In Situ Seedling Baiting to Isolate Plant Growth-Promoting Fungi From Dendrobium Ocinale, An Over-Collected Medicinal Orchid In China

Background: Orchids are highly dependent on mycorrhizae for survival and growth. Traditionally, to obtain orchid mycorrhizal fungi (OMFs) for plant growth, fungi have been isolated from the roots of mature wild plants; however, the species of OMF may change as the plants undergo different developmental stages. In this study, we developed the idea of using in situ seedling baiting to capture seedling growth-promoting fungi from Dendrobium ocinale, an overcollected medicinal orchid in China. Results: In vitro-produced seedlings of D. ocinale were transplanted into their original habitats, and newly established roots of well-growing seedlings were sampled for fungal isolation. Roots were sampled at 6 different times over one year, and ve Tulasnella species and one Fusarium species were obtained and identied morphologically and molecularly. The ability to promote seedling growth was tested in three Tulasnella species TPYD-1, TPYD-2 and TPYD-3, with high isolation frequencies by inoculation onto in vitro-produced D. ocinale seedlings. Although the three fungi were closely related species and clustered together in the phylogenetic tree, they showed different eciencies in promoting D. ocinale seedling growth. Tulasnella TPYD-2 showed a strong ability to promote seedling growth and could be selected for use in restoration plantings of D. ocinale. Conclusions: Our results suggest that using an in situ seedling baiting technique could be an ecient way to obtain seedling growth-promoting fungi, and this approach could have broad applications in orchid mycorrhiza studies and orchid conservation. and fresh and dry weights among the fungal inoculation treatments and control at 120, 150 and 180 d were recorded. Comparisons were made with one-way ANOVA and the least signicant difference (LSD) method when the data showed a normal distribution or the generalized linear model (GLM) when the data were not normally distributed. All statistical analyses were performed using SPSS software (version 25.0).


Background
The genus Dendrobium Swartz is one of the largest genera in Orchidaceae, comprising approximately 1,100 species distributed in the tropical and subtropical areas of Asia and Australia, and approximately 78 species, including 14 endemic species, are native to China [1]. Of these, more than 40 Dendrobium species have a very long history of use in traditional Chinese medicine (TCM) and are known as Shi-Hu in Chinese [2]. Consequently, most of these medicinal Dendrobium species have been exploited to the point of local extirpation, especially in some "hotspot" areas [3,4].
Increasing supplies is considered an important method to reduce harvesting pressures of overexploited wild species [5]. For medicinal Dendrobium species, the Shi-Hu industry has developed rapidly in southern China since the 1990s as a result of massive commercial cultivation; however, due to doubts about the quality and e cacy of the products obtained with this cultivation method, the lack of product standards and the low quantities of products, the utilization of these products in TCM is limited [6]. Moreover, such massive commercial cultivation has not alleviated pressure on wild populations, according to recent surveys on the orchid trade in China [7]. This re ects a consumer preference for wild-collected materials and the fact that wild collection maintains the livelihoods of local suppliers who do not have the capital to invest in rearing facilities. Preference for wild-collected medicines by Chinese consumers re ects the belief that these materials are more effective and contaminant-free than cultivated materials [8]. To balance the conservation of threatened Dendrobium species and the continuing demand for wild-grown materials, a restoration-friendly cultivation model for medicinal orchids, especially Dendrobium species, has been proposed, in which orchids are planted in natural settings [3,8]; this method is considered promising for linking the commercial TCM industry with biodiversity conservation initiatives in China [6].
In the context of restoration-friendly cultivation, future scienti c research on Dendrobium species, including mycorrhizal technology and microbial fertilizers, is urgently needed [6]. In our previous studies focused on the conservation of over-collected Dendrobium species, we obtained germination-enhancing fungi for different Dendrobium species through an in situ/ex situ seed baiting technique [9][10][11][12], and the fungi have also been successfully used in restoration plantings of medicinal Dendrobium species [13]. However, the dynamics of fungal associations at the seedling stage remain largely unknown. In practice, we found that after seedling establishment, seedling growth varied greatly even in a small-scale environment [13], indicating that new fungal partners are established with seedlings and play an important role in plant growth. Orchids have complex symbioses with fungi throughout their lifespan, and the species of orchid mycorrhizal fungi (OMFs) may change in different plant developmental stages [14,15]. Under natural conditions, the broadening and/or changing of a mycorrhizal association may enable orchids to adapt to the varied physiological changes during seedling development, including the switch to partial or full autotrophy, an increase in transpiration, or environmental uctuations [16].
Motivated by the success of in situ/ex situ seed baiting, in this study, we developed the idea of using in situ seedling baiting to obtain seedling growth-promoting fungi, in which in vitro-produced seedlings of D. o cinale were transplanted into their original habitats, and newly established roots of well-growing plants were sampled for fungal isolation. The aim of this study was to understand fungal diversity at the seedling stage and obtain OMFs for seedling growth to support restoration plantings of D. o cinale. Here, we present our results, addressing three principal questions: (1) Does this method work to obtain OMFs associated with seedlings of D. o cinale? (2) What are the effects of the dominant OMFs on seedling growth in D. o cinale? (3) Have the OMFs obtained in the current study also been obtained in other studies in which OMFs were isolated from roots of mature wild plants, and do these fungi have positive effects on seedling growth in D. o cinale?

