Mol. Gen. Genet. 231: 345-352. (1992)
Endogenous and environmental factors influence 35S promoter methylation of a maize A1 gene
construct in transgenic petunia and its colour phenotype.

Meyer, P., Linn, F., Heidmann, I., Meyer, H.Z.A., Niedenhof, I., and Saedler, H. 1992.


30,000 transgenic petunia plants carrying a single copy of the maize A1 gene, encoding a dihydroflavonol reductase, which confers a salmon red flower colour phenotype on the petunia plant, were grown in a field test. During the growing season plants with flowers deviating from this salmon red colour, such as those showing white or variegated phenotypes and plants with flowers exhibiting only weak pigmentation were observed with varying frequencies. While four white flowering plants were shown at the molecular level to be mutants in which part of the A1 gene had been deleted, other white flowering plants, as well as 13 representative plants tested out of a total of 57 variegated individuals were not mutants but rather showed hypermethylation of the 35S promoter directing A1 gene expression. This was in contrast to the homogeneous fully red flowering plants in which no methylation of the 35S promoter was observed. While blossoms on plants flowering early in the season were predominantly red, later flowers on the same plants showed weaker coloration. Once again the reduction of the A1-specific phenotype correlated with the methylation of the 35S promoter. This variation in coloration seems to be dependent not only on exogenous but also on endogenous factors such as the age of the parental plant from which the seed was derived or the time at which crosses were made.


Due to the production of an easily visible flower colour the A1 gene was expected to provide a suitable monitoring system to isolate transposable elements, which should produce white or variegated phenotypes upon integration into the A1 gene. Much to our surprise, however, internal variation in flower pigmentation precluded this approach. It emerged that the A1 system of the salmon red coloured transgenic petunia line RL01-17 can serve as a very sensitive marker for deletions and DNA methylation events in the 35S promoter of the transferred A1 gene construct resulting in changes in colour phenotypes. In the field experiment of summer 1990, four white plants from one capsule, featuring a stable white phenotype in all flowers throughout the season, were shown molecularly to be due to a small deletion within the A1 gene. The frequency of this mutation event was estimated to be approximately 10-4. This is in agreement with mutation frequencies observed in other plant systems (Maddaloni et al. 1990).

However, the most remarkable results obtained in this experiment were the different phenotypes observed and their correlation with methylation of the 35S promoter, suggesting that methylation can occur at any time during the life cycle of the plant:

Premeiotically in the parental plant leading to plants with homogenously white coloured flowers at the beginning of the season that can develop some coloration later on. Capsule no. 109 is an example.

During meiosis in the parental plant. This would result in mixed progeny, some plants having salmon red flowers and others white. Capsule no. 143 is a representative.

Postmeiotically at any time during the development of the progeny plant. Early methylation would result in either a totally white flowering plant or a plant with differing branches, flowers on one branch being salmon red, and on the other white (see Fig. 2 a). At later stages of development sectored flowers would be seen as in Fig. 2b. If the methylation event was frequent and occurred later, variegated phenotypes would be observed as shown in Fig. 2c.

Similar phenotypes have been obtained from glasshouse-grown material and DNA methylation of the 35S promoter has been observed (Linnet al. 1990). While the frequency of the weakly coloured greenhouse plants was about 5% (97 out of 2110 plants), the frequency of such weakly pigmented plants grown in the field increased over the season to about 60% (see Table 1). This increase became apparent after a particular change in weather conditions in the growing season. Even more striking was the observation that the time of seed production was important. The A1 gene construct is obviously almost completely insensitive to DNA methylation in progeny from flowers of young parental plants produced early in the season, but becomes susceptible to methylation within progeny from subsequent crosses (Fig. 7). Similar observations have been made in nontransgenic plants, i.e. for transposable elements in Zea mays where methylation was more pronounced in upper ears and tassels (Federoff and Banks 1988). The most striking example is the behaviour of the maize mutant hcfl06, a pale green homozygous seedling lethal that is caused by the insertion of the Mul transposable element into the promoter region of a gene involved in chlorophyll expression. DNA methylation of Mui sequences results in the expression of the neighbouring hcf gene leading to suppression of the phenotype and formation of green plant sectors. With plant age the size of these sectors increases. Therefore in most tassel tissue Mul is methylated and hence the hcfgene is functional (Martienssen et al. 1990). This shows that DNA methylation in plants is not only seen in the transgenic petunia system but might be a normal component involved in development of a plant.

The observation that DNA methylation increases with age of the meristematic tissue under suitable environmental conditions is interesting and certainly requires more detailed analysis of both endogenous as well as exogenous factors. In any case it seems to be advisable to use first flower progeny if one aims at stable gene expression at least for transgenic plants.

More on methylation and age: Prokofyeva-Belgovskaya (1947)