Co-chaperone p23 Regulates C. elegans Lifespan in Response to Temperature

Temperature potently modulates various physiologic processes including organismal motility, growth rate, reproduction, and ageing. In ectotherms, longevity varies inversely with temperature, with animals living shorter at higher temperatures. Thermal effects on lifespan and other processes are ascribed to passive changes in metabolic rate, but recent evidence also suggests a regulated process. Here, we demonstrate that in response to temperature, daf-41/ZC395.10, the C. elegans homolog of p23 co-chaperone/prostaglandin E synthase-3, governs entry into the long-lived dauer diapause and regulates adult lifespan. daf-41 deletion triggers constitutive entry into the dauer diapause at elevated temperature dependent on neurosensory machinery (daf-10/IFT122), insulin/IGF-1 signaling (daf-16/FOXO), and steroidal signaling (daf-12/FXR). Surprisingly, daf-41 mutation alters the longevity response to temperature, living longer than wild-type at 25°C but shorter than wild-type at 15°C. Longevity phenotypes at 25°C work through daf-16/FOXO and heat shock factor hsf-1, while short lived phenotypes converge on daf-16/FOXO and depend on the daf-12/FXR steroid receptor. Correlatively daf-41 affected expression of DAF-16 and HSF-1 target genes at high temperature, and nuclear extracts from daf-41 animals showed increased occupancy of the heat shock response element. Our studies suggest that daf-41/p23 modulates key transcriptional changes in longevity pathways in response to temperature.


Introduction
Temperature dramatically impacts the lifespan of ectotherms, with lower temperatures typically extending and higher temperatures shortening life [1][2][3]. The conventional view is that temperature passively affects the rate of chemical reactions and metabolism, thereby influencing species longevity. An emerging body of evidence, however, indicates that changes in longevity in response to temperature also reflect a regulated process entailing important organismal adaptations [4][5][6].
Like other ectotherms, the nematode Caenorhabditis elegans shows clear temperature dependent influences on development and lifespan [2]. At the normal cultivation temperature (20°C), animals typically live three weeks. At low temperature (15°C) they live approximately ten days longer, and at high temperature (25°C) ten days shorter. A handful of identified loci have been recently shown to mediate changes in lifespan in response to temperature. At warm temperatures, signaling from thermotaxis neurons promotes normal lifespan, and activates steroid hormone signaling to maintain longevity [4]. Animals compromised in thermotaxis genes or steroid hormone production display shortened life. In addition, animals will mount a heat shock response when exposed to acute heat stress, which helps preserve organismal viability. Thermotaxis neurons regulate the organismal heat shock response through the heat shock transcription factor HSF-1 [7]. Overexpression of hsf-1 and downstream chaperones can extend lifespan [8,9]. In contrast, longevity at cool temperatures requires the cold sensitive TRPA-1 channel [5], which works through Ca 2+ signaling, protein kinase-C PKC-2, and serum and glucocorticoid kinase SGK-1 to activate the forkhead transcription factor DAF-16/FOXO, a crucial regulator of longevity. Evidently the mechanisms governing longevity at low or high temperature appear somewhat different and few components affect both [4].
Temperature effects on organismal longevity are not limited to ectotherms. Notably, temperature sensing neurons involved in homeostatic control of core body temperature affect murine lifespan [6]. Higher temperatures in these neurons trigger a lowering of the core body temperature and correlate with extended lifespan. Several long-lived mouse models including the Ames Dwarf, Growth Hormone Receptor knockout, and FGF21 transgenic mice have associated a lower core body temperature reminiscent of nutrient induced torpor [10,11]. Furthermore cold exposure extends lifespan and suppresses tumorigenesis in rats [12]. These studies suggest an intimate but relatively unexplored relationship between nutrient and thermal sensing, metabolism and longevity.
At the cellular level, chaperones and co-chaperones facilitate protein folding and assembly, often in response to thermal stress [13,14]. One such co-chaperone is p23 [15]. p23 complexes with the HSP90 chaperone and inhibits its ATPase activity [16][17][18], thereby stabilizing association with client proteins such as steroid receptor transcription factors, heat shock factor, and others [19][20][21][22][23][24][25]. These interactions play an important role in regulating transcriptional events. p23 also displays HSP90 independent chaperone-like activity [26], and is implicated in various cellular functions [27]. Additionally the protein reportedly harbors prostaglandin E2 synthase (PGS) activity in vitro [28], although this is debated [29]. Interestingly, p23 upregulation has also been implicated in tumorigenesis presumably through its interactions with HSF1, steroid receptors or growth regulated kinases [30]. Indeed, several anti-cancer drugs being developed target chaperones and co-chaperones, highlighting the clinical importance of these pathways. p23 knockout mice reveal an early peri-natal lethal phenotype, with defects in lung and skin development, but its organismal roles remain largely unknown [29,31].
