Reassembly of the 66 kD neurofilament protein in vitro following isolation and purification from bovine spinal cord

A Possible Cellular Mechanism of Neuronal Loss in the Dorsal Root Ganglia of Dystonia musculorum (dt) Mice

Dystonia musculorum (dt) is a mutant mouse with hereditary sensory neuropathy. A defective bullous pemphigoid antigen 1 (BPAG1) gene is responsible for this mutation. In the present study, we examined the distribution of neuronal intermediate filament proteins in the central and peripheral processes of the dorsal root ganglia (DRG) in adult dt mice using different approaches. We found that not only BPAG1, but also alpha-internexin was absent in the DRG neurons in adult dt mice. To study the relationship between the absence of alpha-internexin and the progressive neuronal loss in the DRG of dt mice, we further cultured DRG neurons from embryonic dt mutants. Immunocytochemical assay of cultured DRG neurons from dt embryos revealed that alpha-internexin was aggregated in the proximal region of axons and juxtanuclear region of the cytoplasma, yet the other intermediate filament proteins were widely distributed in all processes. The active caspase-3 activity was observed in the dt neuron with massive accumulation of alpha-internexin. From our observations, we suggest that the interaction between BPAG1 and alpha-internexin may be one of the key factors involved in neuronal degeneration, and abnormal accumulation of alpha-internexin may impair the axonal transport and subsequently turns on the cascade of neuronal apoptosis in dt mice.

Identification of endogenous phosphorylation sites of bovine medium and low molecular weight neurofilament proteins by tandem mass spectrometry

Neurofilament proteins (NFP) are intermediate filaments found in the neuronal cytoskeleton. They are highly phosphorylated, a condition that is believed to be responsible for the assembly and stability of the filaments. Matrix-assisted laser desorption/ionization (MALDI) time-of-flight (TOF) mass spectrometry (MS) shows molecular masses for bovine NFP subunits of 63, 105, and 125 kDa for NFL, NFM, and NFH. Mass spectrometric de novo sequencing was used to determine the N-terminal sequence of bovine NFM (115 amino acids), which was previously unknown. Molecular mass information shows that there is one-half equivalent phosphate group on NFL and 24 on NFM. For the first time, it is shown that bovine NFL has three phosphorylation sites (Ser(55), Ser(66), and Ser(472)) and NFM has 22 (Ser(512), Ser(546), Ser(554), Ser(560), Thr(627), Ser(629), Ser(634), Ser(639), Thr(646), Ser(649), Ser(654), Ser(664), Ser(669), Thr(676), Ser(679), Ser(684), Ser(694), Ser(726), Ser(750), Ser(756), Ser(770), and Ser(846)) and two tentative sites (Ser(659)/Thr(661) and Thr(840)). Ser(66) was previously not known to be phosphorylated in NFL of other species, while two sites (Ser(55) and Ser(472)) are consistent with the phosphorylations observed in other mammalian NFLs. The three sites, Ser(55), Ser(66), Ser(472), are heterogeneously phosphorylated. Phosphorylation in bovine NFM occurs mainly in the Lys-Ser-Pro (KSP) region, but the Val-Ser-Pro and Ser-Glu-Lys motifs are also phosphorylated. Most of the phosphorylation sites are in accordance with those previously identified in other mammalian NFMs. In bovine NFM, 16 out of the 22 sites are always phosphorylated (Ser(512), Thr(627), Ser(629), Ser(634), Ser(639), Thr(646), Ser(649), Ser(654), Ser(664), Ser(669), Thr(676), Ser(679), Ser(684), Ser(694), Ser(726), and Ser(750)), all of which are contained in the KSP region, and six are sometimes phosphorylated (Ser(546), Ser(554), Ser(560), Ser(756), Ser(770), and Ser(846)). The NFPs have other modifications, including deamidation, oxidation, and N-terminal acetylation. Pyroglutamic acid formation also occurs.

Neurofilament functions in health and disease

Transgenic approaches have recently been used to investigate the functions of neuronal intermediate filaments. Gene knockout studies have demonstrated that neurofilaments are not required for axogenesis and that individual neurofilament proteins play distinct roles in filament assembly and in the radial growth of axons. The involvement of neurofilaments in disease is supported by the discovery of novel mutations in the neurofilament heavy gene from cases of amyotrophic lateral sclerosis and by reports of neuronal death in mouse models expressing neurofilament and alpha-internexin transgenes. However, mouse studies have shown that axonal neurofilaments are not required for pathogenesis caused by mutations in superoxide dismutase and that increasing perikaryal levels of neurofilament proteins may even confer protection in this disease.

