Making microtubules and mitotic spindles in cells without functional centrosomes

Centrosomes are considered to be the major sites of microtubule nucleation in mitotic cells (reviewed in ), yet mitotic spindles can still form after laser ablation or disruption of centrosome function . Although kinetochores have been shown to nucleate microtubules, mechanisms for acentrosomal spindle formation remain unclear. Here, we performed live-cell microscopy of GFP-tubulin to examine spindle formation in Drosophila S2 cells after RNAi depletion of either gamma-tubulin, a microtubule nucleating protein, or centrosomin, a protein that recruits gamma-tubulin to the centrosome. In these RNAi-treated cells, we show that poorly focused bipolar spindles form through the self-organization of microtubules nucleated from chromosomes (a process involving gamma-tubulin), as well as from other potential sites, and through the incorporation of microtubules from the preceding interphase network. By tracking EB1-GFP (a microtubule-plus-end binding protein) in acentrosomal spindles, we also demonstrate that the spindle itself represents a source of new microtubule formation, as suggested by observations of numerous microtubule plus ends growing from acentrosomal poles toward the metaphase plate. We propose that the bipolar spindle propagates its own architecture by stimulating microtubule growth, thereby augmenting the well-described microtubule nucleation pathways that take place at centrosomes and chromosomes.     

Purification of mitotic spindles from cultured human cells

In eukaryotes, both chromosome segregation and the determination of the cell division cleavage plane depend on the mitotic spindle apparatus. Spindle malfunctioning can lead to chromosome mis-segregation and cytokinesis defects and hence result in aneuploidy. Thus, the understanding of the structure and function of mitotic spindles is of interest not only from the perspective of basic science, but has implications also for human health and disease. Until recently, this complex microtubule-based structure was studied mainly by cell biological techniques in mammalian cells, by biochemical assays in Xenopus egg extracts, and by genetic approaches in genetically tractable organisms such as yeast, flies, and nematodes. With the rapid development of mass spectrometry and its increasing application to biological problems, it has become possible to subject highly complex structures, such as the mitotic spindle apparatus, to proteomics approaches. Such studies require the isolation of the mitotic spindle, or its substructures, in sufficient amounts and free of excessive contaminants. A number of methods for the isolation of mitotic spindles from mammalian tissue culture cells have been developed in the past. We have compared these methods and found that protocols based on the stabilization of microtubules by taxol were most efficient and reproducible. Here, we describe the further optimization of a taxol-based method, originally developed by Zieve and Solomon [Cell 28 (1982) 233-242], and its application to the isolation of human mitotic spindles at a scale suitable for mass spectrometric analysis [G. Sauer, R. Korner, A. Hanisch, A. Ries, E.A. Nigg, H.H.W. Sillje, Mol. Cell. Proteomics 4 (2005) 35-43].     

Identification of the 190 kD microtubule-associated protein in cultured fibroblasts and its association with interphase and mitotic microtubules

We previously investigated the biochemical characteristics of microtubule-associated proteins (MAPs) of the adrenal medulla and adrenal cortex and found that they contain a new kind of MAP with a molecular weight of 190,000 (190 kD MAP) as a major species (Kotani, S., H. Murofushi, S. Maekawa, C. Sato, and H. Sakai. Eur. J. Biochem. 156, 23-29, 1986). We now have used an affinity purified anti-(190 kD MAP) antibody and show by indirect immunofluorescent microscopy the association of this MAP with microtubules in situ in TIG-3 cells (human embryonic lung fibroblasts). The 190 kD MAP was present along the interphase and mitotic microtubules, and there was no marked difference between the staining pattern with anti-tubulin and that with anti-(190 kD MAP) antibodies, evidence that the localization of 190 kD MAP is not restricted to the subset of microtubules. We also isolated MAPs from TIG-3 cells and identified their 190 kD MAP as a major heat-stable component. Several other unidentified polypeptides were recovered in the MAP fraction specifically.     

