Abnormal Osmotic Avoidance Behavior in C. elegans Is Associated with Increased Hypertonic Stress Resistance and Improved Proteostasis

Protein function is controlled by the cellular proteostasis network. Proteostasis is energetically costly and those costs must be balanced with the energy needs of other physiological functions. Hypertonic stress causes widespread protein damage in C. elegans. Suppression and management of protein damage is essential for optimal survival under hypertonic conditions. ASH chemosensory neurons allow C. elegans to detect and avoid strongly hypertonic environments. We demonstrate that mutations in osm-9 and osm-12 that disrupt ASH mediated hypertonic avoidance behavior or genetic ablation of ASH neurons are associated with enhanced survival during hypertonic stress. Improved survival is not due to altered systemic volume homeostasis or organic osmolyte accumulation. Instead, we find that osm-9(ok1677) mutant and osm-9(RNAi) worms exhibit reductions in hypertonicity induced protein damage in non-neuronal cells suggesting that enhanced proteostasis capacity may account for improved hypertonic stress resistance in worms with defects in osmotic avoidance behavior. RNA-seq analysis revealed that genes that play roles in managing protein damage are upregulated in osm-9(ok1677) worms. Our findings are consistent with a growing body of work demonstrating that intercellular communication between neuronal and non-neuronal cells plays a critical role in integrating cellular stress resistance with other organismal physiological demands and associated energy costs.


Introduction
Cellular life requires precise control of protein structure and function, which are determined by protein conformation, concentration, assembly and localization. The homeostatic mechanisms that maintain protein function are collectively termed proteostasis. The proteostasis network is evolutionarily conserved and comprises the tightly integrated and regulated activities of gene transcription, RNA metabolism and protein synthesis, folding, assembly, trafficking, disassembly, repair and degradation [1][2][3].
Proteostasis is under constant challenge. Protein structure is inherently unstable [4,5] and readily disrupted by gene mutations and numerous environmental stressors. Random errors in cellular processes such as DNA replication, transcription and protein translation also disrupt protein structure. A poorly understood decline in the capacity of cellular proteostasis networks that repair and degrade damaged proteins is thought to underlie pathophysiology associated with senescence [6,7].
Proteostasis is energetically costly and those costs must be balanced with the energy needs of other cellular and organismal functions as well as with the requirement of the organism to respond to environmental change. A fundamental and emerging question is how do organisms ensure optimal survival by partitioning a finite energy budget to meet the demands of proteostasis and other essential physiological processes including reproduction? Several recent studies in C. elegans have pointed to a critical role for the nervous system and communication between different tissue types in coordinating organismal proteostasis needs. For example, Morimoto and co-workers have demonstrated that the heat shock response of non-neuronal cells is modulated by inhibitory inputs from AFD thermosensory neurons as well as neural circuits that respond to metabolic status and nutrient availability [8,9]. Activation of the endoplasmic reticulum (ER) unfolded protein response (UPR) in neurons activates the UPR in the intestine via a neurosecretory process that in turn increases organismal stress resistance and longevity [10]. Inhibition of protein translation or degradation in non-neuronal cells alters C. elegans behavioral responses that are controlled by chemosensory neurons [11].
The proteostasis network plays a critical role in ensuring optimal survival of C. elegans under dehydrating conditions. When exposed to hypertonic stress, C. elegans rapidly loses water and becomes paralyzed. Water loss is followed by systemic volume recovery and accumulation of the organic osmolyte glycerol [12]. Dehydration causes rapid and widespread protein aggregation and misfolding [13,14]. Stress induced protein damage is minimized by genes that function in protein degradation [13] and by reductions in protein synthesis [15,16]. Hypertonicity induced inhibition of translation also serves as a signal that activates glycerol accumulation pathways [17,18] and possibly other mechanisms that confer increased hypertonic stress resistance.
Given the importance of the proteostasis network to the survival of C. elegans in hypertonic environments, we characterized hypertonic stress resistance in worm strains with defects in osmotic avoidance behavior. C. elegans avoids strongly hypertonic solutions. This avoidance behavior is mediated by ASH chemosensory neurons [19,20]. We demonstrate that disruption of osmotic avoidance behavior via gene mutations or genetic ablation of ASH neurons is associated with enhanced survival in hypertonic environments. Enhanced survival is not due to altered systemic volume regulation or glycerol accumulation and instead may be due to enhanced proteostasis capacity.

