transduction cascade in the neurons. Consistent with this model, we find that neurons that have their FGF receptor function ablated as a consequence of the expression of dominant negative FGF receptors, no longer respond to the soluble CAM. Downstream targets of CAM function have also been studied. Addition of soluble CAMs to isolated growth cone preparations from mouse or rat brain leads to enhanced phosphorylation of the GAP-43 protein providing a link between the cell surface and the cytoskeleton.
The establishment of compositionally distinct fluid compartments by various types of cell sheets is crucial for the development and function of most organs in multicellular organisms. Since these cell sheets consist of two-dimensionally arranged cells, some mechanism is required to seal cells to create a primary barrier to the diffusion of solutes through the paracellular pathway. This mechanism is thus essential for morphogenesis of the multicellular system, and the tight junction, an element of epithelial and endothelial junctional complexes, is now believed to be directly involved in this mechanism (Schneeberger and Lynch, 1992; Gumbiner, 1987, 1993). The brain is an important compartment that is protected by a barrier of endothelial cell sheets bearing well-developed tight junctions. This is referred as the blood-brain barrier.
Dystrophin serves as a link between the subsarcolemmal cytoskeleton and the extracellular matrix. The NH2 terminus attaches to the cytoskeleton, while the COOH terminus attaches to the dystrophin associated protein (DAP) complex, which can be separated into the dystroglycan, sarcoglycan, and syntrophin subcomplexes. While the function of each DAP is not known, the dystroglycan complex binds laminin in the extracellular matrix, and binds the dystrophin COOH terminus in vitro. The syntrophins also bind the dystrophin COOH terminus in vitro, but no evidence has been reported for an interaction between dystrophin and the sarcoglycans. Human mutations have been found in dystrophin, the sarcoglycans and laminin, all of which lead to various types of muscular dystrophy. We have been studying the dystrophin domains necessary for formation of a functional complex by generating transgenic mdx (dystrophin minus) mice expressing internally truncated dystrophins. These mice provide in vivo models to study the localization of truncated dystrophin isoforms, the association of the truncated proteins with the DAP complex, and the functional capacity of the assembled DAP complexes. Expression of a dystrophin deleted for most of the NH2-terminal domain in mdx mice leads to only a mild dystrophy, indicating that dystrophin can attach to the cytoskeleton by multiple mechanisms. Truncation of the central rod domain leads to normal DAP complex formation and almost fully prevents development of dystrophy. Deletion analysis of the COOH-terminal regions indicates that a broad cysteine-rich domain is indispensable for dystrophin function. This region coincides with the in vitro identified beta-dystroglycan binding domain. Mice lacking this latter domain express very low levels of the sarcoglycans, indicating that the sarcoglycan complex binds dystrophin via dystroglycan. All deletion constructs tested lead to normal expression of the syntrophins, indicating that syntrophin associates with the DAP complex via multiple binding partners.
The data presented here are clearly just the beginning of any comprehensive understanding of the set of regulatory and cytoskeletal proteins that interact with membrane receptors in the postsynaptic density. They do, however, indicate that both glutamate channels at central excitatory synapses are involved in complex protein-protein interactions. For example, while NR2A is important for Ca-dependent inactivation of NMDA receptors, studies in several systems suggest that the other major NR2 subunit in hippocampal neurons, NR2B, predominates at critical times during synapse formation. In addition, the COOH terminus of NR2B binds to several novel cytoskeletal proteins. These results provide circumstantial evidence that NR2B may play specific roles in function and localization of receptors at excitatory synapses. The possible role of NR2B in early synaptic function gains additional support from functional data suggesting that NMDA receptors have specific roles during development (Komuro and Rakic, 1993; Rabacchi et al., 1992; Yen et al., 1993). The essential role of NR1 and NR2B in development is graphically demonstrated by the neonatal death of transgenic mice lacking either of these two subunits (Forrest et al., 1994; Kutsuwada et al., 1996) whereas NR2A and NR2C-deficient mice are less severely affected (Sakimura et al., 1995; Ebralidze et al., 1996).
Protein phosphorylation is a primary means of mediating signal transduction events that control cellular processes. Accordingly, the activities of protein kinases and phosphatases are highly regulated. One level of regulation is that the subcellular distribution of several kinases and phosphatases is restricted by association with targeting proteins or subunits. This mechanism promotes rapid and preferential modulation of specific targets within a defined microenvironment in response to diffusible second messengers. The type II cAMP-dependent protein kinase (PKA) is targeted by association of its regulatory subunit (RII) with A-kinase anchoring proteins (AKAPs). To date, 36 unique AKAPs have been identified. Each of these proteins contains a conserved amphipathic helix responsible for AKAP association with cellular structures. Disruption of PKA/AKAP interaction with peptides patterned after the amphipathic helix region blocks certain cAMP responses, including the modulation of glutamate receptor ion-channel activity in neurons and transcription of cAMP-responsive genes. Yeast two-hybrid screening methods have identified neuronal specific AKAP79-binding proteins including the beta isoform of the phosphatase 2B, calcineurin. Biochemical and immunological studies have confirmed the two-hybrid results and identified additional members of this multienzyme signaling complex, including certain protein kinase C isoforms. These findings are consistent with colocalization of CaN, PKC, and type II PKA by AKAP79 and suggest a novel model for reversible phosphorylation in which the opposing kinase and phosphatase actions are colocalized in a signal transduction complex by association with a common anchor protein.
