Introduction

Plant stress responses are generally controlled by a network of specialized genes through intricate regulation by specific transcription factors (TFs). The WRKY TF family, an important member of the stress-related transcription factor families, is involved in the regulation of plant developmental processes and biotic and abiotic stress responses (Eulgem et al. 2000; Rushton et al. 2010). WRKY TF is named after the WRKY domain of ~60 amino acids that contains the highly conserved amino acid sequence WRKYGQK followed by a zinc-finger-like motif C2H2 or C2HC. Additionally, it can recognize and bind TTGAC(C/T) W-box cis-elements (Rushton et al. 2010). WRKY proteins can be divided into three groups (i.e., I, II, and III) on the basis of the number of WRKY domains and the type of the zinc-finger motif (Eulgem et al. 2000). Group I proteins typically have two WRKY domains, including a C2H2 motif. Group II proteins have only one WRKY domain and a C2H2 zinc-finger motif and can be further divided into five subgroups (i.e., IIa–e). Group III proteins contain a single WRKY domain, and their zinc-finger-like motif is C2HC.

The WRKY TF family is a large conserved family of TFs in plants (Eulgem et al. 2000; Agarwal et al. 2011; Chen et al. 2012; Yamasaki et al. 2013). Since the first WRKY gene, SPF1, was isolated from sweet potato (Ipomoea batatas) (Ishiguro and Nakamura 1994), a large number of WRKY genes have been identified from several higher plants and have been found to mediate various physiological processes such as plant growth and development (Zhang et al. 2011; Ulker et al. 2007; Ay et al. 2009), stress responses (Scarpeci et al. 2009; Pandey et al. 2010; Chen et al. 2010), and signal transductions (Antoni et al. 2011; Shen et al. 2008; Rushton et al. 2012). WRKY proteins are known to play a crucial role in plant defenses against various biotic stresses, including bacterial (Kim et al. 2006), fungal (Zheng et al. 2006), and viral pathogens (Chen et al. 2013; Park et al. 2006). WRKY proteins have also been implicated in the regulation of developmental processes such as trichome formation (Johnson et al. 2002), seed development (Jiang and Yu 2009), and leaf senescence (Miao and Zentgraf 2010; Besseau et al. 2012). In addition, WRKY proteins are involved in responses to various abiotic stresses (Wang et al. 2012a; Chen et al. 2012) and some signal transductions mediated by plant hormones such as abscisic acid (Antoni et al. 2011; Chen et al. 2010) and salicylic acid (van Verk et al. 2011). In Arabidopsis thaliana and rice (Berri et al. 2009), microarray analyses have revealed that many WRKY transcripts are strongly induced in response to various abiotic stresses, such as salinity, drought, and cold.

The Chinese cabbage (Brassica rapa ssp. pekinensis) is an important Brassica crop and has nutritional value for humans. It also provides a model for studying gene fractionation because its genome triplication event has been described (Tang and Lyons 2012). Recent completion of the Chinese cabbage chiifu genome sequencing allows the genome-wide identification of specific gene families (Wang et al. 2011). The WRKY transcription factor family has been extensively investigated in many sequenced plants (Ling et al. 2011; He et al. 2012; Huang et al. 2012; Li et al. 2012; Tripathi et al. 2012; Xiong et al. 2013; Wei et al. 2012). However, analyses of gene retention and fractionation, evolutionary divergence, and genome-wide expression patterns of this gene family have not yet been studied.

In this study, we performed a genome-level analysis of WRKY TFs, including genome locations, subgenome distributions, and evolutionary divergence, in the triplicated Chinese cabbage genome. Additionally, Illumina RNA-Seq data showed the tissue-specific and common expressions of BrWRKYs in different tissues. In addition, subsequent quantitative real-time polymerase chain reaction (qRT-PCR) analysis, a precise mechanism, indicated the distinct and coresponsive patterns of BrWRKYs in response to abiotic and biotic stresses in leaves. This study facilitated the identification of tissue-specific and stress-related WRKY candidates and provided some information of the WRKY family members in response to multiple stresses. Taken together, these findings might serve to improve the environmental resistance of the Chinese cabbage through the manipulation of the BrWRKYs.

