Haplotype variation of Glu-D1 locus and the origin of Glu-D1d allele conferring superior end-use qualities in common wheat.

In higher plants, seed storage proteins (SSPs) are frequently expressed from complex gene families, and allelic variation of SSP genes often affects the quality traits of crops. In common wheat, the Glu-D1 locus, encoding 1Dx and 1Dy SSPs, has multiple alleles. The Glu-D1d allele frequently confers superior end-use qualities to commercial wheat varieties. Here, we studied the haplotype structure of Glu-D1 genomic region and the origin of Glu-D1d. Using seven diagnostic DNA markers, 12 Glu-D1 haplotypes were detected among common wheat, European spelt wheat (T. spelta, a primitive hexaploid relative of common wheat), and Aegilops tauschii (the D genome donor of hexaploid wheat). By comparatively analyzing Glu-D1 haplotypes and their associated 1Dx and 1Dy genes, we deduce that the haplotype carrying Glu-D1d was likely differentiated in the ancestral hexaploid wheat around 10,000 years ago, and was subsequently transmitted to domesticated common wheat and T. spelta. A group of relatively ancient Glu-D1 haplotypes was discovered in Ae. tauschii, which may serve for the evolution of other haplotypes. Moreover, a number of new Glu-D1d variants were found in T. spelta. The main steps in Glu-D1d differentiation are proposed. The implications of our work for enhancing the utility of Glu-D1d in wheat quality improvement and studying the SSP alleles in other crop species are discussed.

Introduction

Seed storage proteins (SSPs) occur in diverse plant species. They provide energy for seed germination and seedling growth, and form a major source of dietary protein for mankind [1]. In crop plants, SSPs are frequently crucial determinants of quality traits [1], [2]. Many superior SSP alleles have been identified and used in breeding (e.g., [3]–[5]), and substantial efforts have been devoted in sequencing the genomic regions containing SSP coding genes (e.g., [6]–[8]). In general, SSPs tend to be expressed from complex gene families, and often possess multiple alleles with different phenotypic effects [9], [10]. In order to more effectively use SSPs in improving crop quality traits, it is necessary to understand the variations of the genomic loci harbouring SSP genes in germplasm resources and the origins of superior SSP alleles.

High-molecular-weight glutenin subunits (HMW-GSs) are a family of SSPs essential for the end-use quality control of common wheat (Triticum aestivum, AABBDD) [4], [11]. HMW-GSs, together with the low-molecular-weight glutenin subunits (LMW-GSs) and gliadins, are the major factors conferring viscoelastic properties to wheat doughs. The relative balance of elasticity (controlled mainly by HMW-GSs and LMW-GSs) and viscosity (largely by gliadins) in the doughs determines their suitability for being processed into different types of wheat foods. In hexaploid wheat, the genes encoding HMW-GSs are contained in homoeologous Glu-1 loci (Glu-A1, B1, and D1), which are located on the 1A, 1B and 1D chromosomes, respectively [11]. The loci orthologous to Glu-A1, B1, and D1 have been found in the Triticeae species closely related to common wheat [12]. The genomic regions of Glu-1 loci are structurally complex, with the core of a Glu-1 locus being composed of four genes arranged in the order of Globulin 1−Glu-1-2−Globulin 2−Glu-1-1 [13]. Glu-1-1 and Glu01-2 encode x- and y-type subunits, respectively. Owing to allelic variation and gene inactivation, both Glu-1-1 and Glu-1-2 have multiple alleles [14]. For example, Glu-D1 has several commonly found alleles, Glu-D1a, b, c, d, e and f [4], [15]. The two HMW-GS genes of Glu-D1a are 1Dx2 and 1Dy12 (encoding 1Dx2 and 1Dy12 subunits, respectively), whereas those of Glu-D1d are 1Dx5 and 1Dy10 (specifying 1Dx5 and 1Dy10 subunits, respectively) [15], [16].

The primary structure of HMW-GSs is composed of a signal peptide (removed from mature subunit), a N terminal domain, a central repetitive domain, and a C-terminal domain [9]. The cysteine residues in HMW-GSs are usually conserved in both number and position. However, compared to 1Dx2, 1Dx5 harbors the amino acid substitution at position 118, causing the replacement of a serine residue by cysteine at the beginning of the repetitive domain [17]. Relative to 1Dx2, the presence of this extra cysteine in 1Dx5 is frequently associated with the doughs exhibiting stronger elasticity and superior end-use qualities for breadmaking and noodle processing [17]–[21]. Interestingly, several studies have also shown that the wheat varieties carrying Glu-D1d are generally more tolerant to heat stress-induced decline of dough quality than those having Glu-D1a [19], [20], [22], [23]. This suggests that Glu-D1d is likely to be even more useful for future improvement of wheat end-use qualities, as heat stress may become increasingly frequent in the next 30–50 years owing to global warming [24].

Despite its agronomical importance, the evolutionary origin of Glu-D1d remains a mystery. Since extant common wheat was descended from domestication of the ancestral hexaploid wheat formed about 10,000 years ago through natural hybridization between tetraploid wheat (T. turgidum, AABB) and the D genome donor Aegilops tauschii (DD) [25], [26], it is possible that Glu-D1d may come directly from Ae. tauschii, or be differentiated during hexaploidization. The Glu-D1 genomic region has been sequenced in one Ae. tauschii accession and one common wheat variety [7], [27]. The 1Dx and 1Dy genes in the sequenced Glu-D1 region are usually separated by more than 50 kb with each surrounded by multiple repetitive elements [7]. Furthermore, complex differences were found in not only 1Dx and 1Dy genes but also the repetitive elements between the Glu-D1 region in Ae. tauschii and that of common wheat, indicating close relationships between the molecular variations of 1Dx and 1Dy genes and those of surrounding repetitive elements. Therefore, the main objectives of this work were to obtain a more comprehensive understanding of the haplotype variation of Glu-D1 genomic region, and to shed new light on the origin of Glu-D1d. European spelt wheat (T. spelta thereafter), cultivated as land races, was included in this work because it possessed a hexaploid genome highly similar to that of common wheat, and represents a more primitive hexaploid relative of common wheat [28]–[31].

Materials and Methods

Plant Materials

A total of 208 common wheat varieties (listed in Figure 1 and Tables S1 and S2), along with 355 Ae. tauschii accessions (Table S3) and 215 T. spelta lines (Table S4), were employed in this study to investigate Glu-D1 haplotypes. Additional materials and the main suppliers of the germplasm stocks used in this work are provided in Methods S1.

Figure 1

Development of seven new molecular markers in Glu-D1 region.

