Novel Alleles of FAD2-1A Induce High Levels of Oleic Acid in Soybean Oil
Novel Alleles of FAD2-1A Induce High Levels of Oleic Acid in Soybean Oil
Novel Alleles of FAD2-1A Induce High Levels of Oleic Acid in Soybean Oil
https://doi.org/10.1007/s11032-019-0972-9
Abstract Soybean plays an important role in seed oil alleles; however, the increase in stearic acid was associ-
production for foods and industrial products in the USA. ated with a decrease in the oleic acid content and did not
Chemical hydrogenation of commodity soybean oil in- meet the target of at least 20% stearic acid in the seed oil.
creased functionality but unavoidably created trans fat In addition, existing FAD3 mutant alleles were incorpo-
which has been linked to many health issues in humans. rated into the novel high oleic acid and high oleic/
An alternative to using hydrogenation of the oil to elevated stearic acid soybean lines with four or five
enhance oxidative stability is to genetically increase targeted alleles combined (FAD2-1A, FAD2-1B, FAD3A,
the level of oleic acid in the seed oil. Our goal was to FAD3C, and SACPD-C) to create soybean germplasm
develop a soybean germplasm as a source for more with more functional soybean oil. This study is benefi-
functional soybean oil with a high oleic acid, increased cial for improving the quality of soybean oil based on
stearic acid, and low linolenic acid profile utilizing new nutritional value and oxidative stability.
and existing variant alleles of key fatty acid desaturase
genes. Using forward genetics, novel alleles of FAD2- Keywords Soybean . Oil . Fatty acids . Oleic acid .
1A were discovered and characterized from a mutant Stearic acid . Linolenic acid
soybean population already containing elevated seed
stearic acid content. One of the two new FAD2-1A
alleles, when combined with existing mutant alleles of
Introduction
FAD2-1B created a high oleic acid seed oil phenotype in
soybean germplasm lines. The other new FAD2-1A al-
Soybean [Glycine max (L.) Merr.] is a primary source
lele produced lower and more variable oleic acid levels
for seed oil production in the USA. Because of its
when combined with existing mutant alleles of FAD2-
function and widespread use in foods and industrial
1B. Seed stearic acid increased to ~ 10–11% in lines
products, it is important to genetically improve the fatty
containing combinations of FAD2-1A and FAD2-1B
acid profile of soybean oil to provide an alternative to
mutant alleles plus the SACPD-C missense mutant
chemical hydrogenation. This hydrogenation process
unavoidably creates trans fat as a by-product which
has been linked to health problems in humans
R. Combs
Division of Biological Sciences, University of Missouri, 105
(Crupkin and Zambelli 2008). Trans fats have been
Tucker Hall, Columbia, MO 65211-7400, USA mandated on food nutrition facts labels for over a de-
cade, and partially hydrogenated oils recently were
K. Bilyeu (*)
Plant Genetics Research Unit, USDA-ARS, University of
banned from foods in the USA (FDA 2003; FDA 2015).
Missouri, 110 Waters Hall, Columbia, MO 65211, USA Commodity soybean oil typically has a fatty acid pro-
e-mail: [email protected] file of 13% palmitic acid, 4% stearic acid, 20% oleic acid,
79 Page 2 of 11 Mol Breeding (2019) 39:79
55% linoleic acid, and 8% linolenic acid (Fehr 2007; acid in the oil will increase the functionality of oil
Pham et al. 2010). Soybean oil with a high oleic acid without causing health issues, since stearic acid is
content over 70% and linolenic acid content less than 3% thought to be heart neutral (Crupkin and Zambelli
is desirable for its oxidative stability and health benefits. 2008). However, stearic acid typically only makes up
The delta-twelve fatty acid desaturase-2 (FAD2) enzyme about 3–4% of soybean oil. A single-nucleotide poly-
catalyzes the conversion of precursors of oleic acid (18:1) morphism (SNP) in the stearoyl-acyl carrier protein
into linoleic acid (18:2) (Pham et al. 2010). Previous desaturase C gene (SACPD-C; Glyma.14g121400) was
studies have shown that the two microsomal FAD2-1 identified in a line (194D) derived from an ethyl
desaturases, FAD2-1A (Glyma.10g278000) and FAD2- methanesulfonate (EMS)-induced Williams 82 mutant
1B (Glyma.20g111000), play an important role in oil population used for TILLING (Cooper et al. 2008;
synthesis and the control of oleic acid level in developing Gillman et al. 2014). The SNP in SACPD-C, which
seeds (Fehr 2007; Pham et al. 2010). When mutant alleles results in a missense amino acid substitution of an
of these genes are combined, they produce higher levels of almost invariant residue (V211E), showed elevated
oleic acid in the seed oil due to the reduction of activity in levels of stearic acid in the seed oil (Gillman et al. 2014).
