Genome-Wide Identification of AP2/ERF Transcription Factor Family and Functional Analysis of DcAP2/ERF#96 Associated with Abiotic Stress in Dendrobium catenatum

APETALA2/Ethylene Responsive Factor (AP2/ERF) family plays important roles in reproductive development, stress responses and hormone responses in plants. However, AP2/ERF family has not been systematically studied in Dendrobium catenatum. In this study, 120 AP2/ERF family members were identified for the first time in D. catenatum, which were divided into four groups (AP2, RAV, ERF and DREB subfamily) according to phylogenetic analysis. Gene structures and conserved motif analysis showed that each DcAP2/ERF family gene contained at least one AP2 domain, and the distribution of motifs varied among subfamilies. Cis-element analysis indicated that DcAP2/ERF genes contained abundant cis-elements related to hormone signaling and stress response. To further identify potential genes involved in drought stress, 12 genes were selected to detect their expression under drought treatment through qRT-PCR analysis and DcAP2/ERF#96, a nuclear localized ethylene-responsive transcription factor, showed a strong response to PEG treatment. Overexpression of DcAP2/ERF#96 in Arabidopsis showed sensitivity to ABA. Molecular, biochemical and genetic assays indicated that DcAP2ERF#96 interacts with DREB2A and directly inhibits the expression of P5CS1 in response to the ABA signal. Taken together, our study provided a molecular basis for the intensive study of DcAP2/ERF genes and revealed the biological function of DcAP2ERF#96 involved in the ABA signal.


Introduction
Transcription factors (TFs) are key regulators for plants to respond to various environmental stimuli and transmit stress signals. Many kinds of transcription factors, such as APETALA2/Ethylene Responsive Factor (AP2/ERF) [1], bHLH [2], MYB [3], Dof [4] and WRKY [5], have been proven to play vital roles in biotic and abiotic stress responses [6]. AP2/ERF widely exists in the plant kingdom and is one of the largest transcription factor families in plants [7], which was first discovered in Arabidopsis and soon found similar transcription factors in tobacco [8]. As the most representative domain of AP2/ERF family proteins, the AP2 domain consists of one α-helix and three β-sheets, with a range of 60-70 amino acid residues [9]. The AP2 domain can bind to some special cis-acting elements, such as dehydration responsive elements (DRE), C-repeat element (CRT) and GCC box, to regulate downstream gene expression [7,9,10]. This family is mainly divided into four subfamilies: AP2 (APETALA2), DREB (Dehydration Responsive Element-Binding), ERF (Ethylene Responsive Element Binding protein) and RAV (Related to ABI3/VP) [11]. The AP2 subfamily contains two tandem AP2 domains, which have been shown to be widely involved in floral organ development and regulation [12]. The RAV subfamily contains an AP2 domain and a B3 domain, which are thought to be involved in biotic and abiotic stresses, as well as its known involvement in flowering regulation [13]. The DREB 2 of 21 and ERF subfamily both contain only one AP2 domain, but the 14th and 19th amino acids of their AP2 domains are different, resulting in a distinct affinity for different DNAs. DREB family proteins can specifically bind to DRE/CRT cis-acting elements, thereby regulating ABA, drought and low temperature responses, while the ERF subfamily proteins prefer to bind GCC-box related to ethylene response, disease resistance and abiotic stress [8].
Members of the AP2/ERF family have been identified and scaled in many species, including 146 in Arabidopsis thaliana [11], 163 in Oryza sativa [11], 134 in Fagopyum Tataricum [14], 364 in Salix matsudana [15], 163 in Zingiber officinale [16], 322 in Triticum aestivum [17], 189 in Panax ginseng [18] and 531 in Brassica campestris [19]. AP2/ERF transcription factor family has been confirmed to be involved in a variety of abiotic stress responses; for example, AtCBF1/2/3 control low temperature responses in Arabidopsis [20,21], Zm-DREB2a participates in heat response in maize [22], and GmERF4/7 mediate salt stress response in soybean [23]. Overexpression of OsERF71 in rice has also been found to change the root structure of plants and enhance tolerance to drought stress [24]. However, to date, no studies have clarified the role of AP2/ERF family genes in the response to environmental stress of D. catenatum, so it is very important to identify DcAP2/ERF by genome-wide identification approach.
D. catenatum, as a typical orchid plant, is highly valued because its stem contains many medicinal ingredients [25]. In addition, due to harsh growth circumstances, often attached to rocks, tree trunks and cliffs ( Figure S1), D. catenatum has strong resistance to various stresses [26]. Therefore, studying how D. catenatum resists and resolves stresses from nature may provide new inspiration for studying the mechanism of plants resisting stress. To date, only a few gene families have been identified in D. catenatum, which are involved in stress response, including MYB [15] and GRAS [27]. With the determination of the whole genome sequences of D. catenatum in recent years [28][29][30], it has become possible to systematically analyze the characteristics and functions of AP2/ERF family in D. catenatum.
In this study, we first performed genome-wide identification of DcAP2/ERF gene family using HMMER3.0 and a total of 120 DcAP2/ERF genes were identified. Second, all 120 genes were comprehensively analyzed by phylogenetic analysis, gene structure and domain visualization, and promoter motif prediction. Third, we presented the expression patterns of the AP2/ERF family in 9 different tissues and under low temperature stress. Finally, we detected the expression of 12 DcAP2/ERF genes under drought stress by qRT-PCR analysis and focused on analyzing the characteristics and functions of DcAP2ERF#96. Our research found that the heterologous expression of DcAP2ERF#96 causes at least 9 ABA downstream genes to be significantly inhibited, including P5CS1 and RD29A, which are usually used by researchers to measure the severity of abiotic stress faced by plants [31][32][33]. We also found the protein DREB2A, which interacts with DcAP2ERF#96 in D. catenatum. Together, our study reveals the vital role of AP2/ERF family in D. catenatum and provides several key candidate genes related to abiotic stress.

