A Genome-Wide RegulatorâDNA Interaction Network in the Human

Article pubs.acs.org/jpr

A Genome-Wide Regulator−DNA Interaction Network in the Human Pathogen Mycobacterium tuberculosis H37Rv Jumei Zeng,† Tao Cui,† and Zheng-Guo He* National Key Laboratory of Agricultural Microbiology, Center for Proteomics Research, College of Life Science and Technology, Huazhong Agricultural University, Wuhan 430070, China S Supporting Information *

ABSTRACT: Transcription regulation translates static genome information to dynamic cell behaviors, making it central to understand how cells interact with and adapt to their environment. However, only a limited number of transcription regulators and their target genes have been identified in the pathogen Mycobacterium tuberculosis, which has greatly impeded our understanding of its pathogenesis and virulence. In this study, we constructed a genome-wide transcription regulatory network of M. tuberculosis H37Rv using a highthroughput bacterial one-hybrid technique. A transcription factor skeleton network was derived on the basis of the identification of more than 5400 protein−DNA interactions. Our findings further highlight the regulatory mechanism of the mammalian cell entry 1 (mce1) module, which includes mce1R and the mce1 operon. Mce1R was linked to global negative regulation of cell growth, but was found to be positively regulated by the dormancy response regulator DevR. Expression of the mce1 operon was shown to be negatively regulated by the virulence regulator PhoP. These findings provide important new insights into the molecular mechanisms of several mce1 module-related hypervirulence phenotypes of the pathogen. Furthermore, a model of mce1 module-centered signal circuit for dormancy regulation in M. tuberculosis is proposed and discussed. KEYWORDS: M. tuberculosis, transcription regulation network, Mce1, DevR

■

INTRODUCTION

TR depends on specific physical interactions between transcription factors (TFs) and the promoter sequences of their target genes. Advances in high-throughput genetic techniques have made it possible to analyze the direct interaction between a trans regulator and a cis-promoter at a whole genome scale. Such innovations include DNA microarray,9 ChIP-Seq,10 yeast one-hybrid system11 and bacteria one-hybrid system.12 To identify protein−DNA interactions (PDIs) for known TFs in M. tuberculosis, we screened the regulatory sequence library using a recently developed bacterial one-hybrid (B1H) technique.12 We characterized more than 5400 PDIs, whose overall quality was validated. A genome-wide regulatory network of M. tuberculosis was then constructed and network topology was analyzed. To effectively simplify the original network while preserving its major architecture, a TF−TF skeleton network was extracted, which highlighted the regulatory mechanism of the mammalian cell entry 1 (mce1) module, including mce1R and mce1 operon genes. On the basis of this experimental study and previous reports, we propose a model of mce1 module-centered circuit for regulation of dormancy in M. tuberculosis.

Mycobacterium tuberculosis is the causative agent of tuberculosis, a leading cause of mortality worldwide. During infection, the intracellular pathogen confronts a particular host cell environment, causing reduced oxygen tension and restricted access to nutrients, among others. M. tuberculosis can withstand various stress conditions encountered within phagosomes, but the genetic mechanisms behind this are unknown.1 Although the genome of M. tuberculosis encodes ∼200 putative transcription regulators, the identity and targets of only a few of them are known to date.2 The gene expression profiles of M. tuberculosis under various environmental stresses, such as hypoxia, nitric oxide release, nutrient starvation, low pH and drug exposure, and during infection of the lungs and macrophages have been extensively studied.3 Despite the abundance of studies on expression profiles, the transcription regulation (TR) mechanisms and regulators involved in mediating bacterial response to stress are largely unknown. Several studies have underscored the importance of having a comprehensive TR network for better understanding of the pathogenesis and survival of M. tuberculosis.4,5 By collecting documented regulatory information and mapping the homologous regulatory relationship of Escherichia coli and Corynebacterium glutamicum, several groups6−8 have assembled the M. tuberculosis TR network, which has improved our understanding of transcription regulation in M. tuberculosis. However, the regulators and target genes considered in such studies remain limited. © 2012 American Chemical Society

Received: June 9, 2012 Published: July 18, 2012 4682

dx.doi.org/10.1021/pr3006233 | J. Proteome Res. 2012, 11, 4682−4692

Journal of Proteome Research

■

Article

with the recombinant plasmids were grown at 37 °C in 200 mL of LB medium containing 50 mg/mL of kanamycin. Protein purification was carried out as described previously21 and verified by SDS-PAGE. Protein concentrations were determined by spectrophotometric absorbance at 260 nm.

MATERIALS AND METHODS

Definition of Operons and Promoters

The complete genome sequence of M. tuberculosis and gene location information was downloaded from NCBI (NC_000962.2). Genes with the same transcriptional directions and small intergenic distances (d) between adjacent genes (d ≤ 20) are hypothesized to compose a hypothetical operon that shares the same promoters, and were cotranscripted in vivo. Upstream 350-bp sequences of the first genes of each hypothetical operon were selected as hypothetical promoters for the determination of confidence levels.

RNA Polymerase II-Chromatin Immunoprecipitation (RNApol-ChIP) Assay

Each cotransformant containing both pROMOTER and pTRG plasmids was grown up to OD600 = 0.5, then IPTG was added to a final concentration of 0.5 mM, and the culture was incubated for 4 h at 37 °C with shaking. The culture was fixed with 1% formaldehyde at room temperature and stopped by adding glycine (0.125 M final concentration). Cross-linked cells were pelleted by centrifugation and washed in PBS and TBSTT (150 mM NaCl, 20 mM Tris-HCl pH 7.5, 10 mM EDTA, 0.1% Triton X-100, and 0.1% Tween 20). After resuspending in TBSTT, the sample was sonicated on ice. The average DNA fragment size obtained was approximately 0.5 kb. After sonication, insoluble debris was removed by centrifugation. The resulting extract was incubated under rotation for 3 h at 4 °C, with 10 μL of antibodies against RNA pol-α subunit.13,22 Subsequently, the complexes were immunoprecipitated with 20 μL of 50% protein Asepharose for 1 h under rotation at 4 °C. The immunocomplex was recovered by centrifugation, washed five times with 1× TBSTT, and resuspended with 100 μL of TE (20 mM Tris-HCl pH 7.8, 10 mM EDTA, and 0.5% SDS). Cross-linking was reversed for 1 h at 65 °C. The immunoprecipitated DNA was purified and resuspended in 50 μL of TE. DNA in the samples (1:10 dilution) were analyzed by quantitative PCR with corresponding primers (Table S1, Supporting Information), as described previously.23

Generation of B1H Promoter Prey Library and TFs Bait Strains

Using the pBXcmT vector,12 a pROMOTER vector13 was constructed and integrated with multiple cloning sites. About 2637 promoters were amplified using their respective primers (Table S1, Supporting Information) and cloned into the pROMOTER. The prey library for M. tuberculosis promoters was produced by mixing these recombinant plasmids. About 190 predicted transcription regulatory genes of M. tuberculosis2,14 were amplified using their primers,15 and 184 transcription regulators were successfully cloned into the pTRG vector. Bacterial One-Hybrid Screens

A pair of pROMOTER/pTRG plasmids was cotransformed into the reporter strain as described previously.12 Its growth was then tested together with the self-activation control on selective medium containing 3-AT, Kanr, Strr, and Chlr. Positive growth cotransformants were selected on selective screening medium containing 20 mM 3-AT, 16 μg/mL of streptomycin, 15 μg/mL of tetracycline, 34 μg/mL of chloramphenicol, and 50 μg/mL of kanamycin. The plates were incubated at 30 °C for 3−4 d. Potential positive clones were chosen and verified by sequencing, and interactors were identified by BLAST alignment.

