Rv1675c (cmr) regulates intramacrophage and cAMP-induced gene expression in Mycobacterium tuberculosis-complex mycobacteria

Abstract

Cyclic AMP (cAMP) has recently been shown to be a global regulator of gene expression in Mycobacterium tuberculosis (Mtb). In this study we identified a new cAMP-associated regulon in Mtb and M. bovis BCG, which is distinct from the previously described CRP_Mt regulon. Proteomic comparison of wt M. bovis BCG with a Rv1675c (cmr) knockout strain showed dysregulated expression of four previously identified proteins encoded by the cAMP induced genes (cAIGs) mdh, groEL2, Rv1265 and and PE_PGRS6a. Regulated expression of these four cAIGs also occurred during macrophage infection, and this regulation required cmr in both Mtb and M. bovis BCG. Purified His-Cmr bound to the DNA sequences upstream of three cAIGs (mdh, groEL2, Rv1265) in electrophoretic mobility shift assays, suggesting direct regulation of these genes by Cmr. We also found that low pH stimulated cAMP production in both Mtb and M. bovis BCG, but broadly affected cAIG regulation only in M. bovis BCG. These studies identify Cmr as a transcription factor that regulates cAIGs within macrophages, and suggest that multiple factors affect cAMP-associated gene regulation in TB-complex mycobacteria. cAMP signaling and Cmr-mediated gene regulation during Mtb infection of macrophages may have implications for TB pathogenesis.

Introduction

Tuberculosis (TB) is a serious global health problem, with one third of the world’s population currently infected with Mycobacterium tuberculosis (Mtb), and a new case rate of approximately 8 million per year (WHO Report 2007, http://www.who.int/tb/en/). Infection is further complicated by increasing levels of drug resistance and a deadly synergy with human immunodeficiency virus (HIV) (Nunn et al., 2005). A better understanding of the mechanisms by which Mtb senses and responds to its environment during infection may contribute to the identification of improved antitubercular drug candidates.

cAMP has recently been shown to be a global regulator of gene expression in Mtb, and it is likely to play a role in Mtb-host interactions, including virulence (Agarwal et al., 2006; Gazdik and McDonough, 2005; Lowrie et al., 1975; Lowrie et al., 1979; Rickman et al., 2005). A wide range of cellular responses are controlled by cAMP levels in both prokaryotic and eukaryotic cells (Botsford and Harman, 1992; Peterkofsky et al., 1993; Tang and Hurley, 1998). cAMP signaling is also critical for virulence in a number of protozoan, fungal and bacterial pathogens (Alspaugh et al., 1997; Alspaugh et al., 2002; Gross et al., 2003; Liebmann et al., 2003; Petersen and Young, 2002). In particular, the recent identification of cAMP as a regulator of the virulence-associated plasminogen activator gene, pla, in Y. pestis (Kim et al., 2007), and type III secretion systems in both Pseudomonus aeruginosa and Yersinia enterocolitica (Petersen and Young, 2002; Smith et al., 2004), suggests that the importance of cAMP signaling to bacterial pathogenesis is greater than previously recognized.

The classical model for cAMP regulation in prokaryotes is based on the well characterized cAMP response in Escherichia coli (Botsford and Harman, 1992). E. coli contains a single class I adenylate cyclase (AC) which catalyzes the conversion of ATP to cAMP (Barzu and Danchin, 1994; Botsford and Harman, 1992). The cAMP signal is transduced in E. coli by the cAMP receptor protein (CRP). CRP undergoes a conformational change upon binding cAMP, which activates it as a transcription factor (Botsford and Harman, 1992). By comparison, the Mtb genome contains 15 putative class III ACs (McCue et al., 2000), 10 of which have confirmed biochemical activity (Abdel Motaal et al., 2006; Castro et al., 2005; Guo et al., 2001; Linder et al., 2002; Reddy et al., 2001; Shenoy and Visweswariah, 2006; Sinha et al., 2005; Tews et al., 2005). M. bovis has a similar number of cyclase genes, but the genomes of other Actinobacteria, including Corynebacterium and Streptomyces each contain only a single adenylyl cyclase gene (Shenoy et al., 2004). The surprisingly high number of ACs suggests a large role for cAMP signal transduction in Mtb’s gene regulation. The cAMP signaling paradigm in Mtb is also likely to be far more complex than the classical E. coli model, due to this large number and diversity of AC proteins. However, very little is currently known about cAMP signaling pathways in Mtb.

Two of Mtb’s ten predicted nucleotide binding proteins, CRP_Mt (also called Rv3676) and Rv1675c (referred to hereafter as Cmr, for cAMP and macrophage regulator), were initally classified as members of the CRP/FNR transcriptional regulatory factor family, based on in silico analyses (McCue et al., 2000). Deletion of crp attenuates Mtb virulence in a murine model, although in vitro growth of these mutants is also reduced (Rickman et al., 2005). CRP_Mt’s DNA and cAMP binding properties have recently been experimentally confirmed, and we identified a potential CRP_Mt regulon of 114 members (Bai et al., 2005; Bai et al., 2007). We also separately identified a set of five cAMP-induced genes (cAIGs) in Mtb using an exogenous cAMP supplementation model (Gazdik and McDonough, 2005). However, none of these cAIGs belongs to the predicted CRP_Mt regulon. In this study we show that four cAIGs are regulated by Mtb’s second predicted CRP-like protein, Cmr. Elevated cAMP levels and intramacrophage conditions were identified as key cAIG regulatory signals in both virulent Mtb and attenuated M. bovis BCG.

