Analysing a Group of Homologous BAHD Enzymes Provides Insights into the Evolutionary Transition of Rosmarinic Acid Synthases from Hydroxycinnamoyl-CoA:Shikimate/Quinate Hydroxycinnamoyl Transferases

Abstract

The interplay of various enzymes and compounds gives rise to the intricate secondary metabolic networks observed today. However, the current understanding of their formation and expansion remains limited. BAHD acyltransferases play important roles in the biosynthesis of numerous significant secondary metabolites. In plants, they are widely distributed and exhibit a diverse range of activities. Among them, rosmarinic acid synthase (RAS) and hydroxycinnamoyl-CoA:shikimate/quinate hydroxycinnamoyl transferase (HCT) have gained significant recognition and have been extensively investigated as prominent members of the BAHD acyltransferase family. Here, we conducted a comprehensive study on a unique group of RAS homologous enzymes in Mentha longifolia that display both catalytic activities and molecular features similar to HCT and Lamiaceae RAS. Subsequent phylogenetic and comparative genome analyses revealed their derivation from expansion events within the HCT gene family, indicating their potential as collateral branches along the evolutionary trajectory, leading to Lamiaceae RAS while still retaining certain ancestral vestiges. This discovery provides more detailed insights into the evolution from HCT to RAS. Our collective findings indicate that gene duplication is the driving force behind the observed evolutionary pattern in plant-specialized enzymes, which probably originated from ancestral enzyme promiscuity and were subsequently shaped by principles of biological adaptation.

Keywords: evolution; homologous protein; hydroxycinnamoyl-CoA:shikimate/quinate hydroxycinnamoyl transferase; promiscuity; rosmarinic acid synthase.

Figure 1

(A) In plants, HCT primarily catalyses the transfer of the p-coumaroyl acyl group from p-coumaroyl-CoA to shikimate or quinate to produce p-coumaroyl shikimate or p-coumaroyl quinate, while RAS mainly accepts p-coumaroyl-CoA and 4-hydroxyphenyllactate to produce p-coumaroyl-4′-hydroxyphenyllactate, a main precursor of rosmarinic acid. The final products of these metabolic pathways assume roles in phytochemical defence or structural support. HCT, shikimate/quinate hydroxycinnamoyl transferase; RAS, rosmarinic acid synthase; C3H, p-coumaroyl 3-hydroxylase; C3′H, p-coumaroyl 3′-hydroxylase. (B) The phylogenetic analysis of MlATs. This phylogenetic tree displays the evolutionary connections between MlATs and diverse hydroxycinnamoyl transferases originating from distinct plant taxa. MlAT1 and MlAT3 are grouped within the clade of the HCT family. Other MlATs are adjacent to the Lamiaceae RASs. HQT, quinate hydroxycinnamoyl transferase. MlATs, a general term for RAS homologous proteins derived from Mentha longifolia.

Figure 2

Enzyme activity test of MlATs. Acyl acceptors, namely 4-hydroxyphenyllactate, shikimate, and quinate, can be utilized by MlAT1 (an HCT), MlAT2, MlAT4, and MlAT4 to generate their respective esters with p-coumaroyl-CoA or caffeoyl-CoA at varying degrees of catalytic efficiency. (A) The substrate pair consists of p-coumaroyl-CoA and shikimate, with “a”, “b”, and “c” representing three p-coumaroyl shikimate isomers. (B) The substrate pair consists of p-coumaroyl-CoA and 4-hydroxyphenyllactate, where “e” represents p-coumaroyl-4’-hydroxyphenyllactate molecule, while “d” may be an isomer with an unknown structure. (C) The substrate pair consists of caffeoyl-CoA and shikimate, with “f”, “g”, and “h” representing three caffeoyl shikimate isomers. (D) The substrate pair consists of p-coumaroyl-CoA and quinate, with “i”, “j”, and “k” representing three p-coumaroyl quinate isomers. Control groups: substrate pairs and buffer; enzyme groups: substrate pairs, enzymes, and buffer. Structural formulas in black represent acyl donors, whereas red formulas represent acyl acceptors. LC-MS was used to detect products, molecular ion peaks were extracted in ESI-mode, and fragmentation of the ions assisted in the qualitative analysis.

