The number of days from planting to pollen shed (DPS) is often of critical importance for optimizing crop yields in different environments. The transition from the vegetative to the reproductive phase involves the integration of many environmental and developmental factors. In maize, experimental evidence suggests that genes for plant growth and development are associated into functionally significant clusters (Khavkin and Coe, TAG 93:343-352, 1997). As an example, the 5.06-5.07 region on maize chr. 5 contains structural genes (vp2 and vp7) involved in the synthesis of abscisic acid (ABA) as well as QTLs (quantitative trait loci) for DPS, plant architecture, and pollen growth (Khavkin and Coe, MNL 68:61-62, 1996). A physiological interpretation for a pleiotropic effect on morpho-physiological traits could be attributed to the presence of one or more QTLs controlling the level of particular plant growth regulators (PGRs) influencing the rate of plant development. Little is known about the effects on DPS of naturally-occurring variation in the endogenous levels of PGRs. The overlap of QTLs of different traits provides evidence for a genetic correlation, either by linkage or pleiotropy, among such traits. The objective of this research was to evaluate 151 F3 families of a cross segregating for DPS to identify QTLs for L-ABA and to investigate their overlap with QTLs for DPS evidenced in the same cross by Parentoni (MS thesis, University of Minnesota, 1993).

F3 families of the cross A662 (early flowering) x B73 (late flowering) were evaluated using a 12 x 13 rectangular lattice replicated twice in trials conducted at Rosemount and Waseca, Minnesota. The parent lines and three B73 backcross-derived lines (A679, A680, and A681) were also included. Leaf samples were collected approximately at stages 2 and 3 according to Hanway's scale (Agr. J. 55:487-492, 1963). The concentration of unconjugated ABA was determined using an ABA-specific monoclonal antibody (Quarrie et al., Planta 173:330-339, 1988). The analysis of variance was carried out separately for each one of the four "growth stage x location" (sampling) combinations. Because the lattice design was found more effective than the randomized complete block design, means were adjusted accordingly before checking for normality in their distribution. The linkage map was obtained by Kim (PhD thesis, University of Minnesota, 1992) using 63 PstI RFLP probes. For QTL analysis, the statistical package MAPMAKER-QTL was utilized.

The ANOVA for L-ABA evidenced significant (P < 0.01) differences due to genotypes, growth stages, locations, and their interactions. Due to the significance of the second order interaction, QTL analysis was carried out on each individual sampling. In total, seven unlinked QTLs influenced (LOD > 2.0) L-ABA in at least one of the four samplings. Table 1 reports the main characteristics of the QTLs for L-ABA. In Rosemount, the number of QTLs evidenced at stages 2 and 3 did not vary, although different QTLs were revealed. For L-ABA-2, two QTLs were identified on chr. 3 and 6. The LOD score peak (LOD = 2.47) on chr. 3 was between umc154 and umc10a, while the LOD peak (2.54) on chr. 6 was between umc59a and bnl3.03. In Rosemount, L-ABA-3 was significantly influenced by QTLs on chr. 5 and 8. LOD scores (2.16 and 2.96, respectively) peaked on chr. 5 between umc54 and umc108 and on chr. 8 between umc16b and umc7. A considerable portion (35.8%) of phenotypic variation among F3 families was accounted for by the QTL on chr. 8. In Waseca, the only region with significant effects on L-ABA-2 was on chr. 2, near umc131, while four regions significantly affected L-ABA-3. These four QTLs were localized on chr. 1 (near npi234), chr. 2 (between umc5 and umc88), chr. 5 (between umc54 and umc108, and chr. 8 (between umc16 and umc7). The R2 value varied from 10.3 to 34.1%. The support intervals of the two QTLs on chr. 5 and 8 overlapped with those of the QTLs in the same regions which showed significant effects on L-ABA-3 in Rosemount. With the exception of the first sampling in Rosemount, the alleles increasing L-ABA were those of B73, the late parent line of the cross. In both locations, the absolute effect of allelic substitution was of greater magnitude at stage 3. This, in turn, was paralleled by a substantial increase from stage 2 to stage 3 in the percentage of phenotypic variation for L-ABA that was accounted for by each QTL. Other authors have reported that different QTLs control L-ABA at subsequent growth stages in maize (Lebreton et al., J. Exp. Bot. 46:853-865, 1995; Sanguineti et al., Maydica 41:1-11, 1996). Another important factor which could influence QTLs for L-ABA at subsequent growth stages is the level of water stress experienced by the plants in the time-period prior to sampling. It is worth mentioning that the QTL which was evidenced in Waseca for L-ABA-2 between markers umc131 and umc5 on chr. 2, was mapped very close to the position reported for a major QTL influencing L-ABA in two different maize populations (Lebreton et al., J. Exp. Bot. 46:853-865 1995; Landi et al., Proc. XVII Conf on Genetics, Biotechnology and Breeding of Maize and Sorghum, Thessaloniki, 65-70, 1997). It is likely that other QTLs for L-ABA went undetected due to the low marker density of some regions (eight intervals between adjacent markers were longer than 40 cM) and also in consideration that we estimate that our map covers ca. 75% of the maize genome.

The only overlap between support intervals of the QTLs for DPS and the QTLs for L-ABA consistent in both locations occurred on chr. 5. In this case, the QTL for DPS peaked near umc54, while the QTL peak for L-ABA-3 was localized 16 cM away from umc54. The allele increasing L-ABA was contributed by B73, which also contributed the allele delaying flowering at the nearby QTL for maturity. In Rosemount, an overlap between QTLs for L-ABA-2 and DPS was evidenced on chr. 3, near umc154. The presence of QTLs for L-ABA and DPS was also evidenced on chr. 8. However, in this case, the peaks of the two QTLs were ca. 50 cM apart. These results provide little evidence supporting an association, either by pleiotropy and/or linkage, between QTLs for L-ABA and QTLs for DPS in maize under field conditions. Accordingly, the correlations between L-ABA and DPS were negligible in magnitude in all four samplings (r from 0.08 to 0.18).

Table 1. Main characteristics of the QTLs for leaf ABA concentration in 151 (A662 x B73) F3 families sampled in Rosemount and Waseca at growth stages 2 and 3.
 

Sampling	Chrom.	Flanking markers	LOD	U(1)	R2
Rosemount-2	3	umc154-umc10a	2.47	-1.45	10.2
Rosemount-2	6	umc59a-bnl3.03	2.54	-5.13	33.2
Rosemount-3	5	umc54-umc108	2.16	4.18	7.7
Rosemount-3	8	umc16b-umc7	2.96	18.16	35.8
Waseca-2	2	umc131-umc5	3.42	5.58	19.3
Waseca-3	1	npi234-umc23A	4.14	5.11	34.1
Waseca-3	2	umc5-umc88	3.73	16.04	21.9
Waseca-3	5	umc54-umc108	2.26	5.62	10.3
Waseca-3	8	umc16b-umc7	2.37	13.85	17.8

(1): effect (in ng ABA/g d.w.) of substituting a B73 allele for an A662 allele.

Return to the MNL 72 On-Line Index
Return to the Maize Newsletter Index
Return to the MaizeGDB Homepage