Introduction

Until July 2011 the nomenclature of fungi allowed for the use of multiple names describing asexual and sexual stages of the same fungus (Hawksworth 2011). Prior to the era of sequencing and molecular phylogeny, when only morphological and cultural techniques were used, the association of anamorph and teleomorph life cycles of a given fungus was often established with difficulty. Dual nomenclature was a logical option when morphs of a given fungus were seen to propagate independently, but has become superfluous since the respective stages can be linked with the aid of sequence data. A general agreement was reached among mycologists to officially abolish the dual naming system of fungi, as laid down in the ‘Amsterdam Declaration on Fungal Nomenclature’ and outlined by Hawksworth (2011). The provisions of dual nomenclature, which had been designed especially for fungi, are no longer used. The novel single-name approach certainly has disadvantages in e.g. ecological questions where life cycles and dispersal mechanisms play a role, but will prove to be increasingly useful for (meta)genomic studies, where species are represented by sequence data only.

The present paper discusses the current status of nomenclature of clinically relevant members of the Microascales prevalently producing sticky conidia and, when teleomorphic, cleistothecia with deliquescent asci and smooth-walled, one-celled ascospores. They are currently classified in Scedosporium, Pseudallescheria and related genera. Our aim is to establish a consensus generic naming system for these fungi. In humans, the fungi are primarily responsible for mycetoma and arthritis in healthy individuals, for pulmonary and cerebral infections in near-drowning victims (Tintelnot et al. 2008), and for disseminated infections in patients with severe innate immune disorders or immune suppression (Guarro et al. 2006).

Methods

Index Fungorum (http://www.indexfungorum.org) and MycoBank (http://www.mycobank.org) were used as nomenclatural databases. Original descriptions of the type species of available generic names were analysed. An overview of relevant genera, their type species and their current status is given in Table 1.

Table 1 Generic names of fungi listed in literature as (possibly) belonging to the Scedosporium lineage (Group 1a in Fig. 2) of Microascaceae. Names of accepted genera of the scedosporium-lineage are printed in bold
Full size table

Gross phylogenetic positions were verified on the basis of ribosomal LSU and ITS sequence data. After alignment in BioNumerics v. 4.8 a phylogenetic tree covering Microascales using rDNA D1/D2 Large Subunit sequences was reconstructed from data retrieved from GenBank and CBS databases, using maximum likelihood method implemented in RaxML and applying Tamura 3-parameter model and G + I substitution rate, nearest-neighbour-interchange, with 500 bootstrap replications; >80 % support was considered statistically significant (Fig. 1). Fusarium acutatum CBS 739.97 and F. proliferatum CBS 246.61 were used as out-group taxa. To illustrate gross relationships between genera, an ITS-based tree covering Microascaceae was reconstructed using sequences retrieved from GenBank and CBS databases by alignment with Muscle using the EMBL-EBI web server and handling in BioEdit v. 7.0.9.0 (Fig. 2). Phylogenetic analyses such as Maximum Likelihood and Maximum Parsimony were performed, and the one using neighbour joining with Kimura 2 parameter in the web server above and with the number of bootstraps set at 1,000 was maintained. For detailed information on strains and GenBank numbers see Supplementary Table 1.

figure 1
Fig. 2
figure 2

Phylogenetic overview of Microascaceae based on sequences of ITS region of the rDNA reconstructed with Muscle using the EMBL-EBI web server and handeled in BioEdit v. 7.0.9.0. Analysis was performed using Neighbor joining with Kimura 2 parameter model, midpoint rooting, and with 1000 bootstrap replications. Names with * regarded as ambigious

Full size image

Results and discussion

Six families are currently distinguished within the order Microascales: Ceratocystidaceae, Chadefaudiellaceae, Gondwanamycetaceae, Graphiaceae, and Halosphaeriaceae, in addition to the Microascaceae that encompasses 20 genera with about 200 species. Of the Chadefaudiellaceae and Gwondanomycetaceae only fragmentary sequence data are available, which renders a phylogenetic comparison impossible. The families Graphiaceae and Ceratocystidaceae are monophyletic and supported by high bootstrap values. The Halosphaeriaceae form a separate clade and has been suggested to represent a separate order (Sakayaroi et al. 2011). On the basis of LSU data (Fig. 1) the remaining species form a large cluster (89 % bootstrap support) currently classified as Microascaceae.

