Strains and genetics.

N2 strain worms were used as wild type (WT). The following mutants were used: TQ37C,gbb-1(tm1406) outcrossed eight times; TQ2692C,lgc-36(xu3l) outcrossed four times; TQ2690C,lgc-38(xu2l) outcrossed four times; TQ1247C,lgc-35(tm1444) outcrossed four times; TQ2689C,gab-1(xu1l) outcrossed four times; TQ1490C,exp-1(ox276) outcrossed eight times; TQ40C,unc-49(e407) outcrossed four times; TQ1069C,dvIs19[(pAF15)gst-4::gfp::NLS]; TQ2750C,gbb-1(tm1406);dvIs19[(pAF15)gst-4::gfp::NLS]; TQ2751C,exp-1(ox276);dvIs19[(pAF15)gst-4::gfp::NLS]; TQ1067C,dvIs70[Phsp16.2::GFP+rol-6(+)]; TQ2752C,exp-1(ox276);dvIs70[Phsp16.2::GFP+rol-6(+)]; TQ1071C,Is[sod-3::gfp]; TQ2753C,gbb-1(tm1406);Is[sod-3::gfp]; TQ2754C,exp-1(ox276);Is[sod-3::gfp]; TQ2755C,lgc-36(xu3l);Is[sod-3::gfp], wheregfp is green fluorescent protein,gst-4 is glutathioneS-transferase 4,hsp-16.2 is heat shock protein-16.2, NLS is nuclear localization signal,rol-6 is roller-6, andsod-3 is superoxide dismutase 3. RNA interference (RNAi) clones were from the Ahringer library and confirmed by sequencing. These mutants were extensively outcrossed to our N2 strain four to six times before life span assay.

CRISPR/Cas9 protocol.

Gene editing by clustered regularly interspaced short palindromic repeats (CRISPR) was performed as described previously (19). Targeted mutagenesis was induced via nonhomologous end joining. A CRISPR design tool from the Zhang laboratory at Massachusetts Institute of Technology (seehttps://zlab.bio/guide-design-resources) was used to design specific guide RNAs. The 20-bp target sequence was inserted into the vector pTM55 via PCR, and the PCR product was digested with DpnI (New England Biolabs) at 37°C for 2 h to clean up template DNA and then transformed into competent cells. Plasmids were verified by sequencing. Dumpy-10 (dpy-10, a reporter gene for CRISPR screening) sgRNA,dpy-10 repair template, CRISPR-associated protein-9 (cas9) expression vector, and sgRNA plasmids of target gene were coinjected. F1 worms with dumpy or roller phenotype were picked and screened for CRISPR-generated deletions by PCR. Deletion mutants oflgc-36,lgc-38, andgab-1 were obtained successfully, but not forlgc-37, as PCR amplification of thelgc-37 genomic DNA was unsuccessful because of technical reasons. Deletions in bothlgc-36(xu3l) andlgc38(xu2l) result in frameshift and premature stop codon that truncates the four transmembrane domains, which are crucial for the ion channel function. Thegab-1(xu1l) has a 69-bp deletion in the first exon to delete the signal sequence, which would cause a failure of cell surface trafficking of the ion channel. Thus the deletion mutants oflgc-36,lgc-38, andgab-1 all likely represent null alleles (Supplemental Fig. S1; Supplemental Material for this article is available online athttps://doi.org/10.6084/m9.figshare.9255620).

Worm synchronization.

Strains were grown at 20°C for at least three generations before life span or health span determination. Twentyday 2 adult worms were transferred to fresh 60-mm nematode growth medium (NGM) plates. The adult worms were removed from the plate after laying eggs for 3 h. The plate was placed at 20°C for 2 days. L4 larvae were used to start the life span or health span assays.

Life span assay.

Life span studies were performed at 20°C as previously described (16,50). Approximately 100 worms of each genotype were used for life span assay and transferred every other day to fresh NGM plates. Survival rate was scored every day, and worms were censored if they crawled off the plate, bagged, or exhibited protruding vulva. Life span data were analyzed with GraphPad Prism 5 (GraphPad Software) and IBM SPSS Statistics 19 (IBM). Log-rank (Kaplan-Meier) test was used to calculateP values.

