A novel NAC family transcription factor SPR suppresses seed storage protein synthesis in wheat.

The synthesis of seed storage protein (SSP) is mainly regulated at the transcriptional level. However, few transcriptional regulators of SSP synthesis have been characterized in common wheat (Triticum aestivum) owing to the complex genome. As the A genome donor of common wheat, Triticum urartu could be an elite model in wheat research considering its simple genome. Here, a novel NAC family transcription factor TuSPR from T. urartu was found preferentially expressed in developing endosperm during grain-filling stages. In common wheat transgenically overexpressing TuSPR, the content of total SSPs was reduced by c. 15.97% attributed to the transcription declines of SSP genes. Both in vitro and in vivo assays showed that TuSPR bound to the cis-element 5'-CANNTG-3' distributed in SSP gene promoters and suppressed the transcription. The homolog in common wheat TaSPR shared a conserved function with TuSPR on SSP synthesis suppression. The knock-down of TaSPR in common wheat resulted in 7.07%-20.34% increases in the total SSPs. Both TuSPR and TaSPR could be superior targets in genetic engineering to manipulate SSP content in wheat, and this work undoubtedly expands our knowledge of SSP gene regulation.

Introduction

<span class="Species">Common wheat (Triticum aestivum L., 2n = 6x = 42, AABBDD) is a staple food in the diet of the world’s population and contributes 20% energy and 22% proteins in the daily intake of people (Faris, 2014; Shewry, 2009). The ability of common wheat dough to be processed into kinds of foods, such as bread and noodles, is mainly attributed to its visco‐elasticity, which is determined by seed storage proteins (SSPs; Weegels et al., 1996). Generally, SSPs consist of two groups of polypeptides, glutenins and gliadins. Glutenins are divided into high‐molecular‐weight subunits (HMW‐GSs, 70–90 kDa) and low‐molecular‐weight subunits (LMW‐GSs, 20–45 kDa) according to their molecular weight (D’Ovidio and Masci, 2004), and gliadins are classified into α‐(α/β‐), γ‐ and ω‐types on the basis of the mobility in acid polyacrylamide gel electrophoresis (A‐PAGE; Payne, 1987). SSPs are encoded by multi‐gene families. In each of A, B and D sub‐genomes of common wheat, there are one x‐type and one y‐type HMW‐GS genes at the Glu‐1 loci, 3–8 LMW‐GS genes at the Glu‐3 loci, and dozens of gliadin genes at the Gli‐1 and Gli‐2 loci (Nieto‐Taladriz and Carrillo, 1996; Payne et al., 1981; Zhang et al., 2013).

The syntheses of <span class="Gene">SSPs are spatial‐temporally regulated at the transcriptional level by synergetic interactions between transcription factors (TFs) and cis‐elements distributed in SSP gene promoters (Rubio‐Somoza et al., 2006b, 2006a,2006b, 2006a; Verdier and Thompson, 2008; Xi and Zheng, 2011; Yamamoto et al., 2006). However, only a few TFs, most of which belong to bZIP, Dof and MYB families, have been proved to regulate SSP gene transcription in cereal crops. For example, SPA (bZIP), WPBF (Dof) and TaGAMyb (MYB) are SSP gene regulators in common wheat (Albani et al., 1997; Dong et al., 2007; Guo et al., 2015; Ravel et al., 2009; Ravel et al., 2006); O2 (bZIP), OHP1 (bZIP), OHP2 (bZIP), ZmbZIP22 (bZIP) and PBF (Dof) are regulators in maize (Li et al., 2018; Lohmer et al., 1991; Marzabal et al., 2008; Pysh et al., 1993; Schmidt et al., 1992; Vicente‐Carbajosa et al., 1997); RISBZ1 (bZIP), RPBF (Dof), OsMYB5 (MYB) are regulators in rice (Onodera et al., 2001; Suzuki et al., 1998; Yamamoto et al., 2006) and BLZ1 (bZIP), BLZ2 (bZIP), BPBF (Dof), HvGAMYB (MYB), HvMCB1 (MYB) and HvMYBS3 (MYB) are regulators in barley (Churin et al., 2003; Diaz et al., 2002; Mena et al., 2002; Mena et al., 1998; Oñate et al., 1999; Rubio‐Somoza et al., 2006b, 2006a,2006b, 2006a; Vicente‐Carbajosa et al., 1998). All the above TFs act as activators in SSP synthesis. Intriguingly, the homolog of the activator BLZ1 in barley, SHP was recently found to repress SSP synthesis in common wheat (Boudet et al., 2019). In addition to the bZIP, Dof and MYB TFs, barley Fusca3 and wheat TaFusca3 are also regulators of SSP genes (Moreno‐Risueno et al., 2008; Sun et al., 2017). Due to the complexity of the polyploid genome, all the regulators of SSP synthesis in common wheat were identified through homology‐based cloning and none of them was first characterized in common wheat.

<span class="Gene">NAC is one of the largest TF families in higher plants and comprises more than 100 members in both monocot and dicot species (Murozuka et al., 2018). In common wheat, NAC family TFs are involved in responses to abiotic and biotic stresses such as drought, salt tolerance and stripe rust (Feng et al., 2014; Huang et al., 2015; Mao et al., 2014; Mao et al., 2012; Tang et al., 2012; Wang et al., 2015; Xia et al., 2010a, 2010b,2010a, 2010b; Xue et al., 2006), as well as senescence, nutrient remobilization and grain‐filling duration (Alhabbar et al., 2018b; Uauy et al., 2006; Zhao et al., 2015). Recently, a NAC TF TaNAC2‐5A was found promoting nitrogen use efficiency and seed vigour in common wheat (He et al., 2015; Li et al., 2020), and two NAC TFs ZmNAC128 and ZmNAC130 were identified as activators of zeins in maize (Zhang et al., 2019).

As the A genome donor of <span class="Species">common wheat, T. urartu is diploid and has a high‐quality genome sequence (Ling et al., 2018; Ling et al., 2013). Gene homologs between T. urartu and common wheat always share a conserved function. For example, both TuGASR7 in T. urartu and the homolog TaGASR7 in common wheat control grain length (Ling et al., 2013). Moreover, TFs (1238) in T. urartu have been predicted (http://planttfdb.cbi.pku.edu.cn) at the genome scale, and SSP genes in T. urartu including HMW‐GS genes, LMW‐GS genes and gliadin genes have also been characterized using both gene prediction and traditionally PCR‐based cloning methods in our previous work (Luo et al., 2015; Zhang et al., 2015). Therefore, T. urartu could be a superior model to identify new transcriptional regulators of SSP synthesis in wheat.

In this study, a novel <span class="Gene">NAC family TF TuSPR from T. urartu was preferentially expressed in developing endosperm during grain‐filling stages. The SSP content in mature grains of TuSPR overexpression common wheat was reduced. The molecular mechanisms of TuSPR’s suppression of SSP synthesis were elucidated. Moreover, the function of the homolog in common wheat TaSPR was also investigated. Both TuSPR and TaSPR identified in this work undoubtedly expands our knowledge of the transcriptional regulation of SSP synthesis.

Results

The accumulation of SSPs decreases in common wheat transgenically overexpressing TuSPR

Using the T. urartu genome sequence (<span class="Species">Ling et al., 2018; Ling et al., 2013), 1238 TFs were predicted at the genome scale (http://planttfdb.cbi.pku.edu.cn). We assumed that TFs preferentially expressed in endosperms during grain filling were prone to be involved in the transcriptional regulation of major reserves, such as SSP and starch. Thus, we analysed the tissue‐specific expression of all the 1238 TFs in T. urartu using quantitative reverse transcription PCR (qRT‐PCR) analysis. Among these TFs, a NAC family TF TRIUR3_21467 had the highest expression level in developing endosperms (with the embryo, aleurone, seed coat and pericarp removed) while trace levels in other tissues including roots, stems and leaves in qRT‐PCR analysis (Figure 1a,b). The endosperm preferential expression indicated that TRIUR3_21467 might be involved in the regulation of reserve biosynthesis in endosperms. Considering its inhibition on SSP biosynthesis in subsequent research, TRIUR3_21467 was nominated as a ‘storage protein repressor (TuSPR)’.

