PPE Surface Proteins Are Required for Heme Utilization by Mycobacterium tuberculosis

ABSTRACT Iron is essential for replication of Mycobacterium tuberculosis, but iron is efficiently sequestered in the human host during infection. Heme constitutes the largest iron reservoir in the human body and is utilized by many bacterial pathogens as an iron source. While heme acquisition is well studied in other bacterial pathogens, little is known in M. tuberculosis. To identify proteins involved in heme utilization by M. tuberculosis, a transposon mutant library was screened for resistance to the toxic heme analog gallium(III)-porphyrin (Ga-PIX). Inactivation of the ppe36, ppe62, and rv0265c genes resulted in resistance to Ga-PIX. Growth experiments using isogenic M. tuberculosis deletion mutants showed that PPE36 is essential for heme utilization by M. tuberculosis, while the functions of PPE62 and Rv0265c are partially redundant. None of the genes restored growth of the heterologous M. tuberculosis mutants, indicating that the proteins encoded by the genes have separate functions. PPE36, PPE62, and Rv0265c bind heme as shown by surface plasmon resonance spectroscopy and are associated with membranes. Both PPE36 and PPE62 proteins are cell surface accessible, while the Rv0265c protein is probably located in the periplasm. PPE36 and PPE62 are, to our knowledge, the first proline-proline-glutamate (PPE) proteins of M. tuberculosis that bind small molecules and are involved in nutrient acquisition. The absence of a virulence defect of the ppe36 deletion mutant indicates that the different iron acquisition pathways of M. tuberculosis may substitute for each other during growth and persistence in mice. The emerging model of heme utilization by M. tuberculosis as derived from this study is substantially different from those of other bacteria.

T uberculosis is the leading cause of death from a single infectious disease, resulting in 1.5 million deaths in 2014 (1). Mycobacterium tuberculosis infects humans by aerosols and proliferates in alveolar macrophages by inhibiting macrophage maturation (2). Acquisition of iron within the human host is critical for replication and virulence of pathogenic bacteria (3), because iron is one of the least accessible micronutrients due to sequestration by host proteins (4). To obtain iron, M. tuberculosis produces small molecules with high affinity for iron, the mycobactins and carboxymycobactins (5).
Secretion of siderophores by M. tuberculosis is dependent on the MmpL4/MmpS4 and MmpL5/MmpS5 membrane complexes (6,7). While the initial step of siderophore uptake across the outer membrane is unknown (8), the IrtA/IrtB protein complex is required for efficient uptake of ferric carboxymycobactins across the inner membrane (9). Iron is reductively released from carboxymycobactins by the flavin adenine dinucleotide (FAD)-binding domain of IrtA (10), and the uncomplexed carboxymycobactins are subsequently recycled by secretion through the MmpS4/MmpL4 and MmpS5/ MmpL5 transporters (7). However, it is estimated that approximately 70% of iron in the human body is tightly bound in heme (11), most often in a complex with hemoglobin (12), and cannot be solubilized by bacterial siderophores. Not surprisingly, many bacterial pathogens have developed mechanisms for iron acquisition from heme (13). In Gram-negative bacteria, host hemoproteins or heme are captured by outer membrane receptors that translocate heme into the periplasm in a TonB-dependent manner (14). Then, heme is transported across the inner membrane by specific permeases (11). Recent studies have shown that M. tuberculosis is also capable of utilizing heme as an iron source (15,16). Rv0203 was identified as a heme-binding protein of M. tuberculosis with a limited role in heme utilization by M. tuberculosis (16). Rv0203 was shown to transfer heme to the extracellular domains of the inner membrane proteins MmpL3 and MmpL11. However, MmpL3 and MmpL11 are resistance-nodulation-cell division (RND)type efflux pumps involved in export of trehalose monomycolate, lipids, or other lipid-like molecules for maintenance of the mycobacterial cell wall (17)(18)(19). Thus, the actual roles of MmpL3 and MmpL11 in heme utilization by M. tuberculosis are unclear. MhuD is an oxygen-dependent heme-degrading enzyme of M. tuberculosis involved in iron release from heme (20). Overall, the mechanisms by which M. tuberculosis imports heme into the cell are poorly understood. The main objective of this study was to identify proteins required for heme uptake in M. tuberculosis. Here, we demonstrate that PPE36, PPE62, and Rv0265c are novel heme-binding membrane proteins of M. tuberculosis that have nonredundant roles in heme utilization. We show that PPE36 and PPE62 are cell surface receptors for heme in M. tuberculosis and that PPE36 is essential for heme utilization in vitro.

