Diversity and effect of Trichoderma isolated from the roots of Pinus densiflora within the fairy ring of pine mushroom (Tricholoma matsutake)

Pine mushroom (PM, Tricholoma matsutake) is an important ectomycorrhizal fungus in Asia primarily due to its value as a food delicacy. Recent studies have shown that fairy rings of PM have distinctive fungal communities, which suggests that other fungi influence the growth of PM. Trichoderma is a well-known saprotrophic fungus commonly found in pine roots within PM fairy rings; however, little is known about the diversity of Trichoderma associated with PM and how these species influence PM growth. This study focused on diversity of Trichoderma isolated from pine roots within PM fairy rings and how these species affect the growth of PM isolate. Based on tef1a phylogenetic analyses, nine Trichoderma species (261 isolates) were identified. Trichoderma songyi and T. spirale were the dominant species, and Trichoderma community varied geographically. Growth experiments indicated that metabolites from five Trichoderma species had a significant influence on the growth of PM isolates. Metabolites of two Trichoderma species increased PM growth, while those of three Trichoderma species suppressed the growth. Within the fairy rings, Trichoderma that had a positive or neutral effect comprised the majority of Trichoderma communities. The results of this study suggest that various Trichoderma species co-exist within PM fairy rings and that these species influence PM growth.


Introduction
Ectomycorrhizal fungi have symbiotic relationships with host plants by supplying nutrients (e.g. nitrogen and phosphorus) and receiving carbohydrates in return [1]. Because the root is also inhabited by other microorganisms, dynamic microbial interactions can occur between ectomycorrhizal fungi and soil microorganisms in the root environment [2,3]. Recent studies that focused on bacteria-ectomycorrhizal fungi interactions revealed novel features of these inter-kingdom interactions [4,5]. For example, some bacteria negatively influence ectomycorrhizal fungi, while other bacteria promote growth. However, within the fungi, most studies (N37˚41 0 49@ E127˚53 0 19@) and a forested area in Uljin County (N36˚59 0 05@ E129˚06 0 09@) that is known for high PM production ( Fig 1A). Pinus densiflora was dominant species in both forests, and Quercus species were sparsely occurred with Rhododendron species as shrub vegetation. Soil type was granite-based sandy soil that was similar to other PM productive sites [19]. All sampling was conducted with permission from the National Institute of Forest Science. The front end of PM fairy ring where the PM hyphae are most actively grown was carefully chosen based on the morphological characteristics and abundance of PM ectomycorrhizae. Six lateral roots (> 10 cm in length) were collected from each of three PM fairy rings at each site. Root samples were transferred to the laboratory in an icebox at 4˚C. Soil and organic debris attached to the roots were removed by rinsing with distilled water. Roots were sterilized with 3% NaOCl for 3 min, and washed three times with sterilized distilled water. We confirmed PM colonization on pine roots from ectomycorrhizae attached to lateral root using sequences amplified with PM specific primers [33] and universal fungal primers ITS1F and ITS4 [34]. After detaching ectomycorrhizae, the roots were sliced into pieces of approximately 5 mm in length. Nine root pieces were placed on each Petri dish containing media. Two media were used for isolation: potato dextrose agar (PDA; Difco, USA) was used for isolating fast growing fungi and dichloran rose bengal chloramphenicol agar (DRBC; Difco, USA) was used for slow growing species. Three replicates were used for each fairy ring sample and each media using total of 324 root pieces (9 pieces × 3 plates × 2 media × 3 fairy rings × 2 sites). Plates were incubated at 25˚C for 2-7 days. Single fungal strains were generally obtained from each root piece and transferred to PDA medium. If multiple strains, indicated by distinctive morphology, were observed, individual strains were transferred to separate PDA media and subcultured until pure culture was obtained.