Figure 1

The tissue‐specific expression of TuSPR in Triticum urartu and the overexpression of TuSPR in overexpression common wheat lines. (a) Morphological variations of developing seeds of T. urartu at 5, 10, 15, 20 and 25 days post‐anthesis (DPAs). Bar, 2 mm. (b) Tissue‐specific expression assay of TuSPR in T. urartu using qRT‐PCR analysis. Endosperms at 5, 10, 15, 20 and 25 DPAs are used. Root, stem and leaf are from plant at 15 DPA. Data are means ± SE (n = 3) of three independent biological replicates. (c) The overexpression of TuSPR calculated in RNA‐Seq analysis. (d) The overexpression of TuSPR in RNA‐Seq analysis is verified by qRT‐PCR analysis. Considering the similar phenotype, only the endosperms of overexpression lines planted in Zhaoxian were used in RNA‐Seq and qRT‐PCR analyses. Primers characterizing SNPs are used to distinguish TuSPR and TaSPR in the qRT‐PCR analysis. Data are means ± SE (n = 3) of three independent biological replicates. Asterisks indicate statistically significant differences at P < 0.05 (*) and P < 0.01 (**) between overexpression lines and WT by one‐way ANOVA in the SPSS program.

The tissue‐specific expression of TuSPR in <span class="Species">Triticum urartu and the overexpression of TuSPR in overexpression common wheat lines. (a) Morphological variations of developing seeds of T. urartu at 5, 10, 15, 20 and 25 days post‐anthesis (DPAs). Bar, 2 mm. (b) Tissue‐specific expression assay of TuSPR in T. urartu using qRT‐PCR analysis. Endosperms at 5, 10, 15, 20 and 25 DPAs are used. Root, stem and leaf are from plant at 15 DPA. Data are means ± SE (n = 3) of three independent biological replicates. (c) The overexpression of TuSPR calculated in RNA‐Seq analysis. (d) The overexpression of TuSPR in RNA‐Seq analysis is verified by qRT‐PCR analysis. Considering the similar phenotype, only the endosperms of overexpression lines planted in Zhaoxian were used in RNA‐Seq and qRT‐PCR analyses. Primers characterizing SNPs are used to distinguish TuSPR and TaSPR in the qRT‐PCR analysis. Data are means ± SE (n = 3) of three independent biological replicates. Asterisks indicate statistically significant differences at P < 0.05 (*) and P < 0.01 (**) between overexpression lines and WT by one‐way ANOVA in the SPSS program.

To further investigate its function, the overexpression construct of TuSPR (<span class="Chemical">Glu‐1Bx14) was transformed into the common wheat variety Kenong 199, considering that no robust genetic transformation system had been established in T. urartu. To drive the overexpression of TuSPR specifically in developing endosperms, the promoter of the HMW‐GS gene Glu‐1Bx14 (Geng et al., 2014) was used in Glu‐1Bx14. Three independent TuSPR overexpression lines OE89, OE94 and OE139 were obtained and the T3 generation of these lines were planted in Dishang and Zhaoxian. The following RNA‐Seq analysis showed the overexpression of TuSPR in all three lines, which was verified by qRT‐PCR analysis (Figure 1c,d). SSPs in mature grains of all three TuSPR overexpression lines were separated and quantified using the reverse‐phase high‐performance liquid chromatography (RP‐HPLC) analysis. Compared with the c. 14.40 μg/mg flour in WT, the total HMW‐GSs of all the three overexpression lines planted in both Dishang and Zhaoxian was decreased by 45.59%–67.73% (Figure 2a). Similarly, the total LMW‐GSs of all three overexpression lines in both Dishang and Zhaoxian had 26.71%–45.17% reductions compared with the c. 9.33 μg/mg flour in WT (Figure 2b). The total glutenins of the three overexpression lines in the two planting sites were decreased by 37.56%–58.48% compared with the c. 23.56 μg/mg flour in WT (Figure 2c), which was consistent with reductions in HMW‐GSs and LMW‐GSs. Compared with the c. 42.25 μg/mg flour in WT, the total gliadins of all three overexpression lines in the two planting sites had no obvious changes (Figure 2d). The total SSPs of all three overexpression lines in the two planting sites was decreased by 11.70%–20.92% compared with the c. 65.94 μg/mg flour in WT (Figure 2e). SDS‐PAGE analysis also showed that the bands of HMW‐GSs and LMW‐GSs from all three overexpression lines were less clear than those of WT (Figure S1a), and the bands of gliadins from both overexpression lines and WT were of the similar staining intensity (Figure S1b).

Figure 2

Content variations in SSPs of TuSPR overexpression lines in RP‐HPLC analysis. (a–e) The variations in the total HMW‐GSs, the total LMW‐GSs, the total glutenins, the total gliadins and the total SSPs of mature grains of TuSPR overexpression lines in the RP‐HPLC analysis, respectively, compared with those of wild type (WT), in both Dishang and Zhaoxian. Data are means ± SE (n = 3) of three independent biological replicates. Asterisks indicate statistically significant differences at P < 0.05 (*) and P < 0.01 (**) between overexpression lines and WT by one‐way ANOVA in the SPSS program.

Content variations in <span class="Gene">SSPs of TuSPR overexpression lines in RP‐HPLC analysis. (a–e) The variations in the total HMW‐GSs, the total LMW‐GSs, the total <span class="Chemical">glutenins, the total gliadins and the total SSPs of mature grains of TuSPR overexpression lines in the RP‐HPLC analysis, respectively, compared with those of wild type (WT), in both Dishang and Zhaoxian. Data are means ± SE (n = 3) of three independent biological replicates. Asterisks indicate statistically significant differences at P < 0.05 (*) and P < 0.01 (**) between overexpression lines and WT by one‐way ANOVA in the SPSS program.

Consistent with the reduction of total <span class="Gene">SSPs, the thousand kernel weight (TKW) of all three overexpression lines was decreased by 3.75%–8.18% compared with the c. 44.91 g in WT (Figure S2a). In both Dishang and Zhaoxian, the grain weight per plant of all three overexpression lines was reduced by 11.15%–13.92% compared with the 24.52 g in WT (Figure S2b). These data collectively demonstrated that in <span class="Species">common wheat transgenically overexpressing TuSPR, the accumulation of total SSPs was reduced with a concomitant reduction in total grain weight per plant. Intriguingly, the other agronomically important traits of TuSPR overexpression lines, such as plant height and spike number, had no obvious variations compared with WT (Figure S2c).

The transcription of SSP genes is inhibited in common wheat transgenically overexpressing TuSPR

To explore the mechanisms of <span class="Gene">SSP decline in TuSPR overexpression <span class="Species">common wheat, the transcriptomes of developing endosperms from all three overexpression lines OE89, OE94 and OE139 planted in Zhaoxian at 10, 15 and 20 DPAs were analysed by RNA‐Seq. The developing endosperms of WT at the same stages were used as negative controls. Approximately 39 804 868 clean reads were obtained in individual replicons of each sample, 63.63% of which were mapped to annotated gene‐coding regions, and 18.72% were uniquely mapped (Table S1). The expression level of transcripts was described by the fragments per kilobase of exon model per million reads mapped (FPKM).

