Endogenous hydrogen sulfide production is essential for dietary restriction benefits

Abstract

Dietary restriction (DR) without malnutrition encompasses numerous regimens with overlapping benefits including longevity and stress resistance, but unifying nutritional and molecular mechanisms remain elusive. In a mouse model of DR-mediated stress resistance, we found that sulfur amino acid (SAA) restriction increased expression of the transsulfuration pathway (TSP) enzyme cystathionine γ-lyase (CGL), resulting in increased hydrogen sulfide (H2S) production and protection from hepatic ischemia reperfusion injury. SAA supplementation, mTORC1 activation, or chemical/genetic CGL inhibition reduced H2S production and blocked DR-mediated stress resistance. In vitro, the mitochondrial protein SQR was required for H2S-mediated protection during nutrient/oxygen deprivation. Finally, TSP-dependent H2S production was observed in yeast, worm, fruit fly, and rodent models of DR-mediated longevity. Together, these data are consistent with evolutionary conservation of TSP-mediated H2S as a mediator of DR benefits with broad implications for clinical translation. PAPERFLICK:

Conflict of interest statement

The authors claim no conflicts of interest.

Copyright © 2015 Elsevier Inc. All rights reserved.

Figures

Figure 1. NAC abrogates benefits of DR against acute stress independent of NRF2

(A–H) Effects of 1wk of 50% DR vs. AL feeding +/− NAC as indicated on body weight (n=15–17/group, A); %fat mass (n=5/group, B); serum triglycerides (n=4–7/group, C); hepatic FAO-associated gene expression (n=3/group, D); peroxisomal FAO capacity in liver extracts (n=3/group, E); hepatic RONS (n=6–8/group, F); hepatic NRF2 target gene expression (n=4/group, G); and hepatic GSH levels (n=7/group, H). (I, J) Combined serum markers of liver damage normalized to the average of the AL control group (I) and liver pathology from injured left liver lobes on the microscopic level stained with hematoxylin and eosin (250μm scale bar, above) and the macroscopic level (1x magnification, below) showing fixed tissue 24hrs after reperfusion (J) in mice (n=5–6/group) preconditioned as indicated before hepatic IRI or mock injury. Asterisk indicates the significance of the difference between AL and DR, and pound sign between DR and DR+NAC; */#p<0.05. (K) Serum markers of liver damage in WT (n=5–6/group) or NRF2KO (n=7–12/group) preconditioned as indicated. Asterisk indicates the significance of the difference between AL and DR within genotype; *p<0.05. See also Suppl. Fig. 1.

Figure 2. Sulfur amino acids control the benefits of DR and PR

(A, B) Serum markers of liver damage (A) and liver pathology after reperfusion (B) in mice (n=5/group) preconditioned on complete diets fed AL or 50% DR +/− supplementation with vitamins C&E or Met&Cys as indicated. Asterisk indicates the significance of the difference vs. AL and pound sign vs. DR+2xMet&Cys; */#p<0.05. (C, D) Serum markers of liver damage (C) and liver pathology after reperfusion (D) in mice (n=5/group) preconditioned on complete (Comp) or protein free (Prot. Free) diets fed AL or 35% DR with Met&Cys addition as indicated prior to hepatic IRI. Asterisk indicates the significance of the difference vs. AL complete and pound sign vs. DR Prot. Free+Met&Cys; */# p<0.05. (E, F) Serum markers of liver damage (E) and liver pathology after reperfusion (F) in mice (n=5/group) preconditioned on complete diets fed AL or 50% DR with 2xCys added as indicated. Asterisk indicates the significance of the difference vs. AL and pound sign vs. DR+2xCys; */#p<0.05. See also Suppl. Fig. 2.

