Fungi in the Antarctic Cryosphere: Using DNA Metabarcoding to Reveal Fungal Diversity in Glacial Ice from the Antarctic Peninsula Region

We assessed fungal diversity present in glacial from the Antarctic Peninsula using DNA metabarcoding through high-throughput sequencing (HTS). We detected a total of 353,879 fungal DNA reads, representing 94 genera and 184 taxa, in glacial ice fragments obtained from seven sites in the north-west Antarctic Peninsula and South Shetland Islands. The phylum Ascomycota dominated the sequence diversity, followed by Basidiomycota and Mortierellomycota. Penicillium sp., Cladosporium sp., Penicillium atrovenetum, Epicoccum nigrum, Pseudogymnoascus sp. 1, Pseudogymnoascus sp. 2, Phaeosphaeriaceae sp. and Xylaria grammica were the most dominant taxa, respectively. However, the majority of the fungal diversity comprised taxa of rare and intermediate relative abundance, predominately known mesophilic fungi. High indices of diversity and richness were calculated, along with moderate index of dominance, which varied among the different sampling sites. Only 26 (14%) of the total fungal taxa detected were present at all sampling sites. The identified diversity was dominated by saprophytic taxa, followed by known plant and animal pathogens and a low number of symbiotic fungi. Our data suggest that Antarctic glacial ice may represent a hotspot of previously unreported fungal diversity; however, further studies are required to integrate HTS and culture approaches to confirm viability of the taxa detected.


Introduction
Despite its generally extreme conditions, Antarctica hosts diverse environments dominated by microorganisms, which are present in the most extreme environments of the continent [1][2][3]. Antarctica's continental ice sheets contain the largest volume of glacial ice, inherently characterized by unfavorable conditions to life, including low temperatures, low water activity, low nutrient availability, and, in their surface layers, exposure to high levels of solar radiation [4,5]. Glacial ice is formed through the precipitation, accumulation, compaction and recrystallization of snow. Fungi are among the microorganisms reported from components of the Antarctic cryosphere such as soils, snow, rocks, and associated with plants and animals [6]. However, despite their recognized importance for ecosystem functioning in Antarctica and elsewhere, few studies have attempted to recover and identify fungal species from glacial ice and, until now, few species, mainly representing the phyla Ascomycota, Basidiomycota and Mortierellomycota, have been characterized from this environment [7][8][9][10][11][12].
Glacial ice can contain spores and mycelial fragments of fungi deposited from the air column, both on contemporary and palaeo timescales [5,[12][13][14]. Sonjak et al. [15] suggested that viable fungal cells obtained from Arctic and Antarctic glacial ice may range in age from 10,000 to 140,000 years, which represent aeolian transport of propagules of both local and distant origin. In addition, Rosa et al. [16] studied air samples in the Antarctic Peninsula region and showed the presence of fungi in the airspora, supporting the possibility of dispersal over different geographic scales around Antarctica. To date, very few studies have addressed fungal diversity present in Antarctic glacial ice, with those available being based on cultivation techniques [5]. Furthermore, fungal diversity present in Antarctic glacial ice has not been assessed using cutting edge modern DNA metabarcoding techniques. In the current study, we assessed the fungal diversity, richness, abundance and distribution in glacial ice sampled in the different Antarctic sites using DNA metabarcoding through high-throughput sequencing (HTS).

Ice Sampling
Three "bergy bits" (glacial ice fragments) each of approximately 20 kg mass were collected adjacent to the ice fronts of seven marine-terminating glaciers in the South Shetland Islands and the north-west Antarctica Peninsula during the austral summer season in December 2015 and December 2016 (Fig. 1). Each was collected using sterile suits and gloves to minimize contamination risk. In the microbiology laboratory on board the Brazilian polar research vessel Admiral Maximiano, each sample was broken into smaller pieces, and surface decontamination carried out using 5% sodium hypochlorite (10 s), sterilized distilled water (10 s) and exposure to ultraviolet radiation (10 min) [12,17]. The samples were melted and a total of 12-15 L of the resulting water filtered through 47 mm diameter (Millipore) membranes (three membranes per sampling site, each using 4-5 L) until each membrane became saturated.

