Cystobasidium alpinum sp. nov. and Rhodosporidiobolus oreadorum sp. nov. from European Cold Environments and Arctic Region

Over 80% of the Earth’s environments are permanently or periodically exposed to temperatures below 5 °C. Cold habitats harbour a wide diversity of psychrophilic and psychrotolerant yeasts. During ecological studies of yeast communities carried out in cold ecosystem in the Italian Alps, Svalbard (Norway, Arctic region), and Portugal, 23 yeast strains that could not be assigned to any known fungal taxa were isolated. In particular, two of them were first identified as Rhodotorula sp., showing the highest degree of D1/D2 sequence identity with Cystobasidum laryngis accounted to only 97% with the type strain (C. laryngis CBS 2221). The other 21 strains, exhibiting identical D1/D2 sequences, had low identity (97%) with Rhodosporidiobolus lusitaniae and Rhodosporidiobolus colostri. Similarly, ITS sequences of the type strains of the most closely related species (93–94%). In a 2-genes multilocus D1/D2 and ITS ML phylogenetic tree, the studied strains pooled in two well separated and supported groups. In order to classify the new 23 isolates based on phylogenetic evidences, we propose the description of two novel species Cystobasidium alpinum sp. nov. and Rhodosporidiobolus oreadorum sp. nov.


Introduction
The cryosphere covers about one-fifth of the surface of the Earth, with substantial seasonal variations and a long-term trend of losses in its area and volume due to climate warming [1]. Currently, the terrestrial cryosphere includes nearly 200,000 glaciers and two ice sheets [2]. Glacier ice, due to subfreezing temperatures, low water activity (aw), low nutrients availability, and high hydrostatic pressure and oxidative stress, represents one of the harshest environments of the cryosphere for living organisms. Despite these inhospitable conditions, glacial habitats harbour a wide diversity of adapted microorganisms [3]. In spite of the concerted efforts over the last 10 years, there are still major gaps in our knowledge about the fungal biodiversity inhabiting the worldwide cold regions; as a consequence, a detailed knowledge of the microbial biodiversity and its functions (hydrolysis of duplicates at three separate locations: N 78 • 53, 321/E 12 • 21,699 (samples No. 20 and 21), N 78 • 53, 973/E 12 • 22,093 (samples No.23 and 24), and N 78 • 57, 809/E 12 • 26,947 (samples No. 26 and 27). Ice samples were transported to the laboratory of Natural Environment Research Council Arctic research base, where they were melted aseptically at room temperature, and processed as described by Gunde-Cimerman et al. [7], on the following culture media: synthetic nutrient agar (SNA), malt extract agar (MEA), MEA supplemented with 5% NaCl (MEA5%), malt yeast extract with 20% glucose (MY20%); all culture media were supplemented with 50 mg/L of chloramphenicol. Incubation temperatures were 5 • C and room temperature.
Strains ZP 292 and ZP 295 were isolated form soil covered by snow in the mountain of Serra da Estrela in Portugal [6].
Strains DBVPG 10041 and CCFEE 6272 are preserved in the Industrial Yeasts Collection DBVPG of the University of Perugia, Italy (http://www.dbvpg.unipg.it), and in the Culture Collection of Fungi from Extreme Environments (Mycological Section of the Italian Antarctic Museum, MNA), University of Tuscia, Viterbo, Italy, respectively, as lyophilized (4 • C) and long-term cryopreserved (−80 • C) cultures. Strains with EXF acronym are preserved at −80 • C in the Culture Collection of Extremophilic Fungi Ex (http://www.ex-genebank.com/index.php/en/fungi-2), at the University of Ljubljana, Slovenia. Strains PYCC 8103 and PYCC 8104 are conserved at the Portuguese Yeast Culture Collection (PYCC), Universidade Nova de Lisboa, Portugal.

Physiology
Physiological and biochemical tests were performed according to the protocols described by Kurtzman et al. [17] using the following media: Malt Extract broth (ME, Oxoid), Malt Extract Agar (MEA, Oxoid), Potato Dextrose Agar (PDA, Oxoid), and Corn Meal Agar (CMA, Difco). Before performing the tests the optimum temperatures (defined as the temperature at which the maximum growth was detected) of both species were determined using MEA solid medium [17], inoculating spots of 5 µL of decimal dilutions of each strains. The Petri dishes were incubated at different temperatures (10 • C, 12 • C, 15 • C, 18 • C, 20 • C, 25 • C). The test was performed in triplicate. The temperature at which the most abundant growth at the maximum dilution was observed, was defined as the optimum temperature. Physiological and biochemical tests were done in duplicate at 15 • C for both C. alpinum and R. oreadorum. Results were recorded after 2 and 4 weeks of incubation. Results obtained from duplicate experiments were concordant.

