Fully expanded mature leaves of adult T.cacao plants of the genotypes Catongo (susceptible to M. perniciosa) and CCN-51 (resistant to M. perniciosa) (Gramacho et al., 1992) cultivated under FD and GH conditions were collected and used to extract AWF. The collection of mature leaves on branches infected by M. perniciosa (green witches’ broom) were carried out randomly at the Active Germplasm Bank (BAG) of CEPLAC/CEPEC in Ilhéus, Bahia, Brazil (14°45′40.2″S, 39°14′03.9″W) on plants under FD conditions. Leaves from asymptomatic plants, grown in a GH at CEPLAC/CEPEC, were also used. AWF extraction was performed by vacuum infiltration followed by centrifugation, as previously described (Pirovani et al., 2008; de Oliveira et al., 2024). A total of approximately 200 μL/g fresh weight of T. cacao leaf was collected. For proteomic analysis, approximately 1,200 g of leaves of the Catongo GH genotype, 715 g of Catongo FD leaves, 800 g of CCN-51 GH leaves, and 750 g of CCN-51 FD leaves were collected. For the analysis of chitinase, β-1,3-glucanase, and protease activity, three biological replicates were performed for each sample using a total of five leaves in each AWF extraction.

The collected AWF was lyophilized in aliquots with a volume of 15 mL and used in protein extraction according to Wang et al. (2003), with modifications according to Pirovani et al. (2008). After protein recovery, the precipitate was resuspended in urea (8 mol L⁻¹) and quantified using the 2-D Quant Kit (Cytiva), according to the manufacturer’s recommendations, using BSA (bovine serum albumin) as standard.

After quantification, the proteins were resolved in SDS-PAGE (sodium dodecyl sulfate–polyacrylamide gel electrophoresis) in mini vats electrophoresis (Hoefer), with 8 cm × 10 cm gels at 12.5% acrylamide (Laemmli, 1970). A total of 30 μg of AWF protein from T. cacao genotypes was applied onto gels, along with 30 μg of total leaf protein extract (TE) from CCN-51 FD extracted according to Pirovani et al. (2008), to verify the protein band profiles of each sample (AWF and TE) and their differences.

A mass of 100 µg of AWF proteins from T. cacao genotype samples was treated and subjected to tryptic digestion according to Villén and Gygi (2008) with modifications described by de Almeida et al. (2023). The peptides were then analyzed in technical triplicates on a liquid chromatography system (Agilent 1290 Infinity II HPLC) coupled to a quadrupole/Time-of-Flight mass spectrometer (Agilent 6545 LC/QTOF) with C18 columns (100 mm × 100 mm) (de Almeida et al., 2023).

To analyze the spectral data generated for the identification of peptides, we used the software Spectrum Mill from the Broad Institute (Rev BI.07.08.214). Raw data files were converted to a peak list format (mgf) without summing the scans by the Mascot Distiller software v.2.6.2.0, 2009 (Matrix Science Ldt.). After extracting the mass spectrometry (MS)/MS spectra, a database search of T. cacao and fungi was performed. All databases were downloaded from UniProt (https://www.uniprot.org). A database of the theoretical T. cacao apoplastome (de Oliveira et al., 2024) was also used to identify predicted proteins for the apoplast. The research was carried out according to the following configurations: carbamidomethylation with fixed modifications, methionine oxidation with variable modification, trypsin for cleavage, and 0.1-Da tolerance for precursor and fragment ions. Statistical analyses were performed using the Mass Profiler Professional 15.1 software (MPP; Agilent), determined through analyses in which the inclusion criteria were defined as p ≤ 0.05 and a fold change ≥ 1.5.

The identified proteins were submitted for functional annotation and categorization according to function and cellular processes using the BLAST2GO software (Conesa et al., 2005) available at the OmicsBox platform (https://www.biobam.com/blast2go/).

The extraction efficiency of AWF from T. cacao leaves for recovery of apoplastic proteins was compared and evaluated with the efficiency of total extract (TE) of T. cacao leaves for identification of apoplastic proteins.

