|Home | About | Journals | Submit | Contact Us | Français|
A key to the success of Mycobacterium tuberculosis (Mtb) is the bacteria’s ability to survive and thrive in the presence of numerous stresses mounted by the host. Small, non-coding RNAs (sRNAs) have been shown to modulate numerous stress responses in bacteria, yet to date only two studies have screened the Mtb transcriptome to identify sRNA. Their association with oxidative and acid stress has been demonstrated but the cellular function and role of these sRNAs in the pathogenesis of tuberculosis (TB) remains unknown. Here, we have identified an sRNA, ncrMT1302, in a locus involved in cAMP metabolism and demonstrate that expression of ncrMT1302 responds to changes in pH and cAMP concentration. The differential expression of ncrMT1302 observed in wild-type Mtb during growth is abolished in a strain lacking MT1302, an adenylyl cyclase encoding gene. We report that ncrMT1302 is expressed in Mtb residing in the lungs of mice during an active infection.
Mycobacterium tuberculosis (Mtb), the etiological agent of tuberculosis (TB), is one of the most successful pathogens today as it maintains a reservoir in one-third of the human population and kills close to 1.7 million each year(WHO, 2010). A key to the success of the TB bacillus is its ability to withstand stress and proliferate in the host under adverse conditions. Upon transmission from one host to another, Mtb has to survive in intra and extracellular environments, primarily in the lungs (Manabe et al., 2008), under conditions that evolve during the course of the disease. These include the acidic pH within the phagolysosomal compartment of the macrophage (Fisher et al., 2002) and the hypoxic and nutrient poor environment in granulomatous lesions (Dannenberg, 1993). Mtb is also capable of tolerating a wide range of antibiotics (Wallis et al., 1999). This resilience is largely due to the organism’s ability to sense stress conditions and mount an appropriate response. Sigma factors and two-component signal transduction systems are widely known to coordinate differential expression of transcriptional regulators in response to stress in bacteria (Sachdeva et al., 2010). Perhaps due to this, studies of molecular pathogenesis of TB have been confined to classical genetic analysis of protein encoding genes of Mtb and their function. The existence of sRNAs in Mtb was not known until 2009 (Arnvig and Young, 2009) and therefore their roles have remained poorly understood.
Many sRNAs have been identified in Escherichia coli (E.coli) and other bacteria (Vogel et al., 2003; Gottesman, 2005). While only a small number of sRNAs have been studied, regulation by sRNAs is now widely known to contribute to stress responses and consequently to virulence (Gottesman et al., 2006). In E.coli for example, 6S RNA has been demonstrated to contribute to stationary phase growth by associating with the sigma 70 RNA polymerase holoenzyme (Wassarman and Storz, 2000). These sRNAs often act by base-pairing with specific target mRNAs and altering their translation or stability, while others associate with specific proteins and modulate their function (Waters and Storz, 2009).
Unlike protein encoding open reading frames (ORF), the lack of consensus sequences for the initiation and termination of transcription of sRNAs has meant difficulty in using structured bioinformatic tools to identify them (Vogel and Sharma, 2005). The first sRNAs in E.coli were identified by chance while studying neighboring genes and others were identified in screens of multicopy plasmid screens seeking specific phenotypes (Wassarman et al., 1999). Recently however, genome-wide searches have led to the identification of a larger repertoire of sRNAs in E.coli. To date, sRNAs in Mtb have been identified through high throughput screens: recently Arnvig et al reported identification of nine novel sRNAs with sizes ranging between 20 and 70 nucleotides (Arnvig and Young, 2009), and Dichiara et al identified thirty four novel sRNAs in Mycobacterium bovis BCG and showed that a subset of these existed in Mtb (Dichiara et al., 2010).
As part of our long standing effort to identify essential genetic elements, we report identification of sRNAs of Mtb. To date, there has been no report of a role of sRNAs in the pathogenesis of TB. Here, we performed the first high throughput screen of Mtb RNA in sizes ranging from 70 to 200 nucleotides to identify novel sRNAs in this pathogen. We isolated an sRNA whose genomic locus is flanked by MT1302 and MT1303. MT1302 (Rv1264) has been reported to be an adenylyl cyclase that is active at low pH (Dittrich et al., 2006) while MT1303 (Rv1265) has been shown to be upregulated in macrophages (Hobson et al., 2002) and induced by cAMP under low oxygen conditions (Gazdik and McDonough, 2005). On studying this sRNA, we find that it is in a region previously found to be bound by a transcription factor, Cmr (cAMP and macrophage regulator), that regulates cAMP induced genes when the bacilli infect macrophages (Gazdik et al., 2009). Next, we show that its expression levels are altered by acidic pH, changes in cAMP concentration and in a mutant strain lacking a functional copy of MT1302. This suggests a role for a small, non-coding RNA in Mtb in a potential cAMP modulated regulon. In addition, using the mouse model of TB, we demonstrate that this sRNA is expressed during an active Mtb infection.
