Heterogeneity of Rat Neurofilament Polypeptides Revealed by a Monoclonal Antibody

A monoclonal antibody obtained from mice immunized with a crude neurofilament preparation from newborn rat brain revealed the existence of heterogeneity of the 200,000- and 150,000-dalton neurofilament polypeptides. On immunoblot the monoclonal antibody iC8 reacted with both the 200,000- and 150,000-dalton components in the CNS, but only with the 150,000-dalton polypeptide in sciatic nerve preparations. In addition, the 150,000-dalton polypeptide appeared as a single band in the sciatic nerve, whereas in the CNS a doublet was labeled by iC8. In contrast a second monoclonal antibody (3H5) reacted with the 200,000-dalton peptide and a single 150,000-dalton component in both the central and peripheral nervous system preparations. The differences revealed by iC8 were probably not due to phosphorylation, as the pattern of antibody binding in immunoblots was not changed by pretreatment with alkaline phosphatase. The findings suggest that different isoforms of neurofilament polypeptides are present in the nervous system.     

Immunochemical Characterization of Antisera to Rat Neurofilament Subunits

Abstract: Antisera raised to the 68,000, 145,000 and 200,000 molecular weight subunits of rat neurofilaments were used for immunochemical staining of polypeptides separated by one- and two-dimensional gel electrophoresis. It was found that each antiserum reacts intensely with its corresponding neurofilament subunit and weakly with the other two subunits. All the antisera also react with a polypeptide of molecular weight 57,000 present in neurofilament-rich preparations from both rat spinal cord and peripheral nerve. This polypeptide is different from either tubulin or vimentin and may represent a neurofilament breakdown product, since it varied in amount from preparation to preparation. The three antisera also reacted with the polypeptide subunits of chicken and goldfish neurofilament despite the considerable difference in molecular weight between these subunits and those of mammalian neurofilament. Key Words: Neurofilaments–Antibodies–Immunochemical. Autilio-Gambetti L. et al. Immunochemical characterization of antisera to rat neurofilament subunits. J. Neurochem. 37, 1260-1265(1981).     

Purification of the glial fibrillary acidic protein by anion-exchange chromatography

A procedure for the isolation of assembly-competent glial fibrillary acidic (GFA) protein from 2 m urea extracts of bovine spinal cord by anion-exchange chromatography is reported. The tissue was previously extracted with low-ionic-strength buffer. The procedure allowed the separation of nondegraded GFA protein from GFA protein comprising degraded species. As previously reported for neurofilament preparations obtained from porcine spinal cord (N. Geisler and K. Weber, J. Mol. Biol., 151, 565–571 (1981)), the procedure also allowed the simultaneous separation of the three neurofilament polypeptides (200,000; 150,000; and 70,000 daltons) contained in the 2 m urea extract. Brain filament proteins sequentially eluted at increasing salt concentration (25–200 mm NaCl) according to their isoelectric point. Proteins with higher pI eluted first. Tubulin eluted between the 200,000- and 150,000-dalton neurofilament polypeptides.     

Isolation and characterization of glial filaments from human brain

Intermediate (8--9 nm) filaments of human central nervous system astrocytes were isolated from the gliosed white matter of cases of adrenoleukodystrophy (ALD). This hereditary lipidosis is characterized pathologically by demyelination, loss of axons, and replacement of the white matter of the caudal cerebrum by a glial scar. Glial filaments were composed largely of a single protein component with a mol wt of about 49,000 daltons. Smaller components (44,000--39,000 daltons) were detected in some samples, and appear to represent degradation products of the filament protein. Human neurofilaments were isolated from the normal frontal white matter of ALD cases by the standard myelin-free axon technique. Isolated glial and neurofilament proteins comigrated during acrylamide gel electrophoresis in SDS. Polypeptides resulting from cyanogen bromide cleavage of the two filament proteins were the same. Both proteins reacted with rabbit antisera raised against isolated bovine neurofilament protein and human glial fibrillary acidic protein.     

Identification of the major 68,000-dalton protein of microtubule preparations as a 10-nm filament protein and its effects on microtubule assembly in vitro.

