Novel insights into the composition, variation, organization, and expression of the low-molecular-weight glutenin subunit gene family in common wheat.

Low-molecular-weight glutenin subunits (LMW-GS), encoded by a complex multigene family, play an important role in the processing quality of wheat flour. Although members of this gene family have been identified in several wheat varieties, the allelic variation and composition of LMW-GS genes in common wheat are not well understood. In the present study, using the LMW-GS gene molecular marker system and the full-length gene cloning method, a comprehensive molecular analysis of LMW-GS genes was conducted in a representative population, the micro-core collections (MCC) of Chinese wheat germplasm. Generally, >15 LMW-GS genes were identified from individual MCC accessions, of which 4-6 were located at the Glu-A3 locus, 3-5 at the Glu-B3 locus, and eight at the Glu-D3 locus. LMW-GS genes at the Glu-A3 locus showed the highest allelic diversity, followed by the Glu-B3 genes, while the Glu-D3 genes were extremely conserved among MCC accessions. Expression and sequence analysis showed that 9-13 active LMW-GS genes were present in each accession. Sequence identity analysis showed that all i-type genes present at the Glu-A3 locus formed a single group, the s-type genes located at Glu-B3 and Glu-D3 loci comprised a unique group, while high-diversity m-type genes were classified into four groups and detected in all Glu-3 loci. These results contribute to the functional analysis of LMW-GS genes and facilitate improvement of bread-making quality by wheat molecular breeding programmes.

Introduction

Common wheat (Triticum aestivum L.) is one of the ‘big three’ cereal crops used for human food (Shewry, 2009) since wheat grains confer their viscoelastic properties to wheat dough (Shewry ). These viscoelastic properties allow dough to be incorporated into a wide range of daily food products, which are affected by glutenin and gliadin proteins in wheat seeds (Shewry ; D’Ovidio and Masci, 2004; Juhász and Gianibelli, 2006). Glutenin proteins are composed of two groups of subunits, namely high-molecular-weight and low-molecular-weight glutenin subunits (HMW-GS, 65–90kDa; LMW-GS, 30–45kDa) (Payne, 1987; D’Ovidio and Masci, 2004). The LMW-GS account for about one-third of the seed protein and 60% of glutenin proteins, and play an important role in determining dough properties and the quality of wheat food products (Gupta , 1994; Branlard ; Eagles ; Howitt ). Thus, elucidating the composition and variation of LMW-GS genes in common wheat and investigating the relationship between allelic variants and end-use quality are of interest for wheat quality improvement (Gupta ; He ; Bekes ; Juhász and Gianibelli, 2006; Liu ; Zhang ).

LMW-GS genes form a multigene family in common wheat, generally located at the Glu-A3, Glu-B3, and Glu-D3 loci on the short arms of homoeologous group 1 chromosomes (Jackson ). The copy number of LMW-GS genes was estimated to range from 10–20 to 30–40 (Ikeda ; D’Ovidio and Masci, 2004; Juhász and Gianibelli, 2006; Huang and Cloutier, 2008; Dong ; Zhang , ). In Norin 61, 12 groups of LMW-GS genes were identified by screening the cDNA library (Ikeda ). In Glenlea, among the 12 active genes, one was assigned to chromosome 1A, two to chromosome 1B, and nine to chromosome 1D (Huang and Cloutier, 2008). In Xiaoyan 54, 14 unique LMW-GS genes were identified using BAC (bacterial artificial chromosome) library screening and proteomics analysis, of which four were located at Glu-A3, three at Glu-B3, and seven at Glu-D3. Of the 11 active genes, two, two, and seven were i-, s-, and m-type genes, respectively (Dong ). The above three varieties contained different LMW-GS gene compositions, suggesting that this gene family has high molecular diversity among wheat varieties (Dong ). Moreover, these LMW-GS proteins had similar physical and chemical properties and molecular weights to the gliadins, which are a type of alcohol-soluble, monomeric seed storage protein. The high copy number and their co-migration with gliadins by SDS–PAGE (sodium dodecyl sulfate polyacrylamide gel electrophoresis) and MALDI-TOF MS (matrix-assisted laser desorption ionization-time of flight mass spectrometry) made separation of LMW-GS proteins and isolation of all LMW-GS genes from a particular wheat variety difficult (Howitt ). Thus, characterization of the allelic variation of LMW-GS genes in wheat germplasm remains challenging.

To dissect the LMW-GS complex, a nomenclature system was developed based on their relative mobility in SDS-PAGE (Singh ). Recently, their encoding genes were isolated using PCR with gene-specific primers. One m- and two i-type LMW-GS genes (GluA3-1, GluA3-2, and GluA3-3) were isolated from each Glu-A3 allele (Wang ). At the Glu-B3 locus, four LMW-GS genes (GluB3-1, GluB3-2, GluB3-3, and GluB3-4) and their allelic variants were isolated from nine Glu-B3 alleles (Glu-B3a–Glu-B3i) (Wang ). Also, six Glu-D3 genes were identified from individual wheat varieties containing Glu-D3a–Glu-D3e (Zhao , 2007). Meanwhile, allele-specific markers were developed to discriminate LMW-GS genes and their allelic variants in common wheat (Zhao ; Appelbee ; Wang , 2010). These markers facilitated identification of the known Glu-A3 and Glu-B3 alleles used in breeding programmes (Liu ). However, Glu-D3 genes were highly conserved, and allelic identification using PCR markers was difficult (Liu ). In contrast, these allele-specific primers characterized only one or two genes in individual alleles, and could not be used to determine the exact composition of LMW-GS genes in individual varieties.

To determine the composition of LMW-GS genes in individual wheat varieties, the LMW-GS gene molecular marker system and the full-length gene-cloning method were developed (Zhang , ), which enabled identification and characterization of the complete sequences of all LMW-GS genes in any wheat variety. In the present study, using both methods, LMW-GS genes were investigated in the micro-core collections (MCC) of Chinese wheat germplasm, which covers >70% of the genetic diversity of Chinese wheat germplasm (Hao ). The composition, organization, allelic variation, and expression of LMW-GS genes in 262 MCC accessions were comprehensively investigated.

Materials and methods

Wheat germplasm

The MCC of Chinese wheat germplasm were obtained from the Institute of Crop Science, Chinese Academy of Agricultural Sciences (CAAS). This was a representative sample of Chinese wheat diversity. This collection consisted of 262 accessions including 88 modern varieties, 157 landraces, and 17 foreign varieties, which accounted for >70% of the genetic diversity of the national collection.