Identification of AP2/ERF Genes in D. catenatum
To identify the DcAP2/ERF family members, a combined analysis of genome-wide and full-length transcriptome-wide was performed using HMM software and SMART online. After removing the duplicate transcripts, a total of 120 AP2/ERF genes were finally screened and identified in the D. catenatum genome, and the basic information of these genes was analyzed and summarized in Table 1, including their protein length, MW, PI and subcellular location. Among these genes, the largest protein is encoded by DcAP2ERF#47 with 733 aa, while the smallest protein is encoded by DcAP2ERF#2 with 117 aa. (Table 1). The molecular weight (MW) of proteins ranges from 13.24 to 81.46 kDa and the isoelectric point (PI) varies from 4.35 (DcAP2ERF#39) to 10.10 (DcAP2ERF#64). According to the analysis of the aliphatic index and grand average of hydropathicity (GRAVY), the DcAP2/ERF family proteins are generally hydrophilic. To analyze subcellular localization, all AP2/ERF proteins were predicted using PSORT. The results showed that most DcAP2/ERF proteins (78.3%) are located in the nucleus, which is consistent with the localization characteristics of the transcription factor family. In addition, 18.3% of proteins were predicted to be located in cytoplasm, and 3.3% of proteins were likely to be located in mitochondria (Table 1).

Phylogenetic Analysis of DcAP2/ERF Families
To study the genetic phylogeny of AP2/ERF family in Orchidaceae, phylogenetic trees were constructed using sequences from Arabidopsis thaliana (146 genes), D. catenatum (120 genes) and Phalaenopsis equestris (118 genes) ( Figure 1A). As shown in Figure 1, all AP2/ERFs of the three species can be grouped into four subfamilies, namely, AP2, ERF, DREB and RAV, which suggests that the AP2/ERF family is evolutionarily conserved in the plant kingdom. From the view of family size, with the deepening of evolution, the number of genes in AP2/ERF family has been reduced from 139 in Arabidopsis thaliana to 118 in Phalaenopsis equestris and 120 in D. catenatum. The size of the RAV and DREB subfamily of D. catenatum and Phalaenopsis equestris was significantly smaller than that of Arabidopsis thaliana. The number of ERF subfamily members increased slightly compared with that of Arabidopsis thaliana, while the gene number and tree structures of the AP2 subfamily remained basically unchanged ( Figure 1B). These changes showed that AP2/ERF family is conservative in the history of plant evolution, while making corresponding changes with different plant growth environments. Considering that orchids usually have stronger environmental adaptability than Arabidopsis thaliana, the adjustment of these gene group structures may help them cope with more complicated biotic and abiotic stresses.

