DarR, a TetR-like Transcriptional Factor, Is a Cyclic Di-AMP-responsive Repressor in Mycobacterium smegmatis*

Background: No transcriptional factor has been characterized as a c-di-AMP-responsive regulator in bacteria. Results: A TetR-like regulator, DarR, could specifically bind c-di-AMP and repressed gene expression in M. smegmatis. Conclusion: DarR is the first cyclic di-AMP receptor regulator to be identified in bacteria. Significance: Our findings provided an opportunity to understand the physiological function and regulatory mechanism of c-di-AMP. Cyclic dinucleotides, including cyclic di-AMP (c-di-AMP), are known to be ubiquitous second messengers involved in bacterial signal transduction. However, no transcriptional regulator has been characterized as a c-di-AMP receptor/effector to date. In the present study, using a c-di-AMP/transcription factor binding screen, we identified Ms5346, a TetR family regulator in Mycobacterium smegmatis, as a c-di-AMP receptor in bacteria. Ms5346 could specifically bind c-di-AMP with Kd of 2.3 ± 0.5 μm. Using EMSA and DNase I footprinting assays, c-di-AMP was found to stimulate the DNA binding activity of Ms5346 and to enhance its ability to protect its target DNA sequence. A conserved 14-bp palindromic motif was identified as the DNA-binding site for Ms5346. Further, Ms5346 was found to negatively regulate expression of three target genes including Ms5347 (encoding a major facilitator family transporter), Ms5348 (encoding a medium chain fatty acyl-CoA ligase), and Ms5696 (encoding a cold shock protein, CspA). Ms5346 is the first cyclic di-AMP receptor regulator to be identified in bacteria, and we have designated it as DarR. Our findings enhance our understanding of the function and regulatory mechanism of the second messenger c-di-AMP in bacteria.

Cyclic di-AMP is synthesized by the diadenylate cyclase domain via condensation of two ATP molecules (3). This was observed for the first time in crystallization studies of the DNA damage checkpoint protein DisA, which contains a diadenylate cyclase domain, in Bacillus subtilis (3,14). It has recently been found that overproduction of GdpP (homolog of YybT) phosphodiesterase, which degrades c-di-AMP to pApA (15), greatly enhances the sensitivity of B. subtilis cells to ␤-lactam antibiotics (4). Genetic studies have shown that at least one of three B. subtilis diadenylate cyclase domains is required for bacterial growth (4). More recently, it has been reported that mutation of the phosphodiesterase gene in Staphylococcus aureus led to increased amount of cross-linked peptidoglycans (6). Thus, c-di-AMP could play important roles in controlling bacterial cell size and envelope stress (6). Interestingly, c-di-AMP was found to be secreted by the pathogen Listeria monocytogenes and could stimulate an IFN-␤-mediated host immune response (5). These findings suggested that c-di-AMP could act as either an intracellular or extracellular signal to participate in regulating bacterial physiology and pathogenesis. However, the specific receptors of c-di-AMP and its essential regulatory mechanisms have not been clearly identified yet.
A disA ortholog that can synthesize c-di-AMP (3) exists in almost all mycobacterial genomes with the exception of Mycobacterium leprae. In the human pathogen Mycobacterium tuberculosis, the DisA ortholog Rv3586 is a functional diadenylate cyclase that can convert ATP or ADP to c-di-AMP (16). However, the regulatory function of c-di-AMP in mycobacteria remains to be explored. Mycobacterium smegmatis has been widely used as a model bacterium to study the gene regulation and signal transduction mechanism in mycobacteria (17,18). In this study, using a c-di-AMP/transcription factor binding screen, we have identified the first c-di-AMP receptor regulator, DarR (encoded by Ms5346), in M. smegmatis. DarR was found to specifically associate with c-di-AMP and negatively regulate the expression of its target genes. The binding motif sequence for DarR was further identified, and our results suggest that c-di-AMP could enhance the DNA binding ability of DarR. Our findings provided an opportunity to understand the physiological function and regulatory mechanism of c-di-AMP.

