The unique viscoelastic properties conferred by gluten proteins in bread wheat are the basis of the flexible processing qualities in producing a wide range of food products for a large proportion of the world population. Gluten proteins, also named prolamins, are classically divided into gliadins and glutenins, based on different solubilities in an alcohol/water mixture 1 . The gliadins are generally monomeric proteins, divided into three groups, α/β-, γ- and ω-gliadins, based on their electrophoretic mobilities at low pH 2 . Glutenins form polymeric proteins stabilized by interchain disulfide bonds. Based on different molecular weights, glutenins can be classified into two groups, high-molecular-weight glutenin subunits (HMW-GS) and low-molecular-weight glutenin subunits (LMW-GS) 2 3 . LMW-GS are further divided into B-, C- and D-group subunits according to their mobilities in sodium dodecyl sulphate polyacrylamide-gel electrophoresis (SDS-PAGE) 4 .

In bread wheat, HMW-GS are encoded by genes at the orthologous Glu-1 loci on the long arms of chromosomes 1A, 1B and 1D (Glu-A1, Glu-B1 and Glu-D1). Each locus possesses two paralogous genes encoding one x- and one y-type subunit 5 . LMW-GS genes are located at the Glu-A3, Glu-B3 and Glu-D3 loci on the short arms of group 1 chromosomes. The LMW-GS genes at the Glu-3 loci and the gliadin genes at the Gli-1 loci are tightly linked and form gene clusters covering several centimorgans (cMs) 6 7 8 . Moreover, unlike the simple composition of HMW-GS, LMW-GS are encoded by a complex multigene family without the information of the exact number of genes 9 10 . A large number of genes and abundant allelic variations at Glu-3 loci and their tight linkage with gliadin genes make it difficult to elucidate the composition and function of LMW-GS genes in bread wheat 4 .

SDS-PAGE is widely used to investigate the abundant seed storage proteins in bread wheat. Based on the mobility of proteins in SDS-PAGE gels, whole seed proteins are divided into four groups, HMW-GS, D-group, B-group, and C-group. D-group proteins are mainly composed of ω-gliadin proteins, whereas B-group mostly consists of LMW-GS proteins, and C-group comprise α, β and γ-type gliadins and several LMW-GS proteins 4 . Based on the different electrophoretic patterns, LMW-GS protein alleles encoded by Glu-3 loci are designated alphabetically (e.g., Glu-A3a) 11 . However, identification of the LMW-GS composition in breeding programs remains a significant challenge because determination of LMW-GS alleles with SDS-PAGE needs much experience. This is why the functions of LMW-GS alleles are not well characterized. Gene-specific markers for Glu-A3 and Glu-B3 alleles were developed to identify different LMW-GS alleles. However, molecular markers for Glu-D3 alleles are still not available due to the slight differences among alleles 12 13 14 15 16 17 . Using BAC library screening and proteomics methods, LMW-GS genes in Norin 61 (Glu-A3d, Glu-B3i and Glu-D3c), Glenlea (Glu-A3g, Glu-B3g and Glu-D3c) and Xiaoyan 54 (Glu-A3d, Glu-B3d and Glu-D3c) were identified and characterized 10 18 19 20 . These studies greatly improved our understanding of the unique genes encoding different LMW-GS alleles in bread wheat. Recently, based on the conserved and polymorphic structure of LMW-GS genes, we developed a LMW-GS gene marker system and a full-length gene cloning method 21 22 . They were successfully used to identify and characterize more than 16 LMW-GS genes in individual wheat varieties 21 . Both methods are helpful in elucidating the composition of Glu-3 alleles in LMW-GS genes of bread wheat.

