Fig 1.
Riboflavin biosynthesis and utilization pathways in mycobacteria.
(A) In the biosynthesis pathway, GTP and ribulose-5-P are converted to 2,5-diamino-6-(5-phospho-D-ribosylamino)-pyrimidin-4(3H)-one (DARPP) and 3,4-dihydroxyl-2-butanone 4-phosphate (3,4-DHBP), respectively, by the bifunctional GTP cyclohydrolase II/ DHBP synthase, RibA2. DARPP is deaminated and the side chain subsequently reduced by the bifunctional riboflavin deaminase/ 5-amino-6-(5-phosphoribosylamino) uracil reductase, RibG, to form 5-amino-6-ribitylamino-2,4(1H,3H)-pyrimidinedione-5-phosphate (ArPP). Dephosphorylation of ArPP by an unknown phosphatase generates 5-amino-6-D-ribitylaminouracil (5-A-RU). 5-A-RU together with 3,4-DHBP are condensed by the lumazine synthase, RibH (RibH1/RibH2), to yield 6,7-dimethyl-8-ribityllumazine (DMRL). Two molecules of DMRL are converted to riboflavin and 5-A-RU by the riboflavin synthase, RibC via a dismutation reaction. In the utilization pathway, riboflavin is converted to FMN and FAD by the bifunctional kinase/ FAD synthetase, RibF. The deazaflavin, F420, is produced from 5-A-RU by the sequential action of FbiC, FbiA and FbiB. The deletion mutations are shown in red. (B) Genomic context of the riboflavin biosynthesis genes in Msm and Mtb.
Fig 2.
Growth kinetics of wildtype, knockout and complemented mutant strains of Msm (A-G) and Mtb (H-J) in 7H9 media.
Growth of strains in liquid culture was monitored every 3 hours for 12 hours for Msm, and every day for 7 days for Mtb. Data are plotted as the mean, and error bars represent standard deviation from four biological replicates. Statistical comparisons against wildtype were performed at the 12-hr (for Msm) or 7-day (for Mtb) time point using area under the curve analysis (for Msm, panels A-G) or a one-way ANOVA and Dunnett’s multiple comparison test (for Mtb, panels H-J) whereby statistical significance is represented by p < 0.05, p < 0.001 shown by *, ** respectively. Only statistically significant results (p < 0.05) are shown. RF; riboflavin.
Fig 3.
Quantification of ribA2, ribC and ribG transcript and protein levels in Msm strains.
(A, C, and E) Fold changes in sigA-normalized transcript levels relative to wildtype. Transcript levels of the target genes were normalized to the housekeeping gene sigA (an essential housekeeping gene which is stably expressed) and scaled to the average of wildtype Msm to calculate the ΔΔCt and determine the fold difference in gene expression (n = 2 biological replicates, 2 technical replicates). (B, D, and F) Protein abundance was measured using DIA proteomics (n = 3 biological replicates, 2 technical replicates). Data are shown as mean quantification ± SEM. Statistical comparisons were performed using a one-way ANOVA and Sidak’s multiple comparison test whereby statistical significance is represented by p < 0.05, p < 0.001, p < 0.0005 p < 0.0001, shown by *, **, ***, **** respectively. Only statistically significant relationships are shown. RF, riboflavin. Significance in comparison to wildtype (+/- RF) is shown as symbols as described in legend (p < 0.05). – and + indicates deletion and complementation of the corresponding gene, respectively.
Fig 4.
Quantification of ribH1, ribH2 transcript and protein levels of lumazine synthase genes in Msm strains.
