Phenotypes Associated with Pathogenicity: Their Expression in Arctic Fungal Isolates

Around 85% of the environments on Earth are permanently or seasonally colder than 5 °C. Among those, the poles constitute unique biomes, which harbor a broad variety of microbial life, including an abundance of fungi. Many fungi have an outstanding ability to withstand extreme conditions and play vital ecosystem roles of decomposers as well as obligate or facultative symbionts of many other organisms. Due to their dispersal capabilities, microorganisms from cryosphere samples can be distributed around the world. Such dispersal involves both species with undefined pathogenicity and potentially pathogenic strains. Here we describe the isolation of fungal species from pristine Arctic locations in Greenland and Svalbard and the testing of the expression of characteristics usually associated with pathogenic species, such as growth at 37 °C, hemolytic ability, and susceptibility to antifungal agents. A total of 320 fungal isolates were obtained, and 24 of the most abundant and representative species were further analyzed. Species known as emerging pathogens, like Aureobasidium melanogenum, Naganishia albida, and Rhodotorula mucilaginosa, were able to grow at 37 °C, showed beta-hemolytic activity, and were intrinsically resistant to commonly used antifungals such as azoles and echinocandins. Antifungal resistance screening revealed a low susceptibility to voriconazole in N. albida and Penicillium spp. and to fluconazole in Glaciozyma watsonii and Glaciozyma-related taxon.


Introduction
Despite appearing devoid of life, the poles are now recognized as biomes [1], and it has been repeatedly demonstrated that these unique environments are colonized by a wide variety of microbial life [2]. In addition to bacteria, a great diversity of psychrophilic, as well as psychrotolerant fungi inhabit the cryosphere, where they play important roles in the ecosystem [3]. For example, a strong association between the abundance and diversity of fungi and the blooming of algae on the dark surface ice of the Greenland Ice Sheet has been described [4]. Particularly common are species of the genera Penicillium, Thelebolus, Alternaria, Rhodotorula, Vishniacozyma, Aspergillus, and Cladosporium [3][4][5][6].
Due to their small size, microorganisms generally disperse over large distances, decreasing the impact of geographical barriers [7,8]. At least for some species, a constant exchange of populations between fairly remote habitats is possible [9]. For fungi, and in particular those inhabiting the poles, the Arctic fungal strains displaying phenotypes associated with pathogenicity should be investigated. Accordingly, in the present study we investigated fungal isolates from the Arctic region for three phenotypes commonly associated with pathogenicity in humans, i.e., growth at 37 • C, expression of hemolytic activity on bovine blood agar plates, and resistance to common antifungal therapeutics.

Site and Sample Description
Fieldwork campaigns were conducted in the summer seasons of the years 2016-2018, on the SW coast of Greenland (Greenland Ice Sheet -GrIS, and Russell Glacier) and Svalbard (Ny-Ålesund, Svalbard, Norway). Samples of snow, supraglacial ice and water, subglacial ice, glacial meltwater, and air were obtained from the locations described in Table 1. Supraglacial ice collected on the GrIS was divided based on the visible abundance of living dark glacier algae on its surface into dark ice samples (containing high biomass of glacier algae) and clear ice samples (without visible biomass of glacier algae). Subglacial ice samples were collected from three polythermal glaciers in Svalbard: Midtre Lovénbreen, Vestre Brøggerbreen, and Pedersenbreen. Samples were collected using sterilized tools and sterile nitrile gloves to avoid contamination, and organized into sterile Whirl-Pak ® plastic bags or sterile plastic bottles. All the samples were processed within a few hours of their collection at the primary ice camp on the Greenland Ice Sheet and the NERC Arctic Research Station (Ny-Ålesund).

