Characterization of the WRKY gene family in Akebia trifoliata and their response to Colletotrichum acutatum

Akebia trifoliata, belonging to the Lardizabalaceae family, is a well-known Chinese traditional medicinal plant, susceptible to many diseases, such as anthracnose and powdery mildew. WRKY is one of the largest plant-specific transcription factor families and plays important roles in plant growth, development and stress response, especially in disease resistance. However, little was known about the numbers, characters, evolutionary relationship and expression of WRKY genes in A. trifoliata in response to plant disease due to lacking of A. trifoliata genome. A total of 42 putative AktWRKY genes were identified based on the full-length transcriptome-sequencing data of A. trifoliata. Then 42 AktWRKY genes were divided into three major groups (Group I-III) based on the WRKY domains. Motif analysis showed members within same group shared a similar motif composition, implying a functional conservation. Tissue-specific expression analysis showed that AktWRKY genes could be detected in all tissues, while few AktWRKY genes were tissue specific. We further evaluated the expression of AktWRKY genes in three varieties in response to Colletotrichum acutatum by qRT-PCR. The expression patterns of AktWRKY genes were similar between C01 and susceptible variety I02, but distinctly different in resistant variety H05. In addition, it showed that more than 64 percentages of AktWRKY genes were differentially expressed during fungal infection in I02 and H05. Furthermore, Gene ontology (GO) analysis showed that AktWRKY genes were categorized into 26 functional groups under cellular components, molecular functions and biological processes, and a predicted protein interaction network was also constructed. Results of bioinformation analysis and expression patterns implied that AktWRKYs might play multiple function in response to biotic stresses. Our study could facilitate to further investigate the function and regulatory mechanism of the WRKY in A. trifoliata during pathogen response.


Introduction
The WRKY transcription factor family is one of the largest transcriptional regulatory gene families in plants, which can regulate downstream transcription through specifically recognizing and binding with the cognate cis-element W-box (TTG ACT /C), which is usually located in the promoter region of genes related to growth, development and stress response [1][2][3].
Structurally, WRKY transcription factors contain one or two conserved WRKY domains, which consist of a signature sequence (WRKYGQK) along with a C 2 H 2 or C 2 HC zinc-binding motif. Generally, WRKY transcription factors are classified into three major groups based on their structure. Group I WRKYs contain two WRKY domains at the N-and C-terminal, each of which is Wen et al. BMC Plant Biology (2022) 22:115 followed by a C 2 H 2 zinc-finger, while Group II WRKYs have only one WRKY domain. Similarly, Group III also contains only one WRKY domain, but there is a C 2 HC zinc-finger motif at the C-terminal of the WRKY domain instead of C 2 H 2 zinc-finger [4][5][6][7]. By inferring the results of a nuclear magnetic resonance solution structure of the C-terminal WRKY domain of AtWRKY4, the conserved WRKY domain is composed of a four-stranded beta-sheet with a zinc binding pocket, which is directly involved in DNA binding [8,9]. Previous studies have shown that WRKY domain sequences can directly bind to W-box (C/T)TGAC(C/T) cis-regulatory element, which were found in the promoter region of the target genes [3,10]. These (C/T)TGAC(C/T) sequence elements contain the invariant TGAC core that mediate transcriptional responses to biotic stresses. Thus, plant genes that contain TGAC core in the promoter regions are considered as defense-associated genes [5,11]. Interactions between the WRKY domain and the W box have been demonstrated by numerous binding experiments, both in vitro and in vivo, and these interactions can also be regulated post-translationally, since these bindings can be inhibited by phosphatase and protein-kinase inhibitors [12,13]. Previous studies have proven that WRKYs play essential roles in plant growth, development, and abiotic or biotic stresses responses [14][15][16]. For instance, AtWRKY12 could partly mediate the effect of GA3 in controlling flowering time and regulating the formation of pith secondary wall in Arabidopsis [17][18][19]. Wintersweet WRKY71 was proven to be involved in the regulation of flowering and leaf senescence in Arabidopsis [20]. Interestingly, WRKY transcription factors, although are transcription factors themselves, are also regulated by other WRKYs [21]. For example, WRKY18 was able to bind directly to different W-boxes in the WRKY53 promoter region, thereby repressing the expression of WRKY53. Thus, WRKY18 could act as a positive senescence regulator due to its repressing function on WRKY53 [22]. Furthermore, WRKYs have been found to be involved in various biotic and abiotic stress defense responses, such as viruses, bacterial pathogens, fungi, heat, drought [23][24][25][26][27][28]. For instance, PlWRKY65, as a disease resistancerelated transcriptional activator, could exert a regulatory effect on JA and SA signals to enhances the resistance of Paeonia lactiflora to Alternaria tenuissima [29]. Additionally, it has been reported that VvWRKY30 overexpression lines had higher antioxidant activities and lower reactive oxygen species contents under salinity stress, thus enhancing their resistance to salt stress [30].
Since WRKY transcription factors play critical roles in plant development and stresses resistance, WRKYs have been identified genome-wide from various plant species with the development of high-throughput sequencing, including 74 WRKYs in Arabidopsis, 103 in rice, 86 in Brachypodium distachyon, 197 in soybean, and 54 in pineapple [5,21,[31][32][33][34]. However, the numbers, characters, evolutionary relationships and expressions of WRKY genes in Akebia trifoliata (Thunb.) Koidz. were completely unknown. A. trifoliata belonging to the Lardizabalaceae family, was mainly distributed in the eastern part of Asia, which was a well-known Chinese traditional medicinal plant, as its antiphlogistic, antineoplastic and diuretic characters [35,36]. The wild resources of A. trifoliata were on the verge of exhaustion because of overexploitation. Due to its medicinal and edible value, A. trifoliata has been developing as an artificial cultivation commercial crop in Hunan and Jiangxi province in China. However, the cultivated A. trifoliata seedlings were susceptible to disease, thus it is very important to protect them from pathogen. In this study, a total of 42 putative Akt-WRKY genes were identified based on the full-length transcriptome-sequencing data of A. trifoliata. Subsequently, the characters of AktWRKYs and their expression patterns in response to Colletotrichum acutatum have been further analyzed. Our results could provide the novel insight into protein structures, evolutionary relationships, and expression pattern of WRKYs in A. trifoliata, which could also facilitate to further investigate the biological functions of AktWRKYs under biotic stresses.

