Uncovering major genomic features of essential genes in Bacteria and a methanogenic Archaea

Abstract

Identification of essential genes is critical to understanding the physiology of a species, proposing novel drug targets and uncovering minimal gene sets required for life. Although essential gene sets of several organisms have been determined using large-scale mutagenesis techniques, systematic studies addressing their conservation, genomic context and functions remain scant. Here we integrate 17 essential gene sets from genome-wide in vitro screenings and three gene collections required for growth in vivo, encompassing 15 Bacteria and one Archaea. We refine and generalize important theories proposed using Escherichia coli. Essential genes are typically monogenic and more conserved than nonessential genes. Genes required in vivo are less conserved than those essential in vitro, suggesting that more divergent strategies are deployed when the organism is stressed by the host immune system and unstable nutrient availability. We identified essential analogous pathways that would probably be missed by orthology-based essentiality prediction strategies. For example, Streptococcus sanguinis carries horizontally transferred isoprenoid biosynthesis genes that are widespread in Archaea. Genes specifically essential in Mycobacterium tuberculosis and Burkholderia pseudomallei are reported as potential drug targets. Moreover, essential genes are not only preferentially located in operons, but also occupy the first position therein, supporting the influence of their regulatory regions in driving transcription of whole operons. Finally, these important genomic features are shared between Bacteria and at least one Archaea, suggesting that high order properties of gene essentiality and genome architecture were probably present in the last universal common ancestor or evolved independently in the prokaryotic domains.

Keywords: essential genes; genome evolution; genome organization; operons; prokaryotes; transposon mutagenesis.

Figure 1. Essential genes obtained from 17 dispensability experiments and their correlation to the total gene complement

A) Percentage of essential protein-coding genes; B) Correlation between essential gene set size and genome size. Abbreviations: abayl-min (Acinetobacter baylyi ADP1, minimal medium); bfrag (Bacteroides fragilis 638R); bpseu (Burkholderia pseudomallei K96243); bsubt (Bacillus subtilis 168); bthai (Burkholderia thailandensis E264); ccres (Caulobacter crescentus NA1000); ecoli (Escherichia coli K-12); ftula (Francisella tularensis novicida U112); mgeni (Mycoplasma genitalium G37); mmari (Methanococcus maripaludis S2, rich medium); mmarimin (Methanococcus maripaludis S2, minimal medium); mpulm (Mycoplasma pulmonis CT); mtube-min (Mycobacterium tuberculosis H37Rv, minimal medium); pging (Porphyromonas gingivalis ATCC 33277); sente-SL1344 (Salmonella enterica typhimurium SL1344); sente-Ty2 (Salmonella enterica typhi Ty2); ssang (Streptococcus sanguinis SK36, minimal medium).

Figure 2. Functional categories enriched in essential gene datasets

Squares in magenta represent functional categories enriched in the respective essential gene set (Fisher's exact test; P < 0.05).

Figure 3. Conservation of essential and nonessential gene sets across thousands of species

Boxplot representation of essential gene sets across thousands of species available in the eggNOG database (see methods for details). Unless indicated otherwise, rich media were used in the screenings. For abbreviations of in vitro experiments refer to Figure 1. For in vivo experiments: ftula-invivo (F. tularensis novicida U112, in vivo); mtube-invivo (Mycobacterium tuberculosis H37Rv, in vivo); and mtube-mac (Mycobacterium tuberculosis H37Rv, macrophages). Essential and nonessential gene sets for each condition are side-by-side, in dark and light colors. Proteobacteria, Actinobacteria, Bacteroides, Tenericutes, Firmicutes and Archaea are represented in blue, green, purple, magenta, brown and red, respectively.

Figure 4. Multiple Correspondence Analysis (MCA) of the presence/absence of essential genes in NOGs

MCA analysis of essential gene sets based on the presence/absence profiles of each mapped NOG. The first two dimensions obtained in MCA were dominated by one or two samples and therefore, are not very useful for separation purposes. Dimensions 3 and 4 allowed an evolutionarily coherent clustering, while still accounting for a significant amount of variance. For abbreviations of in vitro experiments refer to Figure 1. For in vivo experiments: ftula-invivo (F. tularensis novicida U112, in vivo); mtube-invivo (Mycobacterium tuberculosis H37Rv, in vivo); and mtube-mac (Mycobacterium tuberculosis H37Rv, macrophages). For color codes, refer to Figure 3.

Figure 5. Association between the number of coding genes and gene essentiality with the presence of homologs

A) Total number of coding genes versus genes in multigene families: genes with same COG/NOG assignment in a genome were considered part of multigene families. B) Gene essentiality versus presence of a homolog in the genome: Fisher's exact tests were performed to assess the enrichment of essential genes in multigene families. Bars with one and two asterisks represent P ≤ 10⁻² and P ≤ 10⁻⁵, respectively. For abbreviations of in vitro experiments refer to Figure 1. For in vivo experiments: ftula-invivo (F. tularensis novicida U112, in vivo); mtubeinvivo (Mycobacterium tuberculosis H37Rv, in vivo); and mtube-mac (Mycobacterium tuberculosis H37Rv, macrophages). For color codes, refer to Figure 3.