alpha-Tubulin limits its own synthesis: evidence for a mechanism involving translational repression

A Chinese hamster alpha-tubulin cDNA was modified to encode an 11-amino acid carboxyl-terminal extension containing the immunodominant epitope from influenza hemagglutinin antigen (to create HA alpha 1-tubulin) and was cloned into a vector for expression in mammalian cells. 12 stable CHO cell lines expressing this HA alpha 1-tubulin were isolated and characterized. HA alpha 1-tubulin incorporated into all classes of microtubules, assembled to the same extent as the endogenous tubulin, and did not perturb the growth of the cells in which it was expressed. However, overexpression of HA alpha 1-tubulin strongly repressed the synthesis of endogenous alpha-tubulin while having little or no effect on the synthesis of beta-tubulin. Treatment of transfected cells with sodium butyrate to induce even greater expression of HA alpha 1-tubulin led to a further decrease in synthesis of endogenous alpha-tubulin that was fully reversible upon removal of the inducer. Decreased synthesis of alpha-tubulin in transfected cells did not result from decreased levels of alpha-tubulin mRNA, as demonstrated by ribonuclease protection assays. On the other hand, colchicine, a drug previously shown to destabilize the tubulin message, caused a clear reduction in both protein synthesis and mRNA levels for transfected HA alpha 1- tubulin and endogenous alpha-tubulin, thus indicating that the decreased alpha-tubulin synthesis observed as a result of HA alpha 1- tubulin overexpression is distinct from the previously described autoregulation of tubulin. The results are consistent with a mechanism in which free alpha-tubulin inhibits the translation of its own message as a way of ensuring stoichiometric synthesis of alpha- and beta- tubulin.     

Physical evidence for cotranslational regulation of beta-tubulin mRNA degradation.

Tubulin synthesis is controlled by an autoregulatory mechanism through which an increase in the intracellular concentration of tubulin subunits leads to specific degradation of tubulin mRNAs. The sequence necessary and sufficient for the selective degradation of a beta-tubulin mRNA in response to changes in the level of free tubulin subunits resides within the first 13 translated nucleotides that encode the amino-terminal sequence of beta-tubulin, Met-Arg-Glu-Ile (MREI). Previous results have suggested that the sequence responsible for autoregulation resides in the nascent peptide rather than in the mRNA per se, raising the possibility that the regulation of the stability of tubulin mRNA is mediated through binding of tubulin or some other cellular factor to the nascent amino-terminal tubulin peptide. We now show that this putative cotranslational interaction is not mediated by tubulin alone, as no meaningful binding is detectable between tubulin subunits and the amino-terminal beta-tubulin polypeptide. However, microinjection of a monoclonal antibody that binds to the beta-tubulin nascent peptide selectively disrupts the regulation of beta-tubulin, but not alpha-tubulin, synthesis. This finding provides direct evidence for cotranslational degradation of beta-tubulin mRNA mediated through binding of one or more cellular factors to the beta-tubulin nascent peptide.     

Autoregulation of tubulin synthesis in hepatocytes and fibroblasts

Microtubule polymer levels in mouse 3T6 fibroblasts and primary cultures of rat hepatocytes can be manipulated by treatment of cells with long term, low doses of colcemid. Such treatment produces a rather uniform population of cells with microtubules of reduced lengths. Using this system, we demonstrate (a) that the rate of tubulin synthesis is sensitive to small changes (10%) in microtubule polymer mass and (b) that the percent of inhibition of synthesis is proportional to the level of soluble tubulin. Experiments with hepatocytes indicate that not only synthesis but the stability of tubulin protein was also regulated to maintain a specific level of tubulin. Treatment of hepatocytes with colcemid or other microtubule-depolymerizing drugs reduced the half-life of tubulin from 50 to 2 h, whereas taxol, which stabilizes microtubules, increased the half-life. To assess the consequences of altering microtubule polymer mass, we have analyzed the effect of controlled depolymerization of microtubules in rat hepatocytes on the processing of endocytosed ligands and found it sensitive to small changes in microtubule polymer levels.     

