OUP user menu

Sexual reproduction and dimorphism in the pathogenic basidiomycetes

Carl A. Morrow, James A. Fraser
DOI: http://dx.doi.org/10.1111/j.1567-1364.2008.00475.x 161-177 First published online: 1 March 2009


Many fungi in the Basidiomycota have a dimorphic life cycle, where a monokaryotic yeast form alternates with a dikaryotic hyphal form. Most of the dimorphic basidiomycetes are pathogenic on plants, animals or other fungi. In these species, infection of a host appears to be closely linked to both dimorphism and the process of sexual reproduction. Sex in fungi is governed by a specialized region of the genome known as the mating type locus that confers cell-type identity and regulates progression through the sexual cycle. Here we investigate sexual reproduction and lifestyle in emerging human pathogenic yeasts and plant pathogenic smuts of the Basidiomycota and examine the relationship among sex, dimorphism and pathogenesis.

  • Basidiomycota
  • yeast
  • mating
  • sexual reproduction
  • pathogenesis


With over 80 000 described species, the ubiquitous Fungi are a spectacularly successful kingdom. While most are saprobes that decay organic matter, symbionts of plants or harmless commensals, the pathogenic species are of particular interest: an estimated 32% of fungi are plant pathogens, while only c. 0.5% (roughly 400 species) are clinically relevant human pathogens (Shivas & Hyde, 1997; De Hoog, 2000; Hawksworth, 2001; Kirk, 2001). The phylum Ascomycota includes the majority of described fungi, and represents most of the plant pathogens and c. 90% (c. 350) of human pathogenic fungi (Hawksworth, 2001; Mitchell, 2005). Their sister phylum is the Basidiomycota, which numbers an estimated 30 000 species. While the saprobic mushrooms are the largest clade, a significant number of basidiomycetes are pathogenic: the rusts and smuts number about 7000 and 1400 species, respectively, and a further 40 yeast species have been reported to infect humans and animals (Kirk, 2001; Mitchell, 2005).

The Basidiomycota are divided into three subphyla (Table 1). The Agaricomycotina contain the majority of described basidiomycete species (c. 20 000), including the mushroom fungi, the jelly fungi and a variety of yeasts (Hibbett, 2006). The Ustilaginomycotina (c. 1500) comprise the majority of the smut fungi plus some dimorphic pathogens, while the Pucciniomycotina (>8000 species) comprise mainly the obligately plant pathogenic rust fungi plus an array of parasites and saprotrophs (Aime, 2006; Begerow, 2006; Hibbett, 2007).

View this table:

Classification of selected pathogenic basidiomycete genera

TremellomycetesCryptococcus/FilobasidiellaBipolar or asexual
UstilaginomycetesUstilagoBipolar or tetrapolar
SporisoriumBipolar or tetrapolar
Incertae sedisMalasseziaBipolar or asexual
Sporobolomyces/SporidiobolusAsexual or bipolar
Rhodotorula/RhodosporidiumAsexual or bipolar
  • * Although Malassezia globosa is hypothesized to be bipolar based on the putative MAT locus, mating has never been observed (Xu, 2007).

A variety of fungi in the Basidiomycota have a single-celled growth form called a yeast, which is derived from the budding of a meiospore (basidiospore). The yeast growth form is a rounded or an elongated cell that reproduces asexually by budding, fission or production of forcefully ejected ballistoconidia (Flegel, 1977; Fell, 2001). Yeast states are widespread across the three subphyla, and the term bears no taxonomic implication. Species that alternate between a single-celled yeast form and a filamentous growth form are termed ‘dimorphic’, in contrast to ‘monomorphic’ species that have only one known growth form (Fell, 2001). The filamentous form consists of long, branching tubular cells known as hyphae, which are divided into compartments by septa and grow at the apical cell (Steinberg, 2007). Dimorphic species are common in the three subphyla of the Basidiomycota, where the unicellular yeast phase is usually monokaryotic and alternates with a dikaryotic hyphal phase. Monomorphic species include the mushrooms and the rusts, where spores germinate by production of hyphae, and a number of exclusively asexual yeasts (Bandoni, 1995; Boekhout, 1998).

Most sexual basidiomycete yeasts have a dimorphic life cycle (Fig. 1). Initially, a haploid basidiospore (either a forcibly discharged ballistospore or a budded sporidium/statismospore) germinates to produce the free-living yeast. The yeast state can colonize a variety of substrata, and may be saprobic on soil or aquatic habitats, or epiphytic on plants. Basidiomycete yeasts can reproduce asexually by enteroblastic budding or via production of ballistoconidia (Fell, 2001). In the presence of a partner of compatible mating type, yeast cells produce conjugation tubes and ultimately fuse to produce a dikaryotic hyphal cell. Basidiomycetes characteristically maintain a prolonged dikaryotic state, where each cell possesses two unfused parental nuclei, and undergo a complex mode of mitosis involving the formation of clamp cells. Nuclear fusion and meiosis occur in a basidium to produce a tetrad of haploid basidiospores (Casselton & Olesnicky, 1998). In fungi, the asexual (mitosporic) and sexual (meiosporic) forms are known as the ‘anamorph’ and the ‘teleomorph’, respectively (Seifert & Gams, 2001).


Two paradigms of dimorphic basidiomycete pathogenesis. The single-celled yeast form of the smut Ustilago maydis (left) is nonpathogenic. Upon mating with a compatible partner, the fungus switches morphology to an infectious hyphal form, where it can invade a host plant. The single-celled yeast form of Cryptococcus neoformans (right) switches to a dikaryotic hyphal form after mating, which may potentially be free-living, phytopathogenic or mycoparasitic. Karyogamy and meiosis later occur in the basidium, where chains of the potentially infectious basidiospores are produced. Crucial to both fungi is the act of sexual reproduction and a dimorphic switch between a yeast and a hyphal form to initiate infection of a plant or an animal host.