Results

cmr is required for regulation of cAMP-induced protein expression

We reasoned that a second putative cAMP-responsive transcription factor, Cmr, might regulate expression of cAIGs that did not belong to the predicted CRP_Mt regulon (Bai et al., 2005). We tested this possibility using a proteomic approach. BCG is often used as a model for gene expression in TB complex bacteria, but several minor differences in cAMP-associated gene regulation between BCG and Mtb have been recently reported (Bai et al., 2007; Hunt et al., 2008). Such differences can be informative with respect to the identification of virulence-associated gene expression, so we also made comparisons between BCG and Mtb throughout these studies.

cmr deletion mutants were generated in BCG and Mtb, using homologous recombination to replace portions of the cmr coding region with a kanamycin resistance marker (Fig. 1). For both BCG and Mtb cultured in mycomedia with ambient air conditions at 37°C, the mutant strains grew at rates similar to those of wt (data not shown, Fig. S1). Therefore, cmr is not essential for growth in standard laboratory conditions.

Figure 1.

cmr knockout construction and verification. (A) Schematic representation of cmr DNA regions in wt and knockout strains of Mtb H37Rv and BCG. Bars under the genes denote the regions of DNA probes used for analysis in panel B. (B) Confirmation of cmr disruption by Southern blot analysis using probes shown in panel A.

Nine of eleven previously identified cAMP-responsive protein spots (Gazdik and McDonough, 2005) were evaluated in the BCGΔcmr proteome using 2D-GE to identify cmr-dependent differences. Proteomes were compared following growth in the presence or absence of exogenously added dibutyrl cAMP (dbcAMP). Four cAMP-dependent spots, including #5 (GroEL2), #7 (PE-PGRS6a), #10 (unknown) and #11 (Rv1265), did not appear in cmr mutant. Two spots, (#’s 1 and 9), were present in the the cAMP-supplemented condition, but with reduced intensities in the cmr mutant relative to wt (Fig. 2, Table 2). Spots #2 (unknown) and 8 (Mdh), were constitutively present in the mutant, even in the absence of cAMP. Spot #6, which included Rv2971, was difficult to assess because it also contained FixB, whose expression is cAMP-independent (Gazdik and McDonough, 2005). Spots #3 and 4 were not evaluated due to intergel variability. Additional proteome changes occurred in the mutant in both the presence and absence of supplemental cAMP, but were not addressed in this study. These results show that cmr regulates protein expression levels in M. bovis BCG. More detailed characterization of these regulatory effects was limited to the five previously identified cAIGs, including Rv1265, groEL2, PE_PGRS6/6a, mdh and Rv2971 (Gazdik and McDonough, 2005).

Figure 2.

2D gel analyses showing cmr-associated differences in protein spots corresponding to cAIGs in the presence or absence of supplemental dbcAMP. Comparison of the regulatory pattern of the cAIG set between wt (top) and ΔCmrBCG (bottom) grown under shaking low oxygen (1.3% O₂, 5% CO₂) with (10 mM) or without (0 mM) exogenous dbcAMP (10 mM). Wild-type data were reported previously, and are included herein for reference only. Circled spots represent proteins regulated similarly in both wt and Δcmr BCG. Squares indicate proteins regulated differently between wt and Δcmr BCG. cAIGs corresponding to the protein spots are as follows: 5, GroEL2; 6, Rv1265; 7, PE_PGRS6; 8, Mdh, malate dehydrogenase; and 11, Rv1265.

Table 2.

Summary of qRT-PCR data^a

	Increased cAMPb		IC 2 hc		IC 24 h		Low pHd
cAIGe	BCG	Mtb	BCG	Mtb	BCG	Mtb	BCG	Mtb
Rv1265	↑f	↑	↑	↑	-	-	↑	-
Rv2971	↑	↑	-	-	-	-	↑	-
mdh	↑	↑	-	↓	↓	-	↓	↑
groEL2	↑	↑	-	-	↑	↑	-	↓
PE_PGRS6	↑	↑	↓	↑	↓	↑	↓	-

^aRelative mRNA levels of cAIG expression measured for M. bovis BCG or Mtb subjected to the specified conditions;

^bLevels of cAMP were increased by overexpression of the Rv1264 AC catalytic domain;

^cIC bacteria were compared with control bacteria in tissue culture media at 2 or 24 h post infection of macrophages;

^dbacteria subjected to pH 5.5 were compared with control bacteria at pH 6.7;

^ecAIGs listed were tested in each of the specified environmental conditions;

^fRelative changes in RNA levels for test condition versus control are designated as: up arrow, increased level of RNA; down arrow, decreased level of RNA; and -, no significant difference in RNA levels.

cAIG expression is induced by in vivo cAMP modulation

The exogenous cAMP supplementation model allowed identification of cAMP-regulated genes (Gazdik and McDonough, 2005), but the permeability and uptake of dbcAMP is difficult to monitor. Therefore, we confirmed the cAMP-induction of cAIGs by directly increasing the endogenous cAMP levels within M. bovis BCG. The catalytic domain of Rv1264, a soluble adenylate cyclase, is constitutively active in the absence of its regulatory domain (Linder et al., 2002). Constitutive expression of the Rv1264 catalytic domain in BCG(p0805:Rv1264cat) increased the level of cAMP within the cytoplasm of bacteria by 33.5 (± 5) fold in 7 day (late log phase) cultures grown shaking in ambient air (~20% O₂, 0.04% CO₂) conditions (Fig. 3A). This increase in cytoplasmic cAMP had no statistically significant effect on bacterial growth, as measured by OD600 (data not shown).

Figure 3.

Expression of cAIGs is induced by in vivo upregulation of M. bovis BCG cAMP levels. A) The catalytic domain of adenylate cyclase Rv1264 was constitutively overexpressed using the strong Rv0805 promoter, and radioimmunoassay was used to determine the bacterial cAMP levels during 7 d growth period. B) Quantitative real-time RT-PCR was used to measure cAIG expression in late log phase (7 day) cultures of wild-type and Rv0805:Rv1264cat BCG grown in ambient air conditions. mRNA levels were normalized to the constitutively expressed sigA mRNA. Data are the means of two independent experiments. * denotes p value < 0.05.

qRT-PCR was used to measure relative changes in cAIG expression between BCG(p0805:Rv1264cat) and a vector-only BCG control strain. Expression of all five cAIGs was increased in the Rv1264-overexpression strain relative to the vector-only control (Fig. 3B, Table 2). Induction levels ranged from 4.5 ± 0.5-fold for Rv1265 to 2.5 ± 0.4-fold for mdh. Similar patterns of regulation occurred when the Rv1264 catalytic domain was expressed in Mtb (Table 2). These data are consistent with our previous results using the exogenous cAMP addition model, and confirm that the regulation of cAIG expression is responsive to cytoplasmic cAMP levels.

