Molecular biology of neuronal intermediate filaments.

In the past few years, several neuronal intermediate filament proteins have been characterized. While ongoing investigations have continued to shed light on their developmental expression, the importance of different domains of the proteins for assembly, the elements in their genes necessary for tissue-specific expression, and the role of phosphorylation of neurofilaments, the function(s) of these structures remain a matter of speculation.

Rapid movement of axonal neurofilaments interrupted by prolonged pauses.

Axonal cytoskeletal and cytosolic proteins are synthesized in the neuronal cell body and transported along axons by slow axonal transport, but attempts to observe this movement directly in living cells have yielded conflicting results. Here we report the direct observation of the axonal transport of neurofilament protein tagged with green fluorescent protein in cultured nerve cells. Live-cell imaging of naturally occurring gaps in the axonal neurofilament array reveals rapid, intermittent and highly asynchronous movement of fluorescent neurofilaments. The movement is bidirectional, but predominantly anterograde. Our data indicate that the slow rate of slow axonal transport may be the result of rapid movements interrupted by prolonged pauses.

alpha-Internexin, a 66-kD intermediate filament-binding protein from mammalian central nervous tissues.

In this paper we describe a 66-kD protein that co-purifies with intermediate filaments from rat optic nerve and spinal cord but can be separated further by ion-exchange chromatography. This protein is distinct from the 68-kD neurofilament subunit protein as judged by isoelectric focusing, immunoblotting, peptide mapping, and tests of polymerization competence. This protein is avidly recognized by the monoclonal anti-intermediate filament antigen antibody, previously demonstrated to recognize a common antigenic determinant in all five known classes of intermediate filaments. Also, when isolated this protein binds to various intermediate filament subunit proteins, which suggests an in vivo interaction with the intermediate filament cytoskeleton, and it appears to be axonally transported in the rat optic nerve. Because of this ability to bind to intermediate filaments in situ and in vitro we have named this protein alpha-internexin. A possible functional role for the protein in organizing filament assembly and distribution is discussed.

Astroglial cells provide a template for the positioning of developing cerebellar neurons in vitro.

Indirect immunocytochemical staining with antisera raised against purified glial filament protein and a neurofilament polypeptide was used to study cell interactions between astrocytes and neurons dissociated from embryonic and early postnatal cerebellum. Staining with antibodies raised against purified glial filament protein revealed that greater than 99% of all processes present in cerebellar cultures during the 1st wk in vitro were glial in origin. After 1 wk in culture, unstained processes that were presumably neuronal were observed. Stained astroglial processes formed a dense network that served as a template for cerebellar neurons, identified by indirect immunocytochemical localization of tetanus toxin. More than 90% of neurons from postnatal days 1 or 7 were positioned within one cell diameter of a glial process. In contrast, less than 40% of the neurons dissociated from early embryonic cerebellum were located adjacent to a glial process. Staining with antibodies raised against purified glial filament protein also revealed differences in astroglial morphology that were under developmental regulation. Astroglial cells from embryonic cerebellum were fewer in number and had thick, unbranched processes. Those from postnatal day 1 were more slender, branched, and stellate. Those from postnatal day 7 were highly branched and stellate. Some veil-like astroglial processes were also observed in cells from postnatal animals. These morphological changes were also observed when cells from embryonic day 13 were maintained for a week in vitro. No specific staining of embryonic or postnatal cerebellum cells was observed with antibodies raised against purified neurofilament polypeptides.

Reexpression of glial fibrillary acidic protein rescues the ability of astrocytoma cells to form processes in response to neurons.

