Soil pH is a Key Determinant of Soil Fungal Community Composition in the Ny-Ålesund Region, Svalbard (High Arctic)

This study assessed the fungal community composition and its relationships with properties of surface soils in the Ny-Ålesund Region (Svalbard, High Arctic). A total of thirteen soil samples were collected and soil fungal community was analyzed by 454 pyrosequencing with fungi-specific primers targeting the rDNA internal transcribed spacer (ITS) region. The following eight soil properties were analyzed: pH, organic carbon (C), organic nitrogen (N), ammonium nitrogen (NH4+-N), silicate silicon (SiO42--Si), nitrite nitrogen (NO2--N), phosphate phosphorus (PO43--P), and nitrate nitrogen (NO3--N). A total of 57,952 reads belonging to 541 operational taxonomic units (OTUs) were found. of these OTUs, 343 belonged to Ascomycota, 100 to Basidiomycota, 31 to Chytridiomycota, 22 to Glomeromycota, 11 to Zygomycota, 10 to Rozellomycota, whereas 24 belonged to unknown fungi. The dominant orders were Helotiales, Verrucariales, Agaricales, Lecanorales, Chaetothyriales, Lecideales, and Capnodiales. The common genera (>eight soil samples) were Tetracladium, Mortierella, Fusarium, Cortinarius, and Atla. Distance-based redundancy analysis (db-rda) and analysis of similarities (ANOSIM) revealed that soil pH (p = 0.001) was the most significant factor in determining the soil fungal community composition. Members of Verrucariales were found to predominate in soils of pH 8–9, whereas Sordariales predominated in soils of pH 7–8 and Coniochaetales predominated in soils of pH 6–7. The results suggest the presence and distribution of diverse soil fungal communities in the High Arctic, which can provide reliable data for studying the ecological responses of soil fungal communities to climate changes in the Arctic.


INTRODUCTION
The Arctic tundra covers approximately 5% of Earth's land surface (Nemergut et al., 2005) and is one of the most extreme environments on Earth, characterized by low temperatures, frequent freeze-thaw and wet-dry cycles, and low organic matter content. The Arctic tundra is particularly sensitive and vulnerable to global climate change that a rapid rate of climate warming has been observed in recent decades. Arctic average temperatures have increased at almost twice the global average rate over the past 100 years (Solomon, 2007). Some marked environmental changes in the Arctic have been observed by researchers, such as retreat of glaciers, melting of permafrost and sea ice, changes of vegetation, alteration of nutrient cycling and energy flow (Overpeck et al., 1997;Schuur et al., 2009;Pearson et al., 2013).
Arctic microbes can be viewed both as sentinels and amplifiers of global change (Vincent, 2010). Understanding the structure of Arctic soil microbial communities is essential for predicting the response of the Arctic tundra to climate change (Blaud et al., 2015a). Along with soil bacteria, fungi play important roles in driving mineral and energy cycles in Arctic soil environments. They can function as mutualists (mycorrhizae, endophytes, lichens) and decomposers and may affect the carbon balance of terrestrial ecosystems subjected to climate change (Timling and Taylor, 2012). Hitherto, a few studies carried out provide only limited insight to fungal diversity in the Arctic soils. To the best of our knowledge, soil fungal diversity has been reported from different Arctic regions, including Franz Joseph Land (High Arctic) (Bergero et al., 1999), Svalbard (High Arctic) (Robinson et al., 2004;Kurek et al., 2007;Geml et al., 2012;Singh et al., 2012;Ali et al., 2013;Blaud et al., 2015b), Toolik Lake site in Alaska (Low Arctic) (Wallenstein et al., 2007;Deslippe et al., 2012), Greenland (Low and High Arctic) (Meyling et al., 2012), Siberian tundra (High Arctic) (Gittel et al., 2014).
Most of these previous studies focused on the diversity of cultured free-living soil fungi in the Arctic using traditional isolation methods (Bergero et al., 1999;Robinson et al., 2004;Kurek et al., 2007;Meyling et al., 2012;Singh et al., 2012;Ali et al., 2013). However, the cultivation-based methods likely do not reflect the actual diversity of soil fungi because of their selectivity. Additionally, a few studies surveyed soil fungi in the Arctic using conventional DNA-based molecular methods (e.g., DGGE, cloning approaches, q-PCR) (Wallenstein et al., 2007;Deslippe et al., 2012;Geml et al., 2012;Blaud et al., 2015b), which are relatively low taxonomic resolution techniques. In recent years, diversity, and community composition of soil fungi and other microbes in the Arctic tundra have been revealed by high-throughput sequencing (Gittel et al., 2014), which can significantly enhance the characterization of fungal diversity compared to traditional methods. Whilst the previous findings have improved our understanding of soil fungal diversity in the Arctic tundra, the environmental factors shaping the soil fungal community composition in the Arctic, especially High Arctic, are not well understood.
Svalbard is an Arctic archipelago, which is geographically isolated from mainland Eurasia and entirely within the High Arctic Zone. About 59% of its land area is covered by glaciers (Liestøl, 1993). Accordingly, the soil environment in Svalbard is a good representation for the study of the soil fungal communities in the High Arctic. Moreover, no pyrosequencing study has yet been carried out to provide insight to the fungal communities that inhabit soil samples from Svalbard (High Arctic). The aim of this study was to use 454 pyrosequencing to investigate the soil fungal communities in the Ny-Ålesund Region (Svalbard, High Arctic) to address the following questions: (1) What are soil fungal diversity and community composition in this High Arctic region? (2) What is the key environmental factor that determines the soil fungal community structure in this region? (3) How does the key factor affect soil fungal taxonomic groups in this region?

