Visit the lab at bioscience company Amyris, and the smell of yeast could kick-start a craving for beer, pizza, or both. But upon further investigation, the robot busily plating yeast colonies clearly won’t yield any delicious rewards for a Pavlovian response.
That’s because Amyris uses baker’s yeast as a catalyst that converts sugar into products ranging from malaria treatments, hand cleaner, and cosmetics to car-engine lubricants, paint compounds, and biofuels.
With help from a Bill & Melinda Gates Foundation grant, the Emeryville, California–based company’s first breakthrough in synthetic biology came in 2005. Its scientists used a yeast-brewing process to produce artemisinic acid, a precursor to the antimalarial medication artemisinin. Until then, the market for the natural artemisinin source—the Chinese Sweet Wormwood plant—was volatile.
“There used to be huge fluctuations in price and amount of artemisinin on the market,” says Sunil Chandran, senior director of R&D. Periods of surplus artemisinin (and a crash in cost) were inevitably followed by a shortage (and a surge in cost).
But that cycle changed with Amyris’s discovery. By October 2014, it shipped 120 million malaria treatments, the price for artemisinin stabilized, and the drug saved lives. “It’s reduced mortality for malaria in Africa and very poor places by hundreds of thousands of cases, and many of those are children 5 and under,” says Darren Platt, vice president of data science. “So it’s a beautiful application of synthetic biology—something that’s useful for people who need it most.”
The Chemical Roadmap
To produce a specific molecule, such as artemisinin, scientists can modify the metabolism of yeast cells and follow the chemical path (that is, rewrite the DNA). “All of the engineering is basically alternating this chemical map to get to where we want to go,” Platt says.
There are many single-celled microorganisms besides yeast that could be used in the process. But people are less comfortable with the idea of handling, say, 50 tons of E. coli for large-scale production. Yeast has a GRAS (generally recognized as safe) designation, so it’s easy to get regulatory approval for commercial production.
And in the end, the yeast smell is gone. “We’re going to take the chemical we made, spin it down, centrifuge it, and distill it, so what gets shipped to you won’t have any trace of yeast or DNA,” Platt says.
Designing DNA Faster
Amyris’s scientists used to spend lab hours handling liquid and petri dishes, but with robots taking over these repetitive tasks, people can spend more time getting creative. Amyris’s DNA-compiler software, Thumper, is used to alter the genetic makeup of yeast strains by deleting and/or adding DNA. Giving access to its Genotype Specification Language (GSL), Amyris collaborated with Autodesk to build a plug-in for a genetic-engineering tool called Genetic Constructor, providing GSL with a visual design interface.
“The entire biology field is ripe for the introduction of more software,” Platt says. “People who used to spend all day in the lab moving liquid around can now sit at a computer and do a lot of that design work using programming languages like GSL. They can think more about the design rather than, ‘Where did I leave my pipette?’ or ‘Do I have clean tips?’”
Though robots and software are useful tools to bring molecules to market faster, it takes problem solving from scientists, too. For example, for the past 10 years, Amyris has been developing farnesene, a hydrocarbon building-block molecule used for renewable and biodegradable jet fuel and industrial oils.
“We invested a lot of effort in taking that molecule to market,” Chandran says. “There are so many molecules like farnesene we could be working on, but the investment for each molecule adds up. The question is, why does it take so long? Can we reduce the time to market for a single molecule? Instead of 10 years, can we do it in two? Instead of $100 million per molecule, can we do it with $5 million? Instead of working on a single molecule at a time, can we work on 100 molecules?”
Amyris signed a Technology Investment Agreement with the U.S. Department of Defense Advanced Research Projects Agency (DARPA) on the Milligrams to Kilograms Project to scale up production of compounds in kilograms or metric tons.
The company is also building a massive library of DNA parts, so when a scientist submits a design, the software can see if it has already been made. If not, it suggests how to build it.
“The analogy I’d give is LEGOs,” Chandran says. “The ends are analogous to each other, which is the same with DNA. If I give you five LEGO parts and say, ‘Make a small car,’ that’s easy. Now let’s say I give you 10,000 LEGO parts and ask you to make 1,000 LEGO cars. That’s difficult. But it’s a lot easier if I give you a database that contains all the LEGO parts and a software tool that matches which parts from the database need to go into which car, and it’s even easier with a robot picking all these LEGO parts and mixing them together in the right order.”
At Amyris, the funnel of activity starts with around 1.2 million yeast strains. They all go through the screening process, and winners move on to fermentation for further testing. “But we are still at times sending strains from a screening lab to a fermentation lab that are false positives, so having the historical expertise—as we do—to rule those out as candidates is mandatory to achieve success,” Chandran says.
To reduce false positives, scientists use a microfermenter as an interim screening step—a miniature lab that fits in the palm of their hands, featuring tiny IoT oxygen and pH sensors, plus channels that feed in sugar, base, or acid manipulations.
Chemistry for Good
Amyris’s first priority is to investors and customers, but it’s also motivated to take on projects that benefit the environment. One goal is to eliminate toxic chemicals, such as petroleum feedstock (found in cell phones), from supply chains.
“If we can find alternate sources of these chemicals through fermentation, where we are feeding sugar to a yeast cell and getting the same molecule with the same properties [without] compromising cost or performance, that’s a win for the environment,” Chandran says.
An example is squalane, a common cosmetics ingredient sourced from shark liver. “The quality of fermentation-derived squalane is the same as that for shark-liver squalane,” Chandran says. “So for the cosmetics industry, it was a very easy switch, with the added benefit that you don’t decimate the shark population.”
But these biotech breakthroughs don’t come easy; they require assuming a fair amount of risk. “I don’t think we’re going to succeed as an industry by just making the safe bets every single time,” Chandran says. “A failed experiment sometimes gives you a lot more answers than an experiment that actually worked. And, in the end, it helps you avoid mistakes going forward and leads to the development of the best yeast strains and innovations possible.” Just don’t expect any pizza.