Pollen-specific expression of EcoRI restriction endonuclease for bioconfinement in Panicum virgatum L.

The dispersal of pollen from genetically engineered (GE) crops can lead to unintended transgene flow, especially in wind-pollinated and obligate-outcrossing species such as switchgrass (Panicum virgatum L.). Various bioconfinement approaches have been studied for their potential to eliminate transgene flow from GE crops including selective male sterility, which aims to inhibit the germination of transgenic pollen grains. In this study, pollen-specific promoters controlling the expression of the EcoRI endonuclease, interrupted by a catalase intron, was evaluated for its efficacy to produce sterile pollen. The TaPSG719, PvPS1, Osg6B, OsRTS, and Zm13 promoters were assessed for pollen-specific expression patterns, none of which have previously been characterized in switchgrass. Here, I report relatively high EcoRI transcript abundance in pollen/anther tissue compared to the switchgrass Ubq1 reference gene when EcoRI is driven by the Zm13 (10% ± 0.9%) and PvPS1 (6% ± 1.3%) promoters. However, the TaPSG719, Osg6B, and OsRTS promoters did not significantly differ from nontransgenic switchgrass in terms of EcoRI transcript abundance in pollen/anther tissue. Furthermore, the highest-expressing transgenic line of the Zm13:EcoRI and PvPS1:EcoRI constructs were used for pollen germination experiments. These lines were presumed to contain a mixture of transgenic and nontransgenic pollen and were observed to exhibit similar pollen germination percentages compared to nontransgenic switchgrass pollen. Upon further investigation, it was determined that the catalase intron was incorrectly spliced from the mRNA, likely resulting in a nonfunctional protein when translated, which may explain why similar pollen germination rates were observed between the transgenic lines and control. In addition to these findings, I observed statistically significant phenotypic differences in many transgenic switchgrass lines when compared to nontransgenic control switchgrass. The most commonly observed adverse trait is decreased plant height, followed by decreased leaf blade width, decreased tiller diameter, and increased tiller count. However, off-target EcoRI expression did not appear correlative with plant height (r= -0.08, p= 0.4134) or leaf blade width (r= -0.23, p= 0.0133), suggesting alternative explanations such as inapt T-DNA integration sites and the high genetic diversity of switchgrass. Overall, the effectiveness of pollen-targeted EcoRI expression to produce sterile pollen was not determined. However, this study identified the Zm13 and PvPS1 promoters as strong candidates for male-specific gene expression and provided valuable insights for the design and production of genetically engineered switchgrass.