The fungal cell wall is a coherent structure formed by microfibrillar polysaccharides and amorphous material made of other polysaccharides and proteins. We performed a phylogenetic analysis of covalent proteins and enzymes that synthesize fungal wall polysaccharides to determine the possible evolution of the wall structure. It is suggested that the components that made up the archaic walls were structural ones, forming a primitive girdle that retained noncovalently bound proteins in the periplasm and allowed cell growth in hypotonic media. The following hypothetical series of events in fungal wall evolution is suggested: (1) Construction of a primitive wall made of chitin and chitosan by division 2 chitin synthases and chitin deacetylases, respectively. (2) Appearance of class II chitin synthase genes (CHS) after separation of Microsporidia. (3) Capture of a gene encoding β-1,3-glucan synthase from an organism related to Plantae or Chromista by horizontal transfer after separation of Chytridiomycota. (4) Appearance or horizontal capture from Chromista of genes involved in β-1,6-glucan synthesis after separation of Zygomycota. (5). Appearance of class III CHS genes. (6) After split of Dikarya phyla, appearance in Ascomycota of class I CHS genes and the capacity to synthesize covalently bound wall proteins.
The cell wall is the rigid outermost layer that covers the cells of a number of organisms belonging to different taxa, both prokaryotic and eukaryotic. In either type of cells, the general functions of the wall are similar, being responsible for the protection of cell integrity by providing support to the cell membrane to withstand the difference in osmotic pressure between the cytoplasm and the external medium. If the cell wall is removed by enzymatic treatment or inhibition of its synthesis by selective drugs, the resulting wall-less cells (protoplasts) can survive only in isotonic or hypertonic media. Besides this crucial role, the wall also protects the cell from the action of enzymes and different deleterious compounds, it is responsible for the shape and morphogenesis of the cell, and contains molecules responsible for the recognition of predators or hosts, or involved in the formation of biofilms (see Ruiz-Herrera, 1992; Sentandreu, 2004).
If the functions are similar, the structure and chemical composition of the cell walls of prokaryotes and eukaryotes are basically different. Thus, the structure of the cell walls from eukaryotic organisms is based on the presence of microfibrils formed by the association through hydrogen bonding of long chains of linear polysaccharides. Characteristically, microfibrillar polysaccharides are polymers of a single type of sugar bound through glycosidic β-linkages. On the other hand, the wall of prokaryotes does not contain microfibrillar components, probably because of their smaller dimensions; instead, its structure is made up in most prokaryotes of a tridimensional net with a characteristic composition, the sacculus. This is made of short oligosaccharides covalently bound to oligopeptides: the peptidoglycan of bacteria, and the archean pseudo-peptidoglycan.
Repeatedly, the structure of the eukaryotic cell wall has been compared with synthetic composites made of a structural component responsible for the resistance to tensions, and an amorphous component that confers resistance to pressure and protects the structural component from fracture and from the action of deleterious components of the environment. The main structural microfibrillar components, depending on the type of organism, are cellulose, chitin and β-1,3-glucans. The chemical components that constitute the amorphous part of the cell wall are different nonmicrofibrillar polysaccharides, proteins and lipids, plus salts and pigments in small amounts (for reviews, see Ruiz-Herrera, 1992; Sentandreu, 2004).
Prevalent ideas sustain the concept that all eukaryotic organisms descend from an ancestral uniciliate eukaryote (Cavalier-Smith, 2002, 2004; Steenkamp, 2006). Accordingly, at the time of appearance of cell-walled eukaryotic organisms, the prokaryotic cell wall had to be substituted by a completely different structure. The antecessors of the two most abundant in number wall-bearing eukaryotic groups, Plantae and Fungi kingdoms, separated early during eukaryote evolution (Cavalier-Smith, 2002, 2004), and developed cell walls based on structurally similar, but chemically different microfibrillar components: cellulose for plants and chitin for fungi. Both of them are linear polysaccharides made of a large number (normally >2000 U) of either glucose for cellulose or N-acetylglucosamine [2-acetamido-2-deoxy-d-glucose (GlcNAc)] for chitin, joined by β-1,4-linkages. Because of the crystalline arrangement of both polysaccharides that eliminates water from their structure, they are extremely insoluble, and with a high tensile strength, higher for chitin than for man-made fibers (steel, or carbon or boron fibers) (reviewed by Ruiz-Herrera, 1992; Ruiz-Herrera & Ruiz-Medrano, 2004).
The structure and chemical composition of the fungal cell walls has been the subject of a large number of studies. The majority of these studies have been focused on ascomycetes, and mostly on different yeasts, especially Saccharomyces cerevisiae, giving rise to an almost generally accepted model of the fungal cell wall. It has been concluded that the most important structural polysaccharides of the fungal cell wall are chitin and β-1,3-glucans that associate with β-1,6-glucans, and with covalently and noncovalently bound mannoproteins (see reviews and representative schemes in Lipke & Ovalle, 1998; Chaffin, 2008). Other polysaccharides present in lower amounts appear to be specific of groups or species: α-glucans, chitosan, polyuronides, galactans, etc. As indicated above, lipids, pigments and salts are present in minor amounts (see Ruiz-Herrera, 1992; Sentandreu, 2004).
Recently, and due to their important roles in the cell wall, considerable attention has been focused on fungal wall proteins, recognizing in general two different classes according to the way in which they associate with other wall components: noncovalently bound and covalently bound proteins. Noncovalently bound proteins are retained in the wall by hydrogen, hydrophobic and/or ionic bonds, being released into the medium in different proportions. Three types of covalently bound proteins are generally recognized: (1) glycosyl phosphatidyl inositol (GPI) proteins, which contain the residue of a GPI substituent, and are bound to β-1,6-glucans by a glycosidic linkage; (2) Pir proteins (proteins with internal repeats), bound to β-1,3-glucans through an alkali-labile linkage; and (3) proteins retained in the wall by binding to other proteins through disulfide linkages (Sentandreu, 2004; Chaffin, 2008).
Although the ascomycete wall constitutes the paradigm of the fungal cell wall, it may be indicated that our experimental and in silico analyses of the structure of the cell wall of the basidiomycete Ustilago maydis revealed significant differences from the established model of the cell wall of ascomycetes (Ruiz-Herrera, 2008), raising doubts on the existence of a universal model of the fungal cell wall.