Database search and identification of WRKY transcription factors
All 42 putative WRKY proteins were retrieved from full-length transcriptome sequencing data of A. trifoliata (unpublished data). Arabidopsis WRKY protein sequences were downloaded from the database of The Arabidopsis Information Resource (TAIR, https:// www. arabi dopsis. org/). The AktWRKY proteins were identified by blastp method using Geneious software with Arabidopsis WRKY proteins as query sequences. The identified AktWRKY proteins in A. trifoliata were rechecked and confirmed to avoid repetition and verify the reliability of our results: a) short AktWRKY sequences with an incomplete WRKY domain have been removed, b) all putative non-redundant sequences were assessed with UniProt and SMART (http:// smart. embl-heide lberg. de/) analyses, respectively. The neighbor-joining (NJ) phylogenetic tree based on 42 putative WRKY proteins were constructed to classify the AktWRKYs. All putative WRKY gene family members in A. trifoliata were designed their names base on the homologs in Arabidopsis.

Protein Motifs, Structure Analysis and Phylogenetic Analysis
The conserved motifs in the WRKY proteins were predicted using MEME suite (http:// meme. sdsc. edu/ meme/ cgi-bin/ meme. cgi). The parameters were set as follows: maximum number, 6; site distribution, any number of repetitions; minimum width, 10; and maximum width, 80. The graph was generated by TBtools v1.068. Subsequently, multiple alignment analyses of the WRKY domains sequence were performed by ClustalW (www. ebi. ac. uk/ clust alw/). The phylogenetic tree based on the alignment of WRKY domains in rice, Arabidopsis, and A. trifoliata was constructed using MEGA X software with NJ method and following parameters: p-distance and pairwise deletion. Bootstrap analysis was performed with 1000 replicates.