RNA interference assay.

RNA interference (RNAi) was performed using the RNAi-compatible OP50 bacterial strain OP50(xu363), as previously described (49). RNAi plates included carbenicillin (100 μg/mL) and isopropyl-β-d-thiogalactoside (1 mM). OP50(xu363) bacteria with vector or RNAi plasmid were seeded on RNAi plates 2 days before experiment. Worms were fed RNAi bacteria, beginning at the egg stage.

Oxidative stress.

Synchronized populations of worms were grown at 20°C.Day 2 adult worms were transferred to NGM plates with 0.01%tert-butyl hydroperoxide (tBHP, no. 458139; Sigma) with 30 worms per plate. Survival was scored every hour until all animals died. At least three replicates were performed for each experiment (4).

Heat stress.

Synchronized populations of worms were grown at 20°C.Day 2 adult worms were heat shocked as described previously (36). Heat shock was performed in a water bath at 36.2°C for 2.5 h, and worms were allowed to recover at 20°C for 10 h. The survival rate was scored every 2 h with eyelash pick. At least three replicates were performed for each experiment.

Ultraviolet stress.

Twenty synchronized L4 larvae were transferred to seeded NGM plates with 2 μg/mL 5-fluoro-2′-deoxyuridine and cultured at 20°C for 2 days. Worms were then exposed to ultraviolet (UV) radiation (18 J/m²) in an ultraviolet cross-linker (SCIENTZ 03-II). After recovery for half an hour, worms were transferred to freshly seeded NGM plates and scored for viability every 12 h at 20°C. At least three replicates were performed for each experiment (40).

C. elegans slow-killing assay.

To examinePseudomonas aeruginosa PA14 infection, a slow-killing assay was used as previously described (42). Briefly, overnight PA14 culture was seeded on NGM plates and incubated at 37°C and 25°C for 24 h each. Thirty synchronized L4 larvae (WT or mutants) were transferred to each PA14 plate and incubated at 25°C. Infected worms were scored for survival every 12 h. At least three replicates were performed for each experiment.

Imaging.

To quantifyPhsp-16.2::gfp,Pgst-4::gfp, andsod-3::gfp fluorescence, worms were immobilized on an agarose pad with 10 mM tetramisole, and images were acquired on an Olympus upright microscope (BX53) under ×10 objectives by MetaMorph (Molecular Devices) and processed by ImageJ as previously described (49).

Loss of the GABA_B receptor gene gbb-1 and GABA_A excitatory receptor gene exp-1 extends life span.

GABA signals through both ionotropic and metabotropic receptors. TheC. elegans genome encodes a group of ionotropic GABA_A receptor family members:unc-49,lgc-38,lgc-36,lgc-37,gab-1,exp-1, andlgc-35, with the first five encoding inhibitory chloride-selective channels and the last two encoding excitatory cation-selective channels (21). Besides ionotropic GABA_A receptors, theC. elegans genome also encodes two metabotropic GABA_B receptor genes:gbb-1 andgbb-2 (Fig. 1A;17). Among these nine GABA receptor genes, onlyunc-49,exp-1,lgc-35,gbb-1, andgbb-2 have mutants available. Thus, as a first step, we attempted to generate null mutants of the rest of the GABA receptor genes using the CRISPR/Cas9 approach and succeeded in obtaining deletion mutants oflgc-36,lgc-38, andgab-1 (Fig. 1B).

We assayed the life span of these GABA receptor mutants. Thegbb-1 mutant worms were long-lived, consistent with our previous report (Fig. 1C;16). Notably, we found thatexp-1 mutant worms were also long-lived (Fig. 1D). By contrast, mutant worms lackinggbb-2,unc-49,lgc-38,lgc-36,gab-1, orlgc-35 all showed normal life span (Fig. 1,E–J). These results identify a novel function for the excitatory GABA_A receptor geneexp-1 in longevity beyond its conventional role in behavioral control.

