For the evolutionary analysis of the TaSnRK2.8 promoter sequences, 10 wheat species were used: hexaploid wheat (Triticum aestivum, AABBDD) cultivars “Hanxuan 10” and “Chinese Spring”; tetraploid wheat (AABB) PS9 (Triticum persicum) and DM 50 (Triticum dicoccum); diploid wheat Triticum urartu (AA) UR201 and UR204; diploid wheat Aegilops speltoides (BB) Y2001 and Y2041; and diploid wheat Aegilops tauschii (DD) Y85 and Ae46. Leaves were harvested for DNA isolation.

To isolate TaSnRK2.8 promoter, its genomic sequence (Zhang et al., 2013) was used as a query to screen the Wheat Genome Browser¹. The approximately 3000-bp upstream sequence of the translation start site was considered to be the putative TaSnRK2.8 promoter region. It was polymerase chain reaction (PCR)-amplified using the primers listed in supplementary Supplementary Table S1.

The SnRK2.8 promoter sequences from wheat (MF351624), rice (LOC_Os03g55600), and maize (AC206916.1) were aligned for element prediction using MegAlign in the DNAStar software, and then predicted using the PlantCARE software². To gain insight into the evolution and origin of the TaSnRK2.8 promoter sequences, the hexaploid wheat varieties and their relatives were used to construct a phylogenetic tree using MEGA 7.0.

The transcriptional start site was determined by 5′ rapid amplification of cDNA ends (RACE) using the GeneRacer Kit (Invitrogen, China). Total RNA was isolated using TRIzol reagent (Sangon, China). Total RNA (3.5 μg) was used as a template for 5′-RACE. The gene-specific primers used for the nested PCR are listed in Supplementary Table S1. The nested PCR products were cloned into the pGEM-T-Easy vectors and sequenced.

Successive 5′ truncations of the TaSnRK2.8 promoter were PCR-amplified using the primers shown in Supplementary Table S1. Nine promoter fragments (-2631 to ATG, Dp2947; -1481 to ATG, Dp1797; -821 to ATG, Dp1137; -774 to ATG, Dp1090; -693 to ATG, Dp1009; -568 to ATG, Dp884; -463 to ATG, Dp779; -408 to ATG, Dp724; Dp724 carrying the 125-bp fragment from position -693 to -568, named as Δ125) were separately fused with the GUS reporter gene in pCAMBIA1391, yielding the P_TaSnRK2:GUS constructs (Figure 1). These expression constructs and the pCAMBIA1391 vector (CaMV35S:GUS) were transformed into Agrobacterium, and then transferred into the wild-type Arabidopsis plants by floral infiltration. The transgenic plants were initially screened on hygromycin plates and subsequently characterized by PCR (Supplementary Table S1) and GUS (beta-glucuronidase) staining. The T₃-generation plants were selected for subsequent functional analyses.

Ten similarly sized 10-day-old seedlings from each transgenic line were used for total RNA isolation using the TRIzol reagent (Sangon) and then treated with RNase-free DNase (Takara, China). The quantitative real-time PCR (qRT-PCR) assays were performed in triplicate with the SYBR Premix Ex Taq (Takara, Shiga, Japan) using an ABI PRISM 7000 system (Applied Biosystems, Foster City, CA, United States). The relative transcript levels of GUS gene were determined using the 2^-ΔΔCT method (Livak and Schmittgen, 2001). The actin transcript was used to quantify the expression level.

Total genomic DNA was extracted from 10 similarly sized 10-day-old seedlings collected from each transgenic line using the Qiagen DNeasy Mini Kit (Qiagen, Valencia, CA, United States). Approximately 10 μg DNA was digested with EcoRI, segregated in 0.8% agarose gels, and then blotted onto nylon membranes (Hybond N+; Amersham Biosciences, Little Chalfont, United Kingdom). A 993-bp GUS construct labeled with digoxigenin (DIG) was used as a probe. Probe hybridization was performed using the AlkPhos Direct Labeling and Detection System (Amersham Biosciences) according to the manufacturer’s protocol.

Seeds were germinated on Murashige and Skoog (MS) medium solidified with 1.0% agar, and cultured in a growth chamber (25°C, 14 h light/10 h dark cycle, and 150 μMM^-2 S^-1 light intensity). After 7 days, the seedlings were planted on MS medium containing 150 mM mannitol, 150 mM NaCl, or 50 μM ABA, which were shown to constitute a significant stress in our pilot experiments. In cold stress treatment, seedlings were planted on MS medium and cultured in a cold growth chamber (4°C). Control plants were grown under non-stress conditions. After 3 days, twenty 10-day-old seedlings of each transgenic line were collected and subjected to GUS activity assays. All experiments were performed in triplicate.

Histochemical and fluorometric analyses of the GUS activity were performed as described previously (Jefferson, 1987). Seedlings were incubated in GUS reaction buffer at 37°C for 24 h, and then rinsed in an ethanol series to remove chlorophyll. Tissue observations were performed using a Leica MZ12 binocular microscope and images were captured using a Leica DC300 camera (Leica Microsystems, Wetzlar, Germany).

