Heat Shock and Caloric Restriction Have a Synergistic Effect on the Heat Shock Response in a sir2.1-dependent Manner in Caenorhabditis elegans*

Background: The heat shock response (HSR) maintains proteostasis and may be under metabolic control. Results: Caloric restriction (CR) enhances induction of hsp70 genes in Caenorhabditis elegans and prevents toxicity in an aggregation model, requiring sir2.1. Conclusion: CR synergizes with heat to activate the HSR. Significance: The metabolic state of an organism modulates the HSR. The heat shock response (HSR) is responsible for maintaining cellular and organismal health through the regulation of proteostasis. Recent data demonstrating that the mammalian HSR is regulated by SIRT1 suggest that this response may be under metabolic control. To test this hypothesis, we have determined the effect of caloric restriction in Caenorhabditis elegans on activation of the HSR and have found a synergistic effect on the induction of hsp70 gene expression. The homolog of mammalian SIRT1 in C. elegans is Sir2.1. Using a mutated C. elegans strain with a sir2.1 deletion, we show that heat shock and caloric restriction cooperate to promote increased survivability and fitness in a sir2.1-dependent manner. Finally, we show that caloric restriction increases the ability of heat shock to preserve movement in a polyglutamine toxicity neurodegenerative disease model and that this effect is dependent on sir2.1.

The heat shock response (HSR) 2 is a cytoprotective response that increases longevity and protects against diseases of aging in model organisms (1,2). This response enables an organism to manage protein-damaging stress through activation of the transcription factor HSF-1. HSF-1 is responsible for transcriptionally up-regulating the heat shock protein (HSP) genes. In mammalian cells, flies, nematodes, and plants, this transcription factor exists as an inactive monomer; however, in response to denaturing stress, it will trimerize and bind to heat shock elements located in the promoter regions of the hsp genes (3)(4)(5)(6). Increased expression of HSPs, such as HSP70 family members, results in cellular protection from a variety of stressors, including elevated temperatures, oxidative stress, heavy metals, proteasome inhibitors, and infection (7).
A characteristic of many HSR activators is the ability to elicit hormesis, a biological process that occurs when a low level stress is applied that promotes cytoprotection against a subsequent exposure to a more severe stress (8). For instance, exposure to a moderate heat shock (HS) can protect against exposure to a subsequent high temperature HS in Caenorhabditis elegans (9). Hormesis occurs, in part, through the up-regulation of molecular chaperones during the first mild stress treatment, which then protect cells from subsequent acute protein damage.
Caloric restriction (CR), a 30 -40% decrease in dietary intake, increases longevity and protects against diseases of aging (10). An association between CR and longevity was established as early as 1935 through studies with calorically restricted rats (11). CR has now been shown to increase longevity in many other models, including Saccharomyces cerevisiae (12,13), C. elegans (14,15), Drosophila melanogaster (16,17), and Mus musculus (18). In addition to increasing longevity, CR slows the progression of many age-related diseases, including neurodegenerative diseases (10). For instance, in various C. elegans models of protein aggregation diseases, CR has been shown to protect against age-associated paralysis (19). Genetic experiments in model organisms have implied that CR mediates its effects through a number of genes, including the sirtuins (20).
The sirtuins are a family of NAD ϩ -dependent deacetylases that have been characterized to play a role in a number of biological processes. The first family member identified was the yeast Sir2␣, based on its responsibility for establishing transcriptional silencing of mating-type loci (21). The mammalian homolog to yeast Sir2␣, SIRT1, deacetylates histones and many stress-inducible transcription factors, including p53, FOXO, and NF-B (22)(23)(24). Recent findings have indicated that SIRT1 also plays a critical role in the HSR by directly deacetylating HSF-1 within its DNA-binding domain to promote HSF-1 occupancy at hsp promoters (25). As SIRT1 is an important metabolic regulator and because HSF-1 and SIRT1 function together to protect cells from stress, we were interested in examining a direct link between the HSR and metabolism.
