Colchicine-binding sites of brain tubulins from an antarctic fish and from a mammal are functionally similar, but not identical: implications for microtubule assembly at low temperature.

The tubulins of Antarctic fishes possess adaptations that favor microtubule formation at low body temperatures (Detrich et al.: Biochemistry 28:10085-10093, 1989). To determine whether some of these adaptations may be present in a domain of tubulin that participates directly or indirectly in lateral contact between microtubule protofilaments, we have examined the energetics of the binding of colchicine, a drug thought to bind to such a site, to pure brain tubulins from an Antarctic fish (Notothenia gibberifrons) and from a mammal (the cow, Bos taurus). At temperatures between 0 and 20 degrees C, the affinity constants for colchicine binding to the fish tubulin were slightly smaller (1.5-2.6-fold) than those for bovine tubulin. van't Hoff analysis showed that the standard enthalpy changes for colchicine binding to the two tubulins were comparable (delta H degrees = +10.6 and +7.4 kcal mol-1 for piscine and bovine tubulins, respectively), as were the standard entropy changes (delta S degrees = +61.3 eu for N. gibberifrons tubulin, +51.2 eu for bovine tubulin). At saturating concentrations of the ligand, the maximal binding stoichiometry for each tubulin was approximately 1 mol colchicine/mol tubulin dimer. The data indicate that the colchicine-binding sites of the two tubulins are similar, but probably not identical, in structure. The apparent absence of major structural modifications at the colchicine site suggests that this region of tubulin is not involved in functional adaptation for low-temperature polymerization. Rather, the colchicine site of tubulin may have been conserved evolutionarily to serve in vivo as a receptor for endogenous molecules (i.e., "colchicine-like" molecules or MAPs) that regulate microtubule assembly.     

Antarctic fish tubulins: heterogeneity, structure, amino acid compositions and charge

1. Tubulins purified from the brain tissues of three Antarctic fishes (Notothenia gibberifrons, Notothenia coriiceps neglecta, and Chaenocephalus aceratus) contain equimolar quantities of the alpha and beta chains and are free of microtubule-associated proteins (MAPs) and other non-tubulin proteins. 2. When examined by isoelectric focusing and by two-dimensional electrophoresis, brain tubulins from the Antarctic fishes were found to be highly heterogeneous; each was resolved into 15-20 distinct variants. The range of isoelectric points displayed by the Antarctic fish tubulins (5.30-5.75) is slightly more basic than that of bovine brain tubulin (5.25-5.60). 3. Peptide mapping demonstrated that tubulins from the Antarctic fishes and the cow differ in structure. 4. The amino acid compositions of piscine and mammalian tubulins are similar, but the Antarctic fish tubulins apparently contain fewer glutamyl and/or glutaminyl residues than do tubulins from the temperate channel catfish (Ictalurus punctatus) and the cow. 5. Native tubulin from N. coriiceps neglecta possesses 1-2 fewer net negative charges per tubulin dimer than does bovine tubulin. 6. We suggest that the enhanced assembly of Antarctic fish tubulins at low temperatures (-2 to +2 degrees C) results from adaptive, perhaps subtle, changes in their tubulin subunits.     

Stathmin Potentiates Vinflunine and Inhibits Paclitaxel Activity

Cell biology and crystallographic studies have suggested a functional link between stathmin and microtubule targeting agents (MTAs). In a previous study we showed that stathmin increases vinblastine (VLB) binding to tubulin, and that conversely VLB increases stathmin binding to tubulin. This constituted the first biochemical evidence of the direct relationship between stathmin and an antimitotic drug, and revealed a new mechanism of action for VLB. The question remained if the observed interaction was specific for this drug or represented a general phenomenon for all MTAs. In the present study we investigated the binding of recombinant stathmin to purified tubulin in the presence of paclitaxel or another Vinca alkaloid, vinflunine, using Isothermal Titration Calorimetry (ITC). These experiments revealed that stathmin binding to tubulin is increased in the presence of vinflunine, whereas no signal is observed in the presence of paclitaxel. Further investigation using turbidity and co-sedimentation showed that stathmin inhibited paclitaxel microtubule-stabilizing activity. Taken together with the previous study using vinblastine, our results suggest that stathmin can be seen as a modulator of MTA activity and binding to tubulin, providing molecular explanation for multiple previous cellular and in vivo studies showing that stathmin expression level affects MTAs efficiency.     

