INTRODUCTION

Top Previous Next References

Candida albicans is a serious agent of infection, particularly in immunocompromised patients. The delicate balance between the host and this otherwise normal commensal fungus may turn into a parasitic relationship, resulting in the development of infection, called candidiasis. The nature and extent of the impairment of normal host defense influence the manifestation and severity of infection. In general, superficial mucocutaneous candidiasis is frequent in patients with T-cell deficiencies, such as AIDS patients. The more serious, life-threatening, deep-seated or disseminated candidiasis is normally found in a spectrum of severely immunocompromised patients (29, 390). The fungus is not a mere passive participant in the infectious process, and a hypothetical set of virulence factors for C. albicans has been proposed and supported by various studies. These fungal attributes include the production of secreted hydrolytic enzymes, dimorphic transition (morphogenetic conversion from budding yeast to the filamentous growth form or hypha), antigenic variability, the ability to switch between different cell phenotypes, adhesion to inert and biological substrates, and immunomodulation of host defense mechanisms (for a review of these topics, see reference 96).

Initially, the cell wall was considered an almost inert structure that supplies rigidity and protection to the protoplast. Today, the cell wall is well established as being essential to almost every aspect of the biology and pathogenicity of C. albicans (64). The cell wall acts as a permeability barrier and is the structure that maintains the characteristic shape of the fungus. Also, as the most external part of the cell, the wall mediates the initial physical interaction between the microorganism and the environment, including the host. For these reasons, the cell wall of C. albicans is the focus of study by numerous research groups. Their objectives are the elucidation of both basic biological processes and functional mechanisms regulating the synthesis, organization, and environmental interactions of this complex macromolecular structure. Proteins have been implicated in most of the cell wall functions. Extensive reviews exist on different aspects of the cell wall of C. albicans (64, 478-482, 491, 493); however, recent reviews that focus on proteins have been limited primarily to the function of proteins as adhesins (49, 50, 110, 111, 151, 209, 406). This review focuses on both general characteristics of the cell wall and secreted proteins and specific aspects of individual proteins.

Although hydrolytic enzymes such as acid phosphatase were examined previously, studies on the identity and function of protein components began in the early to mid 1980s with studies in 1983 by Chaffin and Stocco (71), in 1985 by Elorza et al. (125) and Sundstrom and Kenny (524), and in 1986 by Pontón and Jones (414). In the first part of this decade, there has been an explosive growth in the number of studies of the cell wall proteins. These studies have been driven by presumed virulence functions of specific proteins and fueled by the realization that this is a complex, dynamic "organelle." Out of such impetus has come the identification of specific proteins not yet associated with specific pathogenic function and observations with more general import for the cell and cell wall proteins. This increasing knowledge of the protein component of the cell wall may result in a better understanding of the pathogenic mechanisms of the fungus and also may contribute to the design of innovative therapeutic regimens and diagnostic procedures (175, 333). At times, these studies have revealed several surprises and unexpected findings, which only add more attraction to the study of this fascinating microorganism. In writing this review, we have focused on the protein component with three objectives: (i) to summarize general aspects of proteins; (ii) to summarize studies on specific proteins or protein families; and (iii) to consider the implications, unanswered questions, and future research directions suggested by these studies.

Although the terms "dimorphism" and "dimorphic fungus", i.e., existing in two morphological forms, are well established and commonly accepted when referring to C. albicans, strictly speaking this fungus has the ability to adopt a spectrum of morphologies, and thus C. albicans could be considered a "polymorphic" or "pleomorphic" organism (244, 390). Since changes in the cell wall determine the shape of the whole fungal cell, the cell wall is the structure ultimately responsible for a given morphology. C. albicans can reproduce by budding, giving rise to the formation of yeast cells (also designated blastospores or blastoconidia). The production of germ tubes results in the conversion to a filamentous growth phase or hypha, also called the mycelial form. The formation of pseudohyphae occurs by polarized cell division when yeast cells growing by budding have elongated without detaching from adjacent cells. Under certain nonoptimal growing conditions, C. albicans can undergo the formation of chlamydospores, which are round, refractile spores with a thick cell wall. These morphological transitions often represent a response of the fungus to changing environmental conditions and may permit the fungus to adapt to different biological niches. The transition from a commensal to a pathogenic lifestyle may also involve changes in environmental conditions and dispersion within the human host. The ultrastructure, composition, and biological properties of the cell wall are affected by these morphological changes (64). Although progress has been achieved in the recent years, the molecular mechanisms governing these morphogenetic conversions are still not fully understood, partly due to the difficulty of genetic manipulations in this fungus (274, 275, 474). Recent reports that may herald rapid advances in this area have identified transcriptional regulatory genes, a general transcriptional repressor TUP1 (38), a putative transcriptional factor RBF1 (220), and a myc-like transcriptional factor EFG1 (516) that affect cellular morphology when their expression is altered. Most of the observations from these studies have been incorporated by Magee (310) into a model for the regulation of pseudohyphal growth.

Adhesion is a prerequisite for colonization and an essential step in the establishment of infection. C. albicans adheres to epithelial cells, endothelial cells, soluble factors, extracellular matrix, and inert materials implanted in the body of the host. Multiple adherence mechanisms appear to be used by C. albicans cells (49, 50, 110, 111, 151, 209, 252, 406). Physical interactions of this fungus with the host are mediated at the cell surface, and cell wall constituents implicated in binding have been designated adhesins (49). The large repertoire of adhesins displayed by this fungus may reflect the variety of host sites that it can invade (49, 50, 110, 111, 209). Specific characteristics of individual cell wall moieties participating in adhesion events are discussed later in this review.

Another important aspect of interactions with the host, with direct implications for pathogenesis, is the potential of this fungus to modulate the immune response mounted by the host (64, 96, 107). The capacity of cell wall constituents, including glucan, chitin, and mannoproteins, to modulate (activate or depress) the immune response is well documented (11, 64). Mannans and mannoproteins display the most potent immunomodulatory activity, being able to regulate the action of virtually all arms of the immune system (natural killer cells, phagocytic cells, cell-mediated immunity, and humoral mechanisms) (64, 107, 402, 412). Although individual cell wall moieties with immunomodulatory properties are described below, readers are referred to excellent reviews on this topic (11, 64, 107).

CELL WALL COMPOSITION AND ORGANIZATION

Top Previous Next References

Composition

Approximately 80 to 90% of the cell wall of C. albicans is carbohydrate. Three basic constituents represent the major polysaccharides of the cell wall: (i) branched polymers of glucose containing -1,3 and -1,6 linkages (-glucans); (ii) unbranched polymers of N-acetyl-D-glucosamine (GlcNAc) containing -1,4 bonds (chitin); and (iii) polymers of mannose (mannan) covalently associated with proteins (glyco[manno]proteins). In addition, cell walls contain proteins (6 to 25%) and minor amounts of lipid (1 to 7%) (50, 64, 490, 493).

The microfibrillar polymers (-glucans and chitin) represent the structural components of the wall. They form a rigid skeleton that provides strong physical properties to the cell. From a quantitative point of view, -glucans are the main constituent, accounting for 47 to 60% by weight of the cell wall. Chitin is a minor (0.6 to 9%) but important component of the C. albicans wall, particularly of the septa between independent cell compartments, budding scars, and the ring around the constriction between mother cell and bud (126, 360).

On the other hand, mannose polymers (mannan), which do not exist as such but are found in covalent association with proteins (mannoproteins), represent about 40% of the total cell wall polysaccharide and are the main material of the cell wall matrix (50, 64, 480, 490, 493). The term "mannan" has been used also to refer to the main soluble immunodominant component present in the outer cell wall layer of C. albicans, called phosphomannoprotein or phosphopeptidomannan complex. This cell wall fraction contains homopolymers of D-mannose (as the main component), 3 to 5% protein, and 1 to 2% phosphate (436). The general features of cell wall mannoproteins in C. albicans are basically identical to those found for Saccharomyces cerevisiae, one of the most thoroughly investigated yeasts in this regard. Several studies have resulted in a detailed knowledge of the structure of this cell wall constituent in C. albicans (12, 262-265, 494-498). Thus, mannose polymers are linked to the protein moiety through asparagine (by N-glycosidic bonds through two GlcNAc [di-N-acetylchitobiose] residues) and threonine or serine (by O-glycosidic, alkali-labile linkages) residues. The N-glycosidically linked carbohydrate is composed of backbone chains of -1,6-linked mannopyranosyl residues to which oligosaccharide side chains are attached. The side chain mannopyranosyl residues contain -1,2, -1,3, -1,2, -1,6, and phosphodiester linkages as well as branches (-1,6) that are oversynthesized under acidic growth conditions (150, 152, 261-267, 494-498). The O-glycosidically-linked sugar component consists of single mannose residues and short, unbranched mannose oligosaccharides (412). Several studies raise the question of additional sugars present in cell wall constituents. These observations include the following: (i) not all proteinaceous moieties present in cell wall extracts from this fungus react with concanavalin A, a lectin recognizing -mannosylpyranose, or with polyclonal and monoclonal antibodies that recognize other mannan epitopes, such as factor 6, a mannooligosaccharide that confers serotype A specificity (57); (ii) differences in glycosylation and in sensitivity to neuraminidase have been detected in candidal receptors for complement (4, 563); and (iii) treatment with neuraminidase affects the electrostatic surface properties of C. albicans as detected with a fluorescent probe (227). As suggested in these studies, the observations raise the possibility that additional sugars are cell wall constituents. However, the observations could reflect the existence of contaminating proteases in the glycosidase preparation. Sugar residues other than mannose may define either additional functional or antigenic motifs or both in cell wall glycoproteins.

The percent composition of walls from yeast cells and filamentous forms are similar, although the relative amounts of -glucans, chitin, and mannan vary according to the C. albicans growth form considered (50, 480, 491). Hyphal cells contain at least three times as much chitin as yeast cells do (77, 127, 518). Chitin is the first polymer to appear in regenerating protoplasts (124, 375). Although the ratio of -1,3- to -1,6-glucan in the insoluble fraction is similar in yeast and hyphal cells, the insoluble glucan in the initial period of germ tube formation contains considerably more -1,3 linkages than that found in yeast and mature hyphal cells (518). The literature contains several reports on the identification of morphology-specific proteins and mannoproteins that are discussed later in this review.

The different cell wall components interact with each other to give rise to the overall architecture of the cell wall. Besides hydrogen and hydrophobic bonds, there is also experimental evidence for the presence of covalent linkages between different components (453, 482). Surarit et al. (527) reported the presence of glycosidic linkages between glucan and chitin in the nascent wall of C. albicans. Recent evidence indicates that mannoproteins may also establish covalent associations with -glucans (237, 238, 462, 463). It is suggested that -1,3- and -1,6-glucans are linked to proteins by phosphodiester linkages, a process that may involve the participation of a GPI (glycosyl phosphatidylinositol) anchor (238) (see below). Protein and mannoprotein species that are released only after digestion of the glucan cell wall network with -glucanases may play a key role in configuring the final cell wall structure characteristic of each growth form (yeast and mycelium) of C. albicans (453, 479-482). Interactions between glyco(manno)proteins and chitin also appear to exist in the wall of C. albicans cells as deduced from two lines of evidence: (i) chitinase treatment of isolated cell walls solubilizes protein moieties, and (ii) the kinetics of incorporation of protein and mannoprotein constituents into the walls of regenerating protoplasts is altered in the presence of nikkomycin, an antibiotic that blocks chitin synthesis (124, 319).

Cell wall architecture has been studied most extensively in S. cerevisiae and is likely to be a model for C. albicans since there are some similar observations, in particular sensitivity to enzymatic digestion, glucan-mannoprotein linkages, and candidate proteins, that fit the same model (237, 238, 246, 247, 268, 462, 463). In a very recent study, Kollár et al. (268) detected the presence of material containing all four major cell wall components, -1,3-glucan, -1,6-glucan, chitin, and mannoprotein. Their analysis indicated that -1,6-glucan has some -1,3-glucan branches that may be linked to the reducing end of chitin. The -1,6-glucan and mannoprotein are attached through a remnant of the mannoprotein GPI anchor. Reducing ends of -1,6-glucan may also be attached to the nonreducing end of -1,3-glucan. The proportion of cell wall polysaccharide involved in this type of structure is not clear. The following cell wall building block, where the linkages are indicated by the long dashes, is proposed (247, 268): MannoproteinGPI remnant-1,6-glucan-1,3-glucanchitin. The authors point out that these linkages are likely to be formed in the periplasmic space as a common end of the individual biosynthetic pathways. Chitin and -1,3-glucan are synthesized at the plasma membrane and extruded into the periplasm, mannoprotein is synthesized in the cytoplasm and transported through the secretory pathway, and -1,6-glucan synthesis may occur partially in the endoplasmic reticulum or Golgi complex (268). Not all components are necessarily present in a complex; therefore, the authors suggest that more chitin may be present in inner cell wall layers and more mannoprotein may be present in the outer layers.

Layering. Since polysaccharides are poorly reactive to the ordinary fixatives and stains used for transmission electron microscopy, only a few well-defined ultrastructural details are obtained by conventional protocols (Fig. 1A and B). However, transmission electron microscopy studies performed with more special techniques or with cytochemical stains and contrasting agents show several layers in the cell wall of C. albicans (Fig. 1C to F). The appearance of these layers is variable and seems to be related to the strain examined, growth conditions, morphology, and preparation of the specimens (50, 64, 422). Thus, there is no consensus about the number of layers present in the cell wall. Different authors have reported the presence of three to eight different layers (28, 64, 186, 421, 438). The outer cell wall layer appears as a dense network with a fibrillar or flocculent aspect (64, 480), whereas the inner wall layer appears contiguous with the plasmalemma with extensive membrane invaginations involved in anchoring of the cell wall to the membrane (192, 276). The microfibrillar polysaccharides glucan and chitin, the components that supply rigidity to the overall wall structure, appear to be more concentrated in the inner cell wall layer, adjacent to the plasma membrane. In contrast, proteins and mannoproteins appear to be dominant in the outermost cell wall layer (Fig. 1B), although they are also present through the entire wall and at the inner regions of the cell wall. Some of the latter proteins may be covalently associated with glucans. Evidence from several cytochemical and cytological studies indicate that the cell wall layering may be due to the distribution of mannoproteins at various levels within the wall structure (64). In any case, it seems clear that layering may be the result of quantitative differences in the proportions of the individual wall components (-glucans, chitin, and mannoproteins) in each layer rather than of qualitative differences (389).