Fungal Cultivation and Isolation
The outer surface of subglacial and supraglacial ice was discarded after melting at room temperature. Snow and ice samples were then melted at 4 • C prior to analysis. All the samples were filtered through Milli-pore membrane filters (0.45 µm pore size) in triplicate, and filters were placed onto different media. Minimal medium (MM) and synthetic nutrient-poor agar (SNA) [29,30] were used to isolate oligotrophic fungi; dichloran rose bengal chloramphenicol agar (DRBC) [31] (Biolife, Milan, Italy) was used as a general-purpose medium; Reasoner's 2A agar (R2A) [32] (Biolife) was used to isolate bacteria; dichloran 18% glycerol (DG18) medium [33] (Biolife, Milan, Italy) was used to obtain moderate xerophilic and xerotolerant species. To prevent bacterial growth, the media contained chloramphenicol (50 mg/L), except for R2A. Plates were incubated at 15 • C for up to two months in sterile plastic bags. All strains obtained were deposited in the Ex Culture Collection of the Infrastructural Centre Mycosmo (MRIC UL) at the Department of Biology, Biotechnical Faculty, University of Ljubljana, Slovenia.

Fungal Identification
Fungi were transferred onto fresh MEA (malt extract agar) or PDA (potato dextrose agar, Biolife) media for a week to obtain pure cultures. Filamentous fungi DNA was extracted following the protocol described in [4], while yeast DNA was extracted using the PrepMan Ultra reagent (Applied Biosystems, Foster City, CA, USA) according to the manufacturer instructions.
Amplification of the ITS region using the primers ITS5 and ITS4 [34] was performed for filamentous fungi, while amplification of the D1 and D2 domains of the LSU gene using NL1 and NL4 primers [35] was performed for yeasts as described in [6]. For isolates belonging to Penicillium and Cladosporium spp., additional marker genes were amplified, such as β-tubulin gene (benA) with Ben2f and Bt2b primers [36], and partial sequence of actin gene (act) with ACT-512F and ACT-783R primers [37], respectively. Subsequently, amplicons were Sanger-sequenced (Microsynth AG, Balgach, Switzerland) and the resulting sequences were aligned using MUSCLE software [38], implemented in the MEGA7 package [39] and compared against the GenBank database using the nBLAST platform. Sequences are available in the NCBI GenBank nucleotide database ( Table 2).

Thermotolerance Assay
Fungal isolates were cultivated onto MEA and SDA (Sabouraud dextrose agar, Biolife, Milan, Italy) media and incubated at 37 • C for up to seven days to test their ability to grow at high temperatures.

Hemolytic Assay on Blood Agar
Fungal isolates were cultured on blood agar base (Honeywell Fluka, Charlotte, NC, USA) containing 5% of sterile bovine blood and incubated at both 15 • C and 37 • C for a week. Candida albicans (EXF-14635; ATTC90028) was used as a positive beta-hemolytic control [40,41]. An isolate was considered alpha-hemolytic when a green zone of discoloration was observed in the media, beta-hemolytic when a clear zone was observed denoting complete lysis of the erythrocytes of the blood agar plate, and gamma-hemolytic (not hemolytic) when there was no change noted in the media [42].

Antifungal Susceptibility Test (AFST)
Fungi were cultured at 15 • C for up to four days in SDA prior to the test. AFST was done using MIC-gradient strip method according to the manufacturer's instructions (bioMérieux, Marcy-l'Étoile, France). Inoculum for the AFST was prepared by homogenizing well-isolated colonies/spores from the SDA plate in sterile 0.85% NaCl solution (with the addition of 0.05 g/L of Tween 20 for filamentous fungi) to obtain a turbidity equivalent to 0.5 McFarland standard, approximately corresponding to 1-5 × 10 8 CFU/mL. Suspensions were spread evenly over the entire surface of RPMI-1640 agar plates supplemented with 2.0% glucose (bioMérieux, Marcy-l'Étoile, France) using a sterile cotton swab. Etest gradient strips were then applied to the agar surface. Inverted plates sealed with parafilm were incubated in sterile plastic bags at 15 • C to allow the optimal growth of the Arctic strains. Fungal growth was checked after two and seven days. The antifungals tested belonged to two different antifungal classes: azoles and echinocandins. The azoles used were voriconazole (VO-range: 32-0.002 µg/mL), ketoconazole (KE-range: 32-0.002 µg/mL), and fluconazole (FL-range: 256-0.016 µg/mL). The echinocandins used were caspofungin (CS-range: 32-0.002 µg/mL), micafungin (MYC-range: 32-0.002 µg/mL), and anidulafungin (AND-range: 32-0.002 µg/mL). Given their intrinsic resistance to echinocandins [43], basidiomycetes were not tested for that antifungal class. The reference strain, Candida albicans ATCC ® 90028™, was used as a control. MIC (minimum inhibitory concentration) values were determined and are reported in Table 3.

