Structural Characterization and Evolutionary Relationship of High-Molecular-Weight Glutenin Subunit Genes in Roegneria nakaii and Roegneria alashanica

The Roegneria of Triticeae is a large genus including about 130 allopolyploid species. Little is known about its high-molecular-weight glutenin subunits (HMW-GSs). Here, we reported six novel HMW-GS genes from R. nakaii and R. alashanica. Sequencing indicated that Rny1, Rny3, and Ray1 possessed intact open reading frames (ORFs), whereas Rny2, Rny4, and Ray2 harbored in-frame stop codons. All of the six genes possessed a similar primary structure to known HMW-GS, while showing some unique characteristics. Their coding regions were significantly shorter than Glu-1 genes in wheat. The amino acid sequences revealed that all of the six genes were intermediate towards the y-type. The phylogenetic analysis showed that the HMW-GSs from species with St, StY, or StH genome(s) clustered in an independent clade, varying from the typical x- and y-type clusters. Thus, the Glu-1 locus in R. nakaii and R. alashanica is a very primitive glutenin locus across evolution. The six genes were phylogenetically split into two groups clustered to different clades, respectively, each of the two clades included the HMW-GSs from species with St (diploid and tetraploid species), StY, and StH genomes. Hence, it is concluded that the six Roegneria HMW-GS genes are from two St genomes undergoing slight differentiation.


Introduction
High-molecular-weight glutenin subunits (HMW-GSs) are the major seed storage proteins in the endosperm of wheat and its related species. In common wheat (Triticum aestivum L., AABBDD), HMW-GSs are the critical determinants of wheat processing quality as their composition affects the viscoelastic properties of dough [1,2]. It is well-known that the common wheat HMW-GSs are encoded by three Glu-1 loci, namely Glu-A1, Glu-B1, and Glu-D1, located on the long arm of the group 1 homeologous chromosomes [3,4], and each locus comprises of two tightly linked genes, encoding one x-type and one y-type subunit, respectively [5,6].
Since the 1980s, a series of HMW-GS genes have been isolated from wheat cultivars, and their structural characteristics have been molecularly documented in great detail [7][8][9][10][11]. Orthologous HMW-GSs have also been found, and some of their coding genes have been isolated from several Triticeae grasses, such as Aegilops, Australopyrum, Crithopsis, Elymus, Eremopyrum, Leymus, A striking feature of these Roegneria HMW-GSs was their considerably faster electrophoretic mobilities compared to the subunits from common wheat. Additionally, the mobility of the largest subunit (in GT1 of R. nakaii) was faster than that of the 1Dy12 subunit of Chinese Spring, and these were distributed in the same region wherein low-molecular-weight glutenin subunits (LMW-GSs) or ω-gliadins appeared ( Figure 1). In addition, there were some other protein bands with slower electrophoretic mobility on SDS-PAGE than those of the HMW-GS homologs which were not recognized by the antibody. We speculated that those protein bands might be gliadin homologs. A striking feature of these Roegneria HMW-GSs was their considerably faster electrophoretic mobilities compared to the subunits from common wheat. Additionally, the mobility of the largest subunit (in GT1 of R. nakaii) was faster than that of the 1Dy12 subunit of Chinese Spring, and these were distributed in the same region wherein low-molecular-weight glutenin subunits (LMW- GSs) or ω-gliadins appeared ( Figure 1). In addition, there were some other protein bands with slower electrophoretic mobility on SDS-PAGE than those of the HMW-GS homologs which were not recognized by the antibody. We speculated that those protein bands might be gliadin homologs.

