Major insights into sexual development and cryptic sexuality within filamentous fungi have been gained from investigations using Aspergillus species. Here, an overview is first given into sexual morphogenesis in the aspergilli, describing the different types of sexual structures formed and how their production is influenced by a variety of environmental and nutritional factors. It is argued that the formation of cleistothecia and accessory tissues, such as Hülle cells and sclerotia, should be viewed as two independent but co-ordinated developmental pathways. Next, a comprehensive survey of over 75 genes associated with sexual reproduction in the aspergilli is presented, including genes relating to mating and the development of cleistothecia, sclerotia and ascospores. Most of these genes have been identified from studies involving the homothallic Aspergillus nidulans, but an increasing number of studies have now in addition characterized ‘sex-related’ genes from the heterothallic species Aspergillus fumigatus and Aspergillus flavus. A schematic developmental genetic network is proposed showing the inter-relatedness between these genes. Finally, the discovery of sexual reproduction in certain Aspergillus species that were formerly considered to be strictly asexual is reviewed, and the importance of these findings for cryptic sexuality in the aspergilli as a whole is discussed.
Sexual reproduction in the filamentous fungal genus Aspergillus is reviewed, with detailed descriptions of a series of over 75 genes known to be associated with sexual morphogenesis and a new schematic developmental network presented, together with discussion of the recent discoveries of cryptic sexuality in certain aspergilli.
‘Sexuality in fungi has long been recognized as one of the more perplexing yet intriguing facets of the biology of this large and varied group of micro-organisms’ (Raper, ). Despite a further 40 years of research since John Raper wrote this statement in his book Genetics of Sexuality in Higher Fungi, the process of sexual reproduction in fungi is still perplexing mycologists. However, insights have been gained into many aspects of fungal sexuality following the application of modern molecular genetic techniques. For example, genes involved in the determination of breeding systems have been identified together with a repertoire of genes involved in sexual development, and such ‘sex-related’ genes have been used to investigate the genetic basis of sexuality and asexuality in fungi. Many of these advances have come through or have been aided by the study of Aspergillus species, which will form the focus of the present review.
The genus Aspergillus comprises approximately 250 species (Samson & Varga, ) that are collectively termed the ‘aspergilli’. It should be noted that this number is probably an underestimation, and conflicting reports exist on the actual number of accepted species owing to the complex taxonomy of the genus (Peterson, ). Aspergillus species have traditionally been recognized by the presence of a common morphological feature, the ‘aspergillum’, an asexual reproductive structure consisting of a characteristic conidiophore culminating in an expanded bulbous region on which is borne phialides and metullae that generate chains of conidia (Bennett, ). Phylogenetic analysis has shown that they are essentially a monophyletic group (Peterson, ). They have a ubiquitous distribution, being present in decaying vegetation, soils and dust worldwide (Klich, ; Bennett, ). The aspergilli are of particular importance because they include in their number species that are highly beneficial to mankind, but others that are highly detrimental. Thus, certain species are used in food production, in industry for metabolite production including as a source of drugs, and as laboratory model organisms for basic and applied research. By contrast, other Aspergillus species can cause contamination of food stocks and give rise to life-threatening infections in immunocompromised hosts (Bennett, , ).
The aspergilli have proved especially valuable in studies of fungal sexuality because they include species with a range of different reproductive modes. Some ‘mitosporic’ species are only known to reproduce by asexual means and have been traditionally classified in the fungal phylum the Deuteromycota, which encompasses fungi lacking a known sexual state (Taylor et al., ). By contrast, ‘meiosporic’ species can also reproduce by sexual means and so have been traditionally classified in the fungal phylum the Ascomycota, whose characteristic feature is sexual reproduction involving the production of ascospores within asci. In the case of meiosporic Aspergillus species, they exhibit either heterothallic (obligate outbreeding) or homothallic (self-fertile) sexual breeding systems. The division of Aspergillus species into either the Deuteromycotina or the Ascomycotina is arguably misleading because phylogenetic analysis shows that they form a united group (Peterson, ), and the term ‘Deuteromycotina’ has now largely been replaced by the mitosporic or meiosporic terminology (Kück & Pöggeler, ). The majority of accepted Aspergillus species (approximately two-thirds of taxa) are only known to reproduce by asexual means, whilst those that do exhibit sexual cycles are overwhelmingly homothallic in nature with few heterothallic species described (Table 0001) (Dyer, ; Kwon-Chung & Sugui, ). Elucidating the genetic basis of such differences in reproductive mode is proving a fascinating challenge.
|Genus||Number of species||Examples||Homothallic : heterothallic|
|Chaetosartorya||3||C. chrysella, C. cremea||3 : 0|
|Dichotomomyces||2||D. albu, D. cejpii||2 : 0|
|Emericella||43||E. nidulans, E. heterothallica||42 : 1|
|Eurotium||64||E. herbariorum, E. chevalieri||64 : 0|
|Fennellia||3||F. flavipes, F. nivea||3 : 0|
|Neocarpenteles||1||N. acanthosporum||1 : 0|
|Neopetromyces||1||N. muricatus||1 : 0|
|Neosartorya||40 (46)||N. fischeri, N. fumigata||33 : 7|
|Penicilliopsis||2||P. dybowskii, P. flavidus||2 : 0|
|Petromyces||5||P. alliaceus, P. flavus||1 : 4|
|Sclerocleista||2||S. ornata, S. thaxteri||2 : 0|
|Warcupiella||1||W. spinulosa||1 : 0|
|Genus||Number of species||Examples||Homothallic : heterothallic|
|Chaetosartorya||3||C. chrysella, C. cremea||3 : 0|
|Dichotomomyces||2||D. albu, D. cejpii||2 : 0|
|Emericella||43||E. nidulans, E. heterothallica||42 : 1|
|Eurotium||64||E. herbariorum, E. chevalieri||64 : 0|
|Fennellia||3||F. flavipes, F. nivea||3 : 0|
|Neocarpenteles||1||N. acanthosporum||1 : 0|
|Neopetromyces||1||N. muricatus||1 : 0|
|Neosartorya||40 (46)||N. fischeri, N. fumigata||33 : 7|
|Penicilliopsis||2||P. dybowskii, P. flavidus||2 : 0|
|Petromyces||5||P. alliaceus, P. flavus||1 : 4|
|Sclerocleista||2||S. ornata, S. thaxteri||2 : 0|
|Warcupiella||1||W. spinulosa||1 : 0|
All legitimate species registered at MycoBank (http://www.mycobank.org) on 5 May 2011. Note that the number of Aspergillus species registered at MycoBank is higher than the figure reported by Samson & Varga ().
There have been a number of dedicated reviews over the past decade dealing with different aspects of sexual reproduction in Aspergillus species, some reviewing the aspergilli in general (e.g. Dyer, ; Geiser, ) whilst others have focussed specifically on Aspergillus nidulans (e.g. Braus et al., ; Han et al., ; Han, ; Han & Han, ). The present review will first give an overview of sexual reproduction in the aspergilli as a whole and then provide a summary of the current state of knowledge of genes involved with sexual development in Aspergillus species. This includes updating information from previous reviews and covering new material relating to recently characterized genes, and the involvement of sex-related genes in sclerotial development. Finally it will describe the recent discoveries of sexual cycles in Aspergillus fumigatus, Aspergillus flavus, Aspergillus parasiticus and Aspergillus nomius and discuss the significance of these findings in terms of cryptic sexuality in the aspergilli given their economic and medical importance.
Overview of sexual development in the aspergilli
Morphology of teleomorph states
The Pezizomycotina encompasses those Ascomycota that grow by production of filamentous hyphae. In this subphylum, sexual spores (ascospores) are housed in one of four main types of fruiting body (ascomata/ascocarps): cleistothecia, perithecia, apothecia or pseudothecia. They differ in their size, shape, style and organization of the asci, and presence and type of interascal sterile hyphae (Pöggeler et al., ). Members of the genus Aspergillus produce their ascospores in cleistothecia, which are the only ascomatal type that fully encloses the asci and ascospores (Gre. kleistos, closed + Gre. thēkion, small case). Cleistothecia may contain up to 100 000 asci, each enclosing (with very rare exceptions) eight ascospores; in the case of A. nidulans, an average cleistothecium may contain around 80 000 viable ascospores (Pontecorvo, ; Braus et al., ). Rather than being forcefully discharged, ascospores are released following the natural breakdown of the ascus wall and the outer wall (peridium) of the cleistothecium in their natural environment such as soil or decaying vegetation. One exception is A. athecius, which fails to form true cleistothecia; instead, asci develop directly from ascogonial coils (coiled hyphae containing the maternal nuclei that are fertilized at the onset of sexual reproduction) and are borne naked on undifferentiated mycelium (Raper & Fennell, ).
Following conventional fungal nomenclature, when sexual reproduction occurs the sexual phase is termed the ‘teleomorph’ state, whereas when asexual reproduction occurs the asexual phase is termed the ‘anamorph’ state. Thus, sexually reproducing species of Aspergillus will have two Latin binomials names, one for the anamorphic state and one for the teleomorphic state under the rules of ‘dual nomenclature’, for example A. fumigatus is the anamorphic state of the teleomorph Neosartorya fumigata (O'Gorman et al., ). Aspergillus species, as defined by their anamorphic state, are phylogenetically linked to 12 different teleomorphic genera (Table 0001) (Peterson, ; Samson & Varga, ). This diversity might appear surprising, but is attributed to the fact that the aspergilli exhibit a wide range of differing sexual fruit body morphologies, despite sharing similar asexual morphology, and hence a variety of teleomorphic genera have been erected to distinguish the different sexual states. Strictly speaking, a teleomorph name (where available) should be used in preference to the anamorph name according to the rules of dual nomenclature (Bennett, ). However, it should be noted that the future application of dual nomenclature is currently under debate [see discussion in Bennett (), supplementary materials of O'Gorman et al. (), and Hawksworth et al. ()]. In this review, where an anamorph name is particularly well established [e.g. A. fumigatus (teleomorph N. fumigata) and A. nidulans (teleomorph Emericella nidulans)], this epithet will be referred to by convention.
Eight of the teleomorphic genera associated with the aspergilli contain less than five species (Table 0001). The four most common and widely distributed teleomorphic genera are Eurotium, Emericella, Neosartorya and Petromyces, which are distinguished by the morphology of their cleistothecia as illustrated in Fig. 0001. One of the main differences between them is the composition and often colour of the cleistothecial wall (the ‘peridium’) (Benjamin, ; Geiser, ). In Eurotium, a single layer of frequently yellow, large flattened cells forms the cleistothecial wall (Fig. 0001a). In Emericella, the peridium is composed of two layers of often dark-purple flattened cells, which in A. nidulans at least are glued together with an unknown electron-dense substance (‘cleistin’) that fills the intercellular spaces (Fig. 0001b) (Sohn & Yoon, ). By contrast, the cleistothecial walls of Neosartorya consist of a network of layers of interlocking flattened hyphae that are normally white to light yellow in colour (Fig. 0001c). In the genus Petromyces, cleistothecia are formed inside another larger structure known as a ‘sclerotium’. This is a hardened, thick-walled, generally darkly pigmented and spherical structure formed for survival under adverse conditions. Cleistothecia develop within the sclerotium, embedded in a stroma (a mass or matrix of vegetative hyphae) consisting of pseudoparenchymatous hyphae called the stromal matrix. In this latter case, the peridia of the cleistothecia are composed of irregularly flattened cells (Fig. 0001d). Within these teleomorphic genera, the nature of ascospore ornamentation can often be used as a species-specific taxonomic character (Samson & Varga, ), with some exquisite forms present.
