Introduction

Plant resident microbiota comprises epiphytic and endophytic fungi with important functional traits ranging from plant growth promotion to antibiosis to phytopathogens1,2. These fungi are non-pathogenic fungal representatives that grow in symbiotical association with plants3,4,5. Horizontal transmission from the epiphytic to the endophytic system and from the internal system to the plant surface6 has been demonstrated. This is because endophytes can emerge to sporulate during plant senescence or from host tissues7. However, each plant hosts epiphytic and endophytic microbiomes that are still unknown for most plant species. These microbiomes can be important sources of biotechnological solutions for agriculture and medicine due to their functional traits and bioactive compound synthesis8,9.

Some studies have associated the microbiome of medicinal plants with the ability of the plants to synthesize specific metabolites10,11. Studies examining endophytic fungi from Stryphnodendron adstringens, an endemic species of the Cerrado biome, reported that fungal isolates and bark extracts collected from the genus Diaporthe12,13 exhibit antimicrobial activity14,15,16 and cytotoxicity against human cancer cell lines15,17. The analyses of microbiomes in Chinese medicinal plants (Ainsliaea henryi, Dioscorea opposita, Potentilla discolor, Stellera chamaejasme, Ophiopogon japonicus, Juncus effusus var. decipiens, and Rhizoma arisaematis)18,19 showed that each of these plants hosted a specific community of actinomycetes with significantly high and diverse rhizospheric and endophytic Actinobacteria, presenting antimicrobial and antitumor properties20. Based on these data, we decided to study epiphytic and endophytic fungi associated with the medicinal plant Serjania erecta Radlk (Sapindales: Sapindaceae). The Sapindaceae family comprises many plants that produce latex containing saponins in their leaves, roots and/or seeds. Therefore, many species of this family are known for their importance in traditional medicine21. S. erecta is popularly known in Brazil as cipó de cinco folhas, a native plant to the Cerrado, with distribution restricted to areas of this biome, north of the Caatinga and south of the Amazonia22. This species has been reported in the literature for important biological effects such as anti-inflammatory activity23, gastroprotective action24, nematicidal effects25, and antimicrobial effects26. Ethanolic extract of S. erecta leaves and roots can inhibit a range of bacteria and the methanolic extract of the leaves also inhibits the growth and development of insect pests, such as Chrysodeixis includens27,28.

Medicinal plants have a vast microbiome; however, resident species can be affected or selected based on the presence of metabolites that may have bactericidal or fungicidal action11. Studies pinpointed that the microbiome can be influenced by a diverse set of molecules, including: coumarins, glucosinolates, benzoxazinoids, camalexin, and triterpenes29,30. Hence, as medicinal plants contain a range of leaf metabolites such as flavonoids, alkaloids, and essential oils31, we hypothesized low microbial diversity in S. erecta leaves. Freitas et al.31 reported that the internal tissues of S. erecta leaves can be colonized by the biotrophic fungi. But as this work focused on morphoanatomical aspects, our aimed to assess S. erecta leaf colonization by epiphytic and endophytic fungi. We also aimed to estimate the diversity of species forming the endophytic fungal microbiome of this species due to its relevance in traditional medicine.

Despite its great biotechnological potential, no studies have examined S. erecta resident epiphytic and endophytic microbiomes. Therefore, the objectives of this study were to assess the presence of fungi in leaf tissues by scanning electron microscopy (SEM) and identify leaf endophytic fungi using cultivable species isolation and metagenomic method. We aim to contribute to the increasing understanding of plant-microorganism interactions, particularly those involving medicinal species. We considered the perspective of Dantas et al.15, who stated that Cerrado plants constitute a repository of endophytic fungi with the potential to act as a source of new bioactive substances that can be used in several areas. Thus, it is imperative to study the symbiotic microbiome of medicinal species found in Cerrado.

Results

Colonization of epidermal structures by epiphytic fungi

Epiphytic fungi interact symbiotically with S. erecta leaf epidermis structures, exhibiting colonization via horizontal mechanisms, i.e., penetration through the stomata or the base of the glandular and squamous trichomes. This penetration was evidenced in these structures regardless of their location, i.e., in stomata and trichomes located on the rachis (Fig. 1A,B), tooth (Fig. 1C,D), and blade (Fig. 1E,F). We found a high level of hyphae colonization on the entire leaf surface, with the greatest accumulation of hyphae near stomata and trichomes. Spore accumulation was also observed on the blade and spores in the germination process (Fig. 1F).

Figure 1
figure 1

Colonization of S. erecta leaf surface by epiphytic fungi with penetration via the stomata and the base of the glandular and squamous trichomes located on the rachis (A, B), tooth (C, D), and blade (E, F). Black arrows indicate hypha penetration into the stomata, and white arrow indicates a spore noted during the germination process. A and B, 15 µm bar; C, E, and F, 12 µm bar; D, 25 µm bar.

Full size image

Surface electron micrographs of S. erecta epidermis also showed nematodes in the three leaf regions evaluated: rachis, tooth, and blade. However, fungal hyphae were always associated with nematodes (Fig. 2A–D), indicating that epiphytic fungi and nematodes compete to colonize leaf surface of S. erecta. Nematode eggs appear isolated (Fig. 2C) or in large groups (Fig. 2D). In this process, we saw fungi physically trapping nematodes (Fig. 2A,C) for parasitism or predation. This was evident from the presence of hyphae oriented perpendicularly to the eggshell, exerting a nematophagous effect. In addition, some eggs were damaged by the interaction with the hyphae (Fig. 2A,B).

