The Kandelia obovata transcription factor KoWRKY40 enhances cold tolerance in transgenic Arabidopsis

WRKY transcription factors play key roles in plant development processes and stress response. Kandelia obovata is the most cold-resistant species of mangrove plants, which are the important contributors to coastal marine environment. However, there is little known about the WRKY genes in K. obovata. In this study, a WRKY transcription factor gene, named KoWRKY40, was identified from mangrove plant K. obovata. The full-length cDNA of KoWRKY40 gene was 1420 nucleotide bases, which encoded 318 amino acids. The KoWRKY40 protein contained a typical WRKY domain and a C2H2 zinc-finger motif, which were common signatures to group II of WRKY family. The three-dimensional (3D) model of KoWRKY40 was formed by one α-helix and five β-strands. Evolutionary analysis revealed that KoWRKY40 has the closest homology with a WRKY protein from another mangrove plant Bruguiera gymnorhiza. The KoWRKY40 protein was verified to be exclusively located in nucleus of tobacco epidermis cells. Gene expression analysis demonstrated that KoWRKY40 was induced highly in the roots and leaves, but lowly in stems in K. obovata under cold stress. Overexpression of KoWRKY40 in Arabidopsis significantly enhanced the fresh weight, root length, and lateral root number of the transgenic lines under cold stress. KoWRKY40 transgenic Arabidopsis exhibited higher proline content, SOD, POD, and CAT activities, and lower MDA content, and H2O2 content than wild-type Arabidopsis under cold stress condition. Cold stress affected the expression of genes related to proline biosynthesis, antioxidant system, and the ICE-CBF-COR signaling pathway, including AtP5CS1, AtPRODH1, AtMnSOD, AtPOD, AtCAT1, AtCBF1, AtCBF2, AtICE1, AtCOR47 in KoWRKY40 transgenic Arabidopsis plants. These results demonstrated that KoWRKY40 conferred cold tolerance in transgenic Arabidopsis by regulating plant growth, osmotic balance, the antioxidant system, and ICE-CBF-COR signaling pathway. The study indicates that KoWRKY40 is an important regulator involved in the cold stress response in plants.


Background
Mangrove plants are tropical/subtropical communities of xylophyta that grow in the intertidal zones of coastlines [1]. As an important marine wetland ecosystem, mangrove plants play important ecological and economic roles, such as keeping coastlines and beaches away from storm tide and tsunami, remediating contaminated environment, and providing food and shelters for marine organisms [2,3]. As a dominant community, mangrove plants have evolved to adapt and survive in such extreme habitat (high ultraviolet radiation, high salinity, submerged and hypoxia soil) [4]. Generally, plants have developed complex mechanisms to adapt to stress conditions. Among them, some important transcription Open Access *Correspondence: yswang@scsio.ac.cn; chenghao@scsio.ac.cn factors, such as WRKY transcription factor (TF) family, have participated in the response to various tresses by activating or repressing the expression of the related genes [5]. Since mangrove plants are typically materials for investigating plant adaptive traits [6], the WRKY TF genes might provide clues to understand the adaptation and stress resistance mechanisms of mangrove plants. However, the roles of the WRKY genes in mangrove plants responding to stress are still poorly understood.
The WRKY TFs, one of the largest transcription factor families, involved in various biological processes in plants [7]. WRKY family members have a highly conserved 60 amino acids of DNA-binding regions, called WRKY domain. The WRKY domain contains the conserved WRKYGQK sequence at N-terminal and zincfinger structure at C-terminal [8]. The DNA binding domain of WRKY is mostly invariant WRKYGQK, but it still has differences from other domain, such as WKKYGQK, WRKYGMK, WSKYGQK, WKRYGQK, WVKYGQK, and WRKYGKK [9][10][11]. Almost all the known WRKY proteins with WRKYGQK sequence can recognize the W-box (TTG ACC /T sequences) in promoter region [12]. The zinc-finger structures mainly contain C2H2 type (C-X 4-5 -C-X 22-23 -H-X 1 -H) and C2HC type (C-X 7 -C-X 23 -H-X 1 -C) [13]. Based on the number of WRKY domains and the structural feature of zinc-finger motifs, WRKY proteins can be divided into three groups, namely I, II, and III [13]. Group I WRKY proteins have two WRKY domains and a C2H2 zinc-finger structure. Group II WRKY proteins contain only one WRKY domain and a C2H2 zinc-finger structure. Finally, group III WRKY proteins have one WRKY domain and a C2HC-type zinc finger [8,13,14].
Many reports have suggested that WRKY proteins were involved in plant developmental processes and various abiotic defense responses. Studies showed that expression of WRKY genes were induced under cold and salt stresses in plants [15][16][17][18]. Additionally, overexpression of some WRKY genes had regulated the expression levels of many other stress-related genes, and also enhanced tolerance to salt, cold, heat, or drought stress in transgenic plants [18][19][20][21]. However, there was little study about WRKY genes in mangrove plants [6].
Kandelia obovata is widely distributed in the world and is the most cold-resistance species in mangrove plants [22]. In our previous study, the nucleotide fragment (Ko1140) isolated from K. obovata, and showed homology with other WRKY proteins [23]. In order to better understand it, the full-length sequence of this novel gene, named KoWRKY40, has been cloned and analyzed for its structure and function in this study. The evolutionary relationship of KoWRKY40 protein with other WRKY proteins was analyzed. The expression patterns of KoWRKY40 in response to cold stress in different tissues of K. obovata were characterized. The subcellular localization of KoWRKY40 was performed based on the expression of GFP fusion proteins in tobacco. Furthermore, overexpression of KoWRKY40 in Arabidopsis thaliana was carried out for cold resistance analysis. This study provided useful clues for further exploring the functional mechanism of KoWRKY40 in K. obovata.