Cloning and Sequence
Analyzing of HMW-GS Genes from R. nakaii and R. alashanica Since the HMW-GS genes do not harbor introns, we thus used a pair of degenerate PCR primers, P1/P2, to amplify the coding sequence of HMW-GS from the genomic DNA of R. nakaii (GT1 and GT2) and R. alashanica [9]. The amplicons displayed 4-5 bands for each genotype on agarose gel electrophoresis ( Figure 2). The PCR fragments between 1000 and 2000 bp, which were HMW-GS genes, were excised, purified and sequenced. Overall, six HMW-GS genes were obtained from the two Roegneria species. Among them, four genes (named Rny1-Rny4) were derived from R. nakaii (Rny1 and Rny2 from GT2, and Rny3 and Rny4 from GT1) and two genes (named Ray1 and Ray2) from R. alashanica. Since the HMW-GS genes do not harbor introns, we thus used a pair of degenerate PCR primers, P1/P2, to amplify the coding sequence of HMW-GS from the genomic DNA of R. nakaii (GT1 and GT2) and R. alashanica [9]. The amplicons displayed 4-5 bands for each genotype on agarose gel electrophoresis ( Figure 2). The PCR fragments between 1000 and 2000 bp, which were HMW-GS genes, were excised, purified and sequenced. Overall, six HMW-GS genes were obtained from the two Roegneria species. Among them, four genes (named Rny1-Rny4) were derived from R. nakaii (Rny1 and Rny2 from GT2, and Rny3 and Rny4 from GT1) and two genes (named Ray1 and Ray2) from R. alashanica. Sequence analysis revealed that the lengths of Rny1, Rny2, Rny3, Rny4, Ray1, and Ray2 genes were 1371 bp, 1253 bp, 1452 bp, 1280 bp, 1389 bp, and 1217 bp (including the tandem stop codons), respectively. Rny1, Rny3, and Ray1 possessed intact open reading frames (ORF), while Rny2, Rny4, and Ray2 contained one (Rny4) or two (Rny2 and Ray2) in-frame stop codons at the central repetitive domain. Moreover, there was a single base deletion at the C-terminal of each of these three genes (1223rd in Rny2, 1250th in Rny4, and 1187th in Ray2) which led to frame shift. BLAST analysis of the entire sequences indicated that the six novel genes were highly homologous to the wheat HMW-GS genes, and were similar to the y-type HMW-GS genes. The sequences of the six HMW-GS genes from Roegneria species have been deposited into GenBank [30] with the accession numbers KP121399 for Ray2, KP121400 for Ray1, KP121401 for Rny2, KP121402 for Rny1, KP121403 for Rny4, and KP121404 for Rny3 (Table 1). Sequence analysis revealed that the lengths of Rny1, Rny2, Rny3, Rny4, Ray1, and Ray2 genes were 1371 bp, 1253 bp, 1452 bp, 1280 bp, 1389 bp, and 1217 bp (including the tandem stop codons), respectively. Rny1, Rny3, and Ray1 possessed intact open reading frames (ORF), while Rny2, Rny4, and Ray2 contained one (Rny4) or two (Rny2 and Ray2) in-frame stop codons at the central repetitive domain. Moreover, there was a single base deletion at the C-terminal of each of these three genes (1223rd in Rny2, 1250th in Rny4, and 1187th in Ray2) which led to frame shift. BLAST analysis of the entire sequences indicated that the six novel genes were highly homologous to the wheat HMW-GS genes, and were similar to the y-type HMW-GS genes. The sequences of the six HMW-GS genes from Roegneria species have been deposited into GenBank [30] with the accession numbers KP121399 for Ray2, KP121400 for Ray1, KP121401 for Rny2, KP121402 for Rny1, KP121403 for Rny4, and KP121404 for Rny3 (Table 1). Multiple alignment analyses indicated that Rny1, Rny3, and Ray1 were highly identical at the nucleotide sequence level. Sixteen single nucleotide polymorphisms (SNPs) and two DNA fragment indels (63 bp indel was named as Indel1 and 18 bp indel was named as Indel2) were found among them. Concretely, six SNPs and two indels (Indel1 and Indel2) between Rny1 and Rny3, fifteen SNPs and Indel2 between Rny1 and Ray1, and eleven SNPs together with Indel1 between Rny3 and Ray1 were found ( Figure S1a and Figure 3a). Rny2, Rny4, and Ray2 were also highly identical to each other, with ten SNPs and two DNA fragment indels (27 bp indel was named as Indel3 and 36 bp indel was named as Indel4) occurring among them. Eight SNPs and Indel3, four SNPs and Indel4, and eight SNPs together with Indel3 and Indel4 were detected between genes Rny2 and Rny4, Rny2 and Ray2, and Rny4 and Ray2, respectively (Figure S1b and Figure 3b). Multiple alignment analyses indicated that Rny1, Rny3, and Ray1 were highly identical at the nucleotide sequence level. Sixteen single nucleotide polymorphisms (SNPs) and two DNA fragment indels (63 bp indel was named as Indel1 and 18 bp indel was named as Indel2) were found among them. Concretely, six SNPs and two indels (Indel1 and Indel2) between Rny1 and Rny3, fifteen SNPs and Indel2 between Rny1 and Ray1, and eleven SNPs together with Indel1 between Rny3 and Ray1 were found ( Figure S1a and Figure 3a). Rny2, Rny4, and Ray2 were also highly identical to each other, with ten SNPs and two DNA fragment indels (27 bp indel was named as Indel3 and 36 bp indel was named as Indel4) occurring among them. Eight SNPs and Indel3, four SNPs and Indel4, and eight SNPs together with Indel3 and Indel4 were detected between genes Rny2 and Rny4, Rny2 and Ray2, and Rny4 and Ray2, respectively (Figure S1b and Figure 3b).

