GFP illuminates the cytoskeleton.

Until recently, cytoskeleton research has relied primarily on immunofluorescence microscopy techniques, requiring fixation and hence killing of the specimen before the analysis. The sole method for visualizing cytoskeletal dynamics in living cells has been the microinjection of purified and fluorescently labelled protein, but technical difficulties have precluded its widespread use. The recent introduction of green fluorescent protein (GFP) has enabled visualization of proteins and cytoskeletal dynamics with only minimal perturbations of the living cell and has opened new horizons for studying the cytoskeleton.

PC12 cells express juvenile microtubule-associated proteins during nerve growth factor-induced neurite outgrowth.

Microtubule-associated proteins (MAPs) are believed to play an important role in regulating the growth of neuronal processes. The nerve growth factor-induced differentiation of PC12 pheochromocytoma cells is a widely used tissue culture model for studying this mechanism. We have found that contrary to previous suggestions, the major MAPs of adult brain, MAP1 and MAP2, are minor components of PC12 cells. Instead two novel MAPs characteristic of developing brain, MAP3 and MAP5, are present and increase more than 10-fold after nerve growth factor treatment; the timing of these increases coinciding with the bundling of microtubules and neurite outgrowth. Immunocytochemical staining showed that MAP3 and MAP5 are initially distributed throughout the cytoplasm. Subsequently MAP5 becomes associated with microtubules in both neurites and growth cones but MAP3 distribution remained diffuse. Thus MAP3 and MAP5, which are characteristic of developing neurons in the juvenile brain, are also induced in PC12 cells during neurite outgrowth in culture. In contrast MAP1, which is characteristic of mature neurons, does not increase during PC12 cell differentiation. These results provide evidence that one set of MAPs is expressed during neurite outgrowth and a different set during the maintenance of neuronal form. It also appears that the PC12 system is an appropriate model for studying the active neurite growth phase of neuronal differentiation but not for neuronal maturation.

Different forms of microtubule-associated protein 2 are encoded by separate mRNA transcripts.

Brain microtubule-associated protein 2 (MAP2) consists of a pair of high molecular mass (280 kD) polypeptides, MAP2a and MAP2b, and a recently identified 70-kD protein, MAP2c, which is antigenically related to these high molecular mass MAP2's. Using cDNA clones we have analyzed the expression of these three proteins at the nucleic acid level. cDNA probes selective for the high molecular mass MAP2's a and b identified only a 9-kb mRNA, whereas a probe for sequence common to all three MAP2 isoforms, a, b, and c, recognized the 9-kb transcript and additionally a 6-kb mRNA. Southern blot analysis with cDNA probes indicated that there is only one MAP2 gene from which these two distinct mRNAs are derived. The 70-kD MAP2c protein is much more abundant in neurons of developing brain than those of adult tissues. Similarly the expression of the 6-kb MAP2c-related mRNA, is much greater in neonatal than adult rat brain, indicating that the developmental expression of MAP2 is determined by transcriptional regulation from a single MAP2 gene.

Initial phase of dendrite growth: evidence for the involvement of high molecular weight microtubule-associated proteins (HMWP) before the appearance of tubulin.

It has recently been shown that high molecular weight microtubule-associated proteins (HMWP) in the brain are present in dendrites and are absent from axons (Matus et al., 1981, Proc. Natl. Acad. Sci. U. S. A. 78:3010-3014). In this study we followed the appearance of both HMWP and tubulin in the neonatal rat cerebellum by immunoperoxidase staining, concentrating particularly on comparing Purkinje cell dendrites with adjacent granule cell axons. In the axons both immunohistochemically demonstrable tubulin and structurally distinct microtubules are present at all stages of development. By contrast the Purkinje cell dendrites contain better neither tubulin nor microtubules at early stages of their growth. However, immunoperoxidase staining showed that these developing dendrites are rich in HMWP which are particularly concentrated in the dendritic distal regions. HMWP are also present as patches beneath the surface membrane of the cell body before the emergence of dendrites. Based on this data and the well-documented ability of HMWP to promote microtubule assembly, we propose the hypothesis that during the initial phase of Purkinje neuron differentiation HMWP form part of a specialized cytoskeletal structure which acts as a specifier for the development of dendrites as opposed to axons.

Phosphorylation determines the binding of microtubule-associated protein 2 (MAP2) to microtubules in living cells.