(A) Organization of homologous Glu-D1 regions in the common wheat variety Renan and the Ae. tauschii accession AUS18913. The diagrams illustrating the organization of the various transposon elements and genes in the two Glu-D1 regions were adapted from previously published works [7], [27]. Only the structures relevant to this work are shown. The letters “a” and “b” denote distinct copies of the duplicated genes. Positions of the seven newly developed DNA markers (Xms1, Xid1, Xrj1, Xrj2, Xms2, Xrj3, and Xrj4) are indicated by filled arrowheads. The “*” and “#” symbols indicate the markers based on microsatellite or indel. The HMW-GS genes Glu-D1d-1 (1Dx5), Glu-D1d-2 (1Dy10), Glu-D1a-1 (1Dx2) and Glu-D1a-2 (1Dy12) are marked by empty arrowheads. The Wis-1 and Wis-2p insertions were unique to Renan Glu-D1. The Angela-4 element, although intact in AUS18913 Glu-D1, had its internal region deleted in Renan Glu-D1 (and thus named as Angela-4p). Compared to the solo-LTR Sabrina-3s in Renan Glu-D1, its counterpart in AUS18913 Glu-D1 underwent further sequence deletion (and thus designated as Sabrina-3sd). (B) SDS-PAGE analysis of HMW-GS subunits from 14 common varieties containing Glu-D1a, b, c, d, e or f alleles. The 1Dx and 1Dy subunit pairs (1Dx5+1Dy10, 1Dx2+1Dy12, 1Dx3+1Dy12, 1Dx4+1Dy12, 1Dx2+1Dy10, 1Dx2.2+1Dy12) encoded by different Glu-D1 alleles are indicated by Arabic numbers. (C) Amplification patterns of seven Glu-D1 markers in the 14 common wheat lines with Glu-D1a, b, c, d, e or f alleles. The size (bp) of the amplicons and the names of the markers are provided on the left and right sides of the graph, respectively.

Marker Development and Haplotype Analysis

The BAC sequences of DQ537337 (from Renan) and AF497474 (from Ae. tauschii accession AUS18913) harboring Glu-D1 locus were retrieved from NCBI database (http://www.ncbi.nlm.nih.gov/). Three types of polymorphisms, including transposable element (TE) insertions or deletions, simple sequence repeats, and short nucleotide indels, between the two sequences were considered for marker development. The software Primer Premier 5.0 (PREMIER Biosoft International, CA, USA) was used for designing the primer pairs specific for each of the seven Glu-D1 markers (Table S5). Genomic DNA samples were extracted from desired plant lines grown in the greenhouse using a CTAB method [32], and used for PCR amplifications (Methods S1). Glu-D1 haplotypes were distinguished by differences in the alleles of the seven diagnostic markers.

Detection of 1Dx and 1Dy Genes using the PCR Markers UMN25 and UMN26

The PCR markers UMN25 and UMN26 developed by Liu et al. [33] were used to detect the 1Dx and 1Dy genes in one set of common wheat lines (Tables S2), and to investigate the 1Dy genes hosted by 12 Glu-D1 haplotypes. PCR amplification and gel analysis were conducted as described previously [33].

Isolation of 1Dx or 1Dy Coding Sequences and the Long Terminal Repeat (LTR) Regions of Sabrina-2

The primers for amplifying 1Dx or 1Dy coding sequences and the left or right LTR regions of Sabrina-2 (Table S5) were developed using Primer Premier 5.0 (PREMIER Biosoft International, CA, USA). Genomic PCR was employed for isolating the desired sequences, which was always conducted using the high fidelity DNA polymerase ExTaq (TaKaRa, Tokyo, Japan), with the cycling parameters identical to those reported previously [34]. The PCR fragments of the expected size were cloned and sequenced (Methods S1). To facilitate recognition, the 1Dx (or 1Dy) genes from T. spelta and Ae. tauschii were labeled by 1D (1D) and 1D (1D), respectively. Additionally, each of the newly characterized 1Dx (1D) gene was tagged by the name and Glu-D1 haplotype of the accession from which it was isolated.

Dating LTR Retrotransposon Insertion Times

The insertion time of retroelement was estimated based on nucleotide substitution rate in the left and right LTR sequences according to a previous study [35]. Kimura-2 parameter distances (K) between the two LTRs of individual elements were calculated by software MEGA 5.0 [36]. The average substitution rate (r) of 1.3×10−8 substitutions per synonymous site per year [37], was used to date the insertions of retroelements. The time (T) since element insertion was calculated using the formula T = K/2r [38].

Multiple Alignment and Phylogenetic Analysis

Sequence alignment was performed using the ClustalW program (www.ebi.ac.uk). The resultant multiple alignment was then used for constructing phylogenetic trees using the software MEGA 5.0 (Methods S1).

Estimation of Approximate Divergence Times of 1Dx or 1Dy Genes

Synonymous substitution (Ks) was estimated with MEGA 5.0 [36], Divergence time (T) was calculated using T = Ks/2r, where r is the grass Adh1 and Adh2 substitution rate of 6.5×10−9 substitutions per synonymous site per year [39].

Phylogenetic Network Analysis

The alleles of the seven Glu-D1 markers and the types of the 1Dx and 1Dy genes were employed for phylogenetic network analysis of 12 Glu-D1 haplotypes with the Median-Joining network algorithm [40], which is installed in the software network 4.6.1.0 (Fluxus Technology Ltd., Suffolk, UK).

Accession Numbers

All novel sequences were submitted to GenBank with reference numbers JX173931–JX173953.

Results

Development of New DNA Markers for Glu-D1 Locus

The genomic sequences available for Glu-D1 regions in the Ae. tauschii accession AUS18913 and common wheat variety Renan allowed for understanding structural variations in this locus between the diploid D and hexaploid D genomes [7], [27]. The Glu-D1 allele of Renan encodes 1Dx5 and 1Dy10 [41], and thus belongs to Glu-D1d, whereas the Glu-D1 allele of AUS18913 is likely Glu-D1a, because the two HMW-GSs in this accession resemble highly 1Dx2 and 1Dy12 [27]. The patterns of transposon insertion and deletion vary considerably between the two Glu-D1 regions (Figure 1A and [7]). We therefore developed seven new DNA markers based on their sequence variations to investigate the haplotype structure of Glu-D1 locus (Figure 1A). Among these markers, two (Xms1 and Xid1) were located upstream of 1Dy (Glu-D1–2), two (Xrj1 and Xrj2) resided between 1Dy and 1Dx (Glu-D1-1), and the remaining three (Xms2, Xrj3 and Xrj4) were downstream of 1Dx (Figure 1A). Xms1 and 2 were two microsatellite markers, and Xid1 was an indel marker. Xrj1, 2, 3 and 4 were repeat DNA insertion site based polymorphism (ISBP) markers [42], [43]. Xrj1 and 2 were designed based on the unique insertion of the retrotransposon Wis-1 and specific deletion of the internal region in the retroelement Angela-4 in Renan Glu-D1 (Figure 1A). Xrj3 was developed based on structural difference between Sabrina-3s and Sabrina-3sd in Renan and AUS18913 Glu-D1 regions, while Xrj4 was derived from the specific presence of another partially deleted retrotransposon (Wis-2p) in Renan Glu-D1 (Figure 1A). The location of the seven markers in the 1D chromosome was confirmed by PCR mapping (Figure S1).

