Phosphorylation determines two distinct species of Tau in the central nervous system

The monoclonal antibody, Tau-1, which had previously been used to localize tau to the axonal compartment in brain has been reutilized for light and electron microscopic immunohistochemistry following phosphatase treatment of tissue. We report here that a significant quantity of tau in the central nervous system is phosphorylated in situ at or near the Tau-1 epitope, preventing the binding of the Tau-1 antibody. Upon removal of this/these phosphate group(s), however, Tau-1 was observed in the somatodendritic compartment of neurons as well as in axons. Furthermore, intense staining was also observed in astrocytes and in perineuronal glial cells. This immunoreactivity was present along the lengths of microtubules and on ribosomes (polysomes). Treatment of immunoblots of extracts of whole cerebral cortex with phosphatase confirmed the immunohistochemical results in that a 50-65% increase in Tau-1 binding to the tau region of the blot was noted. Moreover, a novel monoclonal antibody, Tau-2, was also used in these experiments. This antibody binds only to tau and localizes along microtubules in axons, somata, dendrites, and astrocytes and on ribosomes (polysomes) without phosphatase pretreatment.     

Widespread distribution of the major polypeptide component of MAP 1 (microtubule-associated protein 1) in the nervous system

We prepared a monoclonal antibody to microtubule-associated protein 1 (MAP 1), one of the two major high molecular weight MAP found in microtubules isolated from brain tissue. We found that MAP 1 can be resolved by SDS PAGE into three electrophoretic bands, which we have designated MAP 1A, MAP 1B, and MAP 1C in order of increasing electrophoretic mobility. Our antibody recognized exclusively MAP 1A, the most abundant and largest MAP 1 polypeptide. To determine the distribution of MAP 1A in nervous system tissues and cells, we examined tissue sections from rat brain and spinal cord, as well as primary cultures of newborn rat brain by immunofluorescence microscopy. Anti-MAP 1A stained white matter and gray matter regions, while a polyclonal anti-MAP 2 antibody previously prepared in this laboratory stained only gray matter. This confirmed our earlier biochemical results, which indicated that MAP 1 is more uniformly distributed in brain tissue than MAP 2 (Vallee, R.B., 1982, J. Cell Biol., 92:435-442). To determine the identity of cells and cellular processes immunoreactive with anti-MAP 1A, we examined a variety of brain and spinal cord regions. Fibrous staining of white matter by anti-MAP 1A was generally observed. This was due in part to immunoreactivity of axons, as judged by examination of axonal fiber tracts in the cerebral cortex and of large myelinated axons in the spinal cord and in spinal nerve roots. Cells with the morphology of oligodendrocytes were brightly labeled in white matter. Intense staining of Purkinje cell dendrites in the cerebellar cortex and of the apical dendrites of pyramidal cells in the cerebral cortex was observed. By double-labeling with antibodies to MAP 1A and MAP 2, the presence of both MAP in identical dendrites and neuronal perikarya was found. In primary brain cell cultures anti-MAP 2 stained predominantly cells of neuronal morphology. In contrast, anti-MAP 1A stained nearly all cells. Included among these were neurons, oligodendrocytes and astrocytes as determined by double-labeling with anti-MAP 1A in combination with antibody to MAP 2, myelin basic protein or glial fibrillary acidic protein, respectively. These results indicate that in contrast to MAP 2, which is specifically enriched in dendrites and perikarya of neurons, MAP 1A is widely distributed in the nervous system.     

Proteolysis of tau by calpain

The calpain-induced proteolysis of tau associated with twice-cycled microtubules or from a total brain heat-stable fraction was studied. Twice-cycled microtubule tau was rapidly hydrolyzed by calpain. In contrast, tau purified from the total brain heat-stable fraction was very resistant to degradation by calpain. These results clearly demonstrate that there are at least 2 populations of tau in the brain based on calpain-sensitivity, a calpain-sensitive form that is associated with microtubules and a calpain-resistant form that may represent another population of tau in the brain.     

