INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

AT THE CELLULAR LEVEL, failing energy metabolism and ATP depletion are the earliest cell-damaging factors of ischemic insults. In vivo, severe depletion of ATP is a proteotoxic stress that leads to dysfunction, destabilization, and aggregation of many cellular proteins, including enzymes, ion pumps, and constituents of cytoskeletal and contractile structures (1, 9, 17, 27). Sustained lack of ATP is obviously lethal for the cell. On the contrary, a transient (reversible) drop in cellular ATP can confer tolerance to the next energy-depriving exposure, with heat shock proteins (HSPs) being involved in such an adaptive response (reviewed in Ref. 17). The HSP-involving cellular response appears to contribute to delayed ischemic tolerance (the second window of protection) found after heat or ischemic preconditioning in the myocardium. Like high temperature, cellular ATP depletion activates the heat shock transcription factor 1 (HSF1) that afterward induces HSP expression in the recovering cells (4, 36). Most HSPs are molecular chaperones stabilizing protein molecules under heat shock conditions in vitro and in vivo, and the same chaperone activity may protect HSP-enriched cells in the case of other proteotoxic stresses, e.g., ischemia or ATP-depleting treatments with inhibitors of energy metabolism. Despite the numerous reports describing HSP70-mediated cytoprotection against ischemic injury (for a review, see Refs. 5, 17, and 19), the precise mechanism of this phenomenon remains to be determined. Previously, suppression of ATP depletion-provoked protein insolubilization was found in heat-preconditioned ascites tumor cells (7, 16). Although it was speculated in those reports that such an effect of the preconditioning is due to stress-inducible HSPs (e.g., HSP70), no serious evidence was provided. Likewise, it has never been studied whether excess HSP70 in cells attenuates the proteotoxic effects of ischemia-associated ATP depletion, i.e., protects intracellular proteins from denaturation and aggregation during the stress. So far, it is also unknown how the cell viability under ischemia-like (or energy-depleting) conditions correlates with the in vivo chaperone activity of intracellular HSP70.

These two latter issues were addressed in the present study. We hypothesized that excess HSP70 in stress-preconditioned or HSP70-overexpressing cells can preserve cellular proteins from the ATP depletion-induced aggregation leading to cell death. To examine this hypothesis on an ischemia-relevant model, we used a rat embryonic heart-derived H9c2 line of myoblasts that retain some features of cardiac cells and can acquire tolerance to simulated ischemia (23) and thermal or oxidative stress (10, 34) after HSP-inducing heat pretreatments. Besides heat preconditioning, we also employed metabolic preconditioning to cause the transient (reversible) ATP depletion as an alternative HSP-inducing stress that to some extent mimics ischemic preconditioning in vivo. To evaluate contribution of the HSP induction, we treated the stress-preconditioned cells with quercetin, an inhibitor of HSF1 (13, 26). In parallel, to increase intracellular HSP70 by an HSF1-independent method, we overexpressed human inducible HSP70 in H9c2 cells using plasmid- or virus-based vectors. It was previously found that ATP depletion in cells renders many cellular proteins less extractable with a Triton X-100-containing buffer (14, 15). Such insolubilization of cellular proteins is due to their aggregation and correlates with the intensity of cell death under ATP-depleting stress (14, 17). Therefore, the stress-induced increase in the detergent insolubility of cellular proteins was quantified here to assess how the stressful preconditioning and HSP70 overexpression affect the proteotoxic impact within the energy-deprived cells.

In addition to monitoring of the total protein aggregation in the ATP-depleted myoblasts, we explored in them the catalytic activity and the solubility of a reporter enzyme, firefly (Photinus pyralis) luciferase, transiently expressed following transfection. This thermolabile enzyme is used as an in situ probe, allowing us to study protein denaturation, aggregation, and refolding in heat- or chemically stressed cells (2, 24, 25, 28, 29). Furthermore, firefly luciferase, when expressed in mammalian fibroblasts, aggregates and loses activity during cellular ATP depletion (27) and, thus, is a sensitive marker of the stress-associated proteotoxicity. We therefore introduced firefly luciferase into H9c2 myoblasts to evaluate effects of the preconditioning and the HSP70 overexpression on the proteotoxicity of ischemia-mimicking ATP depletion. Plasmids expressing recombinant forms of luciferase with either cytoplasmic or nuclear localization (25) were used to separately probe the situation within cytoplasmic and nuclear compartments of the ATP-depleted cells. To elucidate the role of inducible HSP70 alone, H9c2 cells were cotransfected with luciferase and human HSP70 simultaneously as previously employed in heat shock studies on other cell lines (24, 29).

