José Ruiz-Herrera, Departamento de Ingeniería Genética, Unidad Irapuato, Centro de Investigación y de Estudios Avanzados del IPN, Apartado Postal 629, 36500 Irapuato, Gto, Mexico. Tel.: +52 462 6239600; fax: +52 462 6245849; e-mail: email@example.com
Candida albicans is one of the most important opportunistic pathogenic fungi. Weakening of the defense mechanisms of the host, and the ability of the microorganism to adapt to the environment prevailing in the host tissues, turn the fungus from a rather harmless saprophyte into an aggressive pathogen. The disease, candidiasis, ranges from light superficial infections to deep processes that endanger the life of the patient. In the establishment of the pathogenic process, the cell wall of C. albicans (as in other pathogenic fungi) plays an important role. It is the outer structure that protects the fungus from the host defense mechanisms and initiates the direct contact with the host cells by adhering to their surface. The wall also contains important antigens and other compounds that affect the homeostatic equilibrium of the host in favor of the parasite. In this review, we discuss our present knowledge of the structure of the cell wall of C. albicans, the synthesis of its different components, and the mechanisms involved in their organization to give rise to a coherent composite. Furthermore, special emphasis has been placed on two further aspects: how the composition and structure of C. albicans cell wall compare with those from other fungi, and establishing the role of some specific wall components in pathogenesis. From the data presented here, it becomes clear that the composition, structure and synthesis of the cell wall of C. albicans display both subtle and important differences with the wall of different saprophytic fungi, and that some of these differences are of utmost importance for its pathogenic behavior.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Invasive fungal infections are important causes of morbidity and mortality, mainly in immunodepressed and hospitalized patients. Among these, the infectious processes caused by Candida albicans have acquired an increasing importance. Different virulence factors are involved in C. albicans pathogenicity, but the role of the cell wall in the pathogenesis of this fungus cannot be overestimated. The wall is the structure that: (1) first comes into contact with host cells; (2) carries important antigenic determinants of the fungus; (3) is responsible for the adherence of the pathogen; and (4) establishes a cross-talk with the host, which depends on what has been referred to as the ‘glycan code’, which includes modifications in the chemical composition and linkages of the cell wall polysaccharides.
From these interactions, the outcome will be the development of a pathogenic state, or the mounting of a resistant reaction by the host (for a discussion, see Poulain & Jouault, 2004). As in fungi in general, the cell wall of C. albicans is a coherent structure, made by the ordered arrangement of its different components. Some of these are linked by covalent bonds, whereas others are retained in the wall by hydrogen bonds, salt-type associations, hydrophilic or hydrophobic interactions. This composite provides protection to the cell against physical, chemical and biological aggression, and is responsible for its morphology.
Microscopic observation of thin sections of fungal cells or isolated walls has revealed the existence of several layers in the wall, distinguished by their electron density in electron micrographs. Depending on the method of analysis, the wall of C. albicans appears to contain from four to eight layers (Poulain et al., 1978) (see Fig. 1a).
Mannoproteins are enriched in the outer surface of the C. albicans wall, including the bud scar (Horisberger & Clerc, 1988). Once considered to be absent in the inner layers, their presence has been demonstrated by different techniques, including labeling with ConA or antibodies (Marcilla et al., 1991; Kapteyn et al., 2000). In different fungi, including C. albicans, bona fide wall proteins (see below) are bound to the β-glucan/chitin inner layer through the lateral chains of β-1,6-glucan or, to a lesser extent, to β-1,3-glucan. The structural polysaccharides accumulate in the innermost layer of the cell wall, which appears to be less electron-dense. After extensive treatment of C. albicans blastoconidia (yeast-form) with α-mannosidase and alkaline phosphatase, β-1,6-glucan, but not β-1,3-glucan or chitin, becomes accessible to labeling (see a schematic representation in Fig. 1).
Labeling with positively charged polycationic colloidal gold–chitosan complexes, and observation by electron microscopy, reveals the presence of anionic complexes on the surface of blastoconidia, germ tubes and hyphae (Horisberger & Clerc, 1988). Apparently, these anionic radicals are homogeneously dispersed over the surface of older blastoconidia, except at the bud scars. Their presence on the surface of emerging buds and young cells depends on the growth conditions. In hyphal cells, surface anionic radicals are more abundant, especially at the apex, and are associated with a fuzzy coat that appears external to the wall and seems to play a role in pathogenicity, phagocytosis and adherence. It is feasible that those anionic compounds correspond to glycoproteins containing abundant phosphodiester linkages that confer a negative charge to the cells at physiological pH, and provide hydrophilic or hydrophobic properties to the cell surface, as occurs in Saccharomyces cerevisiae (Jigami & Odani, 1999). Accordingly, hydrophobic properties of C. albicans cells, which are important for virulence, may depend on the polysaccharides located in the outer part of the wall, specifically the phosphodiester-linked, acid-labile β-1,2-mannan, which determines the serotype classification. It has been found that the acid-labile oligomannosides from hydrophobic cells are longer and potentially present in greater abundance than those from hydrophilic cells, and that switching in the hydrophobic status of C. albicans strains depends on a modification of the chain length of these acid-labile β-1,2-oligomannosides. Surprisingly, however, mutation of the gene coding for the enzyme responsible for mannosyl phosphate transfer, and hence for the attachment of β-1,2-mannose oligosaccharides to the acid-labile N-mannan side chains of wall glycoproteins, did not affect morphogenesis, virulence or recognition of the fungus by macrophages (Hobson et al., 2004). On the other hand, mutation of MNT1 and MNT2 genes, which encode partially redundant α-1,2-mannosyltransferases that catalyze the addition of the second and third mannose residues in the O-linked mannose pentamers, resulted in truncation of O-mannan, reduction in the level of mannosyltransferase activity, reduced virulence and reduced adherence, indicating a more important role for O-mannan α-chains in adhesion to host surfaces (Munro, 2005). The capacity to form biofilms (structured microbial communities with high levels of drug resistance) is also dependent on the hydrophobic properties of the cell surface. This is an important property that permits C. albicans to colonize host tissues, implants and surgical material. Farnesol, a quorum-sensing molecule that inhibits hyphal formation in C. albicans, has been found to prevent biofilm formation by the fungus. Using microarray analysis of farnesol-treated populations, a series of gene products important for biofilm formation were identified (Cao et al., 2005). These included, among others, some hyphae-associated proteins, chitinases and proteins associated with the hydrophobic properties of the cells.