Characterization and sequence analysis of the KoWRKY40
Initially, a fragment with 334 bp was obtained from our previous study, which has the homology with those known WRKY genes. According to RACE technology, the full-length cDNA of this WRKY gene was obtained by sequence assembly and re-amplification. Sequence analysis revealed that the cDNA fragment is 1420 bp in length, containing a 127-bp 5′-untranslated region (UTR), a 336bp 3′-UTR and a 957-bp complete open reading frame (ORF). The gene encoded a protein with 318 amino acid residues with an estimated molecular mass (MW) of 33.59 kDa and isoelectric point (pI) of 8.76. This gene has been deposited in GenBank (GenBank accession No. KP267757.1) and was designated as KoWRKY40 in this study. According to EXPASy Molecular Biology Server, the Ser (10.1%), Lys (7.9%), Thr (7.2%), Val (7.2%), Ala (6.9%), Asn (6.9%) Glu (6.9%) and Leu (6.6%) contents were high, but the Trp occupied the lowest (0.6%) portion in KoWRKY40 amino acid sequence. In the secondary structure of KoWRKY40, α-helix accounted for 41.19%, β-sheet for 16.35%, β-turn for 5.97%, and random coil for 36.48%. Bioinformatics analysis showed that KoWRKY40 contained two transmembrane regions (residue positions 45-60 and 299-314), but had no predicted signal peptide. A putative nuclear localization signal (NLS) RKRK is existed on the residue positions 99-102 ( Fig. 1), showing that KoWRKY40 is located in the nucleus. The sequence alignment indicated that KoWRKY40 protein had high homology comparing with other eight WRKY TFs (Fig. 1). These nine WRKY TFs all contain a WRKYGQK sequence and a C2H2 zinc finger structure, which are the typical features of Group II WRKY proteins, indicating that they belong to the Group II WRKY subfamily (Fig. 1).