We demonstrate here that pretreatments leading to HSP70 accumulation in the cells can reduce the protein aggregation resulting from cellular ATP loss. This attenuation of the ATP depletion-associated proteotoxicity seems to be HSP70 mediated and correlates with the elevated cell resistance to prolonged energy deprivation. Improved formation of nonaggregable (soluble) complexes between excess HSP70 and some damaged or instable proteins in the cytosol and nucleosol is suggested as a probable cause of the attenuated protein aggregation under ATP depletion in the HSP70-enriched cells. Analogous mechanisms may contribute to the phenomenon of delayed ischemic tolerance arising in the stress-preconditioned heart when the level of cardiac HSP70 is transiently increased.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Cells. The embryonic rat heart-derived H9c2 line of myoblasts was obtained from American Type Culture Collection (CRL-146; ATCC, Rockville, MD). The cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with sodium bicarbonate and 10% fetal bovine serum (GIBCO BRL) in a humidified atmosphere of 5% CO₂ in air at 37°C. The preconfluent myoblast cultures were used for the experiments.

Stressful treatments and recovery. For heat preconditioning, dishes or tubes containing the adherent cells were plunged into a thermostatic water bath at 43°C for 30 min and then returned into a CO₂ incubator at 37°C. For preconditioning by metabolic (ATP-depleting) stress, the cells were exposed to 3 h of incubation at 37°C with glucose-free DMEM containing 3% fetal bovine serum and 20 µM carbonyl cyanide m-chlorophenylhydrazone (CCCP), and then this medium was removed and the cells were washed twice with normal DMEM and added to the full growth medium. Recovery periods following either priming stress continued for 16-18 h before the challenging ATP depletion. Quercetin was added at a 30 µM concentration to some samples of the cells before the stressful preconditioning (21, 22); this drug was present in the incubation medium during the priming stress and the next 10 h, and then the drug-containing medium was replaced by the usual growth medium. To perform the challenging ATP depletion, we placed the cells in the CCCP-containing medium supplemented with 10 mM 2-deoxyglucose and incubated them at 37°C in a CO₂ incubator (21, 22).

Plasmids and transfection. Plasmid pRSVLL/V encoding cytoplasmic localized firefly luciferase (cyt-luciferase) was kindly provided by Dr. S. Subramani (University of California, San Diego, CA). Construction of plasmid pRSVnlsLL/V encoding firefly luciferase fused to a nuclear localization sequence (nuc-luciferase) has been described previously (25). For in situ control of the transfection efficiency and the compartment-specific localization of the expressed products, plasmid constructs encoding cyt- or nuc-luciferase fused at their COOH termini to enhanced green fluorescent protein (EGFP) were used (constructed by Dr. E. A. A. Nollen, University of Groningen, Groningen, The Netherlands). Distribution of the EGFP-luciferases and the diamidinophenylindole (DAPI)-labeled cell nuclei was viewed on an Opton III fluorescence microscope (Karl Zeiss, Oberkochen, Germany). Plasmid pCMV70 with fragment encoding cDNA for human inducible HSP70 was constructed as previously described (24).

Infection of the cells with virus-based vectors. Human inducible HSP70 cDNA, or a gene encoding the green fluorescent protein (GFP), under the control of the cytomegalovirus promoter, were inserted into the HSV-1 genome as previously described (37). Cell culture dishes (35 mm) with equal numbers of the cells, in the state of preconfluent monolayer, were washed twice with sterile phosphate-buffered saline (PBS) from the growth medium, and then PBS was replaced with 1.5 ml of serum-free DMEM. The virus-based vectors expressing either HSP70 or GFP were added in the dishes at concentration of 30 plaque-forming units/ml and mixed into the serum-free medium. The cells were in contact with the viruses for 1 h in a humidified atmosphere with 5% CO₂ at 37°C. After that exposure, the virus-containing medium was removed and replaced with normal growth medium for 24 h before to the ATP-depleting treatments (6). The efficiency of infection and expression of the products were tested by flow cytometry and Western blotting.

