DNA metabarcoding of fungal diversity in air and snow of Livingston Island, South Shetland Islands, Antarctica

We assessed fungal diversity present in air and freshly deposited snow samples obtained from Livingston Island, Antarctica, using DNA metabarcoding through high throughput sequencing (HTS). A total of 740 m3 of air were pumped through a 0.22 µm membrane. Snow obtained shortly after deposition was kept at room temperature and yielded 3.760 L of water, which was filtered using Sterivex membranes of 0.22 µm mesh size. The total DNA present was extracted and sequenced. We detected 171 fungal amplicon sequence variants (ASVs), 70 from the air and 142 from the snow. They were dominated by the phyla Ascomycota, Basidiomycota, Mortierellomycota and Mucoromycota. Pseudogymnoascus, Cladosporium, Mortierella and Penicillium sp. were the most dominant ASVs detected in the air in rank order. In snow, Cladosporium, Pseudogymnoascus, Penicillium, Meyerozyma, Lecidea, Malassezia, Hanseniaspora, Austroplaca, Mortierella, Rhodotorula, Penicillium, Thelebolus, Aspergillus, Poaceicola, Glarea and Lecanora were the dominant ASVs present. In general, the two fungal assemblages displayed high diversity, richness, and dominance indices, with the assemblage found in snow having the highest diversity indices. Of the total fungal ASVs detected, 29 were only present in the air sample and 101 in the snow sample, with only 41 present in both samples; however, when only the dominant taxa from both samples were compared none occurred only in the air and, among the rare portion, 26 taxa occurred in both air and snow. Application of HTS revealed the presence of a more diverse fungal community in the air and snow of Livingston Island in comparison with studies using traditional isolation methods. The assemblages were dominated by cold-adapted and cosmopolitan fungal taxa, including members of the genera Pseudogymnoascus, Malassezia and Rhodotorula, which include some taxa reported as opportunistic. Our results support the hypothesis that the presence of microbiota in the airspora indicates the possibility of dispersal around Antarctica in the air column. However, further aeromycology studies are required to understand the dynamics of fungal dispersal within and beyond Antarctica.

Valleys, display limited connectivity to the global microbial pool due the strong selection that occurs during atmospheric transport, resulting in regionally isolated airborne inputs and highly specialized soil communities, with fungi also displaying greater isolation from non-polar sources than bacteria. However, detailed information about the aerial routes by which microorganisms arrive and circulate in Antarctica is lacking 8,9 . Biological dispersal by aerial means can be an important factor shaping patterns of biodiversity 9,10 . Viable organisms or their propagules present in the air column may be in dormant and cryptobiotic states, where they are metabolically inactive due the harsh dry, cold, low nutrient and high irradiance conditions. Diverse groups of microorganisms have been recorded in the few Antarctic aerobiological studies completed to date (reviewed by Pearce et al. 9 ), including viruses, bacteria, microalgae and fungi.
Mycological studies in Antarctica have shown that much of the Antarctic fungal community is represented by cold tolerant (psychrophilic or psychrotolerant) species, many of which have wide and even globally cosmopolitan distributions, with presence in polar, temperate, and tropical environments 11 . de Menezes et al. 12 suggested that the high densities of cosmopolitan fungi present in snow are consistent with them being present in air masses arriving at the Antarctic Peninsula from beyond Antarctica, which are then entrained in snow precipitation, and become concentrated in the snow. Snow and ice can provide an indirect record of the presence and deposition of fungal propagules (e.g. spores or hyphal fragments) from the air column over time 12 . In snow samples obtained from six different regions of the Antarctic Peninsula, de Menezes et al. 13 reported a rich fungal diversity assigned to 51 species in 26 genera and dominated by cold tolerant cosmopolitan fungi. However, in ice from continental Antarctica and the Antarctic Peninsula, Rogers et al. 14 and de Menezes et al. 15 , respectively, reported much lower fungal diversity. In the present study, we assessed fungal diversity present in air and freshly deposited snow samples obtained from Livingston Island, Antarctica, using DNA metabarcoding through high-throughput sequencing (HTS).