Apart from the intensively studied smut fungi, little is known about the hyphal phase of most species in the environment. However, the majority of dimorphic basidiomycetes are parasitic in the hyphal phase, usually on plant or fungal hosts (Bandoni, 1995). The vast majority of the Ustilaginomycotina, including the Ustilaginomycetes and the Exobasidiomycetes, are dimorphic and phytopathogenic, and in the case of the smut fungi, the hyphal phase is an obligate parasite. Plant pathogens are also common in the smuts of the Microbotryomycetes of the Pucciniomycotina, while the only dimorphic members of the Pucciniomycetes, the Septobasidiales, are entomopathogenic (Bauer, 2001). Mycoparasitism appears to be a common strategy among dimorphic basidiomycetes, and is often associated with the presence of haustoria, colacosomes or lenticular body organelles (Bandoni, 1995). Mycoparasites are widespread in the Pucciniomycotina, particularly the Agaricostilbomycetes, Cystobasidiomycetes and the Microbotryomycetes, and many species currently considered to be saprobic may be capable of parasitizing fungi (Swann, 2001). Most of the Tremellomycetes have been proposed to be mycoparasitic in their dikaryotic hyphal form (Bandoni, 1995; Wells & Bandoni, 2001). Intriguingly, these changes between the yeast and the hyphal form may potentially be linked with pathogenicity in the dimorphic basidiomycetes via the process of sexual reproduction (Madhani & Fink, 1998; Nadal, 2008).

Sexual reproduction in pathogenic fungi

The majority of eukaryotes undergo sexual reproduction between two individuals to produce recombinant, genetically distinct progeny (Otto & Nuismer, 2004). Sexual reproduction increases genetic variability within a population and can create individuals with increased or decreased fitness (Barton & Charlesworth, 1998). Despite the substantial energy cost associated with sexual reproduction, it is generally considered to be a broadly beneficial process (Goddard, 2005).

Within the Ascomycota, sexual reproduction is widespread among the plant pathogens, such as Magnaporthe oryzae and Leptosphaeria maculans (Sexton & Howlett, 2006). In contrast, the absence of a sexual cycle appears to be commonplace among animal pathogens. In Candida albicans, the most common human fungal pathogen, evidence for meiosis is yet to be found despite the recent elucidation of the MAT locus, mating and a parasexual cycle (Hull & Johnson, 1999; Hull, 2000; Magee & Magee, 2000; Bennett & Johnson, 2003). Sexual reproduction remains similarly elusive in Aspergillus fumigatus, despite the presence of a sexual cycle in its close relative Neosartorya fischeri and population genetic studies that suggest the occurrence of meiotic recombination outside the host (Paoletti, 2005). Accordingly, infections involving animal pathogenic ascomycetes are generally acquired via inhalation of infectious cells produced through asexual modes of reproduction, such as airborne conidia and arthroconidia.

In contrast to this paradigm is the prevalence of sexual reproduction in the pathogenesis of the dimorphic basidiomycetes. Although the plant-infecting smut fungi can persist in an asexual saprophytic yeast phase, sexual reproduction and host invasion are essential for completion of the life cycle. Similarly, most dimorphic basidiomycetes parasitize plants or fungi only in the sexual hyphal phase (Boekhout, 1998). On the other hand, there are a number of emerging human pathogens from the Basidiomycota that infect in the yeast form. In some of these species, particularly Cryptococcus neoformans, it has been suggested that the infectious agent may be the sexually produced basidiospores, and a dimorphic transition from the hyphal to the yeast form may be important to initiate the infection process (Cohen, 1982; Heitman, 2006).

The molecular basis of sexual compatibility: the MAT loci

Sexual reproduction in fungi is governed by the mating type (MAT) locus, a specialized genomic region that confers cell-type identity. Fungi can be self-compatible for mating, with no need for a genetically distinct mate (homothallic), self-incompatible for mating (heterothallic) or, counter-intuitively, a combination of the two. Homothallism is characterized by dikaryotic hyphal cells that develop in the absence of another mating partner (Whitehouse, 1949; Deacon, 1997). In heterothallic species, two cells of a compatible mating type conjugate and produce a dikaryotic mycelium (Fell, 2001). In many heterothallic basidiomycete fungi, two distinct genetic loci reside in the MAT locus and determine sexual compatibility: one locus encodes homeodomain transcription factors, while the other encodes pheromones and pheromone receptors. These genes establish the ‘sex’ of the organism, enable self-non-self recognition and regulate the formation and progression of the dikaryotic sexual form. When these two loci are unlinked in the genome, the organism is termed ‘tetrapolar’, and four possible mating types can segregate independently at meiosis (Fraser & Heitman, 2003; Giraud, 2008).

Alternatively, if the pheromone/receptor locus and the homeodomain locus form one contiguous locus that segregates as a single unit at meiosis, the organism is termed ‘bipolar’. Bipolar species are widespread in the Basidiomycota and include the oleaginous red yeast Rhodosporidium toruloides of the Pucciniomycotina, the sugarcane smut fungus Ustilago scitaminea of the Ustilaginomycotina and the encapsulated yeast Cryptococcus gattii of the Agaricomycotina (Fischer & Holton, 1957; Fell, 2001). In the Ascomycota and the Glomeromycota, the MAT locus is generally small in size and invariably encodes one or more transcription factors; although other genes may be present in the locus, the pheromone and pheromone receptor genes are absent (Butler, 2004; Idnurm, 2008).