The effects of glutenin alleles on dough properties and processing qualities were mostly studied in two types of populations: structured populations (e.g., recombinant inbred lines (RILs) and doubled haploid lines) derived from biparental crosses, and non-structured populations, general collections of varieties and breeding lines. Due to the simple composition and easy identification of allelic variants of HMW-GS, the contributions of HMW-GS to dough properties and end-use quality were well investigated and widely used in breeding programs 23 . However, HMW-GS alone could not explain the variation in quality among wheat varieties, as LMW-GS also contributed to dough properties 24 25 26 27 28 29 30 31 . For example, LMW-GS alleles made a slightly larger contribution than HMW-GS to dough extensibility 32 33 . Compared with HMW-GS, LMW-GS formed highly polymorphic protein complex and contain abundant allelic variation. Using SDS-PAGE and allele-specific primers, LMW-GS alleles were identified in wheat collections or structured populations, and their effects on processing quality were analyzed and discussed. However, controversies were common in regard to different kinds of populations or collections. For example, Cane et al. 34 and Eagles et al. 33 reported that Glu-A3e was correlated with inferior dough resistance and extensibility, whereas Zheng et al. 35 showed that Glu-A3e was a favorable allele for dough-mixing properties. Due to the complex composition of LMW-GS alleles and difficulties in distinguishing LMW-GS from gliadins in SDS-PAGE gels, the molecular genetic mechanisms behind the functional differences of LMW-GS alleles are not well investigated.

In the present study, a set of near-isogenic lines (NILs) containing five Glu-A3 alleles, eight Glu-B3 alleles and five Glu-D3 alleles, was used to study the effects of LMW-GS on the composition of glutenin macro-polymers (GMP), dough properties, and pan bread making quality. These NILs were investigated using SDS-PAGE, the LMW-GS gene marker system and two-dimensional gel electrophoresis (2D-PAGE) for identifying the composition of LMW-GS genes and proteins in each LMW-GS allele, and analyzing their association with dough properties and bread-making quality.

To analyze the genes encoding LMW-GS alleles in Aroona NILs, we investigated the 16 Aroona NILs using a previously developed LMW-GS gene marker system, that was efficient in separating LMW-GS genes in bread wheat (Figure 2; Table 1) 21 22 . Eighteen LMW-GS genes were identified in Aroona, including four Glu-A3 genes (A3-391, A3-400, A3-502-2 and A3-620), four Glu-B3 genes (B3-530-2, B3-578, B3-607 and B3-621-1), and eight Glu-D3 genes (D3-385, D3-393, D3-394, D3-432, D3-528, D3-575, D3-578-1 and D3-591) (Table 1). The other two novel genes corresponding to DNA fragments 388 and 410 present in all 16 lines (Table 1) were pseudogenes with premature termination codons in the CDS regions. Among 16 Aroona NILs, Aroona-Glu-D3f possessed 14 unique genes, with the least number of LMW-GS genes, whereas Aroona-Glu-A3d had the largest number of LMW-GS genes (20 genes). Comparison of the LMW-GS genes among NILs indicated that each NIL differed from Aroona at only one Glu-3 locus. With the help of the complete gene sequences available in Aroona NILs and some other wheat varieties 10 13 14 17 18 19 21 , the composition of LMW-GS genes in each Aroona NIL was well characterized.

^a Bold numbers indicate active LMW-GS genes.

