The neurofibromatosis 2 protein product merlin selectively binds F-actin but not G-actin, and stabilizes the filaments through a lateral association.

The neurofibromatosis 2 protein product merlin, named for its relatedness to the ezrin, radixin and moesin (ERM) family of proteins, is a tumour suppressor whose absence results in the occurrence of multiple tumours of the nervous system, particularly schwannomas and meningiomas. Merlin's similarity to ERMs suggests that it might share functions, acting as a link between cytoskeletal components and the cell membrane. The N-terminus of merlin has strong sequence identity to the N-terminal actin-binding region of ezrin; here we describe in detail the merlin-actin interaction. Employing standard actin co-sedimentation assays, we have determined that merlin isoform 2 binds F-actin with an apparent binding constant of 3.6 microM and a stoichiometry of 1 mol of merlin per 11.5 mol of actin in filaments at saturation. Further, solid-phase binding assays reveal that merlin isoforms 1 and 2 bind actin filaments differentially, suggesting that the intramolecular interactions in isoform 1 might hinder its ability to bind actin. However, merlin does not bind G-actin. Studies of actin filament dynamics show that merlin slows filament disassembly with no influence on the assembly rate, indicating that merlin binds along actin filament lengths. This conclusion is supported by electron microscopy, which demonstrates that merlin binds periodically along cytoskeletal actin filaments. Comparison of these findings with those reported for ERM proteins reveal a distinct role for merlin in actin filament dynamics.

Mutant torsinA, responsible for early-onset torsion dystonia, forms membrane inclusions in cultured neural cells.

Early-onset torsion dystonia is a hereditary movement disorder thought to be caused by decreased release of dopamine into the basal ganglia, without apparent neuronal degeneration. Recent cloning of the gene responsible for this disease, TOR1A (DYT1), identified the encoded protein, torsinA, as a member of the AAA+ superfamily of chaperone proteins and revealed highest levels of expression in dopaminergic neurons in human brain. Most cases of this disease are caused by a deletion of one glutamic acid residue in the C-terminal region of the protein. Antibodies generated against torsinA revealed expression of a predominant immunoreactive protein species similar to the predicted size of 37.8 kDa in neural, glial and fibroblastic lines by western blot analysis. This protein is N-glycosylated with high mannose content and not, apparently, phosphoryl-ated. Overexpression of torsinA in mouse neural CAD cells followed by immunocytochemistry, revealed a dramatically different pattern of distribution for wild-type and mutant forms of the protein. The wild-type protein was found throughout the cytoplasm and neurites with a high degree of co-localization with the endoplasmic reticulum (ER) marker, protein disulfide isomerase. In contrast, the mutant protein accumulated in multiple, large inclusions in the cytoplasm around the nucleus. These inclusions were composed of membrane whorls, apparently derived from the ER. If disrupted processing of the mutant protein leads to its accumulation in multilayer membranous structures in vivo, these may interfere with membrane trafficking in neurons.

Interdomain interaction of merlin isoforms and its influence on intermolecular binding to NHE-RF.

Merlin, the neurofibromatosis 2 tumor suppressor protein, has two major isoforms with alternate C termini and is related to the ERM (ezrin, radixin, moesin) proteins. Regulation of the ERMs involves intramolecular and/or intermolecular head-to-tail associations between family members. We have determined whether merlin undergoes similar interactions, and our findings indicate that the C terminus of merlin isoform 1 is able to associate with its N-terminal domain in a head-to-tail fashion. However, the C terminus of isoform 2 lacks this property. Similarly, the N terminus of merlin can also associate with C terminus of moesin. We have also explored the effect of merlin self-association on binding to the regulatory cofactor of Na(+)-H(+) exchanger (NHE-RF), an interacting protein for merlin and the ERMs. Merlin isoform 2 captures more NHE-RF than merlin isoform 1 in affinity binding assays, suggesting that in full-length merlin isoform 1, the NHE-RF binding site is masked because of the self-interactions of merlin. Treatment with a phospholipid known to decrease self-association of ERMs enhances the binding of merlin isoform 1 to NHE-RF. Thus, although isoform 1 resembles the ERM proteins, which transition between inactive (closed) and active (open) states, isoform 2 is distinct, existing only in the active (open) state and presumably constitutively more available for interaction with other protein partners.