A clinical isolate of Mtb, strain CDC1551 (Valway et al., 1998), and a previously generated MT1302 transposon insertion mutant JHU1264-780 were used. The mutant has a Himar 1 transposon insertion in the TA dinucleotide located at +780 position from the putative translation start site(Lamichhane et al., 2003). Further details of the transposon insertion mutant are available in this database (http://webhost.nts.jhu.edu/target/). Strains were grown in Middlebrook 7H9 broth (Difco) supplemented with 0.5% glycerol, 10% oleic acid-albumin-dextrose-catalase (OADC) and 0.05% Tween 80 under constant shaking at 37°C. For growth phase studies, cultures at optical density (A600nm) of 0.3 were used for early exponential phase, 0.8 for exponential phase, exponential + 2 days for late exponential cultures, late exponential + 1 day for stationary phase and first day of visible cellular clumping for late stationary, clumped cultures. For stresses, 200 ml cultures at an A600nm of 0.8 were centrifuged, washed twice with phosphate buffered saline (PBS) and resuspended as described below. To induce nitric oxide stress, diethylenetriamine nitric oxide adduct (DETA-NO) was added to the broth at a final concentration of 250 μM. To starve bacilli of nutrients, cells were resuspended in PBS containing 0.05% Tween 80. To induce acid stress, cells were resuspended in 7H9 broth adjusted to a pH of 5.5 with 1M HCl. As a control, a 200 ml culture was resuspended in Middlebrook 7H9 broth. Bacilli were incubated under stress conditions for 6 hours and RNA was extracted with Trizol (Invitrogen) as described (Hatfull and Jacobs, 2000). For exposure to drugs, washed cells were resuspended in Middlebrook 7H9 broth containing 0.5 μg/ml rifampicin or 0.05 μg/ml isoniazid and incubated for 3 hours before extracting RNA. For exposure to cAMP, 110 μM dibutyryl cAMP (Sigma) was added to 200 ml cultures at specific growth phases and total RNA was extracted after 8 hours. Cultures without addition of dibutyryl cAMP were used as a control. All studies were carried out in duplicate.
5 μg total RNA was treated with 2U of TURBO™ DNA-free DNase (Applied Biosystems) according to manufacturer’s instructions for 30 minutes, followed by addition of 2U of DNase for 30 minutes. All samples were tested for residual DNA by quantitative reverse transcription PCR (qRT-PCR) and confirmed to be free of DNA if no amplified product could be detected past 35 cycles. Reverse transcription was carried out using SuperScriptIII Reverse Transcriptase (Invitrogen) and a gene specific reverse primer. qRT-PCR was performed using SYBR® Green PCR Master Mix (Applied Biosystems) in the iCycler (BioRad) and normalized to levels of sigA, a housekeeping gene whose expression has been reported to remain constant throughout growth phases (Manganelli et al., 1999). For ncrMT1302, Primer 1A: GTGACCAATCGAAATGTCCG and Primer 2A: ACATGGGCAGACCCGGCGTGA were used as forward and reverse primers respectively. Similarly, sigA-F: CTCGACGCTGAACCAGACCT and sigA-R: AGGTCTTCGTGGTCTTCGTC were used as forward and reverse primers respectively for sigA. For MT1302, Primer MT1302_F: GAACATCGACGATCTGTTGG and Primer MT1302_R: AATCTCGTCGGGGGTGAT were used. For MT1303, Primers MT1303_F: CAGGAACCGGTGTCACGTT and MT1303_R: CTGGCATCTGGTTGCTGAG were used. The average expression levels, normalized to sigA, and standard deviations were calculated using data from three technical replicates of two independent experiments. A t-test was used to compare the significance of differences between groups.
