The insulin/IGF signaling cascade modulates SUMOylation to regulate aging and proteostasis in Caenorhabditis elegans

Abstract

Although aging-regulating pathways were discovered a few decades ago, it is not entirely clear how their activities are orchestrated, to govern lifespan and proteostasis at the organismal level. Here, we utilized the nematode Caenorhabditis elegans to examine whether the alteration of aging, by reducing the activity of the Insulin/IGF signaling (IIS) cascade, affects protein SUMOylation. We found that IIS activity promotes the SUMOylation of the germline protein, CAR-1, thereby shortening lifespan and impairing proteostasis. In contrast, the expression of mutated CAR-1, that cannot be SUMOylated at residue 185, extends lifespan and enhances proteostasis. A mechanistic analysis indicated that CAR-1 mediates its aging-altering functions, at least partially, through the notch-like receptor glp-1. Our findings unveil a novel regulatory axis in which SUMOylation is utilized to integrate the aging-controlling functions of the IIS and of the germline and provide new insights into the roles of SUMOylation in the regulation of organismal aging.

Keywords: C. elegans; SUMOylation; aging; cell biology; germline; insulin/IGF signaling; lifespan; neuroscience; proteostasis.

Figure 1.. The knockdown of daf-2 modulates the SUMOylation of CAR-1 in C.elegans. 

(A) Global protein SUMOylation patterns in homogenates of daf-2 RNAi-treated and untreated wild-type (N2), CF512 and CF1903 worms were compared by western blot using an anti-SUMO antibody. In all three worm strains, several proteins exhibit enhanced levels of SUMOylation upon the knockdown of daf-2 (arrowheads) and others show decreased levels (arrows). (B) Schematic illustration of pulldown procedure to isolate covalently SUMOylated proteins from NX25 animals that express His-Flag-smo-1 in a smo-1 knockout background. Covalently SUMOylated proteins were pulled down and identified by mass spectrometry. (C) Nematodes expressing GFP-tagged CAR-1 (strain WH346) were treated with daf-2 RNAi or left untreated (EV), harvested at day 1 of adulthood and GFP-CAR-1 was immune-precipitated by a GFP antibody and blotted using a SUMO antibody. daf-2 RNAi treatment reduced the level of SUMOylated GFP-CAR-1 that migrated as two bands. One was migrating as a protein of ~ 60 kDa and the other as a protein of ~ 250 kDa. (D) Reblotting the membrane with a GFP antibody showed that daf-2 RNAi treatment had no effect on the amounts of the precipitated GFP-CAR-1 protein (this blot serves as a loading control for C). (E) WH346 worms were either grown on control bacteria (EV) or on daf-2 RNAi bacteria, harvested at day 1 of adulthood and total GFP-CAR-1 amounts in the worm homogenates were analyzed by a western blot. No difference in the total levels of GFP-CAR-1 was observed. (F) Comparison of CAR-1-GFP signals in three independent experiments as in E. G. CF512 worms that were either treated with daf-2 RNAi or left untreated (EV) express similar levels of car-1 as measured by quantitative real-time PCR.

Figure 1—figure supplement 1.. A Wild-type (N2) and two conditionally sterile worms: rrf-3/fer-1 mut (CF512) and glp-1 loss-of-function mutation (CF1903) that were used for the SUMOylation blot (Figure 1A).

All parental worms (F0) were grown in 15°C. Eggs of each strain were extracted and placed on either control bacteria or on daf-2 RNAi bacteria. The synchronized progeny (F1) were developed at 25°C for 48 hr from hatching, and transferred to 20°C for additional 24 hr. While N2 animals were fertile (see eggs in left panels), CF512 and CF1903 animals were sterile, regardless if they were left untreated or grown on daf-2 RNAi bacteria.

Figure 1—figure supplement 2.. (A–C) NX25 worms expressing smo-1 tagged to a His- and Flag-tag in a smo-1 null background, were grown on control (EV) or on daf-2 RNAi bacteria.

At day 1 of adulthood, the worms were harvested and subjected to a His-tag (A) followed by a Flag-tag (B) immunoprecipitation. Flow through and washing steps were free of tagged proteins while the elution exhibited the purified SUMOylated proteins. To verify the purification of SUMOylated proteins, the final eluted samples were subjected to WB using an anti-SUMO antibody (C).