Morphology
For microscopy, cultures were grown on PDA agar at room temperature and studied with a Nikon Microphot-FX microscope and Olympus BX51 microscope, the latter equipped with differential interference contrast optics. Images of colony morphology were taken using a stereo-microscope equipped a Nikon zoom digital camera, COOLPIX950. For determination of sexual compatibility, pairs of 7-day-old cultures were crossed on corn meal agar, incubated at 16-18 • C for one week and examined for production of mycelium and teliospores after 1-6 weeks. To induce teliospore germination, 6-week-old agar blocks containing them were soaked in demineralized water for 8 weeks at 4 • C. Then the teliospores were transferred to 2% water agar at 20-22 • C.

Molecular Analysis and Phylogeny
The D1/D2 domains of the large subunit (28S) ribosomal gene region and the Internal Transcribed Spacer region (ITS1 and 2) including the 5.8S rRNA gene were amplified using standard primers (NL1 and RLR3R for LSU; ITS1 and ITS2 for ITS regions) [18]. Amplifications were performed using the T personal Combi Thermal Cycler (Biometras Gmbh, Goettingen, Germany), applying amplification protocols reported by Selbmann et al. [5]. Sequences were obtained with an ABI PRISM 3730XL Capillary Sequencer using the standard protocol recommended by the manufacturer. ITS and D1/D2 sequences were blasted in the NCBI database. The most similar sequences were exported, alignments were generated using MAFFT version 6 [19] and improved manually in MEGA6 [20]. ITS and D1/D2 alignments were concatenated using FABOX (http://users-irc.au.dk/biopv/php/fabox/alignment_ joiner.php, [21]). Combined sequences of two-locus datasets alignments were exported and the best-fit substitution model was determined using MODELTEST MrAIC 1.4.3 [22] as implemented in PHYML [23].
Phylogenetic ML trees reconstruction was performed selecting the best substitution model gained with MODELTEST MrAIC 1.4.3 in MEGA6. The robustness of the phylogenetic inference was estimated using the bootstrap method [24] with 1000 pseudoreplicates generated and analyzed with TREEFINDER.

