Candida albicans phospholipomannan promotes survival of phagocytosed yeasts through modulation of bad phosphorylation and macrophage apoptosis.

The surface of the pathogenic yeast Candida albicans is coated with phospholipomannan (PLM), a phylogenetically unique glycolipid composed of beta-1,2-oligomannosides and phytoceramide. This study compared the specific contribution of PLM to the modulation of signaling pathways linked to the survival of C. albicans in macrophages in contrast to Saccharomyces cerevisiae. C. albicans endocytosis by J774 and disregulation of the ERK1/2 signal transduction pathway was associated downstream with a reduction in Bad Ser-112 phosphorylation and disappearance of free Bcl-2. This suggested an apoptotic effect, which was confirmed by staining of phosphatidylserine in the macrophage outer membrane. The addition of PLM to macrophages incubated with S. cerevisiae mimicked each of the disregulation steps observed with C. albicans and promoted the survival of S. cerevisiae. Externalization of membranous phosphatidylserine, loss of mitochondrial integrity, and DNA fragmentation induced by PLM showed that this molecule promoted yeast survival by inducing host cell death. These findings suggest strongly that PLM is a virulence attribute of C. albicans and that elucidation of the relationship between structure and apoptotic activity is an innovative field of research.

Candida albicans is part of the normal microbial flora that colonizes the mucocutaneous surfaces of the oral cavity, gastrointestinal tract, and vagina. The high levels of morbidity and mortality induced by C. albicans in hospitalized patients mean that this species is now one of the most prominent human pathogens (1). Research is currently under way to identify the virulence attributes of C. albicans (2) that explain the success of this species as a human pathogen.
It has previously been shown that C. albicans, in contrast to the closely related but nonpathogenic yeast, Saccharomyces cerevisiae, was able to survive within macrophages (3). Endocytosis of C. albicans by macrophages was specifically associ-ated with the reduced phosphorylation of ERK1/2 and the downstream product, p90 rsk , through activation of a specific phosphatase, MKP-1 (3). Both ERK1/2 and p90 rsk have been shown to regulate survival of different cells (4Ϫ6). They are involved in maintaining the phosphorylated state of Bad, a proapoptotic member of the Bcl-2 family that plays an important role in mediating signal transduction pathways leading to apoptosis. Bad function is regulated by phosphorylation at two sites, serine 112 (Ser-112) and serine 136 (Ser-136) (7). Whereas Ser-136 phosphorylation is associated with activation of Akt, Ser-112 phosphorylation requires activation of the mitogen-activated protein kinase pathway (8). Phosphorylated Bad is held by the 14-3-3 protein, freeing Bcl-x(L) and Bcl-2 to promote survival (7). Unphosphorylated Bad dissociates from 14-3-3 and recruits Bcl-2 and Bcl-x(L) to initiate events that lead to mitochondrial dysfunction and caspase activation (9). Intracellular pathogens have evolved diverse strategies to induce (10 -13) or inhibit host cell apoptosis (14 -16), aiding dissemination within the host or facilitating intracellular survival (17,18). Host cell apoptosis is induced through different virulence mechanisms based on either surface glycolipids (15,19,20) or type III secretion proteins (21,22). In host tissues, C. albicans may be both intra-and extracellular (23). Macrophages undergo apoptotic cell death after infection with C. albicans strains capable of hyphal formation (24), and activation of caspase 3 has been observed after endocytosis of C. albicans by neutrophils (25).
It has recently been demonstrated that the cell wall surface of C. albicans is coated with a phylogenetically unique molecule, phospholipomannan (PLM), 1 composed of oligomannose residues with a unusual type of linkage and phytoceramide (26,27) (Fig. 1A). PLM is shed by C. albicans in contact with macrophages (28) and displays potent activity on the innate immune response (29).
In this study, the modulation of signals downstream from ERK1/2 and p90 rsk was investigated, with special attention to the regulation of cell survival induced after endocytosis of C. albicans. The participation of C. albicans PLM as an inducer of apoptosis was then determined using highly purified and well characterized PLM. The addition of PLM to cells was found to allow survival of the sensitive yeast S. cerevisiae. This was associated with down-regulation of ERK1/2-dependent Bad Ser-112 phosphorylation in host cells and alteration of cell integrity leading to apoptosis.
