When the consumption of the essential amino acid methionine is limited, the health span and life span extension of rodents is observed, similar to interventions reducing caloric intake. Ames dwarf mice live longer than their wild-type siblings when fed ad libitum rodent chow, and further extensions in life span are observed when this chow is reduced by 30% [5, 7]. However, in a recent report, we showed that dietary MET restriction did not impact life span in these GH-deficient mice as well as growth hormone receptor knockout mice suggesting that GH signaling is necessary to discriminate amino acid intake [29]. Indeed, the reductions of 50% and 80% MET as well as 150% supplementation did not alter life span in these mice nor was body weight influenced, contrary to several other reports in normal rodents.

In the current study, we fed Ames and wild-type mice different levels of MET and examined the corresponding metabolic pathways of MET, GSH, and redoxin to better understand the relationships between GH deficiency and amino acid metabolism. We focused on liver tissue as it is the primary organ involved in MET metabolism as well as GSH biosynthesis, detoxification, and stress resistance. We also evaluated kidney as this tissue is responsible for GSH degradation and detoxification of endogenous and exogenous substances and has an active transsulfuration system. Skeletal muscle served as a post-mitotic tissue with limited roles in each of these systems. The results of these two tissues are provided in Additional files 5, 6, 7, 8, 9, and 10 for tissue response comparison. The levels of MET were chosen based on previous literature showing that an 80% reduction extends life span (0.16%) in rodents. The 0.43% MET level represented a 50% reduction in typical rodent chow and a level that could be applied in human interventional studies. We also utilized supplementary MET (50% above typical levels of rodent diet) but incorporated a level well below that shown to be toxic (1.3%).

The body weights of dwarf mice were unaffected by dietary MET, similar to our previous report that examined lifetime body weight on these diets. The body and liver weight responses were reflected in the plasma insulin-like growth factor 1 (IGF1) levels that were shown to decrease in wild-type mice only as dietary MET levels decreased [29]. Other investigations have shown that body weights were either lower or unchanged by dietary MET restriction [13, 22].

In general, this study demonstrated that short-term diets containing restricted, low, or enriched MET produce vastly different liver metabolomic, mRNA, and protein profiles in wild-type mice. Dwarf mice, however, demonstrated fewer statistically significant differences between diets indicating that these mice do not respond in a similar manner to alterations in MET intake. This may reflect the differential MET metabolism in GH-deficient dwarf mice due to the inability to signal through the GH pathway. This lack of responsiveness to the diets on metabolite, gene, and protein levels is reflected in the lack of a difference in life span when dwarf mice are fed these diets throughout life (8 weeks to death) [29].

Dwarf mice did not alter liver MET or cysteine levels in response to different levels of MET intake nor were homocysteine levels changed in these long-living mice. Wild-type mice generated higher levels of MET, cysteine, and homocysteine on MET restriction compared to the MET-enriched diet (1.3%) and in comparison to dwarf mice (0.16%). This finding suggests the conservation of sulfur AA possibly through recycling as the genotype and diet differences in MET are reflected in the folate pathway (folate, THF, 5-methyl THF). In contrast, MET restriction decreased sulfur amino acids in the serum and liver of rats inducing hyperhomocysteinemia [30, 31]. Wild-type mouse levels of the sulfur AA did not differ from dwarf mice when fed a MET-enriched diet even though all of the measured MET metabolic enzymes differed significantly in dwarf mice fed high MET. The dwarf and its wild-type counterpart respond differently to MET restriction when compared to one another and to rats. In response to restricted MET, folate or one-carbon metabolism in wild-type mice was elevated resulting in increased levels of MET compared to higher MET diets and to MET restriction in dwarf mice. One report showed that post-transcriptional downregulation of Cbs occurs when animals are fed MET-deficient diets (no MET) indicating that a specific sensing system in mice determines when MET intake is low and signals the liver to activate a reduction in Cbs to conserve MET [32].