Initially, proteins were recovered from the AWF of T. cacao leaves carried out in the present study obtained as described in the section “Protein extraction from apoplastic fluid and SDS-PAGE.” Then, files containing raw peptide data from TE samples from T. cacao leaves carried out in previous work (de Almeida et al., 2023) were used for comparison. In terms of methodology comparison, only the CCN-51 variety was used. The raw peptide data were analyzed against the T. cacao database and the database of predicted proteins for the apoplast (de Oliveira et al., 2024) as per the previous section (“Mass spectrometry and database analysis”). For this analysis, differences in accumulation among the identified proteins were not taken into account. The data obtained did not need to pass statistical inclusion criteria (p ≤ 0.05 and fold change ≥ 1.5). Verification was based on the number of proteins obtained in each sample rather than on protein abundance.

The transcript pattern for the predicted apoplast proteins was verified from reference files containing transcripts from Theobroma cacao cv. ‘Comum’ (Forastero). The files in FASTA format (GCA_000208745.2_Criollo_cocoa_genome_V2/and GCA_000403535.1_Theobroma_cacaoMatina) were obtained from the GenBank database (https://ftp.ncbi.nlm.nih.gov/genomes/genbank/plant/Theobroma_cacao/latest_assembly_versions/). Relative quantification was performed for the six transcripts exclusive to the Catongo genotype and 19 exclusive to CCN-51, with 14 being common, totaling 39 transcripts, which correspond to the diversity of proteins identified in the apoplast of T. cacao in this study. Therefore, proteins that appeared repeatedly in the samples of both Catongo and CCN-51, regardless of the growth condition, were excluded to prevent duplication. Public data were used from10 T. Cacao RNA-Seq libraries (Teixeira et al., 2014), available at NCBI’s Sequence Read Archive (SRA) (https://www.ncbi.nlm.nih.gov/sra) under accession number SRA066232, with five libraries of the control condition and five of the M. perniciosa–infected condition. The relative quantification of T. cacao transcripts and ranking in descending order of expression was performed with the Galaxy platform (https://rna.usegalaxy.eu), using the Salmon extension (Patro et al., 2017). The corresponding proteins of each analyzed transcript were found with the BlastX search tool (https://blast.ncbi.nlm.nih.gov/).

To visualize the expression profile of the transcripts, a heatmap was plotted using the Complex Heatmap packages in the statistical software R (R Core Team, 2022).

The protein–protein interaction network was created with the proteins identified in the apoplast of the Catongo and CCN-51 genotypes of T. cacao using the homologous sequences of Arabidopsis thaliana as reference. The network was built according to the STRING 11.5 database (https://string-db.org/) based on the following parameters: high level of confidence (0.7), no more than 50 interactions, and addition of nodes until network saturation.

The results obtained were analyzed using the Cytoscape software version 3.9.1 (https://cytoscape.org/) (Shannon et al., 2003). The Igraph package of the statistical tool R Studio was used to group the proteins and calculate the centrality (betweenness) and nodes (degree) parameters. To enrich genetic ontology terms and categories, the Biological Networks Gene Ontology v.3.0.3 (BiNGO) tool was used (Maere et al., 2005) with a Cytoscape plugin (https://apps.cytoscape.org/apps/bingo).