RESULTS
Synthesis and characterization of gallium(III)-porphyrin. To identify unknown components required for heme uptake by M. tuberculosis, we exploited the toxicity of gallium(III) for bacteria. The structural properties and coordination chemistry of Ga(III) are similar to those of iron(III), and Ga(III) can substitute for Fe(III) as a cofactor in many proteins. However, Ga(III) cannot be reduced under physiological conditions and, therefore, cannot catalyze redox reactions (21). The efficacy of Ga(III) as an antimicrobial agent has been demonstrated in several bacterial pathogens (22), including M. tuberculosis (23). Since iron utilization from heme is achieved by removal of iron from the protoporphyrin (PIX) ring, we hypothesized that M. tuberculosis may similarly remove gallium from gallium(III)-porphyrin (Ga-PIX). The subsequent incorporation of Ga(III) into essential proteins would then be toxic for M. tuberculosis. Therefore, M. tuberculosis mutants resistant to Ga-PIX should have disruptions in essential components of the heme uptake and/or utilization pathways. To identify novel components of heme utilization by M. tuberculosis, we synthesized Ga-PIX from protoporphyrin IX (PIX) and gallium chloride as previously described (24) and obtained 30 mg of Ga-PIX. PIX, Fe(III)-PIX (hemin), and Ga(III)-PIX absorb light with a characteristic maximum at 410 nm but have distinct fluorescence emission spectra (25). While hemin does not exhibit any fluorescence, PIX and Ga(III)-PIX showed the characteristic fluorescence emission maxima at 585 nm and 610 nm, respectively (Fig. 1A). The synthesis of Ga(III)-PIX was further confirmed by distinct retention factors and fluorescence on silica gel thin-layer chromatography plates (see Fig. S1 in the supplemental material).
Ga(III)-PIX was highly toxic toward M. tuberculosis with an MIC 90 of~1.5 M in iron-free Middlebrook 7H9 medium supplemented with 2 M ferric carboxymycobactin (cMBT) as the sole iron source (Fig. 1B). Ga(III) salts are known to be toxic for M. tuberculosis (23) and might be present in traces in our Ga-PIX preparation. However, M. tuberculosis was completely resistant to GaCl 3 up to concentrations of 12 M in our experiments (Fig. 1B), probably because the presence of cMBT preloaded with Fe(III) prevented Ga(III) from entering the cell, at least initially, and thereby reduced the toxicity of GaCl 3 . M. tuberculosis was also highly resistant to PIX (Fig. 1B). These results validated the synthesis of Ga(III)-PIX and demonstrate that Ga(III)-PIX is toxic for M. tuberculosis.
Identification of M. tuberculosis genes involved in heme utilization. To identify genes involved in heme utilization by M. tuberculosis, we constructed a transposon mutant library in the avirulent M. tuberculosis strain mc 2 6206 containing approximately 70,000 mutants. The M. tuberculosis transposon library was cultured on iron-free Middlebrook 7H10 agar supplemented with 3 M Ga-PIX (~3ϫ MIC 90 ) and 2 M ferric cMBT as the non-heme iron source. After incubation of the plates at 37°C for~40 days, 10 clones were obtained; these clones were more resistant to Ga-PIX than wild-type (wt) M. tuberculosis as determined by a microplate alamarBlue assay (MABA) (Fig. S1D). The ability of these M. tuberculosis mutants to utilize heme was then examined in growth experiments. The M. tuberculosis Tn mutants grew like the parent M. tuberculosis strain in the presence of ferric cMBT as the sole iron source in liquid medium (Fig. S1C). The M. tuberculosis Tn1, Tn2, and Tn5 mutants exhibited small, but significant, defects in heme utilization compared to the parent strain (Fig. S1C). Sequencing of the chromosomal DNA of the M. tuberculosis transposon mutants Tn1 and Tn2 revealed insertions in the intergenic regions of rv0265c-rv0266c genes and pe22-ppe36 (rv2107-rv2108), respectively, suggesting that rv0265c and ppe36 expression may be downregulated. The insertion in the Tn5 mutant mapped at the start codon of ppe62 (rv3533c gene) (Fig. S1E).
To examine the functions of rv0265c, ppe62, and ppe36, isogenic deletion mutants were constructed by allelic exchange in the avirulent M. tuberculosis strain mc 2 6206 using a gfp-hyg cassette as schematically shown in Fig. S2. Due to the short intergenic region (80 bp) between the rv0264c and rv0265c genes, a promoter was inserted upstream of the rv0264c gene to prevent any polar effect from deletion of the rv0265c gene. After homologous recombination the gfp-hyg cassette was then excised by Cre recombinase, leaving a loxP site in lieu of the gene of interest. This process generated the unmarked M. tuberculosis mutants ML2411 (Δppe36::loxP), ML2412 (Δppe62::loxP), and ML2413 (Δrv0265c::loxP) (Table S3). All mutant strains were verified by PCR (Fig. S2). The role of these genes in heme utilization by M. tuberculosis was analyzed by monitoring the growth of the M. tuberculosis mutants in HdB minimal medium (see Materials and Methods) with 10 M ammonium ferric citrate or 10 M hemin as the sole iron source. The medium containing hemin was supplemented with 20 M 2,2=-dipyridyl (DIP) to prevent utilization of trace iron as described earlier (6). In separate experiments, we demonstrated that 20 M DIP in medium with no iron added  (Fig. S3). Deletion of ppe36 (ML2411) and ppe62 (ML2412) did not affect growth of M. tuberculosis in medium with ammonium ferric citrate as the sole iron source, while the rv0265c deletion mutant had a slightly reduced growth rate compared to wt M. tuberculosis ( Fig. 2A). In contrast, the Δppe36 mutant did not grow in minimal medium with hemin ( Fig. 2B), while growth of the ppe62 and rv0265c deletion mutants was strongly reduced ( Fig. 2C and D). The ability of all M. tuberculosis mutants to utilize hemin as an iron source was restored to wt levels when complemented with the respective gene ( Fig. 2B to D).