A Novel Approach for Studying the Physiology and Pathophysiology of Myelinated and Non-Myelinated Axons in the CNS White Matter

Advances in brain connectomics set the need for detailed knowledge of functional properties of myelinated and non-myelinated (if present) axons in specific white matter pathways. The corpus callosum (CC), a major white matter structure interconnecting brain hemispheres, is extensively used for studying CNS axonal function. Unlike another widely used CNS white matter preparation, the optic nerve where all axons are myelinated, the CC contains also a large population of non-myelinated axons, making it particularly useful for studying both types of axons. Electrophysiological studies of optic nerve use suction electrodes on nerve ends to stimulate and record compound action potentials (CAPs) that adequately represent its axonal population, whereas CC studies use microelectrodes (MEs), recording from a limited area within the CC. Here we introduce a novel robust isolated "whole" CC preparation comparable to optic nerve. Unlike ME recordings where the CC CAP peaks representing myelinated and non-myelinated axons vary broadly in size, "whole" CC CAPs show stable reproducible ratios of these two main peaks, and also reveal a third peak, suggesting a distinct group of smaller caliber non-myelinated axons. We provide detailed characterization of "whole" CC CAPs and conduction velocities of myelinated and non-myelinated axons along the rostro-caudal axis of CC body and show advantages of this preparation for comparing axonal function in wild type and dysmyelinated shiverer mice, studying the effects of temperature dependence, bath-applied drugs and ischemia modeled by oxygen-glucose deprivation. Due to the isolation from gray matter, our approach allows for studying CC axonal function without possible "contamination" by reverberating signals from gray matter. Our analysis of "whole" CC CAPs revealed higher complexity of myelinated and non-myelinated axonal populations, not noticed earlier. This preparation may have a broad range of applications as a robust model for studying myelinated and non-myelinated axons of the CNS in various experimental models.     

Differential effect of three-repeat and four-repeat tau on mitochondrial axonal transport

Tau protein is present in six different splice forms in the human brain and interacts with microtubules via either 3 or 4 microtubule binding repeats. An increased ratio of 3 repeat to 4 repeat isoforms is associated with neurodegeneration in inherited forms of frontotemporal dementia. Tau over-expression diminishes axonal transport in several systems, but differential effects of 3 repeat and 4 repeat isoforms have not been studied. We examined the effects of tau on mitochondrial transport and found that both 3 repeat and 4 repeat tau change normal mitochondrial distribution within the cell body and reduce mitochondrial localization to axons; 4 repeat tau has a greater effect than 3 repeat tau. Further, we observed that the 3 repeat and 4 repeat tau cause different alterations in retrograde and anterograde transport dynamics with 3 repeat tau having a slightly stronger effect on axon transport dynamics. Our results indicate that tau-induced changes in axonal transport may be an underlying theme in neurodegenerative diseases associated with isoform specific changes in tau's interaction with microtubules.     

Differential Association of Tau With Subsets of Microtubules Containing Posttranslationally-Modified Tubulin Variants in Neuroblastoma Cells

Neuronal cells display different subsets of dynamic microtubules. In axons and extending neurites, this intrinsic dynamics is modulated by the microtubule-associated protein tau. Moreover, posttranslational modifications of tubulin, namely acetylation, tyrosination or glutamylation are directly involved in determining the stability of neuronal microtubules. Studies were carried out to analyze the interaction patterns of tau with subsets of microtubules in N2A neuroblastoma cells, which can differentiate in the presence of dibutyryl cAMP. Double labeling studies showed a differential pattern of tau association with microtubules containing acetylated and tyrosinated tubulin. Furthermore, studies using depolymerizing drugs revealed a selectivity in the association of tau with microtubular polymers and microfilaments, within the organization of the neuronal cytoskeleton. In order to study the association of specific tau isoforms with microtubules containing modified tubulin variants, immunoprecipitation studies were carried out. The coimmunoprecipitation data indicated a selective binding of specific tau isoforms to either modified tubulin variant. To assess the hypothesis on the roles of tau isoforms in the stabilization of microtubules containing modified tubulins, the association of those variants with tau isoforms was analyzed in overlay experiments. A preferential binding of acetylated tubulin from undifferentiated N2A cell extracts, to at least one slow-migrating tau isoform was revealed. However, acetylated tubulin from N2A cells containing long neurites displayed a preferential association with two isoforms of tau. On the other hand, tyrosinated tubulin from N2A extracts bound to the entire set of neuronal tau isoforms. These studies, along with the tau association with microtubules with different stability, indicate that tau segregates into subsets of microtubules in the axonal processes. The studies also suggest that these interactions may respond to a functional versatility of these polymers in differentiating neurons.     

