| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Dep. of Natural Resources and Environmental Sciences, University of Illinois at Urbana-Champaign, 307 ERML, 1201 W. Gregory Dr., Urbana, IL 61801
* Corresponding author (j-juvik{at}uiuc.edu)
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Abbreviations: ANOVA, analysis of variance • MAB, marker-assisted backcrossing • RFLP, restriction fragment length polymorphism • QTL, quantitative trait locus/loci • RIL, recombinant inbred line • sh2, shrunken2, endosperm mutation in maize (Zea mays L.)
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Many theories have been proposed to explain the physiological basis of poor seedling emergence associated with the sh2 mutation. Douglass et al. (1993) proposed that poor seedling emergence was due to the reduced starch concentration in dry kernels, which results in reduction in the energy reserves required for emergence. Reduced kernel starch content results in greater cracking of kernel pericarp during sh2 seed maturation, which is then responsible for more solute leakage during germination (Tracy and Juvik, 1988). Cell membrane damage associated with high osmotic potential generated by elevated sh2 kernel sugar concentration and the resulting rapid influx of water during imbibition (Simon, 1978) are also implicated as mechanisms reducing seedling survival. Harris and DeMason (1989) found an association between poor emergence and the lowered activities and amounts of the starch hydrolytic enzyme, 
-amylase, in sh2 compared with su1 sweet corn. Finally, Headrick et al. (1990) reported that susceptibility of kernels during maturation to infection by fungal pathogens such as Fusarium moniliforme was associated with reduced seeding emergence.
Apparently a number of kernel physiological properties and interactions among them are associated with reduced sh2 seedling emergence (Juvik et al., 1993). Several reports have shown that seedling emergence is a quantitative trait controlled by multiple QTL (Azanza et al., 1996; Han, 1994). Six marker QTL, located on five different chromosomes of a mapped F2:3 sh2 population explained most of the variation in seedling emergence (R2 = 67%) (Han, 1994). These markers were also observed to be associated with kernel chemical and eating quality characteristics. Identification of QTL, useful for improving emergence of sweet corn hybrids, would be beneficial to sweet corn breeding programs.
Marker-assisted backcrossing (MAB) has been suggested as a breeding strategy to introgress a limited number of QTL into elite germplasm (Dudley, 1993; Stuber, 1994; Bernacchi et al., 1998). This can minimize linkage drag while expediting the transfer of targeted chromosomal segments from exotic germplasm into desired backgrounds (Young and Tanksley, 1989; Tanksley and Nelson, 1996; Hospital and Charcosset, 1997). Empirical and simulation studies provide sufficient evidence that MAB of monogenic traits can reduce time and resources sufficiently to justify its use in cultivar development (Lee, 1995). However, breeders are cautiously optimistic about the advantages of MAB for polygenic traits. There is a consensus that identified QTL should be tested in several different genetic backgrounds and evaluated for associated effects on the characters of agronomic or economic importance prior to applied utilization (Tanksley and Hewitt, 1988; Zehr et al., 1992; Danzmann et al., 1999). The feasibility of using marker-assisted selection in breeding programs is determined by the reproducibility of marker-QTL associations across generations, populations, and environments (Dudley, 1993).
Studies have shown that QTL expression is sensitive to environmental conditions. Variability in consistency of QTL effects across a set of test environments has been reported (Guffy et al., 1989; Stuber et al., 1996; Zehr et al., 1992; Schön et al., 1994, Melchinger et al., 1998). While some reports have shown consistent reproducibility of QTL effects in maize and other crops across generations of inbreeding and environments (Austin and Lee, 1996, 1998; Veldboom and Lee, 1996; Lu et al., 1997), other investigations of QTL effects tested in different genetic backgrounds are equivocal (Toojinda et al., 1998; McKendry et al., 1996; Wang, 1997). Beneficial QTL are commonly assumed to exert a background specific effect due to the interaction among QTL. Tanksley et al. (1996) reported that testing backcrossing generations at BC2F1 and BC3 would provide an estimate of the introgressed QTL effect on enhancement of the desirable character(s) in the breeding program.
