Mycobacterium tuberculosis Invasion of Macrophages: Linking Bacterial Gene Expression to Environmental Cues

Cell Host & Microbe

Resource Mycobacterium tuberculosis Invasion of Macrophages: Linking Bacterial Gene Expression to Environmental Cues Kyle H. Rohde,1 Robert B. Abramovitch,1 and David G. Russell1,* 1Cornell University, College of Veterinary Medicine, Department of Microbiology and Immunology, Ithaca, NY 14853, USA *Correspondence: [email protected] DOI 10.1016/j.chom.2007.09.006

SUMMARY

A central feature of Mycobacterium tuberculosis (Mtb) pathogenesis is the ability of Mtb to survive within macrophages (MØ). Despite its critical importance, our appreciation of the interplay between these two cells remains superficial. We employed microarrays to conduct a stepwise dissection of Mtb-MØ interaction during the invasion of resting bone marrow MØ. Contrary to many bacterial pathogens, engagement by MØ receptors without internalization did not alter Mtb gene expression. Subsequently, a high-resolution profile of Mtb invasion-linked gene expression was generated by assaying the Mtb transcriptome at 20 min intervals up to 2 hr postinfection. Transcriptional responses were detected within minutes of phagocytosis, including gene subsets with distinct temporal profiles. Pharmacological manipulation of phagosomal pH and in vitro acid stress studies revealed that vacuole acidification is an important trigger for differential gene expression. Finally, there are marked species-specific differences in the response of Mtb and M. bovis BCG to intraphagosomal cues. INTRODUCTION The success of Mycobacterium tuberculosis (Mtb) as a pathogen is evidenced by its parasitism of approximately one-third of its target population: mankind. Attempts to control TB are exacerbated by long-term persistence of Mtb, TB-HIV coinfection, and the emergence of multidrug-resistant (MDR) Mtb strains. Following inhalation of aerosolized bacilli, Mtb are taken up by alveolar MØ, within which Mtb survive by arresting phagosome maturation and avoiding fusion with lysosomes. The Mtb vacuole is characterized by incomplete acidification due to limited accumulation of vacuolar ATPases and failure to acquire the lysosomal GTPase Rab7 or mature lysosomal hydrolases (Russell, 2001). Although the arrest in phagosome maturation implies a niche of minimal hostility, it is unclear

what pressures intracellular bacilli face. Moreover, this setting is in flux because MØ activation overrides phagosome arrest and delivers Mtb to increasingly inhospitable compartments (Schaible et al., 1998). Perception of sensory cues is critical for bacteria in transition between different environments. In the Mtb-MØ interaction, engagement of the bacilli by phagocytic receptors is the first opportunity for crosstalk. Once internalized, Mtb is exposed to antimicrobial defenses such as reactive oxygen intermediates (ROI), reactive nitrogen intermediates (RNI), and free fatty acids (Adams et al., 1997; Akaki et al., 1997). The induction of genes involved in b-oxidation of fatty acids and the requirement for isocitrate lyase during persistent infections (McKinney et al., 2000) suggest that phagosomes impose nutritional limitations that force a metabolic shift toward a reliance on fatty acids. The level of single ions within the vacuole lumen may also trigger differential gene expression (Wagner et al., 2005). Even the minimal acidification of the Mtb vacuole could influence bacterial growth and gene expression within MØs. As reviewed by Kendall et al., most microarray studies to date examining Mtb regulatory networks have profiled gene expression in response to in vitro stimuli thought to be relevant in vivo (Kendall et al., 2004b). However, these approaches are limited by our knowledge of parameters within the vacuole and are unlikely to mimic in vivo environments adequately. Transcriptional profiling of Mtb in in vivo models of infection has proven to be technically challenging. Recent transcriptome analysis of Mtb isolated from lungs of infected mice (Talaat et al., 2004) and human tuberculosis patients (Rachman et al., 2006) highlighted potential virulence factors expressed in vivo. Schnappinger et al. first characterized the Mtb transcriptome within MØs, which bore testimony to an oxidative, nitrosative, and nutrient-limited environment (Schnappinger et al., 2003). However, the complexity of these models and resulting gene lists do not clearly define the pertinent signals or context of the observed responses. The current study was designed to investigate the dynamics of Mtb global transcriptional responses and perceived cues during MØ invasion. Appreciation of the timing of gene regulation is vital to the interpretation of complex microarray analyses of pathogen-host interactions. First, the temporal patterns of gene expression beginning within minutes of phagocytosis revealed that

352 Cell Host & Microbe 2, 352–364, November 2007 ª2007 Elsevier Inc.

Cell Host & Microbe Macrophage Cues and Mtb Gene Expression

adaptation of Mtb to the phagosome is a rapid and ongoing process. Second, we establish a link between specific environmental signals (adhesion and phagosome pH) and the transcriptional response, providing proof-of-principle for studying the role of individual variables within the complex phagosome environment. Finally, we observed substantial species-specific differences in the response of Mtb CDC1551 and M. bovis BCG to the phagosomal environment. This suggests that altered gene regulation contributes to the attenuation of BCG and that BCG is a poor surrogate for Mtb in gene expression studies. RESULTS AND DISCUSSION RNA Amplification and Intracellular Transcriptional Profiling The isolation of sufficient RNA for microarray analysis presents an obstacle to transcriptional profiling of intracellular pathogens. To overcome this limitation, we validated the use of linear RNA amplification for microarray analysis of Mtb isolated from MØ phagosomes. We routinely attained >1000-fold amplification from 100 ng of RNA derived from intramacrophage Mtb. Upon microarray comparison, the previously documented bias inherent in RNA amplification (Puskas et al., 2002) was evident in the moderate correlation (R = 0.6–0.7) between amplified RNA (aRNA) and unamplified template RNA (data not shown). However, consistent with observations that bias introduced during linear amplification of eukaryotic poly(A) mRNA were highly reproducible (Puskas et al., 2002), aRNA amplified independently from the same template exhibited excellent correlation (R = 0.95–0.98). Accordingly, only amplified samples were used for microarray analyses in this study. Reproducible detection of Mtb genes induced at consecutive time points during MØ interactions further validated this technique. Finally, the expression profiles of selected genes as assayed by microarray or qRT-PCR of unamplified and amplified aRNA at multiple time points postinfection were concordant (data not shown). These data validate bacterial RNA amplification for the transcriptional profiling of an intracellular pathogen and provide a powerful tool with potential application to animal models or clinical specimens in which bacterial RNA is scarce. Does MØ Adherence Trigger Changes in Mycobacterial Gene Expression? To begin a systematic examination of MØ-induced Mtb gene expression, we tested the hypothesis that engagement of Mtb by MØ surface receptors serves as the earliest signal sensed by Mtb. Host cell contact is known to induce virulence gene expression in numerous pathogens including Yersinia pseudotuberculosis and Neisseria meningitidis (Grifantini et al., 2002; Pettersson et al., 1996). Although Mtb was shown to trigger signaling cascades within MØ (Koul et al., 2004), the reciprocal effect of host cell contact on Mtb has not been addressed.

Mtb CDC1551 was bound to MØ in the presence of cytochalasin D (cytoD), an inhibitor of actin polymerization and phagocytosis. In control experiments, MØ treated with 5 mM cytoD bound but did not internalize fluorescein-labeled beads or mycobacteria (data not shown). Microarray comparison of Mtb bound to the MØ surface but not internalized with control Mtb treated with cytoD in the absence of MØ failed to reveal significant changes in gene expression (Figure 1A). These data indicate that, unlike pathogens such as Yersina and Salmonella that respond to cell contact during invasion (Pettersson et al., 1996; Zierler and Galan, 1995), the induction of Mtb genes does not appear to precede internalization by professional phagocytes.

