Seeds of the selected DRE2 TILLING mutants were obtained from the University of California, Davis, and seeds of the common wheat cultivar “Chinese Spring” nullisomic-tetrasomic lines lacking chromosome 5A, 5B, or 5D were obtained from the Wheat Genetics Resource Center (Kansas State University, Manhattan). The seeds of the diploid wheat progenitors [Triticum urartu (AA), Aegilops speltoides (BB), and Aegilops tauschii (DD)] and “Chinese Spring” [Triticum aestivum (AABBDD)] were obtained from the National Small Grains Collection (NSGC), Aberdeen, ID.

Seeds of common wheat cultivar “Express” (WestBred) and durum wheat cultivar “Kronos” (Arizona Plant Breeders) were used to develop wheat TILLING populations. Wheat TILLING populations were generated by exposing Express and Kronos seeds to the mutagen ethyl methanesulfonate (EMS). Kronos seeds were treated with 0.75 or 1% EMS and Express seeds with 0.75, 1, or 1.2% EMS for 18 h [see (27) for details]. In a separate experiment, the Kronos seeds were treated with 0.7% (57 mM) or 0.75% (to 60 mM) EMS [cf. (36) for details]. Leaves from young M₂ plants of each population were sampled for DNA extraction and screening for mutations in the homoeologous wheat DME and DRE2 genes. Two to 20 M₃ seeds (depending on the M₂ plant zygosity) from selected genotypes were cultivated till maturity in 20 × 25 cm pots using Sunshine #1 potting mixture (SunGro Horticulture, Bellevue, USA) in a greenhouse maintained at a daytime temperature of 20–23°C and a night time temperature of 14–16°C with a photoperiod of 18 h. During this period, plants were fertilized weekly with nutrient water containing 200–250 ppm of water-soluble fertilizer [nitrogen (N), phosphorus (P₂O₅), and potassium (K₂O)]. One-month-old M₃ plant leaves were sampled for DNA extraction to determine the zygosity of mutant plants, and M₄ seeds were harvested for protein profiling and determination of other end-use quality parameters.

For large-scale TILLING library screening, DNA from two individual M₂ plants was pooled. The DNA concentration for each individual within the pool was adjusted to about 2 ng/μL with a final concentration of 4 ng/μL for each pool. Twenty nanogram DNA (i.e., 5 μL of the pooled DNA samples) was arrayed on microtiter plates and subjected to gene-specific PCR. Amplification and TILLING were performed as described in Moehs et al. (34). The Kronos TILLING population developed at UC Davis was searched for mutations in the homoeologous DRE2 genes via performing blast searches against the exon capture sequences derived from 1,535 EMS mutants (TILLING Genomic Ref. version 0.4 at https://dubcovskylab.ucdavis.edu/wheat_blast) using highly stringent search criteria [cf. (32)].

Leaves of 3-week-old M₃/M₄ seedlings were collected in 96-well deep-well plates and dried for 72 h in closed tubes with silica gel. DNA was extracted from dried and finely ground samples using a modified cetyltrimethylammonium bromide (CTAB) method following Xin and Chen (37). DNA was treated with RNase and purified using the phenol: chloroform: isoamyl alcohol (25:24:1) precipitation. Then, the concentration of DNA samples was determined and adjusted to 50 ng/μL.

To determine the zygosity of individual M₃/M₄ plants, the target region from each line was amplified via PCR, and the amplicons were gel purified using the Geneclean kit following the manufacturer's instructions (MP Biomedicals, CA, USA). The eluted DNA was used as the template for the sequencing reaction using either forward or reverse primer in separate reactions. Sequencing reactions were carried out in 10 μL volume, and each reaction contained 100 ng DNA, 0.5 μM forward or the reverse primer, and BigDye mixture (Applied Biosystems, CA, USA). The following protocol was used for PCR: 96°C for 10 s, 50°C for 15 s, 60°C for 6 min with 24 iterations. After PCR, the product was column purified and resolved on the ABI 3730 DNA Analyzer. The DNA sequences were analyzed using the Sequencher 4.9 software (Gene Codes Corporation, MI, USA) to determine each line's genotype.

Genomic DNA was extracted from wheat cultivar “Chinese Spring,” the nullisomic-tetrasomic lines lacking chromosome 5A, 5B, or 5D in the Chinese Spring background and month-old plants of diploid wheat progenitors T. urartu, Ae. speltoides, and Ae. tauschii. After quantification, DNA was used to amplify DME genes from specific homoeologous chromosomes using sub-genome specific primers.

