Biochemical and transcriptomic analyses of the symbiotic interaction between Cremastra appendiculata and the mycorrhizal fungus Coprinellus disseminatus

Background Cremastra appendiculata is a rare terrestrial orchid with a high market value as an ornamental and medicinal plant. However, the species depends entirely on fungi for seed germination under natural conditions. In a previous study, we have successfully isolated and identified the mycorrhizal fungus Coprinellus disseminatus which was able to induce the germination of C. appendiculata seeds. We then speculated that C. disseminatus may do so by breaking the testa imposed dormancy of the seeds. In this study, biochemical and transcriptomic analyses were used to characterize the germination of C. appendiculata seeds, collected at different stages of germination, as affected by C. disseminatus. Results The lignocellulose in the seeds coat of C. appendiculata was degraded by the mycorrhizal fungus resulting in facilitated absorption of water. The rate of decline in lignin content was 67 and 73% at 6 and 12 days after sowing, respectively. The water content increased from 13 to 90% during symbiosis. A total of 15,382 genes showing significantly different levels of expression (log2 FPKM≥2.0, Qvalue≤0.05) were successfully identified among all libraries, where the highest number of DEGs was shared between 6 days versus 0 day after symbiotic germination. Gene annotation results suggested that 15 key genes related water-status, such as DHN gene family and Xero 1 were down-regulated. The genes zeaxanthin epoxidase ZEP, 9-cis-epoxycarotenoid dioxygenase NCED3 and β-carotene hydroxylase involved in the biosynthesis of abscisic acid (ABA) were significantly down-regulated in 6 days as compared to 0 day after symbiotic germination. Conclusions This work demonstrates that mycorrhizal fungus C. disseminatus can stimulate C. appendiculata seeds germination through a mechanism of breaking the testa imposed dormancy and inducing water absorption of the embryo. Supplementary Information The online version contains supplementary material available at 10.1186/s12870-021-03388-6.

coats of the orchids arise mainly from the outer and inner integuments, and contain lignocellulose, suberin, polyphenols and cutin [4,20,34]. Lignin is a component of plant cell wall that contains hemicellulose, cellulose and lignin. Lignin binds to cellulose and hemicellulose forming a hydrophobic barrier to water permeation [27,32]. This barrier enhances the survival of orchid seeds in harsh conditions and enhances wind-aided seed dispersal [4,12]. However, the lignified seed coat may interfere with water and nutrient uptake, resulting in 'physically imposed dormancy' [4,5,12,31,34,44]. Thus, effectively breaking the testa-imposed dormancy is a critical step in seed germination which a prerequisite for species conservation and large-scale production of orchids.
Orchid mycorrhizal fungi have been reported to be capable of degrading lignin by producing lignin-modifying enzymes and energetically supporting the establishment of orchid seedlings [34,36,40]. However, some studies have reported that an efficient enzymatic machinery of mycorrhizal fungi toward lignin was not sufficient for effective break down of lignocellulose in the seed coat, and that some fungi can produce appressorium-like structures or utilize conidia to soften the cell walls [11]. Moreover, the mycorrhizal fungi produce cellulases and pectinases, which contribute to effective penetration of the fungal hyphae into plant tissues [35]. White-rot fungi are recognized as one of the most active microorganisms in lignin degradation. The saprobic fungus Coprinellus disseminatus has been documented to be able to degrade lignocellulose [2,39].
The lignified seed coat, which is thought to be one of the factors in seed dormancy, is formed during the final stages of seed development. Concomitantly, the highly hydrophilic dehydrin (DHN) proteins accumulate in the embryo and serve as protectants against desiccation [7,10,16,28]. Dehydrins have multifaceted roles in the protection of plant cells during stress conditions, including drought which triggers ROS accumulation and negatively affects the overall plant performance [37,48]. Phytohormones strongly regulate seed germination and stress responses. Among which ABA is a primary stress hormone that induces seed dormancy, and regulates stomatal closure to restrict water loss by transpiration [13,24,49]. The expression of DHN proteins has also been reported to be induced by both ABA-dependent or ABAindependent mechanisms under drought stress conditions [37].
Cremastra appendiculata, is a rare terrestrial orchid that often grows in the understory of humid and highly shaded forests [45]. It has a high medicinal value because its tubers are effective for cancer treatment, angiogenesis inhibition, heat-clearing, detoxification and detumescence [22,23,38,41]. Due to its commercial and medicinal market value, C. appendiculata plant material often illegally collected from the wild populations. In a previous study, we reported that seeds of C. appendiculata with intense phenolic compounds in the coats showed enhanced germination and protocorm development in the presence of the mycorrhizal fungus C. disseminatus DJF-10 [14]. However, the mechanisms involved in the symbiotic germination remained unknown.
In this study, seeds at 0 (fungus unpenetrated), 6 (fungus penetrating), 12 (fungus penetrated into embryo and formed pelotons) and 25 days (seeds developed into protocorms) after symbiotic germination were used as test materials to investigate the influence of C. disseminatus on the lignocellulose of the seed. We studied the dynamic changes in the lignocellulose content of C. appendiculata seeds during symbiotic germination with C. disseminatus and then, analyzed the lignocellulose degradation by a combination of methods, including Fourier transform infrared spectroscopy (FTIR), pyrolysis-gas chromatography mass spectrometry (Py-GC/MS) and gas chromatography mass spectrometry (GC-MS). Finally, we analyzed transcriptome of C. appendiculata during symbiotic germination, which indicated that C. disseminatus possibly degrades lignin in the seed coat, with concomitant increase in water uptake.

