| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Dep. of Agronomy, Univ. of Wisconsin, 1575 Linden Drive, Madison, WI 53706
b Dep. of Genetics, North Carolina State Univ. Raleigh, NC 27695-7614
c Syngenta Biotechnology, Inc., Research Triangle Park, NC 27709
* Corresponding author (smkaeppl{at}facstaff.wisc.edu)
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Abbreviations: G x E, genotype x environment • LA, leaf angle • LOD, likelihood of odds ratio • QTL, quantitative trait loci • RFLP, restriction fragment length polymorphism • RIL, recombinant inbred line • SSR, simple sequence repeat • TBA, tassel branch angle • TBN, tassel branch number
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Tassels can affect grain yield by reducing light interception into the canopy as well as by utilizing carbohydrate resources. A predominantly negative effect of tassel shading on maize yield was confirmed experimentally by Hunter et al. (1969). In this study, detasseled plants yielded 19% more than plants with intact tassels or plants that had tassels cut off and then reattached. Since the plants with intact tassels and the plants with reattached dead tassels both exhibited the same yield reduction, this study showed that shading is the predominant effect of the tassel on maize yield. Lambert and Johnson (1978) also demonstrated the yield enhancing effects of detasseling and tassel branch removal with detasseled plants having a 5% yield increase and debranched tassels having a 2% yield increase over the control.
Leaf angle effects on maize yield have also been documented. Pendleton et al. (1968) compared liguleless2 and normal hybrids and showed up to a 41.2% yield increase in upright leaf genotypes relative to normal isogenic hybrids, depending upon plant density. Lambert and Johnson (1978) evaluated the effect of leaf orientation and plant density on maize yield using hybrids B14 x Oh43 and Oh43 x R177 with normal, liguleless1, and liguless2 genotypes at planting densities of 60 000, 75 000, and 90 000 plants ha-1. The liguleless2 hybrids had significantly higher yields relative to the normal leaf hybrids at the two highest planting densities. The studies by Pendleton et al. (1968) and Lambert and Johnson (1978) implied that the upright leaf angles of the liguleless2 hybrids result in higher yield by allowing greater light penetration into the canopy.
Several studies have shown that tassel traits have a narrow sense heritability ranging from 46 to 89%. Using a cross between BSSS-11 and BSSS-26, Mock and Schuetz (1974) estimated a minimum of eight genetic factors involved in determining maize tassel branch number (TBN), a primary determinant of tassel size. Heritability estimates were reported to be 0.53 on a single plant basis (broad sense) and 0.89 on an F3 family mean basis (narrow sense). Similar heritability estimates were found by Schuetz and Mock (1978) and by Geraldi et al. (1985). In an analysis of the inheritance of tassel traits in the Illinois High Oil x Illinois Low Oil (Early Maturity) cross, Berke and Rocheford (1999) found six QTL associated with tassel branch angle (TBA), three QTL associated with TBN, and seven QTL associated with tassel weight. Highly significant phenotypic and genotypic correlations were observed between tassel branch number and tassel weight with two QTL common between the traits. No heritability estimates or putative QTL for leaf angle in maize have been reported to date.
As the literature supports a strong relationship between yield and LA, TBA, and TBN, it is of interest to study the inheritance of these leaf and tassel traits. The objectives of this experiment were to identify genomic regions associated with the morphological traits of TBN, TBA, and LA and to identify any phenotypic correlations and overlapping QTL which may exist between these traits in a B73 x Mo17 recombinant inbred population. The parents of this population were selected based on the knowledge that B73 has upright leaves and tassel branches and low TBN while Mo17 has more horizontal leaves and a larger number of tassel branches.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
Data Collection
Data were collected approximately 5 wk after pollination on 29 and 30 August at Madison, 7 and 8 September at Arlington, and 12 and 13 September at Marshfield. Tassel branch number was determined by counting the total number of tassel branches per plant, including primary and secondary branches. Tassel branch angle was measured in degrees by visually determining the average angle from vertical of all tassel branches and measuring that single angle with a handheld clinometer. Leaf angle was determined for all leaves above the uppermost ear as the angle of each leaf from a plane defined by the stalk below the node subtending the leaf. This was necessary because, in some cases, the angle of the internode subtending the leaf was not exactly perpendicular to the plane of the soil. Tassel branch number, TBA, and LA were scored on two plants per plot in each environment. Plants measured were randomly chosen within a row, but the first, the last, and the adjacent plants were not used. While all plants within a RIL are genetically identical, two plants were measured to reduce the effect of any environmental influences within the plots.