Three-dimensional model of KoWRKY40
Based on the deduced amino acid sequence, the predicted three-dimensional (3D) model of KoWRKY40 was constructed using SWISS MODEL software (Fig. 2). There is a 50% similarity between the amino acid sequences that need for the formation of the 3D model [24]. Modeling results showed that the sequence similarity was 51.35% between the KoWRKY40 and the template (AtWRKY1, SNYL id: 2ayd.1.A) [25], demonstrating the 3D model of KoWRKY40 was reasonable and credible. According to the surface charge distribution of 3D structure, positive charges occupied more area than negative charges in KoWRKY40 protein (Fig. 2B), indicating KoWRKY40 was a positive charge protein. Sequence analysis indicated that KoWRKY40 contained 38 negatively charged residues (Glu and Asp) and 43 positively charged residues (Lys and Arg), which means that KoWRKY40 was positively charged. In general, a positive charge indicates the hydrophobicity of the protein. Sequence analysis reconfirmed that the protein is indeed hydrophobic. The structure of KoWRKY40 was consisted of a fivestrand anti-parallel β-sheet (β1, 149-153; β2, 168-174; β3, 183-188; β4, 197-202; β5, 210-215. see Fig. 2A). The 147 residues at the N-terminus (including the NLS motif ) and the 94 residues at the C-terminus are not included in the structure (see Fig. 2D). Due to the α-helix (Leu154-Asp157) and the long bridging loop (Thr158-Gln167) between β1 and β2, the structure of KoWRKY40 looked spherical and stable ( Fig. 2A, B). The zinc ion existed as an independent ligand and zinc coordination residues interact with amino acid residues at Cys188, Cys199, His218, His220, which are the core residues of the C2H2 zinc finger structure of KoWRKY40 ( Fig. 2A). The C2H2 structure was located at one end of the β sheet (β5), between strands β3 and β4. The presence of zinc ion was essential for the DNA-binding activity, indicating the importance of the zinc-binding motif [14,26]. In Fig. 1, there were more than 17 well-conserved residues in the area between β2 and β3, including WRKYGQK sequence. Since the WRKY motif is responsible for binding to W-box [14], and the 'WRKYGQK' sequence of WRKY motif spanned the entire β2 strand, indicating the importance of β2 of KoWRKY40 in the DNA-binding activity. Therefore, we infer that the β2 and β3 strands are The similar and selfsame amino acids among WRKY proteins are marked light and dark gray, respectively. The predicted α-helix and five β-strands are indicated at the top of the sequence by a black box and five wide white arrows, respectively. Conserved WRKYGQK of WRKY proteins used to stabilize the structure and recognize DNA has been marked with a red dashed box. The yellow background indicates the conserved zinc finger motif (C2H2, C-X4-5-C-X22-23-H-X1-H). The residue of "RKRK", highlighted in green, is the nuclear localization signal (NLS) of KoWRKY40 likely to participate in DNA binding, and the loop between β2 and β3 may participate in conformational changes of DNA binding [25].

Phylogenic analysis of KoWRKY40
We have downloaded all available Arabidopsis WRKYs by NCBI blast. The phylogenetic analysis of KoWRKY40 protein with Arabidopsis WRKYs was performed as shown in Supplementary Fig. S1. In comparing with the 58 Arabidopsis WRKYs, KoWRKY40 showed the closest phylogenetic relationship to WRKY transcription factor 40 of A. thaliana that belongs to Group II WRKY family. In order to further investigate the evolutionary relationship of KoWRKY40, 26 WRKY TF proteins that showed close homology with KoWRKY40 sequence by NCBI blast were downloaded and a phylogenetic tree was constructed (Fig. 3). These representative WRKY TFs were classified into three groups, and the clusters and groups were well-supported by the bootstrap values. The phylogenetic tree indicated that KoWRKY40 was more closely related to the Group II of WRKY family. Besides, the KoWRKY40 protein showed the closest phylogenetic relationship to BgWRKY, which was a putative WRKY protein from mangrove plant Bruguiera gymnorhiza (GenBank accession No. BAG15874.1) [27].

Nuclear localization of KoWRKY40 in transiently transformed tobacco
As a transcription factor, WRKY proteins always possess NLS motif and were located in nucleus [28]. In order to verify the subcellular localization of KoWRKY40, the control vector 35S-GFP and the fusion expression vector 35S-KoWRKY40-GFP were transiently transformed into the epidermal leaf cells of tobacco Nicotiana benthamiana ( Fig. 4), respectively. DAPI fluorescent dye was used as a nuclear marker. In leaves transformed with 35S-GFP vector, the green fluorescent signal of GFP was distributed through the cell. However, the green fluorescent signal was targeted specifically to the nucleus in leaves transformed with 35S-KoWRKY40-GFP (Fig. 4). These results demonstrated that KoWRKY40 is a nuclear-localized protein, and provided direct evidence for the nuclear localization of KoWRKY40.