Figure 2.. car-1 regulates lifespan through the germline of C.elegans.

(A) The knockout of car-1 (strain WH377) shortens lifespan compared to wild-type worms (N2) (mean LS of 14.81 ± 0.41 and 17.56 ± 0.52 days, p<0.001). (B) daf-2 mutant (e1370) worms were treated with car-1, daf-16 RNAi or left untreated (EV) and lifespans were followed. car-1 RNAi as well as daf-16 RNAi, shortened lifespan compared to control worms, however, the lifespan-shortening effect of daf-16 RNAi was more prominent than that of car-1 RNAi (mean LS of 17.56 ± 0.56, 40.02 ± 1.40 and 50.52 ± 1.38 days, respectively (p<0.004, EV vs. car-1 RNAi)). (C) The lifespans of car-1 RNAi-treated and untreated daf-16 mutant worms (strain CF1038) were indistinguishable (mean LS of 15.74 ± 0.47 and 15.37 ± 0.48 days, p=0.3). (D–E) Worms that express CAR-1 K185R (EHC118) live approximately 53% longer than their wild-type counterparts (mean LS of 22.51 ± 0.60 and 14.70 ± 0.59 days respectively, p<0.001), as shown by a representative experiment (D) and a summary of three independent experiments (E). (F) Worms that express CAR-1 K257R (EHC121) and wild-type animals have nearly identical lifespans (16.32 ± 0.54 and 16.84 ± 0.56 days, respectively, p=0.25). (G) The longevity of CAR-1 K185R expressing worms is DAF-16-dependent (mean LS of 18.32 ± 0.47 (daf-16 RNAi) and 21.73 ± 0.61 (EV) days, p<0.001). However, daf-16 RNAi-treated CAR-1 K185R and wild-type worms had nearly identical lifespans (16.90 ± 0.63 and 16.67 ± 0.52 respectively, p=0.42). (H) daf-16 RNAi shortens the lifespans of CAR-1 K257R expressing worms (mean LS of 14.95 ± 0.47 days) compared to untreated wild-type animals (18.09 ± 0.41 days, p<0.001).

Figure 2—figure supplement 1.. The knockdown of car-1 shortens lifespan.

(A) The lifespans of daf-16 RNAi- and car-1 RNAi-treated CF512 worms that were developed at 25°C and transferred to 20°C at day 1 of adulthood, were shorter than that of untreated animals (EV) (mean LS of 14.87 ± 0.48 compared to 18.01 ± 0.63 days, respectively, p<0.0001). In contrast, daf-2 RNAi extended the animals’ lifespans. (B) The knockdown of car-1 by RNAi tend to shorten the lifespans of daf-2 mutant e1368 worms (mean LS of 28.43 ± 0.65 (EV) and 27.46 ± 0.54 (car-1 RNAi, NS).

Figure 2—figure supplement 2.. The computational GPS-SUMO tool identified the SUMOylation consensus sequences in CAR-1 pointing at two lysine residues, 185 and 257 as potential SUMOylation sites.

Figure 2—figure supplement 3.. Worms that over-express either WT CAR-1 (EHC117), CAR-1 K185R (EHC118) or CAR-1 K257R (EHC121) were harvested at day 1 of adulthood, homogenized and equal protein amounts were subjected to IP using an HA antibody (all exogenous CAR-1 proteins are tagged with a double HA tag).

The sediment proteins were loaded on gels and blotted using a SUMO antibody. Our results show that CAR-1 is less SUMOylated in EHC118 and EHC121 compared to EHC117 worms. These results show that CAR-1 is SUMOylated on more than one site.

Figure 2—figure supplement 4.. CAR-1 K185R extends lifespan.

(A) Lifespans of worms that overexpress the mutated CAR-1 K185R protein were significantly longer than these of wild-type animals (21.30 ± 0.61 and 14.92 ± 0.63 days respectively, p=1.54E–11; 21.12 ± 0.70 and 15.91 ± 0.56 days, respectively, p=8.99E–08). (B) Median lifespans of worms that were included in the experiments that are displayed in Figure 2E. (C) The over-expression of wild-type CAR-1 has no effect on lifespan, as animals that express this protein and wild-type worms had very similar lifespans (16.15 ± 0.46 and 15.82 ± 0.39 days respectively, p=0.297). The knockdown of daf-16 by RNAi shortened the lifespans of these animals (mean LS of 11.74 ± 0.37 days, p<0.001 compared to untreated CAR-1 over-expressing worms).

