Genomic sequences of TabZIP14 were isolated from species of different ploidies, including Triticum urartu accession UR206 (A genome; Lebanon, Asia), Aegilops speltoides accession Y2006 (S genome; Syria, Asia), Aegilops tauschii accession Y2282 (D genome; Armenia, Asia), and Chinese Spring wheat (CS, ABD genome; Sichuan, China).

The complete nucleotide sequence of the TabZIP14-B gene was derived from a Chinese Spring wheat leaf full-length cDNA plasmid library. Gene-specific primers (forward 5′-CGATCAAACCCTCGAA-3′ and reverse 5′-CACAGCCAAAGAACAAA-3′) were designed to obtain the genomic sequences of TabZIP14. The alignment of the three TabZIP14 homeologous genomic sequences with the corresponding cDNA sequences was used to determine the introns and exons of the genomic DNA sequences.

C-bZIP sequences were acquired from wheat and other plant species through protein BLAST searches¹. Amino acid sequence identity analysis was performed using the MegAlign program in DNASTAR. Phylogenetic analysis was carried out using MEGA 5.1 software.

The promoter regions of three TabZIP14 homeologous genes were acquired by comparing their genomic sequence with Chinese Spring genome reference². These three promoter sequences were submitted to the PLACE database and analyzed to identify cis-acting elements³.

A total of 213 wheat accessions, including 31 tetraploid species, 97 landraces, and 85 modern cultivars, were used to detect the SNPs and haplotypes of TabZIP14-B (Supplementary Table S1). The genomic sequence of TabZIP14-B was used to blast (E-value is 0.01) the 630,517 probe sequences from the wheat 660K SNP array, and then the matched probe sequences were selected to screen 213 wheat accessions for SNP discovery in the TabZIP14-B genomic sequence region. Standard statistical analyses of the genetic diversity of TabZIP14-B, including polymorphism information content (PIC) and gene diversity, were carried out using PowerMarker v3.25 (Liu and Muse, 2005).

To detect the expression of three TabZIP14 homeologous genes in response to PEG, NaCl and ABA treatments, germinated wheat seedlings were grown at 25°C in an artificial climate chamber with a 16 h light/8 h dark cycle. Ten-day-old seedlings of the Hanxuan 10 wheat cv. (drought resistant), the Chadianhong wheat cv. (salt-resistant) and CS were soaked in a 16.1% PEG-6000 solution, a 250 mM NaCl solution and a 200 μM ABA solution, respectively. Roots were sampled for expression analysis at 0, 1, 3, 6, 12, 24, and 48 h after treatment. For the low-temperature (4°C) treatment, the CS seedlings were transferred from normal growth conditions to a 4°C artificial climate chamber and leaves were harvested for expression analysis at 0, 1, 3, 6, 12, 24, and 48 h. Plant tissues (roots, stems, leaves, and young spikes) were obtained from CS wheat at the booting stage for tissue expression analysis of TabZIP14-B.

The expression levels of AtRD29A, AtCOR47, AtRAB18, AtRD20, and AtGSTF6 were examined in transgenic plants. Triticum aestivum glyceraldehyde-3-phosphate dehydrogenase (TaGAPDH) (NCBI accession: AF251217.1) and Arabidopsis thaliana Actin (NCBI accession: NM_112764) were selected as internal references. RNA extraction and quantitative real-time PCR (RT-qPCR) were conducted in triplicate as previously described (Zhang et al., 2016). The 2^-ΔΔCT method was used to calculate the relative expression of the three TabZIP14 homeologous genes. The specific gene primer sequences used in this experiment are provided in Supplementary Table S2.

The open reading frame (ORF) of TabZIP14-B without the stop codon was obtained via PCR amplification using gene-specific primers. The purified PCR products were fused to the GFP-pEarleyGate vector containing the 35S promoter (35S::GFP-TabZIP14-B) using the gateway method. The empty 35S::GFP vector was used as the control. The 35S::GFP-TabZIP14-B fusion protein was introduced into Agrobacterium tumefaciens strain GV3101-pMP90, which was then transformed into epidermal cells of N. benthamiana (Sheludko et al., 2007). The transformed N. benthamiana leaves were cultivated for 2–6 days. Confocal laser scanning microscopy (Zeiss Lsm 700, Zeiss, Jena, Germany) was used to observe the transformed tobacco cells.

