Composition and functional analysis of low-molecular-weight glutenin alleles with Aroona near-isogenic lines of bread wheat

Background Low-molecular-weight glutenin subunits (LMW-GS) strongly influence the bread-making quality of bread wheat. These proteins are encoded by a multi-gene family located at the Glu-A3, Glu-B3 and Glu-D3 loci on the short arms of homoeologous group 1 chromosomes, and show high allelic variation. To characterize the genetic and protein compositions of LMW-GS alleles, we investigated 16 Aroona near-isogenic lines (NILs) using SDS-PAGE, 2D-PAGE and the LMW-GS gene marker system. Moreover, the composition of glutenin macro-polymers, dough properties and pan bread quality parameters were determined for functional analysis of LMW-GS alleles in the NILs. Results Using the LMW-GS gene marker system, 14–20 LMW-GS genes were identified in individual NILs. At the Glu-A3 locus, two m-type and 2–4 i-type genes were identified and their allelic variants showed high polymorphisms in length and nucleotide sequences. The Glu-A3d allele possessed three active genes, the highest number among Glu-A3 alleles. At the Glu-B3 locus, 2–3 m-type and 1–3 s-type genes were identified from individual NILs. Based on the different compositions of s-type genes, Glu-B3 alleles were divided into two groups, one containing Glu-B3a, B3b, B3f and B3g, and the other comprising Glu-B3c, B3d, B3h and B3i. Eight conserved genes were identified among Glu-D3 alleles, except for Glu-D3f. The protein products of the unique active genes in each NIL were detected using protein electrophoresis. Among Glu-3 alleles, the Glu-A3e genotype without i-type LMW-GS performed worst in almost all quality properties. Glu-B3b, B3g and B3i showed better quality parameters than the other Glu-B3 alleles, whereas the Glu-B3c allele containing s-type genes with low expression levels had an inferior effect on bread-making quality. Due to the conserved genes at Glu-D3 locus, Glu-D3 alleles showed no significant differences in effects on all quality parameters. Conclusions This work provided new insights into the composition and function of 18 LMW-GS alleles in bread wheat. The variation of i-type genes mainly contributed to the high diversity of Glu-A3 alleles, and the differences among Glu-B3 alleles were mainly derived from the high polymorphism of s-type genes. Among LMW-GS alleles, Glu-A3e and Glu-B3c represented inferior alleles for bread-making quality, whereas Glu-A3d, Glu-B3b, Glu-B3g and Glu-B3i were correlated with superior bread-making quality. Glu-D3 alleles played minor roles in determining quality variation in bread wheat. Thus, LMW-GS alleles not only affect dough extensibility but greatly contribute to the dough resistance, glutenin macro-polymers and bread quality.


Background
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-, γand 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, Bgroup, 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 macropolymers (GMP), dough properties, and pan bread making quality. These NILs were investigated using SDS-PAGE, the LMW-GS gene marker system and twodimensional 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.

Separation of LMW-GS proteins in Aroona NILs using SDS-PAGE
The glutenin alleles in the flour of 16 Aroona NILs were separated by SDS-PAGE ( Figure 1a). Aroona possessed five HMW-GS proteins, viz., 1, 7 + 9 and 2 + 12, encoded by genes at Glu-A1, Glu-B1 and Glu-D1 loci, respectively. All NILs had the same HMW-GS as Aroona, and their unique LMW-GS and gliadin bands were labeled in Figure 1a. Four Glu-A3 NILs possessed unique LMW-GS bands at the B-group region from Aroona (Glu-A3c). Aroona-Glu-A3d and Aroona-Glu-A3e also contained specific gliadin bands (Figure 1a). Among eight Glu-B3 NILs, Aroona (Glu-B3b), Aroona-Glu-B3a, B3f and B3g shared similar B-group proteins, whereas Aroona-Glu-B3c, B3d, B3h and B3i showed another group of electrophoretic patterns (Figure 1a). Among the latter four NILs, Aroona-Glu-B3c possessed the lowest quantity of B-group LMW-GS proteins (Additional file 1: Figure S1), especially the protein with the largest molecular weight. Four Glu-D3 NILs and Aroona shared the same B-group LMW-GS proteins (Figure 1a). Aroona-Glu-D3b and D3d had the same protein bands and contained one unique LMW-GS from Aroona. Except for one gliadin protein, the electrophoretic pattern of Aroona-Glu-D3a was the same as those of Aroona-Glu-D3b and D3d. Aroona-Glu-D3f produced a quite different protein pattern in the C-group region compared with other Glu-D3 NILs. Generally, the Glu-3 loci were tightly linked with Gli-1 loci [6][7][8], and it was difficult to break this linkage through genetic recombination in conventional crosses. Thus, each Glu-3 NIL not only possessed unique LMW-GS but also contained 1 or 2 specific gliadins.

Comparison of whole proteins among LMW-GS NILs using 2D-PAGE
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.

Quality properties of the LMW-GS NILs
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).
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.

Discussion
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.

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.

LMW-GS alleles and bread-making quality
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).

