!

DEPARTMENT OF BIOLOGICAL AND ENVIRONMENTAL SCIENCES

EXPLORING FUNGAL COMMUNITIES IN MARINE AND COASTAL SEDIMENTS THROUGH METABARCODING AND HIGH-THROUGHPUT DNA SEQUENCING

Alice Retter

Degree project for Master of Science (120 hec) with a major in Biodiversity and Systematics BIO707, Examenskurser biologi, avancerad nivå H17-V18, 60 hec Second cycle Semester/year: Autumn 2017/ Spring 2018 Supervisor: Henrik Nilsson Co Supervisor: Sarah Bourlat Examiner: Bengt Oxelman

Abstract

Fungi are osmoheterotrophic, unicellular or filamentous organisms. They produce spores for sexual and/or asexual reproduction, and the fungal kingdom exhibits a wide range of different lifestyles as symbionts, parasites, or saprotrophs. Mycologists traditionally believed that fungi were a nearly exclusively terrestrial group, but it was gradually realized that fungi are present in aquatic environments as well. We know precious little about fungi in limnic and marine systems, including aspects of their taxonomy, ecology, and geographical distribution. This thesis focuses on marine fungi, which we since a few years back know to be abundant – if only very inconspicuous.

In this thesis, we use metabarcoding to explore species diversity in littoral sediments at two locations on the Swedish west coast. Two transects were sampled from a coastal meadow landscape to the intertidal and benthic zones to assess the fungal community from a taxonomic and functional point of view. This study aims to contribute data on the taxonomic diversity and ecological roles of fungal communities in coastal habitats and in marine sediments, and how these are influenced by various chemical and edaphic parameters.

We found that more than 80 % of the sequences recovered from the benthic zones could not be classified and that the marine sediments were equally diverse as the coastal terrestrial sites. The phylum composition at both localities was dominated by Dikarya, which made up around 33 % of the classifiable OTUs. Within Dikarya, Ascomycota was clearly prevalent. Nearly half of the assigned fungal guilds were unclassifiable, followed by undefined saprotrophs which were in turn slightly more abundant in ocean sediments than on land. Our metadata indicates a transition from C3 to C4 plants from land to ocean, and that these ocean sites contain organisms at a lower trophic level. Our data show that there are predominantly endophytic, parasitic, and pathogenic fungi in the marine environments, which hints at the presence of interesting and currently poorly understood ecological processes. It is also clear from our results that a very large number of marine fungi are in urgent need of taxonomic study and formal description.

This thesis shows that aquatic mycology is only in its infancy, and that there is much work ahead of us to be able to thoroughly describe and understand aquatic systems in regard to their fungal players.

Keywords: Biodiversity – Ecology – Marine fungi – Metabarcoding – High-throughput Sequencing – Fungal diversity

!1

Table of Contents Introduction 1 .......................................................................................................................................Aims 4 ..................................................................................................................................................Materials and methods 4 ......................................................................................................................

Sampling and experimental design 4 ...............................................................................................ITS2 amplicon library preparation and PCR 5 .................................................................................Bioinformatics 6 ...............................................................................................................................Statistical Analysis 7 ........................................................................................................................

Results 8 ...............................................................................................................................................Physicochemical soil properties 8 ....................................................................................................Processed amplicon data 9 ...............................................................................................................Relative abundance tables 10 ...........................................................................................................Distribution of OTUs and their taxonomic composition 12 .............................................................Rarefication curves 13 ......................................................................................................................Diversity indices, number of OTUs and reads 13 ............................................................................

Network plots of site similarities 15 ............................................................................................Data ordination 16 ............................................................................................................................Linear regression 16 .........................................................................................................................Ecological roles 18 ..........................................................................................................................

Discussion 23 .......................................................................................................................................Discussion of results 23 ...................................................................................................................Limitations and shortcomings 24 .....................................................................................................Future studies 25 ..............................................................................................................................

Conclusion 25 .......................................................................................................................................Acknowledgements 25 .........................................................................................................................Reference list 26 ...................................................................................................................................Popular science summary 30 ................................................................................................................

Fantastic marine fungi and where to find them 30 ...........................................................................Abbreviations 30 ..................................................................................................................................Supplements 31....................................................................................................................................

!2

Introduction In the kingdom of Fungi, there may be upwards to six million species, but only around 143 000

have been formally described so far (Kirk, 2018; Blackwell, 2011; Taylor et al., 2014). Many fungi are inconspicuous, have largely cryptic life styles, and most of them are not easily detected. What unites them is their osmoheterotrophic way of life, taking up dissolved carbonic matter by absorption or osmosis. The fungal kingdom contains a wide range of ecologically important lifestyles, including decomposers, parasites or pathogens, or symbionts/mutualists. Here, the function behind feeding is esentially the same across all lifestyles: (1) attachment to the substrate (2) secretion of enzymes that break down complex biomolecules in their surroundings, and (3) absorbing the released nutrients. Fungi also participate in mineral weathering processes (Hoffland et al., 2004) and the release of organically bound nutrients. They reproduce by many means, the dominant mode typically being sexually/or asexually by spore formation. Spores are also the primary means of dispersal for most fungi. As fungi have lost the ability to perform phagocytosis, many of them are highly dependent on the organisms they associate with whether it is a symbiotic or parasitic association (Richards et al., 2012). Conversely, fungi are often also essential to the survival of the organisms they grow with (Blackwell , 2011). Traditionally, fungi were thought of as a primarily terrestrial group of organisms. Fungi growing on land include well-known groups such as yeasts, moulds, polypores, and mushrooms. Many of these produce easily recognized fruiting bodies. Moreover, terrestrial fungi are thought to be the link between the above and below ground diversity, as many of them are partners of mycorrhizal plants (Hooper et al., 2000) or are involved in other symbiotic interactions. Indeed, fungi are known to interact with all major groups of organisms (Taylor et al., 2014).

It was only recently that we learned that fungi are abundant in aquatic environments as well. It is believed that fungi existed and evolved in aquatic environments long before they colonized land, and the transition to land could have happened multiple times (James et al., 2006a; Le Calvez et al., 2009; Richards et al., 2012). Aquatic fungi are generally categorized into freshwater and marine taxa. The number of extant freshwater species is unclear, but it is estimated to be around 2 500 species, and they contain one of the deepest branching clades of fungi discovered so far – the phylum Cryptomycota (van Hannen et al., 1999; Wurzbacher et al., 2011).

Figure 1 Distributions of ocean net primary productivity in the period of 1997 to 2002 (Dunbar, 2008).

Oceans cover around 70% of the Earth’s surface, and they are home to a spectacular biodiversity. Marine fungi play a not yet thoroughly understood, but nonetheless important role in shaping its aquatic systems. The coastal areas of the ocean are highly productive (Figure 1) and also very important to human activities such as fishery. Coastal waters are rich in phytoplankton, which contribute for around half of the coastal net primary productivity and is also the major primary producer in pelagic waters. Net primary productivity is the rate of chemical energy fixed by photosynthetic organisms and available to consumers of the trophic chain. A large portion of the biomass thus produced ends up in the microbial decompositional pathway, ready for degradation by fungal saprotrophs into simpler biomolecules that are subsequently absorbed (Raghukumar, 2017). There is also evidence that the zoospores of chytrids are grazed by marine zooplankton, which are primary herbivores and thus have an impact on higher level consumers in the marine trophic chain (Lepelletier et al., 2014). Other fungi are parasitic or pathogenic, and thereby play crucial roles in organic matter build up and decomposition, mineralization, and nutrient cycling in marine environments. Fungi are, in other words, potentially able to structure and control the functioning of aquatic food webs (Grossart et al., 2017). Recent studies also hint at the

!1

potential of marine fungi as a source of new drugs, since they produce completely novel metabolites and structures (Gao et al., 2010; Kjer et al., 2010; Meyer et al., 2010).

Marine fungi can inhabit many different environments within the ocean. These range from coastal areas to offshore pelagic, benthic, and also deep-sea habitats as well as a substantial variety of substrates such as mangroves, salt marsh plants, seaweeds, diseased animals, and drift material that has been washed into the ocean from land (Jones, 2011). They exist as free-living organisms or as attached to other organisms mostly as filaments or unicellular yeasts. They can also be found in the marine sediments. There are two different types of marine fungi: obligate and facultative marine. Members of the former group live and reproduce exclusively in marine environments, whereas members of the latter are terrestrial taxa that are able to sustain the marine environments for shorter or longer periods of time (Raghukumar, 2017). The number of described obligate marine species is below 1 000, but Jones (2011) speculates that the number of marine fungi could be as high as 10 000 species. Aquatic fungal taxa are found scattered across the fungal tree of life (Figure 3) with all major phyla represented (Picard et al., 2017; Tedersoo et al., 2017), although members of Dikarya are the lineages most readily recoverable from marine environments (Richards et al., 2012) – at least so far. While overall diversity is lower than in terrestrial fungi, marine fungi can be quite adaptable (Richards et al., 2015). They have been documented in deep ocean sediments down to 1 740 meters below the seafloor. In these deep sediments diversity was low and primarily composed of basidiomycete yeasts (Redou et al., 2014). In deep sea sediments and benthic habitats, generally only members of Ascomycota and Basidiomycota have been recovered so far (Ciobanu et al., 2014; Redou et al., 2015), even though there is evidence that early diverging are relatively more common in marine environments than previously thought (Picard et al., 2007; Tedersoo et al., 2017). Marine members of early diverging fungal lineages (Figure 3) are, as far as known, primarily parasites or saprobes, and make up an important and to this day little studied part of the diversity of fungi.

Studying marine fungi through traditional lab and cultivation-based methods is a very slow and painstaking process, as many fungi are difficult to raise in culture. This is particularly true for fungi that associate with other organisms, such as endosymbionts and endoparasites. The marine environment adds another layer of complexity to cultivation efforts through aspects such as salt levels, pH, temperature and light regimes. Instead, environmental DNA of microscopic fungal communities is increasingly assessed by metabarcoding, a biodiversity assessment method that combines DNA-based identification and high-throughput DNA sequencing (HTS). Metabarcoding can be used to characterise entire fungal communities using a relatively short DNA barcode, such as the internal transcribed spacer region (ITS) of the rDNA gene cluster. Barcode amplification relies on taxon specific primers that ideally target all fungi. Ill-chosen primers will lead to biases in terms of what taxa are recovered, masking environmental signals and distorting the view of the fungal communities obtained (Tedersoo et al., 2015). What species and what taxonomic levels are recovered is also dependent on the completeness of the reference databases for fungi, the most complete and up-to-date one being UNITE (ITS2; Abarenkov et al., 2010). DNA sequencing of environmental samples regularly reveals fungal lineages yet unknown to science, hinting at the largely untapped taxonomic and ecological discovery potential of such sequencing efforts (Nilsson et al., 2016). A lot of fungal genetic data have been collected over the past few years, but the number of species described as new to science has hovered around an annual ~1 500 for many years (Hibbett et al., 2016). This is partly because new species cannot be described solely from molecular evidence (DNA sequence data) – the current edition of the International Code of Nomenclature for algae, fungi and plants (McNeill et al., 2012) bans any formal description of fungal species from sequence data only (in the absence of a physical type specimen). There are countless fungi that never – or only rarely – produce fruiting bodies or other somatic structures that could be used for a

morphological description and serve as type specimen, but that nevertheless are as real as fungi that do produce them. Thus, exploring habitats and substrates for fungal diversity using DNA-based methods

!2

Figure 2 Partial schematic drawing of the fungal nuclear rRNA cluster. It shows the non-coding, very variable internal transcribed spacer regions (ITS) ITS1 and ITS2 and the highly conserved genes 5.8S, and 28S. Amplification with primers, portrayed as arrows above and below the graphical representation of the cluster results in amplicons of around 300 to 550 bases.

such as metabarcoding is gradually becoming a hot research topic. The genes and spacers of the nuclear ribosomal operon have become particularly popular for the study of fungi during the last 20 years (Begerow et al., 2010; Schoch et al., 2012). The genetic marker used in this metabarcoding study is the highly variable internal transcribed spacer region 2 (ITS2). It is located between the more conserved nuclear rRNA genes 5.8S and 28S (Figure 2). In contrast to ITS2, which is degraded during rRNA maturation, 5.8S and 28S encode for the large ribosomal subunit (LSU) of the ribosomes. Specifically designed primers (Table 1) bind to the flanking sites of ITS2, namely the 5' end of 5.8S and the 3' end of 28S. Consequently, the amplified fragments consist of a large portion of 5.8S (~150 bases), the adjacent

!3

Figure 3 Current status of our understanding of the phylogeny of fungi, compiled, and negotiated from the most up-to-date, and robust phylogenies of Ascomycota (Hibbett et al., 2007; Schoch et al., 2009; Rosling et al., 2011; Gazis et al., 2012; Zhuang & Liu, 2012; Prieto et al., 2013), Basidiomycota (Hibbett et al., 2007; Nguyen, 2015), and early diverging lineages (James et al., 2006a; James et al., 2006b; Hibbett et al., 2007; Spatafora et al., 2016) from different phylogenetic studies. Lineages with at least zoosporic form are indicated by a schematic drawing of a zoospore and groups containing members of obligate marine fungi are indicated by an asterisk (Jones & Pang, 2012).