Structural and biological characterization of the tubulin interaction with dinitroanilines

Characteristics of the interaction of dinitroaniline compounds with tubulin molecules have an extremely high selectivity: these substances efficiently bind to the tubulins of both plant and protozoan origins and practically do not interact with any animal and fungal tubulins despite a very high similarity between the corresponding sequences. This work summarizes and comprehensively analyzes the specific structural features and mechanisms of these interactions, in particular, the patterns of the structure and arrangement of dinitroaniline binding sites on the surface of different tubulin subunits and tubulins of various origins. Dinitroaniline binding sites are localized to the surface of longitudinal contacts between tubulin subunits and contain diamine amino acid residues (lysine or arginine), which bind the nitrile group of dinitroanilines. The localizations of these sites on the surface of identical subunits of different origins (for example, α-tubulins of plants and protozoans) coincide; however, the location of these binding sites on the surfaces of tubulin α- and β-subunits is different. The characterized sites can also be potential binding sites for other antimicrotubule compounds, in particular, cyanoacrylates.     

Interaction of thiabendazole with fungal tubulin.

Thiabendazole, 2-(4'-thiazolyl)benzimidazole, at 80 micrometer completely inhibits mitosis in hyphae of Aspergillus nidulans, growing in liquid culture. DNA and RNA synthesis and mycelial growth are only partially inhibited at this concentration. Binding studies with cell-free mycelial extracts from Penicillium expansum showed that thiabendazole competitively inhibits [14C]carbendazim binding to tubulin, which suggests that the antimitotic activity of thiabendazole is based on interference with microtubule assembly. Tubulin from a thiabendazole-resistant and carbendazim-highly sensitive mutant of P. expansum has a lower affinity to thiabendazole and a higher affinity to carbendazim than tubulin from a wide-type strain. This indicates that in this mutant the structure of the binding site is affected. The data presented suggest that several sites of both the tubulin and ligand molecule are involved in the binding of benzimidazole compounds to fungal tubulin.     

Interaction of thiabendazole with fungal tubulin

Thiabendazole, 2-(4′-thiazolyl)benzimidazole, at 80 μM completely inhibits mitosis in hyphae of Aspergillus nidulans, growing in liquid culture. DNA and RNA synthesis and mycelial growth are only partially inhibited at this concentration.Binding studies with cell-free mycelial extracts from Penicillium expansum showed that thiabendazole competitively inhibits [¹⁴C]carbendazim binding to tubulin, which suggests that the antimitotic activity of thiabendazole is based on interference with microtubule assembly.Tubulin from a thiabendazole-resistant and carbendazim-highly sensitive mutant of P. expansum has a lower affinity to thiabendazole and a higher affinity to carbendazim than tubulin from a wild-type strain. This indicates that in this mutant the structure of the binding site is affected.The data presented suggest that several sites of both the tubulin and ligand molecule are involved in the binding of benzimidazole compounds to fungal tubulin.     

Brain and egg tubulins from antarctic fishes are functionally and structurally distinct.

