Potential Use of Wild Einkorn Wheat for Wheat Grain Quality Improvement: Evaluation and Characterization of Glu-1 , Wx and Ha Loci

: Wild einkorn ( Triticum monococcum L. ssp. aegilopoides (Link) Thell.) is a diploid wheat species from the Near East that has been classiﬁed as an ancestor of the ﬁrst cultivated wheat (einkorn; T. monococcum L. ssp. monococcum ). Its genome (A m ), although it is not the donor of the A genome in polyploid wheat, shows high similarity to the A u genome. An important characteristic for wheat improvement is grain quality, which is associated with three components of the wheat grain: endosperm storage proteins (gluten properties), starch synthases (starch characteristics) and puroindolines (grain hardness). In the current study, these grain quality traits were studied in one collection of wild einkorn with the objective of evaluating its variability with respect to these three traits. The combined use of protein and DNA analyses allows detecting numerous variants for each one of the following genes: six for Ax , seven for Ay , eight for Wx , four for Gsp-1 , two for Pina and three for Pinb . The high variability presence in this species suggests its potential as a source of novel alleles that could be used in modern wheat breeding.


Introduction
In recent decades, the rise in environmental awareness has led to a new paradigm in agriculture, where the concept of sustainability is basic to all productive processes. This sustainability is intimately linked to the conservation and utilization of plant genetic resources; in fact, agriculture cannot be considered to be sustainable if it does not include a suitable program for conservation and evaluation of crop genetic resources [1]. At the same time, global warming has become one of the main challenges facing crop improvement programs [2,3]. The search for genetic materials adapted to the prevailing stressful environmental conditions without reducing the technological quality is key for the development of new wheat cultivars.
Wheat quality is a complex characteristic that depends on consumer preferences, the product and its processing; however, from a technical point of view, three grain components play an important role, with the storage proteins, the starch synthases and the puroindolines being key. Each one of these components has effects on different aspects of wheat quality. The storage proteins are mainly responsible for the dough visco-elastic properties (strength and extensibility of the gluten) [4]; the starch synthases affect the composition and properties of the starch [5]; and the puroindolines are related to the grain texture and indirectly to the capacity of the flour to absorb water due to the damaged starch granules generated during the process of milling [6]. Variation in these grain components, therefore, modulates the properties of the flour.
The degree of variation differs markedly between each of these grain components. While the storage proteins exhibit considerable genetic polymorphism, other components show more moderate levels of variation [4][5][6][7][8][9][10]. Of the storage proteins, the high-molecular weight glutenin subunits (HMWGs) have been the most studied due to their marked influence on the properties of the gluten [11] and the ease with which the proteins encoded by individual alleles can be distinguished. These proteins are synthesized by genes located on the long arm of the chromosome-1 group (Glu-1 locus) [12]. The numerous studies carried out on these genes have shown that the least variation is present at the Glu-A1 locus in cultivated hexaploid wheat [4]. Similar results have been observed for the Wx-A1 gene, one of the three genes that code for the three waxy proteins present in bread wheat (Wx-A1, Wx-B1 and Wx-D1). The waxy proteins are responsible for amylose synthesis in the flour starch, and their variability (between homeologs and between different alleles of the same gene) has a considerable effect on starch properties [5]. With respect to the puroindolines (puroindoline-a and puroindoline-b), these genes (Pina and Pinb) are included in the Ha (hardness) locus, formed by both genes together with Gsp-1 (grain softness protein) and seven other genes without a known function [13]. In hexaploid wheat, these genes are exclusively derived from the D genome because the puroindolines genes from the A and B genomes had been deleted during the evolution event that generated tetraploid wheat [13], and therefore there is no allelic variation for these genes on the group A chromosomes.
The search for alternative gene sources is one of the strategies used to develop cultivars more adapted to perform well under the conditions of global warming. In this context, ancient wheats and wild wheat relatives, which are adapted to be grown in marginal zones under extreme conditions [14], are considered to host interesting genetic variability that could be exploited in breeding programs.