the fatty acid desaturase enzyme, which leads to increased Soybean germplasm containing new combinations of
accumulation of oleic acid (Haun et al. 2014; Hoshino mutant fatty acid desaturase genes may provide sources
et al. 2010; Pham et al. 2011; Pham et al. 2010; Pham et al. to develop soybean varieties that can improve the func-
2012; Sweeney et al. 2017). Downregulation of FAD2-1A tionality of soybean oil in the modern marketplace. We
and FAD2-1B is also the basis for the commercialized identified and characterized novel mutant alleles of
Plenish® and Vistive® Gold soybean varieties with the FAD2-1A alone and in combination with an existing
high oleic acid oil trait, while the Calyxt high oleic soy- mutant FAD2-1B gene and with mutant FAD3A and
beans were developed from genome editing mutation of FAD3C genes that produced high oleic (~ 80%)/low
FAD2-1A and FAD2-1B. linolenic (~ 3%) soybean germplasm (HOLL). Another
In addition to a higher content of oleic acid in soy- objective of this work was to develop soybean germ-
bean oil, lower concentrations of the polyunsaturated plasm that would reach homozygosity for the five mu-
fatty acids, linoleic acid (18:2) and linolenic acid (18:3), tant alleles targeted: FAD2-1A, FAD2-1B, FAD3A,
are desired because they have extremely low oxidative FAD3C, and SACPD-C to attempt to create a high
stability. Studies have revealed that mutations in the stearic acid (> 20%) phenotype. Though the novel com-
soybean microsomal omega-3 fatty acid desaturase bination of mutant alleles produced unique fatty acid
(FAD3) encoding genes contribute to a reduction in profiles of the seed oil, the target oil phenotype was not
linolenic acid levels (Bilyeu et al. 2005). Omega-3 fatty achieved.
acid desaturase enzymes catalyze the conversion of
linoleic acid precursors into linolenic acid precursors
by the incorporation of a third double bond in the carbon Materials and methods
chain. There are three FAD3 genes in soybean, and
studies have shown that FAD3A (Glyma.14 g194300) Mutant population in 194D (Williams 82) background
in combination with FAD3B (Glyma.02g227200) or
FAD3C (Glyma.18g062000) reduced linolenic acid Mutant lines were developed from an EMS treatment of
levels in soybean oil to approximately 3% (Bilyeu 20 pounds of elevated stearic acid soybean line 194D
et al. 2011; Bilyeu et al. 2005). Less than 3% of linolenic which itself was selected from an EMS mutagenesis of
acid in the oil is the current market target to increase oil BWilliams 82^ based on elevated seed stearic acid con-
functionality. tent (Gillman et al. 2014). Seeds were exposed to
With a market demand for healthier foods that still 50 mM EMS for 9 h before rinsing and field planting.
provide desired qualities such as texture from saturated M1 plants were advanced to M3 seeds by a single seed
fats, there have been efforts to increase the stearic acid descent at the Costa Rica Seeds winter nursery near
content of soybean seed oil (Boersma et al. 2012; Upala, Costa Rica; M3 seeds were planted at the Uni-
Carrero-Colón et al. 2014; Gillman et al. 2014; Jeong versity of Missouri South Farm Research Center in
et al. 2018; Lakhssassi et al. 2017; Rahman et al. 1997; Columbia, Missouri in 2014, and M3:4 seeds from single
Ruddle et al. 2013; Zhang et al. 2008). Increasing stearic plant threshes of 2431 plants were harvested.
