Enhancing S-adenosyl-methionine catabolism extends Drosophila lifespan

Abstract

Methionine restriction extends the lifespan of various model organisms. Limiting S-adenosyl-methionine (SAM) synthesis, the first metabolic reaction of dietary methionine, extends longevity in Caenorhabditis elegans but accelerates pathology in mammals. Here, we show that, as an alternative to inhibiting SAM synthesis, enhancement of SAM catabolism by glycine N-methyltransferase (Gnmt) extends the lifespan in Drosophila. Gnmt strongly buffers systemic SAM levels by producing sarcosine in either high-methionine or low-sams conditions. During ageing, systemic SAM levels in flies are increased. Gnmt is transcriptionally induced in a dFoxO-dependent manner; however, this is insufficient to suppress SAM elevation completely in old flies. Overexpression of gnmt suppresses this age-dependent SAM increase and extends longevity. Pro-longevity regimens, such as dietary restriction or reduced insulin signalling, attenuate the age-dependent SAM increase, and rely at least partially on Gnmt function to exert their lifespan-extending effect in Drosophila. Our study suggests that regulation of SAM levels by Gnmt is a key component of lifespan extension.

Figure 1. Gnmt is a predominant regulator of systemic SAM levels.

(a) Schematic view of methionine metabolism in Drosophila melanogaster. Gly, glycine; Sar, sarcosine; Met, methionine; SAM, S-adenosyl-methionine; SAH, S-adenosyl-homocysteine; Hcy, homocysteine; Sams, S-adenosyl-methionine synthase; Gnmt, glycine N-methyltransferase. (b) Lifespan analysis of ubiquitous sams-RNAi male flies under 10% SY diet. In total, 200 μM RU486 (RU) is used to knock down sams after adult eclosion in Tub^GS>sams-RNAi. Statistics, log-rank test, P<0.0001 (N=70 for −RU, N=79 for +RU). (c) Western blot analysis of Sams and Gnmt in day-5 male flies with either sams-RNAi or gnmt-RNAi driven by the fat-body drivers: r⁴-Gal4 and FB-Gal4, or no driver. + Indicates UAS-only or Gal4-only controls. (d–l) UPLC–MS/MS analysis of SAM, methionine (Met) and sarcosine (Sar) levels in day-5 male flies with either sams-RNAi or gnmt-RNAi driven by the fat-body drivers (FB-Gal4, r⁴-Gal4) or no driver. + Indicates UAS-only or Gal4-only controls. Error bars represent mean±s.d. (N=4). Statistics, one-way analysis of variance with Bonferroni's multiple comparison test. *P<0.05, **P<0.01, ***P<0.001 from the biological replicates. NS, not significant.

Figure 2. The effect of overexpression of Sams and Gnmt on metabolite levels.

(a–i) UPLC–MS/MS analysis of SAM, Met and Sar levels in day-5 male flies with sams, sams^R275H, gnmt and gnmt^S145A overexpression by ubiquitous (da-Gal4) or fat-body (r⁴-Gal4) drivers or no driver. + Indicates UAS-only or Gal4-only controls. Statistics, two-tailed unpaired t-test. Error bars represent mean±s.d. (N=5 for da-Gal4, N=3 for r⁴-Gal4, N=4 for no-driver control). *P<0.05, **P<0.01, ***P<0.001 from the biological replicates. NS, not significant.

Figure 3. Gnmt is necessary for lifespan extension by reduced insulin signalling or dietary restriction.

