sign in sign up
<<
Home , Lectin

Fine Discrimination in the Recognition of Individual Species of Phosphatidyl- myo -Inositol Mannosides from Mycobacterium tuberculosis by C-Type Lectin Pattern This information is current as Recognition Receptors of September 29, 2021. Jordi B. Torrelles, Abul K. Azad and Larry S. Schlesinger J Immunol 2006; 177:1805-1816; ; doi: 10.4049/jimmunol.177.3.1805 http://www.jimmunol.org/content/177/3/1805 Downloaded from

References This article cites 50 articles, 34 of which you can access for free at:

http://www.jimmunol.org/content/177/3/1805.full#ref-list-1 http://www.jimmunol.org/

Why The JI? Submit online.

• Rapid Reviews! 30 days* from submission to initial decision

• No Triage! Every submission reviewed by practicing scientists

• Fast Publication! 4 weeks from acceptance to publication by guest on September 29, 2021

*average

Subscription Information about subscribing to The Journal of Immunology is online at: http://jimmunol.org/subscription Permissions Submit copyright permission requests at: http://www.aai.org/About/Publications/JI/copyright.html Email Alerts Receive free email-alerts when new articles cite this article. Sign up at: http://jimmunol.org/alerts

The Journal of Immunology is published twice each month by The American Association of Immunologists, Inc., 1451 Rockville Pike, Suite 650, Rockville, MD 20852 Copyright © 2006 by The American Association of Immunologists All rights reserved. Print ISSN: 0022-1767 Online ISSN: 1550-6606. The Journal of Immunology

Fine Discrimination in the Recognition of Individual Species of Phosphatidyl-myo-Inositol Mannosides from Mycobacterium tuberculosis by C-Type Lectin Pattern Recognition Receptors1

Jordi B. Torrelles, Abul K. Azad, and Larry S. Schlesinger2

The Mycobacterium tuberculosis (M.tb) envelope is highly mannosylated with phosphatidyl-myo-inositol mannosides (PIMs), lipo- mannan, and -capped lipoarabinomannan (ManLAM). Little is known regarding the interaction between specific PIM types and host cell C-type lectin pattern recognition receptors. The macrophage (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 number) and individual PIM species (further separated by fatty acid number) from M.tb H R with human monocyte-derived macrophages (MDMs) and lectin- 37 v Downloaded from 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

PIM2f 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- fusion upon MR blockade. In

contrast with the MR, the PIM2f and lipomannan were recognized by DC-SIGN comparable to higher-order PIMs and ManLAM, http://www.jimmunol.org/ 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 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. The Journal of Immu- nology, 2006, 177: 1805–1816.

ycobacterium tuberculosis (M.tb)3 is an intracellular capped lipoarabinomannan (ManLAM) (Fig. 1) (5, 6). They are pathogen that survives within the macrophage. It has thought to be both incorporated into the M.tb membrane and also developed multiple strategies to enhance its entry into exposed on its cell surface (7–9). In this regard, M.tb ManLAM

M by guest on September 29, 2021 human mononuclear phagocytes by interaction of its surface-ex- with its well-defined structure comprised of a carbohydrate core posed cell wall components with specific host pattern recognition (i.e., D-mannan and D-arabinan), a mannosyl-phosphatidyl-myo- receptors (PRRs) and through these interactions, modulates diverse inositol anchor and various terminal mannose-capping motifs (Fig. biological functions (reviewed in Refs. 1 and 2). 1), has been shown to bind to the macrophage mannose receptor The complex outermost components of the mycobacterial cell (MR) mediating phagocytosis of M.tb by human macrophages wall, comprised predominantly of lipids and carbohydrates, are the (10). ManLAMs from different M.tb strains vary in the degree to first to contact host cellular constituents, and therefore play a major which they bind to the MR pointing to a potential relationship role in facilitating host cell entry and modulating the host cell between the length and/or presentation of the mannose caps and response (2, 3). The M.tb cell wall is dominated by a group of their affinity for the MR (11). Importantly, our recent studies biosynthetically related mannosylated lipoglycoconjugates (3, 4) showed that engagement of the MR by M.tb ManLAM during the including phosphatidylinositol mannosides (PIMs) and their hy- phagocytic process directs M.tb to its initial phagosomal niche permannosylated derivatives, lipomannan (LM) and the mannose- (12). ManLAM caps also bind to dendritic cell-specific ICAM-3- grabbing nonintegrin (DC-SIGN) on dendritic cells (13, 14). Thus, terminal components of ManLAM are very important in host cell Departments of Medicine and Molecular Virology, Immunology, and Medical Ge- recognition. The role of ManLAM in MR- and DC-SIGN-depen- netics, Center for Microbial Interface Biology, and Division of Infectious Diseases, 2ϩ Ohio State University, Columbus, OH 43210 dent cytokine and Ca responses has also been reported (14–16). The PIMs are the major phospholipid components of the my- Received for publication December 9, 2005. Accepted for publication May 1, 2006. cobacterial cell wall. They are a heterogeneous mixture of families The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance (Fig. 1), which differ by the number of mannosyl residues in their ϭ with 18 U.S.C. Section 1734 solely to indicate this fact. structure (PIMxf, x number of mannoses). Each family is com- 1 This work was supported by the National Institutes of Health Grant AI-052458. posed of several species that differ in their fatty acid content 2 ϭ Address correspondence and reprint requests to Dr. Larry S. Schlesinger, Depart- (AcyPIMx, y number of fatty acids where 0, 1, or 2 denotes di-, ment of Internal Medicine, Ohio State University, 420 West 12th Avenue, 200 Med- tri-, or tetra-acylated species, respectively). As shown in Fig. 1, ical Research Facility, Columbus, OH 43210. E-mail address: Larry.Schlesinger@ osumc.edu PIM structures consist of a phosphatidylinositol anchor with a mannosyl unit attached to position C2 of the myo-inositol. Addi- 3 Abbreviations used in this paper: M.tb, Mycobacterium tuberculosis; PRR, pattern recognition receptor; PIM, phosphatidyl-myo-inositol mannoside; LM, lipomannan; tional substitutions at the C6 position of the myo-inositol by one, ManLAM, mannose-capped lipoarabinomannan; MR, mannose receptor; DC-SIGN, two, or three mannosyl units define the PIM2f, PIM3f, and PIM4f, dendritic cell-specific ICAM-3 grabbing non-integrin; P-L, phagosome-lysosome; HSA, human serum albumin; MDM, monocyte-derived macrophage; WT, wild type; respectively. The terminal mannose of the PIM4f may be further CRD, carbohydrate recognition domain; TBST, tuberculostearic acid. substituted at position C2 by one or two mannosyl units leading to

Copyright © 2006 by The American Association of Immunologists, Inc. 0022-1767/06/$02.00 1806 RECOGNITION OF PIMs BY MACROPHAGE C-TYPE LECTINS Downloaded from http://www.jimmunol.org/ by guest on September 29, 2021

FIGURE 1. Structure of the M.tb cell envelope mannose-containing . Diagram depicting structures of PI - PIM6, LM, and ManLAM from M.tb. PIMs as well as LM and ManLAM have four potential sites for fatty acid additions. Commonly, the C1 and C2 positions of the ns-glycerol of the PI anchor are substituted by a tuberculostearic and palmitic acid, respectively (PIMy). Further acylation may occur in position C6 of the mannose ␣ 3 attached to the C2 position of the myo-inositol (Ac1PIMy) and to the C3 position of the myo-inositol (Ac2PIMy). LM is characterized by an (1 6)-mannan core, which has substitutions at the C2 position of some of its mannoses. ManLAM is characterized by its terminal mono-, di-, and trimannoside caps, which interact with the MR and DC-SIGN during M.tb phagocytosis. PIM5f and PIM6f contain mono- and dimannoside cap-like structures, respectively.

the formation of PIM5forPIM6f, respectively. The latter two fam- uate the association of highly purified M.tb PIM families and spe- ilies are the highest ones present in M.tb (Fig. 1) and their nonre- cies with the human macrophage MR. In addition, by comparing ducing termini resemble the mono- and dimannosyl cap present in the PIM association pattern between the MR and DC-SIGN using ManLAM, respectively, raising the possibility that along with mammalian cell lines expressing each of these C-type lectins, we ManLAM, PIMs participate in the phagocytic process. Multiacy- establish the importance of PIM composition in host cell recogni- lated forms (di- to tetra-acylated) of PIMs exist in M.tb. Diacylated tion by these lectins that play a role in tuberculosis pathogenesis. forms have been reported to have both fatty acids located in their Finally, we determine the degree to which the MR is involved in ns-glycerol (denoted PIMx), triacylated forms have an additional phagosome-lysosome (P-L) fusion events for phagosomes contain- fatty acid at the C6 position of their 2-mannosyl unit attached to ing PIM-coated beads. These data provide evidence that higher- the myo-inositol (denoted Ac1PIMx), and tetra-acylated forms have order PIMs with mannose cap-like structures participate along the fourth fatty acid located in the C3 position of their myo-inositol with ManLAM in the phagocytic and trafficking process mediated

