Concomitant alterations in distribution of 70 kDa heat shock proteins, cytoskeleton and organelles in heat shocked 9L cells

The International Journal of

PERGAMON

The International Journal of Biochemistry & Cell Biology 30 (1998) 745±759

Biochemistry & Cell Biology

Concomitant alterations in distribution of 70 kDa heat shock proteins, cytoskeleton and organelles in heat shocked 9L cells Ting-Ting Wang, Ann-Shyn Chiang *, Jao-Jia Chu, Ting-Jen Cheng, TzuMei Chen, Yiu-Kay Lai* Department of Life Science, National Tsing Hua University, Hsinchu 30043, Taiwan, Republic of China Received 22 May 1997; accepted 25 September 1997

Abstract Maintenance of cell architecture and positioning of organelles are major functions of the cytoskeleton. On the other hand, induction of heat shock proteins (HSPs) and reorganization of the cytoskeleton are the most signi®cant changes in heat-shocked mammalian cells. We examine the alterations in HSP70 and its constitutively expressed cognate, HSC70, as well as the cytoskeleton and organelles in 9L rat brain tumor cells upon heat shock. We employed ¯uorescence microscopy and scanning electron microscopy to follow these changes. Levels of HSP70s were quanti®ed by Western blotting. Accumulation of HSC70 was more transient and the protein translocated to and subsequently exited from the nucleus more rapidly than HSP70. Changes in actin micro®laments include the nuclear localization of actin fraction and disappearance of cytoplasmic micro®lament bundles, while the cortical actin micro®laments were almost una€ected. Furthermore, microtubules retracted slightly from the cell periphery but remained largely unchanged. In contrast, the intermediate ®laments collapsed into the perinuclear region. The mitochondria converted from ®lamentous into granular forms and clustered in a region overlapping with the collapsed intermediate ®laments. All of the above alterations are reversible and largely reverted after 8 h of recovery. The e€ect on Golgi organization was very transient and the apparatus assumed a normal appearance within 4 h after the heat treatment. The ER, on the other hand, was totally una€ected by the heat treatment. These observations help correlate the sequential events following a stress like heat shock and suggest possible physiological functions of these essential constituents of a cell under stress. # 1998 Published by Elsevier Science Ltd. All rights reserved. Keywords: Heat shock; Heat shock proteins; Cytoskeleton; Organelles; Cell architecture

1. Introduction

* Corresponding authors. Fax: +886-3-5715934; E-mail: [email protected] or [email protected].

The hallmark of cellular response to heat shock is the induction of heat shock proteins (HSPs) [14, 28, 40]. Amongst all HSPs, members of the 70 kDa family (referred to as the HSP70s)

1357-2725/98/$19.00 # 1998 Published by Elsevier Science Ltd. All rights reserved. PII: S 1 3 5 7 - 2 7 2 5 ( 9 7 ) 0 0 1 3 3 - 7

746

T.-T. Wang et al. / The International Journal of Biochemistry & Cell Biology 30 (1998) 745±759

are often the most prominent products of protein synthesis in heat-shocked cells and they are the most extensively studied [37, 40]. There are at least two cytoplasm/nucleus residing HSP70s in mammalian cells, the HSC70 and the HSP70 [16, 23]. HSC70 is constitutively expressed but only slightly heat-inducible, whereas HSP70 is rarely expressed under normal conditions and highly heatinducible [36]. These two proteins possess weak ATPase activity that is stimulated upon the binding of peptides [10] and can dissociate some protein aggregates coupled with ATP lysis [37]. Thus, the HSP70s may be responsible for the refolding and renaturation of other cellular proteins damaged by heat treatment. More recently, HSC70 and HSP70 (as well as other major HSPs), are demonstrated to function as molecular chaperone both in normal cells and in cells under a variety of stresses including heat shock [1, 39, 40]. Due to a high sequence homology between these two proteins [28], their cellular functions are often assumed to be similar if not identical [4, 40]. However, we have previously shown that, with respect to cytoskeleton association and by biochemical analysis, HSC70 is associated with intermediate ®laments (IFs) while HSP70 is associated with the tubulins, the building blocks of microtubules (MTs) in heat-shocked cells [7, 23]. These observations suggest that HSC70 and HSP70 may function di€erently, probably with distinct but overlapping substrates. However, the total protein levels of these two HSP70s and their kinetics of subcellular localizations in heat-shocked cells are yet to be clearly de®ned. Along with the induction of HSPs, alteration of cell morphology resulting from a reorganization of cytoskeletal systems are the most drastic changes in heat-shocked mammalian cells in culture [7, 46, 53, 54]. Most mammalian cells in culture round up when subjected to heat-shock treatment, however, several cell types assumed a ¯attened and wider shape upon heat shock [46, 47]. The cytoskeleton is a dynamic intracellular matrix composed of three major, independent but interconnected, ®laments [3]. These ®laments are assembled from di€erent protein subunits: micro®laments (MFs) from actin [38], MTs from a,

