The long term stability of transgene expression is vital for commercial-scale production of seeds with transgenic traits. However, as an adaptive defense against foreign nucleic acids, gene silencing is likely to occur inevitably in transgenic plants and to influence stability of transgene expression. The mechanism involved in transgene silencing is usually associated with several factors, including the site of integration in the host genome, the transgene copy number, repetitive sequences, transgene zygosity, DNA methylation and so on.

Dr. OW, David W and his group from South China Botanical Garden of Chinese Academy of Sciences are dedicated to developing a site-specific gene stacking method which is valuable for creating the commercial cultivars with multiple transgenic traits. Although this site-specific integration system could eliminate most of the variables (which are introduced by the random and imprecise insertion of DNA into the host chromosome) associated with gene silencing, the occurrence of the plants with silenced transgenes can not be avoided.

A previous study about testing the gene stacking system in tobacco demonstrated that ~5% of the integration events were site-specific and were usable as substrates for the next round of gene stacking, but a third of the site-specific integrants experienced transgene silencing (Hou et al.). Obviously, recovering a higher percentage of useful lines from the site-specific integrants could greatly improve this site-specific gene stacking method. Among the integrants obtained by Dr. OW’s group from targeting a gfp reporter gene into a tobacco genome by site-specific recombination, one such line (23.C.7) with silenced gfp had a copy of gfp at site-specific locus (S copy) and an extra copy at a random site (R copy). In addition, those copies found in this line showed DNA hypermethylation.

To test whether segregating away the random copy would reactivate the transgene expression in the transgenic line, former Ph.D. student WEI Junjie and his colleague initiated the related project under the supervision by Dr. OW. They backcrossed two T1 plants to the wild type, and found that the silenced status was maintained in the progenies from one (23.C.7-D) backcross, whereas, spontaneous partial reactivation of gfp expression was observed among the progenies from a second (23.C.7-2) backcross.

Unexpectedly, the analyses of gene structure and gene expression showed that this reactivation occurred not only in the S but also in the RS (with both the S and R copies) genotype from 23.C.7-2 backcross. This result suggested that the reactivation did not correlated with loss of the second random copy. As gfp was found to be hypermethylated in gfp-silenced line 23.C.7 but unmethylated in gfp-active line 23.C.4 (only contains an S copy), an association was considered to be between hypermethylation and gene silencing.

Based on this phenomenon, it would seem plausible to assume that a correlation would also exist between less methylation and reactivation of transgene expression. However, with the examination of DNA methylation level, this was not found, and although the expression of the various individual plants from 23.C.7-2 backcross were not identical, they nevertheless showed a pattern and level of hypermethylation that were highly similar. In a word, backcrossing transgenic plants to wild type can lead to partial reactivation of transgene expression in progenies, and this reactivation does not necessarily involve loss of DNA homology or methylation.

The related study was published in Journal of Experimental Botany entitled “Spontaneous reactivation of a site-specifically placed transgene independent of copy number or DNA methylation”. For further reading, please refer to https://doi.org/10.1093/jxb/erz514.