View larger version (123K): [in this window] [in a new window]   FIG. 1.   Cell wall structure. (A) Transmission electron micrograph of a section of a C. albicans cell prepared by freeze-substitution, showing the cell wall as a thick, electron-dense, homogeneous structure. The presence of distinct layers was not evident in this preparation. Bar, 1 µm. Reprinted from reference 3 with permission of the publisher. (B) Thin sections of cells treated with gold-conjugated concanavalin A, showing an intense labeling with gold particles of the external wall surface. The surface exhibits a fibrillar appearance (arrows), suggesting that concanavalin A-reactive cell wall components, i.e., mannoproteins, are particularly abundant at the most external wall layers. The remaining wall structure also appeared as a homogenous structure in this transmission electron micrograph. Bar, 0.5 µm. Reprinted from reference 549 with permission of the American Society for Microbiology. (C) Other procedures for transmission electron microscopy examination of thin sections of C. albicans cells revealed more clearly the presence of an outer floccular layer (arrow) and showed that the remaining cell wall structure is not homogeneous and that some layering exists. Bar, 200 nm. Reprinted from reference 240 with permission of the publisher. (D to F) Complexity of the wall ultrastructure and presence of distinct layers in the cell wall of C. albicans as revealed by different scanning electron microscopy-based procedures such as cryo-scanning electron microscopy (D) and freeze-fracture, freeze-etch analysis (E and F). The presence of well-ordered, regularly arranged, radiating fibrils in the outer layer is particularly evident in the micrographs shown in panels E (hydrophilic cells) and F (hydrophobic cells). Bar, 0.3 µm in both panels. Panel D reprinted from reference 246 with permission of the publisher. Panels E and F reprinted from reference 191 with permission of the publisher.

Fimbriae. The outer cell wall layer that is composed mainly of mannoproteins appears as a dense network of radially projecting fibrils (28, 64), designated fimbriae (154, 583). These fibrils extend for 100 to 300 nm (190, 276) and are approximately 5 nm in diameter (28). Both filamentous forms and blastospores exhibit this characteristic feature (28). C. albicans fimbriae consist of many subunits assembled through noncovalent hydrophobic interactions (583). The major structural subunit of fimbriae is a glycoprotein with an apparent molecular mass of 66 kDa, while the unglycosylated protein has an approximate molecular mass of 8.64 kDa (583). In crude extracts, in addition to the 66-kDa moiety, components migrating with an electrophoretic mobility equivalent to proteins of 54, 47, and 39 kDa reacted with monoclonal antibodies (MAbs) raised against purified fimbriae, suggesting the presence of species with differing degrees of glycosylation (585). The hydrophobic status of the cells profoundly affects fimbrial structure. Hydrophilic cells have long, compact, evenly distributed fibrils, while hydrophobic cells have short, blunt fibrils (190) (Fig. 1E and F). The overall hydrophilic status may be due to masking of hydrophobic components by hydrophilic surface fibrils (190). Fimbrial components mediate the adherence of C. albicans to glycosphingolipid receptors on human epithelial cells (279, 583, 584, 586) as discussed later.

CELL WALL PROTEINS

Top Previous Next References

Cell Wall versus Secreted Proteins

Should location or function determine the classification of a cell wall protein or secreted protein? Are proteins either one or the other? Proteins that are found in the in vitro growth medium are often called secreted or extracellular proteins. To reach this location, these proteins travel through the cell wall, where they coexist with cell wall-bound moieties and by location are proteins that contribute the total cell wall proteinaceous component. However, it seems reasonable to consider transiently associated proteins with an environmental destination as being secreted. On the other hand, how do we classify a protein when the association does not appear to be transient or when the postulated function of the protein is within the cell wall? There are proteins that are cell associated under one growth condition and secreted under another (519). Cytochemical detection of phospholipase activity shows a localized cell wall location in cells grown on yeast extract medium and a development of a more generalized cell wall, cell surface, and secreted localization in cells in contact with the chorioallantoic membrane (425, 427). There are enzymes recovered from culture supernatants whose functions are thought to be cell wall biosynthesis and remodeling (75). For most proteins, extracellular locations have not all been examined, so that some proteins that are reported as cell wall associated might also be found in culture fluids if examined. Several cell wall components that are not thought to be secreted have been detected in supernatants of C. albicans cell cultures (2, 178, 299, 536, 543). The relationship of some of these moieties with the cell wall structure is unclear. They may come from the outer wall layers. Alternatively, they may be released by lysed cells or as a consequence of the controlled degradation of the cell wall structure, required for wall expansion during growth. One of the criteria that has been used to demonstrate a cell surface location is binding of a ligand or antibody. When similar observations are made with extracellular proteins such as secreted acid proteinase (397) or phospholipase (425, 427), should this finding be differently interpreted, particularly when there may be a function for the cell associated protein (147, 433)? This discussion makes clear that for some proteins, classification as cell wall or secreted may be dependent upon the growth conditions of the organism, the conditions under which localization has been examined, and our view of the function of the protein. As the mechanisms of targeting proteins to subcellular locations are elucidated and protein functions more completely examined, these issues should be resolved. In this review, we have included both proteins whose function is thought to be in the cell wall and those whose role is thought be primarily extracellular.

In the next four sections, we consider general questions of cell wall protein extraction, protein composition of the extracts, protein modification, and distribution of proteins within the cell wall. In these studies, the emphasis has been on definition of cell wall protein by location, since the principal concern has been the removal of a cytoplasmic contribution to the extract. In the following sections, we review specific proteins that have been grouped by various functional and identity relationships.

Different techniques have been used to extract cell wall components of C. albicans. These include physical, chemical, and enzymatic methods and a combination of them. The choice of extracting reagents and techniques, the sequence of extraction methods, and the use of either intact cells or purified cell walls as the starting material may affect both qualitative and quantitative solubilization of cell wall components. In general, and due to the insolubility of both chitin and glucans, sequential alkali and acid treatments are required to effect their extraction (140, 141). In early studies, mannans were extracted from whole cells or isolated cell walls by alkali treatment and further precipitated with Fehling's solution as a copper complex (408). A milder procedure involved mannan extraction from cells resuspended in citrate buffer (pH 7) by autoclaving and further purification by precipitation with Fehling's solution (405) or with Cetavlon (66, 376, 392). This topic is covered more extensively in other reviews (141).

Proteinaceous components have been extracted or solubilized from the cell wall of C. albicans by a variety of techniques. Most of the studies have involved either detergents (such as sodium dodecyl sulfate or n-octylglucoside), reducing agents (such as dithiothreitol [DTT] and -mercaptoethanol [ME]), or hydrolases (such as proteases, Zymolyase, or other -glucanases, and chitinases) to release proteins from both isolated cell walls and intact cells (57, 59, 61, 71, 78, 122, 125, 319, 365, 414, 508, 524, 525). These reagents have been used alone or in combination. Although sulfhydryl compounds such as ME appear to be less efficient that hydrolases such as Zymolyase in releasing cell wall-bound proteins and glycoproteins (295), these chemical agents solubilize a complex array of proteinaceous components from the walls of intact C. albicans cells (57, 58). On the other hand, some -glucanases used to solubilize cell wall moieties actually are enzymatic complexes that may contain other unidentified or uncontrolled hydrolytic activities, which may alter the native characteristics of the released molecules.

Other extraction procedures have been less frequently reported and include chemical, enzymatic, and physical methods alone or in combination. -Elimination with NaOH has been used to release putative structural proteins (369). Ethylenediamine has also been used in structural studies to extract proteins (455). Salt (NaCl) was used to extract the surface determinant of a MAb (40) and a surface adhesin (225). Homogenization has been used to shear fimbriae (583), and -mannosidase treatment followed by sonication has been used to release wall antigens (195).

There has been little comparison of the various methods used to extract the wall proteins. Casanova and Chaffin (57) compared five extracts for yeast cells and germ tubes: (i) ME extract of intact cells at alkaline pH; (ii) the Zymolyase extract of the treated cells; (iii) ME extract of isolated cell walls at alkaline pH; (iv) the Zymolyase extract of the treated cell walls; and (v) a sodium dodecyl sulfate (SDS)-ME extract of isolated cell walls. The extracts were examined by blotting with concanavalin A, two MAbs (MAb 4C12 to a high-molecular-weight component of germ tubes [59] and MAb 24.17 to a mannan epitope of a high-molecular-weight component [72]) and antiserum for factor 6. The authors concluded that the two sequential extracts obtained from intact cells were most satisfactory. In addition, although extraction with reducing agents is frequently thought to release medium to small components from the cell wall, this study showed that ME also released the high-molecular-weight components. These appeared to be larger than the same component present in Zymolyase extracts. ME and other reducing reagents are believed to solubilize mainly components associated with the outermost layers of the cell wall (50, 61, 295). These reagents also increase cell wall porosity and facilitate subsequent action of cell wall degrading enzymes (103, 589). The hydrolysis of glucan by Zymolyase or glucanases may release proteins enmeshed or covalently attached to the glucan. Proteins covalently attached to glucan are postulated to represent species contributing to cell wall structure (59, 122, 123, 125, 482). Covalent attachment of mannoprotein to glucan, perhaps through phosphodiester linkages, has been suggested, as noted below (237, 238, 462, 463).

A valid question is whether the proteins found in these extracts are genuine cell wall components. It has been suggested that treatment of intact C. albicans cells with reducing agents (DTT or ME) may release some intracellular macromolecular components (50). This question has been examined most thoroughly for extracts obtained with ME. As discussed later in this review, receptors or binding proteins for ligands that bind to the intact cell are found in such extracts. On the other hand, several proteins previously associated with a cytoplasmic function have also been found in the wall extracts (6, 8, 161, 303, 520). These observations led to additional experiments using different approaches to demonstrate that the proteins found in the extract were genuine wall components. Transmission electron microscopy demonstrated that each of the moieties was indeed present in the cell wall, including the cell wall interior (6, 8, 303). Chaffin and colleagues (6, 303) also used a more general method to identify genuine cell wall proteins. Intact cells were treated with a derivative of biotin that does not permeate the membrane and therefore does not label cytoplasmic proteins. The extracted biotinylated proteins included those previously thought to be confined to the cytoplasm. This demonstrated that the proteins were present in the cell wall prior to extraction and that their presence in the extract was not due to cytoplasmic contamination. Support for the validity of the extraction procedure was also obtained with parental and mutant strains of S. cerevisiae (294). Two members of the Ssa family of proteins (Ssa1p and Ssa2p) were detected in the cell wall and cytoplasm of the parental strain, whereas in the mutant strain missing these two proteins the remaining members of the family were detected in the cytoplasm but not in the cell wall. The failure to find Ssa proteins in the cell wall of the mutant strain demonstrated that the cell wall extract was not contaminated with the Ssa proteins of the cytoplasm. Hence, current evidence indicates that treatment with sulfhydryl compounds is a suitable method to release autochthonous cell wall protein and glycoprotein components without substantially altering their biological characteristics (6, 60, 61, 293, 295, 296, 300, 302, 377).

Analysis of the protein and glycoprotein constituents solubilized from isolated cell wall preparations and from intact cells of both candidal growth forms by different treatments has revealed both a complex array of protein-containing components and quantitative and qualitative differences in the protein composition of yeast and mycelial cell walls (57-59, 71, 125, 295, 319, 320, 324, 363, 414-416, 524) (Fig. 2). Some components have been characterized as high-molecular-weight mannoproteins (HMWM). The identity of these proteins may vary with the morphology of the organism. Several HMWM are released by treatment with -glucanases and may be covalently attached to structural polysaccharides. These HMWM may play an important role in modulating the organization of the different cell wall constituents to obtain the final supramolecular structure of the wall specific for each C. albicans morphology. In addition, HMWM contain large amounts of carbohydrate and consequently could be major elicitors of anticandidal host immunity (59, 62, 123, 159, 160, 297, 301, 326, 327, 415, 479, 523). In the medium- to low-molecular-weight range, from 20 to more than 40 polypeptide species (depending on the study considered) have been identified (61, 71, 125) (Fig. 2). As discussed above, the evidence suggests that these proteins are bona fide cell wall constituents.