Fluorescent protein aggregate measurement
The number of body wall muscle cell Q35::YFP aggregates were quantified manually in blinded experiments using a Zeiss Stemi SV11 microscope (Chester, VA). Single aggregate volume measurements and fluorescence recovery after photobleaching (FRAP) analysis were carried using confocal microscopy as described previously [14].
RNA interference. RNA interference (RNAi) was performed by feeding worms from the L1 larval stage with bacteria expressing a nonspecific scrambled dsRNA or dsRNA specific to osm-9. Late stage L4 RNAi worms were transferred to control or high NaCl growth plates seeded with dsRNA expressing bacteria.

let-60(ga89) mutant phenotype assay
Temperature sensitive let-60(ga89) mutant worms were maintained at the permissive temperature of 16°C. Defective egg hatching or larval arrest phenotypes were quantified by transferring 1-day old gravid adults to 300 mM NaCl feeding plates and then removing them after 24 h. Eggs were scored for failure to hatch or develop past the L1 larval stage.

Survival assays
Synchronized late L4 worms were exposed to various stressors and survival was determined by prodding animals with a platinum wire. Worms were considered to be dead if they did not respond to repeated prodding. All survival studies were done blinded.

Lifespan analysis
Synchronized L4 larvae were transferred to 51mM NaCl agar plates containing 50 μg/ml 5-fluorodeoxyuracil (FUDR; Sigma-Aldrich, St. Louis, MO). Animals were transferred onto fresh plates containing FUDR every 2 days for the first week, then every 4-5 days thereafter. Survival was scored daily for touch provoked movement. 35 S-methionine labeling of total protein Incorporation of 35 S-methionine into total protein was used to assess rates of protein synthesis and degradation. Radiolabeling was carried out using methods similar to those described by others [24]. Briefly, synchronized L4 worms were fed 35 S-methionine labeled OP50 bacteria for 4 h, washed and incubated with unlabeled OP50 for 1 h to purge radioactive intestinal bacteria, and then washed thoroughly with NGM buffer. Washed worms were flash frozen in liquid nitrogen and stored at -80°C before extraction. Protein was extracted from thawed samples by trichloroacetic acid-ethanol protein precipitation. Total protein concentration and radioactivity incorporation were measured by BCA assay (Pierce Biotechnology) and liquid scintillation counting, respectively.

RNA-sequencing and gene expression analysis
N2 and osm-9(ok1677) worms were carefully staged for gene expression analysis. Staging was carried out using well established methods [25] and took into account the slight developmental delay we observed in osm-9(ok1677) worms. Briefly, bleached eggs were placed on clean NGM plates to synchronize L1 stage larvae. After washing, L1 larvae were transferred to NGM plates seeded with OP50 and grown at 20°C. Cultures were visually inspected frequently during the staging process. Late stage L4 larvae were defined by the appearance of a white crescent shape that surrounds the prospective vulva. To avoid gene expression changes that might be associated with the L4/adult molt, animals with white crescents were transferred to experimental conditions while they were still moving and feeding and had not yet entered the quiescent period that occurs prior to molting [26].
L4 stage N2 and osm-9(ok1677) worms were washed and transferred to control (51mM NaCl) and hypertonic stress (200mM NaCl) agar plates for 6 h. Two to four independent experiments were conducted for each condition. After the exposure period, plates were visually inspected. No evidence of molting or development of worms into gravid adults was detected. Worms were rinsed off plates, washed 4 times, pelleted, and snap frozen for storage at -80°C until processing. Total RNA was isolated using TRIzol (Invitrogen, Carlsbad, CA). RNA quality was assessed by spectrophotometry and by using an RNA 6000 Nano Kit and Bioanalyzer 2100 (Agilent, Santa Clara, CA).
