Increasing Renewable Oil Content and Utility

University of Kentucky UKnowledge

Theses and Dissertations--Plant and Soil Sciences Plant and Soil Sciences

2017

INCREASING RENEWABLE OIL CONTENT AND UTILITY

William Richard Serson University of Kentucky, [email protected] Digital Object Identifier: https://doi.org/10.13023/ETD.2017.243

Right click to open a feedback form in a new tab to let us know how this document benefits ou.y

Recommended Citation Serson, William Richard, "INCREASING RENEWABLE OIL CONTENT AND UTILITY" (2017). Theses and Dissertations--Plant and Soil Sciences. 89. https://uknowledge.uky.edu/pss_etds/89

This Doctoral Dissertation is brought to you for free and open access by the Plant and Soil Sciences at UKnowledge. It has been accepted for inclusion in Theses and Dissertations--Plant and Soil Sciences by an authorized administrator of UKnowledge. For more information, please contact [email protected]. STUDENT AGREEMENT:

I represent that my thesis or dissertation and abstract are my original work. Proper attribution has been given to all outside sources. I understand that I am solely responsible for obtaining any needed copyright permissions. I have obtained needed written permission statement(s) from the owner(s) of each third-party copyrighted matter to be included in my work, allowing electronic distribution (if such use is not permitted by the fair use doctrine) which will be submitted to UKnowledge as Additional File.

I hereby grant to The University of Kentucky and its agents the irrevocable, non-exclusive, and royalty-free license to archive and make accessible my work in whole or in part in all forms of media, now or hereafter known. I agree that the document mentioned above may be made available immediately for worldwide access unless an embargo applies.

I retain all other ownership rights to the copyright of my work. I also retain the right to use in future works (such as articles or books) all or part of my work. I understand that I am free to register the copyright to my work.

REVIEW, APPROVAL AND ACCEPTANCE

The document mentioned above has been reviewed and accepted by the student’s advisor, on behalf of the advisory committee, and by the Director of Graduate Studies (DGS), on behalf of the program; we verify that this is the final, approved version of the student’s thesis including all changes required by the advisory committee. The undersigned agree to abide by the statements above.

William Richard Serson, Student

Dr. David Hildebrand, Major Professor

Dr. Arthur Hunt, Director of Graduate Studies

INCREASING RENEWABLE OIL CONTENT AND UTILITY

______

DISSERTATION ______

A dissertation submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in the College of Agriculture, Food and Environment at the University of Kentucky

By William Richard Serson

Lexington, Kentucky

Director: Dr. David Hildebrand, Professor of Plant and Soil Sciences

Lexington, Kentucky

2017

ABSTRACT OF DISSERTATION

INCREASING RENEWABLE OIL CONTENT AND UTILITY

Since the dawn of agriculture man has been genetically modifying crop plants to increase yield, quality and utility. In addition to selective breeding and hybridization we can utilize mutant populations and biotechnology to have greater control over crop plant modification than ever before. Increasing the production of plant oils such as soybean oil as a renewable resource for food and fuel is valuable. Successful breeding for higher oil levels in soybean, however, usually results in reduced protein, a second valuable seed component. We show that by manipulating a highly active acyl-CoA: diacylglycerol acyltransferase (DGAT) the hydrocarbon flux to oil in oilseeds can be increased without reducing the protein component. Compared to other plant DGATs, a DGAT from Vernonia galamensis (VgDGAT1A) produces much higher oil synthesis and accumulation activity in yeast, insect cells and soybean. Soybean lines expressing VgDGAT1A show a 4% increase in oil content without reductions in seed protein contents or yield per unit land area. Furthermore, we have screened a soybean fast neutrino population derived from M92-220 variety and found three high oil mutants that do not have reduced levels of protein. From the F2 plant populations we quantitatively pooled the high oil and low oil plants and performed comparative genomics hybridization (CGH). From the data it appears that two families have a 0.3 kb aberration in chromosome 14. We are performing further analysis to study this aberration and develop markers for molecular breeding. Mutagenic techniques are also useful for developing other traits such as early flowering varieties and adapting new high oil crops to a new region. Chia (Salvia hispanica) is an ancient crop that has experienced an agricultural resurgence in recent decades due to the high omega 3 fatty acid (ω-3) content of the seeds and good production potential. The area of cultivation has been expanded to Kentucky using mutagenized populations and the composition traits are similar to that of the original regions of cultivation in Central and South America.

KEYWORDS: diacylglycerol acyltransferase, triacylglycerides, mutagenesis, comparative genomics hybridization, chia, soybean

____William R Serson

June 28, 2017

INCREASING RENEWABLE OIL CONTENT AND UTILITY

William R. Serson

David Hildebrand______Director of Dissertation

Arthur Hunt ______Director of Graduate Studies

June 28, 2017______

TABLE OF CONTENTS LIST OF TABLES ...... v LIST OF FIGURES ...... vi Chapter 1 INTRODUCTION ...... 1 Soybean Oil Content and Composition ...... 1 Chia: Salvia hispanica ...... 10 Chapter 2 VgDGAT1A INCREASES RENEWABLE OIL CONTENT AND OLEIC ACID CONTENT IN SOYBEANS (Glycine max) ...... 20 Introduction ...... 20 Materials and Methods ...... 22 cDNA cloning ...... 22 Expression in insect cells ...... 23 Yeast microsome assays ...... 23 Lipid analysis ...... 24 Seed-specific expression vector construction and soybean somatic embryo transformation ...... 26 Segregating Line Analysis ...... 28 Results ...... 28 In vitro assays of VgDGAT1A ...... 28 Oil and protein content of transgenic soybeans expressing VgDGAT1A ...... 29 Segregating Line Analysis ...... 30 Fatty acid composition in high oil lines ...... 30 Discussion ...... 31 Chapter 3 A LARGE COPY NUMBER VARATION ON CHROMOSOME 14 APPEARS TO INCREASE OIL CONTENT IN SOYBEANS ...... 41 Introduction ...... 41 Materials and Methods ...... 43 Genetic Material ...... 43 Backcrossing ...... 43 Tissue Sampling, Seed Composition and DNA Extraction ...... 43 NMR Methods ...... 44 CGH Analysis ...... 44 Results ...... 44 Oil and Protein Content of Backcrosses ...... 44 CGH Analysis ...... 45 Discussion ...... 45 Chapter 4 COMPOSITION ANALYSIS OF A GAMMA MUTAGENIZED EARLY FLOWERING CHIA POPULATION AND DEVELOPMENT OF NIRS CALIBRATIONS FOR CHIA SEED OIL, PROTEIN AND MOISTURE LEVELS ...... 60 Introduction ...... 60 Materials and Methods ...... 63 Chia Seed Samples ...... 63 Nitrogen and Phosphorus Analysis ...... 63 Trace-element analysis ...... 65 Fiber analysis ...... 66 Fatty Acid Analysis ...... 66 Oil Extraction ...... 67 Phytate Analysis...... 67 NIR Methods and Prediction Model Development ...... 68

iii

NMR Methods ...... 68 Amino Acid Analysis ...... 69 Results ...... 69 Constituent Analysis ...... 69 NIR modeling...... 72 Nitrogen rate assessment ...... 73 Discussion ...... 73 Chapter 5 CONCLUSIONS AND FUTURE DIRECTIONS ...... 89 VgDGAT1A Increases Oil Content and Oleic Acid Content in Soybeans ...... 89 A Deletion Event on Chromosome 14 May Increase Oil Content in Soybeans ...... 90 Development of Oil, Protein and Moisture NIRS Calibrations for Chia Seeds ...... 91 Works Cited ...... 93 Vita...... 104

LIST OF TABLES Table 2.1. Summary Statistics of Oil and Protein...... 35 Table 2.2. Fatty Acid Composition of High Oil Lines ...... 36 Table 3.1. Oil Content of Parent Lines...... 54 Table 3.2. Oil Content of F1 Seeds...... 55 Table 3.3.Plants Selected for CGH AnalysiS ...... 56 Table 3.4. Genes Present in Deletion ...... 59 Table 4.1. Fatty Acid Profile of High Omega-3 Seeds ...... 78 Table 4.2. Summary Statistics of Constituent Analysis of Seed Lots ...... 79 Table 4.3. Summary Data of Extended-Analysis Samples ...... 80 Table 4.4. Summary Statistics for Trace Essential Elements ...... 81 Table 4.5. Mineral Composition of Extended Analysis Samples ...... 82 Table 4.6. Potentially Toxic Trace Elements in Chia Seeds...... 83

LIST OF FIGURES Figure 2.1. TAG Biosynthetic Activity of DGAT1s Expressed in Yeast...... 37 Figure 2.2. Protein and Oil Levels of High Performing VgDGAT1A Lines ...... 38 Figure 2.3. Sequence Alignment of DGATs...... 39 Figure 2.4. Price Trends of Source Oils ...... 40 Figure 3.1: Scatter Plot of Oil and Protein Contend of F2 families...... 57 Figure 3.2: CNVs in Chromosome 14 ...... 58 Figure 4.1. NIR profile of Chia Seeds...... 84 Figure 4.2. Oil Prediction Models of (a) Whole and (b) Ground Seeds...... 85 Figure 4.3. Protein Prediction Models of (a) Whole and (b) Ground Seeds ...... 86 Figure 4.4. Moisture Prediction Model for Whole Seeds ...... 87 Figure 4.5. Effect of Nitrogen Application on Proteina nd Oil Accumulation...... 88

Chapter 1 INTRODUCTION

Soybean Oil Content and Composition

In the past twenty years the world has begun to witness a variety of challenges directly stemming from the use of non-renewable fossil fuels, especially climate change and an impending energy crisis. In light of these concerns, many resources have been

poured into the plant biology community to understand and harness the metabolic

pathways which regulate TAG accumulation in plant tissues to provide a renewable and

carbon neutral fuel source for sustainable energy production. The ultimate goal of this

research is to incorporate traits which lead to high oil accumulation into modern crop

species and make them amenable to large scale agricultural production. Due to its large

scale production and oil accumulation, soybean (Glycine max) has been envisioned as

one of the ideal species to incorporate such traits.

Soybean is the second most important crop in the United States with a total

production of 91 million tons in 2013(da Silva and de Almeida D’Agosto, 2013). The

future demand for soybean is predicted to further increase not only from a growing world

population, but also from increase in meat consumption and aquaculture throughout the

world. A need for sustainable fuel and chemical feedstock sources adds pressure to

increase soybean production for biodiesel and renewable chemicals.

The sum of two major components of commercial soybeans, oil and protein,

generally ranges around 60 ~ 64% total dry weight of the seed (Piper and Boote, 1999).

The oil plus protein contents of several edible nuts exceed 75% such as 81% for Brazil

nuts (67% oil and 14% protein), 78% for walnuts (65% oil and 13% protein) and 75% for

pine nuts (62% oil and 13% protein) (Venkatachalam and Sathe, 2006). Therefore an

increase in the oil plus protein content of soybean seeds should be possible without compromising seed and crop viability.

It is well documented that the oil content and protein content of many oil seeds are negatively correlated (Burton, 1987; Wilcox and Shibles, 2001) and influenced by both maternal and seed genotypes (Hayati et al., 1996; Hernández-Sebastià et al., 2005;

Pandurangan et al., 2012; Pipolo et al., 2004). In general, higher supply and uptake of nitrogenous assimilates results in higher protein but low oil profiles and vice versa. In spite of increased knowledge and insight into the biosynthesis of oil and protein, understanding and manipulating the mechanisms regulating protein and oil content of seeds has been elusive (Schwender and Hay, 2012).

While soybeans are valued for their oil, their place as the global leader in protein production will remain unchallenged for decades. Soybean seeds contain 40-41% protein on a dry weight basis, and the main goal for genetic improvement is to produce varieties with high protein content and improved amino acids composition for human and livestock consumption. Soybean lines with up to 57.9% protein on seed dry weight bases can be found in USDA soybean germplasm collection, but the high protein lines are characterized by smaller number of seeds per plant and significantly lower oil content

(Hill JL et al., 2005; Peregrine et al., 2008). Generally those who try to increase oil content in soybeans are not particularly interested in modifying protein, as long as its levels are not reduced.

While central metabolic processes can never by their nature be solely devoted to oil biosynthesis, they provide all necessary precursors to TAG production and therefore ought to be examined. The two key precursors provided by primary metabolism are G3P

2 and acetyl-CoA. G3P has been demonstrated as a potential limiting product, as overexpression of a G3P phosphate dehydrogenase in canola led to a 40% increase in final seed oil content (Vigeolas et al., 2007). Going even further upstream, we discover that these resources are provided by sucrose, the major form which assimilated carbon is shuttled to various pathways, or sinks, such as oil biosynthesis, and the primary carbon resource provided by the maternal plant for developing seeds (Baud and Lepiniec, 2010).

Once it reaches the seed though, it is frequently interconverted, especially in the first 30 to 40 days of seed development, to starch. Ultimately in normal soybean lines, starch reserves drop to >1% of total seed content, the starch being mobilized for energy and carbon precursor to provide energy for the developing seed (Medic et al., 2014). Levels of starch in commercial soybeans has been found to be 10-11% of total seed composition

20 days before harvest occurs (Stevenson et al., 2006). In a time course study, starch was found to accumulate to a maximum of 14% dry weight content around week eight of development and then rapidly decline to 1% by harvest in week 10 (Saldivar et al., 2011).

In the same study, rapid accumulation of oil occurs until week 8, and some high oil cultivars continue to accumulate oil until harvest at week 10. It is possible that these cultivars are able to utilize starch resources available near the end of the seed filling period, directing carbohydrate resources away from the production of galactooligosaccharides, which are generally undesirable products with no economic value (Medic et al., 2014). Ultimately, eliminating anti-nutritional compounds such as oligosaccharides and increasing levels of oil is the ideal outcome of a metabolic engineering scheme.

Another major factor affecting oil and protein accumulation in oilseeds is

competition for resources from the mother plant absorbed through the funiculus. A key

branch point in the partitioning of hydrocarbon (photosynthate) between triacylglycerol

(TAG), the main component of plant oil, and other storage components is the conversion

of phosphoenolpyruvate (PEP) by PEP carboxylase (PEPC) or pyruvate kinase (PK).

PEPC is important in supplying carbon skeletons to amino acid synthesis and its high

activity is co-related with high protein content (Smith et al., 1989; Sugimoto et al., 1989).

When a bacterial PEPC is overexpressed in V. narbonensis in a seed-specific manner, the

transgenic seeds accumulated up to 20% more protein per gram seed dry weight and

reached a higher seed weight (Rolletschek et al., 2004). On the other hand, a reduction of

PK activity, especially of plastidic PK, results in a reduction of TAG accumulation

(Andre et al., 2007; Baud et al., 2007b; Lonien and Schwender, 2009).

The biosynthetic pathway of TAG production, and the enzymes governing it, has been well characterized. Fatty acids are mostly produced in chloroplasts and the endoplasmic reticulum from the 2C precursor acetyl-CoA. Acetyl CoAs are converted to malonyl-CoA by acetyl-CoA carboxylase, then transferred into malonyl-ACP. These units are added one at a time by the KAS family of enzymes to a growing chain, leading to 16:0, 18:0 and 18:1 units, which can be transported to the endoplasmic reticulum where they can be further desaturated into 16:1, 18:2 and 18:3 (Sidorov and

Tsydendambaev, 2014). The Kennedy pathway then assembles fatty acids onto a glycerol-3-phosphate (G3P) via G3Pacyltransferase, forming lysophosphatidic acid

(LPA). Another FA is tacked on using LPA acyltransferase (LPAAT) and phosphate groups are liberated, leading to the formation of a diacylglycerol (DAG). DAG

acyltransferase (DGAT), the only committed step to TAG biosynthesis, then converts

DAG into TAG using acyl-CoA as a donor of the acyl residue (Sidorov and

Tsydendambaev, 2014). As expected, some key regulation points have been found in this

pathway, especially the activity of DGAT (Li et al., 2013).

The final and only committed step of TAG biosynthesis is catalyzed by two enzymes: DGAT and phosphatidylcholine:diacylglycerol acyltransferase (PDAT), with the former playing the major role (Liu et al., 2012; Zhang et al., 2009). It was recently reported that expression of a Tropaeolum majus DGAT1, TmDGAT1, increases seed size

and yield in Arabidopsis in addition to increasing oil content (Xu et al., 2008). Similar

increases in oil content and seed yield were reported with higher expression of a Brassica

napus (canola) DGAT1, BnDGAT1, back in B. napus (Weselake et al., 2008). They also

report enhanced drought tolerance of these transgenic canola with increased DGAT

expression. Lardizabal et al. (2008b) also report a small but significant increase in oil

levels of transgenic soybeans expressing a fungal DGAT2 and an increase in seed yields

in four out five locations in the US and Argentina although these yield increases were not

significant.

Expression of many genes involved in biosynthetic pathways, such as storage

protein synthesis (Verdier and Thompson, 2008) and/or developmental changes like seed development (Santos-Mendoza et al., 2008) are often regulated by small numbers of

transcription factors (TFs) which can directly or indirectly affect seed oil accumulation.

For seed maturation, LEC1, LEC2, FUS3 and ABI3 are considered the master TFs. They

further interact with several other TFs such as WRINKLED1 (WRI1) to control many

aspects of seed maturation. A series of experiments by various labs recently demonstrated

5 that WRI1 is involved in the expression of key genes in fatty acid synthesis except for final steps in oil biosynthesis such as diacylglycerol acyltransferase (DGAT) (Baud and

Lepiniec, 2009; Baud et al., 2007a; Cernac and Benning, 2004; Lonien and Schwender,

2009; Maeo et al., 2009; To et al., 2012). When AtWRI1 or its maize ortholog is overexpressed in Arabidopsis and maize, seed oil content increased without affecting plant growth or seed production (Liu et al., 2010; Shen et al., 2010). However, it has also been reported (Baud et al., 2009; To et al., 2012) that the overexpression of Arabidopsis WRI1 using several seed-specific promoters in Arabidopsis failed to increase oil content. As mentioned above, DGATs apparently are not regulated by WRI1. The combination of increased fatty acid biosynthesis such as by WRI1 and final TAG synthesis such as by

DGAT can further increase oil levels (Vanhercke et al., 2013). Regulation of oil biosynthesis via transcription factors has proven to be deeply complex, but stands to be one of our best chances to improve oil content and composition.

With all of this knowledge about oil biosynthesis and its regulation available, many resources have been dedicated to producing genetically engineered plants with the ultimate goal being to increase total oil content. The mobilization of carbon resources into the TAG biosynthesis pathway is complicated at best, and the most effective schemes utilize several transgenes which act at key points in carbon metabolism, by utilizing a push-pull-protect strategy (Vanhercke et al., 2014). For instance, WRI1 can push carbon into storage lipid pathways (Gutierrez et al., 2007), then enzymes such as diacylglycerol acyltransferases (DGATs), the only committed step of oil biosynthesis, create TAGs using diacylglycerols (Sidorov and Tsydendambaev, 2014). To protect the oil made, RNAi which knocks out lipases, enzymes which degrade TAGs, can also be

incorporated (Vanhercke et al., 2014). These “protection” mechanisms are more recently

discovered but nonetheless provide another avenue for preserving oil already made by the

plant, thus appearing as an increase in overall oil content compared to wild types.

It is well known that seed oil levels peak at the end of the maturation stage, but

decline slightly during the desiccation stage (Baud et al., 2002; Chia et al., 2005; Murphy

and Cummins, 1989). SDP1 is a triacylglycerol lipase responsible for initiating oil

breakdown in the germinating seeds (Eastmond, 2006), but it has shown some activity

during desiccation and contributes to the loss of the storage reserve during desiccation.

Silencing of SDP1 has been shown to increase oil accumulation levels in Brassica napus

and Arabidopsis (Kelly et al., 2013; van Erp et al., 2014). Especially in Arabidopsis, its

additive effect along with WRI1 and DGAT1 overexpression was shown (van Erp et al.,

2014). Schemes which put together several regulation points seem to increase oil

synergistically.

While these modern approaches have been effective at precisely modifying oil

content and composition positively, they remain tightly regulated and restricted use due

to lobby efforts by “anti-GMO” activists distrust from the public, in spite of their proven

safety time and time again. However, non-transgenic approaches remain relatively

unregulated and with modern genetics technology and sequencing technologies, it is

possible to precisely modify plant genomes for quantitative traits such as oil content.

Among growers, altering management strategies and cultivar selection remain the most

popular ways to increase oil content on a land-area basis.

Cultivar effects are well understood to effect protein content and amino acids in soybeans. For instance, N6202 germplasm line produced seeds with 45.7% protein

content, and a 10% reduction in grain yield compared to a control variety, NC-Roy

(Carter TE et al., 2010). Similar results were found in TN03-350 and TN04-5321, which

achieved 43.1-43.9% protein content without sacrificing seed yield. In soybean genotypes

of early maturity groups, average to high protein content (399-476g/ kg-1) was found in years with high air temperature and moderate rates of rainfall during the seed-filling period, whereas seed protein content was drastically reduced (265±347 g /kg-1) in seasons

of insufficient nitrogen fixation or higher amounts of precipitation during seed filling

(Vollmann et al., 2000). Increasing sink strength results in increased oil and protein with

a strong pronounced effect on protein and less on oils (Rotundo et al., 2011). This was

demonstrated by using depodding to increase supply and shading to decrease it, and it

was found that a low availability of assimilate leads to a significant drop in oil and much

less of an impact in protein.

Temperature during seed fill can also alter oil quality and composition. Growth in

high temperatures will result in higher oil saturation, meaning more oleic acid (18:1) and

less linoleic and linolenic acids (18:2 and 18:3). Growth in a low temperature

environment produces the opposite effect, with lower levels of 18:1 oils and high levels

of 18:2 and 18:3 oils (Rennie and Tanner, 1989). When temperatures are high during seed

filling periods one can also expect levels of oil content in the seed to increase (Kane et

al., 1997). While these factors are out of the control of growers, they may be another

factor plant biologists can use to understand oil composition regulation.

Some management strategies, especially planting date, can also affect oil and

protein content. A study using early maturing cultivars found that an early planting

increased oil content while delayed planting increases protein content. This is mostly

related to temperature during the seed filling period; in this study the early planted

soybeans were subjected to higher temperatures during seed filling than the late planted

soybeans (Kane et al., 1997). Herbicides have also been brought into question in regards

to effects on seed composition (Browman et al., 1986). Since then, others have found that

herbicides had no or little effect on seed quality parameters such as oil and protein

(Bradley et al., 2002). Above all, management strategies and selection of appropriate

genotypes are the most popular and easiest ways for growers to control seed quality and

yields.

In plant breeding random mutagenesis, such as with the use of a chemical

mutagen like ethyl methane sulfonate (EMS) or bombardment with gamma radiation, is a

common way to generate mutations and increase genetic diversity for traits with limited

natural variation. While these methods can produce useful traits, such as an early

flowering mutant or increased oil content, it is possible that other important genes could

be mutated; causing undesirable phenotypes such as altered seed composition. However,

mutagenized populations are not brought under the same scrutiny as transgenic

approaches and therefore traits induced this way are much easier to incorporate into

existing germplasm. Therefore in addition to increasing oil content via a transgenic

approach with the VgDGAT1A trait, we have also utilized a fast neutron (FN) population

of soybeans which exhibit 3 to 4% higher oil content than the parent variety with little or no decrease in protein. The fast neutron soybean mutant population is one of the newest available, and has added power with the large variation in mutant sizes, from several bp deletions up to Mb sized deletion events (Bolon et al., 2011).

Once mutant populations are established, modern sequencing tool s can be used in

conjunction to pinpoint deletions and use the probe data to develop markers for marker

assisted selection techniques. Once a genetic marker is developed it can be used to follow

the trait in a breeding program. This method was used to confirm a FAD2 gene deletion

in the parent line, and the high oleic acid content correlated directly with the presence or

absence of a PCR marker when wild type DNA was used as a template (Bolon et al.,

2011).