Expression patterns of KoWRKY40 in response to cold stress
Cold stress usually affects plant growth. The leaves of K. obovata gradually withered with time extension under cold stress (Fig. 5A). Compared with normal condition (CK), the expressions of KoWRKY40 were obviously induced in different tissues under cold stress (Fig. 5B). In leaves, the expressions of KoWRKY40 were significantly higher under cold stress than that under normal condition during the whole tested-period (p* < 0.05). In stems and roots, the expressions of KoWRKY40 were significantly increased under cold stress in comparison with normal condition (p* < 0.05) at 4 d, 7 d, 15 d and 20 d.
In addition, the gene expressions of KoWRKY40 were increased highly in the roots and leaves, but lowly in the stems under cold stress (Fig. 5B). During the tested time period, the expressions of KoWRKY40 were all increased at first, and then decreased in leaves, stems and roots, respectively in cold-treatments. The difference was that the time points of the highest expression levels were different, at 2 d (15.28-fold), 4 d (5.47-fold) and 15 d (106.82-fold) in leaves, stems and roots, respectively. These data indicated that leaves were the first to respond to cold stress, followed by stems and roots. After 4-days cold treatment, the expression levels of KoWRKY40 were higher in roots (17.65-fold) than that in leaves and stems under cold stress. These results suggested that KoWRKY40 mainly reacted in the leaves at early phases (before 2 d), and principally played roles in roots at late phases (after 4 d) under cold stress. After 20-days cold treatment, the expressions of KoWRKY40 in different tissues were decreased to low levels, probably because the plants withered at this time point (Fig. 5A). These results may suggest that the 20-days of cold treatment may have exceeded cold tolerance limit time of K. obovata. It was indicated that KoWRKY40 may play important roles in the signaling network of K. obovata in response to cold stress.

Overexpression of KoWRKY40 enhanced the tolerance to cold stress in transgenic Arabidopsis
To investigate the role of KoWRKY40 in the response to cold stress, transgenic Arabidopsis plants that overexpressed KoWRKY40 were generated. The wild-type (WT) Arabidopsis which has no expression of KoWRKY40 transcript, and three KoWRKY40 transgenic Arabidopsis lines (Line 1, Line 3 and Line 6) which demonstrated relatively high expression of KoWRKY40 (Fig. 6), were selected and analyzed for stress tolerance. As shown in Fig. 7, both transgenic and wild-type (WT) seedlings grew well and showed no significant difference in phenotype, fresh weight, root length, and lateral root number under normal growth condition (CK). Nevertheless, the transgenic lines showed more green leaves, less black or yellow leaves than the wild type under cold stress (Fig. 7B). Besides, the fresh weight (Fig. 7C), root length (Fig. 7V), and lateral root number (Fig. 7E) of KoWRKY40 transgenic Arabidopsis plants were significantly higher than that of WT plants under cold stress condition. The results showed that KoWRKY40 transgenic lines grew better than WT plants under cold stress condition.
To explore the involvement of KoWRKY40 in osmoregulation in Arabidopsis under cold stress condition, the proline content was measured in WT and transgenic plants. As shown in Fig. 8A, the proline content was significantly higher in transgenic lines than in WT lines under cold stress condition, but did not differ significantly between the two types of plants under normal growth condition. These results suggested that overexpression  To further investigate the regulatory role of KoWRKY40 in response to cold stress, the expression levels of important genes involved in osmotic adjustment (AtP5CS1 and AtPRODH1), ROS scavenging (AtMnSOD, AtPOD and AtCAT1) and ICE-CBF-COR signaling pathway (AtCBF1, AtCBF2, AtICE1 and AtCOR47) were examined in KoWRKY40 transgenic and WT plants. The expression levels of AtP5CS1, AtMnSOD, AtPOD, AtCAT1, AtCBF1, AtCBF2, AtICE1 and AtCOR47 were significantly higher in KoWRKY40 transgenic lines than in WT plants (as shown in Fig. 9). However, the expression level of proline dehydrogenase gene, AtPRODH1, was remarkably lower in KoWRKY40 transgenic lines than in WT plants (Fig. 9B). These results suggested that overexpression of KoWRKY40 in Arabidopsis regulated the expression of genes related to osmolytes, antioxidant biosynthesis, and ICE-CBF-COR signaling pathway under cold stress condition, generating transgenic Arabidopsis plants with improved osmoregulation, antioxidant defenses, and cold-related genes activation, thereby conferred to cold tolerance.

