Interaction of tau isoforms with Alzheimer's disease abnormally hyperphosphorylated tau and in vitro phosphorylation into the disease-like protein.

The microtubule-associated protein tau is a family of six isoforms that becomes abnormally hyperphosphorylated and accumulates in neurons undergoing neurodegeneration in the brains of patients with Alzheimer disease (AD). We investigated the isoform-specific interaction of normal tau with AD hyperphosphorylated tau (AD P-tau). We found that the binding of AD P-tau to normal human recombinant tau was tau4L > tau4S > tau4 and tau3L > tau3S > tau3, and that its binding to tau4L was greater than to tau3L. AD P-tau also inhibited the assembly of microtubules promoted by each tau isoform and caused disassembly when added to preassembled microtubules. This inhibition and depolymerization of microtubules by the AD P-tau corresponded directly to the degree of its interaction with the different tau isoforms. In vitro hyperphosphorylation of recombinant tau (P-tau) conferred AD P-tau-like characteristics. Like AD P-tau, P-tau interacted with and sequestered normal tau and inhibited microtubule assembly. These studies suggest that the AD P-tau interacts preferentially with the tau isoforms that have the amino-terminal inserts and four microtubule binding domain repeats and that hyperphosphorylation of tau appears to be sufficient to acquire AD P-tau characteristics. Thus, lack of amino-terminal inserts and extra microtubule binding domain repeat in fetal human brain might be protective from Alzheimer's neurofibrillary degeneration.

Brain plasma membrane Na+,K+-ATPase is inhibited by acetylated tubulin.

Membranes from brain tissue contain tubulin that can be isolated as a hydrophobic compound by partitioning into Triton X-114. The hydrophobic behavior of this tubulin is due to the formation of a complex with the alpha-subunit of Na+,K+-ATPase. In the present work we show that the interaction of tubulin with Na+K+-ATPase inhibits the enzyme activity. We found that the magnitude of the inhibition is correlated with: (1) concentration of the acetylated tubulin isoform present in the tubulin preparation used, and (2) amount of acetylated tubulin isoform isolated as a hydrophobic compound. In addition, some compounds involved in the catalytic action of Na+K+-ATPase were assayed to determine their effects on the inhibitory capability of tubulin on this enzyme. The inhibitory effect of tubulin was only slightly decreased by ATP at relatively low nucleotide concentration (0.06 mM). NaCl (1-160 mM) and KCl (0.2-10 mM) showed no effect whereas inorganic phosphate abolished the inhibitory effect of tubulin in a concentration-dependent manner.

Tubulin carboxypeptidase assay based on the action of the enzyme on [14C]tyrosinated tubulin bound to nitrocellulose membrane.

We have developed a method for the determination of tubulin carboxypeptidase activity which is based on the action of the enzyme on the substrate, [14C]tyrosinated tubulin, previously adsorbed on nitrocellulose membrane. In addition to being two to three times more sensitive than previous carboxypeptidase assays, this method allows the determination of dilute enzyme preparations even containing high salt (inhibitory) concentrations. This is a valuable property specially under circumstances in which numerous high salt-containing fractions with scarce activity should be analyzed (for example after certain chromatographic stages during enzyme purification). Our method is simpler, less time-consuming, and suitable for multiple, simultaneous determinations and the substrate bound to nitrocellulose can be stored for several months without significant alteration of its properties. Peptidases other than tubulin carboxypeptidase can act on [14C]tyrosinated tubulin bound to nitrocellulose, solubilizing radioactive compounds, suggesting the eventual applicability of this method to assay proteases in general. Other features and advantages of the assay as well as its limitations are discussed.

Posttranslational arginylation of soluble rat brain proteins after whole body hyperthermia.