Chia: Salvia hispanica

Chia has experienced an agricultural resurgence in recent years. Although

introduced to the American public in the 1980s as “Chia Pets®” with sprouts forming

green “hair”, this crop served as a vital food staple and experienced cultural and

economic significance in pre-Columbian Mesoamerica among tribal peoples such as the

Aztecs (Cahill, 2003). Chia is also known as Spanish sage and the roasted and ground

seeds, known as chiapinolli, were eaten as a gruel and the oil was used as a body

emollient in those ancient cultures (Reyes-Caudillo et al., 2008). During the Spanish conquest, cultural differences caused a decline in the cultivation of chia as it was replaced with European alternatives (Acuña, 1986). Chia production has increased significantly in the past two decades due to its potential as a renewable, low impact source of omega-3 fatty acids (ω-3 FA), having one of the highest level of ω-3 FAs of any known crop plant

(Cahill, 2004).

Chia’s most valued biomass product is the nutlets, or dried fruit, it produces. The seeds are egg shaped and flow easily in bulk handling, making chia seeds similar to canola among major oilseeds, with mean seed mass ranging from 1 to 1.3 mg/seed

(Ixtaina et al., 2008). So called chia “seeds” are actually dried fruits, or nutlets, but are handled and treated like seeds (Bueno et al., 2010). The seed (fruit) pericarp has three distinct cell layers between the transparent cuticle and testa. There is a single layer of epicarp cells, the mesocarp composed of several amorphous cell layers and the sclerenchyma, which is also composed of 3 cell types including the inner layer, the endocarp (Hedge, 1970). Chia seed coat colors are white or blue-grey with black spots.

Brown or tan seeds are also common and are apparently due to incompletely matured seeds in which full seed coat pigmentation is not developed. Seed coat color (except for brown seeds) is genetically determined with white coat color being recessive although traditional production in Mexico and Central America included both seed coat genotypes in the same field.

Chia (Salvia hispanica L.) is a member of the Laminaceae family. The center of genetic diversity of chia is in the highlands of western Mexico (Cahill, 2004), with a diploid chromosome number of 12 (Estilai et al., 1990). The genus Salvia contains about

900 known species with some species being cultivated and used globally in folk medicine

(Lu and Foo, 2002). Stems are about 1-2 m and obtusely quadrangular. Leaves are opposite, ovate, tapered and sharply serrated. The flowers are produced in terminal and axillary four-cornered spikes protected by small bracts with long sharp points. The blue or white corolla is tubular with four stamens, two of them larger and sterile. Seeds are present in groups of four.

Until recently, chia production was limited to regions of California and Arizona in the United States, due to its photoperiod requirements (day length of 12.5 hours or less for at least two months before frost), and only with the help of irrigation. Outside of the

United States, chia can be grown in most of Mexico and parts of Central and South

America, including Nicaragua, Chile, Argentina and Ecuador (Cahill, 2004). Recently,

long day flowering mutant chia lines have been developed. These mutants flower under

long-day lengths, much earlier in a temperate growing season than chia populations from

tropical latitudes, meaning these lines can reach maturity before frost damage can occur

on maturing plants and seeds (Jamboonsri et al., 2012). While management practices

continue to be optimized for this crop in Kentucky, great potential exists for chia as a

major ω-3 fatty acid (ω-3 FA) source grown in the heartland of the United States.

Many members of this family exhibit high levels of ω-3 FAs. Chia has oil

composition levels similar to its Lamiaceae relative Perilla frutescens (Ciftci et al., 2012) and Dracocephalum moldavica, averaging 68% ω-3 FA. However, only chia and perilla have high seed and oil yield potential, making them two of the most economically feasible members of the family to develop for large scale crop production (Rao et al.,

2008). Chia seeds contain about 20% protein, and oil content ranges from 28.5-32.7%

(Ayerza and Coates, 2004; Ayerza and Coates, 2007). Chia is high in the ω-3 FA α- linolenic acid, 18:3. There are claims that white chia seeds are higher in oil and ω-3 FA levels than dark seeds, though genes regulating seed pigmentation are independent from those regulating oil composition (Ayerza, 2010).

Chia oil has high oxidation potential due to its very polyunsaturated FA nature, but the crude oil obtained by cold pressing initially has a low peroxide index value (2.6 mEq peroxide/kg) as well as high antioxidant activity due to the presence of phenolic compounds, mainly, myricetin, quercetin, kaempferol, chlorogenic acid, 3,4- dihydroxyphenylethanol-elenolic acid dialdehyde (3,4-DHPEA-EDA), tocopherols,

(Capitani et al., 2012; Marineli et al., 2014; Reyes-Caudillo et al., 2008; Sargi et al.,

2013) as well as other antioxidants such as carotenoids (Alvarez-Chavez et al., 2008;

Álvarez et al., 2008; Ixtaina et al., 2011). The carotenoid content in chia varies from 0.5

mg/kg to 1.2 mg/kg (Ixtaina et al., 2011). Chia seed oil contains about 238–427 mg/kg

tocopherols, mainly γ-tocopherol, which is similar to levels found in peanut, 398.6 mg/kg

(Ixtaina et al., 2011), though less than soybean crude oil which typically ranges from

1,205-2,195 mg/kg on a dry basis (Medic et al., 2014) and flax oil, approximately 800

mg/kg oil (Bozan and Temelli, 2008). Tocopherols consist of four forms, α, β, γ and δ

tocopherol (Kamal-Eldin and Appelqvist, 1996). α-Tocopherol has been studied

extensively for its antioxidant activity mainly due to its high uptake (Brigelius-Flohé,

2006; Brigelius-Flohe and Traber, 1999; Zingg and Azzi, 2004). γ-Tocopherol has been reported to play an important role in reducing the risk of cardiovascular disease and cancer (Nijveldt et al., 2001; Stones et al., 2012; Wagner et al., 2004). The high phenolic content including tocopherol levels in chia oil contribute to the health value of chia.

The total dietary fiber from ground chia seeds is approximately 40%, and of that,

17% is reported to be soluble fiber while the remaining 83% is insoluble fiber based on certain measures (Reyes-Caudillo et al., 2008). The most abundant component of

insoluble fiber is “Klason lignin” at 40% which is suggested to protect the unsaturated

fatty acids from oxidation, and may contribute to a strong, impervious structure. The

lignin may be a major contributor to the hypocholesterolemic effect of chia dietary fiber

due to its ability to absorb bile acids (Reyes-Caudillo et al., 2008). Defatted chia flour has

~ 56.5 g/100 g total dietary fiber content composed of ~ 3% SDF and 53.5% IDF

(Vázquez-Ovando et al., 2009). The fiber-rich fraction water-holding capacity is 15.4 g/g.

Copious production of mucilage is one of the properties that contribute to the

value of the seeds. Chia mucilage can be considered to consist of non-cellulose slime and

cellulose mucilage. When seeds are moistened, water is rapidly absorbed, causing the

cuticle to rupture and the exocarp cell contents to be expelled as mucilage in twisted or

coiled helical threads, often, if not universally, in tubes or sheaths. It is thought that the

mucilage together with its very high water content provides an oxygen barrier, preventing

germination and causes the seeds to adhere to the soil substrate tightly.

The seed coat is high in a fiber which becomes mucilaginous and expands

considerably (> 10-fold) when soaked in water. Mucilage yield is on the magnitude of 38

± 1.0 g/kg of seed. Composition analysis of freeze-dried chia mucilage reveals 11.5%

moisture, 11.3% protein, 3.1% fat, 8.4% ash, 13.5% crude fiber and 63.7% carbohydrates

(Capitani et al., 2013). The carbohydrates have been extensively analyzed, and found to

be a polysaccharide mainly composed of xylose, glucose and methyl-glucuronic acid

residues. The molecular weight of the main polysaccharide ranges from 1 – 2 x 106

daltons from linear repeats of a tetrasaccharide with 4-O-methyl-α-D-glucoronopyranosyl residues occurring as branches at O-2 of some β-D-xylopyranosyl residues in the main chain consisting of (1-->4)-β-D-xylopyranosyl-(1-->4)-α-D-glucopyranosyl(1-->4)-β-D- xylopyranosyl units (Lin et al., 1994).

Clinical trials, epidemiological investigations and experimental studies suggest consumption of α-linolenic acid (ALA - 18:3n3, ω-3) has a positive impact on cardiovascular disease (CVD) and may also have a positive impact on diabetes. Flaxseed with its high ALA content has been found, in a limited number of studies, to decrease the risk of CVD and diabetes (Prasad, 2009; Taylor et al., 2010). Most S. hispanica

genotypes have a greater ALA content than flaxseed (~10-15% more ALA). One line of

S. hispanica, Salba®, has shown some risk reduction benefits regarding cardiovascular health and blood glucose regulation (Vuksan et al., 2009; Vuksan et al., 2007) which may be due to its higher ALA, antioxidant and/or fiber content or composition. Another study demonstrates that use of chia oil compared to maize oil in rats which were fed a high sucrose diet and developed insulin resistance, is able to reduce the activity of liver enzymes and molecules responsible for adipose tissue accumulation, including liver triglycerides (TAG), fatty acid synthase (FAS), Acetyl CoA carboxylase (ACC) and glucose 6- phosphate dehydrogenase (G-6-PDH) activities, and increased fatty acid oxidation and carnitine palmitoyl transferase (CPT-1) activities, which oxidize existing fats (Rossi et al., 2013). A follow-up study found that the replacement of corn oil by chia seed in a sucrose rich diet reduced adipocyte hypertrophy, cell volume and size distribution, improved lipogenic enzyme activities, lipolysis and the anti-lipolytic action of insulin. In the skeletal muscle lipid storage, glucose phosphorylation and oxidation were normalized. Chia seed reversed the impaired insulin stimulated glycogen synthase activity, glycogen, glucose-6-phosphate and glucose transporter type 4 (GLUT-4) protein levels as well as insulin resistance and dyslipidemia (Oliva et al., 2013).

Chia diets dramatically decreased TAG levels and increased high density lipoprotein cholesterol and ω-3 FA content in rat serum (Ayerza and Coates, 2005;

Ayerza and Coates, 2007). Chia seed decreased TAG and cholesterol more than chia oil but both were significantly superior to the control diets. Dietary chia seed also improves adiposity and insulin resistance in dyslipidemic (abnormal amounts of lipids in blood) rats (Chicco et al., 2009). Diets supplemented with chia have been found to decrease risks

from some types of cardiovascular diseases, cancers and diabetes. It has been reported

that chia diet decreased the tumor weight and metastasis number and also inhibited

growth and metastasis in a murine mammary gland adenocarcinoma (Espada et al.,

2007). Long-term supplementation with chia attenuated a major cardiovascular risk factor

and emerging factors safely beyond conventional therapy, while maintaining good

glycemic and lipid control in people with well-controlled type 2 diabetes (Vuksan et al.,

2009; Vuksan et al., 2007). A report in Inpharma Weekly indicates benefits from S.

hispanica seed in treating type 2 diabetes mellitus (Anon, 2008). However, this study

conflicts with another study that reported no changes in a number of health parameters

including C-reactive protein, interleukin 6, monocyte chemoattractant protein 1, tumor

necrosis factor alpha, oxidative stress markers and blood pressure with consumption of

50 g/day of chia seed compared to 50g/day of soy for 12 weeks by overweight middle-

aged men and women (Nieman et al., 2009). Plasma ALA increased from 2.8% to 24.4% but EPA and DHA levels did not show a significant change. This could be due to differences in regular chia vs. Salba ®, the fact that the whole seeds that were consumed in a drink without mastication were poorly digested; a masking effect to the health benefits of soy itself in the placebo or that chia has few health benefits compared to more regular American diets. Chia is reported to exhibit some reduction on CVD risk factors

(Ulbricht et al., 2009 ).

Chia may also be useful for treating other ailments. Chia seed oil was found to be beneficial as an adjuvant moisturizing agent for pruritic skin, including patients with end- stage renal disease (Jeong et al., 2010). Omega-3 FA are reported to have benefit in

psychiatric disorders, specifically that ω-3 FAs have significant benefit in prevention

and/or treatment of unipolar and bipolar depression (Freeman et al., 2006).

Though many studies find chia seeds to ease symptoms of diseases including diabetes, CVD, and cancer, evidence for in vivo safety and efficacy is limited (Ali et al.,

2012). A study on the effect of chia seeds on disease risk factors in 62 overweight post-

menopausal females (Nieman et al., 2012), found no significant difference in risk factors like body composition, inflammation, blood pressure between the chia groups and placebo (poppy seeds) groups. However, the group that received milled chia showed a

58% increase in ALA and 39% increase in EPA in plasma than the group that received whole chia seed supplementation. Studies (Nieman et al., 2009) found no benefit of chia

on body composition in overweight individuals when supplemented with 25g chia seed in

250ml of water twice a day. A randomized control trial using a beverage containing soy

protein, nopal, chia seed and oat was also carried out (Guevara-Cruz et al., 2012). In

comparison to a control diet group, the beverage group experienced body weight loss and

reduction of triglyceride and blood glucose levels. In another study, (Vuksan et al., 2010)

where chia was an ingredient in white bread, participants showed reduced postprandial

glycemia at 120 minutes after consumption. In another report (da Silva Marineli et al.,

2015), chia seed and extracted chia oil can reduce oxidative stress in vivo, by improved

antioxidant status and reduced lipid peroxidation in diet-induced obese rats.

Higher in ALA, fiber and minerals than flaxseeds, chia is considered a

“superfood”. Recently recognized as a novel food by the European Union in 2009, chia is

being explored and used in food products, thereby adding nutritional value to bread

products, cakes, smoothies, yogurt, etc. Chia gel has been found to be able to replace one

fourth of the oil content in cake with no significant difference in color, taste or texture

(Borneo et al., 2010). In addition to improving health and reducing diseases, chia is being

studied for applications in sports nutrition. Though it still remains unclear if chia can

provide a performance enhancing food source, a crossover study (Illian et al., 2011)

found there was no significant difference in level of performance in endurance events

lasting >90 min in individuals who were subjected to CHO carbohydrate loading

compared to participants who were loaded with a chia drink (50% chia seeds and 50%

Gatorade). Thus, chia drinks can be a good substitute for high sugar drinks, thereby

reducing dietary intake of sugar and providing athletes with ω-3 FA which also have

beneficial effects on psychiatric health. A feedback questionnaire submitted by a

participant indicated that they performed their activity at a reduced heart rate with the

chia drink compared to only Gatorade. Another reported that it was easier to consume the

chia drink than straight Gatorade especially when a usual meal plan is followed. A study

to determine the effect of chia seed oil supplementation on prolonged intense running

(Nieman et al., 2015) concluded similar run and exhaustion times in both chia seed oil

supplement group and flavored water group.

In addition to direct human consumption, chia has utility as a feed ingredient.

Feeding chia to egg-laying hens results in a significant accumulation of ω-3 FA in the eggs, but without the fishy or off taste associated with other supplements (Ayerza and

Coates, 1999). Chia also has the potential to be a much less expensive source of ω-3 FAs and is much more environmentally sound than harvesting krill for oil, one of the common sources of ω-3 FA feed today. Omega-3 FA-enriched eggs are now widely available in supermarkets and have been marketed successfully, and many farmers are switching to

chia as a source of ω-3 FA omega 3s as a dietary supplement in egg-laying hens

(Simopoulos AP, 2009). Another study finds that feeding chia seed to chickens and quail to supplement their diet increases levels of ω-3 FAs omega 3 fatty acids in breast meat,

thigh meat and eggs (Komprda et al., 2013). Fish and krill do not biosynthesize ω-3 FAs but only accumulate them from dietary sources, making chia a potentially useful ω3 source in fish diets (Silva et al., 2014).

Since its introduction to American culture as the Chia Pet® over 30 years ago,

Salvia hispanica has captured our culture’s imagination. However, people are starting to take a more serious look at this small seed which is starting to have a big impact on global and American agriculture. A member of the mint family, its production of large quantities of ω-3 FAs is causing agronomists and farmers to start taking it seriously as a viable crop. It is among the most sustainable source of ω-3 FAs available on the market.

Combined with high fiber levels and its promise for improving cardiovascular and diabetic health, chia is sure to be a superfood people will talk about and consume for years to come.

Chapter 2 VgDGAT1A INCREASES RENEWABLE OIL CONTENT AND OLEIC

ACID CONTENT IN SOYBEANS (Glycine max)

Introduction

Various renewable oil sources have been used historically with plant oils representing one of the world’s most important renewable resources. In addition to extensive food uses, plant oils also have value as renewable chemicals and fuels, such as biodiesel. World-wide production of plant oil is dominated by palm, soybeans, rapeseed

(including canola) and sunflowers (Wilson and Hildebrand, 2010). Oil palm and soybean production has increased rapidly and rapeseed has also shown steady increases and this trend is expected to continue with total global plant oil production of ~ 500 million MT in

2012 (Inform Sept. 2013).

Genetic improvement of oilseed yield per unit land area has shown fairly steady progress with continuous soybean yield increases being a good example (Egli, 2008a;

Egli, 2008b). This has occurred with little or no increased inputs, making renewable oil production from plants less expensive over time and progressively more competitive with petroleum as an industrial chemical feedstock. Furthermore, it is becoming increasingly possible to alter hydrocarbon flux into oil in various organisms including oilseeds, increasing oil yield per unit land area independent of higher yield. This can be particularly true for low oil oilseeds such as soybeans. A number of studies report higher oil content increases with enhanced expression of TAG biosynthetic genes (Andrianov et al., 2010; Lardizabal et al., 2008b; Rao and Hildebrand, 2009; Taylor et al., 2009).

Elevated expression of regulatory genes that up-regulate multiple enzymes for fatty acid biosynthesis also can result in higher oil levels (Andrianov et al., 2010). Co-expression of

the transcription factor WRI1 with DGAT1 is shown to have a synergistic effect on TAG

biosynthesis in plants (van Erp et al., 2014; Vanhercke et al., 2014; Vanhercke et al.,

2013).

Seed oil is biosynthesized during the second main stage of seed maturation

(Goldberg et al., 1989; Harwood and Page, 1994; Le et al., 2007a), concurrent with

elevated expression of associated biosynthetic enzymes. In studies of expression profiles of TAG biosynthetic enzymes and oil accumulation in developing soybeans, DGAT1 show an expression profile suggesting a dominant role in soybean oil biosynthesis, but

DGAT2 and PDAT do not (unpublished data; Li et al., 2013). Two full-length DGAT1s

have been cloned from developing soybean cDNA designated GmDGAT1A (GenBank #

AB257589) and GmDGAT1B (GenBank # AB257590). Both soybean DGAT1s are highly

expressed at developing seed stages of maximal TAG accumulation with DGAT1B

showing highest expression at somewhat later stages than DGAT1A (Li et al., 2013).

Vernonia galamensis, Euphorbia lagascae and Stokesia laevis accumulate 60%

or more epoxy fatty acids in their seed oil. The DGAT1 genes of V. galamensis and E.

lagascae were cloned (Hatanaka et al., 2003) and it was found that two DGAT1s from V.

galamensis (VgDGAT1A and VgDGAT1B) are expressed in all plant tissues with highest

expression in developing seeds (Yu et al., 2008). When an epoxygenase gene from S.

laevis (SlEPX) was expressed in soybean seeds, the transgenic seeds exhibited some

undesired phenotypic alterations, but these side effects are overcome in the VgDGAT1

co-expressing soybeans (Li et al., 2012).

Expression of a Tropaeolum majus DGAT1, TmDGAT1, was reported to

increase oil content, seed size and yield in Arabidopsis (Xu et al., 2008). Similar

increases in oil content and seed yield were reported with higher expression of a Brassica

napus (canola) DGAT1, BnDGAT1, back in B. napus (Weselake et al., 2008). They also reported enhanced drought tolerance of these transgenic canola with increased DGAT expression. Lardizabal et al.(2008b) reported a 1.5% increase in oil levels of transgenic soybeans expressing a fungal DGAT2. Unlike traditional breeding for high oil seed protein levels were not reduced. A more recent study, which used a Corylus americana

DGAT1 selected from a screening library, resulted in a 3% increase in total oil content and an edited GmDGAT1B caused a similar increase in oil (Roesler et al., 2016).

Evidence from greenhouse experiments and field trails shows VgDGAT1A can increase oil content at least this much.

We find much greater TAG biosynthesis and accumulation in yeast and insect cells expressing VgDGAT1A than several other DGAT1s and report generation of > 20 independent VgDGAT1A transgenic lines which show robustness in the field. Many of these lines show a significant increase in oil content with little or no decrease in protein content compared to the parental line.

Materials and Methods cDNA cloning

We have previously cloned full length DGAT1s from Vernonia galamensis

(VgDGAT1A, VgDGAT1B), Euphorbia lagascae (ElDGAT1A) and soybean cultivar

“Jack” (GmDGAT1A, GmDGAT1B) (Hatanaka et al., 2003; Li et al., 2013; Yu et al.,

2008). The procedures were based on designing degenerate primers and RT-PCR for V. galamensis and E. lagascae, or EST searching for soybean, and a RACE strategy to sequence the full length of cDNAs.

Expression in insect cells

The cDNAs were cloned into the pFastBac1 vector (Invitrogen, CA) and the

expression in Sf9 cells was performed with the Bac-to-Bac expression system (Gibco

BRL), and the recombinant baculovirus was prepared following their instruction manual.

Sf9 cells were then infected by a baculovirus possessing Vernonia or Euphorbia DGAT1

(VgDGAT1A, ElDGAT1A) and cultured for 4 days and the cells were collected. Another

set of cultured cells was infected by a baculovirus without cloned genes as a control. The

cultured cells were freeze-dried and their lipids were extracted with chloroform:methanol

(2:1). The TAG fractions were separated with thin layer chromatography (TLC) in

hexane: ethyl ether: acetic acid (90:10:1, v/v/v) solvent system and analyzed with gas

chromatography (GC) as described below.

Yeast microsome assays

Arabidopsis, soybean and Vernonia DGAT1s (AtDGAT1, GmDGAT1A,

VgDGAT1A and VgDGAT1B) were cloned into the yeast vector pYES2 (Invitrogen, CA).

The constructs along with the void vector as a control were used to transform yeasts

(Saccharomyces cerevisiae) strain INVSc1 (Invitrogen, CA). Transformed yeast cells were cultured and the microsome fractions were prepared as in Dahlqvist et al. (2000).

The reaction mixture (100 µL) contained 20 mM radiolabeled linoleic acid CoA,

300 mM dioleyl diacylglycerol, 0.02% Tween 20, 100 mM Tris-HCl (pH 7.1), 1 mM

MgCl2, 0.5 mM CoASH, 0.5 mM ATP and microsomes (corresponding to 50 µg protein).

The suspension was incubated at 30 °C with shaking (100 rpm) for 1 hour. The reaction

was stopped by first placing the test tubes with the reaction mixture on ice and followed

by adding 100 µg of soybean triacylglycerol as carrier. Lipids were extracted with

23 chloroform: methanol (2:1, v/v). Samples were loaded on TLC plates and the radioactive bands were detected by phosphorimaging and scintillated. For identification of radioactive products, methylated fractions were analyzed by TLC with a hexane: MTBE

(Methyl tert-butyl ether): acetic acid (85:15:1, v/v/v) solvent system.

Lipid analysis

Samples prepared as described above were frozen in liquid N2, stored at – 80˚C and then lyophilized. Samples were transferred to glass test tubes and tri-heptadecanoin

(tri-17:0) was added at 10 µg/mg tissue as a standard. The samples were finely ground and 1-2 mL of chloroform and methanol (2:1) containing 0.001% butylated hydroxytoluene (BHT) was added and samples were ground further. After brief centrifugation, the chloroform phase was transferred into a new glass tube. Samples were divided into two aliquots, where one was used for TLC and the other directly for GC analysis.

For separation of individual lipid classes by TLC, the chloroform extracts were concentrated to ~ 50 – 100 µL. 10 µL of the sample was loaded in a narrow band in lanes of Whatman LK6D Silica gel 60A TLC plates 1 cm from the bottom of the plates. The plates were put in a chamber with chloroform: methanol: water (65:25:4, v/v/v) containing 0.0001% BHT for running until the first solvent reached ½ up the plate (~ 10 cm). Then the plate was moved into the second solvent, hexane: diethyl ether: acetic acid

(100:100:2, v/v/v) containing 0.0001% BHT and developed until the solvent was ~ 1 cm from the top. After development, the plate was dried, and subsequently sprayed with

0.005% primulin in 80% acetone, followed by visualizing under UV light. The bands of interest were scraped and transferred to a Pasteur pipette with a glass wool plug washed

with chloroform: methanol. The lipid samples were eluted with 0.5 mL of chloroform:

methanol containing 0.0001% BHT twice. Finally, eluted lipid samples were analyzed by

GC.