The influence of phosphorylation on the binding of microtubule-associated protein 2 (MAP2) to cellular microtubules was studied by microinjecting MAP2 in various phosphorylation states into rat-1 fibroblasts, which lack endogenous MAP2. Conventionally prepared brain MAP2, containing 10 mol of endogenous phosphate per mol (MAP2-P10), was completely bound to cellular microtubules within 2-3 min after injection. MAP2 prepared in the presence of phosphatase inhibitors, containing 25 mol/mol of phosphate (MAP2-P25), also bound completely. However, MAP2 whose phosphate content had been reduced to 2 mol phosphate per mol by treatment with alkaline phosphatase in vitro (MAP2-P2) did not initially bind to microtubules, suggesting that phosphorylation of certain sites in MAP2 is essential for binding to microtubules. MAP2-P10 was further phosphorylated in vitro via an endogenously bound protein kinase activity, adding 12 more phosphates, giving a total of 22 mol/mol. This preparation (MAP2-P10+12) also did not bind to microtubules. Assay of the binding of these preparations to taxol-stabilized tubulin polymers in vitro confirmed that their binding to tubulin depended on the state of phosphorylation, but the results obtained in microinjection experiments differed in some cases from in vitro binding. The results suggest that the site of phosphate incorporation rather than the amount is the critical factor in determining microtubule binding activity of MAP2. Furthermore, the interaction of MAP2 with cellular microtubules may be influenced by additional factors that are not evident in vitro.

Immunocytochemical localization of microtubule-associated protein 1 in rat cerebellum using monoclonal antibodies.

Immunohistochemical staining with monoclonal antibodies showed that microtubule-associated protein 1 (MAP1) has a restricted cellular distribution in the rat cerebellum. Anti-MAP1 staining was found only in neurons, where it was much stronger in dendrites than in axons. There were striking variations in the apparent concentration of MAP1 in different classes of neurons. Purkinje cells were the most strongly labeled, while granule cell neurons gave a faint, threshold-level reaction with the antibody. The reaction of Golgi neurons was intermediate between these two extremes. Equivalent results were obtained using two different methods of tissue preparation. Thus MAP1 appears to be a neuron-specific protein that is highly concentrated in dendrites and occurs at markedly different levels in different types of neurons. These observations provide further indications of heterogeneity among brain microtubules.

Ultrastructural localization of cyclic GMP and cyclic AMP in rat striatum.

The subcellular localization of cyclic GMP and cyclic AMP in the rat caudate-putamen has been studied using horseradish peroxidase immunocytochemistry. Both of the putative neurotransmitter second messengers were visualized in neurons and glial cells at light microscopic resolutions, but not all cells of either category gave detectable staining. This was confirmed at the ultrastructural level where both stained and unstained elements of the same cell type were found within the same field. A striking variation was seen in cyclic nucleotide staining intensity within individual neural and glial cells. Both of the cyclic nucleotides were detected within postsynaptic terminal boutons and within astroglial processes. Cyclic GMP postsynaptic staining was stronger than glial staining, whereas the localization pattern was reversed for cyclic AMP. The synaptic localization of cyclic AMP and cyclic GMP immunoreactivity adds support to the idea that these compounds have an influential role in synaptic function within the striatum.

Brain postsynaptic densities: the relationship to glial and neuronal filaments.

Preparations of isolated brain postsynaptic densities (PSDs) contain a characteristic set of proteins among which the most prominent has a molecular weight of approximately 50,000. Following the suggestion that this major PSD protein might be related to a similarly sized component of neurofilaments (F. Blomberg et al., 1977, J. Cell Biol., 74:214-225), we searched for evidence of neurofilament proteins among the PSD polypeptides. This was done with a novel technique for detecting protein antigens in SDS-polyacrylamide gels (immunoblotting) and an antiserum that was selective for neurofilaments in immunohistochemical tests. As a control, an antiserum against glial filament protein (GFAP) was used because antisera against GFAP stain only glial cells in immunohistochemical tests. They would, therefore, not be expected to react with PSDs that occur only in neurons. The results of these experiments suggested that PSDs contain both neuronal and also glial filament proteins at higher concentrations than either synaptic plasma membranes, myelin, or myelinated axons. However, immunoperoxidase staining of histological sections with the same two antisera gave contradictory results, indicating that PSDs in intact brain tissue contain neither neuronal or glial filament proteins. This suggested that the intermediate filament proteins present in isolated PSD preparations were contaminants. To test this possibility, the proteins of isolated brain intermediate filaments were labeled with 125I and added to brain tissue at the start of a subcellular fractionation schedule. The results of this experiment confirmed that both neuronal and glial filament proteins stick selectively to PSDs during the isolation procedure. The stickiness of PSDs for brain cytoplasmic proteins indicates that biochemical analysis of subcellular fractions is insufficient to establish a given protein as a synaptic junctional component. An immunohistochemical localization of PSDs in intact tissue, which has now been achieved for tubulin, phosphoprotein I, and calmodulin, appears to be an essential accessory item of evidence. Our findings also corroborate recent evidence which suggests that isolated preparations of brain intermediate filaments contain both neuronal and glial filaments.