Cell wall lipids
Lipids constitute minor components of the cell wall of Candida albicans, and of fungi in general. An interesting lipid in the wall of C. albicans is phospholipomannan, which reacts with antibodies specific to β-1,2-oligomannosides (Mille et al., 2004). Phospholipomannan lacks glucosamine and displays a distinct organization of the glycan chains. Analysis by radiolabeling, methylation-methanolysis and mass spectrometry evidenced a structure made of linear chains of β-1,2-joined mannose residues with degrees of polymerization varying from eight to 18 sugars, bound to phosphoinositol ceramide. Phospholipomannan appears as a novel type of eukaryotic inositol-tagged glycolipid, based upon the absence of glucosamine and the organization of its glycan chains. It has been suggested that this compound may have some relevance in adhesion, protection and signaling in C. albicans. Interestingly, synthesis of phospholipomannan occurs by a unique mechanism diverging from the usual pathway at the mannose-inositol-phospho-ceramide (MIPC) step. Mutation of the MIT1 gene, encoding the glucose degradation product (GDP)-MIPC mannose transferase, reduced virulence in a mouse model and resistance to lysis by macrophages, indicating the importance of the mannolipid in resistance and pathogenicity of C. albicans (Mille et al., 2004).
Chitin is a linear polysaccharide made of more than 2000 units of N-acetylglucosamine (2-acetamido-2-deoxy-D-glucose, GlcNAc) joined by β-1,4-linkages (Fig. 2;Table 1). Chitin chains are associated in an antiparallel fashion through hydrogen bonds to form microfibrils composed of c. 20–400 chains. Owing to this crystalline arrangement, chitin is one of the most insoluble natural products, and this explains why its linkages to β-1,3-glucan form the basic cell wall scaffold (see below) to which mannoproteins are covalently associated.
Peptide N-glycosydic mainly a mannose pyrophosphate
Detergents, reducing agents, HF-pyridine Degradation with alkali and acids
Synthesis of chitin involves a transglycosylation reaction of GlcNAc residues from the universal substrate UDP-N-acetylglucosamine to the growing chain of the polysaccharide. The reaction (catalyzed by ill-known enzymes called chitin synthases, Chsps) requires a divalent metal, generally Mg2+, but does not involve a lipid or a high-energy intermediate (Ruiz-Herrera & Ruiz-Medrano, 2004; Ruiz-Herrera et al., 2004). Early studies revealed that S. cerevisiae contained more than one Chsp, a property found later on to be common to all fungi. This characteristic probably represents a mechanism that permits their adaptation to different environments, and protects the cell in case of the loss of one chitin synthase. Nevertheless, in some specific cases (see below) a distribution of roles of the different enzymes has been revealed. Comparison of fungal Chsps has led to their classification into two divisions (1 and 2), the first with three classes (I, II and III) and the second with two (IV and V) (see review in Ruiz-Herrera & Ruiz-Medrano, 2004). Chitin synthase is accumulated in the cytosol of fungi, C. albicans included, in specialized microvesicles, chitosomes, responsible for the transfer of the enzyme from its site of synthesis to its place of action (see review in Ruiz-Herrera et al., 2004).
In agreement with the multigenic control of chitin synthases in fungi, three genes encoding chitin synthases (CHS genes) were originally described to exist in C. albicans. One of them, CHS2, is preferentially expressed in the hyphal form of the fungus. Nevertheless, its disruption did not affect the dimorphic behavior of the fungus, its levels of wall chitin or its virulence in a mouse model (Gow et al., 1994). Expression of the three genes did not correlate with chitin levels of the cell. CHS2 and CHS3 reached maximal expression 1–2 h after hyphal induction, but CHS1 gene expression remained at low levels in both yeast and hyphae. Chs2p is the most abundant enzyme measured in vitro, but Chs3p is responsible for in vivo synthesis of most chitin in both yeast and hyphae (Mio et al., 1996). Nevertheless, Chs1p was found to be essential for cell integrity and virulence, apparently being involved in septum formation (Mio et al., 1996). Null mutants form long chains without septa, and eventually lyse. Mutants defective in Chs3p are less virulent than the parental strain in a mouse model. By use of a pair of primers corresponding to the CHS1 gene, it was possible to amplify by PCR specific fragments of the homologous CHS1 genes from the medically most important species of Candida: C. albicans, Candida parapsilosis, Candida tropicalis and Candida (Torulopsis) glabrata. The technique offered the possibility to correctly identify the Candida species present in tissue or blood samples (Jordan, 1994). More recently, by use of in silico analysis of the genome sequence, a fourth CHS gene encoding a putative Chsp has been identified in C. albicans. The gene (CaCHS8) encodes a class I chitin synthase with similarities to Chs2p. Null chs8 homozygous mutants display normal growth rates, cellular morphology and chitin content, but their chitin synthase activity is reduced to 25%, and the cells were hypersensitive to Calcofluor white. A double homozygous chs2chs8 mutant possesses less than 3% of the wild-type Chsp activity in vitro but has normal growth rate and morphology. Mutation of the C. albicans homolog of the S. cerevisiae BNI4 gene, encoding the protein responsible for tethering of Chs3p during budding, brought about a reduction in chitin levels and morphological alterations, but not in the chitin ring that separates mother and daughter cell during budding.
Chitin synthase has repeatedly been proposed to be an adequate target for the control of mycoses, taking into consideration the importance of chitin in the structure of the fungal cell wall, and its absence in the host. Two important families of inhibitors of chitin synthase have been described: polyoxins and nikkomycins. Despite their high activity in vitro, their results in vivo have been disappointing. Chemical modification, to obtain derivatives with an in vivo antifungal activity, or isolation of novel compounds appears to be a promising approach (Ruiz-Herrera & San-Blas, 2003).