Materials and Methods

Characterization of WRKY Gene Family in Chinese Cabbage

To identify the WRKY family members in Chinese cabbage, we first retrieved the whole Chinese cabbage genome sequences including 41,019 protein-coding genes and 283.8 Mb of the genome assembly (Version 1.5) from the Brassica genome database (BRAD, http://brassicadb.org/brad/) using local BLASTP (Camacho et al. 2009) (blast2.2.27+, e value 1e−10, -max_target_seqs 2) against 74 A. thaliana WRKY protein sequences from TAIR10 (http://arabidopsis.org/index.jsp). A total of 119 gene-coding proteins were collected and then aligned using MUSCLE (Edgar 2004) to build a Brassica-specific hidden Markov model (HMM) profile using hmmbuild in HMMER v3.0 (http://hmmer.janelia.org/). The complete WRKY family in Chinese cabbage was identified using hmmsearch program (e value, 1e−3) by searching the full protein-coding gene sequences of the Chinese cabbage genome based on the Brassica-specific HMM profile. The candidate sequences were further confirmed using SMART (Letunic et al. 2012) and Pfam (Finn et al. 2010) programs, and the redundant sequences were manually removed. A total of 145 WRKY protein-encoding genes in Chinese cabbage were obtained and named as BrWRKY followed by Arabic numbers 1–145 according to the position of their corresponding genes on chromosomes 1–10 and from top to bottom as well as the orders of the scaffolds (five genes were separately anchored onto the five scaffolds Scaffold000096, Scaffold000203, Scaffold000217, Scaffold000316, and Scaffold000396; Supplementary Table 1). In addition, each of Chinese cabbage WRKY orthologs in A. thaliana was identified using BLASTP (e value, 1e−20) against A. thaliana WRKY protein sequences and the best hits were collected (Supplementary Table 1).

Identification of Protein Properties and Conserved Motif Analysis

To investigate the protein properties of the putative BrWRKY proteins, the molecular weight (MW) and isoelectric point (pI) were calculated using the Pepstats (http://www.ebi.ac.uk/Tools/seqstats/emboss_pepstats), and the protein nuclear localization scores were predicted by nucpred-predictors-1.0 (Brameier et al. 2007). To identify the distributions of the conserved motifs within the BrWRKY proteins, the conserved motifs were detected by MEME Suite (Bailey et al. 2009) based on the BrWRKY protein sequences that were used in phylogenetic analysis with default settings except that the maximum number of motifs to find was defined as 10 and the maximum width was set to 100.

Multiple Sequence Alignment and Phylogenetic Analysis

Multiple sequence alignments of WRKY proteins and domains were separately carried out using MUSCLE (Edgar 2004), and the alignment results were shown and manually modified using Jalview (Clamp et al. 2004). To study the phylogenetic relationships of BrWRKY proteins along with their orthologs in Arabidopsis and rice, the full-length WRKY protein sequences were retrieved from RGAP7 (http://rice.plantbiology.msu.edu/). Phylogenetic analysis was performed using MEGA 5 (Tamura et al. 2011) with the maximum-likelihood (ML) method and bootstrap value was set to 1,000.

Genomic Organization and Syntenic Analysis of BrWRKYs

To visualize BrWRKY loci on Chinese cabbage chromosomes and subgenomes, the data concerning gene position, gene length, chromosome size, and centromere position were obtained from BRAD, and the information of subgenome fractionation of Chinese cabbage genome was extracted from a diploid ancestral genome of B. rapa analysis (Cheng et al. 2013). We drew a distribution map of BrWRKYs on 10 chromosomes and 3 subgenomes based on above data using custom-designed Perl scripts.

For syntenic analysis of BrWRKY genes, we used MCScanX (Wang et al. 2012b) to detect syntenic gene pairs in Chinese cabbage genome. All protein sequences of Chinese cabbage were compared against themselves using BLASTP, with tabular output format (-m 8) and e value of <1e−20. The BLASTP tabular file and a simplified Chinese cabbage gene location file (that contained chromosome, gene symbol, start, and end) served as an input for MCScanX to identify syntenic gene pairs and duplication types with default settings.

Ka/Ks Calculation and Dating the Duplication Events

To estimate the divergence of duplicated BrWRKY genes, protein sequences of the duplicated pairs of BrWRKYs were aligned using MUSCLE (Edgar 2004). Subsequently, protein alignments were translated into their corresponding coding sequence alignments using an in-house Perl script. We calculated the synonymous rate (Ks), nonsynonymous rate (Ka), and evolutionary constraint (Ka/Ks) between the duplicated pairs of BrWRKYs (Table 2) based on their coding sequence alignments, using the method of Nei and Gojobori as implemented in KaKs_calculator (Zhang et al. 2006). Ks values of all duplicated BrWRKY genes were separately used as proxies for the divergence events, and the divergence time was calculated according to the neutral substitution rate of 1.5 × 10−8 substitutions per site per year for chalcone synthase (Koch et al. 2000).

