Structure of tau protein and assembly into paired helical filaments

Over the past few years the systematic investigation of paired helical filament assembly from tau protein in vitro has become feasible. We review our current understanding of the structure and conformations of tau protein and how this affects tau's assembly into the pathological paired helical filaments in Alzheimer's disease.     

Orientations of tyrosines 21 and 24 in coat subunits of Ff filamentous virus: determination by Raman linear intensity difference spectroscopy and implications for subunit packing.

Virions of the Ff group of bacteriophages (fd, f1, M13) are morphologically identical filaments (approximately 6-nm diameter x approximately 880-nm length) in which a covalently closed, single-stranded DNA genome is sheathed by approximately 2700 copies of a 50-residue alpha-helical subunit (pVIII). Orientations of pVIII tyrosines (Tyr21 and Tyr24) with respect to the filament axis have been determined by Raman linear intensity difference (RLID) spectroscopy of flow-oriented mutant virions in which the tyrosines were independently mutated to methionine. The results show that the twofold axis of the phenolic ring (C1-C4 line) of Tyr21 is inclined at 39.5 +/- 1.4 degrees from the virion axis, and that of Tyr24 is inclined at 43.7 +/- 0.6 degrees. The orientation determined for the Tyr21 phenol ring is close to that of a structural model previously proposed on the basis of fiber x-ray diffraction results (Protein Data Bank, identification code 1IFJ). On the other hand, the orientation determined for the Tyr24 phenol ring differs from the diffraction-based model by a 40 degrees rotation about the Calpha-Cbeta bond. The RLID results also indicate that each tyrosine mutation does not greatly affect the orientation of either the remaining tyrosine or single tryptophan (Trp26) of pVIII. On the basis of these results, a refined model is proposed for the coat protein structure in Ff.     

Nucleation-dependent tau filament formation: the importance of dimerization and an estimation of elementary rate constants

Filamentous inclusions composed of the microtubule-associated protein tau are found in Alzheimer disease and other tauopathic neurodegenerative diseases, but the mechanisms underlying their formation from full-length protein monomer under physiological conditions are unclear. To address this issue, the fibrillization of recombinant full-length four-repeat human tau was examined in vitro as a function of time and submicromolar tau concentrations using electron microscopy assay methods and a small-molecule inducer of aggregation, thiazine red. Data were then fit to a simple homogeneous nucleation model with rate constant constraints established from filament dissociation rate, critical concentration, and mass-per-unit length measurements. The model was then tested by comparing the predicted time-dependent evolution of length distributions to experimental data. Results indicated that once assembly-competent conformations were attained, the rate-limiting step in the fibrillization pathway was tau dimer formation. Filament elongation then proceeded by addition of tau monomers to nascent filament ends. Filaments isolated at reaction plateau contained approximately 2 tau protomers/beta-strand spacing on the basis of mass-per-unit length measurements. The model suggests four key steps in the aggregation pathway that must be surmounted for tau filaments to form in disease.     

Thermoregulation in hypergravity-acclimated rats

To determine the effect of hypergravity acclimation on thermoregulation, core temperature (Tc), tail temperature (Tt), and O2 consumption (VO2) were measured in control rats (raised at 1 G) and in rats acclimated to 2.1 G. When the animals were exposed to a low ambient temperature of 9 degrees C, concurrently with a hypergravic field of 2.1 G, Tc of rats raised at 1 G fell markedly by approximately 6 degrees C (to 30.8 +/- 0.6 degrees C) while that of the rats raised at 2.1 G remained relatively constant (falling only approximately 1 degree C to 36.4 +/- 0.3 degrees C). Thus prior acclimation to a 2.1-G field enabled rats to maintain Tc when cold exposed in a 2.1-G field. To maintain Tc, thermogenic mechanisms were successfully activated in the 2.1-G-acclimated rats as shown by measurements of VO2. In contrast, VO2 measurements showed that rats reared at 1 G and then cold exposed at 2.1 G did not activate thermogenic mechanisms sufficiently to prevent a fall in Tc. In other experiments, rats acclimated to either 1 or 2.1 G were found to lack the ability to maintain their Tc when exposed to a 5.8-G field or when exposed to prolonged cold exposure at 1 G. Results are interpreted as showing that when placed in a 2.1-G field, rats acclimated to 2.1 G can more closely maintain their Tc near 37 degrees C when cold exposed than can rats acclimated to 1 G. However, this enhanced regulatory ability of 2.1-G-acclimated rats over 1.0-G-acclimated rats is restricted to 2.1-G fields and is not observed in 1.0- and 5.8-G fields.     

