Fungal Diversity

, 50:121

The evolution of species concepts and species recognition criteria in plant pathogenic fungi


    • State Key Laboratory of MycologyInstitute of Microbiology, Chinese Academy of Sciences
  • Tatiana Giraud
    • Ecologie, Systématique et Evolution
  • Ning Zhang
    • Department of Plant Biology and Pathology, RutgersThe State University of New Jersey
  • Dominik Begerow
    • Ruhr-Universität Bochum
  • Guohong Cai
    • Department of Plant Biology and Pathology, RutgersThe State University of New Jersey
  • Roger G. Shivas
    • Plant Pathology HerbariumAgri-Science Queensland

DOI: 10.1007/s13225-011-0127-8

Cite this article as:
Cai, L., Giraud, T., Zhang, N. et al. Fungal Diversity (2011) 50: 121. doi:10.1007/s13225-011-0127-8


In this paper, we review historical and contemporary species concepts and species recognition criteria for plant pathogenic fungi. Previous incongruent and unstable classification based on subjective and changing criteria have led to some confusion, especially amongst plant pathologists. The goal of systematics is to provide an informative and robust framework that stands the test of time. The taxonomic histories of Cercospora, Colletotrichum, Fusarium, as well as the rust and smut fungi, are used as examples, to show how concepts and criteria used to delimit and recognize species have changed. Through these examples we compare the Genealogical Concordance Phylogenetic Species Recognition, an extension of the Phylogenetic Species Criterion, with other species recognition criteria and show that it provides a better discrimination for delimiting species. A rapidly increasing number of cryptic species are being discovered amongst plant pathogenic fungi using the Genealogical Concordance Phylogenetic Species Recognition, and it is important to determine their host range, the severity of diseases they cause and their biosecurity significance. With rapidly expanding global trade it has become imperative that we develop effective and reliable protocols to detect these previously unrecognized pathogens.


Cryptic speciesSpecies complexMicrobotryumPucciniomycotinaUstilaginomycotinaSpeciationTaxonomy


Plant pathologists are regularly confronted with having to choose a name for their pathogen of interest and mycologists often need to decide when to recognize a new species or apply an existing name. Country specific inventories of plant pathogenic fungi with accurate and accepted names are essential for the development of effective biosecurity and trade policies as well as a prerequisite for pest risk assessments (Hyde et al. 2010). These inventories also facilitate the early identification of invasive fungal pathogens and allow the timely application of appropriate disease control measures (Rossman and Palm-Hernández 2008). The accurate identification of a plant pathogen will in most cases provide a species name, which may then be used to unlock all of our collective knowledge about the organism. This knowledge may include its evolutionary history, life cycle, distribution, host range, resistance to drugs, economic and biosecurity importance as well as control measures (Rossman and Palm-Hernández 2008). Over time, our understanding about how to identify plant pathogenic fungal species has undergone several revolutionary changes.

The criteria used to delimit and identify species, as applied to plant pathogenic fungi, have changed over time, most recently due to the rapid development of molecular tools. The different criteria that allow the delimitation of species may be classified as morphological, physiological, intersterility, host specificity, and phylogenetic. All of these species recognition criteria attempt to identify evolutionary independent lineages (Taylor et al. 2000). The morphological and phylogenetic criteria can further be used to unravel evolutionary relationships between species and arrive at a natural classification. The classification of plant pathogenic fungal species, together with the associated taxonomic nomenclature, as currently defined by the International Code of Botanical Nomenclature, is fundamentally important for plant pathologists and mycologists in all fields (Rossman and Palm-Hernández 2008; Hyde et al. 2010).

The choice and justification of species criteria to identify the 1.5 million fungal species estimated to populate the world (Hawksworth 1991) or the ca 270,000 tropical plant pathogenic fungi (Shivas and Hyde 1997), has significant consequences for our understanding of emergent diseases on plants and animals (Giraud et al. 2010), particularly against a backdrop of global climate change (Chakraborty and Newton 2011).

In this paper, we review the evolution of species concepts and species recognition criteria in plant pathogenic fungi, by using examples from some important groups, namely, Cercospora, Colletotrichum, Fusarium, and the rust and smut fungi. The taxonomic history of each group is reviewed, with emphasis on the changing focus of criteria used to recognize species. In particular, the utility of Genealogical Concordance Phylogenetic Species Recognition in many fungal groups is compared with the other species criteria. The practical implications of changing criteria used to recognise species are discussed. We also discuss the consequences that recent advances in our understanding of fungal speciation have meant for developing robust species criteria, although more extensive reviews on speciation and species recognition in fungi can be found elsewhere (Giraud et al. 2008a; Kohn 2005; Taylor et al. 2000).