EXPERIMENTAL PROCEDURES
Strains, Enzymes, Plasmids, and Chemicals-Escherichia coli BL21 cells and the pET28a vector were purchased from Novagen. All enzymes including restriction enzymes, ligase and DNA polymerase, dNTPs, and all antibiotics were purchased from TaKaRa Biotech. PCR primers were synthesized by Invitrogen (supplemental Table S1). All the plasmids constructed and used in this study are listed in supplemental Table S2. Ni 2ϩ -NTA-agarose was obtained from Qiagen.
Cloning, Expression, and Purification of Recombinant Proteins-M. smegmatis mc 2 155 genes were amplified by PCR from mycobacterial genomic DNA using their respective primers (supplemental Table S1). The darR gene encoded by Ms5346 was amplified and cloned into the prokaryotic expression vector pET28a. E. coli BL21 was used to express recombinant proteins. Recombinant E. coli BL21 cells were grown in 1 liter of LB medium up to A 600 of 0.6 at 37°C. Protein expression was induced by the addition of 1 mM isopropyl ␤-D-1-thiogalactopyranoside at 16°C for 10 h. Proteins were purified on Ni 2ϩ affinity columns as described previously (19). The elution was dialyzed overnight and stored at Ϫ80°C.
Bacterial One-hybrid Assays-The regulatory sequence of the darR gene was cloned into the reporter vector pBXcmT (20). DarR was fused into the N-terminal domain of the ␣-subunit of RNA polymerase in the pTRG vector (Stratagene). Bacterial one-hybrid assays were carried out as described previously (20).
DNA Substrate Preparation and EMSA-DNA fragments for the DNA binding activity assays were amplified by PCR from M. smegmatis mc 2 155 genomic DNA or directly synthesized by Invitrogen (supplemental Table S1). DNA fragments used in this assay were labeled with FITC. The reactions (10 l) for measuring mobility shift contained labeled DNA substrates and increasing concentrations of the protein (as indicated in the legend of the corresponding figure) and were incubated in EMSA buffer (100 mM Tris-HCl, pH 8.0, 100 mM NaCl, 1 mM DTT, and 10% glycerol) for 30 min at room temperature and then subjected to 5% native PAGE. Electrophoresis was performed at 150 V at room temperature for ϳ1.5 h in 0.5ϫ Trisborate-EDTA buffer. Images of gels were acquired using a Typhoon Scanner (GE Healthcare).
UV Cross-linking Assay-[ 32 P]c-di-AMP was synthesized as previously described (3). UV cross-linking experiments with [ 32 P]c-di-AMP were performed as described in a previous study (10). DarR protein was co-incubated with radiolabeled nucleotides in a buffer containing 20 mM Tris-HCl, pH 7.5, and 50 mM NaCl at room temperature for 15 min. Then the samples were exposed to UV light for 30 min, and 2ϫ SDS loading buffer was added to stop the reactions. The mixtures were then directly subjected to 15% SDS-PAGE. Electrophoresis was performed at 120 V at room temperature for 1 h. Radioactive gels were exposed to a phosphor storage screen (GE Healthcare) over-night. Images of gels were acquired using a Typhoon Scanner (GE Healthcare).
Surface Plasmon Resonance (SPR) Analysis for the Binding of c-di-AMP with DarR-To detect the physical interaction between DarR and c-di-AMP, His-tagged DarR proteins were immobilized onto NTA sensor chips. SPR analysis was conducted on a Biacore 3000 instrument according to previously published procedures (21). Purified c-di-AMP, c-di-GMP, or ATP diluted in reaction buffer (50 mM Tris-HCl, pH 7.5, 50 mM NaCl, 5 mM MgCl 2 ) to concentrations of 10, 20, or 50 M was used as the ligand and passed over the chips. An overlay plot was produced using the BIA evaluation 3.1 software to depict the interaction between DarR and c-di-AMP. Each analysis was performed in triplicate.
DNase I Footprinting Assays-DNase I footprinting was performed as described previously (18,22). The 189-bp promoter regions of the darR gene (supplemental Table S1) were amplified by PCR using appropriate primers labeled with FITC (supplemental Table S1). The amplified products were purified with a DNA purification kit (BioFlux) and then subjected to similar binding reaction as in EMSA. The ladders were produced using the Sanger dideoxy method.
