Integration of Transcriptomics and Metabolomics Reveals Organ-Specific Biosynthesis and Accumulation of Pharmacologically Active Flavonoids in Rhododendron yedoense var. poukhanense

Abstract

Rhododendron yedoense var. poukhanense is an important medicinal plant, but still little is known about how its bioactive flavonoids are made in different organs. Comprehensive metabolomic profiling was performed using ultra-performance liquid chromatography–tandem mass spectrometry (UPLC-MS/MS), coupled with reference-based RNA sequencing (RNA-seq), to analyze root, stem, and leaf tissues from three independent biological replicates. Subsequently, we performed two-way orthogonal partial least squares (O2PLS) regression, canonical correlation analysis (CCA), and Pearson correlation analyses. A total of 2182 metabolites were detected. Among these, 1116, 896, and 1264 metabolites exhibited differential accumulation across the three pairwise comparisons. Concurrently, 9836, 5457, and 8007 genes were differentially expressed across the three pairwise comparisons, yielding a total of 13,019 unique differentially expressed genes (DEGs). The bifunctional flavanone 3-hydroxylase/flavonol synthase (F3H/FLS) enzyme exhibited the highest expression level in roots, consistent with root-preferential biosynthesis of flavonoid skeletons and the subsequent accumulation of flavonol glycosides in this tissue. In contrast, dihydroflavonol 4-reductase (DFR) exhibited predominant activity in roots, consistent with its role in farrerol biosynthesis. The multi-omics integration model demonstrated excellent goodness-of-fit to the experimental data. Canonical correlation analysis (CCA) further revealed a robust positive association between dihydrokaempferol accumulation and kaempferol biosynthesis. These findings collectively support flavanone 3-hydroxylase (F3H) as a candidate regulatory node governing organ-specific flavonoid partitioning. However, functional validation is required to substantiate this inference.