| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a formerly University of Illinois, Dep. of Crop Sciences, 1101 S. Goodwin Ave., Urbana, IL 61801
b University of Delaware, Dep. of Chemical Engineering, Newark, DE 19716
c University of Illinois, Dep. of Crop Sciences, 1101 S. Goodwin Ave., Urbana, IL 61801
* Corresponding author (mewassom{at}sio.midco.net).
 |
ABSTRACT |
|---|
 TOPNOTES ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
0.01) and negative correlations between oil and linoleic acid (rp = –0.46**), and between oleic and linoleic acids (rp = –0.99**). Multiple regression models with QTL detected by composite interval mapping (CIM) on a genetic map with length = 1486 cM, explained 15.4, 41.6, 51.0, 59.6, and 47.9% of the phenotypic variation for palmitic, stearic, oleic, linoleic, and linolenic acids, respectively. A 6-cM interval on chromosome 6 (bin 6.04) includes QTL for stearic, oleic, linoleic, and linolenic acids, explains 10.9 to 39.6% of the variation for these fatty acids, and is 10 to 16 cM from QTL for oil. Another region on chromosome 6 (bin 6.01) includes QTL for oleic and linoleic acids and was epistatic with the QTL in bin 6.04. One or both of these two QTL regions on chromosome 6 may be responsible for fatty acid variation previously attributed to linoleic acid1.
Abbreviations: BC, Backcross • BC1S1s, Backcross1-derived S1 lines • CIM, Composite interval mapping • HOC, High-oil corn or maize • IHO, Illinois High Oil • ILO, Illinois Low Oil • ILO(EM), ILO-early maturity • H2, Broad-sense heritability • LOD, Likelihood of odds • 
, Proportion of genotypic variance explained • QTL, Quantitative trait loci • rp, phenotypic correlation • rg, genotypic correlation • R2adj, Coefficient of determination with adjustment for the number of terms in a model
|
ACKNOWLEDGMENTS |
|---|
|
NOTES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Received for publication August 4, 2007.
a formerly University of Illinois, Dep. of Crop Sciences, 1101 S. Goodwin Ave., Urbana, IL 61801
b University of Delaware, Dep. of Chemical Engineering, Newark, DE 19716
c University of Illinois, Dep. of Crop Sciences, 1101 S. Goodwin Ave., Urbana, IL 61801
* Corresponding author (mewassom{at}sio.midco.net).
Maize (Zea mays L.) produces high-quality oil valued for oxidative stability and low concentrations of saturated fatty acids. The nutritional value of maize oil could be improved by increasing the concentration of oleic acid, a "heart-friendly" monounsaturated fatty acid. To identify quantitative trait loci (QTL) for the major fatty acids constituting oil from maize kernels, we produced 150 backcross1-derived S1 (BC1S1) lines from donor parent, Illinois High Oil (IHO), and recurrent parent, B73. There was a positive phenotypic correlation between oil and oleic acid (rp = 0.47**, 0.01) and negative correlations between oil and linoleic acid (rp = –0.46**), and between oleic and linoleic acids (rp = –0.99**). Multiple regression models with QTL detected by composite interval mapping (CIM) on a genetic map with length = 1486 cM, explained 15.4, 41.6, 51.0, 59.6, and 47.9% of the phenotypic variation for palmitic, stearic, oleic, linoleic, and linolenic acids, respectively. A 6-cM interval on chromosome 6 (bin 6.04) includes QTL for stearic, oleic, linoleic, and linolenic acids, explains 10.9 to 39.6% of the variation for these fatty acids, and is 10 to 16 cM from QTL for oil. Another region on chromosome 6 (bin 6.01) includes QTL for oleic and linoleic acids and was epistatic with the QTL in bin 6.04. One or both of these two QTL regions on chromosome 6 may be responsible for fatty acid variation previously attributed to linoleic acid1.