For GC analysis, samples were dried with N2, and 0.5 mL of 0.5 M sodium

methoxide (NaOCH3) in methanol was added and incubated for at least 15 minutes with

shaking at 22 ˚C. 0.5 mL of isooctane containing 0.001% BHT was added to each tube

and mixed well. Phase separation was obtained with centrifugation and adding aqueous

0.9% potassium chloride if needed. The isooctane layer was extracted and transferred into

GC auto-sampler vials. The fatty acid methyl esters (FAMEs) were analyzed with gas

chromatography on a Varian CP-3800 GC with a 24 m x 0.25 mm ID CP-Select CB for

FAME (Agilent Technologies) fused silica column with a 0.25 µm film thickness. The

temperature program was 90˚C for 1 min., then to 155˚C at 20˚C/min. with no hold, then

to 175˚C at 3.6˚C /min. with no hold and finally to 250˚C at 12˚C/min. holding for one

min.

Analysis of protein and oil content was also performed by non-destructive means

on a single seed basis using nuclear magnetic resonance (NMR) and near-infrared

spectroscopy (NIR) (Agelet et al., 2012; Armstrong, 2006). Briefly, the instrument

collects spectra on a seed as it falls though a short length of illuminated glass tubing.

Optic fibers are aligned with the tube axis and connected to a spectrometer for collection

of seed spectra. The instrument is designed for large sample sorting and processes seeds

at about 3 per second although seeds were hand feed for this study to preserve identity.

NMR single-seed oil measurement was accomplished using a Minispec mq 20, (Bruker

BioSpin, The Woodlands, TX) using the manufacturers’ protocol for oil measurement.

Seed and oil mass were measured and converted to a dry basis oil percentage using a

moisture content estimated from room equilibrium moisture conditions.

Seed-specific expression vector construction and soybean somatic embryo transformation

An expression vector for soybean transformation was constructed using pCAMBIA1301 containing the hygromycin resistance gene and the GUS reporter gene

(Cambia, ACT, Australia; GenBank # AF234297). The coding sequence of VgDGAT1A was amplified by a high fidelity polymerase (Invitrogen) using end-specific primers containing restriction sites. The amplification product was then subcloned into the respective sites of pPHI4752 vector containing a phaseolin promoter, which confers strong seed-specific expression of transgenes (Slightom et al., 1983). The phaseolin promoter cassette containing the coding region of each target gene was transferred into the corresponding sites of the binary pCAMBIA1301, T-DNA vector. The recombinant expression vector was subsequently introduced into somatic embryos of soybean (cv.

`Jack’) using the particle bombardment method of transformation.

Soybean somatic embryo induction and culture was carried out using a protocol modified from prior procedures (Collins et al., 1991; Finer and Nagasawa, 1988;

Samoylov et al., 1998; Trick et al., 1997). Immature soybean seeds at 3-5 mm length were dissected, and cotyledons were placed on D40 (40 mg/L 2,4-D in MS media) solid medium for a one-month somatic embryo induction. The induced embryos were transferred to D20 plates for proliferation. The globular embryogenic cultures from D20

(20 mg/L 2,4-D in MS media) plates were then moved into FN (Finer and Nagasawa,

1988) liquid medium for one-month suspension culture. Small embryo clumps were selected for particle bombardment gene delivery.

Plasmid DNA/gold preparation for the particle bombardment was conducted according to standard protocols (Trick et al., 1997). A DuPont Biolistic PDS1000/HE instrument (helium retrofit) was used for all transformations. After bombardment, the embryo clumps were transferred into FN liquid medium containing 30 mg/L hygromycin for selective culture for four to five weeks. The positive transformed embryos obtained by hygromycin selection were then moved into fresh FN liquid medium for culture and simultaneously for GUS test and identification of the transgene presence by PCR. The

PCR-positive transgenic embryo lines were transferred into maturation medium (SHaM)

(Schmidt et al., 2005b) for three to five weeks. For the transgenic lines, one set of matured somatic embryos were sampled for lipid extraction and subsequent GC analysis.

The rest of the matured somatic embryos were desiccated for 4-7 days, and then were placed on half strength MS solid medium for germination. Germinated plantlets were transferred to closed sterile soil cups for growth in a culture room under 23:1 (light: dark) photoperiod cycle and 25˚C. Once seedlings reached 13 cm, they were transferred to a greenhouse for flowering and seed set under a 16:8 (light:dark) cycle, 25/21˚C. Mature seeds were harvested from each regenerated soybean plant separately. Seeds were chipped for genotyping by PCR and fatty acid analysis by GC.

Upon successful transformation of at least one copy of the transgene into background Jack, the seedlings were developed in the greenhouse and subjected to seed increase the next year. In 2011, plants were grown out at Spindletop Farm in Lexington,

KY in 20’ headrows and hand harvested. Seeds were sent to Manhattan, KS for oil and protein evaluation via non-destructive single seed NIR and NMR described above.

Segregating Line Analysis

Nine independent transgenic lines which contained at least one copy of the

VgDGAT1A trait as determined by PCR analysis as described above were grown on at

Spindletop Research Farm in Lexington, KY in the 2012 growing season. Thirty seeds

were grown each year from these lines in 2013 and 2014, selected using a single seed

descent method. In 2014, leaf DNA was randomly collected from 63 of 270 planted

soybeans using phenol-chloroform extraction and analyzed for the presence of the

VgDGAT1A trait by PCR as described above. Seeds from these plants were hand

harvested during October 2014 and analyzed for oil and protein content via NIRS as

described above. The genotype data was connected to the phenotype data and the mean

oil and protein content of each genotype was compared.

Results

In vitro assays of VgDGAT1A

We have found that Glycine max, E. lagascae and V. galamensis have at least two

DGAT1s and have made full-length DGAT1 cDNAs from all three species. We isolated five cDNA clones: Soybean DGAT1s (GmDGAT1A and GmDGAT1B), Euphorbia

DGAT1 (ElDGAT1A) and Vernonia DGAT1s (VgDGAT1A and VgDGAT1B). These cDNA clones are mainly expressed at the stage of seed development which involves high oil accumulation.

VgDGAT1A resulted in high oil accumulation in Sf9 insect cells. In experiments expressing VgDGAT1A and ElDGAT1A in Sf9 insect cells, we found much higher accumulation of TAG in the VgDGAT1A expressing cells than the ElDGAT1A expressing and vector control cells (Hatanaka et al., 2003). One way ANOVA analysis confirms that

VgDGAT1A incorporation results in a statistically significant higher level of oil

(p=0.002), but E1DGAT1A did not (p=0.072). In fact, VgDGAT1A demonstrated 20-fold

increase in total lipids compared to the control, consistent with the results with yeast

microsome assay below.

An improved system for microsomal analysis was used which utilizes 300-fold

lower concentrations of microsomes compared to most studies, resulting in lower

background and more accurate activity determinations than traditional methods (Yu et al.,

2006). In our yeast system, cells expressing GmDGAT1A, VgDGAT1A and VgDGAT1B

were analyzed for TAG biosynthetic activity along with the Arabidopsis thaliana DGAT1

(AtDGAT1) and the vector control. Of these, three showed significant activity in the yeast system: GmDGAT1A, VgDGAT1A and VgDGAT1B (Figure 2.1). GmDGAT1A exhibited twice the activity of controls, while VgDGAT1A and 1B showed five to six fold activity over GmDGAT1A.

Oil and protein content of transgenic soybeans expressing VgDGAT1A

VgDGAT1A was also expressed in soybean somatic embryos. Transformation resulted in about 40 viable lines that stably integrated at least one copy of VgDGAT1A

(unpublished results). Soybean somatic embryos expressing VgDGAT1A were

regenerated, grown out in a greenhouse and mature T2 seeds were collected. Seeds of

these higher oil soybean lines were grown out in both greenhouse and field environments

and progeny again analyzed for protein and oil contents in the next year, and subsequent

years until lines were homozygous for the transgene, at T6. As it is seen in Table 2.1, the

protein and oil content of mature field-grown seeds were measured and many of the

VgDGAT1A transformed soybean seeds showed 3 to 4 percent increases in total oil

content per seed dry weight, most without any significant (p<0.05) decrease in protein

levels compared to the wild type, Jack.

The estimated higher meal protein levels of such lines are also likely to make

defatted meal more valuable for animal feed, food and many industrial applications. A

summary of protein and oil data can be found in Figure 2.2 of the transgenic lines

including 12-B, 10-B and 7-B and the wild type. These independent transgenic lines, 12-

B, 10-B and 7-B had significantly higher oil content than the wild type (WT) (p<0.0001)

but no significant difference in total protein content. Additionally, 10 other transgenic

lines showed no significant change (α=0.05) in protein content compared to the wild type,

but had significantly higher oil content (See Table 2.1). Protein and oil contents have

been calculated on a 0% moisture basis as described in Table 2.1.

Segregating Line Analysis

Of the 63 randomly selected plants, 22 tested positive for the VgDGAT1A trait

via PCR analysis and 41 tested negative for the trait. The average oil and protein content

by NIRS for VgDGAT1A positive soybeans was 23.4% and 45.9% respectively while

VgDGAT1A negative soybeans was 22.8% and 45.4%. Some VgDGAT1A positive

soybean lines have much more oil than others.

Fatty acid composition in high oil lines

All three high oil lines, 7-B, 10-B and 12-B, analyzed had significantly higher levels of 18:0 and 18:1∆9 while 18:1∆11, 18:2, and 18:3 fatty acids were significantly

lower than the wild type, Jack (Table 2.2). Of the increased oil (difference transgenic

line-wild type), 63% to 79% was due to higher 18:1∆9, oleic acid, level.

Discussion

Vernonia DGAT1A and 1B exhibited uniquely high activity compared with the other DGATs (Figure 2.1). VgDGAT1A and 1B therefore could be useful for increasing renewable oil production. This is consistent with the unique phylogenic grouping of

VgDGAT1A and VgDGAT1B peptide sequences compared to other DGATs (Yu et al.,

2008), and recently reported sunflower DGAT1s (HaDGAT1A and HaDGAT1B) are relatively close to Vernonia DGATs as in the same Asteraceae family. These HaDGAT1s

may contribute to the high oil content in sunflower seeds. The microsomal assays

performed in the yeast system imply that VgDGAT1 genes result in up to a tenfold

increase in DGAT activity. This increase in activity may also exhibit similar results in

commercial oilseeds such as soybeans, and we would anticipate higher triacylglycerol

levels as one of the effects of this increased activity.

We have previously reported on the activity of VgDGATs in the yeast system

described above (Yu et al., 2008). Combined with the results above, we had compelling evidence that VgDGAT1A exhibits uniquely high activity and would do so upon stable transformation into and oilseed, plausibly increasing overall oil content. Of the five genes initially examined, VgDGAT1A results in the highest activity and oil accumulation in the yeast and Sf9 insect cell system. With this data, VgDGAT1A, showing the highest activity in both assays, was transferred into a pCAMBIA1301 vector containing a phaseolin promoter, which confers high expression at seed filling stages. Several transgenic lines showed a 3 to 4 percent increase in total oil content without a significant decrease in protein content. Using gene editing technology to modify specific amino acids, a modified GmDGAT1B increased oil content 3 to 4% compared to null segregants

(Roesler et al., 2016). VgDGAT1A results in a similar or greater increase in oil without

amino acid editing or further modification. Among 14 substitutions in GmDGAT1B-

MOD (Figure 2.3), “Y231F” corresponds to F250 in VgDGAT1A and F245 in

VgDGAT1B. Four substitutions in GmDGAT1B-MOD, S58N, S264T, R467K and

L483V, have corresponding differences in VgDGAT1A at D75, N283, Q486 and F502.

The remaining 9 amino acids substitutions in GmDGAT1B-MOD are not present inVgDGAT1A; these amino acids in VgDGAT1A are the same as GmDGAT1B. These 9 amino acids are unique in GmDGAT1B-MOD. There are no reports of any soybean lines from traditional breeding that have shown up to a 4% increase in oil content without exhibiting a significant loss of protein.

Preliminary small scale yield trials indicate no yield penalty for the VgDGAT1A high oil trait as was the case of soybeans with increased oil expressing a DGAT2

(Lardizabal et al., 2008c) but this will need to be verified in multi-year/location trials.

Incorporation of this trait into 50% of soybeans worldwide would result in an increase of

850 million kg oil/year without new land use or inputs and be worth ≥ $800 million/year at 2015 production and market prices assuming the same yields are confirmed. With higher protein levels this also increases the nutritional value of the meal. Historically, plant oils were ~ 4 times the cost of petroleum but in recent years major plant oils such as palm and soybean oil have been much closer to the price of petroleum (Figure 2.4). As supplies are exhausted, the price of petroleum will gradually rise while advances in the technology of producing renewable oils such as reported herein the cost of producing renewable oils including plant oils may rise less or even fall some. Such higher oil producing enzymes can be used in many renewable oil production systems including

algae. DGAT is known to limit TAG accumulation in many organisms and oil can be

further increased by combining DGAT with other TAG accumulation enzymes (Jin and

Jiang, 2015).

The altered fatty acid composition of the VgDGAT1A high oil soybeans can

affect uses of soybean oil although the changes are not large enough to affect this to a

great extent. This alteration might be due to several possibilities. In a yeast expression

system, it was found that VgDGAT1A has higher activity with oleoyl-CoA and oleoyl-

DAG than vernoloyl- CoA and vernoloyl-DAG (Yu et al., 2008) but it is not yet known if

VgDGAT1A is more active with oleoyl-CoA and oleoyl-DAG substrates than other common CoA and DAG substrates in cells of normal developing oilseeds including soybeans. The evidence accumulated to date indicates that GmDGAT1A and

GmDGAT1B appear to be largely responsible for soybean oil biosynthesis in normal soybeans (Li et al., 2013; Li et al., 2010). The main product of plastid fatty acid biosynthesis is oleoyl-ACP which is exported to the cytosol and extra-plastidial membranes as oleoyl-CoA. It is quite possible that higher DGAT activity in VgDGAT1A transgenic soybean intercepts oleoyl-CoA and oleoyl-DAG substrates from CoA and

DAG pools available for TAG biosynthesis before oleoyl substrates are further metabolized such as by membrane desaturases, FAD2 (Oakes et al., 2011). Increased expression of foreign and native DGATs have been found to increase TAG oleoyl (18:1) levels in species as diverse as maize, olive (Banilas et al., 2011) and Arabidopsis (Jaco et al., 2001). As an example of a commercial oilseed under field conditions, it has been reported (Lardizabal et al., 2008b) that increased oil content in soybean seeds by using a fungal DGAT2A in field conditions and Taylor et al. (Taylor et al., 2009) also reported

increased oil content in canola overexpressing Arabidopsis DGAT1 but neither of these studies mention data on fatty acid composition. Wang et al. (2014) reported that

overexpression of a sesame DGAT1 had no significant effect on fatty acid composition in

soybean seeds grown under controlled conditions. The use of an edited CaDGAT1 and

modified GmDGAT1B also showed a significant increase in 18:1 oil content and

observed a correlation between 18:1 affinity and oil content (Roesler et al., 2016).

Lower polyunsaturated fatty acid levels are favorable for most food and fuel uses

of soybean oil due to the higher oxidative stability. Whether the fatty acid composition of

the higher oil soybeans impacts other food and renewable uses depends on the desired

final product properties. There is a movement toward very high oleate soybeans and other

oilseeds for most food uses and it would be interesting to determine how the VgDGAT1

high oil trait affects fatty acid composition in very high oleate oilseeds.

Table 2.1. Summary Statistics of Oil and Protein. Oil values were collected by utilizing non-destructive single seed NMR for 48 randomly selected clean seeds. Protein data was collected by non-destructive NIR for the same 48 seeds, and protein and oil values were calculated by adding NMR oil and NIR protein. The oil values were verified by Soxhlet and protein by Kjeldahl analyses. Line numbers of 1-B to 24-B are independent transgenic lines. P values were calculated by Dunnett’s test against the wild type group, Jack (WT), where α=0.05.

NIR NMR Oil Protein Protein P + O Lines Protein Oil (%) p-Value p-Value + Oil (%) p-Value (%) 3-B 22.9 <.0001 34.1 0.0002 57.0 0.0009 24-B 22.7 <.0001 33.2 <.0001 55.9 0.4877 14-B 22.6 <.0001 33.4 <.0001 56.0 0.3352 12-B 22.2 <.0001 36.4 0.998 58.6 <.0001 10-B 21.8 <.0001 36.8 1 58.6 <.0001 7-B 21.8 <.0001 36.0 0.816 57.9 <.0001 6-B 21.8 <.0001 36.6 1 58.4 <.0001 11-B 21.6 <.0001 37.9 0.7867 59.5 <.0001 8-B 21.6 <.0001 37.1 1 58.7 <.0001 4-B 21.4 <.0001 36.1 0.8857 57.5 <.0001 2-B 21.4 <.0001 34.5 0.0017 55.9 0.4819 5-B 21.4 <.0001 36.9 1 58.3 <.0001 1-B 21.2 <.0001 38.1 0.5756 59.3 <.0001 9-B 21.2 <.0001 38.6 0.1258 59.8 <.0001 20-B 20.8 <.0001 38.5 0.1965 59.3 <.0001 15-B 20.7 <.0001 37.6 0.9963 58.3 <.0001 23-B 18.9 0.0782 36.8 1 55.7 0.7328 18-B 18.7 0.3846 40.0 <.0001 58.7 <.0001 19-B 18.1 1 38.7 0.0795 56.8 0.0039 WT 18.0 1 37.0 1 54.9 1 13-B 16.6 0.0005 40.6 <.0001 57.2 0.0002 LSD 0.23 0.25 0.17

Table 2.2.Fatty Acid Composition of High Oil Lines. Fatty acids were analyzed by GC after direct methylation (see methods for details). The values are percent total of the six main fatty acids, and were analyzed by Dunnett’s test against the wild type, Jack (WT) (n=4). LSDs are for α=0.05, and values in bold with asterisks* differ significantly from the control.

Lines 16:0 18:0 18:1∆9 18:1∆11 18:2 18:3

7-B 11.4 5.3* 28.3* 1.0* 48.1* 6.0*

10-B 11.5 5.3* 29.3* 1.0* 47.6* 5.5*

12-B 11.2 5.6* 32.0* 1.0* 45.3* 5.0*

WT 10.8 4.0 21.4 1.1 55.2 7.4

LSD 0.72 0.39 3.08 0.05 2.57 0.30

160

140 120 100 80 Synthesis 60 /min/μgprotein 40 TAG

pmol 20 0

Figure 2.1. TAG Biosynthetic Activity of DGAT1s Expressed in Yeast. DGAT1s from Arabidopsis thaliana (AtDGAT1), Glycine max (GmDGAT1A), and Vernonia galamensis (VgDGAT1A, 1B) were cloned from cDNA libraries into the pYES2 yeast vector, which were transformed into strain INVSc1. The yeasts were cultured as previously described and microsomes prepped. VC indicates the vector control. *TAG synthesis was measured by amount total fed radiation with oleoyl-diacylglycerol (DAG) + 14C-oleoyl-CoA. Error bars represent one standard error from the mean for n ≥ 4 replicates.

24 39 Oil Protein 23 38 22

37 21 20 36 % % Oil

19 % Protein 35 18 34 17 16 33 WT 12-B 10-B 7-B

Figure 2.2. Protein and Oil Levels of High Performing VgDGAT1A Lines. Protein and oil levels of three of the highest performing VgDGAT1A expressing lines, 12-B, 10-B, 7- B and the wild type control line grown on a Farm in Lexington, KY in the 2011 growing season.

Figure 2.3. Sequence Alignment of DGATs. Peptide sequence alignment comparing VgDGAT1s, GmDGAT1s and GmDGAT1B-MOD (Roesler et al., 2016). The acyl-CoA binding site (Xu et al., 2008) is underlined and it corresponds to positions K113 to G129 in VgDGAT1A.The essential phenylalanine (Zheng et al., 2008) is indicated with red asterisks (F499 in VgDGAT1A).

Figure 2.4. Price Trends of Source Oils. Price trends of the major plant oils vs. crude oil (petroleum) in the past 15 years; US $/kg.

Chapter 3 A LARGE COPY NUMBER VARATION ON CHROMOSOME 14

APPEARS TO INCREASE OIL CONTENT IN SOYBEANS

Introduction

There is a strong genetic component that is well understood for oil content and quality, via the oil biosynthetic pathway and its regulation. Seed oil is biosynthesized during the second main stage of seed maturation (Goldberg et al., 1989; Harwood and

Page, 1994; Le et al., 2007b), at which time the relevant biosynthetic enzymes are highly expressed. In studies of expression profiles of TAG biosynthetic enzymes and oil accumulation in developing soybeans, DGAT1 shows an expression profile suggesting a dominant role in soybean oil biosynthesis, but DGAT2 and PDAT do not (Li et al.,

2013).

It is becoming increasingly possible to alter hydrocarbon flux in soybeans. A number of studies indicate oil content increases with higher expression of TAG biosynthetic genes (Andrianov et al., 2010; Lardizabal et al., 2008a; Rao and Hildebrand,

2009; Taylor et al., 2009). Increased expression of regulatory genes that up-regulate multiple enzymes for fatty acid biosynthesis also can result in higher oil levels

(Andrianov et al., 2010). Co-expression of the transcription factor WRI1 with DGAT1, a key rate limiting enzyme, is shown to have a synergistic effect on TAG biosynthesis in plants (van Erp et al., 2014; Vanhercke et al., 2014; Vanhercke et al., 2013). Overall, increasing sink strength results in increased oil and protein with a strong pronounced effect on protein and less on oils (Rotundo et al., 2011).

Cultivar effects are well understood to effect protein content and amino acids in soybeans, most likely due to heritable differences in TAG biosynthetic genes and

regulatory factors. For instance, N6202 germplasm line produced seeds with 45.7%

protein content, and a 10% reduction in grain yield compared to a control variety, NC-

Roy (Carter TE et al., 2010). Similar results were found in TN03-350 and TN04-5321, which achieved 43.1-43.9% protein content without sacrificing seed yield. In soybean genotypes of early maturity groups, average to high protein content (399-476g/ kg-1) was

found in years with high air temperature and moderate rates of rainfall during the seed-

filling period, whereas seed protein content was drastically reduced (265±347 g /kg-1) in

seasons of insufficient nitrogen fixation or higher amounts of precipitation during seed

filling (Vollmann et al., 2000).

In plant breeding random mutagenesis, such as with the use of a chemical

mutagen like ethyl methane sulfonate (EMS) or bombardment with gamma radiation, is a

common way to generate mutations and increase genetic diversity for traits with limited

natural variation. While these methods can produce useful traits, such as an early flowering mutant or increased oil content, it is possible that other important genes could be mutated, causing undesirable phenotypes such as altered seed composition. However, mutagenized populations are not brought under the same scrutiny as transgenic approaches and therefore traits induced this way are much easier to incorporate into existing germplasm. Here, we have also utilized a fast neutron (FN) population of soybeans which exhibit 3 to 4% higher oil content than the parent variety with little or no decrease in protein. A fast neutron soybean mutant population is one of the newest resources available, and has added power with the large variation in mutant sizes, from several bp deletions up to Mb sized deletion events (Bolon et al., 2011). In group 0 populations developed at the University of Minnesota, we have identified three advanced

42 generation, high oil soybean varieties and backcrossed them to the parental line to find the mutation which is the source of increased oil. We have also identified a large copy number variation event and identified the genes present in that region.

Materials and Methods

Genetic Material

Seeds from three mutant lines in a mutant population from the University of

Minnesota were requested from their Fast Neutron (FN) mutant library in April 2014 and planted in the field in Lexington, KY the summer of 2014 to verify high oil content in the environment. The oil and protein properties of these lines can be seen in table 3.1.

Backcrossing

Two lines also had high oil content in the new environment and were planted and backcrossed to the parent line, M92-220, in 2015. Successful crosses were harvested and

11 F1 crosses were assessed for oil content via single seed NMR. These were then grown in a greenhouse the winter of 2015-2016. In the spring of 2016, sixty F2 seeds from each of the F1 plants were assessed for single seed oil content and then planted at Spindletop

Farm in Lexington, KY.

