Introduction

The processing properties of common wheat flour are mainly affected by high and low molecular weight glutenin subunit proteins (HMW-GS and LMW-GS) that form the disulphide-bonded gluten macropolymer (Gras et al. 2001) and contribute to the fundamental aspects of dough quality such as viscoelasticity and extensibility (Payne 1987; Wesley et al. 1999, 2001; Brites and Carrillo 2001; Luo et al. 2001). HMW- and LMW-GS alleles are therefore important targets for marker-assisted selection in the development of improved wheat cultivars (Gupta et al. 1999; Eagles et al. 2001, 2002; Gale 2005). LMW-GS represents approximately one-third of total seed storage proteins and 60% of the gluten fraction (Bietz and Wall 1973). Their structural definition through nucleotide sequencing has been more problematical than for HMW-GS because they are more numerous and the respective proteins more difficult to purify.

Most of the LMW-GS are encoded by the complex Glu-3 loci (Glu-A3, Glu-B3 and Glu-D3) on the short arms of chromosomes 1A, 1B and 1D (Gupta and Shepherd 1990; Jackson et al. 1983), though other LMW-GS loci have also been reported, such as Glu-B2 and Glu-B4 on chromosome 1B (Jackson et al. 1985; Liu and Shepherd 1995), Glu-D4 on chromosome 1D and Glu-D5 on chromosome 7D (Sreeramulu and Sigh 1997). Currently, six, eleven and five alleles defined by protein electrophoretic mobility have been confirmed at Glu-A3, Glu-B3 and Glu-D3 locus, respectively, in common wheat (Branlard et al. 2003; Gianibelli et al. 2001; Gupta and Shepherd 1990). Recently, seven new alleles, Glu-A3g, Glu-A3h, Glu-B3m, Glu-B3n, Glu-B3o, Glu-B3p and Glu-B3q were designated (McIntosh et al. 2003). Another allele, Glu-D3f was also identified (CIMMYT and Japan NARC, not published) and is currently awaiting verification.

Based on the first amino acid present in the N-terminal sequences of the proteins, eight types of LMW-GS have been identified (D’Ovidio and Masci 2004), which are LMW-s starting with the sequence SHIPGL-, LMW-i starting with sequence ISQQQQ-, three LMW-m types with N-terminal sequences of METSHIPGL-, METSRIPGL and METSCIPGL-, respectively, and three types with N-terminal sequences resembling those of the α-, β- and γ-type gliadins (Kasarda et al. 1988; Tao and Kasarda 1989; Lew et al. 1992; Cloutier et al. 2001; Gianibelli et al. 2001). LMW-GS was further classified into 12 groups by Ikeda et al. (2002) according to deduced amino acid sequences and in particular the number and position of cysteine residues available for inter-molecular disulphide bond formation (Shewry and Tatham 1997). More than 100 sequence tags of genes, partial genes and pseudo-genes of the LMW-GS family have been cloned and sequenced from several common wheat cultivars (Pitts et al. 1988; Cloutier et al. 2001; Ikeda et al. 2002; Zhang et al. 2004). Hai et al. (2005) retrieved 69 known LMW-GS genes from GenBank and classified them into nine groups based on the deduced amino acid sequence of the highly conserved N-terminal domain, and nine corresponding primer sets proved to be LMW-GS group-specific were established. Ikeda et al. (2006) also constructed ten group-specific markers according to the published nucleotide sequences. However, the relationship between different protein mobility alleles and their corresponding allelic variants at the DNA level is difficult to determine. Based on the allelic variation of one LMW-GS gene at the Glu-A3 locus, a set of PCR markers were developed by Zhang et al. (2004), whereas, no marker sets are currently available for the identification of alleles at Glu-B3 and Glu-D3 loci (Gale 2005). In our previous study, three LMW-GS genes were amplified from the Glu-D3 locus and seven haplotypes were characterized in eight common wheat cultivars (Zhao et al. 2006). In this study, we report the identification of more Glu-D3 gene haplotypes and development of a set of STS markers for these haplotypes.

Materials and methods

Wheat stocks

Eight common wheat cultivars (Tasman, Chinese Spring, Silverstar, Sunco, Aroona, Norin61, Hartog, and BT2288A) carrying five Glu-D3 alleles that were defined traditionally by protein electrophoretic mobility (McIntosh et al. 1998) were used to amplify Glu-D3 genes in this study (Table 1). Chinese Spring and its nulli-tetrasomic lines N1AT1B (nullisomic 1A-tetrasomic 1B), N1BT1D (nullisomic 1B-tetrasomic 1D) and N1DT1B (nullisomic 1D-tetrasomic 1B) provided by Prof. R. A. McIntosh at the Plant Breeding Institute, University of Sydney, were used to confirm chromosomal locations of identified genes. Twenty Chinese wheat cultivars with protein mobility alleles Glu-D3a, b, c and f were used to validate the developed molecular markers (Table 2).

