Three putative promoter sequences have been cloned from Australian banana streak virus (BSV) isolates from the Cavendish, Mysore and Goldfinger banana cultivars and have been named Cv, My and Go, respectively. The activity of the unaltered promoters has been tested in transgenic plants, whereas deletion derivatives have been tested under transient transformation conditions. An attempt was made to optimise the Go promoter activity both by addition of an intron and by molecular evolution. A gene silencing strategy was also carried out in an attempt to develop BSV resistant banana plants.
In transgenic sugarcane, the Cv promoter was at least as active as the maize ubiquitin promoter, with GFP accumulation levels of up to 1% of soluble protein in leaves and 1.5% in stems of glasshouse-grown plants, as measured by a fluorometric quantification method developed for this purpose. Strong promoters for sugarcane transformation are scarce and the Cv promoter may be used as an equivalent to the maize ubiquitin promoter, which is the strongest promoter available for sugarcane transformation so far. In contrast to the Cv promoter, neither the My nor the Go promoter was active in mature sugarcane plants. In transgenic banana plants, the My promoter showed a high constitutive activity, which was up to four times higher than the activity of the maize ubiquitin or enhanced cauliflower mosaic virus (CaMV) 35S promoter, and is therefore among the strongest promoters available for banana transformation. In transformed wheat callus, green fluorescent protein (GFP) expression directed by the Cv promoter was at high levels, as observed by fluorescence microscopy, while GFP was expressed at low levels by the Go promoter. In in vitro-grown tobacco plants, the activity of the My promoter was as high as that of the CaMV 35S promoter, while the Cv promoter was half as active as the My and CaMV 35S promoters. My or Cv promoter activity could not be detected in glasshouse-grown tobacco plants and the Go promoter showed no detectable activity in tissue cultured tobacco plants.
The molecular characterization of the promoters included the determination of the promoter transcription start sites and the identification of transcription regulating regions. In silico analyses revealed that, like other promoters that are active in plants, the BSV promoters have a modular structure of cis-acting elements proximal to the transcription start site. Loss-of-function experiments revealed four regions important for Cv promoter activity in embryogenic sugarcane cells: the region between -382 and -319 that includes an as-1-like element, the region around -141 and down to -77 that includes 4 Dof-transcription factor binding sites and a circadian element, the region including a putative CAAT-box and the short untranslated leader region. Molecular evolution was applied to search regulatory elements in the Go promoter. Promoter clones with reduced activity were readily selected after only one round of evolution. In a small scale experiment, 15.6% of the 608 bp Go promoter region was mutated, identifying 7.4% neutral positions and 8.2% that could be important for promoter activity. Functional promoter analysis by molecular evolution could be faster and more powerful than classic analysis by deletions and specific mutations.
It was then attempted to optimise the Go promoter activity by either inclusion of intron sequences or by molecular evolution. The maize ubiquitin-1 intron did not stimulate gene expression and the maize shrunken-1 intron-exon combination actually decreased gene expression in embryogenic sugarcane cells when coupled to the Go promoter. Improvement of Go promoter activity by molecular evolution was unsuccessful after two rounds of molecular evolution by error prone PCR and DNA shuffling.
Part of the conserved RNase H encoding region of BSV, which is included in the Cv and My promoter regions, was used to test a strategy for induction of BSV resistance in transgenic banana plants. BSV is a worldwide problem and hinders propagation of plant material, breeding of new cultivars and the movement of germplasm. A construct for banana transformation was designed to confer BSV resistance by transcription of a BSV intron-spliced hairpin RNA coupled to the 3’ untranslated region of the gus reporter gene. It was expected that degradation of BSV transcripts would be triggered by dsRNA expression from this silencing construct. Despite the presence of the transformation construct in transgenic in vitro-grown banana plants, silencing of the gus reporter gene and silencing of an expressed BSV sequence could not be demonstrated. Further experiments, eventually including field trials, will be required to verify the effectiveness of the intron-spliced hairpin RNA in delivering resistance to the BSV virus.