Initially, healthy samples of stem barks of H. speciosa were superficially sterilized to eliminate epiphytic microorganisms. The endophytic fungi were isolated from fragments of stem barks (1 cm²) using Petri dishes containing potato dextrose agar (PDA; containing 200 g/L potato, 20 g/L dextrose, and 15 g/L agar, pH 6.0). The whole procedure is comprehensively explained in Santos et al. [11].

Endophytic fungi from stem barks of H. speciosa were identified through macroscopic and microscopic evaluation using the instructions previously reported [25-30]. For this, each colony was grown in PDA, Sabouraud-dextrose agar (SDA), Czapek agar and malt extract agar. Strains showing antimicrobial activity were deposited in the URM Culture Collection (WDCM604) of UFPE. The absolute frequency (f) was estimated as the total number of endophyte isolates, while the relative frequency (rf) was the number of endophytes of each species divided by the total number of endophytic fungi. The rate of colonization was estimated as the total number of fragments of leaves colonized by fungi divided by the total number of fragments used for the isolation of endophytes. The number of isolates in each season was used to determine diversity indices: richness (S), evenness (J′), the Simpson (D′), and Shannon-Wiener (H′) [11].

The antimicrobial activity of endophytic fungi was assayed against test microorganisms provided by the URM Culture Collection and UFPEDA Culture Collection (Department of Antibiotics; UFPE). The following fungi were tested: Candida albicans URM-5825, Candida krusei UFPEDA-1002, Candida tropicalis URM-5871, Malassezia furfur URM-4220, Staphylococcus aureus UFPEDA-02, Bacillus subtilis UFPEDA-86, Escherichia coli UFPEDA-224, Klebsiella pneumoniae UFPEDA-396, Pseudomonas aeruginosa UFPEDA-416, and Proteus mirabilis UFPEDA-767.

All endophytic fungi isolated from stem barks of H. speciosa were subjected to a preliminary agar diffusion assay as described by Dos Santos et al. [12] using Petri dishes previously spread with bacteria (Müller-Hinton agar) and fungi [Sabouraud dextrose agar (SDA) for Candida strains and SDA supplemented with 0.5% (SDA+) olive oil for M. furfur]. After incubation (37°C for 24 h for bacteria; and at 30°C for 48 h for fungi), the antimicrobial activity was assayed by measuring inhibition diameter zones (IDZ).

The endophytic fungi showing the highest antimicrobial activity were cultured in SDA medium at 28 ± 2°C. After 7 days, five blocks (5 mm) were transferred to Erlenmeyer flasks (1000 mL) containing rice or corn semisolid media (media preparation was performed as previously described [12]. Each fungus was statically cultivated at room temperature (28 ± 2°C). After 30 days, methanolic extracts were prepared as described by Dos Santos et al. [12], dissolved in dimethyl sulfoxide (DMSO) and stored at -20°C until use.

Minimum inhibitory concentration (MIC) and minimum microbicide concentration (MMC) were determined by broth microdilution method using specific media (Müller-Hinton agar, SBA or SBA+) as described by Dos Santos et al. [12]. Each extract (in 10% DMSO) was tested in concentrations ranging from 9.76 μg/mL to 5000 μg/mL. Each assay was conducted in triplicate.

In total, 116 isolates of endophytic fungi were obtained from the 180 fragments analyzed (90 in each period of the year) (Table 1). These results correspond to a total frequency of colonization (FC) of 64.4%. Regarding the period of the year, 40 isolates were obtained during the dry season (FC of 44.4%), while 76 were obtained during the rainy season (FC of 84.4%). The isolated endophytic fungi were grouped into 14 species: Trichoderma harzianum (9.5%), Phomopsis archeri (8.7%), Aspergillus niger (6.9%), Penicillium fellutanum (6.9%), Nigrospora sphaerica (6%), and Fusarium solani (5.2%). The occurrence of occasional species was also observed in this study, as represented by Aspergillus flavus, Fusarium lateritium, Tritirachium oryzae, Cladosporium cladosporioides, and Marianea elegans.

Notably, some species of fungi were exclusively isolated during the dry period (C. cladosporioides and M. elegans) or rainy season (F. lateritium, F. solani, P. fellutanum, P. archeri, and T. oryzae). The species P. cava, C. gloeosporioides, and L. theobromae showed higher frequencies of colonization. Only a few species were found more frequently compared to the others [31, 32]. The most frequently occurring species are likely those that have a greater affinity to the host plant and perform essential functions such as protection against pathogens [33, 34]. The genera Aspergillus, Fusarium, Nigrospora, Penicillium, Phomopsis, and Trichoderma isolated in this work are frequently reported as endophytes of tropical and subtropical plants [35].

The lowest colonization rates were observed for C. cladosporioides, M. elegans, and T. oryzae. In general, these species can be classified as occasional, as they were represented by a few isolates [32]. In addition, endophytic isolates that did not develop reproductive structures necessary for identification after a certain culture period were also isolated. These isolates were classified as Mycelia sterilia [12, 32, 36, 37] (Table 1).