Measurements of luciferase activity and cellular ATP. The luciferase transfectants adherent in tubes were instantly cooled in ice, briefly washed with ice-cold PBS, and then lysed in 0.5 ml of ice-cold buffer A [25 mM H₃PO₄-Tris, pH 7.8, 10 mM MgCl₂, 1 mM EDTA, 1% (vol/vol) Triton X-100, and 15% (vol/vol) glycerol] containing 0.5% 2-mercaptoethanol (24, 25). The lysates were frozen and kept at 20°C until determination of luciferase activity. The latter was measured in a luminometer for 10 s after addition of 0.1 ml of buffer A containing substrates [1.25 mM ATP and 87 µg/ml luciferin (Sigma)] to 0.15 ml of the cell lysates (24).

Cell fractionation with Triton X-100. The ATP depletion-induced protein aggregation was assessed upon the increase in the cellular protein insolubility (14, 15) by using the method adapted for adherent cells (24, 27). Cell culture dishes (35 mm) with equal numbers of the cells were washed twice with PBS and then lysed and scraped in 0.3 ml of the ice-cold buffer A devoid of glycerol. The lysates were centrifuged at 12,000 g for 15 min at 4°C, which divided them into pellets (Triton X-100-insoluble fractions) and supernatants (Triton X-100-soluble fractions) (24, 27). The pellets were dissolved in 60 µl of 6 M urea, and the protein contents there were determined using the BCA (bicinchoninic acid) kit (Sigma) (7).

SDS gel electrophoresis and Western blot analysis. The samples prepared from the total cell lysates or cellular fractions were run by electrophoresis in a Laemmli system with SDS-10% polyacrylamide gel under reducing conditions. The separated proteins were electrotransferred from the gel slabs onto 0.45-µm nitrocellulose membrane (Bio-Rad). For integral detection of the endogenous HSC70 (heat shock cognate protein 70)/HSP70, monoclonal antibody N27F3-4 (StressGen) was used in a 1:2,000 dilution. The human inducible HSP70 was detected in the cell lysates with monoclonal antibody C92F3A-5 (StressGen) diluted 1:1,500. Cyt- and nuc-luciferases were detected with rabbit polyclonal antibodies (Promega) in a 1:1,500 dilution. Anti-mouse Ig- and anti-rabbit Ig-peroxidase conjugates and an ECL (enhanced chemiluminescence) kit were employed to develop the antigen band tracks on X-ray film (all from Amersham). The track images were digitized by means of a flatbed scanner (Mustek 600 II CD). The digitized images were then quantitatively analyzed using the NIH Image software.

Determination of cell death and survival. Cell death during the challenging ATP depletion was quantified by counting the number of dead cells unable to exclude the dye trypan blue as a percentage of the total number of cells. In the case of EGFP-luciferase/HSP70 cotransfection, the percentage of cells killed by ATP depletion was counted with the use of the fluorescence microscope after staining with acridine orange (AO; 3 µg/ml) and propidium iodide (PI; 10 µg/ml), as described previously (8). At least 10 microscopic fields per plate (300-500 cells) were counted, and the recounting was repeated 3-4 times.

Metabolic ³⁵S labeling and immunoprecipitation. Equal numbers of the cells growing in 60-mm culture dishes were preconditioned by heat or metabolic stress as described in Stressful treatments and recovery. Seven hours after the start of the preconditioning, the treated and control cells were incubated for eight hours at 37°C in methionine-free DMEM (Flow Laboratories) supplemented with 5% normal DMEM, 10% fetal bovine serum, and 40 µCi/ml [³⁵S]methionine (produced with a specific activity of 950 mCi/mmol by Institute of Physical Energetics, Obninsk, Russia). The label-containing medium was then harvested, and the cells were washed three times with PBS and three times with the full growth medium, followed by 3 h of incubation under normal growth conditions. These radiolabeled cells were extracted with 0.6 ml of the ice-cold glycerol-free buffer A containing a protease inhibitor cocktail (Sigma), and then the Triton X-100-soluble fractions were obtained according to the technique described in Cell fractionation with Triton X-100.

