Epichloë Endophytes Shape the Foliar Endophytic Fungal Microbiome and Alter the Auxin and Salicylic Acid Phytohormone Levels in Two Meadow Fescue Cultivars

Plants harbor a large diversity of endophytic microbes. Meadow fescue (Festuca pratensis) is a cool-season grass known for its symbiotic relationship with the systemic and vertically—via seeds—transmitted fungal endophyte Epichloë uncinata, yet its effects on plant hormones and the microbial community is largely unexplored. Here, we sequenced the endophytic bacterial and fungal communities in the leaves and roots, analyzing phytohormone concentrations and plant performance parameters in Epichloë-symbiotic (E+) and Epichloë-free (E-) individuals of two meadow fescue cultivars. The endophytic microbial community differed between leaf and root tissues independent of Epichloë symbiosis, while the fungal community was different in the leaves of Epichloë-symbiotic and Epichloë-free plants in both cultivars. At the same time, Epichloë symbiosis decreased salicylic acid and increased auxin concentrations in leaves. Epichloë-symbiotic plants showed higher biomass and higher seed mass at the end of the season. Our results demonstrate that Epichloë symbiosis alters the leaf fungal microbiota, which coincides with changes in phytohormone concentrations, indicating that Epichloë endophytes affect both plant immune responses and other fungal endophytes. Whether the effect of Epichloë endophytes on other fungal endophytes is connected to changes in phytohormone concentrations remains to be elucidated.


Introduction
Microbes are essential organismal partners that inhabit every living being on the earth, contributing to the vital life-sustaining functions of their host organism [1,2]. Hostassociated microbes interact not only with their host, but also with coexisting microbes, including bacteria, fungi, and viruses [3,4], thus forming a complex "microbial community." The importance of the microbiome constituting the genetic component of these microbial communities and their metabolic activities has been widely acknowledged. Current sustainable approaches to improve the fitness and health of host plants and animals comprise modifying the parameters of their microbial partners [5,6].
Plants are increasingly studied for their microbiome and associated improvement of performance and fitness, particularly under extreme events driven by climate change [7]. The plant holobiont theory defines plants and their microbiome as one entity, and here, plant performance and fitness can be enhanced by association with certain microbial partners [8][9][10]. The plant microbiome can be composed of microbes residing on plant surfaces (epiphytes) or inside tissues (endophytes). Endophytes can be systemic microbes that exist in the plant seed and are vertically transmitted in the host plants. Nonsystemic symbiotic (E+) and nonsymbiotic (E-) plants were determined using targeted sequencing of the 16S rRNA gene and ITS (internal transcribed sequence) regions. Additionally, plant performance and phytohormone concentrations of the E+ and E-host cultivars were studied to understand their role in shaping the structure and composition of endophytic microbial communities associated with Epichloë. We hypothesize that Epichloë endophytes (1) shape the endophytic microbial communities particularly aboveground, (2) which coincides with changes in phytohormone concentrations, and (3) improves plant performance.

Plant Material and Study Setup
We chose two meadow fescue (Festuca pratensis L.) cultivars ('Kasper', and 'Valtteri') as our model because meadow fescue is widely used as pasture and forage grass in Europe and commonly harbors the systemic seed transmitted fungal endophyte Epichloë uncinata [(W. Gams, Petrini and D. Schmidt) Leuchtm. and Schardl.]. We obtained the seeds of the cultivars from seed producers via the Finnish Food Authority (www.ruokavirasto.fi, accessed on 15 March 2017). Through staining and microscopy of 100 seeds, we verified that all the seed lots of the cultivars had both Epichloë-free (E-) and Epichloë-symbiotic (E+) seeds [43]. The endophyte frequency in the seed lots varied between 30% and 90%. We verified the final endophyte status (E-or E+) of each experimental plant in the field using the results from ITS-targeted polymerase chain reaction (PCR) for fungal community analysis. Because of the mortality of the plants, at the end of the experiment, we had 7 Eand 5 E+ 'Kasper' plants and 5 E-and 5 E+ 'Valtteri' plants.
We first grew plants in pots (May 2017) in the greenhouse and then transferred 20 plants from each cultivar to the experimental field in the Turku University Botanical Garden (60 • 26 N, 22 • 10.4 E) (June 2017) in a randomized design. The plants were planted 0.5 m apart from each other, hand-weeded to avoid competition from other plant species, and watered if needed during the first growing season. The field was fenced with a metal net to keep out mammalian herbivores. The soil in the experimental area was~90% clay with some sand and peat. The soil pH was 6.2, and content of phosphorus 4.2 mg/L, potassium 250 mg/L, calcium 1900 mg/L, magnesium 570 mg/L, sulfur 10.6 mg/L, zinc 2.74 mg/L, copper 7.5 mg/L, and manganese 15 mg/L. We did not use fertilizers or pesticides in the area. The average temperature (Jun-Aug) was 17 • C, and the average relative humidity was 70% at the field site during the growing season (data accessed on 15 December 2022 from Finnish Meteorological Institute; www.ilmatieteenlaitos.fi).

