Functional divergence of CYP76AKs shapes the chemodiversity of abietane-type diterpenoids in genus Salvia

Abstract

The genus Salvia L. (Lamiaceae) comprises myriad distinct medicinal herbs, with terpenoids as one of their major active chemical groups. Abietane-type diterpenoids (ATDs), such as tanshinones and carnosic acids, are specific to Salvia and exhibit taxonomic chemical diversity among lineages. To elucidate how ATD chemical diversity evolved, we carried out large-scale metabolic and phylogenetic analyses of 71 Salvia species, combined with enzyme function, ancestral sequence and chemical trait reconstruction, and comparative genomics experiments. This integrated approach showed that the lineage-wide ATD diversities in Salvia were induced by differences in the oxidation of the terpenoid skeleton at C-20, which was caused by the functional divergence of the cytochrome P450 subfamily CYP76AK. These findings present a unique pattern of chemical diversity in plants that was shaped by the loss of enzyme activity and associated catalytic pathways.

Fig. 1. Phylogenetic relationships of Salvia estimated by Astral using five gene sets.

Five levels of bootstrap values are denoted with red, blue, yellow, black, and green solid circles. Green solid circles indicate alternative topologies for the node among the five trees (see Supplementary Figs. 2–6). Species belonging to the same subgenus are marked with the same branch color. Species names are shown to the right of the tree, and clades are indicated on the far right and highlighted with colored labels. Colored solid stars denote genome duplication identified in this study. Mya million years ago, P Pliocene, Q Quaternary, WGD whole-genome duplication, WGT whole-genome triplication. PR subgenera Perovskia and Rosmarinus. Source data are provided as a Source Data file.

Fig. 2. Chemical diversity and distribution of ATDs in Salvia. The most prevalent ATDs identified by LC-MS in both roots and leaves from Salvia.

a Thirty-six ATDs were divided into five groups based on different structural characteristics. ATDs in groups A, C, D, and E (which accumulated specifically in leaves) exhibited higher intensities in the negative ion mode, and ATDs in group B (which accumulated specifically in roots) exhibited higher intensities in the positive ion mode. 1: 7, 20-Dihydroxyabietaquinone; 2: 16, 20-Dihydroxyabietaquinone; 3: 7-Keto, 20-hydroxyabietaquinone; 4: 11, 20-Dihydroxysugiol; 5: 20-Hydroxyabietaquinone; 6: Miltirone; 7: Cryptotanshinone; 8: Tanshinone IIA; 9: 1, 2-Dihydrotanshinquinone; 10: Dihydrotanshinone I; 11: Methylenetanshinquinone; 12: Tanshinone I; 13: Przewaquinone C; 14: 7-Keto-royleanone; 15: Royleanone; 16: 11-Hydroxysugiol; 17: 7-Hydroxyroyleanone; 18: 6-Hydroxyroyleanone; 19: 6-Keto-11-hydroxysugiol; 20: 7, 11-Dihydroxyferruginol; 21: 11, 14-Dihydroxysugiol; 22: 11-Hydroxyferruginol; 23: 20-Keto-2, 11-dihydroxyferruginol; 24: 20-Keto-11-hydroxysugiol; 25: 20-Keto-7-hydroxyferruginol; 26: 20-Keto-6-hydroxyferruginol; 27: 20-Keto-7-acetyferruginol; 28: 7-Hydroxy-20-carboxyabietaquinone; 29: 7-Hydroxyisocarnosol; 30: 6-Hydroxycarnosol; 31: 7-Keto-carnosic acid; 32: 7-Methoxycarnosol; 33: 6-Methoxycarnosol; 34: Carnosol; 35: Carnosic acid; 36: Methyl carnosate. b Distribution patterns of the 36 ATDs in 71 Salvia species. PR, subgenera Perovskia and Rosmarinus. Source data are provided as a Source Data file.

