Introduction

Gluten, the most important storage protein in the endosperm of common wheat (Triticum aestivum L.), comprises glutenins and gliadins (Lindsay and Skerritt 1999; Shewry and Halford 2002). Glutenins are separated into high-molecular-weight glutenin subunits (HMW-GS) and low-molecular-weight glutenin subunits (LMW-GS) according to their mobilities in sodium-dodecyl-sulphate polyacrylamide gel electrophoresis (SDS-PAGE) (Bietz et al. 1975; Payne and Corfield 1979). These proteins are held together by disulphide bonds to form gluten macropolymers (Gras et al. 2001), contribute to fundamental aspects of dough quality such as viscoelasticity and extensibility, and consequently influence the end-use products of wheat (Payne 1987; Luo et al. 2001). The identification of specific HMW-GS and LMW-GS alleles is, therefore, an important target in improving wheat quality (Gupta et al. 1999; Eagles et al. 2001; Gale 2005).

Allelic variation of the HMW-GS and its relationship with processing quality have been studied extensively, and PCR-based DNA markers are available to discriminate the important Glu-1 alleles Ax2*, Bx7, Bx7*, Bx17, By8, By9 and Dx5 (Ma et al. 2003; Butow et al. 2004; Gale 2005; Lei et al. 2006). Compared with HMW-GS, the extensive allelic variations of LMW-GS and their overlapping mobilities with the more abundant gliadin proteins (Singh and Shepherd 1988) make it difficult to discriminate the roles of individual LMW-GS in wheat quality. Gupta and Shepherd (1990) carried out an extensive survey of LMW glutenin proteins in common wheat cultivars by SDS-PAGE and detected 20 banding patterns. Subsequently, six protein alleles were found for the Glu-A3 locus (a, b, c, d, e, f), nine for the Glu-B3 (a, b, c, d, e, f, g, h, i) and five for the Glu-D3 (a, b, c, d, e). With respect to effects on dough quality, various Glu-3 alleles were ranked for R max (maximum dough resistance, an indicator of dough strength), and the rankings of alleles were b > d > e > c at the Glu-A3 locus, i > b = a > e = f = g = h > c at the Glu-B3 and e > b > a > c > d at the Glu-D3 (Gupta et al. 1989, 1991; Metakovsky et al. 1990). For dry white Chinese noodle (DWCN) quality, Glu-A3d and Glu-B3d were considered slightly better than others (He et al. 2005). Cornish et al. (1993) found that the composition bbb for Glu-A3, Glu-B3 and Glu-D3, respectively, gave the best extensibility, and the composition bbc was almost as extensible.

Traditionally, SDS-PAGE (Jackson et al. 1996) or RP-HPLC (Margiotta et al. 1993) is used to determine allelic compositions of LMW-GS in wheat. However, difficulties in resolving the multigene families and the overlapping fractions of LMW-GS hinder their routine use, particularly for testing large populations in the early generations of wheat breeding programs. Therefore, it is important to develop functional markers to identify different LMW-GS genes (Andersen and Lübberstedt 2003). Long et al. (2005) classified 69 LMW-GS genes registered in GenBank into nine groups and established nine group-specific primer sets to discriminate the nine groups. Ikeda et al. (2006) constructed 12 group-specific markers according to the 12 groups of LMW-GS genes detected in cultivar Norin 61. Based on the allelic variation of LMW-GS gene at the Glu-A3 locus, Zhang et al. (2004) developed several PCR markers to distinguish alleles a, b, c, d, e, f and g. Zhao et al. (2007a, b) designed several gene-specific markers for discriminating haplotypes of Glu-D3 genes. Three markers were developed for different Glu-B3 haplotypes at the DNA level (Zhao et al. 2007b). However, these markers do not discriminate the Glu-B3 alleles a, b, c, d, e, f, g, h and i, and the association between Glu-B3 genes and Glu-B3 protein alleles remained unclear. Here, we report the isolation of LMW-GS genes at the Glu-B3 locus from common wheat, characterization of the relationship between the genetic haplotypes and Glu-B3 alleles defined by protein mobility and development of allele-specific STS markers for different Glu-B3 alleles. This will benefit marker-assisted breeding for wheat quality.