Construction of the darR Deletion Mutant of M. smegmatis and Southern Blot Analysis-A pMind-derived (23) suicide plasmid, pKOdarR (supplemental Table S2), carrying a hygromycin resistance gene was constructed, and a sacB-lacZ gene was inserted to confer sensitivity to sucrose as a negative selection marker. Knock-out of the darR gene in M. smegmatis mc 2 155 was performed as described previously (22), and the knock-out strain was selected on LB agar medium containing 50 g/ml hygromycin, 2% sucrose, and 200 g/ml 5-bromo-4chloro-3-indolyl-␤-D-galactopyranoside (X-gal). Deletion of the darR gene was confirmed by Southern blot analysis as described previously (19). The probe consisted of a 270-bp fragment of the upstream region of the darR gene amplified using appropriate primers (supplemental Table S1).
Scanning Electron Microscopy (SEM)-M. smegmatis cells prepared for SEM observation were grown and harvested at late log phase by centrifugation. The bacterial pellets were resuspended and incubated at 4°C for 24 h in 2.5% glutaraldehyde solution. The further cell treatment for observation was performed as described in a previous study (19). The p values of the cell length date were calculated by unpaired two-tailed Student's t test using GraphPad Prism 5.
Construction of darR Overexpression Strains and Determination of Growth Curves of Recombinant Mycobacteria-We used the predicted darR upstream promoter (Pdar) to replace the strong hsp60 promoter of pMV261 to produce a moderate darR overexpression strain (supplemental Table S2). Wild type darR gene and N-terminal helix-turn-helix (HTH) domain deleted mutant darR⌬N gene were then cloned downstream of the PdarR promoter. The resulting pMV261-derived plasmids were further transformed into M. smegmatis, which produced two recombinant strains, Msm/pMVPdar-darR and Msm/ pMVPdar-darR⌬N (supplemental Table S2). Growth patterns of these two recombinant mycobacterial strains were examined according to previously described procedures (17). Briefly, M. smegmatis was grown overnight in Middlebrook 7H9 medium (complemented with 0.05% Tween 80 and 0.2% glycerol) containing 50 g/ml hygromycin and 30 g/ml kanamycin. When cells entered a stationary growth phase (A 600 between 1.5 and 2.0), each culture was diluted (4:100) in 100 ml of fresh 7H9 broth. The cultures were then allowed to grow further at 37°C with shaking at 160 rpm. Aliquots were taken at the indicated times, and the A 600 values of the cultures were determined.
Quantitative Real Time PCR Assays-Isolation of mRNA and cDNA preparation of Msm/pMVPdar-darR⌬N, Msm/ pMVPdar-darR, and darR deletion mutant M. smegmatis strains (supplemental Table S2) were performed, and real time PCR analysis was subsequently carried out according to previously described procedures (19). The darR-complemented M. smegmatis strains were constructed by introducing the darR gene back into the darR-deleted mutant strains using a pMV361Pdar-darR plasmid. The reactions were performed in a Bio-Rad IQ5 RT-PCR machine using the following thermocycling condition as follows: 95°C for 5 min and 40 cycles at 95°C for 30 s, 60°C for 30 s, and 72°C for 30 s. Amplification specificities were assessed by conducting melting curve analyses. Gene expression levels were normalized to the levels of sigA gene transcripts. The degrees of expression change were calculated using the 2 Ϫ⌬⌬Ct method (19,22). The average relative expression levels and standard deviations were determined from three independent experiments.
GC/MS Assays-M. smegmatis strains were grown overnight at 37°C in 1 liter of 7H9 medium, and the cells were harvested and washed. Preparation of mycobacterial fatty acids was performed according to the procedures previously (24). GC/MS analyses were carried out according to the procedures previously with modifications (25). The experiments were performed on a gas chromatograph Agilent 7890A coupled to Agilent 5975C mass spectrometer (Agilent Technologies) using HP-5MS column (5% Phenyl Methyl Silox 30 m ϫ 250 m ϫ 0.25 m). The GC conditions were as follows: the injection volume was 2 l, the flow rate of N 2 as the carrier gas was 1.0 ml/min. For the temperature program, the initial temperature was 120°C (hold 7 min), with a ramp to 170°C at 4°C/min (hold 3 min) and a ramp to 250°C at 3°C/min (hold 5 min), and the injector temperature was 280°C. The relative percentages of intracellular fatty acids were normalized using dodecanoic acid methyl ester (Sigma) as an invariant fatty acid, which was added into the mycobacterial cell samples before extraction.