Figure 3

Structural comparison and sequence alignment of MlAT6 with AtHCT and CsRAS. (A) Structure comparison of MlAT6 (the docking model, purple) and AtHCT (the crystal structure, pink). (B) The interaction of p-coumaroyl-5-O-shikimate with AtHCT and MlAT6. In AtHCT, Arg-356 forms double salt bridges with the carboxyl group of shikimate moiety, while Tyr-361 and Thr-369 form hydrogen bonds with the ligand. In the case of MlAT6, Tyr-302, Arg-355, and Lys-398 interact with the carboxyl moiety of p-coumaroyl-5-O-shikimate, while Thr-368 forms a hydrogen bond with the hydroxyl group. (C) Structural comparison of MlAT6 (the docking model, purple) and CsRAS (the crystal structure, grey). (D) The interaction of p-coumaroyl-4′-hydroxyphenyllactate with CsRAS and MlAT6. Tyr-35, Thr-37, Lys-396, and Tyr-398 in CsRAS serve as potential interaction sites surrounding the carboxyl group of p-coumaroyl-4′-hydroxyphenyllactate, while in MlAT6, Tyr-35, Thr-37, and Lys-398 act together on the ligand. In addition, Gln-278 in MlAT6 is prone to forming a hydrogen bond with phenolic hydroxyl groups. (E) Potential residues in MlAT6, CsRAS, and AtHCT that may be implicated in indirect interactions with acyl donors. (F) Three primary binding motifs and a catalytic motif in MlAT6, CsRAS, and AtHCT. Regions highlighted in red with white text represent sequence identity among the three enzymes, while regions with light red text represent sequence identity between each two of three. Additionally, in the above catalytic pockets, Trp and His are active residues associated with the catalytic cycle, conservatively present in all plant BAHD acyltransferases. The p-coumaroyl-5-O-shikimate is marked in green, and the p-coumaroyl-4′-hydroxyphenyllactate is in cyan.

Figure 4

(A) Structural comparison and multiple sequence alignment of MlAT2, MlAT4, and MlAT6. They are conserved in the acyl receptor binding domain and exhibit similar molecular characteristics overall. Regions highlighted in red with white text represent sequence identity among the three enzymes. The active residues in the catalytic centre are indicated with an asterisk; red indicates good sequence identity; and black indicates that they are not conservative. (B) Calculated volumes of proteins’ active cavities. The volume of the active pocket of CsRAS was significantly larger than that of HCTs and MlATs. It should be pointed out that the CsRAS and HCTs used in the calculation are crystal structures, whereas MlATs correspond to structural models.

Figure 5

Expansion and diversification of the HCT enzyme family in plants. (A) Phylogenetic analysis of HCT and its homologous enzymes in Lamiaceae and its relatives. NtTHT (Hydroxycinnamoyl-CoA: tyramine N-hydroxycinnamoyltransferase) as the outgroup. Enzymes highlighted in red indicate verified enzymes that have undergone functional validation and serve as a point of reference. Enzymes highlighted in green indicate those that are located within the collinear block of chromosomes. (B) Microsyntenty analysis of HCT regions and RAS regions. Red gene names signify that the gene has undergone functional validation, while black denotes the annotation name. The interconnections among all HCT genes on the local chromosomal region are visually emphasized in orange, while the green lines depict the genetic correlation of the SmRAS gene. The links to other homologous genes of HCT are highlighted in a yellow-green colour. Enzymes in the species tree are depicted with solid circles of distinct colours, where orange represents HCT enzymes and green represents RAS enzymes. “Chr” is an abbreviation for chromosome.