Background

Wheat (Triticum aestivum L., 2n = 6x = 42, AABBDD), as a complex allohexaploid species, is one of the most important crops widely cultivated across the world. Wheat grains contain about 10–15% proteins, and are one of the richest protein sources in the human diet. It is well known that wheat breadmaking quality is largely determined by the seed storage proteins present in the grain endosperm, which mainly consist of polymeric glutenins and monomeric gliadins [1],[2]. The polymeric glutenins are further subdivided into high-molecular weight glutenin subunits (HMW-GS) and low-molecular-weight glutenin subunits (LMW-GS) according to their mobilities on a sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) gel, which determine their dough elasticity, viscosity, and strength [2],[3].

LMW-GS can be separated into three groups, the B, C, and D subunits, based on their electrophoretic mobilities on an SDS-PAGE gel. Genetic analysis showed that these subunits are encoded by the Glu-A3, Glu-B3, and Glu-D3 loci on the short arms of the chromosomes 1A, 1B, and 1D, respectively [4],[5]. Some components were also found to be encoded by genes on the short arms of the group 6 and 7D chromosomes [6]. Based on their N-terminal amino acid sequences, LMW-GS are classified into three subclasses, LMW-m, LMW-s, and LMW-i types, according to the first amino acid residue of the mature protein: methionine, serine, and isoleucine, respectively [6]. The LMW-s type subunit seems to be predominant [7],[8]. Typically, the N-terminal amino acid sequence is SHIPGL- in LMW-s type subunits, while LMW-m type subunits have various N-terminal sequences such as METSHIGPL-, METSRIPGL-, and METSCIPGL- [9]-[11]. The LMW-i type subunit, first reported by Pitts et al. [12], lacks the N-terminal domain and starts directly with the repetitive region of ISQQQQ- after the signal peptide. Although the typical N-terminal domain is absent, LMW-i type subunits can be expressed normally, similar to LMW-m and LMW-s, in the wheat endosperm [13],[14]. Most LMW-GSs possess eight cysteine residues, although their positions vary in the different types of subunits, which plays important roles in the formation of intra- and inter-molecular disulfide bonds in the gluten macropolymer [14].

Compared to the Glu-1 loci encoding HMW-GS, Glu-3 loci exhibit more extensive allelic variations that are closely related to gluten quality. Early work by Gupta and Shepherd [15] identified and named six alleles at Glu-A3, nine alleles at Glu-B3, and five alleles at Glu-D3 loci in common wheat. Recently, 14 unique LMW-GS genes in the wheat cultivar **aoyan 54 were identified, four of which were located at Glu-A3, three at Glu-B3, and seven at Glu-D3, based on bacterial artificial chromosome (BAC) library screening and proteomics analysis [16]. The results from a set of Aroona LMW-GS near isogenic lines (NILs) showed that the Glu-A3 locus has two m-type and 2–4 i-type genes [17]. Analysis of the micro-core collections (MCC) of Chinese wheat germplasm identified more than 15 LMW-GS genes from individual MCC accessions, 4–6 of which were located at the Glu-A3 locus [18].

Since extensive allelic variations are present at Glu-3 loci, it is generally difficult to accurately determine the functional properties of individual alleles in different genotypes. To date, the main method used to investigate the effects of different Glu-3 alleles on dough quality has involved determination of their effects and ranks in NILs. Earlier research on the durum wheat NILs Lira 42 and Lira 45 showed that the LMW-2 type subunit in Lira 45 had significantly greater beneficial effects on gluten strength and breadmaking quality than the LMW-1 subunit in Lira 42 [19]. In bread wheat, Glu-A3d possesses three active LMW-GS genes and produces the highest Zeleny sedimentation value (ZSV) and Extensograph maximum resistance (Rmax) [17]. Other reports also showed that the Glu-A3d allele had a superior effect on dough strength [20]-[22]. Recent work on a set of Aroona NILs showed that Glu-A3b contributed to a longer midline peak time (MPT) and better raw white Chinese noodle (RWCN) color [23]. Despite the large number of studies performed on the functions of Glu-3 alleles, more comprehensive and in-depth analyses on the structures and functions of the individual alleles at Glu-3 loci are still lacking.

