Studies in the Stypella vermiformis group (Auriculariales, Basidiomycota)

Stypella vermiformis is a heterobasidiomycete producing minute gelatinous basidiocarps on rotten wood of conifers in the Northern Hemisphere. In the current literature, Stypella papillata, the genus type of Stypella (described from Brazil), is treated as a taxonomic synonym of S. vermiformis. In the present paper, we revise the type material of S. papillata and a number of specimens addressed to S. vermiformis. As a result, the presumed synonymy of S. papillata and S. vermiformis is rejected and the genus Stypella is restricted to the single species S. papillata. Morphological and molecular phylogenetic studies of specimens from the Northern Hemisphere corresponding to the current concept of S. vermiformis uncovered three species from two newly described genera. S. vermiformis s.str. is distributed in temperate Europe and has small-sized basidia and basidiospores, and it is placed in a new genus, Mycostilla. Another genus, Stypellopsis, is created for two other species, the North American Stypellopsis farlowii, comb. nov., and the North European Stypellopsis hyperborea, sp. nov. Basidia and basidiospores of Stypellopsis spp. are larger than in Mycostilla vermiformis but other morphological characters are very similar. In addition, Spiculogloea minuta (Spiculogloeomycetes, Pucciniomycotina) is reported as new to Norway, parasitising basidiocarps of M. vermiformis and Tulasnella spp.


Introduction
The Heterobasidiomycetes are an artificial group of fungi encompassing basidiomycetous taxa with septate basidia and (or) repetitive basidiospores (Weiss et al. 2004a). Their traditional division was based on features of basidial septation and sterigmata (Tulasne and Tulasne 1873) and survived almost unchanged for over a century. First studies on cell ultrastructure (Bandoni 1984) and then DNA studies (Fell et al. 2000;Weiss and Oberwinkler 2001;Bauer et al. 2006) questioned the reliability of these morphological characters for a higher-level taxonomy of basidiomycetes. Currently, the class and order arrangement of heterobasidiomycetes is more or less well-established (summarized in McLaughlin and Spatafora 2014). However, their lower-level taxonomy remains unstable, often due to the lack of sequence data or reference material Wang et al. 2015). In many cases, this is also a result of a high morphological similarity of taxa appearing unrelated or only distantly related in phylogenetic studies (Millanes et al. 2011;Liu et al. 2015;Wang et al. 2015). An utmost case of this similarity is dealt with in the present paper.
The genus Stypella was described by Möller (1895) from Brazil for two newly introduced species with four-celled basidia, Stypella papillata and Stypella minor. It was not in use until Donk (1958Donk ( , 1966 accepted it and selected S. papillata as the generic type. Next, Martin (1934Martin ( , 1952 synonymized the European species Heterochaetella crystallina Bourdot with S. papillata. It was shown later that Dacrymyces vermiformis Berk. & Broome was an older name for H. crystallina and S. papillata as they were understood at that time (Reid 1974). Therefore, Reid (1974) created a new combination, Stypella vermiformis (Berk. & Broome) D.A. Reid, and placed both H. crystallina and S. papillata among its synonyms. This opinion about identity of S. vermiformis has been widely accepted in the modern taxonomy of heterobasidiomycetes (Oberwinkler 1982;Reid 1990;Roberts 1998).
In the protologues of S. papillata and S. minor, Möller (1895: 76) described and illustrated peculiar basidia; in his interpretation, each basidium is a terminal part of a much narrower hyphal segment. Studies of basidial morphology in Exidia nucleata (Schwein.) Burt showed that the hypha-like segment bearing four terminal basidial cells represents an integral part of the basidium in that species-the socalled 'enucleate stalk' (Wells 1964). Various opinions on the taxonomic value of this feature persist in the mycological literature. In particular, Donk (1966) accepted it as an important generic character although he refused to include all taxa with stalked basidia into one genus and considered them as belonging to four genera, Myxarium, Stypella, Protodontia and Heterochaetella (with a few species among incertae sedis). On the contrary, Roberts (1998) placed all species with effused basidiocarps and stalked basidia into a redefined Stypella, whereas he retained the type species of Myxarium, M. nucleatum Wallr., in Exidia. However, none of these authors studied the authentic material of S. papillata.
Recent molecular phylogenetic studies (Weiss and Oberwinkler 2001;Wells et al. 2004;Spirin et al. 2017) mostly confirm the generic splitting advocated by Donk (1966). Therefore, the genus Stypella has been limited to the single species, S. vermiformis (= S. papillata sensu auct.). At the same time, stalked basidia are observed in some other genera of Auriculariales (Wells and Raitviir 1980;Weiss and Oberwinkler 2001;Malysheva and Spirin 2017). Thus, the taxonomic significance of this character may have been overemphasized. In the present paper, material of Stypella vermiformis from Northern Hemisphere is revised based on morphology and DNA data, and new information about S. papillata is provided.