Characterization and Comparative Analysis of the Roegneria HMW-GSs
Analyses of the deduced primary structures of the six novel Roegneria subunits revealed that they shared the same essential structure with the HMW-GSs characterized previously, composed of the signal peptide, N-terminal domain, central repetitive domain, and C-terminal domain ( Figure 3, Table 1).
The subunits Rny2, Rny4, and Ray2 had a common signal peptide sequence composed of 21 residues. However, in subunits Rny1, Rny3 and Ray1, the signal peptide had three deleted residues and consisted of 18 residues (Figures 3 and 4a). The N-terminal domain of Rny2, Rny4, and Ray2 contained 105 residues, while that of Rny1, Rny3, and Ray1 encompassed 102 residues. The shorter signal peptide and the N-terminal domain of Rny1, Rny3, and Ray1 were attributed to a hexapeptide deletion (boxed in Figure 4a) crossing the signal peptide and N-terminal domain. Compared to the wheat HMW-GSs, the N-terminal domains of the six Roegneria subunits contained an extra glutamine residue (marked by an inverted solid triangle in Figure 4a), which was absent in the y-type subunits of common wheat, but present in all of the known x-type subunits. The N-terminal regions of all the six subunits possessed five conservative cysteine residues, located exactly parallel to those of the typical y-type subunits in wheat. The N-terminal domains of Rny1, Rny3, and Ray1 ended with a hexapeptide, SSQPVQ, which was found only in a few HMW-GSs derived from the St genome, while Rny2, Rny4, and Ray2 ended with the y-type subunit specific hexapeptide SSQTVQ ( Figure 4a).
The conserved C-terminal region of the subunits Rny1, Rny3, and Ray1 was normally composed of 42 amino acid residues, while that of Rny2, Rny4, and Ray2 showed a single base deletion on the 34th triplet codon, which led to the frame shift of the ORF. The undecapeptide LAAQLPAMCRL associated with x-type subunits was also found in the C-terminal of all the six Roegneria subunits (Figure 4b).
The central repetitive domains of the six novel subunits from Roegneria species were similar to those of the representative y-type subunits of wheat in their structures of tandem and interspersed repeats based on hexapeptide and nonapeptide. Despite the conserved structural characteristics, the following particular aspects were observed in these subunits. (1) The central repetitive domain lengths (Table 1, 238-320 residues) were obviously shorter than those of the representative y-type HMW-GSs in common wheat. The shorter central repetitive region was a radical cause leading to smaller molecular masses and, thus, accelerated SDS-PAGE mobility ( Figure 1); (2) cysteine residue did not occur near the C-terminus of the repetitive region of all the six Roegneria subunits, which was distinct from the typical y-type subunits of common wheat, especially that encoded by the Glu1-B1 and Glu1-D1 loci, but similar to the typical x-type subunits; (3) the two irregular motifs, pentapeptide PGGQQ and decapeptide GYYPTSPHQQ, were found in the central repetitive region of Rny1, Rny3, and Ray1 (Figure 3a), and they could be inferred as the mutants of typical hexapeptide (PGQGQQ) and nonapeptide (GYYPTSPQQ); (4) one or two stop codon(s) appeared in the central repetitive region of Rny2 (two), Rny4 (one), and Ray2 (two) (Figure 3b). Hence, these three were pseudogenes without protein products.
Amino acid sequence comparison of the repetitive regions of the six Roegneria subunits found two amino acid fragment indels, the 21-residue fragment PEQGQQGYYPTSPQQPGKGQQ (corresponding to Indel1, Figure S1a) and hexapeptide fragment PGQGQQ (corresponding to Indel2, Figure S1a) among the subunits of Rny1, Rny3, and Ray1. In addition, two amino acid substitutions occurred in the repetitive regions of Rny1 and Rny3, and five between Rny1 and Ray1 and Rny3 and Ray1, respectively ( Figure 3a). Additionally, two amino acid fragment indels, nonapeptide QGYYPTSPQ (corresponding to Indel3, Figure S1b) and dodecapeptide PGQGQQPGQGQQ (corresponding to Indel4, Figure S1b) were found among Rny2, Rny4, and Ray2, whereas two, one, and three amino acid substitutions occurred in the repetitive region of Rny2 and Rny4, Rny2 and Ray2, and Rny4 and Ray2, respectively ( Figure 3b).