Cell Culture-The mouse macrophage-like cell line, J774 (ECACC 85011428), was derived from a tumor of a female BALB/c mouse. Adherent cells were cultured at 37°C in an atmosphere containing 5% CO 2 in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% heat-inactivated fetal calf serum (Valbiotech, Paris, France), 5 mM L-glutamine, 100 g/ml streptomycin, and 50 g/ml penicillin. Before use, cells were gently scraped off with a rubber policeman and, depending on the experiment, either plated into eight-well Labtek tissue culture chambers (Nunc, Naperville, IL) at a concentration of 0.5 ϫ 10 6 cells/well for microscopic analysis or into 12-well tissue culture dishes at a concentration of 10 6 cells/well in 1 ml of culture medium (for biochemical analysis).
Yeast Culture and PLM Purification-C. albicans VW32 (serotype A) and S. cerevisiae Su1 (3) were used throughout this study. Yeasts were maintained on Sabouraud's dextrose agar (SDA) at 4°C. Before the experiments, yeast cells were transferred onto fresh SDA and incubated for 20 h at 37°C. Yeast cells were then recovered, washed with phosphate-buffered saline, and transferred into DMEM. Heat-killed yeasts were prepared by heating 20 ϫ 10 6 yeasts/ml in sterile water at 95°C for 15 min. Efficiency of killing was determined by culture of treated yeasts on SDA for 48 h at 37°C. The presence of PLM at the cell wall surface of yeasts was examined by Western blot using the anti-␤-1,2 oligomannoside monoclonal antibody AF1 as described previously (28).
PLM from C. albicans was prepared by extensive purification partition and hydrophobic interaction steps as described previously (26). The structure of this molecule was determined by a combination of methanolysis/HPLC, phosphorus/proton NMR, and ion spray mass spectrometry methods and is shown in Fig. 1A. This study involved the batch of C. albicans PLM recovered from these structural studies after analysis by nondenaturing methods.
Co-culture of Yeast Cells with Mammalian Cells and PLM Stimulation-J774 cells were gently scraped with a rubber policeman and distributed into 12-well culture plates at a concentration of 10 6 cells/ well. After 18 h, the adherent cells were washed with culture medium. For co-cultivation studies, plated cells were incubated with yeasts at a concentration of 20 yeasts/J774 cell. After incubation for various times, the cultures were washed with DMEM to remove unbound yeast cells and prepared for either biochemical analysis or fungicidal assays. In some experiments, cells were incubated with different concentrations of PLM in culture medium for 1 h before the addition of yeast cells. For assessment of the effect of PLM on macrophage responses, different concentrations of PLM were added to plated J774 cells and incubated for various periods of time before preparation for either biochemical analysis or fungicidal assays.
Extracted proteins were separated by 10% SDS-PAGE before blotting onto a nitrocellulose membrane (Schleicher and Schuell) for 2 h at 200 mA in a semidry transfer system. After staining with 0.1% Ponceau S in 5% acetic acid to confirm equivalence of loading and transfer, the membrane was blocked by incubation with TNT (10 mM Tris, 100 mM NaCl, 0.1% Tween) containing 5% nonfat dry milk for 1 h at 20°C. Membranes were probed with phosphospecific antibodies (diluted 1:1000) or anti-Bcl-2 IgG (diluted 1:250) in TNT-5% bovine serum albumin overnight at 4°C. After washing several times, the membranes were incubated for 1 h at 20°C with a 1:2000 dilution of horseradish peroxidase-conjugated anti-rabbit IgG in TNT-5% bovine serum albumin. After washing, the membrane was incubated with ECL detection reagents (SuperSignal Chemiluminescent Substrate) (Pierce) and exposed to hyperfilm ECL.
Immunofluorescence Analysis-After stimulation, J774 cells were washed with warm DMEM and fixed and permeabilized with 3.7% formaldehyde and 0.2% Triton X-100 in phosphate buffer at 20°C for 20 min. After three washes with phosphate buffer, 100 l of a 1:100 dilution of anti-Bcl-2 antibody was added for 2 h at 20°C. After five washes, 100 l of a 1:100 dilution of fluorescein isothiocyanate-conjugated goat anti-rabbit IgG in phosphate buffer was added for 1 h at 20°C. Slides were then washed five times and mounted for microscopic examination.