Cmr interacts with the cAIG promoter regions

The BCGΔcmr 2D-GE data indicate that cmr is involved in the regulation of cAIG protein levels in BCG exposed to exogenous cAMP, but they do not establish whether this cmr-mediated regulation is direct. We tested the ability of Cmr to bind to DNA containing cAIG promoter regulatory regions using electrophoretic mobility shift assays (EMSA), to determine whether Cmr could be directly regulating expression of any of these genes.

Intergenic DNA sequences from upstream of the translational start site of each Mtb cAIG were used as DNA probes with purified his-tagged Cmr protein (Fig. 4). An alternative translational start site was used for Rv1265, based on sequence comparisons with Rv1265 orthologues in other species and promoter analyses (Bai et al., in preparation). This site is 112 bp upstream of the annotated site (Cole et al., 1998). All Mtb DNA sequences tested were identical in BCG, with the exception of a single nucleotide change at −112 nt relative to the new translational start of Rv1265 (i.e., −224 relative to annotated start). All cAIG probes, except PE_PGRS6, showed some mobility shift in the presence of Cmr. However, only Rv1265, groEL2, and to a lesser extent mdh, remained bound when a high concentration of unlabeled DNA from an unrelated Mtb intergenic region was used as a competitor (Fig. 4). All reactions also contained poly dI-dC, as described in the Methods. Surprisingly, the addition of cAMP did not measurably affect Cmr’s binding with any of these DNA sequences (data not shown). These EMSA results show that Cmr directly and specifically binds to the promoter regions of Rv1265, groEL2, and possibly mdh. In contrast, interactions between Cmr and the promoter region of Rv2971 likely represent non-specific binding.

Figure 4.

Cmr interacts with the upstream regulatory regions of Rv1265, groEL2, and mdh. EMSA were performed to examine the interaction of his-purified Cmr with the upstream DNA sequence of the cAIGs. ³³P end-labeled DNA probes were incubated with 0.4 μg Cmr and complexes were separated on a 6% nondenaturing polyacrylamide gel. 200-fold excess of unlabeled DNA specific for each DNA probe was added for competition reactions (cold specific DNA). 200-fold unlabeled DNA amplified from a non-binding control sequence (cold nonspecific DNA) was used as a control for specificity.

cAIGs are regulated in intracellular mycobacteria

Little is known about the environmental conditions that affect cAMP-mediated gene regulation in mycobacteria, but cAMP is likely to play a role in Mtb-host interactions. We examined the expression of cAIGs in BCG at 2 or 24 hours post infection of J774.1 macrophages. qRT-PCR was used to compare RNA levels within intracellular (IC) versus extracellular (XC) BCG. Expression of three cAIGs was affected during macrophage infection, including PE_PGRS6a (−2.2 ± 0.3-fold and −2.6 ± 0.2-fold at 2 and 24 h post infection), Rv1265 (2.3 ± 0.2-fold at 2 h post infection), and groEL2 (2 ± 0.5-fold at 24 h post infection) (Fig. 5A).

Figure 5.

cAIGs are regulated during intracellular growth in J774.16 macrophages. A) Quantitative real-time RT-PCR was used to examine cAIG mRNA expression in M. bovis BCG after a 2 or 24 h infection of J774.16 macrophages. Control mRNA levels were also examined in extracellular bacteria (white bar) grown in DMEM plus serum for the same time period. cAIG mRNA levels were normalized to the constitutively expressed sigA. B) Quantitative real-time RT-PCR was used to examine cAIG mRNA levels in Mtb (white bar) compared to M. bovis BCG (black bar) after a 2 h or 24 h macrophage infection. Transcription levels were normalized to sigA before cAIG expression in intracellular bacteria (IC) was compared to that of extracellular bacteria (XC). Data are the means of three independent experiments, and * indicates p value < 0.05.

qRT-PCR was also used to examine cAIG expression in virulent Mtb to determine whether regulation of these genes occurred similarly in Mtb and BCG. Late log phase Mtb was grown within J774.16 macrophages, as described for BCG. In most cases, cAIG expression in intracellular Mtb was similar to what was observed with intracellular BCG (Fig. 5B). Rv1265 and groEL2 were induced with the same temporal expression patterns in both BCG and Mtb upon uptake by macrophages. No statistically significant changes in Rv2971 expression occurred in either intracellular Mtb or BCG. Residence within macrophages downregulated the expression of mdh in both Mtb and BCG, although the temporal patterns and expression levels were not identical in Mtb and BCG. PE_PGRS6 was the only significant exception, in that PE_PGRS6 expression decreased in intracellular BCG, but increased in Mtb upon entry into macrophages at both 2 and 24 h post infection (Fig. 5B, Table 2).

Characterization of cmr’s role in intramacrophage regulation of cAIG expression

We further investigated the role of cmr in cAIG regulation, using qRT-PCR to measure cAIG mRNA levels in bacteria exposed to intramacrophage conditions. cAIG expression levels were compared in wt BCG, BCGΔcmr, and the complemented strain BCG Δcmr::pMBC668, which contains a single copy of the wt cmr gene integrated into the chromosome at the L5 att site.