Probably the first study that correlated the chemical composition of the cell wall with the then accepted taxonomic division of fungi was published by Bartnicki-García (1968). This author reported that the different fungal groups possessed cell walls with distinctive polysaccharides. Although the taxonomy of fungi has changed dramatically (see Lutzoni, 2004; James, 2006; Hibbett, 2007), the general concepts of that study are still valid. A different type of study was reported by Bernabé (2002), who analyzed the evolution of characteristic polysaccharides of chemotaxonomic value present in the cell walls of ascomycetes.
More recently, Lipke and colleagues (Coronado, 2006, 2007a, b) devised computational procedures to correctly analyze the homology among fungal cell wall glycoproteins, and utilized the data to determine the existence of a number of motifs in common among cell wall proteins, and their conservation in different fungal groups, in an attempt to understand the evolutionary origin of the fungal cell walls.
In the present study, we have followed a different approach in order to analyze the phylogenetic relationships of the cell walls from the several fungal groups. Accordingly, we have analyzed the homology of the enzymes responsible for the synthesis of different wall polysaccharides, and the covalently bound proteins that make up the fungal cell wall in members of the different fungal phyla. Noncovalently bound wall proteins were excluded from this analysis because of the difficulty in deciding whether they truly form part of the wall structure, are fortuitously retained in the wall because of some physicochemical characteristics (size, charge, etc.) or are mainly secreted into the medium. The data reported here reveal that the model of the cell wall structure of ascomycete yeasts cannot be extended to the walls of the rest of the fungal taxa, and constitute an initial attempt to understand the evolution of the cell wall structure in the kingdom Fungi.
Chitin is the basic structural polysaccharide of the cell wall from most fungi. As indicated above, chitin is a large polymer made of N-acetylglucosaminyl residues bound through β-1,4-glycosidic likages. Chains of the polysaccharide associate by hydrogen bonds to form microfibrils. Three different forms of chitin exist in nature, depending on the arrangements of the chains in the microfibrils: β, with parallel chains [i.e. all in the same nonreducing→reducing ends orientation (polarity)], α, with antiparallel chains (in which the polarity of the chains alternates) and γ, where two parallel chains are adjacent to one antiparallel chain. The most abundant form in nature, also present in the fungal cell walls, is α-chitin (reviewed in Ruiz-Herrera, 1992; Ruiz-Herrera & Ruiz-Medrano, 2004).
Chitin synthesis is catalyzed by enzymes known as chitin synthases (Chs), which utilize UDP-GlcNAc as a sugar donor. Despite some unconfirmed early reports (Kang, 1984; Montgomery, 1984), chitin synthases have not yet been isolated, possibly because of their hydrophobic characteristics. Accordingly, their catalytic properties are known from the use of membrane fractions or chitosomes (see Ruiz-Herrera & Ruiz-Medrano, 2004), and their structure from their encoding genes (reviewed by Ruiz-Herrera, 2002). A characteristic of fungi is that they contain more than one chitin synthase-encoding gene (CHS) in a number that varies from three in S. cerevisiae to >20 in several mucoromycete species. [According to the revised fungal taxonomy, the former Zygomycota phylum was found to be polyphyletic, and was therefore divided into several phyla. The designation Mucoromycotina for the former Mucorales has been suggested (James, 2006; Hibbett, 2007). In this paper, we used either designation for facility.] Sequence analysis of chitin synthases revealed the existence of five classes (reviewed by Ruiz-Herrera, 2002), which some authors have later on extended to seven (in this study, we used the classification into five classes). A thorough analysis of the structure of Chs that identified conserved motifs in chitin synthases revealed the existence of two groups of enzymes, denominated divisions: 1, including classes I–III, and 2, which includes classes IV and V (Ruiz-Herrera, 2002). Division 1 enzymes possess an activity-related motif named ‘chitin synthase 1,’ different from the equivalent motif of enzymes belonging to division 2, named ‘chitin synthase 2’ (see http://pfam.sanger.ac.uk/).
Taking into consideration that chitin is probably the most characteristic compound of the fungal cell wall, we proceeded to analyze the phylogenetic relationships of chitin synthases from fungi belonging to the different phyla. The neighbor-joining method was used for the analysis (see details in the legend of Fig. 1). The first interesting observation of this analysis concerned Chs from division 1 (Fig. 1). Thus, it was observed that its three classes are present only in ascomycetes (although some species, such as S. cerevisiae, contain members of only two classes). On the other hand, basidiomycetes do not possess class I enzymes. Zygomycetes (members of the new phylum Mucoromycotina) and the chytridiomycete Batrachochytrium dendrobatidis contain only class II Chs, whereas the microsporidian Encephalitozoon cuniculi does not contain any Chs belonging to division 1. Regarding class II enzymes, two subclasses (a and b) are distinguished in basidiomycetes and mucoromycetes, but not in ascomycetes. The latter appearance of class III is apparent. With regard to the new phylum Glomeromycota (previously included among zygomycetes), no genomes have been sequenced or are accessible. Nevertheless, a single complete CHS gene, belonging to class IV of division 2, has been isolated and sequenced in Glomus versiforme (Lanfranco, 1999).
Evolutionary relationships of fungal chitin synthases belonging to division 1. The mega4 program (Tamura, 2007) was used to obtain the dendrogram generated by the ‘neighbor-joining’ method (Saitou & Nei, 1987), with 1000 bootstraps (Felsenstein, 1985). Data in the phylogenetic dendrogram were drawn to scale with the branch length in the same proportion as the evolution distance. The evolution distance was computed using the Poisson correction method (Zuckerkandl & Pauling, 1965). The abbreviations used are indicated in Table 1. Asco, ascomycetes; basidio, basidiomycetes; Chytridio, chytridiomycetes; and Mucoro, mucoromycetes (zygomycetes).
Abbreviations after the name of each species were used to identify them in the dendrograms.
Genes belonging to both Chsp classes, IV and V from division 2, are represented in members of all phyla, except the microsporidian E. cuniculi, whose single CHS gene belongs to class IV (Fig. 2). This result may agree with the phylogenetic analysis suggesting that class IV is more primitive than class V. The rest of fungal phyla contain two IV subclasses, but Mucoromycotina species (zygomycetes) possess the largest numbers of enzymes and two V subclasses the same as basidiomycetes, as compared with members of the other phyla. The existence of subclasses may be indicative of gene duplication and differentiation occurring at early periods during evolution. The absence of two IV subclasses in ascomycetes may be attributed to loss of the ancestor gene in this fungal group.