LMW-GS gene analysis

Genomic DNAs of 262 MCC accessions were extracted from young leaves of seedlings with the cetyltrimethyl ammonium bromide (CTAB method) following Saghai-Maroof . The LMW-GS genes in all the MCC accessions were separated using the LMW-GS gene molecular marker system (Zhang ). Based on data from the marker system, 45 accessions containing almost all allelic variants were selected for RNA analysis. Total RNA was prepared from developing seeds at 15–21 dpa (day post anthesis) using TRIzol® Reagent, according to the manufacturer’s protocol (Invitrogen, Carlsbad, CA, USA). The RNA was converted into cDNA using Moloney murine leukemia virus (M-MLV) reverse transcriptase (Promega, Madison, WI, USA). The expressed LMW-GS genes were detected using the molecular marker system (Zhang ). To obtain the full-length sequence of these LMW-GS genes, 30 representative varieties containing all the main allelic variants of each gene were selected. All genes were cloned and identified using the full-length gene cloning method (Zhang ). To clone rare allelic variants, gene-specific primers were developed (Supplementary Table S1 available at JXB online). Sequence analysis and characterization were performed using Lasergene software (DNAStar; http://www.dnastar.com/), ClustalW2 (http://www.ebi.ac.uk/Tools/msa/clustalw2/), and MEGA 5 software (Kumar ).

Results

Composition and variation of LMW-GS genes in MCC accessions

LMW-GS genes in the MCC of Chinese wheat germplasm were amplified using the LMW-GS gene molecular marker system consisting of three independent sets of conserved primers (Zhang ). In general terms, at least 15 LMW-GS genes were identified in each wheat variety (Table 1). To characterize these LMW-GS genes further, 30 representative accessions containing the main allelic variants were selected, and all of their LMW-GS genes were cloned and sequenced using the full-length gene cloning method and gene-specific primers (Table 1; Supplementary Table S1 at JXB online) (Zhang ). In total, 466 LMW-GS gene sequences were identified and deposited in GenBank (JX877778–JX878243).

Table 1.

LMW-GS genes and their allelic variants identified from MCC accessions using the LMW-GS gene molecular marker system.

Genes and allelic variantsa				LMWGS1b				LMWGS2b				LMWGS3b
A3-391	A3-353	A3-370	A3-373	391.8	353.1	370.8	373.2	484.8	460.4	477.5	480.1	375.7	350.2	368.8	371.2
A3-400	A3-374	A3-388	A3-394	400.1	374.5	388.1	Nc	506.3	481.4	494.7	N	399.1	N	387.2	N
	A3-402	A3-408	A3-411		402.8	408.3	411.1		509.5	514.7	517.6		402.5	408.5	411.4
A3-502	A3-480	A3-484	A3-487	N	N	484.3	N	N	N	590.2	N	538.1	513.8	N	522.7
	A3-508				N				N				543.8
A3-565				565.3				671.4				606.7
A3-568	A3-567			568	567.2			674.1	673.2			609.6	607.2
A3-620	A3-590	A3-626	A3-640	620.1	589.5	626.2	N/640.6	725.3	693.7	730.5	N/745.4	666.6	632.3	673.3	619.4/688.5
	A3-643	A3-646	A3-649		643.2	646	649.5		N	750.6	753.6		691	695.6	698.6
	A3-662				662.6				766.8				715.8
B3-530	B3-510	B3-516		530.8	510.5	516		636.9	616.6	622.4		538.2	516.7	522.4
B3-548	B3-557			548.8	557.6			654.8	663.7			556.5	565.7
B3-570				570.8				676.2				N
B3-578				577.7				683.4				587.8
B3-544	B3-590	B3-593	B3-596	544.2	590.4	593.4	596.2	650.2	696.1	699	701.6	554.7	605.6	608.7	611.8
	B3-598	B3-601	B3-604		598.9	601.7	604.6		704.5	706.8	709.6		615.1	618.4	621.5
	B3-607	B3-610			607.4	610.2			712.4	715.6			625	628.1
B3-688	B3-621	B3-624		688.2	621.1	624.5		791.9	726.1	729.1		707.5	635.3	638.5
B3-691	B3-813			691.3	813.8			794.6	916.3			710.4	855.6
D3-385				385.7				492.3				383.6
D3-393	D3-385’			393.1	N			499.5	N			390.7	N
D3-394	D3-397			394.6	397.3			500.7	503.3			393.5	395.8
D3-441	D3-432	D3-444		441.5	432.7	444.4		547.5	538.7	550.3		441.8	432.6	444.6
D3-525	D3-522	D3-528		525.4	522.4	528.2		631.8	628.9	634.5		532.8	529.3	535.9
D3-575				575				681				584
D3-578				577.7				683.4				587.8
D3-586	D3-583	D3-589	D3-591	586.5	583.6	589	591.8	691.3	688.5	694	696.8	597.9	594.6	601	604.3
	D3-594	D3-597			594.8	597.5			699.6	702.6			607.6	610.5

LMW-GS genes and allelic variants were named according to the size of the DNA fragments amplified using the primer LMWGS1. For each gene, the major allelic variant was designated as the LMW-GS gene and the remainder as its variants.

Three sets of conserved primers in the LMW-GS gene molecular marker system (Zhang ).

Not detected with the specific primers.

For each LMW-GS gene, several allelic variants were detected from the MCC (Table 1), some of which were identified previously (Ikeda ; Zhao , 2007; Huang and Cloutier, 2008; Wang , 2010; Dong ; Zhang , ). Using available mapped LMW-GS genes and their allelic relationship with the genes identified from the MCC, all these genes were assigned to specific wheat chromosomes. In individual accessions, 4–6 genes were located at the Glu-A3 locus, 3–5 at the Glu-B3 locus, and eight at the Glu-D3 locus (Table 1). These genes were named according to their DNA fragment size and chromosomal location. For example, the gene corresponding to DNA fragment 441.5, located at the Glu-D3 locus, was designated D3-441. Most of these genes contained several allelic variations across the MCC. To simplify description, the major allelic variant was selected to represent each LMW-GS gene and it was named according to the following scheme: ‘DNA fragment+gene’ (e.g. the A3-402 gene), while the allelic variants of each gene were named according to ‘DNA fragment+allele’ (e.g. the A3-402 allele).

