Temporal Expression Analysis of Barley Disproportionating Enzyme 1 (DPE1) during Grain Development and Malting

Abstract Disproportionating enzyme (D-enzyme) is a 4-α-glucanotransferase (EC 2.4.1.25) that disproportionates α-glucans and maltooligosaccharides and preferentially disproportionates maltotriose into maltopentaose and glucose. Maltotriose is a maltooligosaccharide that accumulates in barley malt and wort without being digested further by starch degrading enzymes. Furthermore, not all brewer’s yeast strains can ferment maltotriose. This research was undertaken to determine if D-enzyme is expressed in developing and/or malting barley grains and thus providing evidence of an inherent enzymatic mechanism capable of disproportionating maltotriose into maltopentaose that can be further degraded into fermentable sugars by amylolytic enzymes such as β-amylase. A partial genomic sequence of barley disproportionating enzyme 1 (DPE1) was obtained that was comprised of 16 exons and 15 introns totaling 4680 bp. The 5’ region of the DPE1 gene was recalcitrant to Sanger sequencing owing to its localization in a highly repetitive region of the barley genome. The DPE1 gene is expressed during grain development and the protein stored in the mature grain. Additionally, the DPE1 gene is de novo expressed during malting in both a 2- and 6-row malting cultivar with significant variation observed amongst 13 elite malting cultivars representing spring and winter growth habits. During grain development, DPE1 mRNA levels peak at 17 days after anthesis, which coincides with a large increase in proteins that react to anti-DPE1 polyclonal antibodies. These proteins appear as a doublet on immunoblots during initial stages of malting and as a singlet as malting progresses indicating proteolytic processing (e.g., transit peptide removal) or differential isoform/splice variant expression.


Introduction
Disproportionating enzyme (D-enzyme, EC 2.4.1.25) is a 4-α-glucanotransferase that catalyzes the transfer of a segment of (1,4)-α-D-glucan to a new position in an acceptor, which can be either a glucose or a (1,4)-α-glucan. D-enzyme was first discovered in potato tubers and has since been characterized in many species including, pea, Arabidopsis, rice, and wheat. [1][2][3][4][5][6] In plants, there are two known isoforms of D-enzyme, encoded by DPE1 and DPE2, that are differentially localized. [2,7] DPE1 encodes a protein that is localized in plastids whereas DPE2 is found in the cytosol. [7,8] In wheat and rice, DPE1 is localized in endosperm amyloplasts. [2,9] The D-enzyme encoded by DPE1 best utilizes maltotriose as a substrate and is unable to utilize maltose whereas DPE2 utilizes maltose as a substrate. [5,8] D-enzyme has the highest activity when using maltotriose as a substrate, in the reaction of maltotriose + maltotriose ←→ glucose + maltopentaose, with weakening activity as the maltose chain increases thus preferentially utilizing maltotriose. [4] In potato, D-enzyme can also work on large chain maltooligosaccharides such as maltopentaose producing larger chain molecules similar to short chain amylose. [6] Dual roles in starch metabolism have been postulated for D-enzyme. DPE1 has been shown to be necessary for normal starch synthesis in rice. [9] In Arabidopsis, D-enzyme is critical in leaf diurnal transitory starch metabolism. [3] Conversely, in potato and Arabidopsis, D-enzyme appears to play a role in starch degradation [3,6] by disproportionating maltooligosaccharides creating suitable substrates for both phospholytic and hydrolytic starch degradation. D-enzyme knockouts also exhibit increased maltotriose levels while slowing the rate of starch degradation in leaves. [3] The efficiency of starch degradation during the brewing process is of the utmost importance to the brewing industry. Malting is a controlled industrial germination process whereby barley grains are first soaked in water (steep) to initiate germination followed by incubation at cool temperatures with periodic rotation. During malting, starch degrading enzymes are de novo expressed or stored enzymes are activated. In the brewing process, the primary responsibility of the starch degrading enzymes is to provide enough diastatic power to breakdown starch into sugars CONTACT  that can be fermented by brewer's yeast. The four primary starch degrading enzymes during malting are α-and β-amylase, limit dextrinase, and α-glucosidase. The α-and