| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Dep. of Horticulture, Michigan State Univ., East Lansing, MI 48824 USA
b USDA-ARS, Corn Insect and Crop Genetics Research Unit, Iowa State Univ., Ames, IA 50011 USA
c Dep. of Crop Sciences, Univ. of Illinois, Urbana, IL 61801 USA
bdiers{at}uiuc.edu
 |
ABSTRACT |
|---|
 TOPNOTES ABSTRACT INTRODUCTIONMaterials and methodsResults and discussionREFERENCES |
|---|
Abbreviations: BC, backcross • cM, centimorgan • LG, linkage group • LOD, likelihood of odds • QTL, quantitative trait locus • RFLP, restriction fragment length polymorphism • SSR, simple sequence repeat
|
INTRODUCTION |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMaterials and methodsResults and discussionREFERENCES |
|---|
The potential for using G. soja as a source of genes for increased protein concentration has been recognized since the 1940s. Glycine soja is believed to be the wild progenitor of soybean and these species are generally interfertile (Hymowitz and Singh, 1987). Weber (1950) studied the inheritance of protein concentration in a population developed from a cross between a G. soja plant introduction (PI) and the G. max cultivar Dunfield. The G. soja parent had nearly a 100 g kg-1 greater seed protein concentration on a moisture-free basis than the G. max parent. In the population, the broad-sense heritability of protein concentration was 0.70, and Weber estimated that 2.9 genes controlled this trait.
Diers et al. (1992) mapped two major QTL controlling protein and oil concentration with RFLP markers. This mapping was done in a population of F2-derived lines developed by crossing a G. max experimental line and a G. soja plant introduction. The QTL mapped to what were then labeled linkage groups (LG) A and K of the soybean map. The map has been revised and LG A has been renamed LG E, and LG K has been renamed LG I (Shoemaker and Specht, 1995). Diers et al. (1992) found that the G. soja allele for the most significant marker from each LG was associated with an increase in protein concentration of 17 g kg-1 for LG A and 24 g kg-1 for LG K. The high protein alleles from G. soja were associated with reduced oil concentration for both QTL.
There have been other reports of genes controlling protein concentration mapping to the same region on LG I (previously LG K) as reported by Diers et al. (1992). Brummer et al. (1997) mapped a QTL for increased protein concentration to this region using a population in which one parent was 25% G. soja germplasm. This suggests that their high protein gene also came from G. soja. Hegstad et al. (2000) mapped the wp locus, which confers pink flower color in the presence of W1 (Stephens and Nickell, 1991), to the same region on LG I. Greater seed protein concentration has been observed in pink flowered lines suggesting either wp or a closely linked gene controls protein concentration (Stephens et al., 1993).
Although the Diers et al. (1992) mapping study is a useful first step, it is important to determine the effect of QTL alleles in new environments, in other genetic backgrounds, and on related traits. For example, Tanksley and Hewitt (1988) mapped three QTL from a wild relative of tomato (Lycopersicon esculentum Mill.) that were associated with increased soluble solids content of fruit. Further testing showed that one QTL was dependent on its genetic background and that these QTL had a negative effect on other production traits.
The first objective of our study was to test the G. soja QTL alleles on LGs E and I in a backcross (BC) population to reexamine their effect on protein and oil concentration and to test the effect of these alleles on other agronomic traits. The second objective was to evaluate the effect of the G. soja QTL allele on LG I in three genetic backgrounds.
|
Materials and methods |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMaterials and methodsResults and discussionREFERENCES |
|---|
Fifty-three BC3F4-derived lines and A81-356022, the recurrent parent, were field evaluated during the summers of 1996 and 1997 in randomized blocks at two locations each year with one replicate at each location. Plots were 4 m long, two rows wide, with a 76-cm row spacing, and were sown at a rate of 23 seeds m-1 row. During 1996, F4:6 lines were sown on 16 May at East Lansing and on 23 May at Britton, MI. In 1997, F4:7 lines were sown on 4 June at Mason, MI, and 10 June at Britton.
Plots were rated both years for plant height, maturity date, and lodging. Maturity date was recorded as the number of days after 31 August when 95% of the plants in a plot reached their mature pod color (R8) (Fehr et al., 1971). Plant height and lodging were recorded when the plots were mature. Plant height was measured as cm from the ground to the average terminal node of plants. Plots were rated for lodging on a scale of one to five with one designated as plants standing erect and five as plants lying prostrate to the ground.
Plots were not end-trimmed during the growing seasons and were harvested with a combine to measure seed yield. Seed yields are reported on a 130 g kg-1 moisture basis. Seed size was determined by weighing 100 seeds from each plot. Seed protein and oil concentration were measured by near infrared transmittance on whole beans at the USDA Northern Regional Research Center at Peoria, IL, and reported on a moisture-free basis. Measurements were taken on a 21 to 25 g sample harvested from each plot.