In this study, a recombinant inbred line (1657-90 sh2) developed from a previously mapped F2:3 population (Han, 1994) was used as a donor parent for three marker-QTL alleles that enhance seedling emergence in sweet corn. These three marker alleles were introgressed into three elite sh2 sweet corn inbreds utilizing marker-assisted backcrossing. The objectives were: (i) to evaluate the effects of these introgressed QTL on seedling emergence in different sh2 genetic backgrounds, and (ii) test for the reproducibility of marker associated effects observed in the mapping population with BC2F1 lines developed from this population following eight generations of selfing and two generations of backcrossing.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Validation Tests for the Detected QTL
Validation tests for the detected QTL number, position, and effect are becoming essential requirements to obtain realistic and unbiased estimates of the associated effects with marker loci (Utz et al., 2000). Composite interval mapping (CIM) analysis with the PLABQTL program (Utz and Melchinger, 1995) was used to confirm the map locations and effects of the putative QTL previously identified (Han, 1994) by single factor analysis. The three loci were found to be associated with seedling emergence using CIM as was observed with single factor analysis. In further testing, the original mapping population (117 F2:3 families) was re-evaluated for seedling emergence in three replications in two more environments, Illinois 1994 and Wisconsin 1993. Data from these two environments (as well as data from across the three environments, using family means of three environments) were analyzed with PLABQTL to test for the location, number, and effect of the detected QTL in the original mapping population. Also, the "cross" command of PLABQTL was used to generate 100 randomly selected partial data sets containing 4/5 of the families in the original data set from the Illinois 1993 environment (the mapping data). Composite interval analysis was conducted on these partial data sets to provide a validation of QTL presence and position. A frequency distribution of significant QTL position (on the basis of maximum likelihood scores) was generated from these partial data sets and compared with those chromosomal positions where significant QTL were detected with the complete data set. The percentage of partial data sets where maximum and significant LOD scores fell within the confidence interval of the QTL mapped in the original data and linked to php200689, umc139, and bnl9.08 marker loci, was 62, 88, and 81% of the validation runs, respectively. Validation of the effect and location of these QTL is provided in a separate investigation where marker-assisted selection with these same donor alleles was successful at enhancing seedling emergence (Yousef and Juvik, 2001a).
These marker-QTL alleles differed in regard to their influence on seedling emergence and eating quality of sweet corn. While the allele linked to umc139 improved both emergence and eating quality, the alleles linked to bnl9.08 and php200689 affected emergence only (Han, 1994).
Introgression of Marker QTL into Elite Sweet Corn Inbreds
In a marker-assisted backcrossing program, the RIL (1657-90, F8 sh2) containing the three marker-QTL alleles was used as a donor parent. Nine elite sh2 sweet corn inbreds provided by Illinois Foundation Seed, Inc. (Tuscola, IL) were screened for polymorphisms at the three marker loci. These sh2 inbreds are genetically unique inbreds and used as parents of commercial hybrids. Three of the inbreds with distinct genetic backgrounds (I-1, I-5, and I-6) that displayed polymorphisms at all three marker loci were selected for use as recurrent parents. These inbreds differed in percent seedling emergence, flowering time, plant vigor (growth mm/day), plant height, and eating quality characteristics. They displayed relatively poor seedling emergence and were presumed to lack any of the beneficial alleles associated with the selected markers.