View larger version (64K): [in this window] [in a new window]   FIG. 2.   Polypeptide composition of cell wall extracts and whole protoplast homogenates from C. albicans as revealed by different electrophoretic techniques (A and B). C. albicans cells from which samples were obtained were incubated in the presence of 14C-labelled protein hydrolysate and subsequently tagged with biotin. Double-labelled cell wall proteins and glycoproteins were extracted from intact blastoconidia (lanes 1 and 3) and germinated blastoconidia (lanes 2 and 4) by sequential treatment with ME (lanes 1 and 2) and digestion with Zymolyase 20T (lanes 3 and 4). Samples of protoplast homogenates from blastoconidia and germinated blastoconidia were run in lanes 5 and 6, respectively. Polypeptides were separated by SDS-PAGE and detected by fluorography (A) or transferred to nitrocellulose and detected with an avidin-peroxidase conjugate (B; lane S in this panel shows a mixture of prestained molecular weight standards run in parallel). Numbers and letters are used to identify and compare bands detected in the cell wall extracts by the different experimental procedures used. Although qualitative differences were observed (i.e., some polypeptides exhibited a strong radioactive label but were weakly biotinylated [band 6; star]), surface labelling of cells with biotin appeared to be a suitable technique to detect proteins in the wall of C. albicans. Thus, from the complex polypeptide pattern found in the protoplast homogenate samples as revealed by fluorography (A, lanes 5 and 6), only few species were labelled with biotin, indicating that most proteins released by ME and Zymolyase from intact cells are bona fide cell wall components. Brackets indicate a cluster of bands within a molecular weight range where many candidal moieties that represent receptors for host ligands have been identified (see the text). Reprinted from reference 61 with permission of the American Society for Microbiology. (C) The complexity of the polypeptide pattern of the cell wall extracts was clearly evidenced when analysis was performed by two-dimensional PAGE and silver staining (the polypeptide pattern shown corresponds to the ME extract from blastoconidia). Reprinted from reference 486 with permission of the American Society for Microbiology.

There is a growing body of experimental evidence indicating that the propertiesexpression, distribution, and chemical characteristicsof cell wall proteins and glycoproteins observed in vitro and in vivo are dependent on multiple factors. These include growth conditions, organism-related factors (such as growth state, morphology of the cells, strain and serotype, phenotypic switching, cell surface hydrophobic or hydrophilic status), and the nature of the biological specimens (intact cells or isolated wall preparations) that are subjected to analysis (5, 7, 24, 44, 57, 63, 106, 159, 190, 191, 195, 210, 284, 297, 301, 326, 416, 422, 477, 513, 529). Iron availability, which has been shown to be important for pathogens in establishing infection (46, 400, 566), affects the cell surface (529). There are quantitative but not qualitative changes in the profile of surface proteins associated with growth at different iron concentrations. Yeast cells of most strains grown in limiting or excess iron do not adhere as well to human buccal epithelial cells as do organisms grown at intermediate concentrations that support optimum growth. The effects of growth conditions on the expression of specific proteins are discussed for each protein in later sections. The cell wall may be thus envisaged as a highly dynamic "organelle." The fungus is capable of expressing differentially variable wall constituents that may be useful for switching between commensal and pathogenic lifestyles and for modulating and/or evading the immune host defense.

Glycosylation. Without any doubt, glycosylation is the most important modification of the proteins in the fungal cell wall. Attachment of sugar moieties to proteins results in the formation of the glycoproteins, which in the case of C. albicans are mainly mannoproteins. The general features of mannoproteins have been discussed above. However, the presence of nonglycosylated proteins has also been found in the cell wall of C. albicans (6, 8, 58, 61, 161). Mannosylated proteins can be broadly divided into two classes. The HMWM, most of which are larger than 200 kDa, are postulated to have structural functions within the cell wall. Some medium- to low-molecular-weight proteins also react with concanavalin A, indicating their mannan content. Within this broad division, the amount of carbohydrate attached to the same polypeptide may vary (59, 122, 364). Within the high-molecular-weight class, there is a difference in the size of the side chains associated with morphology. Oligosaccharides obtained from mannoproteins from yeast cells average 600 residues, and those from germ tubes average 300 residues (122). Cell wall proteins may also be O glycosylated with unbranched mannose chains containing one to a few residues (412). Elorza et al. (123) have suggested that some proteins are initially secreted as O-glycosylated proteins and become cross-linked with glucan and/or other N-glycosylated proteins only after incorporation into the cell wall structure.

Phosphorylation. Phosphorus is a minor component of the cell wall of C. albicans (77). It has been assumed that it is present in cell wall mannoproteins in phosphodiester linkages between mannose residues (17, 454). Bulk mannan from C. albicans can be fractionated into five fractions that differ in the amount of phosphate (393). Phosphomannoprotein complexes from cells of both C. albicans A and B serotypes have been characterized (261-263, 494-498). This material contained 0.9 to 1.6% phosphate depending on the morphology and strain considered, and the authors concluded that -(1,2) oligomannosaccharides were attached by phosphodiester linkage to other branching moieties. -(1,2) oligomannosidic epitopes were further observed on a C. albicans 14- to 18-kDa phospholipomannan moiety (545), a glycolipid with important immunologic properties (133-135, 228, 229, 424), whose mannose residues may be added differently from mannan (546). Not all the glycoproteins in the cell wall of C. albicans contain phosphate, and some proteins may contain phosphate but not carbohydrate (58).

Ubiquitination. Ubiquitin is a small (approximately 8,500-Da) polypeptide first isolated from bovine thymus (167). Sequence analysis of various ubiquitin genes has revealed striking evolutionary conservation among species (138). Ubiquitin plays important roles in protein modification, protein degradation, gene transcription, organization of chromatin structure, and stress resistance in higher eukaryotes (138, 139, 197). It is also associated with some cell surface protein and signaling functions (69, 368, 404, 504, 553). In yeast, a role for ubiquitination in endocytosis and/or turnover of plasma membrane protein receptors including -receptor has been demonstrated (198, 269, 442). The C terminus of ubiquitin is covalently attached to -amino groups of lysine in protein substrates by an enzymatic conjugation system. A large number of enzymes responsible for the formation and processing of ubiquitin-protein conjugates have been described (138), including in C. albicans (98). We have cloned a polyubiquitin gene (UBI1) of C. albicans that contains three tandem copies, head-to-tail spacerless repeats, of the sequence coding for the 76 amino acids of the ubiquitin protein (485). The gene has also been cloned from a different strain (15). Northern blot analysis revealed a single mRNA population of about 1 kb present in similar amounts in both yeast and mycelial cells (485). Indirect immunofluorescence demonstrated that ubiquitin determinants were located on the cell surface, and Western blot analysis of a ME extract demonstrated that several cell wall proteins contained ubiquitin-like epitopes. The cell wall species that are ubiquitinated are discussed below with the individual proteins. The role of ubiquitin in the cell wall whether in protein degradation, stress protection, or perhaps even modulation of activity of receptor-like molecules remains to be assessed.

As described above, transmission electron microscopy studies of the C. albicans cell wall show the existence of several layers. The structural appearance of these layers is variable and seems to be related to the strain examined, the growth conditions, the morphology (yeast cells or germ tubes) exhibited by the microorganism, and the sample preparation protocol (50, 64, 422) (Fig. 1). After total removal of proteins and mannoproteins by treatment with strong alkali or heating, the cell wall appeared to be significantly thinner with loss of any appreciable layering. These structural changes were paralleled by the absence of electron-dense components detectable with ordinary electron microscopy dyes, concanavalin A binding sites, and positive staining with reagents specific for mannoproteins (periodate-silver). Thus, the cell wall layering appears attributable to the distribution of mannoproteins at various levels within the wall structure (64). Treatment of cells with sulfhydryl agents and hydrolytic enzymes, coupled with specific cytochemical staining, has consistently shown that mannoprotein constituents are preferentially located at the outermost layer of the wall of C. albicans cells. The surface has a fibrillar or flocculent aspect with thin, delicate filaments or fimbriae (Fig. 1; see below). This material is present mostly in virulent strains, and it is also more abundant in isolates exhibiting increased adherence to host tissues (64). Proteins whose biological function is at the cell surface or the extracellular environment may nevertheless be found at the innermost layer of the wall and through the wall as they travel from the plasma membrane and periplasmic space to their destination (478). Thus, some proteins destined for the extracellular environment may also be obtained in cell wall extracts. Evidence for this mannoprotein traffic in C. albicans has been reported (423). Immunoelectron microscopy shows that proteins on the cell surface visualized by indirect immunofluorescence are also detected at the cell surface as well as near the plasma membrane with some protein distributed through the wall (6, 303). On the other hand, a protein that is not detected at the cell surface by indirect immunofluorescence but is present in cell wall extracts is located only in the interior of the wall (8). Using a panel of MAbs to localize proteins, Pontón et al. (416) demonstrated various distribution classes of cell wall proteins: (i) expressed only on the germ tube surface, (ii) expressed on the germ tube surface and within the yeast cell wall, (iii) expressed on both yeast cell and germ tube surfaces, and (iv) expressed within the wall of both germ tubes and yeast cells. A fifth category, expressed only on yeast surfaces, is also reported (296). Proteins that are associated with -glucans should be concentrated near the plasma membrane with the structural polysaccharide. In any case, since asymmetry in mannoprotein distribution is evident in C. albicans walls (64), layering is more likely to be the result of quantitative differences in the proportions of the individual wall components (-glucans, chitin, and mannoproteins) in each layer rather than qualitative differences (389).

Differences in the distribution of proteins and glycoproteins at the cell surface are also noted. This asymmetry may be related to the physiologic role played by each particular moiety. High-molecular-weight mannoproteins that may play an important and active morphogenetic role in modulating the organization of the cell wall (59, 62, 123, 160, 326, 327, 480) are homogeneously distributed on the cell surface (59, 62, 159). Some proteins and mannoprotein moieties that are receptors for different host ligands exhibit clustering or asymmetric cell surface distribution (60, 296, 328). The distribution of common or morphology-associated, homogeneously or heterogeneously distributed cell surface antigens of C. albicans as revealed by immunofluorescence microscopy is shown in Fig. 3.

View larger version (135K): [in this window] [in a new window]   FIG. 3.   Surface expression of cell wall proteins. Phase-contrast (A and C) and immunofluorescence (B, D, and E to I) of C. albicans blastoconidia (B) and mycelial filaments (M) reacted with different polyclonal and monoclonal antibody preparations raised to protein and glycoprotein cell wall constituents. Some antibodies recognized antigens that appeared to be specific for or preferentially expressed in germ tubes (A and D) or blastoconidia (E and F). Arrows in panels A to D point to the location of mother blastoconidia (A and C) that exhibited no fluorescence (B and D). Some antigens appeared to be homogenously distributed on the surface of mycelial filaments (B and D) or blastoconidia (G). However, patches of greater fluorescence intensity were observed with other antisera preparations (H and I), suggesting that antigens recognized by such antisera were heterogenously and randomly distributed within the cell wall structure. The pictures in panels B and D to G are from standard immunofluorescence microscopy observations. Panels H and I show single-focal-plane sections of different cells obtained by confocal fluorescence microscopy and associated software. Bar, 10 µm (except for panel G, which is 1.25 µm). Panels A to D reprinted from reference 59 with permission of the American Society for Microbiology. Panels E and F reprinted from reference 24 with permission of the American Society for Microbiology. Panel G reprinted from reference 72 with permission of the American Society for Microbiology. Panels H and I reprinted from reference 328 with permission of the American Society for Microbiology.

At least three cell moieties appear to have posttranslational regulation of their localization in the cell wall. A MAb that recognized a hyphal surface protein detected a smaller protein in the plasma membrane of yeast cells (394). The C3d binding protein was detected as a 60-kDa moiety in germ tube cell walls, while a 50-kDa component was found in yeast cell membranes (563). The third example is a 30-kDa protein found in the cell wall of germ tubes but not yeast cells (5). However, the gene is expressed in yeast cells and presumably translated, although the cellular location of the protein is not known. Another group of proteins to be discussed below, i.e., enolase, hsp70, 3-phosphoglycerate kinase, and glyceraldehyde-3-phosphate dehydrogenase (GAPDH), well-known cytoplasmic proteins, have recently been described as also present in the cell wall. They may represent proteins with regulation of partition between two cellular compartmentscytoplasm and cell wall.

Other proteins, such as hydrolytic enzymes [discussed below], are altered in their expression by environmental substrates. Another example of an apparently environmentally regulated protein is the 58-kDa fibrinogen binding protein that is expressed by cells growing on Lee medium but not by cells growing on yeast extract-peptone-glucose medium (5). The foregoing discussion makes clear that the expression and localization of proteins in the cell wall are complex and dynamic processes. The presence and location of proteins and glycoproteins in the cell wall are likely to be affected by several mechanisms including gene expression, posttranslational regulation, subcellular partitioning, and targeting of destination within the cell wall.

As discussed above, a number of hydrolytic enzymes have been recovered from both cell-associated locations (cell wall and periplasm) and culture medium whose function is postulated to be within the cell wall (Table 1). These enzymes are thought to be involved in cell wall biosynthesis or the remodeling that accompanies growth and division of cells.

Exo--(1,3)-glucanase. Secretory exo--glucan hydrolases (-glucanases or -glucosidases) are widely occurring enzymes in many yeast and fungal species. Although the exact physiological roles of these enzymes are unknown (141, 386), they participate in the metabolism of -glucan, which is the main structural microfibrillar polymer of the cell wall in C. albicans (64). The most widely accepted biological role of glucanases is limited hydrolysis of cell wall glucan during morphogenetic events (141, 386). -Glucanases have been described to be associated with the C. albicans cell wall (387, 428).

-1,3-Glucan transferase. Many yeasts, including C. albicans and S. cerevisiae, contain a highly conserved protein with a size ranging from 31.5 to 34 kDa depending on the species (196). It is a major cell wall mannoprotein in S. cerevisiae (466). In C. albicans, a protein of 34 kDa is secreted by protoplasts and observed as an aggregate in gel filtration (122). When released by Zymolyase, it eluted as a low-molecular-weight species. The protein appeared to have a single N-linked oligosaccharide of 2.5 kDa and a residual protein moiety of 31.5 kDa. A cell wall protein with an approximate molecular mass of 34 kDa was isolated as a by-product during purification of an endo-(1,3)--glucanase from the material secreted to the medium by C. albicans (178). The enzyme displayed a unique glucanosyl transferase activity and did not contain any exo- or endo--glucanase activity. The authors suggested that this 34-kDa protein is a glucan-branching enzyme responsible for the transformation of the initial linear -(1,3)-glucan into the branched -(1,3)--(1,6)-glucan that is found in the cell wall of the fungus. The C. albicans BGL2 gene encoding the -1,3 glucan transferase has been cloned (471) and is similar to the S. cerevisiae BGL2 gene (245). The S. cerevisiae enzyme has been described as an exoglucanase (245) and as an endoglucanase (373). However, more recently, the S. cerevisiae enzyme has been shown to be homologous to the sequence of the C. albicans enzyme and corresponds to Bgl2p (165). More sensitive assays revealed that at low concentrations of glucose oligosaccharides, glucanase activity was observed, while at higher concentrations, glucosyltransferase activity predominated. The enzyme transfers -1,3 glucan oligosaccharides from a donor -1,3 glucan to an acceptor -1,3 glucan, forming a linear polymer joined by a -1,6 linkage (587). This activity would permit the enzyme to participate in cross-linking or repair of glucan within the cell wall.

