Pre-harvest sprouting (PHS) is the germination of grain prior to ripening in the spike when there is excessive moisture before harvest. PHS has become a recurring worldwide problem since it causes a severe reduction in crop yield and flour quality due to starch and protein degradation (Olaerts et al., Reference Olaerts, Roye, Derde, Sinnaeve, Meza, Bodson and Courtin2016). Seed dormancy accounts for up to 60% of the variation in PHS tolerance, and PHS in wheat is mainly caused by the lack of adequate seed dormancy (DePauw and McCaig, Reference DePauw and McCaig1991; Li et al., Reference Li, Ni, Francki, Hunter, Zhang, Schibeci, Li, Tarr, Wang, Cakir, Yu, Bellgard, Lance and Appels2004). The level of wheat grain dormancy partly depends on abscisic acid (ABA) sensitivity before and after the grain reaches physiological maturity (Gubler et al., Reference Gubler, Millar and Jacobsen2005; Sun et al., Reference Sun, Zhang and Xiao2005; Shu et al., Reference Shu, Liu, Xie and He2016). One well-characterized positive regulator of ABA signalling, ABSCISIC ACID-INSENSITIVE 4 (ABI4), was initially identified in screens for mutants exhibiting ABA-resistant germination (Finkelstein, Reference Finkelstein1994). ABI4 is a member of the APETALA 2 (AP2/ERF) transcription factor family and can activate or repress gene expression by binding to specific cis-elements in gene promoters via its AP2 DNA-binding domain (Wind et al., Reference Wind, Peviani, Snel, Hanson and Smeekens2013). It has been documented that ABI4 interacts with target genes to regulate seed dormancy and germination. For example, ABI4-dependent temporal regulation of PTR2 expression influences water status during seed germination, promoting the germination of imbibed grain (Choi et al., Reference Choi, Kim, Song, Choi, Cho, Park, Kang and Park2020). ABI4 is indispensable for inhibiting the expression of three family members of Arabidopsis response regulators (ARRs), namely, ARR6/7/15, which are involved in seed dormancy (Huang et al., Reference Huang, Xiaoyan, Zhizhong, Shuhua and Yiting2017). Moreover, ABI4 is a primary positive regulator of ABI4, ABI5 and starch branching enzyme 2.2 (SBE2.2), activating transcription by binding the CACCG-box (CE1-like) in the promoter regions during seed development (Bossi et al., Reference Bossi, Cordoba, Dupré, Mendoza, Román and León2009). Apart from ABI4 itself, various transcription factors may regulate ABI4 transcription, including several WRKY transcription factors that can bind to the W-box sequence in the ABI4 promoter region (Shang et al., Reference Shang, Yan, Liu, Cao, Mei, Xin, Wu, Wang, Du, Jiang, Zhang, Zhao, Sun, Liu, Yu and Zhang2010; Antoni et al., Reference Antoni, Rodriguez, Gonzalez-Guzman, Pizzio and Rodriguez2011; Liu et al., Reference Liu, Yan, Wu, Mei, Lu, Yu, Liang, Zhang, Wang and Zhang2012). MYELOBLASTOSIS 96 (MYB96) induces ABI4 expression by binding to its promoter during seed germination and seedling development (Lee and Seo, Reference Lee and Seo2015).