Early Transcriptional Response to Invasion of Resting Macrophage Real-time assays using model particles have shown that changes in pH and hydrolytic activity occur within minutes of phagosome formation (Yates et al., 2005). To monitor global gene expression of Mtb during its passage into the phagosome, we isolated Mtb from infected MØ at 20 min intervals from 5 min to 2 hr postinfection for microarray analysis. The early response of Mtb to the phagosome milieu consisted predominantly of gene induction, with few genes downregulated 2 hr postinfection (Figure 1B). We identified 68 early genes significantly upregulated (>1.5-fold, p < 0.05) within MØ at the time course endpoint (2 hr) (Table S1 in the Supplemental Data available with this article online). Microarray results were validated by qRT-PCR analysis (data not shown) of select phagosome-induced genes (Rv1405c, Rv1403c, MT2466, MT2467, whiB7). The kinetic expression profiles shown in Figure 1C illustrate the surprising rapidity of the Mtb response to phagosomal cues. Applying static statistical cutoffs alone to a time course study, however, ignores the temporal relationship of the data. With this in mind we utilized the EDGE (extraction and analysis of differential gene expression) methodology of Storey et al. to identify time-dependent transcriptional changes (Storey et al., 2005). Because this method is computationally intensive, the input data set included genes (176) upregulated marginally at 2 hr postinfection with expression ratios >1 (p < 0.1). This approach confirmed the induction of most 2 hr genes identified using static thresholds and identified an additional 75 early phagosome-induced genes (Table S2). This data set included multiple genes involved in fatty acid metabolism (fadD8–9, fadE26–28, fadE34, Rv3537, ufaA2, icl, umaA1) as well as mpt70 and mpt83, prominent secreted antigens known to be MØ induced at later times (Schnappinger et al., 2003). Isocitrate lyase, encoded by icl, is a key enzyme in the glyoxylate shunt pathway for fatty acid b-oxidation that is required for survival in IFN-g activated MØ and persistent infection of mice (McKinney et al., 2000). Temporal microarray data coupled with EDGE analysis allowed for highly sensitive detection of gene induction at lower levels and earlier time points than is feasible using static, single-time point analyses.

Cell Host & Microbe 2, 352–364, November 2007 ª2007 Elsevier Inc. 353

Cell Host & Microbe Macrophage Cues and Mtb Gene Expression

Figure 1. Transcriptional Response of Mycobacteria to Early Macrophage Interactions (A) Mtb expression profile after 2 hr binding to cytochalasin-D-treated MØ. The y axis plots expression ratios on a log10 scale. Black horizontal lines indicate 1.5-fold changes in expression. (B) Mtb intraphagosomal expression profile 2 hr postinfection (average of four biological replicates). Genes up >1.5-fold (p < 0.05) compared to time-matched extracellular controls are indicated in red. (C) Kinetics of early Mtb global transcription in phagosome at 20 min intervals (normalized to t0 = 5 min) from 5 min to 2 hr postinfection. Each line depicts the expression level of a single gene. For clarity, only a subset of early induced genes is shown.

Among the earliest and most highly induced genes were the adjacent ORFs MT3290.1 and MT3290.2 (Table S1). MT3290.1 encodes WhiB7, a member of the WhiB family of putative transcriptional regulators unique to actinomycetes (Soliveri et al., 2000). The drug sensitivity of a whiB7 mutant led Morris et al. to propose that WhiB7,

induced 70-fold by aminoglycoside antibiotics, regulates a multidrug resistance system in mycobacteria (Morris et al., 2005). However, the only 3-fold induction of whiB7 within MØ and the lack of antibiotics during infections suggest alternate cues for whiB7 expression in vivo. Notably, it was shown that whiB7 was also induced in vitro by heat shock, iron starvation, growth phase, and fatty acids (Geiman et al., 2006). Two additional members of the whiB family, whiB3 and whiB6, were also induced by phagosomal cues (Table S1). The reported induction of whiB3 in mouse lungs after 2 weeks and in naive murine MØ after 6 hr corroborates our results (Banaiee et al., 2006). WhiB3 binds to the principle sigma factor RpoV and contributes to the rate of killing and tissue damage in the lung (Steyn et al., 2002). The role of WhiB6 in Mtb pathogenesis remains to be elucidated. Like whiB7, both whiB3 and whiB6 can be induced by several stressful in vitro stimuli (Geiman et al., 2006). Proteins of the WhiB family may facilitate an integrated transcriptional response to multiple stresses encountered within the phagosome. Two-component signal transduction systems (TCS) are prime candidates for the regulation of Mtb virulence factors within macrophages, with 6 of 16 known TCS response regulators implicated directly in Mtb virulence (Ewann et al., 2002; Malhotra et al., 2004; Parish et al., 2003a; Zahrt and Deretic, 2001). DosR mediates Mtb survival during hypoxia-induced dormancy and is required for virulence in the guinea pig model (Boon and Dick, 2002; Malhotra et al., 2004). We noted that a subset of the DosR regulon was induced within 2 hr of MØ invasion (Figure 1B; Table S1), extending previous reports of MØspecific expression of DosR-controlled genes at later times (Schnappinger et al., 2003). It is unlikely that the rapid, low-level induction of DosR in resting MØ is due to hypoxia or nitric oxide (NO) (Voskuil et al., 2003). In this context, carbon monoxide (Kumar et al., 2007) and hydrogen peroxide (Kendall et al., 2004a) are alternate plausible cues for DosR. The PhoPR regulon was also induced by 2 hr, with 25 of 44 genes reported by Walters et al. to be activated by PhoP (Walters et al., 2006) significantly upregulated (Tables S1 and S2). MØ invasion induced two PhoPR-dependent polyketide synthase loci required for the synthesis of acyltrehaloses and sulfolipids, methylbranched cell wall components restricted to pathogenic mycobacteria (Gonzalo Asensio et al., 2006). Thus, remodeling of the cell wall within phagosomes may be an important adaptation for Mtb intracellular survival. Additional PhoPR-dependent genes induced in early phagosomes included lipF, a lipase necessary for survival in mice (Camacho et al., 1999), and the putative MT2466MT2467-Rv2396 operon with limited homology to proton or metal transporting ATPases (Table S1). Mtb mutants lacking PhoPR are markedly attenuated in macrophages and mice, highlighting the importance of this TCS for Mtb fitness in vivo (Perez et al., 2001). The related PhoPQ TCS of Salmonella typhimurium reportedly detects multiple phagosome-relevant stimuli, including low Mg2+, low pH, and antimicrobial peptides (Groisman and Mouslim, 2006). However, the intracellular condition(s) that

354 Cell Host & Microbe 2, 352–364, November 2007 ª2007 Elsevier Inc.

Cell Host & Microbe Macrophage Cues and Mtb Gene Expression

Figure 2. Transcriptome Dynamics during Sustained Intracellular Survival (A) Mtb expression profile 24 hr postinfection compared to timematched extracellular control (average of five biological replicates). Genes up or down >1.5-fold (p < 0.05) are indicated in red. (B) Temporal expression profiles of early (2 hr) phagosome-induced genes at 24 hr. Subsets of early genes returned to control levels by 24 hr (red) continued to increase (data not shown) or plateau between 2 and 24 hr (data not shown). A fourth set of delayed early genes were not induced significantly until after 2 hr (black).