Results
The complete list of the strains, their isolation sources, and their sequences accession numbers are shown in Table 1. Phylogeny was based on the two-gene dataset that included 42 strains which sequences could be confidently aligned with the ones of studied strains; the dataset was based on 1348 positions (ITS: 1-707; D1/D2: 708-1, 348) including gaps.
The resulting multilocus phylogenetic tree, generated using a GTR+G(4) model selected using the Akaike's information criterion with a Maximum-likelihood approach, is shown in Figure 1. The backbone of the tree and branches of genera are well resolved and supported. The tree topology is concordant with the most recent phylogenetic studies on yeasts in Pucciniomycotina [14] and on the genus Cystobasidium in particular [25].
The two strains DBVPG 10041 and CCFEE 6272 pooled in a separated and well supported branch (bootstrap 99%) in the genus Cystobasidium, confirming the high divergence observed with BlastN analysis. The D1/D2 sequences were, in fact, only 97% similar to sequences available in the public domain and the closest species for strain DBVPG 10041 was C. laryngis, 16 bp different. Surprisingly, a lower divergence with known species was found when the target ITS (1 and 2) sequence was blasted in the public domain, despite this gene being, as barcode for fungi and hence for the genus Cystobasidium too, extremely variable [16]. In particular, DBVPG 10041 and CCFEE 6272 have the same ITS sequence that differ by 12 nucleotide substitutions from the type strain of Cystobasidium pinicola and 16 with C. laryngis. Only 10 nucleotide substitutions differentiate DBVPG 10041 and CCFEE 6272 from two sequences deposited in Genbank as Rhodotorula sp. BC22 [26] and Antarctic yeast CBS 8923 (unpublished), both isolated from Antarctica and not yet described.
Based on these considerations and on the phylogenetic position in the multilocus ML tree, these two strains are here described as new species and Cystobasidium alpinum is proposed to accomodate the strains DBVPG 10041 and CCFEE 6272.
The main macroscopic characteristic that distinguishes C. alpinum from the closest species is the colour of its colonies: cream to brownish-gray while all the species of C. laryngis clade are coloured from light pink to salmon, to red (Table 2). Additionally, optimum growth temperature for C. alpinum is 15 • C and maximum 25 • C, while all the closest species had optimum temperature from 15-17 • C to 25 • C, all of them duplicated at 25 • C and, in few cases, even at 30-35 • C ( Table 2). Other physiological differences are shown in Table 2.    The C. laryngis clade is mainly composed of species which originate in cold environments. In this clade three new Antarctic species were recently described: Cystobasidium tubakii, Cystobasidium ongulense, and Cystobasidium portillonense [25,28]. Additionally, some strains isolated from a glacier in the Alps, first identified as Rhodotorula cf. laryngis, were later identified as Cystobasidium ongulense [4]. Moreover strains of C. laryngis, C. pinicola, and Cystobasidium benticum were also all isolated from cold habitats: Antarctica, Greenland, cold lake in Argentina, and deep sea. All listed species, including C. alpinum, share psychrophilic or psychrotolerant aptitude, showing optimum temperature between 15 • C and 20 • C. The most defining characteristics of C. laryngis clade are therefore their origin in cold environment and adaptations to the habitat temperatures.
The remaining strains listed in Table 1 share similar D1/D2 and ITS sequences. PYCC 8104 had one mismatch in D1/D2 sequences when compared to the other strains. Nucleotide differences in ITS sequences ranged from one to three. The D1/D2 sequence of the type strain EXF-3880 showed three nucleotide substitutions in comparison to the closest species R. lusitaniae, and eight substitutions in comparison to R. colostri; a larger divergence with 27 and 28 mismatches was found in ITS sequences. The ML multilocus phylogeny placed all these strains in a well-supported (100%) and separate branch, with R. lusitaniae and R. colostri as closest neighbours. Based on the evidences above, we here allocate these strains in a new species and propose the name Rhodosporidiobolus oreadorum R. oreadorum isolates differed from the closest species also in their abilities to assimilate cellobiose and nitrite, in the production of starch like compounds and in the formation of capsule (Table 3). Additionally, after crossing sexually compatible strains of R. oreadorum, true hyphae with complete clamp connections were observed together with spherical teliospores formed terminally and intercalarly in the mycelium. The sexual structures of R. oreadorum show some differences when compared with those of R. lusitaniae, its closest sexual relative. In the mycelium of R. lusitaniae, clamp connections are not formed, whereas those structures are formed by R. oreadorum. Interestingly, R. lusitaniae is homothallic while R. oreadorum is heterothallic. Another distinctive feature is the presence of an elongated stalk connecting the basidium to the teliospore in R. lusitaniae, whereas such stalk is absent in R. oreadorum [29]. Maximum growth temperature for R. oreadorum was 30 • C, characterized by weak growth, while all the other species displayed richer growth at the same or higher maximum temperatures (Table 3). Other physiological differences are shown in Table 3. All the Arctic strains of R. oreadorum were isolated exclusively from subglacial ice, and not from adjacent samples like ice at the margin of the glacier, stream water on the glacier, glacial melt water, or adjacent sea water. The two strains isolated in Portugal were found in soil under snow in a mountainous region.