All cAIGs except Rv2971 were dysregulated in BCGΔcmr relative to wt BCG when these bacteria resided within macrophages (Fig. 6A, B). However, the cmr deletion did not affect cAIG expression by bacteria in mycomedia or in tissue culture media alone (data not shown, Fig. S2). This indicates that Cmr specifically regulates cAIG expression under the biologically relevant intramacrophage condition. In contrast to wt BCG, induction of Rv1265 expression at 2 h, and groEL2 expression at 24 h, post infection did not occur in BCGΔcmr. While PE_PGRS6a was downregulated in wild-type BCG at both sampling times, regulation of PE_PGRS6a was affected by the loss of cmr only at the 24 h time point. cmr was also needed to maintain constitutive expression of genes that were not regulated in wild-type BCG during macrophage infection. mdh (−3.5 ± 0.2-fold) and groEL2 (−4.3 ± 3.0-fold) were each downregulated at 2 h post infection in BCGΔcmr, but not wt BCG. Similarly, Rv1265 (−3.4 ± 0.4-fold) was downregulated at 24 h post infection only in the cmr mutant (Fig. 6A, B). Complementation of the mutant returned the regulatory trends to the wt patterns, and survival of the mutant within macrophages was similar to that of wt during these assays (not shown).

Figure 6.

cmr is involved in intramacrophage-mediated cAIG regulation. cAIG mRNA levels were measured by quantitative real-time RT-PCR after wild-type (white bar), Δcmr (black bar), or complemented Δcmr::pMBC668 (horizontal lines) M. bovis BCG (A,B) or Mtb (C, D) were used to infect J774.16 macrophages for 2 h (A, C) or 24 h (B, D). Expression levels were normalized to sigA before comparing experimental and control conditions. * indicates a statistically significant difference compared to expression in the ΔCmr mutant (p value < 0.05). Grey lines represent a 2-fold change in expression, which was considered relevant in this study. Data are the means of three independent experiments.

The role of cmr in Mtb cAIG expression during macrophage infection was addressed using an MtbΔcmr mutant and the corresponding complemented strain. As with BCG, the temporal induction of Rv1265 and groEL2 was abolished in the mutant strain, while regulation of PE_PGRS6 was lost only at 24 h post infection (Fig. 6C, D). These results indicate that cAIG regulation occurs similarly in Mtb and BCG during macrophage infection, and that cmr likely plays a common role in this regulation in both species.

cAIG regulation under environmental stress conditions

Regulation of cAIGs was further characterized by exploring the different environmental stress conditions that Mtb may encounter during host infection. These conditions are expected to include low pH, hypoxia and possibly nitric oxide (NO) -enriched conditions within macrophage phagosomes, and within the caseous centers of granulomas (Smith, 2003; Sturgill-Koszycki et al., 1994). Transcriptional lacZ-reporter fusions were used to measure relative cAIG promoter activity in BCG under conditions of low pH (pH 5.5), elevated NO (250 μM DETA), or hypoxia (1.3 % O₂, 5 % CO₂). Low pH, but not hypoxia or NO, affected expression of most cAIG family members (data not shown). qRT-PCR was used to further investigate the low-pH mediated regulation of cAIGs by comparing RNA levels from wild-type BCG, BCGΔcmr, and the complemented strain grown for 6h at pH 5.5 or pH 6.7. Rv1265 (4.2 ± 1.2-fold) and Rv2971 (3.2 ± 0.08-fold) were each upregulated at pH 5.5 versus pH 6.7, while mdh (−12.0 ± 2.1-fold) and PE_PGRS6 (−4.3 ± 2.2-fold) were downregulated (Fig. 7, Table 2). Expression of groEL2 was unaffected by the change in pH. These data are consistent with the promoter:reporter fusion results, and indicate that four of five cAIGs are also regulated in response to pH in BCG.

Figure 7.

cmr is involved in low pH-mediated cAIG regulation in M. bovis BCG. cAIG mRNA levels were examined by quantitative real-time RT-PCR. Wild-type (white bar), Δcmr (black bar), or complemented Δcmr:pMBC668 (horizontal lines) M. bovis BCG was incubated for 6 h in standard mycomedia at pH 6.7 or pH 5.5. Expression levels were normalized to sigA before comparison between experimental and control conditions. * indicates a statistical difference compared to expression in Δcmr (p value < 0.05). Data are the means of three independent experiments.

cmr’s role in this pH-associated gene regulation was assessed by comparing cAIG expression in wt versus mutant BCG. Low pH induction of Rv1265, Rv2971, mdh, and PE_PGRS6a expression did not occur in the cmr mutant (Fig. 7). However, groEL2 expression, which did not respond to low pH in wild-type M. bovis BCG, was downregulated 3.9-fold in the mutant. Regulation of all five cAIGs returned to wild type trends when cmr was reintroduced into the knockout strain. This suggests that cmr is involved, either directly or indirectly, in pH-responsive cAIG expression, including maintaining constitutive groEL2 expression levels at pH 5.5 in wild type BCG.

Surprisingly, regulation of cAIGs in Mtb exposed to low pH was significantly different from that of the cAIGs in BCG (Fig. 8A), despite the near identity of the gene sequences in both species. None of the four acid regulated genes in BCG, (Rv1265, Rv2971, mdh, and PE_PGRS6), was regulated by the lowered pH in Mtb. In addition, groEL2 RNA levels, which were not affected by pH in BCG, decreased five fold in Mtb at pH 5.5. The expression of two control genes, ompA and gltA, both of which have been reported by others to be upregulated at pH 5.5 in Mtb (Fisher et al., 2002; Senaratne et al., 1998), was measured to confirm that our test conditions induced an acid response in Mtb. Expression of each gene was induced in both BCG and Mtb after 6 h at pH 5.5, indicating that the pH-associated gene regulatory differences we observed in BCG versus Mtb are specific to the cAIG set (Fig. 8B).

Figure 8.