activates the Mtb PhoPR regulon within the MØ phagosome is unknown. The expression profiles of other TCS regulons and the signals they sense within MØ remain to be identified. Transcriptome Dynamics during Sustained Intracellular Survival Long-term survival of Mtb within its dynamic niche likely requires continual modulation of gene expression. To place the early ‘‘knee-jerk’’ reaction to MØ invasion in the context of an enduring host-pathogen interaction, we compared the transcription profiles of Mtb isolated from MØ 24 hr postinfection with the early (2 hr) data set. Not surprisingly, a broader transcriptional response was evident at 24 hr, with 115 genes significantly upregulated (>1.53, p < 0.05) (Figure 2A; Table S3). Only a subset of these genes (50%) was upregulated >1.5-fold in the 24 hr phagosomal transcriptome reported by Schnap-

pinger et al. (2003), likely due to technical differences (i.e., Mtb strain, array platform). The distinct fates of early-induced genes at 24 hr illustrate the complex dynamics of the evolving response of Mtb to intracellular cues (Figure 2B). Transcripts of one subset of 13 genes remained relatively constant up to 24 hr while a second group of 10 genes continued to increase over the 24 hr time course (Table S3). Genes with this latter profile, including members of the DosR regulon, ahpD, whiB6, icl, and genes of unknown function (MT2466), likely facilitate long-term residence within MØ through resistance to antimicrobial effectors, manipulation of host cell biology, or metabolic adaptation to the vacuolar niche. Finally, we observed a set of 38 genes transiently upregulated at 2 hr whose expression declined thereafter (Figure 2B), including whiB7, whiB3, bfrB, pks2, the Rv3612c–3616c operon reported to be regulated by the PhoPR TCS (Walters et al., 2006), and a pair of putative S-adenosylmethionine-dependent methyltransferases (Rv1403c, Rv1405c). The growth defect of a mutant in the transiently induced PrrAB TCS during the initial phase of MØ infection indicates that subtle, transient changes in gene expression can impact Mtb pathogenesis (Ewann et al., 2002). The largest set of genes (76) upregulated in the 24 hr phagosome exhibited delayed kinetics (Figure 2B; Table S3), with no increases in their expression apparent 2 hr postinfection. The delayed induction of fadB2, fadD26, fadE5, fadE29, echA20, glta1, and desA2 are indicative of ongoing modulation of fatty acid metabolic pathways. Prolonged exposure to intracellular stressors also led to upregulation of genes induced in vitro by oxidative stress (furA, katG, rubAB, Rv1460-Rv1466), general stress response (hsp, dnaK, grpE, groEL1, groEL2), and nutrient deprivation (lat). Subsets of genes within the DosR (Park et al., 2003), PhoPR (Walters et al., 2006), MprAB (He et al., 2006), and SenX3-RegX3 (Parish et al., 2003b) TCS regulons displayed distinct temporal expression profiles. Thus, it is likely that additional regulators are acting on the genes of TCS regulons within multitiered regulatory networks to integrate multiple stimuli into an appropriate response. Though often overshadowed by emphasis on gene induction, the coordinated repression of genes likely contributes to survival within MØ by conserving energy and precursors under nutrient-limiting conditions and/or minimizing expression of potential antigens within MØ. Whereas the initial Mtb response to phagocytosis predominantly involved gene induction, transcript levels of 71 genes were reduced (>1.5-fold, p < 0.05) after 24 hr exposure to the phagosome environment (Figure 2A; Table S4). The decreased expression of transcriptional regulators whiB2 (Rv3260c) and Rv3676 suggests a role in the regulation of these genes. The attenuation of a mutant lacking the cAMP receptor protein (CRP) Rv3676 supports the idea that specific downregulation of genes within the phagosome is important for Mtb virulence. Finally, of the 21 lipase (lip) genes and 37 fatty acid CoA ligase (fadD) genes encoded by Mtb, only a few members of these enzyme families (lipL, lipR, lipU, fadD5, and

Cell Host & Microbe 2, 352–364, November 2007 ª2007 Elsevier Inc. 355

Cell Host & Microbe Macrophage Cues and Mtb Gene Expression

fadD30) were repressed within the late phagosome. Diverse expression profiles for genes involved in lipid metabolism may reflect fluctuations in the nature (chain length, degree of saturation) and abundance of fatty acid substrates available to Mtb in vivo (Zhang et al., 2005). Virulence factors required for growth and dissemination in vivo are often differentially regulated during infection, prompting us to compare the early intramacrophage transcriptome and genes identified in published Mtb mutant screens. Only a small set of genes induced early after MØ invasion in this study were critical for fitness or survival in in vivo models of infection. For example, mutants in icl, ufaA2, Rv2396, and Rv3027c were attenuated for growth in MØ (Stewart et al., 2005), whereas mutants lacking Rv1128c, Rv1405c, Rv1460, Rv1465, Rv2707, Rv3502, Rv3534c, Rv3551, fadD26, and lipF exhibited growth defects in mouse lung or spleen (Camacho et al., 1999; Sassetti and Rubin, 2003). Mutants in a putative lipid degradation operon (Rv3540c–Rv3545c) upregulated at 2 hr and 24 hr postinfection were attenuated in both MØ and mouse in vivo models (Chang et al., 2007; Rengarajan et al., 2005; Sassetti and Rubin, 2003). This limited degree of overlap seemingly supports the conclusion by Rengarajan et al. that a majority of Mtb genes required for intracellular survival are constitutively expressed (Rengarajan et al., 2005). Alternatively, gene expression may be differentially regulated during infection of MØ or whole animals under as yet untested time points or conditions. Incomplete genome coverage of mutant libraries used for in vivo screens, including the absence of essential genes, also reduces the input gene set being assayed. Finally, functional redundancy, compensatory regulatory changes, or subtle mutant phenotypes not revealed by survival assays may contribute to this apparent disparity between gene expression and requirement for in vivo fitness. Effect of Phagosome Acidification on Mtb Intracellular Gene Expression The intramacrophage transcriptome of Mtb is a composite response assimilating multiple cues such as pH, ion concentrations, and nutritional and oxidative stress. Thus, studying the effect of a single variable on Mtb gene expression in vitro is unlikely to yield a complete picture. To begin dissecting the role of specific variables in the context of the phagosome, we examined the contribution of pH to intracellular Mtb gene expression. Although the inhibition of vacuole acidification is a hallmark of Mtb cell biology, Mtb phagosomes do acidify to pH 6.4 within minutes of uptake (Sturgill-Koszycki et al., 1994). To explore the effects of this limited acidification, Mtb were isolated from MØ treated with concanamycin A (CcA), a specific inhibitor of vacuolar ATPases, or untreated control MØ 2 hr postinfection. Consistent with a previous study of model bead phagosomes, treatment with 100 nM CcA effectively blocked the acidification of phagosomes containing bacilli labeled with pH-sensitive carboxyfluoroscein (Yates et al., 2005 and data not shown). We found a substantial number of early phagosomal genes (30/68) that were not induced in Mtb isolated from CcA-

treated MØ (Figure 3A; Table 1). However, the continued induction of some 2 hr genes, including whiB7, MT3290.2, and ahpCD, even in vacuoles with neutral pH, demonstrated that additional intraphagosomal cues are detected by the bacteria. Next, we examined the effect of pH on Mtb gene expression in vitro to verify that CcA-sensitive genes are responding to phagosome acidification. To mimic the acidity encountered by Mtb in intracellular vacuoles, Mtb gene expression profiles were determined in buffered, rich medium at pH 6.5 (resting MØ) or 5.5 (activated MØ) and compared to gene expression of Mtb incubated at pH 7.0. We observed 117 and 291 genes that were significantly upregulated (>1.53, p < 0.05) at pH 6.5 or pH 5.5, respectively (Figure 3B; Figure S1; Tables S19 and S21), and we observed 45 and 388 genes that were significantly downregulated at pH 6.5 or pH 5.5, respectively (Tables S20 and S22). Many of these genes showed a greater up- or downregulation at pH 5.5 as compared to pH 6.5 (Figure S1), demonstrating a clear dependence of gene expression on relative acidity of the environment. Previous reports of the induction of lipF and Rv0834c (Saviola et al., 2003) and Rv1130, Rv1131 (gltA1), MT0196, and Rv3614c– 3616c (Fisher et al., 2002) by in vitro acid stress corroborate our results. The pH-regulated genes we identified only show limited overlap with a prior microarray study of Mtb transcriptional responses to acid stress (Fisher et al., 2002), likely attributable to significant differences in Mtb strains, growth conditions (shaking versus standing), and duration of acid stress. However, our data revealed a striking overlap between genes that are upregulated by pH in vitro and CcA sensitive in the MØ (Table 1; Figures 3B and 3C). Of the 24 CcA-sensitive genes in vivo, 16 were also upregulated by acidic pH in vitro (Figure 3B). Thus, at early times following invasion of resting MØ, the mild acidification of Mtb phagosomes appears to serve as an important cue for transcriptional adaptation to this intracellular niche. While the mechanisms by which Mtb responds to pH are poorly characterized, the global transcriptional response of Mtb to acid stress points to several regulators that may respond to the cue of phagosomal acidification (Figure 4). Several of the pH- and CcA-sensitive genes, including lipF, pks2, pks3, Rv2331, Rv2633, Rv2396, and Rv3612c–3616c, were previously reported to be induced by PhoP (Walters et al., 2006), suggesting that this TCS may also regulate pH-dependent gene expression in Mtb. Interestingly, phoP itself is downregulated under similar conditions (Figure 4), consistent with the ability of Mtb PhoP to autoinhibit its own transcription (Gupta et al., 2006). The in vitro induction of whiB3 at pH 4.5 (Geiman et al., 2006), pH 5.5 and pH 6.5, coupled with the inhibition of whiB3 phagosomal induction by CcA, suggests an in vivo role for this regulator in the response of Mtb to phagosome acidification. The sigma factor genes sigB and sigH, induced by multiple stresses including oxidative stress, heat shock, and MØ infection (Rodrigue et al., 2006), were upregulated at pH 5.5, but not at pH 6.5 (Tables S19 and S21). Along similar lines, exposure to