Discussion
It is commonly hypothesized that microbial cells that originate from geographically close and distant places are deposited with snow and gradually become embedded in the ice layers of glaciers or surface of soil covered with snow [4]. After this, the cells concentrate in the veins of liquid water between the forming ice crystals, where the micro-conditions can support the survival of the best adapted species despite the hostile surrounding environment [32].
Microorganisms that grow in such habitats are commonly referred to as psychrophiles, if they are unable to grow at temperature >20 • C, or as psychrotolerant, if they tolerate low temperatures, but have a growth temperature optimum above 15 • C [33].
In recent years, a few studies [8,34] discovered the presence of fungi, including yeasts, beneath and in meltwater draining off from polythermal glaciers, where ice melting occurs due to changes in pressure. Since global warming is predicted to be most pronounced at high latitudes, Arctic and Alpine glaciers are and will be habitats intensively threatened by climate change.
All the strains of the present study were isolated from cold environments: Svalbard, Italian Alps and soil under snow layer in Portuguese mountains (Table 1). Physiology denounces a cold-adaptation of the strains examined; in fact, they showed a psychrotolerant aptitude, having an optimum growth temperature at 15-18 • C, and weak or no growth at 25 or 30 • C (Tables 2 and 3). In the past, cryospheric environments were considered abiotic; however, they are now recognised as ecosystems containing an enormous reservoir of microbial diversity, as exemplified by the description of the two new species. Their recognition will improve our understanding of the dynamics of the extremely cold environments and also provide additional crucial information for predicting the consequences of the massive changes of the Arctic glacial ecosystems caused by global climate changes. Additionally, the global warming will change the climatic condition of cold ecosystems and will determine the melting of glaciers. These habitats are now considered in danger of disappearing [35] as well as the psychrophilic microbial biodiversity associate to them. In this context, the isolation, description, and ex situ conservation of psychrophilic microorganisms can be considered a way for avoiding their possible extinction. Cystobasidium alpinum (al.pi.'num N.L. neut. adj., 'alpinum' referring to Alps, the mountain complex from which the strains were isolated).
Novel yeast species belonging to phylum Basidiomycota, subphylum Pucciniomycotina, class Cystobasidiomycetes, order Cystobasidiales, family Cystobasidiaceae. Telomorph: not observed. MycoBank: MB#824354 (Figure 2). After 14 days on malt extract agar (MEA), corn meal agar (CMA), and potato dextrose agar (PDA) at 15 • C, the colonies were cream to brownish-gray, smooth and glistening, with entire margin, flat profile, and viscous texture. On YPDA (yeast extract 1%, peptone 1%, dextrose 2%, agar 2%) the colonies were bigger and dull and showed butyrous texture and raised profile. The cells were ovoidal, 2-5 µm long, and 1-2 µm wide, and the budding was polar. In none of the strains, pseudohyphae were observed. Sediment was produced when the strains grew in Malt extract broth (ME) and YPD at 15 • C, after 7 days. After 14 days, a light superficial ring was also present. No positive mating reactions were observed among the two strains using different media.
Sucrose and l-malic acid gave different results in the two tested strains, and thus can be considered variably assimilated. Assimilation of nitrogen compounds was positive for L-lysine and ethylamine (HCl), weak for cadaverine and nitrite, and negative for nitrate.
Growth at 10 • C, 15 • C, 20 • C was good, while it was delayed at 4 • C and very weak at 25 • C; no growth was shown at 30 • C. Optimum growth temperature is 15 • C.
Growth YE agar with 50% glucose was absent while was weak and delay on YNB broth with 10% NaCl +5% glucose. Starch-like compounds were not produced. In 0.01% cycloheximide, growth was weak and absent at 0.1% concentration. Urease activity was absent, while Diazonium Blue B reaction was positive.
The type strain of Cystobasidium alpinum sp. nov. was isolated from supraglacial sediments of Miage glacier, Mont Blanc, Italian Alps, Italy (45 • 84 70 N; 68 • 52 00 E), and has been deposited in the Industrial Yeasts Collection DBVPG, Department of Agricultural, Food and Environmental Sciences, University of Perugia, Italy, under the codes DBVPG 10041T, and in CBS KNAW Collection as CBS 14809T.
After 7 days on malt extract/yeast extract agar at 18 • C, the colonies are reddish, smooth and shiny, with an entire margin. On PDA, the red colony color was more intense. The cells are ovoidal to ellipsoidal, 3-9.5 µm long, in average 5.8 µm (SD 1.3) and 2.5-5 µm wide, in average 3.3 µm (SD 0.7), with a capsule, and they reproduce by bilateral budding. In Dalmau plates after 2 weeks on cornmeal agar, pseudohyphae and true hyphae are not formed in single strains. Heterothalic, after crossing sexually compatible strains on CMA or SG agar true hyphae, 1 µm in diameter, with complete clamp connections and terminal as well as intercalary teliospores, 14-18 µm in diameter, developed after one week. Teliospores germinated after a resting stage of 6 weeks at 4 • C and formed transversely septate basidia with four compartments. Basidiospores were ovoidal and measured 5-6 × 2-3 µm. The germination of teliospores was observed in a cross involving strains from Portugal.