Mycobacteria use their surface-exposed glycolipids to infect human macrophages through a receptor-dependent process

Two subfamilies of the polar glycopeptidolipids (GPLs) located on the surface of Mycobacterium smegmatis , along with unknown phospholipids, were recently shown to participate in the nonopsonic phagocytosis of mycobacteria by human macrophages (Villeneuve, C., G. Etienne, V. Abadie, H. Montrozier, C. Bordier, F. Laval, M. Daffe, I. MaridonneauParini, and C. Astarie-Dequeker. 2003. Surface-exposed glycopeptidolipids of Mycobacterium smegmatis specifically inhibit the phagocytosis of mycobacteria by human macrophages. Identification of a novel family of glycopeptidolipids. J. Biol. Chem. 278: 51291–51300). As demonstrated herein, a phospholipid mixture that derived from the methanol-insoluble fraction inhibited the phagocytosis of M. smegmatis . Inhibition was essentially attributable to phosphatidylinositol mannosides (PIMs), namely PIM 2 and PIM 6 , because the purified phosphatidylethanolamine, phosphatidylglycerol, and phosphatidylinositol were inactive. This was further confirmed using purified PIM 2 and PIM 6 from M. bovis BCG that decreased by half the internalization of M. smegmatis . Both compounds also inhibited the uptake of M. tuberculosis and M. avium but had no effect on the internalization of zymosan used as a control particle of the phagocytic process. When coated on latex beads, PIM 2 and polar GPL (GPL III) favored the particle entry through complement receptor 3. GPL III, but not PIM 2, also directed particle entry through the mannose receptor. Therefore, surface-exposed mycobacterial PIM and polar GPL participate in the receptordependent internalization of mycobacteria in human macrophages. —Villeneuve, C., M. Gilleron, I. Maridonneau-Parini, M. Daffé, C. Astarie-Dequeker, and G. Etienne. Mycobacteria use their surface-exposed glycolipids to infect human macrophages through a receptor-dependent process. J. Lipid Res. 2005. 46: 475–483. Supplementary key words mycobacterium • phospholipids • phosphatidylinositol mannosides • phagocytosis The outcome of mycobacterial infections is thought to be critically dependent on the nonopsonic invasion and colonization of macrophages during primary infection of inhaled bacteria. This is why the initial step of mycobacteria phagocytosis by macrophages received considerable attention, including efforts to delineate the molecular mechanisms that underlie this process. Nonopsonic internalization of mycobacteria in macrophages can be mediated by different kinds of receptors that specifically recognize ligands expressed at the surface of bacilli (1, 2). Glycophosphoinositol-anchored receptors such as CD14 (3, 4), scavenger receptor (5), and transferrin receptor (6, 7) have been suggested as putative routes of entry for mycobacteria into macrophages. Additionally, complement receptor 3 (CR3) (8–10) and mannose receptor (MR) (8, 11) are both well-known receptors for the binding and internalization of mycobacteria in macrophages. Both receptors exhibited lectin-like properties. Indeed, MR is a monomeric transmembrane protein with an extracellular domain containing eight carbohydrate recognition domains with C-type lectin properties that recognize mannose, fucose, and N -acetylglucosamine residues present on the cell surface of pathogens, such as mycobacteria (12–14). CR3 is a heterodimeric type 1 transmembrane glycoprotein consisting of a CD18 subunit noncovalently associated with the CD11b chain. The extracellular portion of CD11b contained distinct functional binding domains. One of these, called the I domain, is essential for the recognition and internalization of C3bi opAbbreviations: CR3, complement receptor 3; GPL, glycopeptidolipid; MALDI-TOF, matrix-assisted laser-desorption/ionization time of flight; Man-BSA, mannosylated bovine serum albumin; MDM, macrophages derived from monocytes; MR, mannose receptor; PE, phosphatidylethanolamine; PG, phosphatidylglycerol; PI, phosphatidylinositol; PIM, phosphatidylinositol mannoside. 1 To whom correspondence should be addressed. e-mail: catherine.astarie-dequeker@ipbs.fr (C.A-D.); gilles.etienne@ipbs.fr (G.E.) Manuscript received 13 August 2004 and in revised form 4 November 2004. Published, JLR Papers in Press, December 1, 2004. DOI 10.1194/jlr.M400308-JLR200 by gest, on N ovem er 7, 2017 w w w .j.org D ow nladed fom

