Distribution of the endophytic fungi community in leaves of Bauhinia brevipes (Fabaceae)

(Distribution of the endophytic fungi community in leaves of Bauhinia brevipes (Fabaceae)). Endophytic fungi represent large, yet unexplored components of biodiversity. Th is work evaluated the richness and the distribution of endophytes in the leaves of Bauhinia brevipes (Fabaceae). A total of 1110 colonies were recovered from the samples and grouped by their morphological traits into 126 taxa. Th e total number of taxa according to leaf development was: 102 in mature leaves, 93 in recently expanded leaves and 79 for unfolded leaves. Th e major endophyte genera were Phomopsis, followed by Dothiorella, Pestalotiopsis and Acremonium. Th e richness and the isolate numbers of endophytes were not statistically aff ected by leaf region. However, some taxa were leaf-age specifi c; six were isolated only from unfolded leaves, nine from recently expanded leaves and 17 were exclusively found in mature leaves. Th e composition of endophytes varied with leaf region; the similarities (Jaccard’s Index) among the leaf regions of diff erent leaf ages ranged from 0.36 to 0.46, indicating a high spatial variation in the community of endophytic fungi inside the leaves. Th e high richness of endophytes in this host plant highlights a signifi cant contribution of fungi to tropical biodiversity and the need for further research in this area.


Introduction
Microorganisms that asymptomatically inhabit internal plant tissues for at least one period of their life cycle are known as endophytes.Endophytic fungi have been found in every plant species examined to date (Arnold et al. 2000;Arnold et al. 2001;Tejesvi et al. 2007), have been found in all major plant lineages (Arnold et al. 2007), and colonize vegetative and reproductive parts of their hosts (Carroll 1988;Arnold et al. 2003).Endophytes can be transmitted not only vertically through seeds or other vegetative propagules, but also horizontally, via sexual and asexual spores (Wilson 1996), and the role of insects in endophyte dispersal has also been suggested (Devarajan & Suryanarayan, 2006).
Endophytes are considered a ubiquitous and important component of fungal diversity, although understanding the scale of their diversity, host range, and geographic distributions remains unclear (Arnold et al. 2000).Given the high spatio-temporal variation in the composition and structure of endophytic fungi communities (Arnold et al. 2000;Gamboa & Bayman 2001;Toofanee & Dulymamode 2002;Arnold & Herre 2003;Suryanarayanan & Th ennarasan 2004) tropical plants are argued to be hotspots of fungal diversity (Arnold & Lutzoni 2007).
Investigations, such as from Arnold et al. (2000;2001), have highlighted the high diversity of endophytes in moist tropical forests.In contrast, some studies report that dry tropical forests do not support a high diversity of endophytes.Th e low endophyte diversity in dry forests, in comparison to the high diversity in moist tropical forests, is a refl ection of the infl uence of abiotic (annual rainfall, fi re regime) and biotic (herbivore diversity, canopy openness) factors (Murali et al. 2007).Th us, further studies on the diversity of endophytes in tropical forests are needed to clearly estimate how this diversity varies in diff erent tropical biomes.In addition, information on the eff ect of ontogenetic changes and distributions within plants could contribute to the understanding the role of fungal endophytes.
Understanding the distribution, richness and abundance of endophytes and host affi nity over the geographic ranges of their hosts is the key to knowing the ecology and evolutionary context of endophyte-plant associations.Many researchers have highlighted the importance of the endophytes in plant resistance to climatic (Rodriguez et al. 2004) and soil conditions (Zaurov et al. 2001), pathogens (Mejía et al. 2008) and herbivores (Cheplick & Faeth 2009), and such important roles can infl uence the diversity of an ecosystem as a whole.