Fungal Isolation and Identification
A total of 320 fungal isolates were obtained from the previously described samples after incubation [6]: 45.9% belonging to Basidiomycota (147/320), 53.8% to Ascomycota (172/320), and only 0.3% belonging to Mucoromycotina (1/320). In Greenland, dark ice was the environment with the most diverse mycobiome, hosting 23 different fungal genera and 28 species, followed by snow with 18 genera and 22 species. Unlike the other sampled environments, dark ice supported the presence of several genera that contain some common plant pathogen and endophyte species, such as Acrodontium, Epicoccum, Bjerkandera, and Microdochium nivale. In Svalbard, samples of subglacial ice collected from three different glaciers hosted, in total, 27 genera and more than 35 different species. The most commonly found species in different samples and environments were Cladosporium sp., Penicillium sp., Venturia sp., Glaciozyma-related taxon, Mrakia sp., Phenoliferia sp., and Rhodotorula svalbardensis pro. tem. The most commonly found and most abundant species in the different environments were selected for further tests (Table 4). Rhodotorula mucilaginosa Basidiomycota αβ 13647 Aureobasidium melanogenum Ascomycota β γ − (little growth) 14635 CTRL Candida albicans Ascomycota γ β α: alpha-hemolytic, β: beta-hemolytic, γ: gamma-hemolytic, + signifies a stronger phenotype, − signifies a weaker phenotype, NG: no growth. * α-hemolysis is noted for this isolate even though the phenotype does not correspond to the exact description of an α-hemolytic organism.

Hemolytic Assay on Blood Agar
Results of the hemolytic assay on bovine blood agar are listed in Table 3. The tested fungal species showed both alpha-and beta-hemolytic phenotypes. More than half of the species (13/25) were not hemolytic at 15 • C. Aureobasidium pullulans and A. melanogenum were the only isolates that presented a clear and strong beta-hemolytic reaction at 15 • C (Figure 1), whereas the remaining strains (eight out of 25), such as Preussia sp., Ph. psychrophenolica, Neocucurbitaria sp., and Oleoguttula mirabilis displayed alpha-hemolysis. Penicillium bialowiezense-like had zones of both alpha-and beta-hemolysis (Figure 1), while P. chrysogenum showed only a (weak) alpha-hemolytic reaction. Pseudogymnoascus sp. presented a brownish discoloration of the agar not attributable to any of the known hemolytic reactions. Rhodotorula mucilaginosa was the only isolate with a beta-hemolytic reaction at 37 • C. No strains were alpha-hemolytic at this temperature.

Antifungal Susceptibility Test (AFST)
Antifungal susceptibility was evaluated in all the selected species and results of the test are reported in Table 3.

Antifungal Susceptibility Test (AFST)
Antifungal susceptibility was evaluated in all the selected species and results of the test are reported in Table 3.