Cell wall glucans
Candida albicans does not contain α-glucans. It only contains β-glucans, and these are the most abundant polysaccharides of the fungal cell wall in general. They are polymers of glucose moieties joined by β-1,3 and/or β-1,6-glycosidic linkages (Fig. 2;Table 1). β-1,3-Glucan chains possess a helical or spiral backbone, existing as a single helical polymer strand or as a complex of three polymer chains (triple helix) that is stabilized by extensive hydrogen bonding at the C-2 hydroxyl residue. The triple helical form appears to be the preferred and most common molecular conformation in nature. Unbranched β-1,3-glucans have a microfibrillar structure, as revealed by electron microscopy (Kreger & Kopeck, 1975). Candida albicans cell walls contain both β-1,3 and β-1,6-glucans, but no mixed intrachain β-1,3/1,6 linkages. An acid-soluble fraction was found to consist mainly of highly branched β-1,6-glucan. The alkali-insoluble glucan from either yeast or hyphal cells contains 30–39%β-1,3 and 43–53%β-1,6 linkages, whereas that from the germ tubes has these proportions reversed: 67%β-1,3 and 14%β-1,6 linkages. Analysis by high-resolution, solution-state proton nuclear magnetic resonance spectroscopy (NMR) of glucans isolated from yeast or hyphal forms of C. albicans demonstrated that they were different from S. cerevisiae glucans in side-chain branching and reducing termini. Glucans from the yeast-like form have a mean Mr of over 106 Da, whereas the Mr of the mycelial glucan is slightly higher. The existence of a covalent bond between β-1,6-glucan and chitin through a glycosidic linkage at position 1 of glucose and 6 of N-acetylglucosamine was demonstrated in C. albicans (Surarit et al., 1988). Glucans and complexes of β-glucans and mannans are released by C. albicans in synthetic medium, and apparently also into the blood of infected patients. These compounds are toxic and may induce anaphylactic shock and coronary arteritis in murine models. In addition, C. albicans glucans suppress monocyte functions directly, and T-cell function indirectly, suggesting that they may play a role in the development of candidiasis (Nakagawa et al., 2003). On the other hand, anti-β-glucan antibodies participate in the immune response as a result of their capacity to recognize pathogenic fungi (Ishibashi et al., 2005). Recent evidence has confirmed the existence of glucan-specific receptors on cells outside the immune system. This aspect is important, because it is known that hormones and cytokines may act as regulatory messengers between the neuroendocrine and immune systems. In this way, the innate immune system identifies infectious agents by means of pattern-recognition receptors. These receptors recognize pathogen-specific macromolecules called ‘pathogen-associated molecular patterns’. Fungal cell wall glucans nonspecifically stimulate various aspects of innate immunity via interaction with membrane receptors on immune-competent cells. Breuel et al. (2004) have hypothesized that glucans may directly interact with pituitary cells as an early signaling event in fungal infections, and evidence was obtained that pituitary cells directly recognized and responded to fungal cell wall glucans resulting in an increase in the expression of TLR4 and CD14 genes. Additionally, glucans stimulated secretion of prolactin, a hormone that plays an important role in the response to infection. It has been suggested repeatedly that the hyphal form of C. albicans is the invasive stage of the fungus. In this regard, a possible relationship between C. albicans morphogenesis and Dectin-1 deserves attention. Dectin-1 is a receptor that binds β-glucans and is important for phagocytosis of fungi by macrophages. The receptor also collaborates with Toll-like receptors for inflammatory activation of phagocytes by fungi. It has been observed that glucans remain protected from the receptor by other wall components, except during budding, when scar glucans become accessible to binding. On the other hand, hyphae, which lack a budding mechanism, escape from the receptor, and can colonize the host tissues more effectively (Gantner et al., 2005). Synthesis of β-glucans is a complex reaction involving several enzymes located at different cell compartments. Chain growth of β-1,3-glucan is the best-known process. It involves a transglycosylation reaction of glucosyl residues from the universal donor UDP-glucose (UDPG) to the growing polysaccharide chain (for a review, see Ruiz-Herrera et al., 2004). The reaction is stimulated by guanine triphosphate (GTP) and does not involve a lipid intermediate. The final product obtained in vitro may be huge in size, with an Mr of over 20 × 106 kDa, and with a microfibrillar structure. Studies carried out with C. albicans and S. cerevisiae provided evidence that initiation of synthesis of the polymer requires a protein acceptor (Ruiz-Herrera et al., 2004).
Identification of the catalytic β-1,3-transglycosidase moiety was achieved in Neurospora crassa by labeling a highly purified sample with the photosensitive UDPG analog 5-azido-(β-[32P]-UDPG). The results led to the unequivocal identification of one of the labeled peptides as β-1,3-glucan synthase (Schimoler-O'Rourke et al., 2003). Genes encoding the enzyme, named FKS, have been isolated from different fungi, including C. albicans, where three FKS homologs with high similarity to yeast FKS1 were described: CaGSC1, CaGSL1 and CaGSL2 (Mio et al., 1997). The main activity appears to result from Gsc1p (CaFks1p). Inability to inactivate both CaFKS1 alleles, the target for glucan synthesis inhibitors, has been taken as evidence that CaFks1p is essential. It is interesting to recall that Fks proteins lack the motifs characteristic of β-glycosyl transferases: three Asp residues and the QXXRW motif (Campbell et al., 1997), associated with the catalytic activity, and the proposed UDP glucose-binding site (R/K) XGG. All these data suggest that β-1,3-glucan synthases constitute a different family of glucosyltransferases, probably with a distinct phylogenetic origin.
Activity of glucan synthase requires a small GTPase, referred to as Rho1, that is also involved in other essential functions of the cell, including wall integrity and cell morphogenesis. Originally described in S. cerevisiae (Qadota et al., 1996), the genes encoding the Rho1 protein homologs in a range of other fungi have since been identified. Rho1 protein from C. albicans is 82.9% identical to S. cerevisiae Rho1p and contains all the domains conserved among Rho-type GTPases from other organisms. Recombinant C. albicans Rho1p copurifies with the β-1,3-glucan synthase catalytic subunit, and can rescue inactivated β-1,3-glucan synthase from C. albicans and S. cerevisiae membranes. Rho1p has been found to be essential in C. albicans. Accordingly, by use of mutants of the fungus that conditionally expressed Rho1p, it has been observed that depletion of the protein in either yeast or hyphal cells leads to aggregation, death and lysis. Such mutants are also avirulent in a mouse model (Smith et al., 2002).
