Publications

The Mycobacterium tuberculosis (M.tb) envelope is highly mannosylated with phosphatidyl-myo-inositol mannosides (PIMs), lipomannan, and mannose-capped lipoarabinomannan (ManLAM). Little is known regarding the interaction between specific PIM types and host cell C-type lectin pattern recognition receptors. The macrophage mannose receptor (MR) and dendritic cell-specific ICAM-3-grabbing nonintegrin on dendritic cells engage ManLAM mannose caps and regulate several host responses. In this study, we analyzed the association of purified PIM families (f, separated by carbohydrate number) and individual PIM species (further separated by fatty acid number) from M.tb H(37)R(v) with human monocyte-derived macrophages (MDMs) and lectin-expressing cell lines using an established bead model. Higher-order PIMs preferentially associated with the MR as demonstrated by their reduced association with MDMs upon MR blockade and increased binding to COS-1-MR. In contrast, the lower-order PIM(2)f associated poorly with MDMs and did not bind to COS-1-MR. Triacylated PIM species were recognized by MDM lectins better than tetra-acylated species and the degree of acylation influenced higher-order PIM association with the MR. Moreover, only higher-order PIMs that bind the MR showed a significant increase in phagosome-lysosome fusion upon MR blockade. In contrast with the MR, the PIM(2)f and lipomannan were recognized by DC-SIGN comparable to higher-order PIMs and ManLAM, and the association was independent of their degree of acylation. Thus, recognition of M.tb PIMs by host cell C-type lectins is dependent on both the nature of the terminal carbohydrates and degree of acylation. Subtle structural differences among the PIMs impact host cell recognition and response and are predicted to influence the intracellular fate of M.tb.

A crucial step in infection is the initial attachment of a pathogen to host cells or tissue. Mycobacterium tuberculosis has evolved multiple strategies for establishing an infection within the host. The pulmonary microenvironment contains a complex milieu of pattern recognition molecules of the innate immune system that play a role in the primary response to inhaled pathogens. Encounters of M. tuberculosis with these recognition molecules likely influence the outcome of the bacillus-host interaction. Here we use a novel fluid shear assay to investigate the binding of M. tuberculosis to innate immune molecules that are produced by pulmonary epithelial cells and are thought to play a role in the lung innate immune response. Virulent and attenuated M. tuberculosis strains bound best to immobilized human fibronectin (FN) and surfactant protein A (SP-A) under this condition. Binding under fluid shear conditions was more consistent and significant compared to binding under static conditions. Soluble FN significantly increased the adherence of both virulent and attenuated M. tuberculosis strains to human primary small airway epithelial cells (SAEC) under fluid shear conditions. In contrast, SP-A and SP-D effects on bacterial adherence to SAEC differed between the two strains. The use of a fluid shear model to simulate physiological conditions within the lung and select for high-affinity binding interactions should prove useful for studies that investigate interactions between M. tuberculosis and host innate immune determinants.

The mycobacterial D-arabinofuran is a common constituent of both cell wall mycolyl-arabinogalactan (AG) and the associated lipoarabinomannan (LAM), and is thus accorded critical structural and immunological roles. Despite a well-recognized importance, progress in understanding its full structural characteristics beyond the nonreducing terminal motifs has hitherto been limited by available analytical tools. An endogenous arabinanase activity recently isolated from Mycobacterium smegmatis was previously shown to be capable of releasing large oligoarabinosyl units from AG. Advanced tandem mass spectrometry utilizing both low and high energy collision induced dissociation now afforded a facile way to map and directly sequence the digestion products which were dominated by distinctive Ara18 and Ara19 structural units, together with Ara7 and lesser amount of Ara11 and Ara12. Significantly, evidence was obtained for the first time which validated the linkages and branching pattern of the previously inferred Ara22 structural motif of AG, on which the preferred cleavage sites of the novel arabinanase could be localized. The established linkage-specific MS/MS fragmentation characteristics further led to identification of a galactosamine substituent on the C2 position of a portion of the internal 3,5-branched Ara residue of the AG of Mycobacterium tuberculosis, but not that of the nonpathogenic, fast growing M. smegmatis.