Table 1 Relationship between GluD3 mobility alleles and GluD3 gene haplotypes
Full size table
Table 2 Validation of the 7 STS markers with 20 Chinese common wheat cultivars
Full size table

Development of PCR primers for identifying new Glu-D3 variants

Gene-specific primers were developed based on the method described by Zhang et al. (2003, 2004). Eight reference Glu-D3 genes, X13306, AB062851, AB062872, AB062873, AB062874, M11077, U86026 and X84961 available in GenBank were used for primer development (http://www.ncbi.nlm.nih.gov). Each gene sequence was divided into two parts to design corresponding primers and to obtain the accurate sequence results by reassembling two PCR products of normal length. The forward and reverse primers for the 5′ region of the gene were designed first. The amplified sequences were used to design the forward primers of 3′ region. A total of 48 primers were designed and 96 primer sets (48 for 5′ region and 48 for 3′ region) were tested. Primer screening was conducted according to Zhao et al. (2006). Based on the result of sequence alignments, six pairs of primer sets for three Glu-D3 genes were confirmed with annealing temperature of 58°C. Primer sequences (5′–3′) and their locations within the reference genes are shown in Table 3.

Table 3 Primers used for amplifying the three newly described Glu-D 3 genes
Full size table

DNA extraction and PCR amplification

Genomic DNA was extracted from seedlings or seeds using modified CTAB procedure (Gale et al. 2001). PCR was performed using 3U of TaKaRa Taq polymerase in 40 μl of reaction buffer (1.5 mM MgCl2) containing 60 ng of genomic DNA, 200 μM of each of dNTPs and 10 pmoles of each PCR primer. PCR cycling was 94°C for 5 min followed by 38 cycles of 94°C for 40 s, 58°C for 40 s, 72°C for 90 s, and a final extension at 72°C for 5 min.

Sequencing of PCR products

PCR fragments were sequenced by the Invitrogen Biotechnology Co. Ltd (Beijing). In order to eliminate errors in sequencing, PCR reaction and sequencing procedure were repeated two to four times. Some PCR products that were difficult to sequence directly were purified, cloned, and sequenced using pGEM®-T Easy Vector cloning System (Promega). Sequence analysis and characterization were performed using software DNAMAN (http://www.lynnon.com).

Marker development and validation

STS marker primers were developed based on the sequence alignments of 12 Glu-D3 gene haplotypes detected in this and previous studies (Zhao et al. 2006) and screened by amplifying the eight wheat cultivars with known alleles. The PCR products of gene-specific primer sets were sequenced and compared with their target gene haplotypes. The STS markers were then validated by amplifying 20 Chinese wheat cultivars with different protein mobility alleles (Table 2).

Results

New DNA variations at the Glu-D3 locus

In addition to the three Glu-D3 genes reported in our previous study (Zhao et al. 2006), three additional LMW-GS genes including five allelic variants were identified at the Glu-D3 locus on chromosome 1D in eight common wheat cultivars. The first gene, designated as GluD3-4, was amplified with a primer set S2F21/S2R21 for the 5′ region and S5F54/S5R55 for the 3′ region of the gene (Table 3), which generated 701- and 959-bp products, respectively (Fig. 1). The complete sequence of GluD3-4 was assembled with a size of 1,384 bp. Likewise, the second gene, designated as GluD3-5 with 1,292 bp, was amplified using primer sets S1F11/S1R11 for the upstream region and S4F41/S4R43 for the downstream region with fragment sizes of 880 and 770 bp, respectively (Fig. 1). The third one, designated as GluD3-6 in 1,300 bp, was amplified with primer sets S3F31/S3R33 for the upstream region, and S6F61/S6R64 for the downstream region, which resulted in 855- and 987-bp fragments, respectively (Fig. 1).