Species concepts versus species criteria

The apparent diversity of concepts as to what constitutes a species (De Queiroz 2007; Hey 2006) may lead one to think that there is no general agreement amongst biologists about what defines a species. This view stems from confusion between the concept of a species, i.e. a description of the kind of entity that constitutes a species, and the criteria that delimit a particular species, i.e. practical standards for the recognizing whether individuals should be considered members of the same species. Many so-called “species concepts” actually correspond to species criteria (De Queiroz 2007; Hey 2006; Taylor et al. 2000). The Biological Species “Concept” for instance is most often meant to emphasize the criterion of intersterility, the Morphological Species “Concept” emphasizes the criterion of morphological divergence, the Ecological Species “Concept” emphasizes adaptation to a particular ecological niche, and the Phylogenetic Species “Concept” emphasizes nucleotide divergence between monophyletic lineages (Giraud et al. 2008a; Taylor et al. 2000). These species criteria correspond to the different events that occurred during lineage separation and divergence, rather than to fundamental differences in what represents a species. To the contrary, it has been argued that most modern biologists agree on a common “species concept” or “species definition”, specifically segments of evolutionary lineages that have evolved independently from one another (de Queiroz 1998).

Why are there conflicts over which species criteria we adopt? There are three main reasons why species criteria cannot be universal. Firstly, speciation is a temporally extended process, and one that varies considerably in pace for different types of organisms. Secondly, several modes of speciation can occur, during which the phenomena used for species recognition do not necessarily appear in the same chronological order (Fig. 1). Thirdly, the characteristics of certain organisms render some species criteria difficult to apply (Giraud et al. 2008a). The most useful criteria for the recognition of species in nature will depend on the type of organism, its history of speciation and the degree of achieved divergence. Searching for a single species criterion applicable in all cases is fundamentally impossible (Giraud et al. 2008a).
Fig. 1

Schematic divergence of two species, in two hypothetical cases of respectively allopatric and sympatric speciation, with the progressive appearence of various criteria traditionnaly used to recognize species

The most commonly used species criterion for fungi has long been the Morphological Species Criterion. Recently many cryptic species have been recognised using the intersterility criterion, a derivative of the Biological Species Criteria (Anderson and Ullrich 1978). Mayr (1963) defined biological species as groups of actually or potentially interbreeding natural populations which are reproductively isolated from other such groups. A weakness of the intersterility criterion is that it cannot be applied to homothallic or asexual fungi (Reynolds 1993; Taylor et al. 2000).

The Phylogenetic Species Criterion has been responsible for a surge in the number of cryptic species recognized in recent years (Schubert et al. 2007; Damm et al. 2009; Wulandari et al. 2009; Aveskamp et al. 2010; Summerell et al. 2010). The Phylogenetic Species Criterion relies on phylogenetic analysis of variable characters, usually DNA sequences of selected genes or genomes. Phylogenetic Species Criterion was originally defined as the smallest monophyletic clade of organisms that share a derived character state (Cracraft 1983). A weakness with this approach is that single gene analyses, as compared to whole genome analyses, are dependent on the genes having an evolutionary history that reflects that of the entire fungus, which is often not the case (Aguileta et al. 2008). Taylor et al. (2000) further developed a Genealogical Concordance Phylogenetic Species Recognition, as an objective way to define the limits of sexual species. The Genealogical Concordance Phylogenetic Species Recognition uses the phylogenetic concordance of multiple unlinked genes to indicate a lack of genetic exchange and thus evolutionary independence of lineages. Species have been identified with Genealogical Concordance Phylogenetic Species Recognition that cannot otherwise be recognized due to the lack of distinguishing morphological characters or incomplete intersterility. The Genealogical Concordance Phylogenetic Species Recognition criterion has proved immensely useful in fungi, because it is more finely discriminating than the other criteria in many cases, or more convenient, e.g. with species that are unable to be crossed (Reynolds 1993; Taylor et al. 2000). Genealogical Concordance Phylogenetic Species Recognition is currently more widely used for fungi than any other organisms, because fungi often have a simpler morphology and it is difficult to demonstrate in vitro crosses for many fungi (Dettman et al. 2003a; Fournier et al. 2005; Johnson et al. 2005; Koufopanou et al. 2001; Le Gac et al. 2007a; Pringle et al. 2005; Prihastuti et al. 2009; Glienke et al. 2011).