The multitubulin hypothesis proposes that chemically distinct tubulins may possess different polymerization properties or may form functionally different microtubules. To test this hypothesis, we have examined the functional properties and the structures of singlet-specific nonneural and neural tubulins from Antarctic fishes. Tubulins were purified from eggs of Notothenia coriiceps neglecta, and from brain tissues of N. coriiceps neglecta or N. gibberifrons, by DEAE ion-exchange chromatography and cycles of microtubule assembly/disassembly. At temperatures between 0 and 20 degrees C, each of these tubulins polymerized efficiently in vitro to yield microtubules of normal morphology. Critical concentrations for polymerization of egg tubulin ranged from 0.057 mg/ml at 3 degrees C to 0.002 mg/ml at 18 degrees C, whereas those for brain tubulin at like temperatures were 4-10-fold larger. Polymerization of both tubulins was entropically driven, but the apparent standard enthalpy and entropy changes for microtubule elongation by egg tubulin (delta Happ0 = +33.9 kcal/mol, delta Sapp0 = +151 entropy units) were significantly greater than values observed for brain tubulin (delta Happ0 = +26.5 kcal/mol, delta Sapp0 = +121 entropy units). Egg tubulin was composed of approximately six alpha and two beta chains and lacked the beta III isotype, whereas brain tubulin was more complex (greater than or equal to 10 of each chain type). Furthermore, egg alpha tubulins were more basic, and their carboxyl termini more resistant to cleavage by subtilisin, than were the alpha chains of brain. We conclude that brain and egg tubulins from the Antarctic fishes are functionally distinct in vitro, due either to qualitative or quantitative differences in isotypic composition, to differential posttranslational modification of shared isotypes, or to both.     

Tetrahymena tubulins and in vitro translation of Tetrahymena RNA.

Tetrahymena outer doublet tubulin was compared with neurotubulin and Chlamydomonas flagellar tubulin on SDS-polyacrylamide gels. Tetrahymena alpha tubulin did not comigrate with either brain or flagellar alpha tubulins, although brain, flagellar, and ciliary beta tubulins all comigrated. Axonemal tubulin from Tetrahymena strain ST was compared with this tubulin from strains W, S, HSM, and E, and all were found to have the same mobilities. Poly-A containing RNA was separated from whole cell Tetrahymena RNA by oligo-dT cellulose chromatography. Poly-A+ RNA from 24-h cultures (early exponential growth) stimulated greater incorporation of amino acids into polypeptides in the wheat germ cell-free translation system than did poly-A+ RNA from 36-h and 49-h cultures. When separated on SDS-polyacrylamide gels, the translation products of the 24-h poly-A+ RNA had 2 prominent protein bands which comigrated with alpha and beta tubulin isolated from Tetrahymena cilia. These bands were not found in the translation products of poly-A+ RNA isolated from 49-h cultures or in the translation products of poly-A- RNA.     

Phosphorylation of tubulin by a calmodulin-dependent protein kinase

Calmodulin-dependent protein kinase was purified from porcine brain cytosol through sequential steps involving acid precipitation, DEAE-chromatography, and calmodulin-Sepharose chromatography. The purified enzyme contained a major Mr 50,000 and a minor Mr 60,000 peptide. Porcine brain tubulin was a major substrate for this kinase. Under optimal conditions 2.6 mol of phosphate were incorporated per mol of tubulin. The kinase phosphorylated both tubulin subunits at their carboxyl-terminal region. Limited proteolysis, using trypsin and chymotrypsin, of phosphorylated and unphosphorylated tubulins resulted in different cleavage patterns as determined by peptide mapping. Phosphorylated tubulin was unable to bind to microtubule-associated protein or to polymerize, but regained its assembly capacity after phosphatase treatment.     

