Diversity and distribution of the endophytic fungal community in eucalyptus leaves

1 Departamento de Medicina Universidade Federal de Viçosa, Viçosa, Minas Gerais– 36570-900 Brazil. 2 Coordenação Geral de Ensino e Extensão Instituto Federal do Triângulo Mineiro, Ituiutaba, Minas Gerais – 38035-200 Brazil. 3 Departamento de Ciências Básicas da Vida, Universidade Federal de Juiz de Fora, Governador Valadares, Minas Gerais 35020-220, Brazil. 4 Departamento de Microbiologia/Instituto de Biotecnologia Aplicada à Agropecuária, Universidade Federal de Viçosa, Viçosa, MG 36570-900, Brazil. 5 Departamento de Solos Universidade Federal de Viçosa, Viçosa, Minas Gerais– 36570-000 Brazil. 6 Bolsista do CNPq, Conselho Nacional de Desenvolvimento Científico e Tecnológico, Brasília, DF, Brazil .


INTRODUCTION
Eucalyptus are the most widely used trees in planted forests, due to their growth characteristics, their adaptability to climate conditions and distinct soil types, and the value of their wood (FAO, 2015).Eucalyptus wood can be used in several ways, including the production of paper and cellulose, panels and mechanically processed wood, and in the metalworking industry as plant charcoal (IBA, 2013).Additionally, Eucalyptus may also provide a profitable source of lignocellulose for energy production and advanced biofuels (Rockwood et al. 2008).In Brazil, Eucalyptus grandis x Eucalyptus urophylla hybrid ("urograndis" eucalyptus) plants exhibit uniform growth and high cellulose production, characteristics that have driven the growth of planted forests since the 1990s (Iglesias -Trabado and Wilstermann, 2008).Plant shoots are a common habitat for various microorganisms (Vandenkoornhuyse et al., 2015), and interactions with these microorganisms are important in maintaining the equilibrium of the biogeochemical cycles, gas flows and other determinant processes in ecosystems (Lindow and Brandl, 2003).Endophytes can confer beneficial effects on the plant: protection against invading pathogens and herbivores, or via antibiosis or induced resistance and plant growth.They can still confer the host plant greater tolerance to salinity and drought (Hardoim et al., 2015).Thus, the agronomic and environmental significance of the microorganisms that inhabit plant shoots may be reflected in the adaptability of plant populations and also in crop quality and productivity (Turner et al. 2013).On the global scale, fungal diversity is greater in tropical forests, where terrestrial plant diversity is also greater; however, the true scale of associated endophytic diversity is still not well known (Luo et al., 2014).
Endophytic fungi are present on terrestrial plants and are especially abundant and diverse on the leaves of several tropical (Arnold, 2005) and subtropical trees as compared to other climate zones (Banerjee, 2011).However, multiple patterns have already been reported (Zhang and Yao, 2015), which means that fungal diversity patterns in plants are complex.These microorganisms are considered important components of global biodiversity (Arnold, 2005).The endophytic fungi may affect plant adaptability and evolution in their environment (Vandenkoornhuyse et al., 2015).
Characterizing the microbial community associated with eucalyptus plants in Brazilian commercial forests, in addition to providing a greater understanding of plant-microorganism interactions, is extremely important for maximizing the productivity and optimizing the management of crops that are significant to the Brazilian economy.The use of culture-dependent methods in diversity studies allows for the assessment of only a small fraction of this diversity (Torsvik and Ovreas, 2002).Microbial diversity can also be studied by analyzing the total DNA of the microbiota extracted directly from the plant for processing using electrophoretic techniques (Oliveira et al., 2013).Studies using these techniques have contributed to a better understanding of the microbial community structure and have led to new perspectives and advances in ecological studies (Hoshino and Matsumoto, 2007;Oliveira et al., 2013;Da Silva et al., 2014).Thus, the present study aimed at evaluating the composition and diversity of the endophytic fungal community in eucalyptus leaves at the onset of the rainy period and during the rainy and dry periods.