Tissue Sampling, Seed Composition and DNA Extraction

Leaf tissue was collected and frozen at -80 C for comparative genomic hybridization (CGH) analysis. Seeds were harvested from each plant in the fall and each

F2 plant was analyzed in bulk with a Perten DA7200 NIRS for oil, protein and moisture content. This data was used to generate anti-correlation scatter plots of the F2 sibs. Strong anti-correlations of oil and protein content suggesting segregation of the mutant trait were found in three F2 populations.

NMR Methods

Oil content was determined by NMR using a Minspec 20 (Bruker Biospin, The

Woodlands, TX, USA). The instrument accommodates a twenty mm sampling tube

diameter. Seeds were weighed and placed into the tube and allowed to warm to 40o C

before insertion into the instrument. The standard oil seed measurement procedure

supplied with the instrument controller was used. Calibration of the NMR instrument was

done using weighed amounts of extracted soybean oil encompassing the range of oil

weights of the seed samples. Four different weights were used and were expressed onto

tissue paper at the bottom of the 20 mm sample tube.

CGH Analysis

The highest oil and lowest oil plants’ leaf DNA was extracted using QIAGEN

DNEasy Kit and CGH was performed as in Bolon et al, using a custom NimbleGen CGH

microarray with over 700K probes, approximately one probe every 1 kb (Bolon et al.,

2011). On this array, a large deletion was detected on chromosome 14 and the genes in

that region were analyzed on soybase.org and cross-referenced to homologous genes in

the TAIR database.

Results

Oil and Protein Content of Backcrosses

Three lines developed by backcrossing an M8 line to M92-220, the wild type

parent of the mutants, were found to have high oil content that were suitable for CGH

analysis, as seen in Table 3.1. Of the three lines grown in 2012 in MN, two of them also had high 2% oil content greater than the wild type in 2014 in KY, and one was <1% greater than the parent line. In 2015, 11 crosses were successful, though none of them

were successful for one line, 5R16. The oil content of these F1 seeds ranged from 17.3 to

22.5%, and was determined via single seed NIRS and can be seen in Table 3.2. F1 seeds

were grown in a greenhouse over winter and in the spring, sixty seeds from each F1

family were space planted 30 cm apart from each other, and leaf tissue was collected

during the two-leaf stage and seeds harvested at maturity. A scatter plot of the bulk-seeds harvested from each F2 plant is shown in Figure 3.1. Populations A1, A2, and A3, all crosses from IR22 and the WT, showed good anti-correlations between protein and oil and a wide range of oil and protein values, which are indicative of a heritable change segregating for a qualitative trait. The DNA from the highest oil content and lowest oil content plants can be seen in Table 3.3 was extracted and then pooled and subjected to

CGH analysis.

CGH Analysis

One large copy number variation event (CNV) was detected in lines A1 and A2, as seen in Figure 3.2, with the genes contained in that region described in Table 4. Of the

20 genes detected, none were related to the oil-biosynthesis pathway, one was related to the beta oxidation pathway of fatty acids, two had unknown function and three were characterized as transcription factors, though one of those putative factors actually encodes for a ceramide synthesis gene.

Discussion

The use of mutant populations as a forward screening tool for seed composition traits has been utilized with success in many crop species and the population of fast neutron soybean mutant families is one of the newest available, and has added power with the large variation in mutant sizes, from several bp deletions up to Mb sized deletion

events (Bolon et al., 2011). We utilized this resource and a forward screening approach to

isolate three mutant populations with high oil content and little decrease in protein to gain

greater insight into potential genes which may regulate and alter oil content with no

subsequent decrease in protein, a problem that commonly troubles conventional breeders

and selection techniques (Carrera et al., 2011). These mutants were then backcrossed to

the parent variety, which in theory should produce heterozygous offspring for the mutant

traits in the F1 population. After a generation of self-crossing, we would expect to observe a genotypic segregation of the mutant trait ratio of homozygous mutant: heterozygous:homozygous wild type 1:2:1, and for this to be reflected somewhat in the

phenotype, even though phenotypic variation is variable for qualitative traits.

In the populations A1, A2 and A3, we notice the segregation of the phenotype as

well, with oil content from a range of 19 to 24% spread over a normal distribution. If we

assume that this is a dominant or co-dominant trait, then the top 25% oil content ought to

be homozygous mutants, and the bottom 25% homozygous wild type. Therefore, CGH

ought to show the deletion event when comparing the highest and lowest oil trait, which

we observe in the A1 population. Furthermore, the size of the deletion detected is a

typical size deletion in a FN population, about 300 kb, which is in the size range we

anticipate and is frequently observed in CGH mutants (Bolon et al., 2011).

While the size of the deletion event may imply that it is the source of observed

phenotypic variation it is also important to examine the deleted genes to begin to

hypothesize about the mechanism which causes the observed variability. To date, many

efforts have been made through mutagenesis, conventional breeding, and biotechnology

to increase oil content of seeds. It has been well established that the ratio of

46 sucrose:asn+gln from the mother plant significantly alters oil and protein content, but results in an anti-correlation of these traits.

There is a genetic component to this as high oil is heritable when selected in this manner but usually results in a corresponding loss of protein. Metabolic engineering efforts have been effective at elevating oil content without the corresponding loss in protein by utilizing push, pull and protect mechanisms. First, the “push” mechanism directs hydrocarbon resources toward the oil biosynthetic pathway, creating an abundant source of metabolic precursors. The transcription factor WRI1 is known to operate in this fashion (Focks and Benning, 1998; Vanhercke et al., 2014). Next, “pull” mechanisms occur later in the pathway and use downstream metabolites at a faster rate, thereby causing upstream resources to be re-directed into the pathway (Vanhercke et al., 2013).

The enzyme diacylglycerol acyltransferase (DGAT) catalyzes the final and only dedicated step to TAG synthesis via the Kennedy pathway, by binding a diacylglycerol molecule with an acyl-CoA (Sidorov and Tsydendambaev, 2014). Biochemical studies have also determined this is a rate-limiting step in many species, so increasing the speed of TAG formation via DGAT increases the speed of the entire pathway (Weselake et al.,

2008). DGAT overexpression studies confirm this phenomenon and DGAT overexpressed plants have significantly higher oil levels with no decrease in protein

(Hatanaka et al., 2016). Lastly, “protect” mechanisms ensure that TAGs already formed do not degrade. For example, lipase knockouts (Kelly et al., 2013) exhibit increased oil content, as do oleosin overexpressors, which are proteins that stabilize storage TAGs

(Vanhercke et al., 2014). Better yet, plants which are engineered with two or more of

these steps integrated exhibit synergistic increases in oil, such as tobacco leaves with up

to 17% on a dry weight basis of storage lipids(Vanhercke et al., 2014).

Of the 20 genes detected, none were related to the oil-biosynthesis pathway, one was related to the beta oxidation pathway of fatty acids, two had unknown function and three were characterized as transcription factors. Due to their importance in regulation of the oil biosynthesis pathway it is critical to first examine the transcription factors.

One soybean gene was found be homologous to the WRKY 40 transcription factor in Arabidopsis. To date there have been 72 WRKYs discovered in Arabidopsis, which interact with DNA via the WRKY domain which forms a wedge shape that inserts itself perpendicularly in the major groove of DNA, and binding the W box element through a conserved RKYGQ motif on the beta strand (Schluttenhofer and Yuan, 2015).

These are unique domains to plant species and regulate responses to a wide array of biotic and abiotic stressors as components of signaling cascades for phytohormones including abcisic acid, auxin, brassinosteroids, cytokinin, ethylene, jasmonate, and salicylic acid (Schluttenhofer and Yuan, 2015). More recently, WRKYs have been found to play vital roles in the production of secondary metabolites from phenylpropanoids, alkaloids and terpenes. Overall, it appears that WRKYs are involved in complex defense and stress response pathways such as MAPK cascades (Schluttenhofer and Yuan, 2015).

Therefore, it is impossible to predict the function of a putative WRKY and its response to

oil and protein content in seeds, but it in general it seems these levels would not be

affected by a WRKY knockout.

WRKYs have also been attributed to tolerance of abiotic stress factors including

salt, nutrient deficiency, osmotic, cold, heat, oxidative and UV-B (Schluttenhofer et al.,

2014). A comprehensive study which examined all known Arabidopsis WRKYs shows our orthologue, AtWRKY40, positively regulates the jasmonic acid pathway and negatively regulates abscisic acid pathways. However, it appears that different WRKYs are differentially expressed when exposed to different treatments with the hormone, with

AtWRKY40 being significantly altered in all five sub treatments, and the greatest response with a nine-fold increase after one hour of treatment with methyl jasmonate

(Schluttenhofer et al., 2014). We would anticipate that this mutant may be less responsive to methyl jasmonate treatment, although great redundancy was observed in the

Arabidopsis treatment, so other WRKYs may be able to compliment the loss of function in this deletion. In a study involving 20 WRKYs in a Populus simonii x Populus nigra, it was found that these genes expressed increased expression under multiple stress conditions, and orthologs to AtWRKYs have similar changes in expression levels (Zhao et al., 2015). Therefore, it is challenging to pinpoint a WRKY ortholog to have the same function in a different organism. Nonetheless, there is little evidence to date to suggest that WRKYs alter oil and protein content during seed fill, so it is unlikely this deletion is responsible for the high oil phenotype.

Another putative transcription factor has Arabidopsis orthologs which are actually similar to the LOH family, of which there are three in Arabidopsis, ceramide synthase genes. Overexpression of LOH1 and LOH3 had enhanced biomass, attributed to an increase in very long chain fatty acid/trihydroy LCB ceramides enhances vegetative growth via cell division and elongation (Luttgeharm et al., 2015). Overexpression of

LOH2 resulted in a dwarf phenotype resulting from an increased accumulation of sphingolipids with C16 fatty acid/dihydroxy LCB ceramides, which induces programmed

cell death (Luttgeharm et al., 2015). These genes are expressed in arabidopsis rosettes and may be responsible for increases or decrease in biomass, but seem to have little effect on seed size and quality.

The third transcription factor observed, BEE1, seems to respond in the cold stress response and encodes a brassinosteroid enhanced expression transcription factor which is involved in regulation of the phenylpropanoid pathway. Under low temperature growing conditions, knockout mutants had reduced levels of quercetins and scopolin, with an increase in anthocyanin production by inhibiting anthycyanin synthesis genes by suppressing the TFs TT8 and GL3 genes (Luttgeharm et al., 2015). While it is well established that cold stress can affect the fatty acid composition of plant oils, the effects of BEE1 on this phenomenon have yet to be established. BEE1 can also interact with other brassinosteroid regulators such as CESTA, another example of the complex regulatory networks that exist in higher plants (Poppenberger et al., 2011). Keeping this in mind, it is possible that any transcription factor could result in downstream effects that may ultimately affect oil and protein content but these would be difficult to observe and pinpoint.

These three transcription factors do not directly modulate TAG biosynthesis, but transcription factors which are master regulators and control other transcription factors and metabolic pathways can have unanticipated results when they are knocked out. There is one enhancer present which is also able to increase transcriptional activity thought it is also unknown if this one affects TAG biosynthesis. There are also two unknown proteins that have yet to be characterized so it is difficult to speculate about their importance in the TAG biosynthetic pathway. There are several other enzymes which appear to be

uninvolved with TAG in any foreseeable way. There seem to be many genes in this

deletion which have been knocked out but fortunately do not cause any other obvious

phenotypic changes, and if the gene responsible for increased oil accumulation is

contained here, it would be ideal as it is a deletion that comes with no other crucial

effects.

In this deletion there is one gene that may be related to TAG accumulation/

catalysis in seeds. A putative enoyl-CoA hydratase, an enzyme which is directly involved

in the beta oxidation pathway which digests fatty acids may be responsible for the high

oil phenotype. Specifically, this enzyme catalyzes the syn-addition of a water molecule across the double bond of a trans-2-enoyl-CoA thioester, resulting in the formation of a

β-hydroxyacyl-CoA thioester (Agnihotri and Liu, 2003). TAGs are stored as an energy dense molecule and beta oxidation is downstream of lipase activity, which liberates the fatty acid tails from glycerol via a hydrolysis reaction and directs fatty acids toward the beta oxidation pathway (Fan et al., 2014). Therefore, in the metabolic scheme of push, pull, protect, we may view the beta oxidation pathway as “pulling” the breakdown of fatty acids, which could be related to the breakdown of TAGs in the seed. Therefore in relation to TAG biosynthesis knocking out this enzyme may be seen as a protection mechanism, preventing already assembled TAGs from being pushed into lipase activity and ultimately the beta oxidation pathway. In an arabidopsis the knockout of a beta oxidation enzyme, 3-ketoacyl-CoA thiolase (KAT2) which is expressed post- germination, results in maintenance of TAG seed levels throughout the first five days post imbibition, although this knockout significantly inhibits seedling growth (Germain et al., 2001). However, it would seem that effects of beta-oxidation alteration during the

51 post imbibition period produces variable effects on multiple arabidopsis accessions (Khan et al., 2012), so it is plausible that alternative phenotypes may be apparent in soybeans, as there was no significant differences in emergence of seedlings in the field. Of all the genes examined in this deletion, a knockout of this beta oxidation enzyme seems to be the most plausible cause of the oil increase.

There seem to be several plausible hypotheses for how this deletion event may be influencing oil content in seeds, though further studies are needed to confirm this. First, it is essential to establish a genetic marker for this deletion to confirm that this location is the source of the high oil phenotype. In future generations, if this marker is responsible, then we would expect to see high oil content in mutants which would not contain the marker and lower oil content in those with the marker, similar though opposite of when a

VgDGAT marker was tracked in transgenic lines (Hatanaka et al., 2016). In other CGH mutant lines, this method was used to confirm a FAD2 gene deletion in the parent line, and the high oleic acid content correlated directly with the presence or absence of a PCR marker when wild type DNA was used as a template (Bolon et al., 2011). In addition, cloning of the genes in this region and inserting them into a vector or with a transgenic approach may be able to rescue the wild-type phenotype and precisely pinpoint the gene responsible for increased oil content. Ultimately, breeders and growers are not very concerned with the exact gene or mechanism which increases oil content, only that that change is established as heritable, easily identifiable with standard assays, and having limited other detrimental effects on the phenotype.

To summarize, the finding of a deletion event in high oil soybean mutant lines yielded genes which may be directly or indirectly able to alter oil content in seeds.

Ultimately more work, especially establishing a genetic marker in which the phenotype correlates with the marker will be necessary to establish this mutation as the source of high oil, and to be able to effectively track it in a breeding program.

Table 3.1. Oil Content of Parent Lines. Mean oil content (on a dry weight basis) and the standard error in parentheses, of three FN lines from the field in Minnesota and Lexington, KY and the parental line, determined on a dry weight basis via bulk seed NIRS. n=3 plots in one location.

MN KY ID Gen Notes Oil Oil 22.0 20.4 High oil, short, bushy, 1R22C28Cgadbr355aMN13 M8 (0.6) (1.3) indeterminate, late maturity High oil, erect petioles and lateral 22.4 20.1 5R12C21Dar387dMN13 M5 branches, slightly chlorotic small (0.8) (0.9) lanceolate leaves, late maturity High oil, short, slightly chlorotic 21.9 22.3 smaller lanceolate leaves, petioles 5R16C01Dar388eMN13 M5 (1.1) (1.2) long compared to plant ht., late maturity 19.0 19.4 Parent of mutant lines. M92-220 P (0.7) (0.6)

Table 3.2. Oil Content of F1 Seeds. Each F1 seed was assigned a population name for reference to the future generations raised from that seed. The first parent mentioned is the maternal plant, second mention is the father. Oil content was determined by single seed NMR on a dry weight basis, n=3 technical replications on each seed and SEs are in parentheses. Through the rest of the MS and figures the lines will be referred to by “Pop name” rather than parents for clarity.

Pop Name Parents Mass (g) OIL (% db) A1 IR22xWT 0.2544 22.5 (0.19) A2 IR22xWT 0.2181 22.8 (0.30) A3 IR22xWT 0.2401 22.5 (0.26) A4 5R12xWT 0.1742 19.3 (0.26) A5 5R12xWT 0.2092 19.6 (0.33) A6 5R12xWT 0.2027 21.6 (0.11) A7 5R12xWT 0.2741 18.9 (0.27) A8 WTxIR22 0.2734 19.3 (0.21) B1 WTxIR22 0.1562 17.3 (0.37) B2 WTxIR22 0.2771 18.9 (0.22) WT N/A 0.2051 20.5 (0.23) IR22 N/A 0.1716 22.9 (0.36) 5R12 N/A 0.1706 20.5 (1.3)

Table 3.3.Plants Selected for CGH Analysis. Bold letters indicate DNA pooled for the mutant population and non-bold are the ones used for the WT population. Oil, protein and moisture were determined from bulked F3 seed from F2 plants via NIRS using a Perten DA7200 spectrometer.

Plant ID Mean Protein Mean Oil Mean Moisture MNA1: 37.5 40.1 24.2 11.2 MNA1: 11 39.3 24.2 10.8 MNA1: 14 40.2 24.1 11.4 MNA1: 16 39.0 24.0 11.5 MNA1: 44 38.6 24.0 11.1 MNA1: 33 39.0 23.3 11.7 MNA1: 15 43.9 21.7 12.9 MNA1: 3 45.3 21.4 11.3 MNA1: 26 43.9 21.3 11.7 MNA1: 12 45.2 21.0 12.5 MNA1: 3 43.8 20.6 11.2 MNA1: 41.5 46.3 20.2 12.7 MNA2: 8 40.1 24.1 11.5 MNA2: 28 39.9 23.9 11.0 MNA2: 29 39.6 23.8 11.4 MNA2: 40 38.9 23.8 11.5 MNA2: 4 39.2 23.7 11.0 MNA2: 43 39.8 23.7 11.1 MNA2: 31 38.6 23.4 11.2 MNA2: 5 39.5 23.4 11.1 MNA2: 25 40.2 23.2 13.6 MNA2: 33 39.8 23.2 11.8 MNA2: 21 45.3 22.4 18.6 MNA2: 30 43.9 21.9 11.2 MNA2: 44.5 45.1 21.7 10.3 MNA2: 10 44.9 21.7 11.1 MNA2: 33.5 43.8 21.5 17.0 MNA2: 15 45.0 21.5 13.3 MNA2: 42 45.2 21.0 11.3 MNA2: 44 44.5 20.4 15.0 MNA3: 7 39.5 24.3 13.7 MNA3: 42.5 39.0 24.0 11.6 MNA3: 3 40.2 23.6 12.0 MNA3: 8 40.9 23.5 12.1 MNA3: 14 39.1 23.4 12.4 MNA3: 34 44.0 21.2 12.7 MNA3: 38.5 44.5 21.1 12.3 MNA3: 30.5 44.4 20.9 12.2 MNA3: 4 45.1 20.7 12.4 MNA3: 37.5 44.5 20.4 12.9 MNA3: 25 44.3 20.4 12.0 MNA3: 16 44.9 20.2 13.8 MNA3: 33 46.0 19.2 12.6 MNA3: 34.5 45.1 19.2 11.6

Figure 3.1. Scatter Plot of Oil and Protein Contents of F2 families. Scatter plots of the oil and protein content of F2 families. Seeds were analyzed for oil (x-axis), protein (y-axis), and moisture content on a dry weight basis via bulk seed NIRS. 57

Figure 3.2. CNVs in Chromosome 14. A large copy number variation (CNV) event detected in chromosome 14 of the A1 and A2 population, crosses of WT pollen to an IR22 emasculated flower, which exhibited strong anti-correlations of oil and protein content. The x-axis scale is of #SD’s from the average, with +/- 3 from the mean indicating the ratio of MUT:WT signal detection, the low ratio indicating a deletion. This event was detected on Chr14, from bp 9994086 to 10301954, a span of approximately 308kb. The y-axis is the location on the chromosome relative to bp#1 when mapping. No similar event was detected in the A3 population.

Table 3.4. Genes Present in Deletion. A large deletion event was detected on Chr14, from bp 9994086 to 10301954, a span of approximately 308kb in A1 and the A2 population. No major events were detected in the A3 population. The 20 putative genes in this region were assessed using soybase.org and the TAIR database. *= putative transcription factor

Annotatio Arabidopsis Gene Function Biological relevance n homolog Glyma.14G101000 enhancer of rudimentary putative AT5G10810. enhancer of protein 1 transcription Glyma.14G101100 Major facilitator superfamily homology AT5G10820. substrate transporter domain-containing protein 1 Glyma.14G101200 glutathione peroxidase 7 homology AT4G31870. glutathione 1 peroxidase Glyma.14G101300 glutathione peroxidase 7 homology AT4G31870. glutathione 1 peroxidase Glyma.14G101400 protein phosphatase 2C homology AT4G31860. protein 1 serine/threonine phosphatase activity Glyma.14G101500 Aspartyl-tRNA synthetase homology AT4G31180. aspartyl tRNA 1 synthetase Glyma.14G101600 BCS1 AAA-type ATPase homology AT2G18193. ATPase activity 1 Glyma.14G101700 Sas10/Utp3/C1D family homology AT5G25080. strand break repair 1 Glyma.14G101800 early nodulin-like protein 15 homology AT4G31840. electron carrier, 1 copper binding Glyma.14G101900 unknown unknown AT4G31830. unknown 1 Glyma.14G102000 Endomembrane protein 70 homology AT5G10840. integral membrane protein family 1 protein Glyma.14G102100 unknown unknown none unknown Glyma.14G102200 Sterol regulatory element- homology AT1G18400. brassinosteroid * binding protein 1 signaling Glyma.14G102300 unknown unknown AT1G21280. unknown 1 Glyma.14G102400 TTF-type zinc finger protein homology AT1G19260. Ceramide synthase with HAT dimerisation 1 domain Glyma.14G102500 Phototropic-responsive NPH3 homology AT4G31820. auxin response, family protein 1 cotyledon development Glyma.14G102600 Enoyl-CoA hydratase homology AT4G31810. beta oxidation 1 Glyma.14G102700 chorismate mutase 2 homology AT5G10870. shikimate pathway 1 Glyma.14G102800 Actin-binding FH2 protein homology AT2G25050. unknown 1 Glyma.14G102900 WRKY DNA-binding protein homology AT1G80840. pathogen resistance * 40 1

Chapter 4 COMPOSITION ANALYSIS OF A GAMMA MUTAGENIZED EARLY

FLOWERING CHIA POPULATION AND DEVELOPMENT OF NIRS

CALIBRATIONS FOR CHIA SEED OIL, PROTEIN AND MOISTURE LEVELS

Introduction

Chia (Salvia hispanica L.) is in the midst of an agricultural resurgence. This crop

served as a vital food staple and had cultural and economic significance in pre-

Columbian Mesoamerica among the Aztecs (Cahill, 2003). During the Spanish conquest,

cultural differences caused a decline in cultivation (Acuña, 1986). Chia production

started increasing in the 1990s due to its high ω-3 content (Cahill, 2004). Until recently, commercial chia farming in the United States was limited to California and Arizona, due to its photoperiod requirements (day length of 12.5 hours or less for at least two months before frost). However, researchers at the University of Kentucky recently developed long-day flowering mutant chia lines (Jamboonsri et al., 2012). These lines significantly

expanded the area of cultivation to include most of the heartland.

Clinical trials, epidemiological investigations and experimental studies suggest

consumption of α-linolenic acid (ALA - 18:3n3, ω-3) has a positive impact on cardiovascular disease (CVD) and may also have a positive impact on diabetes mellitus.

Flaxseed with its high ALA content has been found to decrease the risk of CVD and diabetes (Prasad, 2009; Taylor et al., 2010). A particularly rich source of ALA is Salvia hispanica, which has greater ALA content than flaxseed, (Table 4.1) but there is very little research on its beneficial effects and the data that are available are inconsistent

(Nieman et al., 2009; Vuksan et al., 2009; Vuksan et al., 2007). One line of S. hispanica,

Salba®, has shown some risk reduction benefits regarding cardiovascular health and

blood glucose regulation (Vuksan et al., 2009; Vuksan et al., 2007) which may be due to

its higher ALA, antioxidant and/or fiber content or composition.