Here, we used C. elegans to explore the role of p23 function in metazoan biology. We found that daf-41, the C. elegans homolog of co-chaperone p23, has a novel role in regulation of lifespan at both high and low temperatures, as well as in the formation of the long lived dauer stage. Remarkably, daf-41 mutants provoked longevity at warm temperatures, but short lived phenotypes at cold temperatures, thus equalizing the temperature response. daf-41 interacted with insulin signaling, heat shock factor, and steroidal signaling to regulate lifespan by distinct mechanisms at different temperatures. Our findings implicate daf-41 as a central player in the thermal regulation of longevity.
Under conditions of food scarcity, overcrowding and elevated temperature, C. elegans larvae enter the dauer diapause, a stage specialized for survival and dispersal. We found that both daf-41(ok3052) and daf-41(ok3015) alleles were constitutively prone to form dauer larvae (Daf-c phenotype) at elevated temperature, yielding approximately 25% dauer larvae at 25°C, and nearly 100% dauer larvae at 27°C (Fig. 1C-D). By contrast, mutants containing deletions of other putative prostaglandin synthase homologs, pges-2/mPGES2, gst-4/PGDS, had little observable dauer phenotype at these temperatures. By inference, the co-chaperone function of daf-41 may be more important for Daf-c phenotypes.
Many Daf-c loci in the dauer signaling pathways, such as the daf-2/Insulin receptor (InsR) mutant, provoke resistance to various forms of stress [32,33]. Similarly daf-41 mutants displayed resistance to oxidative stress induced by H 2 O 2 challenge comparable in strength to daf-2/InsR mutants (Fig. 1E). daf-41 mutants exposed to heat stress at 35°C were also significantly resistant (Fig. 1E). Other PGS mutants had little effect on oxidative and heat resistance. Altogether these results demonstrate that daf-41 is a novel Daf-c gene involved in stress tolerance.
We found that mutations in the chemotaxis genes, daf-10, osm-1 and osm-3, significantly suppressed Daf-c phenotypes of daf-41(ok3052) at 25°C and partially at 27°C (Fig. 2B). (Precedence for weaker Daf-c mutants suppressing stronger Daf-c mutants has been seen previously [45]). Consistent with a role proximal to the chemotaxis machinery, daf-41(ok3052) worms were chemotaxis defective for isoamylalcohol, benzaldehyde and 2,4,5-trimethylthiazoline (S2A Fig Neurosensory cilia normally contact the environment through sensilla in head and tail, and typically fill with the lipophilic dye DiI. Mutants with defective neurosensory cilia structure, including daf-10 and osm-3 fail to fill with DiI because dendritic endings are not exposed [46]. We found that daf-41 null mutants had normal DiI filling similar to wild type worms (S2B Fig). Altogether these results show that daf-41 affects chemotaxis function, but not sensory cilia structure.

Interactions with HSP90 and cGMP signaling
In its capacity as co-chaperone, p23 is known to complex with HSP90 [16][17][18]47]. The C. elegans homolog of HSP90, daf-21, functions early in the dauer signaling pathways at the level of chemosensory processing, similar to daf-41. daf-21(p673) is a weak gain-of-function (gf) amino acid substitution that renders animals Daf-c [48]. Genetic epistasis studies show that Daf-c phenotypes are suppressed by daf-10 and other chemosensory mutants [40]. cGMP signaling represents another branch of the dauer signaling pathways that works at a similar level as daf- 21. Mutations in the daf-11/transmembrane guanylyl cyclase provoke Daf-c phenotypes, which are similarly suppressed by chemosensory mutants [40,48]. Finally both daf-21 and daf-11 exhibit chemosensory deficits [49].

daf-41 regulates longevity in response to temperature
Because daf-41 mutants showed clear temperature dependent dauer and stress resistance phenotypes, we wondered whether daf-41 would influence ageing at various temperatures. Increasing temperature is well known to reduce longevity in ectotherms, including wild type C. elegans ( Fig. 3A-C, Table 1). daf-41 mutants exhibited an altered temperature dependent longevity, revealing a leveling out of the lifespan curves to those typically seen at 20°C: animals were long lived at 25°C, normal lived at 20°C, and short lived at 15°C relative to wild type ( Fig. 3A-E). Other PGS mutants showed normal temperature dependent lifespan phenotypes ( Fig. 3A-C, Table 1).