The capacity of the newly developed markers for revealing potential haplotype variation of Glu-D1 was tested by examining 14 common wheat varieties carrying Glu-D1a, b, c, d, e or f alleles (Figure 1B). Clearly, two distinct haplotypes (H1 and H2) were detected, with H1 unique to the two varieties (Attila and Bobwhite) carrying Glu-D1d and H2 shared by the remaining 12 varieties harboring Glu-D1a, b, c, e or f (Figure 1C). H1 was characterized by positive amplifications of all seven markers. In contrast, H2 was characterized by positive amplifications of Xid1, Xrj2 and Xms2, but not Xms1 or Xrj1, 3 and 4. Moreover, the alleles of Xid1, Xrj2 or Xms2 amplified for the varieties carrying Glu-D1a, b, c, e or f were identical, indicating that the three markers were monomorphic, and could not differentiate, among the five Glu-D1 alleles. Among the three markers co-dominant for H1 and H2, the amplicon size of Xrj2 was smaller in H1 than in H2 (Figure 1C), which was due to sequence deletion in Angela-4p in the Glu-D1 region represented by Renan (Figure 1A). The amplicon lengths of Xms2 in H1 and H2 were 368 and 346 bp, respectively (Figure S2), which corresponded to a 22 bp indel between the two amplicons of this microsatellite marker in Renan and AUS18913. Finally, the amplicon size of Xid1 did not differ between the two Glu-D1 haplotypes detected in the 14 common wheat genotypes.

Detection and Analysis of Glu-D1 Haplotypes

The haplotype variation of Glu-D1 locus in common wheat was further investigated using two larger sets of materials. The first set was composed of 60 common wheat varieties from diverse geographic locations and with known information on their Glu-D1 alleles (Table S1). After genotyping with the seven diagnostic Glu-D1 markers, H1 was specifically detected in the varieties carrying Glu-D1d, whereas H2 was shared by the remaining having other Glu-D1 alleles (Table S1). In the second set of 134 common wheat varieties, the status of their Glu-D1 alleles was unknown before this work. After screening by PCR amplification of the co-dominant markers UMN25 (co-dominant for 1Dx2 and 1Dx5) and UMN26 (co-dominant for 1Dy12 and 1Dy10) [33], these lines were divided into two groups, one carrying Glu-D1d and the other with Glu-D1a. Upon genotyping with the seven Glu-D1 markers, the varieties having Glu-D1d were all found to belong to H1, while those harboring Glu-D1a alleles could all be assigned to H2 (Table S2). Clearly, H1 and H2 were the two main haplotypes of Glu-D1 locus in common wheat. H1 was specific for Glu-D1d, whereas H2 was shared by Glu-D1a, b, c, e or f.

Using the seven markers, a total of nine Glu-D1 haplotypes was identified among 355 diverse Ae. tauschii accessions collected from over 20 countries (Table 1, Table S3). Compared with common wheat, Xid1 detected additional polymorphisms in Ae. tauschii, but Xrj2 and Xrj4 were both monomorphic (Table 1). Among the nine haplotypes found in Ae. tauschii, H3 was most common (found in 49.3% accessions), followed by H2 (in 23.7% accessions). The only difference between H2 and H3 was the positive amplification of Xrj3 in the latter (Table 1). H4 was the third most frequent haplotype (in 9.6% accessions), and differed from H3 in the size of the product of Xms2, which was 368 bp in H4, but 346 bp in H3 (Table 1). Collectively, H2, H3 and H4 were found in 82.6% of the Ae. tauschii accessions examined. The remaining six haplotypes were detected in only small numbers of accessions (Table 1). The marker alleles of Ae. tauschii H2 were identical to those of common wheat H2 (Table 1). Surprisingly, we did not find H1 in the large number of Ae. tauschii accessions analyzed in this work.

Table 1

Glu-D1 locus haplotypes detected in T. aestivum, T. spelta and Ae. tauschii populations.

Haplotype	Marker							Species
	Xms1	Xid1	Xrj1	Xrj2	Xms2	Xrj3	Xrj4
H1	176a	866	351	428	368	460	631	T. spelta (9, 4.2%)c T. aestivum (99, 47.7%)
H2	–b	866	–	1085	346	–	–	Ae. tauschii (84, 23.7%)T. spelta (175, 81.4%)T. aestivum (109, 52.3%)
H3	–	866	–	1085	346	460	–	Ae. tauschii (175, 49.3%)
H4	–	866	–	1085	368	460	–	Ae. tauschii (34, 9.6%)
H5	176	866	–	1085	368	460	–	Ae. tauschii (3, 0.8%)
H6	176	866	–	1085	346	–	–	Ae. tauschii (10, 2.8%)
H7	176	663	–	1085	346	–	–	Ae. tauschii (20, 5.6%)
H8	176	663	351	1085	346	–	–	Ae. tauschii (4, 1.1%)
H9	–	866	351	1085	346	–	–	Ae. tauschii (6, 1.7%)
H10	176	866	351	1085	346	–	–	Ae. tauschii (19, 5.4%)
H11	–	866	–	1085	346	–	631	T. spelta (30, 14.1%)
H12	176	866	351	1085	368	460	631	T. spelta (1, 0.5%)

Length (bp) of amplified fragment by the corresponding marker.

Null allele.

Items in the brackets indicate the number and percentage of lines in which the given Glu-D1 haplotype was detected.

A total of 215 T. spelta accessions (Table S4) was genotyped with the seven Glu-D1 markers, resulting in the identification of four haplotypes (H1, H2, H11, and H12, Table 1). H1, although not found in Ae. tauschii, was detected in T. spelta. On the other hand, H11 and H12 were found in T. spelta, but not Ae. tauschii or common wheat. H2 was the predominant Glu-D1 haplotype, followed by H11. H11 differed from H2 in the positive amplification of Xrj4, and H12 differed from H1 in the size of the product of Xrj2, which was 428 bp in H1, but 1085 bp in H12 (Table 1).

All together, 12 unique haplotypes were identified for Glu-D1 locus in the T. aestivum, Ae. tauschii, and T. spelta materials in this study (Table 1). An important feature shared by H1, H11, and H12 was the positive amplification of Xrj4, which was not the case for the nine Glu-D1 haplotypes in Ae. tauschii. Among the 12 haplotypes, only H1 had the 428 bp allele at Xrj2 (Table 1).

Approximate Timing for the Differentiation of Glu-D1 Haplotypes

To estimate the time periods in which the different Glu-D1 haplotypes were differentiated, transposon insertion and deletion events found in subsets of the 12 haplotypes were investigated, followed by molecular clock analysis using nucleotide substitutions in the LTRs of retroelements. Based on structural differences between the Glu-D1 locus sequences of Renan and AUS18913 (Figure 1A), and the polymorphic amplification patterns of Xrj1, 2 and 3 (Table 1), the 12 haplotypes could be divided into five categories (Table 2). The H2 to H7 haplotypes in category I all contained the insertion of an intact Angela-4 element, but lacked those by Wis-1, Wis-2p and Angela-4p. Molecular dating with Angela-4 indicated that the insertion of this element into Glu-D1 region took place approximately 1.42 MYA (Table 2), suggesting that the six haplotypes in category I were differentiated around 1.42 MYA. H8 to H10 in category II were characterized by possessing both Angela-4 and Wis-1 (Table 2). The insertion of Wis-1 into Glu-D1 region was estimated to be approximately 0.46 MYA based on nucleotide substitutions in the two LTRs of this element (Table 2). Therefore, H8 to H10 arose possibly around 0.46 MYA. The haplotypes in category III (H11), IV (H12) or V (H1) all contained the Wis-2p element with only one LTR. Therefore, the insertion time of Wis-2p could not be dated. However, considering that the Glu-D1 haplotypes harboring Wis-2p were found in only common wheat (H1) and T. spelta (H1, H11 and H12), but not in Ae. tauschii (Table 1), Wis-2p insertion might occur after the hexaploidization event that gave rise to the ancestral hexaploid wheat (around 0.01 MYA). Thus, the Glu-D1 haplotypes in categories III, IV or V might all be differentiated relatively recently compared to those in categories I and II, possibly around 0.01 MYA (Table 2).