β-amylases both work on α-glucans and maltooligosaccharides by breaking α-1,4 glucosidic bonds releasing a variety of maltooligosaccharides such as maltose (2 glucose moieties) through maltoheptaose (7 glucose moieties). Both α-amylase and α-glucosidase can attack the native starch granule and release maltooligosaccharides by acting as an endohydrolase, whereas β-amylase is an exohydrolase that releases maltose from maltooligosaccharides created by other diastatic enzymes and cannot breakdown the native starch granule. [10] Limit dextrinase breaks 1,6 linkages thus allowing the complete breakdown of amylopectin into glucose by other enzymes. α-Glucosidase is a maltase that can break down maltose into two glucoses as well as being able to attack larger maltooligosaccharides releasing glucose. [10] The smallest maltooligosaccharides that β-amylase can effectively use as a substrate is maltotetraose, thus, it cannot break down maltotriose. [11] α-Amylase metabolizes maltotriose but at an extremely slow rate while α-glucosidase is capable of degrading maltotriose but is extremely thermolabile and would thus be unlikely to survive mashing temperatures. [12][13][14] Mashing is the brewing step where ground malted barley is incubated in hot water allowing for the enzymatic breakdown of macromolecules into their component parts creating a sweet wort that will be utilized by brewer's yeast during fermentation. Maltotriose accumulates during malting and mashing, which can be problematic downstream because not all brewer's yeast can metabolize maltotriose. [15,16] Inability to fully utilize all the sugars present can lead to reduced fermentation efficiency, loss of raw materials (barley starch), and flavor and stability issues. Some types of yeast can utilize maltotriose more efficiently than others based at least in part on their ability to transport maltotriose across their cells with some strains having genes encoding for transporters with more affinity for maltotriose than maltose. [16,17] Fermentation performance is postulated to be tied to the ability of the yeast to transport maltose and maltotriose. [17] Typically, brewer's yeast prefers to metabolize monomeric sugars such as glucose via facilitated diffusion and requires 40-50% of glucose to be expended before activation of transporters that move disaccharides such as maltose and sucrose. [18][19][20] Larger α-glucosides, like maltose and maltotriose, can be actively transported into yeast but not uniformly across yeast strains as ale strains are better able to metabolize maltotriose compared to lager strains. [19] Additionally, maltotriose can inhibit the uptake of the most prevalent sugar found in wort, maltose, and maltotriose is typically the most abundant sugar in the later stages of fermentation further complicating matters. [21] This inability and inefficiency of the yeast to utilize maltotriose can lead to this sugar remaining in wort after fermentation. [22] One way that breweries can circumvent the problem with maltotriose is to use specific yeast strains capable of efficiently fermenting maltotriose, but for some breweries that is not an option due to proprietary yeast strains needed for specific taste profiles. Thus, altering the sugar profile by enzymatically decreasing the amount of maltotriose in the wort seems to be a viable strategy to maximize fermentation efficiency without compromising taste profiles.
Very little is known about barley D-enzyme expression, accumulation, enzyme activity, and function but some work has been conducted in rice and wheat. [1,2,9] In potato, D-enzyme mRNA is found in many tissues but is most prominently found in developing and mature tubers. [6] In wheat, D-enzyme increases during grain development and germination. [2] In contrast, rice D-enzyme is expressed early in grain development (3-5 days after flowering) with transcript levels declining only a few days later. [1,9] Barley DPE1 and DPE2 gene expression was shown in developing grains with DPE1 transcript levels peaking later than in rice (~10 days after flowering) and DPE2 levels relatively constant throughout grain development early with expression levels much lower levels than DPE1. [23] The primary objectives of this study were to (1) determine if DPE1 is expressed during malting and if so, determine the temporal expression pattern and (2) determine if there is natural variation in the expression of D-enzyme in commercial barley varieties.