The outcome of mycobacterial infections is thought to be critically dependent on the nonopsonic invasion and colonization of macrophages during primary infection of inhaled bacteria. This is why the initial step of mycobacteria phagocytosis by macrophages received considerable attention, including efforts to delineate the molecular mechanisms that underlie this process.
Nonopsonic internalization of mycobacteria in macrophages can be mediated by different kinds of receptors that specifically recognize ligands expressed at the surface of bacilli (1,2). Glycophosphoinositol-anchored receptors such as CD14 (3,4), scavenger receptor (5), and transferrin receptor (6,7) have been suggested as putative routes of entry for mycobacteria into macrophages. Additionally, complement receptor 3 (CR3) (8)(9)(10) and mannose receptor (MR) (8,11) are both well-known receptors for the binding and internalization of mycobacteria in macrophages. Both receptors exhibited lectin-like properties. Indeed, MR is a monomeric transmembrane protein with an extracellular domain containing eight carbohydrate recognition domains with C-type lectin properties that recognize mannose, fucose, and N -acetylglucosamine residues present on the cell surface of pathogens, such as mycobacteria (12)(13)(14). CR3 is a heterodimeric type 1 transmembrane glycoprotein consisting of a CD18 ␤ subunit noncovalently associated with the CD11b ␣ chain. The extracellular portion of CD11b contained distinct functional binding domains. One of these, called the I domain, is essential for the recognition and internalization of C3bi op-sonized particles (15,16). Another one, located on the carboxyl-terminal end of the extracellular portion of CD11b, is a nonopsonic binding site with lectin properties (17). This latter domain is mainly involved in phagocytosis of particles containing ␤ -glucan, such as zymosan (18), but also interacts with mannose, glucose, and N -acetyl-d -glucosamine (17). More recently, a second nonopsonic binding domain distinct from the lectin one has been proposed to bind and internalize mycobacteria (19).
We recently demonstrated that mycobacteria express molecules at their surface that play a role in their internalization by human macrophages (20). Using a biological test of phagocytosis inhibition, we found that lipids, but not proteins, isolated from the outermost layer of Mycobacterium s megmatis by a mechanical treatment decrease the nonopsonic phagocytosis of mycobacteria by human macrophages without affecting the internalization of particles such as zymosan or latex beads (20). A detailed analysis of the lipid fractions has indicated that the major inhibitory compounds belong to the glycopeptidolipid (GPL) and phospholipid families (20). When the former family was chemically and structurally analyzed, several inhibitory C-type GPLs were identified, among which a novel class of succinylated GPLs ( Fig. 1 ) passed unnoticed until now (20). We have also proposed that their capacity to inhibit the phagocytosis of mycobacteria is dependent on common structural features. Identification of these molecules and their putative role as surface-exposed ligands of phagocytic receptors would help us understand the host-pathogen interactions at the molecular level. The present study was designed first to identify the mycobacteria surfaceexposed phospholipids that play a role in their internalization by human macrophages and second to examine the roles of GPLs and phospholipids as potential ligands of phagocytic receptors.

Mycobacteria
M. smegmatis (ATCC 607) and M. tuberculosis H37Rv (ATCC 27294) were grown as surface pellicles on liquid Sauton's medium at 37 Њ C without agitation; M. avium (serovar 4), known to form more clumps when grown as surface pellicles than the two former species, was cultured under shaking at 250 rpm in Middlebroock 7H9 medium supplemented with 10% albumin-dextrose-catalase enrichment (Difco). It is important to point out that the qualitative composition of the surface-exposed materials is independent of growth conditions (i.e., surface pellicles or shaken cultures) (21). Single cell suspensions were prepared with late log-phase cultures (6 days for M. smegmatis , 3 weeks for M. tuberculosis and M. avium ) as previously described (20). Briefly, pellicles ( M. smegmatis and M. tuberculosis ) were harvested by pouring off the medium, dispersed by gentle shaking for 30 s with 5 g of glass beads (4 mm diameter), and resuspended in PBS, pH 7.4. In the case of M. avium , the culture was recovered by centrifugation at 10,000 g for 10 min. To remove the remaining clumps, the bacterial suspensions were centrifuged for 10 min at 200 g . Up to 90% of mycobacteria were individualized, and their viability averaged 85%. When required, mycobacteria were labeled with FITC, as previously described (22). We checked that their rate of internalization and their route of entry within macrophages derived from monocytes (MDM) were identical to those obtained by green fluorescent protein-expressing M. smegmatis (data not shown), ruling out a side effect of FITC labeling.