Statistical analysis. All quantitative results are expressed as means ± SE of 4-6 separate experiments. Statistical differences between compared groups of the cells were analyzed using ANOVA (a multiway analysis of variance or covariance). ANOVA significance was applied when P < 0.05 and confirmed with the F-test.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Elevated resistance to ATP depletion-induced cell death in the stress-preconditioned and HSP70-overexpressing cells. Fluctuations in the ATP levels in H9c2 cells during and after heat preconditioning (43°C for 30 min) were insignificant (data not shown). Preconditioning of the cells by metabolic stress (glucose starvation + uncoupling of oxidative phosphorylation with CCCP) caused severe but reversible decrease in cellular ATP (Fig. 1A). Both preconditioning treatments followed by a 16- to 18-h period of recovery resulted in an approximately fourfold increase in the intracellular content of HSC70/HSP70 (Fig. 2A, Table 1). Probably, the content of other HSPs also increases in the cells following either stressful preconditioning, but their expression was not analyzed in the present study.

View larger version (15K): [in this window] [in a new window]   Fig. 1.   Relative ATP content in H9c2 cells subjected to metabolic preconditioning (A) or challenging ATP depletion (B). Note that 30 µM quercetin does not affect the ATP replenishment following metabolic stress (A) and the delayed stressful preconditioning. The infection with virus-based vectors expressing green fluorescent protein (GFP) or heat shock protein 70 (HSP70) does not attenuate the decline of cellular ATP during the challenging ATP depletion (B).

View larger version (20K): [in this window] [in a new window]   Fig. 2.   Accumulation of HSP70 in H9c2 cells following delayed stressful preconditioning or virus infection and the effects of quercetin. Cells were subjected to heat or metabolic preconditioning (A) or were infected with the virus-based vectors (B) without or with 30 µM quercetin, and then, after an 18-h recovery period, cells were lysed as described in MATERIALS AND METHODS. Aliquots of the lysates from equal amounts of the untreated (control) and preconditioned cells were loaded in each well and analyzed by Western blotting with enhanced chemiluminescence (ECL) by using antibody recognizing both heat shock cognate protein 70 (HSC70) and HSP70 (top blots) or antibody with specificity to HSP70 (bottom blots).

View larger version (31K): [in this window] [in a new window]   Fig. 3.   Effects of delayed stressful preconditioning and virus infection (without or with quercetin) on the ability of H9c2 cells to exclude trypan blue dye during challenging ATP depletion. Cells were subjected to heat or metabolic preconditioning in the absence or presence of 30 µM quercetin, and then, after an 18-h recovery period, cells were exposed to challenging ATP depletion as described in MATERIALS AND METHODS. The percentage of dead cells at 5, 6, and 7 h of the ATP-depleting treatment was counted using a trypan blue exclusion test. Data presented are means ± SE of 5 independent experiments (* P < 0.05).

View larger version (41K): [in this window] [in a new window]   Fig. 4.   Effects of delayed stressful preconditioning and virus infection (without or with quercetin) on the ability of H9c2 cells to form colonies after sustained ATP depletion. The numbers of surviving colonies for each group were normalized to the untreated cells: nonpreconditioned (control) cells exposed to 6-h ATP depletion were taken as 100% of possible surviving colonies. Colony survival in the groups of preconditioned cells is expressed as a percentage of that in the nonpreconditioned (control) cells. Data presented are means ± SE of 4 separate experiments (* P < 0.05).