Study Sites and Sample Collection
The study site was located in the Ny-Ålesund Region (78 • 55 N, 11 • 56 E), which is on the west coast of Spitsbergen, the largest island of the Svalbard archipelago. It has the lowest temperatures during February, at an average of -14 • C, and a high average of 5 • C during July 1 . The tundra around Ny-Ålesund experiences a short growing season (June to August) when the temperature typically rises only a few degrees above the freezing point. The local area has a silty clay and sand gley soil that has a thin organic soil cover. The soil thawing begins in early June, and the thaw depths reach their maximum (1.6-2.0 m) in the beginning of September and remain approximately constant until at least the mid of the month (Zhu et al., 2012).
Sampling was performed during China's Arctic expedition in July 2013. The sample collection was under ethical approval of the Svalbard Science Forum, Kings Bay As, and the Chinese Arctic and Antarctic Administration (CAA), the State Oceanic Administration (SOA) of China. Surface soil samples from bare ground without vegetation (∼10 cm 2 ) were collected from the top 5 cm at each sampling sites using a sterile shovel. A total of thirteen samples were collected and directly put into TWIRL'EM sterile sampling bags (Labplas Inc., Sainte-Julie, QC, Canada). Soil samples were placed in centrifuge tubes at −20 • C in the Yellow River Station (China) for about 20 days and taken to the home laboratory in China by plane. Soil samples were then frozen at −80 • C until nucleic acid extraction. The locations of the thirteen soil samples are shown in Table 1.

Physical and Chemical Properties of Soil
A total of eight soil properties were assessed, including pH, organic carbon (C), organic nitrogen (N), ammonium nitrogen (NH 4 + -N), silicate silicon (SiO 4 2− -Si), nitrite nitrogen (NO 2 − -N), phosphate phosphorus (PO 4 3− -P) and nitrate nitrogen (NO 3 − -N). Soil pH was measured by adding 10 ml of distilled water to 4 g of soil and recording pH using a pH electrode (PHS-3C, Shanghai REX Instrument Factory, Shanghai, China). Analysis of organic C and organic N was performed using an Elemental Analyzer (EA3000, Euro Vector SpA, Milan, Italy). The other five properties were analyzed using a High Performance Microflow Analyzer (QuAAtro, SEAL Analytical GmbH, Norderstedt, Germany).

DNA Extraction and PCR Amplification
Genomic DNA was extracted from an aliquot of 0.25 g of wet soil from each sample using a PowerSoil DNA Isolation Kit (MO BIO Laboratories, San Diego, CA, USA) according to the manufacturer's instructions (three replicates for each sample). The combined DNA extracts were then used for the subsequent PCR and sequencing experiments. The primer pair

Pyrosequencing
After purification using the AxyPrep DNA Gel Extraction Kit (Axygen Biosciences, Inc., USA) and quantification using QuantiFluor-ST (Promega Corporation, USA), equimolar mixtures of multiple amplicons were used for pyrosequencing on a Roche 454 GS FLX+ Titanium platform (Roche 454 Life Sciences, US) according to standard protocols. The raw sequences were deposited in the NCBI sequencing read archive (SRA) under Accession No. SRP067367.