Chia has utility in the feed industry which could exceed human consumption.

Feeding chia to egg-laying hens results in a significant accumulation of ω-3 fatty acids in the eggs, but without the fishy or off taste associated with other supplements (Ayerza and

Coates, 1999). Chia also has the potential to be a less expensive source of ω-3 s, as well as more environmentally sound than harvesting krill for oil and some unsustainable fish sources. ω-3 -enriched eggs are now available in a wide variety of supermarkets and have been marketed successfully (Simopoulos AP, 2009).

Many members of the Lamineaceae family exhibit high levels of ω-3 FAs. Chia

has oil composition levels similar to its Lamiaceae relative Perilla frutescens (Ciftci et

al., 2012), among the highest ω-3 sources (Table 4.1). Dracocephalum moldavica of this

family averages 68% ω-3 FA (Rao et al., 2008; Western, 2012). However, only chia and

Perilla have high seed and oil yield potential, making them two of the most economically

feasible members of the family to develop for large scale crop production (Rao et al.,

2008). Chia seeds contain about 20% protein, and oil content ranges from 28.5-32.7%

(Ayerza and Coates, 2004; Ayerza and Coates, 2007). There are claims that white chia

seeds are higher in oil and ω-3 FA levels than dark seeds, though genes regulating seed

pigmentation are independent from those regulating oil composition (Ayerza, 2010).

Though chia is primarily valued for its ω-3 content, it is also a good source of

protein with a high level of essential amino acids (Olivos-Lugo et al., 2010). Of the

complete amino acid profile, about 30 to 34% are essential amino acids and it serves an

adequate source of all except lysine (Ayerza, 1995; Olivos-Lugo et al., 2010). This amino

61 acid profile is considered well balanced in regards to the ratio of essential to non-essential amino acids and the appreciable levels of essential amino acids could supplement a balanced diet (Ullah et al., 2015).

In plant breeding, random mutagenesis is a common way to generate mutations and increase genetic diversity for traits with limited natural variation. While efforts were made to screen out undesirable phenotypes in this population such as plant architecture, it remains unknown if the early flowering lines have other aberrant phenotypes such as altered oil content or other changed seed composition traits, factors which may seriously affect the value of the crop. To address these potential concerns, we have screened populations grown in 2010 and 2011 for a variety of seed characteristics, including protein, oil, fiber and mineral analysis.

Near-infrared spectroscopy has been successfully utilized as a seed analysis technique for many major crop species, including corn and soybeans. These methods are fast, accurate and repeatable, requiring little to no sample preparation, while allowing for simultaneous observations of several composition traits, and these methods are non- destructive so seeds can be planted after assessment (Armstrong et al., 2011). One objective of this study was to determine oil, protein, fatty acid, amino acid, and trace element values for 120 chia seed lots. A second objective was to assess the variability of these parameters with respect to their NIRS wavelengths, and use partial least squares 2

(PLS2) regression to develop calibrations for whole and ground chia seed oil, protein and moisture. To this end we were able to develop quality calibrations for protein, oil and moisture content of whole and ground chia seeds.

Materials and Methods

Chia Seed Samples A total of 120 chia seed lots were analyzed. Several commercial chia samples were obtained locally or from internet sources, including samples from Argentina,

Australia, Mexico and Peru. Chia samples from Kentucky field grown material came from several years of production. In 2010, over 50 new long day flowering chia lines, developed through ethyl methanesulfonate (EMS) and gamma radiation mutagenesis, were grown in a seed yield trial and in increase blocks. These plots had no fertilization and were grown in Maury silt loam soil, sown at 6 kg/ha seeding rate in plots measuring

1.5 x 6 m, with 7 rows at 18 cm between rows and planted in late May. In addition, several lines were placed in a fertilization trial at different N rates. Seed was harvested after frost using a Hege 140 small plot combine with screens for small seeds. To measure composition variability, seed from a G8 field production block grown in Lexington, KY in 2011 was run on a gravity table and divided into 9 fractions, labeled as GT 1-9, where

GT1 was the densest fraction of seed.

Nitrogen and Phosphorus Analysis

Samples were prepared for analysis by weighing 100 mg of dried material into

25x200 Pyrex glass ignition tubes marked at 50 mL. Five mL of concentrated sulfuric acid containing 0.05 g of salicylic acid/mL were added and the samples reacted for one hour at room temperature. This step caused any inorganic nitrate present in the sample to form nitrosalicylic acid. Next, 0.5 g of sodium thiosulfate was added and the samples were placed in a Technicon BD-40 block digester set at 180 °C for one hour. This resulted in reduction of the nitrosalicylic acid to the less refractory compound aminosalicylic acid. Then 1.8 g of potassium sulfate and boiling chips were added and the

digestion was continued for 2.5 more hours at 360°C. All forms of nitrogen were

converted to ammonia and all forms of phosphorus were converted to orthophosphate

during this process. The samples were allowed to cool then diluted to 50 mL with

deionized water and analyzed.

The instrument used for colorimetric determination of total nitrogen and total

phosphate was a dual Technicon System II Autoanalyzer which was configured to

perform both analyses simultaneously. The wavelength was 660 nm for each procedure.

The method for ammonia was a modification of the Berthelot reaction developed by

Chaney and Marbach (1962). Two reagents were used, one containing 0.5% sodium

hydroxide and 0.042% sodium hypochlorite in deionized water, and the other containing

1.0% phenol and 0.02% sodium nitroprusside in deionized water. The sample was

introduced into a bubble segmented stream followed by the reagents. The reaction took

place inside the instrument, and the blue indophenol formed was passed through a

colorimeter for final determination of ammonia concentration. The original manual

method was modified to speed up the reaction rate in order to make it compatible with the

constraints of the automated system. It was necessary to increase the nitroprusside

catalyst concentration to four times the recommendation and pass the reaction stream

through a 60°C heating bath, though the protocol was otherwise similar to the manual

method.

The phosphorus technique was based on the method of Fiske and Subbarow

(1925) and is the same as Technicon Industrial Method 348 R 6-3 1-5, except the dialysis step was not necessary for digested samples. A solution of ammonium molybdate (7.5 g/L) in 1.92 molar sulfuric acid was reacted with the samples in the segmented stream to

form a hetropoly phosphomolbdate complex. This compound was then reduced by adding

a solution containing 150g of sodium bisulfite, 5.0 g sodium sulfite and 2.5 g 1-amino-2-

naphthol-4-sulfoninc acid in 1 L deionized water and heating the reagent stream to 95°C in an oil bath. The reaction resulted in the formation of an intense blue color proportional to the phosphate concentration. Four standards and a blank were run before and after each set of 15 samples to minimize baseline drift.

Trace-element analysis

For the elements Ca, Mg, P, K, Cu, Fe, , Mn, Mb, Na, Zn and Al, chia seeds were

ground to a fine powder, ashed in a muffle furnace for 4 hours at 450 C, then digested

using a 14:5:1 ratio of water:hydrochloric acid:nitric acid, with final acid concentration of

4% HCl and 1% of HNO3. All values listed are reported in mg/kg. Samples were acid

digested in duplicate and analyzed using AOAC method 968.08 with a Varian Vista Pro

ICP-OES.

For Se, B, V, Cr, As, Cd, and Pb, chia seeds were ground to a fine powder and

100 mg sub-samples were digested in 1 mL of a 3:1 mixture concentrated trace-metal

grade HNO3 and 20% H2O2. The samples were heated in metal-free polypropylene

centrifuge tubes to 110 °C over 30 min and refluxed at that temperature for 10 min using

a temperature-controlled microwave reaction system (MARS Xpress, CEM, Matthews,

NC, USA). Standard reference materials (SRM 1573a, tomato leaves, National Institute

of Standards and Technology, Gaithersburg, MD, USA) reagent blanks and duplicate

digestions for randomly selected samples were included with each digestion set. The

samples were then brought to 15 mL with 18 MΩ de-ionized water (DI). Aliquots were

spiked with an internal standard mixture to achieve 1 ug/L of Sc, Ge, In, Tb and Bi.

Randomly selected samples were fortified with calibration standard to determine spike

recovery. The samples were analyzed using an inductively coupled plasma mass

spectrometer equipped (ICP-MS; 7500cx, Agilent Technologies, Santa Clara, CA, USA)

operating in standard mode for B, V, Cr, As, Se, Cd and Pb. Where possible, multiple

isotopes were used for quantification to check for spectral interferences. Calibration was

performed by serial dilution of a certified reference standard in a matched matrix (5%

ultrapure HNO3 and 5% n-butanol in DI) and was verified by analyzing a dilution of an

independent certified standard of a different lot number (Inorganic Ventures,

Christiansburg, VA, USA). Butanol was added to the calibration standards to normalize

the C content, since C causes positive interferences for Cr and Se. Samples where both

reps were beneath the MDL (minimum detection limit) are reported as “BDL”, beneath

detection limit. For strains where only one rep exceeded the MDL, the value for the

detected rep is given and there is and * in the SE section in Table 4.2. For strains with

each rep above the MDL, means and SEs were calculated.

Fiber analysis

Chia seeds were ground to a fine powder in a burr mill analyzed for crude fiber by

AOAC Method 978.1 and for Acid Digestible Fiber via AOAC Method 973.18 (Latimer,

2012).

Fatty Acid Analysis

10 mg of seeds were ground using a mortar and pestle and with tri-17:0 to quantify acyl-lipids. Oil was extracted from seed chips using 500 uL of diethyl either with 0.0001% BHT twice and dried. Once dry, 500 uL sodium methoxide was added and shaken for 10 minutes with 1 mL isooctane containing 0.001% BHT added after. 200 uL

was pulled off the upper layer and transferred into a GC vial with an additional 1 mL of

isooctane added.

GC samples were then run on a Varian CP-3800 Gas Chromatograph using a 25m x 0.25 mm ID fused silica column with a Varian (chrompack) CP=Select CB for FAME,

with a film thickness of 0.25 um. The temperature program ran from 90 °C to 250 °C

with 25 °C ramp for a total of an 8 minute run time with a constant column flow mode of

0.9 mL/min utilizing a splitless injection. Quantification was performed by using a flame

ionization detector and peaks quantified using Star Chromatography Workstation Version

6.00, with peak area being used to calculate relative percentages of FAMEs.

Oil Extraction

Oil extraction was performed according to AOCS Method Am 2-93 (Anon, 2013)

using petroleum ether (bp 30-60° C) as the extraction solvent and defatted, oven dried

paper towels in substitution for thimbles. Approximately 2 g of seeds ground in a burr

mill were used for extraction in a Soxhlet extraction apparatus for at least 30 cycles.

Phytate Analysis

Phytate was estimated by subtracting total phosphate from total phosphorus.

Phosphate was extracted using a protocol from (AL-Amery et al., 2015). 30 mg chia seed

was ground to a fine powder and defatted using 2 mL of petroleum ether twice. 10 mg of

defatted seed was treated with 50 µL extraction buffer (12.5% trichloroacetic acid and 25

mM MgCl2) per mg and samples are incubated 16 h at 37˚C with gentle shaking, and

then centrifuged at 100 g for 3 min. Three 10 µL subsamples are taken from each

extracted sample and diluted with 90 µL of H2O and mixed with 100 µL of Chen’s

Reagent (10% ascorbic acid, 3 M sulfuric acid, 2.5% Ammonium molybdate and water

(1:1:1:2) (Chen et al., 1956). The reaction was allowed to proceed at room temperature

for 1 h. The absorbance was measured at 882 nm and phosphate levels determined from a

phosphate standard curve. Total phosphorus levels were determined as described above

(see Nitrogen and Phosphorus Analysis method section).

NIR Methods and Prediction Model Development

Near infrared reflectance (NIR) spectra of bulk whole and ground chia seed were

obtained from a DA 7200 (Perten Instruments, Springfield, IL, USA). Spectra ranged

from 950 to 1650 nm at 5 nm intervals at a bandwidth of 1 nm. Three spectra readings

with repacks were obtained for each of the 120 samples. A 75 mm, white sample cup

supplied with the instrument was used for spectra collection. GRAMS/AI, Thermo

Scientific, Waltham, MA was used for model development using PLS2 regression. No

pre-processing was done to spectra and replicates were treated as individual data, i.e. a total of 120 x 3 spectra were used for models. All prediction models were developed

using leave-one-out cross-validation analysis to select factor levels and reporting of

model statistics. Oil and protein models used replicates as unique observations while fatty

acid models used the average spectra of the replicates. Data for all 120 samples were not

always available due to either reference analysis not being done, lack of sample material

or suspect reference analysis.

NMR Methods

Oil content was determined by NMR using a Minspec 20 (Bruker Biospin, The

Woodlands, TX, USA). The instrument accommodates a twenty mm sampling tube diameter. Approximately 1.5 g of whole chia seed was used. Seeds were weighed and placed into the tube and allowed to warm to 40o C before insertion into the instrument.

The standard oil seed measurement procedure supplied with the instrument controller was

used. Samples were then oven dried at 103o C for 24 hrs to obtain moisture content.

These were used for moisture adjustment to dry basis oil content. Calibration of the NMR instrument was done using weighed amounts of extracted chia oil encompassing the range of oil weights of the seed samples. Four different weights were used and were expressed onto tissue paper at the bottom of the 20 mm sample tube. The coefficient of determination between the NMR signal calibration and oil weights was 0.9998.

Amino Acid Analysis

Seed samples defatted with petroleum ether were ground to a fine powder using a burr mill on the finest setting. To determine the amino acid composition of the seeds, 100

– 200 mg of defatted, whole ground samples were mixed with 300 μL of the internal

standard 100 mM norleucine, and the samples were hydrolyzed for 24 hours at 110 C in

6N HCl, in accordance with AOAC procedure 994.12. To determine methionine content,⁰

~50 mg of ground sample underwent performic acid oxidation procedures, as described

in AOAC method 994.12, prior to the acid hydrolysis procedures. Following hydrolysis

procedures, samples were converted to their phenylisothiocyanate derivatives and

analyzed using reverse phase HPLC (3.9 X 300 mm PICO-TAG reverse phase column;

Waters Corporation, Milford, MA)(Bidlingmeyer et al., 1984; Cohen and Strydom,

1988). All samples were run in duplicate, with an inter-sample variation of less than 10%.

Results

Constituent Analysis

Chemical analysis of seed lots shows wide variation in oil content, fatty acid

composition, protein and fiber. Protein content of all seed lots ranged from 18.2% to

28.2% with a mean content of 22.8%. Our analyses show crude fiber composition ranged

from 20.9% to 27.5% with a mean of 23.6%. Total phosphorus was a minor component

with 0.55 to 1.0% range. Most of the phosphorus was bound in the form of phytate,

averaging 0.51% but with a range of 0.34 to 0.66%. Free phosphorus, which would be

only form available to monogastric animals, was only 0.35% overall, with a range of 0.19

to 0.54%.

Oil content of these chia lots ranged from 21.4% to 35.3% with a mean of 31.3%.

There was notable variation in fatty acid composition, with the most important being 18:3

(ω-3) ranging from 33.9% to 66.4% with a mean of 56.7%. 16:0 had the largest coefficient of variation at 23.5% and a range from 6.8 to 19.5%. 18:1 and 18:2 had ranges of 6.3 to 13.9 and 17.1 to 30.5. 18:0 was a minor constituent in all samples with a range of 2.1 to 5.4%.

Of 120 seed lots, 15 were used for extended compositional analysis in Table 4.3.

Samples were selected based on seed quantity and variation of constituents relevant to

NIRS calibrations including oil, protein, fatty acid composition, amino acid variation and

seed color. Oil content ranged from 27.9% to 33.8%, and ω-3 FA content ranged from

52.2% to 62.5%. Protein ranged from 18.7 to 23.4%, fiber ranged from 20.9% to 27.1%

and the 100 seed weight was between 69 and 112 mg.

Concentrations of nutritionally important elements in the seeds are presented in in

Table 4.4. The major elements were analyzed by ICP-OES, including Ca (6.3 g/kg), Mg

(6.9 g/kg), K (6.9 g/kg) and P (9.4 g/kg). Meanwhile, micronutrients including, Co (0.15

mg/kg), Cu (19.29), Fe (99.8 mg/kg), Mn (3.9 mg/kg), Mb (38.7 mg/kg), Na (8.5 mg/kg)

70 and Zn (54.8 mg/kg) were also assessed via ICP-OES and Se (0.09 mg/kg) was assessed via ICP-MS.

Elemental data for each individual strain is reported in Table 4.5. Of the macronutrients, Ca ranged from 5.4 to 7.3 g/kg, Mg 3.6 to 4.3 g/kg, P from 7.6 to 9.3 g/kg, and K 6.5 to 7.3 mg/kg. All of these elements had concentrations with <10% coefficient of variation. However, micronutrients had higher CVs and wider ranges. The lowest variation was observed in Mn, with a range of 3.6 mg/kg to 4.3 mg/kg.

Table 4.6 lists the concentrations of potentially toxic trace-elements found in these chia seed lots. Only Al was tested via ICP-OES and all other elements were measured via ICP-MS. B, Al, V, Cr, As and Cd, concentrations below recommended daily intake levels when consuming a 24 g serving. Only one sample was above the recommended daily intake for Pb.

The SRM used for ICP-MS analysis had certified concentrations for B, V, Cr, As,

Se and Cd. Recovery of these trace elements was generally within 15% of the certified concentrations with a CV of <20% with a few exceptions. Recovery of V and Cr were only 54, and 67% respectively, likely due to incomplete solubilizing of these elements.

Mean recoveries of As and Se were 145% and 266% respectively. This is likely due to spectral/matrix interferences which were important at the low concentrations of these elements present in the SRM (0.112 mg/kg for As and 0.05 mg/kg for Se). The certified concentration of Se was 0.05 mg/kg which was above MDL 0.02 mg/kg but below the

MQL of 0.06 mg/kg. The MDLs ranged from 0.01 mg/kg for Pb to 6.59 mg/kg for Ca.

NIR modeling

NIR prediction models were developed using PLS2 regression for oil, protein,

moisture and fatty acid composition of whole and ground chia seeds. NIR spectra

obtained from a DA 7200 (Perten Instruments, Springfield, IL) were regressed against

Soxhlet and NMR observations for oil, micro-Kjeldahl method for protein determination and GC-FID for fatty acid composition.

Figure 4.1 shows the NIR spectra obtained for whole seed samples. The spectra

show consistent profiles for whole seeds with significant variability in reflectance for

individual samples. There are pronounced differences from 950 nm and 1200 nm

between samples which are uncommon for many other seeds types. A profile such as this

shows consistency and variability between samples, which is necessary to generate a

robust prediction model.

Oil and protein models for whole seeds (Figure 4.2a and 4.3a) and ground seeds

(Figure 4.2b and 4.3b) show good correlations between predicted and analytical values

yielding R2s equal or greater than 0.76 for cross validation. Similar results are seen for

moisture calibrations in Figure 4.4, though ground seed models were inaccurate due to oil

oxidation during drying which is common in highly unsaturated seed such as chia. In

practice this results in an accuracy of approximately +/- 1% composition for oil and protein values; RMSE (root mean square error) values were near 1% for oil and protein models at PLS2 model factor levels of 12 to 14. RMSE values for the moisture model was 0.26% at a factor level of 12.

Nitrogen rate assessment

A subset of these samples was subject to different rates of nitrogen application: 0,

37, and 74 kg/ha. A randomized complete block design was utilized and analyzed via

ANOVA, followed by means grouping using Tukey’s assessment via JMP software. The nitrogen application rate was found to significantly affect protein levels (p=0.003) in a ten cultivar field scale study in 2010 (Figure 4.5). No other composition traits were affected significantly, though oil levels decreased by 1% as nitrogen increased by 1%.

Like many other oil crops, a significant negative correlation between protein and oil levels was observed in the N application study.

Discussion

We screened populations grown in 2010 and 2011 and from various commercial sources for a variety of seed characteristics, including protein, oil, fiber and mineral analysis. For major components, including oil, protein and fiber, the level of variation in these populations corresponds with variation observed in populations of the crop’s natural range. However, when comparing fatty acid composition, there is a large range for the ω-3 fatty acid, (ALA 18:3), with some seeds lots containing less than 40% 18:3, which would affect the value and nutritional properties of the oil by decreasing the the ω-

3 level and ratio of ω-3 to ω-6 fatty acids (Monroy-Torres et al., 2008). The amino acid profile is comparable to other cultivars grown in the crop’s native regions and has low variation in this population. Elemental composition was established for chia of the major elements Ca, Mg, K and P. There are also detectable levels of Cu, Fe, Mn, Mo, Na, Se and Zn, though studies need to be conducted to determine how bioavailable these would be to humans and animals. There are also detectable levels of potentially toxic elements

such as V, Cr, As, Al, Cd and Pb. The NIRS calibrations allow for rapid, non-destructive

screening of small seed lots, as little as a single plant’s yield, to measure oil, protein and

moisture to +/- 1% accuracy.

Chia lots grown in Lexington from 2010 to 2011 had a mean oil content of 31%,

though ranging as low as 21% with some cultivars up to 35%. A USDA assessment

places average oil content across cultivars and locations at 30.7% (Anon, 2016; Suri et

al., 2016), very close to the mean value obtained for Kentucky chia. A five location study

in northwest Argentina showed a mean of 36% oil content throughout all locations, only

one of which was irrigated (Ayerza, 1995). However, other multi-location studies have

shown much lower contents, from 27% in Jalisco and Sinaloa cultivars from Mexico

(Alvarez-Chavez et al., 2008), as well as this study’s mean, 31%, from 9 locations

throughout South America (Ayerza and Coates, 2004). In Ecuador and Bolivia, the mean

ranged from 27% to 36% in different locations and environments (Ayerza, 2016). While

more extensive studies will be necessary to determine genetic heritability of oil content in

these mutagenized cultivars, many of them seem to have exhibited little or no genetic

losses or oil yield.

Compared to chia grown in traditional locations, major fatty acids are similar to

those noted in the studies above. However, 18:3 fatty acids are near the low end of the

spectrum compared to other locations and 18:1 is near the higher end. Several accessions

are less than 50% 18:3, which could result in a great loss in value of the crop. However, some lots are as high as 66%, with a mean of 57%. Many studies report 60% to 62% 18:3 over several locations (Alvarez-Chavez et al., 2008; Ayerza, 1995; Ayerza, 2009; Ayerza and Coates, 2004). This may lead to greater oxidative stability of the oil compared to chia

grown in traditional areas, while producing insignificant difference in health outcomes.

Environmental effects observed in the Tzotzol variety show a range of 58.5% to 64.5%,

and a mean difference of 3% between two locations in Ecuador and Bolivia (Ayerza,

2016). Individual lines will require greater scrutiny to establish ranges for oil

composition.

Fiber analysis showed chia contains crude fiber content of 22-27%. In one study, chia flour was separated into fibrous and protein fractions which showed that the total crude fiber content was around 29% (Vazquez-Ovando et al., 2010). Acid digestible fiber in KY chia ranged from 27% to 60% with a mean of 40%. Acid digestible fiber in

Mexican chia samples was around 45%. Total dietary fiber of chia was not assessed but chia typically it contains 36 to 40 g per 100g, much higher than corn, carrots, spinach, banana, pear apple and kiwi. Of common oilseeds, flax is the closest with 27.3% dietary fiber (Suri et al., 2016). Fiber inheritance is a complex trait involving multiple genes, so genetic effects on fiber accumulation are unlikely, though little is known about environmental effects on fiber accumulations.

Overall protein content of chia seed is high and it contains high levels of essential

amino acids. As a total proportion of all amino acids, essential amino acids in this study

ranges from 36% to 42%, which is similar to values reported to soybeans. From various

accessions across cultivars and locations, the range of protein is 15 to 23 percent, much

greater than other grains such as wheat, oats, barley, rice and corn (Suri et al., 2016). One

study found the percentage of essential amino acids in chia to be as great as 46%, which

are levels that cover up to 76% of infants needs and enough to meet adults’ dietary needs

(Sandoval-Oliveros and Paredes-López, 2013). Chemical scoring of chia protein shows

that Lys, Leu and Thr will be the limiting amino acids, so it is recommended to

supplement them with additional sources of protein (Olivos-Lugo et al., 2010). Protein

values from Mexican chia are similar to those found in this study and corresponds with

the studies above (Suri et al., 2016; Ullah et al., 2015), indicating that the geography and

mutagens did not cause significant changes in the amino acid profile of these seed lots.

