Versatile CYP98A enzymes catalyse meta-hydroxylation reveals diversity of salvianolic acids biosynthesis

Abstract

Salvianolic acids (SA), such as rosmarinic acid (RA), danshensu (DSS), and their derivative salvianolic acid B (SAB), etc. widely existed in Lamiaceae and Boraginaceae families, are of interest due to medicinal properties in the pharmaceutical industries. Hundreds of studies in past decades described that 4-coumaroyl-CoA and 4-hydroxyphenyllactic acid (4-HPL) are common substrates to biosynthesize SA with participation of rosmarinic acid synthase (RAS) and cytochrome P450 98A (CYP98A) subfamily enzymes in different plants. However, in our recent study, several acyl donors and acceptors included DSS as well as their ester-forming products all were determined in SA-rich plants, which indicated that previous recognition to SA biosynthesis is insufficient. Here, we used Salvia miltiorrhiza, a representative important medicinal plant rich in SA, to elucidate the diversity of SA biosynthesis. Various acyl donors as well as acceptors are catalysed by SmRAS to form precursors of RA and two SmCYP98A family members, SmCYP98A14 and SmCYP98A75, are responsible for different positions' meta-hydroxylation of these precursors. SmCYP98A75 preferentially catalyses C-3' hydroxylation, and SmCYP98A14 preferentially catalyses C-3 hydroxylation in RA generation. In addition, relative to C-3' hydroxylation of the acyl acceptor moiety in RA biosynthesis, SmCYP98A75 has been verified as the first enzyme that participates in DSS formation. Furthermore, SmCYP98A enzymes knockout resulted in the decrease and overexpression leaded to dramatic increase of SA accumlation. Our study provides new insights into SA biosynthesis diversity in SA-abundant species and versatility of CYP98A enzymes catalytic preference in meta-hydroxylation reactions. Moreover, CYP98A enzymes are ideal metabolic engineering targets to elevate SA content.

Keywords: CYP98A; biosynthesis; danshensu; rosmarinic acid; salvianolic acid.

Figure 1

SA are widely distributed in Lamiaceae as well as Boraginaceae plants and the biosynthesis pathway of SA was verified by isotope‐labelled feeding studies. (a) Different acyl donors and acceptors as well as their ester‐forming products exist extensively in RA‐abundant species. The content of SA determined by UHPLC–MS/MS in these plants was used for comparisons. The ratio of each component was proportional to the content of RA. The phylogenetic tree indicates the evolutionary relationship of representative Lamiaceae and Boraginaceae plants. DSS accumulation was higher in the clade including Rosmarinus officinalis, Melissa officinalis, and S. miltiorrhiza. (b) UPLC‐Q‐TOF/MS analysis of SA after feeding of isotope‐labelled L‐phenylalanine and L‐tyrosine to 3‐month‐old S. miltiorrhiza. Exogenously fed isotope‐labelled L‐phenylalanine and L‐tyrosine were successfully incorporated into different SA structures. Mass fragmentation spectra [M‐H]⁻ for SA are shown in blue line and isotope‐labelled SA are shown in red lines.

Figure 2

Sequence analysis, expression pattern study, and phylogeny of SmCYP98A enzyme with other function‐known CYP98A members. (a) Protein sequences used in the analysis include CYP98A enzymes from S. miltiorrhiza, Arabidopsis thaliana (OAP09214.1/AtCYP98A3, OAP19075.1/AtCYP98A8 and OAP16063.1/AtCYP98A9), Coffea canephora p‐coumaroyl 3′‐hydroxylase (ABB83676.1/CcCYP98A35, ABB83677.1/CcCYP98A36), Cynara cardunculus p‐coumaroyl ester 3′‐hydroxylase (ACO25188.1/CcCYP98A49), L. erythrorhizon (BAC44836.1/LeCYP98A6), p‐coumaroyl ester 3′‐hydroxylase (AGQ48118.1/LjC3H), Manihot esculenta p‐coumaroyl shikimate/quinate 3′‐hydroxylase (UWV81047.1/MeC3H), O. basilicum p‐coumaroyl shikimate 3′‐hydroxylase (AAL99200.1/ObCYP98A13‐1, AAL99201.1/ObCYP98A13‐2), Phacelia campanularia (QDF44409.1/PcCYP98A111/p‐coumaroyl shikimate 3′‐hydroxylase, QDF44410.1/PcCYP98A112/4‐coumaroyl‐3‐(3,4‐dihydroxyphenyl)lactate 3′‐hydroxylase, QDF44411.1/PcCYP98A113/caffeoyl‐3‐(4‐hydroxyphenyl)lactate 3‐hydroxylase), Ruta graveolens (AEG19446.1, RgCYP98A22), Triticum aestivum (CAE47489.1/TaCYP98A10, CAE47490.1/TaCYP98A11), Trifolium pratense p‐coumaroyl‐shikimate 3′‐hydroxylase (ACV91106.1/TpCYP98A44), Glycine max (NP_001235563.1/GmCYP98A2), Anthoceros agrestis (MT119883/AaCYP98A147) and S. bicolor (AAC39316.1/SbCYP98A1). The NJ method was used to study the evolutionary history. The percentages of replicate trees in the bootstrap test (1000 replicates) are shown next to the branches. The tree is drawn to scale, with branch lengths in the same units as those of the evolutionary distances used to construct the phylogenetic tree. The function of enzymes was listed on the right of NJ‐tree. (b) Compared the amino acid sequences encoded by the four SmCYP98As and CbCYP98A14. Sequence alignments were performed with the ClustalW. Black boxes enclose amino acids that are identical in the four SmCYP98As and are common with CbCYP98A14. The P450‐conserved domains are marked with red rectangles: (1): Proline‐rich membrane hinge (PPGP), (2): I‐helix involved in oxygen binding and activation (A/G‐G‐X‐E/D‐T‐T/S), (3): ERR triade (E‐X‐X‐R‐R), (4): Clade signature (PERF), and (5): Heme‐binding region (F‐X‐X‐G‐X‐F‐X‐C‐X‐G). (c) Expression level of four identified SmCYP98A enzymes in S. miltiorrhiza roots, stems, leaves, and flowers.

