JA signal-mediated immunity of Dendrobium catenatum to necrotrophic Southern Blight pathogen

Dendrobium catenatum belongs to the Orchidaceae, and is a precious Chinese herbal medicine. In the past 20 years, D. catenatum industry has developed from an endangered medicinal plant to multi-billion dollar grade industry. The necrotrophic pathogen Sclerotium delphinii has a devastating effection on over 500 plant species, especially resulting in widespread infection and severe yield loss in the process of large-scale cultivation of D. catenatum. It has been widely reported that Jasmonate (JA) is involved in plant immunity to pathogens, but the mechanisms of JA-induced plant resistance to S. delphinii are unclear. In the present study, the role of JA in enhancing D. catenatum resistance to S. delphinii was investigated. We identified 2 COI1, 13 JAZ, and 12 MYC proteins in D. catenatum genome. Subsequently, systematic analyses containing phylogenetic relationship, gene structure, protein domain, and motif architecture of core JA pathway proteins were conducted in D. catenatum and the newly characterized homologs from its closely related orchid species Phalaenopsis equestris and Apostasia shenzhenica, along with the well-investigated homologs from Arabidopsis thaliana and Oryza sativa. Public RNA-seq data were investigated to analyze the expression patterns of D. catenatum core JA pathway genes in various tissues and organs. Transcriptome analysis of MeJA and S. delphinii treatment showed exogenous MeJA changed most of the expression of the above genes, and several key members, including DcJAZ1/2/5 and DcMYC2b, are involved in enhancing defense ability to S. delphinii in D. catenatum. The findings indicate exogenous MeJA treatment affects the expression level of DcJAZ1/2/5 and DcMYC2b, thereby enhancing D. catenatum resistance to S. delphinii. This research would be helpful for future functional identification of core JA pathway genes involved in breeding for disease resistance in D. catenatum.

in the process of large-scale cultivation of D. catenatum, Southern Blight disease caused by Sclerotium delphinii leads to heavy yield losses by infecting its leaves, stems, and roots [5].
Southern Blight is a broad-spectrum disease caused by the necrotrophic pathogen S. delphinii and S. rolfsii [5,6]. Athelia rolfsii is the sexual state of the fungus, we rarely see it, while its asexual state S. delphinii and S. rolfsii are more common [6]. Compared with S. rolfsii, S. delphinii is more pathogenic [5]. In warm and high humidity environments, S. delphinii infects the host and secretes many cell wall-degrading enzymes (CWDEs) to degrade the host cell wall [7], and then hyphae penetrate the plant tissue causing local necrosis, with subsequent formation of abundance of white mycelium and rapeseed-like sclerotia [5]. Southern Blight is considered one of the most destructive diseases worldwide for its vast host range of over 500 plant species, including various crops, vegetables, fruits, ornamental flowers, medicinal plants, etc., resulting in an average economic loss of 25% [8], and even 80% in severe cases [7]. Furthermore, its sclerotia can survive in soil for over five years even in the absence of hosts [7]. At present, biological control agents including Rhizobium [9], Trichoderma [10], and Streptomyces [7], and chemical control fungicides like tebuconazole and flutolanil [11], have been reported. However, biocontrol has the problem of inefficiency and fungicide utilization leads to bio-magnification in the food chain and environmental pollution [9]. Therefore, new control strategy is urgently needed.
Here, we employed a genetic approach to identify the members of COI, JAZ, and MYC family of D. catenatum, respectively, and subsequently assessed the phylogenetic relationship, gene structure, and domain architecture, and then showed exogenous MeJA enhanced D. catenatum resistance to S. delphinii through phenotype identification and gene expression profiling in the core JA pathway. The results provide the basis for further functional characterization of DcCOIs, DcJAZs, and DcMYCs, and in particular, their role in D. catenatum response to S. delphinii.

Exogenous jasmonates enhanced D. catenatum resistance to S. delphinii
Pre-treated with 1 mM MeJA for 4 h, four-month-old D. catenatum plantlets were inoculated with S. delphinii mycelia suspensions. Different responses to the disease infection were observed in control and MeJA-pretreated plantlets (Fig. 1). The symptoms, water-stained and brown necrotic lesions, were much more severe in the control than in pre-treated plantlets at 60 hpi (Fig. 1A). With prolonged hpi, the disease became more and more serious, but disease development was much slower in the pre-treated plantlets compared with the control. At 144 hpi (6 dpi), the disease index was 82.50 and 33.33% for the control and MeJA pre-treated plantlets, respectively (Fig. 1B). While almost all control plantlets were nearly dead and covered with white sclerotia at 240 hpi (10 dpi), pre-treated plantlets showed less necrotic lesions and no visible mycelia or sclerotia (Fig. 1A).