LMW-GS genes at the Glu-A3 locus

At the Glu-A3 locus, 4–6 LMW-GS genes were detected in each accession and several allelic variants were identified for each gene in the MCC (Table 1; Fig. 1). With regard to the A3-391 gene, five allelic variants, A3-353, A3-370a, A3-370b, A3-373, and A3-391, shared >97% identity (Supplementary Fig. S1 at JXB online). The A3-391 allele predominated in 196 accessions, while A3-353 and A3-373 were rare variants, each present in only two MCC accessions (Fig. 1a). Sequence analysis of the A3-391 genes of 30 varieties confirmed that allelic variants showed length polymorphisms in the repetitive regions, and that each variant contained its own single nucleotide polymorphisms (SNPs) (Supplementary Fig. S1). A3-353, A3-373, and A3-391 were highly conserved across the MCC population, whereas the A3-370 allele could be further divided into two variants (A3-370a and A3-370b) due to SNPs in all available sequences (Supplementary Fig. S1). A3-353, A3-370a, A3-370b, and A3-391 alleles contained immature stop codons, and only the rare allele A3-373 possessed an intact open reading frame (ORF) encoding an m-type subunit (Supplementary Table S2). Thus, the A3-391 gene was universal in common wheat, even though only five sequences with >98% identities were deposited in GenBank.

Fig. 1.

Composition of LMW-GS genes at the Glu-3 loci in micro-core collections (MCC) of Chinese wheat germplasm. The diagrams illustrate the LMW-GS genes and their allelic variants at the Glu-A3, B3, and D3 loci identified from MCC accessions. The horizontal axis of each diagram shows the allelic variants of individual genes or haplotypes identified from the MCC. The vertical axis displays the composition of unique genes and haplotypes in individual accessions. The length of line segments represents the number of accessions containing the corresponding allelic variants. Underlined allelic variants were rare in the MCC, and allelic variants in red were active in common wheat. The allelic variants indicated by numbers and letters, such as 370a and 370b, shared the same DNA fragment but had different nucleotide sequences. (A) The compositions of genes and their allelic variants at the Glu-A3 locus. The i-type genes showed high diversity and were tightly linked, forming specific haplotypes. (B) The compositions of genes and their allelic variants at the Glu-B3 locus. The s-type genes were tightly linked and formed specific haplotypes. (C) The composition of eight genes and their allelic variants at the Glu-D3 locus.

The A3-400 gene was also common in wheat varieties. Seven allelic variants with different repetitive region lengths, A3-374, A3-388, A3-394, A3-400, A3-402, A3-408, and A3-411, were identified from MCC accessions (Fig. 1a). Sequence alignments suggested that the A3-374, A3-388, A3-400, A3-408, and A3-411 alleles were conserved among wheat varieties, whereas both A3-394 and A3-402 comprised two variants, A3-394a and A3-394b, and A3-402a and A3-402b, respectively, due to indels and SNPs in the available sequences (Supplementary Fig. S2 at JXB online). Sequence analysis also demonstrated that A3-402a and A3-400 shared the same SNP, containing a premature stop codon, while all remaining allelic variants (i.e. A3-374, A3-388, A3-394a, A3-394b, A3-402b, A3-408, and A3-411) contained intact coding sequences that might encode m-type LMW-GS in common wheat (Supplementary Fig. S2).

Except for the m-type genes above, all the others identified at the Glu-A3 locus were i-type genes (Fig. 1a; Supplementary Table S2 at JXB online). The coding sequences of alleles A3-480, A3-484, A3-487, A3-502, and A3-508 contained a specific length of repetitive regions and unique SNPs. For the major variant A3-502, eight allelic variants (A3-502a–A3-502h) were recognized due to SNPs and indels in the coding sequences (Fig. 1a; Supplementary Fig. S3). For the other i-type genes, eight genes/haplotypes, A3-620, A3-626, A3-643, A3-646, A3-649, A3-573/A3-640, A3-567/A3-590, and A3-565/A3-568/A3-662, were detected, of which alleles A3-626 and A3-646 were further divided into two allelic variants, respectively, due to SNPs in the coding sequences. Moreover, two genes, A3-649-1 and A3-649-2 with unique SNPs and indels, which shared a 649bp DNA fragment, were identified from individual accessions; the A3-567-1 and A3-567-2 genes exhibited similar commonality. After analysing the composition of i-type genes in 262 MCC accessions, it was found that all A3-502 variants were tightly linked with other unique i-type genes. The A3-502a and A3-502b alleles were coupled with A3-620, A3-502c with A3-626a, A3-502d with A3-643, A3-502e with A3-646a, A3-502f with A3-646b, A3-502g with A3-573 and A3-640, and A3-502h with A3-649-1 and A3-649-2. In total, 12 i-type haplotypes were identified in MCC accessions (Fig. 1a).

LMW-GS genes at the Glu-B3 locus

Generally, excluding 11 1BL/1RS translocation lines, 3–5 Glu-B3 genes were identified in each MCC accession, and B3-530 and B3-548 genes were universal (Fig. 1b). With regard to the B3-530 gene, three allelic variants, B3-510, B3-516, and B3-530, were identified (Table 1; Fig. 1b), and shared high sequence identity (>99%). The B3-530 allele was further divided into three variants (B3-530a, B3-530b, and B3-530c) due to the unique SNPs (Supplementary Fig. S4 at JXB online). All allelic variants of the B3-530 gene possessed intact ORFs and their deduced proteins belonged to m-type LMW-GS. For the B3-548 gene, the B3-548 allele predominated in 241 accessions, while the rare allelic variant B3-557 was detected in only one (Table 1; Fig. 1b). However, both variants contained premature stop codons in their coding sequences, suggesting them to be pseudogenes.

The other Glu-B3 genes were identified only in partial MCC accessions (Fig. 1). The B3-570 gene was identified in 44 MCC accessions, and its intact ORF contained 344 amino acid residues and encoded an m-type gene with the novel N-terminal sequence, METSQIPGLEKPS. The B3-578, B3-621, and B3-544 genes were tightly linked at the Glu-B3 locus and formed several haplotypes (Fig. 1b). The B3-578 gene was classified into two variants (B3-578a and B3-578b) based on two SNPs. The B3-544 genes were conserved among accessions, and the allelic variants B3-544 and B3-587–B3-607 shared >99% sequence identity (Supplementary Fig. S5 at JXB online). Also, B3-621 and B3-624 differed in only a CAA indel and an SNP (Supplementary Fig. S6). Moreover, B3-578 contained two premature stop codons in its coding sequences, while the other genes possessed intact ORFs, which might be active in common wheat. Sequence analysis showed that all three genes were s-type, the deduced protein sequences of which contained a MENSHIPGLERPS peptide at the N-terminus. B3-688 could be divided into three groups (B3-688a, B3-688b, and B3-688c) in MCC accessions, although several irregular SNPs were found in the coding sequences (Supplementary Fig. S6). These irregular SNPs were tightly linked with B3-691 or B3-813 at the Glu-B3 locus, forming different haplotypes, namely B3-688a/B3-691, B3-688b/B3-813, and B3-688c/N (Fig. 1b). B3-688a was coupled with B3-691, and shared >99% identity, the only differences being a CAA indel and an SNP. B3-813 was identified for the first time in common wheat and contained a premature stop codon in the ORF, while both B3-688 and B3-691 had unbroken ORFs encoding s-type LMW-GS in common wheat.