Localisation of the major high-molecular-weight protein on microtubules in vitro and in cultured cells

An antiserum has been produced which is specific for the major high-molecular-weight protein (HMW) associated with pig brain microtubules assembled in vitro. The HMW protein can be localised on the surface of brain microtubules assembled in vitro as a corase helix by using a peroxidase labelling technique. In cultured ovarian granulosa cells, by using indirect immunofluorescence, the HMW is present on cold- and colchicine-sensitive structures which have an intracellular distribution similar to microtubules. It seems likely, therefore, that the brain protein is representative of a class of proteins associated with non-neuronal cytoplasmic microtubules.     

Non-random association of mitotic cells from cultured human lymphocyte suspensions.

By using chromosome markers in two separate cell lines of a human dispermic chimera, it was shown that 4.9% of metaphases in suspensions of PHA-cultured lymphocytes were paired because of non-random factors. A similar amount of metaphase pairing occurred in cultured lymphocytes of normal donors, and evidence from the relative mitotic cycles of the paired cells indicated that some of this pairing was non-random. Such non-random pairing could be a source of bias in cell kinetic and other studies involving metaphase cells in lymphocyte cultures.     

Axin localizes to mitotic spindles and centrosomes in mitotic cells

Wnt signaling plays critical roles in cell proliferation and carcinogenesis. In addition, numerous recent studies have shown that various Wnt signaling components are involved in mitosis and chromosomal instability. However, the role of Axin, a negative regulator of Wnt signaling, in mitosis has remained unclear. Using monoclonal antibodies against Axin, we found that Axin localizes to the centrosome and along mitotic spindles. This localization was suppressed by siRNA specific for Aurora A kinase and by Aurora kinase inhibitor. Interestingly, Axin over-expression altered the subcellular distribution of Plk1 and of phosphorylated glycogen synthase kinase (GSK3β) without producing any notable changes in cellular phenotype. In the presence of Aurora kinase inhibitor, Axin over-expression induced the formation of cleavage furrow-like structures and of prominent astral microtubules lacking midbody formation in a subset of cells. Our results suggest that Axin modulates distribution of Axin-associated proteins such as Plk1 and GSK3β in an expression level-dependent manner and these interactions affect the mitotic process, including cytokinesis under certain conditions, such as in the presence of Aurora kinase inhibitor.     

Analysis of stable protein methylation in cultured cells.

It was demonstrated recently that substrates for protein N-methyltransferases (J. Najbauer and D. W. Aswad, 1990, J. Biol. Chem. 265, 12,717-12,721) and protein carboxyl methyltransferases (J. Najbauer, B. A. Johnson, and D. W. Aswad, 1991, Anal. Biochem. 197, 412-420) accumulate when rat PC12 cells are cultured in the presence of the methylation inhibitor, adenosine dialdehyde. In the present report, we have further characterized this phenomenon in PC12 cells and in two other, widely used cell types. Adenosine dialdehyde was found to increase the methyl-accepting capacity of proteins in human skin fibroblasts and mouse Sp2/0 myeloma cells. However, both the level of methyl incorporation in untreated cells and the amount of stimulation afforded by inhibitor treatment were substantially lower in these cells than in PC12 cells. All three cell lines accumulated methyl acceptor(s) at 17-21 kDa. The PC12 cells and the fibroblasts also exhibited stimulation of three apparently similar proteins in the 33- to 38-kDa region, where several arginine-methylated proteins involved in RNA processing would be expected. The optimal conditions for methylation of PC12 cell extracts with regard to pH, time of methylation, and S-[methyl-3H]adenosyl-L-methionine concentration were characterized. Increased methyl incorporation was detected after adenosine dialdehyde treatments as short as 2 h, and methylation of most substrates continued to increase as the time of treatment was extended to 72 h. The kinetics of accumulation varied from substrate to substrate. Fluorograms of two-dimensional gels of extracts from untreated PC12 cells incubated in the presence of S-[methyl-3H]adenosyl-L-methionine revealed patterns of methyl incorporation similar to those of treated cells, but longer exposure times were necessary (e.g., 35 days vs 7 days). These findings suggest that the inhibitor treatment works mainly by inhibiting the post- or cotranslational methylation of a "normal" array of cellular proteins.     