SAM is an important precursor for three key biochemical pathways: methylation, transsulfuration, and polyamine synthesis. Mat catalyzes the formation of SAM from MET and is upregulated in dwarf mice independent of diet agreeing with our earlier reports in these mice fed typical rodent chow [24, 25]. MET restriction elevated SAM in both genotypes regardless of the level of the Mat enzyme in the liver. Genotype differences in downstream methyltransferases resulted in elevated SAH levels in dwarf mice. SAH is metabolized to homocysteine and adenosine via Ahcy, an enzyme that remains elevated in Ames mice on both diets when compared to wild-type mice. The low homocysteine levels in dwarf mice were maintained, yet corresponding adenosine levels were elevated. These observations suggest that the actions of enhanced Ahcy increased the levels of both products in dwarf mice but that homocysteine was likely rapidly utilized via transsulfuration as evidenced by the elevated levels of downstream Cbs and Cth. We have previously shown that MET flux through transsulfuration in dwarf mouse liver occurs at three times the rate of wild-type mice [25]. Also, the activity of Ahcy is regulated in part by GH; thus, in the absence of this hormone, the metabolic pathway of homocysteine may be dysregulated [33]. The MET-conserving enzymes (MatI, Ahcy, Bhmt, Mtr, Mthfr) tend to be inhibited by their products and other metabolites, and rats fed ample MET exhibit reduced hepatic levels of these enzymes. In dwarf mice, the opposite effect is observed with high levels of each enzyme when consuming the MET-enriched diet. Wild-type mice appear to respond to the normal feedback inhibition of the MET-conserving enzymes. The MET-catabolizing enzymes, Gnmt, Cbs, and Cth, are normally activated by the cycle metabolites, and liver levels increase with additional MET intake. In our study, these enzymes were higher in dwarf mice consuming the MET-enriched diet, and Cth was also higher in the MET-restricted dwarf when compared to their wild-type counterparts. Ahcy has been categorized both as a MET-conserving and MET-catabolizing enzyme as it has intermediate properties. Under normal circumstances, the thermodynamics of this reaction ensure the conversion of excess homocysteine to SAH, unless the concentration of adenosine is limiting [34]. Adenosine is not limiting in the dwarf mouse in this study. Homocysteine levels are low in dwarf mice on both diets likely due to the upregulated gene expression and the activities of both the MET recycling and especially the transsulfuration pathways [24].

When MET is restricted, dwarf mice maintain lower levels of both homocysteine and serine, which combine to form cystathionine in reduced amounts, a reaction catalyzed by Cbs. Cbs levels were affected by diet and genotype, yet the resulting levels of cystathionine were unaffected by diet. Cystathionase, also known as γ-cysteine lyase (or CSE) expression was elevated in dwarf mice independent of diet and similar to that found in our previous reports suggesting further dysregulation of these pathways in Ames mice [24, 25]. Although liver cysteine levels did not appear to differ, the manner in which cysteine was utilized differed significantly by genotype. Cysteine is used for de novo GSH biosynthesis, hypotaurine, and taurine biosynthesis, and it is incorporated into a variety of proteins, as well as hydrogen sulfide. Our previous work showed that GH-deficient dwarf mice maintain higher levels of GSH compared to wild-type mice [35]. In the dwarf, GSH, GSSG, and the redox pair ratio were unaltered by diet. GSH levels in wild-type mice did not respond similarly but matched those reported in rats showing that hepatic GSH was low in animals fed restricted MET [1, 2]. Selected metabolite levels within the GSH pathway (including γ-glutamyl cycle) are supportive indicating few dietary effects in dwarf mice in contrast to wild-type mice (Figure 4). Additionally, the dwarf mice fed 0.16% MET maintained lower levels of many of these metabolites in comparison to the wild type further suggesting that the mutant mice do not discriminate changes in MET intake. GSH biosynthesis requires adequate amounts of cysteine, glutamate, and glycine to maintain appropriate levels, and each of these intersect at metabolic ‘centers’ that mirror sulfur and nitrogen metabolism, one-carbon metabolism, and overall energy levels in a tissue [36]. The availability of these precursors reflects the overall metabolic status of cells and tissues [37].

The γ-glutamyl cycle allows GSH to serve as a continuous source of cysteine, and thus appears to contribute to the stable GSH levels in dwarf mice independent of dietary MET intake [38]. In contrast, levels of the γ-glutamylamino acids, 5-hydroxyproline, and GSH are significantly impacted by MET intake in wild-type mice conserving cysteine during low MET intake and utilizing cysteine primarily for GSH and hypotaurine synthesis when MET levels are high. These findings are consistent with the increased recycling of GSH via the γ-glutamyl cycle. As dwarf mice maintain greater GSH under restricted MET along with the decreased demand for growth due to GH deficiency, it makes metabolic ‘sense’ that cysteine may be shifted toward defense and cellular protection mechanisms (GSH, GST…).