In general, distinct gametangia (e.g. ascogonia and antheridia) have not been observed in any of the aspergilli except for the genus Fennellia (Geiser, ). However, this might be an artefact because of the overwhelming number of homothallic aspergilli, which might be expected to exhibit reduced mating apparatus concomitant with other genomic adaptations (Paoletti et al., ). Evidence from both classical cytological studies and more recent genetic and molecular analyses indicates that nuclei from one mating partner (or nucleus in a homothallic species) normally contribute entirely to the formation of maternal tissues forming the accessory tissues and cleistothecial wall, whilst the other paternal partner merely contributes a compatible fertilizing nucleus that undergoes division and passes into the ascogenous hyphae (the specialized multinucleate hyphae contained within the cleistothecia where compatible nuclei fuse to give rise to crozier tips and asci) and ultimately the diploid zygote (Apirion, ; Zonneveld, ; Bruggeman et al., ; Todd et al., ). However, ‘mosaics’ where both partner nuclei contribute to the cleistothecium wall have also been observed in A. nidulans, although only infrequently such as the specialized instance of sex developing from a preformed heterokaryotic mycelium (Bruggeman et al., ).
Emericella is the only known teleomorphic genus that produces Hülle cells (Ger. die Hülle, casing/envelope) (Raper & Fennell, ). These are large (c. 10–15 μm diameter), thick-walled, globose cells that surround the cleistothecia during development, with several hundred Hülle cells surrounding the cleistothecia of A. nidulans for example (Sarikaya Bayram et al., ). They are multinucleate, but interestingly at maturity a macronucleus with a volume 20× that of normal nuclei is formed in A. nidulans, apparently from the grouping of smaller nuclei (Carvalho et al., ). The precise morphology of Hülle cells varies according to species grouping and can be used as a valuable diagnostic character, and Hülle cells can germinate to initiate new hyphal growth (Raper & Fennell, ). Hülle cells have a specialized physiology, exhibiting, for example, laccase and chitin synthase activity (Bayram & Braus, ), and it has long been believed that they ‘nurse’ the cleistothecia during development, such as through the production of α-1,3 glucanase that mobilizes carbon resources required for fruiting body development (Wei et al., ). Indeed, it was recently shown that a reduction in the number of Hülle cells surrounding ascomata of A. nidulans to only 2–5 per cleistothecium in a ΔlaeA mutant (see later) resulted in significantly smaller cleistothecia, which were only 40 μm in diameter unlike wild-type cleistothecia (c. 200 μm diameter) (Sarikaya Bayram et al., ). Hülle cells are also present in some supposedly asexual species in the sections Nidulantes (e.g. A. multicolour, A. subsessilis, A. sydowii) and Usti (e.g. A. puniceus) (Raper & Fennell, ; Klich, ). Intriguingly, their presence in these species suggests that they are vestigial remnants of sexual cycles that were presumably lost over the course of evolution. Similarly, a number of supposedly asexual species in sections Flavi (e.g. Aspergillus caelatus) and Nigri (e.g. A. sclerotioniger) produce sclerotia composed of hardened masses of hyphae (McAlpin, ; Samson et al., ). It was previously thought that sclerotia provide an evolutionary advantage by functioning as a survival mechanism during adverse environmental conditions. However, it has been suggested that, under the right conditions, such sclerotia would also be capable of acting as repositories for cleistothecia (Rai et al., ; Geiser et al., ).
Given that the genus Aspergillus forms an essentially monophyletic clade (Peterson, ; Samson & Varga, ), there has been speculation that the ancestral type of ascomata might have been characterized by a well-developed pseudoparenchymatous wall enclosing the asci, borne within a loose hyphal stroma. Divergent evolution within the aspergilli then led to a gradual reduction and simplification of the peridium in some cases, together with either specialization or loss of stroma tissues during evolution (Malloch & Cain, ). Indeed, in later parts of the present review, it will be observed that some of the initial developmental pathways relating to sex, such as Hülle cell and sclerotial production, can proceed independently of the formation of ascogenous hyphae and meiosis. Thus, two successive sexual developmental programmes are apparent in the aspergilli; the first relating to the initial formation of supporting accessory tissues needed for sex in some species, for example Hülle cells and sclerotia, and the second relating to subsequent cleistothecia and ascospore production. Normally, these developmental programmes would be expected to be synchronized and follow on from one another. However, sometimes they may become unlinked. This is illustrated by species that form accessory tissues independent of the later stages of sex, for example Hülle cell formation in A. multicolour, A. subsessilis, A. sydowii and A. puniceus (see above); sclerotial formation in A. caelatus and A. sclerotioniger (see above); and various mutants of A. nidulans (see later) that can produce copious amounts of Hülle cells despite the absence of meiosis and ascospore development. It is significant to note that members of the genus Petromyces may produce from single to multiple cleistothecia, each with a distinct peridium, within a single sclerotium – this fact emphasizing the independence of the two developmental pathways. (e.g. see Fig. 3 of Horn et al., ). It will be interesting to compare the developmental pathways of stromatic accessory tissues in the aspergilli with those seen in more traditional ascoloculate fungi (i.e. Ascomycota where the asci develop in cavities in a preformed stroma) (Debuchy et al., ).
Environmental factors influencing sexual reproduction
Within the aspergilli, and indeed the Pezizomycotina as a whole, sexual and asexual spore production can be viewed as two competing forms of reproduction with one developmental process inhibiting the other and vice versa. The type of reproduction that occurs is largely governed by environmental factors, with conidia or ascospore production favoured as the result of a sum of environmental influences (Dyer et al., ). In the case of the aspergilli, there is no single set of conditions that favours sexual reproduction; rather this is both species and subgenus specific and relates to the ecology of the species involved. As most of our knowledge comes from studies of the model organism A. nidulans (as will be described below), it should therefore be cautioned that these findings may not be applicable to all aspergilli.
A first key factor influencing sexual development is the absence or presence (including wavelength) of light. Given that many Aspergillus species are natural soil dwellers (Klich, ), the need to determine whether they are above or below ground can be seen as essential because this can favour the production of airborne conidia or more environmentally resistant ascospores, respectively. In A. nidulans, the sexual cycle can take place in both the light and dark, but darkness is preferential, as light delays initiation of the pathway by 15 h (Yager, ). This ‘light-sensitive period’ lasts from 20 to 50 h after inoculation. Furthermore, Zonneveld () showed that incubation in darkness for 24 h directly after inoculation significantly increases cleistothecial numbers. Han et al. () found that cleistothecia were not formed in intense light and suggested that this is a stress response, as vegetative growth is also blocked at high light intensities. In particular, the sexual cycle is differentially regulated by exposure to red light, blue light and far-red light, with red and blue light inhibiting sexual development and far-red light enhancing development (Blumenstein et al., ; Bayram et al., ). Meanwhile, light has also been shown to influence sclerotial formation (a required precursor for cleistothecial development) in the genus Petromyces. In general, increased numbers of sclerotia are produced in darkness by species including A. flavus and A. parasiticus (Rai et al., ; Bennett et al., ; Calvo et al., ; Duran et al., ), although some strains of A. flavus and black aspergilli have been reported to produce equal or greater numbers of sclerotia in the light (Rai et al., ; Calvo et al., ). Light of blue and green wavelengths was particularly inhibitory to sclerotial development in A. parasiticus, whereas red light had no effect (Bennett et al., ).
A second major factor determining whether sex can occur is the composition of the growth medium such as nutrient content and pH. In the case of A. nidulans, sexual development requires ‘well-nourished’ conditions for abundant cleistothecial production (Han et al., ), in contrast to nutrient limitation that triggers sex in many other pezizomycete fungi (Dyer et al., ). Once above a threshold carbon level, cleistothecial numbers in A. nidulans are directly related to the amount and type of carbon that is available (Zonneveld, ; Han et al., , ). High carbon concentrations (e.g. 6% glucose) inhibit sex, but this is because nitrogen then becomes the limiting factor, whilst some carbon sources (e.g. lactose) can promote sex even in the light whereas other sources (e.g. acetate) are inhibitory (Han et al., ). In general, a balanced carbon/nitrogen ratio is most favourable for sex, and when suitable nitrogen sources are available, sexual reproduction is preferred over asexual sporulation (Zonneveld, ; Han et al., ). A slight increase in glucose and decrease in nitrogen concentration was claimed to favour cleistothecial development in A. nidulans (Swart et al., ). By contrast, development of cleistothecia in xerophilic Eurotium species occurs best at media containing up to 30% glucose (Blaser, ). Another aspect of well-nourished conditions is the presence of a suitable amino acid source(s), with development of cleistothecia found to be blocked specifically at the microcleistothecial stage when A. nidulans is cultivated under amino acid starvation (Hoffmann et al., ). The importance of phosphorus has been much less studied, but low phosphorus concentrations suppress sexual development, presumably due to a requirement for phosphorus in the generation of ATP (Bussink & Osmani, ). Manganese is the only trace element to date that has been shown to be essential for cleistothecium formation in A. nidulans (Zonneveld, , ). A lack of manganese impairs the activity of the enzyme phosphoglucomutase, which is required for the synthesis of α-1,3-glucan (Zonneveld, ). Indeed, A. nidulans is acleistothecial without α-1,3-glucan (Polacheck & Rosenberger, ). Further factors influencing sexual reproduction include the pH of the growth medium, maximum numbers of cleistothecia in A. nidulans being produced around neutral pH values (Rai et al., ; Thakur, ), and the salt concentration – relatively high levels of salt inhibiting the sexual cycle although vegetative growth remains possible (Lee et al., ; Kim et al., ; Wang et al., ). Meanwhile, growth medium composition also has a major influence on the induction and maturation of sclerotia in the aspergilli. For example, numbers of sclerotia produced and percentage of stroma bearing ascocarps is greatly influenced by the type of nitrogen source in Petromyces alliaceus, with certain amino acids enhancing the formation of sclerotia but not necessarily leading to a concomitant increase in numbers of ascocarps (McAlpin & Wicklow, ). The source and concentration of carbon, nitrogen and sulphur also impact on sclerotial production in certain members of the black aspergilli (Agnihotri, , ). Similarly, sclerotial formation by A. caelatus and Aspergillus japonicus is affected greatly by the type and concentration of carbon and nitrogen sources (Heath & Eggins, ; McAlpin, ). Sclerotial development in Aspergillus ochraceus is greatly affected by both the nutrient source and pH of the medium (Paster & Chet, ), and sclerotial formation is pH dependent in a number of other aspergilli (Rai et al., ; Thakur, ). Varying medium depth and water availability also has a dramatic impact on sclerotial production in A. flavus (Thakur, ; Okuda et al., ; Nesci et al., ). Furthermore, attempts to induce the sexual cycle of A. fumigatus in vitro were only successful on one particular medium (oatmeal agar) of various tried, illustrating the ‘fastidious’ nature of certain aspergilli in their sexual requirements (Kwon-Chung & Sugui, ; O'Gorman et al., ). Finally, it is noted that addition of linoleic acid can significantly promote the production of both cleistothecia and sclerotia in A. nidulans and A. flavus, respectively (Calvo et al., ; Brown et al., ), possibly linked to a role as both an energy reserve and precursor for oxylipin biosynthesis (see below) (Dyer et al., ; Brodhun & Feussner, ).