RNA-sequencing was carried out by Ocean Ridge Biosciences (Palm Beach Gardens, FL). llumina paired-end and barcoded TrueSeq mRNA sequencing libraries were prepared from RNA samples with an RNA Integrity Numer (RIN) of 8.0, pooled and sequenced for 76 cycles on two Illumina HiSeq2500 flow cell lanes (Illumina, Inc., San Diego, CA). Following diagnostic analyses of the 532,462,712 sequence reads using FastQC (http://www.bioinformatics. babraham.ac.uk/projects/fastqc/), sequences were aligned to the WS235 version of the C. elegans genome assembly using Tophat version 1.4.1 [27] and mapped to Ensembl (version 71) [28] annotated genes using Cufflinks version 2.0.1 [29]. Differentially expressed genes were determined using read counts per gene and a P-value threshold of 0.05 with edgeR version 3.2.4 [30] for each pairwise comparison among sample groups. Functional annotation for genes was obtained using WormBase [31] and Ensembl BioMart [32]. Enrichment analysis of DAF-16 and SKN-1 identified by Tepper et al. (33) and Niu et al. [34] was conducted using Fisher's Exact Test in R 3.2.0. RNA-Seq data are available at the NCBI Gene Expression Omnibus (http://www. ncbi.nlm.nih.gov/geo/) under accession GSE73589 and at the NCBI Short Read Archive (http://www.ncbi.nlm.nih.gov/sra) under accession SRP064324.

Statistical analysis
All data are presented as means ± S.E. Statistical significance was determined using Student's two tailed t-test when two means were compared or ANOVA with Tukey post hoc test when comparing multiple means. P values of 0.05 were taken to indicate statistical significance.

Results
Mutant worms with defective osmotic avoidance behavior are termed Osm for "osmotic avoidance abnormal". Ten osm genes have been identified. Loss of osm-7, osm-8 and osm-11 function to activate gpdh-1 expression and induce constitutive glycerol accumulation [17,[35][36][37]. All other osm genes are required for the normal function or development of chemosensory neurons. osm-9 and osm-12 encode a TRPV cation channel [20] and a protein required for the biogenesis of sensory neuron cilia [38], respectively. Both genes are expressed in ASH osmosensory neurons as well as other chemosensory neurons [39].
To determine whether defects in osmotic avoidance behavior impact hypertonic stress resistance, we initially examined the ability of wild type N2 and the osm-9(ky10) mutant to survive a 24 h exposure to high NaCl growth plates. The ky10 allele is a widely studied loss-of-function mutation that disrupts a variety of chemosensory behaviors including avoidance of hypertonic environments [20]. As shown in Fig 1A, osm-9(ky10) mutants had significantly (P<0.05) greater survival at 400, 500 and 600 mM NaCl compared to wild type animals. Effect of osm-9 and osm-12 mutations and genetic ablation of ASH osmosensory neurons on survival during hypertonic stress. (A) Survival of wild type worms and osm-9(ky10) loss-of-function mutants exposed to different concentrations of NaCl for 24 h. Values are means ± S.E. of four independent and blinded experiments. *P<0.05 and **P<0.001 compared to wild type animals. (B) Survival of wild type worms and osm-9(ok1677) and osm-12(n1606) loss-of-function mutants exposed to different concentrations of NaCl for 24 h. Values are means ± S.E. of four independent experiments. *P<0.05 for both osm-9(ok1677) and osm-12(n1606) mutants compared to wild type animals. (C) Survival of wild type and ASH ablated (i.e., Is[sra-6p::mCasp1] transgenic strain) worms exposed to 500 or 600 mM NaCl for 24 h. Values are means ± S.E. of 6-12 independent experiments. *P<0.002 and **P<0.02 compared to wild type worms. (D) Survival of wild type, gpa-4(pk381) mutants and ASI ablated (i.e., oyIs84[gpa-4p::TU#813 + gcy-27p::TU#814 + gcy-27p:: GFP + unc-122p::dsRed] transgenic strain) worms exposed to 500 or 600 mM NaCl for 24 h. Values are means ± S.E. of 8-9 experiments. *<P<0.05 compared to wild type and **P<0.001 compared to ASI ablation.