β-1,3-Glucan synthases constitute the target of important inhibitors of glucan synthesis, and therefore of wall growth. Among these, the most important are echinocandins, synthetically modified lipopeptides that were originally derived from products synthesized by different fungi. These include: aculeacin A (from Aspergillus aculeatus), echinocandin B (from Aspergillus rugulovalvus), pneumocandin B (from Zalerion arboricola), enfumafungin (from a Hormonema-like fungus) and papulacandins (from Papularia sphaerosperma) (for a review, see Denning, 2002). The mutation responsible for spontaneous echinocandin resistance in C. albicans was pinpointed to CaFKS1, the structural gene encoding the most important glucan synthase of the fungus (Mio et al., 1997a). Today, the clinical pharmacology of different echinocandin derivatives for treatment of patients suffering from fungal invasive infections is an important topic of research.
Enzymes belonging to a family of transglycosidases classified into a new group of glycoside hydrolases, Family 72, appear to be important for the structural organization of the cell wall. These were originally described as being involved in the pH-regulated morphogenesis of C. albicans. Later on, they were found to play a role in β-1,6 and β-1,3-glucans bonding in the wall of the fungus (Fonzi, 1999). The corresponding enzymes – Phr1p and Phr2p from C. albicans, Phr1p and Phr2p from Candida dubliniensis, Gasp1 to 5 from S. cerevisiae, and Phr1 from Pneumocystis carinii– belong to the glycosyl phosphatidylinositol (GPI) family of wall proteins. They internally split β-1,3-glucan molecules and transfer the newly generated fragments containing a reducing end to the nonreducing end of other β-1,3-glucan molecules to form a new β-1,3-linkage (Mouyna et al., 2000). Mutants defective in the genes encoding these proteins display abnormal morphologies and alterations in cell wall properties and structure (Muhlschlegel & Fonzi, 1997).
Regarding synthesis of β-1,6-glucan, most of our knowledge comes from studies with S. cerevisiae. Because β-1,6-glucan is the receptor for killer toxin K1, toxin (kre)-resistant mutants became the ideal tool for the identification of genes involved in its synthesis. By these procedures, at least 10 genes involved in β-1,6-glucan biosynthesis were identified in yeast: KRE1, KRE5, KRE6, KRE9, KRE11, CNE1, CWH41/GLS1, KNH1, ROT2/GLS2 and SKN1 (reviewed in Shahinian & Bussey, 2000). Homologs of KRE1, KRE5, KRE6, CNE1, CWH41/GLS1, ROT2/GLS2 and SKN1 have been identified in the C. albicans genome.
Genetic and structural studies suggest that β-1,6-glucan synthesis may occur at the cell surface, but several intracellular events involving the secretory pathway are crucial for the synthesis of the polymer. Accordingly, several gene products involved in the synthesis of β-1,6-glucan have an intracellular location. Kre5 and Cwh41 proteins are located in the endoplasmic reticulum, whereas Kre6 and Skn1 are transmembrane Golgi proteins (Shahinian & Bussey, 2000). The C. albicans homologs of several KRE genes have been isolated, providing evidence that the mechanism of synthesis of β-1,6-glucan occurs by similar mechanisms in both fungi. The CaKRE1 gene was isolated by its ability to complement a kre1 mutation in S. cerevisiae. The predicted protein encoded by CaKRE1 is structurally similar to that encoded by the yeast gene (Boone et al., 1991). kre5 Mutants of C. albicans have reduced β−1,6-glucan levels and severe wall defects and do not form hyphae on solid medium even when serum is added, although they do it in the presence of GlcNac; besides, they are avirulent (Herrero et al., 2004). cDNAs of KRE6 and SKN1 from C. albicans have been isolated (Mio et al., 1997b). In the yeast phase, KRE6 expression was higher than that of SKN1, but expression of the latter increased at the onset of hyphal growth. Homozygous skn1 null mutants were not affected in the levels of β-1,3 or β-1,6-glucan, whereas in heterozygous kre6 null mutants the levels of β-1,6-glucan were reduced by more than 80% without affecting the amount of β-1,3-glucan. It was not possible to isolate the homozygous mutant, thus suggesting the essential role of Kre6. This was confirmed by inhibition of KRE6 expression through the use of the HEX1 promoter. Repressed cells exhibited a partial defect in cell separation and increased susceptibility to Calcofluor white (Smith et al., 2002).
Cell wall proteins
It is beyond discussion that identification of cell wall proteins (CWPs) has advanced rapidly in recent years as a result of the introduction of three novel methodological approaches: (1) sequencing of the whole genomes of fungi; (2) in silico analysis of the genomes with the help of ingenious programs and algorithms; and (3) extremely sensitive proteomic techniques of analysis (mainly mass spectrometry). However, despite these advances, the definition of cell wall proteins has become an elusive matter because different proteins normally considered to be of cytoplasmic location have consistently been found joined to the cell walls in different fungi, C. albicans included (see below). Accordingly, our definition of bona fide wall-bound proteins (Sentandreu et al., 2004) has included only those proteins that have a secretory motif, are N- and/or O-glycosylated, and possess other specific characteristics, such as a GPI-binding motif or specific inner repeats (see below; Table 1 and 2). In C. albicans, these analyses have identified hundreds of ORFs encoding putative wall proteins (De Groot et al., 2003; Garceráet al., 2003).