Structure Analyses of the AP2/ERF Gene Family in D. catenatum
To further understand the structural characteristics of DcAP2/ERF family genes, the conserved motif, domain, intron and exon of DcAP2/ERFs were further analyzed ( Figure  2). Using motif analysis, eight different motifs were predicted based on DcAP2/ERF protein sequences ( Figure S2). We summarized the different motif patterns of each subfamily, and the results showed that motif 1, motif 3 and motif 4 exist in all four DcAP2/ERF subfamilies ( Figure 2A). Compared with the ERF subfamily, the RAV subfamily lacks motif 2, while the DREB subfamily adds motif 7. Motif 7 only exists in the DREB subfamily and it is located outside the AP2 domain ( Figure S3A). For the AP2 subfamily containing two AP2 domains, these proteins were mainly divided into two categories. Category I is composed of motif 3, motif 2, motif 1, motif 4 (first AP2 domain) and motif 5, motif 1, motif 4 (second AP2 domain). Category II is composed of motif 8, motif 1, motif 4 (first AP2 domain) and motif 3, motif 2, motif 6 (second AP2 domain). These data suggest that the double AP2 domain of the AP2 subfamily is not a simple duplication of the single AP2 domain, but has its own unique characteristics. For domain and structure analysis (Figure

Structure Analyses of the AP2/ERF Gene Family in D. catenatum
To further understand the structural characteristics of DcAP2/ERF family genes, the conserved motif, domain, intron and exon of DcAP2/ERFs were further analyzed ( Figure 2). Using motif analysis, eight different motifs were predicted based on DcAP2/ERF protein sequences ( Figure S2). We summarized the different motif patterns of each subfamily, and the results showed that motif 1, motif 3 and motif 4 exist in all four DcAP2/ERF subfamilies ( Figure 2A). Compared with the ERF subfamily, the RAV subfamily lacks motif 2, while the DREB subfamily adds motif 7. Motif 7 only exists in the DREB subfamily and it is located outside the AP2 domain ( Figure S3A). For the AP2 subfamily containing two AP2 domains, these proteins were mainly divided into two categories. Category I is composed of motif 3, motif 2, motif 1, motif 4 (first AP2 domain) and motif 5, motif 1, motif 4 (second AP2 domain). Category II is composed of motif 8, motif 1, motif 4 (first AP2 domain) and motif 3, motif 2, motif 6 (second AP2 domain). These data suggest that the double AP2 domain of the AP2 subfamily is not a simple duplication of the single AP2 domain, but has its own unique characteristics. For domain and structure analysis ( Figure 2B,C), the characteristics of the gene members in each subfamily were consistent with previous reports [34]. It is worth noting that the intron structures of AP2 subfamily members are very complex, which is conducive to their flexible cutting and assembly so as to cope with different situations. The diversity of the AP2 domain model may be the result of gene evolution and selection. This partly explains why different subfamilies have preferences for different cis-acting elements [8]. 2B,C), the characteristics of the gene members in each subfamily were consistent with previous reports [34]. It is worth noting that the intron structures of AP2 subfamily members are very complex, which is conducive to their flexible cutting and assembly so as to cope with different situations. The diversity of the AP2 domain model may be the result of gene evolution and selection. This partly explains why different subfamilies have preferences for different cis-acting elements [8].  Combined with the secondary structure of the AP2 domain, we further analyzed the motif of a typical AP2 domain ( Figure S3B). Motif 3, motif 2 and the front half of motif 1 are distributed with one β-sheet, respectively. The back half of motif 1 and the first half of motif 4 form one α-helix. These secondary structures together form a complete AP2 domain. These findings are consistent with the previous description of the AP2 domain [35]. In addition, these results showed that although AP2/ERF family is complex at the motif level, it is still conserved in domain composition.

Cis-Acting Element Analysis of DcAP2/ERF Gene Promoters
To explore the putative functions of AP2/ERF family genes in D. catenatum, the 2000-bp upstream sequences of 120 genes were extracted to analyze the potential cis-elements. The predicted TF-binding motifs were classed into three categories: hormone-related elements, stress-related elements, and growth and development-related elements ( Figure 3). Hormone response, such as auxin-responsiveness, ABA-responsiveness, GA-responsiveness, MeJA-responsiveness and SA-responsiveness. From the view of hormone response-related elements, a total of 113 genes have hormone response elements in the promoter region (Table S3), indicating that DcAP2/ERF family genes may be widely involved in hormone response pathways. Among these elements, the number of elements responding to MeJA and ABA was enriched, reaching 185 and 149, respectively, which indicated that DcAP2/ERF genes may respond to biotic and abiotic stress rapidly. Stress response-related elements were associated with anaerobic induction, drought induction, low temperature responsiveness and anoxic specific induction. In addition, the promoters of 68 genes in DcAP2/ERF family contain growth-and development-related response elements associated with meristem expression, zein metabolism, endoperm expression, circadian control, cell cycle regulation, flavonoid biosynthetic gene regulation, and others. Meristem expression elements and zein metabolism regulation elements accounted for the largest proportion, reaching 32 and 31, respectively (Figure 3).

