Fungal systematics: is a new age of enlightenment at hand?

| Fungal taxonomists pursue a seemingly impossible quest: to discover and give names to all of the world’s mushrooms, moulds and yeasts. Taxonomists have a reputation for being traditionalists, but as we outline here, the community has recently embraced the modernization of its nomenclatural rules by discarding the requirement for Latin descriptions, endorsing electronic publication and ending the dual system of nomenclature, which used different names for the sexual and asexual phases of pleomorphic species. The next, and more difficult, step will be to develop community standards for sequence-based classification. PERSPECTIVES NATURE REVIEWS | MICROBIOLOGY ADVANCE ONLINE PUBLICATION | 1 Nature Reviews Microbiology | AOP, published online 3 January 2013; doi:10.1038/nrmicro2942 © 2013 Macmillan Publishers Limited. All rights reserved Nature Reviews | Microbiology and Penicillium camemberti and Penicillium roqueforti (used to make Camembert and Brie, and Roquefort cheeses, respectively). However, Penicillium spp., as traditionally delimited, are paraphyletic as well as pleo‐ morphic, so these well‐known species cannot all remain in this historic genus15. Under the revised code, any of the exist‐ ing valid names for a species can be selected as its correct name, with preference given to the oldest name. However, this libertarian view is tempered by two additional revi‐ sions, both involving review by the General Committee (GC) of the ICN, which is empowered to vote on proposals to conserve or reject names of fungal taxa, as well as to modify the ICN itself 3. First, in situations in which both the anamorph and teleomorph names for the same taxon are widely used — for example, Fusarium (anamorph) and Gibberella (teleomorph) at the genus level — the teleomorph name can be chosen with‐ out approval of the GC, but selection of the anamorph name, even if it is the older name, requires approval. Apparently, it is hard for systematists to abandon the primacy of sex‐ ual characteristics. Second, the GC has the authority to approve lists of names, which presumably will be generated by committees of mycologists with expertise in particular taxonomic groups. However, mycologists have retained the right to appeal any deci‐ sion about names through the established process of conservation of names. No one has had a chance to choose a name for a pleomorphic fungal species under the new code, which only came into effect on 1 January 2013, but the nomenclatural changes mentioned above illustrate what might lie ahead. Another example is the work of Gräfenhan et al.9 on the taxonomy of the anamorphic genus Fusarium, one of the largest genera of fungi, containing nearly 1,500 species, subspecies, varieties and formae speciales. Fusarium spp. include important plant and animal pathogens and mycotoxin producers and have been linked to as many as seven tele‐ omorph genera. On the basis of sequence analyses for RNA polymerase II and ATP citrate lyase genes, Gräfenhan et al.9 identi‐ fied 15 clades with Fusarium‐like asexual forms and gave six of them names based on anamorphs, although five of these six have known teleomorphic forms. The reclassifica‐ tion of Fusarium by Gräfenhan et al. is based on robust phylogenies and would be nomen‐ claturally valid under the forthcoming ICN. Nevertheless, name changes in the genus Fusarium sensu lato might confuse and inconvenience user communities and regula‐ tory bodies in agriculture and medicine, and it remains to be seen how these constituencies will react to the new taxonomy. The complex nomenclatural history of many groups of pleomorphic fungi, coupled with phylogenetic uncertainty and the some‐ times passionate opinions of stakeholders, presents a very challenging taxonomic prob‐ lem. The code provides guidance, but many decisions about names cannot be reduced to ‘legal’ algorithms. As a follow‐up to the ‘One fungus, one name’ movement that led to the repeal of dual nomenclature, a ‘One fungus, which name?’ conference was held in Amsterdam in April 2012 (REF. 16). This time, the goal was to begin working through the myriad options for the classification of pleomorphic fungi