RNA was extracted from developing T₄/T₅ grains 17 days after anthesis (±3 days) when the immature grains were 50–75% filled. The spikes were harvested in liquid nitrogen and about 0.3 g of the developing grains was pulverized to fine powder and total RNA was extracted with TRIzol reagent (15596-018, Invitrogen) according to the manufacturer's recommendations. RNA was quantified using Bio-Rad SmartSpec 3000 (Bio-Rad Laboratories), and ~1 μg of RNA was converted to cDNA using a reverse transcription system with random oligos following the manufacturer's instructions (A3500, Promega).

The qRT-PCR analysis was performed using the LightCycler DNA Master SYBR Green I system (12158817001, Roche Diagnostics) on the LightCycler^® 480 Real-Time PCR System (05015278001, Roche Diagnostics) with the wheat DME primers (non-specific to homoeologous genes) 5′ TGTGCGTCTTTTGACACTCC 3′ and 5′ GCTCGTACAATGTCCGTTGA 3′ and Actin specific primers 5′ CACACTGGTGTTATGGTAGG 3′ and 5′ AGAAGGTGTGATGCCAAAT 3′ strictly following Wen et al. (21). DEMETER mRNA level was normalized to Actin using the DDC_T method. Transcript levels were expressed as a DME transcripts (normalized to Actin) ratio in control Kronos or Express plants to the EMS mutants.

Gluten proteins were extracted from the mature grains of the M₄/M₅ lines and wild type controls (Kronos and Express). In each case, three to four random grains were used for gluten extraction. For this purpose, seeds were dissected into the endosperm part and the germ part, the germ part was saved for plant propagation in the greenhouse, and the endosperm part was used for sequential gliadin and glutenin extraction.

The endosperm of the single seed was ground to flour using pestle and mortar. The amount of flour was determined by weighing, and 210 μL of 60% ethanol (v/v) was added to the flour samples weighing >20 mg and 100 μL of 60% ethanol to flour samples weighing <20 mg. In most cases, the endosperm-half of a single seed weighed around 20–80 mg. After ethanol addition, tubes were vortexed for 45 s to mix the flour with ethanol and incubated at room temperature with constant shaking at 200 rpm for 1 h in an orbital shaker. After incubation, samples were centrifuged at 8,200 rpm for 20 min, and the supernatant (carrying gliadins) was transferred to a new tube and stored at 4°C.

The pellet from the previous step was used for glutenin extraction. Two hundred and thirty microliter of freshly prepared glutenin extraction buffer [60% 1-propanol (v/v), 2 M urea (w/v), 0.05 M Tris-HCL (pH 7.5), 2% dithiothreitol (DTT) (w/v)] was added to the tube. Since DTT is unstable at room temperature, a fresh solution was made each time and added to the extraction buffer. The pellet was resuspended in the buffer by vortexing the tubes for 45–60 s or until a homogeneous mixture was achieved and incubated at 60°C with constant shaking at 200 rpm for 2.5 h in an orbital shaker. Subsequently, the mixture was centrifuged at 8,000 rpm for 20 min, then at 13,000 rpm for 2 min at room temperature. The higher centrifugation speed used in the second step allowed the pellet to stick to the tube wall and facilitated supernatant collection. The supernatant was collected (carrying glutenins) to a new tube and stored at 4°C until used.

The gliadins (the DRE2 TILLING mutants and wild type Kronos) and glutenins extracted from mutant and wildtype lines (three biological replicates) were resolved on a 10% SDS-PAGE gel following Mejías et al. (38). Briefly, SDS-PAGE gels used in the present study consisted of two layers, stacking gel and separating gel. The separating gel was prepared by mixing 4 mL of acrylamide-bisacrylamide stock solution [30% acrylamide (w/v) and 0.78% bisacrylamide (w/v)] with 5 mL of distilled water, 3 mL of 3 M TRIS-HCL (pH 8.8), 120 μL of 10% SDS (w/v), 120 μL of 10% APS (w/v), and 6 μL of TEMED (Tetramethylethylenediamine). The stacking gel was prepared by mixing 1 mL of acrylamide-bisacrylamide stock solution with 4.25 mL of distilled water, 750 μL of 1M TRIS-HCL (pH 6.8), 60 μL of 10% SDS (w/v), 60 μL of 10% APS (w/v), and 4 μL of TEMED. After electrophoresis and staining, the gels were visualized on a lightbox and documented as mentioned in Mejías et al. (38).

Aluminum lactate polyacrylamide gels were prepared following Mejías et al. (38) with minor modification and used to profile the gliadin extracts from the DME TILLING mutants and wild type Kronos and Express. Gliadin samples were prepared for electrophoresis by adding 8 μL of 5 × loading buffer (containing 5 g sucrose and 2 mg of methyl violet in 10 mL of water) to 15 μL of gliadin extract before loading. Electrophoresis was performed at 500 volts (40 mA/gel) for 120 min using the SE 600 Ruby system (80-6479-57, GE Healthcare Life Sciences) in the lactate buffer, pH 3.1 at 15°C. After electrophoresis, the gels were equilibrated for 30 min in 10% trichloroacetic acid (TCA). A 0.5% solution Coomassie Brilliant Blue R-250 (BP101-25, Fisher Scientific) in ethanol was then added. The gels were stained with gentle shaking for 18 h at room temperature, and then each gel was destained in 450 mL of distilled water with 4 drops of Triton X-100 for 30 min. After staining, the gels were visualized on a lightbox and documented with a scanner.

