TY - JOUR
AU - Meyer, Knut
AB - Abstract Kinetically improved diacylglycerol acyltransferase (DGAT) variants were created to favorably alter carbon partitioning in soybean (Glycine max) seeds. Initially, variants of a type 1 DGAT from a high-oil, high-oleic acid plant seed, Corylus americana, were screened for high oil content in Saccharomyces cerevisiae. Nearly all DGAT variants examined from high-oil strains had increased affinity for oleoyl-CoA, with S 0.5 values decreased as much as 4.7-fold compared with the wild-type value of 0.94 µm. Improved soybean DGAT variants were then designed to include amino acid substitutions observed in promising C. americana DGAT variants. The expression of soybean and C. americana DGAT variants in soybean somatic embryos resulted in oil contents as high as 10% and 12%, respectively, compared with only 5% and 7.6% oil achieved by overexpressing the corresponding wild-type DGATs. The affinity for oleoyl-CoA correlated strongly with oil content. The soybean DGAT variant that gave the greatest oil increase contained 14 amino acid substitutions out of a total of 504 (97% sequence identity with native). Seed-preferred expression of this soybean DGAT1 variant increased oil content of soybean seeds by an average of 3% (16% relative increase) in highly replicated, single-location field trials. The DGAT transgenes significantly reduced the soluble carbohydrate content of mature seeds and increased the seed protein content of some events. This study demonstrated that engineering of the native DGAT enzyme is an effective strategy to improve the oil content and value of soybeans. Soybeans (Glycine max) contribute a major part of the world supply of vegetable oil and protein meal. For the 2014/2015 growing season, 48 million metric tons of soybean oil and 204 million metric tons of soybean meal were produced, accounting for about 27% and 68% of world plant oil and plant protein meal production, respectively (www.fas.usda.gov/data/oilseeds). The oil is used primarily for food, with some also used for biodiesel, and the meal is used and valued for livestock feed because of its high protein content. Achieving a substantial increase in either the oil or protein content of soybeans, without any reduction in the other component or in yield, would significantly increase the value of this important crop. Conceptually, an increase in the compositional value of soybeans via altered carbon partitioning seems feasible, because a considerable portion of the soybean dry weight is composed of less valuable fiber and the poorly digestible oligosaccharides raffinose and stachyose (Padgette et al., 1996). However, neither conventional breeding nor quantitative trait locus mapping approaches for increasing soybean oil, while maintaining protein content, has been very successful, because increases in one of these components is usually associated with decreases in the other (Smith and Weber, 1968; Burton and Brim, 1981; Feng et al., 2004; Eskandari et al., 2013; Sonah et al., 2015). Consequently, the oil and protein contents of about 20% and 40% for commodity soybeans have changed little in the last several decades. Diacylglycerol acyltransferase (DGAT) catalyzes the final step in triacylglycerol (TAG) or oil production and transfers an acyl group from acyl-CoA to diacylglycerol (DAG) to form the TAG product. Four very different types of DGAT genes have been characterized (Liu et al., 2012). The type 1 DGAT genes encode polypeptides of about 500 amino acids that form tetramers (Cheng et al., 2001; Weselake et al., 2006) and are predicted by bioinformatics methods to have eight to 10 transmembrane domains, although biochemical studies have suggested as few as three (McFie et al., 2010). The type 2 DGAT genes encode polypeptides of about 300 to 350 amino acids that have almost no sequence identity with type 1 DGATs (Lardizabal et al., 2001) and are predicted to have only two transmembrane domains. The type 2 DGATs play a major role in plant species with unusual fatty acid compositions in their seed oil, such as ricinoleic acid in the castor plant Ricinus communis (Kroon et al., 2006; Burgal et al., 2008), α-eleostearic acid in the tung tree Vernicia fordii (Shockey et al., 2006), and vernolic acid in Vernonia galamensis (Li et al., 2010a). The type 1 and 2 DGATs are localized in different subdomains of the endoplasmic reticulum (Shockey et al., 2006). In addition to the type 1 and 2 DGATs, which are integral membrane proteins, a soluble DGAT from peanut (Arachis hypogaea) also has been identified (Saha et al., 2006). Aside from these three DGAT types, bifunctional wax ester synthases also can catalyze the DGAT reaction (Li et al., 2008). Depending on the species, DGAT enzymes can have strong preferences for different acyl-CoA or DAG substrates (Cao and Huang, 1986, 1987; Griffiths and Harwood, 1991; Mañas-Fernández et al., 2009; Furmanek et al., 2014). Increasing oil content in plant seeds by overexpression of DGAT genes was first demonstrated in Arabidopsis (Arabidopsis thaliana; Jako et al., 2001) and subsequently confirmed by numerous controlled environment studies, mostly with Arabidopsis. Considering only field studies with crop plants, DGAT overexpression has resulted in statistically significant seed oil increases in canola or rapeseed (Brassica napus; Weselake et al., 2008; Xu et al., 2008; Taylor et al., 2009), maize (Zea mays; Zheng et al., 2008; Oakes et al., 2011), and soybean (Lardizabal et al., 2008). In the soybean example, an oil increase of 1.5 percentage points was achieved by overexpressing a type 2 DGAT from the fungus Umbelopsis ramanniana. Small but statistically significant protein decreases also were observed in three of five field seasons, however, suggesting that the lower-protein issue that soybean breeders have faced while increasing soybean oil content also might be present with a DGAT overexpression strategy. Overexpression of a type 1 DGAT from Sesamum indicum in soybeans in a controlled environment study also resulted in mean oil increases of 1.75 and 1.39 percentage points, respectively, in T2 and T3 transgenic soybean lines (Wang et al., 2014), but the effects on protein content were not reported. Despite these successful examples of DGAT overexpression, some important questions still remain about what the most effective DGAT strategy might be for increasing soybean oil content. Is it better to use wild-type DGAT amino acid sequences, or would creating DGAT variants with improved kinetic properties be a better approach? High-throughput methods to screen mutagenized libraries have been demonstrated (Siloto et al., 2009), suggesting that DGAT engineering may be a viable option. If a wild-type DGAT is used, then would a DGAT from a species with high oil and high oleic acid (18:1; i.e. 18 carbons and one double bond in the fatty acid) content be more effective than a native soybean DGAT? High 18:1 soybeans have been developed for their human health benefits and industrial advantages (Clemente and Cahoon, 2009; Gillingham et al., 2011; Brink et al., 2014), and any DGAT genes to be overexpressed ideally would work well with 18:1 substrates. Yet, the wild-type soybean DGATs may not be optimal kinetically with 18:1 substrates, because soybean DGAT activity was greater with 18:2-CoA or 16:0-CoA than with 18:1-CoA (Cao and Huang, 1986). If the DGAT engineering approach is chosen, then what magnitude of change in which kinetic parameters would be needed to optimally improve soybean composition? To begin to answer these questions, we searched the literature for plant tissues with the best combination of high oil content and high 18:1 composition. Hazelnuts have an oil content of 60%, an extremely high 18:1 content of 79%, and an enrichment of 18:1 at the sn-3 position of TAG (Parcerisa et al., 1995), suggesting that this tissue may possess DGAT genes with useful properties. The American hazelnut shrub Corylus americana is native to North America and grows as far north as Canada, implying a short growing season and a fast rate of oil synthesis. C. americana is not cultivated extensively as a crop and, consequently, has received less breeding selection pressure than the widely cultivated common hazelnut Corylus avellana, which is native to Europe and western Asia. In this study, a type 1 DGAT complementary DNA (cDNA) from C. americana was isolated, improved variants were created, and then the corresponding amino acid substitutions were made in a soybean type 1 DGAT (Glyma.17G053300). Effects on soybean oil content and composition were then determined following expression in soybean somatic embryos of either the wild type or the engineered variants of each DGAT. The best soybean DGAT variant was then advanced to field trials to determine the effect on soybean seed composition. RESULTS Isolation of CaDGAT1 cDNA and Comparison with Other Plant Type 1 DGAT Amino Acid Sequences A type 1 DGAT cDNA from developing C. americana seeds was obtained by a PCR method and named CaDGAT1. This cDNA encoded a polypeptide of 539 amino acids with a calculated molecular mass of 61.7 kD and a pI of 8.93. Amino acid sequence identities and an alignment of CaDGAT1 with 20 plant type 1 DGAT sequences are presented in Table I and Supplemental Figure S1, respectively. CaDGAT1 had overall amino acid sequence identities ranging from 58% with the sorghum (Sorghum bicolor) and rice DGATs to 74% and 73%, respectively, with the two closely related soybean type 1 DGATs GmDGAT1a and GmDGAT1b (Table I). If the less conserved N-terminal regions evident in Supplemental Figure S1 were excluded and only the regions corresponding to CaDGAT1 from Lys-115 to the C terminus (Glu-539) were examined, then much higher sequence identities ranging from 68% to 86% were observed. Drastic differences in DGAT function could not be easily predicted from an examination of these sequence identities or alignments. For example, the type 1 DGAT from Euonymus alatus has a very different substrate specificity from the other DGATs, attaching an acetyl group rather than a long-chain fatty acyl group to the third position of TAG (Durrett et al., 2010), and yet this DGAT still had 70% sequence identity with CaDGAT1 and 84% identity if excluding the N-terminal region, which was identical to the 84% identity obtained when comparing CaDGAT1 with the DGAT of the other extremely high-oleic acid species of Table I, olive (Olea europaea). This inability to predict such large differences in DGAT substrate specificity from sequence identities or alignments suggested that a rational, site-directed engineering approach to improve DGAT properties would be difficult. Therefore, the alternative approach of screening libraries of variants containing numerous amino acid substitutions in diverse combinations appeared to be a more logical approach for DGAT improvement. Plant type 1 DGAT amino acid sequence identities compared with CaDGAT1 Table I. Plant type 1 DGAT amino acid sequence identities compared with CaDGAT1 The Needleman-Wunsch algorithm was used, with a gap creation penalty of 8 and a gap extension penalty of 2. DGAT . Plant Species . Percentage Identity with CaDGAT1, Full Sequence . Percentage Identity with CaDGAT1, from Lys-115 to the C Terminus . GenBank No. . CaDGAT1 Corylus americana 100 100 KU744408 GmDGAT1a Glycine max 74 86 AAS78662.1 GmDGAT1b Glycine max 73 84 BAE93461.1 LjDGAT1 Lotus japonicus 73 86 AAW51456.1 OeDGAT1 Olea europaea 72 84 AAS01606 MtDGAT1 Medicago truncatula 71 82 ABN09107.1 PfDGAT1 Perilla frutescens 71 82 AAG23696.1 RcDGAT1 Ricinus communis 71 83 XP_002514132.1 VvDGAT1 Vitis vinifera 71 86 XP_002279345.1 EaDGAT1 Euonymus alatus 70 84 AAV31083.1 VgDGAT1 Vernonia galamensis 70 82 ABV21945.1 JcDGAT1 Jatropha curcas 69 84 ABB84383.1 TmDGAT1 Tropaeolum majus 69 81 AAM03340.2 VfDGAT1 Vernicia fordii 69 83 DQ356680.1 NtDGAT1 Nicotiana tabacum 68 80 AAF19345.1 AtDGAT1 Arabidopsis thaliana 67 79 NP_179535.1 EpDGAT1 Echium pitardii 65 80 ACO55635.1 BnDGAT1a Brassica napus 64 77 AAD45536.1 ZmDGAT1-2 Zea mays 61 73 EU039830 OsDGAT1 Oryza sativa 58 68 NP_001054869.2 SbDGAT1b Sorghum bicolor 58 70 XP_002439419.1 DGAT . Plant Species . Percentage Identity with CaDGAT1, Full Sequence . Percentage Identity with CaDGAT1, from Lys-115 to the C Terminus . GenBank No. . CaDGAT1 Corylus americana 100 100 KU744408 GmDGAT1a Glycine max 74 86 AAS78662.1 GmDGAT1b Glycine max 73 84 BAE93461.1 LjDGAT1 Lotus japonicus 73 86 AAW51456.1 OeDGAT1 Olea europaea 72 84 AAS01606 MtDGAT1 Medicago truncatula 71 82 ABN09107.1 PfDGAT1 Perilla frutescens 71 82 AAG23696.1 RcDGAT1 Ricinus communis 71 83 XP_002514132.1 VvDGAT1 Vitis vinifera 71 86 XP_002279345.1 EaDGAT1 Euonymus alatus 70 84 AAV31083.1 VgDGAT1 Vernonia galamensis 70 82 ABV21945.1 JcDGAT1 Jatropha curcas 69 84 ABB84383.1 TmDGAT1 Tropaeolum majus 69 81 AAM03340.2 VfDGAT1 Vernicia fordii 69 83 DQ356680.1 NtDGAT1 Nicotiana tabacum 68 80 AAF19345.1 AtDGAT1 Arabidopsis thaliana 67 79 NP_179535.1 EpDGAT1 Echium pitardii 65 80 ACO55635.1 BnDGAT1a Brassica napus 64 77 AAD45536.1 ZmDGAT1-2 Zea mays 61 73 EU039830 OsDGAT1 Oryza sativa 58 68 NP_001054869.2 SbDGAT1b Sorghum bicolor 58 70 XP_002439419.1 Open in new tab Table I. Plant type 1 DGAT amino acid sequence identities compared with CaDGAT1 The Needleman-Wunsch algorithm was used, with a gap creation penalty of 8 and a gap extension penalty of 2. DGAT . Plant Species . Percentage Identity with CaDGAT1, Full Sequence . Percentage Identity with CaDGAT1, from Lys-115 to the C Terminus . GenBank No. . CaDGAT1 Corylus americana 100 100 KU744408 GmDGAT1a Glycine max 74 86 AAS78662.1 GmDGAT1b Glycine max 73 84 BAE93461.1 LjDGAT1 Lotus japonicus 73 86 AAW51456.1 OeDGAT1 Olea europaea 72 84 AAS01606 MtDGAT1 Medicago truncatula 71 82 ABN09107.1 PfDGAT1 Perilla frutescens 71 82 AAG23696.1 RcDGAT1 Ricinus communis 71 83 XP_002514132.1 VvDGAT1 Vitis vinifera 71 86 XP_002279345.1 EaDGAT1 Euonymus alatus 70 84 AAV31083.1 VgDGAT1 Vernonia galamensis 70 82 ABV21945.1 JcDGAT1 Jatropha curcas 69 84 ABB84383.1 TmDGAT1 Tropaeolum majus 69 81 AAM03340.2 VfDGAT1 Vernicia fordii 69 83 DQ356680.1 NtDGAT1 Nicotiana tabacum 68 80 AAF19345.1 AtDGAT1 Arabidopsis thaliana 67 79 NP_179535.1 EpDGAT1 Echium pitardii 65 80 ACO55635.1 BnDGAT1a Brassica napus 64 77 AAD45536.1 ZmDGAT1-2 Zea mays 61 73 EU039830 OsDGAT1 Oryza sativa 58 68 NP_001054869.2 SbDGAT1b Sorghum bicolor 58 70 XP_002439419.1 DGAT . Plant Species . Percentage Identity with CaDGAT1, Full Sequence . Percentage Identity with CaDGAT1, from Lys-115 to the C Terminus . GenBank No. . CaDGAT1 Corylus americana 100 100 KU744408 GmDGAT1a Glycine max 74 86 AAS78662.1 GmDGAT1b Glycine max 73 84 BAE93461.1 LjDGAT1 Lotus japonicus 73 86 AAW51456.1 OeDGAT1 Olea europaea 72 84 AAS01606 MtDGAT1 Medicago truncatula 71 82 ABN09107.1 PfDGAT1 Perilla frutescens 71 82 AAG23696.1 RcDGAT1 Ricinus communis 71 83 XP_002514132.1 VvDGAT1 Vitis vinifera 71 86 XP_002279345.1 EaDGAT1 Euonymus alatus 70 84 AAV31083.1 VgDGAT1 Vernonia galamensis 70 82 ABV21945.1 JcDGAT1 Jatropha curcas 69 84 ABB84383.1 TmDGAT1 Tropaeolum majus 69 81 AAM03340.2 VfDGAT1 Vernicia fordii 69 83 DQ356680.1 NtDGAT1 Nicotiana tabacum 68 80 AAF19345.1 AtDGAT1 Arabidopsis thaliana 67 79 NP_179535.1 EpDGAT1 Echium pitardii 65 80 ACO55635.1 BnDGAT1a Brassica napus 64 77 AAD45536.1 ZmDGAT1-2 Zea mays 61 73 EU039830 OsDGAT1 Oryza sativa 58 68 NP_001054869.2 SbDGAT1b Sorghum bicolor 58 70 XP_002439419.1 Open in new tab Creation and Screening of CaDGAT1 Variant Libraries Libraries of CaDGAT1 variants containing multiple amino acid substitutions were created using recursive sequence recombination methods, also known as DNA shuffling (Stemmer, 1994; Ness et al., 2002). Specific amino acid substitutions were planned, but additional diversity was provided from random mutations inherent in these PCR-based methods. A summary of the libraries is presented in Table II. Library A targeted 36 amino acids for substitution, with the aim toward more conservative and, presumably, lower risk substitutions. To further reduce risk, libraries B and C were created, with library B containing 17 of the 36 planned substitutions and library C containing the other 19 planned substitutions of library