C. elegans is a useful model organism for testing the relationship between the HSR and metabolism, as these animals can easily be calorically restricted via bacterial limitation. Here, we show that CR and HS synergize to induce the HSR and that this effect depends on the C. elegans SIRT1 homolog Sir2.1.
Bleach Synchronization-Animals were bleach-synchronized to obtain the same developmental time point prior to any experimental conditions as described previously (30). Briefly, a Petri dish of C. elegans was washed with M9 buffer to dislodge the animals for transfer into a 15-ml conical tube for centrifugation. The pellet was washed with 20% alkaline hypochlorite solution to kill all animals except for the eggs, which are resistant to the bleach. These were then centrifuged and washed three times with M9 buffer, and the final pellet was resuspended in 7 ml of M9 buffer and incubated at room temperature at 220 rpm to allow the eggs to reach the L1 larva stage. Synchronized L1 larvae were used for all experimental conditions. CR and HS Conditions-Synchronized L1 larvae were cultured in S-basal medium with E. coli OP50 under either ad libitum or CR conditions, 1.9 ϫ 10 10 or 2.6 ϫ 10 9 bacterial cells/ml, respectively (31). Cultures were incubated at 23°C at 220 rpm. The life cycle progression of the animals was monitored daily under a dissecting microscope until they reached adulthood. Once the animals reached adulthood, each ad libitum and CR flask was equally divided into two flasks (one for the control and one for HS). HS was performed for 15 min or 1 h in a 33°C circulating water bath, and worms were allowed to recover at 23°C for 30 min or 6 h as indicated.
Quantitative RT-PCR-qRT-PCR was performed to quantify mRNA levels for the hsp70 family members C12C8.1, F44E5.4, and F44E5.5 as well as sir2.1 using gene-specific primers. Animals were collected via centrifugation, and RNA was extracted using TRIzol using the standard protocol. RNA was reversetranscribed using a high capacity cDNA reverse transcription kit (catalog number 4368814, Applied Biosystems) according to the manufacturer's protocol. The samples were diluted to 50 ng/l and used as a template for qRT-PCR. qRT-PCR was per-formed with a StepOnePlus real-time PCR system (Applied Biosystems) using iTaq TM Fast SYBR Green Supermix with ROX (catalog number 172-5101, Bio-Rad) according to the manufacturer's protocol. Statistical data analysis and determination of relative -fold increase from control samples were performed according to standard calculations (32).
Fluorescence Microscopy-Animals were analyzed for GFP expression using an EVOS fluorescence microscope. Animals were photographed individually by pipetting onto a layer of 1% agarose on top of a glass slide. Animals were paralyzed prior to microscopy by adding 1-2 drops of 1 mM levamisole onto the slide.
Immunoblot Analysis-For each treatment condition, cultured C. elegans animals were centrifuged at 5000 rpm for 5 min, and the resulting pellet was repeatedly washed with phosphate-buffered saline to clear bacteria from the extraction. Animals were lysed in 20 mM HEPES (pH 7.9), 25% (v/v) glycerol, 0.42 M NaCl, 1.5 mM MgCl 2 , and 0.2 mM EDTA with the addition of Halt TM protease inhibitors (catalog number 78430, Thermo Scientific) and sonicated for 15 min, cycling on and off for 30-s intervals, using a Diagenode Bioruptor 300 device. Protein extracts were quantified and run on a 10% SDS-polyacrylamide gel. Blots were probed with anti-GFP antibody (catalog number sc-9996, Santa Cruz Biotechnology) at a 1:1000 dilution and anti-actin antibody (catalog number JLA20-C, Amersham Biosciences) at a 1:750 dilution.