To determine if the observed longevity phenotypes were due to lesions in daf-41, we performed rescue experiments with the daf-41(+) transgene. As expected, daf-41(ok3052) mutant animals harboring daf-41(+) were readily rescued for Daf-c phenotypes, while the pges-2(+) control transgene expressed under daf-41 regulatory elements did not rescue (Fig. 3F). We next tested the influence of the transgenes on ageing. First, we found that overexpression of daf-41 and pges-2 had no effect on ageing in both N2 and daf-41(ok3052) at 20°C ( Fig. 3G and S5A  Fig). However, the daf-41 transgene clearly reversed the daf-41(ok3052) longevity phenotype at 25°C, and partially rescued the short-lived phenotype at 15°C ( Fig. 3G and S5B-C Fig). In sum, these results reveal that daf-41(+) regulates longevity in response to temperature. Consistent with a role in thermal regulated processes, daf-41 mRNA increased modestly with temperature ( Fig. 3H).
Because metabolic rates increase with temperature and vary inversely with longevity, we wondered whether daf-41 mutants altered mitochondrial metabolism. When we measured O 2 consumption of daf-41 mutants at different temperatures, however, we saw no significant differences from wild type (S5D Fig). We also examined reproductive potential of daf-41(ok3052)
Given the genetic interactions with hsf-1 for dauer formation, longevity, and gene transcription described above, we wondered whether daf-41 affects assembly of HSF-1 nuclear complexes. To measure this, we prepared nuclear extracts from wild type and daf-41 mutants and performed electrophoretic mobility shift assays on the heat shock factor response element as established previously [58]. First we found that daf-41 mutation had no effect on mRNA or protein levels of HSF-1 (S8C Fig). Second, we observed that daf-41 nuclear extracts showed a clear and reproducible 1.5-2 fold higher occupancy of the heat shock factor response element, compared to WT controls (Fig. 5F). These results suggest that DAF-41(+) normally affects the assembly or disassembly of HSF-1 transcriptional complexes.
Because p23 and HSP90 are known to work together, we asked whether they would have similar effects on lifespan. At 25°C, both daf-41(ok3052) and daf-21(p673)gf worms were longer lived than wild type; however, the daf-41;daf-21 double mutant had longevity phenotypes more similar to the daf-21 single mutant (Fig. 5G, S9A-B Fig, Table 2). The convergent behavior at 25°C suggests they could work together at this temperature, similar to dauer. By contrast at 15°C, daf-21(p673) worms were extremely long-lived; this longevity was additive to the shortlived phenotype of daf-41(ok3052) mutant animals. The distinct behavior of daf-41/p23 and daf-21/hsp90 at 15°C, suggests they may regulate lifespan at this temperature through different pathways (S9C Fig).   The steroid receptor DAF-12 also influenced the longevity phenotypes of daf-41(ok3052) (Fig. 6A-D, Table 2). At 25°C, daf-12 mutation partially reduced the longevity of daf-41 mutants in an additive manner, but was not epistatic, i.e. daf-12 daf-41 did not give the same life span as daf-12 itself. The expression levels of the DA biosynthetic gene, daf-36/Rieske, and two DAF-12 target genes, cdr-6 and fard-1, were also increased at higher temperatures. Therefore, daf-12(+) could work in parallel, or function as a minor branch of daf-41 signaling to promote longevity. More interestingly at 15°C, daf-12 mutation suppressed the short lived phenotypes of daf-41, restoring near normal life span. At this temperature, cdr-6 and fard-1 genes decreased in expression in daf-41 mutants relative to WT (Fig. 6E). These interactions suggest that daf-41(+) could prevent life shortening properties of DAF-12 steroidal signal transduction at lower temperatures.

