Sex and the Imperfect Fungi

ABSTRACT

Approximately 20% of species in the fungal kingdom are only known to reproduce by asexual means despite the many supposed advantages of sexual reproduction. However, in recent years, sexual cycles have been induced in a series of emblematic “asexual” species. We describe how these discoveries were made, building on observations of evidence for sexual potential or “cryptic sexuality” from population genetic analyses; the presence, distribution, and functionality of mating-type genes; genome analyses revealing the presence of genes linked to sexuality; the functionality of sex-related genes; and formation of sex-related developmental structures. We then describe specific studies that led to the discovery of mating and sex in certain Candida, Aspergillus, Penicillium, and Trichoderma species and discuss the implications of sex including the beneficial exploitation of the sexual cycle. We next consider whether there might be any truly asexual fungal species. We suggest that, although rare, imperfect fungi may genuinely be present in nature and that certain human activities, combined with the genetic flexibility that is a hallmark of the fungal kingdom, might favor the evolution of asexuality under certain conditions. Finally, we argue that fungal species should not be thought of as simply asexual or sexual, but rather as being composed of isolates on a continuum of sexual fertility.

INTRODUCTION

Sexual reproduction is a ubiquitous feature of the eukaryotic kingdom with the many benefits of sex in generating genetic diversity as substrates for evolutionary selection being well known. When two different partners come together, there is the generation of genetic variation in the offspring, through the processes of crossover and recombination during meiosis, enabling response of future generations to environmental selection pressures (1–4). Sexual reproduction also allows the repair of random epigenetic or conventional genetic damage by recombination with homologous chromosomes and can mask lethal mutations (4, 5). In addition, sexual recombination alleviates clonal interference and prevents deleterious mutations hitchhiking to fixation (6). Indeed, there are so many benefits to sexual reproduction that exceptions that are purely asexual have been termed “evolutionary scandals” (7). As a result, supposed ancient asexual species such as the bdelloid rotifers (an exclusively female class of over 460 rotifer species thought to date back several million years) and darwinulid ostracods (a family of around 30 crustacean species thought to have been exclusively female and asexual for over 200 million years, but for which very rare living males have recently been described) have gained notoriety (8–10). It therefore comes as a great surprise that, until recently, approximately 20% of all fungal species were considered to reproduce only by asexual means, with no recognized sexual cycle, based on knowledge of described fungal species (11, 12). Indeed, in some phylogenetic groupings such as the Ascomycotina, up to 40% of taxa surveyed were deemed to be asexual (13). This is despite the fact that sexual reproduction in fungi can have additional benefits such as the production of fruit bodies and sexual spores that are resistant to adverse environmental conditions, thereby promoting survival of sexual offspring; it can provide a transient diploid arena for selection of genes; and sex can favorably impact genome evolution (14–17). Asexual species are also proposed to be short-lived evolutionary “dead ends” subject to rapid extinction (12, 18). Fungal species lacking a known sexual cycle have been referred to as “imperfect” or “mitosporic” fungi and have been grouped into the “Fungi Imperfecti” or “Deuteromycota,” although phylogenetic analysis has shown that these are artificial groupings not based on taxonomic relationship (19). The terms “asexual” and “imperfect” are used synonymously in this review.

Therefore, the imperfect fungi appear to present an evolutionary paradox. However, in the past few decades, there has been accumulating evidence that many of these supposedly asexual species might actually have the potential for sexual reproduction and thus exhibit what has been termed “cryptic” (20, 21), “clandestine” (22), or “covert” (23) sexuality, i.e., having the presence of a sexual cycle that has not yet been directly observed. There are various lines of evidence to support this assertion such as analysis of population genetic markers and the recent, fast-growing number of sequencing data sets from asexual yeasts and filamentous fungi indicating that they harbor mating-type loci, and thus have the potential for a sexual life cycle (see further details below). Indeed, the suggestion of cryptic sexuality has been fulfilled or “consummated” recently for several previously considered asexual fungal species when mating and sexual cycles were conclusively demonstrated for the first time. Prominent examples include the discovery of mating in the opportunistic yeast pathogen Candida albicans (24, 25), and both mating and complete sexual cycles in the opportunistic filamentous fungal pathogens Aspergillus flavus, Aspergillus fumigatus, and Aspergillus parasiticus (26–28), as well as the biotechnologically relevant species Penicillium chrysogenum, Penicillium roqueforti, and Trichoderma reesei (29–32; S. S. Swilaiman, H. Darbyshir, J. Houbraken, R. Samson, and P. S. Dyer, unpublished results).

This article will review how these recent advances were made. It will also speculate on the evolution of asexuality in fungi, including consideration of whether there are any truly imperfect fungi, what characteristics they might have, and the factors that could favor their appearance. Special attention is given to the ascomycete group of fungi from which most discoveries have been made.

EVIDENCE FOR SEXUAL POTENTIAL

Various complementary approaches or “sex tests” (33) have been used to provide evidence of the potential for sexual reproduction in asexual species as will now be described.

Population Genetics

One of the first indirect lines of evidence for sexual recombination in supposed asexual species came from population genetic studies. It would be expected that the genetic diversity, both in terms of genotype and phenotype, is higher in sexually propagating populations than in exclusively asexually propagating clonal species. Various molecular and physiological markers have been used in such population genetic analyses. For example, sequencing with arbitrary primer pairs (SWAPP) provides a molecular method to find and characterize genetic markers with nucleotide-level resolution. In the human pathogen Coccidioides immitis this technique provided molecular evidence for recombination in a fungus for which no sexual stage has so far been described (34). Alternatively, the distribution of alleles of unlinked loci or microsatellite markers has been analyzed in field populations. It would be expected that if a species were freely sexually recombining, then this would result in most unlinked loci being associated randomly in populations, whereas if there were a predominantly clonal and asexual reproduction, then nonrandom association of alleles (“linkage disequilibrium”) would be observed. Thus, evidence of recombination through random association can be taken as an indication of a sexual life cycle. For example, a total of 236 isolates from Diplodia pinea, a haploid opportunistic plant pathogen, were characterized using 13 microsatellite markers. Analysis of these markers confirmed previous results that D. pinea has a high level of gene and genotypic diversity (35). In parallel, sequencing of seven molecular markers from 31 isolates of the animal-associated asexual Penicillium dipodomyis provided evidence of genetic diversity and recombination suggesting cryptic promiscuity (36). In previous work, evidence of recombination had also been observed in populations of A. flavus, A. parasiticus, and A. fumigatus based on DNA fingerprinting and the large number of vegetative compatibility groups present within populations (reviewed in reference 37), and in P. chrysogenum based on microsatellite analysis (38)—all species that were, at the time, considered purely asexual. However, it is cautioned that there are possibly other explanations for genetic diversity in populations such as parasexual reproduction (see Daskalov et al. [176]).