We have previously reported the posttranslational addition of [14C]-arginine in the N-terminus of several soluble rat brain proteins. One of these proteins was identified as the microtubule-associated protein, the stable tubule only polypeptide (STOP). However, despite the fact that the biological significance of arginylation is not completely understood, some evidence associates it with proteolysis via the ubiquitin pathway. Since this degradative via is exacerbated as a response to stress, we studied in vitro the posttranslational [14C]-arginylation of cytosolic brain proteins of rats subjected to hyperthermia in vivo. Immediately after subjecting the animals to hyperthermia, a minor reduction (16%) in the acceptor capacity of [14C]-arginine into proteins was observed in comparison with animals maintained at 28 degrees C. However, in the animals allowed to recover for 3 h, an increase (46%) in the arginylation was observed concomitantly with a significant accumulation of the heat shock protein (70 kDa; hsp 70) when compared to the control animals. These findings suggest that the posttranslational arginylation of proteins participate in the heat shock response. The STOP protein of the soluble brain fraction of control animals, which in Western blot appears as a doublet band (125 and 130 kDa, respectively), is seen, after the hyperthermic treatment, as a single band of 125 kDa. The amount of 125 kDa protein, as well as the in vitro incorporation of [14C]-arginine, increases after hyperthermia in comparison with control animals. Following hyperthermic treatment, we observed a decrease in the amount of in vivo [35S]-methionine-labeled brain proteins. We speculate that, as observed for STOP protein, the increase in the degradation of protein that occurs in hyperthermia, would produce an increase in the amount of arginine acceptor proteins.

Association of tubulin carboxypeptidase with microtubules in living cells.

Tubulin carboxypeptidase is the enzyme that releases the C-terminal tyrosine from alpha-tubulin, converting tyrosine-terminated (Tyr) to detyrosinated (Glu) tubulin. The present study demonstrates that this enzyme is associated with microtubules in living cells. We extracted cultured cells (COS-7) with Triton X-100 under microtubule-stabilizing conditions and found tubulin carboxypeptidase activity in the cytoskeleton fraction. We ruled out, by using several control experiments, the possibility that this result was due to contamination of the isolated cytoskeletons by non-associated proteins contained in the detergent fraction or to an artifact in vitro during the extraction procedure. The associated carboxypeptidase activity showed characteristics similar to those of brain tubulin carboxypeptidase and different from those of pancreatic carboxypeptidase A. In comparison with cultures at confluence, those at low cell density contained small (if any) amounts of carboxypeptidase activity associated with microtubules. In addition, the enzyme was shown to be associated only with cold-labile microtubules. The tubulin carboxypeptidase/microtubule association was also demonstrated in Chinese hamster ovary, NIH 3T3 and PC12 cells. Interestingly, this association was not observed in cultured embryonic brain cells. Our results demonstrate that tubulin carboxypeptidase is indeed associated with microtubules in living cells. Furthermore, the findings that this association occurs with a subset of microtubules and that its magnitude depends on the degree of confluence of the cell culture indicate that it could be part of the mechanism that regulates the tyrosination state of microtubules.

Na+,K+-ATPase was found to be the membrane component responsible for the hydrophobic behavior of the brain membrane tubulin.

We have previously described that the tubulin isolated from brain membranes as a hydrophobic compound by partitioning into Triton X-114 is a peripheral membrane protein [corrected]. The hydrophobic behavior of this tubulin is due to its interaction with membrane protein(s) and the interaction occurs principally with the acetylated tubulin isotype. In the present work we identified the membrane protein that interacts with tubulin as the Na+,K+-ATPase alpha subunit by amino acid sequencing. Using purified brain Na+,K+-ATPase we were able to isolate part of the total hydrophilic tubulin as a hydrophobic compound which contains a high proportion of the acetylated tubulin isotype.

The association of tubulin carboxypeptidase activity with microtubules in brain extracts is modulated by phosphorylation/dephosphorylation processes.