Study sites
The tw o study sites are forests belonging to the Celulose Nipo-Brasileira company (CENIBRA) planted w ith Eucalyptus "urograndis" located in the municipalities of Catas Altas (Site 1 -Catas Altas I Project) and Santa Bárbara (Site 2 -Serra do Pinho Project), Minas Gerais, Brazil.The forest in Catas Altas is currently in the seventh farming cycle, first implemented in December 1970, and the forest in Santa Bárbara is currently in the third farming cycle, first implemented in December 1989.The mean productivity at the tw o sites is 340 m 3 ha -1 /six-year rotation (Table 1), and the soils are highly w eathered, w ith an active, undulating, and strongly undulating relief and w ith yellow Oxisol as the most representative soil class.
The natural vegetation of these sites w as a semideciduous seasonal forest, w hich w as replaced w ith coffee crops and pastures.Subsequently, those crops w ere replaced w ith eucalyptus (CENIBRA).The plants w ere 18 months old in Catas Altas (Site 1) and 72 months old in Santa Bárbara (Site 2) at the time of sampling.
At both sites, the climate type, according to the Köppen classification, is the mesothermal Cw a (Köppen, 1948), w ith a dry w inter and rain in the summer (Table 2).The mean maximum, average and minimum temperatures in recent years w ere 26.4,16.9 and 21.6°C, respectively; the relative humidity w as 67%, and  the rainfall and w ater deficit w ere 122.9 and 15.3 mm, respectively (CENIBRA, Gaspar meteorological station.

Sam pling
Leaves w ere sampled from selected plants that w ere 18 (Catas Altas) and 72 (Santa Bárbara) months old and had average height of 6.0 and 18.0 m, respectively.The trees w ere localized in a subarea of 81 m 2 (approximately 8 trees), w ith a spacing of 3.33 x 3 m betw een trees.After the tree harvest, the leaves in three regions [upper (UPP), middle (MID) and low er (LOW)] of the canopy w ere sampled.To ensure a more representative sample from the w hole canopy, leaves w ere collected from proximal, median and distal parts of the stem in each region (in triplicate and mixed on composite samples).The samples w ere collected at the onset of the rainy period-ORP (October, 2011), during the rainy period-RP (December 2012), and during the dry period-DP (April 2012).Approximately, 50 leaves w ithout symptoms of disease (150 g) w ere collected separately from the upper, middle and low er thirds of the canopy.The samples w ere placed in a box containing ice for transport to the Laboratory of Microbial Ecology (Laboratório de Ecologia Microbiana -LEM) of the Department of Microbiology (Departamento de Microbiologia), Federal University of Viçosa (Universidade Federal de Viçosa), Minas Gerais, Brazil.In the LEM, the samples w ere stored and vacuum-packed at -20°C for approximately one month before being processed for a diversity analysis of their filamentous fungi using independent cultivation methodology.

Surface sterilization of the eucalyptus leaves
Surface sterilization of eight healthy leaves from each third of the sampled tree crow ns w as performed after the leaves w ere w ashed under running w ater and distilled w ater.Next, the material w as immersed tw ice in distilled w ater and phosphate buffer (0.05 mmol•L -1 ), pH 7, immersed in 70% ethanol (v/v) for one minute, kept in a container filled w ith sodium hypochlorite (5%) + 0.05% (v/v) Tw een-80 for five minutes, and then immersed for 30 s in 70% ethanol (v/v) before being immersed again in sodium hypochlorite + Tw een-80 for 15 min (Miguel et al., 2013(Miguel et al., , 2016;;Oliveira et al., 2013).This process w as repeated once.This sterilization/w ashing w as performed to reduce the surface DNA.Finally, the leaves w ere placed in sterile distilled w ater and individually placed into tubes containing 10 mL of R2A culture medium (Reasoner and Gelrdreich, 1985) and incubated at 28°C for 72 h.Aliquots (100 µL) of the final w ash w ater w ere transferred into Petri dishes containing agar-R2A, w hich w ere then incubated at 28°C for 72 h to demonstrate the absence of microbial grow th (Oliveira et al., 2013).