Abbreviations: BC, Backcross • BC1S1s, Backcross1-derived S1 lines • CIM, Composite interval mapping • HOC, High-oil corn or maize • IHO, Illinois High Oil • ILO, Illinois Low Oil • ILO(EM), ILO-early maturity • H2, Broad-sense heritability • LOD, Likelihood of odds • , Proportion of genotypic variance explained • QTL, Quantitative trait loci • rp, phenotypic correlation • rg, genotypic correlation • R2adj, Coefficient of determination with adjustment for the number of terms in a model
|
INTRODUCTION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Maize fatty acid composition is variable and heritable. The major fatty acids of maize oil are palmitic (16:0), stearic (18:0), oleic (18:1), linoleic (18:2), and linolenic (18:3). Oil from conventional maize hybrids has about 110 mg palmitic, 20 mg stearic, 240 mg oleic, 620 mg linoleic, and 7 mg linolenic acid g–1 oil (Lambert, 2001). These proportions varied among 418 commercial maize hybrids and 98 inbred lines, with ranges for palmitic, 67 to 165; stearic, 7 to 66; oleic, 162 to 438; linoleic, 395 to 695; and linolenic, 0.0 to 37 mg fatty acid g–1 total fatty acids. (Dunlap et al., 1995a). Similar variation was found in Central American land races and other exotic materials (Dunlap et al., 1995b, Cheesbrough et al., 1997). Broad-sense heritabilities (H2) for palmitic, stearic, oleic, linoleic, and linolenic acids were estimated to be 39, 72, 85, 83, and 67%, respectively, in an IHO x Illinois Low Oil Early Maturity [ILO(EM)] population (Alrefai et al., 1995).
Genomic regions influencing oleic and linoleic acid concentrations have been located, using monosomic, trisomic, and translocation stocks, to chromosome arms 1S (Widstrom and Jellum, 1984), 2 (Plewa and Weber, 1975), 4L (Widstrom and Jellum, 1984), and 5L (Shadley and Weber, 1980; Widstrom and Jellum, 1984). Oleic acid content1 (olc1), a mutation induced in B73 that caused oleic acid concentration to increase from 27% in normal B73 to 52% in mutants, was mapped to chromosome arm 1L (Wright, 1995). Mikkilineni and Rocheford (2003), using molecular methods, mapped cDNAs of fatty acid desaturation 2 (fad2) and fad6, which encode 
12 fatty acid-dehydrogenases that convert oleic to linoleic acid (Okuley et al., 1994), to bins 1.07, 4.06, 5.06, and 10.03. Linoleic acid1 (ln1) was hypothesized to be a major gene controlling relative levels of oleic and linoleic acids (Poneleit & Alexander, 1965, de la Roche et al., 1971). Poneleit (1976), using translocation and inversion stocks, estimated that ln1 is near yellow endosperm1 which is mapped to bin 6.01 in the Maize Genetics and Genomics Database (MaizeGDB, 2007).
Quantitative trait loci for palmitic, stearic, oleic, linoleic, and linolenic acid concentrations were identified in 12 clusters on 10 chromosomes in IHO x ILO(EM) (Alrefai et al., 1995). IHO and ILO(EM) differed greatly for kernel composition traits, aiding detection of genetic variation for oil concentration. But IHO and ILO(EM) are not representative of modern maize hybrids or breeding lines. To perform an oil composition QTL study in a population more genetically relevant to practical plant breeding, we analyzed fatty acids in a backcross population developed from a high-oil donor and a production-oriented recurrent parent with normal kernel composition. Our objectives were to analyze variation for palmitic, stearic, oleic, linoleic, and linolenic fatty acids of maize kernel oil; and to detect, determine, and model the effects of QTL associated with concentrations of these fatty acids.
|
MATERIALS AND METHODS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Field Technique
Field experiments were conducted at the University of Illinois Crop Sciences Research and Education Center at Urbana, IL. The BC1S1s were grown in 1993 and 1994 in one-row plots 4.6 m long and 760 mm apart in an irrigated nursery with two replications of each BC1S1, B73, and IHO90 population (random plants of the IHO cycle 90 population) in an -incomplete block design. Mean daily low and high temperatures during the approximate period of kernel development, 15 July through 15 Sept. were 17 and 27°C in 1993 and 15 and 27°C in 1994. Daily climate data are available from the Illinois State Climatologist Office (2007).