Among the potential variation sources associated with greater adaptation to adverse conditions, the wild wheat relatives carrying the A genome of polyploid wheat could be good candidates [15]. Wild diploid wheat is represented by two main species: Triticum monococcum L. ssp. aegilopoides (Link) Thell. (syn. T. boeoticum Boiss.) and T. urartu Thum. ex Gandil. Both species contain the A genome, which is closely related to the A genome of durum (T. turgidum L. ssp. durum (Desf.) Husn.) and common wheat (T. aestivum L. ssp. aestivum), although reproductive isolation between them has been indicated [16,17]. Later studies at the molecular level have suggested that, whereas T. monococcum spp. aegilopoides was the species from which cultivated einkorn (T. monococcum L. spp. monococcum L.) was domesticated, the A genome of polyploid wheats is equivalent to that of T. urartu [18].
The revival of the interest in the ancient wheats has increased the number of surveys carried out on cultivated einkorn, mainly with respect to the nutritional and health aspects of this ancestral crop [19][20][21][22][23][24][25]. As the decline in this crop began in antiquity [26][27][28], the variation retained from domestication until the present day could be relatively scarce. This relationship would be in agreement with the low variation detected in some genes related to technological aspects, although the wild ancestor species could contain greater variation, which could be transferred to modern wheat (see [15] for a review). In these cases, although the linkage drag of deleterious alleles linked to desirable alleles in the exotic parent, as a result of the wild nature of this source material, might be a handicap [29], the notable variation in both technological quality aspects and adaptive traits could compensate for any negative impact of the linkage drag. In this respect, the development of DNA markers that facilitate the selection of alleles of interest in a breeding program by marker-assisted selection will be key [30,31].
In the present study, we studied the variation in the loci Glu-A m 1, Wx-A m 1 and Ha in a collection of 170 accessions of wild einkorn (T. monococcum ssp. aegilopoides) from Iran, Iraq and Turkey, with the aim of establishing the potential value of this wild species as a gene donor source for wheat quality improvement. Parallelly, the main allelic variants of each gene were characterized by diagnostic markers for evaluating this species' utility in wheat breeding.

Plant Material
One hundred and seventy accessions of wild einkorn wheat from Turkey, Iran and Iraq were analyzed in the current study (Table S1). These materials were kindly supplied by the National Small Grain Collection (Aberdeen, ID, USA).

Glutenin Analysis
Proteins were extracted from crushed endosperm according with the procedure described by Alvarez et al. [32]. Precipitate glutenin subunits were solubilized in buffer (ratio 1:5 mg/µL to wholemeal) and fractionated by electrophoresis in vertical SDS-PAGE slabs in a discontinuous Tris-HCl-SDS buffer system (pH: 6.8/8.8) at a polyacrylamide concentration of 8% (w/v, C: 1.28%), using the Tris-HCl/glycine buffer system. Electrophoresis was performed at a constant current of 30 mA/gel at 18 • C for 45 min after the tracking dye migrated off the gel. The gels were stained overnight with 12% (w/v) trichloroacetic acid solution containing 5% (v/v) ethanol and 0.05% (w/v) Coomassie Brilliant Blue R-250. De-staining was carried out with tap water.

Waxy Protein Analysis
For waxy proteins, whole grain flour was mixed with 1 mL of distilled water and incubated at 4 • C for 24 h. The homogenate was filtered through Miracloth and centrifuged at 14,000 g for 1.5 min. The pellet was washed with 1 mL of buffer A (55 mM Tris-HCl pH 6.8, 2.3% (w/v) sodium dodecyl sulphate, 2% (w/v) dithiotreitol, 10% (v/v) glycerol) according to Echt and Schwartz [33]. Then, 1 mL of buffer A was added to the pellet and left for 30 min at room temperature. The pellet was washed three times with distilled water and once with acetone and then air dried. The residue was mixed with 80 µL of buffer A containing 0.02% (w/v) bromophenol blue, heated in a boiling bath for 2 min, cooled in ice and centrifuged.