Identification of D. catenatum core JA signaling pathway genes
In view of the above-stated phenotypic results, we screened and identified the JA pathway genes in D. catenatum. The COI protein sequences of Arabidopsis and rice were used as queries to BLASTP search against the D. catenatum and its close relative P. equestris and A. shenzhenica datasets. In total, two COI candidate protein sequences were obtained in the D. catenatum genome sequences, while only one was identified in the other two orchid species, respectively, suggesting that specific duplication of COI family genes happened in D. catenatum (Tables 1, S1, and S2). To understand the role of gene functions, subcellular localization prediction for DcCOI, PeCOI and AsCOI genes using WoLE PSORT were performed. It was showed that DcCOI and PeCOI genes exhibited cytoplasmic localization and AsCOI exhibited chloroplast localization. Moreover, the MW and pI of the COI proteins were calculated using the Protparam tool in ExPASy. The MW of COIs varied from 66,194.57 to 76,380.33 Da, and the pIs ranged from 6.31 to 8.11. The length, MW, pI of DcCOIs and PeCOI proteins were similar, and lower than that of AsCOI, suggesting that COI protein was more conservative in epiphytic orchid.
A genome-wide search of TIFY family genes in D. catenatum, P. equestris, and A. shenzhenica datasets yielded 18, 20, and 13 non-redundant genes that were found to harbor the TIFY domain (Tables 1, S1, and S2). Of these TIFY domain-containing genes, 13, 14, and 7 contained both the TIFY domain and the Jas motif (PF09425) in D. catenatum, P. equestris, and A. shenzhenica, respectively, which were designated as JAZ proteins. The JAZ genes in D. catenatum and P. equestris were almost twice the number of A. shenzhenica, which is the original group of orchid [46]. It was observed that the length of the DcJAZ proteins ranged from 123 (LOC110097296) to 401 (LOC110103933) amino acid (aa) residues with an average length of 225 aa. The molecular weight ranged from 13,912.86 to 43,134.7 Da and the pI values varied from 5.50 to 9.34. Subcellular location prediction showed that 10 DcJAZ proteins were localized in the nucleus. Similarly, most JAZ proteins of the other two orchid species were predicted in the nucleus, indicated that JAZ protein mainly functions in the nucleus.
After the removal of redundant gene, 12, 14, and 9 non-redundant MYC genes were identified in D. catenatum, P. equestris, and A. shenzhenica, respectively (Tables 1, S1, and S2). The lengths of 12 DcMYC proteins were ranged from 381 to 662 aa with an average length of 554 aa. The predicted MWs of each DcMYC protein were ranged from 42,301.88 Da to 74,540.42 Da, and the corresponding pIs were changed from 5.18 to 6.92. The pI values of DcJAZs and PeJAZs were much lower than that of AsJAZs, showed that members of the MYC family are also relatively conservative in epiphytic orchids. The predicted subcellular location of DcMYC proteins suggested that all the D. catenatum MYCs were localized in the nucleus, consistent with their major roles as transcriptional factors. In addition, all these JA signal pathway genes were named after their Arabidopsis homologs.

