Temporal Variation in Community Composition of Root Associated Endophytic Fungi and Carbon and Nitrogen Stable Isotope Abundance in Two Bletilla Species (Orchidaceae)

Mycorrhizae are an important energy source for orchids that may replace or supplement photosynthesis. Most mature orchids rely on mycorrhizae throughout their life cycles. However, little is known about temporal variation in root endophytic fungal diversity and their trophic functions throughout whole growth periods of the orchids. In this study, the community composition of root endophytic fungi and trophic relationships between root endophytic fungi and orchids were investigated in Bletilla striata and B. ochracea at different phenological stages using stable isotope natural abundance analysis combined with molecular identification analysis. We identified 467 OTUs assigned to root-associated fungal endophytes, which belonged to 25 orders in 10 phyla. Most of these OTUs were assigned to saprotroph (143 OTUs), pathotroph-saprotroph (63 OTUs) and pathotroph-saprotroph-symbiotroph (18 OTUs) using FunGuild database. Among these OTUs, about 54 OTUs could be considered as putative species of orchid mycorrhizal fungi (OMF). For both Bletilla species, significant temporal variation was observed in the diversity of root endophytic fungi. The florescence and emergence periods had higher fungal community richness of total species and endemic species than did other periods. Both Bletilla species were dominated by Agaricomycetes and Basidiomycota fungi throughout the whole year; however, their abundances varied between two Bletilla species and among phenological stages. Meanwhile, the ranges of 13C and 15N natural abundance were also highly dynamic across all growth stages of Bletilla species. Compared with the surrounding autotrophic plants, significant 13C enrichments (ε13C) were found across all phenological stages, while significant 15N enrichment in the florescence period and strong 15N depletion during the fruiting period were found for both Bletilla species. We can deduce that both Bletilla species obtained carbon from root endophytic fungi during the whole year. Additionally, the temporal varying tendency of root endophytic fungal diversity was consistent with 13C enrichments, which was also accord with the nutritional requirement of plant.


Introduction
Orchidaceae is one of the largest plant families, including almost 10% of all angiosperm species and about 28,000 species comprising 736 genera worldwide [1,2]. Members of this family are widely distributed among terrestrial ecosystems, with the notable exception of extremely cold and dry areas and accordingly have a great variety of life history strategies, ranging from epiphytic to terrestrial and from evergreen to achlorophyllous species [3,4]. Zettler et al. reported that unique characteristics of the orchid family, including its high tures that may reflect changes in plant nutritional requirements and variation in nutrient availability throughout the growing season. Meanwhile, mycorrhizal community compositions are often studied in steady-state situations [14,33,36]. It is unclear whether these fungal communities are temporally stable and whether seasonal changes in mycorrhizal associations could also lead to differences in the nutritional status of orchids.
In this study, we used stable isotope natural abundance analysis and DNA molecular identification methods to investigate temporal variation in C and N stable isotope abundance and root endophytic fungi assemblages associated with two mature greenish Bletilla spp., Bletilla striata (Thunb. ex A. Murray) Rchb. f. and Bletilla ochracea Schltr. Our main objectives are to answer the following questions: (1) What are the root endophytic fungi of the two Bletilla spp.? (2) Do the composition and abundance of root endophytic fungi and C and N stable isotope abundance of Bletilla vary among phenological stages or between the two species?

Temporal Variation in Diversity of Root Endophytic Fungi
Fungal community richness (Sobs) and alpha diversity (Shannon and Simpson's) index values were compared among five phenological stages for the two Bletilla species at the operational taxonomic unit (OTU) level ( Figure 2). For B. striata, the florescence period had significantly higher Sobs and Shannon index values compared with the fruiting and dormancy periods. For B. ochracea, the emergency period had significantly higher Sobs value while the relaxation period had significantly higher Shannon index value compared with the fruiting and dormancy periods. In contrast, for both B. striata and B. ochracea, the dormancy and fruiting periods had higher Simpson's index values than did the other phenological stages, though the differences were not very significant ( Figure 2C). Good's coverage scores of all phenological stages for the two Bletilla species were very high, ranging from 99.95% to 99.99%, indicating that the sequencing depth was adequate to reliably describe the root endophytic fungi associated with Bletilla species at different phenological stages, and no significant differences were observed among the five phenological stages for each species (Table S1).

