Patterns of Fungal Diversity in Needles, Rootlets and Soil of Endemic Pinus Peuce

Pinus peuce Griseb is ve-needle pine native in high-altitude montane habitats of the Balkans. The aim was to assess the diversity and composition of fungal communities associated with the soil, rootlets and living needles of P. peuce at three high-altitude forest sites with different edaphic conditions and stand characteristics in southeastern Montenegro. In total, 90 needle, 90 rootlet and 90 soil samples were sampled. DNA amplication using ITS2 rDNA as a marker and high-throughput sequencing resulted in 17,620 high-quality reads, representing 825 fungal taxa. There were 52.5% Basidiomycota, 43.9% Ascomycota and 3.6% Mucoromycotina. There were 118 unique fungal taxa in the rootlets, 230 in the soil and 113 in the needles, with 8 taxa were shared. The most common fungi in rootlets were Rhizopogon mohelensis (11.0%), Suillus sibiricus (8.4%), R. fallax (6.9%), in the soil – Phlebiopsis gigantea (5.1%), Tylospora asterophora (2.9%), Sollicocossima terricola (2.7%), and in the needles – Dothideomycetes sp. 3360_7 (17.6%), Dothideomycetes sp. 3360_10 (10.7%), Leotiomycetes sp. 3360_16 (6.5%). The results showed that the functional tissues and the rhizosphere soil of P. peuce were inhabited by a high diversity of fungi, which were largely specic to each particular substrate, while. the site conditions had only limited effect on associated fungal communities. Associated fungal communities were largely determined by the host tree species, substrate properties and functional capabilities of associated fungi.


Introduction
Pinus peuce is one of the two European pine species from the subgenus Strobus. It is tertiary relict and endemic to the Balkans, where it is rare and has limited distribution between the latitudes 41° and 43° (Fig. 1), i.e. in the areas that include parts of Montenegro, Serbia, Albania, North Macedonia, Greece and Bulgaria (Jovanović 2007; Alexandrov and Andonovski 2011). Pinus peuce is adapted to high altitudes and grows between 1200 m and 2300 m i.e. in areas characterised by cold and high humidity mountain climate. It is predominantly found on silicate soils with a varying soil depth and fertility. On carbonate soils, it is usually present on deeper and more humid soils of northern expositions (Janković et al. 1987). Pinus peuce characteristically forms dense and pure forest stands, but at the upper tree line it often grows in small and scattered stands (Jovanović 2007) (Fig. 2). It can also grow in admixture with Picea abies, Abies alba, and on lower altitudes with Fagus sylvatica. At the upper tree line, it can form mixed forest stands with Pinus heldreichii. In these areas, P. heldreichii inhabits more xerotherm and dryer sites, while P. peuce − more humid and fertile sites (Janković et al. 1987). Historical data suggests that in the past both P. peuce and P. heldreichii formed a well-developed and continuous forest belt in the Balkans (Janković et al. 1987; Stevanović et al. 1995), but as a consequence of extensive exploitation, nowadays the total area occupied by P. peuce is only about 20,500 ha (Alexandrov and Andonovski 2011). Pinus peuce is one of the most valuable conifer tree species in the Balkan (Jovanović 2007;Stevanović et al. 1995; Alexandrov and Andonovski 2011) due to its high ecological adaptability, the provision of ecosystem services such as protection of soil and water as well as biodiversity conservation. In Montenegro, at the optimum growth conditions P. peuce reaches up to 30 m in height and up to 1 m in diameter. Commonly, stems have branches down to the ground, while root systems have a characteristic central root with large lateral roots, which penetrate deep into the soil. In preserved forests, trees can reach the age of ca. 200 years, and individual specimens over 300 years. Even at high altitudes, P. peuce produces high yields of wood, which is very durable. In the past, wood was used in constructions, furniture production, woodcarving and cooperage (Jovanović 2007; Alexandrov and Andonovski 2011), but nowadays it is not harvested. The resin of P. peuce provides high quality derivatives and in the past was used in the chemical industry, optics and pharmacy. In traditional medicine, the resin was used to cure wounds, pectoral, skin and stomach diseases, varicose veins and other illnesses (Alexandrov and Andonovski 2011).