Tubulin carboxypeptidase, the enzyme which releases the COOH terminal tyrosine from the alpha-chain of tubulin, remains associated with microtubules through several cycles of assembly/disassembly (Arce CA, Barra HS: FEBS Lett 157: 75-78, 1983). Here, we present evidence indicating that in rat brain extract the carboxypeptidase/microtubules association is regulated by the relative activities of endogenous protein kinase(s) and phosphatase(s) which seem to determine the phosphorylation state of the enzyme (or another entity) and in some way the affinity of the enzyme for microtubules. The presence of 2.5 mM ATP during the in vitro microtubule formation resulted in a low recovery of carboxypeptidase activity in the microtubule fraction. This ATP-induced effect was not due to alteration of the enzyme activity or to inhibition of microtubule assembly but to a decrease of the association of the enzyme with microtubules. We found that the ATP-induced effect was not mediated by modifications on the microtubules but, presumably, on the enzyme molecule. The non-hydrolyzable ATP analogue, AMP-PCP, did not reproduce the effect of ATP. The inclusion of phosphatase inhibitors in the homogenization buffer also led to a decrease in the amount of tubulin carboxypeptidase associated with microtubules. Finally, we found that, in concordance with the mechanism hypothesized, the magnitude of the carboxypeptidase/microtubule association correlated well with the different incubation conditions created to favor maximal, minimal or intermediate protein phosphorylation states.

Conversion of hydrophilic tubulin into a hydrophobic compound. Evidence for the involvement of membrane proteins.

Brain membranes contain tubulin that can be isolated as a hydrophobic compound by partitioning into Triton X-114. We have previously postulated: (a) that this kind of tubulin is a peripheral membrane protein that arises from microtubules that in vivo interact with membranes and (b) that the hydrophobic behaviour is due to the interaction of tubulin with a membrane component. Here we report the in vitro conversion of hydrophilic into hydrophobic tubulin by incubating microtubule associated proteins (MAPs) free taxol-stabilized microtubules with Triton X-100 solubilized membranes. After incubation, the microtubules were sedimented, depolymerized and subjected to partition into Triton X-114. Part of the tubulin was isolated in the detergent phase and contained, as observed in native membranes, a high proportion of the acetylated isotype. Because of the high proportion of acetylated tubulin the 'in vitro' conversion resembles the 'in vivo' interaction. Electrophoretic analysis of the detergent phase shows, besides tubulin, two major protein bands of 29 and 100 kDa molecular mass. The ability of the solubilized membranes to convert hydrophilic into hydrophobic tubulin is greatly diminished if the solubilized membrane preparation is preincubated in the presence of trypsin or heated at 90 degrees C for 5 min, indicating that the membrane component that confers the hydrophobic behaviour to tubulin is of proteinaceous nature.

A brain protein (P30) that immunoreacts with a polyclonal anti-pancreatic carboxypeptidase A antibody shows properties that are shared with tubulin carboxypeptidase.

A preparation of tubulin carboxypeptidase partially purified from bovine brain was found to contain a protein of molecular mass 30 kDa (P30) as determined by SDS-PAGE, that is recognized by a polyclonal anti-bovine pancreatic carboxypeptidase A. However, this protein is different from pancreatic carboxypeptidase A as judged by the isoelectric point and the pattern of peptides produced by trypsin digestion. The isoelectric point of P30 was similar to that found for tubulin carboxypeptidase (9 +/- 0.2). When the tubulin carboxypeptidase preparation was subjected to gel filtration chromatography under low salt concentration, P30 behaved as a protein of molecular mass 38 kDa whereas tubulin carboxypeptidase eluted at a position of 75 kDa molecular mass. However, when the chromatography was performed at relatively high salt concentration they behaved as proteins of 49 and 56 kDa, respectively. We considered that P30 may be an inactive monomeric form of the dimeric tubulin carboxypeptidase. However we can not rule out the possibility that it represents another carboxypeptidase not yet described.

Tubulin tyrosine ligase: protein and mRNA expression in developing rat skeletal muscle.