Chitinases. Chitinases are produced by many organisms including chitin-containing organisms that produce both chitin synthases and chitinases. In C. albicans, the most likely role for these enzymes, like glucanases, is limited hydrolysis of cell wall chitin during morphogenetic events (141). C. albicans contains three chitinase genes, CHT1, CHT2, and CHT3 (346). More than half the chitinase activity detectable in whole-cell extracts is associated with secreted enzyme (periplasmic and cell wall) (170). Reactivity is detected in supernatants of washed cells and with intact cells by using a substrate that does not enter the cell (346). Chitinase production increases during exponential growth and is greater in cells grown on yeast extract-peptone-dextrose (YEPD) medium than on a minimal medium. One of the roles of chitinase in S. cerevisiae, where chitin is confined primarily to the septum, is mother-daughter separation. Disruption of S. cerevisiae CTS1 results in clumps of cells due to failure of the cells to separate (273). Treatment of C. albicans yeast cells with a chitinase inhibitor also leads to inhibition of cell separation and clumps of cells (170).

-N-Acetylglucosaminidase. Production of -N-acetylglucosaminidase by C. albicans cells is induced by the presence of GlcNAc in the medium (519). GlcNAc, the -N-acetylglucosaminidase reaction product, also induces the synthesis of other enzymes required for the metabolism of this amino sugar (171, 492, 519). The enzyme may function as a chitobiase that, in conjunction with chitinase, completes the hydrolysis of chitin to provide both nitrogen and carbon sources from chitin. The enzyme, for which two molecular forms with different glycosylation levels appear to exist (364), is secreted and deposited into different regions of the cell envelope of both the yeast and mycelial forms (365, 426, 519). The enzyme is also released to the culture medium of both growth forms but to a greater extent (at least fourfold greater) during hyphal formation (519). It is suggested that the germ tube wall is more porous than that of yeast cells and, consequently, that release of extracellular enzymes to the medium is facilitated by the hyphal formation process. -N-Acetylglucosaminidase acts on a number of natural and synthetic substrates including diacetylchitobiose and triacetylchitotriose. In its native form, the enzyme appears to be a monomer with a molecular mass of 66 kDa (519).

Glutaminyl-peptide--glutamylyl-transferase. The activity of the enzyme glutaminyl-peptide--glutamylyl-transferase (transglutaminase) has been detected in cell extracts obtained from both growth phases of C. albicans. The activity was associated mostly with the cell wall fraction, whereas the cytosol contained almost negligible amounts of enzyme activity (456). This distribution suggested an extracellular (cell wall-bound) location for transglutaminase in intact cells. Although the wall component displaying this enzymatic activity has not been characterized, the postulated transglutaminase activity could be involved in the formation of covalent bonds between different wall proteinaceous moieties (456). In this context, it has been suggested that formation of such covalent linkages could play a role in maintaining the spatial organization of the cell wall during the biogenesis and assembly of this structure.

The enzymes in the previous section are postulated to find substrates and to have their primary function within the cell wall. This section considers enzymes whose substrates are not associated with the cell wall but are found in the environment (Table 1). The action of these enzymes may provide access to nutrients for the organism. When hydrolysis of these substrates or action of extracellular proteins affects the function and viability of the host, the enzymes may be considered virulence factors that contribute to the establishment of infection. These proteins are at least transiently associated with the cell wall during their translocation across the cell wall to the external environment. However, in some cases, the association is less transient. The distribution of some of these enzymes is variable and, under some growth conditions, may be primarily cell associated in the periplasm and cell wall. Some of these proteins have also been localized to the cell surface by the same criteria that have established a cell surface location for adhesins whose function is at the cell surface. One of these extracellular proteins, Sap, also found at the cell surface, may contribute to adherence (209, 406). We have also included in this section a discussion of a secreted protein with immunomodulation properties whose relationship with other secreted moieties has not been established. Additional studies of the function, localization, and molecular mechanisms of targeting proteins to specific locations will help resolve questions about the cell-associated and extracellular distribution of these proteins.

Acid proteinase. The extracellular proteolytic activity is one of the several hydrolytic enzyme activities described for C. albicans (390) and is due to aspartyl proteinase enzymes. The candidal secreted aspartyl proteinase was first identified by Staib (514, 515), and over the last 25 years its biological characteristics, including its role as a potential virulence factor of C. albicans, have been studied by a number of laboratories. The proteolytic activity is associated with a 42- to 45-kDa acid carboxyl proteinase. The enzyme has broad substrate specificity and is active in the range of pH 2.0 to 7.0, with the pH optimum varying from 2.5 to 4.5 depending on the substrate (109, 451). As discussed below, there is a family of aspartyl proteinases whose expression appears to be regulated by the strain, cell type, and environment. As a protease, the enzyme may have a variety of substrates, and these substrates may vary depending upon the host organ, e.g., skin or blood, that is colonized or infected. This potential spectrum of substrates and expression of isoenzymes may account for the different roles that have been postulated for the enzyme as a virulence factor. Purification of Sap1, Sap2, and Sap3 showed isoenzyme differences in antigenic similarity, thermal stability, and activity at low pH (509). The crystal structure of the enzyme showed details of the binding site, suggesting the possibility of structural differences among isoenzymes that might affect substrate specificities (1, 94). The techniques and strains that are being developed and constructed and understanding isoenzyme differences should contribute to elucidation of the role of the enzyme in colonization and infection.

Phospholipase. Extracellular (secreted) phospholipase activities that have been reported for C. albicans include phospholipases A, B, and C (19, 23, 174, 217, 425, 427, 533). C. albicans also secretes lysophospholipase and lysophospholipase-transacylase (19, 23, 174, 217, 355, 425, 427, 533, 534); however, these activities appear to be associated with the same enzyme (355, 533, 534). Phospholipase A and lysophospholipase activities have been found in the cell wall of yeast cells and hyphae by cytochemical techniques (425, 427). Enzyme activity was associated more closely with the walls of older yeast cells than of younger cells and was more prominent at the tip of growing hyphae than in the lateral walls (425, 427). When yeast cells and germ tubes were grown in the same medium but at a different pH, the specific activity of extracellular phospholipase A was similar for yeast cells and germ tubes while that of lysophospholipase was higher for the yeast form (174).

Esterase. Esterase activity has been found in a variety of pathogenic Candida species (56, 419, 452). Recently, induction of an extracellular esterase from C. albicans in culture media containing different Tweens (polyoxyethylenesorbitan compounds) as the sole carbon source was reported (551). Levels of enzyme activity correlated with fungal growth and substrate concentration. Activity was stimulated by activators of lipase activity. The induced esterase was heat labile and had maximum activity at pH 5.5. Thin-layer chromatography of the reaction products suggested that the enzyme is a monoester hydrolase and thus not a lipase in strict sense. Esterase activity was present in a variety of Candida clinical isolates, with no apparent correlation between enzymatic activity and relative pathogenicity. Its biological significance remains to be established.

Glucoamylase. The C. albicans gene encoding secreted glucoamylase [-(1,4)-D-glucan glucohydrolase] has been cloned by conferring on S. cerevisiae cells the ability to grow on media containing starch as the sole carbon source. The enzyme is efficiently secreted in the heterologous host (82). Coding and promoter regions in the sequence were identified. The gene, which contains no introns, codes for a protein of 946 amino acids with a deduced molecular mass of 105 kDa. However, there are 17 potential sites for N glycosylation, which may explain the fact that the partially purified enzyme has an apparent molecular mass of 200 kDa. Importantly, the first 20 amino acid residues are characteristic of a typical signal sequence found in secreted proteins. Because of the highly efficient translation and secretion of the gene product, this gene promoter and its signal sequence have been used for the expression of heterologous proteins in S. cerevisiae (292). The sequence was also used in vaccine development for the expression of glycoprotein antigens (83).

Hemolytic factor. The ability of pathogenic organisms to acquire iron in the mammalian host has been shown to be critical in establishing infection (46, 400, 566). Iron availability is a limiting growth factor for C. albicans in human serum, where iron may be sequestered by transferrin (118, 367). In this context, it has been recently reported that C. albicans exhibits hemolytic activity when grown on glucose-enriched blood agar (317). The activity is present on intact organisms, and hyphae display greater activity than yeast cells. The factor is also secreted into the culture medium. However, the identity of the cell surface moiety responsible for this activity remains to be characterized. In any case, this hemolytic activity may be a putative virulence factor for C. albicans, allowing iron acquisition from hemoglobin released from lysed host erythrocytes and restoring the ability to grow in human serum.

Acid phosphatase. Among the catalytic mannoproteins detected outside the plasma membrane barrier in C. albicans, acid phosphatase was one of the first to be characterized. It is a candidal hydrolase, for which some role in the pathogenesis of candidiasis has been suggested but not confirmed (390, 493). C. albicans acid phosphatase is an inducible enzyme that has been purified to homogeneity; the purified enzyme is a 125- to 130-kDa mannoprotein with a pH optimum of 3.6 to 4.5 (78, 391). The location (or distribution) within the cell wall structure of acid phosphatase in C. albicans (78, 105, 547) is similar to that of acid phosphatase in S. cerevisiae, where the enzyme was found to be located in the outermost and innermost cell wall layers (137, 290).

Miscellaneous. Lipase activity is secreted by C. albicans cells growing in medium containing Tweens 80, 60, 40, and 20 as the carbon source (148). A candidal gene, LIP1, conferring lipase activity on S. cerevisiae has been isolated. The deduced sequence contained 351 amino acids with a predicted molecular mass of 38 kDa and five potential N-glycosylation sites. Southern blot analysis at low stringency suggested the presence of a lipase family.

C. albicans is generally considered a dimorphic organism, capable of reproducing by budding, leading to the formation of yeast cells, or by production of germ tubes, resulting in filamentous growth (mycelium). Since the cell wall is the structure ultimately responsible for shape, and since the final steps of cell wall assemblage occur externally to the plasma membrane, the research efforts of a number of investigators in different laboratories have concentrated in the study of this morphogenetic event, as well as on the characterization of cell wall components (especially proteins and mannoproteins) that are growth phase specific. These studies have been fueled by four important considerations: (i) the interest in morphogenesis itself as an important biological phenomenon, (ii) the commonly postulated relationship between growth in the mycelial form and pathogenicity, (iii) the coordination between expression of adhesion factors and germ tube formation, and (iv) the idea that characterization of mycelium-specific antigens would provide the basis for the development of improved serodiagnostic test for the detection of invasive candidiasis (50, 64, 79, 96, 209, 352, 416, 418, 453, 480).

However, characterization of morphology-specific cell wall components is complicated by the fact that expression of cell wall components is a dynamic process. This process is influenced by environmental and nutritional factors and varies among strains (5, 42-44, 422, 550). Thus, for most phase-specific antigens, it is not clear whether they represent de novo synthesis of proteins associated with morphogenesis or, rather, constitute spatial and/or temporal rearrangements of preexisting components that were in a cryptic state before the morphogenetic event. There is evidence indicating the presence of specific antigens associated with yeast or mycelial cells, as reported by several groups who investigated the antigenic determinants of C. albicans during cellular proliferation (41-43, 50, 59, 62, 63, 72, 159, 207, 283, 394, 414-416, 422, 508, 525, 526, 534). On the other hand, several apparently morphology-specific proteins are also present in the other form in a cryptic state. These include the molecule recognized by a MAb, which is a surface protein in hyphae but a plasma membrane protein in yeast cells (394); a 30-kDa cell wall protein of germ tubes but not yeast cells, although mRNA was detected in both yeast cells and germ tubes (5); and differences in size and location reported for the receptor of the complement fragment C3d (563). In a final example, Gozalbo et al. (175) reported isolation of a gene that directed the synthesis of a 1.5 kb transcript in both yeast cells and germ tubes. Affinity-purified antibody to the gene product was used to determine cellular location and size. Two proteins of 25 and 48 kDa were detected in the membrane fraction of both yeast cells and germ tubes and the material secreted by protoplasts. In cell wall extracts, reactivity with two proteins of approximately 40 and 60 kDa was detected only in extracts from yeast cells obtained with Zymolyase. All of the observations indicate that the same gene products may be expressed in yeast and mycelial cells but located differentially within the cell. Thus, dimorphism may involve, in addition to differential gene expression and cell wall organization, a distinct pattern of both posttranscriptional and posttranslational processing of some gene products leading to differential location of the processed proteins.

Cell wall proteins associated with morphology are summarized in Table 2 and the following sections.

Epitope recognized by MAb 4C12. We have identified two major HMWM components solubilized from isolated mycelial-phase walls after degradation of the glucan network with Zymolyase (59). The proteins have an electrophoretic mobility corresponding to apparent molecular masses of 260 and 180 kDa (HMWM-260 and HMWM-180). These moieties were not detected in the wall of blastoconidia. These components behaved as a highly disperse population of mannoproteins with molecular masses ranging from about 150 kDa to more than 600 kDa on gel chromatography (122).

