Antifungal Potential of Diaporthe sp. Endophytes from Antillean Avocado Against Fusarium spp.: From Organic Extracts to In Silico Chitin Synthase Inhibition

1. Introduction

Species of the Fusarium genus rank among the most detrimental phytopathogens globally, causing toxin contamination as well as severe diseases and economic losses across a broad spectrum of crops, including cereals, vegetables, and tropical fruits. [1,2] These pathogens induce chlorosis and cotyledon necrosis, water-soaked lesions on the crown and lower stem, pre-emergence and post-emergence growth retardation, brown to black wilt in the lower main root and lateral roots, vascular damage and, in some cases, produce mycotoxins harmful to human and animal health. [3,4]

Current strategies for controlling Fusarium infections encompass the use of resistant crop varieties, implementation of best agricultural practices, and application of synthetic fungicides like phosphine, formaldehyde, carbendazim, thiophanate-methyl or propiconazole + prochloraz, among others. [5,6] However, chemical interventions raise critical concerns regarding environmental sustainability, potential toxicity, and the accumulation of harmful residues in food products. [7] In response to these challenges, plant-microbiome interactions, particularly those involving fungal endophytes, have emerged as a promising avenue for sustainable disease management, due to their ability to enhance plant resilience against both biotic and abiotic stressors. [8,9,10]

Notably, the selection of host plants plays a key role in the successful isolation of endophytes, with highest bioactive potential often found in plants from unique ecosystems like tropical forests, endemic species with unusual longevity, or those that have adapted to extreme environmental conditions. [11,12] That is the case of Antillean avocados (Persea americana var. americana) growth in the Montes de María region (Colombian Caribbean). These plants have been propagated by seed over five decades leading to their classification as endemic varieties well adapted to the region's distinct climatic and geographic conditions (0-1000 m.a.s.l. and 18-27°C). [13,14] The endemic ecotypes—Cebo, Leche, and Manteca— are distinguished based on fruit characteristics, including pulp texture and fat content. However, their cultivation has been severely affected by the Avocado Wilt Complex (AWC), a group of phytopathogens including Fusarium, Verticillium, and Phytophthora, which aggressively compromise the root systems of most avocado trees. [15,16] Remarkably, the survival of some avocado plants suggests the presence of intrinsic defense mechanisms, potentially mediated by a diverse and functionally significant endophytic community. In this context, the present study aims to investigate the microbiota associated with Antillean avocado trees in Montes de María and to evaluate their potential as antifungal metabolites producers, with inhibitory activity against F. solani and F. equiseti.

In the same way, understanding the mechanisms of action of antifungal compounds is fundamental for advancing plant disease management and addressing the growing challenge of fungal resistance. Antifungal compounds target essential structures and metabolic processes in phytopathogenic fungi through multiple mechanisms of action like intracellular reactive oxygen species accumulation, cell membrane disruption, lipid peroxidation, mitochondrial dysfunction [17,18], inhibition of cell wall biosynthesis [19,20], etc. Particularly, the unique composition of the fungal cell wall—dominated by polysaccharides such as glucans and chitins, along with glycoproteins—distinguishes it fundamentally from human cellular structures. [21] Consequently, Chitin Synthase (CS), the enzyme responsible for chitin polymerization, has emerged as a strategic molecular target for the development of selective antifungal agents. [22,23] Despite its clinical and agricultural relevance, structural data for CS in specific phytopathogens such as Fusarium solani and Fusarium equiseti remain limited, often hindering the rational elucidation of inhibition mechanisms. To address this gap and investigate the molecular basis of the antifungal activity observed in this study, an in silico approach was integrated. We performed homology modeling to generate 3D structures of the CS enzymes for both pathogens, followed by molecular docking and interaction fingerprint (PLIF) analyses. This computational strategy was employed to evaluate whether bioactive metabolites, previously isolated from endophytic Diaporthe species, could act as potential inhibitors of this critical enzymatic target.