Experimental materials and symbiotic germination
The mature seeds of C. appendiculata and the mycorrhizal fungus C. disseminatus DJF-10 were obtained and stored in the Institute of Plant Physiology and Molecular Biology of Guizhou University, Guiyang city, Guizhou Province, China. The seeds were sterilized with 75% ethanol (30 s) and then with sodium hypochlorite (1% effective chloride, 3 min). The seeds were germinated in Petri dishes containing oatmeal agar medium (OMA: 4 g L − 1 rolled oats, 8 g L − 1 agar, at natural pH). Then a 5 mmplug of actively growing C. disseminatus mycelium, precultured on PDA medium, was placed in each OMA Petri dish. The plates were incubated at 22 ± 2 °C in full darkness.

Phenotypes of C. appendiculata seeds during symbiotic germination with C. disseminatus
The symbiotically germinated seeds were samples at four stages according to a previous histological study [14]. First, at symbiotic sowing (0 day) (before fungus invasion) where the embryo was surrounded with lignified seed coat. Second, at 6 days (fungus penetrating) where the mycelium penetrated from the suspensor. Third, at 12 days (fungus colonization) where some well-developed pelotons formed and the swollen embryo broke through the testa. Fourth at 25 days (pelotons degradation) where more than 70% of the seeds germinated and formed protocorms with protomeristem at the basal cells were degraded. Therefore, samples were collected at 0 ( Fig. 1 0 day), 6 ( Fig. 1 6 days), 12 ( Fig. 1 12 days) and 25 days ( Fig. 1 25 days) after symbiotic sowing (CA, SY1, SY2, and SY3, respectively) and stored at − 80 °C for transcriptome research and other subsequent analyses.
Total lignocellulose content and water uptake in C. appendiculata germinating seeds

Lignocellulose content
The dried samples were milled using a pestle and stored in sealed plastic bags. The lignocellulose contents were examined using lignin content assay kit (Solarbio, China). The hemicellulose and cellulose contents were measured using respective assay kits (Sangon Biotech, China) according to manufacturer's instructions.

Water uptake by seeds at different stages of germination
The symbiotic seeds at different stages of germination were dried at 40 °C until constant dry weight was reached (48-72 h). The water content was expressed as a percentage of the seed fresh weight for four to five biological replicates.

FTIR and Py-GC/MS analysis of lignocellulose degradation
The FTIR was used to follow the degradation of lignocellulose with a Vertex70 (Bruker, Germany) FTIR spectrometer. Before the FTIR, the dried samples were embedded in KBr pellets at room temperature. FTIR spectrophotometer were recorded in a range of 4000 cm − 1 to 400 cm − 1 . The spectra were recorded in an absorption mode of 16 scans per sample with a resolution of 2 cm − 1 [12,46].
The chemical composition of lignocellulose from symbiotic seeds was analyzed by Py-GC/MS according the method of Chen et al. [8] with some modifications. Helium was used as the carrier gas at a constant rate of 1 mL min − 1 . The pyrolysis was carried out at 550 °C, the chromatograph was programmed from 50 to 250 °C at a rate of 15 °C min − 1 and the final temperature was maintained for 10 min. The chemical compounds were identified by comparing their mass spectra with those at the National Institute of Standards and Technology (NIST) library.