Acclimation of C. elegans to relatively low levels of hypertonic stress increases survival under more extreme conditions. Increased survival is due to reduced water loss that results from organic solute accumulation and to enhanced proteostasis capacity [12][13][14][15]. We acclimated wild type, osm-9(ok1677) and osm-12(n1606) worms beginning at the L1 larval stage to 200 mM NaCl and then assessed survival in late L4 and young adult animals exposed to increasing NaCl levels. As expected, acclimation improved hypertonic stress resistance, but enhanced resistance was similar in wild type and mutant worms (data not shown).
Yoshida et al. [21] recently described a transgenic worm strain in which ASH neurons are ablated by expression of mouse caspase 1 driven by the promoter sra-6. sra-6 is expressed highly in ASH neurons and weakly in ASI sensory and PVQ interneurons [39]. The Is[sra-6p:: mCasp] transgenic has defects in avoidance of noxious chemicals [21] and also exhibits defects in avoidance of hypertonic environments similar to osm-9(ky10), osm-9(ok1677) and osm-12 (n1606) worms (Table 1).
We tested the effect of hypertonic stress on survival of wild type worms and the Is[sra-6p:: mCasp] transgenic strain. The transgenic strain showed significantly (P<0.02) increased survival when exposed to 500 or 600 mM NaCl for 24 h (Fig 1C).
osm-9, osm-12 and sra-6 are expressed in ASH as well as other neuron types [39]. While not definitive, the only cell types in which expression of these three genes has been shown to overlap are ASH and ASI chemosensory neurons [39]. It is thus conceivable that the enhanced survival of osm mutants and the ASH ablation strain is due to defects in ASH and/or ASI neuron function. To assess the role of ASI neurons in hypertonic stress resistance, we quantified survival of gpa-4(pk381) loss-of-function mutants and the PY7505 transgenic strain in 500 and 600 mM NaCl. gpa-4 encodes a G-protein alpha subunit and has been reported to be expressed selectively in ASI neurons [40]. In the PY7505 strain, the ASI neurons are ablated by transgenic caspase expression [41]. At 500 mM NaCl, ASI ablation worms showed significantly (P<0.05) reduced survival compared to either wild type N2 worms or gpa-4(pk381) mutants ( Fig 1D). No significant (P>0.2) differences in survival between strains were observed at 600 mM NaCl (Fig 1D). Taken together, data in Fig 1 suggest that loss of ASH neuron functions required for avoidance of hypertonic environments is associated with enhanced hypertonic stress resistance. Table 1. Osmotic avoidance behavior in wild type and ASH ablation worms and osm mutants.
When exposed to hypertonic conditions, C. elegans rapidly loses water and becomes paralyzed. However, over a period of a few hours, systemic volume and normal motility is recovered. As shown in Fig 2A, N2 worms and osm-9(ok1677) and osm-12(n1606) mutants exhibit similar degrees of shrinkage and rates of volume recovery when exposed to agar plates containing 200 mM NaCl. Increased transcription of gpdh-1 catalyzes constitutive accumulation of the organic osmolyte glycerol in worms harboring loss-of-function mutations in certain osm genes [17,[35][36][37]. Increased glycerol levels will reduce water loss, which in turn reduces cellular and molecular damage and increases survival under hypertonic conditions. To directly assess the role glycerol accumulation plays in the enhanced osmotolerance of osm-9(ok1677) and osm-12(n1606) mutants, we quantified gpdh-1 expression and whole animal glycerol levels under control conditions (51mM NaCl). As shown in Fig 2B, gpdh-1 mRNA levels were not significantly (P>0.7) different in N2 and osm-9(ok1677) mutant. In osm-12(n1606) mutants, gpdh-1 mRNA levels were significantly (P<0.002) reduced compared to wild type worms. Glycerol levels (Fig 2C) mirrored gpdh-1 expression and were not significantly (P>0.2) different in osm-9(ok1677) mutants. In osm-12(n1606) mutants, glycerol levels were reduced significantly (P<0.007) compared to wild type worms.