We observed that the frequency of isolated endophytes in the rainy season was higher than in the dry period. This results may be related to some aspects such as (i) increase of spore dissemination promoted by rain; (ii) facilitation of spore germination by high humidity [38]. Numerous environmental factors are known to influence the frequency of colonization in a plant, including climatic conditions and soil composition at the collection site [10, 11, 38]. Other factors such as fog, rainfall, and dew also play important roles in endophyte diversity, as they can serve as vehicles for fungal spores that penetrate plant tissues and colonize the plant [39].

The diversity indices of isolated fungal populations in each period are shown in Table 2. The richness (S) values were 1.491 and 1.581 for the rainy season (RS) and dry season (DS), respectively. The RS showed higher values of Shannon-Wiener and evenness indices (H′: 2.506 and J: 0.977) than the DS (H′: 2.097 and J: 0.911). In contrast, the Simpson index was higher during the DS than RS (D = 0.146 and 0.086, respectively). Furthermore, a weak correlation was observed when the number of strains isolated for each species was analyzed (ρ = 0.053); a Sørensen–Dice coefficient of 0.696 was found between both seasons. The richness (S) values were 1.491 and 1.581 for the RS and DS, respectively. The RS showed higher values of Shannon-Wiener and evenness indices (H′: 2.506 and J: 0.977) than the DS (H′: 2.097 and J: 0.911). In contrast, the Simpson index was higher during the DS than the RS (D = 0.146 and 0.086, respectively). These results showed that although different relative frequencies were observed during each season, species diversity was similar.

Endophytic fungi have a higher potential to produce a wide variety of bioactive molecules that belong to various structural classes (such as alkaloids, peptides, steroids, terpenoids, phenols, quinones, and flavonoids). These chemical variety of fungi metabolites is associated with the inhibition of a wide variety of pathogenic microorganisms [7-9, 12, 13].

We examined the antimicrobial action of all 116 endophytic fungi isolated from stem barks of H. speciosa. A group of 39 strains (33.6%) displayed antimicrobial activity against Gram-positive bacteria, Gram-negative bacteria, and pathogenic yeasts under the tested conditions. The IDZ values ranged from 11 to 39 mm (Table 3). In this study, we considered IDZ values smaller than 10 mm as low active, IDZs between 10 and 20 mm moderately active, and IDZs greater than 20 mm as high active [40]. Among the active fungi, 21 were isolated during the RS, while 18 were isolated during the DS. Therefore, the relative frequency of strains with antimicrobial potential was higher in the DS (45%) than in the RS (27.63%) (considering the total number of isolated strains in each season). However, strains isolated in the rainy season had a higher antimicrobial potential because they inhibited the growth of at least two pathogen species.

Regarding the anti-bacterial activity, all strains isolated in the RS showed the capacity to simultaneously inhibit gram-positive and gram-negative bacteria. Bacillus subtilis was inhibited by all strains isolated during the RS, with more than 80% presenting IDZ values larger than 20 mm. Similarly, approximately 76% of the endophytic fungi during this period exhibited IDZ larger values than 20 mm against E. coli. A significant number of endophytic strains were active against P. aeruginosa (67%), K. pneumoniae (62%), S. aureus (52%), and P. mirabilis (43%). In contrast, the bacteria most effectively inhibited by strains isolated in the dry period were P. aeruginosa (38.9%), followed by B. subtilis (33.3%), E. coli, P. mirabilis, S. aureus (22.2%), and K. pneumoniae (11.1%). Finally, only three isolates showed anti-yeast potential. Two strains were effective against C. albicans (A. flavus FHS012 and A. niger FHS059, isolates from the rainy and dry period, respectively) and another strain was active against M. furfur (F. solani FHS105 from the RS). Other studies have also observed the antimicrobial activity of endophytic fungi from the same genera or species of the strains isolated in our work [12, 31, 32, 41, 42]. The isolates designated as M. sterilia were also active against some microorganisms tested, suggesting that antimicrobial potential is independent of the reproductive processes of each isolate, as previously reported [12].

Our results demonstrate the diversity of endophytic fungi in H. speciosa bark and importance of these microorganisms as promising targets for bioactive compounds with antimicrobial potential. The frequency of colonization of endophytes in H. speciosa bark was higher during the rainy season, although diversity was similar in both periods. Crude extracts of the endophytes obtained from fermentation in rice medium showed higher antimicrobial activity. The best results were observed for A. niger FHS058, A. niger FHS061, and F. solani FHS106. Therefore, the isolates described above constitute attractive targets for other studies to optimize the culture conditions for the production of bioactive compounds, as well as their chemical characterization. These compounds may be use directly or as a basis for the synthesis of new agents with antimicrobial activity.