Temporal Variation in Diversity of Root Endophytic Fungi
Fungal community richness (Sobs) and alpha diversity (Shannon and Simpson's) index values were compared among five phenological stages for the two Bletilla species at the operational taxonomic unit (OTU) level ( Figure 2). For B. striata, the florescence period had significantly higher Sobs and Shannon index values compared with the fruiting and dormancy periods. For B. ochracea, the emergency period had significantly higher Sobs value while the relaxation period had significantly higher Shannon index value compared with the fruiting and dormancy periods. In contrast, for both B. striata and B. ochracea, the dormancy and fruiting periods had higher Simpson's index values than did the other phenological stages, though the differences were not very significant ( Figure  2C). Good's coverage scores of all phenological stages for the two Bletilla species were very high, ranging from 99.95% to 99.99%, indicating that the sequencing depth was adequate to reliably describe the root endophytic fungi associated with Bletilla species at different phenological stages, and no significant differences were observed among the five phenological stages for each species (Table S1).

Figure 2.
Alpha diversity estimates of mycorrhizal fungi associated with Bletilla striata and B. ochracea at different growth stages. Significant differences (p < 0.05) among different growth stages of each species for each index are indicated with lowercase letters.

Community Composition of Root Endophytic Fungi among Different Phenological Stages
There were 10 and 16 fungal OTUs shared among all five phenological stages in association with B. striata and B. ochracea, respectively ( Figure 3). The shared 10 OTUs for B. striata belong to unclassified_c_Agaricomycetes (42.44%), Exophiala (19.35%), un-classified_o_Sebacinales (13.05%), Rhizoctonia (8.55%), Neocosmospora (5.92%), Dactylonectria (3.00%), Cylindrocarpon (2.75%) and Aspergillus (1.97%). The shared 16 OTUs for B. ochracea belong to unclassified_c_Agaricomycetes (70.29%), Exophiala (11.66%), unclassi-fied_p_Basidiomycota (7.97%), Paraphoma (4.12%), Fusarium (3.21%), Neocosmospora (2.42%), and others (0.34%) ( Figure S1). Additionally, there were 106, 15  The predominant species in root endophytic fungal communities associated with the two Bletilla species were largely consistent among the five phenological stages at the phylum level. The fungal phyla with the highest relative abundances were Basidiomycota and Ascomycota. However, differences in relative abundances between the two species for a specific phenological stage or among different phenological stages for each Bletilla species were observed. Other minor phyla, such as Rozellomycota, Glomeromycota, Mortierellomycota and Chytridiomycota, were also found for B. striata, while only Rozellomycota was found for B. ochracea at some phenological stages (Figure 4). At the florescence stage, B. striata and B. ochracea fungal communities were dominated by Ascomycota, with percentages of 78.93% and 67.07%, respectively, while during the fruiting, dormancy and relaxation stages, B. striata and B. ochracea fungal communities were dominated by Basidiomycota, with percentages ranging from 63.94% to 86.71% and 76.53% to 97.39%, respectively. However, during emergence, associations with B. striata were dominated by Ascomycota (49.04%) and Basidiomycota (42.30%) while associations with B. ochracea were domianted by Basidiomycota (77.95%). The predominant species in root endophytic fungal communities associated with the two Bletilla species were largely consistent among the five phenological stages at the phylum level. The fungal phyla with the highest relative abundances were Basidiomycota and Ascomycota. However, differences in relative abundances between the two species for a specific phenological stage or among different phenological stages for each Bletilla species were observed. Other minor phyla, such as Rozellomycota, Glomeromycota, Mortierellomycota and Chytridiomycota, were also found for B. striata, while only Rozellomycota was found for B. ochracea at some phenological stages (Figure 4). At the florescence stage, B. striata and B. ochracea fungal communities were dominated by Ascomycota, with percentages of 78.93% and 67.07%, respectively, while during the fruiting, dormancy and relaxation stages, B. striata and B. ochracea fungal communities were dominated by Basidiomycota, with percentages ranging from 63.94% to 86.71% and 76.53% to 97.39%, respectively. However, during emergence, associations with B. striata were dominated by Ascomycota (49.04%) and Basidiomycota (42.30%) while associations with B. ochracea were domianted by Basidiomycota (77.95%). At the family level, the florescence stage for both Bletilla species had the most diverse associated fungal species, with B. striata dominated by Herpotrichiellaceae, Phaeosphaeriaceae. Diversisporales_fam_Incertae_sedis, Pleosporaceae, Nectriaceae and Glomerellaceae, and B. ochracea dominanted by Herpotrichiellaceae, Serendipitaceae, Nectriaceae, Phaeosphaeriaceae and unclassfied_o_Sebacinales. Meanwhile, unclassifed_c_Agaricomycetes, Herpotrichiellaceae, Nectriaceae and Phaeosphaeriaceae were present among all five phenological stages for both Bletilla species. During the whole physiological period, B. ochracea had higher unclassifed_c_Agaricomycetes abundance and lower Herpotrichiellaceae abundance compared with B. striata. The florescence stage had higher unclass-fied_o_Sebacinales abundance compared with other physiological stages for both Bletilla species (Figure 4). At the family level, the florescence stage for both Bletilla species had the most diverse associated fungal species, with B. striata dominated by Herpotrichiellaceae, Phaeosphaeriaceae. Diversisporales_fam_Incertae_sedis, Pleosporaceae, Nectriaceae and Glomerellaceae, and B. ochracea dominanted by Herpotrichiellaceae, Serendipitaceae, Nectriaceae, Phaeosphaeriaceae and unclassfied_o_Sebacinales. Meanwhile, unclassifed_c_Agaricomycetes, Herpotrichiellaceae, Nectriaceae and Phaeosphaeriaceae were present among all five phenological stages for both Bletilla species. During the whole physiological period, B. ochracea had higher unclas-sifed_c_Agaricomycetes abundance and lower Herpotrichiellaceae abundance compared with B. striata. The florescence stage had higher unclassfied_o_Sebacinales abundance compared with other physiological stages for both Bletilla species (Figure 4).
At the genus level, as shown by the heatmap in Figure 5, the samples were divided into two groups, and root endophytic fungi communities were divided into four groups. The fungal genera associated with the emergence and florescence stages of B. striata and the florescence stage of B. ochracea were clustered into one group (group 1), while the other samples were clustered into another group (group 2). Group 1 had higher abundances of Exophiala, Dactylonectria, unclassified_p_Rozellomycota and Fusarium than group 2, while group 2 had higher abundances of unclassified_c_Agaricomycetes than group 1. At the genus level, as shown by the heatmap in Figure 5, the samples were divided into two groups, and root endophytic fungi communities were divided into four groups. The fungal genera associated with the emergence and florescence stages of B. striata and the florescence stage of B. ochracea were clustered into one group (group 1), while the other samples were clustered into another group (group 2). Group 1 had higher abundances of Exophiala, Dactylonectria, unclassified_p_Rozellomycota and Fusarium than group 2, while group 2 had higher abundances of unclassified_c_Agaricomycetes than group 1.

