Dr. Anat Ben-Zvi, Department of Life Sciences, National Institute for Biotechnology in the Negev, Ben-Gurion University of the Negev, 1 Ben Gurion Blvd., Beer Sheva 84105, Israel. Tel.: +972 8 647 9059; fax: +972 8 647 9041; e-mail: email@example.com
All cells rely on highly conserved protein folding and clearance pathways to detect and resolve protein damage and to maintain protein homeostasis (proteostasis). Because age is associated with an imbalance in proteostasis, there is a need to understand how protein folding is regulated in a multicellular organism that undergoes aging. We have observed that the ability of Caenorhabditis elegans to maintain proteostasis declines sharply following the onset of oocyte biomass production, suggesting that a restricted protein folding capacity may be linked to the onset of reproduction. To test this hypothesis, we monitored the effects of different sterile mutations on the maintenance of proteostasis in the soma of C. elegans. We found that germline stem cell (GSC) arrest rescued protein quality control, resulting in maintenance of robust proteostasis in different somatic tissues of adult animals. We further demonstrated that GSC-dependent modulation of proteostasis requires several different signaling pathways, including hsf-1 and daf-16/kri-1/tcer-1, daf-12, daf-9, daf-36, nhr-80, and pha-4 that differentially modulate somatic quality control functions, such that each signaling pathway affects different aspects of proteostasis and cannot functionally complement the other pathways. We propose that the effect of GSCs on the collapse of proteostasis at the transition to adulthood is due to a switch mechanism that links GSC status with maintenance of somatic proteostasis via regulation of the expression and function of different quality control machineries and cellular stress responses that progressively lead to a decline in the maintenance of proteostasis in adulthood, thereby linking reproduction to the maintenance of the soma.
The correct folding and assembly of proteins and protein complexes is essential for cellular function. Accordingly, cells have evolved specialized machines to maintain proper protein homeostasis (proteostasis). These protein quality control networks are comprised of molecular chaperones and degradation machineries, such as the proteasome and the autophagy pathway, that ensure proper folding and the efficient clearance of misfolded or damaged proteins, respectively (Finley, 2009; Gidalevitz et al., 2010; Haynes & Ron, 2010; Tyedmers et al., 2010; Arias & Cuervo, 2011; Walter & Ron, 2011).
When the proteostatic capacity of the cell becomes limiting, an imbalance in the folding environment ensues, leading to the induction of cytoprotective stress responses designed to suppress protein misfolding and to enhance clearance, such as the heat-shock response (HSR; Akerfelt et al., 2010; Haynes & Ron, 2010; Walter & Ron, 2011). Activation of the heat-shock transcription factor (HSF-1) adjusts the expression of chaperones and other cytoprotective genes to match the elevated levels of damaged proteins, thereby enabling the cell to rebalance proteostasis and increase survival and recovery from stress (Akerfelt et al., 2010). Yet, despite these protective mechanisms, the accumulation of damaged proteins is a well-established marker of aging, with many age-dependent diseases being associated with protein misfolding and aggregation (Gidalevitz et al., 2010; Taylor & Dillin, 2011).
Two possible mechanisms can explain the failure of protein quality control networks to adjust to the cellular folding demands of aged animals. Protein damage and misfolding in such individuals could result from a limited efficiency of the cellular quality control networks in repairing or removing misfolded proteins, leading to a gradual accumulation of damaged proteins over time. Alternatively, the ability of the cellular proteostasis network to rebalance itself may be differentially regulated during the lifespan of the organism. Given that the composition of the cellular proteostasis machinery in Caenorhabditis elegans is modulated during early adulthood (Liu et al., 2011; Taylor & Dillin, 2011; Twumasi-Boateng et al., 2012; Volovik et al., 2012), we hypothesized that the capacity of cellular protein quality control networks may be differentially regulated upon transition to adulthood. This idea is supported by the observed decline in the ability of C. elegans to keep meta-stable proteins folded and mount effective stress responses early in adulthood (Ben-Zvi et al., 2009; David et al., 2010; Taylor & Dillin, 2011). Because this change in proteostasis composition and capacity coincides with the onset of oocyte biomass production, we asked whether disrupting the function of the reproductive system would affect the observed decline in proteostasis maintenance seen during adulthood.
