Certain microscopic fungi live at least a part of their life cycle inside the tissues of the plants without causing visible signs or symptoms and, therefore, are named endophytes ( Petrini, 1996 ). Fungal endophytes are a taxonomically and ecologically heterogeneous group and seem to make up a large fraction of the fungal biodiversity ( Petrini et al ., 1992 ; Saikkonen et al ., 1998 ; Arnold et al . 2000 , 2003 ). Endophytes in plants can play important ecological roles, e.g. mediating plant defense reactions against pathogens and herbivores or influencing host responses to abiotic stressors such as drought ( Costa Pinto et al ., 2000 ; Arnold et al ., 2003 ; Schardl et al ., 2004 ; Arnold & Engelbrecht, 2007 ; Mejia et al ., 2008 ; Estrada et al ., 2013 ). However, some endophytic fungi have proven to be latent pathogens of plant hosts. Furthermore, the role of some endophytes in host plants is still unclear ( Mirabolfathy, 2013 ).

Previous studies have shown that the diversity, abundance, and species composition of endophytic fungi can be highly affected by the locality in which a specific plant occurs ( Carroll & Carroll, 1978 ; Petrini et al ., 1982 ; Bills & Polishook, 1992; Fisher et al ., 1994 ; Hata & Futai, 1996 ; Bayman et al ., 1998 ; Arnold, 2001 ; Higgins et al ., 2007 ). At larger geographical scales, diversity of endophytic fungi varies due to latitude and annual rainfall ( Arnold & Lutzoni, 2007 ), although the impact of co-varying factors, such as plant diversity, remains to be studied. Similarly, due to history of land use, plantation, and other factors, species diversity of endophytes differs at small scales ( Gamboa & Bayman, 2001 ). Further­more, the endophyte composition differs because of localities ( Fisher et al ., 1995 ; Frohlich & Hyde, 1999 ; Arnold et al ., 2003 ). For example, Arnold et al . (2003) reported a distinctive endophytic composition associated with Theobroma cacao at different sites in Panama. Many studies investigating host associations of endophytic fungi have focused on distantly related plants, which grow within the same geographic areas. Contradictory results, however, have been reported about the predominance of host specificity ( Sieber 1989 ; Suryanarayanan & Kumaresan 2000 ; 2005 ; Arnold et al ., 2000 ; Cannon & Simmons, 2002 ; Mohali et al ., 2005 ; Higgins et al ., 2007 ).

Quercus macranthera Fisch. & C.A. Mey ex Hohen (black oak) and Q. brantii Lindl. which are the most common plant species in Iran have never been investigated before in term of composition of cultivable fungal endophytic populations. Therefore, the aim of this research was to characterize fungal endophytic communities of barks in Q. macranthera and Q. brantii and understand if the plant host species or the geographical sites of growth are responsible for shaping the culturable fungal endophytic populations.

A total of 94 fungal isolates comprising of 30 species were isolated from Q. macranthera and Q. brantii (Table 1). The majority of identified fungal species (29 species) belonged to the phylum Ascomycota, besides one basidiomyceteous isolate, Phlebia radiata (Table S2 [suppl.]). At least one representative of each taxonomic group (identified based on preliminary morphological features) was subjected to molecular identification based on ITS-rDNA or TUB sequence analysis. This allowed the placement of our sequenced isolates into ten orders (Pleosporales, Xylariales, Hypocreales, Sordariales, Diaporthales, Botryosphaeriales, Trichosphaeriales, Eu­ro­tiales, Pezi­za­les and Polyporales), which belon­ged to 30 species (Fig. 1 and 2).

In phylogeny analysis, ITS-rDNA dataset (except Fusarium spp.) included 98 different in-group taxa and Ganoderma tornatum (CBS 109679) as the out-group taxon. The final single locus dataset comprised 972 characters (including alignment gaps), of which 635 characters were unique site patterns. MrModelTest v. 2.3 software recommended general time reversible (GTR) substitution as the best evolutionary model with gamma distribution, invariable sites and Dirichlet base frequencies. Bayesian inference of ITS-rDNA region resided our strains in 26 species, with the highest posterior probability (Fig. 1).