Discussion
Arctic fungi are extremophiles able to withstand environmental conditions that are too harsh for most other organisms. Their great adaptability and stress tolerance, together with the ability of yeasts and fungal spores to spread in the atmosphere and water, and the selective pressure induced by global environmental changes, may be increasing the risks they present as emerging fungal pathogens. Exoenzyme production, resistance to antifungals, and the ability to adapt to host body temperature are some of the virulence-associated characteristics shared by pathogenic fungi [19,[44][45][46]. At least one of the abovementioned virulence-associated phenotypes was expressed by several of our Arctic isolates.
Thermotolerance is a crucial virulence factor for establishing the invasive type of infection in humans and is also the trait lacking in most fungi. Typically, psychrotolerant strains able to survive at higher temperature might bear a significant pathogenic potential [15]. All the selected fungal species were isolated at an incubation temperature of 15 • C, showing a psychrotolerant or psychrophilic nature. Growth tests at higher temperature (37 • C) revealed the abilities of certain strains to adapt and grow in broad temperature ranges. Naganishia albida, Rhodotorula mucilaginosa, and Aureobasidium melanogenum are known polyextremotolerant and emerging pathogenic species with clinical relevance, and together with Penicillium chrysogenum showed their ability to grow at 37 • C despite their isolation from a very cold habitat.
The first screening of hemolytic phenotypes on blood agar in environmental Arctic fungi reported here revealed that Aureobasidium pullulans, A. melanogenum (at 15 • C) and Rhodotorula mucilaginosa (at 37 • C) had the strongest beta-hemolytic reaction, an ability already reported in strains of these fungi isolated from hypersaline water and subglacial ice [47]. Also, penicillia (Penicillium bialowiezense-like and P. chrysogenum) exhibited some degree of hemolysis, in agreement with other studies on airborne fungi [48,49]. P. chrysogenum strains from indoor air were found to produce chrysolysin, a proteinaceous hemolysin, only when incubated at 37 • C [50]. The same hemolysin was not always expressed in outdoor air strains of the same species or at lower temperatures [51]. Based on our results, no reaction was observed in P. crustosum, suggesting that hemolytic abilities are species-or even strain-specific. The latter statement is supported by the authors of [49], who studied airbone fungi in a Russian Arctic settlement and found that Penicillium aurantiogriseum, but not P. canescens and P. simplicissimum, showed strong hemolytic activity. Rhodotorula mucilaginosa was the only isolate with a strong beta-hemolytic activity at 37 • C, and a weaker activity at 15 • C, again highlighting the importance of temperature as a triggering factor of virulence-associated traits [50]. Pseudogymnoascus spp. are known to have many extracellular enzymes that can cause blood cells lysis [51]. Neocucurbitaria sp., Preussia sp., Phenoliferia psychrophenolica, and Oleoguttula sp. showed some degree of hemolysis. The knowledge of hemolysis in fungal species is extremely limited compared to bacteria. Fungal hemolysins are pore-forming exotoxins capable of lysis of red blood cells and nucleated cells [52]. Little is known about their role and mechanisms of action, but research highlighted that many well-known pathogenic fungi, such as Aspergillus flavus and A. fumigatus, as well as Candida albicans and Cryptococcus neoformans, produce several cytolysins [52]. The expression of fungal hemolysins has been proposed as a virulence factor and as a survival strategy for opportunistic fungi [53][54][55]. The iron released by the lysis of red blood cells is an essential factor in fungal growth [56]; it can facilitate hyphal infection [53] and favors the transition from fungi with a commensal lifestyle to invasive pathogens [57]. However, while fungal hemolysin expression can provide an opportunity for colonization and infection, research suggests that the compounds might have functional roles unrelated to pathogenesis, such as hyphal aggregation for fruiting body formation [58,59] and regulation of fungal growth [52]. Moreover, their functional role as a competitor in specific ecological niches [60] and plant pathogenesis has also been proposed [61]. Due to the limited nutrient sources in the Arctic, hemolysins could also provide an ecological advantage to fungi, which are well-known decomposers, and help with obtaining limiting nutrients by disrupting the membranes of other microorganisms and thus accessing their intracellular content.