In addition to ABI4, two other transcription factors (ABI3/VP1 and ABI5) have been characterized that regulate ABA response during seed development (Finkelstein, Reference Finkelstein1994; Finkelstein and Lynch, Reference Finkelstein and Lynch2000; Osa et al., Reference Osa, Kato, Mori, Shindo, Torada and Miura2003). It has been reported that some cross-regulation of expression existed among ABI3, ABI4 and ABI5, which function in a combinatorial network, rather than a regulatory hierarchy, controlling seed development and ABA response (Soderman et al., Reference Soderman, Brocard, Lynch and Finkelstein2000). Moreover, ABI3, ABI4 and ABI5 have similar effects on seed dormancy and the expression of maturation-specific seed proteins (Finkelstein, Reference Finkelstein1994). However, ABI4 is a focal point in the signal transduction pathways of ABA (Niu et al., Reference Niu, Helentjaris and Bate2002). Orthologues of ABI4 have been reported in many other plant species, including maize, rice and lotus (Niu et al., Reference Niu, Helentjaris and Bate2002; Ming et al., Reference Ming, Vanburen, Liu, Mei and Shen-Miller2013; Wang et al., Reference Wang, Liu, Mao, Li, Lu, Wang, Liu, Wei and Zheng2015). In maize, ZmABI4 is seed specific, reaching maximum expression at 20 days post-anthesis (DPA) (Niu et al., Reference Niu, Helentjaris and Bate2002). In the rice database, a single sequence shares significant homology with the AtABI4 AP2 domain, indicating that a single ABI4 homologue exists in rice (Yu et al., Reference Yu, Hu, Wang, Wong, Li, Liu, Deng, Dai, Zhou, Zhang, Cao, Liu, Sun, Tang, Chen, Huang, Lin, Ye, Tong, Cong, Geng, Han, Li, Li, Hu, Huang, Li, Li, Liu, Li, Liu, Qi, Liu, Li, Li, Wang, Lu, Wu, Zhu, Ni, Han, Dong, Ren, Feng, Cui, Li, Wang, Xu, Zhai, Xu, Zhang, He, Zhang, Xu, Zhang, Zheng, Dong, Zeng, Tao, Ye, Tan, Ren, Chen, He, Liu, Tian, Tian, Xia, Bao, Li, Gao, Cao, Wang, Zhao, Li, Chen, Wang, Zhang, Hu, Wang, Liu, Yang, Zhang, Xiong, Li, Mao, Zhou, Zhu, Chen, Hao, Zheng, Chen, Guo, Li, Liu, Tao, Wang, Zhu, Yuan and Yang2002). However, there is limited information available for ABI4 orthologues in wheat.
Synthetic hexaploid wheat SHW-L1 obtained from the hybridization of Triticum turgidum and Aegilops tauschii is a useful genetic resource and shows significant tolerance to PHS (Yang et al., Reference Yang, Liu, Pu, Zhang, Yuan and Chen2014). To investigate the regulatory factors that interact with TaABI4 and the role of TaABI4 in the ABA-induced seed dormancy pathway, we performed a conservation analysis on ABI4 in wheat ancestral species and modern cultivars and subsequently cloned this gene. We analysed the expression pattern of TaABI4 at different grain developmental stages. Furthermore, we carried out expression QTL analysis (eQTL) to detect regions regulating the expression of TaABI4 in recombinant inbred lines (RILs). Finally, candidate genes were also predicted and evaluated in the eQTL interval, providing further insight into the role of TaABI4 in ABA signal transduction pathways and into the regulatory framework that controls seed germination in wheat.
Materials and methods
Chuanmai32 (CM32, PHS susceptible), synthetic hexaploid wheat (SHW-L1, PHS-resistant) and their derived RILs (138 lines) were grown under glasshouse conditions (16 house light at 22°C, 8 h dark at 12°C, 70% relative humidity). Days to flowering was measured for each spikelet based on the anther extrusion at 50% of the spike. Developing grains from 5 to 30 DPA were collected at 5-d intervals from the centre florets for subsequent gene expression profiling. Young leaves of SHW-L1 and CM32 were used for DNA extraction. Each sample had biological replicates and was immediately frozen into liquid nitrogen and stored at −80°C for RNA extraction.
Sequence characterization and in silico promoter analysis
Based on the results of BLASTP searches, we obtained coding sequences of TaABI4 in Chinese Spring using EnsemblPlants (http://plants.ensembl.org/index.html). Protein domains of genes were predicted using the SMART tool (http://smart.embl-heidelberg.de/). The coding sequences of TaABI4 were used to query the target database (ViroBLAST, http://126.96.36.199/blast/viroblast.php, The Wheat ‘Pan Genome’, http://www.10wheatgenomes.com/data-repository/, and The Aegilops tauschii genome, http://aegilops.wheat.ucdavis.edu/ATGSP/data.php) to download homologous genes and 2 kb upstream sequences from translational initiation codon in 17 wheat cultivars and 3 wheat ancestors (Altschul et al., Reference Altschul, Madden, Schäffer, Zhang, Zhang, Miller and Lipman1997; Luo et al., Reference Luo, Gu, Puiu, Wang, Twardziok, Deal, Huo, Zhu, Wang, Wang, McGuire, Liu, Long, Ramasamy, Rodriguez, Van, Yuan, Wang, Xia, Xiao, Anderson, Ouyang, Liang, Zimin, Pertea, Qi, Bennetzen, Dai, Dawson, Müller, Kugler, Rivarola-Duarte, Spannagl, Mayer, Lu, Bevan, Leroy, Li, You, Sun, Liu, Lyons, Wicker, Salzberg, Devos and Dvořák2017; Ling et al., Reference Ling, Ma, Shi, Liu, Dong, Sun, Cao, Gao, Zheng, Li, Yu, Du, Qi, Li, Lu, Yu, Cui, Wang, Chen, Wu, Zhao, Zhang, Li, Zhou, Zhang, Hu, van Eijk, Tang, Witsenboer, Zhao, Li, Zhang, Wang and Liang2018; Zhu et al., Reference Zhu, Wang, Rodriguez, Deal, Avni, Distelfeld, McGuire, Dvorak and Luo2019; supplementary Table S1).
