| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Dep. of Community Health Sciences (Human Nutrition), School of Public Health, Univ. of California, Los Angeles, CA 90095
b International Maize and Wheat Improvement Center (CIMMYT), P.O. Box MP163, Harare, Zimbabwe
c Dep. of Plant Breeding, 252 Emerson Hall, Cornell Univ., Ithaca, NY 14853-1902
* Corresponding author (mes25{at}cornell.edu)
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Abbreviations: GCA, general combining ability • SCA, specific combining ability
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
As staple crops contribute substantially to daily caloric intake among people in developing countries, there has been a resurgence of interest in addressing human malnutrition through breeding of staple crops, specifically to address micronutrient malnutrition (Gregorio et al., 2000; Monasterio and Graham, 2000; Beebe et al., 2000). Maize comprises an average of 30 to 50% of the daily caloric intake of people in most southern African countries (FAO, 2001). In Zambia, and potentially in other areas of southern Africa, poorer people tend to derive an even greater proportion of their daily calories from cereal sources (Kumar, 1994). Iron- and zinc-enriched, or biofortified, maize would serve as a logical vehicle for providing iron and zinc in the diets of people in southern Africa.
Research on the nutritional quality of maize for human consumption is not a new approach to addressing human malnutrition. Nutrition-related research in maize was seriously initiated with work on quality protein maize in the 1960s and 1970s to address protein-energy malnutrition (Pradilla et al., 1972). Early research on micronutrient concentrations in crops seems to have emerged from concern about contaminants in the soil and their possible uptake by crops (Arnold and Bauman, 1976). The two areas of research converged when differences in the micronutrient concentrations of opaque-2 and normal maize led to the study of micronutrient concentration as a possible selection criterion for high-lysine maize (Arnold et al., 1977). Though mineral concentrations did not improve selection for lysine-rich maize, opaque-2 kernels had significantly greater iron and zinc concentrations than normal kernels, although on a kernel weight basis, the opaque-2 kernels were significantly greater only for zinc (Arnold et al., 1977). It is important to note that the materials chosen to be studied in the 1960s and 1970s for grain micronutrient concentration, all U.S. Corn Belt-adapted lines, were not adapted to areas where human micronutrient malnutrition would be addressed with improved varieties of maize. Furthermore, the breeding materials studied were a sample of commonly used U.S. Corn Belt inbreds but were not specifically chosen because of their particularly high or low concentrations of micronutrients (Gorsline et al., 1964; Arnold and Bauman, 1976).
Research on the genetics of kernel micronutrient density of maize was reported in the 1960s and 1970s (Gorsline et al., 1964; Arnold and Bauman, 1976). Additive gene action was more important than nonadditive gene action for both grain and ear-leaf iron and zinc concentrations (Gorsline et al., 1964). This research indicated that there was no relationship between ear leaf and grain nutrient concentrations for either iron or zinc, leading to the hypothesis that the "genetic factors controlling element concentration in these two plant tissues (were) different and independent" (Gorsline et al., 1964). Similar findings were described by Arnold and Bauman (1976), who reported that general combining ability (GCA) effects for grain iron and zinc concentrations were more important than specific combining ability (SCA) effects.
All studies on the genetics of kernel micronutrient density of maize were conducted under recommended input conditions. Most farmers in southern Africa use very low-input farming systems necessitated, in part, by the high cost of fertilizer (Jones and Wendt, 1995; Edmeades et al., 1997). Maize yields in southern Africa are also frequently affected by drought (Zambezi and Mwambula, 1996). Drought and nitrogen stress affect the genetics of grain yield in maize (Betrán et al., 1997) and similar effects on grain micronutrient concentration need to be assessed if nutritionally improved maize varieties are to be recommended to farmers in this region.
Through increased understanding of the general and specific combining abilities of these inbreds for kernel iron and zinc concentration and content, breeders in the region can more efficiently develop hybrids and open-pollinated varieties with increased kernel iron and zinc density. The objectives of the diallel study presented here were to evaluate the inheritance of grain iron and zinc density of maize breeding lines adapted to southern Africa, since there is a high percentage of people in this region with iron, and possibly zinc, deficiency who are predominantly maize consumers. The inbreds used represent the highest and lowest fractions of grain micronutrient concentration among a large sample of current breeding materials in southern Africa (Bänziger and Long, 2000).