In the current work, we conducted the first comprehensive investigation on the molecular characteristics and functional properties of the LMW-GS allele Glu-A3a by using a Glu-A3 deletion line in the Chinese Spring (CS) wheat cultivar in combination with various proteomics and molecular biology approaches. Our results demonstrate that the deletion of Glu-A3a significantly reduces wheat dough strength and breadmaking quality. In addition, we demonstrated that Glu-A3a results in a longer repetitive domain and more α-helices in the encoded subunit, as well as a higher expression level and accumulation rate during grain development, which could help to improve the formation of a stronger dough structure and superior quality.

Results

Identification and characterization of seed proteins in CS and the Glu-A3deletion line CS-n

A Glu-3 deletion line of CS was screened and developed in our laboratory, and named CS-n. Compared to CS, the morphological characteristics of plants, spikes, and seeds, as well as the growth and development traits of CS-n showed no significant differences (Additional file 1: Figure S1, Additional file 2: Figure S2, and Additional file 3: Table S1). The grain protein compositions of CS and CS-n were identified by using various proteome approaches (Figure 1 and Additional file 4: Figure S3). The results indicated that CS-n showed the same albumin and globulin compositions as CS, while gliadins displayed minor differences between CS-n and CS; only one gliadin band obtained by acidic polyacrylamide gel electrophoresis (A-PAGE) was absent in CS-n (Additional file 4: Figure S3).

Figure 1
figure 1

Identification of Glu-A3a in Chinese Spring (CS) and Glu-A3 deletion line CS-n. a. SDS-PAGE: the Glu-A3a encoded B-subunit as well as LMW-GS and HMW-GS were indicated. b. 2-DE: two differentially expressed protein spots between CS and CS-n encoded by Glu-A3a were marked by ① and ②. c. RP-UPLC: two protein peaks encoded by Glu-A3a in CS as well as LMW-GS and HMW-GS were indicated.

Full size image

Glutenin subunits identified by SDS-PAGE indicated that HMW-GS in CS-n were the same as those in CS (N, 7 + 8, 2 + 12), and most LMW-GS bands were also identical, except that one clear B-type LMW-GS encoded by Glu-A3a was absent in CS-n (Figure 1a). Two-dimensional electrophoresis (2-DE) analysis revealed that Glu-A3a encodes two proteins (spots 1 and 2 in Figure 1b), which were further determined to be one LMW-i type subunit by liquid chromatography-tandem mass spectrometry (LC-MS/MS), as shown in Table 1. Reversed-phase ultra-performance liquid chromatography (RP-UPLC) analysis further confirmed that Glu-A3a encodes two protein components (peaks 1 and 2 in Figure 1c), which were eluted at 15.5 min and 16 min, respectively. Both peaks accounted for 22.58% of the total LMW-GS in CS.

Table 1 LC-MS/MS analyses of peptides obtained after tryptic digestion of the isolated spot and bands
Full size table

To obtain the accurate molecular mass of the Glu-A3a-encoded B-subunit, the expected protein band on the SDS-PAGE gel indicated in Figure 1a was collected and then analyzed by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF-MS). As shown in Additional file 5: Figure S4, the Glu-A3a-encoded LMW-GS B-subunit was easily identified, and its molecular mass was determined to be 41,701.2 Da.

Confirmation of Glu-A3adeletion in CS-n with a sequence-tagged site polymerase chain reaction (STS-PCR) marker

To further confirm the deletion of the Glu-A3 locus in CS-n, a pair of STS primers developed from the single nucleotide polymorphisms (SNPs) in Glu-A3 allelic variants [24] were used to amplify the Glu-A3a gene. As shown in Figure 2, one specific PCR product of 529 bp was amplified in CS, the CS-1Sl/1B substitution line, the CS-1Sl addition line, and Aroona, which contain the Glu-A3a allele, whereas no such fragments were obtained in the other materials without Glu-A3a, such as CS-n. The specific amplified 529-bp fragment was sequenced, and the sequence was the same as those from the upstream 140–395 bp of the Glu-A3a-coding sequence shown in Additional file 6: Figure S5. Thus, these results confirmed that the Glu-A3 locus was deleted in CS-n.