The major 68,000-dalton protein present in cycled microtubule preparations from bovine brain can be isolated in a rapidly sedimenting fraction consisting of filaments 10 nm in diameter. This 68,000-dalton protein remains in the filament fraction after gel filtration, phosphocellulose chromatography, or salt extraction of microtubule protein. Microtubule protein devoid of 10-nm filaments contains ring structures under depolymerizing conditions, and it polymerizes into microtubules with a characteristically low critical concentration, although all of the 68,000-dalton protein has been removed from it. When cycled microtubule protein is subjected to chromatography on phosphocellulose, the tubulin fraction (PC-tubulin) assembles into microtubules only at concentrations greater than 2 mg/mL. The other fraction, eluted from phosphocellulose at high ionic strength, contains the major 68,000-dalton protein and can be further resolved into two components by centrifugation. The supernatant, which consists mainly of high molecular weight microtubule-associated proteins, stimulates low concentrations of PC-tubulin to assemble. The pellet contains all of the 68,000-dalton protein, consists of 10-nm filaments, and does not stimulate assembly of PC-tublin. Boiling of purified filaments, however, releases several proteins, including the 68,000-dalton protein, and these released proteins stimulate the assembly of PC-tubulin. The morphology and protein composition of the filaments isolated from microtubule preparations by these techniques are very similar to those of mammalian neurofilaments. These results suggest that the major 68,000-dalton protein in cycled microtubule preparations, which may correspond to tubulin assembly protein [Lockwood, A.H. (1978) Cell 13, 613--627], is a constituent of neurofilaments.     

Study of the 10-nm-filament fraction isolated during the standard microtubule preparation.

The cold non-depolymerizable fractions obtained during the standard procedure for the isolation of microtubules from ox brain stem-cerebral hemispheres and spinal cord have been studied. The cerebral-hemisphere preparation was composed of 10-nm filaments but also contained large amounts of membranes. The polypeptide content included tubulin, microtubule-associated proteins and minor proteins corresponding to the neurofilament triplet of proteins of mol.wt. 210 000, 160 000 and 70 000 respectively. The brain-stem preparation contained more 10-nm filaments than membranes. The polypeptide content consisted of the neurofilament triplet (35%), tubulin (30%) and minor proteins. In contrast, the spinal-cord preparation was mainly composed of 10-nm filaments, free of membranes and containing essentially the neurofilament protein triplet (64%). These filaments appeared very similar to the peripheral-nervous-system neurofilaments described by several authors. Since the best neurofilament from the central nervous system often contained less than 15% of the neurofilament protein triplet, our spinal-cord preparation is an improvement on the usual neurofilament preparation. This simple and rapid method gave large amounts of 10-nm filaments (100 mg per 100 g of spinal cord) characterized by the absence of membranous material, a low content of tubulin and the 50 000-mol.wt.-protein component, and a high content of neurofilament peptides. Thus, the presence of tubulin in 10-nm filament preparations seems to be related to the contaminant membranous material and not to be linked to the interaction in vitro of tubulin or microtubules with neurofilaments, as has been suggested previously.     

Synthesis of Cytoskeletal Proteins in Bulk-Isolated Neuronal Perikarya Takashi Nakayama

Abstract: Neuronal perikarya were isolated from young rat brain by sucrose density gradient centrifugation of the tissue, dissociated with a low concentration of trypsin. The isolated cells retained their endogenous proteins, and were capable of active protein synthesis. After incubation with L-[³⁵S]methionine, perikarya were homogenised and separated into soluble and particulate fractions by centrifugation at 70,000 g. Newly synthesised polypeptides in each fraction were resolved by SDS-gel and two-dimensional gel electrophoresis coupled with fluorography. Neuronal perikarya synthesised predominantly actin, and α¹-, α² and β-tubulin. In addition, polypeptides with molecular weights of 35,000, 68,000 and 85,000 were heavily labelled. On two-dimensional electrophoresis, microheterogeneities were seen in soluble actin as well as in soluble tubulins, indicating that heterogeneities reported for brain actin and tubulins are inherent in neuronal actin and tubulins, but not owing to the heterogeneity of cells in the brain tissue. Structural differences between soluble tubulins and those associated with the particulate fraction were indicated by two-dimensional gel electrophoresis and also by one-dimensional peptide maps. The 68,000 molecular weight polypeptide synthesised in neuronal perikarya in vitro yielded a peptide map virtually identical with that generated from the major component of the neurofilament triplet polypeptides that were synthesised in situ. The 160,000 and 200,000 components of the neurofilament triplet were also synthesised in perikarya in vitro , but to disproportionately weaker extents compared with the 68,000 component.     