Although dozens even hundreds of <span class="Gene">SSP genes exist in the genome of common wheat, only the ones predicted in the genome sequence and had an intact open reading frame (ORF) were considered in the FPKM analysis. At 10 DPA, the transcription levels of all the analysed glutenin and gliadin genes in all three overexpression lines were decreased with varying magnitudes compared with those of genes in WT according to the FPKMs (Figure 3a). In more detail, the transcription of HMW‐GS genes, including Glu‐1Ax1, Glu‐1Bx7, Glu‐1By9, Glu‐1Dx2 and Glu‐1Dy12, decreased by 50.55%–62.35%; the transcription of LMW‐GS genes including A3‐620, D3‐394, D3‐441 and D3‐578 had reductions of 46.88%–56.74% and the transcription of gliadin genes including TaGli‐α‐7, TaGli‐γ‐1 and TaGli‐ω‐2 declined by 16.79%‐47.40%. At 15 DPA, the transcription of all the analysed HMW‐GS and LMW‐GS genes was still reduced, but the reduction amplitudes were lower than those at 10 DPA (Figure 3b). However, the transcription of almost all the analysed gliadin genes had no obvious variations even 24.18%–62.55% increase in all overexpression lines at this stage. At 20 DPA, the transcription of analysed HMW‐GS and LMW‐GS genes continued to decline except for those of Glu‐1Dy12 in all three lines, Glu‐1By9 in OE89 and A3‐620 and D3‐394 in OE94, all of which had 20.36%–46.75% increases (Figure 3c). Nevertheless, the transcription of all analysed gliadin genes in all three overexpression lines was increased by 17.69%–92.50%. The transcription decreases of gliadin genes at 10 DPA might be compensated by the transcription increases at 15 and 20 DPAs, which resulted in the no reduction in the total gliadins in mature grains of overexpression common wheat (Figure 2d). qRT‐PCR was conducted to validate the results in the RNA‐Seq analysis and overall correlation was found between the two datasets (Figure S3). These dynamic transcriptomic data demonstrated that all the active SSP genes predicted in common wheat genome were reduced throughout the endosperm development in TuSPR overexpression common wheat, and the reduction strength was the highest at the early stage (10 DPA), then decreased with endosperm maturation, even the transcription of some SSP genes, especially gliadin genes, increased at later stages (15 and 20 DPAs).

Figure 3

The transcription changes of SSP genes in TuSPR overexpression common wheat endosperms during grain filling in RNA‐Seq analysis. (a–c) The transcription changes of SSP genes in TuSPR overexpression common wheat endosperms at 10, 15 and 20 DPAs, respectively, compared with those of WT. Considering the similar variations in the contents of SSPs between Dishang and Zhaoxian, only the endosperms of lines planted in Zhaoxian were used. Since dozens even hundreds of SSP genes exist in the genome of common wheat, only the ones predicted in the genome sequence and had an intact open reading frame (ORF) were considered in the FPKM analysis. Data are means ± SE (n = 3) of three independent biological replicates. Asterisks indicate statistically significant differences at P < 0.05 (*) and P < 0.01 (**) between overexpression lines and WT by one‐way ANOVA in the SPSS program.

The transcription changes of <span class="Gene">SSP genes in TuSPR overexpression common wheat endosperms during grain filling in RNA‐Seq analysis. (a–c) The transcription changes of SSP genes in TuSPR overexpression common wheat endosperms at 10, 15 and 20 DPAs, respectively, compared with those of WT. Considering the similar variations in the contents of SSPs between Dishang and Zhaoxian, only the endosperms of lines planted in Zhaoxian were used. Since dozens even hundreds of SSP genes exist in the genome of common wheat, only the ones predicted in the genome sequence and had an intact open reading frame (ORF) were considered in the FPKM analysis. Data are means ± SE (n = 3) of three independent biological replicates. Asterisks indicate statistically significant differences at P < 0.05 (*) and P < 0.01 (**) between overexpression lines and WT by one‐way ANOVA in the SPSS program.

TuSPR suppresses both the promoter activities and the transcription of SSP genes

As a candidate regulator of <span class="Gene">SSP genes, molecular characteristics of TuSPR were analysed. Based on the IWGSC BLAST search (https://urgi.versailles.inra.fr/blast_iwgsc/?dbgroup=wheat_iwgsc_refseq_v1_chromosomes&program=blastn), TuSPR was located on the long arm of group 2 chromosome (2AL). The BLASTP (https://blast.ncbi.nlm.nih.gov/Blast.cgi?PROGRAM=blastp&PAGE_TYPE=BlastSearch&LINK_LOC=blasthome) search showed that the fragment of 7–133 amino acids (AAs) in TuSPR was the no apical meristem (NAM) domain, which is the signature motif of NAC family TFs (Figure 4a). In the transrepression activity analysis, the relative LUC activity was significantly reduced when the reporter was co‐transformed with the TuSPR overexpression plasmid into Arabidopsis leaf protoplast cells, indicating that TuSPR was a transcriptional repressor (Figure 4b,c). The NAM domain (1–133 AA) and the region of 134–392 AA of TuSPR were further truncated in the transrepression activity analysis. Only the region of 134–392 AA had a decreased relative LUC activity, demonstrating that it was the domain of transrepression activity (Figure 4b,c). Fused with the green fluorescent protein (GFP), the GFP signal was detected in the nucleus of common wheat leaf protoplast cells (Figure 4d), demonstrating the nuclear localization of TuSPR.

Figure 4

The influences of TuSPR on the promoter activities and the transcription of SSP genes in both the in vitro and in vivo assays. (a) Graphical summary of TuSPR. Red indicates the no apical meristem (NAM) domain (7–133 amino acids) predicted in NCBI Protein BLAST (BLASTP). AA, amino acid. (b) Schematic diagrams of the effector and reporter constructs used in the transrepression activity analysis. 35Spro, the 35S promoter; ter, terminator; 5 × GAL4, five GAL4 binding domain; LUC, firefly luciferase; REN, Renilla luciferase. (c) The transrepression activity analysis of TuSPR. GAL4‐VP16 fusion protein is a positive control in which GAL4 would bind to the GAL4 binding domain and VP16 would burst the expression of LUC. GAL4‐BD is a negative control. In addition to the full‐length cDNA, the fragment of 1–133 AA (TuSPR1–133) containing the NAM domain and the fragment of 134–392 AA (TuSPR134–392) of TuSPR were used. (d) The subcellular localization of TuSPR in common wheat leaf protoplast cells. Ubipro is the maize ubiquitin (Ubi‐1) promoter. Bars, 5 μm. (e) The influence of TuSPR on the promoter activities of SSP genes in the dual‐luciferase reporter assay. Empty construct is used as control. The ratio of firefly LUC activity to Renilla LUC activity (LUC/REN) is used to reveal the trans‐activation ability of TuSPR. 35Spro is the 35S promoter. (f) The influence of TuSPR on the transcription of SSP genes in the endosperm transient overexpression analysis. The immature endosperms of common wheat at 15 DPA are used. Empty vector is used as the negative control. Ubipro is the maize ubiquitin promoter (Ubi‐1). Data are means ± SE (n = 3) of three independent biological replicates. Asterisks indicate statistically significant differences at P < 0.05 (*) and P < 0.01 (**) by one‐way ANOVA in the SPSS program.

The influences of TuSPR on the promoter activities and the transcription of <span class="Gene">SSP genes in both the in vitro and in vivo assays. (a) Graphical summary of TuSPR. Red indicates the no apical meristem (NAM) domain (7–133 amino acids) predicted in NCBI Protein BLAST (BLASTP). AA, amino acid. (b) Schematic diagrams of the effector and reporter constructs used in the transrepression activity analysis. 35Spro, the 35S promoter; ter, terminator; 5 × GAL4, five GAL4 binding domain; LUC, firefly luciferase; REN, Renilla luciferase. (c) The transrepression activity analysis of TuSPR. GAL4‐VP16 fusion protein is a positive control in which GAL4 would bind to the GAL4 binding domain and VP16 would burst the expression of LUC. GAL4‐BD is a negative control. In addition to the full‐length cDNA, the fragment of 1–133 AA (TuSPR1–133) containing the NAM domain and the fragment of 134–392 AA (TuSPR134–392) of TuSPR were used. (d) The subcellular localization of TuSPR in common wheat leaf protoplast cells. Ubipro is the maize ubiquitin (Ubi‐1) promoter. Bars, 5 μm. (e) The influence of TuSPR on the promoter activities of SSP genes in the dual‐luciferase reporter assay. Empty construct is used as control. The ratio of firefly LUC activity to Renilla LUC activity (LUC/REN) is used to reveal the trans‐activation ability of TuSPR. 35Spro is the 35S promoter. (f) The influence of TuSPR on the transcription of SSP genes in the endosperm transient overexpression analysis. The immature endosperms of common wheat at 15 DPA are used. Empty vector is used as the negative control. Ubipro is the maize ubiquitin promoter (Ubi‐1). Data are means ± SE (n = 3) of three independent biological replicates. Asterisks indicate statistically significant differences at P < 0.05 (*) and P < 0.01 (**) by one‐way ANOVA in the SPSS program.