Plots were thinned to 15 plants row–1 (equivalent to 43000 plants ha–1). Plants were self-pollinated by hand using shoot and tassel bags. The low population density facilitated hand pollination. At maturity ears were individually harvested, kernels were removed from the centers of individual ears, and kernel samples from five to seven ears from each plot of each BC1S1 line were combined to form balanced bulk samples for analysis of oil concentration, as reported in Wassom et al. (2008).
Fatty Acids Analysis
Fatty acid composition was measured at Pioneer Hi-Bred International Inc. (Johnston, IA, USA) by Jan Hazebroek and associates (Hazebroek, 1997). Liquid was mechanically pressed at 4.1 mPa from two random kernels of each BC1S1 plot, oil was extracted in hexane, and the fatty acids were trans-methylated. Fatty acid methyl esters were separated by capillary gas chroma- tography with flame ionization detection. The integrated areas of the peaks corresponding to C16 and C18 methyl esters were grouped, normalized, and expressed as concentrations of the individual fatty acids in total fatty acid methyl esters (mg fatty acid g–1 total fatty acids).
Oil Concentration
Oil concentration (mg oil g–1 kernel mass) was reported in Wassom et al. (2008) and was measured on bulk samples from field plots (Berke & Rocheford, 1995) using near-infrared light reflectance (Dudley and Lambert, 1992).
Statistical Analysis
Statistical analyses of traits were performed using SAS (SAS, 1999; SAS Institute, Cary, NC) and PLABSTAT computer programs (Utz, 2005; University of Hohenheim, Stuttgart, Germany).
Homogeneity of variance was tested by the SAS GLM procedure using the hovtest option and the Bartlett and Levene-Welch tests. Heterogeneity of error variance among BC1S1 lines for each of the fatty acids was significant. Log or arc sin transformations did not correct the heterogeneity, so we proceeded with the analyses using the unconverted data with the understanding that the power of some tests might be reduced.
The PLABSTAT LATTICE and ANOVA programs were used for analysis of variance, correlation, and estimation of variance components and H2. A PLABSTAT LATTICE analysis was performed within years and ANOVA over years, treating years and lines as random variables. The PLABSTAT LATTICE analysis estimated variance components within years using an intra-block analysis, and PLABSTAT ANOVA estimated variance components in the combined analysis over years. Treatment means were adjusted for block effects (Cochran and Cox, 1957). Broad-sense heritability (H2) and 90% confidence intervals were calculated by PLABSTAT on an entry-mean basis using mean squares from the ANOVA and the F statistic as described by Knapp et al. (1985). Because it is not possible to separate additive and dominance variance in the backcross design, the heritability is broad sense. Variance components were also obtained with the SAS VARCOMP procedure using the MIVQUE0 method to verify the PLABSTAT estimates. The PLABSTAT ANOVA program used block-adjusted line means by year to calculate phenotypic Pearson correlation coefficients. Genotypic correlations (rg) were calculated by PLABSTAT ANOVA in an analysis of covariance as described by Baker (1986).
Molecular Marker Analysis
DNA was extracted from a bulk sample of fresh leaf tissue from 20 to 30 seedlings of each BC1S1 line by a cetyl trimethylammonium bromide (CTAB) method (Doyle and Doyle, 1990). Molecular marker profiles of the BC1S1s were produced using RFLP (Goldman et al., 1994) and simple sequence repeat (SSR) PCR markers (Senior et al., 1996) as described in Wong et al. (2003). The RFLP and SSR markers are described in the Maize Genetics and Genomics Database (MaizeGDB) (2007) except fatty acid desaturase-6 cDNA clones, Cfbbi58 and Cqrak57, which were used as RFLP probes (Mikkilineni and Rocheford, 2003) and were obtained from Pioneer Hi-Bred International, Inc. (Johnston, IA). Primers for SSR PCR were obtained from Research Genetics, Inc. (ResGen, Inc., Huntsville, AL).
A linkage map was constructed (Wassom et al., 2008) using the JoinMap Version 3 computer program (Van Ooijen and Voorrips, 2001; Plant Research International, Wagenignen, the Netherlands). The linkage map was constructed using Haldane's mapping function and 110 molecular markers (38 RFLP and 72 SSR), and had length equal to 1486 cM, average distance between markers equal to 14.9 cM, and 96.8% of the genome within 20 cM of a marker.
