Combined analysis of transcriptome and metabolite data reveals extensive differences between black and brown nearly-isogenic soybean (Glycine max) seed coats enabling the identification of pigment isogenes

BACKGROUND: The R locus controls the color of pigmented soybean (Glycine max) seeds. However information about its control over seed coat biochemistry and gene expressions remains limited. The seed coats of nearly-isogenic black (iRT) and brown (irT) soybean (Glycine max) were known to differ by the presence or absence of anthocyanins, respectively, with genes for only a single enzyme (anthocyanidin synthase) found to be differentially expressed between isolines. We recently identified and characterized a UDP-glycose:flavonoid-3-O-glycosyltransferase (UGT78K1) from the seed coat of black (iRT) soybean with the aim to engineer seed coat color by suppression of an anthocyanin-specific gene. However, it remained to be investigated whether UGT78K1 was overexpressed with anthocyanin biosynthesis in the black (iRT) seed coat compared to the nearly-isogenic brown (irT) tissue. In this study, we performed a combined analysis of transcriptome and metabolite data to elucidate the control of the R locus over seed coat biochemistry and to identify pigment biosynthesis genes. Two differentially expressed late-stage anthocyanin biosynthesis isogenes were further characterized, as they may serve as useful targets for the manipulation of soybean grain color while minimizing the potential for unintended effects on the plant system. RESULTS: Metabolite composition differences were found to not be limited to anthocyanins, with specific proanthocyanidins, isoflavones, and phenylpropanoids present exclusively in the black (iRT) or the brown (irT) seed coat. A global analysis of gene expressions identified UGT78K1 and 19 other anthocyanin, (iso)flavonoid, and phenylpropanoid isogenes to be differentially expressed between isolines. A combined analysis of metabolite and gene expression data enabled the assignment of putative functions to biosynthesis and transport isogenes. The recombinant enzymes of two genes were validated to catalyze late-stage steps in anthocyanin biosynthesis in vitro and expression profiles of the corresponding genes were shown to parallel anthocyanin biosynthesis during black (iRT) seed coat development. CONCLUSION: Metabolite composition and gene expression differences between black (iRT) and brown (irT) seed coats are far more extensive than previously thought. Putative anthocyanin, proanthocyanidin, (iso)flavonoid, and phenylpropanoid isogenes were differentially-expressed between black (iRT) and brown (irT) seed coats, and UGT78K2 and OMT5 were validated to code UDP-glycose:flavonoid-3-O-glycosyltransferase and anthocyanin 3'-O-methyltransferase proteins in vitro, respectively. Duplicate gene copies for several enzymes were overexpressed in the black (iRT) seed coat suggesting more than one isogene may have to be silenced to engineer seed coat color using RNA interference.