The full-length TabZIP14-B cDNA was fused to the pDEST32 vector using the gateway method. The pGAL4 and pDEST32 vectors were used as positive and negative controls, respectively. The fused constructs were transformed into yeast strain AH109 with the His3 and Ade2 reporter genes. Transformed yeast cells were cultured on medium without histidine, leucine, and adenine to identify the transactivation abilities of the constructs.

The bait plasmid pRS315, which contains six copies of the ABRE/mABRE sequence and the Leu reporter gene, was provided by Professor Jun Zhao (Biotechnology Research Institute, Chinese Academy of Agricultural Sciences, Beijing, China). The full-length TabZIP14-B gene was introduced into the pDEST22 vector to generate the recombinant plasmid pDEST22- TabZIP14-B, which was used as the effector plasmid. The bait plasmid and effector plasmid were co-transformed into the yeast strain YM4271. The transformants were screened on SD/Leu^-Trp^- at 30°C for approximately 2 days. The selected yeast cells were coated onto SD/Leu^-Trp^-His^- with 0.5 mM 3-AT to examine cell growth. The pDEST22 vector and the bait plasmid used as the control were transformed into the YM4271 yeast strain, and the growth status of the transformants was checked.

The coding region of TabZIP14-B was ligated into the pEarleyGate 100 vector to obtain the 35S::TabZIP14-B recombinant plasmid, which was transferred into the GV3101-90RK Agrobacterium tumefaciens strain and then introduced into A. thaliana ecotype Col-0 via the floral dip method (Clough and Bent, 1998). The transgenic lines were sprayed with Basta solution and subsequently confirmed via RT-qPCR to screen homozygous lines.

For the freezing tolerance assay, WT and transgenic Arabidopsis seedlings were grown in potted soil for approximately 2 weeks, after which the plants were transferred to -10°C for 3 h. The plants were subsequently placed at 5°C for 2 h and then transferred back to normal conditions.

For the NaCl tolerance assay, WT and transgenic Arabidopsis seeds were transferred from Murashige and Skoog (MS) plates to the potted soil and well watered. Water was halted for 3 weeks, and the seedlings were watered with NaCl solution (250 mM) from the bottom of the plates. When the volume of the NaCl solution no longer reduced, the NaCl solution was removed, and the plants were grown under normal conditions.

To study the response of the transgenic plants to ABA, WT and transgenic Arabidopsis seeds were planted in MS medium for 5 days, transferred to new MS medium with 0 or 24 μM ABA and cultured vertically for 7 days. After that period, the root lengths of the plants were assessed. All experiments were repeated independently three times.

Free proline concentrations were determined based on a previously described method (Zhang et al., 2015). Electrolyte leakage and the soluble sugar content were measured according to a method described previously (Lehner et al., 2006; Cao et al., 2007).

All measurements were conducted in triplicate, and Student’s t-test was performed for statistical analysis.

The complete TabZIP14-B nucleotide sequence was obtained by sequencing a cDNA library. Sequence analysis showed that TabZIP14-B contained a 1576 bp gene fragment encoding a polypeptide of 400 amino acids including the conserved bZIP domain (279–393 aa), with a predicted molecular mass of 41.9 kDa and a pI of 5.06. The TabZIP14-B gene was submitted to GenBank under accession number KY113325.