Beta -tubulin dataset for the phylogenetic analysis of Fusarium spp. consisted of 23 in-group taxa, Peni­cillium araracuarense (CBS 113149) as out-group taxon, and a total of 731 characters including 332 uni­que site patterns. MrModelTest v. 2.3 software selec­ted Hasegawa-Kishino-Yano (HKY) substitution mo­del as the best evolutionary model with gamma dis­tribution and Dirichlet base frequencies. Based on the results, the identity of our strains was determined as F. avenaceum , F. oxysporum , F. solani and F. proliferatum (Fig. 2).

In this study, across the seven sampling counties, the numbers of 94 fungal isolates were recovered from both Q. macranthera (70 strains) and Q. brantii (24 strains) (Table 1 and Table S2). Chi-square analysis showed this frequency is significantly different between both hosts (Tables 2, 3). Proportional to the numbers of isolates, the most species diversity (24 taxa) was found among fungal community obtained from Q. macranthera (Table 4, Fig. 3) further corroborated by higher species diversity indices of Shannon–Wiener index (Hʹ) and Margalef richness (D _marg ) (Table 4). On the contrary, evenness (E _H ) index for fungal community recovered from Q. macranthera was lower than Q. brantii (Table 4). It showed that the frequency of some taxa was higher among fungal community recovered from Q. macranthera (Tables 1 and 4, Fig. 3). Generally, these results highlight that barks of Q. macranthera is probably more preferable to be colonized by endophytic fungi than barks of Q. brantii .

Between provinces, 70 isolates were recovered from East Azerbaijan (67 isolates from Q. macranthera and 3 isolates from Q. brantii ) and 24 isolates from Lorestan (3 isolates from Q. macranthera and 21 isolates from Q. brantii ). This observation was further corroborated using Chi-squared analysis, so that a significant difference was detected between provinces and even counties in terms of the numbers of fungal strains recovered from Q. macranthera and Q. brantii besides Veisian county (Table 3). In scale of counties, the highest frequency of isolates was found in Kaleibar and Hatam-baig counties in East Azerbaijan and followed by Veisian in Lorestan province (Table 4). Moreover, the highest species diversity was also detected in fungal community of Hatam-baig and Kaleibar in East Azerbaijan and followed by Veisian further confirmed by biodiversity indices (Table 4). This highlights the significant effect of growth location on the frequency and diversity of fungal endophyte community of Quercus in Iran.

Overall, the observations of this study suggest that both plant species and plant growth location are involved in distribution and diversity of fungal endophytic communities of Quercus in Iran. It appears that higher frequency and diversity of fungal endophy­tic community on Q. macranthera in East Azerbaijan is probably due to either old establishment of Q. macranthera in East Azerbaijan or more favorable atmospheric condition of East Azerbaijan (mountai­nous and temperate climate) for establishment of this plant species and fungal communities. In the present study all samplings and isolations were made during summer 2014, thus, differences between isolation frequencies cannot be due to date of sampling. Giauque and Haw­kes (2013) have examined the relative importance of environmental and spatial factors in structuring endophyte communities of Panicum hallii Vasey and P. virgatum L. They concluded that environmental factors related to historical and current precipitation were the most important predictors of endophyte communities. In a survey of endophytic fungal com­munities in leaves of Metrosideros polymorpha Gaudich. across wide environmental gradients in Hawai­ian landscape, among-site variation in endophyte community composition was found to be correlated strongly with temperature and rainfall ( Zimmerman & Vitousek, 2012 ).

The most frequent fungal species recovered from across the counties were Biscogniauxia mediterranea , Alternaria alternata , Trichothecium roseum , Sorda­ria sibutii and Paecilomyces formosus . Biscogniauxia mediterranea had the highest relative frequency (18.28%) recovered from Q. macranthera and Q. brantii in all counties (Table 1). This fungus has been shown to be a latent pathogen, with potential to cause major losses to oak industry in Iran ( Mirabolfathy, 2013 ). Al­ter­naria alternata which is frequently identified as endophyte ( Ragazzi et al ., 2001 ; Selim et al ., 2011 ; Maheswari & Rajagopal, 2013 ; Nalini et al ., 2014 ) was the second most frequent endophyte and followed by Trichothecium roseum . Different endophytic fungal taxa showed different relative frequencies in two oak species (or different locations). Quercus macranthera yielded the greater fungal diversity, with 24 different taxa being isolated (Tables 1, 4 and Table S2). Some of the endophytic species were found in only one host species, some are cosmopolitan, not specific to oak and some are rarely found. It shows that these fungal taxa could either restrict only to those counties or may have spread recently across those counties. For example, the only isolate of Curvularia neergardii came from Q. brantii . Furthermore, three species of Daldi­nia with a relative frequency of 4.28% were only obtained from Q. macranthera . The composition and abundance of the endophytes varied according to the host tested. Although the data may indicate that, some of fungal endophytes dominate in mycobiota of Quercus spp., whether it is a result of natural selection or not, awaits detailed investigations.