Total RNA was separated by electrophoresis on an 8% or 10% Tris-Borate EDTA (TBE)-Urea polyacrylamide gel (Sequagel, National Diagnostics). The gel was stained and viewed under UV with 0.5 μg/ml ethidium bromide to ensure equal loading. The RNA Century™ Marker (Applied Biosystems) was used as an RNA ladder to identify 5S rRNA on the polyacrylamide gel. As 5S rRNA is 115 nucleotides in length and our largest ncrMT1302 clone was 109 nucleotides in length, we used 5S rRNA as a loading control and size marker for the northern blot. Gels were then electroblotted onto Brightstar membranes (Applied Biosystems) and UV crosslinked. Membranes were incubated overnight in ULTRAhybOligo at 54°C with 0.1 μm/ml biotin-tagged probes. Membranes were washed according to manufacturer’s instructions (NorthernMax, Applied Biosystems) and the Brightstar Biotin kit (Applied Biosystems) was used for detection.
0.5 μg plasmid DNA with 0.4M NaOH and 10mM EDTA was heated at 95°C for 20 minutes and placed on ice. The sample was then dotted onto a Brightstar membrane and dried. The membrane was rinsed in 2X SSC, air dried and UV crosslinked. Prehybridization, hybridization, washing and detection were carried out according the DIG-High Prime DNA Labeling and Detection Starter Kit II (Roche). 20 ηg/ml of DIG-11-dUTP labeled DNA probes for 5S rRNA, 23S rRNA and tRNA-Serine were used.
Mtb cultures at five growth phases, determined by A600nm, were obtained, centrifuged and the bacterial pellet was resuspended in Trizol (Invitrogen). This mixture was transferred to O-ring tubes containing 0.5 ml of 0.1 mm zirconia beads (BioSpec Products). Cells were incubated at 25°C for 10 minutes and lyzed by bead-beating using a mini-beadbeater at 4,800 RPM. Lysed cells were centrifuged for 5 minutes at 13,000 RPM, the supernatant was transferred to a fresh microfuge tube and RNA was then extracted as described (Hatfull and Jacobs, 2000). RNA was resuspended in 30μl of RNase free water and stored at −80°C.
The total RNA preparation (70μg) was electrophoresed on an 8% TBE-urea polyacrylamide gel. RNA ranging from 70 to 200 nucleotides was excised and eluted in 1x TBE and extracted using DTR columns (Edge Biosystems). RNA was then cloned using the miRCat™ microRNA Cloning Kit (IDT) according to manufacturer’s instructions. Briefly, a 3′ linker was ligated to the gel-extracted RNA using T4 RNA ligase. RNA with the 3′ linker was then reverse transcribed using a 5′ primer and purified using polyacrylamide gel electrophoresis (PAGE). A second 3′ linker with a different sequence from the first linker was then ligated to cDNA using T4 RNA ligase. cDNA with unique linkers at both 5′ and 3′ ends was PAGE purified, amplified by PCR using primers complementary to both linkers and cloned into the pCR2.1 TOPO vector (Invitrogen) according to manufacturer’s instructions. Plasmid DNA was extracted using the Miniprep Kit (Qiagen) and screened for presence of cDNA corresponding to serine-tRNA, 23S rRNA and 5S rRNA using dot-blot analysis. The plasmids that lacked these sequences were chosen and sequenced using the M13 reverse primer. BlastN was used for analyzing the sequences(Altschul et al., 1997). Two DNA linkers with different but known sequences were used to capture the sRNA transcripts. Therefore, the 5′ and 3′ ends of the sRNAs could be readily identified and the orientation of each sRNA determined based on the unique sequences for the first linker at the 3′ end and the second linker at the 5′ end. Sequences of the linkers and primers can be found in the miRCat™ microRNA Cloning Kit manual (IDT DNA).
Five 4 to 5 weeks old female BALB/C mice were infected with an aerosol of Mtb CDC1551 to deliver ~103 CFU to the lungs. At 3 weeks following infection, mice were sacrificed, lungs obtained and total RNA was prepared using the Trizol (Invitrogen) extraction method as described(Hatfull and Jacobs, 2000). 8μg of total RNA from each sample was used in each lane of the northern blot.