LMW-GS genes at the Glu-D3 locus

Eight LMW-GS genes (D3-385, D3-393, D3-394, D3-441, D3-525, D3-575, D3-578, and D3-586) were detected at the Glu-D3 locus in individual wheat varieties (Table 1; Fig. 1c). Neither D3-385 nor D3-575 preserved any allelic variants and were universal in all MCC accessions. In terms of the D3-393 gene, the D3-393 allele was present in 259 MCC accessions, while the other rare allelic variant, D3-385’, was found in only three (Fengkang 2, Guinong 10, and Lovrin 10). For the D3-394 gene, both allelic variants, D3-394 and D3-397, shared >99% sequence identities, the difference being an indel (CAA) and two SNPs; the latter allele was detected in only a single landrace, Yizhimai. Three conserved allelic variants of D3-441 (D3-432, D3-441, and D3-444) were identified with CAA indels in the repetitive region (Supplementary Fig. S7 at JXB online). The three allelic variants of the D3-525 gene, namely D3-522, D3-525, and D3-528, exhibited a similar phenomenon (Supplementary Fig. S8). For the D3-578 gene, D3-578a was identical in length but had a different nucleotide sequence from that of the the D3-578b variant (Supplementary Fig. S9). This gene encoded the only s-type LMW-GS at the Glu-D3 locus in common wheat. The nucleotide sequences of D3-586 and its allelic variants were identical, except for the CAA indels in the repetitive region that contributed to the length polymorphism (Supplementary Fig. S10). Among the eight LMW-GS genes at the Glu-D3 locus, D3-393 was a pseudogene due to a frameshift mutation and D3-586 was a pseudogene due to a premature stop codon. The remaining six genes contained intact ORFs that encoded one s-type (D3-578) and five m-type LMW-GS at the Glu-D3 locus, which contained the largest number of active genes of all the Glu-3 loci.

Organization of LMW-GS genes in MCC accessions

Using the LMW-GS gene molecular marker system and the full-length gene cloning method (Zhang , ), almost all LMW-GS genes were detected in the MCC and were cloned and characterized in 30 MCC accessions, which facilitated investigation of the organization of LMW-GS genes and their linkage relationship.

At the Glu-A3 locus, 4–6 genes were generally isolated from individual MCC accessions, including A3-391, A3-400, and 2–4 i-type genes (e.g. A3-502a and A3-620; A3-502g, A3-573, and A3-640; and A3-484, A3-565, A3-568, and A3-662; Figs 1a, 2a). Although several allelic variants were identified for each gene, only 11 main types of Glu-A3 genotypes were detected in MCC accessions, suggesting that these genes were tightly linked (Fig. 2a). Two genotypes containing A3-391, A3-400 or A3-402a, A3-502, and A3-620 accounted for ~70% of the MCC accessions, while each of the remaining genotypes accounted for <7% (Fig. 2a). Moreover, concerning the distribution of different genotypes in foreign accessions, Chinese modern varieties, and landraces, genotypes containing the A3-649-1/-2 genes were detected only in landraces, while those containing A3-394a/b were present in both foreign accessions and Chinese modern varieties (Fig. 2a).

Fig. 2.

Main genotypes of LMW-GS genes at Glu-3 loci identified from the MCC. Genotypes present in more than three MCC accessions are shown. LMW-GS genes at the same locus were generally linked and formed limited types of genotypes. Genes in red were active in common wheat. Moreover, the modern varieties, landraces, and foreign varieties are indicated by different colours. The genotypes most common in landraces are indicated by red asterisks, and those found only in modern and foreign varieties by blue asterisks. (A) Eleven genotypes at the Glu-A3 locus. (B) Thirteen genotypes at the Glu-B3 locus. Of these, 11 accessions without LMW-GS genes belong to 1B/1R translocation lines. (C) Fourteen genotypes at the Glu-D3 locus.

Based on the genotype data of MCC accessions, the linkage relationship among LMW-GS genes at the Glu-A3 locus was analysed (Fig. 2a). First, four main haplotypes of i-type genes were detected in MCC accessions, namely A3-502a/b/A3-620, A3-502/A3-646, A3-502h/A3-649-1/A3-649-2, and A3-484/A3-565/A3-568/A3-662 (Figs 1a, 2a). The i-type genes were completely linked and formed specific haplotypes in common wheat. Secondly, the A3-391 allele was coupled with alleles A3-400 and A3-402a, and the A3-370a/b alleles co-segregated with A3-394b, A3-402b, A3-408, and A3-411 (Fig. 2a). m-Type genes may have been tightly linked with each other at the Glu-A3 locus. Thirdly, the A3-370a/b alleles were generally coupled with A3-502(e/f)/A3-646 and A3-484/A3-565/A3-568/A3-662, while the A3-391 allele was linked with A3-502(a/b)/A3-620, A3-502h/A3-649-1/A3-649-2, and A3-502g/A3-573/A3-640 (Fig. 2a). Thus, the A3-391 gene generally linked with i-type genes in MCC accessions.

At the Glu-B3 locus, although 3–5 genes were present in individual accessions, their allelic variants consisted of 12 main genotypes in the MCC accessions (Figs 1b, 2b). Genotypes containing allelic variants B3-621, B3-624, or B3-688 covered all MCC accessions, excluding 11 1BL/1RS translocation lines, and genotypes containing the haplotypes B3-688/N or B3-688/B3-691 accounted for 54% of the MCC (Fig. 2b). Tight linkage of LMW-GS genes at the Glu-B3 locus was also observed. The B3-510 allele was tightly coupled with the B3-570 gene, forming the haplotype B3-510/B3-570, which was generally linked with another haplotype B3-688/B3-691 (Fig. 2b). In contrast, s-type genes consisted of two groups of haplotypes at the Glu-B3 locus; one contained the B3-621 gene and the other the B3-688 gene (Fig. 1b). The former group formed various haplotypes, which contained three unique LMW-GS genes, namely B3-578, B3-544 and B3-621, as well as their allelic variants. Alleles B3-544, B3-601, B3-604, and B3-607 regularly co-segregated with B3-621, while the other variants B3-590, B3-593, and B3-596 were usually coupled with B3-624 (Fig. 2b). In terms of the distribution of these genes in MCC, 1BL/1RS lines and genotypes containing B3-601 and B3-604 were identified only in foreign or modern varieties, whereas the genotypes containing B3-624 and the genotype B3-530/B3-548/B3-688/B3-691 were detected mostly in landraces (Fig. 2b).