Inhibition of NF2-negative and NF2-positive primary human meningioma cell proliferation by overexpression of merlin due to vector-mediated gene transfer.

OBJECT: The absence of in vitro models of neurofibromatosis Type 2 (NF2)-defective meningiomas has limited investigative efforts to study the biological effects of this gene in the pathogenesis of these tumors. The goals of this report are to show that gene transfer vectors can efficiently express the wild-type NF2 transgene into primary meningioma cells and to determine effects on cellular proliferation. METHODS: In this study, the authors have compared the transducing capacities of a retrovirus, an adenovirus, and a herpes simplex virus amplicon vector for use in primary human meningioma cells harvested from human tumors excised from patients with and without NF2. Transduction efficiencies with the latter vector approached 100% and it was selected to transfer the wild-type NF2 transgene into these cells. Western blot analysis confirmed that vector-mediated gene transfer mediated the expression of the NF2-encoded polypeptide merlin. Overexpression of merlin significantly inhibited the proliferation of both NF2-negative and NF2-positive human meningioma cells when compared to the proliferation of cells transduced with a control vector. CONCLUSIONS: This study demonstrates the feasibility of using vector-mediated gene transfer to study wild-type NF2 gene function in short-term cultures of primary human meningioma cells.

Analysis of molecular domains of epitope-tagged merlin isoforms in Cos-7 cells and primary rat Schwann cells.

The Neurofibromatosis 2 gene product, merlin, has striking similarity to ezrin, radixin, and moesin (ERM), members of the protein 4.1 family which have been demonstrated to connect proteins in the plasma membrane to the cytoskeletal components. The recent localization of merlin to the motile regions in cultured cells such as membrane ruffles further supports the notion that merlin represents a new class of tumor suppressors. Here we describe the localization of full-length and truncated polypeptides of merlin expressed as Flag-tagged proteins in transfected cells. Similar to endogenous merlin, the epitope-tagged full-length merlin localizes to the membrane ruffles in transfected Cos-7 cells and rat Schwann cells. In addition, the over-expressed merlin localizes to other actin-rich cortical structures, such as microvilli and filopodia. The amino-terminal half of merlin is seen dispersed throughout the cells and in membrane ruffles. Compared to the amino-terminal half of merlin, its carboxy-terminal half localizes more distinctly to membrane ruffles. The full-length and the carboxy-terminal portion of merlin co-localize with F-actin at the membrane ruffles. However, distinct from the ERM proteins, the carboxy-terminal-truncated merlin and F-actin do not co-localize with each other at the stress fibers. Our results suggest that both the amino- and the carboxy-terminal domains of merlin contribute to its membrane ruffle localization.

NHE-RF, a regulatory cofactor for Na(+)-H+ exchange, is a common interactor for merlin and ERM (MERM) proteins.

We have identified the human homologue of a regulatory cofactor of Na(+)-H+ exchanger (NHE-RF) as a novel interactor for merlin, the neurofibromatosis 2 tumor suppressor protein. NHE-RF mediates protein kinase A regulation of Na(+)-H+ exchanger NHE3 to which it is thought to bind via one of its two PDZ domains. The carboxyl-terminal region of NHE-RF, downstream of the PDZ domains, interacts with the amino-terminal protein 4.1 domain-containing segment of merlin in yeast two-hybrid assays. This interaction also occurs in affinity binding assays with full-length NHE-RF expressed in COS-7 cells. NHE-RF binds to the related ERM proteins, moesin and radixin. We have localized human NHE-RF to actin-rich structures such as membrane ruffles, microvilli, and filopodia in HeLa and COS-7 cells, where it co-localizes with merlin and moesin. These findings suggest that hNHE-RF and its binding partners may participate in a larger complex (one component of which might be a Na(+)-H+ exchanger) that could be crucial for the actin filament assembly activated by the ERM proteins and for the tumor suppressor function of merlin.

Universal absence of merlin, but not other ERM family members, in schwannomas.