Evolutionary relationships of fungal chitin synthases belonging to division 2. See details in the legend of Fig. 1.
It is interesting to note that a large number of class V enzymes possess a motif related to a myosin head of an unknown function. The fact that enzymes with or without a myosin head coexist in the same subclasses shown in Fig. 2 indicates that they do not belong to separate groups.
A phylogenetic comparison of Chs from fungi with those from other organisms that also contain chitin, at least during specific stages of their life cycles, Animalia, Chromista and Protozoa kingdoms, was very illustrative. These enzymes formed groups clearly separated from the fungal enzymes, but maintaining a close association with those from division 2 (Fig. 3). In silico structural analysis of the sequence of chitin synthases present in members of Protozoa and Animalia showed that all of them possessed a chitin synthase motif with a homology closer to the ‘chitin synthase 2’ motif from fungal enzymes, and none contained the myosin head motif, suggesting that they are more related to fungal class IV chitin synthases. These data support the concept that the primitive Chs genes were related to modern class IV. Interestingly, chitin synthase genes present in Choroviruses that infect Chlorella (Kawasaki, 2002) are also related to fungal class IV enzymes, possessing the ‘chitin synthase 2’ and lacking the myosin motifs.
Evolutionary relationships of chitin synthases from fungi and other taxonomic groups. See details in the legend of Fig. 1.
Noticeable exceptions to these concepts are genes encoding chitin synthases from Oomycetes (Chromista). These are related to fungal division 1 (see Ruiz-Herrera, 2002), and it is important to notice that those from Aphanomyces euteiches encode smaller fungal enzymes that synthesize a peculiar form of water-soluble chitin of a very low Mr (Badreddine, 2008). Whether this is a general characteristic of chitin synthases and chitin of oomycetes remains to be determined.
Chitosan is a polysaccharide of a basic nature made up mostly of glucosamine and a variable number of GlcNAc residues all bound through β-1,4-linkages. Chitosan is not synthesized by a transglycosylation reaction from a glucosamine-activated donor, but through the deacetylation of nascent chitin by specific deacetylases (chitin deacetylases) before this polymer reaches its microfibrillar crystalline structure (Davis & Bartnicki-Garcia, 1984a, b; Calvo-Mendez & Ruiz-Herrera, 1987). The existence of genes encoding this enzyme may therefore be taken as evidence suggesting the presence of chitosan in an organism, at least at some developmental stage. Chitin deacetylases are not specific of fungi; genes encoding the enzymes are also present in animals and protozoa, revealing their ancient origin.
Regarding fungi, chitosan is the characteristic component of the cell wall from Mucoromycotina (zygomycetes) species, and was originally identified in a member of this taxon (Phycomyces blakesleeanus; Kreger, 1954). Nevertheless, this polysaccharide is not specific of this group of fungi, and it was also found to be present in ascomycetes and basidiomycetes. In ascomycetes, it has been detected in the ascospores of S. cerevisiae (Briza, 1988) and Schizosaccharomyces pombe (Matsuo, 2005), but not in the vegetative walls of S. cerevisiae or Candida albicans (Banks, 2005), whereas in contrast, small amounts of chitosan were detected in the walls of the vegetative forms of S. pombe (Sietsma & Wessels, 1990) and Aspergillus niger (Pochanavanich & Suntornsuk, 2002); besides, it is likely to exist in other ascomycete species where genomic analysis revealed the presence of a variable number of genes encoding chitin deacetylases. In basidiomycetes, there are reports citing their presence in Lentinus edodes and Pleurotus sajo-caju (Crestini, 1996; Pochanavanich & Suntornsuk, 2002), and in the cell walls of the vegetative form of Cryptococcus neoformans that contained higher amounts of chitosan than chitin (Banks, 2005; Baker, 2007). Genome analyses of other basidiomycete species also showed the existence of genes encoding chitin deacetylases, suggesting the wide distribution of chitosan in this group of fungi. In silico analysis of the U. maydis genome revealed the existence of several genes encoding chitin deacetylases (Ruiz-Herrera, 2008), although previous analyses of the walls from its vegetative forms (yeast-like and mycelium) failed to reveal the presence of chitosan (Ruiz-Herrera, 1996). These data suggest the possibility that in this and other basidiomycete species, the polysaccharide may be synthesized at specific stages of the life cycle, other than the vegetative ones, just as occurs with S. cerevisiae.
Our in silico analyses revealed that the single chytridiomycete and microsporidian species analyzed also contained genes encoding chitin deacetylases. These results are further evidence of the ancient origin of chitin deacetylases and chitosan, but as would be expected from the respective levels of chitosan in the cell wall, Mucoromycotina (zygomycetes) species possessed not only the greatest number of enzymes but also the largest number of subclasses of chitin deacetylases (followed by basidiomycetes) (see Fig. 4). These features are suggestive of repeated processes of gene duplication along evolution in these phyla.
Evolutionary relationships of fungal chitin deacetylases. See details in the legend of Fig. 1.
β-1,3-Glucans are polysaccharides widely distributed in fungi and with different cellular locations. There exist intracellular and cell wall-bound β-1,3-glucans with different degrees of polymerization and physicochemical characteristics, but in the majority of fungal species belonging to the clade Dikarya (see Lutzoni, 2004; James, 2006; Hibbett, 2007), this polysaccharide constitutes the most abundant component of the cell walls (see Ruiz-Herrera, 1991, 1992 for a discussion), the structure of the S. cerevisiae cell wall β-1,3-glucan being the most studied (see Manners, 1973). Most of the cell wall β-1,3-glucan is soluble in alkali, leaving an insoluble residue that was originally found to be covalently bound to chitin in A. niger (Stagg & Feather, 1973) and later on in other fungi (Surarit, 1988). β-1,3-Glucan synthesized by protoplasts or in vitro by cell-free extracts from S. cerevisiae appears in the form of microfibrils (Kreger & Kopecká, 1975; Larriba, 1981; Kopecká & Kreger, 1986), and it is also considered to exist in this microfibrillar form within the cell wall. It is important to notice that Mucoromycotina (zygomycetes) do not contain β-1,3-glucans in their vegetative stages, but only in the sporangiospore wall (see Bartnicki-Garcia & Reyes, 1964; Bartnicki-García, 1968), and that in contrast to chitin and chitosan, β-1,3-glucans are restricted to fungi in the clade Opisthokont, being also present in members of the Plantae and Chromista kingdoms.