Electrophoretic Mobility Shift Assay (EMSA)

DNA-binding assays of M. tuberculosis TF proteins were performed using a modified EMSA, as previously described.21 When required, acetylphosphate (ACP) (Sigma) was used to phosphorylate PhoP (Rv0757) or DevR (Rv3133c, also known as DosR) in vitro. Briefly, the reactions (10 μL) for measuring the mobility shift contained 200 fmol 32P-labeled DNA and various TF proteins diluted in a buffer containing 20 mM Tris-HCl (pH 7.5), 100 mM NaCl, 0.5 mM MgCl2, 10 μg/mL of sonicated salmon sperm DNA, 0.7 mM 2-mercaptoethanol and 5% glycerol. Gels were exposed to a storage-phosphor screen and the images were acquired using a Typhoon scanner (GE Healthcare).

In Silico DNA-Binding Motif Analysis

Conserved binding motifs for TFs were analyzed in silico with the MEME suite.16 The latest MEME suite software (4.3.0) was downloaded from the MEME web site (http://meme.sdsc.edu/) and run on a local Linux platform. Minimum and maximum motif widths were set to 8 and 12, respectively. The distribution of motifs was set to OOPS mode. The remaining algorithm-related parameters were kept at default settings. GO Annotation and GO Enrichment

The GO terms17 annotation of proteins encoded by the M. tuberculosis H37Rv genome was obtained from TBDB.6,18 The GO enrichment of target genes was performed by the online enrichment tool provided by TBDB (http://genome.tbdb.org/ an notation /genome/tbdb/BatchSelect.htm l?target= GeneEnrichment.html).18

Surface Plasmon Resonance (SPR) Analysis

Interactions between regulator proteins of M. tuberculosis and their target DNA were assayed using SPR. Biotin-labeled promoter DNA was immobilized onto a Streptavidin (SA) chip (BIAcore) by a previously published procedure.12 DNA−protein interaction assays were performed at 25 °C. An overlay plot was generated to show the interactions.

Network Visualization and Analysis of Topological Properties

M. tuberculosis Growth, Gene Overexpression and Antisense Inhibition Assays

Networks were visualized using Cytoscape.19 The topological properties of networks were analyzed using the NetworkAnalysis plug-in of Cytoscape.20 Edges in the regulatory networks were analyzed as directed.

M. tuberculosis H37Rv was incubated at 37 °C in 7H9 media with 0.05% Tween 80, 0.2% glycerol, and OADC supplement (Becton Dickinson). Cells were harvested when grown to midlog phase, and ChIP analysis was performed as described previously.24 Vectors for overexpressing Mce1R (Rv0165c), DevR, or PhoP were constructed using pMV261. To analyze the effects of an inhibitive expression of mce1R on its target genes, antisense mce1R was cloned into pMind (a TetR-controlled expression

Expression and Purification of Recombinant TF Proteins of M. tuberculosis

The transcription regulatory genes of M. tuberculosis were cloned into the pET28a overexpression vectors to produce corresponding recombinant vectors. E. coli BL21 (DE3) cells transformed 4683



Article

system).25 All recombinant plasmids were transformed into M. tuberculosis H37Rv. The strain containing the empty pMind/ pMV plasmid was used as the negative control. Heat-shock/ tetracycline was used to induce gene expression. Cells were grown at 37 °C with aeration in 7H9-Kan-Tw (7H9 medium supplemented with 0.5% Tween 80, 30 mg/mL of kanamycin and 0.2% glycerol) to an OD600 of 1.5−2.0. Cells were harvested at midlog phase for qRT-PCR assays. RNA Isolation

Total RNA was isolated from M. tuberculosis culture pellets using a Pure RNA Mini kit (Bioteke) as described previously.25 Briefly, the pellet was resuspended in an appropriate volume of the lysis buffer provided in the kit. Then the cell debris was removed by centrifugation at 2000g for 5 min at 4 °C. The supernatant was collected and processed according to the manufacturer’s instructions. To remove residual DNA from the RNA preparation, the RNA sample was incubated at 37 °C for 1 h with 100 units of DNase in the presence of DNase buffer, and then extracted with phenol/chloroform and precipitated with ethanol. The precipitated RNA was recovered by centrifugation at 13000g for 30 min at 4 °C and resuspended in 50 μL of RNasefree water. The A260 values of 1 μL samples were checked to assess yield. Quantitative Real-Time PCR (qRT-PCR)

For qRT-PCR analysis, gene-specific primers (Table S2, Supporting Information) were used, and first-strand cDNAs were synthesized using SuperScript II reverse transcriptase (Invitrogen) according to the manufacturer’s instructions. Each PCR reaction (20 μL) contained 10 μL of 2× SYBR Green Master Mix Reagent (Applied Biosystems), 1.0 μL of cDNA samples, and 200 nM primers. The reactions were performed in a Bio-Rad IQ5 RT-PCR machine. The thermocycling conditions were 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 specificity was assessed by conducting a melting curve analysis. Gene expression levels were normalized to the levels of 16S rRNA gene transcripts.22 The degrees of change in expression were calculated using the 2−ΔΔCt method.26

■

Figure 1. Schematic representation of genome-wide protein−DNA interaction screening and construction of transcriptional networks. A genome-wide protein−DNA interaction screen of M. tuberculosis was performed by the B1H method, and PDI and TR networks were constructed.