Reduced aggregation of endogenous cellular proteins during ATP depletion in the tolerant (stress-preconditioned or HSP70-overexpressing) cells. Severe depletion of cellular ATP evokes massive damage and aggregation of proteins within the stressed cells as revealed by the accumulation of many initially soluble proteins in the Triton X-100-insoluble cellular fraction (1, 14-16, 27). Here, the stress-induced aggregation of endogenous proteins was quantified by comparing the relative content of protein in the Triton X-100-insoluble fractions from the ATP-depleted cells with that from the unstressed ones (7, 16). The basal level of total Triton X-100-insoluble protein in the cells following the delayed stressful preconditioning or virus infection, or the quercetin treatment, was the same as in the nonpretreated control cells (not shown). The data presented in Table 2 demonstrate that the detergent-insoluble protein component does increase during cellular ATP depletion. The extent of this insolubilization was significantly reduced in the stress-preconditioned and HSP70-overexpressing cells compared with control (the nonpreconditioned cells and the GFP-overexpressing cells, respectively).

Stressful preconditioning and HSP70 overexpression diminish the rate of inactivation and extent of insolubility of cyt- and nuc-luciferase in the tolerant ATP-depleted cells. Firefly luciferase is an ATP-dependent reporter enzyme that, when expressed in mammalian cells, is inactivated and insolubilized during heat or chemical stress (2, 24, 25, 28, 29) or depletion of cellular ATP (27). For probing of the situation within the cytoplasm and the nucleus separately, we transfected the cells with plasmid constructs encoding luciferases engineered for expression in either the cytoplasm (cyt-luciferase) or the nucleus (nuc-luciferase) (25). This specific compartmentalization of the expressed products and the transfection efficiency were checked by microscopic analysis of the distribution of EGFP-cyt- and EGFP-nuc-luciferases in the cells with DAPI-labeled nuclei (Fig. 5). The transfection usually yielded up to 20-25% positive cells (see Fig. 5), and its procedure did not affect the time course of the subsequent challenging ATP depletion (not shown).

View larger version (80K): [in this window] [in a new window]   Fig. 5.   Double-label fluorescence images showing expression of enhanced GFP (EGFP)-cytoplasmic (cyt)-luciferase and EGFP-nuclear (nuc)-luciferase in the preparations of diamidinophenylindole (DAPI)-stained H9c2 cells. A and B: EGFP fluorescence of cyt-luciferase distributed in the cytoplasm of 2 transfectants (A) among the cell population stained with DAPI for total labeling of the nuclei (B). C and D: EGFP fluorescence of nuc-luciferase localized to the nuclei of 2 transfectants (C) among the cell population stained with DAPI for total labeling of the nuclei (D). Arrows denote locations of the same nuclei in the EGFP- and DAPI-stained cells (transfectants) in couples of the double-labeled images. Original magnification, ×500.

View larger version (35K): [in this window] [in a new window]   Fig. 6.   Effects of delayed stressful preconditioning (without or with quercetin) and HSP70 overexpression on the rate of inactivation of cyt- (A) and nuc-luciferase (B) in ATP-depleted H9c2 cells. The control and stress-preconditioned transfectants were subjected to 4 h of ATP depletion, and aliquots of the ATP-depleted cells were lysed by the hour. The catalytic capability of cyt- and nuc-luciferase was then measured in the lysates with addition of exogenous ATP and luciferin (see MATERIALS AND METHODS). Note the retarded inactivation of both forms of luciferase during ATP depletion in the stress-preconditioned or HSP70-overexpressing transfectants and the abolishing effects of quercetin in the groups with stressful preconditioning. Data presented are means ± SE of 6 independent experiments (* P < 0.05).

View larger version (21K): [in this window] [in a new window]   Fig. 7.   Effects of delayed heat preconditioning (without or with quercetin) and HSP70 overexpression on insolubilization of cyt- (left) and nuc-luciferase (right) in ATP-depleted H9c2 cells. The non-ATP-depleted transfectants (0 h) and transfectants exposed to 2 or 3 h of ATP depletion were fractionated with the Triton X-100-containing buffer as described in MATERIALS AND METHODS. After the cellular fractions were run in electrophoresis, cyt- and nuc-luciferase were immunodetected in the supernatants (S) comprising Triton X-100-soluble cellular material and in the Triton X-100-insoluble pellets (P) by Western blotting with ECL. Note the attenuated insolubilization of both forms of luciferase during ATP depletion in the heat-preconditioned transfectants and transfectants overexpressing HSP70, and the abolishing effects of quercetin toward the heat preconditioning. Although very similar results including the abolishing effects of quercetin were obtained with metabolic preconditioning, no influence of quercetin was found in the case of HSP70 overexpression (not shown).

