TnpA trans-activates de novo methylated Suppressor-mutator transposable elements in transgenic tobacco
--Michael Schläppi and Nina Fedoroff

We report that the maize Suppressor-mutator (Spm) transposable element is subject to epigenetic inactivation in transgenic tobacco, as it is in maize (Schläppi et al., Genetics, 1993, in press). Spm inactivation in tobacco is correlated with increased methylation of sequences near the element's transcription start site. To determine whether element-encoded gene products can promote the reactivation of an inactive element, as has been reported in maize, we investigated the effects of introducing individual cDNAs for tnpA, tnpB, tnpC, and tnpD, the element's four known protein-coding sequences. Each cDNA was expressed from the strong 35S CaMV promoter and introduced into plants containing one or more copies of the Spm element and an excision assay plasmid with an internally deleted dSpm-disrupted ¸-glucuronidase (GUS) gene. Introduction of the CaMV 35S-tnpA cDNA into the transgenic tobacco plants promoted the reactivation of the inactive resident Spm element, as judged by the appearance of regenerants with very early excision events and transposed elements. By contrast, none of the other CaMV 35S controlled cDNAs affected the activity of the resident Spm element. Similar results were obtained when the element-encoded cDNAs were introduced either by Agrobacterium-mediated retransformation or by a genetic cross. Reactivation of an inactive Spm by CaMV 35S-tnpA is accompanied by reduced methylation of several methylation-sensitive restriction sites near the ele-ment's transcription start site, but not elsewhere in the sequence. Maintenance of the reactivated Spm element in an active state requires the continued presence of the CaMV 35S-tnpA cDNA. Elimination of the CaMV 35S-tnpA cDNA locus by genetic segregation generally results in decreased element activity, as judged by its ability to promote excision of the dSpm element from the excision assay plasmid, and is accompanied by increased methylation of the element's 5' end. Exceptions resembling the phenomenon of "presetting" are also observed in which progeny plants that did not receive the CaMV 35S-tnpA cDNA locus maintain high excision activity and exhibit low methylation levels. Together with the finding in maize that a weak Spm element can transcriptionally activate an inactive Spm (Kolosha and Fedoroff, MNL66:9, 1992), we hypothesize from the tobacco reconstitution experiments that tnpA activates transcription from the Spm 5'-end. The following observations, summarized below, support the hypothesis.

The transcription start site of Spm has been located at nucleotide 209 from the element's 5' end (Pereira et al., EMBO J. 5:835, 1986). TnpA binds in vitro to a 12bp motive repeated in direct and inverted orientation several times within 200bp of the element's 5' end and within 600bp of its 3' end (Gierl et al., EMBO J. 7:4045, 1988). In a transient expression assay recently established in the laboratory, it has been shown that the luciferase reporter gene fused to different Spm 5' segments is active after introduction into NT1 tobacco suspension cells by microprojectile bombardment (Cook and Fedoroff, MNL66:11, 1992). A 600bp Spm 5' end segment containing sequences both upstream and downstream from the transcription start site, as well as a 220bp DNA segment without the G+C-rich downstream sequences of the untranslated leader, are active in the transient expression system. In addition, minimal promoter activity can be detected using deletions to within -41 from the transcription start site, but deletions including the transcription start site abolished promoter activity. All of these constructs are also active in transgenic tobacco (M. Schläppi, unpublished), but become inactivated as rapidly as full length Spm elements. It will be determined whether this inactivation correlates with methylation of the Spm 5' sequences as it does in the full-length element.

Constitutive expression is observed when the 5' Spm-luciferase construct is introduced into tobacco suspension culture cells by microprojectile bombardment (Cook and Fedoroff, MNL66:11, 1992; Raina and Fedoroff, this issue). The same is true when the Spm 3' end is added to the 3' end of the luciferase construct (R. Raina, unpublished). A similar 5' Spm-luciferase-Spm 3' construct was linked to the GUS excision assay plasmid and introduced with the CaMV 35StnpD cDNA into tobacco by Agrobacterium mediated co-transformation. Small plantlets removed from selected calli that were analyzed for luciferase activity at a very young stage give low to moderate levels of luciferase gene activity (two- to five-fold above background). In older leaves, the luciferase activity is on average only two-fold above background. Such leaves were retransformed with the tnpA cDNA controlled by promoters that differ in strength. Since the CaMV 35S-tnpD cDNA had previously been introduced into the retransformed plants, the simultaneous effect of tnpA on dSpm excision from the excision assay locus and on transcription activation from the Spm 5' end (fused to luciferase) can be monitored in this experiment, because tnpA and tnpD are necessary and sufficient for Spm excision (Frey et al., EMBO J. 9:4037, 1990; Masson et al., Plant Cell 3:73, 1991). The results show a positive correlation between excision of the linked dSpm and trans-activation of the luciferase gene controlled by the Spm 5' end. That is, whenever a high frequency of dSpm excision is observed in the presence of the tnpA cDNA, a correspondingly high luciferase activity, five- to sixty-fold above background, is observed. In some cases, little or no dSpm excision correlates with a high luciferase activity. But we have never observed excision in plants that exhibit no luciferase activity. Hence, the results indicate that tnpA introduced by transformation can trans-activate 5' Spm-luciferase constructs that are stably integrated into the tobacco genome. In contrast, when a tnpA cDNA is co-bombarded into cultured cells with an 5' Spm-luciferase construct in the transient assay, it inhibits luciferase expression (Cook and Fedoroff, MNL66:11, 1992). It is therefore possible that Spm 5' sequences are targets for DNA methylation and that TnpA activates the Spm promoter in vivo indirectly by interfering with methylation. Experiments are underway to determine the TnpA concentrations required for promoter activation or repression and to directly analyze the TnpA interaction with cis-acting sequences in vivo. 

Please Note: Notes submitted to the Maize Genetics Cooperation Newsletter may be cited only with consent of the authors

Return to the MNL 67 On-Line Index
Return to the Maize Newsletter Index
Return to the Maize Genome Database Page