Mating-Type Genes

Mating-type (MAT) genes have provided fundamental insights into the reproductive status of asexual fungi through studies both of their distribution and also functional activity. By way of background, two main types of fungal sexual breeding systems can be distinguished: (i) heterothallism, involving two individuals of opposing mating type, and (ii) homothallism, which refers to sexual reproduction by selfing of an individual strain. Heterothallic sex maintains genetic variation due to mixing of genotypes, whereas homothallic sex can lead to clonality, similar to that observed in asexual reproduction (12). In heterothallic species, sexual recombination occurs between morphologically identical partners containing opposite MAT loci, which, in the case of ascomycetes, consist of dissimilar sequences occupying the same locus on the chromosome. In contrast, homothallic species normally contain genes for both MAT loci, which can either be located at a single MAT locus or on different chromosomes (21, 39). In most fungi, the mating-type locus represents a relatively small region of the genome, less than a few thousand base pairs, expressing transcription factors that determine both the haploid cell type specificity and/or the diploid zygote fate. Only very few fungi possess larger dimorphic sex chromosomes analogous to those of animals (reviewed by Heitman [40]).

For the most part, our general understanding of mating-related processes in ascomycetes is based on knowledge obtained from studies with the hemiascomycete baker’s yeast Saccharomyces cerevisiae and the filamentous euascomycete (Pezizomycotina) Neurospora crassa (39). In S. cerevisiae, two alternative mating-type (MAT) loci, namely MATα and MATa, determine the sex of individual strains (41). These loci are termed idiomorphs to indicate that they do not represent alleles of a single gene (42). In addition, each yeast strain carries silent mating-type loci of both MATα and MATa, thus allowing infrequent switching of the mating-type identity. In N. crassa, MAT loci are designated mata and matA, whereas in most other euascomycetes the terms MAT1-1 and MAT1-2 are used (21, 43–45). MAT loci from euascomycetes harbor one or more open reading frames (ORFs), of which at least one codes for a MAT transcription factor (TF) (21, 46). In contrast to yeasts, euascomycetes normally have only one active mating-type locus, with no extra silent mating-type loci. MAT loci have been characterized from a broad range of homo- and heterothallic euascomycetes (21, 39, 43, 47). Mating-type loci in sordariomycetes seem to be more complex with more than one open reading frame per mating-type locus. Heterothallic members of the Eurotiales are usually less complex and normally contain only a single functional open reading frame as depicted for various Aspergillus and Penicillium species in Fig. 1. This latter group can be considered to provide ideal experimental systems for gaining mechanistic and functional insights into MAT TFs and their various cellular functions as will be focused on below (48–51).

FIGURE 1.

Comparison of MAT loci from homo- and heterothallic members of the Eurotiales. Blue arrows indicate a MATα-domain gene, red arrows indicate a MAT high-mobility group gene, black bars indicate intronic sequences, gray bars homologous sequences (48, 49, 51, 173). For A. nidulans, the gene designation is as previously published by Paoletti et al. (49). Note that, whereas isolates of heterothallic species contain only one MAT idiomorph (either MAT1-1 or MAT1-2), isolates of homothallic species contain both types of MAT gene in the same genome (i.e., both MAT1-1 and MAT1-2).

As a rule, the MAT1-1 locus encodes an α-domain TF, and the alternative idiomorph, MAT1-2, is characterized by a gene coding for a TF with a high-mobility group (HMG) domain. The corresponding genes are generally referred to as MAT1-1-1 and MAT1-2-1 (45, 47). While DNA-binding HMG proteins are ubiquitous and found in a wide range of eukaryotes, α1-domain proteins seem to be limited to the fungal kingdom. The prototype of an α-domain protein is Matα1p from S. cerevisiae, whose function has been studied by genetic dissection (52). From this study the α domain was shown to act as a degradation signal for a pathway defined by the SUMO-targeted ligase Slx5–Slx8, which suggested a coordinated regulation in the turnover of regulatory transcription factors to ensure rapid mating-type switch in yeast. Extensive sequence comparisons, phylogenetic analyses, and in silico predictions of secondary and tertiary structures of α and HMG transcription factors supported the hypothesis that the α1 domain is related to the HMG domain (53). This study concluded that fungal α1-box genes originated from an ancestral HMG gene.

Distribution of mating-type genes

First, mating-type genes have provided insights into the nature of asexual fungi through a population genetics approach investigating the distribution of MAT genes in natural populations of supposedly asexual fungi. It might be predicted that truly asexual species would be composed of near-clonal populations consisting of isolates of only one mating type as a result of speciation from a founder of a single mating type. Alternatively, genetic drift could lead to loss of a complementary mating type because there would be no selection pressure to maintain both mating types within a species. And rarely, some fungal species have been found in which isolates of only one mating type have been found (reviewed by Dyer et al. [21]). For example, all isolates of the plant pathogens Rhynchosporium lolii and Rhynchosporium orthosporum from a diverse range of geographical locations and plant hosts throughout Europe were exclusively of the MAT1-1 genotype (54), whereas all individuals of the lichen-forming fungus Thamnolia vermicularis sampled from the Northern, and parts of the Southern, Hemisphere were entirely of the MAT1-2 genotype (55). However, almost always, when studies of MAT distribution have been made in asexual fungal species, it has been found that isolates of both mating types can be found (i.e., the presence of both MAT1-1 and MAT1-2 idiomorphs), and often these are present in a near 1:1 distribution (reviewed in reference 21). This is consistent with a heterothallic sexual breeding system and has been used as strong evidence for the potential for sexual reproduction in the respective asexual species. Such studies have been made with a wide taxonomic diversity of fungi including Aspergillus, Cochliobolus, Fusarium, Penicillium, Rhynchosporium, and Trichoderma species as listed in Table 1. As an example, both MAT1-1 and MAT1-2 isolates of P. chrysogenum have been found worldwide in near equal distribution (Fig. 2). It is noted that the term heterothallic should only strictly be used for outcrossing species where a sexual state has been observed, and the term “proto-heterothallic” has been suggested to describe such asexual species where genetic evidence (e.g., the presence of complementary MAT loci) indicates the presence of a sexual cycle (56).

TABLE 1.

Evidence for mating-type loci, their distribution, functional characterization, and induction of a sexual cycle in representative euascomycete species that have been presumed to be asexual^a