ITS2, and a small part of 28S rDNA. The overall fragment size normally lies between 300 and 550 bases. This size variation is primarily attributable to ITS2 itself, as its length can vary within fungal groups from around 120 bases in archaeorhizomycetes (Taphrinomycotina – Ascomycota) to around 280 or more bases in some chytrids (Chytridiomycota), mortierellomycetes (Mucormycota), and other fungal lineages (Tedersoo et al., 2015).

In this project, field sampling of two littoral soil/sediment gradients was undertaken to examine whether changes in fungal community composition and function can be detected along the gradient from land to ocean. This study aims to contribute data on the taxonomic diversity and ecological roles of fungal communities in coastal habitats and in marine sediments, and how these are influenced by various chemical and edaphic parameters.

Aims In this project, we explore the following questions: • Are there any fungi present in the soil samples of the two gradients, especially in the intertidal

and benthic sediments? • If we find fungi, will we be able to identify them, and to which taxonomic level? What

proportion of operational taxonomic units (OTUs) is unassignable? • We should, through our metabarcoding approach, be able to observe a change in fungal

communities over the littoral gradients from terrestrial habitat to benthic sediment. How will the fungal community composition change along the gradient?

• Can we characterise the ecological roles and/or functional capacity of the fungal lineages we can assign taxonomically, to cast additional light on the ecology and nutrient cycling role of fungi in littoral sediments?

• Can we find correlations between the taxonomic composition of the fungal communities and measured environmental parameters such as pH and conductivity?

Materials and methods Sampling and experimental design

We essentially followed the soil sampling protocol of Tedersoo et al. (2014) with some slight adaptions to assess a spatial gradient from land through an intertidal shoreline to water. Additionally, pH, and salinity were measured at all sampling points. For comparison and to reduce stochastic noise, sediments were sampled from two different places of similar ecology and habitat. Both localities are situated in the coastal region of Gothenburg, Sweden, and were sampled at the beginning of September 2017. The first sampling was carried out at Askimsbadet (AB) (Figure 4), Gothenburg Municipality, Västra Götalands county (57° 37’ 16.973”N × 11° 55’ 49.985” E). This locality is characterized by a

!4

Figure 4 Sampling schema and sampling sites at Askimsbadet (AB) (left) and Stora Amundön (SA) (right), showing sampling points and the length of terrestrial gradient in orange, and marine gradient in blue (Google Maps).

shallow saltwater bay with a short intertidal shoreline and a sandy beach becoming meadow-like with high grass and reeds. The gradient is around 120 m long. The second sampling spot was located on the island Stora Amundön (SA) (Figure 4), Gothenburg Municipality, Västra Götalands county (57° 35’ 29.25”N × 11° 54’ 15.897” E), and was comparatively smaller than the first one. It consisted of a shallow sandy cove turning into a grassy meadow landscape. The grassland was surrounded by rocks and shrubs and was replaced by mixed woodland after a stretch of around 60 m. At each of the two localities, soil was collected along a transect. The meadow stretch was so long that around 7 sampling sites each 20 m apart and approximately at sea-level were accessible. Also, two more sites were sampled at the seabed, approximately 20 m apart at 1 to 2 m below the ocean surface were designated along the gradient. The GPS coordinates and altitude of the sampling area were recorded, and the ectomycorrhizal plant cover at each site was noted. Sampling points are indicated by parenthesis in the following plant cover description. The vegetation at AB (Figure 4) consisted of mostly dried seaweed at the shoreline, but was going over into a natural reed, and grass landscape after around 3 m, bearing a few grazing plants such as Lysimachia maritima, and Argentina anserine (7). These plants, along with Rumex and Anthriscus sylvestris could also be found further up in the meadow landscape (6, 5). They were later replaced by Taraxacum, Cirsium (4, 1), and Urtica (3) although Anthriscus sylvestris and Rumex could be found all along the gradient. Furthest up, genera and species such as Equisetum, Taxus, Thuja, Malus, Rubus, and Calystegia sepium could be spotted (1). The vegetation at SA consisted of dried seaweed at the shoreline (7) that was replaced by a mown grass landscape (6, 5, 4) where Erica, Taraxacum, Juniperus (5), and additionally young Quercus (4) could be found. Further up, the landscape transformed into a mixture of deciduous, and coniferous forest that consisted of Quercus, Juniperus, Populus, Betula, Pinus (1, 2, 3), and some fern taxa such as Pteridium (3) and Polypodium vulgare (2).

At each of the 9 sites, 2 samples were taken and diluted 1:1 with water to measure the pH and conductivity of the soil. The average of these measurements was used as an estimate of the true values. At each sampling point (1 – 9), 10 experimental replicates were selected according to a predetermined sampling scheme (Figure 4) all within a radius of 20 m, where 9 spots were in a row of three, and the 10th (in red) was taken randomly. At each of the 10 replicate spots, soil was sampled down to about 5 cm with a metal bulb planter with a diameter of 5 cm in diameter. Between each round of sampling, the bulb planter was cleaned to avoid carryover of DNA between sampling rounds. Where applicable, loose litter was removed from the surface of points to be sampled. The sample itself thus consisted of the organic and upper mineral layer when sampled from land and sandy to clay mineral layer from the ocean floor sediment, respectively. From the soil cores obtained, a subset of soil from all sides of the core was picked and placed in a clean plastic tray (Tedersoo et al., 2014). The remainder of the soil core was discarded. Within 4 hours of collection, samples were rid of coarse roots and stones, and air-dried in a drying room at room temperature. Fully dried samples were placed in Ziploc bags with silica gel and placed in a dark, dry room for storage and subsequent analysis. We sent subsamples of our soil samples to the department of forest ecology and management at SLU (Umeå, Sweden) for analysis of soil isotopes and mass fraction of carbon and nitrogen with an isotope ratio mass spectrometer (DeltaV,Thermo Fisher Scientific, Bremen, Germany).

ITS2 amplicon library preparation and PCR To prepare the samples for DNA extraction, the Ziploc bags containing dried soil were rubbed

between both hands vigorously for 3 minutes until the soil had transformed into fine dust following the protocol of Tedersoo et al. (2014). From this fine dust, 300 g were subsampled for later DNA extraction. Sandy samples and samples with minute stones could be separated from the desired soil dust by shaking the sample rhythmically, which made the coarse particles rise to the surface for manual removal. For total DNA isolation we used the DNeasy PowerSoil Kit (Qiagen, Hilden, Germany), following the manufacturer’s instructions. The kit is intended to be used for environmental samples with high humic acid contend and other difficult soil types such as sediment. The DNA concentrations after extraction of all samples were measured with Qubit 3.0 Fluorometer (Thermo Fisher Scientific, Waltham, Massachusetts, USA), ranging from a few ng /µl up to ~ 180 ng /µl. We measured DNA purity with NanoDrop (Thermo Fisher Scientific, Waltham, Massachusetts, USA) prior to PCR, with both ocean samples of SA and AB and the forest samples of SA being slightly out of range. To measure the sample concentration with Qubit 3.0, we worked with the iQuant™ Quantitation Kit and samples were prepared following the manufacturer’s recommendations. We used a volume of 1 µl of DNA sample, and measured concentrations in both the broad (2 - 1 000 ng dsDNA) and the high sensitivity (0,2 - 100 ng dsDNA) ranges. The NanoDrop blank was measured with the template DNA eluate from the DNeasy PowerSoil Kit, which confirmed that it yielded a reproducible zero. After each sample was loaded and measured, the instrument was cleaned.

!5

Multiplexed amplicon libraries were constructed according to the two-step PCR protocol described in Bourlat et al. (2016). This method consists of dual PCR amplification, the first PCR using amplicon specific primers including an Illumina adapter overhang (amplicon PCR) and a second cycle limited PCR for the incorporation of Illumina index adapters for multiplexing (index PCR) (Bourlat et al., 2016).

The amplicon primers used in this study to amplify the ITS2 barcode region were designed and previously used for the detection of both terrestrial and aquatic fungi (Tedersoo et al., 2015; Wurzbacher et al., 2017). They contain a Nextera-Illumina-Adapter overhang, a mismatch spacer, and the marker specific sequence (Table 1). Small letters within the primer sequence represent modified nucleotides, so called PTOs (Phosphothioate oligonucleotides). The PTO’s orthophosphate non-bridging oxygen within the DNA backbone is substituted with a sulphur atom, one of the most widely used nuclease resistant oligonucleotides for antisense applications. PTOs prevent mismatch corrections by the proofreading polymerase. The forward primer is degenerated and an equimolar mix of the reverse primers is used to improve taxonomic coverage within fungi. Table 1 Structure of the forward primer ITS3-Mix2 (Tedersoo et al., 2015) and reverse primers ITS4-cwmix1, and ITS4-cwmix2 (Wurzbacher et al., 2017) for amplification of the fungal ITS2 region. The adapter, spacer, and marker-specific sequence are separated by a hyphen. DNA ambiguity symbols follow Cornish-Bowden, 1985.

Amplicon PCR was conducted with the KAPA HiFi HotStart ReadyMix (Roche, Basel, Switzerland), containing an engineered B-family DNA polymerase for fast and versatile high-fidelity PCR that reduces the number of nucleotide incorporation errors produced during PCR amplification (Lindahl et al., 2013). Primers were diluted to a final concentration of 20 pmol/µl, and the reverse primers (ITS4-cwmix1 & ITS4-cwmix2) were mixed at equimolar concentration. DNA template concentration of each sample was adjusted to between 20 and 50 ng/µl to standardize samples. The PCR cocktail of 25 µl reaction volume comprised 12,5 µl KAPA HiFi HotStart ReadyMix (Roche, Basel, Switzerland), 1 µl of forward (TCGTCGGCAGCGTCAGATGTGTATAAGAGACAG-AAC-caWCGATGAAGAACGCAg), and 1 µl o f r e v e r s e p r i m e r ( G T C T C G T G G G C T C G G A G AT G T G TATA A G A G A C A G - A A -TCCTCCGCTTAyTgATAtGc & GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAG-AA-TCCTCCGCTTAtTrATAtGc) at 20 pmol/µl, 9,5 µl of Nuclease-Free Water (Qiagen, Hilden, Germany), and 1 µl of adjusted template DNA (20 – 50 ng/µl). We additionally used a negative control. Three replicate reactions were carried out for each sample with the following program on a MyCycler™ thermal cycler (Bio-Rad, Hercules, California, USA): Initial denaturation 3 min at 95 °C followed by 30 cycles of 30 sec at 95 °C, 30 sec at 57 °C, 1,5 min at 70 °C, and a final elongation cycle for 5 min at 72 °C. We kept the number of PCR cycles as low as possible in order to reduce PCR incorporation error and to avoid reaching a limiting stage due to decreased accessibility of template DNA or PCR mixture components (Opel et al., 2010). To check for amplification products, as well as their size and concentration, we ran 3 µl of the first replicate PCR product on a 2% agarose gel using 1xTAE Buffer, and GelRed® (Biotium, California, USA) for DNA staining. We ran the gel for 30 min and 45 min at 100V for basic and higher resolution respectively. We also verified the individual fragment size with TapeStation (Agilent Technologies, Santa Clara, California, USA), using the genomic DNA ScreenTape assay. The desired fragments had assize of 300 – 550 bases. As no other, undesired bands showed up on any of the gels or in the TapeStation run, the remaining two PCR runs were conducted, and the products of the three PCR reactions were pooled. To quantify the amplicon library, the pooled products were measured with Qubit 3.0 Fluorometer (Thermo Fisher Scientific, Waltham, Massachusetts, USA) Pooled products were then sent to Macrogen (Seoul, Korea) for further Index PCR as well as Illumina MiSeq sequencing (San Diego, California, USA), providing paired-end reads of 2 x 300 bases in a single run. Around 1,3 Gb of raw data were returned to us from Macrogen. Bioinformatics