Assessment of Apoptosis by Fluorescence Microscopy-Detection of apoptotic and necrotic cells after incubation with yeasts or PLM was performed by fluorescence microscopy with the annexin V-propidium iodide apoptosis detection kit (R&D Systems, Minneapolis, MN) on unfixed cells as recommended by the manufacturer.
DNA Fragmentation Assay-J774 cells (2 ϫ 10 6 ) were collected after incubation at 37°C for 16 h with PLM and centrifuged at 400 ϫ g for 10 min at 20°C. Pelleted cells were incubated for 1 h at 50°C in hypotonic lysis buffer (10 mM Tris-HCl, pH 7.4, 10 mM EDTA, 1% Triton-X 100, 100 g/ml proteinase K). The lysates were clarified by centrifugation at 400 ϫ g for 30 min at 20°C, and the supernatants were deproteinized twice by phenol-chloroform (1:1) extraction and once by chloroform extraction. The mixtures were treated with 50% isopropyl alcohol, 0.5 M NaCl at Ϫ20°C overnight for DNA precipitation. Precipitates were pelleted, washed with 70% ethanol, air-dried, and reconstituted with 20 l of TE buffer (10 mM Tris, 1 mM EDTA, pH 7.4). Aliquots (10 l) were applied to horizontal agarose gels (2%) and subjected to a standard electrophoresis procedure. Gels were stained with 5 g/ml ethidium bromide and photographed under UV light.
Flow Cytometry Analysis-Mitochondrial alteration during apoptosis was examined in J774 cells after incubation for 120 min without or with 10 g/ml PLM. Treated cells were incubated for 15 min with 10 M YOPRO-1 (Molecular Probes, Inc., Eugene, OR), a vital fluorescent DNA dye that allows visualization of the alterations to the plasma membrane (30), and with 100 nM chloromethyl-X-rosamine (CMX-Ros) to reveal the integrity of the mitochondrial transmembrane potential (31). Stained cells were analyzed by flow cytometry (Coulter; Beckman).
Fungicidal Assays-J774 cells were incubated for 30 min at 37°C with yeast cells in the presence or absence of PLM, washed with DMEM, and then recultured at 37°C for a further 90 min. The cultures were washed with DMEM, and endocytosed yeast cells were released by lysing the J774 cells with sterile water. The yeast cells recovered were counted, and 100 individual yeast cells in 1 ml of phosphate-buffered saline were plated onto SDA. After incubation for 24 h, the number of colony-forming units was determined.
Densitometry-Autoradiograms were scanned, and densitometry analysis was performed using the public domain NIH Image program (developed at the National Institutes of Health and available on the World Wide Web at rbs.info.nih.gov/nih-image/).

C. albicans Endocytosis by J774 Cells Is Associated with
Down-regulation of Bad Ser-112 Phosphorylation-It has previously been shown that endocytosis of C. albicans by J774 cells is associated with modulation of the MEK-ERK signal transduction pathway, leading to decreased phosphorylation of ERK1/2 and its downstream product, p90 rsk (3). Phosphorylated p90 rsk targets a Ser residue at position 112 of Bad in contrast to Akt-1/protein kinase B, which phosphorylates Bad at Ser-136 (8). As ERK1/2 and p90 rsk phosphorylation were altered after ingestion of C. albicans by J774 cells, the effect of yeast engulfment on Bad phosphorylation was investigated. Fig. 2 shows that after incubation of J774 cells with S. cerevisiae blastoconidia, p90 rsk phosphorylation was associated with high levels of phosphorylation of Bad at both Ser-112 and Ser-136. In cells incubated for 15 min with C. albicans, a similar degree of phosphorylation of p90 rsk and Bad at both Ser-112 and Ser-136 was observed. However, after a 60-min incubation with C. albicans, cells displayed a dramatic decrease in phosphorylation of p90 rsk and a simultaneous and significantly lower phosphorylation of Bad, specifically at Ser-112.