Amino-acid sequences were aligned using DNAMAN (Version. 5.2.10, Lynnon Biosoft, Quebec, Canada). Putative cis-acting regulatory elements located in the promoter regions were predicted using PLANTCARE (Lescot et al., Reference Lescot, Déhais, Thijs, Marchal, Moreau, Van de Peer, Rouzé and Rombauts2002; http://bioinformatics.psb.ugent.be/webtools/plantcare/html/) and PLACE (Higo et al., Reference Higo, Ugawa, Iwamoto and Korenaga1999; http://www.dna.affrc.go.jp/PLACE/). An analysis of conserved motifs in 17 wheat cultivars was obtained using the MEME suite (Bailey et al., Reference Bailey, Boden, Buske, Frith, Grant, Clementi, Ren, Li and Noble2009; http://meme-suite.org/tools/meme). This program was used to search for the top five cis-motifs with consensus patterns of 6–50 base width and E-value < 0.01, on the forward strand of the input sequences only.
Prediction of proteins and PEST motifs
Generated coding sequences were translated to predicted proteins using DNAMAN with default parameters. Searches for potential PEST sequences were performed using the ePESTfind (http://www.bioinformatics.nl/cgi-bin/emboss/epestfind). We used the input parameters in all cases and defined that a score above zero denoted a possible PEST sequence (Gregorio et al., Reference Gregorio, Hernandez-Bernal, Cordoba and Leon2014).
According to the TaABI4 nucleotide sequences of Chinese Spring, specific primers for the gene were designed online (https://www.ncbi.nlm.nih.gov/tools/primer-blast/) and shown in supplementary Table S2. Genomic DNA was isolated from SHW-L1/CM32 young leaves using the CTAB method (Zhang et al., Reference Zhang, Wang, Pan and Peng2013) and was used as templates to amplify the DNA sequences of TaABI4. PCR was performed using high-fidelity Prime STAR Polymerase (TaKaRa, Dalian, China) under the following conditions: 98°C for 3 min, 35 cycles of 98°C for 50 s, 60–65°C for 50 s and 72°C for 90 s, followed by a final extension step of 72°C for 10 min. The PCR amplification products were ligated into the pEASY-blunt Cloning Vector (TransGen, Beijing, China), and the resulting ligation mixtures were transformed into E. coli Trans1-T1 chemically competent cells (TransGen, Beijing, China) to obtain positive clones for sequencing.
RNA extraction and expression analysis
Primer pairs in the relevant conserved exon regions of TaABI4 among A, B and D genomes in SHW-L1 and CM32 were used to amplify 151 bp amplicons (supplementary Table S2). The expression level of TaABI4 was measured in the parents at six seed development stages (5, 10, 15, 20, 25 and 30 DPA). RNA was extracted from each sample using the total RNA extraction kit (Biofit, China), and genomic DNA was removed with DNaseI.
Three seed-developing stages (10, 20 and 30 DAP) of SHW-L1/CM32 were selected to carry out RNA sequencing (RNAseq). RNA quantity and quality were assessed using a NanoDrop 2000 spectrophotometer (Thermo Scientific, Fremont CA, USA) and checked for integrity on an Agilent 2100 bioanalyser (Agilent Technologies, Palo alto CA, USA) by denaturing agarose gel electrophoresis with ethidium bromide staining. Equimolar amounts of the libraries were constructed and sequenced by BerryGenomics (Beijing, China) using the Illumina HiSeq-2000 and HiSeq X Ten platform (Illumina, Hayward CA, USA). Gene transcript levels were estimated using transcripts per million (TPM; Zhao et al., Reference Zhao, Ye and Stanton2020).