Tetrapolarity is a feature unique to the Basidiomycota. An estimated 55–65% of the agaricomycete mushrooms are tetrapolar, often with dozens of alleles at each locus, leading to potentially thousands of mating types (Raper, 1966). A number of other species from all three basidiomycete subphyla have a tetrapolar mating system, such as the mycoparasitic jelly fungus Tremella mesenterica of the Agaricomycotina, the psychrophilic yeast Leucosporidium scottii of the Pucciniomycotina and the phytopathogenic smut Ustilago maydis from the Ustilaginomycotina (Fell, 2001). Although the idea is contentious, tetrapolarity has been proposed to be the ancestral state of the Basidiomycota, and various models have proposed the derivation of a bipolar mating type from the tetrapolar state (Raper, 1966; Burnett, 1975; Bakkeren, 1992; Hibbett & Donoghue, 2001; Fraser, 2004; Fraser & Heitman, 2004).

To attract a partner, fungi can produce signalling molecules to indicate both their presence in the environment and their mating compatibility. Such pheromone production has been studied extensively in the model ascomycete Saccharomyces cerevisiae. The first stage involves the production of a-factor or α-factor precursor molecules, short polypeptides that undergo extensive post-translational modifications to yield the mature pheromone (Kurjan & Herskowitz, 1982). Following export, the pheromone molecule binds with a cognate heterotrimeric G-protein-coupled receptor found on cells of the opposite mating type; the a-factor binds to the Ste3 receptor on α cells, while the α-factor binds to the Ste2 receptor on a cells (Bender & Sprague, 1986). Binding of the ligand initiates a mitogen-activated protein (MAP) kinase cascade that leads to cell cycle arrest, the formation of a projection from the cell in the direction of the pheromone gradient and cell fusion (Leberer, 1997). The S. cerevisiae genes encoding pheromones and receptors do not lie in a specialized region of the genome, and while all cells have the genetic material required for the production of both pheromones, they only produce the one appropriate for their mating type (Herskowitz, 1992).

In the basidiomycetes, the pheromone and receptor genes are often encoded within a MAT locus, and are a determinant of mating type. Only homologues of the a-factor pheromone and its corresponding Ste3 receptor type are found, and never the alternate α-factor/Ste2 receptor system, which appears to be unique to the Ascomycota (Fraser, 2007). Surprisingly, even the agaricomycete mushrooms use a pheromone/receptor system (Fig. 2), despite the fact that mushroom hyphae fuse indiscriminately (Wendland, 1995; O'Shea, 1998). In this case, coupling of a compatible pheromone and receptor in the mushrooms appears to be essential for maintenance of the dikaryotic cell and not initial mate attraction (Vaillancourt, 1997; Casselton & Olesnicky, 1998).


Tetrapolar MAT loci of the Basidiomycota. In the tetrapolar system, mating type is determined by two unlinked genomic loci – the pheromone/receptor locus (a in smuts and B in mushrooms) and the homeodomain locus (b in smuts and A in mushrooms). Where present in the smut fungi, the pheromone locus has few alleles (biallelic in Ustilago maydis and triallelic in Sporisorium reilianum), while the separate homeodomain locus is multiallelic (19 alleles in U. maydis and 5 alleles in S. reilianum). The tetrapolar mating type is uncommon among the smuts. In contrast, tetrapolarity is widespread among the mushroom fungi, where multiple alleles at both loci lead to thousands of possible mating types, as in Coprinopsis cinerea and Schizophyllum commune. Interestingly, differing complements of pheromones and receptors and homeodomain proteins are often incorporated into the alleles of each locus, presumably through segmental duplication and pseudogenization events. Pheromone receptor genes are represented here as red arrows, pheromone genes as solid black arrows, homeodomain genes as white arrows and non-mating type genes as faded arrows.

After cell fusion, the compatibility of the two mating partners is determined intracellularly by two MAT-encoded regulatory transcription factors. In the Ascomycota, the MAT locus can encode transcription factors of three different types: homeodomain, α-box or HMG (Butler, 2004). The S. cerevisiae MAT locus for example encodes one of two distinct homeodomain proteins, HD1 (α2) in MATα and HD2 (a1) in MATa, which dimerize in diploid cells to form a transcriptional repressor of haploid-specific genes (Ho, 1994). In the Basidiomycota, the transcription factors encoded in MAT always contain the homeodomain DNA-binding motif, and the system encodes genes of the two classes found in S. cerevisiae, HD1 and HD2 (Casselton & Challen, 2006). The tetrapolar mushroom fungi Coprinopsis cinerea and Schizophyllum commune both use multiallelic HD1 and HD2 homeodomain proteins encoded at two subloci to generate multiple mating specificities, as each protein can only dimerize with alternate alleles from a different mating type (Fig. 2) (Stankis, 1992; Tymon, 1992; Banham, 1995).

Maintenance of the ensuing dikaryon is under the regulation of both the homeodomain and the pheromone/receptor loci. The homeodomain locus represses asexual sporulation, coordinates nuclear pairing and division, septation and clamp cell formation during mitosis (Raper, 1966; Tymon, 1992). The pheromone/receptor locus regulates nuclear migration and dissolution of the septa separating cells, as well as clamp cell fusion during mitosis of dikaryotic cells (Raper, 1966; Wendland, 1995; O'Shea, 1998).

Semi-obligate pathogenesis: the smut fungi

The smuts are a widespread group of dimorphic phytopathogenic basidiomycetes belonging to either the Ustilaginomycotina for example Ustilago spp., or the Pucciniomycotina for example Microbotryum spp. (Table 1), which almost exclusively infect angiosperms (flowering plants) (Fischer & Holton, 1957). The majority of smuts are heterothallic, with a saprobic yeast state that can propagate asexually by budding (Bauer, 2001). The yeast state can colonize a variety of substrata, but little is known regarding their ecology or distribution in the environment (Boekhout, 1998). Two cells of compatible mating types can fuse to develop an infectious, dikaryotic hyphal appressorium, which can proceed to invade an appropriate host plant (Vanky, 1987). During infection, the fungus usually grows asymptomatically inside the host plant in hyphal form. The smut proliferates, possibly in response to signals from the host plant, and eventually produces masses of sooty or ‘smutty’ teliospores (Banuett & Herskowitz, 1996). The teliospores form in a specialized fruiting body termed a sorus, which generally replaces either the flowers or the seeds of the host. These dikaryotic sexual spores can then disperse before germination, which entails karyogamy and meiosis (Fischer & Holton, 1957; Vanky, 1987).