Molecular identification
Trichoderma isolates were determined by growth morphologies and microscopic features, and were grouped based on morphological characteristics. Among morphological groups, one to three strains from each Trichoderma group were selected as representative if possible. Genomic DNA was extracted from representative strains using a modified CTAB extraction method [35]. PCR was conducted in order to amplify the tef1a gene using the primers EF1-728F [36] and TEF1rev [37], using the conditions of Park et al. (2014). PCR products were checked on a 1% agarose gel and purified using Expin TM PCR Purification Kit (GeneAll Biotechnology, Seoul, South Korea). Sequencing was performed at Macrogen (Seoul, South Korea) using an ABI Prism 3730 genetic analyzer (Life Technologies, Gaithersburg, USA). Sequence analysis for proof reading and editing was conducted using MEGA v. 5 [38]. After aligning sequences using MAFFT v. 7 [39], a phylogenetic tree was constructed using a neighbor joining method and the Kimura-2-parameter model with 1,000 bootstrap replicates. If species identity of representatives were different within morphological groups, we conducted sequence analysis for all strains within the group. Sequences and phylogenetic tree with alignment were deposited at GenBank under accession numbers MG212672-MG212680 and at TreeBASE under accession number S21708, respectively. Community structures were compared by sampling site (Hongcheon vs. Uljin) and culture medium (PDA vs. DRBC) using Constrained Analysis of Principal coordinates (CAP) analysis based on weighted Unifrac dissimilarity with permutational ANOVA test for CAP model.

Effect of Trichoderma metabolites on PM growth
Effects of Trichoderma on PM growth were evaluated using a paper disc diffusion method with Trichoderma metabolite extract. Metabolites of Trichoderma were extracted from each Trichoderma species following a previous study [40] with minor modifications of culture conditions and solvent volumes. Trichoderma species were cultured on PDA in 90 mm Petri dish for 10 days. Five replicates from each culture plate were chopped in 300 mL of 80% methanol and incubated for a day. After filtering the solution through 150 mm Whatman filter paper (Advantec, Japan), the solvent was concentrated to 10 mL in vacuum using an EYELA rotary vacuum evaporator N-N series (Tokyo Rikakikai, Japan). 'Tricholoma matsutake' media (TMM) (glucose 20 g/L, yeast extract 1.5 g/L, soytone 1.5 g/L, and agar 20 g/L) [41]    isolate at the center of plate. A total of 50 μl of Trichoderma extract was inoculated on a sterilized paper disc (8 mm; Advantec, Japan) twice, and air dried in order to evaporate the methanol. Dried paper disc was placed 15 mm away from the center of the plate. All tests were performed in triplicate, and then incubated at 25˚C for 1 month. Radial growth (i.e. diameter) of PM isolates was measured twice and averaged. Differences in PM growth was compared between PM cultures with the metabolite disc (treatment) and culture with 80% methanol disc (control). Significance was tested using a pairwise Student t-test adjusted by the false discovery rate of Benjamini and Hochberg [42].

Species identification and composition
A total of 376 fungal isolates were obtained from root pieces, and 261 isolates were identified as Trichoderma species based on morphological characters (Fig 1B). Trichoderma species were isolated from most of the lateral roots within the six fairy rings, and nine distinct species belonging to four clades were identified using tef1a sequence analysis (Fig 1C). Six of these were identified at the species level based on the phylogenetic tree, but we were unable to identify three Trichoderma species due to ambiguous phylogenetic relationships. The largest number of Trichoderma species was found in Harzianum clade (5 species) followed by Viride clade (2 species). Trichoderma spirale was the most dominant species (n = 97), followed by T. songyi (n = 56) and T. hamatum (n = 52) ( Table 1). The number of Trichoderma species or strains was similar between culture media. Among non-Trichoderma strains, a total of 22 species were identified, and Penicillium had the largest number of species (7 species) followed by Mortierella (3 species) (S1 Table). The number of strains was largest in Penicillium (n = 69) followed by Umbelopsis (n = 18) (S1 Table). CAP analysis based on weighted Unifrac dissimilarities showed that Trichoderma communities were significantly different between the sampling sites (P = 0.002; 61.5% explanatory power) (Fig 2A), while it were not different between culture media (P = 0.961; 0.3% explanatory power). In the Hongcheon samples, T. spirale (57.8%) was the most dominant species, followed by T. crassum (14.5%) and T. hamatum (9.6%) (Fig 2B). In Uljin samples, however, T. songyi (50.5%) was the most dominant species, followed by T. hamatum (37.9%). Only four species were isolated from both sites: T. hamatum, T. songyi, T. spirale, and Trichoderma sp. 2 ( Fig 1C; Table 1). The geographical distribution of non-Trichoderma species was different between Hongcheon and Uljin (S1 Table). The number of non-Trichoderma strains was higher in Uljin (n = 91) than in Hongcheon (n = 24). At the fairy ring level, the Hongcheon 3 (n = 16) and Uljin 1 fairy rings (n = 37) showed largest number of non-Trichoderma strains within each sampling sites. Umbelopsis nana was most abundant in Hongcheon (n = 9), and Penicillium bissettii was in Uljin (n = 32).

