Disruption of microtubules inhibits cytoplasmic ribonucleoprotein stress granule formation

Available online at www.sciencedirect.com R

Experimental Cell Research 290 (2003) 227–233

www.elsevier.com/locate/yexcr

Disruption of microtubules inhibits cytoplasmic ribonucleoprotein stress granule formation Pavel A. Ivanov,a,b,1 Elena M. Chudinova,a,1 and Elena S. Nadezhdinaa,b,* a b

Institute of Protein Research, Russian Academy of Sciences, 142292 Pushchino, Moscow Region, Russian Federation A.N. Belozersky Institute of Physico-Chemical Biology, Moscow State University, 119899 Moscow, Russian Federation Received 9 January 2003, revised version received 17 April 2003

Abstract 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-2␣ 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. © 2003 Elsevier Inc. All rights reserved. Keywords: Translation initiation factor-3; eIF3a; PABP; Stress granules; Arsenate; Microtubule; Actin; Nocodazole; Paclitaxel; Latrunculin B

Introduction Environmental stress provokes the formation of phase dense particles in the cytoplasm of both plant and animal cells [1– 6]. These particles are called stress granules (SGs)2 and consist of RNA and proteins. It has been established that SG formation is a result of stress-induced arrest of translation, probably, in response to the phosphorylation of translation initiation factor eIF2␣. SGs accumulate untranslated mRNAs together with RNA-binding proteins HuR (ELAV), TIAR, poly(A)-binding protein (PABP). They also * Corresponding author. Fax: ⫹7-095-939-3181. E-mail address: [email protected]; elena [email protected] (E.S. Nadezhdina). 1 These authors contributed to the work equally. 2 Abbreviations used: Ab, antibody; DMSO, dimethyl sulfoxide; eEF, eukaryotic elongation factor; eIF, eukaryotic initiation factor; MT, microtubule; PABP, poly(A)-binding protein; SG, stress granules. 0014-4827/$ – see front matter © 2003 Elsevier Inc. All rights reserved. doi:10.1016/S0014-4827(03)00290-8

contain most components of the 48S preinitiation complex (i.e., small, but not large, ribosomal subunits, translation initiation factors eIF3, eIF4E, and eIF4G) [6]. Factors eIF5 and eIF2 usually do not accumulate in SGs [4]. SGs are dynamic cellular microcompartments that influence the fate of the untranslated mRNAs accumulated in stressed cells [4 – 6]. The mRNAs shuttle between SGs and polysomes during stress, and polysome stabilization with emetine causes SGs dissociation [5]. In the cells recovering from stress, SG-associated mRNAs rapidly move to polysomes [2]. Moreover, photobleaching studies indicated that some of protein components of SGs (TIAR and PABP) are in constant and extremely rapid flux, despite the relatively stable appearance of individual SGs [5]. Therefore, formation and persistence of SGs depend on the movement of RNA and proteins in the cytoplasm and their accumulation in particular cytoplasmic foci. How can SG constituents (macromolecules) find each

228

P.A. Ivanov et al. / Experimental Cell Research 290 (2003) 227–233

other in particular sites in cytoplasm? SGs have no specific scaffold proteins or scaffold RNA. A scaffold function of TIAR proposed [5], however, has never been directly shown. Another problem for SG formation which involves rapid macromolecule movement is cytoplasm viscosity. The diffusion and partitioning of macromolecules and vesicles in the cytoplasm are highly restricted by steric hindrance and low-affinity binding interactions [7]. The cells usually resolve the macromolecule transportation problem by cytoskeleton-dependent active transport. The idea that transport along microtubules (MTs) facilitates coming together of cellular components was experimentally confirmed at least for membrane vesicles [8]. We have hypothesized that SGs could be arranged on the surface of cytoskeletal structures, which serve as a scaffold for SG organization, and that SG formation is facilitated by the movement of mRNP and/or some of SG proteins along MTs or actin filaments. It is known that at least some mRNPs associate with MTs [9, 10] and can move along microtubules driven by molecular motors [11,12]. Some other mRNP species were found to bind to actin filaments [13]. Our preliminary data indicated that translation initiation factor eIF3 could bind to MTs [14]. Translation elongation factor eEF1␣ interacts with both MTs and actin filaments [15–17]. The affinity of mRNP-binding protein p50 (YB-1) [18] and translation elongation factor eEF2 [19,20] to actin filaments was also demonstrated previously. In the present study we have demonstrated the influence of various MT-interacting drugs (nocodazole, vinblastin, paclitaxel) and actin-depolymerizing drug latrunculin B to SG formation in cultured cells under the action of sodium arsenate. We have shown that an intact MT system is essential for SG formation, though actin filaments are not required.