To analyse community composition and assign taxonomy to the amplicons, we used the software pipeline micca - MICrobial Community Analysis (Albanese et al., 2015) v. 1.6 that utilizes the third-party software tools Cutadapt (Martin, 2011) for primer trimming and VSEARCH (Rognes et al., 2016) v. 2.7.1 for merging, filtering, OTU picking, and taxonomic classification of sequences. We analysed the

!6

gradient samples from SA and AB separately to maintain integrity in the later statistical analysis. For each gradient, we assembled the paired-end reads of the adapter trimmed data from Macrogen, requiring a minimal overlap of 50 bases, and strict assessment with zero mismatches. Paired reads that did not contain both forward and reverse primers were discarded and primer sequences were then removed from the merged reads. We filtered all remaining sequences by discarding all that were shorter than 260 bases, and had an expected error rate of more than 0,5 %. OTU clustering was carried out using a complete-linkage clustering algorithm for denovo greedy clustering. This type of algorithm tends to be sensitive to read length and frequencies of sequences (Lindahl et al., 2013). Sequence reads were assigned to operational taxonomic units (OTUs; Blaxter et al., 2005) using the denovo greedy clustering approach of VSEARCH in micca, similar to UPARSE (Edgar, 2013) (Figure 5) using a sequence similarity cut-off of 97 %. Chimeric OTUs, which are formed when two or more biological sequences are joined together and amplified during PCR, were removed using the default approach in micca as they represent artificial species. Also, singleton (one-sequence) OTUs were excluded from further analyses, as they most probably are the product of PCR errors. They very likely do not represent actual species and instead lead to inflated diversity estimates in the downstream statistical analysis. Taxonomic assignment was done using the VSEARCH-based consensus classifier of micca utilizing the latest UNITE database reference release (version 7.2, doi 10.15156/BIO/587481; Abarenkov et al., 2010b) and a micca sequence identity threshold of 0,9.

We made an effort to scrutinize the OTU tables for large OTUs of non-fungal origin, as we wanted our statistics to be based primarily on the fungal component of the microbiomes of these sites. OTU tables of both gradients were therefore checked for non-fungal and ambiguous taxonomic assignments. For this, we used a tailored BLAST parser Perl script that returns those OTUs in the table that have primarily non-fungal BLAST matches. These results, as well as the 50 most abundant OTUs of each gradient were manually checked using BLAST in GenBank (https://blast.ncbi.nlm.nih.gov/Blast.cgi), following the guidelines of Nilsson et al. (2012), which apply to full-length ITS sequences. In order not to be excluded from subsequent analysis, our ITS2 reads therefore had to produce matches that comprised more or less the full length of the query sequence. Also, they should feature most of the full 5.8S gene, which should be fairly similar over all ITS sequences. If matches were ambiguous, the next thing to do was to be sure that the sequence indeed came from a fungus. This can typically be ascertained beyond reasonable doubt, if the taxonomic annotations of the majority of matches (ideally stemming from several independent studies) suggest, or at least do not contradict a membership in the fungal kingdom. To arrive at a more detailed conclusion on the taxonomy, the query sequence should group with at least a few other sequences of comparable taxonomic ranks, have a high query cover, and an E value close to zero. OTUs that could not be assigned to the kingdom of fungi for certain were

excluded from the OTU- as well as the taxonomy table for further analysis. From the original 4 355 OTUs in the OTU table of AB, and 3 510 OTUs in SA, each sampling locality saw the exclusion of 40 OTUs. Most of them were either animals or plants, and a few other were deemed too ambiguous to be included. Out of the modified OTU and taxonomy tables we finally computed relative abundance tables of sequences at different taxonomic levels with micca. To account for differences in sequencing depth, rarefication of the

OTU tables was done with micca. These rarefied OTU tables were set up by subsampling without replacement to the number of sequences recovered by the sample with the fewest number of reads. Statistical Analysis

For statistical analysis of the gradients, we used the “vegan” (Oksanen et al., 2018) v. 2.4-6 and “DESeq2” (Love et al., 2014) v. 1.18.1 packages in R (R Core Team, 2014) v. 3.4.4. As before we kept the data analysis of both gradients separate, to avoid losing differences between them. Rarefied and non-rarefied OTU tables from micca were loaded into R, as were taxonomy data and metadata. We visualized the OTUs found per site to get a first impression of the distribution of the number of species.

We then assessed species diversity and computed the rarefication curves of the gradients, which showed us to what extent we managed to sample the entire communities in the gradients. We computed each site’s richness S from the non-rarefied OTU tables,

!7

Figure 5 Greedy clustering schema (modified from https://www.drive5.com/usearch/manual/uparseotu_algo.html)

which is also known as the actual number of species. We also calculated the Shannon index H’ (1) from the rarefied OTU tables. The Shannon index gives an estimate of the species composition, which is the effective species number, at a given site (alpha diversity) (Tuomisto, 2010; Jurasinski & Koch, 2011). We further used the Shannon index to calculate the gradients Pielou's evenness J (2), which gives an estimate of the even distribution of species. Chao1 estimates were calculated from the rarefied OTU tables. These estimates take undiscovered rare species into account and are thus also described as the estimate of the true species richness (Chao, 1984). We computed the distances between sites using the Jaccard dissimilarity index (3). It is a measure of the actual number of distinct compositional units (beta richness), and we calculated it from our presence/absence transformed rarefied OTU tables (Tuomisto, 2010).

Alpha diversity was computed from the Shannon index H’, where ! is the proportion of species i, S is the species richness, and ! is the beta-richness (Hill, 1973):

! (1)

Pielou's evenness J:

! (2)

Beta richness was calculated with the Jaccard dissimilarity index ! :

! (3)

For ordination of the data, we computed a principle component analysis (PCA) of the rarefied OTU table using log(x+1) transformed counts to account for zeros in the untransformed data and to improve normality. PCA reduces multivariate data stored in a matrix by identifying the dimensions that explain the largest part of the variability. We assigned the similar sites into ecological groups (e.g., “SA forest”) that we later used for the determination of significance between them and the metadata. We also used these groups later on to allocate them into their ecological guilds.

Additionally, we computed a regression of the normalized OTU table with DESeq2 and fitted it to a negative binomial general linear model (GLM) using a Wald – test (Wald, 1945; Love et al., 2014). Here, we removed OTUs with clusters that had ≤ 3 sequences prior to calculation with the intention of a strict assessment of the data to be able to offer robust interpretation of the results. The regression was done to test if there was any significant correlation between the OTUs and the relative position within the gradient. When listing the results, we removed the post filtering-step and the resetting function of p-values that is normally done by default by DESeq. The OTUs were deemed significant based on a FDR (false discovery rate) cut-off of the adjusted p-values of < 0,05. We also assessed if some of the most sequence-rich phyla changed significantly over the distance of the gradient. This was tested by creating a 2 x 2 matrix of the number of significantly and non-significantly changing members of the respective phylum and the significantly and non-significantly changing members outside the respective phylum by performing a Fisher's exact test of independence. The null hypothesis of each matrix could be rejected if the test assigned a p-value smaller than 5 %. We tested this for Ascomycota, Basidiomycota, Mortierellomycota, and Chytridiomycota. We also tested if there were any OTUs whose abundance went significantly up or down in correlation to our metadata. Additionally, we used FUNGuild (Nguyen et al., 2016) to assign the fungal OTUs of each ecological group into functional categories, or guilds.

Results Physicochemical soil properties

Gradients from both localities (SA and AB) show overall similar physicochemical soil properties (Table 2). Exceptions are the total carbon and nitrogen contents of the forest sites of the SA gradient (SA1, SA2 and SA3), which are higher than the others, and the ! N ratios, which are close to zero. The shoreline site SA7 stands out as the total carbon and nitrogen contents are lower than in all the other sites. Moreover, the ! C ratio of the SA ocean site 9 stands out as lower than what one would expect when comparing with the ocean sites of AB. The pH and conductivity both increase towards the ocean sediment. The FC content increases slightly towards the ocean in both gradients, while FN stays the same at all sites. Conductivity is directly related to the concentration of salt ions in the soil and was

piSβ

H ′ �=   −S

∑i=1

pilogb pi

J = H′�/ log(S ) CJ

CJ = 2/Sβ − 1 

δ15

δ13

!8

reaching a few mS in the seabed sediments. The forest sediments in SA are more acidic than meadow and ocean sites. The benthic sites are both slightly alkaline. Table 2 Metadata summary of both gradients Stora Amundön and Askimsbadet. Here, C is the mass fraction of carbon, N is the mass fraction of nitrogen, ! C is the isotopic ratio of the to isotopes, ! N is the isotopic ratio of the to isotopes, FC is the isotopic amount of the fraction , FN is the isotopic amount of the fraction , and C/N is the ratio of Carbon to Nitrogen. Measurement units are written in parenthesis.

Processed amplicon data Table 3 shows the number of sequences between each amplicon processing step with micca.

Between merging paired end sequences and primer trimming, not many sequences were lost. More loss can be observed after quality filtering our sequences and discarding reads with less than 260 bases, which left us with around 97 % of the original primer-trimmed sequences or a loss of around 13 000 sequences in SA and 8 900 sequences in AB. Only a minor fraction of sequences were chimeras. Raw sequence statistics can be found in the supplementary Table 10. Table 3 Number of reads after merging the reads, trimming them from their primer sequences and quality filtering of reads that were shorter than 260 bases and had an expected error rate of more than 0,5%, as well as the number of discarded chimeric sequences.

δ13 δ15

C (%) FC (%)

C/N N (%)

FN (%)

CONDUCTIVITY (ΜS/M)

PH DISTANCE (M)

AB1 4,75 -28,98 1,08 13,60 0,35 5,04 0,37 121,5 5,97 0

AB2 7,87 -28,96 1,08 15,70 0,50 5,01 0,37 130,0 6,13 20

AB3 4,39 -28,70 1,08 14,60 0,30 4,99 0,37 144,5 5,30 40

AB4 7,44 -28,58 1,08 14,60 0,51 5,95 0,37 170,5 5,80 60

AB5 7,32 -27,61 1,08 12,80 0,57 6,47 0,37 207,0 6,67 80

AB6 6,23 -19,07 1,09 15,60 0,40 6,80 0,37 276,5 7,40 100

AB7 0,79 -16,47 1,09 15,20 0,05 7,50 0,37 3830,0 6,79 120

AB8 0,90 -14,73 1,10 13,90 0,07 4,85 0,37 2810,0 7,92 140

AB9 0,72 -14,70 1,10 12,40 0,05 4,92 0,37 2315,0 7,94 160

SA1 44,68 -28,50 1,08 21,90 2,04 -0,06 0,37 220,0 4,32 0

SA2 32,59 -29,48 1,08 23,40 1,39 -0,01 0,37 257,5 4,02 20

SA3 35,69 -28,57 1,08 23,20 1,54 0,25 0,37 145,5 3,98 40

SA4 5,17 -27,37 1,08 14,40 0,36 3,77 0,37 133,5 5,28 60

SA5 8,19 -27,40 1,08 14,60 0,56 3,96 0,37 150,0 5,51 80

SA6 6,07 -28,87 1,08 14,10 0,43 3,98 0,37 271,5 6,68 100

SA7 0,08 -21,08 1,09 8,00 0,01 5,45 0,37 398,0 7,14 120

SA8 0,28 -12,34 1,10 14,00 0,02 4,82 0,37 5755,0 8,43 140

SA9 0,87 -5,75 1,10 29,00 0,03 4,65 0,37 5015,0 8,42 160

! N (‰)δ15! C

(‰)δ13

SA AB

READS AFTER MERGING 687600 789758

READS AFTER PRIMER TRIMMING 686338 788012

READS AFTER FILTERING 673711 779899

CHIMERIC SEQUENCES DISCARDED 546 729

!9

Relative abundance tables The relative abundances of sequences are visualized in Figure 6 to Figure 11, each figure showing

a different taxonomic level. One striking aspect of the relative abundance of sequences in the following figures is the number of unclassified sequences of ≥ 80% at the benthic sites SA8-SA9 and AB8-AB9, regardless of the taxonomic level. In the forest sites (Figure 17) we only find around that a modest 5 % of our sequences (Figure 6) were not identifiable at any taxonomic level. In the meadow/scattered trees sites (Figure 17) we found that around 8 % (AB2, AB3, AB4) to 16 % (AB1) of our sequences (Figure 9) could not be identified at any taxonomic level. In the meadow sites (Figure 18) we found that between 8 % to 20 % (SA4, SA5, SA6), and around 60 % (SA7) of our sequences (Figure 6) were not identifiable at any taxonomic level. In the meadow/reed sites (Figure 18) we found that around 20 % (AB5, AB6), and around 8 % (AB7) of our sequences (Figure 9) could not be identified at any taxonomic level.