There are several reasons why the Genealogical Concordance Phylogenetic Species Recognition is better at revealing cryptic species than the Biological Species Criterion (intersterility criterion). Firstly, intersterility often evolves slowly in allopatric divergences, in particular the prezygotic barriers most often tested in fungi (Coyne and Orr 1997; Le Gac and Giraud 2008). The divergence of DNA sequences used under the Genealogical Concordance Phylogenetic Species Recognition criterion may then occur before intersterility has evolved and thus be more useful to distinguish closely related sibling species. Among the numerous complexes of sibling species recently uncovered using the Genealogical Concordance Phylogenetic Species Recognition criterion, many in fact appear consistent with allopatric divergence, because the cryptic species occupy non-overlapping areas separated by geographic barriers (Taylor et al. 2006). This is the case for the species complexes of the model organism Neurospora crassa (Dettman et al., 2003a, 2003b), the yeast Saccharomyces paradoxus, (Kuehne et al., 2007), the plant pathogen Fusarium graminearum (O’Donnell et al., 2004), and the mushrooms Schizophyllum commune (James et al. 1999) and Armillaria mellea (Anderson et al. 1980; Anderson et al. 1989).

Even in cases of sympatric speciation, certain mechanisms of reproductive isolation may allow intersterility to evolve much later than the divergence of DNA, again rendering the Genealogical Concordance Phylogenetic Species Recognition more finely discriminating than the Biological Species Criterion (Giraud et al. 2010; Giraud et al. 2008a, b; Le Gac and Giraud 2008). For many pathogenic fungi, sex must occur within the host after mycelial development. This means that only individuals able to grow within the same host can mate. Adaptation to a new host can in these cases be sufficient to restrict gene flow in sympatry, without requiring active assortative mating, i.e. prezygotic intersterility (Giraud 2006; Giraud et al. 2006). In such cases, close species may remain interfertile for some time, making in vitro crosses a poor criterion for recognizing species. An example is provided by the plant pathogenic genus Ascochyta, in which recent multilocus phylogenetic analyses of a worldwide sample of Ascochyta causing blights of chickpea, faba bean, lentil, and pea revealed that each of these hosted distinct species (Peever 2007). Experimental inoculations demonstrated that infection was highly host specific, yet in vitro crosses showed that the species were completely interfertile. The host specificity of these fungi may therefore constitute the sole reproductive barrier (Peever 2007), resulting in sympatric speciation through the pleiotropic effect of host adaptation (Giraud 2006; Giraud et al. 2006; Giraud et al. 2010). More generally, there exist many close species of ascomycete pathogens that are sympatric but isolated by weak intersterilty barriers (Le Gac and Giraud 2008).

Some other pre-mating barriers to gene flow may allow genetic divergence in sympatry without assortative mating and before intersterility evolves. For organisms depending on biotic vectors, specialization of these vectors can prevent contact between two populations even if they lie close to one another, yielding ecological isolation, e.g. in the Microbotryum violaceum complex of anther smut fungi, the insect vectors are different to some extent between host species, which leads to a reduction in mating opportunities among strains from different plants (van Putten et al. 2007). Another type of pre-mating barrier is allochrony, i.e. differences in the time of reproduction may promote premating isolation, e.g. in the powdery mildew mycoparasite Ampelomyces, the phenology of the host plant of the parasitized fungus provides some reproductive isolation (Kiss et al. 2011). In addition, a high rate of selfing may be efficient in limiting inter-specific matings as seen in some plants (Fishman and Wyatt, 1999). Selfing has been proposed as a reproductive barrier in the anther smut fungus Microbotryum (Giraud et al. 2008b).

Evolution of species criteria in Cercospora

Members of the ascomycete genus Cercospora (Mycosphaerellaceae, Capnodiales, Dothideomycetes) occur world-wide and cause leaf spots on most dicot and monocot plant families, as well as some gymnosperms and ferns (Pollack 1987; Crous and Braun 2003). These fungi rank among some of the most destructive of plant pathogens (To-Anun et al. 2011). Cercospora was first described by Fresenius in Fuckel (1863) with C. apii as the type species. For many years, Cercospora was used for naming any cercosporoid fungus, i.e. a dematiaceous hyphomycete with filiform conidia (Pons and Sutton 1988). As a result, it became one of the largest and most heterogeneous genera of hyphomycetes (Crous and Braun 2003). Chupp (1954) also adopted the broad morphological definition of Cercospora in his monograph.