RNA-Seq Data Analysis

We collected the Illumina RNA-Seq data from BRAD that were previously generated by Cheng et al. (2012). The RNA-Seq transcriptomic data of three tissues (i.e., roots, stems, and leaves of Chiifu-401/42 seedlings were harvested from five-leafed plants growing in greenhouse) were generated using the Illumina HiSeqTM2000 platform. Gene expression levels were calculated as reads per kilobase of exon model per million mapped reads (RPKM) units (Mortazavi et al. 2008). The RPKM values of Chinese cabbage WRKY genes are shown in Supplementary Table 2. Heat maps for only those genes were generated, which have RPKM values in at least one of the three tissues (i.e., roots, stems, and leaves). Heat maps were generated and hierarchical clustering was done using Cluster 3.0 (de Hoon et al. 2004).

Plant materials, Growth Conditions, and Stress Treatments

Chinese cabbage (B. rapa spp. pekinensis) cultivar YANZA03 was used for all experiments. Seedlings were germinated in plastic Petri dishes containing a filter paper saturated with distilled water in darkness at 22 °C for 2 days. Seedlings were transferred to the pots containing a soil growth medium under an artificial growth conditions set at 22 °C and approximately 120 μmol photons m−2 s−1, photoperiod of 16/8 h, and relative humidity of 60 %. The 1/2 Murashige and Skoog liquid solution (pH 5.8, without agar and sugar) was irrigated once every 3 days. Four-week-old (5-leafed) plants were preformed in different abiotic stress treatments and a biotic stress treatment under continued time course. For abiotic stresses including salt and osmotic treatments, pots were irrigated with 250 mM NaCl and 15 % (w/v) polyethylene glycol (PEG), respectively, and kept standing in the irrigation solution for 30 min under normal growth conditions. For cold treatment, pots were exposed to low-temperature (4 °C) conditions. For a biotic-stress-treated Chinese cabbage, plants were infected using downy mildew (Peronospora parasitica) according to the method described previously (McDowell et al. 2000). Inoculations were performed by spraying the seedlings with a spore suspension (100,000 spores mL−1 in ddH2O) through collecting conidia from infected leaves of susceptible plants. The plants were exposed to the normal artificial growth conditions as controls. The seedlings were harvested under a continuous time course (0, 1, 6, 12, 24, 48, and 72 h) in three biological replicates for RNA preparation.

RNA Isolation and qRT-PCR

Total RNA was isolated from treated leaves using TRIzol (Invitrogen, San Diego, CA) according to the manufacturer's instructions. Total RNA was treated with the DNase I (Invitrogen) and 1 μg treated RNA was reverse-transcribed using PrimeScript™ RT Reagent Kit (Perfect Real Time) for RT-PCR (Takara, Dalian, China). cDNA was then diluted 1:20 with ddH2O as template for qRT-PCR. The SYBR® select Master Mix (Invitrogen) was used to detect gene expression according to the manufacturer's recommendations on the 7500 Fast Real-Time PCR System (Applied Biosystem). qRT-PCR was carried out in 96-well optical reaction plates, and PCR reaction was performed in a total volume of 20 μL containing 0.5 μM of each primer (1 μL), 20 ng/μL cDNA (1 μL), and 1× Power SYBR® Select Master Mix (10 μL) and ddH2O was added up to 20 μL. Gene-specific primers that were used to detect transcripts are listed in Supplementary Table 3. The Chinese cabbage glyceraldehydes-3-phosphate dehydrogenase gene (GAPDH, Bra026904) was used as an internal control (Qi et al. 2010). The PCR conditions and relative gene expression calculation were as previously described (Li et al. 2013).