Plant materials and expression analysis
A. trifoliata, which is not a protected plant, was identified by Prof. Liao Liang according to Flora of China. The original samples (C01, I02, and H05) were obtained from Yuncheng City, Shanxi Province, Lu'an City, Anhui Province, and Xiangyang City, Hubei Province, respectively. The voucher specimens deposited in Jiujiang University (accession number JJU130C01, JJU130I02, and JJU130H05, respectively). The seedlings were planted in Mutong yard in Jiujiang University, Jiujiang, Jiangxi province, China. A. trifoliata tissues of buds, young leaves, mature leaves, stems, female flowers, and male flowers were collected from the Mutong yard. For pathogen infection analysis, A. trifoliata seedlings were sprayed with spores of C. acutatum. After infection for 6 h, the leaves were collected for genes expression analysis. Total RNA was extracted from plant tissues by using Trizol reagent (Invitrogen, USA) according to the manufacturer's protocol. A RNase-free DNaseI (TaKaRa, Japan) was used to remove genomic DNA contamination. Firststrand cDNA was synthesized using PrimeScript ™ 1st Strand cDNA Synthesis Kit (TaKaRa, Japan) according to the manufacturer's protocol. The expression of the all AktWRKY genes was assessed upon the qPCR result analysis. For tissue-specific analysis, the average of total ΔCTvalue (ΔCT. average) was subtracted from all other ΔCT values to obtain second normal standardization, according to the previous method [37]. For phytopathogen infection analysis, the expression level of test genes was calculated with the 2 −ΔΔCT method. The data were statistically analyzed using an OriginPro 7.5 software. Gene specific primers for quantitative real-time PCR are listed in Additional file 4. Public Arabidopsis expression datasets were obtained from the Gene Expression Omnibus (GEO) and the AtGenExpress Consortium (Arabidopsis eFP Browser).

Gene ontology annotation and protein interaction analysis
To perform the functional classification, all putative Akt-WRKY genes were analyzed using Blast2GO basic software. And then, the results were performed by using an online tool WEGO (http:// wego. genom ics. org. cn/) to compare and plot Gene Ontology annotation results. Protein-protein interactions were predicted by the STRING 11.0 program (https:// string-db. org/) based on an Arabidopsis association model with the confidence parameter set at a threshold of 0.35.

Identification of WKRY transcription factors in A. trifoliata
As one of the largest gene families of transcriptional regulators in plant, WRKY transcription factor family plays critical roles in regulating plant growth and development as well as abiotic or biotic stress responses [38][39][40]. Although the functions and evolutionary relationships of WRKYs in several model plants have been investigated, little is known about this family genes in the Chinese traditional medicinal plant, A. trifoliata. To systematically explore the organizational structure, evolutionary relationship and function of WRKYs in A. trifoliata, the full-length transcriptome-sequencing data of this species were applied to identify WRKY genes in silico. A total of 42 AktWRKY genes were identified by searching the transcriptome-sequencing datasets using total Arabidopsis WRKY genes as queries (Table 1 and Additional file 1). All 42 deduced AktWRKY proteins contained at least one highly conserved WRKY domains, while 11 of them had two WRKY domains. As shown in Table 1, the deduced AktWRKY proteins contained amino acid residues between 145 (AktWRKY51) to 747 (AktWRKY34), the range of which was similar to that of other dicotyledons, such as Glycyrrhiza glabra and Santalum album [41,42]. Their molecular weight (MW) varied between 16.77 kDa (AktWRKY51) to 81.41 kDa (AktWRKY25), while the isoelectric point (pI) of 21 AktWRKYs were acidic and the other 21 were basic proteins. A neighborjoining phylogenetic tree was preformed to investigate the phylogenetic classification of the WRKY proteins in A. trifoliata according to previous report [5]. As shown in phylogenetic tree, all 42 deduced AktWRKY proteins were clustered into three main groups, namely, Group I, II, and III (Fig. 1). Seven AktWRKYs containing one WRKY domain and a C 2 CH zinc-binding motif at the C-terminal were classified as Group III, while twentythree main clustering proteins with one WRKY domain and a C 2 H 2 motif were clustered into Group II, which was further divided into five subgroups, named Group IIa-IIe. The twelve remaining proteins were included in Group I.
The number of AktWRKY genes belonging to each subgroup was compared to the number of WRKYs in other plant species, in which the WRKY gene family has been fully identified and characterized, including Arabidopsis, Brachypodium distachyon, rice, kiwifruit (Actinidia chinensis), grape and tomato ( Fig. 2 and Additional file 2) [21,43,44]. Comparing the number of WRKY genes in each subgroup, Group I and Group IIc experienced a significant expansion by having majority members in dicot species. In addition, the number of Group I WRKY genes was similar in A. trifoliata, Arabidopsis, B. distachyon, rice, grape and tomato, but it showed a significant expansion in kiwifruit, which might attribute to the two whole-genome duplications of kiwifruit genome [45]. On the contrary, it was apparent that variations in the number of WRKY genes in Group III were the primary cause of the diversity of WRKY gene family size in monocot species, i. e. B. distachyon and rice [21,31]. These results suggested that numerous duplications and diversifications of Group III WRKY genes might be occurred after the divergence of the monocots and dicots. Previous studies have reported that the Group III of WRKY genes, as a newly defined group, was the most dynamic group with respect to gene family evolution [46]. In this study, the number of WRKY genes in Group III was relatively less in A. trifoliata, suggesting that this group had experienced less gene duplications during the evolutionary course, which could attribute to the fact that A. trifoliata located at the basal clade of the phylogenetic tree.