Phylogenetic and structural analyses
To identify the classification and evolutionary patterns of COIs, TIFYs and MYCs in D. catenatum, we used COI, TIFY, and MYC proteins in A. thaliana, O. sativa, and its close relative P. equestris and A. shenzhenica as references for phylogenetic analysis, respectively.
Eight COI proteins from the above five species were clustered into three groups, COII-COIIII, and the four protein sequences from three orchid species clustered into the same group ( Fig. 2A), revealing that COI protein is highly conserved in evolution among orchid species, a similar phenomenon was also observed in the members of TIFY and MYC family as mentioned later. Conserved domain analysis revealed that all the COI proteins contain typical F-box and 10-14 tandem LRR repeats (Figs. 2B and S1). DcCOI1b, PeCOI1, and OsCOI1a contained 13 LRRs with similar location distribution ( Figure  S1), indicating that DcCOI1b and PeCOI1 maybe harbor a similar function with OsCOI1a. The COI genes contain 2 ~ 4 introns (Fig. 2C), which are extremely long in the three orchid species compared with those of Arabidopsis and Oryza, as attributed to a large number of repeated sequences in these three orchid genomes [46][47][48].
As shown in Fig. 3, TIFY proteins of the above five species were classified into eight groups (JAZI, JAZII, JAZIII, JAZIV, JAZV, TIFY, ZML, and PPD). Among them, the classification of JAZ subfamily is consistent with that of Arabidopsis [42]. As expected, the homologous genes in three orchid species always cluster together within one branch. DcTIFYs distributed in seven groups except JAZIII, indicating that the JAZIII branch was lost in the divergence from the D. catenatum and P. equestris. JAZI was the largest group with 5, 6, 2, 10, and 4 genes in D. catenatum, P. equestris, A. shenzhenica, O. sativa, and A. thaliana, respectively. The high number of JAZI genes in D. catenatum, P. equestris, and O. sativa is consistent with the view that significant expansion of JAZ I group genes in monocots [13], but only two members in A. shenzhenica indicated that the basal species in the orchid have undergone gene loss during the evolution [46]. Subsequent structural analysis focused on JAZ subfamily ( Fig. 4B and C). The genomic sequences of most of these genes were less than 5 kb, but DcJAZ6 had the longest genomic sequence of approximately 16 kb. The clade JAZV had the longest open reading frame (ORF), whereas JAZIV clade had the shortest ORF. Insights into the phylogeny and gene structures of JAZ genes indicated that proteins that are closely related phylogenetically have similar gene structures. In addition, all of the JAZ genes had symbolic ZIM domain and Jas motif (Fig. 4B, D and E). The sequence logo for the TIFY domain (Fig. 4D), generated using all DcJAZs, showed that the D. catenatum TIFY domain was Table 1 Characteristics of core JA signaling pathway genes in D. catenatum  remarkably conserved and very similar to other plant species [22,[49][50][51]. Besides, the Jas motifs of DcJAZs display the consensus sequence SLX 2 FX 2 KRX 2 RX 5 PY (Fig. 4E), which involved in the interaction with COI and MYCs [52]. Consist with AtJAZs, the JAZIV clade proteins contained an EAR motif that mediates direct interaction with TPL without NINJA (Fig. 4B). In addition, same to AtJAZI and AtJAZIII group in Arabidopsis, DcJAZI clade showed an N-terminal CMID sequence (Fig. 4B), however, no CMID was identified in DcJAZ III clade. The motif distribution was provided in Fig. 4B. JAZ genes within the same clade were usually found to share a similar motif composition, which was conspicuous in class JAZIV. Motif analysis showed conserved motif was not necessarily within a conserved functional domain. The functions of most of these conserved motifs remain to be elucidated.
A MYC family phylogenetic tree was generated using 12 DcMYCs, 14 PeMYCs, 9 AsMYCs, 8 OsMYCs, and 14 At MYCs. These MYCs were classified into five subgroups with highly bootstrap values (Fig. 5). The class MYCII was the largest, while the class MYCIV only has one member. Four, 5, and 3 DcMYCs were clustered into MYCI, MYCII, and MYCIII, respectively. However, no DcMYC was detected in MYCIV and MYCV, which may be experienced gene loss in evolution. It was worth noting that, in MYCI group, four MYC2-like proteins in D. catenatum were clustered with the other three monocots, while AtMYC2/3/4/5 were clustered in a single branch, indicating that gene replication was carried out after dividing into dicots and monocots, resulting in paralogs (Figs. 5 and 6A). There were five D. catenatum JA-ASSOCIATED MYC2-LIKE (JAM) members in MYCII clade, which orthologs in Arabidopsis is a negative regulator of JA signaling pathway. MYCII clade members had similar motif distribution and gene structure with MYCI clade members. MYCI and MYCII genes usually consist of only one exon, significantly less than the other three clades, revealing the MYC orthologs were remarkably conserved among different species (Fig. 6C). The intron-exon structure of different DcMYC family members was diverse, while the same subclade genes were similar or same, such as MYC2-like (DcMYC2a/2b/2c/2d) and JAM4-like (DcJ AM4a/4b/4c).