Fig. 1
figure 1

Electrophoresis of PCR products amplified from Chinese Spring and its nulli-tetrasomic lines using 6 specific primer sets in agarose gel. 1 Chinese Spring; 2 N1DT1B; 3 N1BT1A; 4 N1AT1B. M, DNA Ladder 2000 (100, 250, 500, 750, 1,000, 1,600, 2,000 bp). a The upstream region, and b the downstream region of GluD3-4, GluD3-5 and GluD3-6 genes

Full size image

GluD3-4 showed three haplotypes or allelic variants at the DNA level in the eight wheat cultivars (Table 1). The first allele, designated GluD3-41, was found in the cultivar Aroona. The second allele, designated GluD3-42, was present in cultivars Chinese Spring, BT2288A, Silverstar, Sunco and Norin 61. The third allele, designated GluD3-43, was presented in cultivars Tasman and Hartog. Compared with GluD3-42, GluD3-41 had a 3-bp insertion at the 857–859 position in gene coding region (Sup-Fig. 1), leading to a glutamine insertion at the 249th position of the C-terminal glutamine-rich region (Sup-Fig. 4); GluD3-43 had a single nucleotide polymorphism (SNP) in the coding region, resulting in an nonsense mutation at the 119th position of deduced peptide, which made the haplotype a pseudogene.

Both GluD3-5 and GluD3-6 had no allelic variation in the eight wheat cultivars (Table 1; Sup-Fig. 2, 3, 5, 6), indicating they were relatively conservative in common wheat.

Development of STS markers for identifying different GluD3 gene haplotypes

In the previous (Zhao et al. 2006) and present studies, we characterized six LMW-GS genes at Glu-D3 locus of common wheat. Among them, GluD3-1, GluD3-2, GluD3-3 and GluD3-4 had two, three, two and three allelic variants or haplotypes, respectively, which were used to design gene specific primers. In total, seven STS markers for the haplotypes of GluD3-2, GluD3-3 and GluD3-4 were confirmed by amplifying the eight wheat cultivars containing mobility alleles GluD3a, b, c, d and e. The primer sequences and their locations in reference genes were listed in Table 6. GluD3-5 and GluD3-6 did not show any allelic variation in all the eight cultivars tested, and thus no gene-specific markers were developed for them. For GluD3-1, the only mutation between its two allelic forms was a CAA indel that occurred in the high repetitive region with 11 CAA repeats. Though both forward and reverse primers were developed based on the indel locus, none of them was specific and useful.

Based on the SNPs between the three haplotypes of GluD3-2, three specific primer sets were selected. Primer pair M2F12/M2R12 generated an 884-bp fragment and was specific to allelic forms GluD3-21/22. M2F2/M2R2 amplified a fragment of 958 bp that was specific to GluD3-22. M2F3/M2R3 amplified a fragment of 725 bp specific to GluD3-23 (Figs. 2, 3).

Fig. 2
figure 2

Haplotype-specific PCR for GluD3-21/22 and GluD3-23 using primer sets M2F12/M2R12 and M2F3/M2R3. 1,9 Tasman (a); 2,10 Chinese Spring (a); 3,11 Silverstar (b); 4,12 Sunco (b), 5,13 Aroona (c); 6,14 Norin61 (d); 7,15 Hartog (e); 8,16 BT2288A (e); M DNA Ladder 2000 (100, 250, 500, 750, 1,000, 1,600, 2,000 bp)

Full size image
Fig. 3
figure 3

Haplotype-specific PCR for GluD3-22 using primer sets M2F2/M2R2. 1 Tasman (a); 2 Chinese Spring (a); 3 Silverstar (b); 4 Amery (b); 5 Sunco(b), 6 Leichhardt (b); 7 Baxter (b); Cunningham (b); 9 Aroona (c); 10 Norin61 (d); 11 Hartog (e); 12 BT2288A (e); M DNA Ladder 2000 (100, 250, 500, 750, 1,000, 1,600, 2,000 bp)

Full size image

Two primer sets were specific for GluD3-3. The first primer set M3F1/M3R1 amplified a fragment of 528 bp specific to GluD3-31 and the second set M3F2/M3R2 amplified a fragment of 334 bp specific to GluD3-32 (Fig. 4). For GluD3-4, only two gene specific markers were verified, of which M4F1/M4R1 was specific to GluD3-41 and M4F3/M4R3 specific to GluD3-43, with 773- and 413-bp PCR products, respectively (Fig. 5).