Chia also appears to be an excellent source of minerals, including chia grown in

Kentucky. Calcium, phosphorus, iron and zinc concentrations are a near match to USDA

values for standard chia reference (Ullah et al., 2015), but there are double the amounts

reported of magnesium and potassium in this study. Sodium concentrations are also much

higher than the USDA reported standards of 1.6 mg/kg, though KY chia can still be

classified as a low sodium food (Suri et al., 2016). Furthermore, potentially toxic trace-

element concentrations were low and far below recommended intake levels for adult

daily consumption based on a 24g daily serving size The elements B (Price et al., 1996),

Al (Golub et al., 2000), V (Domingo et al., 1985), Cr (Dowson, 1992), As, and Cd

(Baars, 2001) were below recommended intake levels. Only one sample, GT3, was above

recommended daily intake of one element, Pb, from a recommended daily serving size of

24 g. However, all others samples tested tenfold lower for Pb than this one.

A new growing region for chia has been established using new long-day

flowering lines. These long-day flowering lines were grown in multiple years and

locations in Lexington, KY. These seed lots and commercial seed lots from multiple locations in the world were analyzed for oil, protein, fatty acid, amino acids, minerals and moisture levels. The mean values and variance of the long-day flowering KY grown seed

lots were similar to chia from other locations and literature values. The mean values for

76 oil content ranged from 21.4 to 35.3%, protein ranged from 18.2% to 28.2%, and the ω-3

α-linolenic acid ranged from 33.9 to 66.4%. Using these seed lots, non-destructive near infrared spectroscopy (NIRS) calibrations were developed for whole and ground seed oil and protein contents and whole seed moisture content with accuracies of +/- 1% for protein and oil contents and 0.25% for moisture content. We were unable to develop accurate NIRS calibrations for fatty acid levels. As more phenotypic and genotypic variation is introduced in chia, we anticipate that the variability will be sufficient to establish calibrations for that as well.

Table 4.1. Fatty Acid Profile of High Omega-3 Seeds. Pinta is the parental variety of chia G8 originally from Mexico, and G8 is an early flowering variety grown in Lexington, KY. Salba® is a trademarked Salvia hispanica variety grown in Peru.

Pinta G8 Salba Flax Perilla China Perilla Korea 16:0 7.7 8.1 10.2 5.9 7.2 7.4 18:0 4.0 4.1 3.8 4.1 2.9 3.6 18:1 8.5 9.2 6.9 19.9 20.8 9.5 18:2 19.4 21.5 12.1 15.0 10.5 16.5 18:3 60.3 57.0 67.0 55.2 52.6 63.0

Table 4.2. Summary Statistics of Constituent Analysis of Seed LotsTable 4.2. Summary Statistics of Constituent Analysis of the seed lots. Constituent Analysis was performed as described in materials and methods. All values are expressed as g/100 g seed, though oil composition numbers are expressed as % fatty acid. 100 seed weight is expressed in grams. SE=Standard Error, CV=Coefficient of Variation and n= number or seed lots surveyed. All units are corrected to a dry weight basis.

Constituent Mean SE CV Min Max n Soxhlet Oil (Recovered) 32.8 0.50 6.1 25.8 34.2 16 NMR Oil 31.3 0.20 6.9 21.4 35.3 116 NMR Moisture 8.1 0.08 10.8 3.1 10.0 116 Total Phosphorus 0.94 0.01 7.3 0.55 1.0 118 Free phosphate .34 .03 32.4 .19 .54 15 Phytate (calculated) .51 .03 21.6 .34 .66 15 Protein 22.8 0.19 9.4 18.2 28.2 120 % 16:0 8.3 0.19 23.5 6.8 19.5 109 % 18:0 4.0 0.06 15.5 2.1 5.4 109 % 18:1 9.5 0.16 17.7 6.3 13.9 109 % 18:2 21.5 0.23 11.0 17.1 30.5 109 % 18:3 56.7 0.52 9.5 33.9 66.4 109 Crude Fiber 23.6 0.52 8.6 20.9 27.5 15 AD Fiber 40.6 2.17 20.7 27.1 60.1 15

Table 4.3. Summary Data of Extended-Analysis Samples. Fifteen samples were selected which had abundant seed material and diverse germplasm traits. Oil, protein, crude fiber and acid digestible fiber are expressed % seeds. 100 seed weight is expressed in milligrams. Fatty acid composition is expressed as % fatty acids. SE presented next to mean as ±.

Oil (soxhlet) Oil (NMR) % Protein % 16:0 % 18:0 % 18:1 % 18:2 % 18:3 Crude Fiber AD Fiber 100 seed wt Pinta 32.7 ±0.4 32 ±0.3 21.4 ±0.2 7.8 ±0.1 4 ±0.1 7.8 ±0.2 19.6 ±0.9 60.8 ±0.9 22.9 ±0.9 42.8 ±5.3 107.5 ±0.96 G3 32.7 ±0.5 33.8 ±0.2 21.8 ±0.1 7.9 ±0.3 4.3 ±0.3 8.7 ±0.6 20.3 ±1 58.8 ±0.8 27.5 ±0.3 45.6 ±9.2 98.2 ±1.98 G6 32.6 ±0.7 31.2 ±1.4 23.4 ±0.2 8.3 ±0 4.5 ±0.4 11.2 ±1.9 22.3 ±1.7 53.6 ±3.8 21.5 ±0.2 27.1 ±1 99 ±3.16 G8 32.7 ±1 33.9 ±0.6 22.1 ±0.3 7.9 ±0 4.7 ±0.2 9.4 ±1.1 20 ±0.4 58 ±1.5 26.5 ±0.4 41.1 ±2.7 104 ±0.86 G15 33.9 ±0.8 32.4 ±0.6 23.2 ±0.4 7.8 ±0 5.1 ±0.2 9.7 ±1 20.7 ±0.4 56.6 ±1.4 24.8 ±1.2 40.1 ±1.6 95.8 ±3.35 E9 32.9 ±0.8 33.1 ±0.4 22.4 ±0.1 7.3 ±0.2 5.1 ±0.1 10.2 ±0.9 20.8 ±0.7 56.6 ±1.7 27.1 ±0 60.1 ±4.3 96 ±1.07 GT1C 33.9 ±0.2 27.9 ±0.5 18.7 ±0.1 7.5 ±0.3 4.7 ±0.1 8.7 ±0.4 18.4 ±0.3 60.8 ±0.6 22.9 ±0.4 33.8 ±4.3 77.2 ±2.22 GT2 32.8 ±0.3 30.7 ±0.3 19.2 ±0.2 8.5 ±0.5 4.6 ±0.3 8.7 ±0.1 20.9 ±1.6 57.4 ±1.6 20.9 ±0.2 34.3 ±4.9 89.5 ±1.55 GT3 34.1 ±0.4 32.4 ±0.4 18.9 ±0 6.9 ±0.4 3.9 ±0.1 8.4 ±0.2 20.4 ±1.6 60.4 ±1.3 23 ±0.8 33.9 ±1.8 94.6 ±1.28 GT4 34.2 ±0.2 33.6 ±0.3 18.7 ±0.2 11.6 ±3.1 3.9 ±0.2 10.7 ±1.5 21.7 ±1.5 52.2 ±6 21.9 ±0.4 37.9 ±0.9 96.3 ±1 GT5 34.2 ±0.8 32.9 ±0.5 18.8 ±0.1 7.6 ±0.2 4.4 ±0 9.5 ±0.4 20.9 ±0.9 57.5 ±0.4 22.4 ±0.2 43.2 ±0.2 98.1 ±0.22 GT6 33.1 ±1 32.9 ±0.2 18.2 ±0 6.8 ±0.1 4.4 ±0.4 9 ±0.6 20.1 ±0.9 59.8 ±1.8 23 ±0.2 36 ±3.5 102 ±0.05 GT7 32.5 ±0.9 33.5 ±0.4 19.3 ±0.1 7.7 ±0.4 4.3 ±0.3 9.5 ±0.4 20.8 ±0.2 57.7 ±0.6 24 ±0.6 36.9 ±0.1 103.8 ±1.46 GT8 34.2 ±0.7 33.8 ±0.2 19.6 ±0.8 7.5 ±0.2 4.2 ±0.1 8.5 ±0.7 18.8 ±1 61.1 ±1.8 23 ±0.2 41.5 ±0.7 112.2 ±2.54 GT9 33.1 ±0.6 34.2 ±0.1 20.6 ±0 7.6 ±0.1 3.7 ±0.2 8.1 ±0.9 18 ±0.5 62.5 ±1.6 22.1 ±0.3 55.2 ±4.3 107.7 ±2.03

Table 4.4. Summary Statistics for Trace Essential ElementsTable 4.4. Summary statistics for the trace essential elements. For all minerals, n=15, each sample value is from the mean of two subsamples. Ca, Mg, K and P are expressed as g/kg. Se, Co, Cu, Fe, Mn, Mb, Na and Zn are expressed as mg/kg. All elements were assessed via ICP-OES, except Se, which was assessed via ICP-MS.

Constituent Mean SE CV Min Max n Macro-minerals(g/kg) Calcium 6.3 0.6 9.2 5.4 7.3 15 Magnesium 6.9 0.27 3.9 3.6 4.3 15 Phosphorus 8.5 0.41 4.8 7.6 9.3 15 Potassium 6.9 0.10 6.1 6.5 7.3 15 Micro- minerals(mg/kg) Selenium 0.09 0.01 7.7 0.04 0.16 15 Copper 19.3 1.5 7.9 16.9 22 15 Iron 99.8 11.7 11.7 86.3 122 15 Manganese 3.9 0.22 5.6 3.6 4.3 15 Molybdenum 38.7 6.9 17.7 33.3 55.9 15 Sodium 8.5 4.1 4.8 7.5 9.3 15 Zinc 54.8 8.1 14.8 48.9 81.8 15

Table 4.5. Mineral Composition of Extended Analysis Samples. Samples were analyzed after ashing in a muffle furnace then subjecting ashes to optical emission spectrometry. Macro-minerals are expressed as g/kg seed and micro-minerals as mg/kg seed. All elements were assessed n=2 technical reps via ICP-OES, except Se, which was assessed via ICP-MS. Data are reported as means ±SE

Sample Ca Co Cu Fe K Mg Mn Mo Na P Se Zn 7.1 0.2 16.9 3.8 55.9 0.25 8.9 E9 ±0.02 ±0.01 ±0.04 100 ±3.5 7.2 ±0.07 ±0.02 ±0.04 ±0.02 20.1 ±0.37 ±0.09 0.07 52.9 ±0.36 5.8 0.19 19.1 3.8 34.5 0.74 7.6 Sales ±0.08 ±0.03 ±0.5 91.3 ±0.6 6.8 ±0.1 ±0.03 ±0.03 ±0.01 13.2 ±0.21 ±0.16 0.04 54.8 ±0.76 5.9 0.17 21.1 102.1 4.2 38.7 0.68 G3 ±0.02 ±0.02 ±0.6 ±8.4 6.6 ±0.07 ±0.02 ±0.38 ±0.23 45.6 ±16.92 8.9 ±0.1 0.04 49.6 ±0.31 7.2 0.2 16.9 3.6 0.34 9.3 G6 ±0.05 ±0.02 ±0.7 122 ±6.6 7.3 ±0.1 ±0.06 45 ±0.66 ±0.01 29.7 ±1.21 ±0.16 0.08 55.1 ±1.28 5.4 0.08 17.3 4.1 34.9 0.47 8.5 G8 ±0.05 ±0.01 ±0.4 96 ±1.1 6.8 ±0.01 ±0.01 ±0.13 ±0.03 18.1 ±0.33 ±0.01 0.11 51.6 ±0.25 7.3 0.2 19.3 3.6 51.5 0.34 8.9 G15 ±0.06 ±0.04 ±0.1 97.7 ±4 7.2 ±0.05 ±0.02 ±0.13 ±0.07 24.7 ±0.75 ±0.07 0.06 54.7 ±0.09 6.9 0.26 107.8 4.3 39.4 0.6 8.3 GT1 ±0.05 ±0.01 22 ±0.1 ±16.2 7.2 ±0.03 ±0.02 ±0.23 ±0.01 125.8 ±0.96 ±0.07 0.14 60.1 ±0.17 6.4 0.23 21 113.5 4.3 38.6 1.04 8.7 0.10 GT2 ±0.01 ±0.05 ±0.01 ±3.7 7.2 ±0.02 ±0.01 ±0.01 ±0.3 208.6 ±5.57 ±0.01 ±.01 81.8 ±1.9 6.1 0.18 19.5 4.1 35.5 8.5 0.14 GT3 ±0.04 ±0.01 ±0.1 93.2 ±1.1 6.8 ±0.07 ±0.03 ±0.01 0.61 ±0 100.2 ±4.89 ±0.09 ±0.01 57.6 ±0.03 6.1 0.11 19.3 4.1 35.2 0.62 8.3 0.12 GT4 ±0.02 ±0.04 ±0.01 91.3 ±1.9 6.8 ±0.04 ±0.03 ±0.36 ±0.03 75.9 ±0.75 ±0.06 ±0.01 53.5 ±0.90 0.11 19.5 4.1 34.7 0.7 8.4 0.07 GT5 6 ±0.05 ±0.01 ±0.2 97.8 ±7.5 6.9 ±0.04 ±0.01 ±0.01 ±0.08 49.8 ±6.09 ±0.01 ±0.04 50.9 ±0.32 6.1 0.09 19.2 34.4 0.86 8.4 GT6 ±0.04 ±0.03 ±0.01 87.8 ±1.3 6.8 ±0.05 4 ±0.02 ±0.25 ±0.2 34.3 ±0.91 ±0.06 0.08 50.1 ±0.40 5.7 0.07 17.9 3.9 33.5 0.6 8.2 GT7 ±0.07 ±0.06 ±0.1 86.3 ±0.4 6.6 ±0 ±0.02 ±0.22 ±0.02 33.9 ±0.62 ±0.02 0.06 48.9 ±0.12 6.1 0.12 20.6 122 3.9 35.3 0.69 8.2 GT8 ±0.11 ±0.02 ±0.9 ±33.1 6.5 ±0.04 ±0.02 ±0.66 ±0.13 36.2 ±13.19 ±0.11 0.08 49.9 ±0.20 6.1 0.07 19.7 3.9 33.3 0.59 8.6 0.06 GT9 ±0.01 ±0.01 ±0.2 87.7 ±0.7 6.6 ±0.02 ±0.01 ±0.01 ±0.03 24.5 ±2.89 ±0.01 ±0.01 50.5 ±0.01

Table 4.6. Potentially Toxic Trace Elements in Chia. Heavy metals were determined with two subsamples serving as replications. Units are all expressed as ppm, where BDL= Below Method Detection Limit for all replications. All samples were run for two reps. * indicates that only one rep exceeded the MDL (minimum detection limit) and other #s are SEs. All elements were assessed via ICP-MS, except Al, which was assessed via ICP- OES.

Sample B Al V Cr As Cd Pb E9 8.6 ±1.3 1.5 ±0.22 BDL 0.1 * BDL BDL 0.01 * G15 8.7 ±1.6 2 ±0.32 BDL 0.1 ±0.01 BDL 0.01 * BDL 0.97 G3 8.6 ±2.6 6.5 ±0.07 BDL BDL 0.06 * 0.01 * ±0.94 G6 8.4 ±2.8 25.1 ±1.6 0.07 * 0.04 * 0.04 ±0.02 BDL 0.06 * G8 6.5 ±2 4.1 ±0.23 BDL BDL BDL 0.01 * 0.01 * 0.04 GT1 9.6 ±2.1 10.9 ±1.2 0.01 * 0.15 * 0.03 ±0.02 0.02 ±0.01 ±0.01 0.02 GT2 7.7 ±2.7 8.9 ±0.15 BDL BDL 0.01 * 0.01 ±0.01 ±0.01 0.1 GT3 8.8 ±1.5 5.2 ±0.23 0.13 * 0.12 * 0.03 ±0.02 0.01 * ±0.09 GT4 7.6 ±2.9 3.6 ±0.55 BDL BDL 0.01 * BDL 0.01 * GT5 7.5 ±3.1 4.9 ±0.61 BDL BDL 0.05 * 0.02 * 0.02 * GT6 7.2 ±3.3 17.3 ±4.5 BDL BDL 0.05 * 0.02 * 0.05 * GT7 7.7 ±3.1 2.4 ±0.15 BDL BDL 0.06 * 0.01 * 0.02 * GT8 7.4 ±2.8 3.9 ±0.33 BDL BDL 0.06 * 0.01 * BDL GT9 7.7 ±3 2.6 ±0.82 BDL BDL 0.01 * BDL BDL 0.03 Pinta 9.1 ±1.6 4.3 ±0.05 BDL 0.09 * 0.01 * 0.01 * ±0.01

Figure 4.1. NIR profile of Chia Seeds. Near infrared profile of n=120 whole seed chia samples. Samples were run as whole seeds on Perten DA 7200 near infrared spectrometer. Units are reported as reflectance on the y-axis and as wavelength (nm) on the x-axis. Each line is the average of three runs per sample. Wavelengths from 950 nm to1200 nm are highly variable.

a. Whole Seed Oil

38 y = 0.8614x + 4.5312

36 R² = 0.8057

seed) oil/100g (g Predicted 22 20 20 25 30 35 40 Observed (g oil/100 g seed)

Ground Seed Oil b. 38 y = 0.8324x + 5.3415 36 R² = 0.7832 34 32 30 28 26

24 22 seed) oil/100g (g Predicted 20 20 25 30 35 40 Observed (g oil/100g seed)

Figure 4.2. Oil Prediction Models of (a) Whole and (b) Ground Seeds. A partial least squares regression was run for ground seed and whole seed chia. The regression equation was applied to a calibration file on the DA 7200 and the population was cross-verified against the known standards. The resulting correlations are of observed measurements versus NIRS predictions.

a. Whole Seed Protein

31 y = 0.8324x + 3.8172 29 R² = 0.7996 27

23 21 19

(g seed) Predicted prot/100g 15 15 17 19 21 23 25 27 29 Observed (g prot/100g seed)

b. Ground Seed Protein

31 29 y = 0.8233x + 4.0257 R² = 0.7936 27 25 23 21 19 17

(g seed) Predicted prot/100g 15 15 17 19 21 23 25 27 29 Observed (g prot/100 g seed)

Figure 4.3. Protein Prediction Models of (a) Whole and (b) Ground Seeds. A partial least squares regression was run for ground seed and whole seed chia. The regression equation was applied to a calibration file on the DA 7200 and the population was cross-verified against the known standards. The resulting correlations are of observed measurements versus NIRS predictions.

Whole Seed Moisture 11 y = 0.9009x + 0.7252 10 R² = 0.8779 9

Predicted (g/100g (g/100g Predicted seed) 6

5 5 6 7 8 9 10 11 Observed (g/100g seed)

Figure 4.4. Moisture Prediction Model for Whole Seeds. NMR calibration set for moisture was derived by drying whole seeds in oven for 48 hours. Drying of ground seeds resulted in oxidation of oil which affected calibration model so that model was excluded.

26.5 * * 26

25.5 25 24.5 24 23.5 % Protein (g/100g seed) 23 22.5 0 37 74 N Rate (kg N/ha)

56.5

55.5 Proein+Oil (g/100g seed)

55 0 37 74 N rate (kg N/ha)

Figure 4.5 Effect of Nitrogen Application on Protein and Oil Accumulation. Chia which was treated with 37 and 74 kg/ha had significantly higher protein levels than was seen in unfertilized chia, about 1 to 1.5 percent greater. This effect was tested on 10 mutagenized cultivars in one growing season. Stars indicate significantly higher protein levels than the control at α=0.05.

Chapter 5 CONCLUSIONS AND FUTURE DIRECTIONS

VgDGAT1A Increases Oil Content and Oleic Acid Content in Soybeans

Through our studies we have determined that VgDGAT1A aligns with several amino acids in a high activity modified soybean DGAT, including F499, F250 and F502.

VgDGAT1A shows a fivefold rate of substrate turnover in yeast microsomes, which results in five to ten times more oil accumulation in SF9 insect cell assays. In the metabolic pathway, the DGAT step is often a rate-limiting step so introducing a high- activity DGAT increases the rate of oil accumulation over the same time period, thus increasing the amount of oil overall (Hatanaka et al., 2016). In addition, the oleic acid content likely increases for the same reasons; increased rate of oil accumulation makes fatty acids unavailable metabolically to desaturases which would decrease the amount of linoleic and linolenic acid in the soybean oil.

While increasing oil content is a worthwhile goal on its own, understanding more about this DGAT could yield valuable information about specific amino acid sites and modifications, and those amino acid substitutions could be generated in other DGATs.

The major drawback to our approach is that it utilizes transgenic technology, resulting in regulatory hurdles that few universities and breeding programs can afford. However, modifying DGATs already present in an organism is not considered a transgenic approach and therefore would not be subject to the same regulatory scrutiny, but would have the same effects on oil accumulation. Other groups have established amino acids that are essential for activity such as F499 using amino acid substation studies(Zheng et al., 2008), while others have found changes that correlate with increase DGAT activity and oil accumulation by screening DNA shuffling libraries such as S58N, S264T and

R467K (Roesler et al., 2016). It would therefore be beneficial to study the effects of the unique amino acid differences in VgDGAT1A. Using site directed mutagenesis, it is possible to make changes to individual nucleotide bases, known as single nucleotide polymorphisms (SNPs). We could then perform yeast microsome and Sf9 lipid accumulation assays to determine what effects each modified amino acid has on enzyme performance.

A Deletion Event on Chromosome 14 May Increase Oil Content in Soybeans

By using comparative genomics hybridization to analyze copy number variation in segregating backcrossed mutant populations, we have discovered a 300 kb deletion event on chromosome 14 which corresponded with variation in seed protein and oil content which is indicative of segregation of a single gene, codominant trait. Analyzing this region reveals 20 different functional genes, including an enhancer, three transcription factors, two of unknown function and one involved in fatty acid beta oxidation. This gene, an enoyl CoA-hydratase, hydrolyzes breakdown products of beta oxidation to produce acetyl-CoA. Other beta oxidation knockouts in arabidopsis result in a slower rate of lipid breakdowns in germinating seedlings(Germain et al., 2001).

Therefore, this may serve as a protection mechanism that prevents oils that have already been produced from being hydrolyzed.

Much work needs to be done to verify that this mutation, and this gene is particular, is responsible for the increased oil accumulation. First, developing a marker will help to establish a correlation between the genotype for this deletion and corresponding phenotypes (Bolon et al., 2011). From there, a variety of tests can be carried out on the segregating and mutant soybeans. First, it would be interesting to study

90 the differences in lipid breakdown during germination between germinating seeds of wild type and mutant beans. It has been well established that knocking out beta oxidation enzymes in arabidopsis affects oil quantity in seeds, though different enzymes were affected in those studies(Khan et al., 2012). If this trait is also affecting lipolysis during germination similar to the arabidopsis beta oxidation knockouts, then we would expect to see the mutant soybeans maintain their oil content at a higher level during later stages of seed germination (Germain et al., 2001). Furthermore, the arabidopsis ortholog has mutant varieties available and it would be possible to test these for a similar pattern in seed oil content during germination. In arabidopsis, we could also demonstrate that this trait is responsible by attempting to rescue the wild type phenotype by adding the gene back into the arabidopsis genome by doing a floral dip transformation, and perform similar complementation studies in soybeans as well by using gene gun transformation techniques.

Development of Oil, Protein and Moisture NIRS Calibrations for Chia Seeds

A new growing region for chia has been established using new long-day flowering lines. These long-day flowering lines were grown in multiple years and locations in Lexington, KY (Jamboonsri et al., 2012). These seed lots and commercial seed lots from multiple locations in the world were analyzed for oil, protein, fatty acid, amino acids, minerals and moisture levels. The mean values and variance of the long-day flowering KY grown seed lots were similar to chia from other locations and literature values. The mean values for oil content ranged from 21.4 to 35.3%, protein ranged from

18.2% to 28.2%, and the ω-3 α-linolenic acid ranged from 33.9 to 66.4%. These values are within similar ranges to the non-mutagenized cultivars from native growing regions

(Ayerza, 2016). Using these seed lots, non-destructive near infrared spectroscopy (NIRS) calibrations were developed for whole and ground seed oil and protein contents and whole seed moisture content with accuracies of +/- 1% for protein and oil contents and

0.25% for moisture content. We were unable to develop accurate NIRS calibrations for fatty acid levels. As more phenotypic and genotypic variation is introduced in chia, we anticipate that the variability will be sufficient to establish calibrations for fatty acids as well.