Astroglial cells play an important role in orchestrating the migration and positioning of neurons during central nervous system development. Primary astroglia, as well as astrocytoma cells will extend long stable processes when co-cultured with granule neurons. In order to determine the function of the glial fibrillary acidic protein (GFAP), the major intermediate filament protein in astroglia and astrocytoma cells, we suppressed the expression of GFAP by stable transfection of an anti-sense GFAP construct in human astrocytoma U251MG cells. The resulting AS2-U251 cells can no longer extend stable processes in the presence of granule neurons. To show that this effect is due specifically to the absence of GFAP, we reintroduced a fully encoding rat brain GFAP cDNA into these AS2-U251 cells. The resulting rat GFAP appeared as a filamentous network and the reexpression of GFAP rescued the ability of these astrocytoma cells to form stable processes when co-cultured with neurons. From these results, it is clear that the glial specific intermediate filament protein, GFAP, is required for process extension of these astrocytoma cells in response to granule neurons.

Assembly of type IV neuronal intermediate filaments in nonneuronal cells in the absence of preexisting cytoplasmic intermediate filaments.

We report here on the in vivo assembly of alpha-internexin, a type IV neuronal intermediate filament protein, in transfected cultured cells, comparing its assembly properties with those of the neurofilament triplet proteins (NF-L, NF-M, and NF-H). Like the neurofilament triplet proteins, alpha-internexin coassembles with vimentin into filaments. To study the assembly characteristics of these proteins in the absence of a preexisting filament network, transient transfection experiments were performed with a non-neuronal cell line lacking cytoplasmic intermediate filaments. The results showed that only alpha-internexin was able to self-assemble into extensive filamentous networks. In contrast, the neurofilament triplet proteins were incapable of homopolymeric assembly into filamentous arrays in vivo. NF-L coassembled with either NF-M or NF-H into filamentous structures in the transfected cells, but NF-M could not form filaments with NF-H. alpha-internexin could coassemble with each of the neurofilament triplet proteins in the transfected cells to form filaments. When all but 2 and 10 amino acid residues were removed from the tail domains of NF-L and NF-M, respectively, the resulting NF-L and NF-M deletion mutants retained the ability to coassemble with alpha-internexin into filamentous networks. These mutants were also capable of forming filaments with other wild-type neurofilament triplet protein subunits. These results suggest that the tail domains of NF-L and NF-M are dispensable for normal coassembly of each of these proteins with other type IV intermediate filament proteins to form filaments.

The intermediate filament protein peripherin is the specific interaction partner of mouse BPAG1-n (dystonin) in neurons.

The dystonia musculorum (dt) mouse suffers from severe degeneration of primary sensory neurons. The mutated gene product is named dystonin and is identical to the neuronal isoform of bullous pemphigoid antigen 1 (BPAG1-n). BPAG1-n contains an actin-binding domain at its NH2 terminus and a putative intermediate filament-binding domain at its COOH terminus. Because the degenerating sensory neurons of dt mice display abnormal accumulations of intermediate filaments in the axons, BPAG1-n has been postulated to organize the neuronal cytoskeleton by interacting with both the neurofilament triplet proteins (NFTPs) and microfilaments. In this paper we show by a variety of methods that the COOH-terminal tail domain of mouse BPAG1 interacts specifically with peripherin, but in contrast to a previous study (Yang, Y., J. Dowling, Q.C. Yu, P. Kouklis, D.W. Cleveland, and E. Fuchs. 1996. Cell. 86:655-665), mouse BPAG1 fails to associate with full-length NFTPs. The tail domains interfered with the association of the NFTPs with BPAG1. In dt mice, peripherin is present in axonal swellings of degenerating sensory neurons in the dorsal root ganglia and is downregulated even in other neural regions, which have no obvious signs of pathology. Since peripherin and BPAG1-n also display similar expression patterns in the nervous system, we suggest that peripherin is the specific interaction partner of BPAG1-n in vivo.

Two forms of cerebellar glial cells interact differently with neurons in vitro.