DarR Is a Potential Regulator for Sensing c-di-AMP in
M. smegmatis-Approximately 500 predicted transcriptional regulatory genes (GenBank TM accession number CP000480) in the genome of M. smegmatis were cloned to identify potential regulators that specifically associate with c-di-AMP. All of these recombinant plasmids were divided into 17 groups, and each group of recombinant proteins was co-purified for the c-di-AMP binding screen (supplemental Fig. S1, upper panels). Using a UV cross-linking assay, we detected a positive c-di-AMP-binding signal only in one group (supplemental Fig. S1, lower panels, lane 16). This group of regulators was then purified and used to conduct a further 32 P-labeled c-di-AMP/pro-tein binding screen. The B. subtilis DisA protein (BsuDisA) was used as a positive control (3). Finally, DarR was found to be a candidate regulator (Fig. 1A, fifth lane) that binds to c-di-AMP, as evidenced by a specific and strong autoradiograph signal observed on the gel. BsuDisA was also observed to bind to c-di-AMP (Fig. 1A, first lane).
DarR Could Specifically Associate with c-di-AMP-We conducted a competition experiment to measure the binding specificity of c-di-AMP with the DarR protein. As shown in Fig. 1B, when DarR was co-incubated with radioactively labeled c-di-AMP, a strong autoradiograph signal corresponding to DarR was clearly seen on a PAGE gel (Fig. 1B, first lane), indicating that DarR can bind to c-di-AMP. Further, the addition of unlabeled c-di-AMP at 60, 120, and 250 M to the reaction mixtures competitively inhibited the binding of DarR to [ 32 P]c-di-AMP, indicating specificity of DarR binding to c-di-AMP. Importantly, the addition of a different nucleotide (ATP) at the same concentration did not alter the binding of DarR to c-di-AMP ( Fig. 1B, fifth, sixth, and seventh lanes). In addition, other unlabeled dinucleotide or nucleotides such as c-di-GMP, ATP, GTP, dATP, and dGTP could not competitively inhibit the binding of DarR to [ 32 P]c-di-AMP (Fig. 1C). These results indicate that DarR can bind specifically with c-di-AMP.
Furthermore, SPR assays also confirmed the binding specificity of c-di-AMP with DarR. As shown in Fig. 1D, when an increasing amount of c-di-AMP was passed over the His 6 -DarR immobilized NTA chip, a corresponding increase in response was observed (left panel). In particular, when 1 mM c-di-AMP was passed over the His 6 -DarR immobilized NTA chip, a response of ϳ400 response units was observed (Fig. 1D). Furthermore, K d for the specific interaction between DarR and c-di-AMP was measured to be 2.3 Ϯ 0.5 M. No significant response was observed if the same concentration of c-di-GMP or ATP was passed over the NTA chip (right panel), which is consistent with the results of cross-linking assays described above. Taken together, these results strongly indicate that DarR can specifically bind with c-di-AMP.
DarR Is a TetR Family Regulator and Specifically Recognizes Its Upstream DNA Sequence-The M. smegmatis darR gene encodes a 210-residue protein containing a typical TetR-like N-terminal HTH domain and a QacR-like C-terminal domain ( Fig. 2A). Sequence alignment analysis showed that DarR shared a highly conserved TetR family DNA-binding domain and a variable C-terminal receptor domain with QacR in S. aureus, AcrR in Pelotomaculum thermopropionicum, and FadR in B. subtilis (supplemental Fig. S2). Genomic location analysis indicated that darR gene shares the same upstream DNA region with the Ms5347-5348 operon (Fig. 2B), and the two genes had previously been predicted to be a major facilitator family transporter and a medium chain fatty acyl-CoA ligase (GenBank TM accession number CP000480), respectively.