Fungal isolation and identi cation
In total, 64 puri ed fungal strains were successfully obtained from 600 root fragments of D. o cinale that were sampled at 6 different times. All fungal strains were morphologically and molecularly identi ed as six fungal species, including ve Tulasnella species and one Fusarium species (Additional le 1; Table 1). Among the 5 Tulasnella species, TPYD-1, TPYD-2, and TPYD-3 were isolated 15, 12 and 7 times and were present in 6, 5 and 3 samples, respectively ( Fig. 1D-F), while TPYD-4 and TPYD-5 were isolated three and one times, respectively, and were only present in one sample each. Fusarium oxysporum TPYD-6 was isolated from all 6 samples with a high isolation frequency of 26 out of 64 times (Additional le 1). Phylogenetic analysis A total of 28 fungal species/strains from 11 studies were reported to have positive effects on seedling growth in D. o cinale (Additional le 2). Among them, 12 fungi were isolated from the protocorms or roots of D. o cinale, 10 fungi were isolated from the roots of other Dendrobium species, and 6 fungi were obtained from the roots of non-Dendrobium orchid species. Interestingly, 17 fungi were OMFs, and the other 11 fungi were non-OMFs (Additional le 2). The ITS-rDNA sequences of 21 out of 28 fungi were downloaded from the GenBank database and used to generate the phylogenetic tree with the fungal sequences of ve OMFs obtained in the current study. In the phylogenetic tree, the fungi TPYD-1, TPYD-2 and TPYD-3 were clustered together, while TPYD-4 and TPYD-5 were scattered with other species of Tulasnella and Epulorhiza (Fig. 2).
Testing the capacity of fungi to support seedling growth In the incubation experiments, the three fungi TPYD-1, TPYD-2 and TPYD-3 with high isolation frequencies were examined in terms of their capacity to support seedling growth in D. o cinale. At 30 and 60 d after incubation, no pelotons were observed in the roots under all treatments (Fig. 1G), and only scarce pelotons could be found in the roots under the TPYD-1 and TPYD-2 treatments at 90 d. All three fungi clearly started to colonize the roots from 120 d after incubation, and the numbers of pelotons in the roots under the three fungal treatments increased greatly at 150 and 180 d (Fig. 1H).