Table 2. Classification and characteristics of cell wall proteins
NCL-CWP. Proteins extractable by ionic detergents or chaotropic agents. Secretory domain (signal peptide), Ser/Thr rich functional and structural domains (N- and/or O-glycosylation). Retained by noncovalent bonds
Proteins solubilized after degradation of the structural polysaccharides, or by breakage of specific bonds
2a ‘True’ wall proteins
GPI-CWP. GPI proteins. Secretory domain (signal peptide). Structural domain (N- and/or O-glycosylation). GPI binding domain. Linked through β-1,6-glucan to β-1,3-glucan or chitin
ASL-CWP. Alkaline soluble (sensitive) wall proteins. Secretory domain (signal peptide). Ser/Thr rich functional and/or functional structural domain (N- and/or O-glycosylation). Most have also internal repeats. Four Cys residues in most terminal parts. Often processed by Kex2. Lack GPI-binding domain. Linked to β-1,3-glucan by unknown linkages
RAE-CWP. Reducing agents and extractable wall proteins. Secretory domain (signal peptide). Ser/Thr rich functional and/or structural domain (N- and/or O-glycosylation). Some with internal repeats. Connected to other proteins through disulfide linkages
2b ‘Atypical’ wall proteins
Do not possess the canonical signal peptide and other characteristics of the ‘true’ cell wall proteins and have been previously identified as members of different cytoplasmic pathways. Unknown mechanism of association to cell wall
In contrast to these ‘true’ CWPs, mass spectrometric analyses have disclosed the presence of some nonglycosylated proteins in the cell wall of C. albicans (Pitarch et al., 2002), whose relevance and mechanism of retention remain obscure. For these reasons, an operational criterion was introduced for a systematic description to include true (see above) and nontypical CWPs (Table 2;Fig. 1) (for a review, see Sentandreu et al., 2004). Accordingly, CWPs were separated into two classes: class 1, proteins which can be extracted by hot water, but mainly by ionic detergents or chaotropic agents (NCL-CWP); and class 2, proteins resistant to this treatment, and solubilized only after digestion of the structural polysaccharides, or by breakage of specific bonds through which they are bound to wall polysaccharides. This second group of proteins was divided into two subclasses. Proteins from subclass 2a are glycoproteins covalently linked to the cell wall – in other words, ‘true’ CWPs. Proteins from subclass 2b (nontypical proteins) are devoid of a carbohydrate moiety and are retained in the wall by unknown mechanisms. Three different types of covalently bound glycoproteins have been described in subclass 2a. Group I includes proteins bound to β-1,6-glucans through a GPI moiety (GPI-CWP) (De Nobel & Lipke, 1994; De Groot et al., 2005). Group II corresponds to Pir proteins (proteins with internal repeats), characterized as containing repetitive sequences (Toh-e et al., 1993) and being highly O-glycosylated. Pir proteins are attached to β-1,3-glucan by unknown alkali-sensitive bonds (possibly O-glycosidic linkages) (ASL-CWP). Recently, some proteins without internal repeats but bound through alkali-sensitive bonds have been described (Castillo et al., 2003; De Groot et al., 2005). Group III is formed by mannoproteins (RAE-CWP) retained by disulfide bridges, which are extracted by treatment with reducing agents such as β-mercaptoethanol (β-ME) or dithiothreitol (Moukadiri & Zueco, 2001).
Glycosyl phosphatidylinositol proteins are rich in Ser and Thr residues, and are highly O-glycosylated (Fig. 2;Table 2). The structure linking the C-terminal end of GPI proteins to the lipid moiety is identical in GPI anchors from all organisms analyzed so far, namely protein-CO-NH-(CH2)-PO4-Man-α-1,2-Man-α-1,6-Man-α-1,4-GlcN-α-1,6-inositol-PO4-lipid. The core contains branches of α-1,3- and α-1,2-linked mannose units (Sipos et al., 1995). However, the GPI anchors from various species differ widely with regard to the side chains attached to this core structure, as well as to the lipid moieties of the anchor. It was suggested (Frieman & Cormack, 2003) that the amino acids immediately upstream of the site of GPI anchor addition (the ω site) serve as the signal determining whether a GPI protein localizes to the cell wall or to the plasma membrane. This signal consists of a region of hydrophobic amino acids, followed by another short region of more hydrophilic amino acids and a binding site formed by three amino acid residues named ω, ω+1 and ω+2. Cleavage of the protein takes place between ω and ω+1, and the GPI anchor remains bound to the ω amino acid (Nuoffer et al., 1993). Transfer of the GPI moiety to the protein, which takes place in the lumen of the endoplasmic reticulum, is carried out by a transamidation reaction involving cleavage of a carboxy terminal hydrophobic sequence, with the concomitant formation of an amide linkage between the ethanolamine phosphate of the GPI and the new carboxy-terminal amino acid (Udenfriend & Kodukula, 1995). Transfer is catalysed by a GPI-transamidase. Assembly of S. cerevisiae GPIs includes the addition of a fourth, side-branching mannose to the third mannose of the core GPI glycan by the Smp3 mannosyltransferase. The gene encoding this enzyme has been cloned in C. albicans, and was found to be essential for growth, suggesting that C. albicans utilizes the same pathway as S. cerevisiae for GPI synthesis (Grimme et al., 2004). In cell wall proteins, the above described anchor is trimmed, and only a part is retained at their C-termini, which participate in the binding of the proteins to β-1,6-glucan (Lipke & Ovalle, 1998).
Quantitatively, GPI proteins appear to be the most important ones in C. albicans, where they account for about 88% of all covalently linked wall proteins. These proteins are bound via β-1,6-glucan to β-1,3-glucan (90%), or to chitin (about 10%) (Marcilla et al., 1991). In Δmnn9 and Δpmt1 mutants that are defective in protein N- or O-glycosylation, respectively, the levels of GPI proteins bound to chitin were increased (Kapteyn et al., 2000). As indicated above, in silico analyses have disclosed the presence of a large number of GPI proteins in Candida species. Using the big-ii predictor for fungal systems, a total of 234 putative GPI proteins, bound to either the plasma membrane or the cell wall, have been identified in C. albicans (Eisenhaber et al., 2004). In Candida glabrata, the number was lower, namely 106 (Weig et al., 2004). It was calculated that 50 of these were adhesive in nature, 11 were glycoside hydrolases, others had a different enzymatic activity, and the rest were probably structural proteins. Among the different GPI proteins identified in Candida spp., we may cite adhesins, such as Hwp1, the agglutinin-like sequence (ALS) family of C. albicans, and Epa1p of C. glabrata (Sundstrom, 2002). Some GPI proteins are essential. Accordingly, double mutants affected in Dfg5p and Dcw1p, two similar GPI proteins, are synthetically lethal (Spreghini et al., 2003). In C. albicans, a proteomic analysis of cell wall extracts obtained from exponentially growing yeast-like cells was performed (De Groot et al., 2004). By use of hydrolysis with hydrofluoric acid (HF)-pyridine, which cleaves phosphodiester bonds, followed by liquid chromatography/mass spectrometry, 12 GPI-CWPs were identified. Five of them (Cht2, Crh11, Pga4, Phr1 and Scw1) have different enzymatic activities related to carbohydrates, two are adhesive proteins (Als1 and Als4), one is related to flocculins (Pga24) and another one appears to be a superoxide dismutase (Sod4/Pga), whereas the functions of the other three (Ecm33.3, Rbt5 and Ssr1) are unknown. In the pathogenic related species C. glabrata, similar proteomic analysis revealed the presence of two structural GPI proteins (CWP1.1p and CWP1.2p) and Crh1p, a putative 1,3-β-glucan remodelling enzyme (Weig et al., 2004).