Organism	Identification of MAT loci	1:1 distribution of MAT locib	Functional MAT locusc	Induction of a sexual cycled	Reference(s)
Acremonium chrysogenum	MAT1-1	No	Yes	No	77
Aspergillus flavus	MAT1-1, MAT1-2	Yes		Yes	26, 80, 158
Aspergillus fumigatus	MAT1-1, MAT1-2	Yes		Yes	28, 48
Aspergillus niger	MAT1-1			No	70
Aspergillus nomius	MAT1-1, MAT1-2			Yes	104
Aspergillus oryzae	MAT1-1, MAT1-2	Yes			57
Aspergillus parasiticus	MAT1-1, MAT1-2	Yes		Yes	27, 80, 159
Aspergillus sclerotiocarbonarius	MAT1-1, MAT1-2	Yes		Yes	108
Aspergillus terreus	MAT1-1, MAT1-2			Yes	106
Aspergillus tubingensis	MAT1-1, MAT1-2			Yes	107
Coccidiodes immitis, C. posadasii	MAT1-1, MAT1-2	Yes	Yes		160, 161
Cochliobolus victoriae	MAT1-2			No?	162
Diplodia pinea (Sphaeropsis sapinea)	MAT1-1, MAT1-2	Yes			163
Fusarium avenaceum, F. culmorum, F. poae, F. senitectum	MAT1-1, MAT1-2		Yes		164
Fusarium tucumaniae	MAT1-1, MAT1-2			Yes	165
Fusarium azukicola, F. brasiliense, F. phaseoli	MAT1-1, MAT1-2	Yes		No	150
Fusarium virguliforme	MAT1-1	No			150
Penicillium biforme, P. camemberti	MAT1-2				117
P. nalgiovense
Penicillium chrysogenum	MAT1-1, MAT1-2	Yes	Yes	Yes	30, 51, 61
Penicillium digitatum	MAT1-1	No			166
Penicillium expansum	MAT1-1, MAT1-2	Yes			166
Penicillium fuscoglaucum, P. paneum,	MAT1-1				117
Penicillium pinophilum	MAT1-1 MAT1-2			Yes?	167
Penicillium roqueforti	MAT1-1 MAT1-2	Yes		Yes	31, 32
Phialocephalia fortinii – Acephala applanata	MAT1-1 MAT1-2	Yes			168
Talaromyces (Penicillium) marneffei	MAT1-1, MAT1-2			No	169
Rhynchosporium agropyri, R. secalis	MAT1-1, MAT1-2	Yes	Yes		54
Rhynchosporium lolii, R. orthosporum	MAT1-1	No	No		54
Septoria passerinii	MAT1-1, MAT1-2	Yes		Yes	170
Talaromyces amestolkiae	MAT1-1, MAT1-2	Yes		Yes	171
Trichoderma reesei (syn. Hypocrea jecorina)	MAT1-1, MAT1-2			Yes	29
Ulocladium spp.	MAT1-1, MAT1-2		Yes		172

^a This list is not exhaustive but aims to provide an overview of the taxonomic diversity of species that have been examined.

^b Where studies have been undertaken of the distribution of MAT1-1 and MAT1-2 idiomorphs in sample populations.

^c As assessed by gene expression and/or functional characterization by gene manipulation.

^d A question mark indicates some signs of sexual development, but no viable sexual spores formed.

FIGURE 2.

Occurrence of both MAT idiomorphs in wild-type isolates from Penicillium chrysogenum. Blue and red dots represent strains with the MAT1-1 or MAT1-2 locus, respectively (C. M. O’Gorman and U. Kück, unpublished data).

Function and regulatory impact of mating-type locus encoded transcription factors

Second, mating-type genes have provided insights into the nature of asexual fungi through an investigation of their expression and functional activity. The majority of studies have been limited to examining whether or not MAT genes are expressed at the mRNA level as an indication of activity. In almost all cases, MAT gene expression has indeed been demonstrated in imperfect fungi where assayed, which has been interpreted as providing evidence of sexual potential because the MAT genes would appear to be conferring sexual identity in the respective species under examination (see examples listed in Table 1). There have been some rare exceptions. For instance, it was not possible to detect MAT1-1-1 gene expression in R. lolii and R. orthosporum, although MAT gene expression was detected in other supposed asexual Rhynchosporium species (54), but such cases are highly unusual, so far.

However, more in-depth research over recent years has revealed a rather more complex scenario, with expression analysis, providing evidence for a regulatory impact of MAT-encoded TFs beyond specifically sexual reproduction. Therefore, it has become clear that MAT gene expression alone does not necessarily indicate sexual potential, and caution should be exercised in interpreting such data, because MAT genes are likely to have other important functions in asexual fungi beyond mating and sexual development. This was demonstrated by pioneering MAT-locus-dependent expression analysis in the asexual Aspergillus oryzae, which showed that MAT1-1 and MAT1-2 specifically regulated over a thousand genes, including many of unknown function, with 33 genes differentially regulated over 10-fold between the different mating types (57). This was interpreted, in part, as evidence for MAT gene functional activity and therefore sexual potential, but this study also indicated that MAT loci are involved in the regulation of genes beyond those strictly involved in a putative sexual cycle. Similarly, gene manipulation of the MAT1-1-1 gene from the then considered asexual P. chrysogenum revealed that the MAT1-1 α-domain TF impacted developmental processes of biotechnological relevance such as penicillin biosynthesis, light-dependent asexual sporulation, and hyphal morphology, in addition to any role in sex (30). This work was extended, when the MAT1-1-1 gene from P. chrysogenum was analyzed by ChIP-seq analysis. This investigation led to the identification of 254 putative direct MAT1-1-1 target genes and a highly specific MAT1-1-1 DNA-binding sequence (58). Interestingly, some of the most highly bound regions in ChIP experiments occurred near homologs of known functional targets of the S. cerevisiae MATα1 protein, such as Pcppg1, a homolog of MFα1, encoding the α-pheromone, and Pcpre1, a homolog of STE3, coding for the a-factor receptor (59, 60). In addition, this analysis confirmed a large number of new MAT1-1-1 target genes, which had never been related to any MAT TF before and were assigned to functional categories beyond mating, such as asexual development, morphogenesis, amino acid metabolism, secondary metabolism, and iron metabolism (58). A summary of MAT1-1-1 target genes is depicted in Fig. 3. In parallel work, it was later shown that the MAT1-2 HMG domain TF from P. chrysogenum also impacted a variety of developmental processes, including light-dependent asexual sporulation, conidiospore germination, and surface properties, again in addition to any role in sex (61). A summary of MAT-controlled processes in P. chrysogenum is shown in Fig. 4. Comparable results were obtained with the homothallic cereal pathogen Fusarium graminearum which contains both MAT1-1 and MAT1-2 loci in the same genome. Unlike Sordaria macrospora, another well-studied homothallic euascomycete (62), F. graminearum requires all the transcripts from both MAT loci for sexual development. The regulatory mechanisms controlled by MAT genes in F. graminearum were explored by several strategies, including genome-wide transcriptional profiling in various genetic backgrounds, and high-throughput gene deletions (63). Genome-wide microarray with total RNAs from F. graminearum mutants that lacked each MAT locus individually or together revealed 1,245 differentially expressed genes, including those involved in metabolism, cell wall organization, cellular response to stimuli, cell adhesion, fertilization, development, chromatin silencing, and signal transduction.

FIGURE 3.

Target genes of the MAT1-1-1 locus encoded transcription factors from Penicillium chrysogenum, deduced from functional genomics experiments (58, 61). In particular, ChIP-seq analysis has shown that MAT1-1-1 regulates gene expression far beyond their described function as regulator of sexual development (modified from reference 174).

FIGURE 4.

Summary of the regulatory functions of MAT locus encoded transcription factors MAT1-1-1 and MAT1-2-1 in Penicillium chrysogenum (modified according to reference 175).

Another example of MAT genes regulating nonsexual processes comes from Fusarium verticillioides, where deletion of the MAT1-2-1 gene led to a drastic reduction in carotenoid production in parallel with a severe decrease in photo-induced expression of genes encoding key enzymes of the carotenoid biosynthesis pathway (64). Moreover, F. graminearum MAT1-1-1 and MAT1-2-1 deletion mutants were shown to be reduced in virulence, although none of the MAT locus genes was important for plant infection, indicating that the MAT proteins may play a host-specific role in colonization of corn stalks (65). Another interesting demonstration of MAT TFs acting outside the sexual life cycle was found for the human pathogenic basidiomycete Cryptococcus neoformans, where the heterodimers of MAT TFs Sxi2a and Sxi1α were shown to be involved in regulation of several well-studied virulence genes (66). Based on these observations, it can be proposed that MAT TF target genes in hemiascomycetes are restricted to genes relevant for mating, whereas MAT-mediated regulation beyond sexual development might be a common feature in euascomycetes. It is conceivable that comprehensive rewiring of gene regulatory networks controlled by MAT TFs has occurred since the division of these two subclasses.