In C. elegans, reproduction is specifically linked to aging by endocrine signaling from germline stem cells (GSCs), a process that alters fat metabolism and extends lifespan (Antebi, 2012). Several pathways have been suggested to respond to life-extending signals sent from the GSC to the soma. One pathway is a lipophilic hormone-signaling pathway that is dependent on daf-12, a nuclear hormone receptor, and on the enzymes that synthesize or modify the daf-12 ligands daf-9/cytochrome P450 and the daf-36/Rieske oxygenase (Hsin & Kenyon, 1999; Gerisch et al., 2007). A second responsive pathway activates daf-16 via a regulatory mechanism that is distinct from that employed in insulin-like signaling pathway (ILS) and which requires kri-1, encoding an intestinal ankyrin-repeat protein, and tcer-1, encoding a putative transcription elongation factor (Berman & Kenyon, 2006; Ghazi et al., 2009). Such regulation alters the expression of genes encoding proteasome components (Vilchez et al., 2012) and can, furthermore, modulate stress survival (Libina et al., 2003). GSC signaling also requires hsf-1 (Hansen et al., 2005), although it is not known whether such regulation occurs independently of the ILS. Finally, GSC signaling modulates fat metabolism in a manner that depends on the daf-12 and daf-16 pathways, as well as on the nuclear receptor nhr-80 (Wang et al., 2008; Goudeau et al., 2011; McCormick et al., 2012) and on the activation of TOR signaling, which regulates autophagy through pha-4 and independently of daf-16 (Lapierre et al., 2011). Thus, GSC proliferation coordinates a regulatory switch that alters endocrine and metabolic signaling to modulate fatty acid metabolism, proteasomal function and autophagy, as well as lifespan (Antebi, 2012). We therefore asked whether reproduction, and specifically, signaling from GSCs, regulates the decline in protein quality control seen early in adulthood.
To determine whether the onset of reproduction is linked to the decline in cellular proteostasis maintenance that occurs in early adulthood, we considered C. elegans strains presenting diverse mutations in the reproductive system to assess whether interfering with different aspects of reproductive function could rescue proteostasis capacity. We found that inhibiting GSC proliferation resulted in robust proteostasis in different somatic tissues of adult animals. Of the known pathways linking signals from GSCs with lifespan, signaling via hsf-1 and daf-16/kri-1/tcer-1 modulated heat-shock (HS) activation in adulthood, while other signaling pathways were required to rescue polyglutamine-dependent toxicity and temperature-sensitive (ts) metastable protein-associated dysfunction. We propose that GSC signaling regulates a set of quality control machineries and cellular stress responses, thereby promoting a regulated switch between the limited and robust states of proteostatic capacity in the soma upon transition to adulthood.
Inhibition of GSC proliferation rescues the decline in thermo-resistance normally seen early in adulthood
Caenorhabditis elegans thermo-resistance declines following the onset of reproduction. Young, wild-type (wt) adults (grown for 50 h at 25 °C) were more resistant to HS than were day 2 adults (grown for 74 h at 25 °C; Fig. S1A and B). While 76.2 ± 5.6% survived if challenged by HS on day 1 of adulthood, only 12.6 ± 2.3% survived if first challenged on day 2 of adulthood. This shift in HS survival rate was rapid, starting 7–12 h after transition to adulthood (76.6 ± 3.4% to 44.7 ± 6%, respectively, P < 0.005), when animals entered the reproductive phase (Fig. 1A). To assess whether the onset of reproduction is linked to the maintenance of proteostasis in aging, we monitored the effects of different sterile mutations on thermo-resistance. Survival of gonadogenesis-defective mutants declined sharply on day 2 of adulthood (Fig. 1B). Likewise, survival on day 2 of adulthood of both male and female gamotogenesis-defective mutants was limited (Table S1). In contrast, HS resistance and survival rates of GSC proliferation mutant glp-1(e2141ts) (glp-1) animals remained high, namely 71.8 ± 7.6% and 68.6 ± 7.3% on days 2 and 3 of adulthood, respectively (Figs 1B, S1A, and S1B). Prolonging stress resistance was dependent on GSC arrest, as the survival rates of several different GSC proliferation mutants also remained high (Fig. 1B and Table S1). Therefore, as with lifespan, GSC proliferation (but not reproduction per se) is required for modulating stress survival after the onset of reproduction. Specifically, GSC arrest inhibits the rapid and acute decline in the organismal response to stress, suggesting that signals from proliferating GSCs may modulate somatic stress response activation upon transition to adulthood.
Inhibiting GSC proliferation restores efficient activation of the HSR during adulthood
To determine whether stress survival is associated with differential regulation of the HSR, we next compared the induction of representative HS genes by monitoring changes in mRNA levels following HS on days 1 or 2 of adulthood. While the induction of hsp-70 and hsp-16.11 in glp-1 mutant animals remained high, the induction of these genes in wt or gonadogenesis-defective mutant gon-2(q388) (gon-2) animals was reduced by ~65% between the first and second days of adulthood (Figs 1C and S1C). This finding suggests the existence of an early switch in HSR regulation that is mediated by GSC proliferation.