sia) was freshly dissolved in full culture medium at a 2.8 mM concentration for each experiment. Nocodazole, paclitaxel, and vinblastine (Sigma) were added to the culture medium from 10 mg/ml stock solutions in DMSO (ICN Pharmaceuticals). Latrunculin B (Calbiochem) was used as 2 mM stock solution in DMSO. Final drug concentrations and time of incubation are given in the figure legends. If nocodazole, paclitaxel, vinblastine, or latrunculin B treatment preceded arsenate addition, these drugs always were present in the medium during subsequent arsenate treatment. Cells were fixed with 0.5% glutaraldehyde in PBS, treated with sodium borohydride and 0.1% Nonidet P-40 detergent, and immunostained. For immunostaining with anti-PABP antibody, cells were fixed as described [4]. Pictures of immunostained cells were taken using an Axiophot microscope (Zeiss) equipped with a MicroMax CCD camera (Princeton Instruments), lens 40 ⫻ Planapo. The portion of SG-containing cells was scored under the microscope directly; at least 100 cells were analyzed in each experiment. The number of SGs was scored and their size was measured using camera-taken cell pictures.(3– 8) Cells in each of three independent experiments were analyzed for every experimental condition. The statistic analysis was made using the Microcal Origin program. For Western blot, cells were treated as indicated in the figure legends and lysed in sample buffer. The proteins were separated by electrophoresis in 6 –12% gradient gels, electroblotted to Hybond-C nitrocellulose membranes (Amersham Biosciences), blocked with 0,1% cold water fish gelatin (Sigma), and incubated subsequently with primary and secondary antibodies. Horseradish peroxidase conjugated with secondary antibodies was developed in 3,3⬘-diaminobenzinide/H2O2 solution.

Materials and methods

Results and discussion

Rabbit polyclonal antibody (Ab) A170-C was raised to the recombinant fragment of eIF3-A subunit (p170) as described in [14]. Mouse monoclonal anti-tubulin DM-1A and B-5-1-2 Abs, goat anti-mouse IgG Ab conjugated with TRITC, goat anti-rabbit IgG Ab conjugated with FITC, and phalloidin-TRITC were purchased from Sigma. Rabbit polyclonal Ab to PABP was kindly provided by Dr. I. N. Shatsky. Goat anti-mouse IgG and anti-rabbit IgG Abs conjugated with horseradish peroxidase were purchased from Imtek (Moscow, Russia). Ab to phosphorylated (Ser51) eIF-2␣ was purchased from Cell Signaling Technology (Beverly, MA), Cat. No 9721. CV-1 cells (green monkey kidney fibroblasts) were cultured in a medium containing 45% DMEM, 45% F12, 10% newborn calf serum (all ingredients were from Fischer), and 100 ␮g/ml of gentamicin (Sigma). Cells were plated on glass coverslips 1 day prior to experiments. Sodium arsenate (Na2HAsO4) (Reachim, Moscow, Rus-