Alpha tubulin can be post-translationally tyrosinated at the carboxy-terminus by a specific enzyme: tubulin tyrosine ligase. The expression of tubulin tyrosine ligase mRNA and protein during the development of rat skeletal muscle was examined in the present study. A portion of the coding region of the rat ligase cDNA was isolated and sequenced. The nucleotide and amino acid sequences showed about 90% homology with previously reported porcine and bovine ligase sequences. In newborn rats, ligase mRNA and protein were highly expressed in skeletal muscle. During early postnatal development, however, both ligase mRNA and protein dropped down dramatically. Quantitative measurements revealed that ligase protein at postnatal day 20 represented only 10% or less of the level at postnatal day 1. Ligase mRNA expression was also examined during the myogenesis in vitro. A strong ligase mRNA signal was detected in both undifferentiated myoblasts and cross-striated, contractile myotubes. The present results suggest that, during muscle differentiation, ligase function may be regulated by the amount of available mRNA. The discrepancy in the ligase expression between the in vivo and in vitro myogenesis suggests that factors controlling the levels of mRNA in vivo are lost in vitro.

Isolation of alpha and beta brain tubulin subunits after alkaline treatment of the protein.

We demonstrate here that brain purified tubulin can be dissociated into alpha and beta subunits at pH > 10 and that the subunits can be separated by using the Triton X-114 phase separation system. After phase partition at pH > 10, alpha tubulin but not beta tubulin behaves as a hydrophobic compound appearing in the detergent rich phase. After three extractions of the alkaline aqueous phase with Triton X-114, about 90% of the alpha tubulin was recovered in the detergent rich phase. The hydrophobic behavior observed for alpha tubulin after its dissociation at pH 11.5 was not due to an irreversible change of the protein, because when the detergent rich phase containing alpha tubulin was diluted with a buffer solution at pH 7.3 and the solution allowed to partition again, alpha-tubulin is recovered in the aqueous phase. The detergent in the aqueous phase of the alpha and beta tubulin preparations can be removed up to 90% by 12 h dialysis. The alpha and beta subunits of tubulin from kidney and liver behave, in this phase separation system, like those of brain tubulin.

Abnormal phosphorylation of tau and the mechanism of Alzheimer neurofibrillary degeneration: sequestration of microtubule-associated proteins 1 and 2 and the disassembly of microtubules by the abnormal tau.

The microtubule-associated protein (MAP) tau is abnormally hyperphosphorylated in Alzheimer disease and accumulates in neurons undergoing neurofibrillary degeneration. In the present study, the associations of the Alzheimer-hyperphosphorylated tau (AD P-tau) with the high molecular weight MAPs (HMW-MAPs) MAP1 and MAP2 were investigated. The AD P-tau was found to aggregate with MAP1 and MAP2 in solution. The association of AD P-tau to the MAPs resulted in inhibition of MAP-promoted microtubule assembly. However, unlike the coaggregation of AD P-tau and normal tau, the association between AD P-tau and the HMW-MAPs did not result in the formation of filaments/tangles. The affinity of the tau-AD P-tau association was higher than that of HMW-MAPs-AD P-tau because normal tau inhibited the latter binding. The association between AD P-tau and the HMW-MAPs also appeared to occur in situ because these proteins cosedimented from the Alzheimer brain extracts, and, in the sediment, the levels of the HMW-MAPs correlated with the levels of AD P-tau. These studies suggested that the abnormally phosphorylated tau can sequester both normal tau and HMW-MAPs and disassemble microtubules but, under physiological conditions, can form tangles of filaments only from tau.

The post-translational incorporation of arginine into a beta-amyloid peptide increases the probability of alpha-helix formation.

The beta-amyloid peptide (beta AP1-40) inhibited the in vitro post-translational incorporation of [14C]arginine at the N-terminus of brain soluble proteins and was labelled by the incorporation of [14C]arginine. Addition of arginine at the N-terminal position of beta AP1-40 is predicted to increase the probability of an alpha-helix structure being formed on the first residues with a higher hydrophilic characteristic, increasing the possibility of these residues being exposed to the aqueous environment. Unmodified beta AP1-40 has a low alpha-helix content and a higher probability of beta-turn formation. Accumulation of beta AP1-40 in Alzheimer's disease may therefore be due to a reduced arginylation reaction and consequently to a decrease in its normal degradation by the ubiquitin pathway.

The relationship of hydrophobic tubulin with membranes in neural tissue.