Figure 2
figure 2

Identification of Glu-A3a by STS-PCR markers. 1. CS ( Glu-A3a ), 2. CS-n; 3. CS-1Sl/1B; 4. CS 1Sl addition line; 5. Aroona-A3a (Glu-A3a); 6. Aroona-A3b (Glu-A3b); 7. Aroona (Glu-A3c); 8. Aroona-A3d (Glu-A3d); 9. Aroona-A3e (Glu-A3e); 10. Aroona-A3f (Glu-A3f); 11. Glenlea (Glu-A3g); 12. CB037A. M. molecular mass marker: 2000 bp, 1500 bp, 1000 bp and 500 bp. Glu-A3a fragment with 529 bp was arrowed.

Full size image

Comparison of gluten quality properties between CS-n and CS

Dough strength and breadmaking quality testing showed that the main gluten quality parameters in CS-n were significantly reduced compared to those of CS (Tables 2 and 3). In general, flour yield, water absorption, final viscosity, and peak viscosity between CS-n and CS showed no apparent differences. However, deletion of Glu-A3a in CS-n increased the ash content by 15.39%. Ash content is an important indicator of flour quality, which has a moderately negative effect on noodle color [25]. In addition, the deletion of Glu-A3a in CS-n resulted in a significant decrease of the gluten index (4% reduction) and an increase in the flour falling number (5.05% increase), as shown in Table 1. The gluten index was shown to have a positive relationship with strong dough property [26].

Table 2 Quality parameters of dough and bread slices in CS and CS-n
Full size table
Table 3 Comparison of C-cell parameters of bread slices between CS and CS-n
Full size table

Farinograph analysis indicated that development time, stability time, tolerance index, and farinograph quality number in CS-n were significantly lower than those in CS (Table 2). These properties led to a decrease in loaf volume of CS-n from 760 to 735 cm3 (Table 2 and Figure 3). Bread texture analysis showed that the hardness and resilience of bread in CS were superior to those in CS-n. Further cell size analysis of the bread demonstrated that the quality in CS-n was significantly reduced (Table 3). For example, wrapper length, slice brightness, and wall thickness of CS-n bread slices were much lower than those of CS. The cell diameter and elongation in CS-n were also reduced as a result of Glu-A3a deletion.

Figure 3
figure 3

The loaves baking pictures and C-cell pictures of CS, CS-n. (a) The loaves baking pictures of CS and CS-n. (b) the C-cell pictures of CS and CS-n.

Full size image

Molecular characteristics of the LMW-GS allele Glu-A3a

To further understand the molecular mechanisms underlying the significant effects of Glu-A3a on gluten and breadmaking quality, the complete coding sequence of Glu-A3a was amplified and sequenced by allelic-specific (AS) PCR. Based on the previously characterized Glu-A3 genes, a pair of specific primers (A3-F and A3-R) for the Glu-A3 locus was designed and used to amplify the Glu-A3a allele from CS. As shown in Additional file 7: Figure S6, a single band of approximately 1100 bp was obtained from CS, whereas no product was amplified from CS-n. Since most of the complete coding sequences of LMW-GS genes vary in length between 909 and 1167 bp [6],[27]-[29], the size of the amplified band corresponded well to the known LMW-GS gene sizes. After sequencing of the amplified product, a complete open reading frame of 1134 bp was obtained. Sequence alignment showed that the cloned gene had no internal stop codons and contained typical structural features of LMW-GS, and therefore was named as Glu-A3a (Additional file 7: Figure S6). After searching the GenBank database, we found that the cloned Glu-A3a gene had the same sequence as GluA3-11 from cultivar Aroona-A3a (GenBank accession number FJ549928). The deduced amino acid sequence of Glu-A3a showed the presence of an isoleucine as the first amino acid residue in the N-terminal of the mature protein, indicating that it belongs to the LMW-i type subunit [6].