356 Cell Host & Microbe 2, 352–364, November 2007 ª2007 Elsevier Inc.

Cell Host & Microbe Macrophage Cues and Mtb Gene Expression

Figure 3. Phagosome Acidification Modulates Mtb Gene Expression (A) Prior to infection, macrophages were treated with 100 nM concanamycin A (CcA) to inhibit acidification of Mtb-containing phagosome from pH 7 to pH 6.4. Vertical axis depicts expression ratios in untreated phagosome (pH 6.4) relative to CcA-treated phagosome (pH 7.0) 2 hr postinfection. Genes whose induction at 2 hr was sensitive to CcA (>1.5-fold, p < 0.05) are shown in red. Red arrows indicate examples of genes induced at 2 hr in untreated macrophages whose expression was insensitive to CcA. (B) Venn diagram showing overlap of genes induced by in vitro acid stress (pH 6.5 and 5.5) with CcA-sensitive phagosome-induced genes. Genes induced >1.5-fold (p < 0.05) in at least one condition and whose array signals passed quality filters in all three conditions were included. (C) Gene tree depicting the expression of CcA-sensitive phagosome-induced genes under in vitro acid stress. Includes genes from Table 1 also present in at least 10 of 12 in vitro acid array experiments.

pH 5.5 but not pH 6.5 repressed genes required for protein synthesis, replication, and energy metabolism (Figure S2A) and induced additional genes involved in general stress responses (Figure S2B). These acid-induced transcription profiles are consistent with the robust growth of Mtb in mildly acidic phagosomes and inhibited replication in more acidic, more stressful compartments. Species-Specific Programs of Intraphagosomal Gene Expression Mycobacterium bovis BCG is used frequently as an avirulent surrogate for Mtb. However, the significant genetic and regulatory differences between Mtb and BCG (Behr et al., 1999; Mostowy et al., 2004; Pym et al., 2002) call

into question how appropriate BCG may be for certain experimental questions. Thus, we examined the transcriptional response of BCG to the MØ phagosome to assess the impact of species-specific gene regulation on host interactions and BCG attenuation. Total BCG RNA was isolated from intracellular bacteria at 20 min intervals from 5 min to 2 hr and at 24 hr postinfection and analyzed by microarray as described previously. Extracellular BCG were used as species-matched controls to ensure that transcriptional profiles reflect differential responses to MØ-derived cues and not species-specific variation in baseline expression patterns. Like Mtb CDC1551, BCG responded to MØ invasion rapidly, with genes showing transient, sustained, and

Cell Host & Microbe 2, 352–364, November 2007 ª2007 Elsevier Inc. 357

Cell Host & Microbe Macrophage Cues and Mtb Gene Expression

Table 1. Comparison of In Vivo and In Vitro Acid-Responsive Genes In Vitro pH Gene Name Ratio p Value Induced Description MT1122.1

1.73

2.20E-02

+

HP

MT2466

1.53

2.13E-01

+

HP, possible H+-transporting ATPase

MT2467

1.57

9.70E-02

MT2467.1 (Rv2396)

1.40

1.23E-01

HP, homologous to heavy metal-translocating P-type ATPase ND

PE-PGRS family

ND

PE family

MT3102

1.51

7.80E-02

MT3106.1

1.95

4.70E-02

HP

MT3223

2.20

1.01E-02

Rv0574c

1.51

1.60E-01

ND

HP

Rv0575c

1.51

4.60E-02

ND

putative FAD-dependent oxidoreductase, energy production and conversion

Rv1180 (pks3)

2.07

8.00E-03

+

putative polyketide b-ketoacyl synthase

HP

a

Rv1181 (pks4)

1.42

1.22E-01

+

Rv1182 (papA3)

1.51

4.15E-02

+a

putative polyketide b-ketoacyl synthase probably conserved polyketide synthase-associated protein

Rv1403c

1.45

1.40E-02

+

SAM-dependent methyltransferase

Rv1405c

1.76

4.00E-02

SAM-dependent methyltransferase

Rv1535

1.60

6.00E-03

C terminus homologous to eukaryotic histone-lysine Nmethyltransferase

Rv1682

2.14

1.00E-03

ND a

putative coiled-coil structural protein PE family protein

Rv1806

1.74

5.20E-02

+

Rv1807

1.63

9.20E-02

+

Rv2007c (fdxA)

1.68

3.20E-02

Rv2030c

1.79

1.35E-01

ND

C-terminal homology to phosphoribosyltransferase, N-terminal homology to erythromycin esterase

Rv2031c (hspx)

2.40

3.20E-02

+a

a-crystallin, Hsp20 family of small stress-induced chaperone

a

PPE family protein ferredoxin

+

a-crystallin coregulated gene, nitroreductase

Rv2032 (acg)

1.87

2.20E-02

Rv2329c (nark1)

1.57

3.60E-02

Rv2331

1.92

5.00E-03

+

putative nitrate reductase, molybdopterin-binding (MopB) superfamily

Rv2389c

1.39

1.18E-01

+

putative transglycosylase

Rv2390c

1.80

2.30E-02

+

membrane protein, similar to Rv1362c and Rv1363c mce proteins (50% similar)

Rv2626c

1.52

2.00E-01

Rv2627c

2.03

2.70E-02

probale nitrite extrusion protein, 12 transmembrane domains, major facilitator superfamily

contains pair of CBS domains putative hydrolase a

HP

Rv2628

1.69

8.40E-02

+

Rv2632c

1.32

7.00E-02

+

CHP

Rv2633c

1.61

7.00E-03

+

similar to hemerythrin HHE cation-binding domain, NO response and cell wall physiology

Rv3021c

1.98

5.20E-02

Rv3130c

1.56

1.83E-01

ND

putative triacylglycerol synthase

Rv3416 (whiB3)

1.67

1.20E-02

+

whiB family transcriptional regulator

Rv3487c (lipF)

1.59

3.00E-03

+

putative lipase/esterase

Rv3532

1.75

3.50E-02

ND

PPE family

Rv3612c

1.43

1.21E-01

ND

CHP

PPE family protein

358 Cell Host & Microbe 2, 352–364, November 2007 ª2007 Elsevier Inc.

Cell Host & Microbe Macrophage Cues and Mtb Gene Expression

Table 1. Continued Gene Name

Ratio

p Value

Rv3613c

1.25

8.60E-02

Rv3614c

1.39

8.80E-02

In Vitro pH Induced

Description CHP

+

a a

CHP

Rv3615c

1.40

2.50E-02

+

Rv3616c

1.53

4.10E-02

+

CHP

CHP

Rv3745c

1.54

3.70E-02

ND

CHP

Rv3746c

1.74

2.00E-02

+

putative PE family protein, but without the glycine-rich C-terminal part

Rv3824c (papA1)

1.40

2.60E-02

+

putative polyketide synthase-associated protein

Rv3825c (pks2)