in light of the new rules. Similar meetings and workshops on the taxonomy of the genus Fusarium, the order Hypocreales and other groups were held in association with meetings of the Mycological Society of Japan (May 2012), the Mycological Society of America (July 2012) and the Mycological Society of China (August, 2012). Classification of environmental sequences Now that dual nomenclature has been abolished, the next major challenge for fungal taxonomy is to develop strategies for classifying environmental sequences (FIG. 2). Nobody knows how many unnamed species have already been detected through metagenomic studies (and this fact alone indicates the need for a centralized database of species that are based on environmental sequences), but as early as 2007 the number of clusters of closely related rRNA genes being discovered with Sanger chemistry approached the number of species being described from specimens5, and the rate of molecular species discovery has surely increased with the application of next‐ generation sequencing in metagenomics. Environmental studies have revealed not only individual species, but also major clades of fungi, such as the class Archaeorhizomycetes17, containing a diverse group of soil‐inhabiting fungi from the phylum Ascomycota. Sequences of Archaeorhizomycetes members have been reported in more than 50 independent stud‐ ies, and they can be grouped into more than 100 species‐level entities17. Nevertheless, only one species, Archaeorhizomyces finlayi, has been formally described, based on a culture that was obtained from coni‐ fer roots. A similar example is provided by the phylum Rozellomycota18 (also known as Cryptomycota19), a large clade of aquatic and soil‐inhabiting fungi that is known almost entirely from environmen‐ tal sequences. The phylum Rozellomycota has been shown to contain the previously described chytrid genus Rozella19, but most of the diversity of this phylum is in groups that are known only from environmental sequences and have not been named. These examples, and many others from fungal molecular ecology, illustrate the profound disconnect that now exists between formal taxonomy and species discovery through environmental sequences. Barriers to the naming of such species include a perceived conflict with the code, and errors and Figure 1 | Two names, one fungus. Eurotium herbariorum is a pleomorphic fungus that has a sexual phase, reproducing by ascospores (the teleomorphic form; left), as well as a conidium-producing asexual phase (the anamorphic form; right) that has been named Aspergillus glaucus. Images courtesy of Paul F. Cannon, Royal Botanical Gardens, Kew, London, UK, and the Centre for Agricultural Bioscience International (CABI). P E R S P E C T I V E S 2 | ADVANCE ONLINE PUBLICATION www.nature.com/reviews/micro © 2013 Macmillan Publishers Limited. All rights reserved incomplete taxon sampling in reference sequence databases. The perceived incompatibility of the code with sequence‐based taxonomy is a consequence of the requirement for type specimens. However, the code places no restrictions on the form of type specimens, which need not be complete or representa‐ tive; all that is required of a type specimen is that it should be a physical specimen. In principle, an aliquot of DNA extracted from an environmental sample, or a por‐ tion of the substrate from which the DNA was isolated, can serve as a legitimate type specimen. To prove this point, Kirk et al.20 recently described a new species of rumen chytrid, Piromyces cryptodigmaticus, based on sequence data, and typified it with a sample from the fermenter from which the DNA was extracted. The new taxon name was validly published, even though the fungus was never directly observed. In the future, if purely sequence‐based taxonomy is incorporated into the code, it may be pos‐ sible to forego the deposition of physical type materials altogether. In the meantime, the publication of P. cryptodigmaticus pro‐ vides a model for environmental molecular biologists who would like to formalize their discoveries through code‐compliant taxonomic names. Errors and incomplete taxonomic sampling in sequence databases, such as GenBank, present a psychological barrier