(denoted Ac2PIMx) (17) (Fig. 1). by the macrophage MR. Mixtures of PIMs have been found to regulate cytokine, oxidant, and T cell responses (18–20), and to stimulate early endosomal Materials and Methods fusion (21). In all of these studies, the nature of the PIMs used was , media, and Abs not clearly defined particularly at the level of acylation. There are very few studies regarding the immunologic responses of macro- M.tb H37Rv, M.tb H37Ra, and M.tb Erdman were grown on Middlebrook phages to individual families of highly purified PIMs (22–24) or to 7H11 (Difco) with 5% oleic acid, albumin, dextrose, and catalase enrich- ment for 14 days at 37°C. Mycobacterium smegmatis mc2155 was grown purified species, i.e., PIMs separated by the degree of acylation under identical conditions described above for M.tb strains but for 5 days. (19, 22). In all these studies, PIMs were used free in solution with RPMI 1640 medium with L-glutamine was purchased from Invitrogen Life murine bone marrow-derived macrophages. In this study, we eval- Technologies. RPMI 1640 medium was used alone or with 20 mM HEPES The Journal of Immunology 1807

(Difco) and 1 mg/ml human serum albumin (HSA; ZLB Bioplasma; pH reagent (specific for phospholipids) and ␣-naphthol (specific for 7.2). Anti-MR mAb (CD206, clone no. 19.2) used for immunofluorescence ). studies was purchased from BD Pharmingen. For receptor blocking, CD206 (clone no. 15-2) mAb was purchased from Serotec. Sample preparation and MALDI-TOF mass spectrometry Analyses by MALDI-TOF were conducted on a Bruker Daltonic Reflex III Purification of M.tb PIMs, LM, and ManLAM (Bruker Daltonic) mass spectrometer using DE-reflectron mode. Ionization was effected by irradiation with pulsed UV light (337 nm) from a N laser. For experiments with monocyte-derived macrophages (MDMs) and COS-1 2 PIMs and PIM species were analyzed by the instrument operating at 22.5 cells, M.tb H R cells were harvested and directly delipidated at 37°C for 37 v kV in the positive ion mode using an extraction delay time set at 200 ns. 12 h using CHCl :CH OH (2:1, v/v) followed by CHCl :CH OH:H O (10: 3 3 3 3 2 Typically, spectra from 100 to 250 laser shots were summed to obtain the 10:3, v/v/v) for an additional 12 h. Total lipid extracts were further pre- final spectrum. Saturated ␣-cyano-4-hydroxycinnamic acid (Bruker Dal- cipitated in cold acetone for 24 h at Ϫ20°C, resulting in a pellet that tonik) in 50% acetonitrile/0.1% aqueous trifluoroacetic acid (50:50) was contained a mixture of phospholipids. PIMs were separated by silica col- used for the matrix. Typically, 1 ␮l of PIM sample (mixture, families or umn chromatography using a gradient from 100% chloroform to 100% species) (10 ␮g) in a CHCl :CH OH:H O (10:10:3, v/v/v) solution and 5 methanol and eluted in the 60% methanol in chloroform fraction. PIM 3 3 2 ␮l of the matrix solution were deposited on the stainless steel target, mixed families and species were further purified by preparative TLC and purified with a micropipet, and allowed to air-dry, forming a cocrystalline sample/ components were visualized as described below for one-dimensional TLC. matrix complex. The measurements were externally calibrated at two A mixture of PIMs, PIM families (separated by their mannose content) and points with PIMs. PIM species (separated by their mannose content and the degree of acylation) were used to coat 1 or 6 ␮m in diameter beads (Polysciences) Monocyte-derived macrophages to determine their association with MDMs, COS-1-MR, and COS-1-DC-SIGN. PBMC were isolated from healthy donors (using an approved Institutional

Review Board protocol for the human subjects at the Ohio State Univer- Downloaded from ManLAM and LM from M.tb H37Rv were extracted and purified from the delipidated cells after PIM extraction as previously described (25). sity) and day 5 MDMs were cultivated as previously described (29). Mono- Briefly, the dried biomass was suspended in breaking buffer containing layers on glass coverslips in 24-well tissue culture plates contained ϳ2.0 ϫ protease inhibitor mixture (pepstatin A, PMSF, leupeptin), DNase, and 105 MDMs. RNase in PBS and disrupted mechanically using a bead beater. Triton MR and DC-SIGN-transfected COS-1 cell lines X-114 (Sigma-Aldrich) was added to the lysed cells to a final concentration of 8% (v/v), and after cooling on ice, the solution was mixed at 4°C over- COS-1-transfected cells in this study were obtained as previously described night. The cell wall was removed by centrifugation at 27,000 ϫ g for1h

(12). Briefly, full-length MR cDNA, derived from MDM mRNA, was http://www.jimmunol.org/ at 4°C, and the supernatant was incubated at 37°C to induce biphasic sep- cloned into the mammalian expression vector, pcDNA3.1/V5-His-TOPO aration (26). The upper aqueous layer was mixed with the cellular debris (Invitrogen Life Technologies). This expression construct was transfected and re-extracted as described above. The detergent layers were combined, into COS-1 cells. Stable clones of COS-1 expressing the MR were selected and the lipoglycans were precipitated by the addition of 9 volumes of cold by limiting dilution and expanded. An MR-positive clone was further ethanol (95%, Ϫ20°C). The precipitate was collected and treated with pro- sorted by flow cytometry (designated COS-1-MR) using a FITC-conju- teinase K for2hat60°C. The solution containing ManLAM and LM was gated MR mAb (BD Pharmingen). Full-length DC-SIGN cDNA was PCR dialyzed and lyophilized. For purification of ManLAM, HPLC was per- amplified from Human Placenta QUICK-Clone cDNA mix (BD Bio- formed on a Beckman liquid chromatography system fitted with a sciences/BD Clontech) and cloned into the mammalian expression vector Sephacryl S-200 HiPrep 16/60 column in tandem with a HiPrep 16/60 pSecTag2 (Invitrogen Life Technologies). The vector was transfected into Sephacryl S-100 column (Amersham Biosciences) equilibrated with 0.2 M COS-1 cells and cell surface expression of DC-SIGN was confirmed by

NaCl, 0.25% deoxycholate, 1 mM EDTA, 0.02% sodium azide, and 10 flow cytometry using an anti-DC-SIGN mAb. Further functional activity of by guest on September 29, 2021 mM Tris (pH 8.0) at a flow rate of 1 ml/min. SDS-PAGE and periodic the MR and DC-SIGN was assessed by flow cytometry using FITC-labeled acid-silver staining (27) were used to monitor the elution profile of the -BSA and zymosan in the absence or presence of mannan to measure fractions containing ManLAM and LM, which were then pooled and dia- mannan-inhibitable pinocytosis and phagocytosis, respectively. lyzed at 37°C without detergent followed by 1 M NaCl and water for several days. Purified ManLAM and LM fractions recovered were reana- Preparation of HSA- and lipoglycoconjugate-coated polystyrene lyzed by SDS-PAGE, sugar analysis and [1H]nuclear magnetic resonance beads to check for purity before detailed analysis. All steps of purification were performed using endotoxin-free water (Hospira). Endotoxin levels were Beads coated with HSA (sham control), ManLAM, LM, mixture of PIMs, calculated for all Ags in this study and were Ͻ18 pg/mg/sample. PIM families and PIM species were prepared as previously described (10) using Polybead polystyrene beads (1 or 6 ␮m). Briefly, 1.5 ϫ 109 Polybead One- and two-dimensional TLC analyses of total lipids from polystyrene beads were washed twice in 0.05 M carbonate-bicarbonate buffer (pH 9.6) and then incubated with 50 ␮g of various purified M.tb different mycobacterial strains H37Rv lipoglycoconjugates or buffer alone for1hat37°C. Beads were then For analysis of mannose-containing lipids located in the cell wall of dif- blocked with 5% HSA, washed repeatedly in 0.5% HSA, and finally ad- ϫ 8 ferent mycobacterial strains, M.tb strains and M. smegmatis were harvested justed to 4.0 10 /ml in 0.5% HSA before being used in the MDM or from Middlebrook 7H11 plates and lysed. quantification was per- COS-1 cell association assays. formed by the bicinchoninic acid method following the manufacturer’s Association assay using MDMs and COS-1 cells instructions (Bio-Rad). Lysates totaling 100 mg of protein from each strain