DNA Extraction, Data Analyses, and Fungal Identification
The three membranes resulting from filtering the melted ice from each sampling site were processed together in order to increase DNA yield due the low microbial biomass usually present in the glacial ice of Antarctica [12]. Total DNA was extracted using 0.5 mL extraction buffer [sodium dodecyl sulfate (SDS) 10%], left at 55 °C for 18 h, followed by 165 µL NaCl (5 M) and 165 µL cetyltrimethylammonium bromide (CTAB, 10%), then 600 µL chloroform was added and the mixture centrifuged (Eppendorf/Germany) at 13,000 rpm for 10 min. The supernatant was cleaned using the QIAGEN DNeasy PowerClean cleanup Kit. Extracted DNA was used as a template for generating PCR-amplicons. The internal transcribed spacer 2 region (ITS2) of the nuclear ribosomal DNA was used as a DNA barcode for molecular species identification [18,19]. PCR-amplicons were generated using the universal primers ITS3 and ITS4 [20] and were sequenced at Macrogen Inc. (South Korea) on an Illumina MiSeq sequencer, using the MiSeq Reagent Kit v3 (600cycle) following the manufacturer's protocol.

Diversity, Distribution, and Ecological Analysis
To quantify species diversity, richness, and dominance, we used the following indices: (i) Fisher's α, (ii) Margalef's, and (iii) Simpson's to assess alpha diversity. In addition, the Sorensen and Bray-Curtis similarity indices were used to assess beta diversity among the fungal assemblages. The relative abundance of the ASVs was used to quantify the fungal taxa present in the glacial ice sampled as described by Rosa et al. [28]. Fungal ASVs with relative abundance > 10% were considered dominant, ASVs with relative abundance of 1-10% intermediate, and ASVs with < 1% minor (rare) components of the fungal community. All of the results were obtained with 95% confidence, and bootstrap values were calculated from 1000 iterations. Taxon accumulation curves were obtained using the Mao Tao index. All diversity index calculations and t tests were performed using PAST, version 1.90 [29]. To prepare Kronar charts, QIIME2 taxonomy classifications and the table of taxa abundance were converted to tsv and biom format, respectively. The table of fungal abundance was converted to tsv by using biom convert and combined with taxonomy classification with a custom script krona_qiime.py (https:// github. com/ lokes hbio/ Ampli con_ course/ blob/ master/ krona_ qiime. py). The Krona Tools (v. 2.7.1) [30] program, ktImportText.pl, was used to provide interactive visualization of identified fungi species. Heat map comparison of fungal phyla data between Antarctic islands and Antarctic Peninsula sites was performed with the "pheatmap" using the R-package (https:// www.R-proje ct. org/). Venn analysis to compare the fungal diversity obtained from the different sampling locations was carried out using the program available at http:// bioin forma tics. psb. ugent. be/ webto ols/ Venn/. The functional assignments of fungal ASVs at species and generic levels were assessed using FunGuild [31].

Fungal Taxonomy
We detected a total of 353,879 fungal DNA reads, representing 94 genera and 184 distinct taxa in glacial ice obtained from the seven sampling locations in the South Shetland Islands and north-west Antarctic Peninsula (Suppl. Table 1). The phylum Ascomycota was dominant in all fungal assemblages, followed by Basidiomycota and Mortierellomycota. A single Mucoromycota taxon (Rhizopus arrhizus) was detected at low abundance (Fig. 2). Penicillium sp., Cladosporium sp., Penicillium atrovenetum, Epicoccum nigrum, Pseudogymnoascus sp. 1, Pseudogymnoascus sp. 2, Phaeosphaeriaceae sp. and Xylaria grammica (Ascomycota) were the most dominant taxa (all with > 10% of DNA reads), in rank order. A further 30 taxa were detected at intermediate abundance (< 1% DNA reads). The majority of the fungal ASVs detected (146 taxa; 79.3%) were classified as rare. Thirty-seven taxa could only be assigned to higher taxonomic levels (phylum, class, order, family).