to naming environmental sequences; if an environmental sequence has no match in GenBank, it could still represent a described but unsequenced species. Faced with such uncertainty, fungal taxonomists might be reluctant to describe new species based on environmental sequences. They should not be; current estimates of the actual diversity in the kingdom Fungi range from as few as 500,000 species to millions of species2, suggesting that most unmatched environmental sequences probably do rep‐ resent new species5. Even if some environ‐ mental species prove to be redundant, taxonomists are accustomed to resolving synonymy based on the principle of priority. Finally, the solution to the GenBank prob‐ lem is conceptually straightforward — that is, generate well‐documented reference sequences21 — and is already being pur‐ sued through the fungal bar‐coding initia‐ tive22 and the creation of custom‐curated databases of well‐documented reference sequences, such as the RefSeq collection within GenBank, and the UNITE database for mycorrhizal fungi23. Lessons from prokaryotic taxonomy Many of the taxonomic challenges faced by mycologists parallel those faced by researchers studying prokaryotes, but the nomenclatural practices adopted by the two groups are often divergent. For example, the expanded power of the GC to rule on the legitimacy of choices among existing names under the forthcoming ICN might worry some mycologists, who could fear a loss of taxonomic freedom, but the new system for fungi might seem familiar to prokaryote taxonomists, who have long used a Judicial Commission to accept or reject newly proposed names24,25. Another key difference between the nomenclatural codes for prokaryotes26 and fungi3 is that Nature Reviews | Microbiology Uncultured fungus clone unisequences#37-3808_2763 ITS2, PS Uncultured Agaricomycotina clone 6_g19 18S rRNA gene Uncultured fungus clone MOTU_4043_GVUGB5B04JK5N2 18S rRNA gene, PS, ITS2 Uncultured fungus clone MOTU_1778_GVUGB5B04IF01X 18S rRNA gene, PS Uncultured fungus clone LT5P_EUKA_P5H04 18S rRNA gene, 18S–25/28S rRNA gene Uncultured fungus clone F66N0BQ02H1NX5 18S rRNA gene Uncultured fungus clone MOTU_43 Uncultured fungus clone unisequences#69-3466_2373 ITS2, PS Uncultured fungus clone MOTU_4349_GOKCVYYY06GR7WA 18S rRNA gene, PS, ITS2 Uncultured fungus clone unisequences #65-3574_00447, ITS2, PS Trichosporonales sp. LM559 18S rRNA gene Uncultured Tremellales clone LTSP_EUKA Uncultured fungus clone unisequence Fungi 3 leaves Uncultured fungus clone unisequence#65-3936_0554 ITS2, PS Uncultured basidiomycete ITS Uncultured fungus clone MOTU_601_GOK Uncultured Tremellales clone LTSP_EUKA_P4L03 18S rRNA gene, PS, ITS Uncultured fungus clone MOTU_141_GOKCVYYY06G5FYL 18S rRNA gene, PS, ITS Fibulobasidium murrhardtense strain CB59109 18S rRNA gene Uncultured fungus clone MOTU_2930_GOKCVYYY06G7201 18S rRNA gene, PS, ITS Uncultured fungus clone MOTU_2993_GOKCVYYY06HH12J 18S rRNA gene, PS, ITS Uncultured fungus clone MOTU_1888_GVUGV5B04JJTLJ 18S rRNA gene Uncultured fungus clone MOTU_3006_GVUGV5B04JIHT 18S rRNA gene Uncultured fungus clone MOTU_2635_GVUGVSB04J56R4 18S rRNA gene, PS, ITS gi|22497358|gb|FJ761130.1| uncultured fungus clone singleton_70-3063_2201 18S rRNA gene Uncultured fungus clone singleton_70-3063_2201 18S rRNA gene, PS, ITS Uncultured fungus clone OTU_403_GW5CJXV07IOX5A 18S rRNA gene Uncultured fungus clone U_QM_090130_240_B_plate1a12.b1 18S rRNA gene, PS, ITS1 Uncultured fungus clone OTU_1445_1GW5CJXV07HXDTO 18S rRNA gene Uncultured fungus clone U_QM_090130_127_1A_plate1g12.b1 18S rRNA gene, PS, ITS1 Uncultured fungus clone MOTU_3163_GYUGV5B0412KQP 18S rRNA gene, PS, ITS1 Uncultured fungus clone MOTU_533_GOKCVYYY06GU3JA18S rRNA gene, PS, ITS1 Uncultured Tremellales clone 5_D20 18S rRNA, ITS1, 5.8S rRNA gene, ITS1 Uncultured Rhodotorula IT51, 5.8S rRNA, ITS2 and partial 28S rRNA, clone MNIB2FAST_K1 Uncultured fungus clone MOTU_3797_GOKCVYYY06HBZ1X 18S rRNA gene, PS, ITS2 Uncultured fungus clone MOTU_2412 Figure 2 | Unnamed diversity. A demonstration of the problem posed by unnamed fungi that are known only from environmental DNA sequences. When a new environmental sequence (the bottom-most operational taxonomic