Specific interactions between neurons and glia dissociated from early postnatal mouse cerebellar tissue were studied in vitro by indirect immunocytochemical staining with antisera raised against purified glial filament protein, galactocerebroside, and the NILE glycoprotein. Two forms of cells were stained with antisera raised against purified glial filament protein. The first, characterized by a cell body 9 microns diam and processes 130-150 microns long, usually had two to three neurons associated with them and resembled Bergmann glia. The second had a slightly larger cell body with markedly shorter arms among which were nestled several dozen neuronal cells, and resembled astrocytes of the granular layer. Staining with monoclonal antisera raised against purified galactocerebroside revealed the presence of immature oligodendroglia in the cultures. These glial cells constituted approximately 2% of the total cell population in the cultures and, in contrast to astroglia, did not form specific contacts with neurons. Staining with two neuronal markers, antisera raised against purified NILE glycoprotein and tetanus toxin, revealed that most cells associated with presumed astroglia were small neurons (5-8 microns). After 1-2 d in culture, some stained neurons had very fine, short processes. Nearly all of the processes greater than 10-20 micron long were glial in origin. Electron microscopy also demonstrated the presence of two forms of astroglia in the cultures, each with a different organizing influence on cerebellar neurons. Most neurons associated with astroglia were granule neurons, although a few larger neurons sometimes associated with them. Time-lapse video microscopy revealed extensive cell migration (approximately 10 microns/h) along the arms of Bergmann-like astroglia. In contrast, cells did not migrate along the arms of astrocyte-like astroglia, but remained stationary at or near branch points. Growth cone activity, pulsating movements of cell perikarya, and ruffling of the membranes of glial and neuronal processes were also seen.

Preferential phosphorylation of the 150,000 molecular weight component of neurofilaments by a cyclic AMP-dependent, microtubule-associated protein kinase.

Highly purified preparations of bovine brain and rabbit nerve root neurofilaments were found to be lacking in protein kinase activity when either histone FIIA or the neurofilaments themselves were used as acceptors. There was no augmentation of activity in the presence of cyclic AMP. Addition of microtubule proteins prepared by cycles of assembly and disassembly resulted in phosphorylation of histone, phosphorylation of tubulin and the microtubule-associated proteins, and phosphorylation of neurofilament subunits. The phosphorylation of neurofilaments was predominantly in the 150,000-dalton species and was completely cyclic AMP dependent.

Regulation of a high molecular weight microtubule-associated protein in PC12 cells by nerve growth factor.

PC12 rat pheochromocytoma cells respond to nerve growth factor (NGF) protein by shifting from a chromaffin-cell-like phenotype to a neurite-bearing sympathetic-neuron-like phenotype. Comparison of the phosphoprotein patterns of the cells by SDS PAGE after various times of NGF treatment revealed a high molecular weight (Mr greater than or approximately 300,000) band whose relative intensity progressively increased beyond 2 d of NGF exposure. This effect was blocked by inhibitors of RNA synthesis and did not require neurite outgrowth or substrate attachment. The enhancement by NGF occurred in serum-free medium and was not produced by exposure to epidermal growth factor, insulin, dibutyryl cAMP, or dexamethasone. Several different types of experiments indicated that this phosphoprotein corresponds to a high molecular weight (HMW) microtubule-associated protein (MAP). These included cross-reactivity with antiserum against brain HMW MAPs, co-cycling with microtubules and co-assembly with tubulin in the presence of taxol. The affected species also co-migrated in SDS PAGE gels with brain MAP1 and, unlike MAP2, precipitated upon boiling. Studies with [35S]-methionine-labeled PC12 cells indicated that at least a significant proportion of this effect of NGF was due to increased levels of protein rather than to mere enhancement of phosphorylation. On the basis of the apparent effects of MAPs on the formation and stabilization of microtubules and of the importance of microtubules in production and maintenance of neurites, it is proposed that induction of a HMW MAP may be one of the steps in the mechanism whereby NGF promotes neurite outgrowth. Furthermore, these findings may lead to an understanding of the role of MAP1 in the nervous system.

Suppression by antisense mRNA demonstrates a requirement for the glial fibrillary acidic protein in the formation of stable astrocytic processes in response to neurons.

