Selection for novel metabolic capabilities in Salmonella enterica

Abstract

Bacteria are known to display extensive metabolic diversity and many studies have shown that they can use an extensive repertoire of small molecules as carbon- and energy sources. However, it is less clear to what extent a bacterium can expand its existing metabolic capabilities by acquiring mutations that, for example, rewire its metabolic pathways. To investigate this capability and potential for evolution of novel phenotypes, we sampled large populations of mutagenized Salmonella enterica to select very rare mutants that can grow on minimal media containing 124 low molecular weight compounds as sole carbon sources. We found mutants growing on 18 of these novel carbon sources, and identified the causal mutations that allowed growth for four of them. Mutations that relieve physiological constraints or increase expression of existing pathways were found to be important contributors to the novel phenotypes. For the remaining 14 novel phenotypes, whole genome sequencing of independent mutants and genetic analysis suggested that these novel metabolic phenotypes result from a combination of multiple mutations. This work, by virtue of identifying the genetic and mechanistic basis for new metabolic capabilities, sheds light on the properties of adaptive landscapes underlying the evolution of novel phenotypes.

Keywords: Carbon sources; experimental evolution; new metabolic functions; novel phenotypes.

Figure 1

Selection of mutants with novel metabolic functions. An inducible error prone polymerase was used to increase the mutation rate of the ancestral Salmonella Typhimurium LT2 strain. The mutagenized population was spread on agar plates containing each of 124 novel carbon compounds. Mutants were selected, whole genome sequenced and causal mutations were identified. Fitness trade‐offs caused by the novel phenotypes were measured.

Figure 2

Utilization of L‐isoleucine as sole carbon source. (A) Common enzymes involved in biosynthesis of L‐isoleucine, L‐leucine and L‐valine: Pathway showing common enzymes for synthesis of branched chain amino acids L‐isoleucine, L‐valine, and L‐leucine. All amino acids exert feedback inhibition on these common enzymes. Isoleucine can also enter the isoleucine degradation pathway. (B) Gene expression analysis for genes in ilvN and ilvE: qPCR measurements of genes ilvN and ilvE in wild‐type ancestral strain and the IlvN N35K mutant in different media (shown as different colored bars). The values are normalized to the expression levels observed in the wild‐type strain when grown in the respective media. Error bars represent standard deviation. Two‐tailed Student's t test was used to calculate significant differences (denoted by * when applicable). No difference was observed in expression level of ilvE gene in different media conditions.

Figure 3

Isoleucine toxicity and rescue of ancestral strain in presence of valine/leucine. Relative exponential growth rates of the llvN N35K mutant compared to the wild‐type strain in different media: 0.1% minimal glucose media, 0.1% minimal glucose media with 0.1% isoleucine and 0.1% minimal glucose media with 0.1% each of isoleucine, valine, and leucine. Error bars represent standard deviation. Two‐tailed Student's t test was used to calculate significant differences at P < 0.01 (denoted by * when applicable).

Figure 4

Mutations allowing utilization of L‐threonine as sole carbon source. (A) Mutations in lrp and upstream of kbl allowing utilization of L‐threonine as sole carbon source. (B) Gene expression analysis for genes kbl and tdh from threonine degradation pathway: qPCR measurements of genes kbl and tdh in Lrp A63E and kbl ‐60 G>A mutants in different media. The values are normalized to the expression levels observed in the wild‐type strain when grown in 0.1% glucose M9‐minimal media. Error bars represent standard deviation. Two‐tailed Student's t test was used to calculate significant differences at P < 0.01 (denoted by * when applicable).

Figure 5

Relative exponential growth rates for llvN N35K mutant on alternate native carbon sources. Exponential growth rates of the mutants were measured and normalized with respect to growth rate of wild‐type strain in minimal media containing 0.2% of different carbon sources (glucose, acetate, citrate, or glycerol).

Figure 6

Fitness trade‐offs observed in Lrp A63E and kbl ‐60 G>A mutants under different growth conditions. Relative exponential growth rate for Lrp A63E and kbl ‐60 G>A mutants, with respect to the wild‐type strain, measured in different media: 0.1% minimal glucose media, 0.1% minimal glucose media with 0.1% threonine and 0.1% minimal glucose media with 0.1% leucine. Two‐tailed Student's t test was used to calculate significant differences at P < 0.01 (denoted by * when applicable).