Tolerant (stress-preconditioned or HSP70-overexpressing) cells, when depleted of ATP, have larger amounts of soluble HSP70 and its complexes with other proteins. To prove the direct involvement of HSP70 in the described effects, we compared the behavior of this chaperone during ATP depletion in the control and tolerant cells. It was previously shown in various cell lines that a major part of HSC70/HSP70 is insolubilized in response to ATP depletion (14, 15, 27), but this effect was never explored in cells rendered tolerant. Figure 8 demonstrates that, whereas H9c2 myoblasts are deprived of ATP, HSC70/HSP70 also accumulates into the Triton X-100-insoluble cellular fraction. There were not significant differences in the levels of the insolubilized HSP70 between the control and tolerant cells (Fig. 8). However, compared with control, both the stress-preconditioned and HSP70-overexpressing cells contain much more HSP70 in the Triton X-100-soluble fractions isolated at different times of the ATP-depleting exposure. In other words, the tolerant cells, being HSP70-enriched before ATP depletion, retain the soluble pool of this chaperone longer during sustained ATP depletion (Fig. 8).

View larger version (12K): [in this window] [in a new window]   Fig. 8.   Effects of delayed stressful preconditioning and HSP70 overexpression on the stress-induced redistribution of HSC70/HSP70 between the Triton X-100-soluble and -insoluble cellular fractions during ATP depletion in H9c2 cells. Equal numbers of the non-ATP-depleted cells (0 h) and the cells subjected to 1, 2, or 3 h of ATP depletion were fractionated with the Triton X-100-containing buffer as described in MATERIALS AND METHODS. After the cellular fractions were run in electrophoresis, HSC70/HSP70 was immunodetected in the supernatants (S) comprising Triton X-100-soluble cellular material and in the insoluble pellets (P) by using Western blotting with ECL along with the monoclonal antibody N27F3-4. Note the longer sojourn of HSC70/HSP70 in the soluble cellular fractions of the tolerant (stress preconditioned or HSP70 overexpressing) cells deprived of ATP. Besides these blots, very similar results were obtained in 4 independent experiments.

View larger version (69K): [in this window] [in a new window]   Fig. 9.   Immunoprecipitation patterns showing the effects of delayed stressful preconditioning, HSP70 overexpression and exogenous ATP on formation of soluble complexes between HSC70/HSP70 and other proteins in ATP-depleted H9c2 cells. The Triton X-100-soluble fractions isolated from equal numbers of the non-ATP-depleted cells (lane 1) and the cells depleted of ATP for 2 h (lanes 2 and 3) were used for immunprecipitation with the anti-HSC70/HSP70 antibodies. The patterns in lane 3 for each group are derived from the soluble fractions of the ATP-depleted cells in which immunoprecipitation was performed in the presence of exogenous ATP (1 mM). After electrophoresis was run, the immunoprecipitate patterns were visualized by either 35S autoradiography (A) or Western blotting with ECL (B) as described in MATERIALS AND METHODS. Note the more abundant protein material coprecipitated with excess HSC70/HSP70 from the soluble fractions of the tolerant (stress preconditioned or HSP70 overexpressing) cells deprived of ATP (A and B, lane 2 for each group). In A, arrows mark protein bands whose intensities increased in the patterns of the tolerant cells (see lane 2 for each group); the same bands disappeared if immunoprecipitation was performed in the presence of exogenous ATP (see lane 3 for each group). Autoradiographs and blots shown were obtained in 1 of 4 separate experiments with very similar results; luc, luciferase.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Preconditioning of H9c2 cells by nonlethal heat or metabolic stress transiently enhances the HSP70 expression. In parallel, cell resistance to the cytotoxic and proteotoxic effects of the challenging ATP depletion is transiently acquired, thus suggesting a casual link between these events. Here we discuss molecular mechanisms of the in vivo ATP depletion-associated proteotoxicity, its link to the stress-induced cell death, and the role of HSP70 in protection from the protein aggregation in energy-deprived cells.