Metagenom ic DNA extraction from leaves
Leaves sampled from each third of the crow ns w ere surface sterilized, incubated in R2A medium, ground in liquid nitrogen, macerated, and transferred into 2.0-mL polypropylene or microcentrifuge tubes.Extraction buffer [(2% (p/v) cetyl trimethyl ammonium bromide (CTAB), 1.4 mol•L −1 NaCl, 20 mmol•L −1 EDTA, 100 mmol•L −1 Tris-HCl, pH 8.0, and 1 g of polyvinylpyrrolidone, and 0.2% (v/v) β-mercaptoethanol] w as added to the tubes containing the ground samples.Next, 1000 µL of extraction buffer, 0.5 g of 106 µm beads, 50 µL of 4% sodium dodecyl sulfate (SDS) and 400 µL of phenol-chloroform (1:24) w ere added to the tubes.The mixture w as stirred in a homogenizer for 10 min and placed in a w ater bath at 60°C for 10 min.The tubes w ere centrifuged at 15,000 g, and the supernatant w as transferred into tubes containing 400 µL of phenolchloroform, follow ed by an additional centrifugation at 15,000 g for 5 min.The DNA w as precipitated by mixing the supernatant w ith 0.6 volumes of isopropanol, follow ed by centrifugation at 15,000 g for 20 min.The DNA pellet w as w ashed tw ice in 70% ethanol and resuspended in 100 µL of sterile Milli-Q w ater after drying under a laminar flow hood.The concentration and purity of the extracted DNA w ere confirmed via optical density at 260 and 280 nm (NanoDrop® ND-1000, Thermo Fisher Scientific, Inc.).

Analysis of the endophytic fungal diversity
Denaturing gradient gel electrophoresis (DGGE) and nested PCR w ere used to examine the endophytic fungal diversity of leaves.In the first PCR, the total DNA w as used as a template to amplify the V1-V9 region of the fungal 18S rRNA gene.The oligonucleotide primers NS1 (May et al., 2001) andEF3 (Oros-Sichler et al., 2006) w ere used for the first reaction.The resulting fragments w ere used as templates for a second PCR, and the V7-V8 region w as amplified using the primers FF390 and FR1GC (Vainio and Hantula 2000), w hich contain a GC clamp incorporated into the oligonucleotide's forw ard region.
The first PCR w as performed in a final volume of 25 μL, containing 5 μL of GoTaq Flex® Reaction Buffer, 200 μM dNTPs, 2.0 U of GoTaq Flex DNA polymerase, 3.0 mM magnesium chloride, 0.16 μM NS1 primer, 0.16 μM EF3 primer, approximately 50 ng of total DNA and sterile deionized w ater (Milli-Q).The amplifications w ere performed under the follow ing conditions: initial denaturation at 94°C for 4 min, follow ed by 35 cycles of denaturation at 94°C for 1 min, annealing at 47°C for 1 min, and extension at 72° C for 2 min, and a final extension at 72°C for 10 min.The second PCR w as performed using 1.0 μL of the first reaction as the template, and the primers NS1 and EF3 w ere replaced w ith the pair FF390/FR1GC.The reaction conditions w ere as follow s: initial denaturation at 94°C for 4 min, follow ed by 35 cycles of denaturation at 94°C for 1 min, annealing at 50°C for 1 min, and extension for 1 min at 72°C, and a final extension at 72°C for 10 min.
To analyze the endophytic fungal community, individual bands that show ed better signal under UV light (300 nm) w ere excised from the polyacrilamide gels, eluted into polypropylene tubes containing 30 μL of sterile Milli-Q w ater, and kept overnight at 4°C.A 7-μL aliquot of the eluate from each band w as used as a template for PCR w ith the oligonucleotide primers FF390 and FR1 (w ithout GC clamp).The 132 amplicons obtained from PCRs w as visualized on an agarose gel (0.8% w /v) stained w ith Gel Red® 1000X, and images w ere obtained using L-pix Chemi (Loccus Biotechnology, São Paulo, São Paulo, Brazil).The 65 amplicons obtained from these reactions containing 100 ng/μL w ere sequenced by Macrogen, Inc. Korea, and the sequences obtained w ere compared w ith those available in the GenBank database (NCBI).For each sequence, an identity search w as performed w ith the BLASTn algorithm (Basic Local Alignment Search Tool) (http://w w w .ncbi.nlm.nih.gov/BLAST) for nucleotides (Altschul et al. 1990).The sequences reported in this study have been submitted to GenBank under the accession numbers KU663411 to KU663476.
The DGGE band profiles w ere compared using BioNumerics® softw are, version 7.1 (Applied Maths, Kortrijk, Belgium).The fungal richness variable w as estimated using the program based on a binary matrix, in w hich the presence of one band corresponding to an operational taxonomic unit (OTU) w as encoded as one (1) and its absence as zero (0).The structure of this community w as evaluated based on the Dice similarity coefficient and the unw eighted pair group method w ith arithmetic mean (UPGMA).The richness and diversity analyses w ere performed using the softw are PAST (Hammer et al., 2001), in w hich diversity is estimated using the Shannon index, and statistical analyses w ere performed in Minitab version 15 (Minitab, 2006) (Minitab Inc., State College, Pennsylvania, USA) using Tukey's test at 5% probability.The correlation of the endophytic fungal communities in Catas Altas and Santa Bárbara at the onset of the rainy period, during the rainy period and the dry period determined by DGGE w as determined using Principal Component Analysis (PCA) in Canoco softw are (version 4.5, Biometris, Wageningen, Netherlands).Rarefaction curves w ere calculated using Analytic Rarefaction 1.3 softw are (http://strata.uga.edu/software/anRareReadme.html).

