Molecular Phylogeny of Endophytic Fungi from Rattan (Calamus castaneus Griff.) Spines and Their Antagonistic Activities against Plant Pathogenic Fungi

Calamus castaneus is a common rattan palm species in the tropical forests of Peninsular Malaysia and is noticeable by the yellow-based spines that cover the stems. This study aimed to determine the prevalence of fungal endophytes within C. castaneus spines and whether they inhibit the growth of fungal pathogens. Twenty-one genera with 40 species of fungal endophytes were isolated and identified from rattan palm spines. Based on molecular identification, the most common isolates recovered from the spines were Colletotrichum (n = 19) and Diaporthe spp. (n = 18), followed by Phyllosticta spp., Xylaria sp., Trichoderma spp., Helminthosporium spp., Penicillium spp., Fusarium spp., Neopestalotiopsis spp., Arthrinium sp., Cyphellophora sp., Cladosporium spp., Curvularia sp., Bionectria sp., and Acremonium spp. Non-sporulating fungi were also identified, namely Nemania primolutea, Pidoplitchkoviella terricola, Muyocopron laterale, Acrocalymma fici, Acrocalymma medicaginis, and Endomelanconiopsis endophytica. The isolation of these endophytes showed that the spines harbor endophytic fungi. Most of the fungal endophytes inhibited the growth of several plant pathogenic fungi, with 68% of the interactions resulting in mutual inhibition, producing a clear inhibition zone of <2 mm. Our findings demonstrate the potential of the fungal endophytes from C. castaneus spines as biocontrol agents.


Introduction
Endophytic fungi are ubiquitous and found in almost all plant parts, including stems, leaves, and roots, and colonize the host plants without causing any disease symptoms throughout their life cycle [1]. These microorganisms have shown the potential to enhance host resistance to pathogens and pests as well as tolerance to abiotic stress [2]. Bilal et al. (2008) [3] reported that endophytic Aspergillus fumigatus and Fusarium proliferatum produce growth regulators and promote plant growth under abiotic conditions. Some endophytic fungi have been reported to improve plant growth and reduce the severity of plant diseases; therefore, these fungi have the potential to be used in plant disease management strategies [4]. For example, fungal endophytes from cocoa (Theobroma cacao) inhibit the growth of several major pathogens of the crop [5]. Endophytic fungi may be antagonistic and inhibit the growth of other fungi, and many have been reported as potential biocontrol agents [5,6]. Biological control using endophytic fungi is an alternative method for sustainable plant disease management and contributes to environmental conservation.
Plants use several sharp structures, such as spines, thorns, and prickles, for defense. Spines are modified leaves, whereas thorns are a modification of branches, and prickles result from the outgrowth of cortical tissues in the bark [7]. Calamus castaneus Griff. is a common rattan species that grows in the Malaysian tropical rainforest and is classified in the palm family, Palmae or Areceae. Calamus castaneus is recognized by its yellow-based spines, which cover the stems and the middle part of the upper leaves. The spines are arranged as a single line on the stem, while at the bottom of the leaves, the spines are arranged in two parallel lines [8]. These sharp structures may harbor various types of fungi as the presence of endophytic fungi, particularly dermatophytes in spines, thorns, and prickles, has been reported by Halpern et al. (2011) [9]. As C. castaneus is common and relatively easy to find in the forests, studying the presence of endophytic fungi in the spines of this rattan species is of interest. Novel endophytic fungal isolates that have the potential to be developed as biocontrol agents against several plant pathogenic fungi might also be recovered from spines of C. castaneus. As there is a lack of information on the fungal endophytes from spines, the objectives of this study were to determine the occurrence of endophytic fungi in the spines of C. castaneus and identify the endophytic fungi through molecular methods. The antagonistic activity of the fungal endophytes from the spines to inhibit growth of several plant pathogenic fungi was also tested using a dual culture method. Knowledge on the endophytic fungal community in spines of C. castaneus contributes to in-depth information on the occurrence of fungal endophytes in various plant parts as well as identifying potential biocontrol agents against plant pathogens.

