| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Dep. of Crop Science, North Carolina State Univ., Box 7620, Raleigh, NC 27695
b USDA-ARS, Plant Science Research, 3127 Ligon St., Raleigh, NC 27607. This project was funded by the United Soybean Board
* Corresponding author (andrea_cardinal{at}ncsu.edu).
 |
ABSTRACT |
|---|
 TOPNOTES ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Abbreviations: 16:0, palmitate fatty acid • 18:0, stearate fatty acid • 18:3, linolenate fatty acid
a Dep. of Crop Science, North Carolina State Univ., Box 7620, Raleigh, NC 27695
b USDA-ARS, Plant Science Research, 3127 Ligon St., Raleigh, NC 27607. This project was funded by the United Soybean Board
* Corresponding author (andrea_cardinal{at}ncsu.edu).
Palmitate is the predominant saturated fatty acid in soybean oil. Major fap alleles that reduce palmitate content in seed oil also reduce seed yield. Breeders are interested in estimating the genotypic correlation between palmitate content and agronomic traits to predict unfavorable correlated responses to selection. The main objective of this study was to estimate the genotypic and phenotypic correlations between palmitate and linolenate contents and other traits in three populations segregating for the fapnc, fap1, and fan alleles and modifier genes. The populations derived from crosses of high-yielding lines and improved low-palmitate and low-linolenate lines were grown in replicated trials in three environments. Significant positive genetic correlations between palmitate and yield and between palmitate and plant height were observed in all three populations. Linolenate content was genetically positively correlated with lodging in two populations and negatively correlated with oil content in three populations. Our results support the observation that the major fapnc or fap1 or both alleles reduced plant height and had a major negative effect on yield. These effects could be due to pleiotropy or linkage with unfavorable yield or height genes. The relative importance of pleiotropy and linkage has very different implications for oil quality breeding.
Abbreviations: 16:0, palmitate fatty acid • 18:0, stearate fatty acid • 18:3, linolenate fatty acid
|
INTRODUCTION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
in et al., 1998; Wilcox and Cavins, 1990; Rahman et al., 1996). Other lines, including N79-2077 (64 g kg–1 16:0) and its derived line, N79-2077-12, were developed by selection of a natural mutation (Burton et al., 1983, 1994b).
Major efforts have been exerted by several scientific groups to perform inheritance studies and allelism tests involving the low-palmitate germplasm mentioned above. The allele conferring low palmitate in C1726 was designated fap1 (Wilcox and Cavins, 1990); in A22, fap3 (Schnebly et al., 1994; Stojin et al., 1998); in N79-2077, fapnc (Burton et al., 1983, 1994b); in ELLP2, fapx (Primomo et al., 2002); and in J3, sop1 (Rahman et al., 1996). Three alleles are at independent loci from each other (fap1, fap3, and fap*) (Schnebly et al., 1994; Stojin et al., 1998; Wilcox et al., 1994; Kinoshita et al., 1998; Primomo, 2000; Primomo et al., 2002). Alleles sop1 and fapnc are independent of the fap1 locus, but their allelic relationship to the other loci is unknown (Kinoshita et al., 1998; Wilcox et al., 1994).
In addition to the major genes, minor effects of modifier genes also influence the palmitate content in soybean oil (Horejsi et al., 1994; Rebetzke et al., 1998b, 2001; Stojin et al., 1998; Li et al., 2002). Heritability estimates for 16:0 content of populations segregating for minor genes ranged from 0.37 to 0.5 on a plot-mean basis and from 0.83 to 0.91 on an entry-mean basis (Rebetzke et al., 1998b). By comparison, the heritability estimate for 16:0 content on an individual F2 plant basis in a population segregating for the fap1, fapnc, and fan loci obtained from the regression of F2:4 progeny rows on F2 plants was 0.65 (Cherrak et al., 2003). Therefore, selection based on family means across a few environments should be successful in changing allele frequencies at both major and minor genes (Rebetzke et al., 1998b).
Breeders are also interested in the potential for unintentional correlated changes in other traits when selection is exerted for the trait of interest. Correlated changes in the other saturated and unsaturated fatty acids in soybean oil have been observed in low-palmitate lines that carry the major and minor genes described above. For example, when lines homozygous for the fap1 and fap3 alleles or the fapnc allele were compared with normal lines derived from the same population, the low-palmitate lines tended to have reduced stearate (18:0) and increased oleate (18:1) and linolenate (18:3) contents (Ndzana et al., 1994; Rebetzke et al., 1998a, 1998b, 2001). In addition, one study reported a significant genetic variation within each major fapnc allelic class due to the segregation of minor genes (Rebetzke et al., 1998a), and within those allelic classes, there was a significant negative genetic correlation between palmitate and oleate and a significant positive correlation between palmitate and linolenate (Rebetzke et al., 1998a). Palmitate and linoleate contents were negatively correlated, and palmitate and oleate contents were positively correlated in a population that was segregating for the fap1 and fan loci (Nickell et al., 1991). Therefore, correlations among fatty acid contents can change, depending on the particular fatty acid mutations segregating in the population under study.