Composite interval mapping of QTL (Haley and Knott, 1992) was performed with the PLABQTL computer program (Utz and Melchinger, 2006; University of Hohenheim, Stuttgart, Germany.) The CIM model is yi = µ + bxi + 
bkxik + eI; with yi the phenotypic value of backcross line i for trait y, xi the coded genotype variable at a putative QTL for i, µ the mean, b the effect of the QTL on the trait, and ei, the residual variation in i (Zeng et al., 1999). Models for regression of kernel traits on QTL included additive and additive x additive digenic epistatic effects. The initial detection of QTL with CIM was performed with a LOD threshold of 2.5.
The proportion of phenotypic variance explained by all the QTL in the multiple regression models with adjustment for the number of terms in the models, the adjusted R2 (R2adj), was calculated as described by Hospital et al. (1997): R2adj = R2– [z/(n – z – 1)](1 – R2). R2 is the coefficient of determination, z is the number of QTL and interaction terms, and n is the number of individuals. The proportion of genotypic variance explained by all QTL in the models, , was calculated according to Utz et al. (2000): = R2adj/H2. Effects were declared significant if p 0.05.
Cofactors for CIM were selected by PLABQTL in stepwise regression, adding cofactors to the regression model with F to enter = 3.5. Cofactor sets were empirically evaluated and modified to maximize the R2adj. Cross validation with the data set randomly divided among lines in 200 five-fold detection and validation runs was employed to check that QTL effects were not falsely detected due to bias in the data set (Utz et al., 2000). The QTL analyses for palmitic and linolenic acids each excluded one BC1S1 line because, on the basis of the studentized residual or the ANDREWS-PREGIBON statistic second factor, PLABQTL indicated they had an extreme influence on the analysis (Draper and John, 1981).
Detection of QTL in CIM was performed with a like- lihood of odds (LOD) threshold of 2.5. For model development the QTL included in the regression models were limited to those detected with LOD thresholds equivalent to an = 0.05 genome-wide error rate, similar to the approach of Cassady et al. (2001). The LOD thresholds for = 0.05 were determined by testing 1000 permutations of the data (Churchill and Doerge, 1994) and for each trait were: palmitic acid, 3.16; stearic acid, 3.05; oleic acid, 2.94; linoleic acid, 3.00; and linolenic acid, 3.00.
Regression models for kernel traits on QTL were based on additive and additive x additive epistatic effects. The direction of QTL effects was defined relative to the IHO90 allele. That is, if the sign of a QTL effect is positive, then the IHO90 allele is associated with higher levels of the trait and if a QTL effect is negative, the B73 allele is associated with higher levels of the trait.
|
RESULTS AND DISCUSSION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
0.05) among (IHO90 x B73) x B73 BC1S1 lines,B73, and IHO90 population (Table 1
). Palmitic and stearic, but not oleic, linoleic, and linolenic acids were significantly affected by years. Oleic and linoleic acids combined (oleic+linoleic) also varied significantly among lines and years. Only palmitic acid was significantly affected by the interaction of line with year. Oleic and linoleic acid were found in a wider range of concentrations among lines than the other fatty acids. But the range for oleic+linoleic was somewhat smaller than the ranges of oleic or linoleic individually. B73 and IHO90 population differed significantly for each fatty acid except stearic, and oleic+linoleic. Differences between parents and BC1S1s were significant at both the low and high ends of the range among BC1S1s for oleic acid, at the high end of the ranges for stearic, linoleic, and linolenic acids, and at the low end for oleic+linoleic combined.
|
2g) for oleic (1403.3) and linoleic (1514.5) acids were much greater than for the other fatty acids (1.0 to 8.1) (Table 1). Jellum and Worthington (1966) also reported greater 2g for oleic and linoleic acids than for palmitic, stearic, or linolenic acids. The 2g for oleic+linoleic acid (23.5) was much less than for either oleic or linoleic alone, but was larger than for the other fatty acids (1.0 to 8.1). PLABSTAT produced negative estimates of the genotype x environment variance (2ge) of oleic (–250.7), linoleic (–260.1), and linolenic (–0.03) acids. Estimates of 2ge were also performed using SAS VARCOMP, which produced similar 2ge estimates for oleic, –225.0; linoleic, –238.6; and linolenic acid, –0.03. Heritabilities ranged from 60% for palmitic to 86% for both oleic and linoleic acids (Table 1).