Since the molecular functions of PPE36, PPE62, and Rv0265c are unknown and the Δppe62 or Δrv0265c mutants did not reveal a complete phenotype, we expressed all genes in the respective M. tuberculosis mutants and examined whether these genes could functionally substitute for each other. While providing a copy of the missing gene fully restored growth of each of the three M. tuberculosis mutants on heme, none of the other genes recovered growth ( Fig. 2B to D). These results indicated that the PPE36, PPE62, and Rv0265c proteins have distinct functions in heme utilization by M. tuberculosis. This finding is consistent with the strong phenotypes of the single mutants with heme as the sole iron source. Increasing the hemin concentration to up to 20 M did not increase growth of these mutants. However, the viability of all strains was reduced at 40 M hemin, which was attributed to the toxicity of heme (Fig. S4). Altogether, these results demonstrate that PPE36 is essential for heme utilization by M. tuberculosis. In contrast, M. tuberculosis requires PPE62 and Rv0265c for efficient heme utilization but may have other proteins of similar functions.
PPE36, PPE62, and Rv0265c are novel heme-binding proteins of M. tuberculosis. Possible functional roles of PPE36, PPE62, and Rv0265c include heme binding, uptake, and/or degradation or accessory functions for other proteins directly involved in heme utilization. To examine whether any of these proteins binds heme, we performed surface plasmon resonance (SPR) experiments. For positive and negative controls, we chose the known heme-binding protein MhuD of M. tuberculosis (26) and the iron-binding transcriptional regulator IdeR (27), respectively. The hexahistidinetagged recombinant proteins of MhuD His6 , IdeR His6 , PPE62 His6 , and Rv0265c His6 were produced in Escherichia coli and purified to apparent homogeneity by nickel affinity chromatography (Fig. 3A). No expression was obtained in E. coli with a vector contain- ing the ppe36 gene alone. Since it has been reported that individual pe and ppe genes did not express well or were expressed in insoluble form (28), we expressed ppe36 together with its cognate pe22 gene. The PE22-PPE36 His6 complex was produced in milligram quantities and was purified by nickel affinity chromatography (Fig. 3A). Then, the control proteins MhuD His6 and IdeR His6 were immobilized on a nitrilotriacetic acid sensor chip and analyzed in SPR experiments as shown in Fig. S5A. Addition of hemin resulted in a rapid signal increase indicative of hemin binding and a slow dissociation of hemin from MhuD (Fig. S5B). In contrast, IdeR showed only a marginal increase in signal upon hemin addition and an immediate dissociation (Fig. S5B). These results indicated that SPR can be used to detect interactions of proteins with heme. PE22-PPE36 His6 , PPE62 His6 , and Rv0265c His6 showed signals similar to that of the MhuD control protein after the addition of hemin. This signal intensity was dependent on the hemin concentration for all proteins (Fig. 3). Kinetic analysis resulted in dissociation constants K d of 4 ϫ 10 Ϫ3 M, 4 ϫ 10 Ϫ4 M, and 3 ϫ 10 Ϫ4 M for PE22-PPE36 His6 , PPE62 His6 , and Rv0265c His6 , respectively (Table 1). These results show that PE22-PPE36, PPE62, and Rv0265c bind heme; however, their binding affinities are low compared to most other heme-binding proteins except for IsdB (Table 1). This might be an intrinsic property of  these particular M. tuberculosis proteins or might result from immobilization of hexahistidine-tagged proteins on a chip in the SPR experiments. Subcellular localization and surface accessibility of the novel heme-binding proteins of M. tuberculosis. Next, we determined the subcellular localization of these novel heme-binding proteins in M. tuberculosis using vectors encoding C-terminally hemagglutinin (HA)-tagged PPE36 (pML3731), PPE62 (pML3732), or Rv0265c (pML3733) proteins. Cells of the M. tuberculosis strains ML2408, ML2409, and ML2410 (Table S3) expressing the ppe36 HA , ppe62 HA , and rv0265c HA genes, respectively, were lysed. The membrane fraction was separated from the water-soluble fraction by ultracentrifugation as indicated by MctB and LpqH, which served as positive controls for membrane proteins, and RNA polymerase, which served as a positive control for cytosolic proteins (Fig. 4A). While PPE36 HA localized exclusively in the membrane fraction, PPE62 HA was detected at similar levels in both membrane and cytosolic fractions. Rv0265c HA localized primarily in the membrane fraction (Fig. 4A). To examine whether the HA tag interfered with protein functions, growth of M. tuberculosis mutants expressing the corresponding HA-tagged genes was determined in minimal medium with 10 M hemin as the iron source. Expression of the HA-tagged genes recovered growth of the mutants to wild-type levels, demonstrating that the HA-tagged proteins were fully functional in heme utilization (Fig. 4B).