Synthesis of β-1,3-glucan is catalyzed by specific synthases that utilize UDP glucose as a glucosyl donor. In contrast to chitin synthases, some fungal glucan synthases have been purified to a reasonable degree (see Ruiz-Herrera, 2004 for a review). Interestingly, β-1,3-glucan synthases, independent of their origin, display important structural differences from the rest of the β-glucosyl transferases, suggesting their probable distinct phylogenetic origin (Campbell, 1997, reviewed by Douglas, 2001). The number of β-1,3-glucan synthase-encoding genes in fungal species is variable; there are species that contain a single β-1,3-glucan synthase such as U. maydis, Sporobolomyces roseus and Malazessia globosa, whereas others contain as many as four, as occurs in S. pombe. Noticeably, our in silico analyses revealed that the chytridiomycete and microsporidian species analyzed do not possess homologues of genes encoding β-1,3-glucan synthases.
Phylogenetic analysis of fungal β-1,3-glucan synthases revealed that in contrast to the results obtained with chitin and chitosan, ascomycetes contain a larger number of enzyme classes than the other fungal taxa, revealing several processes of gene duplication and differentiation only in that phylum (Fig. 5). A phylogenetic analysis of the relationships existing between the fungal enzymes and those from the rest of the taxonomic groups showed that the former appeared as a compact group clearly separated from the rest, their closest relatives being β-1,3-glucan synthases from algae (not shown).
Evolutionary relationships of fungal β-1,3-glucan synthases. See details in the legend of Fig. 1.
As described above for β-1,3-glucans, β-1,6-glucans constitute a large group of polysaccharides with different molecular sizes, structures, chemical characteristics and cellular locations (for a review, see Ruiz-Herrera, 1992). The most widely studied β-1,6-glucan is the one present in the yeast wall, a highly branched glucose polymer with an average of c. 350 monomers, covalently linked to β-1,3-glucan and corresponding to about 10% of the total glucans in the S. cerevisiae cell wall (Shahinian & Bussey, 2000). The presence of β-1,6-glucans is restricted to fungi and Chromista, being present in the latter kingdom in species belonging to oomycetes, for example Phytophthora (Bartnicki-Garcia & Lippman, 1967; Zevenhuizen & Bartnicki-Garcia, 1969), Pythium (Blaschek, 1992) and Achlya (Reiskind & Mullins, 1981) (reviewed by Ruiz-Herrera, 1992). β-1,6-Glucans in fungi have been studied mostly in ascomycetes, and only a few analyses have reported their presence in basidiomycetes: higher basidiomycetes such as Sclerotium rolfsii and Schizophyllum commune (see Rau, 2004), some in relation to their putative therapeutic properties (e.g. Daba & Ezeronye, 2003) and in the cell wall of U. maydis (Ruiz-Herrera, 2008). No report exists on the presence or absence of β-1,6-glucans in chytridiomycetes or microsporidians. Preliminary experiments showed their absence in the cell walls of sporangiospores (walls from their vegetative stages have no glucose polymers) from P. blakesleeanus and Rhizopus oryzae (unpublished data).
A search for genes encoding enzymes involved in β-1,6-glucan synthesis was performed in S. cerevisiae through the isolation of mutants resistant to the killer toxin K1 (kre) (Al-Aidroos & Bussey, 1978; Boone, 1990; Brown, 1993; reviewed by Shahinian & Bussey, 2000), taking into consideration that β-1,6-glucan is the receptor for this toxin (Bussey, 1979; Hutchins & Bussey, 1983). The results obtained demonstrated that the synthesis of β-1,6-glucan is a complex process where at least 10 distinct proteins (some with several homologues) are involved: Kre1p, Kre5p, Kre6p, Kre9p, Kre11p, Cne1p, Cw41p/gls1p, Knh1p, Rot2p/Gls2p and Skn1p. These proteins have different cellular locations, and their specific functions are still controversial and unclear. Mutation of several of them leads to reduction in the levels of β-1,6-glucans, severe alterations in morphology and growth and in some cases they are synthetically lethal (see Table 2). Considerable confusion exists in the field because an acceptable mechanism for β-1,6-glucan synthesis is still missing, and due to the fact that to date the real β-1,6-glucosyltransferase(s) involved and even the glucosyl donor are unknown.
Characteristics of the proteins described to be involved in β-1,6-glucan biosynthesis in Saccharomyces cerevisiae
Putative function and/or characteristics
Surface protein. Involved in glucan assembly. K1 killer toxin receptor?
Plasma membrane, cell wall GPI protein
Similarity to UDPG: glycoprotein glucosyltransferase. Protein initial glycosylation?
Similarity to glucosyl transferases. Type II membrane protein. Homologue of Skn1
Mutations result in an aberrant morphology and growth and mating defects
Subunit of transport protein particle involved in ER to Golgi transport
Calnexin/calreticulin-like. ER membrane protein involved in protein folding control
ER glucosidase I. Involved in β-1,6 glucan assembly and N-glycolylation
Functional homologue of Kre9. Overproduction restores kre9 defects
ER glucosidase II. Involved in β-1,6 glucan assembly and N-glycosylation
Similarity to glucosyl transferases. Type II membrane protein. Homologue of Kre6
Homologues of all the above-described proteins have been described in other ascomycetes (e.g. Mio, 1997) and we identified them by an in silico search. On the other hand, all the basidiomycete species analyzed lacked homologues for genes encoding Kre1p, Kre9p, Kre11p and Cne1p. In the zygomycete and the chytridiomycete species analyzed, in addition to these genes, homologues of KRE6 and SKN1 genes were also absent. Interestingly, E. cuniculi, the single member of the Microsporidia phylum analyzed, did not contain homologues of any of the enzymes putatively involved in β-1,6-glucan synthesis (Table 3). It is interesting to recall that as mentioned above, β-1,6-glucan was identified in the wall of the basidiomycete U. maydis (Ruiz-Herrera, 2008), suggesting that not all of the 10 proteins identified in S. cerevisiae and ascomycetes in general are strictly required for the synthesis of the polysaccharide. Phylogenetic analyses of the four proteins putatively involved in the synthesis of β-1,6-glucans present in all fungal phyla revealed their separation in the fungal groups analyzed (not shown).
