Regulation of Spm promoter activity by the Spm-encoded tnpA gene product and DNA methylation

--Douglas Cook and Nina Fedoroff

The maize suppressor-mutator transposable element (Spm) can exist in both an active and an inactive form. Inactive Spm elements differ from active ones by being methylated at certain cytosine residues surrounding the element's transcription start site (Banks, JA et al., Genes Dev. 2:1364-1380, 1988; Banks and Fedoroff, NV, Dev. Genet. 10:425-437, 1989). Specifically, DNA methylation 5' of the transcription start site, in the so-called upstream control region (UCR), is correlated with element inactivity, while further DNA methylation 3' of the transcription start site, in the so-called downstream control region (DCR), is correlated with stably inactive elements. Despite the strong correlation between DNA methylation and Spm inactivity, it remains uncertain whether DNA methylation has a role in regulating Spm function. The major Spm-encoded gene product, TnpA, has been implicated in the reactivation of inactive elements (Banks et al., Genes Dev. 2:1364-1380, 1988), an event that is correlated with demethylation. Gierl, A et al. (EMBO J. 7:4045-4053, 1988) have shown that TnpA binds, in vitro, to an oligonucleotide with homology to the 12bp subterminal repeats located in the 5' UCR and in the extreme 3' end of the element, and binding affinity is reduced upon DNA methylation. Binding of TnpA to the subterminal repeats is likely responsible for the Spm-suppressible phenotype observed with certain Spm insertion alleles, since TnpA could suppress expression of a b-glucuronidase gene that had been engineered to include the TnpA binding site (Grant et al., EMBO J. 9:2029-2035, 1990). The present experiments provide direct evidence that, at high levels, TnpA can also down regulate the Spm promoter (see also the report by Schl�ppi, m and Fedoroff, NV, this volume).

In this study, we employed a transient assay to define the promoter of Spm transcription and to examine the effect of DNA methylation and Spm-encoded gene products on promoter activity. The 5' end of the Spm element was able to drive expression of the luciferase reporter gene following introduction into tobacco cell suspensions by microprojectile bombardment. When normalized to a co-bombarded chloramphenicol acetyl transferase (CAT) internal control, luciferase activity provided a means to quantify Spm promoter activity.

Sequences essential for at least basal promoter activity are contained within the 5' 220bp fragment of the element, coincident with the UCR and including the proposed transcription initiation site at nucleotide 209 (Periera et al., EMBO J. 5:835-841, 1986). Inclusion of the unmethylated GC-rich first exon, which is common to all known Spm transcripts and contains the largest number of methylatable sites, has no effect on promoter activity. As measured in the transient expression assay, Spm promoter strength is weak, being roughly two to three orders of magnitude below that of the CaMV 35S promoter.

The entire UCR fragment is likely to be important for promoter activity. Deletion of the 5'-terminal inverted repeat was associated with a small but reproducible increase in promoter strength. As the extent of the deletion into the UCR increased, there was a corresponding gradual decrease in promoter activity. Deletion to within 146bp of the transcription start site eliminated half of the promoter activity, while deletion to within 50bp eliminated 90% of the measurable activity. No detectable promoter activity remained upon deletion to -35bp.

DNA methylation inhibited transient gene expression. When plasmid DNA was methylated in vitro by a CpG-specific DNA methylase, transient gene expression was inhibited by three-fold as compared to a non-methylated control. This negative effect on gene expression, however, was not specific to the Spm promoter since it was also observed upon methylation of a promoterless luciferase gene control and of a CaMV 35S-CAT gene fusion. In no case did DNA methylation completely eliminate Spm promoter activity, even when the methylated DNA contained the highly methylatable, GC-rich first exon.

We have also determined that expression of the tnpA cDNA from a strong promoter suppresses Spm promoter activity. When Spm-luciferase gene fusions were co-bombarded with a CaMV 35S-tnpA cDNA, expression from the Spm promoter was completely blocked, even when the 35S-tnpA plasmid was diluted 100-fold relative to the Spm-luciferase plasmid. Not surprisingly, this effect appears to be due to an interaction between TnpA and sequences within the promoter-containing UCR fragment which includes nine potential TnpA binding sites. Step-wise deletion of 5' UCR sequences removed progressively more TnpA binding sites but did not eliminate suppression by TnpA until the last subterminal repeat was removed. This result is consistent with the hypothesis that suppression of the Spm promoter results from an interaction between TnpA and the 12bp subterminal repeats. In co-bombardment experiments, the 35S-tnpA cDNA had no effect on expression of a promoterless luciferase control or of a 35S-CAT gene fusion, both of which were sensitive to the effects of DNA methylation. A second Spm cDNA, 35S-tnpD, had no effect on Spm-luciferase activity in this assay.

The results from these experiments provide direct evidence that Spm-encoded gene products and DNA methylation can affect activity of the Spm promoter (see also Kolosha, V and Fedoroff, NV, and Schl�ppi, M and Fedoroff, this volume). Because of the strong correlation between DNA methylation and Spm inactivity in maize, it seemed likely that the Spm promoter function might be unusually sensitive to the effects of DNA methylation. Based on the present results this does not appear to be the case, however, it is possible that the extrachromosomal context of the Spm-promoter in these experiments had a significant effect on the results of the assay. We have shown that at high levels TnpA can down-regulate the activity of the Spm promoter. It is uncertain, however, whether this constitutes an autoregulatory mechanism in vivo. In the model proposed by Fedoroff et al. (MNL, this volume), we suggest that low levels of TnpA may stimulate Spm activity, while high levels of TnpA would inhibit transcription (also see Gierl, A et al., EMBO J. 7:4045-4053, 1988). Depending on the strength of the Spm promoter in vivo and its sensitivity to TnpA concentration, this may provide a mechanism by which to regulate TnpA protein levels and hence Spm function.

Return to the MNL 66 On-Line Index
Return to the Maize Newsletter Index
Return to the MaizeGDB Homepage