Physicochemical and metabolic constraints for thermodynamics-based stoichiometric modelling under mesophilic growth conditions

Abstract

24Metabolic engineering in the post-genomic era is characterised by the development of new 25 methods for metabolomics and fluxomics, supported by the integration of genetic engineering 26 tools and mathematical modelling. Particularly, constraint-based stoichiometric models have 27 been widely studied: (i) flux balance analysis (FBA) (in silico), and (ii) metabolic flux analysis 28 (MFA) (in vivo). Recent studies have enabled the incorporation of thermodynamics and 29 metabolomics data to improve the predictive capabilities of these approaches. However, an 30 in-depth comparison and evaluation of these methods is lacking. This study presents a thorough 31 analysis of four different in silico methods tested against experimental data (metabolomics and 32 13 C-MFA) for the mesophile Escherichia coli and the thermophile Thermus thermophilus. In 33 particular, a modified version of the recently published matTFA toolbox has been created, 34 providing a broader range of physicochemical parameters. In addition, a max-min driving force 35 approach (as implemented in eQuilibrator) was also performed in order to compare the 36 predictive capabilities of both methods. 37Validating against experimental data allowed the determination of the best 38 physicochemical parameters to perform the TFA for E. coli, whereas the lack of metabolomics 39 data for T. thermophilus prevented from a full analysis. Results showed that analytical 40 conditions predicting reliable flux distributions (similar to the in vivo fluxes) do not necessarily 41 provide a good depiction of the experimental metabolomics landscape, and that the original 42 matTFA toolbox can be improved. An analysis of flux pattern changes in the central carbon 43 metabolism between 13 C-MFA and TFA highlighted the limited capabilities of both approaches 44 for elucidating the anaplerotic fluxes. Finally, this study highlights the need for standardisation 45 in the fluxomics community: novel approaches are frequently released but a thorough 46 comparison with currently accepted methods is not always performed. 47 Keywords 48 Constraint-based modelling, fluxomics, metabolomics, thermodynamics. Predictive capabilities of thermodynamics-based stoichiometric approaches 3 49 Author summary 50 Biotechnology has benefitted from the development of high throughput methods characterising 51 living systems at different levels (e.g. concerning genes or proteins), allowing the industrial 52 production of chemical commodities (such as ethylene). Recently, focus has been put on 53 determining reaction rates (or metabolic fluxes) in the metabolic network of certain 54 microorganisms, in order to identify bottlenecks hindering their exploitation. Two main 55 approaches can be highlighted, termed metabolic flux analysis (MFA) and flux balance analysis 56 (FBA), based on measuring and estimating fluxes, respectively. While the influence of 57 thermodynamics in living systems was accepted several decades ago, its application to study 58 biochemical networks has been only recently enabled. In t...