Figure 3. DR stimulates endogenous H2S production via the TSP

(A) Model of the transmethylation and TSP. Arrows trace sulfur from Met to Cys and downstream cellular processes via Cystathionine Beta-Synthase (CBS) and Cystathionine Gamma-Lyase (CGL). Metabolites in green (taurine, GSH and H2S) have demonstrated potential to protect against IRI. MAT: methionine adenosyl transferase, SAM: S-Adenosylmethionine, SAH: S-Adenosylhomocysteine, SAHH: S-adenosylhomocysteine hydrolase, MS: Methionine synthase, BHMT: Betaine homocysteine methyltransferase. (B) Hepatic CBS and CGL gene expression upon DR +/− 2xMet&Cys or NAC as indicated; n=3/group. (C) Immunoblots of CGL, CBS and actin in liver extracts from AL or DR mice as indicated; NS, non-specific protein band. (D) Liver metabolites in WT or LTsc1KO mice fed 35% DR on a protein free diet relative to the AL fed complete diet group; n=5/group. (E, F) H2S production capacity from liver extracts of AL or DR mice (E) +/− 2xMet&Cys (F) as detected by the black precipitate, lead sulfide. (G) H2S production capacity as in (E), but using a micro-sulfide probe to detect H2S. (H) Endogenous hepatic H2S in mice on the indicated diet using a micro-sulfide probe inserted into liver lobes; n=3 mice/group, 2 lobes/mouse. Asterisk indicates the significance of the difference vs. AL; *p<0.05. (I, J) Immunoblots of liver CGL (I) and H2S production (J) in WT and LTsc1KO mice fed AL or 35% DR on a protein free diet. See also Suppl. Fig. 3.

Figure 4. H2S is necessary and sufficient to confer DR benefits against hepatic IRI

(A, B) Serum markers of liver damage (A) and liver pathology after reperfusion (B) in mice (n=4–5/group) treated with vehicle or H2S 30 min prior to surgery. Asterisk indicates the significance of the difference vs. AL; *p<0.05. (C, D) Serum markers of liver damage (C) and liver pathology after reperfusion (D) in mice (n=3–4/group) treated with PAG during AL or DR preconditioning with a single H2S injection prior to surgery as indicated. Asterisk indicates the significance of the difference vs. AL and pound sign vs. DR+PAG; */#p<0.05. (E) Immunoblot of CGL in liver extracts of WT or CGLKO mice as indicated. NS, non-specific protein. (F) H2S production capacity of liver extracts from WT or CGLKO mice preconditioned as indicated. (G, H) Serum markers of liver injury (G) and liver pathology after reperfusion (H) in WT or CGLKO mice (n=5–8/group) preconditioned as indicated. Asterisk indicates the significance of the difference vs. WT AL and pound sign vs. CGLKO DR; */#p<0.05. (I) Immunoblot of CGL in liver extracts from WT mice infected with control (Ad-Null) or CGL-overexpressing (Ad-CGL) adenovirus as indicated; NS LC, non-specific loading control protein. (J) H2S production capacity of liver extracts of Ad-Null or Ad-CGL infected mice. (K, L) Serum markers of liver damage (K) and liver pathology after reperfusion (L) in Ad-Null or Ad-CGL infected mice (n=6/group). Asterisk indicates the significance of the difference vs. Ad-Null; *p<0.05. See also Suppl. Fig. 4.

Figure 5. Mitochondrial SQR is required for cytoprotective effects of H2S during ischemia