The target promoter of the DarR transcription factor has not yet been identified. We examined the potential of DarR to bind to the upstream 198-bp region of DarR and the Ms5347-5348 operon (Fig. 2B). This region was amplified by PCR and used to conduct an EMSA assay. As shown in Fig. 2C, when 3 nM DNA substrates were co-incubated with increasing concentrations of DarR (0, 0.12, 0.25, and 0.5 M), a specific and clear shift in the A c-di-AMP-responsive Regulator FEBRUARY 1, 2013 • VOLUME 288 • NUMBER 5 bands was observed (lanes 2-4). In contrast, heat-denatured DarR protein failed to cause shifts in the DNA bands (Fig. 2C, lane 5). Additionally, DarR could not bind the promoter DNA of an unrelated gidB gene (Fig. 2C, lanes 6 -9). A competition assay further confirmed the specificity of DarR binding with its upstream DNA region. As shown in Fig. 2D, cold darR upstream DNA, but not gidB promoter DNA, could competitively inhibit the binding of DarR to the labeled darR upstream DNA substrate. In addition, bacterial one-hybrid assays confirmed the specific binding of DarR with its upstream DNA fragment (supplemental Fig. S3). These results indicate that DarR can specifically bind to its own upstream DNA region.
DarR Binds with DNA Fragments Containing a Palindrome Sequence Motif-Several truncated DNA substrates covering the upstream region of darR, designated as Regions A through D (supplemental Fig. S4), were produced to characterize the DNA motif recognized by DarR. An obvious DNA binding activity was observed with the 154-bp substrate Region A and the 92-bp Region B, but not with Region C or D in EMSA assays (supplemental Fig. S4). The binding motif for the recognition of DarR was further characterized by DNase I footprinting assays. As shown in Fig. 3A, when increasing amounts of DarR protein (0 -2 M) were co-incubated with DNase I, the region around ACCAGGGATACTACGGAGTATTACGGTAA was clearly protected on the coding strand. This result indicated that this DNA fragment contained a potential binding motif for DarR. The protected DNA region extended from Ϫ120 to Ϫ92 (relative to the translational start codon) in the coding strand (Fig.  3B). A palindromic motif formed by two inverted repeats (IR, 5Ј-ATACT-3Ј) separated from each other by four nucleotides (Fig. 3B) was found in this protected sequence.
Further EMSA assays confirmed the significance of this motif for specific recognition by DarR (Fig. 3, C and D). We designed a series of DNA substrates in which one of the inverted repeats (darR-S2 or darR-S3) or both inverted repeats (darR-S2 and darR-S3) were mutated (Fig. 3C). As shown in Fig.  3D, DarR was incapable of binding with all three mutated substrates. In addition, mutations in the region located between the IR or elsewhere did not affect the DarR binding (supplemental Fig. S5). These results indicate that DarR recognizes and binds to an essential palindromic sequence motif in its own promoter region.
We used the 14-bp sequence motif (ATACTNNNNAG-TAT) to search for DarR target genes in the M. smegmatis genome, allowing only two mismatches in the underlined palindromic sequences at the most. Interestingly, we identified an additional target site in the upstream sequence of the cspA gene, which encodes a predicted cold shock protein (Fig. 4A). We further confirmed the DNA binding activity of DarR with the upstream DNA sequence of cspA, cspA_up (Fig. 4B, lanes  2-4), in an EMSA assay. Furthermore, using a ChIP assay, the upstream target promoters of Ms5347-5348, and cspA genes could be specifically recovered by a DarR antibody in a wild type strain (Fig. 4C, left panel) but not in darR-deleted mutant strains (Fig. 4C, right panel). A negative control upstream sequence of the gidB gene could not be recovered by the antibody under the same conditions.
Thus, in addition to the darR gene itself, we identified three potential target genes of DarR, namely Ms5347 (which encodes a major facilitator family transporter), Ms5348 (which encodes a medium chain fatty acyl-CoA ligase), and cspA (a cold shock protein gene encoded by Ms5696). DarR can specifically bind with the upstream DNA regions of these target genes.
c-di-AMP Specifically Stimulates the DNA Binding Ability of DarR-We then determined the regulatory effect of c-di-AMP on the DNA binding activity of DarR. As shown in Fig. 5A, 60 nM DarR could bind the darR-S1DNA substrate in an EMSA assay (lane 1). When increasing amounts of c-di-AMP (125-500 M) were added into the reaction mixture, a corresponding increase in the amounts of shifted DNA substrates was clearly observed (Fig. 5A, lanes 3-5). This indicated that c-di-AMP stimulates the DNA binding activity of DarR in a concentration-dependent manner. In contrast, c-di-GMP had no effect on the DNA binding activity of DarR (Fig. 5A,  lanes 9 -11), whereas ATP had a modest effect on the activity of DarR (Fig. 5A, lane 8). These results indicate that c-di-AMP specifically stimulates the ability of DarR to bind to its target DNA.
