Full-scale landscape metabolome map provides insights to convergent metabolite divergence and promotes edible maize breeding

Abstract

Abstract Ediblemaize is an important food crop, providing energy and nutrients to meet human health and nutritional requirements. However, how environmental pressures and human activity have shaped the ediblemaizemetabolome remains unclear.In this study, we collected 452 diverse edible maize accessions worldwide, comprising waxy, sweet and field maize. A total of 3020 non-redundant metabolites, including 802 annotated metabolites, were identified by a two-step optimized approach, which generated the most comprehensive annotated metabolites dataset in plants to date. Although specific metabolite divergencewas detected in Field-Sweet and Field-Waxy divergences, convergent metabolite divergencewas the dominant divergence pattern. We identified hub genes in all metabolite classes by mGWAS hotspot analysis. Seventeen and 16 hub genes were selected as the key divergence genes for flavonoids and lipids, respectively. Surprisingly, almost all of these genes were under non-parallel selection, which indicated non-parallel selection was the main genetic mechanism of convergent metabolic divergence. Furthermore, UGT1 and C1 in the flavonoid pathway, and KCS1 and LPP2 in the lipid pathway, played different roles in convergent metabolite divergence. Based on our research, we established the first edible maize metabolome database, EMMDB. We successfully applied EMMDB for precision improvement of nutritional and flavor traits, and an elite inbred line 6644_2 was bred with greatly improved in contents of flavonoids, lysophosphatidylcholines, lysophosphatidylethanolamines, and vitamins. These findings provide insights into the underlying genetic mechanisms of edible maize metabolite divergence and provide a database for the breeding improvement of edible maize flavor and nutritional traits by metabolome precision design.