Genetic analysis of mutagen-induced flowering time variation in Arabidopsis thaliana (L.) Heynh

Arabidopsis thaliana (L.) HEYNH. was chosen as a model plant to study the genetic system of flowering time (start of flowering) and, often closely correlated to it, number of rosette leaves (vegetative production). The duration of the vegetative phase is of interest from the point of view of both natural selection (local adaptation) and artificial selection (plant breeding).

Genotypic variation was induced by seed treatment with EMS (ethyl methanesulfonate) and with X-rays. Two pure lines were used as a starting material, viz. the very early flowering Landsberg-'erecta' (line C) and a later flowering mutant line derived from it (line 51). In total 24 flowering time mutants with good fertility and vigour, and of independent origin, were obtained through sector selection in M 1 (for fertility), and individual selection in M 2 , followed by line selection in M 3 and further generations (for flowering time and fertility).

After grouping the 24 lines according to parent of origin, and direction and magnitude of effect, diallel crosses (F 1 and F 2 ) were made within the groups, in order to identify the individual mutations by means of classical Mendelian methods, rather than to describe and analyse the induced quantitative variation by the statistical methods of quantitative inheritance. In nearly all cases, including the small-effect lines, single gene differences with the parent of origin, could be successfully identified with the experimental and analytical procedures used for this purpose.

The analysis of the flowering time mutants led to the following conclusions:1. No significant differences could be detected between EMS and X-rays, with respect to magnitude and direction of effect. There is some indication that X- rays induce more (small-effect late) dominant mutations. With respect to mutant frequency at equal levels of M 2 -fertility, EMS is 2 to 3 times as efficient as X-rays.2. The majority of the mutant lines selected differ in one single recessive gene from the parent of origin. Selections containing two or more unidentified mutations (micromutations) are relatively rare.3. Mutagenic treatment of the medium early line 51, releases variability in both directions: towards early and towards late. However, no earlier mutant lines were obtained from the very early line C. Moreover, line C did not respond to vernalization treatment, and early mutant alleles when transferred to line C did not come to expression. Therefore, it is concluded that line C represents a 'physiological limit' towards early.4. There was no evidence for allelism of mutations of independent origin. Several linkage groups were established. On the one hand, the loci for flowering time in Arabidopsis are not distributed at random over the genome, on the other hand they are not restricted to a few chromosome segments either.5. Except for occasional evidence for non-allelic interaction (epistasis), the effects at the individual loci are additive over loci. Crosses between recessives and intermediates give, in connection with the relative magnitudes of the gene contrasts, interesting numerical examples for the dominance theory of heterosis.6. Pronounced genotype x environment interactions were frequently met: homozygotes effect x season, dominance effect x season, and also genotype x vernalization x season interactions were described.