Figure 2
figure 2

Colonization of S. erecta leaf surface by epiphytic fungi and nematodes; fungal hyphae adhered to the surface of the nematodes. * indicates nematodes, and arrows indicate hyphal adhesion and penetration through the eggshells. (A), 15 µm bars; (B), 6 µm bars; (C, D), 20 µm bars.

Full size image

Colonization of leaf tissues by endophytic fungi

Paradermal sections of different S. erecta leaf regions showed endophytic fungi, with hyphae interacting with internal tissues in the rachis, tooth, and blade (Fig. 3A). Fungal spores were seen in glandular regions of the leaf teeth (Fig. 3B). Endophytic fungi were also seen in S. erecta conducting vessels, specifically in the phloem of primary and secondary veins in the leaf rachis (Fig. 3C), where microsclerocytes colonized the internal space of sieve tube elements.

Figure 3
figure 3

Paradermal section showing S. erecta leaf tissue colonization by endophytic fungi. Hyphae colonized internal tissues in the blade (A), and spores colonized glandular cavities in the teeth (B). Longitudinal section showing phloem colonization in the leaf rachis region; black arrows indicate the presence of fungal structures (C). A and B, 20 µm bar; C, 30 µm bar.

Full size image

Isolation and identification of endophytic fungi

A total of nine fungal isolates of the phylum Ascomycota were obtained, and the following fungal species were identified: Colletotrichum gigasporum (one isolate), Diaporthe schini (one isolate), Lasiodiplodia theobromae (four isolates), Macrophomina pseudophaseolina (one isolate), and Nigrospora sphaerica (one isolate) (Figs. 4, 5) (Genbank accesses described in Table 1S). An isolate from the genus Pseudofusicoccum was also found. Although characterization of the species was not possible, this isolate formed a stable and highly sustained cluster with Pseudofusicoccum adansoniae, Pseudofusicoccum violaceum, and Pseudofusicoccum stromaticum.

Figure 4
figure 4

Similarity tree of endophytic fungi species isolated from Serjania erecta leaves. Phylogeny was recovered based on the internal transcribed spacer (ITS) and the TUB2 gene. Numbers on the nodes represent the probability, and the bar at the bottom of the tree represents the genetic distance.

Full size image
Figure 5
figure 5

Endophytic fungi isolated from S. erecta leaves. Colletotrichum gigasporum (A), Diaporthe schini (B), Lasiodiplodia theobromae (C), Macrophomina pseudophaseolina (D), Nigrospora sphaerica (E), and Pseudofusicoccum sp. (F). Cultures observed at 7 days of growth on potato dextrose agar (PDA).

Full size image

Metagenomics of leaf endophytic colonization

The metagenome sequencing of three S. erecta samples generated > 1,081 million paired reads or > 332 Mbp of sequence data. After trimming and filtering, > 1,014 million paired reads remained in the dataset with an average of > 338,000 paired reads per sample. Thus, an average of 6.19% of the reads were excluded (Figure 1S). Quality control resulted in high-quality raw data, with Phred scores ranging from 30–40. The FastQC software showed alterations attributed to the removal of adapters, identifying a low percentage of unique reads, which indicated that these species are commonly found in the samples (Figure 2S). The percentage of single reads was higher in SE-1, followed by SE-3 and SE-2. The rarefaction curve showed that the Shannon diversity index, which quantifies species diversity, reached a plateau at a sequencing depth below 10,000 (Figure 3S). This plateau suggests that the sequencing process effectively captured the full extent of species richness in the samples analyzed. SE-1 had a higher Shannon diversity index than SE-3, and SE-3 had a higher index than SE-2.

A total of 150 ASVs were identified from the 16S and 18S sequencing reads (data deposited in the SRA database, accession PRJNA1110285 and BioSample, accession SAMN41342363, SAMN41342364, SAMN41342365). These ASVs were classified into 90 species. There was a tendency toward a higher number of fungal species (43.53, 48.22, and 48.21% of the sequences observed in SE-1, SE-2, and SE-3, respectively) (Fig. 6A). Bacteria were observed at a lower frequency than fungi, with 38.82, 13.79, and 41.07% of sequences being found in SE-1, SE-2, and SE-3. Archaea were found in low percentages in SE-1 and SE-2 sequences (7.06 and 6.90%, respectively). Metazoan eukaryotes were seen in SE-2 and SE-3 (3.45 and 1.79%, respectively); however, all the sequences showed nematodes of the class Chromadorea, classified as Halicephalobus sp. (Table 2S). The percentage of novelty was higher in SE-2, with 27.59% of the sequences classified as not belonging to any of the known domains. This percentage was 10.59 and 8.93% in SE-1 and SE-3, respectively.

Figure 6
figure 6

Metagenomic study of the endophytic colonization of S. erecta leaves. The three samples (SE-1, SE-2 and SE-3) were analyzed for the relative frequency of sequences attributed to organisms from different known domains (A), genus Archaea (B), Archaea species (C), and bacterial classes (D).

