Level of tissue differentiation influences the activation of a heat-inducible flower-specific system for genetic containment in poplar (Populus tremula L.).

KEY MESSAGE: Differentiation level but not transgene copy number influenced activation of a gene containment system in poplar. Heat treatments promoted CRE gene body methylation. The flower-specific transgene deletion was confirmed. Gene flow between genetic modified trees and their wild relatives is still motive of concern. Therefore, approaches for gene containment are required. In this study, we designed a novel strategy for achieving an inducible and flower-specific transgene removal from poplar trees but still expressing the transgene in the plant body. Hence, pollen carrying transgenes could be used for breeding purposes under controlled conditions in a first phase, and in the second phase genetic modified poplars developing transgene-free pollen grains could be released. This approach is based on the recombination systems CRE/loxP and FLP/frt. Both gene constructs contained a heat-inducible CRE/loxP-based spacer sequence for in vivo assembling of the flower-specific FLP/frt system. This allowed inducible activation of gene containment. The FLP/frt system was under the regulation of a flower-specific promoter, either CGPDHC or PTD. Our results confirmed complete CRE/loxP-based in vivo assembling of the flower-specific transgene excision system after heat treatment in all cells for up to 30 % of regenerants derived from undifferentiated tissue cultures. Degradation of HSP::CRE/loxP spacer after recombination but also persistence as extrachromosomal DNA circles were detected in sub-lines obtained after heat treatments. Furthermore, heat treatment promoted methylation of the CRE gene body. A lower methylation level was detected at CpG sites in transgenic sub-lines showing complete CRE/loxP recombination and persistence of CRE/loxP spacer, compared to sub-lines with incomplete recombination. However, our results suggest that low methylation might be necessary but not sufficient for recombination. The flower-specific FLP/frt-based transgene deletion was confirmed in 6.3 % of flowers.