Pyrosequencing Data Processing
Raw sequence data from pyrosequencing were processed using QIIME 1.8.0 software (Caporaso et al., 2010). Briefly, the sequence libraries were split and denoised to avoid diversity overestimation caused by sequencing errors, such as sequences with an average quality score <20 over a 50 bp sliding window, sequences shorter than 200 bp, sequences with homopolymers longer than six nucleotides, and sequences containing ambiguous base calls or incorrect primer sequences. Operational Taxonomic Units (OTUs) were clustered with a 97% similarity cutoff using UPARSE (Edgar, 2013), and chimeric sequences were identified and removed using UCHIME (Edgar et al., 2011). Singleton OTUs, and OTUs that were assigned to non-fungal organisms and had unreliable BLAST matches (max score below 200 or aligned query sequence below 200 bp) were removed. These OTUs were then used as a foundation for calculating alphadiversity and beta-diversity metrics using QIIME 1.8.0 software (Caporaso et al., 2010).

Statistical Analyses
Sequences representing the OTUs were subjected to BLASTn search in UNITE 2 to determine their taxonomic affiliation and GenBank 3 to determine their originally reported habitat. Statistical analyses of the OTU diversity of each soil sample via Chao1, Good's coverage estimator, and Shannon's index (H ) were performed using QIIME 1.8.0 software (Caporaso et al., 2010). The abundance-based Bray-Curtis similarity coefficient was used to examine the dissimilarity of different soil samples. The relevance of environmental factors in explaining the distribution patterns of fungal communities in different soil samples was analyzed by distance-based redundancy analysis (db-RDA) using the R 3.1.1 statistical software. An analysis of similarities (ANOSIM) was performed using QIIME 1.8.0 software (Caporaso et al., 2010) to determine whether different soil types had significantly different fungal communities. A linear discriminant analysis effect size (LEfSe) method was used to identify the significantly different fungal groups in different soil types (Segata et al., 2011).

Physical and Chemical Properties of Soil
The pH of soil samples collected were greatly different form each other, which were in the range of 6.00 to 8.54. The organic carbon contents ranged from 0.25 to 12.15%, whereas organic nitrogen contents were below detection to 1.41%. The ammonium nitrogen (NH 4 + -N) concentrations ranged from 1.146 to 88.325 µg/g soil, which was much higher than the concentrations of nitrate nitrogen (NO 3 − -N, 0.256-4.601 µg/g soil) and nitrite nitrogen (NO 2 − -N, 0.048-0.792 µg/g soil) at all sampling sites. The concentrations of phosphate phosphorus (PO 4 3− -P) were negligible (0.001-0.134 µg/g soil). The concentrations of silicate silicon (SiO 4 2− -Si) were from 0.679 to 8.277 µg/g soil ( Table 1). In addition, significant correlations were found between pH and some soil properties, including silicate silicon (r = 0.768, p < 0.01), organic C (r = −0.596, p < 0.05), and organic N (r = −0.666, p < 0.05) (Supplementary Table S1).

Soil Fungal Diversity and Community Structure
The raw data from the 13 soil samples consisted of 106,003 sequence reads, of which 57,952 reads representing 541 fungal OTUs were included in the final matrix after quality filtering and processing. The number of OTUs in different soil samples ranged from 56 to 214 ( Of the 541 OTUs, 343 belonged to Ascomycota, 100 to Basidiomycota, 31 to Chytridiomycota, 22 to Glomeromycota, 11 to Zygomycota, 10 to Rozellomycota, whereas 24 belonged to unknown fungi (Supplementary Table S2). Ascomycota dominated in the 12 soil samples, whereas Basidiomycota dominated in only one sample S10.3 ( Figure 1A).
Among the 95 known genera detected in the present study, the most commonly identified genera included Tetracladium  Table S2).
In terms of original reported habitats, most of matching sequences with high similarity (≥97%) in GenBank were derived from fungi found in soils the Arctic tundra, including Svalbard and North American Arctic. Additionally, some matching sequences were also reported from other habitats (e.g., plant tissues, glacier ice, sediments) in and beyond the Arctic regions (Supplementary Table S2).