Thermotolerance Assay-Animals grown under ad libitum or CR conditions were either left untreated or heat-shocked for 1 h at 33°C in liquid culture in a circulating water bath and allowed to recover overnight at 23°C. The next day, 100 animals from each of the four liquid cultures were plated to separate nematode growth medium agar plates without bacteria. These were then exposed to a lethal HS at 36°C in a circulating water bath for 45 min with a 3-h recovery. Survivability of the animals was then manually quantified using a dissecting microscope. Animals were scored as alive if they responded to stimuli by poking with a platinum wire.
Thrashing Assay-After animals were scored in the thermotolerance assay, lethal HS survivors were individually transferred to a drop of M9 buffer on a glass slide. The animals were allowed to recover from the transfer for 1 min. Thrashing was judged as a mid-body bend. The number of thrashes was counted for 1 min after recovery from the transfer.
Paralysis Assay-Q24 animals grown under ad libitum or CR conditions were either left untreated or heat-shocked for 1 h at 33°C in a circulating water bath and allowed to recover overnight at 23°C. At day 4, paralysis was determined by transferring individual animals to a seeded nematode growth medium plate. Animals that did not move through the bacterial lawn and create a trail within 5 min were scored as paralyzed. Animals were scored as alive if they responded to stimuli by poking with a platinum wire.
Polyglutamine Protein Aggregation Assay-Q35-YFP animals were cultured as described for the paralysis assay. At day 3, animals were visualized to analyze GFP expression. Animals were photographed as described above, and the number of protein aggregates was scored.

RESULTS
CR and HS Synergize to Activate the hsp70 Promoter Reporter-CR has previously been reported to induce hsp70 expression (33,34). To verify that our CR conditions induce a HSR in C. elegans, we visualized GFP expression, which is under the regulatory control of the hsp70 promoter (C12C8.1::GFP) by microscopy (26). Animals were maintained under ad libitum or CR conditions with either 1.9 ϫ 10 10 or 2.6 ϫ 10 9 bacterial cells/ml, respectively. As expected, HS-treated animals showed induction of HS-responsive GFP expression compared with untreated animals (Fig. 1A). CR also resulted in an increase in expression of the HS-responsive reporter; however, GFP levels were not as robust as in animals receiving HS treatment. Interestingly, animals subjected to both CR and HS displayed an increase in reporter activation compared with either treatment condition alone. The increase in GFP expression observed by fluorescence microscopy was corroborated by quantification of GFP levels (Fig. 1B) and by Western analysis using anti-GFP antibody (Fig. 1C). Although immunoblotting was not sensitive enough to detect GFP expression in the animals treated with CR alone, quantitation of GFP fluorescence confirmed that GFP expression in animals subjected to CR and HS was increased by 2-fold compared with that in animals treated with HS alone.
CR and HS Synergize to Activate Endogenous HS Gene Expression-To obtain a more accurate quantification of the synergy between CR and HS, endogenous HS gene expression was measured in the C. elegans N2 strain. Wild-type N2 animals cultured under either ad libitum or CR conditions were exposed to a 33°C HS for 15 min with a 30-min recovery. Induction of C12C8.1(hsp70) transcripts was quantitated by qRT-PCR (Fig. 2). As expected, animals maintained under ad libitum conditions showed a 40-fold increase in C12C8.1 mRNA levels upon HS ( Fig. 2A). Consistent with results shown in Fig. 1A, N2 animals cultured under CR conditions also showed a 20-fold increase in C12C8.1 expression. Strikingly, animals cultured under CR and HS conditions together exhibited a 4000-fold increase in C12C8.1 mRNA levels, indicating a synergistic effect between HS and CR. To determine whether similar results were observed for multiple members of the hsp70 family, we also performed qRT-PCR using primers against F44E5.4 and F44E5.5. Similar to C12C8.1, F44E5.4, and F44E5.5 expression significantly increased in response to CR and HS treatment compared with either treatment alone (Fig. 2,  B and C).
sir2.1 Is Required for the Synergy between HS and CR upon hsp70 Induction-We have previously established that mammalian cells require the NAD ϩ -dependent deacetylase SIRT1 for the full transcriptional activity of HSF-1 (25). As CR stimulates SIRT1 activity by changing the NAD ϩ /NADH ratio, we tested whether the synergistic effect observed between HS and CR was dependent on sir2.1. The sir2.1(ok434) strain possesses a homozygous sir2.1 partial gene deletion that encodes a nonfunctional protein.