These are highly O-glycosylated proteins characterized by the presence of internal repeats in variable number (Fig. 2;Table 2) (Toh-e et al., 1993). Their presence in ascomycetes and deuteromycetes appears to be universal, and their organization is similar, including the presence of a signal peptide, a Kex2 sensitive site, a domain with two to 11 repetitive sequences, and a C-terminal sequence with four Cys residues at identical positions (repeat(s)-Cys–66aa–Cys–16aa–Cys–12aa–Cys–COOH). Pir proteins do not contain a GPI anchor motif and are attached to the cell wall by unknown alkali-labile bonds, possibly O-glycosidic linkages with β-1,3-glucan. At least some of the Pir proteins are retained in the wall exclusively by disulfide bridges, given that some of them are released by reducing agents such as β-ME or DTT (Castillo et al., 2003; Toh-e et al., 1993). In C. albicans, an antibody directed to the S. cerevisiae Pir protein Hsp150 recognized two proteins extracted by alkali or β-1,3-glucanase, and a high-molecular-mass protein secreted to the growth medium, demonstrating the existence of Pir-related proteins in the fungus (Kandasamy et al., 2000). In similar experiments, Western blot analysis using an antiserum directed against S. cerevisiae Pir2p/Hsp150 revealed the presence of at least two differentially expressed Pir2 homologs in the cell surface of C. albicans (Kapteyn et al., 2000). It was also observed that in Δmnn9 and Δpmt1 mutants, which are defective in N- or O-glycosylation, respectively, as well as in a Δkre6 mutant, the amounts of Pir proteins were slightly up-regulated (Kapteyn et al., 2000). Recent results have revealed that C. albicans contains a single Pir-protein-encoding gene. By use of mass spectrometry and in silico analyses, two Pir proteins encoded by nonidentical alleles of a single gene (CaPIR1) were identified (Martinez et al., 2004). Both encoded proteins contained a single N-mannosylated chain, four Cys residues and seven repeats, but one of them was 21 amino acids shorter. Homozygous mutants were impossible to obtain, suggesting that the gene is essential for growth. Heterozygous mutants displayed an abnormal phenotype associated with wall alterations (Martinez et al., 2004). In this regard, in a parallel study (De Groot et al., 2004) similar analysis of NaOH-released proteins led to the identification of two Pir proteins (named Pir1 and Pga29). It is feasible that these correspond to the two allelic products of CaPIR1 described above.
Regarding atypical wall proteins belonging to subgroup 2b (see above and Table 2) in C. albicans, numerous examples exist. However, it is important not to confuse the two types: (i) the highly glycosylated proteins that contain a signal peptide and are secreted through the normal pathway mostly to the medium, but retained in the wall in variable proportions where they have different functions, and (ii) those proteins that have been amply recognized to be of cytoplasmic origin and to lack the characteristics of a bona fide secreted protein, and that have been recently detected in significant amounts in the cell wall. Only the latter are the so-called ‘atypical proteins’.
Proteins related to the hsp70 and hsp90 families of conserved stress proteins were found in the cell wall of C. albicans, apparently as bona fide components (reviewed in Chaffin et al., 1998). Higher levels of enolase and two subforms of phosphoglyceromutase were detected in the wall of a fluconazol-sensitive strain, as compared to a resistant one, which displayed higher levels of two exoglucanases (Angiolella et al., 2002). Using a λgt11 cDNA library from C. albicans to screen sera from patients with systemic candidiasis, a cDNA has been identified corresponding to the gene encoding 3 phosphoglycerate kinase (Alloush et al., 1997). The protein is located at the surface of the wall mainly of yeast cells, being released by treatment with β-ME. A similar protocol using antibodies directed to cell wall proteins led to the identification of C. albicans genes encoding several nontypical wall proteins, which were located in the cell wall later on by standard protocols. One of these genes encoded 3-phosphoglycerate kinase, and another one was a novel gene encoding two products: a cytochrome c haem lyase targeted to the mitochondria, and a cell wall protein (Cervera et al., 1998). Whether these proteins are adventitiously trapped in the cell wall, as has been suggested by some authors, or somehow are actively released and associate to the wall polymers, as suggested by others (see above), is still a matter of discussion. Until a decisive answer is obtained, it is advisable to treat this matter cautiously.
Nature of the carbohydrate moieties of wall proteins
Different forms of glycosylation have been described in relation to wall proteins: N-glycosylation, O-glycosylation, and attachment of a GPI anchor. Normally, these proteins are rich in Ser and Thr residues where O-glycosylation occurs, and contain the sequons, tripeptides of N-Asn-Xaa-Ser/Thr, where N-glycosylation takes place. The level of glycosylation is variable, often as high as 50–95% by weight. Branched oligosaccharide moieties of high molecular weight made mostly of mannosyl units are connected to the peptide by N-glycosidic bonds between N-acetylchitobiosyl and asparagine moieties. Two distinct parts have been detected in the glycosidic moiety of glycoproteins, an inner core of 10–14 residues (Asn-GlcNAc2-Man10–14), a large outer chain containing a backbone of α-1,6-linked mannose residues, and a variable number of α-1,2-linked mannose side chains. The side chains in turn contain a variable number of α-1,2 and α-1,3-bound mannosyl moieties, and occasionally phosphomannosyl units that confer a net negative charge on the cell wall, which contributes to the properties of the cell surface.
Regarding O-glycosylated proteins, chains of up to five α-1,2 and α-1,3-bound mannose units are linked to Ser or Thr residues. This bond is sensitive to weak alkali treatment (β-elimination). These proteins are characterized by their rigid structure that protrudes from the wall surface in a manner similar to antennae (for a review, see Orlean, 1997).
In relation to the carbohydrate moieties of the wall proteins of C. albicans, special mention should be made to a lipid-modified polysaccharide called phospholipomannan, which is a modified member of the MIPC family, but due to its hydrophilic properties given by the large polymannose moiety it locates in the cell wall. This molecule seems to be the center of the host cell response following deep C. albicans infections, but the mechanisms for this have not yet been elucidated (Jouault et al., 2003).
