A phenylalanine in DGAT is a key determinant of oil content and composition in maize.
ABSTRACT Plant oil is an important renewable resource for biodiesel production and for dietary consumption by humans and livestock. Through genetic mapping of the oil trait in plants, studies have reported multiple quantitative trait loci (QTLs) with small effects, but the molecular basis of oil QTLs remains largely unknown. Here we show that a high-oil QTL (qHO6) affecting maize seed oil and oleic-acid contents encodes an acyl-CoA:diacylglycerol acyltransferase (DGAT1-2), which catalyzes the final step of oil synthesis. We further show that a phenylalanine insertion in DGAT1-2 at position 469 (F469) is responsible for the increased oil and oleic-acid contents. The DGAT1-2 allele with F469 is ancestral, whereas the allele without F469 is a more recent mutant selected by domestication or breeding. Ectopic expression of the high-oil DGAT1-2 allele increases oil and oleic-acid contents by up to 41% and 107%, respectively. This work provides insights into the molecular basis of natural variation of oil and oleic-acid contents in plants and highlights DGAT as a promising target for increasing oil and oleic-acid contents in other crops.
- SourceAvailable from: Changming Lu[Show abstract] [Hide abstract]
ABSTRACT: Oilseed rape (Brassica napus) is one of the most important oilseed crops globally. To meet increasing demand for oil-based products, the ability to enhance desirable oil content in the seed is required. This study assessed the capability of five genes in the triacylglyceride (TAG) synthesis pathway to enhance oil content. The genes BnGPDH, BnGPAT, BnDGAT, ScGPDH and ScLPAAT were overexpressed separately in a tobacco (Nicotiana benthamiana) model system, and simultaneously by pyramiding in B. napus, under the control of a seed specific Napin promoter. ScLPAAT transgenic plants showed a significant increase of 6.84% to 8.55% in oil content in tobacco seeds, while a ~4% increase was noted for BnGPDH and BnGPAT transgenic seeds. Seed-specific overexpression of all four genes in B. napus resulted in as high a 12.57% to 14.46% increased in seed oil content when compared to WT, equaling close to the sum of the single-gene overexpression increases in tobacco. Taken together, our study demonstrates that BnGPDH, BnGPAT and ScLPAAT may effectively increase seed oil content, and that simultaneous overexpression of these in transgenic B. napus may further enhance the desirable oil content relative to single-gene overexpressors. Copyright © 2014. Published by Elsevier B.V.Gene 12/2014; · 2.20 Impact Factor
- Crop Science 01/2013; 53(6):2518. · 1.48 Impact Factor
- Crop Science 01/2011; 51(2):433. · 1.48 Impact Factor
A phenylalanine in DGAT is a key determinant of oil
content and composition in maize
Peizhong Zheng1, William B Allen1, Keith Roesler1, Mark E Williams2, Shirong Zhang1, Jiming Li1,
Kimberly Glassman1, Jerry Ranch1, Douglas Nubel1, William Solawetz1, Dinakar Bhattramakki1,
Victor Llaca2, Ste ´phane Deschamps2, Gan-Yuan Zhong1,3, Mitchell C Tarczynski1& Bo Shen1
Plant oil is an important renewable resource for biodiesel
production and for dietary consumption by humans and
livestock. Through genetic mapping of the oil trait in plants,
studies have reported multiple quantitative trait loci (QTLs)
with small effects, but the molecular basis of oil QTLs remains
largely unknown1–5. Here we show that a high-oil QTL (qHO6)
affecting maize seed oil and oleic-acid contents encodes an
acyl-CoA:diacylglycerol acyltransferase (DGAT1-2), which
catalyzes the final step of oil synthesis. We further show that
a phenylalanine insertion in DGAT1-2 at position 469 (F469)
is responsible for the increased oil and oleic-acid contents.
The DGAT1-2 allele with F469 is ancestral, whereas the
allele without F469 is a more recent mutant selected by
domestication or breeding. Ectopic expression of the high-oil
DGAT1-2 allele increases oil and oleic-acid contents by up to
41% and 107%, respectively. This work provides insights into
the molecular basis of natural variation of oil and oleic-acid
contents in plants and highlights DGAT as a promising target
for increasing oil and oleic-acid contents in other crops.
Most vegetable oil is used directly for human consumption, but the
nonfood use of oil is growing rapidly, especially in Europe, where it is
used as a primary feedstock for biodiesel production. Adding to the
increasing demand for vegetable oil production, studies continue to
identify benefits associated with oil quality or composition. For
example, oleic acid, a monounsaturated fatty acid, has been linked
to considerable health and cooking benefits when compared to
saturated and polyunsaturated fatty acids6,7. Increasing seed oil
quantity and oleic-acid content would be an important step toward
addressing these market needs; moreover, these traits are key targets
for plant breeding and biotechnology.
