The Enigmatic Snow Microorganism, Chionaster nivalis, Is Closely Related to Bartheletia paradoxa (Agaricomycotina, Basidiomycota)

Chionaster nivalis is frequently detected in thawing snowpacks and glaciers. However, the taxonomic position of this species above the genus level remains unclear. We herein conducted molecular analyses of C. nivalis using the ribosomal RNA operon sequences obtained from more than 200 cells of this species isolated from a field-collected material. Our molecular phylogenetic analyses revealed that C. nivalis is a sister to Bartheletia paradoxa, which is an orphan basal lineage of Agaricomycotina. We also showed that C. nivalis sequences were contained in several previously examined meta-amplicon sequence datasets from snowpacks and glaciers in the Northern Hemisphere and Antarctica.

Food webs in thawing snow and glaciers are comprised of cold-adapted microorganisms, such as algae, heterotrophic protists (e.g., ciliates and heterokonts), rotifers, fungi, and bacteria (Hoham and Duval, 2001;Hoham and Remias, 2020). The ecosystems in these harsh environments have long attracted the interest of biologists, resulting in ongoing efforts to elucidate their biodiversity. Recent highthroughput sequencing techniques revealed comprehensive communities within these ecosystems, particularly in microalgae and bacteria (e.g., Lutz et al., 2015Lutz et al., , 2016Segawa et al., 2018;Krug et al., 2020). The outcomes of highthroughput sequencing for the identification and classification of detected microorganisms depend on the quality of the reference databases (Lutz et al., 2019). Therefore, it is important to increase the availability and taxonomic accuracy of molecular data on microorganisms inhabiting snow and glaciers.
Chionaster nivalis is one of the typical unicellular microorganisms found within thawing snowpacks and glaciers, occurring in blooms of snow-inhabiting microalgae (i.e., colored snow). This species is characterized by a starshaped cell wall that is formed by three to five protuberances (Bohlin, 1893;Wille, 1903). Possibly due to its unique morphological traits, this species has frequently been detected in snowpacks and glaciers in Europe, North Amer-ica, Japan, Australia, and Russia as well as in the Arctic region (Kol, 1968;Marchant, 1982;Müller et al., 1998;Uetake et al., 2006). Therefore, C. nivalis is regarded as a widely distributed species. Despite extensive studies, the taxonomic position of C. nivalis above the genus level remains unclear. This species was originally described by Bohlin (1893) based on the red snow material collected in Sweden. Although the majority of cells within the material appeared to lose their color due to the preservation solution, Bohlin classified the organism as the third species in the unicellular green algal genus, Cerasterias because the cell morphology of this organism was consistent with the concept of the genus being characterized by the presence of several protuberances on cells (Reinsch, 1866; but see also Komárek and Fott, 1983). Wille (1903) subsequently transferred this species to the new fungal genus, Chionaster, due to the observation that the organism was completely colorless. However, Wille only reported that this species needed to be listed as a new family and did not state its taxonomic position above the genus level. MycoBank (www.mycobank.org) currently treats C. nivalis as Fungi incertae sedis, whereas Index Fungorum (http:// www.indexfungorum.org/) assigns this species to Chromista incertae sedis. There are no molecular data available for this species, possibly due to the lack of cultures of C. nivalis and also because this species does not form monospecific blooms in nature (Fukushima, 1963;Kol, 1968).