DNA isolation and sequencing
High molecular weight DNA from a spring 6-row malting cultivar, Legacy, was extracted from one-week old leaves following a standard CTAB DNA extraction method. The GenomeWalker Universal Kit (Takara Bio USA, Inc) was used to sequence the DPE1 gene following the manufacturer's instructions with primers used listed in Supplemental Table 1. Amplified bands using the gene-specific primers were cloned using TOPO-TA Cloning (Invitrogen) following manufacturer's protocol. Sequencing reactions were performed using BigDye® Terminator v3.1 Cycle Sequencing Kit following manufacturer's protocol (Applied Biosystems) and sequenced by the DNA Sequencing Facility at the University of Wisconsin Biotechnology Center. The DPE1 contigs were built using the Lasergene Molecular Biology suite (DNAstar) and the assembled DPE1 sequence was submitted to GenBank (Accession: MW553786).

Seed varieties, malting, and sample collection
Developing grain samples were collected from the malting cultivar Legacy grown under greenhouse conditions in Madison, WI. Anthesis was defined as the emergence of the awn from the boot. Samples were harvested from the middle of the spike. Micromalting samples were generated from the two-row malting cultivar Conrad and the six-row malting cultivar Legacy grown and collected from four environments (Aberdeen, ID; Morris, MN; Crookston, MN; Fargo, ND) over four crop years (2012-2015) with each replication representing a specific crop year and location. [24] Micromalting samples were generated from a subset of American Malting Barley Association, Inc's (AMBA's) recommended malting cultivars and included five spring 2-row varieties (AAC Synergy, AC Metcalfe, CDC Copeland, Conrad, and Hockett), three spring 6-row varieties (Innovation, Lacey, and Tradition), three winter 2-row varieties (Charles, Endeavor, and Wintmalt), and one winter 6-row (Thoroughbred). AMBA micromalting samples were created using a pooled method whereby five 100 g samples originating from similar but different barley growing regions were combined, homogenized, and divided into three micromalting replications. [25] All samples were micromalted by the Cereal Crops Research Unit using their traditional malting system (Madison, WI; https://www.ars.usda.gov/ARSUserFiles/50900500/barleyreports/CCRU%20Malting%20Systems.pdf). The malting procedure begins with imbibition of barley grains (i.e., steeping) in a tank with alternating four-hour water immersion at 16 °C and four-hour air rest at 18 °C over a 36-hour period ending with a target moisture of 45%. After steeping, imbibed barley seeds were transferred to germination cans, placed in germinators, and incubated at 17 °C (>98% humidity) for five days with regular 3 min rotation every 30 min. Sampling of malting barley occurred at 6 time points: 0 Days of Germination (0 DoG, aka out of steep) with subsequent sampling occurring every day (1)(2)(3)(4)(5). Samples were immediately flash frozen in liquid N 2 , and stored at −80 °C.

RNA isolations and RT-qPCR analysis
Total RNA was isolated using the PureLink TM Plant RNA Reagent (Invitrogen) from approximately 100 mg of ground seed tissue. Total RNA isolation was carried out according to the manufacturer's protocol except 700 µL of PureLink TM Plant RNA reagent was used. Further purification was accomplished using Qiagen RNeasy columns with an on-column DNase I treatment (Qiagen). RNA concentration and integrity were determined using a Nanodrop Spectrophotometer (ThermoScientific) and an Agilent BioAnalyzer, respectively. The iScript Advanced cDNA Synthesis kit (BioRad) was used to synthesize cDNA from one microgram of total RNA from each of the samples according to the manufacturer's protocol. Two technical replications of each synthesis were performed on each sample. Quantitative PCR (qPCR) reactions were performed using the QuantStudio 6 Flex Real-Time PCR System in 96-well Micro-Amp optical reaction plates with optical adhesive film (Applied Biosystems). Reactions were carried out in duplicate technical replications using SYBR ® Premix Ex Taq TM according to the manufacturer's instructions for a total of four reactions per RNA (Takara Bio USA, Inc). The relative expression ratio (RER) was calculated using the 2 -ΔΔCt formula. Reference gene primer sequences were obtained from previous research. [24] DPE1 gene-specific primer sequences were created with the Lasergene Molecular Biology Suite (DNAstar; Madison, WI) using GenBank# MW553786, HORVU2Hr1G028940.14 and HORVU2Hr 1G021700 sequences.

Gene expression data analysis
Dunnett's test was used to analyze gene expression data using SAS version 9.4 with the analysis considered significant at the P < 0.05 level. Student's T-test (2-tail, alpha = 0.05, Two-sample equal variance) was conducted using Microsoft Excel ® . Fisher's LSD test performed using SAS 9.4 with analysis considered significant at the P < 0.05 level.