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).     

Autoregulation of tubulin expression is achieved through specific degradation of polysomal tubulin mRNAs

We have utilized protein synthesis inhibitors to investigate the autoregulatory mechanism that uses the concentration of unpolymerized tubulin subunits to specify tubulin mRNA content in animal cells. Puromycin and pactamycin, both of which remove RNAs from polysomes, completely unlink tubulin RNA content from the level of free subunits, whereas pretreatment of cells with cycloheximide, which traps mRNAs onto stalled polyribosomes, enhances the specific degradation of tubulin RNAs in response to increases in the subunit content. Moreover, in the absence of protein synthesis inhibitors, the tubulin RNAs that are lost from cells with elevated free tubulin subunit levels are those that are associated with polyribosomes. Further, beta-tubulin mRNAs encoding a truncated translation product of only 26 amino acids (and that cannot be polyribosomal) are not substrates for autoregulation. We conclude that autoregulation of tubulin synthesis is achieved by specifically altering the stability of tubulin RNAs that are bound to polyribosomes.     

Direct measurement of tubulin and bulk message distributions on polysomes of growing, starved and deciliated Tetrahymena using RNA gel blots of sucrose gradients containing acrylamide.

A method was developed using sucrose gradients containing acrylamide which greatly simplifies the measurement of the polysomal distribution of messages. After centrifugation, the acrylamide was polymerized, forming a "polysome gel". RNA gel blots of polysome gels were used to determine the polysomal distributions of alpha-tubulin and total polyadenylated mRNA in growing, starved (nongrowing) and starved-deciliated Tetrahymena and the number of messages loaded onto polysomes was calculated. These measurements indicated that the translational efficiencies of alpha-tubulin mRNA and total polyadenylated mRNA are largely unaffected when the rates of tubulin and total protein synthesis vary dramatically. Thus, differential regulation of alpha-tubulin mRNA translation initiation does not contribute to the greater than 100-fold induction of tubulin synthesis observed during cilia regeneration and in growing cells. The major translation-level process regulating tubulin synthesis in Tetrahymena appears to be a change in message loading mediated by a non-specific message recruitment or unmasking factor.     

Human chromosomes and centrioles as nucleating sites for the in vitro assembly of microtubules from bovine brain tubulin

Treatment of HeLa cells with Colcemid at concentrations of 0.06-0.10 mug/ml leads to irreversible arrest in mitosis. Colcemid-arrested cells contained few microtubules, and many kinetochores and centrioles were free of microtubule association. When these cells were exposed to microtubule reassembly buffer containing Triton X-100 and bovine brain tubulin at 37 degrees C, numerous microtubules were reassembled at all kinetochores of metaphase chromosomes and in association with centriole pairs. When bovine brain tubulin was eliminated from the reassembly system, microtubules failed to assemble at these sites. Similarly, when EGTA was eliminated from the reassembly system, microtubules failed to polymerize. These results are consistent with other investigations of in vitro microtubule assembly and indicate that HeLa chromosomes and centrioles can serve as nucleating sites for the assembly of microtubules from brain tubulin. Both chromosomes and centrioles became displaced from their C-metaphase configurations during tubulin reassembly, indicating that their movements were a direct result of microtubule formation. Although both kinetochore- and centriole- associated microtubules were assembled and movement occurred, we did not observe direct extension of microtubules from kinetochores to centrioles. This system should prove useful for experimental studies of spindle microtubule formation and chromosome movement in mammalian cells.     

Modulation of tubulin mRNA levels by interferon in human lymphoblastoid cells.