Data Analysis
Tassel branch number, TBA, and LA mean values were calculated for each plot in the three environments for each RIL and the parents as the mean of the values of the two plants measured in each plot. The distribution of phenotypes based on means of genotypes across environments was not normal as tested for normality by the Sewell Wright test statistic in the SAS univariate procedure (SAS, 1990). Natural log, log 10, square root, square root (Y + 0.5), and Box-Cox transformations were tested for their ability to improve normality. The best transformation for a normal distribution was obtained by the square root transformation for LA and a natural log transformation for TBN. Tassel branch angle could not be transformed to fit a normal distribution at a significance level of P 
0.05, so untransformed data were used for analysis. Using the General Linear Model procedure of SAS with parents analyzed separately from the RILs, we conducted analysis of variance for TBN, TBA, and LA with the following linear model: ß = 
+ 
i + 
j(i) + 
k + ik + 
, where ß represents the phenotypic mean of a genotype, is the effect of the ith environment, is the effect of the jth replication in the ith environment, is the effect of the kth genotype, and represents residual error with all main effects being random with inferences not limited to these three sites or 180 RILs (SAS, 1990). The results were used to calculate narrow sense heritability by means of the following equation: 
2G/
where 2G represents genotypic variance, 2GE represents genotype x environmental variance, and 2e represents error variance. Transgressive segregation in the recombinant inbred lines was studied by determining whether any RILs were significantly more diverse than the parental values at an LSD0.05. Phenotypic correlations were calculated among traits by the mean values for each genotype combined across environments.
QTL analysis was conducted with the PlabQTL Version 1.1 mapping program (Utz and Melchinger, 1996) with means combined across environments for TBA and TBN, and with the Arlington and Marshfield environments combined for LA and the West Madison environment individually for LA. Composite interval mapping employing the covariate SELECT option of PlabQTL was performed for detection of QTL. This option uses step-wise multiple regression to select cofactors. For QTL model building a LOD threshold of 3.56 was used, corresponding to an experiment-wise error rate of P 0.05 and an individual test-error rate of 0.000275 based on analysis of 162 intervals. The additive effect of a marker was calculated by PlabQTL as [(mean of the homozygous Mo17 class - mean of the homozygous B73 class)/2]. To convert these values from transformed to original data units, the mean of the transformed data at the nearest marker was added to the reported additive value. This was reverse-transformed to reflect the original values and subtracted from the mean of the original data. This transformation provided the additive effect of each QTL in units of the original trait measurement. QTL support intervals are calculated as the point along the significance peak at which the LOD score is 1.0 units less than the peak LOD score. The phenotypic variance (2p) explained by a single QTL was estimated by the square of the partial correlation coefficient (R2). The phenotypic variance (2p) explained by the QTL model was estimated by the adjusted correlation coefficient (R2adj), which accounts for the number of predictors in the QTL model. QTL x environment interactions were tested by the PlabQTL ENV option for all traits at the P 0.05 significance level. Additive x additive epistatic interactions were identified by the Model A AA statement in the PlabQTL program at the P 0.05 significance level.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
Significant Variation among Recombinant Inbred Lines Found for Leaf and Tassel Traits
Narrow sense heritability estimates were 0.73 for TBA, 0.81 for TBN, and 0.71 for LA. Mean values of RILs across locations ranged from 5 to 62.5° for TBA, 3.0 to 28.8° branches for TBN, and 7.5 to 69.9° for LA (Table 1). Transgressive segregation was observed for TBN with nine RILs significantly greater than B73. For TBA, two RILs had significantly greater angles than Mo17. Three RILs significantly exceeded the mean of Mo17 for LA.
No significant G x E interaction was detected for TBN or TBA, but the interaction was significant for LA. Subsequent analysis indicated that genotypes responded similarly at Arlington and Marshfield, but a narrower range of phenotypic scores at West Madison resulted in G x E interactions between the Arlington and Marshfield data with data from that location. The G x E interaction was due to differences in the magnitude of phenotypic differences and not due to crossover interactions. QTL analysis for LA was, therefore, done with an average across Arlington and Marshfield (LAA,M) and separately for West Madison (LAW).
Phenotypic correlations were calculated for the three traits from means across locations for TBA and TBN and for LA at Arlington and Marshfield and LA at West Madison (Table 2)
. LAA,M and LAW were significantly correlated at P 0.01 with an r2 of 0.81. Significant correlations (P 0.05) were found between TBA and TBN (r2 = 0.17), between TBN and LAA,M (r2 = 0.46), and between TBN and LAW (r2 = 0.47). The correlation between TBA and TBN is consistent with parental means and the general observations that Mo17 has fewer tassel branches and more horizontal tassel branch angles than B73.
|
Detection of QTL for Leaf and Tassel Traits
Test for QTL x environment interactions were calculated for TBA, TBN, and LAA,M Estimates of QTL x environment interactions were not significant for TBA or TBN, so RIL values were combined across locations to give the greatest inference for mapping. No significant QTL x environment interaction was detected with the LA data from the Arlington and Marshfield environments; therefore, QTL analysis was conducted for the West Madison leaf angles individually and Arlington and Marshfield leaf angles combined. No significant epistatic interactions were found for any of the traits.