2. Materials and Methods

2.1. Collection of Plant Material

Leaves and secondary roots from the Cebo, Manteca, and Leche ecotypes of avocado (Persea americana var. americana) cultivated in the Caribbean region of Colombia (09°34′42″N 75°16′15″W), were collected in 2017. The collection of plant material and the study of the isolated fungi were covered by Permission No. 121 of January 22, 2016 (Amendment No. 21), under the Access to Genetic Resources and Derivative Products Agreement No. RGE 46 (Article 6 - Law 1955 of 2019), granted by Ministerio de Ambiente y Desarrollo Sostenible de Colombia. Fungal isolates were placed in the QUIPRONAB strain collection, located at the Chemistry Department of Universidad Nacional de Colombia, Sede Bogotá.

2.2. Avocado Phytopathogens Isolation

Fusarium pathogens were isolated from secondary roots of trees exhibiting symptoms of AWC. Plant material was surface sterilized following established protocols [24], and fungal isolates were identified through a combination of morphological and molecular characterization [25].

Two Fusarium isolates were selected based on an in vivo pathogenicity test. Avocado seedlings, germinated from seeds and cultivated under controlled glasshouse conditions (~18-20 °C day, ~12-15 °C night) for four months, reached 40-50 cm and developed at least four true leaves [26]. Each treatment included ten plants. Conidial suspensions (1 × 10⁷ CFU/mL) were prepared from 14-day-old cultures grown on potato dextrose agar (PDA), and roots were immersed for 2 min prior to transplanting into sterile soil. An additional 1 mL of inoculum was applied near to the root zone. Sterile distilled water served as a negative control. Plants were irrigated weekly for six weeks, and disease severity was evaluated using a 0-3 scale: 0 (no lesion), 1 (chlorosis/slight withering), 2 (severe withering/defoliation), and 3 (plant mortality). The disease index (DI) was calculated using DI = 100 * [(a.X0) + (b.X1) + (c.X2) + (d.X3)] / (N*3), where a-d represents the number of plants at each infection level, and N is the total plants per treatment. Following the experiment, seedlings were uprooted, and the causal agents were re-isolated via root surface sterilization and subsequent culture on PDA.

2.3. Fungal Endophytes Isolation

Healthy leaves and secondary roots were collected from three Antillean avocado ecotypes -Cebo, Leche, and Manteca- that had survived adverse phytosanitary conditions in the Colombian Caribbean region. Plant material was surface sterilized following the previously described protocol. The samples were excised into 5 mm² segments [27] and plated onto different solid culture media, including V8 juice agar, malt extract agar, yeast glucose extract agar, and potato dextrose agar, all supplemented with chloramphenicol (400 ppm). Following the incubation period, individual fungal morphotypes were subcultured onto PDA to obtain pure isolates. To establish a well-characterized fungal collection and ensure culture purity, axenic cultures were obtained [28]. Endophytic fungal strains were subsequently preserved under cryogenic storage conditions [29] for further study.

2.4. Inhibitory Activity of Fungal Endophytes and Their Organic Extracts

Dual-culture assays were conducted on PDA plates (ø=90 mm) by placing 5 mm agar plugs of the entophytic isolate and the phytopathogen (F. solani or F. equiseti) 50 mm apart. Cultures were incubated at 28°c for 14 days and radial fungal growth was measured daily. Control plates consisting of the pathogen grow under identical conditions without endophyte interaction, were included for comparison [30]. The experiment was performed in triplicate, and only endophytes exhibiting significant inhibitory effects on pathogen growth were selected for subsequent organic extraction.

Selected endophytic isolates were cultured in 250 mL Erlenmeyer flasks containing 100 mL of 2% yeast extract broth and incubated at 28°C with constant agitation (150 rpm) for 14 days. Biotechnological products were extracted via three consecutive ultrasound-assisted extractions using 300 mL of ethyl acetate (EtOAc) per flask. The organic phase was subsequently filtered, separated, and concentrated under reduced pressure to obtain crude extracts (Liu & Liu, 2018).