For Glu-A3 alleles, 4–6 genes were identified in Aroona (Glu-A3c) and four Glu-A3 NILs (Table 1; Figure 2). Except for Aroona-Glu-A3d, these NILs possessed LMW-GS genes A3-391 and A3-400 (Table 1; Figure 2), which were m-type pseudogenes with premature termination codons. The other genes were i-type genes. The A3-502 gene contained four allelic variants with unique SNPs and InDels, i.e., A3-502-1, A3-502-2, A3-502-3 and A3-502-4. All A3-502 allelic variants and the A3-640 gene were pseudogenes, containing premature termination codons in the coding sequences. The i-type genes, A3-643, A3-620, A3-646 and A3-573, were the only active genes in Aroona-Glu-A3b, A3c, A3e and A3f, respectively. Sequence alignments showed that A3-620 and A3-643 shared high identity (> 99%) differing only in one InDel and one SNP, whereas A3-573 and A3-646 were greatly different from each other and the other i-type genes (Additional file 1: Figure S2). A3-573 was much shorter than the other i-type proteins due to deletion of three repetitive units and several glutamine residues in repetitive regions, and lacking GTFLQPH in the C-terminal domain (Additional file 1: Figure S2). Aroona-Glu-A3d contained six unique LMW-GS genes (Table 1; Figure 2). For two m-type genes, A3-370 was a pseudogene, whereas A3-402 contained an intact coding sequence that could be expressed in the developing grain 18 . The other four genes, A3-484, A3-565, A3-568 and A3-662 were i-type genes, and among them A3-568 and A3-662 were active. Both genes contained unique SNPs in the C-terminal domain and different lengths of repetitive region compared to other i-type genes (Additional file 1: Figure S2). Thus, compared to only one LMW-GS gene expressed in the other Glu-A3 alleles, Glu-A3d contained three active genes, one m-type (A3-402) and two i-type (A3-568 and A3-662) LMW-GS genes. Except for A3-402, the active genes at the Glu-A3 locus were i-type genes. These i-type genes encoded B-group LMW-GS, producing different electrophoretic patterns in the B-group region among Glu-A3 NILs (Figure 1).

For Glu-B3 alleles, 3–5 LMW-GS genes were identified in Aroona (Glu-B3b) and seven Glu-B3 NILs. Gene B3-530 was present in all Aroona NILs and four allelic variants (B3-530-1, -2, -3 and B3-510) were identified (Table 1). Sequence alignment showed that the deduced protein sequences of the three B3-530 allelic variants differed in only four amino acids, whereas B3-510 contained a deletion in the repetitive domain and differed by seven amino acids from the B3-530 proteins (Additional file 1: Figure S3). B3-544, B3-593, B3-601 and B3-607 were allelic variants, sharing over 99% of identity and their deduced protein sequences differed only in the repetitive domain with the deletion of glutamine residues or one repetitive unit (Additional file 1: Figure S4). B3-621-1, B3-621-2 and B3-624 genes were highly conserved and their proteins differed only in one InDel of a glutamine residue in the repetitive domain and one amino acid at the C terminal region (Additional file 1: Figure S5). The B3-688 gene was quite distinct, differing from B3-621 and B3-624 in several SNPs and InDels (Additional file 1: Figure S5). Moreover, B3-688 varied among Aroona-Glu-B3c, B3d, B3h and B3i, and the protein products differed in seven amino acids and two InDels (Additional file 1: Figure S5). Among Glu-B3 genes, B3-548, B3-578 and B3-813 were pseudogenes with premature termination codons, whereas the other genes contained intact ORFs (Open Reading Frame). Sequence analysis showed that B3-530, B3-510, B3-548 and B3-570 were m-type genes and the others were s-type genes. Based on the differences in s-type genes, Glu-B3 NILs were classified into two groups. Aroona-Glu-B3a, B3b, B3f and B3g contained B3-578, B3-544/593/601/607, and B3-621/624, and the others had longer sequences of genes, B3-688 and B3-691/813/Null (Table 1). This classification was consistent with the different electrophoretic patterns contrasting the two groups (Figure 1a).

Aroona-Glu-B3a, B3b, B3f and B3g had similar LMW-GS compositions, but each of them possessed unique allelic variants (Table 1; Figure 2). Both Aroona-Glu-B3f and B3g contained B3-530-3, but they differed in B3-544, B3-601, B3-621-1 and B3-621-2 genes (Table 1; Additional file 1: Figures S4 and S5).Aroona-Glu-B3f and Aroona (Glu-B3b) shared the same B3-621-1, but possessed different genes B3-601/B3-530-3 and B3-607/B3-530-2, respectively (Table 1; Additional file 1: Figures S3 and S4). On the other hand, Aroona-Glu-B3c, B3d, B3h, and B3i also differed in active genes and their allelic variants. Aroona-Glu-B3c and B3d contained B3-530-1, but their B3-688 proteins differed in five amino acids (Table 1; Additional file 1: Figure S5). Aroona-Glu-B3i was unique, containing four active LMW-GS genes B3-510, B3-570, B3-688-4 and B3-691 (Table 1; Figure 2). B3-570 was only present in Glu-B3i, encoding an m-type LMW-GS with the N-terminal domain METSQIPGLEKPS. Although B3-688-4 and B3-691 were different genes, they shared about 99% identity and B3-691 is formally reported for the first time.

