1. Introduction

Symbiotic relations are present among many organisms in nature, contributing to the success of the involved. Plants associate with many microorganisms, such as fungi, bacteria and virus, varying from mutualisms to parasitisms [1] [2]. The microorganisms associated with plants develop competition relations between themselves, for resources or predation and by the release of volatiles and antimicrobial compounds, having consequences in the structure and stability of the microbial community and in its homeostasis with the host plant [3].

The endophytic fungi are found between the plant’s leaf tissues and stablish a mutualistic relation with the host, without causing symptoms [4]. These fungi are abundant in the tropical regions [5] and have a pathogenic potential if the host plant goes through a stressful situation, for natural causes or for the use of products in its maintenance, which can disequilibrate the endophytic mycobiota and potentialize these fungi’s pathogenicity [6].

Van Bael and colaborators [7] suggested that the endophytic fungi have a defense function in the plant by altering the leaf’s morphology and chemistry. In addition, plants can use directly the endophytic fungi as a defense against these ants [8], as they answer in different ways to each fungus’s signals [9]. The diversity of endophytic fungi that are introduced in the nest by the ant’s foraging activity is not yet widely known, as well as the symbiotic interactions between these endophytic, the ants and their fungal garden [10]. In one of the few studies on the subtopic Van Bael and collaborators [7] verified a tendency for a greater fungi mass per worker in colonies fed with low endophytic leaves and that a high quantity of endophytic fungi of Cucumissativus and Manihotesculenta leaves transported to new nests of Atta colombica (Guérin-Méneville, 1844) could affect its development by limiting the colony’s productivity, while the high endophytic load was beneficial to old colonies, by increasing their defenses against parasitic fungi.

Ants of the genus Atta (Fabricius, 1804) and Acromyrmex (Mayr, 1865), classified as Attine: Attina, are popularly known as leaf-cutting ants and fungus growing ants [11]. Some species are considered plagues in the neotropical region due to the intense cutting of leaves of important agricultural cultures [12]. Another important and obligatory symbiotic relation is the association between the ants of theAtta genus and the fungus Leucoagaricusgongylophorus, which produces biomass, enzymes and nutrients to the colony [13]. The ants execute a quality control of the leaf material that is transported, selecting those which will be incorporated to the fungal garden and removing the unwanted ones [14]. The leaf-cutting ants have a hygiene behavior specific to the different species of endophytic fungi [9], important because filamentous fungi can have antagonistic actions to the attine ant’s cultivar [15]. The mutualist fungus L. gongylophorus, such as the workers, also has defense mechanisms against parasite fungi [10], as the Trichoderma genus fungi [8].

Leaf-cutting ants have a preference for certain plants, like Ligustrum spp., Psidiumguajava, Hyparrheniarufa and Acalypha sp. [16], which is used for maintenance of colonies in laboratories [17]. Acalypha is the most diverse genus of the Euphorbiaceae family with mostly arboreal species [18]. The species Acalyphawilkesiana (Müll.) Arg. is widely used for ornamental purposes and is a potential source of bioactive compounds for cancer’s treatment [19]. This species is very attractive to the leaf-cutting ants, since it can complement the colony’s nutrition by liquids extracted directly from the leaves [16] [20]. Among its many varieties, Acalyphawilkesianahas the “Marginata” variety and the “Musaica” variety. The cultivar “Marginata” has coppery green leaves with a light pink variant borderand is known to be more attractive to the leaf-cutting ants when compared with the cultivar “Musaica”, which has green, red and orange spotted leaves [21].

On the other hand, the leaves of Taro (Colocasiaesculenta (L.) Schott var. esculenta) are less attractive to the leaf-cutting ants, classified as “moderately accepted” [22]. Previous works have shown that C. esculenta lectins have an anti-insect potential [23] and the ability of blocking cysteine proteases of fungal mycelium [24]. The lectins present in this plant are considered insecticidal agents [25] and a compound of its roots (2, 3-Dimethylmaleic anhydride) is biofumigant [26]. The Colocasiaesculenta (L.) Schott is the main species of the Araceae family, has rhizomes rich in starch and is used as human food [27]. Atta sexdens workers are fewly attracted for this plant, and its younger leaves are more cut compared to senescent leaves [28]. Preliminary tests in laboratory showed that the leaf-cutting ants Atta sexdens are more attracted to A. wilkesiana and less attracted to C. esculenta (pers. obs.). However, endophytic diversity can contribute to the choice of material foraged by leaf-cutting ants as demonstrated by Rocha et al. (2017) [8].