Regulation of cAIGs under low pH differs between Mtb and M. bovis BCG. A) Quantitative real-time RT-PCR was used to examine cAIG mRNA levels in Mtb (white bar) after a 6 h exposure to low pH. Mtb expression levels were compared to cAIG expression in M. bovis BCG (black bar). mRNA levels were normalized to sigA before comparing expression at pH 5.5 to expression at pH 6.7. Data include three independent experiments. * statistical difference between Mtb and BCG transcription (p value < 0.05). B) Quantitative real time RT-PCR was used to measure the transcription of two previously reported Mtb acid-induced genes. RNA was extracted from late log phase cultures of Mtb or M. bovis BCG incubated in either standard mycomedia at pH 6.7 (black bar) or mycomedia at pH 5.5 (white bar). Levels of ompA or gltA transcription were normalized to sigA levels. Data are the means of two independent experiments. * p value > 0.05

Effects of low pH on mycobacterial cAMP levels and pHi

cAMP levels of both BCG and Mtb increase significantly upon entry into macrophages (Bai et al., in press). This is consistent with the similar patterns of Mtb and BCG cAIG expression we observed in response to both elevated cAMP and the intramacrophage environment. However, the discordance in acid regulated cAIG expression in Mtb versus BCG suggested differences in the cmr-associated low pH responses of these organisms.

A diverse set of signals can be generated when the external pH is altered, including corresponding changes in intracellular pH (pHi) (Olson, 1993). As noted earlier, the activity of at least one of Mtb’s soluble ACs, Rv1264, is directly responsive to pH (Linder et al., 2002), so we reasoned that cAMP levels could be affected by pHi changes within Mtb or BCG. We measured the pHi of BCG and Mtb in mycomedia at pH 6.7, or after a 1, 30, 60, or 90 min incubation at pH 5.5 using BCECF-AM, a pH-sensitive fluorescent dye. Adjustment of the external pH to 5.5 resulted in an immediate drop of 0.7 pH units in BCG. In contrast, Mtb’s pHi dropped only 0.3 units in response to an external pH of 5.5 (Fig 9). The lowered pHi’s did not affect viability of the bacteria, as both BCG and Mtb survived similarly over a 3 or 6 h incubation at pH 5.5 (data not shown). cAMP levels were also compared in BCG and Mtb exposed to acidic conditions. Exposure of either BCG or Mtb to pH 5.5 caused an approximate 2–3 fold increase in total cAMP levels, with similar increases occurring in both the bacterial and secreted portions of cAMP (Fig. 10). These results indicate that factors other than cAMP mediate the differences between Mtb and BCG with respect to cAIG regulation and pHi modulation in low pH conditions.

Figure 9.

Intracellular pH (pHi) levels of BCG are more affected by changes to the external pH than Mtb’s pHi. BCECF, a pH sensitive fluorescent dye, was used to monitor changes in the internal pH of BCG (A) or Mtb (B) after a 1, 30, 60, or 90 minute incubation at pH 6.7 (solid line) or pH 5.5 (grey line). The pHi of both M. bovis BCG and Mtb in standard mycomedia was approximately 7.6–7.7. These readings are for the purpose of measuring relative changes in pHi upon treatment, rather than absolute values. Recorded readings are typically more alkaline than the actual pHi due to incomplete equilibration of the internal and external pH in the nigericin-based BCECF standard curve.

Figure 10.

Low pH conditions increase cAMP levels in Mtb and M. bovis BCG. Radioimmunoassays were used to determine the amount of cAMP in the culture supernatants (secreted cAMP) or bacterial pellets (bacterial cAMP) of Mtb or M. bovis BCG subjected to different pH conditions. Late log phase cultures of Mtb (black bar) or M. bovis BCG (white bar) were incubated in standard mycomedia at pH 6.7 or pH 5.5 for 2 h before cAMP levels were determined. cAMP amounts are normalized to 1 OD₆₀₀ of treated sample culture.

Discussion

The success of Mtb as a pathogen relies on its ability to sense and respond to changing environmental conditions over the course of an infection. This study identifies at least part of a second cAMP-responsive regulon in Mtb, and establishes Cmr as a cAMP- and intramacrophage-responsive transcription factor. cmr was needed for the regulated expression of at least four cAIGs in response to elevated cAMP levels or intramacrophage conditions. This Cmr-associated regulon is distinct from the previously reported CRP_Mt regulon (Bai et al., 2005) with respect to gene membership, and most likely the environmental conditions that control its expression.

Intramacrophage regulation of cAIGs

As noted, four cAIGs were regulated within macrophages, and this regulation required cmr. These data are consistent with previous reports, which also showed intramacrophage upregulation of Rv1265 and groEL2, as measured by microarray and northern blot analyses (Hobson et al., 2002; Monohan et al., 2001). Surprisingly, PE_PGRS6/PE_PGRS6a expression was induced in Mtb, but downregulated in BCG, upon macrophage infection (Fig 5B). This discordant regulation of PE_PGRS6/PE_PGRS6a in Mtb versus BCG may provide clues to virulence-associated differences between these bacteria, and warrants further investigation. Other Mtb PE_PGRS family members are upregulated during intracellular growth of Mtb (Cappelli et al., 2006), and some have roles in immune modulation or Mtb’s survival within macrophages and granulomas (Banu et al., 2002; Brennan et al., 2001; Delogu and Brennan, 2001). A single base insertion in the BCG PE_PGRS6 DNA sequence causes a frameshift in the PE_PGRS6 orf that results in the potential for expression of PE_PGRS6a (581 amino acids) and PE_PGRS6b (75 amino acids) (BoviList). Loss of an advantageous function in PE_PGRS6a relative to PE_PGRS6 could reduce the selective pressure to maintain a specific pattern of gene regulation in BCG versus Mtb.

cmr-mediated regulation of the cAIGs

Regulation of PE_PGRS6/PE-PE-GRS6a expression in both Mtb and BCG was dependent upon cmr at 24 h, but not 2 h, post infection. In both cases, the ability of PE_PGRS6/PE-PE-GRS6a expression to modulate in response to the intramacrophage environment at 24 h post infection was lost in the absence of cmr. This regulatory pattern, along with the lack of Cmr binding to PE-PGRS6/PE-PGRS6a promoter sequences, is most consistent with an indirect role for Cmr in regulating PE-PGRS6/PE-PGRS6a expression within macrophages. Similarly, the lack of Cmr binding to Rv2971 upstream DNA sequences is consistent with an indirect role for Cmr regulation of this gene in BCG subjected to low pH conditions. Cmr may indirectly regulate these cAIGs by affecting the expression of other activators or repressors. These regulatory factors could also be influenced by other signaling pathways, resulting in the complex cAIG regulatory patterns found in this study.