At the Glu-D3 locus, eight genes were identified from individual NILs except Aroona-Glu-D3f (Table 1). D3-393 and D3-583/586/591 had premature termination codons. All the other genes were widely reported and were expressed in developing grains 10 17 18 20 . D3-385, D3-394 and D3-575 were highly conversed among the Aroona NILs. D3-432 and D3-441 were allelic variants with over 99% identity, and differed in only two SNPs and two InDels. Allelic variants D3-525 and D3-528 were identical, except for one InDel. Two allelic variants, D3-578-1 and D3-578-2 differed by eight SNPs (Table 1). Compared with the active Glu-D3 genes in Aroona (Glu-D3c), D3-441 and D3-578-2 were present in Aroona-Glu-D3b and D3d, which possessed the same active LMW-GS genes (Table 1; Figure 2). Besides D3-441 and D3-578-2, Aroona-Glu-D3a contained unique gene D3-525 in contrast to Aroona. In Aroona-Glu-D3f, only four LMW-GS genes were identified; the other four genes might be absent due to deletion, substitution or other mutations (Table 1; Figure 2).

Each Aroona NIL possessed unique LMW-GS alleles. The gene compositions of these LMW-GS alleles were dissected, all the LMW-GS gene sequences were characterized, and the proteins sequences were deduced (Table 1; Figure 2). Based on the size of the deduced LMW-GS and the LMW-GS 2D-PAGE spots identified from Xiaoyan 54, Jing 411 and Norin 61 18 19 , LMW-GS genes were assigned to protein bands in SDS-PAGE gels (Figure 1b). The active i-type genes at the Glu-A3 locus, all the active genes at Glu-B3 and the genes D3-525/528, D3-575 and D3-578 at Glu-D3 encoded B-group LMW-GS, whereas the protein products of the other active genes belonged to the C-group. Moreover, B3-621/624, B3-593/601/607, D3-575 and D3-578 encoded proteins with similar molecular weights (38–40 kDa), and ran synchronously on SDS-PAGE, forming the thickest protein bands in Aroona-Glu-B3a, B3b and B3f. B3-688 genes encoding the longest protein conferred the unique protein pattern on Aroona-Glu B3d, B3h and B3i. (Figure 1).

The data from SDS-PAGE and the LMW-GS marker system showed that each Glu-3 allele contained several unique LMW-GS and gliadin genes. To analyze the composition of LMW-GS and gliadin proteins in individual NILs, we performed 2D-PAGE analysis of the whole flour proteins (Figures 3, 4 and 5). In the previous studies, using 2D-PAGE coupled with MS or N-terminal sequencing technology, LMW-GS proteins in Xiaoyan 54, Jing 411, Norin 61 and Butte 86 were successfully identified 18 19 36 . These data greatly contributed to the identification of flour proteins in the Aroona NILs.

Among five Glu-A3 NILs, Aroona-Glu-A3b and Aroona (Glu-A3c) shared the same protein patterns, except for one gliadin spot, which was in agreement with the similar composition of LMW-GS genes and the high identity between A3-620 and A3-643 (Figure 3a). Compared with Aroona, Aroona-Glu-A3d contained three unique LMW-GS, A3-402, A3-568 and A3-662, and two gliadin spots were absent (Figure 3a). For Aroona-Glu-A3e, the protein product of the unique i-type gene A3-646 was too little to be detected, but three unique gliadin spots were present in 2D-PAGE (Figure 3b). Aroona-Glu-A3f also contained one unique i-type protein (A3-573) and one gliadin spot different from Aroona (Figure 3b). Thus, accompanying the unique LMW-GSs in individual Glu-A3 NILs, generally 1–2 gliadin spots were different among them. The only exception was Aroona-Glu-A3e which possessed the highest number of gliadins and the lowest quantity (none) of i-type LMW-GS among the Glu-A3 NILs (Figure 3b).