Aliquots of 15 µL supernatant were loaded in vertical SDS-PAGE slabs in a discontinuous Tris-HCl-SDS buffer system (pH: 6.8/8.8) at a polyacrylamide concentration of 12% (w/v, C: 0.44%). The Tris-HCl/glycine buffer system was used. Electrophoresis was performed at a constant current of 30 mA/gel and 18 • C, continuing for 4 h after the tracking dye migrated off the gel. Protein bands were visualized by silver staining.

PCR Amplification of Genes from the Glu-A1, Wx and Ha Loci
Genomic DNA was extracted from approximately 100 mg of young leaves of single plants using the CTAB (cetyl-trimethyl-ammonium bromide) method [34].
For the amplification of the genomic sequence of each gene, different strategies were used. For Ax and Ay genes, the complete coding regions of 2475 bp and 1800 bp, respectively, were amplified using the primers designed by D'Ovidio et al. [35]. The genomic sequence of the Wx gene contains twelve exons and eleven introns, with a coding region around 2800 bp. This genomic sequence was amplified in three fragments using the primers designed by Guzmán and Alvarez [36] and Ayala et al. [37]: the first fragment includes the 1st to 3rd exons (Wx1Fw/Wx1.3Rv); the second extends from the 3rd to the 6th exon (Wx2Fw/Rv); and the last fragment covers the region spanning the 6th to the 11th exon (Wx3Fw/Rv). These fragments overlapped between them because the Wx1.3Rv primer is the complementary sequence of the Wx2Fw primer, whereas the Wx3Fw primer is located inside the second fragment. For the Pina and Gsp-1 genes, the primers designed by Massa et al. [38] were used, which generated amplicons of 516 bp and 564 bp, respectively. For the Pinb gene, the best results (595 bp) were obtained with the primers designed by Lillemo et al. [39].
All amplifications were performed in a 20 µL final reaction volume containing 50 ng of genomic DNA, 1.25 mM MgCl 2 , 0.2 µM of each primer, 0.2 mM dNTPs, 4 µL 10× PCR buffer and 0.75 U GoTaq ® G2 Flexi DNA Polymerase (Promega, Madison, WI, USA). PCR conditions as well as primer sequences are available in Table S2. Amplification products were separated by vertical PAGE in 8% (w/v, C: 1.28%) polyacrylamide concentration gels in a discontinuous Tris-HCl buffer system (pH 6.8/8.8). The bands were stained with GelRed™ nucleic acid stain (Biotium, Fremont, Canada) and then visualized under UV light.

DNA Diagnotic Marker Analysis
The amplicons were digested with the specific endonucleases for each gene to detect internal differences between the different alleles. These endonucleases were selected according to previous studies carried out with these genes in wheat and other grasses [40][41][42][43][44]: HaeIII and MboII for the Ax gene; DdeI and PstI for the Ay gene; DdeI (fragments 1 and 2) and NcoI (fragment 3) for the Wx gene; DdeI for the Gsp-1 gene; RseI for the Pina gene; and BsrBI for the Pinb gene.
Digested fragments were analyzed by polyacrylamide gel electrophoresis in a discontinuous Tris-HCl buffer system (pH: 6.8/8.8) with a 10% polyacrylamide concentration (C: 3.0%). The Tris/glycine buffer was used. The bands were stained with GelRed™ nucleic acid stain (Biotium, Fremont, Canada) and then visualized under UV light.

Results
Due to the nature of this study, as an initial exploration of the variability in genes related to the technological quality in wheat, this germplasm collection was analyzed, using a scaled strategy. Initially, all accessions were analyzed by SDS-PAGE for variation in the HMWGs (Table S1); subsequently, 14 representative accessions, carrying the available allelic variation for HMWGs, were screened to identify variation in the other two loci.