Cold-stable microtubules from Antarctic fishes contain unique alpha tubulins

The cytoplasmic microtubules of Antarctic fishes assemble from their tubulin subunits at physiological body temperatures in the range -2 to +2 degrees C. Our objective is to determine the structural features that enhance the assembly of Antarctic fish tubulins at low temperatures. Here we compare the structures of tubulin subunits from three Antarctic fishes (Notothenia gibberifrons, Notothenia coriiceps neglecta, and Chaenocephalus aceratus), from three temperate fishes (the dogfish shark Mustelus canis, the channel catfish Ictalurus punctatus, and the goosefish Lophius americanus), and from a mammal (the cow Bos taurus). When reduced, carboxymethylated, and examined by polyacrylamide gel electrophoresis, multiple alpha chains were observed in tubulins from the Antarctic fishes, the catfish, and the goosefish; dogfish and bovine alpha tubulins migrated as single components on this gel system. Prominent in the Antarctic fish tubulins was an alpha variant that migrated more rapidly than the bovine alpha chain; smaller amounts of a rapidly migrating alpha chain were also present in catfish and goosefish tubulins. The beta tubulins of the fishes, with the exception of the goosefish, resolved into major and minor variants with mobilities similar to those of beta 1 and beta 2 tubulins from bovine brain. Peptide mapping demonstrated that the alpha tubulins of Antarctic fishes were similar in structure, yet differed from the alpha chains of the dogfish and the cow (which, in turn, were similar to each other). In contrast, the beta tubulins from these organisms gave peptide patterns of near identity. Finally, the alpha chains of native tubulins from N. coriiceps neglecta and the cow differed in the sensitivity of their C-terminal domains to digestion by subtilisin. These results demonstrate that the alpha tubulins of Antarctic fishes (but not their beta chains) differ structurally from those of temperate fishes and a mammal.     

Differential binding of methyl benzimidazol-2-yl carbamate to fungal tubulin as a mechanism of resistance to this antimitotic agent in mutant strains of Aspergillus nidulans

The antimitotic compound methyl benzimidazol-2-yl carbamate (MBC) formed a complex in vitro with a protein present in mycelial extracts of fungi. The binding protein of Aspergillus nidulans showed a set of properties which is unique for tubulin. Binding occurred rapidly at 4 degrees C and was competitively inhibited by oncodazole and colchicine. Other inhibitors of microtubule function such as podophyllotoxin, vinblastine sulfate, melatonin, and griseofulvin did not interfere with binding of MBC. Electrophoretic analysis of partially purified preparations of the binding protein revealed the presence of proteins with similar mobilities as mammalian tubulin monomers. Hence it is concluded that the binding protein is identical with fungal tubulin. The effect of MBC on mycelial growth of mutant strains of A. nidulans was positively correlated with the affinity of the binding sites for this compound. The apparent binding constant for MBC and tubulin from a wild type was estimated at 4.5 X 10(5), from a resistant strain at 3.7 X 10(4), and from a strain with increased sensitivity to MBC at 1.6 X 10(6) liters/mol. Mutants showing resistance and increased sensitivity to MBC are candidates to have alterations in tubulin structure. Affinity of tubulin for MBC is probably a common mechanism of resistance to this compound in fungi. Low affinity of tubulin for MBC is probably a common mechanism of resistance binding constant of 2.5 X 10(3) liters/mol.     

Heterogeneity and structure of brain tubulins from cold-adapted Antarctic fishes. Comparison to brain tubulins from a temperate fish and a mammal

Tubulins purified from brain tissue of Antarctic fishes assemble in vitro to form microtubules at the low temperatures experienced by these extreme psychrophiles (Williams, R. C., Jr., Correia, J. J., and DeVries, A. L. (1985) Biochemistry 24, 2790-2798). We have initiated studies to determine the structural requirements for assembly of Antarctic fish tubulins at low temperatures. As a first step we have compared the heterogeneity, structures, amino acid compositions, and net charge of brain tubulins purified from three Antarctic fishes (Notothenia gibberifrons, Notothenia coriiceps neglecta, and Chaenocephalus aceratus), from the temperate channel catfish (Ictalurus punctatus), and from a mammal (the cow). Each preparation contained the alpha- and beta-tubulins and was free of microtubule-associated proteins. When examined by isoelectric focusing and by two-dimensional electrophoresis, brain tubulins from the Antarctic fishes were found to be highly heterogeneous; each was resolved into approximately 20 isoelectric variants. The distributions of the isotubulins from the cold-adapted fishes were similar but differed significantly from those of tubulins from catfish and cow. The average isoelectric points of the alpha- and beta-tubulins from the Antarctic fishes were more basic than the isoelectric points of the corresponding tubulins from bovine brain. Peptide mapping confirmed that tubulins from the Antarctic fishes and the mammal differed in structure. The amino acid compositions of fish and mammalian tubulins were similar, but Antarctic fish tubulins apparently contained fewer Glx residues than did catfish or bovine tubulins. Finally, native tubulins from an Antarctic fish and the cow differed slightly in net negative charge. Thus, brain tubulins from the cold-adapted fishes differ structurally from the tubulins of a temperate fish and of a mammal.     