Deighton (1967; 1973; 1976; 1979) tried to clarify the taxonomy of Cercospora by segregating Cercospora species into smaller and morphologically more similar units. Many Cercospora species were reclassified into Cercosporella, Cercosporidium, Paracercospora, Pseudocercospora, Pseudocercosporella, Pseudocercosporidium, and other genera. Braun (1995) recognized close to 50 genera in the Cercospora-complex. Members of Cercospora sensu stricto are currently recognized as having hyaline or subhyaline, solitary (rarely catenate) conidia formed on pigmented (rarely hyaline to subhyaline) conidiophores (Braun 1995, Crous & Braun 2003, Crous et al. 2009). This morphological criterion of Cercospora has been accepted by most taxonomists in the last 20 years (Hsieh and Goh 1990; Guo and Hsieh 1995; Crous and Braun 1996; Braun and Melnik 1997; To-Anun et al. 2011).

While Cercospora was defined at genus level by morphology until approximately two decades ago, species definition in this genus was based largely on host association. Chupp (1954) considered species of Cercospora to be generally host specific and listed more than 1,900 species names in his monograph. By 1987 more than 3,000 names had been published in Cercospora (Pollack 1987). Crous and Braun (2003) challenged this concept of raising new names for morphologically indistinguishable Cercospora collections based on new host genera, when they assigned 281 morphologically indistinguishable species to synonymy under C. apii senso lato and recognized 659 Cercospora species. The results of some earlier inoculation experiments (Vestal 1933; Johnson and Valleau 1949; Fajola 1978) and molecular sequence data (Crous et al. 2000; Goodwin et al. 2001) had also raised doubt about narrow host specificity in the Cercospora complex. Using host species as a basis for recognizing species of Cercospora also failed the pogo stick hypothesis of Crous and Groenewald (2005) formulated by observing some species of Mycosphaerella, which proposed that host specific fungal plant pathogens often colonised non-host tissue or other substrates, forming fertile fruiting bodies.

In Cercospora, the application of the criterion of intersterility was particularly limited because only a few species in this genus have a known sexual stage (Chupp 1954; Corlett 1991). Groenewald et al. (2006b) detected the two mating type genes in approximately even proportions in C. beticola, C. zeae-maydis and C. zeina populations, and speculated that a sexual cycle may occur regularly in these species. However, the actual sexual stage was not observed.

The use of DNA sequence data and the adoption of Phylogenetic Species Criterion have started to clarify some of the confusion in Cercospora taxonomy. The Cercospora complex has been shown to form a well-defined clade in the Mycosphaerellaceae (Crous et al. 2009) supporting earlier molecular analyses (Stewart et al. 1999; Crous et al. 2001; Goodwin et al. 2001). Furthermore, only species in this group produce cercosporin, a phytotoxin that enhances virulence (Goodwin et al. 2001).

Two examples of the application of Genealogical Concordance Phylogenetic Species Recognition in the Cercospora complex follow. Firstly, two phylogenetically supported species, C. apii and C beticola, were identified among the 281 synonyms placed in C. apii sensu lato by Crous and Braun (2003), even though they were morphologically similar and capable of infecting the same hosts in inoculation experiments (Groenewald et al. 2005; Groenewald et al. 2006a, b). Secondly, C. piaropi and C. rodmanii, which both infect the aquatic plant water hyacinth, were considered to differ from each other by conidial morphology and virulence (Tharp 1917). A multilocus DNA phylogeny did not support the separation of these two species (Tessmann et al. 2001) and a more detailed study using a collection of isolates showed that morphological characters also did not reliably separate them. Consequently C. rodmanii was reduced to synonymy with C. piaropi.

The adoption of Genealogical Concordance Phylogenetic Species Recognition has profound implications for disease control and quarantine. For example, considering the 281 synonyms in C. apii sensu lato as individual species may lead to unnecessary biosecurity measures and trade restrictions, yet considering them as a single species may miss the opportunity to contain some diseases, e.g. those caused by C. beticola.

Evolution of species criteria in Colletotrichum

The ascomycete genus Colletotrichum (Glomerellaceae, Sordariomycetes) contains many well-known plant pathogens that cause anthracnose and a range of diseases worldwide on economic crops and ornamental plants (Crouch & Beirn 2009, Crouch et al. 2009, Damm et al. 2009, Hyde et al. 2009a, b). They are amongst the most important plant pathogens as they cause latent or quiescent infections at the pre-harvest and post-harvest stages (Sutton 1992). The first report of Colletotrichum was by Tode (1790) in the genus Vermicularia, while the genus name Colletotrichum was introduced by Corda (1831). Colletotrichum encompasses species with endophytic, epiphytic, saprobic and phytopathogenic lifestyles (Kumar and Hyde 2004; Photita et al. 2001; 2003; 2005; Liu et al. 2007; Prihastuti et al. 2009), as well as human pathogens (Cano et al. 2004).