Variation and Characterization of the Ax and Ay Alleles for HMWGs
SDS-PAGE analysis of the HMWGs in grains from the germplasm collection showed that these materials contained both subunit types (x and y) of the proteins encoded by the Glu-A m 1 locus ( Figure 1). The A m x subunits showed greater staining intensity than the A m y subunits, which, in some cases, appeared as one major band together with several minor, more lightly stained bands, probably due to post-translational modifications, an effect which has been reported by other authors as being specific to einkorn [45,46]. In general, the A m x subunits exhibited a mobility intermediate between that of the Ax1 and Ax2* subunits, although some accessions showed variants with mobility faster than either the Ax2* or the Dx5 subunit. The A m y subunits have an electrophoretic mobility that is faster than the Bx subunits but lower than the by subunits ( Figure 1). As the synthesized genes of these proteins are intronless, the variation showed by DNA amplification was not different to that observed in protein SDS-PAGE separation. However, the analysis of the internal structure is easier using the nucleotide sequences rather than the extracted protein.
Consequently, fourteen representative accessions containing the detected A m x and A m y variants were evaluated by PCR amplification, followed by their digestion with specific endonucleases, which would allow the internal variation to be detected along with the relationships between these allelic variants, as previous studies have suggested [40][41][42][43]. In these accessions, both A m x and A m y amplicons were digested, showing a notable variation within the amplified sequence (Figures 2 and 3). tected along with the relationships between these allelic variants, as previous studies have suggested [40][41][42][43]. In these accessions, both A m x and A m y amplicons were digested, showing a notable variation within the amplified sequence (Figures 2 and 3).
The Ax sequences were independently digested with HaeIII and MboII endonucleases, which have been previously used to determine variation in Ax genes from durum and common wheat [40,41]. In this case, the variation was similar between both endonucleases, being clearer due to the size of digested fragments for HaeIII endonuclease ( Figure  2). This enzyme generates eleven fragments in Ax2* subunits and nine in Ax1 subunits, although the six larger fragments (Figure 2, lane 1) are useful to differentiate between both subunits. The main differences among both alleles are the fusion in the unique fragment of the two larger fragments at the Ax1 allele, and the enlargement of the fourth fragment size (335 vs. 317 bp). All of these three fragments are located inside the central repetitive domain. The cleavage patterns of the Ax2* amplicon from cv. Cheyenne (Figure 2, lane 1) and the A m x amplicons detected in wild einkorn were clearly different. The A m x patterns were more similar to the Ax1 or Axnull amplicons, as shown by Alvarez et al. [41], although with some notable differences due to the absence of cut-off points. Thus, the fifth fragment of Ax2* that is common to Ax1 and Axnull [41] is absent in the wild einkorn alleles, with the exception of the PI 554504 accession (Figure 2, lane 5). Furthermore, these variants showed the 1279 bp fragment from the Ax1 subunit and the band equivalent to 358 bp present in all A m x types. However, a mutation was detected in these variants, which eliminated the cut-off point between 335 and 180 bp at the beginning of the repetitive domain (marked with an arrow in Figure 2). The use of this cleavage analysis allowed detecting up to five variants similar to Ax1 subunits (Table 1), with some differences within the ) was used as control. synthesized genes of these proteins are intronless, the variation showed by DNA amplification was not different to that observed in protein SDS-PAGE separation. However, the analysis of the internal structure is easier using the nucleotide sequences rather than the extracted protein. Consequently, fourteen representative accessions containing the detected A m x and A m y variants were evaluated by PCR amplification, followed by their digestion with specific endonucleases, which would allow the internal variation to be detected along with the relationships between these allelic variants, as previous studies have suggested [40][41][42][43]. In these accessions, both A m x and A m y amplicons were digested, showing a notable variation within the amplified sequence (Figures 2 and 3).