Poleward microtubule flux mitotic spindles assembled in vitro

In the preceding paper we described pathways of mitotic spindle assembly in cell-free extracts prepared from eggs of Xenopus laevis. Here we demonstrate the poleward flux of microtubules in spindles assembled in vitro, using a photoactivatable fluorescein covalently coupled to tubulin and multi-channel fluorescence videomicroscopy. After local photoactivation of fluorescence by UV microbeam, we observed poleward movement of fluorescein-marked microtubules at a rate of 3 microns/min, similar to rates of chromosome movement and spindle elongation during prometaphase and anaphase. This movement could be blocked by the addition of millimolar AMP-PNP but was not affected by concentrations of vanadate up to 150 microM, suggesting that poleward flux may be driven by a microtubule motor similar to kinesin. In contrast to previous results obtained in vivo (Mitchison, T. J. 1989. J. Cell Biol. 109:637-652), poleward flux in vitro appears to occur independently of kinetochores or kinetochore microtubules, and therefore may be a general property of relatively stable microtubules within the spindle. We find that microtubules moving towards poles are dynamic structures, and we have estimated the average half-life of fluxing microtubules in vitro to be between approximately 75 and 100 s. We discuss these results with regard to the function of poleward flux in spindle movements in anaphase and prometaphase.     

Intracellular hyaluronan in arterial smooth muscle cells: association with microtubules, RHAMM, and the mitotic spindle.

Although considered a pericellular matrix component, hyaluronan was recently localized in the cytoplasm and nucleus of proliferating cells, supporting earlier reports that hyaluronan was present in locations such as the nucleus, rough endoplasmic reticulum, and caveolae. This suggests that it can play roles both inside and outside the cell. Hyaluronan metabolism is coupled to mitosis and cell motility, but it is not clear if intracellular hyaluronan associates with cytoskeletal elements or plays a structural role. Here we report the distribution of intracellular hyaluronan, microtubules, and RHAMM in arterial smooth muscle cells in vitro. The general distribution of intracellular hyaluronan more closely resembled microtubule staining rather than actin filaments. Hyaluronan was abundant in the perinuclear microtubule-rich areas and was present in lysosomes, other vesicular structures, and the nucleolus. Partially fragmented fluorescein-hyaluronan was preferentially translocated to the perinuclear area compared with high-molecular-weight hyaluronan. In the mitotic spindle, hyaluronan colocalized with tubulin and with the hyaladherin RHAMM, a cell surface receptor and microtubule-associated protein that interacts with dynein and maintains spindle pole stability. Internalized fluorescein-hyaluronan was also seen at the spindle. Following telophase, an abundance of hyaluronan near the midbody microtubules at the cleavage furrow was also noted. In permeabilized cells, fluorescein-hyaluronan bound to RHAMM-associated microtubules. These findings suggest novel functions for hyaluronan in cellular physiology.     

STOP (stable-tubule-only-polypeptide) is preferentially associated with the stable domain of axonal microtubules

Axonal microtubules consist of two distinct domains that differ in tyrosinated-tubulin staining. One domain stains weakly for tyrosinated-tubulin, while the other stains strongly, and the transition between these domains is abrupt; the tyrosinated-tubulin-poor domain is at the minus end of the microtubule, and the tyrosinated-tubulin-rich domain extends from the plus end of the tyrosinated-tubulin-poor domain to the end of the microtubule. The tyrosinated-tubulin-poor domain is drug- and cold-stable, whereas the tyrosinated-tubulin-rich domain is drug-labile, but largely cold-stable. STOP (stable-tubule-only-polypeptide) has potent microtubule stabilizing activity, and may contribute to the cold and drug stability of axonal microtubules. To evaluate this possibility, we examined STOP association with the different types of microtubule polymer in cultured sympathetic neurons. By immunofluorescence, STOP is present in the cell body and throughout the axon; axonal staining declines progressively in the distal portion of the axon, and reaches lowest levels in the growth cone. Growth cone microtubules, which are drug and cold labile, do not stain detectably for STOP. To examine individual axonal microtubules for STOP, we used a procedure that causes microtubules to splay out from the main axonal array so that they can be visualized for relatively long distances along their length. Both tyrosinated-tubulin-rich and tyrosinated-tubulin-poor polymer stain for STOP, but STOP is several-fold more concentrated on tyrosinated-tubulin-poor polymer than on tyrosinated-tubulin-rich polymer. These results are consistent with STOP dependent stabilization of axonal microtubules, with the difference between cold-stable polymer versus cold- + drug-stable polymer determined by the amount of STOP on the polymer.     