Previous studies to identify sRNAs in mycobacteria have focused on sRNA from Mtb in the size range of 20 to 70 nucleotides (Arnvig and Young, 2009) and in the ranges of <80 nucleotides or 100 to 200 nucleotides from non-pathogenic M. bovis BCG (Dichiara et al., 2010). We created a cDNA library using RNA isolated from Mtb at the late exponential phase of growth in order to identify potential sRNA species in the size range of 70 and 200 nucleotides. We selected this size range based on sRNAs identified in E.coli which are predominantly of these sizes (Waters and Storz, 2009). As 5S rRNA falls in this range, we removed the ribosomal RNA with a complementary, single-stranded DNA oligonucleotide followed by treatment with RNaseH. RNA molecules were size fractionated and directly cloned and sequenced using the method outlined in Figure 1. Briefly, a single stranded DNA linker was ligated to the 3′ end of the RNA, the samples were then reverse transcribed and a second linker was ligated to the 3′ end of the cDNA. The 5′-ligation independent method we used ensures that RNA modified at the 5′ end would be captured in our screen. This is of particular importance as little is known about sRNA processing in Mtb and sRNA species in some organisms are known to be tri-phosphorylated at the 5′ end (Pak and Fire, 2007). cDNA with known linker sequence tags on both ends were PCR amplified and cloned into a TA cloning vector to create a cDNA library.
Initial sequencing of 276 clones revealed that the cDNA library was enriched in serine-tRNA, 23S rRNA and 5S rRNA. We excluded these undesired clones from the library with a dot blot using single stranded DNA probes complementary to serine-tRNA, 23S rRNA and 5S rRNA. Analysis of the sequences revealed 238 clones of tRNA or rRNA, 26 clones represented sequences within known open reading frames (ORFs) and 12 clones represented intergenic regions (Table 1). Sequences representing ORFs (Table 2) could be degraded mRNA or regulatory byproducts of a processed mRNA. While some sRNAs may be encoded by regions within ORFs, their characterization is often more complex as it is difficult to ascertain if they are independent transcriptional units or a regulatory side product of the mRNA encoded by that ORF. Hence these clones were not investigated further. Of the 12 clones representing intergenic sequences (Table 3), 4 were novel sRNAs identified for the first time in this study and 8 contained sequences from a region between MT1302 and MT1303 (Figure 2). The presence of homologous sRNA in M.bovis BCG was recently reported by DiChiara et. al.(Dichiara et al., 2010). We annotated this sRNA as ncrMT1302 (non-coding RNA MT1302) hereafter and selected it for further study due to its high abundance in our screen.
Next we studied the ncrMT1302 expression profile during the growth cycle of Mtb in a nutrient rich defined medium in vitro. Mtb was grown in Middlebrook 7H9 broth and total RNA was extracted at five physiologically distinct growth phases. A northern blot analysis was performed to verify expression of ncrMT1302 and the presence of any additional transcripts (Figure 3A). A band representing ncrMT1302 was detected around the 115 nucleotide 5S rRNA band, suggesting that there is a single functional species of the sRNA of a similar size. The blot also revealed an increase in the expression of ncrMT1302 from early exponential to stationary phase of growth.
To further assess this observation independently, we performed quantitative reverse transcription PCR (qRT-PCR) to determine the sRNA levels and normalized these levels to sigA, a gene that has been reported to be expressed at constant levels throughout all growth phases in culture (Manganelli et al., 1999) (Figure 3B). We observed 200 to 12,000 fold higher levels of ncrMT1302 transcript relative to that of sigA. ncrMT1302 levels were low during the early exponential phase of growth, elevated in stationary phase but highest in exponential phase and in late stationary phase, clumped cultures. ncrMT1302 expression was approximately 60 fold higher in clumped cultures compared to the early exponential phase of growth. The high levels of ncrMT1302 suggest that the sRNA may be vital for adaptation to physiological conditions unique to these growth phases.
Bacterial sRNAs have been shown to be involved in regulating response to various stresses (Gottesman et al., 2006). The increased expression of ncrMT1302 in late stationary phase, clumped cultures, a stage of growth that is defined by stress due to limiting nutrients and increased accumulation of toxic metabolic waste(Smeulders et al., 1999), led us to hypothesize that ncrMT1302 may be involved in responding to stresses that are prevalent in vivo during infection and also during exposure to antibiotics. The presence of increased levels of reactive nitrogen species (Nathan and Shiloh, 2000), limiting nutrition and low pH have been shown to prevail in tissues harboring Mtb (Dannenberg, 1993). Therefore, we exposed cultures of Mtb at the exponential phase of growth to nitric oxide, low pH and limited nutrition for six hours to assess any functional role of ncrMT1302 under these conditions. Total RNA was extracted from Mtb cultures and ncrMT1302 expression was assessed by qRT-PCR. We observed an increase and a decrease in sRNA levels under nutrient limitation and exposure to DETA-NO respectively (Figure 4A). The increase in sRNA levels in nutrient limiting conditions is consistent with the growth phase expression study in which we observed elevated levels of ncrMT1302 during stationary phase and in clumped cultures (Figure 3). The largest difference, a 7 fold decrease in ncrMT1302 levels, was observed in cultures with reduced pH. This suggests that expression or stability of the sRNA is most strongly associated with a pH stress response.