Discussion
How temperature influences animal lifespan is not well understood. In this work, we demonstrate that daf-41/p23 co-chaperone PTGES3 homolog modulates longevity in response to temperature, and regulates entry and exit from the long-lived dauer stage. On the one hand, daf-41 (+) promotes normal adult longevity at cold temperature (15°C). On the other hand, daf-41(+) limits lifespan at warm temperature (25°C), but has little influence on longevity at 20°C. Thus the overall effect seen in daf-41 mutants is an equalization of the lifespan curve so that animals appear as if they are living at 20°C. Surprisingly, no difference in respiratory rates (O 2 consumption) between wild type and daf-41 animals was detected at these temperatures, suggesting that daf-41's effect on longevity is likely through mechanisms independent of mitochondrial metabolism. These intriguing observations reveal that daf-41 plays a unifying role in regulatory mechanisms that modulate lifespan in reaction to temperature, and imply that co-chaperone/chaperone complexes may mediate this response.
Although the detailed molecular mechanism by which daf-41 mediates these temperature dependent effects is not entirely understood, our genetic epistasis experiments suggest that daf-41 may directly or indirectly impinge on several transcriptional longevity regulatory mechanisms, including DAF-16/FOXO, HSF-1, and DAF-12 steroidal signaling, to regulate lifespan in response to temperature. Several lines of evidence argue that daf-16/FOXO is a critical mediator in these circuits. At warm temperature, lifespan extension of daf-41 mutants was abolished in the daf-16 mutant background, at low temperatures the daf-41;daf-16 double mutants did not live shorter than daf-41 alone, suggesting convergent mechanisms. Overall daf-16 mutation restored temperature dependent lifespan regulation to daf-41 mutants. As further support for a link to FOXO function, daf-16 partly suppressed daf-41 Daf-c phenotypes, placing daf-16 downstream for both dauer formation and longevity. Consistent with working in a signaling pathway, we found that daf-41(+) negatively regulated transcription of DAF-16 targets (sod-3, dod-3, lipl-4) in response to warm temperature although no clear effect was seen at low temperatures with these genes. Several insulin-like peptides were up or down regulated in daf-41 mutants in response to warm temperature, suggesting that daf-41 has the potential to affect IIS. However, we could not detect obvious differences in DAF-16 nuclear localization, as is often seen with mutations that only weakly affect signaling (Henderson et al. 2001). Conceivably daf-41 could regulate components of IIS or DAF-16 complexes themselves.
Heat shock factor, hsf-1, often works in tandem with daf-16/FOXO [8,9] and consistently, was also required for daf-41 induced longevity at elevated temperatures. Accordingly, several hsf-1 target genes (hsp-16.2, hsp-70, hsp-4) were expressed in daf-41 mutants under these conditions. Although HSF-1 mRNA or protein levels were unchanged in the daf-41 background, nuclear extracts from daf-41 mutants showed increased occupancy of the heat shock response element. Importantly, this result suggests that daf-41(+) influences either directly or indirectly the assembly/disassembly of the HSF-1 complex. This finding is consistent with observation the p23-HSP90 complex has been shown to inhibit the activity of the HSF-1 complex in other systems [23,25]. The observed short-lived phenotypes of daf-41 with hsf-1 at low temperatures were additive, possibly because the hsf-1 allele is temperature sensitive and non-null [56].
daf-41's interactions with nuclear receptor daf-12 suggest an intriguing role for steroid signaling. At 15°C, daf-12 mutation suppressed daf-41 short-lived phenotypes, suggesting that daf-41 modulates life extending effects of daf-12. Steroid signaling has previously been implicated in lifespan regulation albeit at high temperatures: daf-9 hypomorphs as well as thermotaxis mutants are particularly short-lived at 25°C, in a manner dependent upon daf-12 [4]. At these temperatures, thermotaxis loci as well as hsf-1(+) promote expression of the hormone biosynthetic gene daf-9/CYP27A1 [4,50]. These observations suggest that normal lifespan at 25°C depends on adequate stimulation of steroidal signaling, thereby preventing life shortening activities of the unliganded DAF-12. At 25°C, daf-12 mutation shortened the life span of daf-41 to a similar extent as wild type, suggesting parallel pathways. Alternately daf-12 could contribute partially toward daf-41 longevity, since daf-12 target genes as well as daf-36 increase expression at this temperature. At this point, it is unclear whether the activating or repressing functions of the receptor are responsible for longevity.
In mammals, p23 together with HSP90, immunophilins and other chaperones, binds to unliganded steroid receptors in the cytosol, maintaining the receptor in a primed state competent for rapid ligand binding; upon binding hormone the steroid receptor enters the nucleus where interacts with transcriptional coregulators. Thereafter, p23 helps disassemble transient transcriptional complexes to facilitate hormone sampling and repeated rounds of transcription [15,[20][21][22]26]. The p23-HSP90 complex also regulates type II nuclear receptors, such as the thyroid receptor [20], which constitutively reside in the nucleus, a situation more resembling that of DAF-12. Further studies on the role of steroidal signaling at low and high temperature may help clarify the whether similar mechanisms are at work in C. elegans.