Extraction and purification of glycolipids
Glycolipids and phospholipids are present in both the outermost layer of the cell envelope and whole mycobacteria (21,23). Therefore, the present study was performed on lipid extract from M. smegmatis whole cells to get enough material for further fractionation experiments (20). Mycobacterial wet cells were extracted with CHCl 3 /CH 3 OH (1:2, v/v) and then at last three times with CHCl 3 /CH 3 OH (1:1, v/v). The organic phases were pooled, extensively washed with water, and evaporated to dryness. Lipids were then resuspended in a minimal volume of chloroform and precipitated with trickling methanol. After standing for 2 h at 4 Њ C, methanol-insoluble lipids were recovered by centrifugation at 4 Њ C for 20 min (8,000 g ). Lipids were further separated using an anion-exchange QMA-silica gel (Chromabond SB; Macherey-Nagel). Phospholipids were eluted using a stepwise gradient: 0.1 M ammonium acetate in 33% methanol in chloroform (fractions 9 and 10), 0.2 M ammonium acetate in 33% methanol in chloroform (fractions 11-13), and 0.2 M ammonium acetate in methanol (fraction 14). Purification of phosphatidylethanolamine (PE) and phosphatidylglycerol (PG) was achieved by preparative TLC. Plates were developed in CHCl 3 /CH 3 OH/H 2 O (56:38:10, v/v/v); silica bands were scraped and extracted three times in the same solvent, adding trace amounts of acetic acid.
The purification of GPL has been described in detail (20). Briefly, the methanol-soluble lipids were chromatographed on a Florisil column irrigated with chloroform and then with a stepwise gradient of increasing concentrations of methanol and water in chloroform. GPL I was eluted with CHCl 3 /CH 3 OH (90:10, v/v), whereas GPL III was eluted in more polar fractions with a mixture of CHCl 3 /CH 3 OH/H 2 O (65:25:4, v/v/v); the latter were further separated from contaminating phospholipids using an anion-exchange QMA-silica gel (Chromabond SB; Macherey-Nagel).
All of the purification steps were monitored by TLC on silica gel 60-precoated plates (0.25 mm thickness; Merck) developed with CHCl 3 /CH 3 OH/H 2 O (60: 35:8) for the phospholipid-containing fractions. Sugar-containing compounds were visualized by spraying plates with 1.5% orcinol in 30% aqueous sulfuric acid, followed by heating, whereas the Dittmer-Lester reagent and ninhydrin were used to detect phosphorus-and amino group-containing substances. Phosphatidylinositol mannosides type 2 and 6 (PIM 2 and PIM 6 ) ( Fig. 1) were purified on a QMA-Spherosil M (BioSepra) column as previously described (24).

Mass spectrometry
Matrix-assisted laser-desorption/ionization time of flight (MALDI-TOF) mass spectrometry analysis of lipids was performed as previously described (25). MALDI-TOF spectra were acquired on a Voyager-DE STR mass spectrometer (PerSeptive Biosystems) equipped with a pulsed nitrogen laser emitting at 337 nm and were analyzed in the Reflectron mode using an extraction delay time set at 100 ns and an accelerating voltage operating in negative ion mode of 20 kV.

Human MDM
Human peripheral blood monocytes were isolated as previously described (11) and cultured for 7 days on sterile glass coverslips on 24-well tissue culture plates (5 ϫ 10 5 cells/well) containing RPMI medium with 10% heat-inactivated fetal calf serum in 5% CO 2 . The culture medium was renewed on the third day. Before use, MDM were washed twice with fresh RPMI medium.