Dynamics of Isotopic Abundances over the Bletilla Growth Season
We compared the δ 13 C and δ 15 N values among the two Bletilla species and corresponding autotrophic reference plants at different phenological stages ( Figure 6). The mean δ 13 C values varied significantly among the three groups across the four phenological stages. At the florescence and fruiting periods, the mean δ 13 C values of B. striata were highest, followed by B. ochracea and lastly the reference plants, with significant difference

Dynamics of Isotopic Abundances over the Bletilla Growth Season
We compared the δ 13 C and δ 15 N values among the two Bletilla species and corresponding autotrophic reference plants at different phenological stages ( Figure 6). The mean δ 13 C values varied significantly among the three groups across the four phenological stages.
At the florescence and fruiting periods, the mean δ 13 C values of B. striata were highest, followed by B. ochracea and lastly the reference plants, with significant difference among each of the three groups. At the emergence and relaxation periods, the mean δ 13 C values of B. ochracea and B. striata were significantly higher than that of reference plants, but the two species did not significantly differ (Table 1). For the mean values of δ 15 N, significant differences among the three groups only existed in the florescence and fruiting periods. At florescence, B. striata and B. ochracea had significantly higher δ 15 N values than that of reference plants, while at fruiting, the two Bletilla species had significantly lower mean δ 15 N values than that of reference plants (Table 1). s 2021, 10, x FOR PEER REVIEW 9 of 20 among each of the three groups. At the emergence and relaxation periods, the mean δ 13 C values of B. ochracea and B. striata were significantly higher than that of reference plants, but the two species did not significantly differ (Table 1). For the mean values of δ 15 N, significant differences among the three groups only existed in the florescence and fruiting periods. At florescence, B. striata and B. ochracea had significantly higher δ 15 N values than that of reference plants, while at fruiting, the two Bletilla species had significantly lower mean δ 15 N values than that of reference plants (Table 1).  The δ 13 C and δ 15 N values among different phenological stages were also compared for each plant group (Table 1). For the values of δ 13 C, significant differences among the four phenological stages only existed in B. ochracea and autotrophic reference plants. For B. ochracea, the emergence and relaxation periods had significantly higher mean δ 13 C values than that of fruiting period. For the values of δ 15 N, significant differences among the four phenological stages existed for all the three plant groups. The fruiting period had significantly lower δ 15 N values than other three periods for both Bletilla species.  The δ 13 C and δ 15 N values among different phenological stages were also compared for each plant group (Table 1). For the values of δ 13 C, significant differences among the four phenological stages only existed in B. ochracea and autotrophic reference plants. For B. ochracea, the emergence and relaxation periods had significantly higher mean δ 13 C values than that of fruiting period. For the values of δ 15 N, significant differences among the four phenological stages existed for all the three plant groups. The fruiting period had significantly lower δ 15 N values than other three periods for both Bletilla species.