Conservation of enzymes involved in the synthesis of the mannoproteins outer chain
Phylum (species analyzed)
Number of species with homology
In species with more than one homologue, only the one with the highest homology is shown.
All bona fide extracellular proteins are glycosylated, hence their denomination as glycoproteins. In eukaryotic organisms, there exist two main types of binding of the carbohydrate moieties to the proteins: O-glycosylation, where a short chain of sugars binds to serine or threonine residues, and N-glycosylation, in which the sugar chain binds to asparagine. The relative proportion of the carbohydrate moiety may vary from <10% to >95% of the whole glycoprotein Mr. Glycoproteins are essential for all eukaryotes, as shown by the observation that mutations or inhibitors that block the early steps in N- or O-glycosylation are lethal. Protein O- or N-glycosylation initiates cotranslationally during protein synthesis in the ER, and continues during the transit of the protein through the Golgi. O-glycosylation involves the addition of only a limited number of different sugars depending on the organism (at most five mannosyl residues in fungi) (see the review by Goto, 2007). Synthesis of the most internal moiety (the inner core) of the N-bound oligosaccharide is universal, but further modifications and extension of this core depend on the type of organism. Characteristically, in fungi, the size of the carbohydrate fraction of the glycoprotein may extend, mostly by the action of mannosyl transferases, to give rise to high Mr molecules, constituting the wall fraction known as mannan. The main sugar present in fungal glycoproteins is mannose, but diacetylchitobiose is universally present, forming the bridge between the asparagine residue of the protein and the carbohydrate fraction. Other sugars may also be present in variable amounts, depending on the fungal species (see reviews in Ruiz-Herrera, 2004; Sentandreu, 2004; Jigami, 2008). The first step in protein O-glycosylation in S. cerevisiae (the best-studied model) involves transfer of a mannosyl residue using dolichyl-phosphate mannose as a substrate, whereas the substrate used for transfer of the rest of the mannosyl units is GDP mannnose. The classical studies from Ballou (e.g. Frevert & Ballou, 1985; Ballou, 1990) with S. cerevisiae as a model demonstrated that synthesis of the outer chain of the mannan moiety of the glycoproteins was catalyzed by specific mannosyl transferases, most of them different from the ones responsible for the synthesis of the inner core oligosaccharide. Further work has revealed that elongation of the inner core is initiated by addition of an α-1,6-linked mannosyl residue by the enzymatic action of Och1p mannosyltransferase, and it is further extended by the sequential action of different mannosyl transferases to form an oligosaccharide of variable length made up of α-1,6-linked mannose, with α-1,2- and α-1,3-linked mannose branches, plus some variable modifications (for a review of the enzymes involved, see Jigami, 2008). In S. cerevisiae, growth of the outer chain is catalyzed by an α-1,6-mannosyl transferase encoded by the MNN9 gene, and its branching is catalyzed by two α-1,2-mannosyl transferases: Mnn2, catalyzing the addition of the first mannosyl residue, and Mnn5, responsible for the addition of further mannosyl residues, whereas Mnn1 is the mannosyl transferase involved in the attachment of α-1,3-bound mannosyl residues.
In this study, we proceeded to the comparison of the genes encoding the enzymes involved in the synthesis of the outer chain of glycoproteins (mannan) in the several fungal groups included in our analyses using the sequences of the whole S. cerevisiae enzymes for comparison. As expected, homologues of all the genes encoding the above-mentioned enzymes were found to be present in the analyzed ascomycete species with homology values between e−33 and e−144. On the other hand, no homologue for any of the genes was identified in the microsporidian E. cuniculi. Analysis of homology in species of the rest of the phyla yielded variable results (see Table 3).
GPI proteins are glycoproteins that contain a GPI anchor, which is added post-translationally to a motif located at their C-terminus. The GPI anchor remains associated with the cell membrane and its bound polypeptide protrudes into the periplasmic space. GPI proteins are present in all eukaryotic groups, mostly associated with the plasmalemma. De Nobel & Lipke (1994) indicated their possible role within the fungal cell wall, and later on a variable number of GPI proteins were found to be covalently associated by a glycosidic linkage to β-1,6-glucans in the cell wall of most of the ascomycete species analyzed (e.g. Yin, 2005; Castillo, 2008). Regarding the factors involved in their selective location at the plasma membrane or the cell wall, Mao (2008) obtained evidence that a short stretch of amino acids at the C-terminus of GPI proteins was involved in providing their specificity in targeting to either location in C. albicans. Whether this is a general characteristic in other fungi remains to be determined.
As would be expected, in silico analysis revealed the presence of a number of GPI proteins in all species of the ascomycetes, basidiomycetes, mucoromycetes and chytridiomycetes analyzed (see Ruiz-Herrera, 2008). In the microsporidian E. cuniculi, at least one GPI protein extracted from the spore wall with reducing agents has been described (Xu, 2006). These data are evidence that GPI proteins are distributed in all the fungal phyla. Nevertheless, repeated proteomic analyses have failed to demonstrate the presence of covalently bound GPI proteins in the cell wall of U. maydis (J. Ruiz-Herrera & R. Sentandreu, unpublished data).
Synthesis of the GPI moiety occurs in the ER by a pathway that is conserved in eukaryotic organisms, and independent of the synthesis of the glycoprotein. In the final step occurring at the plasmalemma, the characteristic C-terminal signal of the protein is removed by a GPI transamidase and replaced by the GPI moiety (for reviews, see Brown & Waneck, 1992; Orlean & Menon, 2007; Pittet & Conzelmann, 2007). The transamidase complex is made up of five components: Gaa1/GPAA, Gpi8/PigK, Gpi16/PigT, Gpi17/PigS and Gab1/PigU (see Orlean & Menon, 2007). We identified homologues for all the encoding genes in the ascomycetes, basidiomycetes, zygomycetes and chytridiomycete species analyzed. The exception was the microsporidian E. cunicili, which contained only a homologue of the Gpi8 gene (also involved in the synthesis of the Gpi moiety), a result in agreement with its phylogenetic distance.