GPDH-1 catalyzes the rate limiting step in glycerol accumulation during hypertonic stress [17]. The similar rates of volume recovery between wild type, osm-9(ok1677) and osm-12 (n1606) animals shown in Fig 2A indicate that enhanced survival of osm mutants is not due to increased gpdh-1 expression and increased glycerol accumulation. To address this issue directly, we quantified temporal changes in gpdh-1 expression during exposure of worms to 200 mM NaCl. gpdh-1 expression increases rapidly during hypertonic stress and then declines to new steady state levels [18]. As shown in Fig 2D, gpdh-1 expression showed similar rates of increase and steady state levels in wild type and osm mutants. However, the decline in gpdh-1 expression observed 2 h after hypertonic stress was significantly (P<0.01) more rapid in osm-9 (ok1677) and osm-12(n1606) animals. We conclude that enhanced volume and glycerol homeostasis do not contribute to increased short term hypertonic stress resistance in worms with defects in osmotic avoidance behavior.
In a recent genome-wide RNAi screen, we identified 40 genes whose function is essential for survival of C. elegans in hypertonic environments. Loss of these genes results in a hypertonic sensitive or Hos phenoytype. Twenty hos genes play central roles in trafficking and destroying damaged proteins [13]. We subsequently demonstrated that hypertonic stress causes rapid, diverse and widespread protein damage [13][14][15][16]. Worms acclimated to mild hypertonic stress show enhanced survival and greatly reduced protein damage during exposure to more extreme hypertonicity [13][14][15]. Experimental maneuvers that increase protein damage reduce survival under hypertonic conditions [13,16]. Together, these studies indicate that detection and repair and/or destruction of damaged proteins is required for optimal survival during water loss, and suggest that enhanced proteostasis capacity may account for the increased osmotolerance of osm-9(ok1677) and osm-12(n1606) mutants.
To assess whether proteostasis capacity is altered in worm strains with defects in hypertonic avoidance behavior, we crossed osm-9(ok1677) mutant worms with a worm strain expressing polyglutamine (Q35) containing yellow fluorescent (YFP) protein in their body wall muscle cells. As shown in Fig 3A, wild type N2 and Q35::YFP worms showed similar survival when exposed to 500 or 600 mM NaCl. In contrast, survival was strongly enhanced by crossing osm-9(ok1677) into the Q35 strain, a finding consistent with results shown in Fig 1. Q35::YFP is normally fully soluble in the muscles cells of young worms, but undergoes rapid aggregation when the animals are exposed to hypertonic media [13][14][15]. We quantified Q35::YFP aggregation in osm-9(ok1677) mutants exposed to 400 mM NaCl for 24 h or exposed to 500 mM NaCl for 1 h followed by a 3 h after recovery on control medium. As shown in Fig  3B, Q35::YFP aggregation was reduced 40-50% (P<0.005) in osm-9(ok1677) mutants. Total Q35:YFP fluorescence was not significantly (P>0.9) different in wild type versus osm-9 (ok1677) mutants (total YFP fluorescence expressed as arbitrary units was 465 ± 41 and 469 ± 33 in wild type and osm-9(ok1677) worms, respectively; n = 11) demonstrating that differences in aggregation were not due to differences in Q35:YFP expression levels.
We also characterized the functional properties of the Q35::YFP aggregates. Aggregate volume increases rapidly when worms are exposed to hypertonic stress [14]. As shown in Fig 3C, aggregate growth was inhibited~50% (P<0.0004) in osm-9(ok1677) mutants. However, the osm-9(ok1677) mutation had no effect on the solubility of aggregated Q35::YFP proteins. The fluorescence of regions within aggregates did not recover after photobleaching demonstrating that the aggregated proteins are insoluble ( Fig 3D).
Finally, we quantified aging induced protein aggregation. Q35::YFP undergoes a slow, progressive aggregation as C. elegans ages [42]. As shown in Fig 3E, the osm-9(ok1677) mutation had no effect on the accumulation of Q35::YFP aggregates in aging worms. Taken together, data in Fig 3 demonstrate that hypertonicity induced Q35::YFP aggregation is reduced in osm-9(ok1677) mutant worms.
Temperature sensitive (ts) mutations give rise to proteins that fold and function correctly at low or 'permissive' temperatures [43][44][45]. let-60(ga89) encodes a ts mutant of ras GTPase. We have shown previously that hypertonic stress causes apparent misfolding of the let-60(ga89) encoded protein under permissive temperature conditions (16°C) and gives rise to an egg hatching and larval arrest phenotype [14,16].