Enrichment Factors and N Concentrations of Two Bletilla Species at Different Phenological Stages
Relative enrichment factors (ε) for 13 C and 15 N were calculated for the two Bletilla species at four phenological stages. For B. striata, the ε 13 C value at florescence was significantly higher than those at emergence, fruiting and relaxation, and there were no significant differences among the latter three. For B. ochracea, the situation is rather different, with the ε 13 C values during relaxation, florescence and emergence periods being significantly higher than that during fruiting (Figure 7). However, the ε 15 N values at the different phenological stages for the two Bletilla species had the same temporal trend, with the values at florescence significantly higher than those at emergence and relaxation, and the values at emergence and relaxation being significantly higher than that at fruiting (Figure 8).
The ε 13 C values confirmed a significant 13

Enrichment Factors and N Concentrations of Two Bletilla Species at Different Phenological Stages
Relative enrichment factors (ε) for 13 C and 15 N were calculated for the two Bletilla species at four phenological stages. For B. striata, the ε 13 C value at florescence was significantly higher than those at emergence, fruiting and relaxation, and there were no significant differences among the latter three. For B. ochracea, the situation is rather different, with the ε 13 C values during relaxation, florescence and emergence periods being significantly higher than that during fruiting (Figure 7). However, the ε 15 N values at the different phenological stages for the two Bletilla species had the same temporal trend, with the values at florescence significantly higher than those at emergence and relaxation, and the values at emergence and relaxation being significantly higher than that at fruiting ( Figure  8). The ε 13 C values confirmed a significant 13 C enrichment for the two Bletilla species at the four phenological stages. The ε 15 N values confirmed significant 15 N enrichment during florescence, with significant 15 N depletion during fruiting for the two Bletilla species.   We also compared the ε 13 C and ε 15 N differences of the two Bletilla species at each phenological stage. At florescence and fruiting, B. striata had higher ε 13 C values than B. ochracea, while B. striata had lower ε 13 C values than B. ochracea at the emergence and relaxation stages. Meanwhile, based on ε 15 N values, B. ochracea had higher enrichment during florescence and less depletion at fruiting, emergence and relaxation stages compared with B. striata, though the differences were rather small. Nitrogen concentrations were compared among different phenological stages for B. striata and B. ochracea. For the two species, the nitrogen concentrations at the emergence and relaxation periods were significantly higher than those during florescence, while the values during florescence were significantly higher than those during fruiting. Meanwhile, there were no significant differences between the values during the emergence and relaxation periods (Figure 9). We also compared the ε 13 C and ε 15 N differences of the two Bletilla species at each phenological stage. At florescence and fruiting, B. striata had higher ε 13 C values than B. ochracea, while B. striata had lower ε 13 C values than B. ochracea at the emergence and relaxation stages. Meanwhile, based on ε 15 N values, B. ochracea had higher enrichment during florescence and less depletion at fruiting, emergence and relaxation stages compared with B. striata, though the differences were rather small. Nitrogen concentrations were compared among different phenological stages for B. striata and B. ochracea. For the two species, the nitrogen concentrations at the emergence and relaxation periods were significantly higher than those during florescence, while the values during florescence were significantly higher than those during fruiting. Meanwhile, there were no significant differences between the values during the emergence and relaxation periods (Figure 9).

