Regulation of the one carbon folate cycle as a shared metabolic signature of longevity

Abstract

The metabolome represents a complex network of biological events that reflects the physiologic state of the organism in health and disease. Additionally, specific metabolites and metabolic signaling pathways have been shown to modulate animal ageing, but whether there are convergent mechanisms uniting these processes remains elusive. Here, we used high resolution mass spectrometry to obtain the metabolomic profiles of canonical longevity pathways in C. elegans to identify metabolites regulating life span. By leveraging the metabolomic profiles across pathways, we found that one carbon metabolism and the folate cycle are pervasively regulated in common. We observed similar changes in long-lived mouse models of reduced insulin/IGF signaling. Genetic manipulation of pathway enzymes and supplementation with one carbon metabolites in C. elegans reveal that regulation of the folate cycle represents a shared causal mechanism of longevity and proteoprotection. Such interventions impact the methionine cycle, and reveal methionine restriction as an underlying mechanism. This comparative approach reveals key metabolic nodes to enhance healthy ageing.

Fig. 1. Regulation of the folic acid cycle is altered in longevity mutants.

a Untargeted metabolomic analysis of daf-2(e1370), eat-2(ad465), isp-1(qm150), and glp-1(e2141)ts worms at day 1 of adulthood. glp-1(e2141)ts worms are compared with wild-type (WT) control, which undergoes the same 25 °C thermal shift as the mutant. Heat map containing all biological replicates, indicates the relative abundance of metabolite concentrations relative to the wild-type average, including both significant and non-significant changes, (listed in Supplementary Data 1). Metabolites are manually grouped into different functional categories. b Venn diagram of the significantly changed metabolites (adj p < 0.05) for each genotype showing unique and overlapping compounds. c Metabolic-protein network of unknown and known features created by the PIUMet algorithm (http://fraenkel-nsf.csbi.mit.edu/piumet2/). The degree of confidence of the PPI algorithm is represented by node diameters. Additional parameters are found in Supplementary Table 5. Dotted circle indicates the region of the network chosen for further investigation. Abbreviations in the chart: SNF-5 (Sodium: Neurotransmitter symporter Family, orthologous to SLC6A8), BTD (biotinidase), MCT-1 (Mono carboxylate Transporter family, orthologous to SLC16A14), BCAT-1 (branched amino acids transporter), DLD (dihydrolipoamide dehydrogenase), FPGS (folylpolyglutamate synthase), TYMS-1 (thymidylate synthetase), MEL-32 (orthologue of SHMT1 serine hydroxymethyl transferase 1) THK-1 (thymidine kinase-1). d Quantitation of folic acid intermediates using targeted metabolomic analysis in longevity mutants (day 1 adult). FA, THF, and ML accumulate in three out of four longevity mutants. 5MTHF significantly decreases in three longevity mutants. Abbreviations in the chart: FA (folic acid), DHF (dihydrofolic acid), THF (tetrahydrofolic acid), 5MTHF (5‐methyl‐tetrahydrofolic acid), ML (5,10‐methylene‐tetrahydrofolic acid), MN (5,10‐methenyl‐tetrahydrofolic acid), FO (formyl‐tetrahydrofolic acid). a, d N = 5 independent biological replicates. a, d Normalized metabolite concentrations are converted to log2 for heat map generation. a, b Statistics were performed using one-sided Fisher test and Benjamini–Hochberg correction for multiple comparisons (adj p < 0.05) and d using one-way-ANOVA and Dunnett’s multiple comparison *p < 0.5, **p < 0.01, ***p < 0.001 (Supplementary Table 6, for statistics).

Fig. 2. dhfr-1i prolongs nematode life span in a 5MTHF dependent manner.

adhfr-1 and tyms-1 RNAi treatment increase wild-type life span (from L4 stage). b Supplementation of 10 nM 5MTHF abolishes dhfr-1i longevity (from L4 stage). c Quantitation of folic acid intermediates using targeted mass spectrometry in wild-type worms (day 1) with luci and dhfr-1i treatment in the presence or absence of folic acid and 5MTHF. dhfr-1i increases folic acid and decreases 5MTHF and downstream intermediates. d Increasing concentrations of FA inhibit ¹³C¹⁵N labeled 5MTHF incorporation within a 2 h time period (day 1 adult). e Incorporation of ¹³C¹⁵N labeled 5MTHF over time in worms with luci (black line) and dhfr-1i (orange line) (day 1 adult) each dot represents a single biological replicate. f Thrashing assay of polyQ35 worms (day 7 adult). g Protein aggregate quantitation in polyQ40 model (day 7 adult). dhfr-1i or FA supplementation are beneficial, whereas 5MTHF is detrimental for motility and aggregate accumulation. Each dot represents a single worm. a, b n = 150 worms per repeat per condition, N = 3 biological replicates. c N = 5 biological replicates. d, e N = 4 independent biological replicates. f, g n = 30 worms, N = 3 independent biological replicates, only one biological replicate is shown. a, b Statistics were performed with the two-sided Mantel–Cox log-rank test (Supplementary Table 7 for statistics). c, f, g Significance was assessed using one-way ANOVA and Dunnett’s multiple comparisons test. *p < 0.5, **p < 0.01, ***p < 0.001. Data are presented as mean ± S.D. (Supplementary Table 8 for statistics).