Therefore, in this study we aimed to verify how the presence of endophytic fungi could influence the foraging by the leaf-cutting ants in plants (Figure 1). For this, two vegetal species were used, one highly attractive and other less attractive and the following items were done: to 1) isolate and identify the fungi communities present in each cultivar of A. wilkesiana and in young and senescent leaves of C. esculenta, 2) compare the endophytic communities present in each plant, to verify a possible relation with the attractiveness of A. wilkesiana and less attractiveness of C. esculenta to the Atta sexdens ants.

2. Material and Methods

2.1. Collect, Isolation and Achievement of Pure Cultures

The study was carried out at the Center of Studies of Social Insects, in São Paulo State University (UNESP), Institute of Bioscience, Campus of Rio Claro, from June of 2018 to December of 2019. Individuals of Acalyphawilkesiana “Marginata”, Acalyphawilkesiana “Musaica” and Colocasiaesculenta variety esculentawere collected at Rio Claro, SP (22˚24'48"S, 47˚34'11"W) and kept in laboratory in 20-liter pots for two weeks. Three branches were collected from 4 individuals from each plant. After collection, the plant material was washed under running water. The external sterilization of the material was done by sequential immersion

in 70% (v/v) ethanol for 5 minutes, hypochlorite (3% - 5% active chlorine) for 5 minutes, 70% (v/v) ethanol for 30 seconds and twice in deionized sterile water (1 minute each). After removal of water excess, the leaves were cut into 0.5 cm squares, which were transferred to Petri Plaques containing Potato Dextrose Agar (PDA) culture medium. In each plaque, 4 leaf squares were placed. The plaques were incubated in B.O.D. in 25˚C and observed daily. Once detected the endophytic fungus growth, samples were taken from the growing hyphae to obtain axenic cultures. To confirm that the sterilization process was successful, aliquots of deionized water, used in the final wash, were plated by the plate spreading method, in PDA medium and incubated in 25˚C [29] (Figure 2).

2.2. DNA Extraction, PCR, Purification, Sequencing and Identification

DNA extraction was made following the CTAB method for filamentous fungi adapted from Möller et al. [30] and Gerardo et al. [31]. Each morphologic group was molecularly identified by the amplification and sequencing of the internal transcribed spacer region (ITS) of the ribosomal nuclear DNA.

For the amplification of the ITS region, it was used 1 μL (10 uM) of primers ITS1-F (forward: 5’-CTTGGTCATTTAGAGGAAGTAA-3’) [32], and ITS4 (reverse: 5’-TCCTCCGCTTATTGATATGC-3’) [33], 2 μL of DNA (1:49), 5 μL of GoTaq 5× Reaction Buffer, 4 μL of dNTPs (1.25 mM each), 0.2 μL of GoTaq

DNA polymerase (5 U/μL), 2 μL of MgCl₂ (25 mM), 1 μL of BSA (10 mg/mL) and adjusting the final reaction volume to 25 μL with sterile Milli-Q^® water. The reaction was conducted in thermocirculator Veriti™ 96 Well of Applied Biosystems according to the following program: 94˚C for 10 min, followed by 30 cycles of 94˚C for 1 min, annealing at 52˚C for 1 min and 72˚C for 2 min, and final extension at 72˚C for 10 min. The PCR amplified products were analyzed by electrophoresis in agarose gel 0.81%. The amplified fragments presented around 500 base pairs (Figure 2).

The PCR product’s purification was made with the illustra™ PCR DNA and Gel Band Purification Kit (GE Healthcare), according to the manufacturer’s protocol. The purified samples were quantified in the NanoDrop 2000 Thermo Scientific—Uniscience. The sequencing reactions were made with the BigDye Terminator kit, according to the manufacturer’s orientation. Sequences were obtained using the automatic sequencer 3130 Genetic Analyzer (Applied Biosystems) [34]. Forward and reverse sequences were edited in the Bioedit Sequence Alignment Editor [35] and aligned with the Clustal W tool [36] (Figure 2). With the Blastn tool from NCBI, the consensus sequences were compared to those with the higher similarity deposited on the GenBank. Those comparisons varied from 95% to 100% of similarity rate with E.value equal to 0.0. Only the sequences with more than 95% of similarity rate were considered in the comparisons. The communities of A. wilkesiana varieties were compared as one endophytic community with the communities of the C. esculenta leaf ages, which were considered as one community as well.