were delipidated at 37°C for 12 h using CHCl3:CH3OH (2:1, v/v) followed Before incubation with mannosylated lipoglycoconjugate-coated beads, ϩ by CHCl3:CH3OH:H2O (10:10:3, v/v/v) for an additional 12 h. Lipids were MDM monolayers were washed and repleted with RPMI 1640 10 mM analyzed by one- and two-dimensional TLC. For one-dimensional TLC, HEPES and 1 mg/ml HSA (UniZLB Bioplasma). M.tb H37Rv lipoglyco- strips (10 ϫ 20 cm) were cut from aluminum-backed thin layer sheets of conjugate-coated beads were added to cells at a ratio of 100:1 and mono- Kieselgel 60 (0.2 mm; Merck). Lipids from each strain were loaded based layers were incubated for 3 h with shaking at 100 rpm as previously de- ␮ on amounts of protein (100 g) from the bacterial lysates in CHCl3: scribed (10). In some cases, MDMs were preincubated with mannan CH3OH:H2O (10:10:3, v/v/v) and separated with CHCl3:CH3-COOH: (Sigma-Aldrich) at 2.5 mg/ml for 30 min at 37°C or with anti-MR mAb ␮ CH3OH:H2O (40:25:3:6, v/v/v/v) (28). Plates were dried and lightly (Serotec) at 10 g/ml for 20 min at 37°C to block the activity of the MR. sprayed with a solution of 10% concentrated sulfuric acid in absolute eth- These conditions have been shown to effectively inhibit binding to the MR anol, or ␣-naphthol or Dittmer reagent to screen for total lipids, glycolipids, in this assay (12, 30). Monolayers were fixed in 10% formalin and washed and phospholipids, respectively, and heated at 110°C until lipid bands ap- three times in Dulbecco’s PBS before coverslips being mounted on glass peared. For two-dimensional TLC analysis, total crude lipids from each slides. The average number of beads per cell was determined by counting strain were loaded onto the origin of 10 cm ϫ 10 cm TLC plates based 200–300 cells per coverslip using a ϫ100 oil-immersion objective with a upon equal amounts of protein content (100 ␮g) from the bacterial lysates wide bandwidth 570 nm dichroic mirror on a BX51 Olympus fluorescence Ϯ and run in the first dimension using CHCl3:CH3OH:H2O (60:30:6, v/v/v) microscope. The mean SEM was determined for triplicate coverslips in as a solvent system. The TLC plate was then rotated 90° to the left and run each test group. For each experiment, the mean ratio was tested for a

in the second dimension using CHCl3:CH3-COOH:CH3OH:H2O (40:25: significant difference from 1 using T statistics. The number of experiments 3:6, v/v/v/v) as a solvent system. TLC plates were developed as described cited were independent experiments performed using MDMs derived from above. Spots were identified as PIMs when positively reacted for Dittmer different donors. 1808 RECOGNITION OF PIMs BY MACROPHAGE C-TYPE LECTINS

For COS-1 cell association assays, freshly grown COS-1-wild-type somes in each test group in Ն2 independent experiments using a minimum (WT), COS-1-MR, or COS-1-DC-SIGN cells were seeded onto coverslips of two different donors. (1 ϫ 105 cells/well, triplicate wells/test group) in wells of a 24-well tissue culture plate and incubated overnight to form monolayers. Cell monolayers Statistical and three-dimensional prediction analyses were washed and repleted with PBS-HHG medium (PBS ϩ 10 mM ϩ ϩ Statistical analyses were performed using GraphPad Prism version 4.0 HEPES 1 mg/ml HSA 0.1% ) to which HSA- or lipoglyco- (͗www.graphpad.com͘). Three dimensional predictions of PIM structures conjugate-beads were added (ratios of 50–100:1) in the presence of ab- depending on the number and nature of their fatty acids were performed by sence of mannan (Sigma-Aldrich) as described for the MDM assays. The using ACD/ChemSketch version 8.0 (Advanced Chemistry Development, cells were incubated at 37°C for 3 h without shaking. After washing, the ͗www.acdlabs.com͘). cells were fixed with 2% paraformaldehyde for 10 min followed by several washes using Dulbecco’s PBS before the coverslips being mounted on glass slides. Bead association analysis was performed as described for the Results MDM assays. Purification of ManLAM, LM, and PIM families and species from virulent M.tb H37Rv Quantitative P-L fusion assay in MDMs using confocal microscopy To assess the association profile of individual PIM types with the human macrophage MR, we purified PIMs based on their mannose 5 MDM monolayers (2 ϫ 10 cells/well) on coverslips were incubated with content as well as degree of acylation. M.tb H R cells were har- isotype control IgG1 or anti-MR mAb at 37°C for 20 min as described for 37 v vested and delipidated to obtain total lipids. An enriched phospho- the binding experiments. Then, M.tb H37Rv lipoglycoconjugate-coated green fluorescent beads (4 ϫ 107/well) were added to the monolayers at lipid fraction containing mainly PIMs was obtained after total lipid 37°C for 2 h (12). Cell monolayers on coverslips were washed with PBS precipitation in cold acetone (17). To separate PIMs from other buffer, fixed with 2% paraformaldehyde, and permeabilized with 100% mycobacterial phospholipids further purification was necessary by Downloaded from methanol for 5 min at room temperature. MDMs were blocked overnight at silica column chromatography using a chloroform/methanol gra- 4°C in blocking buffer (Dulbecco’s PBS ϩ 5 mg/ml BSA ϩ 10% heat- inactivated FBS), stained with mAb against late endosomal/lysosomal dient. A mixture of PIM families was obtained in the 60% meth- marker CD63 (0.5 ␮g/ml final concentration), and next stained with Alexa anol in chloroform fraction. PIM families and species were finally Fluor 647-conjugated goat anti-mouse IgG secondary Ab (4 ␮g/ml final purified by TLC, visualized by one-dimensional TLC (Fig. 2A) and concentration). After extensive washing in blocking buffer, the coverslips identified by MALDI-TOF for each individual PIM species puri- were dried, mounted on glass slides, and examined by confocal microscopy (Zeiss LSM 510). fied (Fig. 2C). The molecular ions observed for Ac1PIM2 (m/z http://www.jimmunol.org/ The percentage of bead phagosomes that colocalized with the lysosomal 1475.9), Ac2PIM2 (m/z 1756.2), Ac1PIM5 (m/z 1985.9), Ac2PIM5 marker was quantified by counting over 1000 consecutive bead phago- (m/z 2154.4), Ac1PIM6 (m/z 2108.2), and Ac2PIM6 (m/z 2404.5) by guest on September 29, 2021

FIGURE 2. Purification of PIMs, LM, and ManLAM from M.tb. A, TLC showing the indicated purified PIM species and (B) a 15% SDS-PAGE showing the purified ManLAM and LM used in this study. C, Identification of each PIM species and its fatty acid composition by MALDI/mass spectrometry in a positive ion mode. The Journal of Immunology 1809 were attributed to their sodiated forms [M-H ϩ 2Na]ϩ with an mg/ml), a soluble carbohydrate inhibitor of MR activity on mac- additional sodium atom, presumably on the phosphate moiety. rophages (30) (Fig. 3C). As in previous studies, ManLAM beads Ͻ Other mannosylated components of the M.tb H37Rv cell wall, bound well to macrophages ( p 0.01) in a highly mannan-inhib- ManLAM and LM, were extracted and purified from the delipi- itable manner (63% inhibition, p Ͻ 0.0005). In contrast, LM bead dated cells as previously described (25) (Fig. 2B). association was low and not mannan-inhibitable when compared with HSA beads, as previously reported (10). The magnitude of Recognition of different PIM families by the MR on human association of PIM5f-coated beads with MDMs was comparable to macrophages that observed with ManLAM beads ( p Ͻ 0.05), followed in mag- To establish the association profile of PIM families with the mac- nitude by beads coated with a mixture of PIM families (PIMs) and rophage MR, we used our established bead model (10). MDM PIM6f. The association of PIM2f beads was poor, pointing toward ϫ 5 monolayers (2 10 ) on glass cover slips were incubated with the importance of the nonreducing termini of the PIM5f and PIM6f 2 ϫ 107 beads coated with a mixture of PIMs or individual PIM for association with MDMs. families (PIM2f, PIM5f and PIM6f) as well as ManLAM (positive Mannan inhibition indicative of MR involvement was greatest Ͻ control) and LM or HSA (negative controls (10)). The total num- for the PIM6f beads (61.3% inhibition, p 0.0005; comparable to ber of beads associated with the MDMs was enumerated by mi- ManLAM) followed by the mixture of PIMs (43.2% inhibition, p Ͻ Ͻ croscopy (Fig. 3, A and B) and MR-dependent association was 0.005) and PIM5f (42.4% inhibition, p 0.05) indicating that in ad- defined as the percent reduction in the presence of mannan (2.5 dition to ManLAM, higher-order PIMs preferentially associate with Downloaded from http://www.jimmunol.org/ by guest on September 29, 2021