Coating of latex beads
Latex beads were coated as described by Schlesinger, Hull, and Kaufman (26). Briefly, 2 ϫ 10 8 polystyrene microspheres of 1 m diameter were washed twice (10,000 g , 10 min) in 0.05 M carbonate-bicarbonate buffer, pH 9.6. Beads were then incubated for 1 h at 37 Њ C under gentle agitation with either 50 g of various glycolipids, Man-BSA, or BSA in a reaction volume of 1 ml. The microspheres were washed twice, incubated in 5% BSA in PBS for 2 h at 37 Њ C, resuspended in 5% BSA in PBS, and stored at 4 Њ C until the experiments .

Infection of adherent macrophages and phagocytosis assay
When specified, MDM were first pretreated for 15 min at 37 Њ C with the purified lipid fractions, which were previously resuspended in sterile apyrogenic water and sonicated for 10 min, and then added with particles at the appropriate multiplicity of infection. At the end of infection, cells were washed twice with fresh medium to remove unbound particles. Phagocytosis of mycobacteria, zymosan, and latex beads was assessed by immunofluorescence as previously described (11). For each set of conditions, duplicate experiments were performed, and at least 100 cells per slide were counted by fluorescence microscopy. The percentage of phagocytic cells having ingested at least one particle or bacterium was calculated as follows: number of phagocytic cells/number of total cells ϫ 100. Data are expressed as the percentage of phagocytosis inhibition determined as follows: number of phagocytic cells in the control minus number of phagocytic cells under treated conditions/number of phagocytic cells in the control ϫ 100.

Statistics
Data are presented as means Ϯ SEM of the indicated number of experiments (n) performed in duplicate. The significance of the differences was determined by paired or unpaired Student's t -test.

Inhibitory activity of surface-exposed phospholipids on the phagocytosis of M. smegmatis
In the present study, we first attempted to identify the surface-exposed phospholipids implicated in the internalization of mycobacteria. For this purpose, lipid extract from M. smegmatis was first enriched in phospholipids by methanol precipitation. As depicted in Fig. 2 , the methanol-insoluble fraction, analyzed by TLC, contained several compounds: PIMs, in particular PIM 2 and PIM 6 ( Fig. 1) as the main subspecies, PI ( Fig. 2A), PG (Fig. 2B), and PE (Fig. 2B). Examination of the blocking effect of the methanol-insoluble fraction on phagocytosis was the approach chosen to determine the role of phospholipids in the phagocytic process of mycobacteria. MDM were pretreated for 15 min with the phospholipid-enriched extract at concentrations ranging from 10 pg to 100 g lipid/ml and then put into contact with M. smegmatis for 45 min. At the lowest tested concentration (i.e., 10 pg lipid/ml), treated and untreated macrophages equally internalized M. smegmatis (62 Ϯ 6% and 61 Ϯ 5%, respectively; n ϭ 2). In contrast, increasing lipid concentration to 10 ng/ml resulted in a 30% decrease in the uptake of M. smegmatis (from 62 Ϯ 6% to 45 Ϯ 1%; n ϭ 2). This inhibitory effect remained stable up to 100 g/ml.

Purification and identification of the active phospholipid fractions of M. smegmatis
To establish the nature of the active phospholipids, we individually tested the effect of the addition of 10 g/ml PI, PE, and PG on the phagocytosis of M. smegmatis . Preincubation of MDM for 15 min with 10 g/ml commercial-ized PI did not influence the internalization of M. smegmatis compared with untreated cells (52 Ϯ 7% vs. 50 Ϯ 5%, respectively; n ϭ 3). Similarly, 10 g/ml PE or PG, purified from mycobacteria cell extracts by preparative TLC, did not affect the phagocytosis of M. smegmatis compared with untreated macrophages (43 Ϯ 4% and 49 Ϯ 3%, respectively, vs. 50 Ϯ 5%; n ϭ 3). These data strongly suggested that the observed inhibitory activity of the phospholipid-rich methanol-insoluble fraction was essentially attributable to PIM. This hypothesis was further confirmed using PIM-enriched fractions obtained from a M. smegmatis lipid extract as follows: methanol-insoluble lipids were fractionated by ion-exchange chromatography into 14 fractions eluted with increasing concentrations of ammonium acetate in chloroform-methanol. TLC and MALDI-TOF analysis showed that the first 8 fractions contained traces, if any, of phospholipids (data not shown), whereas fractions 9-14 contained essentially PIM and PI ( Fig. 3A ). Therefore, these last six fractions were tested on phagocytosis of M. smegmatis at 10 ng lipid/ml, the lowest active concentration for the phospholipid mixture. As shown in Fig. 3B, the tested fractions exhibited an inhibitory effect of ‫ف‬ 50%. This set of experiments indicated that PIMs were the compounds responsible for the activity of the phospholipid mixture and therefore were the essential phospholipids involved in the phagocytic process of mycobacteria.

