E. M. Chudinova

Disruption of microtubules inhibits cytoplasmic ribonucleoprotein stress granule formation.

Stress granules are RNP-containing particles arising in the cytoplasm in response to environmental stress. They are dynamic structures assembling and disassembling in the cytoplasm very rapidly. We have studied whether the cytoskeleton is involved in the formation of stress granules. Stress granules were induced in CV-1 cells by sodium arsenate treatment and visualized by immunofluorescent staining with antibodies either to the p170 subunit of eIF3 or to poly(A)-binding protein. Treatment with sodium arsenate for 30-120 min led to assembling of stress granules in a majority of CV-1 cells. Disruption of MT array with nocodazole treatment abolished arsenate-induced formation of stress granules. A similar effect was induced by the microtubule-depolymerizing drug vinblastine, though the influence of the microtubule-stabilizing drug paclitaxel was opposite. Nocodazole treatment did not prevent arsenate-induced phosphorylation of the eIF-2alpha factor, essential for stress granule formation, suggesting that the presence of intact MT array is required for granule assembly. Unexpectedly, treatment of cells with the actin filament-disrupting drug latrunculin B slightly enhanced stress granule formation. We propose that stress granule formation is microtubule-dependent process and likely is facilitated by the motor protein-driven movement of individual stress granule components (e.g., mRNP) along microtubules.

Microtubules govern stress granule mobility and dynamics

Stress granules (SGs) are ribonucleoprotein (RNP)-containing assemblies that are formed in the cytoplasm in response to stress. Previously, we demonstrated that microtubule depolymerization inhibited SG formation. Here, we show that arsenate-induced SGs move throughout the cytoplasm in a microtubule-dependent manner, and microtubules are required for SG disassembly, but not for SG persistence. Analysis of SG movement revealed that SGs exhibited obstructed diffusion on an average, though sometimes SGs demonstrated rapid displacements. Microtubule depolymerization did not influence preformed SG number and size, but significantly reduced the average velocity of SG movement, the frequency of quick movement events, and the apparent diffusion coefficient of SGs. Actin filament disruption had no effect on the SG motility. In cycloheximide-treated cells SGs dissociated into constituent parts that then dissolved within the cytoplasm. Microtubule depolymerization inhibited cycloheximide-induced SG disassembly. However, microtubule depolymerization did not influence the dynamics of poly(A)-binding protein (PABP) in SGs, according to FRAP results. We suggest that the increase of SG size is facilitated by the transport of smaller SGs along microtubules with subsequent fusion of them. At least some protein components of SGs can exchange with the cytoplasmic pool independently of microtubules.

Translation initiation factor eIF3 probably binds with microtubules in mammalian cells

Association of the translation apparatus with the cytoskeleton is essential for its transportation within the cell and probably also for translation regulation. Very little is known about the involvement of particular proteins of this association. A polypeptide homologous with the heavy chain of translation initiation factor eIF3 p170 was found earlier in a microtubule preparation from adrenal cells. Antibody A167 directed against the recombinant fragment of p170 has been generated to study eIF3 interaction with microtubules in mammalian cells. This antibody was shown to recognize a single 170-kDa polypeptide in eIF3 preparations as well as in homogenates of various cell types. A167 allowed detection of the 170-kDa polypeptide in microtubule preparation from bovine brain and confirmation of its presence in microtubule preparations from adrenal cells. As shown by immunofluorescence microscopy using A167, the 170-kDa polypeptide is mainly located in the endoplasm within numerous small and some large granules. Cell treatment with cycloheximide resulted in growth and clustering of the large granules, and partial antigen redistribution along intracellular microtubules. These new experimental data indicate that mammalian translation factor eIF3 may bind with microtubules.

RNP stress-granule formation is inhibited by microtubule disruption.