Fig. 3. Cladogram of the CYP76AK subfamily members.

The branch color represents the lineage to which the species containing these CYP76AK genes belong. See Supplementary Fig. 32 for the annotated phylogram. PR, subgenera Perovskia and Rosmarinus. Source data are provided as a Source Data file.

Fig. 4. Catalytic activity of targeted CYP76AK subfamily members with two substrates.

a Extracted ion chromatogram (EIC) analysis of the CYP76AK catalytic reaction products in vitro with 11-hydroxyferruginol. To obtain 11-hydroxyferruginol, all plasmids contained SmCYP76AH3, which converts ferruginol to 11-hydroxyferruginol. A sample with only SmCYP76AH3 protein functioned as the negative control. Peaks corresponding to potential products are indicated by the m/z value of 11, 20-dihydroxyferruginol (m/z 317.2115), and carnosic acid (m/z 331.1900). b EIC overlays showing in vitro catalytic activity of CYP76AKs with 11-hydroxysugiol. The sample with the empty vector (pESC-His) functioned as the negative control. Peaks corresponding to potential products are indicated by the m/z value of 11, 20-dihydroxysugiol (m/z 331.1900), 20-keto-11-hydroxysugiol (m/z 329.1745), and 7-keto-carnosic acid (m/z 345.1695). c Enzyme activity summary of CYP76AKs.

Fig. 5. Expression profiles of the CYP76AK subfamily members in Salvia.

Gene expression heatmap for CYP76AK subfamily members from representative species of each branch of Salvia, displaying the relative expression levels in leaves (L) and roots (R). The heatmap shows three biological replicates, each plotted individually. Expression values represent Z score-transformed log (TPM + 1) values. Species abbreviations can be found in Supplementary Data 1. Source data are provided as a Source Data file.

Fig. 6. Evolution of the CYP76AK subfamily.

a Syntenic relationships among CYP76AK genes in Salvia genomes revealed the divergence of CYP76AK clades. The phylogenetic tree in the left panel represents the phylogeny of six Salvia species with N. cataria as the outgroup. WGD events are indicated by red dots. Microsynteny of CYP76AK genes in the synteny blocks. Red polygons show syntenic relationships between CYP76AK genes, and gray polygons show other genes. Anc-CYP76AK, Ancestral CYP76AK gene. b Reconstruction of ancestral CYP76AK proteins. Ancestral nodes (ancNodes) 1−6 show the resurrected enzyme of each major branch point of the CYP76AK phylogenetic tree. Functional characters of each branch are distinguished by different colors. c, d Ancestral function of CYP76AKs. Extracted ion chromatograms (EICs) showing in vitro catalytic activity of each ancestral CYP76AK using 11-hydroxyferruginol and 11-hydroxysugiol as the substrate, respectively. e Evolution of ATDs biosynthesis in Salvia. The phylogenetic tree shows the speciation of the clades of Salvia, and the diamonds represent the predicted ancestral metabolic traits for each speciation node. Square frames represent the realistic metabolic traits of each Salvia clade as shown in Fig. 3. Green, red, and light blue represent the existence of carnosic acid, tanshinones, and 20-keto ATDs, respectively. The time scale on the left reflects the time of speciation for each clade. Branches highlighted in light blue indicate clades in which the ancestral metabolic traits have been retained. f Chronology of CYP76AK genes. The phylogenetic tree represents the evolution of CYP76AK clades according to the MCMCtree model. The speciation time for each clade is labeled. Branches highlighted in blue indicate CYP76AK clades that have retained their ancestral catalytic functions, and branches highlighted in red indicate clades that have undergone a loss of function. The arrows at the bottom indicate the CYP76AK clade that contributes to producing the type of ATDs in the corresponding lineage. Green, red, and indigo gray arrows indicate the biosynthesis of carnosic acid, tanshinones, and 20-keto ATDs, respectively. Source data are provided as a Source Data file.