MAP3: characterization of a novel microtubule-associated protein.

Using monoclonal antibodies we have characterized a brain protein that copurifies with microtubules. We identify it as a microtubule-associated protein (MAP) by the following criteria: it copolymerizes with tubulin through repeated cycles of microtubule assembly in vitro; it is not associated with any brain subcellular fraction other than microtubules; in double-label immunofluorescence experiments antibodies against this protein stain the same fibrous elements in cultured cells as are stained by antitubulin; and this fibrous staining pattern is dispersed when cytoplasmic microtubules are disrupted by colchicine. Because it is distinct from previously described MAPs we designate this novel species MAP3. The MAP3 protein consists of a closely spaced pair of polypeptides on SDS gels, Mr 180,000, which are present in both glial (glioma C6) and neuronal (neuroblastoma B104) cell lines. In brain the MAP3 antigen is present in both neurons and glia. In nerve cells its distribution is strikingly restricted: anti-MAP3 staining is detectable only in neurofilament-rich axons. It is not, however, a component of isolated brain intermediate filaments.

Direct observations of the mechanical behaviors of the cytoskeleton in living fibroblasts.

Cytoskeletal proteins tagged with green fluorescent protein were used to directly visualize the mechanical role of the cytoskeleton in determining cell shape. Rat embryo (REF 52) fibroblasts were deformed using glass needles either uncoated for purely physical manipulations, or coated with laminin to induce attachment to the cell surface. Cells responded to uncoated probes in accordance with a three-layer model in which a highly elastic nucleus is surrounded by cytoplasmic microtubules that behave as a jelly-like viscoelastic fluid. The third, outermost cortical layer is an elastic shell under sustained tension. Adhesive, laminin-coated needles caused focal recruitment of actin filaments to the contacted surface region and increased the cortical layer stiffness. This direct visualization of actin recruitment confirms a widely postulated model for mechanical connections between extracellular matrix proteins and the actin cytoskeleton. Cells tethered to laminin-treated needles strongly resisted elongation by actively contracting. Whether using uncoated probes to apply simple deformations or laminin-coated probes to induce surface-to-cytoskeleton interaction we observed that experimentally applied forces produced exclusively local responses by both the actin and microtubule cytoskeleton. This local accomodation and dissipation of force is inconsistent with the proposal that cellular tensegrity determines cell shape.

Surface antigens of brain synapses: identification of minor proteins using polyclonal antisera.

Antigenic proteins of brain synaptic plasma membranes (SPM) and postsynaptic densities (PSD) were characterized using antisera raised against SPM. Immunostaining of brain sections showed that the antigens were restricted to synapses, and electron microscopy revealed staining at both presynaptic terminals and PSDs. In primary brain cell cultures the antisera were also neuron-specific but the antigens were distributed throughout the entire neuronal plasma membrane, suggesting that some restrictive influence present in whole tissue is absent when neurons are grown dispersed. The antigenic proteins with which these antisera react were identified using SDS gel immunoblots. SPM and PSD differed from one another in their characteristic antigenic proteins. Comparison with amido-black stained gel blots showed that in both cases most of these did not correspond to known abundant proteins of SPM or PSDs revealed by conventional biochemical techniques. None of the antigens revealed by the polyclonal antisera were detected by any of a large series of monoclonal antibodies against SPM.

Domains of neuronal microtubule-associated proteins and flexural rigidity of microtubules.