Fungal extracts were tested for antifungal activity using agar dilution method in 12-well plates. A stock solution (50 mg/mL) was prepared in ethanol, from which 40 µl was combined with 10 µl of MTT (2 mg/mL) and 1950 µL of PDA, resulting in a final concentration of 1 mg/mL per well. Pathogens were inoculated using a sterile toothpick from a seven-day PDA culture and incubated at 28°C for 72 h. Growth inhibition (% MGI) was quantified using ImageJ™ and compared to untreated controls [31]. Extracts exhibiting mycelial growth inhibition (% MGI) ≥ 50% at 1000 µg/mL were selected for further dose-response analysis to determine IC₅₀ values, employing agar dilution with concentrations between 25-3000 µg/mL. All experiments were conducted in triplicate, and IC₅₀ values were calculated using GraphPad Prism software.

2.5. Morphological and Molecular Characterizations of Endophytes

Fungal endophytes exhibiting ≥50 % MGI against Fusarium spp. were identified through a combination of morphological and molecular analyses. Molecular identification was conducted via DNA sequence analysis of the internal transcribed spacer (ITS-1) region of ribosomal DNA, following established protocols comparisons for the internal transcribed spacer ITS-1 of the ribosomal DNA [32]. The resulting sequences were queried against GenBank using the Basic Local Alignment Search Tool (BLAST) hosted at NCBI (https://blast.ncbi.nlm.nih.gov/Blast.cgi) to identify the closest taxonomic matches [33]. All ITS-rDNA sequences were submitted to GenBank to obtain accession numbers (Table S1).

2.6. Homology Modeling and Structural Selection

The amino acid sequences of Chitin Synthase (CS) for F. solani and F. equiseti were retrieved from the UniProtKB database. [34] To identify a suitable structural template, a sequence alignment was performed using BLASTp against the Protein Data Bank (PDB) to find a crystallographic structure with high identity and query coverage. Three independent automated prediction servers were employed to generate the 3D structural models: Robetta[35], AlphaFold[36], and I-TASSER [37]. To select the most reliable structure among the generated predictions, the models were structurally superimposed onto the identified template. The final selection was determined by calculating the Root Mean Square Deviation (RMSD) of the backbone atoms relative to the template; the model exhibiting the lowest RMSD value was chosen.

2.7. Molecular Docking and Interaction Analysis

The molecular docking simulations were performed using AutoDock Vina to predict the binding affinity and orientation of the isolated metabolites within the catalytic site of CS. [38] The selected 3D models of F. solani and F. equiseti CS were prepared by adjusting the protonation states at pH 7.4 were assigned using PDB2PQR, and partial charges were added. [39] The 3D structures of the 15 fungal metabolites were generated and energy-minimized using RDKit (MMFF94 force field). [40] Gasteiger partial charges were assigned, and rotatable bonds were defined before conversion to PDBQT format using OpenBabel of all structures. [41]

The search space was defined centered on the coordinates of the reference inhibitor Nikkomycin Z (derived from the template PDB: 7WJO) and creating a cubic grid box of 8Å in each dimension to cover the active site pocket. [42] Docking runs were executed with an exhaustiveness parameter of 16. To validate the protocol, a re-docking assay was performed using the co-crystallized ligand, calculating the Root Mean Square Deviation (RMSD) between the predicted and experimental poses using spyrmsd. [43] The resulting protein-ligand complexes were analyzed to identify key molecular contacts. Interaction fingerprints (IFP) were generated using ProLIF to map hydrogen bonds, hydrophobic contacts, and pi-stacking interactions. [44] To rank the compounds, a similarity analysis was conducted by calculating the Cosine coefficient between the interaction fingerprint of each metabolite and that of the reference inhibitor (Nikkomycin Z).

To identify the most promising candidates, a consensus scoring function (consensus score) was established by integrating the binding affinity and the interaction similarity. Since the Vina score and the PROLIF similarity possess different units and scales, both variables were normalized using the Min-Max scaling method. The consensus score for each molecule was calculated as the arithmetic mean of these normalized values.