To investigate the molecular mechanisms of the <span class="Gene">SSP gene transcription reduction in overexpression common wheat, the regulation of TuSPR on SSP gene promoters was analysed using the dual‐luciferase reporter assay. Compared with those of promoters co‐transformed with the empty plasmid, the relative LUC activity of each tested SSP gene (including HMW‐GS genes TuGlu‐1Ax and TuGlu‐1Ay, LMW‐GS genes TuA3‐520 and TuA3‐538a and gliadin genes Gli‐α‐8, Gli‐γ‐1 and Gli‐ω‐1) promoter co‐transformed with 35S was decreased by c. 374 times (Figures 4e and S4a). To investigate its suppression on SSP gene transcription, the overexpression plasmid (Ubi:TuSPR) of TuSPR was transformed into endosperms of common wheat at 15 DPA (Figure S4c). Compared with those in endosperms transformed with the empty plasmid, the transcription levels of each examined SSP gene in the endosperms, where TuSPR was transiently overexpressed, were repressed by 2‐ to 3‐fold in the qRT‐PCR analysis (Figure 4f). These data indicated that TuSPR hampered the promoter activities in vitro and inhibited the transcription of SSP genes in vivo, which was consistent with the transcription reduction of SSP genes in developing endosperms of overexpression common wheat.

TuSPR binds to SSP gene promoters both in vitro and in vivo

To explore TuSPR’s interactions with <span class="Gene">SSP genes, the candidate binding motifs of NAC family TFs in the SSP gene promoters were analysed. In the New PLACE (https://www.dna.affrc.go.jp/PLACE/?action=newplace), four to fifteen occurrences of the cis‐element 5′‐CANNTG‐3′, which was supposed to be bound by the NAC family TFs, were found distributed in the promoters (0~−2000 bp) of all the SSP genes predicted in the genomes of both T. urartu and common wheat (Figure S5). And one to four occurrences of the cis‐element 5′‐CANNTG‐3′ were located within the functional promoter, which was from 0 to −500 bp (Korkuc et al., 2014; Ravel et al., 2014). Considering the identical cis‐element in all SSP gene promoters, only the binding of TuSPR to the cis‐element 5′‐CATGTG‐3′ distributed at −933 bp in the promoter of the HMW‐GS gene TuGlu‐1Ay was tested using electrophoretic mobility shift assay (EMSA). In the EMSA analysis, a shifted band was observed when TuSPR‐GST was mixed with the biotin‐labelled probe, which was the promoter oligo containing 5′‐CATGTG‐3′ in the middle (Figure 5a,b). Added with the non‐labelled competitive probe, the shift was turning less clear. No shifted band was detected when TuSPR‐GST was mixed with the biotin‐labelled mutant probe in which the 5′‐CATGTG‐3′ was exchanged with 5′‐AAAAAA‐3′ (Figure 5b). When the biotin‐labelled probe was mixed with the GST tag, no shifted band was observed. The EMSA results indicated that TuSPR bound to the cis‐element 5′‐CATGTG‐3′ in vitro.

Figure 5

The interaction between TuSPR and SSP gene promoters in EMSA assay and ChIP‐PCR analysis. (a) Location of the probe containing the cis‐element 5′‐CATGTG‐3′ (−933 bp) in the promoter region of HMW‐GS gene TuGlu‐1Ay. (b) The binding of TuSPR to the probe containing the cis‐element 5′‐CATGTG‐3′ and the mutant in the EMSA analysis. (c) The binding of TuSPR to the promoter regions (0~−500 bp) of SSP genes containing the cis‐element 5′‐CANNTG‐3′ in the ChIP‐PCR assay. AB, the antibody. Only the promoters of genes which are active and could represent most of the sequence diversity in each the HMW‐GS, LMW‐GS and gliadin gene sub‐family are selected in the ChIP‐PCR. Data are means ± SE (n = 3) of three independent biological replicates. Asterisks indicate statistically significant differences at P < 0.05 (*) and P < 0.01 (**) by one‐way ANOVA in the SPSS program.

The interaction between TuSPR and <span class="Gene">SSP gene promoters in EMSA assay and ChIP‐PCR analysis. (a) Location of the probe containing the cis‐element 5′‐CATGTG‐3′ (−933 bp) in the promoter region of HMW‐GS gene TuGlu‐1Ay. (b) The binding of TuSPR to the probe containing the cis‐element 5′‐CATGTG‐3′ and the mutant in the EMSA analysis. (c) The binding of TuSPR to the promoter regions (0~−500 bp) of SSP genes containing the cis‐element 5′‐CANNTG‐3′ in the ChIP‐PCR assay. AB, the antibody. Only the promoters of genes which are active and could represent most of the sequence diversity in each the HMW‐GS, LMW‐GS and gliadin gene sub‐family are selected in the ChIP‐PCR. Data are means ± SE (n = 3) of three independent biological replicates. Asterisks indicate statistically significant differences at P < 0.05 (*) and P < 0.01 (**) by one‐way ANOVA in the SPSS program.

To verify the binding to the cis‐element 5′‐CANNTG‐3′ in vivo, chromatin immunoprecipitation (ChIP) assay was conducted using endosperms at 15 <span class="Chemical">DPA from TuSPR overexpression lines. The promoter regions containing the cis‐element 5′‐CANNTG‐3′ distributed within 0~−500 bp of HMW‐GS (Glu‐1Ax1, Glu‐1Bx7, Glu‐1By9, Glu‐1Dx2 and Glu‐1Dy12), LMW‐GS (A3‐620 and D3‐575) and gliadin genes (TaGli‐α‐6 and TaGli‐ω‐10), which were active and could represent most of the sequence diversity in each of the HMW‐GS, LMW‐GS and gliadin gene sub‐families in the phylogenetic analysis, respectively, were determined (Figure S6b). Primer pairs, whose last three nucleotides were based on single nucleotide polymorphisms (SNPs), were used to distinguish the promoter regions of SSP gene members sharing high sequence identities. In the ChIP‐PCR analysis, the enrichment of the promoter regions of all the above‐listed SSP genes in the experimental sample (TuSPR‐AB) was 2–32 times higher than that in the negative control (No‐AB; Figure 5c). Taken together, these data indicated that TuSPR bound to SSP gene promoters both in vitro and in vivo. Previously identified regulatory elements of SSP genes were always distributed within a region <500 bp upstream of the start codon (Forde et al., 1985; Halford et al., 1989; Hammond‐kosack et al., 1993; Muller and Knudsen, 1993). However, regulatory elements were also predicted recently in the regions much far from 500 bp upstream of the start codon of HMW‐GS and LMW‐GS genes (Juhász et al., 2011; Makai et al., 2015; Ravel et al., 2014). Whether the cis‐element 5′‐CANNTG‐3′ predicted here in the regions 500–2000 bp upstream of the start codon was involved in the regulation of SSP genes still needs more evidence.

TaSPR inhibits both the promoter activities and the transcription of SSP genes

The application of this gene in breeding would be broadened if the homolog of TuSPR in <span class="Species">common wheat TaSPR also shared a conserved function with TuSPR on SSP synthesis suppression. To resolve this question, TaSPR in Chinese Spring was identified through homology‐based cloning. In protein sequence alignment, the three sub‐genome copies of TaSPR, TaSPR‐A (GenBank accession No. MN022886), TaSPR‐B (GenBank accession No. MN022887) and TaSPR‐D (GenBank accession No. MN022888), shared 90.30%–94.70% identity with TuSPR, and they had the identical NAM domain (Figure S7). Like TuSPR, all TaSPR‐A, TaSPR‐B and TaSPR‐D were preferentially expressed in developing endosperms and had trace expressions in roots, stems and leaves during grain‐filling stages (Figure 6a). To investigate the transrepression activity of TaSPR on SSP genes, the dual‐luciferase reporter assay was performed. In the dual‐luciferase reporter assay, the relative LUC activity of each representative SSP gene promoter co‐transformed with each 35S, 35S and 35S was decreased by c. 169 times, compared with those of promoters co‐transformed with the empty plasmid (Figure 6b). To explore their regulation on SSP gene transcription, each TaSPR‐A, TaSPR‐B and TaSPR‐D was transiently overexpressed in endosperms of common wheat at 15 DPA. qRT‐PCR analyses showed that the transcription levels of all representative SSP genes were repressed by 2‐ to 12‐fold compared with those in endosperms transferred with the empty vector (Figure 6c). These data indicated that all three TaSPR copies suppressed the SSP gene promoter activities in vitro and inhibited the transcription of SSP genes in vivo.

