Seasonality affects the community of endophytic fungi in coconut ( Cocos nucifera ) crop leaves

The diversity of endophytic fungi in healthy coconut palm leaves ( Cocos nucifera ) was assessed by analyzing fungal isolates from three coconut cultivars (yellow dwarf, green dwarf, and a hybrid cultivar PB121) used agriculturally in Brazil. The influence of season (rainy or dry) on the endophytic fungal community was also analyzed. Overall, 318 fungal isolates were obtained from 972 coconut leaf fragments. The rDNA ITS region was sequenced from representative species of the isolated endophytic fungi and the most common species identified were Nigrospora oryzae , Pestalotiopsis sp., and Zasmidium musae . The alpha diversity of the endophytic fungi was also calculated. Nonmetric multidimensional scaling (NMDS) ordination and permutational analysis of variance (PERMANOVA) revealed significant seasonal effects on the composition of the endophyte community. However, practically no influence was observed on the fungal communities for the different cultivars of coconut.


Introdution
Coconut (Cocos nucifera) belongs to the palm family Arecaceae and is one of the most important tropical crops in the world (Lamdande et al. 2018). It is commonly used to produce coconut oil, husk, fiber, and water, but it also has non-food uses in various industries, such as for brick, furniture, and coir production, as well as in the adsorption industry (Roopan & Elango 2015). Additionally, coconuts are important for health promotion and disease prevention (DebMandal & Mandal 2011), and they have been studied for their potential to prevent and treat Alzheimer's disease (Fernando et al. 2015), as well as for their anti-inflammatory activity (Silva 2013).
Endophytic fungi are microorganisms that inhabit plant tissues without causing disease symptoms (Azevedo & Araujo 2007). They are helpful to many different plants by acting as biocontrol agents (Mejía et al. 2008;Larran et al. 2016;Saad et al. 2019;NI Silva et al. 2019), promoters of antimicrobial activity (Malhadas et al. 2017), and contributing to abiotic stress tolerance (Hubbard et al. 2014).
Several new species of endophytic fungi have been described (Bezerra et al. 2017a;Oliveira et al. 2016;Silva et al. 2019a; and many studies have been carried out on the endophytic fungal communities of important crop plants, such as coffee, rice, grapes, and soybeans (Pimentel et al. 2006;Sette et al. 2006;Naik et al. 2009;Lima et al. 2014;Oliveira et al. 2014a;Vaz et al. 2014;Varanda et al. 2016). Initial studies of endophytic fungi in palm species (Arecaceae) were carried out by Rodrigues & Samuels (1990) in Australia, who isolated fungi from the leaves of Licuala ramsayi. Further studies were performed by Taylor et al. (1999), who characterized the endophytic mycobiota of Trachycarpus fortunei; Rodrigues (1994), who isolated endophytic fungi from the Amazonian palm (Euterpe oleracea); and Mariano et al. (1997), who compared the epiphytic and endophytic communities of Cocos nucifera. In addition, other studies on endophytic fungal isolates from palm species have been carried out on the Bermudian palmetto (Sabal bermudana), the Chinese fan palm (Livistona chinensis), and Wallichia caryotoides (Southcott & Johnson 1997;Lumyong et al. 2009). Many studies have already investigated the influence of seasonality on endophytic fungal communities (Arnold & Lutzoni 2007;Mishra et al. 2012;U'Ren et al. 2012;Ek-Ramos et al. 2013;Yadav et al. 2016). However, little is known about the influence of seasonality on endophytic fungi in species of palm crops (such as coconut).
Studies on the composition of endophytic fungal communities, in relation to the influence of seasonality and the differences between plant cultivars of the same species, are poorly understood, especially in palm crops. We hypothesize that the endophytic fungal communities of different cultivars of Cocos nucifera are influenced by both seasonal factors and cultivar type. The aim of the present study was to evaluate the influence of both seasonality and coconut cultivar on the community of endophytic fungi in healthy leaves of plants cultivated in Northeastern Brazil.

Sampling area
The study was conducted in three coconut crops located in the Instituto Agronômico de Pernambuco (IPA) in Goiana, Pernambuco state, Northeastern Brazil (7°33'45" S, 35°0'0" W). The precipitation and temperature for the area are shown in Figure 1.