Post-translationally modified tubulins in Artemia: prelarval development in the absence of detyrosinated tubulin

The synthesis of post-translationally modified tubulins was examined during Artemia development. Tubulin, either purified to homogeneity or in cell-free extracts, was blotted to nitrocellulose and probed with a panel of antibodies. When purified tubulin was examined, tyrosinated tubulin underwent a large decrease as development progressed and this was accompanied by the appearance of detyrosinated tubulin in samples from organisms developed 24 hr. The inclusion of carboxypeptidase inhibitors had a small effect on the relative amounts of tyrosinated and detyrosinated tubulins in 24-hr preparations. The amount of alpha- and beta-tubulin in cell-free extracts of Artemia either remained relatively constant during development or increased slightly. The same result was obtained for acetylated and tyrosinated tubulin. Detyrosinated tubulin first appeared in 24-hr cell-free extracts and was only post-translationally modified tubulin to increase, relative to the total amount of tubulin, as the brine shrimp developed. As revealed by immunofluorescence staining, detyrosinated tubulin occurred in many cell types of developing nauplii and was prominently displayed in mitotic figures. Artemia, a complex metazoan animal, is thus able to grow for an extended period of time in the absence of detyrosinated tubulin. This isoform is however, synthesized in early larvae and may be required for the development of elongated cells including those which encircle the gut. Detyrosination remains as the only developmentally related change observed for brine shrimp tubulin.     

Classical ligands interact with native and recombinant tubulin from Onchocerca volvulus with similar rank order of magnitude

The alpha- and beta-tubulin genes from Onchocerca volvulus were individually expressed for the first time in Escherichia coli (DH5alpha). The recombinant tubulins were purified, renatured and reconstituted into oligomers, probably dimers, which were competent to bind three classical tubulin ligands: mebendazole (MBZ), taxol (TAX) and vinblastine (VBN). A new charcoal-dependent binding assay allowed accurate discrimination between specific and non-specific ligand binding in crude cell extracts. To compare the magnitude of binding of both native and recombinant forms of tubulin, we developed an ELISA assay for estimating the amount of tubulin in soluble protein extracts of O. volvulus. Binding assays were performed; both the maximum binding at saturating ligand concentrations (B(max)) and the equilibrium dissociation constants (K(d)) were determined. The B(max) values of the different ligands were significantly different from one another (P<0.05), but the order of the B(max) and K(d) for each drug were VBN > TAX > MBZ for both native and recombinant tubulin. Indeed, B(max) values for MBZ with native and recombinant tubulins were similar. On average, native tubulin had higher or similar binding capacity (B(max)) but a consistently higher affinity (lower K(d)) than the recombinant tubulin. We conclude that at least some of the recombinant molecules form receptors that are similar to those in native tubulin dimers. These data suggest that recombinant tubulin can be used to develop a molecular screen for novel anti-tubulin ligands to develop into drugs against onchocerciasis.     

Identification of multiple tubulins in taxol microtubules purified from carrot suspension cells

Tubulin has been purified from carrot suspension cells by ion-exchange chromatography and assembled into microtubules in the presence of 20 microM taxol. One-dimensional SDS-PAGE suggested that the alpha band migrated faster than the beta band (as has been established for some lower eukaryotic tubulins) and this heterology with brain tubulins was confirmed by peptide mapping. When subjected to two-dimensional gel electrophoresis, the plant tubulins could be separated into multiple alpha and beta isotypes. Immunoblotting, using monoclonal anti-tubulins, confirmed that the tubulin isotypes identified in taxol microtubules represent all of the tubulins present in homogenates of unsynchronised log-phase carrot suspension cells. All identified tubulins are therefore assembly-competent under these conditions. Plant cells can contain four different microtubule arrays, but cells arrested in G0/G1 contain only cortical microtubule arrays; such cells, however, exhibit the same tubulin profile as non-synchronised cells, thereby showing no restriction in the number of subunits during this phase of the cell cycle.     