The taxonomy of many groups of plant pathogenic fungi, including Colletotrichum, has been based on host association (von Arx 1957; Sutton 1980). If a pathogenic fungus was found on a host from which no records of that pathogen were known, it was described as a new species (von Arx 1957). This species criterion has failed to reliably reflect evolutionary independence of the lineages of Colletotrichum and many other groups of fungi, as many pathogenic species have a facultative saprobic ability, with the exception perhaps of the obligate plant pathogenic fungi, e.g. rusts, smuts, downy and powdery mildews (Cummins and Hiratsuka 2003; Vánky 2002, Yamaoka 2002).

The work done by mycologists during the 19th and early 20th century resulted in numerous fungal names with specific epithets based on the scientific names of the host plant. These names cannot be ignored in modern systematic studies and pose a huge challenge for modern researchers who will need to determine whether these names represent distinct species or synonyms of other names (Hyde et al. 2009a; Cai et al. 2009). Sutton (1992) suggested that “In every large genus like Colletotrichum there needs to be a degree of systematics catharsis resulting in a severe reduction in the number of accepted species before any real advances in identification can be made”. Such an approach was taken by von Arx (1957) who reduced the number of Colletotrichum species from several hundred to 11 based on morphological characters, with many taxa treated as synonyms of C. gloeosporioides (ca. 600 synonyms) or C. dematium (86 synonyms). Several additional species have been accepted since von Arx (1957), based on Morphological Species Criterion (Sutton, 1980; 1992). Sutton (1980) also built a key that has provided a practical identification tool and standard reference for identifying species of Colletotrichum for many years. In the last 30 years, and until recently, the number of newly described species of Colletotrichum has substantially slowed down and species were defined by their distinguishable morphological characters.

Although the application of Morphological Species Criterion has resulted in an important and revolutionary progress in Colletotrichum systematics, some flaws persisted. The species criteria of von Arx (1957) were very broad, and most of his taxonomic treatments were based on literature descriptions, that lead to considerable inaccuracy and generality. For example, the treatment of synonymizing ca. 600 names to C. gloeosporioides has been questionable, as it contains a lot of physiologically and genetically distant lineages. Recent application of Genealogical Concordance Phylogenetic Species Recognition in Colletotrichum has revealed that many distinct species exist in the C. gloeosporioides complex. Some of these have been formally described, such as C. asianum, C. siamense, C. fructicola, and C. cordylinicola (Prihastuti et al. 2009; Phoulivong et al. 2010), while others have been typified, including C. gloeosporioides sensu stricto, C. horii and C. musae (Cannon et al. 2008; Weir and Johnston 2010; Su et al. 2011). In the C. acutatum complex, 4 distinct lineages were recognized and three of them have been assigned species names (Shivas and Tan 2009).

The application of Genealogical Concordance Phylogenetic Species Recognition in Colletotrichum has had an important impact on species discovery, plant breeding, disease control and biosecurity protocols, all of which depend on accurate pathogen identification. Identification of a specimen as C. gloeosporioides sensu lato or C. dematium sensu lato has little practical value. The many species hidden in the C. gloeosporioides complex will certainly have different biosecurity significance. For example, the C. gloeosporioides complex on coffee berries has been well characterized and several distinct genetic and phenotypic species have been established (Prihastuti et al. 2009; Waller et al. 1993). Among these, C. kahawae is a strongly aggressive pathogen specific to coffee in Africa (Waller et al. 1993) and the application of strict quarantine protocols is justified to prevent its spread to coffee growing regions on other continents where it is not present. On the other hand, C. asianum, C. fructicola and C. siamense are opportunistic pathogens of coffee berries (Prihastuti et al. 2009) and appear to have a wide host range and little biosecurity significance. Phoulivong et al. (2010) indicated that morphologically similar isolates from chilli, mango, papaya, rose apple and jujube comprised more than one distinct species. Very little is known about whether these records, and many other worldwide records of C. gloeosporioides sensu lato, represent saprobes, weak or opportunistic pathogens or severe pathogens (Hyde et al. 2010).

Evolution of species criteria in Fusarium

The ascomycete genus Fusarium (Nectriaceae, Hypocreales, Sordariomycetes) represents a large group of ascomycetes ubiquitously distributed in soil and in association with plants. Although most members are saprobic, Fusarium is better known for its toxigenic and plant pathogenic species, which significantly impact agriculture (Marasas et al. 1984). Fusarium produces secondary metabolites, such as fumonisins, trichothecenes and zearalenone, which are toxins that threaten food safety and human health. Recently, Fusarium species also have emerged as opportunistic human pathogens causing ocular or systemic infections (Dignani and Anaissie 2004; Zhang et al. 2006).

