Enzymes are essential to a wide range of industrial processes. As biocatalysts, they play a key role in producing active pharmaceutical ingredients, specialty chemicals and ingredients for cosmetics, besides being vital to the food industry. Enzymes allow many chemical reactions to occur under milder conditions, thereby saving energy and making existing processes more
efficient and environmentally friendly.
Molecular biology, genetic engineering and bioinformatics can be used to optimize enzymes for industrial applications. Targeted changes to an enzyme's genetic blueprint can make it work faster, keep it reliable even under demanding process conditions or help it control chemical reactions with greater precision. Modern computational methods and artificial
intelligence can support this process by helping identify suitable enzyme variants more quickly.
New approaches to enzyme design
Nevertheless, developing custom enzymes remains a resource-intensive process. The complex relationships between genetic sequence, three-dimensional structure and enzymatic function make accurate predictions difficult. One major bottleneck is the lack of experimental data on how genetic changes affect enzyme function.
This is where the ZenBio project comes in. Researchers at the Fraunhofer Institute for Cell Therapy and Immunology, Branch Bio-analytics and Bioprocesses IZI-BB in Potsdam are developing an automated platform that uses cell-free protein synthesis to generate large volumes of sequence-function data for AI-assisted enzyme design. This will accelerate the process of producing and analyzing enzyme variants and developing them into high-performance
biocatalysts for industrial applications.
The project addresses the specific demands of industrial processes. “Enzymes are naturally adapted to the conditions in the organisms where they occur,” explains Simon Krebs, a biotechnologist at Fraunhofer IZI-BB. “Industrial processes, on the other hand, often impose very different demands. Enzymes may need to withstand higher temperatures, altered pH levels or organic solvents. In the ZenBio project, we are tailoring enzymes to these applications, thereby unlocking new manufacturing processes that are difficult to achieve using conventional chemistry.”
Cell-free protein synthesis and AI accelerate enzyme design
“AI can only learn from what has been measured experimentally,” explains Benedikt Dechamps, an industrial engineer at Fraunhofer IZI-BB. “That’s why we aim to produce as many enzyme variants as possible and study how they work. This will allow us to systematically build a data set that the AI can learn from and use to propose promising new variants.”
What sets this approach apart is its use of cell-free protein synthesis, which allows the team to screen large numbers of enzyme variants in a short time. The researchers combine the genetic blueprint for the desired enzyme with a cell extract containing all the essential components needed to produce it. The enzyme is then made directly in the extract. In contrast, conventional cell-based methods take several days because the enzyme-producing cells must first be cultured. In addition, some enzyme variants are difficult to produce in living cells or do not fold into a functional form.
Because the enzymes are produced in a cell extract rather than in living cells, the researchers can miniaturize the process and fine-tune the reaction conditions. In the future, a robotic pipetting station will produce and analyze up to 1,500 different variants of an enzyme in parallel. Within just a few hours, the team can find out if each variant withstands factors such as temperature, pH or a solvent. The AI platform developed in the project uses the resulting data to identify relationships between enzyme sequences and properties and to propose promising new variants.
“Every experiment teaches us something new that we can apply to the next design cycle,”
explains Holly Corbett, an AI expert at Fraunhofer IZI-BB. “This allows us to engineer enzymes step by step for industrial applications, thereby making an important contribution to a high-performing, sustainable chemical industry. To me, that’s engineering craftmanship at a
molecular level.”
The team has already completed the first phase of the automated workflow. They will now further miniaturize and automate the platform and test it with an initial class of enzymes.
Fraunhofer researchers plan to launch a company
Krebs and Dechamps met through AHEAD, a Fraunhofer funding program for startups. Corbett later joined the team as lead for the AI platform. The German Federal Ministry for Economic Affairs and Energy (BMWE) is funding the project through its exist Research Transfer program. This funding will enable the team to further develop and automate the platform and validate it using an initial class of enzymes. The team then plans to launch its first pilot projects with industry clients and establish the spin-off ProEvo.
Fraunhofer IZI-BB
The biotechnology solutions developed by Fraunhofer IZI-BB in Potsdam focus on human, animal and environmental health. The institute combines bioanalytics, biotechnology and data science to create integrated platform technologies that support diagnostics, drug development, food and environmental analysis. Fraunhofer IZI-BB collaborates with industry, academia and clinical partners to develop robust, sustainable solutions that meet regulatory requirements and facilitates their transition into practical applications.
Fraunhofer-Gesellschaft