Double fluorescent staining for the separate demonstration of chromosomes and microtubules in mitotic cells in vitro.

A simple fluorescent method for double staining of mitotic cells using a rhodamine B indirect immunofluorescent method for tubulin and the DNA-specific fluorescent dye Hoechst 33258 for nuclei and chromosomes is described. This procedure enables one through the use of appropriate excitation filters to view at will either chromosomes and nuclei or tubulin within the same cell.     

Physiological and ultrastructural analysis of elongating mitotic spindles reactivated in vitro

We have developed a simple procedure for isolating mitotic spindles from the diatom Stephanopyxis turris and have shown that they undergo anaphase spindle elongation in vitro upon addition of ATP. The isolated central spindle is a barrel-shaped structure with a prominent zone of microtubule overlap. After ATP addition greater than 75% of the spindle population undergoes distinct structural rearrangements: the spindles on average are longer and the two half-spindles are separated by a distinct gap traversed by only a small number of microtubules, the phase-dense material in the overlap zone is gone, and the peripheral microtubule arrays have depolymerized. At the ultrastructural level, we examined serial cross-sections of spindles after 1-, 5-, and 10-min incubations in reactivation medium. Microtubule depolymerization distal to the poles is confirmed by the increased number of incomplete, i.e., c-microtubule profiles specifically located in the region of overlap. After 10 min we see areas of reduced microtubule number which correspond to the gaps seen in the light microscope and an overall reduction in the number of half-spindle microtubules to about one-third the original number. The changes in spindle structure are highly specific for ATP, are dose-dependent, and do not occur with nonhydrolyzable nucleotide analogues. Spindle elongation and gap formation are blocked by 10 microM vanadate, equimolar mixtures of ATP and AMPPNP, and by sulfhydryl reagents. This process is not affected by nocodazole, erythro-9-[3-(2-hydroxynonyl)]adenine, cytochalasin D, and phalloidin. In the presence of taxol, the extent of spindle elongation is increased; however, distinct gaps still form between the two half-spindles. These results show that the response of isolated spindles to ATP is a complex process consisting of several discrete steps including initiation events, spindle elongation mechanochemistry, controlled central spindle microtubule plus-end depolymerization, and loss of peripheral microtubules. They also show that the microtubule overlap zone is an important site of ATP action and suggest that spindle elongation in vitro is best explained by a mechanism of microtubule-microtubule sliding. Spindle elongation in vitro cannot be accounted for by cytoplasmic forces pulling on the poles or by microtubule polymerization.     

BimC motor protein KLP61F cycles between mitotic spindles and fusomes in Drosophila germ cells.

KLP61F in Drosophila is a member of the BimC family of kinesins and, as for other family members [1], is required for spindle assembly [2] [3]. KLP61F is a bipolar homotetramer that cross-links spindle microtubules [4]. It is not known, however, whether the function of KLP61F is dedicated to mitosis or whether KLP61F interacts exclusively with microtubules. Previous work suggested that KLP61F functions during interphase in proliferating germ cells [3]. Cytokinesis is incomplete in germ cells and a branched cortical structure known as a fusome extrudes through intercellular bridges called ring canals. Here I show that, in germ cells, KLP61F cycles between spindles during mitosis and fusomes during interphase. Inspection of fusome-deficient hu-li tai shao (hts) mutants indicated that KLP61F gains fusome-dependent interactions near telophase that mediate its incorporation into these structures. KLP61F proved to be maintained in fusomes by microtubule-independent, detergent-resistant interactions. Inspection of KLP61F mutants indicated that KLP61F is required to recruit fusome material to spindle midbodies near telophase and for normal fusome organization. These observations suggest that KLP61F is bifunctional in germ cells, with microtubule-dependent functions in spindle assembly and microtubule-independent functions in fusome organization. Cytological analyses with antibodies against phosphorylated Eg5 peptide [4] suggest that cycling of KLP61F might reflect phosphorylation.     

Specific association of an M-phase kinase with isolated mitotic spindles and identification of two of its substrates as MAP4 and MAP1B.