unit, gi|22497358; blue box) was used in a BLAST search of the GenBank database and the result displayed using the BLAST distance tree tool, only two of the 35 most closely related sequences were from cultured organisms (green boxes), and only one was named (Fibulobasidium murrhardtense). Without names, the information content of this tree leaves much to be desired. ITS, internal transcribed spacer; PS, partial sequence. P E R S P E C T I V E S NATURE REVIEWS | MICROBIOLOGY ADVANCE ONLINE PUBLICATION | 3 © 2013 Macmillan Publishers Limited. All rights reserved the prokaryotic code specifies the technical means to recognize new species, and all new species are recorded in the International Journal of Systematic and Evolutionary Microbiology, whereas the ICN specifies no particular technique for the recognition of fungal species, which can be published in diverse venues. Under the ICN, acceptance of fungal species is left to the mycology community; new names are picked up by other mycologists and appear in the litera‐ ture, or they are simply ignored. The highly regulated system for prokaryotes promotes uniformity in the species recognition cri‐ teria and preserves the stability of names, but it can also limit the rate of species description. By contrast, the laissez‐faire system for fungi results in non‐uniform spe‐ cies recognition criteria (for example, many new species descriptions lack supporting molecular data5), extensive synonymy, an ongoing challenge in compiling new names (although the new requirement for name registration will solve this problem) and frequent changes in species‐level classifica‐ tions. At the same time, the fungal system promotes rapid taxonomic updates to reflect new discoveries and advances in phylogenetic reconstruction. Changes in fungal species classifica‐ tions often occur when evidence for genetic diversity is discovered within morphological taxa. For example, it might have surprised readers to learn that Alexander Fleming’s Penicillium species, Penicillium chrysogenum, is now known as P. rubens27, but the change was necessitated when phylogenetic and population genet‐ ics data showed that the P. chrysogenum of old harboured several genetically isolated species28. Older mycologists may grumble about having to learn a new name, but the new classification reflects the current state of knowledge, and new students will not be bothered by the change. By contrast, the archaeon Sulfolobus islandicus was shown to comprise several genetically isolated species according to population genetics techniques, which showed genetic isolation by distance29 and also evidence of ecologi‐ cal speciation30, but these species were left unnamed, in part because the now passé technique of DNA–DNA hybridization would have been required for formal spe‐ cies descriptions24. Admittedly, there are huge challenges in determining species limits in bacteria and archaea, particularly in the face of extensive horizontal gene transfer31. Nonetheless, the differences in nomenclatural practices for bacteria and archaea versus fungi may be part of the reason why the number of new spe‐ cies described per year is about twice as many for fungi as it is for prokaryotes5,32. The ICN will increase the centralization of taxonomic authority for fungi, although the basic criteria for fungal species recognition will remain unrestricted. It is important that as the new rules of the ICN are imple‐ mented, the GC acts with restraint and does nothing to impede progress in fungal species description. Mycologists can also learn from the expe‐ rience of bacterial and archaeal researchers with regard to the classification of environ‐ mental sequences. The requirement for a living type culture for describing bacterial or archaeal species26 is comparable to the requirement for a physical type specimen for naming fungal species. To enable the naming of bacteria that lack cultures but are known by “more than a mere sequence” (REF. 33), Murray and Schleifer34 suggested that the prefix Candidatus be used, indicating that the name is provisional. This recommen‐ dation has been appended to the bacterial code25, but fewer than 400 bacteria and archaea have been described as Candidatus