Mycobacterium tuberculosis (M. tb) pathogenesis involves the interaction between the mycobacterial cell envelope and host macrophage, a process mediated, in part, by binding of the mannose caps of M. tb lipoarabinomannan (ManLAM) to the macrophage mannose receptor (MR). A presumed critical step in the biosynthesis of ManLAM, and other mannose-containing glycoconjugates, is the conversion of mannose-6-phosphate to mannose-1-phosphate, by a phosphomannomutase (PMM), to produce GDP-mannose, the primary mannose-donor in mycobacteria. We have identified four M. tb H37Rv genes with similarity to known PMMs. Using in vivo complementation of PMM and phosphoglucomutase (PGM) deficient strains of Pseudomonas aeruginosa, and an in vitro enzyme assay, we have identified both PMM and PGM activity from one of these genes, Rv3257c (MtmanB). MtmanB overexpression in M. smegmatis produced increased levels of LAM, lipomannan, and phosphatidylinositol mannosides (PIMs) compared with control strains and led to a 13.3 +/- 3.9-fold greater association of mycobacteria with human macrophages, in a mannan-inhibitable fashion. This increased association was mediated by the overproduction of higher order PIMs that possess mannose cap structures. We conclude that MtmanB encodes a functional PMM involved in the biosynthesis of mannosylated lipoglycans that participate in the association of mycobacteria with macrophage phagocytic receptors.

Mycobacterium tuberculosis (M.tb) survives in macrophages in part by limiting phagosome-lysosome (P-L) fusion. M.tb mannose-capped lipoarabinomannan (ManLAM) blocks phagosome maturation. The pattern recognition mannose receptor (MR) binds to the ManLAM mannose caps and mediates phagocytosis of bacilli by human macrophages. Using quantitative electron and confocal microscopy, we report that engagement of the MR by ManLAM during the phagocytic process is a key step in limiting P-L fusion. P-L fusion of ManLAM microspheres was significantly reduced in human macrophages and an MR-expressing cell line but not in monocytes that lack the receptor. Moreover, reversal of P-L fusion inhibition occurred with MR blockade. Inhibition of P-L fusion did not occur with entry via Fcgamma receptors or dendritic cell-specific intracellular adhesion molecule 3 grabbing nonintegrin, or with phosphatidylinositol-capped lipoarabinomannan. The ManLAM mannose cap structures were necessary in limiting P-L fusion, and the intact molecule was required to maintain this phenotype. Finally, MR blockade during phagocytosis of virulent M.tb led to a reversal of P-L fusion inhibition in human macrophages (84.0 +/- 5.1% vs. 38.6 +/- 0.6%). Thus, engagement of the MR by ManLAM during the phagocytic process directs M.tb to its initial phagosomal niche, thereby enhancing survival in human macrophages.

The repertoires of CD1- and MHC-restricted T cells are complementary, permitting the immune recognition of both lipid and peptide Ags, respectively. To compare the breadth of the CD1-restricted and MHC-restricted T cell repertoires, we evaluated T cell responses against lipid and peptide Ags of mycobacteria in leprosy, comparing tuberculoid patients, who are able to restrict the pathogen, and lepromatous patients, who have disseminated infection. The striking finding was that in lepromatous leprosy, T cells did not efficiently recognize lipid Ags from the leprosy pathogen, Mycobacterium leprae, or the related species, Mycobacterium tuberculosis, yet were able to efficiently recognize peptide Ags from M. tuberculosis, but not M. leprae. To identify a mechanism for T cell unresponsiveness against mycobacterial lipid Ags in lepromatous patients, we used T cell clones to probe the species specificity of the Ags recognized. We found that the majority of M. leprae-reactive CD1-restricted T cell clones (92%) were cross-reactive for multiple mycobacterial species, whereas the majority of M. leprae-reactive MHC-restricted T cells were species specific (66%), with a limited number of T cell clones cross-reactive (34%) with M. tuberculosis. In comparison with the MHC class II-restricted T cell repertoire, the CD1-restricted T cell repertoire is limited to recognition of cross-reactive Ags, imparting a distinct role in the host response to immunologically related pathogens.

D-Arabinans, composed of D-arabinofuranose (D-Araf), dominate the structure of mycobacterial cell walls in two settings, as part of lipoarabinomannan (LAM) and arabinogalactan, each with markedly different structures and functions. Little is known of the complexity of their biosynthesis. beta-D-Arabinofuranosyl-1-monophosphoryldecaprenol is the only known sugar donor. EmbA, EmbB, and EmbC, products of the paralogous genes embA, embB, and embC, the sites of resistance to the anti-tuberculosis drug ethambutol (EMB), are the only known implicated enzymes. EmbA and -B apparently contribute to the synthesis of arabinogalactan, whereas EmbC is reserved for the synthesis of LAM. The Emb proteins show no overall similarity to any known proteins beyond Mycobacterium and related genera. However, functional motifs, equivalent to a proline-rich motif of several bacterial polysaccharide co-polymerases and a superfamily of glycosyltransferases, were found. Site-directed mutagenesis in glycosyltransferase superfamily C resulted in complete ablation of LAM synthesis. Point mutations in three amino acids of the proline motif of EmbC resulted in marked reduction of LAM-arabinan synthesis and accumulation of an unknown intermediate and of the known precursor lipomannan. Yet the pattern of the differently linked d-Araf units observed in wild type LAM-arabinan was largely retained in the proline motif mutants. The results allow for the presentation of a unique model of arabinan synthesis.