GO Annotation Analysis of DcAP2/ERF Family Proteins
To further understand the molecular function of DcAP2/ERF proteins, GO annotations were conducted in this study. According to gene ontology, genes or gene products have three main characteristics: cell composition, molecular function and biological pro-

GO Annotation Analysis of DcAP2/ERF Family Proteins
To further understand the molecular function of DcAP2/ERF proteins, GO annotations were conducted in this study. According to gene ontology, genes or gene products have three main characteristics: cell composition, molecular function and biological process. Their terminology labels can help us understand the protein functions [36]. Here, a total of 75 DcAP2/ERF proteins were assigned 20 GO terms, including 4 cellular component terms, 2 molecular function terms and 14 biological process terms ( Figure S4, Table S2). Under the cellular component category, 32 proteins were identified as 'cell' (GO: 0005623), 'cell part' (GO: 0044464) and 'organelle' (GO: 0043226). Under the molecular function category, 50 and 75 proteins were annotated for 'binding' (GO: 0005488) and 'transcription regulator activity' (GO:0140110), respectively. Under the biological process category, all annotated proteins are involved in 'metabolic processes' (GO: 0008152), 'cellular processes' (GO: 0009987) and 'biological regulatory' (GO: 0065007). In addition, there are 53 proteins involved in 'response to stimulus' (GO: 0050896). In summary, DcAP2/ERF proteins are a typical family of transcription factors with the potential ability to bind downstream DNA and regulate various physiological processes.

Expression Patterns of DcAP2/ERF Genes under Different Stresses
To study the expression pattern of DcAP2/ERF family genes in response to hormone and environmental stresses, the transcriptomic data of D. catenatum under MeJA and lowtemperature treatments were analyzed. In chilling stress, a total of 29 genes showed significant changes under chilling treatment, of which 16 genes were significantly upregulated, especially DcAP2ERF#52, DcAP2ERF#7 and DcAP2ERF#98 (Table S4), and 13 genes were downregulated ( Figure 4B), especially DcAP2ERF#84, DcAP2ERF#109 and DcAP2ERF#61. These data suggest that these genes may play important roles in the cold stress response and can be considered candidate genes for further study of cold stress biology.
For the MeJA treatment, the repetitions of three experimental groups and three control groups were clustered into two groups ( Figure 5A). Under MeJA treatment, 9 genes were upregulated, and 11 genes were downregulated ( Figure 5B). Among these genes, the expression of DcAP2ERF#44 and DcAP2ERF#71 was highly increased (about 18 times and 9 times, respectively) compared with the control (Table S5). This suggests that they may be strongly induced by JA signaling and perform important functions in D. catenatum.

Expression Patterns of DcAP2/ERF Genes in Different Organs
To study the tissue expression pattern and reveal the related function of DcAP2/ERF genes, we analyzed the transcriptome data of DcAP2/ERFs in different tissues ( Figure S5). Among all 120 DcAP2/ERF genes, 14 DcAP2/ERF genes were constitutively expressed, 11 genes tended to be expressed in roots (including green root tip and white part of root), stem or leaves, and 34 genes tended to be expressed in flower organs (including sepal, labellum, gynostemium and pollinia). In addition, by the cluster analysis of tissues and organs, the expression pattern of flower organs was distinguished from roots, stems, leaves and their related tissues. The expression pattern of gynostemium is similar to that of sepal among these DcAP2/ERF genes, except DcAP2ERF#74, DcAP2ERF#64, DcAP2ERF#55, DcAP2ERF#50, DcAP2ERF#22 and DcAP2ERF#26. It is worth mentioning that the expression pattern of DcAP2/ERF genes in pollen is obviously different from those in other tissues. Some genes, such as DcAP2ERF#117, DcAP2ERF#24, DcAP2ERF#40, DcAP2ERF#41 and DcAP2ERF#59, are specifically expressed in pollen, indicating that these genes play important roles in pollen development.

