Systematic biotechnological production of isoprenoid analogs with bespoke carbon skeletons

Abstract

Natural products are widely used as pharmaceuticals, flavors, fragrances, and cosmetic ingredients. Synthesizing and evaluating analogs of natural products can considerably expand their applications. However, the chemical synthesis of analogs of natural products is severely hampered by their highly complex structures. This is particularly evident in isoprenoids, the largest class of natural products. Here, we develop a yeast cell-based biocatalytic method that enables the systematic biotechnological production of analogs of different classes of isoprenoids (including monoterpenoids, sesquiterpenoids, triterpenoids, and cannabinoids) with additional carbons in their skeletons. We demonstrate the applicability of this approach through two proof-of-concept studies: the biosynthesis of the highly valued aroma ingredient ethyllinalool, and the production of cannabinoid analogs with improved cannabinoid receptor agonism. This method is simple, readily adaptable to any cell factory, and enables the tailored expansion of the isoprenoid chemical space to identify molecules with improved properties and the biotechnological production of valuable compounds.

Fig. 1. Overview of the developed approach.

The developed isopentenol utilization pathway is shown in the light blue box. In this pathway, prenol and isoprenol are phosphorylated successively by AtFKI and AtIPK to generate DMAPP and IPP, respectively, which are then used to synthesize canonical terpenes. To produce terpenoid analogs, we used seven isopentenol-like alcohols (3M2E, 3MH2E, 3,4-DMP, 3E2E, 3MP, 4E3E and 3MH1E; see main text for full compound names), which can be accepted by AtFKI and AtIPK to generate non-canonical terpenoid building blocks (4-methyl-DMAPP, 4-ethyl-DMAPP, 4,4-dimethyl-DMAPP, 4,5-dimethyl-DMAPP, 4-methyl-IPP, 4,5-dimethyl-IPP and 4-ethyl-IPP, respectively). These non-canonical terpenoid building blocks condense with IPP or DMAPP to give rise to non-canonical prenyl diphosphate precursors (e.g., 8-methyl-GPP). Subsequently, terpene synthases or prenyltransferases utilize these non-canonical prenyl diphosphate precursors to produce terpenoid analogs (e.g., ethyllinalool) and high-value compounds (e.g., non-canonical CBGA analogs). OA olivetolic acid.

Fig. 2. Identification of an efficient kinase to support the first step of the constructed isopentenol utilization pathway.

Co-expression of kinase candidates and AtIPK in yeast cells to identify the optimal kinase for the first step of the constructed pathway. Yeast cells containing the empty vectors (strain LW20; black) or yeast cells with AtIPK and AtFKI (strain LW01) incubated in the absence of prenol or isoprenol (light blue) are used as controls. a Linalool production evaluated by SPME analysis of the headspace of yeast cultures supplied with 0.1% v/v prenol. b Linalool production in the headspace of yeast cultures supplied with 0.1% v/v isoprenol. c Comparison of linalool production by yeast cells containing the full-length AtFKI and or the truncated form Δ65AtFKI and supplied with 0.1% v/v prenol. Values represent the mean of three biological replicates (n = 3) and error bars correspond to standard deviation. Source data are provided as a Source Data file.

Fig. 3. Synthesis of monoterpene, sesquiterpene, and triterpene analogs in yeast.

a Isopentenol-like alcohols evaluated in this study. Alcohols that were efficiently converted by the established isopentenol utilization pathway into prenyl diphosphate analogs are indicated by a light green box. Structural differences from the corresponding canonical isopentenol are shown in red. b SPME-GC-MS extracted ion chromatogram (m/z 150; corresponding to C₁₁H₁₈) showing the production of 1 when the limonene synthase-containing yeast strain LW13 was supplied with 0.01% 3M2E. Strain LW13 incubated in the absence of the corresponding alcohol is shown as a control (blue). c SPME-GC-MS extracted ion chromatogram (m/z 164; corresponding to C₁₂H₂₀) showing the production of 2 by LW13 cells supplied with 0.005% v/v 3MH2E and 0.01% v/v isoprenol to facilitate the conversion. d–f SPME-GC-MS extracted ion chromatogram showing that strain LW35 (co-expressing SfCarS1 and ERG20(F96C)) produced one C₁₆ analog (3) and two C₁₇ analogs (4 and 5) when supplied with 0.05% v/v 3MP, 4E3E, or 3MH1E, respectively, in the presence of 0.05% v/v isoprenol. g GC-MS extracted ion chromatogram showing the production of 26-methyl-squalene (6) by strain LW36 (co-expressing ERG9 and ERG20(F96C)) when supplied with 0.05% v/v 3MP and 0.05% v/v isoprenol.

Fig. 4. Synthesis of ethyllinalool and other linalool analogs in yeast.

GC-MS ion-extracted chromatograms (m/z 107) of the headspace of strain LW37 grown in the presence of (a) 0.01% v/v 3M2E, (b) 0.005% v/v 3,4-DMP and 0.01% v/v isoprenol, (c) 0.005% v/v 3E2E and 0.01% v/v isoprenol, or (d) 0.005% v/v 3MH2E and 0.01% v/v isoprenol, showing the production of ethyllinalool and linalool analogs (pink). An ethyllinalool standard sample is shown in green in panel (a). Chromatograms from yeast cells (strain LW37) not supplied with the respective alcohols are shown as controls (blue). The sensory characteristics of each of the isolated linalool analogs were evaluated by a professional flavourist and five quality attributes (floral, fruity, spicy, zesty, and agrestic) were scored in the scale of 1-6 and plotted in the form of a spider diagram (panels e–h).

Fig. 5. Production of CBGA analogs in yeast.

UPLC–HRMS analysis of extracts of strain LW16 (containing the geranyl transferase CsPT4) revealed the production of analogs of CBGA (12–18) (pink), when supplied with 0.5 mM OA and: (a) 0.01% v/v 3M2E, (b) 0.005% v/v 3,4-DMP and 0.01% v/v isoprenol, (c) 0.005% v/v 3E2E and 0.01% v/v isoprenol, or (d) 0.005% v/v 3MH2E and 0.01% v/v isoprenol, respectively (total ion chromatograms are shown). LW16 yeast cells supplied only with OA and isoprenol but not any of the prenol-like alcohols are shown as a control (blue). The retention time of CBGA in panels (a) and (b) differs from panels (c) and (d) because these experiments were carried out using different instruments.

Fig. 6. Bioactivity evaluation of cannabinoid analogs.

a Dose-response curves of the activation of the cannabinoid receptor CB₂ by CBGA and the cannabinoid analogs 12, 13, and 17 as determined by measuring the resulting luminescent signal using a plate reader. Blue lines indicate the receptor response to the cannabinoid analog, while the black lines show the response induced by CBGA. Gray lines show the response of the control strain expressing the A_2A receptor, KM207, to the corresponding cannabinoid analog. b Dose-response curves of the activation of CB₂ by CBG and the heat-treated cannabinoid analogs 12, 13, and 17. Blue lines indicate the receptor response to the heat-treated cannabinoid analog, while the black lines show the response induced by CBG. Gray lines show the response of the control strain expressing the A_2A receptor. Data are represented as the mean of three biological replicates (n = 3) and error bars correspond to standard deviation. Source data are provided as a Source Data file.