The Illinois high-oil (IHO) population and the Alexho single-kernel
(ASK) synthetic population are examples of high-oil maize developed
by breeders through recurrent selection to increase energy density for
animal feed8,9. Although seed oil content in these populations has
reached as high as 22%, commercial high-oil maize hybrids have not
been released, mainly because a significant reduction in grain yield
and other undesirable agronomic traits are associated with high-oil
germplasm. However, these and other high-oil maize populations
have provided valuable materials for oil QTL mapping. Recent
studies indicate that maize seed oil content is a quantitative trait
affected by several QTLs, including one on chromosome 6 in bin
6.04 (refs. 1–5). Despite a good understanding of the plant oil
biosynthetic pathway, and although many genes involved in the oil
pathway have been isolated, the molecular basis for oil QTL has
remained largely unknown.
Most maize seed oil is located in the embryo6. Seed oil content is
thus primarily determined by the oil concentration of the embryo and
the proportion of the seed occupied by the embryo. Using a BC2
(second backcross generation) population derived from a cross
between a high-oil inbred line, ASKC28IB1 (ASK cycle 28 inbred 1,
18.8% and 56.4% seed and embryo oil concentration, respectively),
and a normal-oil inbred line, PH09B (3.2% and 26.1% seed and
embryo oil concentration, respectively), we identified a high-oil QTL
on chromosome 6 (qHO6) as a key oil QTL controlling both seed oil
content and embryo oil concentration, but not embryo size and seed
weight. qHO6 was mapped to 49.81 cM—near the marker UMC1918
(Fig. 1a)—with a lod score of 7.5, and it explained 11% and 9.5% of
the variation for seed oil content and embryo oil concentration,
respectively (Supplementary Table 1 online). Using marker-assisted
selection, we developed overlapping near-isogenic lines (NILs) that
contained different lengths of qHO6 segments from ASKC28IB1 for
fine mapping. Using embryo oil concentration data from these NILs,
we mapped qHO6 to a region of 4.2 cM in the BC3S2population,
between SSR marker UMC1145 and SNP marker MZA5002 (Fig. 1b
and Supplementary Fig. 1 online).
In the BC4S2population, NILs homozygous for the ASKC28IB1
allele (qHO6+/+) showed a 17.1% increase in seed oil content, a 18.7%
increase in embryo oil concentration and no significant changes in
seed or embryo weight (Table 1). Because a locus associated with high
oleic-acid content, ln1, has been reported in the same region as qHO6
in the IHO line10,11, we also determined the fatty-acid composition in
Received 1 October 2007; accepted 4 December 2007; published online 17 February 2008; doi:10.1038/ng.85
1Pioneer Hi-Bred International Inc., A DuPont Company, 7300 NW 62nd Avenue, PO Box 1004, Johnston, Iowa 50131, USA.2DuPont Crop Genetics Research,
Experimental Station, PO Box 80353, Wilmington, Delaware 19880, USA.3Present address: United States Department of Agriculture-the Agricultural Research
Service (USDA-ARS), Grape Genetics Research Unit, Geneva, New York 14456, USA. Correspondence should be addressed to B.S. (Bo.Shen@pioneer.com).
NATURE GENETICS VOLUME 40 [ NUMBER 3 [ MARCH 2008367
© 2008 Nature Publishing Group http://www.nature.com/naturegenetics
BC4S2NILs and found that the qHO6+/+lines showed a 61.3%
increase in oleic-acid content and a 24.1% decrease in linoleic-acid
content (Table 1). As oleic-acid concentration showed a larger
increase than oil content and can be measured more accurately, we
used oleic-acid content in addition to oil content in subsequent
generations to detect the presence of qHO6 in fine mapping. Using
overlapping NILs, we mapped qHO6 to a small region between SNP
markers BAC18 and BAC32 (Fig. 1b), corresponding to three over-
lapping BAC clones. By sequencing the three BAC clones, we identified
five putative genes that encode a potassium efflux system protein
(KESP), a type I acyl-CoA:diacylglycerol acyltransferase (DGAT1-2), a
40S ribosomal protein S24, an amino acid permease (AAP) and the 5¢
half of a PDR-like ABC transporter (ABCT) (Fig. 1b).
DGAT catalyzes the final step in the Kennedy pathway for triacyl-
glycerol (TAG) biosynthesis; this step has been implicated as a rate
limiting step12,13. In dairy cattle, a QTL affecting milk yield and
composition has been genetically mapped to a type I DGAT and
functionally confirmed14. Thus, DGAT1-2 was a strong candidate for
qHO6. We isolated full-length cDNAs for DGAT1-2 from developing
embryos of qHO6 NILs homozygous for the ASKC28IB1 high-oil
allele or the PH09B normal-oil allele. Their deduced amino acid
sequences contained four residue changes: three near the N terminus
(V45G, P55S and Q67 deletion), and a phe-
C terminus in the ASKC28IB1 allele (Fig. 1b
and Supplementary Fig. 2 online). To con-
firm that qHO6 and DGAT1-2 were allelic, we
screened an additional 4,000 BC5S1seedlings
to isolate new recombinants using SNP mar-
kers developed from the three Mo17 BAC
114 recombinants, including five plants with
crossovers within DGAT1-2. Recombinant
plants were self-pollinated to produce BC5S2
seeds, and oleic acid and oil contents were
determined in selected recombinant lines for
fine mapping. As a result, we mapped qHO6
to a 4.8-kb fragment (between SNP markers
BAC22 and BAC17) located in the 3¢ end of
DGAT1-2 (Fig. 1b), with no segregation of
the high-oil and high-oleic-acid phenotypes.