A. Library D was designed to target 23 less conservative amino acid substitutions compared with those of library A. Libraries E and F were designed to have 11 and 12 substitutions, respectively, of the library D planned substitutions. Considering only planned mutations, the theoretical complexity of the libraries ranged from 211 or 2,048 for library E to 236 or 69 billion for library A. Portions of the plasmid libraries were first transformed into Escherichia coli, and then plasmid preparations comprising the DNA from about 4,000 to 5,000 E. coli colonies per library were generated. The plasmid DNA was then transformed into a Saccharomyces cerevisiae (yeast) double null strain (ƊDGA1/ƊLRO1) that was deficient in both DGAT and phospholipid:diacylglycerol acyltransferase (PDAT) activities and, consequently, was deficient in TAG synthesis (Zheng et al., 2008). Yeast strains transformed with DGAT variants were screened for high oil content using the fluorescent stain Nile Red (Greenspan et al., 1985; Siloto et al., 2009), with or without a prior enrichment of low-density yeast strains, as described in “Materials and Methods.” The low-density enrichment method was expected to result in the loss of some desirable variants present in less aggressive strains during the 2- to 3-d growth periods, but the enrichment also was expected to reduce the number of strains needed for Nile Red staining. Because the amplified library sizes never exceeded 5,000, only a very small fraction of the theoretical complexity of the largest original libraries was screened in yeast. Summary of CaDGAT1 variant libraries, and approximate number of yeast strains screened by Nile Red staining, with or without prior enrichment of low-density strains Table II. Summary of CaDGAT1 variant libraries, and approximate number of yeast strains screened by Nile Red staining, with or without prior enrichment of low-density strains Parameter . Library . A B C D E F No. of planned amino acid substitutions 36 17 19 23 11 12 Theoretical size of library (planned substitutions only) 6.9 × 1010 1.3 × 105 5.2 × 105 8.4 × 106 2,048 4,096 No. of variants amplified in E. coli 4,000 5,000 5,000 5,000 5,000 5,000 No. of yeast strains screened by Nile Red staining, without prior low-density enrichment 200 200 200 600 1,300 0 No. of yeast strains enriched for low density 5,000 6,000 10,000 7,000 0 8,000 No. of yeast strains screened by Nile Red staining, following low-density enrichment 700 400 400 600 0 900 Parameter . Library . A B C D E F No. of planned amino acid substitutions 36 17 19 23 11 12 Theoretical size of library (planned substitutions only) 6.9 × 1010 1.3 × 105 5.2 × 105 8.4 × 106 2,048 4,096 No. of variants amplified in E. coli 4,000 5,000 5,000 5,000 5,000 5,000 No. of yeast strains screened by Nile Red staining, without prior low-density enrichment 200 200 200 600 1,300 0 No. of yeast strains enriched for low density 5,000 6,000 10,000 7,000 0 8,000 No. of yeast strains screened by Nile Red staining, following low-density enrichment 700 400 400 600 0 900 Open in new tab Table II. Summary of CaDGAT1 variant libraries, and approximate number of yeast strains screened by Nile Red staining, with or without prior enrichment of low-density strains Parameter . Library . A B C D E F No. of planned amino acid substitutions 36 17 19 23 11 12 Theoretical size of library (planned substitutions only) 6.9 × 1010 1.3 × 105 5.2 × 105 8.4 × 106 2,048 4,096 No. of variants amplified in E. coli 4,000 5,000 5,000 5,000 5,000 5,000 No. of yeast strains screened by Nile Red staining, without prior low-density enrichment 200 200 200 600 1,300 0 No. of yeast strains enriched for low density 5,000 6,000 10,000 7,000 0 8,000 No. of yeast strains screened by Nile Red staining, following low-density enrichment 700 400 400 600 0 900 Parameter . Library . A B C D E F No. of planned amino acid substitutions 36 17 19 23 11 12 Theoretical size of library (planned substitutions only) 6.9 × 1010 1.3 × 105 5.2 × 105 8.4 × 106 2,048 4,096 No. of variants amplified in E. coli 4,000 5,000 5,000 5,000 5,000 5,000 No. of yeast strains screened by Nile Red staining, without prior low-density enrichment 200 200 200 600 1,300 0 No. of yeast strains enriched for low density 5,000 6,000 10,000 7,000 0 8,000 No. of yeast strains screened by Nile Red staining, following low-density enrichment 700 400 400 600 0 900 Open in new tab Oil Content and DGAT Amino Acid Substitutions of Yeast Strains with Increased Fluorescence The oil content and fatty acid composition of yeast strains with greatest fluorescence from Nile Red staining were determined (Table III). Because strains expressing the CaDGAT1 variants from libraries A, B, and C were grown and analyzed separately from those expressing variants from libraries D, E, and F, the wild-type CaDGAT1 controls were grown and analyzed with each group. The two library A strains had a mean oil content of 22.5%, nearly double the CaDGAT1 mean of 12.3%, and they also had substantial decreases in 16:0 and 18:0, and increases in 16:1 and 18:1, compared with CaDGAT1 values. The library C strains had oil contents ranging from 20.5% to 17.6%, with small increases in 16:0 and small decreases in 16:1, compared with the CaDGAT1 controls. The vector control strain oil content of 3.1% was low, as expected for a DGAT/PDAT null strain. The underrepresentation of library B variants in Table III may reflect either a greater proportion of deleterious substitutions and/or a lesser proportion of beneficial substitutions in library B compared with the other libraries. The library D and E strains had mean oil contents of 17.4% and 16.7%, respectively, compared with the CaDGAT1 mean value of 10.2%. Small increases in 16:0 and 18:0, and small decreases in 16:1 content, also were observed for the library D variants. The data of Table III confirmed that Nile Red staining was effective in identifying high-oil yeast strains and that the engineered DGAT variants were more effective than the wild-type CaDGAT1 in increasing oil content in yeast. The data also showed that targeting different amino acid substitutions in different libraries resulted in different effects on fatty acid composition. Enrichment for low density was clearly not essential for obtaining high-oil strains, as evident from the multiple high-oil strains expressing variants from library E, the only library screened without density enrichment (Tables II and III). Oil content and fatty acid composition of yeast strains expressing CaDGAT1 variants Table III. Oil content and fatty acid composition of yeast strains expressing CaDGAT1 variants Single and double asterisks denote statistical significance for CaDGAT1 variants compared with CaDGAT1 in two-tailed Student’s t tests at P < 0.05 and P < 0.01, respectively. DGAT Variant . Percentage Oil . Percentage 16:0 . Percentage 16:1 . Percentage 18:0 . Percentage 18:1 . From libraries A to C  CaDGAT1-A2 24.2 22.5 41.8 6.6 29.1  CaDGAT1-A3 20.8 23 41.8 6.9 28.4  CaDGAT1-C10 20.5 32.9 30.4 12.2 24.6  CaDGAT1-C8 20.5 31.5 31 12.2 25.3  CaDGAT1-C12 19.8 31.6 31.1 12.2 25.1  CaDGAT1-C9 19.5 31.8 31.1 12.3 24.8  CaDGAT1-C11 19 31.9 31 12.5 24.7  CaDGAT1-C13 18.5 31.8 31 12.6 24.6  CaDGAT1-C3 17.9 31.8 32.1 10.7 25.5  CaDGAT1-C15 17.6 33 30.9 12.2 23.9  CaDGAT1, rep2 13 30.2 32.2 12.7 24.9  CaDGAT1, rep1 11.5 29.4 32.5 12.2 26  Vector control 3.1 18.9 39.9 13.8 27.4  CaDGAT1 mean ± se 12.3 ± 0.75 29.8 ± 0.40 32.4 ± 0.15 12.5 ± 0.25 25.5 ± 0.55  CaDGAT1 library A mean ± se 22.5 ± 1.7* 22.8 ± 0.25** 41.8 ± 0** 6.75 ± 0.15** 28.8 ± 0.35*  CaDGAT1 library C mean ± se 19.2 ± 0.39** 32.0 ± 0.20** 31.1 ± 0.17** 12.1 ± 0.21 24.8 ± 0.18 From libraries D to F  CaDGAT1-D2 19.3 32.6 31.1 12.2 23.4  CaDGAT1-E4 18.5 33.1 29 13.4 24.5  CaDGAT1-D16 18.2 32.5 31.4 12.7 23.4  CaDGAT1-E3 18 30.2 28.8 13.5 27.4  CaDGAT1-D19 17.5 32.6 31.7 12.4 23.2  CaDGAT1-D15 16.2 32.4 30.7 12.9 24  CaDGAT1-D5 15.8 30.4 27.1 14.9 27.2  CaDGAT1-E1 15.3 31.7 35.2 11.8 21.4  CaDGAT1-F8 15.2 31.8 32.3 12.5 23.3  CaDGAT1-E2 15.1 29.6 32 11.7 26.6  CaDGAT1, rep1 10.6 29.6 34.9 10.6 24.8  CaDGAT1, rep2 9.7 29.2 35.1 10.7 25  CaDGAT1 mean ± se 10.2 ± 0.45 29.4 ± 0.20 35.0 ± 0.10 10.7 ± 0.05 24.9 ± 0.10  CaDGAT1 library D mean ± se 17.4 ± 0.64** 32.1 ± 0.43* 30.4 ± 0.84* 13.0 ± 0.49* 24.2 ± 0.75  CaDGAT1 library E mean ± se 16.7 ± 0.89** 31.2 ± 0.79 31.3 ± 1.5 12.6 ± 0.49 25.0 ± 1.3 DGAT Variant . Percentage Oil . Percentage 16:0 . Percentage 16:1 . Percentage 18:0 . Percentage 18:1 . From libraries A to C  CaDGAT1-A2 24.2 22.5 41.8 6.6 29.1  CaDGAT1-A3 20.8 23 41.8 6.9 28.4  CaDGAT1-C10 20.5 32.9 30.4 12.2 24.6  CaDGAT1-C8 20.5 31.5 31 12.2 25.3  CaDGAT1-C12 19.8 31.6 31.1 12.2 25.1  CaDGAT1-C9 19.5 31.8 31.1 12.3 24.8  CaDGAT1-C11 19 31.9 31 12.5 24.7  CaDGAT1-C13 18.5 31.8 31 12.6 24.6  CaDGAT1-C3 17.9 31.8 32.1 10.7 25.5  CaDGAT1-C15 17.6 33 30.9 12.2 23.9  CaDGAT1, rep2 13 30.2 32.2 12.7 24.9  CaDGAT1, rep1 11.5 29.4 32.5 12.2 26  Vector control 3.1 18.9 39.9 13.8 27.4  CaDGAT1 mean ± se 12.3 ± 0.75 29.8 ± 0.40 32.4 ± 0.15 12.5 ± 0.25 25.5 ± 0.55  CaDGAT1 library A mean ± se 22.5 ± 1.7* 22.8 ± 0.25** 41.8 ± 0** 6.75 ± 0.15** 28.8 ± 0.35*  CaDGAT1 library C mean ± se 19.2 ± 0.39** 32.0 ± 0.20** 31.1 ± 0.17** 12.1 ± 0.21 24.8 ± 0.18 From libraries D to F  CaDGAT1-D2 19.3 32.6 31.1 12.2 23.4  CaDGAT1-E4 18.5 33.1 29 13.4 24.5  CaDGAT1-D16 18.2 32.5 31.4 12.7 23.4  CaDGAT1-E3 18 30.2 28.8 13.5 27.4  CaDGAT1-D19 17.5 32.6 31.7 12.4 23.2  CaDGAT1-D15 16.2 32.4 30.7 12.9 24  CaDGAT1-D5 15.8 30.4 27.1 14.9 27.2  CaDGAT1-E1 15.3 31.7 35.2 11.8 21.4  CaDGAT1-F8 15.2 31.8 32.3 12.5 23.3  CaDGAT1-E2 15.1 29.6 32 11.7 26.6  CaDGAT1, rep1 10.6 29.6 34.9 10.6 24.8  CaDGAT1, rep2 9.7 29.2 35.1 10.7 25  CaDGAT1 mean ± se 10.2 ± 0.45 29.4 ± 0.20 35.0 ± 0.10 10.7 ± 0.05 24.9 ± 0.10  CaDGAT1 library D mean ± se 17.4 ± 0.64** 32.1 ± 0.43* 30.4 ± 0.84* 13.0 ± 0.49* 24.2 ± 0.75  CaDGAT1 library E mean ± se 16.7 ± 0.89** 31.2 ± 0.79 31.3 ± 1.5 12.6 ± 0.49 25.0 ± 1.3 Open in new tab Table III. Oil content and fatty acid composition of yeast strains expressing CaDGAT1 variants Single and double asterisks denote statistical significance for CaDGAT1 variants compared with CaDGAT1 in two-tailed Student’s t tests at P < 0.05 and P < 0.01, respectively. DGAT Variant . Percentage Oil . Percentage 16:0 . Percentage 16:1 . Percentage 18:0 . Percentage 18:1 . From libraries A to C  CaDGAT1-A2 24.2 22.5 41.8 6.6 29.1  CaDGAT1-A3 20.8 23 41.8 6.9 28.4  CaDGAT1-C10 20.5 32.9 30.4 12.2 24.6  CaDGAT1-C8 20.5 31.5 31 12.2 25.3  CaDGAT1-C12 19.8 31.6 31.1 12.2 25.1  CaDGAT1-C9 19.5 31.8 31.1 12.3 24.8  CaDGAT1-C11 19 31.9 31 12.5 24.7  CaDGAT1-C13 18.5 31.8 31 12.6 24.6  CaDGAT1-C3 17.9 31.8 32.1 10.7 25.5  CaDGAT1-C15 17.6 33 30.9 12.2 23.9  CaDGAT1, rep2 13 30.2 32.2 12.7 24.9  CaDGAT1, rep1 11.5 29.4 32.5 12.2 26  Vector control 3.1 18.9 39.9 13.8 27.4  CaDGAT1 mean ± se 12.3 ± 0.75 29.8 ± 0.40 32.4 ± 0.15 12.5 ± 0.25 25.5 ± 0.55  CaDGAT1 library A mean ± se 22.5 ± 1.7* 22.8 ± 0.25** 41.8 ± 0** 6.75 ± 0.15** 28.8 ± 0.35*  CaDGAT1 library C mean ± se 19.2 ± 0.39** 32.0 ± 0.20** 31.1 ± 0.17** 12.1 ± 0.21 24.8 ± 0.18 From libraries D to F  CaDGAT1-D2 19.3 32.6 31.1 12.2 23.4  CaDGAT1-E4 18.5 33.1 29 13.4 24.5  CaDGAT1-D16 18.2 32.5 31.4 12.7 23.4  CaDGAT1-E3 18 30.2 28.8 13.5 27.4  CaDGAT1-D19 17.5 32.6 31.7 12.4 23.2  CaDGAT1-D15 16.2 32.4 30.7 12.9 24  CaDGAT1-D5 15.8 30.4 27.1 14.9 27.2  CaDGAT1-E1 15.3 31.7 35.2 11.8 21.4  CaDGAT1-F8 15.2 31.8 32.3 12.5 23.3  CaDGAT1-E2 15.1 29.6 32 11.7 26.6  CaDGAT1, rep1 10.6 29.6 34.9 10.6 24.8  CaDGAT1, rep2 9.7 29.2 35.1 10.7 25  CaDGAT1 mean ± se 10.2 ± 0.45 29.4 ± 0.20 35.0 ± 0.10 10.7 ± 0.05 24.9 ± 0.10  CaDGAT1 library D mean ± se 17.4 ± 0.64** 32.1 ± 0.43* 30.4 ± 0.84* 13.0 ± 0.49* 24.2 ± 0.75  CaDGAT1 library E mean ± se 16.7 ± 0.89** 31.2 ± 0.79 31.3 ± 1.5 12.6 ± 0.49 25.0 ± 1.3 DGAT Variant . Percentage Oil . Percentage 16:0 . Percentage 16:1 . Percentage 18:0 . Percentage 18:1 . From libraries A to C  CaDGAT1-A2 24.2 22.5 41.8 6.6 29.1  CaDGAT1-A3 20.8 23 41.8 6.9 28.4  CaDGAT1-C10 20.5 32.9 30.4 12.2 24.6  CaDGAT1-C8 20.5 31.5 31 12.2 25.3  CaDGAT1-C12 19.8 31.6 31.1 12.2 25.1  CaDGAT1-C9 19.5 31.8 31.1 12.3 24.8  CaDGAT1-C11 19 31.9 31 12.5 24.7  CaDGAT1-C13 18.5 31.8 31 12.6 24.6  CaDGAT1-C3 17.9 31.8 32.1 10.7 25.5  CaDGAT1-C15 17.6 33 30.9 12.2 23.9  CaDGAT1, rep2 13 30.2 32.2 12.7 24.9  CaDGAT1, rep1 11.5 29.4 32.5 12.2 26  Vector control 3.1 18.9 39.9 13.8 27.4  CaDGAT1 mean ± se 12.3 ± 0.75 29.8 ± 0.40 32.4 ± 0.15 12.5 ± 0.25 25.5 ± 0.55  CaDGAT1 library A mean ± se 22.5 ± 1.7* 22.8 ± 0.25** 41.8 ± 0** 6.75 ± 0.15** 28.8 ± 0.35*  CaDGAT1 library C mean ± se 19.2 ± 0.39** 32.0 ± 0.20** 31.1 ± 0.17** 12.1 ± 0.21 24.8 ± 0.18 From libraries D to F  CaDGAT1-D2 19.3 32.6 31.1 12.2 23.4  CaDGAT1-E4 18.5 33.1 29 13.4 24.5  CaDGAT1-D16 18.2 32.5 31.4 12.7 23.4  CaDGAT1-E3 18 30.2 28.8 13.5 27.4  CaDGAT1-D19 17.5 32.6 31.7 12.4 23.2  CaDGAT1-D15 16.2 32.4 30.7 12.9 24  CaDGAT1-D5 15.8 30.4 27.1 14.9 27.2  CaDGAT1-E1 15.3 31.7 35.2 11.8 21.4  CaDGAT1-F8 15.2 31.8 32.3 12.5 23.3  CaDGAT1-E2 15.1 29.6 32 11.7 26.6  CaDGAT1, rep1 10.6 29.6 34.9 10.6 24.8  CaDGAT1, rep2 9.7 29.2 35.1 10.7 25  CaDGAT1 mean ± se 10.2 ± 0.45 29.4 ± 0.20 35.0 ± 0.10 10.7 ± 0.05 24.9 ± 0.10  CaDGAT1 library D mean ± se 17.4 ± 0.64** 32.1 ± 0.43* 30.4 ± 0.84* 13.0 ± 0.49* 24.2 ± 0.75  CaDGAT1 library E mean ± se 16.7 ± 0.89** 31.2 ± 0.79 31.3 ± 1.5 12.6 ± 0.49 25.0 ± 1.3 Open in new tab The DGAT amino acid substitutions of these yeast strains are presented in Figure 1. Substitutions were observed at a total of 63 different positions in the DGAT amino acid sequence. For any one variant, the number of substitutions ranged from two for CaDGAT1-F8 to 17 for CaDGAT1-C11. Two kinds of substitutions were observed at positions Glu-86 (Asp or Asn) and Val-329 (Leu or Met). Unambiguous sequences were not determined for CaDGAT1-C3 and CaDGAT1-C12; therefore, they were not included in Figure 1. The CaDGAT1-A2 and CaDGAT1-A3 amino acid sequences were identical, in good agreement with the similar values for oil content and fatty acid composition observed for these two strains (Table III). Figure 1. Open in new tabDownload slide Amino acid substitutions in CaDGAT1 variants that gave high oil content in yeast. The wild-type CaDGAT1 amino acids and positions are presented in the top row. The total number of substitutions for each variant is given in the right column. Different colors denote variants from different libraries. Figure 1. Open in new tabDownload slide Amino acid substitutions in CaDGAT1 variants that gave high oil content in yeast. The wild-type CaDGAT1 amino acids and positions are presented in the top row. The total number of substitutions for each variant is given in the right column. Different colors denote variants from different libraries. Kinetic Parameters of CaDGAT1 Variants Yeast microsomal membrane preparations were used to determine kinetic parameters and DGAT protein abundance by immunoblots for nine CaDGAT1 variants that gave high oil content in yeast (Table IV). The immunoblots were probed with rabbit polyclonal antibodies prepared against the peptide NGNDGGEKIANGEDR (positions 93–107 of CaDGAT1). No amino acid substitutions that might affect signal strength on an immunoblot were present in this peptide region in the CaDGAT1 variants. Using endogenous DAG and varying the 18:1-CoA concentration, sigmoidal kinetics were observed for CaDGAT1 and the variants. Most variants had greater affinity for 18:1-CoA, with S 0.5 values decreasing as much as 4.7-fold for the CaDGAT1-C11 variant, compared with the CaDGAT1 value of 0.94 µm. After adjusting for protein abundance, V max values of the variants were less consistently changed, ranging from 55% to 323% of the CaDGAT1 value of 3,798 pmol min−1 mg−1 microsomal protein. There was a trend for more pronounced cooperativity (slightly larger Hill coefficients) in the variants compared with CaDGAT1. As evident in Table IV, most variants had decreased protein abundance in microsomes relative to CaDGAT1. The decreased abundance may be due to either decreased expression or decreased protein stability of the variants; alternatively, it may mean that proportionately more of the DGAT