Protein extraction and immunoblot analysis
Protein extraction, quantification, SDS-PAGE, and immunoblots were performed on developing grain and micromalting samples and were performed as previously described. [26] Electrophoresis parameters were adjusted slightly for the micromalting samples and were performed at 68 V (constant) for 10 min followed by 30 min at 200 V (constant). Developing grain samples (Legacy) and micromalting samples (Conrad and Legacy) were loaded with 6 µg of protein per lane. Polyclonal antibodies were raised against two short peptide sequences (GRKSGEDGSPYSGQDANC and PATQKGNWRWRIPSC) as recommended by Anaspec (Fremont, CA) using the predicted amino acid sequence of DPE1 (GenBank# BAK07657 derived from accession AK376462). The first peptide aligns to amino acids 131-148 and the second peptide aligns to amino acids 544-558. Neither peptide aligned to DPE2 amino acid sequence from transcript 2 (Blastp, HORVU2Hr1G021700). Immunoblots were visualized on a UVP AutoChemi TM System.

Sequencing of DPE1
Genomic DNA sequencing was employed to identify non-coding sequences of DPE1 for designing accurate  Figure 2). Numerous attempts were made to extend the genomic DNA sequence though the entire DPE1 gene into the promoter region to no avail. The extremely repetitive nature of the barley genome has been reported with over 80% of the genome consisting of transposable elements. [27] In fact, the 5′ region directly upstream of the 4680 bp DPE1 sequence is highly repetitive and several attempts to sequence this region resulted in amplification of numerous other genomic regions. The 4680 bp sequence we generated for DPE1 aligns with the HORVU2Hr1G028940 genomic DNA sequence with 99% identity. Interestingly, this region appears to have been recalcitrant to assembly by the International Barley Sequencing Consortium because the ISBCv2 genome assembly contains a long list of Ns in the middle of the first exon of DPE1. When these N's are removed there is only a 28 bp difference between HORVU2Hr1G02890 and our sequence (GenBank #MW553786) at the 5′ region (Supplemental Figure 2). Additionally, the HORVU2Hr1G028940 sequence contained an additional 67 bp at the 3′ end, which would not have been found in this study because we only extended the sequence from the 3′ end to the 5′ end (Supplemental Figure 2). The barley draft genome viewer, Barlex (https:// apex.ipk-gatersleben.de/apex/f?p=284:10::::::), contains the newest genome assembly build (Morex v3 gene models) and searching this build revealed five 4-α-glucanotransferases. The five new gene builds were compared to the DPE1 GenBank sequence #MW553786 and the DPE2 gene stable ID HORVU2Hr1GO21700 (Table 1) (Table 1).

DPE1 gene expression during caryopsis development
DPE1 gene expression during development was determined using gene specific RT-qPCR primers using cDNA created from RNA isolated from developing barley caryopsis of the malting cultivar Legacy (Figure 1). The developing grain time-course spanned the three phases of barley grain  development with the first time point in the pre-storage phase (5 DAA), the next four time-points in the storage phase (9-21 DAA), and the last three time-points in the desiccation phase (25-33 DAA). [28] DPE1 transcript levels stay relatively stable between 5 and 9 DAA but begin to significantly increase thereafter with levels peaking at 17 DAA, which was 5.2 times higher than 5 DAA. ( Figure 1A). In contrast, rice DPE1 expression peaks at 3 days after flowering and declines rapidly thereafter. [1,9] DPE1 transcript levels steadily decline after peaking at 17 DAA (Figure 1). Radchuk et al. [23] also observed DPE1 expression that exhibited a unimodal pattern except they observed the peak earlier, at 10 days after flowering. Similarly, Collins et al. (2021) [29] observed the highest DPE1 levels at 10 DAA in a combined endosperm/embryo fraction but also reported high expression at 14 and 18 DAA especially in the embryo, albeit about half as much as the peak, with transcript levels declining thereafter. When comparing DPE1 expression levels to highly expressed barley endosperm genes such as Bmy1 and Hor2, DPE1 transcript levels were higher at 5 DAA with levels 25 times higher than Hor2 and 274 times higher than Bmy1 ( Figure 1B). However, during peak DPE1 expression, levels of Bmy1 