To assay the molecular structures of the TabZIP14 homeologous genes in wheat, we isolated the genomic sequences from diploid wheat genomes A (UR206), S (Y2006), and D (Y2282) and three homeologous sequences from the hexaploid CS wheat. After comparison with the diploid TabZIP14 genomic sequences, the three hexaploid sequences from subgenomes A, B and D of hexaploid wheat were determined and denoted as TabZIP14-A, TabZIP14-B, and TabZIP14-D, respectively. To determine the gene structure of the TabZIP14, the cDNA sequences of each TabZIP14 gene were aligned with the gDNAs. The results indicated that the three homeologous genes all contained six exons and five introns; the intron length varied, specifically for intron 1, which varied from 1159 to 1709 bp; and the exon size was relatively conserved, with exons 2, 3, and 5 having the same lengths (Supplementary Table S3). TabZIP14-A, TabZIP14-B, and TabZIP14-D showed 94.3–96.2% similarity at the amino acid level. They grouped together in one group (Figure 1). Phylogenetic analysis with other species indicated that TabZIP14 homeologous proteins formed a phylogenetic tree with the members group C of the bZIP family. For instance, the amino acid sequence alignment showed that TabZIP14-B shared 72% identity with Zmopaque2 (NP_001105687.1), 41.4% identity with AtbZIP63 (NP_568508.2), 31.5% identity with AtbZIP25 (NP_567003.2), and 28.4% identity with AtbZIP10 (NP_192173.1) (Figure 1).

The 1800-bp promoter sequences of three TabZIP14 homeologous genes were obtained and analyzed for cis-acting regulatory elements individually. Seven cis-elements, including TATA-box, CAAT-box, G-box, GCN4-motif, LTR, MBS, and P-box, were identified in all three promoter regions. In addition to the seven cis-elements, O2-site involved in zein metabolism regulation was detected in the promoter region of TabZIP14-A, ERE involved in ethylene-responsive element was found in the promoter region of TabZIP14-B, and ABRE cis-element involved in ABA responsiveness was present in the promoter region of TabZIP14-D (Supplementary Table S4).

Genomic sequence variations of TabZIP14-B among 213 wheat accessions were also analyzed in the present study. Three SNPs were found in TabZIP14-B, including two in the first intron and one in the third exon, and three haplotypes, denoted Hap1, Hap2, and Hap3, were formed based on these three SNPs (Supplementary Figure S1A). Haplotype frequency in tetraploids, landraces, and modern cultivars showed that Hap1 increased but Hap2 decreased from tetraploids to landraces, while the pattern was opposite from landraces to modern cultivars (Supplementary Figure S1B). Moreover, gene diversity of TabZIP14-B was significantly reduced in landraces and modern cultivars compared with that in tetraploid species (Supplementary Figure S1C).

The tissue expression patterns of TabZIP14-B were analyzed via RT-qPCR. TabZIP14-B was expressed in all of the analyzed tissues, including the roots, stems, leaves, and young spikes. The relative expression levels were highest in the leaves (Supplementary Figure S2).

The expression levels of three TabZIP14 homeologous genes in wheat seedlings exposed to PEG, NaCl, ABA, and low temperature were also tested using the RT-qPCR method. The results showed that TabZIP14-A, TabZIP14-B, and TabZIP14-D had similar expression patterns. The expression levels of three TabZIP14 homeologous genes began to increase for 3 h and reached their highest levels 12 h after exposure to NaCl treatment, and the expression level of TabZIP14-B was higher than that of TabZIP14-A and TabZIP14-D 12 h post-treatment (Figure 2A). Under PEG and low temperature stresses, the expression of three TabZIP14 homeologous genes was induced and peaked 1 h post-treatment (Figures 2B,C). In addition, the expression of the three TabZIP14 homeologous genes was also induced under ABA conditions and peaked 24 h post-treatment (Figure 2D), and the expression level of TabZIP14-B was higher than that of TabZIP14-A and TabZIP14-D 24 h post-treatment.

Given that the expression patterns of the three TabZIP14 homeologous genes in wheat were similar in response to abiotic stress and that TabZIP14-B was obtained from a wheat full-length cDNA library, TabZIP14-B was chosen for further study.

We checked the subcellular localization of the TabZIP14-B protein and found that the cells transformed with 35S::GFP alone exhibited green fluorescence signals in both the cytoplasm and the nucleus. However, the expression of 35S::GFP-TabZIP14-B was only observed in the nucleus (Figure 3). These results demonstrated that the TabZIP14-B protein was localized to the nuclei.