NF2 (neurofibromatosis 2, encoding the merlin protein) gene mutations and chromosome 22q loss have been demonstrated in the majority of sporadic and NF2-associated schwannomas, but many schwannomas fail to demonstrate genetic evidence of biallelic NF2 gene inactivation. In addition, the role of the merlin-related ERM family members (ezrin, radixin, and moesin) remains unclear in these tumors. We therefore studied expression of NF2-encoded merlin as well as ezrin, radixin, and moesin in 22 vestibular and peripheral schwannomas that had been evaluated for NF2 mutations and chromosome 22q loss. Western blotting and immunohistochemistry with antibodies directed against the amino and carboxy termini of merlin demonstrated loss of merlin expression in all studied schwannomas, including 12 tumors lacking genetic evidence of biallelic NF2 gene inactivation. Western blotting with antibodies directed against ezrin, radixin, and moesin, however, showed expression of these proteins in all schwannomas. In addition, immunohistochemistry with an antibody to moesin revealed widespread expression in tumor and endothelial cells. These data indicate that the specific loss of merlin is universal to schwannomas and is not linked to loss of ezrin, radixin, or moesin expression.

Expression of NF2-encoded merlin and related ERM family proteins in the human central nervous system.

Germline mutations of the neurofibromatosis 2 (NF2) gene are associated with an increased incidence of gliomas and glial harmartomas, suggesting a role for the NF2-encoded protein, merlin, in glial growth control. Using monoclonal and polyclonal anti-merlin antibodies for Western blotting and immunohistochemistry, we evaluated the cellular pattern of merlin expression in the normal human central nervous system (CNS), reactive gliosis; and NF2-associated glial hamartomas. In the normal CNS, merlin is widely expressed in coarse cytoplasmic granules in both glia and neurons, with less pronounced expression in other cells. Merlin is also expressed in reactive astrocytes and in the astrocytes of NF2-associated glial hamartomas. In reactive astrocytes, however, merlin is also present at the cell membrane and in cellular processes, suggesting redistribution of the protein in activated cells. Merlin is structurally related to ezrin, radixin and moesin, which are also expressed in the CNS, as demonstrated by Western blotting. The pattern of merlin expression, however, is distinct from that of ezrin, which has been previously described, and that of moesin, in which immunohistochemistry with an anti-moesin antibody showed expression in endothelial cells, glia and neurons in a membranous or diffuse cytoplasmic pattern. These findings imply that merlin has widespread and specific functions in the human central nervous system.

The merlin tumor suppressor localizes preferentially in membrane ruffles.

Merlin is a tumor suppressor whose inactivation underlies the familial schwannomas and meningiomas of neurofibromatosis 2 and their sporadic counterparts. It bears striking similarity to the ERM proteins, ezrin, radixin and moesin, members of the protein 4.1 superfamily that link proteins in the cytoskeleton and the plasma membrane. We have generated polyclonal and monoclonal antibodies that detect merlin as an approximately 66 kD protein in many different cell types. Using indirect immunofluorescence we have for the first time visualized endogenous merlin and localized it to the motile regions, such as leading or ruffling edges, in human fibroblast and meningioma cells. Merlin co-localizes with F-actin in these motile regions but is not associated with stress fibers. Merlin does not localize to the same structures as either ezrin or moesin in human meningioma cells, suggesting a function distinct from these ERMs. Thus, merlin is associated with motile regions of the cell and its participation in these structures may be intimately involved in control of proliferation in Schwann cells and meningeal cells.

Response of radixin to perturbations of growth cone morphology and motility in chick sympathetic neurons in vitro.

The ERM protein--ezrin, radixin, moesin--localize to a variety of cortical structures, where they may participate in connecting the cytoskeleton to components of the plasma membrane. Antibodies that recognize the ERM proteins specifically stain growth cones of various neurons [Goslin et al., 1989: J. Cell Biol. 109:1621-1631; Birgbauer et al., 1991: J. Neurosci. Res. 30:232-241]. To probe the function of ERM proteins in growth cones, we studied the consequences of perturbing growth cone morphology and motility of cultured chick sympathetic neurons. We demonstrate that radixin is present in these growth cones. Withdrawal of nerve growth factor (NGF) induces rapid collapse of the growth cones; concomitantly, radixin staining in these growth cones are greatly diminished. Upon readdition of NGF, rapid growth cone formation is accompanied by relocalization of radixin. Induction of growth cone collapse by either growth cone-growth cone contact or exposure to brain membrane extract results in a similar diminution of radixin staining. We induced a more subtle change in the organization of the growth cones by subjecting them to an electric field. These growth cones rapidly orient toward the cathode. We show that the radixin staining of the growth cones is also asymmetrically localized toward the leading edges in the new direction of growth. The results suggest that the localization of radixin may be essential for the normal expression of growth cone morphology and function.