Brain membrane preparations contain tubulin that can be extracted with Triton X-114. After the extract is allowed to partition, 8% of the total brain tubulin is isolated as a hydrophobic compound in the detergent-rich phase. Cytosolic tubulin does not show this hydrophobic behaviour since it is recovered in the aqueous phase. Membrane tubulin can be released by 0.1 M Na2 CO3 treatment at pH > or = 11.5 in such a way that the hydrophobic tubulin is converted into the hydrophilic form. These results suggest that tubulin exists associated with some membrane component that confers the hydrophobic behaviour to tubulin. If the tissue is homogenized in microtubule-stabilizing buffer containing Triton X-100, the hydrophobic tubulin is isolated from the microtubule fraction. This result indicates that the hydrophobic tubulin isolated from membrane preparations belongs to microtubules that in vivo are associated to membranes. Therefore, hydrophobic tubulin (tubulin-membrane component complex) can be obtained from membranes or from microtubules depending on the conditions of brain homogenization.

Some common properties between a brain protein that is modified by posttranslational arginylation and the microtubule-associated STOP protein.

Properties so far studied of the 125-kDa 14C-arginylated protein from rat brain show remarkable similarities with those of the STOP (stable tubule only polypeptide) protein. On sodium dodecyl sulfate-polyacrylamide gel electrophoresis the 125-kDa 14C-arginylated protein moves to the same position as the STOP protein. The 125-kDa 14C-arginylated protein was immunoprecipitated by the monoclonal Mab 296 antibody specific for neuronal STOP protein. The 125-kDa 14C-arginylated protein was retained by a calmodulin column like STOP protein. As occurs with the STOP protein, the 125-kDa 14C-arginylated protein is found in higher proportion in cold-stable than in cold-labile microtubules. However, the modified protein associates with microtubules in a lower proportion than the STOP protein. We conclude that the STOP protein incorporates arginine by a posttranslational reaction but that only a small fraction of the STOP protein shows acceptor capacity in vitro.

Tyrosinatable and non-tyrosinatable tubulin subpopulations in rat muscle in comparison with those in brain.

Using immunobinding and enzymatic assays we determined in rat muscle extracts the proportion of tyrosinatable tubulin, that is, tubulin that participates in the tyrosination/detyrosination cycle. We found that in muscle, in contrast with nervous tissue, practically all tubulin molecules are tyrosinatable. In the case of rat brain the non-tyrosinatable tubulin pool accounts for about 50% of the tubulin. In addition, isolectrofocusing of 14C-tyrosinated tubulin from brain and muscle extracts revealed a different composition in tyrosinatable tubulin isotypes. One of the isotypes, which in muscle accounts for 86% of the 14C-tyrosinated tubulin species, was detyrosinated by the action of tubulin carboxypeptidase faster than the rest of the 14C-tyrosinated tubulin isotypes taken in whole. In the case of brain extract, that isotype accounts for only 16% of the labeled tubulin.

Immunodetection of tubulin carboxypeptidase activity on nitrocellulose membrane after gel electrophoresis and blotting.

It was found that the detyrosination of tyrosinated tubulin by tubulin carboxypeptidase can occur when both the enzyme and the substrate are adsorbed on nitrocellulose. This, and the use of a specific antibody that recognizes detyrosinated tubulin allowed us to localize tubulin carboxypeptidase on a nitrocellulose membrane after agarose gel electrophoresis and blotting. The method was also extended to detect pancreatic carboxypeptidase A.

Tyrosinated, detyrosinated and acetylated tubulin isotypes in rat brain membranes. Their proportions in comparison with those in cytosol.