The Ax sequences were independently digested with HaeIII and MboII endonucleases, which have been previously used to determine variation in Ax genes from durum and common wheat [40,41]. In this case, the variation was similar between both endonucleases, being clearer due to the size of digested fragments for HaeIII endonuclease ( Figure  2). This enzyme generates eleven fragments in Ax2* subunits and nine in Ax1 subunits, although the six larger fragments (Figure 2, lane 1) are useful to differentiate between both subunits.
The main differences among both alleles are the fusion in the unique fragment of the two larger fragments at the Ax1 allele, and the enlargement of the fourth fragment size (335 vs. 317 bp). All of these three fragments are located inside the central repetitive domain. The cleavage patterns of the Ax2* amplicon from cv. Cheyenne (Figure 2, lane 1) and the A m x amplicons detected in wild einkorn were clearly different. The A m x patterns were more similar to the Ax1 or Axnull amplicons, as shown by Alvarez et al. [41], although with some notable differences due to the absence of cut-off points. Thus, the fifth fragment of Ax2* that is common to Ax1 and Axnull [41] is absent in the wild einkorn alleles, with the exception of the PI 554504 accession (Figure 2, lane 5). Furthermore, these variants showed the 1279 bp fragment from the Ax1 subunit and the band equivalent to 358 bp present in all A m x types. However, a mutation was detected in these variants, which eliminated the cut-off point between 335 and 180 bp at the beginning of the repetitive domain ) was used as control.
The Ax sequences were independently digested with HaeIII and MboII endonucleases, which have been previously used to determine variation in Ax genes from durum and common wheat [40,41]. In this case, the variation was similar between both endonucleases, being clearer due to the size of digested fragments for HaeIII endonuclease (Figure 2). This enzyme generates eleven fragments in Ax2* subunits and nine in Ax1 subunits, although the six larger fragments (Figure 2, lane 1) are useful to differentiate between both subunits. The main differences among both alleles are the fusion in the unique fragment of the two larger fragments at the Ax1 allele, and the enlargement of the fourth fragment size (335 vs. 317 bp). All of these three fragments are located inside the central repetitive domain.
The cleavage patterns of the Ax2* amplicon from cv. Cheyenne (Figure 2, lane 1) and the A m x amplicons detected in wild einkorn were clearly different. The A m x patterns were more similar to the Ax1 or Axnull amplicons, as shown by Alvarez et al. [41], although with some notable differences due to the absence of cut-off points. Thus, the fifth fragment of Ax2* that is common to Ax1 and Axnull [41] is absent in the wild einkorn alleles, with the exception of the PI 554504 accession (Figure 2, lane 5). Furthermore, these variants showed the 1279 bp fragment from the Ax1 subunit and the band equivalent to 358 bp present in all A m x types. However, a mutation was detected in these variants, which eliminated the cut-off point between 335 and 180 bp at the beginning of the repetitive domain (marked with an arrow in Figure 2). The use of this cleavage analysis allowed detecting up to five variants similar to Ax1 subunits (Table 1), with some differences within the repetitive domain, together with one variant that could be more related to the Ax2* subunit (Figure 2, lane 5). The different variants detected in each accession were named with Roman numbers. Further information on the internal structure of the A m y genes was obtained by the digestion of amplified fragments obtained with the AyFw/Rv primers with two restriction enzymes, DdeI and PstI (Figure 3a,b, respectively).