Phenotypically, two approximate groups can be distinguished within the family Microascaceae: species with prevalently scedosporium-like anamorphs with slimy conidia, and species with prevalently scopulariopsis-like anamorphs with hygrophobic conidia; Petriellidium desertorum has an exceptional type of anamorph reproduction by arthroconidia (von Arx 1973). However, these phenotypic groups lack statistical supported in LSU data (Fig. 1). Also some subgroups, i.e. a cluster containing Wardomyces and relatives (group 2), one with prevalently Microascus species (group 3) and one with prevalently Scopulariopsis species (group 4) have low support. Species identified as Kernia with Enterocarpus take an intermediate position between the phenotypic series, except for K. hyalina, which clusters among Scopulariopsis species. Most Kernia species have hygrophobic, scopulariopsis-like anamorphs, but K. hippocrepida produces slimy conidia similar to those prevalent in the Scedosporium lineage (Malloch and Cain 1971). Thus the phenotypic separation of two lineages within Microascaceae on the basis of conidial type is not unambiguous, which is in line with the absence of support in the phylogenetic tree.

The ITS tree is suitable to construct an overview of genera within the family Microascaceae, with accent on the scedosporium-lineage ‘group 1’ (Fig. 2). The main clade (95 % bootstrap support), includes Enterocarpus, Lophotrichus, Parascedosporium, Petriella, Petriellopsis, and Scedosporium, The Wardomyces relationship (group 2) has 99 % bootstrap support, the genus Microascus with purported Scopulariopsis anamorphs splits into several entities (groups 3 and 4), and groups with Doratomyces and Kernia species are paraphyletic to the Scedosporium lineage. Lineage 1a (95 % bootstrap support) is the group of taxa with prevalently scedosporium-like anamorphs, and the focus of the present paper. Thus, species with slimy conidia form a derived cluster, but do not unambiguously segregate into a separate taxonomic group, confirming LSU data above. The lineages with scedosporium-like and scopulariopsis-like species should therefore remain to be classified in the single family Microascaceae (Table 1).

Within the unsupported ‘group 1’ of Figs. 1 and 2, Scedosporium is the oldest valid generic name, published in 1919 (Castellani and Chalmers 1919) with the type species Scedosporium apiospermum (Table 1). Main generic names that have been applied to the dimorphic genus Pseudallescheria/Scedosporium are the following: Allescheria Sacc. & P. Syd. 1899, Allescheria R. Hartig 1899, Monosporium, Pseudallescheria, Scedosporium, and Petriellidium. Based on the identity of type species of each genus, the genera Allescheria R. Hartig, Allescheria Sacc. & P. Syd. and Monosporium, can be excluded, as these are either invalid or represent entirely different fungi (Table 1). For reasons of priority the generic name Pseudallescheria (1944; Table 1) should be treated as a synonym of Scedosporium. This deviates from the proposed list of protected generic names under the Melbourne Code of Botanical Nomenclature (Kirk et al. 2013). It implies transfer of Pseudallescheria minutispora to Scedosporium. In addition to the latter species, the generic name Scedosporium applies to S. apiospermum, S. aurantiacum, S. boydii, and S. dehoogii. Scedosporium deficiens (Rainer and Kaltseis 2010) was described without indication of a type specimen and is therefore invalid. The taxonomic status of P. angusta, P. ellipsoidea, and S. deficiens, is still under investigation (Lackner) and further name changes are therefore considered to be premature. In our ITS data (Fig. 2), based on a small number of strains per species, S. apiospermum and S. boydii are clearly different, but the two groups are known to form a widely variable cloud of strains without medically relevant differences in e.g. pathology or antifungal susceptibility between entities (Lackner et al. 2014). In routine identification in the clinical laboratory, these molecular siblings might be taken together as a ‘complex’ (de Hoog et al. 2013), which in this case should be indicated as the ‘Scedosporium apiospermum complex’. Petriella is found in two separate clusters. With LSU data the genera Lomentospora, Parascedosporium, and Petriellopsis are part of Petriella, but with ITS unambiguous distinction is achieved (Fig. 2).