FIGURE 3. Mannan-inhibitable association of PIM beads with MDMs. MDM monolayers (2 ϫ 105) on glass coverslips were incubated with 2 ϫ 107 PIM-, LM-, ManLAM-, or HSA (control)-coated beads. The total number of beads associated with the cells was enumerated by microscopy. Specific association of coated beads was measured in the presence of mannan (2.5 mg/ml), a soluble carbohydrate inhibitor of MR activity on macrophages. A, Representative phase-contrast microscopy images of MDM monolayers after incubation with control HSA- or M.tb lipoglycoconjugate-coated beads in the presence or absence of mannan. Beads associated with MDM monolayers (arrows) were visualized at ϫ1000 as round refractive objects. B, Association p Ͻ 0.01). C, Association of PIMf with the ,ءء ;p Ͻ 0.05 ,ء ;of PIMf with MDMs is shown as fold increase relative to control HSA beads (Student t test p Ͻ 0.0005). The data shown are from four ,ءءء ;p Ͻ 0.005 ,ءء ;p Ͻ 0.05 ,ء ;MDM MR as indicated by the percent mannan inhibition (Student t test Ϯ ϩ separate experiments (mean SEM). PIMs: mixture of PIM2f, PIM5f and PIM6f; PIM2f: mixture of Ac1PIM2 Ac2PIM2; PIM5f: mixture of Ac1PIM5 ϩ ϩ Ac2PIM5; PIM6f: mixture of Ac1PIM6 Ac2PIM6. 1810 RECOGNITION OF PIMs BY MACROPHAGE C-TYPE LECTINS the MR on macrophages (Fig. 3C). In this regard, the nonreducing have shown that the type and degree of acylation play a key role termini of the PIM5f and PIM6f resemble the mono-and dimanno- in regulating the bioactivity of LPSs from other bacteria (31–35). side cap of ManLAM, respectively. The low level of PIM2f asso- We purified individual PIM species (based on their degree of ciation was also mannan-inhibitable but to a lesser degree (12.6% acylation) and assessed their association with macrophages and the inhibition) suggesting that a small proportion of PIM2f association MR. Fig. 5A shows a representative experiment for the association with the macrophage surface is C-type lectin dependent assay in the absence (Ⅺ) or presence (f) of mannan. Data com- To complement the results we obtained with mannan, we further bined from four independent experiments shown in Fig. 5B pro- evaluated involvement of the macrophage MR in PIM association vide evidence that Ac1PIM6 beads (triacylated) associate with using an anti-MR mAb (12). MDMs were preincubated with anti- MDMs to a much greater degree than Ac2PIM6 (tetra-acylated) MR mAb or IgG1 isotype control before incubation with lipogly- beads ( p ϭ 0.001). Similarly, triacylated forms of other PIM spe- coconjugate-coated beads. Results in Fig. 4A show a representative cies associated with MDMs to a greater degree than their respec- Ⅺ f experiment in the absence ( ) or presence ( ) of anti-MR mAb tive tetra-acylated forms (Fig. 5B). For example, higher association and Fig. 4, B and C, show combined data from three independent was also observed for Ac PIM followed by Ac PIM , and experiments. In this set of experiments, bead association with 1 5 2 5 Ac PIM association was higher than Ac PIM ; however, in both MDM monolayers was greatest when beads were coated with 1 2 2 2 of the latter cases, the baseline association with MDMs was poor ManLAM ( p Ͻ 0.05) followed by a mixture of PIMs and the as was seen with the PIM f (Fig. 4B). PIM f, and this association was markedly inhibited when MDMs 2 5 Mannan-inhibitable association indicative of the MR involve- were pretreated with MR mAb (Fig. 4C). ManLAM association Ͻ

ment (Fig. 5C) was high for Ac PIM (53.3% inhibition, p Downloaded from was inhibited by 76.2% ( p Ͻ 0.0001) comparable to the level of 1 6 Ͻ inhibition observed in the presence of mannan (Fig. 3C). Similarly, 0.001) and Ac1PIM5 (56.7% inhibition, p 0.05), at the same Ͻ the association of beads coated with higher-order PIM families level as that seen with ManLAM (56.3% inhibition, p 0.005) Ͻ was MR dependent, with a reduction in association between 40 and followed by Ac2PIM5 (37.8% inhibition, p 0.005). Ac1PIM6 has 45% in the presence of anti-MR mAb (Fig. 4C). In contrast, MR- three fatty acids (two in the glycerol and the third in the C6 po- sition of the mannose attached to the C2 position of its myo-ino- dependent association of PIM2f beads was low (17.5% inhibition in the presence of anti-MR mAb) (Fig. 4C). Together, these data sitol) and Ac2PIM6 has four fatty acids (the fourth located in the http://www.jimmunol.org/ C3 position of its myo-inositol). Our data suggest that the addition provide evidence that the higher-order PIM families (PIM5f and

PIM6f) along with ManLAM preferentially associate with the mac- of the fourth fatty acid bound to beads alters the conformation of rophage MR, whereas the PIM2f does not. Ac2PIM6 and results in lower association of this lipoglycan with the MR. This fact may account for the variability of the macro- The degree of acylation of PIMs impacts their recognition by phage association seen with the PIM6f which contains a mixture of the macrophage MR Ac1PIM6 and Ac2PIM6. This variability is also observed to a lesser

PIMs vary in the degree and nature of their fatty acids. How this degree with the PIM5f (Ac1PIM5 and Ac2PIM5), but not with the impacts recognition by PRRs such as the MR is unknown. Studies PIM2f (Ac1PIM2 and Ac2PIM2) suggesting that in the case of the by guest on September 29, 2021

FIGURE 4. MR-dependent association of higher-order PIMs with MDMs. A, A representative experiment of MR-mediated association of lipoglyco- conjugate beads with MDMs. MDM monolayers in triplicate were preincubated with 10 ␮g/ml anti-MR mAb (f) or IgG1 isotype control (Ⅺ) before incubation with lipoglycoconjugate-coated or control HSA beads. B, Results from three independent experiments (mean Ϯ SEM) show the fold increase (relative to HSA beads) in association of ManLAM and PIM beads with MDM monolayers; and C, MR-dependent association of lipoglycoconjugate beads ;p Ͻ 0.005 ,ء ,with MDMs (represented by the percent inhibition in the presence of anti-MR mAb relative to the subtype control mAb). In A, Student t test Ͻ ءءء Ͻ ء in B and C, , p 0.05; , p 0.0001 when compared with control HSA beads. PIMs: mixture of PIM2f, PIM5f and PIM6f; PIM2f: mixture of Ac1PIM2 ϩ ϩ ϩ Ac2PIM2; PIM5f: mixture of Ac1PIM5 Ac2PIM5; PIM6f: mixture of Ac1PIM6 Ac2PIM6. The Journal of Immunology 1811 Downloaded from http://www.jimmunol.org/

FIGURE 5. The nature of the acyl group of PIMs bound to beads affects the ability of these lipoglycans to interact with the MR. A, A representative experiment performed in triplicate. Individual PIM species (based on their degree of acylation) were purified and assessed for their association with the MR in the presence (f) or absence (Ⅺ) of mannan. B, Results from four independent experiments (mean Ϯ SEM) show the fold increase (relative to HSA beads) in association of ManLAM and PIM species beads with MDM monolayers; and C, results from four independent experiments (mean Ϯ SEM) show p Ͻ 0.0005; in B, Studentءءء p Ͻ 0.001; and ,ءء ;p Ͻ 0.05 ,ء ,the percent mannan inhibition of ManLAM and PIM species beads. In A, Student t test .p Ͻ 0.005 ,ءء ;p Ͻ 0.05 ,ء ,p Ͻ 0.0001; and in C, Student t test ,ءءء ;p ϭ 0.001 ,ءء ,t test by guest on September 29, 2021

PIM2f, the fatty acids do not play a role in the low association of bition seen with ManLAM (66.3% inhibition) (Fig. 6, A and C). this family with the MR. Association profiles for Ac1PIM5 and Ac2PIM5 with the COS- 1-MR cells were also high and significantly inhibited in the pres- Mannan-inhibitable association of ManLAM and PIM beads ence of mannan (36.5% inhibition for Ac1PIM5 and 58% inhibition with COS-1-MR or COS-1-DC-SIGN cell lines Ͻ Ͻ for Ac2PIM5, p 0.05 and p 0.01, respectively). To confirm the direct involvement of the MR in PIM association In contrast to the results obtained with the COS-1-MR cells, with MDMs, we assayed for PIM association using a mammalian beads coated with ManLAM, LM, and all PIM types (especially cell line expressing the MR (COS-1-MR) and compared the level Ac1PIM5,Ac2PIM2, and Ac1PIM2) associated with the COS-1- of association to that seen with a cell line expressing DC-SIGN DC-SIGN cells (Fig. 6B) and in all cases, the association was (COS-1-DC-SIGN), another C-type lectin known to bind to the significantly reduced in the presence of mannan (Fig. 6D). Asso- terminal mannose caps of ManLAM (13). DC-SIGN is expressed ciation of lipoglycoconjugate-coated beads with COS-1 WT cells in abundance on DCs (36). In contrast, recent studies using flow was negligible and not mannan-inhibitable (data not shown). Thus, cytometry and Western blotting (37–39) together with our own both carbohydrate and fatty acid composition play important roles work (our unpublished data) indicate that there is little or no DC- in enabling the recognition of PIMs by the MR, whereas specific SIGN expression on the surface of MDMs. Lipoglycoconjugate- recognition of PIMs by DC-SIGN appears to be relatively inde- coated beads were added to COS-1-MR or COS-1-DC-SIGN cell pendent of the nature of the carbohydrates and fatty acids. monolayers and specific association of coated beads with the MR or DC-SIGN was measured in the absence (Fig. 6, A and B, re- spectively) or presence (Fig. 6, C and D, respectively) of mannan The role of the MR in directing P-L fusion events in (2.5 mg/ml). Similar to the data with MDMs, Ac2PIM2 and macrophages for higher-order PIM-coated beads Ac1PIM2 beads associated poorly with the COS-1-MR cells (Fig. 6A) and the association was not mannan-inhibitable (Fig. 6C). ManLAM is a critical regulator of phagosome maturation in mu- rine macrophages and a human monocytic cell line (40, 41). Re- Ac2PIM6 beads also associated poorly with the cells (Fig. 6A)as was seen with the MDMs (Fig. 5B); however in this case, the cently, we have shown that engagement of the MR by ManLAM association was mannan-inhibitable (43.5% inhibition) (Fig. 6C). through its mannose caps during the phagocytic process directs