In contrast, Cmr more likely functions directly as a transcription factor that regulates groEL2, Rv1265 and mdh expression. Purified Cmr interacted specifically with intergenic DNA upstream of each of these genes, as measured by EMSA (Fig. 4). However, Cmr’s in vitro DNA binding characteristics differed significantly from those of CRP_Mt (Bai et al., 2005; Bai et al., 2007). In contrast to CRP_Mt, Cmr’s DNA binding affinity was insensitive to the presence of cAMP, and was generally much lower than that of CRP_Mt when measured by EMSA. These results may indicate that Cmr’s activity is highly sensitive to environmental conditions, and our in vitro assays lacked a required cofactor or included suboptimal DNA and/or cAMP binding conditions. We also cannot rule out the possibility that Cmr’s cAMP-responsive regulation of these cAIGs occurs indirectly, although we think this less likely. We are currently investigating these possibilities.

cAMP signaling and gene regulation

The upregulatory patterns of Rv1265 and groEL2 expression by both Mtb and BCG within macrophages was similar to their regulation in response to elevated cAMP conditions. However, the cmr-associated regulatory patterns of other cAIGs were more complex in the macrophage environment than in the excess cAMP models. These results suggest that Cmr responds to environmental signals in addition to, or in lieu of, cAMP while within the macrophage. Elevated cAMP consistently caused only upregulated expression of cAIGs, while the effects of the intramacrophage environment on cAIG expression included both positive and negative regulation. In addition, the responses of Rv1265 and groEL2 to the intramacrophage environment varied with time, a pattern that was also observed by others for Rv1265 (Hobson et al., 2002). This complex regulatory pattern suggests that the cAIGs comprise a heterogenous group of cAMP-regulated genes, and that multiple competing pathways within the macrophage environment affect regulation of individual cAIGs differently. This possibility is consistent with the observation that only some of the five cAIG regulatory regions were directly bound by Cmr, as noted above.

It is also possible that localized cAMP concentrations and/or availability within the cell vary depending upon the cAMP source, e.g. the specific cyclase that produced it. In this case, artificially elevating cAMP levels within the cell using the exogneous cAMP or Rv1264 overexpression models could flood signaling pathways and cause a loss of regulatory specificity. We are currently investigating this potential role of cAMP localization and signaling specificity.

The differing cAIG regulatory responses that occurred between BCG and Mtb in a low pH environment (Fig. 8) is also consistent with additional levels of cAMP signaling specificity. The level of cAMP increased similarly in response to low pH in both Mtb and BCG, but this increased cAMP level did not affect cAIG expression in Mtb the way it did in BCG. The lack of cAIGs’ responsiveness to low pH conditions in Mtb is consistent with the results of a previous microarray study (Fisher et al., 2002). The acid-responsive control genes were similarly regulated in both BCG and Mtb (Fig. 8B), so the pH-associated regulation of the cAIGs in BCG may be due to dysregulation of an overlapping pathway. While the biological significance of these differences in Mtb versus BCG’s response to low pH remain to be determined, these different gene regulatory responses will be useful for identifying the signaling pathways that overlap with cAMP pathways in BCG versus Mtb. In addition, the reduced impact on pHi suggest that Mtb’s homeostasis was less readily perturbed by changes in the external pH. This phenotype could be important during host infection, and warrants further investigation.

Understanding Mtb’s response to its environment will provide important information about its requirements for establishing a successful infection. The present study identifies a new transcription factor and provides important new insights into cAMP-associated gene regulation during macrophage infection. Future work is needed to better define the individual components of the signaling pathway(s) that control cAIG expression. In addition, the broader questions of whether cAIG regulation is mediated by cAMP produced from a specific cyclase, and whether there is specificity of signaling through selected Mtb ACs in response to different environmental conditions, are likely to be fruitful areas of future investigation.

Experimental Procedures

Bacterial and Tissue Cultures

M. bovis BCG (Pasteur strain, Trudeau Institute) or Mtb H37Rv (ATCC 25618) was grown in mycomedia (7H9 liquid medium (Difco) supplemented with 0.5% glycerol, 10% oleic acid-dextrose-albumin-catalase, and 0.05% Tween-80). Cultures were grown at 37°C under conditions of ambient air (~20% O₂, 0.04% CO₂) or low oxygen (1.3% O₂, 5.0% CO₂) in 25 cm² tissue culture flasks left either standing undisturbed or shaking with constant gentle rocking as described previously (Florczyk et al., 2001; Purkayastha et al., 2002). For environmental gene expression assays late log phase cultures were exposed to low pH (mycomedia adjusted to pH 5.5 with 125 mM HCl), hypoxia (1.3 % O₂, 5.0% CO₂), or nitric oxide (250 μM diethylenetriamine [Sigma]) for 6 h, and gene expression was compared to non-exposed cultures.

The murine macrophage cell line J774.16 was maintained in antibiotic-free Dulbecco’s Modified Eagles’s Medium (DMEM) (Gibco) supplemented with 20% (v/v) fetal bovine serum, 5% (v/v) NCTC-109 (Gibco), 1% (v/v) nonessential amino acids (NEAA) (Gibco), and 1% (v/v) glutamine, as described (McDonough et al., 1993).

Mutant and gene reporter construction

cmr knockout strains of both BCG and Mtb were generated by using homologous recombination to replace a portion of the cmr open reading frame with a kanamycin cassette. Δcmr strains were confirmed using PCR and southern blot analyses (Fig. 1). The cmr ORF and its corresponding upstream intergenic promoter sequence was PCR amplified and cloned into the integrating expression vector pMBC409, which contains a hygromycin-resistance marker, to generate the single copy complemented strain Δcmr::pMBC668.