In the last 50 years, only limited natural regeneration was reported from P. peuce forests, though recent observations show more extensive regeneration in the abandoned mountain areas (Fig. 2). Pinus peuce is classi ed as nearly threatened by the International Union for Conservation of Nature Red List of Threatened Species (IUCN) and it is protected in the Balkan countries.
As a protected and highly value trees species, P. peuce requires special attention, i.e. the development and application of conservation measures (Janković et al. 1987; Stevanović et al. 1995). It is expected to adapt well to the climate change and can be a promising tree species for forestry and nature conservation in high altitude mountain areas (Jovanović 2007; Alexandrov and Andonovski 2011).
Fungi play key roles in montane forest ecosystems and represent an essential part of biodiversity (Baldrian 2017;Lazarević and Menkis 2020). They directly in uence several physiological processes in trees and contribute to carbon, nutrient and water cycling (Stewart et al. 2018). Despite their importance and possible impact on tree health, information on fungi associated with P. peuce is scarce. Moreover, the knowledge about fungal communities associated with pines from subgenus Strobus is also limited. Among different fungi associated with trees, ectomycorrhizal (ECM) fungi directly in uencing the uptake of nutrients and water in pines forests, affect soil characteristics and overall plant vitality. They can be essential for the successful regeneration, The aim of this study was to assess the diversity and composition of fungal communities associated with the soil, rootlets and living needles of P. peuce from three high-altitude forest sites with different edaphic conditions and stand characteristics in south-eastern Montenegro.

Study sites
The study sites were at Bogićevica, Visitor and Zeletin (detailed description of each site is below) in south-eastern Montenegro ( Fig. 1, Table 1). These sites represented characteristic P. peuce forest stands growing at ca. 1600-2100 m altitudes and included mixed coniferous or coniferous and deciduous forest stands. The geographical distance between the Bogićevica site and two other sites (Visitor and Zeletin) was about 12−15 km. The distance between Visitor and Zeletin sites was about 5 km. All three sites differed from each other in terms of forest stand composition and age, soil characteristics and altitude (Table 1). Climatic conditions were similar for all three sites.
The climate at the study sites, according to the Köppen climate classi cation, is a humid cold temperate boreal climate with cool summer and cold winter (Dfc, Dfwcx) (Burić et al. 2014). The winter minimum is below − 30°C and the summer maximum is above 10°C. The arid period is absent. However, the primary precipitation maximum is in the late autumn or winter. The secondary precipitation maximum occurs in the late spring or early summer. The winters are usually cold and snowy, while summers are usually cool. In summer, at the altitudes above 1500−1600 m the average air temperature can be above 10 C for up to four months. At the Bogićevica site, at the altitude is of ca. 1700 m, P. peuce is growing in admixture with P. abies and A. alba ( at lower altitudes Abieti -Picenion forest type is situated), while at the altitude of ca. 1800 m, it is growing in admixture with P. abies. Above this altitude, P. peuce forms pure forest stands, which extends up to the uppermost mountain ridges. The sampling site was located in pure, old growth P. peuce forest stand with the majority of trees being between 100 and 200 year-old. Trees were 25-30 m in height and 50-60 cm in diameter. The sampling site was situated on the slope and between altitudes 1970 m and 2100 m, i.e. from the mountain plateau with a glacial lake Ridsko up to the top of the mountain. The slope has north and northeast-facing exposition. The forest site is characterised as a Pinetum peucis montenegrinum Bleč. & Tat. type. The site at Bogićevica has the silicate geological structure: quaternary conglomerates composed of quartz and calcite. Brown acid soil type on quartz sandstone (Fuštić and Đuretić 2000), is variable in depth, but being positioned on the mountain slopes and ridges, is very shallow and skeletal (more than 50% of the skeleton). Transient A(B) and (B)C soil horizon occurs, but horizon (B) is rarely present. The humus layer is thin, dark in colour and densely colonised by the roots of grasses, shrubs and forest trees. A or transient A (B) horizon is brown or dark brown in colour and is 8-12 cm deep.