Structure and biochemical composition of desmosomes and tonofilaments isolated from calf muzzle epidermis

Complexes of plasma membrane segments with desmosomes and attached tonofilaments were separated from the stratum spinosum cells of calf muzzle by means of moderately alkaline buffers of low ionic strength and mechanical homogenization. These structures were further fractionated by the use of various treatments including sonication, sucrose gradient centrifugation, and extraction with buffers containing high concentrations of salt, urea, citric acid, or detergents. Subfractions enriched in desmosome-tonofilament-complexes and tonofilament fragments were studied in detail. The desmosome structures such as the midline, the trilaminar membrane profile, and the desmosomal plaque appeared well preserved and were notably resistant to the various treatments employed. Fractions containing desmosome- tonofilament complexes were invariably dominated by the nonmembranous proteins of the tonofilaments which appeared as five major polypeptide bands (apparent molecular weights: 48,000; 51,000; 58,000; 60,000; 68,000) present in molar ratios of approx. 2:1:1:2:2. Four of these polypeptide bands showed electrophoretic mobilities similar to those of prekeratin polypeptides from bovine hoof. However, the largest polypeptide (68,000 mol wt) migrated significantly less in polyacrylamide gels than the largest component of the hoof prekeratin (approximately 63,000 mol wt). In addition, a series of minor bands, including carbohydrate-containing proteins, were identified and concluded to represent constituents of the desmosomal membrane. The analysis of protein-bound carbohydrates (total 270 microgram/mg phospholipid in desmosome-enriched subfractions) showed the presence of relatively high amounts of glucosamine, mannose, galactose, and sialic acids. These data as well as the lipid composition (e.g., high ratio of cholesterol to phospholipids, relatively high contents of sphingomyelin and gangliosides, and fatty acid pattern) indicate that the desmosomal membrane is complex in protein and lipid composition and has a typical plasma membrane character. The similarity of the desmosome-associated tonofilaments to prekeratin filaments and other forms of intermediate- sized filaments is discussed.     

Purification of a calcium-activated neutral proteinase from bovine brain

A calcium-activated neutral proteinase (CANP) resolved into three components has been partially purified from bovine brain. The method of isolation has resulted in 22,000, 7,100, and 8,000-fold purification for CANP I, II and III respectively. All three fractions require Ca2+ for activation. The characterization of the purified CANP I has shown that it is activated by 250 microM Ca2+ and the enzyme loses its activity when incubated in the presence of Ca2+ without substrate. Mg2+ is ineffective. The enzyme degrades neurofilament triplet proteins, tubulin and casein efficiently. The myelin basic protein is hydrolyzed after longer incubation. Bovine serum albumin and histones are unaffected. The enzyme is active at pH 5.5 to 9.0 with optimum between pH 7.5 and 8.5. It has a Km of 1.8 X 10(-7) M for the 69,000 dalton neurofilament protein. The enzyme is inhibited by sulphydryl blocking reagents and also by EGTA, leupeptin and E-64c. The SDS-PAGE analysis of the enzyme fractions has shown a major band at 66-68,000 daltons and two minor bands at 60,000 and 48-50,000 daltons for CANP I; a major band at 48-50,000 daltons and a minor band at 30-32,000 daltons for CANP II and a predominant doublet at 30-32,000 daltons with a minor band at 48-50,000 daltons for CANP III. The degradation of neurofilament proteins suggests that the CANP(s) may be involved in the turnover of these proteins.     

Intermediate filaments: analysis of filamentous aggregates induced by griseofulvin, an antitubulin agent

Mice fed griseofulvin, an antibiotic with antimicrotubular activity, formed hepatocellular aggregates of intermediate filaments, which resembled those associated with human alcoholic liver disease. These aggregates, termed Mallory bodies, were isolated from both human and mouse liver and the composition of these structures compared. Electrophoretic analysis indicated that the mouse filaments were composed of four major polypeptides (51,000, 47,000, 37,000, and 36,000 daltons). Human Mallory bodies possessed a similar number of components but of different molecular weights (56,000, 51,000, 50,000, and 38,000 daltons). Guinea pig antisera prepared against both whole human Mallory bodies and the major human polypeptide (56,000 daltons) crossreacted with mouse Mallory body material in both immunochemical and immunocytochemical systems. Our findings suggest that the two filament systems possess similar biochemical and immunological properties.     