Roles of wall proteins
Wall proteins are absolutely essential for the life of C. albicans, as with all fungi. Proteins have extremely important and varied roles, among which we can cite the following five.
Some wall enzymes play a role in the degradation of large impermeable molecules, making the products accessible for cell nutrition; others are involved in the degradation of cell wall polymers or in their synthesis, being necessary for wall and therefore for cell growth. Accordingly, several degradative enzymes have been found located in the wall of C. albicans (Chaffin et al., 1998; Pedreño et al., 2004). Also, a number of GPI proteins display enzymatic activities related to the synthesis and modulation of wall components (chitinases, glucanases, etc.).
Other wall proteins are involved in the interaction with other cells. Glycoproteins present in fimbriae take an active role in this sense. Fimbriae are external protruding filamentous structures that provide the initial contact of pathogenic fungi such as C. albicans with host cells (Tokunaga et al., 1990). Fimbriae mannoproteins are made of 80–85% D-mannose and 10–15% protein. Among other surface proteins involved in interactions with other cells, or inert substrates, we may cite wall proteins involved in adhesion of C. albicans to host tissues and some of their products such as fibrinogen, complement fragments and several extracellular matrix components. Attachment and adherence of C. albicans has been described to depend on at least four recognition systems, classified according to the type of adhesins, the kind of host cells (epithelial, endothelial or platelets) and the type of host cell ligand (carbohydrate or protein) (for a review, see Calderone, 1993). System I adhesin corresponds to a mannoprotein with lectin-like properties that recognizes fucosyl or glucosaminyl glycosides of epithelial cells. System II adhesin functionally resembles the ‘integrin’ receptors of mammalian cells. It is also a mannoprotein, but it recognizes proteins with the RGD ligand of platelets, endothelial cells or RGD (Arg-Gly-Asp acid) domains of extracellular matrix proteins of endothelial cells. System III adhesin (mannan) promotes the adherence of the fungus to epithelial cells. However, this adhesin clearly differs from the system I adhesin, and it utilizes the protein component to recognize the ligand. System IV adhesin (mannoprotein) seems to be associated with the colonization of splenic tissue by the pathogen. The C. albicans ALS gene family is composed of eight genes (ALS1 to ALS7 and ALS9) encoding cell wall glycoproteins involved in adhesion to host surfaces. Allelic variations in some of these genes suggest that they may be related to the evolutionary stress to which the fungus has been subjected. Analysis of the structure of these proteins, their cloning and expression in S. cerevisiae cells, and analysis of the adherence characteristics of the transgenic yeasts have provided evidence of the extreme capacity of adaptation in C. albicans (Sheppard et al., 2004).
Different wall proteins are antigenic, as occurs with several C. albicans glycoproteins that are differentially present in the yeast or mycelial forms (Sundstrom et al., 1988). A 47-kDa protein component of different strains of C. albicans was identified as an immunodominant antigen present in the cytoplasm and cell wall of both yeast and mycelial cells, but mainly exposed on the outer surface and unable to bind to Concanavalin A (Matthews et al., 1988). An interesting wall glycoprotein of 260 kDa, specific to the mycelial form of C. albicans, was isolated and monoclonal antibodies raised against its polypeptide moiety (Casanova et al., 1989). (It is relevant to recall that antigenicity of different wall glycoproteins has been associated not with the protein but with their mannan moiety.) Interestingly, Fab fragments from these monoclonal antibodies inhibited the yeast-to-mycelium dimorphic transition of the fungus (Casanova et al., 1990).
An important role of some wall proteins is their involvement in the establishment of the pathogen in the host, or in the response of the latter to invasion. The notion that in pathogenic fungi wall proteins frequently constitute virulence factors is widely accepted. In relation to this, it is known that the wall proteins extracted by β-ME exacerbate collagen-induced arthritis in mice, and, as described above, a number of mutants deficient in proteins located or related to cell wall construction display reduced virulence. This type of information provides evidence on the role of wall proteins as pathogenicity or virulence factors, either directly or indirectly (Fradin et al., 2005)
Wall structure and morphogenesis
Finally, other glycoproteins may be important from a structural point of view, and for the morphogenetic response of the fungus. An example of this is a putative surface glycosidase (Csf4), identified as an important factor for cell wall integrity and maintenance. Deletion of CSF4 reduced hyphal growth, adherence to mammalian cells and virulence in a mouse model (Alberti-Segui et al., 2004). A further example is the 260-kDa protein that, as described above, was essential for mycelial growth. GPI protein Dfg5p has also been found to be involved in dimorphism. Mutants lacking this protein are defective in hypha formation at alkaline pH, suggesting that Dfg5p is involved in the transmission of the surface signal required for hyphal growth (Casanova et al., 1990). Mutants defective in another GPI protein, CaEcm33p, display morphological alterations, including reduced hyphal formation, defects in wall structure and reduced virulence in a murine model (Martinez-Lopez et al., 2004). Finally, we must recall the CaPir1 protein cited above, which is apparently required for the correct formation of the cell wall, because null mutants are inviable (Martinez et al., 2004).
Cell wall maturation and organization
It is not an exaggeration to suggest that organization of the cell wall components is initiated at the level of transcription. In Saccharomyces cerevisiae, two transcription factors, Ace2p and Swi5p, are key regulators of cell wall synthesis. In agreement with these results, mutation of the gene CaACE2, which is the sole homolog of these genes in C. albicans, provoked severe alterations in functions that depend on the normal structure of the cell wall (Kelly et al., 2004). Mutant cells displayed defects in cell separation, hyphal growth, adhesion and virulence in a mouse model, suggesting that cell wall organization depended on the orchestrated synthesis, controlled by the transcription factor encoded by CaACE2, of a number of cell products.
