A procedure adapted from that described by Mitchison and Kirschner [Nature 312:232-237, 1984] was used to isolate centrosomes from human lymphoid cells. High yields of homogeneous centrosomes (60% of the theoretical total, assuming one centrosome per cell) were obtained. Centrosomes were isolated as pairs of centrioles, plus their associated pericentriolar material. Ultrastructural investigation revealed: 1) a link between both centrioles in a centrosome formed by the gathering in of a unique bundle of thin filaments surrounding each centriole; 2) a stereotypic organization of the pericentriolar material, including a rim of constant width at the proximal end of each centriole and a disc of nine satellite arms organized according to a ninefold symmetry at the distal end and; 3) an axial hub in the lumen of each centriole at the distal end surrounded by some ill-defined material. The total protein content was 2 to 3 X 10(-2) pg per isolated centrosome, a figure that suggests that the preparations were close to homogeneity. The protein composition was complex but specific, showing proteins ranging from 180 to 300 kD, one prominent band at 130 kD, and a group of proteins between 50 and 65 kD. Actin was also present in centrosome preparations. Functional studies demonstrated that the isolated centrosomes were competent to nucleate microtubules in vitro from purified tubulin in conditions in which spontaneous assembly could not occur. They were also very effective at inducing cleavage when microinjected into unfertilized Xenopus eggs.

The lck gene product, p56lck, is a member of the src‐related family of protein tyrosine kinases. It is known as lymphocyte specific and involved in thymocyte development and in the immune response mediated by the T cell receptor. We report that the lck gene is also expressed in adult mouse CNS and that brain p56lck is similar to the thymus protein. In situ hybridization and immunohistochemistry show that the lck gene is expressed in neurons throughout the brain in distinct regions, including hippocampus and cerebellum. In primary cultures from fetal mouse brain, neuronal cells are immunoreactive to Lck antiserum. This suggests that the lck gene product might be involved in a new signal transduction pathway in mouse brain.

Highly enriched preparations of centrosomes from human T‐lymphoblasts KE 37 were analyzed for their protein content. The specific pattern of polypeptides was characterized by an abundant subset of high mol. wt proteins and a major group of proteins with mol. wt ranging from 50 to 65 kd. Several immunoreactive proteins were identified, using a rabbit serum spontaneously reacting with human centrosomes. They include a family of high mol. wt ranging from 180 to 250 kd, a 130‐kd protein and a 60‐65 kd doublet. These antigens have the following properties: they are localized within the pericentriolar material; their abundance, as judged by centrosome labelling, changes significantly during the cell cycle, the maximum being observed at the pole of the metaphasic spindle; in Taxol‐treated cells where the centrosome is no longer acting as a nucleating center, they redistribute at one end of the microtubule arrays in both mitotic and interphasic cells, as expected for nucleating, or capping, proteins. All these properties are compatible with their involvement in microtubule nucleation.

The diversity of microtubular networks was analyzed in quail oviduct and in Paramecium cells using conventional and confocal immunofluorescence as well as pre- and post-embedding EM immunocytochemistry with a variety of anti-tubulin antibodies. The 6-11B-1 monoclonal antibody, specific for the post-translational acetylation of Lys 40 of alpha-tubulin, and a polyclonal antibody raised against Paramecium axonemal tubulin (anti-PA tubulin antibody) both decorated stable microtubular arrays in Paramecium ie ciliary axonemes and a set of microtubular bundles associated with the cortex, suggesting that the two antibodies may be directed against the same epitope. However, several differences in the immunocytological patterns yielded by each antibody on the two cell types were evident. For example, in quail, as in all other Metazoa, the anti-PA tubulin antibody only decorated axonemes enclosed in normal ciliary membrane while it was unreactive on cytoplasmic tubulins. Immunoblotting of peptide maps of axonemal tubulins demonstrated that the epitopes of the two antibodies were indeed completely different. Double immunolabelling of dividing paramecia using a universal anti-tubulin antibody and the anti-PA tubulin one revealed that all newly assembled microtubular arrays were first detected by the universal antibody and, only shortly afterwards, by the anti-PA tubulin one. This provided a strong indication that the anti-PA tubulin antibody is directed against a post-translational modification taking place on already assembled microtubules (MTs) (as previously known to be the case for acetylation and detyrosination). In taxol-treated quail cells undergoing ciliogenesis, massive assembly of MTs and even axonemes occurred in the cytoplasm. These MTs were not decorated by the anti-PA tubulin antibody however, suggesting that in Metazoa the post-translational modification can only take place within the ciliary lumen. The present work provides one further mechanism for generating MT immunological and biochemical diversity post-translationally; this may account for the high multiplicity of tubulin isoforms observed in ciliates which contain very little if any genetic diversity of tubulin genes.