In preparation for HPLC analysis, both gliadin and glutenin extracts from the DME TILLING mutants and wild type Kronos and Express were passed through 0.45 μm Durapore^® 13 mm membrane filters. The filtered samples were then injected onto a reversed-phase C8 analytical column (Zorbax 300SB-C8, Agilent Technologies) with 5 μm particle size and 30 nm microporous silica diameter (250 mm length, 4.6 mm I.D.), using a 1200 Series Quaternary LC-System liquid chromatograph (Agilent Technologies), with a diode array UV-V detector. The column temperature was maintained at 60°C, and a linear elution gradient was implemented using two mobile solvents. The polar solvent A consisted of a mixture of 0.1% trifluoroacetic acid (TFA) (v/v) and type-I ultrapure water (18 MΩ·cm specific resistance), and the non-polar solvent B contained 0.1% TFA (v/v) and acetonitrile (ACN). Absorbance was monitored at a detection wavelength of 210 nm. The injection volumes for both the gliadin and the glutenin fractions were 30 μL, and the flow rate was adjusted to 1.0 mL min⁻¹. The elution gradient conditions were as described in Mejías et al. (38).

HPLC was used to determine relative quantities of gliadins and glutenins in a sample. To obtain a standard curve using HPLC, increasing concentrations of bovine serum albumin (BSA; 0, 10, 25, 50, 75, 100, 150, and 300 mg/mL) were loaded onto the C8 reversed-phase analytical column (Zorbax 300SB-C8, Agilent Technologies) and resolved following Mejías et al. (38). Two peaks were observed, respectively, at 16.78 and 17.98 min retention times. Peaks were integrated, and the obtained peak areas in milli-absorbance units (mAU) were regressed against the BSA concentrations loaded onto the HPLC column to obtain a standard curve. The gliadin and glutenin extracts of the control and DME TILLING mutants were analyzed analogously as mentioned above for BSA, and the peak areas were calculated. The mAU values were plotted on the standard curve to determine relative protein concentrations, and the values obtained for each peak were summed to calculate a cumulative concentration for each of the glutenin and gliadin fractions from control and mutant lines.

Gel images were captured on a lightbox using a Canon DS127721 DSLR camera. Gel images were all set to the same brightness and contrast setting using Adobe Photoshop ver. 20.0.06. Densitometric quantification of the protein bands was performed using ImageJ software (39) following (38). Briefly, each image was transformed to 8-bits. Subsequently, the image background was subtracted, and the optical density (OD) for each protein band was determined. The OD of each band shows up as a peak on the densitogram, where the area under each peak corresponds with the OD of the protein bands.

Seven seeds each of the wild type control and the DME TILLING mutant lines were used for studying the total grain protein content (GPC) using the LECO FP-228 Combustion N analyzer (FP228, LECO Corporation). The thousand-kernel weight was determined as an estimation of the grain yield.

Three procedures were applied to estimate the baking quality of the wildtype control and the DME TILLING mutants. (i) 500 mg of ground seeds from each genotype were used in the SDS sedimentation test as described by Carter et al. (40). (ii) 200 mg of ground seeds from each line were used in the sodium carbonate solvent retention capacity (SRC) test. The requisite weighed sample was added to a pre-weighed tube, and then 1 mL of a 5% (w/w) sodium carbonate solution was added. The mixture was vortexed and then allowed to gently and continuously agitate on a Labquake rotator for 20 min. The samples were then centrifuged at 1,000 × g for 15 min and the supernatant decanted. The tubes were inverted for 10 min and then swabbed with cotton to remove any extra solvent. The tubes were then weighed again, and the weight of the flour was determined by subtracting the tube's weight from that of the tube and flour. The weight of the dry flour and the solvated flour or gel's weight was then used to calculate the SRC for each sample. (iii) 200 mg of ground seeds from each genotype were used in the sucrose SRC test. This test resembles the sodium carbonate SRC test, except in this case, instead of adding 1 mL of a 5% (w/w) sodium carbonate solution, 1 mL of a 50% (w/w) sucrose solution was added to samples, and the retention of sucrose in flour was studied.