days after incubation
All three fungi showed positive effects on seedling growth ( Fig. 3A-E). The lengths of the longest roots in all three fungal treatment groups were signi cantly longer than that in the control group (F = 4.970, P = 0.003) with the highest growth rate in the TPYD-2 treatment (R = 29.78%; Fig. 3A). TPYD-1 and TPYD-2 signi cantly increased the biomass of seedlings as the fresh and dry weights in the two treatments were signi cantly higher than those in the control treatment ( 180 days after incubation: Clearly, TPYD-2 promoted seedling growth more effectively than TPYD-1 and TPYD-3 ( Fig. 1I; Fig. 3K & O). Except for the longest leaf length, the longest root length, plant height, fresh weight and dry weight were signi cantly longer or higher in the TPYD-2 treatment than in the TPYD-1 and TPYD-3 treatments and the control treatment (all P < 0.001). For the longest leaf length, there were no signi cant differences between the TPYD-2 and TPYD-3 treatments (P = 0.204), but the leaves were signi cantly longer than those in the TPYD-1 and control groups (all P < 0.05; Fig. 3L). The biomasses of seedlings were also signi cantly higher under the TPYD-3 treatment than under the control (fresh weight: P < 0.001; dry weight: P = 0.018; Fig. 3N & O).

Discussion
Orchid mycorrhizae are symbiotic associations between orchid plant roots and fungi, in which fungal hyphae grow within living plant cells and form intracellular pelotons [17,18]. Orchid mycorrhizal fungi (OMFs) are considered to belong to the so-called rhizoctonia aggregate, a polyphyletic group of fungi belonging to Tulasnellaceae, Ceratobasidiaceae, and Serendipitaceae [19]. Orchids are highly dependent on mycorrhizae for survival and growth: OMFs not only supply carbohydrates to facilitate the growth of nonphotosynthetic protocorms during the seed germination stage but also provide nutrients and growth factors to plants, conferring plant metal tolerance and inhibiting the development of pathogens [17,[20][21][22][23].
Mycorrhizal technologies and microbial fertilizers could simplify orchid seedling production and promote plant growth and have been considered important aids to the sustainable development of the Dendrobium industry in China [6,24,25]. As the most popular medicinal plant species in the genus Dendrobium, D. o cinale has received much research attention regarding its mycorrhizal symbionts [25]. Traditionally, to obtain the optimal source of fungal mycobionts for symbiotic germination or plant growth, fungi have been isolated and screened from the roots of mature wild plants [26,27]. Among the 28 fungal species reported in 11 D. o cinale studies, only ve fungal species in one study were obtained from protocorms via ex situ seed baiting [28], while the other 23 fungi in the remaining 10 studies were isolated from the roots of wild mature plants, including D. o cinale and other orchid species (Additional le 2). The 28 fungal species/strains, originally obtained from different orchid species and belonging to a wide range of taxonomic groups (17 OMFs and 11 non-OMFs), were found to have positive effects on the growth of D. o cinale seedlings (Additional le 2). Seemingly, a randomly obtained fungus could promote seedling growth in D. o cinale; however, all of the above fungi were tested in vitro, and there have been no reports on the successful application of these fungi in practice so far. Theoretically, a high diversity of fungi may associate with the roots of an adult orchid plant, an orchid will utilize different OMFs at different life history stages, and the mycorrhizal symbionts may change in different developmental stages [14,15,29,30]. It is unclear whether the fungi involved in the seedling stage remain until the plants reach adulthood in D. o cinale.
The in situ/ex situ seed baiting technique has been suggested as an effective and easy way to obtain seed germination-enhancing fungi [31,32]. Using this method, we successfully obtained e cient germination-enhancing fungi for different Dendrobium species [9][10][11][12], as well as other terrestrial orchids [33,34]. This led to the development of the idea of using seedling-trap experiments to capture seedling growth-promoting fungi in the current study. After in vitro-produced seedlings of D. o cinale were transplanted for more than one year in their original habitats, newly established roots of well-growing plants were sampled six times across one year in different seasons (Additional le 1). In total, ve OMFs and one non-OMFs were obtained and identi ed (Table 1). Tulasnella species TPYD-1 and TPYD-2 were the dominant OMFs with a higher frequency of isolation than other OMFs, and TPYD-1 was present in all six samples, while TPYD-2 was present in ve samples (Additional le 1). The three fungi TPYD-1, TPYD-2 and TPYD-3 with high isolation frequencies were closely related species, as they were clustered together in the phylogenetic tree (Fig. 2). Fusarium oxysporum TPYD-6 was also dominant, with a high isolation frequency and presence among the samples (Additional le 1).