Dephosphorylation of Bad at Ser-112 by C. albicans Coincides with the Disappearance of Bcl-2 Staining in J774 Cells-Phosphorylation of Bad at Ser-112 has been shown to be critical for its binding to 14-3-3 (8). The absence of such phosphorylation frees Bad, which may complex with the antiapoptotic protein Bcl-2. The effect of endocytosis of C. albicans on the distribution of Bcl-2 was therefore investigated in an immunofluorescence assay. A similar distribution of Bcl-2 was observed in untreated J774 cells (Fig. 3A) and in cells that had ingested S. cerevisiae blastoconidia (Fig. 3B). In contrast, a strong decrease in Bcl-2 staining was observed in cells that had ingested C. albicans yeasts (Fig. 3C). Analysis by Western blotting, to detect both free and complexed protein, revealed similar levels of Bcl-2 expression in cell extracts (Fig. 4). Together, these data show that the disappearance of Bcl-2 observed in the immunofluorescence assay could not be related to down-modulation of protein expression and suggest the formation of heterodimeric complexes between Bcl-2 and unphosphorylated Bad.
C. albicans-induced Proapoptotic Signal in J774 Cells-Bad dephosphorylation, together with the disappearance of Bcl-2 staining in cells which had ingested C. albicans, suggest that the cells underwent apoptosis. The effect of ingestion of C. albicans (either alive or heat killed) and S. cerevisiae on the expression of phosphatidylserine at the outer leaflet of macrophages was investigated using propidium iodine staining to discriminate between necrosis and apoptosis. Staining with propidium iodine was similar irrespective of how the cells had been treated, showing that the cells were not affected by necrosis. No binding of annexin V was observed with control cells (Fig. 5A) or with cells that had ingested S. cerevisiae (Fig. 5B). In contrast, binding of annexin V to cells that had ingested live C. albicans yeast cells, revealed that phosphatidylserine was exposed at the plasma membrane of these cells (Fig. 5C). When killed yeasts were used, no annexin V binding at the J774 cell membrane could be observed (Fig. 5D).
PLM Induces Disregulation of the ERK1/2 Signal Transduction Pathway-Presence of PLM at the surface of C. albicans yeast cells is a characteristic of this pathogenic yeast, which makes it different from other strains of Candida or from S. cerevisiae (26,27). Comparison of the presence of PLM at the surface of live and heat-killed C. albicans (Fig. 6) revealed that the molecule was weakly detected in extracts from heat-killed yeasts compared with live yeast cells. We thus hypothesized that PLM could play a role in the alteration of the J774 cell response seen with C. albicans whole live yeast cells. The effect of PLM on signal transduction was examined by incubating cells with PLM before the addition of either S. cerevisiae or C. albicans blastoconidia. As shown in Fig. 7A, the addition of PLM to J774 cells led to a decreased capability of the cells to phosphorylate ERK1/2 in response to ingestion of S. cerevisiae. This inhibition depended on the dose of PLM used, a 10 g/ml concentration of PLM being sufficient to obtain 40% inhibition of the signal. Incubation of J774 cells with PLM before the addition of C. albicans also led to the extinction of ERK1/2 phosphorylation when the highest concentration of PLM was used (50 g/ml). The influence of PLM on the ERK1/2 pathway and downstream products was confirmed by the decreased phosphorylation of p90 rsk observed in cells pretreated with 50 g/ml PLM before the addition of yeasts (Fig. 7B).
PLM Induces Apoptotic Signals in J774 Cells after Ingestion of the Sensitive Yeast S. cerevisiae and Allows Yeasts to Escape the Lytic Activity of Cells-The ability of PLM to alter the signaling response induced in cells during endocytosis and its effect on the capability of treated cells to kill endocytosed yeasts were then investigated. As shown in Fig. 8, treatment of J774 cells with PLM before the addition of yeasts led to a decreased capability of J774 cells to kill S. cerevisiae. The inhibitory effect of PLM depended on the dose used for pretreating the cells (Fig. 8A) and was maximal with 50 g/ml PLM. With this dose, 28 Ϯ 2% of blastoconidia that had been ingested by the cells survived compared with 5 Ϯ 5% of blastoconidia that survived after ingestion by untreated control cells (Fig. 8B).