The population was tested with RFLP and simple sequence repeat (SSR) markers using leaf tissue collected in 1996 from several plants of each BC3F4:6 line grown in the field trial near East Lansing, MI. DNA was extracted from the leaf tissue according to Kisha et al. (1997) and Southern blotting, hybridization, and autoradiography were performed as described by Diers and Osborn (1994) with modifications according to Kisha et al. (1997). Simple sequence repeat marker data were collected according to Cregan and Quigley (1997). The SSR markers were developed by Dr. P.B. Cregan (USDA-ARS, Beltsville, MD) and the primers were obtained from Research Genetics (Huntsville, AL). Soybean RFLP and SSR markers from LG E and I were screened against parental DNA to identify polymorphisms. Each RFLP marker was screened against parental DNA digested with the enzyme previously used to map it (SoyBase, 1995). The entire population of 53 BC3F4-derived lines was then tested with the polymorphic markers.
All data collected were analyzed by standard analysis of variance procedures with PROC GLM of SAS (SAS Institute, 1987). Genotypes, locations, and years were treated as random effects. Associations between traits and markers were tested with one-factor analysis of variance by PROC GLM of SAS. The R2 value was used to describe the proportion of the variance in the population explained by individual markers. Lines heterozygous for the genetic markers were included in the analysis but their means were not presented in the summary tables (Tables 1–3) . Pearson product-movement correlations were calculated with PROC CORR of SAS to test for correlations among traits. The correlation values were calculated from the means of genotypes across locations.
|
|
|
Test Populations
A BC3F4-derived line from the BC population described above was crossed to the cultivars Parker (Orf and Kennedy, 1994) and Kenwood (Cianzio et al., 1990) and the experimental line C1914 (Wilcox, 1995). The BC3F4-derived line was selected because, based on RFLP markers, it was homozygous for the G. soja region on LG I where the protein QTL mapped. Populations developed from each cross were inbred by single seed descent to the F3 generation. The population from the cross with Parker (Parker population) included 100 lines, and the populations from the crosses with Kenwood (Kenwood population) and C1914 (C1914 population) included 98 lines each.
The F3:4 lines from each population and parents were sown on 9 June 1997 near Mason, MI. Lines were sown in single-row plots, 1 m long with a 76-cm row spacing and a seeding rate of 30 seeds m-1 of row. Each population was grown in a separate block. In 1998, lines from each population were sown in plots replicated twice using a randomized complete block design at Urbana, IL, and East Lansing, MI. Lines at the Urbana location were sown on 27 April in two-row plots, 3.2 m long, with a 76-cm row spacing and a seeding rate of 39 seeds m-1 of row. Both rows were harvested to measure seed yield. The East Lansing location was sown on 12 May in six-row plots, 4.3 m long with a 38-cm row spacing and a seeding rate of 25 seeds m-1 of row. The middle four rows from each plot were harvested to measure seed yield. Plots were not end trimmed at either location.
The seed harvested from both years was evaluated for seed weight and protein and oil concentration. In addition, plots were evaluated for plant height, lodging, maturity and seed yield in 1998. The seed and agronomic traits were measured according to the methods listed for the backcross population. Leaf tissue was collected from the 1997 plots for RFLP marker analysis. The DNA extractions, marker testing, and data analysis were done with the methods described for the backcross population.
|
Results and discussion |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMaterials and methodsResults and discussionREFERENCES |
|---|
|
The population was evaluated in field trials in two locations over 2 yr. Because of wet field conditions, the East Lansing location in 1997 was not harvested for seed yield but seed samples were taken from plots to measure seed weight and conduct protein and oil analyses. There was significant (P < 0.05) genotypic variation in the population across environments for all traits except lodging. The location x genotype and year x genotype interactions were not significant for any trait although these interactions were not testable for yield because there was no yield data for the East Lansing location in 1997.
The marker pb on LG E was not significantly (P < 0.01) associated with the seed and agronomic traits at any environment or across environments. In contrast, the markers on LG I were significantly associated with all measured traits except lodging. Across environments, BARC-Satt127 had the greatest R2 value for protein concentration, seed yield, maturity date, and plant height, whereas IaSU-A144H-1 had the greatest R2 value for oil concentration, and seed size (Table 1).