Discussion
Many researches have reported that the WRKY proteins play important roles in response to abiotic stresses [29][30][31] Although there have been many studies on WRYK gene in other plants [8,13,14], little was known about WRKY genes in mangrove plants [6]. In this paper, based on our previous study [23], we cloned the full-length sequence of KoWRKY40 gene from K. obovata. The deduced amino acid sequence showed that KoWRKY40 was a group II WRKY protein.
In general, WRKY proteins are characterized by their domains, which are typically about 60 residues in length and has a conserved WRKYGQK motif. The WRKYGQK motif is a W-box element (C/T) TGAC (C/T) located in the promoter region of the target genes and is essential for DNA binding activity [14]. The 3D structure of WRKY4 exactly showed that conserved WRKYGQK residues can participate in DNA binding [32]. However, the core WRKYGQK sequence of WRKY protein is sometimes replaced by WRKYGKK, which is common variant in canola [33], tomato [9], and pepper [34]. It has been found that each amino acid residue of WRKYGQK sequence was replaced, the binding affinity with DNA was significantly decreased [26]. In our study, the KoWRKY40 possessed the WRKYGQK motif. Therefore, we speculated that the KoWRKY40 may recognize the W-box based on its typical WRKYGQK motif. The transmembrane helix is required for membraneassociated proteins related to cell signal transduction and substrate transport [35]. Sequence analysis showed that the KoWRKY40 protein contained two transmembrane helixes, indicating the KoWRKY40 might actively participate in signal transduction. Previous researches had reported that most WRKY proteins contained a basic NLS [36], and GFP-WRKY fusion protein have demonstrated that diverse WRKY TFs were located in nucleus [20,36,37]. In this study, KoWRKY40 protein contained the NLS sequence RKRK and was predicted to be localized in the nucleus. Some studies have reported that merged fluorescence signals of GFP and DAPI were visualized for subcellular localization [38,39], the DAPI staining was used to detect the florescent signals of nuclei for further confirmed the nuclear localization of KoWRKY40. Subcellular localization analysis further confirmed that KoWRKY40 was nuclear-localized protein in this study. These data indicated that our results were consistent with previous studies.
Some studies have reported that the expression of WRKY was tissue specific under abiotic stresses. For example, the LcWRKY5 gene was expressed only in leaves and roots in sheepgrass [40]. In Arabidopsis, the AtWRKY25 gene was mainly expressed in roots [41]. The expression of CsWRKY2 in leaves was higher than in other organs in defense response in tea plant [20]. In our study, the KoWRKY40 gene was highly induced in the roots and leaves under cold stress in K. obovata. Thus, we speculated that KoWRKY40 may play an important role in K. obovata under cold stress.
It was worth mentioning that the expression level of KoWRKY40 was increased to the highest at 15 d (106.82fold) in roots under cold stress, but that was dramatically decreased at 20 d (2.64-fold). This variation tendency of KoWRKY40 expression levels was usually agreement with morphological changes. Seedlings that had been cold-treated (5 °C) for 15 days were still alive after being cultured for 20 d at the recovery temperature (25 °C). However, seedlings that had been cold-treated (5 °C) for 20 days have withered and fallen (data not shown here) under the same recovery condition. These results suggested that the expression of KoWRKY40 may be closely related to the survival of K. obovata, and play important roles in protecting K. obovata from cold stress. Based on the high consistency between expression of KoWRKY40 at the molecular level and recovery survival at morphological level, we speculated that 15-20 days may be the upper limit of survival time of K. obovata seedlings at low temperature (5 °C).
WRKY proteins have involved in the response of plants to various stresses, and their homologs have been found in various plant species [29]. Several WRKY proteins were found to be induced in response to cold stress in maize [42], cotton [43] and barley [44]. However, there was an evidence reported that one tobacco WRKY gene was specifically induced during collaborative pressure of drought and heat, but not by drought or heat stress alone, indicating this WRKY genes were induced only by a combination of different stresses [45]. In this study, the KoWRKY40 expression was induced during the entire process under cold stress, indicating the KoWRKY40 gene could be induced by cold stress individually. Since mangrove plants generally grow in the intertidal zones and were subjected to various stresses [1][2][3], whether other WRKY genes of mangrove plants were induced by individual or combined stresses remains to be studied. WRKY protein can activate or inhibit the transcription of downstream by combining with W-box cisacting elements of target genes [8,46]. Since W-box elements existed in most WRKYs, the WRKYs can bind with their own promoters to achieve self-regulation or cross-regulation networks [47]. CaWRKY6 transcriptionally activated CaWRKY40, and conferred the plant more tolerant to high temperature and humidity in pepper [48]. AtWRKY34 had a negative regulatory effect on cold response pathway, inducing enhanced resistance to cold stress [49]. WRKY proteins could be quickly and differentially expressed, regulating the expression of downstream genes and promoting signal transduction [41]. For example, the antioxidant enzyme genes AtSOD, AtCAT , and AtAPX1, which can scavenge reactive oxygen species (ROS) to protect plants from oxidative damage [50], are important downstream genes regulated by WRKY TFs. In our study, the expression levels of AtMnSOD, AtPOD, and AtCAT1 were up-regulated in KoWRK40 transgenic Arabidopsis plants under cold stress, acceleration of which could increase tolerance to cold stress in transgenic Arabidopsis plants.
The plant response to cold stress is rather complex. A frequent plant response to abiotic stress is to accumulate certain osmolytes, particularly proline [51][52][53], which then function as osmoprotectants. In Arabidopsis, the genes encode P5CS isoforms involved in proline biosynthesis, AtP5CS1 (AT2G39800), and the genes encode ProDH isoforms involved in the degradation of proline, AtPRODH1 (AT3G30775) [54][55][56]. In this study, the proline biosynthesis gene AtP5CS1 was up-regulated, and the proline degradation gene AtPRODH1 was down-regulated in KoWRKY40 transgenic Arabidopsis, compared with their expression in WT plants (Fig. 9E, F). This suggested that KoWRKY40 increased the proline content to enhance the osmoregulatory capacity of transgenic Arabidopsis in response to cold stress by inducing proline biosynthesis genes and inhibiting proline degradation genes. Cold stress also can induce the rapid generation and accumulation of ROS, resulting in secondary oxidative stress in plants [57]. Antioxidant enzymes (e.g., SOD, POD and CAT) can scavenge ROS to protect plants from oxidative damage [50]. The increase of cold tolerance in plants may due to the high expression of these genes [58][59][60]. In the present study, KoWRKY40 transgenic plants exhibited higher activities of SOD, CAT, POD, and lower contents of MDA and H 2 O 2 than WT Arabidopsis under cold stress conditions (Fig. 8B-F). Three genes related to ROS scavenging, AtMnSOD, AtPOD, and AtCAT1, were up-regulated in KoWRKY40 transgenic Arabidopsis (Fig. 9C-E). Thus, our data suggested that KoWRKY40 increased the activities of antioxidant enzymes, and decreased membrane system damage in transgenic Arabidopsis plants under cold stress condition. In A. thaliana, AtCBF1, AtCBF2, AtICE1 and AtCOR47 genes involved in ICE-CBF-COR signaling pathway have been proved to play important roles in cold tolerance [61][62][63][64][65]. In our study, overexpression of KoWRKY40 enhanced cold tolerance in transgenic Arabidopsis plants and these cold-resistance genes were all significantly increased in KoWRKY40 transgenic Arabidopsis plants. Thus, we suggested that accumulation of which may help to develop cold acclimation and protect transgenic plants from cold damage.