For the fluorometric assay, plants were homogenized in GUS extraction buffer. The homogenate was centrifuged at 12,000 × g for 10 min at 4°C, and then the supernatant activity was detected via an assay buffer containing 1 mM 4-methylumbelliferyl-β-glucuronide (Sigma, United States) at 37°C. The reaction was terminated by the addition of 200 mM Na₂CO₃ to a final concentration of 180 mM. Fluorometric estimations were performed using a PerkinElmer LS45 spectrofluorimeter (PerkinElmer).

The Student’s t-test, variance analysis, and Chi-square test, were performed using the statistical software SAS 9.2 to determine significant differences (P < 0.05) between the means.

The putative transcription start site was determined by two independent 5′-RACE experiments, indicating that the nucleotide position 313 bp upstream of the ATG start codon was the putative transcription start site of TaSnRK2.8 (Supplementary Figure S1).

The possible cis-elements were predicted by the multiple alignments of the SnRK2.8 promoter sequences from wheat, rice, and maize, and then determined using the PlantCARE software. As shown in Table 1, 11 predicted elements, including one ABA response element (ABRE), two binding sites for MYB transcription factors, and one low-temperature response element (LTRE), were identified in the TaSnRK2.8 promoter.

The hexaploid wheat T. aestivum originated from hybridization between the diploid goat grass (Ae. tauschii, DD) and the tetraploid emmer wheat (Triticum turgidum, AABB) (Dubcovsky and Dvorak, 2007). To gain insight into the origin of the TaSnRK2.8 promoter sequences, the hexaploid wheat varieties and their relatives were used to construct a phylogenetic tree. Consistent with the genomic sequence analysis of TaSnRK2.8 (Zhang et al., 2010), these TaSnRK2.8 promoter sequences were divided into three distinct groups. As shown in Figure 2, two genotypes (P_TaSnRK2.8-A and P_TaSnRK2.8-B) were present in all hexaploid and tetraploid wheats: P_TaSnRK2.8-A and P_TaSnRK2.8-B originated from the A and B diploids, respectively. No TaSnRK2.8 promoters were detected in the D diploids. Moreover, no genetic diversity of TaSnRK2.8 promoters existed in the B and D genomes. The TaSnRK2.8 promoter in B genome was further examined through the GUS activity assays.

To characterize the function of the TaSnRK2.8 promoter, a series of 5′ promoter deletions were performed and four derivatives of TaSnRK2.8 promoter fragments were fused to the GUS (beta-glucuronidase) reporter gene (Figure 1). Hygromycin segregation in the T₁ and T₂ seeds was analyzed through the Chi-square tests. The T₃ homozygous lines were used for performing the GUS activities assays. The integration of DNA in transformed plants was identified via PCR (Figure 3A). The qRT-PCR performed to determine the transcript levels of GUS under non-stress conditions showed that different promoter constructs have different abilities to regulate gene expression, and no GUS transcription was detected in the Dp724 and Dp1137 plants (Figure 3B). Southern blotting was used for selecting two independent transgenic lines with a single gene copy, which were further used in the GUS activity assays (Figure 3C). Expression patterns of GUS in various tissues showed that GUS was strongly expressed in all tissues of 35S:GUS seedlings, weakly in Dp2947 seedlings, and marginally in seedling stalks of Dp1797 plants under non-stress conditions. For Dp2947 plants, the transcript levels of GUS in stalks and roots were higher than those in leaves (Figure 3D).

Under non-stress conditions, the CaMV35S plants displayed the highest GUS expression intensity in all tissues, whereas only a slight GUS expression was observed in the controls of the TaSnRK2.8 promoters (Dp2947). Moreover, different expression patterns and activities were observed among the deletion segments of the TaSnRK2.8 promoter(Figure 4).

No GUS activity was observed in the Dp724 and Dp1137 plants, and only a faint expression was detected in the stalks of the Dp1797 seedlings under non-stress conditions. Although Dp2947 induced GUS expression in all tissues, weak GUS activity was observed in the leaves. These results indicated that the 660-bp segment (from position -1481 to -821) might harbor the stalk-specific expression elements under non-stress conditions, and that some enhancers were conserved in the region from position -2631 to -1481.

Under ABA treatment, the GUS expression was detected in Dp1797, but not in Dp724 and Dp1137 plants. When the promoter length was increased to position -2631, stronger GUS activities were detected in the leaves and roots of the Dp2947 seedlings (Figure 4). This finding suggested that the region from position -821 to -2631 might harbor some ABREs.

In the abiotic stress assays, no GUS activity was observed in the Dp724 (-408) seedlings. When the promoter length was increased to position -821 (Dp1137), a weak GUS expression was observed in the leaves and stalks in response to drought and salt stresses (Figure 4), suggesting the likely presence of osmotic stress-response elements in the region from position -821 to -408. The Dp1797 (-1481) seedlings showed relatively higher GUS expression levels in the leaves and stalks under various stresses, whereas no GUS activity was observed in the roots. Under stress treatments, the longest promoter, Dp2947, conferred the highest levels of GUS expression in all tissues, predominantly in the leaves. The results indicated that the fragment from position -1481 to -821 contained positive regulatory elements that enhanced gene expression in the leaves and stalks. Furthermore, the TaSnRK2.8 promoter (Dp2947) conferred the highest gene expression, suggesting the likely presence of positive regulatory motifs in the region from position -2631 to -1481.