The effect of PLM treatment on Bcl-2 staining after ingestion of S. cerevisiae by J774 cells was then determined (Fig. 9). As shown above (Fig. 3), cells that had ingested S. cerevisiae presented strong Bcl-2 immunostaining (Fig. 9A). Treatment with PLM before the addition of yeasts led to reduced, if any, staining of the protein (Fig. 9B), which was comparable with that observed after endocytosis of C. albicans blastoconidia (Fig. 3).
PLM Induces Apoptosis of J774 Cells through a Signal That Affects Mitochondrial Integrity-The direct effect of PLM on J774 cell apoptosis was then determined. J774 cells were incubated with PLM, and evidence of apoptosis was sought by different methods. The effect of incubation with PLM on cell membrane permeability and mitochondrial potential was examined by flow cytometry (Fig. 10). In the absence of PLM treatment, 86.3% of cells were negative for YOPRO-1, revealing plasma membrane integrity of the cells (Fig. 10A), and positive by CMX-Ros staining, revealing intact mitochondrial potential (Fig. 10B). After treatment with PLM, although most cells were unlabeled with YOPRO-1 (Fig. 10C), more than 34% of cells presented negative staining with CMX-Ros, showing that the mitochondria were altered by the treatment (Fig. 10D).
The apoptotic process induced by PLM was confirmed using annexin V to reveal phosphatidylserine expression on the external side of the treated cell membrane. Compared with untreated cells (Fig. 11, A and C), binding of annexin V to the cell membrane was most obvious after incubation of the cells with PLM (Fig. 11, B and D).
Apoptosis of cells after incubation with PLM was also confirmed by the DNA fragmentation observed in these cells, which was related to the dose of PLM used (Fig. 12). DISCUSSION Several pathogens are known to interfere with host cell apoptotic control (32). Depending on their parasitic behavior, differential triggering of cell survival or cell death is used by microbes to promote their survival. The yeast C. albicans has been reported to inhibit tumor necrosis factor-␣-induced DNA fragmentation in macrophages (33) and to induce apoptosis of macrophages (24) and neutrophils (25), but the mechanisms and C. albicans molecules involved are unknown. It has previously been shown that, in contrast to the nonpathogenic yeast S. cerevisiae, C. albicans was able to specifically alter signal transduction of the MEK-ERK pathway (3). This down-modulation of the cell machinery decreased phosphorylation of p90 rsk , resulting in survival of ingested yeast cells. In this study, the alteration of ERK1/2 and p90 rsk phosphorylation was observed to end with the specific dephosphorylation of Bad at Ser-112. Dephosphorylation of Bad, which allows it to complex with anti-apoptotic Bcl-2 homologues, is known to induce mitochondrial events leading to apoptosis (9,35).
In our C. albicans model, heat-killed yeasts did not induce apoptosis. However, due to PLM hydrophylicity (conferred by its glycan moiety), this molecule is not retained within the cell wall of dead cells (see Fig. 6); conventional heat treatment resulted in a dramatic decrease in the PLM associated with heat-killed yeasts placed in contact with macrophages. This prompted us to consider PLM, rather than any other mechanism triggered by living yeasts, as the molecule responsible for the proapoptotic activity of C. albicans on macrophages. The PLM proapoptotic effect was strong enough to promote the survival of the sensitive yeast S. cerevisiae when the C. albicans-derived molecule PLM was added together with S. cerevisiae to macrophages (Fig. 1B).
Extensive literature exists on the effects of surface glycolipids from pathogens on the control of host cell apoptosis (36). The lipoarabinomannan from the prokaryotic intracellular pathogen Mycobacteria tuberculosis promotes macrophage survival by mediating phosphorylation of Bad at Ser-136 in a phosphatidylinositol 3-kinase/Akt-dependent manner (15). The lipophosphoglycan from Leishmania donovani, which is a glycolipid present at the cell surface of an eukaryotic intracellular pathogen belonging to the family Trypanosomatideaea, also promotes host cell survival (19). However, depending on the structure of the glycolipid, opposite effects may also be observed. The glycosylinositol phospholipid from Trypanosoma cruzi, another member of the Trypanosomatideaea with an extracellular parasitic phase, induces host cell apoptosis (20). However, in contrast to lipophosphoglycan, the lipid core of glycosylinositol phospholipid belongs to a family of ceramides whose proapoptotic effects are well known (37,38).