Within this region, the only difference in
amino acid sequence between the two alleles
was the extra phenylalanine (F469) in the
ASKC28IB1 allele, indicating that F469 may
have a critical role in increasing oil and oleic-
acid concentrations (Fig. 1b).
We confirmed this result transgenically by transforming constructs
containing the embryo-preferred 16-kDa oleosin promoter driving
either the ASKC28IB1 or PH09B alleles of the DGAT1-2 cDNA into
maize. The constructs also contained a DS-RED2 gene driven by an
aleurone-specific lipid-transfer protein 2 (LTP2) promoter to facilitate
identification of transgenic and null kernels for phenotypic analysis.
Both constructs produced T1lines that increased seed oil and oleic-
acid contents (Fig. 2a,b). For the 29 lines created and analyzed for
each construct, we found that the construct containing the ASK-
C28IB1 DGAT1-2 allele produced a higher percentage of lines with a
considerable increase in oil compared to the construct containing the
PH09B DGAT1-2 allele (Fig. 2a,b). Ectopic expression of the PH09B
DGAT1-2 allele (normal oil) led to average increases of 9.3%, 12.4%
and 40.5% in seed oil, embryo oil and oleic-acid concentrations,
respectively, and a 13.7% decrease in linoleic-acid concentration. In
comparison, ectopic expression of the ASKC28IB1 DGAT1-2 allele
(high oil) led to average increases of 27.9%, 26.1% and 84.5% in seed
oil, embryo oil and oleic-acid concentrations, respectively, and a
28.1% decrease in linoleic-acid concentration—a more than a twofold
increase in the four measured traits relative to the PH09B allele. The
ASKC28IB1 line with the highest increases shows a 41% increase in
seed-oil content and a 107% increase in oleic-acid content. Notably,
Promoter3′ UTRDGAT1-2 coding
Q67 delF469 ins
Figure 1 Mapping and cloning of qHO6. (a) Map
position of the high-oil QTL qHO6 in the BC2
generation. Positions for the first and last markers
on chromosome 6 used are indicated. (b) Fine
mapping of qHO6. Approximate lengths (cM or
kb) between the left and right flanking markers
are shown on the right. Locations of the four
polymorphic residues between the two mapping
parents in DGAT1-2 protein are indicated. Letters
to the right of the amino acid–position number
represent the amino acid in the ASKC28IB1
allele; letters to the left represent the PH09B
allele. Deletion (del) and insertion (ins) also refer
to the ASKC28IB1 allele.
Table 1 Effects of qHO6 on different grain traits in the BC4S2generation
3.0 ? 10–7
2.8 ? 10–5
2.5 ? 10–3
1.4 ? 10–7
Data are averages of five homozygous ASKC28IB1 and five homozygous PH09B NILs, with ten seeds per line analyzed.
A Student’s t-test was used to generate the P value.
368VOLUME 40 [ NUMBER 3 [ MARCH 2008 NATURE GENETICS
© 2008 Nature Publishing Group http://www.nature.com/naturegenetics
the increases in seed oil content and embryo oil concentration caused
by ectopic expression of the ASKC28IB1 allele were greater than those
caused by the native ASKC28IB1 allele of qHO6, as observed in BC4S2
homozygous NILs (Fig. 2a and Fig. 1b). Furthermore, homozygous
T3seeds from five ASKC28IB1 allele transgenic lines showed a similar
increase in seed oil and oleic content as T1seeds without changes in
seed weight, indicating that oil and oleic-acid traits are stable in
transgenic lines (Fig. 2c,d and Supplementary Fig. 3 online).
Together, these data indicate that DGAT1-2 is the causal gene for
qHO6 and is responsible for the increased oil and oleic-acid contents.
To determine further whether the extra phenylalanine at position
469 (F469) in the ASKC28IB1 allele was responsible for the qHO6
phenotype, we measured the enzymatic activities of DGAT1-2 alleles
following their expression in a yeast strain deficient in oil biosynthesis.