The glial fibrillary acidic protein (GFAP) is a glial-specific intermediate filament protein, which is expressed in astrocytes in the central nervous system, as well as in astrocytoma cell lines. To investigate the function of GFAP, we have studied the human astrocytoma cell line, U251, which constitutively expresses GFAP and vimentin in the same 10-nm filaments. These cells respond to neurons in vitro in the same way as primary astrocytes: they withdraw from the cell cycle, support neuronal cell survival and neurite outgrowth, and they extend complex, GFAP-positive processes. To determine the role of GFAP in these responses, we have specifically suppressed its expression by stably transfecting the U251 cells with an antisense GFAP construct. Two stable antisense cell lines from separate transfections were isolated and were shown to be GFAP negative by Northern and Western blot analyses, and by immunofluorescence studies. The antisense cell lines were inhibited in their ability to extend significant glial processes in response to neurons. In culture with primary neurons, the average increase in process length of the U251 cells was nearly 400%, as compared to only 14% for the antisense transfectants. The other neuron induced responses of astrocytes, i.e., proliferative arrest and neuronal support, were not affected in these cell lines. These data support the conclusion that the glial-specific intermediate filament protein, GFAP, is required for the formation of stable astrocytic processes in response to neurons.

Interactions between neurofilaments and microtubule-associated proteins: a possible mechanism for intraorganellar bridging.

Mammalian neurofilaments prepared from brain and spinal cord by either of two methods partially inhibit the in vitro assembly of microtubules. This inhibition is shown to be due to the association of a complex of high molecular weight microtubule-associated proteins (MAP1 and MAP2) and tubulin with the neurofilament. Further analysis of the association reveals a saturable binding of purified brain MAPs to purified neurofilaments with a Kd of 10(-7) M. Purified astroglial filaments neither inhibit microtubule assembly nor show significant binding of MAPs. It is proposed that the MAPs might function as one element in a network of intraorganellar links in the cytoplasm.

Astrotactin: a novel neuronal cell surface antigen that mediates neuron-astroglial interactions in cerebellar microcultures.

A microculture system for mouse cerebellar cells has been used to identify an immune activity, raised in rabbits against postnatal cerebellar cells, that blocks neuron-glial interactions in vitro. In the presence of blocking antibodies, stable neuron-glial contacts did not form and neuronal induction of glial process outgrowth did not occur. Subsequently, neurons were randomly arranged in the cultures rather than organized along the arms of astroglia. We have named the immune activity that blocks neuron-astroglial interactions anti-astrotactin. Partial purification of the anti-astrotactin blocking antibodies was obtained by cellular absorption with PC12 cells, a clonal cell line which expresses both the N-CAM and NILE (Ng-CAM, L1) glycoproteins. Subsequent absorption with purified cerebellar granule cells, but not with astroglial cells, removed the blocking activity, suggesting that the antigen(s) bound by blocking antibodies are neuronal. Immunoprecipitation of [35S]methionine- or [3H]fucose-radiolabeled Triton extracts of early postnatal cerebellar cells showed that the unabsorbed antiserum recognized a large number of proteins. Among these were bands with apparent molecular masses of N-CAM (180 and 140 kD) and NILE (230 kD). After absorption of the immune serum with PC12 cells, the number of bands recognized by the antiserum was reduced to a prominent band at 100 kD and a diffuse smear of material between 80 and 90 kD. The prominent band at 100 kD was removed by subsequent absorption of the immune serum with granule cells, a step which removed the blocking activity in the cerebellar microculture assay. Further evidence suggests that the astrotactin activity is missing or defective on granule cells from the neurological mutant mouse weaver, an animal that suffers a failure of glial-guided neuronal migration. When anti-astrotactin Fab fragments were pre-absorbed with weaver cerebellar neurons and then tested in the functional assay of neuron-glial interactions, the immune blocking activity was not removed. In contrast, wild-type cerebellar neurons removed the anti-astrotactin blocking activity under the same conditions. Subsequently, when [3H]fucose-radiolabeled Triton extracts of weaver and normal cells were immunoprecipitated with whole or PC12-absorbed anti-astrotactin antiserum, the intensity of the band at 100 kD was reduced by 95% in weaver cells.

beta-Internexin, a ubiquitous intermediate filament-associated protein.