The donor parent was crossed to the three commercial inbreds (I-1, I-5, and I-6) and the resulting F1s were backcrossed to generate three BC1 populations. Seventy-five BC1 plants in each population were backcrossed to the respective recurrent parent to create BC2 families. BC1 plants were then screened for marker loci and one BC2 family in each population generated from crossing BC1 plants heterozygous for all three donor-marker alleles was selected and planted. Due to segregation, each selected BC2 family contained plants with none, one, two, or three of the donor marker-QTL alleles in the heterozygous condition. According to Mendelian theory for unlinked loci, the BC2 progeny should contain eight genotypic classes that occur in equal frequency (0.125%): ADARBDBRCDCR, ADARBDBRCRCR, ADARBRBRCDCR, ADARBRBRCRCR, ARARBDBRCDCR, ARARBDBRCRCR, ARARBRBRCDCR, and ARARBRBRCRCR, where A, B, and C refer to the three marker alleles, and D and R refer to the donor and recurrent parent as the source of allele. Individual plants of the selected BC2 families (75 plants) were screened for marker loci and crossed to the non-recurrent commercial inbreds as testers. This resulted in the development of three BC2F1 populations. The BC2 plants developed from the recurrent parent I-1 [I-1(BC2), 
] were crossed to I-6 
and I-5 resulting in the two BC2F1 populations [I-6 x I-1(BC2)] and [I-5 x I-1(BC2)], respectively. The BC2 plants developed from the recurrent parent I-6 [I-6(BC2), ] were crossed to I-5 resulting in the third BC2F1 population [I-5 x I-6(BC2)]. Each BC2 plant was crossed to five non-recurrent inbred plants to produce sufficient seed for field evaluation.
At any given marker locus, BC2F1 families in each population consisted of 50% plants heterozygous and 50% plants lacking the donor marker allele. Hereafter, BC2F1 families that contained any of the marker alleles will be designated "segregating for donor allele", while families homozygous for commercial inbred alleles will be designated "lacking donor allele" at any specific marker locus (loci). When considering all three loci simultaneously, genetic expectation would predict that 1/8 of the BC2F1 families in each population are lacking any of the donor marker alleles with 7/8 of the families containing the marker donor allele(s) in all possible combinations in the heterozygous condition.
A total of 50 BC2F1 families with sufficient seed for replicated evaluation were produced; 13 families in each of the [I-6 x I-1(BC2)] and [I-5 x I-1(BC2)] populations and 24 families in the [I-5 x I-6(BC2)] population. It was observed that BC2 plants containing the donor marker alleles were more vigorous, faster growing, and flowered earlier than the commercial inbreds. This unintentionally resulted in a higher frequency of crosses using BC2 plants as male parents that lacked donor marker alleles. Chi-square tests (
2) for the expected versus observed number of BC2F1 families with and without donor alleles were significant at P < 0.01, suggesting that this unintentional parental selection resulted in a higher than expected frequency of BC2F1 families without the donor marker alleles.
The commercial inbreds were also hybridized, in the same direction as described above, to generate three unmodified F1 hybrids for use as controls. I-1 was used as male parent and crossed to I-6 and I-5 as females, while I-6 was used as the male and crossed to I-5 as the female parent. The same crossing directions were used in both modified and the unmodified hybrids. This conforms with the use of these inbreds in the commercial production of hybrid seed.
DNA Analysis and Marker Screening
DNA was isolated from finely ground frozen samples of fresh leaf tissue according to the procedures described by Hoisington (1991) and modified by Mikkilineni (1997). DNA was extracted with 2% CTAB, 1.4 M NaCl, 100 mM Tris HCl (pH 8.0), 20 mM EDTA (pH 8.0), and 0.2% 2-mercaptoethanol. Ten micrograms of DNA were digested with 30 units of EcoRI and subjected to Southern Blotting analysis as described by Hoisington (1991). The three RFLP markers are maintained in plasmid vectors. These clones originated from collections of mapped maize clones developed and provided by the University of Missouri-Columbia (umc), Brookhaven National Laboratory (bnl), and Pioneer Hi-Bred International (php). Genomic cDNA clones were oligolabeled (Feinburg and Vogelstein, 1983) and hybridized to membranes according to Hoisington (1991). Hybridized membranes were wrapped in plastic covers, placed in cassettes, and exposed to X-ray film at -80°C for 2 to 5 d depending on the intensity of P32 labeling and scored. Families were screened for presence/absence of the donor marker allele(s).
