The impact of genetic diversity on gene essentiality within the Escherichia coli species

Abstract

Bacteria from the same species can differ widely in their gene content. In E. coli, the set of genes shared by all strains, known as the core genome, represents about half the number of genes present in any strain. While recent advances in bacterial genomics have enabled to unravel genes required for fitness in various experimental conditions at the genome scale, most studies have focused on model strains. As a result, the impact of this genetic diversity on core processes of the bacterial cell largely remains to be investigated. Here, we developed a new CRISPR interference platform for highthroughput gene repression that is compatible with most E. coli isolates and closely-related species. We applied it to assess the importance of ~3,400 nearly ubiquitous genes in 3 growth media in 18 representative E. coli strains spanning most common phylogroups and lifestyles of the species. Our screens highlighted extensive variations in gene essentiality between strains and conditions. Unlike variations in gene expression level, variations in gene essentiality do not recapitulate the strains' phylogeny. Investigation of the genetic determinants for these variations highlighted the importance of epistatic interactions with mobile genetic elements. In particular, we showed how mobile genetic elements can trigger the essentiality of core genes that are usually nonessential. This study provides new insights into the evolvability of gene essentiality and argues for the importance of studying various isolates from the same species in bacterial genomics..

Essential gene acquisition destabilizes plasmid inheritance

“…The set of essential genes we tested here comprises 504 protein-coding genes (S2 Table ); of those, 394 protein-coding genes were identified as essential in E. coli K-12 under standard laboratory conditions via gene knockout [24] or transposon mutagenesis [25]. Additional 110 proteincoding genes were identified as essential in a collection of 18 E. coli isolates by CRISPR interference of gene expression ( [18]; only genes identified as essential in at least one out of the three tested growth media were included; S2 Table ). Searching for plasmid-encoded homologs to the essential genes using sequence similarity revealed that only 17 (3%) of the essential genes had a homolog on plasmids (Fig 1A and Table 1).…”

“…In the second route, termed stability first, a plasmid that initially evolved a stable reproduction cycle may subsequently evolve into an essential plasmid following an essential gene gain. Notwithstanding, as essential genes are typically encoded in the chromosomes [18], the presence of an essential gene on a plasmid may be comparable to the effect of gene duplication. In addition, the expression level of plasmid-encoded genes may be amplified following an increase in plasmid copy number (e.g., [19]).…”

Genome-scale metabolic modelling when changes in environmental conditions affect biomass composition

“…In addition, we performed a semi-quantitative comparison with experimental gene knockout data of E. coli K12 MG1655 [ 37 ] in three different nutrient environments. The approach is described in Materials and Methods with resulting data in Fig C in S1 Text .…”

“…We compared the computational single-gene knockout phenotype predictions for the different biomass compositions with experimental data from Rousset et al [ 37 ] for E. coli K12 MG1655 grown in three different media: minimal, rich, and gut microbiota. We divided the experimental data into three groupings based on the given scores: The first group consisted of genes with score s < −3, where the gene knockout could be considered essential.…”

“…Fig 3 ; Fig A in S1 Text) and provides a visual mapping of environmental space coordinates used in the single-gene knockout analysis (Fig B in S1 Text). In Fig C in S1 Text, we show the gene essentiality similarity of the model predictions in relation to experimental data by Rousset et al [ 37 ], the raw data is shown in S2 Table .…”