Sample Collection and Isolation of Endophytic Fungi
The spines of C. castaneus were randomly collected from rattan trees found in three rainforests, in two states of the Peninsula Malaysia, namely in Bukit Panchor State Park, Penang (5. 30.7 E). The spines were kept in an envelope and transported to the laboratory. The spines were placed in a beaker, covered with a net cloth, and placed under running tap water overnight to remove any debris, dirt, and epiphytes adhered to the surface. Thereafter, the spines were surface sterilized by soaking in 70% ethanol for 5 min, followed by 5% sodium hypochlorite (NaOCl) for 5 min. Then, the samples were washed with sterile distilled water three times for 2 min and blotted dry using sterile filter papers to remove excess water. The sterilized spines were plated onto potato dextrose agar (PDA, HiMedia Laboratory, Maharashta, India) plates and incubated at room temperature (27 ± 1 • C) until there was visible mycelial growth from the spine tissues ( Figure 1). Sixty spine samples were used for isolation. The efficiency of the surface sterilization technique was determined using an imprint method [1]. The surface sterilized spines were imprinted or dabbed on the surface of a PDA plate and the plate was incubated at room temperature. Surface sterilization is considered effective if no fungal colony grows on the imprint plate. Mycelia growing from the spine tissue were sub cultured onto new PDA plates. A pure culture of the isolate was obtained using the spore suspension method and the plates were incubated at room temperature for seven days.
The fungal isolates were sorted into their respective groups or genera based on the appearance of the colonies and microscopic characteristics.

DNA Extraction and PCR Amplification
The fungal isolates were grown in potato dextrose broth and incubated at room temperature for six days. Mycelia were harvested and ground with liquid nitrogen in a sterile mortar and pestle to a fine powder. The DNeasy ® Plant Mini kit (Qiagen, Hilden, Germany) was used to extract genomic DNA, according to the manufacturer's instructions.
The internal transcribed spacer (ITS) region was used to identify all endophytic fungal isolates recovered from the spines except Xylaria. The primers used were ITS1 and ITS4 [10]. After amplification of the ITS, species identity was obtained based on the basic local alignment search (BLAST) and a combination of at least two genes/regions was used for further confirmation of the species (Table 1). However, for several fungal genera, the analysis of the ITS region was not sufficient to differentiate closely related species.  A 1% agarose gel (Promega, Middleton, WI, USA) was used to detect the PCR products in 1 × Tris-Borate-EDTA (TBE) buffer stained with FloroSafe DNA stain (Axil Scientific, Singapore). PCR products were sent to a service provider for Sanger DNA sequencing.

Molecular Identification and Phylogenetic Analysis
The DNA sequences were aligned manually and edited using the Molecular Evolution Genetic Analysis version 7 (MEGA7 version 7) [18]. Forward and reverse sequences were aligned with ClustalW using pairwise alignments. The aligned forward and reverse sequences were edited when necessary to form a consensus sequence. For species identity, a BLAST search was used to analyze the number of bases and determine the maximum identity of the consensus sequences from the GenBank database.
A phylogenetic analysis was also conducted, particularly for species that are known to belong to a species complex or for isolates whose ITS sequences cannot be used to confidently identify the isolates to the species levels. Multiple sequence alignments were generated and used to construct phylogenetic trees based on combined sequences. A maximum likelihood (ML) tree was constructed with 1000 bootstraps replicates. The heuristic method used in ML was the nearest neighbor interchange (NNI) and the initial tree for ML was generated automatically. The best model for ML tree was determined from the model search with number of discrete gamma categories 5. The results show that the Kimura 2 parameter model was the best model. Missing data or gaps were treated as complete deletion.