controlled by the promoter. The final regulatory network contains 184 TFs, 1116 target genes, and 8510 TF-gene interactions (Table S4, Supporting Information). The entire network is highly interconnected without any isolated nodes (Figure 1). The average path length of the network is 2.619, which suggests the “small-world” effect of the network. In the directed TR network, in-degree of a gene refers to the number of transcriptional factors regulating its expression, while out-degree of a regulator shows the number of genes under its direct control. The in-degree of M. tuberculosis B1H TR network ranges from 1 to 142 and follows a power-law distribution (Figure S1A, Supporting Information). The number of genes decline rapidly when in-degree increases. A total of 712 genes (63.79% of the network) are regulated by a single TF, which reflects specific regulation. The out-degree of regulators ranges from 1 to 130 and follows an almost Gaussian-like distribution (Figure S1B, Supporting Information). Rv2021c has the highest out-degree and regulates the expression of 130 genes. Degree analysis of a TR network can reflect its global topological properties, while network motifs can reveal local properties of a network.27 Motifs are small connected subnetworks that a network displays in significantly higher frequencies than would be expected for a random network. Motifs in TR networks represent the basic building blocks of the network and thus are essential for uncovering structural design principles that control gene expression in cells.28 Using fanmod to compare 100 derived random networks,29 we calculated the significance profiles for the M. tuberculosis B1H TR network focusing on the motifs of 3 nodes. Out of 13 possible directed 3-nodes motifs, motif 13 was not found in the B1H TR network. The first six motifs, which correspond to open motifs, were found more frequently than the random version of the TR network, while

RESULTS

Identification of Regulator−Promoter DNA Interactions in M. tuberculosis

We employed the well-established B1H system to generate a genome-wide PDI network of M. tuberculosis (Figure 1). In accordance with the defined operon map of M. tuberculosis H37Rv (see the Materials and Methods), 2637 promoters were amplified using their respective primers (Table S1, Supporting Information) and were cloned into the modified pBXcmT vector,12 designated as pROMOTER.13 A promoter library of M. tuberculosis was then generated by mixing these promoter recombinant plasmids. About 190 genes encoding M. tuberculosis TFs were cloned into the pTRG vector. All of those clones were further subjected to extensive screening against the M. tuberculosis promoter library using the B1H technique. Positive cotransformants were sequenced to identify PDIs. The final data set contains 184 TFs, 749 promoters and 5414 TF-promoter interactions (Table S3, Supporting Information). Construction and Topological Properties Analysis of M. tuberculosis TR Network

The M. tuberculosis TR network was then built by mapping each promoter in the PDI database to the genes whose expression was 4684



Article

Figure 2. Analysis of the TF−TF skeleton network. (A) In-degree distribution of the TF−TF network. (B) Out-degree distribution of the TF−TF network. (C) High in-degree nodes in the TF−TF network. (D) High in-degree nodes in the TF−TF network had a high degree of cross-regulation. (E) A simplified framework of the TF−TF network. Nodes in the TF−TF network were divided into five groups according to their in-degree. Each group is indicated with a different color. The sizes of the nodes correspond to the number of regulators in the groups. The height of the edges corresponds to the number of total interactions between two groups. (F) Loop paths in the M. tuberculosis TR network. (G) Rv2034-centered loop paths.

interaction or groups of genes coregulated by groups of regulators were identified (Figure S2, Supporting Information). Uncovering the biological function of these clusters will require wide range information combined with experimental confirmation in the future. In the present study, the regulatory mechanism of the mammalian cell entry 1 (mce1) module was further studied as a model of TF−TF interaction.

closed motifs were mostly underrepresented in the B1H TR network (Figure S1C, Supporting Information). Modules in the network are formed by groups of genes that play roles in related cellular function and are relatively isolated from other parts of the network. As full connectivity feature, no clearly physically separated modules could be directly identified in our current TR network. Several tools have been used in the past to find highly interconnected regions and identify modules in protein interaction networks. We performed a modules search in our current TR network using the MCODE tool that was originally designed for analysis of protein interaction networks.30 Many clusters formed by either relatively intensive TF−TF

Quality Evaluation of the PDI Network

Two independent methods, RNA polymerase II-chromatin immunoprecipitation (RNAPol-ChIP)13 and electrophoretic mobility shift assay (EMSA) were chosen for evaluating the quality of the PDI data. The high-degree feature in our network is 4685



Article

Figure 3. Mce1R as a global negative regulator. (A) ChIP assays for the binding of Mce1R with its putative target promoters in M. tuberculosis. ChIP using preimmune or anti-Mce1R antibody raised against Mce1R. DNA samples of the input and ChIP were purified and resuspended in 50 μL of TE. DNA recovered from immunoprecipitates was amplified with primers specific for target genes or to an unrelated mycobacterial promoter SigM. (B) The sequence motif predicted from the promoters of Mce1R target regulators. (C) qRT-PCR assays for the levels of target gene expression in M. tuberculosis overexpressing the antisense mce1R. Relative expression levels of genes were normalized using the 16S rRNA gene as an invariant transcript. Fold change in expression was calculated using the 2ΔΔCt method. All target genes were amplified using specific primers. Representative data are shown. (D) qRTPCR assays for target gene expression in M. tuberculosis overexpressing mce1R. (E) EMSA assays for the binding of Mce1R with its promoter DNA. 200 fmol 32P-labeled DNA substrates were coincubated with various amounts of the Mce1R protein (100−400 nM). Heat-denatured Mce1R protein was used as the negative control. (F) SPR assays for the binding of Mce1R with its own promoter. Biotin-labeled mce1R promoter was immobilized on an SA chip (BIAcore). Following a period of stabilization, Mce1R, heated-Mce1R or GST was passed over the chip. An overlay plot was then produced to look for interactions.

similar to that of previously published networks.31,32 However, high out-degree regulators and high in-degree promoters likely contain potential false positive interactions. Therefore, 10 high out-degree regulators, including Mce2R (Rv0586), IdeR (Rv2711), DevR, CarD (Rv3583c), Rv0275c, and Rv1219c,

and 3 high in-degree promoters, including dnaAp (Rv0001p), dnaKp (Rv0350p), and fadE26p (Rv3504), were first selected for validation by the RNAPol-ChIP approach (Figure S3, Supporting Information). In addition, 33 pairs of interactions involved in continuous promoters were randomly selected and evaluated. In 4686



Article

Figure 4. Regulation analysis of mce1R and the mce1 operon. (A) EMSA assays for the binding of DevR with the promoter of mce1R. 200 fmol 32Plabeled DNA substrates were coincubated with various amounts of the DevR protein (100−400 nM). The effect of phosphorylation of DevR on binding with the mce1R promoter was assayed by adding 50 mM ACP to the reactions. (B) SPR assays for binding of DevR with the mce1R promoter. Biotinlabeled mce1R promoter was immobilized on an SA chip (BIAcore). DevR, heat-denatured DevR, or GST was passed over the chip. An overlay plot was then produced to look for interactions. Assays for the effects of phosphorylation of DevR on binding with the mce1R promoter were carried out by adding 50 mM ACP to the reactions. (C) ChIP assays for the binding of DevR and PhoP with their target promoters in M. tuberculosis. ChIP using preimmune or specific antibody raised against DevR or PhoP. DNA recovered from immunoprecipitates was amplified with primers specific for target genes or to an unrelated mycobacterial promoter SigM. (D) qRT-PCR assays for changes in gene expression levels upon overexpression of DevR or PhoP. (E) EMSA assays for binding of PhoP with the mce1 promoter. (F) SPR assays for binding of PhoP with the mce1 promoter. Biotin-labeled mce1 promoter was immobilized on an SA chip. The effects of phosphorylation of PhoP on binding with the mce1 promoter were assayed by adding 50 mM ACP to the reactions.