Current knowledge on the structure of lipoarabinomannan (LAM) has resulted primarily from detailed studies on a few selected laboratory strains of Mycobacterium tuberculosis, Mycobacterium bovis BCG, and Mycobacterium smegmatis. Our previous work was the first to report on the salient structural features of M. tuberculosis clinical isolates and demonstrated significant structural variations. A prime effort is to correlate a particular structural characteristic with observed differences in eliciting an immunobiological response, especially in the context of CD1-restricted presentation of LAM to T cells. T cell clones derived from the cutaneous lesions of leprosy patients have been shown to recognize specifically LAM from Mycobacterium leprae and not from M. tuberculosis Erdman or H37Rv. Herein we provide further fine structural data on LAM from M. leprae (LepLAM) and a tuberculosis clinical isolate, CSU20 (CSU20LAM), which was unexpectedly recognized by the supposedly LepLAM-specific CD1-restricted T cell clones. In comparison with the de facto laboratory LAM standard from M. tuberculosis H37Rv (RvLAM), LepLAM derived from in vivo grown M. leprae is apparently simpler in its arabinan architecture with a high degree of exposed, non-mannose-capped termini. On the other hand, CSU20, an ethambutol-resistant clinical isolate, makes a vastly heterogeneous population of LAM ranging from rather small and non-mannose-capped to full-length and fully capped variants. LepLAM and CSU20LAM contain a higher level of succinylation than RvLAM, which, in the context of truncated or less elaborated arabinan, may contribute to selective recognition by T cells. LAM from all species could be resolved into discrete forms by isoelectric focusing based apparently on their arabinan heterogeneity. In the light of our current and more recent findings, we reason that all immunobiological data should be cautiously interpreted and that the actual LAM variants that may be present in vivo during infection and pathogenesis need to be taken into consideration.

The arabinans of the mycobacterial cell wall are key structural and immunological polymers in the context of arabinogalactan (AG) and lipoarabinomannan (LAM) respectively. The three homologous membrane proteins EmbA, EmbB and EmbC are known to be involved in the synthesis of arabinan but their biochemical functions are not understood. Herein we show, that synthesis of LAM, but not AG, ceases after inactivation of embC in Mycobacterium smegmatis by insertional mutagenesis. LAM synthesis is restored upon complementation with the embC wild-type gene. Previously we have shown that the synthesis of the arabinan of AG is affected by embA or embB disruption. Thus the Emb proteins are capable of differential recognition of the galactan or mannan acceptors prior to appropriate arabinosylation. In addition, a combination of genetic and biochemical approaches have allowed us to assign some specific functions to the regions of emb gene products. Complementation of the embCmacr; mutant with a hybrid gene encoding the N-terminus of EmbC and the C-terminus of EmbB resulted in LAM with a lower molecular weight than the wild-type LAM. Structural studies involving enzyme digestion, chromatography and mass spectrometry analyses revealed that the arabinan of the 'LAM' formed in the hybrid was of AG kind rather than LAM type of arabinan.

Intracellular mycobacteria release cell wall glycolipids into the endosomal network of infected macrophages. Here, we characterize the glycolipids of Mycobacterium bovis BCG (BCG) that are released into murine bone marrow-derived macrophages (BMMØ). Intracellularly released mycobacterial lipids were harvested from BMMØ that had been infected with 14C-labelled BCG. Released BCG lipids were resolved by thin-layer chromatography, and they migrated similarly to phosphatidylinositol dimannosides (PIM2), mono- and diphosphatidylglycerol, phosphatidylethanolamine, trehalose mono- and dimycolates and the phenolic glycolipid, mycoside B. Culture-derived BCG lipids that co-migrated with the intracellularly released lipids were purified and identified by electrospray ionization mass spectrometry. When delivered on polystyrene microspheres, fluorescently tagged BCG lipids were also released into the BMMØ, in a manner similar to release from viable or heat-killed BCG bacilli. To determine whether the released lipids elicited macrophage responses, BCG lipid-coated microspheres were delivered to interferon gamma-primed macrophages (BMMØ or thioglycollate-elicited peritoneal macrophages), and reactive nitrogen intermediates as well as tumour necrosis factor-alpha and monocyte chemoattractant protein-1 production were induced. When fractionated BCG lipids were delivered on the microspheres, PIM2 species reproduced the macrophage-activating activity of total BCG lipids. These results demonstrate that intracellular mycobacteria release a heterogeneous mix of lipids, some of which elicit the production of proinflammatory cytokines from macrophages that could potentially contribute to the granulomatous response in tuberculous diseases.