b, g-tubulins [18, 32] and IFs from a large family of related protein subunits, such as desmin, keratins and vimentin [12, 19, 44]. In conjunction with respective arrays of associated proteins, these ®laments constitute the cellular organization and participate in various cellular activities. For instance, it has been reported that the association of intracellular organelles with the cytoskeletal systems is essential for intracellular organelle transportation and may result in the modi®cations of organelle activities [8, 18, 20, 31]. The cytoskeletal architecture and the distribution of a number of intracellular membraneous organelles are profoundly a€ected by heat stress. It has been shown that, in heat-shocked mammalian cells, a fraction of actins appears in the nucleus [35, 53], the IFs collapse into a perinuclear ring [7, 53, 55]; whereas the MTs are only slightly perturbated [53]. Concomitantly, mitochondria redistribute along with the collapsed IFs and the Golgi complex appears fragmented [53]. However, the above observations are made by using di€erent cell lines and under di€erent treatment protocols. It is expected that the kinetic changes in cytoskeletal architecture and positioning of intracellular organelles in heatshocked cells may vary depending on the cell types as well as the severity of the heat stress [9]. Herein we studied the e€ect of heat shock on 9L rat brain tumor cells with regards to the following: (1) accumulation and subcellular distribution of HSC70 and HSP70; (2) cell morphology as well as organization of micro®laments, microtubules and intermediate ®laments and (3) morphology and distribution of membraneous organelles including mitochondria, Golgi apparatus and endoplasmic reticulum. The results allow a correlation of the sequential events that are essential constituents of a `heat-shock response' in mammalian cells. 2. Materials and methods 2.1. Immunochemical reagents and ¯uorescence dyes HSC70, HSP70, tubulin and vimentin were, respectively, immunolabeled with mouse or rat

T.-T. Wang et al. / The International Journal of Biochemistry & Cell Biology 30 (1998) 745±759

monoclonal antibodies. Antibodies against HSC70 and HSP70 were purchased from StressGen (Victoria, BC, Canada), those against b-tubulin from Boehringer Mannheim (Postfach, Germany) and those against vimentin from Amersham (Little Chalfont, UK). Rhodamineor ¯uorescein isothiocyanate (FITC)-conjugated goat anti-mouse/rat IgG were both from Boehringer Mannheim. Actins were labeled with rhodamine-conjugated phalloidin purchased from Molecular Probes (Eugene, OR). The nuclei were counterstained with Hoechst 33342 (Molecular Probes). Vital-dye staining for mitochondria and endoplasmic reticulum was done by using di€erent concentrations of DiOC6 (Molecular Probes). The Golgi complex was stained with BODIPYceramide (Molecular Probes).

2.2. Cell culture and heat-shock treatment The 9L rat brain tumor cells, originated from rat gliosarcoma, were a gift of Dr. M.L. Rosenblum (University of California at San Francisco, CA) [49]. The cells were maintained in Eagle's minimum essential medium (MEM) containing 10% fetal calf serum, 100 units/ml penicillin G and 100 mg/ml streptomycin. Stock cells were cultured in 25 cm2 culture ¯asks in a humidi®ed 378C incubator set at 5% CO2 and 95% air. Prior to each experiment, stocked cells were seeded at a density of 4 to 6  104 cells per cm2 on sterilized glass coverslips if the cells were to be used for microscopic studies. All experiments were performed using exponentially growing cells at 80 to 90% con¯uency. For heat-shock treatment, the dishes were sealed with Para®lm and submerged in a thermostat waterbath at 45 2 0.18C for 15 min. The designated temperature of the medium in the heating protocol was reached within 3 min and the time for equilibrium was included in the treatment duration. After treatment, the cells were returned to the incubator for various durations as indicated.

747

2.3. SDS-PAGE and immunoblotting analysis of the 70 kDa heat-shock proteins After treatment, the cells were washed with PBS and lysed in sample bu€er and the cell lysates were applied to 10% SDS-polyacrylamide gels with 4.75% stacking gels [21]. After electrophoresis, the gels were stained for 1 h (in 0.1% Coomassie brilliant blue R250 in 10% acetic acid and 50% methanol) and then destained. Alternatively, the gels were soaked in transfer bu€er (50 mM Tris±borate, pH 8.3, 1 mM EDTA) for 10 min and the proteins were electrotransferred onto nitrocellulose membranes (Hybond-C Extra, Amersham) using the semi-dry method (OWL Scienti®c Plastic, Cambridge, MA). The membranes were incubated for 1 h with Tween 20 containing Tris±bu€er saline (TTBS: 20 mM Tris±HCl, pH 7.4, 500 mM NaCl, 0.05% Tween 20) supplemented with 5% nonfat milk powder and then rinsed with TTBS brie¯y. Subsequently, the membranes were incubated with monoclonal antibodies to HSC70 or HSP70 (both diluted 1:200 in TTBS containing 1% milk powder) at room temperature for 2 h. After three washes with TTBS, immunocomplexes on the membranes were reacted with goat anti-mouse antibody conjugated with alkaline phosphatase (diluted 1:2,000 in TTBS containing 5% milk powder) at room temperature for 30 min. The membrane was then rinsed three times with TTBS, dried and subjected to detection at room temperature in developing bu€er (15 mg of nitroblue tetrazolium, 0.7% N,Ndimethylformamide, 30 mg of 5-bromo-4-chloro3-indolyl phosphate per 10 ml, 1 mM MgCl2 and 100 mM NaHCO3, pH 9.8). 2.4. Scanning electron microscopy For scanning electron microscopy, cells on coverslips were ®xed directly in 2.5% glutaraldehyde solution containing 4.5% glucose and buffered with 75 mM cacodylate (pH 7.2) at 48C for 40 min. After three washes with 100 mM cacodylate bu€er, cells were ®xed in 1% osmium tetroxide bu€ered with 50 mM cacodylate (pH 7.2) at 48C for 40 min. Subsequently, cells were washed