Although the Brazilian Cerrado (savanna) represents one of the most important areas for biodiversity conservation (Myers et al. 2000), it remains one of the most unexplored biomes in terms of the fungal endophyte community.Th e aim of this study was to evaluate the richness and abundance of endophytic fungi found in the native shrub Bauhinia brevipes, a species we have been studying for the last 15 years (Cornelissen et al. 1997;Fernandes 1998;Cornelissen & Fernandes 2001;Santos et al. 2008).We addressed the following questions. 1) Does endophyte abundance and richness increase with leaf age?Studies have demonstrated that leaf age infl uences the density of endophyte infection in leaves of tropical forest trees (Arnold & Herre 2003).Th erefore, variation in leaf traits during leaf maturation should aff ect plant colonization by endophytes (Duong et al. 2006).2) Does endophyte abundance and richness vary among leaf regions of the host plant?Th e distribution of endophytic fungi in leaves is typically not homogenous (Cannon & Simmons 2002) and if the community of endophytic fungi is spatially structured (Gamboa & Bayman 2001;Arnold & Herre 2003), it should be expected that diff erent leaf parts support different communities and/or richness and abundance.Th e greater colonization of certain leaf parts may be related to more complex anatomical structure and/or susceptibility to infection (Cannon & Simmons 2002).

Material and methods
Th is study was performed at the Estação Ecológica de Pirapitinga (EEP), in southeastern Brazil.Th e EEP is a human-made island formed during the construction of a water reservoir for a hydro-electrical power plant in 1965.Th e reservoir is located near of the city of Três Marias (18°23'S, 45°20'W), at an elevation of 560 m above sea level.Th e average annual temperature of the study site varies from 21 to 25°C and the mean annual precipitation is 1200mm.Th e EEP has an area of 1,100 ha and its vegetation is primarily Cerrado (Gonçalves-Alvim & Fernandes 2001).
Bauhinia brevipes Vog.(Fabaceae) is a deciduous shrub up to 3m tall (Cornelissen et al. 1997) that loses its leaves at the onset of the dry season.At the end of the dry season (September) all individuals are leafl ess.Leaf growth starts synchronously at the onset of the rainy season (October), and leaf maturation may be observed until mid-rainy season (Cornelissen et al. 1997).Bauhinia brevipes is attacked by several free-feeding and galling insects (Cornelissen & Fernandes 2001;Santos et al. 2008).
In order to evaluate the composition of endophyte communities, 15 individuals of B. brevipes were randomly selected and sampled in January 2004.Sampled plants were located at least 20 m from each other in order to avoid identical genets, as growth of ramets belonging to the same genet has been observed (G.W. Fernandes, unpublished data).From each plant, we randomly sampled six leaves belonging to three age classes from around the shrub canopy, two unfolded leaves, two recently expanded leaves, and two mature leaves, for a total of 90 leaves.
Leaves were immediately taken to the laboratory, kept at 4ºC and stored for no longer than 24 h before the analyses.Th e leaf surface was sterilized by immersion in 70% ethanol for 1 min, 3% sodium hypochlorite for 3 min and 70% ethanol for 30 s. Leaves were then rinsed in sterile distilled water for 30 s (Cannon & Simmons 2002).Nine leaf fragments (5 x 3 mm) were excised from each leaf, three from the leaf blade, three from the petiole, and three from the midvein (n= 810), for endophyte assessment.Th e fragments of each leaf region were placed onto PDA (Potato-Dextrose-Agar supplemented with cloramphenicol 100mg.L -1 to inhibit the growth of bacteria) Petri plates and incubated at 25ºC for 5-15 days.Aft er hyphae proliferation (from seven to 15 days) colonies were submitted to microculture on glass slides for examination of reproductive structures (Riddell 1950).For sporulating fungi, we used morphology-based taxonomy (Barnett & Hunter 1972).Th e sterile isolates were separated according to similarity of morphological traits, including aerial mycelium form, colony and medium color, colony surface texture and margin characters.
For molecular identifi cation, the sterile isolates were grown on PDA Petri plates and incubated at 20°C for 10 days.Th e DNA extraction was performed according to De Hoog et al. (2005).Th e primers ITS 1 and ITS 4 were used to amplify the ITS 1 and ITS 2 sequences surrounding the 5.8S-coding sequence of the ribosomal operon.Th ese primers amplify a wide range of fungal targets and work well to analyze DNA isolated from individual organisms (White et al. 1990).PCR was run using a Gene Amp PCR System 2400 thermocycler (5 min at 94° C; 30 cycles of 30 s at 94° C, 30 s at 55° C and 30 s at 72° C, followed by a fi nal 7 min at 72° C).Amplicons were analyzed in a 1% agarose gel stained with ethidium bromide.