Works Cited

A. B. Charette, J.-F. M., 1995. Synlett. 1197–2017. Acuña, R., 1986. Relaciones geográficas del siglo XVI. Serie antropológica. Universidad Nacional Autónoma de México. Instituto de Investigaciones Antropológicas, México City. Agelet, L. E., Armstrong, P. R., Romagosa Clariana, I., Hurburgh, C. R., 2012. Measurement of single soybean seed attributes by near-infrared technologies. A comparative study. J. Agric. Food Chem. 60, 1314-1322. Agnihotri, G., Liu, H.-w., 2003. Enoyl-CoA hydratase: Reaction, mechanism, and inhibition. Bioorganic & Medicinal Chemistry. 11, 9-20. AL-Amery, M., Fukushige, H., Hildebrand, D., 2015. Single Seed Selection for Low Phytate Lines. Journal of the American Oil Chemists' Society. 92, 1119-1123. Ali, N. M., Yeap, S. K., Ho, W. Y., Beh, B. K., Tan, S. W., Tan, S. G., 2012. The Promising Future of Chia, Salvia hispanica L. Journal of Biomedicine and Biotechnology. Alvarez-Chavez, L. M., Valdivia-Lopez, M. D., Aburto-Juarez, M. D., Tecante, A., 2008. Chemical characterization of the lipid fraction of Mexican chia seed (Salvia hispanica L.). International Journal of Food Properties. 11, 687-697. Álvarez, L. M., Chávez, M. A., Valdivia-López, M.L. Aburto-Juárez, A. T., 2008. Chemical Characterization of the Lipid Fraction of Mexican Chia Seed (Salvia hispanica L.). International Journal of Food Properties 11. Andre, C., Froehlich, J. E., Moll, M. R., Benning, C., 2007. A heteromeric plastidic pyruvate kinase complex involved in seed oil biosynthesis in Arabidopsis. Plant Cell. 19, 2006- 2022. Andrianov, V., Borisjuk, N., Pogrebnyak, N., Brinker, A., Dixon, J., Spitsin, S., Flynn, J., Matyszczuk, P., Andryszak, K., Laurelli, M., Golovkin, M., Koprowski, H., 2010. Tobacco as a production platform for biofuel: overexpression of Arabidopsis DGAT and LEC2 genes increases accumulation and shifts the composition of lipids in green biomass. Plant Biotechnol. J. 8, 277-87. Anon, 2008. Food for thought for treating type 2 diabetes mellitus. Inpharma Weekly. 18-18. Anon, 2013. AOCS Official Method Am 2-93. In: Firestone, D., (Ed.), Official Methods and Recommended Practices of the AOCS. vol. 6. American Oil Chemists Society. Anon, Basic Report: 12006, Seeds, chia seeds, dried National Nutrient Database for Standard Reference Release 28 United States Department of Agriculture Agricultural Research Service, 2016. Armstrong, P. R., 2006. Rapid single-kernel NIR measurement of grain and oil-seed attributes. Appl. Eng. Agric. 22, 767-772. Armstrong, P. R., Tallada, J. G., Hurburgh, C., Hildebrand, D. F., Specht, J. E., 2011. Development of Single-Seed Near-Infrared Spectroscopic Predictions of Corn and Soybean Constituents using Bulk Reference Values and Mean Spectra. T Asabe. 54, 1529-1535. Ayerza, R., 1995. Oil content and fatty acid composition of chia (Salvia hispanica L.) from five northwestern locations in Argentina. Journal of the American Oil Chemists' Society. 72, 1079-1081. Ayerza, R., 2009. The Seed's Protein and Oil Content, Fatty Acid Composition, and Growing Cycle Length of a Single Genotype of Chia (Salvia hispanica L.) as Affected by Environmental Factors. Journal of Oleo Science. 58, 347-354. Ayerza, R., 2010. Effects of Seed Color and Growing Locations on Fatty Acid Content and Composition of Two Chia (Salvia hispanica L.) Genotypes. Journal of the American Oil Chemists' Society. 87, 1161-1165.

Ayerza, R., 2016. Crop year effects on seed yields, growing cycle length, and chemical composition of chia (Salvia hispanica L) growing in Ecuador and Bolivia. Emirates Journal of Food and Agriculture. 28, 196-200. Ayerza, R., Coates, W., 1999. An omega-3 fatty acid enriched chia diet: Influence on egg fatty acid composition, cholesterol and oil content. Can J Anim Sci. 79, 53-58. Ayerza, R., Coates, W., 2004. Composition of chia (Salvia hispanica) grown in six tropical and subtropical ecosystems of South America. Tropical science. 44, 131-135. Ayerza, R., Coates, W., 2005. Ground chia seed and chia oil effects on plasma lipids and fatty acids in the rat. Nutr. Res. 25, 995-1003. Ayerza, R., Jr., Coates, W., 2007. Effect of dietary alpha -linolenic fatty acid derived from chia when fed as ground seed, whole seed and oil on lipid content and fatty acid composition of rat plasma. Annals of Nutrition and Metabolism. 51, 27-34. Baars, A. J., Re-evaluation of human-toxicological maximum permissible risk levels. In: Environment, N. I. o. P. H. a. t., (Ed.). RIVM, Bilohoven, The Netherlands, 2001. Banilas, G., Karampelias, M., Makariti, I., Kourti, A., Hatzopoulos, P., 2011. The olive DGAT2 gene is developmentally regulated and shares overlapping but distinct expression patterns with DGAT1. J. Exp. Bot. 62, 521-32. Baud, S., Boutin, J.-P., Miquel, M., Lepiniec, L., Rochat, C., 2002. An integrated overview of seed development in Arabidopsis thaliana ecotype WS. Plant Physiol. Biochem. 40, 151- 160. Baud, S., Lepiniec, L., 2009. Regulation of de novo fatty acid synthesis in maturing oilseeds of Arabidopsis. Plant Physiol. Biochem. 47, 448-455. Baud, S., Lepiniec, L., 2010. Physiological and developmental regulation of seed oil production. Progress in Lipid Research. 49, 235-249. Baud, S., Santos Mendoza, M., To, A., Harscoët, E., Lepiniec, L., Dubreucq, B., 2007a. WRINKLED1 specifies the regulatory action of LEAFY COTYLEDON2 towards fatty acid metabolism during seed maturation in Arabidopsis. Plant J. 50, 825-838. Baud, S., Wuillème, S., Dubreucq, B., de Almeida, A., Vuagnat, C., Lepiniec, L., Miquel, M., Rochat, C., 2007b. Function of plastidial pyruvate kinases in seeds of Arabidopsis thaliana. Plant J. 52, 405-419. Baud, S., Wuillème, S., To, A., Rochat, C., Lepiniec, L., 2009. Role of WRINKLED1 in the transcriptional regulation of glycolytic and fatty acid biosynthetic genes in Arabidopsis. Plant J. 60, 933-947. Bidlingmeyer, B. A., Cohen, S. A., Tarvin, T. L., 1984. Rapid Analysis of amino acids using pre- column derivitization. Journal of Chromatography. 336, 93-104. Bolon, Y.-T., Haun, W. J., Xu, W. W., Grant, D., Stacey, M. G., Nelson, R. T., Gerhardt, D. J., Jeddeloh, J. A., Stacey, G., Muehlbauer, G. J., Orf, J. H., Naeve, S. L., Stupar, R. M., Vance, C. P., 2011. Phenotypic and Genomic Analyses of a Fast Neutron Mutant Population Resource in Soybean. Plant Physiology. 156, 240-253. Borneo, R., Aguirre, A., León, A. E., 2010. Chia (Salvia hispanica L) Gel Can Be Used as Egg or Oil Replacer in Cake Formulations. Journal of the American Dietetic Association. 110, 946-949. Bozan, B., Temelli, F., 2008. Chemical composition and oxidative stability of flax, safflower and poppy seed and seed oils. Bioresour. Technol. 99, 6354-6359. Bradley, C., Hartman, G., Wax, L., Pedersen, W., 2002. Quality of harvested seed associated with soybean cultivars and herbicides under weed-free conditions. Plant disease. 86, 1036- 1042. Brigelius-Flohé, R., 2006. Bioactivity of vitamin E. Nutrition research reviews. 19, 174-186. Brigelius-Flohe, R., Traber, M. G., 1999. Vitamin E: function and metabolism. The FASEB Journal. 13, 1145-1155.

Browman, J. E., Hartman, G. L., McClary, R. D., Sinclair, J. B., Hummel, J. W., Wax, L. M., 1986. effect of weed control and row spacing in conventional tillage, reduce tillage and no tillage on soybean seed quality. Plant Dis. 70, 673-676. Bueno, M., Di Sapio, O., Barolo, M., Busilacchi, H., Quiroga, M., Severin, C., 2010. Quality tests of Salvia hispanica L. (Lamiaceae) fruits marketed in the city of Rosario (Santa Fe province, Argentina). B Latinoam Caribe Pl. 9, 221-227. Burton, J. W., 1987. Quantitative genetics results relevant to soybean breeding. In: Wilcox, J. R., (Ed.), Soybeans: Improvement, production and uses. vol. 16. ASA-CSSA-SSSA, pp. 211- 248. Cahill, J. P., 2003. Ethnobotany of Chia, Salvia hispanica L. (Lamiaceae). Econ Bot. 57, 604-618. Cahill, J. P., 2004. Genetic diversity among varieties of Chia (Salvia hispanica L.). Genetic Resources and Crop Evolution. 51, 773-781. Capitani, M., Spotorno, V., Nolasco, S., Tomás, M., 2012. Physicochemical and functional characterization of by-products from chia (Salvia hispanica L.) seeds of Argentina. LWT- Food Science and Technology. 45, 94-102. Capitani, M. I., Ixtaina, V. Y., Nolasco, S. M., Tomas, M. C., 2013. Microstructure, chemical composition and mucilage exudation of chia (Salvia hispanica L.) nutlets from Argentina. Journal of the Science of Food and Agriculture. 93, 3856-3862. Carrera, C., Martínez, M. J., Dardanelli, J., Balzarini, M., 2011. Environmental Variation and Correlation of Seed Components in Nontransgenic Soybeans: Protein, Oil, Unsaturated Fatty Acids, Tocopherols, and Isoflavones. Crop Science. 51, 800-809. Carter TE, Rzewnicki PE, Burton JW, Villagarcia MR, Bowman DT, Taliercio E, Kwanyuen P, 2010. Registration of N6202 soybean germplasm with high protein, favorable yield potential,large seed, and diverse pedigree. . J Plant Reg 4, 73–79. Cernac, A., Benning, C., 2004. WRINKLED1 encodes an AP2/EREB domain protein involved in the control of storage compound biosynthesis in Arabidopsis. Plant J. 40, 575-585. Chen, P. S., Toribara, T. Y., Warner, H., 1956. Microdetermination of Phosphorus. Anal Chem. 28, 1756-1758. Chia, T. Y. P., Pike, M. J., Rawsthorne, S., 2005. Storage oil breakdown during embryo development of Brassica napus (L.). J. Exp. Bot. 56, 1285-1296. Chicco, A. G., D'Alessandro, M. E., Hein, G. J., Oliva, M. E., Lombardo, Y. B., 2009. Dietary chia seed (Salvia hispanica L.) rich in alpha-linolenic acid improves adiposity and normalises hypertriacylglycerolaemia and insulin resistance in dyslipaemic rats. British Journal of Nutrition. 101, 41-50. Christie, W. W., 1970. Cyclopropane and cyclopropene fatty acids. In: Gunstone, F. D., (Ed.), Topics in Lipid Chemistry. vol. 1. John Wiley & Sons, New York, NY, pp. 1-49. Ciftci, O. N., Przybylski, R., Rudzińska, M., 2012. Lipid components of flax, perilla, and chia seeds. European Journal of Lipid Science and Technology. 114, 794-800. Cohen, S. A., Strydom, D. J., 1988. Amino Acid Analysis Utilizing Phenylisothiocyanate Derivatives. Analytical Biochemistry. 174, 1-16. Collins, G. B., Hildebrand, D. F., Lazzeri, P. A., Adams, T. R., Parrott, W. A., Hartweck, L. M., Transformation, somatic embryogenesis and whole plant regeneration method for Glycine species. University of Kentucky, USA, 1991. da Silva Marineli, R., Lenquiste, S. A., Moraes, É. A., Maróstica Jr, M. R., 2015. Antioxidant potential of dietary chia seed and oil (Salvia hispanica L.) in diet-induced obese rats. Food Research International. in press, (available online). da Silva, M. A. V., de Almeida D’Agosto, M., 2013. A model to estimate the origin–destination matrix for soybean exportation in Brazil. Journal of Transport Geography. 26, 97-107. Dahlqvist, A., Ståhl, U., Lenman, M., Banas, A., Lee, M., Sandager, L., Ronne, H., Stymne, S., 2000. Phospholipid:diacylglycerol acyltransferase: An enzyme that catalyzes the acyl-

CoA-independent formation of triacylglycerol in yeast and plants. Proc. Natl. Acad. Sci. USA. 97, 6487-6492. Domingo, J. L., Llobet, J. M., Tomas, J. M., 1985. Short Term Toxicity Studies of Vanadium in Rats. Journal of Applied Toxicology. 5, 418-421. Dowson, M. L., 1992. The Chromium Reference Dose. In: Mertz, W., Abernathy, C. O., Olin, S. S., Eds.), Risk Assessment of Essential Elements. ILSI Press, Washington, USA, pp. 207- 212. Eastmond, P. J., 2006. SUGAR-DEPENDENT1 encodes a patatin domain triacylglycerol lipase that initiates storage oil breakdown in germinating Arabidopsis seeds. Plant Cell. 18, 665- 675. Egli, D. B., 2008a. Comparison of corn and soybean yields in the United States: Historical trends and future prospects. Agron. J. 100, S79-S80. Egli, D. B., 2008b. Soybean yield trends from 1972 to 2003 in mid-western USA. Field Crops Res. 106, 53-59. Espada, C. E., Berra, M. A., Martinez, M. J., Eynard, A. R., Pasqualini, M. E., 2007. Effect of Chia oil (Salvia hispanica) rich in omega-3 fatty acids on the eicosanoid release, apoptosis and T-lymphocyte tumor infiltration in a murine mammary gland adenocarcinoma. Prostaglandins Leukotrienes and Essential Fatty Acids. 77, 21-28. Estilai, A., Hashemi, A., Truman, K., 1990. Chromosome number and meiotic behavior of cultivated chia, Salvia hispanica (Lamiaceae). HortScience. 25, 1646-1647. Fan, J., Yan, C., Roston, R., Shanklin, J., Xu, C., 2014. Arabidopsis Lipins, PDAT1 Acyltransferase, and SDP1 Triacylglycerol Lipase Synergistically Direct Fatty Acids toward β-Oxidation, Thereby Maintaining Membrane Lipid Homeostasis. The Plant Cell. 26, 4119-4134. Finer, J. J., Nagasawa, A., 1988. Development of an embryogenic suspension culture of soybean (Glycine max Merrill.). Plant Cell Tiss. Org. Cult. 15, 115-136. Focks, N., Benning, C., 1998. wrinkled1: A Novel, Low-Seed-Oil Mutant of Arabidopsis with a Deficiency in the Seed-Specific Regulation of Carbohydrate Metabolism. Plant Physiol. 118, 91-101. Freeman, M. P., Hibbeln, J. R., Wisner, K. L., Davis, J. M., Mischoulon, D., Peet, M., Keck, P. E., Marangell, L. B., Richardson, A. J., Lake, J., Stoll, A. L., 2006. Omega-3 fatty acids: Evidence basis for treatment and future research in psychiatry. . J. Clin. Psychiat. 67, 1954-1967. Germain, V., Rylott, E. L., Larson, T. R., Sherson, S. M., Bechtold, N., Carde, J.-P., Bryce, J. H., Graham, I. A., Smith, S. M., 2001. Requirement for 3-ketoacyl-CoA thiolase-2 in peroxisome development, fatty acid β-oxidation and breakdown of triacylglycerol in lipid bodies of Arabidopsis seedlings. The Plant Journal. 28, 1-12. Goldberg, R. B., Barker, S. J., Perez-Grau, L., 1989. Regulation of gene expression during plant embryogenesis. Cell. 56, 149-160. Golub, M. S., Germann, S. L., Han, B., 2000. Lifelong feeding of a high Al diet to mice. Toxicology 150, 107-117. Guevara-Cruz, M., Tovar, A. R., Aguilar-Salinas, C. A., Medina-Vera, I., Gil-Zenteno, L., Hernández-Viveros, I., López-Romero, P., Ordaz-Nava, G., Canizales-Quinteros, S., Guillen Pineda, L. E., Torres, N., 2012. A Dietary Pattern Including Nopal, Chia Seed, Soy Protein, and Oat Reduces Serum Triglycerides and Glucose Intolerance in Patients with Metabolic Syndrome. The Journal of Nutrition. 142, 64-69. Gutierrez, L., Van Wuytswinkel, O., Castelain, M., Bellini, C., 2007. Combined networks regulating seed maturation. Trends in Plant Science. 12, 294-300. Harwood, J. L., Page, R. A., 1994. Biochemistry of oil synthesis. In: Murphy, D. J., (Ed.), Designer Oil Crops. VCH, Weinheim, pp. 165-194.

Hatanaka, T., Serson, W., Li, R., Armstrong, P., Yu, K., Pfeiffer, T., Li, X.-L., Hildebrand, D., 2016. A Vernonia Diacylglycerol Acyltransferase Can Increase Renewable Oil Production. Journal of Agricultural and Food Chemistry. 64, 7188-7194. Hatanaka, T., Yu, K., Hildebrand, D. F., 2003. Cloning and expression of a Vernonia and Euphorbia diacylglycerol acyltransferase cDNAs. In: Murata, N., Yamada, M., Nishida, I., Okuyama, H., Sekiya, J., W, H., Eds.), Advanced research on plant lipids. Kluwer Academic Publishers, Dordrecht, The Netherlands, pp. 155-158. Hayati, R., Egli, D. B., Crafts-Brandner, S. J., 1996. Independence of nitrogen supply and seed growth in soybean: studies using an in vitro culture system. J. Exp. Bot. 47, 33-40. Hedge, I., 1970. Observation on the mucilage of Salvia fruits. Notes from the Royal Botanic Garden Edinburgh. 30, 79-95. Hernández-Sebastià, C., Marsolais, F., Saravitz, C., Israel, D., Dewey, R. E., Huber, S. C., 2005. Free amino acid profiles suggest a possible role for asparagine in the control of storage- product accumulation in developing seeds of low- and high-protein soybean lines. J. Exp. Bot. 56, 1951-1963. Hill JL, Peregrine EK, Sprau GL, Cremeens CR, Nelson RL, Orf JH, DA, T., 2005. Evaluation of the USDA soybean germ-plasm collection: maturity groups 000–IV (PI507670-PI 574486). US Department of Agriculture Technical Bulletin No. 1914, 1–131. Illian, T. G., Casey, J. C., Bishop, P. A., 2011. Omega 3 Chia Seed Loading as a Means of Carbohydrate Loading. Journal of Strength and Conditioning Research. 25, 61-65. Ixtaina, V. Y., Martinez, M. L., Spotorno, V., Mateo, C. M., Maestri, D. M., Diehl, B. W. K., Nolasco, S. M., Tomas, M. C., 2011. Characterization of chia seed oils obtained by pressing and solvent extraction. J Food Compos Anal. 24, 166-174. Ixtaina, V. Y., Nolasco, S. M., Tomas, M. C., 2008. Physical properties of chia (Salvia hispanica L.) seeds. Industrial Crops and Products. 28, 286-293. Jaco, C., Kumar, A., Wei, Y., Zou, J., Barton, D. L., Giblin, E. M., Covello, P. S., Taylor, D. C., 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. Jamboonsri, W., Phillips, T., Geneve, R., Cahill, J., Hildebrand, D., 2012. Extending the range of an ancient crop, Salvia hispanica L.—a new ω3 source. Genetic Resources and Crop Evolution. 59, 171-178. Jeong, S. K., Park, H. J., Park, B. D., Kim, I. H., 2010. Effectiveness of Topical Chia Seed Oil on Pruritus of End-stage Renal Disease (ESRD) Patients and Healthy Volunteers. Ann Dermatol. 22, 143-148. Jin, H.-H., Jiang, J.-G., 2015. Phosphatidic Acid Phosphatase and Diacylglycerol Acyltransferase: Potential Targets for Metabolic Engineering of Microorganism Oil. Journal of Agricultural and Food Chemistry. 63, 3067-3077. Kamal-Eldin, A., Appelqvist, L.-Å., 1996. The chemistry and antioxidant properties of tocopherols and tocotrienols. Lipids. 31, 671-701. Kane, M. V., Steele, C. C., Grabau, L. J., MacKown, C. T., Hildebrand, D. F., 1997. Early- maturing soybean cropping system. III. Protein and oil contents and oil composition. Agron J. 89, 464-469. Kelly, A. A., Shaw, E., Powers, S. J., Kurup, S., Eastmond, P. J., 2013. Suppression of the SUGAR-DEPENDENT1 triacylglycerol lipase family during seed development enhances oil yield in oilseed rape (Brassica napus L.). Plant Biotechnology Journal. 11, 355-361. Khan, B. R., Adham, A. R., Zolman, B. K., 2012. Peroxisomal Acyl-CoA oxidase 4 activity differs between Arabidopsis accessions. Plant Molecular Biology. 78, 45-58. Komprda, T., Zornikova, G., Rozikova, V., Borkovcova, M., Przywarova, A., 2013. The effect of dietary Salvia hispanica seed on the content of n-3 long-chain polyunsaturated fatty acids

in tissues of selected animal species, including edible insects. J Food Compos Anal. 32, 36-43. Kyriakidis, N. B., Katsiloulis, T., 2000. Calculation of iodine value from measurements of fatty acid methyl esters of some oils: Comparison with the relevant American Oil Chemists Society method. Journal of the American Oil Chemists' Society. 77, 1235-1238. Lardizabal, K., Effertz, R., Levering, C., Mai, J., Pedroso, M. C., 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. Latimer, G., 2012. AOAC Method 973.18: Fiber (Acid Detergent) and Lignin in Animal Feed. Official Methods of Analysis: 2012. vol. 19. AOAC. Law, J. H., 1971. Biosynthesis of cyclopropane rings. Accounts of Chemical Research. 4, 199- 203. Le, B. H., Wagmaister, J. A., Kawashima, T., Bui, A. Q., Harada, J. J., Goldberg, R. B., 2007a. Using genomics to study legume seed development. Plant Physiology. 144, 562-574. LeGoff, E., 1964. Cyclopropanes from an Easily Prepared, Highly Active Zinc—Copper Couple, Dibromomethane, and Olefins. The Journal of Organic Chemistry. 29, 2048-2050. Li, R., Hatanaka, T., Yu, K., Wu, Y., Fukushige, H., Hildebrand, D., 2013. Soybean oil biosynthesis: role of diacylglycerol acyltransferases. Functional and Integrated Genomics. 13, 99-113. Li, R., Yu, K., Hildebrand, D. F., 2010. DGAT1, DGAT2 and PDAT expression in seeds and other tissues of epoxy and hydroxy fatty acid accumulating plants. Lipids. 45, 145-57. Li, R., Yu, K., Wu, Y., Tateno, M., Hatanaka, T., Hildebrand, D. F., 2012. Vernonia DGATs can complement the disrupted oil and protein metabolism in epoxygenase-expressing soybean seeds. Metab. Eng. 14, 29-38. Lin, K. Y., Daniel, J. R., Whistler, R. L., 1994. Structure of chia seed polysaccharide exudate. Carbohydr. Polym. 23, 13-18. Liu, J., Hua, W., Zhan, G., Wei, F., Wang, X., Liu, G., Wang, H., 2010. Increasing seed mass and oil content in transgenic Arabidopsis by the overexpression of wri1-like gene from Brassica napus. Plant Physiol. Biochem. 48, 9-15. Liu, Q., Siloto, R. M. P., Lehner, R., Stone, S. J., Weselake, R. J., 2012. Acyl-CoA:diacylglycerol acyltransferase: Molecular biology, biochemistry and biotechnology. Prog. Lipid Res. 51, 350-377. Lonien, J., Schwender, J., 2009. Analysis of metabolic flux phenotypes for two Arabidopsis mutants with severe impairment in seed storage lipid synthesis. Plant Physiol. 151, 1617- 1634. Lu, Y., Foo, L. Y., 2002. Polyphenolics of Salvia--a review. Phytochemistry. 59, 117-40. Luttgeharm, K. D., Chen, M., Mehra, A., Cahoon, R. E., Markham, J. E., Cahoon, E. B., 2015. Overexpression of Arabidopsis Ceramide Synthases Differentially Affects Growth, Sphingolipid Metabolism, Programmed Cell Death, and Mycotoxin Resistance. Plant Physiology. 169, 1108-1117. Maeo, K., Tokuda, T., Ayame, A., Mitsui, N., Kawai, T., Tsukagoshi, H., Ishiguro, S., Nakamura, K., 2009. An AP2-type transcription factor, WRINKLED1, of Arabidopsis thaliana binds to the AW-box sequence conserved among proximal upstream regions of genes involved in fatty acid synthesis. Plant J. 60, 476-487. Marineli, R. d. S., Moraes, É. A., Lenquiste, S. A., Godoy, A. T., Eberlin, M. N., Maróstica Jr, M. R., 2014. Chemical characterization and antioxidant potential of Chilean chia seeds and oil (Salvia hispanica L.). LWT - Food Science and Technology. 59, 1304-1310. Medic, J., Atkinson, C., Hurburgh, C., Jr., 2014. Current Knowledge in Soybean Composition. Journal of the American Oil Chemists' Society. 91, 363-384.