3. Results and Discussion

3.1. Avocado Phytopathogens Isolation

Six fungal strains were isolated from the secondary roots of Antillean avocado trees exhibiting symptoms of root rot. These isolates were subjected to morphological and molecular characterization. DNA sequencing was performed to facilitate taxonomic identification (Table 1).

After pathogenicity assay on avocado seedlings, two Fusarium strains could induce disease symptoms. F. solani and F. equiseti exhibited a 100% disease incidence, with disease severity indices of 77% and 80%, respectively, six weeks post-inoculation. Koch’s postulates were confirmed through the successful re-isolation of the pathogens following surface disinfection of infected seedling roots and subsequent culture on PDA. Morphological identification was confirmed based on macroscopic and microscopic characteristics of the re-isolated strains.

These findings are consistent with previous reports on the phytopathogenic potential of Fusarium species. Indeed, Fusarium spp. are among the most prevalent pathogens affecting avocado crops globally, with F. solani, F. equiseti, and F. oxysporum recognized as the most economically significant. Typical disease symptoms include chlorosis, necrosis, water-soaked lesions, growth retardation, and root wilt [45]. In addition, previous studies have reported the disease incidence of Fusarium species in avocado seedlings. For instance, a study conducted in Michoacán, Mexico, on Persea americana var. drymifolia identified F. oxysporum and F. solani isolates as pathogenic, with 63% causing wilting and 16% inducing necrosis [46]. Similarly, in Kenya, F. solani, F. oxysporum, and F. equiseti were associated with stem-end rot in Hass avocados, with disease severity ranging from 19.2% to 60.8% [47]. All about, and our findings further establish Fusarium as a major pathogen in avocado cultivation.

Although Phytophthora cinnamomi is a primary causal agent of root rot in Colombian avocado orchards [48,49], it was not recovered in this study. Previous research has reported the presence of oomycetes like P. cinnamomi, P. citricola, and P. heveae, as well as fungal pathogens including Lasiodiplodia theobromae, Pythium spp., Fusarium spp., and Verticillium spp., in Hass avocado plantations in Antioquia, with disease incidence ranging from 10% to 65% [50]. Additionally, Phytophthora heveae, Calonectria spp., and Fusarium spp. have been identified in avocado nurseries in Valle del Cauca [51]. In Persea americana var. americana (Antillean avocado), P. cinnamomi has been identified as the primary causal agent of wilt, with an incidence exceeding 40% in 2009 [16]. Moreover, recent studies suggest that Colletotrichum spp. contribute to abnormal seedling development and formation of necrotic acervuli in nurseries settings. Additionally, other pathogenic organisms including P. cinnamomi, Sclerotium rolfsii, and Nectria spp., have been implicated in symptoms such as wilting, necrosis, and terminal stem/root rot [52].

3.2. Avocado Fungal Endophytes Isolation

Following surface disinfection and incubation of plant material, a total of 88 strains were selected as different morphotypes based on their growth on PDA. Of these, 19 endophytic fungi were isolated from avocado roots and 69 were recovered from leaves. Most isolates were obtained from Cebo and Manteca avocado ecotypes, whereas the Leche ecotype exhibited the lowest percentage of fungal endophytes (17%). This reduced endophyte diversity in Leche ecotype may be related to observations reported by local avocado growers, who indicate that this ecotype has been the most susceptible to phytopathogenic infections. Consequently, the Leche ecotype has become increasingly rare in the Montes de María region (Oral communication, Asociación de Productores de Aguacate Tecnificado de los Montes de María, Asproatemom, 2017).