To the best of our knowledge all of the species identified in this study, except B. mediterranea ( Dava­ri et al ., 2003 ; Mirabolfathy, 2013 ), are reported for the first time from Q. macranthera and Q. brantii . Recently, Hajizadeh et al ., (2015) have studied spe­cies diversi­ty of fungal endophytes of Q. brantii in Kurdistan province, Iran. They reported Cladospo­rium tenellum , Paecilomyces formosus , Petriella guttulata , Preussia australis , and Sordaria sibutii . This is the first report of Pyronema domesticum and Valsa persoonii for the mycobiota of Iran. To the best of our knowled­ge, this is the first survey of cultivable endophytic fungal community of Q. macranthera . Several investigations have been conducted regarding fungal endophytes of different oak species ( Ragazzi et al ., 2001 ; Anselmi et al ., 2004 ; Kwasna et al ., 2016 ). Kwasna et al . (2016) characterized root fungal endophytes of Q. rubor . They identified a more diverse fungal species including 126 taxa (Zygomycota, Ascomycota and Basidiomycota), and number of species was higher in roots subjected to floods. It seems that the studied tis­sue (root) had an effect on species diversity of isolated endophytes. In 2001, endophytes of current-year twigs, buds and leaves of Q. cerris were investigated and the results revealed organ specificity for endophytic fungi ( Ragazzi et al ., 2001 ).

In the assemblage of endophytic fungi in healthy tissues of oak trees, some of them may be possible latent pathogens of oak. Our data revealed a low proportion of strains of oak phytopathogenic fungi. However, B. mediterranea and Ph. radiata , usually associated with oak decline were isolated ( Boddy & Rayner, 1983 ; Mirabolfathy, 2013 ). Biscogniauxia mediterranea is mainly related to charcoal dise­ase ( Mirabolfathy, 2013 ). Interestingly, no wood-decaying basidiomycetes associated with oak trees were recovered in East Azerbaijan province. Some of the recovered genera in this study have previously been reported as potential biocontrol agents, which draws attention to further clarification of their antimicrobial properties ( Gonzalez & Tello, 2011 ). Of those, several species belonging to genera such as Chaetomium ( Ch. globosum ), Epicoccum ( E. nigrum ) and Fusarium ( F. proliferatum ) have been here obtained. Neoscytalidium dimidiatum was only isolated from Quercus brantii in this study. Bakhshizadeh et al ., (2014) have reported N. dimidiatum as a human pathogen from Iran. This highlights that further investigations are needed to fully elucidate the ecology and putative use of wood-inhabiting endophytes.

Although endophytic fungi are known to be ubiquitously distributed in terrestrial plants and the plant itself benefits from these hidden inhabitants as they modulate host nutrition, metabolites, and stress response ( Yuan et al ., 2010 ; Soltani et al ., 2016 ), only recently, intense research efforts have been sought to build a more detailed understanding of biodiversity and bioprospecting of endophytic fungi ( Aly et al ., 2010 ; Soltani et al ., 2016 ). Herein, we focused on cul­tivable fungal species, however uncultivable strains could be a big portion of endophytic fungal commu­nity. Since those strains could be, for example the candidate fungi for production bioactive molecules ( Tejesvi et al ., 2011 ), future surveys should focus on me­ta­genomics and transcriptomics approaches to study the functional role of those hidden members of the microbial population.

The frequency and diversity of fungal community recovered from Q. macranthera and East Azerbaijan province was far higher than Q. brantii and Lorestan province respectively. Accordingly, our data and analyses demonstrate that both oak species and growth locations play a prominent role in shaping the frequency and diversity of fungal endophyte community of Quercus in Iran.

The authors would like to thank the Research Deputy of the University of Tabriz for financial support.