Morphological study
Collections and type specimens from several herbaria have been studied: University of Helsinki, Finland (H); Botanical Museum of the University of Oslo, Norway (O); University of Hamburg, Germany (HBG); University of Gothenburg, Sweden (GB); Farlow Herbarium at Harvard University, USA (FH); National Museum of Natural History, France (PC), and the private herbarium of Heikki Kotiranta in Helsinki, Finland (H.K.). Herbarium acronyms are given according to Thiers (2018). Morphological study follows Miettinen et al. (2012). The abbreviations used in microscopic descriptions are: L-mean basidiospore length, W-mean basidiospore width, Q-mean L/W ratio, n-number of basidiospore measurements per specimens studied.

DNA extraction and sequencing
For DNA extraction, small fragments of dried basidiocarps were used. Extractions were done using the NucleoSpin Plant II Kit (Macherey-Nagel GmbH and Co. KG, Düren, Germany) following the manufacturer's instructions. PCR amplification and sequencing of the nrITS region was performed using primers ITS1F (Gardes and Bruns 1993) and ITS4 (White et al. 1990). Primers JS1 (Landvik 1996) and LR5 (Vilgalys and Hester 1990) were used to amplify and sequence approximately 700 bp of nrLSU region. Sequencing was performed with an ABI model 3130 Genetic Analyzer (Applied Biosystems, CA, USA). Raw data were edited and assembled in MEGA 6 (Tamura et al. 2013).

Phylogenetic analyses
For this study, eight nrITS and seven nrLSU sequences were generated. In addition to the sequences published here, 4 nrITS sequences and 43 nrLSU sequences were retrieved from GenBank (www.ncbi.nlm.nih.gov/ genbank/). Sequences were aligned with the MAFFT version 7 web tool (http://mafft.cbrc.jp/alignment/ server/) using the Q-INS-i option for nrITS and nrLSU. Before the phylogenetic analyses, the best-fit substitution models for the alignments (GTR) were estimated based on the Akaike Information Criterion (AIC) using FindModel web server (http://hiv.lanl. gov/content/sequence/findmodel/findmodel.html).
Two different phylogenetic analyses were performed for the nrLSU dataset: (1) Maximum likelihood (ML) analyses were run on the PhyML server v.3.0 (Guindon et al. 2010), with 100 rapid bootstrap (BS) replicates; (2) Bayesian inference analyses (BI) were run using MrBayes 3.2.5 software (Ronquist and Huelsenbeck 2003) for 5 million generations, under a GTR model, with four chains, and trees sampled every 100 generations. To check for convergence of MCMC analyses and to get estimates of the posterior distribution of parameter values, Tracer v1.6 was used (Rambaut et al. 2014). In total, 100,002 trees were read. Credible sets of trees contained 30,093 trees sampled. Burn-in was 1000 iterations. We accepted the result where the ESS (Effective Sample Size) was above 200 and the PSRF (Potential Scale Reduction Factor) was close to 1. For the nrITS dataset, we only performed ML analyses, again on the PhyML server v.3.0. Newly generated sequences have been deposited in GenBank (Table 1)