A mass of 30 μg of AWF protein from cacao genotypes and 7 μL of molecular weight marker (Kaleidoscope, pre-stained, 0.45 μm pore, Bio-Rad, USA) were applied to SDS-PAGE in electrophoresis minivats (Hoefer), with 8 cm × 10 cm gels containing 12.5% acrylamide and run at a constant amperage of 30 mA as described by Oliveira et al. (2020). The membranes were incubated with PR-3 primary antibodies (8:20,000) for 1 h. Subsequently, the membrane was washed three times, for 15 min each wash, with 1× Tris-buffered saline (TBS-T) buffer followed by incubation with the secondary antibody (anti-rabbit Immunoglobulin G (IgG)) conjugated to alkaline phosphatase (7.5 μ:2000 of 1× TBS-T) for 2 h. After incubation, the membrane was washed again three times and finally incubated for 15 min with development buffer [MgCl₂, 5 mmol L⁻¹; NaCl, 100 mmol L⁻¹; Tris-HCl (pH 9.8), 100 mmol L⁻¹; and H₂O]. The development was carried out in the absence of light with Nitroblue tetrazolium/5-bromo-4-chloro-3-indolyl phosphate (NBT/BCIP) substrates for approximately 5 min. Finally, the apparent accumulation was determined from images of the nitrocellulose membrane, using the Gel Quant Net 1.8.2 program (http://biochemlabsolutions.com/GelQuantNET.html).

For enzyme activity analysis, AWF from Catongo and CCN-51 GH and FD genotypes was collected by infiltration in ice-cold distilled water as described by de Oliveira et al. (2024). A 30-μL aliquot of each sample was separated for quantification by the Bradford method (Bradford, 1976) using BSA as standard.

Obtaining a good yield of apoplast proteins, which can be soluble in AWF or linked to the cell wall, is still a challenge, due to the difficulties of avoiding disruption of the cell wall and membrane. As a result, they are still poorly characterized (Delaunois et al., 2014).

The technique for obtaining apoplastic proteins, which initially involves the extraction of AWF, demonstrated a high protein yield in samples of T. cacao leaves, despite presenting cytoplasmic contamination of 56% ( Figure 3 ). The number of apoplastic proteins identified (27) was greater from AWF extraction compared to obtaining apoplastic proteins from TE (19). For the recovery of apoplastic proteins from the TE of T. cacao leaves ( Figure 3 ), cytoplasmic contamination of 87% was observed, corresponding to 132 proteins out of a total of 151, due to the leaf maceration process.

The total number of proteins recovered is reduced using the optimized AWF extraction technique from T. cacao leaves (62 proteins), compared to the number of proteins in the TE sample (151 proteins). However, it is important to highlight that the concentration of proteins in the apoplast is lower compared to those present in the internal space of the cells (Delaunois et al., 2014), additionally obtaining AWF also reduces cytoplasmic contamination compared to TE samples. The centrifugal force used in AWF extraction is essential to reduce this contamination, and the centrifugation-infiltration method that was used in this work has already been adopted in several works (Soares et al., 2007; Delannoy et al., 2008; Pirovani et al., 2008; Pechanova et al., 2010; de Oliveira et al., 2024), being adapted for each species and enabling better investigation within the area of subcellular proteomics.

Most of the proteins identified in the AWF of T. cacao genotypes are related to defense and stress, with emphasis on pathogenesis-related proteins (PR-proteins).

PR-proteins make up an important class of proteins that play a crucial role in the plant defense response against pathogens. The production of these proteins is increased when the plant is under stress, and they perform several functions, ranging from directly inhibiting the growth of pathogens to strengthening the plant’s defenses (Sels et al., 2008; Sudisha et al., 2012; dos Santos and Franco, 2023).

Currently, 19 families of PRs have been described (dos Santos and Franco, 2023), among which we identified representatives of the families PR-3, PR-4, PR-5, PR-9, and PR-14. Based on the analysis of cacao transcripts using public data carried out in a previous work (Teixeira et al., 2014), some transcripts belonging to apoplastic PR-proteins were up-accumulated in infected plants compared to healthy T. cacao (Forastero) plants used by Teixeira et al. (2014), as members of PR-3, PR-4, PR-5, PR-9, and PR-14 ( Figure 4 ). This corroborates findings in the present work, due to the importance that PR-proteins have in host defense.