748

T.-T. Wang et al. / The International Journal of Biochemistry & Cell Biology 30 (1998) 745±759

with milli-Q water, dehydrated in an ethanol series, critical-point dried, mounted on a stub, coated with gold±palladium and examined by a Hitachi S-2300 SEM at 15 kV. 2.5. Indirect immuno¯uorescence microscopy for the HSP70s To visualize the distribution of HSC70 and HSP70, cells were washed in phosphate-bu€ered saline (PBS) at room temperature and ®xed in 4% paraformaldehyde for 20 min at 48C. Subsequent steps were carried out at room temperature. Blocking of non-speci®c sites and dilution of antibodies were performed by in 1% bovine serum albumin (BSA) in PBS. After ®xation, cells were rinsed in PBS and permeabilized with 0.1% Triton X-100 in PBS for 10 min, rinsed again in PBS, blocked for 1 h and incubated with mouse monoclonal antibody against HSC70 (diluted 1:40) or HSP70 (diluted 1:50). The cells were washed three times with PBS for 15 min, blocked again for 30 min, allowed to react with rhodamine-conjugated goat anti-mouse IgG (diluted 1:50) for 1 h and counterstained with 10 mM Hoechst 33342 for 30 s. All ¯uorescence staining procedures (same as in Section 2.6) were performed in the dark. The coverslips were washed extensively with PBS for 15 min and then mounted in an anti-bleaching reagent (5% propylgallate dissolved in 15% PBS/85% glycerol) and stored in the dark at 48C until viewed. The samples were examined under a Zeiss Axiophot microscope equipped with epi¯uorescence optics and all images were recorded under a Plan-neo¯uar 100X (N.A. 1.3) objective. 2.6. Fluorescence microscopy for cytoskeletal components Actin ®laments were ¯uorescent-labeled by rhodamine-conjugated phalloidin as follows. After rinsing in PBS, the cells were ®xed in 4% paraformaldehyde at 48C for 20 min, permeabilized with 0.05% Triton X-100 at room temperature for 10 min and then labeled with rhodamineconjugated phalloidin (diluted 1:20) for 20 min. For tubulin labeling, cells on coverslips were

washed in PBS and then ®xed and permeabilized with absolute methanol at ÿ208C for 5 to 10 min. Subsequent steps were performed as described for the HSP70s except that mouse monoclonal antibodies against b-tubulin (diluted 1:50) were used. Vimentin stains were presented as double ¯uorescence together with rhodamine±phalloidin stain to actins. The cells were rinsed in PBS, ®xed in paraformaldehyde and permeabilized with Triton X-100 as described for actin staining. Subsequently, vimentin IFs were ¯uorescent labeled by mouse monoclonal antibodies against vimentin (diluted 1:50) and FITC-conjugated goat anti-mouse IgG (diluted 1:50) as primary and secondary antibodies, respectively. After further washing in PBS, the samples were labeled with rhodamine-conjugated phalloidin as described. 2.7. Vital-dye staining of mitochondria, ER and Golgi apparatus For staining of mitochondria and ER, cells were washed with Hanks' balanced salt solution (HBSS) to remove the culture medium. The living cells were incubated with 0.1 mg/ml DiOC6 in HBSS at 378C for 1 min for staining mitochondria [17]. Cells were rinsed in HBSS and mounted in a wet-chamber with dye-free HBSS for examination. Vital-dye staining of the ER was performed similarly except that the concentration of DiOC6 was increased to 2.5 mg/ ml [45]. For staining of the Golgi complex, cells were washed several times with 10 mM Hepesbu€ered MEM (HMEM) to remove the culture medium. The living cells were incubated with 5 mM BODIPY-ceramide at 378C for 5 to 10 min [29], washed and mounted in dye-free HMEM for examination. 3. Results 3.1. Accumulation of cellular HSP70s in heatshocked 9L cells Accumulation of the HSC70 and HSP70 in the heat-shocked cells was studied by immunoblot-

T.-T. Wang et al. / The International Journal of Biochemistry & Cell Biology 30 (1998) 745±759

749

Fig. 1A and B demonstrate that the antibodies against HSC70 and HSP70 were not cross-reactive. 3.2. Nuclear localization and distribution of HSC70 and HSP70 in heat-shocked 9L cells

Fig. 1. Immunoblotting analysis of HSC70 and HSP70 in normal, heat-shocked and recovering 9L RBT cells. Cells were heat-shocked at 458C for 15 min and allowed to recover under normal growing condition for up to 16 h. At di€erent time intervals, the cells were harvested and cellular proteins were resolved by SDS-PAGE. The resolved protein were electrotransferred onto membranes and respectively probed by antibodies to HSC70 (A) and HSP70 (B). Subsequently, immunoblots in (A) and (B) were quanti®ed by a computer densitometer (C). Channels 1: untreated control cells; 2±8: immediate, 1, 2, 4, 8, 12 and 16 h after the heat-shock treatments, respectively. The data are means2 S.D. from three independent experiments.

ting. As shown in Fig. 1, there was a basal level expression of HSC70 and this protein was heatinduced as the protein content rose signi®cantly after heat-shock. However, the accumulation of HSC70 continued for only 4 h after heat-shock and the protein level subsided thereafter (Fig. 1A and C). The immunoblots also indicate the absence of HSP70 in control cells and a vigorous induction as well as a dramatic accumulation of this protein in cells recovering from heat-shock (Fig. 1B and C). Single immunoreactive bands in

