End-to-end continuous E. coli biomanufacturing of recombinant proteins via integrated and automated digital control

Abstract Background Continuous biomanufacturing is increasingly recognized for its potential to improve sustainability, efficiency, and production continuity in manufacturing of biologics. However, its implementation for the production of recombinant proteins in microbial systems, particularly Escherichia coli , remains limited by persistent challenges, including genetic instability, limited secretion capacity, and process heterogeneity. These constraints have so far impeded the development of robust and scalable continuous E. coli -based production platforms. Results Here, we demonstrate the development of a fully integrated, continuous, digitally controlled and automated biomanufacturing platform utilizing E. coli as microbial production host. As model protein, recombinantly expressed Protein A was used, which was released into the culture supernatant. The end-to-end system comprises upstream cultivation, continuous primary recovery, multi-column chromatography, advanced process automation, and model-based control. The upstream module employs a two-stage chemostat using a genetically engineered and stabilized strain, achieving robust long-term productivity (> 500 h), genetic stability, and minimal cell lysis during Protein A production. An intensified Design of Experiments (iDoE) strategy enabled efficient screening of the design space in a single continuous fermentation, allowing determination of process conditions associated with the highest Protein A titers. Downstream, the culture supernatant was continuously collected by removing cells via Cleaning in Place (CIP) - enabled microfiltration with vibrating membranes. The clarified supernatant was subsequently subjected to real-time inline product analytics and automated continuous chromatography adapting to dynamic process conditions. A modular digital process twin was developed to enable model-based decision support across the entire process chain. The modeling architecture integrates time-resolved sensor data with mechanistic mass-balance relationships to describe dynamic behavior across upstream and downstream unit operations. This framework enabled model-based determination of chromatography switch times by incorporating dynamically varying upstream conditions and inline Protein A measurements. A techno-economic analysis showed 25–39% reductions in capital investment and 26–37% lower operating costs compared to fed-batch strategies, with 45% decrease in facility energy demand. Conclusion This work represents the first demonstration of a fully digitized continuous E. coli manufacturing platform and provides a scalable blueprint for biomanufacturing transformation. The system offers a modular architecture and demonstrates sustainability benefits, providing a compelling foundation for industrial adoption.