The smut lifestyle, involving a switch between a free-living yeast form and an infectious, plant pathogenic hyphal form, is enormously successful (Nadal, 2008). Crucially, a morphogenetic transition and sexual reproduction are directly associated with pathogenesis. The change in morphology is a requirement for both penetration of the plant surface and for differentiation and proliferation within the host. Notably, the fungus can only sporulate in planta; without the host, sexual teliospores cannot be produced (Banuett & Herskowitz, 1996; Feldbrugge, 2004). The dimorphic switch exemplified by the smuts is a fundamental determinant of pathogenicity, and a common theme among pathogenic fungi (Bakkeren, 2008).

Ustilago maydis

In the corn smut U. maydis, fusion of compatible yeast cells is governed by the pheromone signalling system of the biallelic a locus (Fig. 2) (Kahmann & Kamper, 2004). Both a1 and a2 encode a lipoprotein pheromone and a heterotrimeric G-protein-coupled receptor (Bolker, 1992; Spellig, 1994). Binding of the secreted pheromone to its appropriate receptor triggers cell cycle arrest and the production of conjugation tubes in the direction of the pheromone (Durrenberger, 1998; Muller, 2003). Once two conjugation tubes fuse, continued growth is mediated by the b locus, which encodes the HD1 orthologue bE and the HD2 orthologue bW (Schulz, 1990; Gillissen, 1992). Two different alleles of bE and bW are required to produce the active heterodimer, but unlike the simple biallelic system of the a locus, the b locus has 35 known alleles (Kamper, 1995). The bE/bW heterodimer allows for the formation of the invasive appressorium and the initiation of infection. The a and b loci are on separate chromosomes and segregate independently at meiosis, giving rise to the archetypal tetrapolar mating type.

Ustilago hordei

Ustilago hordei causes covered smut of barley and oats. The bipolar U. hordei MAT locus is unusually large, spanning 526 kb for MAT-1 and 430 kb for MAT-2. The a and b regions are located at opposite ends of the locus, which is suppressed for recombination (Fig. 3) (Bakkeren & Kronstad, 1994; Lee, 1999). The MAT-1 locus contains 47 predicted ORFs, although many remain uncharacterized, and any possible relationship with pathogenicity is unclear (Bakkeren, 2006). The bipolar mating type found in U. hordei and most smuts is thought to promote an inbreeding lifestyle, as with only two mating types, a given meiotic progeny can mate with half of its siblings (Raper, 1966).


Bipolar MAT loci of the Basidiomycota. In contrast to the small sex-determining regions of the Ascomycota, all the confirmed or putative bipolar MAT loci of the Basidiomycota are complex genomic structures of c. 100 kb or greater. The Cryptococcus neoformans alleles are highly rearranged between mating type, and contain a cohort of mating- and virulence- associated genes, but only one homeodomain gene per locus. The Ustilago hordei MAT-1 locus harbours a number of repeated elements, functional genes and pseudogenes, with the pheromone/receptor and homeodomain loci at either end. The putative MAT locus of Malassezia globosa contains predicted homeodomain genes, an STE3-type pheromone receptor gene and a pheromone gene. Analysis of the Sporobolomyces sp. genome (v1.0) revealed the presence of divergently transcribed bE and bW orthologues on a separate scaffold from a predicted pheromone/receptor locus; the two scaffolds may represent the same chromosome if the organism is bipolar. Pheromone receptor genes are represented here as red arrows, pheromone genes as solid black arrows, homeodomain genes as white arrows and non-mating type genes as faded arrows. Uncharacterized genes from M. globosa and Sporobolomyces sp. are named using the closest Saccharomyces cerevisiae or Schizosaccharomyces pombe homologues.

Sporisorium reilianum

Sporisorium reilianum causes head smut of both corn and sorghum (Martinez, 2002). Classical genetic studies revealed that S. reilianum is tetrapolar, although molecular characterization of the MAT locus revealed that while the b locus encoding the homeodomain proteins is multiallelic (five alleles), the pheromone/receptor a locus appears to be triallelic, with three different pheromone/receptor loci, a1, a2 and a3 (Fig. 2) (Schirawski, 2005). Interestingly, each allele appears to encode not one but two pheromone genes plus one receptor, in a manner similar to the mushroom fungi. These two pheromones are the ligands for the receptors of the other two alleles, and leads to the scenario where each mating type can successfully detect a pheromone from the other two, but not itself, and can also produce a pheromone detectable by both (Schirawski, 2005).

Microbotryum violaceum

The anther smut M. violaceum from the Pucciniomycotina infects plants of the carnation family. The smut lifestyle appears to have evolved twice within the Basidiomycota, as smuts from the Ustilaginomycotina and the Pucciniomycotina display many similarities with regard to host infection, morphology and teliosporogenesis, and represent a remarkable example of convergent evolution (Bauer, 2001). The mating type locus of M. violaceum appears to extend over an entire chromosome (2.8–3.1 Mb for A1 and 3.4–4.2 Mb for A2) and harbours a single pheromone receptor plus orthologues of bE and bW (Hood, 2002; Yockteng, 2007). Populations show a mating type ratio bias, which is due to the presence of lethal alleles in the A2 mating type. As this phenotype is masked in the dikaryon, intertetrad mating is the preferred and dominant form of mating (Hood & Antonovics, 2000).