Epitope recognized by MAb 3D9. Protein from a germ tube culture purified by affinity chromatography on fibrinogen was used as immunogen in the production of MAbs. One of these antibodies, MAb 3D9 of the IgM class, reacted only with germ tube surfaces as judged by indirect immunofluorescence (324). The fluorescence was homogeneous along the germ tube. However, after 48 h of growth, the fluorescence was less intense and heterogeneous, with its greatest reactivity at the tip. This observation suggested that expression was greater at the point of active growth. Whether the diminution of reactivity behind the growing tip was a result of shedding of the determinant or alterations and rearrangements in the mature hyphal wall is unclear. While the presumed presence of the determinant in the culture medium points to the first explanation, the possibility that excess antigen is secreted to the exocellular environment during growth can not be eliminated. Western blot analysis of EDTA-ME and Zymolyase extracts from yeast cells and germ tubes confirmed reactivity with germ tube moieties. The reactive moiety was polydisperse. The determinant was larger than 210 kDa in the chemical extract and ranged from 110 to 220 kDa in the Zymolyase extract. The larger size in extract containing reducing agent parallels that observed with the determinant of MAb 4C12, as discussed above. The sensitivity of the determinant to proteases and insensitivity to periodate indicated that the epitope was a peptide. Analysis by SDS-polyacrylamide gel electrophoresis (PAGE) of the protein purified from Zymolyase extracts of germ tubes revealed a single disperse band of 110 to 170 kDa that was reactive with concanavalin A (325). The final product was enriched 126-fold for the protein and 16-fold for carbohydrate. MAb 3D9 did not react with Rhodotorula or Saccharomyces or any of 40 isolates distributed among six other Candida species (324). Reactivity was observed with all 58 isolates of C. albicans examined. The determinant appears to be specific to C. albicans and the hyphal form.

Hwp1p. HWP1 (hyphal wall protein) was cloned by screening a cDNA library with adsorbed antiserum that reacted only with hyphal surfaces (513). Further analysis showed that the ORF encoded a 234-amino-acid sequence with a calculated molecular mass of 22,750 Da containing a typical signal sequence and signal site for yeast KEX2-encoded endoprotease. More interestingly, the sequence revealed a series of tandem repeats containing a 10-amino-acid motif rich in proline and glutamine that covered much of the remaining sequence. No consensus N-glycosylation sites were found. The C terminus was rich in serine and threonine, providing potential O-glycosylation sites. A 2.3-kb mRNA was detected only under growth conditions where hyphae were produced. Antiserum raised to the recombinant protein reacted only with hyphal surfaces. Similar expression of Hwp1p was detected on the surface of 20 other C. albicans strains. The antiserum also reacted with fungi in tissue sections of colonized mice. Again, only hyphal surfaces were reactive. Western blot analysis of Zymolyase extracts of hyphal forms showed a pattern of polydisperse reactivity at approximately 34 kDa. No reactivity was detected with extracts of yeast cells. A recombinant protein produced in Pichia pastoris was about twice the calculated size. The authors suggest that the differences may be attributable to anomalous migration of proline-rich proteins, glycosylation, and protease activity of Zymolyase. As a proline-rich protein, Hwp1p may share properties with similar proteins in which proline residues are proposed function to maintain polypeptide chains in extended conformations and to mediate noncovalent interactions between chains. As noted by the authors, salivary proline-rich proteins are found in this group. These proteins are a substrate for transglutaminase of buccal epithelial cells to which C. albicans can adhere. This protein on candidal surfaces may also be a substrate. The repeats also contain cysteine that may be involved in covalent linkages.

Hyr1p. A hyphally regulated gene, HYR1, was isolated and characterized (16). Northern analysis confirmed that the 3-kb transcript of the gene was expressed under several conditions favoring formation of germ tubes but not under yeast growth conditions. The ORF encoded a predicted 937-amino-acid polypeptide of 94.1 kDa. The sequence contained an N-terminal signal sequence and C-terminal sequence for attachment of a GPI anchor. The sequence also contained 17 N-glycosylation sites and a domain rich in serine and threonine. Another domain was rich in serine, glycine, and asparagine and contained seven copies of a 4-amino-acid motif, NEGS, of unknown function. The presence of an N-terminal signal sequence, a C-terminal GPI signal sequence, and a serine- and threonine-rich region is shared with several other sequences that are associated with the S. cerevisiae cell wall (Cwp1p, Cwp2p, Tip1p, Aga1p, Ag1p, and Flo1p) (291, 444, 538, 555) or membrane sequences (388). Other C. albicans sequences have similar structural characteristics: the putative cell wall protein Als1p (discussed in a subsequent section) (212), the protein encoded by clone 8M (477), membrane protein Phr1p (467), and possibly the putative Hyr1-like protein associated with HYR99 (473).

Cassone and colleagues (45, 65, 169, 542, 543) have studied the main mannoprotein components of C. albicans implicated in immunomodulation of host defense. Among the mannoproteins present in an acidic extract (MP-F2), a 65-kDa mannoprotein (MP65) is a main target of human T-cell response. The proliferative response stimulated by this component was of an antigenic nature rather than a mitogenic one, and the response was targeted mainly to polypeptide epitopes (65, 542). A similar constituent was present in the material released to the culture medium by C. albicans (45, 543). MP65 was further purified by immunoaffinity chromatography with MAbs generated against putative protein epitopes of the molecule (169). MP65 has a pI of 4.1 and a protein-to-polysaccharide ratio of 1.8:1. The moiety appears to contain only O-linked sugar residues, as deduced from its sensitivity to dilute alkali but not to several endoglycosidases. Nanogram doses of purified MP65 induced extensive T-cell proliferation of human peripheral blood mononuclear cells.

One of the most highly evolutionarily conserved features of living organisms is the ability to respond to a sudden change of temperature (heat shock response) and to other adverse environmental conditions. This response involves the increased production of a set of proteins collectively referred to as heat shock proteins (hsps), whose production presumably contributes to the protection and damage repair of cells following stress (312). Initially identified as proteins whose synthesis was enhanced by an increase in temperature, several of the major hsps have been described since then to play important roles in all major growth-related processes, such as cell division, DNA synthesis, transcription, translation, protein folding and transport, and membrane translocation (312). Several hsps are synthesized constitutively, reflecting the important cellular functions performed by these proteins under nonstress conditions. hsps are also immunodominant antigens and major targets of host immune response during different types of infection (240, 241, 323, 332, 581). The response to hsps may also be involved in autoimmunity (241, 332, 582). Several families of hsps exist which are designated according to their average apparent molecular mass (e.g., hsp90, hsp70, and hsp60). Some of these families are divided into different subfamilies, and each subfamily may have different members.

The response of C. albicans to heat shock includes the synthesis of a set of proteins with an ample array of molecular masses (588). Genes coding for proteins belonging to the hsp90 and hsp70 families have been found in C. albicans (131, 278, 303, 316, 335, 532). Interestingly, the cellular location of some of these gene products is not confined to the cytosolic compartment, since they are also present in the cell wall and at the cell surface (303, 340). This extracellular localization could be a common feature of several fungal hsps, since it has been described also for Histoplasma capsulatum (168) and S. cerevisiae (294). As components of fungal cell walls, hsps may play roles similar to those described in the cytoplasm.

hsp90. The structure of hsp90 is highly conserved in organisms ranging from bacteria to humans. Members of the hsp90 family of proteins play important roles in protein biogenesis and have been referred to as molecular chaperones. hsp90s interact with a variety of cellular proteins, such as steroid hormone receptors (410) and tyrosine kinases (95), and have been associated with morphological changes in pathogenic fungi and protozoa (322, 488, 554). S. cerevisiae contains two genes encoding hsp90, HSC82 and HSP82. HSC82 is a constitutively expressed gene only weakly induced upon stress exposure, while HSP82 is expressed at a much lower basal level and is strongly activated after exposure to heat shock (32).

hsp70. hsp70s are highly conserved in organisms ranging from bacteria to humans. The DnaK chaperones are the prokaryotic homologues of hsp70s (437), having about 50% identity to all eukaryotic hsp70s. The eukaryotic proteins are between 50 and 98% identical. Sequence similarity between hsp70 proteins extends over the entire protein, but particularly highly conserved regions are present in the N-terminal domain (30, 89). The hsp70s are not restricted to heat shock protection; they also play a role in protein folding, translocation of proteins across membranes, and gene regulation (90, 91, 155, 156, 431). In S. cerevisiae, at least 10 HSP70-like genes have been described and classified into five families from SSA to SSE (288, 374, 440, 502). Most of the proteins encoded by these genes are present at considerable levels under all growth conditions, although their synthesis is increased after exposure to stress. The SSA family includes four members, SSA1 to SSA4, with 80 to 97% DNA sequence similarity (218). Cell viability requires moderate to high levels of at least one of the proteins encoded by this family (567). Under normal growth conditions, both SSA1 and SSA2 genes were expressed at moderate levels, while heat shock resulted in increased expression of SSA1 and strong induction of both SSA3 and SSA4 expression (218). Inactivation of either SSA1 or SSA2 did not result in an obvious phenotype. However, ssa1 ssa2 double mutants were unable to form colonies at 37°C (88). They were viable and grew more slowly than the parental strain at temperatures lower than 37°C as a result of a high expression of SSA4 (567). Until recently, the gene products of the SSA subfamily of hsp70 have been considered to be located exclusively in the cytosol of S. cerevisiae (91, 312). However, we have described the presence of Ssa1p and Ssa2p in the cell wall of this budding yeast, where they could play a similar role to that in the cytoplasm (294).

Heat shock or stress mannoproteins. Heat shock mannoproteins with approximate molecular masses of 180 to 200, 130 to 150, 90 to 110, and 60 to 70 kDa have been identified as being involved in the secretory immune response during mucosal candidiasis (413, 417). These molecules were expressed after a temperature shift from 25 to 37°C and were detected by indirect immunofluorescence. The antigenic determinants recognized by secretory IgAs in saliva and vaginal washings were polysaccharide in nature (413). Subculturing on agar at 25°C switched off expression of the antigens from mucosal isolates, while several transfers were required to reduce expression by isolates from deep infections (417). Since factors other than temperature can influence the in vitro and in vivo expression, these cell wall components should be considered "stress" proteins. The 200-kDa component appeared to be the main component recognized by vaginal secretory IgA. The 180- to 200-kDa component also enhanced tumor necrosis factor secretion by a murine macrophage cell line (411).

Glycolytic enzymes, which are highly conserved through evolution, appear to be important during C. albicans pathogenesis, since they act as major inducers of host immune responses and are major allergens during candidiasis (161, 219, 221, 489, 517, 530). Glycolytic enzymes can constitute up to 30% of total soluble proteins in S. cerevisiae (201). The presence of glycolytic enzymes in the cell wall of C. albicans has been reported recently. Enolase was found to be associated with glucan in the inner layers of the cell wall (8). Another report localized this enzyme on the surface of C. albicans cells, probably due to adventitious binding of enolase released from lysed cells (132). Phosphoglycerate kinase (PGK) was found in the cell wall and at the surface (6). We have also recently reported the presence of a cell wall-bound form of GAPDH (161). The presence of alcohol dehydrogenase (ADH) in the cell wall also has been suggested (406) as the fibronectin receptor. Together, localization of presumably "intracellular" glycolytic enzymes in the cell wall of C. albicans is rather intriguing, as are the possible roles that these molecules could play as part of the cell wall (e.g., energetic role, antigenicity, surface receptors), as well as the mechanisms involved in their secretion. However, the presence of glycolytic enzymes on microbial surfaces is not unprecedented. GAPDH has been reported as a major surface protein on group A streptococci (Streptococcus pyogenes), where it not only is enzymatically active but also serves as a binding protein for fibronectin, lysozyme, myosin, actin, and plasmin (403, 572). In Kluyveromyces marxianus, the same enzyme protein has been identified as a constitutive protein of the cell wall, and its level increases substantially upon induction of flocculation (136). ADH is also a surface protein of Entamoeba histolytica, where it has been reported to display receptor activities for a number of extracellular matrix (ECM) components (579). PGK, GAPDH, and triose-phosphate isomerase have been found on the surface of Schistosoma mansoni (173, 280, 503) and have been suggested as potential candidates for vaccine development.

Enolase. Enolase (2-phospho-D-glycerate hydrolyase) is an enzyme of the glycolytic pathway, where it catalyzes the dehydration of 2-phosphoglycerate to phosphoenolpyruvate. It also catalyzes the reverse reaction during gluconeogenesis. It is among the most abundant proteins in C. albicans and S. cerevisiae (201). In S. cerevisiae, enolase is encoded by two genes whose expression is differentially regulated depending on the carbon source and growth phase (343) and is coordinately regulated with the expression of other glycolytic enzymes (314). Different groups have reported the cloning and characterization of the C. albicans enolase gene (143, 330, 521). Although only one gene was initially detected by Southern analysis (143, 330), further experiments demonstrated the presence of two enolase gene loci (420). However, it is not known whether both loci provide functional genes.

Phosphoglycerate kinase. PGK catalyzes the hydrolysis of 1,3-bisphosphoglycerate to 3-phosphoglycerate with the production of ATP. Alloush et al. (6) reported cloning PGK1 by screening a cDNA expression library with antiserum to cell wall proteins. Antibody affinity purified to product of the clone detected a 40-kDa constitutively expressed cell wall protein and bound to the surface of intact cells. The presence of PGK in the cell wall was confirmed by two methods. Cell wall proteins of whole cells were biotinylated with a derivative which does not permeate the membrane, and PGK was found among the biotinylated proteins. Immunoelectron microscopy revealed that the protein was present at the outer surface of the cell membrane and in the cell wall as well as in the cytoplasm.