Fusarium was first described by Link (1809) as species with fusiform spores borne on a stroma. This asexual genus was validated by Fries (1821) in terms of the IBCN. With increased knowledge of fungal morphological identification, the presence of fusoid macroconidia with a basal foot cell became accepted as the key character of the genus instead of the presence of a stroma (Booth 1971). The teleomorphs of Fusarium, when known, belong to either Haematonectria or Gibberella.

The history of the taxonomy of Fusarium has been unstable. In the first century after the genus was established, over 1,000 species were defined on the basis of superficial observations, often based on host association (Toussoun and Nelson 1975). Wollenweber and Reinking (1935) in their monograph Die Fusarien reduced the genus to 142 species, varieties and forms in 16 sections. Based solely on morphological characters, Snyder and Hansen (1941; 1945) further reduced Fusarium to nine species. Successor taxonomists conducted revisions based on either the Wollenweber and Reinking system, e.g. Booth (1971), or the Snyder and Hansen system, e.g. Nelson et al. (1983). This resulted in many taxonomic incongruences. Currently, there are over 100 valid Fusarium species names according to the Dictionary of The Fungi (Kirk et al. 2008). The frequent conflicts and instability in Fusarium systematics have resulted from the absence of clear morphological characters to separate species as well as the existence of phenotypic variation in cultures (Geiser et al. 2004).

The root of the problem in Fusarium systematics lies in the use of the Morphological Species Criterion as adopted in traditional fungal taxonomy. As with many other fungi, Fusarium taxonomy was mostly based on the Morphological Species Criterion until two decades ago, when Morphological Species Criterion and Phylogenetic Species Criterion were applied (Summerell et al. 2010). Intersterility barriers have been detected in some Fusarium groups based on crossing experiments, e.g. the genetically distinct mating populations in the F. solani species complex (Matuo and Snyder 1973). Clearly, the Biological Species Criterion cannot be applied to the majority of Fusarium lineages, which are homothallic or apparently asexual (Taylor et al. 1999). Nonetheless, the BSC enabled reliable identification for some sexual species of Fusarium, especially for those in the Gibberella fujikuroi and F. solani species complexes (Leslie and Summerell 2006; Kvas et al. 2009).

For species of Fusarium for which sexual reproduction is difficult to induce in vitro, the Phylogenetic Species Criterion (including the Genealogical Concordance Phylogenetic Species Recognition) has proven to be highly informative. Several studies have shown that Phylogenetic Species Criterion and Biological Species Criterion are congruent, although the Biological Species Criterion appears more finely discriminating than the Morphological Species Criterion, particularly in Fusarium (O’Donnell 2000; O’Donnell et al. 1998; Zhang et al. 2006). Distinct lineages recognized by the Phylogenetic Species Criterion can be used as a guide for finding diagnostic morphological or ecological differences among fungal species, which otherwise did not appear obvious. For example, phylogenetic analyses have revealed the existence of cryptic species in F. graminearum, occurring in different parts of the world, resulting in the description of 13 new species (O’Donnell et al. 2000; Starkey et al. 2007; Kvas et al. 2009). The most frequently used gene for species recognition and phylogenetic analysis in Fusarium is the translation elongation factor 1 α gene (EF-1α). Other useful loci include the internal transcribed spacer (ITS) region of the rRNA gene repeat and β-tubulin gene (Geiser et al. 2004; Summerell et al. 2010; Park et al. 2011).

An example of the benefits of modern Phylogenetic Species Criterion-based systematics in Fusarium can be seen in soybean sudden death syndrome (Rupe 1989), a disease occurring throughout the world. Previously, the causal agent was referred to as F. solani f. sp. glycines or Fusarium solani sensu lato (Gao et al. 2004). Multiple gene sequence analyses revealed a diversity of cryptic taxa, leading to the recognition of four different species of Fusarium as causes of this disease. Those isolates known to cause soybean sudden death syndrome in North America have been segregated as Fusarium virgulifore, while three related but different species of Fusarium are associated with this disease in South America (Aoki et al. 2005). Studies based on experimental crosses further confirmed the distinction of these phylogenetic species (Covert et al. 2007). The intersterility and allopatry of these cryptic species were discovered only after their existence was revealed using phylogenetic analyses. The geographical and genetic information associated with these newly recognized species will enable the development of precise control strategies and the formulation of appropriate biosecurity policies.