The digestion of the ghost-Ay allele linked to the Ax2* subunit in cv. Cheyenne (Figure 3a, lane 1) with DdeI endonuclease generates up to six fragments of different sizes (in order: 5-72-164-160-564-839 bp). However, the A m y sequences from wild einkorn present one major pattern formed by three fragments (Figure 3a, lanes 2-7), due to two point mutations, one between the 72 and 164 bp fragments and another between the 164 and 160 bp fragments [47]. These variants presented in wild einkorn did not show variation in the size of the 72-164-160 and 839 bp fragments, with the exception of the variant detected in PI 554504 (Figure 3a, lane 5). The larger variation was detected in the central fragment (564 bp) located within the repetitive domain. The small differences in this central fragment among variants were confirmed by the use of PstI endonuclease (Figure 3b). On the basis of this, the combined use of both endonucleases permitted detecting up to seven different restriction patterns associated with the same number of alleles (Table 1).

Variation and Characterization of Wx Variants
Analysis of the waxy proteins in the 12 accessions showed that these materials did not exhibit any variation in electrophoretic mobility with respect to this protein. In all Agronomy 2021, 11, 816 7 of 14 cases, the accessions showed a waxy protein with an electrophoretic mobility faster than the Wx-A1 protein present in tetra-and hexaploid wheats (Figure 4). Nevertheless, although proteins' mobility was the same, the waxy genes could carry some variation at the molecular level. Due to the structure of these genes (12 exons + 11 introns), they tend to present conservative regions (mainly exons), together with variable regions (introns). As the combined use of the PCR amplification and restriction pattern of these PCR products has shown to be a useful tool to evaluate this possible variation [36,37,44], this strategy was used here ( Figure 5).
The first fragment was digested with DdeI endonuclease that only cuts the Wx-A1 genes, but Wx-B1 or Wx-D1 (Figure 5a, lanes 1 and 11). In this case, any of the variants detected in wild einkorn were digested with this restriction enzyme (Figure 5a, lanes 2-10), being consequently different to Wx-A1 variants from durum or common wheat. This same endonuclease (DdeI) was used to digest the second fragment (Figure 5b). In this case, both Wx-B1 and Wx-D1 products were not digested (Figure 5a, lanes 1 and 11), and Wx-A1 showed a cut-off point that generates two fragments: a very small one with 136 bp (not shown in the gel) and another with 1037 bp (Figure 5b, lanes 1 and 11). On the contrary, the Wx-A m 1 variants presented one restriction pattern very different to the reference pattern, forming two fragments with 636 bp and 550 bp (Figure 5b, lanes 2-10). Up to six different patterns were also detected: P-1, lanes 2, 6 and 8; P-2, lane 3; P-3, lanes 4-5; P-4, lane 7; P-5, lane 9; and P-6, lane 10.
The combination of three fragments with their respective digestions suggested the presence of up to eight different alleles (Table 1), although the active protein of each of them showed, as mentioned above, a similar size (Figure 4).

Variation and Characterization of Gsp-1, Pina and Pinb Variants
As a first approach, the variability of the three main genes (Gsp-1, Pina and Pinb) from the Ha locus was analyzed through amplification using the gene-specific primers designed by Massa et al. [38] and Lillemo et al. [39] (Figure 6a,b, respectively). The Massa et al. [38] primers allow simultaneous amplification of the Pina and Gsp-1 genes (Figure 6a), and both genes showed variation in the accessions evaluated here, presenting, in both cases, PCR products with a similar size to those observed in common wheat (Figure 6a). All accessions evaluated contained the Pina gene, but not the Gsp-1 gene (Figure 6a, lane 4). According to the mobility, four variants were detected for Gsp-1 (I: lanes 1 and 3; II: lanes 2, 5 and 7; III: lane 4; and IV: lane 6) and two for Pina (I: lanes 1, 2, 4, 5 and 7; II: lanes 3 and 6). For the Pinb gene (Figure 6b), the accessions also exhibited up to three different variants (I: lanes 1 and 5; II: lanes 2, 4 and 6; and III: lanes 3 and 7). The sum of the partial variations (Gsp-1 + Pina + Pinb) indicated that these accessions showed marked variation for the Ha locus (Table 1).