Tissue-specific expression patterns of core JA signaling pathway genes
To gain insights into the potential roles of core JA pathway genes during development in D. catenatum, the published transcriptomic data were investigated in various tissues and organs, including vegetative (leaf, root, green root tip, the white part of root, and stem) and reproductive (flower bud, sepal, labellum, pollinia, and gynostemium) organs. According to hierarchical clustering (Fig. 7A), DcCOI1a and DcCOI1b exhibited constitutive expressions with high levels in all tissues and organs except pollinia. Results of JAZ family members indicate that the expression profiles of DcJAZs could be split into two groups, TJ1 and TJ2 (Fig. 7B). Members of the TJ1 group (DcJAZ1/2/6/8) usually featured high expressions in most tissues. Exceptionaly, all TJ1 genes showed low expressions in pollinia, and DcJAZ1and DcJAZ2 exhibited low expression in leaf. The TJ2 genes (DcJAZ3/4/5/7/9/10/11/12/13) witnessed low expression levels in the vegetative organs, while some of these genes showed high transcriptional levels in special reproductive organs, such as DcJAZ7 in all reproductive organs except pollinia, DcJAZ3 in pollinia and gynostemium, DcJAZ12 in pollinia, and DcJAZ5/11/13 in sepal and gynostemium. These findings suggested that TJ1 genes might play essential roles in plant growth and development and TJ2 genes displayed prominent functions in flower development. The expression characteristics of MYC family genes in different tissues were quite diverse (Fig. 7C). The transcript abundance of DcMYCs could be also divided into two groups, TM1 and TM2. The majority of TM1 genes showed intermediate expressions in most of the detected tissues, whereas several expressed at high levels in specific tissues, such as DcGL3 in leaf, DcJAM4b and DcEGL3 in gynostemium, DcMYC2c in pollinia and gynostemium, DcJAM1a in sepal and gynostemium, DcJAM1b and DcMYC2b in pollinia, suggesting that these genes might play essential roles in specific tissue and organ development. TM2 genes (DcMYC2d, DcJAM4a, DcJAM4c, and DcTT8) showed low expressions in most tissues. However, DcMYC2d are highly expressed in sepal and gynostemium, illustrating it might perform a dominant function during sepal and gynostemium development.

Transcriptome sequencing
To gain further insights into the role of exogenous MeJA application in the enhanced resistance to S. delphinii in D. catenatum, a comparative transcriptome analysis was performed. Four leaf samples were obtained at 24 h after inoculated with sterile distilled water (mock) or S. delphinii (P1) from plantlets pre-treated with 0 or 1 mM MeJA (CK, P1, JA, and P1 + JA), respectively. All samples used for transcriptome sequencing have three biological replicates (Table S3). Using the HiSeq4000 sequencing platform, 107.83-148.06 million raw reads were generated for each RNA sample. Based on the D. catenatum genome, the number of mapped clean reads was 47.31 ~ 69.65 million (Table S4). Comparison of the transcript abundances in the leaves of CK, P1, JA, and P1 + JA, 5255 differentially expressed genes (DEGs) were detected (Table S5). Comparison analysis of the plantlets inoculated with S. delphinii pre-treated by MeJA (P1 + JA) and those only MeJA pre-treated (JA), 982 DEGs were identified, with 825 significantly upregulated and 157 significantly downregulated genes. Comparing the control plantlets and MeJA-pretreated plantlets after S. delphinii inoculation (P1 vs P1 + JA) uncovered 848 genes significantly up-regulated in the latter (Table S5).

JA pathway genes significantly responds to S. delphinii
Comparative transcriptome changes for genes involved in JA pathway response to pathogen and MeJA were investigated (Table S6). Figure 8 shows the S. delphinii (P1) and MeJA-induced expression pattern changes of 31 related genes in this pathway. Gene expression levels were illustrated using a heatmap and estimated by log scale. The transcription levels of the two DcCOI1 genes did not shown significantly changes after MeJA treated or S. delphinii infected ( Fig. 8 and Table S6). This result may be attributable to the fact that COI1 as JA receptor is functionally regulated mainly in protein level but not in transcription level. Most of the DcJAZ genes were significantly upregulated in three treatment samples (P1, JA, P1 + JA) compared with the control (CK). The expression pattern of these JAZ genes can be divided into four categories (Fig. 8). The PJ1 class genes (DcJAZ1/2/5/6/13) were significantly induced by MeJA or S. delphinii, respectively, and the expression had an additive effect when MeJA and S. delphinii were both treated; the PJ2 class genes (DcJAZ4/11/12) were significantly upregulated by MeJA or S. delphinii respectively, but it was more significantly response to S. delphinii than MeJA. However, there was no evidently expression difference between P1 and P1 + JA, indicating that these genes are responsive to S. delphinii but not MeJA; the PJ3 class genes (DcJAZ3/8/9/10) were only induced by MeJA, suggesting that these type genes did not respond to S. delphinii, and may be involved in the growth and development process related to JA pathway; the fourth type of genes (DcJAZ7) was not detected obvious transcriptional changes. The expression of MYC family genes was more complicated, and its expression profiles could also be divided into four types (Fig. 8). The expression patterns of PM1 (DcMYC2b and DcGL3) and PM3 (DcJAM1a) group were consistent with PJ1 and PJ2 classes of JAZ genes, respectively. The PM2 (DcMYC2c/2d) class genes showed that S. delphinii inhibited but MeJA significantly induced expression, indicating that this type of genes was a negative regulator in D. catenatum response to S. delphinii; the PM4 class genes were significantly downregulated in samples P1, JA, and P1 + JA, which was consistent with their role as negative regulators of JA pathway. However, JA-responsive genes, DcPR3 (LOC110093420) and DcLOX2 (LOC110106362) showed significantly induced in Real-time quantitative PCR of 13 representative genes confirmed the RNA-sequencing data (Fig. 9). DcCOI1a and DcCOI1b had no markedly expression changes. DcJAZ1, DcJAZ2 and DcJAZ5 were tightly co-expressed with DcMYC2b, with their transcript levels significantly upregulated in P1, JA, and P1 + JA treatment samples. DcJAM4b showed negative regulation, while the transcript levels of DcJAZ4 and DcJAM1a were high in P1 and P1 + JA samples but low in JA treated sample. DcPR3 and DcLOX2 were significantly induced in three treated samples. DcMYC2c induced by JA but not seriously respond to S. delphinii, while MYC2c significantly downregulated after P1 treatment in transcriptome data. The DcJAZ8 was significantly upregulated in P1 + JA, which was inconsistent with transcriptome expression data.