Mol Breeding (2019) 39:79 Page 3 of 11 79
Seed sampling for fatty acid phenotyping downloaded from the raw sequence data to analyze the
read at single base pairs.
Seeds were produced in three field seasons at the South
Farm Research Center in Columbia, Missouri. A three Population development for allele combinations
seed composite from each M3:4 mutant lines from 2014
was analyzed for fatty acids in the oil with gas chroma- The 1478 and 1975 lines were each crossed with KB14-
tography as previously described (Beuselinck et al. 30#882 (FAD2-1AA, FAD2-1bb, and FAD3aacc) in the
2006). The population was assessed to have a contam- summer of 2016 at South Farm Research Center in
ination rate of less than 1% after identifying lines with Columbia, Missouri. Lines 1478 and 1975 were the
stearic acid less than 6% of the seed oil. Two lines were result of mutagenesis of the Williams 82 mutant line
identified (1478 and 1975) with elevated oleic acid 194D identified to have elevated stearic acid due to a
compared to the population that also had elevated stearic missense mutation (V211E) in the SACPD-C gene in
acid. Three single seeds were subsequently analyzed addition to the novel FAD2-1A mutant alleles (Gillman
independently from the two original mutant lines. In et al. 2014). In experimental line KB14-30#882, the
2015, seeds from the two mutant lines and control lines FAD2-1A alleles were wild-type and the FAD2-1B mu-
were produced in the field and analyzed for fatty acids tant alleles contained the missense mutation P137R
as three seed composite samples from each of three from PI 283327 (Pham et al. 2010). For line KB14-
individual mutant plants for line 1975 but one plant for 30#882, the FAD3A and FAD3C mutant alleles (splice
line 1478. In 2015, the control lines 194D (a selection of site and missense G128E, respectively) were derived
one mutagenized line from the population) and Williams from the mutant line CX1512-44 (Bilyeu et al. 2005).
82 (W82) were produced as part of a separate study as The F1 seeds from the resulting crosses, named KB16-
three plots of ten seeds each in a randomized complete 14 or KB16-15 for the parents 1478 or 1975 respective-
block design with seeds bulked from each plot. Williams ly, were harvested and sent to the Costa Rica Seeds
82 is considered wild-type for the alleles in fatty acid winter nursery in Upala, Costa Rica to grow F1 plants,
biosynthesis. Line 194D was selected from a mutagen- followed by self-fertilization to produce F2 seeds after
esis of Williams 82 for its elevated stearic acid and was one generation. Individual F2 seeds (225 from the 1478-
found to contain a V211E missense allele of the derived population and 325 from the 1975-derived pop-
SACPD-C gene. In 2016, the mutant lines and controls ulation) were chipped to analyze the fatty acid profiles
were part of a field study also as three plots of ten seeds by extracting seed oil from the chip for lipid gas chro-
each in a randomized complete block design with seeds matography (Beuselinck et al. 2006). The identified 13
bulked from each plot. or 18 F2 seeds produced in the Costa Rica environment
with a high oleic acid content (> 70%) from the KB16-
14 and KB16-15 populations, respectively, were germi-
FAD2-1A and FAD2-1B amplification and sequencing nated in moist seed packets, and surviving seedlings
were transplanted and grown at South Farm Research
DNA was isolated from the two mutant lines, named Center in Columbia, Missouri in the summer of 2017.
1478 and 1975. Primers and PCR amplification for the Not all seeds germinated or survived transplanting.