Materials and methods

Plant materials

Aroona (Glu-B3b) and its seven near-isogenic lines (NIL), Aroona-B3a (Glu-B3a), Aroona-B3c (Glu-B3c), Aroona-B3d (Glu-B3d), Aroona-B3f (Glu-B3f), Aroona-B3g (Glu-B3g), Aroona-B3h (Glu-B3h) and Aroona-B3i (Glu-B3i), and Cheyenne (Glu-B3e), with different Glu-B3 alleles defined by protein mobility, were used to isolate Glu-B3 genes and develop molecular markers (Table 1). Certain homologous Group 1 Chinese Spring nulli-tetrasomic lines and five Chinese varieties with the 1BL.1RS translocation were used to confirm the chromosomal locations of identified genes. Twenty varieties from our laboratory collection and 141 wheat varieties and advanced lines from the International Maize and Wheat Improvement Centre (CIMMYT) with different Glu-B3 protein alleles detected by SDS-PAGE were employed to validate the allele-specific markers.

Table 1 Primers used for amplifying Glu-B3 genes
Full size table

DNA extraction and PCR amplification

Genomic DNA was extracted from seedlings or seeds using a modified CTAB procedure (Gale et al. 2001). PCR was performed using TakaRa Taq DNA polymerase (1.0 unit) in 20 μl reaction volumes containing approximately 50 ng of genomic DNA, 1× PCR buffer (1.5 mM MgCl2), 100 μM of each dNTPs and 10 pmoles of each PCR primer. PCR cycling conditions for gene-specific primers were 94°C for 5 min followed by 38 cycles at 94°C for 45 s, 56–61°C for 45 s, 72°C for 90 s and a final extension at 72°C for 8 min. PCR conditions for allele-specific markers are shown in Table 2.

Table 2 Allele-specific PCR markers for the discrimination of Glu-B3 alleles defined by protein mobility in common wheat
Full size table

Development of locus-specific primers for isolation of Glu-B3 genes

Locus-specific primers for cloning Glu-B3 genes were developed from the descriptions of Zhang et al. (2003, 2004) and Zhao et al. (2006). Eight reference genes with complete coding regions, including seven LMW-GS genes (GenBank accessions AB119006, AB164415, AB164416, AB262661, Y14104, AB062852 and AJ007746) located on the short arm of chromosome 1B, and one (AY542898) with high similarity to Glu-B3 available in GenBank, were used for primer development (http://www.ncbi.nlm.nih.gov). The genes AB062852, AB164416 and AB262661 were selected as probes for the in silico cloning of Glu-B3 genes according to He et al. (2007).

Based on the alignment of the reference genes, 63 primers were designed and 378 primer combinations were tested with the NILs and Cheyenne with different Glu-B3 alleles. Primer screening was conducted as described by Zhao et al. (2006). Finally, four Glu-B3 genes were isolated by four pairs of primers with annealing temperatures ranging from 56 to 61°C (depending on primer set) (Fig. 1). Primer sequences (5′–3′) and their relative positions in reference genes are shown in Table 1.

Fig. 1
figure 1

Electrophoresis of PCR products of four gene-specific primer sets on agarose gels. a GluB3-1, GluB3-4 and GluB3-2; 1 Chinese Spring, 2 N1AT1B, 3 N1BT1A, 4 N1BT1D, 5 N1DT1B. b GluB3-3; 1, 2, 5–7 1BL.1RS lines, 3 Aroona-B3c, 4 Aroona-B3d, 8 Aroona-B3h, 9 Aroona-B3i, 10 N1AT1B, 11 N1BT1A, 12 N1BT1D, 13 N1DT1B. M DNA Ladder 2000 (100, 250, 500, 750, 1,000 and 2,000 bp)

Full size image

Sequencing of PCR products

PCR fragments with expected sizes were recovered from agarose gels and cloned into the pGEM-T Easy Vector. Recombinant clones with expected sizes were sequenced after a PCR test. To eliminate cloning pitfalls (Masci et al. 1998), the recovered fragments were also sequenced directly using the corresponding PCR primers. Each PCR and sequence procedure was repeated three to six times to avoid any technical errors. All the sequencings were performed by the Sangon Biotechnology (Shanghai, China). Sequence analysis and characterization were performed using software DNAMAN (http://www.lynnon.com).