Results and Discussion

Identification of WRKY Orthologs and Genomic Location in Chinese Cabbage

To identify all possible members of the WRKY family in Chinese cabbage, we first used Arabidopsis WRKY proteins to search the Chinese cabbage genome and then built a Brassica-specific HMM profile on the basis of the results of the searches to identify all the members of the WRKY family in Chinese cabbage (see “Materials and Methods”). The candidates were then subjected to Pfam and SMART databases to confirm the presence of the WRKY domains. In Chinese cabbage, there were 145 protein-coding genes, which contained the complete WRKY domains and open reading frames (ORFs) ranging from 432 to 3249 bp (Supplementary Table 1). Thus, the WRKY TF family in Chinese cabbage also consisted of >100 members, as has been reported for rice (Ross et al. 2007) and maize (Wei et al. 2012). For the better identification and classification of the WRKY TFs in B. rapa, a uniform nomenclature has been assigned to the WRKY genes in B. rapa. The 145 WRKY genes from Chinese cabbage are uniformly designated as BrWRKY followed by Arabic numbers 1–145 according to the position of their corresponding genes on chromosomes 1–10, from top to bottom (Fig. 1), and the order of scaffolds (Supplementary Table 1). Subsequently, the predicted gene models of the BrWRKYs were analyzed on the basis of sequence characteristics, multiple sequence alignments, and subcellular localization predictions. All of the predicted, putative WRKY proteins were divided into three groups (i.e., I, II, and III) and were compared to the published classification system of the WRKY family in flowering plants, where three groups of WRKY TFs were defined according to Rushton et al. (2010). The isoelectric points, molecular weights of the 145 members, and nuclear localization scores >0.5 (NucPred score) of most BrWRKY TFs (128/145) are shown in Supplementary Table 1, and 31 BrWRKY proteins were detected such that each of them contained two WRKY domains (Supplementary Table 1).

Fig. 1
figure 1

The distribution of Chinese cabbage WRKY genes identified in this study from 10 chromosomes and 3 subgenomes. The chromosome number is indicated on the top of each chromosome. The subgenome is shown on the right of each chromosome, and subgenome types are indicated at the bottom. The different color blocks represent different subgenomes. BrWRKY genes are showed on the left of each chromosome. The positive (+) and negative (−) signs following each gene represent forward and reverse orientation of the respective gene, respectively, on the chromosome or subgenome. The black ring represents centromere of each chromosome, and the dotted line indicates tandemly duplicated gene pairs. Gene position and each chromosome size can be estimated using the scale on the left of the figure

Full size image

The BrWRKYs were mapped on 10 chromosomes and 3 subgenomes in Chinese cabbage (Fig. 1). In total, 145 BrWRKYs were separately mapped onto the chromosomes from A01–A10, except for the five members on the scaffolds (see “Materials and Methods”). Most of the WRKY genes (83/140) were concentrated on chromosomes 2, 3, 4, and 9 (Fig. 1). Chromosome A03 contained the highest number of BrWRKYs (i.e., 26/140 members or ~18 %), whereas chromosome A10 only contained four BrWRKYs (i.e., <3 %; Fig. 1). The almost complete triplication of the B. rapa genome has been reported relative to A. thaliana, and the ortholog divergence event is known to occur between five and nine million years ago (MYA) (Wang et al. 2011). They proposed the hypothesis that the three differentially fractionated subgenomes were consistent with the least fractionated (LF), medium fractionated (MF1), and most fractionated (MF2) of the subgenomes. Thus, 140 BrWRKYs that could be mapped onto the chromosomes were fractionated into three subgenomes (i.e., LF, MF1, and MF2), including 50 in LF, 52 in MF1, and 38 in MF2 (Fig. 1 and Supplementary Table 1). Therefore, the 145 putative WRKY genes, ranging from 432 to 3249 (BrWRKY54 and BrWRKY140, respectively) ORF lengths, were divided into three groups (i.e., I, II, and III; Supplementary Table 1), and they shared typical characters that contained one to two WRKY domains previously reported for Arabidopsis, rice, poplar, and maize. Genes in different genomes that evolved from a common ancestral gene by speciation are considered orthologs, which are known to retain similar functions during evolution (Ross et al. 2007). WRKY orthologs across B. rapa and A. thaliana were detected using BLASTP (e value 1e−20). Most members of the WRKY orthologs in Arabidopsis presented three copies that were linked to their corresponding BrWRKYs (Supplementary Table 1).