Among eight Glu-B3 NILs, Aroona-Glu-B3c, B3d, B3h and B3i shared similar protein spot patterns on 2D-PAGE, consistent with their similar electrophoretic patterns on SDS-PAGE (Figures 1a and 4). Compared with the protein spot pattern of Aroona (Glu-B3b), Aroona-Glu-B3c, B3d, B3h and B3i contain the unique B3-688 proteins (Figure 4). The B3-688 protein spot in Aroona-Glu-B3c was much weaker than those in the other three NILs (Additional file 1: Figure S6). Except for LMW-GS proteins, only one or two small gliadin spot differences were identified among the four NILs (Figure 4). The other three Aroona NILs (i.e., Aroona-Glu-B3a, B3f and B3g) had similar protein spot patterns with Aroona (Glu-B3b; Figure 4). Among these four NILs, a unique LMW-GS protein B3-544 was detected in Aroona-Glu-B3g, whereas the other three alleles shared the same LMW-GS protein spot patterns (Figure 4). Thus, the 2D-PAGE patterns of Aroona and seven Glu-B3 NILs suggested that their different proteins mainly were unique LMW-GS proteins rather than gliadins.

For the Glu-D3 NILs, the protein spots encoded by the D3-441 gene in Aroona-Glu-D3a, D3b and D3d had larger molecular weights and pIs than its allelic variant D3-432 in Aroona (Figure 5). No other different LMW-GS spots were detected among these Glu-D3 NILs, and Aroona-Glu-D3a and D3d, each contained only one unique gliadin spot (Figure 5). The Aroona-Glu-D3f allele was unique, lacking two LMW-GS proteins (i.e., D3-385 and D3-441) and at least three medium gliadin spots (Figure 5). The absence of four LMW-GS genes was also observed using the LMW-GS gene marker system (Table 1). Thus, except for Aroona-Glu-D3f, only one or two different protein spots were detected among Glu-D3 NILs, and the proteins encoded by genes at the Glu-D3 locus were highly conserved.

Dough properties such as Zeleny-sedimentation value (ZSV), Farinograph and Extensograph parameters, GMP parameters, and pan bread quality parameters, were measured on the Aroona NILs (Table 2). The 16 genotypes showed significant differences (P < 0.05 or P < 0.01) in most quality parameters, suggesting that genotype had an important influence on variation in wheat quality properties (Additional file 2: Table S1). Significant differences in some parameters (e.g., kernel protein, ZSV, Farinograph water absorption, glutenin and gliadin contents, and external bread color and structure) were also observed between the two locations (Additional file 2: Table S1). However, no significant interaction effects between genotypes and locations were detected for most quality parameters, except for external color of bread and Farinograph development time (Additional file 2: Table S1). Analysis of variance (ANOVA) and multiple comparisons of all quality parameters were performed among all NILs within each Glu-3 group (Table 2, Additional file 2: Tables S2, S3 and S4).

* Significant at P < 0.05; ** Significant at P < 0.01.

^a LV: Loaf volume (cm³), LVS: Loaf volume score, EC: External color, Sha: Shape, IC: Inner color, Smo: Smoothness, Spr: Springiness, Str: Structure, TF: Taste flavor, PBTS: Pan bread total score, KP: Kernel protein (%, 14% m.b.), ZSV: Zeleny sedimentation value (mL), ST: Farinograph stability time (min), DT: Farinograph development time (min), Wab: Farinograph water absorption (%), Rmax: Extensograph maximum resistance (B.U.), EA: Energy area (cm²), Ext: Extensograph extensibility (mm), EPP: Extractable glutenin polymeric protein; UPP: Unextractable glutenin polymeric protein; Gli/Glu: Ratio of gliadin to glutenin; %UPP = UPP/(UPP + EPP) × 100.