Conclusions
To summarize, KoWRKY40, a new WRKY transcription factor gene from mangrove plant K. obovata was reported in this study. The KoWRKY40 protein was a nuclear-localized protein and a member of group II WRKY family. The expression of KoWRKY40 was highly induced in the roots and leaves in K. obovata under cold stress. In addition, KoWRKY40 enhanced cold tolerance of transgenic Arabidopsis lines by activating different stress responsive genes involved in osmotic adjustment (AtP5CS1 and AtPRODH1), ROS scavenging (AtMnSOD, AtPOD and AtCAT1) and ICE-CBF pathway (AtCBF1 ,  AtCBF2, AtICE1, AtCOR47). These results provide key insight into the role of this gene in K. obovata that KoWRKY40 acts as a positive regulator of cold stress tolerance. This study suggested that KoWRKY40 may be a valuable genetic resource in molecular breeding programs of plants. Future studies are needed to elucidate the functional mechanism of KoWRKY40 against abiotic stresses.

Plant material, growth conditions and treatments
The hypocotyls of K. obovata were provided by Guangdong Mangrove Ecological Technology Co. Ltd. (China). The hypocotyls were surface-disinfected and sown in clean sand at room temperature. The 3-months seedlings were transferred into growth chamber with normal condition (25 °C, relative humidity 75%, 14 h light/10 h dark cycle). After 7 days, the seedlings were treated under cold condition (5 °C) for 0 h, 6 h, 12 h, 1 d, 2 d, 4 d, 7 d, 15 d and 20 d. Seedlings treated at 5 °C for 0 h were used as the control. All treatments contained at least three seedlings. The harvested leaves, stems and roots were immediately frozen in liquid nitrogen, and then transferred to − 80 °C refrigerator until use.