Phylogenetic analysis
The obtained sequences after sequencing w ere compared w ith those from the NCBI Nucleotide database using the BLAST algorithm (Altschul et al., 1990).The 18S rRNA sequences that w ere distinct from each other in the database and sharing more than 97% identity w ere imported w ith Mega 6.0 and aligned using ClustalW.
The alignments w ere manually adjusted, and a phylogenetic analysis w as performed using the neighbor-joining method (Saitou and Nei, 1987).The phylogenetic distance w as computed using the p-distance method, and the robustness of the resulting trees and Miguel et al. 95 the statistical significance levels of the interior nodes w ere obtained by bootstrap analysis w ith 1000 replicates, and the values above 50% were show n.

RESULTS
The protocol for DNA extraction and 18S rRNA gene amplification resulted in amplicons with distinct electrophoretic migration patterns in DGGE, allowing the evaluation of the endophytic fungal diversity in the leaves of eucalyptus (Figure 1).The electrophoretic patterns obtained by DGGE showed more intense bands in the same relative positions (same location in the gels) and OTUs distincts were detected in the leaves analyzed.The presence of lower-and higher-intensity OTUs indicates that the nested PCR provided the resolution necessary for the diversity analyses (Figure 1).This resolution (Carmona et al., 2012) was interpreted as a single band after electrophoresis on acrylamide gel.The DNA fragments in the bands excised from different positions in the DGGE gel were identified as belonging to the phyla Basidiomycota and Ascomycota.The band-excision technique was useful in assessing the endophytic fungal diversity of eucalyptus in the present study (Figure 1).
Analysis of the fungal 18S rRNA gene fragments present in the leaves revealed distinct fungal communities with respect to the cultivation sites (Catas Altas and Santa Bárbara).DGGE allowed the detection of differences between the endophytic communities in eucalyptus farmed in Catas Altas and Santa Bárbara (Figures 2 and 3).The comparative analysis between the two areas showed a smaller number co-occurring groups in relation to the analysis of individual areas.The highest bootstrap (98%) corresponded to samples from Santa Bárbara at the top of the canopy at the beginning of the rainy season (Figure 3).
In the eucalyptus leaves collected within the Catas Altas region, UPGMA analysis generated five distinct groups, where the highest similarity value (52.3%) corresponded to the collection performed at the onset of the rainy period in leaves from the lower portion of the tree canopy (Figure 2A).The highest similarity found within the Santa Bárbara region was 55.3% during the dry period, also from the lower portion of the tree canopy (Figure 2B).The occurrence of common OTUs (23 and 22) in the eucalyptus leaves is independent of the sampling period.Other OTUs exhibit distinct distribution profile between the sampling periods, such as a higher incidence of specific OTUs during the rainy period in Catas Altas, whereas this occurred during the dry period at the Santa Bárbara location (Figure 4).The Shannon diversity indices within Catas Altas ranged from 2.56 to 3.02, and the richness indices ranged from 13 to 21 (Table 3).In Santa Bárbara, the variation was smaller, with diversity indices ranging from 2.09 to 2.4 and richness indices ranging from 7.5 to 11.3 (Table 3).Although, there are variations in Shannon diversity index and richness in Catas Altas and Santa Bárbara individually, the difference between them is not significant according to the Tukey test at 5% probability.However, when comparing the averages of these indices between the two areas, Catas Altas shows higher diversity than Santa Bárbara according to the Tukey test at 5% probability (Table 3).The first and second axis of the principal component analysis (PCA) explained 25.1 and 22.6% of the variation in the community of endophytic fungi in Catas Altas and Santa Bárbara, respectively (Figure 5).