2-Hydroxybutyrate (2-OHB) is released as a byproduct when cystathionine is cleaved to the cysteine that is incorporated into GSH. Elevated 2-OHB and further metabolism to ophthalmate (a structural analog of GSH) and norophthalmate reveal that restricted dietary MET increased the abundance in wild-type mice, and enriched MET suppressed formation in both genotypes suggesting the conservation of cysteine for protein synthesis under conditions of MET restriction (Figure 4). During metabolic stress, cysteine is diverted from transmethylation (to form MET) into transsulfuration to form cystathionine. However, we know from MET flux studies that the rates of transmethylation and transsulfuration are two and three times greater, respectively, in dwarf versus wild-type mice [25]. The enhanced MET enzyme activity and mRNA expression help to explain the increased flux through the transmethylation and transsulfuration pathways. Thus, lower levels of 2-OHB, ophthalmate, and norophthalmate in dwarf mice consuming altered MET diets are not reflective of the enhanced liver Gclc mRNA observed in this study. This observation contributes to the hypothesis that deficient GH signaling undermines normal allosteric and regulatory mechanisms affecting multiple metabolic pathways.

The transmethylation and MET salvage (5-methylthioadenosine; MTA) pathways are also involved in the regulation of MET metabolism. Both SAM and MTA work to reduce pro-growth signals including decreasing the mitogenic effects of IGF1 [39]. MTA is part of the MET salvage pathway that involves polyamine synthesis. The levels of MTA tended to be lower in dwarf mice independent of diet (Additional file 1). Others have reported that mice with increased MTA have lower polyamine synthesis [40]. In turn, ornithine levels were decreased in MET-restricted dwarf mice compared to the wild type likely reflecting the known stimulatory effects of GH on ornithine decarboxylase activity and polyamine synthesis [4143]. An increase in transmethylation in both genotypes consuming restricted MET is also supported by increased SAM, betaine, and dimethylglycine. In dwarf mice, other methyltransferases, such as Gnmt and Bhmt as well as DNA methyltransferases, are upregulated when fed normal chow [44, 45]. Thus, the metabolite data indicate that elevated methyltransferase enzyme function is maintained in dwarf mice and increased in wild-type mice consuming excess MET indicative of sensitivity to dietary MET in the presence of GH.

The first step toward cysteine utilization in the synthesis of hypotaurine, taurine, and H2S occurs via cysteine dioxygenase. In dwarf mice, increased cysteine dioxygenase (CDO), Cbs, and Cth suggest increased cysteine, although our metabolite analysis did not detect higher cysteine levels in these mice. Hepatic protein levels of CDO are significantly elevated in dwarf mice on normal rodent chow (Brown-Borg, unpublished data). CDO is expressed in high levels in rodent liver and kidney and is strongly regulated in response to cysteine availability such that in animals fed low protein; CDO is low (conserving cysteine) and is increased (up to 45-fold) in hepatic tissue of animals fed excess sulfur AA, to remove high cysteine and prevent cytotoxicity [46, 47]. The process of desulfuration via CDO dominates when cysteine is deficient (MET restriction), whereas oxidative catabolism dominates when cysteine is in excess (MET enrichment). When adequate sulfur AAs are provided to rats, two thirds of cysteine catabolism is oxidative (CDO) and one third undergoes desulfuration (Cbs, Cth). Both the oxidative and desulfurative mechanisms of cysteine metabolism were upregulated in Ames mice suggesting that more cysteine is being routed through these pathways.

In addition to the enhanced antioxidative enzyme defense observed in dwarf mice, protein glutathionylation likely plays a role in these mice in protecting protein thiols from oxidative-induced damage [48]. We have preliminary evidence indicating greater glutathionylation in Ames mice (Brown-Borg, unpublished data). Glutaredoxins remove glutathionylated modifications from proteins, are dependent on GSH pools, and function to maintain a reduced state of cysteines in cellular protein [4951]. In this study, the elevated hepatic Grx, Trx, and thioredoxin reductase (TrxR) reflect similar relationships between dwarf and wild-type mice reported previously and indicate that altered MET intake does not perturbate this redox system markedly [52]. In addition, GH suppresses the activity of Grx and Trx [52]. These results support the hypothesis that deficient GH signaling heightens oxidant defense mechanisms.