Evolution of species criteria in rust and smut fungi

The rust and smut fungi belong to the subphyla Pucciniomycotina and Ustilaginomycotina respectively, in the phylum Basidiomycota (Hibbett et al. 2007). At the species level there are almost 7,000 species of rust fungi in 163 genera (Kirk et al. 2001) and about 1,675 smut fungi in 95 genera (Vánky, pers. comm.), which collectively account for about 10% of all known fungi. The rust and smut fungi both contain many economically and agriculturally important species and their profound influence on human history is well documented (Carefoot and Sprott 1967).

The traditional definition of the rust and smut life form was the presence of teliospores that germinate to produce basidia and these fungi were classified together in the Teliomycetes (Jülich, 1981). However ultrastructural and molecular studies have shown that the rust and smut fungi are only distantly related with separate monophyletic origins (Bauer et al. 1997; Begerow et al. 1997). Furthermore, the rust fungi have a pleomorphic life cycle with up to five spore states (spermatia, aeciospores, urediniospores, teliospores and basidiospores) which differentiates them from the smut fungi that mostly produce two types of spores (teliospores and basidiospores).

Rust and smut fungi are essentially obligate plant pathogens, although the smut fungi may have a short stage of saprobic growth on non-living substrates, and about 30 rusts have been cultivated on artificial media (Yamaoka 2002). The classification of rust and smut species has been traditionally based on both Morphological Species Criterion, with emphasis on sori and spore stages, as well as on Ecological Species Criterion, with emphasis on pathogenicity on specific hosts (Vánky 2002, Cummins and Hiratsuka 2003). Consequently these groups have been relatively stable, and easily classified and identified using morphology and host associations for over 200 years. The recent application of molecular phylogenetic analyses to the rust and smut fungi has mostly supported previous classifications at the level of genus (Maier et al. 2003; Bauer et al. 2006; Aime et al. 2006; Begerow et al. 2006). In contrast, higher taxonomic ranks have been shuffled in part. While the Pucciniales (rust fungi) are monophyletic, the smut fungi in the traditional sense correspond to at least two phylogenetically distant clades. Those smut fungi in the Ustilaginomycotina cluster together with other plant pathogens of the Exobasidiales, Microstromatales as well as human pathogens of Malassezia (Begerow et al. 2000; Begerow et al. 2006). The smut fungi in the Microbotryales cluster together with other plant pathogens and mycoparasites in the Microbotryomycetes in the Pucciniommycotina (Bauer et al. 2006). Ultrastructural characters and cell wall compounds have subsequently supported these modern classifications (Bauer et al. 1997; 2006).

Early species descriptions of rust and smut fungi were usually based on morphology and separation of the two groups was not always clear. The earliest descriptions of rust fungi date back to Micheli (1729) and Persoon (1801), and smut fungi to Tillet (1755) and Prévost (1807). They were followed by detailed studies on their life cycle including teliospore formation and germination (de Bary 1853; Tulasne and Tulasne 1847). Interestingly, most of these early treatments included quite comprehensive studies comprising morphological, physiological and systematic aspects. Most of the early species lists included information about host species and the lists were even organized according to host (Fischer von Waldheim 1869). It is commonly accepted that rust and smut species are each restricted to a family or even a narrower plant taxon and there are many examples describing the rust or smut species of a given plant family (Vánky 2006; Vánky and Lutz 2007) or a given host genus (Bauer et al. 1999; Vánky 2003). Several genera of heteroecious rust fungi on the other hand need two different hosts to complete their life cycle, which may be from completely different plant families.

Considering that over 8,000 species of rust and smut fungi are known, there has been little application of DNA sequence data to identify rust and smut fungi in terms of the Genealogical Concordance Phylogenetic Species Recognition. This is certainly because the morphological characteristics of rust and smut fungi have appeared stable and reliable, allowing confident identification of species that are supported, in most cases, by narrow host ranges. However complexes composed of closely related reproductively isolated cryptic species are still common in the rust fungi (Gaümann 1959) and also some groups of smut fungi. Consequently, molecular studies have been used to resolve phylogenetic relationships between morphologically similar taxa of rust and smut fungi. For example, the host-specificity, morphology and DNA sequence data of two microcyclic rusts species Puccinia melampodii and P. xanthii (Seier et al. 2009) lead to the establishment of a new morphospecies, P. xanthii var. parthenii-hysterophorae, to accommodate records of P. melampodii associated with the host Parthenium hysterophus. Another example is Karnal bunt of wheat, caused by the smut fungus Tilletia indica, an important pathogen absent from Australia. ITS sequence data have been used to separate it from morphologically similar species that may also occur as contaminants in consignments of wheat seed (Levy et al. 2001, Pascoe et al. 2005). Overall, the application of Phylogenetic Species Criterion supported Morphological Species Criterion in many of the genera of smut fungi (Hendrichs et al. 2005; Begerow et al. 2000; Castlebury et al. 2005) and only recently the large Ustilago / Sporisorium / Macalpinomyces species complex has been substantially resolved using a phylogeny derived from molecular data that reflected morphological synapomorphies and host associations (Stoll et al. 2003, 2005; McTaggart 2010).