Proteins from the PR-3 family (chitinases) were detected only in the resistant genotype (CCN-51), with most representatives showing a greater accumulation in green brooms (FD) ( Figure 1 ). Recently, chitinases (EC 3.2.1.14) were identified in the AWF of T. cacao only in the FD-grown CCN-51 genotype (de Oliveira et al., 2024). This corroborates the data revealed on PR-3 accumulation in CCN-51 FD ( Figure 7 ). Although not identified by proteomic analysis, the protein was also accumulated in the Catongo GH and FD genotypes (susceptible) and verified by Western blot ( Figure 7 ). A previous study demonstrated that the Catongo genotype contained a greater amount of upregulated defense and stress-related proteins, such as an acid chitinase isoform, at 45 days after inoculation with M. perniciosa, compared to the downregulated proteins (dos Santos et al., 2020). The genotype of T. cacao resistant to WBD (TSH1188) showed upregulated chitinase isoforms 42 h after inoculation and 45 days after inoculation with M. perniciosa (dos Santos et al., 2020).

Chitinases play an effective role in breaking down chitin, an essential component in the constitution of the fungal cell wall. One of the main functions of chitinases in plants is associated with the defense response against pathogenic fungi, but they can also have antibacterial and antiviral actions (dos Santos and Franco, 2023). In the T. cacao apoplast of the CCN-51 genotype (resistant), chitinases were identified, represented in cluster 1 and are involved in the defense response ( Figure 6 ). According to the network of interactions, these proteins have a high betweenness value and can be considered bottleneck proteins. Bottleneck proteins play an important role in the regulation and control of metabolic pathways and/or cell signaling and are, therefore, essential in the communication and integration of different parts of the network (Nithya et al., 2023).

Chitinases are important in cell signaling, due to their chitinolytic activity and the formation of oligosaccharides, which act as elicitor molecules that are recognized by the host plant, favoring a cascade of signals (Kaur et al., 2022; dos Santos and Franco, 2023). Extracellular chitinases can have direct antimicrobial action by the degradation of invading hyphae (Stangarlin and Pascholati, 2000) and can be acidic or basic (Vaghela et al., 2022).

PR-4 was identified in both T. cacao genotypes, being up-accumulated in the FD condition (green witch’s broom stage) ( Figure 1 ). When comparing protein accumulation between genotypes under the same condition, PR-4 was up-accumulated in the CCN-51 FD (resistant) genotype and down-accumulated in Catongo FD (susceptible) ( Figure 2 ). PR-4 can act directly on the pathogen’s cell wall due to its ability to bind to chitin and may present chitinolytic activity (Ali et al., 2018; Butassi et al., 2022). Recently, this protein showed increased abundance in the apoplast of peach plants inoculated with the fungus Taphrina deformans (Butassi et al., 2022). The antifungal activity of PR-4 was proven in wheat against fungi of the genus Fusarium (Caruso et al., 1996), in transgenic grapevines in response to powdery mildew disease (Dai et al., 2016) and in response to abiotic stress in rice (Wang et al., 2011).

The PR-4 protein identified in Catongo is represented in cluster 1, involved in the carbohydrate metabolic process. In the CCN-51 genotype, it is represented in cluster 1 and has a high betweenness value and can be considered a bottleneck protein ( Figure 6 ).

Thaumatin-like protein (PR-5) was detected in the Catongo (susceptible) and CCN-51 (resistant) genotypes exclusively in the FD condition ( Figure 1 ). The PR-5 protein is recognized for its role in plants’ defense response against a variety of pathogens. It can accumulate in the apoplast or vacuole and act directly as an antifungal agent, inhibiting the growth and spread of invasive fungi (Stintzi et al., 1993; de Jesús-Pires et al., 2020; Manghwar and Hussain, 2021). Previous studies have demonstrated that the PR-5 protein delayed the disease caused by Phytophthora infestans in potato (Liu et al., 1994). In garlic, PR-5 genes are involved in defense against Fusarium infection (Anisimova et al., 2021). Silencing of a PR-5 gene in the wheat apoplast leads to compromised resistance against the pathogen Puccinia triticina (Zhang et al., 2018). Recently, an osmotin (PR-5) was identified in the apoplast of the WBD-resistant genotype of T. cacao (de Oliveira et al., 2024). The Catongo thaumatin-like protein is represented in cluster 1, being involved in the carbohydrate metabolic process ( Figure 6 ). In CCN-51, it is involved in the defense response (cluster 1) representing a bottleneck protein within the network ( Figure 6 ). Accordingly, it is suggested that this protein plays a crucial role in immune responses in T. cacao, contributing to the reinforcement and tentative resistance of T. cacao to pathogenic threats.