Opportunistic pathogenesis: the yeasts

The human pathogenic basidiomycetous yeasts are a phylogenetically diverse group, united by a shared growth form. Single-celled yeasts have the ideal morphology to effect pathogenesis in an animal host – for ease of entry, for transport around the body or to penetrate certain tissues – as demonstrated by the large number of yeast pathogens compared with the filamentous fungal pathogens (De Hoog, 2000). Systemic infections by basidiomycete yeasts are generally acquired in one of two ways: via inhalation of an infectious cell from the environment, or alternatively when the yeast is allowed passage into the body, such as through an indwelling catheter. In both cases, efficient transport throughout the body is achieved via the bloodstream.

While there are some yeasts such as C. neoformans and C. gattii in which sexual reproduction is thought to play a role in the pathogenic process, in many other basidiomycetous yeasts the role of sex is poorly understood, with no known teleomorph or sexual cycle. However, recent molecular analyses of fungi previously considered to be strictly asexual, such as the ascomycete A. fumigatus and the basidiomycete Malassezia globosa, have revealed predicted MAT loci and suggested cryptic sexual behaviour. Where sexual states are known, very little is known about the hyphal state in the environment (Boekhout, 1998).

The best-known and by far the most clinically significant basidiomycete pathogen of animals is C. neoformans, although the commensal yeasts of the genus Malassezia also cause a variety of common cutaneous and deep-seated infections. With the growth in the immunocompromized population worldwide, a number of additional basidiomycete yeasts have also recently emerged as increasingly important opportunistic pathogens of humans, including species of Trichosporon, Rhodotorula, Sporobolomyces and Cryptococcus (Hazen, 1995; Khawcharoenporn, 2007).

Cryptococcus neoformans

Cryptococcus neoformans is a ubiquitous human yeast pathogen of the Tremellomycetes that primarily infects immunocompromized individuals, particularly AIDS patients. Infection is primarily acquired via inhalation of infectious cells, which disseminate to the central nervous system to cause life-threatening meningoencephalitis (Casadevall & Perfect, 1998). The sexual teleomorph, Filobasidiella neoformans, and a complete sexual cycle have been described (Kwon-Chung, 1975). In C. neoformans, sexual reproduction may potentially be associated with the infection process, because sexual development produces spores. Although the idea remains highly controversial, the small and easily aerosolized sexual basidiospores are thought to be the most likely infectious propagule (Powell, 1972; Ruiz & Bulmer, 1981; Cohen, 1982; Zimmer, 1984; Sukroongreung, 1998).

The primary environmental niche of C. neoformans is pigeon guano. In the presence of pheromones and appropriate environmental cues, compatible haploid mating partners fuse and undergo a developmental transition to a dikaryotic filamentous growth form. Virtually nothing is known about the filamentous phase in the environment, which has been proposed to be free living, plant parasitic or mycoparasitic (Bandoni, 1995). However, successful mating and sporulation has been observed on both pigeon guano media and live plant surfaces (Nielsen, 2007; Xue, 2007). The dikaryon undergoes karyogamy and meiosis in a terminal basidium, before budding off four chains of haploid spores (Kwon-Chung, 1975).

Cryptococcus neoformans is heterothallic and bipolar (a and α), with a large mating type locus spanning >100 kb (Fig. 3). The MAT locus contains c. 25 genes in total and includes the pheromone/receptor and homeodomain genes (Lengeler, 2002; Fraser, 2004). Significantly, the locus contains additional genes relating to mating and the sexual cycle, including elements of the pheromone-responsive MAP kinase cascade, meiosis and sporulation genes. In contrast to the basidiomycete paradigm where each mating type encodes two homeodomain genes (one each of the HD1 and HD2 class), C. neoformans has adopted a format similar to the ascomycetes: MATα cells encode only Sxi1α (HD1) while MATa cells encode only Sxi2a (HD2) (Hull, 2005).

A peculiar feature of C. neoformans is the presence of a highly skewed population ratio: nearly all isolates are of the α mating type (Kwon-Chung & Bennett, 1978; Lengeler, 2000). Furthermore, up to 50% of the population is recalcitrant to mating, and thought to be sterile. However, a number of highly fertile a strains and a predicted recombinant population structure (much closer to 1 : 1) have been identified from isolates from sub-Saharan Africa, suggesting that mating may play an important role in certain regions (Litvintseva, 2003, 2005). Cryptococccus neoformansα isolates have been found to be more virulent in certain genetic backgrounds, and α isolates have predominated in the CNS in mixed-infection experiments with a and α isolates (Kwon-Chung, 1992; Nielsen, 2003, 2005).

Interestingly, C. neoformans can utilize an alternate pathway to traditional sexual development. Mating can occur between two cells of the same mating type, resulting in hyphal growth with unfused clamp connections, meiosis and the production of recombinant spores in a process known as monokaryotic fruiting (Wickes, 1996; Lin, 2005). This would be advantageous for a species with such a skewed population, and enables basidiospore production and dispersal even when mating partners are scarce (Lin, 2005; Nielsen, 2007).

Cryptococcus gattii

Cryptococcus gattii is a sibling species of C. neoformans in the Tremellomycetes, generally restricted to tropical and subtropical climates, that predominantly infects immunocompetent hosts (Kwon-Chung, 1976; Fraser, 2003). As in C. neoformans, the proposed mode of infection is via inhalation of spores, and a complete sexual cycle and sexual teleomorph Filobasidiella bacillispora have also been identified (Kwon-Chung, 1976; Campbell, 2005a). The a and αMAT loci show a structure similar to the C. neoformans locus: both are c. 100 kb in size and contain the same cohort of genes, albeit highly rearranged (Fraser, 2004).