Table 2

Approximate differentiation times of five categories of Glu-D1 haplotypes.

TE event	Haplotype	TE used forestimation	Rough timing ofdifferentiation (MYA)	Species
+Angela-4 a- Wis-1- Wis-2p-Angela-4p	H2 to H7	Angela-4	1.42±0.18	Ae. tauschii (H2 to H7)T. spelta (H2)T. aestivum (H2)
+Angela-4+Wis-1- Wis-2p-Angela-4p	H8 to H10	Wis-1	0.46±0.09	Ae. tauschii
+Angela-4- Wis-1+Wis-2p-Angela-4p	H11	Wis-2p	0.01	T. spelta
+Angela-4+Wis-1+Wis-2p-Angela-4p	H12	Wis-2p	0.01	T. spelta
− Angela-4+Wis-1+Wis- 2p+Angela-4p	H1	Wis-2p	0.01	T. speltaT. aestivum

The plus and minus symbols indicate the presence or absence of TE insertion.

The differentiation of H1 and H12 occurred around similar times was further investigated by analyzing the terminal sequences of the LTR element Sabrina-2, which existed in the Glu-D1 region of both Renan and AUS18913 (Figure 1A). This element was inserted into Glu-D1 locus around 2.01–2.30 MYA based on molecular dating using its LTR sequences (Table S6). No nucleotide substitution was found between the LTR sequences of the Sabrina-2 elements resided in H1 and H12, although such substitutions existed between the Sabrina-2 elements hosted by H1 and H5 or by H1 and H10 (Table S7). Further examination revealed that the number of nucleotide substitutions in Sabrina-2 LTR sequences between H1 and H5 was substantially higher than that between H1 and H10 (Table S7). This observation was consistent with the findings that H5 was differentiated earlier than H10, and H10 was formed earlier than H1 (Table 2).

Characterization of 1Dx Genes and their Deduced Proteins

Because a major difference between Glu-D1d and Glu-D1a lies in their 1Dx subunits, we conducted sequence analysis of the 1Dx genes in a range of Ae. tauschii and T. spelta accessions with their Glu-D1 locus haplotyped in this study. A total of 20 unique 1Dx ORFs was analyzed at both coding sequence and deduced protein levels by comparing with the two well characterized common wheat 1Dx genes, 1Dx5 and 1Dx2 that belonged to Glu-D1d or Glu-D1a and were carried by Glu-D1 haplotypes H1 and H2, respectively. The 20 1Dx ORFs were all intact and active because their products could be found in the grains by SDS-PAGE analysis of seed protein extracts (Figure S3).

Like the coding sequences of 1Dx5 and 1Dx2, the newly cloned 1Dx ORFs contained no intron. The 20 deduced 1Dx proteins from these accessions shared a highly similar primary structure consisting of a signal peptide (21 residues), a N-terminal domain (89 residues), a central repetitive domain (681 to 702 residues), and a C-terminal domain (42 or 43 residues) (Table 3). Their central repetitive domains were composed of mainly tandem and interspersed repeats with base units of tripeptide, hexapeptide, and nonapeptide motifs (Figure S4). For eight 1Dx subunits, the number of conserved cysteine residues and their relative positions in the protein sequence were identical to those in 1Dx5 (Table 3, Figure S4). For the remaining 12 1Dx subunits, the number of conserved cysteine residues and their relative positions in the protein sequence were the same as those in 1Dx2 (Table 3, Figure S4). Interestingly, the eight accessions with 1Dx5-like subunits all belonged to T. spelta, and their Glu-D1 haplotype was H1; for the 12 accessions with 1Dx2-like subunits, nine belonged to Ae. tauschii and three to T. spelta (Table 3). Subsequently, we compared the substitution and indel patterns among 1Dx5, 1Dx2, and the 20 1Dx subunits. A total of 15 polymorphic sites, including 10 substitutions (S1 to S10) and five indels (ID1 to ID5), were analyzed. As shown in Table S8, each of the ten substitutions occurred between two residues. The S1 substitution, S118C, was linked to functional difference between 1Dx2 and 1Dx5 (see Introduction). The indels occurred at the five sites caused removal or addition of short peptides (Table S8). In general, the 1Dx subunits specified by Glu-D1 haplotypes H3 to H5 showed the highest dissimilarities from both 1Dx5 and 1Dx2 in the 15 sites (Table S8). Remarkably, the 1Dx2-like subunit 1Dsx-TRI19057H12, although lacking S118C substitution, strongly resembled 1Dx5 and 1Dx5-like subunits in the majority of the remaining sites; three 1Dx2-like subunits (1Dtx-PI511368H2, 1Dsx-PI348360H2 and 1Dsx-TRI5008H11) were either highly similar, or identical to, 1Dx2 in these 15 sites (Table S8). Apart from the substitutions and indels described above, there also existed additional changes that were either specific to individuals, or shared by subsets, of the 20 compared 1Dx protein sequences (Figure S4). For example, extra and unique cysteine residues were found in the repetitive domain of the 1Dx5-like subunits 1Dsx-TRI9883H1 and 1Dsx-TRI16607H1 and the 1Dx2-like subunit 1Dsx-TRI5008H11.

Table 3

Main features of the 20 newly isolated 1Dx subunits.