Genome Studies

Whole-genome sequencing projects have been completed for several asexual fungi over the past decade. This has allowed the genomes to be screened for the presence of genes known to be required for sexual reproduction as a way to evaluate whether the species are truly asexual. The rationale for this work is that, if a species is genuinely asexual, then it would be predicted that genes required for sexual reproduction would either accumulate deleterious mutations, or even be lost, in asexual species due to random mutation and lack of selective pressure to maintain the functional forms of the gene. Furthermore, those genes involved with the earliest stages of mating might be the most obvious candidates for mutation and/or loss to avoid unnecessary downstream metabolic investment in sex. This is based on the understanding that asexual species have evolved from sexual ancestors (e.g., reference 67), sex being a basal characteristic of the fungal kingdom (68).

Perhaps surprisingly, almost all genome projects to date have revealed that supposedly asexual fungi do have apparently intact sets of genes known to be required for sexual reproduction, with no evidence of gene mutation or loss. In the first study for the Pezizomycotina, the genomes of the then considered asexual A. fumigatus and A. oryzae were screened for the presence of over 200 genes linked to sex (from mating through to meiosis), and it was found that all genes were present. Furthermore, a heterothallic mating system was indicated based on the presence of a characteristic MAT idiomorph (69). Soon afterward, genome sequencing of the asexual Aspergillus niger, used widely in the biotechnological industries, again revealed the presence of a suite of genes linked to mating and sexual development (70). An apparent mutation in the pro1 gene was found in the latter study, but this was later discounted because of an incorrect annotation of an intron. Concurrent genome screening of the then asexual A. flavus also revealed a full complement of apparently functional genes required for sexual reproduction (C. E. Eagle and P. S. Dyer, unpublished results). Furthermore, BLAST analysis of the genome sequence of the basidiomycete yeast Malassezia globosa, a causal organism of dandruff, revealed the presence of a MAT locus containing pheromone, pheromone receptor, and sex-linked homeodomain genes, together with other pheromone signaling pathway genes elsewhere in the genome of this presumed asexual species (71). Most recently, genome analysis of asexual Rhynchosporium species has again failed to detect any loss or mutation in genes related to sexual reproduction (54; International Rhynchosporium Genome Project, unpublished results). A further revelation from genome analysis of P. chrysogenum was the discovery of mutations apparently caused by a RIP-like (repeat-induced point mutation) process, again linked to sexual reproduction (72).

However, some caution must again be exercised in interpreting such evidence. Genome analysis of eight Candida genomes revealed that certain species with known sexual states were missing some orthodox components of mating and meiotic pathways. For example, the homothallic Lelongisporus elongisporus lacked any known MAT gene, having a syntenic MAT locus but which only contained noncoding DNA, and also lacked a pheromone precursor and receptor (73). Similarly, evidence of a degenerated MAT1-1-1 pseudogene has been reported for the heterothallic Cordyceps takaomontana (74). Therefore, the lack of sex-related genes is not conclusive proof of asexuality, whereas the presence of such genes is arguably circumstantial evidence for sexual potential.

Functionality of Sex-Related Genes

Further evidence that imperfect fungi have the potential for sexual reproduction has come from studies where sex-related genes from certain asexual species have been expressed heterologously in model fungal species known to be sexual, and have been shown to be functionally active. This provides evidence of a conserved sexual function in the respective asexual host species. Genes assayed in this way include MAT genes and members of pheromone signaling pathways as will now be described.

Cross-species mating-type gene exchange provides a powerful method to demonstrate the functionality of mating-type locus-encoded proteins. Different approaches are feasible. For example, mating-type loci of the α domain or HMG type can be introduced into isolates of a heterothallic species carrying the opposite mating-type locus. In this case, the triggering of sexual development leading to formation of fruiting bodies (although not necessarily sexual spore formation) provides good evidence of MAT gene functionality. This approach was successfully demonstrated in pioneering studies of known sexual species when MAT loci from the homothallic fungus S. macrospora were shown to induce sexual development (although not ascospore formation) in the heterothallic fungus Podospora anserina (75). An alternative approach is to use deletion mutants, lacking a mating-type locus. Complementation of the deletion mutant may be performed with a DNA fragment carrying the MAT locus from a heterologous host. For example, MAT1-2-1 from A. fumigatus, encoding a HMG box mating transcriptional factor, was introduced into a mating-type deletion mutant of the homothallic fungus A. nidulans (76). Using a MAT1-2-1 gene under the expressional control of the Aspergillus nidulans MAT2 (matA) promoter, it was demonstrated that the MAT1-2-1 gene from A. fumigatus, at that point only known to be asexual, was able to drive sexual reproduction in A. nidulans. In parallel studies, it was shown that the MAT1-1-1 gene from A. fumigatus was able to complement an A. nidulans MAT1 (matB) deletion mutant and restore sexual fertility (50). Comparable results were reported for Acremonium chrysogenum, the fungal producer of the pharmaceutically relevant β-lactam antibiotic cephalosporin C. This filamentous fungus is classified as asexual, because no direct observation of mating or meiosis has yet been reported. The investigated strains were shown to carry a MAT1-1 mating-type locus with a MAT1-1-1 α-domain gene and a MAT1-1-2 gene, encoding an HPG domain protein, defined by the presence of the three invariant amino acids histidine, proline, and glycine. In addition, a MAT1-1-3 gene was detected, encoding a high-mobility group (HMG) domain protein (77). So far, only a single wild-type isolate has been characterized, and no reports are available that strains with opposite mating-type loci exist. To assess the potential of A. chrysogenum for sexual reproduction, the entire A. chrysogenum MAT locus was transferred into a MAT1-1 deletion strain of the heterothallic ascomycete P. anserina. After fertilization with a P. anserina MAT1-2 (MAT2) strain, the corresponding transformants developed fruiting bodies with mature ascospores. These results demonstrated that the MAT genes of A. chrysogenum were functional and supported the hypothesis of an extant sexual cycle in this species.

Other evidence for a sexual cycle stems from functional studies of pheromone and pheromone receptor genes, using the heterologous S. cerevisiae system as a genetic tool. Successful pheromone binding and signaling would be expected to result in cell cycle arrest and change in cell morphology and lead to formation of a halo in lawns of S. cerevisiae (78). Based on similarity to the S. cerevisiae MFα proteins, the P. chrysogenum Pcppg1 gene was predicted to produce a decapeptide pheromone of sequence KWCGHIGQGC, expected to bind to the cognate PcPRE2 receptor protein. And indeed, S. cerevisiae wild-type cells (ScSTE2p) or yeasts heterologously expressing PcPRE2 exhibited polarized growth, leading to pear-shaped forms (shmoos, originally named after a cartoon character of this shape drawn by Al Capp) of unconjugated haploid cells, in response to either the native S. cerevisiae α factors or the synthetic decapeptide pheromone PcPPG1, respectively (30). This result was complemented by bioassays in which adding synthetic PcPPG1 pheromone to lawns of yeast cell, expressing PcPRE2, resulted in halo formation. Thus, PcPRE2 and PcPPG1 represent a functional pheromone-receptor pair likely involved in the mating of P. chrysogenum, which was later observed by Böhm et al. (30). These data were consistent with microarray analyses demonstrating the expression of the genes for a pheromone precursor (Pcppg1) and two pheromone receptors (Pcpre1, Pcpre2) in P. chrysogenum (30).