To distinguish between inner and outer membrane proteins, we used flow cytometry to detect proteins on the cell surface of M. tuberculosis as shown before (29,30). The HA-tagged surface sphingomyelinase SpmT of M. tuberculosis was used as a marker protein (30). SpmT HA was detected on the cell surface of M. tuberculosis mc 2 6206 in flow cytometry experiments using a monoclonal HA antibody, while only background fluorescence was observed for M. tuberculosis containing the empty vector and mbtG HA (Fig. 5A). This result showed that the C-terminal HA tag of SpmT is surface accessible and indicated that the M. tuberculosis cells were mostly intact in this experiment, since the periplasmic MbtG (6) was not detected. M. tuberculosis cells producing PPE36 HA and PPE62 HA exhibited a similar shift in signal as SpmT HA relative to M. tuberculosis without any HA-tagged protein ( Fig. 5B and C). No increased fluorescence was observed for M. tuberculosis cells producing Rv0265c HA , indicating that the HA tag of Rv0265c is not surface accessible in these experiments (Fig. 5D). Taken together, these experiments demonstrate that PPE36 and PPE62 are heme-binding cell surface proteins of M. tuberculosis.
PPE36 is not required for virulence of M. tuberculosis in mice. To determine the role of heme utilization by M. tuberculosis in vivo, we infected C56BL/6 mice with wt H37Rv, the Δppe36 mutant which cannot utilize heme anymore (Fig. 2B), and the complemented Δppe36 mutant. All strains were grown in Middlebrook 7H9 medium and then cultured in iron-free Middlebrook 7H9 medium for five more days to deplete intracellular iron pools. The similar bacterial loads of lungs and spleens 21 and 42 days after exposure of mice to aerosols with these M. tuberculosis strains (Fig. S6) indicate that different iron acquisition pathways of M. tuberculosis may substitute for each other during growth and persistence in mice.

DISCUSSION
Identification of proteins required for heme utilization by M. tuberculosis. In this study, we identified three proteins of previously unknown functions as important for heme utilization by M. tuberculosis by selecting for mutants resistant to the toxic heme analog gallium(III)-porphyrin. Two of these proteins, PPE36 and PPE62, belong to an unusual protein family which is found only in mycobacteria. These PPE proteins share a conserved N-terminal domain of approximately 180 amino acids containing characteristic proline-proline-glutamate motifs and a highly variable C-terminal sequence (31). Growth in a medium with heme as the sole iron source was completely abrogated in the ppe36 mutant and strongly reduced in the ppe62 mutant (Fig. 2). These phenotypes are in contrast to the slight growth defect of the M. tuberculosis Rv0203 mutant in medium with heme as the sole iron source (16). Rv0203 was shown to bind heme and was proposed to be a secreted hemophore of M. tuberculosis (16). The minor role of Rv0203 in heme utilization in M. tuberculosis is likely the reason why Rv0203 was not identified in our selection assay using toxic gallium(III)-porphyrin.
PPE36 is essential for heme utilization by M. tuberculosis in vitro. The PPE36 (Rv2108) protein is essential for heme utilization by M. tuberculosis and consists of 243 amino acids comprising the PPE domain of~180 amino acids and a short C terminus of unique sequence. PPE36 belongs to the PPE sublineage III, comprising small proteins such as PPE41 (31). ppe41 and many other ppe genes are expressed only as folded proteins in a complex with their cognate PE proteins (28). Similarly, the pe22 (rv2107) gene is located upstream of the ppe36 (rv2108) gene and is required for expression of ppe36 in E. coli. On the basis of these observations, it has been proposed that these PE and PPE proteins are exported as folded heterodimers in M. tuberculosis (28,32) with their final localization either at the cell surface of M. tuberculosis and/or in the culture medium (33). Flow cytometry experiments showed that PPE36 is a surfaceaccessible protein consistent with previous observations for several other PPE proteins (34,35). PPE36 is exclusively associated with membranes in subcellular fractionation experiments in contrast to PPE62 and Rv0265c. However, neither PPE36 nor any other small PPE protein have a predicted transmembrane helix. Furthermore, the structure of the similar PPE41/PE25 complex also does not show any known hydrophobic domain (28,32). Thus, it is unclear how PPE36 is anchored to the cell surface. Our observation that PPE36 is completely associated with membranes when M. tuberculosis was grown in a shaking culture in the presence of detergents appears to exclude the possibility that PPE36 is attached to the polysaccharide capsule as proposed by Ates et al. for other PPE proteins (33). It is also unknown whether PPE36 requires PE22 for heme binding and/or uptake. PPE17 is a PPE protein that does not require its cognate PE proteins for stability and surface localization (35). Further experiments are needed to examine whether this is the case for PPE36.