Leaf sampling
Between May 2012 and April 2013, six leaf collections were performed (three in the dry season, September to March, and three in the rainy season, April to August). Samples were collected from three subareas, each of which corresponded to the location of a specific coconut cultivar. The cultivars used were as follows: 1. Yellow dwarf, 2. Green dwarf, and 3. A hybrid cultivar (PB121). In each subarea, three random points were delimited. Three individuals of coconut were chosen at each point, from which three leaves were collected, giving a total of 27 leaves from each subarea. Eighty-one leaves were obtained per collection from across all of the subareas, giving an overall total of 486 leaves for the study (243 leaves from the dry season and 243 from the rainy season). The samples were processed within a maximum of 24 h from collection.

Isolation of endophytic fungi
In the laboratory, the collected leaves were first washed carefully with running water and soap. Leaf discs (6 mm diameter) were then cut at random from all parts of the leaves. The discs were then decontaminated with 70 % ethanol for 1 min and 2 % sodium hypochlorite (NaOCl) for 2.5 min. Finally, they were washed again with 70 % alcohol for 30 s to remove excess hypochlorite. To complete the sterilization process, the material was rinsed with sterilized distilled water (Araújo et al. 2002).
The leaves from each collection point were mixed to create composite samples. In total, 972 leaf fragments were obtained (equating to 324 fragments per cultivar or 486 fragments per season). Six leaf discs were transferred to each of three Malt Extract Agar (MEA) Petri dishes supplemented with chloramphenicol (50 mg L-1), which were incubated at room temperature (28 ± 2 °C) and observed daily for 15 days to record the development of fungal colonies from each sample. As an aseptic control, 50 µL of the final rinsing water was plated on MEA as evidence of surface disinfection (Pereira et al. 1993). Isolates were first grouped into morphotypes based on morphological characteristics, and then confirmed by DNA sequencing.

Extraction, amplification, and sequencing of rDNA
Biomass was obtained from cultures grown on MEA after 7 days at 25 °C. DNA was extracted using the cetyltrimethylammonium bromide (CTAB) method based on the protocol described by Oliveira et al. (2016). To amplify the internal transcribed spacer (ITS) sequence of nuclear rDNA, the primer pair ITS1/ITS4 (White et al. 1990) were used, following the PCR protocol described by Oliveira et al. (2014b). Sequencing was performed by the Plataforma Multiusuária de Sequenciamento de DNA, Centro de Biociências, Universidade Federal de Pernambuco, Recife, Brazil. The sequences obtained were deposited in the National Center for Biotechnology Information (NCBI), under accession numbers MK508790 to MK508819, and were compared by percentual of identity, with homologous sequences of several fungi taxa (by BLASTn) already deposited in the NCBI (Tab. 1). Species sequences with multiple coincidences or ambiguous results were only classified to the genus level. Cladosporium dominicanum (KJ000287) and Phaeosphaeria nodulispora (KR092904) were identified from previous studies (Oliveira et al. 2014c;Oliveira et al. 2016).

Data analysis
The sampling efficacy was assessed using individualbased rarefaction (interpolation) and extrapolation analyses with 95 % unconditional confidence intervals in the iNEXT R software package (Hsieh et al. 2016). The rarefied and extrapolated alpha diversities of taxa were displayed in relation to the sample unit. The general linear model (GLM) was used to test for significant differences in fungal endophyte diversity (alpha diversity and effective alpha diversity as the exponential of the Shannon index) for each cultivar and season (Jost 2006). A general linear mixed model (GLMM) was used to test for cultivar specificity of endophytic species abundance, with plant species as fixed factors and plots as random factors, using the lme4 software package (Bates et al. 2014). The generalized additive model (GAM) was used to test for the influence of sampling time on endophytic species diversity, using the mgcv R software package (Wood 2017). Post hoc multiple comparison tests were conducted using the agricolae software package with the Bonferroni-adjusted kruskal function (Mendiburu 2016). The diversity and overlap of endophytic taxa across the three cultivars were visualized by assembling a bipartite network, implemented with the bipartite package in R (Dormann et al. 2008).
The endophytic community data were standardized according to the Hellinger method, using the decostand function of the vegan R software package, to avoid variable sampling intensity due to variation in the number of species. A variety of other statistical tests were also performed using various functions of the vegan software package, as described below. The variation in endophytic fungal communities between host cultivars was determined using multivariate analysis (Oksanen et al. 2015), and differences in fungal community composition between the samples were calculated using the Bray-Curtis dissimilarity method. Nonmetric multidimensional scaling (NMDS) was used to graphically visualize the organization of the samples in two-dimensional space using the metaMDS function. To test the overall effects of host identity and seasonality on the endophytic fungal community, permutational multivariate analysis of variance (PERMANOVA) (Anderson 2001;McArdle & Anderson 2001) was performed using the adonis function (9,999 permutations). The turnover of species composition among communities (beta diversity) was evaluated using the betadisper function to perform analyses of multivariate homogeneity of group dispersion (Anderson 2006). Post hoc pairwise PERMANOVA analyses were used to test for significant differences in each of the factors.