Heterologous Expression of the Roegneria HMW-GS Genes in Escherichia coli
To confirm the expression of the Roegneria HMW-GS genes, they were cloned and overexpressed in Escherichia coli. The nucleotide sequences encoding the signal peptides in the ORF sequences were removed for the comparable electrophoretic mobilities of the mature proteins. SDS-PAGE analysis of the proteins extracted from the transformed E. coli cells showed that the three genes Rny1, Rny3, and Ray1 successfully expressed the specific protein subunits and that they matched with their counterparts derived from R. nakaii and R. alashanica endosperm proteins, respectively ( Figure 5). In contrast, no protein was expressed in the transformants of the three constructs pET30a-Rny2, pET30a-Rny4, and pET30a-Ray2 ( Figure 5), which established them as non-expressional pseudogenes.

Heterologous Expression of the Roegneria HMW-GS Genes in Escherichia coli
To confirm the expression of the Roegneria HMW-GS genes, they were cloned and overexpressed in Escherichia coli. The nucleotide sequences encoding the signal peptides in the ORF sequences were removed for the comparable electrophoretic mobilities of the mature proteins. SDS-PAGE analysis of the proteins extracted from the transformed E. coli cells showed that the three genes Rny1, Rny3, and Ray1 successfully expressed the specific protein subunits and that they matched with their counterparts derived from R. nakaii and R. alashanica endosperm proteins, respectively ( Figure 5). In contrast, no protein was expressed in the transformants of the three constructs pET30a-Rny2, pET30a-Rny4, and pET30a-Ray2 ( Figure 5), which established them as non-expressional pseudogenes. The electrophoresis patterns showed the three isopropyl β-D-1-thiogalactopyranoside (IPTG)-induced Roegneria HMW-GSs in E. coli, Rny1, Rny3, and Ray1 had the same mobility with those of HMW-GSs extracted from Roegneria seeds. The three pseudogenes, Rny2, Rny4 and Ray2, did not express protein under the IPTG-induced condition. The HMW-GSs extracted from Roegneria seeds (lanes R. nakaii GT1, GT2, and R. alashanica) and expressed in E. coli (lanes pET30a-Rny3I, pET30a-Rny1I, and pET30a-Ray1I) are marked by arrows. The Chinese Spring HMW-GSs (1Dx2, 1Dy12, 1Bx7, 1By8) extracted from seeds were used as controls.

Phylogenetic Relationships of the New Roegneria HMW-GSs with Orthologous Subunits from Other Species
To investigate the evolutionary relationships among the six Roegneria HMW-GSs characterized in this study and some of the representative HMW-GS in wheat and related species with different genomes (E, J, R, C, U, Ta, and K), two maximum likelihood (ML) phylogenetic trees were constructed. ML was based on the amino acid sequences of the N-terminal with the first three motifs of the central repetitive domain and the C-terminal plus the last six motifs of the central repetitive domain of these subunits, respectively. The two phylogenetic trees exhibited very small differences in the topological structure and support values for the branches ( Figure 6).
In accordance with the phylogenetic tree constructed based on the C-terminal plus motifs of central repeat domain, the HMW-GSs analyzed in this study could also be classified into four groups. Although small differences in the topological structure and support values for the clades were found between the phylogenetic trees based on the C-and N-terminals, they were highly The electrophoresis patterns showed the three isopropyl β-D-1-thiogalactopyranoside (IPTG)-induced Roegneria HMW-GSs in E. coli, Rny1, Rny3, and Ray1 had the same mobility with those of HMW-GSs extracted from Roegneria seeds. The three pseudogenes, Rny2, Rny4 and Ray2, did not express protein under the IPTG-induced condition. The HMW-GSs extracted from Roegneria seeds (lanes R. nakaii GT1, GT2, and R. alashanica) and expressed in E. coli (lanes pET30a-Rny3I, pET30a-Rny1I, and pET30a-Ray1I) are marked by arrows. The Chinese Spring HMW-GSs (1Dx2, 1Dy12, 1Bx7, 1By8) extracted from seeds were used as controls.