Expression Pattern of DcAP2/ERFs under PEG6000 Treatment
Many ABA response elements in DcAP2/ERF family gene promoters and the excellent drought resistance of D. catenatum allow us to suspect that some DcAP2/ERF family genes are involved in the drought-tolerant response. Based on the expression of these genes in different transcriptomes (Figures 4, 5 and S5), we finally selected 12 genes of interest to test their expression levels under drought stress ( Figure 6). After 20% PEG6000 treatment for 0 h, 2 h, 4 h and 8 h, respectively, leaves, stems and roots of D. catenatum were collected for RNA extraction and qRT-PCR analysis. As shown in Figure 6, most of the genes responding to drought are found in leaves, with 11 genes, while only 7 and 5 genes showed increased expression in stems and roots, respectively, under drought stress. Among these genes, four genes (DcAP2ERF#16, DcAP2ERF#50, DcAP2ERF#52, DcAP2ERF#78) showed obvious increased expression in all three tissues under drought stress. Interestingly, DcAP2ERF#1 and DcAP2ERF#86 only increased in leaves and stems, while they had no changes in roots, indicating that these genes may play different roles in D. catenatum in response to drought stresses. In addition, according to the differences in the expression levels and response times of these genes in the three tissues, we can infer that the three kinds of tissues showed different sensitivities in the face of drought stress.

Expression Patterns of DcAP2/ERF Genes under Different Stresses
To study the expression pattern of DcAP2/ERF family genes in response to hormone and environmental stresses, the transcriptomic data of D. catenatum under MeJA and lowtemperature treatments were analyzed. In chilling stress, a total of 29 genes showed significant changes under chilling treatment, of which 16 genes were significantly upregulated, especially DcAP2ERF#52, DcAP2ERF#7 and DcAP2ERF#98 (Table S4), and 13 genes were downregulated ( Figure 4B), especially DcAP2ERF#84, DcAP2ERF#109 and DcAP2ERF#61. These data suggest that these genes may play important roles in the cold stress response and can be considered candidate genes for further study of cold stress biology.  trol groups were clustered into two groups ( Figure 5A). Under MeJA treatment, 9 genes were upregulated, and 11 genes were downregulated ( Figure 5B). Among these genes, the expression of DcAP2ERF#44 and DcAP2ERF#71 was highly increased (about 18 times and 9 times, respectively) compared with the control (Table S5). This suggests that they may be strongly induced by JA signaling and perform important functions in D. catenatum.

DcAP2ERF#96 Encodes a Conserved AP2 Domain Transcription Factor Localized in the Nucleus
To further confirm the related genes involved in drought stress in the AP2/ERF family, we measured the biomass differences in several tissues of D. catenatum under drought stress and normal conditions (Figure 7). The results showed that the growth of the stem was significantly inhibited under three-month drought conditions ( Figure 7A). In addition, stems, as the main medicinal parts of D. catenatum, had the most obvious changes in biomass under drought stress ( Figure 7B). Based on this finding, we selected DcAP2ERF#96 as a potential gene for further study, which showed specifically decreased expression in the stem under PEG6000 treatment, as shown in Figure 6. Sequence analysis revealed that DcAP2ERF#96 is an AP2-like ethylene-responsive transcription factor. Subcellular localization in D. catenatum protoplasts and tobacco leaf cells showed that the DcAP2ERF#96 protein was located to the nucleus ( Figure 8A,B). Previous studies have shown that AP2/ERF transcription factors can bind to GCC (AGCCGCC), DRE (GCCGAC), and CRT (ACCGAC) boxes to regulate downstream gene expression [6]. To confirm this, a Y1H assay was performed, and the results showed that DcAP2ERF#96 has a strong ability to bind to the CRT box. In addition, it also has weak binding activity to DRE and GCC box ( Figure 8C), indicating that the binding motifs of DcAP2ERF#96 were conserved in D. catenatum. A transactivation activity assay showed that DcAP2ERF#96 had self-activating properties, and the self-activating regions existed in the N terminal (1-185 aa) and C-terminal (280-429 aa) of DcAP2ERF#96, while not the middle region (186-279 aa), which contain two AP2 domains ( Figure 8D). These results confirm that DcAP2ERF#96 is a nuclear-localized transcription factor.