total, 536 of 598 interactions were confirmed, and the rate of false positives was less than 10% (Figure S3, Supporting Information). The transcriptional regulators KmtR (Rv0827c), IdeR, WhiB3 (Rv3416), and Rv3066, as well as the 33 pair interactions were also independently validated by EMSA experiments. Addition-

ally, 230 of 261 DNA-binding activities were confirmed, and the average false rate was appropriate 12%. A summary of the regulator-gene interactions validated by the two different methods described here is provided in Table S5 (Supporting Information). 4687



Article

activity of Mce1R with its own promoter was confirmed by EMSA and surface plasmon resonance (SPR) assays (Figure 3E,F). To further investigate the potential function of Mce1R, all of its target genes were analyzed in the context of the GO catalog (Table S7, Supporting Information). None of the biological processes were found to be significantly enriched, which suggested a wide regulatory role of Mce1R in cells. However, it is notable that many target genes (Rv0282, Rv0283, Rv0284, Rv0285, Rv0286, Rv0674, Rv0675, Rv0700, Rv0701, Rv0702, Rv0703, Rv1194c, Rv2500c, Rv2827c, Rv3177, Rv3597c) are involved in the process of cell “growth” (GO: 0040007). Notably, one of the target proteins Rv0284 is homologous to the FtsK/SpoIIIE family of proteins, suggesting that Rv0284 may play important roles during cell division. Taken together, our network analysis and expression assays indicate that Mce1R represents a global repressor in M. tuberculosis, likely by regulating genes involved in cell growth and cell cycle.

Taken together, about 800 pairs of PDIs were evaluated, and more than 80% of the interactions were confirmed (Table S6, Supporting Information). Extraction of the TF−TF Skeleton Network and Its Characteristics

The extraction of a TF−TF skeleton network from the TR network can effectively simplify it while preserving its architecture and information on the regulatory signal stream. A TF−TF skeleton network of M. tuberculosis was produced, and the degree of each TF was calculated. Interestingly, it followed power-law and had Gaussian distribution (Figure 2A,B). There were about 18 TFs with in-degree higher than 10 and comprised the high in-degree nodes (Figure 2C). A certain number of input regulation of these high in-degree group TFs came from members of the group itself; these nodes were thus highly regulated internally by each other (Figure 2D). For a more intuitive exploration of network architecture and signal stream, a simplified framework of the TF−TF network was drawn by group TFs according to their in-degree (Figure 2E). The framework showed that direct regulation of targets was the most common mode of regulation in the network. The multicascade portion decreased along with increase in cascade hierarchy (Figure 2E). Further path analysis of the TF−TF network identified the number of loop paths (Figure 2F). In general, more than one loop path exists for a TF. For example, when the maximum path length was limited to three, there were a total of 12 loop paths for the ArsR-type repressor protein Rv2034c (Figure 2G). These regulators are possibly responsible for the multifeedback paths and stability of the global network. Figure S4 (Supporting Information) shows the individual loop path network for each regulator.

Dormancy Regulator DevR Positively Regulates the Expression of mce1R

The upstream regulator of mce1R was analyzed further in the current regulatory network to explore the potential actions of Mce1R in cell processes. The mce1R gene was found to be regulated by a two-component response regulator (DevR) involved in the regulation of M. tuberculosis dormancy (Table S4, Supporting Information). This interaction was confirmed by B1H spot assay (Figure S5C, Supporting Information). An EMSA assay was performed to further validate the specific DNAbinding activity of DevR in vitro. As shown in Figure 4A, a band of retarded mobility, indicating protein bound to the DNA, was evident when mce1R promoter DNA was mixed with DevR. DevR is the response regulator of the DevRS two-component system and can be phosphorylated by the cognate histidine sensor kinase DevS.33 To test the effect of phosphorylation on the ability of DevR to bind to the mce1R promoter, an EMSA assay was performed using acetylphosphate (ACP)-preincubated DevR protein. Interestingly, phosphorylation of DevR resulted in a weaker DNA-binding activity as evidenced by the fewer DevR/ DNA complexes observed on the gel compared with the reaction with unphosphorylated DevR (Figure 4A). The interaction of DevR with the mce1R promoter and the effect of phosphorylation were further confirmed by chip assay (Figure 4B). When an increasing concentration of DevR protein (100−500 nM) was passed over the chip surface, an increasing trend in response values was observed (Figure 4B, upper panel). By contrast, both heated inactive protein and GST showed no response (Figure 4B, upper panel). This result is consistent with those of the EMSA assays. More interestingly, a conserved binding motif (TTSGGGACTWWAGTCCCSAA) of DevR was also identified in the promoter of mce1R.34 This suggests that the potential binding region possesses a sequence motif that specifically interacts with unphosphorylated DevR (Table S8, Supporting Information). DevR was also found bind to the mce1R promoter in M. tuberculosis in a ChIP assay. Figure 4C shows that the mce1R promoter DNA was specifically recovered by immunoprecipitation using the specific anti-DevR antiserum. In contrast, the preimmune serum failed to precipitate significant amounts of DNA. Furthermore, overexpression of DevR (by about 10-fold) increased the expression of mce1R by about 5-fold in M. tuberculosis (Figure 4D, left panel).

Mce1R Functions as a Global Negative Regulator in the Network

In the TF−TF skeleton network, a GntR family repressor, Mce1R, directly regulated 12 regulators including Mce1R itself, VirS (Rv3082c), Lsr2 (Rv3597c), FurA (Rv1909c), three WhiB family regulators (WhiB2 (Rv3260c), WhiB5 (Rv0022c) and WhiB7 (Rv3197A)), DNA replication initiation protein DnaA, and four other TFs (Rv1049, Rv1990c, Rv1219c, Rv2034). This suggests that Mce1R is a potential high-level regulator. The interactions between Mce1R and most of these target promoters were reconfirmed by B1H spot assay and validated by EMSA using purified Mce1R protein (Figure S5A,B, Supporting Information). In addition, using ChIP with anti-Mce1R antibody, we confirmed the recognition by Mce1R regulator of these target promoter DNA in M. tuberculosis H37Rv (Figure 3A), except for the promoter of Rv2034. The sequence of these promoters was further analyzed in silico with MEME, and an 8-bp conserved DNA motif was identified (Figure 3B). By RT-PCR assays, up-regulation in the expression of target regulator genes was clearly observed when the antisense mRNA of the mce1R gene was expressed through a pMind recombinant plasmid (Figure 3C). Conversely, a significant down-regulation in the expression of these target genes was observed when Mce1R was overexpressed using the pMV expression vector (Figure 3D). Interestingly, a significant up-regulation of mce1R (about 4.5-fold) was observed one day after induction. However, down-regulation of most of the target regulators did not occur in that time frame (Figure 3D, upper panel). At a later time point (after seven days), the mce1R gene itself was also down-regulated (Figure 3D, lower panel). This phenomenon can be attributed to feedback regulation of Mce1R because the specific DNA-binding 4688