Phenotypic correlations among the 18-carbon fatty acids were all significant, but correlations between palmitic acid, a 16-carbon fatty acid, and each of the 18-carbon fatty acids were all nonsignificant, except the correlation with linoleic acid, which was significant, but weak (rp = –0.17*, 0.05; rg = –0.24). The relatively weak correlations of palmitic with the 18-carbon fatty acids might result from the divergence of palmitic acid from biochemical pathways involving the 18-carbon fatty acids on export from plastids. The 18-carbon fatty acids, however, share a common step-by-step desaturation pathway from stearic (18:0) to oleic (18:1), to linoleic (18:2), and to linolenic (18:3). There was a relatively strong correlation between palmitic and oleic+linoleic acids (rp = –0.65**, rg = –0.56). This might be explained by the predominance of oleic and linoleic acids among 18-carbon fatty acids, the stepwise lengthening of acyl chains by two-carbon units in plastids, and the cessation of further elongation after export to the cytoplasm where assembly into triglycerides occurs (Voelker and Kinney, 2001).
There was a near one-to-one negative correlation between oleic and linoleic acids (rp = –0.99**). This is similar to previous reports by Cheesbrough et al. (1997) (rp = –0.95*), Alrefai et al. (1995) (rp = –0.97**), and Pamin et al. (1986) (rp = –0.96**). The genetic correlation between oleic and linoleic was also negative and near unity (rg = –0.99**), indicating that most or all of the genes responsible for variation in the accumulation of oleic and linoleic acids are closely linked or pleiotropic.
Palmitic, stearic, or oleic acids were positively correlated with kernel oil concentration (mg oil g–1 kernel mass) (rp = 0.13ns, rg = –0.20; rp = 0.27**, rg = –0.26; and rp = 0.47**, 0.55, respectively), but linoleic and linolenic acids were negatively correlated with oil (rp = –0.46**, rg = –0.53; rp = –0.65**, rg = –0.79) (Table 2 ). Other studies, representing several genetic backgrounds, also reported positive correlations of oil with oleic acid and negative correlations with linoleic or linolenic acid. Oil concentration as a proportion of kernel mass and oleic acid concentration as a proportion of total fatty acids were positively correlated in IHO x ILO(EM) (rp = 0.46**) (Alrefai et al., 1995) and in a Krug variety-related population (rp = 0.51**) (Pamin et al., 1986). Oil was negatively correlated with linoleic (rp = –0.48**) and linolenic acid (rp = –0.45** in the Krug population (Pamin et al., 1986).
|
Quantitative Trait Loci
Quantitative trait loci were detected for each of the fatty acids (Table 3
). Three QTL influencing palmitic acid were detected at LOD 2.5 on chromosomes 2 and 7. Two, both on chromosome 7, were included in the regression model, which explained 15.4% of the phenotypic (R2adj) and 25% of the genotypic variation (). The models for the other fatty acids explained more variation, with R2adj ranging from 41.6% for the stearic acid model to 59.6% for the linoleic acid model, and ranging from 54.4% for stearic acid to 69.8% for linoleic acid.