SG formation in CV1 cells under arsenate action The eukaryotic initiation factor 3 large subunit, p170, has diffused distribution in the cytoplasm of cultured fibroblastlike CV-1 cells (Fig. 1A) [14]. Treatment of cells with 2.8 mM sodium arsenate (for 30 and 120 min) induced formation of multiple p170-contaning spots in the central part of cells (Fig. 1B, C, and F). We have suggested that these spots were SGs because it has been published previously that eIF3 is a regular component of SGs [4 – 6] and the distribution and shape of SGs [4 – 6] were very similar to the pattern we have observed. Similar multiple spots were revealed in the cytoplasm of arsenate-treated cells with anti-PABP antibody (Fig. 2A); the presence of PABP in SGs was also shown before [5]. The distribution of SGs in the cytoplasm was not random; most often SGs were arranged in a kind of circle surrounding the area occupied by the nucleus and putative Golgi region (Figs. 1B, C, and F and 2A). Arsenate

P.A. Ivanov et al. / Experimental Cell Research 290 (2003) 227–233

229

rupted MTs (by treating with nocodazole or vinblastine), increased number of MTs (by treating with paclitaxel), and disrupted actin filaments (by treating with latrunculin B). Stock solutions of anti-cytoskeleton drugs were dissolved in DMSO. To examine DMSO action on MTs and SG formation, we treated cells with pure DMSO at 1:500 v/v concentration (it was the final concentration of DMSO when 6.6 ␮M nocodazole was used for cell treatment). We found that DMSO did not influence the MT network (data not shown) and the total number of SGs in arsenate-treated cells (Table 1); however, the portion of cells containing SGs was slightly decreased (Table 1). In our further analysis we used DMSOtreated cells as a reference point. We cannot explain the influence of DMSO to SG formation. Effect of microtubule disruption (nocodazole or vinblastine treatment) on SG formation

Fig. 1. Stress granules and cytoskeletal structures in CV-1 cells treated with sodium arsenate. Double immunostaining with Ab A170-C which recognized translation initiation factor eIF3-A subunit p170 (A–C, E, F) and with anti-tubulin DM-1A Ab (D). (G) Rodamine-phalloidin staining. (A) Control cells; (B–G) cells after treatment with sodium arsenate for 30 min (B) or 120 min (C–G). (C, D and F, G) The same cells. Bars, 10 ␮m. The boxed region in C is shown increased in E. Small SGs are pointed with thin arrows, and large SGs with short thick arrows.

treatment did not influence either cellular MT system (Figs. 1D and 2B) or actin stress-fibers (Fig. 1G). The distribution of SGs in cells might correlate with the distribution of MTs (Figs. 1C and D; Fig. 2A and B). However, the density of MTs in the cell core was too high to determine precisely whether SGs were bound to MTs. A subpopulation of cells did not always contain SGs after treatment with arsenate even during 120 min (Table 1; Fig. 1C). We have estimated that under arsenate action the average number of SGs was about a hundred per cell containing SGs (Table 1). The size of SGs was variable (Fig. 1E). We measured the SG diameter and found that SGs ranged from 0.2 to more than 4 ␮m. A category of large SGs (larger than 1.0 ␮m) was defined for the special analysis. The bigger the SGs, the more difficult its formation through simple diffusion. We calculated that in CV1 cells treated with arsenate for 30 min large SGs represented about 10 –20% of all SGs (Table 1). Increasing the duration of arsenate treatment of cells caused an increase of the portion of large SGs up to 30 – 40% though the overall number of SGs per cell decreased (Table 1; Fig. 4). To confirm the idea of dependence of SG formation on cytoskeleton, we induced SG formation in cells with dis-

To depolymerize MTs, we treated cells with the anti-MT drug nocodazole. In control nocodazole-treated CV-1 cells MTs were depolymerized, and no SGs were formed: p170 was distributed diffusely (Fig. 3A and B). If cells were treated with arsenate (plus nocodazole) after 120 min of nocodazole treatment, we also did not see SGs in the cytoplasm of a majority of cells immunostained with A170-C Ab (Fig. 3C and D; Table 1). It indicated that either SG formation was strongly inhibited or SGs did not contain p170 (eIF3-A). To distinguish between these two possibilities we immunostained cells with anti-PABP Ab. PABP binds to poly(A) with a high affinity; poly(A)-mRNAs are the key component of SGs [4 – 6]. However, we did not observe PABP-containing granules in the cells treated with nocodazole and then with arsenate plus nocodazole (Fig. 2C