A third factor influencing the extent of sexual reproduction is the presence of various atmospheric gases. Development of cleistothecia in A. nidulans will normally only occur when grown on a medium with an air interface, for example an agar surface. If grown in submerged liquid culture, the mycelium normally remains entirely undifferentiated, that is, there is a fundamental influence of dimensionality (Champe et al., ). On solid media, cleistothecial formation in A. nidulans is significantly enhanced by elevated levels of carbon dioxide (Zonneveld, , ). This is commonly achieved in the laboratory by sealing the plates to limit air exchange. As CO2 is needed for both the synthesis and breakdown of α-1,3-glucan, a lack of CO2 reduces cleistothecial production (Zonneveld, ). Plate sealing in addition functions to lower the oxygen concentration, causing irreversible entry into the sexual cycle. This appears to arise because of the partial blockage of the electron transport system from the lack of oxygen (Han et al., ). By contrast, plate sealing impedes cleistothecium formation in Eurotium species (A. Ashour & P.S. Dyer, unpublished results), possibly reflecting the different ecological niches of each genus – xerophilic Eurotium species often being found on exposed surfaces c.f. soil-dwelling Emericella species. Plate sealing can also prevent the formation of sclerotia in some strains of A. flavus and black aspergilli (Okuda et al., ; H. Darbyshir & P.S. Dyer, unpublished results). Relative humidity (RH) also greatly influences both sclerotia and cleistothecia production in the aspergilli, maximum numbers being produced at 100% RH (Rai et al., ). Most recently, nitric oxide has been shown to promote cleistothecial production in A. nidulans (Baidya et al., ; Marcos et al., ).
A final environmental factor influencing sexual development is temperature. Traditionally, A. nidulans is incubated at 37 °C to induce the sexual cycle (Pontecorvo, ), whereas an almost twofold increase in the number of cleistothecia can be achieved by incubation at 32 °C (Dyer, ). Production of cleistothecia by A. fumigatus is also temperature dependent, fruit bodies being observed only at one (30 °C) of the various temperatures assayed (O'Gorman et al., ). Similarly, highest numbers of cleistothecia are produced by Eurotium species between 27 and 33 °C (Blaser, ), and sclerotial production in A. flavus is temperature dependent (Rai et al., ).
In addition, there are other factors that can influence the success of crosses within a laboratory setting. In the case of heterothallic aspergilli, it is becoming apparent that much greater success can be achieved by inoculating mating partners in a barrage formation rather than using a single point or spread mixed spore inoculum (O'Gorman et al., ; S. Swilaiman & P.S. Dyer, unpublished results). And it has been found that the density of spore inoculum used to establish Petri dish cultures has a major impact on formation of sclerotia by A. flavus and A. parasiticus, the greatest mass of sclerotia being produced using low spore concentrations in the range 101–103 spores per plate, possibly relating to quorum-sensing oxylipin signalling (Brown et al., , ).
Benefits of sexual reproduction
There are many potential reasons why a number of aspergilli, especially facultatively sexual species such as A. nidulans, have retained the ability to undergo sexual reproduction despite the increased metabolic costs compared to purely asexual reproduction (Rice, ). Sexual reproduction involving outcrossing (which is possible even in homothallic species) can generate large amounts of genetic variation and produce novel genotypes much faster than by asexual means, as illustrated by the many novel genotypes that were observed to arise from crossing strains of A. fumigatus (O'Gorman et al., ). This can increase the mean fitness of the next generation, quickening adaptation to changes in the environment and improving their chances for long-term survival. Such an increase in fitness associated with sex has been shown experimentally for A. nidulans (Schoustra et al., ). In parallel, sex increases gene flow, bringing together and subsequently proliferating advantageous mutations as well as removing deleterious mutations that would otherwise persist and accumulate in asexual populations (i.e. the concept of ‘Muller's ratchet’) (Dyer & Paoletti, ). Indeed, sexual reproduction has been shown to slow down the accumulation of deleterious mutations in A. nidulans, possibly through a ‘selection arena’ operating at the dikaryotic stage to favour the proliferation of nuclei with higher fitness (Bruggeman et al., , ). In addition, the ascospores of many Aspergillus species have a far superior ability compared to conidia to survive harsh environmental stresses such as desiccation and high temperatures, in part because of their thickened cell walls (Baggerman & Samson, ; Dijksterhuis, ).
However, it is important to note that there are also many benefits to asexual reproduction in the aspergilli. These include the ability to produce prolific numbers of conidia for dispersal of the species in a shorter time than is required for ascospore production (Champe et al., ), the lower metabolic costs associated with asexual sporulation, and the ability to produce asexual propagules on a wider range of substrates and under a broader set of environmental conditions than the often more fastidious requirements for sexual reproduction. In addition, asexual reproduction may allow better selection and adaptation for local niches and limit the spread of parasitic transposable elements (Wright & Finnegan, ).
Genes involved with sexual development in the aspergilli
The ease with which the homothallic sexual cycle of A. nidulans can be induced under laboratory conditions, together with availability of genome sequence data and a variety of tools allowing genetic manipulation (e.g. Archer & Dyer, ; Galagan et al., ; Todd et al., ), has made this species an ideal subject for studies aiming to identify genes involved in sexual development. Indeed, almost all research efforts in this area relating to the aspergilli have been focussed on A. nidulans. It is therefore cautioned that results from A. nidulans might not necessarily be generally applicable to all aspergilli. Also, most studies using A. nidulans have been made with veA1 strains (described below), and therefore results may not fully reflect wild-type functionality. However, there are now also an increasing number of studies on other aspergilli to substantiate results with A. nidulans, notably A. fumigatus which exhibits a heterothallic breeding system (O'Gorman et al., ). The latter species is therefore a more suitable organism for studying the genetic basis of the mating process in obligate outcrossing Aspergillus species, especially given that it is now possible to cross isolates in only 4–8 weeks (J. Sugui, C.M. O'Gorman, P.S. Dyer & K.J. Kwon-Chung, unpublished results). In addition, a limited number of studies have investigated the molecular genetic basis of sclerotial development in the aspergilli, this process now known to be a vital early stage of sexual reproduction in Aspergillus species with teleomorphs in the genus Petromyces.
To date, at least 78 genes have been identified which are associated with sexual reproduction in A. nidulans, A. fumigatus, A. flavus, A. parasiticus and most likely the aspergilli in general. Table 0002 lists all genes currently known to have a proven role in the sexual cycle. Some of these genes are essential for sexual development to occur, whereas others appear to moderate levels of fertility (‘fertility’ here used to refer to numbers of cleistothecia produced containing mature ascospores; high fertility meaning high numbers produced and vice versa for low fertility). Many of these genes have been described in depth in the review of Dyer () to which readers are referred for extra detail. An updated version now follows, including details of genes identified since that review was written and incorporating new data from Aspergillus species not previously included. These ‘sex-related’ genes are involved at various stages of the sexual developmental pathway, and a schematic genetic network showing relationships between the various genes is proposed in Fig. 0002. A key principle of this is that ultimately sex involves bringing together two nuclei to form a diploid zygote, which then undergoes meiosis and all the ‘sexual circuitry’ works to enable this to occur. Thus, some sets of genes are involved with environmental sensing of suitable conditions for sex; these genes then activate the early stages of sexual development with the formation of maternal hyphae with receptive nuclei, the ascogonia. Most likely, these environmental sensing genes then also activate/upregulate genes controlling mating processes designed to attract and bring about the union of complementary nuclei with those in the ascogonia. Once fertilization and karyogamy have occurred, later signalling processes occur to bring about the synchronous development of the fruit body and meiotic tissues. The precise details of the relative stages of gene action and importance of certain genes may vary according to whether species exhibit homothallic or heterothallic breeding systems, but the overall schematic model in Fig. 0002 should provide a useful guide for the aspergilli in general, as well as being valid for aspects of sexual reproduction in other pezizomycete fungi. In the following section, these sets of genes are categorized under various subheadings according to their particular site of action in the sexual cycle or nature of encoded protein.
Genes listed as ‘Activators’ are required for sexual reproduction, and gene overexpression may enhance sexual fertility. Expression of genes listed as ‘Repressors’ may reduce sexual fertility in general or operate only under particular environmental conditions. See main text for full details.
Comprehensive details of gene and protein sequence are as available from the AspGD website: http://www.aspergillusgenome.org/. Also see listing of Aspergillus gene names at: http://www.fgsc.net/Aspergillus/gene_list/loci.html.
Acts as a repressor or activator under different environmental conditions and/or at different stages of the sexual cycle. Effects might only be seen under forced gene expression.
Data not yet available.
Based on marked upregulation during sexual morphogenesis.
It is noted that other genes required for general metabolism may also be necessary for sexual development [e.g. genes for arginine, histidine, tryptophan, riboflavin and phosphatidylcholine biosynthesis, and sumoylation (Champe et al., ; Eckert et al., ; Todd et al., ; Tao et al., ; Sarikaya Bayram et al., )], but these are not referred to here as these roles are arguably not specific to sex. Also, many of the genes described below have been detected by blast analysis in the genomes of presumed asexual aspergilli such as Aspergillus oryzae, Aspergillus niger and Aspergillus terreus (Galagan et al., ; Dyer, ; Pel et al., ), but their functional role in these species is as yet undetermined so will not be discussed further.
Perception of environmental signals – light
As described above, the sexual cycles of A. nidulans and A. fumigatus are induced only under specific environmental conditions, and a variety of genes have been identified linked to sensing of external factors. Regarding photobiology, a series of genes that are responsive to light have been identified in A. nidulans: the fphA gene encodes a red light–sensing phytochrome (Blumenstein et al., ), the lreA and lreB genes encode blue light–sensing white-collar homologues (Purschwitz et al., ), and the cryA gene encodes a blue light- and UVA-sensing cryptochrome (Bayram et al., ). Analysis of A. nidulans deletion mutants of each of these genes indicates that FphA and CryA repress sexual reproduction under red and UVA/blue light wavelengths, respectively (Blumenstein et al., ; Bayram et al., ). For example, upregulation of both veA and nsdA (see below) was seen in a ΔcryA mutant, suggesting that cryA acts as an upstream repressor of these genes (Bayram et al., ). By contrast, LreA and LreB activate sexual reproduction in the dark and also promote limited sex in the light (Purschwitz et al., ). Furthermore, bimolecular fluorescence assays suggest that there is physical interaction between lreA, lreB, FphA and an additional key protein VeA (see below), which together form a nuclear bound ‘light regulator complex’, and it is the interaction of proteins within this complex that leads to the balance between asexual and sexual development under various light/dark conditions (Purschwitz et al., ; Bayram et al., ).