First-strand cDNA was synthesized using a PrimeScriptTM 1st Strand cDNA Synthesis Kit (TaKaRa, Dalian, China). cDNA sampling was performed in duplicate and SsoFast™ EvaGreen® Supermix (Bio-Rad, Hercules CA, USA) was used for real-time quantitative PCR (RT-qPCR) (CFX96 Touch™ Real-Time PCR Detection System, Bio-Rad, USA). Each reaction contained approximately 50 ng first-strand cDNA, 0.5 μl, 10 μmol l−1 gene-specific primers and 10 μl real-time PCR SYBR Green (TIANGEN, Beijing, China). Amplification conditions were 5 min at 95°C, followed by 40 cycles of 30 s at 95°C, 30 s at 60°C, 40 s at 72°C and a final extension of 10 min at 72°C. Seven 1/10 dilutions of the recombinant plasmid cDNA template were used to make a standard curve for amplification efficiency (E) calculation. Three housekeeping genes, TaActin, Ta.14126.1 and Ta.7894.3.a1_at, were used as internal controls (Long et al., Reference Long, Wang, Ouellet, Rocheleau, Wei, Pu, Jiang, Lan and Zheng2010). Gene expression data were analysed using the Bio-Rad CFX Manager (Bio-Rad, Hercules CA, USA) software. The expression profile of the target gene was normalized to that of the internal control genes, and the geometric mean was calculated. The relative gene expression quantity of each sample was calculated using the E −ΔΔCt method (Pfaffl, Reference Pfaffl2001).
Expression QTL (eQTL) analysis
In order to characterize regions that regulate TaABI4 expression levels, we conducted eQTL mapping analysis within the RIL population using the previously constructed high-density genetic map (Yang et al., Reference Yang, Xu, Chen, Li, Xia, Chen, Liu and Luo2019). eQTL analysis was achieved using the WinQTLcart2.5 software (North Carolina State University, Raleigh, NC, USA) with the composite interval mapping (CIM) method (Wang and Basten, Reference Wang and Basten2007). The analysis was done by setting the control parameters to model 6 (standard model), forward regression, 10-cm windows and five makers as the control. The threshold was set at 4.0 to detect eQTLs. The wheat reference genome ‘Chinese Spring’, IWGSC RefSeq v1.0 (International Wheat Genome Sequencing Consortium, 2018), was used to query marker positions using the blastn2.2.26+ package (Camacho et al., Reference Camacho, Coulouris, Avagyan, Ma, Papadopoulos, Bealer and Madden2009). On the basis of eQTL intervals, the gene annotation was conducted using Wheat Gmap (http://www.wheatgmap.org.cn/tools/gene/information/). Subsequent candidate genes were validated by querying all of the predictions against the Nr-NCBI (http://www.ncbi.nlm.nih.gov/) and EnsemblPlants (http://plants.ensembl.org). Genes relating to the ABA signalling pathway were compared and mapped to the genome reference sequence of Chinese Spring v1.0 to identify candidate genes that may underlie eQTL. The CORREL function in Excel was used to calculate the correlation coefficient between the expression pattern of TaABI4 and RNAseq data of candidate genes in SHW-L1 and CM32. The correlation coefficient was used to measure the strength of the relationship between two variables.