Antagonistic Activity
The ability of the fungal endophytes to inhibit the mycelial growth of several plant pathogenic fungi was determined with a dual culture method using PDA. Several endophytic fungi from C. castaneus spines were selected to assess their antagonistic activity against several plant pathogenic fungi. The endophytic fungi were chosen based on fungal genera or species that have been reported as antagonists against plant pathogens, such as Xylaria cubensis, Penicillium indicum, Penicillium oxalicum, Trichoderma harzianum, and Trichoderma koningiopsis. Endophytic fungal species that have not been reported as antagonists were also tested, namely Endomelanconiopsis endophytica, Neopestalotiopsis saprophytica, Colletotrichum endophytica, Colletotrichum siamense, Colletotrichum boninense, Diaporthe arengae, Diaporthe tectonae, Diaporthe cf. nobilis, and Diaporthe cf. heveae.
Selected plant pathogenic fungi were obtained from the culture collection at the Plant Pathology Laboratory, School of Biological Sciences, Universiti Sains Malaysia, Penang, Malaysia. The pathogenic fungi included two anthracnose chili pathogens, C. truncatum and C. scovellei; two pathogens that cause dragon fruit stem rot, Fusarium proliferatum and F. fujikuroi; and F. solani and F. oxysporum, which are associated with crown disease in oil palm. Four pathogens associated with mango diseases were also included: Lasiodiplodia theobromae and Pestalotiopsis mangiferae, which are the causal pathogens of the mango leaf spot, and L. pseudotheobromae and D. pascoei, which cause mango stem-end rot.
A combination of the endophytic fungi and plant pathogenic fungi tested in dual culture test is shown in Table 2. A control plate harbored only plant pathogenic fungi without the endophytes. Mycelial plugs (5 mm) of the pathogen and endophyte were cultured 6 cm apart. The plates and three replications were incubated at room temperature for seven days. The experiment was repeated twice.
After seven days, the percentage of the pathogen growth inhibition (PGI) was calculated according to the method described by Skidmore and Dickinson (1976) [19]: R1-radial growth of plant pathogenic fungi in control plate. R2-radial growth of plant pathogenic fungi in dual culture plate. R1 was measured from the point of inoculation to the pathogen colony margin on the control plate and R2 was measured from the point of inoculation to the colony margin on the dual culture plate in the direction of the endophytes.
Statistical analysis of the PGI value was performed using ANOVA in SPSS statistical software version 24. Interactions between plant pathogens and endophytic fungi were assigned in a range of interactions from types A to E, according to the interactions described by Skidmore and Dickinson (1976) [19]. Type A interactions occurred when the pathogens and endophytic fungi displayed intermingling growth; type B interactions represented the overgrowth of pathogens by endophytic fungi; type C interactions represented the overgrowth of endophytic fungi by pathogens; type D interactions represented mutual inhibition with a clear inhibition zone at small distance (<2 mm); and type E interactions represented mutual inhibition with a clear inhibition zone at a greater distance (>2 mm).

Molecular Identification
A total of 108 isolates of endophytic fungi comprising 21 genera with 40 species were recovered from the C. castaneus spines (Table 3). Fungi isolated from the spines were confirmed as endophytes as no fungal growth on the imprinted plates was observed. The imprint method was used as an indication that the epiphytes from the surface of the spines had been removed. A successful and correct procedure of surface sterilization removes epiphytes from the surface of the spines, which results in no fungal growth and must be used in all studies concerning endophytes [20,21]. Table 3. Molecular identification of endophytic fungi isolated from C. castaneus spines. Table 3. Cont.

Genbank Accession Number
Isolates Note: Colletotrichum endophytica is synonymous with Colletotrichum endophyticum.
Endophytic fungal species recovered from C. castaneus spines identified using ITS and other additional markers are shown in Table 2. Most of the isolates were successfully identified to the species levels except for three isolates of Diaporthe. The most common isolates recovered from the spines were Colletotrichum spp. (n = 19) and Diaporthe spp. Six species of Colletotrichum were identified using ITS and GAPDH sequences, namely C. horii (n = 4), C. siamense (n = 3), C. fructicola (n = 2), C. cliviae (n = 2), C. endophytica (n = 7), and C. boninense (n = 1) ( Table 3). All the species identified are members of the C. gloeosporioides species complex. In addition to ITS, the GAPDH gene was included as an additional marker as the gene is among the most effective secondary markers to distinguish species in the genus Colletotrichum. Moreover, GAPDH is the easiest gene to amplify and sequence [22,23]. The phylogenetic analysis showed that isolates from the same species were grouped in the same clade as their epitype strains (Figure 2), which confirmed the identity of the endophytic Colletotrichum species obtained from C. castaneus spines.