To confirm the regulatory effect of c-di-AMP on DarR, we performed a footprinting assay in the presence of different concentrations of c-di-AMP. As shown in Fig. 5B, c-di-AMP was able to enhance the protection of the DNA region by 0.5 M DarR (Fig. 5B, lanes 2-5) in a concentration-dependent manner. This observation is consistent with the stimulatory effect of c-di-AMP on the DNA binding activity of DarR described above (Fig. 5A). In addition, c-di-AMP also stimulated the DNA binding activity of DarR with the upstream sequence of cspA (Fig. 4B, lanes 5-7), as indicated by the higher amount of protein-DNA complex produced in the presence of c-di-AMP in the EMSA reactions. Taken together, these results indicate that c-di-AMP can stimulate the ability of DarR to bind to its target DNA sequences.
Overexpression of DarR Protein Is Toxic to M. smegmatis-To characterize the regulatory function of DarR, we examined the effect of changing the expression levels of darR in recombinant M. smegmatis strains. A darR-deleted mutant strain of  5-8). The protected regions are indicated. B, sequence and structural characteristics of the protected darR upstream region. The regions protected by DarR are underlined. The 14-bp sequences containing IRs separated by 4 bp are indicated by a pair of arrows. The translation start codon of darR is indicated in bold type. C, DNA substrates designed for assaying the DNA binding activity of DarR. Both the wild type (darR-S1) and mutated (darR-S2-S4) DNA substrate sequences are shown. D, EMSA assays for the DNA binding activity of DarR on DNA substrates with wild type IR sequence (lanes 1-4) and mutated IR sequences (lanes 5-16). The DNA substrate was co-incubated with 0.12-0.5 M of DarR protein.
M. smegmatis was first generated by a gene replacement strategy (22) (supplemental Fig. S6). We first used SEM to compare the morphological features of wild type and ⌬darR strain. A distinguishing morphological feature for mutant strain in comparison with wild type was increased cell length (Fig. 6A, middle  panel). We calculated the lengths of the bacterial cells from representative fields as visualized by SEM. Wild type M. smegmatis cells had lengths of 4.0 Ϯ 1.0 m, and ⌬darR cells had lengths of 6.7 Ϯ 1.5 m. The p values of the cell length data were calculated and were Ͻ 0.05. Strikingly, ⌬darR complementation cells restored the wild type growth morphology (Fig. 6A,  right panel).
We also used the mycobacterial expression vector pMV261 containing the strong Hsp60 promoter to overproduce DarR or DarR⌬N, in which the N-terminal DNA-binding HTH domain of darR had been deleted. Surprisingly, we could not obtain transformants containing pMV261-darR on the 7H10 medium. In contrast, transformants containing pMV261-darR⌬N were easily obtained (Fig. 6B). To obtain a moderate darR overexpression strain to facilitate investigation of the regulatory mechanism of DarR, we used the predicted darR upstream promoter to replace the strong hsp60 promoter of pMV261. Two new M. smegmatis strains, Msm/pMV261Pdar-darR and Msm/ pMV261 Pdar-darR⌬N, were successfully constructed (Fig.  6C). We observed that growth of the new darR overexpression M. smegmatis strains was still lower than that of the darR⌬N overexpression and wild type strains (Fig. 6D). These results indicate that overexpression of the darR gene is inhibitory to the growth of M. smegmatis at moderate levels and is toxic at high levels.