1Dx	ORF size (bp)	Deduced protein (aa)
		SP	ND	RD	CD	Total
1Dx5	2550	21	89 (3)a	696 (1)a	42 (1)a	848 (5)b
1Dx2	2523	21	89 (3)	687 (0)	42 (1)	839 (4)
1Dsx-PI15865H1	2550	21	89 (3)	696 (1)	42 (1)	848 (5)
1Dsx-PI361813H1	2550	21	89 (3)	696 (1)	42 (1)	848 (5)
1Dsx-TRI9883H1	2550	21	89 (3)	696 (2)	42 (1)	848 (6)
1Dsx-TRI16607H1	2550	21	89 (3)	696 (2)	42 (1)	848 (6)
1Dsx-TRI16981H1	2550	21	89 (3)	696 (1)	42 (1)	848 (5)
1Dsx-PI347904H1	2550	21	89 (3)	696 (1)	42 (1)	848 (5)
1Dsx-PI348150H1	2550	21	89 (3)	696 (1)	42 (1)	848 (5)
1Dsx-PI348171H1	2550	21	89 (3)	696 (1)	42 (1)	848 (5)
1Dsx-PI348360H2	2523	21	89 (3)	687 (0)	42 (1)	839 (4)
1Dtx-PI511368H2	2532	21	89 (3)	690 (0)	42 (1)	842 (4)
1Dtx-IG46663H3	2541	21	89 (3)	693 (0)	42 (1)	845 (4)
1Dtx-PI603224H4	2541	21	89 (3)	693 (0)	42 (1)	845 (4)
1Dtx-IG48561H5	2541	21	89 (3)	693 (0)	42 (1)	845 (4)
1Dtx-PI603236H6	2505	21	89 (3)	681 (0)	42 (1)	833 (4)
1Dtx-CIAE24H7	2541	21	89 (3)	693 (0)	42 (1)	845 (4)
1Dtx-TA2527H8	2541	21	89 (3)	693 (0)	42 (1)	845 (4)
1Dtx-PI349047H9	2568	21	89 (3)	702 (0)	43 (1)	855 (4)
1Dtx-PI603223H10	2568	21	89 (3)	702 (0)	43 (1)	855 (4)
1Dsx-TRI5008H11	2523	21	89 (3)	687 (1)	42 (1)	839 (5)
1Dsx-TRI19057H12	2523	21	89 (3)	687 (0)	42 (1)	839 (4)

ORF: open reading frame; SP: signal peptide; ND: N-terminal domain; RD: repetitive domain; CD: C-terminal domain. The 1Dx5-like and 1Dx2-like subunits are shown in bold and underlined, respectively.

Items in the brackets indicate the number of cysteine residues in the ND, RD or CD regions of each 1Dx protein.

The number of total cysteine residues in each 1Dx subunit is provided in the brackets.

Phylogenetic Analysis of 1Dx Genes and Estimation of 1Dx Divergence Times

Phylogenetic analysis was conducted to investigate the relatedness among the 20 1Dx genes. In a typical neighbor joining tree constructed with full length nucleotide sequences (Figure 2), there were three well supported clades (C1 to C3), with C1 and C2 in the upper and parallel positions and C3 located at the basal side. Six major branches, including three (B1 to B3) in C1, two (B4 to B5) in C2, and one (B6) in C3 (Figure 2), could be recognized. Within C1, the top branch B1 was formed by 1Dx5 from common wheat and the seven 1Dx5-like genes from T. spelta, the middle branch B2 was composed of the 1Dx2-like gene 1D and the 1Dx5-like gene 1D, and the basal branch B3 contained two 1Dx2-like genes from Ae. tauschii (Figure 2). Within C2, B4 was composed of exclusively 1Dx2-like genes from Ae. tauschii, and B5 of 1Dx2 from common wheat and three 1Dx2-like genes from Ae. tauschii and T. spelta (Figure 2). C3 contained the 1Dx2-like genes from only Ae. tauschii (Figure 2). The 1Dx genes aggregated in the three clades were derived from separate sets of Glu-D1 haplotypes, namely, H1, H9, H10 and H12 for the genes in C1, H2, H6, H7, H8 and H11 for those in C2, and H3, H4 and H5 for those in C3 (Figure 2). Phylogenetic trees were also constructed using full length amino acid sequences and an alternative program (i.e., minimum evolution). The result was identical to that displayed in Figure 2.

Figure 2

Phylogenetic analysis of 1Dx genes.

The tree shown was constructed using the multiple alignment of nucleotide sequences, and the neighbor joining program. Three distinctive clades (C1 to C3) and six major branches (B1 to B6) were observed. Highly similar trees were obtained with multiple alignment of deduced protein sequences, and an alternative tree building programs (i.e., minimum evolution). The 1Dx5-like and 1Dx2-like genes are shown in bold and underlined, respectively. 1Ax1 from common wheat Glu-A1 locus was used as an outgroup control. Bootstrap values were obtained with 1000 permutations. The GenBank accession numbers for 1Dx5, 1Dx2 and 1Ax1 are X12928, X03346 and X61009, respectively.

On the basis of the phylogenetic analysis, the divergence times between the 1Dx genes of hexaploid wheat and Ae. tauschii were estimated. In the C1 cluster, the averaged divergence time between 1Dx5 (representing hexaploid wheat 1Dx5 and 1Dx5-like genes) and the two Ae. tauschii 1Dx2-like genes (1D and 1D) was found to be around 0.29 MY (Table S9). In C2, the averaged divergence time between 1Dx2 (representing hexaploid wheat 1Dx2 and 1Dx2-like genes) and the four Ae. tauschii 1Dx2-like genes (1D, 1D, 1D and 1D) was approximately 0.13 MY, with the specific time between 1Dx2 and 1D being only about 0.04 MYA (Table S9). Interestingly, the three Ae. tauschii 1Dx2-like genes in C3 exhibited relatively the longest averaged divergence times from both 1Dx5 (0.37 MY) and 1Dx2 (0.36 MY) (Table S9).

Analysis of 1Dy Genes and their Deduced Proteins

Initially, the 1Dy genes hosted by different Glu-D1 haplotypes were analyzed using the UMN26 marker that permits the distinction of 1Dy10 from 1Dy12 [33]. By comparing with the fragments amplified from the common wheat varieties Bobwhite (carrying 1Dy10) and CS (harboring 1Dy12), 1Dy10 (or 1Dy10-like) was found in T. spelta haplotypes H1 and H12 and Ae. tauschii haplotypes H9 and H10, with 1Dy12 (or 1Dy12-like) present in the remaining haplotypes (Figure S5). To verify the presence of intact 1Dy10 ORF, the 1Dy coding sequence in the Ae. tauschii accession PI603223 (with haplotype H10) and the T. spelta accession TRI19057 (with haplotype H12) were cloned. To facilitate further comparisons, the 1Dy coding sequence in the two Ae. tauschii accessions PI511368 and IG48561 (with haplotypes H2 and H5, respectively) and the T. spelta accession PI348360 (representing the H2 haplotype in spelt wheat) were also isolated. The expression of these 1Dy genes was verified by SDS-PAGE analysis of seed protein extracts (Figure S3).

Like 1Dy10 and 1Dy12, the five 1Dy sequences (designated as 1D, 1D, 1D, and 1D, respectively) did not contain intron, and were all terminated by tandem stop codons. The primary structure of their deduced proteins was identical to that of 1Dy10 and 1Dy12 (Figure S6). Nucleotide sequence comparisons revealed that 1D was identical to 1Dy10, and 1D showed higher identity to 1Dy10 (99.6%) than to 1Dy12 (96.7%). In contrast, 1D was identical to 1Dy12, and 1D exhibited higher identity to 1Dy12 (99.7%) than to 1Dy10 (96.7%). 1D displayed comparable identities to both 1Dy10 (96.2%) and 1Dy12 (97.9%).