Full size image

Most archaea sequences observed in SE-1 were assigned to the genus Methanobacterium (57.14%), with all the genera and species related to methanogenesis (Fig. 6B,C). All archaea obtained in SE-2 were assigned to unknown organisms belonging to these domains. SE-1 showed relatively more diverse bacterial classes compared to the other samples, but regardless of the sample, the class Alphaproteobacteria predominated (40.63, 50, and 75.00% of the bacteria in SE-1, SE-2, and SE-3, respectively) (Fig. 6D). Sequences from SE-1 and SE-3 were assigned to class Gammaproteobacteria (6.25 and 16.67%, respectively), and the percentage of novelty was high in SE-2 sequences (50%).

Most bacterial sequences in SE-1 could not be assigned to known bacterial genera or species (17.65%) (Fig. 7A,B). The genera Methylobacterium and Sphingomonas were the most frequent (11.76 and 8.92%), followed by Lacticaseibacillus, Lactobacillus, and Lentilactobacillus (5.88% each). Sequences of the genus Methylobacterium were classified as Methylobacterium sp. Sphingomonas were classified as Sphingomonas insulae and Sphingomonas sp. (5.98 and 2.94%, respectively). Lacticaseibacillus were classified as Lacticaseibacillus paracasei and Lactiplantibacillus plantarum, Lactobacillus as Lactobacillus delbrueckii and Lactobacillus helveticus, and Lentilactobacillus as Lentilactobacillus hilgardii and Lentilactobacillus parafarraginis.

Figure 7
figure 7

Metagenomic study of the endophytic colonization of S. erecta leaves. The three samples (SE-1, SE-2 and SE-3) were analyzed for the relative frequency of sequences attributed to bacterial genera and species in SE-1 (A, B), SE-2 (C), and SE-3 (D, E).

Full size image

The microbiome of SE-2 was less biodiverse, with 50% of the bacteria classified as unknown, 25% classified as Novosphingobium sp., and 25% as Atribacter laminatus (Fig. 7C). Most sequences in SE-3 were assigned to the genera Sphingomonas and Methylobacterium (34.79 and 17.75%, respectively), followed by Acinetobacter and Aureimonas (8.70% each) (Fig. 7D). Most Methylobacterium sequences were classified as Methylobacterium sp. (13.40%), but Methylobacterium nodulans sequences were also found (4.35%) (Fig. 7E). Sphingomonas sequences were classified as Sphingomonas sp. (21.74%), Sphingomonas insulae (4.35%), Sphingomonas montanisoli (4.35%), and Sphingomonas phyllosphaerae (4.35%). Acinetobacter species were classified as Acinetobacter lwoffii and Acinetobacter ursingii (4.35% each), Aureimonas as Aureimonas psammosilene and Aureimonas sp. (4.35% each).

SE-1 contained sequences of the species Atribacter laminatus, while SE-2 and SE-3 harbored bacteria of the genus Novosphingobium. Additionally, SE-1 and SE-3 exhibited the presence of the genera Methylobacterium and Sphingomonas, with the presence of the species Sphingomonas insulae being particularly noteworthy in both samples.

Sequences of endophytes of two fungal phyla were found in S. erecta SE-1 and SE-1 (Ascomycota and Basidiomycota). On the other hand, the samples also demonstrated sequences belonging to fungi not classified within the known phyla (3.45%) (Fig. 8A). Ascomycetes were the most frequent fungi, with sequences present in 57.50, 50.00, and 48.28% of the sequences classified in SE-1, SE-2, and SE-3, respectively. The fungal class Dothideomycetes was the most frequently found in the samples (25.00, 25.00, and 37.93% of sequences in SE-1, SE-2, and SE-3, respectively), followed by the classes Malasseziomycetes (15.00, 18.75, and 10.34%), Saccharomycetes (12.50, 6.25, and 3.45%), and Tremellomycetes (5.00, 6.25, and 20.69%) (Fig. 8B).

Figure 8
figure 8

Metagenomic study of the endophytic colonization of S. erecta leaves. The three samples (SE-1, SE-2 and SE-3) were analyzed for the relative frequency of sequences attributed to fungal phyla (A) and classes (B). Fungal genera (C) and species (D) were analyzed in SE-1.

Full size image

Most fungal sequences in SE-1 were assigned to the genera Malassezia (15.79%), Leucosporidium (10.53%), Gomphillus (7.89%), and Candida (7.89%) (Fig. 8C). The genus Malassezia included Malassezia globosa (7.84%), Malassezia restricta (5.20%), and Malassezia furfur (2.75%), and the genus Leucosporidium included Leucosporidium scottii (5.26%) and Leucosporidium sp. (5.26%). The genus Gomphillus included only one species, Gomphillus americanos (7.89%), and the genus Candida included Candida sake (5.26%) and Candida sp. (2.63%) (Fig. 8D).

Most sequences in SE-2 were of the genus Malassezia (18.78%) (Fig. 9A), the only one in this sample represented by more than one species, including Malassezia furfur, Malassezia globosa, and Malassezia restricta (6.26% each) (Fig. 9B). Most sequences in SE-1 were assigned to the genera Derxomyces (14.29%), Didymella (14.29%), and Malassezia (10.71%) (Fig. 9C). Derxomyces sequences included Derxomyces bifurcus (10.64%) and Derxomyces napiformis (3.65%). All Didymella sequences were classified as Didymella bellidis (14.29%). Malassezia sequences included the sequences of the species Malassezia globosa (7.15%) and Malassezia restricta (3.56%) (Fig. 9D).