Interactions with chemosensory mutants suggest a role within the sensory apparatus. Indeed, daf-41 longevity at warm temperatures converged with daf-10/IFT122, and both mutants stimulate daf-16, suggesting they could work through a similar chemosensory mechanism [54]. Consistent with a proximal role in chemosensory signaling, daf-41 mutants exhibited chemosensory deficits and chemosensory mutants suppressed the daf-41 Daf-c phenotypes. Accordingly, daf-41 was expressed in several chemosensory neurons including ASE, AWC, ADL, and ASI but was not obviously expressed in the main thermosensory neurons AFD and AIY, although AWC and ASI also reportedly contribute to thermosensation [61]. Ultimately, tissue specific dissection of daf-41 activities may help illuminate what cells mediate these interactions. It is also noteworthy that daf-41 did not further shorten the lifespan of the gcy triple mutant at 15°C, suggesting it could work through thermotaxis circuits at low temperature.
p23 is known to interact directly with HSP90 to modulate various client proteins [15,16], and recently, C. elegans p23 and HSP90 have been shown to physically interact in vitro [47]. Consistent with the possibility of working together, both daf-41/p23 and daf-21/HSP90 act at the level of chemosensory processing with respect to dauer formation (this work; [49]), and modulated one anothers' dauer and longevity phenotypes at 25°C. That daf-41 null and daf-21 gain of function have similar phenotypes could indicate that the wild type activities work in opposition for these processes. On the other hand, daf-41 and daf-21 had several divergent phenotypes: whereas daf-41 Daf-c phenotypes were enhanced by hsf-1 mutation (Fig. 2H, this work), daf-21 Daf-c phenotypes were suppressed [50]. Furthermore, whereas daf-41 mutants were short-lived at 15°C, daf-21 mutants were long-lived, and regulated life span differently. Thus some p23 phenotypes might arise from HSP90 dependent as well as independent processes, as has been noted previously [26,27]. We interpret these experiments with caution since daf-21 mutations are gain-of-function and non-null. Intriguingly, HSP90 in Candida albicans has been shown to regulate the switch to pseudohyphal growth in response to temperature [62], perhaps analogous to the thermal regulation of dauer formation or longevity.
Based on the observed genetic interactions we suggest the following model for longevity regulation. At elevated temperatures, p23 directly or indirectly inhibits the transcriptional activities of HSF-1 and perhaps DAF-16, thus limiting lifespan at these temperatures, while DAF-12/FXR and thermotaxis signaling work in parallel (Fig. 7C). Given the convergence with daf-10 longevity, we suggest that daf-41(+) might work through the chemosensory apparatus to impinge upon DAF-16. More speculatively, at lower temperatures, p23 stimulates DAF-16/FOXO and inhibits DAF-12/FXR, possibly through the thermotaxis apparatus, while HSF-1 works in parallel to promote lifespan extension.
Given the intriguing genetic interactions described here, it would be interesting to investigate whether DAF-41 and/or HSP90 binds and regulates the activities of HSF-1, DAF-16, or DAF-12 by protein-protein interaction. Alternately DAF-41 could interact with upstream components of these signaling pathways, including kinases, temperature sensitive channels, guanylyl cyclases, or cilia proteins. Conceivably, thermal regulation in these circuits could result from thermal influences on protein-protein interactions.
The results here and elsewhere reveal that regulation of longevity at different temperatures works by distinct mechanisms. This is perhaps not surprising, given that different stresses challenge the organism at low and high temperatures. Moreover, ectotherms must have evolved optima for growth and reproduction within a temperature range. With this in mind, we suggest that daf-41 could play distinct roles at low and high temperatures. Alternately daf-41 may be part of an adaptive response to temperature in which optima shifted towards higher temperatures have a consequent tradeoff at lower temperatures, and vice versa. Further elucidation of daf-41/p23 complexes and physiology in C. elegans should help illuminate the mechanism of thermal regulation of metazoan longevity.