! Figure 6 Relative abundance of reads at phylum level (SA)

! Figure 7 Relative abundance of reads at order level (SA)

!10

! Figure 8 Relative abundance of reads at genus level (SA)

! Figure 9 Relative abundance of reads at phylum level (AB)

! Figure 10 Relative abundance of reads at order level (AB)

!11

! Figure 11 Relative abundance of reads at genus level (AB)

Distribution of OTUs and their taxonomic composition

!12

2 4 6 8

4.4

4.6

4.8

5.0

Sites

log(

Otu

) per

site

2 4 6 8

4.4

4.6

4.8

5.0

Sites

log(

Otu

) per

site

Figure 13 visualizes the taxonomical composition of the transects sampled at SA and AB by compartmentalizing them into areas showing the proportions of their most abundant OTUs. Out of 3 470 fungal OTUs, 1 872 OTUs could not be classified to the phylum level in the SA gradient (~53 %). The remaining 1 598 OTUs comprised mainly Dikarya (composed of 21,8 % Ascomycota and 12 % Basidiomycota) followed by species of Glomeromycota (3,1%), whereas Mortierellomycota, Chytridiomycota, and Rozellomycota were about equally abundant (~ 1,5 %). Within the AB gradient 2 371 OTUs out of 4 315 fungal OTUs (~55 %), could not be classified to the phylum level. AB gives a similar picture of the remaining 1 998 OTUs dominated by Dikarya (20,6 % representing Ascomycota and 11,2 % for Basidiomycota), followed by Glomeromycota (3 %), Chytridiomycota (2,1 %), whereas Mortierellomycota, and Rozellomycota were fairly equally abundant (~ 1,5 %). Rarefication curves

T h e

rarefication curves in Figure 14 represent the means of repeated re-sampling of samples within SA and AB. Levelling off of the slopes indicates a sufficiently large sampling effort. The lowest sequencing depth of both gradients is around 20 000 reads, and is attributable to the ocean sediment samples (SA8, SA9, AB8 and AB9). Diversity indices, number of OTUs and reads

The number of OTUs and reads, as well as diversity indices per site, and gradient are summarized in Table 4 (SA) and Table 5 (AB). The lowest number of reads that each of the tables were rarefied to can be taken from SA8 (18 996 reads), and AB9 (22 101 reads) respectively. The number of OTUs recovered is on average lower for the SA gradient (744 ± 400) than the AB gradient (1 232 ± 500). The estimate of richness (570 ± 275) differs from the true species richness chao1 (722 ± 388) by around 150 OTUs in SA. In AB, richness (945 ± 326) differs from Chao1 (1214 ± 468) by around 270 OTUs. Alpha diversity increases slightly towards the benthic sediments in SA, but it is similar overall. For AB it is somewhat constant, except for the intertidal site AB7, which is strikingly lower than the average of the gradient (5±1). Table 4 Number of sequences, OTUs, and diversity indices of Stora Amundön.

SA1 SA2 SA3 SA4 SA5 SA6 SA7 SA8 SA9

READS 82577 66043 104082 79767 81595 77016 52967 18996 19139

OTUS 447 746 865 693 688 1371 883 426 574

RICHNESS 373,94 516,77 568,23 491,13 472,64 1070,33 637,80 426,00 573,55

EVENNESS 0,59 0,62 0,55 0,65 0,63 0,76 0,73 0,74 0,76

CHAO1 424,93 694,98 770,26 699,77 622,27 1298,00 931,90 451,24 602,02

SHANNON 3,49 3,83 3,50 4,03 3,91 5,27 4,73 4,51 4,81

!13

0 20000 40000 60000 80000 100000 120000

050

010

0015

00

Sample Size

OTU

s

AB1

AB2

AB3AB4

AB5

AB6

AB7AB8

AB9

0e+00 2e+04 4e+04 6e+04 8e+04 1e+05

020

040

060

080

010

0012

0014

00

Sample Size

OTU

s

SA1

SA2

SA3

SA4SA5

SA6

SA7

SA8

SA9

Figure 14 Rarefication curves of SA (left) and AB (right).

!14

Table 5 Number of sequences, OTUs, and diversity indices of Askimsbadet.

Network plots of site similarities Figure 15 Network graph of beta richness (left: SA, right: AB).

Thick l i n e s i n Figure 1 5

indicate stronger similarities between sites than thinner lines. Beta richness was computed on the basis of the Jaccard index of dissimilarity that is estimated based on the presence/absence transformed, rarefied OTU tables. The Jaccard index ! is a popular measurement for beta diversity and proportional species turnover (Figure 15). It indicates the dissimilarities among all sites, where the actual species turnover is 1 - ! .

AB1 AB2 AB3 AB4 AB5 AB6 AB7 AB8 AB9

READS 77433 70132 63519 93882 65999 64737 124351 23485 22101

OTUS 1561 1310 1411 1458 1742 1425 771 752 655

RICHNESS 1153,9 963,35 1069,02 988,68 1339,56 1087,3 496,72 747,51 655

EVENNESS 0,71 0,7 0,68 0,67 0,76 0,72 0,55 0,76 0,76

CHAO1 1483,51 1279,37 1548,71 1264,35 1752,95 1469,3 674,34 782,86 670,45

SHANNON 5,00 4,82 4,78 4,64 5,50 5,05 3,43 5,05 4,92

CJ 

CJ

!15

Component 1 (18.7% of the total variance)

Com

pone

nt 2

(17.

8% o

f the

tota

l var

ianc

e)

SA1

SA2

SA3

SA4SA5

SA6

SA7

SA8SA9

−1.0 −0.5 0.0 0.5 1.0

−1.0

−0.5

0.0

0.5

1.0

Component 1 (23.4% of the total variance)

Com

pone

nt 2

(16.

4% o

f the

tota

l var

ianc

e)

AB1

AB2

AB3 AB4

AB5

AB6

AB7

AB8AB9

−1.0 −0.5 0.0 0.5 1.0

−1.0

−0.5

0.0

0.5

1.0

Figure 16 Principal component analysis (PCA) of the log(1+x) transformed OTUs along the gradient (left: SA, right: AB). The first two dimensions explain the most variance between samples, and their relative position is indicative of the similarities between samples.

1

2

3

4

5

6

7

8

9

●

●

●

●

●

●

●

●

●

SA1SA2SA3SA4SA5SA6SA7SA8SA9

SA1SA2SA3SA4SA5SA6SA7SA8SA9

1

2

3

4

5

6

7

8

9

●

●

●

●

●

●

●

●

●

AB1AB2AB3AB4AB5AB6AB7AB8AB9

AB1AB2AB3AB4AB5AB6AB7AB8AB9

Data ordination According to Figure 16, we could group the sites in groups of their principle components. This

resulted in assigning the sites into groups representative of different environments. For SA these were: forest group (SA1, SA2, SA3), meadow group (SA4, SA5, SA6, SA7), and ocean (SA8, SA9). For AB they were: meadow/scattered tree group (AB1, AB2, AB3, AB4), meadow/reed group (AB5, AB6, AB7), and ocean group (AB8, AB9), although AB1 could not be so clearly assigned. Linear regression

Here, the significance of coefficients was tested with a Wald – test based on a differential expression analysis of a generalized linear model (GLM). In our case, the model describes random variation by a negative binomial distribution, where occurrences have two possible outcomes that are dependent on each other. False discovery rate (FDR) adjusted p-values (q-values) were deemed significant or not significant based on a cut-off of < 0,05. The only phylum that changed significantly along the distance of the gradient was Mortierellomycota within the Askimsbadet gradient (Table 6). Table 6 Changing phyla over the gradient with p-values from fisher exact test (significant if p-value < 0,05), AB.

Table 7 Changing phyla over the gradient with p-values from fisher exact test (significant if p-value < 0,05), SA.

ASCOMYCOTA BASIDIOMYCOTA MORTIERELLOMYCOTA CHYTRIDIOMYCOTA

P-VALUE 1,0 0,730 0,002 0,138

ASCOMYCOTA BASIDIOMYCOTA MORTIERELLOMYCOTA CHYTRIDIOMYCOTA

P-VALUE 0,069 0,478 0,085 1,0

!16

Table 8 Testing for significant correlation of OTUs and metadata, and inferring how many OTUs significantly went up or down in abundance along the gradient (nS= not significant).

In Table 8 no clear pattern is observable that could indicate a correlation with the occurrence of total OTUs at each transect SA and AB, although the metadata seem to have a significant effect on the increase or decrease of species. Exceptions are the amounts of the isotopic fractions FC and FN, as well as the C/N ratio in AB, which are not significantly correlating to OTU abundance.

SA AB

going down going up going down going up

PH 71 311 226 105

31 93 207 29

N 158 31 105 147

C 117 27 55 83

GROUP 2 VS 3 48 155 19 64

GROUP 1 VS 3 31 66 256 127

GROUP 1 VS 2 38 139 119 41

DISTANCE 18 86 153 34

CONDUCTIVITY 0 3 225 137

40 153 296 113

C/N 159 45 nS nS

FC nS nS nS nS

FN nS nS nS nS

! Nδ15

! Cδ13

!17

Ecological roles

!18

Ecological roles of the fungal communities at each site were inferred using FUNGuild and are visualized in Figure 17 to Figure 19. We assembled the sampling sites into ecologically similar groups, supported by their site similarities (Figure 15) and site groupings according to their PCA (Figure 16).Thus, we assumed that one site is representable for each group. In each representative group, around 45% out of the identified OTUs (Figure 13) could not be classified into functional categories by FUNGuild. This leaves us with ~ 22,5 % of the total OTUs that we were able to assign to functional guilds.

In the forest sites (Figure 17), the classifiable OTUs are mainly allocated into guilds of undefined saprotrophs (genera such as Mortierella, Umbelopsis, Xenasmatella, Cladophialophora, Arachnopeziza, Peniophorella, Exophiala, Filobasidium, Mucor, and Botryobasidium, and the orders Eurotiales, Hypocreales (Pezizomycotina) Saccharomycetales (Saccharomycotina) and Mucorales (Mucormycotina)), ectomycorrhizal fungi (genera such as Cortinarius, Russula, Amanita, Pseudotomentella, Elaphomyces, Lactarius, Tricholoma, Tuber, and Cenococcum), and plant pathogens (Coccomyces, Venturia, Adisciso, Kochiomyces, Verticillium, and Rhizophydium) (Figure 17). These account for around 41 % of the total number of guilds, as the rest are either very small or unclassifiable. The guilds with the highest number of sequences are mostly ectomycorrhizal that grow symbiotrophically.