DcAP2ERF#96 Interacted with DcDREB2A and Negatively Regulate ABA Signaling in Arabidopsis
To further investigate the function of DcAP2ERF#96 in response to drought, the two overexpression lines of Arabidopsis thaliana #3 and #6 with the highest gene expression levels were selected for further study. Compared with wild-type (Col-0), the leaf area and root length of the two overexpression lines were significantly reduced in 1/2 MS with 10 µM ABA, suggesting that DcAP2ERF#96 had an ABA-sensitive phenotype ( Figure 9A-C). To further explore whether DcAP2ERF#96 participates in the ABA signaling pathway in response to drought, the expression of 9 genes related to the ABA signaling pathway was selected and detected under ABA treatment ( Figure 9D). Results showed that all 9 genes except AtADH1 were significantly upregulated after 10 µM ABA treatment in Col-0. However, the relative expression levels of these genes were significantly reduced in DcAP2ERF#96 overexpression lines compared with Col-0, which confirmed our hypothesis that DcAP2ERF#96 acts as a transcriptional repressor to negatively regulate ABA signaling under drought stress in plants. Cis-elements analysis of these 9 gene promoters showed that three gene promoters contain DRE, CRT and GCC box: The AtP5CS1 promoter contains one GCC box; the AtRD29A promoter contains three CRT boxes and one DRE box; and the AtRAB18 promoter contains one DRE element. All the elements are located within the range of 500-bp upstream sequences of genes. (Figure 10A) Y1H results showed that DcAP2ERF#96 can directly bind to the promoter of P5CS1 and RD29A in yeast, but there is no indication that it can combine with RAB18 promoter (Figure 10B). At the same time, our Y2H results confirmed that DcAP2ERF#96 interacts with DREB2A-1 (LOC110112950) and DREB2A-2 (LOC110113533), which have been reported to regulate the expression of RD29A in Arabidopsis [37] (Figure 10C). These results suggested that DcAP2ERF#96 negatively regulates the expression of P5CS1 and RD29A in Arabidopsis thaliana, thereby affecting its response to the ABA signal ( Figure 10D). a strong ability to bind to the CRT box. In addition, it also has weak binding activity to DRE and GCC box ( Figure 8C), indicating that the binding motifs of DcAP2ERF#96 were conserved in D. catenatum. A transactivation activity assay showed that DcAP2ERF#96 had self-activating properties, and the self-activating regions existed in the N terminal (1-185 aa) and C-terminal (280-429 aa) of DcAP2ERF#96, while not the middle region (186-279 aa), which contain two AP2 domains ( Figure 8D). These results confirm that DcAP2ERF#96 is a nuclear-localized transcription factor.

Discussion
D. catenatum, a precious Chinese herbal medicine of Orchidaceae, has great medicinal and economic value [27]. AP2/ERF family genes are widely involved in various physiological activities and biotic/abiotic stress responses of plants. However, the role of AP2/ERF family genes in D. catenatum remains uncertain. In this study, 120 DcAP2/ERF genes were detected, and their phylogenetic relationship, domain composition, gene structure, cisacting elements and gene ontology were analyzed. Furthermore, the tissue expression profile and different responses of DcAP2/ERF family genes under cold environment, MeJA treatment and simulated drought were also explored. Based on these results, we discussed the potential functions of some genes in D. catenatum.
In phylogenetic analysis, we found obvious structural similarities of phylogenetic trees and adjustment of subfamily proportion among Arabidopsis thaliana, D. catenatum and Phalaenopsis equestris (Figure 1), which indicated that orchidaceae had a conserved evolutionary process. We also found that the number of introns in the AP2 subfamily was quite greater than those in other subfamilies (Figure 2). Previous studies have shown that introns may be gradually lost in the process of evolution [38]. According to this theory, we speculate that the AP2 subfamily may be one of the origins of the evolution of other subfamilies. Other studies have shown that the AP2 domain in plant genes may be derived from bacterial or viral HNH endonucleases, and various subfamilies have gradually been generated during the long evolutionary process [39]. In general, the evolutionary relationship between the various AP2/ERF subfamilies still remains to be explored.
Transcriptional regulation plays a leading role in the regulation of gene expression, which is mainly controlled by its cis-acting element on the gene promoter [40,41]. We analyzed the cis-acting elements in all DcAP2/ERF family promoter regions ( Figure 3) and described their functions by gene ontology ( Figure S4). The results showed that the cis-acting elements responding to MeJA and ABA accounted for the largest proportion, which was similar to the promoter characteristics of AP2/ERF family in other plants like wheat [17] and ginger [16]. At the same time, gene ontology also analyzed DcAP2/ERF proteins that have typical features of the general transcription factor families and have the ability to respond to stimuli and signal transduction.
Drought is a common abiotic stress that can cause serious damage to the growth and development of plants [42]. Drought stress will directly lead to cell dehydration, which produces an ABA signal and causes a series of tolerance responses in plants, including promoting stomatal closure of guard cells and regulating the expression of related genes [43]. Previous studies have suggested that some AP2/ERF family genes are closely related to drought stress and ABA [44,45]. We selected 12 genes of interest of DcAP2/ERF family, and tested their expression through PEG6000 treatment simulating an arid environment ( Figure 6). The results confirmed our hypothesis that the expression of most selected genes increased under simulated drought conditions. As the most strongly upregulated gene, the homologous gene of DcAP2ERF#1 in Arabidopsis has been confirmed to control the synthesis of epidermal wax [46], and the increase of epidermal wax in D. catenatum leaves may be beneficial in reducing water loss and surviving longer under drought stress. In the stem, the changes in the expression levels of all tested genes were significantly smaller than those in leaves and roots, which may be due to the fact that stems are rich in polysaccharides, polyphenols and other substances and have strong resistance to drought stress. After comparing the biomass of different tissues of D. catenatum under drought conditions, the stem is considered to play a key role in the process of D. catenatum resisting drought stress (Figure 7). At the same time, we noticed a decrease in the expression of DcAP2ERF#96 in stems under simulated drought conditions ( Figure 6). Molecular biology and transgenic techniques including subcellular localization, Y1H, Y2H and transgenic experiments confirmed that DcAP2ERF#96 functions as a transcriptional repressor to regulate the ABA signaling pathway and interacts with DREB2A in D. catenatum. The downregulation of its expression in stems may be one of the self-rescue responses of D. catenatum under drought stress, which is conducive to the transmission of ABA signals in the whole plant. (Figures 8-10). In addition, the expression of some genes downstream of ABA, such as AtABI1, AtABI5 and AtADH1, were also significantly reduced ( Figure 9D), which may be caused by DcAP2ERF#96 indirect regulation, but this speculation remains to be further tested. In general, this study systematically analyzed the DcAP2/ERF family of D. catenatum and preliminarily studied the role of DcAP2ERF#96 in the ABA signaling pathway. Further functional verification should be carried out to understand the role of DcAP2/ERF family genes under different conditions.