Mutations in the single homoeologous DME genes on the “A” and “B” sub-genome of bread and durum wheat were crossed in combinations to obtain DME double mutants. All crosses were made reciprocally and in duplicates. The F₁ and later F₂ seeds were obtained for 3 and 4 different mutant combinations, respectively, in Kronos and Express backgrounds. The mutations for stacking were selected based on the type (substitution, splice site variation, or premature stop codon), locations in the DME active site, and effect on gluten accumulation in the wheat endosperm. The F₂ grains obtained from the aforementioned crosses were propagated in 48 well-flats and checked for zygosity of double mutations by PCR followed by DNA sequencing.

A pair of DME-specific primers (potentially amplifying the homoeologous wheat DME genes) covering nucleotide positions 13,263–13,618 on barley DME sequence FM164415.1 was designed and used to amplify fragment(s) from wheat genomic DNA. The PCR product was used to screen the common wheat (Triticum aestivum L.) cv. “Chinese Spring” BAC library, with ca. 1.3 million clones. Macroarray hybridizations led to the identification of three unique BAC clones, namely 1946D08, 2106P11, and 2159B03 (Supplementary Figure S1). Subsequently, the BAC clones were sequenced at >60-fold coverage by the 454-sequencing method. A total of 38.9 Mb of good quality sequence was obtained. Analysis of these BAC sequences revealed that each of them harbors a full-length DME sequence (accession numbers JF683316-JF683318). However, the three DME sequences differed from each other in length and composition. The observed differences in the length of DME homoeologues were mostly due to insertions and deletions (InDels) in the introns. Many point mutations [135 Homoeologous Sequence Variants (HSVs)] and small InDels observed among the DME homoeologous belonged to exons. These HSVs were used to design homoeologue-specific primers by tagging their 3′-ends at HSVs, and these primers were used to assign the DME homoeologues to specific sub-genomes (Supplementary Table S1). For this purpose, each homoeologue-specific primer pair was used on the genomic DNA of diploid wheat progenitors [T. urartu (AA), Aegilops speltoides (BB), and Ae. tauschii (DD)] with Chinese Spring [T. aestivum (AABBDD)] as control. These primers allowed unambiguous assignment of 2159B03 to chromosome 5A, 1946D08 to chromosome 5B, and 2106P11 to chromosome 5D (Figure 1). Sub-genome assignment of the DME homoeologues was further validated by the use of wheat group 5-specific nullisomic-tetrasomic lines (Figure 1).

The DRE2 or Derepressed for Ribosomal protein S14 Expression gene facilitates the deposition of the iron-sulfur (Fe-S) cluster into the DME apoenzyme, which is vital for its interaction with genomic DNA and the DNA's subsequent demethylation. Given DRE2's role in DME activation, we decided to test its effect on prolamin accumulation. As a first step, we obtained the full-length DNA sequences of the wheat DRE2 homoeologues following a comparative genomics approach. A two-step procedure was adopted as the Arabidopsis DRE2 sequence (AT5G18400) did not yield a wheat sequence with apparent homology. Therefore, first, a putative Arabidopsis DRE2 homolog was identified in rice and Brachypodium, then these sequences were used to identify homologous sequence(s) in common wheat. These searches allowed the identification of two homologs in rice (OS04G0674400 and OS04G0682050) and two homologs in Brachypodium (BRADI1G25576 and BRADI5G25970). When rice and Brachypodium sequences were blasted against the wheat reference genome sequence, the results suggested two homologs, one on group 2 chromosomes with a copy each on 2AL (TraesCS2A02G553500), 2BL (TraesCS2B02G585100), and 2DL (TraesCS2D02G555000), and another on group 5 chromosomes with a copy each on 5AL (TraesCS5A02G262300) and 5DL (TraesCS5D02G269900) (Figure 1). The copies on group 2 chromosomes closely resembled the rice and Brachypodium DRE2 genes. Hence only this gene was considered further in this study. These DRE2 homoeologues were amplified from durum wheat cv. “Kronos” in three overlapping fragments that covered the area from the transcription start site to the termination codon of the gene (Figure 1). The amplified fragments were cloned in the pGEM^®-T Easy vector (Promega, Madison), and 20 clones per fragment were sequenced to have a fair representation of each homoeologue in the sequences. Obtained sequences were aligned with each other and the reference sequence. The sequence analysis revealed that we obtained the full-length sequence of the wheat DRE2 homoeologues.