Figure 9
figure 9

Metagenomic study of the endophytic colonization of S. erecta leaves. The three samples (SE-1, SE-2 and SE-3) were analyzed for the relative frequency of sequences attributed to fungal genera and species in SE-2 (A, B) and SE-3 (C, D).

Full size image

The genus Malassezia was found in the three samples analyzed, represented by the species Malassezia globosa and Malassezia restricta. The genera Leucosporidium and Peniophora were found in SE-1 and SE-2, with sequences classified as Leucosporidium scottii and Leucosporidium sp. in SE-1 and as Leucosporidium scottii in SE-2. The genus Peniophora was represented in these two samples by the species Peniophora versicolor. SE-2 and SE-3 also had species from the genera Didymella, Lenzites, Meyerozyma, Neoascochyta, and Hannaella. The genus Didymella was represented by Didymella bellidis in SE-2 and SE-3, and Lenzites sequences included those of the species Lenzites betulinus. The genera Meyerozyma, Neoascochyta, and Hannaella were represented by the species Meyerozyma guilliermondii, Neoascochyta exitialis, and Hannaella pagnoccae, respectively, in these two samples.

Discussion

Epiphytic fungi interact with S. erecta leaf tissues, penetrating through the stomata or the base of the trichomes and trapping nematodes

Some studies showed that the internal colonization of plant tissues begins with the branching of hypha and penetration through the epidermis32. Evidence of epidermal penetration and prevalence of endophytic fungi in S. erecta leaf tissues indicates the ability of these fungi to survive the effects of metabolites and resistance mechanisms of this plant. Cell or spore transmission through stomata is one of the main routes by which microorganisms access interior plant systems33,34. We observed hyphal penetration through open stomata. The results showed symbiotic relationships between S. erecta and leaf surface epiphytic microorganisms since stomatal immunity against fungal invasion includes chitin-induced stomatal closure35. A study by Kumar et al.36 showed that Colletotrichum gloeosporioides germ tubes and its hyphae were oriented toward open stomata, avoiding closed stomata, with the hyphae entering the leaves through stoma openings. Some studies showed that fungal chitinases can convert chitin into chitosan, a compound that does not interact with immune response-inducing receptors, to counteract stomatal immunity37,38.

Trichome penetration may be another mechanism for the horizontal transmission of fungi in S. erecta. A substantial accumulation of hyphae was evident in the proximity of these structures, with concurrent penetration into their interior. Marques et al.39 also demonstrated that fungal hyphae can use the base of glandular trichomes to penetrate leaf tissues. Data shows that fungal hyphae can use the base, side, or top of trichomes for colonization, spreading to adjacent cells40. Trichome penetration can be asymptomatic or mildly symptomatic38, triggering symbiotic processes. Wang et al.41 showed that Rosa roxburghii exhibit glandular trichome cells with thin cell walls and plasmodesmata with large intercellular spaces. On the other hand, the outer wall of these trichomes seems to exhibit a thinner configuration at the base, as reported by Argyropoulou et al.42, who analyzed secretory trichomes in Lippia citriodora. This may help understand the high concentration of epiphytic hyphae associated with the base of S. erecta leaf glandular trichomes.

We also visualized fungi physically trapping nematodes for parasitism or predation on the epidermal surface of S. erecta. Filamentous fungi capable of capturing free-living nematodes from capture structures are known as nematode-trapping fungi43. According to Jones44, hyphae-hyphae fusions adhered to nematodes form three-dimensional traps. Conversely, similar to hyphae, spores also adapt to adhere to pathogens by secreting adhesive compounds44. These compounds help fungi capture live nematodes and their eggs. The corresponding process involves several steps: attraction, recognition, trap formation, adhesion, penetration, and digestion45. Nematodes are attracted to mechanical (constrictor rings) or adhesive fungal traps (adhesive nets, adhesive buttons, adhesive columns, non-constrictor rings) where they are trapped or activate fungal capture mechanisms (restriction traps). The nematode is then paralyzed and the hyphae pierce its cuticle, growing inside the nematode body and completely degrading it by eliminating digestive enzymes45,46.

The cultivable endophytic microbiota of S. erecta leaves include phytopathogenic fungi

Electron micrographs showed endophytic fungi in S. erecta leaf tissues. Previous studies have also shown biotrophic interactions involving leaf endophytic fungi in the epidermis of this plant, particularly Bipolaris and Curvularia species31. The literature reports the isolation of endophytic fungi from S. erecta due to their phytopathogenic habit; however, these fungi express symbiotic traits when associated with S. erecta. Endophytic fungi positively affect host plants by producing phytohormones, improving nutrient acquisition, and protecting against invading pathogens, either via direct antibiosis or activation of induced resistance mechanisms in the plant47,48.