Plant Performance Parameters
We recorded the growth, reproduction, and chlorophyll content (SPAD) of the experimental plants at the beginning of August 2019. To estimate the growth, we measured the average height of the plant, height of the longest leaf, circumference of the tuft at 5 cm above the ground, and number of tillers. For the reproduction estimate, we counted the number of flower heads, and for the chlorophyll content, we measured the SPAD values (Minolta SPAD-502 Plus meter) from three randomly chosen leaves per experimental plant.

Sample Collection
We collected plant samples for phytohormone and microbiome analysis from the experimental field in mid-August 2019. For phytohormone analysis, we sampled two to three healthy leaves from each plant, which were weighed and immediately stored in liquid nitrogen until further processing.
We sampled two to three healthy leaves and approximately 100 mg of root from five to seven replicates each of E+ and E-'Valtteri' and 'Kasper' cultivars in sterile plastic bags on ice. In the laboratory, the samples were washed with tap water, dried, and weighed to ensure 100 mg samples. The samples were then surface sterilized in a laminar air flow hood using 70% ethanol (1 min), followed by 3% sodium hypochlorite solution (3 min), rinsed thrice with sterile distilled water (3 × 1 min), and air-dried. The samples were transferred to 2 mL microcentrifuge tubes and stored at −80 • C until further processing.

Microbiome Analysis
The frozen samples were homogenized using a bead mill homogenizer (Bead Ruptor 96-Well Plate Homogenizer, Kennesaw, GA, OMNI International US), and DNA was extracted using an Invisorb Spin Plant Mini Kit (STRATEC Biomedical AG, Birkenfeld, Berlin, Germany). Following DNA extraction, the 16S rRNA gene and ITS regions from the DNA samples were amplified by PCR.
From the DNA samples, the variable regions V6-V8 of bacterial 16S rRNA genes were amplified by a nested PCR approach. We performed the first round of PCR reactions with 30 ng DNA, 1X PCR buffer, 0.2 mM dNTPs, 0.3 µM each of primers 799F (AACMGGATTA-GATACCCKG) [44], and 1492R (GGYTACCTTGTTACGACTT) (modified from [45]) and 2000 U/mL GoTaq DNA Polymerase (Promega, Madison, WI, USA) in a 30 µL reaction volume. This was followed by a second round of PCR with 1:10 diluted PCR product from round 1, 1X PCR buffer, 0.2 mM dNTPs, 0.3 µM and M13-1062F (TGTAAACGACG-GCCAGTGTCAGCTCGTGYYGTGA, [46,47] and 1390R (ACGGGCGGTGTGTRCAA) [48] primers. A third round of PCR was done with 1:1 dilution of the second PCR product, IonA-barcode-M13 primers [47] for sample tagging, and 1390R and1390R-P1 for IonTorrent PGM sequencing. The PCR profile was as follows: 3 min initial denaturation at 95 • C, denaturing at 95 • C for 45 s, annealing at 54 • C for 45 s, and extension at 72 • C for 1 min, and final extension at 72 • C for 5 min was set on a thermocycler (BIORAD). The same protocol was followed with 35 cycles for first round, and 25 and 8 cycles for the second and third rounds, respectively.
The ITS regions were amplified using primers M13-ITS7F (TGTAAAACGACGGCCAGT-GTGARTCATCGAATCTTTG) and ITS4R (TCC TCC GCT TAT TGA TAT GC) [49]. The 30 µL reaction mixture contained 30 ng of sample DNA, 1x PCR buffer, 0.2 mM dNTPs, 0.3 µM of each primer, and 1250 U/mL GoTaq DNA Polymerase (Promega, Madison, WI, USA). We performed the second round of PCR using 1:10 dilution of the first PCR product as a template, 1x PCR buffer, 0.2 mM dNTPs, 0.3 µM of each primer, and 1250 U/mL GoTaq DNA Polymerase, barcode-M13 forward primer, and ITS4-P1 (CCTCTCTATGGGCAGTCG-GTGATTCCTCGCTTATTGATATGC) as the reverse primer. The amplification profile was 5 mins initial denaturation at 95 • C, followed by denaturing, annealing, and extension at 95 • C for 30 s, 55 • C for 30 s, and 72 • C for 1 min, repeated at 35 cycles for the first round and 8 cycles for the second round. The final extension was carried out at 72 • C for 7 mins.
The PCR products were run on 1.5% agarose gel and those without good quality DNA and low amplification were not processed further. Three to seven replicates from each sample were quantified on an Agilent 2100 Bioanalyzer system, pooled to obtain a sequence library, size-fractionated on 2% agarose gel cassette (Marker B) using Pippin Prep (Sage Science, Beverly, MA, USA), and sequenced on Ion 314™ Chip v2 in Ion Personal Genome Machine™ (ThermoFisher Scientific Ltd., Waltham, MA, USA).

