Evolutionary implications of Liebig's law of the minimum: Selection under low concentrations of two nonsubstitutable nutrients

Abstract

Interactions between different axes of an organism's niche determine the evolutionary trajectory of a population. An extreme case of these interactions is predicted from ecological theory in Liebig's law of the minimum. This law states that in environments where multiple nutrients are in relatively low concentrations, only one nutrient will affect the growth of the organism. This implies that the evolutionary response of the population would be dictated by the most growth-limiting nutrient. Alternatively, it is possible that an initial adaptation to the most limiting nutrient results in other nutrients present in low concentration affecting the evolutionary dynamics of the population. To test these hypotheses, we conducted twelve evolution experiments in chemostats using Escherichia coli populations: four under nitrogen limitation, four under magnesium limitation, and four in which both nitrogen and magnesium are in low concentrations. In the last environment, only magnesium seems to limit growth (Low Nitrogen Magnesium Limited environment, LNML). We observe a decrease in nitrogen concentration in the LNML environment over the course of our evolution experiment indicating that nitrogen might become limiting in these environments. Genetic reconstruction results show that clones adapted to magnesium limitation have genes involved in nitrogen starvation, that is, glnG (nitrogen starvation transcriptional regulator) and amtB (transport protein) to be upregulated only in the LNML environment as compared to magnesium-limiting environments. Together, our results highlights that in low-nutrient environments, adaptation to the growth-limiting nutrient results in other nutrients at low concentrations to play a role in the evolutionary dynamics of the population.

Keywords: Escherichia coli; Liebig; experimental evolution; magnesium limitation; nitrogen limitation.

Figure 1

Identification of concentrations of nitrogen and magnesium that are growth limiting. The stationary phase population density of cultures was measured for twelve combinations of nitrogen and magnesium concentrations. The nitrogen concentrations used are 0, 0.07, 0.7, and 14 mmol/L. Each curve represents a given magnesium ion concentration ( for 0 mmol/L, ‐ ‐ for 0.4 mmol/L and – – for 0.8 mmol/L). Nitrogen‐limiting concentrations represent conditions where ammonium was not detected using an indophenol assay (■). Δ represents environments where nitrogen was detected. Experimental evolution was performed at concentrations denoted by A (0.7 mmol/L nitrogen, 0.8 mmol/L magnesium), B (14 mmol/L nitrogen, 0 mmol/L magnesium), and C (0.7 mmol/L nitrogen, 0 mmol/L magnesium), which represents nitrogen‐limitation, magnesium‐limitation, and LNML environment, respectively

Figure 2

glnG and amtB gene expression levels in Escherichia coli MG1655 in environment A, B, and C. qPCR analysis shows statistically different expression levels for genes glnG and amtB as compared between environment A and environment B, as well as between environment A and environment C (p < .01). The difference between environment B and environment C is not statistically significant (p > .01). Samples were taken after populations reached stationary phase in the appropriate nutrient‐limiting environments

Figure 3

Fitness changes in 12 independent populations over four time points: 72, 168, 240, and 400 generations. Selection (Relative fitness −1) was measured by performing competition experiments against the ancestor in the same media in which the population evolved, regressing the ratios of population densities on time, and by measuring the slope. Error bars represent standard error. Different line types represent different replicates for each nutrient‐limiting environment

Figure 4

Fitness of clones compared to the fitness of the populations from which they were isolated. Evolved clones and evolved populations were competed against the ancestral population in the nutrient‐limiting conditions in which they evolved. p1 is the probability that the fitness of the clone and the population from which it evolved are the same. It is predicted that the fitness of the clone (●) and the population from which it evolved (■) will be the same (p1 > .05), but in seven of twelve cases the fitness of the clone is significantly higher than the fitness of the population from which it came. The fitness of the clone was also measured under nutrient replete conditions (▲) to infer general adaptation that is not the result of nutrient limitation. It is expected that the fitness of the clone under nutrient replete conditions should be significantly less that the fitness of either this clone or the population from which it was isolated under appropriate nutrient‐limitation environment. p2 is the comparison between the fitness of the clone in a nutrient replete environment and the appropriate nutrient‐limited environment. p3 is the comparison between the fitness of the clone in a nutrient replete environment and the population in a nutrient‐limited environment. The expectation is met in ten of the twelve replicates. In one replicate (#3 on LNML), all three fitnesses (the clone, the population, and the clone in nutrient replete environment) are the same suggesting that all the fitness increase is unrelated to evolution to the limiting nutrient

Figure 5

The adaptation of clones evolved in LNML to a limiting nitrogen environment (■) and a limiting magnesium environment (▲) compared with the adaptation in LNML(●). p1 is the probability that the adaptation to LNML (●) and to the nitrogen‐limiting environment (■) is the same. p2 is the probability that the adaptation to LNML(●) and to the magnesium‐limiting environment (▲) is the same

Figure 6

Fitness increase of three populations evolved in LNML over time in different environments in a LNML environment (), in a magnesium‐limited environment () and in a nitrogen‐limited environment (). The samples were taken at 72, 168, 240, and 400 generations

Figure 7

Changes in ammonium ion concentration during long‐term evolution experiments. Changes in concentration of ammonium ion under nitrogen‐limitation (■), magnesium‐limitation (▲), and LNML environment (●) over 400 generations. Under each limitation four replicate populations are represented as Population 1 (), Population 2 (), Population 3 (), and Population 4 ()

Figure 8

(a) Relative fitness measurements for a ΔyhaV phoQ L467P mutant under nitrogen‐limitation, magnesium‐limitation, and LNML environment. (b) Fold expression change for genes glnG and amtB under magnesium limitation for ΔyhaV phoQ L467P mutant relative to ancestral strain. (c) Fold expression change for genes glnG and amtB in LNML environment for ΔyhaV phoQ L467P mutant relative to ancestral strain. Asterisk indicates statistically significant difference (p < .01) between gene expression levels for the ancestral strain and mutant strain in a given environment