variant proteins were in the fat pad, rather than the microsomal membrane pellet, compared with the CaDGAT1 protein. We observed extremely large fat pads for some of the yeast strains expressing DGAT variants, consistent with the high oil contents reported in Table III. Kinetic parameters of CaDGAT1 variants that gave high oil content in yeast Table IV. Kinetic parameters of CaDGAT1 variants that gave high oil content in yeast Means ± se were determined by the SigmaPlot 12.5 enzyme kinetics module. ND, Not determined. DGAT Variant . S 0.5 . V max . DGAT Protein Abundance in Microsomes (Relative to CaDGAT1) . V max, Adjusted for Protein Abundance . . Hill Coefficient . µm 18:1-CoA pmol min−1 mg−1 microsomal protein pmol min−1 mg−1 protein % of CaDGAT1 CaDGAT1 0.94 ± 0.12 3,798 ± 284 1 3,798 100 1.8 ± 0.38 CaDGAT1-C11 0.20 ± 0.036 2,517 ± 301 0.2 12,585 331 2.6 ± 1.1 CaDGAT1-C10 0.39 ± 0.025 2,467 ± 70 0.83 2,972 78 2.5 ± 0.40 CaDGAT1-C13 0.52 ± 0.076 1,851 ± 124 0.29 6,383 168 2.3 ± 0.71 CaDGAT1-C9 0.53 ± 0.079 2,210 ± 148 0.18 12,278 323 2.1 ± 0.62 CaDGAT1-E3 0.57 ± 0.088 2,295 ± 168 0.2 11,475 302 2.5 ± 0.81 CaDGAT1-D16 0.59 ± 0.075 1,989 ± 117 0.95 2,094 55 1.8 ± 0.45 CaDGAT1-D2 0.69 ± 0.076 3,003 ± 163 1.25 2,402 63 2.1 ± 0.48 CaDGAT1-D19 0.75 ± 0.11 3,133 ± 246 0.6 5,222 137 2.5 ± 0.86 CaDGAT1-A2 1.11 ± 0.20 7,584 ± 737 0.92 8,243 217 1.6 ± 0.46 ZmDGAT1-2 2.8 ± 0.19 1,138 ± 61 ND ND ND 2.7 ± 0.44 DGAT Variant . S 0.5 . V max . DGAT Protein Abundance in Microsomes (Relative to CaDGAT1) . V max, Adjusted for Protein Abundance . . Hill Coefficient . µm 18:1-CoA pmol min−1 mg−1 microsomal protein pmol min−1 mg−1 protein % of CaDGAT1 CaDGAT1 0.94 ± 0.12 3,798 ± 284 1 3,798 100 1.8 ± 0.38 CaDGAT1-C11 0.20 ± 0.036 2,517 ± 301 0.2 12,585 331 2.6 ± 1.1 CaDGAT1-C10 0.39 ± 0.025 2,467 ± 70 0.83 2,972 78 2.5 ± 0.40 CaDGAT1-C13 0.52 ± 0.076 1,851 ± 124 0.29 6,383 168 2.3 ± 0.71 CaDGAT1-C9 0.53 ± 0.079 2,210 ± 148 0.18 12,278 323 2.1 ± 0.62 CaDGAT1-E3 0.57 ± 0.088 2,295 ± 168 0.2 11,475 302 2.5 ± 0.81 CaDGAT1-D16 0.59 ± 0.075 1,989 ± 117 0.95 2,094 55 1.8 ± 0.45 CaDGAT1-D2 0.69 ± 0.076 3,003 ± 163 1.25 2,402 63 2.1 ± 0.48 CaDGAT1-D19 0.75 ± 0.11 3,133 ± 246 0.6 5,222 137 2.5 ± 0.86 CaDGAT1-A2 1.11 ± 0.20 7,584 ± 737 0.92 8,243 217 1.6 ± 0.46 ZmDGAT1-2 2.8 ± 0.19 1,138 ± 61 ND ND ND 2.7 ± 0.44 Open in new tab Table IV. Kinetic parameters of CaDGAT1 variants that gave high oil content in yeast Means ± se were determined by the SigmaPlot 12.5 enzyme kinetics module. ND, Not determined. DGAT Variant . S 0.5 . V max . DGAT Protein Abundance in Microsomes (Relative to CaDGAT1) . V max, Adjusted for Protein Abundance . . Hill Coefficient . µm 18:1-CoA pmol min−1 mg−1 microsomal protein pmol min−1 mg−1 protein % of CaDGAT1 CaDGAT1 0.94 ± 0.12 3,798 ± 284 1 3,798 100 1.8 ± 0.38 CaDGAT1-C11 0.20 ± 0.036 2,517 ± 301 0.2 12,585 331 2.6 ± 1.1 CaDGAT1-C10 0.39 ± 0.025 2,467 ± 70 0.83 2,972 78 2.5 ± 0.40 CaDGAT1-C13 0.52 ± 0.076 1,851 ± 124 0.29 6,383 168 2.3 ± 0.71 CaDGAT1-C9 0.53 ± 0.079 2,210 ± 148 0.18 12,278 323 2.1 ± 0.62 CaDGAT1-E3 0.57 ± 0.088 2,295 ± 168 0.2 11,475 302 2.5 ± 0.81 CaDGAT1-D16 0.59 ± 0.075 1,989 ± 117 0.95 2,094 55 1.8 ± 0.45 CaDGAT1-D2 0.69 ± 0.076 3,003 ± 163 1.25 2,402 63 2.1 ± 0.48 CaDGAT1-D19 0.75 ± 0.11 3,133 ± 246 0.6 5,222 137 2.5 ± 0.86 CaDGAT1-A2 1.11 ± 0.20 7,584 ± 737 0.92 8,243 217 1.6 ± 0.46 ZmDGAT1-2 2.8 ± 0.19 1,138 ± 61 ND ND ND 2.7 ± 0.44 DGAT Variant . S 0.5 . V max . DGAT Protein Abundance in Microsomes (Relative to CaDGAT1) . V max, Adjusted for Protein Abundance . . Hill Coefficient . µm 18:1-CoA pmol min−1 mg−1 microsomal protein pmol min−1 mg−1 protein % of CaDGAT1 CaDGAT1 0.94 ± 0.12 3,798 ± 284 1 3,798 100 1.8 ± 0.38 CaDGAT1-C11 0.20 ± 0.036 2,517 ± 301 0.2 12,585 331 2.6 ± 1.1 CaDGAT1-C10 0.39 ± 0.025 2,467 ± 70 0.83 2,972 78 2.5 ± 0.40 CaDGAT1-C13 0.52 ± 0.076 1,851 ± 124 0.29 6,383 168 2.3 ± 0.71 CaDGAT1-C9 0.53 ± 0.079 2,210 ± 148 0.18 12,278 323 2.1 ± 0.62 CaDGAT1-E3 0.57 ± 0.088 2,295 ± 168 0.2 11,475 302 2.5 ± 0.81 CaDGAT1-D16 0.59 ± 0.075 1,989 ± 117 0.95 2,094 55 1.8 ± 0.45 CaDGAT1-D2 0.69 ± 0.076 3,003 ± 163 1.25 2,402 63 2.1 ± 0.48 CaDGAT1-D19 0.75 ± 0.11 3,133 ± 246 0.6 5,222 137 2.5 ± 0.86 CaDGAT1-A2 1.11 ± 0.20 7,584 ± 737 0.92 8,243 217 1.6 ± 0.46 ZmDGAT1-2 2.8 ± 0.19 1,138 ± 61 ND ND ND 2.7 ± 0.44 Open in new tab As sigmoidal kinetics had not been reported previously for plant type 1 DGAT activity, the question arose whether this kinetic behavior was unique to CaDGAT1. Therefore, we determined kinetics for another plant DGAT1, the maize high-oil DGAT allele (ZmDGAT1-2, ASK allele) reported previously (Zheng et al., 2008). ZmDGAT1-2 also had sigmoidal kinetics, with moderately more pronounced cooperativity compared with CaDGAT1 (Table IV). ZmDGAT1-2 had poorer affinity for the 18:1-CoA substrate, with an S 0.5 value about 3-fold greater than that of CaDGAT1. The ZmDGAT1-2 V max could not be compared directly with that of CaDGAT1 in this data set, as the peptide sequence used to prepare antibodies against CaDGAT1 was not conserved in the maize DGAT. The observation of sigmoidal kinetics for ZmDGAT1-2 confirmed that this kinetic behavior was not unique to CaDGAT1. In previous studies, the N-terminal fragments of rapeseed and mouse type 1 DGAT proteins were shown to bind acyl-CoA substrates in a sigmoidal manner (Weselake et al., 2006; Siloto et al., 2008), consistent with the sigmoidal kinetics for activity observed here. Oil and Fatty Acid Composition of Soybean Somatic Embryos Expressing CaDGAT1 Variants CaDGAT1 and four engineered variants with differing kinetic properties were expressed in soybean somatic embryos to determine the effects on oil and fatty acid composition (Fig. 2). Three of the four engineered variants gave oil contents of 10.1%, 11.5%, and 12%, significantly greater than the CaDGAT1 value of 7.6%. There were also significant increases in 18:0 and 18:1 contents, and decreases in 16:0, 18:2, and 18:3 contents for the three variants with increased oil contents, compared with the values for CaDGAT1. The largest change in fatty acid composition was observed for 18:1, with values of 27% to 28% for three variants, compared with 19.4% for CaDGAT1. The affinity for 18:1-CoA was strongly correlated with oil content (Fig. 3). V max values were not correlated with oil content, but we did not determine and correct for DGAT protein abundance in soybean embryos, which may have obscured meaningful correlations of oil and V max . Figure 2. Open in new tabDownload slide Oil and fatty acid composition of soybean somatic embryos expressing CaDGAT1 variants. Single and double asterisks denote statistically significant differences from CaDGAT1 values at P < 0.05 and P < 0.01, respectively, in two-tailed Student’s t tests. Error bars indicate se. Figure 2. Open in new tabDownload slide Oil and fatty acid composition of soybean somatic embryos expressing CaDGAT1 variants. Single and double asterisks denote statistically significant differences from CaDGAT1 values at P < 0.05 and P < 0.01, respectively, in two-tailed Student’s t tests. Error bars indicate se. Figure 3. Open in new tabDownload slide Correlation of DGAT affinity for 18:1-CoA with oil content in soybean somatic embryos expressing CaDGAT1 variants. Figure 3. Open in new tabDownload slide Correlation of DGAT affinity for 18:1-CoA with oil content in soybean somatic embryos expressing CaDGAT1 variants. Creation of GmDGAT1b Variants, and Their Effects on Oil and Fatty Acid Composition of Soybean Somatic Embryos The amino acid substitutions present in CaDGAT1-C9, CaDGAT1-C10, and CaDGAT1-C11 were highly effective in increasing oil content in soybean somatic embryos (Fig. 2). To determine whether these amino acid substitutions also could improve another plant type 1 DGAT, the corresponding substitutions were made in the soybean type 1 DGAT, GmDGAT1b (Glyma.17G053300), at positions where the new amino acid was not already present in the GmDGAT1b protein (Supplemental Table S1). The corresponding soybean variants GmDGAT1b-C9, GmDGAT1b-C10, and GmDGAT1b-C11 contained five, five, and 11 substitutions, respectively. A fourth variant, GmDGAT1b-MOD, was designed with all 14 substitutions that were collectively present in the three CaDGAT1 variants that were not already present in GmDGAT1b. The four GmDGAT1b variants were expressed in soybean somatic embryos to determine the effects on oil and fatty acid composition (Fig. 4). All four soybean DGAT variants gave increased oil contents compared with that achieved by GmDGAT1b. The GmDGAT1b-MOD variant was the most effective in increasing oil content, to a level of 10%, which was double that of the GmDGAT1b value of 5%. Increases in 18:1 and 18:0, and decreases in 16:0, 18:2, and 18:3, also were observed for the GmDGAT1b-MOD variant. The largest changes in fatty acid composition were observed for 18:1, with a value of 28.3% for GmDGAT1b-MOD compared with 20% for GmDGAT1b, and for 18:3, with a value of 11% for GmDGAT1b-MOD compared with 17.9% for GmDGAT1b. Considering all GmDGAT1b variants plus the wild type, there was a particularly strong inverse correlation of oil content and 18:3 content (r 2 = 0.99). Figure 4. Open in new tabDownload slide Oil and fatty acid composition of soybean somatic embryos expressing soybean DGAT variants. Single and double asterisks denote statistically significant differences from GmDGAT1b values at P < 0.05 and P < 0.01, respectively, in two-tailed Student’s t tests. Error bars indicate se. Figure 4. Open in new tabDownload slide Oil and fatty acid composition of soybean somatic embryos expressing soybean DGAT variants. Single and double asterisks denote statistically significant differences from GmDGAT1b values at P < 0.05 and P < 0.01, respectively, in two-tailed Student’s t tests. Error bars indicate se. Seed Composition of Soybean Events Expressing the Engineered Soybean DGAT Recombination-mediated cassette exchange (Li et al., 2009, 2010b) was used to create transgenic soybean events that expressed the GmDGAT1b-MOD variant under the control of the promoter of the soybean β-conglycinin α-prime subunit gene (Glyma.10G246300) known to confer strong seed-preferred gene expression (Beachy et al., 1985). A total of four events were created that harbor a single-copy transgene at the same location in the soybean genome (Supplemental Fig. S2D). T0 plants were grown to maturity in controlled environments. T1 seeds were harvested, and the seed compositional traits shown in Table V were predicted using SS-NIR spectroscopy (Haefele et al., 2007; Janni et al., 2008; Agelet et al., 2012). For each event, 10 or more T1 plants homozygous for the DGAT transgene were grown alongside 10 or more null segregant T1 plants in controlled environments, and T2 seeds were harvested from null isoline plants or lines homozygous for the GmDGAT1b-MOD transgene and subjected to prediction of compositional traits, summarized in Table VI, using FT-NIR spectroscopy of seed samples derived from individual plants. Finally, a field trial was performed during the summer of 2015 in Johnston, Iowa, using 10 transgenic and null-segregant T2 seed sources from each of the four events. Every seed source was tested in three replications of 10-foot rows comprising 50 seeds. In this manner, the seed compositional performance of 120 rows of soybean lines homozygous for the GmDGAT1b-MOD transgene was compared with that of 120 rows of null segregant isolines devoid of the transgene. Individual rows were bulk harvested, and seed composition, summarized in Table VII, was predicted using NIT spectroscopy. Composition of T1 seeds expressing GmDGAT1b-MOD Table V. Composition of T1 seeds expressing GmDGAT1b-MOD A total of four transgenic events expressing GmDGAT1b-MOD were generated using particle bombardment and recombination-mediated cassette exchange. For every event, two T0 plants were generated and grown to maturity in controlled environments. Thirty-six seeds from every T0 source were subjected to nondestructive compositional analysis using single-seed near-infrared reflectance spectroscopy (SS-NIR). In this manner, 72 seeds from each of the four events were analyzed. Seeds were classified as null or transgenic based on oleic acid and total oil contents. Mean comparisons that were not significant at the 5% level (probability t test, homoscedastic Student’s t test P ≤ 0.05; Microsoft Excel) are indicated by ns. TG, Transgenic segregant; Ɗ % pts, change in percentage points. Event . Percentage Oleic Acid (18:1) . Percentage Oil . Percentage Protein . χ2 . . Null . TG . Ɗ % pts . Null . TG . Ɗ % pts . Null . TG . Ɗ % pts . . Soil 111 SOY 3615.1.1 22.7 32.7 10.0 18.3 23.6 5.3 40.6 39.3 −1.3 (ns) 1.7 Soil 111 SOY 3615.2.2 18.3 28.4 10.1 19.6 24.4 4.8 33.6 34.6 1.0 (ns) 0.3 Soil 111 SOY 3615.4.1 18.2 29.2 10.9 17.8 22.9 5.1 36.8 37.5 0.7 (ns) 0.6 Soil 111 SOY 3615.9.1 20.2 30.6 10.3 19.2 24.3 5.1 36.8 37.6 0.8 1.2 All event average 19.9 30.2 10.3 18.7 23.8 5.1 37.0 37.3 0.3 Event . Percentage Oleic Acid (18:1) . Percentage Oil . Percentage Protein . χ2 . . Null . TG . Ɗ % pts . Null . TG . Ɗ % pts . Null . TG . Ɗ % pts . . Soil 111 SOY 3615.1.1 22.7 32.7 10.0 18.3 23.6 5.3 40.6 39.3 −1.3 (ns) 1.7 Soil 111 SOY 3615.2.2 18.3 28.4 10.1 19.6 24.4 4.8 33.6 34.6 1.0 (ns) 0.3 Soil 111 SOY 3615.4.1 18.2 29.2 10.9 17.8 22.9 5.1 36.8 37.5 0.7 (ns) 0.6 Soil 111 SOY 3615.9.1 20.2 30.6 10.3 19.2 24.3 5.1 36.8 37.6 0.8 1.2 All event average 19.9 30.2 10.3 18.7 23.8 5.1 37.0 37.3 0.3 Open in new tab Table V. Composition of T1 seeds expressing GmDGAT1b-MOD A total of four transgenic events expressing GmDGAT1b-MOD were generated using particle bombardment and recombination-mediated cassette exchange. For every event, two T0 plants were generated and grown to maturity in controlled environments. Thirty-six seeds from every T0 source were subjected to nondestructive compositional analysis using single-seed near-infrared reflectance spectroscopy (SS-NIR). In this manner, 72 seeds from each of the four events were analyzed. Seeds were classified as null or transgenic based on oleic acid and total oil contents. Mean comparisons that were not significant at the 5% level (probability t test, homoscedastic Student’s t test P ≤ 0.05; Microsoft Excel) are indicated by ns. TG, Transgenic segregant; Ɗ % pts, change in percentage points. Event . Percentage Oleic Acid (18:1) . Percentage Oil . Percentage Protein . χ2 . . Null . TG . Ɗ % pts . Null . TG . Ɗ % pts . Null . TG . Ɗ % pts . . Soil 111 SOY 3615.1.1 22.7 32.7 10.0 18.3 23.6 5.3 40.6 39.3 −1.3 (ns) 1.7 Soil 111 SOY 3615.2.2 18.3 28.4 10.1 19.6 24.4 4.8 33.6 34.6 1.0 (ns) 0.3 Soil 111 SOY 3615.4.1 18.2 29.2 10.9 17.8 22.9 5.1 36.8 37.5 0.7 (ns) 0.6 Soil 111 SOY 3615.9.1 20.2 30.6 10.3 19.2 24.3 5.1 36.8 37.6 0.8 1.2 All event average 19.9 30.2 10.3 18.7 23.8 5.1 37.0 37.3 0.3 Event . Percentage Oleic Acid (18:1) . Percentage Oil . Percentage Protein . χ2 . . Null . TG . Ɗ % pts . Null . TG . Ɗ % pts . Null . TG . Ɗ % pts . . Soil 111 SOY 3615.1.1 22.7 32.7 10.0 18.3 23.6 5.3 40.6 39.3 −1.3 (ns) 1.7 Soil 111 SOY 3615.2.2 18.3 28.4 10.1 19.6 24.4 4.8 33.6 34.6 1.0 (ns) 0.3 Soil 111 SOY 3615.4.1 18.2 29.2 10.9 17.8 22.9 5.1 36.8 37.5 0.7 (ns) 0.6 Soil 111 SOY 3615.9.1 20.2 30.6 10.3 19.2 24.3 5.1 36.8 37.6 0.8 1.2 All event average 19.9 30.2 10.3 18.7 23.8 5.1 37.0 37.3 0.3 Open in new tab Composition of T2 seeds expressing GmDGAT1b-MOD Table VI. Composition of T2 seeds expressing GmDGAT1b-MOD Forty-eight seeds were planted for every T1 seed source. The transgene zygosity of every plant was determined using quantitative PCR analysis of genomic DNA. For every event, T1 null isolines were grown alongside lines homozygous for the GmDGAT1b-MOD transgene. T2 seeds were harvested by plant, and bulk seed samples of every T1 plant were subjected to compositional analysis of oil, protein, and oleic acid content using Fourier-transform near-infrared (FT-NIR) spectroscopy. Mean comparisons that were not significant at the 5% level (probability t test, homoscedastic Student’s t test P ≤ 0.05; Microsoft Excel) are indicated by ns. TG, Transgenic segregant; Ɗ % pts, change in percentage points. Event . n . Seed Mass per Plant . Percentage Oleic Acid (18:1) . Percentage Oil . Percentage Protein . . Null . TG . Null . TG . Ɗ % pts . Null . TG . Ɗ % pts . Null . TG . Ɗ % pts . Null . TG . Ɗ % pts . g Soil 111 SOY 3615.1.1 17 28 32.2 32.7 1.6 (ns) 26.3 35.4 9.1 19.0 22.7 3.7 34.5 34.9 0.4 (ns) Soil 111 SOY 3615.2.2 18 24 31.2 33.7 8.0 (ns) 25.3 35.4 10.1 18.7 22.7 4.0 34.6 35.4 0.8 Soil 111 SOY 3615.4.1 16 20 39.0 41.2 5.6 (ns) 23.6 33.9 10.3 19.8 23.1 3.3 33.9 34.8 0.9 Soil 111 SOY 3615.9.1 13 17 34.1 31.0 −9.1 (ns) 26.6 37.4 10.8 19.1 22.3 3.2 35.9 36.6 0.7 (ns) All event average 34.1 34.7 1.5 25.5 35.5 10.1 19.2 22.7 3.5 34.7 35.4 0.7 Event . n . Seed Mass per Plant . Percentage Oleic Acid (18:1) . Percentage Oil . Percentage Protein . . Null . TG . Null . TG . Ɗ % pts . Null . TG . Ɗ % pts . Null . TG . Ɗ % pts . Null . TG . Ɗ % pts . g Soil 111 SOY 3615.1.1 17 28 32.2 32.7 1.6 (ns) 26.3 35.4 9.1 19.0 22.7 3.7 34.5 34.9 0.4 (ns) Soil 111 SOY 3615.2.2 18 24 31.2 33.7 8.0 (ns) 25.3 35.4 10.1 18.7 22.7 4.0 34.6 35.4 0.8 Soil 111 SOY 3615.4.1 16 20 39.0 41.2 5.6 (ns) 23.6 33.9 10.3 19.8 23.1 3.3 33.9 34.8 0.9 Soil 111 SOY 3615.9.1 13 17 34.1 31.0 −9.1 (ns) 26.6 37.4 10.8 19.1 22.3 3.2 35.9 36.6 0.7 (ns) All event average 34.1 34.7 1.5 25.5 35.5 10.1 19.2 22.7 3.5 34.7 35.4 0.7 Open in new tab Table VI. Composition of T2 seeds expressing GmDGAT1b-MOD Forty-eight seeds were planted for every T1 seed source. The transgene zygosity of every plant was determined using quantitative PCR analysis of genomic DNA. For every event, T1 null isolines were grown alongside lines homozygous for the GmDGAT1b-MOD transgene. T2 seeds were harvested by plant, and bulk seed samples of every T1 plant were subjected to compositional analysis of oil, protein, and oleic acid content using Fourier-transform near-infrared (FT-NIR) spectroscopy. Mean comparisons that were not significant at the 5% level (probability t test, homoscedastic Student’s t test P ≤ 0.05; Microsoft Excel) are indicated by ns. TG, Transgenic segregant; Ɗ % pts, change in percentage points. Event . n . Seed Mass per Plant . Percentage Oleic Acid (18:1) . Percentage Oil . Percentage Protein . . Null . TG . Null . TG . Ɗ % pts . Null . TG . Ɗ % pts . Null . TG . Ɗ % pts . Null . TG . Ɗ % pts . g Soil 111 SOY 3615.1.1 17 28 32.2 32.7 1.6 (ns) 26.3 35.4 9.1 19.0 22.7 3.7 34.5 34.9 0.4 (ns) Soil 111 SOY 3615.2.2 18 24 31.2 33.7 8.0 (ns) 25.3 35.4 10.1 18.7 22.7 4.0 34.6 35.4 0.8 Soil 111 SOY 3615.4.1 16 20 39.0 41.2 5.6 (ns) 23.6 33.9 10.3 19.8 23.1 3.3 33.9 34.8 0.9 Soil 111 SOY 3615.9.1 13 17 34.1 31.0 −9.1 (ns) 26.6 37.4 10.8 19.1 22.3 3.2 35.9 36.6 0.7 (ns) All event average 34.1 34.7 1.5 25.5 35.5 10.1 19.2 22.7 3.5 34.7 35.4 0.7 Event . n . Seed Mass per Plant . Percentage Oleic Acid (18:1) . Percentage Oil . Percentage Protein . . Null . TG . Null . TG . Ɗ % pts . Null . TG . Ɗ % pts . Null . TG . Ɗ % pts . Null . TG . Ɗ % pts . g Soil 111 SOY 3615.1.1 17 28 32.2 32.7 1.6 (ns) 26.3 35.4 9.1 19.0 22.7 3.7 34.5 34.9 0.4 (ns) Soil 111 SOY 3615.2.2 18 24 31.2 33.7 8.0 (ns) 25.3 35.4 10.1 18.7 22.7 4.0 34.6 35.4 0.8 Soil 111 SOY 3615.4.1 16 20 39.0 41.2 5.6 (ns) 23.6 33.9 10.3 19.8 23.1 3.3 33.9 34.8 0.9 Soil 111 SOY 3615.9.1 13 17 34.1 31.0 −9.1 (ns) 26.6 37.4 10.8 19.1 22.3 3.2 35.9 36.6 0.7 (ns) All event average 34.1 34.7 1.5 25.5 35.5 10.1 19.2 22.7 3.5 34.7 35.4 0.7 Open in new tab Composition of field trial-derived T3 seeds expressing GmDGAT1b-MOD Table VII. Composition of field trial-derived T3 seeds expressing GmDGAT1b-MOD T2 seeds from 10 transgenic and null-segregant seed sources from each of the four events were planted in 2015 in a field in Johnston, Iowa. Every seed source was planted in three replications of randomly distributed 10-foot rows composed of 50 seeds. In this manner, seed agronomics and seed compositional performance of 120 rows of soybean lines homozygous for the GmDGAT1b-MOD transgene were compared with those of 120 rows of null-segregant isolines devoid of the transgene. Individual rows were bulk harvested, and seed composition was predicted using near-infrared transmission (NIT) spectroscopy. Mean differences of trait versus null were determined based on Fisher’s lsd approach at P < 0.05. Mean comparisons that were not significant at the 5% level are indicated by ns. TG, Transgenic segregant; Ɗ % pts, change in percentage points. Event . Percentage Oleic Acid (18:1) . Percentage Oil . Percentage Protein . . Null . TG . Ɗ % pts . Null . TG . Ɗ % pts . Null . TG . Ɗ % pts . Soil 111 SOY 3615.1.1 20.2 30.1 9.9 19.5 22.4 2.9 34.5 35.4 1.0 Soil 111 SOY 3615.2.2 20.5 28.2 7.7 19.1 22.5 3.4 35.1 35.3 0.2 (ns) Soil 111 SOY 3615.4.1 20.5 29.1 8.7 19.7 22.5 2.8 34.8 35.4 0.7 Soil 111 SOY 3615.9.1 23.1 30.5 7.5 18.9 21.7 2.8 36.1 36.3 0.1 (ns) All event average 21.1 29.5 8.4 19.3 22.3 3.0 35.1 35.6 0.5 Event . Percentage Oleic Acid (18:1) . Percentage Oil . Percentage Protein . . Null . TG . Ɗ % pts . Null . TG . Ɗ % pts . Null . TG . Ɗ % pts . Soil 111 SOY 3615.1.1 20.2 30.1 9.9 19.5 22.4 2.9 34.5 35.4 1.0 Soil 111 SOY 3615.2.2 20.5 28.2 7.7 19.1 22.5 3.4 35.1 35.3 0.2 (ns) Soil 111 SOY 3615.4.1 20.5 29.1 8.7 19.7 22.5 2.8 34.8 35.4 0.7 Soil 111 SOY 3615.9.1 23.1 30.5 7.5 18.9 21.7 2.8 36.1 36.3 0.1 (ns) All event average 21.1 29.5 8.4 19.3 22.3 3.0 35.1 35.6 0.5 Open in new tab Table VII. Composition of field trial-derived T3 seeds expressing GmDGAT1b-MOD T2 seeds from 10 transgenic and null-segregant seed sources from each of the four events were planted in 2015 in a field in Johnston, Iowa. Every seed source was planted in three replications of randomly distributed 10-foot rows composed of 50 seeds. In this manner, seed agronomics and seed compositional performance of 120 rows of soybean lines homozygous for the GmDGAT1b-MOD transgene were compared with those of 120 rows of null-segregant isolines devoid of the transgene. Individual rows were bulk harvested, and seed composition was predicted using near-infrared transmission (NIT) spectroscopy. Mean differences of trait versus null were determined based on Fisher’s lsd approach at P < 0.05. Mean comparisons that were not significant at the 5% level are indicated by ns. TG, Transgenic segregant; Ɗ % pts, change in percentage points. Event . Percentage Oleic Acid (18:1) . Percentage Oil . Percentage Protein . . Null . TG . Ɗ % pts . Null . TG . Ɗ % pts . Null . TG . Ɗ % pts . Soil 111 SOY 3615.1.1 20.2 30.1 9.9 19.5 22.4 2.9 34.5 35.4 1.0 Soil 111 SOY 3615.2.2 20.5 28.2 7.7 19.1 22.5 3.4 35.1 35.3 0.2 (ns) Soil 111 SOY 3615.4.1 20.5 29.1 8.7 19.7 22.5 2.8 34.8 35.4 0.7 Soil 111 SOY 3615.9.1 23.1 30.5 7.5 18.9 21.7 2.8 36.1 36.3 0.1 (ns) All event average 21.1 29.5 8.4 19.3 22.3 3.0 35.1 35.6 0.5 Event . Percentage Oleic Acid (18:1) . Percentage Oil . Percentage Protein . . Null . TG . Ɗ % pts . Null . TG . Ɗ % pts . Null . TG . Ɗ % pts . Soil 111 SOY 3615.1.1 20.2 30.1 9.9 19.5 22.4 2.9 34.5 35.4 1.0 Soil 111 SOY 3615.2.2 20.5 28.2 7.7 19.1 22.5 3.4 35.1 35.3 0.2 (ns) Soil 111 SOY 3615.4.1 20.5 29.1 8.7 19.7 22.5 2.8 34.8 35.4 0.7 Soil 111 SOY 3615.9.1 23.1 30.5 7.5 18.9 21.7 2.8 36.1 36.3 0.1 (ns) All event average 21.1 29.5 8.4 19.3 22.3 3.0 35.1 35.6 0.5 Open in new tab Tables V to VII demonstrate that an increase of seed oil content of at least 3 percentage points (i.e. 16% oil increase on a relative basis) was consistently observed across three generations in controlled as well as field environments. There was also a small increase in protein content that was not always statistically significant. Segregation of the high-oil trait among T0 plant progeny reflects the segregation pattern of a dominant trait (χ2 = 0.3–1.7). Mature seeds derived from field testing were subjected to compositional analysis of soluble sugars and fatty acids (Meyer et al., 2012), and test weight as a measure of seed density also was determined. Table VIII shows that the greatest absolute changes in sugar composition accompanying the oil increase were reductions in Suc (31%, 1.5 percentage points) and raffinosaccharides (7%, 0.4 percentage points). There were also small absolute reductions in other soluble sugars such as Fru and Glc. Thus, a significant fraction of the oil increase observed in response to expression of the engineered DGAT1 was accounted for by a concomitant reduction of soluble sugars. It should be of note that the small (i.e. 1%) change in test weight is most likely related to the difference in density between seed storage lipid (0.93 g cm−3) and Suc (1.59 g cm−3). The change in the fatty acid profile associated with expression of the engineered soybean DGAT1 protein mostly agreed with changes observed previously in the soybean somatic embryo system. The greatest absolute changes were a reduction of the 18:2 content and an increase of the 18:1 content (Table IX; Supplemental Table S2). There were also small but significant increases of the saturated fatty acids 16:0 and 18:0. The increase of the former was not observed in soybean somatic embryos. Finally, like previously observed soybean somatic embryos, the 18:3 content was reduced significantly. Seed composition (soluble sugars) Table VIII. Seed composition (soluble sugars) For each event, field trial-derived seed samples from two homozygous transgenic and null segregant lines were subjected to analysis of soluble carbohydrate content in triplicates. Test weight was determined by triplicate determination of the weight of a 50-mL volume of packed seeds. Fatty acid composition was determined by gas chromatography analysis of extracted oils. Mean comparisons that were not significant at the 5% level (probability t test, homoscedastic Student’s t test P ≤ 0.05; Microsoft Excel) are indicated by ns. TG, Transgene. Event Name . Type . Percentage Moisture . 50-mL Weight . Percentage of Bean at 13% Moisture . . . . . Fru . Glc . Suc . Galactinol + Raffinosaccharides . Total Soluble Sugars . g 3615.1.1 TG 5.6 33.4 0.02 0.04 3.1 4.5 7.9 sd 0.17 0.09 0.02 0.02 0.08 0.06 0.12 3615.1.1 Null 6.0 33.4 0.02 0.08 4.7 5.1 10.1 sd 0.13 0.14 0.02 0.04 0.07 0.08 0.09 Percentage change −6.1 −0.2 (ns) −2.8 (ns) −41.2 (ns) −33.2 −11.6 −21.7 3615.2.2 TG 5.5 33.3 0.01 0.03 3.2 4.8 8.4 sd 0.05 0.14 0.01 0.01 0.09 0.16 0.25 3615.2.2 Null 6.2 33.7 0.03 0.10 4.5 4.9 9.9 sd 0.20 0.16 0.01 0.02 0.10 0.18 0.25 Percentage change −12.4 −1.4 −71.4 −65.1 −29.2 −2.0 (ns) −15.4 3615.4.1 TG 5.6 34.2 0.02 0.05 3.3 4.6 8.2 sd 0.07 0.42 0.02 0.02 0.07 0.12 0.16 3615.4.1 Null 5.8 34.2 0.03 0.07 4.7 5.0 10.1 sd 0.13 0.71 0.01 0.03 0.04 0.22 0.25 Percentage change −3.6 0.2 (ns) −42.7 (ns) −36.1 (ns) −30.8 −7.5 −18.5 3615.9.1 TG 5.6 33.1 0.00 0.03 3.2 4.6 8.1 sd 0.07 0.35 0.01 0.00 0.12 0.10 0.17 3615.9.1 Null 6.1 34.0 0.02 0.06 4.7 4.9 10.0 sd 0.23 0.19 0.01 0.02 0.20 0.15 0.35 Percentage change −8.7 −2.7 −75.9 −57.8 −32.2 −7.0 −19.1 Average percentage change −7.7 −1.0 −48.2 −50.1 −31.3 −7.0 −18.7 Event Name . Type . Percentage Moisture . 50-mL Weight . Percentage of Bean at 13% Moisture . . . . . Fru . Glc . Suc . Galactinol + Raffinosaccharides . Total Soluble Sugars . g 3615.1.1 TG 5.6 33.4 0.02 0.04 3.1 4.5 7.9 sd 0.17 0.09 0.02 0.02 0.08 0.06 0.12 3615.1.1 Null 6.0 33.4 0.02 0.08 4.7 5.1 10.1 sd 0.13 0.14 0.02 0.04 0.07 0.08 0.09 Percentage change −6.1 −0.2 (ns) −2.8 (ns) −41.2 (ns) −33.2 −11.6 −21.7 3615.2.2 TG 5.5 33.3 0.01 0.03 3.2 4.8 8.4 sd 0.05 0.14 0.01 0.01 0.09 0.16 0.25 3615.2.2 Null 6.2 33.7 0.03 0.10 4.5 4.9 9.9 sd 0.20 0.16 0.01 0.02 0.10 0.18 0.25 Percentage change −12.4 −1.4 −71.4 −65.1 −29.2 −2.0 (ns) −15.4 3615.4.1 TG 5.6 34.2 0.02 0.05 3.3 4.6 8.2 sd 0.07 0.42 0.02 0.02 0.07 0.12 0.16 3615.4.1 Null 5.8 34.2 0.03 0.07 4.7 5.0 10.1 sd 0.13 0.71 0.01 0.03 0.04 0.22 0.25 Percentage change −3.6 0.2 (ns) −42.7 (ns) −36.1 (ns) −30.8 −7.5 −18.5 3615.9.1 TG 5.6 33.1 0.00 0.03 3.2 4.6 8.1 sd 0.07 0.35 0.01 0.00 0.12 0.10 0.17 3615.9.1 Null 6.1 34.0 0.02 0.06 4.7 4.9 10.0 sd 0.23 0.19 0.01 0.02 0.20 0.15 0.35 Percentage change −8.7 −2.7 −75.9 −57.8 −32.2 −7.0 −19.1 Average percentage change −7.7 −1.0 −48.2 −50.1 −31.3 −7.0 −18.7 Open in new tab Table VIII. Seed composition (soluble sugars) For each event, field trial-derived seed samples from two homozygous transgenic and null segregant lines were subjected to analysis of soluble carbohydrate content in triplicates. Test weight was determined by triplicate determination of the weight of a 50-mL volume of packed seeds. Fatty acid composition was determined by gas chromatography analysis of extracted oils. Mean comparisons that were not significant at the 5% level (probability t test, homoscedastic Student’s t test P ≤ 0.05; Microsoft Excel) are indicated by ns. TG, Transgene. Event Name . Type . Percentage Moisture . 50-mL Weight . Percentage of Bean at 13% Moisture . . . . . Fru . Glc . Suc . Galactinol + Raffinosaccharides . Total Soluble Sugars . g 3615.1.1 TG 5.6 33.4 0.02 0.04 3.1 4.5 7.9 sd 0.17 0.09 0.02 0.02 0.08 0.06 0.12 3615.1.1 Null 6.0 33.4 0.02 0.08 4.7 5.1 10.1 sd 0.13 0.14 0.02 0.04 0.07 0.08 0.09 Percentage change −6.1 −0.2 (ns) −2.8 (ns) −41.2 (ns) −33.2 −11.6 −21.7 3615.2.2 TG 5.5 33.3 0.01 0.03 3.2 4.8 8.4 sd 0.05 0.14 0.01 0.01 0.09 0.16 0.25 3615.2.2 Null 6.2 33.7 0.03 0.10 4.5 4.9 9.9 sd 0.20 0.16 0.01 0.02 0.10 0.18 0.25 Percentage change −12.4 −1.4 −71.4 −65.1 −29.2 −2.0 (ns) −15.4 3615.4.1 TG 5.6 34.2 0.02 0.05 3.3 4.6 8.2 sd 0.07 0.42 0.02 0.02 0.07 0.12 0.16 3615.4.1 Null 5.8 34.2 0.03 0.07 4.7 5.0 10.1 sd 0.13 0.71 0.01 0.03 0.04 0.22 0.25 Percentage change −3.6 0.2 (ns) −42.7 (ns) −36.1 (ns) −30.8 −7.5 −18.5 3615.9.1 TG 5.6 33.1 0.00 0.03 3.2 4.6 8.1 sd 0.07 0.35 0.01 0.00 0.12 0.10 0.17 3615.9.1 Null 6.1 34.0 0.02 0.06 4.7 4.9 10.0 sd 0.23 0.19 0.01 0.02 0.20 0.15 0.35 Percentage change −8.7 −2.7 −75.9 −57.8 −32.2 −7.0 −19.1 Average percentage change −7.7 −1.0 −48.2 −50.1 −31.3 −7.0 −18.7 Event Name . Type . Percentage Moisture . 50-mL Weight . Percentage of Bean at 13% Moisture . . . . . Fru . Glc . Suc . Galactinol + Raffinosaccharides . Total Soluble Sugars . g 3615.1.1 TG 5.6 33.4 0.02 0.04 3.1 4.5 7.9 sd 0.17 0.09 0.02 0.02 0.08 0.06 0.12 3615.1.1 Null 6.0 33.4 0.02 0.08 4.7 5.1 10.1 sd 0.13 0.14 0.02 0.04 0.07 0.08 0.09 Percentage change −6.1 −0.2 (ns) −2.8 (ns) −41.2 (ns) −33.2 −11.6 −21.7 3615.2.2 TG 5.5 33.3 0.01 0.03 3.2 4.8 8.4 sd 0.05 0.14 0.01 0.01 0.09 0.16 0.25 3615.2.2 Null 6.2 33.7 0.03 0.10 4.5 4.9 9.9 sd 0.20 0.16 0.01 0.02 0.10 0.18 0.25 Percentage change −12.4 −1.4 −71.4 −65.1 −29.2 −2.0 (ns) −15.4 3615.4.1 TG 5.6 34.2 0.02 0.05 3.3 4.6 8.2 sd 0.07 0.42 0.02 0.02 0.07 0.12 0.16 3615.4.1 Null 5.8 34.2 0.03 0.07 4.7 5.0 10.1 sd 0.13 0.71 0.01 0.03 0.04 0.22 0.25 Percentage change −3.6 0.2 (ns) −42.7 (ns) −36.1 (ns) −30.8 −7.5 −18.5 3615.9.1 TG 5.6 33.1 0.00 0.03 3.2 4.6 8.1 sd 0.07 0.35 0.01 0.00 0.12 0.10 0.17 3615.9.1 Null 6.1 34.0 0.02 0.06 4.7 4.9 10.0 sd 0.23 0.19 0.01 0.02 0.20 0.15 0.35 Percentage change −8.7 −2.7 −75.9 −57.8 −32.2 −7.0 −19.1 Average percentage change −7.7 −1.0 −48.2 −50.1 −31.3 −7.0 −18.7 Open in new tab Seed composition (fatty acids) Table IX. Seed composition (fatty acids) Fatty acid composition was determined by gas chromatography analysis of extracted oils. Mean comparisons that were not significant at the 5% level (probability t test, homoscedastic Student’s t test P ≤ 0.05; Microsoft Excel) are indicated by ns. TG, Transgene. Event Name . Type . Percentage of Extractable Fatty Acids . . . Palmitic (16:0) . Stearic (18:0) . Oleic (18:1) . Linoleic (18:2) . Linolenic (18:3) . Other . 3615.1.1 TG 12.1 5.0 31.8 44.9 5.6 0.6 sd 0.08 0.03 0.10 0.17 0.03 0.01 3615.1.1 Null 10.4 3.7 23.7 54.2 7.6 0.5 sd 0.07 0.13 0.16 0.19 0.10 0.01 Percentage change 16.7 35.2 34.3 −17.0 −26.9 25.1 3615.2.2 TG 12.1 5.1 31.6 45.3 5.4 0.6 sd 0.05 0.04 0.59 0.54 0.10 0.01 3615.2.2 Null 10.4 3.8 21.7 56.1 7.7 0.5 sd 0.05 0.01 0.22 0.19 0.07 0.01 Percentage change 17.0 35.2 45.7 −19.3 −30.1 25.7 3615.4.1 TG 11.8 5.0 32.5 45.0 5.2 0.6 sd 0.04 0.02 0.07 0.11 0.04 0.01 3615.4.1 Null 10.3 3.8 23.1 55.0 7.4 0.5 sd 0.09 0.01 0.59 0.40 0.09 0.02 Percentage change 14.0 33.5 40.6 −18.2 −29.6 23.1 3615.9.1 TG 11.5 5.4 34.4 43.2 4.9 0.6 sd 0.12 0.14 0.56 0.46 0.12 0.01 3615.9.1 Null 10.0 3.9 25.9 52.5 7.2 0.5 sd 0.09 0.03 0.33 0.28 0.16 0.01 Percentage change 14.3 39.9 32.8 −17.7 −31.7 23.0 Average percentage change 15.5 35.9 38.3 −18.1 −29.6 24.2 Event Name . Type . Percentage of Extractable Fatty Acids . . . Palmitic (16:0) . Stearic (18:0) . Oleic (18:1) . Linoleic (18:2) . Linolenic (18:3) . Other . 