Male-sterile and cleistogamous phenotypes in tall fescue induced by chimeric repressors of SUPERWOMAN1 and OsMADS58

Since tall fescue (Festuca arundinacea Schreb.) is an anemophilous (wind-pollinated) grass species, male sterility is strongly desired for transgenic tall fescue to prevent pollen dispersal. To create male-sterile tall fescue, we applied Chimeric REpressor gene-Silencing Technology (CRES-T) based on rice APETALA3 (AP3) and AGAMOUS (AG) orthologues that specify the formation of stamens. We fused the coding regions of rice AP3 orthologue SUPERWOMAN1 (SPW1), and rice AG orthologues, Os12g0207000, Os01g0886200 and OsMADS58, respectively with the artificial sequence encoding the modified EAR-like motif repression domain (SRDX). We first introduced Os12g0207000SRDX, Os01g0886200SRDX and OsMADS58SRDX into rice for evaluation of their abilities to induce male sterility. The transgenic rice expressing OsMADS58SRDX had reiterated formation of lodicule-like organs instead of stamens and carpel, a typical phenotype of ag mutant. Thus, we found that OsMADS58SRDX was most suitable for our purpose. Next, we introduced SPW1SRDX and OsMADS58SRDX into tall fescue. Although the transgenic tall fescue did not have the stamen alterations seen in SPW1SRDX and OsMADS58SRDX rice, they either produced no pollen or produced immature pollen; thus, the anthers were not dehiscent and the plants were male-sterile. In addition to the male sterility, SPW1SRDX tall fescue showed a cleistogamous (closed) phenotype in which anthers were not observed outside the glumes, with thin, abnormally elongated lodicules. Some lines of OsMADS58SRDX tall fescue showed a cleistogamous phenotype in which the lodicules were homeotically transformed into lemma-like organs. In both cases, cleistogamous phenotype was associated with morphological changes to the lodicules. We also obtained a mild phenotype of OsMADS58SRDX tall fescue, which exhibited only the male sterility. In this study, we produced novel male-sterile phenotypes using chimeric repressors and thus suggest CRES-T as a tool for transgenic improvement of forage and turf grasses.