|
Because of their proximity and the direction of the allelic effects, some QTL detected for individual fatty acids appear to indicate single genes affecting amounts of more than one fatty acid (Table 3). For example, linoleic and linolenic acid QTL on chromosome 1 at positions 4 and 16 had higher concentrations of linoleic and linolenic acids with IHO90 alleles. And an oleic acid QTL on chromosome 1 at position 2, which was detected at LOD 2.5 but not included in the model because the LOD was too low, had higher oleic acid concentration with the B73 allele. Because linoleic and linolenic acids were positively correlated (Table 2) and oleic acid was negatively correlated with both linoleic and linolenic acids, it is logical that a common QTL in this region on chromosome 1 from positions 2 to 16 affected the three fatty acids. Oleic and linoleic acid QTL on chromosome 6 at position 20 had higher concentrations of oleic acid with the IHO90 allele and higher linoleic acid with the B73 allele. The opposite direction of the oleic and linoleic acid QTL effects and the negative correlation of oleic and linoleic acids are consistent with this being a common QTL affecting both fatty acids. The 6-cM interval on chromosome 6 from positions 48 to 54, which included QTL for stearic, oleic, linoleic, and linolenic acids, might be a single QTL affecting the four fatty acids. The individual QTL in this interval had positive effects for stearic and oleic acids, but negative effects for linoleic and linolenic acids, which is consistent with their correlations. Therefore, directional effects of the QTL in these intervals are consistent with the existence of common QTL for multiple fatty acids.
We detected QTL for fatty acids at or near some of the markers that Alrefai et al. (1995) found significantly affected fatty acid concentrations in IHO x ILO(EM). These are listed by chromosome bin (MaizeGDB, 2007) in Table 4 and include QTL in five bins on four chromosomes. Our (IHO90 x B73) x B73 BC1S1 analysis shows that IHO alleles of fatty acids QTL also affect fatty acids in a B73 genetic background. Two of the bins (3.04 and 6.04) with fatty acid QTL detected in both (IHO90 x B73) x B73 and IHO x ILO(EM) also have oil QTL for both populations (Wassom et al., 2008; Berke and Rocheford, 1995).
|
A QTL on chromosome 6 in the interval from position 48 to 54, on the basis of LOD and partial R2, could be the most influential QTL associated with stearic, oleic, and linoleic acids. And a nearby QTL on chromosome 6 at position 20 was included in the oleic and linoleic acid models. The oleic model (Table 3) included an epistatic interaction term involving the chromosome 6 QTL at positions 20 and 50, and the linoleic model included three epistatic interactions including one or both of the chromosome 6 QTL: Chromosome 6 positions 20 x 50, Chromosome 6 position 20 x Chromosome 1 position 4, and Chromosome 6 position 50 x Chromosome 1 position 4. Thus, our oleic and linoleic acid models indicate a strong influence by two QTL from this region of chromosome 6.
One of the markers flanking QTL for oleic and linoleic acid concentrations on chromosome 6 at position 20 is y1ssr, a microsatellite 11 base pairs upstream of the yellow endosperm1 (y1) start codon (MaizeGDB, 2007). The region of chromosome 6 linked to y1 has been considered to include ln1, a putative gene strongly influencing concentrations of oleic and linoleic acids (Poneleit and Alexander, 1965; de la Roche et al., 1971; Poneleit, 1976). On the basis of an analysis using translocation and inversion stocks, Poneleit (1976) predicted that ln1 was near y1. But because these stocks affect large chromosome segments, Poneleit may have been observing effects from more than one gene in this region of chromosome 6. Alrefai et al. (1995), using single-factor analysis, found marker umc65a of chromosome 6 was significantly associated with variation for palmitic, stearic, oleic, linoleic, and linolenic acid concentrations and suggested it might be linked to ln1. Marker umc65a is a flanking marker for our oleic acid QTL on chromosome 6 at position 50. Therefore, the QTL on chromosome 6 at position 20 and in the interval from 48 to 54, and their epistatic interaction could collectively explain the effects of the putative ln1 gene.
Three of the intervals with QTL for fatty acids are at or near QTL affecting oil concentration in this population (Wassom et al., 2008), and two intervals are near oil QTL in IHO x ILO(EM) (Table 4). For example, a stearic acid QTL and an oil QTL are both mapped at position 14 on chromosome 3 (flanking markers bnlg1019a and npi247). The B73 allele of this QTL increased accumulation of stearic acid and the IHO90 allele increased accumulation of oil. The opposite directions of the allelic effects for stearic acid and oil and the weak positive correlation (r = 0.27**) between them imply that the stearic and oil effects are due to separate QTL, though they were mapped to the same position. Stearic, oleic, linoleic, and linolenic acid QTL on chromosome 6 in the interval from 48 to 54 (flanking markers umc1006 and umc65a, and umc65a and nc009) are 10 to 16 cM from an oil QTL at position 64 (flanking markers umc65a and nc009). The QTL in this interval had positive effects on oil and oleic acid, but a negative effect for linoleic acid. This is consistent with the correlations between oil, oleic, and linoleic acid concentrations and the directions of allelic effects expected from a common QTL. A linolenic acid QTL on chromosome 8 at position 90 was increased by the B73 allele, and is at nearly the same position as a QTL for oil on chromosome 8 at position 88, which was increased by the IHO90 allele. Because oil and linolenic acid are negatively correlated, the opposite sign of the effects is consistent with this being a common QTL for oil and linolenic acid.