To determine whether GSC proliferation comes at the cost of maintaining a robust stress response in somatic tissues, we utilized a HS transcriptional reporter in which the hsp-16.2 promoter regulates green fluorescent protein (GFP) expression in a stress-dependent manner. We then followed GFP expression patterns and expression levels in intact and germline-less animals. When animals were challenged on day 1 of adulthood, GFP fluorescence was detected in various somatic tissues, predominantly in the intestinal cells of both wt and glp-1 mutant individuals (Fig. 1D). While the ability to induce hsp-16.2-regulated GFP expression in wt animals declined strongly by day 3 of adulthood, with only 8.2 ± 6.6% of the animals expressing GFP in the intestine (P < 0.05), the ability to induce hsp-16.2-regulated GFP in somatic tissues of glp-1 mutant animals remained high up to day 10 of adulthood (Figs 1D,E, and S1D). Similar results were observed upon expression of an hsp-70 (C12C8.1) promoter-regulated reporter and depended on glp-1 loss-of-function (Fig. S1E). No GFP fluorescence was detected in untreated wt or glp-1 animals, with hsp-16.2-regulated GFP induction declining if the glp-1 animals were fertile (Fig. S1F and S1G), suggesting that hsp-16.2-regulated GFP induction can be attributed to glp-1 loss-of-function. These results suggest that cell-nonautonomous signals from proliferating GSCs regulate the cellular HSR in somatic tissues, whereas interfering with such signals prevents the decline in HSR activation seen in adult animals, thereby increasing stress survival.
Dissecting the signaling pathways required to reset the HSR in adulthood
To understand the mechanism by which GSCs regulate the activation of HS response in adult somatic tissues, we asked whether genes that are required for GSC lifespan-extending signaling also play a role in resetting proteostasis. As discussed above, several signaling pathways are specifically required for GSC-dependent longevity. Inactivation of daf-16 (and of its GSC-specific modifiers kri-1 and tcer-1), hsf-1, daf-12, daf-9, daf-36, nhr-80, or pha-4 suppresses the enhanced longevity of glp-1 animals, with most of these pathways showing little effect on the lifespan of wt animals (Antebi, 2012). We therefore examined whether inactivation of these genes also affected somatic proteostasis. We utilized the hsp-16.2-regulated GFP expression reporter and followed GFP expression patterns in glp-1 animals treated with RNA interference (RNAi) for genes involved in GSC-dependent lifespan regulation. When challenged on day 3 of adulthood, 98.2 ± 1.2% of the glp-1 animals grown on control RNAi were able to induce hsp-16.2-regulated GFP in the intestine. This ability to induce hsp-16.2-regulated GFP levels was not affected by RNAi knockdown of daf-12, daf-9, daf-36, nhr-80, or pha-4 (Figs 2A and S2A). In agreement, HS survival rates of glp-1 animals harboring a mutation in daf-12(rh61rh411), daf-9(daf-9(rh50)), daf-36(daf-36(K114)) or nhr-80(tm1011) on day 2 of adulthood remained high (Fig. 2B), while RNAi or mutations in any of the above genes alone did not affect HS activation or survival (Fig. S2B and S2C). To extend our observations, we also examined the effect of RNAi-mediated knockdown of daf-12, nhr-80, and pha-4 on the ability of glp-1 mutant animals to respond to HS by examining the expression levels of the HS genes, hsp16 (hsp16.2 and hsp16.11) and hsp70 (C12C8.1 and F44E5.4), following HS on the second day of adulthood. The induction of these genes was not significantly affected (Fig. S2D–G). In contrast, RNAi-mediated knockdown of the stress transcription factors daf-16 or hsf-1 reduced hsp-16.2-regulated GFP induction in the intestine of glp-1 mutant animals on day 3 of adulthood (33.3 ± 7% and 4.7 ± 2.7%, respectively, P < 0.005); while reducing hsf-1 levels affected all HS genes examined, daf-16 knockdown affected only some (Figs 2A and S2). However, the daf-16(mu86) deletion mutant repressed glp-1-mediated thermo-resistance on day 2 of adulthood, reducing survival to 19.6 ± 6.9% (P < 0.05; Fig. 2B). The effects of daf-16 and hsf-1 knockdown on the HSR of glp-1 animals were similar to those seen on wt animals on day 1 of adulthood (37.3 ± 4.1% and 16.1 ± 3.8%, respectively, Figs 2A and S2B). This finding suggests that signals from proliferating GSCs converge on the cellular stress response of somatic tissues and modulate DAF-16 and HSF-1 function in a cell-nonautonomous manner.