The Correlation Between Soil Fungal Communities and Geochemical Properties
Distance-based redundancy analysis (db-RDA) (Figure 2) and Monte Carlo permutation test (Table 3) were performed to examine the relationship between the eight soil geochemical factors and soil community composition. The combination of the eight geochemical factors showed a significant correlation with soil fungal community structure (F = 1.175068, p = 0.031). These factors explained 70.15% of the soil fungal community variation, while 29.85% of the variation was not explained by any of the selected eight geochemical factors. Among the selected geochemical factors, pH (r 2 = 0.8851, p = 0.001) was the most significant geochemical factor that drive the soil fungal communities in this region. The factors SiO 4 2− -Si (r 2 = 0.7159, p < 0.01), organic C (r 2 = 0.5844, p < 0.05), and organic N (r 2 = 0.5576, p < 0.05) were also significantly correlated with the soil fungal communities in this region. The other four factors, including NH 4 + -N, NO 2 − -N, NO 3 − -N, and PO 4 3− -P were not significantly correlated with soil fungal community composition ( Table 3).

DISCUSSION
The aim of the present study was to analyze the soil fungal communities in the High Arctic tundra using high-throughput pyrosequencing. This study demonstrated a close relationship  between soil fungal communities and certain soil properties in the Ny-Ålesund Region, which was not clarified before in this High Arctic region.

Soil Fungal Communities in High Arctic Tundra are Highly Diverse
Despite the geographic isolation and High Arctic climate of Svalbard, the soil fungal communities are surprisingly diverse in this High Arctic region. Members of Ascomycota were more frequently identified in soils than those of Basidiomycota and Chytridiomycota. Additionally, Zygomycota, Glomeromycota, and Rozellomycota were less abundant in the Arctic soil environments. These findings were consistent with previous studies of soil fungi in Arctic tundra. Gittel et al. (2014) reported that the most abundant phyla were Ascomycota and Basidiomycota and other phyla (i.e., Chytridiomycota, Glomeromycota, and Zygomycota) accounted for only a minor fraction in Siberian tundra (High Arctic). Wallenstein et al. (2007) reported that soil fungal communities were dominated by Basidomycota and there were also a large proportion of Ascomycota, and lesser relative amounts of Zygomycota, Chytridmycota, and Glomeromycota in Alaska (Low Arctic).
The dominant fungal orders detected in this study were Helotiales (Leotiomycetes) in Ascomycota, and Agaricales (Agaricomycetes) in Basidiomycota. The order Helotiales was previously observed in plant rhizosphere in alpine and Arctic regions (Bjorbaekmo et al., 2010;Walker et al., 2011). This order represents the largest and the most diverse group in the Leotiomycetes (Wang et al., 2006) and members of the order can function as mycorrhizae, ectomycorrhizal parasites, terrestrial saprobes, and root symbionts (Vrålstad et al., 2002). Deslippe et al. (2012) found that Ascomycota were dominated by Helotiales, and Basidomycota were dominatead by Agaricales in Alaska (Low Arctic). In Siberian tundra (High Arctic), Ascomycota were dominated by Leotimycetes, whereas Basidiomycota were dominated by Agaricomycetes (Gittel et al., 2014).