We transformed yeast cells with plasmids containing either the
ASKC28IB1 or PH09B alleles or their mutant forms of DGAT1-2
driven by a constitutive 3-phosphoglycerate kinase (PGK) promoter
(Fig. 3a). Yeast cells containing the ASKC28IB1 DGAT1-2 allele had
higher enzyme activity than those containing the PH09B allele
(Fig. 3b). Of note, when F469 was deleted from the ASKC28IB1 allele,
the enzyme activity was reduced to a level similar to that of cells
containing the PH09B allele. Conversely, inserting F469 into the
PH09B allele increased enzyme activity to the level of cells containing
the ASKC28IB1 allele (Fig. 3b). The amount of TAG accumulated in
these cells correlated well with the level of DGAT activity (Fig. 3c).
Protein blots of yeast microsomal protein consistently showed lower
amounts of DGAT protein for the ASK28IB1 allele than for the PH09B
allele (Fig. 3d). Thus, the difference in specific activity between the
two alleles may be greater than suggested in Figure 3b. Protein
structure analysis15predicts that the critical F469 is located in the
last of eight transmembrane domains in the DGAT1-2 protein
(Fig. 3e). A deletion of F469 in a transmembrane helix would shift
the side-chain positions of all residues from the F469 to the end of the
helix (positions 469–482). Such a major change could alter the nature
(for example, hydrophobicity) of the helix, which could alter overall
protein structure or oligomeric state in a way that would influence
both activity and stability. These results show that a single amino-acid
difference (F469) is responsible for the different enzyme activities and
TAG contents observed in yeast.
Previous studies10,11have reported a key locus associated with oleic-
and linoleic-acid content (ln1) from the IHO line in the same region
as qHO6. We sequenced part of the IHO DGAT1-2 covering the four
polymorphic residues described earlier, and found that it was identical
to the ASKC28IB1 allele. This was not surprising, because IHO was
one of the 56 varieties that comprised the base population of ASK9.
These data suggest that qHO6 (DAGT1-2) is also responsible for the
Embryo oilSeed oil
ASK PH09B ASK
Oleic acidLinoleic acid
ASK PH09B ASK
Seed oil (%)
Oleic acid (%)
Null Transgenic seed
Figure 2 Transgenic ectopic expression of DGAT1-2 in maize. (a) Effects of ectopic expression of the PH09B allele (filled triangles) or the ASKC28IB1 allele
(open circles) of DGAT1-2 on embryo oil concentration and seed oil content. (b) Effects of ectopic expression of the PH09B allele (filled triangles) or the
ASKC28IB1 allele (open circles) of DGAT1-2 on seed oleic acid and linoleic-acid content. Symbols (triangles and circles) represent the relative change in
mean between T1transgenic and null seeds for a single line. For each line, we analyzed ten transgenic and ten null seeds harvested from the same ear. The
horizontal bar represents the mean of all 29 transgenic lines. (c) Seed oil content of T3homozygous transgenic ASKC28IB1 lines and null segregants. For
each line, we analyzed ten seeds per ear for five homozygous transgenic or null ears. Data are shown as mean ± s.d. All five transgenic lines showed a
significant increase in seed oil content compared to their null segregants (P o 0.01, t-test). (d) Seed oleic-acid content of T3homozygous transgenic
ASKC28IB1 lines and null segregants. For each line, we analyzed three seeds per ear for five homozygous transgenic or null ears. Data are shown as
mean ± s.d. All five transgenic lines showed a significant increase in seed oleic acid content compared to their null segregants (P o 0.001, t-test).
Yeast (∆yDGAT, ∆yPDAT)
VectorASK PH09B+FASK-F PH09B
460 VGNMIFWFFFSIVGQPMCVLLYY 482
Figure 3 Expression of maize DGAT 1-2 variants in yeast. (a) Yeast strain
and construct. PGK1 PRO and PGK1 TER, S. cerevisiae phosphoglycerate
kinase (PGK1) promoter and terminator. (b) DGAT enzymatic activity assay.
ASK, the original ASKC28IB1 allele of DGAT1-2 cDNA; PH09B, the original
PH09B allele of DGAT1-2 cDNA; ASK-F, ASKC28IB1 allele with F469
deletion; PH09B+F, PH09B allele with F469 insertion. Data represent mean
± s.d. and are expressed as pmol of14C-labeled TAG produced per minute
per mg of microsomal protein. (c) TAG accumulation in yeast cells. Data
represent mean ± s.d. (d) Protein blot using antibodies against maize
DGAT1-2. (e) Predicted transmembrane domains of DGAT1-2. Amino acid
sequence and position of the last transmembrane domain are shown. F469
is in red.