At the Glu-D3 locus, although eight LMW-GS genes were identified from individual accessions, few haplotypes were characterized throughout the entire MCC population because of the high conservation of LMW-GS genes at this locus (Fig. 2c). Among them, D3-578a was linked with the D3-432 allele, whereas D3-578b was generally coupled with D3-441 (Fig. 2c). The D3-586 gene showed high length polymorphisms among MCC accessions for six allelic variants D3-583/586/589/591/594/597, which resulted in detection of 14 main genotypes at the Glu-D3 locus in MCC accessions (Fig. 2c). Two genotypes, D3-385/D3-393/D3-394/D3-441/D3-525/D3-575/D3-578/D3-586 and D3-385/D3-393/D3-394/D3-441/D3-525/D3-575/D3-578/D3-589, differed only in the pseudogene D3-586, which accounted for 64.8% of the MCC accessions (Fig. 2c).

Expression of LMW-GS genes in the MCC accessions

Since only the active genes in this family affected bread-making quality, the mRNA of the developing seeds was investigated (Fig. 3) and active LMW-GS genes were identified by comparing the mRNA and genomic DNA data (Fig. 3).

Fig. 3.

Expression analysis of LMW-GS genes in MCC accessions. Electropherograms show the patterns of DNA fragments detected in genomic DNA and cDNA of individual accessions using the LMWGS1 primers. The results of four of 45 representative accessions are shown. The horizontal axis shows the size of the detected DNA fragments, while the vertical axis displays the signal intensity (i.e. the concentration of DNA fragments in the PCR products). The orange peaks indicate the GeneScan 1200 LIZ size standard fragments, while the blue peaks represent the DNA fragments in the PCR products. In the genomic DNA electropherogram, peaks indicated by arrows and numbers were detected only in genomic DNA, all of which were pseudogenes. Peaks in the cDNA electropherograms indicated by numbers correspond to active LMW-GS genes. Glu-A3 genes are shown in black, Glu-B3 genes in green, and Glu-D3 genes in pink.

At the Glu-A3 locus, among the 4–6 genes detected in genomic DNA (Fig. 2a), the A3-370a/b and A3-391 alleles were not detected in mRNA of developing seeds (Fig. 3), while their allelic variant A3-373 may have been active in the intact ORF. In terms of the A3-400 gene, the A3-400 and A3-402a alleles were inactive, whereas both A3-402b and A3-408 were expressed during seed filling (Fig. 3). Also, the rare allelic variants A3-374, A3-388, A3-394a/b, and A3-411 may encode LMW-GS proteins in their intact ORFs. None of the allelic variants of the A3-502 gene was expressed due to premature stop codons, except for the active allele A3-502h. Among the other i-type genes, A3-573, A3-620, A3-646, A3-649-2, A3-568, and A3-662 were detected in developing seeds (Fig. 3), and the rare allelic variants A3-626a/b, A3-643, A3-567-1, A3-567-2, and A3-590 contained intact ORFs, which might also be active in common wheat. Thus, in a particular wheat variety, 1–3 LMW-GS genes might be active at the Glu-A3 locus (Fig. 2a). For example, accessions with the genotype A3-391/A3-400/A3-502/A3-620 contained only one active LMW-GS gene, A3-620, while those with the i-type haplotype A3-484/A3-565/A3-568/A3-662 possessed three active genes at the Glu-A3 locus (Fig. 2a).

At the Glu-B3 locus, 3–5 genes were detected in the genomic DNA of individual accessions (Fig. 2b). Based on the electropherogram of the LMW-GS gene marker system, the B3-548, B3-578a/b, and B3-813 genes were not detected in developing seeds (Fig. 3), while the remaining m-type genes (B3-530 and the newly identified B3-570) and s-type genes (B3-544, B3-621, and B3-688) were generally expressed in wheat varieties, although B3-570 was only present in partial MCC accessions (Figs 2b, 3). With regard to active genes of different haplotypes, the number of active Glu-B3 genes varied from two to four in one wheat variety, including one or two m-type-encoding genes and one or two s-type-encoding genes (Fig. 2b).

At the Glu-D3 locus, of the eight LMW-GS genes detected in genomic DNA, D3-385, D3-394, D3-441, D3-525, D3-575, and D3-578 were detected in developing seeds of MCC accessions (Fig. 3), and the rare allelic variants D3-397, D3-444, and D3-522 might be active in their intact ORF. The remaining two genes, D3-393 and D3-586, were not identified in the developing seeds probably due to premature stop codons (Fig. 3). Thus, one s-type (D3-578) and five m-type genes (D3-385, D3-394, D3-441, D3-525, and D3-575) were generally expressed at the Glu-D3 locus in individual wheat varieties (Fig. 2c). These expression analyses revealed that individual accessions with different allelic variants or haplotypes might possess different numbers of active genes. Generally, none or one m-type and one or two i-type active genes were detected at the Glu-A3 locus, and one or two m-type and one or two s-type genes at the Glu-B3 locus were expressed, whereas one s-type and five m-type genes comprised the active genes at the Glu-D3 locus. Thus, the number of active genes in individual MCC accessions varied from nine to 13.

Characteristics of LMW-GS genes identified from MCC accessions

All proteins encoded by the genes identified in the present study were typical of LMW-GS, which had similar structures to previously characterized LMW-GS (D’Ovidio and Masci, 2004; Juhász and Gianibelli, 2006). Each deduced protein contained four main structural domains: a signal peptide, a short conserved N-terminal domain, a repetitive domain, and a C-terminal domain, except for i-type proteins, which lacked the N-terminal domain (Supplementary Fig. S11 at JXB online). Based on the N-terminal sequence of mature proteins, three types of LMW-GS (m-, s-, and i-types) were recognized. The m-type proteins were the most abundant in all genotypes analysed, and their molecular mass varied from 31.8kDa (D3-385) to 39.6kDa (D3-575). The s-type proteins generally had a higher molecular mass than did m-type subunits, which ranged from 37.0kDa (B3-544) to 42.5kDa (B3-691). Also, the i-type proteins had higher molecular weights (39.2–43.0kDa), despite lacking the N-terminal sequences.

Cysteine residues played a vital role in determining the structural and functional characteristics of wheat proteins (Shewry ; D’Ovidio and Masci, 2004). All deduced proteins identified in this study possessed eight cysteine residues, except the putative amino acid sequences from pseudogenes A3-502d and D3-385’, which contained seven and nine cysteine residues, respectively. However, both pseudogenes do not play a role in glutenin polymers and bread-making quality. The locations of the first (or third for i-type genes) and seventh cysteines were highly diverse, while the remaining six cysteines were conserved among all LMW-GS genes (Supplementary Fig. S11 at JXB online). Based on the relative locations of cysteines, LMW-GS proteins were divided into six groups (Supplementary Fig. S11).