The species Colletotrichum gigasporum, Diaporthe schini, Lasiodiplodia theobromae, Macrophomina pseudophaseolina, Nigrospora sphaerica, and Pseudofusicoccum sp. were isolated from S. erecta leaf tissues. Silva et al.49 suggested the genus Colletotrichum has different lifestyles, including phytopathogenic, endophytic, and hemibiotrophic fungi. Species of the genus Colletotrichum have already been isolated as endophytes from various plants, including Nothapodytes pittosporoides, Morus alba, and Dendrobium spp.50,51,52. The species Colletotrichum gigasporum appears to be widely distributed in tropical regions, being isolated from Centella asiatica, Stylosanthes guianensis, and Coffea arabica53.

The species Diaporthe schini may have phytotoxic activity. Brun et al.54 demonstrated that a bio-herbicide produced from D. schini metabolites effectively suppressed the growth of Bidens pilosa L., Amaranthus viridis L., Echinochloa crusgalli (L.) Beauv., and Lolium multiflorum Lam. weeds. In addition, the data presented by Zhou et al.55 showed that strains of the genus Diaporthe consistently reduced gall formation by nematodes. The nematicidal behavior of this genus corroborates the presence of nematode-trapping fungi on the leaf surface of S. erecta.

Lasiodiplodia theobromae is a fungal species recognized for its pathogenicity, leading to the demise of tropical fruit trees in the northeastern region of Brazil. When introduced to mango fruits and young cashew, Annona spp., and Spondias spp. plants, it induces necrotic lesions of varying severity levels26,56. Initially regarded as a pathogen with limited impact, primarily affecting stressed plants, this species has evolved into a prominent pathogenic agent significantly influencing the viability and productivity of tropical fruit trees in Brazil. Its effects encompass stem canker, Leucostoma canker, plant mortality, and postharvest fruit decay57. Lasiodiplodia can infect a wide range of host plants or survive as saprophytes or endophytes in seeds and other living tissues58,59.

Fungi of the genus Macrophomina have a wide geographical distribution. They have been described as one of the most destructive phytopathogenic fungi60. However, Macrophomina pseudophaseolina does not trigger pathogenic symptoms in S. erecta. In sensitive plants, this pathogen grows rapidly and produces a large number of sclerotia, which clog the vessels, resulting in plant wilting61. A study assessed the nematophagous potential of M. phaseolina. The results showed 98% mortality in Meloidogyne javanica after 48 h of exposure62. Considering that D. schini and M. pseudophaseolina are a part of the endophytic microbiome of S. erecta, we suggested the development of extensive examination methods to discern whether the nematophagous fungi observed by SEM in this study correspond to D. schini or M. pseudophaseolina.

Nigrospora sphaerica leads to the development of rust on Camellia sinensis leaves and spots on Vaccinium corymbosum leaves, branches, and shoots63,64. However, an endophytic N. sphaerica isolate obtained from the medicinal plant Euphorbia hirta demonstrated the ability to synthesize phenolic compounds and flavonoids65. Pseudofusicoccum sp. is associated with death, canker, and fruit rot in several tropical hosts, including mango and Carya illinoinensis66,67. In contrast, endophytic Pseudofusicoccum sp. isolates can synthesize phenolic compounds and cyclopeptides68,69.

The metagenomic study of the endophytic colonization of S. erecta leaves showed low microbial diversity and a tendency toward a higher number of fungal species

We recovered a total of 90 species from S. erecta leaf samples, with 57 ASVs being assigned to bacteria. Other studies with medicinal plant leaves recovered a greater number of OTUs, such as 174 bacterial OTUs in Aloe vera and 114 OTUs for endophytic bacteria in Mentha longifolia70,71. Liu et al.72 revealed that the leaf microbiome of Paris polyphylla var. yunnanensis harbours 910 OTUs. In Hamamelis virginiana L., the endophytic microbiome of the leaves totalled 501 bacterial species and 68 fungi, while in Achillea millefolium L. these values reached 155 bacteria and 52 fungi respectively73, that is, microbiomes more diverse than that found in S. erecta. Work with medicinal plants suggests that the Shannon index in the leaf endophytic microbiome is lower than in the root microbiome, given the selective pressure induced by the presence of secondary metabolites in the leaves72.

Antimicrobial secondary metabolites in S. erecta leaves restrict colonization and reduce the diversity of leaf endophytic microbiome, especially of bacteria. These bioactive compounds include flavonoids, alkaloids, and essential oils31. Cardoso et al.28 reported that S. erecta leaf extract exhibited antibacterial/antifungal activity against Mycobacterium tuberculosis, Staphylococcus aureus, Pseudomonas aeruginosa, Salmonella setubal, Candida albicans, Saccharomyces cerevisiae, and Escherichia coli. The extract also inhibited the growth of Mycoplasma arginini, M. hominis, and Ureaplasma urealyticum74 due to its toxicity75. Thus, the endophytic conditions facilitated by S. erecta could potentially exert a subtle influence on fungi, rendering them more susceptible to epiphytic (as observed by SEM) and endophytic colonization (as seen in the metagenomic study). On the other hand, fungi that inhabit this plant may be prone to inducing pathways associated with defense responses. Studies attest that endophytes can induce the phenylpropanoid pathway, in which several defense compounds are formed, including pathogenesis-related proteins (PR)76,77. Thus, since the endophytic microbiome of medicinal plants can have a significant impact on the production of unique secondary metabolites and pharmacologically active substances78,79,80, we suggest that S. erecta endophytic fungi be examined extensively to elucidate the effects of these microorganisms on the potential of this species to produce beneficial bioactive compounds.