Gene expression data across wheat grain development was accessed from the Wheat Expression Browser (http://www.wheat-expression.com) to understand the transcriptional interplay among the wheat DME and DRE2 genes and genes encoding immunogenic gluten proteins (http://www.allergenonline.org/celiacproteinbrowse.shtml). This data is largely based on the study of Ramírez-González et al. (41). The DME and DRE2 genes did not express in concert with each other. Similarly, the homoeologous copies of DME and DRE2 exhibited asymmetrical expression patterns throughout the studied grain development period, i.e., 6–30 days post-anthesis (DPA) (Figure 2). The DME homoeologue on 5DL exhibited a dominant expression pattern throughout this period in the starchy endosperm and the aleurone layer (Figure 2). On the contrary, the DRE2 homoeologues on 2AL and 2DL exhibited a developmental stage-specific asymmetrical expression pattern, where its 2AL-copy exhibited the highest expression at the 6 and 14 DPA in both starchy endosperm and aleurone and the 2DL-copy exhibited the highest expression at 12 DPA. In the starchy endosperm, the expression patterns of DME and DRE2 seemed to somewhat follow a similar trend up to 20 DPA, whereas they exhibited opposite expression patterns in the aleurone. The expression patterns of these genes were compared with that of the expression patterns of immunogenic gluten proteins. An opposite expression trend was observed, which means the peaks of high gene expression for DME and DRE2 opposed valleys of low gene expression for the immunogenic prolamins (Figure 2). For instance, the α-gliadins exhibited the highest expression during grain development with a peak at 12 DPA, whereas both DME and DRE2 exhibited the highest expression, respectively, at 9 and 6 DPA. The second peak of α-gliadin expression was observed at 20 DPA and the third peak at 30 DPA, whereas for both DME and DRE2, their gene expression had started to decline at these developmental time points. The expression of other gliadins and glutenins started to rise at 30 DPA (Figure 2). In contrast, an opposite trend was observed for DME and DRE2. Among gluten proteins, the expression pattern of HMWgs genes somewhat followed the trend of the α-gliadins (Figure 2). The γ- and ω-gliadins showed an opposite expression pattern, and LMWgs showed a gradual increase in their expression between 6 and 30 DPA (Figure 2). In the aleurone layer, the expression patterns of DRE2 and the immunogenic prolamin genes somewhat overlapped, but their expression was half that of their expression in the starchy endosperm (Figure 2).

Total RNA and gluten proteins were extracted from the developing wheat grains at 5, 7, 8, 10, and 13 DPA to validate the results of in-silico expression analysis. For the sake of precision about the anthesis date, wheat cv. “Express” spikes were emasculated and artificially pollinated with its pollen. Later, we collected the developing kernels in liquid nitrogen for expression analysis and protein profiling. For expression profiling of the homoeologous wheat DME genes, a set of primers (capable of amplifying all three DME homoeologues) were used to amplify products from cDNA, and the amplified product was cloned in a TA-cloning plasmid. Thirty colonies were sequenced via Sangers sequencing, and the expression level of different homoeologues was determined via sequence comparison with the genomic DNA sequences of the wheat DME homoeologues. At 5 and 8 DPA, the three homoeologous DME genes were expressed equally, but at the rest of the studied time points, the “A” and “D” subgenome DME homoeologues showed opposite expression patterns. In contrast to the “A” and “D” homoeologues, the “B” subgenome DME homoeologue maintained a constant expression level until 10 DPA (Figure 3).

Gliadins and glutenins were extracted from the developing grains, as described in the materials and methods, and resolved via HPLC. The protein quantity for seed samples collected at 5 and 7 DPA was too low to be resolved via HPLC; therefore, those samples were excluded from the analysis. At 8 DPA the γ-gliadins and LMWgs accumulated, and at 10 DPA γ- and α/β-gliadins and LMWgs accumulated, and by 13 DPA all prolamins accumulated (Figure 3).

Our earlier research established a general connection between the wheat DME expression and prolamin accumulation in the endosperm (21, 22). However, the relationship between the expression pattern of DME homoeologues and the accumulation of different prolamins during grain development was not studied. When the accumulation of prolamins was compared with the DME expression levels, it suggested that “B” and “D” subgenome DME homoeologues are largely responsible for initiating and maintaining transcription of genes encoding γ-gliadins at 8 DPA (Figure 3). By contrast, the “A” subgenome DME homoeologue, whose expression increases with time at 8 and 10 DPA, triggers the expression of genes encoding ω-gliadins and LMWgs (Figure 3). At the 13th DPA transcript of the “D” subgenome DME homoeologue reaches its maximum to induce expression of α/β-gliadin genes (Figure 3). This analysis suggests that the DME homoeologues have sub-functionalized to some extent during wheat evolution, but still retain some overlapping function.

Tetraploid “Kronos” and hexaploid “Express” TILLING populations were screened for mutations in the wheat DME homoeologues. The average mutation density in the EMS mutagenized Kronos M₂ population was 1 mutation per 40 kb DNA, and in the Express M₂ population it was 1 mutation per 24 kb DNA (3). The mutations are primarily single nucleotide polymorphisms or small deletions.

