Taxonomic study of polymorphic basidiomycetous fungi Sirobasidium and Sirotrema: Sirobasidium apiculatum sp. nov., Phaeotremella translucens comb. nov. and rediscovery of Sirobasidium japonicum in Japan

Species in the genera Sirobasidium and Sirotrema (Tremellales, Tremellomycetes, Agaricomycotina, Basidiomycota) have been described based solely on the morphology of teleomorph, and many of them lack both isolates of anamorphic yeast state and nucleotide sequence data. Strains of Sirotrema translucens and Sirobasidium japonicum were established for the first time from basidiocarps collected in Japan. Also, an undescribed species in the genus Sirobasidium was isolated. Sirobasidium sp. was characterized by its apiculate epibasidia and 2-celled basidia divided by a longitudinal septum, which is a unique combination of characteristics in the genus. Although the phylogenetic placement of Sb. japonicum within the Tremellales was not resolved in our analysis, Sirobasidium sp. formed a well-supported monophyletic clade with Sb. magnum and Fibulobasidium spp., and Sirotrema translucens was located in the genus Phaeotremella. Mating experiments using single-basidiospore strains showed that Sb. japonicum produced basidia, epibasidia, and basidiospores on a nutrient-poor medium, and the life cycle was successfully completed in controlled conditions. In conclusion, we propose Sirobasidium apiculatum sp. nov. and Phaeotremella translucens comb. nov.


Introduction
The genera Sirobasidium and Sirotrema are basidiocarp-forming fungal taxa belonging to the order Tremellales (Tremellomycetes, Agaricomycotina). The genus Sirobasidium is characterized by the morphological characteristics: basidia formed in chains and deciduous primary spores defined as epibasidia (Bandoni 1957). Basidiocarps of Sirobasidium spp. usually closely associate with pyrenomycetes, especially Xylariales, suggesting that they are mycoparasites (Roberts and Meijer 1997;Chen 1998). A parasitic structure called tremelloid haustorium (= haustorial branch) is common in the Tremellales (Bandoni 1987;Grube and de los Ríos 2001). However, as the structure was reported only one time in the genus Sirobasidium (Bandoni 1987;Bandoni et al. 2011), the parasitic nature of Sirobasidium species remains unclear. There are around 10 described species in the genus, but only two species have been cultured and few nucleotide sequences data are available. Based on the morphological features, there have been many discussions about the species delimitation or their taxonomic position in higher rank for the species without strains (Dämon and Hausknecht 2002;Roberts and Spooner 1998). For instance, transversally septate basidia of Sirobasidium japonicum Kobayasi caused the suspicion that the species is closely related to another order, Auriculariales (Agaricomycetes; Dämon and Hausknecht 2002).
The other genus, Sirotrema is characterized by an occasional formation of basidia in chains and parasitizing rhytismataceous fungi via tremelloid haustoria (Bandoni 1986). Based on its similarity in basidial ontogeny, Bandoni (1986) speculated that Sirotrema is a close relative of the genus Sirobasidium. Because the genus Sirotrema lacks both strains and nucleotide sequence data, the phylogenetic relationship between these two genera has been unknown for years.
In more recent studies, however, the sequences of Sirobasidium spp., including those obtained in this study and already published, were added to the dataset of phylogenetic analyses, supporting their attribution to Tremellales (Liu et al. 2015;Kachalkin et al. 2019;Li et al. 2020). The sequence of Sirotrema obtained in this study has also been included in the above phylogenetic analyses, supporting the attribution to the genus Phaeotremella (Kachalkin et al. 2019;Li et al. 2020). On the other hand, information on their morphology, sexual reproduction, and host interactions is not sufficient. In this study, we established strains and intensively observed morphological features of Sirobasidium japonicum, one undescribed species of Sirobasidium, and Sirotrema translucens collected in Japan. Based on the phylogenetic analysis, morphological observations, and physiological characteristics, we discuss the taxonomic status of the species. In addition, mating experiments using singlebasidiospore strains were performed for each species, and the results are reported.

Morphological observations, specimens, and strains
Basidiocarps growing on fallen branches or pine needles were collected in the middle and the southern part of Japan (Table 1). To study microscopic characteristics, handmade sections were mounted in sterile water or 3% (w/v) potassium hydroxide stained with or without phloxine. The sections were observed under a Zeiss Axioskop Microscope (Carl Zeiss, Oberkochen, Germany) or a BX53 upright microscope (Olympus, Tokyo, Japan). For the comparison, a herbarium specimen of Sirobasidium japonicum (TNS-F-196751) was observed in the same way. Single-basidiospore isolates of each taxon were established following Wong et al. (1985) using Malt Agar (MA; Nissui, Tokyo, Japan). Morphological features of the anamorphic yeast states were observed according to Kurtzman et al. (2011). Yeast cells for these observations were cultivated on 5% (w/v) malt extract agar (MEA; Kurtzman et al. 2011) at 25 °C for 3 days. Dried specimens of basidiocarps were deposited in the National Museum of Nature and Science, Tokyo (TNS), and strains were in the Japan Collection of Microorganisms (JCM), the Westerdijk Fungal Biodiversity Institute (formerly known as CBS), and the Portuguese Yeast Culture Collection (PYCC) as shown in Table 1.