C. albicans PLM is expressed at the cell wall surface of live yeasts and binds to and stimulates macrophages through a process of shedding from the yeast cell wall initiated by contact (26 -29, 39, 40). Recent structural investigations have shown that the PLM glycan moiety composed of long linear chains of ␤-1,2-oligomannosides (40) coupled by a specific arm to phytoceramide, is derived from the mannose inositol phosphoceramide biosynthetic pathway (26). Although ␤-1,2oligomannosides per se have been demonstrated to act as adhesins (41,42) and stimulate macrophages (43, 44), the proapoptotic properties of PLM are more likely to be related to its lipid core (38,45).
Ceramides have been shown to interfere either with the ERK1/2 signaling pathway (46) or the PI-3k and Akt/protein kinase B pathway, which regulates Bad activity at Ser-136 (47,48). Bad dephosphorylation at Ser-136 by prolonged inactivation of Akt/protein kinase B (49) has been described as ceramide-induced apoptosis in Bad-expressing COS-7 cells through a mechanism involving Ras-MEK1. In this study, no alteration of phosphorylation of Ser-136 was observed. This argues against a role for Akt/protein kinase B in C. albicans-induced interference. This was confirmed by treatment of cells with wortmanin, a well known inhibitor of PI-3k and consequently of Akt/protein kinase B and Bad phosphorylation at Ser-136 (50). This treatment did not change the viability of yeasts after ingestion by J774 cells (data not shown). The hypothesis that the PLM ceramide moiety may be responsible for apoptosis is nonetheless coherent with the observed transduction pathways, since (i) ceramides have been shown to accelerate dephosphorylation of ERK1/2 (46) and to inhibit cell growth through a mechanism involving protein kinase C-⑀ (51) and (ii) the down-regulation of ERK1/2 previously observed after endocytosis of C. albicans was related to prolonged activation of MKP-1, a MEK-ERK-specific phosphatase whose expression also depends on n-protein kinase C activation (3). However, the mechanisms by which ceramide may induce apoptosis are multiple. It has recently been shown that ceramide may induce mitochondrial activation through Bax, a proapoptotic member of the Bcl-2 family closely related to Bad. This apoptosis was independent of caspase activation (52). Whether PLM-induced apoptosis depends on caspase activation remains to be determined. Endocytosis of Escherichia coli by macrophages has recently been described to activate the proapoptotic signal (53). In this case, signaling was initiated by Toll-like receptor-2 and involved the adaptor, myeloid differentiation factor 88, and the caspase-9/-3-dependent mitochondrial amplification loop. Myeloid differentiation factor 88 may also directly recruit caspase-8 (54). Interestingly, both the mitochondrial amplification loop and caspase-8 have recently been shown to be modulated by ERK phosphorylation (55). These differential mechanisms by which ceramide influences the balance between cell survival and cell death are probably closely related to their structure. In this respect, the phytoceramide nature of the C. albicans PLM lipid moiety raises an important question (26). Although our knowledge of phytoceramide and phytosphingosine synthesis or hydrolysis is still embryonic, the presence of these molecules in mammalian cells has been demonstrated, and several studies have suggested that, like ceramide, phytosphingosine may play an important role in cell growth or death of eukaryotic cells (56,57). Due to their critical role in cell integrity, it is interesting to note that ceramide biosynthetic pathways are also critical for the establishment of fungal pathogens in the host. For the intracellular pathogen Cryptococcus neoformans, it has been recently shown that down-regulation of C. neoformans inositolphosphoryl phytoceramide synthase (an early step in mannose inositol phosphoceramide synthesis) significantly decreased intracellular growth and resistance of this yeast in J774 macrophages (34).
In conclusion, PLM was shown to cause by itself an alteration to host cells, which resulted in the cells undergoing apoptosis, as revealed by mitochondrial potential alteration, phosphatidylserine exposure at the plasma membrane, and DNA fragmentation. The apoptotic pathway induced is probably of pathophysiological significance, since killed yeasts did not display such activity and the addition of PLM allowed S. cerevisiae to survive to macrophage phagocytosis. The mech-anisms involved in induction of host cell death remain obscure. Although participation of other yeast factors cannot be excluded, the unique phylogenetic structure of PLM makes it a worthy candidate for further investigation.