Viscoelasticity of Tau Proteins Leads to Strain Rate-Dependent Breaking of Microtubules during Axonal Stretch Injury: Predictions from a Mathematical Model

The unique viscoelastic nature of axons is thought to underlie selective vulnerability to damage during traumatic brain injury. In particular, dynamic loading of axons has been shown to mechanically break microtubules at the time of injury. However, the mechanism of this rate-dependent response has remained elusive. Here, we present a microstructural model of the axonal cytoskeleton to quantitatively elucidate the interaction between microtubules and tau proteins under mechanical loading. Mirroring the axon ultrastructure, the microtubules were arranged in staggered arrays, cross-linked by tau proteins. We found that the viscoelastic behavior specifically of tau proteins leads to mechanical breaking of microtubules at high strain rates, whereas extension of tau allows for reversible sliding of microtubules without any damage at small strain rates. Based on the stiffness and viscosity of tau proteins inferred from single-molecule force spectroscopy studies, we predict the critical strain rate for microtubule breaking to be in the range 22–44 s⁻¹, in excellent agreement with recent experiments on dynamic loading of micropatterned neuronal cultures. We also identified a characteristic length scale for load transfer that depends on microstructural properties and have derived a phase diagram in the parameter space spanned by loading rate and microtubule length that demarcates those regions where axons can be loaded and unloaded reversibly and those where axons are injured due to breaking of the microtubules.     

Viscoelasticity of Tau Proteins Leads to Strain Rate-Dependent Breaking of?Microtubules during Axonal Stretch Injury: Predictions from a Mathematical Model

The unique viscoelastic nature of axons is thought to underlie selective vulnerability to damage during traumatic brain injury. In particular, dynamic loading of axons has been shown to mechanically break microtubules at the time of injury. However, the mechanism of this rate-dependent response has remained elusive. Here, we present a microstructural model of the axonal cytoskeleton to quantitatively elucidate the interaction between microtubules and tau proteins under mechanical loading. Mirroring the axon ultrastructure, the microtubules were arranged in staggered arrays, cross-linked by tau proteins. We found that the viscoelastic behavior specifically of tau proteins leads to mechanical breaking of microtubules at high strain rates, whereas extension of tau allows for reversible sliding of microtubules without any damage at small strain rates. Based on the stiffness and viscosity of tau proteins inferred from single-molecule force spectroscopy studies, we predict the critical strain rate for microtubule breaking to be in the range 22–44 s⁻¹, in excellent agreement with recent experiments on dynamic loading of micropatterned neuronal cultures. We also identified a characteristic length scale for load transfer that depends on microstructural properties and have derived a phase diagram in the parameter space spanned by loading rate and microtubule length that demarcates those regions where axons can be loaded and unloaded reversibly and those where axons are injured due to breaking of the microtubules.     

Immunochemical and biochemical characterization of tau proteins in normal and Alzheimer's disease brains with Alz 50 and Tau-1

Microtubule-associated protein tau was characterized in 5 Alzheimer and 5 control brains using two monoclonal antibodies, Alz 50 and Tau-1. Quantitative analysis of immunoblots with the antibodies showed that both homogenate and supernatant fractions (12,000 x g) from Alzheimer brains contained 38-65% less tau immunoreactivity compared to normal brains. The reduction was found in all brain regions studied (frontal and temporal lobes and thalamus) and in both gray and white matter. In partially purified tau preparations, the yield of protein was lower in Alzheimer (by 35%) than in control brain. Incubation of brain proteins, transferred onto nitrocellulose paper, with alkaline phosphatase had either no effect or slightly increased the antibody binding to tau proteins from both brain tissues. Immunoblots of tau-enriched preparations subjected to two-dimensional gel electrophoresis showed no major changes in the staining pattern of tau isoforms in Alzheimer samples except for a weaker reactivity of the basic isovariants as compared to non-Alzheimer samples. The elution volume of tau from Alzheimer brain supernatant on a Sepharose CL-6B column was similar to that from non-Alzheimer brain and equal to that of aldolase (Mr = 158,000). Our data suggest that most of tau proteins from both types of brain have similar biochemical properties. The reduction in tau reactivity in Alzheimer tissue may be due to a reduction in neuronal cell population or incorporation of soluble tau into stable structures such as neurofibrillary tangles, since the tangles have been shown to react with anti-tau antibodies.     