Phylogenetic Network Analysis of Glu-D1 Haplotypes and a Possible Homologous Recombination Event between H5 and H10

Based on the variations of 1Dx and 1Dy genes and the similarities and differences in the alleles of the seven DNA markers (Figure S7), a phylogenetic network analysis was conducted to investigate relationships among the 12 haplotypes. In Ae. tauschii, direct network connection was found among H2, H3, H4, H5, H6, H7 and H8, as well as between H9 and H10 (Figure 3). In common wheat and T. spelta, direct connection was found between H12 and H1, and between H2 and H11 (Figure 3). Two hypothetical intermediate haplotypes were predicted by the analysis program, one (X) connecting H6, H7 and H8 to H9 and H10, and the other (Y) linking H5 and H10 to H12 (Figure 3).

Figure 3

Phylogenetic network analysis of 12 Glu-D1 haplotypes.

The haplotypes detected in Ae. tauschii were circled, whereas those in T. spelta and common wheat were boxed. H2 was a shared haplotype. Two hypothetical intermediate haplotypes (X and Y) were predicted by the network analysis program.

The linkage of both H5 and H10 to H12 was of particular interest considering the direct relationship between H12 and H1 that carries Glu-D1d. A simple explanation could be that both H5 and H10 might contribute to the differentiation of H12 through a recombination event. Judging from the alleles of 1Dx, 1Dy and the seven diagnostic markers, H12 resembled H10 in the upstream portion and including Xrj2 (Figure 4). In the downstream portion of Xrj2, H12 shared two identical marker alleles (specified by Xms2 and Xrj3, respectively) with H5, although the two haplotypes differed in the alleles of 1Dx and Xrj4 (Figure 4). We thus speculated that a homologous recombination event might have happened between H5 and H10 at a position downstream of 1Dx, resulting in the prototype of H12 (Figure 4). In the prototype H12, the segment with the 1Dx and 1Dy genes was dissented from H10, whereas the remaining segment was derived from H5. Since neither H5 nor H10 were detected in extant common wheat and T. spelta (Table 1), the recombination event yielding the prototype H12 occurred in Ae. tauschii.

Figure 4

A possible homologous recombination event between Glu-D1 haplotypes H5 and H10.

This event might take place in Ae. tauschii between H5 and H10 downstream of the 1Dx gene (as indicated by the cross). The descendent was introduced to the ancestral hexaploid wheat via hexaploidization, giving rise to H12 after further differentiation (as evidenced by the positive amplification of Xrj4 due to Wis-2p insertion, boxed area). The types of 1Dx and 1Dy genes and the alleles of the seven Glu-D1 markers hosted by the different haplotypes were indicated.

Consistent with the homologous recombination hypothesized above, it was found that 1D and 1D showed higher identities to 1Dx5 and 1Dy10 (98.8% and 99.6%, respectively) than to 1Dx2 and 1Dy12 (96.2% and 96.7%, respectively). On the other hand, 1D and 1D, derived from the Ae. tauschii accession with haplotype H5, exhibited lower levels of identities to 1Dx5 and 1Dy10 (97.5% and 96.2%, respectively).

Discussion

In this study, three complementary approaches were taken in order to investigate haplotype variation of Glu-D1 and the origin of Glu-D1d. The development and application of seven diagnostic markers distributed in Glu-D1 region permitted a rapid survey of the haplotype variation of this locus in several hundred germplasm lines. Molecular cloning of 1Dx and 1Dy coding sequences provided accurate gene sequence information in important Glu-D1 haplotypes. Phylogenetic analysis and molecular dating were deployed to understand the relationships and divergence times of different Glu-D1 haplotypes and the associated 1Dx and 1Dy genes. The new insights gained and their implications for further research are discussed below.

Major Characteristics of Glu-D1 Haplotypes

First, the number of Glu-D1 haplotypes is highest in Ae. tauschii (totally nine), intermediate in T. spelta (four), and relatively low in common wheat (only two). Considering that more than two hundred diverse common wheat varieties had been haplotyped, the low number of Glu-D1 haplotypes in common wheat is unlikely due to a biased selection of accessions. Since more than 200 T. spelta lines were genotyped, the four haplotypes revealed is likely a reasonable representation of the variation of Glu-D1 region in this species.

Second, most Glu-D1 haplotypes are not shared among Ae. tauschii, T. spelta, and T. aestivum, and haplotypes H1, H11 and H12 are not present in Ae. tauschii. Among the 12 haplotypes detected, only H2 was shared among common wheat, T. spelta and Ae. tauschii, and H1 by common wheat and T. spelta. H2 was the dominant haplotype in both common wheat and T. spelta, and is the second most abundant haplotype in Ae. tauschii. In contrast, H1 was not detected in Ae. tauschii, and a relatively rare haplotype (compared to H2 and H11) in T. spelta. The failure in detecting H1 in Ae. tauschii is consistent with the fact that typical 1Dx5 subunit with the S118C substitution has not been found in this species by this and previous studies [44], [45]. Therefore, it is highly probable that H1 (and the Glu-D1d allele contained therein) is absent in Ae. tauschii. A potential concern on this suggestion is that H1 may exist in Ae. tauschii in a very low frequency and was not detected by the present study. However, the failure of detecting H1 in 355 diverse Ae. tauschii accessions collected from over 20 countries (Table 1, Table S3), the presence of Xrj2 and Xrj4 in H1 but none of the Ae. tauschii Glu-D1 haplotypes (Table 1, Figure S7), and the coincidence between the times of H1 differentiation and hexaploidization (Table 2) are strongly in favor of the emergence of H1 in hexaploid wheat but not Ae. tauschii. Although detected in T. spelta, H11 and H12 were also not found in Ae. tauschii. Because H11 and H12 shared the Wis-2p insertion with H1 (Table 1), and may be differentiated around the same time as H1, it is possible that, like H1, H11 and H12 are also absent in Ae. tauschii.

Third, some Glu-D1 haplotypes are relatively ancient, whereas others may have emerged only recently. H2 to H10 were already present in Ae. tauschii before the hexaploidization event around 10,000 years ago. In contrast, H1, H11, and H12 may be emerged in more recent times (around 0.01 MYA, Table 2), coinciding with the time frame in the formation of the ancestral hexaploid wheat. Among H2 to H7 (differentiated around 1.42 MYA), H3 to H5 might be differentiated earlier, and H2, H6 and H7 might be derived from H3 to H5 (after further sequence deletion in Sabrina-3s). The direct derivation of H2 from H3 is also supported by a straight connection between the two haplotypes in the phylogenetic network analysis (Figure 3). On the other hand, H8 to H10 were differentiated later than H2 to H7 (around 0.46 MYA, Table 2).