Glyceraldehyde-3-phosphate dehydrogenase. GAPDH catalyzes the conversion of glyceraldehyde 3-phosphate to 1,3-bisphosphoglycerate and the concomitant reduction of NAD⁺. We have recently isolated a cDNA clone by immunoscreening an expression library with pooled sera from two patients with systemic candidiasis and five neutropenic patients with a high level of anti-C. albicans IgM antibodies (161). The encoded polypeptide was reactive only with sera from patients with systemic candidiasis. The nucleotide sequence of the coding region of the genomic clone showed homology (78 to 79%) to S. cerevisiae TDH1 to TDH3 genes coding for GAPDH, and their amino acid sequences showed 76% identity; thus, the gene has been designated TDH1. Western blot analysis with a polyclonal antiserum against the purified cytosolic C. albicans GAPDH (PAb anti-GAPDH) revealed reactivity of a single 33-kDa band in the ME cell wall extracts. Indirect immunofluorescence demonstrated the presence of GAPDH at the candidal cell surface. Semiquantitative flow cytometry analysis showed sensitivity of the GAPDH form to trypsin and its resistance to removal with 2 M NaCl or 2% SDS. The specificity of the reaction was indicated by decreased fluorescence in the presence of soluble GAPDH. A cell concentration-dependent GAPDH activity was detected in intact blastoconidia and germ tubes. Activity was reduced by pretreatment with trypsin, formaldehyde, and PAb anti-GAPDH. These observations indicate that an immunogenic, enzymatically active cell wall-associated form of the glycolytic enzyme GAPDH is found at the cell surface of C. albicans (161).

Alcohol dehydrogenase. ADH catalyzes the reduction of acetaldehyde to ethanol with the formation of NAD⁺. A possible relationship between C. albicans ADH and receptors for vitronectin and fibronectin has been suggested (406). However, its putative surface location, deduced from its role as an adhesin, has not been further studied. The nucleotide sequence of the gene ADH1 encodes a 350-amino-acid polypeptide with high homology to other yeast ADHs (27, 489).

Since the cell wall is the outer surface of the organism, it is inescapable that physical interactions between the fungal cell and the host will be mediated by cell wall moieties. The potential interactions involve contact with phagocytic cells, host cells of every organ infectable by the fungus, extracellular matrix, and soluble proteins. In the balance of the relationship between the fungus and the host, these interactions may serve to promote the "interest" of the fungus in maintaining itself as a human commensal or establishing metastatic disease or the "interest" of the host in protecting itself against invasion of sterile sites and overgrowth in areas of normal colonization. It is not clear that all of these interactions have been identified, and it is even less clear that we know the roles played by these interactions in the development of infection. The task of deciphering the interactions and their role in pathogenesis is complicated, since there are multiple virulence factors and their importance may be influenced by the strain of the fungus, by the host tissue that is infected, and by underlying factors associated with host status. The concept of a virulence set hypothesis has been discussed by Cutler (96) in his review of putative virulence factors.

Although other cell wall components such as chitin, -glucan, and lipid play roles in the adhesion of C. albicans (158, 398, 476), proteins and mannoproteins are unquestionably the major mediators of adhesion. Early studies in the mid-1980s that sparked interest in the host ligands bound by C. albicans and the proteins that mediated such interactions were reports of the binding of human fibrinogen (34) and of complement fragments iC3b and C3d (194). Since then, additional host ligands and some of the fungal binding proteins or receptors for these ligands have been identified. The most relevant features of candidal cell wall adhesins and receptors for host ligands are summarized in Table 3. Most of these proteins appear to be among the medium-size polypeptide components present in the cell wall of C. albicans. For some ligands, several candidal proteins are able to bind the host protein, and it appears that some fungal proteins may be able to bind more than one ligand. The relationships between these multiple binding entities that have been detected in vitro and among isolated components and their functional role in the intact organism remain to be elucidated.

Serum proteins. The fact that C. albicans can interact with and respond to serum proteins has been recognized for many years. The ability of serum to stimulate the in vitro formation of germ tubes is well established and has been used as an element of species identification (535). The ability of the organism to stimulate the alternative complement pathway was also described 20 years ago (432). An effect on coagulation proteins was also noted during infection (409) and in vitro a decade later by Maisch and Calderone (313). To a large extent, interest in the interactions of C. albicans with serum proteins initiated the cascade of the studies of interactions between the fungus and host ligands at the molecular level.

(i) Serum albumin and transferrin. The ability of intact C. albicans cells to bind different serum proteins in vitro (including albumin, transferrin, and fibrinogen) was initially described by several authors on the basis of optical (immunofluorescence) and electron microscopy observations (33, 34, 401, 548). While the binding of fibrinogen has been examined more extensively, as discussed below, the binding of serum albumin and transferrin has not been investigated further.

(ii) Fibrinogen. Interaction of serum proteins with C. albicans has been particularly well characterized for fibrinogen. In vitro binding of human fibrinogen was reported initially to occur exclusively on short germ tubes and mature hyphae of the organism and with higher avidity than other serum proteins such as albumin and transferrin (34, 401, 548). Binding was not detected to yeast cells grown on Sabouraud medium. Among different cell surface mannoproteins that appeared to be responsible for adhesion of C. albicans cells to plastic surfaces, two species with molecular masses of 68 and 60 to 62 kDa were shown to interact with fibrinogen. They also interacted with laminin and C3d, the terminal degradation product of the complement C3 component (35, 550). This multiple interaction lead to the suggestion that candidal fibrinogen-binding proteins may be related to mammalian integrins of the ₂ or ₃ subset (33, 34, 548) (see below). Using iodinated whole fibrinogen, and D and E fibrinogen fragments, Annaix et al. (9) showed that fibrinogen and the D fragment bound to C. albicans germ tubes. The D fragment is an 85-kDa moiety containing the carboxy-terminal portions of fibrinogen. The binding was time dependent, saturable, and reversible. An average of 6,000 binding sites were determined per germ tube, with a dissociation constant (K_d) of 5.2 × 10⁸ M. Binding was not inhibited by several sugars, suggesting that the interaction was not a lectin type of binding. Ligand affinity blotting of a DTT-iodoacetamide extract with the D fragment identified a single 68-kDa candidal moiety with binding activity.

(iii) Complement fragment C3d. The ability of C. albicans cells to rosette antibody-sensitized erythrocytes coated with C3d as well as iC3b complement fragments was first described by Heidenreich and Dierich (194) and latter reported by other investigators (for a review of this topic, see reference 50).

(iv) Complement fragment iC3b. The presence of iC3b-binding components on the surface of hyphal forms of C. albicans was reported by Heidenreich and Dierich (194) and subsequently confirmed for both yeast and hyphal forms (115, 162). Very early in the study of the iC3b binding protein, various laboratories hypothesized that the functional similarity between the fungus and host cells in binding iC3b might extend to a structural similarity of the receptors involved. Antibodies to the human receptors were used as probes of candidal surfaces. Human complement receptors 3 and 4 (CR3 and CR4) bind iC3b and are two members of the human ₂ subset of the integrin family of receptors (86, 460). These heterodimeric proteins differ in their  subunits. CR3 contains the _M subunit, and CR4 contains the _X subunit. _M and _X are polypeptides of 165 and 150 kDa, respectively, and, in mammals, they each combine with a common ₂ chain of 95 kDa. At least eight MAbs that recognize _M exhibit high to moderate reactivity with C. albicans, and one MAb recognizing _X shows high reactivity with the fungus (25, 115, 117, 146, 162, 210; for a review, see reference 209).

Extracellular matrix proteins. The initiation of studies on the ability of C. albicans to bind and adhere to components of ECM is attributable largely to Klotz (248, 249, 251). Klotz showed that C. albicans yeast cells bound at the junctions of endothelial cells in confluent layers and that binding was enhanced when endothelial cells were removed from the substrate or contracted, resulting in ECM exposure. ECM proteins, collagen type IV, laminin, and fibronectin immobilized on plastic plates mediated the binding of yeast cells to the coated wells. ECM components form a network by interactions with each other. For example, entactin (nidogen) has the ability to interact with other ECM components such as fibronectin, fibrinogen, laminin, and type IV collagen (142, 575, 576). Thus, multiple binding targets are simultaneously present and adjacent in the ECM. Another potential contributor to the interaction are the collageneous domains present in candidal proteins, e.g., p37 laminin binding protein of yeast cells and mp58 fibrinogen binding protein (484). Some ECM proteins have collagen binding sites, and these may bind to candidal proteins containing collagenous domains. Therefore, in vivo binding to ECM is likely to involve multiple receptors binding to different ligands. These may be candidal receptors that bind to similar binding sites on ECM components as do integrins (see below), as well as ECM components that have collagen binding sites that bind to candidal proteins via collagenous domains. In general, the receptors are expressed extensively on hyphae, and this has been postulated as one of the ways by which germ tubes contribute to the development of infection. Although yeast cells are the most likely form to be hematogenously disseminated and extravasate, only a few yeast cells express the various binding capacities in vitro. An obvious question, then, is whether interactions between extravasating yeast cells and ECM contribute to establishing and maintaining infection. Three possibilities may be considered. First, not all extravasating yeast cells may need to bind for an infection to be initiated. Second, more cells may express the various receptors in vivo. Third, with the multiplicity of binding interactions that are available to the fungus, most yeast cells may express at least one of these, and that may be sufficient for binding.

(i) Laminin. Laminin, a major component of the basement membrane (539), is a large multidomain glycoprotein that seems to play a critical role not only in normal cell adhesion but also during tissue invasion and metastasis by tumor cells and pathogenic microorganisms. Contemporaneous with the study by Klotz (249) on the binding of the fungus to immobilized ECM components, Bouchara et al. (35) reported the binding of laminin to germ tubes grown in medium 199. As determined by indirect immunofluorescence, no binding to yeast cells grown in Lee medium was detected. Immunoelectron microscopy showed that this binding was in the outermost fibrillar layer. Radiolabeled laminin was used to show that binding was saturable and specific. Further analysis indicated about 8,000 binding sites per organism with a K_d of 1.3 × 10⁹ M. Binding was reduced by fibrinogen but not fibronectin, serum albumin, or several sugars, suggesting that binding was not lectin-like.

(ii) Fibronectin. C. albicans and other Candida species adhere to fibronectin. Fibronectin is a large (440-kDa) dimeric glycoprotein that circulates in plasma and is found as part of ECM. Various functional domains of the molecule have been identified; they include binding domains denoted as fibrin, collagen, DNA, cell (containing an RGD sequence), and heparin. Initial studies of the adhesion of Candida cells to endothelium suggest a role for fibronectin as the most prominent host ligand for candidal adhesion (248, 257, 472, 507), and binding of C. albicans to immobilized fibronectin was demonstrated by Skerl et al. (507). The binding was inhibited by mannan and was reduced by proteolytic treatment of fungal cell surfaces. The fibronectin receptor appeared to be distinct from C3d and iC3b binding activities when fibronectin binding in strains deficient in the other activities was examined (380). Several studies in different laboratories have examined the candidal receptor and its ligand binding characteristics. However, there are differences between these studies that may reflect the presence of multiple adhesins. A contribution by strains, experimental conditions, or assay methods is also a possible explanation for the differences in results reported. Comparisons between studies are also difficult, since the same parameters were not determined in each study. While a number of observations are consistent with homology between the candidal fibronectin receptor(s) and a mammalian integrin, some are not.

8

9

7

9

(iii) Entactin. López-Ribot and Chaffin (293) examined the interaction between C. albicans and entactin (nidogen), a novel adhesion glycoprotein molecule present in the ECM, especially in the basement membrane (239). Entactin is a sulfated glycoprotein that contains an RGD sequence and that forms a tight complex with laminin and interacts with collagen IV, fibronectin, fibrinogen, and itself. The interaction of entactin with intact cells and the binding moieties present in cell wall extracts was examined. Cells of both the yeast and the hyphal morphologies of the fungus were found to bind entactin, as detected by an indirect immunofluorescence assay. Most germ tubes bound entactin, and binding was heterogenously distributed on the hyphal surface with no reactivity on the parent yeast cell. A small percentage of nongerminated yeast cells in a germ tube culture bound entactin, and about 10% of cells in a yeast culture expressed the entactin binding capacity.

(iv) Vitronectin. C. albicans cells can bind vitronectin (serum-spreading factor) (222). Vitronectin is a constituent of vascular walls and dermis. It is present in serum and is involved in the regulation of blood coagulation (97). It contains a number of domains capable of interacting with microorganisms and mammalian cells, including an RGD sequence that binds eukaryotic integrin receptors and a glycosaminoglycan-binding domain that interacts with complex carbohydrates and glycoconjugates (182, 528). The characteristics of vitronectin binding to Candida were found to be the same as and to differ significantly from those of other ECM components such as fibronectin. The binding of vitronectin increased during late exponential growth of the organism (222). Two studies that have examined the effect of growth temperature and cell surface hydrophobicity (CSH) on the interaction have reached somewhat different conclusions. Jakab et al. (222) found that cells grown at 37°C bound 35% more soluble vitronectin than did cells grown 21°C and that cells with increased CSH bound somewhat more vitronectin than did cells with more hydrophilic surfaces. On the other hand, Silva et al. (505) found that hydrophobic cells grown at 23°C did not differ significantly in their adhesion to immobilized vitronectin from cells grown at 37°C. The interaction with soluble vitronectin was optimal at pH 4 and was enhanced substantially by calcium at pH 7 (222). A reduction in the binding capacity of cells following heating or treatment with several proteases suggested the presence of a protein receptor. Among potential binding inhibitors, a 64% reduction was observed with fibronectin and a minimal reduction of 10% with fibrinogen and collagen type I, while no effect was noted with type IV collagen and gelatin. Unlike other ECM ligands, no clear role for the RGD motifs present in the vitronectin molecule was established, since RGD and RGDS were ineffective inhibitors while GRGDS was somewhat effective (20% inhibition). Heparin was the most effective inhibitor (50 to 85% inhibition) (222, 286). This suggests that binding may be through the glycosaminoglycan binding region of vitronectin. Analysis of binding of plasma vitronectin revealed the presence of both high- and low-affinity receptors (286). For the high-affinity receptor, analysis yielded a determination of 98,000 binding sites per cell with a K_d of 3.5 × 10⁷ M. This dissociation constant is similar to the low-affinity binding constant reported in one study for fibronectin (380), although the number of binding sites is larger than reported for any other ligand.