Blot hybridization with labeled tubulin cDNA showed that treatment of Ramos cells, a human cell line of lymphoblastoid origin, with either alpha or beta interferon (IFN) induced a marked increase in the amount of tubulin mRNA sequences. The level of tubulin mRNA sequences increased rapidly after exposure of cells to IFN-alpha and reached a maximum after 1 h of treatment, which was four times the control level. Treatment with IFN-beta induced a maximal increase after 4 h; the amount of tubulin mRNA sequences was seven times higher than the control level. The mRNA extracted from IFN-treated and nontreated cells was translated in vitro in a reticulocyte lysate cell-free system containing [35S]methionine. Electrophoretic analysis of the labeled cell-free products showed an increase in the amount of translatable tubulin mRNA that parallels the time course of induction of tubulin mRNA sequences. Two-dimensional gel electrophoresis of the labeled protein products directed by mRNA indicates that IFN caused a more pronounced increase in the level of alpha-tubulin than beta-tubulin mRNA. Treatment with colchicine, which disrupts the cell microtubules, caused a marked decrease in the tubulin mRNA content. Concomitant treatment of the cells with colchicine and IFN abolished the interferon-dependent induction of tubulin mRNA.     

Role of membrane colchicine binding proteins in the transmission of prolactin message to casein genes in the rabbit mammary gland

Previous work demonstrated that tubulin binding drugs specifically inhibit the capacity of prolactin to initiate casein and DNA synthesis in the mammary cell. It was concluded that microtubules or other tubulin containing cellular structures were involved in the transmission of the prolactin message to genes. In the present work, it is shown that griseofulvin, an antimitotic drug which alters microtubule structure and function, does not prevent prolactin actions. Autoradiographic studies showed that [3H]colchicine binds preferentially to plasma and Golgi membranes in the mammary cell. Short term cultures of mammary explants with [3H]colchicine demonstrated that the labelled drug binds to membranous cellular structures which were isolated from explants at the end of the culture. Fractions containing plasma and Golgi membranes contained the highest amount of radioactivity. Solubilisation of the membranes by Triton X-100 dissociated the [3H]colchicine from the prolactin receptors as judged by a chromatography of the soluble fraction on a Sepharose 6 B column. On the column, the labelled colchicine remains associated with a molecular entity which may be free tubulin. In all cases, the binding of [3H]colchicine was greatly attenuated by an excess of unlabelled colchicine but was only slightly affected by the competition with lumicolchicine. These results suggest that mammary membranes contain tubulin and that binding of drugs to this molecule inhibits the generation of the prolactin second messengers eliciting the hormonal actions in the mammary cell. This also suggests that microtubules are probably not involved in the mechanism of prolactin action.     

Effects of estramustine, a new anti-microtubule drug, on the induction of casein gene expression by prolactin

Estramustine, a new anti-microtubule drug, was added to the culture medium of rabbit mammary explants with lactogenic hormones. In the absence of the drug, prolactin with insulin and cortisol stimulated DNA synthesis and it induced beta-casein and beta-casein mRNA accumulation in the tissue. As opposed to other anti-microtubule agents such as colchicine, estramustine was unable to prevent prolactin actions. An examination of the mammary cells by immunofluorescence revealed that the microtubule network was significantly altered under the influence of estramustine. These data indicate that the integrity of microtubules is not required for prolactin to deliver its message to the mammary cell. These data also suggest that other anti-microtubule drugs such as colchicine which prevent prolactin action act through their binding to tubulin molecule unrelated to microtubule structures.     

Local anesthetic-induced inhibition of collagen secretion in cultured cells under conditions where microtubules are not depolymerized by these agents