Promoter:lacZ reporter strains were generated for gene expression analyses. The intergenic DNA sequences upstream of each cAIG (Rv1265, Rv2971, mdh, groEL2, and PE_PGRS6) and the constitutive control gene tuf were amplified by PCR using primers specified in Table 1. Amplified DNA was cloned into the single copy vector, pLacInt, directly upstream of the promoterless reporter gene β-galactosidase (lacZ) (Purkayastha et al., 2002).

Table 1.

Primers used in this study

Gene	Oligonucleotide sequencea
Promoter amplification
Rv1265	F - GGATCCCGTGCATGTGATGGTGCC
	R - GGATCCCGGCCGGCCTTCACCC
Rv2971	F - GGATCCCCAGCACGTCATCTCACC
	R - GGATCCCCTACAAAACTCTGTCACGC
mdh	F - GGATCCCGTCAACCTCCGATCGCGG
	R - GGATCCGGACTAGCGCTCACGTCGG
groEL2	F - GGATCCGGTCTTGTTGTCGTTGGCGG
	R - GGATCCTTCGTACGCAATTGTCTTGGC
PE_PGRS6	F - GGATCCGGCCCGGCTGCGTGGC
	R - GGATCCGCACCGATAGCCGACCC
tuf	F - GGATCCACACCCGAGGACTACATGGG
	R - GGATCCTGGTCCCGATGTTGACGTGG
Rv0805	F - GATATCGATGTCGGCTCGTGAGTTGC
	R - GGATCCCGCGGCCCTAAGTCTATGC
RT-PCR
Rv1265	F – ATGTAATCATGTAATTATGAGGC
	R - TTGTCCAGCAGCCTGCC
Rv2971	F - GCGCAATTGCAGCCTCCG
	R - GCGATAGTCATTCGCTCCG
mdh	F - GGATGACGTTCGCGTCGGG
	R - GAATTCATCCTCGATCCAGGC
groEL2	F - CATCGCCGGACGAGTGGC
	R - CGATCTGCTTCAGCGGGG
PE_PGRS6	F - GGATTTGGCGGGTATTGGG
	R - CTGCTGCTCCACGTTGGC
tuf	F - GTGCGGAAGTAGAACTGCGC
	R - AGGAAGTTGAGATCGTCGGC
16S rRNA	F - GCGAACGGGTGAGTAACACG
	R - TGAAAGAGGTTTACAACCCG
ORF of catalytic domain
Rv1264	F - GGATCCCCGGGAGCGCGACAGGTC
	R - AAGCTTCATGTGTACCGGTGTGCCTG

^aF – forward primer; R – reverse primer

The catalytic domain of adenylate cyclase Rv1264 was expressed using the Rv0805 promoter to increase in vivo cAMP levels in both BCG and Mtb. The Rv0805 promoter provides strong constitutive expression in the conditions tested (not shown). DNA sequences encoding the catalytic domain (631 nt – 1194 nt) of the adenylate cyclase Rv1264 (Linder et al., 2002) were amplified from the Mtb genome using primers with BamHI or HindIII ends, as listed in Table 1. The Rv0805 promoter sequence was amplified using primers with EcoRV or BamHI ends. Both Rv0805 promoter and Rv1264 catalytic domain sequences were joined with the BamHI linkers, and ligated into pCRII.oriM at unique EcoRV and HindIII unique sites. This construct was transformed into BCG and Mtb. Rv1264 catalytic domain was constitutively expressed from the Rv0805 promoter and the in vivo cAMP levels were measured by radioimmunoassay, as described below.

Sample preparation and 2D-gel electrophoresis of Δcmr M. bovis BCG proteins

Bacteria were harvested and lysed, as described previously (Gazdik and McDonough, 2005). Briefly, washed cells were resuspended in Tris-SDS buffer (0.3% [w/v] SDS, 50mM Tris-HCl pH8.0) and lysed by several cycles of freeze/thaw and 2 rounds of sonication using a Virsonic 475 Ultrasonic Cell Disrupter with a cup horn attachment (VirTis Company) for 10 min at setting 8. Protein samples were precipitated with 10% trichloracetic acid (TCA) to remove any contaminating salts or lipids before isoelectric focusing.

Protein samples were separated by 2-D SDS-PAGE as described previously (Gazdik and McDonough, 2005). Approximately 200μg of protein were separated by isoelectric focusing in polyacrylamide tube gels 1.5 mm (i.d), 15 cm (length). Focusing was performed using a constant voltage of 667 V over 18 h. The second dimension was run on 12% SDS-PAGE gels 1.5 mm thick, 16×20 cm. Gels were stained by standard silver staining methods, and spot patterns were compared visually (Gazdik and McDonough, 2005).

lacZ-Reporter fusion assay

Cells were incubated with the fluorescent substrate 5-acetylamino-fluorescein di-β-D-galactopyranoside (C₂FDG) (Molecular Probes) for 2 h at 37°C and the level of fluorescence was detected using the CytoFluor multi-well plate reader 4000 (PerSeptive Biosystems) at 485 nm excitation and 530 nm emission, as described (Rowland et al., 1999). Fluorescence levels were normalized to 10⁶ bacteria, determined by either cfu or OD650. An empty vector containing the promoterless LacZ was used as a negative control.