Temporal Variation in Root-Associated Fungal Diversity for Bletilla Species
In this study, significant temporal variation was observed in the diversity of root endophytic fungi associated with the two Bletilla species. Compared with the fruiting and dormancy periods, the florescence period of B. striata and the relaxation period of B. ochracea had significantly higher Shannon index values. This finding may be associated with dynamic changes in plant physiological process and interactions between root endophytic fungi and orchids because metabolic rate of plant and nutritional requirement vary among different physiological stages. Additionally, pelotons formed by orchid endophytic fungi are short-lived structures that undergo rapid turnover within orchid cells, allowing new colonization events [37]. Thus, the diversity of symbiotic fungi can also change quite rapidly [25]. Accordingly, temporal variation in fungal diversity across seasons has been found in many cases [38,39]. For example, Koide et al. [38] found variation in fungal diversity of ectomycorrhizae roots, and Jumpponen [39] found temporal changes in arbuscular mycorrhizal roots. In our study, Bletilla plants become dormant with extremely low metabolic rate during dormancy period, which may lead to the lowest root endophytic fungal diversity in this period. The fruiting period had lower root endophytic fungal diversity, which may be due to the weaker physiological activity, higher temperature or lower soil water content during summer. The senescence of plant accompanying weaker physiological activity during fruiting periods, which do not need too much nutrition supply, may reduce the dependence of plants on fungi [40]. Meanwhile, high temperature and drought would reduce microbial population size and thus restrain community diversity and complexity [41]. Tedersoo et al. [42] found that although host plant families had strong effects on the phylogenetic community composition of fungi, temperature and precipitation mostly affected ectomycorrhizal fungal richness. Han et al.

Temporal Variation in Root-Associated Fungal Diversity for Bletilla Species
In this study, significant temporal variation was observed in the diversity of root endophytic fungi associated with the two Bletilla species. Compared with the fruiting and dormancy periods, the florescence period of B. striata and the relaxation period of B. ochracea had significantly higher Shannon index values. This finding may be associated with dynamic changes in plant physiological process and interactions between root endophytic fungi and orchids because metabolic rate of plant and nutritional requirement vary among different physiological stages. Additionally, pelotons formed by orchid endophytic fungi are short-lived structures that undergo rapid turnover within orchid cells, allowing new colonization events [37]. Thus, the diversity of symbiotic fungi can also change quite rapidly [25]. Accordingly, temporal variation in fungal diversity across seasons has been found in many cases [38,39]. For example, Koide et al. [38] found variation in fungal diversity of ectomycorrhizae roots, and Jumpponen [39] found temporal changes in arbuscular mycorrhizal roots. In our study, Bletilla plants become dormant with extremely low metabolic rate during dormancy period, which may lead to the lowest root endophytic fungal diversity in this period. The fruiting period had lower root endophytic fungal diversity, which may be due to the weaker physiological activity, higher temperature or lower soil water content during summer. The senescence of plant accompanying weaker physiological activity during fruiting periods, which do not need too much nutrition supply, may reduce the dependence of plants on fungi [40]. Meanwhile, high temperature and drought would reduce microbial population size and thus restrain community diversity and complexity [41]. Tedersoo et al. [42] found that although host plant families had strong effects on the phylogenetic community composition of fungi, temperature and precipitation mostly affected ectomycorrhizal fungal richness. Han et al. [43] found that Paphiopedilum spicerianum was associated with obviously more mycorrhizal fungi OTUs with higher Shannon values during wet seasons (274 and 1.88, respectively) relative to dry seasons (143 and 1.54, respectively).
Significant temporal variation in root-associated fungal endophyte community structure across seasons was also found in this study. During florescence, B. striata and B. ochracea were dominated by Ascomycota, while during fruiting, dormancy and relaxation Basidiomycota were dominant. However, during emergence, B. striata fungal associations were dominated by Ascomycota, and B. ochracea fungal associations were dominated by Basidiomycota. Meanwhile, the abundance of dominant species at various phenological stages varied among Bletilla species. The reason root endophytic fungi community composition and its abundance varied among phenological stages requires further analysis. There are three possible reasons. First, some root endophytic fungi require metabolites produced by orchid plants at some particular period for its growth [19]. Second, mycorrhizal tissue structure varies among physiological periods, underlying different mycorrhizal fungi infection rates and fungi types. Third, the ecological factors extrinsic to host plant roots, such as precipitation, light irradiation, soil temperature and humidity, can vary among physiological periods, greatly influencing the root endophytic community composition. The reasons for different responses of fungi community composition and abundances to phenological changes by two species are difficult to identify. One most likely explanation may be genetic differences in the plant species. Tedersoo et al. [42] found that host plant family had the strongest effect on the phylogenetic community composition of fungi.
At the family level, the florescence stage for both Bletilla species had the highest fungal diversity, likely owing to orchid plants requiring considerable nutrients to bloom, exceeding the nutrients supplied by photosynthesis; accordingly, they may need root endophytic fungi to supply nutrients to enable blooming [44]. Data of ε 13 C and ε 15 N in this study substantiate this finding. Bellino et al. [45] found that phototrophic and mycotrophic nutrition alternate change with the seasons and nutritional requirements of plants. Another possible explanation is that the outer velamen tissue of orchids is destroyed to some extent during the fast-growing stages through friction with soil, which results in massive fungal invasion during later physiological stages (e.g., florescence). Thus, root cells are heavily colonized and have high activity of invading hyphae, which leads to the highest fungal infection rate occurring at the florescence stage. Rasmussen et al. [46] found that mycorrhizal infection reached its maximum intensity 2-6 months after mycotrophic root development.
In this study, the composition and abundance of total species, dominant species and endemic species varied between the two Bletilla species at each phenological stage and also varied among phenological stages for both species. This may reflect distinct mycorrhizal preferences for specific host species, or alternatively, this pattern could result from hosts undergoing natural selection for different fungal endophytes enabling better adaption to their environments [47,48].