Physiological tests
Physiological tests of the strains were performed according to Kurtzman et al. (2011). Assimilation of carbon and nitrogen compounds were tested using liquid and solid media, respectively. Growth at various temperatures were tested in YM broth (Difco).

DNA sequencing and phylogenetic analysis
Genomic DNAs were extracted according to Ishida et al. (1999). The D1/D2 region of large subunit ribosomal RNA gene (LSU D1/D2), ITS-5.8S ribosomal RNA gene (ITS), and small subunit ribosomal RNA gene (SSU) were amplified by PCR with Ex Taq (Takara Bio, Otsu, Japan). The LSU D1/D2 was amplified using a primer pair of LR0R (Rehner and Samuels 1994) and LR5 (Vilgalys and Hester 1990), and a pair of ITS1F (Gardes and Bruns 1993) and ITS4 (White et al. 1990) was used for the ITS region.
For the SSU, a primer pair NS1 (White et al. 1990) and NS8 (White et al. 1990) was used. PCR amplifications were carried out as follows: for the primer pairs LR0R/LR5 and ITS1F/ITS4, initial denaturation at 94 °C for 3 min, followed by 30 cycles of 94 °C for 1 min, 51 °C for 30 s, 72 °C for 1 min, and then final extension at 72 °C for 15 min. For the primer pair NS1/NS8, initial denaturation at 94 °C for 2 min followed by 10 cycles of 98 °C for 10 s, 55 °C for 30 s dropping by 0.5 °C per cycle, and extension at 72 °C for 2 min. Those 10 cycles were followed by 25 cycles of 98 °C for 10 s, 50 °C for 30 s, 72 °C for 2 min. The PCR products were purified by polyethylene glycol precipitation. Sequence reactions containing BigDye terminator v3.1 (Applied Biosystems) and primers, namely LR0R and LR5 for the LSU D1/ D2, ITS1F, and ITS4 for the ITS region, and NS1, NS3, NS5, and NS8 for the SSU, were carried out following manufacturer's instructions. DNA sequences were analyzed with ABI PRISM 3130 Genetic Analyzer (Applied Biosystems). The dataset used for molecular phylogenetic analysis is shown in Table S1. Sequences were aligned using the online version of MAFFT v.7.490 (Katoh and Standley 2013;Katoh et al. 2019) and ambiguous sites were manually trimmed in SeaView v.5.0.4 (Gouy et al. 2010) after using Gblocks v.0.91b (Castresana 2000), allowing less strict flanking positions. The alignments of three genes (SSU, ITS, and LSU D1/D2) were concatenated in SeaView, and maximum likelihood (ML) analysis was performed using IQ-TREE v.2.1.2 (Minh et al. 2020). The final alignment comprised of 1480 (SSU), 220 (ITS), and 426 (LSU D1/D2) sites. Among them, 284 (SSU), 89 (ITS), and 173 (LSU D1/D2) sites were parsimonyinformative, and 1047 (SSU), 108 (ITS), and 192 (LSU D1/D2) sites were invariable. The concatenated dataset was partitioned by each gene and the best-fit model of nucleotide substitution for each partition based on the Bayesian information criterion (BIC) was estimated by ModelFinder (Kalyaanamoorthy et al. 2017) as implemented in IQ-TREE. The applied substitution models for ML analysis were as follows: TN + F + R3 for SSU, TIM2e + I + I + R3 for ITS, and TIM3e + I + I + R4 for LSU D1/D2. Branch support was assessed using 1,000 replicates of an ultrafast bootstrap approximation approach (Hoang et al. 2018) in combination with a nonparametric Shimodaira-Hasegawa approximate likelihood-ratio test (SH-aLRT). Tree topology was constrained with the well-supported (> 85%) bipartitions in the sevengene-based phylogenetic tree inferred by Li et al. (2020).