Sequence characterization of TaABI4
The DNA sequence of ABI4 (AT2G40220) from Arabidopsis was used as a query sequence to carry out BLAST searches in EnsemblPlants. Three homologues of TaABI4 were identified on the A, B and D sub-genomes of 18 wheat cultivars (TaABI4-1A, TaABI4-1B and TaABI4-1D). All the 50 TaABI4 sequences were found to be represented by a single exon. The coding sequences (CDS) of three homologues of Chinese Spring were conserved with 97.53% nucleotide identity. Compared with TaABI4-1A, TaABI4-1B and TaABI4-1D had two 3–6 bp deletions as well as 28 single-nucleotide polymorphisms (SNPs), 13 SNPs of which caused non-synonymous mutations (supplementary Fig. S1). The three homologue-encoded proteins with 260, 256 and 257 amino-acid residues, respectively. The proteins have highly conserved AP2 domains that were also found in previously annotated AtABI4 in Arabidopsis and ZmABI4 in maize (Zea mays) (Fig. 1). In addition, ten amino acids (KGGPENAKFR) were contiguous to the AP2 domain (designated as the AP2-associated motif). Additionally, a stretch of eight amino acids (LRPLLPRP) identified as the LRP motif was located nearby (Fig. 1). TaABI4 proteins revealed 100% identity in these common regions, while TaABI4-1A contained three additional amino acids, His171, Leu196 and Ala197 (Fig. 1). Putative proteins were predicted from Ae. tauschii, T. dicoccoides cv. Zavitan and T. urartu. The protein sequence of AetABI4 obtained from Ae. tauschii showed 100% identity with the TaABI4-1D sequence. TuABI4-1A obtained from T. urartu showed 99.23% amino acid identity with TaABI4-1A. TdABI4-1B of T. dicoccoides cv. Zavitan shared 98.08% identity with TaABI4-1B (Fig. 1).
ABI4 proteins and putative motif analysis in wheat cultivars
The AP2 domains, the AP2-associated motifs and the LRP motifs were conserved in 50 putative ABI4 homologous proteins in terms of their position and sequence identity (Fig. 2). Putative PEST degradation signals at the terminus of wheat ABI4 proteins with a positive probability value (>0) were detected using the PEST-find program (Rice et al., Reference Rice, Longden and Bleasby2000), which were in agreement with a previous report (Gregorio et al., Reference Gregorio, Hernandez-Bernal, Cordoba and Leon2014). This demonstrated that potential PEST sequences were detected in all of these proteins, with probability scores ranging from +0.44 (ABI4-1A) to +3.52 (ABI4-1B) (Table 1). For ABI4-1B and ABI4-1D proteins, one PEST sequence was detected at the C-terminal with a length of 60AA. For ABI4-1A proteins, a shorter PEST motif of 41 amino acids was detected, sharing 99.8% identity among the 17 cultivars in addition to another PEST sequence predicted at the N-terminus that was also identified in TaABI4-1B. Although some variant amino acids were detected in the proteins of each genome, as shown in grey boxes in Fig. 2, they did not locate in the region of crucial motifs. This demonstrates that the ABI4 proteins are conserved in their protein architecture, coinciding with their central role in wheat hormone signalling.
Potential cis-acting regulatory elements of ABI4 promoters in wheat ancestors
The presence of potential cis-regulatory elements in the upstream (≥2000 bp) region of TaABI4 homologues from wheat cultivar Chinese Spring was analysed. Eleven types of potential cis-acting regulatory elements were identified in the upstream region (Fig. 3). This region was also isolated from T. dicoccoides cv. Zavitan, T. urartu and Ae. tauschii. A putative TATA-box was detected 190 bp upstream of the start codon. A binding site (CE1-like motif, CACCGCCCC) was present immediately downstream from a putative W-box (TTGACY). In addition, RY elements with CATGCATG involved in seed-specific regulation were predicted. ABRE elements known to be involved in ABA response, with CACGTG core motif, were recognized nearby the 5′-termini. ARE elements with an AAACCA core motif that are essential for the anaerobic induction also existed in all ABI4 proteins. Additionally, conserved motifs such as CAAT-box, CAT-box and A-box were detected. One Myb and one Myc element, known to be involved in ABA signalling (Lin, Reference Lin2009), were predicted in the TaABI4-1D and AetABI4 promoter regions. The detected cis-acting regulatory elements were conserved among the wheat and its ancestral species.
The putative motif analysis of ABI4 genes in wheat cultivars
The top five motifs identified by this analysis were found in almost all of the ABI4 genes in wheat cultivars and were highly conserved in terms of number and position (Fig. 4). Although motif 2 did not exist in the A sub-genome of Kronos, it shared 99.4% identity among 50 upstream regions and was regarded as a novel cis-motif with no current description in the PLACE database (Table 2). As shown in Table 2, motif 1 with a W-box as its core element was also conserved in all sequences with 100% identity. Although there were some variable SNPs in motif 3, motif 4 and motif 5, they did not exist in the core region of each motif. Overall, putative motifs within the upstream of ABI4 genes were almost completely conserved in wheat cultivars.