Avocado plants in the Colombian Caribbean, particularly those in Montes de María, serve as reservoirs of microbiological diversity due to their longevity and endemic characteristics. Three key traits distinguish these ecotypes: (1) seed-based propagation, (2) adaptation to tropical environments, and (3) resilience to adverse phytosanitary conditions. These factors contribute to the establishment of endophytic communities with significant bioactive potential. Vertical transmission through seeds facilitates the persistence of endophytes across generations, thereby enhancing host survival and conferring resistance against pathogens. Moreover, the high plant diversity in tropical regions has been correlated with abundant and chemically diverse endophytes [53]. Additionally, endophytes modulate host phenotypes through mutualistic interactions, enhancing stress tolerance by producing antimicrobial metabolites and activating systemic defense mechanisms against pathogens [54].

The inhibitory activity of 88 fungal endophytes was assessed using dual culture assays. After 14 days of incubation, avocado-associated endophytes exhibited four distinct antagonistic interactions: 1). Competition for space and nutrients, promoting high colonization density (Figure 1A); 2). Mycoparasitism, characterized by endophytic overgrowth on the pathogen’s mycelium and the formation of appressoria on hyphae (Figure 1B); 3). Fumigant activity via volatile organic compounds (VOCs), validated through separate inoculation and confrontation assays (Figure 1C) [55]; and 4). Antibiosis, wherein extracellular enzymes, volatiles, and antibiotics inhibited pathogen growth (Figure 1D). In this study, antibiosis was the primary selection criterion for identifying isolates with antifungal potential, with 35 strains exhibiting inhibitory effects against F. solani and 29 against F. equiseti.

The production of secondary metabolites is typically more pronounced during the late exponential or stationary growth phases, requiring adjustments in harvest time to optimize yield. Notably, most filamentous fungi undergo a 7-day culture for metabolite extraction [56]. Given this, the 35 endophytes isolated from Antillean avocado leaves and roots that exhibited antifungal activity against Fusarium spp. in dual culture assays were incubated for 14 days prior to organic extraction using ethyl acetate (EtOAc). In an agar dilution assay, eight EtOAc extracts demonstrated >50% inhibition of at least one Fusarium pathogen at 1000 µg/mL. However, extracts UN76, UN104, and UN106, derived from Leche and Cebo ecotypes, unexpectedly promoted F. solani growth (Table 2, Figure 2).

The antifungal activity of the extracts evaluated in this study was comparable to that of other extracts described in the literature as promising inhibitors of Fusarium species. For example, the crude extract of the endophyte Fusarium proliferatum, obtained with EtOAc, showed 75.51% inhibition of F. oxysporum at a concentration of 1000 μg/mL [57]. Similarly, the EtOAc extract of the endophyte Chaetomium globosum exhibited mycelial growth inhibition percentages of 59.06%, 64.90%, and 59.41% against F. graminearum, F. oxysporum, and F. moniliforme, respectively [58].

The inhibitory concentrations (IC₅₀) of the most promising extracts, those with MGI higher than 50% at 1000 µg/mL were determined using a standard dose-response model, yielding values ranging from 199 to 688 µg/mL. Notably, the UN310 extract exhibited the highest antifungal activity, with an IC₅₀ of 362 µg/mL against F. equiseti and 204 µg/mL against F. solani. These values were not significantly different from those of the UN93 extract, which displayed an IC₅₀ of 200 µg/mL against F. solani. In contrast, the endophytic fungus UN51 demonstrated the weakest antifungal activity, with an IC₅₀ exceeding 3000 µg/mL against both Fusarium pathogens (Table 3).

Values followed by the same lowercase letters in the columns of % MGI for each pathogen at 1000 µg/mL do not show significant differences according to Tukey's test (p > 0.05). Similarly, means followed by the same uppercase letters in the rows are not significantly different in a one-way ANOVA analysis (p > 0.05).

3.3. Morphological and Molecular Characterizations of the Most Promising Endophytes

Figure 3 illustrates the macroscopic characteristics of the eight selected endophytic strains cultured on PDA. Colonies of UN22, UN39, UN92, UN93, UN95, UN99, and UN310 exhibited white to gray, woolly aerial mycelium, accompanied by yellow-orange exudates and a dark gray to black reverse pigmentation, with distinct lobed margins. In contrast, strain UN51 displayed the most divergent macroscopic features forming a colony with entire margins, a light-yellow velvety mycelium, and occasional dark brown exudates.