Isolated mammalian (Chinese hamster ovary [CHO]) metaphase spindles were found to be enriched in a histone H1 kinase whose activity was mitotic-cycle dependent. Two substrates for the kinase were identified as MAP1B and MAP4. Partially purified spindle kinase retained activity for the spindle microtubule-associated proteins (MAPs) as well as brain and other tissue culture MAPs; on phosphorylation, spindle MAPs exhibited increased immunoreactivity with MPM-2, a monoclonal antibody specific for a subset of mitotic phosphoproteins. Immunofluorescence using an anti-thiophosphoprotein antibody localized in vitro phosphorylated spindle proteins to microtubule fibers, centrosomes, kinetochores, and midbodies. The fractionated spindle kinase was reactive with anti-human p34cdc2 antibodies and with an anti-human cyclin B but not an anti-human cyclin A antibody. We conclude that spindle MAPs undergo mitotic cycle-dependent phosphorylations in vivo and associate with a kinase that remains active on spindle isolation and may be related to p34cdc2.     

A microtubule-associated protein antigen unique to mitotic spindle microtubules in PtK1 cells

Microtubule-associated proteins (MAPs) that copurify with tubulin through multiple cycles of in vitro assembly have been implicated as regulatory factors and effectors in the in vivo activity of microtubules. As an approach to the analysis of the functions of these molecules, a collection of lymphocyte hybridoma monoclonal antibodies has been generated using MAPs from HeLa cell microtubule protein as antigen. Two of the hybridoma clones secrete IgGs that bind to distinct sites on what appears to be a 200,000-dalton polypeptide. Both immunoglobulin preparations stain interphase and mitotic apparatus microtubules in cultured human cells. One of the clones (N-3B4.3.10) secretes antibody that reacts only with cells of human origin, while antibody from the other hybridoma (N-2B5.11.2) cross-reacts with BSC and PtK1 cells, but not with 3T3 cells. In PtK1 cells the N-2B5 antigen is associated with the microtubules of the mitotic apparatus, but there is no staining of the interphase microtubule array; rather, the antibody stains an ill-defined juxtanuclear structure. Further, neither antibody stains vinblastine crystals in either human or marsupial cells at any stage of the cell cycle. N-2B5 antibody microinjected into living PtK1 cells binds to the mitotic spindle, but does not cause a rapid dissolution of either mitotic or interphase microtubule structures. When injected before the onset of anaphase, however, the N-2B5 antibody inhibits proper chromosome partition in mitotic PtK1 cells. N-2B5 antibody injected into interphase cells causes a redistribution of MAP antigen onto the microtubule network.     

Specific visualization of tubulin-containing structures in tissue culture cells by immunofluorescence. Cytoplasmic microtubules, vinblastine-induced paracrystals, and mitotic figures.

Antibody against tubulin from the outer doublets of sea urchin sperm flagella reacts with tubulin-containing structures in mammalian cells. Thus cytoplasmic microtubules, vinblastine-induced paracrystals and the full spectrum of mitotic figures can be visualized by immunofluorescence. These results show that the tubulin structure has been highly conserved during evolution.     

Dynamic and stable populations of microtubules in cells

Using a new immunocytochemical technique, we have visualized the spatial arrangement of those microtubules in cells that are stable to biotin-tubulin incorporation after microinjection. Cells fixed at various periods of time after injection were exposed to antibody to biotinylated tubulin and several layers of secondary antibodies; these layers prevented reaction of biotin-containing microtubules with antitubulin antibodies. The microtubules that had not incorporated biotin-tubulin could then be stained with anti-tubulin and a fluorescent secondary antibody. In BSC1 cells, most microtubules in the cell exchange with a half-time of 10 min. A separate population of microtubules can be detected, using the above techniques, that are stable to exchange for 1 h or more; these have a characteristic pericentrosomal spatial arrangement as compared to the majority of dynamic microtubules. Unlike the dynamic microtubules, most of the stable microtubules are nongrowing. The average BSC-1 cell contains approximately 700 microtubules: approximately 500 growing at 4 micron min-1, 100 shrinking at approximately 20 micron min-1, and approximately 100 that are relatively more stable to exchange. The potential significance of these stable microtubules is discussed.