Studies of the immunohistochemical and biochemical localization of the cytochrome P-450scc-linked monooxygenase system in the adult rat brain

Immunohistochemical and biochemical studies were performed on the brains of adult female and male rats using a specific antibody against bovine adrenocortical cytochrome P-450scc. The results showed that in both male and female rats, the myelinated regions of the white matter are selectively immunostained throughout the brain and that even in rats pretreated with colchicine, there is never positive staining of neuronal cell bodies and their dendrites in any brain region. Western immunoblotting with the P-450scc antibody and enzymatic assays revealed that P-450scc and cholesterol side-chain cleavage activity were present in a homogenate derived from the cortical white matter, but not detectable in that from the cerebral cortex. Furthermore, quantitation of the P-450scc protein in the immunoblots indicated that the concentration of P-450scc in the cortical white matter of both female and male rat brains is approx. 3-4 pmol per mg tissue protein. Thus it could be concluded that in the adult rat brain, P-450scc and cholesterol side-chain cleavage activity are selectively localized only in the myelinated region of the white matter.     

Preferential binding of a kinesin-1 motor to GTP-tubulin�Crich microtubules underlies polarized vesicle transport

Polarized transport in neurons is fundamental for the formation of neuronal circuitry. A motor domain–containing truncated KIF5 (a kinesin-1) recognizes axonal microtubules, which are enriched in EB1 binding sites, and selectively accumulates at the tips of axons. However, it remains unknown what cue KIF5 recognizes to result in this selective accumulation. We found that axonal microtubules were preferentially stained by the anti–GTP-tubulin antibody hMB11. Super-resolution microscopy combined with EM immunocytochemistry revealed that hMB11 was localized at KIF5 attachment sites. In addition, EB1, which binds preferentially to guanylyl-methylene-diphosphate (GMPCPP) microtubules in vitro, recognized hMB11 binding sites on axonal microtubules. Further, expression of hMB11 antibody in neurons disrupted the selective accumulation of truncated KIF5 in the axon tips. In vitro studies revealed approximately threefold stronger binding of KIF5 motor head to GMPCPP microtubules than to GDP microtubules. Collectively, these data suggest that the abundance of GTP-tubulin in axonal microtubules may underlie selective KIF5 localization and polarized axonal vesicular transport.     

Subcellular compartmentalization of phosphorylated neurofilament polypeptides in neurons

Immunocytochemistry and polyacrylamide gel electrophoresis have been used to study the distribution of phosphorylated forms of neurofilament antigens in rat brain. Immunostaining of tissue with an antisera produced against phosphatase-sensitive domains of the 200-kilodalton (kd) neurofilament polypeptide showed that phosphorylated forms of this polypeptide were present in virtually all axons and certain somata and dendrites of neurons in different brain regions. Immunoblots of whole brain homogenate or a neurofilament preparation from rat revealed that the affinity-purified anti-200-kd sera used to immunostain tissue labeled the neurofilament-associated 200-kd band in a phosphatase-sensitive manner. Fine structural analysis of this immunoreactivity in tissue showed that whenever the labeled organelle could be identified, it was a microtubule. In contrast, immunoblot analysis of twice-cycled microtubules from porcine brain revealed that microtubules in vitro did not possess the 200-kd antigen that was observed in situ. The results suggest that our antibody recognizes a phosphorylated domain on the neurofilament involved in cross-linking neurofilaments and microtubules, and that in vivo, phosphorylated epitopes of the 200-kd neurofilament polypeptide are capable of associating with microtubules.     

Selective stabilization of tau in axons and microtubule-associated protein 2C in cell bodies and dendrites contributes to polarized localization of cytoskeletal proteins in mature neurons

In mature neurons, tau is abundant in axons, whereas microtubule- associated protein 2 (MAP2) and MAP2C are specifically localized in dendrites. Known mechanisms involved in the compartmentalization of these cytoskeletal proteins include the differential localization of mRNA (MAP2 mRNA in dendrites, MAP2C mRNA in cell body, and Tau mRNA in proximal axon revealed by in situ hybridization) (Garner, C.C., R.P. Tucker, and A. Matus. 1988. Nature (Lond.). 336:674-677; Litman, P., J. Barg, L. Rindzooski, and I. Ginzburg. 1993. Neuron. 10:627-638), suppressed transit of MAP2 into axons (revealed by cDNA transfection into neurons) (Kanai, Y., and N. Hirokawa. 1995. Neuron. 14:421-432), and differential turnover of MAP2 in axons vs dendrites (Okabe, S., and N. Hirokawa. 1989. Proc. Natl. Acad. Sci. USA. 86:4127-4131). To investigate whether differential turnover of MAPs contributes to localization of other major MAPs in general, we microinjected biotinylated tau, MAP2C, or MAP2 into mature spinal cord neurons in culture (approximately 3 wk) and then analyzed their fates by antibiotin immunocytochemistry. Initially, each was detected in axons and dendrites, although tau persisted only in axons, whereas MAP2C and MAP2 were restricted to cell bodies and dendrites. Injected MAP2C and MAP2 bound to dendritic microtubules more firmly than to microtubules in axons, while injected tau bound to axonal microtubules more firmly than to microtubules in dendrites. Thus, beyond contributions from mRNA localization and selective axonal transport, compartmentalization of each of the three major MAPs occurs through local differential turnover.     