At the Visitor site, P. peuce is growing from ca. 1600 m altitude, where it is mixed with P. abies and A. alba ( at lower altitudes Abieti -Picenion forest type is situated). Pure P. peuce forest stand, which is dominated by 100-200 year-old trees, is growing above the altitude of ca. 1800 m and extends up to the uppermost mountain ridges (alt. 2210 m). The sampling site was on the exposed mountain slope (alt. 1900-2000 m), on an abandoned mountain pasture, where forest of P. peuce has regenerated. The site has north and northwest-facing exposition and was dominated by 10-60 year-old trees of P. peuce. On the south-facing slope, an old forest stand of P. heldreichii is situated, which has also regenerated and is mixed with P. peuce on the sampling site. The Visitor site is dominated by silicate geological soil structure, where the keratophyres, quartz keratophyres, andensitis, dacities and diabases are the most common eruptive rocks, but the highest mountain peaks are composed of Triassic limestone. On the sampling site, the soil is dystric cambisol on igneous rock (Fuštić and Đuretić 2000). As basic eruptive rocks are compact and resistant to weathering, the soil stays in initial stage for a very long time and its development is very slow. It is mainly shallow and skeletal (with large fragments of stones) and acid. The characteristic soil pro le is A (B) C. A and B horizons are poorly differentiated in terms of colour and mechanical composition. The topsoil layer is dark. Mineralization of humus is very slow, resulting in its accumulation. The B horizon has fragments of brown, yellow or orange colour, which is due to the release of iron.
At the Zeletin site, the forest vegetation is differing compared to vegetation on near neighbouring areas. At 1200-1600 m altitude, Fagus sylvatica forest or mixed F. sylvatica and A. alba forest is situated (forest types Fagetum moesiacae Bleč. et. Lakš, and Ass. Abieti -Fagetum). The sampling was carried out at the altitude of 1600-1800 m, where P. peuce is mixed with F. sylvatica and A. alba, in a dense, old-growth forest stand (100-150 year.) on the mountain slope with east and south-east-facing exposition and was characterised by the well-developed soil. Pure P. peuce forest stands occur at ca. 2060 m altitude and up to the top of the mountain, where these stands alternate with mountain pastures. At the sampling site, the soil is dystric cambisol on chert (Fuštić and Đuretić 2000). In the dense, old-growth and mixed coniferous and deciduous forest, soil is well developed and it has a A (B) C pro le. The A horizon is dark brown in colour, and with a high humus content, loose, and with a crumbly structure; sand-loam or loamy. The B horizon is brown or dark brown in colour, with orange-grey shades and loamy. The soil is waterproof and dry.

Experimental design and sampling
At each study site, stands of P. peuce were healthy-looking i.e. showed no signs of damage or decline. The sampling was carried out in autumn 2015 and at each site included needles, rootlets and the rhizosphere soil of P. peuce. For the sampling of needles, at each site a twig up to 15 cm long, which included the current and the previous year living needles, was randomly collected using secateurs up to 2 m from the ground from each of 30 mature trees, which were situated at a distance of ca. 50 m from each other. Collected needles were visually inspected for the presence of disease symptoms, placed in individual plastic bags, transported to the laboratory and stored at -20 °C before used for DNA extraction.
For the sampling of rootlets and surrounding ne fraction soil, at each site 30 samples were randomly collected using a spade under 5-15 year-old trees of P. peuce growing in the vicinity to mature P. peuce trees used for sampling of needles. Soil samples with rootlets were taken down to 20 cm depth and included lateral rootlets with attached ne roots. Samples were individually packed into plastic bags, transported to the laboratory and kept at 4 °C for a maximum period of one week before processed. For each sample, rootlets were separated from the soil, soaked in cold water for 12 h and gently washed in tap water to remove any of the remaining soil. Fine roots were separated from lateral rootlets, placed in 1.5 ml centrifugation tubes and stored frozen at -20 °C before used for DNA extraction. Individual samples of ne fraction soil, which was sieved using a sieve (mesh size 2 mm × 2 mm), were placed in plastic bags and stored at -20°C before used for DNA extraction. Taken together, the sampling resulted in 90 needle, 90 rootlet and 90 soil samples of P. peuce.