In this study, to test and screen the fungi that could be used in real practice for growth of D. o cinale seedlings, in vitro-produced seedlings were moved to mixed cultivation substrates in open environments, in which seedlings normally undergo an acclimatization stage. This could be a possible reason why the three fungi TPYD-1, TPYD-2 and TPYD-3 started to colonize seedling roots after a long period of 120 d. However, the corresponding increases in pelotons in the roots and in the ve measured indices in all three fungal treatments at 150 and 180 d suggested that all three fungi could promote seedling growth in D. o cinale but showed different e ciencies (Fig. 3). At 180 d after incubation, the longest root length, plant height, fresh weight and dry weight of seedlings in the TPYD-2 treatment were signi cantly longer or higher than those in the other two fungal treatments and the control treatment (Fig. 3). TPYD-2 could be selected as an ideal OMF for seedling growth in the restoration-friendly cultivation of D. o cinale.
Tulasnellaceae is regarded as one of the main fungal families of orchid mycorrhizae, and many of these species are found to support seed germination and seedling growth in Dendrobium species [9,10,19,24].
The other two Tulasnella species, TPYD-1 and TPYD-3, were closely related to TPYD-2 but showed less pronounced effects on seedling growth in D. o cinale. Seedlings of D. o cinale might need a longer time to acclimatize to fungal infection [35], or the two fungi might have other ecological functions [21,22]. In the current study, we focused on the effects of fungi obtained from seedling baiting on seedling growth, but it would also be worthwhile to explore whether these fungi could effectively promote seed germination in D. o cinale. In another study, we conducted comparisons of symbiotic germination of seeds inoculated with the TPYD-2 obtained in this study and six other fungal strains isolated from protocorms of D. o cinale via in situ seed baiting. At 90 d after incubation, the percentage of seedlings in the LQ treatment was 70.09 ± 3.2%, while no protocorms or seedlings were found in the TPYD-2 treatment (Wang et al. unpublished data). The results also suggested that D. o cinale could associate with different fungi in different life stages.
In addition to OMFs, other root-associated nonmycorrhizal endophytes have also been recorded and identi ed from a wide range of orchid species [36,37]. In this study, Fusarium oxysporum TPYD-6 was isolated from all 6 samples with high isolation rates. Fusarium species have been reported to be associated with different orchid species [33,38]. Interestingly, although Fusarium species have been reported as pathogens in many orchid species [39], especially F. oxysporum, which was reported to cause wilt disease in D. o cinale [40], other studies have suggested that Fusarium species can stimulate seed germination in the terrestrial orchid Cypripedium reginae [41] and enhance resistance to pathogens and promote plant growth in Dendrobium species [42,43]. In a recent study, F. oxysporum KB-3, obtained from the roots of Bletilla striata, was considered an OMF because it could promote seed germination of B. striata, establish colonization and produce coiled hyphal structures within the cortical cells in the roots of B. striata and Dendrobium candidum [44]. As noted in many studies, the border between endophytic and mycorrhizal fungi could be di cult to de ne, and some fungus-plant interactions can easily shift from mutualism to parasitism depending on the plant's physiology and environmental conditions [45,46,47].
For the current study, it is worth exploring the effect of F. oxysporum TPYD-6 alone as well as the possible synergistic effect of TPYD-6 with other OMFs on the growth of D. o cinale seedlings.