The complete coding sequence of Glu-A3a was aligned with 15 other known LMW-i type genes to detect SNP and insertion/deletion (InDel) variations, and the results are listed in Table 4. These LMW-i genes originated from different Triticum species, including T. aestivum and T. dicoccoides. Six SNPs at different positions, resulting from G-A or C-T transitions and two deletions at nucleotides 81 and 854, were identified in Glu-A3a. Six SNPs could produce amino acid substitutions, and thus are considered nonsynonymous SNPs.

Table 4 The positions of SNPs and InDels identified between Glu-A3 and other LMW-i type gene*
Full size table

The deduced amino acid sequence of Glu-A3a had 376 amino acid residues with a predicted molecular mass of 41,346.1 Da, corresponding well to that determined by MALDI-TOF-MS (41,701.2 Da). Multiple alignment of the deduced amino acid sequences of Glu-A3a with the other 14 LMW-i type subunits (Figure 4) showed that all have conserved signal peptides and four domains in the mature protein sequences, including a repetitive domain, cysteine-rich region, glutamine-rich region, and C-terminal conservative region, as reported by Cassidy et al. [27]. Similar to other LMW-i type subunits, the Glu-A3a-encoded subunit contained eight cysteine residues at relatively conserved positions (Additional file 8: Table S2). It is speculated that the first and seventh cysteines of the LMW-GS form the inter-molecular disulfide bond, while the rest form three intra-molecular disulfide bonds [30],[31].

Figure 4
figure 4

Multiple alignment of the deduced amino acid sequences of Glu-A3a and other 14 LMW-i glutenin genes. These genes including GenBank number AB062877 [14], AY542896 [13], DQ307386 [32], EU189087 [33], EU594335 and EU594336 [34], FJ549929, FJ549931, FJ549932 and FJ449933 [24], FJ876819 (Han, 2009), GQ870245, GQ870249 [35] and GU942731 [36]. Signal represents signal peptide (I), repetitive domain (II) and three sub-regions of C-terminal domain were indicated, respectively. The first amino acid residue of the mature proteins and cysteine residues were highlighted by black box and red shading, respectively. Deletions were indicated by dashes. Polyglutamine stretches were indicated by broken line frames.

Full size image

The number of repeats present in the repetitive domain is mainly responsible for the length variation and the general hydrophilic character of LMW-GS [30]. The Glu-A3a-encoded subunit contained the typical repeat motif of LMW-GS: P1–2FP/SQ2–6. Our results showed that Glu-A3a has a rather large and regular repeated sequence domain that includes a high proportion of glutamine residues (about 46%) in the repeats (consensus sequence PPFSQQQQ), and two polyglutamine stretches with 11 and 12 continuous glutamine residues in the repetitive and C-terminal domains, respectively. Repeat motif numbers in LMW-i subunits are much higher than those in the LMW-m and LMW-s subunits, ranking them the longest protein subunits among all Glu-3 loci.

Secondary structure and function prediction of the Glu-A3a-encoded protein

The secondary structures of the Glu-A3a-encoded protein (FJ549928) and five other LMW-i type subunits from bread wheat (AY724436, AY724437, AY263369, AY831866, and AY542896) were predicted by the PSIPRED server, as shown in Table 5. The results showed that the α-helices and β-strands were dispersed in the normal configuration in C-terminal I and were highly conserved in C-terminal III. FJ549928 contained seven α-helices, mainly located at the C-terminal, and one β-strand dispersed in the conserved C-terminal region. Thus, the number of α-helices in FJ549928 was much higher than that of the other five subunits, which contain only 0–3 α-helices. For example, the LMW-i type glutenin subunit AY542896, assigned to the 1A chromosome, only has one α-helix, which was confirmed to co-migrate with the LMW-50 subunit that plays an important role in determining good quality characteristics of Glenlea [13] and the XYGluD3-LMWGS1 subunit (AY263369), with only 3 α-helices, is also considered to have a positive effect on dough quality [37].