GC-MS analysis of low-molecular-weight degradation products of lignocellulose during symbiotic germination
To evaluate the lignocellulose degradation efficiency of the mycorrhizal fungus C. disseminatus, products from lignocellulose of C. appendiculata seeds were analyzed by GC-MS, since this system is appropriate for analysis of the low-molecular-weight compounds. The compounds Fig. 1 Phenotypes of the seeds of C. appendiculata symbiotically germinated with C. disseminatus DJF-10 at different stage after sowing. Zero day, six days, twelve days, twenty-five days indicated symbiotic seeds at 0 day, 6 days, 12 days and 25 days after sowing, respectively were harvested by crushing the symbiotic medium with a rod and then stirring on shaker at 120 rpm at 10 °C for 4 h [2]. After filtering the contents through four layers of cheese cloth, the samples centrifuged at 10,000 rpm for 10 min at 4 °C. The supernatant was harvested and stored at − 20 °C. The pre-treatment of the supernatant was performed according to Zang et al. [46] with some modifications. The supernatants were acidified with concentrated HCl to pH 2-3. Then, the samples were extracted three times with ethyl acetate and the organic layers were combined. The combined fractions were dewatered over anhydrous Na 2 SO 4 passed through a filter paper and evaporated down to 2 mL at 40 °C. Consequently, 100 μL of pyridine and 400 μL of N, O-bis (trimethylsilyl) trifluoroacetamide (BSTFA) were added to 500 μL samples of the extracted products. The samples were silylated with heating at 60 °C for 30 min and filtered through 0.45 μm membranes. Aliquots of 1 μL of the silylated samples were injected into the GS-MS system. Three replicates were used for each sample.
The GC-MS analysis achieved using a 5975C-7890A system (Agilent, USA) with a DB-5 capillary column (30 m × 0.25 mm, 0.25 μm film thickness). In addition, helium was used as the carrier gas at a constant flow rate of 1.0 mL min − 1 . The column temperature program was set to 60 °C (2 min) and 60-260 °C (10 °C min − 1 , hold time of 10 min). The injection and detection temperatures were maintained at 200 and 250 °C, respectively. Electron ionization (EI) mass spectra were acquired in full-scan mode from 50 to 55 m/z. The compounds were identified by comparing their mass spectral data to those in NIST library.

Preparation of RNA for transcriptome analysis
The four samples of symbiotic seed germination were collected for transcriptome analysis. Additionally, samples of the pure cultured C. disseminatus on OMA medium were collected for RNA de novo assembly. Total RNA was extracted from 100 mg using plant RNeasy Kit (Omega Bio-Tek, USA) according to the manufacture's protocol. The extracted RNA was treated with RNase-free DNase I to remove genomic DNA. The quality and quantity of the purified RNA were tested by using NanoDrop One (Nanodrop Technologies Inc., CA, USA), Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, U.S.A.) and 1% agarose gel electrophoresis. Library construction and sequencing were performed using Illumina HiSeq4000 platform (Illumina Inc., CA, USA) at Beijing Genomics Institute (Shenzhen, China).

De novo assembly of RNA-seq data and functional annotation of unigenes
The total clean reads were obtained by removing reads containing adapters, ambiguous reads (N > 5%), and low-quality reads from the raw reads. De novo assembly of the trimmed reads derived from each sample was conducted using Trinity V2.0.6 with default settings. The complete series of fungal and plant transcriptomic data were deposited in the National Center for Biotechnology information (NCBI) database (accession: No. PRJNA762946).
Unfortunately, reference genomes of C. appendiculata and C. disseminatus have not been published yet. We were unable to remove the intracellular hyphae from the symbiotic protocorms. Therefore, the fungus C. disseminatus de novo group was assembled individually using Trinity at default parameters. The coding sequences of symbiotic protocorms were first mapped to the fungal de novo assembly nonredundant protein (NR) database. Then, all the mapped C. disseminatus reads were derived from the symbiotic group [15,47]. The remaining unpaired reads were classified as the de novo reference assembly of C. appendiculata and were prepared to detect the expression levels. The paired-end reads were mapped to each plant and fungal reference genome assembly using Bowtie2 V2.2.5 [19]. The transcript abundance was estimated using RSEM V1.2.28 [21]. The gene expression levels were estimated by applying the fragments per kilobase per million mapped reads (FPKM). Functional annotation of the unigenes was subjected to NR, Nt, KOG, GO, KEGG, Swissprot and Pfam public databases.