Article

Figure 5. Schematic representation of regulation of the mce1 module and a signal circuit involved in dormancy regulation in M. tuberculosis. (A) A schematic representation of regulation of mce1R and mce1 operons. (B) Dormancy signal circuit model. Production of the mce1 operon stimulates and activates the host immune system. Immune response induces activation of the two-component response regulator DevR, which activates the global repressor Mce1R, which then inhibits active growth and facilitates dormancy in M. tuberculosis. PhoP negatively regulates expression of the mce1 operon.

obtained for the heat-denatured protein or GST (Figure 4F, lower panel). The binding of PhoP with the promoter of the mce1 operon in M. tuberculosis was also confirmed by ChIP assays. As shown in Figure 4C, the mce1 operon promoter DNA was specifically recovered by immunoprecipitation using anti-PhoP antiserum. In contrast, the preimmune serum failed to precipitate significant amounts of DNA. The physiological significance of the interaction between PhoP and mce1p was studied further by overexpression experiments in M. tuberculosis. When phoP was cloned into the pMV261 vector and overexpressed (about 7fold) in M. tuberculosis (Figure 4D, right panel), about 2-fold down-regulation of expression of the mce1 operon was observed (Figure 4D, right panel). These results strongly suggest that PhoP can physically interact with the mce1 promoter, and this interaction can be inhibited by phosphorylation. Consistent with this, we observed that PhoP can negatively regulate expression of the mce1 operon in M. tuberculosis.

Our results suggest that DevR can specifically bind with the promoter of mce1R both in vitro and in vivo and positively regulates mce1R expression. Virulence Regulator PhoP Regulates the Expression of mce1 Operon

In our network, the two-component response regulator PhoP, which is a key virulence regulator,35 participates in the regulation of the mce1 operon (Table S4, Supporting Information), which is tightly associated with the activation of host immune response during infection.36 Binding of PhoP to the mce1 promoter was confirmed by gel-shift assays in which a characteristic protein/ DNA complex appeared while the amount of free promoter DNA decreased with increasing PhoP concentration (Figure 4E). The negative effect of phosphorylation on the ability of PhoP to bind to mce1p was clearly observed when protein samples were preincubated with ACP before using for assays. As shown in Figure 4E, phosphorylated PhoP had very low DNA binding ability. The interaction between PhoP and mce1p DNA as well as the effect of PhoP phosphorylation on their interaction was confirmed by SPR assay. As shown in Figure 4F, when an increasing concentration of the PhoP protein (100−500 nM) was passed over the chip surface, a corresponding increase in response values was observed (Figure 4F, upper panel). By contrast, ACP-preincubated PhoP protein showed almost no response when passed over the chip, which is similar to the result

■

DISCUSSION TR translates static genome information to dynamic cell behaviors, making it central to understand how cells interact and adapt to their environment. In the case of pathogens, a comprehensive understanding of their transcription regulatory network would enable the identification of new regulators, and aid the tracing of regulatory circuits that allow them to respond to 4689



Article

mce1 operon has been shown to be tightly associated with the activation of host immune response during infection.36 In the current study, Mce1R was found to act as a global negative regulator of cell growth. Additionally, Mce1R was identified as a downstream target of DevR, and PhoP was determined to be a negative regulator of the mce1 operon (Figure 5A). These findings reveal a signaling pathway that links the initial dormancy signal to the final dormancy-associated cellular processes (Figure 5B). In this model, PhoP regulates expression of the M. tuberculosis mce1 operon, which stimulates immune response. DevR detects the immune response signals such as NO or hypoxic stress. If a robust immune system exists, devR is activated and dormancy is induced via activation of a global negative regulator Mce1R, which participates in regulating dormancy by repressing the expression of cell cycle and metabolism-associated genes. In this signaling model, mutation of the mce1 operon, devR, or mce1R breaks the signal stream and prevents the activation of dormancy, leading to active bacterial growth and replication. Several M. tuberculosis mutants exhibit rapidly growing hypervirulence phenotypes.42 However, the mechanism of such phenotypes remains poorly understood. Interestingly, the components of the signaling pathway identified here, namely devR, mce1R and the mce1 operon, have previously been reported to be associated with hypervirulence. According to our model, hypervirulence is caused by the failure of regulatory circuits to cooperate in controlling M. tuberculosis dormancy lifestyles, thereby causing uncontrolled replication. This is consistent with the known function of DevR, which is to respond to hypoxia and induce dormancy in bacteria.41 Deletion of devR generates a hypervirulent mutant that grows more rapidly in immunocompetent mice and in gamma interferon-activated macrophages.43 The mce1 operon is responsible for stimulating host immune responses, and its disruption causes a decrease in host immune response during infection.36 A robust immune response is required for inducing and maintaining the dormancy of M. tuberculosis in vivo. Despite lacking a direct role, the mce1 operon contributes to dormancy regulation when an immune system exists. The hypervirulence phenotype in mutants of the mce1 operon may result from failures in stimulating the immune response, thereby causing failure to activate dormancy. Mce1R negatively regulates expression of the mce1 operon.44 That mutation of mce1R also causes hypervirulence is unexpected45 considering the hypervirulence phenotype of the mce1 operon mutation.36 In the current study, we provide evidence to support that Mce1R negatively regulates bacterial growth and replication as a global regulator. In this model, hypervirulence caused by the mce1R mutation may be caused by a different mechanism that is independent of the mce1 operon. One possible explanation is that mutation in the mce1 operon reduces the host immune response and inhibits dormancy at an earlier stage. By contrast, Mce1R mutation leads to derepression and tends to activate infection, leading to failure in execution of dormancy signals at a later stage. Another phenotype of the mce1R mutation is that it can be more rapidly reactivated from dormancy than wild-type strains in a drug-treated mouse model.46 A possible explanation for this is that mce1R mutants do not actually enter dormancy but simply have temporary inhibition of growth. PhoP has been shown to be an essential virulence regulator in M. tuberculosis.35 A single nucleotide polymorphism within the PhoPR operon has been confirmed to be the major factor contributing to avirulence in H37Ra.47 However, mechanisms for the contribution of PhoP to virulence are largely unknown