!19

UnclassifiedUndefined SaprotrophEctomycorrhizalPlant PathogenEricoid MycorrhizalEctomycorrhizal-Endophyte-Ericoid Mycorrhizal-Litter Saprotroph-Orchid MycorrhizalWood SaprotrophAnimal Pathogen-Clavicipitaceous Endophyte-Fungal ParasiteAnimal Pathogen-Dung Saprotroph-Endophyte-Epiphyte-Plant Saprotroph-Wood SaprotrophFungal ParasiteLeaf Saprotroph-Plant Pathogen-Undefined Saprotroph-Wood SaprotrophPlant Pathogen-Wood SaprotrophSoil SaprotrophAnimal PathogenAnimal Pathogen-Endophyte-Epiphyte-Plant Pathogen-Undefined SaprotrophAnimal Pathogen-Endophyte-Fungal Parasite-Plant Pathogen-Wood SaprotrophAnimal Pathogen-Endophyte-Plant Pathogen-Wood SaprotrophBryophyte Parasite-Ectomycorrhizal-Ericoid Mycorrhizal-Undefined SaprotrophBryophyte Parasite-Ectomycorrhizal-Ericoid Mycorrhizal-Undefined Saprotroph-Wood SaprotrophBryophyte Parasite-Leaf Saprotroph-Soil Saprotroph-Undefined Saprotroph-Wood SaprotrophDung Saprotroph-Wood SaprotrophEctomycorrhizal-Lichenized-Wood SaprotrophEndomycorrhizal-Plant Pathogen-Undefined SaprotrophEndophyteEndophyte-Litter Saprotroph-Undefined SaprotrophEndophyte-Litter Saprotroph-Wood SaprotrophFungal Parasite-Lichen ParasiteFungal Parasite-Undefined SaprotrophLichen Parasite-Undefined SaprotrophLichenized-Undefined SaprotrophOrchid MycorrhizalPlant Saprotroph-Wood SaprotrophNULL

UnclassifiedUndefined SaprotrophArbuscular MycorrhizalPlant PathogenWood SaprotrophFungal Parasite-Undefined SaprotrophEctomycorrhizalEndophyte-Plant Pathogen-Wood SaprotrophLichenized-Undefined SaprotrophPlant Pathogen-Soil Saprotroph-Wood SaprotrophEndophyteSoil SaprotrophAnimal Pathogen-Dung Saprotroph-Endophyte-Lichen Parasite-Plant Pathogen-Undefined SaprotrophAnimal Pathogen-Plant Pathogen-Undefined SaprotrophEctomycorrhizal-Fungal Parasite-Soil Saprotroph-Undefined SaprotrophEndomycorrhizal-Plant Pathogen-Undefined SaprotrophEricoid MycorrhizalFungal ParasiteAnimal PathogenAnimal Pathogen-Dung Saprotroph-Endophyte-Epiphyte-Plant Saprotroph-Wood SaprotrophAnimal Pathogen-Endophyte-Plant Pathogen-Wood SaprotrophEndophyte-Plant PathogenLeaf Saprotroph-Plant Pathogen-Undefined Saprotroph-Wood SaprotrophPlant Pathogen-Wood SaprotrophPlant SaprotrophAnimal Parasite-Fungal ParasiteAnimal Pathogen-Endophyte-Fungal Parasite-Plant Pathogen-Wood SaprotrophAnimal Pathogen-Undefined SaprotrophDung SaprotrophDung Saprotroph-Endophyte-Litter Saprotroph-Undefined SaprotrophDung Saprotroph-Plant Saprotroph-Soil SaprotrophDung Saprotroph-Wood SaprotrophEndophyte-Litter Saprotroph-Wood SaprotrophFungal Parasite-Lichen ParasitePlant Pathogen-Undefined SaprotrophPlant Saprotroph-Wood SaprotrophUndefined Saprotroph-Wood SaprotrophAnimal Endosymbiont-Animal Pathogen-Undefined SaprotrophAnimal Pathogen-Clavicipitaceous Endophyte-Fungal Parasite

In the meadow/scattered tree sites (Figure 17), the classifiable OTUs are mainly allocated into guilds of undefined saprotrophs (genera such as Mortierella, Umbelopsis, Tetracladium, Luellia, Rhizophlyctis, Marasmiellus, Apodus, Exophiala, Dactylella, Ophiosphaerella, and Coprinopsis, and orders such as Eurotiales, Hypocreales (Pezizomycotina) Saccharomycetales (Saccharomycotina), and Mucorales (Mucormycotina)), arbuscular mycorrhizal fungi (genera such as Claroideoglomus, Scutellospora, Archaeospora, Rhizophagus, and Acaulospora as well as three orders of Glomeromycota (Archaeosporales, Paraglomerales, and Glomerales)), and plant pathogens (genera such as Rhizophydium, Entorrhiza, Drechslera, Ustilago, Waitea, Venturia, Itersonilia, and Pseudopeziza) (Figure 17). These account for around 33 % of the total number of guilds, and the rest are either very small or unclassifiable. The guilds with the highest number of sequences are a diverse selection of saprotrophic, lichenized, ectomycorrhizal, fungal parasites, and/or plant pathogens that show a variety of saprotrophic, pathotrophic, and/or symbiotrophic.

In the meadow sites (Figure 18), the classifiable OTUs are mainly allocated into guilds of undefined saprotrophs (genera such as Mortierella, Umbelopsis, Geoglossum, Tetracladium, Clavulinopsis, Pholiota, Gymnostellatospora, Glutinoglossum, Leohumicola, Brevicellicium, and Clavaria, and orders such as Eurotiales, Hypocreales (Pezizomycotina) Saccharomycetales (Saccharomycotina) and Mucorales (Mucormycotina)), arbuscular mycorrhizal fungi (genera such as Claroideoglomus, and Rhizophagus as well as two orders of Glomeromycota (Gigasporales, and Glomerales)), and plant pathogens (genera such as Entorrhiza, Devriesia, Drechslera, Ustilago, Thanatephorus, Clonostachys, Itersonilia, Plectosphaerella, and Septoria)

(Figure 18). These account for around 30 % of the total number of guilds, and the rest are either very small or unclassifiable. The guilds with the highest number of sequences are lichenized and/or saprotrophs that grow

!20

UnclassifiedUndefined SaprotrophArbuscular MycorrhizalPlant PathogenFungal Parasite-Undefined SaprotrophEctomycorrhizal-Fungal Parasite-Soil Saprotroph-Undefined SaprotrophLichenized-Undefined SaprotrophEctomycorrhizalEndophyte-Plant PathogenSoil SaprotrophAnimal Pathogen-Dung Saprotroph-Endophyte-Epiphyte-Plant Saprotroph-Wood SaprotrophAnimal Pathogen-Dung Saprotroph-Endophyte-Lichen Parasite-Plant Pathogen-Undefined SaprotrophDung Saprotroph-Endophyte-Litter Saprotroph-Undefined SaprotrophEndophyteEricoid MycorrhizalLeaf Saprotroph-Plant Pathogen-Undefined Saprotroph-Wood SaprotrophUndefined Saprotroph-Undefined BiotrophWood SaprotrophAnimal Pathogen-Endophyte-Plant Pathogen-Wood SaprotrophAnimal Pathogen-Plant Pathogen-Undefined SaprotrophEndomycorrhizal-Plant Pathogen-Undefined SaprotrophEndophyte-Plant Pathogen-Wood SaprotrophPlant Pathogen-Soil Saprotroph-Wood SaprotrophPlant Pathogen-Wood SaprotrophAnimal Parasite-Fungal ParasiteAnimal Pathogen-Endophyte-Epiphyte-Plant Pathogen-Undefined SaprotrophAnimal Pathogen-Endophyte-Ericoid Mycorrhizal-Plant Pathogen-Wood SaprotrophAnimal Pathogen-Endophyte-Fungal Parasite-Plant Pathogen-Wood SaprotrophAnimal Pathogen-Soil SaprotrophAnimal Pathogen-Undefined SaprotrophDung Saprotroph-EndophyteDung Saprotroph-Plant Saprotroph-Soil SaprotrophDung Saprotroph-Wood SaprotrophEctomycorrhizal-Fungal Parasite-Plant Saprotroph-Wood SaprotrophEndophyte-Plant Pathogen-Undefined SaprotrophFungal ParasiteFungal Parasite-Lichen ParasiteLichen Parasite-Undefined SaprotrophNULLOrchid Mycorrhizal

UnclassifiedUndefined SaprotrophArbuscular MycorrhizalPlant PathogenWood SaprotrophEctomycorrhizalFungal Parasite-Undefined SaprotrophEctomycorrhizal-Fungal Parasite-Soil Saprotroph-Undefined SaprotrophEndomycorrhizal-Plant Pathogen-Undefined SaprotrophFungal ParasiteEndophyte-Plant Pathogen-Wood SaprotrophPlant Pathogen-Soil Saprotroph-Wood SaprotrophAnimal Pathogen-Plant Pathogen-Undefined SaprotrophEndophyteAnimal Pathogen-Dung Saprotroph-Endophyte-Epiphyte-Plant Saprotroph-Wood SaprotrophPlant Saprotroph-Wood SaprotrophSoil SaprotrophUndefined Saprotroph-Wood SaprotrophDung SaprotrophEndophyte-Plant PathogenLichenized-Undefined SaprotrophPlant Pathogen-Wood SaprotrophAnimal Parasite-Fungal ParasiteAnimal PathogenAnimal Pathogen-Dung Saprotroph-Endophyte-Lichen Parasite-Plant Pathogen-Undefined SaprotrophAnimal Pathogen-Endophyte-Plant Pathogen-Wood SaprotrophAnimal Pathogen-Undefined SaprotrophDung Saprotroph-Endophyte-Litter Saprotroph-Undefined SaprotrophEndophyte-Litter Saprotroph-Wood SaprotrophPlant Pathogen-Undefined SaprotrophPlant SaprotrophAnimal Endosymbiont-Undefined SaprotrophAnimal Pathogen-Endophyte-Ericoid Mycorrhizal-Plant Pathogen-Wood SaprotrophAnimal Pathogen-Endophyte-Fungal Parasite-Plant Pathogen-Wood SaprotrophBryophyte Parasite-Ectomycorrhizal-Ericoid Mycorrhizal-Undefined SaprotrophDung Saprotroph-Plant Saprotroph-Soil SaprotrophDung Saprotroph-Wood SaprotrophDung Saprotroph-Wood SaprotrophEricoid MycorrhizalFungal Parasite-Lichen ParasiteLeaf Saprotroph-Plant Pathogen-Undefined Saprotroph-Wood SaprotrophLeaf Saprotroph-Plant Pathogen-Undefined Saprotroph-Wood SaprotrophAnimal Endosymbiont-Animal Pathogen-Undefined SaprotrophAnimal Pathogen-Clavicipitaceous Endophyte-Fungal ParasiteAnimal Pathogen-Dung Saprotroph-Endophyte-Plant Saprotroph-Soil Saprotroph-Wood SaprotrophAnimal Pathogen-Endophyte-Epiphyte-Plant Pathogen-Undefined SaprotrophBryophyte Parasite-Leaf Saprotroph-Wood SaprotrophDung Saprotroph-Leaf SaprotrophDung Saprotroph-Soil Saprotroph-Wood SaprotropDung Saprotroph-Undefined SaprotrophEndophyte-Epiphyte-Fungal Parasite-Insect ParasiteEndophyte-Lichen Parasite-Plant Pathogen-Undefined SaprotrophEndophyte-Undefined SaprotrophFungal Parasite-Litter SaprotrophFungal Parasite-Plant PathogenLeaf SaprotrophLeaf Saprotroph-Wood Saprotroph

saprotrophic or saprotrophic-symbiotrophically. In the meadow/reed sites (Figure 18), the classifiable OTUs are mainly allocated into guilds of

undefined saprotrophs (genera such as Mortierella, Preussia, Exophiala, Ophiosphaerella, Rhizophlyctis, Tetracladium, Clitopilus, Arthrobotrys, Coprinopsis, Lycoperdon, and Mucor, and orders such as Eurotiales, Hypocreales (Pezizomycotina) Saccharomycetales (Saccharomycotina), and Mucorales (Mucormycotina)), arbuscular mycorrhizal fungi (genera such as Claroideoglomus, Diversispora, and Scutellospora as well as the order Paraglomerales and five orders of Glomeromycota (Paraglomerales, Diversisporales, Archaeosporales, Gigasporales, and Glomerales)), and plant pathogens (genera such as Ustilago, Urocystis, Plectosphaerella, Neonectria, Clonostachys, Drechslera,

Rhizophydium, Volutella,

and Itersonilia) (Figure 18). These account for around 30

% of the total number of guilds, and the rest are either very small or unclassifiable. The guilds with the highest number of sequences are saprotrophs, plant-, and animal pathogens that grow pathotrophically and/or saprotrophically.

In the AB transect, the classifiable OTUs of the benthic sites (Figure 19) are mainly allocated into guilds of undefined saprotrophs (genera such as Hypholoma, Pseudeurotium, Mortierella, Umbelopsis, Lycoperdon, Teichospora, Geoglossum, Leohumicola, Clitopilus, Exophiala, and Mucor, and orders such as Eurotiales, Hypocreales (Pezizomycotina) Saccharomycetales (Saccharomycotina), and Mucorales (Mucormycotina)), plant pathogens (Urocystis, Ustilago, Paradendryphiella, Leptosphaeria, Volutella, and Microbotryum), and ectomycorrhizal fungi (Byssocorticium, Chloridium, Suillus, and Amphinema). These account for about 38 % of the total number of guilds, and the rest are either very small or unclassifiable. In the SA transect, the classifiable OTUs of the benthic sediment (Figure 19) are mainly allocated into guilds of undefined saprotrophs (genera such as Hypholoma, Pseudeurotium, Mortierella, Umbelopsis, Paraphaeosphaeria, Glutinoglossum, Lycoperdon, Teichospora, Sarocladium, Exophiala, Leohumicola, and Mucor, and orders such as Eurotiales, Hypocreales (Pezizomycotina) Saccharomycetales (Saccharomycotina), and Mucorales (Mucormycotina)), and lichenized-undefined saprotrophs (Clavariaceae), and plant pathogen-wood saprotrophs (Phoma, Stemphylium, and Fusarium) (Figure 19). These account for about 35 % of the total number of guilds, and the rest are either very small or unclassifiable. Of both gradients, guilds with the highest number of sequences are endophyte plant-, and/or animal pathogens that grow pathotrophically or pathotroph-saprotroph-symbiotrophically.