DarR Negatively Regulates Expression of Its Target Genes and Synthesis of Fatty Acids in M. smegmatis-We then compared the expression levels of the target genes in wild type, darRdeleted mutant, darR-complemented strain, and darR overexpression M. smegmatis strains using quantitative real time PCR assays. As shown in Fig. 7A, expression of three target genes were significantly up-regulated (p Ͻ 0.05) in the mutant strains compared with that in the wild type strain. However, expression level of the negative control gene Ms0030, which lacked the conserved DarR-binding motif in its promoter, was comparable with that in the wild type strain. Strikingly, we were able to complement the up-regulation of target genes in strain Msm⌬darR/pMV361Pdar-darR (Fig. 7A). This finding suggested to us that DarR could function as a negative regulator of its target genes in M. smegmatis. Further overexpression experiments also confirmed this observation. As shown in Fig. 7B, expression of three target genes was significantly down-regulated (p Ͻ 0.05) in the darR overexpression strains (which had ϳ5-10-fold lower expression) compared with that in the wild type strain or darR⌬N overexpression strain. Importantly, expression level of Ms0030, an unrelated gene, did not change upon overexpression of DarR.
DarR inhibited expression of two target genes associated with fatty acid metabolism and transportation, suggesting that DarR could be involved in the regulation of the synthesis of mycobacterial fatty acids. To test this hypothesis, we measured the contents of fatty acids in the darR and darR⌬N overexpression strains. As shown in Fig. 8 and supplemental Fig. S7, GC/MS analysis revealed that contents of six fatty acids: including hexadecanoic acid, 9-octadecenoic acid, octadecenoic acid, 10-methyl-octadecenoic acid, docosanoic acid, and tetracosa-  5-7) of 500 M c-di-AMP. Shifts in the DNA substrate caused by c-di-GMP were quantified, and the mean values of three experiments are shown in the lower panel. C, ChIP assays for the in vivo association of DarR with the upstream promoter sequences of target genes using preimmune (P) or immune sera (I) raised against DarR in wild type (left panels) and darR-deleted mutant strains (right panels). DNA recovered from the immunoprecipitates was amplified. gidB_up was used as a negative control. noic acid, were significantly down-regulated (p Ͻ 0.05) in the darR overexpression strain compared with that in the darR⌬N overexpression strain. This is consistent with our above data that DarR inhibited expression of fatty acid metabolism and transportation genes. Based on these results, we conclude that DarR functions as a repressor that negatively regulates expression of its target genes and synthesis of fatty acids in M. smegmatis.

DISCUSSION
Accumulating evidence shows that c-di-AMP plays an important role in controlling a variety of cellular processes in bacteria, mainly cell size, cell wall peptidoglycan homeostasis, and host immune response (4 -6). However, the essential regulatory mechanism, potential target genes, and specific receptors of c-di-AMP are still unknown. In the current study, we identified a TetR-like transcriptional factor, DarR, to be a bacterial c-di-AMP receptor regulator in M. smegmatis. c-di-AMP was found to stimulate the DNA binding activity of the DarR protein, which recognizes a 14-bp conserved sequence motif. Furthermore, DarR was found to function as a repressor and to negatively regulate expression of its target genes. To our knowledge, DarR is the first identified c-di-AMP responsive regulator in bacteria, and this study also represents the first description of an interaction between a regulator and c-di-AMP in bacteria.
In recent years, it has become clear that cyclic dinucleotides can act as second messengers and play important roles in multiple bacterial physiological processes and even the interaction of bacterial pathogens with host cells (1)(2)(3)(4)(5)(6). Much of these findings were made in recent studies on c-di-GMP receptor/effectors and their signaling mechanisms in bacteria. For example, FleQ, which regulates the expression of pel and other EPS genes that are necessary for biofilm formation in Pseudomonas aeruginosa, was the first transcriptional regulator to be identified as a c-di-GMP receptor/effector (12). In addition, c-di-GMP has been shown to control a variety of cellular processes including motility, differentiation, and virulence of bacterial pathogens (1,2). However, the significance of c-di-AMP, another cyclic dinucleotide second messenger, has remained to be explored. Identifying proteins that directly respond to c-di-AMP is critical for further understanding the signaling pathways it regulates. In the present study, we screened for potential c-di-AMP receptors/effectors in a M. smegmatis transcription factor library and successfully identified the first c-di-AMP responsive regulator DarR and found evidence for its interaction with c-di-AMP. Our findings provide further opportunities to explore the regulatory mechanism of c-di-AMP and its associated signal transduction pathway in bacteria.