A similar strategy to that used in the Wx gene analysis was applied here. The amplicons were digested with three endonucleases (DdeI for Gsp-1; RseI for Pina; and BsrBI for Pinb), which were successfully used to identify variation in these genes in previous studies [43,48].
No additional variation for Gsp-1/Pina amplicons was detected with RseI and BsrBI endonucleases; however, the digestion with DdeI was more successful (Figure 7). This endonuclease specifically cuts Gsp-1 but not Pina in common wheat (

Discussion
Wild relatives of cultivated wheat have been used as sources of genes in wheat breeding programs, mainly in the improvement of biotic stress resistance [49][50][51]. However, the role of the primary wheat gene pool, such as the accessions studied here, in grain quality improvement has been more limited [15]. In addition to the fact that quality is a more complex character, involving many genes, than, say, disease resistance, the main difficulty is to evaluate grain quality in species that exhibit low grain yield and a very small grain. Therefore, the analysis and characterization of the genes of these wild accessions and their comparison with those of cultivated wheats may be a useful strategy, which could later lead to the introgression of these exotic genes into elite germplasms to analyze their effects on grain quality.
The differences in variation among these genes could be related to both their physiological functions in plants and to the genomic structure of each plant. The glutenin and puroindolines play roles as food reserves, structural materials or defense against pathogens in wheat grain, whereas the waxy proteins are enzymes involved in the synthesis of starch. On the other hand, intronless genes, such as those encoding glutenins or puroindolines, have a tendency to be more variable because the changes are easily fixed and translated to the mature protein. However, Wx genes are fragmented genes (with introns and exons), where many of the mutations occur in the introns and their effects on the properties of the protein end product are hence more limited. These differences may be useful in terms of diversity or phylogenetic studies; in fact, the Wx gene has been considered to be a valuable tool for this type of analysis, due to its fragmented structure and its ubiquity [52].
The variation detected for HMWGs in the wild einkorn accessions evaluated in this study was high, although some variants were present at very low frequencies and showed a clear risk of loss by random genetic drift. This variation was similar to that found in a collection of T. urartu accessions, which had previously been evaluated [53,54], and was clearly greater than the variation shown in some of the studies carried out with cultivated wheat species. For example, Alvarez et al. [55], using Spanish cultivated einkorn materials, detected only three Glu-A m 1 variants. This low variability in the cultivated species relative to the wild ones could be related to how these materials have been used. The cultivated einkorn, although abandoned in ancient times, was cultivated for a specific purpose, probably bread making [28]. This implied a selection pressure that fixed those alleles best adapted to the use in question, while the rest of them were progressively lost, due to the fact that these genotypes were discarded, resulting in low variability [27,28]. This artificial selection process obviously did not take place in wild species (wild einkorn or T. urartu), and the modifications of these genes and their frequencies were regulated by only stochastic events.
Gluten strength has been positively correlated with the number of HMWGs [56,57]; in this respect, both durum and common wheat lack the alleles encoding the active Glu-1 Ay subunits and many durum wheats are also Ax-silent. However, wild diploid and tetraploid wheats exhibit active Ay subunits and, consequently, they could be good sources to increase the number of alleles encoding active subunits at the Glu-A1 locus. This procedure was exploited by Rogers et al. [58] with two lines of T. boeoticum Boiss. ssp. thaoudar for introducing active Ay subunits into common wheat, in order to achieve increases in the gluten strength. This was also observed in a cross between a T. urartu accession and durum wheat cv. Yavaros [59]. Although further studies should be carried out, this opens the way to introgress other Glu-A1 alleles, such as the ones found in the current study.
For the Wx-A1 gene, our previous studies, carried out with cultivated einkorn and T. urartu, had shown that variation in this gene was higher in wild species than in cultivated ones. Ortega et al. [60] analyzed the genomic sequences of the Wx-A1 genes, detecting up to five alleles in T. urartu, one of which had previously been described [36], but only one of which was found in the cultivated einkorn accessions analyzed, confirming the findings of Guzmán et al. [61]. In general, the studies carried out on this gene in cultivated einkorn have shown that variation is scarce [62,63].