Fig. 3. dhfr-1i affects the methionine cycle and mimics methionine restriction.

a Schematic of the methionine cycle. b, c Quantitation of methionine cycle intermediates using targeted mass spectrometry in dhfr-1i and luci, with or without FA and 5MTHF supplementation (day 1 adult). dhfr-1i decreases methionine and S-adenosyl methionine levels and increases homocysteine and adenosyl-homocysteine levels. This increase is reversed by supplementation with 5MTHF. d dhfr-1i and metr-1i. show similar regulation of mRNA expression of selected genes implicated in methionine restriction (dgn-1, let-23, ign-3, dapk-1, and daf-18), day 1 adults. e Methionine restriction specific gene signature is increased by dhfr-1i and reversed by 5MTHF supplementation, day 1 adults. f Methionine supplementation (20 mM, 40 mM) suppresses dhfr-1i longevity. b, c N = 5 independent biological replicates. d, e N = 3 independent biological replicates. f n = 150 worms per repeat per condition, N = 3 biological replicates. b–e Significance was assessed using one-way anova with Dunnett’s multiple comparisons test *p < 0.5, **p < 0.01, ***p < 0.001. f Statistics were performed with the two-sided Mantel–Cox log-rank test. *p < 0.5, **p < 0.01, ***p < 0.001. Data are presented as mean ± S.D. (Supplementary Table 8 for statistics).

Fig. 4. dhfr-1 and 5MTHF act within insulin/IGF and mitochondrial signaling pathways.

a Life span of daf-2(e1370) supplemented with 10 nM 5MTHF (from L4 stage). b Life span of isp-1(qm150) supplemented with 10 nM 5MTHF (from L4 stage). 5MTHF does not alter WT life span but reduces daf-2 and isp-1 longevity. c, d Relative mRNA expression of folate cycle genes in daf-2(e1370) and isp-1(qm150) (day 1 adult). dhfr-1 mRNA expression is lower in both genotypes. e Life span experiment of wild-type and daf-16(mgDf50) with luci and dhfr-1i in the presence or absence of 5MTHF (from L4 stage). dhfr-1i extends daf-16 mutant life span and supplementation with 5MTHF abolishes this extension. f Relative mRNA expression of folate cycle genes in daf-16 and daf-16,daf-2 backgrounds. daf-16 mutation has little or no effect on dhfr-1 or mthf-1 mRNA expression in the daf-2 background. a, b, e n = 150 per repeat per condition, N = 3 independent biological replicates. c, d, f N = 3 biological replicates. a, b, e Statistics were analyzed by the two-sided Mantel–Cox log-rank test (Supplementary Table 7 for statistics). c, d, f Significance was assessed using one-way ANOVA and Dunnett’s multiple comparisons test. *p < 0.5, **p < 0.01, ***p < 0.001. Data are presented as mean ± S.D. (Supplementary Table 6 for statistics).

Fig. 5. Reduced insulin/IGF signaling alters folic acid and methionine cycle intermediates.

a Relative abundance of methionine cycle intermediates in daf-2 mutants and WT animals using targeted metabolomics. b Relative mRNA expression level of methionine restriction gene signature in daf-2 and daf-2;daf-16 backgrounds. dgn-1, ing-3, dapk-1, and daf-18 are upregulated and let-23 and ing-3 downregulated in the daf-2 background, independent of daf-16. c, e Targeted metabolomics of folic acid intermediates in brain and liver in wild-type and Irs1^−/− full body knockout mice. Folic acid and THF increase in Irs1^−/−, whereas 5MTHF decreases in the mutant. d, f Quantitation of methionine cycle intermediates in brain and liver in wild-type and Irs1^−/− knockout mice. Methionine and SAM decrease and homocysteine increases in the mutant. a N = 4 independent biological replicates. b N = 3 independent biological replicates, c–f N = 5 mice, each dot represents a single animal. a–f Significance was assessed using one-way ANOVA Dunnett’s multiple comparisons test. *p < 0.5, **p < 0.01, ***p < 0.001. Data are presented as mean ± S.D. (Supplementary Table 6 for statistics).