Conserved motifs of AktWRKYs
To better understand the conservation and diversity of AktWRKYs, the conserved motifs of all the putative AktWRKY proteins were predicted by MEME online program. The distribution of motifs in each group and the diversity of conserved motifs were shown in Fig. 3. Generally, most members within the same group or subgroup shared a similar motif composition, which implied that AktWRKYs homologs located under the same group might have similar functions. However, AktWRKY members of different groups had no common conserved motifs except for the WRKY domain at the C-terminal. The motif 1 and 2 of group I contained conserved heptapeptides WRKYGQK sequence, representing the C-terminal and N-terminal WRKY domains, respectively. Obviously, the C-terminal and N-terminal WRKY domains were differentiated, suggesting that these two WRKY domains might be different in origin or functional differentiation. It was consistent with the fact that the specific binding to W-box was mediated mainly by the C-terminal WRKY domain, whereas N-terminal WRKY domain showed weak binding activity [9]. Further, insight into motif analysis of group II indicated that most of members Most members of group III contained motif 1, which shared a tripeptide HTC residue at the C-terminal of the motif. Overall, these results present group-based motif analysis of AktWRKYs, while the functions of most of these motifs need to be further investigated.

Comparative analysis of WRKY domains
To understand the diversity and evolutionary relationships of the AktWRKY domains, we compared the WRKY domains from the two other sequenced plant genomes (O. sativa and Arabidopsis). An unrooted neighbor-joining comparative phylogenetic tree was constructed by using MEGA software version 7 from the 252 conserved WRKY domains among these three plant species. As shown in Fig. 4, the complete WRKY domains Fig. 2 The distribution of WRKY transcription factors from Arabidopsis, B. distachyon, rice, A. trifoliata, kiwifruit (A. chinensis), grape and tomato. The width of the band represents the percentage of WRKYs in each group. Green represents group I, blue represents group IIa, orange represents group IIb, cyan represents group IIc, yellow represents group IId, purple represents group IIe, and red represents group III were divided into three major groups (I, II and III), and all groups were present in monocots and eudicots. Of the three major groups, group II was the largest major group in the phylogenetic tree, with 118 WRKY domains distributed in five subgroups, including 9 in IIa, 19 (Fig. 5). The highly conserved heptapeptide sequence WRKYGQK was found within a total of 37 AktWRKYs, while five proteins (AktWRKY20, Akt-WRKY30, AktWRKY46, AktWRKY50 and AktWRKY51) were different due to one or two amino acid substitution. The protein AktWRKY50 and AktWRKY51 were found to contain a WRKYGKK sequence, while AktWRKY20, AktWRKY30 and AktWRKY46 contained a WCKYGRK, WMKYGQK and WEKYGQQ sequence, respectively. In addition, CX 4 CX 22-23 HXH and CX 4 CX 23 HXH zinc finger motifs were found in the N-terminal and C-terminal of group I AktWRKYs, respectively. CX 5 CX 23 HXH motifs were found in subgroup IIa, IIb, IId and IIe AktWRKYs, and CX 7 CX 23 HTC motifs were found in group III Akt-WRKYs, while the zinc finger motifs in subgroup IIc were same with C-terminal of group I AktWRKYs.

