A single domestication origin of adzuki bean in Japan and the evolution of domestication genes.

Adzuki is a central legume in East Asian culinary culture, yet its domestication origin remains debated. Using ~700 accessions across Asia, we show that the initial domestication happened three to five thousand years ago in central Japan during the Jomon period, followed by a range expansion into China and secondary hybridization with Chinese wild populations. We mapped, validated, and dated key genes associated with seed coat color evolution (VaPAP1 for loss of mottled black and VaANR1 for gain of red colors). The frequency increases of variants affecting key domestication syndrome substantially predated the wild-cultigen divergence. Together, our results resolve the conflict between genetic and archaeological evidence about adzuki origins and reconstruct the evolutionary trajectory of archaeobotanically unobservable traits, consistent with a role of early weak selection during domestication. Editor's summary: Adzuki beans are an important East Asian legume given the frequent use of red bean paste in many confections. However, the history and domestication of this plant has been fraught with contradicting evidence. Chien et al. analyzed genomic data from 693 wild and cultivated adzuki bean accessions. Cultivated accessions showed high similarity in their chloroplast genomes to Japanese wild accessions, although Chinese varieties displayed higher levels of nuclear genome diversity, which is often indicative of a source population. These results were reconciled in that domestication of adzuki beans likely occurred in Japan, with introgression from wild Chinese varieties being responsible for the complex genomic patterns. —Corinne Simonti INTRODUCTION AND RATIONALE: The adzuki bean (Vigna angularis), with its characteristic red seed coat color, is a crucial legume of East Asian culinary culture. Although archaeological records suggest the earliest traces of cultivation in Japan, previous genetic evidence identified the highest genetic variation in central Chinese cultigens. Because the hunter-gatherer-fisher Jomon populations inhabited the Japanese archipelago before the arrival of the rice-growing Yayoi populations about three thousand years ago (3 kya), the spatiotemporal origin of adzuki domestication remains debated. Furthermore, the evolution of some domestication traits (e.g. pod shattering and seed coat color) cannot be traced by archaeological records, requiring input from genetics studies. RESULTS: We investigated about 700 adzuki accessions, covering wild and cultigen populations across the species range. The wild populations of this species likely originated near the Himalayas and naturally spread to central China and Japan. We identified genetic evidence of domestication 3 to 5 kya in Japan, with subsequent cultigen expansion back to China, followed by hybridization with Chinese wild populations and generation of admixed cultigen population in southern China. This complex demography likely thwarted previous attempts to trace adzuki domestication origin, and our results explain previous archaeological and genetic observations. Wild adzuki seed coats have mottled black pigment spots on a pale background, and most cultigens have a characteristic uniform red background without black spots. We identified one major effect gene for each of the two traits. The gain of red background was caused by an amino acid mutation in the anthocyanidin reductase gene VaANR1. Our molecular characterization of the wild and cultigen alleles confirmed that this amino acid mutation altered enzyme function, disrupting the processing of anthocyanidins and the accumulation of downstream colorless proanthocyanidins. For the locus conferring mottled black pigment spots, the wild-progenitor haplotype contained three homologous MYB transcription factor genes, and only the central copy, VaPAP1, expressed during seed development. In cultigens, nonallelic homologous recombination likely happened between the two flanking copies, deleting VaPAP1 and resulting in a chimeric copy not expressed during seed development. We modeled their evolutionary trajectories with another gene, VaMYB26, whose frameshift mutation was previously shown to cause pod nonshattering in cultigens. Because all three traits have a Mendelian genetic basis, the time-course mutant allele frequency reflects the domestication phenotype evolution in this predominately self-fertilizing species. Mutant frequency of the three genes increased much earlier than 5 kya, suggesting that weak selection existed before the inferred wild-cultigen divergence time. CONCLUSION: In this study, we resolved the complex demography of adzuki domestication, identified and validated key genes for domestication traits, and inferred the evolutionary trajectory of these traits, which could not be directly observed from archaeological records. These results complement the growing archaeological evidence that Jomons were not exclusively hunter-gatherer-fishers from the genetics perspective and support the recent archaeobotanical view that weak selection may have existed in local populations thousands of years before clear traces of sophisticated agricultural activities. The domestication of adzuki bean.: (A) The expansion route of wild and cultivated adzuki. [Map based on data from Stadia Maps; inset globe adapted from FreeVectorMaps.com] (B) Central Japanese wild accessions as the most likely progenitor of adzuki cultigens. (C) The two-step evolution from wild to cultivated adzuki and associated genetic changes. (D) The timeline of Japanese human demography and adzuki trait evolution. Horizontal black bars indicate mutation age; asterisks mark timing of initial mutant frequency increase and when frequency reached 0.1; gray band marks wild-cultigen divergence. [ABSTRACT FROM AUTHOR]