Two runs, one each with a set of subgenome-specific primers, were executed on “Kronos” and “Express” M₂ DNA bulks. The subgenome specific primers amplified 1,050 bp from the “A” subgenome of “Kronos” and “Express” and 1,044 and 855 bp from the “B” subgenome of “Kronos” and “Express,” respectively (Supplementary Table S1). The amplified region covered exon 5 to 9 of the genomic sequence containing the active site of the DME enzyme. In total, 42 mutations in “A” homoeologue and 35 mutations in “B” homoeologue of the durum wheat DME genes (TdDME) were detected. Among 42 mutations detected in the “A” homoeologue and 35 mutations detected in the “B” homoeologue, 4 and 5 mutations, respectively, coincided with the conserved motifs in the DNA glycosylase domain [consisting of a helix–hairpin–helix (HhH) motif, a glycine/proline-rich loop, and a conserved aspartic acid (GPD) residue]. A mutation resulting in a premature stop codon falling between the two conserved domains was also detected in the “A” homoeologue of TdDME. Similarly, 42 mutations in “A” homoeologue and 56 mutations in “B” homoeologue of the common wheat DME genes (TaDME) were detected (Table 1, Figure 4). Out of 42 mutations detected for the “A” homoeologue and 56 mutations detected for the “B” homoeologue, 6 and 5 mutations, respectively, coincided with the conserved domains, whereas four mutations, two each detected in “A” and “B” homoeologues resulted in either premature stop codons or splice site variants, that fell between the two conserved domains. Similarly, a “D” subgenome specific primer pair amplifying a 1,008 bp fragment from the common wheat genome was used to screen “Express” M₂ DNA bulks. Three rounds of screening allowed the identification of 93 mutations in the “D” homoeologue of TaDME. Out of these 93 mutations, 11 coincided with the glycosylase domain (Table 1, Figure 4). One mutation resulted in a premature stop codon, and fell between the two conserved domains. Interestingly a point mutation at a conserved aspartic acid residue, repeatedly shown to disrupt DME enzyme activity in Arabidopsis, was also identified (43).

A total of 344 M₃ plants, including 130 plants, respectively, representing 3 and 7 (either in homozygous or heterozygous state) mutations in “Kronos” and “Express” backgrounds, underlying the conserved domains of “B” subgenome DME homoeologue, and 214 plants representing 5 and 9 mutants, respectively, in “Kronos” and “Express” backgrounds underlying the conserved domains of “A” subgenome DME homoeologue, were planted in the greenhouse. Similarly, a total of 44 M₃ plants representing 9 (either in homozygous or heterozygous state) mutations in the “Express” background, underlying the conserved domains of “D” subgenome DME homoeologue, were propagated in the greenhouse. DNA was extracted from the leaves of 2 weeks old plants, used to amplify subgenome-specific DME products, and to determine the zygosity of plants by sequencing the PCR products (Supplementary Figure S2). The analysis allowed the identification of at least one homozygous M₃ plant for each heterozygous M₂ stock, except for the four cases where no homo/heterozygous mutant plant could be recovered (Supplementary Table S2).

A total of 36 mutations were identified in the 2AL DRE2 homoeologue. Of these 36 mutations, ten mutations fell in introns, and 26 in exons−11 are synonymous variations, 12 are missense variations, two are amber mutations, and one is a splice site variant. Similarly, a total of 15 mutations were detected in the 2BL copy of the DRE2 gene. Of these mutations, six fell in introns, five in exons, and four in the 3′-untranslated region. Two of these five mutations are synonymous variations, two are missense variations, and one is a splice site variant. Seeds of 13 DRE2-2AL mutants, including two amber mutants and 11 missense mutants, and three DRE2-2BL mutants with two missense mutants and a splice site variant were procured from the Dubcovsky Lab at UC Davis. Four to eight seeds of each line were propagated in the greenhouse, and the genotype of each mutant line was confirmed by DNA sequencing. At least one homozygous plant in each mutant was identified. All mutant lines were allowed to self-pollinate, and seeds from homozygous lines were used to determine the gluten profile.

Immature seeds (17 ± 3 DPA) from the homozygous M₃ plants were used for RNA extraction, followed by qRT-PCR analysis. In reference to the wheat Actin gene, five DME mutants, namely 36505 (W42^*), 42773 (V29M), and 33412 (splice site variant) in the “Kronos” background and 50324 (M89I) and 21599 (G93R) in the “Express” background exhibited transcriptional suppression (Supplementary Figure S3). On the other hand, none of the DRE2 mutant lines exhibited any difference in their DRE2 transcript levels.

