Yeast Cells Coaxed Into Making Medical Cannabinoids

Scientists have successfully reprogrammed baker’s yeast cousin to manufacture cannabis compounds in a laboratory, potentially bypassing the need for acres of hemp plants and unpredictable growing seasons. The engineered microbes produced cannabigerolic acid (CBGA), a precursor to CBD and other therapeutic molecules, at levels that could eventually support commercial production.

The research team, led by Peng Xu at the Guangdong Technion-Israel Institute of Technology, focused on Yarrowia lipolytica, an oleaginous yeast naturally gifted at accumulating oils and generating the molecular building blocks cannabinoids require. Unlike Cannabis sativa plants, which need months to mature and remain vulnerable to drought, pests, and regulatory restrictions, yeast can churn out compounds in stainless steel fermentation tanks year-round.

Engineering a Cellular Assembly Line

The cannabinoid synthesis pathway resembles a molecular assembly line with three distinct stations. First, the yeast must produce olivetolic acid (OLA), a ring-shaped molecule constructed from fatty acid precursors. Second, it generates geranyl pyrophosphate (GPP) through the mevalonate pathway. Finally, specialized enzymes called prenyltransferases stitch these components together to form CBGA, the grandmother molecule that plants convert into THC, CBD, and dozens of other cannabinoids.

Getting yeast to execute this botanical choreography required solving multiple metabolic puzzles. The researchers imported a gene from the soil bacterium Pseudomonas putida to help convert hexanoic acid into the proper starting material. They systematically knocked out genes that diverted carbon away from cannabinoid production, including DGA1, which normally shuttles acetyl-CoA toward fat storage. Each genetic modification pushed OLA production higher, from an initial 0.33 milligrams per liter to 6.73 mg/L.

“By optimizing the precursor supply, engineering biomolecular condensate-like dual prenyltransferase expression and expanding endogenous metabolism with a noncanonical polyketide synthase, we achieved the de novo biosynthesis of various cannabinoids and their analogs.”

But the prenyltransferase step remained stubbornly inefficient. The cannabis enzyme CsPT4, even when stripped of its chloroplast-targeting sequence, could not keep pace with the upstream supply of OLA and GPP. The breakthrough came when researchers paired CsPT4 with NphB, a bacterial enzyme with complementary activity. Confocal microscopy revealed something unexpected: the two enzymes clustered together on the endoplasmic reticulum surface, forming what the researchers describe as “condensate-like” assemblies that dramatically boosted CBGA output.

From Trace Amounts to Industrial Promise

The dual-enzyme approach elevated CBGA production to 2.85 mg/L in strains growing on simple glucose. When researchers supplemented the medium with pre-made OLA and added a drug that blocks competing sterol synthesis, the yeast produced 15.7 mg/L CBGA, the highest titer reported in any oleaginous yeast. That represents roughly a 39 percent conversion efficiency from OLA to CBGA.

The team also demonstrated the platform’s versatility by producing orsellinic acid (OSA), a precursor to cannabidiorcol and other minor cannabinoids with potential therapeutic applications. They introduced ArmB, a large multi-domain enzyme from the honey mushroom Armillaria mellea, which had never before been expressed in yeast. The engineered strains produced up to 18.87 mg/L OSA in fed-batch cultures, though the downstream conversion to cannabigerorcinic acid (CBGOA) remained less efficient at 541 micrograms per liter.

“Our engineered Y. lipolytica produced approximately 3.5 mg/L cannabigerolic acid, 18.8 mg/L orsellinic acid, and 0.5 mg/L cannabigerorcinic acid.”

Current titers remain far below what pharmaceutical manufacturing would require. Competing metabolic pathways still siphon away critical precursors, and hexanoic acid concentrations above 1 millimolar proved toxic to cells. The prenyltransferases, while improved, have not been optimized for their non-native substrates. Adaptive laboratory evolution and directed enzyme engineering could address these limitations, potentially pushing production into economically viable territory.

The work represents more than an alternative source for CBD. CBGA shows promise as an anti-epileptic agent and demonstrated activity against the SARS-CoV-2 protease in preclinical studies. Microbial production could provide consistent access to these compounds while enabling systematic exploration of cannabinoid analogs that plants never evolved to make. The researchers envision using their platform to generate designer molecules with tailored therapeutic profiles, a capability that remains out of reach with agricultural production.

Whether engineered yeast can displace cannabis cultivation depends on economic factors the study does not address. But the research demonstrates that synthetic biology can recreate botanical complexity in surprisingly compact cellular packages, transforming what was once the exclusive domain of Cannabis sativa into a problem of metabolic engineering.

BioDesign Research: 10.1016/j.bidere.2025.100021