Identification of the binding region of basic calponin on alpha and beta tubulins

Calponin is a basic smooth muscle protein capable of binding to actin, calmodulin, tropomyosin, and phospholipids. We have found that the basic calponin interacted with brain tubulin under polymerized and unpolymerized conditions in vitro [Fujii, T., Hiromori, T., Hamamoto, M., and Suzuki, T. (1997) J. Biochem. 122, 344-351]. We examined the calponin-binding site on the tubulin molecule by sedimentation, limited digestion, chemical-cross linking, immunoblotting, and delayed extraction matrix-assisted laser desorption ionization time-of-flight mass spectrometric (DE MALDI-TOF) analyses. Calponin interacts with both the alpha and beta tubulins and only slightly with the tyrosinated and acetylated form of alpha tubulin. The binding of calponin to microtubules was blocked by adding poly(L-aspartic acid) (PLAA) or MAP2. After digestion of microtubule proteins with subtilisin, the amount of calponin binding to alphabetas microtubules was reduced compared to native microtubules, but no further reduction was observed in the case of alphasbetas microtubules. The chemical cross-linked products of calponin and synthesized peptides (KDYEEVGVDSVEGE; alpha-KE) derived from the C-terminal region of alpha tubulin and (YQQYQDATADEQG; beta-YG) and (GEFEEEGEEDEA; beta-GA) from that of beta tubulin were detected by mass spectrometry. One kind of calponin-peptide complex was formed in the presence of alpha-KE or beta-YG, while five complexes (calponin:peptide = 1:1-5) were generated in the presence of beta-GA. Peptides alpha-KE and beta-GA inhibited the binding of calponin to tubulin produced by EDC in a concentration-dependent manner. These findings suggest that basic calponin interacts with both tubulin subunits and that their C-terminal regions, which also contain the binding sites of MAP2, tau, and kinesin, may be involved in calponin-binding.     

Defining the eukaryotic cytosolic chaperonin-binding sites in human tubulins

The actins and tubulins are the obligate substrates in vivo of the chaperonin-containing TCP-1 (CCT). The precise elements of recognition between the chaperonin and its substrates remain largely unknown. We have used a solid phase peptide binding assay to screen the human alpha, beta and gamma-tubulin sequences for CCT recognition. Multiple regions seem to be implicated in interactions between tubulins and CCT. These potential CCT-binding sites are highly dispersed throughout the primary sequences of the human tubulins. In addition, using site-directed mutagenesis we assessed the contribution of the selected residues in the C-terminal domain of beta-tubulin to CCT binding. Various hot spots have been identified even though, in each case, their replacement by alanine does not reduce dramatically the total affinity of beta-tubulin for CCT. The CCT-binding information in the tubulins is probably confined to multiple specific regions each having weak or moderate affinity for CCT apical domains. The main binding region seems to be located between residues 263 and 384, but there are no single amino acid residues in this region, which make large contributions to the binding energy, although we have detected a minor contribution by F377. These biochemical results are understandable in the context of our recent structural analysis of CCT-tubulin complexes by cryo-electron microscopy and image reconstruction, which shows that, in one stage of an in vitro binding reaction between apo-CCT and tubulin diluted from guanidinium chloride, ten major, stable contacts between tubulin and CCT are involved. Therefore, specificity is achieved through the co-operation of many specific, albeit weak, interactions.     