Fig. 2. Effect of nocodazole to stress granule formation in CV-1 cells. Double immunostaining with Ab to PABP (A, C) and with anti-tubulin DM-1A Ab (B, D). (A, B) Cells treated with arsenate for 30 min; (C, D) cells subsequently treated with 1.6 ␮M nocodazole for 120 min and then 30 min with arsenate plus 1.6 ␮M nocodazole. Bars, 10 ␮m.

230

P.A. Ivanov et al. / Experimental Cell Research 290 (2003) 227–233

Table 1 Cytoskeletal structures and stress granules in CV-1 cells under various experimental conditions Experimental conditions

Experimental results

120 min preincubation with:

Arsenate added and cells fixed after:

Microtubules

Actin filaments

Portion of cells with SGs, %

Number of SGs per one cell exhibiting SGs

Portion of large SGs, %

None None DMSO, 0.2% DMSO, 0.2% Noc, 1.6 ␮M Noc, 6.6 ␮M Noc, 1.6 ␮M Noc, 6.6 ␮M Pacl, 11.7 ␮M Pacl, 11.7 ␮M Vinbl, 11 ␮M LatB, 2 ␮M*

30 min 120 min 30 min 120 min 30 min 30 min 120 min 120 min 30 min 120 min 30 min 30 min

Radial array Radial array Radial array Radial array Partially disrupted Disrupted Disrupted Disrupted Stabilized, disorganized Stabilized, disorganized Partially disrupted Radial array

Stress fibers Stress fibers Stress fibers Stress fibers Stress fibers Stress fibers Stress fibers Stress fibers Stress fibers Stress fibers Stress fibers Disrupted

74 ⫾ 16 76b 58 ⫾ 17 41 ⫾ 1 18 ⫾ 5a 8 ⫾ 2a 3 ⫾ 2a 2 ⫾ 1a 49 ⫾ 19 85 ⫾ 5 24 ⫾ 22b 76 ⫾ 4

119 ⫾ 56 62 ⫾ 16 102 ⫾ 42 86 ⫾ 36 132 ⫾ 52a 118 ⫾ 49a 198 ⫾ 35a 91 ⫾ 30 136 ⫾ 75 70 ⫾ 16 169 ⫾ 62 41 ⫾ 9a

N.D. N.D. 6 ⫾ 1.8 19 ⫾ 6.9 2 ⫾ 1.4a 1 ⫾ 0.3a 19 ⫾ 6.5 5 ⫾ 5.4a 13 ⫾ 5.3a 25 ⫾ 10.2 2 ⫾ 0.9a 27 ⫾ 7.2a

Noc, nocodazole; Pacl, paclitaxel, Vinbl, vinblastine; LatB, latrunculin B. Numbers in the table are mean ⫾ SD. a According to independent t test, means are different from DMSO treatment with P ⬍ 0.05. b Portion of SG-exhibiting cells was analyzed in one or two experiments only, and t test is not applicable. * Preincubation with latrunculin was for 15 min.