Antagonistic Activity
In general, most of the endophytic fungi from C. castaneus spines inhibited mycelial growth of the plant pathogenic fungi tested (Table 4). Only three species of Diaporthe, D. cf. nobilis, D. cf. heveae, and D. tectonae, as well as two isolates of X. cubensis did not show antagonistic activity against L. theobromae and L. pseudotheobromae (Table 4). Both pathogens overgrew the endophytic fungi as L. theobromae and L. pseudotheobromae are fast growing fungi able to compete for space and nutrients. Based on the observation of the dual culture plates, the most common interactions between the fungal endophytes and plant pathogenic fungi were type D interaction, which is mutual inhibition with a clear inhibition zone (<2 mm).
Both endophytic T. harzianum and T. koningiospsis overgrew the pathogens on the 7th day of incubation. Endomelanconiopsis endophytica and D. tectonae moderately inhibited all tested plant pathogens ( Figure 11). The results showed that the pathogens were lysed and subsequently killed as no growth was observed when the hyphae from the contact point of both fungi in the dual culture test were transferred onto PDA. A high percentage of growth inhibition was shown by the endophytic T. harzianum and T. koningiopsis that inhibited the mycelial growth of all tested plant pathogens (Table 4).

Discussion
A total of 108 isolates of endophytic fungi comprising 21 genera with 40 species were recovered from C. castaneus spines. The results showed that endophytic fungi residing in the spines are mostly Ascomycetes, class Sardariomycetes, order Glomerellales (Colletotrichum), Diaporthales (Diaporthe), Xylariales (Xylaria), Hypocreales (Trichoderma, Fusarium), as well as several other classes and orders. The present study demonstrated that endophytic fungi isolated residing in C. castaneus spines may be considered as cosmopolitan fungal isolates.
The endophytic fungi from C. castaneus spines were identified using ITS and other suitable markers. Despite the advantages of the ITS region for fungal identification, the region may not be useful to distinguish species in a species complex or closely related species, such as Colletotrichum and Diaporthe. This may be due to lower sequence variation in many closely related species, the presence of sequence heterogeneity among the ITS copies, and the inability of some groups of fungi to amplify the ITS region resulting in poor sequencing success [24,25]. Hence, several genes were also used to accurately identify the fungal isolates and for phylogenetic analysis. The gene chosen depends on the fungal genera; TEF-1α, β-tubulin, GAPDH, and ACT genes were used in this study. Introns in protein-coding genes are highly variable, which make them useful for species identification and phylogenetic analyses. Several of these genes are considered secondary barcode markers with adequate intra-and interspecies variation often used as part of identification using multiple gene phylogeny [25].
The endophytic fungal species from genera Colletotrichum, Trichoderma, Penicillium, Phomopsis, Phyllosticta, and Xylaria are among common fast-growing culturable fungi, which might be one of the reasons these genera were mostly recovered as endophytic fungi from the spines. Moreover, the methods used in this study were culture-dependent methods of which only culturable isolates were recovered from the spines. In culture-dependent methods, several growth parameters including temperature, light, nutrient, and aeration contribute to the growth of the endophytic fungi [31]. By using culture-dependent methods, fast-growing fungal isolates commonly inhibit the growth of slow-growing isolates and thus many fast-growing fungi were recovered [32]. Unculturable endophytic fungi could not grow or were difficult to grow on culture media. Thus, unculturable endophytic fungi are commonly analyzed using culture-independent methods such as denaturing gradient gel electrophoresis and high-throughput sequencing methods [33,34]. These methods can directly amplify endophytic fungi residing in the plant tissues.
Colletotrichum spp. (n = 19) and Diaporthe spp. (n = 18) were the most common endophytes isolated from C. castaneus spines. Species from both genera have been reported as endophytes in the roots, leaves, and stem of several plants, including mangrove tree leaves (Acanthus ebracteatus and Phoenix paludosa) [35], leaves of Sapindus saponaria [36], and twigs of a woody tree (Acer truncatum) [37]. Therefore, the endophytic fungal species from both genera isolated from C. castaneus spines are similar to those previously reported from other types of plants that harbor fungal endophytes [35][36][37].
Although numerous endophytic species from C. castaneus spines are common endophytes, several species have not been reported as endophytes from any plant. These endophytes are P. carochlae, P. indicum, Arthrinium urticae, C. guyanensis, A. hennebertiiennebertii, and P. terricola. Among these endophytic fungi, P. terricola is a rare species and was only reported in the rhizosphere of Quercus rubra in Ukraine [38] and from earthworm casts in Domica Cave, Slovakia [39].