The metagenomic study of the endophytic colonization of S. erecta leaves showed association between the nematodes of the genus halicephalobus with leaf tissues, the presence of methanogenic archaea, and a predominance of Alphaproteobacteria among the sequences

The classification of Halicephalobus sp. sequences corroborated the presence of nematode eggs on the epidermal surface of S. erecta, as observed by SEM. This genus exhibits an almost cosmopolitan distribution81, with studies showing insects as vectors for the entry of Halicephalobus nematodes into fungal colonies82 or plant tissues83. However, the exclusive presence of methanogenic archaea may indicate the presence of organisms reducing the availability of CO2 in S. erecta tissues for methane84. This process would constitute a mechanism of metabolic competition with S. erecta, but the physiological benefits of these archaea should symbiotically balance the cost of maintaining them. An analysis of the rhizospheric and root endophytic microbiome of Oryza longistaminata identified a great diversity of methanogenic archaea85.

The CH4 generated by these archaea may be conveyed through S. erecta leaf tissues, potentially creating a microenvironment rich in both inorganic and organic substances. This environment, inclusive of C1 compounds like methane, could support the proliferation of methanotrophs utilizing CH4 and methanol as carbon sources86,87. Thus, a considerable number of methanotrophs such as Methylobacterium can develop in the tissues. It is key to maintain Alphaproteobacteria of the genus Methylobacterium in plant tissues because they can fix atmospheric nitrogen and produce the hormone cytokinin and the enzymes pectinase and cellulase, thus increasing plant growth due to nitrogen availability and systemic resistance induction88,89,90. Bacteria of the genus Methylobacterium were found in S. erecta SE-1 and SE-3.

Associated with this, the Alphaproteobacteria Atribacter laminatus present in SE-1 and SE-2 belongs to the phylum Atribacterota, which commonly has species in anoxic sediments rich in methane. Genetic analyses suggest a heterotrophic metabolism producing fermentation products such as acetate, ethanol, and CO2. These products support methanogenic substances within the microbial community, explaining their co-occurrence with methanogenic archaea91,92. Alphaproteobacteria of the genus Sphingomonas were also found in SE-1 and SE-3 sequences. These bacteria have been described as common and abundant in plant tissues, being found in the microbiome of 26 plant species in 11 families. The maximum level of Sphingomonas was 108 g−1 (wet weight) of plant tissue, demonstrating large populations93. Some studies showed that species in this genus improve plant growth under stressful conditions such as drought, salinity, and the presence of heavy metals in agricultural soils. This role has been attributed to its potential to produce plant growth hormones such as gibberellins and indole-acetic acid94.

Metagenomic studies support the idea that the microbiome residing in medicinal plants may be considerably variable between species95. Our work corroborates these studies, as we observed in the leaf microbiome of S. erecta, a generic composition different from that observed, for example, in Aloe vera, where the genera Pseudomonas and Bacillus predominated96. Cyanobacteria and Rhizobium were the most frequent genera in the leaf microbiome of Paris polyphylla var. yunnanensis. Here, Rizobium appears at low frequency, and only in SE-372. Streptophyta was the dominant genus found in Senna italica leaf samples97. In the leaves of Bouvardia ternifolia, the predominant genera were Erwinia, Propionibacterium and Microbacterium, genera that did not appear in the S. erecta samples98. In the case of Dicoma anomala, Cutibacterium sequences were predominant in the leaves, but also the genera Acinetobacter and Methylobacterium99, observed at high frequency in S. erecta.

The metagenomic study of endophytic fungi in S. erecta showed a predominance of the phylum ascomycota, mainly comprising plant growth-promoting yeasts

Ascomycota species form symbiotic relationships with plants, with important functions such as nutrient acquisition and disease suppression100,101. In this study, the three samples of S. erecta analyzed had fungi of the genus Malassezia, including the species Malassezia globosa and Malassezia restricta. Malassezia was also one of the most frequent genera in the microbiome of stigmas and pistils in Orobanche alsaticae flowers102, with Malassezia restricta being a very common species. Elhady et al.103 showed that Malassezia globosa was the most frequent species in the microbiome of nematode-suppressive agricultural soil. Malassezia restricta and Malassezia globosa were found in the leaf microbiome of Astragalus canadenses and in several developmental stages of the butterfly Lycaeides melissa, which hosts A. canadenses104.

S. erecta leaves showed a tendency to accumulate yeasts in their endophytic microbiome, especially of the genera Malassezia, Leucosporidium, Meyerozyma, and Hannaella. The genera Leucosporidium was also sampled as part of the leaf microbiome of pasture plants105, but studies have shown that these yeasts may also be present in the phyllosphere microbiome106. Meyerozyma is considered a yeast that promotes plant growth107, being observed in the endophytic microbiome of grains108 and also in fruits of plants in the Cerrado biome109. The species Meyerozyma guilliermondii is promising in reducing the effects of abiotic stress, increasing tolerance, and improving crop performance110,111. This species is well known for its antagonistic effects against phytopathogens such as Fusarium equiseti112. The genus Peniophora was present in SE-1 and SE-2. Wu et al.113 suggested that species of this genus promote plant growth. Peniophora and Lenzites synthesize laccases, which are important enzymes that act in plant morphogenesis, fungal plant-pathogen/host interaction, defense against stress, pigment formation, and phenolic compound detoxification114,115,116. The genus Hannaella and the bacteria Sphingomonas and Methylobacterium are the most abundant in the seed microbiome of six different Oryza sativa genotypes, consistent with the findings of the present117.