Methods Strains
Strains were obtained from the Caenorhabditis Genetics Center (CGC, USA) and National Bio Resource Project (NBRP, Japan). All strains were outcrossed at least 4 times to wild type N2 before further analysis. All strains in this work are itemized in S1 Table. Genotypes of mutants were confirmed by PCR, sequencing and phenotyping. Primer sets for genotyping are listed in S2 Table. Ageing experiments All lifespans were measured as previously described [52]. Strains for ageing experiments were maintained at the respective cultivation temperature of 15°C and 20°C for more than 3 generations before analysis, unless indicated otherwise. Progeny were collected by egg laying and bleaching. Ageing experiments were performed with and without 2 0 fluoro-5 0 deoxyuridine (FUdR, Sigma) to prevent contamination of next generation progeny and to reduce bagging phenotypes of Egl mutants. Strains were treated with FUdR as described previously [63]. For ageing experiments at warm temperature, worms were cultivated at 20°C until the young adult stage, then moved to 25°C to start the ageing analysis in order to bypass dauer formation and minimize internal hatching phenotypes. Each experiment started with more than 150 worms. Sterile, escaped, internally hatched, and exploded worms were censored on the day of loss. Experiments were performed at least 3 times and the mean lifespan calculated. Worms were transferred onto new OP plates every 2-3 days from the end of reproductive period, and scored for survival every 2-4 days. Statistical analyses were performed by Kaplan-Meier method with GraphPad Prism software (GraphPad Software, Inc.)

Dauer formation assays
For dauer formation assays, all strains were maintained at 20°C and eggs collected by egg laying or bleaching. Greater than 50 eggs were transferred onto 3cm OP plates and cultured at 20°C, 22.5°C, 25°C and 27°C, respectively. Dauer formation fraction was typically scored at 60 hrs, and dauer exit ratio was scored at 84 hrs.

Stress resistance analysis
All strains were maintained at 20°C and Day 1 young adults were used for analysis. For heat stress analysis, 50-100 adult worms were transferred onto fresh 6 cm OP plates and shifted to 35°C. Fraction survival was scored 8 hrs after heat shock. For oxidative stress analysis, 50-100 adult worms were collected by washing off plates and transferred into 24 well plastic plates filled with 20mM of hydrogen peroxide (SIGMA) in M9 buffer. Experiments were performed with 5 biological replicates with 3 technical replicates for each mutant.

qRT-PCR analysis
Synchronized worms were prepared at different temperatures. Worms were collected in TRIzol (Invitrogen) at L4 stage and frozen in liquid nitrogen. Total RNA was extracted by RNeasy Mini kit (QIAGEN) and Superscript III First Strand Synthesis System (Invitrogen) was used for cDNA generation. qRT-PCR was performed with Power SYBR Green master mix (Applied Biosystems) on a 7900HT Fast Real-Time PCR System (Applied Biosystems). ama-1 was used as internal control for mRNA quontification. For each analysis, qRT-PCR was performed with at least three biological replicates. Primer sequences are listed in S2 Table. Preparation of nuclear protein extracts WT and daf-41(ok3052) worms were harvested and frozen down at day 1 of adulthood. Frozen worm pellets were first homogenized in an equal volume of 2X NPB buffer (20 mM HEPES, pH 7.6, 20 mM KCl, 3 mM MgCl 2 , 2mM EDTA, 0.5 M sucrose, 1 mM dithiothreitol, protease inhibitors, and phosphatase inhibitors) using a Kontes Pellet Pestle tissue grinder. The suspension was then centrifuged (4000 g, 5 min, 4°C) and the pellets were further homogenized by 20 strokes with pestle A of a Dounce homogenizer. The pellet was resuspended in NPB buffer with 0.25% NP-40 and 0.1% Triton-X100, centrifuged again, and washed three more times with the same buffer. The nuclear pellet was extracted with four volumes of HEG buffer (20 mM HEPES, pH 7.9, 0.5 mM EDTA, 10% glycerol, 0.42 M NaCl, 1.5 mM MgCl 2 , and protease inhibitors) at 4°C for 45 min. Finally, the nuclear fraction was collected by centrifugation at 14,000 g for 15 min at 4°C. Protein concentrations were determined with a Bradford assay kit (Bio-Rad, Hercules, CA).
Additional details for DiI staining, chemotaxis analysis, oxygen consumption measurement and microscopy analysis are described in S1 Text.  (XLSX) S1 Text. Supplemental materials and methods. Additional details for DiI staining, chemotaxis analysis, oxygen consumption measurement and microscopy analysis are described in S1 Text. (DOCX)