Several studies have shown that the presence of major fap alleles that reduce the palmitate content in seed oil also reduces seed yield. Low-palmitate lines homozygous for the fap1 and fap3 alleles yielded significantly less, had less oil content, were taller, and had greater protein content (in two of three populations) when compared with normal lines derived from the same population (Ndzana et al., 1994). When a low-palmitate line homozygous for the fapnc allele was crossed to two high-yielding normal palmitate lines, progenies that inherited the major fapnc allele had a 10% decrease in yield and no differences in protein content in all crosses and an increase in oil content in only one cross (Rebetzke et al., 1998a). The effects of minor genes affecting palmitate content within the major fapnc allelic classes were not correlated with yield (Rebetzke et al., 1998a). Small but significant positive correlations between palmitate content and seed oil or seed yield were observed in a population derived from the cross of a low-palmitate, low-linolenate line homozygous for the fapnc, fap1, and fan alleles to cultivar Anand (Cherrak et al., 2003). None of the previously discussed studies reported genotypic correlation between these traits in populations that are segregating at both major and minor genes for palmitate content.
The objectives of this study were (i) to estimate genetic variances and heritabilities for fatty acid concentrations and agronomic traits in three populations segregating for the fapnc, fap1, and fan alleles as well as alleles at modifier genes grown in replicated trials, and (ii) to estimate the genotypic and phenotypic correlation of palmitate and linolenate contents with other fatty acids and agronomic traits in those populations.
|
MATERIALS AND METHODS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Phenotypic Evaluation of Populations
Each population was evaluated in separate field experiments. Ninety-eight F4-derived lines and the parents of the FADD00 population, and 99 F4-derived lines and one parent (N97-3681-11) of the FAHH00 population were grown in Plymouth, NC, in 2002 and in Caswell and Plymouth, NC, in 2003. Ninety-nine F4-derived lines and one parent (Anand) of the FACC00 population were grown in Clinton and Plymouth, NC, in 2002 and in Plymouth, NC, in 2003. Within each environment, the experimental design for each population was a 10 by 10 lattice with two replications. Experimental units for the FADD00 and FAHH00 populations were four 4.88-m rows, where only the middle two rows, end trimmed to 3.96 m, were harvested. Experimental units for the FACC00 population were three 5.79-m rows, where only the middle row, end trimmed to 4.88 m, was harvested. There was a 0.97-m spacing between rows.
Flowering date was recorded for FADD00 and FAHH00 at the R2 stage as number of days after planting (Fehr and Caviness, 1977). Maturity date was recorded for FADD00 and FAHH00 populations at the R8 stage as number of days after planting (Fehr and Caviness, 1977). Plots were mechanically harvested and yield and moisture content were measured on every plot in FADD00 and FAHH00. Yield was adjusted to 13% moisture content. Plots were mechanically harvested and yield was measured on every plot in FACC00 after several weeks of moisture stabilization in a seed laboratory. Ratings for lodging were taken whenever it was possible to observe variation for this trait in the three populations. A five-point lodging scale was used, where a rating of 1 was equivalent to a perfectly erect plant and a lodging rating of 5 was equivalent to plants lying on the ground. Height from the soil to the tip of a plant in centimeters was measured on a medium plant in a plot both years in populations FADD00 and FAHH00, but in population FACC00 only in 2002.
A 30-g subsample of seed from each plot was analyzed for fatty acid content, using gas–liquid chromatography of the methyl esters, at the USDA-ARS Nitrogen Fixation Laboratory, Raleigh, NC. Fatty acid content is reported as gram per kilogram of total lipids. Protein and oil content (dry-weight basis) in all the populations were analyzed by near infrared reflectance spectroscopy at the USDA North Regional Research Center, Peoria, IL, in 2003.