To determine whether disruption of genes that control hypertonic avoidance behavior modulates hypertonic stress induced protein misfolding, we examined the effect of loss of osm-9 function on the let-60(ga89) mutant phenotype. osm-9 and let-60 are both located on chromosome IV and we were unable to successfully cross worm strains carrying mutant alleles of the two genes. We therefore utilized an RNA interference (RNAi) approach. let-60(ga89) mutants were crossed with the worm strain TU3311 uIs60[Punc-119::yfp + Punc-119::sid-1], which exhibits increased sensitivity of neurons to RNAi [46]. let-60(ga89);uIs60[Punc-119::yfp + Punc-119::sid-1] worms were fed bacteria expressing scrambled (control) or osm-9 dsRNA. As expected, exposure of worms fed scrambled dsRNA at 16°C to 300 mM NaCl caused a striking and significant (P<0.006) increase in egg hatching defects and larval arrest. Silencing of osm-9 function by RNAi reduced the expression of the mutant phenotype by~70% (P<0.005) (Fig 4).
Acclimation to hypertonic stress increases lifespan and resistance to other environmental stressors [47]. Given the improved hypertonic stress resistance and proteostasis capacity of osm-9(ok1677) mutant worms, we therefore also characterized the lifespan of this strain and its resistance to heat, heavy metal and oxidative stress. Median and maximum lifespan were not significantly (P>0.3) different in wild type and osm-9(ok1677) mutants (Fig 5A). The osm-9 (ok1677) mutant exhibited sensitivity to heat shock similar to that of wild type animals ( Fig 5B) and increased sensitivity to the heavy metal cadmium (Fig 5C). Resistance to high concentrations of the quinone juglone was significantly (P<0.04) increased in osm-9(ok1677) worms Hypertonic stress induced aggregation of Q35::YFP in body wall muscle cells of osm-9(ok1677) mutant worms. (A) Survival of wild type N2, Q35::YFP and Q35;osm-9(ok1677) worms exposed to 500 or 600 mM NaCl for 24 h. Values are means ± S.E. of 6 experiments. *P<0.001 and **P<0.01 compared to wild type N2 and Q35 worms. (B) Hypertonic stress induced Q35::YFP aggregation in wild type Q35 worms and Q35;osm-9(ok1677) mutants. Values are means ± S.E. of 4 experiments with a total of 32-78 worms. The number of Q35::YFP aggregates was quantified in worms exposed to 400 mM NaCl for 24 hour or 500 mM NaCl for 1 h followed by 3 h recovery on control medium. Values are means ± S.E. (n = 3). *P<0.005 and **P<0.001 compared to wild type Q35 worms. (C) Q35::YFP volume in wild type Q35 worms and Q35;osm-9(ok1677) mutants. Worms were exposed to 500 mM NaCl for 1 h and then allowed to recover on control medium before aggregate volumes were quantified. Values are means ± S.E. (n = 20 aggregates in 5-6 worms). ( Fig 5D). Quinones like juglone generate reactive oxygen species and form adducts with diverse macromolecules [48].
The enhanced proteostasis capacity of osm-9(ok1677) mutants could result from a number of different physiological changes. When exposed to hypertonic stress, protein synthesis is rapidly reduced in C. elegans and this reduction in translation functions in part to minimize protein damage [15,16,18]. More extensive reductions in translation could reduce protein damage in osm-9(ok1677) mutants. However, reductions in rates of translation were not significantly different (P>0.7) in wild type and osm-9(ok1677) worms exposed for 2 h to 200 mM NaCl ( Fig  6A). Net rates of total protein degradation, measured as the decrease in 35 S-methionine labeling following a 6 h exposure to 200 mM NaCl in the presence of 500 μg/ml cycloheximide, were also not significantly (P>0.7) different in the two groups of animals ( Fig 6B).