It has been suggested that transfer of the protein to the β-1,6-glucan present in the wall occurs by a transglycosylation reaction in which the GPI moiety is cleaved between the mannosyl 1 and the glucosaminyl residues and transferred to the terminal or an internal glucosyl residue of the glucan, leaving the hydrophobic GPI residue attached to the plasmalemma. Nevertheless, the enzymes involved and the mechanism of the reaction are still unknown (for a review, see Orlean & Menon, 2007).
Pir proteins are characterized by the presence of several glutamine-rich repeats with the consensus sequence Q[IV]XDGQ[IVP]Q, a signal peptide, a Kex2 cleavage site and a C-terminal domain with four cysteine residues with the structure-C-65/66-C-16-C-12-C. Pir proteins were initially identified in S. cerevisiae (Toh-e, 1993), and later on reported to be present in a number of ascomycete species (e.g. Martínez, 2004). It has been demonstrated that Pir proteins play an important role in the structure and stability of the cell wall of ascomycetes (Mrsa & Tanner, 1999; Martínez, 2004). Pir proteins can be extracted from the cell walls by treatment with dilute alkali at a low temperature. The alkali-sensitive bond has been identified as an ester linkage between the γ-carboxyl group of glutamic acid and the hydroxyl groups of glucose (Ecker, 2006).
Our blast analyses using the most conserved residue (SQIGDGQVQATSAAT) failed to identify Pir proteins sensu stricto in basidiomycetes or in other fungal species (Ruiz-Herrera, 2008). In fact, these proteins appear to be specific of ascomycetes in the entire living world. Nevertheless, it must be stressed that other proteins, unrelated to Pir, and containing different types of repeats, are present in the fungal walls. Perhaps the best known of these are S. cerevisiae flocculins (Flo1, Flo5, Flo9, Flo10 and Flo11), and Aga1, Tir1, Tir4 and Dan4 (Verstrepen, 2005) and mating-specific yeast agglutinins (Lipke & Kurjan, 1992; Cappellaro, 1994). Further examples of surface proteins containing multiple repeats have been identified in U. maydis, among them, repellent Rep1 and the atypical hydrophobin Hum3, which confer hydrophobic properties to the cell wall of this fungus (Teertstra, 2006), and Hum3 and Rsp1, proteins involved in its pathogenic behavior (Müller, 2008). It is therefore feasible that Pir-unrelated proteins containing different repeats may be present in the walls of species belonging to other fungal groups.
Modern analysis of the phylogeny and taxonomy of fungi has revealed that the former Chytridiomycota and Zygomycota groups are not monophyletic (Lutzoni, 2004; James, 2006; Hibbett, 2007). In the present analysis, we mostly made use of data corresponding to those fungi whose genomes have been sequenced and are available. Because of this limitation, chytridiomycetes are represented by a single species of the new phylum Chytridiomycota (now separated from phyla Neocallimastigomycota and Blastocladiomycota), and zygomycetes by species belonging to the new phylum Mucoromycotina. Also, a single species is representative of microsporidians. On the contrary, the two phyla belonging to the clade Dikarya, Basidiomycota and Ascomycota, are over-represented.
Despite this possible source of criticism, we consider that the data described here are valuable in the analysis of the two main questions approached in this study: (1) is the cell wall of S. cerevisiae a true paradigm of the structure of the fungal cell wall? Or do we have to consider that the cell walls of members of the different fungal groups have significant differences and variations? and (2) can we obtain information on the evolution of the structure of the fungal cell wall by means of this type of analysis?
Regarding the first question, we can conclude that the data presented here are indicative that the use of the wall from S. cerevisiae and related ascomycete yeasts as the paradigm of the fungal cell wall is probably inadequate. The only common feature of the cell walls from organisms belonging to the several fungal phyla is their architecture, which, as mentioned in the Introduction, depends on the association of microfibrils of structural polysaccharides with an amorphous material made up mainly of proteins and other types of polysaccharides. For the second question, it is clear that the analysis of the enzymes involved in the synthesis of the most important polysaccharides and the nature of the covalently linked proteins provided information that may suggest the steps occurring in the differentiation of the cell wall along the evolution of fungi.
The example of chitin synthases is illustrative of this matter. Chitin is a polysaccharide specific of eukaryotes, and its presence in eukaryote kingdoms (Fungi, Animalia, Chromista and Protozoa) suggests that the appearance of the enzymes responsible for its synthesis was an early event in evolution, probably deriving from the ancestor of different β-glycosyl-transferases, as suggested by their common structural features (see Ruiz-Herrera, 2002). Data of the phylogenetic analysis of chitin synthases, and the relationship of division 2 Chs with chitin synthases from members of other chitin-containing kingdoms, suggest that this type of enzymes were the first ones to appear during evolution. This suggestion is supported by the observation that all fungal phyla contain enzymes belonging to this division, whereas only ascomycetes contain all Chsp classes from division 1, basidiomycetes lack one of them (class I) and zygomycetes, and chytridiomycetes possess only one (class II), whereas Microsporidia contain none. This observation is interesting, because it has been indicated that microsporidians probably belong to the first diverging fungal branch during fungal evolution (James, 2006). Also, the absence of a class V chitin synthase in the microsporidian species analyzed may be taken as further evidence, in addition to those discussed above, that class IV preceded the appearance of class V chitin synthases. The possibility that this deficiency in E. cunicili could be attributed to the loss of a class V gene is possible, but it is not supported by the phylogenetic analysis. Accordingly, it may be suggested that the original class IV gene gave rise, through duplication and modification, to a class V gene that most probably acquired the myosin head motif, lost later on in some of the enzymes belonging to different species.