Microtubules are flexible polymers whose mechanical properties are an important factor in the determination of cell architecture and function. It has been proposed that the two most prominent neuronal microtubule-associated proteins (MAPs), tau and MAP2, whose microtubule binding regions are largely homologous, make an important contribution to the formation and maintenance of neuronal processes, putatively by increasing the rigidity of microtubules. Using optical tweezers to manipulate single microtubules, we have measured their flexural rigidity in the presence of various constructs of tau and MAP2c. The results show a three- or fourfold increase of microtubule rigidity in the presence of wild-type tau or MAP2c, respectively. Unexpectedly, even low concentrations of MAPs promote a substantial increase in microtubule rigidity. Thus at approximately 20% saturation with full-length tau, a microtubule exhibits >80% of the rigidity observed at near saturating concentrations. Several different constructs of tau or MAP2 were used to determine the relative contribution of certain subdomains in the microtubule-binding region. All constructs tested increase microtubule rigidity, albeit to different extents. Thus, the repeat domains alone increase microtubule rigidity only marginally, whereas the domains flanking the repeats make a significant contribution. Overall, there is an excellent correlation between the strength of binding of a MAP construct to microtubules (as represented by its dissociation constant Kd) and the increase in microtubule rigidity. These findings demonstrate that neuronal MAPs as well as constructs derived from them increase microtubule rigidity, and that the changes in rigidity observed with different constructs correlate well with other biochemical and physiological parameters.

Actin-based plasticity in dendritic spines.

The central nervous system functions primarily to convert patterns of activity in sensory receptors into patterns of muscle activity that constitute appropriate behavior. At the anatomical level this requires two complementary processes: a set of genetically encoded rules for building the basic network of connections, and a mechanism for subsequently fine tuning these connections on the basis of experience. Identifying the locus and mechanism of these structural changes has long been among neurobiology's major objectives. Evidence has accumulated implicating a particular class of contacts, excitatory synapses made onto dendritic spines, as the sites where connective plasticity occurs. New developments in light microscopy allow changes in spine morphology to be directly visualized in living neurons and suggest that a common mechanism, based on dynamic actin filaments, is involved in both the formation of dendritic spines during development and their structural plasticity at mature synapses.

Dietary cadmium and benzo(a)pyrene increased intestinal metallothionein expression in the fish Fundulus heteroclitus.

Fish were individually fed food pellets containing cadmium, benzo(a)pyrene, or a combination of the two, then analyzed for metallothionein mRNA expression in the intestine, liver, and gill using real-time RT-qPCR. An initial experiment using only cadmium showed that ingestion of pellets varied in individual fish, and estimates of cadmium dose from the numbers of ingested pellets indicated considerable individual variability in cadmium dose. Induction of intestinal metallothionein mRNA was apparent, however, and a linear dose-response relationship was observed for metallothionein expression and cadmium dose in the intestine, but not the other organs, which showed no induction. In a second experiment, the entire daily cadmium dose was provided in a single contaminated pellet that was consumed by all treated fish, effectively eliminating the effect of variable ingestion rates on dose, and the interaction between cadmium and benzo(a)pyrene was also investigated. The intestine was again the primary organ for metallothionein induction by cadmium. When benzo(a)pyrene was administered together with cadmium, induction of metallothionein was potentiated by the presence of benzo(a)pyrene, with the main effect seen in the intestine, where already high levels of induction by cadmium alone increased by 1.74-fold when benzo(a)pyrene was present.

Cytoskeletal plasticity in cells expressing neuronal microtubule-associated proteins.

MAP2 and tau are the two most prominent neuron-specific microtubule-associated proteins. They have been implicated in the stabilization of microtubules and consequently of neurite morphology. To investigate their influence on microtubule dynamics, we have tagged both proteins with green fluorescent protein and expressed them in non-neuronal cells. Time-lapse recordings of living cells showed that MAP2 and tau did not significantly affect the rates of microtubule growth and shrinkage. Longer recordings revealed the growth and disappearance of MAP-induced microtubule bundles coinciding with changes in cell shape. This supports the idea that microtubule dynamics are influenced by the cortical cytoskeleton. The dynamics-preserving stabilization of microtubules by MAP2 and tau thus provides a molecular basis for the morphological plasticity reported to exist in established neurites.

In situ localization of microtubule-associated protein mRNA in the developing and adult rat brain.