In contrast, Ac1PIM6 beads associated with COS-1-MR cells to a M.tb to its initial phagosomal niche in human macrophages by high degree (equivalent to that seen with ManLAM) and the as- limiting P-L fusion (12). Blockade of the MR with a mAb reversed sociation was reduced to control levels in the presence of mannan the inhibition of P-L fusion. The fact that Ac1PIM5,Ac2PIM5, and

(62.4% inhibition), also equivalent to the degree of mannan inhi- Ac1PIM6 have greater association with the MR when compared 1812 RECOGNITION OF PIMs BY MACROPHAGE C-TYPE LECTINS

FIGURE 6. Mannan-inhibitable association of ManLAM and PIM beads with COS-1-MR or COS-1- DC-SIGN cell lines. Cells (2 ϫ 105) were adhered to glass coverslips in 24-well plates and washed. Lipogly- coconjugate-coated beads were added to cell monolayers (multiplic- ity of infection 50–100:1). Specific association of coated beads with the MR (n ϭ 3) (A) or DC-SIGN (repre- sentative experiment) (B) was mea- sured in the absence or presence of mannan (2.5 mg/ml). Percent mannan inhibition was calculated and is shown in C for COS-1-MR (n ϭ 3) ϭ

and D for COS-1-DC-SIGN (n 3) Downloaded from (mean Ϯ SEM). In A, Student t test, p Ͻ ,ءءء ;p Ͻ 0.005 ,ءء ;p Ͻ 0.05 ,ء p Ͻ ,ء ,in B, Student t test ;0.0001 p Ͻ 0.0005; in C, Student ,ءءء ;0.05 p Ͻ 0.005; and ,ءء ;p Ͻ 0.05 ,ء ,t test ,ءء ;p Ͻ 0.05 ,ء ,in D, Student t test

p Ͻ 0.001. http://www.jimmunol.org/

with the other M.tb PIMs, and that the nonreducing termini of focal microscopy. Results (Fig. 7) showed that when the MR was these PIMs resemble the mono- and dimannoside caps of Man- blocked, ManLAM- (used as positive control; p Ͻ 0.0001), Ͻ Ͻ LAM, directed us to examine the degree to which MR blockade on Ac2PIM5-(p 0.05), and Ac1PIM6-(p 0.05) coated beads macrophages by anti-MR Ab affects P-L fusion of beads coated demonstrated a significant increase in P-L fusion when compared by guest on September 29, 2021 with several PIM species. with the sham control HSA beads. In contrast, blockade of the MR

MDM monolayers were preincubated with anti-MR or control had minimal effects on P-L fusion for Ac2PIM2- and Ac2PIM6- IgG1 mAbs followed by the addition of either ManLAM-, coated beads (Fig. 7B). Thus, these results indicate that the MR

Ac2PIM2- (as a representative of the PIM2f), Ac2PIM5- (as a rep- regulates P-L fusion events for higher-order PIMs but not lower- resentative of the PIM5f), Ac1PIM6-orAc2PIM6-coated fluores- order PIMs and therefore correlate with MR recognition by the cent beads and subsequent assessment for P-L fusion using con- former PIM types.

FIGURE 7. Higher-order PIMs that associate with the macrophage MR demonstrate increased P-L fusion upon MR blockade. MDM monolayers on coverslips were incubated with isotype control IgG1 or anti-MR

mAb at 37°C for 20 min before adding M.tb H37Rv li- poglycoconjugate-coated green fluorescent beads (4 ϫ 107/well) and allowing for phagocytosis to occur at 37°C for 2 h. MDM monolayers were fixed, permeabil- ized, and then immunostained with the late endosomal/ lysosomal marker CD63. Coverslips were mounted on glass slides and examined by confocal microscopy. A, Representative fluorescence microscopy images of MDM monolayers after incubation with control HSA- or M.tb lipoglycoconjugate-coated beads in the presence or absence of anti-MR mAb. Compartments that stain positive for CD63 are red, beads in unfused phagosomes are green, and beads in fused phagolysosomes are yel- low. B, Percent increase in P-L fusion of HSA- or M.tb lipoglycoconjugate-coated beads in the presence of anti- MR mAb is shown as analyzed by confocal microscopy. P-L fusion was enumerated by counting over 1000 con- secutive phagosomes in each test group (mean Ϯ SEM, Ͻ ء ϭ ϭ n 3, except for Ac2PIM2, n 2), Student t test, , p .p Ͻ 0.0001 ,ءءء ;0.05 The Journal of Immunology 1813

M.tb strains contain greater amounts of higher-order PIMs in tive immune responses against M.tb. In addition to ManLAM, their cell envelope than the nonpathogenic M. smegmatis other cell wall mannosylated lipoglycoconjugates of M.tb are be- mc2155 strain lieved to interact with the MR and/or DC-SIGN, thereby initiating To further explore the importance of higher-order PIMs in M.tb host cell responses and influencing the fate of the bacillus within recognition by the MR and their potential involvement in M.tb the host. Based on their structures and functional domains, it is pathogenesis, we purified cell envelope PIMs from the virulent predicted that there will be differences in recognition of manno- sylated lipoglycoconjugates between these C-type lectins. In this M.tb Erdman and M.tb H37Rv strains, the attenuated M.tb H37Ra strain and the saprophytic M. smegmatis mc2155 strain, and com- study, we evaluated the lectin association pattern of the PIMs, pared the relative amounts of higher-order and lower-order PIMs. major phospholipids present in the M.tb cell wall. To accomplish Bacteria were grown under identical conditions (except 12 days for this, we purified PIM families based on their mannose content M.tb strains and 5 days for the M. smegmatis strain due to the followed by purification of PIM species based on the number of different growth rate between both species), homogenized, and fatty acids. This proved to be a powerful approach in enabling us PIMs resolved by two-dimensional TLCs for each strain (loading to demonstrate for the first time that there is preferential engage- was based on protein equivalents of cell lysates). A representative ment of the MR by higher-order PIMs in contrast to the engage- experiment (Fig. 8) shows a clear difference in the amounts of ment of DC-SIGN by both higher-order and lower-order PIMs and higher-order PIMs between strains. In all cases, M.tb strains have LM. Preferential engagement of the MR by higher-order PIMs was found to regulate P-L fusion events in macrophages. Thus, this significantly more Ac1PIM6,Ac1PIM5, and Ac2PIM5 (Fig. 8, spots 1, 2, and 3, respectively) when compared with M. smegmatis work makes the fundamental observation that recognition of PIMs 2 Downloaded from mc 155, while lower-order PIMs (Ac1PIM2 and Ac2PIM2) are by the MR and DC-SIGN is dependent on both the nature of their much less predominant in M.tb than in M. smegmatis (Fig. 8 spots mannosyl units and the fatty acid content which impacts the host 4 and 5, respectively). In addition, virulent M.tb strains (M.tb Er- response. dman and M.tb H37Rv) have quantitatively more higher-order The human macrophage has high MR activity and negligible PIMs (especially Ac1PIM6) than the attenuated M.tb H37Ra strain. DC-SIGN activity (37, 43). To demonstrate specific interaction The abundance of exposed higher-order PIMs along with Man- between PIMs and the MR, we used a model established in this LAM likely facilitates the interaction of M.tb with the MR and/or laboratory in which M.tb lipoglycoconjugates are bound to poly- http://www.jimmunol.org/ DC-SIGN. Further, our data provide evidence that surface man- styrene beads by their lipid tails with the carbohydrates preferen- nosylation of virulent strains of M.tb promotes the interaction with tially exposed for engagement of macrophage C-type lectins (10). human macrophages because we have found that M. smegmatis This technique has been used successfully in several studies with associates poorly with these cells (42). mycobacterial lipoglycoconjugates (23, 29, 44). MR activity is blocked with mannan (10) and an anti-MR mAb (12). Using these Discussion approaches with human macrophages and an MR-expressing The MR and DC-SIGN are two major C-type lectins involved in mammalian cell line, we demonstrate that higher-order PIM fam-

the recognition of M.tb ManLAM by macrophages (10) and den- ilies (PIM5f and PIM6f) associate well with the MR, whereas the by guest on September 29, 2021 dritic cells (13, 14), respectively, influencing both innate and adap- lower-order PIM2f does not. We confirmed that ManLAM is a