(A, B) Cell autonomous increase in TSP enzymes and H2S production. (A) CBS and CGL gene expression in Hepa1-6 cells cultured overnight in complete or -Met&Cys media. Asterisk indicates the significance of the difference vs. complete; *p< 0.05. (B) H2S production in live Hepa1-6 cells preconditioned in complete or -Met&Cys media in quadruplicate for 16–24hrs as indicated before readdition of Cys and PLP for H2S detection. (C, D) Cell autonomous effects of exogenous H2S on simulated IRI in Hepa1-6 cells in vitro; H2S was added during the ischemic phase and removed upon simulated reperfusion. (C) LDH release during the ischemic phase consisting of 3hr incubation in saline (simulated nutrient/energy deprivation) under normoxic or hypoxic (simulated ischemia) conditions +/− exogenous H2S as indicated. (D) MTT activity during the reperfusion phase consisting of readdition of complete media under normoxic conditions (simulated reperfusion). (E, F) Cell autonomous effects of overnight Met&Cys withdrawal (preconditioning phase conditions) on simulated IRI in primary hepatocyes in vitro. LDH release (E) during simulated ischemia and MTT activity (F) upon simulated reperfusion. (G) Schematic of H2S oxidation to thiosulfate by SQR in mitochondria. Q, Coenzyme Q; IMS, inner membrane space; I and II, Complex I and II of the mitochondrial ETC; SD, Sulfur Dioxygenase; ST, Sulfur Transferase; S2O3−2, thiosulfate. (H–J) Effects of exogenous H2S on simulated IRI upon SQR KD in Hepa1-6 cells. LDH release during the ischemic phase (H) in saline under the indicated normoxic or hypoxic conditions +/− SQR KD, and during the reperfusion phase (I) in complete media expressed relative to respective group not receiving H2S during the ischemic phase; and MTT activity (J) during the reperfusion phase. Asterisk indicates the significance of the difference in the indicated group +/− H2S treatment; hash mark indicates the significance of the difference between control and SQR KD within a given treatment group; */#p<0.05. (K) Cell autonomous effects of exogenous thiosulfate on MTT activity during the reperfusion phase following simulated IRI in Hepa1-6 cells +/− SQR KD. Asterisk indicates the significance of the difference in the indicated group +/− thiosulfate treatment; hash mark indicates the significance of the difference between control and SQR KD within a given treatment group; */#p<0.05. See also Suppl. Fig. 5.

Figure 6. Increased endogenous H2S production in diet-induced longevity extension models crosses evolutionary boundaries

(A–B) H2S production capacity of liver and kidney extracts from mice with AL access to complete or MetR diets as indicated for 4mos (A) and in mice with AL, every-other-day (EOD) fasting or 20–30% restricted diets for 6wks (B). Each circle represents H2S production from an individual animal, normalized for extract protein content. (C) H2S production capacity of whole fly lysates normalized for protein content from populations subject to the indicated dietary amino acid (AA) and Met concentrations. Maximal H2S production capacity correlated with the optimal diet for longevity extension (0.4x AA with 0.15mM Met) (Lee et al., 2014). (D) H2S production capacity of N2 and eat-2 whole worm lysates in quadruplicate. (E) Lifespan of N2 and eat-2 worms subject to RNAi-mediated KD of the TSP genes cbs-1 or cbl-1 as indicated. (F) Lifespan analysis of WT or transgenic worms overexpressing CBS-1. (G) Chronological lifespan of budding yeast grown in 2% or 0.5% glucose as indicated. (H) H2S production in yeast cultures grown in 2% or 0.5% glucose as detected by black lead sulfide accumulation on lead acetate papers inserted into the caps of the growth flasks and removed at the indicated time point. (I) Schematic of transmethylation, TSP and H2S production in budding yeast. (J) H2S production in WT and sulfur assimilatory pathway mutant strains grown in 2% or 0.5% glucose for the indicated time. (K) H2S production in the met14 sulfur assimilatory mutant grown in 0.5% glucose and the indicated relative amount of Met or Cys. (L) Chronological lifespan in 2% glucose +/− H2S donors NaHS and GYY4137 added during early culture growth at indicated time points (arrows). See also Suppl. Fig. 6.

Figure 7. Model of DR-mediated benefits through increased H2S production

DR regimens leading to extended longevity across evolutionary boundaries include PR, SAA restriction and Cys restriction, leading to increased TSP activity via CGL and/or CBS family members, and increased endogenous H2S production. Specific SAA addition, increased activity of the nutrient/energy sensing kinase, mechanistic target of rapamycin (mTOR), or the pharmacological agent, PAG, can block TSP upregulation and H2S production. H2S can readily diffuse through tissues and has pleiotropic, overlapping beneficial effects on the cellular, tissue and organismal levels with the potential to contribute to stress resistance and longevity benefits of DR.