Conclusions
In this study, in situ seedling baiting was used to capture orchid seedling growth-promoting fungi for the rst time. Five Tulasnella species and Fusarium oxysporum TPYD-6 were successfully isolated from seedlings of D. o cinale that were produced in vitro and transplanted for more than one year into their original habitats. Tulasnella sp. TPYD-2 showed a great ability to promote plant growth in D. o cinale and could be selected for practical use in restoration plantings of D. o cinale. For any given orchid species, it is still uncertain whether the fungi involved in the seedling stage remain until plant adulthood.
Our results suggest that using an in situ/ex situ seedling baiting technique could be an e cient approach to obtain plant growth-promoting fungi. This method might be more reliable for overcollected orchids for which mature plants are hard to nd in the wild but their original habitats are known. A similar idea has been used in studies on the interactions between orchids and fungi [48] and suggests that this approach could have broad applications in orchid mycorrhiza studies and orchid conservation.

Study species and study site
Dendrobium o cinale Kimura & Migo (synonym of D. catenatum and D. candidum) is a lithophytic orchid widely distributed in subtropical areas, especially karst regions at altitudes from 500 to 1600 m in China [1]. D. o cinale is one of the most popular TCM herbs and is known as Tie-Pi-Shi-Hu in Chinese. The wild populations of D. o cinale are very small and sparsely distributed due to severe overharvesting over the past 30 years [3], and this species is listed as critically endangered on the IUCN Red List [49]. The study site was located at a limestone mountain near Lengdong village (24°54′13″N; 104°58′43″E; alt. 1100 m; Fig. 1A) in Xingyi, Guizhou Province, China. This region is a typical karst landform, and more than 20 species of Dendrobium including D. o cinale have been reported in the area but are seriously threatened due to habitat loss and overcollection [50]. According to local residents, wild plants of D. o cinale could be found in the mountains of the study site several years ago.
In situ seedling baiting and sampling An in situ seedling baiting experiment was conducted using in vitro-produced seedlings of D. o cinale, in which mature seeds were obtained via hand-outcross pollination of owers between different individuals of D. o cinale and asymbiotically geminated on MS medium. The seedlings were removed from the asks, and the roots were cleaned and then transplanted to the cracks and crevices of calcareous rocks at the study site during May and June at the start of the rainy season in 2016 (Fig. 1B). After more than one year of natural growth, root samples were taken six times at the study site during August 2017 and May 2018 (Additional le 1). For each sampling time, 10 clusters of seedlings containing 3-5 individuals with robust growth and newly established roots were randomly selected. Ten roots from each cluster were randomly sampled. Collected root samples were placed in an icebox and transferred to the laboratory for fungal isolation on the same day.

Fungal isolation and identi cation
The sampled roots were washed with running water, surface sterilized with 75% ethanol for 2 min and 2% sodium hypochlorite for 3 min, and then washed three times with distilled water. Mycorrhizal fungi were isolated from roots of D. o cinale following the method described by Bayman et al. [51]. The super cial uncolonized root tissue was removed using a thin sterile blade, and root fragments containing intracellular pelotons were obtained by scrapping the roots using needles and forceps. The root fragments were scraped as small as possible with an anatomical needle under a 10 * 20 magni cation microscope (XS-A, Shanghai Pudan Co., Ltd.), transferred to PDA medium (200 g/L potato, 20 g/L dextrose and 20 g/L agar) with 0.05 g/L penicillin and 0.05 g/L streptomycin and incubated at 25 ± 2 °C. The root fragments were monitored daily during incubation on PDA. When hyphal growth from the root fragments exceeded 0.5 cm in length, tips of the hyphae were cut and transferred to new PDA medium for puri cation. After repeating this puri cation step 4-5 times, puri ed strains were obtained.
All obtained fungal strains were morphologically identi ed and then grouped based on culture characteristics such as colony morphology, hyphal characteristics, hyphal colour, and the colours associated with secondary metabolites [52]. For each group, a strain obtained from each sampling time was randomly selected for molecular identi cation. The rDNA region containing two ITS regions and the 5.8S gene was ampli ed using the ITS1 and ITS4 primers [53], and PCR was performed following Selosse et al. [54]. All ITS-rDNA sequences obtained were compared with those deposited in the GenBank database (National Center for Biotechnology Information, NCBI 2012) using the Basic Local Alignment Search Tool (BLAST), which allows the identi cation of isolates at the genus or species level when the ITS sequence similarity exceeds 95% or 97%, respectively [55]. Later, one sequence per species was deposited in NCBI GenBank under the accession numbers listed in Table 1.

