Advanced workflows for the systematic identification of metabolic optimization targets in DBTL-cycles: A demonstrator for producing aromatic compounds in C. glutamicum

The establishment of rational Design - Build - Test – Learn (DBTL) cycles based on miniaturization, parallelization, automation and digitalization enables a reduction in development times and an increase in reproducibility and effectiveness of microbial strain construction for competitive biomanufacturing.

Here, we present an advanced strain design and analysis workflow to systematically improve the understanding of microbial production pathways and to enable faster identification of metabolic optimization targets (Fig. 1). As a first demonstrator, we focus on aromatic compound production in Corynebacterium glutamicum with the specific goal to enhance the carbon flux towards l-tyrosine through the native shikimate pathway.

The workflow harnesses the power of in silico metabolic modelling with parallelized strain phenotyping. A standardized library of small- to genome-scale stoichiometric network models of C. glutamicum in combination with thermodynamic data and available kinetic information was established. For the simulation of intracellular metabolic states, using parsimonious flux balance analysis, a python-based workflow was set up. Quantitative and fast characterization of growth and production phenotypes is realized by employing robotic-assisted micro-cultivation experiments with (un)targeted metabolomics via LC/ GC-ToF-MS. Raw data processing and modelling is performed using recently developed python tools [1]. Resulting specific rate estimates are applied to further constrain reaction fluxes for in silico strain design. The wider range of potential metabolic engineering targets contributing to enhanced l‑tyrosine production will then be constructed using MoClo-based workflows on our AutoBioTech platform [3].

Ultimately, a standardized and validated modelling toolbox enables the design and analysis of a broad range of production strains and thus represents an essential component for the operation of biofoundries.

Strohmeier D et al. 2022, estim8: Parameter estimations for Dymola and FMU models. (https://pypi.org/project/estim8/)https://www.biooekonomierevier.de/Innovationslabor_AutoBiotech