As well as the examples above, a STE12 homolog from the asexual Penicillium marneffei (renamed Talaromyces marneffei) was able to complement a steA gene deletion mutation of A. nidulans and restore sexual fertility (79). Meanwhile, overexpression of the A. fumigatus nsdD gene led to increased development of sexual structures (cleistothecia) in A. nidulans (50). Finally, before the identification of any sexual stage, pheromone precursor and receptor genes were shown to be expressed in A. fumigatus, A. flavus, A. parasiticus, A. oryzae, and P. chrysogenum in what was termed “sexual posturing” (22, 48, 51, 57, 80). Most recently, similar expression of pheromone signaling genes has been demonstrated in several other asexual aspergilli including Aspergillus aculeatus, A. brasiliensis, A. clavatus, A. niger, A. sydowii, A. wentii, and A. zonatus (178). However, as above, some caution must be used when interpreting such results because it appears that some aspects of the cellular sexual machinery can be subverted for other purposes in fungi. For example, the STE2 pheromone receptor of the asexual Fusarium oxysporum has been shown to play a role in detection of plant exudates (81). Therefore, detecting expression of sex-related genes is not necessarily a sign of cryptic sexuality.

Formation of Sex-Related Structures and Lichen Species Pairs

One final line of evidence that imperfect fungi have the potential for sexual reproduction comes from the observation that certain asexual species have been observed to form developmental structures representing the initial stages of sexual reproduction. This suggests that they might have the potential for sexual fertility if correct conditions can be identified to induce the complete sexual cycle (56). For example, several asexual Aspergillus and Penicillium species are known to form sclerotia, a necessary prerequisite for sexual reproduction in closely related species (82, 83). This phenomenon was recently observed for certain isolates of A. niger which were able to form sclerotia when they were incubated on certain organic substrates including raisins (Fig. 5) (84) (G. Ashton and P. S. Dyer, unpublished results).

FIGURE 5.

Sclerotia formation (arrowed large gray-brown spheres) in Aspergillus niger, an indication of the potential for sex in this biotechnologically important species? Scale bar indicates 500 μm. Note that this species is predominantly of the MAT1-1-1 genotype (H. Darbyshir, G. Ashton, and P. S. Dyer, unpublished data).

A particularly intriguing phenomenon related to this is the occurrence of so-called “species pairs” in lichen biology. This refers to the situation where two morphologically and phylogenetically very closely related pairs of lichen species have been described which differ solely in the fact that one species bears sexual structures on the thallus, whereas the sister species lacks any sexual structures (85–87). It is tempting to speculate that the reason for asexuality in the sterile partner species might be principally because of the lack of a suitable mating partner to induce sex, rather than these being genuinely separate species as per received orthodoxy. In related examples from lichen-forming fungi, Seymour et al. (86) reported finding the first fertile example of Usnea sphacelata (as evidenced by formation of sexual fruiting bodies [ascomata] on the thalli), and Greenaway (88) described the first fertile thalli of Normandina pulchella; both species had previously been considered to reproduce only by asexual means.

DISCOVERY OF MATING AND SEXUAL REPRODUCTION—AND IMPLICATIONS

The above lines of evidence have provided, and these continue to give, an indication of the potential for sexual reproduction in the imperfect fungi, although such evidence has at best been circumstantial because the elusive sexual states were still missing. However, a series of major breakthroughs have been reported since 2000 for yeasts and 2009 for filamentous fungi, where modern molecular approaches combined with classical microbiology laboratory culture have been used to demonstrate the existence of either mating and/or a complete functional sexual reproductive cycle in a series of supposed asexual species from different genera. In most cases, a key first step was the use of molecular diagnostic tools to identify potential sexual mating partners and then the setting up of directed crosses between parental isolates of complementary mating type under a range of incubation conditions (56). Some notable examples are described below.

Mating in C. albicans

There had been strong indications that the opportunistic human pathogen C. albicans might have sexual potential following the identification of a mating-type locus similar to that seen in the model sexual yeast S. cerevisiae, together with the detection of other genes that had high homology to sexual pathway genes in S. cerevisiae (89) (see also Bennett and Turgeon [177]). These indications were confirmed when two different groups were able to induce mating between a or α strains of C. albicans to produce a tetraploid state. Hull et al. (24) made artificial a or α strains by deletion of the alternative MTL α or a locus, respectively, and crossed these under selective conditions in mice, while Magee and Magee (25) induced chromosome loss to produce a or α strains by deletion of one of the MTL loci and then crossed auxotrophic strains on agar media using selective conditions. Although both groups therefore reported the first success in mating strains of C. albicans, the rates of mating were rather low and led to a tetraploid state without any meiotic division. A second major advance then followed, being the discovery that homozygous a/a or α/α C. albicans cells must undergo a phenotypic transition from a “white” traditional yeast-like morphology to an “opaque” morphology to achieve mating competency (90–92). However, even under these conditions, meiosis has still not yet been observed in C. albicans. Instead, tetraploids appear to undergo stochastic concerted chromosome loss to return to a diploid or near-diploid state, particularly under stressful conditions (93, 94). Furthermore, even when mating-competent haploid strains of C. albicans were produced under in vitro and in vivo conditions, it was still not possible to induce meiosis (95). Therefore, true sex involving meiotic nuclear recombination has still not yet been observed in C. albicans; instead, this is best described as mating and a parasexual process. It is also noted that the intriguing phenomenon of same-sex mating has also been observed in C. albicans, but again with no evidence of meiosis (96).

Sex in Aspergillus Species

The genus Aspergillus comprises over 340 species (97) and includes a series of species of great economic, medical, and scientific importance. Members of the genus share the common feature of an “aspergillum,” a branching head conidiogenous structure for the production of asexual spores. In addition, about one-third of species (36% according to reference 33) are known to reproduce sexually. All the latter species produce fruiting bodies known as cleistothecia, but a variety of cleistothecial forms are evident, meaning that under traditional taxonomic classification, different teleomorphic genera have been assigned according to morphological features such as whether cleistothecia are enclosed, the color of cleistothecia, and whether they are surrounded by accessory cells (33, 37, 97). The fact that the remaining majority of Aspergillus species (64% according to reference 33) are only known as imperfect fungi has attracted research interest both from an applied and fundamental perspective, because many species of biotechnological and pathological interest were only known as asexual organisms. If it were possible to induce sexual cycles in such species, then this would provide a valuable tool for strain improvement and classical genetic analysis, as well as providing insights into the population biology and evolutionary potential of species. Research into this area has been reviewed (see reference 37), with a summary and updates provided below.