We suggest further studies with S. erecta to identify other microbiomes, such as rhizospheric and root, and the leaf endophytic microbiome of other plant species. This will help us understand the effective role of the microorganisms found in this study in the microbiomes of different plant species, especially in medicinal plants such as S. erecta. However, none of the endophytic species isolated from S. erecta leaf tissues were identified in metagenomic sequences, indicating that isolation and metagenomic techniques are not mutually exclusive, contributing to a comprehensive understanding of microbial diversity found in the microbiome of medicinal plants.

Conclusions

In the presented study it was depicted that epiphytic fungi interact with S. erecta leaf tissues. They are horizontally transmitted to internal tissues via stomata or the base of the trichomes and express functional traits for trapping nematodes. Cultivable endophytic fungi isolated from S. erecta are known for their phytopathogenic habits; nevertheless, their association with S. erecta did not elicit dysbiosis symptoms. We confirmed the hypothesis of low microbiome diversity in S. erecta, with a tendency toward a higher number of fungal species, suggesting that this is an effect of antibacterial secondary metabolites present in the leaves. In contrast, the classification of Halicephalobus sp. sequences corroborated the presence of nematode eggs on the epidermal surface of S. erecta. In addition, we confirmed the presence of methanogenic archaea and a considerable number of methanotrophs of the genus Methylobacterium. The methanogenic study of endophytic fungi showed plant growth-promoting yeasts, mainly of the genera Malassezia, Leucosporidium, Meyerozyma, and Hannaella. Endophytic fungi and other microbiomes isolated from S. erecta should be examined to understand the effects of these microorganisms on the capacity of the species to produce beneficial bioactive compounds.

Materials and methods

Collection of biological materials

Healthy leaves were sampled from three S. erecta samples (SE-1, SE-2 and SE-3) collected in a field at Fontes do Saber farm, Rio Verde, GO, Brazil (-17.783262° S -50.967928° W), an area of transitional vegetation between strict Cerrado and Cerradão (Fig. 10A,B). The collection of adult shrub-like samples was previously authorized by the SISBio (Biodiversity Authorization and Information System) under license no. 92592/1 and by the Instituto Federal Goiano under registration no. 011/2021. Thus, all necessary licenses and permissions were obtained for the development of the study, and the plant collection and use was in accordance with all the relevant guidelines. The samples were carefully packed in previously sterilized plastic bags and stored in thermal boxes containing ice. The material was immediately sent to the Laboratory of Agricultural Microbiology, Instituto Federal Goiano, Rio Verde campus, for processing. The leaves were carefully sectioned from the stem using a sterilized scalpel and sent for analysis of leaf surface colonization by epiphytic fungi, internal leaf tissue colonization by endophytes, and metagenomic study. A voucher sample was deposited for identification confirmation at the herbarium of the Instituto Federal Goiano, Rio Verde campus with a catalog number of 545. The material was then sent to Dr. Germano Guarim Neto at the Federal University of Mato Grosso for identification. The identification of the material as Serjania erecta Radlk was confirmed.

Figure 10
figure 10

Sampling area of S. erecta samples (SE-1, SE-2 and SE-3) at Fontes do Saber farm, Rio Verde, GO, Brazil (A), Sampling points in the experimental area, (B) and location of the different leaf regions evaluated: rachis, teeth, and blade (C). Map constructed using ArcGIS Pro 3.1.5 (ESRI) software, obtained from https://www.esri.com/en-us/arcgis/products/arcgis-pro.

Full size image

Epidermal surface epiphytic colonization was assessed by SEM, and endophytic colonization in internal leaf tissues was evaluated by isolation in culture medium. Metagenomic analyses with next-generation sequencing were performed to determine the resident endophytic microbiome.

Scanning electron microscopy

Leaf fragments of approximately 3 mm2 were obtained from the rachis, blade, and teeth of S. erecta leaves (Fig. 10C). The fragments were fixed in formalin, acetic acid, and 70% ethyl alcohol and analyzed at the SEM laboratory of the Regional Center for Technological Development and Innovation (CRTI) (for details on S. erecta leaf anatomy, see Freitas31). The samples underwent a critical point drying process, compositional analyses, and coating with gold, which served as a conductive element for image acquisition.

Images from the epidermal surface of the three samples were captured using a Jeol JSM7100F field emission SEM microscope (SEM-FEG) with an electron acceleration voltage of 5 keV in secondary electron detection (SED) mode. Alternatively, images from longitudinal and transverse sections were used to assess the interaction of hyphae and fungal structures with S. erecta internal leaf tissues.

Isolation of endophytic fungi

Endophytic fungi were isolated following the methodology proposed by dos Reis et al.118. Thus, the S. erecta leaf samples were superficially disinfected to eliminate epiphytic microorganisms. They were then successively rinsed in ethanol (70%), sodium hypochlorite 2.5% (active chlorine), and ethanol (70%) solutions for 1 min, 5 min, and 30 s, respectively. At the end of the process, the samples were rinsed four times in autoclaved distilled water, and a 100 µL aliquot was collected during the last rinse for inoculation in nutrient broth (3 g meat extract, 5 g peptone) at 28 °C for 24 h to test the efficiency of the disinfestation process.

