Molecular markers represent an efficient tool for detecting the changes in allelic frequencies occurring at the loci controlling the response of a quantitative trait subjected to selection. We have investigated the effects of two cycles of divergent selection for leaf abscisic acid (L-ABA) concentration on the frequency of alleles at RFLP loci in a maize population derived from the cross between Os420 (high L-ABA) and IABO78 (low L-ABA).

In 1992, 480 random F2 plants were selfed and evaluated for L-ABA under conditions of partial drought stress at the stage of tassel emergence and pollen shed. In 1993, the F3 families derived from the 30 F2 plants with the highest L-ABA and the 30 F2 plants with the lowest L-ABA were tested in a replicated trial conducted under conditions of partial water deficit. The 8 F3 lines with the highest L-ABA and the 8 F3s with the lowest L-ABA were identified; selection was also conducted within these F3 families. The corresponding F4 families showing, respectively, the highest and the lowest L-ABA were then intercrossed according to the diallel scheme in order to obtain the high L-ABA and the low L-ABA populations. The two populations were evaluated in 1995 in a replicated field trial under conditions of partial water deficit. The high L-ABA population, as compared to the low L-ABA population, showed significantly higher L-ABA concentration (413 vs. 254 ng ABA/g d.w.), thus indicating the effectiveness of the divergent selection procedure.

The two groups of divergently-selected F4 lines were subjected to RFLP analysis with the same probes utilized in the preceding article. So far, the F4 families have been screened with 29 well-spaced probes localized on chromosomes 1, 2, 3, and 4. A significant departure (X2 test) of RFLP allele frequencies from the expected ones was observed in two chromosomal regions (csu133-csu4 on chr. 2 and umc31-umc193 on chr. 4). In both cases, the low L-ABA parental allele (from IABO78) was fixed in the eight low L-ABA families; among the eight high L-ABA families, there was only one family which was heterozygous at each region, while the remainder were homozygous for the high L-ABA parental allele (see Fig. 1 for csu4 on chr. 2). The QTL analysis carried out on 80 random F4 families derived from the same cross (see preceding article) also indicated that the two regions on chromosomes 2 and 4 contain two of the strongest and more stable QTLs for L-ABA. Therefore, the results of the divergent selection support the validity of the QTL analysis.

In order to ascertain with more precision the effects associated with the QTLs for L-ABA on chromosomes 2 and 4, the heterozygosity at these regions will be maintained through three-four selfing cycles. Sets of nearly isogenic lines differing for the parental alleles at the chr. 2 and chr. 4 regions will then be derived and evaluated per se and in hybrid combination under different water regimes for L-ABA, yield, and other physiological traits.

Return to the MNL 71 On-Line Index
Return to the Maize Newsletter Index
Return to the MaizeGDB Homepage