QTL were detected for each of the traits measured in this study. Three QTL were detected for TBA accounting for 35.6% of the phenotypic variation (Table 3) . Two QTL were detected on chromosome 5 with the Mo17 allele contributing greater TBA. LOD1.0 confidence intervals did not overlap for these QTL, indicating that chromosome 5 likely contains two QTL for this trait.
|
Analysis of LA data from Arlington and Marshfield revealed seven QTL that explained 39.2% of the total phenotypic variation. A QTL on chromosome 7 had an LOD score of 12.59 and a partial r2 of 27.7%. This QTL had an additive effect of 3.10° with the Mo17 allele causing a greater LA. QTL on chromosomes 2, 5, and 7 were identified for LAW, and these QTL accounted for 31.2% of the phenotypic variation (Table 3). The QTL for LAW on chromosome 5 had overlapping support intervals for a QTL detected for LAA,M. While the QTL on chromosome 1 near umc58 was not significant for LAW, LOD plots (Fig. 2) indicate a QTL peak for LAW below the threshold at an LOD of 2.8 which is at the same position as the QTL for LAA,M. Also on chromosome 1, a QTL for LAW with a maximum LOD score of 2.91 appears to fall directly under the QTL for LAA,M near marker phi120 which had a LOD score of 9.03 and explained 20.6% of the phenotypic variation for this trait (Fig. 2).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
From the standpoint of light interception at current planting densities, the ideal maize genotype would have a smaller tassel, sufficient only for reproduction and seed set, and average LA in the range of 10 and 19. Selection for all favorable (few tassel branches and upright leaves) alleles identified in our QTL study has the potential to produce a plant with three tassel branches and 9° leaves in the B73 x Mo17 genetic background. Clearly, the optimum leaf angle is greater than zero, so a plant ideotype for LA must be based on factors including average light angle and plant density. Duncan (1971) determined that the optimum maize plants with vertical, rather than horizontal, leaves maximized photosynthetic activity when average leaf-area index (LAI) was greater than 3.0. Using computer simulations, we found that, at LAI greater than 3.0, vertical leaves in the upper layers of the crop canopy and horizontal layers in the lower canopy gave the greatest values for photosynthesis for maize leaf types. As LAI increased, a greater proportion of vertical than horizontal layers would be needed for maximizing photosynthesis. By understanding both the physical and genetic parameters needed for maximum light interception, precise engineering of maize plant architecture to give the optimum LA patterns and tassel traits could be accomplished. Previous field observations in this population suggested a pattern of more erect leaves immediately below the tassel with lower leaves becoming increasingly horizontal.
A viable tassel is required for adequate pollination both in the production of hybrid seed as well as in grain production fields, and breeders have conflicting interests in which tassel traits to select for. From the standpoint of light interception, a smaller tassel is best. However, tassel size may be particularly important in stress environments where pollen shed is often reduced. DuPlesis and Dijkhuis (1967) and Berbecel and Eftimescu (1973) found that temperature and moisture stress before and during flowering caused the time between pollen shed and silk emergence to be extended. If the stress is severe, the majority of pollen may be shed before the silks first appear, causing barrenness and poor grain filling. By selecting for tassel traits, breeders must balance the shading effect of the tassel with the need for adequate pollen, particularly in stressed environments where silking is delayed.
The effect of TBA on light interception has not been studied, although it seems reasonable that tassels with very upright branches and a cylindrical shape would minimize the amount of shading per unit of biomass. Therefore, altering tassel angle may be a strategy to minimize shading without sacrificing tassel size. Fixation of the QTL found in this study would potentially allow tassel angle to be decreased up to 8° relative to the tassel of Mo17. The narrow angle allele of the QTL on chromosome 4 is from Mo17 indicating that there is genetic potential to improve even upon the upright tassel of B73.
By understanding both the physical and genetic parameters necessary for maximum light interception, precise engineering of maize plant architecture to provide the optimum LA patterns and tassel traits could be accomplished. This study has identified chromosome regions controlling the inheritance of LA, TBA, and TBN in a B73 x Mo17 recombinant inbred population of maize. Future research utilizing genetic stocks as tools to define optimum leaf angles and tassel sizes on the basis of grain yield and seed set under stress is necessary to refine our models of maize ideotype for leaf and tassel traits. In addition, other traits such as plant height and leaf area will be necessary to consider in these modeling exercises. Research such as that reported in this paper provides pieces of information which will ultimately cumulatively define optimum plant morphology and provide a genetic explanation for phenotypic variation observed among maize germplasm. Rethinking optimum maize morphology may lead to more drastic changes in the structure of the maize plant through breeding. The phenotypic plasticity of maize is clearly great enough to produce a vast array of commercial products. The key to these innovations lies in combining genetic tools with physiological and production research to define optimum phenotypic targets.
|
ACKNOWLEDGMENTS |
|---|
Received for publication September 21, 2001.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  M. Gonzalo, T. J. Vyn, J. B. Holland, and L. M. McIntyre Mapping Density Response in Maize: A Direct Approach for Testing Genotype and Treatment Interactions Genetics, May 1, 2006; 173(1): 331 - 348. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| 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 |
||||
3.56 was used for QTL detection.