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
The 14 parent lines and 91 F1 hybrids were evaluated in adjacent trials at six locations in Zimbabwe with two replications per location during the 1999–2000 growing season. Parent trials were randomized in an 
-lattice design (2 x 7) and the F1 trials were randomized in an -lattice design (12 x 8), both with one-row plots. Potential intergenotypic competition from single-row plots is unlikely to affect the results of this diallel (Bänziger et al., 1995). Plot areas for parent and F1 trials were 3.375 m2 (4.5-m rows at 0.75-m row spacing) except for the Harare low-N trial site (one parent trial and one F1 trial) that had plot areas of 3 m2 (4-m rows at 0.75-m row spacing). All trials were bordered with maize hybrids or inbred lines of similar vigor, and grown at a planting density of 5.3 plants m–2.
Trial Sites
The trial sites were at Glendale (17.08°S, 31.03°E, 1200 m above sea level, asl; commercial farm), Kadoma (18.32°S, 30.90°E, 1155 m asl; research station of the Ministry of Agriculture), Matopos (20.38°S, 28.50°E, 1347 m asl, research station of ICRISAT), and three sites in close proximity to Harare (17.80°S, 31.05°E, 1468 m asl): ART Farm (research station of the Agricultural Research Trust), and Harare high N and Harare low N (both at the research station of CIMMYT at the University of Zimbabwe Farm). The diallel F1s at Harare high N were hand-pollinated with a bulk of pollen from within the row to allow an assessment of xenia, while all other trials were allowed to open-pollinate.
Long-term seasonal rainfall averages approximately 800 mm at Glendale, 700 mm at Kadoma, 550 mm at Matopos, 800 mm at ART Farm, and 750 mm at Harare. Potential evapotranspiration averages around 700 mm (Glendale and Kadoma), 750 mm (Matopos), and 650 mm (ART Farm and Harare), indicating that crop water deficit is most likely at Matopos followed by Kadoma, while the risk of drought is low at the other sites. The Glendale, Kadoma, ART Farm, and Harare sites had reddish clay soils while the Matopos site had black clay soil.
The Harare low N site received no N fertilizer application, and had been depleted of nitrogen for several years before the trial by continuously growing maize (main season) and irrigated wheat (dry season) without N fertilizer application and removing all biomass after each season. All other sites received N applications following site-specific fertilizer recommendations. Planting dates in 2002 were 15 November (Harare low N), 19 November (ART Farm), 24 November (Harare high N and Kadoma), and 2 December (Glendale and Matopos). Phosphorus was applied to all trials following site-specific recommendations. The pre-emergence herbicides atrazine (6-chloro-N2–ethyl-N4–isopropyl-1,3,5-triazine-2,4-diamine) 50 flowable (atrazine 50%) at 4.5 L/ha and Dual Magnum [Syngenta Crop Protection, Inc., Greensboro, NC; 960EC S-metolachlor 96%—2-chloro-N-(2-ethyl-6-methylphenyl)-N-(2-methoxy-1-methylethyl)acetamide] at 1.8 L/ha were applied at all trial sites. Basagran 3 L/ha [bentazon 48%—3-isopropyl-1H-2,1,3-benzothiadiazin-4(3H)-one 2,2-dioxide] was applied 6 wk post-emergence to control broadleaf weeds while hand weeding was used to control grass weeds.
Measurements
After kernel maturation and plant dry down, ears were hand harvested and dried in the sun or in charcoal burning ovens to lower post-harvest grain moisture content. Ears were shelled and the grain bulked from each plot. Grain moisture was measured with a moisture meter. Grain yield was recorded at 125 g kg–1 H2O. Grain yield was not determined for materials at the Harare high-N site where plants were hand-pollinated. For micronutrient analyses, grain was rinsed in distilled water for <15 s to remove surface contamination resulting from ears being exposed to soil and other contaminants during the harvesting and drying processes. Grain was dried in an oven overnight at 80°C and a sample of >40 kernels per plot was ground into a fine powder in an aluminum mill.
The grain from the parent trials was analyzed for grain micronutrient concentrations with an Inductively Coupled Plasma Optical Emission Spectrometer (ICPOES) (ARL, Switzerland) at the University of Adelaide in Australia. The grain harvested from the F1 trials was analyzed for grain micronutrient concentration at the USDA Plant, Soil and Nutrition Laboratory at Cornell University using an Inductively Coupled Argon Plasma Emission Spectrometer (ICPES) (Thermo Elemental, Franklin, MA), a system equivalent to the ICPOES. No inter-laboratory differences were detected.