ANOVA showed that the Glu-A3 NILs were significantly different in most quality parameters (Table 2). This suggested that Glu-A3 alleles affected the bread-making quality of bread wheat. Subsequent multiple comparisons indicated that Aroona-Glu-A3e always produced the worst quality parameters, e.g., ZSV, Farinograph stability time (ST), Extensograph maximum resistance (Rmax), Extensograph extensibility (Ext), percentage of SDS-unextractable fraction in total polymeric protein (%UPP), and pan bread total score (PBTS), whereas the other five Glu-A3 NILs had similar values for Rmax, ST, %UPP and Ext (Figure 6). Although Aroona-Glu-A3d had the largest ZSV, it showed only moderate PBTS (Figure 6).

Glu-B3 NILs showed large differences in the most important quality parameters (Table 2), indicating that Glu-B3 alleles were also significantly associated with bread-making quality. Among eight Glu-B3 NILs, Aroona-Glu-B3g produced the highest PBTS, ZSV, ST, Rmax, %UPP and some other parameters (Figure 6, Additional file 2: Tables S2, S3 and S4). Similarly, Aroona (Glu-B3b) and Aroona-Glu-B3i had higher quality parameters than the other Glu-B3 NILs, excluding Aroona-Glu-B3g. In contrast, Aroona-Glu-B3c performed the lowest in ZSV, ST, Rmax, %UPP, PBTS, and some other parameters (Figure 6; Additional file 2: Tables S2, S3 and S4), suggesting that Glu-B3c was an undesirable allele for bread-making quality. The other Glu-B3 NILs (Aroona-Glu-B3a, B3d, B3f and B3h) possessed similar moderate quality parameters, without significant differences.

Five Glu-D3 NILs produced similar quality parameters, without significant differences, except for Farinograph water absorption and the ratio of gliadin to glutenin (Table 2). This was consistent with the similar composition of LMW-GS genes and proteins among Glu-D3 NILs.

In the present study, SDS-PAGE, 2D-PAGE and LMW-GS gene molecular marker system were used for characterization of LMW-GS genes and proteins in 16 Aroona NILs. Each NIL contained a unique LMW-GS allele. The genetic and protein compositions of each LMW-GS allele were dissected. In addition, we analyzed the functional differences among the LMW-GS alleles in bread-making quality using these NILs. The molecular mechanisms behind the functional differences were discussed based on the characterization of LMW-GS genes and proteins in the NILs.

SDS-PAGE is a useful method to accurately identify HMW-GS in bread wheat, but is inefficient for separating LMW-GS alleles, because of the presence of large numbers of LMW-GS proteins with similar mobilities with each other and with gliadins (Figure 1b). In the present study, the LMW-GS gene marker system was successfully used to separate 14–20 LMW-GS from each NIL, and also distinguished 16 Glu-3 NILs, except for Aroona-Glu-B3c and B3d without the length polymorphisms among B3-688-1 and B3-688-2 (Table 1; Figure 2). These results indicated that the LMW-GS gene marker system was efficient and accurate not only in separating members of the LMW-GS gene family 21 , but also in distinguishing allelic variants of individual LMW-GS genes. Recently, allele-specific markers were widely used in distinguishing Glu-A3 and Glu-B3 alleles 12 13 14 . However, the high conservation among Glu-D3 allelic variants made it difficult to develop allele-specific markers for discriminating different allelic variants. Because these allelic variants showed DNA sequence polymorphisms in length, the LMW-GS gene marker system worked well in dissecting the complex genes and allelic variants at the Glu-D3 locus (Table 1). On the other hand, compared with only one gene in individual alleles identified with gene-specific primers 13 14 , almost all the genes in Glu-3 alleles were displayed using the LMW-GS gene marker system (Table 1). Identification of all genes in each allele will greatly contribute to an understanding of the molecular mechanism determining functional differences among Glu-3 alleles in bread wheat.