and Hor2 mRNA were significantly higher than DPE1 (249 and 4327-times higher, respectively) ( Figure 1C). Data mining of HORVU2Hr1G028940 gene expression using Expression Atlas (https://www.ebi.ac.uk/ gxa/home) corroborated the DPE1 expression trend observed in this study (Figure 1). DPE1 TPM levels increased 40% between 5-and 15-days post anthesis according to Expression Atlas experiment E-MTAB-2809. [30] D-enzyme is implicated in starch synthesis as well as starch degradation so high DPE1 expression levels during early grain development could be explained by D-enzymes role in starch synthesis. [9] During grain development, anti-DPE1 antibodies reacted with three proteins of different MWs although the lower MW band disappeared by 27 DAA (Figure 2, Supplemental  Figure 1). The higher MW doublet band was present throughout grain development but when the band intensity increased, as development proceeded, the lower MW band appeared less distinct and was difficult to distinguish between the top band in the unmalted sample ( Figure 2). The bottom band is slightly larger than 50 kDa, which coincides with the predicted molecular weight of the HORVU2Hr1G028940.14 protein (55.4 kDa, Supplemental Figure 2). Interestingly, the top two bands appear to be around 60 kDa or larger, which corresponds with the predicted molecular weight of the HORVU.MOREX. r3.2HG0122840 protein (64.3 kDa) from the 2021 barley gene build. Wheat DPE1 is localized to developing grains amyloplasts and contains a 45 amino acid transit peptide. [2] Similarly, Arabidopsis DPE1 is plastid-localized, and mutants exhibit an increase in leaf maltotriose and reduced leaf starch degradation. [3] TargetP-2.0, an algorithm that predicts N-terminal presequences, predicted barley DPE1 (HORVU. MOREX.r3.2HG0122840) to localize to the thylakoid lumen with a predicted cut site between amino acids 46 and 47. [31] Two additional transit peptide sequence prediction tools were used, Predator and DeepLoc-2.0, both predicting localization to chloroplast and plastid, respectively. [32,33] The predicted MW of barley DPE1 with the 46 amino acids removed to simulate processing was 59.3 kDa. The polyclonal antibodies raised against two DPE1-specific peptides react with two proteins around the predicted molecular weight of the full gene (64.3 kDa) and the processed protein (59.3 kDa) ( Figure 2, Supplemental Figure 1).
The three proteins detected by the anti-DPE1 polyclonal antibodies increased at 17 DAA corroborating the transcript data (Figure 1). These proteins were observed at low levels during early grain development with increased levels observed between 17 and 25 DAA (Figure 2). Between 29 and 33 DAA the putative DPE1 proteins appear more pronounced than earlier in grain development and are stored in the mature grain indicated by the continued presence at 35 DAA and in unmalted grain ( Figure 2, Supplemental Figure 1). Similarly, in wheat, DPE1 protein accumulates in the endosperm during grain development with levels increasing as development progresses until 25 days post anthesis. [2] DPE1 gene expression during malting DPE1 gene expression during malting was tested in both a two-row spring malting cultivar, Conrad, and a six-row spring malting cultivar, Legacy ( Figure 3A and B, Supplemental Figures 3 and 4). Legacy DPE1 gene expression increased 4-fold during malting with transcript levels highest at 4 DoG. When DPE1 transcript levels were calibrated to 0 DoG every other time point was significantly higher. Also, DPE1 transcript levels at 2-5 DoG are significantly higher than 1 DoG. DPE1 mRNA levels at 4 and 5 DoG are significantly higher than 2 or 3 DoG but are not different between them (Supplemental Figure 4). Conrad exhibits almost the same trend as Legacy except with some statistical differences. DPE1 transcript levels at 3 through 5 DoG are not significantly different than 2 DoG despite levels peaking at 4 DoG there were not any significant differences between 2 through 5 DoG (Supplemental Figure 3). Gene expression levels of DPE1 between the six-row malting cultivar Legacy and the two-row malting cultivar Conrad were tested ( Figure 4A). Legacy had significantly lower DPE1 transcript levels at all days of germination. At both their peak expression, 4 DoG, Conrad had 1.4 times more DPE1 transcript ( Figure 4A).