All LMW-GS genes and their allelic variants were subjected to cluster analysis using ClustalW2 and MEGA 5. Also, the six main groups were further divided, which was consistent with the grouping data based on the cysteine positions of the deduced proteins (Fig. 4). The i-type genes located at the Glu-A3 locus formed a single group (iA), and all the s-type genes at the Glu-B3 locus and Glu-D3 locus were located in a single branch (sBD) (Fig. 4). All the remaining LMW-GS genes were m-type, which were further divided into four groups (Fig. 4). Variants of both the D3-441 and D3-525 genes formed single groups (mD-2 and mD-1, respectively), which were unique at the Glu-D3 locus. The mBD group was composed of three genes (B3-530, B3-548, and B3-570) from the Glu-B3 locus and two (D3-575 and D3-586) from the Glu-D3 locus. In contrast, the mAD group contained five m-type genes, two of which were located at the Glu-A3 locus and three at the Glu-D3 locus. Collectively, genes from the Glu-A3 locus contained all the i-type genes (iA) and m-type genes (mAD), genes from the Glu-B3 locus were associated with two groups, including s-type (sBD) and m-type (mBD) genes, and genes from the Glu-D3 locus were distributed into five of the six groups and showed higher diversity than those at the Glu-A3 and Glu-B3 loci (Fig. 4).

Fig. 4.

Phylogenetic reconstruction of all LMW-GS genes and their allelic variants identified from MCC accessions. The phylogenetic tree of LMW-GS genes was constructed using MEGA 5 (Kumar ). All LMW-GS genes were divided into six groups. The i-type genes located at Glu-A3 were special and formed a single group (iA). The s-type genes at Glu-B3 and Glu-D3 shared high identity and were located in a single branch (sBD). The other LMW-GS genes were of the m-type, and were divided into four groups (mAD, mD-1, mBD, and mD-2). LMW-GS genes at the Glu-D3 locus were assigned to five groups, and showed a higher diversity than those at the Glu-A3 and Glu-B3 loci. (This figure is available in colour at JXB online.)

Of the LMW-GS genes, i-type genes were the most complex and 12 haplotypes were detected, each of which contained unique LMW-GS genes. Sequence alignments and phylogenetic analysis of all i-type genes demonstrated that the A3-502 gene was conserved (>85% diversity) among the haplotypes, whereas the other genes could be divided into five subgroups (Supplementary Fig. S12 at JXB online). A3-626 and A3-643 shared high identity (>98%) and formed the subgroup iA-1 with the A3-502 gene. The haplotypes A3-502e/f and A3-646 were named iA-2. The A3-502g/A3-573/A3-640 and A3-502h/A3-649-1/A3-649-2 haplotypes contained three i-type genes and were classified into subgroups iA-3 and iA-4, respectively. A3-484/A3-565/A3-568/A3-662 and A3-487/A3-567-1/A3-567-2/A3-590 represented a unique i-type genotype (iA-5) in common wheat (Supplementary Fig. S12).

Discussion

LMW-GS genes in common wheat are complex, and their exact copy number remains unclear (Cassidy ; Ikeda ; Juhász and Gianibelli, 2006; Huang and Cloutier, 2008; Dong ). Recently, using BAC library screening, 14 and 19 genes were isolated from the common wheat varieties Xiaoyan 54 and Glenlea, respectively (Huang and Cloutier, 2008; Dong ). Meanwhile, LMW-GS genes at the Glu-A3, Glu-B3, and Glu-D3 loci were identified using gene-specific primers, which suggested that at least 12 genes are present in the common wheat genome (Zhao , 2007; Wang , 2010). Based on the conserved and polymorphic structures of these genes, the LMW-GS gene marker system and the full-length gene cloning method were developed, which can identify >15 members of this gene family in common wheat (Zhang , ). In the present study, both methods were used to investigate the MCC of Chinese wheat germplasm, and the complex LMW-GS gene family in common wheat was successfully dissected.

Dissection of LMW-GS genes at individual Glu-3 loci

Glu-A3 locus

Two m-type genes and 2–4 i-type LMW-GS genes were generally identified at the Glu-A3 locus, which was the highest number reported for this locus in individual wheat varieties (Figs 1a, 2a). The m-type gene, A3-391, and its allelic variants shared high identities (>99%) with a few sequences in GenBank derived from T. macha, T. durum, and T. timopheevii (Supplementary Table S2 at JXB online), which suggests that A3-391 might be widely present in Triticum. The other m-type gene at the Glu-A3 locus, A3-400, was reported by several groups, corresponding to GluA3-2 genes from Aroona near-isogenic lines (NILs) (Wang ), the group 6 type IV gene from Norin 61 (Ikeda ), and A3-1 from Xiaoyan 54 (Supplementary Table S2) (Dong ). The present results provided direct evidence for the presence of m-type genes at the Glu-A3 locus in common wheat, and this gene showed high diversity among MCC accessions with several novel allelic variants (i.e. A3-374, A3-388, A3-394a, and A3-411; Fig. 1a). Moreover, the new allelic variants, A3-388, A3-394a/b, A3-408, and A3-411, contained intact ORFs and may make specific contributions to wheat bread-making quality.

In the present study, using conserved primers, 2–4 i-type genes were identified in individual wheat varieties. (Supplementary Table S2 at JXB online). In previous studies, 1–3 i-type genes in only a few wheat varieties were characterized (Supplementary Table S2) (Zhang ; Ikeda ; Huang and Cloutier, 2008; Dong ), which made it difficult to analyse the relationships among haplotypes of these genes. Here, the MCC of Chinese wheat germplasm were investigated in terms of LMW-GS gene composition. The i-type genes were present in the wheat genome as haplotypes rather than single genes, and 12 haplotypes of i-type genes were detected in the MCC. Nucleotide sequence comparisons showed that genes in six of 12 haplotypes identified in this study were similar (>99%) to those isolated from seven Glu-A3 alleles, for example haplotype A3-502d/643 corresponding to GluA3-32/Glu-A3-12 from Glu-A3b (Supplementary Table S2) (Wang ; Zhang ). In addition, haplotype A3-484/A3-565/A3-568/A3-662 contained the three i-type genes identified in Norin 61 and Xiaoyan 54 (A3-2, A3-3, and A3-4), and the haplotype A3-502f/646b covered the i-type gene detected in Glenlea (EU189087) (Huang and Cloutier, 2008; Dong ). This confirms that i-type genes in common wheat exist as haplotypes at the Glu-A3 locus and exhibit high genetic diversity. The identification and characterization of these haplotypes will facilitate the functional analysis of i-type genes and the selection of specific genes using haplotype-specific markers.