Phylogenetic Relationships of the New Roegneria HMW-GSs with Orthologous Subunits from Other Species
To investigate the evolutionary relationships among the six Roegneria HMW-GSs characterized in this study and some of the representative HMW-GS in wheat and related species with different genomes (E, J, R, C, U, Ta, and K), two maximum likelihood (ML) phylogenetic trees were constructed. ML was based on the amino acid sequences of the N-terminal with the first three motifs of the central repetitive domain and the C-terminal plus the last six motifs of the central repetitive domain of these subunits, respectively. The two phylogenetic trees exhibited very small differences in the topological structure and support values for the branches ( Figure 6).
In accordance with the phylogenetic tree constructed based on the C-terminal plus motifs of central repeat domain, the HMW-GSs analyzed in this study could also be classified into four groups.
Although small differences in the topological structure and support values for the clades were found between the phylogenetic trees based on the C-and N-terminals, they were highly similar with each other. Rny1, Rny3, and Ray1 together with the other eight subunits derived from the St genome composed the St group 1; Rny2, Rny4, Ray2 and other five St subunits belonged to St group 2 (Figure 6b). Among the subunits from the St genome, the subunit St1 (accession No. JN680047) from Pd. libanotica and Glu-1St1 (accession No. DQ344028) from Pd. stipifolia had their N-terminals clustered together with St subunits (St group 2), but their C-terminals were more similar to those of the typical x-type subunit (Figure 6b, x group).  (St group 2), but their C-terminals were more similar to those of the typical x-type subunit (Figure 6b, x group).

Discussion
Recently, a number of HMW-GS genes from wheat-related grasses were molecularly investigated. However, until now, the HMW-GS of the Roegneria genus were rarely studied. In the present study, we isolated six novel HMW-GS genes from two tetraploid Roegneria species endemic to China, R. nakaii and R. alashanica. Among them, the genes Rny1, Rny3, and Ray1 possessed intact ORFs, and Rny2, Rny4, and Ray2 were pseudogenes containing in-frame stop codon(s). The analyses of the coding regions and the corresponding amino acid sequences indicated that these Roegneria HMW-GSs shared the same essential molecular structures with those of the common wheat and other related species. Each was composed of four domains, a signal peptide (which would be