Collinearity and Evolutionary Divergence of Duplicated BrWRKY Genes

To investigate the collinearity of the WRKY gene family in B. rapa, we first searched all gene models (41019) of B. rapa by using BLASTP (e value of <1e−20, tabular output), and the resulting BLAST hits and gene location files of B. rapa were compiled as the input for MCScanX to classify duplication types under the default settings. We found 23,284 collinear gene pairs (56.76 %; Supplementary Table 4) and 2,278 tandem duplicated pairs (Supplementary Table 5) in the Chinese cabbage genome. The collinear relationships of the duplicated pairs in the Chinese cabbage WRKY gene family are shown in Fig. 2. Among BrWRKY genes, 144 were found to be segmentally duplicated and are located on duplicated segments on 10 chromosomes and 3 subgenomes in Chinese cabbage (Fig. 2 and Table 1). A minimum of six BrWRKYs are located in duplicated segments on each chromosome (Fig. 2) and two BrWRKYs on chromosome 8 presented tandem duplication (Fig. 1 and Table 1). Interestingly, all the BrWRKY genes containing subgenome LF had one or more duplicated BrWRKY genes in other subgenomes (i.e., MF1 and MF2), suggesting that all the BrWRKY genes had been retained in Chinese cabbage after genome triplications, which might have also contributed to the expansion of the BrWRKY gene family.

Fig. 2
figure 2

Depiction of duplicated BrWRKY genes on 10 Chinese cabbage chromosomes. Gray lines indicate collinear blocks in whole Chinese cabbage genome, and black lines indicate duplicated BrWRKY gene pairs

Full size image
Table 1 Ka/Ks calculation and divergent time of the duplicated BrWRKY gene pairs
Full size table

Further, the selection types and divergence timings of duplicated genes were estimated by calculating the synonymous (Ks) and nonsynonymous substitutions (Ka) per site between duplicated pairs. The Ks, Ka, and Ka/Ks of 73 duplication pairs that occurred in the Chinese cabbage WRKY gene family were calculated (Table 1). Ka/Ks = 1 indicates neutral selection, Ka/Ks < 1 indicates purifying selection, and Ka/Ks > 1 indicates accelerated evolution with positive selection. The results indicated a total of 72 segmental duplication pairs and 1 tandem pair (BrWRKY105 vs. BrWRKY106), suggesting that most of the duplicated BrWRKY gene pairs (68/73) have Ka/Ks ratios <1, representing purifying selection on these duplicated pairs (Table 1). The tandem pair showed a Ka/Ks ratio slightly equal to 0.22, suggesting that purifying selection was also present in this duplicated pair. Thus, the gene pair (BrWRKY54 vs. BrWRKY35) was the only one (out of all WRKY duplicated pairs) in Chinese cabbage that showed a Ka/Ks ratio slightly equal to 1.0, indicating neutral selection. For the other four duplicated pairs (i.e., BrWRKY9 vs. BrWRKY35, BrWRKY54 vs. BrWRKY9, BrWRKY105 vs. BrWRKY86, and BrWRKY105 vs. BrWRKY106), the Ka/Ks ratios were >1, suggesting positive selection after triplication of the genome.

According to previous reports, the evolutionary timescale in Brassicaceae is estimated on the basis of a synonymous substitution rate (Koch et al. 2000). The divergence dates of the duplicated BrWRKY genes were also calculated (Table 1). The results indicated that most of the duplicated BrWRKY gene pairs (68/72) underwent purifying selection, and the divergence time spanned 0.46~6.66 million years, which indicates that the duplicated divergence of the WRKY family members in B. rapa occurred after the triplication events (i.e., 5~9 MYA) (Wang et al. 2011; Cheng et al. 2012). Moreover, all of the WRKY duplicated pairs and the three subgenome fractionations of the WRKY gene family in B. rapa were clearly identified using gene collinearity analysis, and B. rapa was confirmed to have undergone three rounds of whole-genome duplication, as suggested by the recent whole-genome triplication events.

Identification of Conserved Motifs and Phylogenetic Analysis of the BrWRKY Gene Family