Tissue-specific expression patterns of AktWRKY genes
Numerous studies have demonstrated that the abundance of transcription factor genes varied greatly in different tissues and at different developmental stages, and played critical roles in regulating plant growth and development. For instance, WRKY12 was specifically expressed in pith and cortex cells of stem and hypocotyls, playing a critical role in pith secondary wall formation, while WRKY13 transcripts were highly abundant in the juvenile phase but decreased over time, implying that WRKY13 was involved in the control of age-mediated pathway [17][18][19]. Thus, tissue-and developmental stagespecific gene expression patterns might provide clues to gene functional divergence during evolution [49]. To investigate the patterns and expression levels of putative AktWRKY genes, the expression levels of 42 A. trifoliata WRKY genes in six tissues (buds, young leaves, mature leaves, stems, female flowers and male flowers) were ClustalW amino acid sequence alignment of typical A. trifoliata WRKY domains. Gaps (dots) have been inserted for optimal alignment. Black and gray shading indicate the presence of identical and conserved amino acid residues, respectively. Consensus amino acid residues are shown below the alignment determined by qRT-PCR (Fig. 6). Our results showed that AktWRKY genes could be detected in all test tissues, and exhibited distinct expression patterns. Generally, most AktWRKY genes were relatively highly abundant in leaves, especially in young leaves, while most Akt-WRKY genes were expressed at low levels in stem and flowers. However, tissue-specific expression of WRKY genes was also observed in A. trifoliata. For instance, AktWRKY21 was particularly highly expressed in stems but less expressed in female flowers. AktWRKY17 exhibited extremely high levels in young leaves, but low in the other tissues, implying this gene might be involved in leaf development. AktWRKY28 showed high levels of expression in buds and female flowers, while AktWRKY19 were highly expressed in young leaves and male flowers. Previous reports demonstrated that WRKY genes with high expression in plant tissues were often found to be able to regulate target genes involved in the relevant processes of plant growth and organs development [50,51]. Thus, tissue-specific AktWRKY genes in this study might provide some useful clues for further investigation of their biological functions in the growth and organs development of A. trifoliata. In addition, some clustered gene pairs showed the same expression pattern, such as AktWRKY03/20, AktWRKY26/33, AktWRKY07/11, AktWRKY27/65, etc., implying that they might be functionally redundant. On the other hand, AktWRKY gene pairs with different expression patterns might executed different biological functions in plant growth and development.

Expression analysis of AktWRKY genes in response to phytopathogen
As crucial components of plant defense signaling networks, WRKY genes were known to exhibit complex response patterns [16,52]. The function of WRKY genes in the regulation of plant response to biotic stresses has been studied in many plant species, especially in Arabidopsis, rice and tomato [7]. It is well known that the functions of AtWRKY genes response to phytopathogens have been well studied. Expression profiles of AtWRKY genes under various phytopathogen stresses indicated that most of AtWRKYs were involved in various biotic stresses response (Additional file 5) [53][54][55][56][57][58][59][60]. For example, AtWRKY3, AtWRKY4, AtWRKY53 and AtWRKY70 played positive roles in disease defense against the necrotrophic fungal pathogen and the biotrophic pathogen Pseudomonas syringae [61][62][63]. AtWRKY33, which was activated by two sigma factor binding proteins, SIB1 and SIB2, was essential for defense toward the necrotrophic fungus Botrytis cinerea [62,64,65], while AtWRKY18 and PtrWRKY18 could activate pathogenesis-related genes, and increase resistance to the biotrophic pathogens [66,67]. Moreover, some WRKY genes, such as AtWRKY7 and AtWRKY48, had direct negative effects on plant defense responses [68,69]. In regards to Akt-WRKY expression in response to biotic stresses, three different varieties of A. trifoliata (C01, susceptible variety I02 and resistant variety H05) were infected with a polyphagous fungal plant pathogen, C. acutatum. Then, qRT-PCR was performed on C. acutatum infected leaves in order to investigate the expression profiles of A. trifoliata in response to C. acutatum (Additional file 3). As shown in Fig. 7, a large number of AktWRKY genes were induced by C. acutatum infection. It has been shown that four WRKY genes (AktWRKY03, 12, 28 and 33) were highly expressed in all three tested A. trifoliata varieties, which were consistent with the previous reports that their homologs (AtWRKY3, BrWRKY12, BnWRKY15 and AtWRKY33) enhanced resistance to phytopathogens through transcriptional activation of defense-related genes [61,65,70,71]. These results suggested that, at least, these four genes might be involved in plant disease resistance. Furthermore, AktWRKY28 showed an expression pattern different from AktWRKY03 and Akt-WRKY33, with high abundance 6 hpi in susceptible variety, but low abundance in the resistant one, although the expression level of which were almost similar in the control samples (Additional file 6). Such divergent behaviors were previously found in rice and cacao [72,73]. For instance, knock-out of OsWRKY28 led to a two-fold increase in resistance to Magnaporthe oryzae, while overexpression of OsWRKY28 resulted in enhanced susceptiblility to M. oryzae [72,74] [63,67]. Although analysis of WRKY expression was helpful in discriminating the role and function of these proteins at the tissue and organism levels, further molecular and biological experiments are needed to investigate their biological function.