Some enzymes have been shown by other researchers, using genetic and molecular genetic methods, to affect major seed storage fatty acids. Seeds of Arabidopsis (Arabidopsis thaliana L.), cotton (Gossypium hirsutum L.), or peanut (Arachis hypogaea L.) from plants with reduced fad2 activity due to mutation or RNA-mediated suppression had increased levels of oleic acid (Jung et al., 2000; Liu et al., 2002; Stoutjesdijk et al., 2002). Peanut seeds did not accumulate oleic acid to higher than normal concentrations except with mutations in both of two homologous fad2 genes (Jung et al., 2000). Mikkilineni and Rocheford (2003) mapped fad2 in maize to three loci on three chromosomes and sequenced fad2 clones from an embryo cDNA library. Sequence data indicated two fad2 isoforms in developing maize seeds. If both fad2 isoforms are enzymatically effective, the redundancy might explain why we did not detect QTL in any of the regions where Mikkilineni and Rocheford mapped fad2. β-ketoacyl-acyl carrier protein (ACP) synthase II (KASII) elongates 16:0-ACP to 18:0-ACP and is encoded by fatty acid biosynthesis1 (fab1). Seeds of Arabidopsis from fab1 mutants with reduced KASII activity had higher than normal levels of palmitic acid (Pidkowich et al., 2007). Fab1 is mapped to bin 10.04 in maize (MaizeGDB, 2007), but we detected no QTL in bin 10.04 for palmitic acid or any of the other fatty acids except linolenic acid. Diacylglycerol acyltransferase, encoded by triacylglycerol1 (TAG1), facilitates addition of a fatty acid to diacylglycerol to form triacylglycerol (Ohlrogge and Jaworski, 1997; Voelker and Kinney, 2001; Zou et al., 1999). A TAG1 mutation in Arabidopsis reduced diacylglycerol acyltransferase activity, total oil production, and oleic acid concentration; but increased linoleic and linolenic acid concentrations (Katavic et al., 1995; Jako et al., 2001). This combination of effects is similar to the combination of effects on fatty acids and oil concentrations from the QTL at positions 50, 54, and 64 on chromosome 6 in our maize population. We do not have enough experimental information and precision to determine whether the oil and fatty acid effects in this region are caused by a pleiotropic gene, or by closely linked genes affecting oil and fatty acids separately. Further study will be needed to determine whether variation for oil and fatty acids can be separated in these regions.
|
CONCLUSIONS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
This research was supported by grants from the Consortium for Plant Biotechnology Research (USDA) and Illinois-Missouri Biotechnology Alliance (USDA) with matching support from Pioneer Hi-Bred International, Inc, Cargill Seeds, Inc., and Monsanto Corp. We acknowledge support from the University of Illinois Agricultural Experiment Station and Department of Crop Sciences. We are grateful for assistance in the field and laboratory from Jeff Wong, Jerry Chandler, and many others.
All rights reserved. No part of this periodical may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Permission for printing and for reprinting the material contained herein has been obtained by the publisher.
Received for publication August 4, 2007.
|
REFERENCES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  H.-W. Wang, B.-Y. Wu, T.-M. Song, and S.-J. Chen Effects of Long-Term Selection for Kernel Oil Concentration in KYHO, a High-Oil Maize Population Crop Sci., March 17, 2009; 49(2): 459 - 466. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Vadose Zone Journal | |||
| Journal of Natural Resources and Life Sciences Education |
Soil Science Society of America Journal | ||||
| Journal of Plant Registrations | Journal of Environmental Quality |
The Plant Genome | |||