Glu-A3 alleles
Among five Glu-A3 alleles, Glu-A3e (A3-391, A3-400, A3-502-3 and A3-646) performed the poorest in almost all quality properties (Table 1, Figure 6; Additional file 2: Tables S2, S3 and S4). This was consistent with previous reports in which Glu-A3e was associated with lower extensibility and Rmax than Glu-A3d, A3b and A3c [33]. The negative effect of Glu-A3e on dough rheological properties was reported in several studies previously [24,33,37,38]. No unique i-type protein band was detected from the Glu-A3e allele using SDS-PAGE (Figure 1a) [39]. The protein product of the A3-646 gene was also not identified in 2D-PAGE although the i-type gene A3-646 in Aroona-Glu-A3e contained the intact ORF (Figure 3b, Additional file 1: Figure S2). Less i-type proteins and more gliadins in Glu-A3e increased the ratio of gliadin to glutenin and greatly reduced %UPP ( Figure 6; Additional file 2: Table S4), resulting in the worst performance of the Glu-A3e genotype in dough resistance and extensibility and pan bread total score [39]. The other Glu-A3 NILs produced similar quality parameters, including Rmax, ST, Ext and %UPP. These data indicated that Glu-A3a, A3c, A3d and A3f all had equivalent positive effects on UPP content and dough resistance and extensibility ( Figure 6). Among them, Glu-A3d had a significant effect on high ZSV, which was consistent with results from Xiaoyan 54 and Jing 411 RILs [18]. Some other studies also reported that the Glu-A3d allele had a superior effect on dough strength [40][41][42]. Compared with only one active gene at the Glu-A3 locus in the other alleles, the Glu-A3d allele possessed three active LMW-GS genes and produced the highest ZSV, Rmax, and %UPP ( Figure 6). The large number of active genes in Glu-A3d might be the basis of the superior performance in wheat quality properties [18].

Glu-D3 alleles
The five Glu-D3 alleles produced similar values for almost all quality properties in the present study ( Figure 6; Table 2), which confirmed previous reports that Glu-D3 alleles produced similar Rmax and extensibilities among large collections of wheat varieties [24,33,48], although some studies indicated different effects of Glu-D3 alleles on dough strength or mixing properties [27,29,38,49]. Glu-D3a, D3b, D3c and D3d contained similar 2D-PAGE spot patterns, and all six active genes were highly conserved among the alleles (> 99% identities). Their similarity in LMW-GS genes and whole proteins was consistent with their equivalent effects on all quality properties. Although Glu-D3f allele lacks two LMW-GS proteins and three gliadin spots, it produced similar quality properties to the other Glu-D3 alleles ( Figure 5; Table 2). These results suggested that D3-385, D3-432 and the three missing gliadins were not related to quality improvement. However, the lack of active genes D3-394 and D3-528 at the Glu-D3 locus in Jing 411 (Glu-D3l) showed significant negative effects on ZSV [18]. Thus, except for Glu-D3l, Glu-D3 alleles appeared to play only minor roles in determining quality variation among bread wheat varieties, and they should be given the lowest priority among LMW-GS alleles in selecting for improved bread-making quality in wheat.

Conclusion
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.

Plant materials
The wheat variety Aroona and 15 near isogenic lines (NILs) were kindly provided by Dr. Marie Appelbee and Prof. Ken Shepherd, SARDI Grain Quality Research Laboratory, South Australia. Each NIL contains a unique LMW-GS allele from a donor variety added to Aroona (Additional file 2: Table S5). They were planted at the Xinjiang Academy of Agri-Reclamation Sciences, Shihezi, and Xinjiang Academy of Agricultural Sciences, Urumqi, Xinjiang province, in randomized complete blocks with two replications during the 2010 cropping season.

Isolation and separation of LMW-GS proteins
Glutenin extraction was performed according to the method described by Singh et al. [11]. These proteins were separated by SDS-PAGE using the method described by Sunbrook and Russell [53]. Whole seed proteins were isolated from wheat flour based on the SDS/Phenol method [54] with some modifications. Briefly, proteins in 0.12 g flour were precipitated with 10% TCA/acetone at −20°C overnight. After centrifuging at 20,000 g at 4°C for 15 min, the pellet was washed three times with 80% acetone, then dried at 50°C. Whole proteins were extracted with SDS/ Phenol buffer (50% Tris-phenol pH 8.0, 30% sucrose, 2% SDS, 0.1 M Tris-HCl pH 8.0 and 2% DTT). The upper phenol phase was transferred into a new 2 mL tube. A fivefold volume of methanol containing 0.1 M ammonium acetate was added to the tube. The proteins were deposited at −20°C for 10 min or overnight. After centrifuging at 20,000 g for 5 min at 4°C, the pellet was washed once with 100% methanol and twice with 80% acetone before briefly drying in air. The proteins were dissolved in isoelectric focusing (IEF) sample extraction solution and used in 2D-PAGE analysis according to Dong et al. [18]. The images of SDS-PAGE were analyzed using NIH ImageJ software program (http://rsb.info.nih.gov/ij/).

Quality testing and evaluation of pan bread
Measures of grain hardness, protein content, Zeleny sedimentation values, Farinograph and Extensograph parameters, and pan bread qualities were performed by methods reported in He et al. [55]. The glutenin macropolymer compositions were measured following Zhang et al. [56].

Statistical analysis
The SAS statistical package (SAS Institute, Cary, NC) was used for data analysis. All statistical analyses were based on averaged data from two locations.

Additional files
Additional file 1: Figure S1. Quantitative gel densitometry measurements of B-group proteins in Glu-B3 NILs as determined using NIL ImageJ software program. Data points were normalized with respect to the Aroona value. Figure S2: Sequence alignments of i-type proteins identified in Aroona NILs. Figure S3: Sequence alignments of the B3-530 protein and its allelic variants identified in Aroona NILs. Figure S4: Sequence alignments of the B3-544/593/601/607 proteins identified in Aroona NILs. Figure S5: Sequence alignments of the B3-621/624 and B3-688 proteins identified in Aroona NILs. Figure S6: The differential display of B3-688 spots among Aroona-Glu-B3c, B3d and B3h. Spot volume values are expressed in percentages (%vol) of the total proteome. Additional file 2: Table S1. F values of one way ANOVA of wheat quality properties by locations and genotypes.