Expression analysis of DEGs
The differentially expressed genes (DEGs) in library size were identified with DESeq2 and moderated estimation of the fold change [1,25] in the expression level with a fold change ≥2.0 and Qvalue ≤0.05. The GO and KEGG enrichment were performed using the obtained DEGs.

Statistical analysis
All experiments were repeated at least three times and the results were presented as means ± SD. The statistical analysis was performed using one-way repeated measures ANOVA, and significant differences were tested by Duncan's test at p < 0.05.

Lignocellulose degradation and water uptake of C. appendiculata seeds during symbiotic germination
To investigate the degradation of lignocellulose by C. disseminatus, the loss of lignocellulose dry weight and water absorption during symbiosis were measured. After 6 days of symbiotic germination of C. appendiculata with C. disseminatus, the lignin content significantly dropped from 119.48 mg g − 1 (CA) to 38.55 mg g − 1 (SY1) and the lignin degradation rate in the seeds was 73.15% at 12 days (SY2) (Fig. 2). Subsequently, with cell differentiation and protocorm development, lignin accumulation increased to 68.84 mg g − 1 at SY3, but remain much lower than that of CA. The hemicellulose content was also decreased from 55.04 mg g − 1 in CA to 38.65 mg g − 1 in SY1. The hemicellulose degradation rate was 29.78% in SY1, and significantly increased to 61.61 mg g − 1 in SY2. With protocorm formation, the hemicellulose content increased to 126.89 mg g − 1 in SY3. The cellulose content did not change significantly among different stages.
Under symbiosis, the water content increased sharply significantly from CA to SY1 (13.94 and 91.66%, respectively) and then unchanged at SY2 and SY3 (92.54 and 91.58%, respectively) (Fig. 2). These results revealed that the lignified seed coat was broken by C. disseminatus, resulting in enhanced water permeation to the embryo. The results also demonstrate that the seed coat limitation is one of the important components of seed dormancy.

Identification of lignocellulose in symbiotic seeds
To further validate the lignocellulose degradation in C. appendiculata seeds by C. disseminatus, the chemical compositions of different samples was characterized by FTIR and Py-GC/MS. The samples showed remarkably different spectra (Fig. 3). Bands of C-lignin in CA were evident at 1650 (strong) and 782 cm − 1 . Bands of G/S lignin existed at valleys of 1592, 1515 (sharp) and 871 cm − 1 (sharp). Under symbiosis with C. disseminatus, significant reduction in the lignin signals were recorded at 1650, 1592 and 1515 cm − 1 in SY1, SY2 and SY3, respectively as compared to CA. The xylan band of 2923 cm − 1 was significantly decreased at SY1, SY2 and SY3. The absorption band at 1621 cm − 1 (arrow 1) may be associated with carbonyl and acetyl groups in xylan degradation intermediates. The band, which was associated with cellulose, disappeared at 1000 cm − 1 (arrow 2) during symbiosis. In addition, strong and sharp signals were recorded at 1766 and 1749 cm − 1 , though it was not clear whether or not they were associated with lignocellulose.
Pyrolysis of the seed symbiotic samples released various families of compounds ( Table 1). The fraction of phenolic compounds decreased significantly, while the concentrations furanics and other compounds varied in different samples. These results suggested that the lignocellulose fraction of the seeds was degraded by the symbiotic fungus. In particular, lignin degradation was more significant as demonstrated. Other non-phenolic compounds such as acetic acid and ketone groups resulted from lignin and from the efficient release of phenolic monomers.