Figure 6

The influences of TaSPR on the promoter activities and the transcription of SSP genes in both the in vitro and in vivo assays. (a) Tissue‐specific expression analysis of TaSPR in common wheat (Chinese Spring). Endosperms at 5, 10, 15, 20 and 25 DPAs are used. Root, stem and leaf are from plant at 15 DPA. (b) The influence of TaSPR on the promoter activities of SSP genes in the dual‐luciferase reporter assay. Empty construct is used as a negative control. The ratio of firefly LUC activity to Renilla LUC activity (LUC/REN) is used to reveal the trans‐activation ability of TaSPR. (c) The influence of TaSPR on the transcription of SSP genes in the endosperm transient overexpression analysis. The immature endosperms of common wheat at 15 DPA are used. Data are means ± SE (n = 3) of three independent biological replicates. Asterisks indicate statistically significant differences at P < 0.05 (*) and P < 0.01 (**) by one‐way ANOVA in the SPSS program.

The influences of TaSPR on the promoter activities and the transcription of <span class="Gene">SSP genes in both the in vitro and in vivo assays. (a) Tissue‐specific expression analysis of TaSPR in common wheat (Chinese Spring). Endosperms at 5, 10, 15, 20 and 25 DPAs are used. Root, stem and leaf are from plant at 15 DPA. (b) The influence of TaSPR on the promoter activities of SSP genes in the dual‐luciferase reporter assay. Empty construct is used as a negative control. The ratio of firefly LUC activity to Renilla LUC activity (LUC/REN) is used to reveal the trans‐activation ability of TaSPR. (c) The influence of TaSPR on the transcription of SSP genes in the endosperm transient overexpression analysis. The immature endosperms of common wheat at 15 DPA are used. Data are means ± SE (n = 3) of three independent biological replicates. Asterisks indicate statistically significant differences at P < 0.05 (*) and P < 0.01 (**) by one‐way ANOVA in the SPSS program.

The SSP accumulation and bread‐making quality are enhanced in TaSPR knock‐down common wheat

To investigate its suppression on <span class="Gene">SSP synthesis in common wheat, TaSPR in Kenong 199 was knocked down by RNA interference (RNAi). To determine the transcription changes of TaSPR copies and SSP genes during grain filling, the endosperms of all three obtained independent RNAi lines (RNAi#1, RNAi#2 and RNAi#3) at 15 DPA, which was at the middle of grain‐filling stage (6–24 DPA; Simmonds and O’Brien, 1981), were used for qRT‐PCR analysis. Considering the similar variations in SSP content, only the endosperms of the three TaSPR knock‐down common wheat lines (RNAi#1, RNAi#2 and RNAi#3) planted in Zhaoxian was used. The transcription levels of TaSPR in all three RNAi lines were c. 7 times lower than that in WT, indicating the successful knock‐down (Figure S8b). Compared with WT, the transcription levels of almost all representative SSP genes including HMW‐GS genes (Glu‐1Ax1, Glu‐1Bx7, Glu‐1Dx2 and Glu‐1Dy12), LMW‐GS genes (A3‐620 and D3‐575) and gliadin genes (TaGli‐α‐7 and TaGli‐ω‐10) were elevated by 22.92%–126.19% in RNAi lines (Figure S8c). Intriguingly, the transcription levels of glutenin genes (c. 69.40%) had higher increases that those of gliadin genes (c. 48.73%). The accumulation of SSPs in mature grains of RNAi lines was determined by RP‐HPLC analysis. In the RP‐HPLC analysis, the total HMW‐GSs of all three RNAi lines was increased significantly with amplitudes of 9.38%–20.64% in Dishang and of 35.41%–50.96% in Zhaoxian compared with that of WT (Figure 7a). Compared with that in WT, the total LMW‐GSs of all three RNAi lines was elevated by 11.28%–16.65% in Dishang and by 16.57%–30.13% in Zhaoxian (Figure 7b). Owing to the increases in both HMW‐GSs and LMW‐GSs, the total glutenins of all three RNAi lines was elevated with amplitudes of 7.07%–15.02% in Dishang and of 27.53%–41.79% in Zhaoxian, compared with that of WT (Figure 7c). Although increased in both the two planting sites, the total gliadins of all RNAi lines rose more (by 15.77%–29.01%) in Dishang than in Zhaoxian (by 2.80%–9.56%), which was in contrast with the elevation in the total glutenins (Figure 7d). In Zhaoxian, the higher increase in glutenin accumulation than that of gliadins was consistent with the higher elevations in the transcription of glutenin genes during grain filling. Attributed to the increases of both glutenins and gliadins, the total SSPs of all three RNAi lines was elevated by 10.89%–20.34% in Dishang and by 7.07%–17.72% in Zhaoxian compared with that of WT (Figure 7e). Altogether, transcription of SSP genes during endosperm development and accumulations of SSPs in mature grains of TaSPR knock‐down common wheat were elevated. Attractively, the other agronomically important traits of TaSPR knock‐down lines had no obvious variations compared with WT (Figure S9a).

Figure 7

The variations in SSP content and bread‐making quality of TaSPR knock‐down common wheat. (a–e) The variations in the total HMW‐GSs, the total LMW‐GSs, the total glutenins, the total gliadins and the total SSPs of mature grains of TaSPR knock‐down common wheat lines in the RP‐HPLC analysis, respectively, compared with those of wild type (WT), in both Dishang and Zhaoxian. (f–h) The variations in the HMW‐GS/LMW‐GS ratio, the Gliadin/Glutenin ratio and the SDS sedimentation volume of TaSPR knock‐down common wheat lines, respectively. Data are means ± SE (n = 3) of three independent biological replicates. Asterisks indicate statistically significant differences at P < 0.05 (*) and P < 0.01 (**) by one‐way ANOVA in the SPSS program.

The variations in <span class="Gene">SSP content and bread‐making quality of TaSPR knock‐down common wheat. (a–e) The variations in the total HMW‐GSs, the total LMW‐GSs, the total glutenins, the total gliadins and the total SSPs of mature grains of TaSPR knock‐down common wheat lines in the RP‐HPLC analysis, respectively, compared with those of wild type (WT), in both Dishang and Zhaoxian. (f–h) The variations in the HMW‐GS/LMW‐GS ratio, the Gliadin/Glutenin ratio and the SDS sedimentation volume of TaSPR knock‐down common wheat lines, respectively. Data are means ± SE (n = 3) of three independent biological replicates. Asterisks indicate statistically significant differences at P < 0.05 (*) and P < 0.01 (**) by one‐way ANOVA in the SPSS program.

The HMW‐GS/LMW‐GS ratio, Gliadin/<span class="Chemical">Glutenin (Gli/Glu) ratio and SDS sedimentation volume are pivotal indicators of bread‐making quality (Blackman and Gill, 1980; Blumenthal et al., 1994; Doekes and Wennekes, 1982; Gupta et al., 1992; MacRitchie, 1987; MacRitchie and Gupta, 1993; Pechanek et al., 1997). The HMW‐GS/LMW‐GS ratios, the Gli/Glu ratios and SDS sedimentation volumes of TaSPR knock‐down common wheat were calculated. In all the three TaSPR knock‐down lines planted in Zhaoxian, the HMW‐GS/LMW‐GS ratios were increased by 23.49% while the Gli/Glu ratios were decreased by 24.27%, compared with those in WT (Figure 7f,g). The HMW‐GS/LMW‐GS ratios and the Gli/Glu ratios of TaSPR knock‐down lines planted in Dishang had no obvious variations (Figure 7f,g). Attractively, the SDS sedimentation volumes of all TaSPR knock‐down lines planted in both the two sites were increased by 14.16% (Figure 7h). Considering the HMW‐GS/LMW‐GS ratio and SDS sedimentation volume were positively corrected with bread‐making quality while the Gli/Glu ratio was negatively correlated, TaSPR knock‐down wheat should have improved bread‐making quality. In concert with the increase in SSPs, the TKW and the grain weight per plant of TaSPR knock‐down common wheat were increased by 4.16% and 12.38%, respectively, in both the two planting sites (Figure S9b,c). Taken together, the suppression of TaSPR could improve both the bread‐making quality and the yield in common wheat.