Infection of J774.16 monolayers for intracellular mycobacterial RNA extraction

J774.16 macrophages were seeded at 7.5 × 10⁵ cells/ml in 150 cm² flat bottom tissue culture flasks. Macrophage monolayers were then infected for 2 or 24 h with either M. bovis BCG or Mtb at a multiplicity of infection (MOI) of 100:1 (bacteria:macrophage). This high MOI was used in order to increase bacterial yield to allow for successful RNA extraction. After infection, cultures were washed 3 times with calcium and magnesium-free phosphate buffered saline (PBS-CMF) to remove extracellular bacteria. Intracellular (IC) bacteria were collected by lysing macrophages with 0.5% deoxycholate/0.5% Triton X-100 in PBS-CMF. IC bacteria were collected in a 10% guanidinium thiocyanate (GTC) solution (4 M guanidinium thiocyanate, 25 mM sodium citrate, 0.5% (w/v) sodium-N-lauroylsarcosine, and 0.1 M β-mercaptoethanol) in PBS-CMF to halt transcription. The macrophage debris/IC bacteria mixture was sonicated in a closed container using the Virsonic 475 Ultrasonic Cell Disrupter with a cup horn attachment (VirTis Company) at 4°C using setting 4 for 6 cycles (M. bovis BCG) or 8 cycles (Mtb) of 5 s on, 5 s off. Bacteria were pelleted to recover them from the macrophage debris and lysed for RNA extraction following the protocol described below. Extracellular (XC) bacteria were grown in DMEM supplemented with 20% (v/v) fetal bovine serum for 2 or 24 h for use as a control.

RNA Preparation

M. bovis BCG or M. tuberculosis H37Rv late log phase culture was pelleted, washed with 0.5% Tween-80, and resuspended in RNase free water. RNA was extracted as previously described (Gazdik and McDonough, 2005). Briefly, cells were disrupted mechanically using a bead beater (BioSpec Products) in a mixture of 0.1 mm zirconia-silica beads (BioSpec Products), 45% Divolab no. 1 (Diversey), 45% acid phenol, and 10% chloroform-isoamyl alcohol (24:1). RNA was reextracted with an equal volume of chloroform-isoamyl alcohol, precipitated with isopropanol/3 M sodium acetate (pH 5.2), and resuspended in RNase-free water. DNA contamination was removed using the RNeasy Mini Kit and the RNase-free DNAse set following manufacturer specifications (QIAGEN).

Quantitative Real-time Reverse Transcription – Polymerase Chain Reaction (qRT-PCR)

cDNA was prepared from 0.5 μg of RNA and 0.125 μg of random primers RPA00, RPT00, RPC00, and RPG00 (Table 1) as described previously (Gazdik and McDonough, 2005). 2 μl of a 1:5 dilution of cDNA was used in 25 μl PCR reactions containing 12.5 μl SYBR green supermix (BioRad), 6.3 μl 10 μM forward and reverse primers (Table 1), 4.2 μl H₂O, and 5.6 μl 50% glycerol. Primer sets were optimized using a standard curve produced by amplifying serial dilutions of genomic DNA template. Real time PCR analysis was performed in a MyiQ light cycler (BioRad) at 1 cycle of 95°C for 3 min, 40 cycles of 95°C for 10 sec, 58°C for 30 sec, 1 cycle of 95°C for 1 min, and 1 cycle of 55°C for 1 min. Gene expression was normalized to the reference gene sigA, which shows constitutive expression in the environmental conditions analyzed (Manganelli et al., 1999). Normalized gene expression was obtained using the ΔC_T method with the equation; Normalized expression (ratio of ref:target) = E^{Ct(ref) – Ct(target)}. E stands for the amplification efficiency of the PCR reaction and was determined using the linear regression equation from the standard curve, E = 10^−1/slope (Bio-Rad).

Electrophoretic mobility shift assay

Electrophoretic mobility shift assays (EMSA) were performed as described previously (Bai et al., 2005). Briefly, ³³P-end-labeled DNA probes (0.05 pmol) were incubated with various concentrations of purified His-Cmr for 30 min at room temperature in DNA binding buffer (10 mM Tris-HCl pH 8.0, 50 mM KCl, 1 mM EDTA, 50 μg/ml BSA, 1 mM DTT, 0.05% NP-40, and 5% glycerol). 200-fold excess of unlabeled DNA fragments were used for competition experiments. 20 μg/ml poly dI-dC was used in all reactions as a nonspecific competitive inhibitor. Reactions were electrophoresed on a non-denaturing 6% polyacrylamide gel for 2 to 3 h at 14 V/cm and visualized with a Storm 860 PhosphorImager (Molecular Dynamics).

Radioimmunoassay for cAMP detection

Late log phase bacteria were pelletted and cAMP content of the supernatant was measured to determine the amount of secreted cAMP. The bacterial pellet was lysed by boiling in 500 μl 50 mM sodium acetate for 5 min. Bacterial debris was removed and cAMP content of the supernatant of lysed bacteria was measured to determine bacterial levels of cAMP. For experiments that included virulent Mtb, both bacterial and secreted samples were heat killed at 95°C for 1 h before processing. cAMP levels were measured by radioimmunoassay using a rabbit anti-cAMP antibody (Calbiochem) and [¹²⁵I]cAMP tyrosine methyl ester, as previously described (Steiner et al., 1972; Swift and Dias, 1987). Anti-cAMP antibody, [¹²⁵I]cAMP tyrosine methyl ester, and acetylated cAMP standard (0 – 512 fmol) or unknown samples were mixed and incubated for 18 h at 4°C, followed by co-precipitation with γ-globulin using (NH₄)₂SO₄. Precipitated radioactivity was counted in a Wallac 1470 Wizard gamma counter (PerkinElmer Life Sciences) and results were normalized by OD600 readings.

pHi meaurement

Intracellular pH (pHi) was determined by using 2′,7′-bis-2-carboxyethyl-5-(and 6)-carboxyfluorescein acetoxymethyl ester (BCECF-AM). Late log phase M. bovis BCG was resuspended in 5 mM BCECF in DPBS and incubated at 37°C for 2 h. BCECF-loaded mycobacterial cultures were then exposed to either pH 5.5 or pH 6.7 and BCECF fluorescence was measured using the CytoFluor multi-well platereader 4000 (PerSeptive Biosystems) with a dual excitation at 450 nm (pH-insensitive) and 480 nm (pH-sensitive) and a fixed emission at 530 nm. The ratio of 480 nm: 450 nm was then compared to fluorescence ratios determined from an in vivo standard curve generated by the K⁺/H⁺ ionophore nigericin (Sigma).