Diversity, Distribution, and Ecology
Mao Tao's rarefaction curves approached a plateau for all sampling locations, indicating that the DNA reads obtained gave a good representation of the fungal sequence diversity present at each (Suppl. Figure 1). Alpha diversity indices across the sampled locations indicated generally high diversity (Fisher α) and richness (Margalef) and moderate dominance (Simpson) indices (Table 1), varying between the different sites. The sequence diversity detected in the Leonardo-Blanchard region (Antarctic Peninsula) was most diverse and rich, and included a wider range of dominant taxa, followed by those of Greenwich Island, when compared with the other sampling locations. The sequence diversity detected in the Sikorsky region (Antarctic Peninsula) displayed the lowest diversity indices. The Leonardo/ Blanchard (Antarctic Peninsula) location had the lowest values of conductivity and salinity.
The beta diversity of the fungal assemblages varied across the different sampling locations (Fig. 3). The presence-absence-based Sorensen index showed that the most similar fungal assemblages were found at Greenwich Island, Traub and Livingston Island, Huron, followed by Antarctic Peninsula, Leonardo-Blanchard and King George Island, Ajax-Stenhouse. However, the abundancerelated Bray-Curtis index indicated that the fungal assemblages from Greenwich Island, Fuerza Aérea and Antarctic Peninsula, Sikorsky showed the highest similarity. In addition, a heat map was used to show the fungal phyla abundance between the glacial ice samples obtained in the Antarctic Island and Antarctic Peninsula sites (Suppl. Figure 2).
The physicochemical properties of all the ice samples were generally similar, except for those from Antarctic Peninsula, Leonard-Blanchard and Antarctic Peninsula, Sikorsky, which displayed the extreme values of conductivity and  Fig. 4). The fungal assemblage detected in the ice sampled in the Antarctic Peninsula, Leonard-Blanchard site, which had the lowest physicochemical parameters, included the lowest number of DNA reads and the highest number of ASVs, and had the highest diversity (Fisher α) and richness (Margalef) indices. In contrast, the fungal assemblage detected in the ice from Antarctic Peninsula, Sikorsky displayed the lowest values of the same diversity parameters. PCA analysis indicated that the conductivity, total dissolved solids, oxidation-reduction potential, pH, and salinity showed negative correlation with the number of taxa, Fisher α, Margalef, and Simpson indices. Twenty-six of the 186 fungal taxa detected were present at all sampling locations (Suppl. Table 2), while 82 taxa were detected at only a single location. However, when the fungal communities detected in the glacial ice samples from the South Shetland Islands and Antarctic Peninsula were compared, 91 (48%) were shared, including the dominant taxa Penicillium sp., Cladosporium sp., Penicillium atrovenetum, Epicoccum nigrum, and Pseudogymnoascus sp. (Suppl. Figure 3). Ecological functional assignments of the taxa detected at generic level are given in Suppl. Table 3. Taxa of 94 genera

Fungal Taxonomy
Glacial ice is considered an extreme and ultra-oligotrophic environment and one of the most challenging natural environments for life globally [2]. Representatives of bacteria, archaea, and fungi have been detected in glacial ice from different cold regions of the planet [15,32]. However, among the microorganisms present in the glacial ice, fungi remain poorly known and few taxa have been reported to date [6,14] in studies based on traditional culturing methods. Many factors, such as extraction, PCR, and primer bias, can influence the outcomes of metabarcoding studies and the numbers of reads obtained [33], thus leading to misinterpretation of absolute abundance [34]. However, Giner et al. [35] concluded that such biases did not affect the proportionality between reads and cell abundance, implying that more reads are linked with higher abundance [36,37].
Our data revealed the presence of rich and diverse fungal sequence diversity in glacial ice collected from the seven different sampling locations. The total sequence diversity was dominated by a relatively small number of taxa of the genera Penicillium, Cladosporium, Epicoccum, Pseudogymnoascus and Xylaria, all members of the Ascomycota. However, the majority of the diversity identified comprised intermediate and rare members of the phyla Basidiomycota, Mortierellomycota and Mucoromycota.
The genera Cladosporium and Penicillium include wellknown cosmopolitan species often detected in the airspora. In Antarctica, different species of Cladosporium have been detected in association with plants and soil [3]. Species of Penicillium are widespread across Antarctica and have been reported in studies of multiple terrestrial substrates including soils [38][39][40], permafrost [41,42], associated with marine macroalgae [43], invertebrates [44], sediments [45,46] and seawater [47]. Penicillium atrovenetum was detected as the dominant fungal sequence present in the gypsum encrustations and carbonate veins of rocks in a polar desert region of continental Antarctica [48]. Members of Pseudogymnoascus occur widely in cold polar, alpine and temperate environments [49][50][51][52]. In Antarctica, they have been reported from soils [49,53,54], associated with plants [55][56][57] and marine macroalgae [58], in freshwater lakes [45] and associated  [59]. Cladosporium, Penicillium and Pseudogymnoascus sequences were detected as dominant fungal sequences present in air and snow samples from the South Shetland Islands [16,28,60]. The dominance of sequences in these genera in the glacial ice examined here is consistent with these fungi being abundant in the air, being deposited (possibly facilitated by snow precipitation) on the glacier surface, and progressively incorporated in the glacial ice as it becomes compacted over time. It is important, however, to note that metabarcoding methodologies detect the presence of DNA sequences, with identification still limited by the available sequence databases, and do not provide any confirmation of viability. Therefore, further specific studies are necessary to determine whether the fungal taxa detected are present in a viable form.
Members of the genus Epicoccum are commonly present in air, soil, decaying vegetation and as endophytes in living plant tissues, with some also being documented producers of bioactive compounds [61]. In Antarctica, species of Epicoccum have been documented in aerobiological studies on Signy Island in the South Orkney Islands [62] and associated with Antarctic marine sponges [63]. The DNA of Epicoccum nigrum was also recently detected in rock surface gypsum encrustations and carbonate veins in the Ellsworth Mountains [48].
The genus Xylaria contains between 570 and 670 recognized species [64], but may include many more yet to be described [65]. Species of Xylaria are important saprophytic fungi found on decomposing wood in temperate and tropical ecosystems [66] and also as plant endophytes [67]. Additionally, members of the genus are among the most prolific secondary metabolite producers [64]. In Antarctica, Xylaria has been reported from soil exposed by glacial retreat on King George Island [68].

