Taxonomy of Aspergillus series Versicolores : species reduction and lessons learned about intraspecific variability

Abstract

Aspergillus series Versicolores members occur in a wide range of environments and substrates such as indoor environments, food, clinical materials, soil, caves, marine or hypersaline ecosystems. The taxonomy of the series has undergone numerous re-arrangements including a drastic reduction in the number of species and subsequent recovery to 17 species in the last decade. The identification to species level is however problematic or impossible in some isolates even using DNA sequencing or MALDI-TOF mass spectrometry indicating a problem in the definition of species boundaries. To revise the species limits, we assembled a large dataset of 518 strains. From these, a total of 213 strains were selected for the final analysis according to their calmodulin (CaM) genotype, substrate and geography. This set was used for phylogenetic analysis based on five loci (benA, CaM, RPB2, Mcm7, Tsr1). Apart from the classical phylogenetic methods, we used multispecies coalescence (MSC) model-based methods, including one multilocus method (STACEY) and five single-locus methods (GMYC, bGMYC, PTP, bPTP, ABGD). Almost all species delimitation methods suggested a broad species concept with only four species consistently supported. We also demonstrated that the currently applied concept of species is not sustainable as there are incongruences between single-gene phylogenies resulting in different species identifications when using different gene regions. Morphological and physiological data showed overall lack of good, taxonomically informative characters, which could be used for identification of such a large number of existing species. The characters expressed either low variability across species or significant intraspecific variability exceeding interspecific variability. Based on the above-mentioned results, we reduce series Versicolores to four species, namely A. versicolor, A. creber, A. sydowii and A. subversicolor, and the remaining species are synonymized with either A. versicolor or A. creber. The revised descriptions of the four accepted species are provided. They can all be identified by any of the five genes used in this study. Despite the large reduction in species number, identification based on phenotypic characters remains challenging, because the variation in phenotypic characters is high and overlapping among species, especially between A. versicolor and A. creber. Similar to the 17 narrowly defined species, the four broadly defined species do not have a specific ecology and are distributed worldwide. We expect that the application of comparable methodology with extensive sampling could lead to a similar reduction in the number of cryptic species in other extensively studied Aspergillus species complexes and other fungal genera.