Tertiary amine local anesthetics previously have been shown to influence some microtubule-dependent cellular functions. Since several cell secretion processes, including secretion of collagen, have been shown to be inhibited by microtubule-disrupting drugs such as colchicine, we determined whether local anesthetics affect collagen secretion. Six local anesthetics inhibited collagen and non-collagen protein secretion (up to 98%) into the extracellular medium of 3T3 cells and human fibroblasts, an effect apparently independent of influences on proline transport and total protein synthesis. A combination of colchicine and cytochalasin B did not duplicate the effects of local anesthetics. The effects of subsaturating concentrations of colchicine and procaine on secretion were additive, suggesting that both drugs act on the secretory pathway at the level of microtubules, but other effects of the two types of drugs were strikingly different. In comparing the mechanisms of action of colchicine and local anesthetics, it was seen that, in contrast to colchicine, radioactive procaine and lidocaine were slowly transported into 3T3 cells, did not bind to the tubulin-containing TCA-insoluble fraction, and did not bind to purified tubulin in vitro. The fraction of cellular tubulin present as microtubules (47% in normal cells) was determined by measuring tubulin in stabilized, sedimentable microtubules compared to total tubulin, using a [3H]colchicine binding assay. Pretreatment of cells in the cold or with colchicine led to depolymerization of microtubules, but pretreatment with five local anesthetics tested did not. Therefore, in contrast to colchicine, local anesthetics in concentrations that inhibit secretion do not directly interact with or depolymerize microtubules. These drugs, however, do affect a microtubule-dependent process and may do so by detaching the microtubular system from the cell membrane.     

Regulation of tubulin synthesis and cell cycle progression in mammalian cells by gamma-tubulin-mediated microtubule nucleation

We have previously shown that gamma-tubulin, the third member of the tubulin family that functions in microtubule nucleation, when overexpressed, accumulates throughout the cytoplasm and forms numerous ectopic microtubule nucleation sites in mammalian cells (Shu and Joshi [1995] J. Cell. Biol. 130:1137-1147). We now show that overexpression of gamma-tubulin differentially upregulates the synthesis of alpha- and beta-tubulins in mammalian cells. Surprisingly, despite a dramatic increase in the level of gamma-tubulin protein in transfected cells, there is no obvious alteration in the level of endogenous gamma-tubulin mRNA, suggesting that synthesis of gamma-tubulin might employ a regulatory mechanism other than the autoregulatory pathway shared by alpha- and beta-tubulins. Interestingly, a significant number of mammalian cells transfected with gamma-tubulin fail to form normal bipolar mitotic spindle during mitosis; instead, numerous microtubules occur in the cytoplasm intermingled with the condensed chromosomes. In addition, they reduplicate their DNA after an abnormal mitotic exit. These results thus suggest that the number of microtubule nucleation sites, or even gamma-tubulin itself, might play an important role in the regulation of tubulin synthesis as well as cell cycle progression.     

Microtubules, mitochondria, and molecular chaperones: a new hypothesis for in vivo assembly of microtubules

In Chinese hamster ovary cells, a number of independent mutants selected for resistance to antimitotic drugs have been found to be specifically altered in two major cellular proteins, designated P1 (relative mass (Mr) approximately 60-63 kilodaltons (kDa] and P2 (Mr approximately 69-70 kDa), which appeared microtubule related by a number of genetic and biochemical criteria. Antibodies to P1 have been found to bind specifically to mitochondria that showed specific association with microtubules in interphase cells. Biochemical and cDNA sequence studies on P1 showed that this protein, which is localized in the matrix compartment, is the mammalian homolog of the highly conserved chaperonin family of proteins (other members include the GroEL protein of Escherichia coli, the 60-kDa heat-shock protein of yeast, and the rubisco subunit binding protein of plant chloroplasts). The chaperonin proteins in various systems play a transient but essential molecular chaperone role in the proper folding of polypeptide chains and their assembly into oligomeric protein complexes. Our studies on P2 protein established that it corresponds to the constitutive form of the major 70-kDa heat-shock protein of mammalian cells (i.e., hsc70), which also acts as a molecular chaperone in the intracellular transport of nascent proteins to organelles such as mitochondria and endoplasmic reticulum. To account for the above, as well as a number of other observations (e.g., binding of fluorescent-labeled antimitotic drugs to mitochondria, association of tubulin with mitochondria as well as other membranes, and high affinity binding of antimitotic drugs to free tubulin but not to assembled microtubules), a new model for the in vivo assembly of interphase microtubules is proposed. The model ascribes a central role to the mitochondrially localized chaperonin (i.e., P1) protein in the intracellular formation of tubulin dimers and in their addition to the growth sites in microtubules. The proposed model also explains a number of other observations related to microtubule assembly in the literature.