Soil chemical analyses and texture
Before analysis of soil physical and chemical properties, at least 100 g of soil from each sample was dried at room temperature (ca. 21°C) for 24 h. Soil pH was determined using a combined glass-electrode in 1 M KCl of soil suspension 1:2.5 (w/v). The humus content was determined by wet oxidation with 0.02 M KMnO 4 . Total nitrogen was determined by Kjeldahl method (Džamić et al. 1996). The available phosphorus and potassium (extraction with ammonium-acetate-lactate solution at pH = 3.7) was determined according to Egner-Riehm-Domingo method using the spectrophotometry and ame photometry, respectively. Total carbonates was determined by volumetric method (Džamić et al. 1996). isolated from the needles was additionally puri ed using JetQuick DNA puri cation kit (Genomed GmbH, Leinfelden, Germany). The DNA concentration of each sample was determined using a NanoDrop™ One spectrophotometer (Thermo Scienti c, Rodchester, NY, USA) and adjusted to 10 ng/µL. Ampli cation by PCR of the ITS2 rDNA region was done using barcoded fungalspeci c primer gITS7 (Ihrmark et al. 2012) and barcoded universal primer ITS4 (White et al. 1990). All samples of the same substrate (needles, rootlets, or soil) and site were ampli ed using primers with the same barcode, resulting in 9 different barcodes (3 substrates × 3 sites). Ampli cation of multiple samples with the same barcode was done owing to get a broader representation of fungal communities per each substrate and site. Ampli cations were performed using the Applied Biosystems 2720 thermal cycler (Foster City, CA, USA). An initial denaturation step started at 95°C for 2 min, followed by 27 ampli cation cycles of denaturation at 95°C for 30 s, annealing at 55°C for 30 s, and extension at 72°C for 60 s. The thermal cycling was ended by a nal extension step at 72°C for 7 min. The PCR products were analysed using gel electrophoresis on 1% agarose gels stained with Nancy-520 (Sigma-Aldrich, Sweden). PCR products were puri ed using a sodium acetate protocol (Menkis et al.

2015)
. Puri ed PCR products were quanti ed using a Qubit uorometer 4.0 (Thermo Fisher Scienti c, Waltham, MA, USA), and an equimolar mix of all PCR products was used for high-throughput sequencing using a Paci c Biosciences platform (Menlo Park, CA, USA) and one SMRT cell at the SciLifeLab (Uppsala, Sweden).

Bioinformatics
Principles of bioinformatics followed Lynikienė et al. (2020). The sequences obtained were subjected to quality control and clustering in the SCATA NGS sequencing pipeline. The initial procedure started with quality ltering of the sequences that included the removal of sequences shorter than 200 bps, sequences with low read quality, primer dimers and homopolymers, which were collapsed to 3 bps before clustering. Only sequences containing a barcode and primer were retained. Then, the primer and sample barcodes were removed from the sequence, but information on the sample and sequence association was stored as meta-data. A single-linkage clustering based on 98% similarity was used to cluster sequences into different taxa. For each cluster, the sequence of the most common genotype was used for taxonomic identi cation. For clusters containing only two sequences, a consensus sequence was produced. The taxa were taxonomically identi ed using the GenBank database and the Blastn algorithm (Altschul et al. 1997). The following criteria were used for identi cation: sequence coverage > 80%; 94-97% similarity to genus level and ≥ 98% similarity to species level. Sequences deviating from these criteria were identi ed only to a high taxonomic rank and were given unique names. Representative sequences of fungal non-singletons are available from GenBank under accession numbers MZ441403 -MZ442202.