Cloning and qRT-PCR analysis of TaABI4 in SHW-L/CM32 developing seeds
The TaABI4 sequences were cloned from SHW-L and CM32, which were highly conserved in these two cultivars (supplementary Fig. S2). According to RNAseq analysis, the expression level of TaABI4 in CM32 was higher than that in SHW-L1 at each detected stage (Fig. 5A). Then, RT-qPCR assays were performed using cDNA from five time points (5, 15, 20, 25 and 30 DPA) to detect the expression level variation of TaABI4 between SHW-L1 and CM32. During seed development, TaABI4 expression began as early as 10 DPA, increasing between 10 and 15 DPA as the transition from growth to storage phase of grain development (starting after 12 DPA) took place, and peaked at 20 DPA, with a decline in expression until 30 DPA. The expression of TaABI4 in CM32 was higher than that in SHW-L1 in most of the measured stages. Two significant differences in relative expression were detected at 15 DPA (5.07-fold) and 20 DPA (1.39-fold) (Fig. 5B).
The significant difference between CM32 and SHW-L1 in the expression levels of TaABI4 at 15 and 20 DPA enabled the detection of eQTLs. Based on the consensus genetic map and corresponding SNP marker positions, six significant eQTLs (P < 0.05, LOD > 4) were identified (Table 3 and Fig. 6). One eQTL detected on chromosome 2A at 15 DAP was designated as eQABI4.15DPA.2A.1, with LOD scores at 4.53. Two eQTL regions located on chromosome 2D designated as eQABI4.20DPA.2D.1 and eQABI4.20DPA.2D.2 were detected at 20 DAP, showing 9.63 and 6.38 LOD scores, respectively. eQABI4.20DPA.4A.1 and eQABI4.20DPA.4A.2 were located on chromosome 4D with negative alleles from SHW-L1. They explained 38.2 and 46.1% of the phenotypic variation, respectively. The physical mapping of 3B eQTL was designated as eQABI4.20DPA.3B.1, showing that the corresponding interval location was Chr.3B: 667902308-669428443. All identified eQTLs had negative additive effects, indicating that eQTLs that could reduce the expression of TaABI4 were derived from synthetic wheat SHW-L1. Genes involved in regulatory processes, including hormone response and biosynthesis, signal transduction, protein phosphorylation, starch and sucrose metabolism, were selected in those eQTL regions. Subsequently, genes expressed in seeds and ABA related were highlighted, resulting in five candidate genes being identified (Table 4). Further correlation coefficient analysis of gene expression was carried out (Table 4). Correlation coefficients between the expression of TaABI4 and candidate genes ranged from 0.61 to 0.9. TaABI4 expression positively correlated with the expression of TraesCS2A02G089400, TraesCS2A02G099400, TraesCS4A02G094300 and TraesCS4A02G114400, while it negatively correlated with the expression of TraesCS4A02G093600.
a Correlation coefficient (r) was calculated between the expression pattern of TaABI4 and RNAseq data of candidate genes in SHW-L1 and CM32. A correlation of −1.0 shows a perfect negative correlation, while a correlation of 1.0 shows a perfect positive correlation.
In this study, we presented the characterization of the wheat ABI4, a gene involved in ABA responsiveness during seed development and germination. TaABI4 proteins from three wheat sub-genomes were conserved with an AP2 domain required for nuclear localization (AP2-associated motif), as well as regions for transcriptional activation (LRP motif). The conserved domains are used as hallmarks to identify ABI4 orthologues in different species (Gregorio et al., Reference Gregorio, Hernandez-Bernal, Cordoba and Leon2014). Although the protein sequences for the three homologues had slight polymorphisms, the overall identity was high (96.9%). Our results suggested that the AP2 proteins presented in wheat are the orthologues of the Arabidopsis ABI4 and should be considered as TaABI4-1A, TaABI4-1B and TaABI4-1D.