Total grain protein from the 16 selected mutant lines, all carrying mutations in the active site region of TaDME and TdDME, were extracted from mature seeds and profiled using SDS-PAGE followed by densitometry. All genotypes showed reduced accumulation of one or more immunogenic gluten proteins. In particular, lines carrying premature stop codons (“Express” mutant “20396” in the “B”-subgenome and “12680” in the “A”-subgenome), splice site variant (“Kronos” mutant “33412” in the “B”-subgenome), or a non-synonymous substitution (“Express” mutant “50322” with M89I substitution in the “B”-subgenome, “13105” with L9F substitution, and “50449-6” with D30N substitution in the “D”-subgenome) showed significant reductions in gliadin and glutenin content (Figure 4, Supplementary Table S3). Gliadins and glutenins extracted from these mutants were also resolved using RP-HPLC. The area under peaks representative of different protein families on the chromatogram was converted to protein quantity in mg mL⁻¹ compared to a standard curve developed using bovine serum albumin (BSA) resolved under similar conditions on RP-HPLC. The analysis suggested that in comparison to the wildtype “Express,” the single mutant “12680” (stop gained in “A” subgenome DME copy) accumulated significantly reduced amounts of HMWgs, LMWgs, and α/β-gliadins (about 81, 79, and 81% reductions, respectively). A similar great reduction in ω-gliadin content was observed in “Express” mutant “50322” (an M89I substitution in the “B”-subgenome DME copy) and γ-gliadins in “Express” mutant “50449” (a D30N substitution in the “D”-subgenome DME copy) (Table 2). Likewise, compared to wildtype “Kronos” we observed significant reductions in the content of HMWgs, LMWgs, ω-gliadins, and α/β-gliadins (68, 66, 60, and 44%, respectively) in mutant “33412” (a splice site variant in the “B”-subgenome DME copy) (Table 2).

Three seeds (biological replicates) from 16 mutant lines (homozygous for mutations in the DRE2 gene) were used to extract gliadins and glutenins, and the proteins were profiled using SDS-PAGE followed by densitometry. Protein profiles of mutants were compared with wild-type “Kronos” to study the effects of mutations in the individual homoeologous DRE2 genes on chromosome arm 2AL and 2BL. Two types of differences were observed: qualitative differences and quantitative differences, where qualitative differences are the presence or absence of specific protein bands, and quantitative differences are under-accumulation or over-accumulation of the specific protein(s) (visible as reduced or high-intensity bands). Compared to “Kronos”, nine of 16 mutants showed quantitative differences, and seven mutants showed qualitative differences. Most if not all DRE2 mutants exhibited reduced accumulation of ω- (27% reduction to complete lack) and γ-gliadins (40–93% reduction) and over-accumulation (range from 1.1 to 2.7 fold) of α/β-gliadins (except a non-synonymous T121I substitution in the DRE2 2AL homoeologue). Further, some mutants exhibited over-accumulation of HMWgs (range from 1.3 to 5.5 fold) and LMWgs (range from 1.1 to 2.2 fold), whereas some showed reduced HMWgs (14–78%) and LMWgs (6–44%) accumulation (Table 3, Figure 5).

We performed micro SRC and SDS sedimentation tests and determined the total grain protein content (GPC)—using the LECO FP-228 Combustion N analyzer (FP228, LECO Corporation)—of reduced-gluten DME mutants to predict their end-use performance. Kernel weight was determined for each line (from the seven randomly selected grains, also used later to determine the GPC); it ranged from 0.17 g in “Express” mutant “12680” to 0.31 g in “Express” mutant “20396”, which is significantly less or more than 0.29g in “Express”. The protein content expressed in percentage ranged from 13.17% in “Express” mutant “50449” to 26.45% in “Kronos” mutant “33412”, which is, respectively, less and more than “Express” (17.72%) and “Kronos” (22.68%). The components for diagnostic SRC solvents (water, dilute aqueous lactic acid, dilute aqueous sodium carbonate, and concentrated aqueous sucrose solution) were calculated using standard formula and assuming 14% water content of the flour. Values obtained using 50% sucrose ranged between 136% in “Express” mutant “50449” (similar to “Express”) to 178% in “Express” mutant “12680” and 143% in “Kronos” mutant “33412” (significantly <168% in “Kronos”), indicating higher water retention capacity in the finished bread. The results of the sodium carbonate test of the lines ranged from 129% in “Express” mutant “20396” to 196% in “Express” mutant “12680” (139% in “Express”) (Supplementary Table S4). The value of 146% for “Kronos” mutant “33412” is at par with “Kronos” (149%). It is an indicator of damaged starch, which leads to bread staling and reduced shelf life. Values for the lactic acid SRC ranged from 149% for “Express” mutant “50449” to 231% for “Express” mutant “12680,” which is significantly lower and higher than “Express” (180%). Similarly, for water SRC lowest (131%) and highest (214%) values were observed for “Express” mutants “50449” and “12680”, respectively. The gluten performance index (GPI) was calculated using the values of micro SRC. It is an indicator of the overall performance of flour glutenin in the presence of other flour polymers. The GPI ranges from 0.53 (“Express” mutant “50449”) to 0.76 (“Express” mutant “51464”) in the analyzed samples, lower than what is desired for bakers' flour (personal communication, Craig F. Morris; deceased October 25, 2021). Only two lines obtained values acceptable for bakers' flour, lines “20396” and “51464” with 0.71 and 0.76, respectively (Supplementary Table S4).