FAD2-1A and FAD2-1B genes were performed as ex- There were 12 F2 selected high oleic plants from
plained by Pham et al. (2010). The PCR product was KB16-14 and 17 plants from KB16-15 that produced
confirmed by agarose gel electrophoresis, purified, and F2:3 seeds. Following the harvest of single plants, five
Sanger sequenced by the DNA core center at the Uni- single F2:3 seeds from each line were analyzed for fatty
versity of Missouri. Sequence alignments were analyzed acids by gas chromatography. Genotypes of the F2
using a multiple sequence alignment from CLUSTALW plants were determined from DNA extracted from a
http://www.genome.jp/tools-bin/clustalw. Sequences mixture of three F2:3 seeds (see below). For samples
were aligned with Williams 82 reference to look for that were segregating for one or more of the
nucleotide variations. Protein translation was done target alleles, individual F3 seeds were also chipped for
using a six-frame translation http://www.bioline. gas chromatography, and the remainder of the seed was
com/media/calculator/01_13.html and protein utilized for DNA extraction and genotyping. Fatty acid
alignment with CLUSTALW. Chromat files were also analysis was on single F3 seeds or chipped F3 seeds in
79 Page 4 of 11 Mol Breeding (2019) 39:79
which DNA was made from the remaining seed section SACPD-C (Zhang et al. 2008). For SACPD-C, the
to identify samples which had segregated to the desired primers were 194Dsacpdexf: ACCCTGGGACAGAC
homozygous alleles by genotype assay. Single seed AACAAC and ZSACPDCrev: AGTCACAA
samples from control plants that had all wild-type alleles AGCCAAAAACCTG, and the 194Dsacpdexf primer
or just the SACPD-C V211E alleles were taken from was used for sequencing. PCR products were verified by
single plant threshes of Williams 82 or four independent agarose gel electrophoresis, followed by post-PCR pu-
194D plants produced at South Farm Research Center in rification. Samples were then Sanger sequenced at the
Columbia, Missouri in the summer of 2017. DNA Core Facility at the University of Missouri. Se-
quences were analyzed as described previously by
Genotyping and selection of soybean lines homozygous alignment with Williams 82 to assign genotypes for
for targeted mutant alleles FAD2-1A [wild-type (WT), C366R from 1478, or
T316I from 1975, and SACPD-C (WT or V211E)] at
The selected new high oleic acid soybean lines were the respective allele locations. The SNP in SACPD-C
genotyped for their fatty acid desaturase alleles after from 194D was the thymine to adenine change 632 bp
harvest of F2:3 seeds, so no agronomic evaluations were from the start codon responsible for a V211E missense
performed. One high oleic acid F2:3 line designated 53B mutation reported earlier (Gillman et al. 2014). Se-
from the KB16-14 population had the following geno- quence traces were downloaded in Chromat files to
type: mutant alleles of FAD2-1A (C366R), FAD2-1B, analyze clean reads and base pairs. A genotype of mu-
FAD3A, FAD3C, and SACPD-C. Five additional seeds tant or wild-type was assigned according to the base pair
from this line were chipped for gas chromatography present at the allele location, but was assigned hetero-
following gas chromatography analysis of five single zygous if both bases were present at that location in the
F2:3 seeds from each line. Two high oleic acid F2 plants Chromat file. For some high oleic acid F2 plants, it was
from the KB16-14 population were identified by geno- determined that there were S117N mutant alleles segre-
type to potentially produce F3 seeds for the high oleic/low gating (presumably from a parental donor plant of
linolenic acid (HOLL) trait containing four homozygous KB14-30#882 that was wild-type and mutant S117N
mutant genes. The genotype of line 100A was mutant heterozygous for FAD2-1A, even though the line was
alleles of FAD2-1A (C366R), FAD2-1B, and FAD3A, and thought to contain only wild-type FAD2-1A alleles)
segregating for FAD3C, whereas line 101A contained (Dierking and Bilyeu 2009). Subsequent to this discov-
mutant alleles of FAD2-1A (C366R), FAD2-1B, and ery, the DNA of all F2:3 samples was analyzed by
FAD3C, but was segregating for FAD3A. Twenty F3 seeds SimpleProbe assay for the 17D S117N FAD2-1A alleles
of each line 100A and 101A were chipped for fatty acid (Pham et al. 2012). Soybean lines containing FAD2-1A
analysis, and a set of seven was selected with the lowest S117N alleles were eliminated from our analysis (one
linolenic acid levels for germination of the remainder of from KB16-14 and six from KB16-15).
the seed, transplanting to the field and generation of tissue SimpleProbe assays were used for FAD3A and
for DNA extraction. Of the surviving plants, DNA was FAD3C as described by Bilyeu et al. (Bilyeu et al.
extracted, genotyping was performed, and four plants 2011), as well as FAD2-1B (P137) and FAD2-1A
(three from 100A and one from 101A) were confirmed (17D) as described by Pham et al. (Pham et al. 2012).
to be homozygous for the four targeted mutant alleles. These genotyping assays were run using a Lightcycler
The fatty acid phenotypes of seed chips for the four F3 480 II real-time PCR instrument (Roche Applied Sci-
seeds with the mutant alleles of FAD2-1A (C366R), ences). Genotypes were assigned according to their
FAD2-1B, FAD3A, and FAD3C were combined. characteristic melting curves. The details of the assays
that distinguish the alleles used for sequence analysis
DNA extraction and genotype analysis and genotyping are provided (Table 1).