Samples for the ICPES analysis were digested with 1 mL concentrated nitric acid (HNO3) at 80°C for 1 h and were then digested to dryness at 100°C. One additional mL of nitric acid was added, the temperature was increased to 150°C, and samples were digested to dryness. Samples were further digested to dryness with a 50% perchloric acid (HClO4)/50% HNO3 solution on a heated block with the temperature gradually increased overnight to 240°C to ensure complete digestion. Samples were then dissolved in 10 mL of 5% HNO3 and left for one hour before analysis. Once kernel iron and zinc concentrations were determined, 100-kernel weights were recorded for all trials, and per kernel iron and zinc content were determined by dividing iron or zinc concentrations by 100-kernel weights.
Statistical Analysis
Grain yield, per kernel iron and zinc content, and flour iron and zinc concentration mean values for each genotype were adjusted using spatial statistics developed by Federer (1998). Spatial variation patterns differ for each trait and each field, and as this variation is not known in advance, we used exploratory trend analyses to model the variation present for each trait in each field (Federer, 2003). We used orthogonal polynomial regression coefficients to account for spatial variation and significant regression effects were identified (Bozovich et al., 1956; Federer, 1998; Federer and Wolfinger, 1998; Federer et al., 2001). The resulting spatially adjusted means for each trait at each trial site were used in a multisite diallel analysis.
As trial sites ( = environments) were considered random effects, the error terms in the analysis of variance F-tests for genotypes, GCA, and SCA were the interaction terms with the environment for each of these components. Since the spatial analysis model generated only one adjusted mean value per genotype per site, within-site error variation was not available for the diallel analysis. Therefore, to calculate the F-tests for the interaction terms in the diallel (genotype x environment, GCA x environment, and SCA x environment), the error terms were calculated by averaging the error sums of squares from the spatial analyses across all sites. The degrees of freedom for the interaction F-tests were calculated by summing the degrees of freedom for all error terms from the spatial analyses across all sites (Federer, personal communication). Parent-line mean comparisons were analyzed by the Tukey-Kramer method across all six trials.
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
Combining Ability Analysis
The analysis of variance for flour iron and zinc concentration and per kernel iron and zinc content across all six hybrid trials showed that the variation among genotypes and the genotype x environment interaction were highly significant (Table 3). When the genotype sum of squares was partitioned into GCA and SCA effects, only GCA effects were found to be significant for iron and zinc flour concentration and per kernel iron content, while both GCA and SCA effects were significant for per kernel zinc content. As expected, the analysis of variance for grain yield indicated that both GCA and SCA effects were significant sources of variation among genotypes.
|
|
The 14 lines in this study were selected from among over 1000 inbreds on the basis of grain iron and zinc concentration measured in test-crosses of those lines to common testers. As per se line performance significantly differed in this study and GCA effects emerged as most important, it follows that breeding has the potential to permit selection of maize with higher kernel iron and zinc density, though the results of this study cannot be extrapolated to other populations of maize. GCA effects are a general indicator that additive gene action is important in the inheritance of these traits, while SCA effects are an indicator of the importance of nonadditive gene action. The significant contribution of GCA in explaining genotypic variation indicates that, in general, per se line performance should be a good indicator of hybrid performance for flour iron and zinc concentration among these inbreds.
GCA effect estimates for flour iron and zinc concentration and per kernel iron and zinc content across the five high yielding sites were positive and significant for most of the high parent lines, and significantly negative for an even higher number of the low parent lines (Table 5). For the 14 inbred parent lines, the correlation between GCA effect estimates and flour iron concentration was 0.88, while the correlation between GCA effect estimates and flour zinc concentration was 0.92. The strong correlation between inbred flour iron and zinc concentrations and their GCA effect estimates indicates that selection based on per se line evaluation should identify lines that would contribute high micronutrient density to their progeny.
|
At least 78% of the SCA effect estimates across all five high yielding sites for grain yield, flour iron and zinc concentration, and per kernel content were significantly different from zero (data not shown). There was no association between positive or negative SCA effect estimates across the five high yielding sites, regardless of whether hybrids were formed from a high x high, high x low, or low x low parent combination. The large number of significant SCA effect estimates indicates that while per se line selection can be the primary strategy for selecting materials, one still needs to evaluate final cross combinations. Furthermore, this signifies that there may be dominance or interaction effects that could be exploited in the breeding process.