The function of daf-16 and hsf-1 in enhancing proteostasis is also regulated cell-nonautonomously by the ILS signaling pathway (Libina et al., 2003; Taylor & Dillin, 2011). We thus asked whether daf-16 and hsf-1 regulation also occur independently of the ILS pathway. For hsf-1, no GSC-specific effectors are known. In contrast, down-regulation of kri-1 or tcer-1 was shown to affect daf-16 function in germline-less animals but not in animals with decreased ILS signaling (Berman & Kenyon, 2006; Ghazi et al., 2009). We therefore examined whether knockdown of kri-1 or tcer-1 modulated HS activation and survival of germline-less animals. Silencing of kri-1 or tcer-1 by RNAi reversed hsp-16.2-dependent GFP induction in glp-1 animals by 68% and 56%, respectively (P < 0.005), and reduced the induction of other HS genes, yet did not affect the ability of wt animals to induce a HSR on day 1 of adulthood (Figs 2A and S2). Likewise, a kri-1(ok1251) deletion mutant completely suppressed glp-1-mediated thermo-resistance on day 2 of adulthood, reducing survival to 19.7 ± 11.3% (P < 0.05; Fig. 2B). Our data suggest that GSC signaling modulates DAF-16 (and possibly HSF-1) function upon transition to adulthood. This claim is supported by the observation that germline ablation can extend the lifespan of daf-2 mutants (Hsin & Kenyon, 1999). Indeed, microarray analysis of DAF-16 targets identified many nonoverlapping targets between genes showing GSC-regulated expression and genes previously identified in a daf-2(-) expression analysis (McCormick et al., 2012). Likewise, whereas hsp16.2 induction was shown to be independent of daf-16 in a daf-2(-) background (McColl et al., 2010), hsp16.2 was affected by GSC-regulated DAF-16 function (Figs 2A, S2A and S2D). Nonetheless, we cannot rule out a possible contribution of ILS-dependent regulation of DAF-16 and HSF-1 upon transition to adulthood.
Inhibiting GSC proliferation postpones the onset of age-associated misfolding of meta-stable proteins
The ability of C. elegans to maintain meta-stable proteins that are dependent on the cellular folding capacity for proper folding declines early in adulthood. Because the inhibition of GSC proliferation affects the expression of genes encoding for proteasome and autophagy pathway components, as well as modulating the function of HSF-1 and DAF-16 (Antebi, 2012), we asked whether signals from GSCs compromise somatic proteostasis and lead to the age-dependent misfolding of meta-stable proteins observed during C. elegans adulthood (Ben-Zvi et al., 2009). We first examined the function of perlecan unc-52(e669,su250) (unc-52(ts)), a previously characterized folding sensor. This sensor contains a temperature-sensitive missense mutation that shows age-dependent paralysis due to disruption of body wall muscle myofilament anchoring during adulthood (Ben-Zvi et al., 2009). While unc-52(ts) mutant animals grown under permissive conditions (during adulthood) were partially paralyzed on day 4 of adulthood (48 ± 7.7%), the motility of unc-52(ts) in a glp-1 or mes-1 mutant background was significantly improved (7.4 ± 3% and 7.9 ± 3.2%, respectively, P < 0.005). In contrast, the motility of animals expressing unc-52(ts) in the gon-2 mutant background was not significantly different (55.3 ± 8.1%) from that of unc-52(ts) mutant animals (Fig. 3A).
We then examined the folding of the naturally occurring meta-stable proteins, myosin and dynamin, which show age-dependent misfolding and aggregation later in adulthood (Ben-Zvi et al., 2009; David et al., 2010). We first monitored the age-dependent myofilament disruption (sarcopenia) and motility decline associated with myosin misfolding (Ben-Zvi et al., 2009). As expected, by the seventh day of adulthood, the myofilaments of wt animals were mostly disrupted (83.6 ± 5.3%). In contrast, the integrity of the myofilaments of glp-1 animals was well maintained (14.3 ± 1.8%, P < 0.005; Fig. 3B and C). Accordingly, the motility of wt animals quickly declined, whereas the glp-1 mutant animals were mostly unaffected (Fig. 3D). Germline arrest also altered the reported age-dependent mislocalization of the metastable protein dynamin-1 (DYN-1), a protein that participates in neuronal synaptic vesicle recycling (Ben-Zvi et al., 2009). While DYN-1 was mostly mislocalized by the fifth day of adulthood in wt animals (90 ± 5.8%, P < 0.05), such mislocalization was rarely observed in glp-1 animals (14 ± 4.7%; Fig. 3E). Thus, inhibition of GSC proliferation before transition to adulthood abrogated age-dependent decline in protein quality control in muscle, neuronal and intestinal cells (Figs 1E and 3), resulting in the maintenance of robust proteostasis and stress resistance in these somatic tissues during adulthood.