The transactivation abilities of TabZIP14-B were analyzed using a yeast assay system. All transformants grew well on SD plates with SD/Leu^- medium. The yeast that expressed full-length TabZIP14-B and pGAL4 grew on SD plates with SD/Leu^-His^-Ade^- medium, but the yeast harboring the pDEST32 vector alone failed to grow on SD/Leu^-His^-Ade^- medium (Figure 4). These results indicated that the TabZIP14-B protein could activate the transcription of the reporter genes Ade and His in the yeast system, which suggests that the TabZIP14-B protein exhibits transcriptional activation capability.

Plant bZIP proteins bind to DNA sequences with an ACGT core cis-element. We examined whether the TabZIP14-B protein could bind to the ABRE cis-element using the yeast one-hybrid assay system. Figure 5 demonstrates that all of the transformants grew on medium containing SD/Leu^-Trp^-. However, only the transformants harboring the effector plasmid pAD-TabZIP14-B and the bait plasmid with the ABRE cis-element were able to grow on SD/Leu^-Trp^-His^- containing 3-AT medium, which suggests that the TabZIP14-B protein had an affinity for the ABRE cis-element.

We transformed Arabidopsis with TabZIP14-B and generated more than 10 transgenic lines. The expression levels of lines 1, 2, and 3 (denoted L1, L2, and L3, respectively) are shown in Supplementary Figure S3. These three lines were used to study gene function.

As TabZIP14-B expression is induced by ABA, we assessed whether overexpression of TabZIP14-B in Arabidopsis affects ABA sensitivity. ABA sensitivity was scored by measuring root length in transgenic and WT plants in the presence of 24 μM ABA for 7 days. The root growth of the transgenic plants was hindered more severely compared with that of the WT plants. This result indicated that the overexpression of TabZIP14-B resulted in severe ABA sensitivity and that TabZIP14-B participated in the ABA signaling pathway (Figure 6).

Many studies have indicated that group C-bZIP TFs are involved in freezing and salt tolerance. Therefore, we evaluated the ability of TabZIP14-B transgenic plants to resist abiotic stress. After treatment at -10°C for 3 h, nearly 90% of the WT seedlings were dead. However, the survival rates of the transgenic lines were approximately 60–90% (Figures 7A,B). Furthermore, electrolyte leakage in the transgenic plants was lower than in the WT plants when the plants were subjected to freezing stress (Figure 7C). The transgenic plants engendered obviously high levels of proline content compared with those of WT plants under freezing conditions (Figure 7D). However, the soluble sugar content changed little before and after freezing treatment (Figure 7E).

We also evaluated salt resistance between TabZIP14-B transgenic Arabidopsis and WT plants in the soil. There were no significant differences between WT and transgenic plants under normal conditions. However, when the Arabidopsis seedlings and WT plants were treated with 250 mM NaCl solution, 3 days later, most of the leaves of WT plants began to turn yellow and shrink. However, no evident shrinking was observed for most transgenic Arabidopsis seedlings. The transgenic lines looked greener than did the WT plants. Seven days later, most of the WT plants were dead. The survival rates of transgenic seedlings were 54–93%, significantly higher than in the WT plants (Figures 8A,B). In addition, after salt stress, TabZIP14-B-overexpressing lines exhibited lower electrolyte leakage than did WT plants (Figure 8C). Moreover, the proline content and soluble sugar levels were significantly higher in TabZIP14-B-overexpressing plants than in the WT plants (Figures 8D,E).

As the transgenic plants showed enhanced salt and freezing tolerance, we investigated whether abiotic stress-responsive genes were altered in the TabZIP14-B-transgenic plants under salt stress. As shown in Supplementary Figure S4, the expression levels of AtRD29A, AtCOR47, AtRD20, AtGSTF6, and AtRAB18 were enhanced in TabZIP14-B transgenic plants compared with WT plants. The promoter regions of these up-regulated genes were obtained using the Arabidopsis genome database. Our results showed that these genes also contained the bZIP ACGT-element within their promoter regions, suggesting that TabZIP14-B may confer stress tolerance by regulating various stress-responsive genes.