Statistical analyses
The statistical analysis of the soil data was performed using IBM SPSS Statistics 23.0 (New York, USA). The signi cant differences between the means were determined with the one-way ANOVA and Duncan's test at p < 0.05. Differences in richness of fungal taxa in different study sites of P. peuce were compared by nonparametric chi-square testing (Sokal and Rohlf 2005). As each of the datasets was subjected to multiple comparisons, con dence limits for p-values of chi-square tests were reduced the corresponding number of times as required by the Bonferroni correction. The Shannon diversity index, qualitative Sørensen similarity index and principal coordinate analysis (PCA) in Canoco 5 were used to characterise the diversity and composition of fungal communities (Shannon 1948 Soil analyses showed that P. peuce forests were growing on very acidic to acidic soils, which were rich in humus and nitrogen (Table 2). Furthermore, soils were poor in available P, had moderate availability of K, high content of Fe and Mn, optimal content of Zn and low to optimal of Cu. Values for electrolytic conductivity (EC) covers the range from infertile to very fertile soils. The exchangeable Ca and Mg were in a range from low to optimal level. Among the three sites, the most favourable growth conditions was found to be at the Visitor site due to higher soil pH, the highest level of humus as well as of nutrients such as N, P, K, Ca, Mg and Zn ( Table 2). The soil at the Zeletin site had lowest pH value, signi cantly higher level of available Fe and EC value, but signi cantly lower content of available P and Mn. The soil at the Bogićevica site had the lowest content of humus, N and the EC value, but the highest level of Mn and Cu (Table 2).  Table 3. In rootlets, the richness of fungal taxa was signi cantly higher at the Zeletin site than at the Visitor site (p < 0.05), but in this respect, both of these sites did not differ signi cantly from the Bogićevica site (p > 0.05) ( Table 3). In the soil, the richness of fungal taxa was similar between the Zeletin and Visitor sites (p > 0.05), but differed signi cantly from those at the Bogićevica site (p < 0.05) ( Table 3). In the needles, the richness of fungal taxa was signi cantly higher at the Zeletin site vs. the Visitor site (p < 0.05), and both of these showed a signi cantly higher richness of fungal taxa than at the Bogićevica site (p < 0.05) ( Table 3). The number of unique fungal taxa differed among different substrates (roots, soil or needles) as shown in Fig. 3. When all sites were taken together, there were 118 unique fungal taxa in the rootlets, 230 in the soil and 112 in the needles (Fig. 3). There were 340 fungal taxa shared between the rootlet and soil samples, but only 18 taxa were shared between the rootlet and needle samples, and 23 between the soil and needle samples. Overall, only 8 fungal taxa were common to all substrates (rootlets, soil and needles) (Fig. 3).
Within each substrate (soil, rootlets or needles), the composition of fungal classes was similar among different sites (Fig. 4).
Principal coordinates analysis (PCA) of fungal communities explained 35.5% variation on axis 1 and 17.4% on axis 2. PCA showed that fungal communities from the same substrate (samples representing different sites) clustered together (Fig. 5). PCA also showed that fungal communities in different substrates (needles, rootlets and soil) were separated from each other (Fig. 5), showing that fungal communities in each of these substrates were largely different. In the needles, the Shannon diversity index was between 3.3 and 3.5, in rootlets -between 3.1 and 4.1 and in the soil -between 4.1 and 4.8 ( Table 3). The Mann-Whitney test showed that the Shannon diversity index among different substrates did not differ signi cantly (p > 0.05). The Sørensen similarity index of fungal communities was: needles vs. rootlets -0.06, needles vs. soil -0.06, and rootlets vs. soil -0.6.

Discussion
Five-needle pines, which belong to the subgenus Strobus (section Quinquefoliae), are unique in their ecology and distribution. There are 24 Pinus spp. in this subgenus, which are native to Europe, North and Central America and Asia (Gernandt et al. 2005). The majority of these species are characterized by a narrow distribution and high-altitude habitats i.e. growing under harsh environmental conditions. Information about fungal diversity associated with those Pinus spp. is still limited, while the available knowledge is mainly on the above ground sporocarps or on ECM fungi. For example, ECM communities of P. cembra were The results of the present study have expanded the available knowledge on endemic P. peuce, demonstrating that needles, rootlets and the rhizosphere soil are inhabited by taxa-rich communities of fungi ( Table 3). The detected fungal communities were largely speci c to each particular substrate (Fig. 3-5), showing their adaptation and substrate preferences. In support, the qualitative Sørensen similarity index was very low when compared between the above-(needles) and belowground (rootlets or soil) substrates, repeatedly demonstrating the potential importance and functional preferences of associated fungi. Interestingly, the site conditions had only limited effect on associated fungal communities as in different sites these were similar (Fig. 4-5). The latter may suggest that associated fungal communities were largely determined by the host tree species, substrate properties and functional capabilities of associated fungi. This may also suggest that fungi detected at one particular site could be used in different habitats (e.g. for ECM inoculation, or for biocontrol of pests or diseases) i.e. within the distribution of P. peuce.