A taxol-dependent procedure for the isolation of microtubules and microtubule-associated proteins (MAPs)

The effect of the antimitotic drug taxol on the association of MAPs (microtubule-associated proteins) with microtubules was investigated. Extensive microtubule assembly occurred in the presence of Taxol at 37 degrees C. at 0 degrees C, and at 37 degrees C in the presence of 0.35 M NaCl, overcoming the inhibition of assembly normally observed under the latter two conditions. At 37 degrees C and at 0 degrees C, complete assembly of both tubulin and the MAPs was observed in the presence of Taxol. However, at elevated ionic strength, only tubulin assembled, forming microtubules devoid of MAPs. The MAPs could also be released from the surface of preformed microtubules by exposure to elevated ionic strength. These properties provided the basis for a rapid new procedure for isolating microtubules and MAPs of high purity from small amounts of biological material. The MAPs could be recovered by exposure of the microtubules to elevated ionic strength and subjected to further analysis. Microtubules and MAPs were prepared from bovine cerebral cortex (gray matter) and from HeLa cells. MAP 1, MAP2, and the tau MAPs, as well as species of Mr = 28,000 and 30,000 (LMW, or low molecular weight, MAPs) and a species of Mr = 70,000 were isolated from gray matter. Species identified as the 210,000 and 125,000 mol wt HeLa MAPs were isolated from HeLa cells. Microtubules were also prepared for the first time from white matter. All of the MAPs identified in gray matter preparations were identified in white matter, but the amounts of individual MAP species differed. The most striking difference in the two preparations was a fivefold lower level of MAP 2 relative to tubulin in white matter than in gray. The high molecular weigh MAP, MAP1, was present in equal ratio to tubulin in white and gray matter. These results indicate that MAP 1 and MAP2, as well as other MAP species, may have a different cellular or subcellular distribution.     

Beta-secretase-cleaved amyloid precursor protein in Alzheimer brain: a morphologic study

β-amyloid (Aβ) is the main constituent of senile plaques seen in Alzheimer's disease. Aβ is derived from the amyloid precursor protein (APP) via proteolytic cleavage by proteases β- and β-secretase. In this study, we examined content and localization of β-secretase-cleaved APP (β-sAPP) in brain tissue sections from the frontal, temporal and occipital lobe. Strong granular β-sAPP staining was found throughout the gray matter of all three areas, while white matter staining was considerably weaker. β-sAPP was found to be localized in astrocytes and in axons. We found the β-sAPP immunostaining to be stronger and more extensive in gray matter in Alzheimer disease (AD) cases than controls. The axonal β-sAPP staining was patchy and unevenly distributed for the AD cases, indicating impaired axonal transport. β-sAPP was also found surrounding senile plaques and cerebral blood vessels. The results presented here show altered β-sAPP staining in the AD brain, suggestive of abnormal processing and transport of APP.     

The GDVII Strain of Theiler’s Virus Spreads via Axonal Transport

Following intracerebral inoculation, the DA strain of Theiler's virus sequentially infects neurons in the gray matter and glial cells in the white matter of the spinal cord. It persists in the latter throughout the life of the animal. Several observations suggest that the virus spreads from the gray to the white matter by axonal transport. In contrast, the neurovirulent GDVII strain causes a fatal encephalitis with lytic infection of neurons. It does not infect the white matter of the spinal cord efficiently and does not persist in survivors. The inability of this virus to infect the white matter could be due to a defect in axonal transport. Using footpad inoculations, we showed that the GDVII strain is, in fact, transported in axons. Transport was prevented by sectioning the sciatic nerve. The kinetics of transport and experiments using colchicine suggested that the virus uses microtubule-associated fast axonal transport. Our results show that a cardiovirus can spread by fast axonal transport and suggest that the inability of the GDVII strain to infect the white matter is not due to a defect in axonal transport.     