Immunochemical comparison of prekeratin and alcoholic hyalin intermediate filaments

Alcoholic hyalin is an hepatocellular aggregate composed of filaments apparently related to the prekeratin intermediate filament subclass. The relationship between these two filament preparations was determined immunochemically using guinea pig antisera derived against alcoholic hyalin, prekeratin, and major prekeratin polypeptides. Immunocrossreactivities were determined using sensitive solid-phase enzyme-immunoassays. These assays indicated that antisera derived against a given filament preparation reacted 10–1000 times better with that preparation than with the other system. The nature of crossreactive meterial was determined using antisera derived against the larger prekeratin polypeptides (M_r 61,000 and 51,000). When tested against these two antisera, alcoholic hyalin appeared to react better with the serum derived against the larger prekeratin component. Moreover, anti-alcoholic hyalin antiserum bound four to five times better to the 61,000 dalton component than to the 51,000 dalton polypeptide in the enzyme-immunoassay. Our results indicate that antigenic determinants related to prekeratin can be detected in alcoholic hyalin, but that these determinants are present in relatively low concentrations in purified alcoholic hyalin. In addition, it appears that the relative concentrations of prekeratin components in alcoholic hyalin do not reflect those in purified prekeratin.     

An isoelectric variant of the 150,000-dalton neurofilament polypeptide. Evidence that phosphorylation state affects its association with the filament

A soluble isoelectric variant of the 150,000-dalton neurofilament protein was isolated from bovine brain by treating a partially purified filament preparation with a low-ionic-strength high-pH buffer. The protein (S150) had similar peptide maps to the neurofilament component of the same molecular weight (NF150) and was recognized by a polyclonal antibody made against the NF150 polypeptide. However, only half the anti-NF150 activity could be removed with the S150 protein. In addition, the S150 protein had a higher isoelectric point than the NF150 protein. Phosphate analysis indicated that the S150 protein was considerably lessened in phosphate content, which could account for the higher isoelectric point of the protein. It appears, therefore, that the S150 protein may be a precursor of NF150 or the result of phosphatase activity during the isolation procedure. Assembly studies showed that the S150 protein, unlike the NF150 protein, could not assemble with the 70-kDa neurofilament protein, indicating that the phosphate groups which were removed are important in the association of this protein to the neurofilament. When filaments containing all three triplet neurofilament polypeptides or those composed of the 70- and 150-kDa neurofilament proteins were subjected to acid phosphatase, a soluble fraction was obtained, which contained isoelectric variants with higher pI values than the NF150 polypeptide. Only unmodified NF150 protein was found in the insoluble fraction. These results support the argument that removal of phosphate groups results in the dissociation of this protein from the filament.     

Slowly migrating axonal polypeptides. Inequalities in their rate and amount of transport between two branches of bifurcating axons

Polypeptides in the dorsal root ganglion (L5) of the adult rat were radioactively labeled, and components slowly migrating in the sciatic nerve (peripheral axons) and dorsal root (central axons) were analyzed, using SDS-polyacrylamide slab gel electrophoresis and fluorography. In particular, the transport rates and amounts of six major polypeptides, i.e., the triplet (reference 15; with mol wts of 200,000, 160,000, and 68,000 daltons), alpha- and beta-tubulins and actin were compared between the two axon branches. In peripheral axons, fronts of the triplet, tubulins, and actin migrate at 2-3 mm/d, 9-13 mm/d and approximately 19 mm/d, respectively. The corresponding values in central axons are 1-2 mm/d, 3-4 mm/d, and approximately 4 mm/d, indicating an obvious asymmetry in the transport rate between the two branches of bifurcating axons. A greater amount of labeled triplet, tubulins, and actin each is found to migrate in peripheral than in central axons. Another striking aspect of asymmetry between the two branches relates to the tubulins/triplet ratio which is significantly higher in the peripheral branch. Considerable proportions of radioactivities associated with tubulins and actin in the ganglion are nonmigratory, which are thought to derive mostly from periaxonal satellite cells. In contrast, most if not all of the labeled triplet is migratory, suggesting a virtual absence of triplet polypeptides in satellite cells. The possible significance of peripheral-central inequalities in slow axoplasmic transport is discussed from the viewpoints of axon volume and axonal outgrowth.     