and D). PABP-positive granules were nicely expressed in arsenate-treated MT-containing cells which were not treated with nocodazole (Fig. 2A and B). So, SG formation was abolished by nocodazole treatment of cells. In 1.6 ␮M nocodazole some residual MTs remained, and 6.6 ␮M nocodazole disrupted all cellular MTs. The more profoundly MTs were disrupted, the fewer number of cells exhibited SG formation (Table 1). If a nocodazole-treated cell contained any SGs, the total number of SGs was greater than that of control, though the portion of large SGs dramatically decreased (Table 1). However, SGs and in particular large SGs occurred very rarely in nocodazole- and arsenate-treated cells (Fig. 4). When we washed out nocodazole in the presence of arsenate, MT array rapidly restored. In these cells washed after nocodazole treatment SGs developed in the cytoplasm in parallel with the MT growth (Fig. 3E–H). Sometimes SG distribution might correlate with newly formed MTs. To evaluate whether SG formation was inhibited by MT disruption or by nocodazole per se, we treated cells with another MT-disrupting drug vinblastine. This drug caused MT depolymerization and polymerization of tubulin paracrystals (data not shown). In vinblastine-treated cells the formation of SGs was inhibited similarly to nocodazoletreated cells (Table 1, Fig. 4). The tubulin paracrystals potentially could prevent visualization of some SGs; however, the score of SGs per one cell having SGs was close to that of the control (arsenate plus DMSO). The occurrence of large SGs in vinblastine plus arsenate-treated cells was significantly decreased similarly to nocodazole-treated cells (Fig. 2). Thus, we suggested that MT-disruption drugs inhibited SG formation.

Effect of nocodazole and arsenate treatments on eIF-2␣ phosphorylation SG formation depends on eIF-2␣ phosphorylation [4]. We questioned whether eIF-2␣ was phosphorylated in nocodazole plus arsenate-treated cells or nocodazole treatment prevented eIF-2␣ phosphorylation. We have examined the level of eIF-2␣ phosphorylation in treated cells by Western blot with commercially available Ab-specific to phosphoeIF-2␣ (with phosphoylated Ser51). We have found that DMSO-treated CV-1 cells exhibited a weak signal of eIF-2a phosphorylation, though in arsenate (plus DMSO)-treated cells eIF-2␣ was strongly phosphorylated (Fig. 5). The rise of eIF-2␣ phosphorylation was evident also in nocodazoletreated cells (Fig. 5), though it was less expressed as in arsenate-treated cells. Nocodazole did not abolish the elevation of eIF-2a phosphorylation under arsenate action (Fig. 5). We probed immunoblot also with other Abs used in this work and confirmed that neither treatment of cells influenced the amount and electrophoretic mobility of p170, PABP, and ␣-tubulin (Fig. 5). The only exception was the appearance of an anti-PABP Ab-recognized second band in arsenate- and in nocodazole plus arsenate-treated cells. Probably, a fraction of PABP is modified (e.g., phosphorylated) under arsenate treatment, and nocodazole does not influence it. Effect of microtubule stabilization (paclitaxel treatment) on SG formation Anti-MT drugs can induce a signal transduction pathway, which leads to activation of particular kinases and transcription factors [21–23]. So, we had to reveal whether

P.A. Ivanov et al. / Experimental Cell Research 290 (2003) 227–233

231

Fig. 4. The occurrence of large SGs in CV-1 cells under various experimental conditions. The occurrence of large SGs was estimated as the portion of SG-containing cells multiplied to the average number of SGs in these cells. The occurence of SGs in CV-1 cells treated 120 min with DMSO and then 30 min with arsenate plus DMSO was taken as 100%. Column bars are SD. Vb1, vinblastine; Paclitax, paclitaxel; LatB, latrunculin B. Concentrations of drugs and time of treatment as in Table 1.

localization of SGs in the paclitaxel-treated cells was changed: SGs were more randomly distributed in the cytoplasm (Fig. 3J) as in control cells. They were excluded from dense MT clusters though they sometimes followed Mt bundles. The total number of SGs per one paclitaxel-treated SGcontaining cell was similar to that of the control, though the occurrence of large granules was increased (Table 1; Fig. 4). After 2 h of incubation with sodium arsenate in the presence of paclitaxel, the portion of cells containing SGs increased twice that of the control (incubation with arsenate plus Fig. 3. Effect of nocodazole washout and of paclitaxel treatment to stress granules formation. (A, C, E, G, J) Immunostaining with Ab A170-C; (B, D, F, H, K) Immunostaining of the same cells with anti-tubulin Ab DM-1A. Cells were treated with 6.6 ␮M nocodazole for 120 min (A, B); with 1.6 ␮M nocodazole for 120 min and then with arsenate plus nocodazole for 30 min (C, D); (E, F) 15 min of nocodazole removal in the presence of arsenate; (G, H) 30 min of nocodazole removal in the presence of arsenate. (J, K) Cells were treated with paclitaxel for 120 min, and then with arsenate plus nocodazole for 30 min. Bars, 10 ␮m.