The improved proteostasis capacity of the osm-9(ok1677) worm strain could be due to altered expression of genes encoding components of the proteostasis network. To begin testing this possibility, we performed RNA-sequencing gene expression analyses on wild type and osm-9(ok1677) worms under control and hypertonic stress conditions. In unstressed osm-9 (ok1677) worms we found significant (P<0.05) upregulation of genes that have well defined or likely roles in protein degradation, synthesis and folding as well as genes with RNAi phenotypes of increased protein aggregation, altered sensitivity to protein aggregation and/or decreased hypertonic stress resistance. For example, in control osm-9(ok1677) worms we identified 23 upregulated (1.4-to 3.2-fold) genes that play known or presumptive roles in protein degradation including proteases, proteins associated with lysosome function, E3 ubiquitin ligases and E2 ubiquitin-conjugating enzymes ( Table 2). RNAi silencing of one of these genes, vha-3, a vacuolar proton ATPase subunit likely required for lysosome function, results in increased protein aggregation during hypertonic stress and decreased hypertonic stress resistance [13]. Knockdown of C17H11.6, which encodes a predicted E3 ubiquitin ligase, increases sensitivity to protein aggregation-induced paralysis [49].  [14,16]). let-60(ga89) mutants were crossed with uIs60[Punc-119::yfp + Punc-119::sid-1] worms to increase the sensitivity of neurons to RNAi (46). let-60(ga89);uIs60[Punc-119::yfp + Punc-119::sid-1] worms were fed bacteria expressing scrambled (control) or osm-9 dsRNA and maintained at 16°C on agar plates containing either 51 or 300 mM NaCl.
Values are means ± S.E. (n = 3-5 experiments with 100-300 eggs). *P<0.006 compared to control worms exposed to 51 mM NaCl. **P<0.005 compared to control worms exposed to 300 mM NaCl. Twenty-two genes with diverse functions that are not associated with proteostasis in obvious ways were upregulated 1.5-to 9.9-fold in control osm-9(ok1677) worms (Table 2). RNAi knockdown of these genes results in increased protein aggregation, reduced resistance to hypertonic stress and/or changes in sensitivity to protein aggregation-induced paralysis.
osm-9(ok1677) worms placed under hypertonic stress also exhibited significant (P<0.05) differential upregulation of genes associated with proteostasis ( Table 3). Eight of these genes have RNAi phenoptypes of increased protein aggregation, altered sensitivity to protein aggregation and/or decreased hypertonic stress resistance.

Discussion
ASH chemosensory neurons mediate avoidance behavior to hypertonic solutions as well as other noxious chemical and mechanical stimuli [59]. Animals harboring mutations that disrupt hypertonic avoidance behavior (Fig 1 and Fig 3A) or in which ASH neurons have been ablated (Fig 1) exhibit improved survival during hypertonic stress. osm-9, osm-12 and sra-6, which drives ASH ablation [21], are expressed in ASH as well as other neuron types [39]. The only cell types in which expression of these three genes has been shown to overlap are ASH and ASI chemosensory neurons [39]. ASI neurons function in dauer formation, chemotaxis and navigation [59]. However, genetic ablation of ASI or loss-of-function mutation of a gene required for ASI function had no effect on survival on hypertonic environments (Fig 1D). Therefore, we propose that loss of ASH neuron functions required for avoidance of hypertonic environments enhances basal hypertonic stress resistance.
Enhanced hypertonic stress resistance in animals with defective ASH neuron function is not due to alterations in systemic volume regulation or glycerol metabolism (Fig 2). These conclusions are consistent with our earlier findings demonstrating that dramatic reductions in glycerol accumulation do not reduce short term survival during hypertonic stress [17]. As we have shown previously, proteostasis capacity is a critical limiting factor for survival during hypertonic stress [13]. Consistent with this, we found that osm-9(ok1677) mutant and osm-9(RNAi) worms exhibit reduced protein damage during hypertonic stress (Figs 3 and 4).
