Incorporation of Cryptic Plasmid Energetics Improves Genome-Scale Metabolic Predictions in Probiotic E. coli Nissle...

Escherichia coli Nissle 1917 (EcN) is a well-characterized Gram-negative probiotic distinguished by its unique, strain-specific physiology. Genome-scale metabolic models (GEMs) are powerful tools for elucidating metabolic traits and predicting genotype-phenotype relationships. Although several EcN GEMs have been published, none have explicitly exploited its probiotic physiology. Here, we present a plasmid-specific module that can be incorporated into EcN GEMs, which, for the first time, considers the energetic costs associated with its cryptic plasmids. Using COBRA methodologies and possibilistic metabolic flux analysis, we show how the plasmid-module inclusion improves biomass and overall flux predictions in an additional, manually curated EcN model as well as in previous EcN reconstructions. Then, the different EcN reconstructions were systematically compared to evaluate the trade-off between model refinement depth and predictive performance. The analysis revealed that once expanded to include the plasmid-related costs, increased level of curation in the different GEMs does not necessarily enhance quantitative accuracy and that predictive reliability depends on both computational methodology chosen and experimental contexts. Metabolomic profiling under gut microbiota medium and anaerobic conditions further showed that EcN exhibits a distinctive metabolic phenotype, characterized by elevated amino acid consumption and enhanced short-chain fatty acid production, which is captured by the evaluated metabolic models. These findings highlight the unique probiotic physiology of EcN and demonstrate the utility of metabolic modelling for reproducing and exploring such traits. Overall, this study provides a quantitatively reliable and physiologically relevant framework for modelling E. coli Nissle 1917 while considering its cryptic plasmids, supporting advances in probiotic engineering, synthetic biology and bioprocess design.