γ-Tubulin is assumed to be a typical cytosolic protein necessary for nucleation of microtubules from microtubule organizing centers. Using immunolocalization and cell fractionation techniques in combination with siRNAi and expression of FLAG-tagged constructs, we have obtained evidence that γ-tubulin is also present in nucleoli of mammalian interphase cells of diverse cellular origins. Immunoelectron microscopy has revealed γ-tubulin localization outside fibrillar centers where transcription of ribosomal DNA takes place. γ-Tubulin was associated with nucleolar remnants after nuclear envelope breakdown and could be translocated to nucleoli during mitosis. Pretreatment of cells with leptomycin B did not affect the distribution of nuclear γ-tubulin, making it unlikely that rapid active transport via nuclear pores participates in the transport of γ-tubulin into the nucleus. This finding was confirmed by heterokaryon assay and time-lapse imaging of photoconvertible protein Dendra2 tagged to γ-tubulin. Immunoprecipitation from nuclear extracts combined with mass spectrometry revealed an association of γ-tubulin with tumor suppressor protein C53 located at multiple subcellular compartments including nucleoli. The notion of an interaction between γ-tubulin and C53 was corroborated by pull-down and co-immunoprecipitation experiments. Overexpression of γ-tubulin antagonized the inhibitory effect of C53 on DNA damage G(2) /M checkpoint activation. The combined results indicate that aside from its known role in microtubule nucleation, γ-tubulin may also have nuclear-specific function(s).

γ-Tubulins are highly conserved members of the tubulin superfamily essential for microtubule nucleation. Humans possess 2 γ-tubulin genes. It is thought that γ-tubulin-1 represents a ubiquitous isotype, whereas γ-tubulin-2 is found predominantly in the brain, where it may be endowed with divergent functions beyond microtubule nucleation. The molecular basis of the purported functional differences between γ-tubulins is unknown. We report discrimination of human γ-tubulins according to their electrophoretic and immunochemical properties. mutagenesis revealed that the differences in electrophoretic mobility originate in the C-terminal regions of the γ-tubulins. Using epitope mapping, we discovered mouse monoclonal antibodies that can discriminate between human γ-tubulin isotypes. Real time quantitative RT-PCR and 2-dimensional-PAGE showed that γ-tubulin-1 is the dominant isotype in fetal neurons. Although γ-tubulin-2 accumulates in the adult brain, γ-tubulin-1 remains the major isotype in various brain regions. Localization of γ-tubulin-1 in mature neurons was confirmed by immunohistochemistry and immunofluorescence microscopy on clinical samples and tissue microarrays. Differentiation of SH-SY5Y human neuroblastoma cells by all- retinoic acid, or oxidative stress induced by mitochondrial inhibitors, resulted in upregulation of γ-tubulin-2, whereas the expression of γ-tubulin-1 was unchanged. Fractionation experiments and immunoelectron microscopy revealed an association of γ-tubulins with mitochondrial membranes. These data indicate that in the face of predominant γ-tubulin-1 expression, the accumulation of γ-tubulin-2 in mature neurons and neuroblastoma cells during oxidative stress may denote a prosurvival role of γ-tubulin-2 in neurons.-Dráberová, E., Sulimenko, V., Vinopal, S., Sulimenko, T., Sládková, V., D'Agostino, L., Sobol, M., Hozák, P., Křen, L., Katsetos, C. D., Dráber, P. Differential expression of human γ-tubulin isotypes during neuronal development and oxidative stress points to γ-tubulin-2 prosurvival function.

No abstract

In previous studies, we have shown overexpression and ectopic subcellular distribution of gamma-tubulin and betaIII-tubulin in human glioblastomas and glioblastoma cell lines (Katsetos et al., 2006, J Neuropathol Exp Neurol 65:455-467; Katsetos et al., 2007, Neurochem Res 32:1387-1398). Here we determined the expression of gamma-tubulin in surgically excised medulloblastomas (n = 20) and in the human medulloblastoma cell lines D283 Med and DAOY. In clinical tissue samples, the immunohistochemical distribution of gamma-tubulin labeling was pervasive and inversely related to neuritogenesis. Overexpression of gamma-tubulin was widespread in poorly differentiated, proliferating tumor cells but was significantly diminished in quiescent differentiating tumor cells undergoing neuritogenesis, highlighted by betaIII-tubulin immunolabeling. By quantitative real-time PCR, gamma-tubulin transcripts for TUBG1, TUBG2, and TUBB3 genes were detected in both cell lines but expression was less prominent when compared with the human glioblastoma cell lines. Immunoblotting revealed comparable amounts of gamma-tubulin and betaIII-tubulin in different phases of cell cycle; however, a larger amount of gamma-tubulin was detected in D283 Med when compared with DAOY cells. Interphase D283 Med cells exhibited predominantly diffuse cytoplasmic gamma-tubulin localization, in addition to the expected centrosome-associated distribution. Robust betaIII-tubulin immunoreactivity was detected in mitotic spindles of DAOY cells. Our data indicate that overexpression of gamma-tubulin may be linked to phenotypic dedifferentiation (anaplasia) and tumor progression in medulloblastomas and may potentially serve as a promising tumor marker.

The expression, cellular distribution, and subcellular sorting of the microtubule (MT)-nucleating γ-tubulin small complex (γTuSC) proteins, GCP2 and GCP3, were studied in human glioblastoma cell lines and in clinical tissue samples representing all histologic grades of adult diffuse astrocytic gliomas (n = 54). Quantitative real-time polymerase chain reaction revealed a significant increase in the expression of GCP2 and GCP3 transcripts in glioblastoma cells versus normal human astrocytes; these were associated with higher amounts of both γTuSC proteins. GCP2 and GCP3 were concentrated in the centrosomes in interphase glioblastoma cells, but punctate and diffuse localizations were also detected in the cytosol and nuclei/nucleoli. Nucleolar localization was fixation dependent. GCP2 and GCP3 formed complexes with γ-tubulin in the nucleoli as confirmed by reciprocal immunoprecipitation experiments and immunoelectron microscopy. GCP2 and GCP3 depletion caused accumulation of cells in G2/M and mitotic delay but did not affect nucleolar integrity. Overexpression of GCP2 antagonized the inhibitory effect of the CDK5 regulatory subunit-associated tumor suppressor protein 3 (C53) on DNA damage G2/M checkpoint activity. Tumor cell GCP2 and GCP3 immunoreactivity was significantly increased over that in normal brains in glioblastoma samples; it was also associated with microvascular proliferation. These findings suggest that γTuSC protein dysregulation in glioblastomas may be linked to altered transcriptional checkpoint activity or interaction with signaling pathways associated with a malignant phenotype.