NATURE GENETICS VOLUME 40 [ NUMBER 3 [ MARCH 2008369
© 2008 Nature Publishing Group http://www.nature.com/naturegenetics
high oleic content in IHO and that ln1 and qHO6 are likely the same
gene. The increase in oleic-acid content of the qHO6+/+line may be
explained by the fact that diacylglycerol is both a substrate for DGAT
to produce TAG, and a substrate for choline phosphotransferase to
produce phosphatidyl choline. The substrate for FAD2 is oleoyl–
phosphatidyl choline. If a higher activity DGAT increased flux of
oleoyl-DAG to TAG and decreased flux to oleoyl–phosphatidyl cho-
line, then FAD2 would be deprived of substrate and less linoleate
would result16,17. Because ectopic expression of the PH09B allele also
significantly increased oleic-acid content and reduced linoleic-acid
content in transgenic plants (Fig. 2b), it is less likely that the
ASKC28IB1 DGAT1-2 protein has a specific substrate preference for
oleic acid compared to linoleic acid.
A phylogenic analysis of known plant type I DGAT protein
sequences shows that monocot and dicot type I DGATs fall into two
main groups. Within the monocot group, DGAT1-1s and DGAT1-2s
form two distinct clusters (Fig. 4a). Notably, F469 is conserved in all
plant type I DGATs. Within maize, however, the F469 deletion allele is
found frequently in inbred lines, such as PH09B, Mo17 and B73
(Fig. 4b and Supplementary Table 2 online). To determine which
allele is ancestral in maize, we amplified and sequenced a 751-bp DNA
fragment around F469 from maize wild relatives, Zea mays ssp.
heuhuetenangenis, Zea mays ssp. hexicana, Zea mays ssp. parviglumis,
Zea diploperennis and Zea luxurians. All 46 accessions of teosinte lines,
including 20 accessions of the maize progenitor, Zea mays ssp.
parviglumis, contain F469 (Fig. 4b and Supplementary Table 2).
This suggests that the ASKC28IB1 allele with F469 is ancestral and that
the PH09B allele is a more recent deletion mutant selected either by
domestication or breeding.
A previous study using genome-wide association mapping with
8,590 markers in 553 maize inbred lines did not report a strong
association of markers at the qHO6 region with oleic-acid content,
because of the low frequency of the high-oil allele in modern inbred
lines18. In order to determine whether F469 is associated with seed oil
and oleic-acid variations in maize, we selected 71 maize lines with high
or normal embryo oil and oleic-acid concentrations and resequenced
the DGAT1-2 polymorphic region. The results showed that maize lines
with F469 all had high concentrations of seed oil and oleic acid,
whereas none of the lines with normal concentrations of seed oil or
oleic acid contained F469 (Fig. 4c). In contrast to the ASKC28IB1
haplotype for F469, the ASKC28IB1 haplotypes for the other three
polymorphic amino acids, V45G, P55S and the Q67 deletion, were not
always associated with high concentrations of oil and oleic acid. For
example, two low-oil lines also contained the ASKC28IB1 G45
haplotype. In selected 71 maize lines, lines with F469 showed a
significantly higher embryo oil and oleic-acid content than those
lines without F469 (Fig. 4d). This result is consistent with our earlier
conclusions that F469 is responsible for increasing corn oil and oleic-
Positional cloning of the oil qHO6 provides insights into the
molecular basis of natural and human selection of kernel oil and
oleic-acid contents in maize, and opportunities for uncoupling oil
from other agronomic traits by selecting high-oil DGAT alleles
directly. In addition, identification of a DGAT allele responsible for
the qHO6 phenotype raises questions regarding how this oil pathway
enzyme controls seed oil and oleic acid content, and whether further
increases in these two traits can be made by engineering of DGAT in
Oil, fatty acid and TAG analyses. For seed oil, we used air-dried seeds for
direct NMR measurements as described previously19. For embryo oil, we
soaked seeds in water overnight. Embryos were then dissected from endo-
sperms, freeze-dried and subjected to NMR analysis. We determined fatty acid
profile analysis as described previously18. For yeast TAG analysis, a 20-ml yeast
culture was grown for 54 h and harvested. Lipases were inactivated by heating
at 80 1C for 10 min in the presence of 1-ml isopropanol. We lysed cells by
vortexing in the presence of 0.5-ml glass beads and 0.5 ml of 0.15 M acetic acid.
As an internal control, 50 mg of TAG (C17:0) was added to each sample. Total
lipids were extracted with 3 ml extraction buffer (chloroform:methanol ¼ 2:1),
and washed with 1.5 ml of chloroform. The organic phase was combined and
dried under a nitrogen stream, and then dissolved in 100-ml chloroform.
We loaded one-half of each sample (50 ml) onto TLC plates, and developed
them with hexane/ether/acetic acid (70:30:1) followed by a brief stain with
iodine vapor. The corresponding bands were scraped off for GC analysis
similarly to the protocol described above. We calculated total TAG contents
on the basis of the sum of experimental values for C16:0, C16:1, C18:0, and
C18:1, and expressed as mg of TAG in 1 OD (optical density) cells (approxi-
mately 2 ? 107cells).