Isotopic Abundance Trends over the Growth Season
Our results demonstrate that variation in 13 C and 15 N enrichment can be highly dynamic throughout the whole growth season for Bletilla spp. For B. striata, the florescence period had significantly higher ε 13 C values, while the ε 13 C value during fruiting in B. ochracea was significantly lower than that during the other three periods. For the two Bletilla species, the ε 15 N values during florescence were significantly higher than those during emergence and relaxation periods, and the values at the emergence and relaxation periods were significantly higher than that of the fruiting period. This indicates that Bletilla plants substantially change their isotopic content across phenological stages, which is inconsistent with the results of Ercole et al. [25], who reported only slight changes in 13 C and 15 N natural abundances in the evergreen species Anacamptis morio throughout all its growth stages.
In this study, the two Bletilla species were relatively enriched in 13 C across all phenological stages (except dormancy), indicating that they obtained carbon from their root endophytic fungi throughout the whole year. Because fungal tissues have higher 13 C stable isotope abundance compared with neighbouring autotrophic plants [26,27], according to the isotopic differences between plant-and fungus-derived C, we can trace the source of carbon by 13 C enrichment across different phenological stages. Additionally, B. striata had higher ε 13 C values at florescence and fruiting and lower ε 13 C values at emergence and relaxation compared with B. ochracea, which indicates that the carbon obtained from root endophytic fungi varied among Bletilla species and phenological stages. This variation in 13 C enrichment may be owing to variation in climate, genetic effects, soil nutrients and/or mycorrhizal associations [17,49,50]. However, in this study, soil total carbon, nitrogen, phosphorus content and pH did not differ substantially among B. striata and B. ochracea sites (Table S2). The observed interspecific variation in 13 C enrichment at same time might owing to genetic variation within Bletilla spp. Meanwhile, the variation in 13 C enrichment among phenological periods might be owing to variations in mycorrhizal tissue structure and ecological factors such as soil temperature, soil humidity and light radiation, which greatly influenced the ease with which fungi infect roots as well as root endophytic fungi types, thus affecting root endophytic fungal community composition. Additionally, 13 C enrichment was positively correlated with the abundances of Ascomycota at the phylum level across all growth stages. Selosse et al. [51] found that Ascomycota species were often found in orchid roots and could form typical orchid mycorrhizae.
Obvious enrichment of 15 N during the florescence period and significantly strong 15 N depletion at the fruiting period were found for the two Bletilla species, indicating that Bletilla species obtained N from their mycorrhizal fungi during florescence and that a significant transfer of nitrogen from orchids to their mycorrhizal associates occurred during the fruiting period. Researchers found that a transfer of nutrients from the orchids to their mycorrhizal associates was consistent with such 13 C and 15 N depletions [52,53]. Meanwhile, compared with B. striata, B. ochracea had relatively higher enrichment during florescence and lower depletion of 15 N at other phenological stages, indicating that B. ochracea obtained more nitrogen from their associated mycorrhizal fungi during the florescence period and transferred less N to mycorrhizal fungi during other phenological stages compared with B. striate.
For both B. striata and B. ochracea, leaf total nitrogen concentrations were significantly lower during the florescence and fruiting periods compared with the emergence and relaxation periods, which is most probably explained by nitrogen investments in flower and seed tissues. Liebel et al. [17] reported that compared with non-flowering/non-fruiting individuals, flowering/fruiting Goodyera repens individuals had lower leaf total nitrogen and chlorophyll concentrations, which is most probably explained by the plants using leaf nitrogen to form flowers and seeds. Andersson [54] showed that floral investment by Nigella sativa could cause reduced allocation to other plant functions. Based on the obvious enrichment of 15 N during florescence and significantly strong 15 N depletion during fruiting, we can infer that Bletilla species might compensate for low leaf total nitrogen concentrations by using 15 N enriched fungal sources during flowering but that there was no such compensation during fruiting. One explanation is a relative paucity of root endophytic fungal species and lower activity owing to drought stress during the fruiting period. Another likely reason is that photosynthesis can mostly support fruiting costs and thus maintain the same seed production by increasing leaf, stem and fruit photosynthesis without increases in carbon fluxes from fungus to orchid [16,17].