A number of drugs are used to treat TB. While the method of action of some drugs is understood, the complete mechanism has been elusive for many drugs. sRNA genes, such as micF, have been found to be upregulated on exposure to cationic peptide antimicrobials (Oh et al., 2000). Hence we sought to investigate if ncrMT1302 could be involved in responding to stresses induced during exposure to rifampicin and isoniazid as these are the two drugs that constitute the backbone of TB treatment(WHO, 2010). Total RNA was extracted from Mtb cultures exposed to the minimal inhibitory concentrations (MIC) (Lorian, 2005) of either isoniazid or rifampicin for three hours and ncrMT1302 expression was assessed by northern blot analysis (Figure 4B). RNA isolated from Mtb cultured in the absence of these drugs was used as a negative control and 5S rRNA was used as a loading control. Northern blots revealed a diminished level of expression of ncrMT1302 upon exposure to rifampicin while there was no change on exposure to isoniazid. This was independently confirmed by qRT-PCR (data not shown). Rifampicin is known to bind the RpoB subunit of RNA polymerase and inhibit new transcription (Campbell et al., 2001). Therefore the quantity of any RNA species in a cell following exposure to rifampicin is expected to remain unchanged and any changes due to degradation would reflect its stability. The relative decrease of ncrMT1302 compared to 5S rRNA indicates that it is less stable than the ribosomal RNA.
sRNAs, such as GadY, regulate genes at the same genetic locus (Opdyke et al., 2004). The two genes flanking the ncrMT1302 locus are MT1302, which encodes an adenylyl cyclase that converts ATP to cAMP (Dittrich et al., 2006), and MT1303, which encodes a hypothetical protein whose transcription is induced by cAMP under low-oxygen conditions (Gazdik and McDonough, 2005) (Figure 7). Therefore, it is reasonable to postulate that the locus may be involved in cAMP metabolism. We thus hypothesized that ncrMT1302 expression could be affected by changes in cAMP concentration. To study this, we exposed cultures of wild-type (WT) Mtb to 110 uM dibutyryl-cAMP (dbcAMP) as described (Gazdik and McDonough, 2005) for 8 hours, using cultures at the same growth phase but incubated in 7H9 broth without dbcAMP as controls. Total RNA was extracted from these cultures and ncrMT1302 transcript levels were studied using qRT-PCR. During early exponential phase, the level of ncrMT1302 was low and remained unchanged on exposure to dbcAMP. ncrMT1302 levels were elevated during exponential phase and the addition of dbcAMP reduced transcript levels 5 fold, while during stationary phase, when ncrMT1302 expression was lower, dbcAMP increased transcript levels 5 fold (Figure 5A). Hence, cAMP supplementation results in the same fold change but is able to both increase and decrease ncrMT1302 levels. Furthermore, ncrMT1302 levels in untreated exponential phase cultures and cAMP supplemented stationary cultures are similar, suggesting that an optimal cellular concentration of the sRNA is required for it to function. This data implies that there is specific cAMP moderated regulation of ncrMT1302 expression.