Figure 3.. The knockdown of car-1 modulates the activity of glp-1.

(A) Untreated (EV) and car-1 RNAi-treated, long-lived glp-1 mutant worms (strain CF1903) show no difference in lifespans (mean LS of 17.76 ± 0.72 and 18.13 ± 0.85 days, respectively p=0.37). In contrast, daf-16 RNAi reduced the lifespan of these animals (10.71 ± 0.0.33 days, p<0.001). (B) car-1 RNAi treatment has no significant effect on the lifespans of kri-1 mutant worms (mean LS of 16.59 ± 0.43 (EV) and 15.96 ± 0.39 (car-1 RNAi), p=0.14). In contrast, daf-16 RNAi shortened the lifespans of these animals (mean LS of 13.37 ± 0.26 days (daf-16 RNAi), p<0.001). (C) DAPI stained image of gonads of worms of the indicated genotypes. Bar = 50 mm. (D) The over expression of CAR-1 (strain EHC117) or of the mutated K185R CAR-1 (EHC118) resulted in significantly reduced number of germ cells in the worms’ gonads. The expression of K185R CAR-1 (in EHC118 animals) reduces the number of germ cells by ~ 22% compared to the number that was observed in worms that express the wild-type CAR-1 (EHC117) (p<0.01). (E) Quantification of germline apoptosis by acridine orange staining. car-1 RNAi elevates the average number of apoptotic cells in the gonads of wild-type worms by ~ 2.5-fold. The average numbers of apoptotic cells in the gonads of untreated N2 and of EHC117 worms are nearly identical (3.36 and 3.48, respectively). In contrast, the expression of K185R CAR-1 elevates the average number of apoptotic cells by ~ 80% compared to untreated N2 and EHC117 (p<0.0001, bars represent ± SEM). (F–H) Worms expressing CAR-1 K185R (EHC118) have reduced number of progeny compared to N2 animals (F). The average total number of progeny of EHC118 animals was 105 while control animals had an average of 271 offspring (G). No significant difference in the brood size of wild-type (N2) and EHC117 worms (H). (I) The knockdown of car-1 increases the expression levels of sygl-1 compared to the levels detected in untreated daf-2 (e1370) but decreases the levels of lst-1. (J) The knockdown of glp-1 (in CF1903 worms that were exposed to 25°C) lowers the expression levels of car-1. No such effect was observed in wild-type animals.

Figure 3—figure supplement 1.. CAR-1 does not appear to function through signals from the somatic gonad.

(A) The lifespans of daf-9 mutant worms (strain CF2531) were shortened by daf-16 RNAi (14.53 ± 0.36 days, p=8.54E–14), as well as by the knockdown of car-1 compared to the control animals (17.48 ± 0.45 compared to 19.16 ± 0.44 days, respectively, p<0.01). (B) Similarly, the lifespans of daf-12 mutant nematodes (strain AA86) were reduced by daf-16 RNAi (13.70 ± 0.60 days, p<0.001) and car-1 RNAi compared to untreated animals (15.03 ± 0.34 and 17.19 ± 0.38 days, respectively, p<0.001).

Figure 3—figure supplement 2.. The knockdown of car-1 affects egg-laying patterns of daf-2, daf-16 and kri-1 mutants but not of daf-9 and daf-12 mutant animals.

