Proteasomal control of cytokinin synthesis protects Mycobacterium tuberculosis against nitric oxide

Abstract

One of several roles of the Mycobacterium tuberculosis proteasome is to defend against host-produced nitric oxide (NO), a free radical that can damage numerous biological macromolecules. Mutations that inactivate proteasomal degradation in Mycobacterium tuberculosis result in bacteria that are hypersensitive to NO and attenuated for growth in vivo, but it was not known why. To elucidate the link between proteasome function, NO resistance, and pathogenesis, we screened for suppressors of NO hypersensitivity in a mycobacterial proteasome ATPase mutant and identified mutations in Rv1205. We determined that Rv1205 encodes a pupylated proteasome substrate. Rv1205 is a homolog of the plant enzyme LONELY GUY, which catalyzes the production of hormones called cytokinins. Remarkably, we report that an obligate human pathogen secretes several cytokinins. Finally, we determined that the Rv1205-dependent accumulation of cytokinin breakdown products is likely responsible for the sensitization of Mycobacterium tuberculosis proteasome-associated mutants to NO.

Figure 1. Identification of NO-resistant suppressor mutants

(A) Genomic regions surrounding transposon insertions that suppressed the NO-sensitive phenotype of a Δmpa::hyg mutant. Black arrows on triangles represent the direction of neo expression in the transposon. Gene data are from http://genolist.pasteur.fr/TubercuList/. (B) NO susceptibility and complementation of suppressor mutants. All strains harbor an integrated pMV306.strep vector with or without the indicated gene. (−) indicates empty vector. Colony forming units (CFU) of surviving bacteria after exposure to acidified media without (grey bars) or with (black bars) NaNO₂. Open bars show input CFU. Data show means ± standard deviation (SD), n = 3, from one of three independent experiments with similar results. All samples exposed to NaNO₂ were compared to wt with two-tailed Student’s t-test; * p < 0.05; ** p < 0.01; ns: not statistically significantly different. (C) Growth curves comparing wt, mpa and sup1–4 strains in 7H9 broth. (D) NO susceptibility assay as described in (B) of the Rv1205 mutant and the complemented strain. (E) and (F) Disruption of Rv1205 partially restored virulence to an mpa mutant. Bacterial CFU after infection of wt C57BL6/J mice with wt, mpa, sup1 and Rv1205 strains. CFU from the lungs (E) and spleens (F) from two independent experiments at days 1 (n = 6), 21 (n = 8) and 49 (n = 8). Data were analyzed for statistical significance by a two-way ANOVA followed by a Bonferroni post-test. For spleens: day 21 CFU were not significantly different; day 49, p <0.05 between mpa and sup1. For lungs: day 21 and day 49 p <0.001 and p <0.01, respectively, between mpa and sup1.

Figure 2. Rv1205 is a newly identified proteasome substrate

(A) Immunoblot for GlpK in total cell lysates of wt, prcBA, pafA, mpa, and sup4 strains. (B) Immunoblot for Rv1205 (20 kD) in total cell lysates of wt, mpa, pafA, and prcBA strains. (C) Quantitative-real time-PCR (qRTPCR) shows Rv1205 transcripts were unchanged in the mpa mutant when compared with wt M. tuberculosis. Fold changes relative to wt M. tuberculosis are plotted on the y-axis. This experiment represents one biological replicate performed in triplicate. Data show means ± SD. dlaT (dihydrolipoamide acyltransferase) was used as a control for a gene whose expression does not change in PPS mutants. (D) In vitro pupylation assay of recombinant histidine-tagged Rv1205 and Rv1205_K74A. See also Figure S1. Proteins were detected with polyclonal antibodies to Rv1205-His₆. (E) Immunoblot for Rv1205 in total cell lysates of wt, mpa, and Rv1205 strains harboring integrated vector with or without the Rv1205 gene. (−) indicates empty vector. The loading control for all immunoblots is DlaT. For all panels the molecular weight (MW) markers are indicated on the left. The asterisk represents a cross-reactive protein in panels (B) and (E).