Discussion
Southern Blight is one of the most destructive fungal diseases against D. catenatum planting, causing almost no yield [5]. However, an understanding of the mechanism(s) underlying D. catenatum resistance to S. delphinii has not been reported. After decades of research, JA has been firmly established as a central role in mediating defense response and facilitate the adaptation of plants to a wide range of biotic and abiotic stresses, especially in defenses to necrotrophic pathogens and chewing insects [12,42]. In this study, D. catenatum plantlets pre-treated with MeJA obviously alleviated S. delphinii infection symptoms, indicating that the MeJA was useful for enhancing D. catenatum resistance to S. delphinii (Fig. 1). The F-box protein COI1 was isolated in 1990s, which is the first milestone in JA signaling pathway [22,53]. Both JA related pathways are depend on the COI1 gene. In Arabidopsis, a single COI1 (At2939940) gene was characterized and coi1 mutant exhibits male sterile, resistant to a coronatine-producing strain of pseudomonas syringae, insensitive to MeJA, and sensitive to B. cinerea [53,54]. In the rice genome, three COI homologs were identified [55], i.e., OsCOI1a (Os01g0853400), OsCOI1b (Os05g0449500), and OsCOI2 (Os03g0265500). OsCOI1a and OsCOI1b share over 80% amino acid sequence similarity, while OsCOI2 shows fewer sequence identities (63%) with the two OsCOI1 proteins. Moreover, OsCOI1 RNAi lines, targeting a sequence conserved in OsCOI1a and OsCOI1b, show altered plant height, internode length, grain length, and increased susceptibility to herbivores [56,57]. Here, 2 COI1 genes were identified in D. catenatum. Compared with Arabidopsis, the number change of COI genes in D. catenatum may attribute to undergone two wholegenome duplication (WGD) events (a τ WGD shared by most monocots and a recent WGD specific to the Orchidaceae lineage [46,58]. The gene structure and motif distribution of the pair of DcCOI1 genes, DcCOI1a/1b, were conserved. Moreover, DcCOI1a/1b expressed in all examined tissues with similar expression patterns (Fig. 8), indicating that they might have experienced subfunctionalization [59]. It is intriguing that DcCOI1a/1b expressed in pollen were significantly lower than the other tissues. Previous study found that COI performs functions at the protein level and a proper protein level of COI1 is important to maintain distinct JA responses [60].
The plant-specific group JAZ proteins plays a pivotal role in MYC TFs activity change and has emerged as nodes for signal crosstalk in hormone response pathways. Arabidopsis genome contains 13 JAZ members that are classified into five groups (I-V) [42]. In the present study, 13 DcJAZ proteins were identified and classify based on the homology with that of Arabidopsis ( Fig. 4A and Table 1). However, no DcJAZs were found in JAZIII group. This result is inconsistent with the view that all the five groups exist in all angiosperms [13]. We speculated that group III genes might have been experienced specifically lost during the evolution of D. catenatum. The domain architecture of JAZI group proteins (DcJAZ1-5) was more similar to AtJAZ1 and AtJAZ2, containing a CMID domain but no EAR domain [42]. Among them, DcJAZ1/2/5 were simultaneously significantly induced by P1, JA, and both (PJ1 expression mode, Fig. 8), indicating that these three DcJAZs played a critical part in enhancing the resistance to S. delphinii in D. catenatum. DcJAZ4 exhibited specific significantly respond to S. delphinii but weaker respond to MeJA treatment (PJ2 expression mode, Fig. 8), indicating that DcJAZ4 acted more as a hub to interact with other pathways (ethylene-JA, salicylic acid (SA)-JA etc.) to enhance the defense responses to pathogens in D. catenatum. DcJAZ3 showed PJ3 expression pattern and was significantly induced by JA instead of S. delphinii, speculating that DcJAZ3 might be involved in JA-mediated growth and development pathways. The single member of JAZII group, DcJAZ6, exhibited PJ1 expression pattern but no significant changes detected. Its Arabidopsis homologous gene mutant jaz12-1 shows normal JA responses and performs a redundant function in the JA response. Therefore, DcJAZ6 might function redundantly with DcJAZ1/2/5 in D. catenatum. The JAZIV group with five members had the shortest protein length in all the five groups ( Fig. 4B and Table 1). Its close Arabidopsis homologues AtJAZ7, AtJAZ8, and AtJAZ13 harbors a divergent Jas motif that interacts weakly with COI1 but is able to interact with MYC2, and behaves as a constitutive repressor in the jasmonate pathway [25,32,33]. The DcJAZs in this group showed three expressing patterns (PJ1, PJ2, and PJ3), meaning they functions in different JA pathways act as inert JAZ proteins. The JAZV group members (DcJAZ7 and DcJAZ8) showed different structures with the other four groups with the longest protein length and the highest number of exons. In Arabidopsis, group JAZV proteins includes AtJAZ3, AtJAZ4, and AtJAZ9 [42]. Atjaz9-1 mutant has normal JA-mediated responses but a specific role in the JA-GA crosstalk, acting functional redundancy in JA responses [33,61]. The knockout lines Atjaz4-1 were hypersensitive to JA-induced leaf senescence [62]. Combined with the expression profile of DcJAZ7 (no obvious change) and DcJAZ8 (significantly response to MeJA but not S. delphinii), suggesting that these two genes mainly participated in JA-mediated growth and development in D. catenatum. All in all, further in-depth molecular dissection of DcJAZ1, DcJAZ2, and DcJAZ5 will enrich the defense mechanisms of D. catenatum response to S. delphinii.