In the present study, Aroona and its NILs contained most of the Glu-3 allelic variations identified from the worldwide wheat germplasm in a cooperative program 12 . Each LMW-GS allele was encoded by several linked genes or haplotypes 13 14 16 17 21 22 . The dissection of LMW-GS genes in Aroona NILs would facilitate the haplotype analysis of LMW-GS genes in common wheat. At the Glu-A3 locus, besides two m-type genes, 2–4 i-type genes were identified from individual Glu-A3 alleles. Unlike the conserved m-type genes, i-type genes were completely different among five alleles. These i-type genes might be tightly linked, forming unique haplotype in each Glu-A3 allele, e.g. A3-484/A3-565/A3-568/A3-662 in Glu-A3d (Table 1) 13 18 19 21 . At the Glu-B3 locus, 1–2 m-type gene(s) and 1–3 s-type gene(s) were characterized in individual alleles. Generally, these s-type genes could be divided into two groups, one containing B3-578/B3-544/593/601/607/B3-621/624, the other having B3-688/B3-691/813/N (Table 1, Figure 2). These s-type genes in each group might cosegregate and form unique haplotypes in bread wheat 14 . Among four allelic variants of m-type gene B3-530, B3-530-2 was present in both Aroona (Glu-B3b) and Aroona-Glu-B3h, containing difference s-type haplotypes B3-578/B3-607/B3-621-1 and B3-688/B3-813, respectively. Thus, at the Glu-B3 locus, the m- and s-type genes appeared not to be tightly linked (Table 1; Figure 2) 6 18 . All the Glu-D3 alleles except Glu-D3f contained eight LMW-GS genes (Table 1). Only one or two allelic variants (D3-432/441, D3-525/528 and D3-578-1/-2 were identified for each LMW-GS gene and only a few SNPs were detected among allelic variants. Thus, Glu-D3 genes were highly conserved among wheat varieties 16 17 , wherease Glu-A3 and Glu-B3 genes showed high diversities, which were derived from multiple allelic variations of i- and s-type genes, respectively.

To analyze differences in seed storage proteins among LMW-GS NILs, flour proteins were displayed using 2D-PAGE (Figures 3, 4 and 5). LMW-GS proteins shared some similar characteristics, such as mostly located in the B-group region and larger pIs than gliadins. Although it is difficult to detect small variation in molecular weight and quantity of each protein spot in 2D-PAGE, the composition and allelic variations of seed storage proteins for each Glu-3 allele were successfully characterized. These results were in agreement with the 2D protein patterns of Xiaoyan 54, Jing 411, Norin 61, Butte 86 and some other varieties identified with LC-MS or N-terminal sequencing 18 19 36 . For example, the Glu-A3d and Glu-B3d alleles showed similar electrophoretic patterns to Xiaoyan 54 and Norin 61 in i-type (A3-568 and A3-662) and s-type (B3-688) proteins, respectively 18 19 , and Glu-B3g shared the same spot pattern of B3-544 with Glenlea 19 . Moreover, comparison of 2D-PAGE patterns showed that, except for the different LMW-GS proteins, only 1–3 unique small gliadin spots were present in individual Aroona NILs (Figures 3, 4 and 5). These results suggested that the differences among the Aroona NILs were mainly derived from allelic variation of LMW-GS genes.

The NIL population was used to study the effects of LMW-GS on dough properties, GMP parameters, and pan bread quality parameters. The results confirmed that the LMW-GS played important roles in determining variation in wheat quality properties. Among the three Glu-3 loci, alleles Glu-A3 and Glu-B3 were of major importance in determining differences in processing qualities among the NILs (Table 2).