The phylogenetic relationships and structural features of the WRKY proteins in Chinese cabbage were identified by constructing an unrooted ML phylogenetic tree of the 145 putative BrWRKY proteins by using the MEGA 5 software (using the ML method and bootstrap at 1,000), and the conserved motifs were detected using the MEME program (Fig. 3, Table 2). In general, the 145 BrWRKYs were clustered into seven large subgroups (i.e., I, IIa–e, and III) that contained different numbers of WRKY members, which shared similar motif compositions, indicating functional similarities among members of the same subgroup (Fig. 3). According to a prior classification system, 31 BrWRKYs belonged to group I, and most contained two WRKY domains. A total of 89 BrWRKYs were classified as WRKY group II members, and each had a single WRKY domain containing a C2H2 zinc finger and 25 BrWRKYs had C2HC zinc fingers and belonged to group III (Supplementary Fig. 1). Interestingly, after the distributions of the different motifs from the different subgroups were assessed, most of the conserved motifs were continually detected lying around the WRKY domain, indicating that these motifs may be essential for the function of the WRKY proteins, such as motifs, 1, 2, 3, and 7 (Fig. 3). Of the 10 motifs, motifs 1, 2, 3, and 7 were localized in the WRKY-box domain, which partly represented the distribution of the C- or N-terminal WRKY domains (Supplementary Table 3). In addition to motifs 1, 2, and 6, some subgroups of IIb and III had their own typical motifs (Fig. 3). Of the seven groups of BrWRKY proteins, group I was separated by groups IIc and III, and the division of the remaining groups was rather rigorous. Most of the 31 members (27/31) from group I, which contained two WRKY domains, and all the members of group I shared conserved motif 1 (Fig. 3). These results indicated the evolutionary tendency of the WRKY genes in Chinese cabbage to originate from group I. For example, BrWRKY84, BrWRKY86, BrWRKY105, and BrWRKY106 from group I, which contained only one WRKY domain, were further clustered into group IIc on the phylogenetic tree of the full-length proteins, whereas, on the phylogenetic tree of the WRKY domain, they were completely placed in group ICT. Group II contained five subgroups (IIa–e) that were classified on the basis of zinc-finger types. It consisted of 89 WRKY members that contained several common and specific motifs, such as motifs 3 and 5 that were shared in subgroups IIa–e, while motifs 1 and 6 were continually detected in group IIc (Fig. 3). In addition, alignment of all WRKY domains from Chinese cabbage showed that, of the 145 putative BrWRKY proteins, 142 contained the WRKYGQK sequence and 3 possessed naturally altered WRKY motifs, including WRFD (BrWRKY14), WMKY (BrWRKY24), and WRK (BrWRKY119, Supplementary Fig. 1). In addition, the variations of the zinc fingers in the WRKY domains also emerged in some BrWRKY proteins, such as BrWRKY19, BrWRKY24, and BrWRKY73 (Supplementary Fig. 1). These results suggested that the expansion of the WRKY genes in Chinese cabbage occurred through changes in the number of and structural variations in the WRKY domains and led to the derivation of novel members from the ancient form group I (Supplementary Fig. 1) and further supported group I WRKY genes as the ancestral form in the evolutionary history of the WRKY gene family in plants (Zhang and Wang 2005).

Fig. 3
figure 3

Phylogenetic relationship and conserved motif compositions of BrWRKY proteins. The left panel shows the multiple sequence alignment of 145 full-length BrWRKY proteins, which was done using MUSCLE, and the phylogenetic tree was constructed using MEGA 5 by the maximum-likelihood method with 1,000 bootstrap replicates. The tree was divided into seven phylogenetic subgroups designated as I, IIa–e, and III marked with different color backgrounds. The right panel is a schematic representation of the conserved motifs in the BrWRKY proteins as detected by MEME analysis. Gray lines represent the nonconserved sequences, and each motif is represented by a color box numbered at the bottom. The length of protein can be estimated using the scale at the bottom

Full size image
Table 2 Conserved motif compositions of aligned WRKY proteins in Chinese cabbage
Full size table

Over the past few decades, the original phylogenetic classification of the WRKY TFs by Eulgem et al. (2000) has proven to be robust. In order to better understand the phylogenetic relationship of the WRKY gene family in Chinese cabbage, the WRKY genes from two other sequenced plant genomes were selected for comparative analyses, including the rice (model plant for C3 monocot) and Arabidopsis (model plant for dicot) genomes. An unrooted phylogenetic tree from alignments of the conserved WRKY domain sequences among Arabidopsis, Chinese cabbage, and rice was constructed using the ML method (Fig. 4). It showed that the complete WRKY domains were divided into typical eight subgroups previously found in flowering plants, namely groups INT, ICT, IIa–e, and III, and the ancestral groups INT or ICT were separately located at the start and end of the phylogenetic tree (Fig. 4 and Supplementary Fig. 2). These results indicated that the WRKY genes originated from a common ancestor and that the WRKY domains INT and ICT of group I, as common ancestor nodes, evolved into other members through structural variations and changes in the number of the WRKY domains (Fig. 4 and Supplementary Fig. 1). Unlike Chinese cabbage, rice and Arabidopsis have less WRKY proteins in group IIc; the WRKY domains from Chinese cabbage and Arabidopsis were clustered into the corresponding subgroups, thus showing a similar topological tree (Supplementary Fig. 2a, b). Additionally, in the WRKY domains from group III, the C2HC zinc-finger motif replaced other WRKY family proteins with a C2H2 zinc-finger motif (Supplementary Fig. 1). The patterns of the WRKY domains and zinc-finger motifs of the BrWRKY proteins were similar to those reported in previous studies (Rushton et al. 2010) on groups II (C-X4-5-C-X22-23-H-X1-H) and III (C-X7-C-X23-H-X1-C). Interestingly, the classification system of the WRKY family in the Chinese cabbage more closely resembles that of Arabidopsis, suggesting a close relative with similar functional organization of the gene family. Taken together, these results suggest that BrWRKYs shared the same origin with the AtWRKYs and that they may be expanded rapidly after speciation from A. thaliana by genome triplication.