A recent outbreak of this fungus has occurred on Vancouver Island, Canada (Hoang, 2004). Interestingly, it appears that an unusual, more virulent population of C. gattii has managed to expand beyond its normal environmental niche into temperate climates. Despite the fact that most isolates worldwide are sterile, almost all samples from Vancouver are highly fertile, and are almost entirely of the α mating type. The population is also nearly clonal, although recent studies suggest that this more virulent strain may have arisen recently due to a recombination event, possibly between two parents of the same mating type (Fraser, 2003, 2005). Alternatively, the outbreak strain may have originated in South America, where this rare genotype is more common and may be recombining in mixed a–α populations (Kidd, 2004; Trilles, 2008).

Unusual population structures and variable mating competency have also been observed in other populations of C. gattii. Population genetic studies from Australia have revealed broadly clonal populations, but also identified isolated recombining, highly fertile populations (Campbell, 2005b; Saul, 2008). A skewed population structure was also evident: sampled areas included populations with c. 1 : 1 mating type ratios, a hallmark of recombining populations, but an α mating type bias in others (Halliday, 1999). Remarkably, surveys of C. gattii from Colombia have revealed populations predominantly of a isolates (c. 97%) of both clinical and environmental samples (Escandon, 2006). Although the fertility and genetic variability of these specimens were not ascertained, a population almost entirely of the a mating type would be highly unusual, particularly in comparison with the fertile and highly virulent clade of α mating type in Vancouver. Intriguingly, diploid or aneuploid hybrids between C. gattii and C. neoformans have recently been described from clinical specimens that may potentially display unique virulence characteristics (Bovers, 2006, 2008).

Cryptococcus laurentii and Cryptococcus albidus

A variety of Cryptococcus species in the Tremellomycetes have been implicated in a range of both localized and systemic infections in humans, primarily C. laurentii and C. albidus. Both C. laurentii and C. albidus are generally saprobic and can be widely isolated from environmental sources, as well as among the skin flora (Krajden, 1991). Recent phylogenetic analyses of these species have revealed them to be polyphyletic, which unfortunately obscures causality between phenotype and pathogenesis (Fell, 2000; Fonseca, 2000; Sugita, 2000; Takashima, 2003).

Infections with these species generally involve immunocompromized individuals, although invasive devices are a predisposing factor for C. laurentii infection. Both C. laurentii and C. albidus have been reported in cases of cutaneous infections, as well as fungaemia, meningitis and pulmonary infections (Kordossis, 1998; Averbuch, 2002; Burnik, 2007). These species complexes share two classical virulence factors with their more pathogenic distant relatives; both possess a polysaccharide capsule and synthesize the antioxidant melanin (Ikeda, 2000, 2002). However, the ability to grow at 37 °C is rare in both species, perhaps explaining why C. laurentii and C. albidus are less successful pathogens than C. neoformans and C. gattii, which can routinely grow at 37 °C (Casadevall & Perfect, 1998).

Cryptococcus laurentii and C. albidus are considered to be anamorphic. While a complete sexual cycle has not been elucidated, compatible cell fusion, formation of mycelia and possible teliospores have been reported for C. laurentii (Kurtzman, 1973). Both complexes harbour teleomorphs in species closely related to these pathogens, which may imply the presence of cryptic sexuality (Fonseca, 2000; Takashima, 2003). As cells of both species are commonly found in the air, the suggested route of infection is via the respiratory system in a manner similar to C. neoformans (Krajden, 1991; Burnik, 2007).


The genus Malassezia consists of a small group of common human commensal yeasts classified with the primarily phytopathogenic Ustilaginomycotina. Besides C. neoformans, Malassezia spp. are the most clinically prevalent basidiomycete human pathogens, responsible for a range of cutaneous infections including pityriasis versicolor and seborrhoeic dermatitis, and also causing a number of systemic mycoses in immunosuppressed patients, including catheter-related fungaemia, peritonitis and meningitis (Ashbee & Evans, 2002). Recent taxonomic revision of the genus has led to the recognition of a number of new species, with the most commonly isolated pathogenic species including Malassezia furfur, M. globosa, Malassezia sympodialis, Malassezia pachydermatis, Malassezia restricta, Malassezia slooffiae and Malassezia obtusa (Gueho, 1996). All species in the genus, except for M. pachydermatis, are fatty acid auxotrophs and are therefore obligate commensals. As a consequence, these yeasts require an external lipid source in culture, which often has a confounded diagnosis (Shifrine & Marr, 1963; Xu, 2007).

Malassezia species are dimorphic with cutaneous infections, often presenting a characteristic ‘spaghetti and meatballs’ form, where both yeast and hyphal phases of the fungus are found. Malassezia species possess virulence factors similar to those of C. neoformans, including melanin production and a capsule-like lipid layer that is thought to have an immunomodulatory effect on the host immune system during asymptomatic carriage (Ashbee, 2006; Thomas, 2008). The entire genus is anamorphic, with no identified sexual cycle. However, the taxonomic position of the genus Malassezia within Ustilaginomycotina– a group of anamorphic animal pathogens among a clade of nearly exclusively sexual plant pathogens – is peculiar; it is possible that Malassezia species are phytopathogenic in an as-yet undetermined sexual dikaryophase (Begerow, 2006).

The recent completion of the genome of the dandruff-causing M. globosa has revealed the presence of a putative MAT locus (Fig. 3) (Xu, 2007). The locus contains genes encoding both pheromones and pheromone receptors (Ste3 type) plus homeodomain-containing bE and bW homologues. The c. 173-kb locus also contains homologues of two genes found in the C. neoformans locus (STE12 and CID1) and appears to be quite gene-rich (82 predicted ORFs), but repeated element-poor in comparison with other large MAT loci. Identification of another MAT allele from M. globosa could potentially enable the elucidation of the process of mating and sexual reproduction in the species, as the mating type genes do not appear to be degraded as may be expected from a strictly asexual species.


