Selection for anther culture ability
--V.H. Beaumont and J.M. Widholm

Anther culture ability is the ability of a genotype to induce microspore division and embryo formation from microspores, once the anthers are plated in appropriate conditions. This trait is strongly genotype-dependent in maize with most genotypes being unresponsive (Petolino and Jones, Crop Sci. 26:1072-1074, 1986). It is generally explained with a few dominant genes (Theor. Appl. Genet. 80:459-464, 1990). The F1 hybrid Pa91 x FR16 was previously found to be responsive and some F2 plants from this cross gave high yields in our anther culture system. Therefore, we started a selection program with the progeny.

All the anthers plated came from donor plants grown in the field (Champaign, Illinois) during the summers 1992 to 1994. The anther culture protocol has been described elsewhere (Beaumont et al., MNL66:114-115, 1992). The variables E/100A (number of embryos produced for 100 anthers plated) and % Resp.P (percentage of plants producing at least one embryo) were recorded. Because of the non-normality of the distributions, we chose non-parametric tests to analyse the data: the Wilcoxon (= Mann-Whitney) test was used to analyse E/100A and the F2 test was used to analyse %RESP.P.

The number of plants and anthers plated are given in Table 1. In 1992, the percentage of responding plants was not recorded for F1 hybrids. The F2 generation was anther cultured in 1992 but we did not select in this generation since the tassels were used for anther culture and the plants could not be self-pollinated. 43 F3 families were produced in 1992 and evaluated for their anther culture ability in 1993. Five F3 families were selected for their anther culture ability (marked F3Sel). Two plants in each selected F3 family were selfed, giving rise to ten F4 subfamilies (marked 5 x 2 in the table). All F4 families from selected F3's were anther cultured in 1994. Results from the best F4 (F4Sel) are given in Table 1.

This progeny was anther cultured during three different years. Therefore, we had to compare the results to a common control. The parents FR16 and Pa91 do not produce enough embryos to be evaluated precisely. The F1 hybrid could have been a good control but the lack of seeds did not allow us to plate anthers in 1993. We chose to compare the results obtained to the genotype HFI (haplodiploid from anther culture of H99 x FR16) during the corresponding year. Since HFI might not react the same way as the PA91 x FR16 progeny, the comparisons obtained must be interpreted cautiously (Fig. 1).

The yields of the 5 F4 families are higher than their F3, for the E/100A (P < 0.1) and %RESP.P (P < 0.001). Altogether, the F4 families gave 36% responding plants (Table 1), but all families produced embryos (100% responding F4 families). Thus, the selected F3's represent a good starting material for further selection. The genotype PFF4#6B seems very interesting since its number of embryos (E/100A) and percentage of responding plants (%RESP.P) are higher than HFI for the same year. These results demonstrate that breeding for anther culture ability is possible. We will continue this program and also cross HFI with PFF4#6B for further selection.

Table 1: Selection for anther culture ability from the cross Pa91 x FR16 (PF). Geno: genotype, E/100A: number of embryos for 100 anthers plated, %RESP. P: percentage of responding plants, ND: not determined.

Figure 1: Anther culture ability of successive generations. The results are corrected assuming HFI to be constant over the years. Sel: selected families (see Table 1). The results obtained in the F3 generation were compared to those obtained in the F4, with the Wilcoxon test for the variable E/100A and the F2 test for the variable %RESP.P. Significant differences are noted. *:P < 0.1; **: P < 0.01; ***: P < 0.001.

Return to the MNL 69 On-Line Index
Return to the Maize Newsletter Index
Return to the MaizeGDB Homepage