Glu-B3 locus

Three to five Glu-B3 genes were detected in individual varieties, of which 1–3 were s-type and two or three were m-type (Figs 1b, 2b). The m-type gene B3-530 shared >99% sequence identity with GluB3-4 genes from Aroona and its near-isogenic lines, the B3-1 gene from Xiaoyan 54 and Jing 411, 1557N24-M from Glenlea, and the group 2 type I gene from Norin 61 (Supplementary Table S2 at JXB online) (Ikeda ; Huang and Cloutier, 2008; Wang ; Dong ), while B3-548 has been reported only rarely since it is a pseudogene. The third m-type gene, B3-570, was newly identified from wheat varieties and was detected in partial MCC accessions containing B3-510 (Fig. 2b). Thus, at least two m-type genes were present at the Glu-B3 locus, rather than the one reported previously (Ikeda ; Huang and Cloutier, 2008; Dong ). The other genes at the Glu-B3 locus were of s-type, and were divided into two subgroups based on their gene composition; one containing the B3-688 gene and the other containing B3-621 (Figs 1b, 2b). The B3-688 subgroup of s-type haplotypes corresponded to B3-2 from Xiaoyan 54 as well as Jing 411 and GluB3-3 from Aroona-Glu-B3c, B3d, B3h, and B3i (Supplementary Table S2) (Wang ; Dong ; Zhang ). The other subgroup contained two active genes, B3-544 and B3-621, and one pseudogene, B3-578. B3-544 had 99% sequence identity with GluB3-1 from Aroona-Glu-B3a, B3b, B3f, and B3g. Also, B3-621 genes were present in Aroona-Glu-B3a, B3b, B3f, and B3g, corresponding to GluB3-2 (Wang ; Zhang ) (Supplementary Table S2). Thus, s-type genes existed as haplotypes in common wheat. However, both subgroups of s-type genes displayed significant differences in terms of gene composition and sequences, and thus might make different contributions to dough quality. Overall, evaluation of Glu-B3 genes/haplotypes will enable development of haplotype-specific primers for marker-assisted selection.

Glu-D3 locus

One s-type and seven m-type genes were identified at the Glu-D3 locus from a single wheat genotype, which was by far the highest number of LMW-GS genes reported for this locus (Fig. 2c). Pseudogene D3-586 was newly detected in common wheat, whereas the other seven genes have been investigated extensively (Supplementary Table S2 at JXB online), and covered all Glu-D3 genes identified in wheat varieties, including Norin 61, Glenlea, Xiaoyan 54, Jing 411, and Aroona NILs (Supplementary Table S2) (Ikeda ; Zhao , 2007; Huang and Cloutier, 2008; Dong ; Zhang , ). These Glu-D3 genes were highly conserved, with only a few allelic variants (>99% identities), of which three novel active allelic variants (D3-397, D3-444, and D3-522) were detected with unique SNPs or indels in MCC, but functional analysis of these variants has been limited. Moreover, these Glu-D3 genes shared >97% identities with LMW-GS genes isolated from Aegilops tauschii (Johal ; Dong ), which further confirmed the conservation of Glu-D3 genes.

Relative genetic locations of LMW-GS genes at the Glu-3 loci

Typical LMW-GS genes were located at the Glu-A3, Glu-B3, and Glu-D3 loci on the homoeologous group 1 chromosomes. However, little is known about the relative location of LMW-GS genes at individual loci due to the complexity of gene composition and the lack of appropriate methods of investigating this gene family. Recently, the recombination of 14 LMW-GS genes at Glu-3 loci was analysed and the relative genetic position of these genes was determined (Dong ). Subsequently, four more genes were detected and sequenced in Xiaoyan 54 (Zhang , ). In the present study, based on the allelic relationship with the genes in Xiaoyan 54, all the LMW-GS genes in the MCC were located at a specific position in homoeologous group 1 chromosomes (Fig. 5). At the Glu-A3 locus, two groups of LMW-GS genes, mAD and iA gene clusters, were found and little recombination was detected within groups, of which the iA group was distal and the mAD group was proximal to the centromere (Figs 2, 5). At the Glu-B3 locus, although the relative position of A3-548 was not determined, the m- and s-type genes might exist as two gene clusters (mBD and sBD). Also, m-type genes were more proximal to the centromere than s-type genes (Figs 2, 5). At the Glu-D3 locus, tight linkage was identified only between D3-441 and D3-578a or D3-432 and D3-578b (Fig. 5), which could be explained by the close physical proximity (15.9kb) of these genes (Dong ). Additionally, D3-385, D3-393, D3-394, and D3-525 genes were located in close proximity and had a high identity (Figs 4, 5). Based on the location and sequence analysis, LMW-GS genes that show high identity or belong to the same group may be tightly linked and located at the same position in the Glu-3 loci.

Fig. 5.

Organization of the LMW-GS genes in homoeologous group 1 chromosomes. The relative locations of genes or haplotypes were determined based on Dong . The main allelic variants are displayed as representatives. The distances among genes do not represent the genetic or physical distances. (A) Relative genetic positions of LMW-GS genes at the Glu-A3 locus. The i-type genes were tightly linked at the Glu-A3 locus and formed five haplotype subgroups, and both m-type genes also generally co-segregated in the MCC. (B) Relative genetic positions of LMW-GS genes at the Glu-B3 locus. The s-type genes were coupled at the Glu-B3 locus and were of two principal haplotypes. (C) Relative genetic positions of the eight LMW-GS genes at the Glu-D3 locus. Of these, only D3-411 and D3-578 were tightly linked.

In addition, twelve i-type haplotypes at the Glu-A3 locus could be divided into five subgroups (iA-1 to iA-5; Fig. 5; Supplementary Fig. S12 at JXB online), and s-type haplotypes at the Glu-B3 locus were divided into two subgroups, revealing the high diversity of the i- and s-type genes/haplotypes among wheat varieties. These subgroups were significantly different in terms of gene numbers and sequences, which suggests that the Glu-A3 and Glu-B3 loci in common wheat might be derived from several unique ancestors or have been involved in significant mutational or recombination events during the course of their evolution (Figs 1b, 5).

The complex LMW-GS gene family in Chinese wheat germplasm

The LMW-GS gene family was investigated using the MCC of Chinese wheat germplasm, which consists of 262 accessions with an estimated 70% genetic diversity compared with the full collection (Hao ). Using the MCC, most (>15) LMW-GS genes were identified in individual wheat varieties. This allowed investigation of the classification and relationship of these genes in Chinese wheat germplasm.