Genotyping and QTL mapping. We developed SNP markers unique to this
project from Pioneer proprietary sequences or from public BAC ends from the
inbred B73. DNA fragments were amplified from ASKC28IB1 or PH09B
parents to identify SNPs. We collected eight leaf disks from each seedling
Oleic acid (%)
Embryo oil (%)
P = 9.9 × 10–7
P = 5.8 × 10–7
Figure 4 Plant type I DGAT phylogeny and the association of F469 with oil and oleic-acid content in maize. (a) Phylogenic analysis using PHYLIP package.
(b) Amino acid sequence around F469 from plant species. Accession numbers for published sequences are shown. The teosinte sequence is based on the
consensus of 46 accessions, with all accessions containing F469. Maize DGAT1-2 (EU039830); wheat DGAT1-1 (TaDGAT1-1) assembled from public EST
clones (accession codes listed in Methods). (c) Distribution of F469 positive and F469-negative haplotypes in maize lines. (d) Association of F469 with oil
and oleic-acid content. Data are shown as mean ± s.d.
370VOLUME 40 [ NUMBER 3 [ MARCH 2008 NATURE GENETICS
© 2008 Nature Publishing Group http://www.nature.com/naturegenetics
(4–6 weeks) in 96-well plates. DNA was extracted using a robotic system. SSR
genotyping was done at Pioneer. SNP genotyping were done by Genaissance
Pharmaceuticals or at Pioneer.
Mapping populations were developed as shown in Supplementary Figure 4
online. PH09B is the recurrent parent. Composite interval mapping of embryo
or seed oil traits for BC2populations was done using the Windows QTL
Cartographer program (see URLs section below). The QTL significance level
for each trait was determined by 300 permutations at P o 0.05. For BC4S2
and later generations, we used oleic data to determine the presence of
qHO6, because it could more reliably differentiate the three classes of indivi-
duals (homozygous for ASKC28IB1, PH09B and heterozygous), followed
by confirmation by oil trait data. Map positions used throughout this
paper were based on the proprietary Pioneer High Density (PHD) map,
version 1.2 (2006).
Vector construction and maize transformation. Standard restriction fragment
preparation and ligation techniques were used to position each ZM-DGAT1-2
coding sequence (ASK and PHO9B) behind the embryo-preferred promoter
from the 16-kDa oleosin gene of maize (GenBank no. BD235503, including the
81-bp 5¢-untranslated region of Oleosin, U13701). Similarly, each gene cassette
included a 277-bp fragment of the NOS (nopaline synthase) gene from
Agrobacterium tumefaciens as transcriptional termination/polyadenylation
signal sequence. Each completed gene cassette was flanked by Gateway
(Invitrogen) homologous recombination sites ATT L1 and ATT L2. These
were used to mobilize the DGAT gene expression cassettes into Gateway-
modified pSB11-derived T-DNA vectors (Japan Tobacco)20. These T-DNA
vectors contained both a selectable marker (a Ubi::moPAT::PinII expression
cassette consisting of the maize ubiquitin-1 promoter (Ubi, including the
5¢-untranslated region and first intron21), a maize-optimized PAT gene (US
patent no. 6,096,947) and potato PINII terminator22and a screenable marker,
the DS-RED2 gene (Clontech), under the control of the aleurone-specific
LTP2 promoter (US patent no. 5,525,716) and potato PINII terminator).
We transformed each confirmed T-DNA vector via electroporation into
Agrobacterium tumefaciens LBA4404 (pSB1) cells and confirmed the resulting
cointegrate plasmid by extensive restriction digest analysis. We introduced
constructs into the maize Hi-II line using the Agrobacterium-mediated trans-
formation method as described23. T0plants were crossed with nontransgenic
inbred lines to produce T1seeds.
DGATactivity assays. We purchased a Saccharomyces cerevisiae DGA1 (DGAT)
deletion strain (Clone ID: 12501) from Invitrogen and used it to create a
double mutant (dga1D lro1D) by removing the LRO1 (PDAT) gene using
homologous recombination. The double mutant is deficient in oil biosynthesis.
Maize DGAT1-2 genes were cloned into pRS426 (Stratagene) containing a
1.0-kb promoter and a 0.5-kb terminator from S. cerevisiae 3-phosphoglycerate
kinase (PGK1). The resulting plasmids and a vector-only control were used to
transform the yeast double mutant.
Microsomal protein preparation was as described with minor modifica-
tions24. Yeast cultures were grown to early stationary phase in SC media minus
uracil. Following harvest, the yeast pellets were resuspended in 4 ml of 20 mM
Tris-HCl pH 8, 10 mM MgCl2, 1 mM EDTA, 5% glycerol, 1 mM DTT, and
0.3 M (NH4)2SO4. We added 2 ml of glass beads and lysed the cells by vortexing
for 5 min. The lysate was centrifuged for 15 min at 1,500 g at 6 1C. The
supernatant was then centrifuged at 100,000g for 1.5 h at 6 1C. The microsomal
pellet was resuspended in 500 ml of 100 mM potassium phosphate pH 7.2
containing 10% glycerol and frozen in liquid nitrogen before storage at –80 1C.