In the ITS tree (Fig. 2) a polytomy is observed of Parascedosporium (100 % bootstrap support), S. prolificans (99 % bootstrap support) and two Petriella groups (100 % bootstrap support). Within group 1a (Fig. 2), Scedosporium prolificans is separate from the remaining Scedosporium species, as noted earlier by Issakainen et al. (Issakainen et al. 1999; Issakainen et al. 2003). The species clusters close to Petriella setifera, which also produces scedosporium-like anamorphs, i.e. one-celled, subhyaline, mucoid conidia produced by annellidic conidiogenesis. The isotype culture of Petriella asymmetrica, CBS 258.31—the type species of Petriella—has been classified in Melanospora (von Arx and Müller 1954), but sequence data showed that they are identical with the older species Petriella sordida, which hence has been listed as generic type of Petriella (Guarro et al. 2012). The position of Petriella setifera outside Petriella sensu stricto needs further attention. For S. prolificans the oldest generic name is Lomentospora (Hennebert and Desai 1974). Its significant phylogenetic distance from the remaining Scedosporium species is underlined by its microscopic morphology: (a) inflated versus tubular conidiogenous cells in verticillate arrangement, and (b) different colony texture and coloration. The distinction of S. prolificans from the remaining Scedosporium species is meaningful for clinical practice, because in contrast to Scedosporium (c) the species is highly resistant to voriconazole (Lackner et al. 2012). Moreover, (d) S. prolificans causes a different disease spectrum, with primarily disseminated infections in immunocompromised patients versus occurrence of the Scedosporium apiospermum complex being associated with near-drowning syndrome (Guarro et al. 2006), and as causative agent of eumycotic mycetomas. In animal models (e) S. prolificans was shown to be more virulent (Ortoneda et al. 2002). As such, the separation of Lomentospora from Scedosporium is supported from an evolutionary as well as from a medical perspective. The recommended name for S. prolificans is therefore Lomentospora prolificans.

Many species of lineage 1a (Fig. 2) are polymorphic, strains exhibiting differentially named morphological stages: Scedosporium for a hyphomycetous anamorph, ‘Graphium’ for a synnematous anamorph, and Pseudallescheria for the teleomorph. Species that regularly exhibit a teleomorph have been described in the respective teleomorph genus (e.g., Pseudallescheria minutispora), whereas close relatives without known teleomorph are known under their anamorph genus name (e.g., Scedosporium aurantiacum). An overview of the presence or absence, as well as of the names of anamorphic, synanamorphic and teleomorphic stages is given in Table 2. With the priority of Scedosporium over Pseudallescheria, this name will now be applied for the holomorph. It may be useful, however, to maintain the former anatomic system to indicate ecologically different types of dispersal. The pseudallescheria-stage (indicated with roman, non-capitalized letters) observed in many species of Scedosporium serves sexuality, and a graphium-like stage has been speculated to have another vector of dispersal (Dowding 1969; Guarro 2012). It should be noted, however, that the genus Graphium sensu stricto is a member of the family Graphiaceae and thus unrelated.

Table 2 Proposed generic and specific names and subsequent synonyms in Scedosporium and related genera
Full size table

In conclusion, Lomentospora, Lophotrichus, Parascedosporium, Petriella, Petriellopsis, and Scedosporium will remain the correct generic names in lineage 1a prevalently covering members of Microascaceae with slimy conidia. Pseudallescheria minutispora and Scedosporium prolificans need to be renamed. The ambiguous taxonomic validity of Pseudallescheria angusta as an intermediate between S. apiospermum and S. boydii (Lackner et al. 2014) to be the result of an ongoing process of sympatric speciation and hence difficult to classify. Based on the above, the following name changes are proposed:

Scedosporium minutisporum (Gilgado et al.) Lackner & de Hoog, comb. nov., MycoBank MB 807326

Basionym: Pseudallescheria minutispora Gilgado et al.—J. Clin. Microbiol. 43: 4938, 2005.

Scedosporium desertorum (v. Arx & Moustafa) Lackner & de Hoog, comb. nov., MycoBank MB807329

Basionym: Petriellidium desertorum v. Arx & Moustafa—Persoonia 7: 371, 1973 ≡ Pseudallescheria desertorum (v. Arx & Moustafa) McGinnis et al.—Mycotaxon 14: 98, 1982.