Analysis of the dynamic properties of Bacillus circulans xylanase upon formation of a covalent glycosyl-enzyme intermediate

NMR spectroscopy was used to search for mechanistically significant differences in the local mobility of the main-chain amides of Bacillus circulans xylanase (BCX) in its native and catalytically competent covalent glycosyl-enzyme intermediate states. 15N T1, T2, and 15N[1H] NOE values were measured for approximately 120 out of 178 peptide groups in both the apo form of the protein and in BCX covalently modified at position Glu78 with a mechanism-based 2-deoxy-2-fluoro-beta-xylobioside inactivator. Employing the model-free formalism of Lipari and Szabo, the measured relaxation parameters were used to calculate a global correlation time (tau(m)) for the protein in each form (9.2 +/- 0.2 ns for apo-BCX; 9.8 +/- 0.3 ns for the modified protein), as well as individual order parameters for the main-chain NH bond vectors. Average values of the order parameters for the protein in the apo and complexed forms were S2 = 0.86 +/- 0.04 and S2 = 0.91 +/- 0.04, respectively. No correlation is observed between these order parameters and the secondary structure, solvent accessibility, or hydrogen bonding patterns of amides in either form of the protein. These results demonstrate that the backbone of BCX is well ordered in both states and that formation of the glycosyl-enzyme intermediate leads to little change, in any, in the dynamic properties of BCX on the time scales sampled by 15N-NMR relaxation measurements.     

Plain and Complex Flagella of Pseudomonas rhodos: Analysis of Fine Structure and Composition

Cells of Pseudomonas rhodos 9-6 produce two morphologically distinct flagella termed plain and complex, respectively. Fine structure analyses by electron microscopy and optical diffraction showed that plain flagellar filaments are cylinders of 13-nm diameter composed of globular subunits like normal bacterial flagella. The structure comprises nine large-scale helical rows of subunits intersecting four small-scale helices of pitch angle 25 degrees . Complex filaments have a conspicuous helical sheath, 18-nm wide, of three close-fitting helical bands, each about 4.7-nm wide, separated by axial intervals, 4.7 nm wide, running at an angle of 27 degrees . The internal core has similar but not identical substructure to plain filaments. Unlike plain flagella, the complex species is fragile and does not aggregate in bundles. Mutants bearing only one of two types of flagellum were isolated. Cells with plain flagella showed normal translational motion, and cells with complex flagella showed rapid spinning. Isolated plain flagella consist of a 37,000-dalton subunit separable into two isoproteins. Complex filaments consist of a 55,000-dalton protein; a second 43,000-dalton protein was assigned to complex flagellar hooks. The results indicate that plain and complex flagella are entirely different in structure and composition and that the complex type represents a novel flagellar species. Its possible mode of action is discussed.     