Fig. 4
figure 4

Haplotype-specific PCR for GluD3-31 and GluD3-32 using primer sets M3F1/M3R1 and M3F2/M3R2. 1,9 Tasman (a); 2,10 Chinese Spring (a); 3,11 Silverstar (b); 4,12 Sunco (b), 5,13 Aroona (c); 6,14 Norin61 (d); 7,15 Hartog (e); 8,16 BT2288A (e); M DNA Ladder 2000 (100, 250, 500, 750, 1,000, 1,600, 2,000 bp)

Full size image
Fig. 5
figure 5

Haplotype-specific PCR for GluD3-41 and GluD3-43 using primer sets M4F1/M4R1 and M4F3/M4R4. 1,11 Tasman (a); 2,12 Chinese Spring (a); 3,13 Silverstar (b); 4,14 Sunco (b), 5,15 Aroona (c); 6 Dagger (c); 7 Halberd; 8,16 Norin61 (d); 9,17 Hartog (e); 10,18 BT2288A (e); M DNA Ladder 2000 (100, 250, 500, 750, 1,000, 1,600, 2,000 bp)

Full size image

Validation of the seven GluD3 STS markers

In order to verify the accuracy of the developed markers, the PCR products of each primer set were sequenced and compared with its corresponding gene haplotype. DNA sequence analysis indicated that all the seven markers were completely matched to their target gene haplotypes. Validation with 20 Chinese wheat cultivars (Table 2; Sup-Fig. 7–13) indicated that the 10 cultivars containing alleles a and b were positive to both M2F12/M2R12 and M3F1/M3R1, indicating that they have haplotypes of GluD3-21 or GluD3-22 and GluD3-31; the 10 cultivars containing alleles c and f were positive to both M2F3/M2R3 and M3F2/M3R2, exhibiting that they contain the GluD3-23 and GluD3-32 haplotypes; and all cultivars were negative to both M4F1/M4R1 and M4F3/M4R3, suggesting that they may contain another haplotype of this gene GluD3-42 (Note, STS marker for GluD3-42 was unsuccessful). For the primer set M2F2/M2R2, two cultivars with protein allele b (Shaan 229 and CA9550) and two cultivars with allele a (Zhengzhou 9023 and Jingdong 8) gave the 528 bp band (Sup-Fig. 8), indicating that they contain haplotype GluD3-22.

Discussion

The relationship between the GluD3 gene haplotypes and GluD3 mobility alleles

The differences in electrophoretic patterns for LMW-GS that were observed between the cultivars with different mobility alleles (Lew et al. 1992; Gupta and Shepherd 1990) were conditioned by several Glu-3 genes (Zhao et al. 2006; D’Ovidio and Masci 2004). In this and previous studies (Zhao et al. 2006), six different Glu-D3 genes were identified from Glu-D3 locus on chromosome 1D of common wheat. Among these, four showed allelic variations in the wheat cultivars tested in this study. In most cases, the cultivars with the same mobility allele (a, b, c, d or e) always had the same allelic variant at DNA level (Table 1) (Zhao et al. 2006). For example, cultivars with allele b contain gene haplotypes GluD3-12, -21/22, -31, -42, -5 and -6. However, there were also exceptions in that cultivars Chinese Spring with allele a and BT2288A with allele e had the same haplotype of GluD3-42, and Tasman with allele a and Hartog with allele e had the same haplotype of GluD3-43. This may due to the presence of gamma-gliadins in the LMW glutenin fraction, which link Glu-D3 locus and are having similar molecular weights with LMW-GS. Meanwhile, the LMW-GS encoded at Glu-D3 locus may be complicated by post-translational modification leading to changes at protein mobility level that is not related to differences in the gene sequence (Liu et al. 2005; He et al. 2005). Based on the deduced amino acid sequences of the 12 haplotypes of 6 genes, 5 types of allelic combinations (i.e. 5 alleles at DNA level) were found across the 8 wheat cultivars (Table 1), although they were not completely consistent with the traditional Glu-D3 alleles defined by protein electrophoretic mobility (McIntosh et al. 1998). The results indicated that the so-called alleles (protein electrophoretic patterns) were, in fact, controlled jointly by the combinations of haplotypes of six Glu-D3 genes at least. This is consistent with Ikeda et al. (2006).