Immunolabeling of HSC70 and HSP70 was performed by monoclonal antibodies that are not cross-reactive, therefore, it was possible to separately investigate the distribution of these two closely related proteins by the immuno-¯uorescence microscopic technique. The experimental procedures for immunostaining of HSC70 and HSP70 were carried out simultaneously and the ®nal ¯uorescence images were recorded under identical conditions. In control cells, a de®nite staining pattern was observed with anti-HSC70 indicative of a basal expression level of the protein. Moreover, the protein was found to be concentrated in the nucleus, particularly in the nucleolar regions (Fig. 2A). In contrast, expression of the HSP70 was hardly detected (Fig. 3A). The immuno¯uorescent pattern for HSC70 gradually increases, demonstrating the induction of this protein upon heat-shock (Fig. 2B± D). Although a portion of HSC70 already reside in the nucleus/nucleoli, further concentration of HSC70 into the organelle is evident after heat treatment. Upon heat shock, the protein redistributed from the cell periphery to the center and became more concentrated in the nucleus as well as at the perinuclear region (Fig. 2B±D). After 4 h of recovery, HSC70 was found to be highly concentrated in the nucleoli as implicated by bright ¯uorescence (Fig. 2E). Thereafter, the distribution of this protein was restored to normal and the ¯uorescence level returned to basal after 8 h of cell recovery (Fig. 2F±H). As mentioned earlier, HSP70 was not detected in the untreated cells and would not become detectable until 1 h past treatment (Fig. 3A±C). The majority of the protein was localized to the nucleus soon after its synthesis (Fig. 3C and D) followed by a more wide-spread cytoplasmic localization of this protein (Fig. 3E and F). Moreover, the immuno¯uorescence of HSP70 continued to increase during the ®rst 8 h of recovery, showing the vig-

750

T.-T. Wang et al. / The International Journal of Biochemistry & Cell Biology 30 (1998) 745±759

Fig. 2. Distribution of HSC70 in normal, heat-shocked, and recovering 9L RBT cells. Cells were heat-shocked at 458C for 15 min and allowed to recover under normal growing condition for up to 16 h. At di€erent time intervals, the distribution of HSC70 was visualized by indirect immuno¯uorescent staining (upper panels). The cells were also counterstained with Hoechst (lower panels). (A) Untreated control cells; (B) immediately following the heat-shock treatment; (C) 1 h, (D) 2 h, (E) 4 h, (F) 8 h, (G) 12 h, (H) 16 h after the heat-shock treatment. Shown are typical photographs from three independent experiments. Bar = 20 mm.

orous induction and accumulation of this protein under the present experimental condition (Fig. 3A±E). Unlike HSC70, where the accumulation subsided quickly after 4 h of cell recovery, the cellular amount of HSP70 remained at a relative high level for at least 24 h (Fig. 3H). 3.3. Changes in cell shape and distribution of cytoskeletal components in heat-shocked 9L cells Fig. 4 shows the morphological changes of the 9L cells after the heat-shock treatment (458C, 15 min). Before treatment, the cells appear well

spread on the substratum (Fig. 4A). Upon heatshock, while most cytosol contracted to the cell center, a thin layer remained attached to the substratum. These cells assumed their maximum contracted morphology at 2 h after the treatment (Fig. 4B±D). Later, the cells gradually spread outwards and assumed their normal appearance (Fig. 4E and F). The process of cell rounding up and respreading was assumed to result from the reorganization of the cytoskeleton which was subsequently con®rmed by ¯uorescence microscopy. In untreated cells, actin MFs were found to be dis-

T.-T. Wang et al. / The International Journal of Biochemistry & Cell Biology 30 (1998) 745±759

751

Fig. 3. Distributions of HSP70 in normal, heat-shocked and recovering 9L RBT cells. Cells were heat-shocked at 458C for 15 min and allowed to recover under normal growing condition for up to 16 h. At di€erent time intervals, the distribution of HSP70 was visualized by indirect immuno¯uorescent staining (upper panels). The cells were also counterstained with Hoechst (lower panels). (A) Untreated control cells; (B) immediately following the heat-shock treatment; (C) 1 h, (D) 2 h, (E) 4 h, (F) 8 h, (G) 12 h, (H) 16 h after the heat-shock treatment. Shown are typical photographs from three independent experiments. Bar = 20 mm.

tributed mainly in two major areas, suggested by rhodamine-conjugated phalloidin staining. Concentrated groups of MFs were associated with the cell cortex, especially at several leading edges. In addition, cytoplasmic localized parallel bundles of MFs were also evident (Fig. 5A). After heat-shock, the peripheral MFs remained largely unchanged but the cytoplasmic MF bundles swiftly disassembled and almost completely disappeared at 1 hour post-treatment (Fig. 5B and C). Simultaneously, a di€used pattern of actin staining was found inside the nucleus (Fig. 5C and D). Subsequently, the MF bundles began reassembling and normal distribution of

MFs was restored after 8 h of recovery (Fig. 5D± F). In control cells, the cytoplasm was ®lled with well-spread and evenly distributed MTs and vimentin IFs that extended from the perinuclear region to the cell peripherals (Fig. 5G and M). After heat-shock, the MT network retracted slightly from the cell periphery, no other major change was observed (Fig. 5G±L). In contrast, the vimentin IFs collapsed and aggregated in the perinuclear region (Fig. 5P). Coincident with the changes in cell shape, perinuclear aggregation of vimentin IFs was most evident 2 h after the heatshock treatment and the network gradually

752

T.-T. Wang et al. / The International Journal of Biochemistry & Cell Biology 30 (1998) 745±759

Fig. 4. Morphological changes of heat-shocked 9L RBT cells. Cells were heat-shocked at 458C for 15 min and allowed to recover under normal growing condition for up to 8 h. At di€erent time intervals, the cells were ®xed and processed for scanning electron microscopy and viewed by a Hitachi S-2300 SEM at 15 kV. (A) Untreated control cells; (B) immediately following the heat-shock treatment; (C) 1 h, (D) 2 h, (E) 4 h, (F) 8 h after the heat-shock treatment. Shown are typical photographs from three independent experiments. Bar = 20 mm.