In this article we show a Triton-insoluble, intermediate filament-associated protein of approximately 70 kD to be expressed ubiquitously in diverse mammalian cell types. This protein, assigned the name beta-internexin, exhibits extreme homology in each of the various cell lines as demonstrated by identical limited peptide maps, similar mobilities on two-dimensional gels, and detection in Triton-soluble and -insoluble extracts. beta-Internexin also shares some degree of homology with alpha-internexin, an intermediate filament-associated protein isolated and purified from rat spinal cord, which accounts for the immunologic cross-reactivity displayed by these polypeptides. Light microscopic immunolocalization of beta-internexin with a monoclonal antibody (mAb-IN30) reveals it to be closely associated with the vimentin network in fibroblasts. The antigen is also observed to collapse with the vimentin reticulum during the formation of a juxtanuclear cap induced by colchicine treatment. Ultrastructural localization, using colloidal gold, substantiates the affinity of beta-internexin for cytoplasmic filaments and, in addition, demonstrates its apparent exclusion from the intranuclear filament network. We examine also the resemblance of beta-internexin to a microtubule-associated polypeptide and the constitutively synthesized mammalian heat shock protein (HSP 68/70).

Intermediate filaments in nervous tissues.

Intermediate filaments have been isolated from rabbit intradural spinal nerve roots by the axonal flotation method. This method was modified to avoid exposure of axons to low ionic strength medium. The purified filaments are morphologically 75-80 percent pure. The gel electrophoretogram shows four major bands migrating at 200,000, 145,000, 68,000, and 60,000 daltons, respectively. A similar preparation from rabbit brain shows four major polypeptides with mol wt of 200,000 145,000, 68,000, and 51,000 daltons. These results indicate that the neurofilament is composed of a triplet of polypepetides with mol wt of 200,000, 145,000, and 68,000 daltons. The 51,000-dalton band that appears in brain filament preparations as the major polypeptide seems to be of glial origin. The significance of the 60,000- dalton band in the nerve root filament preparation is unclear at this time. Antibodies raised against two of the triplet proteins isolated from calf brain localize by immunofluorescence to neurons in central and peripheral nerve. On the other hand, an antibody to the 51,000-dalton polypeptide gives only glial staining in the brain, and very weak peripheral nerve staining. Prolonged exposure of axons to low ionic strength medium solubilizes almost all of the triplet polypeptides, leaving behind only the 51,000- dalton component. This would indicate that the neurofilament is soluble at low ionic strength, whereas the glial filament is not. These results indicate that neurofilaments and glial filaments are composed of different polypeptides and have different solubility characteristics.

The BPAG1 locus: Alternative splicing produces multiple isoforms with distinct cytoskeletal linker domains, including predominant isoforms in neurons and muscles.

Bullous pemphigoid antigen 1 (BPAG1) is a member of the plakin family with cytoskeletal linker properties. Mutations in BPAG1 cause sensory neuron degeneration and skin fragility in mice. We have analyzed the BPAG1 locus in detail and found that it encodes different interaction domains that are combined in tissue-specific manners. These domains include an actin-binding domain (ABD), a plakin domain, a coiled coil (CC) rod domain, two different potential intermediate filament-binding domains (IFBDs), a spectrin repeat (SR)-containing rod domain, and a microtubule-binding domain (MTBD). There are at least three major forms of BPAG1: BPAG1-e (302 kD), BPAG1-a (615 kD), and BPAG1-b (834 kD). BPAG1-e has been described previously and consists of the plakin domain, the CC rod domain, and the first IFBD. It is the primary epidermal BPAG1 isoform, and its absence that is the likely cause of skin fragility in mutant mice. BPAG1-a is the major isoform in the nervous system and a homologue of the microtubule actin cross-linking factor, MACF. BPAG1-a is composed of the ABD, the plakin domain, the SR-containing rod domain, and the MTBD. The absence of BPAG1-a is the likely cause of sensory neurodegeneration in mutant mice. BPAG1-b is highly expressed in muscles, and has extra exons encoding a second IFBD between the plakin and SR-containing rod domains of BPAG1-a.