A second protein complex known as the ‘velvet’ complex also has a pivotal role in light/dark regulation of sexual development, most likely acting downstream of the previous light regulator complex (Fig. 0002). The core components of the velvet complex are three proteins VeA, VelB and LaeA, together with associated proteins VosA, VipA, VipB and VipC (Bayram et al., , ; Bayram & Braus, ). The overall action of the velvet complex in A. nidulans is to activate sexual development in the dark whilst suppressing asexual reproduction. However, analysis of the individual protein components has revealed differences in their specific activities. The veA gene was first to be studied, with gene deletion resulting in a complete inability to reproduce sexually, whereas overexpression resulted in sexual morphogenesis under high salt and liquid conditions normally unfavourable for sex (Kim et al., ). By contrast, overexpression of veA in A. fumigatus had no obvious phenotype (Krappmann et al., ). Deletion of veA in A. nidulans also prevented expression of the oxylipin gene ppoA (see below) and a regulatory feedback loop has been suggested (Stinnett et al., ). Tandem affinity purification studies then identified that VeA interacts with two further proteins VelB and LaeA, with VeA thought to act as a putative ‘bridge’ via interactions with the N- and C-terminal portions of VeA, respectively (Bayram et al., ). The velvet-like protein VelB shares homology with VeA, and the proteins appear to be functionally interdependent with a ΔvelB gene deletant unable to form cleistothecia even in the dark; this defect could not be rescued by veA overexpression (Bayram et al., ). However, unlike veA, velB overexpression did not result in increased numbers of cleistothecia, so it was suggested that veA determines the production of the number of cleistothecia whereas velB has complementary developmental functions (Bayram et al., ). In particular, VelB was found to interact with the protein VosA, which is required for production of trehalose in both asexual conidia and sexual ascospores, hence affecting the viability of spores (Ni & Yu, ; Bayram et al., ; Sarikaya Bayram et al., ). The VelB–VosA heterodimer was suggested to both repress asexual development and mediate ascospore maturation (Sarikaya Bayram et al., ).
Meanwhile, it was a major discovery that VeA also interacts with the putative methyltransferase LaeA, itself a global regulator of much secondary metabolism in A. nidulans (Bayram & Braus, ). Further analysis into the role of LaeA in sexual development revealed that LaeA is required for light-mediated inhibition of sex, with deletion of laeA resulting in elevated numbers of cleistothecia being produced both in the light and dark (Sarikaya Bayram et al., ). However, the cleistothecia formed were significantly smaller in size than the wild type. This was most likely due to a drastic reduction in the number of Hülle cells produced, and lack of concomitant mutanase activity, thought to be needed for nursing of the developing cleistothecia, that is, LaeA is required for correct Hülle cell development (Fig. 0002). Furthermore, overexpression of LaeA in the dark led to a twofold increase in the production of cleistothecia. Thus, LaeA appears to have bimodal activity, suppressing sexual development in the light but promoting it in the dark. LaeA possibly acts by chemically modifying VeA, which was found to have two different cellular isoforms (Sarikaya Bayram et al., ; Bayram & Braus, ). Finally, it was observed that VeA also interacted with a nuclear importin protein KapA (Stinnett et al., ; Bayram et al., ). Taken together, this has led to the current model whereby VeA, VelB and KapA are thought to be transported to the nucleus in the dark, where resulting interaction with LaeA (possibly via an epigenetic effect) results in sexual development and altered secondary metabolism (Sarikaya Bayram et al., ; Bayram & Braus, ). This provided evidence of a molecular link between secondary metabolism and development, with Sarikaya Bayram et al. () speculating that these processes might be interconnected, with the production of toxic secondary metabolites of benefit to protect fruiting bodies in the soil environment.
It is noted that the role of LaeA in sexual development had not been identified earlier because, as mentioned above, many laboratory strains of A. nidulans contain a particular mutation in the veA gene leading to the production of a truncated protein lacking the first 37 amino acids (Kim et al., ). This veA1 mutation leads to a lack of red light sensitivity and increased asexual sporulation and reduced sexual development in the dark (these being advantageous traits for many laboratory projects). The mutation also largely eliminates the formation of aerial hyphae, whereas these are formed by wild-type veA strains that have a velvety appearance, hence the gene name. And critically it was this N-terminal portion of the protein that appears to be required for interaction with LaeA, possibly due to nonfunctionality of a nuclear localization signal in the truncated protein (Stinnett et al., ; Sarikaya Bayram et al., ). Unfortunately, this veA1 effect could cause similarly misleading conclusions to be drawn for other genes impacting on sexual development. LaeA has also been shown to be required for the development of sclerotia in A. flavus, with a complete loss of sclerotial formation in a ΔlaeA strain, whereas an overexpressing strain showed a c. sevenfold increase in mass of sclerotia in the dark (Kale et al., ). There was also evidence that laeA negatively regulates veA expression (Kale et al., ) and that multicopy expression of laeA (but not veA) could obviate the density-dependent suppression of sclerotial formation both in the light and dark when using high spore concentrations as the starting inoculum (Amaike & Keller, ).
A protein kinase ImeB has also been shown to be involved in light-mediated inhibition of sexual development (Bayram et al., ). This is perhaps surprising because the yeast homologue Ime2 is involved specifically in meiosis and the production of ascospores. ImeB deletion strains of A. nidulans produced three- to fourfold increased numbers of cleistothecia in the light compared to controls and were observed to form Hülle cells in submerged culture. Expression of the VeA, VelB, stuA and nsdD genes (see below) was increased in the ΔimeB strain, suggesting that ImeB acts upstream of these genes (Fig. 0002). Interactions with fphA were also investigated, and it was concluded that ImeB and FphA are involved in different light-response pathways (Bayram et al., ). However, somewhat surprisingly overexpression of imeB led to ‘enormous numbers’ of sexual structures independent of any light effect, so Bayram et al. () suggested that ImeB also had a later role in promoting fruit body formation.
A few final comments on light sensing are that there is possible redundancy in receptor function and crosstalk between the red light- and blue light–sensing systems (Bayram et al., ). Also, a VeA homologue has also shown to be required for sclerotial development in both A. parasiticus and A. flavus (Calvo et al., ; Duran et al., ), and given that these structures are required for later development of cleistothecia, it can be seen that VeA activity is involved in the co-ordination of sexual morphogenesis in the aspergilli in general. Furthermore, a velvet-related VelC protein has recently been identified in A. nidulans with activity in enhancing sexual development in A. nidulans. The velC gene is predicted to encode a 524-amino acid polypeptide and is one of many potential VosA-interacting proteins identified via yeast two-hybrid (Y2H) screens. Overexpression of velC caused elevated numbers of sexual fruiting bodies to be formed, whilst deletion of velC resulted in increased asexual spore formation and decreased fruit body production (Sarikaya Bayram et al., ; H.S. Park & J.H. Yu, pers. commun.). Finally, a mutant screen has identified a series of isolates able to undergo sex in the light, but the underlying genetic basis has been determined for only two of these mutants. First, deletion of the silA gene (which encodes a zinc-finger domain protein) resulted in a blind phenotype showing strong induction of cleistothecia in the light (Min et al., ; Han et al., ). Second, deletion of the silG gene (also encoding a zinc-finger domain protein) similarly resulted in a mutant able to produce high numbers of cleistothecia under visible light (Han et al., ), that is, both SilA and SilG were repressors of sex in the light.
Perception of environmental signals – nutrient and stress levels
The cross-pathway-control genes cpcA and cpcB are involved in the sensing of amino acid levels (via interaction with a CpcC kinase) and can modulate sexual development in A. nidulans. Under limiting amino acid conditions, the c-Jun related transcription factor CpcA suppresses development beyond the formation of microcleistothecia, even though Hülle cells are present, whereas the Gβ subunit protein CpcB represses cpcA expression in the presence of suitable amino acid levels to allow sexual development (Fig. 0002) (Hoffmann et al., ). Other genes involved in nutrient sensing are the lsdA gene, which can suppress sex under high salt concentrations (Lee et al., ), imeB (see above), which has a possible role in inhibiting sex under low glucose conditions although this remains to be proven (Bayram et al., ), phoA, which encodes a cyclin-dependent kinase that is inhibitory under low phosphorus conditions (Bussink & Osmani, ), and An-pho80, which encodes a cyclin that conversely promotes sexual development at low phosphorus concentrations (Wu et al., ). A further esdC gene involved in early sexual development has been identified, which encodes a protein containing a glycogen binding domain, suggesting a role in regulating developmental processes according to nutrient/metabolic status. Deletion of esdC led to a total failure to form cleistothecia even under conditions normally favouring sex, although gene overexpression had no obvious effect (Han et al., ). Significantly, only very low levels of esdC expression were seen in a ΔveA strain, and reduced levels in a ΔnsdD strain, suggesting that VeA and NsdD positively regulate esdC expression (Fig. 0002).
More generally, various environmental stressors can trigger elevated sexual reproduction. Recently, the addition of the nitric oxide (NO)-generating compound diethylenetriamine-NoNoate has been shown to result in an increase in cleistothecial production and higher expression of the regulatory genes nsdD and steA (see below) (Baidya et al., ; Marcos et al., ). Linked to this observation, the fhbA and fhbB genes encode flavohaemoglobin proteins involved in the reduction and detoxification of NO. Deletion of fhbA resulted in increased Hülle cell production (Baidya et al., ), suggesting that fhbA and fhbB are likely to suppress sexual development under stressful conditions if able to mitigate these circumstances.
Genes involved in the mating process
Studies of other filamentous ascomycete species (Pezizomycotina) have established that mating-type (MAT) genes have a key role in determining sexual identity in heterothallic species and are also important in sexual developmental processes in both heterothallic and homothallic species (Debuchy et al., ). MAT genes have now been identified from a range of Aspergillus species, including both known sexual and supposed asexual species (Dyer et al., ; Paoletti et al., ; Dyer, ; Pel et al., ; Ramirez-Prado et al., ).
MAT genes were first described from the homothallic A. nidulans, which was found to contain both alpha- and high mobility group (HMG) classes of MAT genes within the same genome; though, these were unlinked and found on chromosomes 6 and 3, respectively (Dyer et al., ; Galagan et al., ; Paoletti et al., ). Functional analysis of these genes, termed MAT1 and MAT2 in recognition of their distinct loci according to standard nomenclature for MAT genes, revealed that they were essential for correct sexual development, with deletion of either gene resulting in the production of significantly decreased numbers of cleistothecia that were smaller in diameter and entirely sterile (containing no ascospores) (Paoletti et al., ). However, ΔMAT1 and ΔMAT2 strains were able to outcross to each other, that is, the key observation that manipulation of the MAT genes had lead to heterothallic (self-sterile) strains from a homothallic parent. The ΔMAT2 strains also showed occasional prolific production of Hülle cells. Overexpression of both MAT1 and MAT2 together resulted in vegetative growth arrest and development of cleistothecia on media normally unfavourable for sex, further confirming their key role in sexual development (Paoletti et al., ). A preliminary study by Miller et al. () also reported regulation of fruit body formation and meiosis in A. nidulans by the unlinked MAT loci. A further preliminary study by Alcocer et al. () reported the presence of a noncoding RNA region termed ‘jgaA’ adjacent to MAT2 (matA), whose expression was co-regulated with MAT2 (matA), deletion of which led to a 42% decrease in production of cleistothecia. Given all these findings in the homothallic A. nidulans, it is intriguing to speculate whether there might be differential expression of the MAT genes in the nuclei that form the original sexual dikaryon and eventual diploid zygote, such that two mating competent nuclei may come together, one expressing MAT1 and the other MAT2 to allow nuclear fusion. Indeed, it has been suggested that MAT genes might have a role in determining nuclear identity (Debuchy, ; Debuchy et al., ) and could be involved in the process of ‘relative heterothallism’ in A. nidulans whereby a preference for nuclear outcrossing is evident despite the self-fertile nature of this species (Pontecorvo, ; Hoffmann et al., ; Scazzocchio, ; Paoletti et al., ).