The marker alleles from G. soja for IaSU-A144H-1 and BARC-Satt127 were associated with greater protein and lower oil concentration than the G. max alleles (Table 1). This is consistent with the phenotypic correlation of -0.86 between these traits in our population and with observations by other researchers (Burton, 1984; Wehrmann et al., 1987; Wilcox, 1998). The associations between the significant markers and protein and oil concentration were greater in 1996 than in 1997 (Table 1). For protein, the R2 value for BARC-Satt127 was 0.80 across locations in 1996 and 0.41 in 1997 (Table 1). This is consistent with difference between the means of the homozygous classes for BARC-Satt127. This difference dropped from 30 g kg-1 in 1996 to 12 g kg-1 in 1997. For oil, the R2 value for IaSU-A144H-1 declined from 0.51 across environments in 1996 to 0.11 in 1997. It is unclear why such a large difference in the marker association between years was observed. Although there was a significant difference between years for the average protein concentration, this difference was only 3 g kg-1 of seed. One factor that may have contributed to the year difference was the later planting and maturity of lines in 1997 than in 1996. Compared with 1996, the average planting date in 1997 was over 17 d later and the average maturity was 19 d later. The later planting and maturity may have reduced the relative effect of the G. soja gene or genes controlling these traits.
The G. soja alleles for IaSU-A144H-1 and BARC-Satt127 were significantly associated across years with earlier maturity, taller plants, reduced yield, and smaller seed than the G. max alleles (Table 1). This is consistent with the significant phenotypic correlation in the population between protein concentration and plant height (0.36), maturity (-0.44), seed size (-0.77), and yield (-0.62). Burton (1984) also reported significant correlations between greater protein concentration and lower yield.
This research confirms the finding of Diers et al. (1992) that the G. soja line PI 468916 has a major QTL allele on LG I that increases seed protein concentration. The effect of this QTL allele was not diminished when it was backcrossed into the A81-356022 background. The homozygous G. soja and G. max classes for IaSU-A144H-1 differed for protein concentration by 18 g kg-1 in the original F2 population (Diers et al., 1992) compared with 17 g kg-1 across both years for the backcross population.
The resolution of our mapping is insufficient to shed light on whether the same gene is controlling all of the significant traits on LG I, or if they are controlled by different genes. The peaks from the Mapmaker analysis for the traits can not be separated in our current study (Fig. 1). Much larger populations and/or fine mapping using lines with recombination close to the QTL (Paterson et al., 1990) would be needed to further study whether these traits are controlled by one or more genes.
Our research is inconclusive for the protein QTL on LG E mapped by Diers et al. (1992). Segregation of pb was not significantly associated with protein concentration in the backcross population. However, because RFLP markers flanking this gene were not segregating in the population, the lack of association may be the result of recombination between the protein QTL and pb during the backcrossing process.
Test Populations
The high protein QTL on LG I was further evaluated in three test populations developed by crossing a line from the BC3 population with the cultivars Parker and Kenwood and the experimental line C1914. The cultivars are high yielding, have average protein concentration, and are representative of cultivars in commercial production. The experimental line C1914 was developed by J.R. Wilcox, USDA-ARS (Wilcox, 1995). The line has a high protein concentration, but below average yield. Lines in all three populations were tested with the RFLP markers IaSU-A144H-1, IaSU-A515I-1, and IaSU-A688D-1.
Parker Population
There was significant genetic variation in the Parker population for all seed and agronomic traits (P < 0.05). Across locations, all three markers were significantly associated with protein concentration and the markers IaSU-A144H-1 and IaSU-A688D-1 were significantly associated with oil concentration. The RFLP marker IaSU-A144H-1 had the greatest association with these traits with an R2 value of 0.44 for protein concentration and 0.15 for oil concentration across environments (Table 2). Lines homozygous for the G. soja alleles of IaSU-A144H-1 had 20 g kg-1 greater protein and 9 g kg-1 less oil concentration than the homozygous G. max lines across environments.
Of the three markers, only IaSU-A688D-1 was significantly associated with seed yield and maturity date and only IaSU-A515I-1 was significant for seed size. For these markers, lines homozygous for the G. soja allele had lower yield, smaller seed, and earlier maturity than lines homozygous for Parker alleles. There were no significant associations between the markers and lodging and plant height.
Kenwood Population
The results from this population were similar to the Parker population. There was significant (P < 0.05) variation in the population for all seed and agronomic traits. All three markers were significantly associated with protein and oil concentration across environments. The RFLP marker IaSU-A144H-1 had the greatest association with these traits with an R2 value of 0.41 for protein and 0.23 for oil concentration across environments (Table 3). Lines homozygous for the G. soja allele of IaSU-A144H-1 had a seed concentration of 19 g kg-1 greater protein and 9 g kg-1 less oil than the homozygous Kenwood lines across environments.