Microtubule formation from maternal tubulins during sea urchin embryogenesis: measurement of soluble and insoluble tubulin pools

The mass of tubulin protein in developing embryos of the sea urchin Lytechinus pictus was measured using a radiodilution immunoassay based on densitometric analysis of immunoprecipitated tubulins resolved electrophoretically. The tubulins constitute an average of 360 +/- 35 pg per egg, or 0.66% of the total protein, and there is no significant change in their concentration during embryogenesis. The masses of soluble and polymerized tubulin were measured for extracts prepared under conditions that stabilize microtubules. In eggs, a maximum of 14% of the tubulin is insoluble, and this increases throughout embryogenesis to 67% at pluteus stage (72 hr). The concentration of tubulin in eggs is at least 500 micrograms/ml, well above the critical concentration for tubulin assembly in vitro, yet microtubules have not been observed in eggs. The mass of newly synthesized tubulin, estimated from the mass of tubulin mRNA per embryo, accounts for a small fraction of the total tubulin by the end of gastrulation but for over half of the tubulin by the 72-hr pluteus stage. These observations are consistent with a model in which the declining level of unpolymerized tubulin controls the stability of tubulin mRNa, providing an autogenous regulation of the ontogenetic pattern of tubulin synthesis during sea urchin embryogenesis (Gong and Brandhorst, Development 102: 31-43).     

Posttranslational modification of brain tubulins from the Antarctic fish Notothenia coriiceps: reduced C-terminal glutamylation correlates with efficient microtubule assembly at low temperature

We have shown previously that the tubulins of Antarctic fish assemble into microtubules efficiently at low temperatures (-2 to +2 degrees C) due to adaptations intrinsic to the tubulin subunits. To determine whether changes in posttranslational glutamylation of the fish tubulins may contribute to cold adaptation of microtubule assembly, we have characterized C-terminal peptides from alpha- and beta-tubulin chains from brains of adult specimens of the Antarctic rockcod Notothenia coriiceps by MALDI-TOF mass spectrometry and by Edman degradation amino acid sequencing. Of the four fish beta-tubulin isotypes, nonglutamylated isoforms were more abundant than glutamylated isoforms. In addition, maximal glutamyl side-chain length was shorter than that observed for mammalian brain beta tubulins. For the nine fish alpha-tubulin isotypes, nonglutamylated isoforms were also generally more abundant than glutamylated isoforms. When glutamylated, however, the maximal side-chain lengths of the fish alpha tubulins were generally longer than those of adult rat brain alpha chains. Thus, Antarctic fish adult brain tubulins are glutamylated differently than mammalian brain tubulins, resulting in a more heterogeneous population of alpha isoforms and a reduction in the number of beta isoforms. By contrast, neonatal rat brain tubulin possesses low levels of glutamylation that are similar to that of the adult fish brain tubulins. We suggest that unique residue substitutions in the primary structures of Antarctic fish tubulin isotypes and quantitative changes in isoform glutamylation act synergistically to adapt microtubule assembly to low temperatures.     

Epothilone and paclitaxel: unexpected differences in promoting the assembly and stabilization of yeast microtubules

Paclitaxel (Taxol) and the epothilones are antimitotic agents that promote the assembly of mammalian tubulin and stabilization of microtubules. The epothilones competitively inhibit the binding of paclitaxel to mammalian brain tubulin, suggesting that the two types of compounds share a common binding site in tubulin, despite the lack of structural similarities. It is known that paclitaxel does not stabilize microtubules formed in vitro from Saccharomyces cerevisiae tubulin; thus, it would be expected that the epothilones would not affect yeast microtubules. However, we found that epothilone A and B do stimulate the formation of microtubules from purified yeast tubulin. In addition, epothilone B severely dampens the dynamics of yeast microtubules in vitro in a manner similar to the effect of paclitaxel on mammalian microtubules. We used current models describing paclitaxel and epothilone binding to mammalian beta-tubulin to explain why paclitaxel apparently fails to bind to yeast tubulin. We propose that three amino acid substitutions in the N-terminal region and at position 227 in yeast beta-tubulin weaken the interaction of the 3'-benzamido group of paclitaxel with the protein. These results also indicate that mutagenesis of yeast tubulin could help define the sites of interaction with paclitaxel and the epothilones.