The sequencing reaction was carried out by using ET Dye Terminator Kit for MegaBACE (GE Healthcare) combined with the automatic sequencing system Mega-BACE TM 1000, at the Laboratório de Biodiversidade e Evolução Molecular (Instituto de Ciências Biológicas-UFMG).The nucleotide sequencing was analyzed comparatively with the NCBI GenBank database from the BLASTn (http://www.ncbi.nlm.nih.gov/blast).Molecular identification was considered from the sequences that showed the highest similarity (> 97%) with GenBank.The nucleotide sequence data were submitted and registered in GenBank.Some taxa, however, were not possible to determine to species due to the lack of taxonomic and genomic database information.Specimens were deposited in the Collection of Cells and Cultures at UFMG (CCC-UFMG).
Th e relative abundance curves were compared to a theoretical model of taxa abundance, lognormal distribution, through the Kolmogorov-Smirnov test for goodness of fi t (Gamboa & Bayman 2001).Moreover, taxon accumulation curves and bootstrap estimates of total taxa richness were inferred with EstimateS 8.0 (Colwell 2006).
In order to evaluate variation of endophyte abundance and richness among leaf regions within each leaf age, the data were analyzed by the non-parametric test of Kruskal-Wallis with the average ordination (rank mean) compared by Tukey's test (α= 0.05) (Conover 1980;Zar 1996).Th e statistical program used was SigmaStat version 3.5 (Copyright © 2006 Systat Soft ware, Inc.).Th e same procedures were used to evaluate variation in abundance and richness among leaf age.In order to compare the community of endophytes in B. brevipes, the similarity between leaf age and leaf regions was evaluated using Jaccard's index -JI (Magurran 1998): JI = a / (a + b + c) where "a" represents the number of taxa occurring in both samples, "b" represents the number of taxa restricted to sample 1 and "c" represents the number of taxa exclusive to sample 2. JI ranges from 0.0 (no taxa shared) to 1.0 (all taxa shared).A similarity matrix was built using the JI, which was then clustered using UPGMA as the clustering algorithm.

Results
A total of 1110 axenic cultures were obtained from 810 leaf fragments (an average of 74.3 colonies per plant).Fungi abundance conformed to a log-normal distribution (Kolmogorov-Smirnov test D = 0.1163, p < 0.01).Th e taxon accumulation curve for individual plants did not reach an asymptote, suggesting the taxa richness of endophytic fungi was underestimated (Fig. 1).
Endophytic fungi colonized 81.7% of the leaf fragments of 88 (97.8%) leaves from the sampled plants (77.8% in unfolded, 86.3% in recently expanded and 81.1% in leaves).From the 1110 isolates recovered, 653 isolates (58.83%) remained as mycelia sterilia in culture.Th e other 457 isolates did sporulate in culture and were identifi ed into 18 taxa, whereas non-sporulating endophytes were grouped into 108 colonial taxa.Th irty-one taxa were commonly found in the leaves, comprising 60% of total colonies.Ninety-fi ve taxa were rarely encountered (with less than 10 colonies).Th e most common genus was Phomopsis with 378 colonies (around 35% of total colonies) and 12 taxa.Th e genus Dothiorella (36 colonies, about 3.2% of total colonies) was the second most abundant, followed by Pestalotiopsis (25 colonies, 2.3% of the total colonies), taxon 91 (24 colonies, 2.16% of the total colonies), taxon 96 (22 colonies, 1.98% of the total colonies), and Acremonium (20 colonies, 1.8% of the total colonies).Furthermore, Phomopsis sp. 1, was the dominant taxon, with 99 colonies (8.9% of total colonies) isolated from 13 out of the 15 sampled individual plants (Table 1).
Most of the endophytes were found in more than two leaf regions.Th e 31 most abundant endophyte taxa were infl uenced by leaf regions of the same leaf age (p> 0.05).Th e richness of endophytes did not diff er statistically among leaf age or leaf region (p> 0.05).

