Activator (Ac) transposes following replication from only one of the two daughter chromatids. It has been suggested that DNA methylation in conjunction with methylation-sensitive transposase (TPase) binding to DNA may control the association of Ac transposition and replication. This mechanism requires that the TPase binding sites within Ac are methylated prior to replication. By restriction analysis of genomic maize DNA with methylation sensitive enzymes it has been shown that the three HpaII sites and the PvuII site at the 3'-end of Ac in the wx-m9::Ac allele are methylated, whereas no methylation could be detected at the 5'-end. In contrast, during the inactive state of Ac in the wx-m9::Ds-cy allele the 5'-end of the element is also hypermethylated (Chomet et al., EMBO J. 6:295-302, 1987; Schwartz and Dennis, Mol. Gen. Genet. 205:476-482, 1986). The TPase binding sites are not accessible by any restriction enzymes, however. We have therefore determined the methylation state of these sites at both Ac ends by genomic sequencing. We used the positive display protocol which is based on the conversion of unmethylated cytosine residues to uracil by bisulfite treatment. This procedure allows the methylation state of individual molecules to be determined (Frommer et al., PNAS 89:1827-1831, 1992). We have meanwhile completed the analysis of the active Ac in the wx-m9::Ac allele, and the analysis of the inactive Ac in the wx-m9::Ds-cy allele is in progress.

The active Ac elements in wx-m9::Ac endosperm exhibit intriguing methylation patterns at their ends and fall into two distinct groups. Half of the elements are unmethylated throughout the 256 residues at the 5'-end (the promoter end). The other half is partially methylated between Ac residues 27 and 92. In contrast, at the 3'-end all Ac molecules are heavily methylated between residues 4372 and 4554, including the CpG sequences within the TPase binding sites (AAACGG). The more internally located Ac sequences and the flanking Waxy DNA are unmethylated. In addition, methylation of non-symmetrical cytosines (C's in other than CpG or CpNpG sequences) in the hypermethylated regions of Ac is common. The observed methylation pattern suggests that the Ac element is a "methylation island" which contains certain regions whose methylation (and demethylation?) is governed by signals within the Ac sequence. These signals seem to act specifically on Ac as the hypermethylation of the Ac 3'-end remains restricted to Ac and is not extending into the flanking CpG-rich Waxy DNA.

Preliminary results indicate that the methylation pattern of the inactive Ac in the wx-m9::Ds-cy allele partially differs from the active Ac. The 3'-ends of both elements are hypermethylated to a similar degree. In contrast to the active Ac , however, the inactive element is also hypermethylated throughout the 5'-end except the terminal inverted repeat. Obviously, 5'-end methylation of the inactive element is not restricted to the HpaII restriction sites that are predominantly located in the 5'-untranslated region (Schwartz and Dennis, Mol. Gen. Genet. 205:476-482, 1986), but includes the TPase binding sites.

The inactive Ac behaves like a non-autonomous Ds element, i.e. it is mobilized if TPase is provided in trans. Thus, methylation of TPase binding sites at both ends of the element does not inhibit transposition, although TPase does not bind to fully methylated target sites (Kunze and Starlinger, EMBO J. 8:3177-3185, 1989). However, after replication the TPase binding sites will transiently be hemimethylated and can be bound in this state by TPase. Our data are compatible with the hypothesis that DNA methylation in conjunction with methylation-dependent DNA binding of TPase is responsible for replication-dependent transposition and the strand selectivity of transposition.

Return to the MNL 70 On-Line Index
Return to the Maize Newsletter Index
Return to the MaizeGDB Homepage