Division 2 Chs enzymes have an average Mr larger than division 1 Chs, and contain a large N-terminal tail without conserved motifs (Ruiz-Herrera, 2002). It would not be a wild speculation to consider that the first gene encoding a division 1 enzyme (most probably belonging to class II CHS, as supported by the phylogenetic analysis and the distribution of the different classes of Chs in the various fungal phyla) also originated by gene duplication and modification, a phenomenon amply represented in gene evolution. In the present case, the modification would have involved deletion of the region encoding the N-tail from a division 2 gene, which left its catalytically important motifs untouched, and its further variation during evolution. The later appearance of genes encoding class III enzymes occurring only in the common ancestor of Dikarya, and class I CHS genes only in Ascomycota after its separation from Basidiomycota, would have taken place by a similar duplication–modification mechanism. As indicated above, in contrast to animals and protists, members of the group of oomycetes (Chromista) do not contain Chs related to division 2 fungal enzymes; they only possess enzymes related to division 1. Two possible explanations exist for this situation: their common ancestor acquired the gene that gave rise to the two classes of Chs present in the group by horizontal transfer from the common ancestor of Chitridiomycota, Zygomycota and Dikarya, or less likely, they lost the ancestor genes encoding Chs related to division 2 and class III enzymes. Modern concepts on the evolution of eukaryotes strongly support the first possibility (Cavalier-Smith, 2004). In this sense, it is also important to recall the existence of the division 1-related CHS gene present in chloroviruses (Kawasaki, 2002), which reminds us of CHS gene mobility from fungal ancestors.
Chitin deacetylases exist in all fungal phyla, but only zygomycetes appear to contain their product, chitosan, in all their life-cycle forms, being, in some of them, the most important wall component. In ascomycetes, chitosan has been identified mostly in the sexual spores, interestingly being synthesized from the product of enzymes belonging to division 2 (Pammer, 1992), although isolated reports have described its presence in the vegetative forms of S. pombe and A. niger. Only a few reports have described the presence of chitosan in the vegetative forms of some basidiomycetes. The most thorough analyses correspond to the pathogenic basidiomycete C. neoformans, which contains higher amounts of chitosan than chitin in its cell wall, interestingly also being synthesized from the product of a division 2 enzyme (Chs3; Banks, 2005). It is therefore not surprising that zygomycetes contain the highest number of genes encoding chitin deacetylases. The presence of chitin deacetylases in all the fungal groups analyzed, as well as in some insects and in amoeba, suggests that their encoding genes appeared early in evolution, before the separation of protozoa and opisthokonts, but, after their split from the rest of the eukaryotic kingdoms, in which its presence has not been reported. This suggestion is in line with evolutionary data from Cavalier-Smith (2002).
Regarding the role of this apparently superfluous enzyme, which might even be considered as deleterious because it converts the stiff component responsible for the structure of the cell wall into a less resistant (chemically and mechanically) component, it might be argued that the presence of a chitosan layer in the cell wall of zygomycetes makes them more resistant to chitinases, which are probably more abundant in nature than chitosanases, and perhaps to some other deleterious factors, because of its physicochemical characteristics. Similarly, in C. neoformans, it was described that chitosan is necessary for the integrity of the cell wall (Baker, 2007) in analogy to yeast ascospores. In yeasts, the polysaccharide is mostly restricted to the cell wall of ascospores, and chitin deacetylases are required for its assembly, and for providing its rigidity (Christodoulidou, 1996, 1999). Mutation of the gene encoding the deacetylase involved in chitosan formation in S. pombe impairs ascospore formation (Matsuo, 2005). This characteristic may be related to the resistance of a structure that must remain dormant for long periods of time. Accordingly, ascospores are more resistant than vegetative cells to acid, alkali, ether, salt or heat stress as described by Coluccio (2008). These authors also reported that ascospores from S. cerevisiae and S. pombe were resistant to the passage through the digestive tract of Drosophila due to their chitosan layer (and a polymer containing dityrosine in S. cerevisiae), and suggested that this resistance is a mechanism for spore dispersal by means of insects that feed on them. Accordingly, it may be hypothesized that the capacity to turn on the gene(s) encoding chitin deacetylases during sporulation, mostly in ascomycetes, appeared as a selective factor during evolution.
In contrast to chitin and chitosan, β-1,3-glucans are more abundant in the cell walls of ascomycetes and basidiomycetes than in zygomycetes, where the presence of the polysaccharide is restricted to the sporangiospore wall. In some chytridiomycetes, the existence of glucose-containing polymers has been described (e.g., Sikkema & Lovett, 1984), although their chemical structure has not been analyzed. Because in silico analysis did not reveal the presence of a homologue of a β-1,3-glucan synthase in B. dendrobatidis, it may be suggested that the polysaccharide detected in other chytridiomycete species is probably not β-1,3-glucan. This result and the absence of β-1,3-glucan synthase also in E. cuniculi suggest an interesting possibility of the presence of the enzyme only in zygomycetes, ascomycetes and basidiomycetes.
Fungal β-1,3-glucan synthases, the same as those from plants, algae and members of the kingdom Chromista, have significant structural differences from the rest of the glycosyltransferases: they lack the characteristic conserved amino acids (three Asp residues and the QXXRW motif), associated with the catalytic activity of the rest of the β-glycosyl transferase families including cellulose and chitin synthases (Campbell, 1997; reviewed by Douglas, 2001), and the UDP glucose-binding site (R/K) XGG from glycogen synthases (Farkas, 1990). These characteristics and the absence of β-1,3-glucan synthases in the rest of the opisthokont species make unlikely the possibility that acquisition of these enzymes by fungi occurred by a parallel and independent process. It is more feasible that after separation of microsporidians and chytridiomycetes from the fungal evolutionary line, a gene encoding β-1,3-glucan synthase from a precursor of the Plantae or Chromista kingdoms was captured by the common ancestor of the rest of the fungal phyla through a horizontal transfer process. As with chitosan in ascomycetes, the capacity to synthesize β-1,3-glucans remained restricted to spores in zygomycetes as a relic process of probable ecological advantage. On the other hand, the polymer became an important component of the cell wall in the Dikarya species. Knowledge of the factors involved in the regulation of chitosan and glucan synthesis, respectively, in ascomycetes or zygomycetes, would undoubtedly provide important information on the prevailing ecological conditions under which the antecessors of these fungi survived.