Inhibition of mycobacterial internalization by PIM 2 and PIM 6 subfamilies
PIM 2 and PIM 6 ( Fig. 1) represent the main mycobacterial glycophospholipids described to date (24,27,28). As expected, these two subclasses of PIM constituted the majority of glycophospholipids of M. smegmatis and were differently distributed in fractions 9-14 (Fig. 3A). PIM 2 and triacylated PIM 2 (Ac 1 PIM 2 ) were preferentially eluted in fractions 9-11. PIM 6 and triacylated PIM 6 (Ac 1 PIM 6 ) appeared mainly in fractions 12-14. However, we failed to properly separate PIM 2 from PIM 6 by the purification protocol used. This prompted us to use purified PIM 2 and PIM 6 subfamilies isolated from M. bovis BCG by Gilleron et al. (24), which are structurally identical to those of M. smegmatis (27). As depicted in Fig. 4A , both PIM 2 and PIM 6 subfamilies reduced phagocytosis of M. smegmatis . Their inhibitory activity increased from 100 pg to 10 ng of lipid per milliliter, PIM 6 being slightly more efficient than PIM 2 . A plateau of inhibition was then reached at 10 ng/ ml up to 10 g/ml. Importantly, 10 ng/ml PIM 2 and PIM 6 subfamilies did not affect zymosan internalization, used as a control of phagocytic process (33 Ϯ 9% and 36 Ϯ 11%, respectively, vs. 35 Ϯ 9%; n ϭ 3 and 4), indicating a specific inhibition of the MDM-particle interactions by structurally defined molecules. We then addressed the question of whether PIM 2 and PIM 6 interfered with the phagocytosis of the pathogenic mycobacteria (e.g., M. avium and M. tuberculosis ). When tested at 10 ng/ml, both compounds reduced the internalization of bacteria (Fig.  4B). However, micromolar concentrations of PIM were required to achieve an inhibition of phagocytosis of the pathogenic species (Fig. 4B) comparable to what was observed with nanomolar concentrations of M. smegmatis (Fig. 4A). Nevertheless, these observations clearly demonstrated that subfamilies of PIM are involved in the nonopsonic phagocytic process of mycobacteria.

