We have isolated drug-resistant variants from adenovirus-transformed rat cells that had concomitantly lost their transformed phenotype. Our aim was to determine the reason for reversion, to attempt retransformation with 5-azacytidine (5-AzaC) and to study the mechanism of retransformation. Of the three cell lines studied, one (G4F) had lost the integrated E1a genes, whereas the other two (G2a and G5) failed to synthesize E1a RNA or proteins. Incubation of these cell lines with 3 microM-5-AzaC for 2 days, followed by passaging in the absence of drug, gave rise to transformed foci in all of the cell lines. The efficiency of transformation was typical of each cell line. Surprisingly, retransformation was not accompanied by the reappearance of detectable levels of E1a gene activity in the G2aAza and G5Aza cell lines. In search of a mechanistic explanation for the loss of gene activity in the revertants and its reappearance in the retransformants, we examined the state of methylation of the E1a gene region in these cells. Neither the E1a promoter nor its upstream region was methylated in the revertants or the 5-AzaC retransformants. These results suggest that E1a transcription was suppressed by mechanisms other than DNA methylation and that 5-AzaC could retransform these cells without lifting the E1a-suppressed state.