3. Results

From a total of 112 isolates, 82 endophytic fungi were isolated from the attractive plant A. wilkesiana, which represents 73% of the total of isolates, being 32% from “Marginata” and 43% from “Musaica”. After a morphological grouping of A. wilkesiana isolates, it was obtained 50 morphotypes. Some of the isolates are shown in Figure 3.

The varieties ofA. wilkesiana presented a bigger endophytic community, with 82 isolates, when compared to the C. esculenta community, with 30 isolates. In A. wilkesiana, nine isolates, being eight from the variety “Marginata” and 1 from the variety “Musaica”, could not be genetically identified due to impossibility of the ITS region sequencing or because of a low rate of similarity (<95%) with the date available at the GenBank. The identified isolates of the attractive plant (A. wilkesiana) are listed in Table 1, in which is evident that there are differences in the fungal composition of the endophytic communities of the Acalyphawilkesiana varieties, “Marginata” and “Musaica”. In the variety “Marginata”, the genus Alternariasp. and Diaporthe sp. were the most frequent (with four isolates each) and in the variety “Musaica”, the genus Colletotrichumsp. (with four isolates), Alternaria sp. and Nigrosporasp. (with three isolates each) were the most found in the community.

Thirty endophytic fungi were isolated from the leaves of C. esculenta, from which five were recovered from young leaves and thirteen from the senescent ones. Three isolates from young leaves and one from the older leaves could not be identified because, although the ITS amplification was successful, the sequencer was not able to sequence the TV6, TN2, TN4 and TN5 samples, and we were not able to obtain consensus sequences for comparison to GenBank.

In Table 2 are listed the identified genus from the isolates of young and old leaves of the less attractive plant C. esculenta. In the table, it is observable the difference in quantity and composition of the endophytic communities of the two leaf ages.

Comparison of the fungal communities.

Figure 1 shows the endophytic diversity of Acalyphawilkesiana (of both cultivars

“Marginata” and “Musaica” together) and Colocasiaesculenta (of both leaf ages together). A. wilkesiana presented a larger diversity of fungi genus when compared with the less attractive plant, C. esculenta. This difference can be verified in Figure 4 because of the different colors that composes the pie charts, since each color below represents a different endophytic fungus genus that was found in this work. It is also evident in Figure 4 that the plants share some genus, represented by the colors orange, light yellow and blue.

The two endophytic communities have some genus in common. The cultivar “Marginata” shares the genus Xylaria sp. with the two leaf ages of Taro (C. esculenta) and the genus Nemania sp. with the senescent leaves of Taro. The cultivar “Musaica” shares the genus Nigrospora sp. with the senescent leaves of the less attractive plant. Although both A. wilkesiana cultivars have the same number of fungi genus, they share only 5 of them, which are not present in C. esculenta leaves, being the most frequent Alternaria sp. and Colletotrichum sp., followed by Phomopsis sp., Diaporthe sp. and Curvulariasp. The genus Coniotyrium sp.

and Preussia sp. were found only in “Marginata” leaves, and Bipolaris sp., Cochiliobolus sp. and Pestalotiopsis sp. were found only in “Musaica” leaves. In C. esculenta, the less attractive plant, the genus found exclusively in its leaves were Fusarium sp. in the old leaves and Hypoxylon sp. in the young leaves.

4. Discussion

In the present study, it was done the analysis of the microbiota of cultivable endophytic fungi present in the plant species A. wilkesiana (“Marginata” and “Musaica”) and C. esculenta var. esculenta(young and senescent leaves). It was put in evidence that the endophytic fungi communities differ between the attractive plant to the leaf-cutting ants Atta sexdens (A. wilkesiana) and the less attractive plant (C. esculenta). Among the 300 thousand plant species described till today, each one is a host of, at least, one endophytic microorganism [37], what can constitute a type of plant defense against the leaf-cutting ant’s [7] and other pathogen microorganism’s attack, since the endophytic fungi can have an antibiotic action via its extracellular bioactive metabolites [38]. There are evidences that the leaf-cutting ants prefer entering the colony with leaves that have endophytic fungi with antifungal properties not directed to the symbiont Leucoagaricusgongylophorus, and by that protecting the colony from parasite fungi if necessary [14].