It is important to recall that the cell wall is a coherent structure. Accordingly, in order to achieve its final structure, the establishment of connections among its different components becomes necessary. Some of these associations involve hydrogen, hydrophobic or polar bonding, but others take place through covalent linkages. These associations must occur in the cell wall itself, where they produce changes in its mechanical properties (for a review, see Ruiz-Herrera et al., 2004). The Family 72 of transglycosidases may function in the cross-linking of β-glucans in the cell wall (Fonzi, 1999; Mouyna et al., 2000). Also a transglycosylation reaction catalyzed by a wall-located β-1,3-glucanase was reported to occur in C. albicans, apparently being involved in branching of β-1,3-glucan (Goldman et al., 1995). In vitro, the enzyme cleaved laminaribiose from the reducing end of a linear β-1,3-glucan and transferred the remainder to another laminarioligosaccharide, to give a product containing a β-1,3-β-1,6-branchpoint (Goldman et al., 1995). Mutation of BGL2, the gene encoding this β-1,3 glucosyltransferase in a clinical strain of C. albicans, gave rise to reduction in growth rate, cell aggregation during stationary phase, decrease in virulence in a murine model and increase in sensitivity to nikkomycin Z (Sarthy et al., 1997). The observation that null mutants still displayed β-1,3-glucosyltransferase activity made the authors suggest the presence of additional enzymes exercising the same activity. As described above, β-1,3-glucan and chitin in C. albicans (Surarit et al., 1988) are covalently linked, and the formation of such a bond must occur in the wall itself. Another type of association of wall macromolecules takes place among proteins. This may occur through the formation of disulfide bonds, a redox reaction which is not fully understood, whereas other reactions may account for the association between O-linked and N-linked mannoproteins. Accordingly, the covalent bonding of an O-glycosylated epitope to N-glycosylated mannoproteins was reported to occur in the cell wall of C. albicans (Elorza et al., 1989). Synthesis of the cell wall in fungi is under the control of mechanisms that monitor its integrity and characteristic resistance. In S. cerevisiae, a decrease in the synthesis of some wall polysaccharides gives rise to a significant increase in the levels of chitin. For example, mutants deficient in O-glycosylation (pmt mutants) or N-glycosylation (mnn mutants) contained twice the normal levels of chitin (Gentzsch & Tanner, 1996). These mutants also contained higher amounts of Pir and GPI proteins (Kapteyn et al., 2000). Mutants of C. albicans and C. glabrata deficient in β-1,6-glucan synthesis also had increased levels of chitin in the wall. It has been suggested that these alterations in the chemical composition of the wall may operate as compensatory mechanisms to guarantee its integrity. Their mode of action involves mitogen activated protein kinase (MAPK) cascades, as occurs with other signaling pathways. In C. albicans, a MAPK named Mkc1p has been identified as a homolog of Mpk1p (Slt2) from S. cerevisiae (Navarro-Garcia et al., 1998), the MAPK involved in the regulation of cell wall integrity. Homozygous mkc1 mutants showed alterations in the cell surface, evidenced by scanning electron microscopy, increased amounts of specific cell wall epitopes, and higher sensitivity to inhibitors of β-glucan and chitin synthesis. Using the two-hybrid system, it was demonstrated that Mkc1p was able to interact specifically with yeast Mkk1p and Mkk2p, the MAPK kinases of the Pkc1p-mediated route. MAPK may not be the sole pathway involved in the regulation of cell wall integrity. Recent data have provided evidence that the phenomenon also involves a signaling pathway dependent on phosphoinositides (Schorr et al., 2001).
With regard to the sensing mechanisms, it seems relevant that at least three putative two-component histidine kinase signal transduction proteins, including Chk1p and a response regulator protein (Cssk1p), exist in C. albicans. To determine the role of these systems, chk1 mutants have been obtained. These became avirulent in a murine model, and displayed significant changes related to cell wall composition and structure. Accordingly, the alkali-soluble carbohydrate levels, as well as the Mr of acid-stable mannan species, were reduced, the O-linked oligosaccharides appeared truncated, and the degree of polymerization and the ratio of β-1,3/β1,6 linkages in β-glucans were all reduced (Kruppa et al., 2003). In further analyses, immunoelectron microscopy with specific antibodies directed to surface isotopes revealed alterations of the cell surface, not only in chk1 mutants but also in mutants defective in the other two genes encoding histidine kinases (HK) of the fungus (SLN1 and NIK1). Analysis of the expression of 29 genes involved in mannan biosynthesis revealed similar alterations in all mutants (Kruppa et al., 2004). All these data have been taken as evidence for a role of the two-component histidine kinase signal transduction proteins in the regulation of cell wall structure in C. albicans. Finally, it is also significant that the repair mechanisms induced by the perturbations in the cell wall construction of S. cerevisiae mutants (fks1, Kre6, mnn9, gas1 and knr4) integrate three major regulatory systems: the PKC1-SLT2 MAPK signaling unit, the Ca2+/calcineurin-dependent signaling conduct and the general stress system (Lagorce et al., 2003). Recently, a global model integrating different pathways and processes in relation with the fungal status (commensal or pathogenic) has been proposed by the European Consortium ‘Novel approaches for the control of fungal diseases’ (Fig. 3).
From the data presented above, it is evident that our concepts on the structure, synthesis and organization of the cell wall of Candida albicans have rapidly evolved. The introduction of different modern techniques has allowed comparative analyses of the structural components, and in silico analyses of the genome and sequencing of the enzymatic and structural proteins have deepened our knowledge on how the cell wall is organized. Comparison of the results uncovered in C. albicans with those obtained with other fungi has demonstrated the similarities in the mechanisms of cell wall synthesis and construction in different fungi, but also the existence of important differences that reveal the uniqueness of the C. albicans species. Among the important differences disclosed, we may cite the structure of β-1,3-glucan and Pir proteins, the biosynthetic mechanism of β-1,6-glucans, the different number and roles of chitin synthases and Pir proteins, and the antigenic and adhesive characteristics of the cell wall surface. Also, the observation that some components of the C. albicans cell wall play important roles in pathogenicity and virulence imposes a strong distinction between the cell wall of this species and that of their saprophytic relatives. Finally, it should be stressed that our extended knowledge of the role of wall components in the pathogenic behavior of C. albicans is leading to a better comprehension of the molecular bases of the invasiveness and aggressiveness of the fungus, and accordingly to the design of better strategies for its control.
Original work from the laboratories of J.R.H., R.S., E.V. and M.V.E. was supported, respectively, by CONACYT, Mexico, and the European Union (QLK2CT-2000-0795 and MRTN-CT-2003-504148), Spanish Ministerio de Ciencia y Cultura (BMC2003-01023) and Agència Valenciana de Ciència i Tecnologia de la Generalitat Valenciana (Grupos 03/187). We acknowledge the helpful discussions with our colleagues in the EC Galar Fungail Consortium.