Microtubule actin cross-linking factor (MACF): a hybrid of dystonin and dystrophin that can interact with the actin and microtubule cytoskeletons.

We cloned and characterized a full-length cDNA of mouse actin cross-linking family 7 (mACF7) by sequential rapid amplification of cDNA ends-PCR. The completed mACF7 cDNA is 17 kb and codes for a 608-kD protein. The closest relative of mACF7 is the Drosophila protein Kakapo, which shares similar architecture with mACF7. mACF7 contains a putative actin-binding domain and a plakin-like domain that are highly homologous to dystonin (BPAG1-n) at its NH(2) terminus. However, unlike dystonin, mACF7 does not contain a coiled-coil rod domain; instead, the rod domain of mACF7 is made up of 23 dystrophin-like spectrin repeats. At its COOH terminus, mACF7 contains two putative EF-hand calcium-binding motifs and a segment homologous to the growth arrest-specific protein, Gas2. In this paper, we demonstrate that the NH(2)-terminal actin-binding domain of mACF7 is functional both in vivo and in vitro. More importantly, we found that the COOH-terminal domain of mACF7 interacts with and stabilizes microtubules. In transfected cells full-length mACF7 can associate not only with actin but also with microtubules. Hence, we suggest a modified name: MACF (microtubule actin cross-linking factor). The properties of MACF are consistent with the observation that mutations in kakapo cause disorganization of microtubules in epidermal muscle attachment cells and some sensory neurons.

An intermediate filament-associated protein, p50, recognized by monoclonal antibodies.

Monoclonal antibodies (mAB) were raised to be used as probes to identify cytoplasmic components associated with intermediate filaments (IF). Four hybridomas (B27, B76, B78, and B100) secreting mAB were generated by fusing mouse myeloma cells with the spleen cells of mice immunized intraperitoneally with Triton-high salt insoluble materials from BHK-21 cells. This insoluble material consists mostly of IF, a small number of microfilaments, and some polyribosomes. Biochemical studies show that the Triton-insoluble materials contain many proteins, including vimentin (decamin) and desmin. Immunofluorescence microscopy of BHK-21 cells stained with the four mAB showed that these mAB decorate the IF in a dotted pattern. Double staining with polyclonal antibody to vimentin confirmed the reactivity of the mAB with the IF. These mAB also stained the vimentin-containing filament system in a variety of other cells including epithelial cells (PTK1 and HeLa) and cells of astroglial origin. Histological studies showed that mAB-B100 stained many types of tissue including epidermis, smooth muscle, and subdermis pericytes, but not the white matter nor the gray matter of the cerebellum and spinal cord. Immunoelectron microscopy with colloidal gold has shown that the mAB-B100 decorated the IF in clusters or aggregates around proteinaceous materials associated with the filaments. Results of immunoprecipitation indicate that mAB-B100 reacted with a protein of 50,000 daltons. These findings suggest that the mAB-B100 we have developed recognizes one of the many components of what appears to be an integrated cytoskeletal structure connected with intermediate filaments.

Expression of rat neurofilament proteins NF-L and NF-M in transfected non-neuronal cells.

Two cDNA clones fully encoding the rat neurofilament proteins NF-L and NF-M were subcloned into eukaryotic expression vectors behind the strong constitutive viral promoters from SV40 and Rous sarcoma viruses. Transient transfection of L tk- and Cos cell lines with these expression constructs resulted in cells expressing the neurofilament proteins in an intermediate filament-type pattern. Additionally, a putative juxtanuclear organizing center or region was observed in the transfected cells, most noticeable shortly after the transfection procedure. Stable transfections were performed on mouse L tk- and Swiss 3T6 cells using NF-L and NF-M constructs bearing an SV40 early promoter driven neomycin selectable marker. Although G418-resistant clones were recovered with both the NF-L and the NF-M constructs, only clones expressing immunofluorescently stainable amounts of NF-M were detected and established. Immunoelectron microscopic analysis revealed NF-M and vimentin proteins to be colocalized on the same intermediate filaments.