Tuning microtubule-based transport through filamentous MAPs: the problem of dynein

We recently proposed that regulating the single-to-multiple motor transition was a likely strategy for regulating kinesin-based transport in vivo . In this study, we use an in vitro bead assay coupled with an optical trap to investigate how this proposed regulatory mechanism affects dynein-based transport. We show that tau's regulation of kinesin function can proceed without interfering with dynein-based transport. Surprisingly, at extremely high tau levels – where kinesin cannot bind microtubules (MTs) – dynein can still contact MTs. The difference between tau's effects on kinesin- and dynein-based motility suggests that tau can be used to tune relative amounts of plus-end and minus-end-directed transport. As in the case of kinesin, we find that the 3RS isoform of tau is a more potent inhibitor of dynein binding to MTs. We show that this isoform-specific effect is not because of steric interference of tau's projection domains but rather because of tau's interactions with the motor at the MT surface. Nonetheless, we do observe a modest steric interference effect of tau away from the MT and discuss the potential implications of this for molecular motor structure.     

Action of cytochalasin D on cytoskeletal networks

Extraction of SC-1 cells (African green monkey kidney) with the detergent Triton X-100 in combination with stereo high-voltage electron microscopy of whole mount preparations has been used as an approach to determine the mode of action of cytochalasin D on cells. The cytoskeleton of extracted BSC-1 cells consists of substrate-associated filament bundles (stress fibers) and a highly cross-linked network of four major filament types extending throughout the cell body; 10-nm filaments, actin microfilaments, microtubules, and 2- to 3-nm filaments. Actin filaments and 2- to 3-nm filaments form numerous end- to-side contacts with other cytoskeletal filaments. Cytochalasin D treatment severely disrupts network organization, increases the number of actin filament ends, and leads to the formation of filamentous aggregates or foci composed mainly of actin filaments. Metabolic inhibitors prevent filament redistribution, foci formation, and cell arborization, but not disorganization of the three-dimensional filament network. In cells first extracted and then treated with cytochalasin D, network organization is disrupted, and the number of free filament ends is increased. Supernates of preparations treated in this way contain both short actin filaments and network fragments (i.e., actin filaments in end-to-side contact with other actin filaments). It is proposed that the dramatic effects of cytochalasin D on cells result from both a direct interaction of the drug with the actin filament component of cytoskeletal networks and a secondary cellular response. The former leads to an immediate disruption of the ordered cytoskeletal network that appears to involve breaking of actin filaments, rather than inhibition of actin filament-filament interactions (i.e., disruption of end-to-side contacts). The latter engages network fragments in an energy-dependent (contractile) event that leads to the formation of filament foci.     

Three- and four-repeat Tau coassemble into heterogeneous filaments: an implication for Alzheimer disease

Tau filaments are the pathological hallmark of numerous neurodegenerative diseases including Alzheimer disease, Pick disease, and progressive supranuclear palsy. In the adult human brain, six isoforms are expressed that differ by the presence or absence of the second of four semiconserved repeats. As a consequence, half of the tau isoforms have three repeats (3R tau), whereas the other half of the isoforms have four repeats (4R tau). Tauopathies can be characterized based on the isoform composition of their filaments. Alzheimer disease filamentous inclusions contain all isoforms. Pick disease filaments contain 3R tau. Progressive supranuclear palsy filaments contain 4R tau. Here, we used site-directed spin labeling of recombinant tau in conjunction with electron paramagnetic resonance spectroscopy to obtain structural insights into these filaments. We find that filaments of 4R tau and 3R tau share a highly ordered core structure in the third repeat with parallel, in-register arrangement of β-strands. This structure is conserved regardless of whether full-length isoforms (htau40 and htau23) or truncated constructs (K18 and K19) are used. When mixed, 3R tau and 4R tau coassemble into heterogeneous filaments. These filaments share the highly ordered core in the third repeat; however, they differ in their overall composition. Our findings indicate that at least three distinct types of filaments exist: homogeneous 3R tau, homogeneous 4R tau, and heterogeneous 3R/4R tau. These results suggest that individual filaments found in Alzheimer disease are structurally distinct from those in the 3R and 4R tauopathies.     