!21

UnclassifiedUndefined SaprotrophLichenized-Undefined SaprotrophPlant Pathogen-Wood SaprotrophAnimal Pathogen-Dung Saprotroph-Endophyte-Epiphyte-Plant Saprotroph-Wood SaprotrophArbuscular MycorrhizalEctomycorrhizalPlant PathogenUndefined Saprotroph-Undefined BiotrophUndefined Saprotroph-Wood SaprotrophAnimal Endosymbiont-Undefined SaprotrophAnimal Pathogen-Dung Saprotroph-Endophyte-Lichen Parasite-Plant Pathogen-Undefined SaprotrophAnimal Pathogen-Endophyte-Epiphyte-Plant Pathogen-Undefined SaprotrophAnimal Pathogen-Endophyte-Lichen Parasite-Plant Pathogen-Wood SaprotrophAnimal Pathogen-Endophyte-Plant Pathogen-Wood SaprotrophAnimal Pathogen-Plant Pathogen-Undefined SaprotrophEctomycorrhizal-Endophyte-Ericoid Mycorrhizal-Litter Saprotroph-Orchid MycorrhizalEctomycorrhizal-Fungal Parasite-Soil Saprotroph-Undefined SaprotrophEctomycorrhizal-Lichen Parasite-Lichenized-Plant PathogenEndophyteEndophyte-Plant PathogenEndophyte-Plant Pathogen-Wood SaprotrophEndophyte-Wood Saprotroph-Animal PathogenEricoid MycorrhizalFungal Parasite-Undefined SaprotrophLeaf SaprotrophNULLPlant Pathogen-Soil Saprotroph-Wood SaprotrophWood Saprotroph

UnclassifiedUndefined SaprotrophPlant PathogenEctomycorrhizalFungal Parasite-Undefined SaprotrophAnimal Pathogen-Endophyte-Plant Pathogen-Wood SaprotrophAnimal Pathogen-Plant Pathogen-Undefined SaprotrophFungal ParasiteAnimal Pathogen-Clavicipitaceous Endophyte-Fungal ParasiteEndophytePlant Pathogen-Plant SaprotrophPlant SaprotrophSoil SaprotrophWood SaprotrophAnimal Endosymbiont-Undefined SaprotrophAnimal PathogenAnimal Pathogen-Endophyte-Epiphyte-Plant Pathogen-Undefined SaprotrophAnimal Pathogen-Endophyte-Fungal Parasite-Plant Pathogen-Wood SaprotrophEctomycorrhizal-Lichen Parasite-Lichenized-Plant PathogenEndophyte-Epiphyte-Fungal Parasite-Insect ParasiteEndophyte-Plant Pathogen-Wood SaprotrophEndophyte-Wood Saprotroph-Animal PathogenFungal Parasite-Litter SaprotrophLeaf SaprotrophNULLPlant Saprotroph-Wood SaprotrophUndefined Saprotroph-Wood Saprotroph

Figure 19 Ecological roles of sampling site 9, representative for the benthic sites in SA (most abundant guilds: 43,3 %

!22

Discussion Discussion of results

Recent studies point to the existence of diverse, but not yet thoroughly understood communities of fungi recovered from environmental sequencing efforts of different marine habitats (Jones, 2011; Ciobanu et al., 2014; Redou et al., 2015; Richards et al., 2015; Zhang et al., 2015; Grossart et al., 2016; Tisthammer et al., 2016; Picard, 2017). In this study, we explored two littoral gradients that were assessed for their fungal abundance, taxonomic composition, diversity, and ecological roles.

The physiochemical properties of the soil constitute an important part of ecosystem characterization, and many geochemical processes within the soil lead to changes in ratios of stable isotopes of carbon and nitrogen. Forest sites in the SA gradient (Table 2) stand out by their increased C and N contents, which could be interpreted as consistent with the amount of organic material available in forests compared to grass landscapes. The range of ! C is indicative for the ratio of C3 to C4 plants. Values between -22 ‰ and - 35 ‰ point towards a predominant growth of C3 plants, which discriminate against the heavier isotope (Tiunov, 2007). C3 plants are for example small seeded cereal crops, most trees and lawn grasses, conifers, deciduous trees and shrubs as well as weedy plants. Many C4 plants are confined to the herbaceous growth form, and most of them are monocots (Still et al., 2003). Their discrimination against the heavier carbon isotope is less strong ( ! C ranges from -11 ‰ to -17 ‰) (Tiunov, 2007). Our results (Table 2) indicate a transition from C3 plants to C4 plants along the gradient from terrestrial towards the sea sediments. The fraction ! N indicates how much the two stable isotopes of nitrogen were fractionated during nitrogen assimilation and can give a clue to the changes caused by processes such as nitrification and ammonification (Tiunov, 2007) and also the trophic level of the organisms in the soil. Bulk tissues of organisms higher up in the trophic chain usually are accumulating the heavier isotope and vice versa (Griffith, 2004). The close-to-zero numbers of ! N measurements in the SA forest sites indicate a relatively higher level of the lighter isotope. This suggests that these soils potentially contained organisms lower in the food chain such as saprotrophs. It also seems to correlate with the higher carbon content in these sites, and not so much the amount of total nitrogen available. There are studies that suggest a relatively higher ! N content when carbon is low, as in mineral soils (Tiunov, 2007; Craine et al., 2015). Our results follow this pattern.

PCR amplification of the ITS2 barcode does not fully reflect the true abundance of species, as it has different copy numbers in different organisms due to a tandem repeat of the whole rDNA gene cluster. Metabarcoding approaches are also sensitive to primer choice as well as PCR biases and thus can only ever be a semi quantitative approach for environmental sampling efforts. To overcome quantification issues, ribosome sequencing could be used. It does not rely on a PCR amplification, as ribosomes are highly abundant in cells and is potentially quantitative. Ribosome sequencing could also prevent the sequencing of metabolically inactive cells such as spores, as ribosomes are built for protein synthesis, and thus are only found within living cells. The IT2 barcode, although robust, is difficult to align across genera or larger taxonomic expanses and so is not always useful for higher-level classification, but it works reasonably well for identification at the species level (Nilsson et al., 2014; Tedersoo et al., 2015). Instead, rRNA gene sequences such as the whole or parts of the LSU could be used to build a more robust phylogenetic backbone (Tedersoo et al. 2018). To account for differences in copy numbers across fungi, we transformed the read numbers into a binary presence/absence matrix for beta diversity analysis.

One suggestion why mostly members of Dikarya are recovered from environmental sequencing efforts of marine fungi (Figure 13) is that they are more readily available due to e.g. methodological shortcomings and biases in sample preparation or analysis. The study of Zhang et al. (2015) reveals a similar taxonomical composition within marine fjord sediments, although they recover a lower number of unclassified fungi of around 18 %. Dini-Andreote et al. (2016) recovered a similar composition within salt marsh sediments at different succession stages. In sparsely vegetated soils with a low nutrient, silt and clay content, a somewhat higher number of sequences (50 %) could not be classified in that study, decreasing with ongoing ecological succession.

It appears that members of early diverging lineages of fungi have to be looked for more specifically, as they seem to be more divergent and particularly distantly related to already described fungi (Richards et al., 2012). Both Richards et al. (2012), and Le Calvez et al. (2009) show results consistent with the assumption that marine environments host numerous, unclassified deep branching lineages of fungi. The results of the present study support these claims (Figure 6, Figure 9), as more than

δ13

δ13

δ15

δ15

δ15

!23

80 % of our sequences cannot be classified at the phylum level, suggesting they may belong to new, undescribed fungal phyla or make up entirely new lineages that are yet unknown to science (Figure 6, Figure 9). Even though we do not know to which phylum the 80 to 100 % of unclassified sequences belong to, we can with relative certainty place them in the fungal kingdom since we used fungal specific primers and thorough data processing to scrutinize for artificial and non-fungal OTUs.

According to Tuomisto (2010) “True diversity is the effective number of types in a classification of interest, and this criterion is fulfilled equally well by alpha, beta, and gamma diversity”, and adds that they just correspond to different measurement units. Diversity consists of two components, namely, richness, and evenness (Magurran, 2004). Although species richness is related to species diversity, the two are not the same thing. Richness is an estimate of the actual number of species, but diversity represents the effective number of species. The Shannon index we used to assess alpha diversity is one of a few conceptually different diversity indices, and is very commonly used (Hill, 1973). The difference between these indices lies in the order of diversity, which results in a different weight given to abundant versus rare species. When calculating richness, all species are treated as equally abundant no matter their actual proportions. Richness, thus, is an estimate of the actual number of species. The Shannon index gives abundant species more weight and quantifies the uncertainty in the species identity when picked randomly from a dataset (Tuomisto, 2010). It is preferred over other indices, as it takes species richness and equitability (evenness) into account, and is thus the closest to the definition of diversity. Our diversity data (Table 4, and Table 5) should be reasonably authentic as indicated by the rarefication curves (Figure 14), which are near the asymptote. They thus indicate a sufficient sampling effort and close-to-complete accounting of fungal richness. To some extent, the slope of the curve represents species diversity. Evenness as well as the Shannon index are higher in the foreshore, and seabed sites of SA than in the forest sites, but quite similar to the soil sampled from the meadow landscape. This suggests that overall diversity in the foreshore and benthic sites are higher than in the forest. Lower evenness numbers would suggest that a few species are relatively more abundant than others, which would have a negative impact on species diversity estimates. In the AB gradient, evenness and richness are consistent, with the exception of seashore site AB7, which seems to be dominated by, and composed of fewer species than the other sites. Lower diversity estimates could for instance be caused by a shortage in DNA extraction from mineral/sandy or very humus rich soils, as the purity measurement of the shoreline and forest sites suggest (Table 9) but could also be due to the bland nature of the site itself.

Although both gradients were geographically close and of overall similar appearance there is no observable pattern when it comes to the correlation of OTUs changing significantly with our metadata (Table 8). Other studies have suggested a strong correlation of some communities of fungi with soil pH (acidic, non-acidic) and site moisture (Taylor et al., 2014), which could potentially be further investigated with our data. The only phylum changing significantly over the distance of our gradient from terrestrial to benthic sediments is Mortierellomycota. More than half of the known lineages of this phylum have been described only within the last 13 years by molecular ecologists (Hibbett, & Glotzer, 2011). They are found mostly as saprobes in soil, but also other organic material such as leaves or dung.

When looking at the sites grouping together in the PCA (Figure 16) and the network plot of site similarities (Figure 15), they fit to the actual habitats from where the samples were taken from well. This suggests that sites of different habitats are indeed more different among each other than within them. It also indicates that it is reasonable to use one of the sites as a habitat representative when it comes to the assignment of OTUs in different guilds. Limitations and shortcomings

DNA extraction and PCR set-up play crucial roles in recovering robust results to our research questions, as targeted fungal DNA is present in very low concentrations before PCR, and contaminations may have a grave influence at this stage. The purity measurements of our samples (Table 9) suggested that the DNA extraction step could be improved by: (1) producing finer soil powder prior to extraction. This could be achieved by using mortar, and pestle or a mill for grinding the soil samples more thoroughly, in case they contain firm material such as minute stones. Better grinding of the soil could provide a more optimal starting point for subsequent DNA extraction as it promotes homogenization, and break down of cell-wall material (Clemmensen et al., 2016). (2) Optimizing the DNA extraction itself, where it seems that the DNeasy PowerSoil Kit did not work optimally for the volumes of marine sediment samples usedas well as for very humus-rich soil. In the latter case, organic soil volume exceeded the cup volume, and the extraction buffer was not able to dissolve the samples completely, thus potentially preventing the extraction of the full amount of DNA. This step could be optimized by

!24

weighting in less organic soil and more mineral soil than proposed by the protocol when using the same kit for multiple soil types, as noted by Clemmensen et al. (2016).