2.00

1.40E-02

+

putative polyketide synthase, involved in lipid metabolism

Includes concanamycin A-sensitive phagosome-induced genes with relative expression ratios (untreated/CcA-treated) of 1.5-fold (p < 0.1) plus adjacent CcA-sensitive genes below these cutoffs (underscored). +, genes with induction ratio >1.5 (p < 0.05) at pH 6.5 and/or 5.5; +a, genes induced by in vitro acid stress but below 1.5-fold (p < 0.05) cutoff; , genes not induced by in vitro acid stress; ND, no data passing quality filters from in vitro acid stress arrays; see Figure 3 legend for additional details. In addition, Tables S1– S24 containing microarray data and analysis are included in the Supplemental Data online. CHP, conserved hypothetical protein.

delayed temporal expression profiles (data not shown and Tables S5 and S6). A small core of 24 genes was upregulated similarly in both species by 2 hr (Figure 5A), including whiB7, MT3290.2, gmhA, nark1, Rv1806, and members of the DosR regulon (Table S7). However, even within 2 hr of MØ entry there was a striking level of divergence in the global response of Mtb and BCG (Figure 5A) (Tables S8– S11). Figure 5B shows a subset of genes with speciesspecific expression patterns at 2 hr postinfection. A putative operon including Rv3019c-Rv3022c and Rv3022A (MT3106.1) was induced in BCG but not Mtb within 2 hr of phagocytosis. BCG-specific activation of this locus encoding ESAT6-like secreted virulence factors (Brodin et al., 2004) highlights the need for a better understanding of the regulation and role of multiple, seemingly redundant ESAT6-like ESX loci. Genes involved in cell wall synthesis (pks2, papA1, pks3, papA3), short-chain fatty acid metabolism (lipF) and regulation (whiB6) were upregulated significantly more in Mtb than BCG. Some genes (i.e., MT2466-MT2467) exhibited phagosomal induction only in Mtb, whereas they were constitutively expressed at high levels by both intracellular and control BCG, evidence that BCG lacks some of the regulatory pathways active in Mtb. Although Mtb and BCG inhabit a similar niche, the attenuation of BCG indicates a very different host interaction and ultimate outcome of infection. After 24 hr of growth inside MØ, both Mtb and BCG had upregulated a common set of stress-induced genes (Tables S12 and S18) including whiB7, the DosR regulon, and heat-shock response (hsp, dnaK, grpE, groEL2). However, it appears the species-specific patterns of gene expression at early time points (0–2 hr) became even more disparate as the infection progressed (Tables S12–S18). As illustrated in Figure 5A, 65 (Mtb) and 132 (BCG) genes were upregulated (>1.5-fold up, p < 0.05) by only one species within the vacuolar environment at 24 hr. One would expect

the virulence of an intracellular pathogen to correlate with the ability to regulate virulence factors, modulate the host cell, acquire nutrients, and defend against

Figure 4. Transcriptional Regulators Modulated by In Vitro Acid Stress Gene tree of putative transcriptional regulators and signaling Ser/Thr protein kinases induced (24 genes) or repressed (6 genes) significantly (>1.5-fold, p < 0.05) at pH 6.5 and/or pH 5.5 relative to pH 7.0 controls.

Cell Host & Microbe 2, 352–364, November 2007 ª2007 Elsevier Inc. 359

Cell Host & Microbe Macrophage Cues and Mtb Gene Expression

Figure 5. Species-Specific Programs of Intraphagosomal Gene Expression (A) Venn diagram comparing Mtb and M. bovis BCG genes significantly upregulated (>1.5fold, p < 0.05) 2 hr and 24 hr after MØ infection. (B) Gene tree of species-specific gene induction 2 hr postinfection. Included are genes that are induced >1.5-fold (p < 0.05) in at least one species with >1.5-fold different expression levels (p < 0.05) between Mtb and BCG. Four biological replicates of BCG 2 hr and 24 hr infections were conducted. (C) Gene tree comparing species-specific induction levels of select stress-associated genes.

antimicrobial effectors. Consistent with this idea, we noted the enhanced induction in Mtb of genes known to be required for persistent infections (icl-fadB2-umaA1) at 24 hr (Tables S14 and S16). Likewise, our data suggest that BCG either is more susceptible to similar stresses or inhabits a more hostile environment (Figure 5C). BCG exhibited a more robust transcriptional response based on the number of genes induced (Figure 5A) and induction levels (Tables S15–S17). The BCG-specific upregulation of two key stress response regulators, sigB and mprA, provides a readout of increased stress-related cues (He et al., 2006; Manganelli et al., 1999). Elevated expression of glbN, encoding a dimeric hemoglobin that protects Mtb from RNI (Couture et al., 1999), and the putative furAkatG-Rv1907c catalase-peroxidase operon suggest BCGspecific differences in exposure or response to oxidative stress. Mtb encodes 40 toxin-antitoxin loci containing MazF homologs that help cells cope with unfavorable growth conditions (Gerdes et al., 2005). Two putative mazF-like mRNA interferases, Rv0659c and Rv1989c, are upregulated significantly higher in BCG than Mtb. In E. coli, induction of mazF-like mRNA interferases by nutritional stress inhibits protein synthesis, yielding cells in a metabolically active, nonreplicative state (Suzuki et al., 2005). Interestingly, Zhu et al. recently reported that expression of Rv0659 (mazF-mt4) and Rv1989c (mazF-mt3) in E. coli caused growth inhibition and proposed that Mtb mRNA interferases mediate the onset of the dormancy phenotype under stressful conditions (Zhu et al., 2006).

Based on conserved features of cell biology (arrest of phagosome maturation and acidification), BCG is often used as a model to study Mtb-MØ interactions. However, it should be considered that their unique responses to intramacrophage growth may reflect unappreciated differences in innate defense mechanisms (i.e., cell wall), gene regulation, or the actual phagosomal environment encountered by Mtb and BCG. As discussed above, the intracellular transcription profiles suggest that BCG experiences elevated stress levels within the phagosome. Virulence factors and regulatory mechanisms lost during the attenuation of BCG might impair its ability to manipulate phagosome maturation for optimal access to nutrient sources and/or cope with antimicrobial effectors. Concluding Remarks Our examination of Mtb-host interactions identified genes responsive to intracellular cues and provided insight into conditions within the MØ phagosome. The novelty stems from key aspects of the design of this microarray study. First, the application of RNA amplification was a technical advance crucial for undertaking such an intensive analysis of intracellular gene expression. Applying this technique to a temporal dissection of Mtb phagosome adaptation emphasized the importance of understanding the timing and dynamics of gene regulation during host interactions. Initially, we chose to profile Mtb responses immediately after phagocytosis by resting MØ to mirror the kinetics of phagosome remodeling (Yates et al., 2005). Over time, this ‘‘knee-jerk’’ reaction becomes more complex