MAT genes were next identified from A. fumigatus, which exhibited a standard organization for heterothallic Pezizomycotina with either a MAT1-1 alpha domain or MAT1-2 HMG domain gene being found at a single MAT locus in the genome, which showed an ‘idiomorphic’ structure with highly dissimilar regions of DNA sequence being present in the MAT locus of the different mating types, unlike the more normal allelic variants found at a gene locus (Paoletti et al., ). These genes were later shown to be functional from heterologous expression in A. nidulans hosts lacking functional MAT1 (MATB) and MAT2 (matA) genes (Groβe & Krappmann, ; Pyrzak et al., ), from mating studies (O'Gorman et al., ), and most recently from gene deletion work where ΔMAT1-1 and ΔMAT1-2 strains were unable to form fruiting bodies (Szewczyk & Krappmann, ; Szewczyk et al., ). Complementary alpha- and HMG domain MAT genes have since also been characterized from Neosartorya fischeri (Rydholm et al., ) and A. flavus, A. parasiticus and P. alliaceus (Ramirez-Prado et al., ). MAT1-1 and MAT1-2 idiomorph regions have in addition been identified from A. oryzae, and most recently altered gene expression was seen in an A. oryzae mating-type replacement strain in which the MAT1-1 gene was substituted by MAT1-2. Many genes, including the putative α-pheromone precursor gene AoppgA, were expressed more abundantly in the MAT1-1 strain, and several genes were upregulated in the MAT1-2 strain, suggesting functionality of these mating-type genes (R. Wada, H. Yamaguchiu, N. Yamamoto, Y. Wagu, M. Paoletti, D.B. Archer, J. Maruyama, P.S. Dyer & K. Kitamoto, unpublished results).
One key role of MAT genes in heterothallic pezizomycetes is to regulate the production of peptide pheromones, which are involved in the chemotactic attraction of mating partners. This involves the production of pheromones (from precursor molecules) in a mating-type-specific manner and the consequent export of either hydrophilic α-type pheromones (from MAT1-1 isolates) or hydrophobic a-type pheromones (from MAT1-2 isolates) (Debuchy et al., ). These peptide pheromones are then recognized by cognate G protein-coupled receptors (GPCRs), also produced in a mating-type-specific fashion with α-type pheromones docking to a Pre2 receptor (in MAT1-2 isolates) and a-type pheromones docking to a Pre1 receptor (in MAT1-1 isolates) (Debuchy et al., ). Research in the aspergilli first focussed again on the homothallic A. nidulans, with the identification of a ppgA α-type pheromone and both preA and preB receptor genes arising from genome analysis (Dyer et al., ). Subsequent experimental work by Seo et al. () revealed that both preA and preB (co-named gprB and gprA, respectively) were required for normal sexual development. Deletion of either gene alone resulted in the production of only a few, small cleistothecia containing significantly reduced numbers of cleistothecia, whereas a double deletion strain was completely sterile with no cleistothecia produced (although Hülle cells were formed). These results indicate that pheromone signalling is required for fertilization and nuclear karyogamy even in a homothalic species, despite the repeated failure to detect trichogynes or differentiated ascogonia and antheridia in A. nidulans (Sohn & Yoon, ; Dyer, ). Genes for pheromone precursors (ppgA) and pheromone receptors (preA and preB) have also been described from A. fumigatus and have been shown to be expressed on solid agar media (Dyer et al., ; Paoletti et al., ; Szewczyk & Krappmann, ). In the first study of their expression, there was no perceptible difference in expression of preA and preB between MAT1-1 and MAT1-2 isolates under the conditions assayed, although slightly increased ppgA expression was seen in a MAT1-1 isolate (Paoletti et al., ). Similar results were obtained in a subsequent study using wild-type isolates, and in addition, a significant decrease in ppgA expression was seen in a ΔMAT1-1 strain whereas increased expression was seen in a ΔMAT1-2 strain, consistent with ppgA regulation by the MAT loci (Szewczyk & Krappmann, ). However, despite much bioinformatic effort, no gene encoding a hydrophobic a-type ppgB pheromone has yet been detected in the aspergilli. This is most likely due to high sequence divergence amongst such pheromones and the relatively short length of product, although potential a-factor efflux genes atrC and atrD, encoding ABC transporters, have been identified from A. nidulans with homologues in A. fumigatus and A. flavus (Andrade et al., ). Other genes for enzymatic processing and maturation of pheromones (Saccharomyces cerevisiae ste13, ste14, ste24, kex-1, ram1, ram2 and rce1 homologues) have been detected by blast analysis of the A. nidulans genome (Dyer et al., ).
Signal transduction pathways
A variety of G protein-coupled signal transduction pathways have been identified in A. nidulans (Han et al., ; Galagan et al., ). One particular MAP kinase cascade has been identified as being linked to pheromone sensing in A. nidulans, based on homology to the same transduction pathway in the S. cerevisiae yeast model, and is essential for sexual reproduction to occur (Dyer, ). The PreA and PreB receptor proteins are linked to a heterotrimeric G protein composed of alpha (α), beta (β) and gamma (γ) subunits. Deletion of any of the genes encoding these subunits (fadA, sfaD and gpgA, respectively) results in the failure to form cleistothecia, although Hülle cells continued to be formed in abundance under some conditions (Rosén et al., ; Seo et al., ), that is, emphasizing the possible independence of Hülle cell formation from development of ascogenous hyphae and meiosis (see above). The FlbA protein, a regulator of G protein signalling because of its GTPase-enhancing activity, is also required for the development of both cleistothecia and Hülle cells as highlighted by the fact that a ΔflbA mutant showed only a fluffy-autolytic phenotype even under conditions favourable for sex (Han et al., ). Deletion of flbA resulted in minimal nsdD expression (see below), although no clear effect was seen in fadA and sfaD deletion strains (Han et al., ); expression of esdC was also dependent on flbA (Han et al., ). A phosducin-like protein PhnA is also required for correct Gβγ-mediated signalling, possibly acting as a molecular chaperone. Deletion of phnA resulted in a block in sexual development (although enhanced Hülle cell formation was noted), and no effect on nsdD expression was observed suggesting a downstream role or a different regulatory branch (Fig. 0002) (Seo & Yu, ).
A complete MAP kinase cascade in A. nidulans was then identified downstream of Gβγ-mediated signalling, based on both genomic and experimental findings (Dyer et al., ; Galagan et al., ; Dyer, ) (Fig. 0002). It consists of a series of consecutive elements with kinase activity [STE20 (MAPKKKK), steC (first identified as steB from a partial EST sequence) (MAPKKK), STE7 (MAPKK), mpkB (MAPK)] that activate a final homoeodomain protein SteA, which then triggers further stages of sexual development. Deletion of individual genes in the pathway has been shown to result in sterility, with failure to form cleistothecia and ascogenous hyphae, although Hülle cells were sometimes produced (Vallim et al., ; Wei et al., ; Paoletti et al., ). Significantly, Paoletti et al. () found that expression of the whole MAP kinase cascade, together with a ste50 kinase regulator, was markedly upregulated during sexual development. Thus, selfing in A. nidulans was interpreted to involve the activation of the same mating processes characteristic of sex in heterothallic species, that is, self-fertilization does not bypass the requirements of outcrossing sex but instead involves activation of these pathways within a single individual (Paoletti et al., ). However, unlike heterothallic species, certain aspects of pheromone signalling appeared to be independent of MAT gene expression, suggesting that at least in A. nidulans the MAT genes are primarily required for later stages of sexual development (Paoletti et al., ).
Other signalling cascade components have been identified in A. nidulans in addition to the above pathway elements. A second GTP/GDP binding developmental switch has been reported involving a small G protein encoded by a ras gene homologue rasA (or A-ras). RasA levels modulate asexual sporulation, whilst overexpression of rasA results in an acleisothecial phenotype, suggesting that RasA has to be predominantly inactivated before fruit body formation can occur (Hoffmann et al., ; Braus et al., ). Meanwhile, Han et al. () identified an extra GPCR encoding gene termed gprD, deletion of which resulted in restricted vegetative growth and production of small colonies densely packed with cleistothecia. This implied that GprD-mediated signalling negatively regulates sexual development, and indeed nsdD transcripts were strongly elevated in a ΔgprD deletion strain. GprD is thought to act upstream of GprA/GprB (PreB/PreA) signalling (Seo et al., ). A further receptor GprK also appears to be essential for the maturation of cleistothecia. A ΔgprK deletion strain showed accumulation of Hülle cell aggregates but a block in further development of cleistothecia, and gprK transcripts were highly expressed during sexual development (J.H. Yu, pers. commun.). Also related to G protein activity, a novel Gβ-like RACK1 homologue termed GibB has been identified in A. nidulans with similarity to Gib2 in Cryptococcus neoformans (Palmer et al., ; Q. Kong & J.H. Yu, pers. commun.). GibB is thought to bind directly to Gpα1 as a Gβ-like protein to stabilize Gpα1 and facilitate its activation and inactivation cycle, and to regulate cAMP signalling. A gibB deletion mutant produces small colonies and only c. 2% of the number of cleistothecia compared to the wild type, and these cleistothecia contain very few (if any) ascospores. The cleistothecia of the gibB deletion mutant are also smaller in size and colourless with abnormal morphology compared to the wild type. This indicates that GibB plays an important role in sexual development of A. nidulans, governing a distinct stage of sexual fruiting (Q. Kong & J.H. Yu, pers. commun.). In addition, a homologue of the RIC-8 animal protein, which is crucial for GDP/GTP exchange in the absence of GPCRs or other guanine nucleotide exchange factors, has been identified in A. nidulans. A ricA deletion mutant was found to exhibit extremely restricted hyphal growth and failed to produce any sexual structures (either Hülle cell or cleistothecia) even under conditions favourable for sexual development. These results indicate that RicA functions upstream of the cascade controlling sexual development in A. nidulans (Fig. 0002) (N.J. Kwon & J.H. Yu, pers. commun.).
Finally, a putative MAP kinase gene sakA (hogA) has been identified as part of a likely transduction pathway sensing environmental osmotic and oxidative stress (Kawasaki et al., ). Given that ΔsakA mutants undergo precocious sexual development and increased production of cleistothecia, it is likely that SakA is a repressor of steA-dependent morphogenesis under environmentally unfavourable conditions for sex, but has minimal effect under conducive conditions (Kawasaki et al., ). SakA is thought to interact with a further transcription factor AtfA, with deletion of the atfA gene also resulting in derepressed sexual development with over a twofold increase in numbers of Hülle cells and cleistothecia (Lara-Rojas et al., ).