Among the principal fungi identi ed in the soil and rootlets, there were taxa from genera Suillus and Rhizopogon, which are closely related and almost exclusively restricted to Pinaceae (Bruns et al. 2002). A long co-evolutionary history between Suillus and Pinus species (Wu et al. 2000) represents an example of host speci city and adaptation (Jenkis 2018). Moreover, a limited number of Suillus fungi appear to be speci c to ve-needle pines (Klofac 2013) (Karadelev 1998), Bulgaria (Boev 2011) and Montenegro (Kasom and Karadelev 2012), where it was recorded in native forest stands of P. peuce. In agreement, the results of the present study provided the evidence that S. sibiricus is a common and an important symbiont of P. peuce as it was detected in both the rootlets and the soil ( Rhizopogon mohelensis was one of the most commonly detected fungi in this study, particularly in pure P. peuce forests, but less abundantly found in mixed forests. By contrast, R. falax was commonly detected in all study sites (Table 4). Rhizopogon mohelensis was reported from many countries in Europe (Holec et al. 2013). It belongs to R. roseolus group and sometimes can be confused with R. rubescens. For example, R. salebrosus is exclusively associated with P. strobus (Kohout et al. 2011), what limits such misidenti cation.
It appears that both Suillus and Rhyzopogon fungi possess speci c ecological adaptations important for the establishment of Pinus spp. on marginal habitats and after forest disturbance (Kjøller and Bruns 2003;Mohatt et al. 2008). Such host specialists may often represent the dominant ECM species in high-altitude habitats characterised by extreme conditions (Bruns et al. 2002;Antibus and Cripps 2010). Certain suilloid fungi can be of special importance to ve-needle pines due to host specialisation, high e ciency of nutrient and water uptake and transfer between the symbiosis partners, and an exclusion of mycoheterotrophy (Cripps and Antibus 2011). The common occurrence of suilloid fungi was particularly notable at the Visitor site containing newly regenerated P. peuce trees, but their abundance was lower in old-growth stands (Tables 1 and 3).
Among other fungi commonly detected in rootlets, there were Thelephora terrestris and Phialocephla fortinii (Table 4). Thelephora terestris is a typical ECM fungus of long-distance exploration type. This property allows e cient transportation of nutrients and water over long distances, which appears to be especially suitable for undisturbed, but very stony habitat with high soil heterogeneity (Reiner at al. 2015). Phialocephala fortinii belongs to a complex of dark septate endophytes, which forms nonspeci c associations with many plant hosts. The complex includes ECM-forming fungi as well as numerous pathogenic fungi (Tedersoo et al. 2008). Interestingly, P. fortinii was also commonly detected in rootlets of old growth P. heldreichii in high-altitude habitats, but not in its pioneer forests (Lazarević and Menkis 2017).
Tomentella brasodellea and Tylospora asterophora were the other two ECM fungi found in common association with P. peuce rootlets (  Horton et al. (1999), the sharing of ECM fungi between coniferous trees and plants from Ericaceae may play a major role in plant community dynamics. Soil microorganisms associated with arbutoid members of Ericaceae may enhance growth, survival, mycorrhizal root formation, and nitrogenase activity of conifer tree seedlings (Amaranthus et al. 1990). Fungal networks shared between those hosts (Pinaceae-Ericoid) remain even after disturbance events as these fungi are able to form associations with different plants (Perry et al. 1989). Ericaceous plants was shown to play an important role in the formation of ECM communities associated with P. strobus (Kohout et al. 2011). Besides, ericoid plants are very important in Mediterranean forests as these support ECM and ericoid fungi after re events (Bergero et al. 2003).