Bioinformatics and Statistical Analys
The sequence reads from the microbiome analysis were processed using CLC Genomics Workbench 11.0 with a Microbial Genomics Module (https://digitalinsights.qiagen. com/products-overview/discovery-insights-portfolio/analysis-and-visualization/qiagenclc-microbial-genomics-module/ accessed on 10 October 2020). After filtering low-quality and <150 bp sequence reads, high-quality reads were aligned and clustered into OTUs (Operational Taxonomic Units) at 97% sequence identity. The OTUs were taxonomically classified using reference databases RDP 16S rRNA training set 16 for bacteria and UNITE Fungal ITS trainset 7.1 for fungi (https://rdp.cme.msu.edu accessed on 1 December 2020) [52]. OTUs representing plant genes with less than a total count of 10 reads were eliminated. Rarefaction curves were constructed in Paleontological Statistics (PAST) software version 4.12 (https://www.nhm.uio.no/english/research/resources/past/ accessed on 27 December 2022) for each individual sample, showing the number of observed OTUs to the number of acquired reads. Statistical analyses were conducted separately for bacterial and fungal communities using PRIMER-7 software with PERMANOVA+ add-on (https://www.primer-e.com/ accessed on 6 January 2021). Permutational multivariate analysis of variance (PERMANOVA) tests and principal co-ordinate analysis (PCoA) to find the effect of Epichloë on the overall microbial structure community were performed on Bray-Curtis distance matrices of square root transformed abundance data. With the similarity percentages-species contribution (SIMPER) analysis, we identified OTUs or species majorly contributing to the differences between community structures. The statistical analyses were repeated on fungal datasets after removing OTUs assigned as Epichloë.
Plant performance parameters showed normal distribution and were analyzed with a Student's t-test. Phytohormone concentrations were not normally distributed and analyzed with the generalized linear model (GLM). A graphical illustration was done with the ggplot2 package in the software R.