Discussion
Endophytes were abundant and diverse in the leaves of B. brevipes.Th e community of endophytes fi t a lognormal model of taxon distribution, with few abundant taxa and several rare ones.Th e lognormal distribution may represent a general pattern for tropical fungal assemblages (Gamboa & Bayman 2001).However, it should be pointed out that the community was under-sampled.Th e bootstrap analysis found in all leaf regions.Similarities among the communities of endophytic fungi colonizing B. brevipes leaves were near 0.50.Th e dendrogram produced three well-resolved groups, mature, unfolded and recently-expanded leaves based on leaf age (Fig. 2).Th e analysis also shows high similarity between unfolded and recently expanded leaves.Th e endophytes community colonizing mature leaves displayed the greatest diff erentiation (Fig. 2).
Endophyte abundance was only aff ected by leaf age (p < 0.001).Unfolded leaves supported lower abundance of endophytes compared to recently expanded (p = 0.002), and mature leaves (p < 0.001), though abundance in mature and recently expanded leaves did not diff er statistically (p = 0.373).Otherwise, the abundance of endophytes was not  indicates that, although most of the expected taxa were recovered by sampling 15 B. brevipes individuals, some taxa were not sampled.Th erefore, the lack of an asymptote indicates that saturation in the taxon accumulation curve was not achieved.Th at fact was also observed in other surveys, and normally is explained by the hyper diversity reported in tropical endophytic fungi (Arnold et al. 2000;Arnold et al. 2001;Arnold 2005;Gamboa & Bayman 2001).Th e majority of B. brevipes leaves sampled in this study were colonized by endophytes.Our data corroborate the fi ndings of Arnold et al. (2000) that suggested that the proportion of leaves colonized in tropical hosts might exceed those of temperate hosts.Th is prediction was supported in a recent study by Arnold & Lutzoni (2007) who demonstrated that the percentage of colonized tissue decreases with increasing latitude.
Th e community of B. brevipes endophytes is diverse when compared to temperate species and even to some tropical plants.Th e richness of endophytes in B. brevipes is remarkably greater than those found in grasses (Sanchez-Marquéz et al. 2007) and other tropical species (Rodrigues 1994;Gamboa &Bayman 2001;Toofanee & Dulymamode 2002;Suryanarayanan & Th ennarasan 2004).Indeed, our estimate of endophyte richness is comparable with the assemblages recovered from the most endophyte speciesrich tropical trees (Fröhlich et al. 2000;Arnold et al. 2001) but lower than that reported for Th eobroma cacao (Arnold et al. 2003).
Th e richness of endophytic fungi in B. brevipes is also high or similar when compared to phylogenetic related species.Other Fabaceae species harbor lower richness of endophytes (Pereira et al. 1993;Cannon & Simmons 2002) when compared to B. brevipes, suggesting that phylogeny alone may not explain the high richness of leaf endophytes in B. brevipes.Th e reason for such richness is unknown and further studies about the ecology and the biodiversity of endophytes in the tropics are necessary to clarify our fi ndings.As B. brevipes is a widespread species occurring in the Cerrado and in the seasonally dry tropical forests of Brazil (Caatinga) and Argentina (Chaco) (Vaz & Tozzi 2003), the role of habitat in mediating endophyte species colonization and richness should be addressed in the future.
Nearly 25% of the B. brevipes endophyte taxa were isolated on three or less occasions.Th e community of endophytic fungi in tropical hosts typically comprises many species represented by only a single isolate (Arnold et al. 2000;Arnold & Lutzoni 2007).In B. brevipes, the genus Phomopsis was the most abundant with 12 taxa.Th is is a common genus recovered mainly from tropical plants as an endophyte (Rodrigues 1994;Mariano et al. 1997;Taylor et al. 1999;Gamboa & Bayman 2001;Cannon & Simmons 2002;Silva et al. 2006;Arnold & Lutzoni 2007) or as a pathogen (Moricca 2002).Phomopsis sp. 1 was the unique taxon found for all leaf ages, all leaf regions and in almost every individual sampled, suggesting that this fungus is widespread in B. brevipes leaves in the studied area.Its dominance may be related to its higher potential to colonize B. brevipes leaves, which favor its establishment and growth (Duong et al. 2006) or to negative eff ects on other fungi (Herre et al. 2007).