360 Cell Host & Microbe 2, 352–364, November 2007 ª2007 Elsevier Inc.

Cell Host & Microbe Macrophage Cues and Mtb Gene Expression

as various regulatory networks required for sustaining a long-term infection are engaged. Even the 24 hr profiles likely represent a transcriptome in flux, given that an initial period of growth lag or killing, in the case of Mtb CDC1551 (Pethe et al., 2004), occurs before intracellular replication begins. Schnappinger et al. reported that changes in Mtb gene expression within MØ appeared to plateau between 24 and 48 hr (Schnappinger et al., 2003). Our unpublished observations suggest that Mtb continues to adapt to its intracellular niche long beyond these time points, and studies are in progress to monitor Mtb gene expression during long-term, up to 14 day, MØ infections. An underappreciated facet of Mtb cell biology that one must factor into interpretations of the intracellular transcriptome is the heterogeneity of intracellular bacteria. Although the majority of Mtb reside in nonfused vacuoles, as much as 20%– 30% of Mtb traffic deeper into the endosomal/lysosomal continuum and survive, perhaps even grow in more hostile environments (Armstrong and Hart, 1975; McDonough et al., 1993). A complete description of the Mtb-MØ interface will require studies that define the metabolic and transcriptional adaptation of bacilli within distinct intracellular compartments. We have also begun to systematically analyze the contribution of specific signals to Mtb gene regulation in the context of MØ infections. First, our data demonstrated that engagement of Mtb by MØ receptors without phagocytosis does not alter Mtb gene expression. Thus, it appears that the first signal of host cell invasion perceived by Mtb occurs within the forming phagosome. By blocking phagosomal acidification with CcA and validating pHinduced genes by in vitro acid stress, we revealed a direct, causative link between a vacuole stimulus (pH) and a transcriptional response. Because the complex effects of pH can intersect with multiple environmental factors, it is perhaps surprising that we observed such a substantial overlap between genes regulated by pH in rich medium in vitro and phagosome-induced genes affected by CcA in vivo. This observation hints that, early in Mtb-phagosome interactions, pH may play a major role as an environmental cue. Continuing this subtractive approach using pharmacological inhibitors and MØ from appropriate knockout mice should shed light on Mtb responses to additional stimuli in vivo. Finally, the disparity between the transcriptional responses of Mtb and BCG are both intriguing and disturbing. Many of the genes upregulated in Mtb in MØ are actually constitutively expressed to high levels in BCG, implying that appropriate regulation and timing of gene expression may contribute to virulence. Thus, speciesspecific differences in transcriptional regulation may cause Mtb and BCG to respond differently to similar stimuli in vivo. The global transcriptional response of BCG to MØ phagocytosis suggests that BCG is less capable of establishing a protected niche within the host cell or is more sensitive to antimicrobial effectors. This is supported by the increased expression of stress-related genes, which correlates well with the reduced survival and virulence of the attenuated BCG strain. The data

also raise some concerns with respect to the use of BCG as an appropriate surrogate for Mtb in MØ-interaction experiments. Microarray analysis is an extraordinarily powerful tool, but it functions best with small questions and carefully defined and controlled conditions. We feel that the incremental approach to the infection of the MØ described in this study affords the best opportunity to identify functionally significant themes and link these transcriptional themes to specific environmental stimuli. EXPERIMENTAL PROCEDURES Bacterial Strains and Cells The clinical isolate strain CDC1551 of MTb and M. bovis BCG Pasteur (BCG) were used in this study. Details of culture conditions and murine bone marrow macrophage isolation and culture are described fully in the Supplemental Data. In Vitro Acid Stress of Mtb The transcriptional responses of Mtb CDC1551 to low pH were examined in buffered 7H9-OADC medium. The pH 5.5 and 6.5 adjusted media were buffered with 100 mM 2-(4-morpholino)-ethane sulfonic acid (MES), and the pH 7.0 adjusted medium was buffered with 100 mM 3-(N-morpholino) propanesulfonic acid (MOPS). Early log phase Mtb (OD0.25) were exposed to pH 7.0, 6.5, and 5.5 for 2 hr at 37 C before RNA isolation, amplification, and analysis by microarray as detailed below (see Supplemental Experimental Procedures). Microarray Fabrication A detailed description of the microarray platform is included in the Supplemental Data. Macrophage Infections and RNA Isolation MØ were infected (MOI 20:1) with Mtb or BCG grown to mid-log phase (OD = 0.6–0.8) in standing 75 cm2 vented tissue culture flasks. Infections were initiated by centrifugation (1000 3 g, 10 min) of bacteria onto confluent MØ monolayers followed by incubation at 37 C for various times as indicated. Intracellular bacteria were stabilized and isolated using a guanidine thiocyanate-based lysis buffer as previously described (Butcher et al., 1998; Mangan et al., 2002). Pelleted mycobacteria were lysed using 5 mg/ml lysozyme, hot Trizol, and physical disruption with silica beads in a BeadBeater. Total RNA was isolated by chloroform extraction followed by direct application to QIAGEN RNeasy column purification. Where indicated, MØ were pretreated with 5 mM cytochalasin D or 100 nM concanamycin A for 20 min prior to infection to inhibit phagocytosis or phagosome acidification, respectively, and maintained throughout the infection period. Control mycobacteria not exposed to MØ were treated with identical inhibitor concentrations. Refer to the Supplemental Data for further details of infection and RNA isolation protocols. Linear Amplification of Mycobacterial RNA Using the MessageAmp-II Bacteria RNA Amplification system (Ambion), we amplified 100 ng total RNA to generate sufficient RNA from intracellular mycobacteria for microarray hybridization. Briefly, total bacterial RNA was first polyadenylated by E. coli poly(A) polymerase. Using an oligo(dT) primer that incorporated a T7 promoter, poly(A) tailed RNA was reverse transcribed to yield double-stranded cDNA. The resulting cDNA served as the template for in vitro transcription (IVT) by T7 RNA polymerase to generate multiple antisense RNA (aRNA) copies of each transcript. Amino-allyl UTP was incorporated into aRNA during the IVT reaction to permit fluorophore post-labeling. Refer to the Supplemental Data for complete RNA amplification protocol details.

Cell Host & Microbe 2, 352–364, November 2007 ª2007 Elsevier Inc. 361

Cell Host & Microbe Macrophage Cues and Mtb Gene Expression

Target Preparation and Microarray Hybridization In assays of binding to cytochalasin D-treated MØ, 5 mg of unamplified RNA was used to generate labeled cDNA targets for microarray hybridization. CyDye fluors were conjugated to aa-cDNA using CyDye Post-Labeling Reactive Dye reagents (Amersham). For all other arrays, 10 mg of amino-allyl modified aRNA were labeled with Alexa Fluor 555 and Alexa Fluor 647 (Invitrogen) before overnight hybridization (16 hr) at 45 C. Complete description of target preparation, buffers, hybridization, and wash conditions can be found in the Supplemental Data. Microarray Data Analysis Microarrays were scanned with a GenePix 4000B instrument (Axon Instruments, Inc.) with preliminary image analysis conducted using Imagene software (version 6.0, Biodiscovery). Subsequent normalization, statistical analysis, and visualization of array data were performed with Genespring 7.3 (Agilent). Time-dependent changes in gene expression were analyzed using the EDGE methodology developed by Storey et al. (2005). All figures represent averaged data from duplicate biological replicates unless otherwise specified. All transcriptional profile files have been submitted to the GEO database at NCBI (accession number GSE8827). Supplemental Data The Supplemental Data include Supplemental Experimental Procedures, two supplemental figures, and 24 supplemental tables and can be found with this article online at http://www.cellhostandmicrobe. com/cgi/content/full/2/5/352/DC1/. ACKNOWLEDGMENTS The authors acknowledge the support of the National Institutes of Health through the award AI067027 to D.G.R. We thank Paul Debbie (Center For Gene Expression Profiling [CGEP], Boyce Thompson Institute, Ithaca, NY) for array printing and Dr. David Lin (Cornell Big Red Spots Core Facility) for array scanning. Received: May 18, 2007 Revised: August 17, 2007 Accepted: September 18, 2007 Published: November 14, 2007 REFERENCES Adams, L.B., Dinauer, M.C., Morgenstern, D.E., and Krahenbuhl, J.L. (1997). Comparison of the roles of reactive oxygen and nitrogen intermediates in the host response to Mycobacterium tuberculosis using transgenic mice. Tuber. Lung Dis. 78, 237–246. Akaki, T., Sato, K., Shimizu, T., Sano, C., Kajitani, H., Dekio, S., and Tomioka, H. (1997). Effector molecules in expression of the antimicrobial activity of macrophages against Mycobacterium avium complex: Roles of reactive nitrogen intermediates, reactive oxygen intermediates, and free fatty acids. J. Leukoc. Biol. 62, 795–804. Armstrong, J.A., and Hart, P.D. (1975). Phagosome-lysosome interactions in cultured macrophages infected with virulent tubercle bacilli. J. Exp. Med. 142, 1–16. Banaiee, N., Jacobs, W.R., Jr., and Ernst, J.D. (2006). Regulation of Mycobacterium tuberculosis whiB3 in the mouse lung and macrophages. Infect. Immun. 74, 6449–6457. Behr, M.A., Wilson, M.A., Gill, W.P., Salamon, H., Schoolnik, G.K., Rane, S., and Small, P.M. (1999). Comparative genomics of BCG vaccines by whole-genome DNA microarray. Science 284, 1520–1523. Boon, C., and Dick, T. (2002). Mycobacterium bovis BCG response regulator essential for hypoxic dormancy. J. Bacteriol. 184, 6760– 6767. Brodin, P., Rosenkrands, I., Andersen, P., Cole, S.T., and Brosch, R. (2004). ESAT-6 proteins: Protective antigens and virulence factors? Trends Microbiol. 12, 500–508.