The i-type genes were reported only at the Glu-A3 locus (Zhang ; Ikeda ; Huang and Cloutier, 2008; Dong ; Wang ; Zhang , ) or the A genome, for example T. urartu and T. monococcum (An ; Ma ; Caballero ; Long ). The findings also indicate that all i-type genes detected in MCC accessions were located at the Glu-A3 locus (Fig. 4). Since i-type genes lacked the sequences encoding the N-terminal domain, m- and s-type genes without the N-terminal domain-coding sequences were used for the phylogenetic analysis. It was found that i-type genes had a closer relationship with s-type genes (sBD) than m-type genes, excluding the mD-2 group (Supplementary Fig. S13 at JXB online). This results confirmed that the i-type genes may be the result of a deletion event of s-type genes (Gao ), and the i-type genes comprised a relatively young group of LMW-GS genes (Juhász and Gianibelli, 2006). The s-type genes were distributed at the Glu-B3 and Glu-D3 loci in common wheat and the progenitor of the wheat A genome, T. urartu (data not shown), but not at the Glu-A3 locus in common wheat (Fig. 4). This suggests that their disappearance from the Glu-A3 locus might be the result of elimination at the polyploid level. The m-type genes were common at the Glu-3 loci, and the difference between the s- and m-types was not significant (D’Ovidio and Masci, 2004). The D3-441 gene (mD-2) and s-type genes were located at the same main branch, although they belonged to different groups (Fig. 4). These data confirm that the s-type genes probably originated from m-type genes due to mutation of MET to MEN in the N-terminal region (Masci ; D’Ovidio and Masci, 2004). This also suggests that the m-type genes might be the oldest type of LMW-GS gene.

Genome sequence analysis of Glu-3 loci revealed that both i-type and s-type genes/haplotypes existed together with the Pm3 analogue and genetic marker SFR159, while most m-type genes, mAD, mD-1, and mBD, were tightly linked with another genetic marker, WHS179 (Wicker ; Gao ; Dong ). Moreover, at the Glu-A3 and B3 loci, i-type and s-type genes were distal, while the mAD and mBD groups of genes were proximal to the centromere (Fig. 5) (Dong ). At the Glu-D3 locus, the s-type gene D3-578 was also more distal to the centromere than the mBD gene, D3-575 (Fig. 5) (Dong ). Thus, the phylogenetic analysis (Fig. 4; Supplementary Fig. S13 at JXB online), together with the linked genes/markers and their relative locations on chromosomes, suggest that i-type and s-type genes may have been derived from similar ancestral genes, and the mAD (only Glu-A3 genes) and mBD groups of genes were orthologues among the A, B, and D subgenomes.

After sequence alignment and clustering analysis, the LMW-GS genes detected in MCC accessions were divided into six groups (Fig. 4). The genes at the Glu-A3 locus were assigned to the iA and mAD groups, and those at the Glu-B3 locus were divided into the sBD and mBD groups, whereas those at the Glu-D3 locus were distributed widely among the mD-1, mD-2, sBD, mBD, and mAD groups (Fig. 4). Thus, the Glu-D3 genes showed higher diversity than those at the Glu-A3 or Glu-B3 loci in individual wheat varieties. The Glu-A3 locus did not share the same group of LMW-GS genes as the Glu-B3 locus (Figs 4, 5), which suggested that genes at the Glu-A3 and Glu-B3 loci evolved through different routes and had a distant evolutionary relationship. All genes at the Glu-B3 locus and three Glu-D3 genes comprised two groups (sBD and mBD); these homoeoalleles showed close relationships between the Glu-B3 and Glu-D3 loci, which were consistent with the derivation of BB and DD subgenomes from Aegilops.

Analysis of the LMW-GS gene family was performed using MCC accessions, consisting of foreign varieties, Chinese modern varieties, and landraces. The novel i-type haplotype, A3-502h/A3-649-1/A3-649-2 was detected only in landraces and were absent from Chinese modern or foreign wheat varieties (Fig. 2a). This also occurred for genotypes containing B3-624 and the genotype B3-530/B3-548/B3-688/B3-691 (Fig. 2b). Although these haplotypes possessed an equal number of active genes to other haplotypes, they have not been selected for use in modern wheat breeding programmes. This may be because these genes or those linked to them have a detrimental effect on bread-making quality or yield potential. In contrast, A3-394a/b, B3-601/604, and the genotypes D3-385/D3-376/-/D3-432/-/D3-575/D3-578a/-, and 1BL/1RS lines were present only in modern or foreign varieties. The presence of these genes/genotypes in Chinese modern varieties might be the result of incorporation of foreign germplasm in breeding programmes in the past several decades. For example, the elite 1BL/1RS translocation lines were introduced into China in the 1970s and were exploited in modern wheat breeding since they contain several disease resistance and yield improvement genes. Also, the other genes/genotypes were probably introduced into Chinese wheat germplasm since they (or linked genes) increased the yield potential or bread-making quality.

Using the LMW-GS gene marker system and full-length gene cloning method, a representative population (MCC) was investigated, and the composition, organization, variation, and expression of LMW-GS genes were evaluated. Furthermore, the LMW-GS genes corresponding to all DNA fragments from the LMW-GS gene marker system were identified (Table 1; Supplementary S1 at JXB online). The expression profile of these genes was revealed by comparing the genomic DNA and cDNA data. These data will facilitate the update of the LMW-GS gene marker system which can be used to separate, identify, and characterize LMW-GS genes efficiently in common wheat.

Supplementary data

Supplementary data are available at JXB online.

Figure S1. Sequence alignments of the A3-391 gene identified in the MCC.

Figure S2. Sequence alignments of the A3-400 gene identified in the MCC.

Figure S3. Sequence alignments of the A3-502 gene identified in the MCC.

Figure S4. Sequence alignments of the B3-530 gene identified in the MCC.

Figure S5. Sequence alignments of the B3-544 gene identified in the MCC.

Figure S6. Sequence alignments of the B3-621 and B3-688 genes identified in the MCC.

Figure S7. Sequence alignments of the D3-441 gene identified in the MCC.

Figure S8. Sequence alignments of the D3-525 gene identified in the MCC.

Figure S9. Sequence alignments of the D3-578 gene identified in the MCC.

Figure S10. Sequence alignments of the D3-586 gene identified in the MCC.

Figure S11. Sequence alignments of the deduced proteins of 15 representative LMW-GS genes from MCC accessions.

Figure S12. Phylogenetic reconstruction of all i-type LMW-GS genes and their allelic variants identified from MCC accessions.

Figure S13. Phylogenetic reconstruction of all LMW-GS genes and their allelic variants with removed sequences coding for N-terminal domains.

Table S1. Gene-specific primers used for cloning rare allelic variants.

Table S2. Nucleotide sequence identities of LMW-GS genes from MCC to the previously reported Glu-A3, B3, and D3 alleles/genes.