host environments. In the present study, we constructed a genome-wide transcription regulatory network of the human pathogen M. tuberculosis on the basis of the B1H assay. Following a TF-centered screen strategy, almost all the TFs encoded by M. tuberculosis were included in the global network, covering about one-third of the defined hypothetical promoters. Studies on specific PDIs and large-scale TR networks have become central topics in both prokaryotes and eukaryotes in the postgenomic era.31,32,37,38 However, most studies focus primarily on changes in the expression of individual genes, single regulons, or distinct subunits of the complex TR network of M. tuberculosis. In addition, most interactions derived from expression profiles are indirect protein−DNA interactions, and only a handful of interactions have been confirmed experimentally. Some previously or recently reported interactions were found in our global regulatory network (Table S9, Supporting Information). Importantly, a significant proportion of our interaction data was confirmed through various independent methods (80% success rate). The success rate in our network is similar to that of other published networks.31,32 Several resources such as TBDB6 and MycoRegNet systems39 have been designed as data repositories and provide bioinformatic platforms for transcriptional regulatory network visualization, analysis, and reconstruction. We will release the data reported in this study for public use by integrating them into existing platforms in the near future. The entire network is highly interconnected without any isolated nodes. Two kinds of regulation tend to facilitate this connectivity feature of a TR network. Direct regulation between regulators could link different regulators and genes under their control, while indirect links between different regulators could be established if these regulators shared common target genes. In our analysis of the B1H TR network, both of these possibilities were found to have contributed significantly to this connectivity feature on the basis of successful identification of many genes regulated by more than one regulator and considerable number of TF−TF interactions (see module analysis part). The peculiarly high in-degrees were found in the M. tuberculosis B1H TR network and a single M. tuberculosis gene or gene cluster could be regulated by as many as 142 regulators (Figure 2). While similar in-degrees have been observed for human PDI network,31 the observation is still unique for microbes. From a cellular function perspective, genes regulated by more than one regulator may be responsible for a common routine activated by a variety of signals. On the other hand, TF−TF interactions may underlie signaling cascades and feedback regulation. We calculated the significance profiles of the M. tuberculosis B1H TR network using fanmod. Our results are consistent with that of a previous study on M. tuberculosis TR network in which motif 13 was absent from the Z-scores representation.8 However, the significance profiles of open motifs and closed motifs were almost opposite in the two studies, except that motif 11 was found more frequently in both TR networks. This discrepancy may be due to differences in the source of the data as well as dense difference of the network. The emergence of closed motifs depends on the intense interaction between transcriptional factors, which may occur more frequently in data that contains indirect regulatory interactions, such as the case of a study by Sanz et al. that integrated a diversity of experimental sources.8 Cell cycle arrest and downregulation of growth are two main features of dormant bacteria. M. tuberculosis enters dormancy when it encounters various stressful conditions exerted by the host immune system. The regulator DevR can respond to host environments and induce dormancy of M. tuberculosis.40,41 The 4690



■

ABBREVIATIONS ACP, acetylphosphate; PCR, polymerase chain reaction; TB, tuberculosis; SPR, surface plasmon resonance; TR, transcription regulation; TFs, transcription factors; PDIs, protein−DNA interactions; B1H, bacterial one-hybrid; SA, Streptavidin; GO, gene ontology; EMSA, electrophoretic mobility shift assay; GST, glutathione S-transferase; IPTG, isopropyl-D-1-thiogalactopyranoside

because of limited knowledge on its direct target genes. In the present study, PhoP was found to bind to the promoter of the mce1 operon, and overexpression of PhoP caused downregulation of expression of the mce1 operon. Thus, a possible mechanism for the contribution of PhoP to virulence of M. tuberculosis is repression of the expression of the mce1 operon. Interestingly, a recent study showed that PhoP controls M. tuberculosis ESAT-6 secretion and specific T cell recognition and may contribute to virulence in part by facilitating pathogen−host interaction.48 Similar to the ESAT-6 system, production of the mce1 operon is another essential component that participates in interaction with host immune systems.36 Thus, both our results and those of previous studies support the model that the importance of PhoP in virulence may largely depend on its regulation of genes that interact with host immune systems. In addition, we found that phosphorylation negatively regulates the binding of PhoP to the mce1 operon. It is possible that an unknown physiological condition dissociates phosphorylated PhoP from the mce1 operon, thereby allowing expression of the mce1 operon.

■

CONCLUSION In the present study, a genome-wide TF-promoter DNA interaction screen was conducted, and a global TR network of M. tuberculosis was constructed. The network contains more than 5400 pairs of PDIs, covering almost all TFs and one-third of M. tuberculosis genes. Further analysis of the network, in combination with experiments, uncovered a molecular regulatory pathway of the mce1 module. A signal circuit for dormancy regulation of M. tuberculosis is proposed. Our findings provide insight into the molecular mechanisms of several mce1 modulerelated hypervirulence phenotypes of the pathogen. The TR network reported here should serve as a significant resource for further studies on gene regulatory pathways in M. tuberculosis. ASSOCIATED CONTENT

* Supporting Information S

Figures S1−S5 and Tables S1−S9. This material is available free of charge via the Internet at http://pubs.acs.org.