Plant Species
Bletilla Rchb. f. is a genus of terrestrial orchids distributed across northern Burma, China and Japan, and it belongs to the tribe Epidendreae in the subfamily Orchidoideae within Orchidaceae [55]. B. striata and B. ochracea are two relatively widespread Bletilla species in China that are widely used for garden landscaping as well as traditional Chinese medicine [56]. Mature Bletilla plants enter dormancy in the winter and begin to bolt in the early spring under natural conditions. Adult plants form mycorrhizal associations, mainly with Ascomycota and Basidiomycota and are characterized by the phenological stages described in Table 2.

Study Site
Samples were collected from the Bletilla germplasm resource nursery located at Shanghai Chenshan Botanical Garden (31 • 04 N and 121 • 11 E). The area has a subtropical monsoon climate with a cumulative mean annual precipitation of 1213 mm and a mean annual temperature of 15.6 • C. The soil matrix (within the top 0-5 cm) was composed of mountain clay, river sand and vermiculite, with a ratio of 1.5:1:1 and a pH of 6.5-7.5. The Bletilla plants examined had been cultivated for three years and were already fully established at the site.

Sampling
Sampling was conducted from May 2018 to April 2019 in the germplasm resource nursery of Shanghai Chenshan Botanical garden. Root samples were collected at five time points, while leaf samples were collected at four time points (outside of the dormancy period) spanning the different phenological stages (Table 2). Three 1 × 1 m plots were randomly established in the B. striata and B. ochracea sites. In each plot, two roots per plant from five orchid plants were collected. Collected roots from each plot were pooled and kept cold during transport to the laboratory for further analysis. For each species, we sampled fresh top leaves from several Bletilla individuals at four time points (five individuals at florescence and fruiting periods and three individuals at emergence and relaxation periods). Meanwhile, leaves of three to five autotrophic reference plants under the same microclimate were also sampled at each phenological stage. The autotrophic reference species were chosen based on the criteria described by Gebauer and Meyer [23]. For detailed information on reference species, please see Table 1. In total, 30 root samples and 32 leaf samples from two Bletilla species and 59 leaf samples from autotrophic reference species were collected in this study.

Identification of Root Endophytic Fungi
Roots were rinsed with tap water, sonicated to remove any adhering soil and dirt and sterilized as follows: Roots were rinsed with sterile water for 30 s and then 70% ethyl alcohol for 2 min, soaked in 2.5% sodium hypochlorite for 5 min, transferred to 70% ethyl alcohol for 30 s and finally washed with sterile water three times. Roots were then cut into small pieces with sterile scissors, and ten to twelve sections per sample were selected for genomic DNA extraction and purification using the FastDNA Spin Kit for Soil (MP Biomedicals, Irvine, CA, USA) according to the manufacturer's protocol. The nuclear ribosomal internal transcribed spacer (ITS) region was amplified with the fungal-specific primers ITS1F and ITS4 [17]. The PCR amplification was performed as follows: initial denaturation at 95°C for 3 min, followed by 35 cycles of denaturing at 95°C for 30 s, annealing at 55°C for 30 s and extension at 72°Cfor 45 s, single extension at 72°C for 10 min, and end at 10°C. The PCR reactions were performed in triplicate 20 µL mixture containing 2 µL of 10 × Pyrobest Buffer, 2 µL of 2.5 mM deoxyribonucleotide triphosphates (dNTPs), 0.8 µL of each primer (5 µM), 0.2 µL of Pyrobest DNA Polymerase (TaKaRa), 10 ng of template DNA, and finally ddH 2 O up to 20 µL. The PCR product was extracted from 2% agarose gel and purified using the AxyPrep DNA Gel Extraction Kit (Axygen Biosciences, Union City, CA, USA) according to manufacturer's instructions and quantified using Quantus™ Fluorometer (Promega, Madison, WI, USA).
All positive PCR products were purified with the AxyPrep DNA Gel Extraction Kit (Axygen Biosciences, Union City, CA, USA) and sequenced bidirectionally using an Illumina MiSeq PE300 platform (Illumina, San Diego, CA, USA) according to the standard protocols from Majorbio Bio-Pharm Technology Co. Ltd. (Shanghai, China). The generated paired-end reads were merged once, but because the reads were longer than 300 bp, the paired-end reads could not be merged without overlap. Thus, we used single-end long reads for further analysis. The raw reads were deposited into the NCBI Sequence Read Archive (SRA) database (Accession Number: SRP225764). Raw fastq files were demultiplexed and quality-filtered with Trimmomatic according to the following criteria: 300 bp reads that were truncated at any site receiving an average quality score of <20 over a 50 bp sliding window, and the truncated reads shorter than 50 bp were discarded; reads containing ambiguous characters were also discarded. Operational taxonomic units (OTUs) were clustered with a 97% similarity cut-off using UPARSE (version 7.1, http://drive5.com/uparse/), and chimeric sequences were removed using UCHIME. The taxonomy of each ITS rRNA gene sequence was analysed with the RDP Classifier algorithm (http://rdp.cme.msu.edu/) against the Unite 8.0 ITS rRNA database using a confidence threshold of 70%. Fungal OTUs were classified into different putative trophic strategies following the classification of the FunGuild v1.0 ( http://www.stbates.org/guilds/app.php).