Recognition of PIM and GPL by nonopsonic phagocytic receptors
Recently, we reported that subfamilies of another surface-exposed glycolipid of M. smegmatis , GPL (Fig.1), in-hibit the phagocytosis of mycobacteria (20). We have also demonstrated that the nonopsonic phagocytosis of both pathogenic and nonpathogenic mycobacteria by human phagocytes is mediated at least in part by two major receptors, CR3 and MR (11,29). Therefore, the ability of PIM and GPL to inhibit mycobacteria internalization by blocking their entry through these receptors was suspected and investigated. To this aim, latex beads were coated with purified PIM or GPL using the method previously described by Schlesinger, Hull, and Kaufman (26). As PIM 2 and . The various PIM species and their relative abundances were determined by matrix-assisted laser-desorption/ionization time of flight mass spectrometry in the negative mode. PI was present in each fraction. PIM 2 and the triacylated form of PIM 2 (Ac 1 PIM 2 ) were enriched in fractions 9-11 and to a lesser extent in fractions 11-14. PIM 6 and the triacylated form of PIM 6 (Ac 1 PIM 6 ) appeared in fractions 12-14. B: Macrophages derived from monocytes (MDM) were preincubated with fractions 9-14 used at 10 g lipid/ml for 15 min and then challenged with FITC-stained M. smegmatis for 45 min at 37ЊC (25 particles per cell). Phagocytosis was determined as the number of macrophages having engulfed at least one particle. Phagocytosis inhibition is expressed as the percentage effect compared with the control. Results are means Ϯ SEM of three independent experiments performed in duplicate. Significance of the treatment effect was assessed by paired Student's t-test. * P Ͻ 0.05, ** P Ͻ 0.01 compared with untreated conditions. PIM 6 displayed comparable activities on the phagocytosis of mycobacteria, PIM 2 , the quantitatively major constituent of PIM, was isolated from M. bovis BCG and used for the subsequent experiment as representative of the PIM family. Beads were also coated with either GPL III or GPL I from M. smegmatis, two classes of GPL that have been shown to exhibit different inhibitory activities on MDM (20).
The capacity of the coated beads to enter macrophages through MR or CR3 was then investigated using a competition approach with blocking antibodies directed against receptors or soluble ligands. Monoclonal antibody 2LPM was selected for its capacity to prevent both opsonic and nonopsonic internalization through CR3 (19,30,31). Serum-opsonized and nonopsonized zymosan were used as positive CR3 binding particles (19) and BSA-latex beads as a negative control. We did not observe any difference be-tween the internalization of PIM 2 -or GPL-and BSA-coated beads (data not shown). As expected, pretreatment of MDM with 2LPM (5 g/ml) for 30 min at 37ЊC significantly decreased the internalization of serum-opsonized and nonopsonized zymosan but did not affect the phagocytosis of BSA-coated beads (Fig. 5A). Such pretreatment also reduced by approximately one-third the entry of both GPL III-and PIM 2 -coated beads but not that of GPL I (Fig. 5A), indicating that entry of PIM 2 -and GPL III-coated beads, but not that of beads coated with GPL I, involves CR3.  The entry of beads through the MR was also evaluated using a competitive ligand, Man-BSA (11), or a goat antihuman MR antiserum selected for its capacity to block MR-dependent functions (32,33). Man-BSA-and BSAcoated latex beads were used as positive and negative controls of MR-dependent phagocytosis, respectively (11). In this set of experiments, GPL III-coated beads were internalized as efficiently as beads coated with Man-BSA (45 Ϯ 5% vs. 33 Ϯ 7%; n ϭ 3), whereas PIM 2 -coated beads were slightly better internalized (58 Ϯ 1% vs. 33 Ϯ 7%; n ϭ 3). As expected, pretreatment of MDM with anti-MR antiserum (1:50) or 300 g/ml Man-BSA for 30 min at 37ЊC dramatically reduced the internalization of Man-BSA-coated beads (Fig. 5B). Such pretreatments did not affect the phagocytosis of either GPL I-or PIM 2 -coated beads. In contrast, both treatments slightly but significantly decreased the phagocytosis of GPL III-coated beads (Fig. 5B), indicating that only GPL III-coated beads enter MDM through the MR. Taken together, these data demonstrate that both PIM 2 and GPL III participate to the internalization of mycobacteria and thereby are potential ligands of phagocytic receptors expressed at the surface of MDM.