MARKing tau for tangles and toxicity

In healthy neurons, tau proteins regulate microtubule function in the axon. In the brains of individuals with Alzheimer's disease, tau is hyperphosphorylated and aggregated into intraneuronal deposits called neurofibrillary tangles (NFTs). Hyperphosporylation dislodges tau from the microtubule surface, potentially resulting in compromised axonal integrity and the accumulation of toxic tau peptides. Recent biochemical and animal model studies have re-evaluated tau phosphorylation and other aspects of neurofibrillar pathology. The results indicate that phosphorylation of tau's microtubule-binding domain by the protein kinase MARK primes tau for hyperphosphorylation by the kinases GSK-3 and Cdk5, which in turn triggers the aggregation of tau into filaments and tangles. Toxic consequences for the neuron might be exacerbated by tangle formation but are already evident during the early steps of the process.     

In vitro assembly of tau protein: mapping the regions involved in filament formation

Unraveling the mechanism of self-assembly of the protein tau into paired helical filaments (PHFs) is a crucial step toward the understanding of Alzheimer's and other neuropathological diseases at the molecular level. In an effort to map the role of different regions of tau in the mechanism of self-assembly, we have studied the polymerization ability of different tau fragments using an in vitro assay. Our results indicate that the N-terminal domain interferes with tau's ability to polymerize in vitro. The effect seems to be size dependent. Particularly, an isoform of tau from the peripheral nervous system, which has a much larger N-terminal domain, was found unable to form filaments in our in vitro assay. This finding can explain why in Alzheimer's patients PHFs only accumulate in the neurons from the central nervous system. We also report that a short segment of tau located in the third microtubule binding repeat (residues 317 to 335, peptide 1/2R) is probably the minimal segment of that region able to grow into filaments in vitro and in the presence of heparin. In contrast with whole peptide 1/2R, peptides corresponding to either the N-terminal or C-terminal halves of this segment were unable to form filaments. Finally, our polymerization studies of peptides from the C-terminal domain reveal a short sequence spanning residues 391 to 407 that grows into filaments in vitro. This tau segment forms filaments regardless of whether is incubated with heparin. Moreover, such filaments differ in diameter and morphology, suggesting a different mechanism of self-assembly.     

Three-dimensional image reconstruction of insect flight muscle. II. The rigor actin layer

The averaged structure of rigor cross-bridges in insect flight muscle is further revealed by three-dimensional reconstruction from 25-nm sections containing a single layer of thin filaments. These exhibit two thin filament orientations that differ by 60 degrees from each other and from myac layer filaments. Data from multiple tilt views (to +/- 60 degrees) was supplemented by data from thick sections (equivalent to 90 degrees tilts). In combination with the reconstruction from the myac layer (Taylor et al., 1989), the entire unit cell is reconstructed, giving the most complete view of in situ cross-bridges yet obtained. All our reconstructions show two classes of averaged rigor cross-bridges. Lead bridges have a triangular shape with leading edge angled at approximately 45 degrees and trailing edge angled at approximately 90 degrees to the filament axis. We propose that the lead bridge contains two myosin heads of differing conformation bound along one strand of F-actin. The lead bridge is associated with a region of the thin filament that is apparently untwisted. We suggest that the untwisting may reflect the distribution of strain between myosin and actin resulting from two-headed, single filament binding in the lead bridge. Rear bridges are oriented at approximately 90 degrees to the filament axis, and are smaller and more cylindrical, suggesting that they consist of single myosin heads. The rear bridge is associated with a region of apparently normal thin filament twist. We propose that differing myosin head angles and conformations consistently observed in rigor embody different stages of the power stroke which have been trapped by a temporal sequence of rigor cross-bridge formation under the constraints of the intact filament lattice.     

Sarcomere length dispersion in single skeletal muscle fibers and fiber bundles.