Analysis of Stable Isotope Abundance and N Concentration
Leaf samples were washed with deionized water, oven-dried at 105 • C, ground into a fine power and stored in a desiccator fitted with silica gel until subsequent analysis. Relative C and N isotope abundances and N content were measured using an elemental analyser (vario PYRO cube; Elementar Analysensysteme GmbH, Langenselbold, Germany) coupled with a continuous flow isotope ratio mass spectrometer (IsoPrime100, Elementar UK Ltd., Stockport, UK), as described by Bidartondo et al. [57]. Relative isotope abundances are denoted as δ values, which were calculated according to the equation: where R sa and R st are the ratios of heavy isotopes to light isotopes in the samples and the respective standards. Standard gases were calibrated with USGS40 and USGS41a for carbon and nitrogen isotopes, provided by the United States Geological Survey (USGS). Reproducibility and accuracy of the isotope abundance measurements were routinely controlled by measures of laboratory standard acetanilide. The calculation of N concentrations in the samples followed the protocol by Gebauer and Schulze [58]. For relative C and N isotope natural abundance analyses, acetanilide was routinely analysed with variable sample weights once every 12 samples. Enrichment factors (ε s ) for all samples were calculated according to the equation: where δ S is the relative isotope abundance of a Bletilla sample, and δ ref is the mean isotope abundance of all autotrophic reference plants.

Statistical Analyses
Isotopic data analyses were performed using SPSS 16.0 for Windows (SPSS Inc., Chicago, IL, USA). Before analysis, all variables were checked for normality by Shapiro-Wilks test and for homogeneity of variance by Levene's test. Two-way ANOVA with Tukey HSD post hoc comparisons or Tamhane's T2 test were used when data were normally distributed. We checked the differences in mean δ 13 C and δ 15 N values among the autotrophic reference plants, B. striata and B. ochracea as well as differences among different physiological stages for each species. Additionally, values of ε 13 C, ε 15 N and N concentration among the different physiological stages were also compared for each Bletilla species. Significant differences in three alpha diversity indexes (Sobs, Shannon and Simpson) and coverage among different physiological stages were also compared in this study. Significance was defined at the 95% confidence level throughout this article. Additionally, microbial species composition analysis was performed using the free online Majorbio I-Sanger Cloud Platform (www.i-sanger.com).

Conclusions
Identifying the trophic relationship of endangered Bletilla with its root endophytic partners is critical for understanding plant growth and ultimately restoring wild populations. We investigated temporal variation in root endophytic fungal diversity, as well as the natural abundance of carbon and nitrogen stable isotopes in two Bletilla species (B. striata and B. ochracea) at different phenological stages. Our results suggest that variations in 13 C and 15 N natural abundance and root endophytic fungi for Bletilla spp. can be highly dynamic across all phenological stages. Both Bletilla species obtained carbon from their mycorrhizal fungi during the whole year. B. striata obtained more carbon during the florescence and fruiting periods but obtained less carbon during the emergence and relaxation periods compared with B. ochracea. For both Bletilla species, the florescence stage had relatively more 13 C and 15 N enrichment compared with the other phenological stages. Due to the two-way selection of plant species and root endophytic fungi, the community structure and abundance of total species, dominant species and endemic species varied between the two Bletilla species and among phenological stages. Additionally, the varying tendency of root endophytic fungal diversity across the whole growth stage was consistent with 13 C enrichments, which was also accord with the nutritional requirement of plant. Our results have important implications for the current understanding of fungus-host relationships and also provide practical information for Bletilla conservation efforts.