nocodazole and vinblastine effects on SGs formation depended on MT disruption or on some signal transduction events. In contrast to nocodazole, paclitaxel stabilizes cellular MTs, though it activates the same signal transduction pathways [23]. We treated cells with paclitaxel, and MTs became chaotically distributed and bundled (Fig. 3K). Arsenate induced SG formation in paclitaxel-treated cells (Fig. 3J), and a subpopulation of cells having SGs was equal to that of control (DMSO plus arsenate treated) (Table 1). The

Fig. 5. Nocodazole does not abolish phosphorylation of eIF-2␣ in CV-1 cells under arsenate action. Immunoblots of CV-1 cell homogenates probed with A170-C antibody (p170), anti-PABP antibody (PABP), anti-␣-tubulin antibody (tubulin), or anti-phospho-eIF-2a antibody (p-eIF2␣). Horseradish peroxidase was developed with 3,3⬘-diaminobenzidine/H2O2. Cells were treated 120 min either with DMSO (DMSO) or nocodazole (Noc). Ars: cells were subsequently treated for 120 min with DMSO, and then for 30 min with arsenate plus DMSO. Noc ⫹ Ars: cells were subsequently treated for 120 min with 6.6. ␮M nocodazole, and then for 30 min with arsenate plus 6.6. ␮M nocodazole.

232

P.A. Ivanov et al. / Experimental Cell Research 290 (2003) 227–233

Fig. 6. Effect of latrunculin B to stress granule formation. (A, C) Immunostaining with Ab A170-C; (B, D) staining of the same cells with rhodamine-phalloidin. (A, B) A 15-min incubation with latrunculin B; (C, D) as in A and B, and then arsenate was added for 30 min (together with latrunculin B). Bars, 10 ␮m. Arrows point to SGs.

studied yet, although the existence of such an interaction could be hypothesized. We speculate that interaction with MTs could be essential not only for mRNP and other translational machinery component transport, but for facilitation of translation initiation as well. MTs can help translation initiation components to meet each other in the cytoplasm. Likely, mRNP and initiation factors move along MTs driven by MT-dependent motors. After initiation occurs, polysomes loose the association with MTs and tether in the cytoplasm, probably, on actin filaments as supposed previously [25]. Our results suggest that SGs are gatherings of preinitiation complexes that cannot finish initiation because of cell stress and therefore accumulate at MTs. An additional study of SG formation mechanism is required to define the molecular components involved in the association of translation initiation components with MTs and especially in their transport along MTs.

Acknowledgments DMSO), and the occurrence of large SGs also increased (Table 1, Fig. 4). Thus, paclitaxel enhanced SGs formation rather than inhibited it, and the effect of anti-MT drugs on SGs most probably is a result of their action on MTs but not a result of signal transduction pathway activation.

This work was financially supported by the Russian Foundation for Basic Research, Grant 02-04-48784. We are very grateful to Dr. Ivan N. Shatsky for the kind gift of antibody, Dr. Vladimir I. Rodionov for fruitful discussions, and Dr. Yulia A. Komarova for critical reading of the manuscript.