Relationship between the GluD3 genes identified and those registered in GenBank

So far, more than 100 entries related to LMW-GS genes in common wheat have been registered in GenBank, including complete genes, partial genes and pseudo-genes (Okita et al. 1985; Colot et al. 1989; Van et al. 1995; Masci et al. 1998; Ikeda et al. 2002; Ozdemir and Cloutier 2005; Zhao et al. 2006). In this study, 18 GluD3 and 16 undefined LMW-GS gene sequences were selected to compare with the 6 genes identified. Results indicated that GluD3-4 was highly similar to AB062872 and M11077; GluD3-41 had 3 bp difference from M11077 at position 79, 106 and 1,010 with the former two being in start codon region and the third in encoding region, respectively; GluD3-42 had only one base difference from AB062872 (G to A) at position 1,051 but was 300 bp longer in the downstream region than the later; GluD3-5 is highly identical to AB062851 and X84961, with only 3 bp difference from AB062851 at position 843 (G to A), 1,002 (G to T) and 1,076 (C to T); and GluD3-6 shared 99.4–99.7% identity with AB062873, AB062874 and U86029. The identity with the other 11 Glu-D3 genes from GenBank was all under 90%. Further analysis showed that all the 18 Glu-D3 and seven undefined genes in GenBank could also be classified into six groups that matched the haplotypes of the six Glu-D3 genes studied in this manuscript even though one or few bp difference existed between each groups (Table 7). Overall these results indicate that the Glu-D3 locus is a multiple gene locus consisting of at least six different LMW-GS genes that all have allelic variants among different genotypes.

GluD3 gene variation and STS marker development

Within the six LMW-GS genes, including the 12 allelic variants or haplotypes identified at the Glu-D3 locus, two types of mutations were detected. These mutations were either base substitutions or indels. Here we also found an indel of the triplet code CAA in the repeat region of GluD3-4, the same as in GluD3-1 and GluD3-2, and which may influence protein feature (Zhao et al. 2006). It was interesting to note that although the base sequences among the six Glu-D3 genes varied significantly (with identity of 80.3–92.4%, Table 4), the differences between the allelic forms or haplotypes of each gene were relatively small (with similarity of 99.3–100%). In addition, no allelic variation for GluD3-5 and GluD3-6 was found in the eight wheat cultivars tested. The results indicated that the LMW-GS genes at Glu-D3 locus were relatively conservative compared with HMW-GS genes (Lei et al. 2006; Ma et al. 2003; Gianibelli et al. 2001).

Table 4 Similarity comparison of 12 haplotypes of six GluD3 genes (below diagonal) and their deduced amino-acid sequences (above diagonal) (%, irrespective of the sequence length)
Full size table
Table 5 the N-terminal amino acid sequences of the six deduced Glu-D3 proteins
Full size table
Table 6 Seven pairs of PCR primers for the identification of different GluD3 gene haplotypes
Full size table
Table 7 The relationship between GluD3 haplotypes identified in this study and GluD3 genes from Genbank
Full size table

Allelic variation at LMW-GS loci is related to differences in dough quality in common wheat (Gupta et al. 1989; Gupta and MacRitchie 1994) and durum wheat (Pogna et al. 1990; Ruiz and Carrillo 1993). Some allelic forms of LMW-GS show even greater effects on dough strength and extensibility than HMW-GS (Payne 1987). Pogna et al. (1996) reported that the durum genotypes with the Gli-D1/GluD3 translocation showed increase in dough strength and extensibility and decrease in tenacity compared with its wild-type durum wheats. Ma et al. (2005) showed that the Glu-D3 locus played multifaceted effects on dough physical properties. Ikeda et al. (2006) also reported that the abundance of LMW-GS encoded by Glu-D3 might contribute more to the gluten viscoelasticity of common wheat.

Based on the present study, the effects of Glu-D3 subunits on quality could not be unambiguously traced among different genotypes using the currently established mobility allele system. For example, the results of M2F2/M2R2 and M4F1/M4R1 with 20 wheat cultivars was not consistent with the expected results, even though a consistent result was obtained using the 10–12 Australian cultivars (Figs. 3, 5). For M2F2/M2R2, the three cultivars with protein mobility allele b (Wanmai 33, Yumai 70 and Yannong 15) were expected to have the 958 bp band missing, but in fact a product was detected. For the two cultivars containing allele a (Zhengzhou 9023 and Jingdong 8), a PCR band of 958 bp was present, which is contrary to the expectation that they should be missing. For M4F1/M4R1 (a marker for GluD3-41), all the six Chinese cultivars with allele c did not contain the specific haplotype PCR band but the three Australian cultivars with c allele showed this PCR product. This may be due to the complexity of relations between protein alleles and their coding haplotypes, or the ambiguity in identification of the Glu-D3 protein subunits by SDS-PAGE. Due to the difficulty in directly and correctly identifying LMW-GS proteins, it is of great importance to clarify the gene composition of the Glu-D3 locus and to develop markers for these genes. Until now, no molecular marker for distinguishing Glu-D3 alleles has been available. In this study, we developed and validated seven STS markers for different Glu-D3 gene haplotypes. Our markers will be useful in accurately dissecting the effects of the LMW Glu-D3 locus on wheat quality at the gene level, and make it possible to utilize this information in wheat breeding.