Figure 3. Rv1205 is a homologue of LOG

(A) LOG activity and structures of select cytokinins. R = isoprene-derived or aromatic side chain. iP: isopentenyladenine; tZ: trans-zeatin. p-topolin: para-topolin. (B) Comparison of the crystal structures of proteins from M. marinum (left, PDB code 3SBX) and A. thaliana (right, PDB code 1YDH). The two subunits in the homodimers are colored green and yellow. AMP molecules in the M. marinum structure are shown in stick representation with carbon atoms colored orange, oxygen atoms red, nitrogen atoms blue, and phosphorus atoms black. (C) Phylogenetic analysis of LOG-like genes. See also Figure S2. Clades with bootstrap values greater than 90% have a red circle. Well-supported monophyletic clades are shown as filled triangles. Predicted operons and domain architectures are illustrated to the right. Names of the LOG-like genes are indicated below.

Figure 4. Rv1205 is a phosphoribohydrolase

(A) Rv1205-His₆ phosphoribohydrolase activity towards the substrates iPRMP and tZRMP as measured by HPLC, (means ± SD, n = 3). ATP and AMP were used as controls. (B) and (C) Determination of Rv1205 kinetic parameters for iPRMP (B) and AMP (C) conversion to iP and adenine, respectively. (D) Three-dimensional model of Rv1205 constructed using SWISS-MODEL (Arnold et al., 2006) based on M. marinum PDB ID: 3SBX, which has a sequence identity of 84% to Rv1205. The viewing angle is the same as in Figure 3B. The ribbon diagram is centered on the presumed active site with iPRMP placed according to the position of AMP in the M. marinum structure. Rv1205 is predicted to be a homodimer (subunits are colored green and yellow). Selected side chains are shown in stick representation. Atom colors: carbon in Rv1205 (green) or iPRMP (orange), oxygen (red), nitrogen (blue), sulfur (yellow) and phosphorus in iPRMP (black). (---) indicates potential hydrogen bonding. (E) Separation by native gel electrophoresis of recombinant wt or mutant Rv1205-His₆ proteins. (F) Activity of point mutant alleles of Rv1205-His₆ with indicated substrates. See also Figure S4. Data show means ± SD, n = 3. iPRMP: isopentenyladenosine monophosphate; iP: isopentenyladenine; tZRMP: trans-zeatin-riboside monophosphate; tZ: trans-zeatin; ATP: adenosine triphosphate and AMP: adenosine monophosphate.

Figure 5. Site-directed mutagenesis of Rv1205 abrogates NO sensitization in M. tuberculosis

Point mutant alleles of Rv1205 could not fully restore NO-sensitivity to the sup1 mutant. Data show mean ± SD, n = 9 (upper panel). All samples exposed to NaNO₂ were compared to wt with a two-tailed Student’s t-test and statistical differences are indicated with * p < 0.05; *** p < 0.001. ns: not statistically significantly different. Samples showing statistical difference to wt were also compared to mpa and significant differences are similarly indicated; these data suggest partial complementation. See also Figure S5. Immunoblot for Rv1205 (20 kD) in total cell lysates of indicated strains. DlaT is the loading control. MW markers are indicated on the right. The asterisk represents a cross-reactive protein. A: alanine; M: methionine; R: arginine; D: aspartate; E: glutamate; T: threonine.

Figure 6. Aldehyde accumulation in M. tuberculosis causes NO sensitivity

(A) Metabolomics analysis revealed several molecules at elevated levels in an mpa mutant. Results of relevant compounds from cell lysates of wt, mpa, sup1 and log M. tuberculosis strains (n = 4). HODE: hydroxy octadecadienoic acid. See also Tables S3 and S4. (B) Detection of pHBA, a degradation product of p-topolin, in M. tuberculosis lysates. Absorption spectra of the 4-aminophenol assay, which detects aldehyde release, are shown. p -topolin without M. tuberculosis lysate ( ), 500 μM p-topolin with M. tuberculosis lysate (●), and 300 μM pHBA with M. tuberculosis lysate was used as a positive pHBA control (▲). M. tuberculosis lysate was used as a blank for the latter two experiments. (C) pHBA sensitized wt M. tuberculosis to NO in a dose-responsive manner. Data show mean ± SD, n = 3. (D) 2-methyl-3-butanal sensitized wt M. tuberculosis to NO in a dose-responsive manner. Data show mean ± SD, n = 3. Also See Figure S6.