The endophytic fungal distribution in eucalyptus leaves in the Catas Altas region differs depending on the position of the leaves in the tree canopy and between the rainy and dry periods.At this site, 14 species were identified; the greatest number of species was found at the onset of the rainy period (Table 4).
The endophytic community of the Santa Bárbara leaves comprises seven species (Table 4), which are mostly the same as those found in Catas Altas.However, Anomoloma albolutescens, Rhodotarzetta rosea and Rhizoctonia solani were exclusive to Santa Bárbara (Table 4).
Although, leaf position and seasonality did not affect the diversity and richness of endophytic fungi (Table 3), these factors affected the endophytic fungal distribution of Catas Altas more than that of Santa Bárbara (Table 4).The highest endophytic prevalence in Catas Altas was found at the onset of the rainy period (Table 4), whereas in Santa Bárbara, it was found during the rainy period (Table 4).At the latter site of the fungal species identified by sequencing, only Boletus rubropunctus was found in more than one third of the tree canopy and during more than one of sampling period (Table 4).In eucalyptus, there are differences in the colonization and persistence Table 3. Richness and diversity of endophytic fungi at the onset of the rainy period, during the rainy period, and the dry period in leaves of the upper, middle and low er thirds of the tree canopy of eucalyptus in 18 and 72-month-old plants grow n at Catas Altas (CA) and Santa Bárbara (SB). of endophytic fungi as a function of seasonality (Table 4), and Basidiomycota is the fungal phylum that predominates in eucalyptus leaves (Table 4).Phylogenetic analysis of the sequences revealed that they all belong to the phyla Basidiomycota and Ascomycota, forming distinct clades (Figure 6).Most of the groupings formed exhibited bootstrap values above 70, which are considered moderate to strong (Schneider, 2007).These findings indicate the robustness of the analysis.In Catas Altas, bands 1 (B1), 4 (B4), B8 (B8), 25 (B25), 62 (B62) and 87 (B87) formed the groupings with the greatest phylogenetic support with bootstrap values greater than 80, with most of them between 98 and 100 (Figure 6).These bootstrap values are considered strong (Schneider, 2007) and indicate the robustness of the phylogenetic analysis.

Onset of the rainy period
The OTUs from the amplicons extracted from the bands that corresponded to leaves collected within the Santa Bárbara region formed groups with the greatest bootstraps for bands 99 (B99) and 101 (B101), whose bootstrap values were 97 and 99, respectively.The OTUs were grouped with high phylogenetic support into two distinct clades, both belonging to the phylum Ascomycota.Phylogenetic tree support is ensured by a value of 99 for the outermost node (Figure 6), although some of the bootstrap values could be considered moderate and low.
The rarefaction curve calculated for the samples from the beginning of the rainy season, rainy season and dry season of Catas Altas and Santa Bárbara tended to reach a plateau, showing that the number of OTUs screened in the fungal community of both areas was   FJ480426.1 sufficient to reveal most of the sequence types within the community and to reasonably describe the diversity of group (Figure 7).