Allele-specific PCR marker design and validation

Allele-specific PCR markers were designed based on the allelic variants of Glu-B3 following the method of Zhang et al. (2003). These markers were firstly validated with the eight Aroona NILs and Cheyenne, and then with 161 wheat varieties and advanced lines from CIMMYT, Australia and France with Glu-B3 protein mobility alleles previously identified in SDS-PAGE by other workers.

Results

Allelic variants at the Glu-B3 locus

Four Glu-B3 genes, designated GluB3-1, GluB3-2, GluB3-3 and GluB3-4, including 17 allelic variants at the DNA level were identified at the Glu-B3 locus in the eight NILs and Cheyenne.

The GluB3-1 gene had five haplotypes or allelic variants, designated as GluB3-11, GluB3-12, GluB3-13, GluB3-14 and GluB3-15 (GenBank accessions EU369699, EU369700, EU369701, EU369702 and EU369703), amplified with the primer set LB1F/LB1R from Aroona-B3a (Glu-B3a), Aroona (Glu-B3b), Cheyenne (Glu-B3e), Aroona-B3f (Glu-B3f) and Aroona-B3g (Glu-B3g), respectively. Compared with GluB3-12, four, one and two triplet-nucleotide (CAA) deletions at positions 319–330, 328–330 and 325–330 in the coding region were found in GluB3-11, GluB3-13 and GluB3-14 (Appendix Fig. A1 in the Electronic Supplementary Material), leading to four, one and two glutamine deletions in the glutamine-rich repetitive domains of the deduced peptides BP1-1, BP1-3 and BP1-4, respectively (Fig. A2). At position 292–354, GluB3-15 had a 63-bp deletion, resulting in a 21-amino acid deletion in the deduced peptide BP1-5. GluB3-11 and GluB3-15 showed an additional 3-bp (CAA) deletion at position 556–558, leading to a glutamine deletion. In addition, GluB3-11 had two SNPs, one with A–G transition at position 1,092 and the other at position 1,113 with a C–T transition; GluB3-13 had a SNP at the position 69 with a C–A substitution; GluB3-15 showed four SNPs in the coding region at positions 360, 363, 918 and 1,056, respectively; the 360th SNP of GluB3-15 resulted in a change of phenylalanine to leucine at position 105 in the N-terminal domain of the deduced peptides, whereas all the others represented synonymous changes.

The GluB3-2 gene, amplified with primer set LB2F/LB2R, had three allelic variants at the DNA level, designated as GluB3-21, GluB3-22 and GluB3-23 (GenBank accessions EU369704, EU369721 and EU369705), respectively. GluB3-21 was detected in Aroona-B3a, GluB3-22 in genotypes with protein mobility alleles b, e and f and GluB3-23 was present in Aroona-B3g. Compared with GluB3-21, both GluB3-22 and GluB3-23 had a triplet nucleotide deletion (CAA) at position 378–380 in the coding region (Fig. A3), leading to a glutamine deletion in the repetitive domain of the deduced peptides BP2-2 and BP2-3, respectively (Fig. A4). In addition, GluB3-22 contained two SNPs at positions 671 and 1,246, and the latter resulted in an amino acid change from serine to asparagine in the C-terminal conserved region.

GluB3-3 had four allelic variants, designated as GluB3-31, GluB3-32, GluB3-33 and GluB3-34 (GenBank accessions EU369715, EU369716, EU369717 and EU369718, respectively), amplified with the primer set LB3F/LB3R from Aroona-B3c, Aroona-B3d, Aroona-B3h and Aroona-B3i, respectively. Compared with GluB3-31, four SNPs were detected at positions 197, 505, 836 and 1,030 in the coding region of both GluB3-32 and GluB3-34 (Fig. A5). GluB3-34 and GluB3-32 contained an additional SNP at positions 662 and 692, respectively. GluB3-33 had SNPs at positions 186, 505, 836 and 1,180 and a double-base substitution at positions 1,030 and 1,031. In addition, GluB3-34 showed a 3-bp (CAA) deletion at position 603–605 and a triplet code (CAA) insertion between positions 651 and 652, leading to a glutamine deletion and insertion in the deduced peptide BP3-4, respectively (Fig. A6).