The genus Trichosporon contains a number of yeast species responsible for both superficial and systemic mycoses in humans, characterized by the presence of true hyphae and arthroconidia (Middelhoven, 2004). Following recent taxonomic reclassification of Trichosporon beigelii, six species of Trichosporon are implicated in human infection: Trichosporon asahii, Trichosporon mucoides, Trichosporon cutaneum, Trichosporon asteroides, Trichosporon inkin and Trichosporon ovoides (Gueho, 1992). Infections range from superficial mycoses (particularly white piedra) to invasive and systemic mycoses, including fungaemia, pneumonia and meningitis. The major cause of disseminated trichosporonosis is T. asahii, which has emerged as an increasingly important pathogen primarily of immunocompromized patients; note that since reclassification, most previous cases of T. beigelii-invasive trichosporonosis have been ascribed to T. asahii (Anaissie & Bodey, 1991; Walsh, 2004).

Trichosporon asahii can be found as both an environmental saprophyte and as part of the skin flora, and infection is associated with haematological malignancies, AIDS or corticosteroid usage. Although infection is commonly associated with the presence of invasive devices such as catheters, T. asahii is also thought to enter via the respiratory and gastrointestinal tracts (Walsh, 1992, 2004; Tashiro, 1995). Trichosporon asahii accounts for nearly 10% of disseminated fungal infections, and is associated with a generally poor prognosis and high mortality rate (approaching 80%) with a variable response to antifungal treatment (Krcmery, 1999; Di Bonaventura, 2006).

Infection with T. asahii presents with physical and histopathological symptoms similar to disseminated candidiasis, which can often result in misdiagnosis. Significantly, the fungus is able to change morphology among the yeast, hyphal and pseudohyphal forms during infection in a manner similar to C. albicans, although the presence of arthroconidia in T. asahii infections is characteristic (Tashiro, 1994). Trichosporon asahii also expresses an antigen similar to cryptococcal glucuronoxylomannan, which has an immunosuppressive effect on the host (Lyman, 1995). No sexual teleomorph has been identified in the genus, however; only asexual reproduction has been found to date (Middelhoven, 2004).


The Rhodotorula species are a polyphyletic group of yeasts distributed in both Pucciniomycotina and Ustilaginomycotina, which can be isolated widely from the environment and normal human biota. Three pucciniomycete species have been identified as causing infections in humans: Rhodotorula mucilaginosa (formerly Rhodotorula rubra), Rhodotorula glutinis and Rhodotorula minuta. Rhodotorula mucilaginosa is most commonly isolated from patients (c. 70%) (Tuon & Costa, 2008). Similar to other emerging pathogenic yeasts, all three species are primarily pathogens of the immunocompromized, and are frequently associated with indwelling catheters. Infections commonly associated with R. mucilaginosa include endocarditis, peritonitis and meningitis, but fungaemia is most common (Thakur, 2007; Tuon, 2007). Rhodotorula mucilaginosa and R. glutinis can be recalcitrant to treatment, and are known for being highly resistant to fluconazole; mortality rates for R. mucilaginosa infection have been reported to be as high as 15% (Diekema, 2005; Neofytos, 2007).

Rhodotorula mucilaginosa and R. glutinis are both anamorphic yeasts of the Microbotryomycetes, while R. minuta is an anamorphic yeast of the Cystobasidiomycetes. Recent taxonomic re-evaluation of R. glutinis, however, revealed significant genetic heterogeneity, and a number of isolates previously ascribed to R. glutinis have been reclassified as five separate sexual teleomorphs in the genus Rhodosporidium, rendering any correlations between lifestyle and infection in previous clinical cases extremely difficult (Gadanho, 2001; Sampaio, 2001; Gadanho & Sampaio, 2002). All three anamorphic Rhodotorula species are closely related to a number of teleomorphic species, however, and future study may yet reveal the presence of cryptic sexuality (Biswas, 2001; Gadanho & Sampaio, 2002).

Recently, a portion of the MAT locus of the closely related bipolar microbotryomycete yeast R. toruloides was sequenced, revealing the presence of multiple pheromone precursor genes, a p21-activated kinase (PAK) gene similar to the pheromone-responsive MAP kinase component STE20 located in the MAT locus of C. neoformans and an STE3-type pheromone receptor gene (Coelho, 2008). No putative homeodomain genes have been located thus far in close proximity to the pheromone locus.


The genus Sporobolomyces is a polyphyletic assemblage of yeasts within the Pucciniomycotina that produce forcibly discharged asexual ballistoconidia. Sporobolomyces is found as both a widespread environmental saprophyte and as a common human commensal, and three species have been reported to be pathogenic in humans: Sporobolomyces roseus, Sporobolomyces holsaticus and Sporobolomyces salmonicolor. The latter species has primarily been implicated in human infection, including reports of dermatitis, fungaemia, lymphadenitis and endophthalmitis, in association with both immunocompetent and AIDS patients (Plazas, 1994; Sharma, 2006). Sporobolomyces salmonicolor has shown resistance to both fluconazole and amphotericin B in vitro, although standard antifungal therapies have often resolved infections (Serena, 2004).

Sporobolomyces salmonicolor has an identified teleomorph, Sporidiobolus salmonicolor (Van der Walt, 1970). The very closely related species Sporidiobolus johnsonii, considered a synonym of S. holsaticus, is primarily homothallic and may be in the process of speciation from Sporobolomyces salmonicolor due to limited sexual compatibility between the two (Valerio, 2008b). Sporobolomyces roseus, which has long been considered asexual, has a recently described sexual teleomorph, Sporidiobolus metaroseus, although only self-fertile isolates have been found so far (Valerio, 2008a). It is unclear whether the asexual or the sexual state is the infectious form.