Light diffraction patterns produced by single skeletal muscle fibers and small fiber bundles of Rana pipiens semitendinosus have been examined at rest and during tetanic contraction. The muscle diffraction patterns were recorded with a vidicon camera interfaced to a minicomputer. Digitized video output was analyzed on-line to determine mean sarcomere length, line intensity, and the distribution of sarcomere lengths. The occurrence of first-order line intensity and peak amplitude maxima at approximately 3.0 mum is interpreted in terms of simple scattering theory. Measurements made along the length of a singel fiber reveal small variations in calculated mean sarcomere length (SD about 1.2%) and its percent dispersion (2.1% +/- 0.8%). Dispersion in small multifiber preparations increases approximately linearly with fiber number (about 0.2% per fiber) to a maximum of 8-10% in large bundles. Dispersion measurements based upon diffraction line analysis are comparable to SDs calculated from length distribution histograms obtained by light micrography of the fiber. First-order line intensity decreases by about 40% during tetanus; larger multifibered bundles exhibit substantial increases in sarcomere dispersion during contraction, but single fibers show no appreciable dispersion change. These results suggest the occurrence of asynchronous static or dynamic axial disordering of thick filaments, with a persistence in long range order of sarcomere spacing during contraction in single fibers.     

Architecture of Ure2p prion filaments: the N-terminal domains form a central core fiber

The [URE3] prion is an inactive, self-propagating, filamentous form of the Ure2 protein, a regulator of nitrogen catabolism in yeast. The N-terminal "prion" domain of Ure2p determines its in vivo prion properties and in vitro amyloid-forming ability. Here we determined the overall structures of Ure2p filaments and related polymers of the prion domain fused to other globular proteins. Protease digestion of 25-nm diameter Ure2p filaments trimmed them to 4-nm filaments, which mass spectrometry showed to be composed of prion domain fragments, primarily residues approximately 1-70. Fusion protein filaments with diameters of 14-25 nm were also reduced to 4-nm filaments by proteolysis. The prion domain transforms from the most to the least protease-sensitive part upon filament formation in each case, implying that it undergoes a conformational change. Intact filaments imaged by cryo-electron microscopy or after vanadate staining by scanning transmission electron microscopy (STEM) revealed a central 4-nm core with attached globular appendages. STEM mass per unit length measurements of unstained filaments yielded 1 monomer per 0.45 nm in each case. These observations strongly support a unifying model whereby subunits in Ure2p filaments, as well as in fusion protein filaments, are connected by interactions between their prion domains, which form a 4-nm amyloid filament backbone, surrounded by the corresponding C-terminal moieties.     

Brain Levels of Microtubule-Associated Protein τ Are Elevated in Alzheimer''s Disease: A Radioimmuno-Slot-Blot Assay for Nanograms of the Protein

The microtubule-associated protein tau, which stimulates the assembly of alpha-beta tubulin heterodimers into microtubules, is abnormally phosphorylated in Alzheimer's disease (AD) brain and is the major component of paired helical filaments. In the present study, the levels of tau and abnormally phosphorylated tau were determined in brain homogenates of AD and age-matched control cases. A radioimmuno-slot-blot assay was developed, using a primary monoclonal antibody, Tau-1, and a secondary antibody, antimouse 125I-immunoglobulin G. To assay the abnormally phosphorylated tau, the blots were treated with alkaline phosphatase before immunolabeling. The levels of total tau were about eightfold higher in AD (7.3 +/- 2.7 ng/micrograms of protein) than in control cases (0.9 +/- 0.2 ng/micrograms), and this increase was in the form of the abnormally phosphorylated protein. These studies indicate that the abnormal phosphorylation--not a decrease in the level of tau--is a likely cause of neurofibrillary degeneration in AD.     