In 1994, Nguyen and Bensaude (27) described two reporter enzymes, Eschericia coli -galactosidase and firefly luciferase, that, both being Triton X-100-soluble in unstressed cells, are insolubilized during ATP depletion. Although those foreign enzymes were artificially expressed in mammalian cells, their insolubilization appears to reflect the in situ response of many endogenous cellular proteins to the stress. The data of Table 2 reveal that rather a large percentage of total cellular protein lost solubility because of ATP depletion. Such reaction implies at least two processes that may occur in parallel: some initially soluble proteins 1) form aggregates that precipitate in Triton X-100-containing solutions and/or 2) undergo translocation to the detergent-resistant cellular compartments such as the cytoskeleton, chromatin, and nuclear matrix. Probably, both the former and latter are caused by structural changes in some instable protein molecules resulting from ATP depletion per se or its harmful consequences (e.g., ionic imbalance including acidosis). These structural changes promote undesirable intermolecular interactions leading to aggregation and abnormal translocation of a number of affected intracellular proteins. This may result in dysfunction of the involved proteins through the loss of their native conformations and through the inability of aggregated or abnormally localized proteins to interact with their natural substrates and cofactors.

The question arises, which intracellular proteins aggregate and are insolubilized in response to ATP depletion? Among mature cytosolic proteins, this feature is seen for the 68-kDa double-stranded RNA-dependent protein kinase (17, 27) and a major glycolytic enzyme, glyceraldehyde-3-phosphate dehydrogenase (1). The latter, being an abundant cytosolic enzyme in unstressed cardiomyocytes, associates with myofibrils as a result of ischemia-provoked ATP depletion (1). Within the nuclear compartment, proteins of the chromatin and/or nuclear matrix may coaggregate with the stress-sensitive nucleosolic enzymes; otherwise, the latter may aggregate with each other without sticking to the originally insoluble structures. As for nonenzymatic proteins, the cytosolic pools of cytoskeletal proteins such as G-actin, myosin, vinculin, and -actinin can also supply material for aggregating in ATP-depleted cells (9, 12, 14-17). Despite the fact that the preexisting cytoskeleton undergoes disintegration in ATP-depleted cells, its debris remains insoluble and may entrap surrounding proteins, thus aggravating the total aggregation (17).

Importantly, the proteotoxic effects of ATP depletion correlate with its cytotoxicity as indicated by comparison of the data of Tables 2 and 3 and Figs. 3, 4, 6, and 7. The same suggestion came from the previous studies performed on murine tumor cells (7, 14, 16). Quercetin completely abolished both the HSP70-inducing and the cytoprotective and anti-proteotoxic effects of the stressful preconditioning, suggesting that the inducible HSP(s) can alleviate injury of ATP-depleted cells by reducing the protein aggregation. Although quercetin could exert other effects besides the HSF1 inhibition, we found that, in our model, this drug did not affect the time course of in vivo ATP depletion/replenishment, the basal level of the insoluble cellular protein, or the cell viability and protein aggregation under the challenging ATP depletion in the cells of control groups (non-stress-preconditioned or GFP-overexpressing H9c2 cells). While the HSF1-independent (i.e., plasmid or viral vector induced) overexpression of inducible HSP70 was already sufficient to moderate the cytotoxic and proteotoxic influence of ATP depletion in H2c9 cells, the quercetin treatment was not able to prevent these HSP70 overexpression-mediated effects. Taken together, these results allow us to attribute the cytoprotective and anti-proteotoxic effects of the stressful preconditioning to stress-induced HSP(s) (e.g., HSP70). Although other inducible HSPs besides HSP70 can also be involved in cellular defense/repair from injurious consequences of ATP depletion, HSP70 appears to play a major role in the attenuation of the ATP depletion-induced protein aggregation revealed after the stressful preconditioning. Such a conclusion follows from the experiments with overexpression of HSP70 alone, which yielded both the intracellular HSP70 accumulation and the impairment of the ATP depletion-associated proteotoxicity quite comparable with those in the stress-preconditioned cells.