Material and methods
Snow and air sampling. Air and snow samples were collected at Punta Polaca (62°40′16″ S; 60°22′43″ W), Hurd Peninsula, Livingston Island, South Shetland Islands, near to the Spanish station Juan Carlos I (Fig. 1). Two air samples were collected with a high flow glass impinger following Šantl-Temkiv et al. 16,17 . The chamber was filled with 2 L of sampling liquid (ddH 2 O) and the sampler was run for 5 min, so that the liquid came in contact with the entire chamber, after which 0.5 L of the sampling liquid was removed, stored as a control, and analyzed along with the samples. The control represented a field blank to certify that the samples were not contaminated by external organisms. The resulting solution was filtered directly on the Sterivex filter units for the air, as described by Lever et al. 18 . Air was collected over c. 5 h on March 11th 2019. In addition, the two separate air DNA extractions were combined together in order to increase DNA yield. Two freshly deposited snow samples were collected on March 20 th 2019 at the same site using a sterilized shovel. Both pairs of samples were separately   22 , to trim ends 3′ or 5′ with low Phred quality score. Sequences shorter than 50 bases were discarded. These sequences were imported to QIIME2 version 2019.10 (https ://qiime 2.org/) for bioinformatics analyses 23 . The qiime2-dada2 plugin is a complete pipeline that was used for filtering, dereplication, turning paired-end fastq files into merged reads, and removal of chimeras 24 . Taxonomic assignment was carried out for the amplicon sequence variants (ASVs) using qiime2-feature-classifier 25 classify-sklearn against the UNITE fungal ITS database version 7.2 26 and trained with Naive Bayes classifier. A confidence threshold of 98.5% was used. All raw sequences have been deposited in the NCBI database under the codes SRR12830238, SRR12830240 and SRR12830239.
Many factors, including extraction, PCR, and primer bias, can affect the number of reads 27 , and thus lead to misinterpretation of abundance 28 . However, Giner et al. 29 concluded that such biases did not affect the proportionality between reads and cell abundance, implying that more reads are linked with higher abundance 29,30 . Therefore, for comparative purposes we used the number of reads as a proxy for relative abundance.
All sequences obtained from air and snow samples were matched with sequences present in the list of the top 50 'most wanted' fungi according to Nilsson et al. 31 . The sequences were merged, filtered, dereplicated, and clustered into > 97% identity ASVs using USEARCH version 10 32 . Nucleotide-Nucleotide BLAST 2.6.0 + was used to compare these ASVs against the top50_release_04.02.2020.fasta 33 , considering just subject matches with aligned length longer than 250 bp and > 98% identity.
Fungal diversity and distribution. To quantify species diversity, richness, and dominance, we used the following indices: (i) Fisher's α, (ii) Margalef 's, and (iii) Simpson's, respectively. The numbers of DNA reads of the amplicon sequence variants (ASVs) were used to quantify the fungal taxa present in the air sampled, where fungal ASVs with more than 1,000 reads were considered dominant and < 1,000 minor components (rare) of the fungal community. All of the results were obtained with 95% confidence, and bootstrap values were calculated from 1,000 iterations. Taxon species accumulation curves were obtained using the Mao Tao index. All diversity indices and species accumulation curves calculations were performed using PAST v. 1.90 34 . Venn diagrams were prepared according to Bardou et al. 35 to compare the fungal assemblages present in both air and snow samples. The functional assignments of fungal ASVs at species and genera levels are shown using FunGuild 36 .

Results
Fungal taxonomy. The number of reads in the air sample was 162,038 and that in snow 268,710. From these, we detected 171 fungal amplicon sequence variants (ASVs), 70 in 740 m 3 of air and 142 in 3.76 L of snow from Livingston Island, Antarctica ( were detected less frequently (< 1,000 reads) and may represent the rare portion of the fungal assemblages. In addition, 78 ASVs could only be assigned to higher taxonomic levels (phylum, class, order or family). A total of 29,069 sequences from the air and 6,223 from the snow samples were matched with the sequences of 11 unidentified species hypotheses in the list of the top 50 most wanted fungi 31 with the alignment length longer than 250 bp and > 98% identity (Suppl. Table 1).
