Alexandre Zanghellini can’t help but think about what makes up the world around him. Sitting in a conference room, Zanghellini considered the paint on the walls, the table, the window shades, the plastic chairs. It’s all oil.
“The entire world is made from oil. We just don’t realize it,” he said.
Zanghellini’s job, as the CEO of Seattle-based synthetic biology company Arzeda, is to reconsider how we make the basic molecules that go into anything and everything in the human world. And he has a bias for processes that use living organisms. “The tools of biology, proteins, are better at doing chemistry than chemists,” he said.
If Amazon is “the everything store,” Arzeda is a kind of everything lab. The startup uses synthetic biology to make industrial ingredients — molecules that could go into anything from fertilizer to nylon clothes — with a 35-person team made up of scientists and computer engineers.
What’s at stake is the future of how people make everyday items in a world that doesn’t depend on oil. The contrarian approach has helped Arzeda land chemical conglomerates and eco-friendly startups as customers. But it also faces competition from the chemical industry and better-funded synthetic biology firms like Ginkgo Bioworks and Zymergen.
Zanghellini formed the company out of the University of Washington’s Institute for Protein Design, a department led by Prof. David Baker that’s also responsible for spinning out Cyrus Biotechnology, PvP Biologics and A-Alpha Bio.
Arzeda’s challenge is to prove that cells really can outperform chemists by making industrial inputs cleaner, faster, cheaper and at scale.
Brewing with machine learning
Zanghellini’s go-to metaphor to explain Arzeda’s process is brewing beer.
The company takes a bunch of cells (barley and hops) and puts them through a series of natural processes using enzymes (the malting, mashing and fermenting stage) to make something new (beer). Arzeda also uses powerful machine learning algorithms to think up new recipes (the brewmaster).
During a recent tour of Arzeda’s lab, Zanghellini explained that he’s changing how we make chemically-derived products, which tend to be molecules that:
- Don’t exist, but might be useful
- Are harmful to produce
- Can only be found in small quantities
Arzeda’s method had appealed to a range of industries. The Seattle-based company worked with DuPont Pioneer on enzymes that improved seed performance. It devised a process that uses cells, rather than oil, to create the building blocks of nylon for Invista, a subsidiary of Koch Industries. Amyris, the maker of sustainable beauty products, worked with Arzeda on a fragrance.
The startup has also received funding from multiple government organizations, including the National Science Foundation, the USDA, DARPA and the Department of Energy.
If you lose sleep over genetic engineering, this probably sounds like a nightmare. The staff seems to have a sense of humor about the whole tinkering-with-the-technology-of-life thing. During the visit, a sign on the kitchen counter read “Welcome to Arzeda! 7 days without a zombie out-brainz.”
Scientists have pointed out that the tools of synthetic biology could be used to create bioweapons. Moreover, environmentalists might object to partnerships with Koch Industries and DuPont, which represent the very chemical industry Arzeda hopes to disrupt.
But Zanghellini, who started the company in 2008 alongside Drs. Eric Althoff, Daniela Grabs and David Baker, insists that synthetic biology could be a force for good.
“This technology is really applicable to a wide variety of products,” he said. “We set out to have a positive impact, whether it’s sustainability, performance or better products.”
Designing proteins in the cloud
Arzeda’s headquarters are nestled in a line of auto repair shops in Seattle’s Interbay neighborhood. Behind the building is a network of train tracks where the fuselages of Boeing jets sit waiting to be transported. The setting seems out of place for a biotech company until you realize that Arzeda is fundamentally an industrial enterprise.
The office is divided into two floors: the scientists are downstairs, and the engineers and business team are upstairs. The engineers, who comprise about a third of the staff, work on an open floor that overlooks the lab below.
Arzeda’s algorithms think up ways to throw proteins at cells in order to create a molecular ingredient, thus turning cells into microscopic factories — which is exactly what they are.
The massive leaps in our ability to sequence and understand genetics has helped make Arzeda possible. But the developments in cloud computing and machine learning are arguably just as important.
Yih-En (Andrew) Ban, VP of computing at Arzeda, tried to explain the scale of the math behind the company’s algorithms. “We’re looking at 20 raised the power of ‘N,’ where ‘N’ is the length of the protein sequence.”
Ban told me that proteins can be composed of hundreds of amino acids, but it must have been obvious that I wasn’t not grasping the mathematical point. Zanghellini chimed in: “There are more combinatorial sequences here than there are atoms in the universe.”
Arzeda’s process works a bit like Google Maps, showing not just a route but multiple ones. “We can find shortcuts as well enable a new route,” said Rudesh Toofanny, Arzeda’s director of computational metabolic pathway design.
Options are important because biology is finicky. Not every process that the computers think up will work, but the algorithms improve over time.
Arzeda calls itself “the protein design company,” but the design is only the first step. After the engineers come up with a method for making a molecule, the scientists downstairs have to test it.
Once I’m downstairs, wearing an oversized lab coat and safety goggles, the team explains just how far our ability to tinker with DNA has come.
For starters, if you want to make your own genes, you can order a tray of DNA fragments online. And if you want to arrange those fragments to make any gene you want, you can use a robot that carefully assembles the DNA using sound energy.
None of this is cheap, and neither are the mass spectrometers that the company uses to test its designer proteins. Arzeda raised more than $15 million in 2017, which allowed it to build the lab, among other investments.
Arzeda focuses on proteins because they’re the worker bees of the cellular world, constantly changing the chemical makeup of their environment. You digest food and fight infections with proteins. Proteins are in most biologics, a class of drugs that made up a third of new medicine approved by the FDA last year.
There’s plenty of trial and error, but Arzeda says it can design, make and test 10,000 proteins per week. Once the company has a winning sequence, it’s time to brew.
At this stage, the beer metaphor becomes reality. The engineered cells are fed and left to ferment. Rather than producing alcohol, they create ingredients to be used in industrial products.
The last step, called downstream processing, is where chemistry comes into play. Arzeda’s chemists turn the “beer” into a usable product. They also need to convert the home-brew method to an industrial one, which is essential to bring down the costs.
Arzeda needs to create processes that are as cheap as those used in the chemical industry to be successful. Years ago, efforts to use synthetic biology to create biofuels from algae drew plenty of early-stage investments. Then the bubble burst, most of those companies disappeared, and investors became wary.
Arzeda plans to avoid that fate by going after higher value molecules or ones that companies are willing to pay to develop. Arzeda’s development deals with such companies involve royalties that could be lucrative if the process ever takes off.
Zanghellini, for his part, seems obsessed with proving the supremacy of cells.
“For 150 years, what we could make was limited by two things: could it be made from oil, and could you make it with chemistry,” he said. “Now we’re asking the question, ‘Can we make better molecules with biology instead of chemistry?’”