Biosynthesis of Rauscher leukemia viral proteins

Antisera to disrupted Rauscher leukemia virus (RLV) or to the purified Rauscher viral 30,000 dalton polypeptide were used to specifically precipitate newly synthesized intracellular viral polypeptides from extracts of infected NIH Swiss mouse cells (JLS-V16). Analysis by SDS-polyacrylamide gel electrophoresis (SDS-PAGE) of extracts from cells pulse-labeled for 10–20 min with ³⁵S-methionine showed that immune precipitates contained none of the nonglycosylated internal structural polypeptides of mature viruses. The major viral-specific polypeptides labeled in 10 min included polypeptides of 180,000, 140,000, 110,000, 80,000, and 60,000 daltons with minor polypeptides of 65,000, 50,000, and 40,000 daltons. Labeling the intracellular virus-specific polypeptides with ¹⁴C-glucosamine indicated that the 180,000, 110,000, 80,000, and 60,000 dalton polypeptides were glycosylated, and all but the 110,000 dalton polypeptides are contained in the mature virions. Based on pulse-chase experiments, it appears that at least 3 of the large polypeptides (140,000, 65,000, and 50,000 daltons) are precursors to the three major internal structural polypeptides of the mature virions.     

Antibody decoration of neurofilaments

We have decorated neurofilaments with antibodies against three polypeptides (designated here as H [mol wt = 195,000], 45[mol wt = 145,000], and 46[mol wt = 73,000]) in an effort to understand the arrangement of these polypeptides within neurofilaments. The three polypeptides were purified and antibodies were raised against each. The cross-reactivity of the antibodies suggested that each polypeptide contains both shared and unique antigenic determinants. The differential reactivities of each antibody preparation were enhanced by adsorption with the two heterologous polypeptides, and the resulting preparations were used to decorate purified neurofilaments, which were then negatively stained and examined in an electron microscope. The appearance of the antibody-decorated structures led to the following conclusions: All three polypeptides are physically associated with the same neurofilament. However, the disposition of H and 46 within a filament is different; 46 antigens appear to be associated with a "central core" of the filament, whereas H antigens compose a structure more loosely and peripherally attached to the central core and periodically arranged along its axis. The antibody-decorated H- containing structure assumes variable configurations; in some cases it appears asa bridge connecting two filaments; in other cases it appears as a helix wrapping the central core with a period of approximately 1,000 A and an apparent unit length of approximately 1.5 periods. These configurations suggest several functional implications, including the possibility that H is a component of the cross-bridges observed between filaments in situ. We also note that the central core-helix relationship could be used in the design of an intracellular transport motor.     

Isolation and partial characterization of a cage of filaments that surrounds the mammalian mitotic spindle

Mitotic cells have been detergent extracted under conditions that support microtubule assembly. When HeLa cells are lysed in the presence of brain tubulin, mitotic-arrested cells nucleate large asters and true metaphase cells yield spindles that remain enclosed within a roughly spherical cage of filamentous material. Detergent-extracted mitotic Chinese hamster ovary (CHO) cells show a similar, insoluble cage but the mitotic apparatus is only occasionally stabilized. In later stages of mitosis, HeLa cages are observed in elongated and furrowed configurations. In the terminal stages of cell division, two daughter filamentous networks are connected by the intercellular bridge. When observed in the electron microscope the cages include fibers 7-11 nm in diameter. The polypeptide composition of cages isolated from mitotic HeLa cells is complex, but the major polypeptides are a group with mol wt ranging from 43,000-60,000 daltons and a high molecular weight polypeptide. CHO cells contain a subset of these proteins which includes a major 58,000-dalton and a high molecular weight polypeptide. Two different antisera directed against the vimentin-containing intermediate filaments bind to polypeptides in the electrophoretic profiles of isolated HeLa and CHO cages and stain the cages, as visualized by indirect immunofluorescence. These results suggest that the HeLa and CHO cages include intermediate filaments of the vimentin type. The polypeptide composition of HeLa cages suggests that they also contain tonofilaments. The cages apparently form as the cells enter mitosis. We propose that these filamentous cages maintain the structural continuity of the cytoplasm while the cell is in mitosis.     

The polypeptides of isolated brain 10nm filaments and their association with polymerized tubulin.

Brain 10 nm filaments were isolated from bovine, rabbit and rat brains by a modification of an existing procedure. The overall polypeptide composition of these preparations was similar to that previously reported for brain neurofilaments. In addition to the major polypeptide component, which has mol. wt. approx. 50 000, three other polypeptides with chain mol. wts. approx. 210 000, 155 000 and 70 000, which correspond to peripheral-nerve neurofilament polypeptides, were consistently found to be present. The mol. wt.-50 000 species was found to be heterogeneous and may contain a component derived from the mol. wt. 70 000 polypeptide. The three higher-molecular-weight polypeptides did not appear to be obviously homologous or to be homologous with myosin or Myxicola neurofilament polypeptides. These same three higher-molecular-weight components were shown to be identical with the polypeptides probably responsible for the 10 nm filaments formed during the early cycles of the tubulin-purification protocol.     