All three markers were significantly associated with seed weight. The marker IaSU-A515I-1 had the greatest R2 value across environments (Table 3). The lines homozygous for the G. soja allele had smaller seed than the homozygous Kenwood lines. The only significant marker association with seed yield was IaSU-A688D-1 at Urbana (Table 3). Lines homozygous for the G. soja allele had reduced seed yields compared with the homozygous Kenwood lines. The markers were not significantly associated with maturity, lodging, or plant height at any environment or across environments.
C1914 Population
There was significant (P < 0.05) variation in the population for all traits. Although the genetic variation for protein content was significant in this population, this variation was 3.3 times less than in the Kenwood population and 2.4 times less than in the Parker population. There were no significant associations between the markers and protein and oil concentration or any of the agronomic traits.
Conclusions for the Test Populations
The protein QTL allele from G. soja was associated with an increase in protein concentration in the Parker and Kenwood populations but not in the C1914 population. This suggests that there may be a gene conferring increased protein concentration in C1914 but not in Parker or Kenwood that is allelic with the G. soja gene on LG I. The source of the high protein concentration in C1914 was Pando (J.R. Wilcox, 1999, personal communication), a high protein G. max plant introduction from Korea (Bernard et al., 1987). These results suggest that the high protein QTL allele from G. soja that we studied could already exist in Pando and other high protein soybean genotypes.
The high protein QTL allele from G. soja was associated with lower yield in the backcross, Kenwood and Parker populations but not in the C1914 population. This provides evidence that the high protein QTL allele from G. soja also reduces yield. If the high protein QTL allele was associated with less yield only because of coupling linkage with a yield reducing gene, this coupling linkage would have to be present in both PI468916, the G. soja parent, and C1914. If the coupling linkage was not present in C1914, a yield QTL would have been detected on LG I in the C1914 population. The coupling linkage seems less likely than the high protein allele causing the yield reduction.
We were successful in improving the protein concentration of soybean genotypes through genetic mapping coupled with marker assisted selection during backcrossing. Marker assisted selection allowed us to backcross the G. soja genes into a soybean genotype much more rapidly than can be done with tradition backcross methods. We were able to produce the BC1F1, BC2F1, and BC3F1 all within 1 yr. In contrast, it took Wilcox and Cavins (1995) more than 10 yr to produce backcross three populations with increased protein content. A disadvantage of the marker assisted selection approach is that the genes must be mapped prior to backcrossing which is not required in the traditional backcross approach.
Although we have evidence that the QTL we introgressed from G. soja was already in soybean germplasm, our introgression methods allowed us to determine this rapidly with the test populations. Without the marker associations, it would have been difficult to conclusively determine this. The methods we used in this study could be applied to other quantitative traits as our germplasm collections are mined for new genes in the future.
|
NOTES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMaterials and methodsResults and discussionREFERENCES |
|---|
Received for publication October 27, 1999.
|
REFERENCES |
|---|
|
TOPNOTESABSTRACTINTRODUCTIONMaterials and methodsResults and discussionREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  D. Li, T. W. Pfeiffer, and P. L. Cornelius Soybean QTL for Yield and Yield Components Associated with Glycine soja Alleles Crop Sci., March 19, 2008; 48(2): 571 - 581. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. Liu, T. Fujita, Z.-H. Yan, S. Sakamoto, D. Xu, and J. Abe QTL Mapping of Domestication-related Traits in Soybean (Glycine max) Ann. Bot., October 1, 2007; 100(5): 1027 - 1038. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. M. Nichols, K. D. Glover, S. R. Carlson, J. E. Specht, and B. W. Diers Fine Mapping of a Seed Protein QTL on Soybean Linkage Group I and Its Correlated Effects on Agronomic Traits Crop Sci., February 24, 2006; 46(2): 834 - 839. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Schauer, D. Zamir, and A. R. Fernie Metabolic profiling of leaves and fruit of wild species tomato: a survey of the Solanum lycopersicum complex J. Exp. Bot., January 1, 2005; 56(410): 297 - 307. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
V. A. Fasoula, D. K. Harris, and H. R. Boerma Validation and Designation of Quantitative Trait Loci for Seed Protein, Seed Oil, and Seed Weight from Two Soybean Populations Crop Sci., July 1, 2004; 44(4): 1218 - 1225. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. Chung, H. L. Babka, G. L. Graef, P. E. Staswick, D. J. Lee, P. B. Cregan, R. C. Shoemaker, and J. E. Specht The Seed Protein, Oil, and Yield QTL on Soybean Linkage Group I Crop Sci., May 1, 2003; 43(3): 1053 - 1067. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. A. Hoeck, W. R. Fehr, R. C. Shoemaker, G. A. Welke, S. L. Johnson, and S. R. Cianzio Molecular Marker Analysis of Seed Size in Soybean Crop Sci., January 1, 2003; 43(1): 68 - 74. [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 | |||