A significant challenge in endophytic fungal identification is the absence of conidia or other reproductive structures when cultured in conventional media, as widely reported in scientific literature [59]. To overcome this limitation, molecular characterization was performed on the eight selected endophytes. Sequence analysis revealed a 99–100% similarity with previously deposited sequences in GenBank, except for strain UN51, which exhibited 96% correspondence (Table 4). All strains coded as UN22, UN39, UN92, UN93, UN95, UN99, and UN310 were closely related to species of the genus Diaporthe. In contrast, strain UN51 was more distantly and showed greater similarity to the ITS1 region of species within the genus Nodulisporium (Figure 3). Given its IC₅₀ value, UN51 was reclassified as an inactive endophytic strain against F. solani and F. equiseti.

3.4. Some Antifungal Compounds from Endophytic Diaporthe sp.

Notably, the endophytic isolates exhibiting the highest antifungal activity against F. solani and F. equiseti were phylogenetically related with the genus Diaporthe (Table 4). This genus, classified within the phylum Ascomycota, comprises over 950 described species that function as endophytes, saprophytes, or pathogens in a diverse range of plant and mammalian hosts [8]. Species within this genus are characterized by their ability to produce a broad spectrum of bioactive metabolites, including terpenoids, steroids, macrolides, alkaloids, flavonoids, and polyketides [60]. These metabolites exhibit cytotoxic and antineoplastic effects [61], anti-inflammatory [62], hypocholesterolemic [63], cytotoxic [64], herbicide [65], antibacterial and antifungal activities [66].

Among antifungal metabolites, cytochalasin-like compounds are the most abundant in Diaporthe genus (Figure 4). These secondary metabolites, derived from polyketide-amino acid fusion, have demonstrated biocontrol potential. They are known as microfilament targeting molecules which exhibit a wide range of biological activities, by interfering with several cellular processes involving cytoskeleton formation [67]. For example, Huang et al. reported the cytochalasins E (1), H (2), and 7-acetoxy-cytochalasin (3) from Diaporthe sp., exhibiting strong antifungal activity against Alternaria oleracea, Pestalotiopsis theae, and Colletotrichum capsici (MIC= 3.125–12.5 μg/mL) [68]. Similarly, cytochalasin H (2), N (4), and epoxy-cytochalasin H (5) from Phomopsis sp. By254 inhibited Sclerotinia sclerotiorum, Fusarium oxysporum, Botrytis cinerea, and Rhizoctonia cerealis (IC₅₀= 0.1–5.0 μg/mL). [69] In addition, cytochalasin J (6) from D. miriciae showed antifungal activity against the fungal plant pathogens Phomopsis obscurans and P. viticola at 300 µM. [70]

Diaporthe species produce another important antifungal compounds that had been evaluated on plant phytopathogens (Figure 4). Among them, phomompsolide A (7), phomopsolide B (8) and phomopsolide C (9), isolated from D. maritima cultures, demostrated growth inhibition on Microbotryum violaceum and Saccharomyces cerevisiae at 25-250 µM. [71] Likewise, phomopsolide G (10) was isolated from Diaporthe sp. AC1 and showed moderate antifungal activity against Fusarium graminearum, F. moniliforme, and Botrytis cinerea. [72] Additionally, the monoterpene verbanol (11) from D. terebinthifolii showed inhibitory activity on the germination of Phyllosticta citricarpa conidia. [73] In the same way, two xanthone dimers, diaporxanthone A (12) and diaporxanthone B (13), produced by D. goulteri L17 exhibited moderate antifungal activities against Nectria sp. and Colletotrichum musae. [74] Gao et.al (2020) isolated from D. eucalyptorum a rare polyketide fatty acid, eucalyptacid A (14),with MIC values ranging from 12.5-50.0 µM against Alternaria solani, B. cinerea, F. solani, and Gibberella saubinettii. [75] The dihydroisocoumarin (−)-(3R,4R)-cis-4-hydroxy-5-methylmellein (15) produced by D. cf. heveae, drastically reduced the growth of Phyllosticta citricarpa and Colletotrichum abscissum, two important citrus diseases worldwide. [76]

