The intestinal TORC2 signaling pathway contributes to associative learning in Caenorhabditis elegans

Abstract

Several types of associative learning are dependent upon the presence or absence of food, and are crucial for the survival of most animals. Target of rapamycin (TOR), a kinase which exists as a component of two complexes, TOR complex 1 (TORC1) and TOR complex 2 (TORC2), is known to act as a nutrient sensor in numerous organisms. However, the in vivo roles of TOR signaling in the nervous system remain largely unclear, partly because its multifunctionality and requirement for survival make it difficult to investigate. Here, using pharmacological inhibitors and genetic analyses, we show that TORC1 and TORC2 contribute to associative learning between salt and food availability in the nematode Caenorhabditis elegans in a process called taste associative learning. Worms migrate to salt concentrations experienced previously during feeding, but they avoid salt concentrations experienced under starvation conditions. Administration of the TOR inhibitor rapamycin causes a behavioral defect after starvation conditioning. Worms lacking either RICT-1 or SINH-1, two TORC2 components, show defects in migration to high salt levels after learning under both fed and starved conditions. We also analyzed the behavioral phenotypes of mutants of the putative TORC1 substrate RSKS-1 (the C. elegans homolog of the mammalian S6 kinase S6K) and the putative TORC2 substrates SGK-1 and PKC-2 (homologs of the serum and glucocorticoid-induced kinase 1, SGK1, and protein kinase C-α, PKC-α, respectively) and found that neuronal RSKS-1 and PKC-2, as well as intestinal SGK-1, are involved in taste associative learning. Our findings shed light on the functions of TOR signaling in behavioral plasticity and provide insight into the mechanisms by which information sensed in the intestine affects the nervous system to modulate food-searching behaviors.

Fig 1. TOR-specific inhibitor rapamycin disrupts normal taste associative learning.

Chemotaxis indices of wild-type worms after conditioning on low (25 mM) or high (100 mM) salt-containing agar plates with (+) or without (–) food. (B) Schematic representation of the timing of rapamycin treatment. Rapamycin or vehicle was administered to worms after 4 days of cultivation. Error bars, s.e.m.; **p < 0.01 (EtOH vs. each condition, one-way ANOVA with Dunnett’s post hoc test, N ≥ 4).

Fig 2. TORC1 signaling regulates migration to low salt concentrations through S6K.

(A) A schematic diagram of putative TORC1 signaling in C. elegans. (B) Salt chemotaxis of wild-type N2 worms and rsks-1(ok1255) mutants after salt conditioning. (C) Expression of rsks-1 cDNA under the pan-neuronal H20 promoter, but not ASER-specific gcy-5 promoter or intestinal ges-1 promoter, rescued the chemotaxis abnormality of rsks-1(ok1255) mutants. Salt chemotaxis was tested after high-salt/food(–) conditioning. (D) Salt chemotaxis of wild-type N2 animals and atg-13(bp417) mutants after salt conditioning. *p < 0.05, **p < 0.01, n.s. = not significant (one-way ANOVA with Dunnett’s post hoc test, N ≥ 9).

Fig 3. TORC2 components regulate migration to high salt concentrations.

(A) A schematic model of putative TORC2 signaling in C. elegans. (B) Salt chemotaxis of wild-type N2 worms and mutants of TORC2 components. TORC2-signaling mutants migrated to lower salt levels than did wild-type worms. (C) Genomic structure of sinh-1. Pink boxes indicate exons. The location of the pe420 deletion is shown below the sinh-1 locus. (D) Calcium responses of ASER of wild-type animals and sinh-1 mutants to an NaCl down-step from 50 to 25 mM and back to 50 mM (traces represent the means, the shadings the s.e.m.) after 5-h low-salt/food(–) conditioning. (E) Bar graphs showing averaged yellow/cyan fluorescence ratios during 10 s after downstep stimulation. Error bars, s.e.m.; n.s. = not significant (Student’s t test, N ≥ 9). (F) Benzaldehyde chemotaxis of wild-type N2 and mutant worms after benzaldehyde conditioning or in the naive state. rict-1(ft7), sinh-1(pe420), and sgk-1(ok538) mutants, but not pkc-2(ok328) and rsks-1(ok1255) mutants, showed defects in food-odor associative learning. Error bars, s.e.m.; **p < 0.01, n.s. = not significant (wild-type vs. each mutant, one-way ANOVA with Dunnett’s post hoc test, N ≥ 9).