3615.1.1 TG 12.1 5.0 31.8 44.9 5.6 0.6 sd 0.08 0.03 0.10 0.17 0.03 0.01 3615.1.1 Null 10.4 3.7 23.7 54.2 7.6 0.5 sd 0.07 0.13 0.16 0.19 0.10 0.01 Percentage change 16.7 35.2 34.3 −17.0 −26.9 25.1 3615.2.2 TG 12.1 5.1 31.6 45.3 5.4 0.6 sd 0.05 0.04 0.59 0.54 0.10 0.01 3615.2.2 Null 10.4 3.8 21.7 56.1 7.7 0.5 sd 0.05 0.01 0.22 0.19 0.07 0.01 Percentage change 17.0 35.2 45.7 −19.3 −30.1 25.7 3615.4.1 TG 11.8 5.0 32.5 45.0 5.2 0.6 sd 0.04 0.02 0.07 0.11 0.04 0.01 3615.4.1 Null 10.3 3.8 23.1 55.0 7.4 0.5 sd 0.09 0.01 0.59 0.40 0.09 0.02 Percentage change 14.0 33.5 40.6 −18.2 −29.6 23.1 3615.9.1 TG 11.5 5.4 34.4 43.2 4.9 0.6 sd 0.12 0.14 0.56 0.46 0.12 0.01 3615.9.1 Null 10.0 3.9 25.9 52.5 7.2 0.5 sd 0.09 0.03 0.33 0.28 0.16 0.01 Percentage change 14.3 39.9 32.8 −17.7 −31.7 23.0 Average percentage change 15.5 35.9 38.3 −18.1 −29.6 24.2 Open in new tab Table IX. Seed composition (fatty acids) Fatty acid composition was determined by gas chromatography analysis of extracted oils. Mean comparisons that were not significant at the 5% level (probability t test, homoscedastic Student’s t test P ≤ 0.05; Microsoft Excel) are indicated by ns. TG, Transgene. Event Name . Type . Percentage of Extractable Fatty Acids . . . Palmitic (16:0) . Stearic (18:0) . Oleic (18:1) . Linoleic (18:2) . Linolenic (18:3) . Other . 3615.1.1 TG 12.1 5.0 31.8 44.9 5.6 0.6 sd 0.08 0.03 0.10 0.17 0.03 0.01 3615.1.1 Null 10.4 3.7 23.7 54.2 7.6 0.5 sd 0.07 0.13 0.16 0.19 0.10 0.01 Percentage change 16.7 35.2 34.3 −17.0 −26.9 25.1 3615.2.2 TG 12.1 5.1 31.6 45.3 5.4 0.6 sd 0.05 0.04 0.59 0.54 0.10 0.01 3615.2.2 Null 10.4 3.8 21.7 56.1 7.7 0.5 sd 0.05 0.01 0.22 0.19 0.07 0.01 Percentage change 17.0 35.2 45.7 −19.3 −30.1 25.7 3615.4.1 TG 11.8 5.0 32.5 45.0 5.2 0.6 sd 0.04 0.02 0.07 0.11 0.04 0.01 3615.4.1 Null 10.3 3.8 23.1 55.0 7.4 0.5 sd 0.09 0.01 0.59 0.40 0.09 0.02 Percentage change 14.0 33.5 40.6 −18.2 −29.6 23.1 3615.9.1 TG 11.5 5.4 34.4 43.2 4.9 0.6 sd 0.12 0.14 0.56 0.46 0.12 0.01 3615.9.1 Null 10.0 3.9 25.9 52.5 7.2 0.5 sd 0.09 0.03 0.33 0.28 0.16 0.01 Percentage change 14.3 39.9 32.8 −17.7 −31.7 23.0 Average percentage change 15.5 35.9 38.3 −18.1 −29.6 24.2 Event Name . Type . Percentage of Extractable Fatty Acids . . . Palmitic (16:0) . Stearic (18:0) . Oleic (18:1) . Linoleic (18:2) . Linolenic (18:3) . Other . 3615.1.1 TG 12.1 5.0 31.8 44.9 5.6 0.6 sd 0.08 0.03 0.10 0.17 0.03 0.01 3615.1.1 Null 10.4 3.7 23.7 54.2 7.6 0.5 sd 0.07 0.13 0.16 0.19 0.10 0.01 Percentage change 16.7 35.2 34.3 −17.0 −26.9 25.1 3615.2.2 TG 12.1 5.1 31.6 45.3 5.4 0.6 sd 0.05 0.04 0.59 0.54 0.10 0.01 3615.2.2 Null 10.4 3.8 21.7 56.1 7.7 0.5 sd 0.05 0.01 0.22 0.19 0.07 0.01 Percentage change 17.0 35.2 45.7 −19.3 −30.1 25.7 3615.4.1 TG 11.8 5.0 32.5 45.0 5.2 0.6 sd 0.04 0.02 0.07 0.11 0.04 0.01 3615.4.1 Null 10.3 3.8 23.1 55.0 7.4 0.5 sd 0.09 0.01 0.59 0.40 0.09 0.02 Percentage change 14.0 33.5 40.6 −18.2 −29.6 23.1 3615.9.1 TG 11.5 5.4 34.4 43.2 4.9 0.6 sd 0.12 0.14 0.56 0.46 0.12 0.01 3615.9.1 Null 10.0 3.9 25.9 52.5 7.2 0.5 sd 0.09 0.03 0.33 0.28 0.16 0.01 Percentage change 14.3 39.9 32.8 −17.7 −31.7 23.0 Average percentage change 15.5 35.9 38.3 −18.1 −29.6 24.2 Open in new tab DISCUSSION Creating improved DGAT variants was clearly a more effective strategy to increase soybean oil content than was using wild-type DGATs, because variants of both soybean and C. americana DGATs gave substantially greater oil contents in soybean somatic embryos than did either wild-type DGAT. Thus, neither the wild-type DGAT from an oilseed crop subjected to decades of intense breeding pressure for yield, nor that from a naturally high-oil plant tissue, was optimal for high oil production. Breeding for increased yield would not necessarily be expected to optimize DGAT, because the high energy content of oil means that a strong positive correlation of yield and oil content is unlikely. However, the fact that a DGAT from a tissue with 60% oil content fell so far short of optimal was more surprising. The comparisons made here were done with typical high-18:2 soybean somatic embryos, so it is possible that the C. americana DGAT performance would have been better in a high-18:1 genotype. The strong inverse correlation of oil contents and S 0.5 values for 18:1-CoA suggested that the DGAT affinity for acyl-CoA is an important determinant of oil synthesis in soybeans, and it also implied that the in vivo concentration of acyl-CoA must be quite low in order to make these S 0.5 values of less than 1 µm relevant. The concentration of the total acyl-CoA pool in developing B. napus seeds was determined to be 6 µm (Larson and Graham, 2001). If a similar concentration is present in soybeans, and only one-fifth of the total is 18:1-CoA, and only a fraction of that is available to DGAT, then even a few-fold decrease in S 0.5 values for 18:1-CoA, compared with a wild-type value of 0.94 µm, would be expected to have a significant impact on oil synthesis, as observed here. DGAT affinity for acyl-CoA appeared to be important for oil synthesis in yeast, as well as soybeans, because nearly all DGAT variants examined from high-oil yeast strains had decreased S 0.5 values. If DGAT affinity for acyl-CoA is indeed a limitation for oil synthesis, then an increase in acyl-CoA concentration, as an alternative to a DGAT with a better S 0.5 value, also would be expected to increase DGAT activity and oil synthesis. In a previous study, the addition of exogenous fatty acids to developing plant seed embryos increased the rate of oil synthesis (Bao and Ohlrogge, 1999) and may have worked at least partly via an increase in the concentration of the acyl-CoA substrate for DGAT. Aside from the importance of DGAT affinity for acyl-CoA, the kinetic data contributed additional insights concerning plant lipid metabolism. First, the sigmoidal DGAT kinetics, rather than hyperbolic, appeared metabolically meaningful. When acyl-CoA concentrations are low, less would be used by DGAT and more would be available for membrane synthesis, but when concentrations are high, the greater DGAT activity would result in more acyl-CoA used for storage oil production. Another implication of sigmoidal kinetics is that a DGAT variant with better affinity for 18:1-CoA, in a tetramer with three wild-type subunits, could potentially result in greater product being formed from the wild-type subunits also, as binding of the first substrate molecule by the engineered subunit would increase the substrate affinity of the other subunits. Second, it was interesting and logical that the S 0.5 value for 18:1-CoA was about 3 times lower for the DGAT from the high-18:1 Corylus spp. compared with the DGAT from the high-18:2 species, maize. DGATs from other high-18:1 plant tissues such as olive also may have high affinity for 18:1-CoA, as there would have been little selection pressure to optimize DGAT for other acyl-CoA substrates in high-18:1 tissues such as Olea or Corylus spp. Third, these plant DGAT1 kinetic properties differed markedly from published values for a fungal type 2 DGAT (Lardizabal et al., 2008). The U. ramanniana DGAT2a had K m values of 13 to 19.4 µm 18:1-CoA, depending on the DAG species used, which were substantially greater than either the S 0.5 values of 0.94 and 2.8 µm 18:1-CoA determined here for plant type1 DGATs or the K m values of less than 1 µm reported previously for DGAT activity in developing soybean seeds (Settlage et al., 1998). Similar DGAT kinetics would not necessarily be expected for heterotrophic organisms that might periodically encounter abundant food sources to be stored as oil compared with plants that make oil from photosynthetically derived carbon. Improved DGAT variants increased the 18:1 content, and decreased the 18:2 and 18:3 contents, of oil in both soybean somatic embryos and seeds, in marked contrast to previous studies in which overexpressed wild-type DGATs had little impact on soybean fatty acid composition (Lardizabal et al., 2008; Wang et al., 2014). One possible explanation for the increased 18:1 content is based on evidence that most newly made 18:1 fatty acyl groups in wild-type soybeans are initially transferred from the acyl-CoA pool to phosphatidylcholine (PC), where they can undergo further desaturation to produce 18:2 and 18:3, rather than being incorporated directly into TAG by the Kennedy pathway (Bates et al., 2009). A DGAT variant with increased affinity for 18:1-CoA would compete more effectively for the newly made 18:1-CoA, resulting in more 18:1 fatty acyl groups going directly for TAG synthesis without going through PC first, thus avoiding further desaturation. An alternative hypothesis for the higher 18:1 content considers that DAG and PC can be interconverted by choline phosphotransferase (Slack et al., 1985). A kinetically improved DGAT variant would be more likely to pull DAG toward TAG synthesis and away from PC-based desaturation reactions, even in the presence of a strong soybean FAD3 desaturase activity, as discussed previously (Zhang et al., 2013). The increased incorporation of saturated fatty acids into TAG, on the other hand, may be caused by increased flux through a distinct route of TAG biosynthesis with preference for nascent saturated acyl-CoAs (Bates et al., 2009). Clearly, elucidation of the substrate specificity of the GmDGAT1b-MOD variant for DAG and acyl-CoA substrates alongside analysis of acyl fluxes of developing soybeans expressing this soybean DGAT variant are required to clarify the basis of the altered seed fatty acid profile. The demonstration that beneficial amino acid substitutions in one DGAT also could improve another DGAT was especially helpful in this work. When we expressed the wild-type soybean DGAT in yeast, we observed significantly slower growth compared with yeast transformed with the C. americana DGAT or maize DGAT. The reason for the slower growth was unknown, but any amino acid substitutions or codon changes that allowed faster growth would result in larger colonies that would be more likely to get picked and also would result in strain enrichment during the growth periods done prior to low-density enrichment, even if the DGAT variant was not kinetically better. Thus, initially screening libraries of soybean DGAT variants in yeast might have yielded a high proportion of good growers rather than good DGATs. Therefore, the ability to obtain significantly improved soybean DGAT variants, without directly screening for them, was quite helpful. The soybean DGAT variants have not yet been expressed in yeast, but if the best variant allows more normal yeast growth than observed previously with the wild-type soybean DGAT, then it may be an ideal starting place for further engineering that could target an entirely different set of amino acid substitutions. Although we focused on plant type 1 DGATs in this study, a similar engineering approach could be taken with either type 1 or 2 DGATs from diverse sources. For example, many algal species have multiple type 2 DGAT genes (Chen and Smith, 2012), some of which may be amenable to improvement by DNA-shuffling technology. A powerful advantage of screening enzyme variants with multiple simultaneous amino acid substitutions is that beneficial substitutions can be discovered that never would have been predicted. For example, most CaDGAT1 variants examined had increased affinity for 18:1-CoA, yet no substitutions were observed in the acyl-CoA-binding site identified previously in the nasturtium (Tropaeolum majus) DGAT1 (Xu et al., 2008) corresponding to positions Lys-129 to Gly-145 of CaDGAT1. Without a DGAT three-dimensional structure available, it is difficult to interpret the effects of amino acid substitutions on function, but it is well known from other studies that substitutions far removed from the active site of an enzyme can still have large effects on function (Morley and Kazlauskas, 2005). Expression of GmDGAT1b-MOD using a strong seed-preferred promoter increased the oil content of mature seeds by 16% under field conditions, a much larger increase than that observed under field testing conditions in a previously published attempt at increasing soybean oil content through metabolic engineering (Lardizabal et al., 2008). Additional features that distinguish the GmDGAT1b-MOD-based high-oil trait are the absence of a reduction in seed protein content and the very obvious tradeoff between oil and soluble sugars. The compositional shift cannot be entirely explained by the concomitant reduction in soluble sugars. Specifically, a 3.5 percentage point increase of the seed storage compounds oil and protein is accompanied by a 1.9 percentage point reduction of soluble sugars. Preliminary analysis of the cell wall polysaccharide fraction of soybeans expressing GmDGAT1b-MOD indicates that there is also a significant reduction of the insoluble fiber fraction, namely pectin-derived Gal, GalUA, and arabinose (J. Everard, unpublished data). It is clear from the compositional analysis that the expression of the modified DGAT exerts significant metabolic pull in developing soybeans that alters the partitioning between soluble sugars and seed storage lipids. The compositional shift in mature seeds related to expression of the engineered DGAT enzyme indicates that increased glycerolipid assembly effectively competes with other pathways fed by the nucleotide sugar pools, such as cytosolic Suc synthesis, cell wall biosynthesis, and raffinosaccharide biosynthesis. The reduction in soluble sugars further suggests that seed oil accumulation in GmDGAT1b-MOD-expressing events is limited by source strength and should be sensitive to any manipulation of cytosolic Suc supply. Raffinosaccharides still constituted 4% to 5% of the transgenic seed weight, suggesting that expression of the engineered DGAT, in combination with methods to decrease raffinosaccharide synthesis, may allow even greater oil increases than observed here. On the applied side, this study highlights a set of amino acid modifications in the highly conserved scaffold of plant DGAT1 proteins that could be employed to improve the enzyme performance of any given plant DGAT1 that would be expected to lead to increased oil content. As such, it could provide multiple, almost native, transgene-derived proteins with the ability to increase the oil content of seeds or vegetative tissues. Recently developed gene-editing technology (Li et al., 2015) also could be used to change the amino acid sequence of endogenous DGAT1 proteins from multiple plant species of interest to achieve similar goals. CONCLUSION Our study applied DNA-shuffling and medium-throughput screening technology in a heterologous microbial host to the improvement of a key enzymatic step controlling seed storage lipid accumulation. Even though only a very small fraction of the created diversity was screened, several variants with improved kinetic properties were identified and validated by plant expression. Finally, it was shown that multiple changes in the amino acid sequence shown to benefit one DGAT1 protein can be transferred to a DGAT1 protein from a different plant family and continue to cause improved enzyme performance. As such, our study provided critical sequence information for the modification of the native DGAT1 sequence, enabling a soybean high-seed-oil trait that, under field conditions, showed improved trait efficacy compared with previous attempts and that did not lead to a reduction in seed protein content. MATERIALS AND METHODS Isolation of Corylus americana cDNA Developing C. americana seeds were collected locally, deshelled, and frozen in liquid nitrogen. The frozen tissue was ground under liquid nitrogen with a mortar and pestle, and total RNA was isolated using the Purescript RNA kit (from Gentra Systems at the time, now part of Qiagen). First-strand cDNA synthesis was performed using total RNA as template, with a SuperScript III kit from Invitrogen, using oligo(dT) as primer and following the manufacturer’s protocol. To obtain a partial-length DGAT cDNA, PCR was done with first-strand cDNA as template using forward