When the liver metabolite analysis is expanded to other key pathways, one can observe a more global picture of genotype and diet effects (Additional files 11 and 12). Several components of glycolysis reveal significant genotype-specific responses. Glucose, glucose 6-phosphate, fructose 6-phosphate, the hexose diphosphates, and glycerol 3-phosphate were higher in dwarf compared to wild-type mice while other glycolytic metabolites (glycerate, PEP, 3-phosphoglycerate, lactate) were lower (Additional file 11). However, different levels of dietary MET did not impact levels of the measured glycolytic components in either genotype. These findings are in contrast to those reported in calorie restriction where the glycolytic pathway was downregulated, and acetyl CoA was unaltered when compared to ad libitum-fed animals [53]. Major components within the Krebs cycle appeared to be similarly unaffected by diet suggesting that MET intake does not impact their functioning, that metabolite levels are not indicative of function, or that the static picture we obtained did not capture potential alterations (Additional file 12).

Coenzyme A is an indispensible cofactor in all organisms functioning as an acyl carrier and carbonyl activating groups in key metabolic pathways including the Krebs cycle and fatty acid metabolism [54]. Many thiol-containing proteins are elevated in dwarf mice including GSH, metallothionein, GSH peroxidase, and selenoprotein R (Msrb), and now, we show that the thiol, coenzyme A, is also higher in dwarf mice regardless of diet (Additional file 12) [55]. These high levels are in concert with the enhanced fat utilization suggested by low respiratory quotients observed in dwarf mice [56, 57]. Long-term MET restriction increases energy expenditure limiting fat deposition despite increased food intake suggesting that these diets lead to increased metabolic flexibility [22]. Since dwarf mice already appear to rely on fat as a metabolic fuel (RQ < 1 and high VO2), an additional mechanism, such as restricted MET intake, may not further impact these systems. These conditions in dwarf mice likely reflect the increased energy demand for thermogenesis [57, 58]. Fatty acid metabolism was impacted by restricted MET in wild-type mice but not in dwarf mice suggesting that the reduced mobilization of fatty acids from lipid stores with low MET may be dependent on GH signaling. Glycerophospholipid metabolism was also driven by restricted dietary MET in a GH-dependent manner with several phospholipid metabolites elevated in MET restriction compared to MET enrichment in wild type but not in dwarf mice (data not shown). Supporting evidence indicating enhanced fatty acid oxidation is derived from studies showing that PGC1-α, PPAR-α, acyl-CoA oxidase (rate-limiting enzyme in β-oxidation), Cyp4a, and others are significantly elevated in dwarf versus wild-type mice [5961].

Many aspects of mitochondrial function are altered in the long-living Ames mice. The OXPHOS enzymes are upregulated, liver mitochondrial H2O2 production is reduced, and both cytosolic and mitochondrial antioxidative enzymes are increased providing the basis for the lower oxidative damage levels to DNA and proteins [55, 62]. MET is sensitive to ROS-mediated oxidation, and the product MET sulfoxide is reduced back to MET by thioredoxin. The increased potential for repair of protein oxidative damage is suggested by higher Msrb protein levels and the increased Trx in dwarf mice [35]. The dwarf maintained lower levels of methionine sulfoxide levels possibly due to altered Trx and Grx (Additional file 1).

Several differences in metabolite levels that report on glucose metabolism were observed with diet in wild-type mice (Additional file 11). Glucose and the six-carbon glycolytic intermediates, fructose 6-phosphate, fructose 1,6-diphosphate, and myo-inositol (1,4 or 1,3)-diphosphate were lower with MET restriction compared to MET enrichment. Glucose polymers that arise from glycogen metabolism (maltopentaose, maltotetraose, maltotriose, and maltose) were significantly lower in wild-type mice consuming 0.16% MET compared to those fed 1.3% MET (data not shown). These metabolites are thought to be reflective of glycogen breakdown [63]. The pentose phosphate pathway metabolites (ribose-5-phosphate, ribulose-5-phosphate, xylulose-5-phosphate, ribulose) were also lower with the 0.16% diet (data not shown).

Although all of the liver metabolomic data is not presented, in total, the data indicates that dwarf mice are able to maintain their distinctive MET metabolism despite changes in MET levels as indicated by relatively few diet-specific alterations when compared to wild-type mice. The responsiveness of the wild-type mouse is reflected in changes in life span when fed altered MET and much like previous studies.