Aseptic leaf fragments of approximately 1 cm2 (1 fragment per leaf region, analyzed in triplicate per sample, totaling 9 fragments per sample) were placed in Petri dishes containing potato dextrose agar (PDA) medium (potato infusion broth, 200 ml; dextrose, 20 g; agar, 17 g; q.s.p. 1,000 ml, final pH = 5.6 ± 0.2) supplemented with azithromycin (500 mg L−1). The plates were incubated at 30 °C and monitored for seven days. Fungal colonies possibly associated with leaf tissues were purified by removing mycelial fragments using an inoculation loop and transferring them to plates containing PDA.

Molecular identification of fungal isolates

For molecular identification, fungal isolates were grown in potato dextrose (PD) broth (potato infusion broth, 400 ml; dextrose, 20 g; q.s.p. 1,000 ml, final pH = 6.6 ± 0.2) for seven days. Subsequently, genomic DNA was extracted in triplicate from each isolate, following the method proposed by Cheng and Jiang119 and using a Minirep extraction kit (Axygen biosciences, USA) according to the manufacturer’s instructions. Identification was carried out by sequencing the ITS-1 and β-tubulin regions after amplification and purification. The Sanger method was used for sequencing, and the sequences were paired by similarity with sequences from the GenBank for phylogenetic inference using the nucleotide Basic Local Alignment Search Tool (BLASTn)120 considering homology greater than 98%. The fungal sequences obtained were concatenated and aligned with the sequences of nine fungal species, also obtained from the GenBank (Colletotrichum gigasporum, Diaporthe schini, Lasiodiplodia theobromae, Pseudofusicoccum adansoniae, Pseudofusicoccum violaceum, Pseudofusicoccum stromaticum, Macrophomina pseudophaseolina, Macrophomina phaseolina, and Nigrospora sphaerica). The sequences were aligned using the Clustal Omega software121.

The sequence evolution model was selected according to the Bayesian Information Criterion (BIC) using the jModelTest 2 software122. The selected model was HKY + G with a gamma shape of 1.8010. Phylogenetic trees were independently inferred for bacteria and fungi using Bayesian inference methods in the MrBayes v.3.2.6. software123. Each tree underwent four independent runs, with 10 × 106 generations assigned to the chains and a posteriori probability distribution obtained every 500 generations. The first 2,500 trees sampled were discarded before calculating the consensus trees, one for bacteria and one for fungi, to ensure chain convergence. The phylogenetic tree with the highest Bayesian probability was visualized and edited using FigTree v 1.4.4124.

Genetic data for metagenomic analysis

The community of total endophytic fungi was assessed using S. erecta leaves previously subjected to the disinfestation process described above, ensuring total epiphyte removal. The leaves were then placed in liquid nitrogen until analysis. DNA was extracted using the PowerSoil Pro Kit (QIAgen), from 200 mg of leaf tissue per replicate, each sample (SE-1, SE-2 and SE-3) was analyzed in triplicate. DNA concentration and purity were monitored by electrophoresis on 1.0% (w/v) agarose gel. V3-341F (CCTACGGGNGGCWGCAG) and V4-785R (GACTACHVGGGTATCTAATCC) primers125 were used for amplification of the 16S rRNA gene region of bacteria and archaea, Euk1391F (GTACACACCGCCCGTC) and EukBR (TGACCTTCTGCAGGTTCACCTAC) were used to amplify the 18S rRNA126 gene region of the fungi, and the amplification products were visualized on 1.5% (w/v) agarose gel. After quantification, qualification, grouping, and purification, polymerase chain reaction (PCR) amplification products were sequenced using the KAPA Library Quantification kit for Illumina (Roche) on an Illumina MiSeq sequencer with a coverage of 300,000 reads per sample.

Bioinformatics analysis

We utilized the Nextflow pipeline ampliseq v.2.7.1 (available at [https://github.com/nf-core/ampliseq]) for the analysis of 16S and 18S rRNA sequencing data. Initially, the integrity of raw sequencing reads was assessed using FastQC127, followed by the removal of primer sequences using cutadapt v.2.7128. Subsequent processing involved denoising, dereplication, and chimeric sequence filtration through DADA2129. The denoised paired-end reads were truncated from position 233 (forward) and 229 (reverse) after a thorough manual inspection of the sequencing error patterns. Reads that did not meet specific length criteria were excluded. The truncated sequences were merged with at least a 20 bp overlap, resulting in exact amplicon sequence variants (ASVs). These ASVs were taxonomically classified from phylum to species level and clustered with 99% similarity using the SILVA v132 database130. Following the taxonomic classification of ASVs into OTUs (Operational Taxonomic Units) by applying a Naïve Bayes classifier implemented in QIIME2131 trained on the preprocessed database. Subsequently, ASVs associated with mitochondrial or chloroplastic origins were filtered. For downstream analysis, we exclusively considered ASVs with a read frequency of ≥ 5 in at least one sample. Furthermore, rarefaction curves were employed to evaluate whether the sequencing depth adequately captured the full extent of species richness within the samples.