Clustering and high expression of 1,8-cineole synthase genes in S. rosmarinus leaf tissue
The co-expression results showed that limonene synthases in S. rosmarinus were highly correlated with the synthesis of monoterpenes (p-value = 0.0102, Table S21). A total of three limonene synthases and three 1,8-cineole synthases were identified in S. rosmarinus . All genes encoding limonene synthases (SrLS-1 , SrLS-2 ,SrLS-3 ) and 1,8-cineole synthases (SrCinS-3 ,SrCinS-4 ) are clustered on pseudochromosome 3, within a 1.1 Mb region (Figure 5b, Table S25). Transcriptome analysis revealed that the expression levels of the three limonene synthases (SrLS-1 ,SrLS-2 , SrLS-3 ) were 8.44, 9.80 and 6.67 times higher in leaves than in roots (Figure 5c), respectively, while the expression levels of the two 1,8-cineole synthases (SrCinS-3 ,SrCinS-4 ) were 5.57 and 8.81 times higher in leaves than in roots (Figure 5c). Additionally, covariance analysis was performed on S. rosmarinus , S. miltiorrhiza and S. splendens , revealing that the genes encoding 1,8-cineole synthase on pseudochromosome 2 did not have homologues in S. miltiorrhiza and S. splendens.
We constructed an evolutionary tree of 24 species to analyze limonene synthases and 1,8-cineole synthases, which revealed that SrCinSsand SrLSs formed two clades with 40.47%–99.16% sequence identity ((Figure S19, Table S29). 3D models of SrCinSs andSrLSs (pdbid: 2ong) were generated and validated using Ramachandran plots (Figure S21). The models of both enzymes were highly similar in stereospecificity, as indicated by the average root means square displacements (RMSDs) (0.38Å–1.03Å) between the predicted models (Figure S22, S21a). Docking studies showed that. intermediate terpinol cations are located near the active pockets of SrCinS-3 andSrCinS-4 (Figure 5d). And eight amino acid residues of the active pocket (Cys-250, Trp-253, Asn-274, Thr-278, Met-458, His-502, Tyr-496 and Ser-454) that all lie within 10 Å distance of the docking site and have a direct effect on biosynthesis of 1,8-cineole were examined (Kampranis et al., 2007; Piechulla et al., 2016; N. Srividya et al., 2015; J. Xu et al., 2017). Cys-379, Trp-382, Tyr-626, and Thr-278 maintained their original polarity in SrCinS-3 andSrCinS-4. Mutations S512G and A278T in silico analysis indicated that these residues retained their original polarity in SrCinS-3and SrCinS-4 . However, A278T became more hydrophilic and M622I lost its original polarity (Figure S24, Table S27). Based on these findings, we hypothesize that SrCinS-3 and SrCinS-4 are responsible for 1,8-cineole biosynthesis. Furthermore, we found that S512G and A278T, which promote the accumulation of 1,8-cineole in tobacco terpene synthases (Piechulla et al., 2016), were retained in SrCinS-3 and SrCinS-4 in rosemary, their association with high 1,8-cineole accumulation in rosemary.
The limonene synthases in S. rosmarinus were observed to dock with terpinyl cations near the active pocket (Figure S23), indicating that the crystallographic structures of limonene synthase could accept terpinyl cation intermediate. Mutations in the key sites of the active pocket can affect the product diversity of terpene synthases as the product profile is determined by the conformation of the substrate or intermediate. Analysis of the active pockets of SrLSs showed that they deviate from the ancestral limonene synthase pattern (N. Srividya et al., 2015; Narayanan Srividya, Lange, & Lange, 2020), with changes in polarity observed for several important sites (Thr-278, His-502, Tyr-496 and Ser-454, see Table S27). M519I and T279V mutations directly led the loss of polarity of original residues, but compensatory mutations were observed for Asn-274 (to Phe-274) and Cys-250 (to Asn-250), which maintained the polarity of the active site (J. Xu et al., 2017) (Figure S23, Table S27). The polarity changes in the active sites resulted in a larger active pocket, potentially enhancing the production of more abundant terpenoids by attenuating the stability of the carbon positive ion.
The transcriptional expression of genes started from the specific binding of the promoter region upstream of the gene to RNA polymerase; therefore, we extracted the promoter sequences of SrCinSs andSrLSs and analyzed the promoter elements using Promoter 2.0 (Knudsen, 1999). TATA-boxes are essential for binding to RNA polymerase and activating transcription among the various elements (Orphanides, Lagrange, & Reinberg, 1996). The results showed the number of TATA-boxes nearly doubled in the promoter sequence of SrCinS-3(Table S28), greatly enhancing its expression and promoting the accumulation of 1,8-cineole. Additionally, we examined G-box elements in the promoter region of SrCinSs, and five G-boxes were targeted onSrCinS-3 , significantly increasing its binding probability with the bHLH gene family. Furthermore, we discovered twoSrbHLH143 near the gene cluster (Figure 5b), and their expression was 2.24-fold and 1.63-fold higher in leaves than in roots (Figure 5c), leading to a speculation that SrbHLH143 may play a regulatory role in the biosynthesis of 1,8-cineole. The expression of the bHLH family was reported to be significantly correlated with changes in 1,8-cineole content in Artemisia absinthium (Yi et al., 2021), and G-box was cis-acting DNA regulatory element which could bind with bHLH transcription factor(Qian et al., 2007). Conversely, the promoter region of other 1,8-cineole synthases had incomplete G-box elements, resulting in low expression.