Grain texture is the main character controlled by the genes present at the Ha locus, although this locus has been mainly associated with the puroindoline genes (Pina and Pinb), whereas the role of the Gsp-1 gene remains controversial [6,64], and the true function of the seven other genes is unknown [13]. While the diversity of puroindoline genes has been evaluated [48,[65][66][67], there is less information available on variation in the other genes at the Ha locus. In the current study, the variation for both Pin genes was high in wild einkorn, being comparable with the variability detected in cultivated einkorn, but clearly lower than that detected in the other wild diploid wheat tested, T. urartu [67].
On the other hand, other notable changes have been described, such as the loss of Pin genes, located on the short arm of chromosome 5, in both the A and B genomes in tetra-and hexaploid wheats [13]. One clear opportunity to increase genetic variability is by re-incorporating these genes by the full or partial introgression of chromosome 5A. This process has been carried out with cultivated einkorn, the Ha locus of which, located on the 5A m S chromosome, was transferred to cv. Chinese Spring, as either an entire chromosome [68,69] or only the Ha locus [70], with the grains of the resulting materials being softer than 'Chinese Spring'. This could be very interesting in durum wheat, where these materials could show a soft texture, similar to cv. Soft Svevo obtained by the 5BS-5DS translocation [71]. The incorporation of puroindoline genes from other species in durum wheat could be more effective in the A genome, given the necessity for retaining the 5BL chromosome, on which the Ph1 gene (which controls homeolog chromosome pairing) is located, together with its lower tolerance of aneuploidy than common wheat.
The transfer of this valuable variability to the modern wheat gene pool could be achieved by different routes. This is conditioned by the homology among the A chromosomes of polyploid wheat and those of the ancestral A genome, due to the marked chromosomal reorganization which occurred after the polyploidization of wheat. One clear example of this was described by Devos et al. [72] in relation to chromosomes 4A, 5A and 7B. According to those authors, during the generation of tetraploid wheat, a reciprocal translocation occurred between chromosomes 4AL and 5AL, followed by a pericentric inversion of chromosome 4A. Finally, after a reciprocal translocation between chromosomes 4AL and 7BS, and a paracentric inversion of the original 5AL region within chromosome 4AL, the current chromosome 4AL, present in durum and common wheat, was generated. As a consequence of this reorganization, the Wx-B1 gene, which should be on chromosome 7BS, was actually located on chromosome 4AL.
On the other hand, some differences between the genome of einkorn (A m ) and the A genome of polyploid wheats, derived from T. urartu (A u ), have been indicated. For this reason, crosses between einkorn (wild and cultivated) and tetraploid wheats produce offspring with A m and A u chromosomes. This has generated new natural species, such as T. zhukovskyi Menabde & Ericzjan (A u A u GGA m A m ), an amphiploid derived from a spontaneous cross between einkorn and T. timopheevii (Zhuk.) Zhuk. ssp. timopheevii [73]. A similar occurrence was also reported by Gill et al. [74] in crosses between wild einkorn and durum wheats, which have shown potential for breeding [75]. In our previous work using T. urartu, several backcrosses were achieved between the hybrid and the durum wheat parent [59].
Although the data obtained in the present study should be understood as an approximation of the variability present in this wild species, our results show that the variability was high for Glu-A1 and Ha loci in wild einkorn. However, for waxy proteins, these results suggest that this species is probably not the best choice of species in which to find novel variation.

Conclusions
Wild einkorn is an interesting source of genes related to wheat quality. Although the effects of these allelic variants on the technological quality properties in cultivated wheat should be evaluated further, the information revealed in this study may be of interest to wheat breeders in order to select parental accessions to generate recombinant lines with different gluten and texture properties, using wild einkorn as the donor species.