The occurrence of trophic groups such as ectomycorrhizal fungi within the marine sediment could be explained by dead DNA from dispersed spores, fragmented fungal hyphae, or perhaps soil washed into the ocean from land. This relic DNA – DNA that stems from dead or metabolically inactive cells, can sometimes make up to 40 % of ITS sequences recovered from soil (Carini et al., 2016) and is often hard to discriminate. Moreover, the seabed we took the benthic sediment samples from was in both transects particularly close to diverse mixed forests. Another possibility to explain the presence of ectomycorrhizal fungi within marine sediments could be a shortcoming in correct guild assignment in FUNGuild – a resource known to be incomplete, just like all other databases covering the fungal kingdom. Future studies

The fungal communities in our marine samples were similar across the two sites, as were our two terrestrial samples across the two sites. The marine samples were clearly different from the non-marine ones, not least by revealing a large number of abundant OTUs that could not even be identified to the phylum level in spite of our various attempts. Since we used fungus-specific primers, it is reasonable to assume that these organisms indeed are fungi. At the same time, our primers also did recover a few plants and algae, hinting at the possibility that these large OTUs could possibly represent the other kingdoms instead. More data are clearly needed here – ideally data comprising the small ribosomal subunit (SSU) and/or LSU genes for robust and higher-level phylogenetic assignment, or whole ribosomal sequences that is potentially quantitative and reflects activity. Longer sequence data up to a million bases could potentially be produced by novel sequencing techniques of the third sequencing generation, which is currently under active development. Similarly, data from more sampling sites would help secure statistical support in the deduction of trends of correlation and co-variation of the fungal communities and the edaphic factors.

Conclusion What we can confirm with this study is that there is a disproportionately large number of fungal

lineages waiting to be discovered and described especially in marine environments. This study could be used as a starting point for an estimate and characterization of as-yet undescribed fungal species (or even phyla) that could potentially be found in other temperate zones of the northern hemisphere. It seems reasonable to think that other, more exotic localities than the Gothenburg area will be found to host much more diverse fungal communities.

We could resolve most of our terrestrial OTUs to fairly refined taxonomic levels, showing that ‘molecular’ mycology has done a decent job of characterizing terrestrial fungal diversity up to now. Interestingly, we found many “terrestrial” lineages also in the marine samples, showing that aquatic mycology cannot be pursued in isolation from traditional mycology. Similarly, our edaphic data hinted at processes and correlations that we could not fully explain. It seems clear that improved collaboration among scientific disciplines is needed to tackle the explosion of new data emerging in the wake of fungal community metabarcoding.

Our data show that there are predominantly endophytic, parasitic, and pathogenic fungi in the marine environments, which hints at the presence of interesting and currently poorly understood ecological processes in the marine environment. It is also clear from our results that a very large number of marine fungi are in urgent need of taxonomic study and formal description.

Acknowledgements My personal thanks goes to Henrik Nilsson, for the inexhaustible support und always being

available for discussions, questions and inspiration. Also, I want to thank Sarah Bourlat, who has been a huge support in the lab and very appreciated source of knowledge and experience in all fields. Thank you also, Erik Kristiansson, for your indispensable help with the statistical analysis. Furthermore, I want to thank Anna Ansebo for her help in the lab and Camila Ritter for kindly sharing her R-script with me. Last but not least, I want to say thank you to Emanuel for your encouragement and being there for me during the time of my Master’s degree.

!25

Reference list Abarenkov K, Nilsson H, Larsson KH, Alexander IJ, Eberhardt U, Erland S, et al., 2010. The UNITE

database for molecular identification of fungi – recent updates and future perspectives. New Phytologist, 186(2), 281-285.

Albanese D, Fontana P, De Filippo C, Cavalieri D, & Donati C, 2015. MICCA: A complete, and accurate software for taxonomic profiling of metagenomic data. Scientific Reports, 5(1), 9743.

Begerow D, Nilsson H, Unterseher M, & Maier W, 2010. Current state and perspectives of fungal DNA barcoding, and rapid identification procedures. Applied Microbiology, and Biotechnology, 87(1), 99-108.

Blackwell M, 2011. The Fungi: 1, 2, 3 ... 5.1 million species? American Journal of Botany 98(3), 426-438.

Blaxter M, Mann J, Chapman T, Thomas F, Whitton C, Floyd R, & Abebe E, 2005. Defining operational taxonomic units using DNA barcode data. Philosophical Transactions of the Royal Society B: Biological Sciences, 360(1462), 1935-1943.

Bourlat SJ, Haenel Q, Finnman J, & Leray M, 2016. Preparation of amplicon libraries for metabarcoding of marine eukaryotes using Illumina MiSeq: The dual-PCR method. Marine Genomics - Methods and protocols. Springer: New York.

Carini P, Marsden PJ, Leff JW, Morgan EE, Strickland MS, & Fierer N, 2016. Relic DNA is abundant in soil and obscures estimates of soil microbial diversity. Nature Microbiology, 2, 16242.

Chao A, 1984. Non-parametric estimation of the number of classes in a population. Scandinavian Journal of Statistics, 11, 265-270.

Ciobanu MC, Burgaud G, Dufresne A, Breuker A, Rédou V, Ben Maamar S, et al., 2014. Microorganisms persist at record depths in the subseafloor of the Canterbury Basin. The ISME Journal, 8(7), 1370-1380.

Clemmensen K, Ihrmark K, Durling M, & Lindahl BD, 2016. Sample preparation for fungal community analysis by high-throughput sequencing of barcode amplicons. In: Martin F., Uroz S. (eds) Microbial Environmental Genomics (MEG). Methods in Molecular Biology, 1399, 61-88.

Cornish-Bowden A, 1985. Nomenclature for incompletely specified bases in nucleic acid sequences: Recommendations 1984. Nucleic Acids Research, 13(9), 3021-3030.

Craine J, Brookshire M, Cramer E, Hasselquist N, Koba K, Marin-Spiotta M, & Wang D, 2015. Ecological interpretations of nitrogen isotope ratios of terrestrial plants and soils. Plant and Soil, 396(1), 1-26.

Dini-Andreote F, Pylro VS, Baldrian P, Van Elsas JD, & Salles JF, 2016. Ecological succession reveals potential signatures of marine–terrestrial transition in salt marsh fungal communities. The ISME Journal, 10(8), 1984-1997.

Dunbar B, 2008. NASA, Ocean plant life slows down, and absorbs less carbon, https://www.nasa.gov/centers/goddard/news/topstory/2003/0815oceancarbon.html, Nov. 2017.

Edgar RC, 2013. UPARSE: Highly accurate OTU sequences from microbial amplicon reads. Nature Methods, 10(10), 996-998.

Gao SS, Li XM, Du FY, Li CS, Proksch P, & Wang BG, 2010. Secondary metabolites from a marine-derived endophytic fungus Penicillium chrysogenum QEN-24S. Marine Drugs, 9(1), 59-70.

Gazis R, Miadlikowska J, Lutzoni F, Arnold AE, & Chaverri P, 2012. Culture-based study of endophytes associated with rubber trees in Peru reveals a new class of Pezizomycotina: Xylonomycetes. Molecular Phylogenetics and Evolution, 65(1), 294-304.

Griffith G, 2004. The use of stable isotopes in fungal ecology. Mycologist, 18(4), 177-183. Grossart HP, Wurzbacher C, James TY, & Kagami M, 2016. Discovery of dark matter fungi in aquatic

ecosystems demands a reappraisal of the phylogeny and ecology of zoosporic fungi. Fungal Ecology, 19, 28-38.

!26

van Hannen EJ, Mooij WM, van Agterveld MP, Gons HJ, & Laanbroek HJ, 1999. Detritus-dependent development of the microbial community in an experimental system: Qualitative analysis by denaturing gradient gel electrophoresis. Applied and Environmental Microbiology, 65(6), 2478-2484.

Hibbett DS, Binder M, Bischoff JF, Blackwell M, Cannon PF, et al., 2007. A higher-level phylogenetic classification of the Fungi. Mycological Research, 111, 509-547.

Hibbett D, & Glotzer D, 2011. Where are all the undocumented fungal species? A study of Mortierella demonstrates the need for sequence- based classification. New Phytologist, 191, 592-596.

Hibbett D, Abarenkov K, Kõljalg U, Öpik M, Chai B, Cole J, et al., 2016. Sequence-based classification and identification of Fungi. Mycologia, 108(6), 1049-1068.

Hill MO, 1973. Diversity and evenness: A unifying notation and its consequences. Ecology, 54, 427-473. Hoffland E, Kuyper T, Wallander H, Plassard C, Gorbushina A, Haselwandter K, et al., 2004. The role of

fungi in weathering. Frontiers in Ecology and the Environment, 2(5), 258-264. Hooper D, Bignell D, Brown V, Brussard L, Mark Dangerfield J, Wall D, .et al., 2000. Interactions

between aboveground, and belowground biodiversity in terrestrial ecosystems: Patterns, mechanisms, and feedbacks. BioScience, 50(12), 1049-1061.

James TY, Kauff F, Schoch CL, Matheny PB, Hofstetter V, Cox CJ, et al., 2006a. Reconstructing the early evolution of Fungi using a six-gene phylogeny. Nature, 443(7113), 818-822.

James TY, Letcher PM, Longcore JE, Mozley-Standridge SE, Porter D, Powell MJ, Griffith GW, & Vilgalys R, 2006b. A molecular phylogeny of the flagellated Fungi (Chytridiomycota), and description of a new phylum (Blastocladiomycota). Mycologia, 98, 860-871.

Jones EBG, 2011. Are there more marine fungi to be described? Botanica Marina, 54(4), 343-354. Jones EBG, Pang KL, 2012. Marine Fungi and Fungal-like Organisms. Walter de Gruyter GmbH & Co.

KG: Berlin, Boston. Jurasinski G, & Koch M, 2011. Commentary: Do we have a consistent terminology for species

diversity? We are on the way. Oecologia, 167(4), 893-902. Kirk PM, 2018. Species Fungorum (version Oct 2017). In: Roskov Y, Abucay L, Orrell T, Nicolson D,

Bailly N, Kirk PM, Bourgoin T, DeWalt RE, Decock W, De Wever A, van Nieukerken E, Zarucchi J, & Penev L, (eds) Species 2000 & ITIS Catalogue of Life. Species 2000: Naturalis, Leiden, the Netherlands.

Kjer J, Debbab A, Aly AH, & Proksch P, 2010. Methods for isolation of marine-derived endophytic fungi and their bioactive secondary products. Nature Protocols, 5(3), 479-490.

Le Calvez T, Burgaud G, Mahe S, Barbier G, & Vandenkoornhuyse P, 2009. Fungal diversity in deep-sea hydrothermal ecosystems. Applied and Environmental Microbiology, 75(20), 6415-6421.

Lepelletier F, Karpov SA, Alacid E, Le Panse S, Bigeard E, Garcés E, et al., 2014. Dinomyces arenysensisgen. et sp. nov. (Rhizophydiales, Dinomycetaceae fam. nov.), a chytrid infecting marine dinoflagellates. Protist, 165(2), 230-244.

Lindahl BD, Nilsson H, Tedersoo L, Abarenkov K, Carlsen T, Kjøller R, et al., 2013. Fungal community analysis by high-throughput sequencing of amplified markers – a user's guide. New Phytologist, 199(1), 288-299.

Love MI, Huber W, & Anders S, 2014. Moderated estimation of fold change dispersion for RNA-seq data with DESeq2. Genome Biology 15(12), 550.

Magurran A, 2004. Measuring biological diversity. Oxford: Blackwell. Martin M, 2011. Cutadapt removes adapter sequences from high-throughput sequencing reads.

EMBnet.journal, 17(1), 10-12. McNeill J, & International Association for Plant Taxonomy, 2012. International code of nomenclature

for algae, fungi and plants (Melbourne code) adopted by the Eighteenth International Botanical Congress Melbourne, Australia, July 2011 (Regnum vegetabile ; 154). Koeltz Scientific Books: Konigstein, Germany.

!27

Meyer S, Mordhorst T, Lee C, Jensen P, Fenical W, & Köck M, 2010. Penilumamide, a novel lumazine peptide isolated from the marine-derived fungus, Penicillium sp. CNL-338. Organic & Biomolecular Chemistry, 8(9), 2158-2163.

Nguyen NH, 2015. Taxonomy, phylogeny, and genomics of the Wallemiomycetes, and a newly discovered class of extremophilic fungi (PhD thesis, University of Ottawa).

Nguyen NH, Song Z, Bates ST, Branco S, Tedersoo L, Menke J, Schilling JS, & Kennedy PG, 2016. FUNGuild: An open annotation tool for parsing fungal community datasets by ecological guild. Fungal Ecology 20, 241-248.