Initial investigations involved the use of population genetic approaches, which revealed evidence of high levels of genetic diversity and recombination in nature for the spoilage and aflatoxin producing A. flavus and A. parasiticus (98, 99), and the opportunistic pathogen A. fumigatus (28, 48, 100, 101). Mating-type genes were then identified for the first time from the aspergilli (102, 103), which allowed PCR diagnostics to be developed. This led to the identification of alternative MAT1-1 and MAT1-2 idiomorphs and the discovery of a near 1:1 distribution of compatible mating types both within global and regional populations of A. fumigatus (28, 48). Similar results were later reported for both A. flavus and A. parasiticus for a local field population from the United States (80). All these observations provided evidence of cryptic sexuality in these species.

A major breakthrough then came when laboratory work led to a sexual stage being induced for the first time in A. fumigatus. This was achieved by crossing known MAT1-1 and MAT1-2 isolates from a population in Ireland (which had shown evidence of recombination) on oatmeal agar in a barrage formation under specific environmental conditions. Cleistothecia characteristic of the teleomorph genus Neosartorya were produced after 6 to 12 months, and the sexual stage named Neosartorya fumigata under existing taxonomic rules. Recombination was demonstrated in the ascospore offspring using molecular markers (28). Soon afterward, similar crossing efforts were rewarded with A. flavus and A. parasiticus, when incubation of strains of opposite mating type for between 6 and 12 months led to the production of hardened sclerotia found to contain sexual ascospores, although recombination was not assayed. The two sexual stages were assigned to the teleomorph genus Petromyces (26, 27).

The studies with A. fumigatus, A. flavus, and A. parasiticus provided a precedent for ensuing research in which attempts were made to cross MAT1-1 and MAT1-2 isolates of other asexual Aspergillus species under conditions conducive to sex in the aspergilli. A sexual cycle was subsequently induced in A. nomius (104), A. lentulus (105), A. terreus (106; A. E. Ahmed and P. S. Dyer, unpublished results), A. tubingensis (107), A. sclerotiicarbonarius (108), and A. clavatus (S. Swilaiman and P. S. Dyer, unpublished results). The discovery of sexual reproduction in A. tubingensis and A. sclerotiicarbonarius is particularly notable, because these are members of the black aspergilli and therefore related to the economically important species A. niger, which is much used in the biotechnology sectors. However, all efforts to induce sexual reproduction have so far failed in this species, although the ability to induce sclerotial production in certain strains, together with identification of a TF linked to this process, is a promising sign (Fig. 5) (84; T. R. Jørgensen, J. Frisvad, K. F. Nielsen, P. S. Dyer, and A. F. J. Ram, unpublished results).

Sex in Penicillium Species

The genus Penicillium comprises over 360 species that have a ubiquitous distribution (83). Morphologically, they are characterized by the production of an asexual reproductive structure, called the penicillius (83, 109). Together with species of the genus Aspergillus, penicillia belong to the order Eurotiales (97) and are among the most prevalent fungi on earth. Taxonomically, Penicillium species are the asexual (anamorphic) forms of sexual (teleomorphic) species from the genera Eupenicillium and Talaromyces (83, 97). However, the majority of Penicillium species (73% according to reference 33) have no known sexual state.

Because of their widespread beneficial usage in food production and biotechnology, penicillia are of outstanding economic and medical importance. For example, P. camemberti and P. roqueforti are used in the making of white mold (e.g., Brie and Camembert) and blue mold-ripened cheeses (e.g., Roquefort and Stilton) (110). In the pharmaceutical industry, P. brevicompactum is used for large-scale production of the immunosuppressant mycophenolic acid, P. griseofulvum for the industrial production of the broad-spectrum antibiotic griseofulvin, and P. citrinum for production of the cholesterol-lowering drug compactin (111–113). Last but not least, P. chrysogenum, perhaps the most renowned representative of the genus Penicillium, constitutes the only industrial producer of the β-lactam antibiotic penicillin. Today, penicillin is the most commonly used drug in the treatment of bacterial infections and one of the most valued products in the global anti-infective market. As with the aspergilli, the economic importance of certain asexual Penicillium species has meant that they have attracted research interest regarding the possibility of inducing sexual reproduction for strain improvement and classical genetic applications as will now be described for P. chrysogenum and P. roqueforti.

Initial investigations of the sexual potential of P. chrysogenum (reclassified recently as P. rubens by Houbraken et al. [97, 114]) were made using DNA and genome-sequencing approaches (51, 115). Interestingly, the strain originally discovered by Sir Alexander Fleming (116) was found to contain a MAT1-2 locus, while later, industrially used strains were found to carry the MAT1-1 locus. The latter observation can be explained by the fact that they all are derivatives of a wild-type isolate, obtained in the late 1940s in Peoria, Illinois. Hoff and coworkers showed that these strains retained transcriptionally expressed pheromone and pheromone receptor genes required for sexual reproduction (51). A further survey showed that P. chrysogenum isolates from diverse global locations contain either the MAT1-1 or the MAT1-2 locus with an almost equal distribution (Fig. 1).

In order to induce the sexual life cycle in P. chrysogenum, 17 MAT1-1 wild-type or mutant strains were crossed in various combinations with 10 MAT1-2 wild-type isolates using a range of different growth media and conditions. Finally, a MAT1-1 strain derived from a strain development program in the 1950s was found to generate cleistothecia in crosses with a MAT1-2 wild-type strain (30). The culture conditions were similar to those described for A. fumigatus (28), but the oatmeal agar had to be supplemented with biotin, which accelerated the formation of mature cleistothecia, which appeared after 5 weeks. Measurement of the penicillin titer on plate test and examination of restriction fragment length polymorphism of 11 genes, which differed between the two parental strains, proved that genetic recombination has occurred in all ascospore isolates investigated. In a subsequent analysis, the chromosome constitution of selected ascospore isolates was determined by whole genome sequencing (61). Included in these studies were the functional analyses of deletion strains, lacking either the MAT1-1 or the MAT1-2 locus. From these studies, it became evident that both mating-type loci have functions beyond sexual development (see discussion above).

Meanwhile the sexual potential of P. roqueforti has also been investigated. P. roqueforti is widespread in the natural environment and is a common spoilage agent in stored foods and meat products. In food production, this ascomycete is used as a starter culture for blue-veined cheese production. A survey of several world isolates from P. roqueforti revealed the presence of both mating types, and the detection of repeat-induced point mutation (RIP) within repeated sequences and transposable elements indicated strongly that sexual recombination had occurred or was ongoing in P. roqueforti (31, 117). This indirect evidence for a sexual cycle was recently confirmed when a sexual cycle was induced under laboratory conditions (31, 117; Swilaiman et al., unpublished). In one study, four MAT-1-1 and MAT1-2 strains were tested in eight possible mating combinations. Although many combinations were not fully fertile (only empty fruiting bodies were produced), typical sexual structures such as cleistothecia, asci, and ascospores were detected in some crosses (32). In a parallel study, sexual reproduction was induced in a higher number of crosses, including when crossing isolates from different types of blue cheese (31; Swilaiman et al., unpublished). Interestingly, the conditions developed for P. chrysogenum (30) to induce sex were identical, indicating that these conditions are valid for a wide range of Penicillium species. Recombination of ascospore progeny from P. roqueforti was demonstrated with either microsatellite or DNA fingerprint markers (31, 32; Swilaiman et al., unpublished). The sexual cycle of P. roqueforti offers the exciting potential to select for progeny with altered flavor and mycotoxin characteristics for novel blue cheese strains (H. Darbyshir, J. Frisvad, and P. S. Dyer, unpublished results).