In our recent field work on melting snowpacks in Japan, we collected several samples of colored snow containing cells that were morphologically identifiable as C. nivalis (see Method S1). These cells were not dominant in the samples, as previously described (Fukushima, 1963;Kol, 1968). Viewed under a light microscope, cells were solitary and generally had four protuberances that formed a cross ( 1a). There were sometimes three ( Fig. 1b) or five protuberances, and each protuberance was 20-30 μm in length with an obtuse apex. The cellular protoplasm was colorless and amorphous in shape, and was generally positioned near the bases of the protuberances. These characteristics were consistent with descriptions of C. nivalis from Europe (Bohlin, 1893;Wille, 1903) and Japan (Fukushima, 1963). An attempt was made to establish a culture of Japanese C. nivalis by isolating field-collected cells into several conventional culture media for fungi and freshwater and snowinhabiting microalgae in a cold experimental room at 5°C. However, neither cell division nor spore germination by C. nivalis was observed during the study and attempts completely failed. These findings are consistent with previous descriptions of these cells apparently being hypnospores (Bohlin, 1893;Wille, 1903;Fukushima, 1963), suggesting that additional factor(s) for the induction of germination are required to establish cultures of this species. Alternatively, the culture media used in this study were not suitable for C. nivalis.
Since C. nivalis was uncultured and did not dominate within samples, we attempted to extract total DNA from an isolated single cell or a few (between 6 and 11) cells of field-collected C. nivalis in order to sequence the internal transcribed spacer 2 (ITS-2) region by nested PCR (see Method S2). Only three out of 30 PCR products showed amplification (Fig. S1); however, we obtained faint PCR products and their sequences were of low quality (data not shown). This may have been due to the low quantity and quality of DNA extracted from a single to a few cells. Therefore, we isolated more than 200 cells of the species from field-collected materials to obtain a sufficient quantity of total DNA for PCR (see Method S3 for details). To eliminate contaminants, the surface of C. nivalis cells within a snow sample was treated with a sterilizing solution (1% -Toyooka et al., 2004]). Using a capillary pipette, approximately 500 and 200 cells of C. nivalis were isolated from colored snow samples collected from Oze National Park (Fig. 1c) and Tambara-kogen Plateau (Fig. 1d), Gunma, Japan, respectively. Isolated cells were grouped by locality and assigned specimen labels of Oze (Fig. S2) and Tambara (Fig. S3), respectively. Each specimen was treated on ice with cold-active nuclease (Cryonase; Takara Bio) to eliminate nontarget nucleotides outside of the cells. According to a previously described method by Nakada et al. (2007), C. nivalis cells in each specimen were broken using ceramic beads to allow total DNA extraction.
Using extracted DNA from the two specimens, we amplified part of the rRNA operon (including 18S rRNA, internal transcribed spacer 1 [ITS-1], 5.8S rRNA, ITS-2, and 28S rRNA) by PCR, and then directly Sanger-sequenced the PCR products (see Method S3). In the PCR of each DNA sample, we used universal primers for eukaryotic rRNA operons (Table S1 and Fig. S4). The 18S rDNA sequences of the Oze and Tambara specimens harbored four and three group I introns, respectively ( Fig. S4 and S5). However, the sequences of the 18S and 28S rRNA-coding regions and rapidly evolving ITS-1 and ITS-2 regions were identical between the two specimens, supporting the hypothesis that the obtained sequence of the rRNA operon originated from a single species, C. nivalis.
To construct a phylogenetic tree, Bayesian inference (BI) using MrBayes 3.2.7 (Ronquist et al., 2012) and the maximum likelihood (ML) analysis using RAxML-NG 0.9 (Kozlov et al., 2019) were performed, as previously described by Nozaki et al. (2010) and Matsuzaki et al. (2021), respectively. Phylogenetic analyses based on 18S rRNA (1,610 bp) and 28S rRNA (549 bp) were conducted for comparisons ( Fig. S6 and S7). Since robust discrepancies in phylogenetic relationships within the phylum Basidiomycota were not detected between the trees, the concatenated 2,159bp data matrix of 18S and 28S rRNA was subjected to the same methods as those described above (see Method S4 and Table S2 for details).
Several meta-amplicon sequencing datasets of 18S rRNA and ITS-2 from previous studies on snowpacks and glaciers are available in DDBJ/ENA/GenBank (Table S3); therefore, we attempted to detect the sequences assignable to C. nivalis in these datasets using Vsearch 2.15.1 (Rognes et al., 2016). Thresholds for identical matches in 18S rRNA and ITS-2 were set as 0.99 and 0.97, respectively (see also Method S5).