RNA extraction and reverse transcription
Total RNA was extracted from leaves, stems and roots of K. obovata using the Tiangen RNA plant Plus Reagent (Tiangen Biotech, Beijing) according to the method as described [66,67]. The 1.0% agarose gel was used to analyze the integrity and purity of total RNA. The quality and content of RNA were determined by Nanodrop 1000 spectrophotometer (Thermo Scientific, USA). The RNase-free DNase I (Promega, USA) was used to remove potential genomic DNA contamination of total RNA. The RNA samples was then subjected to synthesize the first strand cDNA by SMART ™ reverse transcription Kit (Clontech, USA) following the manufacturer's protocol. The cDNA samples were used for cloning the full-length of KoWRKY40 gene and RT-qPCR analysis.

Cloning the complete sequence of KoWRKY40 gene
The rapid amplification of cDNA ends (RACE) PCRs was performed from the first cDNA strand of K. obovata leaves. A complete cDNA sequence was assembled by combining the 5′-RACE and the 3′-RACE fragments. Based on the partial sequences (GenBank accession number: JZ585678.1) from our previous study [23], the gene-specific primers (GSP1, 5′-GCT ACT AGG ACT GAA ACA TCC TCC GCAC-3′, GSP2, 5′-GGT GGC AAT ACT GAA ACC AGC TGT AGCGA-3′) were used as primary PCR to obtain 5′ and 3′ end sequences, respectively. The nested PCR reactions were performed by nested primers (NGSP1, 5′-ATC CTC CGC ACT TCT CTG GAC CTT CTTC-3′, NGSP2, 5′-GAA ACC AGC TGT AGC GAC GAA GAG TCAA-3′). All steps of RACE were carried out using SMART ™ RACE cDNA Amplification Kit (Clontech, USA) on the basis of the manufacturer's protocol. The RACE products were purified by agarose gel and then were cloned into pMD19-T Vector for sequencing. The 3′-and 5′-nucleotide sequences were assembled by DNAMAN software through overlap to acquire the fulllength KoWRKY40 sequence. The assembled sequence was used to design the primers, and the full-length sequence of KoWRKY40 was cloned and sent to the company (BGI, China) for sequencing. Finally, the complete cDNA sequence of KoWRKY40 was confirmed, and was submitted to GenBank to obtain the accession number KP267757.1.

Expression analysis by RT-qPCR
To analyze KoWRKY40 expression in K. obovata under cold stress, the real-time quantitative PCR (RT-qPCR) was performed to determine the transcription levels in leaves, stems and roots of K. obovata. The samples were collected from different tissues of K. obovata at 0 h, 6 h, 12 h, 1 d, 2 d, 4 d, 7 d, 15 d and 20 d, respectively, under cold condition (5 °C). For comparison, the samples were also collected at same time points under normal condition (25 °C). The samples collected at 0 h were considered as the control. Total RNA was extracted and reversetranscribed into cDNA in the above experiment. The internal reference gene was selected with the 18S (Ko18S) rRNA of K. obovata. The RT-qPCR reactions were carried out by iCycler iQ5 real time PCR detection system (Bio-Rad, CA, USA) according to SYBR Premix Ex TaqTM II (Takara, Dalian, China) manufacturer's instructions.
The RT-qPCR reaction parameters were 95 °C for 1 min, 40 cycles at 95 °C for 10 s, 55 °C for 30 s, and 72 °C for 40 s. The 2 -△△CT method was used to calculate the relative expression levels of gene [68,69] comparing with 0 h. Sequences of primers used for RT-qPCR analysis are listed in Supplementary Table S1. Three biological replicates and three technical replicates were performed for each sample. The data was expressed as mean ± standard deviation (x ± SD). Statistical analyses were carried out by Student t-test, and diagramming was performed with GraphPad Prism 7.0 (GraphPad Software, San Diego, California). One-way ANOVA was selected to quantify the significant differences among expressions of KoWRKY under cold stress and normal condition by SPSS statistics 25.

Subcellular localization analysis
To determine the subcellular localization of KoWRKY40, the ORF of KoWRKY40 termination-codon free was cloned into vector pCAMBIA2301-35S-GFP. The Nicotiana benthamiana leaf preparation and transformation were performed as described [39]. The GFP fusion expression vector pCAMBIA2301-35S-KoWRKY40-GFP was analyzed and sequenced, and the successful fusion was confirmed. The recombinant plasmid was introduced into Agrobacterium tumefaciens strain EHA105, and transformed transiently into the leaf epidermal of N. benthamiana by Agrobacterium infection. The 35S-GFP vector was taken as the control. To locate the fluorescent proteins in nuclei, the N. benthamiana leaves were infiltrated with PBS containing 4′,6-diamidino-2-phenylindole (DAPI) for 20 min, and the fluorescence microscopic image was observed by Zeiss LSM710 laser scanning confocal microscope to determine the subcellular location of KoWRKY40-GFP.