Dermatophytes of animals and humans have been reported from spines, thorns, and prickles [40]. Dermatophytes causing subcutaneous mycosis and infection may occur by inoculation of the dermatophytes into subcutaneous tissues by penetration of spines and thorns [41,42]. Among the dermatophytes from plants, Fonsecaea pedrosoi was reported in thorns of Mimosa pudica isolated from the site of infection [43]. Cladophialophora carrionii has also been isolated from plants. Another dermatophyte, Sporothrix schenckii, is commonly transmitted through a prick from roses [44,45]. However, in the present study, dermatophytes were not recovered from C. castaneus spines, which might be due to different host plants, environmental conditions, and geographical location. These factors may contribute to the endophytic fungi occurrence and diversity in the host plant [46,47].
An antagonistic activity assay was conducted to assess the ability of the fungal endophytes from C. castaneus spines to be used as antagonists that inhibit the growth of plant pathogens. Among the endophytic fungi recovered from C. castaneus spines, T. harzianum, and T. koningiospsis highly inhibited growth of all tested plant pathogens. Other endophytic fungi tested produced low to moderate inhibition. The results of the present study indicated endophytic T. harzianum and T. koningiopsis showed strong antagonistic effects against all the pathogens tested and successfully inhibited the growth of the pathogens. Trichoderma harzianum has been reported to inhibit growth of C. truncatum, causal pathogen of strawberry anthracnose [48], and mango anthracnose [49]. So far, there are no reports on antagonistic activity of T. koningiopsis against anthracnose pathogens, but this species has strong antagonistic activity against F. oxysporum, Rhizoctonia solani, and Botrytis cinerea that infected tomato and cucumber seedlings [50]. Trichoderma koningiopsis was also reported as strong antagonistic fungus, showing 85% growth inhibition of Calonectria pseudonaviculata causing blight of boxwood plant [51].
Several reports are available on the antagonistic activity of T. harzianum against plant pathogenic Fusarium spp. Trichoderma harzianum inhibited growth of F. proliferatum, causing basal rot of onion bulb [52] and stalk rot of maize [53] as well as inhibiting growth of F. solani, causal pathogen of root rot of olive tree [54]. As for T. koningiopsis, this fungus exhibited strong antagonistic activity against F. proliferatum, causal pathogen of soybean damping-off [55].
As one of the effective antagonistic fungi, Trichoderma spp. have several mechanisms of inhibition, which include competition for space and nutrients, antibiosis by secretion of antifungal compounds, mycoparasitism, and induced resistance [56]. These mechanisms may occur with T. harzianum and T. koningiospsis as both grew faster than the pathogens.
Endomelanconiopsis endophytica and D. tectonae may also be considered as effective antagonistic fungi. Both endophytic fungi moderately inhibited the mycelial growth of all tested plant pathogens except for L. theobromae and L. pseudotheobromae, whereby both pathogens grew faster than the endophytes. The inhibition mechanisms might be similar to that of Trichoderma spp., in which the mycelial growth of the tested pathogens was inhibited by competition, antibiosis, or mycoparasitism.
Antagonistic activity of E. endophytica against other plant pathogenic fungi has not been reported, but in a study by Ferreira et al. (2015) [26], the extract of this endophytic fungus displayed trypanocidal activity against amastigote forms of Trypanosoma cruzi. For endophytic D. tectonae, this fungus moderately inhibited growth of Phytopthora palamivora, pathogen of cocoa black pod [57].
Endophytic fungi residing in the spines exhibited antagonistic activity, indicating their ability to produce bioactive compounds. These bioactive compounds may be involved in defense mechanisms against pathogen infections, chemical defense [6,58], and adaption and survival in the host plant [26].
As a conclusion, a total of 108 isolates of endophytic fungi were isolated from C. castaneus spines and 40 species were identified. The results demonstrate that C. castaneus spines harbor diverse groups of endophytic fungi with an antagonistic activity against several plant pathogenic fungi. Among the endophytic fungi, T. harzianum and T. koningiopsis inhibited all plant pathogens tested with a high percentage of inhibition. The antagonistic activity against plant pathogenic fungi indicated that the endophytic fungi have the potential to be developed for use as biocontrol agents. Therefore, further studies should be performed to detect and identify bioactive compounds produced by the endophytic fungi as well as to understand the mechanism the endophytes used to inhibit the pathogen growth. To the best of our knowledge, the present study is the first to determine the occurrence and diversity of filamentous fungi in spines of rattan palm.