Genes encoding transcription factors and other regulatory proteins
In addition to the genes listed above, a diverse range of regulatory proteins, many with likely transcription factor activity, have been shown to influence sexual development in the aspergilli (Table 0002; Fig. 0002). Some of these were first identified owing to their influence on asexual development. For example, the genes stuA and medA with are best known for their roles in regulating asexual conidiation. They encode proteins with a basic helix-loop-helix (bHLH) region APSES domain and no clearly conserved domain, respectively, However, they are also required for correct sexual development, with medA and stuA mutants of A. nidulans failing to produce ascospores and cleistothecia, although medA mutants did produce unorganized masses of Hüllle cells (Busby et al., ; Wu & Miller, ). The devR gene encodes a second bHLH protein, which has also been most intensely characterized for its role in asexual conidiation, but is additionally essential for sexual development because ΔdevR strains entirely failed to form Hülle cells or cleistothecia (Tüncher et al., ). In parallel, the leucine zipper-like protein DopA is needed for normal asexual sporulation, but sexual reproduction was also completely abolished in a ΔdopA strain (Pascon & Miller, ). In a similar fashion, the WD40 family protein RcoA is required for both normal asexual development and sexuality, as illustrated by the finding that a ΔrcoA strain was completely self-sterile and unable to form maternal tissues in outcrosses (Todd et al., ). Overexpression of the veA gene in a ΔrcoA background failed to restore sexual development, indicating that RcoA acts downstream of VeA in development (Fig. 0002). Conversely, the zinc-finger transcription factor FlbC, which is required for asexual conidiation in A. nidulans, acts mainly as a repressor of sexual development as shown by the abundant formation of Hülle cells and cleistothecia in a ΔflbC deletant strain (Kwon et al., ). Similarly, the associated regulator of conidiation FlbE also represses sexual development as illustrated again by the increased formation of Hülle cells and cleistothecia, particularly in a veA1 background, in a ΔflbE deletant strain (Kwon et al., ). These latter two reports emphasize the competing nature of asexual and sexual reproductive pathways.
Meanwhile, a series of genes have been identified specifically due to their requirement for sexual reproduction. One with perhaps the greatest significance for the initial stages of sexual development is nsdD, so called as it arose from a screen of ‘never in sexual development’ (NSD) mutants (Han & Han, ). It encodes a zinc-finger GATA-type transcription factor, and ΔnsdD mutants of A. nidulans failed to produce cleistothecia or Hülle cells under conditions favourable for sex, whereas overexpression of nsdD led to a dramatic increase in cleistothecia and fruiting bodies were produced even on high-salt media that normally repressed sex (Han et al., ). Thus, NsdD was concluded to act as a positive activator of sex, possibly acting downstream of VeA (Han et al., ). A Y2H screen later revealed at least two proteins interacting with NsdD (Kwon et al., ). These interactor of NsdD proteins ‘IndB’ and ‘IndD’ exhibited possible VeA-dependent expression (repressed in a veA background, but highly induced in a ΔveA mutant), suggesting a possible role in the inhibition of NsdD function possibly due to binding to the zinc-finger region, thereby preventing DNA binding and transcription by NsdD. An nsdD homologue has also been identified in A. fumigatus, whose overexpression was found to decrease asexual sporulation and induce the production of coiled hyphal structures possibly representing ascogonia, although no further sexual development was observed (Groβe & Krappmann, ). A subsequent study determined that deletion of the A. fumigatus nsdD gene resulted in loss of crossing ability and also impaired ability to form heterokaryotic mycelium. Therefore, it was speculated that NsdD might have a role in regulating cell wall integrity, which could be critical in facilitating hyphal anastomosis during early mating (Szewczyk & Krappmann, ). These findings are consistent with nsdD mediating early stages of sexual development in homothallic and heterothallic aspergilli as a whole.
A second A. nidulans NSD-related gene, nsdC, has also been found to encode a zinc-finger DNA-binding protein, which features a novel fungal C2HC motif (Kim et al., ). As with NsdD, the NsdC protein also appears to play an important role as a positive regulator of the initial stages of sexual development and repressor of asexual sporulation. Deletion of nsdC resulted in the complete loss of fruit body formation and formation of Hülle cells, whilst overexpression overcame the inhibitory effects of certain stressors, allowing sexual development when cultures were grown under white light, on high-salt media or were subjected to other osmotic stresses (Kim et al., ). Possible links to VeA and NsdD activity were investigated, but it was found that overexpression of nsdC failed to restore sexual development in ΔveA and ΔnsdD strains and vice versa, indicating that all three genes are critical for sex and that they act independently of each other – there being no obvious change in expression levels of nsdC in the ΔveA and ΔnsdD strains.
Two further possibly paralogous genes encoding zinc-finger proteins have been identified from A. nidulans, which also influence sexual development. The first, nosA, is required for completion of sexual development, with ΔnosA mutants failing to develop beyond sexual primordia, although very occasional cleistothecia containing ascospores were formed (Vienken & Fischer, ). By contrast, the second, rosA, inhibits sexual development under certain conditions, with a ΔrosA mutant able to undergo sexual reproduction under low glucose and high osmolarity conditions that normally inhibit sex (Vienken et al., ). Further studies with various manipulations of the nsdD and nosA genes suggested that NosA most likely acts downstream of NsdD, or in a parallel pathway (Vienken & Fischer, ). This is consistent with the proposal of Dyer () that VeA, NsdD, RosA and now NsdC act early in development to control genes inducing sexual primordial (Fig. 0002), whereas NosA is required for the maturation of primordia that already exist, that is, downstream in the developmental pathway. Meanwhile, a forkhead protein transcription factor encoded by the fhpA gene was shown to be required for the production of cleistothecia in A. nidulans, although Hülle cell formation was possible in an fhpA disruptant strain (Lee et al., ).
An additional major regulatory contributor is the multiprotein COP9 ‘signalosome’ complex (CSN), which is involved with various developmental processes throughout the eukaryota (Braus et al., ). It consists of an eight-subunit complex in A. nidulans, and it appears to have contrasting effects on sexual development. Research efforts have focussed mainly on two subunits, CsnD and CsnE (Busch et al., , ; Nahlik et al., ). It was first found that a ΔcsnD gene deletion mutant was blocked in sexual development at the primordial stage (although Hülle cells were formed as normal), but that the mutant was simultaneously ‘blind’ and able to develop primordia even in the presence of light (Busch et al., ), that is, CsnD was required not only for correct sexual development but also for the suppression of sex in the light. Similarly, the subsequent characterization of a ΔcsnE gene deletion mutant revealed a block in sexual development at the primordial stage, but the mutant again demonstrated constitutive formation of primordia even in the light. This confirmed that CSN activity is required for sexual morphogenesis but is conversely involved with light suppression of sex, noting that the CsnE subunit appears to have a particular key role in the assembly of the overall CSN complex (Nahlik et al., ). Both of these effects were likely to be mediated partly via changes in levels of Psi factors regulating sporulation (see below) as increased ppoA gene expression and altered hormone balance were observed in the ΔcsnE mutant. Also, reduced expression of cell wall–degrading enzymes was seen in the ΔcsnE strain, with the suggestion that normal CSN activity is required for the release of materials needed for wall remodelling and development of mature fruit bodies (Nahlik et al., ). Most of the CSN effects are assumed to arise from the CsnE deneddylase activity linked to protein turnover (Busch et al., ), and it has been speculated that the CSN might be involved with VeA degradation in the light (Braus et al., ; Nahlik et al., ). Overexpression of veA in a ΔcsnD background failed to overcome the sexual developmental block (Busch et al., ). Other CSN subunits have also been investigated. A UV mutagenesis screen yielded a sterile, aconidial mutant acoB202 (Butnick et al., ), and later complementation and gene rescue work identified a mutation in the acoB gene as responsible for the failure to reproduce sexually (Lewis & Champe, ). This gene was subsequently found to be part of the CSN and was co-named csnG (Busch et al., ). Meanwhile, it was found that deletion of CsnA and CsnB also led to identical blocks in development at the primordia stage as seen in the ΔcsnD and ΔcsnE mutants, illustrating that formation of the whole CSN complex is required for correct sexual development (Busch et al., ). In parallel to the CSN investigations, a further protein Cand1 has recently been described from A. nidulans, which also appears to be involved with protein neddylation. Deletion of N- and C-terminal components, encoded by genes candA-N and candA-C, respectively, resulted in a block in sexual development at the initial stage of early nest formation (although Hülle cells were formed), suggesting related developmental functions between Cand1 and the CSN (Helmstaedt et al., ).
Finally, two transcription factors involved in the formation of sclerotia have been described. First, a bHLH transcription factor SclR (for sclerotium regulator) has been identified from A. oryzae, and deletion of the gene found to result in sparse sclerotial production whilst overexpression of sclR led to greater than fivefold production of sclerotia with increased formation of branched aerial hyphae (Jin et al., , ). Second, deletion of a zinc-finger calcineurin-response gene crzA was found to result in the production of mainly immature sclerotia in A. parasiticus, that is, a delay in development (Chang, ).
Genes linked to endogenous physiological processes
Various physiological changes occur during the switch from asexual to sexual growth in the aspergilli and the subsequent development of cleistothecia and accessory tissues. Some of these changes have been well documented at a biochemical level, and over the past decade, some of the genes involved in these processes have been identified. For example, a chemically related group of oxylipin hormones termed ‘Psi’ factors (for precocious sexual inducers), derived from linoleic acid and related to mammalian prostaglandin hormones, has been identified. These include the members psiBα [8-hydroxy-(9,12)-octadecadienoic acid (8-HODE)] and psiCα (5,8-diHODE), which can trigger a switch from asexual to sexual growth, or psiAα (similar structure to psiCα but with a lactone ring replacing the hydroxyl group at the 5′ position), which can repress sex and induce asexual reproduction (Champe et al., ; Tsitsigiannis et al., ). And three oxylipin biosynthetic genes, ppoA, ppoB and ppoC (mnemonic for psi factor–producing oxygenase), encoding linoleate diol synthases were subsequently identified from A. nidulans (Tsitsigiannis et al., , ; Brodhun & Feussner, ). PpoA has been shown to catalyse the formation of various 8′ oxylipins including psiBα and pisCα (Tsitsigiannis et al., ; Brodhun & Feussner, ). PpoC is mainly involved in the formation of 10′ products (e.g. 10-HPODE) and appears not to form Psi factors directly, but has an indirect role in the production pathway possibly leading to psiA formation (Brodhun & Feussner, ). Thus, ppoA and ppoC have antagonistic roles. Sexual and asexual development are characterized by different ratios between 8-hydroxy-(9)-octadecanoic acid and psiB, the ratio shifting from 1 : 8 in darkness (sexual) to 1 : 3 in light (asexual) (Bayram & Braus, ). The biochemical properties of PpoB remain unclear (Brodhun & Feussner, ). Consistent with these findings, deletion of ppoA resulted in a major reduction in ascospore production, but an increase in asexual sporulation, whilst ppoA overexpression led to increased sexual reproduction (Tsitsigiannis et al., ). Interestingly, virtually no ppoA expression was detected in a ΔveA strain and increased expression was seen in csnD mutant strains, suggesting that VeA and CsnD may regulate ppoA expression (Fig. 0002) (Tsitsigiannis et al., ; Dyer, ). Deletion of ppoB resulted in decreased sex linked to downregulation of ppoA but upregulation of ppoC, that is, suggesting a regulatory loop between the three genes (Fig. 0002). By contrast, deletion of ppoC led to an increase in production of cleistothecia, demonstrating the opposing activity of ppoB and ppoC (Tsitsigiannis et al., ; Brown et al., ). Surprisingly, a triple ΔppoAΔppoBΔppoC mutant exhibited increased activation of sexual development with the production of Hülle cells and even some cleistothecia under water, correlated with greatly increased VeA expression (Tsitsigiannis et al., ). Thus, although oxylipins have an important role in the induction of sex, other parameters can independently trigger morphogenesis (Dyer, ). In the context of asexual and sexual reproduction in A. nidulans, it is intriguing to speculate whether the fluG gene, suggested to regulate asexual conidiation via production of a mysterious extracellular developmental signal (Lee & Adams, ), might mediate its effect partly through an influence on psiA synthesis.