Fungal diversity. The Mao Tao rarefaction curves of the fungal assemblages present in air and snow reached asymptote for both fungal assemblages (Fig. 3), indicating that the data provided a good description of the diversity present. In general, both fungal assemblages displayed high diversity, richness, and dominance indices ( Table 2). The assemblage present in the snow was more diverse, rich, and included a wider range of dominant fungi when compared with that from the air sample. Of the total fungal ASVs detected, 29 were only present in the air sample and 101 in the snow sample, with 41 present in both samples (Fig. 4a). However, when only the dominant ASVs (> 1,000 reads) from both samples were compared, none occurred only in the air (Fig. 4b) and, among the rare portion, 26 occurred in both air and snow (Fig. 4c). In addition, the ecological functional assignments of fungal ASVs in species and genera levels were showed in Suppl. Table 2 and Suppl.   www.nature.com/scientificreports/

Discussion
Fungal taxonomy and diversity. Despite an increase in mycological studies, fungal diversity in Antarctica remains poorly known 11 . According to Bridge and Spooner 37 , around 1,000 fungal species have been described from Antarctica, identified using a range of approaches including traditional methods for cultivable fungi such as macro-and/or micromorphology of colonies and fruiting bodies as well as DNA sequencing of mycelia of cultivable fungi. Most airborne mycological studies in Antarctica have relied on traditional morphological methods. Marshall 4 monitored airborne fungal spores over 13.6 months at three sites on Signy Island (South Orkney Islands) in the maritime Antarctic, reporting that Epicoccum spp. and Cladosporium spp. dominated the diversity present. Duncan et al. 38 sampled air inside the historic wooden huts on Ross Island, finding Cladosporium cladosporioides, Pseudeurotium desertorum, Pseudogymnoascus sp. and Antarctomyces psychrotrophicus as dominant viable fungal propagules and Cadophora sp. and Thebolus sp. as minor components of the outdoor airborne fungal assemblage. Archer et al. 2 compared microbial diversity in near-ground and highaltitude air above the Victoria Land Dry Valleys as well as that of underlying soil microbial communities, finding basidiomycete yeasts to be dominant in the air and unclassified fungi to be common in soils. However, the more recent fungal inventories using metabarcoding approaches have demonstrated that fungal diversity in Antarctica is greater than previously recognised [39][40][41] .
As air and snow are typically ultra-oligotrophic microhabitats, few viable fungal taxa are expected to be present, as reported by de Menezes et al. 12 who, using cultivation techniques, reported only 14 fungal taxa in snow samples from several different Antarctic islands. However, despite analysing only a small a small absolute sample size of air and snow collected in the Livingston Island, use of the HTS approach in the current study revealed the presence of much greater fungal diversity in both air and snow, many of which display mechanisms that render them well-adapted to survive atmospheric transport, such as the production of resistant spores and UV protective compounds 42,43 .
The dominant taxa detected in the air included representatives of Pseudogymnoascus, Cladosporium, Mortierella and Penicillium. However, even though recently deposited snow would be expected to contain microbial airborne particles entrained from the air column as the snow fell, fungal diversity in the snow sampled was very different to that in the air over the same location. In snow sample, the dominant taxa found included representatives of Cladosporium, Pseudogymnoascus, Penicillium, Meyerozyma, Lecidea, Malassezia, Hanseniaspora, Austroplaca, Mortierella, Rhodotorula, Penicillium, Thelebolus, Aspergillus, Poaceicola, Glarea and Lecanora. The  www.nature.com/scientificreports/ diversity present in both the air and snow samples also included dominant taxa that could only be assigned to higher taxonomic levels such as Fungal sp., Ascomycota sp., Basidiomycota sp., Agaricales sp., Chaetothyriales sp., Helotiales sp., Lecanorales sp. and Polyporales sp. These may represent currently undescribed or otherwise unsequenced species, further supporting the assertion that much of the true fungal diversity present in Antarctica is currently unknown. Pseudogymnoascus were detected as dominant fungi in both air and snow samples. Pseudogymnoascus (previously known as Geomyces) is a genus often detected in cold environments including those of polar, alpine, and temperate regions 11,[44][45][46][47] . In Antarctica, it has been reported from soils 44,[48][49][50] , associated with plants [51][52][53][54] and macroalgae 55 , in freshwater lakes 56 , and associated with lichens 57 . Cladosporium and Penicillium also represent common airborne fungi reported globally, including Antarctica. Cladosporium is a dematiaceous fungal group with global distribution 58 . In Antarctic microhabitats, Cladosporium has mainly been detected in association with plants and soil 11 . Penicillium is a ubiquitous genus, again detected in multiple substrates in Antarctica including soils 50,59,60 , permafrost 61,62 and associated with macroalgae 63 . The abundant presence of Pseudogymnoascus, Cladosporium, and Penicillium both in air and snow sampled indicated that these fungi may circulate at least around the Antarctic Peninsula.