Statistical Analysis of Populations
All the traits in each population were analyzed as a lattice design across environments with PROC Mixed, SAS 8.2 (SAS Institute, 1999). Locations, replications, incomplete blocks, and lines were considered random effects. Genetic, genetic x environment, and error variances were estimated in each population and, with those estimates, heritability on a plot- and a line-mean basis and their approximate standard errors were estimated according to Holland et al. (2003). Genetic and phenotypic correlations with their standard errors were obtained by estimating genetic, genetic x environment, and error covariances by treating each pair of traits as repeated measurements in a combined lattice design analysis across environments with the SAS PROC Mixed procedure (Holland, 2006). Genetic and phenotypic correlations were considered significant if their absolute value exceeded 1.96 times their standard error (Holland, 2006).
|
RESULTS AND DISCUSSION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
These results differ from the conclusion by Rebetzke et al. (1996) that a few environments with contrasting temperatures were needed for the selection of low saturated soybean. Our heritability estimates on a plot-mean basis are higher than those estimated by parent–offspring regression (0.65) in one population of the cross Anand x N87-3708-13 (Cherrak et al., 2003) and also those reported for reduced and normal palmitate subpopulations (0.37–0.50) segregating for minor genes (Rebetzke et al., 1998b). A likely explanation for the discrepancy in our heritability results with Rebetzke et al. (1998b) is that, in our populations, both major and minor genes were segregating; thereby, the greater the genetic effect of a particular allele segregating in a population, the greater the genetic variance in that population and possibly the heritability. In Rebetzke et al. (1998b), only minor genes were segregating in the subpopulations. Additionally, there are important differences between our heritability estimator and that of Cherrak et al. (2003). Following Cockerham's (1983) notation, the expectation of the estimator of heritability based on regression of F2:4 offspring on F2 parents obtained by Cherrak et al. (2003) is equivalent to the covariance of relatives C0,0,2 = 
A2 + 0.25D2 + AA2, divided by the phenotypic variance of the parents 
(A2 + D2 + AA2 + AE2 + D E2 + AAE2 + e2) (Cockerham, 1983). In contrast, the expectation of our heritability estimate on a plot-mean basis is equivalent to G2/(G2 + GE2 + e2), where G2 is the covariance of C2,4,4 = 1.75A2 + 0.025D2 + 3.0625AA2. The numerator of our heritability estimator includes a 75% larger proportion of additive variance and three times the additive x additive variance. Also the error variance associated with plots containing multiple progenies of each line in our study is expected to be relatively smaller than the error variance associated with individual F2 plants, as in the Cherrak et al. (2003) study.
The genetic variances for linolenate was very similar across the three populations (Table 1), indicating that the effect of the fan allele is very stable across different genetic backgrounds. Similarly the genetic variances for linolenate were similar in two different crosses involving the low-linolenate line, A5, that carries the fan(A5) allele (Graef et al., 1988). The genetic x environment and error variances are eight to 10 times smaller than the genetic variance for linolenate in the three populations, indicating that the effect of the fan allele is also very stable across the different environments (Table 1). In contrast, Graef et al. (1988) observed larger error variances that were half the magnitude of the genetic variance. This difference in results could be due to the plot size used in their experiment (hills vs. 4.88-m plots in our study) and due to the fact that the error variance in their study includes relatively more genetic variance, as the F2 families in that study had 12 or fewer individuals (Graef et al., 1988). In the current study, heritabilities on a plot-mean basis (0.82–0.87) and on a line-mean basis (0.95–0.96) were very high, indicating that selection in one environment should be effective (Table 1). The heritability estimates on a plot-mean basis were higher than that estimated by regression of F2:4-derived progeny means on F2 plant values (0.73) in the population derived from Anand x N87-3708-13 (Cherrak et al., 2003) and higher than those estimated from replicated hills of F2:3 families (0.32–0.33) derived from A5 x ‘Pella’ and A5 x ‘Weber’ crosses (Graef et al., 1988). These differences could be attributed to the difference in genetic covariances among relatives included in each estimate and differences in experimental errors between plots and individual plants, as explained above.
The greatest differences in variance component estimates for fatty acid contents among the three populations occurred for oleate and stearate. The genetic variance estimate for oleate content in FADD00 was almost half that in the other two populations (Table 1). Similarly, the genetic variance for stearate content in FADD00 was less than half of that estimated for FACC00 and FAHH00. Consequently, the heritability estimates for stearate were lower for FADD00 than for the other two populations. These results indicate that these populations differ for at least some of the genes that affect the stearate and oleate contents. Similarly, large variation in the genetic variances for stearate and oleate were observed in different populations derived from A5 (Graef et al., 1988).