T.-T. Wang et al. / The International Journal of Biochemistry & Cell Biology 30 (1998) 745±759

Fig. 5. (caption overleaf ).

753

754

T.-T. Wang et al. / The International Journal of Biochemistry & Cell Biology 30 (1998) 745±759

assumed their normal appearance after 8 h of recovery (Fig. 5P±R). 3.4. Morphological changes and redistributions of mitochondria, Golgi apparatus and endoplasmic reticulum in heat-shocked 9L cells The membraneous organelles including mitochondria, Golgi apparatus and ER were visualized by vital-dye staining. In normal cells, mitochondria are mostly long and straight and evenly distributed in the cytoplasm (Fig. 6A). Upon heat-shock, these ®lamentous mitochondria changed their morphology rapidly and became shortened and swollen within 15 min (Fig. 6B). Subsequently, they were found to migrate toward the perinuclear region where they became mostly granular (Fig. 6C). The increase in the number of mitochondria around the nucleus was maximum at 1 to 2 h post-treatment, similar to vimentin distribution under the same treatment condition (Fig. 6C and D). After 2 h of recovery, mitochondria started to move away from the cell center and the ®lamentous form reappeared (Fig. 6D±F). Normal appearance and distribution of mitochondria were restored after 8 h of recovery (Fig. 6F). The Golgi apparatus stained by BODIPY-ceramide appeared as a well-packed and organized collection located near the nucleus of normal 9L cells (Fig. 6G). After heat-shock, the Golgi apparatus dissociated into numerous small vesicles while remaining at the same juxtanuclear position (Fig. 6H and I). Subsequently, these membraneous vesicles gradually reorganized and reverted to well-packed Golgi apparatus after 4 h of cell recovery (Fig. 6J±L). The ER, appeared to be una€ected by heat-shock and maintained its reticulated pattern throughout the treatment and recovery stages (Fig. 6M±R).

Although the ¯uorescent dye DiOC6 at high concentration stained all intracellular membrane, ER could be easily recognized by its lace-like reticulum appearance (Fig. 6M). This organelle distributed parallel to the substratum in normal as well as heat-shocked cells. Thus, its distribution was not a€ected by the rounding up response of cells to heat shock. 4. Discussion Induction of HSP70s in heat-shocked cells is almost a universal phenomenon and can be detected in all mammalian cells studied [50]. However, most of the studies have been concentrated in the induction process and are carried out by labeling of the treated cells with radioactive amino acids. It is worthy to note that the technique resolves the synthesis rate but not the accumulation rate of the HSPs. By using Western blotting, together with the immuno¯uorescent microscopic studies, our data demonstrate that the accumulation of HSC70 and HSP70 is completely di€erent in the 9L cells despite the fact that the pro®les of their synthesis rate are similar (see Chen et al. [6] for the [35 S]methionine labeling data). Accumulation of HSC70 in the heatshocked cells seems to be transient but that of HSP70 is more persistent. Such a di€erence may be due to di€erent degradation rates of these two proteins. Moreover, the distribution of these proteins is also di€erent as revealed by the time course studies. Although both proteins are localized to the nucleus after heat shock, the concentration of HSC70 into the nucleolar area is much more abundant compared to that of HSP70. After 8 h of cell recovery, the amount and distribution of HSC70 are similar to that of the

(Fig. 5. on p. 753) Fig. 5. Distribution of actin, tubulin and vimentin in normal, heat-shocked and recovering 9L RBT cells. Cells were heat-shocked at 458C for 15 min and allowed to recover under normal growing condition for up to 8 h. At di€erent time intervals, actins were stained by rhodamine-conjugated phalloidin (left column), tubulins were visualized indirectly by staining with primary antibodies to b-tubulin and rhodamine-conjugated secondary antibodies (center column), vimentins were double stained with FITC-conjugated secondary antibody to the immunolabeled vimentin and rhodamine-conjugated phalloidin to actins (to outline the cell boundary, right column). Nuclei were counterstained with Hoechst. C, Untreated control cells; R0 to R8, 0 to 8 h of recovery after the heatshock treatment. Shown are typical photographs from three independent experiments. Bar = 20 mm.

T.-T. Wang et al. / The International Journal of Biochemistry & Cell Biology 30 (1998) 745±759

Fig. 6. (caption overleaf ).

755

756

T.-T. Wang et al. / The International Journal of Biochemistry & Cell Biology 30 (1998) 745±759

untreated cells whereas there is a further concentration of HSP70 in the nucleus as well as an increase in cytoplasmic distribution. On the other hand, the amount and/or the synthesis rate of HSP70s has been suggested to be responsible for the development of thermotolerance in cells primed with a sublethal dose of heat shock [22, 26, 30]. By comparing the kinetics of cellular accumulation of the HSP70s with that of the development of thermotolerance after the same heat-shock treatment in 9L cells [25], it is obvious that the accumulation of HSP70 but not HSC70 is closely related to the development of thermotolerance. Taken together, these observations lead us to conclude that the substrates/ targets for these two chaperones are not identical, consistent with our previous hypothesis [23, 25]. The reorganization of the cytoskeleton has also been suggested to be responsible for the development of thermotolerance in heat-shocked cells [52]. We have previously shown that the integrity of IFs is associated with the development of acquired thermotolerance in heatshocked 9L cells [25]. Amongst the three major cytoskeletal systems, the MTs and MFs are almost una€ected but for slight retraction of MTs from the cell periphery and the concomitant depolymerization of cytoplasmic MF bundles with the nuclear localization of a small fraction of actins. In contrast, the IFs are concentrated and form a tight ring around the nucleus and this arrangement may protect the nuclear structure. Recently, several lines of evidence indicate that many HSPs are closely associated with, if not the legitimate components of, the cytoskeletal structures [7, 11, 15, 27]. Taken together, it is conceivable that both HSPs and the cytoskeleton are required for the maximum protection of cells under stress.

