Importance of Resolving Fungal Nomenclature: the Case of Multiple Pathogenic Species in the Cryptococcus Genus

Cryptococcosis is a major fungal disease caused by members of the Cryptococcus gattii and Cryptococcus neoformans species complexes. After more than 15 years of molecular genetic and phenotypic studies and much debate, a proposal for a taxonomic revision was made.

Brazil, 26 to 30 March 2017), this matter was once more discussed, and ample evidence was provided that at least seven, and likely even more, species exist.
Cryptococcosis is an important fungal infection, globally affecting immunocompromised and immunocompetent humans and animals (3,4). Annually more than 200,000 HIV-positive individuals develop cryptococcal meningitis with approximately 180,000 casualties (5). The phenotypic heterogeneity within the Cryptococcus neoformans species complex has been known for many years, beginning with the identification of four serotypes, serotypes A to D (6,7). The discovery of an atypical clinical cryptococcal isolate led to the designation of a new variety named C. neoformans var. gattii (serotypes B and C) next to C. neoformans var. neoformans (serotypes A and D) (8,9). The observation of the sexual cycle led to the description of Filobasidiella neoformans and Filobasidiella bacillispora (10)(11)(12). A third variety, C. neoformans var. grubii, was introduced in 1999 for serotype A strains, thus the variety neoformans became restricted to serotype D strains (13). In 2002, C. neoformans var. gattii was raised to species level, and the name C. gattii was given nomenclatural priority over the older name C. bacillisporus (14). At this stage, two species, C. gattii and C. neoformans, were recognized with the latter comprising two varieties, neoformans and grubii. The presence of diploid and aneuploid serotype A and serotype D hybrids (C. neoformans ϫ C. deneoformans) has been known for a long time (7,(15)(16)(17)(18), and they constitute 19 to 36% of the cryptococcal agents in southern Europe (19,20). It is noteworthy that from a nomenclatural point of view, the type strain of C. neoformans CBS132 is a serotype AD hybrid (1,17).
Molecular data revealed that the C. neoformans and C. gattii species complexes were unexpectedly genetically diverse (30). On the basis of four genes, it was calculated that C. neoformans/C. deneoformans separated from the C. gattii species complex 37 million years ago, C. neoformans and C. deneoformans separated 18.5 million years ago, and C. gattii and C. bacillisporus separated 9.5 million years ago (31). These divergence times might be older, as recent calculations based on genomic data fine-tuned the divergence time of the C. neoformans/C. deneoformans and the C. gattii species complex to 80 to 100 million years ago (32). The genomes of C. deneoformans and C. neoformans differ at~10% of nucleotide positions (33). This difference is so large that the same phylogenetic groups have been found no matter which particular isolates were used and despite the increasing resolution of molecular typing tools, such as PCRfingerprinting, amplified fragment length polymorphism (AFLP) fingerprinting, multilocus sequence typing (MLST), and whole-genome sequencing (WGS) (15,30,(34)(35)(36)(37)(38)(39)(40)(41)(42).
Phenotypic, ecological, and geographical variation also supports creating specieslevel taxa in the C. gattii and C. neoformans species complexes (Table 1) (1,. For example, a recent study on virulence attributes such as capsule and melanin of members of the C. gattii species complex concluded with "These findings argue for increased acceptance of the new species and may be useful for informing diagnosis and prognosis in clinical infection" (50).
Genetic methods revealed that intraspecies crosses between C. neoformans and C. deneoformans isolates showed a higher spore viability compared to C. deneoformans ϫ C. neoformans interspecies crosses (33). Twenty-three quantitative trait loci were identified from the analysis of interspecific crosses involved in virulence-associated and azoleresistant phenotype differences between both species (61), and the observed postzygotic isolation mechanisms were explained by Bateson-Dobzhansky-Muller incompatibility affecting basidiospore viability in interspecific crosses (62). Mitotic recombination, causing chromosomal loss and crossing over, seems a further genetic separation mechanism  Present (65) Present (65) Present (65) Lost (65) Present (65) N D