DISCUSSION
The diversity of endophytic fungi in the eucalyptus leaves, as determined by nested PCR and DGGE, demonstrates the appropriateness of this approach in evaluating the endophytic fungal diversity in eucalyptus leaves (Figures 1, 2 and 3).Notably, this method was developed more than 20 years ago (Muyzer et al., 1993) and has been an efficient method for microbial diversity studies in several environments, such as in soil (Bresolin et al., 2010), in plants (Oliveira et al., 2013;Miguel et al., 2016) and in animals (Kittelmann et al., 2012).The different intensities of the bands in the electrophoretic profile of DGGE were interpreted as different community structures.
The DGGE analysis using UPGMA provides current fingerprinting patterns that can be measured quickly (Fromin et al., 2002) and result in dendrograms that graphically show the similarities between samples (Laplante and Derome, 2011).The endophytic fungi were distributed into five distinct groups via UPGMA analysis (Figure 2), where the highest similarity value (52.3%) corresponded to leaves sampled from the lower third of the tree canopy at the onset of the rainy period in the Catas Altas region.This finding indicates changes in the endophytic fungal distribution due to seasonality and leaf position (Figure 5).This change is less pronounced in older leaves from the Santa Bárbara region, where the lowest number of distinct clades was found (Figure 2B).The highest similarity value at the Santa Bárbara site was 55.3%, occurring in the lower third of the tree canopy during the dry period (Figure 2B).
The distribution of most OTUs during the three sampling periods of the eucalyptus leaves (23 and 22) is similar; however, specific OTUs exist, reflecting the differences in endophytic fungal community structure.Additionally, the OTUs also exhibit distinctions in t he rainy and dry periods, such as higher incidence during the rainy period in Catas Altas, in contrast to Santa Bárbara, where the highest incidence occurred during the dry period (Figure 4).The variation in the Shannon diversity indices and the richness indices in Catas Altas, which were between 2.56 and 3.02, and between 13 and 21 (Table 3), respectively, and in Santa Bárbara, where these values were lower, with diversity between 2.09 and 2.4 and richness between 7.5 and 11.3 (Table 3), were interpreted to indicate that the location within the tree canopy and seasonality are not factors that significantly affect diversity (Figure 4).This interpretation is attributed to the fact the average Shannon and richness indices do not differ significantly according to the Tukey test at 5% probability (Table 3).However, when these rates are compared between the fungal communities of Catas Altas and Santa Bárbara, the higher average for Catas Altas indicates that the age of the plants influences the diversity (Table 3).Species diversity is measured in terms of richness and uniformity, and the most common and extensively used index is Shannon-Wiener (H'); typical values range from 1.5 to 3.5 (Gazis and Chaverri, 2010).More diverse communities tend to exhibit more distinct species (Ghimire et al., 2010), which explains the discrepancy in the diversity indices in leaves from the Catas Altas and Santa Bárbara sites (Table 3).
Seasonality, although it did not influence the diversity (Table 3), modulated the distribution of endophytic fungi in Catas Altas more than that in Santa Bárbara, enabling best groups and the distinction between the rainy periods (beginning of the rainy season and the rainy season) and section (Figure 5).This distinction can be attributed to the different species found in each of these areas (Table 4).Seasonality shapes endophytic fungal diversity in eucalyptus, which can be observed based on the presence of common and site-specific OTUs in Catas Altas and Santa Bárbara (Figure 4 and Table 3).Seasonality can also affect the gain, loss, latency, or persistence of a given microbial species in the community (Ghimire et al., 2010).Although, the functional significance of these changes in microbial community structure due to seasonality has not been demonstrated, some authors report that plants are affected by factors such as antagonism among fungi, as well as abiotic variables that can affect the host plant and thus shape the dynamics of the associated microbiota (Ghimire et al. 2010).
Endophytic colonization is usually affected by the ontogeny of the leaves (Arnold and Herre, 2003).