Peroxidases (PR-9) were detected in both genotypes under both growth conditions. Peroxidase 70 was exclusively detected in Catongo FD, peroxidase A2 was up-accumulated in Catongo and CCN-51 GH, and peroxidase 10 was exclusively detected in CCN-51 GH. The expression of two peroxidases was increased in the apoplast of barley infected by the pathogen Pyrenophora teres f. teres (Hassett et al., 2020). This protein plays an important role against pathogens in the initial stages of infection, acting as an antioxidant enzyme, and also actively participates in cell wall synthesis and remodeling (Hiraga et al., 2001; dos Santos and Franco, 2023). In CCN-51, peroxidase 10 is represented in cluster 4 ( Figure 6 ), related to the response to oxidative stress. According to the findings in the present work, the production of peroxidases may be induced in infected FD genotypes. Because some PR-proteins can be produced constitutively (van Loon et al., 2006), it is likely that this is occurring in GH genotypes. Therefore, in normal situations, plants may express peroxidases to manage natural ROS production.

The PR-14 protein (nsLTPs) was detected exclusively in the Catongo GH genotype. PR-14 facilitates the transfer of lipids and has antifungal and antibacterial activity, by acting on the permeability of the fungal/bacterial membrane (Van Loon and Van Strien, 1999; Hoffmann-Sommergruber, 2002; Gorjanović et al., 2005; Patkar and Chattoo, 2006). This protein was represented in cluster 8, involved in lipid transport ( Figure 6 ), but the exact biological role remains to be determined.

Asymptomatic T. cacao plants grown in the GH had a greater number of proteins identified in the apoplastome, both in comparison within and between genotypes in relation to green witches’ broom plants grown in the FD. This may be related to the modulation of the defense system (Nishad et al., 2020). It can be inferred that the defense system of FD plants was previously turned off due to more dynamic interactions with the environment and that M. perniciosa infection favors the reduction of prolonged defense.

In a previous work, it was proven that 45 days after inoculation with M. perniciosa in the Catongo (susceptible) and TSH1188 (resistant to WBD) genotypes, the accumulation of proteins related to stress and defense is reduced in Catongo, whereas, in TSH1188, it was significantly increased (dos Santos et al., 2020). The period of onset and development of WBD in the FD plants used in the present work is unknown. However, presumably more than 45 days had already passed since the onset of the disease, which may explain the reduction in protein accumulation in the genotypes of infected plants (FD).

A cysteine-rich secretory protein showed increased accumulation in the Catongo (susceptible) and CCN-51 (resistant) GH genotypes but decreased in CCN-51 FD and was not detected in Catongo FD. The cysteine-rich repeat protein is apoplastic and plays a fundamental role in plant defense against phytopathogens. A cysteine-rich repeat protein from wheat was proven to be effective against the phytopathogens Rhizoctonia cerealis and Bipolaris sorokiniana, through its proven antifungal activity in inhibiting mycelium growth, as well as regulating the expression of pathogenesis-related genes such as of β-1,3-glucanase and chitinases (Guo et al., 2020). A cotton cysteine-rich repeat protein acted on the stability of a chitinase, protecting it from cleavage by Verticillium dahliae protease. Overexpression of the protein increased resistance against the cotton fungal phytopathogen (Han et al., 2019). The Catongo cysteine-rich secretory protein is represented in cluster 2 and in CCN-51 in cluster 3, being involved in the steroid biosynthetic process in both genotypes ( Figure 6 ). It is likely that this protein acts in the production of important compounds in the stress response of T. cacao. The reduced accumulation of cysteine-rich repeat protein in T. cacao FD (green witch broom stage) genotypes may have been mediated by the strong inhibitory activity of the fungus M. perniciosa, unlike what occurred in T. cacao GH (asymptomatic grown in a GH) genotypes.