Fourth, comparative analyses among 1Dx or 1Dy genes provide further clues on the differentiation of Glu-D1 haplotypes. The 1Dx alleles carried by H3 to H5 clustered together (Figure 2), and exhibited relatively the longest divergence times to both 1Dx5 and 1Dx2 (Table S9), supporting the idea that H3 to H5 may be the most ancient ones among the 12 haplotypes. The 1Dx alleles from H1, H9, H10 and H12 aggregated together, with those from H9 and H10 at the basal position (Figure 2), pointing to the involvement of H9 and H10 in the evolution of H12 and H1. The 1Dx alleles from H2, H6 to H8, and H11 formed another well supported cluster (Figure 2), with the 1Dx allele from the Ae. tauschii H2 haplotype at the most basal position, indicating that the Ae. tauschii H2 is important for the evolution of H6 to H8, the presence of H2 in T. spelta and common wheat, and the emergence of H11 in T. spelta. The discovery of 1Dy10-like gene in H10 and 1Dy10 in H12 further supports the role of the two haplotypes in the evolution of H1 that carries 1Dx5 and 1Dy10. On the other hand, the finding of 1Dy12-like gene in Ae. tauschii H2 and 1Dy12 in T. spelta H2 reinforces the likelihood that the H2 in T. spelta and common wheat was directly derived from the H2 of Ae. tauschii due to hexaploidization. Lastly, the direct relationship between Ae. tauschii H2 and the H2 in hexaploid wheat was also consistent with the very short divergence time between the 1Dx2-like gene (e.g., 1D) in Ae. tauschii and 1Dx2 of common wheat (Table S9).

Finally, homologous recombination may contribute to the diversification of Glu-D1 haplotypes in Ae. tauschii. In higher plants, haplotype variation and diversification of genes and chromosomal loci are usually caused by multiple means [46]. From the data presented in Figure 1 and Table 1, it is clear that haplotype variation of Glu-D1 is caused by several reasons. These include nucleotide sequence alterations in two microsatellites (Xms1 and Xms2), the indel in Xid1, and the mutations associated with four TEs (Wis-1, Angela-4, Sabrina-3s, and Wis-2p). Furthermore, we hypothesized that a homologous recombination event could have happened between H5 and H10 in Ae. tauschii, leading to further diversification of Glu-D1 haplotypes (Figure 4). The involvement of homologous recombination in the differentiation of new haplotypes has also been reported in other plant species [47], [48].

Important Steps in the Differentiation of H1 Haplotype Containing Glu-D1d

Given the discussion above, the network relationships among 12 Glu-D1 haplotypes (Figure 3), and the hypothetical recombination between H5 and H10 (Figure 4), a simplified model could be drawn on possible derivative relationships among 12 Glu-D1 haplotypes. In this model (Figure S8), H3, H4 and H5 are probably the founder haplotypes, H2 (derived from H3), H5 and H10 play an important role in the diversification of Glu-D1 haplotypes. H2 is likely to give rise to a number of haplotypes (H6 to H10) in Ae. tauschii, and to H11 after being introduced into hexaploid wheat through hexaploidization.

Compared to H2 and H11, the route of H1 differentiation is more complex. Three steps may be important for the evolution of H1 containing Glu-D1d (Figure S8). The first one is the differentiation of H10, hosting a 1Dx2-like allele and a 1Dy10-like allele in Ae. tauschii. This step likely occurred around 0.46 MYA. The second step is the differentiation of H12 in the ancestral hexaploid wheat. This step might take place around 0.01 MYA. The source material for H12 differentiation might be the descendant from the homologous recombination event between H5 and H10, which was introduced to the ancestral hexaploid wheat via hexaploidization. The 1Dx2-like allele in H12, although encoding a 1Dx subunit lacking the S118C substitution, was highly similar to 1Dx5 (Figure 2, Table S8). The 1Dy gene in H12 was identical to 1Dy10. The third step is the differentiation of H1, containing Glu-D1d allele and specifying the expression of 1Dx5 (or 1Dx5-like) and 1Dy10 subunits, from H12. This step might happen during or shortly after hexaploidization, with the resultant H1 retained in both common wheat and T. spelta populations.

Impact of Hexaploidization and Domestication on Glu-D1 Haplotypes

As a hexaploid crop species, the genetic structure of common wheat has been greatly influenced by hexaploidization, domestication, and breeding selection [31]. The strong effects of these processes on the birth and dissemination of new haplotypes are well demonstrated by recent studies on the powdery mildew resistance gene Pm3 and the Lr34 gene conferring multi-pathogen resistance [49]-[51]. A preponderance of the haplotype data obtained in this study leads us to suggest that the differentiation and dissemination of Glu-D1 haplotypes might have also been affected by hexaploidization, domestication, and breeding in at least two aspects.

First, as noted above, the number of Glu-D1 haplotypes detected varied substantially among Ae. tauschii, T. spelta, and common wheat. The most probable explanation for the low degree of haplotype diversity in common wheat may be genetic bottlenecks associated with hexaploidization (i.e., formation of the ancestral hexaploid wheat by hybridization between tetraploid wheat and Ae. tauschii) and domestication and breeding (from ancestral hexaploid wheat to extant common wheat varieties). It is well known that nucleotide diversity is substantially higher in Ae. tauschii genome than in the D genome of contemporary common wheat [52], and that relative few types (or individuals) of Ae. tauschii were involved in the formation of ancestral hexaploid wheat [53]–[55]. Moreover, it is widely accepted that human selection tends to further reduce the genetic diversity of plant and animal populations [31]. As land races, T. spelta accessions are less affected by artificial breeding efforts, and may have thus preserved more genetic diversities inherited from the ancestral hexaploid wheat [28], [29], [31]. This may, at least partly, be responsible for the existence of relatively more Glu-D1 haplotypes in T. spelta than in common wheat. Several studies have speculated that T. spelta may be evolved from secondary hybridization between ancient free-threshing hexaploid wheat with hulled emmer wheat [28], [29], [56], [57]. This process might have also affected the differentiation and retention of Glu-D1 haplotypes in T. spelta. Thus, the spectrum of T. spelta Glu-D1 haplotypes revealed by this work might have been shaped by multiple forces (e.g., secondary hybridization, less artificial selection).

Second, the differentiation and dissemination of H1 (carrying Glu-D1d allele) might have been more intimately connected with hexaploidization and human selection. Terasawa et al. [58] found that the frequency of Glu-D1d was exceptionally high in the hexaploid wheat lines collected from Caucasian regions. Considering that common wheat originated near Caucasia [53], it is possible that H1 might be differentiated during or shortly after the hexaploidization event in Caucasia, and was subsequently selectively enriched in the free threshing common wheat populations around Caucasian regions due to the beneficial effects of Glu-D1d on end-use qualities [58]. As found in this work, H1 was a relatively rare haplotype compared to H2 and H11 in T. spelta, but its presence in extant common wheat almost equaled to that of H2 (Table 1). Because common wheat has been subjected to breeder’s selection more strongly than T. spelta, the increased presence of H1 in contemporary common wheat is largely caused by artificial breeding.

Implications for Further Research

Apart from common wheat, there have also been substantial interests in recent years in identifying the elite alleles of SSPs in other crops, such as rice (e.g., [59], [60]), maize (e.g., [61]–[63]), and soybean (e.g., [5], [64]). As exemplified by this work, haplotype analysis with appropriate germplasm populations may be employed as an efficient strategy for increasing the understanding of SSP alleles, especially for those harbored by complex genomic loci and potentially impacted by genome polyploidization events. The resulting new knowledge may facilitate more effective use of SSP alleles in improving the quality traits of crops.