Monroy-Torres, R., Mancilla-Escobar, M. L., Gallaga-Solorzano, J. C., Medina-Godoy, S., Santiago-Garcia, E. J., 2008. Protein digestibility of chia seed Salvia hispanica L. Revista Salud Publica y Nutricion. 9, unpaginated. Murphy, D. J., Cummins, I., 1989. Biosynthesis of seed storage products during embryogenesis in rapeseed, Brassica napus. J. Plant Physiol. 135, 63-69. Nieman, D., Gillitt, N., Meaney, M., Dew, D., 2015. No Positive Influence of Ingesting Chia Seed Oil on Human Running Performance. Nutrients. 7, 3666. Nieman, D. C., Cayea, E. J., Austin, M. D., Henson, D. A., McAnulty, S. R., Jin, F. X., 2009. Chia seed does not promote weight loss or alter disease risk factors in overweight adults. Nutr. Res. 29, 414-418. Nieman, D. C., Gillitt, N., Jin, F., Henson, D. A., Kennerly, K., Shanely, R. A., Ore, B., Su, M., Schwartz, S., 2012. Chia Seed Supplementation and Disease Risk Factors in Overweight Women: A Metabolomics Investigation. Journal of Alternative and Complementary Medicine. 18, 700-708. Nijveldt, R. J., Van Nood, E., Van Hoorn, D. E., Boelens, P. G., Van Norren, K., Van Leeuwen, P. A., 2001. Flavonoids: a review of probable mechanisms of action and potential applications. The American journal of clinical nutrition. 74, 418-425. Oakes, J., Brackenridge, D., Colletti, R., Daley, M., Hawkins, D. J., Xiong, H., Mai, J., Screen, S. E., Val, D., Lardizabal, K., Gruys, K., Deikman, J., 2011. Expression of fungal diacylglycerol acyltransferase2 genes to increase kernel oil in maize. Plant Physiol. 155, 1146-1157. Oliva, M. E., Ferreira, M. R., Chicco, A., Lombardo, Y. B., 2013. Dietary Salba (Salvia hispanica L) seed rich in alpha-linolenic acid improves adipose tissue dysfunction and the altered skeletal muscle glucose and lipid metabolism in dyslipidemic insulin-resistant rats. Prostaglandins Leukotrienes and Essential Fatty Acids. 89, 279-289. Olivos-Lugo, B. L., Valdivia-Lopez, M. A., Tecante, A., 2010. Thermal and Physicochemical Properties and Nutritional Value of the Protein Fraction of Mexican Chia Seed (Salvia hispanica L.). Food Sci Technol Int. 16, 89-96. Pandurangan, S., Pajak, A., Molnar, S. J., Cober, E. R., Dhaubhadel, S., Hernández-Sebastià, C., Kaiser, W. M., Nelson, R. L., Huber, S. C., Marsolais, F., 2012. Relationship between asparagine metabolism and protein concentration in soybean seed. J. Exp. Bot. 63, 3173- 3184. Peregrine, E., Sprau, G., Cremeens, C., Handly, P., Kilen, T., Smith, J., Thomas, D., Sarins, J., RL, N. , 2008. Evaluation of the USDA soybean germplasm collection: maturity group V (FC 30265-PI 612614) and maturity groups VI–VIII (PI 416758-PI606432B). . US Department of Agriculture Technical Bulletin No. 1920. 1-367. Piper, E. L., Boote, K. L., 1999. Temperature and cultivar effects on soybean seed oil and protein concentrations. J. Am. Oil Chem. Soc. 76, 1233-1241. Pipolo, A. E., Sinclair, T. R., Camara, G. M. S., 2004. Protein and oil concentration of soybean seed cultured in vitro using nutrient solutions of differing glutamine concentration. Ann. Appl. Biol. 144, 223-227. Polacheck, J. W., Tropp, B. E., Law, J. H., McCloskey, J. A., 1966. Biosynthesis of cyclopropane compounds. VIII. The conversion of oleate to dihydrosterculate. Journal of Biological Chemistry. 241, 3362-3364. Poppenberger, B., Rozhon, W., Khan, M., Husar, S., Adam, G., Luschnig, C., Fujioka, S., Sieberer, T., 2011. CESTA, a positive regulator of brassinosteroid biosynthesis. The EMBO Journal. 30, 1149-1161. Prasad, K., 2009. Flaxseed and Cardiovascular Health. J. Cardiovasc. Pharmacol. 54, 369-377. Price, C. J., Strong, P. L., Marr, M. C., CH Myers, C. H., Murray, F. J., 1996. Developmental Toxicity and postnatal recover in rats fed boric acid during gestation. Fundamental Applied Toxicology. 32, 179-193.

Rao, S., Abdel-Reheem, M., Bhella, R., McCracken, C., Hildebrand, D., 2008. Characteristics of High α-Linolenic Acid Accumulation in Seed Oils. Lipids. 43, 749-755. Rao, S. S., Hildebrand, D., 2009. Changes in oil content of transgenic soybeans expressing the yeast SLC1 gene. Lipids. 44, 945-51. Rennie, B. D., Tanner, J. W., 1989. Fatty acid composition of oil from soybean seeds grown at extreme temperatures. Journal of the American Oil Chemists' Society. 86, 1622-1624. Repic, O., Vogt, S., 1982. Ultrasound in organic synthesis: cyclopropanation of olefins with zinc- diiodomethane. Tetrahedron Letters. 23, 2729-2732. Reyes-Caudillo, E., Tecante, A., Valdivia-López, M., 2008. Dietary fibre content and antioxidant activity of phenolic compounds present in Mexican chia (Salvia hispanica L.) seeds. Food Chemistry. 107, 656-663. Roesler, K., Shen, B., Bermudez, E., Li, C., Hunt, J., Damude, H. G., Ripp, K. G., Everard, J. D., Booth, J. R., Castaneda, L., Feng, L., Meyer, K., 2016. An Improved Variant of Soybean Type 1 Diacylglycerol Acyltransferase Increases the Oil Content and Decreases the Soluble Carbohydrate Content of Soybeans. Plant Physiol. Rolletschek, H., Borisjuk, L., Radchuk, R., Miranda, M., Heim, U., Wobus, U., Weber, H., 2004. Seed-specific expression of a bacterial phosphoenolpyruvate carboxylase in Vicia narbonensis increases protein content and improves carbon economy. Plant Biotech. J. 2, 211-219. Rossi, A. S., Oliva, M. E., Ferreira, M. R., Chicco, A., Lombardo, Y. B., 2013. Dietary chia seed induced changes in hepatic transcription factors and their target lipogenic and oxidative enzyme activities in dyslipidaemic insulin-resistant rats. British Journal of Nutrition. 109, 1617-1627. Rotundo, J. L., Borrás, L., Westgate, M. E., 2011. Linking assimilate supply and seed developmental processes that determine soybean seed composition. European Journal of Agronomy. 35, 184-191. Saldivar, X., Wang, Y.-J., Chen, P., Hou, A., 2011. Changes in chemical composition during soybean seed development. Food Chemistry. 124, 1369-1375. Samoylov, V. M., Tucker, D. M., Thibaud-Nissen, F., Parrott, W. A., 1998. A liquid-medium- based protocol for rapid regeneration from embryogenic soybean cultures. Plant Cell Rep. 18, 49-54. Sandoval-Oliveros, M. R., Paredes-López, O., 2013. Isolation and Characterization of Proteins from Chia Seeds (Salvia hispanica L.). Journal of Agricultural and Food Chemistry. 61, 193-201. Santos-Mendoza, M., Dubreucq, B., Baud, S., Parcy, F., Caboche, M., Lepiniec, L., 2008. Deciphering gene regulatory networks that control seed development and maturation in Arabidopsis. Plant J. 54, 608-620. Sargi, S. C., Silva, B. C., Santos, H. M. C., Montanher, P. F., Boeing, J. S., Junior, O. O. S., Souza, N. E., Visentainer, J. V., 2013. Antioxidant capacity and chemical composition in seeds rich in omega-3: chia, flax, and perilla. Ciencia e Tecnologia de Alimentos. 33, 541-548. Schluttenhofer, C., Pattanaik, S., Patra, B., Yuan, L., 2014. Analyses of Catharanthus roseus and Arabidopsis thaliana WRKY transcription factors reveal involvement in jasmonate signaling. BMC Genomics. 15, 502. Schluttenhofer, C., Yuan, L., 2015. Regulation of Specialized Metabolism by WRKY Transcription Factors. Plant Physiology. 167, 295-306. Schmidt, M. A., Dietrich, C. R., Cahoon, E. B., 2005a. Biotechnological enhancement of soybean oil for lubricant applications. In: Rudnick, L. R., (Ed.), Synthetics, Mineral Oils, and Bio- Based Lubricants: Chemistry and Technology. Taylor and Francis CRC Press, pp. 389- 397.

100

Schmidt, M. A., Tucker, D. M., Cahoon, E. B., Parrott, W. A., 2005b. Towards normalization of soybean somatic embryo maturation. Plant Cell Rep. 24, 383-391. Schwender, J., Hay, J. O., 2012. Predictive modeling of biomass component tradeoffs in Brassica napus developing oilseeds based on in silico manipulation of storage metabolism. Plant Physiol. 160, 1218-1236. Shen, B., Allen, W. B., Zheng, P., Li, C., Glassman, K., Ranch, J., Nubel, D., Tarczynski, M. C., 2010. Expression of ZmLEC1 and ZmWRI1 increases seed oil production in maize. Plant Physiol. 153, 980-987. Sidorov, R. A., Tsydendambaev, V. D., 2014. Biosynthesis of fatty oils in higher plants. Russian Journal of Plant Physiology. 61, 1-18. Silva, B. e., dos Santos, H., Montanher, P., Boeing, J., Almeida, V. d., Visentainer, J., 2014. Incorporation of Omega-3 Fatty Acids in Nile Tilapia (Oreochromis niloticus) Fed Chia (Salvia hispanica L.) Bran. Journal of the American Oil Chemists' Society. 91, 429-437. Simopoulos AP, D. M. F., 2009. A Balanced Omega-6/ Omega-3 Fatty Acid Ratio, Cholesterol and Coronary Heart Disease. . World Rev Nutr Diet. 100, 110–121. Slightom, J. L., Sun, S. M., C., H. T., 1983. Complete nucleotide sequence of a French bean storage protein gene: Phaseolin. Proc. Natl. Acad. Sci. USA. 80, 1897-1901. Smith, A. J., Rinne, R. W., Seif, R. D., 1989. Phosphoenolpyruvate carboxylase and pyruvate kinase involvement in protein and oil biosynthesis during soybean seed development. Crop Sci. 29, 349-353. Sonntag, N. O. V., 1979. Composition and characteristics of individual fats and oils. In: Swern, D., (Ed.), Bailey's Industrial Oil and Fat Products. vol. 1. John Wiley & Sons, New York, pp. 289–477. Stevenson, D. G., Doorenbos, R. K., Jane, J.-l., Inglett, G. E., 2006. Structures and Functional Properties of Starch From Seeds of Three Soybean (Glycine max (L.) Merr.) Varieties*. Starch - Stärke. 58, 509-519. Stones, W., Campbell, S., Krishnan, K., 2012. the role of vitamin E in prostate cancer.Oxidative stress in cancer biology and therapy. Ocidative stress in applied Basic Research and clinical Practice , D.R.Spitz, K.J.Dornfeld,K.Krishnan and D.Gius, eds.Humana Press, pp.333-354. Sugimoto, T., Tanaka, K., Monma, M., Kawamura, Y., Saio, K., 1989. Phosphoenolpyruvate carboxylase level in soybean seed highly correlates to its contents of protein and lipid. Agric. Biol. Chem. 53, 885-887. Suri, S., Passi, S. J., Goyat, J., Chia Seed-A New Age Functional Food. 4th International Conference on Recent Innovations in Science Engineering and Management, New Delhi, India, 2016. Taylor, C. G., Noto, A. D., Stringer, D. M., Froese, S., Malcolmson, L., 2010. Dietary Milled Flaxseed and Flaxseed Oil Improve N-3 Fatty Acid Status and Do Not Affect Glycemic Control in Individuals with Well-Controlled Type 2 Diabetes. Journal of the American College of Nutrition. 29, 72-80. Taylor, D. C., Zhang, Y., Kumar, A., Francis, T., Giblin, E. M., Barton, D. L., Ferrie, J. R., Laroche, A., Shah, S., Zhu, W., Snyder, C. L., Hall, L., Rakow, G., Harwood, J. L., Weselake, R. J., 2009. Molecular modification of triacylglycerol accumulation by overexpression of DGAT1 to produce canola with increased seed oil content under field conditions. Botany. 87, 533-543. To, A., Joubès, J., Barthole, G., Lécureuil, A., Scagnelli, A., Jasinski, S., Lepiniec, L., Baud, S., 2012. WRINKLED transcription factors orchestrate tissue-specific regulation of fatty acid biosynthesis in Arabidopsis. Plant Cell. 24, 5007-5023. Trick, H. N., Dinkins, R. D., Santarem, E. R., Di, R., Samoylov, V., Meurer, C. A., Walker, D. R., Parrott, W. A., Finer, J. J., Collins, G. B., 1997. Recent adcances in soybean transformation. Plant Tissue Cultre and Biotechnology. 3, 9-26.

101

Tupy, M., Shafer, A., Industrial products solutions from renewable oils. 97th AOCS Annual Meeting & Expo, St. Louis, MO, 2006, pp. 66. Ulbricht, C., Chao, W., Nummy, K., Rusie, E., Tanguay-Colucci, S., Iannuzzi, C., Plammoottil, J., Varghese, M., Weissner, W., 2009 Chia (Salvia hispanica): a systematic review by the natural standard research collaboration. Reviews on Recent Clinical Trials. 4, 168-174. Ullah, R., Nadeem, M., Khalique, A., Imran, M., Mehmood, S., Javid, A., Hussain, J., 2015. Nutritional and therapeutic perspectives of Chia (Salvia hispanica L.): a review. Journal of Food Science and Technology. 1-9. van Erp, H., Kelly, A. A., Menard, G., Eastmond, P. J., 2014. Multigene engineering of triacylglycerol metabolism boosts seed oil content in Arabidopsis. Plant Physiol. 165, 30- 6. Vanhercke, T., El Tahchy, A., Liu, Q., Zhou, X.-R., Shrestha, P., Divi, U. K., Ral, J.-P., Mansour, M. P., Nichols, P. D., James, C. N., Horn, P. J., Chapman, K. D., Beaudoin, F., Ruiz- López, N., Larkin, P. J., de Feyter, R. C., Singh, S. P., Petrie, J. R., 2014. Metabolic engineering of biomass for high energy density: oilseed-like triacylglycerol yields from plant leaves. Plant Biotechnology Journal. 12, 231-239. Vanhercke, T., El Tahchy, A., Shrestha, P., Zhou, X.-R., Singh, S. P., Petrie, J. R., 2013. Synergistic effect of WRI1 and DGAT1 coexpression on triacylglycerol biosynthesis in plants. FEBS Lett. 587, 364-369. Vázquez-Ovando, A., Rosado-Rubio, G., Chel-Guerrero, L., Betancur-Ancona, D., 2009. Physicochemical properties of a fibrous fraction from chia (Salvia hispanica L.). LWT - Food Science and Technology. 42, 168-173. Vazquez-Ovando, J. A., Rosado-Rubio, J. G., Chel-Guerrero, L. A., Betancur-Ancona, D. A., 2010. Dry processing of chia (Salvia hispanica L.) flour: chemical characterization of fiber and protein. CyTA-J. Food. 8, 117-127. Venkatachalam, M., Sathe, S. K., 2006. Chemical composition of selected edible nut seeds. J. Agric. Food Chem. 54, 4705-4714. Verdier, J., Thompson, R. D., 2008. Transcriptional regulation of storage protein synthesis during dicotyledon seed filling. Plant Cell Physiol. 49, 1263-1271. Vickery, J. R., 1980. The fatty acid composition of seed oils from ten plant families with particluar reference to cyclopropene and dihydrosterculic acids. Journal of American Oil Chemists' Society. 57, 87-91. Vigeolas, H., Waldeck, P., Zank, T., Geigenberger, P., 2007. Increasing seed oil content in oilseed rape (Brassica napus L.) by over-expression of a yeast glycerol-3-phosphate dehydrogenase under the control of a seed-specific promoter. Plant Biotechnology Journal. 5, 431-441. Vollmann, J., Fritz, C. N., Wagentristl, H., Ruckenbauer, P., 2000. Environmental and genetic variation of soybean seed protein content under Central European growing conditions. Journal of the Science of Food and Agriculture. 80, 1300-1306. Vuksan, V., Jenkins, A. L., Dias, A. G., Lee, A. S., Jovanovski, E., Rogovik, A. L., Hanna, A., 2010. Reduction in postprandial glucose excursion and prolongation of satiety: possible explanation of the long-term effects of whole grain Salba (Salvia hispanica L.). Eur J Clin Nutr. 64, 436-438. Vuksan, V., Jovanovski, E., Dias, A., Lee, A., Rogovik, A., Jenkins, A., 2009. Comparable Dose- Response Glucose-Lowering Effect with Whole Versus Finely Ground Novel Omega-3 Rich Grain Salba (Salvia hispanica) Baked into White Bread. Pharmaceutical Biology. 47, S13. Vuksan, V., Whitham, D., Sievenpiper, J.-L., Jenkins, A.-L., Rogovik, A.-L., Bazinet, R.-P., Vidgen, E., Hanna, A., 2007. Supplementation of Conventional Therapy With the Novel Grain Salba (Salvia hispanica L.) Improves Major and Emerging Cardiovascular Risk

102

Factors in Type 2 Diabetes: Results of a randomized controlled trial. Diabetes Care. 30, 2804-2810. Wagner, K.-H., Kamal-Eldin, A., Elmadfa, I., 2004. Gamma-tocopherol–an underestimated vitamin? Annals of nutrition and metabolism. 48, 169-188. 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 Tiss. Org. Cult. 119, 399-410. Weselake, R. J., Shah, S., Tang, M., Quant, P. A., Snyder, C. L., Furukawa-Stoffer, T. L., Zhu, W., Taylor, D. C., Zou, J., Kumar, A., Hall, L., Laroche, A., Rakow, G., Raney, P., Moloney, M. M., Harwood, J. L., 2008a. Metabolic control analysis is helpful for informed genetic manipulation of oilseed rape (Brassica napus) to increase seed oil content. J. Exp. Bot. 59, 3543-9. Wilcox, J. R., Shibles, R. M., 2001. Interrelationships among seed quality attributes in soybean. Crop Sci. 41, 11-14. Wilson, R. F., Hildebrand, D., 2010. Engineering Status, Challenges and Advantages of Oil crops. In: Mascia, P., Scheffran, J., Thomas, S., Widholm, J. M., Eds.), Plant Biotechnology for Sustainable Production of Energy and Coproducts. Springer. Xu, J., Francis, T., Mietkiewska, E., Giblin, E. M., Barton, D. L., Zhang, Y., Zhang, M., Taylor, D. C., 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. Yu, K., Li, R., Hatanaka, T., Hildebrand, D., 2008. Cloning and functional analysis of two type 1 diacylglycerol acyltransferases from Vernonia galamensis. Phytochemistry. 69, 1119- 1127. Yu, K. S., McCracken, C. T., Li, R. Z., Hildebrand, D. F., 2006. Diacylglycerol acyltransferases from Vernonia and Stokesia prefer substrates with vernolic acid. Lipids. 41, 557-566. Zhang, M., Fan, J., Taylor, D. C., Ohlrogge, J. B., 2009. DGAT1 and PDAT1 Acyltransferases Have Overlapping Functions in Arabidopsis Triacylglycerol Biosynthesis and Are Essential for Normal Pollen and Seed Development. Plant Cell. 21, 3885-3901. Zhao, H., Wang, S., Chen, S., Jiang, J., Liu, G., 2015. Phylogenetic and stress-responsive expression analysis of 20 WRKY genes in Populus simonii × Populus nigra. Gene. 565, 130-139. Zheng, P., Allen, W. B., Roesler, K., Williams, M. E., Zhang, S., Li, J., Glassman, K., Ranch, J., Nubel, D., Solawetz, W., Bhattramakki, D., Llaca, V., Deschamps, S., Zhong, G. Y., Tarczynski, M. C., Shen, B., 2008. A phenylalanine in DGAT is a key determinant of oil content and composition in maize. Nat. Genet. 40, 367-72. Zingg, J.-M., Azzi, A., 2004. Non-antioxidant activities of vitamin E. Current medicinal chemistry. 11, 1113-1133.

103

Vita

1. PLACE OF BIRTH: Summit, New Jersey

2. EDUCATION:

Millersville University - Bachelor of Science (2008 – 2012)

3. RESEARCH EXPERIENCE

Graduate Research Assistant University of Kentucky (2012 – Present) Undergraduate Research Assistant Millersville University (2010-2012)

PA Department of Agriculture Forest Insect Pest Aide (Summer 2011) PA Department of Agriculture Plum Pox Virus Surveyor (Summer 2010)

4. PUBLICATIONS

Orlowski, John, Gary Gregg, William Serson, Maythem AL-Amery and Chad Lee. “Early- season stress can have small effect on soybean seed protein and oil content”. Crop, Forage, & Turfgrass Management. (accepted).

Orlowski, J. M., G. L. Gregg, C. D. Lee, and W. R. Serson. 2016. “Early-Season Lactofen Application Fails to Increase Soybean Yield under Weed-Free Conditions”. Agronomy Journal, 108:1552-1560.

AL-Amery, Maythem, Hirotada Fukushige, William Serson, and David Hildebrand. "Nondestructive DNA Extraction Techniques for Soybean (Glycine max) Seeds." Journal of Crop Improvement, 30:2, 165-175.

Hatanaka, Tomoko, William Serson, Runzhi Li, Paul Armstrong, Keshun Yu, Todd Pfieffer, Xi- Le Li, and David Hildebrand. "A Vernonia Diacylglycerol Acyltransferase Can Increase Renewable Oil Production." Journal of Agricultural and Food Chemistry, 64:38, 7188- 7194.

Serson, William, Maythem AL-Amery, Shrea Patel, Timothy Phillips, and David Hildebrand. "Emerging Industrial Oil Crops: Chia." Industrial Oil Crops. Elsevier Inc., 2016. 278- 287.

104