Butcher, P.D., Mangan, J.A., and Monahan, I.M. (1998). Intracellular gene expression. Analysis of RNA from mycobacteria in macrophages using RT-PCR. Methods Mol. Biol. 101, 285–306. Camacho, L.R., Ensergueix, D., Perez, E., Gicquel, B., and Guilhot, C. (1999). Identification of a virulence gene cluster of Mycobacterium tuberculosis by signature-tagged transposon mutagenesis. Mol. Microbiol. 34, 257–267. Chang, J.C., Harik, N.S., Liao, R.P., and Sherman, D.R. (2007). Identification of mycobacterial genes that alter growth and pathology in macrophages and in mice. J. Infect. Dis. 196, 788–795. Couture, M., Yeh, S.R., Wittenberg, B.A., Wittenberg, J.B., Ouellet, Y., Rousseau, D.L., and Guertin, M. (1999). A cooperative oxygen-binding hemoglobin from Mycobacterium tuberculosis. Proc. Natl. Acad. Sci. USA 96, 11223–11228. Ewann, F., Jackson, M., Pethe, K., Cooper, A., Mielcarek, N., Ensergueix, D., Gicquel, B., Locht, C., and Supply, P. (2002). Transient requirement of the PrrA-PrrB two-component system for early intracellular multiplication of Mycobacterium tuberculosis. Infect. Immun. 70, 2256–2263. Fisher, M.A., Plikaytis, B.B., and Shinnick, T.M. (2002). Microarray analysis of the Mycobacterium tuberculosis transcriptional response to the acidic conditions found in phagosomes. J. Bacteriol. 184, 4025–4032. Geiman, D.E., Raghunand, T.R., Agarwal, N., and Bishai, W.R. (2006). Differential gene expression in response to exposure to antimycobacterial agents and other stress conditions among seven Mycobacterium tuberculosis whiB-like genes. Antimicrob. Agents Chemother. 50, 2836–2841. Gerdes, K., Christensen, S.K., and Lobner-Olesen, A. (2005). Prokaryotic toxin-antitoxin stress response loci. Nat. Rev. Microbiol. 3, 371–382. Gonzalo Asensio, J., Maia, C., Ferrer, N.L., Barilone, N., Laval, F., Soto, C.Y., Winter, N., Daffe, M., Gicquel, B., Martin, C., and Jackson, M. (2006). The virulence-associated two-component PhoP-PhoR system controls the biosynthesis of polyketide-derived lipids in Mycobacterium tuberculosis. J. Biol. Chem. 281, 1313–1316. Grifantini, R., Bartolini, E., Muzzi, A., Draghi, M., Frigimelica, E., Berger, J., Randazzo, F., and Grandi, G. (2002). Gene expression profile in Neisseria meningitidis and Neisseria lactamica upon host-cell contact: From basic research to vaccine development. Ann. N Y Acad. Sci. 975, 202–216. Groisman, E.A., and Mouslim, C. (2006). Sensing by bacterial regulatory systems in host and non-host environments. Nat. Rev. Microbiol. 4, 705–709. Gupta, S., Sinha, A., and Sarkar, D. (2006). Transcriptional autoregulation by Mycobacterium tuberculosis PhoP involves recognition of novel direct repeat sequences in the regulatory region of the promoter. FEBS Lett. 580, 5328–5338. He, H., Hovey, R., Kane, J., Singh, V., and Zahrt, T.C. (2006). MprAB is a stress-responsive two-component system that directly regulates expression of sigma factors SigB and SigE in Mycobacterium tuberculosis. J. Bacteriol. 188, 2134–2143. Kendall, S.L., Movahedzadeh, F., Rison, S.C., Wernisch, L., Parish, T., Duncan, K., Betts, J.C., and Stoker, N.G. (2004a). The Mycobacterium tuberculosis dosRS two-component system is induced by multiple stresses. Tuberculosis (Edinb.) 84, 247–255. Kendall, S.L., Rison, S.C., Movahedzadeh, F., Frita, R., and Stoker, N.G. (2004b). What do microarrays really tell us about M. tuberculosis? Trends Microbiol. 12, 537–544. Koul, A., Herget, T., Klebl, B., and Ullrich, A. (2004). Interplay between mycobacteria and host signalling pathways. Nat. Rev. Microbiol. 2, 189–202. Kumar, A., Toledo, J.C., Patel, R.P., Lancaster, J.R., Jr., and Steyn, A.J. (2007). Mycobacterium tuberculosis DosS is a redox sensor and

362 Cell Host & Microbe 2, 352–364, November 2007 ª2007 Elsevier Inc.

Cell Host & Microbe Macrophage Cues and Mtb Gene Expression

DosT is a hypoxia sensor. Proc. Natl. Acad. Sci. USA 104, 11568– 11573. Malhotra, V., Sharma, D., Ramanathan, V.D., Shakila, H., Saini, D.K., Chakravorty, S., Das, T.K., Li, Q., Silver, R.F., Narayanan, P.R., and Tyagi, J.S. (2004). Disruption of response regulator gene, devR, leads to attenuation in virulence of Mycobacterium tuberculosis. FEMS Microbiol. Lett. 231, 237–245. Mangan, J.A., Monahan, I.M., and Butcher, P.D. (2002). Gene Expression during host-pathogen interactions: Approaches to bacterial mRNA extraction and labeling for microarray analysis. In Methods in Microbiology, B.W. Wren and N. Dorrell, eds. (London: Academic Press), pp. 137–151. Manganelli, R., Dubnau, E., Tyagi, S., Kramer, F.R., and Smith, I. (1999). Differential expression of 10 sigma factor genes in Mycobacterium tuberculosis. Mol. Microbiol. 31, 715–724. McDonough, K.A., Kress, Y., and Bloom, B.R. (1993). The interaction of Mycobacterium tuberculosis with macrophages: A study of phagolysosome fusion. Infect. Agents Dis. 2, 232–235. McKinney, J.D., Honer zu Bentrup, K., Munoz-Elias, E.J., Miczak, A., Chen, B., Chan, W.T., Swenson, D., Sacchettini, J.C., Jacobs, W.R., Jr., and Russell, D.G. (2000). Persistence of Mycobacterium tuberculosis in macrophages and mice requires the glyoxylate shunt enzyme isocitrate lyase. Nature 406, 735–738. Morris, R.P., Nguyen, L., Gatfield, J., Visconti, K., Nguyen, K., Schnappinger, D., Ehrt, S., Liu, Y., Heifets, L., Pieters, J., et al. (2005). Ancestral antibiotic resistance in Mycobacterium tuberculosis. Proc. Natl. Acad. Sci. USA 102, 12200–12205. Mostowy, S., Cleto, C., Sherman, D.R., and Behr, M.A. (2004). The Mycobacterium tuberculosis complex transcriptome of attenuation. Tuberculosis (Edinb.) 84, 197–204. Parish, T., Smith, D.A., Kendall, S., Casali, N., Bancroft, G.J., and Stoker, N.G. (2003a). Deletion of two-component regulatory systems increases the virulence of Mycobacterium tuberculosis. Infect. Immun. 71, 1134–1140. Parish, T., Smith, D.A., Roberts, G., Betts, J., and Stoker, N.G. (2003b). The senX3-regX3 two-component regulatory system of Mycobacterium tuberculosis is required for virulence. Microbiology 149, 1423– 1435. Park, H.D., Guinn, K.M., Harrell, M.I., Liao, R., Voskuil, M.I., Tompa, M., Schoolnik, G.K., and Sherman, D.R. (2003). Rv3133c/dosR is a transcription factor that mediates the hypoxic response of Mycobacterium tuberculosis. Mol. Microbiol. 48, 833–843. Perez, E., Samper, S., Bordas, Y., Guilhot, C., Gicquel, B., and Martin, C. (2001). An essential role for phoP in Mycobacterium tuberculosis virulence. Mol. Microbiol. 41, 179–187. Pethe, K., Swenson, D.L., Alonso, S., Anderson, J., Wang, C., and Russell, D.G. (2004). Isolation of Mycobacterium tuberculosis mutants defective in the arrest of phagosome maturation. Proc. Natl. Acad. Sci. USA 101, 13642–13647. Pettersson, J., Nordfelth, R., Dubinina, E., Bergman, T., Gustafsson, M., Magnusson, K.E., and Wolf-Watz, H. (1996). Modulation of virulence factor expression by pathogen target cell contact. Science 273, 1231–1233. Puskas, L.G., Zvara, A., Hackler, L., Jr., and Van Hummelen, P. (2002). RNA amplification results in reproducible microarray data with slight ratio bias. Biotechniques 32, 1330–1334, 1336, 1338, 1340.