To initiate experiments, animals were treated with HS, CR, or both as described previously for Fig. 2, and qRT-PCR was performed to evaluate mRNA expression for the hsp70 family members C12C8.1, F44E5.4, and F44E5.5 (Fig.  3). For reasons that are not clear, the hsp70 genes are induced to a higher level upon HS treatment in the sir2.1(ok434) strain than in the N2 strain under ad libitum conditions. Interestingly, the sir2.1(ok434) strain was no longer able to up-regulate C12C8.1 in response to CR (Fig. 3A). Furthermore, the ability of CR and HS treatment to synergistically up-regulate C12C8.1, F44E5.4, and F44E5.5 gene expression was lost upon deletion of sir2.1 (Fig. 3, A-C).
To determine whether the synergistic up-regulation of the hsp70 gene family was dependent on sir2.1 and not some other sirtuin family member, the sir2.3(ok444) knock-out strain was treated with CR, HS, or both, and C12C8.1, F44E5.4, and F44E5.5 expression was measured by qRT-PCR. As shown in Fig. 4, animals treated with both CR and HS showed a synergistic increase in the expression of the hsp70 gene family, indicat-ing the specificity of the response to sir2.1. Thus, expression of the sir2.1 gene product, but not sir2.3, is required to synergistically up-regulate hsp70 genes in response to HS and CR in C. elegans.
CR Enhances the Ability of HS to Induce Thermotolerance-As a prior induction of hsp genes upon mild stress can lead to increased thermotolerance upon subsequent exposure to a lethal HS in a variety of organisms (8,9), we investigated whether CR and HS could heighten thermotolerance in C. elegans above levels produced with HS alone. N2 animals were exposed to a 33°C conditioning HS with an overnight recovery prior to receiving a 36°C lethal HS. After a 3-h recovery, the animals were scored for survivability. As expected, preconditioning with a mild HS provided thermotolerance to lethal HS treatment with a 25% increase in survival rate compared with animals without HS preconditioning (Fig. 5A). CR  endogenous hsp70 genes C12C8.1, F44E5.4, and F44E5.5. A, N2 animals were treated with or without HS and CR as indicated, and mRNA levels were determined for C12C8.1 by qRT-PCR. B, N2 animals were treated as described for A, and mRNA levels were determined for F44E5.4. C, N2 animals were treated as described for A, and mRNA levels were determined for F44E5.5. Results are in technical triplicates and are representative of biological duplicates. Statistical significance was measured by Student's t test compared with ad libitum (AL): **, p Ͻ 0.01; ***, p Ͻ 0.001. sir2.1. A, sir2.1(ok434) deletion animals were treated with or without HS and CR as indicated, and mRNA levels were determined for C12C8.1 by qRT-PCR. B, sir2.1(ok434) animals were treated as described for A, and mRNA levels were determined for F44E5.4. C, sir2.1(ok434) animals were treated as described for A, and mRNA levels were determined for F44E5.5. Results are in technical triplicates and are representative of biological duplicates. Statistical significance was measured by Student's t test compared with ad libitum (AL): *, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001. preconditioning prior to lethal HS treatment also provided modest thermotolerance, with a 14% increase in survival. Similar to what we observed with hsp70 gene induction, animals preconditioned with both CR and HS displayed heightened thermotolerance. Animal survival increased from 48% with no preconditioning to 94% survival upon preconditioning with both CR and HS together. Both sir2.1(ok434) and sir2.3(ok444) deletion strains were examined for responses to thermotolerance. Consistent with our other data, sir2.1, but not sir2.3, was required for thermotolerance. Therefore, as for the induction of hsp70 genes, the ability of CR to enhance HS-induced thermotolerance requires sir2.1.