(a) Lifespan analysis of Gnmt-overexpressing male flies by Tub^GS-Gal4. Statistics, log-rank test, gnmt versus gnmt+RU, P<0.0001 (N=117 for gnmt, N=118 for gnmt+RU). gnmt^S145A versus gnmt^S145A+RU, P=0.7732 (N=111 for gnmt^S145A, N=113 for gnmt^S145A+RU). (b,d,f,j,l) Average lifespan. One-way analysis of variance (ANOVA) with Bonferroni's multiple comparison test was used for statistical analysis. (c) Lifespan analysis of gnmt-overexpressing male and female flies by Tub^GS-Gal4. UAS-gnmt integrated on attP2 site is used. Statistics, log-rank test, gnmt-2 male versus gnmt-2+RU male, P<0.0001 (N=116 for each). gnmt-2 female versus gnmt-2+RU female, P<0.0001 (N=116 for each). (e) Lifespan analysis of male flies overexpressing InR^DN (dominant-negative form of the insulin receptor), gnmt-RNAi^sh or both. Statistics, log-rank test, InR^DN versus InR^DN+RU, P<0.0001 (N=158 for InR^DN, N=154 for InR^DN+RU). gnmt-RNAi^sh versus gnmt-RNAi^sh+RU, P=0.0478 (N=80 for both). InR^DN, gnmt-RNAi^sh versus InR^DN, gnmt-RNAi^sh+RU, P<0.0001 (N=142 for InR^DN, gnmt-RNAi^sh, N=154 for InR^DN, gnmt-RNAi^sh+RU). (g) qRT–PCR analysis of InR in day-7 adult male flies treated RU486 for 5 days. Error bars represent mean±s.e.m. (N=4). Statistics, one–way ANOVA with Bonferroni's multiple comparison test. (h) Lifespan analysis of male flies overexpressing sams. Statistics, log-rank test, sams versus sams+RU, P=0.6381 (N=135 for sams, N=143 for sams+RU). Inset, average lifespan. (i) Lifespan analysis of male gnmt^Mi flies under dietary restriction. Statistics, log-rank test, yw 20 versus 5%, P<0.0001 (N=134 for 20%, N=132 for 5%). gnmt^Mi 20 versus 5%, P=0.5904 (N=133 for 20%, 129 for 5%). (k) Lifespan analysis of female gnmt^Mi flies under dietary restriction. Statistics, log-rank test, yw 20 versus 5%, P<0.0001 (N=140 for 20%, N=133 for 5%). gnmt^Mi 20 versus 5%, P=0.9107 (N=134 for 20%, 136 for 5%). *P<0.05, **P<0.01, ***P<0.001 from the biological replicates. NS, not significant.

Figure 4. SAM levels increased during ageing in spite of induced gnmt expression by dFoxO in fat body.

(a–d) UPLC–MS/MS analysis of SAM and Met levels in wild-type male flies during ageing. Error bars represent mean±s.d. (N=4). Statistics, one-way analysis of variance (ANOVA) with Bonferroni's multiple comparison test. (e) UPLC–MS/MS analysis of sarcosine levels in male w¹¹¹⁸ during ageing. Error bars represent mean±s.d. (N=3). Statistics, one-way ANOVA with Bonferroni's multiple comparison test. (f) Western blot analysis of Sardh and Gnmt during ageing in male w¹¹¹⁸. (g,h) qRT–PCR analysis of gnmt and sardh in young (1 week old) and old (7 weeks old) wild-type male flies. Error bars represent mean±s.e.m. (N=5 for 1 w w¹¹¹⁸, N=6 for 7 w w¹¹¹⁸, N=3 for Canton S). Statistics, two-tailed unpaired t-test 1 versus 7 w. (i) UPLC–MS/MS analysis of sarcosine levels in young and old male flies with lacZ-RNAi or sardh-RNAi. Error bars represent mean±s.d. (N=4). Statistics, One-way ANOVA with Bonferroni's multiple comparison test. (j) Western blot analysis of Gnmt in dFoxO knockdown flies.+ Indicates Gal4-only controls. (k) UPLC–MS/MS analysis of SAM levels in young and old male flies with lacZ-RNAi, dFoxO-RNAi, gnmt or gnmt^S145A. Error bars represent mean±s.d. (N=4). Statistics, one-way ANOVA with Bonferroni's multiple comparison test. w, week. *P<0.05, **P<0.01, ***P<0.001 from the biological replicates.

Figure 5. Age-dependent SAM increase is attenuated by lifespan-extending regimens.

(a,b) UPLC–MS/MS analysis of SAM and Met levels in yw and gnmt^Mi young (1 weeks old) or old (5 weeks old) male flies maintained under dietary restriction or a nutrient-rich condition. Error bars represent mean±s.d. (N=3–4). Statistics: one-way analysis of variance (ANOVA) with Bonferroni's multiple comparison test. (c–e) UPLC–MS/MS analysis of SAM levels in lacZ-, InR^DNor TOR^TED-overexpressing male flies at young (1 week old) or old (5 weeks old) stages. Error bars represent mean±s.d. (N=3–4). Statistics: one-way ANOVA with Bonferroni's multiple comparison test. (f) Schematic view of the relationship between SAM metabolism and longevity control. Compared with young flies, SAM levels are increased despite Gnmt induction in old flies. Strengthening the dFoxO–Gnmt axis rescues age-dependent SAM increases and extends lifespan. *P<0.05, **P<0.01, ***P<0.001 from the biological replicates. NS, not significant.