Generation of KoWRKY40 transgenic Arabidopsis plants
For KoWRKY40 overexpression in wild-type Arabidopsis thaliana (WT; Columbia), the full-length KoWRKY40 sequence without the stop codon was cloned into pCAM-BIA2301 and driven by CaMV 35S promoter. The coding sequence of KoWRKY40 (with KpnIsite added to its 5′ and 3′ ends, respectively) was amplified from pMD19-T-KoWRKY40 using gene-specific primers F1 (5′-GCG GGT CGA CGG TAC CAT GGA ATC AAA ATG GGT GAA C-3′) and R1 (5′-TAG ACA TAT GGG TAC CGA ATG TGG TTC CTG AAAGG-3′). The digested amplicon was inserted into pCAMBIA1301 driven by the CaMV 35S promoter, and confirmed by sequencing. This recombinant vector, named pCAMBIA2301-35S::KoWRKY40, was transformed into A. tumefaciens EHA105 by the freeze-thaw method and then transformed into Arabidopsis plants by the floral dip method [70]. Positive Arabidopsis transgenic lines were harvested and selected by culturing on MS medium agar plates containing 50 mg/L kanamycin, and these were advanced by self-pollinated until obtaining T 3 transgenic plants. Transgenic plants were validated further by RT-PCR (reverse transcription-PCR) analysis with gene-specific primers F2 (5′-GAC GCA CAA TCC CAC TAT CC-3′) and R2 (5′-GAA TGT GGT TCC TGA AAG G-3′). Finally, the T3 or T4 homozygous lines were used for all the subsequent experiments.

Physiological analysis of transgenic A. thaliana lines under cold stress
Seeds of WT Arabidopsis and KoWRKY40 transgenic Arabidopsis lines (Lines 1, 3 and 6) were surface sterilized by soaking in 70% ethanol (v/v) for 5 min, and then rinsed four to five times with sterile distilled water. The sterilized seeds were grown on solidified MS medium for 6 days and transplanted onto new square plates with MS medium under normal conditions for 10 days before treatments. For cold tolerance evaluation, some wildtype and KoWRKY40 transgenic plants were cultured under normal conditions as the control and others were treated under cold stress (5 °C) for 10 days, and their fresh weight, root length, lateral root number and other physiological parameters were measured. The contents of proline, MDA, and H 2 O 2 , and the activities of SOD, POD, and CAT, were measured using the corresponding assay kits (Jiancheng Bioengineering Institute, Nanjing, China) according to the manufacturer's instructions. The data was expressed as mean ± standard deviation (x ± SD). All the experiments were carried out in triplicate for biological replicates. Statistical analysis was performed with GraphPad Prism 7.0 (GraphPad Software, San Diego, California). One-way ANOVA followed by Duncan's test was selected to quantify the significant differences by SPSS statistics 25.

Expression analysis of stress-related genes in transgenic A. thaliana lines
To examine the expression of stress-related genes, cDNA was synthesized from RNA extracted from the leaves of WT Arabidopsis and KoWRKY40 transgenic lines (Lines 1, 3 and 6). The expression levels of coldrelated genes were measured by qRT-PCR as described above. The stress-related genes monitored were AtP5CS1 (AT2G39800), AtPRODH1 (AT3G30775), AtMnSOD (AT3G56350), AtPOD (AT3G49120), AtCAT1 (AT1G20630), AtCBF1 (AT4G25490), AtCBF2 (AT4G25470), AtICE (AT3G26744) and AtCOR47 (AT1G20440). The expression level of KoWRKY40 in transgenic lines was also measured. Arabidopsis AtAC-TIN2 (AT3G18780) was used as a reference gene in the RT-qPCR reactions. The gene primer sequences used are listed in Supplemental Table S1. Relative gene expression values were calculated using the 2 -ΔΔCt Method as described above. Three biological replicates and three technical replicates were performed for each sample. The data was expressed as mean ± standard deviation (x ± SD). All the experiments were carried out in triplicate for biological replicates. Statistical analysis was performed with GraphPad Prism 7.0 (GraphPad Software, San Diego, California). One-way ANOVA followed by Duncan's test was selected to quantify the significant differences by SPSS statistics 25.