Golden liquid churns in six glass vessels, each hooked up to a computer via a bundle of tubes and wires. Air pumps buzz, and the room smells of bananas and old beer. But this facility – I was allowed to visit on the condition I not use its name or location – is no brewery. Its operators are making cannabis products using microbes instead of plants.
Humans have been cultivating cannabis plants for thousands of years, first for their fibres and later for their cannabinoid molecules, some of which have psychoactive and therapeutic properties. But we can now make those molecules without plants.
For at least five years, synthetic biologists have been producing cannabinoids using bioengineered yeast, more recently including tetrahydrocannabinol (THC), the cannabinoid behind ’ psychoactive effects.
Advertisement
Moreover, the yeast can make large quantities of cannabinoids that are present in low concentrations in the plant, some of which . And they have a smaller environmental footprint than farming. “I don’t see how plants ever compete when we get to scale,” says at Cellibre in California, one of a number of start-ups aiming to capitalise on microbial cannabinoids.
But for all this ambition, the industry has struggled: several companies recently went bankrupt and others are abandoning their yeast vats for other projects. Challenges in finding funding and worry about protecting proprietary processes have made some in this space wary of attention at the moment too – hence the terms of my recent visit.
Despite this stuttering start, however, some still believe that microbially brewed cannabis products have potential. What does that mean for the future of cannabis?
Recreating cannabinoids
It first became possible to think about cannabis in terms of cannabinoids in 1964. That year, Raphael Mechoulam and Yechiel Gaoni at the Weizmann Institute of Science in Israel  of delta-9-tetrahydrocannabinol, or delta-9 THC, which was later determined to be the main psychoactive component in the cannabis plant. In their seminal study, the researchers thanked the local police for supplying the hashish.
Mechoulam, who , went on to characterise many other cannabinoids from the plant. Seeking to understand why these molecules affected the human body  to discover that humans, like , make their own set of cannabinoids, which bind to specific receptors in the nervous system. Cannabinoids from the plant affect the mind and body by interacting with this endocannabinoid system (“endo” means “inner”).
Exactly why the cannabis plant produces these compounds is still up for debate, says  at the Weizmann Institute. But researchers have, at least, come to understand how the plants do so. Figuring out this pathway was the first step to solving another question about the special molecules, says Berman. “How can we produce them outside of the plant?”
The science of cannabis
As the use of marijuana and its compounds rises around the world, żěè¶ĚĘÓƵ explores the latest research on the medical potential of cannabis, how it is grown and its environmental impact, the way cannabis affects our bodies and minds and what the marijuana of the future will look like.
By the early 2000s, researchers had started to identify some of the enzymes involved, and had determined the that result in the production of cannabigerolic acid, or CBGA. This is known as “the mother of all cannabinoids” because most other cannabinoids are produced when additional enzymes modify CBGA.
With this pathway mapped, researchers then tried to recreate it using microbes. The first group to successfully do this and publish their findings was led by  at the University of California, Berkeley. In a study published in Nature in 2019, Keasling and his colleagues described how they systematically , hops and a species of bacteria into the common laboratory yeast Saccharomyces cerevisiae.
Their approach involved first developing a strain of S. cerevisiae that could produce sizeable amounts of a CBGA precursor when fed sugar. Modifying this yeast led to a second strain that produced CBGA. Yet more S. cerevisiae strains with further modifications could then produce the enzymes needed to transform the CBGA into THCA or CBDA, which are easily converted to the familiar THC and cannabidiol (CBD).
“Then we put in some really interesting things,” says Keasling. By feeding the modified yeast different types of fatty acids, they found they could make cannabinoids that don’t occur in the cannabis plant – or anywhere else in nature – expanding the already large library of cannabinoids available for study or use. The researchers ended their report looking towards the future: “This work lays the foundation for the large-scale fermentation of cannabinoids, independent of °ä˛ą˛Ô˛Ô˛ą˛úľ±˛őĚýł¦łÜ±ôłŮľ±±ą˛ąłŮľ±´Ç˛Ô.”
A biochemical revolution
Based on this work, Keasling co-founded a company called  aimed at using yeast to produce a large amount of CBG; this cannabinoid is non-psychoactive but shows . As żěè¶ĚĘÓƵ reported at the time, at least two other companies were already working on similar “cannayeasts”, including °ä˛ą±ôľ±´Ú´Ç°ů˛Ôľ±˛ą-˛ú˛ą˛ő±đ»ĺĚý and Canada-based , which said it had made the rare cannabinoid cannabidivarin (CBDV) with yeast as early as 2017.
A number of other biosynthetic cannabinoid companies also popped up around this time, seeing promise in a wave of actual and anticipated cannabis legalisation and decriminalisation, as well as booming demand for cannabinoids like CBD. While some companies, including Hyasynth, focused on competing with plants directly by making CBD with yeast, others, such as , prioritised making rare cannabinoids, either for use in commercial products or to explore new pharmaceutical applications.
The companies also chose to pursue different technical routes. Cellibre, for instance, which was founded in 2017, developed a strain of the less common yeast Yarrowia lipolytica, which it found naturally generates lots of the precursors needed to make cannabinoids. “It’s a very weird Yarrowia,” says Chiarelli. “When we feed it sugar it pukes acetyl-CoA”, a molecule that plays a key role in our metabolism and is also a precursor of cannabinoid compounds.
°ä˛ą±ôľ±´Ú´Ç°ů˛Ôľ±˛ą-˛ú˛ą˛ő±đ»ĺĚý, founded in 2019, did away with cells entirely. Instead, the company extracts and purifies the cell enzymes and uses them to produce cannabinoids in a manner more akin to a chemical process. “[Cells] are not terribly interested in making what you want them to make,” says , a biochemist at the University of California, Los Angeles, and the company’s co-founder.