(v) Collagens. Klotz (249) showed that C. albicans bound to immobilized type IV collagen, a component of basement membrane, and also to type I collagen. Binding to gelatin (denatured type I collagen) was about one-third lower than to the native form in this study, although in a more recent study the binding to native and denatured collagen was similar (255). Binding to type I collagen was reduced by about half to five fold by the removal of calcium and magnesium ions (249, 259). Calcium mediated most of the ion enhancement, since removal of calcium effected nearly a threefold reduction (259).

(i) Mannan adhesins. In addition to the ability of protein moieties to interact with the host, the mannan portion of mannoprotein may be involved in the fungus-host interaction. Two different oligosaccharides of mannan have been implicated in the adhesion of hydrophilic yeast cells to the marginal zone of murine spleen. A mutant strain deficient in the acid-labile portion of mannan was reduced in its ability to bind to marginal zone areas and displayed less specificity in binding, since cells also adhered to white pulp (73). Cutler and colleagues demonstrated that yeast cells bound to macrophages of the marginal zone and that mannan inhibited the binding to spleen and lymph node tissue (236). They identified a -1,2-linked mannotetraose as the entity possessing adhesive activity in the acid-labile mannan (285). An additional adhesin that contributes significantly to the binding of yeast cells to the marginal zone has been found in the acid-stable portion of mannan (234). Factor 6, a mannooligosaccharide that confers serotype A specificity, also contributes to epithelial adhesion (357). A factor 6-deficient strain or serotype B strains showed reduced adhesion to exfoliated human buccal epithelial cells compared to serotype A cells. Mannan from serotype A parental cells but not from the deficient strain, as well as antiserum to factor 6, inhibited the binding of C. albicans A cells. O-linked mannosyl moieties of cell wall proteins may also play a role. The O-linked oligosaccharides of the 58-kDa fibrinogen binding protein have been implicated in the binding of ligand (60).

(ii) Hydrophobic proteins. CSH in C. albicans appears to be related to a plethora of host interactions and fungal functions. These interactions and functions include the ability of the microorganism to adhere to epithelial and endothelial cell surfaces, ECM proteins, and plastic indwelling devices; to evade the action of phagocytic cells; and to enhance the uptake of substances from the medium (10, 39, 130, 185-187, 252, 258, 298, 309, 500, 549, 562).

(iii) Fimbriae. The presence of long, thin filamentous protein cell surface appendages termed fimbriae in C. albicans strains was initially reported by Gardiner et al. (154). Fimbriae may mediate adhesive interactions between the fungus and the host. As discussed above, further characterization of purified fimbrial preparations revealed that the major structural subunit of the fimbriae is a glycoprotein with a molecular mass of approximately 66 kDa (583). The glycoprotein consists of 80 to 85% carbohydrate (primarily D-mannose) and 10 to 15% protein composed of 50% hydrophobic amino acid residues. It is unclear whether there are additional minor components that contribute to the fimbrial structure and function. The fimbriae bind directly to buccal epithelial cells (BECs), and purified fimbrial preparations inhibit C. albicans binding to BECs (583). The ability of C. albicans yeast cells to bind specifically to the glycosphingolipid lactosylceramide has been reported (226). In this regard, C. albicans fimbriae also bind to asialo-GM₁ (gangliotetraosylceramide) in a saturable and concentration-dependent manner (584). Therefore, the fimbrial protein appears to be the moiety mediating binding to glycosphingolipids displayed on the surface of human BECs and could represent an adhesion motif for fungal cells (226). Bacterial pili may also bind to the glycosphingolipid asialo-GM₁, and antibodies to Pseudomonas aeruginosa pili cross-reacted with fimbriae (585). Recent studies showed that the epitope that confers receptor-binding properties (adhesintope) to C. albicans fimbriae appears to be identical to that present in P. aeruginosa pili (279, 586). Antibodies prepared to the bacterial adhesintope inhibited binding (43 to 50%) of the fungus to BECs. Two synthetic peptides from the P. aeruginosa adhesintope were competitive inhibitors of binding to asialo-GM₁ (58 to 69%) and BECs (48 to 59%). The smaller peptide, DEQGIPK, that was the central region of the larger peptide was somewhat less effective than the larger peptide. These conserved adhesive motifs present in adhesins from evolutionarily distant microorganisms may account for the recognition of host ligands bearing GalNAc(1-4)Gal groups. It has been demonstrated that a number of pulmonary pathogens are able to utilize the GalNAc(1-4)Gal sequence as the ligand or receptor to attach to host tissues (271). Hence, this carbohydrate motif may also play a role in the genesis of Candida infections.

(iv) Plastic binding proteins. The ability of C. albicans cells to adhere to plastic medical devices was established some time ago. After adherence the organism may propagate and establish biofilms. The release of microorganisms from biofilms formed on different types of medical implants, prostheses, or catheters may contribute to or initiate acute disseminated nosocomial infections, whose frequency has increased dramatically in the last years due precisely to this reason (119, 166). In this context, a large number of nosocomial infections caused by Candida species derive from the use of different types of medical catheters and prostheses (390). Organisms may also colonize denture materials and adhere to saliva-coated plastic, as noted below. Organisms adhering to plastic may also be less susceptible to antifungal drugs (180, 231), and colonization may contribute to deterioration of the devices (48, 172, 321, 559).

(v) Epithelial binding lectin-like protein. The adhesion of C. albicans to BECs or vaginal epithelial cells facilitates colonization and can be regarded as the first step in the pathogenesis of Candida infections. Several categories of candidal adhesins appear to be involved in this process: (i) integrin analogs, (ii) fimbrial adhesins, (iii) lectin-like adhesins, and (iv) factor 6 (49, 50, 96, 110, 111, 209, 406). Of these, lectin-like interactions between the protein moiety of a candidal lectin-like mannoprotein adhesin and a carbohydrate receptor on the surface of the host cell seem to be a major mechanism of adhesion to epithelial cells. The presence of candidal lectin-like adhesins with specificity for L-fucose or GlcNAc has been reported (37, 53, 92, 93, 544). Fucose was the major inhibitor of binding of one strain of C. albicans to human BECs, while glucosamine and GlcNAc were major inhibitors of the binding of another strain (93). Synthesis of the lectin-like proteins was increased when yeasts were grown in galactose (345). While two different lectin-like proteins were produced, each strain appeared to produce only a single such protein with affinity for either L-fucose or GlcNAc (53, 93). None of these adhesins have been fully characterized, and there is no information on the molecular mass of the native moiety bearing the adhesion motif. However, a fucoside-binding protein fragment of an adhesin was purified from culture supernatants of C. albicans yeast cells by affinity adsorption chromatography on the trisaccharide determinant of the H blood group antigen that terminates in an L-fucose residue (53, 544). Before chromatography, the protein was treated with papain, N-glycanase, and dilute alkali. The purified adhesin fragment was devoid of carbohydrate, although no further biochemical characteristics of this molecule were reported. Analysis of the purified biotinylated adhesin by SDS-PAGE and Western blotting with ¹²⁵I-streptavidin and subsequent autoradiography revealed a band corresponding to a protein with a molecular mass of 15.7 kDa (53). However, the intact mannoprotein moiety containing the fucoside-binding fragment is likely to be considerably larger.

(vi) Agglutinin-like proteins. Several genes have been identified that have homology to fungal agglutinin genes, in particular S. cerevisiae AG1 encoding -agglutinin (291). The proteins encoded by these genes are candidates for surface binding interactions. In C. albicans, the ALS (agglutinin-like sequence) family consists of at least four genes that may encode cell wall proteins (212, 473). In S. cerevisiae, -agglutinin mediates cell-cell interaction during the mating of haploid yeast. It is possible that the ALS genes are remnants of such a cycle in C. albicans or encode proteins that have been subverted to another function. Like Ag1p, the Als proteins had hydrophobic N and C termini, suggestive of processing through the secretory pathway and addition of a GPI anchor. Recently, the GPI anchor of Ag1p was shown to be lost during covalent attachment of the protein to cell wall glucan (306). One unusual property of the Als proteins was a central domain of a tandemly repeated 36-amino-acid motif. These tandem repeats were not found in Ag1p. The number of copies of this repeated motif was highly variable, depending on which C. albicans strain was examined. In addition, ALS alleles in the same strain may exhibit different numbers of copies of the repeat motif. The repeat sequences were rich in serine, threonine, and proline and contained a consensus N-glycosylation site, suggesting that the repeat region would be highly N and O glycosylated. If indeed Als1p is in the cell wall, along with Hwp1p (513) (discussed above), it will be the second cell wall protein to contain a repeat motif. The repeat motifs of the Als proteins and Hwp1p are not the same. While ALS genes were originally isolated due to their expression in RPMI 1640-grown hyphae but not in YEPD-grown yeast forms, the genes do not appear to be regulated by the yeast-hyphal transition. Instead, regulation is dependent on components of the culture media.

(vii) Adherence to Streptococcus spp. and other bacteria. As a commensal of the oral cavity, C. albicans must adhere to surfaces and cells of the oral cavity to prevent clearance by the flushing action of saliva. In addition to adhesins that mediate the binding to BECs discussed in other sections, the organism can bind salivary proteins and oral and nonoral bacteria. These interactions appear to be effected through multiple mechanisms. Macroscopic and microscopic aggregation or agglutination of isolates with C. albicans has been observed with strains of Streptococcus sanguis, S. salivarius, S. mutans, S. mitis, Fusobacterium nucleatum, Actinomyces viscosus, Lactobacillus amylovorus, and Bacteroides gingivalis (13, 213). In general, agglutination with bacteria other than streptococcal species was sensitive to heat treatment of the bacteria and to inhibition by a sugar. For these species, the interaction suggested a lectin-like interaction with candidal surface carbohydrate. The streptococcal coagglutination appeared to involve a different mechanism, which has been examined more extensively in subsequent studies by Jenkinson and colleagues.

(viii) Adherence to salivary proteins. C. albicans can bind several salivary proteins, although very little is known about the cell wall components mediating the binding. C. albicans adheres to human salivary proteins when these proteins are use to coat dental acrylic or lining material or hydroxyapatite beads, a model for pellicle formation on tooth surfaces (54, 113, 385, 560). Several bacterial species, e.g., S. sanguis and A. viscosus, can bind to pellicle. Several C. albicans strains adhered to the beads, although differences in the extent of binding were noted (54). Glucose starvation of yeast cells grown at 28°C enhanced adhesion. Treatment of the bound salivary proteins but not the fungus with neuraminidase increased binding. Deglycosylation of salivary proteins bound to glass slides reduced the binding of the fungus (384). These observations suggest that carbohydrates may be involved in binding. Neuraminidase secreted by oral bacteria could increase the availability or identity of ligands participating in binding by exposing sugar residues that are recognized by candidal adhesins, e.g., the fucose-binding protein discussed above.

(ix) Miscellaneous. A DNA sequence of C. albicans, AAFI, that confers enhanced adhesion to both polystyrene and epithelial cells as well as autoaggregation properties to S. cerevisiae transformants has been isolated (20, 114). A 30-kDa protein that may be responsible for these phenomena in S. cerevisiae transformants was identified on the surface of both yeast and hyphal C. albicans cells (21). The authors suggested that this moiety may be related to the cysteine-rich 30-kDa hydrophobins that have been isolated from the cell surface of Schizophyllum and Aspergillus species (568, 569). The ability of the 30-kDa protein to promote autoaggregation in S. cerevisiae transformants indicates that the cell surface molecule may also be involved in the interaction among the fungal cells, acting as an aggregation factor. In fact, the formation of macroscopic aggregates is commonly observed in C. albicans mycelial cultures. Further work is obviously required to determine the role of this protein in candidal biology and host interactions.

A term that has been used in studies about cell wall proteins and in this review is "dynamic." This term has been used to remind us that the cell wall is not a static, invariant structure but, rather, a structure that can change during the life of a single cell in response to the growth state and a variety of environmental stimuli. This plasticity of an organelle that is required to maintain the integrity of the cell in its normal milieu is perhaps surprising. However, the budding cycle, cell growth, and hyphal formation require that this protective envelope be modulated in even its structural role. As demonstrated by the studies considered in this review, the regulation of expression of components in the cell wall occurs at several levels. Even in the glucan and chitin component, the fraction that one may be tempted to consider most invariant, there is temporal regulation of synthesis during the cell cycle involving multiple enzymes (for recent reviews, see references 80 and 246).

In the previous sections of this review, we have focused primarily on individual components of the cell wall. These observations suggest the direction of studies in the near future. Thus, we can anticipate additional studies to identify cell wall proteins and their functions in vitro, particularly if they are postulated as being contributors to pathogenesis, and we can expect the extension of studies of identified proteins to demonstrate the role of specific components in host interactions. These latter studies are likely to take advantage of the increasing experience with molecular genetics with this organism and emerging views of host immune response. In these concluding remarks, we would like to summarize many of the observations from the standpoint of the questions they have raised about the mechanisms required to produce and regulate the structure and function of the proteins and mannoproteins in the cell wall. Some of the recent observations are quite surprising and raise questions that we would not have considered a few years ago.