EXP-1-dependent life span regulation requires heat shock transcription factor 1.

Longevity genes tend to converge on a handful of transcription factors to regulate life span (34). We previously showed that GBB-1 regulates life span through the FOXO transcription factor abnormal Dauer formation protein-16 (DAF-16;16). We thus asked how EXP-1 controls life span. By examining the major transcription factors known to regulate life span, we found that RNAi of heat shock transcription factor 1 (hsf-1) completely suppressed the life span-extending phenotype ofexp-1 mutant worms (Fig. 2A), whereas RNAi of other transcription factor genes such asdaf-16, skinhead-1 (skn-1), and defective pharyngeal development protein-4 (pha-4) did not (Fig. 2,B–D), suggesting that EXP-1 regulates life span via HSF-1. To obtain further evidence, we assessed whether loss of EXP-1 promotes HSF-1 activity by assaying the expression level ofhsp-16.2::gfp transgene, a commonly used reporter for HSF-1 activity (23);hsp-16.2 is a target gene of HSF-1. We found that the expression level ofhsp-16.2::gfp transgene reporter was upregulated inexp-1 mutant worms, which can be fully suppressed byhsf-1 RNAi (Fig. 2,E andF). These data suggest that EXP-1 regulates life span via HSF-1. As GBB-1 regulates life span through DAF-16, these results indicate that different GABA receptors may regulate life span through distinct transcription factors.

Loss of GBB-1 and EXP-1 enhances oxidative stress resistance via the Nrf2 transcription factor SKN-1.

Though life span is a key parameter of aging, it is not the only one. Health span represents an equally, if not more, important measure in normal aging (27). As animals age, their ability to survive environmental insults deceases. Thus, we next examined stress resistance in GABA receptor mutants. We first checked oxidative stress by exposing worms to tBHP, which produces reactive oxygen species (46). We found thatgbb-1 andexp-1 mutants exhibited enhanced resistance to oxidative stress (Fig. 3,A andB). By contrast, mutant worms lacking other GABA receptors had normal life spans upon tBHP treatment (Fig. 3,C–H). These results demonstrate thatgbb-1 andexp-1 mutants are not only long-lived but also exhibit enhanced resistance to oxidative stress.

We then asked how GBB-1 and EXP-1 regulate oxidative stress resistance. InC. elegans, reactive oxygen species detoxification is primarily mediated by the nuclear factor erythroid 2-related factor 2 (Nrf2) transcription factor SKN-1 (3). We therefore examined SKN-1 and found that RNAi ofskn-1 completely suppressed the oxidative stress resistance phenotype ofgbb-1 andexp-1 mutant worms (Fig. 4,A andE), whereas RNAi of other transcription factor genes such asdaf-16,hsf-1, andpha-4 did not (Fig. 4,B–D andF–H), indicating that SKN-1 is required for GBB-1 and EXP-1 to regulate oxidative stress resistance. To obtain further evidence, we examined the expression level ofgst-4::gfp transgene, a commonly used reporter for SKN-1 activity (15);gst-4 is a target gene of SKN-1. We found that the expression level ofgst-4::gfp reporter was upregulated ingbb-1 andexp-1 mutant worms (Fig. 4,I andJ), providing additional evidence that GBB-1 and EXP-1 regulate oxidative stress resistance via SKN-1. Our results show that GBB-1 and EXP-1 regulate life span via DAF-16 and HSF-1, respectively, but regulate oxidative stress resistance via a different transcription factor, SKN-1. This reveals an interesting phenomenon, namely, that the same GABA receptors may regulate life span and health span through distinct mechanisms.

Loss of GBB-1, EXP-1, and UNC-49 promotes resistance to heat stress.

Heat stress resistance is another important measure of health span. We found that in addition togbb-1 andexp-1 mutants,unc-49 mutant worms also exhibited enhanced resistance to heat stress (Fig. 5,A–C). By contrast, other GABA receptor mutants had normal life spans upon heat treatment (Fig. 5,D–H). It is worth noting that thoughunc-49 mutant worms showed normal life span, this mutant exhibited enhanced resistance to heat stress. This supports the notion that life span and health span are not always tightly coupled (7).

