Spatiotemporal transcriptomic and metabolomic landscapes of wild soybean seed development reveal regulatory mechanisms of nutrient accumulation

Seed development is a pivotal stage of the soybean life cycle, directly determining yield and nutritional quality related to oil and protein contents. However, the spatiotemporal mechanisms underlying cell differentiation and nutrient accumulation during seed growth remain to be resolved, especially in wild soybean (Glycine soja), which harbors rich genetic diversity for quality traits. Here, spatial transcriptomics and metabolomics were combined to dissect the dynamics of cell differentiation and nutrient accumulation in G. soja seeds at the mid-maturity stage. Differential expression analysis revealed distinct patterns of accumulation in adaxial versus abaxial parenchyma cells of the embryo: abaxial cells are enriched in protein metabolism pathways, whereas adaxial cells are focused on lipid metabolism pathways, consistent with previous reports on spatial nutrient accumulation in G. soja seeds. Pseudotemporal trajectory analyses supported a sequential pattern of transcriptional regulation underlying these differences. Analysis of cell-cell communication provided insight into the interactions that may mediate cell-type-specific differences among seed cells. Key genetic regulators and differentially abundant metabolites were identified through the integration of spatial transcriptomics and metabolomics, and GsMAPK23-4 was identified as a core candidate gene linked to nutrient metabolism in the cotyledon. Functional validation confirmed that GsMAPK23-4 modulates seed quality: knockout mutants had significantly higher levels of amino acids and proteins. These findings reveal cellular characteristics and differentiation processes in G. soja seeds at the mid-maturity stage, providing a molecular basis for understanding this phase and generating targets to improve soybean yield and quality.