DISCUSSION
The outermost cell envelope layer, also called a capsule in the case of pathogenic mycobacteria, represents a privileged interface between mycobacteria and host cells. Some of its components, such as glucan (34) and GPL (20), have been implicated in the interaction with macrophages. Other constituents of the mycobacterial cell envelope have also been shown to induce various host cell responses (35), but their localization at the cell surface is still a matter of debate. We show here that the outermost layer of M. smegmatis contained PIMs that also participate in the nonopsonic phagocytosis of mycobacteria. The PIMs, described by Ballou and colleagues (36) in the 1960s, are based on PI and are formed of one to six mannosyl residues attached to the inositol moiety of the phospholipid. The PIMs usually occur as a mixture of compounds differing from one another by the number of mannosyl residues and fatty acyl groups. The nonpathogenic strain M. smegmatis ATCC 607 has been described to contain mainly two PIM subfamilies, PIM 2 and PIM 6 (24,27), in agreement with the composition of the crude PIM fractions reported herein.
Although other forms of PIM have been reported in the literature (24), they were undetectable under our experimental conditions, probably because they occur in too small quantities. Therefore, the crude PIM-dependent inhibition of mycobacterial phagocytosis was essentially attributable to PIM 2 and PIM 6 . This was further confirmed using enriched PIM 2 and PIM 6 fractions isolated from M. bovis BCG. Nanomolar concentrations of PIM decreased the internalization of M. smegmatis by MDM, whereas much higher concentrations of lipids were required to affect M. avium or M. tuberculosis phagocytosis to the same extent. PIMs are found on the bacterial surface of both patho-genic and nonpathogenic mycobacteria (27,37). However, there are strain differences in the surface exposure of PIMs that are inversely correlated with the abundant expression of capsular polysaccharides composed of ␣-glucan that can partially mask capsular PIM and directly act on phagocytic receptor, such as CR3 (30). If we assume that PIMs, exposed at the surface of M. smegmatis, have a lower affinity for the phagocytic receptors than the outermost polysaccharides of pathogenic species (38), this may explain why different PIM concentrations are required to inhibit to a comparable extent the phagocytosis of M. smegmatis and M. avium or M. tuberculosis.
It has to be noted that the occurrence of a carbohydrate portion on phospholipid is critical for the inhibitory activity, because PI displayed no effect but PIM 2 did. However, surprisingly, the length of the oligomannosyl chain does not appear to be important for the activity, because PIM 2 and PIM 6 equally inhibited phagocytosis. Previous studies have reported that soya PI and, to a lesser extent, mycobacterial PI inhibited the nonopsonic binding of M. tuberculosis to murine macrophages (39,40). This is in apparent conflict with the data presented here. Such a discrepancy could be attributed to interspecies differences in the PI mode of action between murine and human macrophages. Our data, together with our previous report (20), collectively indicate that surface-exposed PIM and GPL participate in the internalization of mycobacteria by macrophages and, as such, are potential ligands of phagocytic receptors. It is possible, however, that other constituents of the mycobacterial envelope, such as glucan and unidentified molecules, could also be involved in the interactions of bacilli with macrophages, as illustrated by the variety of routes of entry of mycobacteria into host cells.
The molecular mechanisms by which surface-exposed GPLs and PIMs participate in the phagocytosis of mycobacteria deserve consideration. PIMs have been proposed to participate in the receptor-mediated uptake of mycobacteria (26,41). For instance, they have been shown to act as an adhesin, allowing the binding of M. tuberculosis and M. smegmatis to hamster ovary fibroblasts, either directly or after opsonization with mannose binding protein (41). Here, we demonstrate that coating latex beads with PIM 2 favors their internalization through a CR3-dependent process, probably through the lectin binding site of CR3, which might recognize the mannosyl moiety of PIM, as mentioned above. Like other mannosylated conjugates, PIMs have been suspected to be ligands of the MR. We addressed this question and show, for the first time, that PIMs, at least PIM 2 , do not in fact interact with this receptor in our cell model. As mentioned above, the length of the oligomannosyl chain was not critical for PIM inhibition of phagocytosis, suggesting that PIM 6 was not a partner of MR. The mode of action of GPL as a putative ligand of macrophages has been poorly investigated. We found that GPL III ensures the entry of particles through both CR3 and MR. Although all types of GPLs share the same peptidolipid core, GPL I, which is inactive, lacks the succinyl substituent that occurs in GPL III and is 3,4-O-acetylated on the 6-deoxytalosyl moiety (20). At present, these moieties have not been proposed among those described to be recognized by MR and CR3, and their putative binding to these receptors need to be further investigated. Nonetheless, we propose that surface-exposed PIMs and GPLs coated on latex beads are potential ligands of CR3 and/or MR, thereby mimicking part of the entry process of mycobacteria in macrophages. When used as soluble molecules, PIMs (present study) as well as GPLs (20) affect the internalization of mycobacteria but not that of zymosan, whereas the two processes are well known to involve both MR and CR3. Concerning CR3, the fact that zymosan and mycobacteria use distinct nonopsonic binding sites of the receptor (19) suggests that the site involved in zymosan internalization does not recognize PIMs and GPLs.
In conclusion, we propose that surface-exposed GPLs and PIMs interact with the nonopsonic receptors CR3 and/or MR involved in the phagocytosis of mycobacteria. Characterization of their binding properties is beyond the scope of this report, but it would clearly help to further investigate their signaling pathways. Understanding the mycobacterial strategy to infect macrophages without triggering bactericidal activity is still a challenge. Finally, PIMs and GPLs, as inhibitors of mycobacteria phagocytosis, may also help to design new pharmacological drugs for the control of macrophage infection.