As indicated previously, the presence of β-1,6-glucans in opisthokonts is restricted to fungi, and its mechanism of synthesis is complicated and has not been clarified, apparently involving at least 10 proteins in the process carried out by ascomycetes (see the review by Shahinian & Bussey, 2000). Because homologues of only six of these proteins (Kre5, Kre6, Cne1, Cwh41/Gls1, Rot2/Gls2 and Skn1) are present in basidiomycetes, it may be suggested either that the rest of the proteins have other functions only indirectly related to synthesis of the polysaccharide or that their roles are redundant. The observation that mutation in ascomycetes of some of the genes absent in basidiomycetes leads to an aberrant morphology and growth defects (not observed of course in normal basidiomycetes) suggests that the first possibility is more likely. As indicated above, zygomycetes and chytridiomycetes lack, in addition Kre6 and Skn1, proteins with similarity to glycosyl transferases, whereas Microsporidia lack homologues to all of them. These results suggest that those two proteins might be indispensable for β-1,6-glucan synthesis, and their absence would explain why the polysaccharide is not present in these fungal phyla. According to these data, it is likely that acquisition of the capacity to synthesize this polysaccharide occurred in a common ancestor of Dikarya after departure of zygomycetes from the fungal evolution line. As already mentioned, β-1,6-glucans are also present in chromista, and absent in plants and the rest of the Opisthokont clade. Whether appearance of the genes encoding the hypothetical catalytic peptide(s) for β-1,6-glucan synthesis (homologues of KRE6 and SKN1?) in Fungi and Chromista kingdoms was due to separate and independent processes, or once present in one group they were transferred to the precursor of the other one by a horizontal transfer mechanism (in what direction?) as suggested above for β-1,3-glucans, remains to be determined.
Regarding the synthesis of mannans, the outer chain of mannoproteins, it is apparent that drastic changes occurred during evolution of the genes encoding the enzymes involved, as revealed by the variable degrees of homology among the corresponding genes of the several fungal phyla. These changes most probably gave rise to alterations in the structure of the main types of outer chains from the glycoproteins present in several fungal groups, and show the differences existing with the ascomycete model, best represented by S. cerevisiae.
The presence of covalently bound Pir and GPI proteins in the walls of ascomycetes is an interesting feature. The acquisition of the capacity to synthesize proteins that are linked to the polysaccharide mesh of the cell wall through β-1,6-glucans or β-1,3-glucans, respectively, undoubtedly represented a landmark for a change in the organization of the cell wall. The existence of Pir proteins in other fungi besides ascomycetes has already been clearly discarded, but as indicated, GPI proteins are present in all the fungal groups, bound to the cell membrane through the GPI moiety. Whether GPI proteins covalently bound to β-1,6-glucans exist in fungi other than ascomycetes remains unknown. It is unfortunate that, with the exception of U. maydis, no analyses have been performed to detect the presence of GPI proteins covalently bound to the cell wall in basidiomycetes, zygomycetes, chytridiomycetes or microsporidians. Nevertheless, the absence of β-1,6-glucans in the last three phyla makes the existence of the typical ascomycete cell wall-bound GPI proteins impossible. Identification of the enzymes involved in the covalent attachment of GPI proteins to β-1,6-glucans would be the ideal tool to determine by in silico techniques as to whether they exist in other fungi besides ascomycetes.
Based on the data and ideas described above, tentatively, we may schematically represent the evolution of the organization of the fungal cell wall as illustrated in Fig. 6. According to this hypothetical scheme, the first step in the development of the eukaryotic cell wall would have been the acquisition of the capacity to synthesize microfibrillar polysaccharides, chitin in the case of the common ancestor of the opisthokonts, by chitin synthases most probably related to modern class IV enzymes. This feature would allow the formation of an exoskeleton that permitted cell growth in hypotonic media and retention of noncovalently bound important proteins, some with enzymatic activity, in the periplasmic space. This would have been followed by the appearance of a more plastic wall component (chitosan) by the activity of chitin deacetylases (it is important to recall that these enzymes are also present in protozoa and animals). Once fungi separated from the other Opisthokont phyla, duplication and modification may have occurred of the gene encoding the class IV enzyme ancestor to give rise to genes coding for class V chitin synthases. Next, when Microsporidia departed from the fungal evolutionary line, the appearance of class II CHS genes possibly, as discussed above, through duplication and modification of a gene encoding a division 2 enzyme, might have taken place. After departure of chytridiomycetes from the main fungal evolutionary line, the common ancestor of the rest of the fungal phyla would have captured a gene encoding β-1,3-glucan synthase, and not just after separation of the phylum Zygomycota that genes encoding class III Chs would have evolved probably by the duplication and modification most likely of a class II CHS gene. Also around this time genes encoding catalytic enzymes involved in β-1,6-glucan synthesis would have appeared or would have been acquired by horizontal transfer, as suggested above. Finally, after the split of the Ascomycota and Basidiomycota phyla, the ancestor of the former would have developed the capacity to covalently bind exocellular proteins to β-glucans, and acquired class I CHS genes possibly by the same mechanism as that suggested for the origin of class II and class III CHS genes.
Hypothetical scheme of the evolution of the fungal cell wall structure. CHS, chitin synthases; Cda, chitin deacetylases; and GS, glucan synthases. Arrows with question marks suggest the possible origin of genes encoding β-glucan synthases by horizontal transfer.
In summary, our data point out that cell walls from the different taxonomical groups of fungi have distinct features that characterize them. No general model of the fungal cell walls can be defined, because their structure in the fungal phyla is a recapitulation of the evolutionary history of the different compounds that make them up. This history involves: (1) evolution of polysaccharide synthases, (2) horizontal gene transfer of synthesizing enzymes, (3) development of systems for the selective regulation of specific synthases during the life cycles of fungi, (4) important changes leading to modification of the synthesized products and (5) the capacity to associate different products through covalent linkages, in order to give rise to a coherent structure. Most importantly, although chemically and structurally the fungal cell wall may be different in the several taxonomic groups, it faithfully fulfills its important functions in all of them.
This work was partially supported by CONACYT, Mexico and by CONCYTEG from the State of Guanajuato, Mexico.
(1975) On the nature and formation of the fibrillar nets produced by protoplasts of Saccharomyces cerevisiae in liquid media: an electronmicroscopic, X-ray diffraction and chemical study. J Gen Microbiol 92: 207–220.
(2006) Repellents have functionally replaced hydrophobins in mediating attachment to a hydrophobic surface and in formation of hydrophobic aerial hyphae in Ustilago maydis. Microbiology 152: 3607–3612.