Our molecular phylogenetic analyses demonstrated that C. nivalis (Oze and Tambara specimens) belongs to the phylum Basidiomycota (Fig. 2, S6, and S7). Within Basidiomycota, the monophyly of the subphyla Pucciniomycotina, Ustilaginomycotina, and Wallemiomycotina was well recovered (1.00 posterior probabilities [PP] in BI and 100% bootstrap value [BV] [Felsenstein, 1985] in the ML analysis), whereas that of Agaricomycotina was low (only 52% BV in the ML analysis). C. nivalis was positioned within Agaricomycotina and was a sister to Bartheletia paradoxa (Bartheletiomycetes) with robust statistical support (1.00 PP in BI and 100% BV in the ML analysis). B. paradoxa is a  filamentous fungus associated with fallen Ginkgo biloba leaves, and is known as an orphan basal lineage of Agaricomycotina (Mishra et al., 2018). Within Agaricomycotina, the monophyly of Agaricomycetes, Dacrymycetes, and Tremellomycetes (i.e., Agaricomycotina, excluding B. paradoxa and C. nivalis) was moderately supported (1.00 PP in BI and 86% BV in the ML analysis). We detected sequences assignable to C. nivalis (or at least its relatives) in meta-amplicon sequencing data from snowpacks and glaciers in Arctic regions, Asia (Japan), Europe, North America, and Antarctica (Table 1). To the best of our knowledge, this is the first study on C. nivalis from Antarctica; however, the evidence obtained is based solely on molecular data. The present results strongly support the hypothesis that C. nivalis is a widely distributed or cosmopolitan species. The ratios of the C. nivalis sequences detected in respective sequencing data for each sampling locality were very low (generally less than 1%), which may have been due to the difficulties associated with disrupting the thick cell wall during DNA extraction (see also Fig. S1). Sequence similarities in ITS-2 (known as a highly evolving region) between Japanese C. nivalis and sequences from the Northern Hemisphere were as high as 99.4%-100%, whereas those from Antarctica had similarities of at most 97.1%. Although it is important to note that available data were highly limited and further studies are warranted, these sequence similarities indicate that C. nivalis diversified between Antarctica and the other regions examined (all of which are in the Northern Hemisphere). Potentially similar cases were reported in the snow-inhabiting microalga, Sanguina nivaloides (formerly known as the cyst of Chlamydomonas nivalis) (Segawa et al., 2018;Brown and Tucker, 2020).
In the present study, we collected hundreds of C. nivalis cells from a single source of field material, and performed PCR and Sanger sequencing using DNA extracted from these cells. This method may be applicable to molecular analyses of microorganisms that are not culturable and do not form monospecific blooms in nature. Our molecular data clearly demonstrated that C. nivalis is closely related to B. paradoxa, known as the basal lineage of Agaricomycotina. The former species (C. nivalis) has only been reported from snow and glacier environments, whereas the latter (B. paradoxa) grows only on fallen leaves of G. biloba. Since the morphological, ecological, and life cycle data of C. nivalis are highly limited, further discussions, including the taxonomic position of the species at the class level (Bartheletiomycetes or a new class altogether), are difficult. However, the present results provide insights into this enigmatic snow fungus and the basal lineage of Agaricomycotina. In addition, we demonstrated that the sequences assignable to C. nivalis were actually present in several meta-amplicon sequencing datasets from previously examined snowpacks and glaciers. This strongly supports previous findings by Lutz et al. (2019) indicating the importance of enlarging reference databases with accurate taxonomic information to reveal whole microbial communities in field-collected materials using high-throughput sequencing technologies.

Data availability
All sequence data obtained in the present study have been deposited in DDBJ/ENA/GenBank under the accession numbers LC599386-7, MW404634-6, and MW404638-40. The sequence matrix used in the present study is available from TreeBASE (www.treebase.org/treebase-web/ home.html, study ID: S27454).