Discussion

TuSPR regulates SSP genes directly

Considering the interactions with <span class="Gene">SSP gene promoters in the dual‐luciferase reporter assay (Figures 4e and 6b) and the binding to SSP gene promoters in both EMSA and ChIP‐PCR assays (Figure 5), TuSPR and TaSPR regulated SSP genes directly. A bZIP family TF SHP was recently found to repress SSP synthesis in common wheat (Boudet et al., 2019). However, SHP is not first identified in common wheat, but the homolog of BLZ1, which is an activator of SSP synthesis in barley. Similarly, all the well‐characterized SPA, WPBF and TaGAMyb in common wheat (Albani et al., 1997; Dong et al., 2007; Guo et al., 2015; Ravel et al., 2006) were identified through homology‐based cloning. To the best of our knowledge, TuSPR and TaSPR are transcriptional repressors of SSP synthesis first characterized in wheat.

In our qRT‐PCR analysis, TaSPR‐A, TaSPR‐B and TaSPR‐D had the lowest, medium and highest expression levels, respectively, in endosperms of <span class="Species">common wheat throughout grain filling. However, TaSPR‐A, TaSPR‐B and TaSPR‐D had medium, lowest and highest the activities (regulation strength on SSP genes), respectively, in both the dual‐luciferase reporter assay and the endosperm transient overexpression assay. Therefore, the activity of a single copy was not consistent with its expression level. Transcriptional regulators always form a complex in the regulation of SSP genes. For example, the well‐characterized regulators GAMYB, WBPF and SPA in wheat form a GAMYB‐WBPF‐SPA complex in the activation of SSP genes (Kawakatsu and Takaiwa, 2010). Yeast two‐hybrid assay was conducted to investigate the interactions between TuSPR and each of the three well‐characterized TFs in wheat. The yeast two‐hybrid assay showed that TuSPR interacted with none of the above three TFs, this might indicate that TuSPR and the GAMYB‐WBPF‐SPA complex worked independently in the regulation of SSP synthesis in wheat (Figure S10a,b).

TuSPR and TaSPR might not be involved in N regulation

<span class="Chemical">Nitrate (N) affects SSP accumulation in common wheat grain (Barneix, 2007; Zheng et al., 2018). Several NAC family TFs are involved in the regulation of N use in major cereal crops, such as NAM‐A1, NAM‐B1 and TaNAC2‐5A in common wheat and OsNAP in rice (Alhabbar et al., 2018b; He et al., 2015; Liang et al., 2014; Uauy et al., 2006). This raises the question of whether TuSPR and TaSPR regulate the N uptake, transportation or reduction, which in turn affects SSP accumulation. Distributed in different clades in the phylogenetic analysis, TuSPR and TaSPR had rather distant relationships with the previously identified N regulating NAC family TFs (Figure S11). N acquisition happens in root and N transport occurs mainly in leaf and stem (Tegeder and Masclaux‐Daubresse, 2018). However, during the grain‐filling stages, both TuSPR and TaSPR were preferentially expressed in developing endosperms and had trace expression in leaves, roots and stems (Figures 1b and 6a). Although all NAM‐A1, NAM‐B1 and OsNAP increase the grain protein accumulation, they accelerate leaf senescence by enhancing N allocation to grains (Alhabbar et al., 2018a; Liang et al., 2014; Uauy et al., 2006). Throughout both vegetative and reproductive growth, delayed leaf senescence was not observed in TuSPR overexpression common wheat with decreased SSP content (Figure S2c); and early leaf senescence was not detected in TaSPR knock‐down common wheat with increased SSP content (Figure S9a). In the RNA‐Seq analysis of TuSPR overexpression common wheat, the expression levels of the N transport and assimilation related genes, for example TaNRT2.1 (He et al., 2015), had no obvious changes in endosperms throughout grain‐filling stages compared with WT (Table S2). Altogether, these data indicated that TuSPR and TaSPR are unlikely to be involved in wheat N regulation. As transcription factors regulated SSP synthesis specifically in developing endosperms, the trace expression of TuSPR and TaSPR in leaves was intriguing. Considering the fact that no growth retardation was detected during leaf development and no phenotypic changes were observed in mature leaves when TaSPR was knocked down in common wheat (Figure S9a) and the exclusion of involvement in N regulation, the functions of TuSPR and TaSPR in wheat leaf development remain to be investigated.

Proteome rebalancing between glutenins and gliadins

Proteome rebalancing has long been aware in endosperms of cereal crops. In <span class="Species">maize, the loss of zeins could be compensated by the synthesis of non‐zein proteins, which contributed the stable seed protein content (Holding and Larkins, 2009; Schmidt et al., 2011; Wu and Messing, 2012b, 2014; Wu et al., 2012; Zhang et al., 2015). In addition to the rebalancing between SSPs and non‐SSPs, compensation between glutenins and gliadins is also observed in common wheat endosperms. When the accumulation of glutenins decreases in RNAi or deletion lines, the content of gliadins increases (Dumur et al., 2004; Galili et al., 1986; Yang et al., 2014). Conversely, the reduction of gliadins results in elevation of glutenins (Becker et al., 2012; Gil‐Humanes et al., 2012; Gil‐Humanes et al., 2010; Piston et al., 2011). In TaSPR knock‐down common wheat, the total glutenins was increased more in Zhaoxian (27.53%–41.79%) than in Dishang (7.07%–15.02%). However, the total gliadins was elevated more in Dishang (15.77%–29.01%) than in Zhaoxian (2.80%–9.56%; Figure 7c,d). These data indicated the rebalancing between glutenins and gliadins. Although gliadin genes were suppressed by TuSPR, the total gliadins in TuSPR overexpression common wheat had no obvious reductions, and this might be a compensation for the loss in the total glutenins. At 15 and 20 DPAs, the transcription of a portion of gliadin genes in TuSPR overexpression common wheat was enhanced by 24.18%–92.50%, which might compensate for the transcription reduction of gliadin genes at 10 DPA and resulted in the stable total gliadins in mature grains. Our RNA‐Seq data showed that TuSPR was still overexpressed at 15 and 20 DPAs and the transcription levels were even 2–3 times higher than that at 10 DPA (Figure 1c). Moreover, the expression of previously identified transcriptional activators of SSP genes in common wheat, such as SPA, WPBF and TaGAMyb (Albani et al., 1997; Dong et al., 2007; Guo et al., 2015; Ravel et al., 2006) did not rise at 15 and 20 DPAs (Table S2). Therefore, the driving force of the gliadin gene transcription increases at 15 and 20 DPAs was unclear. Both the total glutenins and the total gliadins of all three TaSPR knock‐down common wheat lines planted in Zhaoxian were lower than those in Dishang. However, both the total glutenins and the total gliadins of WT planted in the same plot with TaSPR knock‐down lines in Zhaoxian were also lower than those of WT planted in the same plot with TaSPR knock‐down lines in Dishang. Considering the similar variations in WT planted in the same plot, the different glutenin and gliadin amounts of TaSPR knock‐down lines between the two planting sites were less likely due to the plasmids or constructs used in the transformation and might be attributed to the environmental condition variations. Despite the compensation between glutenins and gliadins, the total SSPs was decreased in TuSPR overexpression common wheat and was increased in TaSPR knock‐down common wheat, indicating that TuSPR and TaSPR could be elite targets in genetic engineering to manipulate SSP content in wheat.

In conclusion, a novel <span class="Gene">NAC family TF TuSPR in T. urartu was preferentially expressed in developing endosperm. TuSPR bound to the promoters of SSP genes and suppressed the transcription of the latter in both in vivo and in vitro assays. In TuSPR overexpression common wheat, both the transcription of SSP genes and the accumulation of SSPs were hampered. The homolog TaSPR in common wheat also suppressed SSP synthesis. In TaSPR knock‐down common wheat, the content of total SSPs was elevated. Both TuSPR and TaSPR could be optimal targets for genetic engineering to manipulate SSP content in wheat.