Epichloë Shapes Endophytic Fungal Community Composition in Leaves and Not in Roots
A total of 276,216 good quality fungal sequence reads were obtained and classified into 438 OTUs representing three phyla (Ascomycota, Basidiomycota, and Glomeromycota) and belonging to 75 families and 103 genera. The rarefaction curve for each sample for fungal microbial community data is given ( Figure 1). The genus Epichloë dominated 65% and 39% of the relative abundance of the fungal communities in leaves of Epichloë-symbiotic (E+) plants of the 'Valtteri' and 'Kasper' cultivars, respectively ( Figure 2). The next major genera were Mycosphaerella and Cadophora. For Mycosphaerella, the relative abundance was lower in E+ plants ('Valtteri' 13%; 'Kasper' 8%) and higher in Eplants for both cultivars ('Valtteri' 31%; 'Kasper' 47%). The relative abundance of Cadophora was lower in E+ plants compared with E-plants in 'Valtteri' (E+ leaves 5%; E-leaves 20%), while in E+ 'Kasper' plants, their relative abundance was higher (E+ leaves 31%; E-leaves 0.02%) ( Figure 1). The presence of Epichloë clearly impacted the composition of the endophytic fungal communities in leaves for both 'Valtteri' and 'Kasper' cultivars, as evidenced by PERMANOVA analysis (Table 1) and visualized by PCoA (Figure 3a). The PERMANOVA and PCoA analysis on Epichloë-depleted datasets was performed to ensure that the fungal community structures were significantly different in the E+ and E-leaves of both cultivars (Table 2, Figure 3b).
With the SIMPER analysis on leaves, the major species that contributed to dissimilarity between E+ and E-plants in both cultivars were Epichloë, Mycosphaerella tassiana, Cadophora, and Heliotales spp. In the Epichloë-depleted datasets, in addition to the above taxa other than Epichloë, the relative abundances of Phaeosphaeria triglochinicola and Pleosporales sps. were relatively lower in VE+ plants. Another interesting observation was that the relative abundance of Pleosporales sps. was higher in VE+ plants compared with VE-plants, but in contrast, the relative abundance of Pleosporales sps. was lower in the KE+ plants. Furthermore, the relative abundance of Vishniacozyma victoriae was remarkably lower in the KE+ leaves (Table 3). A detailed table on the SIMPER analysis is provided in Table S1

Epichloë Shapes Endophytic Fungal Community Composition in Leaves and Not in Roots
A total of 276,216 good quality fungal sequence reads were obtained and classified into 438 OTUs representing three phyla (Ascomycota, Basidiomycota, and Glomeromycota) and belonging to 75 families and 103 genera. The rarefaction curve for each sample for fungal microbial community data is given (Figure 1). The genus Epichloë dominated 65% and 39% of the relative abundance of the fungal communities in leaves of Epichloësymbiotic (E+) plants of the 'Valtteri' and 'Kasper' cultivars, respectively ( Figure 2).  The next major genera were Mycosphaerella and Cadophora. For Mycosphaerella, the relative abundance was lower in E+ plants ('Valtteri' 13%; 'Kasper' 8%) and higher in Eplants for both cultivars ('Valtteri' 31%; 'Kasper' 47%). The relative abundance of Cadophora was lower in E+ plants compared with E-plants in 'Valtteri' (E+ leaves 5%; E-leaves 20%), while in E+ 'Kasper' plants, their relative abundance was higher (E+ leaves 31%; Eleaves 0.02%) (Figure 1). The presence of Epichloë clearly impacted the composition of the

Epichloë Shapes Endophytic Fungal Community Composition in Leaves and Not in Roots
A total of 276,216 good quality fungal sequence reads were obtained and classified into 438 OTUs representing three phyla (Ascomycota, Basidiomycota, and Glomeromycota) and belonging to 75 families and 103 genera. The rarefaction curve for each sample for fungal microbial community data is given (Figure 1). The genus Epichloë dominated 65% and 39% of the relative abundance of the fungal communities in leaves of Epichloësymbiotic (E+) plants of the 'Valtteri' and 'Kasper' cultivars, respectively ( Figure 2).  The next major genera were Mycosphaerella and Cadophora. For Mycosphaerella, the relative abundance was lower in E+ plants ('Valtteri' 13%; 'Kasper' 8%) and higher in Eplants for both cultivars ('Valtteri' 31%; 'Kasper' 47%). The relative abundance of Cadophora was lower in E+ plants compared with E-plants in 'Valtteri' (E+ leaves 5%; E-leaves 20%), while in E+ 'Kasper' plants, their relative abundance was higher (E+ leaves 31%; Eleaves 0.02%) (Figure 1). The presence of Epichloë clearly impacted the composition of the  idenced by PERMANOVA analysis (Table 1) and visualized by PCoA (Figure 3a). The PERMANOVA and PCoA analysis on Epichloë-depleted datasets was performed to ensure that the fungal community structures were significantly different in the E+ and E-leaves of both cultivars (Table 2, Figure 3b).    For the roots, Epichloë status did not impact the overall structure of the fungal communities. The taxonomic distribution of the fungal communities in the roots (Figure 4) showed major genera, including Cadophora, Ophiosphaerella, Phaeosphaeria, Rhexocercosporidium, and Scytalidium, in both cultivars of the E+ and E-plants. We also compared E+ leaves of 'Valtteri' and 'Kasper' and E-plants of both cultivars to analyze whether cultivar-specific changes occurred in the communities, but found no significant differences.