The protein SOD was detected in both genotypes only in the GH condition. SOD plays a fundamental role in antioxidant defense in plants due to the increase in reactive oxygen species (ROS) (Mase and Tsukagoshi, 2021). SOD acts as the first line of defense against oxygen free radicals in less toxic forms, such as hydrogen peroxide (H₂O₂) and oxygen (O₂), reducing oxidative damage (Qi et al., 2017; Wang et al., 2018). The main type of SOD found in the apoplast is copper-zinc (Cu-Zn SOD) (Alscher et al., 2002). The SOD protein identified in Catongo (susceptible) GH is represented in cluster 3 and may be involved in the generation of precursor metabolites and energy ( Figure 6 ). This protein is classified according to the biological network as a “hub” protein, where it plays an important role in the transmission of signals, the integration of metabolic pathways, and the regulation of biological processes (Nithya et al., 2023).

Increased SOD accumulation may be associated with increased ROS, especially in response to biotic stress (Jwa and Hwang, 2017; Wang et al., 2018). The decrease in SOD in FD genotypes (green witch broom stage) in the present work may be an effect of the plant’s response to infection, where manipulation of antioxidant defenses may be a strategy used by the pathogen M. perniciosa to facilitate disease progression. Therefore, the fungus possibly interferes with the defense mechanisms of T. cacao, reducing the effectiveness of antioxidant enzymes such as SOD, which could result in greater oxidative damage and benefit the pathogen in the plant.

Some PR-proteins that have enzymatic action, such as PR-3 and PR-2, were evaluated in the present study.

The chitinases evaluated, although they demonstrated a similar pattern of activity in all samples, exhibited higher endochitinase activity compared to the exochitinase activity ( Figure 8 ). Plant chitinases are generally endochitinases with a molecular weight of 25 kDa to 36 kDa (Hoffmann-Sommergruber, 2002; Hamid et al., 2013; Vaghela et al., 2022) and are active in plant self-defense in response to phytopathogen attack (Vaghela et al., 2022).

For endochitinase (EC 3.2.1.14) activity, a statistically significant difference was observed between samples of the susceptible genotype (Catongo GH and FD) ( Figure 8 ). The activity in the Catongo GH genotype was significantly higher than in the FD sample. Although proteomic data did not reveal the accumulation of chitinase in Catongo GH, it was possible to measure its ability to catalyze reactions in this genotype. The use of crude AWF extract with proteins in the native state may have favored its dosage in GH samples (asymptomatic plants grown in a GH).

When evaluating PR-2 activity, we observed statistically significant increases between CCN-51 FD and GH and between CCN-51 FD and Catongo FD, approximately twice as high ( Figure 9 ). Increased PR-2 activity after infection was reported in rice infected by Magnaporthe oryzae (Wang et al., 2021), in chickpea AWF infected by Ascochyta rabiei (Hanselle and Barz, 2001), and in resistant cultivars of wheat infected by Diuraphis noxia (Van Der Westhuizen et al., 1998).

PR-2 (EC 3.2.1.39) has beta-1,3-glucanase activity, an enzyme that breaks beta-1,3-glucan bonds, an important component in the cell wall of many fungi (Hoffmann-Sommergruber, 2002). Therefore, beta-1,3-glucanases play a crucial role in the defense response of plants against fungal pathogens, because the breakdown of beta-1,3-glucan compromises the integrity of the fungal cell wall (Balasubramanian et al., 2012; dos Santos and Franco, 2023). This can lead to the death of the fungus or at least weaken its ability to invade and colonize plant tissue.

