Elucidation of the genetic regulation of vicine and convicine in faba bean (Vicia faba L.)

Abstract

Vicine and convicine (v-c) are antinutritional compounds that significantly limit the global use of faba bean (Vicia faba) for food and feed. Consequently, reducing or eliminating These compounds through breeding is highly desirable. However, progress has been hindered by an incomplete understanding of the molecular mechanisms underlying their biosynthesis, which in turn has limited the effective application of genomic tools in breeding programs. Recently, the bifunctional riboflavin gene VC1 was identified as a key enzyme in the v-c biosynthetic pathway. A frameshifting AT insertion in the coding sequence of this gene disrupts its function, leading to reduced, but not eliminated, v-c content. The availability of high-quality Genome assemblies for two cultivars, Hedin2 and Tiffany, which differ in v-c content, has enabled deeper insights into the genetic basis of this trait. The work described in this thesis revealed multiple tandem duplications of the VC1 gene and consequent copy number variation, with different genotypes carrying between two and Five copies. These copies, which carry structural variants, including insertions and deletions in exons 3 and 6 and intron 3, were classified into three major variants: VC1A, VC1B, and VC1C. High v-c genotypes were found to predominantly carry VC1A, while low v-c genotypes were characterized by the presence of VC1B and VC1C. Interestingly, copy number does not correspond to gene expression level. Instead, expression of a single variant appears to determine v-c content, suggesting a regulatory dosage compensation mechanism, possibly involving DNA methylation or microRNA activity. Importantly, low v-c genotypes still carry active VC1 variants that remain unexpressed in seeds during both early and late seed-filling stages. Furthermore, a neighbouring homolog of VC1, denominated as RIBA2, was identified. It Shares high sequence similarity with VC1 but differs structurally. Like VC1 and other bifunctional riboflavin genes, RIBA2 contains two expression domains, RibB and RibA, encoding 3,4-dihydroxy-2-butanone-4-phosphate synthase and GTP cyclohydrolase II, respectively, suggesting that it may function as a secondary riboflavin biosynthetic gene in V. faba. RIBA2 accounts for approximately 5% of riboflavin gene transcripts in immature seeds. While VC1 remains the primary determinant of v-c content, RIBA2 may contribute as a minor-effect locus, potentially explaining residual v-c production in VC1 mutants. However, functional Validation via gene editing is needed to confirm its role. Additionally, multiple reciprocal populations derived from high and low v-c lines were used to investigate maternal inheritance of the trait at the gene expression level. The findings showed that v-c-related gene expression is confined to maternal tissues, with no detectable activity in the embryo. Specifically, gene expression in the seed coat regulates v-c levels in seeds, indicating that maternal genotype determines seed v-c content through developmental stagespecific expression. Building on these insights, a set of reliable molecular markers were developed to support v-c breeding programs. One marker targets a previously described AT insertion in VC1, while four others are based on novel polymorphisms that were detected within RIBA2. To mitigate marker bias from multiple VC1 copies, a TPA-KASP approach was implemented, combining PCR and KASP genotyping in a single assay plate. The method facilitates simple implementation as a diagnostic tool for marker-assisted selection, significantly improving selection accuracy while considerably reducing the cost and effort of screening in breeding populations. Subsequent validation of the marker set in 97 faba bean genotypes revealed 100% accuracy of the assay, effectively distinguishing high and low v-c lines. These markers offer a high-throughput, cost-effective, and early-generation selection tool, critically enabling efficient screening for low v-c genotypes at the F2/F3 stage to enhance breeding efficiency. In conclusion, the methods, tools and results in this thesis advance our understanding of the complex genetic regulation of v-c biosynthesis and provides valuable genomics-based breeding tools to support the development of high-quality faba bean protein.