We determined protein concentrations using the Bradford method, with the
Coomassie Plus reagent (Pierce) and bovine serum albumin as a standard.
DGAT assays were conducted for 1 min at 25 1C, with 50 mM potassium
phosphate pH 7.2, 10 mM 1-14C-labeled oleoyl–coenzyme A (50 mCi/mmol,
Perkin Elmer), and 20 mg of microsomal protein in a total reaction volume of
100 ml. The reaction was started by adding the microsomal protein. We stopped
the assay and extracted the lipids with 2 ml of hexane:isopropanol (3:2)25
containing 4 ml of unlabeled triacylglycerol (triolein, Sigma). Following
vortexing for 10 s, we separated the phases by adding 1 ml of 500 mM sodium
sulfate, and vortexed again for 10 s. After 10 m, the upper phase was transferred
to another tube and dried with nitrogen gas. The lipid was resolubilized in a
small volume of hexane (approximately 100 to 150 ml) and applied to K6 silica
TLC plates, which were developed in 80:20:1 (v/v) hexane:diethylether:acetic
acid. Triacylglycerols were visualized and marked by staining in iodine vapor.
After the stain faded, the triacylglycerol was scraped, and radioactivity was
determined by liquid scintillation counting.
Antibody and protein analysis. Rabbit polyclonal antibodies were prepared
against the peptide RLRRAPSADAGDLAGD, corresponding to residues 27 to
42 of maize DGAT 1-2. This peptide sequence was identical in the ASKC28IB1
and PHO9B alleles. The antibodies were affinity purified against the same
peptide using an AminoLink immobilization kit (Pierce). SDS-PAGE was run
with samples containing 20 mg of yeast microsomal protein, using 4–20% Tris-
Hepes Precise protein gels (Pierce). Proteins were blotted to nitrocellulose with
a semidry transfer apparatus, using a transfer buffer consisting of 25 mM Tris,
192 mM glycine, 20% methanol, and 0.1% SDS. The blot was probed with the
rabbit antibodies to DGAT, followed by a goat anti-rabbit IgG-HRP conjugate
(BioRad). The blot was developed with a chemiluminescent method using ECL
reagents (GE Healthcare).
Phylogeny analysis. The tree is a consensus of 1,000 bootstrap replicates using
the PHYLIP package (see URLs section below) and the neighbor program.
Branch lengths are scaled to evolutionary distance and were calculated using the
protML program (protein maximum likelihood). Branches with o50% sup-
port were collapsed; most of the branches had 490% support.
URLs. Windows QTL Cartographer program, http://statgen.ncsu.edu/qtlcart/
WQTLCart.htm; PHYLIP package, http://evolution.genetics.washington.edu/
Accession numbers. GenBank: maize DGAT1-2, EU039830; wheat DGAT1-1
CJ620038, CK197709 and BJ268713.
Note: Supplementary information is available on the Nature Genetics website.
We thank B. Li for providing the maize physical map and advice on map-based
cloning, S. Zhen for growing corn plants in the greenhouse, C. Li for laboratory
support, J. Hazebroek for fatty acid composition analysis, H. Sullivan for SNP
marker development, the Pioneer molecular marker laboratory for genotyping,
T. Colbert for valuable discussion and D. Selinger and M. Ayele for
bioinformatics support. We are grateful to B. Hitz for his support and
R. Jung for critical review of the manuscript.
B.S. and M.C.T. conceived and directed the project; M.E.W., G.Y.Z., P.Z. and J.L.
conducted oil QTL mapping; P.Z. conducted fine mapping and cloning; W.B.A.
performed the oil analysis; K.R. and S.Z. performed DGAT activity assay and
immunoblot; K.G. conducted vector construction; J.R. conducted corn
transformation; D.N. and W.S. conducted field experiments; D.B. completed
genotyping; V.L. and S.D. conducted BAC sequencing and DGAT resequencing;
B.S., P.Z., K.R. and M.C.T. analyzed the data and wrote the manuscript.
Published online at http://www.nature.com/naturegenetics
Reprints and permissions information is available online at http://npg.nature.com/
1. Berke, T.G. & Rocheford, T.R. Quantitative trait loci for flowering, plant and ear height
and kernel traits in maize. Crop Sci. 35, 1542–1549 (1995).
2. Song, X.-F., Song, T.-M., Dai, J.-R. & Rocheford, T.R. QTL mapping of kernel oil
concentration with high-oil maize by SSR markers. Maydica 49, 41–48 (2004).
3. Mangolin, C.A. et al. Mapping QTLs for kernel oil content in a tropical maize
population. Euphytica 137, 251–259 (2004).
4. Laurie, C.C. et al. The genetic architecture of response to long-term artificial selection
for oil concentration in the maize kernel. Genetics 168, 2141–2155 (2004).