Intriguingly, HSP70 is known as an ATP-dependent chaperone, and, at first glance, it seems unclear how it could combat the proteotoxicity in the case of lack of ATP. However, previous heat shock studies (20, 33) showed that overexpression of the deletion mutant of human HSP70 devoid of the ATP-binding domain confers thermoresistance and reduced intranuclear protein aggregation in heat-shocked Rat-1 cells. Because the inability of the mutant to bind and hydrolyze ATP did not yet abolish its protective effects, the authors suggested that the mutant is still able to form complexes with heat-denatured cellular proteins, thus reducing protein aggregation during heat shock (20, 33). This heat shock model with the deletion mutant of HSP70 rather closely mimics the present situation with ATP depletion in HSP70-enriched H9c2 cells. Taking into account that HSP70 binds ADP much stronger than ATP (11, 31, 32) and the ATP/ADP ratio obviously declines in energy-deprived cells (15, 18), it seems likely that under severe depletion of ATP in vivo, most (if not all) HSC70/HSP70 molecules are in the "ADP state." The latter is characterized as having a slower on-rate but also slower off-rate of substrate than HSP70 in the "ATP state." Therefore, excess HSP70 in the ADP state within the ATP-depleted tolerant cells may form stable complexes with proteins affected by the stress. In favor of this idea, we find 1) the longer sojourn of excess HSP70 in the soluble fractions of the tolerant cells undergoing ATP depletion, 2) the larger amounts of soluble protein material coimmunoprecipitating with excess HSP70 from the tolerant cells deprived of ATP, and 3) the dissociating effect of exogenous ATP on the soluble HSP70-protein complexes formed for lack of cellular ATP (see Figs. 8 and 9). Apparently, the proteins bound to HSP70 are thereby protected from aggregation with each other and/or sticking to the insoluble cytoskeletal/nuclear structures or debris despite the fact that the chaperone cannot undergo an ATP-ADP cycle in the ATP-deprived cells. As for the stress-damaged proteins, their stable association with HSP70 may retain them in a soluble, folding-competent state that should save these proteins from the aggregation during cellular ATP depletion. Besides the stress-damaged mature proteins, the preexisting pool of nascent polypeptide chains can recruit some part of HSP70 in ATP-depleted cells (3), and the chains bound to the chaperone may be better preserved from the stress-induced aggregation. Very similar mechanisms appear to act in the luciferase model: during ATP depletion, the stress-sensitive enzyme can be either involved in the aggregation cascade (inactivation and insolubilization) or trapped by HSP70 (protection). We suppose that during ATP depletion in the HSP70-enriched cells, the large part of inactivated luciferase still remains in a folding-competent state being bound to HSP70, and then, under cell lysis into the ATP-containing buffer, the instant ATP-mediated dissociation of the enzyme-chaperone complexes occurs, with released luciferase being detected as catalytically active in the assay. The immunoprecipitation experiments demonstrating the improved HSP70-luciferase complex formation in the tolerant cells depleted of ATP as well as the complex-disrupting effect of exogenous ATP (Fig. 9B) provide strong arguments in support of such a supposition.

Although the protection of two model enzymes within the HSP70-enriched cells cannot be directly linked to what happens with the cells' vital proteins, our data do indicate that excess HSP70 can diminish the ATP depletion-induced protein aggregation in both the cytoplasm and the nucleus. Moreover, evident correlation is observed between the cell viability under the challenging ATP depletion and the in situ chaperone potential of HSP70 toward cyt- and nuc-luciferase during the stress (see Figs. 6 and 7 and Table 3). Extrapolating the luciferase data to endogenous proteins of the ATP-depleted cells (see Table 2), we suggest that HSP70 can protect stress-sensitive cytoplasmic and nuclear proteins that otherwise are involved in the detrimental aggregation leading to necrotic cell death. The fact of partial migration of cytosolic HSP70 into the nuclei of ischemia-stressed rat cardiomyocytes (35) supports the suggestion that nuclear proteins can also become the chaperone targets under the energy-depriving stress. It seems likely that in the case of ischemia in vivo, excess HSP70 would also be able to minimize the proteotoxic impact of cellular ATP depletion, thereby allowing HSP70-enriched cells to better tolerate an acute phase of ischemic insults. Such speculations generate an additional reason for development of clinically applicable ways enabling to increase the HSP70 level in ischemia-attacked tissues of patients.