Epichloë Does Not Impact Endophytic Bacterial Community Structure
The bacterial community structure and rarefaction curves ( Figure 5) were deciphered from 153,824 final sequence reads divided into 755 OTUs. The community structure consisted of 12 phyla, as represented by 110 families and 212 genera. The major genera in the leaves were Sphingomonas, Hymenobacter, Massilia, and Methylobacterium and in the roots were Roseiflexus, Shinella, Rhizobium, and Rhizobacter. The taxonomic distribution of bacterial communities in the leaves (Figure 6a) and roots (Figure 6b) of the two cultivars

Epichloë Does Not Impact Endophytic Bacterial Community Structure
The bacterial community structure and rarefaction curves ( Figure 5) were deciphered from 153,824 final sequence reads divided into 755 OTUs. The community structure consisted of 12 phyla, as represented by 110 families and 212 genera. The major genera in the leaves were Sphingomonas, Hymenobacter, Massilia, and Methylobacterium and in the roots were Roseiflexus, Shinella, Rhizobium, and Rhizobacter. The taxonomic distribution of bacterial communities in the leaves (Figure 6a) and roots (Figure 6b) of the two cultivars ('Valtteri' and 'Kasper') was not different between the E+ and E-plants. The PERMANOVA (Supplementary material Table S2) and PCoA analysis (Figure 7) showed there were no significant differences in the bacterial community structures of E+ and E-plants, indicating Epichloë does not impact endophytic bacterial community structure in leaves or roots.

Epichloë Does Not Impact Endophytic Bacterial Community Structure
The bacterial community structure and rarefaction curves ( Figure 5) were deciphered from 153,824 final sequence reads divided into 755 OTUs. The community structure consisted of 12 phyla, as represented by 110 families and 212 genera. The major genera in the leaves were Sphingomonas, Hymenobacter, Massilia, and Methylobacterium and in the roots were Roseiflexus, Shinella, Rhizobium, and Rhizobacter. The taxonomic distribution of bacterial communities in the leaves (Figure 6a) and roots (Figure 6b) of the two cultivars ('Valtteri' and 'Kasper') was not different between the E+ and E-plants. The PER-MANOVA (Supplementary material Table S2) and PCoA analysis (Figure 7) showed there were no significant differences in the bacterial community structures of E+ and E-plants, indicating Epichloë does not impact endophytic bacterial community structure in leaves or roots.

Epichloë Symbiosis Alters Plant Parameters
The Epichloë endophyte (E+)-symbiotic plants were taller and had a higher n of flower heads compared with their nonsymbiotic (E-) counterparts in both cultiv ble 4, Figure 8). In contrast, plant circumference was smaller in the E+ plants of the c 'Kasper' showing a trend to be smaller in the E+ 'Valtteri' cultivar (  Figure 8).

Epichloë Symbiosis Alters Plant Parameters
The Epichloë endophyte (E+)-symbiotic plants were taller and had a higher number of flower heads compared with their nonsymbiotic (E-) counterparts in both cultivars (Table 4, Figure 8). In contrast, plant circumference was smaller in the E+ plants of the cultivar 'Kasper' showing a trend to be smaller in the E+ 'Valtteri' cultivar (Table 4, Figure 8). Chlorophyll content was measured as SPAD value and by trend higher in E+ plants compared with E-plants (Table 4, Figure 8).   Table 4. * = p < 0.05; . = p < 0.1.