The genome of S. roseus was recently made available (http://www.genome.jgi.doe.gov); however, the strain sequenced has since been revealed as an undescribed Sporobolomyces species of unknown sexuality and mating type (Valerio, 2008a). Nevertheless, the genome sequence for this closely related species reveals potential unlinked homeodomain and pheromone/receptor loci (Fig. 3). The pheromone/receptor locus contains three copies of a pheromone precursor, plus an STE20 homologue and an STE3-type receptor, and appears syntenic to the partial pheromone locus elucidated in R. toruloides (Coelho, 2008). The homeodomain locus contains both HD1 and HD2 homologues, which appear intact and are divergently transcribed (unpublished data). The two regions could potentially form a large, contiguous locus (≥1700 kb) spanning the majority of the chromosome, in a similar manner to the large MAT loci of both U. hordei and M. violaceum.

Conclusion and outlook: sex, morphology and infection

Comparison between mechanisms of pathogenesis of basidiomycete yeasts on both human and plant hosts reveals vastly different approaches to a pathogenic lifestyle. Central to both modes of infection, however, is a change in the morphology of the fungus between yeast and hyphal growth forms, and potentially, the requirement for sexual reproduction. Links between sexual development and pathogenesis have been suggested previously for a number of dimorphic fungi, particularly in signalling pathways (Madhani & Fink, 1998; Nadal, 2008).

Animal pathogenic yeasts tend to infect in the yeast form, which enables ease of entry into the host and movement via the bloodstream. The respiratory system may be a common portal for many opportunistic basidiomycete yeast pathogens of humans, particularly Cryptococcus species. However, the filamentous sexual form may potentially be required to produce the infectious particle. In the plant pathogens, while the asexual yeast form can survive epiphytically on the host, it must locate an appropriate partner to initiate infection. Pathogenesis is intimately linked to a dimorphic transition, as the fungus changes from a yeast to a hyphal growth form and invades the host plant. The reverse is extremely uncommon within the basidiomycetes – infection of humans with filamentous basidiomycete fungi is limited to sporadic accounts of infection with S. commune and C. cinerea (Rihs, 1996; Lagrou, 2005); similarly, yeast infections of plants are rare (Fell, 2001).

There are few pathogenic fungi that can infect both plant and animal hosts, due to the different mechanisms and morphology required to effect pathogenesis. Within the basidiomycetes, U. maydis has been sporadically implicated in human infections; however, the yeast state appears to be the infectious form (Preininger, 1937; Moore, 1946; Patel, 1995). Yeast anamorphs of the normally filamentous, plant pathogenic Ustilaginomycetes have been shown to infect humans, including Pseudozyma and Malassezia species (Sugita, 2003). Conversely, under laboratory conditions, both C. neoformans and C. gattii have recently been shown to infect Arabidopsis thaliana, causing dwarfing and chlorosis in the host plant. Correspondingly, it was the dikaryotic filamentous form produced via mating of a and α yeast isolates on seedlings that was pathogenic (Xue, 2007). However, many basidiomycete yeast species such as C. neoformans and C. gattii may potentially be mycoparasitic in their hyphal form (Bandoni, 1995).

Further ties between the process of sexual reproduction and pathogenesis are found in the mating type locus itself: in C. neoformans and C. gattii, a number of gene products located in MAT have been implicated in virulence, including the transcription factor gene STE12, the pheromone receptor gene STE3 and the PAK kinase gene STE20 (Chang, 2000;; Wang, 2002). Interestingly, STE12 is also located in the putative MAT locus of M. globosa and c. 80 kb downstream of STE3 in Sporobolomyces sp.; STE20 in Sporobolomyces sp. also lies directly between two pheromone genes. MAT locus genes of C. neoformans including the pheromone genes are also induced during infection, and the two mating types display different virulence properties (Nielsen, 2003, 2005; Heitman, 2006). A number of ongoing basidiomycete genome projects (including the T. mesenterica, C. laurentii, U. hordei and S. reilianum genomes) may reveal further insights into MAT evolution and virulence.

Although a number of the basidiomycetous yeast pathogens are currently considered to be exclusively asexual, many of these ‘asexual’ fungi retain the machinery to undergo sex, but perhaps utilize it only rarely (Nielsen & Heitman, 2007). Remarkably, molecular analysis and genome sequencing are consistently confirming that these asexual species are either engaging in sexual recombination, or possess intact MAT loci. Examples from the Ascomycota include C. albicans, Coccidioides immitis and Coccidioides posadasii, Penicillium marneffei and A. fumigatus (Burt, 1996; Koufopanou, 1997; Hull & Johnson, 1999; Paoletti, 2005; Woo, 2006). Similarly, in the Basidiomycota, the potential MAT locus in the genome of M. globosa suggests the presence of cryptic sex in Malassezia, and a sexual teleomorph has recently been described for S. roseus (Xu, 2007;; Valerio, 2008a).

Drawing correlations between the pathogenesis and the lifestyle of many of the basidiomycete yeasts is challenging; all the genera reported to infect humans have recently undergone extensive phylogenetic reclassification, and many of the original clinical reports now refer to obsolete species, or species that have undergone revision. While dimorphism and sexual reproduction is clearly linked to the pathogenesis of the smut fungi, the exact nature of the infectious particle in the yeasts is less clear; while the sexually produced basidiospore has been implicated in Cryptococcus species, this remains a subject of debate. Additionally, information on the hyphal phases of most yeasts is fragmentary for even the best-studied systems. The emerging molecular data, combined with a greater understanding of the ecology of dimorphic basidiomycetes, should provide a greater insight into the potential role of sexual reproduction in these important emerging pathogens.


  • Editor: Teun Boekhout


View Abstract