Stress-granules (SGs) are dense micrometre sized particles that assemble in the cytoplasm of animal and plant cells in response to cell stress: heat shock, drug application, etc. (Nover et al., 1989). SGs contain components of the translation initiation machinery, poly(A)RNA, some RNA-binding proteins, 40S ribosomal subunit and initiation factors eIF3, eIF4E and eIF4G (Kedersha et al., 2000, 2002). They are defined as dynamic non-membrane cellular compartments existing in dynamic equilibrium with polysomes. Surprisingly, SG formation takes minutes. It is not clear how hundreds of large molecules can find each other so quickly in a large cell volume. We speculated that SG constituents could bind or move along cytoskeletal elements, and this binding/moving could increase the probability of molecules to congregate. Interaction of SGs with microtubules (MTs), actin filaments and intermediate filaments and the dependence of SG formation on the integrity of the MT and actin network were studied in this work. Treatment of cells with sodium arsenate is often used as an approach for SG induction (Kedersha et al., 2000, 2002). Arsenate treatment does not influence cell MTs (Fig. 1A,C) or actin filaments (data not shown). To study the interaction of SGs and cytoskeleton, we treated cultured mammalian cells CV-1 (green monkey kidney fibroblasts) with 2.8 mM sodium arsenate dissolved in growth medium, fixed them with glutaraldehyde, and immunostained with antibody A167 (to p167 subunit of eIF3 factor, described in Shanina et al., 2001) and commercial antibodies to cytoskeletal structures. Immunostained A167 cells demonstrated diffuse antigen distribution or sometimes very small granules in the cytoplasm (Fig. 1E). If cells were incubated with arsenate, the antigen was concentrated in SGs within 30 min (Fig. 1G). Usually, the size of SGs was 0.5–2 μm, and the average number of SGs per cell was 110. SGs were distributed in the middle part of cytoplasm like a kind of a ring with the nucleus and putative Golgi area inside. The cell distribution of SGs did not match the distribution of vimentin filaments and actin stress-fibers. Some correlation of SG and MT localization was found, though SGs are too large to be precisely localized on particular MTs (Fig. 1A,E). The influence of the cytoskeleton on SG formation was studied by the elimination of actin filaments or MTs, with specific drugs, before arsenate application. If cells were treated with latrunculin B (1 μg/ml, 15 min), their actin stress-fibers were destroyed. However, arsenate induced formation of SGs in such cells and these SGs were indistinguishable from control ones. If cells were treated with the MT-depolymerizing drug nocodazole (1 μg/ml, 14 h) (Fig. 1B,F), SGs did not appear in the cytoplasm. Arsenate application (30 min) in this case did not induce SG assembly (Fig. 1B,D,F,H). We have also shown that partial depolymerization of MTs with low doses of nocodazole or brief treatment with this drug resulted in the assembly of small SGs. These small SGs did not organize a ring but were randomly distributed in the cytoplasm. The same results were obtained with vinblastine treatment of cells. Nocodazole removal induced rapid formation of MTs and SGs within 5–10 min. In cells treated with the MTstabilizing drug paclitaxel, MTs were arranged as disordered bundles. The addition of arsenate resulted in rapid formation of small SGs growing very large by aggregation after 3–4 h. Our experimental data indicate that SG formation directly depends on cell MTs and that actin filaments have no influence on the same. So, cell response to stress can be changed with MT-disrupting drugs. We suppose that interaction of some SG components, mRNA (Jansen, 1999) and eIF3 (Shanina et al., 2001) particularly, can greatly increase the probability of interaction of these components 1 These authors contributed equally to the work. * Corresponding author. Tel.: +7-95-135-9786; fax: +7-95-9393181. E-mail address: elena_nad@lycos.com (E.M. Nadezhdina). Cell Biology International 27 (2003) 207–208 Cell Biology International

Nuclear Localization Signals in p170, the Large Subunit of Translation Initiation Factor 3

Apart from their role in translation, eukaryotic translation factors or their individual subunits may perform other functions, in particular, regulating nuclear processes. Primary structure analysis revealed four potential nuclear localization signals (NLS) in the human eIF3 large subunit, p170. NLS were tested for ability to direct p170 into the nucleus. For this purpose, cDNAs coding for p170 fragments fused with the green fluorescent protein were expressed in CV-1 and Cos-1 cultured monkey cells. The location of the expression product was studied by fluorescence microscopy. At least two of the four putative bipartite NLS proved to direct the corresponding p170 fragments into the nucleus. Larger p170 fragments with the same NLS were retained in the cytoplasm. It was assumed that, with the help of some specific factors or after limited proteolysis, p170 enters the nucleus and participates in regulating genome expression. Alternatively, the cytoplasmic function of p170 might be regulated via a reversible binding of integrins to NLS.

Cellular acidosis inhibits assembly, disassembly, and motility of stress granules.