FIGURE 8. Two-dimensional TLC analysis of the total lipid from virulent and attenuated M.tb strains shows increased production of higher-order PIMs when compared with the saprophytic M. smegmatis. PIMs from the indicated strains were extracted from 10 mg of whole cell mycobacterial lysates by organic extraction before being dried and resuspended in CHCl3:CH3OH (2:1, v/v) for two-dimensional TLC analysis. Following separation, plates were sprayed with ␣-naphthol or Ditt- mer reagent (data not shown) and charred to visualize glycolipids and phospholipids, respectively. Spots were identified as PIMs went positive for ␣-naphthol and Dittmer reagent. Results are representative of three in- dependent experiments. Equal amounts of samples were loaded per TLC according to total protein content as described in Materials and Methods. Spot 1, Ac1PIM6; 2, Ac1PIM5;3,Ac2PIM5;4,Ac1PIM2;5,Ac2PIM2. 1814 RECOGNITION OF PIMs BY MACROPHAGE C-TYPE LECTINS major ligand for the MR (via binding of its mannose caps) whereas larger lipoglycans such as ManLAM or LM. This concept is sup- LM demonstrates negligible binding to this receptor (10), a finding ported by a recent study with ManLAM and the human pulmonary which indicates that the MR recognizes the branches [t-Manp- C-type lectin surfactant protein A where the authors suggested that ␣(132)-Manp-␣(136)-] and the linear terminus [t-Manp- the lipid moiety of ManLAM is responsible for the macromolec- ␣(136)-Manp-␣(136)-] present in LM poorly. The fact that ular organization of this lipoglycan that enables C-type lectin in-

PIM5f and PIM6f termini resemble the mono- and dimannoside teraction (49). caps of ManLAM [t-2-Manp and t-Manp(132)-␣-Manp-, respec- In M.tb, the fourth fatty acid of PIMs is most frequently palmitic tively] together with their location on the cell wall surface, allows acid, tuberculostearic (TBST) acid or oleic acid (17, 22). Although us to postulate that both families play an important role in the the importance of the nature of this fourth fatty acid is not clear for interaction of the bacillus with human macrophages through rec- C-type lectin recognition by the macrophage, previous studies per- ognition by the MR. The PIM6f accumulates as a final biosynthetic formed in dendritic cells (48) attribute proinflammatory responses product on the M.tb cell wall (4, 17, 45), and as such, is likely to to the TBST. Three dimensional computer modeling of Ac1PIM6 play a particularly important role in bacterial recognition by this vs Ac2PIM6 (Fig. 9), when both are bound to a flat surface, showed receptor. that Ac1PIM6 adopts a spatial conformation with a different plane Important differences exist in the carbohydrate recognition do- of axis of the nonreducing terminus when compared with Ac2PIM6 main (CRD) structural motifs between the MR and DC-SIGN (Fig. 9). We speculate that this change in axis may explain the making it likely that these are not redundant in their func- marked difference in association with the MR between these two tions, particularly with regard to M.tb carbohydrate recognition. In species from the same family (i.e., the dimannoside termini in Downloaded from this context, the association profile for the PIM families and LM Ac1PIM6 are in a more favorable conformation to bind to the MR). differed markedly between these two C-type lectins. In the case of In addition, when we evaluated how the nature of the fatty acid in the MR, association was highest for the higher-order PIMs. For Ac2PIM6 affects its spatial conformation, there was a slight affect DC-SIGN, mannan-inhibitable association was independent of the on the axis when the fourth fatty acid was substituted with palmitic number and location of the mannose moieties in PIMs and LM. or oleic acids relative to TBST. These findings lend support to the Therefore, DC-SIGN appears to recognize mannose-oligosaccha- importance of the type and nature of fatty acids in dictating the

rides of any length. The multiple linear CRDs located in the ex- spatial conformation of the carbohydrates in the PIMs. http://www.jimmunol.org/ tracellular domain of the MR (46) appear to favor the recognition of the mono- and dimannoside cap-like structures present in high- er-order PIMs. In contrast, DC-SIGN has a single CRD and forms tetramers on the cell surface, where the DC-SIGN tetramer clusters its CRDs to enable high avidity binding to mannose-containing glycoconjugates (46). Together, these results are consistent with the idea that the spatial orientation of MR CRDs limits its sugar recognition repertoire, whereas the clustered CRD conformation of tetrameric DC-SIGN enables a broader range of mannose glyco- by guest on September 29, 2021 conjugate recognition. Our data indicate that the nature of the fatty acids in PIMs play a role in their recognition by the MR. In our bead model, the degree of acylation of the PIM6f markedly affected their ability to ϾϾ associate with the MR (Ac1PIM6 Ac2PIM6) pointing toward the importance of the fatty acids in the conformation and accessi- bility of the terminal mannosyl cap-like structure. This was dem- onstrated using both MDMs and COS-1-MR cells. There are con- tradictory data in the literature regarding the importance of the acyl groups of PIMs in generating proinflammatory cytokines depend- ing on the model used. Some studies showed that when presented free in solution, mixtures of PIMs, PIM2 and PIM6 families and ␣ purified species of PIM2f and PIM6f induce TNF- via a TLR-2- dependent pathway irrespective of their acyl forms (19, 22, 47). In contrast, a recent study using the liposome model showed that the presence of a fourth fatty acid in the C3 position of the myo- FIGURE 9. Carbohydrate linkage and fatty acid composition of M.tb inositol of Ac2PIM2 diminished the production of proinflammatory PIMs determine their recognition by human macrophage C-type lectins. cytokines (48). Based on our findings, it is possible that this ob- Shown is a three-dimensional prediction (0° and 90° rotation around the servation was related to lower binding of Ac2PIM2 to its recep- phosphate, see directional arrow) of Ac1PIM6 vs Ac2PIM6 depicting the tor(s). Our data show lower association of the tetra-acylated PIM role of the acyl groups in determining the spatial conformation of the PIM species with the MR. Thus, the presence of this fourth fatty acid, terminal dimannosyl cap. The change in spatial conformation observed in the cap may explain the marked difference in binding to the CRDs of the especially in Ac2PIM6, may induce a spatial conformational change which diminishes its efficient interaction with the CRDs of MR between Ac1PIM6 and Ac2PIM6.Ac1PIM6 beads associate to a much greater extent with the macrophage MR than Ac PIM beads. Ac PIM has the MR. Although it is possible that our findings could be related 2 6 1 6 three fatty acids (two in the glycerol and the third in the C6 position of the to the bead model used, perhaps with altered coating of Ac PIM , 2 6 mannose attached to the C2 position of its myo-inositol). The three fatty this phenomenon was not observed for association of PIMs with acids of Ac1PIM6 bound to the bead allow for a spatial conformation of the DC-SIGN making this possibility unlikely. Thus, the presence of cap that is optimal for recognition by the MR (equivalent to ManLAM). the fourth fatty acid attached to the myo-inositol may be critical for The addition of a fourth fatty acid located in the C3 position of the myo-

C-type lectin recognition and this would be particularly pro- inositol alters the conformation of the cap of Ac2PIM6 such that the asso- nounced for the PIMs, which are small glycolipids relative to the ciation of this lipoglycan with the MR is markedly reduced. The Journal of Immunology 1815

Evidence is emerging for the importance of mycobacterial PIMs consequently aiding the survival of the bacillus in human in host cell recognition and response (23, 50). These studies macrophages. showed that human macrophages preincubated with PIM2f and In summary, we purified the different PIM families and species PIM6f inhibited phagocytosis of mycobacteria. PIM2f-coated found in M.tb and have characterized their association with two beads were found to associate with complement receptor 3 and not important host C-type lectins, the MR and DC-SIGN, in tubercu- the MR (23). These results complement our observations, where losis pathogenesis. Whereas higher-order PIMs and ManLAM, the PIM2f associated poorly with the MR and is heavily expressed both heavily expressed on the M.tb surface, preferentially engage in the cell wall of M. smegmatis when compared with M.tb. How- the MR and direct P-L fusion inhibition, these lipoglycoconjugates ever, our results showed that the low level of PIM2f association along with lower-order PIMs (expressed in abundance in M. smeg- with macrophages was mannan inhibitable raising the possibility matis) and LM engage DC-SIGN. Association of higher-order for involvement of a C-type lectin. In this regard, we found that PIMs with the MR is dependent on the number of fatty acids in the purified species of the PIM2f (Ac1PIM2 and Ac2PIM2) associate lipid anchor. The generation of highly purified M.tb cell envelope well with DC-SIGN. Because MDMs express negligible DC- lipoglycoconjugates will allow us to do more detailed studies of SIGN, it remains unclear whether this is an important PRR for the the relationship between the MR and DC-SIGN and their signaling

