The role of transcriptional regulation in maintaining the availability of mycobacterial adenylate cyclases

Bacterial culture

M. smegmatis mc²155 was grown in Tryptic Soy Broth (TSB) supplemented with 0.05% Tween-80. Cultures were grown in ambient air or 5% CO₂ at 37°C in 25 cm² tissue culture flasks rocking with gentle agitation. For gene regulation assays, late log phase cultures were exposed to low pH (TSB adjusted to pH 5.5 with 0.1 M HCl), starvation (incubation in Phosphate Buffered Saline), hydrogen peroxide (5 mM), or nitric oxide (10 mM diethylenetriamine/nitric oxide adduct) for 4 h and gene expression was compared to non-exposed cultures.

RNA preparation

Late log phase culture of M. smegmatis mc²155 was pelleted and resuspended in RNase-free water. Cells were mechanically disrupted using a bead beater (BioSpec Products) for four rounds of beating on high for 1 min each, in a mixture of 0.1 mm zirconia-silica beads (BioSpec Products), 45% TRIzol (Invitrogen), 45% acid phenol, and 10% chloroform-isoamyl alcohol (24:1). RNA was precipitated with isopropanol/3 M sodium acetate (pH 5.2) and resuspended in RNase-free water. RNeasy Mini Kit and RNase-free DNAse (Qiagen) were used to remove contaminating DNA following manufacturer’s specifications.

Semi-quantitative RT-PCR

cDNA was prepared from 0.5 µg of RNA using the iScript cDNA synthesis kit according to the manufacture’s specifications (BioRad). PCR was run using a series of cDNA dilutions (0–1:1000) as templates to ensure reactions chosen for quantitation were in the linear range of the PCR (Table 1 for primers). Reactions were performed at 94°C for 1 min, 57°C for 1 min and 72°C for 1 min followed by a 10 min extension at 72°C. Control reactions were performed against 16S rDNA using cDNA diluted 10⁻⁵. PCR products were separated on agarose gels and band densities quantified using ImageJ software (Abramoff, Megalhaes & Ram, 2004). The 16S PCR products from all growth conditions were normalized to one another before quantitation of individual genes, to ensure equal levels of starting RNA in each reaction. 16S rDNA PCR was also performed using total RNA without reverse transcription to ensure the absence of DNA contamination.

Gene reporter construction and assay

Promoter:gfp reporter strains were generated for gene expression analysis of M. tuberculosis promoter regions in a M. smegmatis background. The intergeneic DNA sequences of the adenylate cyclase genes were amplified by PCR (Table 1 for primer sequences) and amplified DNA was cloned into pGFPoriM, which carries a promoterless gfpmut2 gene as previously described (Purkayastha, McCue & McDonough, 2002; Florczyk et al., 2003). Constructed plasmids were electroporated into M. smegmatis mc²155 at 2500 mV (Eppendorf 2510 Electroporator). GFP fluorescence from cultured cells was detected using GloMax Multi + Detection System (Promega) and normalized to 10⁶ bacteria based on OD₆₀₀.

Regulation of M. smegmatis adenylate cyclases

The genome of M. smegmatis contains ten annotated adenylate cyclases, eight of which have orthologs in pathogenic M. tuberculosis (Kapopoulou, Lew & Cole, 2011). In order to determine the role of gene expression in adenylate cyclase availability we systematically examined transcription of all eight orthologs using semi-quantitative RT-PCR. Gene expression was examined with a focus on M. tuberculosis orthologs, using a variety of environments known to be relevant for M. tuberculosis infection. Conditions examined include starvation, low pH, oxidative and nitrosative stress and 5% CO₂. Evidence suggests that non-growing persistent M. tuberculosis is exposed to nutrient starvation in lung granulomas (Betts et al., 2002; Nyka, 1974) while low pH, oxidative and nitrosative environments are known to occur following macrophage phagocytosis of M. tuberculosis (Smith, 2003; Liao, Fan & Bao, 2013; Chan et al., 1992).

Expression under control conditions (growth in TSB + Tween-80) indicated that all eight adenylate cyclase genes are transcribed at the same time in the cell, albeit with varied levels of transcription (Fig. 1). Out of all conditions tested, starvation and oxidative stress (H₂O₂ exposure) affect gene expression the most, with three genes showing statistically significant changes in expression under starvation conditions and two demonstrating statistical changes in expression under oxidative stress. This enhances the evidence of cAMP-mediated gene regulation in nutrient-limiting conditions as 30% of the M. tuberculosis CRP_Mt regulon has previously been linked to nutrient starvation (Bai, McCue & McDonough, 2005). Two genes (MSMEG_0545 and MSMEG_4279) were upregulated 2–3 fold after 4 h of starvation while one gene (MSMEG_3780) was observed to be downregulated approximately 2 fold after starvation and exposure to low pH. Additionally, two genes (MSMEG_0545 and MSMEG_4924) were downregulated approximately 2 fold under oxidative stress (Figs. 1 and 2). No expression differences were seen during nitrosative stress or the presence of CO₂ (data not shown).