Mating experiment of yeast cells
Intraspecific mating experiments were performed to confirm the mating type. In each species, singlebasidiospore strains were established from the same basidiocarp. Pairs of the strains were inoculated on conjugation medium (CJM; Flegel 1981) and kept at 25 °C. Method for the inoculation was according to Flegel (1976), and inoculated cells were covered with flame-sterilized coverslips to reduce yeast cell growth and promote conjugation of yeast cells (Flegel 1981). One month after incubation, mycelia were cut off and inoculated onto weakly nutrient medium (wMY; Spiegel 1990) and kept at 25 °C to promote basidia and basidiocarp production. Those cultivations were directly observed under a light microscope by covering the surface of the media with coverslips, or mounted in phloxine with 3% potassium hydroxide or lacto-cotton blue.

Results and discussion
Isolation and morphological observation We collected 4 and 2 specimens of Sirobasidium and Sirotrema, respectively. Their collecting sites, herbarium numbers, and single-basidiospore derived strains are listed in Table 1. According to the morphological observation, 3 specimens of Sirobasidium were identified as Sb. japonicum Kobayasi. The other specimen of the genus, however, could not be designated to any previously described species.
In newly collected specimens of Sirobasidium japonicum, microscopic features (morphology of basidia, epibasidia, and basidiospores) were fitted with the original description by Kobayasi (1962). Basidiocarps were closely associated with ascomata of Biscogniauxia spp. (Table 1; Fig. 1a), which has not been reported before.
The specimens of the genus Sirotrema were identified as Sirotrema translucens (H.D. Gordon) Bandoni. Their basidiocarps parasitized Lophodermium conigenum (Brunaud) Hilitzer growing on fallen needles of Japanese red pine (Pinus densiflora Siebold et Zucc.; Fig. 3a). Their microscopic features agreed well with the descriptions by Gordon (1938), Reid and Minter (1979), and Bandoni (1986). Formation of basidia in chains was not detected, which also agreed with observations in the previous studies (Gordon 1938;Reid and Minter 1979).
The descriptions of each species are given in the taxonomy section.

Phylogeny
For the strains of Sirobasidium japonicum, Sirobasidium sp., and Sirotrema translucens, SSU, ITS, and LSU D1/D2 were sequenced (Table. 1). It should be noted that these sequences have already been publicly available for several years and have been added to the dataset of phylogenetic analyses in some studies (e.g., Liu et al. 2015;Kachalkin et al. 2019;Li et al. 2020). In our constrained ML analysis based on a concatenated (SSU-ITS-LSU D1/D2) dataset, Sirobasidium sp. (as "Sb. apiculatum") formed a well-supported clade with Fibulobasidium spp. (Fig. 4), which is characterized by basidia produced by the expansion of clamp connection and forming cluster (Bandoni 1979), in consistency with previous studies (Kachalkin et al. 2019;Li et al. 2020). In addition, our analysis also agrees with the previous study (Kachalkin et al. 2019) in that Sb. magnum was resolved as a monophyletic group with the clade described above (Fig. 4). However, there was no agreement among studies for Sb. japonicum and our analysis did not resolve its phylogenetic position within Tremellales as in Li et al. (2020), whereas Kachalkin et al. (2019) found Sb. magnum, Sirobasidium sp., and Fibulobasidium spp. formed a moderately supported clade. Further sampling of closely related species and more multiple DNA markers including protein-coding loci will be needed to resolve phylogenetic relationships of Sb. japonicum in Tremellales. For St. translucens (as "Phaeotremella translucens"), there was strong support for its location within the genus Phaeotremella (Fig. 4).

Physiological tests
The results of assimilation, fermentation, and other physiological tests were shown in Table 2. All the tested species did not ferment glucose. Two Sirobasidium species differed in the assimilation abilities of maltose, raffinose, melezitose, L-arabinose, D-ribose, L-rhamnose, ethanol, methyl-α-D-glucoside, D-glucuronic acid, and N-acetyl-D-glucosamine as a sole carbon source.

Mating experiments
As a result of the mating experiments on CJM, 9 tested strains of Sirobasidium japonicum formed true  (Table 3). The hyphae, which germinated from single yeast cells, also had clamp connections (Fig. 1g). Hyphal formation after conjugation of two yeast cells was detected only in some pairs (Table 3; Fig. 1h). After transferring the mycelia to wMY agar plates, basidia (Fig. 1i), epibasidia, and basidiospores ( Fig. 1j) were produced from the inoculations in which conjugation of yeast cells was detected on CJM (Table 3). These compatible pairs could be categorized into 4 groups ( Table 3), suggesting that the species has a tetrapolar mating system as in Sb. magnum and other basidiomycetes (Flegel 1976;Raudaskoski and Kothe 2010).
In Sirobasidium sp., we performed a mating experiment using 8 single-basidiospore strains. Conjugation of yeast cells was not detected in all the pairs or single inoculations during one month on CJM or wMY. All the inoculations, however, produced true hyphae with false clamp and tremelloid haustoria (Fig. 2h, i). The results suggest that the true hyphae with false clamps derived from single-cell growth, and the strains would require other conditions to differentiate sexual reproduction.
In Sirotrema translucens, 8 single-basidiospore strains were crossed. In all the pairs or single inoculations, neither conjugations of yeast cells, formation of basidia nor hyphal growth were detected on both CJM and wMY during one month. As in Sirobasidium sp., cultural conditions would not be suitable for the species to induce sexual reproduction.