Experimental procedures

Plant materials and growth conditions

The genome‐sequenced T. urartu accession PI 428198 and <span class="Species">common wheat (T. aestivum) variety Chinese Spring were planted in the experimental station (40°11′N, 116°42′E) of the Institute of Genetics and Developmental Biology of the Chinese Academy of Sciences (IGDB, CAS), Beijing, China, during the 2016–2017 standard growing season. The common wheat variety Kenong 199 was used to generate overexpression and RNA interference (knock‐down) lines. All the T3 generation of TuSPR overexpression lines (OE89, OE94 and OE139), T3 generation of TaSPR RNAi lines (RNAi#1, RNAi#2 and RNAi#3) and the wild type (WT) Kenong 199 were planted at two experimental stations located at Dishang (37°95′N, 114°73′E, Hebei, China) and Zhaoxian (37°45′N, 114°46′E, Hebei, China), respectively, during the 2017–2018 standard growing season. Located in the North China Plain, both Dishang and Zhaoxian are two of the main cultivation areas of Chinese winter wheat with a different climate. The cultivation practices in both Dishang and Zhaoxian were identical and were described in the work of Song et al. (2020). For each line, 45 seeds were sown in three 1.50‐m‐long rows, which were spaced 23.00 cm apart. The growth of these transgenic common wheats was under normal conditions to avoid the influence of flood and drought on the accumulation of SSPs.

Vector construction and plant transformation

In the overexpression construct (<span class="Chemical">Glu‐1Bx14), the promoter of the HMW‐GS gene Glu‐1Bx14 (GenBank accession No. AY367771; Geng et al., 2014) was used to drive the overexpression of TuSPR specifically in developing endosperms of transgenic common wheat. In the RNAi vector construction, the fragment (901–1091 bp) distributed in the fourth exon of TaSPR which was amplified from the common wheat variety Kenong 199 was used as the target sequence in the RNAi hairpin to build the Ubi RNAi vector (Figure S8a). Ubipro was the maize ubiquitin (Ubi‐1) promoter. The fourth intron of common wheat Wx (GenBank accession No. LC373577) gene was used as a spacer in the hairpin. Each of the above constructs was transformed into the immature embryos of Kenong 199 using the biolistic particle delivery system following the method of Wang et al. (2014). All the transgenic lines were obtained from independent transformation events. The primers used for vector constructions and transgenic common wheat screening are listed in Table S3.

RNA extraction and quantitative reverse transcription PCR analysis

Total RNA was extracted from wheat developing endosperms (with the embryo, aleurone, seed coat and pericarp removed), roots, stems and leaves using the TRNzol Universal Reagent (TIANGEN, Beijing, China). The first‐strand cDNA was synthesized from 2.00 μg of total RNA by the Fast Quant RT Kit (TIANGEN). qRT‐PCR was performed with the Ligntcycler 480 (Roche, Mannheim, Germany) using the SYBR Green I Master (Roche) kit. For the tissue‐specific analyses of both TuSPR and TaSPR, the relative expression of TuSPR or TaSPR was calculated using . As to the transcription changes of TuSPR in overexpression lines and TaSPR copies in knock‐down lines, the transcription variations of <span class="Gene">SSP genes in both overexpression and knock‐down lines and the transcription changes of <span class="Gene">SSP genes in the endosperm transient overexpression analyses, the normalized relative quantities (NRQs) relative to the reference gene were calculated using the formula (Schmittgen and Livak, 2008). Three reference genes, Ta4045 (Ubiquinol‐cytochrome C reductase iron‐sulphur subunit) (Paolacci et al., 2009), TaActin and TaTubulin, were used in parallel for the best results. And only the data calculated using Ta4045 were shown considering the similar results obtained from all the three reference genes. Primers used for the qRT‐PCR analysis are listed in Table S3.

Transrepression activity analysis of TuSPR

The transrepression activity analysis assay was performed according to the method of Ohta et al. (2001). In the transrepression activity analysis, the full‐length cDNA of TuSPR, the fragment of 1–133 AA (TuSPR1–133) and the fragment of 134–392 AA (TuSPR134–392) were inserted into the 35S vector to generate effector plasmids (Figure 4b). The vector 35S was used as a negative control and the vector (35S) containing the <span class="Gene">GAL4‐VP16 fusion protein expression cassette was used as the positive control (Chasman et al., 1989; Croston et al., 1992; Sadowski et al., 1988). The firefly luciferase (LUC) gene driven by the minimal TATA box of the 35S promoter plus five GAL4 binding elements in series was used as a reporter. The Renilla reniformis luciferase (REN) gene driven by the 35S promoter (35S) was used as an internal control. Each effector was co‐transformed in Arabidopsis leaf protoplasts with the reporter and the internal control, and the fluorescence was measured according to the protocol of Liao et al. (2008). The ratio of firefly LUC activity to R. reniformis luciferase activity was used to reveal the transrepression ability.

Subcellular localization assay

The full‐length cDNA of TuSPR was inserted into the pJIT163‐hGFP vector and was driven by the <span class="Species">maize ubiquitin (Ubi‐1) promoter. The construct was subsequently transformed into common wheat protoplasts following the methods of Shan et al. (2014). After 18 h of dark incubation in a chamber at 25 °C, GFP fluorescence of common wheat protoplasts was observed using a laser confocal microscope (Zeiss LSM 710 NLO, Oberkochen, Germany). The primers used for vector construction are listed in Table S3.

Measurement of the SSP content

Reverse‐phase high‐performance liquid chromatography (RP‐HPLC) was used to measure the content of <span class="Gene">SSPs including glutenins (HMW‐GSs and LMW‐GSs) and gliadins in mature grains of TuSPR overexpression lines and TaSPR knock‐down lines following the methods described elsewhere (Tilley et al., 1993; Yang et al., 2014). For each transgenic line, c. 20.00 g mature grains of four randomly selected plants were ground into whole‐wheat flour, and 45.00 mg whole‐wheat flour of each plant was used for glutenin and gliadin extraction. Both the glutenin and gliadin fractions were filtered through a 0.45 μm organic nylon filter and analysed by RP‐HPLC according to the methods described by González‐Torralba et al. (2011). The amounts of glutenins and gliadins were estimated by integrating the relevant peak areas in the chromatograms and 1.00 mg/mL BSA standard solution was used to quantify the gliadin and glutenin extracts. The content of these SSP components was represented by μg/mg flour. To verify the quantification in RP‐HPLC analysis, the same SSP extracts of transgenic lines were separated in SDS‐PAGE analysis following the methods of Luo et al. (2015).

Measurements of the agronomically important traits

For each TuSPR overexpression and TaSPR RNAi line in both Dishang and Zhaoxian, more than 15 independent plants were randomly selected to measure the agronomically important traits. The grain weight per plant and the thousand kernel weight (TKW) were measured manually.

RNA‐Seq analysis

Total RNAs extracted from endosperms (with the embryo, aleurone, seed coat and pericarp removed) of TuSPR overexpression <span class="Species">common wheat lines OE89, OE94, OE139 and WT at 10, 15, and 20 DPAs were used for RNA‐Seq analysis. Considering the similar variations in the contents of SSPs between Dishang and Zhaoxian, only the endosperms of lines planted in Zhaoxian were used. Each line and WT had three independent plants for RNA extraction. For individual samples, a library with an insert length of c. 350 bp was sequenced using the Illumina HiSeq 2000 (San Diego, CA). All clean reads were mapped to the Chinese Spring reference genome (TGACv1; Clavijo et al., 2017). To verify the transcript abundance calculated in RNA‐Seq, qRT‐PCR analyses of both TuSPR and SSP genes were performed using gene‐specific primers (Table S3).

Electrophoretic mobility shift assay

The full‐length cDNA of TuSPR was cloned into pGEX‐4T‐1 and transferred to <span class="Species">Escherichia coil <span class="CellLine">BL21 (Transseta) to obtain the fusion protein. The induction and purification of the GST (Glutathione S‐transferase) recombinant protein of TuSPR were following the methods of Qiao et al. (2016). EMSA was performed using the LightShift Chemiluminescent EMSA Kit (Thermo Fisher Scientific, Waltham, MA) according to the manufacturer’s instructions. The primers used for vector construction are listed in Table S3.