Discussion
Recently, a number of HMW-GS genes from wheat-related grasses were molecularly investigated. However, until now, the HMW-GS of the Roegneria genus were rarely studied. In the present study, we isolated six novel HMW-GS genes from two tetraploid Roegneria species endemic to China, R. nakaii and R. alashanica. Among them, the genes Rny1, Rny3, and Ray1 possessed intact ORFs, and Rny2, Rny4, and Ray2 were pseudogenes containing in-frame stop codon(s). The analyses of the coding regions and the corresponding amino acid sequences indicated that these Roegneria HMW-GSs shared the same essential molecular structures with those of the common wheat and other related species. Each was composed of four domains, a signal peptide (which would be removed from the matured subunit), conserved N-and C-terminals, and a longer central repetitive region. These novel Roegneria subunit genes expanded the HMW-GS gene family and laid the foundation for investigating their perspective on wheat improvement and evolution of the Glu-1 locus in the tribe Triticeae.
The novel Roegneria HMW-GS genes also showed specific properties in the sequence structures. The DNA sequence analyses indicated that the six genes (1217-1452 bp, Table 1) were notably shorter than the common HMW-GS genes in wheat, which ranged from 1800 to 2500 bp [31]. For the three genes with intact ORFs (Rny1, Rny3, and Ray1), this feature was consistent with their protein productions with fast electrophoretic mobility (Figure 1). The deduced amino acid sequences from the six Roegneria HMW-GS genes revealed that the smaller molecular size was mainly due to the fewer repeat motifs in the central repetitive domain (Figure 4). The smaller HMW-GS subunits have also been found in some other wheat related genera, such as Pseudoroegneria [14,22,32], Elytrigia (=Agropyron) [15,19], and Elymus [16], thereby suggesting that smaller HMW-GS is common in wheat-related species, but little is known about its evolutionary significance and perspective on wheat improvement and, thus, merit further studies.
The analyses of the deduced amino acid sequences from the six Roegneria HMW-GS genes showed that these subunits possessed some essential structural features of the typical y-type HMW-GSs, such as N-terminal composed of 105 (subunits Rny2, Rny4, and Ray2) or 102 (subunits Rny1, Rny3, and Ray1) amino acid residues, five conserved cysteine residues in N-terminal sequences (Figure 4), and no tripeptide GQQ (x-type subunit-specific) in the central repetitive domains (Figure 3). Nevertheless, these Roegneria HMW-GSs also showed some structural characteristics of the x-type subunits, such as the N-terminal domains of all the six HMW-GSs possessing an additional glutamine residue (a solid triangle indicated in Figure 4a), which was typically presented in the x-type subunits of wheat and some y-type subunits derived from the wild species with E, F, K, O, P, Q, Ta, V, and W genomes [13,[17][18][19], and these y-type subunits were deemed to evolutionary precede those without this glutamine residue [18,32]. The C-terminal of all of the six Roegneria HMW-GSs demonstrated a distinctive undecapeptide, LAAQLPAMCRL (Figure 4b), which is associated with x-type subunits. In addition, in the phylogenetic trees constructed based on the N-and C-terminal amino acid sequences, these Roegneria HMW-GSs were closely associated with the subunits derived from species with the St genome (St, StY, or StH). All of these subunits formed an independent clade, which had a long genetic distance with the typical x-and y-type subunits derived from A, B, and D genomes of wheat and E, J, R, C, U, Ta, and K genomes of the related species involved in this study ( Figure 6).
As per the above characteristics, these six Roegneria HMW-GS genes are not a typical y-type gene, and should rather be classed into an intermediate type inclining towards the y-type. Some atypical or intermediate type HMW-GSs have also been found in other Triticeae species, such as Agropyron cristatum [33], Elymus glaucus [16], and several Pseudoroegneria species [14,22], suggesting the atypical HMW-GS is universally existent in the wild species of the tribe Triticeae, and plausibly represents an evolution status of the Glu-1 locus.
Shewry et al. proposed that the x-and y-type HMW-GS genes derived from a common ancestral gene similar to the y-type genes [24]. However, no x-type subunit genes have been identified from the Glu-St1 locus until now [14,22,32]. It was speculated that the St genome differentiated from the modern-day Triticeae genomes earlier than the differentiation between x-and y-type genes [22]. The phylogenetic analysis in this study was in agreement with the hypothesis. It further showed that the HMW-GSs from the species with St, StY, or StH genome(s) clearly clustered as an independent clade that was genetically distal from the typical x-and y-type clusters composed of HMW-GSs from wheat and other related species with seven different genomes ( Figure 6). Thus, based on the evidence, we infer that the Glu-1 locus in R. nakaii and R. alashanica is a very primitive glutenin locus across evolution.
R. nakaii and R. alashanica have the same genome compositions, StStYY [26]. Phylogenetic analysis showed that the six Roegneria HMW-GS genes were split into two groups clustered to different clades, respectively ( Figure 6, St group 1 and St group 2), suggesting that maybe they were derived from two separate genomes. However, it was apparent from the phylogenetic trees that each one of the two clades (St group 1 or St group 2) included the HMW-GSs from species with only St (diploid and tetraploid species), StY, and StH genomes. Hence, we speculate that two group HMW-GS genes in Roegneria are all derived from the St genome rather than Y. Nevertheless, the two genomes of the two groups of Roegneria HMW-GS genes have undergone slight differentiation.

Plant Materials
Two tetraploid Roegneria species, R. nakaii Kitag (No. Z1874) and R. alashanica Keng (No. Z1081), were used in this study, and their mature seeds were kindly provided by Prof. Lihui Li (Institute of Crop Sciences, Chinese Academy of Agricultural Sciences, Beijing, China). The common wheat cultivar Chinese Spring was used as a control in SDS-PAGE and Western blotting experiments, and its mature seeds were from our laboratory collections.

SDS-PAGE
The HMW-GSs were extracted from mature seeds of R. nakaii and R. alashanica and separated by SDS-PAGE as described previously [34]. The HMW-GSs from the common wheat cultivar Chinese Spring (Null, 1Bx7 + 1By8, 1Dx2 + 1Dy12) were used as standards. At least ten individual seeds were examined to ascertain the composition of HMW-GS in the given accessions.