Table 1 Details of genotyping assays distinguishing variant alleles of fatty acid desaturase genes from their wild-type alleles
top 100 best-matched sequences were aligned and used oleic acid, respectively, in the seed oil compared to
as input for sequence LOGO http://weblogo.berkeley. 19.3% oleic acid for the mutant population mean and
edu/logo.cgi. 26.6% oleic acid for the control line Williams 82
(Table 2). There was no obvious segregation for oleic
Statistics acid content in lines 1478 or 1975. Seeds of lines 1478
and 1975 produced in the field in 2015 and 2016 also
Minitab statistics software was used for an ANOVA and contained increased levels of oleic acid in the seed oil
Fisher pairwise comparisons for genotypes in Table 3 compared to a mutant control line with the SACPD-C
with p < 0.05. The Student’s t test with p values less than mutation (194D) and Williams 82.
0.01 was used to determine significant differences in The candidate genes for elevated oleic acid in soy-
mean fatty acid values for different genotypes for the bean seed oil were FAD2-1A and FAD2-1B. These genes
four and five allele combinations in Table 4. were amplified and sequenced from lines 1478 and
1975. Two novel missense alleles were found in the
FAD2-1A gene in lines 1478 and 1975. The SNPs were
Results identified from comparing sequence alignments to the
Williams 82 reference sequence. No new mutations
Discovery of novel missense alleles in FAD2-1A gene were found in FAD2-1B. The missense allele in the
in mutant soybean lines with increased oleic acid 1478 line was a cysteine to arginine amino acid change
at position 366 FAD2-1A (C366R) near the carboxyl
Two mutant soybean lines with increased oleic acid in terminus of the protein from a thymine to cytosine
the seed oil were identified in a forward genetics screen mutation at position 1096 of the coding sequence
of an elevated stearic acid mutant population fixed for a (Fig. 1). The missense allele in the 1975 line was a
missense V211E mutation in SACPD-C (Gillman et al. threonine to isoleucine amino acid change at position
2014). Seeds from lines 1478 and 1975 produced in the 316 FAD2-1A (T316I) from a cytosine to thymine mu-
field in 2014 exhibited a mean of 35.3% and 33.8% tation at position 947 of the coding sequence (Fig. 1).
79 Page 6 of 11 Mol Breeding (2019) 39:79
Table 2 Seed oil fatty acid composition of field-produced soybean mutants 1478, 1975, and control lines over 3 years
To examine the possible effect these alleles might mutant donor either line 1478 or 1975, and the other
have on enzyme function, the amino acid conserva- parent an experimental line that contained mutant alleles
tion at the respective missense locations was ana- for elevated oleic acid and low linolenic acid in the seed
lyzed. The neutral cysteine amino acid at position oil (Fig. 2). The process for population development
366 of FAD2-1A is highly conserved, although it is was to first produce F1 seeds that were heterozygous
near a region of low conservation (amino acid posi- for all five mutant alleles, where the FAD2-1A mutant
tions 370–373). The arginine replacement in line donor was either the C366R from line 1478 or the T316I
1478 is positively charged and polar (Fig. 1). The from line 1975. Lines 1478 and 1975 also contributed to
neutral threonine amino acid at position 316 of V211E mutant alleles of SACPD-C (Fig. 1). Seed chips
FAD2-1A changed to the neutral amino acid isoleu- from individual F2 seeds produced in the winter nursery
cine in line 1975 was found in the middle of the were analyzed for the oleic acid levels, and seeds with
highly conserved region II described to be critical greater than 70% oleic acid were selected for growth and
for enzyme function (Shanklin et al. 1994). genotype and phenotype characterization in our US field
environment. The fatty acid profile was subsequently
Assessment of combinations of novel alleles determined for fatty acids in F2:3 seeds. The experimen-
with increased oleic acid tal line KB14-30#882 parent contained wild-type alleles
of FAD2-1A and SACPD-C and homozygous mutant
To evaluate the ability of the novel FAD2-1A alleles to alleles for FAD2-1B (P137R), FAD3A (splice), and
contribute to high oleic acid and high stearic acid oil FAD3C (G128E) (Pham et al. 2010; Pham et al. 2012).