Effect of Trichoderma metabolite on the growth of PM
Five Trichoderma species showed significant effects on PM growth among the nine Trichoderma species (Fig 3A). Two species (T. songyi and T. spirale) had a positive effect on PM growth: the treatment increased growth by 168-175% of PM grown alone (control); while three species (T. hamatum, T. polypori, and T. pyramidale) had a negative effect: the treatment decreased growth by 25-41% of control. The Trichoderma species showing a positive effect on PM growth were belonging to Spirale (T. spirale) and Viride clade (T. songyi) (Fig 1C). The species showing negative effect were belonging to Harzianum (T. polypore and T. pyramidale) and Viride clade (T. hamatum). Trichoderma species were categorized based on their effect on PM growth, and the predicted proportion of effect type (i.e. positive or negative effect) in Trichoderma communities are presented in Fig 3B. The predicted proportion of Trichoderma that had a negative effect was less than 50% in all sampling site, except for the Uljin 1 fairy ring. At the fairy ring level, Hongcheon 3 (34.9%) and Uljin 1 fairy rings (76.0%) had largest proportion of negative Trichoderma species within each sampling locations.

Trichoderma diversity in the pine roots under PM fairy ring
Trichoderma species are cosmopolitan saprotrophic fungi and are a major component of the mycoflora community in forest soils [43,44]. Most studies of Trichoderma focused on their roles as pathogens [37,45,46] or biocontrol agents [23,47,48], while relatively little is known about the ecology of individual Trichoderma species associated with ectomycorrhizal fungi. Part of the reason for this may be the taxonomic difficulties often encountered within this genus. Traditional taxonomy of Trichoderma was based on morphological characters, however, these morphological characters are often insufficient to differentiate species [37] or unreliable because they can change based on environmental conditions [49]. In recent years,  sequence analysis of the nuclear ribosomal internal transcribed spacer region (ITS), the second largest unit of the RNA polymerase II (rpb2), and the tef1a region have improved species identification [50][51][52] within the Trichoderma. Approximately 250 species of Trichoderma have been detected based on a combination of genetic and morphological data to date [53].
Trichoderma was the most dominant genus in the fungal communities we studied (69.4%) (Fig 1B), which suggests that PM share habitats with Trichoderma and they may interact together. A total of nine Trichoderma species were identified, and T. hamatum, T. songyi, and T. spirale were isolated most frequently (Table 1). Among these, the initial discovery of the species T. songyi was in a PM-associated environment: soil and root within the fairy rings and fruiting bodies of PM [31]. Moreover, five Trichoderma species (T. polypori, T. pyramidale and three unidentified Trichoderma species) detected in this study belong to the Harzianum clade exhibiting fungicolous characteristics such as colonizing mushroom-associated environments like fruiting bodies and mushroom composts [54,55]. In addition, some Trichoderma species in the Harzianum clade (e.g. T. harzianum) showed fungistatic effects on the growth of white button mushrooms (Agaricus bisporus) [56]. These negative effects on fungal growth among the several species in the Harzianum clade are consistent with our results of T. polypori and T. pyramidale (Fig 3A).
Previous study showed that various saprotrophic fungi co-existed with PM colonized rootlets of Pinus densiflora, while Trichoderma was not found [22], likely due to difference of sampling strategy; lateral roots were used in this study, while rootlets were used in You et al. [22]. In this study, non-Trichoderma species were also isolated (S1 Table) and had similar species composition with previous studies conducting from PM fairy rings [16,17,22]. As Penicillium was most abundant species in You et al. [22], the number of Penicillium species and strains was highest among non-Trichoderma species (S1 Table). In addition, Mortierella and Umbelopsis have been abundantly detected previously from PM fairy ring [16,20,22], and we isolated three Mortierella and two Umbelopsis species. Therefore, the characteristics of fungal species isolated in this study is similar to it of previous studies, except for Trichoderma abundance. According to the number of strains (S1 Table), PM fairy rings with low abundance of Trichoderma showed high number of non-Trichoderma strains, which suggests that antagonistic relationship between Trichoderma and non-Trichoderma species can be one of the reasons for relatively small number of non-Trichoderma species in this study.
Trichoderma communities isolated in this study showed geographical differences (Fig 2A). Among the nine species, only four (T. hamatum, T. songyi, T. spirale, and Trichoderma sp. 2) were isolated from both locations (Table 1). On the other hand, the distribution of dominant species showed a clear pattern based on geography (Fig 2B). Trichoderma spirale was most frequently isolated from the Hongcheon site, while T. hamatum and T. songyi were more frequently isolated from the Uljin site. Geographical differences in Trichoderma distribution have been detected in previous studies in China and Tunisia, and this spatial variation may be associated with environmental conditions such as climate, soil properties, and vegetation [57,58]. Soil fungal communities within PM fairy rings were also significantly different depending on geographical location [20], which suggests the importance of geographic effect on the fungal community in the fairy ring of PM.