Although a number of fungi were shared between the rootlet and soil samples as indicated by a high value of Sørensen similarity index, soil samples were characterised by a higher abundance of saprotrophic fungi as compared to the rootlet samples (dominated by ECM fungi) (Tables 4 and 5), which led to the differentiation of fungal communities in these two substrates (Fig. 4). Among the dominant fungi in the soil, there was P. gigantea, which is known as a common saprophytic fungus that causes white rot in conifer logs and stumps (Copenhaver et al. 2014), thereby playing an important role in the decomposition of conifer wood debris. Its common occurrence at the Bogićevica site was likely associated with the vast availability of dead wood, which was absent at the other two sites. Solicoccozyma terricola was another commonly detected saprotrophic fungus known from soils of temperate forests (Mašínová et al. 2016). Hygrocybe intermedia and Neohygrocybe ingrata were also among dominant fungi, which are known to be common in montane grasslands in Montenegro. These fungi, while being red-listed, are often associated with habitats of P. heldreichii (Perić and Perić 2004), which are nutrient-poor, but support high diversity of fungi (Lazarević and Menkis 2018). Interestingly, recently described ubiquitous soil fungi of the genus Archaeorhizomyces (Rosling et al. 2011) were also detected ( Table 5). Although functional properties, reproduction structures and dispersal strategy of these fungi are largely unknown, the current observation expands available knowledge on the host tree species and geographical distribution.
A number of ECM fungi were also commonly detected in the soil, including Tylospora asterophora, Russula vesca, Inocybe whitei, Laccaria laccata and Cenococcum geophilum ( Table 5), showing that these may be important symbionts of forest trees grown in high-altitude habitats. However, as these ECM fungi were mainly detected at the Zeletin site (Table 5), the possibility should not be excluded that their occurrence was also affected by the other three species (A. alba or F. sylvatica) present there.
Living needles are known to be associated with diverse trophic groups of fungi, including endophytes, epiphytes and pathogens.
Many of these can be functionally important and metabolically active taxa, which respond to changes in the environment (Nguyem et al. 2016). They appear to be able to colonise different tree species, but factors driving their distribution remains largely unclear (Tehronen et al. 2019). Among the dominant fungi associated with the needles of P. peuce, there were Dothideomycetes sp. 3360_7, Dothideomycetes sp. 3360_10 and Leotiomycetes sp. 3360_16 (Table 6), which could not be identi ed to the species or genus level, thereby not only posing a challenge to fungal taxonomy, but also limiting the identi cation of their ecology and functional roles. Further, Sydowia polyspora was detected in P. peuce needles (Table 6), which is the fungus with a wide geographical distribution and common occurrence in Europe (Botella and Diez 2010). The pathogenic behaviour of S. polyspora to young conifers (genera Thuja, Abies, Tsuga, Larix, Picea and Pinus) was previously reported (Talgo et al. 2018). Besides, it has been recently reported as one of the most abundant needle pathogens in high-altitude P. heldreichii forests in Montenegro (Lazarević and Menkis 2020), showing that it is not restricted by harsh environmental conditions prevailing in these habitats. Sydowia symptoms include needle discoloration, necrosis and shoot dieback. However, the fungus is favoured by a warm climate, especially if the host is stressed by drought or insect attack (Munoz -Adalia et al. 2017). Celosporium larixicola was another commonly detected fungus (Table 6), which was recently described from needles of Larix lyallii in Canada (Tsuneda et al. 2010) and needles of P. abies in Sweden (Ngyen et al. 2016). Fungi from the genus Celosporium was shown to be commonly recovered from alpine habitats and may be biotrophic or necrotrophic (Brown et al. 2015). Lophodermium pinastri was also common in needles of P. peuce (Table 6) In summary, P. peuce in high altitude mountain habitats harbour diverse communities of fungi, composition of which appears to be largely determined by the host tree species, substrate properties and functional capabilities of these fungi.   Venn diagram showing the diversity of fungal taxa found in rootlets, soil and needles of Pinus peuce, and the number of fungal taxa shared between different substrates. Samples from different sites are combined.