Molecular dissection of radixin: distinct and interdependent functions of the amino- and carboxy-terminal domains.

The ERM proteins--ezrin, radixin, and moesin--occur in particular cortical cytoskeletal structures. Several lines of evidence suggest that they interact with both cytoskeletal elements and plasma membrane components. Here we described the properties of full-length and truncated radixin polypeptides expressed in transfected cells. In stable transfectants, exogenous full-length radixin behaves much like endogenous ERM proteins, localizing to the same cortical structures. However, the presence of full-length radixin or its carboxy-terminal domain in cortical structures correlates with greatly diminished staining of endogenous moesin in those structures, suggesting that radixin and moesin compete for a limiting factor required for normal associations in the cell. The results also reveal distinct roles for the amino- and carboxy-terminal domains. At low levels relative to endogenous radixin, the carboxy-terminal polypeptide is associated with most of the correct cortical targets except cleavage furrows. In contrast, the amino-terminal polypeptide is diffusely localized throughout the cell. Low level expression of full-length radixin or either of the truncated polypeptides has no detectable effect on cell physiology. However, high level expression of the carboxy-terminal domain dramatically disrupts normal cytoskeletal structures and functions. At these high levels, the amino-terminal polypeptide does localize to cortical structures, but does not affect the cells. We conclude that the behavior of radixin in cells depends upon activities contributed by separate domains of the protein, but also requires modulating interactions between those domains.

Analysis of a cortical cytoskeletal structure: a role for ezrin-radixin-moesin (ERM proteins) in the marginal band of chicken erythrocytes.

We are studying how the cytoskeleton determines cell shape, using a simple model system, the marginal band of chicken erythrocytes. We previously identified a minor component of the marginal band by a monoclonal antibody, called 13H9 (Birgbauer and Solomon (1989). J. Cell Biol. 109, 1609-1620; Goslin et al. (1989). J. Cell Biol. 109, 1621-1631). mAb 13H9 also binds to the leading edges of fibroblasts and to neuronal growth cones and recognizes the cytoskeletal protein ezrin. In recent years, two proteins with a high degree of homology to ezrin were identified: moesin and radixin, together comprising the ERM protein family. We now show that the contiguous epitope sufficient for mAb 13H9 binding is a sequence present in each of the ERM proteins, as well as the product of the gene associated with neurofibromatosis 2, merlin or schwannomin. We used biochemical and immunological techniques, as well as PCR to characterize the expression and localization of the ERM proteins in chicken erythrocytes. The results demonstrate that radixin is the major ERM protein associated with the cytoskeleton. Both ezrin and radixin localize to the position of the marginal band. Our results suggest that the ERM proteins play functionally conserved roles in quite diverse organelles.

The contributions of microtubules and F-actin to the in vitro migratory mechanisms of Hydra nematocytes as determined by drug interference experiments.

In order to investigate the contributions of microtubules and of F-actin to the in vitro migration mechanisms of Hydra nematocytes we have studied the effects of agents directed against cytoskeletal structures. Disassembly of microtubules by treatment with the drug nocodazole in moving nematocytes resulted in the loss of all locomotory activity within 20 min after the onset of treatment and in the detachment from the substratum after about 30 min. Depolymerization of microtubules by exposure to low temperatures had the same effect but was reversible in this case. Locomoting cells treated with cytochalasin D, which disrupts the actin filaments, stopped movement 2 min after drug administration and detached from the substratum after 15 min. The pattern of F-actin, alpha-tubulin, and tyrosinated tubulin in drug- or cold-treated cells was determined by immunocytochemical techniques and confocal laser scanning microscopy. These patterns and the reactions of the cells to the various drug treatments suggest that both actin filaments and microtubules play a crucial role in nematocyte locomotion. Analysis of the cytoskeletal pattern in drug-treated cells shows that the microtubules which are involved in locomotion are mostly tyrosinated. Furthermore it is suggested that microtubules and actin filaments interact with each other during the locomotion of nematocytes.