3.5. Molecular Docking of Diaporthe Antifungals with Chitin Synthase

The amino acid sequences of CS for F. solani and F. equiseti were retrieved from the UniProt database under accession codes A0A9P9GUZ8 and A0A8J2II09, respectively. A BLASTp search against the PDB identified the cryo-EM structure of Phytophthora sojae chitin synthase 1 (PDB ID: 7WJO) as the most suitable template for homology modeling, given its high sequence identity and structural coverage. Among the 3D models generated by the different servers, those predicted by Robetta exhibited the highest structural fidelity to the template, displaying the lowest Root Mean Square Deviation (RMSD) values of 1.68 Å for F. equiseti and 2.21 Å for F. solani. Consequently, these Robetta-derived models were selected as the reference structures for the subsequent molecular docking simulations.

Visual inspection of the superimposed structures confirms a remarkable topological similarity between the generated Fusarium models and the P. sojae template. Crucially, the detailed analysis of the active site reveals that the architecture of the ligand-binding pocket is maintained in both F. solani and F. equiseti models. As depicted in Figure 5, the key amino acid residues responsible for stabilizing the reference inhibitor Nikkomycin Z are spatially conserved. This structural preservation suggests that the generated models possess a binding pocket competent for ligand recognition, thereby validating their use for the subsequent molecular docking of the endophytic metabolites.

The normalized binding energy values (Vina score) shown in Table 5 revealed that several isolated metabolites possess a theoretical affinity for the CS active site that rivals or even surpasses the reference inhibitor. Notably, compound 12 exhibited the highest predicted binding stability, achieving a normalized Vina score of 1.0 in F. equiseti and 0.83 in F. solani. These values suggest that compound 12 has the potential to form a complex with the enzyme that is theoretically more favorable than the control Nikkomycin Z (which scored 0.65 and 0.57, respectively). This predictive data implies that the hydrophobic skeleton of the isolated endophytes metabolites likely fits tightly within the catalytic pocket, potentially driven by strong Van der Waals interactions.

While high affinity is crucial, the specific mode of binding determines the biological outcome. The interaction fingerprint analysis (PLIF) quantified how well the candidates mimicked the key contacts established by Nikkomycin Z. The similarity scores for the top candidates ranged between 0.40 and 0.66. While lower than the control (which is 1.0 by definition), these values are significant for non-peptidic small molecules. Specifically, compound 13 showed the highest structural similarity in interaction patterns against F. equiseti (0.66), implying that it engages conserved residues essential for catalysis, such as those involved in uridine binding or translocation. [77,78] The lower similarity scores observed for compound 12 (approx. 0.40–0.46), despite its high energy, suggest it may occupy the binding pocket in a unique orientation, potentially exploiting auxiliary sub-pockets not accessed by the reference inhibitor.

Detailed inspection of the ligand-receptor contacts in F. equiseti reveals that Nikkomycin Z (NZ), anchors itself within the active site through hydrogen bonds with Asp388, Pro462, and Ala464, while relying on a critical hydrophobic stacking interaction with Trp553 as is show in Figure 6. This tryptophan residue is structurally significant, as it corresponds to the conserved aromatic residues found in the catalytic loop of the P. sojae template and other Family 2 glycosyltransferases, acting as a gatekeeper that stabilizes the GlcNAc sugar ring. [22,79] Notably, compound 12 successfully mimics this pharmacophoric feature, preserving the hydrophobic clamp with Trp553 and the hydrogen bond with Pro462, while further stabilizing the complex through additional hydrogen bonds with Asp538 and Gly390. The engagement with Asp538 is particularly relevant, as aspartate residues in this region are typically involved in coordinating the divalent metal ions essential for catalysis. [80,81]