Despite the lack of validated standard antifungal treatment protocols for non-neoformans cryptococcosis, the literature supports the use of fluconazole for consolidation and maintenance therapy [22]. Our results showed that fluconazole and voriconazole were only weakly active against N. albida EXF-12581 in vitro (MIC = 64 µg/mL and MIC = 2 µg/mL, respectively). Ketoconazole was the only azole that showed higher in vitro susceptibility (MIC = 0.047 µg/mL). This yeast is, like all Cryptococcus-like fungi, intrinsically resistant to the echinocandin class.
Filobasidium magnum, formerly Cryptococcus magnus, has rarely been isolated from mammals, with only four case reports [64][65][66]. Compared to other AFST data, it is in vitro susceptible to voriconazole with an at least two 2-fold dilution steps lower MIC [46]. According to other studies on environmental isolates [67,68], Naganishia albida and Filobasidium magnum are known to be fluconazole-resistant.
Penicillium spp. and Cladosporium spp. have their place in clinical mycology and are occasionally reported as agents of infection in immunocompromised patients [76,77]. Like all filamentous fungi, Penicillium spp. and Cladosporium spp. are resistant to fluconazole [20], with MICs ranging from 16 to 256 µg/mL for our isolates. All three Penicillium species tested revealed high in vitro MICs for voriconazole (up to 1.5 µg/mL) and ketoconazole (up to 12 µg/mL). Clinical cases of invasive mycosis caused by voriconazole-resistant Penicillium species previously not recognized as primary pathogens have been reported [77]. This is particularly worrying since voriconazole is the first-line therapy for invasive mold infections. Resistance mechanisms responsible for acquired azole resistance in filamentous fungi include point mutations of the drug target gene, overexpression of the target gene and of genes encoding efflux pump proteins, and the formation of biofilm [78]. Those mechanisms have already been observed in Penicillium digitatum and are related to the massive use of fungicides in agriculture [78]. Whether the high voriconazole MIC observed in Arctic penicillia is a consequence of acquired resistance and an indication of the dispersion of resistant species from temperate areas to the Arctic or even of evolution of resistant strains in the Arctic is an intriguing question for future studies.
In general, compared to other echinocandins, caspofungin exhibited higher MICs for all tested isolates: four 2-step dilutions higher MIC 50 than those of anidulafungin and micafungin (0.125 µg/mL for caspofungin versus 0.008 µg/mL for anidulafungin and micafungin). MIC 50 is usually a species-or strain-specific value and is used here only to facilitate comparisons among different antifungal classes. Such discrepancies are typical of echinocandins. Caspofungin susceptibility is known to show inter-and intra-laboratory variability and should be interpreted in the context of susceptibility to the other two echinocandins [79].

Conclusions
This study was the first effort to explore potentially pathogenic phenotypes of environmental Arctic fungi. To the best of our knowledge, this includes the first screening for hemolytic ability of Arctic environmental fungal species on blood agar, identifying several species with marked alpha-or beta-hemolytic activity, with one expressing the phenotype at 37 • C and several others able to grow at this temperature despite their polar origins. This is also the first assessment of antifungal susceptibility in Arctic fungi isolated from environments considered to be pristine. We report unexpectedly low susceptibilities for voriconazole in N. albida and Penicillium spp. and for fluconazole in Glaciozyma watsonii and the Glaciozyma-related taxon, which should be investigated further to explore their potential clinical implications. The present study contributes to a better understanding of microbiological intrinsic resistance, the development of acquired resistance in environmental isolates, and, most importantly, the emergence of intrinsically resistant species as opportunistic pathogens.
Our research stresses the importance of the identification of virulence-associated phenotypes in Arctic fungi, especially in the context of the increased release of these fungi into the surroundings due to climate change and the recent increment of opportunistic mycoses. The potential hazards these fungi may represent for human and animal health warrant further investigation.