Epichloë Symbiosis Alters Plant Hormone Concentrations
Plants symbiotic to Epichloë showed significantly higher auxin concentrations (IAA and PAA) compared with nonsymbiotic plants in both cultivars (Table 5, Figure 9). Furthermore, ABA increased in Epichloë-symbiotic plants from the Kasper cultivar. In contrast, the phytohormone SA was lower in the Epichloë-symbiotic plants from both cultivars (Table 5, Figure 9).

Discussion
In the present study, Epichloë symbiosis (E+) in the host plant meadow fescue, F. pratensis, clearly impacted the structure of the endophytic fungal community in leaves, but not in roots. This confirms our hypothesis that Epichloë endophyte is one of the major factors shaping the endophytic fungal communities in the aboveground parts of its host plants [38]. Epichloë was relatively the most abundant genus in the leaves of E+ plants of both cultivars 'Valtteri' and 'Kasper'. The endophytic fungal community structures of E+ leaves differed considerably from the Epichloë-uninfected (E-) leaves. Epichloë symbiosis did not alter the structure of the endophytic bacterial community in the leaves or roots of the host plant. Finally, E+ plants were taller, with more flower heads in both cultivars. Both cultivars showed higher auxin and lower SA concentrations in E+ plants, and ABA concentrations were significantly higher in E+ plants than in E-plants of the 'Kasper' cultivar. Our study is the first to examine both the endophytic microbial communities and phytohormone concentrations, while the mechanistic link between both needs to be unraveled.
Epichloë endophytes grow in the intercellular space of the leaves, stems, and reproductive tissue of several cool-season grasses and are typically absent in the roots [53,54], indicating that their effect on shaping fungal community structure is limited to aboveground plant parts. This is in agreement with our previous study, where Epichloë symbiosis in tall fescue (Schedonorus phoenix) impacted only the endophytic fungal community in leaves, not the bacterial community [38]. However, Epichloë symbiosis in tall fescue have also been reported not to impact foliar fungal community structures [39]. Significant differences in the structures of fungal communities were observed in the dataset, with and without depleting Epichloë taxa and, hence, confirming the role of Epichloë in shaping the endophytic mycosphere. The most abundant endophytic fungal genera next to Epichloë were Mycosphaerella and Cadophora. Epichloë symbiosis in Lolium perenne similarly impacted its foliar fungal composition, with Mycosphaerella being the second-most dominant