Mycophenolic acid
Sensitive (66) Sensitive (66) Sensitive (66) Sensitive (66) Sensitive (66) Not sensitive (66) ND Growth on the following medium: between both species (63). One study indicated that C. neoformans (cited as serotype A strains) reproduced mainly clonally, whereas C. deneoformans (cited as serotype D strains) showed recombination. Moreover, genomic differences and MLST analysis separated both species (64). Cryptococcosis is usually diagnosed by microscopy, histopathology, culture, and serology, including lateral flow assays, and by molecular assays (Table 1) (68-92), all of which allow straightforward identification of unknown environmental and clinical cryptococcal isolates. Importantly, the matrix-assisted laser desorption ionization-time of flight mass spectrometry (MALDI-TOF MS) approach can reliably identify the recognized species of Cryptococcus (that may have been cited as genotypes) (1,93,94). Kwon-Chung and coworkers (2) questioned the usefulness of MALDI-TOF MS for the separation of the new species and the hybrids, suggesting that only score values of Ն2.0 indicate a reliable species identification. However, several studies show that yeast and even filamentous fungal isolates can be reliably identified with a score value of Ն1.7 (95)(96)(97), and this is acknowledged in the current Bruker guidelines. The identification of Cryptococcus isolates by MALDI-TOF MS yields comparable results or even outperforms the identification methods used for Candida, Geotrichum, Malassezia, and Trichosporon isolates.
Kwon-Chung and coworkers (2) suggested that methods employed in the 2015 taxonomic proposal are not appropriate because they have been developed for sexually reproducing organisms. One of the first applications of molecular recognition of species was with a fungus that has yet to reveal its sexual morphology, Coccidioides (100). Furthermore, Cryptococcus has a sexual cycle and clearly can reproduce both sexually and asexually. Moreover, the methods used have been applied to identify species-level lineages in asexual taxa (126)(127)(128)(129)(130)(131)(132)(133)(134). Methods using branch length differences to identify thresholds between intra-and interspecific distances (such as the coalescence-based general mixed Yule coalescent method) potentially underestimate species diversity in asexual species, since sexual species are separated by larger genetic gaps than asexual species (135). Individual methods for species delimitation based on molecular data have been shown to either oversplit or underestimate species diversity under specific circumstances (136); understanding the performance of each method is still in its infancy given the recent and rapid development of this field of research. Therefore, three independent approaches were used to delimit species boundaries within the C. neoformans/C. gattii species complexes. In addition, DNA-based approaches were congruent with, for example, MALDI-TOF MS-based data. Sampling of additional loci would certainly be useful, as well as the addition of further genomic data sets. However, studies of other microorganisms repeatedly show that additional loci will either confirm clades found or reveal the presence of new ones. Thus, species delimitation for the seven etiologic agents of cryptococcosis was minimal and conservative (1). Most, if not all, studies that used whole-genome data published before the 2015 taxonomy paper (cited in reference 1), and thereafter, e.g., Farrer and coworkers (36) and those presented at ICCC10 (42,43,(137)(138)(139) identified the same species clades.
The insights that resulted in the 2015 taxonomy proposal (1) were elaborated, presented, and discussed at several related meetings from ICCC4 (London, United Kingdom, 1999) to ICCC10 (Foz do Iguaçu, Brazil, 2017). At ICCC6 (Boston, MA, USA, 2005), a debate entitled "Cryptococcus neoformans: one, two or more species" was held. Two different opinions were presented, namely, for two species or multiple species (at that time, six species). The community strongly supported the name C. neoformans for serotype A strains that are clinically important. The type strain of C. nasalis belongs to serotype D (15); hence, it had nomenclatural priority. However, the community leaders present at ICCC6 to ICCC8 were strongly against the use of this name. Therefore, C. deneoformans was proposed for this clade at ICCC6, as it shows affinity with the epithet neoformans and serotype D (de-neoformans). The name C. gattii received renewed attention, as it was reported as the cause of a number of major outbreaks (35,140,141). The rules of fungal nomenclature do not allow this name to be used for a clade other than the one containing the type strain (and ex-type strain). The clade referred to as AFLP4/VGI represents C. gattii, and the AFLP5/VGIII clade is C. bacillisporus. Three other consistently observed clades in the C. gattii species complex were named using "gattii" in part of the epithet in order to keep reference to the name "gattii." The taxonomy of the species complexes is complicated by various interspecies hybrids (16,20,(142)(143)(144)(145)(146)(147). Hybrids occur among many yeast genera, such as Saccharomyces, where well-recognized species form hybrids and even triple hybrids (147)(148)(149)(150). For Saccharomyces hybrids, a conventional nomenclature has been proposed (150). The species that contribute to the hybrid will be given in alphabetic order, and in cases where the genomic contribution is known, this will be indicated. For instance, the type strain of S. bayanus CBS380 is written as S. cerevisiae Ͻ1% ϫ S. eubayanus 37% ϫ S. uvarum 63%. This convention is also applicable to the genus Cryptococcus. The hybrid type strain of C. neoformans can be thus described as C. deneoformans ϫ C. neoformans.

FOLLOWING THE RULES OF THE INTERNATIONAL CODE OF NOMENCLATURE
The naming of fungi is governed by the International Code of Nomenclature for Algae, Fungi, and Plants, and naming fungi is based on a number of principles (151). Among them, the priority principle implies that the oldest validly given name should be applied to an organism and that the phylogenetic position of the type that determines the name has to be given to a certain clade at a specific taxonomic level. Thus, when a validly described species name exists for a certain species, that name must be used. This was the case for the species that were reinstalled as C. gattii, C. bacillisporus, and in fact also for C. deneoformans (see above).

SUMMARY
The main advantage of recognizing seven species rather than just two "species complexes" (viz., C. gattii sensu lato and C. neoformans sensu lato) is that researchers and clinicians will be stimulated to search for further phenotypic and genetic differences and similarities between the recognized species. This stimulation of research has already yielded new genetic, molecular, and phenotypic features, including differences in drug susceptibility ( Table 1). The recognized species can be identified using a diverse array of molecular diagnostics and MALDI-TOF MS, and some of them can already be identified by phenotypic means. Ignoring the species impedes deciphering the differences among them, which may delay future clinical advances. Finally, it is apparent that more species seem to occur within Cryptococcus, e.g., the Botswana lineage within C. neoformans (18,(137)(138)(139).

ACKNOWLEDGMENTS
V. Arsic Arsenijevic reports research grants and consultation honoraria from Pfizer and received speaker fees from Astellas, Pfizer, and Schering-Plough. O. A. Cornely