cAMP is involved in the function and expression of the two genes adjacent to ncrMT1302 (Dittrich et al., 2006; Gazdik et al., 2009). Having identified cAMP as a modulator of ncrMT1302 levels, we speculated that ncrMT1302 may be functionally linked to genes at this locus. MT1302 encodes the only adenylyl cyclase in Mtb that functions selectively under low pH conditions (Dittrich et al., 2006). Since we observed low pH to affect ncrMT1302 expression, we hypothesized ncrMT1302 to be functionally related to MT1302. To explore this, we compared levels of ncrMT1302 between WT Mtb and a transposon mutant strain lacking a functional copy of MT1302 (MT1302::Tn). As the mutant harbors a Himar 1 transposon insertion at the +780 nucleotide position from the putative translation start site (Lamichhane et al., 2003) and is 408 nucleotides downstream of the ncrMT1302 locus, it is sufficient to disrupt MT1302 function without interrupting the small RNA gene. This transposon has, on average, 30 stop codons in all frames, and therefore effectively disrupts MT1302 translation(Rubin et al., 1999). We grew cultures of both strains and compared ncrMT1302 expression levels at the following four growth phases: early exponential phase, exponential phase, stationary phase and clumped cultures. Total RNA was extracted and assessed using northern blotting and qRT-PCR. The qRT-PCR results show that ncrMT1302 was expressed in the mutant lacking MT1302 but there was no significant differential expression over different phases of growth (Figure 5B). Furthermore, the level of expression is lower compared to that of the WT strain. While the level of expression of ncrMT1302 at early exponential and stationary phases were similar in the WT and MT1302::Tn strains, we observed a 5 fold increase in expression in the WT compared to the MT1302::Tn mutant during the exponential phase and in clumped cultures. The reduced expression of the sRNA in the mutant suggests that MT1302 may be functionally linked to ncrMT1302.
Macrophages infected with Mtb use acidified phagosomes as a form of defense against the bacteria (Rohde et al., 2007). Since we observed a reduced expression of ncrMT1302 during acidic stress we hypothesized that the sRNA could have a functional role during infection. To test this hypothesis, we isolated total RNA from the lungs of naïve mice (negative control) and mice infected with Mtb, and performed northern blotting analysis with a probe against ncrMT1302. As Mtb RNA encompassed only a small fraction of the RNA extracted, we could assume that by loading the same amount of total RNA, there would be similar levels of host RNA in both uninfected and infected tissue derived RNA. 5S rRNA was used as a positive control for the presence of bacterial RNA and was detected only in the samples obtained from the lungs of Mtb infected mice. The analysis however detected an abundance of ncrMT1302 RNA in the lungs of infected mice, indicative of a role of this sRNA during infection in the murine model of TB (Figure 6). While 5s rRNA and ncrMT1302 were not detected in the total RNA samples from uninfected mice, the abundance of eukaryotic RNA resulted in nonspecific bands.
The intergenic sequences of many bacteria have now been demonstrated to contain sequences that encode small, regulatory RNAs which function in virulence, cell cycle and numerous stress responses (Livny and Waldor, 2007; Padalon-Brauch et al., 2008). The first study to predict Mtb sRNA genes was carried out by Livny et al who predicted 56 sRNA genes in Mtb through the sRNAPredict2 program(Livny et al., 2006). However none of these were experimentally verified or identified in the studies that followed suggesting that bioinformatics analysis alone was not sufficient to identify sRNA genes in Mtb. While our studies were underway, Arnvig et al reported identification of 9 sRNA species through a cloning scheme in Mtb in the 20 to 70 nucleotide range (Arnvig and Young, 2009). They demonstrated that 3 of these sRNAs were induced by oxidative stress and a fourth, F6, was induced by both oxidative and acid stress. Overexpression of two of these sRNAs proved lethal in Mtb and over expression of F6 resulted in a slow growth phenotype. This indicates an important role for sRNAs in Mtb physiology and also suggests that sRNAs could be involved in regulation of a more global stress response as indicated by the response of F6 to multiple stresses. Soon thereafter, Dichiara et al identified sRNAs in M.bovis BCG, a non-virulent mycobacteria. Using a screen similar to the one reported here together with computational prediction, Dichiara et al reported the discovery of 34 novel sRNAs. This included an approximately 120 base sequence, Mcr11, from the M.bovis BCG transcriptome that is identical in sequence to the ncrMT1302 RNA we report here. Our annotation is based on the existing unique annotation of the upstream open reading frame. They demonstrated that Mcr11 was expressed in an in vitro culture of Mtb and its expression increased from late log to stationary phase of growth.
Here, we identify five small RNA candidates through a high throughput screen of an Mtb small RNA library consisting of RNA between 70 and 200 nucleotides (Table 1). Eight independent clones in our sRNA library harbored the ncrMT1302 sequence, while other sRNA candidate sequences were observed only once. We propose two likely explanations for this observation: (a) ncrMT1302 is stable with a half-life longer than other sRNA, or (b) it is expressed in higher abundance. The quantitative assessments we carried out show that ncrMT1302 is indeed expressed at high levels in Mtb relative to sigA and that ncrMT1302 is differentially expressed, with levels peaking in clumped cultures which represent late stationary phase. Our results correlate with previous reports as small regulatory RNAs have been implicated in the presence of environmental stresses during growth and this is most evident during stationary phase (Gottesman et al., 2006).