The mannoprotein and protein fraction contains some 40 or more species that, in itself, introduces a measure of complexity to the cell wall. Redundancy appears to be common among binding proteins or receptors, with several moieties binding each ligand or plastic. Identification of proteins in cell wall or cell extracts is not synonymous with surface localization, which is a requirement to be the mediator of the binding observed in intact cells. Thus, with the exception of the plastic binding proteins that were isolated by a method that implies exposure of the fungal surface to the substratum, it is not clear that each of these components is present at the cell surface. For most of these proteins that bind the same ligand, we do not know whether they represent products of different genes, are related by processing, or both. Another complicating factor is the possibility that processing alters the ligand specificity or binding site used by the binding protein.

Another related issue is the suggestion made several years ago that candidal binding proteins are multifunctional. Supporting this contention is the similarity of apparent molecular size for some components, the competition among ligands for isolated components, and the analogy to mammalian integrins that may bind more than one ligand. On the other hand, some ligands bind to more than one integrin. In contrast to these observations, there are some proteins that seem to bind a single ligand, e.g., the mp58 fibrinogen binding protein and the p37 laminin receptor moieties. Two studies that directly addressed this point determined that the C3d binding and mp58 fibrinogen binding proteins were different and that laminin and fibrinogen binding were not colocalized in yeast cells (300, 302). The fungus also appears to use multiple interactions to bind to host cells and substrates. Two mannan adhesins have been identified in mediating the binding to splenic macrophage. C. albicans can bind to multiple ligands present in ECM and do so via more than one binding site, i.e., RGD-peptide dependent and independent. Multiple binding interactions have also been suggested to mediate the binding to epithelial cells, e.g., mannan factor 6, lectin-like protein, and iC3b binding protein. Unraveling the individual contribution of the various binding proteins, those that are singular and those that are promiscuous in their ligand specificity, to the complex substrates presented by the host tissue compared to the use of purified ligands in vitro will be challenging.

Multiple mechanisms affect the expression of various cell wall mannoproteins and proteins. First, expression may regulated at the level of gene transcription as judged from Northern blot analysis. Among the genes whose expression is regulated by nutrient conditions are the genes for mp58 (FBP1) and ALS1 (5, 212). At least two genes, HWP1 and HYR1, appear to be expressed only in hyphae (16, 513). There are also proteins whose level of expression responds to in vitro growth conditions. Thus, more iC3b binding protein is detected at elevated glucose levels and more lectin-like protein is detected when cells are grown on galactose rather than glucose (162, 210, 345). There are also proteins and protein modifications that are expressed on some cells but not others in the population (60, 72, 296). Some determinants are also expressed by a variable number of cells in the population, apparently depending on the stage of culture growth (41, 42, 72).

Expression of the SAP gene family clearly is subject to a variety of controls of environment, strain, and morphology. Determination of the conditions affecting locus-specific expression is under way, and elucidation of the regulatory circuits will surely follow (214, 359). Miyasaki et al. (359) have suggested that there may be different trans-acting elements present in different strains or whose expression is controlled by morphology that contribute to differences in locus expression. For other proteins, variable expression within a population exposed to the same environmental conditions poses some interesting questions. The phenomenon is similar to penetrance, in which not all organisms with an appropriate genotype express the phenotype. Nonexpressing yeast cells can give rise to expressing germ tubes, and subsequent yeast cultures have a population of variable expression. For the mp58 and p37 fibrinogen and laminin binding proteins, the expressing yeast cells were among the largest cells in the population. Since older cells are generally larger than young ones (70), an intriguing possibility is that yeast cell expression is associated either with age or with some alteration in control of cell size.

Multiple posttranslational mechanisms determine cell wall protein expression. Proteins whose cellular location is morphology dependent are examples of one type of regulation (5, 394, 563). Since at least two of these proteins differ in size in the different locations, a posttranslational processing-dependent mechanism of targeting may be hypothesized. Thus, when processed in one way, the protein bears a signal that targets one destination, and when processed in another way, it bears a different destination signal. Such a change could also be effected by a regulation of mRNA splicing that removed signals for processing and targeting under one condition but not the other. In the case of the one protein where gene expression has been examined (5), no obvious difference was noted in the size of the transcript in a Northern blot.

Another type of regulation is represented by proteins that are differentially distributed within the cell wall. First, there are proteins that are secreted through the cell wall to the exocellular environment, and then there are proteins that are retained within the cell wall. However, retained proteins can differ in their location. Some proteins are homogeneously distributed at the surface, and others are heterogeneously distributed (59, 62, 293, 296, 328). At least one protein, enolase, is located internally and is not expressed at the surface (8). Some proteins are thought to be covalently linked to glucan (237, 238, 247, 268, 462, 463). However, proteins that are postulated to be covalently linked, i.e., releasable by -glucanases, may also be released by treatment with reducing agents (57). Among the proteins that have signal sequences are proteins such as Sap, whose presumed major destination is the exocellular environment, as well as proteins that remain predominantly wall associated. Therefore a signal sequence is not sufficient to determine the extracellular location. The retention or release mechanisms have yet to be explored fully.

These examples suggest the likely existence of multiple methods for targeting cell wall proteins to specific locations and retention of proteins in the cell wall. Gene sequences clearly indicate that some proteins are secreted by the classical endoplasmic reticulum-Golgi pathway. Proteins such as Hwp1, Hyr1p, Sap proteins, -N-acetylglucosaminidase, and putative Als1p contain a signal peptide sequence associated with routing through this pathway (55, 215, 292, 359, 571). On the other hand, other cell wall proteins do not appear to contain such a signal peptide. These proteins include hsp70, enolase, PGK, and GAPDH, which that have been demonstrated to be bona fide cell wall components. A few exceptions of secreted proteins without such a signal sequence have been noted in plants, animals, and microbes (272). These proteins in prokaryotic organisms and the model yeast S. cerevisiae appear to be secreted by substrate-specific transport systems (ATP binding cassette [ABC] transporters), while the mechanisms for such proteins in mammalian cells are unknown. The secretion of a-factor in S. cerevisiae involves the ABC transporter Ste6p. Other cell wall proteins appear to be secreted by as yet unknown pathways (81). If this is a valid analogy, the existence of several such systems for this group of C. albicans proteins would be predicted. The presence of an hsp70 among this group raises another possibility. Among functions associated with cytoplasmic hsp70 is translocation across intracellular membranes. hsp70 may assist in its own translocation across the plasma membrane or may contribute to the translocation of other proteins. If members of this group of proteins provide important or essential functions for C. albicans, substrate-specific transporters may constitute a new target for antifungal agents.

One area in which we have minimal information, even from the model system of S. cerevisiae, is how proteins are targeted to specific cell wall locations. As noted above, there are several aspects to localization of a protein within the cell or cell wall. In S. cerevisiae, -agglutinin is modified by addition of a GPI anchor that is lost during covalent attachment to cell wall glucan (306) and a truncated protein missing the GPI anchor is secreted to the medium (573). To date, no C. albicans cell wall proteins have been demonstrated to have a GPI anchor, although Hyr1p and the putative cell wall protein Als1p contain a likely sequence. Two proteins, determinants of which are recognized by MAb 4C12 or 3D9, are both released by -glucanases, but both may also be found in ME extracts. If the enzymatically released moieties are covalently attached to glucan, there are several possibilities to explain this partitioning. First, the moieties could be secreted in both a GPI-tailed form (enzymatically released) and an unmodified form (chemically released). A second possibility would be that some of the bound form is subsequently released endogenously within the wall. A third possibility is that excess moiety is secreted and only a portion is attached.

Another subcellular partitioning problem is found with the hsp70, PGK, GAPDH, and enolase proteins that are found both in the cytoplasm and the cell wall. There must be some mechanism to apportion the protein between the two sites. For enolase, two genes have been identified, and the function of the two loci is unknown (420). One possibility is that the loci specify intra- and extracellular enolase. The cell wall enolase protein is blocked at its N terminus, and this type of modification could be used as a partitioning mechanism. Unlike enolase, only one locus is proposed for PGK (309). Subcellular partitioning may influence the regulatory circuit(s) affecting the regulation of the gene. The promoter region of C. albicans PGK1 did not contain some of the regulatory sequences found in S. cerevisiae (6), and the noncoding regions of other glycolytic enzymes are less highly conserved than the coding regions (530). These differences support the potential for different regulation between the two species that could be related in part to the subcellular partitioning in C. albicans.

A homogeneous and heterogeneous distribution of moieties on the cell surface has been demonstrated. However, mechanisms to effect such differences in distribution have not been addressed. Since there is thought to be minimal mobility within the cell wall, such an asymmetric distribution is likely to occur during cell wall synthesis. Indeed, surfaces of formaldehyde-fixed cells show the same distribution pattern of fibrinogen binding as do viable cells (60). Thus, this asymmetry is not induced by ligand binding. Furthermore, the mp58 and p37 binding proteins, which are both asymmetrically distributed, are not codistributed in cells that express both activities (302). Such a subcellular localization could be introduced by an asymmetric delivery of the proteins to the growing cell wall or by self-aggregation following delivery. However, the last possibility would suggest that different proteins are separately delivered or have different aggregation mechanisms.

Although none of the glycolytic enzyme proteins present in the cell wall have been demonstrated to also bind host ligands, there is a precedent among other microbial systems for this activity. With several proteins of the glycolytic pathway present, at least two of which appear to be enzymatically active, the possibility that the function of these proteins is indeed enzymatic will have to be considered. This is particularly interesting, since the reactions are able to generate ATP and NADH, which could be used for extracellular reactions, such as covalent linkages between cell wall components. However, this would also require that the substrates for the reaction be available.

Lastly, there is tantalizing evidence that cell wall adhesins/binding proteins are environmental sensors capable of initiating an alteration in fungal gene expression. Bailey et al. (14) showed that the profile of abundant proteins was altered by adherence to HeLa cells and that after adhesion, at least one protein appeared to become phosphorylated. Alterations of gene expression are mediated by some integrins (216). If the functional analogy of some surface proteins with integrins extends beyond the identity of the ligand, binding of ligands such as ECM may also initiate a cascade of signals that alter gene expression. If this suggestion is substantiated, a whole new avenue of investigation will open to determine the genes and functions altered by adhesion and the adhesive interactions that are able to initiate such events. In this context, it is provocative that the major ubiquitinated proteins of the cell wall also are binding proteins. There are suggestions that ubiquitination may be involved in the regulation of receptor signaling for some mammalian receptors (368, 404).

Our investigation into the proteins and mannoproteins of the cell wall of C. albicans is little more than a decade old. From these studies, along with those of the glucan and chitin component, has emerged a picture of a complex organelle. The number of proteins and mannoproteins found in wall in itself provides some measure of complexity. To this is added the evidence that proteins and mannoproteins collectively and individually do not have the same patterns of distribution within the wall, at the cell surface, or secreted to the extracellular environment. Furthermore, proteins are not retained in the wall by the same mechanisms, i.e., covalent or noncovalent attachment, and this again applies to some individual proteins that appear to have both mechanisms. A schematic model showing some of these aspects of distribution of cell wall components and the distribution of various protein and glycoprotein species in the wall structure is presented in Fig. 4. The identification and function of some these extracellular proteins as hydrolyases, adhesins, and putative structural proteins were not unexpected. However, the identity of some proteins to intracellular enzymes, the antigenic, functional, or structural relationships that some proteins appear to have with mammalian integrins, and the redundancy of proteins with apparently similar functions were not anticipated. The cell wall proteins are subject to regulation at the transcriptional and posttranslational levels by mechanisms that are responsive to environmental conditions and growth state. In turn, some cell wall proteins may be sensors of environmental contacts that initiate a signal for alterations in fungal gene expression. The localization of proteins within the cell wall is apparently subject to regulatory mechanisms that we are only now recognizing and have not yet begun to explore. This organelle is integral to fungal biology and pathogenesis of candidal infection, and unraveling its mysteries will provide challenges for the next decade.

View larger version (53K): [in this window] [in a new window]   FIG. 4.   Schematic diagram of the cell wall (CW) structure of C. albicans, showing the presence of different layers enriched in particular components. The microfibrillar polymers of -glucans (-g, ) and chitin (c, ) appear to be more heavily concentrated in the inner cell wall domains; -glucan-chitin complexes that appear to be formed by glycosidic linkages between both polymers will be located adjacent to the plasma membrane (PM) and the periplasmic space (ps). Proteins and glyco(manno)proteins (gp) appear to be dominant in the outermost cell wall layer, although they are also distributed through the entire wall structure. Once secreted through the plasma membrane, some protein and glycoproteins species will remain at the periplasmic space, possibly playing enzymatic roles (); some others will establish functional (i.e., -glucanases []) or structural covalent associations with -glucans and possibly also with chitin () adjacent to the plasma membrane; and, finally, other moieties will constitute the most external layer, where the different molecular entities may be homogeneously () or heterogeneously (fimbriae, cluster of receptor-like molecules, etc. []) distributed or specifically released (i.e., extracellular enzymes) to the extracellular medium (EM) (, ). Proteins and glycoprotein species in the outermost wall layer (, ) may establish different types of covalent (disulfide linkages) and noncovalent (hydrophobic and hydrogen ionic bonds) interactions. During their passage through the wall from the plasma membrane and periplasmic space to the outermost cell wall layers (, ) and possibly the extracellular environment (, ), proteins and glycoproteins are most likely to be in equilibrium with other proteinaceous constituents, thus contributing, at least from a functional point of view, to the cell wall layering. In any case, protein and glycoprotein species other than those specifically secreted to the exocellular medium may also be released to such locations by dying (lysed) cells or as a consequence of unbalanced processes of synthesis and degradation of the cell wall structure, required for wall expansion during cell growth. To simplify the scheme, some aspects such as possible interactions of cell wall components with the plasma membrane and proteins retained in the cell wall, apparently by either covalent or noncovalent linkages, are not depicted.