Effect of latrunculin B treatment on SG formation Arsenate treatment did not influence the actin stress-fiber system visualized with rhodamine-phalloidin (Fig. 1F and G). In cells treated with latrunculin B (2 ␮g/ml, 30 min), actin fibers were disrupted and cell bodies were contracted (Fig. 6A and B). In cells treated with arsenate in additionaly to latrunculin B, distinct SGs were formed in the cytoplasm (Fig. 6C and D). We analyzed a 30-min arsenate treatment only because prolongation of both latrunculin B and arsenate treatment caused cell detachment from coverslips. The portion of latrunculin-treated cells that exhibited SGs was equal to the control condition (Table 1). Interestingly, the portion of large SGs in each cell and their occurrence was even increased (Table 1; Fig. 4). This observation supports the suggestion that SG formation strictly depends on MTs. Interestingly, it has been shown that binding to actin can delay MT-based motility [24]. Probably, actin disruption facilitates the motility of SG components in the cytoplasm; therefore, the occurrence of large SGs increases. All our results indicate that SG formation depends on the MT system. MT could serve either as a scaffold or as railway bringing SG components together. It is well known that cellular mRNPs can be transported along MTs by motor proteins [11,12,25–27]. This transport was studied for oocytes and neurons, and poorly studied in fibroblasts. Some translation factors can bind MT also [14]. The interaction of these factors with MT-depending motors proteins was not

References [1] L. Nover, K.D. Scharf, D. Neumann, Formation of cytoplasmic heat shock granules in tomato cell cultures and leaves, Mol. Cell. Biol. 3 (1983) 1648 –1655. [2] L. Nover, K.D. Scharf, D. Neumann, Cytoplasmic heat shock granules are formed from precursor particles and are associated with a specific set of mRNAs, Mol. Cell. Biol. 9 (1989) 1298 –1308. [3] K.D. Scharf, H. Heider, I. Hohfeld, R. Lyck, E. Schmidt, L. Nover, The tomato Hsf system: HsfA2 needs interaction with HsfA1 for efficient nuclear import and may be localized in cytoplasmic heat stress granules, Mol. Cell. Biol. 18 (1998) 2240 –2251. [4] N.L. Kedersha, M. Gupta, W. Li, I. Miller, P. Anderson, RNAbinding proteins TIA-1 and TIAR link the phosphorylation of eIF-2 alpha to the assembly of mammalian stress granules, J. Cell Biol. 147 (1999) 1431–1442. [5] N. Kedersha, M.R. Cho, W. Li, P.W. Yacono, S. Chen, N. Gilks, D.E. Golan, P. Anderson, Dynamic shuttling of TIA-1 accompanies the recruitment of mRNA to mammalian stress granules, J. Cell Biol. 151 (2000) 1257–1268. [6] N. Kedersha, S. Chen, N. Gilks, W. Li, I.J. Miller, J. Stahl, P. Anderson, Evidence that ternary complex (eIF2-GTP-tRNA(i)(Met))deficient preinitiation complexes are core constituents of mammalian stress granules, Mol. Biol. Cell 13 (2002) 195–210. [7] K. Luby-Phelps, Cytoarchitecture and physical properties of cytoplasm: volume, viscosity, diffusion, intracellular surface area, Int. Rev. Cytol. 192 (2000) 189 –221. [8] M. Bomsel, R. Parton, S.A. Kuznetsov, T.A. Schroer, J. Gruenberg, Microtubule- and motor-dependent fusion in vitro between apical and basolateral endocytic vesicles from MDCK cells, Cell 62 (1990) 719 –731.

