Experimental evolution of metabolism under nutrient restriction: enhanced amino acid catabolism and a key role of branched-chain amino acids

Abstract

Periodic food shortage is a common ecological stressor for animals, likely to drive physiological and metabolic adaptations to alleviate its consequences, particularly for juveniles that have no option but to continue to grow and develop despite undernutrition. Here we study changes in metabolism associated with adaptation to nutrient shortage, evolved by replicate Drosophila melanogaster populations maintained on a nutrient-poor larval diet for over 240 generations. In a factorial metabolomics experiment we showed that both phenotypic plasticity and genetically-based adaptation to the poor diet involved wide-ranging changes in metabolite abundance; however, the plastic response did not predict the evolutionary change. Compared to nonadapted larvae exposed to the poor diet for the first time, the adapted larvae showed lower levels of multiple free amino acids in their tissues-and yet they grew faster. By quantifying accumulation of the nitrogen stable isotope ¹⁵N we show that adaptation to the poor diet led to an increased use of amino acids for energy generation. This apparent "waste" of scarce amino acids likely results from the trade-off between acquisition of dietary amino acids and carbohydrates observed in these populations. The three branched-chain amino acids (leucine, isoleucine, and valine) showed a unique pattern of depletion in adapted larvae raised on the poor diet. A diet supplementation experiment demonstrated that these amino acids are limiting for growth on the poor diet, suggesting that their low levels resulted from their expeditious use for protein synthesis. These results demonstrate that selection driven by nutrient shortage not only promotes improved acquisition of limiting nutrients, but also has wide-ranging effects on how the nutrients are used. They also show that the abundance of free amino acids in the tissues does not, in general, reflect the nutritional condition and growth potential of an animal.

Keywords: Drosophila melanogaster; deamination; dietary restriction; experimental evolution; malnutrition; metabolomics.

Figure 1.

Diet quality induces both phenotypically plastic responses (“Diet”) and evolutionary changes (“Regime”) in metabolism. (A) PCA on the abundance of all 174 metabolites; each point corresponds to a population estimate averaged over replicate samples (an animated version is available in Supplementary Figure S1). (B) Venn diagram summarizing the overlaps between significant metabolites. For diet this includes the main effect whereas for regime it includes both the main effect of regime and the regime contrast on poor diet (see text).

Figure 2.

The effect of plastic response and evolutionary adaptation to poor larval diet on abundance of individual metabolites in 3rd instar larvae. For each metabolite, the first column of the heat map represents the estimated effect of diet (poor–standard) on Control larvae; the second column shows the effect of evolutionary regime (Selected–Control) on larvae raised on poor diet. *q < 0.1 for the corresponding contrast; (*) marks metabolites with q < 0.1 for the main effect of evolutionary regime but not for Selected–Control contrast on poor diet; “×” between columns indicates regime × diet interaction (q < 0.1, for patterns of those metabolites see Figure 4 and Supplementary Figure S3). Several metabolites show diet effect with log-fold changes greater than ±1 (Supplementary Figure S2); the color scale only varies between −1 and 1 to optimize resolution within this range. For a more traditional heatmap based on clustering of individual populations see Supplementary Figure S4.

Figure 3.

Plasticity versus evolutionary change. (A) Adaptation to poor diet as modifier of ancestral plastic response: the relationship between metabolite abundance in Control and Selected larvae raised on poor diet, both expressed as a deviation from the ancestral state (Control larvae on poor diet). Colored symbols indicate metabolites with significant effect (q < 0.1) of diet on Control larvae (“diet in CTL”), of evolutionary regime on larvae raised on poor diet (“regime on poor”), and of both evolutionary regime and diet (“regime and diet”). CTL: Controls; SEL: Selected. std: standard diet. (B) The relationship between plastic and evolutionary responses to diet expressed as the main effects of diet and regime from the factorial model (see Methods). Each symbol corresponds to one metabolite, solid symbols are metabolites significant (q < 0.1) for the effects of both evolutionary regime. (C) The relationship between plastic responses of metabolome of Control and Selected populations. Solid symbols: metabolites significant (q < 0.1) for the main effect of diet; solid line represents the major axis regression fitted to all metabolites, dashed line is the diagonal. Correlations were calculated using all metabolites.

Figure 4.

Branched-chain amino acids (BCAA) and adaptation to poor diet. Top row: Patterns of abundance (mean ± SE) of amino acids that show significant regime × diet interaction in the metabolome data: leucine (A), isoleucine (B), valine (C) (BCAA) and proline (D); n.s., q > 0.1; *q < 0.1; **q < 0.05; ***q < 0.01. Bottom row: Effects of supplementing poor diet with BCAA or a mix of tryptophan, histidine, and lysine (WHK) on egg-to-adult survival probability (E), developmental rate (F), female dry weight (G), and female estimated growth rate (H). Symbols indicate means ± SE. N = 3 replicate bottles per population and treatment. CTL: Controls. SEL: Selected. Reg: regime. Treat: treatment. Details of statistics in Supplementary Tables S4–S8.

Figure 5.

Plasticity and evolution of amino acid deamination in response to poor diet. (A) Simplified purine/uric acid synthesis pathway (created in Lucidchart, www.lucidchart.com). Metabolites are colored according to their estimated fold change (Selected – Control) on poor diet; *q < 0.1. XMP was not reliably quantified. (B) Relative concentration of urate in Selected and Control larvae raised on poor and standard diet. (C) Isotope content after feeding on poor or standard diet enriched in ¹⁵N. Δ¹⁵N (‰ N₂-Air) refers to the difference between δ¹⁵N of prepupae or adults and δ¹⁵N of the larval food. Symbols indicate means ± SE. Details of statistics in Supplementary Table S9. N = 2 samples of 20–30 individuals per population, diet, and life stage.