Table 5 The secondary structure prediction of the six deduced LMW-GS
Full size table

Phylogenetic analysis of Glu-A3aand other LMW-GS genes

A homology tree was constructed to reveal the phylogenetic relationships among 25 LMW-GS genes at Glu-3 loci from different species and genomes through nucleotide sequence alignment of their coding regions using MEGA5 software (Figure 5). These sequences comprised 21 LMW-GS genes from different genomes of Triticum diploid, tetraploid, and hexaploid species. The phylogenetic tree displayed two clear branches, which corresponded well to distinguishing the LMW-i type from the LMW-m and LMW-s type subunits. This demonstrated that LMW-i type genes have undergone greater divergence during evolution compared to LMW-s and LMW-m genes, as previously reported [38],[39]. Sine LMW-m and LMW-s type subunit genes generally show higher consistency, they showed close phylogenetic evolutionary relationships. Glu-A3a showed a closer relationship with other LMW-i type genes from common wheat. All of the LMW-i type subunit genes from common wheat and related species shared higher sequence identity, indicating their high evolutionary conservation.

Figure 5
figure 5

Homology tree constructed based on the coding regions of 21 LMW-GS genes. 21 LMW-GS genes named AB062876, AB062877 and AB062878 [14], AB262661 (Takeuchi T, 2006), AB119007 and AB164415 [40], AY453158 and AY453159 [41], AY585355 [42], DQ307389, DQ307387 and DQ345449 [39], DQ457416 [55]. For example, the over-expressed HMW-GS 1Bx7OE has positive effects on dough strength [53],[56]. In addition, the accumulation rates vary between different groups of proteins, suggesting differential regulation of protein biosynthesis and different quality performance. In particular, the wheat biotype with superior HMW-GS 5 + 10 subunits accumulated larger polymers more quickly than the biotype with poor allelic subunits 2 + 12 [54].

The B-subunits of LMW-GS are the most abundant and have the greatest impact on wheat processing qualities [6]. In this work, RP-UPLC analysis revealed a higher expression level and greater proportion of Glu-A3a-encoded B-subunits, accounting for more than 22% of the total LMW-GS in CS (Figure 1c), indicating its major contribution to LMW-GS synthesis and its important roles in determining dough quality. A recent study also found that higher numbers of active LMW-GS genes at the Glu-A3 and Glu-D3 loci in ** grains of hexaploid wheats. J Exp Bot. 1996, 47 (9): 1377-1385. 10.1093/jxb/47.9.1377." href="/article/10.1186/s12870-014-0367-3#ref-CR54" id="ref-link-section-d142980475e3962">54],[57]. LMW-GS, HMW-GS, and ω-gliadins can be detected by gel electrophoresis as early as 7 DPA [54], and 10–18 DPA represents the key stage of storage protein synthesis [58]. In the present study, Glu-A3a transcripts demonstrated an up-down expression pattern during grain development, and the highest expression level occurred at 14 DPA (Figure 7a), similar to a previous report [55]. Thus, the Glu-A3a-encoded B-subunit has a higher accumulation rate during grain development similar to HMW-GS 5 + 10 [54], which could improve the conformation of the regular gluten structure. Some important genes related to storage protein folding and synthesis, such as protein disulfide isomerase (PDI) and binding protein (BiP) genes, generally have higher expression levels at the early grain developmental stages. For instance, the PDI genes PDIL1-1 and PDIL2-1, which are involved in disulfide bond formation, displayed a peak expression level in the early stages (about 10–15 DPA) of grain development [58]. The higher accumulation rate of the Glu-A3a-encoded B-subunit was accompanied by higher expression levels of the genes involved in storage protein synthesis and assembly during early grain development, suggesting that this subunit could improve the conformation of gluten macropolymers (GMP) and result in superior dough quality.

Potential application of Glu-A3ain wheat quality improvement through molecular marker-assisted selection

Characterization of the allelic variations of LMW-GS is important for improvement of wheat-processing quality. Some allelic variations of LMW-GS have greater positive effects on dough properties than others [3],[45],[60]. Marker-assisted selection is an effective supplement to conventional breeding practices. For LMW-GS, because of the low resolution of traditional SDS-PAGE and the tedious operation procedures involved in 2-DE, development of different markers is important for the study and application of target subunits.