(A–B) daf-2 mutant (e1370, strain CB1370) were grown on either control bacteria (EV), daf-16 or car-1 RNAi bacteria. daf-16 null (mu86) worms (strain CF1038) were either grown EV bacteria or on car-1 RNAi. Living progeny per worm were counted. The knockdown of car-1 rendered worms of both strains to near sterility. The average number of live progeny of car-1 RNAi-treated, daf-16 mutant worms was ~ 4 and of daf-2 (e1370) mutant worms was nearly zero. (C–D) Wild-type (strain N2) and worms that carry mutated kri-1 (strain CF2052) were subjected to egg-laying assay as in A. The knockdown of car-1 by RNAi resulted in ~ 46% reduction in the brood size of N2 worms (average numbers of progeny 244 and 133 for untreated and car-1 RNAi treated worms, respectively) and in ~ 91% reduction in CF2052 worms (an average of 116 and 10 living progeny per a single N2 and CF2052 worm, respectively). (E–F) daf-9 (strain CF2531) and daf-12 (strain AA86) mutant animals were grown on either control (EV) or car-1 RNAi bacteria, and living progeny were counted. The knockdown of car-1 affected neither the egg-laying patterns nor the brood sizes of these worms.

Figure 3—figure supplement 3.. (A–B) Wild-type (N2) and CF1903 worms were grown in either 15 or 25°C (CF1903 worms lack glp-1 when grown in 25°C).

The worms were homogenized and qPCR was performed to examine the levels of the glp-1-target genes, sygl-1 (A) and lst-1 (B). The expression levels of both genes were remarkably and significantly decreased in the absence glp-1.

Figure 3—figure supplement 4.. The knockdown of car-1 by RNAi tends to elevate the expression level of sygl-1 in N2 worms.

However, this effect was not significant (p=0.07).

Figure 4.. CAR-1 modulates proteostasis in C.elegans.

(A–B) Worms expressing Aβ_3-42 in their body wall muscles were treated with car-1, daf-2 RNAi or left untreated (EV). While daf-2 RNAi protected from paralysis, car-1 RNAi significantly increased paralysis (N = 5). (C–D) The dilutions of car-1 RNAi (orange) or of daf-2 RNAi (red) bacteria with control bacteria do not significantly change the effects of these treatments on Aβ-mediated paralysis. Concurrent knockdown of daf-2 and car-1 by RNAi only partially protects Aβ worms from paralysis (C, blue). The increased rate of paralysis after car-1 RNAi treatment and the reduced paralysis after car-1 and daf-2 knockdown were significant compared to the level seen in untreated worms (EV) (N = 4, p<0.04) (D). (E–F) The expression of CAR-1 K185R in Aβ worms (strain EHC124) protects the animals from paralysis (E). Three independent experiments confirmed the significance of this observation (F) (p<0.004).

Figure 4—figure supplement 1.. CAR-1 K185R but not CAR-1 K257R modulates proteotoxicity.

(A) An exogenous expression of CAR-1 K185R protects worms from Aβ-mediated proteotoxicity (strain EHC124) compared to Aβ worms. This protective effect is abolished by the knockdown of daf-16 by RNAi. (B) In contrast, the expression CAR-1 K257R does not affect the rate of paralysis in Aβ-expressing worms (strain EHC125) compared to Aβ worms.

Figure 5.. CAR-1 modulates proteostasis through the GLP-1 axis.

(A–B) Worms expressing polyQ35-YFP in their body wall muscles were crossed with CF1903 animals carrying a ts mutant glp-1. PolyQ35-YFP and polyQ35-YFP/glp-1 worms were exposed during development to 25˚C and either left untreated (EV) or treated with daf-2 or car-1 RNAi and subjected to paralysis assay. While daf-2 RNAi protected worms of both strains from paralysis, car-1 RNAi enhanced paralysis of polyQ35-YFP worms but not of polyQ35-YFP/glp-1 animals (A). Three independent experiments confirmed the significance of these phenotypes (B). (C–D) Aβ worms were either left untreated or fed with daf-2, car-1 or cgh-1 RNAi bacteria and rates of paralysis were followed. While daf-2 RNAi protected the worms from proteotoxicity, the knockdown of cgh-1 and car-1 enhanced paralysis compared to control animals (C). Three independent experiments confirmed the significance of this phenotype (D, p<0.035).

Figure 5—figure supplement 1.. The number of body bends of worms expressing polyQ67-YFP in a pan-neuronal fashion, was measured at days 1 and 4 of adulthood.

At day 1 of adulthood, the daf-2 RNAi-treated worms presented a significantly increased number of body bends per 30 s compared to control animals. At day 4 of adulthood, the knockdown of car-1 reduced the number of bends and daf-2 RNAi significantly increased the number of bends compared to the untreated EV worms (p<0.001).