Northern blot and qRT-PCR analyses identified a low pH and rifampicin exposure as conditions that affect ncrMT1302 levels in the cell. This is consistent with published reports that sRNAs function to coordinate stress responses in bacteria (Gottesman et al., 2006). A reduced pH is faced by the bacterium in the macrophage phagosome, suggesting that ncrMT1302 could play a role in sensing and responding to pH changes. This is of particular importance as Mtb must be able to counteract acidification of phagolysosomes during host invasion (Pethe et al., 2004). The smaller expression changes observed under nutrient limitation and DETA-NO treatment could be due to downstream effects of these specific stress responses and not directly linked to ncrMT1302 function. Furthermore, the increase in ncrMT1302 levels under nutrient limitation could explain the elevated levels of the sRNA we observed in clumped cultures. Lastly, as rifampicin inhibits synthesis of new RNA, the reduced level of ncrMT1302 is an indication of its stability relative to 5S rRNA whose levels remain relatively high upon rifampicin exposure (Campbell et al., 2001).
MT1302 encodes an adenylyl cyclase with an auto-inhibitory N-terminal domain (Tews et al., 2005) that makes it the only adenylyl cyclase in Mtb to function selectively at a low pH. MT1303 encodes a hypothetical protein that is upregulated during macrophage infection (Monahan et al., 2001) and by acidic pH in M. bovis BCG but not in Mtb (Gazdik et al., 2009). Our observation that ncrMT1302 levels are sensitive to low pH correlates well with the hypothesis that ncrMT1302 could be functionally linked to genes at this locus. cAMP is another modulator of this locus-with MT1302 being an adenylyl cyclase and MT1303 being a cAMP-induced gene under low oxygen conditions (Gazdik and McDonough, 2005). Our study demonstrates that cAMP supplementation modulates ncrMT1302 levels and this too is consistent with the three genes at this locus being functionally linked. This hypothesis is further supported by the reduced expression and loss of differential expression of ncrMT1302 in the absence of a functional MT1302 gene.
cAMP regulates expression of numerous genes in Mtb by binding to cNMP binding transcription factors. The best characterized of these is the CRP (cAMP receptor protein) transcription factor (McCue et al., 2000). cAMP binding enhances its DNA binding properties (Bai et al., 2005) enabling it to both positively and negatively regulate genes in Mtb such as whiB1, which has an activating and repressing CRP binding site (Stapleton et al., 2010). This paradigm supports the cAMP induced changes we observed in ncrMT1302 expression. Although we found no CRP binding domains upstream of ncrMT1302, there are 11 proteins with cNMP binding domains in Mtb(McCue et al., 2000) that could potentially regulate expression of ncrMT1302 in response to cAMP. A second cAMP associated transcription factor is Cmr (cAMP and macrophage regulator) (McCue et al., 2000). A previous study has reported that Cmr induces MT1303 transcription by binding to the upstream region of MT1303 (Gazdik et al., 2009). The study identified a binding region of Cmr using a gel shift assay. We mapped homology to this sequence to derive the coordinates of this site in Figure. 7 and show that this region significantly overlaps the ncrMT1302 sequence. As such the ncrMT1302 RNA sequence is highly homologous to the Cmr DNA binding sequencing, suggesting that Cmr could be a potential binding target for the ncrMT1302 sRNA.
Herein, we have assigned a stress response to an sRNA in Mtb by demonstrating a reduced expression of ncrMT1302 under low pH condition. This is the first report to identify a potential functional linkage of an sRNA in Mtb and demonstrate expression of an sRNA in vivo using the mouse model. As cAMP contributes to virulence and a pH stress response is vital for the survival of the bacillus, this study demonstrates a new avenue for studying TB disease through sRNAs.
We thank Hani Zaher for expert advice on RNA related studies, Maia Schoonmaker for providing RNA from infected animals and Cara Smith for assistance with cDNA library screening and dot-blot analysis. A gift of the miRCAT™ RNA cloning Kit from Integrated DNA Technologies (IDT) is appreciated. This work was funded by NIH/NIAID AI30036, AI079590, AI37856, AI36973 and DP2OD008459. SP, WRB and GL planned the studies, SP conducted them, SP, WRB and GL analyzed the data and SP and GL wrote the manuscript.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.