To further test if there are varietal differences, a set of 12 malting barleys were malted and DPE1 expression levels were compared at 5 DoG. Five DoG was chosen because there were no significant differences between 4, the observed peak, and 5 DoG (Figure 3). The 12 barley varieties were chosen to represent all the types of malting barley recommended to grow in the US: spring and winter 2-row; spring and winter 6-row. When calibrated to Conrad there were five significance groupings identified using Fisher's LSD test ( Figure 4B). The highest expression was found in the 2-row winter cultivar Endeavor but all the highest expressed cultivars from Conrad (calibrated to 1.0) to Endeavor were in the same group denoted with an a ( Figure 4B). The second group consisted of Hockett, Conrad, and Lacey. Conrad had 1.13-fold more DPE1 transcript than Lacey, while Conrad had 1.16-times more transcript than Tradition, which was significantly different. Lacey, Tradition, CDC Copeland, Charles, and AAC Synergy were in their own significance group, which ranged from 0.78 to 0.88 when compared to Conrad. Innovation had the lowest DPE1 expression (0.71) of the twelve but not as low as Legacy from Figure 4A (0.68). Recently, the transcriptome of malting barley was published and the expression of both D-enzyme genes, DPE1 and DPE2, was determined from that dataset ( Figure 3C). The DPE1 gene  [26] tPm = transcripts per million. asterisks represents significance at the p < 0.05 using the Dunnett's test. (HORVU2Hr1G028940) expression values corroborated the RT-qPCR values reported here showing a significant increase in gene expression during malting (Figure 3). The second D-enzyme gene, DPE2, was not highly expressed with TPM values not exceeding 2.5 while slightly increasing early in malting. During germination, DPE1 was expressed at low levels in the embryo after 4 days (TPM = 4), while DPE2 was expressed more robustly with levels as high as 100 TPM by the end of germination. [34] Three additional datasets were searched for DPE1 expression using Expression Atlas experiments E-GEOD-66024, E-MTAB-2809, and E-MTAB-2764. All three experiments reported TPM values for germinating embryos and all experiments showed DPE1 expressed. Experiment E-GEOD-66024, and the two other experiments, [30,35] reported similar DPE1 expression increases in germinating embryos between 24 and 48 h. Malting transcript levels of DPE1 were higher than reported during germination but malting and germination, despite their similarities, exhibit differences in enzyme activity, fermentable sugar production, and protein accumulation so this difference was not unexpected. [36] Proteins reacting to anti-DPE1 antibodies were present throughout malting in both Conrad and Legacy ( Figure 5, Supplemental Figure 5). The doublet band was observed at the beginning of malting and transitioned into a smaller MW singlet band by the end of malting ( Figure 5). This phenomenon was also observed in unmalted mature grain and malted grain although the doublet is less pronounced in the mature grain compared to 35 DAA and 0 DoG ( Figure 2). Interestingly, the large MW band coincides with the unprocessed HORVU.MOREX.r3.2HG0122840 protein (64.3 kDa) and the lower MW band could be the processed protein with transit peptide sequence removed (59.3 kDa) or the HORVU2Hr1G028940.14 protein (55.5 kDa). The wheat DPE1 protein is localized to the amyloplast but immunoblots only detected one protein. [2] Further inquiry into the origins of these possible isoforms is warranted to determine if these are derived from splice variants, transit peptide removal or other proteolytic processing.

Summary
This is the first report of the presence of disproportionating enzyme in malting barley. Barley DPE1 mRNA and proteins that reacted with anti-DPE1 polyclonal antibodies increased during grain development with the protein stored in the mature grain ( Figure 2). Additionally, there were significant differences in DPE1 transcript abundance amongst 13 elite malting cultivars (Figure 4). Further research is needed to confirm disproportionating activity of the barley DPE1 and subsequently to determine its thermostability and, thus, its ability to withstand mashing temperatures. If this is confirmed, the observed variation in DPE1 mRNA expression amongst elite malting cultivars suggests a plausible strategy to generate variations in maltooligosaccharide profiles that could have a bearing on the fermentation process.