Interestingly, the expression of only four of the eight adenylate cyclases was influenced by the environments examined, and those observed expression changes, though significant, were not dramatic. Combined with the result that all eight genes were transcribed at various levels in all environments tested we conclude that our results counter our hypothesis and indicate that transcriptional regulation plays only a minor role in controlling the availability of various adenylate cyclases in the cell. Three of the four genes with environmentally-altered expression encode for soluble adenylate cyclase proteins suggesting that transcriptional regulation may play a larger role in the availability of soluble cyclases as opposed to the membrane associated and multi-domain structures. It is likely that biochemical regulation of various protein domains has the dominant role in regulation of cAMP production by the redundant cyclases.

Regulation of M. tuberculosis adenylate cyclases

Of the ten annotated M. smegmatis cyclases, eight have reported orthologs in M. tuberculosis and M. bovis, seven in M. marinum, and five in M. avium (Kapopoulou, Lew & Cole, 2011). Phylogenic analysis based on protein sequence alignment was completed using Phylogeny.fr, online software that combines MUSCLE for multiple alignment, PhyML for tree building, and TreeDyn for tree rendering (Dereeper et al., 2008). This analysis confirmed the similarities of the M. tuberculosis and M. smegmatis ortholgos (Fig. 3). Additionally, promoter regions of each ortholog pair, represented by the 500 nucleotides upstream of the putative ATG start site of each gene, were compared for sequence similarities using the sequence alignment program T-Coffee (Notredame, Higgins & Heringa, 2000; McWilliam et al., 2013). Percent identities ranged from 51.94% to 82.08% (Table 2), indicating enough similarity to predict regulation would be similar between orthologs.

In order to determine if regulation in M. smegmatis was similar to that of the pathogenic orthologs we generated gfp:promoter fusions using the intergenic regions amplified from H37Rv chromosomal DNA. Overall, regulation between species orthologs was very similar. Rv1647 showed downregulation in both starvation and low pH similarly to MSMEG_3780, supporting the observation by Dass et al. (2008) who reported comparable regulation (Fig. 4). Additionally, Rv1359 showed similar regulatory patterns of upregulation in starvation and downregulation in oxidative stress as did MSMEG_0545 (Fig. 4). Oxidative stress and starvation are key environments for M. tuberculosis infection and latency respectively. Interestingly, Rv1359 is one of five H37Rv annotated adenylate cyclases that has not been shown to be biochemically active. The catalytic mechanism of class III adenylate cyclases is predicted to be a 2-metal ion reaction similar to DNA polymerase, where one metal ion, typically magnesium, associates with ATP and the second is involved in the catalysis reaction (Zimmermann, 1998). The Rv1359 protein sequence contains a glycine residue in place of the first of two required metal binding aspartates, along with missing a critical arginine residue, making it likely that this protein cannot function as an enzymatically active homodimer (Shenoy et al., 2004). Blastp alignment of the orthologs Rv1359 and MSMEG_0545 indicates that MSMEG_0545 is also missing the first aspartate residue, containing an arginine in this location instead, suggesting that MSMEG_0545 may also be catalytically inactive.

Transcriptional regulation of Rv1359 and MSMEG_0545 suggests that these genes do encode functional proteins. Guo et al. (2009) identified interactions between the Rv1359 protein and the promoter regions of rubredoxin encoding genes rubA and rubB and fatty acid metabolic genes fadD26, fadE26, and fadE27 (Guo et al., 2009). These genes encode proteins that may function in oxidative stress or starvation conditions. Rubredoxins are iron containing proteins involved with electron transfer, which have been shown to be upregulated in phagocytized M. tuberculosis (Buchko et al., 2011), an environment known to expose bacterial cells to oxidative stress. FadD26 is a fatty acid synthetase that is downregulated during starvation of M. tuberculosis (Betts et al., 2002) while FadE26 and FadE27 are involved in lipid degradation, a process likely to be important for phagocytized mycobacteria which rely on fatty acids for energy in the phagosome (Schnappinger et al., 2003). We hypothesize that Rv1359, and likely MSMEG_0545, are transcriptional regulators which could play roles in mediating a cAMP signal under oxidative stress and starvation. These proteins may function individually or in conjunction with other proteins as fadD26 has already been identified as part of the CRP_Mt regulon (Rickman et al., 2005).

The last regulatory similarity observed was the oxidative stress regulation of both MSMEG_4924 and the Rv1320c promoter region. MSMEG_4924 has three predicted orthologs in M. tuberculosis, Rv1318c, Rv1319c and Rv1320c. Rv1319c and Rv1320c comprise an operon and are both represented by the Rv1320c promoter region. The regulatory similarity of MSMEG_4924 to Rv1319c/Rv1320c and not Rv1318c correlates with the higher percent similarity to the promoter (82.08%) compared to Rv1318c’s promoter (54.6%) (Table 2).