Improved proteostasis during hypertonic stress in osm-9(ok1677) mutant worms may be due in part to the upregulation of genes that play known or likely roles in protein synthesis, folding and degradation (Tables 2 and 3). RNAi phenotypes of various upregulated genes such  as the vacuolar proton ATPase vha-3, the E3 ubiquitin-protein ligase C17H11.6 and the eukaryotic translation elongation factor eef-1A.2 include increased protein aggregation, increased sensitivity to hypertonic stress and increased sensitivity to protein aggregationinduced paralysis [13,49,50]. Several studies have shown that diverse proteins with no known role in proteostasis slow or prevent protein aggregation [60][61][62][63][64]. Many genes upregulated in osm-9(ok1677) worms also have no known proteostasis function, but protein aggregation is increased by RNAi knockdown of their activity. Two of these genes showed particularly striking upregulation. Expression of col-176 and let-4 was increased compared to wild type animals 9.9-and 3.6-fold and 9.9-and 5.4-fold under control and hypertonic stress conditions, respectively (Table 2 and 3).
col-176 encodes a cuticle collagen. RNAi silencing of this gene increases aggregation of misfolding-prone mutant human SOD1 protein [51]. Several in vitro studies have shown that collagen interacts with aggregation prone proteins and slows the growth of toxic amyloid fibrils [60][61][62]. Increased expression of col-176 could minimize hypertonic stress induced protein damage by similar mechanisms. In addition, various collagens have been shown previously to play an important role in osmotic stress signaling in C. elegans [17,37]. Remodeling of the extracellular matrix by increased expression of col-176 may alter signaling processes that directly or indirectly regulate proteostasis mechanisms activated in response to hypertonic stress.
let-4 encodes a leucine-rich repeat protein expressed on the apical surface of epidermal and epithelial cells [65]. Knockdown of let-4 increases Q35::YFP aggregation [50]. RNAi screening in mammalian neuronal cells identified a leucine-rich repeat protein that regulates mutant huntingtin protein aggregation [66]. The leucine-rich repeat is a protein motif that mediates protein-protein interactions and thus plays important roles in diverse cellular processes [67]. Proteins containing leucine-rich repeat motifs function in protein degradation [68,69] and have been implicated in protein aggregation and impaired autophagy associated with Parkinson's Disease [70]. In C. elegans, let-4 likely plays an important role in maintaining epithelial integrity [65]. Disruption of epithelial barrier function is expected to disrupt salt and water homeostasis, which in turn could lead to increased protein aggregation. Increased let-4 expression in osm-9 (ok1677) worms may improve epithelial barrier function under hypertonic conditions. While speculative, it is also possible that leucine-rich repeat proteins like collagens may preferentially bind with denatured proteins and slow aggregate formation.
It should be stressed here that while RNAseq has been shown to be an accurate tool for assessing differential gene expression [71][72][73], additional studies will be needed confirm the results shown in Tables 2 and 3. Furthermore, it is conceivable that the changes in gene expression and reduced protein damage observed in osm-9(ok1677) mutant worms may be unique to this strain and not to loss of osm-9 function per se. Therefore, it will also be important to determine if other worm strains with defective ASH neuron function show similar differences in gene expression and proteostasis capacity. If our findings are confirmed, it will be necessary to determine the cellular locations of altered gene expression and the mechanisms by which differentially expressed genes confer increased hypertonic stress resistance.
An animal's nervous system plays a critical role in detecting its complex and constantly changing external environment and integrating those environmental cues to coordinately regulate multiple physiological processes. Because of its relatively simple nervous system and genetic and molecular tractability, C. elegans has begun to provide insights into the underlying molecular mechanisms. For example, Prahlad and Morimoto [9] have shown that calcium activated dense core vesicle neurosecretion from C. elegans thermosensory neurons normally suppresses heat shock transcription factor 1 (HSF-1) dependent chaperone expression in muscle and intestinal cells. Signaling from ASH as well as ASI sensory neurons negatively regulates the innate immune response by MAP kinase dependent inhibition of the expression of genes that are part of the noncanonical unfolded protein response pathway [74]. In a similar fashion, osmosensory signaling from ASH neurons could regulate osmoprotective gene expression in non-neuronal cells. However, additional studies will be needed to more thoroughly test this possibility.
In summary, we have demonstrated that defects in ASH mediated hypertonic avoidance behavior are associated with enhanced hypertonic stress resistance and improved proteostasis. Our findings are consistent with a growing body of work demonstrating that intercellular communication between neuronal and non-neuronal cells plays a critical role in integrating cellular stress responses with other organismal physiological demands and associated energy costs.