PIM2f on macrophages. and trafficking pathways. Our results suggest that the balance be- A study using murine bone marrow-derived dendritic cells has tween the relative amounts of ManLAM plus higher-order PIMs shown that PIM2f liposomes activate a marked proinflammatory (preferentially recognized by the MR) and PIM2f plus LM (rec- cytokine response, a finding that the authors speculated was due to ognized by DC-SIGN and other receptors) exposed on the surface an interaction with TLRs (48). Our findings provide evidence that of different M.tb strains will influence the to Downloaded from the PIM2f is recognized differently from higher-order PIMs by M.tb infection. macrophages and potentially by dendritic cells. Whereas the PIM2f associates with complement receptor 3 (23), DC-SIGN (our study), Acknowledgments and potentially TLRs, we demonstrate that higher-order PIMs pref- We thank Dr. Joanne Turner for careful reading of this manuscript and the M.tb erentially associate with the MR and DC-SIGN. engagement help provided by the Mass Spectrometry and Proteomics Facility in the of the MR and DC-SIGN has been shown to inhibit the proinflam- Campus Chemical Instrument Center at Ohio State University. http://www.jimmunol.org/ matory response of phagocytes (14, 51). We make the important observation that M.tb expresses low quantities of the PIM2fonits surface and, instead, expresses abundant higher-order PIMs. This Disclosures result along with previous studies with ManLAM indicate that vir- The authors have no financial conflict of interest. ulent M.tb has evolved to more heavily decorate its abundant cell wall lipoglycoconjugates with ␣(132) mannosylated termini. References These termini resemble those present in high mannose N-linked 1. Schlesinger, L. S. 1997. The role of mononuclear phagocytes in tuberculosis. In of newly produced in eukaryotic Lung Macrophages and Dendritic Cells in Health and Disease. M. F. Lipscomb and S. W. Russell, eds. Dekker, New York, pp. 437–480. ␣ 3 by guest on September 29, 2021 cells (52). Such heavily (1 2) mannosylated N-glycoproteins 2. Fenton, M. J., L. W. Riley, and L. S. Schlesinger. 2005. Receptor-mediated released into circulation are cleared by macrophages via the MR recognition of Mycobacterium tuberculosis by host cells. In Tuberculosis and the by pinocytosis to maintain homeostasis (53). We speculate that Tubercle Bacillus. S. T. Cole, K. D. Eisenach, D. N. McMurray, and W. R. Jacobs, Jr., eds. ASM Press, New York, pp. 405–426. increased ␣(132) mannosylation on the surface of M.tb allows for 3. Crick, D. C., P. J. Brennan, and M. R. McNeil. 2003. The cell wall of Myco- preferential engagement of the MR on macrophages in a form of bacterium tuberculosis. In Tuberculosis. W. M. Rom and S. M. Garay, eds. Lip- exploitation to enhance its survival. pincott Williams and Wilkins, Philadelphia. 4. Besra, G. S., C. B. Morehouse, C. M. Rittner, C. J. Waechter, and P. J. Brennan. Consistent with this notion, we have recently determined that 1997. Biosynthesis of mycobacterial lipoarabinomannan. J. Biol. Chem. 272: MR-mediated phagocytosis of ManLAM-coated microspheres and 18460–18466. M.tb leads to limited P-L fusion in human macrophages (12). In 5. Chatterjee, D., and K.-H. Khoo. 1998. Mycobacterial lipoarabinomannan: an ex- traordinary lipoheteroglycan with profound physiological effects. Glycobiology 8: this study, we show that in addition to ManLAM, higher-order 113–120.

PIMs that bind to the MR (i.e., Ac2PIM5 and Ac1PIM6 but not 6. Gilleron, M., J. Nigou, B. Cahuzac, and G. Puzo. 1999. Structural study of the Ac PIM ) demonstrate a significant increase in P-L fusion upon lipomannans from Mycobacterium bovis BCG: characterisation of multiacylated 2 6 forms of the phosphatidyl-myo-inositol anchor. J. Mol. Biol. 285: 2147–2160. MR blockade further emphasizing the importance of the MR path- 7. Hunter, S. W., and P. J. Brennan. 1990. Evidence for the presence of a phos- way in directing vesicle trafficking for M.tb within the macro- phatidylinositol anchor on the lipoarabinomannan and lipomannan of Mycobac- phages. Our P-L fusion results with Ac PIM (as a representative terium tuberculosis. J. Biol. Chem. 265: 9272–9279. 2 2 8. Ozanne, V., A. Ortalo-Magne´, A. Vercellone, J. J. Fournie, and M. Daffe. 1996. of the lower-order PIMs) support previous studies where it was Cytometric detection of mycobacterial surface antigens: exposure of mannosyl epitopes and of the arabinan segment of arabinomannans. J. Bacteriol. 178: shown that a mixture of PIM1f and PIM2f did not inhibit P-L fusion (44). The fact that components of the PIM f did not bind to 7254–7259. 2 9. Vercellone, A., J. Nigou, and G. Puzo. 1998. Relationships between the structure the MR adds further evidence for the specificity of the MR path- and the roles of lipoarabinomannans and related glycoconjugates in tuberculosis way in limiting P-L fusion. pathogenesis. Frontiers Biosci. 3: 149–163. M. smegmatis 10. Schlesinger, L. S., S. R. Hull, and T. M. Kaufman. 1994. Binding of the terminal Recently, we have shown that a strain of over- mannosyl units of lipoarabinomannan from a virulent strain of Mycobacterium expressing a phosphomannomutase (manB) involved in the bio- tuberculosis to human macrophages. J. Immunol. 152: 4070–4079. synthesis of GDP-mannose (the mannose donor for PIM biosyn- 11. Schlesinger, L. S., T. M. Kaufman, S. Iyer, S. R. Hull, and L. K. Marciando. 1996. Differences in mannose receptor-mediated uptake of lipoarabinomannan thesis (4)), expressed large amounts of higher-order PIMs on its from virulent and attenuated strains of Mycobacterium tuberculosis by human surface (mainly Ac1PIM5,Ac2PIM5, and Ac1PIM6) (42). As a con- macrophages. J. Immunol. 157: 4568–4575. sequence of overexpression of higher-order PIMs, this strain dem- 12. Kang, B. K., A. K. Azad, J. B. Torrelles, T. M. Kaufman, A. A. Beharka, E. Tibesar, L. E. Desjardin, and L. S. Schlesinger. 2005. The human macrophage onstrated a 13-fold increase in binding to the human macrophage mannose receptor directs Mycobacterium tuberculosis lipoarabinomannan-medi- MR when compared with control strains (42). This fact, along with ated phagosome biogenesis. J. Exp. Med. 202: 987–999. our current findings, further support the idea that higher-order 13. Maeda, N., J. Nigou, J. L. Herrmann, M. Jackson, A. Amara, P. H. Lagrange, G. Puzo, B. Gicquel, and O. Neyrolles. 2003. The cell surface receptor DC-SIGN PIMs expressed on the surface of M.tb will bind to the MR and discriminates between Mycobacterium species through selective recognition of participate together with ManLAM in restricting P-L fusion and the mannose caps on lipoarabinomannan. J. Biol. Chem. 278: 5513–5516. 1816 RECOGNITION OF PIMs BY MACROPHAGE C-TYPE LECTINS