The heterogeneity of alpha-tubulin and the relative proportions of the tubulin isotypes were investigated in brain membranes of rats of 1, 25 and 180 days of age by using four anti-alpha-tubulin antibodies: a) the monoclonal DM1A antibody, specific for alpha-tubulin; b) the monoclonal 6-11B-1 antibody, specific for acetylated tubulin; c) a polyclonal antibody (Glu antibody), specific for detyrosinated tubulin; and d) a polyclonal antibody (Tyr antibody), specific for tyrosinated tubulin. We found that rat brain membranes contain the three tubulin isotypes mentioned above. The proportions of tyrosinated and detyrosinated tubulin relative to total alpha-tubulin were somewhat lower in membrane than in cytosol in animals of 25 and 180 days of age. At day one of development, the proportions in membrane were similar to those found in cytosol. With respect to the acetylated form, it was about 20 times higher in membrane than in cytosol at the three ages studied. The proportion of acetylated tubulin was determined in different subcellular fractions: myelin, synaptic vesicles, mitochondria, microsomes, and plasma membrane. While the amount of total tubulin differed between the different subcellular fractions, the proportion of acetylated tubulin relative to total alpha-tubulin was constant and similar to that found in total membranes. The proportion of acetylated tubulin was also investigated in non-neural tissues (kidney, liver and lung). Although values for cytosol were about 10-fold higher than that found in brain cytosol, no detectable values for membranes could be obtained in these organs.

Peanut agglutinin binding glycoproteins in the chick retina: their presence in Müller glia cells.

The histological and cellular distribution and some biochemical characteristics of components that bind peanut agglutinin (PNA), a lectin that recognizes preferentially terminal galactose-beta (1-3) N-acetyl galactosamine disaccharide residues of glycoconjugates, were studied in chick retinal tissue and in dissociated retinal cells after their differentiation in culture. In sections of retinal tissue from animals 7 days after hatching (Rp7), in addition to the inner and outer segments of the photoreceptor layer, the plexiform and optic fiber layers were stained with rhodamine-labeled PNA, indicating that, besides photoreceptor cells, other cellular types contribute to the PNA staining. We present evidence indicating that at least part of this staining is provided by Müller glia cells. In cultures of dissociated cells from retinas at embryonic day 7 (R7), photoreceptor-like cells and flat Müller glia-derived cells but not neurons were stained with rhodamine-labeled PNA. Furthermore, Müller glia cells isolated from Rp7 were also brightly stained with PNA. Western blot assays of extracts from R7 showed the presence of PNA binding glycoproteins of 31-33 kDa and a component that migrates at the dye front. In addition to the components detected in R7, extracts from R14 and Rp7 showed the presence of a major PNA binding glycoprotein of 175 kDa and a minor glycoprotein of 220 kDa. Extracts from the photoreceptor layer contain the 175 and 220 kDa glycoproteins, indicating their association with photoreceptor cells. The 31-33 kDa components were detected in extracts from the remnant inner retina, suggesting their association with the Müller glia cells. Supporting this view, these components and not those of 175 and 220 kDa were detected in cell cultures enriched in flat Müller glia-derived cells. Only the 31-33 kDa components and the component that migrates at the dye front were detected in extracts from cell cultures enriched in photoreceptor-like cells, suggesting the need of some environmental element for the expression of the 175 and 220 kDa components in the differentiated photoreceptor cells.

The posttranslational arginylation of proteins in different regions of the rat brain.

The posttranslational incorporation of arginine into proteins catalyzed by arginyl-tRNA protein transferase was determined in vitro in different rat brain regions. The incorporation was found in all the regions studied, although with different specific activities (pmol [14C]arginine incorporated/mg protein). Of the regions studied, hippocampus had the highest specific activity followed by striatum, medulla oblongata, cerebellum, and cerebral cortex. Electrophoretic analysis of the [14C]arginyl proteins from the different regions followed by autoradiography and scanner densitometry showed at least 13 polypeptide bands that were labeled with [14C]arginine. The radioactive bands were qualitatively coincident with protein bands revealed by Coomassie Blue. There were peaks that showed different proportions of labeling in comparison with peaks of similar molecular mass from total brain. Most notable because of their high proportions were those of molecular mass 125 kDa in hippocampus, striatum, and cerebral cortex; 112 and 98 kDa in striatum and cerebellum; and 33 kDa in hippocampus and striatum. In lower proportions than in total brain were the peaks of 33 kDa in medulla oblongata and cerebral cortex and of 125 kDa in medulla oblongata.