■

REFERENCES

(1) Rengarajan, J.; Bloom, B. R.; Rubin, E. J. Genome-wide requirements for Mycobacterium tuberculosis adaptation and survival in macrophages. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 8327−32. (2) Cole, S. T.; Brosch, R.; Parkhill, J.; Garnier, T.; Churcher, C.; Harris, D.; Gordon, S. V.; Eiglmeier, K.; Gas, S.; Barry, C. E., 3rd; Tekaia, F.; Badcock, K.; Basham, D.; Brown, D.; Chillingworth, T.; Connor, R.; Davies, R.; Devlin, K.; Feltwell, T.; Gentles, S.; Hamlin, N.; Holroyd, S.; Hornsby, T.; Jagels, K.; Krogh, A.; McLean, J.; Moule, S.; Murphy, L.; Oliver, K.; Osborne, J.; Quail, M. A.; Rajandream, M. A.; Rogers, J.; Rutter, S.; Seeger, K.; Skelton, J.; Squares, R.; Squares, S.; Sulston, J. E.; Taylor, K.; Whitehead, S.; Barrell, B. G. Deciphering the biology of Mycobacterium tuberculosis from the complete genome sequence. Nature 1998, 393, 537−44. (3) Rohde, K. H.; Veiga, D. F.; Caldwell, S.; Balázsi, G.; Russell, D. G. Linking the transcriptional profiles and the physiological states of Mycobacterium tuberculosis during an extended intracellular infection. PLoS Pathog. 2012, 8, e1002769. (4) Graham, J. E.; Clark-Curtiss, J. E. Identification of Mycobacterium tuberculosis RNAs synthesized in response to phagocytosis by human macrophages by selective capture of transcribed sequences (SCOTS). Proc. Natl. Acad. Sci. U.S.A. 1999, 96, 11554−9. (5) Manganelli, R.; Voskuil, M. I.; Schoolnik, G. K.; Smith, I. The Mycobacterium tuberculosis ECF sigma factor sigmaE: role in global gene expression and survival in macrophages. Mol. Microbiol. 2001, 41, 423− 37. (6) Galagan, J. E.; Sisk, P.; Stolte, C.; Weiner, B.; Koehrsen, M.; Wymore, F.; Reddy, T. B.; Zucker, J. D.; Engels, R.; Gellesch, M.; Hubble, J.; Jin, H.; Larson, L.; Mao, M.; Nitzberg, M.; White, J.; Zachariah, Z. K.; Sherlock, G.; Ball, C. A.; Schoolnik, G. K. TB database 2010: overview and update. Tuberculosis 2010, 90, 225−35. (7) Balázsi, G.; Heath, A. P.; Shi, L.; Gennaro, M. L. The temporal response of the Mycobacterium tuberculosis gene regulatory network during growth arrest. Mol. Syst. Biol. 2008, 4, 225. (8) Sanz, J.; Navarro, J.; Arbués, A.; Martín, C.; Marijuán, P. C.; Moreno, Y. The transcriptional regulatory network of Mycobacterium tuberculosis. PLoS One 2011, 6, e22178. (9) Mukherjee, S.; Berger, M. F.; Jona, G.; Wang, X. S.; Muzzey, D.; Snyder, M.; Young, R. A.; Bulyk, M. L. Rapid analysis of the DNAbinding specificities of transcription factors with DNA microarrays. Nat. Genet. 2004, 36, 1331−9. (10) Jothi, R.; Cuddapah, S.; Barski, A.; Cui, K.; Zhao, K. Genome-wide identification of in vivo protein-DNA binding sites from ChIP-Seq data. Nucleic Acids Res. 2008, 36, 5221−31. (11) Deplancke, B.; Dupuy, D.; Vidal, M.; Walhout, A. J. A gatewaycompatible yeast one-hybrid system. Genome Res. 2004, 14, 2093−101. (12) Guo, M.; Feng, H.; Zhang, J.; Wang, W.; Wang, Y.; Li, Y.; Gao, C.; Chen, H.; Feng, Y.; He, Z. G. Dissecting transcription regulatory pathways through a new bacterial one-hybrid reporter system. Genome Res. 2009, 19, 1301−8. (13) Zeng, J.; Li, Y.; Zhang, S.; He, Z. G. A novel high-throughput B1H-ChIP method for efficiently validating and screening specific regulator-target promoter interactions. Appl. Microbiol. Biotechnol. 2012, 93, 1257−69. (14) Camus, J. C.; Pryor, M. J.; Médigue, C.; Cole, S. T. Re-annotation of the genome sequence of Mycobacterium tuberculosis H37Rv. Microbiology. 2002, 148, 2967−73. (15) Wang, Y.; Cui, T.; Zhang, C.; Yang, M.; Huang, Y.; Li, W.; Zhang, L.; Gao, C.; He, Y.; Li, Y.; Huang, F.; Zeng, J.; Huang, C.; Yang, Q.; Tian,

■

■

Article

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]; [email protected]. edu.cn. Tel: +86-27-87284300. Fax: +86-27-87280670. Author Contributions †

These authors equally contributed to this work.

Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The M. tuberculosis H37Rv and MDR10093006 strains were provided by Dr. J. Xiang (Wuhan Medical Treatment Center, Wuhan, China). We thank Prof. G. Zhao (Laboratory of Systems Biology, Chinese Academy of Sciences, Shanghai, China) for his kind help. This study was funded by the National Natural Science Foundation of China (31025002 and 30930003), Creative Research Groups of Hubei, and Hubei Chutian Scholar Program (to Z.-G.H.). 4691



Article

tuberculosis that is hypoxia-responsive in the absence of the DNAbinding domain of DevR. Microbiology 2004, 150, 865−75. (34) Chauhan, S.; Tyagi, J. S. Interaction of DevR with multiple binding sites synergistically activates divergent transcription of narK2Rv1738 genes in Mycobacterium tuberculosis. J. Bacteriol. 2008, 190, 5394−403. (35) Ryndak, M.; Wang, S.; Smith, I. PhoP, a key player in Mycobacterium tuberculosis virulence. Trends Microbiol. 2008, 16, 528− 34. (36) Shimono, N.; Morici, L.; Casali, N.; Cantrell, S.; Sidders, B.; Ehrt, S.; Riley, L. W. Hypervirulent mutant of Mycobacterium tuberculosis resulting from disruption of the mce1 operon. Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 15918−23. (37) Blais, A.; Dynlacht, B. D. Constructing transcriptional regulatory networks. Genes Dev. 2005, 19, 1499−511. (38) Lee, T. I.; Rinaldi, N. J.; Robert, F.; Odom, D. T.; Bar-Joseph, Z.; Gerber, G. K.; Hannett, N. M.; Harbison, C. T.; Thompson, C. M.; Simon, I.; Zeitlinger, J.; Jennings, E. G.; Murray, H. L.; Gordon, D. B.; Ren, B.; Wyrick, J. J.; Tagne, J. B.; Volkert, T. L.; Fraenkel, E.; Gifford, D. K.; Young, R. A. Transcriptional regulatory networks in Saccharomyces cerevisiae. Science 2002, 298, 799−804. (39) Krawczyk, J.; Kohl, T. A.; Goesmann, A.; Kalinowski, J.; Baumbach, J. From Corynebacterium glutamicum to Mycobacterium tuberculosistowards transfers of gene regulatory networks and integrated data analyses with MycoRegNet. Nucleic Acids Res. 2009, 37, e97. (40) Park, H. D.; Guinn, K. M.; Harrell, M. I.; Liao, R.; Voskuil, M. I.; Tompa, M.; Schoolnik, G. K.; Sherman, D. R. Rv3133c/dosR is a transcription factor that mediates the hypoxic response of Mycobacterium tuberculosis. Mol. Microbiol. 2003, 48, 833−43. (41) Sherman, D. R.; Voskuil, M.; Schnappinger, D.; Liao, R.; Harrell, M. I.; Schoolnik, G. K. Regulation of the Mycobacterium tuberculosis hypoxic response gene encoding alpha-Crystallin. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 7534−9. (42) ten Bokum, A. M.; Movahedzadeh, F.; Frita, R.; Bancroft, G. J.; Stoker, N. G. The case for hypervirulence through gene deletion in Mycobacterium tuberculosis. Trends Microbiol. 2008, 16, 436−41. (43) Parish, T.; Smith, D. A.; Kendall, S.; Casali, N.; Bancroft, G. J.; Stoker, N. G. Deletion of two-component regulatory systems increases the virulence of Mycobacterium tuberculosis. Infect. Immun. 2003, 71, 1134−40. (44) Casali, N.; White, A. M.; Riley, L. W. Regulation of the Mycobacterium tuberculosis mce1 operon. J. Bacteriol. 2006, 188, 441−9. (45) Uchida, Y.; Casali, N.; White, A.; Morici, L.; Kendall, L. V.; Riley, L. W. Accelerated immunopathological response of mice infected with Mycobacterium tuberculosis disrupted in the mce1 operon negative transcriptional regulator. Cell Microbiol. 2007, 9, 1275−83. (46) Cheigh, C. I.; Senaratne, R.; Uchida, Y.; Casali, N.; Kendall, L. V.; Riley, L. W. Posttreatment reactivation of tuberculosis in mice caused by Mycobacterium tuberculosis disrupted in mce1R. J. Infect. Dis. 2010, 202, 752−9. (47) Lee, J. S.; Krause, R.; Schreiber, J.; Mollenkopf, H. J.; Kowall, J.; Stein, R.; Jeon, B. Y.; Kwak, J. Y.; Song, M. K.; Patron, J. P.; Jorg, S.; Roh, K.; Cho, S. N.; Kaufmann, S. H. Mutation in the transcriptional regulator PhoP contributes to avirulence of Mycobacterium tuberculosis H37Ra strain. Cell Host Microbe 2008, 3, 97−103. (48) Frigui, W.; Bottai, D.; Majlessi, L.; Monot, M.; Josselin, E.; Brodin, P.; Garnier, T.; Gicquel, B.; Martin, C.; Leclerc, C.; Cole, S. T.; Brosch, R. Control of M. tuberculosis ESAT-6 secretion and specific T cell recognition by PhoP. PLoS Pathog. 2008, 4, e33.