Ppo function has also been investigated in A. flavus and A. fumigatus. In A. flavus, homologues to the A. nidulans ppoA, ppoB and ppoC genes were discovered by genome blast analysis, together with an additional ppoD dioxygenase homologue (Brown et al., ). In A. flavus, oxylipin products appear to play a vital role in mediating spore concentration–dependent quorum sensing, which influences production of sclerotia (as described above). Deletion of an Aflox gene, encoding a lipoxygenase (LOX), was found to minimize the density-dependent inhibition of sclerotial production at high spore concentrations, suggesting that a LOX-derived metabolite is responsible for the activation of asexual sporulation and the simultaneous repression of sex under these conditions (Brown et al., ). Downregulation of the four genes ppoA, ppoB, ppoC and Aflox in a ΔppoD background resulted in a similar outcome, with a dramatic increase in sclerotial production (up to 500-fold) at high spore concentrations (Brown et al., ). It was suggested that ppoC and Aflox were particularly inhibitory to sclerotial formation, whilst ppoD promoted sclerotial production, possibly via production of an opposing oxylipin that functioned at low spore concentration. As noted above, multicopy expression of laeA was able to abolish this quorum-like phenomenon (Amaike & Keller, ). In the case of A. fumigatus, deletion of only ppoC led to a distinct phenotype, with altered condium size, germination and stress tolerance (Dagenais et al., ).
Another physiological signal for early sexual morphogenesis appears to be the cellular oxidation state. Deletion of the noxA gene in A. nidulans, which encodes an NADPH oxidase able to produce reactive oxygen species, resulted in a block in cleistiothecial development at the initial stage (although Hülle cells were present) (Lara-Ortíz et al., ). It was proposed that noxA might contribute to a hyperoxidant state that triggers cell proliferation, apoptosis and cell wall developmental changes during early sexual morphogenesis. NoxA expression was independent of SteA and StuA activity but suppressed by SakA (Fig. 0002). A related catalase–peroxidase gene cpeA has also been identified that possibly offers protection against oxidative damage (Scherer et al., ). Linked to these observations, two protein components TrxA and TrxR forming a classic cytoplasmic thioredoxin system have been identified in A. nidulans. Deletion of trxA resulted in failure to form any sexual structures, although this could be alleviated by addition of low levels of reduced glutathione, again indicating that cellular redox state has a vital role in allowing sexual development to occur (Thön et al., ). Possibly also linked to cellular redox state, Eisendle et al. () found that deletion of the sidC gene, encoding a nonribosomal peptide synthetase required for production of an intracellular siderophore, led to impaired asexual and sexual development even in iron-replete conditions. Cleistothecia failed to develop even though Hülle cells were formed, suggesting that iron homoeostasis is vital for cellular developmental processes. Also linked to cellular redox state, a homologue of the yeast MSN2 transcription factor involved in stress response has been identified in the aspergilli. The NrdA (originally termed MsnA but renamed for being a negative regulator of differentiation) protein appears to play a role in the maintenance of cellular oxidative state, although with contrasting roles in different aspergilli. Deletion of nrdA led to increased production of cleistothecia in A. nidulans, whereas deletion of the msnA homologue led to loss of sclerotial production in A. flavus and A. parasiticus, so subtle differences are apparent (Jeon et al., ; Chang et al., ).
Sexual morphogenesis in A. nidulans has also been shown to be correlated with increased α-glucanase activity, possibly linked to energy release for the development of cleistothecia (Zonneveld, ). Genes encoding an α-1,3 glucanase/mutanase (mutA) and high-affinity hexose transporter (hxtA), which both show elevated levels of expression during sexual development, have subsequently been identified (Wei et al., , ). However, no obvious inhibition of sexual development was seen in mutA and hxtA deletion strains, suggesting redundancy in carbon usage and uptake systems.
Later developmental stages – ascospore production
Certain genes appear to be required for the later stages of the sexual cycle in A. nidulans, being linked specifically to the production and maturation of ascospores. For example, a deletion mutant of the tubB gene, encoding an alpha-tubulin protein, exhibited normal development and formed asci, but these contained only a single nuclear mass consistent with a block in karyogamy and/or meosis following zygote formation (Kirk & Morris, ). Similarly, a deletion mutant of the grrA gene, encoding a fungal F-box protein (likely involved with protein ubiquitinylation and degradation/recycling), was able to form Hülle cells and cleistothecia, but was blocked at the point of meiosis and ascospore formation with only empty asci observed. This particular F-box protein appears to be specifically required for ascosporogenesis, as evidenced by upregulation of its transcript during cleistothecial development (Krappmann et al., ). Deletion of the samB gene, encoding a zinc-finger protein associated with nuclear positioning, also resulted in failure to produce viable ascospores and led to spore lysis (Kuger & Fischer, ). And as mentioned above, the vosA gene is required for the production of trehalose in sexual ascospores, as illustrated by the fact that a ΔvosA mutant strain produced defective, small cleistothecia containing only very few viable semi-transparent ascospores (Ni & Yu, ).
A further gene with a role in ascosporogenesis is the striatin homologue strA, which encodes a putative multidomain scaffolding protein with a possible role in calcium signalling in the endomembrane system (Wang et al., ). Deletion of strA resulted in the production of abnormally small cleistothecia with various defects in ascospore formation, for example failure to develop asci, failure of ascospores to separate correctly in asci, and production of abnormally shaped ascospores. However, in a VeA (although not VeA1) background, the majority of ascospores developed normally and were viable, a good example of a ‘leaky’ mutation ensuring ascospore production as a possible survival mechanism despite mutational damage (Dyer, ). Overexpression of strA resulted in a c. twofold increase in cleistothecial production under sex-inducing conditions, increased production of protocleistothecia under repressive high-salt conditions, and the formation of Hüllle cells even in shaking liquid culture (Wang et al., ). This illustrates the positive role of StrA in sexual development and shows that it promotes other parts of the sexual cycle in addition to ascospore formation (Fig. 0002).
A series of other mutants of A. nidulans have been produced by UV mutagenesis with various defects in recombination ability and ascospore development, for example no croziers, arrest at prekaryogamy or meiotic prophase or metaphase, nondelineation of ascospores, and colourless or blue ascospores. However, the genes involved are as yet unknown although some have been localized to certain chromosomes (Apirion, ; Osman et al., ; Swart et al., ). It has been estimated that 50–100 genes might be involved specifically in ascospore formation based on complementation analysis (Swart et al., ) with the term mei proposed for certain of these sexual sporulation mutants (Bruggeman et al., ). It is noted that blast searching has detected at least 100 genes in A. nidulans with homologues to genes in budding and fission yeasts known to be involved in karyogamy and meiosis (Galagan et al., ).
Insights into cryptic sexuality in the aspergilli
In the final section of this review, recent findings concerning sexual development in the once-presumed ‘asexual’ species A. fumigatus, A. flavus, A. parasiticus and A. nominus will be described and implications for cryptic sexuality in the aspergilli discussed.
The majority of Aspergillus species are only known to reproduce by asexual means. As discussed in the Introduction, this is very surprising given the many perceived benefits of reproducing by both sexual and asexual means, as opposed to obligate asexuality. However, there is now accumulating evidence that many supposedly ‘asexual’ Aspergillus species might in fact have the potential to reproduce by sexual means (Horn et al., , b, c, ; O'Gorman et al., ) and thus exhibit ‘cryptic sexuality’ or ‘covert’ sexual reproduction, with a so far hidden sexual state that has yet to be discovered (Kück & Pöggeler, ; Heitman, ). This is highly significant given that many of the asexual aspergilli are of either medical or industrial importance. Thus, the presence of a sexual cycle would impact on the population biology of a species, perhaps enabling more rapid evolution of antifungal resistance in pathogens, whilst simultaneously providing a valuable genetic tool for classical genetic analyses and for strain improvement by industry.
Evidence for cryptic/covert sexuality has come from data drawn from diverse research areas such as population biology analyses, genome analysis of sex-related genes, the distribution of complementary mating types in nature, expression of sex-related genes and classical laboratory mating studies (Dyer & Paoletti, ; Heitman, ). Such studies have had a great impact, leading to the discovery of sexual states in the following Aspergillus species, which were once considered purely asexual.
Aspergillus fumigatus is one of the most ubiquitous fungi worldwide, colonizing diverse habitats but being typically found in soil and organic debris such as compost. It is well known for the prolific production of asexual conidia (Samson et al., ). Inhalation of these spores by immunocompetent individuals rarely has any adverse effect because of efficient elimination by the innate immune system (Latgé, ). However, since the early 1990s, A. fumigatus has become the most prevalent airborne fungal pathogen worldwide. This has been due to the increase in numbers of immunocompromised patients, whose weakened immune systems allow it to cause opportunistic infections (Latgé, ; Dagenais & Keller, ). Groups at risk include those undergoing chemotherapeutic regimes for bone marrow and organ transplantation and patients with HIV/AIDS. Even with treatment, mortality rates of 50–80% are reported (Lin et al., ; Richardson & Lass-Flörl, ). Manifestations of disease range from saprophytic growth in pre-existing cavities (aspergillomas) to life-threatening invasive aspergillosis. In addition, the airborne conidia of A. fumigatus can trigger hypersensitivity diseases including asthma, allergic sinusitis and allergic bronchopulmonary aspergillosis (Latgé, ; Dagenais & Keller, ). The species also causes aspergillosis in other mammals and birds (Tell, ).
The importance of A. fumigatus to human health led to a clinical isolate (Af293) being the first Aspergillus genome to be publically sequenced. blast analysis revealed that, despite its asexual status, the Af293 genome possessed a complete set of known genes linked to sexual reproduction, 215 in total, that is, a full suite of ‘sexual machinery’ (Galagan et al., ). Importantly, all of these sex-related genes appeared functional, with no mutations within any of their coding regions. The same suite of genes was found when a second A. fumigatus isolate (A1163) was sequenced, providing further evidence for a species-wide sexual potential (Fedorova et al., ). Characteristic conserved α-domain (MAT1-1) and HMG domain (MAT1-2) genes were present at a MAT locus in the A1163 and Af293 isolates, respectively (Pöggeler, ; Galagan et al., ; Fedorova et al., ), consistent with experimental cloning of a MAT1-1 idiomorph (Paoletti et al., ), in the typical arrangement for heterothallic (outcrossing) species. When a worldwide collection of 290 clinical and environmental isolates of A. fumigatus was screened, MAT1-1 and MAT1-2 mating types were found in near-equal proportion, suggesting that sexual reproduction was still occurring frequently enough to maintain the balance of mating types (Paoletti et al., ). In addition, the pheromone precursor (ppgA), pheromone receptor (preA and preB) and mating-type genes (MAT1-1 and MAT1-2) were all shown to be expressed during normal vegetative growth (Paoletti et al., ).