Fig. 4
figure 4

Phylogenetic tree of WRKY proteins of Chinese cabbage, Arabidopsis, and rice. Multiple sequence alignment of WRKY domains was done using MUSCLE, and the phylogenetic tree was constructed using MEGA 5 by the maximum-likelihood method with 1,000 bootstrap replicates. The tree was divided into eight phylogenetic subgroups, designated as INT, ICT, IIa–e, and III. Members of Chinese cabbage, Arabidopsis, and rice were denoted by triangles, circles, and diamonds, respectively

Full size image

Expression Profiling of BrWRKY Genes in Roots, Stems, and Leaves

To identify a tissue-specific expression profile of BrWRKY genes, we utilized Illumina RNA-Seq transcriptomic data from BRAD that were previously generated by Cheng et al. (2012). Tissues (i.e., roots, stems, and leaves) were collected from 4-week-old chiifu-401/42 plants grown in a greenhouse, and paired-end reads were generated using the Illumina HiSeqTM2000 platform. The reads that were mapped completely in exons were represented in the expression level measurement, and the gene expression levels were calculated in the units RPKM. The Chinese cabbage RNA-Seq data provided the expression of >29,000 Chinese cabbage genes in the three types of tissue.

The transcript levels (RPKM values) of 126 of 145 BrWRKYs were obtained from at least one of the three tissues, while the rest of the 19 BrWRKYs either had no expression or had spatial and temporal expression patterns not found in the RNA-Seq libraries. RPKM files for BrWRKY gene expression are shown in Supplementary Table 4. Of the 126 BrWRKYs, >100 (>79 %) were ubiquitously expressed at relatively lower levels in all tissues tested. For instance, the expression of 48 BrWRKYs (~38 %) in the roots, 116 (~92 %) in the stems, and 42 (~33 %) in the leaves was downregulated (Fig. 5). In contrast, the expression of 56 WRKY genes was upregulated only in the roots, 10 in the stems, and 42 (~33 %) in the leaves, including 6 members (BrWRKY14, BrWRKY27, BrWRKY77, BrWRKY67, BrWRKY85, and BrWRKY95) that were highly expressed (Fig. 5). Simultaneously, the 20 BrWRKYs were highly expressed in both the roots and leaves, and 4 BrWRKYs shared high expression in the leaves and stems. Some of the BrWRKYs also exhibited tissue-specific expression. For example, BrWRKY16, BrWRKY72, and BrWRKY114 were expressed only in the root tissues, and BrWRKY71 and BrWRKY120 were specifically expressed in leaf tissues (Fig. 5). These results indicated that various BrWRKYs might be involved in the diversification of the morphological features similar to their Arabidopsis orthologs. For example, WRKY54 (ortholog of BrWRKY72) and WRKY70 (ortholog of BrWRKY131) cooperated as negative regulators of leaf senescence in A. thaliana (Besseau et al. 2012), while the expression profiling is antagonistic to their orthologs in the three tissues of Chinese cabbage (Fig. 5). Previous studies (Ay et al. 2009; Zhang et al. 2011) have shown that the WRKY genes highly expressed in plant organs often play key roles in plant development through transcription activation of the downstream target genes involved in the processes of plant growth and development. Therefore, we speculated that the tissue-specific BrWRKYs reported herein might play integrative roles in transcriptional regulation by controlling the transcription of a broad set of genes to affect the development of Chinese cabbage.