Gene ontology annotation and interaction analysis of AktWRKY proteins
Functional annotation of proteins and prediction of protein-protein interaction could help us predict their possible regulatory functions, and provide great support for further investigation of gene families. Based on the similarity of peptide sequences, the Gene Ontology (GO) annotations of one protein could be coordinately transferred to another [75,76]. Here, GO annotations of 42 AktWRKY proteins were analyzed using the Blast2GO basic tool (Fig. 8). Among the protein sequences annotated in the GO database, AktWRKY proteins were categorized into three main categories (i. e. biological processes, molecular functions, and cellular components) and 26 subcategories. In the category of biological process, most AktWRKYs were identified to be involved in the 'regulation of cellular process' (GO:0009987), 'regulation of biological process' (GO:0050789), biological regulation (GO:0065007), metabolic process (GO:0008152), and so on. However, three AktWRKYs (AktWRKY08, 27, and 28) were predicted to be involved in developmental process, including cell differentiation and leaf senescence, while four AktWRKYs (AktWRKY26, 33, 50, and 51) appeared to be involved in defense response to biotic stimulus. The molecular functions of AktWRKYs were associated mostly with 'binding' (GO:0005488) and 'transcription regulator activity' (GO:0140110). The cellular component of this protein family included organelle (GO:0043226), cell part (GO:0044464) and cell (GO:0005623), besides, all AktWRKYs were predicted to be localized in the nucleus (Table 1). Subsequently, to systematically analyze the interaction of AktWRKY proteins, a predicted protein interaction network containing 16 AktWRKY proteins was constructed based on Arabidopsis homologous genes using STRING 11.0 software with the confidence parameter set at a threshold of 0.35 (Fig. 9). Among these proteins, the interaction between AtWRKY33 (AktWRKY33 and 58) and AtWRKY22 (Akt-WRKY74) were related to the Kyoto Encyclopedia of Genes and Genomes (KEGG) signaling pathway of plant MAPK signaling pathway (ath04016) and plant-pathogen Fig. 9 Protein-Protein interaction of AktWRKYs based on AtWRKYs orthologs as predicted by STRING search tool. The thickness of the lines represents the level of interaction between proteins interaction (ath04626) [77][78][79][80][81]. The interaction network among AtWRKY18 (AktWRKY18), AtWRKY33 (Akt-WRKY33 and 58), AtWRKY40 (AktWRKY40), AtWRKY53 (AktWRKY46 and 53), and AtWRKY70 (AktWRKY54 and 70) enriched significantly GO term in defense response, including bacterium (GO: 0042742), fungus (GO:0050832), chitin (GO:0010200) and salicylic acid (GO:0009751) [22,65,67,82]. In addition, these five AtWRKYs and corresponding ortholog AktWRKYs were also involved in a stronger interaction network with other proteins. Overall, the results showed that there were multiple interactions among AktWRKY proteins, implying that AktWRKY proteins were involved in multiple stress responses.

Conclusion
In conclusion, a total of 42 WRKY genes were identified based on transcriptome sequences, which was the first study on the organizational structure and abundance of WRKY in A. trifoliata. The classification, protein structure, conserved motif composition, and phylogenetic relationship of AktWRKYs were systematically analyzed and compared, which provide a basis for further study of the molecular and functional structure of AktWRKYs. Furthermore, based on qPCR results, the expression patterns of tissues-specific and phytopathogen-responsive AktWRKYs were obtained, providing useful information for further investigation of the function of AktWRKYs in response to biotic stresses. Our study could help researchers better understand the function and regulatory mechanism of WRKYs in A. trifoliata during pathogen response.