One of the biggest entries to the field was , a publicly traded cannabis company in Canada. In 2018, with the Massachusetts-based synthetic biology firm Ginkgo Bioworks to develop strains of yeast that produce eight different cannabinoids, including .
In October 2021, Cronos – and still the only – biosynthetic cannabinoid product currently available to consumers in a popular gummy called SPINACH FEELZ Chill Bliss. The gummy included THC derived from plants mixed with CBG produced by yeast at the company’s  in Winnipeg. Its steel vats can handle 102,000 litres of fermenting yeast broth at a time, and it is currently the largest such facility in operation.
At its height, Keasling says Demetrix made more than 1000 kilograms of pure CBG over the course of four runs in a 180,000-litre bioreactor in California’s Bay Area. He says the company sold the product at a profit. “The technology worked beautifully,” says Keasling. “Then [we] ran into the buzzsaw of the financial crisis.”
The global economic recession caused by the covid-19 pandemic hit the company hard. In December 2022, Demetrix laid off its last employees and was shuttered. In August 2023, , another company co-founded by Keasling making biosynthetic CBG as well as other fermented products, . Others have also struggled. In August, Cronos said it would be  by the end of 2023 to save money in the face of declining revenue and profit. The company – which also farms cannabis plants – did not respond to multiple interview requests.
“Everybody’s dead,” says Chiarelli. He says work on cannabinoids at his company Cellibre is on pause while it focuses on other products. Chiarelli attributes the bust to several factors: alongside covid-19, there have been difficulties finding investors who understand both cannabis and synthetic biology. And he says there are technical challenges still to overcome in order to coax yeast into being productive enough to compete with cannabis plants.
There may also be work to do to persuade consumers to buy cannabinoids made by genetically modified yeast, or to find a market for the still relatively unknown and untested rare cannabinoids.
Commenting on the closure of Demetrix, Keasling also points to the “glut” of cheap cannabinoids now available because so many people have established traditional cannabis farms in recent years to take advantage of new laws and new demand. “Everybody was on a high of cannabinoids,” he says. “It’s not clear that any of them are going to make it or make money.”
While the business of cannabinoid biosynthesis seems to have fallen squarely into the so-called “trough of disillusionment” – the gap between initial enthusiasm and actual real-world roll-out – researchers say there are still reasons to pursue microbial cannabis. Perhaps the most compelling reason is to develop new medical therapies.
Farming rare cannabinoids
The microbes can make large quantities of the rare cannabinoids that the plant can only make in small amounts, says Berman. “If you have access to these rare cannabinoids, people would be able to do much more science to learn about their potential therapeutic benefits,” she says.
There is a deficit of research on the health benefits of the abundant THC and CBD, but even less is known about the rare cannabinoids. “We have just started to scrape the ground,” says Berman. And the possibilities are intriguing. For instance, Berman and her colleagues recently found that the combined effects of CBD and two rare cannabinoids, including one that was previously unknown, . Other early clinical research has found  and conditions such as inflammation, pain and neurodegeneration.
“You need now to go one by one and start understanding what are the medicinal properties these might have,” says Berman.
Some of the cannabis biosynthesis companies are explicitly taking this pharmaceutical approach. For instance, Invizyne is supplying at the University of Sydney in Australia and his colleagues with several types of rare cannabinoids to use in animal trials on treatments for seizures, sleep disorders, anxiety and ageing. Arnold says some of the acidic cannabinoids have shown promise in mouse models of a severe form of epilepsy known as Dravet syndrome. “Having ready access to highly pure and affordable minor cannabinoids at sufficient quantities is imperative to help advance cannabinoid therapeutics,” he says.
Brewing these compounds using microbes isn’t the only way to produce them though. Some strains of the cannabis plant produce like CBG, cannabinol (CBN) or cannabichromene (CBC). And synthetic chemistry can be used to produce compounds too: many cannabinoids now on the market – such as delta-8 THC, which has  – are made using synthetic chemistry to alter CBD derived from hemp plants.
But these approaches have a few drawbacks, including the cost of purification and the difficulty of precisely controlling what molecules are actually made. Synthetic chemistry is particularly problematic when it comes to controlling the stereochemistry, or “handedness”, of reaction products, as a molecule and its mirror image can sometimes have completely different effects in the body. Because the enzymes in yeast produce just the form of the molecule desired and not its mirror image form, biosynthetic cannabinoids don’t have this issue.
Another potential benefit of biosynthesis is the modest size of its climate and environmental footprints – both far lower than the surprisingly large environmental impact associated with farming cannabis. According to an analysis by Cronos of its soon-to-be-mothballed fermentation facility,  than obtaining the same quantity of cannabinoids by growing cannabis plants indoors and then processing them. Fermenting cannabis also requires less land and water than growing plants, and doesn’t involve the use of pesticides. The company leader I met on my visit to a facility in August estimated it would take just a few large fermentation facilities on each continent to supply global demand for cannabinoids, which he estimated is now about 1000 tonnes.
Despite his dire portrayal of where things stand with biosynthesis today, Chiarelli is optimistic about the future for the cannayeasts. It is his job to say so, of course, but he forecasts that, within a few decades, 80 per cent of cannabinoids will be supplied by microbes.
In his view, that doesn’t look like packing bowls and rolling joints with some sort of yeast-derived residue; he believes the cannabis plant will always remain important for recreational users. Instead, he sees a bigger role for cannabinoids in everything else, from beverages to animal feed. The recent explosion of interest in products infused with CBD – a non-intoxicating cannabinoid that is claimed to improve sleep and reduce anxiety – suggests he might have a point. And as we learn more about the rarer cannabinoids and their potential therapeutic properties, it might well be the microbial side of the business that flourishes and grows.