Composition of fungal endophytes
In the 972 leaf fragments, 318 fungal endophyte specimens were found, with an isolation frequency of 32.7 %. Overall, 32 species belonging to 14 orders were identified. The rDNA ITS region was sequenced from representatives of each endophytic fungal species isolated from Cocos nucifera, except for Syncephalastrum racemosum and Xylaria sp. (for which it was not possible to amplify the ITS region). The maximum identity of each species was then calculated using the BLASTn program from the NCBI platform (Tab. 1).  (Oliveira et al. 2014c;Oliveira et al. 2016).
-DNA amplification was not successful, and the identification was performed with morphological analysis.

Variation in the diversity of endophytic fungi between cultivars
The abundance of endophytic species differed between the cultivars (X 2 = 7.39, P = 0.02), with the highest number of species found in the PB121 hybrid cultivar (Fig. S1A in  the supplementary material). The most abundant taxa were Cladosporium sp., Pestalotiopsis sp., Nigrospora oryzae, and Zasmidium musae, which presented similar distribution patterns in the different coconut cultivars throughout the sampling period (Fig. S1B in the supplementary material). The alpha diversity and effective number of species (exponential of Shannon index) of endophytic fungi were similar in the three cultivars (X 2 = 0.40, P = 0.81 and X 2 = 0.56, P = 0.75, respectively) (Fig. S2 in the supplementary  material). The endophytic fungi rarefaction curves indicated that the overall number of endophytic fungi was higher in the yellow dwarf cultivar (mean and SE per plot; 2.74 ± 0.39), followed by the green dwarf (2.56 ± 0.33) and the PB121 hybrid (2.5 ± 0.25) (Fig. S3 in the supplementary material). The extrapolation curves suggest that additional sampling may allow the detection of more endophytic fungi.
The alpha diversity of endophytic fungi varied significantly with the sampling season (edf = 5, F = 2.67, P = 0.03). The alpha diversity of endophytic fungi associated with the green dwarf cultivar was lower (X 2 = 11.3, P = 0.04) at the end of both the rainy and dry seasons, while the alpha diversity of the yellow dwarf (X 2 = 3.2, P = 0.66) and PB121 hybrid cultivars (X 2 = 1.53, P = 0.9) did not differ over time (Fig. 2). Only six fungal taxa were common to the three cultivars, with the highest number of shared taxa (eight) found between the yellow and green dwarf cultivars (Fig.  S4 in the supplementary material). According to a network analysis, the PB121 hybrid presented the highest specificity index (0.29), followed by green dwarf (0.25), with yellow dwarf (0.24) being the most generalist host among the coconut cultivars.
Occurrence and community composition of endophytic fungi according to season NMDS ordination and PERMANOVA analyses revealed significant effects of the season on endophytic fungal community composition (Pseudo-F = 4.38, R 2 = 0.08, P = 0.003). The standardized deviation ellipses showed clear separation of the fungal communities between the rainy and dry seasons (Fig. 3). In the dry season, 21 taxa from 10 orders were recorded, compared to 22 species from 11 orders in the rainy season. The greatest number of orders was associated with the yellow dwarf cultivar (12), followed by the green dwarf (11) and PB121 hybrid (10) cultivars. Capnodiales and Tricosphaeriales were the richest orders found in the three cultivars. The order Mucorales occurred only in the yellow dwarf cultivar. The order Amphisphaeriales was mostly present during the rainy season, with the highest abundance at the beginning of the season in all the cultivars. Cystobasidiales was present only in the yellow dwarf and green dwarf cultivars, and only two isolates of this order occurred during the dry season. Conversely, the order Ustilaginales occurred only during the rainy season (Tab. 1).