Several studies have shown that old leaves support more endophytes than relatively younger leaves (Toofanee & Dulymamode 2002;Suryanarayanan & Th ennarasan 2004).Th e higher abundance of endophytic fungi in mature leaves of B. brevipes suggests three non-mutually exclusive hypotheses.Firstly, mature leaves are more favorable for fungal colonization.Changes in leaf biochemistry infl uence endophytic colonization with consequences for endophyte distribution (Fernandes et al. 2011).Moreover, young leaves are chemically more protected from herbivore attack than mature leaves (Coley 1988).Cornelissen & Fernandes (2001) reported that as the leaves of B. brevipes aged, tannin concentration increased and herbivore attack decreased.Th erefore, the increase in density of endophytes with leaf age suggests that tannins may not have the expected toxic eff ects on the fungus community, as previously argued by Taper & Case (1987).Alternatively, endophyte presence could in turn increase tannin concentration.A lack of a negative correlation between endophyte abundance and leaf age should corroborate the fi ndings of Faeth & Hammon (1996) and Arnold & Herre (2003) who have argued that leaf chemistry plays a minor role in endophyte colonization.
Secondly, leaf exposure time may also have accounted for increased density of endophytic fungi (Fröhlich et al. 2000;Toofanee & Dulymamode 2002;Arnold & Herre 2003).Arnold et al. (2003) verifi ed that fungal endophytes associated with leaves of Th eobroma cacao were horizontally transmitted and accumulated over leaf lifetimes.Moreover, the frequency of endophytes transmitted horizontally is also high in tropical regions (Arnold & Herre 2003;Arnold et al. 2003;Arnold 2005).It is likely that the widespread occurrence of Phomopsis, Dothiorella sp., Pestalotiopsis sp. 1 and Pestalotiopsis sp. 2, in all leaf stages and the low similarity among the leaf ages may be related to the exposure time and horizontal transmission, rather than absolute leaf age or chemistry.Finally, mature leaves may have supported higher endophytic abundance probably due to their higher biomass providing more sites and resources for colonization when compared to young leaves (Toofanee & Dulymamode 2002).Detailed future fi eld and experimental studies are called for to test these hypotheses.
Th e composition of endophytic fungi strongly diff ered among leaf ages while the number of colonies was higher in mature leaves.Th is result not only highlights a wide diversity of endophytic fungi that can coexist in the same leaf depending on the phase of its development, but also demonstrates a temporal variation in species replacement in leaves.Toofanee & Dulymamode (2002) suggested that the intrinsic environment of the leaves undergoes constant changes as they mature and thus the leaves support diff erent endophytes.Endophytic fungi have been shown to be tissue-recurrent in several studies (Cannon & Simmons 2002;Duong et al. 2006), but the mechanisms underlying tissue-recurrence are not entirely understood (Hyde & Soytong 2007).Our data show that although leaf region appears not to have aff ected the abundance and richness of endophytic fungi, it caused changes in the composition of the endophytic community.It remains to be investigated whether spatial niche diff erentiation would result in competition avoidance and consequent endophyte coexistence within leaves.

Figure 1 .
Figure 1.Taxon accumulation curves showing the number of Bauhinia brevipes individuals and the richness of endophytic fungi.

Figure 2 .
Figure 2. Dendrogram showing the relationship among the communities of endophytic fungi in the petiole (PET), mid-vein (MID) and leaf blade (BLA) of unfolded (UN), recently expanded (RE) and mature (MA) leaves of Bauhinia brevipes based on the Jaccard Similarity Index.Cophenetic correlation = 0.88.

Table 1 .
Isolate number of the 33 more abundant endophyte taxa (ranked according to abundance) found in diff erent regions (petiole, mid-vein and leaf blade) of the leaves (unfolded, recently expanded and matu re) from Bauhinia brevipes from Estação Ecológica de Pirapitinga, Três Marias, Minas Gerais, Brazil.Th e accession numbers for GenBank (http://www.ncbi.nlm.nih.gov/sites/entrez?db=nucleotide) are in the parentheses.