The genus Mortierella includes about 85 species, which occur mainly in soils 64 . Mortierella species are found worldwide, particularly in temperate and polar regions. Representatives of the genus are abundant in Antarctica and reported in association with plants 51,52 , macroalgae 63 , lichens 57 , soils 65 , freshwater 56 , and permafrost 62 . Some species of Mortierella are known as snow moulds and have the capability to growth and produce spores at 0°C 66 . They occur abundantly in the interstitial water in Antarctic snow where snow melting occurs in summer, for instance in association with snow algal communities.
The genus Malassezia includes 17 species of basidiomycetous pigmented black yeast species generally present in the skin and mucosa microbiome of humans and other warm-blooded animals 67 . According to Prohic et al. 68 , several Malassezia species found on human and animal skin are commensals, but they can also be associated with Pityriasis versicolor, Malassezia folliculitis, seborrheic dermatitis/dandruff, atopic dermatitis, and psoriasis. The detection of Malassezia in Antarctica is unusual. Rosa et al. 54 detected different Malassezia taxa in soil samples from undisturbed and disturbed (by human activity) sites on Deception Island (South Shetland Islands) using HTS metabarcoding techniques.
The genus Meyerozyma includes species that are typically widely distributed or cosmopolitan 69 . Species of Meyerozyma have previously been isolated from aquatic environments in Antarctica 69,70 and associated with macroalgae 63 . The genus Hanseniaspora (anamorph Kloeckera) includes ascomycete yeast species commonly associated with alcoholic fermentation, but is also recorded from soil, plants, fruit-eating insects, birds, and seafood 71 . Some Hanseniaspora species have been reported as unusual opportunistic superficial mycosis in humans [72][73][74][75] .
The genus Rhodotorula includes cosmopolitan pigmented yeast species and is often dominant in extreme environments 76 , including those of Antarctica 63,70 . Our study represents the first report of high abundance of R. muscilagionsa in Antarctic snow samples, although de Menezes et al. 13 reported the species among the dominant fungi detected in snow samples from several Antarctic islands. The genus Thelebolus is distributed globally and representatives occur in diverse habitats 77 . Species of Thelebolus have been reported in Arctic and Antarctic environments 78,79 , as being abundant in lakes, and in association with birds (skuas) 80 , in freshwater 56,81 and in ice 15 . Finally, from the air and snow sampled in Livingston Island, Antarctica, we detected 11 unidentified species hypotheses in the list of the top 50 most wanted fungi 31 , suggesting the both habitats may shelter rare species that merit further taxonomic attention.

Conclusions
We used DNA metabarcoding to catalogue the fungi present in air and snow samples from Livingston Island, South Shetland Islands. This revealed a diverse fungal community comprising taxa from the phyla Ascomycota, Basidiomycota, Mortierellomycota and Mucoromycota. The assemblages were dominated by cold-adapted and cosmopolitan (psychrophilic) taxa, including members of the genera Pseudogymnoascus, Malassezia and Rhodotorula, which include taxa reported as opportunistic fungi. Our results confirm the presence of fungi in the airspora, supporting the possibility of dispersal over different geographical scales around Antarctica in the air column. Given that many of the taxa identified in this study are known from Antarctic fungal communities, a local source for those present in the air column is plausible. The large proportion of unassigned taxa highlight the poor level of baseline knowledge of Antarctic fungal diversity, and further aeromicrobiology and diversity studies are required to understand the dynamics of fungal dispersal within and beyond Antarctica. However, as metabarcoding detects environmental DNA, the technique can also detect DNA from dead fungi or otherwise non-viable material. Further studies will be necessary to develop strategies to isolate these fungi into culture.