The components of the cell cytoskeleton have been suggested to involve in gene regulation. For instance, we have hypothesized that the disrupted vimentin IFs may represent a form of denatured proteins in stressed cells thus would attract the binding of HSC70/HSP90 by which the heatshock transcription factor (HSF) is activated [7, 24]. On the other hand, nuclear localization of actins in heat-shocked 9L cells deserves attention. In addition to heat shock, translocation of actin into the nucleus has been detected in numerous cells under a variety of stress [13, 34, 42, 53] and it has been shown that nuclear actins are involved in the transcriptional process [43]. More recently, it has been reported that a variety of agents and conditions that depolymerized microtubules activate the sequencespeci®c transcription factor NF-kB and induce NF-kB-dependent gene expression [41]. By computer-aided sequence analysis, we have located NF-kB binding sites in promoter regions of HSP70 in rat and human (Cheng and Lai, unpublished observation) but the functionality of these regulatory elements in the hsp70 genes has not been established. Therefore, whether the retraction of the MT system detected in the heatshocked cells represents a partial depolymerization of the MTs as well as whether the transcription factor NF-kB is actually involved in the transactivation of hsp70s remain to be investigated. Nevertheless, changes in cell shape or architecture resulting from the reorganization of the cytoskeletal systems have long been shown to exert speci®c e€ects on gene expression [2, 51]. The cytoskeleton, especially the MT system, is well-known to be responsible for the intracellular translocation and positioning of organelles [3, 19]. The motility and distribution mitochondria have repeatedly been suggested to be MT- and IFassociated [5, 19, 31]. However, in heat-shocked

(Fig. 6. on p. 755) Fig. 6. Distributions and morphological changes of mitochondria, Golgi and ER in normal, heat-shocked and recovering 9L RBT cells. Cells were heat-shocked at 458C for 15 min and allowed to recover under normal growing condition for up to 8 h. At di€erent time intervals, mitochondria were stained by a low concentration of DiOC6 (0.1 mg/ml) (left column), Golgi complexes were stained by BODIPY-ceramide (center column), ERs were stained by a high concentration of DiOC6 (2.5 mg/ml) (right column). Nuclei were counterstained with Hoechst. C, Untreated control cells; R0 to R8, 0 to 8 h of recovery after the heat-shock treatment. Shown are typical photographs from three independent experiments. Bar = 20 mm.

T.-T. Wang et al. / The International Journal of Biochemistry & Cell Biology 30 (1998) 745±759

cells where IFs but not MTs are a€ected, the distribution of mitochondria is always coincident to that of the IFs [8, 53] (Collier et al., present study). Therefore it is likely that the transportation of mitochondria is MT-dependent but the positioning of the organelles require intact IFs. The clustering of mitochondria around the nucleus in heat-shocked cells is of interest. As mentioned previously, the HSP70s are ATPase and require ATP hydrolysis for their function as molecular chaperones. Therefore, the massive nuclear localization of these HSPs may require a substantial amount of ATP inside the nucleus for the proper functioning of the HSP70s. In addition, the nuclear localized HSP70 are released from the nucleus in an ATP-dependent manner during the late recovery period [33]. The proximate localized mitochondria may be just in the right position for providing the energy needed. In contrast to that of mitochondria, the distributions of the Golgi apparatus and the ER are less a€ected. Acting together, these two organelles are responsible for most of the modi®cations, intracellular tracking and secretion of the translational products. Moreover, the integrity of MTs is absolutely necessary for the normal functions of the Golgi and the ER [3, 31]. It is not known whether any of these processes is a€ected in heat-shocked cells, however, it is known that the GRPs, which are induced by the disruption of any of the above processes, are not induced in the heat-shocked cells [6, 48]. Our ®nding that MTs, Golgi and ER are least a€ected by heat shock is in good agreement with the above observations. In summary, the current studies enable us to further correlate the induction of HSP70s to the morphological and distribution of cytoskeletal structures and the membraneous organelles. The results also suggest the functional aspects of these changes in cells subjected to and recovering from a heat-shock treatment. Acknowledgements This work is supported by ROC National Science Council grants NSC85-2311-B007-007-

757

B12 (to Y.-K.L.) and NSC85-2311-B007-003-B12 (to A.-S.C.). We thank Dr. A.A. Vyas for her valuable comments on this manuscript.