species35. If mycologists wish to adopt a new category similar to Candidatus to accommo‐ date the huge numbers of species discovered through environmental sequences, as has been suggested5, they will need to find ways to facilitate high‐throughput taxonomy, almost certainly involving automated work flows. The future of fungal taxonomy Twenty‐five years after the first descrip‐ tion of PCR, species‐level fungal taxonomy is finally catching up with the molecular revolution. Change has come slowly and has been prompted by the actions of radi‐ cals, who flouted and subverted the code by naming taxa based on anamorphs7,8 or environmental sequences20. Such individual acts of rebellion illuminate the way forward, but ultimately fungal taxonomy is a group enterprise that can succeed only with the support and participation of the broad community of mycologists. Proponents of unitary taxonomy worked effectively as a community to repeal dual nomencla‐ ture and are now organizing themselves to resolve the correct names of scores of pleomorphic fungal species. Supporters of sequence‐based taxonomy have not been so unified, however. The publication of P. cryptodigmaticus demonstrates that it is ‘legally’ possible, under the code, to describe new species based on sequences (as long as a nominal type is deposited somewhere), but community effort will be needed to develop the broadly accepted protocols required for a mass movement towards sequence‐based taxonomy. At least one difficult issue appears to have been resolved: the internal transcribed spacer (ITS) region of the nuclear rRNA gene has been proposed as the fungal bar‐ code locus22 and is being used for sequence‐ based species delimitation in environmental surveys for many groups of fungi. However, other key issues remain problematic. Longer reads that provide sequences for the ITS and the phylogenetically tractable large subu‐ nit (LSU) rRNA cannot be obtained until there are improvements in next‐generation sequencing. The gold standard for species delimitation in fungi is the genealogical concordance method, which uses multiple genetic loci to assess the limits of recombina‐ tion36. Such approaches are not applicable in environmental data sets, which usually use single loci amplified from pooled DNAs. Moreover, in order to carry out species delimitation in environmental samples, the consequences of intragenomic heterogeneity in multicopy rRNA genes, as well as error owing to gene tree versus species tree con‐ flict, will have to be determined empirically in relation to multigene data sets. The names of species known only from environmental sequences might require a new taxonomic category comparable to the Candidatus status for bacteria and archaea5, or an iden‐ tifying suffix (for example, ENAS (environ‐ mental nucleic acid sequence) or eMOTU (environmental molecular operational taxonomic unit))6. The reality of sequenc‐ ing errors might prevent naming until the same sequence is found a second time and by a different research group. Finally, myco‐ logical databases such as MycoBank must prepare for a massive influx of new species, especially if automated work flows are devel‐ oped to describe fungi from environmental nucleic acid sequences. Given the rate of species discovery, mycologists do not have another 25 years to ponder the problem. David S. Hibbett is at The Biology Department, Clark University, Worcester, Massachusetts 01610, USA. John W. Taylor is at The Department of Plant and Microbial Biology, University of California, Berkeley, California 94720, USA. Correspondence to D.S.H. e-mail: [email protected] doi:10.1038/nrmicro2942 Published online 3 January 2013 1. Kirk, P. et al. (eds) Dictionary of the Fungi 10th edn (CABI, 2008). 2. Bass, D. & Richards, T. A. Three reasons to re‐evaluate fungal diversity “on Earth and in the ocean”. Fungal Biol. Rev. 25, 159–164 (2011). P E R S P E C T I V E S 4 | ADVANCE ONLINE PUBLICATION www.nature.com/reviews/micro © 2013 Macmillan Publishers Limited. All rights reserved 3. McNeill, J. et al. (eds) International Code of Botanical Nomenclature (Vienna Code) (Gantner, 2006). 4. Hawksworth, D. L. et al. The Amsterdam declaration on fungal nomenclature. IMA Fungus 2, 105–112