5. Clark, D., Dudley, J.W., Rocheford, T.R. & LeDeaux, J.R. Genetic analysis of corn kernel
chemical composition in the random mated 10 generation of the cross of generations
70 of IHO ? ILO. Crop Sci. 46, 807–819 (2006).
6. White, P.J. & Weber, E.J. in Corn: Chemistry and Technology 2nd edn, Vol. 10 (eds.
White, P.J. & Johnson, L.A.) Lipids of the kernel, 355–395 (American Association of
Cereal Chemists, Inc., St. Paul, Minnesota, 2003).
7. Kinney, A.J., Cahoon, E.B. & Hitz, W.D. Manipulating desaturase activities in trans-
genic crop plants. Biochem. Soc. Trans. 30, 1099–1103 (2002).
NATURE GENETICS VOLUME 40 [ NUMBER 3 [ MARCH 2008 371
© 2008 Nature Publishing Group http://www.nature.com/naturegenetics
8. Moose, S.P., Dudley, J.W. & Rocheford, T.R. Maize selection passes the century mark: a
unique resource for twenty-first century genomics. Trends Plant Sci. 9, 358–364
9. Lambert, R.J., Alexander, D.E. & Mejaya, I.J. in Plant Breeding Reviews Part 1, Vol. 24
(ed. Janick, J.) Single kernel selection for increased grain oil in maize synthetics and
high-oil hybrid development, 153–175 (John Wiley & Sons, Inc., Hoboken, New Jersey,
10. Poneleit, C.G. & Alexander, D.E. Inheritance of linoleic and oleic acids in maize.
Science 147, 1585–1586 (1965).
11. Alrefai, R., Berke, T.G. & Rocheford, T.R. Quantitative trait locus analysis of fatty acid
concentrations in maize. Genome 38, 894–901 (1995).
12. Jako, C. et al. Seed-specific over-expression of an Arabidopsis cDNA encoding a
diacylglycerol acyltransferase enhances seed oil content and seed weight. Plant
Physiol. 126, 861–874 (2001).
13. Lung, S.-C. & Weselake, R.J. Diacylglycerol acyltransferase: a key mediator of plant
triacylglycerol synthesis. Lipids 41, 1073–1088 (2006).
14. Grisart, B. et al. Genetic and functional confirmation of the causality of the DGAT1
K232A quantitative trait nucleotide in affecting milk yield and composition. Proc.
Natl. Acad. Sci. USA 101, 2398–2403 (2004).
15. Krogh, A., Larsson, B., von Heijne, G. & Sonnhammer, E.L.L. Predicting transmem-
brane protein topology with a hidden Markov model: application to complete genomes.
J. Mol. Biol. 305, 567–580 (2001).
16. Ohlrogge, J. & Browse, J. Lipid biosynthesis. Plant Cell 7, 957–970 (1995).
17. Voelker, T. & Kinney, A.J. Variations in the biosynthesis of seed-storage lipids. Annu.
Rev. Plant Physiol. Plant Mol. Biol. 52, 335–361 (2001).
18. Belo ´, A. et al. Whole genome scan detects an allelic variant of fad2 associated with
increased oleic acid levels in maize. Mol. Genet. Genomics 279, 1–10 (2008).
19. Shen, B., Sinkevicius, K.W., Selinger, D.A. & Tarczynski, M.C. The homeobox gene
GLABRA2 affects seed oil content in Arabidopsis. Plant Mol. Biol. 60, 377–387
20. Ishida, Y. et al. High efficiency transformation of maize (Zea mays L.) mediated by
Agrobacterium tumefaciens. Nat. Biotechnol. 14, 745–750 (1996).
21. Christensen, A.H., Sharrock, R.A. & Quail, P.H. Maize polyubiquitin genes: structure,
thermal perturbation of expression and transcript splicing, and promoter activity
following transfer to protoplasts by electroporation. Plant Mol. Biol. 18, 675–689
22. An, G. et al. Functional analysis of the 3¢ control region of the potato wound-inducible
proteinase inhibitor II gene. Plant Cell 1, 115–122 (1989).
23. Zhao, Z.Y. & Ranch, J. Transformation of maize via Agrobacterium tumefaciens
using a binary co-integrate vector system. Methods Mol. Biol. 318, 315–323
24. Milcamps, A. et al. Isolation of a gene encoding a 1,2-diacylglycerol-sn-acetyl-CoA
acetyltransferase from developing seeds of Euonymus alatus. J. Biol. Chem. 280,
25. Hara, A. & Radin, N.S. Lipid extraction of tissues with a low-toxicity solvent. Anal.
Biochem. 90, 420–426 (1978).
372VOLUME 40 [ NUMBER 3 [ MARCH 2008 NATURE GENETICS
© 2008 Nature Publishing Group http://www.nature.com/naturegenetics