Inhibiting GSC proliferation postpones the onset of protein misfolding and toxicity in a polyglutamine disease model
A suggested consequence of proteostatic decline in adulthood is the onset of aging-associated protein misfolding diseases. Accordingly, we monitored protein aggregation propensity and cellular toxicity of polyglutamine-expansion (polyQ) proteins in glp-1 mutant animals. By the second day of adulthood, glp-1 animals expressing polyQ-yellow fluorescent protein (YFP) with 35 repeats (Q35) in body wall muscle developed 45% less visible foci than did Q35-expressing animals (P < 0.005; Fig. 4A and B). In agreement, when Q35 aggregation was monitored using semi-denaturing detergent–agarose gel electrophoresis (SDD–AGE), 67% less insoluble high molecular weight (HMW) species that could only be dissolved by boiling were detected in Q35;glp-1 than in Q35 animals already at day 2 of adulthood (Fig. 4C and D), even though no effect in terms of Q35 protein levels were noted (Fig. S3). When motility was examined as a measure of toxicity, the onset of Q35-mediated paralysis was significantly delayed in glp-1 animals. By the sixth day of adulthood, only 18.3 ± 4.2% of Q35;glp-1 animals were paralyzed, as compared to 88.1 ± 5.1% of Q35 animals (P < 0.05; Fig. 4E). Thus, GSC arrest modulates the onset and progression of protein aggregation in a polyQ disease model.
Misfolding of polyQ proteins was shown to impact many different aspects of the protein quality control machinery, including chaperones, proteasome-mediated degradation, and autophagy (Tyedmers et al., 2010). To examine whether GSC signaling pathways similarly regulate age-dependent toxicity of polyQ proteins, we monitored the roles of the signaling pathways specifically required for GSC-dependent longevity in modulating the Q35-dependent paralysis of germline-less animals. Similar to the regulation of HSR, the delay in the onset of Q35-mediated paralysis in glp-1 animals was largely unaffected by RNAi of daf-12, daf-9, and daf-36 on day 6 of adulthood, as compared to animals fed control RNAi (11.7 ± 2.5%, 14.4 ± 1.3%, 20.7 ± 2.2% and 13.1 ± 2.2% respectively). By contrast, daf-16, kri-1, and tcer-1 RNAi significantly inhibited glp-1-dependent rescue of Q35-mediated toxicity (35.7 ± 3.3%, 54.2 ± 9% and 46 ± 2.4%, respectively, P < 0.005; Fig. 4F). Likewise, knockdown of hsf-1 resulted in a severe phenotype with 65.7 ± 4% (P < 0.005) paralysis on day 5 of adulthood and 100% lethality on day 6 of adulthood (data not shown). In contrast to the regulation of HSR, treatment with RNAi for nhr-80 or pha-4 also inhibited glp-1-dependent rescue of Q35-mediated toxicity (33 ± 2.2% and 27.6 ± 2.3%, respectively, P < 0.005), suggesting that the specific requirements of a given protein determines its susceptibility to GSC-dependent signaling pathways. This is further supported by our observation that knockdown of hsf-1, daf-12, and daf-9 (and to a lesser extent, daf-36, nhr-80 and pha-4) inhibited glp-1-dependent rescue of unc-52(ts) paralysis, while knockdown of daf-16, kri-1, and tcer-1 did not affect the motility of unc-52(ts) animals (Fig. S4). We therefore propose that the effect of GSCs on proteostasis is due to a regulatory switch mechanism that links GSC status with the maintenance of somatic proteostasis by regulating a set of quality control machineries and cellular stress responses, each affecting different cellular functions, yet all required for longevity.