Tau protein expression in adult bovine oligodendrocytes: functional and pathological significance

In tauopathies, overexpression of tau exon 10 is linked to degeneration and abnormal tau deposition in neurons and oligodendroglia (OLGs). To compare exon 10 expression in normal neurons and OLGs, adult bovine brain was examined for the expression of tau in gray matter and cultured OLGs isolated from white matter. Using exon-specific antibodies, we found that both types of tissues abundantly expressed exon 2 but isolated OLGs had a lower expression of exons 3 and 10 when compared to gray matter. Relative expression of exons 3 and 10 did not change significantly during the in vitro maturation of OLGs for 39 days. Using a panel of well-characterized antibodies against tau, we determined that isolated OLGs contained tau phosphorylated at the Tau-1, 12E8, and PHF-1 but not the AT8, AT100, AT180, and AT270 epitopes. Tau phosphorylation status diminished during in vitro maturation, suggesting that healthy OLG processes require regulated phosphorylation of tau at specific sites. We propose that the tau isoform profile and phosphorylation status contribute to the vulnerability of OLGs in degenerative diseases linked to overexpression of exon 10.     

The microtubule-severing proteins spastin and katanin participate differently in the formation of axonal branches

Neurons express two different microtubule-severing proteins, namely P60-katanin and spastin. Here, we performed studies on cultured neurons to ascertain whether these two proteins participate differently in axonal branch formation. P60-katanin is more highly expressed in the neuron, but spastin is more concentrated at sites of branch formation. Overexpression of spastin dramatically enhances the formation of branches, whereas overexpression of P60-katanin does not. The excess spastin results in large numbers of short microtubules, whereas the excess P60-katanin results in short microtubules intermingled with longer microtubules. We hypothesized that these different microtubule-severing patterns may be due to the presence of molecules such as tau on the microtubules that more strongly shield them from being severed by P60-katanin than by spastin. Consistent with this hypothesis, we found that axons depleted of tau show a greater propensity to branch, and that this is true whether or not the axons are also depleted of spastin. We propose that there are two modes by which microtubule severing is orchestrated during axonal branch formation, one based on the local concentration of spastin at branch sites and the other based on local detachment from microtubules of molecules such as tau that regulate the severing properties of P60-katanin.     

alpha-synuclein binds to Tau and stimulates the protein kinase A-catalyzed tau phosphorylation of serine residues 262 and 356.

alpha-Synuclein has been implicated in the pathogenesis of several neurodegenerative disorders based on the direct linking of missense mutations in alpha-synuclein to autosomal dominant Parkinson's disease and its presence in Lewy-like lesions. To gain insight into alpha-synuclein functions, we have investigated whether it binds neuronal proteins and modulates their functional state. The microtubule-associated protein tau was identified as a ligand by alpha-synuclein affinity chromatography of human brain cytosol. Direct binding assays using (125)I-labeled human tau40 demonstrated a reversible binding with a IC(50) about 50 pM. The interacting domains were localized to the C terminus of alpha-synuclein and the microtubule binding region of tau as determined by protein fragmentation and the use of recombinant peptides. High concentrations of tubulin inhibited the binding between tau and alpha-synuclein. Functionally, alpha-synuclein stimulated the protein kinase A-catalyzed phosphorylation of tau serine residues 262 and 356 as determined using a phospho-epitope-specific antibody. We propose that alpha-synuclein modulates the phosphorylation of soluble axonal tau and thereby indirectly affects the stability of axonal microtubules.     

Axonal tubulin and axonal microtubules: biochemical evidence for cold stability

Nerve extracts containing tubulin labeled by axonal transport were analyzed by electrophoresis and differential extraction. We found that a substantial fraction of the tubulin in the axons of the retinal ganglion cell of guinea pigs is not solubilized by conventional methods for preparation of microtubules from whole brain. In two-dimensional polyacrylamide gel electrophoresis this cold-insoluble tubulin was biochemically distinct from tubulin obtained from whole brain microtubules prepared by cold cycling. Cleveland peptide maps also indicated some differences between the cold-extractable and cold- insoluble tubulins. The demonstration of cold-insoluble tubulin that is specifically axonal in origin permits consideration of the physiological role of cold-insoluble tubulin in a specific cellular structure. It appears likely that much of this material is in the form of cold-stable microtubules. We propose that the physiological role of cold-insoluble tubulin in the axon may be associated with the regulation of the axonal microtubule complexes in neurons.