GluB3-4 gene was amplified with the primer set LB4F/LB4R in all the eight NILs and Cheyenne, and had five allelic variants at the DNA level, designated as GluB3-41, GluB3-42, GluB3-43, GluB3-44 and GluB3-45 (GenBank accessions EU369724, EU369719, EU369727, EU369729 and EU369720, respectively). GluB3-41 was detected in NILs with protein mobility alleles a, c and d, GluB3-42 in Aroona-B3b, GluB3-43 in Cheyenne and Aroona-B3h, GluB3-44 in Aroona-B3f and Aroona-B3g and GluB3-45 in Aroona-B3i. Compared with GluB3-41, five, four and two SNPs were present in the upstream noncoding region of GluB3-42, GluB3-43 and GluB3-45, respectively (Fig. A7). In the coding region, a SNP with a C–T base transition was found at position 669 in GluB3-45 and at position 714 in GluB3-42 and GluB3-43, respectively, leading to an amino acid substitution from serine to leucine in the deduced proteins of BP4-5, BP4-2 and BP4-3 (Fig. A8). GluB3-42, GluB3-43, GluB3-44 and GluB3-45 had a common SNP (A–G) at position 731, leading to a change from isoleucine to valine in the deduced glutamine-rich repetitive domain. At the position 770–790, a 21-bp deletion was present in GluB3-45, resulting in a 7-amino acid (QFPQQQQ) deletion at the protein level. In addition, GluB3-45 had SNPs at positions 863, 885, 890 and 894, and GluB3-42 and GluB3-43 had a SNP at position 894, respectively. At positions 1,056 and 1,135, two SNPs were detected in GluB3-44, and the former led to an amino acid mutation from methionine to threonine, whereas the latter was a synonymous mutation in the deduced C-terminal cysteine-rich domain.

Characterization of the GluB3-1, GluB3-2, GluB3-3 and GluB3-4 genes and their deduced amino acid sequences

The 17 identified allelic variants of the four Glu-B3 genes contain a complete coding sequence, including start codon, termination sequence with double-stop codons TAATAA in GluB3-2 and GluB3-3 or TGATAA in GluB3-1 and GluB3-4. GluB3-1, GluB3-2 and GluB3-4 had the AATAAA polyadenylation signals in the 3′ flanking region. GluB3-4 was longer than the other three genes at the 5′ flanking region, comprising also the endosperm boxes, CAAT box and double TATA box. Sequence alignments indicated that the homologies of the DNA sequences were 87.4–94.4% among the four gene sequences, and 99.3–99.9% among different allelic variants within each of the four gene sequences (data not shown). At the protein level, the homologies of deduced amino acid sequences of the four Glu-B3 genes (designated BP1, BP2, BP3 and BP4) were 82.4–93.0% among the four genes and 98.3–100.0% (BP4-2 was the same as BP4-3) among different allelic variants within each of the four genes (data not shown).

All the deduced amino acid sequences of the four genes contained a single open reading frame (ORF) encoding a highly conserved signal peptide of 20 amino acids and a short N-terminal conserved region with 13 amino acids, followed by a repetitive domain rich in glutamine and a C-terminal conserved domain. The C domains had three subregions typical of LMW glutenin subunit proteins (Cassidy et al. 1998). The deduced peptides BP1, BP2 and BP3, encoded by GluB3-1, GluB3-2 and GluB3-3, respectively, were characterized with the amino acids MENSHIP in the N-terminal domain, and their deduced molecular weight ranged from ~39.0 kDa (BP1-5) to ~44.6 kDa (BP3-1). BP4, corresponding to GluB3-4, started with METSHIP in the N-terminal domain with molecular weights of ~39.0 kDa (BP4-5) or ~39.8 kDa (BP4-1, BP4-2, BP4-3 and BP4-4). The deduced amino acid sequences also showed typical eight-cysteine (Cys) residues, and all can be classified into type II based on the distribution of the cysteine residues (D’Ovidio and Masci 2004).

Relationship between Glu-B3 gene haplotypes and Glu-B3 protein mobility alleles

Each of the Glu-B3 protein mobility alleles contained different haplotypes at the DNA level (Table 3). Aroona-B3a possesses the haplotypes GluB3-11, GluB3-21 and GluB3-41; Aroona (Glu-B3b) has allelic variants GluB3-12, GluB3-22 and GluB3-42; Aroona-B3c contains GluB3-31 and GluB3-41; Aroona-B3d has GluB3-32 and GluB3-41; Cheyenne (Glu-B3e) contains GluB3-13, GluB3-22 and GluB3-43; Aroona-B3f possesses GluB3-14, GluB3-22 and GluB3-44; Aroona-B3g contains GluB3-15, GluB3-23 and GluB3-44; Aroona-B3h has GluB3-33 and GluB3-43; Aroona-B3i contains GluB3-34 and GluB3-45.