References [1] J. Becker, E.A. Craig, Heat-shock proteins as molecular chaperones, Eur. J. Biochem. 219 (1994) 11±23. [2] A. Ben-Ze'ev, Animal cell shape changes and gene expression, Bioessays 13 (1991) 207±212. [3] A.D. Bershadsky, J.M. Vasiliev, Cytoskeleton, Plenum Press, New York, 1988. [4] C.R. Brown, R.L. Martin, W.J. Hansen, R.P. Beckmann, W.J. Welch, The constitutive and stress inducible forms of hsp 70 exhibit functional similarities and interact with one another in an ATP-dependent fashion, J. Cell Biol. 120 (1993) 1101±1112. [5] L.B. Chen, Fluorescent labeling of mitochondria, in: Y.L. Wang, S.L. Taylor (Eds.), Methods in Cell Biology, vol. 29: Fluorescence Microscopy of Living Cells in Culture, Part A: Fluorescent Analogs, Labeling Cells and Basic Microscopy, Academic Press, New York, 1989, pp. 103±123. [6] K.D. Chen, J.J. Chu, Y.K. Lai, Modulation of protein phosphorylation and stress protein expression by okadaic acid on heat shock cells, J. Cell. Biochem. 61 (1996) 255± 265. [7] T.J. Cheng, Y.K. Lai, Transient increase in vimentin phosphorylation and vimentin-HSC70 association in 9L rat brain tumor cells experiencing heat-shock, J. Cell. Biochem. 54 (1994) 100±109. [8] N.C. Collier, M.P. Sheetz, M.J. Schlesinger, Concomitant changes in mitochondria and intermediate ®laments during heat shock and recovery of chicken embryo ®broblasts, J. Cell. Biochem. 52 (1993) 297±307. [9] R.A. Coss, W.A.M. Linnemans, The e€ects of hyperthermia on the cytoskeleton: A review, Int. J. Hyperthermia 12 (1996) 173±196. [10] G.C. Flynn, T.G. Cappell, J.E. Rothman, Peptide binding and release by proteins implicated as catalysis of protein assembly, Science 245 (1989) 385±390. [11] Y. Fostinis, P.A. Theodoropoulos, A. Gravanis, C. Stournaras, Heat shock protein HSP90 and its association with the cytoskeleton: A morphological study, Biochem. Cell Biol. 70 (1992) 779±786. [12] E. Fuchs, K. Weber, Intermediate ®lament: Structure, dynamics, function, and disease, Annu. Rev. Biochem. 63 (1994) 345±382. [13] Y. Fukui, Intranuclear actin bundles induced by dimethyl sulfoxide in interphase nucleus of Dictyostelium, J. Cell Biol. 76 (1978) 146±157. [14] C. Georgopoulos, W.J. Welch, Role of the major heat shock proteins as molecular chaperones, Annu. Rev. Cell Biol. 9 (1993) 601±634.

758

T.-T. Wang et al. / The International Journal of Biochemistry & Cell Biology 30 (1998) 745±759

[15] L.A.D. Green, R.K.H. Liem, b-Internexin is a microtubule-associated protein identical to the 70-kDa heatstock cognate protein and the clathrin uncoating ATPase, J. Biol. Chem. 264 (1989) 15210±15215. [16] L.E. Hightower, F.P. White, Cellular-responses to stress: Comparison of a family of 71-kDa proteins rapidly synthesized in rat-tissue slices and canavanine-treated cells in culture, J. Cell. Physiol. 108 (1981) 261±275. [17] L.V. Johnson, M.L. Walsh, B.J. Bockus, L.B. Chen, Monitoring of relative mitochondrial membrane potential in living cells by ¯uorescence microscopy, J. Cell Biol. 88 (1981) 526±535. [18] R.B. Kelly, Microtubules, membrane trac, and cell organization, Cell 61 (1990) 5±7. [19] M.W. Klymkowsky, Intermediate ®laments: New protein, some answers, more questions, Curr. Opin. Cell Biol. 7 (1995) 46±54. [20] T.E. Kreis, Role of microtubules in the organization of the Golgi apparatus, Cell Motil. Cytoskeleton 15 (1990) 67±70. [21] U.K. Laemmli, Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227 (1970) 680±685. [22] J. Landry, P. Chretien, H. Lambert, E. Hicker, L.A. Weber, Heat shock resistance conferred by expression human hsp70 gene in rodent cells, J. Cell Biol. 109 (1989) 7±15. [23] W.C. Lee, K.Y. Lin, K.D. Chen, Y.K. Lai, Induction of HSP70 is associated with vincristine resistance in heatshock 9L rat brain tumor cells, Br. J. Cancer 66 (1992) 653±659. [24] W.C. Lee, Y.C. Lee, M.D. Perng, C.M. Chen, Y.K. Lai, Induction of vimentin modi®cation and vimentin-HSP72 association by withangulatin A in 9L rat brain tumor cells, J. Cell. Biochem. 52 (1993) 253±265. [25] Y.C. Lee, Y.K. Lai, Integrity of intermediate ®laments is associated with the development of acquired thermotolerance in 9L rat brain tumor cells, J. Cell. Biochem. 57 (1995) 150±162. [26] C.G. Li, Z. Werb, Correlation between the synthesis of heat shock proteins and the development of thermotolerance in Chinese hamster ®broblasts, Proc. Natl. Acad. Sci. USA 79 (1982) 3918±3922. [27] J. Liao, L.A. Lowthert, N. Ghori, M.B. Omary, The 70kDa heat shock proteins associate with glandular intermediate ®laments in an ATP-dependent manner, J. Biol. Chem. 270 (1995) 915±922. [28] S. Lindquist, E.A. Craig, The heat-shock proteins, Annu. Rev. Genet. 22 (1988) 631±677. [29] N.G. Lipsky, R.E. Pabano, A vital stain for the Golgi apparatus, Science 228 (1985) 745±747. [30] R.Y. Liu, X.C. Li, L.G. Li, G.C. Li, Expression of human HSP70 in rat ®broblast enhances cell-survival and facilitates recovery from translational and transcriptional inhibition following heat-shock, Cancer Res. 52 (1992) 3667±3673.