Recently, with increasing numbers of LMW-GS alleles being cloned and sequenced from common wheat, different molecular markers have been developed to rapidly screen and select desirable Glu-3 alleles. Zhang et al. [41] developed a set of markers that can be used to discriminate the alleles Glu-A3a, b, c, d, e, f, and g. Long et al. [61] classified 69 LMW-GS genes registered in GenBank into nine groups and established nine group-specific primer sets to identify each group. Ikeda et al. [62] developed 12 specific PCR markers to distinguish 12 groups of LMW-GS genes in the wheat cultivar Norin 61. Ten allele-specific STS markers for Glu-D3 were developed by Zhao et al. [10: Table S3.

Identification of seed proteins

Protein extraction, A-PAGE, SDS-PAGE and RP-UPLC

According to the solubility in a series of solvents, grain albumins, globulins, gliadins and glutenins were extracted according to the established methods [28],[34],[39],[76],[77].

Expression of the cloned LMW-GS gene in E. coli

The gene cloned was re-amplified to remove the signal peptides by designing a new pair of primers CS-F (5’-GGGCATATGATTTCACAGCAACAA-3’) and CS-R (5’-CTCGAGTCAGTAGACACCAACTCCGATG-3’), NdeI and XhoI sites (underlined) were incorporated into the 5’ ends of the CS-F and CS-R, respectively. After purification, the PCR products were ligated into the expression vector pET-30a (Novagen), and transformed into E.coli BL21 (DE3) plysS cells. And then we extract and separate the expressed protein from the E.coli, after that, we carried out them by SDS–PAGE according to Li et al. [75].

mRNA extraction, cDNA synthesis and qRT-PCR

Developmental seeds from three spikes were combined together to extract total RNA from endosperm of CS and CS-n, and cDNA synthesis, qRT-PCR were according to Wang et al. [56]. The primers were: LMW-i-F: TGAAGACCTTCCTCGTCTTTG, LMW-i-R: CTGTGAAATTTGCGCAACG. Gene-specific primers were designed using Primer 5.0 and their specificities were checked by the melting curves of the RT-PCR products. Each qRT-PCR reaction was performed in 20 μl volumes containing 10 μl 2 × SYBR® Premix Ex Taq™ (TaKaRa), 2 μl 50-fold diluted cDNA, 0.4 μl of each gene-specific primer and 7.2 μl ddH2O. PCR conditions were as follows: 95°C for 3 min, 45 cycles of 15 s at 95°C, 57°C for 20s and 72°C for 20s. Three replicates were used for each sample. Reactions were conducted in a CFX96 Real-Time PCR Detection System (Bio-Rad). All data were analyzed with CFX Manager Software (Bio-Rad).

Determination of Glu-A3adeletion in CS by STS-PCR marker

To identify the deletion of Glu-A3a in CS-n, we did the STS-PCR marker of the seven NILs, CS-1Sl/1B, S genome addition line, CB037A, CS, CS-n. We used the marker name of glu-A3a to do PCR as what Wang et al. [24] did before. The primer sets are LA1F: AAACAGAATTATTAAAGCCGG, and SA1R: GGTTGTTGTTGTTGCAG CA. Their PCR cycling conditions were 94°C for 4 min, followed by 35 cycles of 94°C for 35 s, 55°C for 45 s, 72°C for 40 s, and a final extension at 72°C for 8 min.

Development and validation of allele-specific PCR markers for Glu-A3a

To identify the Glu-A3a gene in different genotypes, based on the SNPs we detected in Glu-A3a, we designed the primer named Glu-A3a F: GCAAAGAAGGAAAAGAG GTGG, R: GGTTGTTGTTGTTGCTGCA as the primer to discriminate the gene Glu-A3a from others in CS and CS-n, this also validated in four RILs (CB037A and 1Sl/1B), CS-1Sl/1B, 1Sl genome addition line, 7 NILs of Aroona and 48 varieties. PCR cycling conditions were 94°C for 4 min, followed by 35 cycles of 94°C for 35 s, 60°C for 30 s, 72°C for 30 s, and a final extension at 72°C for 8 min.

Additional files