Table 3 Relationship between Glu-B3 protein mobility alleles and Glu-B3 haplotypes
Full size table

Gene-specific PCR markers for Glu-B3 alleles defined by protein mobility

Based on the sequence alignment of allelic variants among each of the Glu-B3 genes, ten primer sets were developed to amplify different Glu-B3 alleles based on the detected SNPs (Table 2). In the eight NILs and Cheyenne, primer set SB1F/SB1R amplified only a 1,095-bp PCR fragment in Aroona-B3a with the Glu-B3a allele (Fig. 2a). Primer set SB2F/SB2R was designed for the Glu-B3b allele in Aroona with a 1,549-bp PCR fragment (Fig. 2b). For the Glu-B3c allele in Aroona-B3c, SB3F/SB3R generated a unique 472-bp PCR product (Fig. 2c). Primer set SB4F/SB4R was used to identify the Glu-B3d allele in Aroona-B3d, producing a 662-bp band (Fig. 2d). In Cheyenne with the Glu-B3e allele, a specific 669-bp PCR product was generated with the primer set SB5F/SB5R (Fig. 2e). Primer set SB7F/SB7R was used to detect the Glu-B3g allele in Aroona-B3g with an 853-bp PCR fragment (Fig. 2f). To discriminate the Glu-B3h allele from others, primer set SB8F/SB8R was developed; it generated a 1,022-bp band in Aroona-B3h (Fig. 2g). Primer set SB9F/SB9R uniquely amplified a 621-bp PCR fragment in Aroona-B3i with the Glu-B3i allele (Fig. 2h). Since it was difficult to design a specific primer set for Glu-B3f, primer set SB6F/SB6R was developed to amplify Glu-B3f and Glu-B3g in Aroona-B3f and Aroona-B3g, respectively (Fig. 2i). In combination with SB7F/SB7R, this primer set can be used to identify Glu-B3f. In addition, primer set SB10F/SB10R was designed to amplify Glu-B3b, e and f in Aroona, Cheyenne and Aroona-B3f, respectively (Fig. 2j), and this set can be used to verify the former primer sets.

Fig. 2
figure 2

Electrophoresis of PCR products amplified from the eight NILs and Cheyenne on agarose gels using ten allele-specific markers: a gluB3a, b gluB3b, c gluB3c, d gluB3d, e gluB3e, f gluB3g, g gluB3h, h gluB3i, i gluB3fg, j gluB3bef. Materials used as PCR templates were as follows: 1 Aroona-B3a (a), 2 Aroona-B3b (b), 3 Aroona-B3c (c), 4 Aroona-B3d (d), 5 Cheyenne (e), 6 Aroona-B3f (f), 7 Aroona-B3g (g), 8 Aroona-B3h (h), 9 Aroona-B3i (i). PCR product sizes and conditions are as listed in Table 2. M DNA Ladder 2000 (100, 250, 500, 750, 1,000 and 2,000 bp)

Full size image

Validation of Glu-B3 allele-specific markers

Eight NILs and Cheyenne used in cloning the Glu-B3 genes, and additional 161 wheat varieties and advanced lines were used to validate the ten Glu-B3 allele-specific markers (Table 4). The results were in accordance with those detected by SDS-PAGE except for five genotypes (entries 28, 68, 121, 152 and 165). In particular, the entry 121, which was detected to possess Glu-B3j or Glu-B3g allele in SDS-PAGE (Fig. 3), was determined to have Glu-B3g allele by the markers gluB3fg and gluB3g. Thirty-two genotypes with the 1BL.1RS translocation showed no target PCR product in the test with molecular markers, indicating the presence of the protein allele Glu-B3j.

