Seed coat color and structure are related to the seed dormancy and overwintering ability of crop-to-wild hybrid soybean

The possible persistence of genetically modified (GM) crop-to-wild hybrid seeds in the soil seed bank is a major concern in risk assessment and is closely related to seed characteristics such as dormancy. In the present study, we generated F3 hybrids via crosses between GM soybean accessions and wild soybean and evaluated the dormancy, overwintering ability and inheritance of foreign genes in different-colored hybrid seeds (yellow, green, brown, and black). The results revealed that the 5-enolpyruvylshikimate-3-phosphate synthase (EPSPS) transgene may have no influence on crop wild hybrid seed dormancy and overwintering ability, and the dormancy of the hybrid seeds was closely related to seed coat color. F3 hybrid seeds with light colors (yellow and green) were relatively nondormant, while seeds that were dark (brown and black) in color were relatively dormant. Moreover, the hybrid seeds that were dark in color had a much stronger overwintering ability than the lighter-colored seeds, with 21.33% of the black seeds and 33.33% of the brown seeds remaining viable after 240 days of soil burial. In contrast, almost all the F3 yellow and green seeds were no longer viable during winter. Scanning electron microscopy revealed that the lighter-colored seeds had a thin palisade layer and very few surface deposits, while the darker-colored seeds had a thicker palisade layer and a large area of honeycomb-like surface deposits similar to those of wild soybean seeds. Thus, the physical dormancy and overwintering ability of the darker-colored seeds may be related to the seed coat. Our results suggest that transgenes of GM soybean might disperse into wild populations and persist in seed banks.