Genome-wide Phenotypic Profiling Identifies and Categorizes Genes Required for Mycobacterial Low Iron Fitness

Abstract

Iron is vital for nearly all living organisms, but during infection, not readily available to pathogens. Infectious bacteria therefore depend on specialized mechanisms to survive when iron is limited. These mechanisms make attractive targets for new drugs. Here, by genome-wide phenotypic profiling, we identify and categorize mycobacterial genes required for low iron fitness. Mycobacterium tuberculosis (Mtb), the causative agent of tuberculosis (TB), can scavenge host-sequestered iron by high-affinity iron chelators called siderophores. We take advantage of siderophore redundancy within the non-pathogenic mycobacterial model organism M. smegmatis (Msmeg), to identify genes required for siderophore dependent and independent fitness when iron is low. In addition to genes with a potential function in recognition, transport or utilization of mycobacterial siderophores, we identify novel putative low iron survival strategies that are separate from siderophore systems. We also identify the Msmeg in vitro essential gene set, and find that 96% of all growth-required Msmeg genes have a mutual ortholog in Mtb. Of these again, nearly 90% are defined as required for growth in Mtb as well. Finally, we show that a novel, putative ferric iron ABC transporter contributes to low iron fitness in Msmeg, in a siderophore independent manner.

Figure 1

Msmeg in vitro growth requirement analysis. (a) Transposon insertion counts across the Msmeg genome (Middlebrook 7H10-selected library). The height of the black bars represents the number of insertion counts at the respective genome site. (b) Definition of the Msmeg wt in vitro requirement for growth (Middlebrook 7H10-selected library). 5815 genes were identified as non-essential (NE) for growth, 415 as causing growth advantage when disrupted (GA), 306 as essential (ES), and 97 as causing growth defect when disrupted (GD). 83 genes were not defined. (c) Venn diagram illustrating Msmeg required genes (ES and GD) and Mtb required genes (ES, GD and ESD, the latter genes with essential domains) as defined by DeJesus et al., relative to the entire pool of Msmeg (6716) and Mtb (4019) genes and their mutual orthologs (2547). Msmeg (403) and Mtb (625) required genes are shown within the blue and red circle, respectively. 343 genes are required in both species, 44 (Msmeg) and 247 (Mtb) required genes have a non-required mutual ortholog in the other species, and 16 (Msmeg) and 35 (Mtb) required genes do not have mutual orthologs in the other species (d) Venn diagram illustrating Msmeg and Mtb GA (causing growth advantage when disrupted) genes, as described for (c). Msmeg (415) and Mtb (283) GA genes are shown within the blue and red circle, respectively. 27 genes cause growth advantage in both species, 121 (Msmeg) and 123 (Mtb) GA genes have a non-GA mutual ortholog in the other species, and 267 (Msmeg) and 160 (Mtb) GA genes do not have mutual orthologs in the other species.

Figure 2

Schematic overview of screen for low iron fitness genes. (a) A transposon mutant library of Msmeg wt selected on high iron levels was compared to low iron-selected transposon libraries of Msmeg wt, ΔfxbA (exochelin knockout), ΔmbtD (mycobactin knockout) and ΔfxbAΔmbtD (double siderophore knockout) to identify and categorize genes involved in low iron growth. The amplitude of the blue vertical bars represents the hypothetical number of transposon insertions counted at the given TA dinucleotide site. (b) Msmeg wt, ΔfxbA, ΔmbtD and ΔfxbAΔmbtD mutants were prewashed in chelated Sauton’s before diluted and spotted on chelated Sauton-based agar plates with increasing concentrations of FeCl₃. (c) Overview of library size, selective conditions, TA sites hit and total insertion count of the five sequenced libraries of the screen for low iron genes.

Figure 3

Validation of screen for low iron fitness genes. Distribution of transposon insertion counts (blue vertical bars) for all sequenced libraries. (a) Whole genome. Insertion counts in log scale of 1–1500. (b) Exochelin gene cluster. Insertion counts in log scale of 0–50. The red triangles indicate the knocked out gene (msmeg_0014, fxbA). (c) Mycobactin gene cluster. Insertion counts in log scale of 0–50. The red triangles indicate the knocked out gene (msmeg_4512, mbtD). (d) ESX-3 gene cluster. Insertion counts in log scale of 0–25. Plots created using IGV - distributed by the Broad Institute (http://www.broadinstitute.org/igv/). Black dashed boxes show genes significantly under-represented in insertion counts in one or more of the low iron libraries compared to the high iron library.

Figure 4

Identification and categorization of low iron fitness genes. Transposon insertion counts presented relative to counts ratio (control/experimental) per gene between Msmeg wt high iron control library and (a) Msmeg wt, (b) ΔfxbA, (c) ΔmbtD or (d) ΔfxbAΔmbtD mutant low iron libraries (x-axis) and the corresponding P-values calculated by Mann Whitney U-test (y-axis). Genes previously known to be involved in mycobactin-mediated (green), exochelin-mediated (pink), or siderophore-independent (blue) iron uptake are color-coded. Red lines represent a cutoff of genes more than 5 fold under- (to the right) or over-represented (to the left) and with a P-value of less than 0.05. Genes knocked out are circled with pink (fxbA) or green (mbtD). In figure c is msmeg_0019 out of scale with P-value 1.5 × 10⁻⁹ and relative counts ratio 1.478.

Figure 5

msmeg_3635 is important for Msmeg siderophore-independent low iron growth. (a) Growth of Msmeg strains in high (upper panel) or low iron (lower panel) monitored over time (x-axis) by OD₆₀₀ (y-axis). Error bars represent standard error from the mean of three biological replicas. To the ΔfxbAΔmbtD and ΔfxbAΔmbtDΔ3635 curves, 3 μM FeCl₃ was added in the low iron condition. To all other curves, 100 μM 2, 2′-bipyridine was add to low iron, and 150 μM FeCl₃ was added to high iron condition. (b) Growth of ΔfxbAΔmbtD and ΔfxbAΔmbtDΔ3635 in low iron, low zinc (upper left), low iron, high zinc (lower left), high iron, low zinc (upper right), and high iron, high zinc (lower right). For all curves; low iron was supplemented with 3 μM FeCl₃, low zinc with 0 μM ZnSO₄, high iron with 150 μM FeCl₃, and high zinc with 3.67 μM ZnSO₄. (c) Distribution of transposon insertion counts (blue vertical bars) for all sequenced libraries in the msmeg_3630-3636 operon. Insertion counts in log scale of 0–75. Plot created using IGV - distributed by the Broad Institute (http://www.broadinstitute.org/igv/). The black dashed box shows genes significantly under-represented in insertion counts in one or more of the low iron libraries compared to the high iron library. (d) Complementation of the ΔfxbAΔmbtDΔ3635 low iron phenotype. Msmeg ΔmbtDΔfxbA, ΔmbtDΔfxbAΔ3635 and ΔmbtDΔfxbAΔ3635 compl. were grown in either high (150 μM FeCl₃) or low (1.5 μM FeCl₃) iron.