Abstract

Enzyme cascades are attractive alternatives to chemical synthesis to produce complex (natural) compounds. However, computer-aided retrobiosynthesis for predicting such cascades is limited by scarce biochemical reaction data and by models that largely neglect 3D enzyme-substrate interactions and comprehensive enzyme-substrate activity data. This results in missing enzyme candidates for many theoretically feasible biotransformations. In StRetro, we will develop a transferable structure-based machine learning (ML) approach for predicting enzyme-substrate activity based on rigorous encodings of enzyme active sites, substrates, and their interactions, supported by state-of-the-art structure prediction and evaluation. This model can then be used for the identification of enzyme candidates for such cascades. We will train our ML model on uniquely large and uniformly screened datasets of enzyme activity for the two biotechnologically relevant model enzyme systems esterases and methyltransferases. An efficient, transferable, and automated high-throughput screening assay will be established, allowing for a rapid extension to novel enzyme systems. All data will be deposited in open-access platforms following FAIR principles, vastly expanding available high-quality biochemical data. By enabling the automated and accurate prediction of cascaded enzymatic synthesis pathways via modular and sustainable processes, we anticipate that StRetro will fuel the transition to a sustainable and circular bioeconomy.

SEED FUND 3.0 Project coordinator

Dr. Till El Harrar
IBG-4 Structure-based Bioinformatics, Forschungszentrum Jülich/ Computational Pharmaceutical Chemistry and Molecular Bioinformatics, Heinrich Heine University Düsseldorf
email: t.el.harrar[at]fz-juelich.de

Partners

Dr. Till El Harrar, Prof. Dr. Holger Gohlke, Computational Pharmaceutical Chemistry and Molecular Bioinformatics, Heinrich Heine University Düsseldorf/ IBG-4 Structure-based Bioinformatics, Forschungszentrum Jülich
Prof. Dr. Jörg Pietruszka, Institut of Bioorganic Chemistry, Heinrich Heine University Düsseldorf
Dr.-Ing. Stephan Noack, IBG-1 Systems Biotechnology, Forschungszentrum Jülich

Funding period

01.11.2025 - 31.12.2026

Funding

StRetro is part of the NRW-Strategieprojekt BioSC and thus funded by the Ministry of Culture and Science of the German State of North Rhine-Westphalia.