PPE62 is required for efficient heme utilization by M. tuberculosis and has a different function than PPE36. PPE62 (Rv3533c) consists of 582 amino acids and belongs to the PPE-major polymorphic tandem repeat (MPTR) family (31). Heme utilization by M. tuberculosis was strongly reduced by deletion of ppe62, indicating that PPE62 is required for efficient heme utilization but not essential (Fig. 2C). Except for the conserved N-terminal PPE domain, PPE62 does not share any sequence similarity with PPE36. ppe62 expression does not complement the ⌬ppe36 mutant, demonstrating that these proteins have different functions in heme utilization. Interestingly, secondary structure prediction for PPE62 indicated that the tandem repeats may form a righthanded parallel ␤-helix consisting of three parallel ␤-sheets (see Fig. S7 in the supplemental material). This predicted structure has a striking resemblance to HxuA, which is the cell surface receptor of hemopexin in Haemophilus influenzae. The crystal structure of HxuA shows a right-handed ␤-helix with three parallel ␤-sheets that are involved in binding and release of heme from host hemopexin (36). Whether the predicted ␤-helix of PPE62 functions in binding and release of heme in a similar manner as in HxuA remains to be determined. Similar to PPE36, flow cytometry and subcellular fractionation experiments showed that PPE62 is a surface-accessible membrane protein. Although the molecular basis for membrane localization of PPE62 is unknown as for all other PPE proteins, it is interesting that 16 out of 20 PPE-MPTR proteins have been predicted to be outer membrane proteins based on secondary structure predictions of amphiphilic ␤-strands that may form a ␤-barrel with a hydrophobic surface (37,38). This includes the surface proteins PPE62 and PPE34 (34), but not PPE36 and PPE17 (35), which belong to other PPE subfamilies.
Rv0265c is required for efficient heme utilization by M. tuberculosis. The ability of the M. tuberculosis rv0265c mutant to grow in medium with heme as the sole iron source is reduced compared to the parent strain (Fig. 2D), identifying Rv0265c as the third M. tuberculosis protein involved in heme utilization. However, the growth phenotype of the M. tuberculosis rv0265c deletion mutant is weaker than those of the ppe36 and ppe62 mutants, indicating that Rv0265c has a partially redundant role in heme utilization by M. tuberculosis. The rv0265c deletion mutant has a growth defect even in medium with ferric citrate as the sole iron source ( Fig. 2A), indicating that Rv0265c may have functions in addition to its role in heme utilization. Rv0265c consists of 330 amino acids and has a signal peptide targeting the protein to the twin-arginine translocation (TAT) pathway, which is utilized for export of folded proteins (39). Interestingly, Rv0265c also contains a cysteine at the N terminus after the predicted signal peptidase cleavage site, indicating that Rv0265c is a lipoprotein (40). This is consistent with our observation that Rv0265c is primarily localized in the membrane fraction in subcellular fractionation experiments with M. tuberculosis (Fig. 4). In E. coli, lipoproteins with an aspartate residue in position ϩ2 after the cysteine are localized on the outer leaflet of the inner membrane, whereas all other lipoproteins are directed to the outer membrane (40). Rv0265c does not have an aspartate residue following the N-terminal cysteine of the mature protein, which may indicate targeting to the outer membrane. However, it is not clear how lipoproteins are sorted in mycobacteria and whether a system similar to the Lol system of E. coli exists in M. tuberculosis. The crystal structure of Rv0265c was solved (PDB identifier: 4PM4) and shows significant structural similarities to HmuT, the periplasmic component of the heme uptake pathway in Yersinia pestis (41). Taken together, a model in which Rv0265c is located in the inner leaflet of the outer membrane is consistent with the results of this study and our bioinformatic analysis (Fig. 6).