The M. smegmatis genome encodes more than 500 potential regulatory factors (GenBank TM accession number CP000480). Strikingly, only one protein, DarR, from the M. smegmatis transcriptional factor library was found to strongly bind c-di-AMP. DarR was further identified to be a c-di-AMP responsive regulator. DarR is a TetR-like transcriptional factor containing a typical TetR-like N-terminal HTH DNA-binding domain and a QacR-like C-terminal domain (Fig. 2). In the currently available bacterial genomes, we found additional five conserved DarR orthologs with amino acid identities of 56.4 -84.6% by BLAST alignment analysis (supplemental Fig. S8A). DarR shares the highest amino acid residue identity with three mycobacterial orthologs when compared with the orthologs in Rhodococcus erythropolis and Corynebacterium variabile. More interestingly, these DarR homologs seem to recognize the same palindrome sequence motifs as M. smegmatis DarR (supplemental Fig. S8B). Therefore, the interaction between DarR and its target motif sequence, as well as its mode of regulation, could be highly conserved among several bacterial species.
An important finding from our study is that DarR functions as a repressor and negatively regulates expression of its target genes. Based on a 14-bp palindromic motif recognized by DarR, three target genes for DarR were characterized. One of them was Ms5696, which encodes a cold shock CspA family protein that usually regulates genes that respond to low temperature stress in bacteria (26 -28). This suggested that DarR could participate in responding to environmental stresses in M. smegmatis. Strikingly, the other two target genes were found to be asso- FIGURE 7. Differential expression assays of target genes in the wild type, mutant and darR-overexpressed strains. Mycobacterial cDNA were amplified, and quantitative real time PCR assays were performed as described under "Experimental Procedures." The relative expression levels of the genes were normalized using sigA gene as an invariant transcript, and an unrelated Ms0030 gene was used as a negative control. Relative expression levels of target genes in Msm and Msm/⌬darR (A) and in Msm/pMVPdar-DarR (B) recombinant strains were assayed. As a positive control, total DNA of each strain was used as template for PCR amplification. The cDNA of the mutant strains and the recombinant strain containing an empty pMV261 vector were used as templates in the negative controls. The data were analyzed using the 2 ⌬⌬Ct method (19,22). The relative expression data were analyzed for statistical significance by the unpaired two-tailed Student's t test using GraphPad Prism (version 5). FIGURE 8. Relative percentage of intracellular fatty acids. GC/MS analysis identified six intracellular fatty acids from M. smegmatis: hexadecanoic acid, 9-octadecenoic acid, octadecenoic acid, 10-methyl-octadecenoic acid, docosanoic acid, and tetracosanoic acid. Their contents were determined. The relative percentages of these fatty acids between Msm/pMVPdar-darR⌬N and Msm/pMVPdar-darR were calculated. The data were analyzed for statistical significance by the unpaired two-tailed Student's t test using GraphPad Prism (version 5). Dodecanoic acid methyl ester (Sigma) was added into the samples before extraction and used as an invariant fatty acid. The error bars indicate the variant range of the data derived from three biological replicates. ciated with fatty acid metabolism and transportation. Specifically, Ms5347 encodes a major facilitator family transporter, and Ms5348 encodes a medium chain fatty acyl-CoA ligase. Fatty acids are essential components of membranes and are important sources of metabolic energy in all organisms, and acyl-CoA synthetases/ligases are essential for lipid synthesis, fatty acid catabolism, and phospholipid remodeling (29 -32). Thus, DarR could sense c-di-AMP and play an important role in maintaining membrane lipid homeostasis in M. smegmatis by regulating the expression of fatty acid metabolism genes. Interestingly, the DarR homologs in C. variabile (CVAR_2936-9) and R. erythropolis (RER_49790) are also likely involved in regulating fatty acid metabolism, because conserved sequence motifs recognized by DarR are also found in the upstream regions of fatty acid metabolism operons in these organisms (supplemental Fig. S8C). These findings imply that c-di-AMP responsive DarR-like regulators might represent a new group of fatty acid metabolism-associated regulators, and this regulatory model could be conserved across different bacterial species. Recently, Yamamoto et al. (33) found that mutation of a TetRlike regulator resulted in overexpression of a transporter gene, mdrT, and concomitant increase in the cyclic di-AMP secretion in L. monocytogenes. As a major facilitator family transporter, Ms5347 might be involved in transportation of c-di-AMP in M. smegmatis. This remains to be determined in future work.