Discussion
In the present study, Epichloë symbiosis (E+) in the host plant meadow fescue, F. pratensis, clearly impacted the structure of the endophytic fungal community in leaves, but not in roots. This confirms our hypothesis that Epichloë endophyte is one of the major factors shaping the endophytic fungal communities in the aboveground parts of its host plants [38]. Epichloë was relatively the most abundant genus in the leaves of E+ plants of both cultivars 'Valtteri' and 'Kasper'. The endophytic fungal community structures of E+ leaves differed considerably from the Epichloë-uninfected (E-) leaves. Epichloë symbiosis did not alter the structure of the endophytic bacterial community in the leaves or roots of the host plant. Finally, E+ plants were taller, with more flower heads in both cultivars. Both cultivars showed higher auxin and lower SA concentrations in E+ plants, and ABA concentrations were significantly higher in E+ plants than in E-plants of the 'Kasper' cultivar. Our study is the first to examine both the endophytic microbial communities and phytohormone concentrations, while the mechanistic link between both needs to be unraveled.
Epichloë endophytes grow in the intercellular space of the leaves, stems, and reproductive tissue of several cool-season grasses and are typically absent in the roots [53,54], indicating that their effect on shaping fungal community structure is limited to aboveground plant parts. This is in agreement with our previous study, where Epichloë symbiosis in tall fescue (Schedonorus phoenix) impacted only the endophytic fungal community in leaves, not the bacterial community [38]. However, Epichloë symbiosis in tall fescue have also been reported not to impact foliar fungal community structures [39]. Significant differences in the structures of fungal communities were observed in the dataset, with and without depleting Epichloë taxa and, hence, confirming the role of Epichloë in shaping the endophytic mycosphere. The most abundant endophytic fungal genera next to Epichloë were Mycosphaerella and Cadophora. Epichloë symbiosis in Lolium perenne similarly impacted its foliar fungal composition, with Mycosphaerella being the second-most dominant fungi apart from Epichloë [37]. In general, the impact of Epichloë symbiosis on the foliar fungal community was more pronounced in the 'Kasper' cultivars compared with 'Valtteri' ones.
The relative abundance of Epichloë OTUs belonging to Epichloë uncinata, here specifically forming symbiotic relations with meadow fescue [55], slightly differed between the cultivars. Epichloë symbiosis is common, but frequencies of infection can vary among cultivars, depending on the infection status of the mother plants and genetic and biotic and abiotic environment [56][57][58][59]. The noticeable pattern of foliar endophytic fungal assemblage observed in the 'Kasper' cultivar reflects the complexity of the relation of mutualistic endophyte Epichloë with its specific host cultivar. From the SIMPER analysis of the fungal community in leaves, the relative abundance of Mycosphaerella tassiana was lower in the E+ plants compared with the E-plants in both varieties, but more prominent in 'Kasper'. M. tassiana is a commonly known plant pathogen in wheat [60] and other grass species [61]. Our results have indicated that Epichloë symbiosis may be an advantage, keeping pathogen such as M. tassiana levels inactive. Several species of Epichloë restrict or inhibit plant pathogens through fungi-fungi interactions, such as producing fungistatic chemicals, competing for favorable niches, or eliciting changes in the host plant endosphere [62][63][64].
The relative abundance of Cadophora was slightly higher in uninfected 'Valtteri' leaves, but significantly lower in uninfected 'Kasper' leaves. The complex cultivar-specific compatibility of each cultivar with Epichloë may explain this contrasting trend of relative abundance of Cadophora, which is considered a pathogenic fungus. The relative abundance of another microbial species, Vishniacozyma victoriae, was higher in the nonsymbiotic leaves of 'Kasper'. Epichloë-symbiotic plants generally have more fitness than their nonsymbiotic counterparts, which then harbor beneficial microbes, such as V. victoriae because it is one of the cold-adapted endophytic fungi producing various bioactive and antimicrobial metabolites, enzymes, and hormones helping in plant growth and ecological adaptation to cold environments [65].
Epichloë-symbiotic plants showed higher plant performance with increased plant height, number of flower heads, and higher chlorophyll content. Similar to our study, better plant growth of Epichloë-symbiotic grasses has been commonly documented and may be attributed to increased concentrations of auxins, which are common growth-promoting phytohormones throughout the plant kingdom and are connected to physiological changes in endophyte-symbiotic grasses [66]. The inducing effect of endophyte symbiosis on SA concentrations has been connected to a better defense response against biotrophic pathogens and piercing sucking insects [32,67]. On the other hand, SA reduction in symbiotic plants has been reported earlier in Epichloë-symbiotic plants and can be an Epichloë-mediated suppression of the host immune response for a better intercellular establishment [68]. Because of the key role of SA in interactions with biotrophic pathogens, the observed reduction in SA concentrations in endophyte-symbiotic plants is likely to be responsible for changes in the endophytic mycosphere [31]. On the other hand, the Epichloë symbiosis may first alter the endophytic microbial community, hence leading to altered plant responses at the hormonal level [18]. Hence, Epichloë symbiosis shapes the endophytic fungal community, and because of this tripartite symbiont-host plant-endophyte interaction, they mutually coexist and coevolve.

Conclusions
In conclusion, our results have demonstrated the central role of systemic Epichloë endophytes in cool-season grasses affecting the community of the fungal leaf endophytic microbiota and indicating a link to plant hormone concentrations related to resistance against biotrophic pathogens. The endophyte features of improving plant resistance, biomass, and reproduction are increasingly unraveled to be linked to multiple transcription factors and hormone regulation. It remains to be elucidated whether the effect presented here on the fungal microbiome is caused by direct effects of the Epichloë fungi on other fungal species or whether it is indirectly caused by Epichloë-mediated changes in the plant physiology, such as phytohormone concentration causes.

Data Availability Statement:
The data is available upon reasonable request.