Development of a Genome-Scale Metabolic Model and Phenome Analysis of the Probiotic Escherichia coli Strain Nissle 1917

Abstract

Escherichia coli Nissle 1917 (EcN) is an intestinal probiotic that is effective for the treatment of intestinal disorders, such as inflammatory bowel disease and ulcerative colitis. EcN is a representative Gram-negative probiotic in biomedical research and is an intensively studied probiotic. However, to date, its genome-wide metabolic network model has not been developed. Here, we developed a comprehensive and highly curated EcN metabolic model, referred to as iDK1463, based on genome comparison and phenome analysis. The model was improved and validated by comparing the simulation results with experimental results from phenotype microarray tests. iDK1463 comprises 1463 genes, 1313 unique metabolites, and 2984 metabolic reactions. Phenome data of EcN were compared with those of Escherichia coli intestinal commensal K-12 MG1655. iDK1463 was simulated to identify the genetic determinants responsible for the observed phenotypic differences between EcN and K-12. Further, the model was simulated for gene essentiality analysis and utilization of nutrient sources under anaerobic growth conditions. These analyses provided insights into the metabolic mechanisms by which EcN colonizes and persists in the gut. iDK1463 will contribute to the system-level understanding of the functional capacity of gut microbes and their interactions with microbiota and human hosts, as well as the development of live microbial therapeutics.

Keywords: Escherichia coli Nissle 1917; flux balance analysis; metabolic network model; phenome analysis; probiotics.

Figure 1

Comparison of phenotype microarray (PM) tests of E. coli strains Nissle 1917 and K-12 MG1655. (A) Carbon sources (PM1 and PM2). (B) Nitrogen sources (PM3). (C) Inhibitory compounds, such as antibiotics, antimetabolites, and other inhibitors (PM11 to PM20). The numbers indicate the number of nutrients on which cells grew.

Figure 2

Comparison of carbon and nitrogen source utilization of E. coli strains Nissle 1917 and K-12 MG1655. Growth curves in all the cells are shown for E. coli Nissle 1917 (red) and K-12 MG1655 (green). (A) Carbon sources (PM1). (B) Carbon sources (PM2). (C) Nitrogen sources (PM3).

Figure 3

Examples of different metabolic gene clusters in the genomes of E. coli strains Nissle 1917, CFT073, and K-12 MG1655. Gene clusters for N-acetyl-d-galactosamine metabolism (A), 2-deoxy-d-ribose metabolism (B), L-sorbose metabolism (C), 5-keto-d-gluconic acid metabolism (D), O-antigen biosynthesis (E), anaerobic utilization of α-ketoglutarate (AKG) and L-lactate (F), and a part of methionine salvage pathway (G). Gene names in red are for Nissle genes, the reactions of which were included in the Nissle metabolic model.

Figure 4

Model predictions of growth capability in the nutrient sources under anaerobic conditions. (A) Schematics of the predicted flux distributions. The nine carbon sources simulated are in red. Reactions are in the box. Arrows denote directions of the predicted metabolic fluxes, and dashed arrows indicate the multi-step reaction. Metabolites colored above arrows indicate cofactors (ADP, ATP, NAD, and NADH). Some metabolites (pyruvate and acetate) are duplicated on the map for clearer visualization and are labeled with an asterisk (*). (B) Summary of the simulation for the utilization of nine carbon sources. “Reactions for ATP production” and “Reactions for NAD⁺ regeneration” are reactions responsible for >95% of total ATP production and total NAD⁺ regeneration, respectively, in descending order. Abbreviations are given in Supplementary Table S4. For flux balance analysis (FBA) under anaerobic conditions, the upper limits of oxygen and each of the carbon sources were set to zero and 20 mmol gDCW⁻¹ h⁻¹, respectively.