Characterization of the transcriptional regulator Rv3124 of Mycobacterium tuberculosis identifies it as a positive regulator of molybdopterin biosynthesis and defines the functional consequences of a non-synonymous SNP in the Mycobacterium bovis BCG orthologue

Abstract

A number of single-nucleotide polymorphisms (SNPs) have been identified in the genome of Mycobacterium bovis BCG Pasteur compared with the sequenced strain M. bovis 2122/97. The functional consequences of many of these mutations remain to be described; however, mutations in genes encoding regulators may be particularly relevant to global phenotypic changes such as loss of virulence, since alteration of a regulator's function will affect the expression of a wide range of genes. One such SNP falls in bcg3145, encoding a member of the AfsR/DnrI/SARP class of global transcriptional regulators, that replaces a highly conserved glutamic acid residue at position 159 (E159G) with glycine in a tetratricopeptide repeat (TPR) located in the bacterial transcriptional activation (BTA) domain of BCG3145. TPR domains are associated with protein-protein interactions, and a conserved core (helices T1-T7) of the BTA domain seems to be required for proper function of SARP-family proteins. Structural modelling predicted that the E159G mutation perturbs the third alpha-helix of the BTA domain and could therefore have functional consequences. The E159G SNP was found to be present in all BCG strains, but absent from virulent M. bovis and Mycobacterium tuberculosis strains. By overexpressing BCG3145 and Rv3124 in BCG and H37Rv and monitoring transcriptome changes using microarrays, we determined that BCG3145/Rv3124 acts as a positive transcriptional regulator of the molybdopterin biosynthesis moa1 locus, and we suggest that rv3124 be renamed moaR1. The SNP in bcg3145 was found to have a subtle effect on the activity of MoaR1, suggesting that this mutation is not a key event in the attenuation of BCG.

Fig. 1.

(a) Rv3124 structural model based on the EmbR molecular structure. The DNA-binding domain is in green, and the BTA domain in purple. Residue E159, located in the third α-helix of the BTA domain, is shown in red. (b) Close-up of the third α-helix of the BTA domain from a different angle. The side chains of residues that form the helix are in blue, except for E159, which is shown in red. Intra-chain H bonds are represented by green dashed lines. Replacement of Glu159 by Gly perturbs the helix.

Fig. 2.

Confirmation of microarray results by qRT-PCR. The moa1 locus showed a higher level of expression in M. bovis BCG overexpressing Rv3124 than M. bovis BCG wild-type as measured by microarray (white bars) and by qRT-PCR (black bars). Fold changes are the mean ratios±sd of gene expression from two independent experiments.

Fig. 3.

Binding of Rv3124 to the rv3108–moa1A intergenic region. (a) The moaA1 locus showing locus organization and positions of oligonucleotides used in EMSA and lacZ fusions. Primers F1 and R3 were used to generate the Rv3109pro product used in EMSA assays; primers Fg1, Fg2, Fg3 and Rev were used to generate products for lacZ-transcriptional fusions. (b) EMSA was performed in the absence of Rv3124 (lane 1) or in the presence of 0.15, 0.3, 0.6 or 1.2 μg purified Rv3124 (lanes 2–5). Competition was performed with Rv3124 (2.4 μg) with no competitor (lane 6) or with unlabelled specific competitor rv3109 probe (10-fold and 100-fold molar excess, lane 7 and lane 8, respectively). (c) β-Galactosidase activity of promoter-probe constructs in wild-type M. smegmatis mc²155 (denoted by ‘−’) and M. smegmatis overexpressing Rv3124 (denoted by ‘+’). Activity was monitored by plating strains on medium containing X-Gal and by quantitative β-galactosidase assay. Data are shown as Miller units, and are the mean±sd of three replicates. (d) Binding assay with full-length and truncated versions of the Rv3109pro probe (190 bp, 116 bp and 45 bp) in the absence of Rv3124 (lane 1) or in the presence of 0.6 μg and 1.2 μg Rv3124 (lanes 2 and 3 respectively). Decreased Rv3124 binding to the shorter probes correlates with reduced β-galactosidase activity in the corresponding lacZ transcriptional fusions.

Fig. 4.

(a) Fold change in expression of the moa1 locus in BCG strains overexpressing mutated versions of Rv3124. The expression of the moa1 locus in BCG strains complemented with the mutated Rv3124 was compared with M. bovis BCG overexpressing the non-mutated Rv3124. Data from two independent experiments for each strain were first normalized to Rv3124 and then normalized to sigA RNA levels. (b) Residues G90 and Y91 (depicted in blue) are located in one of the two β-sheets that form the DBD of Rv3124. Residue G90 seems not to be critical, explaining why its mutation had no effect on the regulator's activity. However, residue Y91 interacts with a conserved residue (H63) located in a neighbouring helix, showing why mutation of this residue in MoaR1 caused a decrease in the induction of the moa1 locus.