Figure 5—figure supplement 2.. The knockdown of cgh-1 or of daf-16 by RNAi similarly reduces lifespan of CF512 worms compared to the lifespans of control animals (mean LS of 13.15 ± 0.25, 13.11 ± 0.28 and 16.18 ± 0.40 days respectively, p<0.001).

Figure 5—figure supplement 3.. The knockdown of car-1 and of cgh-1 enhance proteotoxicity.

(A) HE250 worms were grown at 15˚C until day 1 of adulthood and then transferred to 25˚C, and the paralysis assay was performed at this temperature. Nematodes treated with RNAi against car-1 or cgh-1 showed an increased rate of paralysis compared to untreated (EV) worms. (B) Three independent experiments were performed. At day 2 of adulthood, the knockdown of car-1 significantly increased the rate of paralysis (p=0.027). At day 3 of the experiment, car-1 and cgh-1 RNAi increased (p<0.0085) and daf-2 RNAi decreased (p<0.0085) the paralysis rate of HE250 worms at 25°C.

Figure 6.. The roles of car-1 in stress resistance.

(A–C) The survival rates of heat-stressed CF512 worms (A) N2 animals (B) that were exposed to 35°C for 11 hr, and of daf-2(e1370) mutant worms (strain CB1370) (C) that were exposed for 19 hr to 35°C, were not significantly affected by car-1 RNAi. In contrast, daf-16 RNAi significantly reduced the survival rates of heat-stressed daf-2 mutant animals (p<0.001). (D) The expression of CAR-1 K185R (EHC118) significantly elevates survival compared to the wild-type animals (average survival of 67.39% and 43.51% respectively, N = 3, p<0.04). A trend but no significant effect was observed in worms that over-express the natural CAR-1 protein (EHC117). (E) car-1 RNAi-treated CF512 worms are more resistant to the pathogenic bacteria Pseudomonas aeruginosa than control worms (mean survival of 6.77 ± 0.21 and 5.88 ± 0.15 days, p<0.001). daf-2 RNAi prolonged and daf-16 RNAi reduced the survival of worms that were grown throughout adulthood with P. aeruginosa (mean survival rates of 11.96 ± 0.42 and 4.28 ± 0.11 days, respectively, p<0.001 for both treatments). (F) The over expression of wild-type CAR-1 (strain EHC117) or of CAR-1 K185R (strain EHC118) shortens survival of worms that were grown during adulthood on P. aeruginosa (mean survival of 6.05 ± 0.17, 5.43 ± 0.14 and 5.66 ± 0.17 days for N2, EHC117 (p<0.005) and EHC118 (p=0.1029), respectively). (G) The survival of CF512 worms that were exposed to sub-lethal dose of UV radiation was significantly increased by daf-2 and car-1 RNAi treatments compared to control animals (EV); (mean survival of 10.45 ± 0.22, 9.23 ± 0.22 and 7.97 ± 0.21 days, respectively, p<0.001). (H) In agreement, the over-expression of wild-type CAR-1 (strain EHC117) or of CAR-1 K185R (strain EHC118) shortens survival of animals that were exposed to UV radiation (mean survival of 12.78 ± 0.33, 8.89 ± 0.19 and 10.87 ± 0.27 days for N2, EHC117 (p<0.001) and EHC118 (p<0.001), respectively).

Figure 7.. A model The IIS negatively regulates its downstream transcription factors DAF-16, SKN-1 and HSF-1 by inhibiting their entrance into the nucleus (I).

Thus, knocking down daf-2 hyper-activates these transcription factors resulting in longevity, proteostasis, stress resistance and modulated reproduction profile. The IIS also governs aging by SUMOylating CAR-1 on lysine 185 to mitigate its regulatory function (II), on GLP-1 (III). The RNA-helicase CGH-1 acts in cooperation with CAR-1 to regulate glp-1. IIS reduction hyper-activates CAR-1 by lowering the level of its SUMOylation on K185. This modulates the activity of GLP-1 to mediate longevity and enhance proteostasis in a DAF-16-dependent manner. Our results also indicate that GLP-1 positively controls the expression of car-1 (IV). CAR-1 also appears to affect lifespan, proteostasis, stress resistance and reproduction by a DAF-16-independent mechanism (V).