Overall, most of the fungal community isolated from A. wilkesiana is known by the production of bioactive compounds, including antibiotics [37] [39] - [44]. Phomopsis sp. is also known for having the capacity to survive the chemical and physical cleaning done by the ants before entering their colonies [45], besides being cited as a fungal genus well accepted by the attine ants, such as the genus Colletotrichumsp., Pestalopsis sp. And Xylaria sp. [14]. Probably, the attraction of a plant is related to the volatile compounds produced by the endophytic fungi. As these fungi can produce antibiotics that can harm the L. gongylophorus, the presence of determined compounds may signalize the toxicity or nutritional potential of a plant and influence in the decision of the ants to forage or not a plant.

This could indicate that the leaf preference by the Atta sp. ants is related to the presence of determined endophytic fungi, and not by the quantity of fungus present in a leaf, contradicting previous works that have put in evidence that these ants prefer leaves with less quantity of endophytic [46] [47], that, in this study, turned out to be C. esculenta, previously classified as less attractive leaves to the leaf-cutting ants. As Hypoxylon sp. and Fusarium sp. were genus found only in C. esculenta, they could be the ones causing the less attractiveness of the host plant to the leaf-cutting ants.

Hypoxylon sp. is a fungal genus that possess an antifungal and cytotoxic potential via its metabolic activities [48] and Fusarium sp., previously identified in the rejected material by the leaf-cutting ants, during the observations of Rocha and collaborators [14].Fusarium sp. was found only in the senescent leaves of C. esculenta, which has not presented fungi of the genus Diaporthe sp., found in A. wilkesiana. These two fungal genera are suggested to have antagonist actions toward each other [49], what can explain the presence of each one in different plants and suggest a relation between the attractiveness of Diaporthe sp. to the leaf-cutting ants, as it is a genus present in A. wilkesiana and not in C. esculenta.

Workers of A. sexdens explore more the plant’s apex when compared to its other regions, that being an innate behavior of the leaf-cutting ants [27]. The plant’s apex is the place where most of the young leaves are, which can provide more nutrients to the ants and, in C. esculenta, the young leaves have a smaller diversity of endophytic fungi when compared to the old leaves of the same plant. The ant’s choice for younger leaves could be related to the ease of cutting, lack of endophytes, their nutritional value or all these factors together, but it is a difficult question to be assessed [50]. Even though ants may remove more material from weaker and thinner leaves, the endophytic load does not differ within these traits, indicating that leaf-cutting ants are not influenced only by physical leaf’s attributes [47] but by multiple factors which are difficult to be assessed alone [46].

Coblentz and Van Bael [46] and Bittleston et al. [47], suggested that the leaf-cutting ants have a preference for leaves with less quantity of endophytic fungi and Van Bael, Estrada and Wscilo [51] found out that these ants prefer to cut young leaves of plants. The presence of different fungi in the analyzed plants suggests that the composition of the plant’s endophytic communities affects its attractiveness to the leaf-cutting ants, and not only the quantity of fungi as, according to Mighell and Van Bael [9], the ants respond differently to each endophytic fungi. They can alter the leaf’s chemical characteristic, which is associated with the ant’s preference for determined plants [52], and could explain why A. wilkesiana has a bigger quantity of endophytic fungi than C. esculenta, but a different diversity, and be more attractive to the leaf-cutting ants A. sexdens.

5. Conclusion

This study’s data suggest that the endophytic fungal communities of the attractive plant A. wilkesiana, and the less attractive plant, C. esculenta are different in the quantity of isolated fungi and in the genus found in each one, suggesting a relation between the endophytic mycobiota and the preference of cutting by the A. sexdens workers. Our results suggest that the leaf-cutting ants may prefer leaves that have specific endophytic fungi, that fungi such as Hypoxylon sp. and Fusarium sp. may be responsible for the less attractiveness of a plant and that the endophytic community composition can influence the attack of those plants by the leaf-cutting ants.

Acknowledgements

We thank Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) and Fundação para o Desenvolvimento da Unesp (FundUnesp); A.A.O thanks this study was financed in part by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior-Brasil (CAPES)-Finance Code 001; M.O.R. thanks U.S. National Science Foundation (NSF DEB 1900357) for financial support.