and reverse primers designed from conserved regions of known type I DGAT genes. The forward primer nucleotide sequence was 5′-CAAGGAGAGTCCGCTTAGCTC-3′, and that of the reverse primer was 5′-CAGAAAATGAACCAGAAGATCATGTT-3′. The PCR program first used three cycles of 94°C for 30 s, 43°C for 30 s, and 72°C for 2 min; followed by 35 cycles of 94°C for 30 s, 55°C for 30 s, and 72°C for 2 min; followed by a final extension of 72°C for 10 min. PCR products were isolated and sequenced to confirm that they had sequence identity with known type I DGAT cDNAs. To obtain the rest of the DGAT coding region, Invitrogen 5′ RACE and 3′ RACE systems were used with gene-specific primers. To amplify the entire coding region, and to include restriction sites for cloning, a final PCR was done with the same program just mentioned, using forward primer 5′-TTTTAGATCTATGGCGATTTCGGATATGCCTGAAAGCACG-3′ and reverse primer 5′-TTTTCCCGGGTTATTCAGTCTTCCCTTTACGGTTCATC-3′. The full-length cDNA was then digested with BglII and SmaI and ligated into the BamHI and SmaI sites of a yeast/Escherichia coli shuttle vector that included the promoter and terminator of the yeast phosphoglycerate kinase gene (PGK1) and was reported previously (Zheng et al., 2008). This wild-type CaDGAT1 clone was used for the initial confirmation of DGAT activity. Subsequently, to facilitate the synthesis and cloning of large libraries of variants, the internal BamHI and EcoRI restriction sites were removed and the cDNA was ligated into the more efficient BamHI and EcoRI sites of the vector. This modified version of the CaDGAT1 construct, which retained the identical amino acid sequence and nearly identical nucleotide sequence as the wild type, was used in this study as the CaDGAT1 comparator versus the engineered variants. Preparation of CaDGAT1 Variant Libraries DNA-shuffling libraries were created as follows. Oligonucleotides encoding the target amino acids were purchased from Integrated DNA Technologies and assembled using PCR to recombine with the natural amino acid diversity represented in the CaDGAT1 fragments. Assembly reactions containing 0.5 to 1 µm library oligosaccharides were assembled in a Herculase II (Stratagene) reaction. Standard PCR conditions were used. Libraries were digested with BamHI/EcoRI and ligated into the yeast/E. coli shuttle vector reported previously (Zheng et al., 2008). The ligation mixture was then transformed into E. coli strain XL1Blue (Stratagene) and plated onto Luria-Bertani agar containing 50 µg mL−1 carbenicillin. Colonies were harvested, and plasmid DNA was collected for subsequent screening. Screening DGAT Variants in Yeast Plasmid DNA encoding CaDGAT1 variants was transformed into the low-oil Saccharomyces cerevisiae DGAT/PDAT double null strain (ƊDGA1/ƊLRO1) described previously (Zheng et al., 2008) using a published transformation method (Gietz and Schiestl, 2007), except that the heat shock was done at 37°C rather than 42°C. The medium used was SC minus uracil and was composed of 20 g of Glc, 6.7 g of Difco yeast nitrogen base without amino acids (BD catalog no. 291940), and 0.77 g of –uracil dropout supplement (Clontech catalog no. 630416) per 1 L. Transformed yeast was plated, and single colonies were picked to start 1-mL cultures in 96-deep-well culture blocks with a 2-mL volume per well capacity. Airpore membranes were used to cover the blocks, and cultures were grown for 2 to 3 d at 30°C with shaking at 300 rpm. Nile Red staining was performed in 96-well microtiter plates by adding 5 µL of a 0.02 mg mL−1 stock of Nile Red dissolved in 95% ethanol to 200 µL of a 1:10 dilution of the yeast culture in phosphate-buffered saline (137 mm NaCl, 2.7 mm KCl, 10 mm Na2HPO4, and 1.8 mm KH2PO4, pH 7.4). Staining was done for 5 min, followed by the determination of fluorescence intensity using an excitation wavelength of 489 nm and an emission wavelength of 581 nm. Fluorescence intensity was divided by A 600 to correct for differences in cell density, and buffer blanks with no yeast were used to correct for background. A Biotek robotics system was used to initially suspend any settled yeast in the culture blocks by pipetting 15 times and then to dilute the yeast and combine with Nile Red on the microtiter plates. Typically, four to six cultures per block with the greatest Nile Red staining were regrown and restained for confirmation as before, except that 3-mL volumes in 15-mL culture tubes were used to avoid any block position bias. Although the surest way to achieve complete coverage of a 5,000 strain library would be to screen all 5,000 strains directly by Nile Red staining, we sometimes enriched for low-density strains first, followed by Nile Red staining of fewer strains. For low-density strain enrichment, typically several thousand yeast colonies were combined by scraping multiple plates with 4 mL of medium per plate, mixing together, and then using a 1:100 dilution to start a 25-mL culture that was grown for 2 to 3 d at 30°C with shaking at 250 rpm. The culture was harvested by centrifugation, suspended in 1 mL of medium, applied to the top of 2 mL of 60% glycerol in a test tube, and centrifuged at 2,500g for 5 min. The low-density yeast strains at or near the top were used to repeat the enrichment step once or twice, followed by plating and screening by Nile Red staining as described above. Microsomal Membrane Preparations and DGAT Activity Microsomal membranes were prepared by a published method (Milcamps et al., 2005) with minor modifications. Yeast cultures expressing DGAT variants were grown in 100-mL volumes to early stationary phase, harvested, and resuspended in 4 mL of 20 mm Tris-HCl, pH 8, 10 mm MgCl2, 1 mm EDTA, 5% glycerol, 1 mm dithiothreitol, and 300 mm (NH4)2SO4. Two milliliters of glass beads, 425 to 600 nm, was added, and cells were lysed by vortexing for 5 min. The lysate was centrifuged at 1,500g for 15 min at 6°C, and the resulting supernatant was then centrifuged at 100,000g for 1.5 h at 6°C. The microsomal membrane pellet was resuspended in 500 µL of 100 mm potassium phosphate, pH 7.2, containing 10% glycerol, divided into smaller aliquots, and frozen in liquid nitrogen, followed by storage at −80°C. Protein concentration was determined with the Bradford method using the Coomassie Plus reagent from Pierce, with bovine serum albumin as standard. To determine the kinetic parameters of Table IV, DGAT activity assays were done for 1 min at 25°C with 100-µL volumes containing 50 mm potassium phosphate, pH 7.2, microsomal membrane protein, and 1-14C-labeled oleoyl-CoA (Perkin Elmer; 50 mCi mmol−1), with no additional DAG added beyond the endogenous DAG already present in the microsomes. The microsomal protein quantity and the range of oleoyl-CoA concentrations used were as follows: for ZmDGAT1-2, 20 µg of microsomal protein and 1 to 10 µm oleoyl-CoA; for CaDGAT1, 5 µg of microsomal protein and 0.5 to 5 µm oleoyl-CoA; for all CaDGAT1 variants except CaDGAT1-C11, 5 µg of microsomal protein and 0.3 to 5 µm oleoyl-CoA; and for CaDGAT1-C11, 2 µg of microsomal protein and 0.1 to 3 µm oleoyl-CoA. Less CaDGAT-C11 microsomal protein was used to avoid excessive substrate depletion at the low-oleoyl-CoA concentrations required for this variant. The reaction was started by the addition of microsomal protein and was stopped by the addition of 2 mL of hexane:isopropanol (3:2; Hara and Radin, 1978) containing 4 µL of unlabeled TAG (triolein from Sigma). Vortexing was done for 10 s to extract lipids, and then phase separation was achieved by adding 1 mL of 500 mm sodium sulfate followed by vortexing for 10 s and waiting 10 min. The upper phase was transferred to a separate tube and dried with nitrogen gas. The lipid was resolubilized in approximately 100 to 150 µL of hexane and applied to K6 silica thin-layer chromatography plates, which were developed in 80:20:1 (v/v) hexane:diethyl ether:acetic acid. TAG was visualized and marked with a pencil by staining briefly in iodine vapor. After the stain faded, the TAG area of the plate was scraped and transferred to scintillation vials to determine radioactivity by liquid scintillation counting. The enzyme kinetics module of SigmaPlot 12.5 was used to determine kinetic parameters. DGAT protein abundance in microsomal membranes was determined by immunoblots probed with rabbit polyclonal antibodies prepared against the CaDGAT1 peptide NGNDGGEKIANGEDR. Yeast microsomes were thawed on ice. Protease inhibitors, 1 µL (Calbiochem catalog no. 539134) diluted 1:1 in 0.5 m EDTA and 0.5 µL of phenylmethylsulfonyl fluoride at 100 mm, were added to tubes on ice followed by 20 µL of yeast microsomes (1–3 µg µL−1 protein), vortexed quickly to mix, and incubated approximately 15 min on ice. 2× SDS sample buffer (NuSep catalog no. BG-165) with 200 mm dithiothreitol was diluted 1:1 in water, and the protease inhibitor-treated microsomes were added to the lids of each of the sample tubes. After all tubes were prepared, samples were centrifuged quickly to mix the microsomes with sample buffer to a final volume of 30 µL. Samples were used for SDS-PAGE, with 4% to 12% Bis-Tris polyacrylamide gels in 1× MES. Proteins were transferred to nitrocellulose membrane for 7 min at approximately 20 V, using the Invitrogen Iblot dry transfer system. The membrane was rinsed briefly in TBST (20 mm Tris-HCl, pH 7.5, 150 mm NaCl, and 0.2% Tween 20) and blocked for 20 min at room temperature in 5% nonfat dry milk/TBST with slow rocking. Primary anti-CaDGAT1 peptide antibodies were added at 1:2,000 in 3% milk/TBST and incubated 1 h at room temperature with slow rocking. Prior to adding secondary antibodies, the membrane was rinsed three times quickly in TBST followed by washing three times for 7 min each in TBST. Secondary horseradish peroxidase-conjugated antibodies at 1:5,000 in 3% milk/TBST were added and incubated for 1 h with slow rocking. The membrane was washed as described previously. SuperSignal West Dura ECL substrate (Pierce catalog no. 34075) was added to the membrane and incubated 5 min. Blots were imaged using a real-time imager, FujiFilm LAS3000 CCD camera. Pixel intensities of the DGAT protein bands were determined using Totallab software by Nonlinear. Standard curves determined with differing amounts of microsomes expressing wild-type CaDGAT1 were used to determine the relative abundance of CaDGAT1 variants. Oil Analysis of Yeast Expressing DGAT Variants Oil content (total fatty acid methyl esters as a percentage of dry weight) and fatty acid composition were determined by quantitative gas chromatography for yeast strains with high Nile Red staining. Approximately 5 to 15 mg of yeast powder from 2-d cultures was weighed into the bottom of a 13- × 100-mm glass culture tube with screw cap and Teflon seal. Five microliters of a stock solution of 17:0 TAG (10 mg mL−1 in toluene) was added followed by the addition of 500 µL of 5% sulfuric acid in methanol (anhydrous). Samples were incubated at 95°C for 1.5 h. Subsequently, tubes were allowed to cool to room temperature, after which 1 mL of 1 m sodium chloride was added followed by mixing. One milliliter of heptane was added, the contents were mixed, and samples were spun briefly to mediate phase separation. Approximately 500 µL of the organic phase was transferred to a gas chromatography vial. Fatty acid methyl esters were analyzed by gas chromatography. Four microliters of heptane extract was analyzed on a Hewlett-Packard 6890 gas chromatograph fitted with an Omegawax 320 fused silica capillary column (Supelco; catalog no. 24152). The oven temperature was programmed to hold at 220°C for 2.7 min, increase to 240°C at 20°C min−1, and then hold for an additional 2.3 min. Carrier gas was supplied by a Whatman hydrogen generator. Retention times were compared with those for methyl esters of commercially available standards (Nu-Chek Prep; catalog no. U-99-A). Expression of DGAT Variants in Soybean Somatic Embryos The expression of DGAT variants in soybean (Glycine max) somatic embryos was described in detail previously (Roesler et al., 2012). Briefly, a soybean expression vector comprising a unique NotI restriction enzyme site (for cloning transgenes) flanked by the promoter for the soybean α′-subunit of β-conglycinin, for seed-specific expression of transgenes (Beachy et al., 1985), and the phaseolin termination sequence (Doyle et al., 1986) as well as a hygromycin B phosphotransferase gene (Gritz and Davies, 1983) under the control of the T7 RNA polymerase promoter (for bacterial selection) and under the control of the cauliflower mosaic virus 35S promoter (for plant selection), were generated as described previously (Cahoon et al., 1999). The wild-type C. americana DGAT gene (CaDGAT1) and variants (A2, C9, C10, and C11) were PCR amplified from the yeast expression vectors with oligonucleotides designed to introduce NotI restriction enzyme sites at the 5′ and 3′ ends of the PCR product. Similarly, GmDGAT1b and variants (C9, C10, C11, and MOD) were designed and synthesized by GENEART with NotI restriction enzyme sites flanking the codon-optimized gene sequence. All DNA fragments containing DGAT1 genes were then cloned into the unique NotI site of the soybean expression vector. Soybean embryogenic suspension cultures were transformed with intact plasmid DNA of soybean expression vector alone or vector plus DGAT1 transgene by the method of particle gun bombardment (Klein et al., 1987) using a DuPont Biolistic PDS1000/HE instrument (helium retrofit). Individual transformed events were selected on hygromycin, and embryos from each event were matured as described previously (Cahoon et al., 1999; Li et al., 2007). Oil Analysis of Soybean Somatic Embryos Oil content (total fatty acid methyl esters as a percentage of dry weight) and fatty acid composition of soybean somatic embryos were determined by quantitative gas chromatography. Dried embryos were ground to a fine powder using a genogrinder vial (0.5- × 2-inch polycarbonate) and a steel ball (SPEX Centriprep). Grinding time was 30 s at 1,450 oscillations min−1. For every event, triplicates of approximately 10 mg of tissue were weighed into Eppendorf tubes. The tissue was extracted using 200 µL of heptane at room temperature under continuous shaking for 2 h. Heptane extracts were cleared by centrifugation, and 25 μL of extract was derivatized to fatty acid methyl esters as follows. One milliliter of a 25% sodium methoxide stock solution was added to 24 mL of HPLC-grade methanol. Sodium methoxide was stored under an inert gas. Five microliters of a 17:0 TAG (Nu-Chek Prep) stock solution (10 mg mL−1) was combined with 25 µL of heptane tissue extract in a glass culture tube, and 500 µL of 1% sodium methoxide was added. Samples were derivatized in a water bath at 50°C for 15 min. Samples were allowed to cool to room temperature, and 1 mL of 1 m NaCl was added followed by brief mixing. Fatty acid methyl esters were extracted into 1 mL of heptane, and 4-µL samples were quantitated by gas chromatography analysis. Means and se were determined for the 10 somatic embryo events with highest oil content for each DGAT variant. Generating Site-Specific Integration Target Event TB5 Transgenic site-specific integration (SSI) target event TB5 was produced as described previously (Li et al., 2009), except that a DNA fragment (QC599A) comprising a hygromycin B phosphotransferase gene (HPT; Gritz and Davies, 1983), flanked by the yeast flippase (FLP)-mediated recombination sites (McLeod et al., 1986) FRT1 (nucleotide sequence GAAGTTCCTATTCCGAAGTTCCTATTCTCTAGAAAGTATAGGAACTTC) and FRT87 (nucleotide sequence GAAGTTCCTATTCCGAAGTTCCTATTCTCCAGAAAGTATAGGAACTTC) and under the control of the constitutive soybean S-adenosyl-l-Met synthetase (SAMS) promoter, was used in the initial transformation. The FRT sites were included for subsequent gene targeting by recombinase-mediated cassette exchange (RMCE; Li et al., 2009, 2010b). Additionally, QC599A contains the nopaline synthase transcription terminator (NOS) after the HPT. The transformation of QC599A into soybean 93B86 embryogenic suspension culture, selection of events on hygromycin, and maturation of somatic embryos were carried out as described herein. Events generated were