Compared with the ABI4 from T. urartu, Ae. tauschii and T. dicoccoides, the amino-acid variation existed only in TaABI4-1A (Thr15/Leu15, Gly218/Arg218) and TaABI4-1B (Ser108/Pro108) and was not located in core regulatory regions (Fig. 1). This indicated that TaABI4 was highly conserved during the polyploidization and domestication processes of wheat. In Arabidopsis, the low accumulation of ABI4 resulted from both post-transcriptional and post-translational regulation (Finkelstein et al., Reference Finkelstein, Lynch, Reeves, Petitfils and Mostachetti2011). PEST sequences are degradation motifs that can affect protein stability (Gregorio et al., Reference Gregorio, Hernandez-Bernal, Cordoba and Leon2014) and are characterized by regions enriched in the amino-acid proline, glutamic acid, serine and threonine (Rogers et al., Reference Rogers, Wells and Rechsteiner1986). Based on the available pan-genome data, we analysed 50 putative ABI4 proteins from 18 wheat cultivars to predict potential PEST motifs. Most of the possible PEST sequences were located in the N-terminal region of the protein and were longer than AtABI4 (Table 1). These differences may cause divergence in post-translational mechanisms compared with Arabidopsis. In fact, ZmABI4 also has two PEST motifs located in the N-terminus and C-terminus, showing score values of +3.04 and +0.68, respectively (Gregorio et al., Reference Gregorio, Hernandez-Bernal, Cordoba and Leon2014).
The discovery of cis-acting regulatory elements in the promoter regions is essential to understanding the spatial and temporal expression patterns of ABI4 genes. The six cis-acting regulatory elements were conserved in terms of position and sequence identity (Fig. 3). TATA-box is regarded as the core promoter element, and transcription factors bind to TATA-proximal regions (W-box, CE-1 like) having been shown to regulate downstream gene transcription (Heins et al., Reference Heins, Frohberg and Gatz1992; Busk et al., Reference Busk, Jensen and Pagès1997; Phukan et al., Reference Phukan, Jeena and Shukla2016). Additionally, A-box and RY element are cis-acting regulatory elements, and CAT-box is related to meristem expression in Arabidopsis (Sakata et al., Reference Sakata, Nakamura, Taji, Tanaka and Quatrano2010). ABRE (ABA-responsive elements) motifs are known to participate in response to ABA (Sarkar and Lahiri, Reference Sarkar and Lahiri2013). TaABI4-1D contained two classical ABRE elements that are necessary to constitute an active ABA-responsive complex because a single ABRE is not sufficient to confer ABA responsiveness (Hobo et al., Reference Hobo, Asada, Kowyama and Hattori1999; Zhang et al., Reference Zhang, Zhao and Zhao2005; Ganguly et al., Reference Ganguly, Roychoudhury, Sarkar, Sengupta, Datta and Datta2011). The identification of conserved cis-acting regulatory elements in ABI4 promoters of wheat revealed that other transcription factors might regulate those homologues.
The expression pattern of TaABI4 was variable between modern wheat cultivar CM32 and synthetic wheat SHW-L1. It is noteworthy that TaABI4 showed a higher transcript accumulation in weakly dormant material (CM32) than in dormant material (SHW-L1) during most periods of seed development (Fig. 5). By contrast, seeds of the Arabidopsis abi4 mutant germinated significantly more quickly than wild type (Shu et al., Reference Shu, Zhang, Wang, Chen, Wu, Tang, Liu, Feng, Cao and Xie2013), indicating that the presence of functional ABI4 is important for resistance to PHS. The expressions of both TaABI3 and TaABI5 in SHW-L1 were significantly higher than those in CM32 (Zhou et al., Reference Zhou, Yang, Wang and Wang2016). These results are consistent with the corresponding research results in Arabidopsis and maize, finding that ABI3 and ABI5 are positive regulators of seed dormancy (McCarty et al., Reference McCarty, Hattori, Carson, Vasil, Lazar and Vasil1991; Hoecker et al., Reference Hoecker, Vasil and McCarty1995; Finkelstein and Lynch, Reference Finkelstein and Lynch2000). The gene expression patterns of TaABI3 and TaABI5 were similar to that of TaABI4 in the early and middle stages of seed development (5–15 DPA), signifying that TaABI4 associated the ABA biosynthetic pathway with TaABI3 and TaABI5 as found in Arabidopsis (Lopez-Molina et al., Reference Lopez-Molina, Mongrand, McLachlin, Chait and Chua2002). From these results, other regulatory factors interacting with TaABI4 are required to complete our understanding of the gene networks involving seed germination.