MYC2 and its closely related paralogs MYC3, MYC4, and MYC5 orchestrated transcriptional reprogramming, which is a central theme of JA pathway [27,63]. Although the molecular components (COI1/JAZ/MYC2 complex) of JA signaling pathway in plants are relatively conserved, but MYC2 regulates downstream genes is species-specific [31,37]. In Arabidopsis, MYC2 differentially regulates the expression of two groups of JA-induced genes. MYC2 activates NAC019 further positively regulates wound response genes (first group) but inhibits the Ethylene-Response-Factor1 (ERF1) to further negatively regulates pathogen response genes (second group) [64]. Arabidopsis myc2/jin1 mutant shows increased resistance to B. cinerea and P. cucumerina, and sensitivity to insect pests [12,29]. In tomato, MYC2 positively regulates both of these two group genes by activating JA2L and ERF.C3 transcription factors, respectively, and the necrosis area of the tomato MYC2-RNAi line inoculated with B. cinerea was significantly larger than those of the wild type [31]. Arabidopsis MYC2 and its tomato homolog have distinct action modes. It might be due to recruiting different partner proteins in transcriptional regulatory activities of pathogen response genes [31,45]. In this study, we identified 12 MYC members with bHLH domain and bHLH-MYC_N domain and classified them into five groups. MYCI group genes (DcMYC2a/2b/2c/2d) had close homology with AtMYC2/3/4/5. The domain architecture and gene structure of DcMYC2a, DcMYC2b, and DcMYC2c were similar, but shorter protein lengths with compact domain architecture was found in DcMYC2d, which consistent with that DcMYC2a, DcMYC2b, and DcMYC2c had a more closely phylogenetic genetic relationship. However, the expression patterns of these four duplicated DcMYC2 genes were dramatic differences. DcMYC2a exhibited PM4 pattern with negative regulation under both MeJA and S. delphinii. DcMYC2c and DcMYC2d showed upregulated by JA, but repressed by S. delphinii, suggesting these two genes might involve in JA-regulated growth and development pathways. DcMYC2b showed PM1 expression pattern and were positively regulated by MeJA and S. delphinii, suggesting that MYC2b may act as positive regulator and play a key role in JA-mediated defense to S. delphinii stress in D. catenatum. We also detected the expression level of MYC2-regulated pathogen defense genes (AtPDF1.2 orthologous gene in D. catenatum was not found), DcPR3, DcLOX2, DcTAT3, and DcVSP2. Transcriptome results showed that DcPR3 and DcLOX2 significantly response to S. delphinii, which was validated in qRT-PCR analysis (Figs. 8 and 9). Therefore, we speculated the action mode of DcMYC2b was same to tomato MYC2. It worth further investigate its direct targets by high-throughput ChIP-seq method or yeast onehybrid system. In MYCII group, five JAM genes, termed the MYC2-TARGETED BHLH1 (MTB) genes in tomato [37], were isolated. In Arabidopsis, the bHLH subclade IIId proteins, bHLH17/JAM1, bHLH13/JAM2, bHLH3/ JAM3, bHLH14/JAM4, were found to compete with MYC2-like TFs by binding to its target gene promoters [37,65,66]. MYC2-like TFs interacts with the JAZ protein through the N-terminal JID domain. Adjacent to JID is the transcriptional activation domain (TAD), which is responsible for the interaction with MED25. MED25 is a subunit of the mediator complex, which recruits general transcription machinery such as RNA polymerase II to achieve the transcription factor activity of MYC and initiate its transcription of target genes. The MID region in the TAD region is necessary for the interaction between MYC2 and MED25. JAMs-like TFs have a normal JID domain and an altered MID domain (AMID) and can interact with JAZ but cannot bind to MED25 [37]. Same to Arabidopsis, the MID-corresponding region of DcJAMs had been altered (Fig. 6D). Sequence alignment indicated that the MID was generally conserved in MYC2-like proteins and that the AMID was conserved in JAMs-like proteins in Arabidopsis and D. catenatum.
Notably, there was some sequence divergence in AMID domain between DcJAM4s and the other DcJAMs-like proteins and their functional divergence is worthy to further investigation. Transcriptome analysis revealed that four of this group genes (DcJAM1b/4a/4b/4c) were downregulated after S. delphinii, MeJA, and both treatment, which was consistent with the role as the negative regulators of JAMs-like TFs. However, DcJAM1a showed a significantly improved expression response to S. delphinii, suggesting a distinct regulated mechanism.
Based on these observations, we propose a model of D. catenatum JA signaling pathway response to S. delphinii. Upon S. delphinii elicitation, D. catenatum elevated JA-Ile levels promote SCF DcCOI1a/1b -dependent degradation of DcJAZ1/2/5 repressors, thereby DcMYC2b rapidly and directly regulates the transcription of downstream DcMYC2b-targeted TFs (DcMTFs), which in turn regulate the expression of wounding-responsive (DcPR3) or pathogen-responsive genes (DcLOX2, Figure S2). Further functional study of the DcMTFs will help to clarify this issue.