Diversity, Distribution, and Ecology
Aside from the eight Ascomycota taxa classified as dominant in the current study, the majority of taxa were of rare or intermediate abundance, and were mostly known as mesophilic fungi. de Menezes et al. [12] reported culturable fungal diversity from the same glacial ice samples as examined here. They documented the presence of 27 taxa belonging to 14 genera. The number of taxa detected and diversity ecological indices calculated using the metabarcoding approach in the current study were approximately seven times greater than those reported by de Menezes et al. [12] ( Table 2). The use of metabarcoding revealed sequence diversity potentially representing a much richer and more diverse fungal community than previously appreciated, including 184 taxa belonging to the 98 genera, among which were fungi not previously reported from Antarctica. The fact that 37 taxa could only be assigned at higher taxonomic levels (phylum, class, order, family) reinforces the evidence that Antarctic environments are likely to host new and/or previously unreported fungal diversity.
Sorensen and Bray-Curtis similarity indices indicated that the beta diversity of the fungal assemblages varied across the different sampling sites, which may be related with ice physicochemical properties. The fungal assemblage detected in the ice sampled in the Antarctic Peninsula, Sikorsky location formed an isolated group based on the fungal taxa present, as well as having the highest values of conductivity and total dissolved solids and the lowest number of ASVs, Fisher's α (diversity) and Margalef index values (richness). However, when the Bray-Curtis index was compared with the physicochemical parameters and alpha diversity indices no other correlations were detected. Our alpha and beta diversity data differed from those reported by de Menezes et al. [12] in their study of culturable fungal diversity from the same glacial ice samples as examined here. The PCA analysis reported by de Menezes et al. [12] showed a positive correlation between pH and the diversity indices only at two sampling sites, differing from the current analyses, which identified a positive correlation between the lowest values of the physicochemical parameters and the highest diversity values. Among the dominant fungi detected in the glacial ice were the genera Penicillium, Cladosporium and Pseudogymnoascus, which have been reported from various different habitats and environments in Antarctica and represent cold-adapted and cosmopolitan fungi [3,5]. Glacial ice originates from snow precipitation, followed by compaction over several years and can represent a cryptic microhabitat for fungi directly related to the atmosphere thousands of years ago that are trapped in the ice matrix [14]. In this context, Cladosporium, Epicoccum and Penicillium are known cosmopolitan genera with high dispersal capabilities and are commonly present in tropical, temperate and polar environments [28]. Their spores and/or mycelial fragments may be constantly deposited by precipitation over many years and, consequently, trapped and preserved in this glacial ice of Antarctica. In contrast, we also detected the DNA of Pseudogymnoascus taxa, which represent psycrotolerant fungi widely reported from different Antarctic and other cold habitat [28]. Finally, the majority of fungi detected as dominant, intermediate and minor components in the glacial ice samples include those known as saprophytic, mutualistic, parasitic and opportunistic taxa [31].

Conclusions
Our data suggest that Antarctic glacial ice may host a hotspot of as yet unreported fungal diversity. Use of a DNA metabarcoding approach revealed the presence of high fungal sequence diversity including a small number of dominant fungi with capabilities for aerial dispersal and a high number of taxa of rare or intermediate abundance. The sequence diversity detected was dominated by saprophytic taxa, followed by known plant and animal pathogens and a small number of symbiotic fungi. The potentially high fungal diversity detected here in glacial ice samples emphasizes the need for further studies characterizing fungal communities of this extreme ecosystem using a combination of culturing and metabarcoding approaches.