CBGA (1) and CBE (8) were purchased from Cayman Chemicals. All other chemicals used are of the highest purity that is required for the different experiments. The 232 homologues chosen from the SSN were codon optimized for yeast expression, synthesized, and cloned into pYES2-CT vector (Thermofisher).

The sequence of Q8GTB6 was used as a query to determine the superfamily the enzyme belongs to (http://www.ebi.ac.uk/interpro/ (accessed on 2 January 2018)) [16]. Thereafter, the EFI-EST tool (https://efi.igb.illinois.edu/efi-est/ (accessed on 2 January 2018)) was used to generate the SSN. A small number of homologues (232 homologues) were chosen to test if CBGA (1) can be used as a substrate for these enzymes to catalyze the biosynthesis of a cannabinoid. The homologues were chosen using the following criteria: (1) Orthologues found in C. sativa and its related organisms such as Nicotiana sp.; (2) Orthologues that are experimentally shown to exist either at the protein or transcript level; (3) Orthologues from other plants that share less than 40% sequence identity to Q8GTB6.

The cloned genes were transformed into Saccharomyces cerevisiae BY4741 by chemical transformation [17,18,19]. The transformed cells were plated onto SC-URA-glucose plates and incubated for two days at 30 °C. Three single colonies were picked and grown in SC-URA-glucose media for 24 h. The cells were harvested and resuspended in SC-URA-galactose media and incubated for another 24 h. Thereafter, cells were harvested and resuspended in 100 mM citrate, pH 5.5, and 1 mM MgCl₂. The cell wall was digested by the addition of lyticase (0.5 mg) for 1 h at 37 °C. Glass beads (425–600 μm) were added and the cells were broken using a tissue homogenizer. The cell supernatant was clarified by centrifuge and used for the biosynthetic activity screen.

A 50-μL reaction mixture was prepared for the cannabinoid biosynthesis activity screen. The mixture contains 0.4 mM CBGA, 1.5 mM FAD, and 48 μL cell lysate. The reaction was incubated for 24 h under ambient conditions. The compounds were extracted using ethyl acetate, dried, and re-dissolved in acetonitrile. The presence of the cannabinoids was analyzed using the Agilent RapidFire 365 High-Throughput system coupled with the Agilent 6495 TQ. As mentioned previously, the potential cannabinoid products using CBGA (1) have two different molecular formulas—C₂₂H₃₀O₄ and C₂₂H₃₀O₅ with m/z values (ESI, negative mode) of 357.2 and 373.2, respectively. Corresponding control experiments in the absence of CBGA (1) were also prepared.

Orthologues that were determined to produce the cannabinoid compounds in S. cerevisiae were expressed in P. pastoris according to published protocols with some modifications [13,14]. A0A1J6KP0 and P93479 were annotated to contain a signaling peptide using the SingalP 5.0 bioinformatic tool. Thus, the signaling peptide was removed and a Strep-tag II and a 6x-histidine tag were added on the C-terminus for purification purposes [20]. The modified genes were cloned into the pPICZα A plasmid (Invitrogen). Cells were transformed by electroporation using 1 mg of PmeI linearized plasmid DNA following published protocols (Invitrogen). Transformants were selected by plating the cells on YPD agar plates containing zeocin (100 mg mL⁻¹). Clones containing the integrated gene were confirmed by following published protocols (Invitrogen).

The orthologue M4DIE5 does not contain a signaling peptide. Thus, the gene was modified by adding a Strep-tag II on the C-terminus for purification purposes [20]. The gene was cloned into the pPICZα A plasmid (Invitrogen). Cells were transformed and the presence of the gene was confirmed using protocols mentioned previously.

Recombinant P. pastoris cells containing the orthologues were grown in BMGY media at 30 °C and 190 rpm for 24 h. Afterwards, cells were harvested by centrifugation at 3000× g for 5 min and resuspended in modified BMMY (mBMMY) media to an OD₆₀₀ of 2. Pichia cells were cultivated at 15 °C and 90 rpm and supplemented with 1% (v) methanol and 0.02% (w/v) riboflavin every 24 h for induction of protein expression. Cells were harvested and resuspended in 100 mL of binding buffer. Cells were cracked using a homogenizer (M110P Microfluidizer^®®) and debris was removed by centrifugation at 20,000 rpm and 4 °C for 30 min. The cleared lysate was loaded onto a Ni²⁺ chelating sepharose column (GE Healthcare) or a Strep-Tactin^®® XT column (IBA Lifesciences). For orthologues purified using a Ni²⁺ chelating sepharose column, enzymes were eluted using increasing amounts of imidazole (4 mM to 300 mM imidazole). For orthologues purified using a Strep-Tactin^®® XT column, enzymes were eluted using 50 mM D-biotin. Fractions containing the protein were pooled, concentrated, and dialyzed vs. 100 mM citrate, pH 5.5, and 0.2 mM FAD. Purified samples were frozen and stored at −80 °C.

Kinetic parameters of the orthologs were determined using an end-point assay under initial rate conditions. A 100-mL reaction containing 100 citrate, pH 5.5, 0.2 mM FAD, 0.2 to 2 mM CBGA, 5 mM CBS orthologue was prepared for each orthologue. A 20-μL aliquot was stopped by the addition of acetonitrile every 10 min. Samples were filtered and analyzed using the Agilent 1290 Infinity HPLC coupled with the Agilent 6550 iFunnel Q-TOF high resolution mass spectrometer.

The purified proteins were used to catalyze the production of cannabinoids on a larger scale to produce enough compounds for ¹H-nuclear magnetic resonance (NMR) and ¹³C-NMR analysis. The steps performed are analogous to the smaller scale experiment mentioned previously. Samples were dissolved in CD₃OD and the spectra were recorded using a Bruker AVANCE 500 MHz NMR spectrometer at the Department of Chemistry, National University of Singapore.

AlphaFold models of the orthologues are taken from their respective UniProt entry pages.