GC/MS analysis of low-molecular-weight products of degradation of lignocellulose of C. appendiculata seeds symbiotically germinated with C. disseminatus
To identify the products of lignocellulose biodegradation, GC-MS was used to detect the low-molecularweight compounds in OMA symbiotic medium collected at different stages during the symbiotic germination of C. appendiculata seeds with C. disseminatus (Table 2; Fig. S1). The degradation products are listed in Table 2. More than 10 degradation products were detected. However, the CA group contained only dibutyl phthalate, suggesting that the compound might have originated from the OMA medium. It is well documented that benzene compounds result from degradation of lignin. Five relative compounds (benzeneethanamine, benzene, benzeneacetic acid, benzamide and benzenedicarboxylic acid) were identified in the OMA symbiotic medium of SY1 and SY2. This suggested that the lignin component of the seed coat was degraded, in agreement with FTIR and Py-GC/MS data. Furfural, acetic acid and butanoic acid are degradation products of xylan. Acetic acid was detected in SY1, SY2 and SY3, whereas butanoic acid was detected only in SY2.
Furfural was not detected at any stage. Some other acids and eaters resulting from lignocellulose degradation were detected in SY1 or SY2. These results confirmed that the lignocellulose fraction of the seed coat was degraded by C. disseminatus during symbiotic germination.

RNAseq analysis, de novo assembly, and functional annotation
Analysis of RNA-seq of C. appendiculata at four developmental stages of symbiotic germination was performed   22:15 to characterize the transcriptome changes during seed germination. A total of 41.06 Gb clean data were generated from each library after filtering out the low-quality data. Among raw reads in all samples, the Q30 values ranged from 91.47 to 92.28% indicating high-quality reads appropriate for further analysis (Table S1). Since no reference genome was available for C. appendiculata, all 359,510,000 reads were do novo-assembled into 48,750 CDS with an N50 length of 1398 bp and 97,800 unigenes with an average length of 1120 bp (Table S2). Samples of different biological replicates were clustered separately based on their distinct developmental stages.
For annotation, 97,800 unigenes were subjected to BLASTX search against the sequences in NR, Nt, Pfam, Swissprot, GO, KOG, and KEGG databases ( Fig. 4A; Table S3). As a result, a total of 65,911 unigenes (67.39% of all unigenes) had at least one putative function from one of these databases. For NR annotation, a total of 61,803 (63.19%) unigenes were annotated. As shown in Fig. S2, 36,107 unigenes were assigned with a best score to Dendrobium catenatum (58.42%) and 14,862 ones were assigned to Phalaenopsis equestris (24.05%). For GO annotation, 45,555 unigenes were annotated into three GO categories, including cellular component (CC), molecular function (MF), and biological process (BP) (Fig. S3). The cellular component class contained cellular anatomical entity and intracellular and protein-containing complex. The most abundant molecular functions were binding and catalytic activity. Among the biological processes, cellular and metabolic processes were more abundant. Other entries such as biological regulation, localization and response to stimulus were relatively high.
Unigenes among different groups were chiefly classified into carbohydrate transport and metabolism, function unknown, general function prediction only, posttranslational modification, protein turnover, chaperones, signal transduction mechanisms and transcription (Fig. 4B).

Functional annotation of differentially expressed genes Differential expression analysis of gene expression at different stages of symbiotic germination
To identify the genes involved in symbiosis, we compared, as pairs, the transcriptomes of the four developmental periods of the symbiotic seeds. The pair SY1-CA, represented the symbiotic seeds at 6 days (fungal invasion) compared to 0 day (fungus-free seeds), while SY2-SY1 was the symbiotic material at 12 days (fungal colonization) compared to the stage of fungal invasion at day 6. The pair SY3-SY2 presented the transition to fungal degradation and protocorms with meristem from the fungal colonization at 12 days. A total of 15,382 genes showed significantly different expression levels (log2 FPKM≥2.0, Qvalue≤0.05) among all libraries (Fig. 4C). Among these differentially expressed genes (DEG), 5050 genes were up-regulated and 4818 ones were down-regulated in SY1-CA (Fig. 4D). In SY2-SY1, 5285 genes were up-regulated and 1893 ones were down-regulated in SY2-SY1. In SY3-SY2, 449 genes were up-regulated and 669 genes were down-regulated ( Fig. 4D; Table S4). We found that the highest number of DEGs was shared between the early mycelium penetration stages (SY1-CA), followed by the early germination stage (SY2-SY1), and then the protocorm stage (SY3-SY2) (Fig. 4D).