The promoter region from position -821 to -408 was used for further deletion analysis (Figure 2). The integration of DNA in the transformed plants was identified via PCR (Figure 5A). Under drought and salt stresses, no GUS activity was observed in the Dp779 and Dp884 seedlings. Increasing the length of the promoter to position -693 (Dp1009) or -774 (Dp1090) led to the detection of GUS activity in the leaves and stalks in response to osmotic stress in a pattern similar to the Dp1137 seedlings (Figures 5B,C). This finding suggested that some osmotic stress-response elements might exist in the 125-bp region from position -639 to -568. To further verify the osmotic stress response element, the 125-bp region was fused to Dp724 through an overlapping PCR, which cannot induce the GUS transcription under non-stress and stress conditions. It was found that the promoter fragment Dp724 carrying the 125-bp region could induce the expression of GUS in the leaves and stalks in response to osmotic stress, very similar to what was observed in the Dp1009, Dp1090, and Dp1137 seedlings (Figures 5B,C). Collectively, these results suggest that the 125-bp TaSnRK2.8 promoter region from position -693 to -568 was required for inducing an osmotic stress response in Arabidopsis plants. The effects of different TaSnRK2.8 promoter deletion segments in response to ABA and abiotic stresses were shown in Supplementary Figure S2.

Previous studies found that TaSnRK2.8 (subclass III) was strongly expressed in the roots, and faintly in the stems and leaves. The over-expression of TaSnRK2.8 in Arabidopsis enhanced its tolerance to abiotic stresses through different stress-signaling routes (Zhang et al., 2010, 2011, 2016). To more thoroughly investigate the signaling and regulation of TaSnRK2.8 under different stresses, we characterized the spatial expression pattern of different TaSnRK2.8 promoter derivatives under ABA and environmental stress conditions.

Gene expression patterns might be a direct indication of a promoter’s function in stress and developmental processes. In Arabidopsis, the GUS gene driven by AtSnRK2.7 promoter was specifically expressed in the roots, whereas the AtSnRK2.8 promoter-mediated GUS activity was observed in all tissues (Umezawa et al., 2004; Mizoguchi et al., 2010). Furthermore, AtSnRK2.6 was confirmed to play a key role in stomatal closure under stress conditions (Mustilli et al., 2002). These data indicate that the SnRK2 family members play different roles in different tissues. The promoter deletion analyses of TaSnRK2.8 indicated that different promoter fragments harbored different elements for tissue-preferential expression. In general, the region from position -1481 to -821 contained the stalk-specific elements, whereas the region from position -2631 to -1481 contained the leaf- and root-specific elements. The spatial variation in expression levels was closely related to the different physiological stages of the plant lines and to the duration of stress treatment.

ABA is a vital hormone involved in plant abiotic stress responses. Plants can respond to stresses via ABA-dependent or -independent pathways. As a stress hormone, ABA acts through the regulatory pathways that control gene expression in response to stress. Increasing evidence shows that the subclass III SnRK2s respond strongly to ABA, the subclass II SnRK2s are only faintly induced by ABA, and the subclass I SnRK2s are moderately induced by ABA (Kobayashi et al., 2004, 2005; Boudsocq et al., 2007; Huai et al., 2008; Fujita et al., 2009). In the present study, the ability of TaSnRK2.8 promoter to induce GUS expression in Arabidopsis revealed that TaSnRK2.8 could drive transgene expression in response to ABA, and the promoter region from position -821 to -2631 might contain ABREs.

The identification of tissue-specific or inducible promoters in expression analysis would be of considerable practical value because it could eliminate the unnecessary burdens on plants by restricting the genetic expression to specific tissues or in response to specific environmental conditions (Hsieh et al., 2002; Zavallo et al., 2010; Hou et al., 2016). Here, we examined the activation of GUS under the regulation of TaSnRK2.8 promoter fragments in response to abiotic stresses and showed that a typical stress-inducible promoter contains some regions for regulating tissue specificity and transcription levels under stress, leading to different regulation patterns. For example, there might be osmotic stress response elements in the region from position -693 to -568 and positive regulatory motifs from -2631 to -1481, whereas the fragment from -1481 to -821 contained the leaf- and stalk-specific enhancers. Further analysis might enable the identification of essential control elements in these functional regions.

The comparative functional analysis of different stress-responsive promoters has provided useful information, leading to enhancements in the specificity and inducibility of stress responsive genes. Therefore, the TaSnRK2.8 promoter examined in the present study might prove beneficial for the overexpression of foreign genes in the plants subjected to abiotic stresses. Moreover, the findings of the deletion analysis reported here provide the basis for identifying the direct upstream regulators of SnRKs and stress-response elements in TaSnRK2.8 promoter.