Figure 3

UPLC‐Q‐TOF/MS profiles of SmRAS and SmCYP98A enzymes‐catalysed reactions in SA biosynthesis. SmRAS as well as yeast microsomes containing SmCYP98A75 and SmCYP98A14 were incubated with the substrates indicated. (a) Reactions of recombinant SmRAS enzyme assays using different acyl donors and acceptors as substrates. (b) Reactions of recombinant SmCYP98A75 and SmCYP98A14 enzymes assays using 4C‐4′‐HPL, 4C‐3′,4′‐DHPL, Ca‐4′‐HPL as substrates. (c) Reactions of recombinant SmCYP98A75 and SmCYP98A14 enzymes assays using 4‐HPL as substrates versus reaction system addition with SmRAS and acyl donor participation. (d) Scheme showing the steps of acyl donors and acceptors ester‐forming reactions in vitro by the recombinant SmRAS. (e) Scheme showing the steps of RA and its precursors synthesized in vitro by the recombinant SmCYP98A75 and SmCYP98A14. (f) Scheme showing the steps of DSS and 4‐HPPA synthesized in vitro by the recombinant SmCYP98A75 and SmCYP98A14. The blue circle indicated the catalytic position of SmCYP98A75, and the red circle indicated the catalytic position of SmCYP98A14. ‘S’ represented the substrate used in reactions and the bold arrow indicates the product of each catalysis assay.

Figure 4

Determination of SA from SmCYP98A75 and SmCYP98A14 knockout and overexpression transgenic hairy root. (a) SmCYP98A mutation introduced by CRISPR–Cas9 and schematic diagram showing the SmCYP98A75 and SmCYP98A14 structure as well as editing sites. Deletions are shown with dashed lines, and insertions are shown in purple colour. (b) CRISPR/Cas9‐introduced SmCYP98A75 and SmCYP98A14 representative mutation chromatography, as determined by Sanger sequencing after PCR amplification. (c) The content of SA in SmCYP98A75‐knockout lines. (d) The content of SA in SmCYP98A75‐knockout lines. (e) The content of SA in SmCYP98A75 and SmCYP98A14 in dual‐targets mutant lines. (f) The content of SA in SmCYP98A75 overexpression lines. SmCYP98A14 overexpression lines. (g) The content of SA in SmCYP98A14 overexpression lines. Data are presented as means ± SD of three biological replicates. Asterisk indicates a significant difference compared to the corresponding control lines with *P < 0.05 (Student's t‐test).

Figure 5

Structural analysis of SmCYP98A75 and SmCYP98A14. (a) Overall structure of SmCYP98A75 and SmCYP98A14. The cofactor Heme is shown as sticks in SmCYP98A75 (aquamarine) and SmCYP98A14 (wheat). (b) Channels in the structure of SmCYP98A75 and SmCYP98A14. The electrostatic surface of SmCYP98A75 and SmCYP98A14 and the two substrate channels are shown and labelled. Blue and red colours represent positive and negative charges, respectively. Red and blue arrows indicate the hypothetical channels for substrate and water, respectively. The residues binding to Heme are shown as sticks (SmCYP98A75, aquamarine; SmCYP98A14, wheat). (c) Interactions of Ca‐4′‐HPL and SmCYP98A75 in the active pocket. Yellow dashes indicate H‐bonds, and the residues involved in interactions are shown as sticks and labelled. Heme is shown as sticks and coloured salmon. (d) Interactions of 4C‐3′,4′‐DHPL and SmCYP98A14 in the active pocket. Heme was shown as sticks and coloured deep orange. (e) Two key motifs for substrate recognition. The substrate selection region is shaded in cyan, the substrate binding motif is shaded in green.

Figure 6

The biosynthesis pathway of SA in S. miltiorrhiza verified in this study. Different acyl donors from phenylalanine pathway as well as acyl acceptors from tyrosine pathway are synthesized by SmRAS to generate precursors of RA. SmCYP98A14 and SmCYP98A75 catalyse the hydroxylation at C‐3 or C‐3′ of aromatic rings to yield RA. dash lines indicate the reaction has been characterized, but the enzyme has not been identified. 4CL, 4‐coumaric acid coenzyme A ligase; HPPR, hydroxyphenylpyruvic acid reductase; PAL, L‐phenylalanine ammonia‐lyase; TAT, L‐tyrosine aminotransferase.