Chromatin immunoprecipitation assay

To investigate the binding of TuSPR to <span class="Gene">SSP gene promoters in vivo, ChIP assay was conducted using the developing endosperm at 15 DPA of TuSPR overexpression lines following the method of Bowler et al. (2004) with minor modifications. Considering the similar phenotype among overexpression lines, endosperms of OE89, OE94 and OE139 were mixed to generate experimental samples. A monoclonal antibody (TuSPR‐AB) was produced in the Abmart antibody company (Shanghai, China) using the peptide NPEGQHGPSSDLTT (263–276 AA of TuSPR). In the ChIP assay, the experimental sample was added with the antibody (TuSPR‐AB) to immunoprecipitate the protein‐DNA complex, and the sample without the antibody was used as a negative control (No‐AB). After reverse cross‐linking, the precipitated DNA was used for ChIP‐PCR analysis. Primer pairs (Table S3) amplifying a c. 100 bp fragment containing the cis‐element 5′‐CANNTG‐3′ in the promoter region (0~−500 bp) of each SSP gene were used in the ChIP‐PCR analysis.

The dual‐luciferase reporter assay

To investigate the interaction between TuSPR/TaSPR and <span class="Gene">SSP gene promoters, the dual‐luciferase reporter assay was performed following the methods of Ohta et al. (2001). Each full‐length cDNA of TuSPR and TaSPR (TaSPR‐A, TaSPR‐B and TaSPR‐D) was inserted downstream of the GAL4‐VP16 fusion protein expression cassette in the backbone to generate individual effector plasmids (Figure S4a,b). For a single reporter plasmid, each SSP gene promoter was inserted at the upstream of five GAL4 binding elements (5 × GAL4) in series, which formed a unit to drive the expression of the firefly luciferase (LUC) reporter gene. For individual promoters, the 2000 bp region upstream of the translation initiation codon (ATG), which was supposed to contain most of the key cis‐elements in SSP gene expression, were used (Juhász et al., 2011; Lamacchia et al., 2001; Ravel et al., 2014; Wang et al., 2020). Considering the large number of SSP gene in the T. urartu genome (https://plants.ensembl.org/Triticum_urartu/Info/Index) and the common wheat genome (http://plants.ensembl.org/Triticum_aestivum/Info/Index), only the promoters of genes that were active and could represent the sequence diversity in each of the HMW‐GS, LMW‐GS and gliadin gene sub‐families in the phylogenetic analysis were selected in the dual‐luciferase reporter assay (Figure S6). In T. urartu, the promoters of HMW‐GS genes TuGlu‐1Ax and TuGlu‐1Ay, LMW‐GS genes TuA3‐520 and TuA3‐538a and gliadin genes Gli‐α‐8, Gli‐γ‐1 and Gli‐ω‐1 were used. In common wheat, promoters of HMW‐GS genes Glu‐1By8, Glu‐1Dx2 and Glu‐1Dy 12, LMW‐GS genes A3‐620, D3‐575 and D3‐578 and gliadin genes TaGli‐α‐1, TaGli‐α‐7 and TaGli‐ω‐10 were selected.

The <span class="Gene">GAL4‐<span class="Gene">VP16 fusion protein would bind to the GAL4 binding elements and burst the expression of LUC, which could compensate for the weak activities of SSP gene promoters in Arabidopsis leaf protoplast cells. The R. reniformis luciferase (REN) gene driven by the 35S promoter (35S) was used as an internal control. The GAL4‐VP16 fusion protein without TuSPR or TaSPR insertion was used as control. In experimental samples, each reporter (SSP) was co‐transformed in Arabidopsis leaf protoplasts with the internal control (35S) and the effector (35S or 35S). In the control experiment, each reporter (SSP) was co‐transformed in Arabidopsis leaf protoplasts with the internal control (35S) and the empty effector (35S). The fluorescence was measured according to the protocol of Liao et al. (2008). The ratio of firefly LUC activity to Renilla LUC activity (LUC/REN) was used to reveal the trans‐activation abilities of TuSPR and TaSPR. Experiments were done with three biological replicates.

Transient overexpression of TuSPR and TaSPR in developing common wheat endosperms

TuSPR and the three sub‐genome copies of TaSPR (TaSPR‐A, TaSPR‐B and TaSPR‐D) were driven by the <span class="Species">maize ubiquitin promoter (Ubi‐1) to generate overexpression constructs (Ubi or Ubi; Figure S4c,d). Each overexpression construct was transferred into the developing endosperms of Chinese Spring at 15 DPA following the method of Wang et al. (2014). The empty overexpression vector (Ubi) was used as a negative control. After 48 h of dark incubation at 25 °C, the total RNA of these bombarded endosperms was extracted and the expression levels of SSP genes were quantified by qRT‐PCR analysis. Considering the numerous SSP gene members in the common wheat genome (https://plants.ensembl.org/Triticum_aestivum/Info/Index), only the genes (HMW‐GS genes Glu‐1Bx7 and Glu‐1Dx2, LMW‐GS genes A3‐620 and D3‐578 and gliadin genes TaGli‐α‐7, TaGli‐γ‐1 and TaGli‐ω‐10) that were active and could represent most of the sequence diversity in each of the HMW‐GS, LMW‐GS and gliadin gene sub‐families in the phylogenetic analysis were selected for qRT‐PCR analysis (Figure S6b).

Phylogenetic analysis

Protein sequences of the 488 <span class="Gene">NAC family TFs predicted in the <span class="Species">common wheat genome (Borrill et al., 2017; Guerin et al., 2019) were used in the phylogenetic analysis following the method of Luo et al. (2018).

Evaluation of the SDS sedimentation volume

The SDS sedimentation volume (SDS‐SV) of TaSPR RNAi lines and WT was measured following the method described earlier (Zeleny, 1947).

Statistical analysis

The one‐way ANOVA with the SPSS 17.0 package for Windows (SPSS, Inc., Chicago, IL) was used for statistical analysis in this work. The data were expressed as mean ± SE. Differences of P < 0.05 were considered statistically significant.

Conflict of interest

The authors declare that they have no conflict of interests.

Author contributions

A.Z., D.L., K.Z. and D.C. conceived and designed the study. L.S., G.L. and Y.S. performed the experiments. G.L., L.S., J.X., J.J., C.Z., E.G., W.Y., X.L. and J.S. analysed the data. G.L. prepared the manuscript, A.Z., D.L. and W.Y. improved the manuscript. K.Z. and D.C. contributed reagents and materials. A.Z. and D.L. provided guidance on the whole study. All authors read and approved the final manuscript.

Accession number

Accession numbers of genes used in this study are listed in Table S4.

Figure S1 Comparison of stained bands of HMW‐GSs, LMW‐GSs and gliadins in SDS‐PAGE analysis between TuSPR overexpression <span class="Species">common wheat and WT.

Figure S2 Yield variations in TuSPR overexpression <span class="Species">common wheat.

Figure S3 The verification of the transcription declines of <span class="Gene">SSP genes in developing endosperms of TuSPR overexpression <span class="Species">common wheat in the RNA‐Seq analysis using qRT‐PCR analysis.

Figure S4 Schematics of the constructs used in both the in vitro and in vivo assays of TuSPR and TaSPR.

Figure S5 Distributions of the cis‐element 5′‐CANNTG‐3′ in <span class="Gene">SSP gene promoters of Triticum urartu and common wheat.

Figure S6 The phylogenetic analysis of <span class="Gene">SSP genes in Triticum urartu and common wheat.

Figure S7 Protein sequence alignment between TuSPR and the three copies of TaSPR.

Figure S8 The transcription changes of <span class="Gene">SSP genes in developing endosperms of TaSPR knock‐down <span class="Species">common wheat in the qRT‐PCR analysis.

Figure S9 Yield variations in TaSPR knock‐down <span class="Species">common wheat.

Figure S10 The interactions between TuSPR and other well‐characterized <span class="Gene">SSP gene regulators in <span class="Species">yeast two‐hybrid analysis.

Figure S11 Phylogenetic analysis of TuSPR, TaSPR and all the <span class="Gene">NAC family TFs predicted in the <span class="Species">common wheat genome.

Click here for additional data file.

Table S1 Mapped reads of the RNA‐Seq data

Table S2 The FPKM variations of <span class="Gene">SSP genes and well‐characterized transcriptional activators of <span class="Gene">SSP synthesis in TuSPR overexpression lines

Table S3 Primers used in this study

Table S4 Accession numbers of genes used in this study

Click here for additional data file.