Table 4 Validation of the Glu-B3 allele-specific markers with eight NILs and Cheyenne, and 161 wheat varieties and advanced lines
Full size table
Fig. 3
figure 3

SDS-PAGE patterns of glutenin subunits of wheat varieties and advanced lines. 1 entry no. 43 in Table 4, 2 no. 48, 3 no. 52, 4 Pavón (Control), 5 no. 30, 6 no. 33, 7 no. 39, 8 no. 41, 9 no. 53, 10 no. 121, 11 no. 40, 12 no. 46, 13 no. 47, 14 Opata (Control)

Full size image

Discussion

The LMW-GS proteins are critical components of the gluten complex of wheat and are encoded by a highly variable gene family. Because of their importance in wheat flour quality and difficulties in discriminating them by traditional SDS-PAGE techniques, it was necessary to develop allele-specific markers to identify different LMW-GS alleles (Gupta et al. 1999; Gale 2005; Bagge et al. 2007). Zhang et al. (2004) reported seven Glu-A3 markers to discriminate different Glu-A3 alleles. Several STS markers were also developed for different Glu-D3 gene haplotypes (Zhao et al. 2007a, b). In this study, ten allele-specific primer sets were successfully designed for the nine Glu-B3 alleles defined by protein mobility. The ten markers were validated with eight Aroona NILs and Cheyenne, and 161 wheat varieties and advanced lines from CIMMYT, Australia and France.

The GluB3-1 and GluB3-2 genes were present in varieties with protein based alleles a, b, e, f and g, and the GluB3-3 allele was present in those with c, d, h and i only. As the three genes were isolated by the gene-specific primer sets LB1F/LB1R, LB2F/LB2R and LB3F/LB3R, respectively, we also designed new primer sets from within the coding regions of these genes to perform PCR, and confirmed the distinctions between the genes. In a previous study, Zhao et al. (2007b) found that the primer set T13F4/T13R3 was null in varieties with Glu-B3c, d, h and i. The GluB3-4 genes were present in varieties with the protein alleles a, b, c, d, f, g, h, and i, and thus the haplotypes of gene variants present in different varieties share common genes as well as carrying unique variants.

Currently, more than 200 LMW-GS genes have been registered in GenBank. About 30 sequences were located on chromosome 1B and 13 of them had complete coding sequences (Van Campenhout et al. 1995; D’Ovidio et al. 1997, 1999; Masci et al. 1998; Ikeda et al. 2002; Maruyama-Funatsuki et al. 2005; Huang and Cloutier 2008). In the present study, four Glu-B3 genes were identified, including the complete coding sequences of 17 allelic variants. Sequence alignments indicated that GluB3-1 was 99.1–99.7% identical in sequence to AB262661 and EU189088. In particular, GluB3-11 had only three SNP differences from AB262661 and EU189088 in the aligned domain. The three allelic variants of GluB3-2 showed lower identities to the genes in GenBank, with the highest similarity of 93.4% between GluB3-23 and AB164415, indicating that GluB3-2 might be a new LMW-GS gene identified at the Glu-B3 locus. GluB3-3 shared 99.5–100.0% identity with AB164415, and GluB3-33 had the same sequence as AB164415. GluB3-4 shared 99.1–99.9% of identities with AB062852 and X84960, and GluB3-41 had only 1 bp difference from X84960 at position 1,303 in the coding region.

Based on the first N-terminal amino acid of the mature protein, LMW-GS were divided into three types: LMW-m, LMW-s and LMW-i, corresponding to methionine, serine and isoleucine, respectively (Lew et al. 1992). In this study, BP4 started with the amino acids METSHIP in the N-terminal domain and could therefore be classified as a LMW-m type. The LMW-m type was found to be the main type by gene sequencing, but the LMW-s type was the predominant type found by N-terminal sequencing of proteins (Lew et al. 1992; Masci et al. 2002). It was suggested that the LMW-s type might originate from post-translational cleavage by an asparaginyl peptidase (Masci et al. 1998; Dupont et al. 2004). Consequently, BP1, BP2 and BP3 starting with MENSHIP in the N-terminal domain could be classified as LMW-s types based on asparagine at the third position (Ikeda et al. 2002, 2006). Previous studies also indicated that such classification of LMW-m and LMW-s type genes might be highly tenuous, as the LMW-m and LMW-s type genes have high homologies in wheat and its relatives (Jiang et al. 2008; Huang and Cloutier 2008).

The overall results of this study showed that the protein mobility alleles determined by SDS-PAGE were consistent with the screening results obtained using the allele-specific markers in 165 of 170 genotypes. The inconsistent results for five genotypes were attributed to the low discrimination power of SDS-PAGE in distinguishing some LMW-GS and the impurity of some materials. The ten Glu-B3 allele-specific markers can be used in the marker-assisted breeding aimed at the improvement of wheat quality.