As metabolic models grow larger and more complex and methods for the thermodynamic analysis of these models emerge, knowledge of the Gibbs free energy of formation (D_fG^'°) and reaction (D_rG^'°) for the compounds and reactions in these models is becoming essential. We introduce an expanded version of the group contribution method for the estimation of D_fG^'° and D_rG^'° in biochemical systems. Gibbs energy contributions are fit using 74 distinct molecular substructures and 11 interaction factors against a training set of 644 reactions and 224 compounds. We utilized bootstrap analysis to determine the accuracy of the methodology in estimating D_rG^'° and D_fG^'° for reactions and compounds not included in the training set of the fitting. Based on the results of the bootstrapping, the standard error involved in these estimations was 2.09 kcal/mol.

The new group contribution method is applied to estimating D_fG^'° and D_rG^'° for the metabolites and reactions in a genome-scale metabolic model of E. coli. The new estimated D_fG^'° and D_rG^'° values were utilized in a Thermodynamics-based Metabolic Flux Analysis (TMFA) of E. coli metabolism (1). TMFA is a form of metabolic flux analysis that includes thermodynamic constraints in addition to the mass balance constraints of metabolic flux analysis. Using TMFA we determined the thermodynamically feasible ranges for the reaction fluxes, reaction , and metabolite activities in the E. coli metabolism. The results indicate that most reactions essential for growth in E. coli have a sufficiently favorable and the physiological range of metabolite activities is sufficiently flexible to allow thermodynamic feasibility of these reactions under a wide range of conditions. A small number of reactions emerge as significant thermodynamic bottlenecks in the metabolic model, and of these reactions and the activities of the metabolites involved in these reactions are constrained to narrow ranges. Overall, the E. coli metabolism was found to be remarkably resilient to the thermodynamic constraints.

1. Henry, C. S., L. J. Broadbelt and V. Hatzimanikatis. 2007. Thermodynamics-based metabolic flux analysis. Biophys J. 92(7):1792-1805.