Résumé
Recent breakthroughs in engineering synthetic apomixis in rice open new avenues for clonal propagation of F1 seeds, offering promising prospects to reach the small farmer fields. The most efficient method to date forconverting rice hybrids to apomixis combines the Mitosis instead of Meiosis (MiMe) triple mutation which induce apomeiosis with egg cell expression of BABYBOOM 1 to trigger parthenogenesis. However, high frequency (95-100%) of clonal seed production remains associated with a low grain filling rate (25-40%), raising questions about the origin and fate of unfilled grains. This bottleneck still prevents further field development of the technology.To address this, we conducted a large-scale screening of ovaries collected at pre-anthesis and post-anthesis stages in fully penetrant apomictic plants of three rice hybrids. Female gametogenesis appears to occur normallyin synthetic apomicts, with an embryo sac abortion rate comparable to that observed in control hybrids. Interestingly, two- to sixteen-celled parthenogenetic embryos are already observed 3-4 days before anthesis(panicles still not emerged from the flag leaf sheath), indicating that parthenogenesis from the unfertilized egg cell can start shortly after completion of female gametophyte formation. At pre-anthesis stage, 16-64 celledparthenogenetic embryos are observed in 90-95% of embryo sacs together with synergids and polar nuclei. However, 48 hours post-fertilization, consistent with the low grain filling rate, only 25-40% of the embryos areaccompanied by endosperm nuclei. This suggests that the fusion between the sperm nucleus and the polar nuclei may be hampered, likely due to the early development of parthenogenetic embryos. The precise timingand underlying cause of this massive failure are currently under investigation.