We have presented ultrastructural and semiquantitative immunofluorescence evidence to support the idea that AChR are clustered in rat myotubes by virtue of their ability to associate with a spectrin-based membrane skeleton. Many of the interactions postulated to be involved in the formation of this skeleton, and in the anchoring of AChR to it, must still be examined at the biochemical level, but the overall similarity of this structure to that of the human erythrocyte is already clear. The ability of different members of the spectrin superfamily to associate in various combinations to form distinct plasmalemmal domains provides some exciting hints as to how the surface membrane can be organized efficiently to subserve multiple purposes. One of the challenges of future research will be to learn how innervation regulates the assembly of the membrane skeleton at the developing NMJ, and how this structure is altered as the junction matures. Another will be to learn if the principles of neuromuscular synaptogenesis are relevant to interactions between neurons in the brain, where cells must distinguish between multiple synaptic inputs and assemble synaptic structures at thousands of distinct sites on the neurolemma. Members of the spectrin superfamily have been identified in synaptic structures in the central nervous system (e.g., Carlin et al., 1983; LeVine and Sahyoun, 1986; Malchiodi-Albedi et al., 1993), so much of what we have learned at the neuromuscular junction may be applicable to central synapses.
Actin filament lengths are precisely regulated and very stable in the sarcomeres of striated muscle, in the erythrocyte membrane skeleton, and in cell protrusions such as microvilli in intestinal epithelial cells and stereocilia in hair cells of the inner ear. In contrast, in motile cells, actin filament lengths are dynamically regulated when cells extend lamellipodia. Control of actin filament lengths and dynamics in cells is expected to be achieved in part by capping proteins that prevent filament growth or shrinkage by blocking subunit exchange at both the fast growing (barbed) and slow growing (pointed) filament ends. Much is known about how barbed end capping proteins control actin filament assembly and length in many cells, but little is known about the significance of regulating actin filament assembly at the pointed end. Tropomodulin is the only known capping protein for the pointed ends of actin filaments and is a approximately 40-kD protein that is expressed in erythrocytes, striated muscle, lens fiber cells, some regions of the adult brain, as well as sensory neurons and epithelial cells of the inner ear. A related isoform (59% identical at the protein level) is expressed principally in neurons of both embryonic and adult brain. In striated muscle and in erythrocytes, tropomodulin is tightly associated with the actin filament pointed ends where it functions to maintain actin filament length in vivo (for recent reviews, see Fowler, 1996; Gregorio and Fowler, 1996). Unlike proteins that cap actin filament barbed ends, tropomodulin also binds tropomyosin and requires tropomyosin for tight capping of actin filament pointed ends. Mapping of functional domains on tropomodulin shows that the COOH-terminal end of tropomodulin is important for actin filament pointed end capping activity while the NH2-terminal portion of tropomodulin contains the tropomyosin binding domain. Searches of protein and EST databases for tropomodulin-like sequences reveal a number of proteins with homologies to both the tropomyosin binding and the actin filament capping portions of tropomodulin. In particular, we have identified a tropomodulin-like 64-kD protein that is principally expressed in smooth muscle cells. We anticipate that tropomodulin and this 64-kD protein are members of a larger family of tropomyosin and actin binding proteins that are responsible for capping actin filament pointed ends and regulating actin filament lengths in muscle and nonmuscle cells.
The purpose of the studies included in this chapter was to examine the role of the actin network in the propagation of insulin action leading to stimulation of glucose transport and activation of the mitogen-activated protein kinase cascade. The active insulin receptor phosphorylates tyrosine residues of intracellular proteins such as the insulin receptor substrate-1 (IRS-1) which acts as docking sites for molecules containing Src homology 2 (SH2) domains. One such molecule is phosphatidylinositol 3-kinase (PI 3-kinase) which becomes activated by binding to IRS-1. PI 3-kinase activity is required for the insulin-stimulation of glucose transport and glycogen synthesis. Grb2, a small adaptor molecule, can bind IRS-1 and, through the guanine nucleotide exchange factor Sos, leads to the activation of the small GTP binding protein Ras. Through a cascade of protein kinases, activation of Ras results in activation of the Erk 1 and 2 mitogen-activated protein kinases (MAPKs) which appear to control important nuclear and metabolic events. To investigate the role of the actin network in the propagation of insulin action leading to stimulation of glucose transport and the activation of the Erk MAPKs, we used the fungal metabolite cytochalasin D which disassembles the actin network. Actin disassembly abolished almost completely the ability of insulin to increase the rate of glucose transport into L6 muscle cells (myotubes) through prevention of the insulin-induced recruitment of glucose transporters to the plasma membrane which is the event that mediates the increase in the rate of transport. Actin disassembly did not affect either the insulin-mediated phosphorylation of IRS-1, the association of PI 3-kinase with this molecule, or the activation of IRS-1-associated PI 3-kinase. These results were also verified in another insulin responsive cell line, the 3T3-L1 adipocytes. In these cells, actin disassembly inhibited the insulin-induced recruitment of PI 3-kinase to intracellular membranes containing glucose transporters. Moreover, actin disassembly abolished the insulin-mediated phosphorylation of the Erk MAPKs. We conclude that the cellular actin network of insulin responsive cells is not required for the activation of PI 3-kinase but prevents its cellular redistribution. In contrast, intact actin filaments are essential for the propagation of insulin signals leading to the the activation of the MAPKs.
Studies on the attachment and spreading of cells in culture have provided valuable insights into the mechanisms by which cells transmit information from the outside to the inside of the cell. This brief review considers recent information on the role of focal adhesion-associated protein tyrosine kinases in integrin-regulated cell signalling.