Fig 4. Effects of rapamycin on taste associative learning of CeTOR signaling mutants.

(A and B) Salt chemotaxis of rapamycin-treated worms after starvation conditioning on low (A) or high (B) salt-containing agar plates; 100 μM rapamycin was used. Error bars, s.e.m.; *p < 0.05, **p < 0.01, n.s. = not significant (EtOH vs. rapamycin, Student’s t test, N ≥ 6).

Fig 5. Putative TORC2 substrates akt-1, pkc-2, and sgk-1 are involved in taste associative learning.

(A) sinh-1 and akt-1 mutations have additive effects on salt chemotaxis after low-salt/food(–) conditioning. (B) akt-1(mg144gf) does not suppress the chemotaxis defect of sinh-1(pe420) after low-salt/food(–) conditioning. (C and D) Salt chemotaxis of the mutants of the putative TORC2 substrates sgk-1 (C) and pkc-2 (D). sgk-1 mutants showed abnormal migration to low salt levels after 25 mM salt/food(–) and 100 mM salt/food(+) conditioning (C), whereas pkc-2 mutants showed migration to low salt levels only after 25 mM salt/food(–) conditioning (D). Error bars, s.e.m.; *p < 0.05, **p < 0.01 (wild-type vs. each mutant, one-way ANOVA with Dunnett’s post hoc test, N ≥ 6).

Fig 6. TORC2 and PKC-2 function in the same pathway and SGK-1 acts downstream of TORC2.

(A) sinh-1(pe420) and sgk-1(ok538) double mutants do not show an additive effect on chemotaxis after low-salt/food(–) conditioning when compared to single mutants. (B) sgk-1(ft15gf) suppresses the sinh-1(pe420) chemotaxis defect after low-salt/food(–) conditioning. (C) sinh-1(pe420) and pkc-2(ok328) mutations do not produce an additive effect on salt chemotaxis after low-salt/food(–) conditioning. (D) daf-16(mgDf47) and sgk-1(ok538) have an additive effect on chemotaxis after low-salt/food(–) conditioning. (E) sgk-1(ft15gf) suppresses the chemotaxis defect of daf-16(mgDf47) after low-salt/food(–) conditioning. Error bars, s.e.m.; **p < 0.01, n.s. = not significant (one-way ANOVA with Dunnett’s post hoc test, N ≥ 6).

Fig 7. TORC2/PKC-2 pathway functions in neurons, whereas TORC2/SGK-1 pathway functions in the intestine.

(A) Neuronal (unc-14 promoter) and intestinal (ges-1 promoter) expression of sinh-1 cDNA rescued the chemotaxis defects of the sinh-1(pe420) mutant after low-salt/food(–) conditioning. (B) Pan-neuronal expression (H20 promoter) and ASER-specific expression (gcy-5 promoter), but not intestine-specific expression (ges-1 promoter), of the pkc-2 cDNA rescued the chemotaxis abnormality of the pkc-2(ok328) mutant after low-salt/food(–) conditioning. (C) Intestine-specific expression (sgk-1B promoter or ges-1 promoter), but not pan-neuronal expression (H20 promoter), of sgk-1 isoform B was sufficient for rescuing the chemotaxis defect of the sgk-1(ok538) mutant after low-salt/food(–) conditioning. (D) Intestine-specific expression of the active form of sgk-1 suppressed the chemotaxis defect of the sinh-1(pe420) mutant after low-salt/food(–) conditioning. Error bars, s.e.m.; *p < 0.05, **p < 0.01, n.s. = not significant (Dunnett’s test, N ≥ 9).