A new model for heme acquisition by M. tuberculosis. Growth of M. tuberculosis is strongly reduced using heme as an iron source versus the siderophore-mediated utilization of ferric citrate (26 and 11 days, respectively [ Fig. 2]). This result is consistent with the previous observation that growth of M. tuberculosis on hemin requires 1,000-fold-higher concentrations compared to ferric cMBT as an iron source (7). However, heme acquisition by M. tuberculosis is probably important due to the abundance of heme-bound iron compared to other iron sources in the human body (42). The findings of this study substantially expand our understanding of the mechanism by which M. tuberculosis utilizes heme as an iron source as depicted in a revised model (Fig. 6). PPE36 may capture heme on the cell surface of M. tuberculosis (Fig. 6). It is unknown whether complex formation with PE22 is required for heme binding and/or surface attachment of PPE36. PPE62 is predicted to be an outer membrane protein (37,38) and may function as a heme receptor similar to HasR of Serratia marcescens (43). Since PPE62 is not essential for heme utilization, it is likely that M. tuberculosis produces at least one additional protein which can partially substitute for PPE62. Ultimately, an atomic structure is required to provide final evidence whether PPE62 forms a membrane-spanning hydrophobic domain such as those found in other mycobacterial outer membrane proteins (44). Our observation that PPE36 and PPE62 are hemebinding surface proteins suggests that they are unlikely to be involved in heme detoxification which is mainly dependent on efflux by inner membrane transporters and/or sequestration by cytoplasmic and periplasmic proteins in other bacteria (45). Rv0265c is an exported, heme-binding protein of M. tuberculosis whose structure is PPE36 is a membrane-bound surface-accessible protein that is essential for heme utilization by M. tuberculosis. The PPE36/PE22 complex binds heme (Hm), but it is unknown whether complex formation is required for heme binding. PPE62 is a surface-accessible, heme-binding protein anchored in the outer membrane (OM). Rv0265c is a periplasmic heme-binding protein that might transfer heme to a yet unknown transporter in the inner membrane (IM). Cytoplasmic heme is degraded to mycobilin by the MhuD oxygenase, thereby releasing iron (small black circle) (20). MmpL3 and MmpL11 are IM efflux pumps which are capable of heme transfer to Rv0203 (46). Heme Utilization by M. tuberculosis similar to that of HmuT, a periplasmic heme-binding protein of Y. pestis (41). It is unclear in which membrane Rv0265c is anchored. Interestingly, none of the genes identified in this study is regulated by iron or by the major iron regulator IdeR (27), indicating that expression of these genes might be specifically induced by heme. We propose that PPE36 and PPE62 are involved in heme translocation across the outer membrane. Heme uptake across the inner membrane is mediated by an unknown transporter. Then, the heme oxygenase MhuD degrades the porphyrin ring and releases iron from heme without generating carbon monoxide (20). Rv0203 is a heme-binding protein (16) which transfers heme to water-soluble domains of the inner membrane proteins MmpL3 and MmpL11 (46) in the periplasm of M. tuberculosis (47). MmpL proteins are efflux pumps (48,49) that export hydrophobic molecules such as lipids (50)(51)(52), azoles (53), and siderophores (6) out of the cell. Thus, it is difficult to reconcile heme uptake, as proposed for MmpL3 and MmpL11 (16), with the opposite transport directionality of MmpL proteins. Hence, we propose that MmpL3 and MmpL11 might be involved in heme efflux to protect M. tuberculosis from the toxicity of excess heme as described for other bacterial pathogens (45). Redundancy of MmpL3 and MmpL11 is also consistent with the slight growth defects of the mmpL3 and mmpL11 mutants in the presence of heme (16).
Conclusions. PPE36 and PPE62 are, to our knowledge, the first PPE proteins of M. tuberculosis with functions in nutrient acquisition. Previously, PPE proteins were thought to be involved in antigenic variation and immune evasion (54,55). While our study identified three novel proteins involved in heme utilization by M. tuberculosis, their molecular functions and how these proteins interact with each other are unknown. Further, crucial components of the heme uptake system are still missing (Fig. 6). The 1,000-fold-lower efficiency of iron acquisition from heme compared to ferric siderophore uptake might be balanced by the much higher abundance of heme in the human body (11). Taken together, the emerging model indicates that M. tuberculosis has evolved a heme uptake system different from those in other bacteria (56)(57)(58)(59)(60).
Synthesis of gallium(III)-porphyrin. Gallium(III)-porphyrin (Ga-PIX) was synthesized as previously described (24). Briefly, 2.4 g of gallium chloride (GaCl 3 ), 5.3 g of protoporphyrin IX (PIX), and 5.0 g of sodium acetate were added to 400 ml of acetic acid, refluxed for 16 h, and then cooled on ice to allow crystallization. All excess acetic acid was removed by evaporation, and excess GaCl 3 and sodium acetate were dissolved in 200 ml water and removed, leaving 30 mg of a brown precipitate containing the insoluble Ga-PIX and any unreacted PIX. The dried precipitate was dissolved in 10 ml dimethyl sulfoxide (DMSO). Ga-PIX and PIX were excited with light with a wavelength of 410 nm, and the fluorescence of synthesized Ga-PIX (at 585 nm) was compared to that of PIX (at 610 nm) and quantified. Ga-PIX and PIX were also characterized by thin-layer chromatography (TLC) on silica gel plates using a mobile phase consisting of methanol and water in a 3:1 volume ratio. PIX, hemin, and Ga-PIX were visualized using UV light.