Breeding Potential
Iron and zinc concentrations are higher in the aleurone layer and lower in the endosperm (Bityutskii et al., 2002). Flour micronutrient concentration is, therefore, influenced by kernel size. As the aleurone layer is a single cell layer surrounding the starchy endosperm cells, the surface area (or aleurone cell volume) to endosperm volume ratio is larger for small kernels. Selection for increased flour iron and zinc concentration may, therefore, result in selection of genotypes with small kernels that, in some instances, may also have lower grain yields. Selection for increased flour iron and zinc concentration, while maintaining or even increasing kernel iron and zinc content, would effectively prevent the selection of genotypes with small kernels. Parent Line 3 proved to be the most promising line in this regard, as GCA effect estimates indicate that it had no significant negative effect on grain yield and it had a significantly positive effect on both per kernel iron and zinc content and flour iron and zinc concentration (Table 5).
Effects of xenia on kernel iron and zinc concentrations could not be examined in this study because of the high variability in kernel micronutrient concentrations among sites. As aleurone traits can be affected by pollen from the male parent, this issue should be further explored.
The materials in this study were selected for both iron and zinc flour concentrations; therefore, these particular materials may not reflect the relationships between grain iron and zinc concentrations in other maize populations. Earlier work with U.S. inbreds did not demonstrate a correlation between grain iron and zinc concentrations (Arnold and Bauman, 1976). In this study, the correlation coefficient of flour iron and zinc concentrations within each trial site ranged from 0.56 to 0.72, while across all six sites, the correlation coefficient was 0.37. The reasonably high within-trial correlations between flour iron and zinc concentration indicate that selection attempts among the southern African inbreds were successful in identifying materials with both high iron and high zinc concentrations. As both iron and zinc deficiencies in humans are of concern in southern Africa, the materials from this study may have real potential to increase intake of these nutrients in the diets of maize consumers.
Impacts of Low Nitrogen Sites on Selection
At the low-N site, the difference between GCA effect estimates for flour iron concentration of high and low lines was not as pronounced (0.8 mg kg–1) as across the five high-N sites (1.7 mg kg–1) (Table 6). The difference between high and low lines was much more similar at low and high N sites for GCA effect estimates for flour zinc concentration: 2.5 mg kg–1 under low N and 2.6 mg kg–1 under high N. Under low N conditions, parent line 14 had a significant positive GCA effect for flour iron concentration, per kernel iron and zinc content, and grain yield with no significant effect on flour zinc concentration (Table 6). Parent Line 14 was originally selected in inbred trials in high yielding environments for its low flour iron and zinc concentration. In inbred line trials in the low-N environment, it also had low flour iron and zinc concentrations (data not shown). Yet in the low-N hybrid trial, it emerged as a highly promising line for the target traits when in hybrid combination. This indicates that materials should be screened in hybrid combinations since the potential of some materials could be overlooked, as with Parent Line 14. Furthermore, when GCA effect estimates were calculated at five high yielding sites combined (Table 5) and individually (data not shown), the GCA effect estimates for Parent Line 14 were not significantly different from zero for the high yielding sites; whereas, the GCA effect estimate was significantly different from zero and positive at the low-N site (Table 6).
|
Eighty percent of the SCA effect estimates of materials at the low-N site were significantly different from zero for grain yield, flour iron and zinc concentration, and per kernel zinc content (data not shown). Seventy-five percent of the SCA effect estimates of materials at the low-N site were significantly different from zero for per kernel iron content. There was no association between positive or negative SCA effect estimates, regardless of whether hybrids were formed from a high x high, high x low, or low x low parent combination. For flour iron concentration, the GCA to SCA ratio of mean square components under low N conditions was 0.12 while for high N conditions (across five sites) it was 0.73. This demonstrates that under low N conditions, parent line performance in inbred trials and hybrid performance for iron concentration can differ markedly, and therefore screening materials in hybrid combinations is essential for identifying materials for low N environments.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
Received for publication December 29, 2003.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
Related articles in Crop Science:
This article has been cited by other articles:
|
|
 |
|
  W. H. Pfeiffer and B. McClafferty HarvestPlus: Breeding Crops for Better Nutrition Crop Sci., December 18, 2007; 47(Supplement_3): S-88 - S-105. [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 | |||
from a diallel analysis of southern African-adapted maize inbreds for flour Fe and Zn concentration, and per kernel Fe and Zn content across five high-yielding sites, and grain yield across four high-yielding sites in Zimbabwe, 1999–2000.