Results and discussion
In total, 40 specimens from Europe, East Asia and North America corresponding to the current concept of S. vermiformis (sensu Reid 1990;Roberts 1998) were selected for morphological study. Basidiocarps are initially represented by minute, sharp-pointed, gelatinous outgrowths irregularly arranged on an extremely thin joint subiculum. These outgrowths quickly fuse together and produce compound resupinate basidiocarps of a very characteristic, reticulate appearance (observable under lens) (Fig. 1). The basidiocarp core consists of several tubular, bluntpointed or tapering, thin-walled cystidia up to 220 lm long. These giant cystidia are often glued together and covered by sparsely arranged hyphae, basidia and occasionally by cystidia-like cells of smaller size (gloeocystidia). Basidiospores are rather small-sized, broadly ellipsoid to subglobose, repetitive, ca. 3.5-7 9 3-6 lm (Fig. 2). All studied specimens were collected from coniferous wood at advanced decay stages in temperate-boreal forests in the Northern Hemisphere.
The authentic specimen of S. papillata (designated as a lectotype in Specimens examined) was detected in HBG by Friedrichsen (1977) and studied by us. This fungus is widely different from S. vermiformis. Its fructifications are continuous, not reticulate, and up to 0.5 mm thick. Microscopically, it is characterized by slightly or distinctly thick-walled and totally clampless hyphae (as illustrated in Möller's original figure) and a presence of globose conidiiferous cells. These cells are arranged in chains connecting with each other by hypha-like outgrowths. Walls of some conidiophores are somewhat shrunken and, if observations are made in normal light and without colouring medium, they make an impression of the inner cell septation (Fig. 2). However, using phase contrast illumination and coloured mountant (Cotton Blue) reveals it merely as an artifact. ''Spores'' described by Möller (1895) in the protologue of S. papillata are in fact broadly ellipsoid or subglobose, slightly thick-walled conidia located on short terminal projections of some conidiiferous cells. A few cystidia-like elements have been Kirschner, also possess clamped hyphae (Kirschner and Chen 2004;Kirschner et al. 2010). The only species with clampless hyphae confirmed as a member of the Auriculariales is Endoperplexa enodulosa (Hauerslev) P. Roberts (Weiss et al. 2004b); however, it is not reminiscent of S. papillata either. Considering these observations, we reject the synonymy of S. papillata and S. vermiformis s.l. The identity of S. papillata should be re-established based on newly collected and sequenced material from the locus classicus.
Eight collections of Stypella vermiformis s.l. were selected for DNA study, and two datasets were assembled for phylogenetic analyses: 1. A nrLSU phylogeny of the Auriculariales (Fig. 3a). The final aligned dataset included 863 characters (including gaps). The overall topologies of the ML and BI trees were nearly congruent. Specimens of S. vermiformis s.l. ended up in two distantly related clades within the order: A. Stypella vermiformis s.str. clade (BS = 100, pp = 1) appeared as a sister group of Pseudohydnum gelatinosum although this     b the best tree from the ML analyses of the nrITS dataset. Voucher numbers are given for newly sequenced specimens and accession numbers for additional sequences retrieved from GenBank. Scale bar shows expected changes per site Auriculariales may vary from five (Eichleriella shearii-E. macrospora complex) to fifteen base pairs (Heteroradulum deglubens-H. kmetii complex) . Therefore, we found the aforementioned genetic difference between S. farlowii and S. hyperborea sufficient enough to introduce them as separate taxa. Their descriptions are presented below.
Taxonomic changes proposed in our study are summarized in Table 2.
Remarks. Mycostilla vermiformis is distributed in temperate forests of Europe (Svrček 1950;Hauerslev 1976;Wojewoda 1981;Reid 1990;Roberts 1998). Reid (1974Reid ( , 1990) studied types of both Dacrymyces vermiformis and Heterochaetella crystallina and concluded that they were conspecific. Smaller basidia and basidiospores help to distinguish M. vermiformis from the similarly looking Stypellopsis species. Brownish colouration of cystidial content in KOH can help in identification of recent Stypellopsis spp. samples; this reaction is absent in all specimens of M. vermiformis studied by us. Attention should be paid to the age and collecting data of a specimen, nonetheless. We studied and sequenced one specimen of M. vermiformis (O  F188059) with abnormally large basidia, 7.7-11.2 9 6.1-8.1 lm. This basidial size might point towards Stypellopsis spp. However, other microscopic structures of this collection, as well as ITS sequence, are identical to other specimens of M.
vermiformis. This morphological deviation can possibly be explained as due to unusual fructification time (the specimen was collected in March).
One specimen of M. vermiformis from Norway was infected by Spiculogloea minuta P. Roberts (Spiculogloeomycetes, Pucciniomycotina), and this is the first record of this mycoparasite in the country. Spiculogloea minuta was originally described as a parasite of Tulasnella violea (Quél.) Bourdot & Galzin (Roberts 1997
Remarks. Burt (1919) mistook shallow pits of a compound, reticulate basidiocarp of this species for minute pores and thus described it as a poroid fungus (see also Martin 1952). We made recent collections of this species from North America, which are identical to the type material as described by Burt (1919). Due to scantiness of the P. farlowii holotype, we decided to designate an epitype here. Basidiospores of Stypellopsis farlowii are on average larger than in M. vermiformis and smaller than in S. hyperborea. Luck-Allen (1960) reported a number of collections from Northeastern USA and Canada identified as Stypella papillata. In fact, they may belong to Stypellopsis farlowii.
Remarks Stypellopsis hyperborea possesses the largest basidiospores of the species complex. It seems to be a truly boreal species distributed in coniferous forests of North Europe. The description and microscopic drawing of Stypella vermiformis by Strid (1986) refer to this species.