The proteolytic digestion of ox neurofilaments with trypsin and alpha-chymotrypsin.

Brief digestion of ox neurofilaments with trypsin liberates fragments that are soluble and have molecular weights ranging from 164 000 to 97 000. Peptide fingerprinting indicates that these regions, termed the tryptic head-regions, arise from the 205 000- and 158 000-mol.wt. components of the triplet. The remains of the parent polypeptides sediment with normal filaments and have been termed tail-regions. Digestion of neurofilaments with chymotrypsin also liberates soluble fragments (chymotryptic head-regions) but these have mol.wts. 171 000 and 119 000, though they too originate from the higher-molecular-weight triplet polypeptides. Tryptic and chymotryptic head-regions have extensive homology, and a low (less than or equal to 20%) helix content. Electron microscopy shows that chymotryptic digestion rapidly reduces the length of filaments, probably because this enzyme preferentially attacks the 72 000-mol.wt. polypeptide. In contrast, brief digestion with trypsin does not reduce filament length even though more than 90% of the two higher-molecular-weight components have been cleaved. These results indicate that the backbone of native filaments is formed from the 72 000-mol.wt. polypeptide together with the tail-regions from the 205 000- and 158 000-mol.wt. polypeptides. The corresponding head-regions of these components, which can represent nearly 75% of each molecule, are not necessary for preserving the backbone of native neurofilaments and are therefore good candidates for being the side arms that connect these filaments in nerve cells.     

Biosynthesis of alpha-fetoprotein in cultured hepatoma cells

Pulse and pulse-chase experiments demonstrated that a heterogeneous polypeptide with an apparent Mr = 68,000 was the first intracellular anti-alpha-fetoprotein (AFP)-precipitable polypeptide synthesized by rat Mc-A-RH-7777 hepatoma cells. The 68,000-dalton polypeptide may consist of polypeptides with apparent molecular weights ranging from 68,000 to 70,000. It was the precursor of two intracellular anti-AFP-precipitable polypeptides of 69,000 and 73,000 apparent molecular weight. The latter were secreted into the medium without further processing. The anti-AFP-precipitable polypeptides in both cells and medium incorporated [3H]glucosamine, indicating that these polypeptides are at least partially glycosylated. The 68,000-dalton polypeptide in cells was bound mostly to concanavalin A-Sepharose, whereas the 69,000-dalton polypeptide was entirely unbound. The 73,000-dalton polypeptide consisted of concanavalin A-bound and -unbound variants. Tunicamycin completely abolished the uptake of [3H]glucosamine into anti-AFT-precipitable polypeptides in both cells and medium, and the resulting polypeptide of apparent Mr = 66,000 did not bind to concanavalin A-Sepharose. Tunicamycin did not affect the synthesis or secretion of AFP by hepatoma cells.     

Immunological and ultrastructural studies of neurofilaments isolated from rat peripheral nerve

Neurofilaments were isolated from desheathed and minced segments of rat peripheral nerve by osmotic shock into 0.01 M Tris-HCI buffer, pH 7.2. Freshly isolated neurofilaments were observed to undergo disassembly by progressive fragmentation upon exposure of dilute tissue extracts to this buffer. Low- and high-speed centrifugations of these tissue extracts separated membranous and particulate constituents and produced a progressive enrichment of 68,000-dalton polypeptide band in successive supernates, as determined by analyses of soluble proteins by SDS-polyacrylamide electrophoresis. The final high-speed supernatant fractions (S3) of nerve extracts, which were predominantly composed of 68,000-dalton polypeptide, were used to raise a specific experimental antisera in rabbits. Utilizing techniques of immune electron microscopy, experimental rabbit antisear was shown to contain antibodies against neurofilaments. Intact neurofilaments isolated from rat nerves and attached to carbon-coated grids became decorated when exposed to experimental rabbit antisera or purified gamma globulin (IgG) derivatives. The decoration of neurofilaments closely resembled the IgG coating seen in immune electron microscopy. Antibody absorption techniques were used to identify the biochemical constituency of neurofilamentous antigenic determinants. The decoration of neurofilament by experimental IgG was not altered by additions of tubulin or bovine serum albumin, but was prevented by additions of S3 fractions as well as the 68,000-dalton polypeptide of this fraction which was eluted and recovered from polyacrylamide gels. These findings are indicative that a 68,000-dalton polypeptide is a constituent subunit of rat peripheral nerve neurofilaments.