To screen BC1 plants for the RFLP markers, a subset of 10 kernels from each BC2 family were sown in the greenhouse for DNA extraction. Bulked BC2 seedling tissue from each family was used to screen BC1 plants. One BC2 family in each population, segregating for all three marker loci, was used for crossing to the commercial inbreds to produce BC2F1 families. BC2 plants used for crossing were screened for the RFLP marker loci using genomic DNA extracted from bulked leaf tissues of 30 seedlings from each of BC2F1 family. The commercial inbreds involved in each specific cross, as well as the donor parent, were included in the RFLP analysis as references.
Experimental and Statistical Analysis
The fifty BC2F1 families, three unmodified F1s, and commercial inbreds (I-1, I-5, I-6, and the donor parent, 1657-90) were evaluated for seedling emergence in two successive years at the University of Illinois' South Farm in Urbana-Champaign. The soil type was Drummer silty clay loam. The average soil temperature at 4-cm depth for the 24-hr period after seeds were planted was 14°C and 23°C in 1998 and 1999, respectively. The experimental design was a randomized complete block (RCBD) with three replications in both years. Using hand planters, 50 kernels per family were individually planted in each replicate at about 4-cm soil depth in 8-m rows on 1 May, 1998 and 25 May, 1999. A total of 300 kernels/family were used in both years. Seedling emergence was determined by counting the emerged seedlings at 4 wk after planting.
Since marker-QTL x environment interaction was not found to be a major source of variation (Table 1), the statistical analysis was performed on combined data from the two environments, 1998 and 1999. Replications were nested within environments. The BC2F1 families were nested within each marker locus class. For statistical analysis, data were analyzed in two ways by the same statistical model described below. The first analysis was designed to compare means of the two marker classes at each marker locus averaged across all families in each population. The second analysis compared the means of the eight possible marker-QTL genotypes.
|
 |
= residual error. Since the number of families within marker classes was not equal, mean comparisons were performed using Lsmeans for marker-QTL genotype classes instead of LSD, which is suitable for balanced designs. The SAS statement, Lsmeans marker/pdiff was used after the model statement, where the analysis output contained paired T-test comparisons between marker classes and probability values used for rejection or acceptance of the null hypothesis which states that class effects or treatments are not different at P < 0.05 (SAS, 1991). The unmodified F1 hybrids and commercial inbreds were not included in any of the statistical analyses, but they are presented in the tables for reference.
Comparing Means of the Two Marker Classes at Each Marker Locus
To investigate marker-QTL associated effects on seedling emergence and in different genetic backgrounds, the two marker-QTL classes in each of the three BC2F1 populations were compared. The first class contained the BC2F1 families segregating for a specific donor allele, and the second had the BC2F1 families lacking the donor allele. In addition, these two marker classes were compared across the three sh2 genetic backgrounds by pooling all the data from all three populations.
To test for the reproducibility of marker-QTL associated effects across generations, the average seedling emergence of marker-QTL classes across all three populations were calculated as the percentage increase over the unmodified F1 hybrids and compared with their associated effects in the original mapping population. The donor allele effects in the original mapping F2:3 population was calculated as the percentage increase in seedling emergence of the heterozygous families over the homozygous families with the unfavorable allele. The associated effect of the donor QTL in the BC2F1 generation was compared with its effect in the original mapping F2:3 population, using 1993 data from Illinois (Han, 1994). This provides an estimate of the reproducibility of the marker-QTL associated effects across generations of selfing and backcrossing.
Since the above analysis was done for each marker locus regardless of the effect of any other loci segregating in the genome, the confounding effects of the other donor alleles in BC2F1 crosses was not taken into account in the analysis. The same condition applies to results from the analysis of the original mapping population where effects of individual QTL would be confounded by random segregation of other genes in the F2 generation.