Western Blotting
The proteins on the polyacrylamide gel were transferred to an Amersham Hybond-P PVDF Membrane (GE Healthcare, Uppsala, Sweden) by electroblotting at 90 V for 3 h. The membrane was washed for 10 min with TBST buffer (0.01 M Tris-HCl, 0.15 M NaCl pH 7.4, and 0.05% Tween 20) and blocked in TBSTM (5% skim milk in TBST) for 2 h. After three subsequent washes, the membrane was incubated in TBSTM with a 1:7000 dilution of the anti-R2 polyclonal antibody [35] for 2 h, followed by horseradish peroxidase (HRP)-labelled goat anti-rabbit IgG (Beyotime, Shanghai, China) for 2 h. Consecutively, the membrane was washed three times and then incubated in precipitating TMB (3,3 1 ,5,5 1 -tetramethylbenzidine) Substrate Solution (Beyotime) for 10-30 min. The reaction was stopped by washing the membrane with distilled water.

Cloning and Sequencing of the Roegneria HMW-GS Genes
Genomic DNA was extracted from fresh leaves of seedlings using the cetyltrimethylammonium bromide (CTAB) method described by Webb and Knapp [36]. A pair of degenerate primers, P1: 5 1 -ATGGCTAAGCGGC/TTA/GGTCCTCTTTG-3 1 and P2: 5 1 -CTATCACTGGCTG/AGCCGACA ATGCG-3 1 [9], were used to amplify the ORFs of HMW-GS genes from R. nakaii and R. alashanica using the high-fidelity PrimeSTAR HS DNA Polymerase with GC buffer II (TaKaRa, Dalian, China) to avoid PCR errors. The PCR was as follows: denaturation for 3 min at 98˝C; followed by 30 cycles of 20 s at 98˝C, 30 s at 69˝C, and 3 min at 72˝C; extended finally for 10 min at 72˝C. The amplicons were separated on 2.0% agarose gel, and the anticipated band was excised and purified using a TIANgel Midi Purification Kit (TIANGEN, Beijing, China). The purified DNA fragments were cloned into pZeroBack/Blunt vector (TIANGEN). The recombinant plasmids were transformed into Escherichia coli DH5α competent cells (TIANGEN). Nucleotide sequencing was performed commercially (Beijing Genomics Institute, Beijing, China) and the final sequence for each HMW-GS gene was determined from the sequencing results of several independent clones.

Sequence Comparison and Phylogenetic Analyses of Cloned HMW-GS Genes
The cloned DNA sequences were assembled, and the ORFs of the HMW-GS genes were translated into amino acid sequences using the DNAMAN software (Version 6.0.3, Lynnon Biosoft, Vaudreuil, QC, Canada). Multiple alignments of the DNA or protein sequences were carried out separately with MEGA6 [37]. In order to investigate the phylogenetic relationships of the new Roegneria HMW-GS with previously-characterized Glu-1 alleles from wheat and related species, the amino acid sequences of the N-terminal with the first three motifs of the central repeat domain and the amino acid sequences of the C-terminal with the last six motifs of the central repeat domain were used to create the ML phylogenetic trees using MEGA6, respectively [37]. A total of 42 HMW-GSs were involved in this phylogenetic analysis, including the six novel Roegneria subunits and the other 36 representative subunits from wheat and related species with different genomes. The D-hordein gene (AY268139) from Hordeum vulgare was used as the outgroup. The clade support values were estimated using 500 bootstrap pseudoreplicates. The DNA or protein sequences used for multiple alignments and phylogenetic analysis in this research were derived from the NCBI database [30].

Conclusions
In the present work we cloned six HMW-GS genes from two tetraploid Roegneria species, R. nakaii and R. alashanica. These novel Roegneria HMW-GS genes extended and enriched HMW-GS gene families. All of the six Roegneria HMW-GSs are not typical x-or y-type subunits and should be classed into an intermediate type inclining towards the y-type, which plausibly represents an evolution status of the Glu-1 locus. The Roegneria HMW-GSs, together with other HMW-GSs from species with St, StY, or StH genome(s) formed an independent clade in phylogenetic trees, varying from the typical x-and y-type clusters. We speculate that the Glu-1 locus in R. nakaii and R. alashanica is a very primitive glutenin locus across evolution. In addition, the six Roegneria HMW-GS genes were phylogenetically split into two groups, each group included the HMW-GSs from species with St, StY, and StH genomes, demonstrating that the six Roegneria HMW-GS genes were from the St genome rather than Y. Nevertheless, the two derived genomes of the two Roegneria HMW-GS gene groups have undergone slight differentiation.