Genomewide Assessment of Mycobacterium tuberculosis Conditionally Essential Metabolic Pathways

Abstract

A better understanding of essential cellular functions in pathogenic bacteria is important for the development of more effective antimicrobial agents. We performed a comprehensive identification of essential genes in Mycobacterium tuberculosis, the major causative agent of tuberculosis, using a combination of transposon insertion sequencing (Tn-seq) and comparative genomic analysis. To identify conditionally essential genes by Tn-seq, we used media with different nutrient compositions. Although many conditional gene essentialities were affected by the presence of relevant nutrient sources, we also found that the essentiality of genes in a subset of metabolic pathways was unaffected by metabolite availability. Comparative genomic analysis revealed that not all essential genes identified by Tn-seq were fully conserved within the M. tuberculosis complex, including some existing antitubercular drug target genes. In addition, we utilized an available M. tuberculosis genome-scale metabolic model, iSM810, to predict M. tuberculosis gene essentiality in silico Comparing the sets of essential genes experimentally identified by Tn-seq to those predicted in silico reveals the capabilities and limitations of gene essentiality predictions, highlighting the complexity of M. tuberculosis essential metabolic functions. This study provides a promising platform to study essential cellular functions in M. tuberculosis IMPORTANCE Mycobacterium tuberculosis causes 10 million cases of tuberculosis (TB), resulting in over 1 million deaths each year. TB therapy is challenging because it requires a minimum of 6 months of treatment with multiple drugs. Protracted treatment times and the emergent spread of drug-resistant M. tuberculosis necessitate the identification of novel targets for drug discovery to curb this global health threat. Essential functions, defined as those indispensable for growth and/or survival, are potential targets for new antimicrobial drugs. In this study, we aimed to define gene essentialities of M. tuberculosis on a genomewide scale to comprehensively identify potential targets for drug discovery. We utilized a combination of experimental (functional genomics) and in silico approaches (comparative genomics and flux balance analysis). Our functional genomics approach identified sets of genes whose essentiality was affected by nutrient availability. Comparative genomics revealed that not all essential genes were fully conserved within the M. tuberculosis complex. Comparing sets of essential genes identified by functional genomics to those predicted by flux balance analysis highlighted gaps in current knowledge regarding M. tuberculosis metabolic capabilities. Thus, our study identifies numerous potential antitubercular drug targets and provides a comprehensive picture of the complexity of M. tuberculosis essential cellular functions.

Keywords: Tn-seq; comparative genomics; metabolic modeling; metabolism; tuberculosis.

FIG 1

Identification of essential genes in MtbYM rich medium. (A) Medium compositions of Mtbminimal and MtbYM rich media. 7H9 medium contains glutamic acid and ammonium sulfate in addition to the nutrients shown in bold. DAP, diaminopimelic acid; FAC, ferric ammonium citrate; PABA, para-aminobenzoic acid; *, 0.5% Casamino Acids and 0.98 mM tryptophan. (B) M. tuberculosis H37Rv growth in MtbYM rich (red), 7H9 (black), and Mtbminimal (blue) media. (C) Bayesian/Gumbel results for M. tuberculosis H37Rv in MtbYM rich medium. (D) HMM results for M. tuberculosis H37Rv in MtbYM rich medium. (E) Venn diagram comparisons of essential genes by Tn-seq in this study and two previous studies (8, 9). Red, this study; blue, Griffin et al. (8); green, DeJesus et al. (9).

FIG 2

Identification of conditionally essential genes in MtbYM rich and Mtbminimal media. (A) Results for comparative analysis between MtbYM rich and Mtbminimal media. Numbers of conditionally essential genes (adjusted P value of <0.05) were identified by the resampling method in TRANSIT (14). (B, C) Essentialities of genes in the pantothenate biosynthesis pathway (B) and methionine biosynthesis pathway (C). Genes in red are essential in both MtbYM rich and Mtbminimal media, genes in blue are essential in Mtbminimal medium but not essential in MtbYM rich medium, genes in orange are essential in MtbYM rich medium but not essential in Mtbminimal medium, and genes in black are nonessential. Information on genes in each pathway was obtained from the BioCyc database (18). (D) Ratios of essential genes in the metabolic pathways. All genes are listed in Table S5. PABA, para-aminobenzoic acid; purine, 5-aminoimidazole ribonucleotide.

FIG 3

Identification of highly conserved essential genes among virulent M. tuberculosis complex strains. (A) Strains used for comparative genomic analysis. (B) Results for pan-genome analysis by Roary. (C) Number of highly conserved genes among virulent M. tuberculosis complex. Core genes are conserved in >99% of strains, soft core genes are conserved in 95 to 99% of strains, shell genes are conserved in 15 to 95% of strains, and cloud genes are conserved in 0 to 15% of strains. (D) Numbers of genes that are essential for M. tuberculosis H37Rv survival in MtbYM rich medium and highly conserved among virulent M. tuberculosis complex strains.

FIG 4

Comparison of Tn-seq-identified essential genes and in silico-predicted essential genes. (A) Results for in silico gene essentiality prediction by iSM810 and iSM810 (modified). (B) Comparison of Tn-seq-identified essential genes (blue) and iSM810-predicted essential genes (red). (C) Comparison of Tn-seq-identified essential genes and iSM810-predicted essential genes in the purine biosynthesis pathway. Genes in red are essential in both MtbYM rich and Mtbminimal media, genes in blue are essential in Mtbminimal medium but not essential in MtbYM rich medium, and genes in black are nonessential. Information on genes in each pathway was obtained from the BioCyc database (18).