Heat shock response is mainly regulated by the transcription factor HSF-1 (23). However, RNAi ofhsf-1 did not completely suppress the enhanced heat stress resistance phenotype detected ingbb-1,exp-1, andunc-49 mutants (Fig. 6,B,F, andJ), suggesting an HSF-1-independent mechanism. We then turned our attention to other transcription factors such as DAF-16, SKN-1, and PHA-4, which have also been shown to regulate stress responses. We found that RNAi ofdaf-16 completely suppressed the heat stress resistance phenotype ofgbb-1 mutants (Fig. 6A), whereas RNAi ofskn-1 orpha-4 did not (Fig. 6,C andD). We also tested the expression level ofsod-3::gfp transgene, a commonly used reporter for DAF-16 activity (25), and found that it was upregulated ingbb-1 mutant worms (Fig. 6,M andN). Thus, GBB-1 regulates heat stress resistance through DAF-16. However, none of the tested transcription factors was required for the enhanced heat stress resistance observed inexp-1 (Fig. 6,E–H) andunc-49 mutants (Fig. 6,I–L), indicating that they may require a different mechanism. This indicates that different GABA receptors may regulate heat stress resistance through different mechanisms, revealing complexity of GABA signaling in heat stress regulation.

Loss of EXP-1 and LGC-36 enhances resistance to UV stress via the FOXO transcription factor DAF-16.

Cells and tissues are constantly attacked by intrinsic and extrinsic genotoxic insults, causing DNA lesions. Accumulation of DNA lesions is a driving factor for aging. The capacity to maintain genome homeostasis is another important health span parameter. UV irradiation causes DNA damage and is a commonly used method to assess a worm’s capacity to maintain genome homeostasis (24). We found thatexp-1 andlgc-36 mutants exhibited enhanced resistance to UV insults (Fig. 7,A andB), whereas other GABA receptor mutants did not (Fig. 7,C–H). These results identify a role for GABA receptors in regulating DNA damage response.

The FOXO transcription factor DAF-16 is known to mediate UV-induced DNA damage response. Indeed, RNAi ofdaf-16, but not other transcription factors, completely suppressed the enhanced UV resistance phenotype observed inexp-1 (Fig. 8A) andlgc-36 mutant worms (Fig. 8E). The expression level ofsod-3::gfp transgene, a reporter for DAF-16 activity, was also upregulated both inexp-1 andlgc-36 mutants (Fig. 8,I andJ). These data suggest that EXP-1 and LGC-36 regulate UV stress resistance through DAF-16.

Loss of GBB-1 promotes innate immunity via the Nrf2 transcription factor SKN-1.

As animals age, they not only are more vulnerable to environmental insults but also become more susceptible to infection by pathogens as their immune system declines with age. Therefore, we assessed whether GABA receptors regulate a worm’s defense againstP. aeruginosa PA14, an infectious bacterial strain commonly used to assay innate immunity in worms (42). We found that loss of GBB-1 protected worms from PA14 infection. Mutations in other GABA receptor genes had no effect (Fig. 9A). RNAi of the Nrf2 transcriptional factorskn-1 completely suppressed the enhanced resistance to PA14 (Fig. 9B), whereas RNAi of other transcriptional factors includingdaf-16,hsf-1,pha-4, andpmk-1 did not (Fig. 9,C–F). This suggests that GBB-1 regulates innate immunity via SKN-1.

Another hallmark of aging is age-dependent motor function decline. We recorded locomotion behavior of GABA receptor mutants throughout the life span and found that they all showed a normal rate of decline in locomotion speed compared with WT worms. Although the locomotion speed ofunc-49 mutant worms was slower than that of WT worms, these mutant worms displayed a similar rate of motor activity decline with age (Supplemental Fig. S5). We also assayed the pharyngeal pumping rate of GABA receptor mutants and found that there was no significant difference between mutant worms and WT worms (Supplemental Fig. S6). These data suggest that GABA receptors regulate health span not through affecting locomotion or feeding behavior.