In the present study, we dissected the genetic and protein composition of 16 LMW-GS NILs, measured the dough property and bread-making quality properties of individual NILs, and performed functional analyses for each allele. Among five Glu-A3 alleles, Glu-A3e (i-type haplotype A3-502-3/A3-646) was inferior with negative effects on all quality properties. Among eight Glu-B3 alleles, Glu-B3b (m-type gene B3-530-2 and s-type haplotype B3-578/B3-607/B3-621-1), Glu-B3g (m-type gene B3-530-3 and s-type haplotype B3-578/B3-544/B3-621-2) and Glu-B3i (m-type haplotype B3-510/B3-570 and s-type haplotype B3-688-4/B3-691) were correlated with superior bread-making quality, whereas Glu-B3c (m-type gene B3-530-1 and s-type haplotype B3-688-1/N) produced inferior quality properties. Among five Glu-D3 alleles, there were no significant differences in all quality parameters measured in the present study. Moreover, all alleles with superior dough properties and pan bread quality also possessed high contents of UPP and %UPP. Thus, it is possible that LMW-GS alleles determine dough viscoelasticity by modifying the size distribution of glutenin polymers and aggregative properties of glutenins 50 . These results significantly enhance our understanding of the composition of LMW-GS, confirm the strong effects of LMW-GS on not only dough extensibility but dough strength, and provide useful information for quality improvement in bread wheat.

2D-PAGE: Two-dimensional gel electrophoresis (IEF × SDS-PAGE); %UPP: UPP/(UPP + EPP) × 100; ANOVA: Analysis of variance; CIMMYT: International maize and wheat improvement center; cMs: Centimorgans; DT: Farinograph development time (min); DTT: Dithiothreitol; EA: Energy area (cm²); EC: External color; EPP: Extractable glutenin polymeric protein; Ext: Extensograph extensibility (mm); Gli/Glu: Ratio of gliadin to glutenin; GMP: Glutenin macro-polymers; HMW-GS: High-molecular-weight glutenin subunits; HPLC: High-performance liquid chromatography; IC: Inner color; IEF: Isoelectric focusing; kDa: Kilodalton; KP: Kernel protein (%, 14% m.b.); LMW-GS: Low-molecular-weight glutenin subunits; LV: Loaf volume (cm³); LVS: Loaf volume score; NIL: Near-isogenic lines; ORFs: Open reading frame; PBTS: Pan bread total score; PCR: Polymerase chain reaction; pI: Isoelectric points; RILs: Recombinant inbred lines; Rmax: Extensograph maximum resistance (B.U.); SDS: Sodium dodecyl sulfate; SDS-PAGE: Sodium dodecyl sulphate-polyacrylamide gel electrophoresis; Sha: Shape; Smo: Smoothness; Spr: Springiness; ST: Farinograph stability time (min); Str: Structure; TCA: Trichloroacetic acid; TF: Taste flavor; Tris–HCl: Tris (hydroxymethyl) aminomethane hydrochloride; UPP: Unextractable glutenin polymeric protein; Wab: Farinograph water absorption (%); ZSV: Zeleny sedimentation value (mL).

XZ carried out SDS-PAGE, 2D-PAGE and LMW-GS gene marker system analyses and drafted the manuscript. HJ measured glutenin macro-polymers, dough properties and pan bread quality parameters and performed statistical analysis. YZ participated in the measurement of processing quality parameters. DL participated in the identification of LMW-GS genes and proteins with SDS-PAGE, 2D-PAGE and LMW-GS gene marker system and revised the manuscript. GL participated in the design of the study and quality parameter analysis. XX participated in the design of the study and revised the manuscript. ZH conceived of the study and participated in its design, coordinated the various research groups, and revised the manuscript. AZ conceived of the study and participated in its design, and revised the manuscript. All authors read and approved the final manuscript.

· In the past five years have you received reimbursements, fees, funding, or salary from an organization that may in any way gain or lose financially from the publication of this manuscript, either now or in the future? Is such an organization financing this manuscript (including the article-processing charge)? If so, please specify.

No.

· Do you hold any stocks or shares in an organization that may in any way gain or lose financially from the publication of this manuscript, either now or in the future? If so, please specify.

No.

· Do you hold or are you currently applying for any patents relating to the content of the manuscript? Have you received reimbursements, fees, funding, or salary from an organization that holds or has applied for patents relating to the content of the manuscript? If so, please specify.

No.

· Do you have any other financial competing interests? If so, please specify.

No.

Are there any non-financial competing interests (political, personal, religious, ideological, academic, intellectual, commercial or any other) to declare in relation to this manuscript? If so, please specify.

No.