[31] M. Mizuno, S.J. Singer, A possible role for stable microtubules in intracellular transport from the endoplasmic reticulum to the Golgi apparatus, J. Cell Sci. 107 (1994) 1321±1331. [32] M. Moritz, M.B. Braunfeld, J.W. Sedat, B. Alberts, D.A. Agard, Microtubule nucleation by gamma-tubulin-containing rings in the centrosome, Nature 378 (1995) 638± 640. [33] K. Ohtsuka, K.R. Utsumi, T. Kaneda, H. Hattori, E€ect of ATP on the release of hsp70 and hsp40 from the nucleus in heat-shock HeLa cells, Exp. Cell Res. 209 (1993) 357±366. [34] M. Osborn, K. Weber, Dimethlsulfoxide and the ionophore A23187 a€ect the arrangement of actin and induce nuclear actin paracrystals in PtK2 cells, Exp. Cell Res. 129 (1980) 103±114. [35] D. Pekkala, B. Heath, J.C. Silver, Changes in chromatin and the phosphorylation of nuclear proteins during heat shock of Achlya ambisexualis, J. Cell Biol. 104 (1984) 1198±1205. [36] H.R.B. Pelham, Speculations on the functions of the major heat shock and glucose-regulated proteins, Cell 46 (1986) 959±961. [37] H.R.B. Pelham, Functions of the hsp70 protein family: An overview, in: R.I. Morimoto, A. Tissieres, C. Georgopoulos (Eds.), Stress Protein in Biology and Medicine, Cold Spring Harbor, New York, 1990. [38] T.D. Pollard, J.A. Cooper, Actin and actin-binding proteins. A critical evaluation of mechanisms and functions, Annu. Rev. Biochem. 55 (1986) 987±1035. [39] J. Rassow, N. Pfanner, Molecular chaperones and intracellular protein translocation, Rev. Physiol. Biochem. Pharmacol. 126 (1995) 199±264. [40] J. Rassow, W. Voo, N. Pfanner, Partner proteins determine multiple functions of Hsp70, Trends Cell Biol. 5 (1995) 207±212. [41] C. Rosette, M. Karin, Cytoskeletal control of gene expression: Depolymerization of microtubule activates NFkappa B, J. Cell Biol. 128 (1995) 1111±1119. [42] J.W. Sanger, J.M. Sanger, T.E. Kreis, B.M. Jockusch, Reversible translocation of cytoplasmic actin into the nucleus caused by dimethyl sulfoxide, Proc. Natl. Acad. Sci. USA 77 (1980) 5268±5272. [43] U. Scheer, H. Hinssen, W.W. Franke, B.M. Jockusch, Microinjection of actin-binding proteins and actin antibodies demonstrates involvement of nuclear actin in transcription of lampbruch chromosome, Cell 39 (1984) 111± 122. [44] O. Skalli, R.D. Goldman, Recent insights into the assembly, dynamics, and function of intermediate ®lament networks, Cell Motil. Cytoskeleton 19 (1988) 67±79. [45] M. Terasaki, J.D. Song, J.R. Wong, M. Weiss, L.B. Chen, Localization of endoplasmic reticulum in living and glutaraldehyde ®xed cells with ¯uorescent dyes, Cell 38 (1984) 101±108. [46] G.P. Thomas, W.J. Welch, M.B. Mathews, J.R. Feramisco, Molecular and cellular e€ects of heat shock

T.-T. Wang et al. / The International Journal of Biochemistry & Cell Biology 30 (1998) 745±759

[47]

[48] [49]

[50] [51]

and related treatments of mammalian tissue-culture cells, Cold Spring Harbor Symp. Quant. Biol. 46 (1982) 985± 996. P.M.P. Van Bergenen Henegouwen, W.J.R.M. Jordi, G. Van Dongen, F.C.S. Ramaekers, H. Amesz, W.A.M. Linnemans, Studies on a possible relationship between alterations in the cytoskeleton and induction of heat shock protein synthesis in mammalian cells, Int. J. Hyperthermia 1 (1985) 69±83. S.S. Watowich, R.I. Morimoto, Complex regulation of heat shock-responsive and glucose-responsive genes in human cells, Mol. Cell Biol. 8 (1988) 393±405. M. Weizsaecker, D.F. Deen, M.L. Rosenblum, T. Gutin, P.H. Gutin, M. Barker, The 9L rat brain tumor: Description and application of an animal model, J. Neurol. 224 (1981) 183±192. W.J. Welch, Mammalian stress response: Cell physiology, structure/function of stress proteins and implication for medicine and disease, Physiol. Rev. 72 (1992) 1063±1081. W.J. Welch, J.R. Feramisco, Disruption of the three cytoskeletal networks in mammalian cells does not a€ect

[52]

[53]

[54] [55]

759

transcription, translation, or protein translocation changes induced by heat shock, Mol. Cell. Biol. 5 (1985) 1571±1581. W.J. Welch, L.A. Mizzen, Characterization of the thermotolerant cell. II. E€ects on the intracellular distribution of heat-shock protein 70, intermediate ®laments and small nuclear ribonucleoprotein complexes, J. Cell Biol. 106 (1988) 1117±1130. W.J. Welch, J.P. Suhan, Morphological study of the mammalian stress response: Characterization of changes in cytoplasmic organelles, cytoskeleton, and nucleoli and appearance of intranuclear acting ®laments in rat ®broblasts after heat shock treatment, J. Cell Biol. 101 (1985) 1198±1211. W.J. Welch, J.P. Suhan, Cellular and biochemical events in mammalian cells during and after recovery from physiological stress, J. Cell. Biol. 103 (1986) 2035±2053. W.J. Welch, J.R. Feramisco, S.H. Blose, The mammalian stress response and the cytoskeleton: Alterations in intermediate ®laments, Ann. N. Y. Acad. Sci. 455 (1985) 57± 67.