P.A. Ivanov et al. / Experimental Cell Research 290 (2003) 227–233 [9] G.J. Bassell, R.H. Singer, K.S. Kosik, Association of poly(A) mRNA with microtubules in cultured neurons, Neuron 12 (1994) 571–582. [10] N.J. Pokrywka, E.C. Stephenson, Microtubules are a general component of mRNA localization systems in Drosophila oocytes, Dev. Biol. 167 (1995) 363–370. [11] E.A. Arn, P.M. MacDonald, Motors driving mRNA localization: new insights from in vivo imaging, Cell 95 (1998) 151–154. [12] J.H. Carson, K. Worboys, K. Ainger, E. Barbarese, Translocation of myelin basic protein mRNA in oligodendrocytes requires microtubules and kinesin, Cell Motil. Cytol. 38 (1997) 318 –328. [13] C.L. Sundell, R.H. Singer, Requirement of microfilaments in sorting of actin messenger RNA, Science 253 (1991) 1275–1277. [14] N.A. Shanina, P.A. Ivanov, E.M. Chudinova, F.F. Severin, E.S. Nadezhdina, Translation initiation factor eIF3 is able to bind with microtubules in mammalian cells, Mol. Biol. (Mosk.) 35 (2001) 638 – 646. [15] R.C. Moore, R.J. Cyr, Association between elongation factor-1 alpha and microtubules in vivo is domain dependent and conditional, Cell Motil. Cytoskel. 45 (2000) 279 –292. [16] G. Liu, W.M. Grant, D. Persky, V.M. Latham Jr., R.H. Singer, J. Condeelis, Interactions of elongation factor 1 alpha with F-actin and beta-actin mRNA: implications for anchoring mRNA in cell protrusions, Mol. Biol. Cell 13 (2002) 579 –592. [17] G. Liu, J. Tang, B.T. Edmonds, J. Murray, S. Levin, J. Condeelis, F-actin sequesters elongation factor 1 alpha from interaction with aminoacyl-tRNA in a pH-dependent reaction, J. Cell Biol. 135 (1996) 953–963. [18] P.V. Ruzanov, V.M. Evdokimova, N.L. Korneeva, J.W. Hershey, L.P. Ovchinnikov, Interaction of the universal mRNA-binding protein, p50, with actin: a possible link between mRNA and microfilaments, J. Cell Sci. 112 (1999) 3487–3496.

233

[19] M. Bektas, R. Nurten, Z. Sayers, E. Bermek, Interactions of elongation factor 2 with the cytoskeleton and interference with DNase I binding to actin, Eur. J. Biochem. 256 (1998) 142–147. [20] E.A. Shestakova, L.P. Motuz, A.A. Minin, V.I. Gelfand, L.P. Gavrilova, Some of eukaryotic elongation factor 2 is colocalized with actin microfilament bundles in mouse embryo fibroblasts, Cell Biol. Int. Rep. 15 (1991) 75– 84. [21] C. Rosette, M. Karin, Cytoskeletal control of gene expression: depolymerization of microtubules activates NF-kappa B., J. Cell Biol. 128 (1995) 1111–1119. [22] T.V. Ivanova, V.N. Ivanov, E.S. Nadezhdina, Transcription factors NF-kappaB and AP-1/c-fos in cell response to nocodazole, Membr. Cell Biol. 14 (2001) 727–741. [23] M.V. Blagosklonny, T. Fojo, Molecular effects of paclitaxel: myths and reality (a critical review), Int. J. Cancer 83 (1999) 151–156. [24] A. Al-Haddad, M.A. Shonn, B. Redlich, A. Blocker, J.K. Burkhardt, H. Yu, J.A. Hammer 3rd, D.G. Weiss, W. Steffen, G. Griffiths, S.A. Kuznetsov, Myosin Va bound to phagosomes binds to F-actin and delays microtubule-dependent motility, Mol. Biol. Cell 12 (2001) 2742–2755. [25] J.K. Yisraeli, S. Sokol, D.A. Melton, A two-step model for the localization of maternal mRNA in Xenopus oocytes: involvement of microtubules and microfilaments in the translocation and anchoring of Vg1 mRNA, Development 108 (1990). [26] R.P. Brendza, L.R. Serbus, J.B. Duffy, W.M. Saxton, A function for kinesin I in the posterior transport of oskar mRNA and Staufen protein, Science 289 (2000) 2120 –2122. [27] R.B. Knowles, J.H. Sabry, M.E. Martone, T.J. Deerinck, M.H. Ellisman, G.J. Bassell, K.S. Kosik, Translocation of RNA granules in living neurons, J. Neurosci. 16 (1996) 7812–7820.