Cell-nonautonomous regulation of cellular quality can restore proteostasis in adulthood
The question of how dynamic is the functional capacity of cellular protein quality control networks during the lifespan of an organism remains open. Several lines of evidence suggest that the machineries responsible for general maintenance of proteostasis undergo significant changes when animals enter fertile adulthood. The ability to activate cellular stress responses, such as the heat-shock response and the unfolded protein response, is disrupted following transition to adulthood (Ben-Zvi et al., 2009). Likewise, the activation of c-Jun N-terminal kinase (JNK) pathway was shown to enhance DAF-16 transcriptional activity during development but to inhibit such activity after the transition to adulthood (Twumasi-Boateng et al., 2012). Finally, epidermal growth factor signaling (EGF) was shown to change the expression of proteostasis components, such as elements of the ubiquitin-proteasome system and chaperone genes, early in adulthood (Liu et al., 2011). This switch is associated with altered requirements of stress transcription factors at the transition to adulthood (Cohen et al., 2010; Volovik et al., 2012). Thus, there is a differential regulation of protein quality control networks at the transition to adulthood. A corresponding change is observed in the functional capacity of the cellular quality control machinery that progressively loses the ability to maintain metastable proteins during adulthood (Ben-Zvi et al., 2009; David et al., 2010; Fig. 5A). Our findings suggest that the switch between the two functional states of proteostatic networks that occurs once the animals enter reproductive adulthood is regulated by signals from proliferating GSCs. These signals, which are mediated via different signaling pathways and which affect the composition and function of the proteostasis network (Lapierre et al., 2011; Vilchez et al., 2012), limit the ability of somatic cells to maintain proteostasis (Fig. 5B and C).
What are the consequences of reduced proteostasis capacity in adulthood? For an aggregation-prone protein that is highly dependent on the proteostasis network (Tyedmers et al., 2010), like Q35, such reduced capacity resulted in the fast accumulation of aggregates as early as day 2 of adulthood (Fig. 4). For metastable proteins that are responsive to the load of misfolded proteins in the cell (Gidalevitz et al., 2010), misfolding was observed during the first few days of adulthood, earlier than when the corresponding wt proteins were affected (Ben-Zvi et al., 2009). In agreement, analysis of aggregate accumulation during adulthood identified several proteins that aggregated early in adulthood and showed that wt proteins increasingly aggregated during adulthood (David et al., 2010). When proteostasis was modified by GSC arrest early in adulthood, the beneficial effects of maintaining robust proteostasis were observed throughout adulthood, with the onset of protein misfolding and aggregation being delayed for aggregation-prone metastable and wt proteins. We therefore propose that a switch in proteostasis composition and function early in adulthood can result in progressive damage accumulation during adulthood, with severity of the phenotype depending on the genetic polymorphism of the organism (Gidalevitz et al., 2006, 2010).
Different signaling pathways mediate cross talk between reproduction and somatic maintenance
Active signaling cascades initiated by proliferating GSCs were first identified as modulators of lipid metabolism and lifespan (Hsin & Kenyon, 1999; Arantes-Oliveira et al., 2002). While inactivation of any of the GSC signaling pathways has only mild effects on the lifespan of wt animals, all pathways are required for GSC-dependent extension of lifespan (Antebi, 2012). Whereas changes in lipid metabolism in GSC-arrested animals are mostly regulated by the daf-12, nhr-80, and pha-4 signaling pathways (Goudeau et al., 2011; Lapierre et al., 2011; McCormick et al., 2012), GSC-dependent regulation of autophagy requires pha-4, yet is independent of daf-16. At the same time, GSC-dependent regulation of proteasome function requires daf-16 and daf-12 but not hsf-1 (Lapierre et al., 2011; Vilchez et al., 2012). Here, we show that different GSC signaling pathways are required to modulate somatic proteostasis. While both daf-16 and hsf-1 are required for the activation of HSR in adulthood (Fig. 2), the reduced toxicity of Q35;glp-1 animals is mediated by hsf-1, daf-16, nhr-80, and pha-4 (Fig. 4), whereas unc-52(ts)-dependent paralysis is modulated by hsf-1, daf-12, and daf-9 (Fig. S4). Autophagy and proteasomal function in germline-less animals was shown to be regulated by pha-4 and daf-16, respectively (Lapierre et al., 2011; Vilchez et al., 2012), suggesting that different pathways affect the expression of a specific set of proteostasis modulators and therefore those proteins that depend on these machineries for their function. This premise implies that GSC signaling pathways do not overlap and cannot functionally complement each other. Indeed, daf-16 and daf-12 were shown to regulate a unique set of genes with very little overlap in their expression patterns, when regulated by GSC signaling (McCormick et al., 2012). We thus propose that GSC signaling activates several different regulatory programs that differentially modulate somatic functions, such as metabolism, cellular quality control, and responses to various stress conditions, resulting in a progressive loss of the organism's ability to maintain somatic proteostasis and other repair and defense functions. The combined actions of all of these pathways are required for lifespan enhancement, although activation of only some of the pathways may suffice to improve the health of the organism during adulthood (Fig. 5C and D).