As one of the best studied models, Microbotryum has provided a good example of the utility of the Phylogenetic Species Criterion. A narrow species criterion based on host use (Zillig 1921; Liro 1924; Baker 1947) long contrasted with a broad Morphological Species Criterion species criterion, the latter defining a single species, Microbotryum violaceum, considered as the pathogen responsible for almost all anther smuts of Caryophyllaceae (Perlin 1996). Population genetics studies (Bucheli et al. 2000) and the use of the Genealogical Concordance Phylogenetic Species Recognition (Le Gac et al. 2007a) revealed an absence of gene flow and an ancient differentiation between populations of Microbotryum found on different host plants, which were confirmed by ITS phylogenies as distinct species on different hosts (Lutz et al. 2005, 2008). As in other fungal species, the Biological Species Criterion was less discriminating (Le Gac et al. 2007b), with little evidence of assortative mating in the form of conjugation initiation in vitro, although hybrid inviability and sterility was observed between the cryptic species (Le Gac et al. 2007b; de Vienne et al. 2009a). Cross-inoculation studies also appeared less discriminating in vitro than host specificity seen in the field (de Vienne et al. 2009b; Gladieux et al. 2011). The studies highlight the value of Genealogical Concordance Phylogenetic Species Recognition to validate Phylogenetic Species Criterion and Morphological Species Criterion that have practical application in the field of plant pathology.

Conclusive remarks

Molecular DNA sequence data have recently been extensively employed in studying the systematics of plant pathogenic fungi. The advantage of using molecular data is that it provides a greater number of heritable characters that allow for convenient information sharing between laboratories. Morphological characters, however, are prone to change under different environmental conditions. In addition, well-developed bioinformatics tools make analysis of data more objective and less controversial when examined by different scientists. The Biological Species Criterion has proved useful in some fungal groups but overall appears less convenient and less discriminating, although congruent, with the Genealogical Concordance Phylogenetic Species Recognition.

The Genealogical Concordance Phylogenetic Species Recognition has been widely accepted in fungal systematics. Multi-gene sequencing and phylogenetic analysis have become a routine procedure in identifying new fungal species, especially for those that lack distinctive morphological characters. Consequently a rapidly increasing number of cryptic species are being discovered amongst plant pathogenic fungi using the Genealogical Concordance Phylogenetic Species Recognition and it is critical to determine their host range, the severity of diseases they cause and their biosecurity significance. With rapidly expanding global trade it has become imperative that we develop effective and reliable protocols to detect these previously unrecognized pathogens.

Based on Genealogical Concordance Phylogenetic Species Recognition, previously applied phenotypic characters that were used to define taxa need re-evaluation. Although the recent molecular advances in multi-locus phylogeny has been able to recognize stable and well-separated phylogenetic species, there is still a long way to go before we can finally establish a natural classification system and species recognition criteria for most fungi using DNA sequence data. A major problem is that only a very small number of species have been deposited in public culture collections and only a fraction of these have had some DNA fragment sequenced. The number of ex-type strains that have been sequenced is even lower. Another potential problem is that most of the well-known plant pathogenic fungi were described based on the Morphological Species Criterion and ex-type cultures are not available. It is inevitable that Phylogenetic Species Criterion and Morphological Species Criterion will still be used together to define species for many years.

Once robust species delimitation and classification is established, the development of high-throughput identification tools like barcoding should be a major improvement for assignment of particular strains to species (Begerow et al. 2010). Together with well-curated databases and regularly updated regional species lists rapid molecular identification is able to efficiently support not only quarantine regulations but the monitoring of new emerging diseases as well.


Dr Alistair McTaggart (Louisiana State University) is thanked for his helpful comments. Ms. Liu Fang is thanked for technical assistance. Lei Cai acknowledges grants CAS KSCX2-YW-Z-1026 and NSFC 31070020. Tatiana Giraud acknowledges the grants ANR 06-BLAN-0201 and ANR 07-BDIV-003. Parts of this chapter are derivatives of articles previously published (Giraud et al., 2008a; Gladieux et al., 2010). Roger Shivas acknowledges Chinese Academy of Sciences for the Visiting Professorship for Senior International Scientists (Grant No. 2010T2S12). Dominik Begerow acknowledges the grants BE 2201/4-2 and BE 2201/8-1 from DFG.

Copyright information

© Kevin D. Hyde 2011