Phylogenetic Tree Analysis of the DcAP2/ERF Protein
The phylogenetic tree was constructed by MEGA7 using the neighbor-joining method with a bootstrap test (1000 replicates). Multiple sequence alignments of all proteins were completed by MUSCLE before phylogenetic analysis [49]. For the graphic display, the final phylogenetic trees were modified using ITOL (https://itol.embl.de/, accessed on 1 November 2022).

Gene Structure, Conserved Domain, Motif and Promoter Analyses of AP2/ERF Family Members
The gene structure information was directly generated by TBtools according to the gene annotation file. The conserved domain information of AP2/ERF family proteins was calculated by Batch CD-Search (https://www.ncbi.nlm.nih.gov/Structure/bwrpsb/bw rpsb.cgi, accessed on 1 November 2022) [50]. The conserved motif of the full length of AP2/ERF proteins was analyzed using the online MEME website (https://meme-suite.org /meme/tools/meme, accessed on 1 November 2022) [51] by the zero or one occurrence per sequence (zoops) strategy and set the number of searchable motifs to 8. Other parameters are website default settings 4.4. Cis-acting element analyses of DcAP2/ERF family genes The promoter sequences (2000-bp upstream of gene) were obtained from the D. catenatum genome. All cis-acting elements were calculated by PlantCARE (https://bioinformati cs.psb.ugent.be/webtools/plantcare/html/, accessed on 1 November 2022) [52].

Plant Materials and Methods of Stress Treatment
'Honggan ruanjiao' a well-known D. catenatum variety, was used in this study. In order to detect the expression pattern of DcAP2/ERF family under drought conditions, 6-month-old seedlings were treated with 20% PEG6000 to mimic drought conditions. After treatment for 0, 2, 4 and 8 h, respectively, three samples (root, stem and leaf) were harvested and frozen in liquid nitrogen for RNA extraction.
In addition, DcAP2ERF#96 was transformed into Arabidopsis thaliana to obtain overexpression lines. All overexpression lines were selected by hygromycin and the relative expression level of these lines was detected by qRT-PCR ( Figure S6).

RNA Extraction and Quantitative/Real-Time-PCR (qRT-PCR) Analysis
The total RNA of plant materials was extracted using Trizol reagent (Coolaber, Beijing, China). One microgram of total RNA was used to synthesize cDNA with HiScript II Q Select RT SuperMix for qRT-PCR (Vazyme, Nanjing, China). qRT-PCR was performed on a CFX384 real-time system (BIO-RAD, Hercules, CA, USA) with ChamQ Universal SYBR qPCR Master Mix (Vazyme, Nanjing, China). The experiments were repeated three times independently. The actin2 gene was used as an internal control. The primers used in this assay are listed in Table S1.