Comparing the Means of the Eight Possible Marker Genotypes
To investigate the individual and combined effects of the donor QTL, data from the BC2F1 families of each population were sorted into the eight possible marker genotypic classes with either none or segregating for one, two, or all three of the donor marker loci. These classes were compared with each other within and across the three populations. This allowed for comparing families without any of the marker alleles with those that were segregating for one or any combination of the three alleles in each of the three populations. While this analysis removed the confounding effects of the other segregating QTL, experimental rigor was lost because of the limited number of families in each of the BC2F1 populations. The number of individual BC2F1 in each of the genotypic classes in each population was uneven, ranging from 0 to 7. However, when families were pooled across the three populations, all marker loci classes were represented by at least 2 BC2F1 families.
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
|
Effects of the Donor QTL in BC2F1 Hybrids
Comparing Means of the Two Marker Classes at Each Marker Locus
Results indicated that seedling emergence was significantly enhanced in all three BC2F1 populations with the addition of the Ia453 allele linked to the umc139 marker on chromosome 2 (Table 3). The BC2F1 families segregating for the donor allele near the umc139 locus showed significantly improved emergence over the families lacking the donor allele at this locus. The average percentage increase in seedling emergence of families segregating at the umc139 QTL was 54% over the families lacking the donor allele in the three populations. The observed variability in the BC2F1 populations indicated that this QTL allele had a substantial effect on improving seedling emergence in sh2 sweet corn. This agrees with earlier investigations (Wang, 1997), where the Ia453 allele linked to this marker locus was observed to improve emergence over one cycle of pedigree selection in one sweet corn genetic background across two environments. Interval mapping analysis using PLABQTL (Utz and Melchinger, 1995) suggests that the most likely location for this QTL is between umc131 and umc139 on chromosome 2 at a distance of 4 and 9 cM from each of these RFLP loci, respectively. The estimates of QTL location were obtained using the LOD score with support intervals with 1.0 drop off on either side. Support interval of 0–13 cM on chromosome 2 including the umc139 locus was observed with PLABQTL. Mapping distances along chromosomes are reported in units of recombination (centiMorgans, cM) and start at the position of the marker on the short arm most distal to the centromere and increase as one moves toward the most distal marker locus on the long arm of the same chromosome. The percentage partial data sets which identified significant QTL within the 0–13 cM interval was 88% with peak logarithms of odds (LOD) score values concentrated in the region from 0–5 cM (80%).
|
Averaged across all three populations, seedling emergence was enhanced by 40.8, 30.2, and 28.2% over the unmodified hybrids with the introgression of the beneficial QTL alleles linked to umc139, bnl9.08, and php200689 marker loci, respectively (Table 4). In the original mapping population, seedling emergence mean of the F2:3 families heterozygous for umc139, bnl9.08, or php200689 QTL alleles was higher by 30.4, 33.3, and 36.4% over families homozygous for unfavorable alleles at these loci, respectively. The percentage increases in seedling emergence in the BC2F1 and the original F2:3 mapping population (Table 4) suggested that marker-QTL associated effects of these loci were reproducible across generations and populations. These results did not take into account the confounding effects of the other QTL segregating in the BC2 generation or F2:3 population. However, the large effects associated with these QTL suggest that introgression of these alleles into the recurrent parent backgrounds can improve seedling emergence even when heterozygous. Currently, near isogenic lines of BC3S1 of the modified inbreds are being developed to evaluate these QTL when homozygous in the three backgrounds.
|
|
The assayed families for seedling emergence were segregating for the marker-QTL, therefore half of the plants in each family was heterozygous for marker QTL and the other half was lacking donor allele(s). If we assume that all plants in the BC2F1 family were heterozygous at a particular donor locus (loci), then the increase over the unmodified F1 would be twice that observed. For example, in the I-6 x I-1(BC2) cross, BC2F1 families segregating for umc139 locus showed a 15.4% increase in seedling emergence over the unmodified F1 (64–48.6%). Thus, if we assume we had all the plants heterozygous for umc139, an increase of seedling emergence of 30.8% is expected over the unmodified F1 hybrid.