characterized for single-copy insertions by quantitative PCR (qPCR) and Southern-blot analysis, and flanking sequence was cloned using the Genome Walker kit (Clontech), sequenced using a capillary DNA analyzer and the dye terminator cycle DNA sequencing kit (Applied Biosystems), and the sequence was analyzed by comparison with the cv Williams 82 reference genome (Schmutz et al., 2010) using BLASTN. In this way, an event called TB5 having the SAMS::FRT1::HPT::NOS::FRT87 transgene cassette and flanking sequence with matching borders, compared with the cv Williams 82 genome, was identified (Supplemental Fig. S2A). Somatic embryos from this event were dried and germinated, and homozygous transgenic plants were obtained. Developing seed from these plants was used to generate new embryogenic suspension culture of the TB5 SSI target event. FLP-FRT-Mediated RMCE Comprising GmDGAT1b-MOD (Soil 111) A donor plasmid, pKR2706, was generated having the trait cassette flanked by FRT1 and FRT87 sites. The trait cassette comprised a promoterless selectable marker and the soybean acetolactate synthase gene encoding a mutant ALS enzyme insensitive to sulfonylurea herbicides and having a P178S mutation in the encoded protein (ALS P178S), which would not be expressed unless a promoter was placed upstream. The trait cassette also contained the GmDGAT1b-MOD variant under the control of the soybean α′-subunit of β-conglycinin (Beachy et al., 1985), similar to the vectors described in the somatic embryo events herein. Additionally, pKR2706 contains the acetolactate synthase transcription terminator after the ALS P178S gene and the bean (Phaseolus vulgaris) phaseolin transcription terminator after the GmDGAT1-MOD variant (Supplemental Fig. S2B). FLP plasmid QC663 contained the yeast FLP recombinase under the control of the constitutive soybean EF1A2 promoter (Li et al., 2009, 2015) followed by the potato (Solanum tuberosum) proteinase inhibitor II transcription terminator, which provided the FLP recombinase for transient expression (Supplemental Fig. S2C). Embryogenic suspension cultures of the TB5 SSI target event were cobombarded with the donor plasmid pKR2706 and FLP plasmid QC663 plasmid DNA at a 9:3 ratio following the same biolistic transformation protocol, except using 90 ng mL−1 chlorsulfuron (DuPont) to select retransformation events. Only those events where an RMCE occurred, placing the ALS selectable marker downstream of the SAMS promoter from the target event, along with the GmDGAT1b-MOD, were recovered, similar to that described previously (Supplemental Fig. S2D; Li et al., 2009, 2010b). Chlorsulfuron-resistant events were screened for single copy by qPCR analysis, and embryos were matured and dried as described. Germinated embryos were grown into T0 plants, and T1 seed was obtained (segregating 3:1 for the GmDGAT1b-MOD transgene). T1 seeds were planted and grown, homozygous plants were identified by qPCR analysis, and T2 homozygous seeds were obtained. T1 Seed Analysis by SS-NIR Briefly, the SS-NIR system was designed in such a way as to introduce an individual soybean into the analytical cell, where it was illuminated from all points in three dimensions. This was achieved by tumbling the seed, with an air stream, within an approximated integrating sphere constructed from a 16-mm-diameter quartz cup coated with 6080 white reflectance coating (Labsphere). During 6 s of data collection time, the seed surfaces from all sides are presented to collection optics. Illumination was provided through 12 optical fibers, connected to four 20-W 8211-002 light bulbs (Welch Allyn), the ends of which were incorporated into the cell cover. The reflected spectral region from 904 to 1,686 nm was collected through the apex of the cover of the sampling cell by an NIR512 spectrometer (Control Development). Separate calibration models were generated for every constituent of interest using partial least-squares analysis coupled with an optimized number of latent variables, spectral range, and spectral preprocessing, prior to the online/offline compositional analysis of oil, protein, and oleic acid contents in individual seeds. The optimized number of latent variables, spectral range, and spectral preprocessing were determined by analyzing a training and monitoring subset of calibration data when the calibration performance reached the optimum in terms of root mean error of calibration and root mean error of cross validation. Once a model had been created, a group of samples was removed from the data set and the influence of their omission was assessed. In robust models, the root mean error of cross validation should be similar to root mean error of calibration. T2 Seed Analysis by FT-NIR Spectroscopy Analyses were performed on a Bruker Multi-Purpose Analyzer FT-NIR spectrometer fitted with a 54-mm-diameter rotating cup assembly. Sample sizes of approximately 100 seeds (20 g) were used for the analysis. The weight of each sample (to an accuracy of 0.01 g) was recorded prior to scanning. The reflected spectra were captured for each sample to a wave number resolution of 8 cm−1 (1.5 μm) in the wavelength range between 833 and 2,778 nm, with the instrument in macro-reflectance mode. The cup was rotated over the source and detector while 64 full spectral scans were collected. The rotation of the cup was stopped, and the soybeans were poured into a foil pan and then returned to the cup prior to scanning for a second time. Three full scan cycles (with complete mixing of the sample between each scan) were found to provide good data quality and sample throughput. Captured spectra were analyzed, and models to predict moisture content, oil content, protein content, and oleic acid content were developed using the Bruker OPUS 7.0 software package. The reference chemistry methods used for the calibration of moisture, oil, and protein were based on AOCS official methods (Ac 2-41 [moisture], Ac 3-44(mod) [crude fat/oil], and Ba 4e-93 [crude protein]). The reference chemistry used for the oleic acid calibrations was developed by gas chromatographic analysis of fatty acid methyl esters of oil extracts derived from the soybean samples, after spectral capture. T3 Seed Analysis by NIT Spectroscopy NIR spectra, from 850 to 1,050 nm (2-nm step and 30-mm path length), for 400- to 500-g bulk samples of intact soybeans were acquired in transmission mode using a Foss Tecator AB model 1241 grain analyzer fitted with a standard instrument hopper and sample transport mechanism. Each batch was analyzed in duplicate using 10 subsample scans, which were saved as the average. All data analysis was performed using the InfraSoft International chemometrics software WinISI II version 1.50e (NIRSystems). Pretreatment of the raw near-infrared (log 1/T) spectral data included multiplicative scatter correction and first derivative transformation over a four-point (8-nm) gap using a four-point smoothing function. Predictions of oil and protein content (corrected to a 13% moisture basis) were based on calibration models developed by the U.S. Department of Agriculture Federal Grain Inspection Service/Grain Inspection, Packers, and Stockyards Administration (https://fgis.gipsa.usda.gov/default_home_FGIS.aspx). Calibration models for oleic acid were developed using partial least-squares regression (Williams and Norris, 1987) utilizing the transformed spectrum captured from material representing a wide range of acid oleic content. Other Analytical Methods Carbohydrate contents were developed using the analytical chemistry techniques described by Meyer et al. (2012). Analyses were performed on 20- to 30-mg (to an accuracy of 0.1 mg) aliquots of ground/sifted (1-mm mesh) soybean powders created using a Foss Knifetec 1095 grinder (FOSS North America). All reported values represent means ± sd of three replicates per sample. Controlled Environment Growth Conditions T1 seeds were planted into flats containing peat-based substrate (Fafard 3B). Watering was with tap water as needed for 2.5 weeks after planting. At that point, liquid fertilizer was added at a concentration of 75 ppm nitrogen based on a 15:5:15 nitrogen:phosphorus:potassium fertilizer mix and fertigated as needed. After transplanting, plants were given 150 ppm nitrogen from a 50:50 blend (w/w) of 15:5:15 and 21:5:20 nitrogen:phosphorus:potassium. One to 2 weeks prior to harvest, fertilizer was stopped based on leaf bronzing appearing on mature, fully expanded leaves. Photoperiod was 16 h for 4 weeks (3 weeks in flats and 1 week after transplanting to pots), 14 h for 1 week, and 12 h for the remaining length of time. Supplemental lighting was provided by a 3:1 mixture of metal halide and high-pressure sodium lamps. Lamps provided 230 μmol m−2 s−1 for seedling cultivation and 250 μmol m−2 s−1 during the remaining vegetative and reproductive growth. Plants were harvested on an individual basis based on when they reached 90% brown pods. Temperature for germination and seedling cultivation had 25°C/20°C day/night set points. After transplanting, set points were 27°C/21°C day/night. Field Trial Methods A field trial was performed during the summer of 2015 in Johnston, Iowa, using 10 transgenic and 10 null-segregant T2 seed sources from each of the four events. The planting date of this field trial was on May 9. Transgenic and null segregants within an event were randomized, and every seed source was tested in three replications of rows 2.3 m in length composed of 50 seeds with a 0.76-m planting width among rows. In this manner, the seed compositional performance of 120 rows of soybean lines homozygous for the GmDGAT1b-MOD transgene was compared with that of 120 rows of null segregant isolines devoid of the transgene. Individual rows were harvested with a ALMACO HP5 single-row plot harvester, and seed composition was predicted using NIT spectroscopy. Field Experimental Design and Statistical Analysis Methods A nested field experimental design was adopted to evaluate seed trait performance, where positive and negative blocks were nested within each respective event and positive and negative isolines were randomly nested within each positive and negative block, respectively. Recorded traits included the content of oil, protein, and oleic acid. Least-squares means for positive and null within each event were calculated using a mixed-model analysis method via the residual maximum likelihood software package ASReml (Gilmour et al., 2009). Event and positive and null trait classes were treated as fixed effects, and isolines were fitted as random effects. The spatial variation of first-order autoregressive (AR) correlation structure for rows and autoregressive correlation for columns (AR1 × AR1) was incorporated in the analysis. Mean differences of trait versus null were determined based on Fisher’s lsd approach at a significance level of P < 0.05. The CaDGAT1 GenBank accession number is KU744408. Supplemental Data The following supplemental materials are available. Supplemental Figure S1. Amino acid sequence alignment of CaDGAT1 and 20 plant type 1 DGATs. Supplemental Figure S2. Schematic for DNA constructs and transgenes used to generate the Soil 111 RMCE event. Supplemental Table S1. Description of amino acid substitutions in soybean DGAT. Supplemental Table S2. Seed composition (fatty acids) including minor constituents. Supplemental Table S3. Specifications of models for the prediction of soybean seed composition. ACKNOWLEDGMENTS We thank Tom Davis for assistance with the robotics system, Zhongsen Li for providing the TB5 SSI target site, Linda Castle for helpful DNA shuffling advice, Carl Simmons for designing the codon-optimized soybean DGAT variant nucleotide sequences, Kevin L. Stecca, Mike McCluskey, and Alfred Ciuffetelli for the analysis of soybean composition, and Weilan Pan and Min Ren for the development of software models supporting the spectroscopic analysis of seed composition. Glossary TAG triacylglycerol DAG diacylglycerol cDNA complementary DNA SS-NIR single-seed near-infrared reflectance spectroscopy FT-NIR Fourier-transform near-infrared NIT near-infrared transmission PC phosphatidylcholine SSI site-specific integration RMCE recombinase-mediated cassette exchange qPCR quantitative PCR LITERATURE CITED Agelet LE , Armstrong PR, Romagosa Clariana I, Hurburgh CR ( 2012 ) Measurement of single soybean seed attributes by near-infrared technologies: a comparative study . J Agric Food Chem 60 : 8314 – 8322 Google Scholar Crossref Search ADS PubMed WorldCat Bao X , Ohlrogge J ( 1999 ) Supply of fatty acid is one limiting factor in the accumulation of triacylglycerol in developing embryos . Plant Physiol 120 : 1057 – 1062 Google Scholar Crossref Search ADS PubMed WorldCat Bates PD , Durrett TP, Ohlrogge JB, Pollard M ( 2009 ) Analysis of acyl fluxes through multiple pathways of triacylglycerol synthesis in developing soybean embryos . Plant Physiol 150 : 55 – 72 Google Scholar Crossref Search ADS PubMed WorldCat Beachy RN , Chen ZL, Horsch RB, Rogers SG, Hoffmann NJ, Fraley RT ( 1985 ) Accumulation and assembly of soybean β-conglycinin in seeds of transformed petunia plants . EMBO J 4 : 3047 – 3053 Google Scholar Crossref Search ADS PubMed WorldCat Brink K , Chui CF, Cressman RF, Garcia P, Henderson N, Hong B, Maxwell CA, Meyer K, Mickelson J, Stecca KL, et al. ( 2014 ) Molecular characterization, compositional analysis, and germination evaluation of a high-oleic soybean generated by the suppression of FAD2-1 expression . Crop Sci 54 : 2160 – 2174 Google Scholar Crossref Search ADS WorldCat Burgal J , Shockey J, Lu C, Dyer J, Larson T, Graham I, Browse J ( 2008 ) Metabolic engineering of hydroxy fatty acid production in plants: RcDGAT2 drives dramatic increases in ricinoleate levels in seed oil . Plant Biotechnol J 6 : 819 – 831 Google Scholar Crossref Search ADS PubMed WorldCat Burton J , Brim C ( 1981 ) Recurrent selection in soybeans. III. Selection for increased percent oil in seeds . Crop Sci 21 : 31 – 34 Google Scholar Crossref Search ADS WorldCat Cahoon EB , Carlson TJ, Ripp KG, Schweiger BJ, Cook GA, Hall SE, Kinney AJ ( 1999 ) Biosynthetic origin of conjugated double bonds: production of fatty acid components of high-value drying oils in transgenic soybean embryos . Proc Natl Acad Sci USA 96 : 12935 – 12940 Google Scholar Crossref Search ADS PubMed WorldCat Cao YZ , Huang AH ( 1986 ) Diacylglycerol acyltransferase in maturing oil seeds of maize and other species . Plant Physiol 82 : 813 – 820 Google Scholar Crossref Search ADS PubMed WorldCat Cao YZ , Huang AH ( 1987 ) Acyl coenzyme A preference of diacylglycerol acyltransferase from the maturing seeds of cuphea, maize, rapeseed, and canola . Plant Physiol 84 : 762 – 765 Google Scholar Crossref Search ADS PubMed WorldCat Chen JE , Smith AG ( 2012 ) A look at diacylglycerol acyltransferases (DGATs) in algae . J Biotechnol 162 : 28 – 39 Google Scholar Crossref Search ADS PubMed WorldCat Cheng D , Meegalla RL, He B, Cromley DA, Billheimer JT, Young PR ( 2001 ) Human acyl-CoA:diacylglycerol acyltransferase is a tetrameric protein . Biochem J 359 : 707 – 714 Google Scholar Crossref Search ADS PubMed WorldCat Clemente TE , Cahoon EB ( 2009 ) Soybean oil: genetic approaches for modification of functionality and total content . Plant Physiol 151 : 1030 – 1040 Google Scholar Crossref Search ADS PubMed WorldCat Doyle JJ , Schuler MA, Godette WD, Zenger V, Beachy RN, Slightom JL ( 1986 ) The glycosylated seed storage proteins of Glycine max and Phaseolus vulgaris: structural homologies of genes and proteins . J Biol Chem 261 : 9228 – 9238 Google Scholar Crossref Search ADS PubMed WorldCat Durrett TP , McClosky DD, Tumaney AW, Elzinga DA, Ohlrogge J, Pollard M ( 2010 ) A distinct DGAT with sn-3 acetyltransferase activity that synthesizes unusual, reduced-viscosity oils in Euonymus and transgenic seeds . Proc Natl Acad Sci USA 107 : 9464 – 9469 Google Scholar Crossref Search ADS PubMed WorldCat Eskandari M , Cober ER, Rajcan I ( 2013 ) Genetic control of soybean seed oil. II. QTL and genes that increase oil concentration without decreasing protein or with increased seed yield . Theor Appl Genet 126 : 1677 – 1687 Google Scholar Crossref Search ADS PubMed WorldCat Feng L , Burton J, Carter T, Pantalone V ( 2004 ) Recurrent half-sib selection with testcross evaluation for increased oil content in soybean . Crop Sci 44 : 63 – 69 Google Scholar Crossref Search ADS WorldCat Furmanek T , Demski K, Banaś W, Haslam R, Napier J, Stymne S, Banaś A ( 2014 ) The utilization of the acyl-CoA and the involvement PDAT and DGAT in the biosynthesis of erucic acid-rich triacylglycerols in Crambe seed oil . Lipids 49 : 327 – 333 Google Scholar Crossref Search ADS PubMed WorldCat Gietz RD , Schiestl RH ( 2007 ) Large-scale high-efficiency yeast transformation using the LiAc/SS carrier DNA/PEG method . Nat Protoc 2 : 38 – 41 Google Scholar Crossref Search ADS PubMed WorldCat Gillingham LG , Harris-Janz S, Jones PJ ( 2011 ) Dietary monounsaturated fatty acids are protective against metabolic syndrome and cardiovascular disease risk factors . Lipids 46 : 209 – 228 Google Scholar Crossref Search ADS PubMed WorldCat Gilmour AR , Gogel BJ, Cullis BR, Thompson R ( 2009 ) ASReml User Guide Release 30. VSN International , Hemel Hempstead, UK Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Greenspan P , Mayer EP, Fowler SD ( 1985 ) Nile red: a selective fluorescent stain for intracellular lipid droplets . J Cell Biol 100 : 965 – 973 Google Scholar Crossref Search ADS PubMed WorldCat Griffiths G , Harwood JL ( 1991 ) The regulation of triacylglycerol biosynthesis in cocoa (Theobroma cacao) L . Planta 184 : 279 – 284 Google Scholar Crossref Search ADS PubMed WorldCat Gritz L , Davies J ( 1983 ) Plasmid-encoded hygromycin B resistance: the sequence of hygromycin B phosphotransferase gene and its expression in Escherichia coli and Saccharomyces cerevisiae . Gene 25 : 179 – 188 Google Scholar Crossref Search ADS PubMed WorldCat Haefele D , Sevenich D, Jones D, Janni J, Wright S ( 2007 ) NIR-based analytical systems enable development of corn seed products and corn grain markets that are specific for dry-grind ethanol production . Int Sugar J 109 : 154 – 156 Google Scholar OpenURL Placeholder Text WorldCat Hara A , Radin NS ( 1978 ) Lipid extraction of tissues with a low-toxicity solvent . Anal Biochem 90 : 420 – 426 Google Scholar Crossref Search ADS PubMed WorldCat Jako C , Kumar A, Wei Y, Zou J, Barton DL, Giblin EM, Covello PS, Taylor DC ( 2001 ) Seed-specific over-expression of an Arabidopsis cDNA encoding a diacylglycerol acyltransferase enhances seed oil content and seed weight . Plant Physiol 126 : 861 – 874 Google Scholar Crossref Search ADS PubMed WorldCat Janni J , Weinstock BA, Hagen L, Wright S ( 2008 ) Novel near-infrared sampling apparatus for single kernel analysis of oil content in maize . Appl Spectrosc 62 : 423 – 426 Google Scholar Crossref Search ADS PubMed WorldCat Klein T , Wolf E, Sanford J ( 1987 ) High-velocity microprojectiles for delivering nucleic acids into living cells . Nature 327 : 70 – 73 Google Scholar Crossref Search ADS WorldCat Kroon JT , Wei W, Simon WJ, Slabas AR ( 2006 ) Identification and functional expression of a type 2 acyl-CoA:diacylglycerol acyltransferase (DGAT2) in developing castor bean seeds which has high homology to the major triglyceride biosynthetic enzyme of fungi and animals . Phytochemistry 67 : 2541 – 2549 Google Scholar Crossref Search ADS PubMed WorldCat Lardizabal K , Effertz R, Levering C, Mai J, Pedroso MC, Jury T, Aasen E, Gruys K, Bennett K ( 2008 ) Expression of Umbelopsis ramanniana DGAT2A in seed increases oil in soybean . Plant Physiol 148 : 89 – 96 Google Scholar Crossref Search ADS PubMed WorldCat Lardizabal KD , Mai JT, Wagner NW, Wyrick A, Voelker T, Hawkins DJ ( 2001 ) DGAT2 is a new diacylglycerol acyltransferase gene family: purification, cloning, and expression in insect cells of two polypeptides from Mortierella ramanniana with diacylglycerol acyltransferase activity . J Biol Chem 276 : 38862 – 38869 Google Scholar Crossref Search ADS PubMed WorldCat Larson TR , Graham IA ( 2001 ) A novel technique for the sensitive quantification of acyl CoA esters from plant tissues . Plant J 25 : 115 – 125 Google Scholar PubMed OpenURL Placeholder Text WorldCat Li F , Wu X, Lam P, Bird D, Zheng H, Samuels L, Jetter R, Kunst L ( 2008 ) Identification of the wax ester synthase/acyl-coenzyme A:diacylglycerol acyltransferase WSD1 required for stem wax ester biosynthesis in Arabidopsis . Plant Physiol 148 : 97 – 107 Google Scholar Crossref Search ADS PubMed WorldCat Li R , Yu K, Hatanaka T, Hildebrand DF ( 2010 a ) Vernonia DGATs increase accumulation of epoxy fatty acids in oil . Plant Biotechnol J 8 : 184 – 195 Google Scholar Crossref Search ADS PubMed WorldCat Li Z , Liu ZB, Xing A, Moon BP, Koellhoffer JP, Huang L, Ward RT, Clifton E, Falco SC, Cigan AM ( 2015 ) Cas9-guide RNA directed genome editing in soybean . Plant Physiol 169 : 960 – 970 Google Scholar Crossref Search ADS PubMed WorldCat Li Z , Moon BP, Xing A, Liu ZB, McCardell RP, Damude HG, Falco SC ( 2010 b ) Stacking multiple transgenes at a selected genomic site via repeated recombinase-mediated DNA cassette exchanges . Plant Physiol 154 : 622 – 631 Google Scholar Crossref Search ADS PubMed WorldCat Li Z , Xing A, Moon BP, Burgoyne SA, Guida AD, Liang H, Lee C, Caster CS, Barton JE, Klein TM, et al. ( 2007 ) A Cre/loxP-mediated self-activating gene excision system to produce marker gene free transgenic soybean plants . Plant Mol Biol 65 : 329 – 341 Google Scholar Crossref Search ADS PubMed WorldCat Li Z , Xing A, Moon BP, McCardell RP, Mills K, Falco SC ( 2009 ) Site-specific integration of transgenes in soybean via recombinase-mediated DNA cassette exchange . Plant Physiol 151 : 1087 – 1095 Google Scholar Crossref Search ADS PubMed WorldCat Liu Q , Siloto RM, Lehner R, Stone SJ, Weselake RJ ( 2012 ) Acyl-CoA:diacylglycerol acyltransferase: molecular biology, biochemistry and biotechnology . Prog Lipid Res 51 : 350 – 377 Google Scholar Crossref Search ADS PubMed WorldCat Mañas-Fernández A , Vilches-Ferrón M, Garrido-Cárdenas JA, Belarbi EH, Alonso DL, García-Maroto F ( 2009 ) Cloning and molecular characterization of the acyl-CoA:diacylglycerol acyltransferase 1 (DGAT1) gene from Echium . Lipids 44 : 555 – 568 Google Scholar Crossref Search ADS PubMed WorldCat McFie PJ , Stone SL, Banman SL, Stone SJ ( 2010 ) Topological orientation of acyl-CoA:diacylglycerol acyltransferase-1 (DGAT1) and identification of a putative active site histidine and the role of the N terminus in dimer/tetramer formation . J Biol Chem 285 : 37377 – 37387 Google Scholar Crossref Search ADS PubMed WorldCat McLeod M , Craft S, Broach JR ( 1986 ) Identification of the crossover site during FLP-mediated recombination in the Saccharomyces cerevisiae plasmid 2 microns circle . Mol Cell Biol 6 : 3357 – 3367 Google Scholar Crossref Search ADS PubMed WorldCat Meyer K , Stecca KL, Ewell-Hicks K, Allen SM, Everard JD ( 2012 ) Oil and protein accumulation in developing seeds is influenced by the expression of a cytosolic pyrophosphatase in Arabidopsis . Plant Physiol 159 : 1221 – 1234 Google Scholar Crossref Search ADS PubMed WorldCat Milcamps A , Tumaney AW, Paddock T, Pan DA, Ohlrogge J, Pollard M ( 2005 ) Isolation of a gene encoding a 1,2-diacylglycerol-sn-acetyl-CoA acetyltransferase from developing seeds of Euonymus alatus . J Biol Chem 280 : 5370 – 5377 Google Scholar Crossref Search ADS PubMed WorldCat Morley KL , Kazlauskas RJ ( 2005 ) Improving enzyme properties: when are closer mutations better? Trends Biotechnol 23 : 231 – 237 Google Scholar Crossref Search ADS PubMed WorldCat Ness JE , Kim S, Gottman A, Pak R, Krebber A, Borchert TV, Govindarajan S, Mundorff EC, Minshull J ( 2002 ) Synthetic shuffling expands functional protein diversity by allowing amino acids to recombine independently . Nat Biotechnol 20 : 1251 – 1255 Google Scholar Crossref Search ADS PubMed WorldCat Oakes J , Brackenridge D, Colletti R, Daley M, Hawkins DJ, Xiong H, Mai J, Screen SE, Val D, Lardizabal K, et al. ( 2011 ) Expression of fungal diacylglycerol acyltransferase2 genes to increase kernel oil in maize . Plant Physiol 155 : 1146 – 1157 Google Scholar Crossref Search ADS PubMed WorldCat Padgette SR , Taylor NB, Nida DL, Bailey MR, MacDonald J, Holden LR, Fuchs RL ( 1996 ) The composition of glyphosate-tolerant soybean seeds is equivalent to that of conventional soybeans . J Nutr 126 : 702 – 716 Google Scholar Crossref Search ADS PubMed WorldCat Parcerisa J , Boatella J, Codony R, Rafecas M, Castellote AI, Garcia J, Lopez A, Romero A ( 1995 ) Comparison of fatty acid and triacylglycerol compositions of different hazelnut varieties (Corylus avellana L.) cultivated in Catalonia (Spain) . J Agric Food Chem 43 : 13 – 16 Google Scholar Crossref Search ADS WorldCat Roesler K , Meyer K, Damude HG, Shen B, Li C, Bermudez E, Tarczynski MC, inventors. January 24, 2012. DGAT genes for increased seed storage lipid production and altered fatty acid profiles in oilseed plants. US Patent Application No. 8101819 B2 Saha S , Enugutti B, Rajakumari S, Rajasekharan R ( 2006 ) Cytosolic triacylglycerol biosynthetic pathway in oilseeds: molecular cloning and expression of peanut cytosolic diacylglycerol acyltransferase . Plant Physiol 141 : 1533 – 1543 Google Scholar Crossref Search ADS PubMed WorldCat Schmutz J , Cannon SB, Schlueter J, Ma J, Mitros T, Nelson W, Hyten DL, Song Q, Thelen JJ, Cheng J, et al. ( 2010 ) Genome sequence of the palaeopolyploid soybean . Nature 463 : 178 – 183 Google Scholar Crossref Search ADS PubMed WorldCat Settlage S , Kwanyuen P, Wilson R ( 1998 ) Relation between diacylglycerol acyltransferase activity and oil concentration in soybean . J Am Oil Chem Soc 75 : 775 – 781 Google Scholar Crossref Search ADS WorldCat Shockey JM , Gidda SK, Chapital DC, Kuan JC, Dhanoa PK, Bland JM, Rothstein SJ, Mullen RT, Dyer JM ( 2006 ) Tung tree DGAT1 and DGAT2 have nonredundant functions in triacylglycerol biosynthesis and are localized to different subdomains of the endoplasmic reticulum . Plant Cell 18 : 2294 – 2313 Google Scholar Crossref Search ADS PubMed WorldCat Siloto RM , Madhavji M, Wiehler WB, Burton TL, Boora PS, Laroche A, Weselake RJ ( 2008 ) An N-terminal fragment of mouse DGAT1 binds different acyl-CoAs with varying affinity . Biochem Biophys Res Commun 373 : 350 – 354 Google Scholar Crossref Search ADS PubMed WorldCat Siloto RM , Truksa M, Brownfield D, Good AG, Weselake RJ ( 2009 ) Directed evolution of acyl-CoA:diacylglycerol acyltransferase: development and characterization of Brassica napus DGAT1 mutagenized libraries . Plant Physiol Biochem 47 : 456 – 461 Google Scholar Crossref Search ADS PubMed WorldCat Slack C , Roughan P, Browse J, Gardiner S ( 1985 ) Some properties of cholinephosphotransferase from developing safflower cotyledons . Biochim Biophys Acta 833 : 438 – 448 Google Scholar Crossref Search ADS WorldCat Smith R , Weber C ( 1968 ) Mass selection by specific gravity for protein and oil in soybean populations . Crop Sci 8 : 373 – 377 Google Scholar Crossref Search ADS WorldCat Sonah H , O’Donoughue L, Cober E, Rajcan I, Belzile F ( 2015 ) Identification of loci governing eight agronomic traits using a GBS-GWAS approach and validation by QTL mapping in soya bean . Plant Biotechnol J 13 : 211 – 221 Google Scholar Crossref Search ADS PubMed WorldCat Stemmer WP ( 1994 ) DNA shuffling by random fragmentation and reassembly: in vitro recombination for molecular evolution . Proc Natl Acad Sci USA 91 : 10747 – 10751 Google Scholar Crossref Search ADS PubMed WorldCat Taylor DC , Zhang Y, Kumar A, Francis T, Giblin EM, Barton DL, Ferrie JR, Laroche A, Shah S, Zhu W, et al. ( 2009 ) Molecular modification of triacylglycerol accumulation by over-expression of DGAT1 to produce canola with increased seed oil content under field conditions . Botany 87 : 533 – 543 Google Scholar Crossref Search ADS WorldCat Wang Z , Huang W, Chang J, Sebastian A, Li Y, Li H, Wu X, Zhang B, Meng F, Li W ( 2014 ) Overexpression of SiDGAT1, a gene encoding acyl-CoA:diacylglycerol acyltransferase from Sesamum indicum L. increases oil content in transgenic Arabidopsis and soybean . Plant Cell Tissue Organ Cult 119 : 399 – 410 Google Scholar Crossref Search ADS WorldCat Weselake RJ , Madhavji M, Szarka SJ, Patterson NA, Wiehler WB, Nykiforuk CL, Burton TL, Boora PS, Mosimann SC, Foroud NA, et al. ( 2006 ) Acyl-CoA-binding and self-associating properties of a recombinant 13.3 kDa N-terminal fragment of diacylglycerol acyltransferase-1 from oilseed rape . BMC Biochem 7 : 24 Google Scholar Crossref Search ADS PubMed WorldCat Weselake RJ , Shah S, Tang M, Quant PA, Snyder CL, Furukawa-Stoffer TL, Zhu W, Taylor DC, Zou J, Kumar A, et al. ( 2008 ) Metabolic control analysis is helpful for informed genetic manipulation of oilseed rape (Brassica napus) to increase seed oil content . J Exp Bot 59 : 3543 – 3549 Google Scholar Crossref Search ADS PubMed WorldCat Williams P , Norris K ( 1987 ) Near-Infrared Technology in the Agricultural and Food Industries. American Association of Cereal Chemists St. Paul, MN Google Scholar Google Preview OpenURL Placeholder Text WorldCat COPAC Xu J , Francis T, Mietkiewska E, Giblin EM, Barton DL, Zhang Y, Zhang M, Taylor DC ( 2008 ) Cloning and characterization of an acyl-CoA-dependent diacylglycerol acyltransferase 1 (DGAT1) gene from Tropaeolum majus, and a study of the functional motifs of the DGAT protein using site-directed mutagenesis to modify enzyme activity and oil content . Plant Biotechnol J 6 : 799 – 818 Google Scholar Crossref Search ADS PubMed WorldCat Zhang C , Iskandarov U, Klotz ET, Stevens RL, Cahoon RE, Nazarenus TJ, Pereira SL, Cahoon EB ( 2013 ) A thraustochytrid diacylglycerol acyltransferase 2 with broad substrate specificity strongly increases oleic acid content in engineered Arabidopsis thaliana seeds . J Exp Bot 64 : 3189 – 3200 Google Scholar Crossref Search ADS PubMed WorldCat Zheng P , Allen WB, Roesler K, Williams ME, Zhang S, Li J, Glassman K, Ranch J, Nubel D, Solawetz W, et al. ( 2008 ) A phenylalanine in DGAT is a key determinant of oil content and composition in maize . Nat Genet 40 : 367 – 372 Google Scholar Crossref Search ADS PubMed WorldCat Author notes 1 Present address: DuPont Pioneer, 4010 Point Eden Way, Hayward, CA 94545. 2 Present address: Syngenta Biotechnology China Co., Ltd., No. 25 Life Science Park Road, Changing District, Beijing 102206, People’s Republic of China. * Address correspondence to keith.roesler@pioneer.com and knut.meyer@usa.dupont.com. The author responsible for distribution of materials integral to the findings presented in this article in accordance with the policy in the Instructions for Authors (www.plantphysiology.org) is: Keith Roesler (keith.roesler@pioneer.com). B.S. proposed DGAT DNA shuffling and supervised that part of the project; K.R. designed screens, did screening and activity assays, and wrote the article with K.M., using input from all authors; E.B. prepared DGAT variant libraries; C.L. obtained CaDGAT1 cDNA, designed screens, and did screening; J.H. quantitated DGAT protein abundance; H.G.D. designed soy amino acid variants; H.G.D. and K.G.R. directed and performed soybean vector construction and transformation; J.R.B. designed and executed plant characterization in controlled environments; J.D.E. directed spectroscopic calibrations supporting seed analysis; L.C. designed and executed field trials; L.F. performed statistical analysis of field trial data; K.M. supervised the transgenic plant research. [OPEN] Articles can be viewed without a subscription. www.plantphysiol.org/cgi/doi/10.1104/pp.16.00315 © 2016 American Society of Plant Biologists. All Rights Reserved. © The Author(s) 2016. Published by Oxford University Press on behalf of American Society of Plant Biologists. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse, distribution, and reproduction in any medium, provided the original work is properly cited.
TI - An Improved Variant of Soybean Type 1 Diacylglycerol Acyltransferase Increases the Oil Content and Decreases the Soluble Carbohydrate Content of Soybeans
JF - Plant Physiology
DO - 10.1104/pp.16.00315
DA - 2016-06-10
UR - https://www.deepdyve.com/lp/oxford-university-press/an-improved-variant-of-soybean-type-1-diacylglycerol-acyltransferase-4yhLL7r81s
SP - 878
EP - 893
VL - 171
IS - 2
DP - DeepDyve
ER -