One potential interpretation of our results is that GSC signaling regulates proteostasis capacity of somatic cells by modulating the cell-autonomous function of various transcription factors. However, several signaling pathways that are modulated by GSC arrest were shown to function only in specific tissues. For example, nhr-80 was shown to function in intestine and neurons (Goudeau et al., 2011), while kri-1 and tcer-1 were shown to modulate DAF-16 function specifically in intestine (Berman & Kenyon, 2006). Thus, it remains to be examined whether the different signaling pathways function downstream of germline signaling to modulate proteostasis in somatic cells in a cell-autonomous manner or by modulating cell-nonautonomous signaling between different somatic tissues.
Uncoupling somatic proteostasis maintenance from GSC signaling
The link between GSC proliferation and somatic function suggests that the commitment to reproduction is reported to the soma, resulting in altered metabolism and cellular repair that can affect the rate of aging. This interpretation is supported by the fact that different mutations in DNA damage checkpoint genes affect somatic stress resistance and lifespan (Olsen et al., 2006). Given that reproduction is also affected by environmental conditions, it is likely that, in return, signals from the soma can impact GSC proliferation. This idea is further supported by the finding that cell-nonautonomous signaling via kri-1 regulates germline cell death in response to DNA damage (Ito et al., 2010). Thus, kri-1 can integrate signals from both the germline and the soma to determine GSC reproduction potential and somatic cellular repair. It is of note that the reproduction potential of GSCs can modulate both protein homeostasis (this work), repair and defense responses (Arantes-Oliveira et al., 2002; Alper et al., 2010), and lipid metabolism (Wang et al., 2008; Lapierre et al., 2011), thus serving to link reproduction, metabolism, and somatic repair.
The ability demonstrated in this study to restore somatic proteostatic function in adulthood strongly indicates that C. elegans possesses a highly efficient cellular repair system that is actively down-regulated once reproduction is initiated. Identifying these signals and then uncoupling somatic proteostasis from the reproductive system may restore the intrinsic ability of the cell to maintain its proteome. Defining the cell-nonautonomous signals that regulate proteostasis could therefore offer novel approaches for the treatment of age-dependent diseases with different etiologies but similar underlying biology.
Nematodes and growth conditions
A list of strains used in this work and name abbreviations is given in Tables S2 and S3 (Supporting information). Nematodes were grown on NGM plates seeded with the Escherichia coli OP50 or OP50-1 strains at 15 °C. Unless otherwise stated, 15–30 embryos, laid at 15 °C, were transferred to fresh plates and grown at 25 °C for the duration of the experiment. The first day of adulthood (day 1) was set at 50 h after temperature shift, before the onset of egg laying. Animals were moved every 1–2 days during the reproductive period to avoid progeny contamination. HS-treated animals were discarded after scoring.
Experiments were repeated at least three times, and >30 animals per experimental condition were scored. Data are presented as means ± SEM. P-values were calculated using the Wilcoxon–Mann–Whitney rank sum test to compare two independent populations. (*) denotes P <0.05 and (**) denotes P <0.005. For Figs 1E, 3D, and 4E, data were compared with age-matched wt animals. For Figs 2A, 4F, S2, and S4, data were compared with empty vector control. For all other figures, data were compared with wt animals on day 1 of adulthood.
Unless otherwise stated, a total of 30–50 age-synchronized animals grown at 25 °C were used for each assay. Animals were moved to fresh plates, and the plates were then sealed and placed in a 37 °C bath for 90 min. Animals were frozen or fixed immediately following stress.
Animals were picked at the indicated ages and transferred to a 24-well plate containing HS buffer (100 mm Tris–HCl, pH 7.4, 17 mm NaCl and 1% cholesterol supplemented with bacteria). These animals were then subjected to a 37 °C HS for 6 h. HS buffer was supplemented with SYTOX orange (Invitrogen, Carlsbad, CA, USA), and animal survival was scored by monitoring dye uptake, using a Leica M165 FC fluorescent stereoscope with a TXR filter. Fluorescent animals were scored as dead. For comparison, kinetic thermo-resistance assays were performed in which >70 animals/strain were subjected to a 37 °C HS, and survival was tested by manual scoring (Fig. S1A and S1B). In agreement with previously reported data (Libina et al., 2003; Wang et al., 2008), when the HS treatment was extended (9 h at 37 °C), glp-1 animals were significantly more thermo-resistant than were wt animals (42.2 ± 2.9% and 21 ± 2.3%, respectively, P < 0.005). However, given that the HSR in wt animals declines in this timeframe (Fig. 1A), we monitored survival after 6 h.