Stress granules (SGs) are large ribonucleoprotein (RNP)-containing particles that form in cytoplasm in response to a variety of acute changes in the cellular environment. One of the general parameters of the cell environment is pH. In some diseases, as well as in muscle fatigue, tissue acidosis occurs, leading to decrease in intracellular pH. Here we studied whether decrease in pH causes the formation of SGs in cultured animal cells, whether it affects the formation of the SGs under the action of arsenite and, if such effects occur, what are the mechanisms of the influence of acidosis. Acidosis was simulated by decreasing the pH of the culture medium, which acidified the cytoplasm. We found that medium acidification to pH 6.0 in itself did not cause formation of SGs in cells. Moreover, acidification prevented the formation of SGs under treatment with sodium arsenite or sodium arsenite together with the proteasome inhibitor MG132, and it inhibited the dissociation of preformed SGs under the influence of cycloheximide. We established that pH decrease did not affect the phosphorylation of eIF2α that occurs under the action of sodium arsenite, and even caused such phosphorylation by itself. We also found that the velocity of SG motion in cytoplasm at acidic pH was very low, and the mobile fraction of SG-incorporated PABP protein revealed by FRAP was decreased. We suppose that acidic pH impairs biochemical processes favoring assembly of RNPs in stress conditions and RNP dissociation on the termination of stress. Thus, in acidosis the reaction of the cellular translation apparatus to stress is modified.

Interactions between the Translation Machinery and Microtubules

Microtubules are components of eukaryotic cytoskeleton that are involved in the transport of various components from the nucleus to the cell periphery and back. They also act as a platform for assembly of complex molecular ensembles. Ribonucleoprotein (RNP) complexes, such as ribosomes and mRNPs, are transported over significant distances (e.g. to neuronal processes) along microtubules. The association of RNPs with microtubules and their transport along these structures are essential for compartmentalization of protein biosynthesis in cells. Microtubules greatly facilitate assembly of stress RNP granules formed by accumulation of translation machinery components during cell stress response. Microtubules are necessary for the cytoplasm-to-nucleus transport of proteins, including ribosomal proteins. At the same time, ribosomal proteins and RNA-binding proteins can influence cell mobility and cytoplasm organization by regulating microtubule dynamics. The molecular mechanisms underlying the association between the translation machinery components and microtubules have not been studied systematically; the results of such studies are mostly fragmentary. In this review, we attempt to fill this gap by summarizing and discussing the data on protein and RNA components of the translation machinery that directly interact with microtubules or microtubule motor proteins.

MAST-like protein kinase IREH1 from Arabidopsis thaliana co-localizes with the centrosome when expressed in animal cells

AbstractMain conclusionThe similarity of IREH1 (Incomplete Root Hair Elongation 1) and animal MAST kinases was confirmed; IREH1cDNA was cloned while expressing in cultured animal cells co-localized with the centrosome. In mammals and fruit flies, microtubule-associated serine/threonine-protein kinases (MAST) are strongly involved in the regulation of the microtubule system. Higher plants also possess protein kinases homologous to MASTs, but their function and interaction with the cytoskeleton remain unclear. Here, we confirmed the sequence and structural similarity of MAST-related putative protein kinase IREH1 (At3g17850) and known animal MAST kinases. We report the first cloning of full-length cDNA of the IREH1 from Arabidopsis thaliana. Recombinant GFP-IREH1 protein was expressed in different cultured animal cells. It revealed co-localization with the centrosome without influencing cell morphology and microtubule arrangement. Structural N-terminal region of the IREH1 molecule co-localized with centrosome as well.

[Is the microtubule disruption-induced alteration of peroxide concentration a factor inhibiting the assembly of ribonucleoprotein stress granules?].

It has been examined whether the destruction of cell microtubules affects the increase in the intracellular hydrogen peroxide concentration caused by sodium arsenite, which induces the formation of stress ribonucleoprotein granules. As expected, sodium arsenite caused a 50% increase in hydrogen peroxide concentration in HeLa cells; on the other hand, another stress granule inducer tert-butylhydroquinone did not affect the peroxide concentration. The disruption of microtubules by nocodazole or vinblastine also resulted in some increase in the intracellular peroxide concentration, and the microtubule stabilization by taxol did not affect it. The combined treatment of cells with arsenite and antimicrotubule drugs caused an additive effect, and the peroxide concentration increased twice or more. Thus, the inhibition of stress granule formation after microtubule disruption cannot be explained by a decrease in peroxide concentration as compared with the affect of arsenite.