Taxonomy
The ML analysis demonstrated that Sirobasidium sp. is a novel lineage in the order Tremellales. Furthermore, as its morphology is different from any species in the genus Sirobasidium, a new taxon should be established. Whereas Sirobasidium sp. formed a well-supported monophyletic clade with Sb. magnum and Fibulobasidium spp., Sb. japonicum did not form a monophyletic lineage with other available Sirobasidium sequences (i.e., Sb. intermedium and Sb. magnum). Many studies have shown that Sb. intermedium was phylogenetically separated from the other Sirobasidium species (e.g., Boekhout et al. 2011;Millanes et al. 2011;Liu et al. 2015;Kachalkin et al. 2019;Li et al. 2020; this study), and thus the genus is polyphyletic. In a phylogenetic tree in Liu et al. (2015), Sb. intermedium CBS 7805 formed a clade with another Sirobasidium species named Sb. brefeldianum AM71. However, since their sequences have the same origin represented by a single isolate (see Millanes et al. 2011), its phylogenetically isolated position needs to be confirmed by new isolates, as pointed out by Boekhout et al. (2011). Unfortunately, the phylogenetic position of the type species of the genus Sirobasidium, Sb. rubrofuscum (Berk.) P. Roberts (syn. Sb. sanguineum Lagerh. et Pat.;Dämon and Hausknecht 2002), is unknown, making it difficult to discuss Sirobasidium s. str. Therefore, we here follow the morphological definition of Sirobasidium and draw taxonomic conclusions.
Sirotrema translucens has been suggested to be related to the genus Sirobasidium because its basidia are sometimes formed in chains (Bandoni 1986), but the results of the ML analysis did not support this idea and strongly supported the assignment to the genus Phaeotremella (Fig. 4), which was recently emended from the 'foliacea clade' recognized in the genus Tremella (Liu et al. 2015). In St. translucens, both morphological features of teleomorph and physiological traits of anamorph agree with those of the genus Phaeotremella (Liu et al. 2015). Therefore, it is reasonable that the species is transferred to the genus Phaeotremella.
Notes: Although Sb. apiculatum was associated with Eutypella scoparia, we could not observe tremelloid haustoria in a basidiocarp. Nevertheless, there are many reports of co-occurrence of sirobasidiaceous fungus with Xylariales (e.g., Chen 1998;Gilbertson and Adaskaveg 1993), but no reports of tremelloid haustoria were there until Bandoni et al. (2011) mentioned the rare formation of them at the interface with associated ascomycetous stroma in Sb. magnum. Although tremelloid haustoria were not detected during the observation of the basidiocarp in Sb. apiculatum, their formation under the cultural conditions suggests the mycoparasitic nature of the species. Kobayasi, Trans. Mycol. Soc. Japan 4: 29 (1962).
Materials examined: JAPAN, Sugadaira Montane Research Center (now as Sugadaira Research Station), University of Tsukuba, Sugadaira-Kogen, Ueda, Nagano Pref., 36°31′20.7″N 138°21′2.2″E, alt. Notes: The formation of basidia in chains was not detected in the specimen examined, which is consistent with some previous reports (Gordon 1938;Reid and Minter 1979). Bandoni (1986) indicated the different frequencies of basidia in chains among collections. These inconsistent observations suggest that the formation of basidia in chains is an unstable characteristic. In the genus Sirotrema, St. parvula Bandoni and the type species St. pusilla Bandoni have been described in addition to this species (Bandoni 1986). It is necessary to establish these cultures in the future to determine if these two species are also placed in the genus Phaeotremella, as is the case with the here examined species.
In the genus Phaeotremella, many species are known to be parasitic on Stereaceae (Russulales, Agaricomycotina, Basidiomycota), and P. mycetophiloides (Kobayasi) (Kobayasi 1939;Martin 1940;Bandoni and Ginns 1993) and have hyphal swellings near clamp connections similar to P. translucens (Bandoni and Ginns 1993). This analysis revealed that the genus includes S. translucens, with an ascomycete host (Lophodermium) for the first time. While phylogenetic relationships of the mycoparasitic species in the genus are still not clear, this character appears to be common to the mycoparasitic taxa in the genus.