We have used cDNA probes specific for three of the major brain microtubule-associated proteins (MAPs), MAP1, MAP2, and MAP5, to study the timing of appearance, relative abundance, and intracellular compartmentalization of MAP gene transcripts in developing rat brain. The MAP1 probe hybridizes throughout the brain, in both grey and white matter. MAP2 mRNA is detected only in grey matter and appears in cerebral neurons only after they have ceased dividing and have migrated to the cortical plate. The MAP5 cDNA hybridizes throughout the embryonic brain, but by P12, MAP5 mRNA distribution is restricted to relatively immature areas. MAP2 mRNA, found in dendrites in the developing brain, persists in some adult dendrites. MAP5 mRNA, like beta-tubulin mRNA, is found only in the cell bodies of developing neurons, indicating that the protein must be transported from the soma into processes. MAP1 mRNA is found only in the proximal regions of cortical pyramidal cell dendrites in both developing and adult brain. The diverse distributions of MAP gene transcripts emphasize the importance of these proteins in generating heterogeneity of microtubule function and indicate that MAP compartmentalization within neurons is regulated in part by differential mRNA transport.

Rapid actin-based plasticity in dendritic spines.

Dendritic spines have been proposed as primary sites of synaptic plasticity in the brain. Consistent with this hypothesis, spines contain high concentrations of actin, suggesting that they might be motile. To investigate this possibility, we made video recordings from hippocampal neurons expressing actin tagged with green fluorescent protein (GFP-actin). This reagent incorporates into actin-containing structures and allows the visualization of actin dynamics in living neurons. In mature neurons, recordings of GFP fluorescence revealed large actin-dependent changes in dendritic spine shape, similar to those inferred from previous studies using fixed tissues. Visible changes occurred within seconds, suggesting that anatomical plasticity at synapses can be extremely rapid. As well as providing a molecular basis for structural plasticity, the presence of motile actin in dendritic spines implicates the postsynaptic element as a primary site of this phenomenon.

Glutamate receptors regulate actin-based plasticity in dendritic spines.

Dendritic spines at excitatory synapses undergo rapid, actin-dependent shape changes which may contribute to plasticity in brain circuits. Here we show that actin dynamics in spines are potently inhibited by activation of either AMPA or NMDA subtype glutamate receptors. Activation of either receptor type inhibited actin-based protrusive activity from the spine head. This blockade of motility caused spines to round up so that spine morphology became both more stable and more regular. Inhibition of spine motility by AMPA receptors was dependent on postsynaptic membrane depolarization and influx of Ca 2+ through voltage-activated channels. In combination with previous studies, our results suggest a two-step process in which spines initially formed in response to NMDA receptor activation are subsequently stabilized by AMPA receptors.

Bundling of microtubules in transfected cells does not involve an autonomous dimerization site on the MAP2 molecule.

We have searched for putative dimerization sites in microtubule-associated protein 2 (MAP2) that may be involved in the bundling of microtubules. An overlapping series of fragments of the embryonic form MAP2c were created and immunologically "tagged" with an 11 amino acid sequence from human c-myc. Nonneuronal cells were transfected simultaneously with one of these myc-tagged fragments and with full-length native MAP2c. Immunolabeling with site-specific antibodies allowed the two transgene products to be located independently within the cytoplasm of a single double-transfected cell. All transfected cells contained bundled microtubules to which the full-length native MAP2 was bound. The distribution of the tagged MAP2 fragment relative to these MAP2-induced bundles was determined by the anti-myc staining. None of the fragments tested, representing all of the MAP2c sequence in overlapping pieces, were associated with MAP2-induced microtubule bundles. These results suggest that MAP2-induced bundle formation in cells does not involve an autonomous dimerization site within the MAP2 sequence.

Stiff microtubules and neuronal morphology.

Neuronal processes contain high concentrations of two related microtubule-associated proteins, MAP2 and tau. When MAP2 is expressed in non-neuronal cells the microtubules appear to be stiffer than those in control cells that do not express MAP2. A stiffening effect of MAP2 is further suggested by recent experiments with microtubules reassembled in vitro and by the fact that, under appropriate circumstances, MAP2-expressing cells can be induced to form processes that are long and cylindrical. Both MAP2 and tau contain homologous microtubule-binding domains, consisting of three or four repeats of an 18 amino acid sequence, which we believe are responsible for the stiffening effect. Our hypothesis is that each repeat binds to a neighbouring tubulin subunit in the wall of the microtubule, tethering them together and reducing their freedom of movement relative to one another. Based on these considerations, we suggest that MAP2 and tau may contribute to the support of neuronal processes by making the microtubules they contain longer, more stable and stiffer than those in non-neuronal cells.