Effect of Trichoderma metabolite on PM growth
Trichoderma have been well studied because of their beneficial interactions with plants [23]. In contrast to their generally positive effects on the plants, most studies suggest that Trichoderma usually have a negative effect on other fungi [25,26,59]. It has been shown that Trichoderma exhibit biocontrol activity in suppressing plant pathogenic fungi (e.g. Fusarium oxysporum, Pythium ultimum, and Rhizoctonia solani) using cell wall degrading enzymes and secondary metabolites [26]. In addition, Trichoderma does seem to have some level of mycoparasitic ability on an arbuscular mycorrhizal fungus (Glomus intraradices) via the penetration of mycelium [25]. In the case of macrofungi, T. pleuroti and T. pleuroticola cause green mold disease in the oyster mushroom (Pleurotus ostreatus) [46]. In this study, three Trichderma species, T. hamatum, T. polypori, and T. pyramidale, exhibited strong antifungal effects on PM growth ( Fig  3A). Although nothing is known of the antifungal activity and metabolite secretions of T. polypori and T. pyramidale because these are recently recorded species [54,60], T. hamatum has been applied for the suppression of other fungi [61][62][63] and also secretes secondary metabolites with antifungal properties (e.g. gliotoxin, isonitrin, and viridiol) [64][65][66]. Therefore, antifungal substance in secondary metabolites from T. hamatum, as well as T. polypori and T. pyramidale, may suppress PM growth.
Compared to the effects of Trichoderma on saprotrophic or pathogenic fungi, the relationship between Trichoderma and ectomycorrhizal fungi has been relatively overlooked [6]. Although some Trichoderma have positive or neutral relationship with ectomycorrhizal fungi [7,27,28], most previous studies suggested that Trichoderma suppressed the growth of hyphae and mycorrhization of ectomycorrhizal fungi [6,9,29,30]. Our results showed, however, that the proportion of Trichoderma that had a negative effect on PM was low in the fairy rings ( Fig  3B). In addition, the PM growth promoting fungi, T. songyi and T. spirale, were the dominant species in the Hongcheon and Uljin sites, respectively. Therefore, the relationship between Trichoderma and PM is not restricted to negative impacts, as many of these species obviously have significant positive effects. In both sampling locations, PM growth promoting Trichoderma species were dominant, except for Uljin 1 which had a large number of Trichoderma that inhibit growth. This unusual proportion of effect type in the Uljin 1 fairy ring suggests that the PM in this site may be in poor health or that other fungi act as PM growth promoting fungi. Given that various bacterial genera can promote PM growth [67][68][69], PM may be benefited by other fungi that are not belonging to Trichoderma. However, it needs to pay attention to interpret the proportion of effect type in Trichoderma communities because we used single strain from each Trichoderma species, thus intraspecific variation of effect on PM growth was ignored in this study. Therefore, a better understanding of the relationship of PM with multiple strains of Trichoderma species, as well as non-Trichoderma fungi is needed in order to better understanding the overall microbial interactions associated with PM.

Conclusions
In conclusion, nine Trichoderma species were isolated from the roots of Pinus densiflora within PM fairy rings. Trichoderma was dominant and community structure was significantly influenced by geographical locations, which suggests that PM have an intimate relationship with various Trichoderma species. In the PM growth experiment, the metabolites from T. songyi and T. spirale had a positive effect on PM mycelial growth. Therefore, our results suggest that saprotrophic fungi can have positive effects on the physiology of ectomycorrhizal fungi, and that introduction of Trichoderma metabolites may improve the prospects of successful PM cultivation in the future.
Supporting information S1