phenotypes, novel soybean germplasm was created for The other parent was either line 1478 with homozygous
two populations with the FAD2-1A allele and SACPD-C mutant alleles for FAD2-1A (C66R) and SACPD-C
Mol Breeding (2019) 39:79 Page 7 of 11 79
c 1478
weblogo.berkeley.edu
1975
weblogo.berkeley.edu
Fig. 1 Characterization of novel FAD2-1A missense alleles in in conservation for the FAD2-1A enzyme as part of the BLINK
soybean mutant lines 1478 and 1975. a Partial FAD2-1A amino feature at NCBI using GI no. 197111722. The top 100 best-
acid sequence alignment with reference Williams 82. Amino acid matched sequences were aligned and used as input for sequence
position is indicated in front of the alignment. Identical amino LOGO http://weblogo.berkeley.edu/logo.cgi. Each stack of
acids are highlighted in black and the mutations in lines 1975 and symbols represents a position in the amino acid sequence. The
1478 are indicated with no highlight. b Chromas DNA sequence height of the stack relates to the sequence conservation at the
traces for the two mutations. The peaks correspond to the DNA corresponding location, and the individual symbol height
sequence information from the Sanger sequence analysis. The represents the relative frequency of the amino acids at that
positions of the DNA sequence change are indicated above the location. The arrows point to the amino acid missense positions
trace by the red highlight. c Weblogo output of amino acid as indicated in a box for line 1478 and 1975
(V211E) or line 1975 with homozygous mutant alleles addition to the wild-type and mutant versions of
for FAD2-1A (T316I) and SACPD-C (V211E). SACPD-C (Table 3). The status of the FAD3A and
Within the two mutant population selections, a phe- FAD3C alleles was ignored for this analysis. Plants that
notypic comparison of field-produced F3 seeds was were confirmed by genotype to be homozygous mutant
conducted for the two novel mutant alleles of FAD2- for the FAD2-1A (either C366R or T316I) and FAD2-1B
1A in combination with mutant alleles of FAD2-1B in (P137R) alleles and homozygous mutant (V211E) or
79 Page 8 of 11 Mol Breeding (2019) 39:79
Fig. 2 Process for creating novel soybean germplasm segregating One of the objectives of this work was to develop
for mutant alleles of five different genes involved in fatty acid
desaturation of the seed oil. New mutant lines 1478 and 1975 with
soybean germplasm with the high oleic acid/low
FAD2-1A and SACPD-C mutations were individually crossed with linolenic acid (HOLL) trait and also the HOLL trait
the experimental line KB14-30#882 with mutant alleles of FAD2- with the high stearic acid trait. For the HOLL trait,
1B, FAD3A, and FAD3C. The two F2 populations were screened four mutant alleles were targeted, FAD2-1A, FAD2-
for fatty acid phenotype to identify high oleic acid individual F2
seeds that were germinated and confirmed by genotype to be
1B, FAD3A, and FAD3C; for the HOLL plus high
homozygous for the novel combinations of mutations in FAD2- stearic acid trait, the four FAD alleles plus the
1A and FAD2-1B. These high oleic acid F2 plants were segregating SACPD-C alleles were targeted. From F2:3 seeds, a
for the other three genes, and a subset of those plants was identified single plant (53B) from the 1478 FAD2-1A (C366R)
that contained different homozygous alleles of SACPD-C, FAD3A,
and FAD3C. Those plants were evaluated for their F2:3 seed fatty
population was identified with the homozygous com-
acid phenotypes. Most of the lines were still segregating for bination of all five targeted alleles desired for the
SACPD-C, FAD3A, or FAD3C, so additional genotyping was done HOLL plus high stearic acid trait. None of the F2
to identify individual seeds with homozygous genotypes generation plants from the 1478 FAD2-1A (C366R)
population were homozygous for the four targeted
functional (WT) for the SACPD-C alleles were ana- HOLL mutant genes and wild-type for SACPD-C.