Similar results were observed in the I-5 x I-1(BC2) population, although not all of the marker genotypes were recovered in this cross. The BC2F1 families that contained the donor allele linked to php200689 and both alleles linked to umc139 and bnl9.08 displayed significant increases over the BC2F1 families lacking donor alleles. In the third population I-5 x I-6(BC2), seedling emergence was enhanced by the beneficial QTL allele linked to umc139 in all marker combinations. BC2F1 families that were segregating for donor QTL alleles linked to umc139/php200689 showed the greatest enhancement in seedling emergence. In the absence of other donor alleles, the beneficial QTL allele linked to bnl9.08 was not found to exert significant effects in this genetic background, as was observed with php200689 in the first population. This may be the result of interloci interaction between these alleles and other QTL. Since the tested families were in an early generation of backcrossing, large DNA fragments from the donor parent are likely to be present in the segregating population. These fragments could contain other alleles conferring unanticipated effects as proposed by Zeven et al. (1977). With two generations of backcrossing followed by hybridization to the non-recurrent commercial inbreds the percentage of donor genome present in the BC2F1 plants is expected to be 6.25%. This suggests that background could play a minor role as a source of other alleles that could confound the results presented in this study.
When the data were pooled across the three populations (Table 5), seedling emergence was enhanced with the introgression of any single or combination of the donor marker loci compared with families lacking all donor alleles. The magnitude of the effect of these beneficial alleles appeared to be smaller in families displaying relatively improved emergence. BC2F1 families possessing both the beneficial QTL alleles linked to the umc139/php200689 markers had the highest mean emergence followed by umc139 alone and the combination of umc139/bnl9.08/php200689 loci. On the average, families containing all three-donor marker alleles did not perform greater than the combinations of two donor loci. The expected increase in seedling emergence over the unmodified F1 (35.8%) across all three populations, if all plants were heterozygous for umc139, bnl9.8, php200689, um139/bnl9.08, umc139/php200689, bnl9.08/php200689, and umc139/bnl9.08/php200689 would be expected to be 31.4, 8.4, 6.4, 24.2, 52.4, 26.0, and 23.4%, respectively. Mean emergence of the BC2F1 families lacking any of the donor marker alleles was lower than the mean emergence of the three F1 hybrids suggesting the presence of DNA fragments from the donor parent that imposed negative effects, directly or epistatically. In the second analysis, the effects of the marker-QTL backgrounds were eliminated by grouping families on the basis of their genotypes at the marker loci. When the families were classified in this way, the beneficial effect of the QTL linked to bnl9.08 and php200689 on seedling emergence in families with one or the other of these donor alleles was not as dramatic as reported in Table 4.
Understanding the genetic background and QTL expression can help geneticists develop breeding schemes with greater likelihood of success. Variable effects of introgressed QTL in previous studies can be attributed to factors such as large genetic diversity between populations tested (Wang, 1997; McKendry et al., 1996), physiological epistasis among paternally inherited QTL alleles (Danzmann et al., 1999), or an allelic copy of the introgressed QTL may exist in the target population or cultivar (Toojinda et al., 1998; Yousef and Juvik, 2001b). Romagosa et al. (1999) suggested that the economic utility of beneficial QTL includes understanding the physiological and agronomic effects of any QTL and its environmental interactions. Zhu et al. (1999) suggested that for traits such as grain yield, MAS efforts may be better targeted at determining optimum combinations of QTL alleles rather than pyramiding alleles detected in a reference mapping population. In this study, a combination of QTL linked to umc139 and php200689 markers resulted in the highest seedling emergence compared with other combinations including all three of the beneficial marker-QTL alleles together.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
Received for publication August 23, 2000.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| The SCI Journals | Agronomy Journal | Vadose Zone Journal | |||
| Journal of Plant Registrations | Soil Science Society of America Journal | ||||
| Journal of Natural Resources and Life Sciences Education |
Journal of Environmental Quality |
||||