Sex in Trichoderma and Other Genera

The ascomycete T. reesei is used in the biotechnology industry for applications such as the decomposition of diverse substrates through to the production of second-generation biofuels from cellulosic waste. Moreover, efficient biocontrol strains have been developed as biological fungicides (118). T. reesei, the anamorph of the ascomycete Hypocrea jecorina, was believed for more than 50 years to propagate strictly asexually. However, Seibel and coworkers succeeded in inducing sexual development in the industrial production strain QM6a (29). This strain contains the MAT1-2 mating-type locus, but was found to be female sterile, i.e., unable to produce the specialized hyphal tissues needed for fruit body development (in this case, perithecia). Thus, in mating, this strain was only able to act as a male fertilizing partner, but was then able to undergo a heterothallic reproductive cycle with wild-type isolates carrying the MAT1-1 locus. Perithecia embedded in stroma were generated under different light conditions. Interestingly, visible light is dispensable for stroma formation, and constant light even inhibited stroma formation. Further studies indicated that the light-regulatory protein Envoy 1, a homolog of the light-oxygen-voltage (LOV) domain-containing protein vivid from N. crassa, is essential for female fertility (119). In follow-up studies, further parental strains were crossed with T. reesei. The analysis of unusual 16-spore asci, which were produced when parental strains showed chromosomal heterogeneities, revealed the formation of a considerable number of inviable ascospores. This indicated that chromosomal homogeneity is crucial for highest sexual fertility (120). These results correspond well with the sequencing data of ascospore progeny of P. chrysogenum that were derived after crossing of parental strains that exhibited heterogeneous chromosomal configurations (121).

There are further reports of sex being induced in other genera such as Fusarium (122) and Trichophyton (123) species by utilizing MAT diagnostic tools and particular crossing conditions, but a full discussion is beyond the scope of the present review.

Implications of the Discovery of Fungal Sex

The presence of a sexual cycle offers many potential benefits. First, the discovery of a sexual cycle in biotechnologically relevant species provides an exciting new tool in addition to conventional breeding approaches to genetically engineer industrial production strains in the laboratory. Thus, fungal mating systems may be appropriate for the improvement of industrial production strains similar to established animal and plant-breeding programs (124). This strategy was recently implemented for T. reesei, when a female sterile strain was engineered to female fertility (125). By introduction of the ham5 gene, it was possible to restore female fertility of the QM6a T. reesei wild-type strain, which is the progenitor strain of all industrially used derivatives. The generation of this strain, which may serve as male as well as female partner in mating experiments, provides the basis for further breeding of this cellulose-producing fungus. As an indication of the potential for this work, a crossing experiment with different wild-type mating partners of T. reesei showed that ascospore isolates were obtained that produced a higher level of xylanases than either parental strain, despite chromosomal heterogeneity (120). In comparable work, low-penicillin-producing strains could be crossed with P. chrysogenum wild-type isolates to give rise to ascospore isolates with an increased penicillin titer (61). By crossing different isolates of P. chrysogenum it was also possible to select for ascospore progeny with minimal production of a secondary metabolite chrysogenin, which can interfere with penicillin purification (30). However, it might prove to be a general problem in future mating experiments to obtain mating partners with compatible karyotypes. Previous crossing experiments with isolates from different locations of the pseudohomothallic ascomycete Neurospora tetrasperma resulted frequently in sexual dysfunction, with abnormalities ranging from reduced ascospore viability to complete sterility (126). Pulsed-field gel electrophoresis analysis and mating of different geographic strains from the homothallic fungus S. macrospora showed a correlation between karyotype heterogeneity and reduced sexual fertility (127). Similarly, one investigation of mating with P. roqueforti strains from noncheese and cheese habitats belonging to different genetic clusters found that highest fertility was only achieved when strains from the same genetic cluster were crossed. Fertility was significantly reduced when strains isolated from cheese were crossed with those from noncheese habitats, possibly as a result of domestication and clonal selection (128). However, parallel studies found comparable levels of fertility when crossing different isolates of P. roqueforti used in blue cheese production, so this might be a strain-specific effect (31; Swilaiman et al., unpublished). In future strain improvement programs, the combination of conventional mating experiments and contemporary molecular tools, such as ChIP-Seq or forward genetic analysis, will generate novel approaches to engineer genetically industrial producer strains. One prerequisite for the success of these breeding experiments might therefore be to select or design mating partners with compatible karyotypes to ensure breeding success.

Second, the discovery of a sexual cycle provides a very valuable tool for classical genetic analysis. By crossing of partners showing different traits and analysis of the sexual progeny, it is possible to determine whether a trait is mono- or polygenic in basis. The sexual cycle can also be used for gene mapping, verification of gene function, and identification of genes of interest through techniques such as bulk segregant analysis (124).

Finally, the discovery of a sexual cycle provides important insights into the evolutionary potential of species. Species undergoing sexual recombination are likely to generate and maintain considerable genetic diversity within populations. This is predicted to lead to more rapid breakdown of control measures for fungal pathogens (129). Therefore, having knowledge about the reproductive mode of a species can inform disease management strategies. This might, for example, involve removal of suitable substrates to reduce the risk of occurrence of the sexual cycle—as can be achieved by removal of straw stubble after harvest to prevent the sexual cycle of plant pathogenic Tapesia species (130), and melon debris infected with Monosporascus cannoballus after crop termination (131).

POTENTIAL FOR SEX IN OTHER ASEXUAL SPECIES—CONSIDERATIONS

As will be apparent from the examples given above, in almost all cases where asexual fungal species have been studied in depth, evidence can be obtained of cryptic sexuality, and, in many instances, a fully functional sexual cycle can be induced. This raises the question of whether there are any truly imperfect fungi. We offer some considerations about this topic below.

First, we contend that there are some genuinely long-lived asexual fungal species, which is made possible by the genetic flexibility that is an especial characteristic of the fungal kingdom (132). Fungi exist in nature in a variety of ploidy states, from haploid to diploid to tetraploid, they can exhibit accessory chromosomes, and some species can even manipulate the genome to duplicate localized chromosome regions in the face of environmental pressures (132, 133). In tandem with this, fungi have developed alternative methods to heterothallic sex to allow gene flow both within and between species. Fungi are able to undergo vegetative fusion with hyphae of partners of the same heterokaryon compatibility group, thereby allowing nuclear exchange, and this can even lead to nuclear fusion and gene recombination via the parasexual cycle (see Daskalov et al. [176]). In recent years there have been reports of same-sex mating in human pathogenic fungi (92, 96). Also, there is accumulating evidence of horizontal gene transfer in the fungal kingdom both from fungi and also prokaryotic organisms (134). In early reports, transfer of low-molecular extrachromosomal DNA was reported between Neurospora species, between the distantly related ascomycetes Ascobolus immersus and P. anserina, and from the zygomycete mycoparasite Parasitella parasitica to its host Absidia glauca (135–137). Therefore, there is arguably less pressure on fungi to be sexually “perfect” to provide new sources of genetic variation than in many other eukaryotic organisms where there is an obligate requirement for sex.