Fluorescence stress reporters
Animals expressing GFP under control of the hsp-16.2 promoter (phsp-16.2::GFP) or mCherry fluorescent protein under control of the hsp-70 (C12C8.1) promoter (phsp-70::mCherry) were crossed with glp-1(e2141ts) animals. Animals expressing GFP or mCherry were subjected to HS, and fluorescence was monitored 18–24 h later. Animals expressing GFP or mCherry in the gut were scored as HS-induced.
A total of 30–50 eggs were placed on E. coli strain HT115(DE3) transformed with the indicated RNAi vectors (obtained from the Ahringer or Vidal RNAi libraries), as previously described (Ben-Zvi et al., 2009). RNAi knockdown of mRNA levels was controlled by comparing the mRNA levels of the target gene with mRNA levels of animals fed on strain HT115(DE3) bacteria containing the empty vector (pL4440).
Total RNA was extracted from wild-type or glp-1 animals untreated or subjected to HS (see HS treatment, above). RNA was extracted using the TRIzol reagent (Invitrogen). For cDNA synthesis, mRNA was reverse-transcribed using the iScript™ cDNA Synthesis Kit (Bio-Rad, Hercules, CA, USA). Quantitative PCR was performed on a C1000 Thermal Cycler (Bio-Rad) with SsoFas EvaGreen Supermix (Bio-Rad).
Protein expression levels
Synchronized nematodes were collected and lyzed in SDS sample buffer. Samples were analyzed by immunoblot, using anti-paramyosin antibody 5-23 (Hybridoma Bank, Iowa city, IA, USA) and anti-GFP antibodies (Enco Scientific, Petach Tikvah, Israel). Peroxidase-conjugated AffiniPure goat anti-rabbit and goat anti-mouse antibodies (Jackson Immuno-Research, West Grove, PA, USA) were used as secondary antibodies.
Animals expressing punc-54::Q35::YFP were crossed with glp-1(e2141ts) animals. For quantification, the number of bright foci of age-synchronized animals expressing Q35-YFP was counted.
For semi-denaturing agarose gel electrophoresis, 120 animals were collected and extracts were prepared by mechanical disruption under moderately denaturing conditions. Samples were diluted in SDD–AGE buffer (50 mm Tris–HCl, pH 6.8, 2% (w/v) SDS, 0.025% bromophenol blue, 5% glycerol), incubated at room temperature or heated at 98 °C for 10 min, and loaded onto a 1.5% agarose gel as previously described (Eremenko et al., 2013). Samples were analyzed by immunoblot using anti-GFP antibodies (Enco Scientific).
A total of 15–30 age-synchronized animals were used for each assay. Animals were grown at 25 °C for the duration of the experiment. Animals were moved every day, and paralysis was scored by monitoring animals' movement 10 min after transfer to a new plate. Animals that did not move were scored as paralyzed.
Stiff-body paralysis assay
Animals expressing mutant unc-52(ts) were grown at 25 °C until day 1 of adulthood, then shifted to 15 °C, and paralysis-scored.
Animals expressing GFP-tagged myosin heavy chain A (MYO-3::GFP) were crossed with glp-1(e2141ts) animals. wt or glp-1 embryos expressing MYO-3::GFP were grown at 25 °C until the indicated age. To assess MYO-3::GFP mislocalization, synchronized adults expressing the transgene were fixed (Gidalevitz et al., 2006), and GFP fluorescence was monitored. Animals were imaged using an Olympus Fluoview FV1000 confocal microscope (Olympus, Tokyo, Japan) through a 60 × 1.0 numerical aperture objective with a 488-nm line for excitation. More than 210 cells per experimental condition were scored.
Immunofluorescence studies were performed as previously described (Gidalevitz et al., 2006). Animals were stained with anti-dynamin-1 antibodies (Hybridoma Bank). Secondary DyLight 488 goat anti-mouse IgG antibodies (Jackson Immuno-Research, West Grove, PA, USA) were used. Animals were imaged using an Olympus Fluoview FV1000 confocal microscope through a 60 × 1.0 numerical aperture objective with a 488-nm line for excitation.
Some nematode strains used in this work were provided by the Caenorhabditis Genetics Center, which is funded by the NIH National Center for Research Resources (NCRR). The monoclonal antibodies developed by M. Nonet and H.F. Epstein were obtained from the Developmental Studies Hybridoma Bank developed under the auspices of the NICHD and maintained by the Department of Biology, University of Iowa. N. Shemesh was supported by Fay and Bert Harbour award. A.B.-Z. was supported by the Israeli Council for Higher Education Alon Fellowship, by a Marie Curie International Reintegration grant, and by a grant from the Binational Science Foundation.
AB, N. Shemesh, and N. Shai designed the study, performed the experimental work, and wrote the manuscript.