Soil pH is a Key Factor Determining the Soil Fungal Community Structure in High Arctic Tundra
In previous studies, researchers found that Arctic fungal communities in roots and soils could change in response to environmental factors, such as pH, temperature, precipitation, and nutrient availability (e.g., C/N ratio, P content) (Clemmensen et al., 2006;Fujimura and Egger, 2012;. Both Robinson et al. (2004) and Deslippe et al. (2012) found Arctic fungal community structure differed between organic soils and mineral soils, with this variation between soils probably related to the higher C and other nutrient content in the organic soil.
From db-RDA analysis in this study, soil pH was the most important environmental factor (p = 0.001) that was correlated  with soil fungal communities in the Ny-Ålesund Region. An ANOSIM analysis also revealed that fungal communities inhabiting soils of different pH levels were significantly different (p < 0.05). All these results suggested that soil pH was a key factor shaping the fungal community structure in this High Arctic region. Silicate silicon (p < 0.01), organic C (p < 0.05), and organic N (p < 0.05) were also found to be significantly related to soil fungal community composition in this study. It is generally believed that some soil characteristics (e.g., organic matter) are directly or indirectly related to soil pH levels. Interestingly, significant correlations were observed between pH and other soil properties in the present study, including soluble silicate (r = 0.768, p < 0.01), organic C (r = −0.596, p < 0.05), and organic N (r = −0.666, p < 0.05) (Supplementary Table S1). The organic carbon (C) and nitrogen (N) could be transferred into soils by vegetation and animal activities. Stempniewicz et al. (2007) reported that animal feces, carcasses and plants are important contributors to the nitrogen (N) and phosphorus (P) cycles in High Arctic soils. Liu et al. (2015) found that the vegetation had a significant effect on the content of soluble silicon in soils of Ny-Ålesund.

Some Fungal Phylogenetic Groups Show Significantly Different Responses to Soil pH
Soil pH may directly affect fungal community composition by imposing a physiological constraint on fungal survival and growth and some fungal taxa may be unable to grow or survive if the soil pH falls outside a certain range. Fungi generally grow well in acidic conditions (Dix and Webster, 1995), but some fungi (e.g., Mortierella, Peziza) grow well in neutral to slightly alkaline conditions (Warcup, 1951;Yamanaka, 2003). As not all members of the same phylum behaved in the same way, it is interesting to examine the different responses of fungal phylogenetic groups, especially groups at the lower taxonomic levels. In this study, many phylogenetic groups at order, family, and genus levels showed significantly different responses to soil pH levels (as shown in Figure 3). Interestingly, soil samples in glacier foreland (at high altitude) were generally alkaline (pH 8-9), whereas pH values of soil samples, which are far away from glacier forelands, were generally below 8. Glacial retreat leads to subsequent exposure of foreland soils to atmospheric conditions, thus creating a FIGURE 4 | The abundance histogram of major taxonomic units (relative abundance >2%) which are significantly different among three soil pH levels using Lefse analysis. sequence of change in the soil environments, such as the low levels of nutrients present in foreland soils and the decrease in soil pH that occurs along the chronosequence (Hodkinson et al., 2003). We hypothesize those lichen-forming fungi, e.g., Verrucariales (order), Verrucariaceae (family), and Atla (genus), may serve as pioneer organisms and firstly colonize and dominate in the glacier foreland soils. A previous study also reported that some members of Verrucariaceae prefer sites with high pH (calcareous sites) (Shivarov, 2013;Gueidan et al., 2014). In addition, non-lichen forming fungi predominated in soil samples of pH < 8.0. For example, Sordariales (order) predominated in soil samples of pH 7-8, whereas Coniochaetales (order) predominated in soil samples of pH 6-7. Yamanaka (2003) found that many of the saprotrophic species grew well at pH 7 or 8 and the ectomycorrhizal species showed optimum growth at pH 5 or 6. Most members of Sordariales are saprobic and consist of mostly wood-and dung-inhabiting species, whereas members of Coniochaetales (order) always occur on wood, dung or soil (Zhang et al., 2006).
In summary, soil pH was a key factor in determining the soil fungal community composition in this High Arctic region. In addition, there are many unmeasured environmental factors, which may be related to the soil fungal community composition in the High Arctic tundra, such as vegetation composition and productivity. With a warming climate, an enhanced colonization of vascular plants would then affect soil fungal community structure and thus further experiments should be performed to confirm the influence of plant vegetation in the High Arctic tundra. The other question is as follows: what are the possible ecological roles and functions of these soil fungi in the Arctic tundra? Perhaps the use of other molecular tools (e.g., transcriptome analysis using next-generation sequencing) will allow us to clarify their ecological functions in Arctic tundra soils.

AUTHOR CONTRIBUTIONS
TZ planned the study, collected samples, conducted lab work, and wrote the manuscript; NW performed geochemical analyses of soil samples; HL and YZ conducted parts of lab work; LY contributed with planning the project and revising the manuscript. All authors reviewed the manuscript.