Toxicity of gallium(III)-porphyrin and precursors for M. tuberculosis. The MICs of GaCl 3 , PIX, and Ga-PIX were determined against avirulent M. tuberculosis mc 2 6206. Cultures were grown to log phase in Middlebrook 7H9 medium, washed in sterile phosphate-buffered saline (PBS) (137 mM NaCl, 2.7 mM KCl, 10 mM Na 2 HPO 4 , 1.8 mM KH 2 PO 4 ) with 0.02% Tyloxapol (PBS-Tyloxapol) and then depleted of iron for three or four generations in iron-free Middlebrook 7H9 medium. In 96-well plates, washed M. tuberculosis cells were inoculated at an optical density at 600 nm (OD 600 ) of 0.05 into iron-free Middlebrook 7H9 medium containing 2 M ferric carboxymycobactin (cMBT) as the iron source and various concentrations of GaCl 3 , PIX, or Ga-PIX. After 5 or 6 days of growth, the MIC 90 s for all compounds were determined using an alamarBlue assay as described previously (61).
Targeted gene deletion in M. tuberculosis. A schematic representation of targeted gene deletion is shown in Fig. S2 in the supplemental material. For all genes, upstream (U) and downstream (D) sequences were amplified using corresponding primer pairs UF/SpeI-UR/SwaI (F stands for forward, and R stands for reverse) and DF/PacI-DR/NsiI (Table S2), respectively, and cloned into pML2424. pML3715, pML3723, and pML3726 (Table S1) were transformed into avirulent M. tuberculosis mc 2 6206 for deletion of ppe36, ppe62, and Rv0265 genes, respectively. Transformants were selected at 37°C on Middlebrook 7H10 medium supplemented with hygromycin (Hyg) and visually validated by the presence of both green fluorescent protein (GFP) and red fluorescent protein (RFP) fluorescence. Liquid cultures of transformants were then plated on Middlebrook 7H10 medium supplemented with Hyg containing 2% sucrose at 40°C for selection of double crossovers. Putative double crossovers were visually analyzed for the presence of only GFP, and gene deletion was validated by PCR (Fig. S2). For excision of the loxP-flanked gfp 2ϩ m -hyg cassette, pML2714 expressing Cre recombinase was transformed into marked mutants, and unmarked mutants were selected on Middlebrook 7H10 medium supplemented with Kan at 37°C. Putative unmarked mutants were first visually validated through the absence of GFP fluorescence and then through PCR (primers [Table S2]) and loss of growth on hygromycin. Unmarked mutants of ppe36, ppe62, and rv0265c gene were designated ML2411, ML2412, and ML2413, respectively. All strains are described in Table S3.
Growth experiments for determining iron utilization. Strains were first grown in Middlebrook 7H9 medium, then washed in sterile PBS with 0.02% Tyloxapol, and iron depleted for 3 or 4 generations in iron-free Middlebrook 7H9 medium. Strains were then inoculated into HdB minimal medium (62) containing either 10 M hemin or 10 M ammonium ferric citrate as the sole iron source. To prevent growth from traces of iron, 20 M iron chelator 2,2-dipyridyl (DIP) was added to medium containing heme.

Construction of expression vectors and strains for mycobacterial genes.
The open reading frames (ORFs) of ppe36, ppe62, and rv0265c were amplified using corresponding primers, 016Clone/F and 016Clone/R (Table S2), and cloned into pMN016 to construct pML3716, pML3724, and pML3727, respectively. pML3716, pML3724, and pML3727 were transformed into strains ML2411, ML2412, and ML2413 to create strains ML2414, ML2415, and ML2416, respectively. All of these strains were used in subsequent growth experiments for determining heme utilization. For hemagglutinin (HA)-tagged expression of M. tuberculosis proteins, ORFs of ppe36, ppe62, and rv0265c were amplified using the corresponding primers, primers HA/F and HA/R (Table S2), and cloned into pML1391 to construct pML3731, pML3732, and pML3733, respectively. pML3731, pML3732, and pML3733 were transformed into avirulent M. tuberculosis mc 2 6206 to create strains ML2408, ML2409, and ML2410, respectively, which were then used in subcellular localization and surface detection experiments. pML1828 and pML2109 expressing MbtG HA and Rv0888 HA were transformed into avirulent M. tuberculosis mc 2 6206 to create strains ML2435 and ML2436, respectively, which were negative and positive controls for surface detection experiments.
Subcellular fractionation of M. tuberculosis. Strains ML2408, ML2409, and ML2410 were grown in Middlebrook 7H9 supplemented with ADS and Hyg to an OD 600 of 2. Cells were harvested by centrifugation, washed twice with PBS containing 1 mM phenylmethylsulfonyl (PMSF) and lysed by sonication (20 min, 12-W output power). Cell debris was removed from the lysate by centrifugation at 3,200 ϫ g for 10 min at 4°C, and the clear lysate (L) was centrifuged at 100,000 ϫ g for 1 h at 4°C. The supernatant (C1) was transferred to a separate tube, and the pellet was resuspended in the same volume of PBS as C1 and designated M1. Both C1 and M1 fractions were centrifuged at 100,000 ϫ g for 1 h at 4°C. The supernatant containing the cytosolic fraction was transferred to a new tube and labeled C2, and the membrane pellet fraction was resuspended in the same volume of PBS as that used for C2 and designated M2. HA-tagged proteins in L, M2, and C2 fractions were detected in Western blots using a horseradish peroxidasecoupled mouse antibody against HA (catalog no. H6533; Sigma). The membrane control proteins MctB and LpqH and the cytosolic control protein RNA polymerase were detected by monoclonal mouse antibodies. Blots were developed using enhanced chemiluminescence (ECL) Western blotting substrate (Pierce), and luminescence was visualized using LabWorks (UVP, Inc.) imaging system.