The genetic and genetic x environment variances for yield, date of flowering (R2), date of maturity (R8), height, and protein and oil contents are much smaller in FADD00 than in the other two populations (Table 2 ). Therefore, the heritability estimates on plot-mean basis are smaller for all of those traits in FADD00 and very different in the other two populations. The heritability estimates in this study are within the range of those reported by Burton (1987). The smaller genetic variances observed for all the traits in FADD00 (except for the small differences in 16:0 and 18:3) are a logical consequence of the close genetic relationship between the two parents of this population. Fewer genes for all traits are expected to be segregating in this population when compared with the other two. The same major genes for palmitate and linolenate are segregating in the three populations, however, which explains the similarity of the genetic variances for these two traits across populations.
|
|
There were significant small negative genetic and phenotypic correlations between palmitate and flowering date, between palmitate and maturity date, and between palmitate and protein content in FAHH00 only (Table 3), suggesting that the genetic background in a particular cross has an important effect on the correlation between these traits. There were significant positive genetic and phenotypic correlations between palmitate and oil concentration in FACC00 and FAHH00. Similarly, small but significant positive Pearsons' correlations between palmitate content and seed oil were observed in a population independently derived but genetically identical to FACC00 (Cherrak et al., 2003). There were no significant correlations between palmitate and lodging in the three populations.
Linolenate content was genetically (0.416 and 0.276) and phenotypically (0.348 and 0.193) correlated with lodging in FADD00 and FACC00, respectively (Table 3), in our study, unlike the studies of Ross et al. (2000) and Walker et al. (1998), who found no significant correlations between these two traits. There were small negative genetic and phenotypic correlations between linolenate and oil content in all three populations (Table 3), in agreement with Walker et al. (1998). This suggests that the effect, though small, is real and observable in other genetic backgrounds As in previous studies (Ross et al., 2000; Walker et al., 1998), linolenate was neither genetically nor phenotypically correlated with yield, flowering date, maturity date, or height. Thus selection for reduced linolenate content will not indirectly affect most agronomic traits.
Yield was negatively genotypically and phenotypically correlated with protein content and positively correlated with oil content in FAHH00 only (Table 3). The negative correlation between yield and protein has been commonly reported in soybean; however, significant correlations (positive or negative) between oil content and yield are less commonly observed in soybean.
Genetic and Phenotypic Correlations between Fatty Acids
For completeness, genetic and phenotypic correlations between the fatty acids are presented (Table 4
). There were significant positive genetic and phenotypic correlations between palmitate and stearate (0.45–0.77) in the three populations (Table 4). The probable explanation for the positive genetic correlation between stearate and palmitate in these populations is that the 16:0 thioesterase encoded by the GmFATB1a gene that is deleted in the fapnc allele also possesses some activity toward 18:0-acyl carrier protein (ACP) substrates, which is consistent with studies of 16:0 thioesterase enzymes in other plant species and the molecular basis of the fapnc mutation (Jones et al., 1995; Voelker, 1996; Cardinal et al., 2007). A reduction in stearate content in lines homozygous for fapnc was also observed in other studies, and positive phenotypic correlations between palmitate and stearate were observed in populations segregating for the fap1 and fapx alleles or the fap1, fapx, and fan alleles (Rebetzke et al. (1998b), 2001; Primomo et al., 2002).
|
Significant negative genetic correlations between palmitate and linoleate (–0.47 to –0.54) were observed in all three populations in this study (Table 4). Similar negative phenotypic correlations between palmitate and linoleate were observed in two populations segregating the fap1 and fapx alleles and in one population that also segregated the fan allele (Primomo et al., 2002). No significant genetic correlation was observed, however, between palmitate and linoleate in three out of four normal and reduced-palmitate populations segregating for minor genes that affect palmitate content and in lines derived from several cycles of recurrent selection for higher oleate (Rebetzke et al. (1996, 1998a).
There were large significant negative genetic and phenotypic correlations between oleate and linoleate in the three populations (Table 4). These correlations are in agreement with previous studies and the current knowledge of the fatty acid biosynthesis pathway (Ohlrogge et al., 1991; Ohlrogge and Browse 1995; Kinney, 1997; Rebetzke et al., 1996).
The significance of the genetic and phenotypic correlations between linolenate and the other fatty acids varies among the populations, therefore the effect of the fan allele on other fatty acids is population dependent. In the current study and previous studies, correlations between the other fatty acids were variable, depending on the particular breeding population being studied.
|
CONCLUSIONS |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
NOTES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Received for publication September 13, 2006.
|
REFERENCES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
in, D., G.R. Ablett, B.M. Luzzi, and J.W. Tanner. 1998. Use of gene substitution values to quantify partial dominance in low palmitic acid soybean. Crop Sci. 38:1437–1441.This article has been cited by other articles:
|
|
 |
|
  A. J. Cardinal, R. E. Dewey, and J. W. Burton Estimating the Individual Effects of the Reduced Palmitic Acid fapnc and fap1 Alleles on Agronomic Traits in Two Soybean Populations Crop Sci., March 19, 2008; 48(2): 633 - 639. [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 | |||