Variations in diversity and abundance observed in this study may be associated with the nutritional and defense properties at each developmental stage of these organs (Sanchez-Azofeifa et al., 2012).In this study, the largest differences in endophytic fungal diversity among the plants from Catas Altas (younger) and Santa Bárbara (older) can be attributed to plant age (Table 3).In addition, other variables, such as cultivation and rotation cycles (Ellouze et al., 2014), nutrient and sugar levels in the leaves, and other characteristics (Lang et al., 2011), may together have affected the differences between the diversity indices at the two sites.Notably, the forest in Catas Altas is currently in the seventh farming cycle, first implemented in December 1970, and the forest in Santa Bárbara is currently in the third farming cycle, first implemented in December 1989.The crop and rotation cycles can affect the fungal community of the soil A B (Ellouze et al., 2014) and, consequently, endophytic colonization, considering that leaves contain many endophytic microorganisms that originate from the soil (Sprent and Defaria, 1988;Hardoim et al., 2008;Van Der Lelie et al., 2009).Additionally, other factors may also contribute to differences in the endophytic communities, such as changes in leaf physiology and the presence of chemical substances, such as phenolic compounds, that can limit the richness of microbial species.As these compounds are natural inhibitors of fungal colonization, especially by representatives of the phylum Ascomycota, including Aspergillus (Banso and Rai, 2008) and Fusarium (Kaur et al., 2011).An equally likely explanation is the simple absence of these taxa in the older plants from Santa Bárbara.
The differences in the distribution of endophytic fungi in the upper, middle and lower thirds of the tree canopy (Table 4) may indicate that endophytic colonization depends on the site of the plant sampled.The species, Pachylepyrium carbonicola and Malassezia restricta (Table 4), which are present in Catas Altas, and Boletus rubropunctus (Table 4), which is present in Santa Bárbara, can occupy multiple micro-habitats within the plants, indicating more generalist behavior (Table 4).Factors such as altitude, moisture content and canopy density, among others, are reported to affect the level of plant infection (Qi et al., 2012).
The most commonly observed endophytic fungi in eucalyptus farmed in Catas Altas and Santa Bárbara were Fusarium solani, Malassezia restricta, Pachylepyrium carbonicola and Boletus rubropunctus (Table 4).The high identity of the sequences obtained with those in the NCBI database (Table 4) was the criterion used to confirm these species as belonging to the phyla Ascomycota and Basidiomycota (Figure 6).Although, many of the species identified are pathogenic to some plants, they were endophytic in the present study.Notably, disinfection of the surfaces of healthy leaves without symptoms of infection was confirmed by the absence of fungal growth in R2A inoculated with the final rinse water.The strong dominance of some fungal groups (Table 4) indicates that they can play a relevant role in plant physiology.Fungal species can produce a wide variety of growth regulators, such as gibberellins (GAs), abscisic acid (ABA), and auxins (IAA) (You et al., 2012), and they can also confer tolerance to adverse biotic and abiotic factors (Hubbard et al., 2014).
Endophytic microorganisms can colonize plants via wounds at lateral root emergence sites (Hallman et al., 1997) or by releasing hydrolytic enzymes (Robl et al., 2013) that allow them to enter and colonize the plants (Hallman et al., 1997).
Fusarium species are most commonly isolated as pathogens from plants at all latitudes (Zakaria and Ning, 2013) rather than as endophytes in tropical plants (Vega et al., 2010;Zakaria and Ning, 2013).Fungi of the genus, Marasmius are recognized by the production of secondary metabolites that inhibit the growth of Escherichia coli (Rosa et al., 2003).This group has already been described as endophytic (Ngieng et al., 2013).However, according to the literature, there have been no reports of endophytism for the species Marasmius alliaceus.The genus, Boletus contains species that are described as endophytic in the leaves of Pinus sp.(Arnold et al., 2007).However, the species, Boletus rubropunctus is reported here as endophytic for the first time.
The presence of endophytic fungi in leaves reported here expands the understanding of endophytic colonization in eucalyptus.The description of endophytic fungal diversity in this important forest species is an important step in accessing this genetic resource in the search for metabolites and processes that can contribute to improving plant development.