GO enrichment analysis
To characterize the DEGs, we performed a GO enrichment analysis using Rage tool. Among the GO terms of DEGs, there were 51, 100 and 153 terms significantly overrepresented in SY1-CA, SY2-SY1 and SY3-SY2, respectively (Table S5). In total, 25 GO terms were over-represented in the three comparisons (Table 3). In particular, these terms included some significantly overrepresented GO terms related to oxidation reactions such as heme binding, monooxygenase activity, dioxygenase activity, peroxidase activity and iron ion binding. In addition, the terms of response to water, metal iron binding and endo-1,4-β-xylanase activity were annotated in SY1 versus CA. The terms of ATP binding, channel activity and cell wall macromolecule catabolic process were identified in SY2-SY1. However, xyloglucan and carbohydrate metabolic process were also found.

KEGG enrichment analysis
To investigate the differences in metabolic processes at different stages of symbiotic germination in more detail, a KEGG pathway enrichment analysis with the DEGs was conducted, based on the Qvalue≤0.05 (Table S6, S7). At SY1-CA, the categories flavonoid biosynthesis, plant hormone signal transduction, fatty acid degradation and phenylpropanoid biosynthesis were significantly enriched. The pathways of phenylpropanoid and flavonoid biosynthesis were also dramatically enriched in SY2-SY1 and SY3-SY2. Pathways for fatty acid degradation, carbon fixation in photosynthetic organisms, glycosaminoglycan degradation and carbon metabolism were enriched in SY2-SY1. Pathways of protein processing in endoplasmic reticulum, glycosaminoglycan degradation, plant signal transduction and other glycan degradation were significantly enriched between SY3-SY2. Taken together, these findings suggested that the fungus induced the expression of plant hormone signal transduction pathways which may break the seed dormancy. Moreover, it induced the expression of fatty acid degradation, carbon metabolism and endoplasmic reticulum processing, and hence improves the efficiency of utilization of stored nutrients.

Expression patterns of symbiosis-related genes at each stage
The lignified seed coat forms a hydrophobic barrier to water permeation during germination. The lignin content of the seeds was sharply reduced after symbiosis with the fungus C. disseminatus. Once the hydrophobic barrier was removed, the embryo was able to absorb water efficiently. In this study, we paid much attention to terms related to water absorption. Fifteen genes (including 9 CaDHN1 genes, 4 CaXero1 genes, 1 hypothetical protein, and 1 CaDHN4-like) related to the term response to water were down-regulated ( Fig. 5; suggested that these genes regulate the response to fungal stimulation during seed germination of C. appendiculata after symbiosis with the mycorrhizal fungus. Indeed, plant hormones play a vital role in the process of symbiotic germination. The current study revealed that a large number of DEGs was annotated Table 3 Significantly overrepresented gene ontology terms during the symbiotic germination of C. appendiculata with C. disseminatus DJF-10 Round (filled) circles represent the GO terms were overrepresented during mycorrhizal fungus Tulasnella sp. symbioses with Bletilla striata [29]  to signal transduction mechanisms. Carotenoid biosynthesis provides the precursors for the synthesis of ABA. In our study, zeaxanthin epoxidase ZEP (CL5276.Contig3_All), 9-cis-epoxycarotenoid dioxygenase NCED3 (CL6018.Contig1_All) and β-carotene hydroxylase (Unigene9557_All and Unigene9558_All) were significantly down-regulated in SY1-CA ( Fig. 5; Table S8). Furthermore, most of the abscisic acid receptor pyrabactin resistance 1-like (PYLs) were upregulated in SY1-CA. Seven genes of protein phosphatase 2C (PP2C), six of SNF1-related protein kinase 2 (SnRK2) and four basic region-leucine zipper (bZIP) transcription factors were down-regulated in this group ( Fig. 5; Table S8).