14. Geijtenbeek, T. B., S. J. Van Vliet, E. A. Koppel, M. Sanchez-Hernandez, 34. Kawasaki, K., R. K. Ernst, and S. I. Miller. 2004. 3-O-deacylation of lipid A by C. M. Vandenbroucke-Grauls, B. Appelmelk, and Y. Van Kooyk. 2003. Myco- PagL, a PhoP/PhoQ-regulated deacylase of Salmonella typhimurium, modulates bacteria target DC-SIGN to suppress dendritic cell function. J. Exp. Med. 197: signaling through Toll-like receptor 4. J. Biol. Chem. 279: 20044–20048. 7–17. 35. Katz, S. S., Y. Weinrauch, R. S. Munford, P. Elsbach, and J. Weiss. 1999. Deacy- 15. Riedel, D. D., and S. H. E. Kaufmann. 2000. Differential tolerance induction by lation of lipopolysaccharide in whole Escherichia coli during destruction by cel- lipoarabinomannan and lipopolysaccharide in human macrophages. Microbes In- lular and extracellular components of a rabbit peritoneal inflammatory exudate. fect. 2: 463–471. J. Biol. Chem. 274: 36579–36584. 16. Bernardo, J., A. M. Billingslea, R. L. Blumenthal, K. F. Seetoo, E. R. Simons, 36. Geijtenbeek, T. B., R. Torensma, S. J. Van Vliet, G. C. van Duijnhoven, and M. J. Fenton. 1998. Differential responses of human mononuclear phagocytes G. J. Adema, Y. Van Kooyk, and C. G. Figdor. 2000. Identification of DC-SIGN, to mycobacterial lipoarabinomannans: role of CD14 and the mannose receptor. a novel dendritic cell-specific ICAM-3 receptor that supports primary immune Infect. Immun. 66: 28–35. responses. Cell 100: 575–585. 17. Khoo, K.-H., A. Dell, H. R. Morris, P. J. Brennan, and D. Chatterjee. 1995. 37. Puig-Kroger, A., D. Serrano-Gomez, E. Caparros, A. Dominguez-Soto, Structural definition of acylated phosphatidylinositol mannosides from Mycobac- M. Relloso, M. Colmenares, L. Martinez-Munoz, N. Longo, N. Sanchez-Sanchez, terium tuberculosis: definition of a common anchor for lipomannan and lipoarabi- M. Rincon, et al. 2004. Regulated expression of the pathogen receptor dendritic nomannan. Glycobiology 5: 117–127. cell-specific intercellular adhesion molecule 3 (ICAM-3)-grabbing nonintegrin in 18. Chan, E. D., K. R. Morris, J. T. Belisle, P. Hill, L. K. Remigio, P. J. Brennan, and THP-1 human leukemic cells, monocytes, and macrophages. J. Biol. Chem. 279: D. W. H. Riches. 2001. Induction of inducible nitric oxide synthase-NO⅐ by 25680–25688. lipoarabinomannan of Mycobacterium tuberculosis is mediated by MEK1-ERK, 38. Serrano-Gomez, D., A. Dominguez-Soto, J. Ancochea, J. A. Jimenez-Heffernan, MKK7-JNK, and NF-␬B signaling pathways. Infect. Immun. 69: 2001–2010. J. A. Leal, and A. L. Corbi. 2004. Dendritic cell-specific intercellular adhesion molecule 3-grabbing nonintegrin mediates binding and internalization of As- 19. Gilleron, M., C. Ronet, M. Mempel, B. Monsarrat, G. Gachelin, and G. Puzo. pergillus fumigatus conidia by dendritic cells and macrophages. J. Immunol. 173: 2001. Acylation state of the phosphatidylinositol mannosides from Mycobcterium 5635–5643. bovis bacillus Calmette-Guerin and ability to induce granuloma and recruit nat- 39. Dominguez-Soto, A., A. Puig-Kroger, M. A. Vega, and A. L. Corbi. 2005. PU. ural killer T cells. J. Biol. Chem. 276: 34896–34904. 1 regulates the tissue-specific expression of DC-sign. J. Biol. Chem. 280: 20. Barnes, P. F., D. Chatterjee, J. S. Abrams, S. Lu, E. Wang, M. Yamamura, 33123–33131. P. J. Brennan, and R. L. Modlin. 1992. Cytokine production induced by Myco- 40. Hmama, Z., K. Sendide, A. Talal, R. Garcia, K. Dobos, and N. E. Reiner. 2004. Downloaded from bacterium tuberculosis lipoarabinomannan: relationship to chemical structure. Quantitative analysis of phagolysosome fusion in intact cells: inhibition by my- J. Immunol. 149: 541–547. cobacterial lipoarabinomannan and rescue by an 1-␣,25-dihydroxyvitamin D3- 21. Vergne, I., R. A. Fratti, P. J. Hill, J. Chua, J. Belisle, and V. Deretic. 2004. phosphoinositide 3-kinase pathway. J. Cell Sci. 117: 2131–2140. Mycobacterium tuberculosis phagosome maturation arrest: mycobacterial phos- 41. Chua, J., I. Vergne, S. Master, and V. Deretic. 2004. A tale of two lipids: My- phatidylinositol analog phosphatidylinositol mannoside stimulates early endoso- cobacterium tuberculosis phagosome maturation arrest. Curr. Opin. Microbiol. 7: mal fusion. Mol. Biol. Cell 15: 751–760. 71–77. 22. Gilleron, M., V. F. Quesniaux, and G. Puzo. 2003. Acylation state of the phos- 42. McCarthy, T. R., J. B. Torrelles, A. S. Macfarlane, M. Katawczik, B. Kutzbach, Mycobacterium bovis phatidylinositol hexamannosides from bacillus Calmette- L. E. Desjardin, S. Clegg, J. B. Goldberg, and L. S. Schlesinger. 2005. Overex- http://www.jimmunol.org/ Guerin and Mycobacterium tuberculosis H37Rv and its implication in Toll-like pression of Mycobacterium tuberculosis manB, a phosphomannomutase that in- receptor response. J. Biol. Chem. 278: 29880–29889. creases phosphatidylinositol mannoside biosynthesis in Mycobacterium smegma- 23. Villeneuve, C., M. Gilleron, I. Maridonneau-Parini, M. Daffe, tis and mycobacterial association with human macrophages. Mol. Microbiol. 58: C. Astarie-Dequeker, and G. Etienne. 2005. Mycobacteria use their surface-ex- 774–790. posed glycolipids to infect human macrophages through a receptor-dependent 43. McGreal, E. P., L. Martinez-Pomares, and S. Gordon. 2004. Divergent roles for process. J. Lipid Res. 46: 475–483. C-type lectins expressed by cells of the innate . Mol. Immunol. 41: 24. de la Salle, H., S. Mariotti, C. Angenieux, M. Gilleron, L. F. Garcia-Alles, 1109–1121. D. Malm, T. Berg, S. Paoletti, B. Maitre, L. Mourey, et al. 2005. Assistance of 44. Fratti, R. A., J. Chua, I. Vergne, and V. Deretic. 2003. Mycobacterium tubercu- microbial antigen processing by CD1e. Science 310: 1321–1324. losis glycosylated phosphatidylinositol causes phagosome maturation arrest. 25. Torrelles, J. B., K. H. Khoo, P. A. Sieling, R. L. Modlin, N. Zhang, Proc. Natl. Acad. Sci. USA 100: 5437–5442. A. M. Marques, A. Treumann, C. D. Rithner, P. J. Brennan, and D. Chatterjee. 45. Ballou, C. E., and Y. C. Lee. 1964. The structure of a myo-inositol mannoside

2004. Truncated structural variants of lipoarabinomannan in Mycobacterium lep- from Mycobacterium tuberculosis glycolipid. Biochemistry 3: 682–685. by guest on September 29, 2021 rae and an ethambutol-resistant strain of Mycobacterium tuberculosis. J. Biol. 46. McGreal, E. P., J. L. Miller, and S. Gordon. 2005. Ligand recognition by antigen- Chem. 279: 41227–41239. presenting cell C-type lectin receptors. Curr. Opin. Immunol. 17: 18–24. 26. Bordier, C. 1981. Phase separation of integral membrane proteins in Triton X-114 47. Jones, B. W., T. K. Means, K. A. Heldwein, M. A. Keen, P. J. Hill, J. T. Belisle, solution. J. Biol. Chem. 256: 1604–1607. and M. J. Fenton. 2001. Different Toll-like receptor agonists induce distinct mac- 27. Prinzis, S., D. Chatterjee, and P. J. Brennan. 1993. Structure and antigenicity of rophage responses. J. Leukocyte Biol. 69: 1036–1044. lipoarabinomannan from Mycobacterium bovis BCG. J. Gen. Microbiol. 139: 48. Sprott, G. D., C. J. Dicaire, K. Gurnani, S. Sad, and L. Krishnan. 2004. Activation 2649–2658. of dendritic cells by liposomes prepared from phosphatidylinositol mannosides from Mycobacterium bovis bacillus Calmette-Guerin and adjuvant activity in 28. Brennan, P. J., M. Heifets, and B. P. Ullom. 1982. Thin-layer chromatography of vivo. Infect. Immun. 72: 5235–5246. lipid antigens as a means of identifying non-tuberculous mycobacteria. J. Clin. 49. Sidobre, S., G. Puzo, and M. Riviere. 2002. Lipid-restricted recognition of my- Microbiol. 15: 447–455. cobacterial lipoglycans by human pulmonary surfactant protein A: a surface- 29. Kang, B. K., and L. S. Schlesinger. 1998. Characterization of mannose receptor- plasmon-resonance study. Biochem. J. 365: 89–97. dependent phagocytosis mediated by Mycobacterium tuberculosis lipoarabino- 50. Villeneuve, C., G. Etienne, V. Abadie, H. Montrozier, C. Bordier, F. Laval, mannan. Infect. Immun. 66: 2769–2777. M. Daffe, I. Maridonneau-Parini, and C. Astarie-Dequeker. 2003. Surface-ex- 30. Schlesinger, L. S. 1993. Macrophage phagocytosis of virulent but not attenuated posed glycopeptidolipids of Mycobacterium smegmatis specifically inhibit the strains of Mycobacterium tuberculosis is mediated by mannose receptors in ad- phagocytosis of mycobacteria by human macrophages: identification of a novel dition to complement receptors. J. Immunol. 150: 2920–2930. family of glycopeptidolipids. J. Biol. Chem. 278: 51291–51300. 31. Munford, R. S., and C. L. Hall. 1986. Detoxification of bacterial lipopolysaccha- 51. Chieppa, M., G. Bianchi, A. Doni, A. Del Prete, M. Sironi, G. Laskarin, P. Monti, rides (endotoxins) by a human . Science 234: 203–205. L. Piemonti, A. Biondi, A. Mantovani, et al. 2003. Cross-linking of the mannose 32. Teghanemt, A., D. Zhang, E. N. Levis, J. P. Weiss, and T. L. Gioannini. 2005. receptor on monocyte-derived dendritic cells activates an anti-inflammatory im- Molecular basis of reduced potency of underacylated endotoxins. J. Immunol. munosuppressive program. J. Immunol. 171: 4552–4560. 175: 4669–4676. 52. Helenius, A., and M. Aebi. 2001. Intracellular functions of N-linked . 33. Somerville, J. E., Jr., L. Cassiano, B. Bainbridge, M. D. Cunningham, and Science 291: 2364–2369. R. P. Darveau. 1996. A novel Escherichia coli lipid A mutant that produces an 53. Martinez-Pomares, L., S. A. Linehan, P. R. Taylor, and S. Gordon. 2001. Binding antiinflammatory lipopolysaccharide. J. Clin. Invest. 97: 359–365. properties of the mannose receptor. Immunobiology. 204: 527–535.