lyzed. Since there was no initial selection for the stearic However, phenotypic screening and genotypic selec-
acid phenotype, many of the lines were still segregating tion of homozygous mutant FAD3A or FAD3C alleles
for the SACPD-C alleles, so some samples were chipped from two segregating F 2:3 plants from the 1478
F3 seeds in which the remainder of the seed was geno- FAD2-1A (C366R) population resulted in the identi-
typed to determine the SACPD-C allele status. Lines fication of four seeds containing homozygous alleles
with mutant FAD2-1A and FAD2-1B alleles but wild- of the four targeted mutant FAD2 genes with wild-
type SACPD-C alleles are designated herein as HO, type alleles of SACPD-C (designated 100/101A). The
while lines with mutant FAD2-1A, FAD2-1B, and fatty acid profile of the line 53B demonstrated a high
SACPD-C alleles are designated herein as HOS. Seeds oleic acid and low linolenic acid phenotype (< 3%
from lines containing the 1478 FAD2-1A C366R alleles linolenic acid), and stearic acid was 8.4% of the seed
Mol Breeding (2019) 39:79 Page 9 of 11 79
Table 3 Seed oil fatty acid composition of 2017 field-produced F3 seeds selected for different combinations of FAD2-1A, FAD2-1B, and
SACPD-C mutant alleles
oil (Table 4). The fatty acid profile of the line 100/ containing elevated stearic acid in the seed oil due to
101A selections also had a high oleic acid and low missense alleles of the SACPD-C gene. When com-
linolenic acid phenotype, and stearic acid was 3.3% paring the two new FAD2-1A missense alleles, there
(Table 4). The inclusion of homozygous selection for was an obvious distinction between the oleic acid
mutant FAD3A and FAD3C alleles produced a mean levels in the seed oil only after combining with mu-
linolenic acid level of 2.3% of the seed oil when the tant FAD2-1B alleles. Lines containing the 1478
FAD2-1A C366R and FAD2-1B P137 alleles were C366R FAD2-1A alleles had a higher mean oleic acid
combined. seed oil content compared to lines containing the
1975 FAD2-1A allele by ~ 22% when combined with
mutant alleles of FAD2-1B. This result was not un-
Discussion precedented, since we and others have determined
that not all alleles of FAD2-1A or FAD2-1B have the
Two novel missense mutation alleles of the soybean same effects on oleic acid levels, singly or in combi-
FAD2-1A gene encoding an enzyme that converts nation (Bilyeu et al. 2018; Hoshino et al. 2010; Pham
oleic acid precursors into linoleic acid precursors in et al. 2011; Pham et al. 2010; Sweeney et al. 2017;
the seed oil were identified. The novel alleles were Thapa et al. 2016). The 1478 C366R FAD2-1A allele
discovered in a background of a mutant soybean line caused an amino acid change that did not retain
Table 4 Seed oil fatty acid composition of 2017 field produced low linolenic acid soybean mutants 53B and 100/101A
53B HOLLSa 10 6.8 ± 0.4 8.4 ± 1.2**b 75.8 ± 1.6** 6.7 ± 1.6* 2.3 ± 0.2
100/101A HOLL 4 6.6 ± 0.3 3.3 ± 0.3 83.4 ± 1.9 4.5 ± 1.3 2.2 ± 0.3
a
HOLLS indicates the homozygous allele combination FAD2-1A (C366R), FAD2-1B (P137R), FAD3A (splice), FAD3C (G128E), and
SACPD-C (V211E), while HOLL has the same genotype, except it was wild-type for SACPD-C
b
The mean values are indicated followed by ±, the standard deviation, and an * or ** is indicated if the mean values in the column differ at
p < 0.05 or 0.01significance, respectively
79 Page 10 of 11 Mol Breeding (2019) 39:79
similar properties to the wild-type version; the 1975 Soybean breeding for seed oil quality has a long
T316I FAD2-1A amino acid substitution was in a history with variation accessed through mutagenesis,
more conserved region, but the amino acid substitu- biotechnologies, and natural variation. The identifica-
tion was with one with similar properties. The allelic tion of the causative alleles for seed oil traits shifts the
series of FAD2-1A with FAD2-1B genes now avail- breeding scheme so that direct selection of those
able may provide additional information to under- alleles is potentially more efficient than phenotypic
stand how structural features of these enzymes influ- selection during the breeding process. We routinely
ence their biochemical activities. develop segregating populations and use perfect mo-
Our results demonstrated that combining certain lecular markers for genotypic selection of four to
mutant alleles of FAD2-1A with FAD2-1B genes pro- eight alleles. Based on the expected inheritance ra-
duced a high oleic acid phenotype (> 70%) in the tios, determining the appropriate population size and
1478 population containing the C366R FAD2-1A al- fixing as many desired homozygous alleles as early
leles plus the P137R FAD2-1B alleles. The oleic acid as possible is one strategy that has been successful.
