From recyclable plastics to renewable fuels to smarter cellular transport systems, the work unfolding in UCSF School of Pharmacy labs reflects a powerful premise: By designing the molecular machines of life, scientists can help design a more sustainable future.

UCSF School of Pharmacy Professors William DeGrado, PhD, and Tanja Kortemme, PhD, have landed three out of five inaugural awards given by the U.S. National Science Foundation (NSF) that aim to strengthen the bioeconomy by translating advances in AI into real-world biotechnologies that support domestic manufacturing, sustainable materials, and resilient supply chains. The $32 million investment across the five awards is through the NSF Use-Inspired Acceleration of Protein Design initiative.

For decades, scientists have studied proteins as the molecular machines that power biology. Now, with breakthroughs in AI and machine learning, researchers are moving beyond predicting protein structures to designing entirely new proteins with tailor-made functions. That shift is already underway at UCSF, with researchers advancing technologies that improve human health while supporting a more competitive economy.

Designing biology from first principles

DeGrado and Kortemme are co-leading the protein design components of three three-year NSF-funded programs central to the emerging bio-based economy:

• “Transporters with Transformers” (with Koliber Biosciences)
• “Programmable Small Molecule Biosynthesis” (with Purdue University, UC Berkeley, and Stanford University)
• “De Novo Design and Evolution of Enzymes for Biomass Upcycling to Surfactants and Fuels” (with UC Santa Barbara)

DeGrado, professor in the Department of Pharmaceutical Chemistry, has long focused on de novo protein design — building proteins from first principles to test and expand our understanding of how they fold and function. That approach now provides a foundation for industrial biotechnology.

“De novo design allows us to start with a blank slate,” DeGrado said. “Rather than modifying what evolution has already produced, we can build proteins specifically for the task at hand, whether that’s stabilizing a membrane transporter or catalyzing a reaction that doesn’t exist in nature. That level of control opens the door to entirely new kinds of industrial chemistry.”

Reprogramming nature’s assembly lines

In the “Programmable Small Molecule Biosynthesis” project, Kortemme and collaborators are redesigning nature’s molecular assembly lines.

In living systems, some enzymes operate like linked stations on a factory floor assembly line, each carrying out a step in a chemical transformation. But those systems evolved to meet the needs of an organism, not the demands of industry.  

With AI-enabled protein design, researchers can streamline and recombine those biological components to produce entirely new materials. One goal: bacteria that manufacture biodegradable, heat-resistant plastics that can be broken down and recycled into their original building blocks, addressing both performance and sustainability challenges.

“In nature, these systems function like an assembly line, where each part performs a specific job,” said Kortemme, the school’s vice dean of research. “What we want to do is take those parts and recombine them in many different ways, like modular components, so we can build new assembly lines that make exactly the products we want.”

Because proteins govern virtually every biological process, the ability to program them with precision represents a platform technology with implications for medicine, agriculture, advanced materials, and more.

Upcycling biomass, strengthening supply chains

In another NSF-funded collaboration, DeGrado’s team is designing new enzymes to convert plant biomass into higher-value products such as fuels, lubricants, and surfactants. Instead of relying on fossil fuels or energy-intensive chemical processes, these approaches harness engineered biology to transform abundant renewable materials.

The “Transporters with Transformers” project addresses a longstanding bottleneck in biomanufacturing: moving small molecules efficiently across cell membranes to increase production yields.

Across all three programs, the emphasis is translation — bringing AI-based protein design from the laboratory into practical industrial use.

“For this field to impact the bioeconomy, protein design has to move beyond proof-of-concept,” DeGrado said. “We’re focused on building systems that are robust, scalable, and ready for real-world manufacturing environments. If we can do that, we reduce dependence on fragile supply chains and expand domestic capacity for producing essential materials.”

A cross-disciplinary strength

The NSF initiative was launched using an “ideas lab” model, bringing together academic and industry leaders to co-develop ambitious but achievable projects and creating a national network of cross-sector teams.

For Maggie Horst, PhD, a postdoctoral scholar in DeGrado’s lab, the school’s long-standing expertise in enzymology and structural biology is a distinct advantage. Rather than relying on enzymes shaped by millions of years of evolution, her work focuses on designing them from scratch, and tailoring them to withstand industrial conditions.

“Natural enzymes evolved through contingency and for the needs of organisms, not for industrial processes,” Horst said. “When we design enzymes de novo, we can prioritize what industry values — high concentrations, high stability, and precise selectivity — without the evolutionary baggage.”

By integrating machine learning with physics-based models and rapid experimental validation, UCSF researchers can iterate quickly between computational prediction and laboratory testing. That tight feedback loop — computation informing experiment, and experiment refining computation — is what makes AI-driven protein design increasingly practical and scalable.

An inflection point for biotechnology

The school, and UCSF at large, has a history of pioneering technologies that reshaped biotechnology, from genetic engineering to structure-based drug design. Faculty leaders see AI-enabled protein design as the next inflection point. As global competition intensifies, initiatives like NSF USPRD aim to ensure that breakthroughs translate into domestic manufacturing capacity and economic resilience.

“AI protein design isn’t limited to one disease or one application,” said Kortemme, a faculty member in the Department of Bioengineering and Therapeutic Sciences (BTS). “Anywhere biology plays a role, this approach has the potential to make a difference.”