Y.; Zhao, C.; Chen, H.; Zhang, H.; He, Z. G. Global protein−protein interaction network in the human pathogen Mycobacterium tuberculosis H37Rv. J. Proteome Res. 2010, 9, 6665−77. (16) Bailey, T. L.; Elkan, C. Fitting a mixture model by expectation maximization to discover motifs in biopolymers. Proc. Int. Conf. Intell. Syst. Mol. Biol. 1994, 2, 28−36. (17) Ashburner, M.; Ball, C. A.; Blake, J. A.; Botstein, D.; Butler, H.; Cherry, J. M.; Davis, A. P.; Dolinski, K.; Dwight, S. S.; Eppig, J. T.; Harris, M. A.; Hill, D. P.; Issel-Tarver, L.; Kasarskis, A.; Lewis, S.; Matese, J. C.; Richardson, J. E.; Ringwald, M.; Rubin, G. M.; Sherlock, G. Gene ontology: tool for the unification of biology. The Gene Ontology Consortium. Nat. Genet. 2000, 25, 25−9. (18) Reddy, T. B.; Riley, R.; Wymore, F.; Montgomery, P.; DeCaprio, D.; Engels, R.; Gellesch, M.; Hubble, J.; Jen, D.; Jin, H.; Koehrsen, M.; Larson, L.; Mao, M.; Nitzberg, M.; Sisk, P.; Stolte, C.; Weiner, B.; White, J.; Zachariah, Z. K.; Sherlock, G.; Galagan, J. E.; Ball, C. A.; Schoolnik, G. K. TB database: an integrated platform for tuberculosis research. Nucleic Acids Res. 2009, 37, D499−508. (19) Shannon, P.; Markiel, A.; Ozier, O.; Baliga, N. S.; Wang, J. T.; Ramage, D.; Amin, N.; Schwikowski, B.; Ideker, T. Cytoscape: a software environment for integrated models of biomolecular interaction networks. Genome Res. 2003, 13, 2498−504. (20) Assenov, Y.; Ramírez, F.; Schelhorn, S. E.; Lengauer, T.; Albrecht, M. Computing topological parameters of biological networks. Bioinformatics 2008, 24, 282−4. (21) Li, Y.; Zeng, J.; Zhang, H.; He, Z. G. The characterization of conserved binding motifs and potential target genes for M. tuberculosis MtrAB reveals a link between the two-component system and the drug resistance of M. smegmatis. BMC Microbiol. 2010, 10, 242. (22) Sandoval, J.; Rodríguez, J. L.; Tur, G.; Serviddio, G.; Pereda, J.; Boukaba, A.; Sastre, J.; Torres, L.; Franco, L.; López-Rodas, G. RNAPolChIP: a novel application of chromatin immunoprecipitation to the analysis of real-time gene transcription. Nucleic Acids Res. 2004, 32, e88. (23) Sala, C.; Forti, F.; Magnoni, F.; Ghisotti, D. The katG mRNA of Mycobacterium tuberculosis and Mycobacterium smegmatis is processed at its 5′ end and is stabilized by both a polypurine sequence and translation initiation. BMC Mol. Biol. 2008, 9, 33. (24) Zhang, L.; Zhang, L.; Liu, Y.; Yang, S.; Gao, C.; Gong, H.; Feng, Y.; He, Z. G. Archaeal eukaryote-like Orc1/Cdc6 initiators physically interact with DNA polymerase B1 and regulate its functions. Proc. Natl. Acad. Sci. U.S.A. 2009, 106, 7792−7. (25) Yang, M.; Gao, C.; Wang, Y.; Zhang, H.; He, Z. G. Characterization of the interaction and cross-regulation of three Mycobacterium tuberculosis RelBE modules. PLoS One 2010, 5, e10672. (26) Livak, K. J.; Schmittgen, T. D. Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) method. Methods 2001, 25, 402−8. (27) Shen-Orr, S. S.; Milo, R.; Mangan, S.; Alon, U. Network motifs in the transcriptional regulation network of Escherichia coli. Nat. Genet. 2002, 31, 64−8. (28) Alon, U. Network motifs: theory and experimental approaches. Nat. Rev. Genet. 2007, 8, 450−61. (29) Wernicke, S.; Rasche, F. FANMOD: a tool for fast network motif detection. Bioinformatics 2006, 22, 1152−3. (30) Bader, G. D.; Hogue, C. W. An automated method for finding molecular complexes in large protein interaction networks. BMC Bioinf. 2003, 4, 2. (31) Hu, S.; Xie, Z.; Onishi, A.; Yu, X.; Jiang, L.; Lin, J.; Rho, H. S.; Woodard, C.; Wang, H.; Jeong, J. S.; Long, S; He, X; Wade, H; Blackshaw, S; Qian, J; Zhu, H.. Profiling the human protein-DNA interactome reveals ERK2 as a transcriptional repressor of interferon signaling. Cell 2009, 139, 610−22. (32) Deplancke, B.; Mukhopadhyay, A.; Ao, W.; Elewa, A. M.; Grove, C. A.; Martinez, N. J.; Sequerra, R.; Doucette-Stamm, L.; Reece-Hoyes, J. S.; Hope, I. A.; Tissenbaum, H. A.; Mango, S. E.; Walhout, A. J. A genecentered C. elegans protein-DNA interaction network. Cell 2006, 125, 1193−205. (33) Saini, D. K.; Malhotra, V.; Dey, D.; Pant, N.; Das, T. K.; Tyagi, J. S. DevR-DevS is a bona fide two-component system of Mycobacterium 4692


A Genome-Wide RegulatorâDNA Interaction Network in the Human

Recommend Documents

A Genome-Wide RegulatorâDNA Interaction Network in the Human