Phosphorylation of tau at Ser214 mediates its interaction with 14-3-3 protein: implications for the mechanism of tau aggregation

The microtubule associated protein tau is a major component of neurofibrillary tangles in Alzheimer disease brain, however the neuropathological processes behind the formation of neurofibrillary tangles are still unclear. Previously, 14-3-3 proteins were reported to bind with tau. 14-3-3 Proteins usually bind their targets through specific serine/threonine –phosphorylated motifs. Therefore, the interaction of tau with 14-3-3 mediated by phosphorylation was investigated. In this study, we show that the phosphorylation of tau by either protein kinase A (PKA) or protein kinase B (PKB) enhances the binding of tau with 14-3-3 in vitro . The affinity between tau and 14-3-3 is increased 12- to 14-fold by phosphorylation as determined by real time surface plasmon resonance studies. Mutational analyses revealed that Ser214 is critical for the phosphorylation-mediated interaction of tau with 14-3-3. Finally, in vitro aggregation assays demonstrated that phosphorylation by PKA/PKB inhibits the formation of aggregates/filaments of tau induced by 14-3-3. As the phosphorylation at Ser214 is up-regulated in fetal brain, tau's interaction with 14-3-3 may have a significant role in the organization of the microtubule cytoskeleton in development. Also as the phosphorylation at Ser214 is up-regulated in Alzheimer's disease brain, tau's interaction with 14-3-3 might be involved in the pathology of this disease.     

Strong binding of myosin heads stretches and twists the actin helix

Calculation of the size of the power stroke of the myosin motor in contracting muscle requires knowledge of the compliance of the myofilaments. Current estimates of actin compliance vary significantly introducing uncertainty in the mechanical parameters of the motor. Using x-ray diffraction on small bundles of permeabilized fibers from rabbit muscle we show that strong binding of myosin heads changes directly the actin helix. The spacing of the 2.73-nm meridional x-ray reflection increased by 0.22% when relaxed fibers were put into low-tension rigor (<10 kN/m(2)) demonstrating that strongly bound myosin heads elongate the actin filaments even in the absence of external tension. The pitch of the 5.9-nm actin layer line increased by approximately 0.62% and that of the 5.1-nm layer line decreased by approximately 0.26%, suggesting that the elongation is accompanied by a decrease in its helical angle (approximately 166 degrees) by approximately 0.8 degrees. This effect explains the difference between actin compliance revealed from mechanical experiments with single fibers and from x-ray diffraction on whole muscles. Our measurement of actin compliance obtained by applying tension to fibers in rigor is consistent with the results of mechanical measurements.     

The formation of straight and twisted filaments from short tau peptides

We studied fibril formation in a family of peptides based on PHF6 (VQIVYK), a short peptide segment found in the microtubule binding region of tau protein. N-Acetylated peptides AcVYK-amide (AcVYK), AcIVYK-amide (AcPHF4), AcQIVYK-amide (AcPHF5), and AcV-QIVYK-amide (AcPHF6) rapidly formed straight filaments in the presence of 0.15 m NaCl, each composed of two laterally aligned protofilaments approximately 5 nm in width. X-ray fiber diffraction showed the omnipresent sharp 4.7-A reflection indicating that the scattering objects are likely elongated along the hydrogen-bonding direction in a cross-beta conformation, and Fourier transform IR suggested the peptide chains were in a parallel (AcVYK, AcPHF6) or antiparallel (AcPHF4, AcPHF5) beta-sheet configuration. The dipeptide N-acetyl-YK-amide (AcYK) formed globular structures approximately 200 nm to 1 microm in diameter. The polymerization rate, as measured by thioflavin S binding, increased with the length of the peptide going from AcYK --> AcPHF6, and peptides that aggregated most rapidly displayed CD spectra consistent with beta-sheet structure. There was a 3-fold decrease in rate when Val was substituted for Ile or Gln, nearly a 10-fold decrease when Ala was substituted for Tyr, and an increase in polymerization rate when Glu was substituted for Lys. Twisted filaments, composed of four laterally aligned protofilaments (9-19 nm width, approximately 90 nm half-periodicity), were formed by mixing AcPHF6 with AcVYK. Taken together these results suggest that the core of PHF6 is localized at VYK, and the interaction between small amphiphilic segments of tau may initiate nucleation and lead to filaments displaying paired helical filament morphology.     