FIGURE 3. Synergistic effect between HS and CR on hsp70 gene induction is dependent upon
CR Enhances the Ability of HS Preconditioning to Increase the Fitness of Lethal HS Survivors-To measure whether HS and CR preconditioning increased overall fitness after lethal HS exposure, thrashing assays were performed immediately after scoring for thermotolerance by transporting individual animals into a drop of M9 buffer. After a 1-min recovery from the transfer, the number of thrashes (or mid-body bends) was scored for 1 min. Similar to survival assays, preconditioning with either HS or CR alone prior to exposure to lethal HS increased the number of thrashes/min (Fig. 5B). Moreover, animals preconditioned with both CR and HS displayed heightened fitness compared with either treatment alone, in a manner dependent on sir2.1 (Fig. 5, B-E). Collectively, our data indicate that CR works in conjunction with HS to effectively provide increased thermotolerance, survival, and fitness, a process specifically requiring sir2.1.
CR Increases the Ability of HS to Preserve Movement in a C. elegans Model of Polyglutamine Diseases-C. elegans strains genetically engineered to express a polyglutamine (polyQ) repeat tract fused to YFP in the body wall muscle serve as models of polyQ disease, where the expression of polyQ expansions in body wall muscle cells causes animal paralysis that develops in a polyQ length-and age-dependent manner (27). In polyQ models in a variety of organisms, induction of the HSR protects animals from polyQ aggregation and aggregation-induced toxicity (36 -38). We were interested in determining whether CR would enhance the ability of HS to protect from polyQ aggregation and cytotoxicity. Q35-YFP worms produce aggregates that can be easily visualized by fluorescence microscopy by day 3 of adulthood (39). Upon testing the effect of HS and CR on the Q35-YFP strain, we found a statistically significant decrease in the number of aggregates with either CR or HS pretreatment (Fig. 6, A and B). There was no increased effect observed upon treatment with HS and CR together, which could perhaps be due to the strong propensity of Q35-YFP to aggregate. To test whether CR could enhance the cytoprotective effect of HS on a more sensitive polyQ strain, we assayed polyQ-induced paralysis in Q24-YFP animals. Although this strain does not produce visual aggregates, it does result in age-dependent paralysis that is evident at day 4 of adulthood. CR was observed to increase the ability of HS to alleviate paralysis in this strain (Fig. 6C). A Q24-YFP;sir2.1(ok434) strain was created to determine whether the ability of CR to enhance HS-induced cytoprotection is dependent upon sir2.1. The Q24-YFP;sir2.1(ok434) double mutant displayed a higher degree of paralysis and eliminated the ability of CR to enhance HS-induced movement preservation (Fig. 6C). Therefore, we conclude that CR can increase the ability of HS to preserve movement ability in this polyQ model in a sir2.1-dependent manner. The transcriptional regulation of the HSR was further evaluated for the Q24-YFP mutant to determine whether the cytotoxicity of the polyQ repeat altered the HS and CR synergism observed for the hsp70 genes in N2 animals. Q24-YFP animals cultured under either ad libitum or CR conditions were exposed to a 33°C HS for 15 min with a 30-min recovery. Induction of hsp70 family transcripts was quantitated by qRT-PCR (Fig. 7, A-C). We found that the Q24 repeats did not impact the synergy between HS and CR, although the overall inducibility of the hsp70 genes was reduced. We tested to see if the expression of Q24-YFP influenced sir2.1 expression and found that it was decreased by 43% (Fig. 7D). Thus, we conclude that although Q24-YFP expression dampens sir2.1 mRNA expression and the inducibility of the HSR, the ability of CR to synergize with HS is maintained.