Conclusion
In this study, pre-treated with 1 mM MeJA significantly improved the resistance of D. catenatum to S. delphinii. Based on this phenotype, we identified 2 COI1, 13 JAZ, and 12 MYC proteins involved in core JA pathway of D. catenatum. These core JA pathway members of D. catenatum were classified and their evolutionary relationships with Arabidopsis, rice, and its closely related orchid species P. equestris and A. shenzhenica, were evaluated depend on phylogenetic relationship and substrate specificity. The expression profiles of D. catenatum core JA pathway genes in different tissues and organs was further analyzed. Comparative transcriptome analysis of the core JA pathway genes in D. catenatum planets under MeJA, S. delphinii, and combined treatments revealed that DcJAZ1/2/5 and DcMYC2b synergistically promote JA mediated defense response in D. catenatum. Therefore, further studies are required to understand how DcJAZ1/2/5 and DcMYC2b is regulated and whether and how it regulates other genes (DcMTFs) to control D. catenatum responds to S. delphinii. Our findings may help establish available avenues for plant breeding strategies aimed at southern blight disease resistance via JA-mediated enhanced plant immunity.

Biological materials
D. catenatum clonal cultivar 'Jingpin xianshi' (6A2B) is susceptible to S. delphinii, which was obtained by subculture from stem explants, placed in a modified 1/2 Murashige and Skoog (MS) medium [67], and cultivated in a growth room with a 12 h light/12 h dark photoperiod, 60% relative humidity, and 25 °C temperatures. The seedlings used for experiments were in same growth status and about 5 cm high, 6 leaves and 3 ~ 5 roots, and the stem diameter was about 0.3 cm.

MeJA treatment
To study the effects of MeJA on D. catenatum plantlets, four-month-old plantlets were spotted with 0.25% ethanol solution (control) or a solution containing 1 mM MeJA (Sigma-Aldrich, 392,707) prepared in 0.25% ethanol solution (1 mL of per plantlet). After MeJA pre-treatment for 4 h, D. catenatum plantlets were inoculated with S. delphinii (named P1), which was isolated and identified in our laboratory [5].