(A and C) Fold changes in sigA-normalized transcript levels relative to wildtype. Transcript levels of the target genes were normalized to the housekeeping gene sigA (an essential housekeeping gene which is stably expressed) and scaled to the average of wildtype Msm to calculate the ΔΔCt and determine the fold difference in gene expression (n = 2 biological replicates, 2 technical replicates). (B and D) Protein abundance was measured using DIA proteomics (n = 3 biological replicates, 2 technical replicates). Data are shown as mean quantification ± SEM. Statistical comparisons were performed using a one-way ANOVA and Sidak’s multiple comparison test whereby statistical significance is represented by p < 0.05, p < 0.001, p < 0.0005 p < 0.0001, shown by *, **, ***, **** respectively. Only statistically significant relationships are shown. RF, riboflavin. Significance in comparison to wildtype (+/- RF) is shown as symbols as described in legend (p < 0.05). – and + indicates deletion and complementation of the corresponding gene, respectively.
Fig 5.
Quantification of ribA2, ribH and ribC transcript and protein levels in Mtb strains.
(A, C and E) Fold changes in sigA-normalized transcript levels relative to wildtype. Transcript levels of the target genes were normalized to the housekeeping gene sigA (an essential housekeeping gene which is stably expressed) and scaled to the average of wildtype Mtb to calculate the ΔΔCt and determine the fold difference in gene expression (n = 3 biological replicates, 2 technical replicates). (B and D) Protein abundance was measured using DIA proteomics (n = 3 biological replicates, 2 technical replicates). Data are shown as mean quantification ± SEM. Statistical comparisons were performed using a one-way ANOVA and Sidak’s multiple comparison test whereby statistical significance is represented by p < 0.05, p < 0.001, p < 0.0005 p < 0.0001, shown by *, **, ***, **** respectively. Only statistically significant relationships are shown. RF; Riboflavin. Significance in comparison to wild type (+/- RF) is shown as symbols as described in legend (p < 0.05). – and + indicates deletion and complementation of the corresponding gene, respectively.
Fig 6.
Quantification of DMRL in Msm and Mtb strains.
Levels of intracellular DMRL were quantified in Msm (A and B) and Mtb (C) using MRM. Data are shown as mean quantification ± SEM for three biological replicates. Quantification was carried out using chromatographic peak area of the most intense transition ion (m/z 327.1 → 193) of DMRL. Statistical comparisons were performed using a one-way ANOVA and Sidak’s multiple comparison test whereby statistical significance is represented by p < 0.05, p < 0.001, p < 0.0005 p < 0.0001, shown by *, **, ***, **** respectively. Only statistically significant relationships are shown. RF, riboflavin; Sig., significance in comparison to wild type (+/- RF) is shown as symbols, as described in the legend (p < 0.05).
Fig 7.
Impact of riboflavin pathway mutations on MR1T recognition of Msm.
MR1T cell clone (1e4) IFN-γ response to dendritic cells (1e4) incubated with Msm at a MOI of 10. Response was normalized to wildtype Msm. All strains were grown with riboflavin. Wildtype Msm and positive control (PHA) response are shown in S12 Fig. Data are representative of n = 3 independent experiments. Statistical comparisons were performed using a one-way ANOVA and Sidak’s multiple comparison test whereby statistical significance is represented by p < 0.05, p < 0.001, p < 0.0005 p < 0.0001, shown by *, **, ***, **** respectively.
Fig 8.
Differential impacts of riboflavin pathway knockouts on MR1T recognition of Mtb.
The data represent the MR1T cell clone (1e4) IFN-γ response to dendritic cells (1e4) infected at an MOI of 10 of the indicated Mtb strain. The response for each Mtb strain was normalized to the response of D454 H1-2, a classically restricted CD8 T cell clone that recognizes Mtb8.4 [22]. Data are representative of n = 3 independent experiments. Statistical comparisons were performed using a one-way ANOVA and Sidak’s multiple comparison test whereby statistical significance is represented by p < 0.05, p < 0.001, p < 0.0005 p < 0.0001, shown by *, **, ***, **** respectively.
Fig 9.
MAIT cell agonists and antagonists produced via the riboflavin pathway in mycobacteria.
Solid black arrows denote enzyme-catalyzed reactions. Dashed grey arrows denote non-enzymatic reactions. MAIT cell agonists (green) or antagonists (red) are shown in bold italics. RL, ribityllumazine. The production of photolumazines has only been observed in Msm and not in Mtb.