DISCUSSION
Prior work on mammalian HSF-1 has suggested that it may be regulated by metabolism due to control by the NAD ϩ -dependent sirtuin SIRT1 (25,40). SIRT1 activity promotes the HSR, so we therefore reasoned that CR, by changing the NAD ϩ /NADH ratio, would also promote the HSR. We have tested this hypothesis in C. elegans, an organism that is easily amenable to growth under CR conditions by limiting the bacterial food source. The studies presented here demonstrate that CR conditions do indeed enhance the HSR as observed in a C12C8.1(hsp70)::GFP reporter strain. Additionally, striking synergy occurs in the HS-induced levels of hsp70 mRNA in the presence of CR. This effect is dependent on sir2.1, as evidenced by experiments using sir2.1 deletion animals. The synergy that we observe between CR and HS is biologically relevant, as CR can enhance the ability of HS to cause thermotolerance and increase thrashing fitness after a lethal HS in a sir2.1-dependent manner. In addition, CR and HS can work together to preserve movement in a C. elegans polyQ cytotoxicity model. To ensure the survival of the organism, it makes sense that stress induced by CR would enhance the cytoprotective effects of HS-induced stress.
HSP70 is the chaperone that is highly induced by the HSR and protects cells from cytotoxic stress (41)(42)(43). There is an age-related decline in HSP70 levels during aging in many model organisms (44 -50), which correlates with a decline in the ability to respond to stress. Interestingly, CR has been shown to restore the ability of cells to mount a HSR in aged rat hepatocytes (51). In these cells, CR was found to function not by increasing HSF-1 levels, but by increasing the ability of HSF-1 to bind to DNA upon HS. It was proposed that HSF-1 may be post-translationally altered by CR to allow it to be more active. Our studies support a positive influence of CR upon induction of the HSR in a second model organism, C. elegans. The synergy that we observed between CR and HS requires sir2.1.
A direct interaction between the insulin/IGF-1-like signaling pathway and HSR in C. elegans has recently been described. Upon increased insulin/IGF-1-like signaling, an inhibitory complex containing the C. elegans proteins DDL-1, DDL-2, and HSB-1 was found to associate with HSF-1 and sequester it (3). By decreasing insulin/IGF-1-like signaling, CR conditions may thus enhance the HSR via two mechanisms: 1) an increased NAD ϩ /NADH ratio activates Sir2.1 and thus the HSR, and 2) the release of HSF-1 from the DDL-1⅐DDL-2⅐HSB-1 inhibitory complex leads to an increase in free HSF-1 capable of being activated. Homologs of the DDL-1, DDL-2, and HSB-1 proteins are also present in mammals, indicating a possible preservation of this pathway.
A characteristic of various activators of the HSR is that they can often function synergistically with each other. For example, the non-steroidal anti-inflammatory drug indomethacin can cooperate with a mild heat stress to induce increased levels of chaperone gene expression over treatment with just heat alone (52). Other small molecules shown to work synergistically with HS include the inflammatory pathway intermediate arachidonic acid, the hydroxylamine derivative bimoclomal, and the triterpenoid celastrol (35,53,54). We now add CR as another HSR activator that can synergize with HS.  FIGURE 7. polyQ cytotoxicity does not alter the synergistic effect between HS and CR on hsp70 gene induction. A, Q24-YFP animals were treated with or without HS and CR as indicated, and mRNA levels were determined for C12C8.1 by qRT-PCR. B, Q24-YFP animals were treated as described for A, and mRNA levels were determined for F44E5.4. C, Q24-YFP animals were treated as described for A, and mRNA levels were determined for F44E5.5. D, N2 and Q24-YFP animals were analyzed for sir2.1 mRNA levels by qRT-PCR. Results are in technical triplicates and are representative of biological duplicates. Statistical significance was measured by Student's t test compared with ad libitum (AL): *, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001.