eQTL mapping is an efficient approach to identify genetic loci controlling complex crop traits (Chen et al., Reference Chen, Hackett, Niks, Hedley, Booth, Druka, Marcel, Vels, Bayer, Milne, Morris, Ramsay, Marshall, Cardle and Waugh2010; Motomura et al., Reference Motomura, Kobayashi, Iehisa and Takumi2013). In this study, we chose 15 and 20 DPA, which are the middle periods of seed development, to identify six significant eQTLs associated with TaABI4 expression variation on chromosomes 2A, 2D, 3B and 4A (Table 3), suggesting that the observed differences in TaABI4 expression in the RIL population were regulated in part by trans-acting factors (Doss et al., Reference Doss, Schadt, Drake and Lusis2005). Several previous studies mapped the major QTLs for seed dormancy and PHS tolerance to chromosomes 4A (Mares et al., Reference Mares, Mrva, Cheong, Williams, Watson, Storlie, Sutherland and Zou2005; Torada et al., Reference Torada, Ikeguchi and Koike2005; Chen et al., Reference Chen, Cai and Bai2008). In this study, two major eQTLs located on chromosome 4A accounted for 38.2 and 46.1% of the phenotypic variance. This result further confirmed that the chromosome 4A harbours QTL, and eQTL associated with the PHS resistance is important for wheat. The eQTL regions detected in this study may provide candidate genes that play potential roles in regulating PHS through effects on TaABI4 expression. Thus, eQTLs detected in this study suggested that unidentified genes or indirect regulation genes would affect TaABI4, which causes the different expression patterns of TaABI4 compared with Arabidopsis.
Five putative candidate genes were detected in eQTL internals, and the correlation between the expression of each candidate gene and the expression of TaABI4 was analysed according to the available RNAseq database of SHW-L1 and CM32. In this study, the eQTL for TaABI4 at 15 DPA was located close to TraesCS2A02G089400, the orthologues of Pyrabactin resistance 1-like (PYL) abscisic acid receptors 4 (PYL4), approximately 6.16 Mb (Fig. 7). TraesCS2A02G099400 was regarded as an orthologue of OsbZIP62, which was involved in ABA signalling pathways (Yang et al., Reference Yang, Xu, Chen, Li, Xia, Chen, Liu and Luo2019). In particular, TraesCS2A02G099400 displayed a high expressional correlation coefficient with TaABI4 (+0.90), indicating the involvement of this gene in up-regulating TaABI4 expression. Two genes located in the region of eQABI4.20DPA.4A.2 were related to protein phosphatase. As an orthologue of TraesCS4A02G094300, OsPP2C30 is involved in ABA signalling pathway during seed germination (Kim et al., Reference Kim, Hwang, Hong, Lee, Ahn, Yoon, Yoo, Lee, Lee and Kim2012). TraesCS4A02G114400 located in the internals of eQABI4.20DPA.4A.2 was the orthologue of OsPYL, which positively regulated the ABA response during the seed germination (Tian et al., Reference Tian, Wang, Li, Lv, Liu, Wang, Niu and Bu2015). Together, these results suggested that five candidate genes may have a regulatory relationship with TaABI4.
In this study, the characterization of TaABI4, including its conserved protein domains and cis-acting regulatory elements analysis, provides information on the critical nucleotide and amino-acid residues of this gene. Meanwhile, high conservation was found in the amino-acid sequences and promoter regions, but the different expression level of TaABI4 in two wheat cultivars drove us to identify regions linked to candidate genes that function upstream of TaABI4 transcripts. Six potential eQTL regions that may regulate the expression of TaABI4 were detected. Five potential upstream candidate genes that may influence the expression of TaABI4 were also detected. These results can be utilized for future TaABI4 studies on interactions with other transcription factors in response to ABA and the establishment of the co-expressed networks relating to seed germination, which will successfully boost the efficiency of wheat breeding with sufficient seed dormancy to prevent PHS.
To view supplementary material for this article, please visit: https://doi.org/10.1017/S0960258521000015.
We thank Dr Jizeng Jia from the Chinese Academy of Agricultural Science for providing the data of Aikang58.
This research was supported by the National Key Research and Development Program of China (2018YFE0112000 and 2017YFD0100900), the National Natural Science Foundation of China (31871609 and 91935303) and the Sichuan Science and Technology Support Project (2019YFN0141, 2020YFH0154, and 2021YFH0077).
Conflict of interest
The authors declare no conflicts of interest.