phenotype in seed oil of this new soybean germplasm For example, selection of all five alleles for high
was similar to the oleic acid phenotype of other oleic acid, low linolenic acid, and elevated stearic
characterized high oleic acid soybean lines produced acid from the quintuple heterozygote F1 would be 1/
in multiple environments (Bilyeu et al. 2018). As 1024 F2 samples, although fixing four of those genes
expected, combining FAD3A and FAD3C homozy- with a fifth one still segregating increases the ratio
gous mutant alleles also reduced the linolenic acid, slightly to 3/1024. Once the targeted alleles have
to create HOLL lines, comparable with previous ex- been fixed in different genetic backgrounds,
periments (Bilyeu et al. 2018; Pham et al. 2012). The intercrossing those lines can produce populations that
new soybean germplasm developed from line 100/ segregate for other phenotypes without regard for the
101A offers alternative opportunities to develop soy- oil quality trait.
bean varieties with the high oleic acid/low linolenic In conclusion, this research demonstrates that using
acid oil trait. traditional plant breeding to combine novel mutant al-
There have been reports of varying levels of stearic leles of FAD2-1A with FAD2-1B, FAD3A, and FAD3C
acid for soybean lines combining the high oleic acid genes will result in elevated levels of oleic acid and
alleles of FAD2-1A and FAD2-1B with mutant alleles reduced levels of linolenic acid in the seed oil. In com-
of SACPD-C. In one study, HOS combinations pro- bination with SACPD-C, stearic acid seed oil content is
duced either 12.7% or 10.4% stearic acid in the high elevated; however, the trade-off is that oleic acid levels
oleic genetic background (Ruddle et al. 2014). In our decrease. These findings are beneficial as they contrib-
previous work, we characterized HOS lines that pro- ute to our knowledge and ability to enhance soybean oil
duced over 18% stearic acid in a high oleic acid quality.
genetic background, but the stearic acid phenotype
was variable depending on the production environ-
Acknowledgments Christine Cole and Paul Little provided ex-
ment (Bilyeu et al. 2018). One F2 plant (53B) from cellent technical assistance for this research.
our 1478 population was identified to be mutant for all
five alleles of interest: FAD2-1A, FAD2-1B, FAD3A,
FAD3C, and SACPD-C. Although this germplasm Funding information Funding for some of this project was
provided as a grant from the United Soybean Board.
contained less than the target stearic acid level, it still
produced a high oleic acid phenotype of ~ 76% oleic
Compliance with ethical standards
acid with elevated stearic acid in the seed oil. The
functional properties of oil with elevated stearic acid Disclaimer Mention of trade names or commercial products in
and high oleic acid plus low linolenic acid require this publication is solely for the purpose of providing specific
further characterization. It is possible that this unique information and does not imply recommendation or endorsement
by the US Department of Agriculture. USDA is an equal oppor-
combination may provide desirable functionalities in
tunity provider and employer.
foods or for industrial uses. This germplasm will be
important for future studies on the advancement of Conflict of interest The authors declare that they have no con-
soybean oil functionality. flict of interest.
Mol Breeding (2019) 39:79 Page 11 of 11 79