What characteristics might such long-lived imperfect fungal species have? We would suggest factors such as the lack of any observed sexual cycle or developmental structures despite sustained investigative studies; a high degree of clonality in populations; loss of genes normally associated with sexual development or their subversion to other cellular functions; dominance of one mating type (if MAT loci could even be detected); and development of alternative methods for generating or sustaining genetic variation as an adaptation to asexuality. The predicted long-term genomic effects of the loss of sex have even been classified and include possible congruent divergence of nuclear loci (the “Meselson effect” in asexual diploids), decrease in mutation rate, and loss of transposable element activity (see Table 1 in Normak et al. [4]). As a prime example in the fungal kingdom, we propose the group of fungi involved with vesicular-arbuscular mycorrhizal (VAM) symbioses from the phylum Glomeromycota (138). No morphological evidence of sexual structures has ever been observed in this group, which, combined with clonal population structures and conservation of asexual spore morphology, has led to the suggestion that they have been asexual for over 400 million years, even though some core meiotic genes have been detected (7, 139–141). Instead, it appears that individuals can arise from heterokaryotic spores composed of a mosaic of nuclei, which can maintain and promote genetic variation that helps explain occasional signs of recombination seen in the VAM fungi (142, 143). Such nuclear variation might provide a useful way to buffer these apparently asexual organisms against the effects of accumulating mutations (144) and maintain genetic diversity in the absence of sex. Most recently, a putative mating-type locus has been identified from genome sequencing of the model species Rhizophagus irregularis, suggesting a sexual origin for the observed heterokaryosis (145).

Second, we suggest that certain human activities have resulted in an increased likelihood of the evolution of asexual fungal species. Despite the many benefits of sexual reproduction, there are also costs to sexual reproduction leading to fungal “sexual hang-ups” (15). In outcrossing sex for every good, enhanced individual in the sexual offspring, there will be a corresponding poorly adapted one with less favorable sets of genes compared with the parents. Indeed, there is some evidence from comparing offspring of outcrossing and selfing in A. nidulans that outcrossing can result in progeny with an overall mean reduction in fitness, albeit that some offspring have significantly increased fitness (146). Thus, sex involving outcrossing has the risk of breaking up favorable sets of genes, a factor that can provide strong selection pressure for either the evolution of asexuality and/or homothallism in organisms well adapted to their environment (147). Human activities can exacerbate this phenomenon by establishing large, relatively uniform environments in which well-adapted organisms might evolve away from sex to avoid the risk of dilution of favorable genomes. There are also increased metabolic costs to sexual reproduction compared with asexual reproduction as a result of the investment needed for mating and development of fruiting bodies. In addition, the number of sexual spores produced is normally far less than can be produced by an equivalent investment in asexual reproduction, and development of such sexual spores needs a longer time period because of the complexity of sexual morphogenesis, all further reasons for the evolution of asexuality. Thus, there can be benefits to an asexual lifestyle. As Ross (148) wryly commented, the thought that the absence of sex is an imperfection, and that sex is perfection, might be an assumption of the human observer, rather than the fungus!

An example of such human activity perhaps favoring the evolution of asexuality is the deployment over wide geographic regions of crop plants with a narrow genetic base that has provided the opportunity for the clonal spread of fungal plant pathogens, which if well adapted to their host have an incentive to lose sex for all the factors listed above. And, indeed, there is evidence of what has been termed a “slow decline” in sexual fertility in populations of some fungal phytopathogens (149). F. oxysporum has evolved to be pathogenic on a wide variety of plant hosts and exhibits clonal spread on some crops such as banana. Despite concerted efforts, it has not been possible to induce a sexual cycle in this species complex, and pheromone-signaling sex-related genes now appear to be involved with pathogenicity processes detecting a plant host (81). A similar evolution away from sex is apparent for several other Fusarium species (122, 150). Similarly, different Rhynchosporium species (causing leaf blotch disease) are thought to have evolved on various graminaceous hosts as a result of human agricultural activities linked to crop selection, cultivation, and then dispersal over recent millennia (151, 152). These host-specialized species are only known to reproduce asexually, and sampling of R. lolii and R. orthosporum has so far only detected isolates of a single mating type (54). Finally, the rice pathogen Magnaoporthe oryzae is considered to be a heterothallic sexual species. However, there are reports of extensive loss of sexual potential in field populations worldwide, both due to loss of female fertility (i.e., the ability to produce maternal fruit body tissues, while rudimentary male ability to donate nuclei is retained) or total loss of any mating ability (153, 154). The slow decline in sexual ability might be attributable to various factors, as previously discussed (149), and the loss of sexual fertility seen in the field mirrors that seen under laboratory conditions when isolates are subcultured long-term by asexual means (149, 155). The long-term future of such asexual species is unclear; it may be speculated that they are still likely to be relatively short-lived evolutionary dead ends (18) because they are dependent on human activity.

Third, we suggest that although some asexual species do exist and others are currently evolving, overall such species are still likely to be very rare. Therefore there is likely to be the occurrence of a sexual cycle in the majority of 20% of all fungi that are currently only known in the imperfect state. We suggest that there are various reasons for the failure to detect sex in such species. Many might need very specific conditions to induce a sexual cycle, requiring much painstaking work to identify (56). A further obstacle is that almost all are likely to exhibit heterothallic breeding systems, requiring the identification of a possibly rare compatible mating type as is the case for C. neoformans and Cryptococcus gattii (92), whereas it will be much easier to induce sex in homothallic species where no mating partner is needed. For some species, many isolates might have low sexual fertility as a result of a slow decline in sexual potential in the species overall (149). Therefore, it might take dedicated sampling to find rare isolates of a species that are a remnant that maintain sexual potential. But such isolates would be important in evolutionary terms to generate variation episodically within populations. Indeed, given the argument that “a little sex goes a long way,” (7) it may only be necessary to retain a low proportion of isolates that retain the ability to reproduce sexually, i.e., that a very low rate of sexual recombination is sufficient to overcome evolutionary problems associated with asexuality.

CONCLUSIONS

With the advent of the “one species, one name” ruling in fungal biology (156), whether a species is perfect or imperfect has less taxonomic relevance than the past, because only one Latin binomial is needed for a species regardless of sexuality. But a knowledge of the reproductive mode remains a critical feature for an understanding of the biology of a species. There has been a sexual revolution in the understanding of asexuality over recent years, with a number of studies revealing sexual cycles in species that were previously considered imperfect. This implies that the majority of asexual fungal species will be shown to exhibit cryptic sexuality with the potential for sexual reproduction. However, there is a danger of simply thinking of fungi as either “perfect” sexual or “imperfect” asexual species. Indeed, we finish by echoing the comments of Dyer and Paoletti (149) that species should not be viewed as either sexual OR asexual, but, rather, they should be thought of as consisting of isolates on a continuum of sexual fertility, with the potential presence of isolates within a species with a range of abilities to undergo sexual reproduction from high to low fertility to even apparent asexuality. A thought to ponder over a cup of coffee, appropriate given that caffeine can induce sex in some fungal species (157).