Fig. 5
figure 5

Heat map representation and hierarchical clustering of BrWRKY genes across root, stem, and leaf. The Illumina RNA-Seq data were reanalyzed, and the RPKM values were log2-transformed and heat map generated using Cluster 3.0 software. Bar on the left represents log2-transformed values, ranging from −1.36 to 1.36, represent low-to-high expression. NA represents no available data

Full size image

Expression Patterns of BrWRKY Genes Under Abiotic and Biotic Stresses

The WRKY TFs are known to play a crucial role in plant responses to abiotic or biotic stresses. All BrWRKYs from the above expression profiling that were significantly expressed in the leaves, stems, and roots were used for qRT-PCR analysis in response to cold, salinity, and osmosis treatments, and the seedlings were infected with downy mildew. The results showed that 28 BrWRKYs were differentially expressed in response to at least one stressor, as well as several BrWRKYs were simultaneously expressed to mediate abiotic or/and biotic stresses (Fig. 6). Most of the BrWRKYs (26/28) were preferentially or specifically expressed at high levels, except for BrWRKY14 and BrWRKY103 in a continuous time course (i.e., 0, 1, 6, 12, 24, 48, and 72 h) under different abiotic stresses (Fig. 6a, b, and d). The expression levels of BrWRKY70 (ortholog of AtWRKY33) and 141 (ortholog of AtWRKY6) were significantly upgraded by the cold treatment (Fig. 6a). BrWRKY109 (ortholog of AtWRKY11) was rapidly expressed and showed peak expression at the 6-h time point under the salt treatment (Fig. 6b). Additionally, BrWRKY59, BrWRKY79, and BrWRKY109 had high expression level, in the continuous time course under the osmotic stress treatment (Fig. 6c).

Fig. 6
figure 6

Expression patterns of BrWRKY genes under biotic and abiotic stress. Three-week-old (four-leafed) Chinese cabbage plants were preformed in various treatments, including cold (a), salinity (b), osmosis (c), and biotic stress (d), under continued time course (0, 1, 6, 12, 24, 48, and 72 h). qRT-PCR data were normalized using BrGAPDH. The relative expression levels of BrWRKYs in leaves were quantified against BrGAPDH transcript levels as 2−ΔΔCT

Full size image

Biotic stress treatment induced the expression of 22 BrWRKY genes in response to downy mildew (P. parasitica) infection by qRT-PCR analysis. Moreover, the expression patterns of biotic-related BrWRKYs were generally similar to those obtained in response to abiotic stresses (Fig. 6). Among the 22 biotic-related BrWRKYs, 18 were upregulated and 6 (i.e., BrWRKY4, 53, 65, 109, 133, and 141) were significantly upregulated (Fig. 6d). In contrast, BrWRKY18, BrWRKY83, BrWRKY103, BrWRKY104, and BrWRKY124 were significantly downregulated during the continuous time course in response to biotic stress (Fig. 6d). The qRT-PCR data also indicated that several BrWRKYs play important roles in biotic and abiotic stress responses in plants. For example, BrWRKY53, BrWRKY59, BrWRKY70, BrWRKY77, BrWRKY79, BrWRKY109, BrWRKY133, and BrWRKY141 were expressed at relatively high levels with a cross-responsive approach to cold, salinity, and osmosis treatments. Additionally, BrWRKY53, BrWRKY79, BrWRKY109, BrWRKY133, and BrWRKY141 were simultaneously expressed at higher levels in response to abiotic or biotic stresses. These results further support that WRKY genes can co-regulate abiotic and biotic stresses by the manual interactions of WRKY TFs (Chi et al. 2013).

Conclusion

A total of 145 WRKY genes were identified in Chinese cabbage, and all of these genes were anchored on 10 chromosomes and 3 subgenomes. WRKY paralogs were detected by syntenic analysis, and the distributions and divergence of the duplicated gene pairs in BrWRKYs supported the recent triplicated copies of the Chinese cabbage genome. Phylogenetic analysis of the WRKY family among Arabidopsis, Chinese cabbage, and rice showed that the conserved organization of the WRKY family implied that these genes underwent ancient and recent gene duplication events from a common origin and were retained over a long period of domestication for each genome. In addition, the increased number of WRKY genes correlates closely with the frequency of duplication events, and these duplications might contribute to the survival of plants in adverse conditions and the functional diversity of the WRKY genes. Further, tissue expression profiles of the WRKY genes showed several BrWRKY genes that were highly expressed in different tissues, suggesting that they might play more important roles in developmental and metabolic processes. In addition, several BrWRKY genes showed cross-responsive expression patterns under different abiotic and biotic stresses, suggesting that they have complex mechanisms in response to multiple stresses. This study enabled us to further our research and to identify appropriate tissue-specific or stress-related WRKY genes and provide some insights into the response of Chinese cabbage to multiple stresses under field conditions. Determination of the role of the WRKY family members in these responses warrants further work.