Discussion
The isolation frequency (32.7 %) of endophytic fungi was similar to that found in other studies of the Arecaceae family. Rodrigues (1994) isolated endophytic fungi from Euterpe oleracea, obtaining a 21-30 % isolation rate, and Southcott & Johnson (1997) reported an isolation frequency of 20.3 % in two species of palm (Sabal bermudana and Livistona chinensis). However, Taylor et al. (1999) and Lumyong et al. (2009) reported higher isolation frequencies of endophytic fungi, of 60.9 % and 68.7 %, from Trachycarpus fortunei and Wallichia caryotoides.
In this study, we found that alpha diversity was similar between the three cultivars, and the most abundant taxa were the same for all the cultivars. The genetic and physiological differences between the coconut cultivars are insufficient to greatly influence the species richness of the endophytic fungal communities. Mariano et al. (1997), also working with Cocos nucifera, similarly reported that very common species were not specific to any cultivar. Pancher et al. (2012) suggested that differences between grape cultivars (Merlot and Chardonnay) may drive only a minor shift in endophyte composition when compared to crop management. In addition, the composition of the endophytic fungal communities was not found to be significantly different between the Syrah, Cabernet Sauvignon, and Aragonez grape cultivars (Varanda et al. 2016). On the other hand, the endophytic communities were found to be statistically different between two cultivars of pepper (at the seedling stage) in open field and greenhouse trials (Halász et al. 2016). In olive cultivars, endophytic fungi richness and diversity are influenced by changes in season, site, and cultivar (Materatski et al. 2019). The reports above suggest that the differences in endophytic fungal community composition between cultivars depend on the host species.
Although the endophytic fungal community composition was not significantly different between cultivars, NMDS ordination and PERMANOVA analyses revealed significant effects of the season on the fungal community composition. This suggests that environmental factors can influence the fungal community more than the choice of cultivar. It is possible that differences in the environmental humidity favor some groups of endophytic fungi, or that, depending on the season, the plant physiology changes to give an advantage to some taxa of fungi over others. In Eugenia jambolana, the diversity of fungal species was different in summer compared to winter (Yadav et al. 2016). According to Martins et al. (2016), seasonal differences in the endophytic fungal communities were detected in the olive tree, but other factors, such as different tissues of the host, are also important. The fungal communities in the Indian medicinal plant Tinospora cordifolia were more strongly affected by season when compared to geographic location (Mishra et al. 2012). Similarly, according to Yadav et al. (2016), the diversity and colonization frequency of endophytic populations are more influenced by factors such as season.
In the present study, some species were detected only in the rainy season, while others were present only in the dry season. The order Cystobasidiales, for example, was present only during the dry season, while Ustilaginales occurred only during the rainy season. According to Sadeghi et al. (2019), the species richness of endophytic fungi in mandarin trees is probably higher in the autumn than in the spring, due to the higher rainfall. Factors such as precipitation and temperature can impact the release of inoculum and the germination of fungal spores, subsequently influencing the establishment of the fungi. Under favorable conditions (depending on the season), a greater number of fungal morphotypes can be isolated from the foliar tissue (Singh et al. 2017).
In this study, we observed that seasonal differences have an influence on the community of endophytic fungi. However, practically no influence was observed on the fungal communities for different cultivars of coconut. Thus, our hypothesis was only partially confirmed. As coconut is an important crop, studying the diversity of endophytic fungi in different cultivars and seasons is important for better understanding the dynamics of these fungal communities. Environmental factors (i.e., the season) can alter the communities of endophytic fungi in coconut leaves. Considering that some of these endophytic fungi can be latent pathogens, knowledge about their ecology, pertaining to differences between seasons and cultivars, can contribute towards future work on phytopathology in order to assist farmers with the management of this plant. However, more studies are necessary to better understand these relationships and how they can benefit crop systems.