In this work, we found that the deduced protein sequences of the eight T. spelta 1Dx5-like subunits, although all carrying the S118C substitution (Table S8), differed considerably from that of the 1Dx5 subunit in common wheat (Figure S9). These Glu-D1d variants are likely valuable materials for further improvement of common wheat end-use qualities. We are now developing chromosomal introgression lines containing T. spelta Glu-D1d variants in common wheat background, which may be useful for testing the functionality of these new Glu-D1d variants. The seven DNA markers developed for Glu-D1 region are being used for monitoring the transfer of Glu-D1d variants from T. spelta to common wheat. These markers may also be used in other end-use quality improvement programs involving the transfer of Glu-D1d, which are ongoing in many wheat production regions of the world [4], [65].

Validation of the chromosomal specificity of the seven newly developed markers in this study. Genomic DNA samples were extracted from Chinese Spring (CS), the nulli-tetrasomic line N1DT1A (lacking 1D chromosome and associated Glu-D1 locus), the ditelosomic line of chromosomal arm 1DS (Dt1DS, lacking 1DL and associated Glu-D1 locus), three additional common wheat genotypes (Attila, Bobwhite and Kenong 199), two tetraploid wheat genotypes (Langdon and Cofa), and two Ae. tauschii accessions (AUS18913 and AL8/78). They were then used in genomic PCR with the primer pairs for each of the seven markers (Xms1, Xid1, Xrj1, Xrj2, Xms2, Xrj3, and Xrj4). These markers did not amplify products in tetraploid wheat genotypes, N1DT1A, and Dt1DS, confirming that they are D genome specific, and that their target sequences are located on chromosomal 1DL on which Glu-D1 resides. As anticipated, the seven markers showed polymorphisms in common wheat and between common wheat and Ae. tauschii. Lane M contains DNA size standard. The size (bp) of the amplified PCR fragments is shown on the left side of the graph. Items in the brackets indicate genome constitution of the examined wheat or Ae. tauschii genotypes.

(PDF)

Click here for additional data file.

Multiple alignment of the DNA sequences of amplicons from Attila, Bobwhite, Chinese Spring (CS) and Kenong 199, as well as the reference sequences from Renan and AUS18913. Chinese Spring (CS) and Kenong 199, as well as the reference sequences from Renan and AUS18913. The 22 bp deletion in the amplicons from CS, Kenong 199 and AUS18913 was indicated by dashed line.

(PDF)

Click here for additional data file.

Detection of 1Dx and 1Dy subunits expressed by and Haplotypes. The 1Dx and 1Dy HMW-GSs in the two common wheat varieties, Bobwhite (expressing 1Dx5 and 1Dy10) and Chinese Spring (CS, expressing 1Dx2 and 1Dy12), and the T. spelta and Ae. tauschii accessions with different Glu-D1 haplotypes were revealed by SDS-PAGE analysis of seed protein extracts. 1Dx and 1Dy subunits specified by different T. spelta and Ae. tauschii Glu-D1 haplotypes were indicated by arrowheads and arrows, respectively. The protein band labeled by asterisk in H7 is a ω-gliadin protein.

(PDF)

Click here for additional data file.

Multiple alignment of the deduced amino acid sequences of 1Dx5 and 1Dx2 subunits from common wheat, and the 20 1Dx subunits from and . The signal peptide is underlined, while the N- and C-terminal domains are labeled bold and bold italic, respectively. The repetitive domain is situated between the N- and C-terminal domains. The amino acid substitutions (S1 to S10) and indels (ID1 to ID5) between 1Dx5 and 1Dx2, and their variations in the 20 1Dx subunits from T. spelta and Ae. tauschii, are indicated. The four cysteine residues conserved among the 22 compared 1Dx subunits are marked by arrowheads. In addition, the cysteine reside unique to 1Dsx-TRI9883H1, 1Dsx-TRI16607H1 or 1Dsx-TRI5008H11 is boxed.

(PDF)

Click here for additional data file.

Detection of genes in the two common wheat varieties, Bobwhite (harboring ) and Chinese Spring (CS, containing ), and the and accessions with different haplotypes using the PCR marker UMN26. This marker is co-dominant for 1Dy10 and 1Dy12, and allows the distinction between the two genes. Lane M contains DNA size standard (bp).

(PDF)

Click here for additional data file.

Multiple alignment of the amino acid sequences of 1Dy10 and 1Dy12 subunits from common wheat and the five 1Dy subunits from and . The signal peptide is underlined, while the N- and C-terminal domains are labeled bold and bold italic, respectively. The repetitive domain resides between the N- and C-terminal domains. The amino acid substitutions (S1 to S12) and indels (ID1 to ID3) patterns between 1Dy10 and 1Dy12, and their variations in the five 1Dy subunits from Ae. tauschii, are indicated. The seven cysteine residues conserved among the seven compared 1Dy subunits are labeled by arrowheads.

(PDF)

Click here for additional data file.

A diagram illustrating the variations of and genes and the alleles of the seven DNA markers in 12 haplotypes. The types of 1Dx and 1Dy genes and the alleles of the seven Glu-D1 markers were indicated.

(PDF)

Click here for additional data file.

A simplified model illustrating possible derivative relationships among 12 haplotypes. The arrows indicate the likely directions of haplotype differentiation. The approximate differentiation times (MYA) of the haplotypes in Ae. tauschii and hexaploid wheat were indicated. The three haplotypes (H5, H10 and H12) involved in the differentiation of H1 are labeled in bold. H2, detected in both Ae. tauschii and hexaploid wheat, was shown in italic.

(PDF)

Click here for additional data file.

A diagram illustrating the positions of polymorphic sites among the 1Dx5 subunits from two common wheat varieties (Cheyenne and Renan) and the eight 1Dx5-like subunits from genotypes. 1Dx5 protein sequence (GenBank accession CAA31395) from Cheyenne was used as reference for locating the positions of polymorphic sites, where amino acid substitutions were detected. The diagram is not drawn in scale.

(PDF)

Click here for additional data file.

Haplotype variation of locus in the common wheat varieties with known information on their allele.

(XLS)

Click here for additional data file.

Haplotype variation of locus in the common wheat lines from different countries.

(XLS)

Click here for additional data file.

haplotypes detected in 355 accessions.

(XLS)

Click here for additional data file.

haplotypes detected in 215 accessions.

(XLS)

Click here for additional data file.

The oligonucleotide primers used in this study.

(DOC)

Click here for additional data file.

Estimation of approximate insertion time of into locus.

(DOC)

Click here for additional data file.

Number of nucleotide substitutions between the LTR sequences in the elements resided in several haplotypes.

(DOC)

Click here for additional data file.

Amino acid substitutions (S1 to S10) and indels (ID1 to ID5) between 1Dx5 and 1Dx2 of common wheat and their variations in the 20 1Dx subunits from or

(DOC)

Click here for additional data file.

Divergence time estimates between two representative hexaploid wheat genes ( and ) and genes.

(DOC)

Click here for additional data file.

Additional methods.

(DOC)

Click here for additional data file.