Two distinct functions of the carboxyl-terminal tail domain of NF-M upon neurofilament assembly: cross-bridge formation and longitudinal elongation of filaments

Neurofilaments are the major cytoskeletal elements in the axon that take highly ordered structures composed of parallel arrays of 10-nm filaments linked to each other with frequent cross-bridges, and they are believed to maintain a highly polarized neuronal cell shape. Here we report the function of rat NF-M in this characteristic neurofilament assembly. Transfection experiments were done in an insect Sf9 cell line lacking endogenous intermediate filaments. NF-L and NF-M coassemble to form bundles of 10-nm filaments packed in a parallel manner with frequent cross-bridges resembling the neurofilament domains in the axon when expressed together in Sf9 cells. Considering the fact that the expression of either NF-L or NF-M alone in these cells results in neither formation of any ordered network of 10-nm filaments nor cross- bridge structures, NF-M plays a crucial role in this parallel filament assembly. In the case of NF-H the carboxyl-tail domain has been shown to constitute the cross-bridge structures. The similarity in molecular architecture between NF-M and NF-H suggests that the carboxyl-terminal tail domain of NF-M also constitutes cross-bridges. To examine this and to further investigate the function of the carboxyl-terminal tail domain of NF-M, we made various deletion mutants that lacked part of their tail domains, and we expressed these with NF-L. From this deletion mutant analysis, we conclude that the carboxyl-terminal tail domain of NF-M has two distinct functions. First, it is the structural component of cross-bridges, and these cross-bridges serve to control the spacing between core filaments. Second, the portion of the carboxyl- terminal tail domain of NF-M that is directly involved in cross-bridge formation affects the core filament assembly by helping them to elongate longitudinally so that they become straight.     

Physicochemical characterization of the 68 000-dalton protein of bovine neurofilaments

The 68 000-dalton protein from bovine neurofilaments was purified by a combination of chromatography on DEAE-cellulose and on hydroxylapatite in buffers containing 8 M urea. Although the separation of this protein from the other proteins of the neurofilament appeared to be hampered by a mixed association of the several components, a nearly homogeneous product was obtained for study. Sedimentation equilibrium experiments in buffers containing 8 M urea showed the molecule to be a monomer with a molecular weight of 70 600 +/- 2000. Circular dichroic spectra taken under the same conditions gave no evidence of residual alpha-helix. Molecular sieve chromatography in 8 M urea on controlled-pore glass showed that the molecule eluted at an unexpectedly small volume. The small elution volume did not depend significantly on protein concentration and is unlikely to be the result of intermolecular association. Rather, the monomer probably has a conformation more rigid or extended than a classical random coil. When dialyzed into 0.01 M tris(hydroxymethyl)aminomethane/1 mM ethylene glycol bis(beta-aminoethyl ether)-N,N,N',N'-tetraacetic acid/0.1 mM dithioerythritol, pH 8.5, the protein does not assemble into filaments. Sedimentation velocity reveals that under these conditions it consists mainly of a 4.8S molecular species, containing few large particles; sedimentation equilibrium shows that it is composed of oligomers, the smallest present in significant concentration having a molecular weight approximately that of a trimer. Circular dichroism measurements lead to the interpretation that the molecule has refolded in this buffer into a structure that has approximately 55% alpha-helix. Assembly into filamentous particles resembling neurofilaments occurs when the protein is dialyzed against 0.1 M 2-(N-morpholino)ethane-sulfonic acid/0.1% beta-mercaptoethanol/1 mM ethylene glycol bis(beta-aminoethyl ether)-N,N,N',N'-tetraacetic acid/0.17 M NaCl, pH 6.5. We suggest that the oligomeric species present in 0.01 M tris(hydroxymethyl)aminomethane may frequently be present in solubilized preparations of intermediate filaments and may represent an intermediate in the assembly process.