Surface detection of proteins in M. tuberculosis by flow cytometry. Strains ML2408 (PPE36 HA ), ML2409 (PPE62 HA ), ML2410 (Rv0265 HA ), ML2435 (MbtG HA ), and ML2408 (Rv0888 HA ) were used for surface detection experiments. Cultures were grown to mid-log phase and fixed with 4% paraformaldehyde for 30 min at room temperature. The cells were washed three times with PBS-Tyloxapol (0.02%) and incubated with monoclonal rabbit anti-HA antibody at a 1:1 dilution for 2 h for bacterial surface staining. The cells were then washed three times with PBS-Tyloxapol and then stained with fluorescein isothiocyanate (FITC)-labeled anti-rabbit antibodies at a dilution of 1:100 for 2 h. The cells were again washed three times with PBS-Tyloxapol and analyzed via flow cytometry. Surface-accessible proteins were quantified by measuring fluorescence and displayed as histograms.
Construction of protein overexpression vectors and protein purification. The ORFs of all target genes were amplified using the corresponding primers, 21aClone/F and 21aClone/R (Table S2), and cloned into pET-21aϩ. ppe36 (rv2108) was cloned with its cognate pe22 gene, rv2107, to create the expression plasmid pML3720. Finally, the ppe62, ideR, mhuD, and rv0265c genes, excluding its signal peptide, were all cloned into pET-21a to create pML3730, pML3754, pML3755, and pML3729, respectively. Transformants in E. coli BL21(DE3) were selected on LB agar plates containing Amp. Starter cultures of strains were inoculated into 500 ml of fresh LB with Amp and grown to an OD 600 of 0.3. Protein expression was induced with 1 mM isopropyl-␤-D-thiogalactopyranoside (IPTG) at 16°C for 14 h. Cells were harvested by centrifugation and lysed by sonication in ice-cold lysis buffer (20 mM Tris, 300 mM NaCl, 1 mM PMSF [pH 7.4]). Cell lysate was clarified by centrifugation, and the supernatant was loaded on to activated nickel resin and bound overnight at 4°C with shaking. Protein-loaded resin was washed three times with wash buffer (lysis buffer with 25 mM imidazole), and target protein was eluted with elution buffer (lysis buffer with 250 mM imidazole). Purified recombinant proteins were quantified by the Bradford assay. PE22-PPE36 His and PPE62 His were purified in buffers containing 0.5% (vol/vol) n-octylpolyoxyethylene (OPOE).
Surface plasmon resonance spectroscopy. Heme binding by recombinant MhuD His6 , IdeR His6 , PPE36 His6 , PPE62 His6 , and Rv0265 His6 was determined by surface plasmon resonance (SPR) spectroscopy using a Biacore T200 molecular interaction system (GE Healthcare). The SPR experimental scheme is depicted in Fig. S5A. HBS-EP (10 mM HEPES, 150 mM NaCl, 3 mM EDTA, and 0.005% polysorbate 20 [pH 7.4]) was used as the running buffer for the immobilization and kinetic studies. In a Series S Sensor Chip NTA (Biacore), recombinant protein (ligand) was immobilized onto a flow cell at a flow rate of 5 l/min for 5 min until a ligand density of~200 response units (RU) was obtained. Following ligand capture, heme (analyte) at different concentrations (diluted in running buffer) was injected into the flow cell at 10 l/min for 2 min to observe association, and then dissociation was allowed for 5 min. As a control, a reference flow cell without any bound nickel was used before ligand capture, with all the following same steps as the active flow cell. All RU values were normalized to those of the protein capture level, and the binding response was reported as the difference of RU values between the active flow cell and the control flow cell. Within the Biacore evaluation software, a 1:1 binding model was used to fit the binding response curves and dissociation constant K d was calculated for all target proteins.
Mouse infections. Female C57BL/6 mice (Jackson Laboratory) were infected using an inhalation exposure system (Glas-Col) with M. tuberculosis grown to mid-log phase in iron-depleted Middlebrook 7H9 medium to deliver approximately 100 bacilli per mouse. Lungs and spleens were harvested at the indicated time points and homogenized in PBS, serially diluted, and cultured on Middlebrook 7H10 agar to quantify CFU. All animal experiments were performed following National Institutes of Health guidelines for housing and care of laboratory animals and performed in accordance with institutional regulations. The protocol was reviewed and approved by the Institutional Animal Care and Use Committee of Weill Cornell Medical College.

ACKNOWLEDGMENTS
This work was supported by grants from the University of Alabama at Birmingham and from the Alabama Drug Discovery Alliance to M.N. We thank Olaf Kutsch for granting us access to his flow cytometer.