Neuronal intermediate filament protein alpha-internexin facilitates axonal neurite elongation in neuroblastoma cells

We examined the localization and role of alpha-IN vs. other neuronal intermediate filaments before and during differentiation. Vimentin but not alpha-IN localized within filopodia-like neurites of undifferentiated cells. During differentiation, alpha-IN immunoreactivity accumulated within axonal neurites following vimentin but, as previously describe in neurons in situ, before the appearance of NF-L. We therefore manipulated alpha-IN synthesis, accumulation, and function in attempts to determine whether or not this intermediate filament species played a role in axonal development. Intracellular delivery of anti-alpha-IN antisense oligonucleotides and antibodies was permissive for neuritogenesis, yet compromised neurite elongation; this effect was further reflected in diminished levels of stabilized axonal microtubules. These data suggest that alpha-IN plays a role in the development of neuronal polarity. Relatively more alpha-IN than NF-L accumulated within the plastic axonal neurites induced following serum-deprivation, while stable, dbcAMP-induced neurites treatment contained equivalent levels of each. Protease inhibition increased NF-L and NF-H but not alpha-IN immunoreactivity within serum-deprived neurites, suggesting that proteolysis restricts NF-L accumulation pending neurite stabilization. To test the possibility that NF-H accumulation is dependent upon NF-L and cannot be mediated by alpha-IN, we examined levels of NF-H co-precipitated from cells with alpha-IN and NF-L. Virtually all newly synthesized NF-H co-precipitated with NF-L, while only a small percentage co-precipitated with alpha-IN. Finally, NF-H or NF-M were absent from the axon hillock or perikaryal area at the base of neurites, where alpha-IN immunoreactivity is prominent. These data extend earlier cell-free demonstrations that NF-H preferentially associates with NF-L rather than alpha-IN.

No requirement of alpha-internexin for nervous system development and for radial growth of axons

Alpha-Internexin is a type IV intermediate filament protein that is expressed abundantly in neurons during development of the peripheral and central nervous systems as well as in few neurons of the adult central nervous system. It has been suggested that alpha-internexin may act as a scaffold for the formation of neuronal intermediate filaments during early development. In addition, recent reports suggest that alpha-internexin could play a major role in two degenerative neurological disorders. We report here an analysis of mice with a targeted disruption of alpha-internexin gene. Unexpectedly, alpha-internexin -/- mice developed normally and did not exhibit overt phenotypes. Moreover, the absence of alpha-internexin did not interfere with neurite extension of cultured DRG neurons. The number and caliber of L4 ventral root axons remained unchanged in alpha-internexin -/- mice. In the retina, alpha-internexin begins to be expressed in retinal ganglion cells when their first axons reach the optic chiasma. Using HRP tracer, we show that the projection pattern of the RGC axons is not modified by the absence of alpha-internexin. Electron microscopy did not reveal significant differences in axonal calibers, in myelination of axons and in neurofilament structures between alpha-internexin -/- and control mice during development and at adult stage. These data indicate that alpha-internexin is not required for the polymerization of neurofilament in vivo. Mice deficient for both alpha-internexin and neurofilament light chain (NF-L) exhibited no over phenotypes as well. No intermediate filament structures were detectable in optic nerve of alpha-internexin -/-; NF-L -/- mice. Ours results do not support the hypothesis of a role for type IV intermediate filaments in axonal outgrowth during development of nervous system.

The pathway of assembly of intermediate filaments from recombinant alpha-internexin

The pathway of filament assembly from the neuronal intermediate filament alpha-intermexin was investigated. Optimal assembly occurred in solutions of pH 6.5 to 7 and moderate ionic strength at 37 degrees C. Short filaments formed upon dialysis at 24 degrees C, which elongated further when incubated at 37 degrees C. Soluble forms of alpha-internexin were characterized by analytical ultracentrifugation and electron microscopy. In 10 mM Tris, pH 8, conditions that favor formation of tetramers and other small oligomers for other intermediate filament proteins, alpha-internexin formed 10.5 S particles, apparently unit-length half-filaments in the form of rods 10.6 nm in diameter and 68 nm long. Dialysis vs the same buffer with added 10 mM NaCl yielded 16 S rods, probably unit-length filaments, of the same length but 13.0 nm in diameter. At 50 mM NaCl, rods about 13 nm in diameter and heterogeneous in length were observed in electron micrographs, apparently formed from longitudinal annealing of unit-length rods. The results favor a model of assembly in which coiled coil dimers aggregate laterally to form first "unit-length half-filaments" (Herrmann, H., and Aebi, U. (1998) Curr. Opin. Struct. Biol. 8, 177-185) and then "unit-length filaments," which subsequently elongate by annealing.

Developmentally regulated stabilization of neuronal intermediate filaments in rat cerebral cortex

The expression and Triton X-100 (Triton) solubility of neuronal intermediate filament proteins were determined in the developing rat cerebral cortex. The level of expression of alpha-internexin was unchanged from embryonic day 15 (E15) to postnatal day 15 (P15), whereas expression of the mid-sized neurofilament subunit increased continually during this interval concomitant with a reduction in Triton solubility of the two proteins. The low molecular weight neurofilament subunit, first barely detected at P2, was largely insoluble in Triton from the initial time point that its solubility could be assayed, at P5, to P15. Similar expression patterns and Triton solubility profiles were obtained for neuronal intermediate filament proteins in cultured neurons from E15 cerebral cortex. These results suggest that alpha-internexin is expressed earlier than neurofilament proteins to provide a more plastic network in the early developing brain. The incorporation of neurofilament proteins apparently results in the formation of the more stable intermediate filament network found in mature neurons.