Discussion
The orchid C. appendiculata is one of the most difficult-to-propagate species, a condition which severely restricts its large-scale production. The factors that inhibit C. appendiculata seed germination are still largely unknown. During symbiosis, the lignin content abruptly decreased resulting in a significant increase in water uptake of the seeds compared to the control. These results revealed a negative correlation between lignin and germination, suggesting a possible mechanism whereby the testa-imposed dormancy was broken [34,42]. The mycorrhizal fungus C. disseminatus can degrade lignocellulose by releasing various enzymes such as xylanase, laccase and cellulase [2,39]. In the present study, we tested laccase, xylanase and cellulase in the symbiotic OMA medium during symbiosis. The results showed that these enzymes were abundant in SY1 and SY2 but decreased at SY3 (Fig. S4) consistently with the physiological data which indicated that lignocellulose was largely degraded at SY1 and SY2. However, the chemical basis of the mechanism involved has not been clearly explained. Moreover, a layer of the seed coat was strained greenish blue with TBO, indicating the presence of polyphenols and lignin in the seed coat of mature seeds, a barrier that disappeared during symbiotic germination [14]. Moreover, the enzymes of cellulases may play an important role in the penetration of fungal hyphae into orchid seeds [35], suggesting that the mycorrhizal fungus C. disseminatus essentially scarifies and degrades the lignified seed coat by producing hydrolytic enzymes, resulting in facilitating water uptake and stimulating germination. Some microorganisms grow on lignocellulose as carbon source and degrade lignin to access cellulose as a preferred carbon source [11,27]. It is noteworthy that C. disseminatus is a common species that exists on stumps, buried wood and logs with high wood-decaying capacity [26,30,39]. Hynson et al. [17] reported that more than 80% of the C. appendiculata carbon was obtained from fungal partner, which is more than that in many other orchids. In the present study, the reducing sugar concentration was highest at SY1 in the OMA medium, and decreased thereafter (Fig. S5). The mycelium formed pelotons in the embryo that began to degrade after about 2 weeks. Combing these data with these of enzyme activity, lignocellulose analysis and germination experiment, it seems that C. appendiculata was associated with the wood-decaying Coprinellus members [14,45] that obtains nutrients via saprotrophic fungi for seed germination and protocorm development during symbiosis [36,43]. Fungal invasion appears to trigger metabolic changes in the host plant (especially the nutrient accumulation) during the establishment of symbiosis between C. appendiculata and C. disseminatus, a process that requires further research.
Four key stages were selected for transcriptome analysis based on the dynamics of fungus penetration into the embryo and key anatomical features in the embryo development. Unfortunately, it was not possible to remove the intracellular fungal hyphae from the symbiotic samples. This meant that the symbiotic samples contained transcripts of C. appendiculata and the fungus C. disseminatus. Therefore, RNA from the symbiotic fungus library was used for establishing a C. disseminatus reference transcriptome and all the mapped reads were driven from the symbiotic samples. This method could be effectively applied to assign genes identified in the transcriptome of plant-fungi associations [47]. The orchid C. appendiculata has been reported to form mycorrhizal associations with saprotrophic fungi [14,43,45], which have the potential to elicit strong defense responses in plants. However, our transcriptome data provided no evidence of strong defense activation at the GO terms in C. appendiculata symbiotic samples. This result is similar to those of other symbiotic transcriptome studies on orchids, such as B. strilata [29], Serapias vomeracea [33] and to arbuscular mycorrhizal AM symbiosis. These results suggested that plants recruit a signaling pathway to exploit saprotrophic fungi without inducing stress response [29].
The pathway of plant hormone signal transduction was highly enriched in SY1-CA during C. appendiculata symbiotic germination. The data showed that ABA content significantly dropped from 1.38 mg g − 1 (CA) to 0.08 mg g − 1 (SY2) with more than 15-fold reduction, and remained at low level thereafter. During this stage, the genes for key enzymes of ABA synthesis were downexpression. Meanwhile, the genes of DHN family were significantly enriched in SY1-CA, suggesting high expression in CA that effectively alleviate the negative effects of drought stress [37]. During symbiosis, the expression of DHN family was down-regulated. Therefore, we suggest that the low permeability of the seed coat and the high level of ABA in the embryo could act in parallel to hinder the germination of the mature seeds of C. appendiculata.

Conclusion
In the present study, we investigated the biochemical and transcriptomic changes of C. appendiuclata seeds and provided an outline of the important components of the response to symbioses. In summary, the presence of lignified seed coats on the seeds of C. appendiculta suggests testa-imposed dormancy that restricts seed germination. As the lignified seed coat was broken down by mycorrhizal fungus C. disseminatus, the water absorption by the embryo was improved. The maximal reduction in lignin content was 67.34% at the first week after sowing. This result was confirmed by FTIR and Py-GC/MS of lignin degradation. Comparative transcriptome profiling revealed the expression patterns of water uptake and ABA biosynthesis candidate genes, which demonstrated that the testa-imposed dormancy was broken. During symbiosis, all genes related to water uptake and ABA biosynthesis were down-regulated. This study extends our understanding of the barriers to seeds germination in C. appendiculata. Future work should focus on the molecular mechanistic processes in C. appendiculata seed germination as affected by C. disseminatus symbiosis.