Conclusions
DGGE was efficient at assessing the diversity and distribution of endophytic fungi in eucalyptus.Using the DNA fragments in the bands excised from different positions of the DGGE gel was a satisfactory strategy for assessing the endophytic fungal diversity of eucalyptus in the present study.The age of plants affected the diversity of endophytic fungi in Eucalyptus "urograndis".The leaf position and seasonality affected the endophytic fungal distribution of Catas Altas more than that of Santa Bárbara.The phyla Basidiomycota and Ascomycota are predominant components of the endophytic fungal microbiota in eucalyptus.

Figure 1 .
Figure 1.DGGE electrophoretic patterns of the endophytic fungal community extracted from leaves of the low er, middle and upper thirds (in triplicate) of the tree canopy of eucalyptus grow n at distinct sites: (A) 18-month-old trees grow n in the municipality of Catas Altas, (B) 72-month-old trees grow n in the municipality of Santa Bárbara.LOW: leaves from the low er third of the tree canopy; MID: leaves from the middle third of the tree canopy; UPP: leaves from the upper third of the tree canopy.The samplings w ere performed at the onset of the rainy period, during the rainy period, and during the dry period.Letter B combined w ith Arabic numbers indicates the band excision location.The leaf samples collected from the middle part of the crow n at the beginning o f the rainy season and the rainy season from both locations w ere analyzed in duplicate because the amount of DNA extracted from the third sample w as insufficient for analysis.

Figure 2 .
Figure 2. Cluster analysis and similarity indices obtained from the DGGE electrophoretic pattern of the endophytic fungal community extracted from leaves from the low er, middle and upper thirds of the tree canopy of eucalyptus.(A) 18-month-old trees grow n in the municipality of Catas Altas.(B) 72-month-old trees grow n in the municipality of Santa Bárbara.ORP: sampling performed at the onset of the rainy period; RP: sampling performed during the rainy period; DP: sampling performed during the dry period.

Figure 3 .
Figure 3. Cluster analysis and similarity indices obtained from the DGGE electrophoretic patterns of the endophytic fungal samples extracted from leaves of the low er, middle, and upper thirds of the tree canopy of eucalyptus grow n at Catas Altas (CA) (18-month-old trees) and Santa Bárbara (SB) (72month-old trees).ORP: onset of the rainy period; RP: rainy period; DP: dry period.LOW: low er portion of the canopy; MID: middle portion of the canopy; UPP: upper portion of the canopy.

Figure 4 .
Figure 4. OTU distribution by Venn diagram in eucalyptus leaves at the onset of the rainy period, during the rainy period, and during the dry period.(A) OTU distribution in leaves collected at Catas Altas (18month-old trees).(B) OTU distribution in leaves sampled at Santa Bárbara (72-month-old trees).

Figure 5 .
Figure 5. Principal component analysis (PCA) based on PCR-DGGE profiles of the 18S rRNA gene from plants samples from Eucalyptus "urograndis" from (A) Catas Altas and (B) Santa Bárbara at the onset of the rainy period (ORP), during the rainy (RP) and dry periods (DP) in the low er portion of the canopy (LOW); the middle portion of the canopy (MID) and the upper portion of the canopy (UPP).

Table 4 .
Distribution, identity, e-value and NCBI accession number for each endophytic fungal species identified by sequencing the 18S rRNA gene at the onset of the rainy period, during the rainy and dry periods in leaves of the low er, middle and upper thirds of the tree canopy in 18-and 72-month-old eucalyptus plants grow n at Catas Altas (CA) and Santa Bárbara (SB).

Figure 6 .
Figure 6.Phylogenetic tree constructed w ith the neighbor-joining method using fungal 18S rRNA gene sequences identified in the leaves of 18-and 72-month-old eucalyptus plants grow n at Catas Altas and Santa Bárbara, respectively.Bootstrap values above 50% are show n.

Figure 7 .
Figure 7. Rarefaction curves indicating OTUs based on the amplification of 18S rRNA gene diversity.(A) Catas Altas and (B) Santa Bárbara at the onset of the rainy period, during the rainy period, and during the dry period.

Table 1 .
Georeferencing of the sites planted w ith eucalyptus forests, predominant soil class, year of the first planting and mean forest productivity (m 3 ha -1 )/rotation.

Rainy period Dry period Third of the eucalyptus canopy
BaUppercase letters in richness in columns indicate significant differences between means.Uppercase letters in diversity in co lumns indicate significant differences betw een means.The same letters in either richness or diversity in row s indicate no signific ant difference betw een the means.All comparisons used the Tukey test at 5% probability.