Rengarajan, J., Bloom, B.R., and Rubin, E.J. (2005). Genome-wide requirements for Mycobacterium tuberculosis adaptation and survival in macrophages. Proc. Natl. Acad. Sci. USA 102, 8327–8332. Rodrigue, S., Provvedi, R., Jacques, P.E., Gaudreau, L., and Manganelli, R. (2006). The sigma factors of Mycobacterium tuberculosis. FEMS Microbiol. Rev. 30, 926–941. Russell, D.G. (2001). Mycobacterium tuberculosis: Here today, and here tomorrow. Nat. Rev. Mol. Cell Biol. 2, 569–577. Sassetti, C.M., and Rubin, E.J. (2003). Genetic requirements for mycobacterial survival during infection. Proc. Natl. Acad. Sci. USA 100, 12989–12994. Saviola, B., Woolwine, S.C., and Bishai, W.R. (2003). Isolation of acidinducible genes of Mycobacterium tuberculosis with the use of recombinase-based in vivo expression technology. Infect. Immun. 71, 1379– 1388. Schaible, U.E., Sturgill-Koszycki, S., Schlesinger, P.H., and Russell, D.G. (1998). Cytokine activation leads to acidification and increases maturation of Mycobacterium avium-containing phagosomes in murine macrophages. J. Immunol. 160, 1290–1296. Schnappinger, D., Ehrt, S., Voskuil, M.I., Liu, Y., Mangan, J.A., Monahan, I.M., Dolganov, G., Efron, B., Butcher, P.D., Nathan, C., and Schoolnik, G.K. (2003). Transcriptional adaptation of Mycobacterium tuberculosis within macrophages: Insights into the phagosomal environment. J. Exp. Med. 198, 693–704. Soliveri, J.A., Gomez, J., Bishai, W.R., and Chater, K.F. (2000). Multiple paralogous genes related to the Streptomyces coelicolor developmental regulatory gene whiB are present in Streptomyces and other actinomycetes. Microbiology 146, 333–343. Stewart, G.R., Patel, J., Robertson, B.D., Rae, A., and Young, D.B. (2005). Mycobacterial mutants with defective control of phagosomal acidification. PLoS Pathog 1, 269–278. 10.1371/journal.ppat.0010033. Steyn, A.J., Collins, D.M., Hondalus, M.K., Jacobs, W.R., Jr., Kawakami, R.P., and Bloom, B.R. (2002). Mycobacterium tuberculosis WhiB3 interacts with RpoV to affect host survival but is dispensable for in vivo growth. Proc. Natl. Acad. Sci. USA 99, 3147–3152. Storey, J.D., Xiao, W., Leek, J.T., Tompkins, R.G., and Davis, R.W. (2005). Significance analysis of time course microarray experiments. Proc. Natl. Acad. Sci. USA 102, 12837–12842. Sturgill-Koszycki, S., Schlesinger, P.H., Chakraborty, P., Haddix, P.L., Collins, H.L., Fok, A.K., Allen, R.D., Gluck, S.L., Heuser, J., and Russell, D.G. (1994). Lack of acidification in Mycobacterium phagosomes produced by exclusion of the vesicular proton-ATPase. Science 263, 678–681. Suzuki, M., Zhang, J., Liu, M., Woychik, N.A., and Inouye, M. (2005). Single protein production in living cells facilitated by an mRNA interferase. Mol. Cell 18, 253–261. Talaat, A.M., Lyons, R., Howard, S.T., and Johnston, S.A. (2004). The temporal expression profile of Mycobacterium tuberculosis infection in mice. Proc. Natl. Acad. Sci. USA 101, 4602–4607. Voskuil, M.I., Schnappinger, D., Visconti, K.C., Harrell, M.I., Dolganov, G.M., Sherman, D.R., and Schoolnik, G.K. (2003). Inhibition of respiration by nitric oxide induces a Mycobacterium tuberculosis dormancy program. J. Exp. Med. 198, 705–713.

Pym, A.S., Brodin, P., Brosch, R., Huerre, M., and Cole, S.T. (2002). Loss of RD1 contributed to the attenuation of the live tuberculosis vaccines Mycobacterium bovis BCG and Mycobacterium microti. Mol. Microbiol. 46, 709–717.

Wagner, D., Maser, J., Lai, B., Cai, Z., Barry, C.E., III, Honer Zu Bentrup, K., Russell, D.G., and Bermudez, L.E. (2005). Elemental analysis of Mycobacterium avium-, Mycobacterium tuberculosis-, and Mycobacterium smegmatis-containing phagosomes indicates pathogeninduced microenvironments within the host cell’s endosomal system. J. Immunol. 174, 1491–1500.

Rachman, H., Strong, M., Ulrichs, T., Grode, L., Schuchhardt, J., Mollenkopf, H., Kosmiadi, G.A., Eisenberg, D., and Kaufmann, S.H. (2006). Unique transcriptome signature of Mycobacterium tuberculosis in pulmonary tuberculosis. Infect. Immun. 74, 1233–1242.

Walters, S.B., Dubnau, E., Kolesnikova, I., Laval, F., Daffe, M., and Smith, I. (2006). The Mycobacterium tuberculosis PhoPR two-component system regulates genes essential for virulence and complex lipid biosynthesis. Mol. Microbiol. 60, 312–330.

Cell Host & Microbe 2, 352–364, November 2007 ª2007 Elsevier Inc. 363

Cell Host & Microbe Macrophage Cues and Mtb Gene Expression

Yates, R.M., Hermetter, A., and Russell, D.G. (2005). The kinetics of phagosome maturation as a function of phagosome/lysosome fusion and acquisition of hydrolytic activity. Traffic 6, 413–420. Zahrt, T.C., and Deretic, V. (2001). Mycobacterium tuberculosis signal transduction system required for persistent infections. Proc. Natl. Acad. Sci. USA 98, 12706–12711. Zhang, M., Wang, J.D., Li, Z.F., Xie, J., Yang, Y.P., Zhong, Y., and Wang, H.H. (2005). Expression and characterization of the carboxyl esterase Rv3487c from Mycobacterium tuberculosis. Protein Expr. Purif. 42, 59–66.

Zhu, L., Zhang, Y., Teh, J.S., Zhang, J., Connell, N., Rubin, H., and Inouye, M. (2006). Characterization of mRNA interferases from Mycobacterium tuberculosis. J. Biol. Chem. 281, 18638– 18643. Zierler, M.K., and Galan, J.E. (1995). Contact with cultured epithelial cells stimulates secretion of Salmonella typhimurium invasion protein InvJ. Infect. Immun. 63, 4024–4028. Accession Numbers All transcriptional profile files have been submitted to the GEO database at NCBI (accession number GSE8827).

364 Cell Host & Microbe 2, 352–364, November 2007 ª2007 Elsevier Inc.