Expression Pattern Analysis of DcAP2/ERF Genes
To reveal the expression patterns of all AP2/ERF family genes in D. catenatum, the raw RNA-seq data were downloaded from the NCBI Sequence Read Archive (SRA) database. The transcriptome data of root (SRX2938667), stem (SRR4431600), leaf (SRR4431601), white part of root (SRR4431598), Green root tip (SRR4431599), sepal (SRR4431597), labellum (SRR4431602), pollinia (SRR5722145) and gynostemium (SRR4431596) were obtained for tissue expression analyses [29]. The raw RNA-seq data of leaves under 20 • C conditions (SRR3210630, SRR3210635 and SRR3210636) and 0 • C conditions for 20 h (SRR3210613, SRR3210621 and SRR3210626) were downloaded for low temperature stress analyses [53]. By clustering, six groups of transcriptomic data were grouped into two categories according to temperature, indicating that the experimental data were reliable and had significant differences between the control and chilling treatments ( Figure 4A). The raw RNA-seq data of 1mM MeJA (SRR14635790, SRR14635791, SRR14635792) treatment for 4 h and control (SRR14635793, SRR14635796, SRR14635797) were downloaded for hormone treatment analyses [54]. All raw data were converted by the Kallisto method and normalized by transcripts per million (TPM). Differential expression analysis and volcano plot drawing were visualized using Tbtools.

Subcellular Localization of DcAP2ERF#96
The full-length coding sequences of DcAP2ERF#96 were amplified and cloned into the pEarleyGate 101 vector to produce the DcAP2ERF#96-GFP fusion construct. To observe the localization in protoplasts of D. catenatum, young leaves of six-month-old D. catenatum were used for protoplast preparation and transformation. 10 µg plasmid was transformed into 100 µL protoplast suspension, and fluorescence was observed after 12 h [55]. To observe the localization in tobacco, four-week-old N. benthamiana was used for gene transient expression. The activated GV3101 Agrobacterium containing the constructed plasmid was incubated with infection solution (10 mM MES, 10 mM MgCl 2 and 200 µM AS) for 2 h and then injected into the young leaves of tobacco (N. benthamiana). Fluorescence was observed at 488 nm and 560 nm using a confocal microscope after 2 days.

Yeast One-Hybrid (Y1H) and Yeast Two-Hybrid (Y2H) Assays
For the Y1H assay, the full-length CDS of DcAP2ERF#96 was cloned into the pB42AD vector. Three tandem repeats of the motif sequences (DRE, CRT, GCC) were cloned into the pLacZi2µ vector. The combined plasmids were co-transformed into yeast strain EGY48 and selected on the SD/-Ura/-Trp agar medium. Colonies were then plated onto SD/-Ura/-Trp agar medium containing raffinose, galactose and 5-bromo-4-chloro-3-indolyl-b-D-galactopyranoside (X-gal) for blue color development. For the Y2H assay, the full-length CDS and three truncated fragments of DcAP2ERF#96 were cloned into the pGBKT7 vector. The constructed plasmids were transformed into AH109 yeast competent with empty pGADT7 using the Yeastmaker Yeast Transformation System (Clontech, Shanghai, China). Transformation cells were plated on SD/-Trp/-Leu medium to screen positive clones. After 4 days, positive clones were transferred to SD/-Ade/-His/-Leu/-Trp dropout medium to detect self-activation. All relevant primer sequences are listed in Table S1.

Acquisition and Handling of Arabidopsis Transgenic Lines
The full-length CDS of DcAP2ERF#96 was cloned into the pGWB512 vector containing the 35S promoter. Arabidopsis thaliana inflorescences were infected by the floral dip method and the selection of transgenic positive lines by MS medium containing hygromycin. The two lines with the highest expression level were identified and selected by qRT-PCR, and their seeds were placed on MS medium for germination to a root length of 1 cm, then transferred to 1.2% MS medium containing 10 µM ABA, and cultured for 5 days before observation phenotype.

Conclusions
In this study, 120 AP2/ERF genes were identified in D. catenatum for the first time, and their characteristics were further analyzed. These results contribute to our understanding of the classification and evolution trend of AP2/ERF family of orchids. At the same time, according to transcriptome data, we further analyzed the expression pattern of DcAP2/ERF family genes and selected 12 genes of interest from them to test their responses to drought stress. In addition, we preliminarily studied the characteristics of transcription factor DcAP2ERF#96 and found its interacting proteins and promoters according to its inhibition effect on the ABA signaling pathway. However, the specific molecular mechanism of DcAP2ERF#96 in response to abiotic stress in D. catenatum remains to be further studied.