This information together with evidence of gene recombination from population genetic studies (Varga & Tóth, ; Paoletti et al., ; Pringle et al., ; Bain et al., ) and the presence of close taxonomic relatives with sexual states (Samson et al., ) suggested the presence of cryptic or ‘clandestine’ sexuality in A. fumigatus, which was suggested to be ‘holding back the truth about its sexuality’ (Gow, ). Critically, the necessary conditions for a sexual cycle remained unclear. However, O'Gorman et al. () then made a major breakthrough when they were able to induce a fully functional sexual cycle in vitro and show that it leads to the production of recombinant ascospore progeny; the teleomorph was named N. fumigata (Fig. 0003). A key underpinning to this work was the availability of known MAT1-1 and MAT1-2 isolates, as revealed by the PCR diagnostic of Paoletti et al. (), meaning that directed crosses could be set up between isolates with potential sexual compatibility. This was previously not possible, and therefore crossing efforts might otherwise have been confounded because of the abortive presence of isolates of the same mating type. The required conditions for sexual reproduction were growth on oatmeal agar at 30 °C in the dark with restricted aeration for up to 6 months (O'Gorman et al., ). Significantly, the authors have since successfully mated combinations of both clinical and environmental isolates from a wide variety of locations other than the Irish isolates used in their original study, confirming that fertility is neither limited to the environmental strains nor is geographically restricted to Ireland (C.M. O'Gorman, S. Swilaiman, J. Sugui, K.J. Kwon-Chung & P.S. Dyer, unpublished results). Szewczyk & Krappmann () also recently reported the successful crossing of two unrelated clinical strains of A. fumigatus, one of which was the genome-sequenced isolate Af293 (Galagan et al., ).
The discovery of a sexual cycle in A. fumigatus was highly significant as it helped to explain many puzzling aspects of its biology and evolution. These ‘puzzles’ included the presence of numerous different genotypes within populations and the successful defence of the genome against repetitive elements (Latgé, ). Using classical genetic analyses, it will now be possible to gauge the contribution of meiotic recombination to the ongoing evolution of A. fumigatus.
Aspergillus flavus and Aspergillus parasiticus
Aspergillus flavus and A. parasiticus are two very closely related members of Aspergillus section Flavi. Morphologically, the two species are near identical, but they can be easily separated by mycotoxin profiling, AFLP fingerprinting and DNA sequencing (Horn et al., ; Montiel et al., ; Peterson, ). They are the major worldwide producers of aflatoxins, one of the most potent groups of natural toxins known to man with carcinogenic and acute immunosuppressive and neurotoxic effects (John, ). Aflatoxins contaminate a wide variety of agricultural crops including cereals, oilseeds and nuts, and over 100 countries have strict limits on the amount of aflatoxin contamination allowable in these foods (Van Edmond & Jonker, ). In addition to being mycotoxigenic, A. flavus is also an opportunistic human pathogen, being the second most common agent of aspergillosis after A. fumigatus (Krishnan et al., ).
Both species are characterized by very high levels of genetic diversity, as seen by the large number of vegetative compatibility groups (VCG) and DNA fingerprint types that have been found (Horn et al., ; Horn, ). This considerable diversity could not be attributed to asexual reproduction alone, suggesting that either meiotic or parasexual recombination was responsible. Indeed, related population genetic studies have revealed that A. flavus and A. parasiticus both have recombining population structures (Geiser et al., ; Carbone et al., ). A sample population of A. flavus from Australia showed evidence of recombination, with two reproductively isolated clades found (Geiser et al., ), whilst analysis of a field collection of A. parasiticus from Georgia in the USA found evidence of both a history of recombination and contemporary recombination from phylogenetic network analysis of the aflatoxin gene cluster (Carbone et al., ). Genome analysis also revealed that both species have an idiomorphic MAT locus arrangement characteristic of the heterothallic Pezizomycotina, with a single biallelic MAT locus (Ramirez-Prado et al., ). In addition, the MAT genes were shown to be expressed in the two species, and there were near-equal numbers of isolates of each mating type present in close proximity in the Georgia field populations. Thus, both species appeared to exhibit cryptic sexuality. This was re-enforced by genome screening that has revealed a full complement of apparently functional genes required for sexual reproduction in A. flavus (C.E. Eagle & P.S. Dyer, unpublished results).
Following the discovery of sexual reproduction in A. fumigatus (O'Gorman et al., ), similar breakthroughs were then published shortly afterwards concerning A. parasiticus and A. flavus. The sexual cycles of both species were induced using the same method: strains of opposite mating type were paired on mixed cereal agar in darkness at 30 °C in sealed plastic bags for 6–9 months (A. parasiticus) or 6–11 months (A. flavus) (Horn et al., , c). Hardened sclerotia were produced during this period, with the sexual ascocarps found to develop within their matrices. Sclerotial production had previously been observed in A. parasiticus and A. flavus and was postulated to function as a survival mechanism for the fungi during adverse environmental conditions. However, here they acted as a repository for the cleistothecia. It was previously hypothesized that the sclerotia of A. flavus would be capable of harbouring fruiting bodies given the right conditions (Rai et al., ; Geiser et al., ). The two sexual stages were assigned to the genus Petromyces, the teleomorphic genus associated with Aspergillus section Flavi (Horn et al., , b, c). Previously, the homothallic Aspergillus alliaceus was the only other known sexual species within this section (Raper & Fennell, ). Morphologically, the ascospores of the three Petromyces species are extremely similar; Petromyces flavus and Petromyces parasiticus are near indistinguishable, whilst P. alliaceus differs by having slightly smaller ascospores that are smooth instead of finely roughened (Horn et al., , b, c). Significantly, many species in Aspergillus section Flavi produce sclerotia, suggesting that sexual reproduction may also be possible in these species.
It is not surprising that the crossing conditions for A. flavus and A. parasiticus are very similar to the conditions required for A. fumigatus mating as they are all soil saprotrophs (Klich, ). Ascospores are typically highly resistant structures, with a thick cell wall that confers thermotolerance (Baggerman & Samson, ; Dijksterhuis, ). Their long incubation periods for fruit body development may have evolved as a mechanism to outlast unfavourable environmental conditions. As most other known sexual Aspergillus species complete their sexual cycles within several weeks, this may explain why sexual reproduction in these three species remained unobserved for so long.
Aspergillus nomius and beyond
Most recently, a sexual cycle has also been induced in the previously ‘asexual’ aflatoxin producing species A. nomius (Horn et al., ). This is also a member of the section Flavi, and similar protocols were used to induce the sexual stage as were used for A. parasiticus and A. flavus, that is, identification of compatible mating types using a PCR diagnostic, and then crossing used a mixed spore inoculum on mixed cereal agar. As seen with A. parasiticus and A. flavus, the cleistothecia of A. nomius were formed within sclerotia over a long time period of 5–10 months. The discovery of the teleomorph was consistent with earlier findings of cryptic sexual recombination based on patterns of gene polymorphisms (Peterson et al., ). However, A. nomius differed from the other species in that overall fertility of the field isolates studied was markedly lower than A. parasiticus and A. flavus, ascospores only being produced in 24% of crosses. Even in such fertile crosses, in most cases, only 1–5% of sclerotia contained asci with ascospores (Horn et al., ). A similar lack of sexual fertility has been described in various other pezizomycete species (Dyer et al., ; Dyer & Paoletti, ). A variety of factors might contribute to this apparent low fertility. For example, the crossing conditions used in the laboratory may not have been optimal, leading to decreased fertility. Alternatively, differences in VCG between isolates may have influenced mating success (Dyer et al., ). In the case of A. nomius, the VCGs of the paired isolates were unknown, unlike the situation for the A. flavus and A. parasiticus crosses (Horn & Greene, ; Horn et al., ). Finally, it has been suggested that an overall ‘slow decline’ in sexual reproduction and fertility within populations may be linked to evolution towards asexuality, because of factors such as decreased or suppressed expression of sex-related genes (Dyer & Paoletti, ).
Aspergillus nomius also exhibited one other unusual feature in that whilst most field isolates were of either MAT1-1 or MAT1-2 genotype, some occasional isolates were of joint MAT1-1/MAT1-2 genotype. These were not self-fertile, but were able to outcross to both MAT1-1 and MAT1-2 isolates. Such unusual ‘dual maters’ have been reported elsewhere in the Pezizomycotina albeit very rarely (Debuchy et al., ; Amselem et al., ). The exact genetic organisation of this genotype remains to be determined.
From the above, it can be seen that discoveries concerning the sexual biology of the aspergilli have made a major contribution to the understanding of fungal sex in general. Numerous genes involved in the initiation and co-ordination of sexual reproduction have already been described from the aspergilli (Table 0002), and many parts of the developmental pathway proposed in Fig. 0002 might be applicable to the Pezizomycotina as a whole. The availability of considerable genome resources and an ongoing project to systematically create a series of knockout constructs for all genes of A. nidulans (http://www.fgsc.net/Aspergillus/ko_cassettes/cassetteannouncement.htm) provides wonderful opportunities to identify further aspects of the sexual circuitry of the group. Indeed, it has been estimated that up to 2000 genes might be involved in developmental and differentiation processes (Braus et al., ), so many more genes remain to be characterized. For example, classical genetic studies have identified a series of mutants ranging from acleistothecial to dense cleistothecial forms (sgp, aco, acl, lcl, dcl) whose molecular genetic basis remains unknown (Houghton, ; Zonneveld, ; Scherer & Fischer, ). Genomic technologies such as EST and microarray analysis and whole transcriptome shotgun sequencing (‘RNA seq’) offer further opportunities to identify genes differentially regulated during sexual morphogenesis. Already some results are available from pilot studies of cleistothecial development in A. nidulans (Jeong et al., , ) and sclerotial development in A. flavus (Cary et al., ). In addition, there is promise that bioinformatic analyses will yield novel insights into development. For example, a reciprocal smallest distance analysis was recently used to identify 245 orthologous genes shared between the sclerotial-forming species Botrytis cinerea, Sclerotinia sclerotiorum, A. flavus and A. oryzae, but absent from the nonsclerotial-producing A. nidulans and A. fumigatus; the characterization of such genes offers a potential foothold into a broader understanding of sclerotial development and evolution (Amselem et al., ).
Meanwhile a ‘sexual revolution’ is ongoing in the supposedly asexual aspergilli (Horn et al., , c, ; O'Gorman et al., ). The discoveries of sexual cycles in more ‘asexual’ species will further serve to validate the joining of mitosporic fungi with their meiosporic relatives and perhaps would only come as a surprise to the ‘laboratory voyeur’ rather than a population geneticist (Gow, ). Attempts are currently underway to induce the sexual stages of various other asexual aspergilli of industrial and medical importance (S. Swilaiman, C.M. O'Gorman, H. Darbyshir & P.S. Dyer, unpublished data). It is cautioned that the sexual cycles of some species might still elude discovery because of the fastidious nature of their sexual demands (Kwon-Chung & Sugui, ) or genuine evolution to asexuality. Also, species should arguably not be viewed as either sexual or asexual but rather as being composed of isolates present on a continuum of sexual fertility (Dyer & Paoletti, ). However, the availability of MAT diagnostic tools now allows targeted crosses to be set up between compatible MAT1-1 and MAT1-2 isolates of asexual species offering the exciting prospect of further sexual revelations in the near future.
We thank the Wellcome Trust (UK) for research funding, and Jae-Hyuk Yu and Ana Calvo for sharing unpublished results.