Phylogenetic analysis
A phylogenetic tree was generated to visualize the phylogenetic positions of the OMFs obtained in the current study among other known fungal species associated with Dendrobium o cinale. By searching scienti c databases (e.g., Web of Science; Google Scholar; Weipu Chinese journals database) with the key words "Dendrobium o cinale", "Dendrobium catenatum", "Dendrobium candidum" and "mycorrhizal fungi", we collected all the available information on fungi that are considered to have positive effects on D. o cinale seedling growth, including original resources, host orchid species and their effects on seedling growth in D. o cinale (Additional le 2). Then, the available ITS-rDNA sequences of all related representative fungi were downloaded from the GenBank database (https://www.ncbi.nlm.nih.gov/) and used to generate a phylogenetic tree with the fungal sequences obtained in the current study. The ITS sequence of Armillaria sinapina As93 (FJ495039) was set as an outgroup based on previous studies [24]. Alignment of nucleotide sequences was performed by Clustal X version 1.81 [56] followed by manual adjustments using BioEdit [57] (Hall 1999). Maximum likelihood (ML) phylogenetic tree searches and ML bootstrapping were conducted using the web server RAxML-HPC2 on TG ver. 7.2.8 [58,59] with 1000 rapid bootstrap analyses followed by a search for the best-scoring tree in a single run [58]. Testing the capacity of fungi to support seedling growth The three OMFs with the highest isolation rates were selected, and their capacity to support seedling growth in D. o cinale was tested. In vitro-produced seedlings of D. o cinale were transplanted from asks to circular plastic pots (height * diameter = 8 cm * 10 cm) after the medium on roots was washed away (Fig. 1C). For each pot, 5 seedlings were planted together in cultivation substrate with the same proportion of volcanic stone, peat and bark fragments. Once seedling transfer was completed, each pot was inoculated with one cubic centimetre of the three fungal inoculants, which were placed in the centres of the pots, to establish three fungal incubation treatments, and a sterile control without fungal inoculation was used as a control treatment (CK). Each treatment was replicated in 24 pots, and to avoid possible fungal infection among treatments, all pots from a given treatment were placed in separate germination chambers at 25 ± 2 °C with a 12/12-h light/dark cycle.
To determine if and when mycorrhizal symbiosis actually became established, three roots from each treatment were randomly selected and examined for the presence of pelotons at 30, 60, 90, 120, 150 and 180 d after incubation. Pelotons were cleared using 10% KOH solution, washed with 1% HCl solution, stained with 0.05% (w/v) trypan blue in lactic acid glycerol solution overnight [11], and then destained in acetic glycerol solution before observation under a microscope (DM2000, Leica Microsystems GmbH, Wetzlar, Germany).

Data collection and statistical analysis
At the same time as seedling transplanting, 40 seedlings were randomly selected to measure the lengths of the longest root and leaf, plant height, and fresh and dry weights. Later, 40 seedlings from each treatment were randomly selected, and the same measurements were made at 120, 150 and 180 d after incubation. Moreover, to assess the effects of different fungi on seedling growth, the growth parameters

Availability of data and materials
All data analyzed during this study are included in this published article and its supplementary information les. Additional raw data from sequencing and quality control datasets are available from the corresponding author upon reasonable request.
Ethics approval and consent to participate Not applicable.

Consent for publication
Not applicable.