The apoplastic AWF of T. cacao genotypes also revealed a high protease activity through the zymogram gel ( Figure 10 ), confirming the activity of proteases, such as the subtilases identified in the present work: protease secreted by CO₂ response and similar proteases to subtilisin (XP_017981649.1, XP_007017870.2, XP_007018543.2 and XP_007017195.2). Subtilases are very numerous in plants and can be directed to the cell wall and contribute to wall modification, cell signaling, and response to biotic and abiotic stress (Figueiredo et al., 2014; Schaller et al., 2018). A subtilase gene in cotton is involved in the plant’s defense against the pathogen Verticillium dahliae (Duan et al., 2016). The identification of subtilases in AWF has already been reported in coffee leaves (Guerra-Guimarães et al., 2015), Arabidopsis thaliana leaves (Jiang et al., 2023) and leaves of the Catongo and CCN-51 genotypes of T. cacao (de Oliveira et al., 2024). The transcript belonging to two protein isoforms were up-accumulated (lcl_NC_030857.1_mrna_XM_018126160.1_26400 and lcl_NC_030857.1_mrna_XM_007018481.2_26945) ( Figure 4 ). The activity of these proteases in the T. cacao apoplast, together with others identified and shown in Figures 1 , 2 , may be important in the infection process by modulating the plant’s defense responses.

A total of 40 proteins from microorganisms were identified in the AWF of T. cacao genotypes, 23 of which belonged to the fungus M. perniciosa present in the FD genotype samples (plants in the green witch’s broom stage) ( Tables 1 , 2 ). Some functions of these proteins are associated with biosynthetic processes, protein transport, and proteolysis. The FD genotypes were under a high level of stress, evident with the branches at the green witch’s broom stage used in AWF extraction, and, in this context, it was predicted that the fungus M. perniciosa would secrete molecules in the T. cacao apoplast in order to interfere with the plant defense (Fiorin et al., 2018).

It is known that the microbial community can exert a symbiotic relationship with plants, resulting from coevolution between species (Pathak et al., 2022). This relationship is possible because their genomes were shaped to survive together, so they stand out for having characteristics that prevent the recognition of microorganisms by the plants’ immune systems (Khare et al., 2018; Pathak et al., 2022).

A protein belonging to an unidentified fungal species (ID: A0A0S6XHH0.1) was found in the Catongo GH sample ( Table 1 ). This protein’s beta-1,3-glucan inhibitory activity has already been established by in vitro and in vivo studies, showing antifungal activity against the fungi Candida albicans and Aspergillus fumigatus (Fujie et al., 2000a, Fujie et al., 2000b). We suggest that the fungus has a beneficial relationship in Catongo against possible invaders. Additional studies are essential for a more thorough understanding of its action.

In a previous work, the diversity of endophytic fungi in T. cacao (Rubini et al., 2005) was proven, including the fungus Lasiodiplodia theobromae. This corroborates our data from this study, where peptides from L. theobromae were identified. Interestingly, L. theobromae peptides were detected in both GH and FD plants. L. theobromae is a pathogenic necrotrophic fungus important to several crops, including T. cacao (Pereira et al., 2006; de Oliveira et al., 2013). Under controlled conditions, the first symptoms can be visible within 2 weeks after inoculation, such as wilting and death within 10 days (Moreira-Morrillo et al., 2021). Because the GH genotypes (asymptomatic plants grown in a GH) had no visible symptoms of any of the fungi mentioned, it is necessary to perform more comprehensive investigations, using other methods, in order to more precisely elucidate the identification of these peptides and/or functions in the plant, such as metagenome sequencing.

Yeast peptides were also found in both genotypes. Endophytic yeasts have been associated with several plant species: Saccharomyces cerevisiae, considered endophytic of Ficus carica (Ling et al., 2020); and Debaryomyces hansenii, endophytic of wheat (Wachowska et al., 2018) and rice (Tantirungkij et al., 2015), all identified in the apoplast of T. cacao. Although endophytes may play a role in increasing growth, acquiring nutrients and strengthening tolerance to stress in plants (Ling et al., 2020), additional studies are needed. Such studies would offer a more comprehensive understanding of the possible symbiotic relationship between these microorganisms and T. cacao, clarifying the benefits for the host plant.