Theor Appl Genet (1989) 77:325-331
DNA methylation of embryonic carrot cell cultures and its variations as caused by mutation, differentiation, hormones, and hypomethylating drugs.
LoSchiavo, F., Pitto, L., Giuliano, G., Torti, G., Nuti-Ronchi, V., Marazatti, D., Vergara, R., Orselli, S. and Terzi, M.

Discussion

Our results show that in proliferating cell cultures of carrot, there is a positive correlation between exogenously-added auxin and methylation of cytosine: passing from 0.5 to 5 mg/1 2,4-D leads to an increase of MC from 16% up to 70% of total cytosine. These induced changes in the level of methylation are rather fast and reversible.

In the absence of auxin, embryogenesis takes place and shows, in its course, an initial decrease in methylation, followed by an increase during late embryogenesis. This "basal level" of methylation observable in the absence of hormones, with its temporal variation during embryogenesis, cannot be changed without impairing the process: both hypermethylation (as caused by auxin) and hypomethylation (as caused by hypomethylating drugs) immediately and irreversibly stop embryogenesis and the expression of embryo-related functions (one possible exception being greening).

If we pre-treat proliferating cells with ECP, we obtain a drop in the embryogenic potential. If however, the cells are left to recover in the presence of auxin, the embryogenic potential is restored. In a specular way, hypermethylated cells, if left without hormones, lower their level of methylation and embryogenesis can occur.

In a purely formal sense, hypomethylating drugs and 2,4-D are antagonistic. This is shown by the fact that hypermethylated cells are more resistant to ECP and that an abortive differentiation starts when cells are left in the presence of ECP + 2,4-D (not shown).

Azacytidine and ECP do both cause hypomethylation but the target of their action is not the same. Azacytidine acts on DNA transmethylases (Jones and Taylor 1980). ECP in prokaryotes acts as a methyl acceptor (Raugei et al. 1981). In carrot this may not be the case, as the ECP action cannot be completed by methyl donors such as methionine or S-adenosyl-methionine. DNA methyl transferases were not inhibited, in an in vivo assay, by ECP (not shown).

The mutant line E9, which is resistant to ECP and azacytidine shows a higher internal level of IAA (at least in late embryonal stages) and a different metabolism and/or compartmentalization of IAA. At the level of methylation, E9 shows a somewhat higher level in proliferating cell cultures but, perhaps more meaningful, an almost constant level of MC, which does not vary during embryogenesis nor in the presence of ECP. This is at variance with out wt line but also with several other lines of carrot we tested which, in spite of the absolute level of MC which was somewhat variable, showed constancy in behavior in being hypomethylated by ECP and in following the characteristic curve of methylation during embryogenesis.

All these indications lead us to the inescapable conclusion that not all methylations are equivalent. What we would like to propose is the following: each and every tissue has its own level of methylation (we may say "pattern") which is necessary for the maintenance of the differentiative state. This is what we have called "basal state" and might be equivalent to the methylation level we encounter without auxin. The basal level would vary among different tissues and it does so during embryogenesis. If we induce a hypomethylation with drugs, this pattern is disorganized with serious consequences on, e.g., embryogenesis.

There is another type of methylation which occurs, more or less randomly, as a response to auxin. This type of methylation is not tissue-specific and is lost in the absence of its causative hormone. This methylation is, or occurs concomitantly with, a signal to divide, and disorganizes the differentiative patterns. In this light we can understand why in the phenomena involving a change of program, be it acquisition of embryogenic potential (Sung et al. 1984) or wound response, gall formation, tumor formation or adventitious root induction, the presence of auxin is always necessary (for general reference, see Kahl and Schell 1982).

Along these lines, E9, with its higher level of endogenous IAA, resistance to hypomethylation and proneness to tumors, difficulties in regeneration (always accompanied by cell growth) may prove very useful in helping to solve some of the outstanding problems concerning the active forms of IAA and its role in differentiation. We will gladly make it available to everybody willing to tackle these problems.