Nilsson H, Tedersoo L, Abarenkov K, Ryberg M, Kristiansson E, Hartmann M, et al., 2012. Five simple guidelines for establishing basic authenticity, and reliability of newly generated fungal ITS sequences. Mycokeys, 4(4), 37-63.

Nilsson H, Hyde K, Pawłowska J, Ryberg M, TedersooL, Aas A, et al., 2014. Improving ITS sequence data for identification of plant pathogenic fungi. Fungal Diversity, 67(1), 11-19.

Nilsson H, Wurzbacher C, Bahram M, Coimbra VRM, Larsson E, Tedersoo L, et al., 2016. Top 50 most wanted fungi. Mycokeys, 12(12), 29-40.

Oksanen J, Blanchet FG, Friendly M, Kindt R, Legendre P, McGlinn D, Minchin PR, O'Hara RB, Simpson GL, Solymos P, Stevens MHH, Szoecs E, & Wagner H, 2018. vegan: Community ecology package. R package version 2.4-6.

Opel K, Chung D, & McCord B, 2010. A study of PCR inhibition mechanisms using real time PCR* , †. Journal of Forensic Sciences, 55(1), 25-33.

Picard K, 2017. Coastal marine habitats harbor novel early-diverging fungal diversity. Fungal Ecology, 25, 1-13.

Prieto M, Baloch E, Tehler A, & Wedin M, 2013. Mazaedium evolution in the Ascomycota (Fungi) and the classification of mazaediate groups of formerly unclear relationship. Cladistics, 29(3), 296-308.

R Core Team, 2014. R: A language, and environment for statistical computing. R Foundation for Statistical Computing: Vienna, Austria.

Raghukumar S, 2017. Fungi in Coastal, and Oceanic Marine Ecosystems. Springer: Cham, Switzerland. Rédou V, Meslet-Cladiere L, Navarri M, Barbier G, & Burgaud G, 2015. Species richness, and

adaptation of marine fungi from deep-subseafloor sediments. Applied and Environmental Microbiology, 81(10), 3571-3583.

Richards T, Jones M, Leonard G, & Bass D ,2012. Marine fungi: Their ecology, and molecular diversity. Annual Review of Marine Science, 4, 495-522.

Richards T, Leonard G, Mahé F, Del Campo J, Romac S, Jones M, et al., 2015. Molecular diversity, and distribution of marine fungi across 130 European environmental samples. Proceedings of the Royal Society B: Biological Sciences, 282(1819), 20152243.

Rognes T, Flouri T, Nichols B, Quince C, & Mahé F, 2016. VSEARCH: A versatile open source tool for metagenomics. PeerJ, 4, e2584.

Rosling A, Cox F, Cruz-Martinez K, Ihrmark K, Grelet G, Lindahl BD, et al., 2011. Archaeorhizomycetes: Unearthing an ancient class of ubiquitous soil fungi. Science, 333(6044), 876-879.

Schoch CL, Sung GH, López – Giráldez F, Townsend JP, Miadlikowska J, Hofstetter V, et al., 2009. The Ascomycota tree of life: A phylum-wide phylogeny clarifies the origin and evolution of fundamental reproductive and ecological traits. Systematic Biology, 58(2), 224-239.

Schoch CL, KA Seifert, S Huhndorf, Vincent Robert, JL Spouge, CA Levesque, et al., 2012. Nuclear ribosomal internal transcribed spacer (ITS) region as a universal DNA barcode marker for Fungi. Proceedings of the National Academy of Sciences, 109(16), 6241-6246.

Spatafora J, Chang Y, Benny G, Lazarus K, Smith M, Berbee M, et al., 2016. A phylum-level phylogenetic classification of zygomycete fungi based on genome-scale data. Mycologia, 108(5), 1028-1046.

!28

Still C, Berry J, Collatz G, & DeFries R, 2003. Global distribution of C3 and C4 vegetation: Carbon cycle implications. Global Biogeochemical Cycles, 17(1), 6-1-6-14.

Taylor D, Hollingsworth T, MFarland J, Lennon N, Nusbaum C, & Ruess R, 2014. A first comprehensive census of fungi in soil reveals both hyperdiversity, and fine-scale niche partitioning. Ecological Monographs, 84(1), 3-20.

Tedersoo L, Bahram M, Põlme S, Kõljalg U, Yorou NS, Wijesundera R, et al., 2014. Fungal biogeography. Global diversity and geography of soil fungi. Science, 346(6213), 1256688.

Tedersoo L, Anslan S, Bahram M, Põlme S, Riit T, Liiv I et al., 2015. Shotgun metagenomes, and multiple primer pair-barcode combinations of amplicons reveal biases in metabarcoding analyses of fungi. MycoKeys, 10, 1-43.

Tedersoo L, Puusepp R, Nilsson H, & James TY, 2017. Novel soil-inhabiting clades fill gaps in the fungal tree of life. Microbiome, 5(42), 42.

Tedersoo L, Sánchez-Ramírez S, Kõljalg U, Bahram M, Döring M, Schigel D, et al., 2018. High-level classification of the Fungi and a tool for evolutionary ecological analyses. Fungal Diversity, 1-25.

Tisthammer KH, Cobian GM, & Amend AS, 2016. Global biogeography of marine fungi is shaped by the environment. Fungal Ecology, 19, 39-46.

Tiunov A, 2007. Stable isotopes of carbon and nitrogen in soil ecological studies. Biology Bulletin, 34(4), 395-407.

Tuomisto H, 2010. A consistent terminology for quantifying species diversity? Yes, it does exist. Oecologia, 164(4), 853-860.

Wald A, 1945. Sequential tests of statistical hypotheses. Annals of Mathematical Statistics, 16(2), 117-186.

Wurzbacher C, Nilsson H, Rautio M, & Peura S, 2017. Poorly known microbial taxa dominate the microbiome of permafrost thaw ponds. Isme Journal, 11(8), 1938-1941.

Zhang T, Wang NF, Zhang YQ, Liu HY, & Yu LY, 2015. Diversity, and distribution of fungal communities in the marine sediments of Kongsfjorden, Svalbard (High Arctic). Scientific Reports, 5, 14524.

Zhuang WY, & Liu CY, 2012. What an rRNA secondary structure tells about phylogeny of Fungi in Ascomycota with emphasis on evolution of major types of ascus. PLoS ONE, 7(10), e47546.

!29

Popular science summary Fantastic marine fungi and where to find them

The interesting thing about fungi is that they are kind of weird, mysterious organisms. Many people know them as toadstools or mushrooms from when they went mushroom picking with their grandparents as kids. Little did I know when I was young that fungi can be as small as yeasts and most of all – that one can find them not only in the forests or meadows of our childhood, but also within rivers and the entirety of the oceans. Many fungi not only survive, but even thrive in these wet environments where they occur in wide varieties of mostly small forms of livings. Due to their microscopic nature, they are often hard to spot with the naked eye. Modern scientists are using the fungi’s DNA to prove that they are there – which in a scientific terminology is called metabarcoding.

Barcoding provides a quick and relatively straightforward way to identify species, given that there has been somebody before who described the species, so a name can be assigned to the DNA barcode. This is often not the case with fungi, especially in the aquatic environments where there is a lack of described fungal species. Therefore many fungi recovered stay anonymous. But is it true when scientists are claiming they need to be studied more? Why are they important anyway?

I doubt that anyone would argue against it if I say we would have a big problem if penicillin would not have been discovered. Even worse if Penicillium, the fungi that produces this antibiotic, did not have had a name at that time. How would people have communicated their newest achievement in medicine? That is exactly the argument for why it is so urgent to describe all the potentially new species that are still hidden in the dark. Maybe what we never know we will not miss, but it would be naive if we did not confess to ourselves that we at least have a slight idea of what we would miss out on. Especially given everything that has been discovered about fungi so far, for example a number of described species as high as 143 000. Fungi not only play an important role for medical purposes, but also within their ecosystems, on land as well as in the water.

With the study we carried out we wanted to make a stance for the disproportion of unidentified fungal species comparing land and ocean sediment samples. We took the sediment samples in form of a gradient from land to ocean from two very average looking areas near Gothenburg. An amazing 80 – 100 % of the fungal DNA recovered from the ocean sediments could not be identified, as there is no name to put on those sequences from any existing reference databases. The ecological roles of the remaining identifiable fungi were, as expected, life forms as decomposers or pathogens and parasites – all very small and inconspicuous. The diversity would have been expected to be higher on land, but we learned otherwise, as at least for the two areas in Gothenburg, the diversity of fungi on land and in the ocean was nearly the same.

It could be beyond our imagination how many fungi – yet unknown to science - we would stumble upon if one would look into more exotic places like the tropics. Or even just your grandparents’ backyard.

Abbreviations AB Askimsbadet

FDR False discovery rate

GLM Generalized linear model

HTS High throughput sequencing

ITS Internal transcribed spacer

LSU Large subunit

OTU Operational taxonomic unit

PCA Principal component analysis

PCR Polymerase chain reaction

PTO Phosphothioate oligonucleotides

SA Stora Amundön

SSU Small subunit

!30

Supplements Table 9 DNA concentrations and purities measured with Qubit and NanoDrop after DNA extraction. Absorption ratios of 260/280 indicates the purity of DNA and should be around 1,8. A lower value usually indicates the presence of protein, phenol or other contaminants that absorb strongly at or near 280 nm. The 260/230 ratio is a measurement for nucleic acid purity and lies between 2,0 and 2,2 for pure nucleic acid samples. A lower value indicates the presence of contaminants such as Ethylenediaminetetraacetic (EDTA), phenol and carbohydrates.

Table 10 Raw data statistics, showing the total number of bases, reads, GC content GC (%), AT content AT (%), the ratio of reads that have phred quality score of over 20 Q20 (%) and the ratio of reads that have phred quality score of over 30 Q30 (%).

DNA [NG/ΜL] (QUBIT) DNA [NG/ΜL] (NANODROP) 260/280 260/230

AB1 37,80 51,30 1,84 1,80

AB2 20,80 26,10 1,84 1,44

AB3 82,60 77,90 1,87 1,86

AB4 87,60 90,20 1,85 1,85

AB5 189,00 181,00 1,88 2,08

AB6 150,00 141,80 1,87 1,98

AB7 6,92 8,50 1,58 1,14

AB8 10,10 13,90 1,69 1,13

AB9 7,68 10,10 1,67 1,05

SA1 1,35 13,10 1,41 0,62

SA2 10,70 13,10 1,59 1,37

SA3 4,14 4,50 1,48 0,94

SA4 33,20 29,00 1,92 1,81

SA5 30,80 28,40 1,86 1,72

SA6 33,20 47,70 1,84 1,85

SA7 1,36 2,80 1,20 0,65

SA8 4,26 4,30 1,40 0,79

SA9 3,96 4,50 1,37 0,70

Sample Total Bases Read Count GC (%) AT (%) Q20 (%) Q30 (%)

AB1 195217162 648562 48,72 51,28 90,77 85,44

AB2 192447962 639362 50,02 49,98 90,82 85,56

AB3 189912338 630938 50,48 49,52 90,71 85,48

AB4 229122404 761204 50,04 49,96 90,72 85,35

AB5 195539834 649634 50,35 49,65 90,86 85,60

AB6 216373850 718850 50,32 49,68 90,25 84,99

AB7 249300842 828242 49,51 50,49 91,18 86,05

AB8 186239536 618736 47,68 52,32 90,74 85,28

AB9 170912616 567816 46,20 53,80 91,70 86,47

SA1 157920854 524654 48,68 51,32 91,64 86,19

SA2 151210360 502360 49,03 50,97 91,62 86,41

!31

! Figure 20 Relative abundance of reads at family level (SA)

! Figure 21 Relative abundance of reads at species level (SA)

!

SA3 191141020 635020 46,14 53,87 92,62 87,51

SA4 205405410 682410 49,85 50,15 90,41 84,84

SA5 180246024 598824 48,77 51,23 91,82 86,72

SA6 218058848 724448 50,45 49,56 90,25 84,86

SA7 169757980 563980 47,86 52,14 90,98 85,63

SA8 170826530 567530 50,75 49,25 89,93 84,31

SA9 173884690 577690 50,60 49,40 90,75 85,47

!32

Figure 22 Relative abundance of reads at family level (AB)

! Figure 23 Relative abundance of reads at species level (AB)

Additional supplementary material can be found under the following link: http://www.emerencia.org/Retter_supplementary.zip

!33