The single mutants identified in the “A” and “B” subgenome DME homoeologues of common and durum wheat were crossed in combinations to obtain DME double mutants to stack their gluten phenotypes. Crosses were made reciprocally to account for possible maternal influence, which is a prerequisite, specifically when dealing with epigenetic regulators (Supplementary Figure S4). The mutants for crossing were selected based on the mutation type (substitution, splice site variation, or premature stop codon), their respective DME active site locations, and their gluten profiles. F₂ seeds were obtained for three combinations in the “Kronos” background and four combinations in the “Express” background (Supplementary Figure S4). The F₂ grains were planted (in 48-well flats) in the greenhouse and checked for zygosity at two DME loci via DNA sequencing. Sequence analysis allowed the identification of 20 double mutations in the “Kronos” background and 14 double mutants in the “Express” background (Supplementary Table S5). F₃ grains were obtained from the double mutants.

All double mutants were identified as crosses between mutants carrying an amino acid substitution or a splice site variation in single DME homoeologues. However, even after screening a few hundred F₂ plants, no double mutants were identified from the crosses between mutants carrying a premature stop codon in one homoeologue and an amino acid substitution in another or premature stop codons in both homoeologues in both “Express” and “Kronos” backgrounds. However, a double mutant with a splice site variant in one homoeologue and an amino acid substitution in another homoeologue was identified in the “Kronos” background.

Protein profiles of the mature F₃ seeds, specifically a double mutant derived from a cross between “50322” (M89I, DME-5B) × “20548” (V29M; DME-5A) in the “Express” background, exhibited reduced accumulation of gliadins and LMWgs (greater reductions in ω- and α/β-gliadins and LMWgs, and less reductions in γ-gliadins) (Table 2). However, these reductions in gliadins and glutenins in the double mutant were relatively less apparent than in the single mutants with premature stop codons (Figure 4). The HPLC analysis of gluten proteins endorsed these observations. For instance, relative to “Express” (wildtype), a 50.8% reduction in ω-gliadins, a 43% reduction in LMWgs, and a 25.2% reduction in α/β-gliadins was observed (Table 2). On the other hand, the double mutant derived from a cross between “33412” (splice site, DME-5B) × “29375” (A111T, DME-5A) in the “Kronos” background exhibited a gluten profile reduced in gliadin and glutenin content more than any other single mutant (Figure 4). These results were supported by the results of HPLC analysis where >70% content reductions in α/β- and γ-gliadins and LMWgs were observed (Table 2). Preliminary analysis of the end-use performance of the double mutants “50322” × “20548” in the “Express” background and “33412” × “29375” in the “Kronos” background are at par with the respective wild type controls expect for the SDS sedimentation test values, which were twice as much (42 mm) as that of the control (24 mm) for the “50322” × “20548” double mutant and half that of the control (33 mm) for the “33412” × “29375” mutant (16 mm) (Supplementary Table S4).

The results of initial crosses made between the mutations in the “A” and “B” subgenome DME homoeologues showed distorted segregation in most of the combinations except for the cases where crosses were made between genotypes carrying amino acid substitutions (with mild effects on the gluten profiles). To determine the cause of the observed segregation distortion, we studied anther morphology, anatomy, number of viable pollen grains, and pollen germination (in pollen germination media supplemented with azadecalin) (44). Interestingly, in contrast to the barley high-lysine mutants, DME mutants carrying premature stop codons or splice site variants, especially in the tetraploid wheat background, showed reductions in anther size, locule size, number of fertile pollen grains, and pollen germination rate (Figure 4). The reduced pollen germination adversely affects the competitiveness of the mutant pollen grains relative to the wild-type pollen grains causing severe segregation distortion. Indeed, analysis of hundreds of F₂ plants from crosses between “A” and “B” sub-genome DME mutants with premature stop codons yielded either single mutants or wild type alleles at both homoeologues. In addition, a reduction in seedling vigor of the DME mutants with stop codons was also observed.

Crosses between mutants belonging to other homoeologous groups were attempted; for instance, crosses were made between “A” and “D” and “B” and “D” subgenome DME homoeologues. When mutants carrying splice site variants and/or premature stop codons in different DME homoeologues were crossed in combinations, a limited number of grains were obtained. Unfortunately, none of the tested genotypes carried double mutations either in the homozygous or the heterozygous state. Similarly, the DRE2 mutants (except four mutants) in either the 2AL copy or the 2BL copy due to their effect on pollen viability and anther size were maintained in the heterozygous state, making it challenging to obtain homozygous lines to make genetic crosses to stack their gluten phenotypes.