Pathogen challenge assays
S. delphinii grew on PDA medium at 25 °C for 6 days. Five 5 mm agar disks containing mycelia were collected and cultured in 200 mL PDB medium for 5 days at 25 °C with shaking (180 rpm).
Four-month-old D. catenatum plantlets pre-treated by MeJA or 0.25% ethanol solution as indicated above were sprayed with mycelia suspensions (2 mL per plantlet). Control plantlets in each treatment were inoculated with an equal volume of sterile distilled water. S. delphinii bioassays were performed at 28 ºC under a photoperiod of 12 h light/12 h dark.

Disease assessment after S. delphinii inoculation
The progress of S. delphinii infection was followed for 10 days by observing the development of necrosis in the infected leaves, which can be detected at 24 h post inoculation (hpi), and infection ratings were assigned to the inoculated plantlets: 0, no necrosis; 1, one or two leaves showing a small part of necrosis lesion; 2, one or two leaves showing a large area of necrosis lesion; 3, over three leaves showing necrosis and some leaves dropped off; 4, dead/decayed plant [68]. At least 10 plantlets were inoculated in each experiment. Experiments were repeated at least three times with similar results. The disease index was calculated with the following formula: disease index = ((Σ disease grade × the number of infected plantlets) / (total assessed plantlets × 4)) × 100. Photographs were taken at 60 and 240 hpi.
Genome-wide identification of core JA signaling pathway genes in some representative orchid species All the COI, TIFY, MYC sequences including genomic DNAs, coding sequence (CDS), and proteins in A. thaliana and Oryza sativa were downloaded from the plant genomics resource Phytozome v12 (http:// phyto zome. jgi. doe. gov/ pz/ portal. html). D. catenatum, P. equestris and A. shenzhenica genome sequences were retrieved from NCBI Genome (https:// www. ncbi. nlm. nih. gov/ genome/). To acquire the COI domain-containing sequences in D. catenatum, P. equestris, and A. shenzhenica, the protein sequences of A. thaliana [22] and O. sativa [69] were used as queries to search homologs against NCBI database using BLASTp tool. The hidden Markov model (HMM) profile of TIFY domain (Pfam accessions: PF06200) was downloaded from the Protein family (Pfam) database (https:// pfam. xfam. org/) [70], and used as queries to search for potential TIFY proteins in the D. catenatum, P. equestris and A. shenzhenica protein datasets by using HMMER3.0 [71] with an E-value cutoff of 1.0E-05. The MYC family protein was filtered for Pfam database identifiers of the bHLH domain (PF00010) and bHLH-MYC_N domain (PF14215), respectively. All the hits were further confirmed to remove the incomplete and redundant sequences.

Gene structure, domain architecture, conserved motif, and phylogenetic relationship analyses
The conserved motifs were investigated by MEME version 5.3 online tool (http:// meme-suite. org/ tools/ meme). The domain organization was analyzed using the Simple Modular Architecture Research Tool (SMART) and NCBI CD search program. The gene structure, protein domain architecture, and motif composition was visualized using the TBtools [72].

RNA-seq analysis
The leaf samples of D. catenatum plantlets were pretreated with 0 (control) or 1 mM MeJA were harvested for transcriptome sequencing at 24 hpi. Three independent RNA samples (biological replicates) were performed for each treatment. RNA was extracted using the Min-iBEST Plant RNA Extraction Kit (TaKaRa, Japan) according to the manufacturer's instructions. The paired-end sequencing was performed on an Illumina Hiseq4000.
Reads of above-described samples were aligned to the D. catenatum genome [47] using HISAT package [75], which initially remove a portion of the reads based on quality information accompanying each read and then maps the clean reads to the reference genome. StringTie was used to assemble the mapped reads of each sample and estimate the expression levels of all genes by calculating FPKM [76]. The differentially expressed genes were defined with |log 2 (fold change) |> 1 and with an adjusted P-value (q-value) < 0.05 by R package [77]. Heatmap was generated using TBtools software [72].

Quantitative real-time (RT-qPCR) analysis
Total RNA was extracted using the MiniBEST Plant RNA Extraction Kit (TaKaRa, Japan). The cDNA was reverse transcribed with the PrimerScript RT Enzyme Mix I kit (TaKaRa, Japan). The primers used for expression analysis were designed by Primer Premier 5 (Table S7). qPCR analysis was performed with SYBR ® Premix Ex Taq II (TaKaRa, Japan) on CFX96 Touch ™ Real-Time PCR System (BIO-RAD, USA) in three technical replicates each for three independent biological replicates. The DcAC-TIN was used as the internal control gene, which was stably expressed in D. catenatum plants and not affected by treatments and genotypes. The relative expression levels were evaluated automatically by the Bio-Rad CFX Manager (version 2.3) with 2 −ΔΔCT method [3].