The structural basis for the consensus ranking becomes evident when comparing these interaction profiles. While compound 13 establishes a strong hydrogen bonding network involving Asp538, Thr537, and Lys364, it notably lacks the hydrophobic interaction with Trp553 observed in both NZ and compound 12. This absence likely accounts for its slightly lower thermodynamic stability (Vina score) compared to compound 12, despite having a high interaction similarity index. Consequently, compound 12 could have a better activity (Consensus Score = 0.73) because it acts as a dual-modality inhibitor: it not only occupies the catalytic center by interacting with essential aspartates but also replicates the hydrophobic stabilization mechanism exploited by the reference inhibitor.

The docking results for F. solani reveal a distinct interaction landscape, likely attributed to the greater structural divergence of its catalytic domain relative to the template. This conformational variability is reflected in the binding mode of Nikkomycin Z, which shifts to a pose dominated by hydrogen bonds with Asp510, Gln549, Thr537, and Ala464, notably losing the strong hydrophobic anchor observed in the F. equiseti model as is displayed in Figure 7. In contrast, compound 12 demonstrates remarkable adaptability to this distorted pocket. It not only establishes a wide hydrogen bonding network with Pro462, Thr537, Asp538, Ala389, and Gly690, but crucially, it recovers the hydrophobic stacking interaction with Trp553. This ability explains the predicted binding energy (0.83 and 0.57 for compound 12 and Nikkomycin respectively).

Compound 13 relies primarily on electrostatic and polar stabilization, anchoring to the active site through hydrogen bonds with Lys364, Asp388, and Thr537. While Thr537 emerges as persistent contact across all ligands, suggesting a pivotal role in substrate orientation, the absence of the hydrophobic clamp with Trp553 limits the binding stability of compound 13. This comparison further validates the consensus scoring hierarchy, identifying compound 12 as the most robust candidate capable of maintaining key inhibitory interactions despite the structural plasticity of the pathogen's target site for chitin synthase in both organisms.

3.6. Concluding Remarks

Antillean avocado trees that survived adverse phytosanitary conditions in Montes de María serve as reservoirs of diverse endophytic fungi, some capable of producing antifungal metabolites against F. solani and F. equiseti. The most active isolates were found in Manteca and Cebo ecotypes, suggesting a potential correlation between endophyte presence and host resilience. Molecular analyses identified these isolates as Diaporthe species, a genus renowned for its production of bioactive secondary metabolites. Notably, crude extracts from Diaporthe sp. strains UN93 and UN310 demonstrated significant antifungal activity, with IC₅₀ values near 200 μg/mL against Fusarium spp.

This study also identified Epicoccum sp. and Colletotrichum sp. in the secondary roots of symptomatic avocado trees, with Fusarium sp. being the most prevalent. No Phytophthora species were isolated, possibly due to their specific growth requirements, though their absence cannot be confirmed. Future research should expand microbiome characterization using genetic sequencing. In addition, the validation of in vitro results through controlled seedling experiments is crucial, as antagonistic effects do not always translate to effective disease control. The challenges in studying the pathosystem open the door to an alternative biocontrol strategy: the direct application of chemical compounds produced by endophytic microorganisms to protect their host against biotic and abiotic stress factors. Thus, further studies should focus on the chemical characterization of promising extracts to identify the bioactive compounds responsible for their antifungal activity and assess their potential applications in plant disease management.

Furthermore, the integrated in silico analysis proposes a plausible molecular mechanism of action, identifying chitin synthase as a potential target for the bioactive metabolites, particularly candidates 12 and 13. However, these computational findings must be interpreted with caution, as they rely on homology models that, despite their structural quality, inherently carry a degree of uncertainty regarding the precise conformational landscape of the native enzymes. Consequently, these predictive results should be viewed as a theoretical framework to guide future molecular exploration. Subsequent investigations must prioritize direct enzymatic inhibition assays to experimentally validate the affinity of these molecules; alongside structural biology efforts aimed at crystallizing the specific Fusarium CS to corroborate the binding modes proposed in this study.