Growth-resolved genome-scale metabolic modeling of Priestia megaterium SR7 validated by chemostat and 13-C flux analysis
Chang, K. Y. W.; Song, Y.; Hing, N. Y. K.; Vethathirri, R. S.; Wang, Y.; Thompson, J. R.
Show abstract
Priestia megaterium SR7 is a promising candidate chassis for bioprocess engineering, but its development is limited by the availability of condition-grounded, mechanistic models that can translate experimental measurements into predictive design hypotheses. Here, we present PMSR7, a genome-scale metabolic model for SR7, and evaluate it under a growth-resolved chemostat framework spanning a dilution-rate series. Stable steady states were established across the growth regime, with the highest dilution rate (D = 1.1538 h-{superscript 1}) excluded from growth interpretation due to biomass collapse. Extracellular carbon fluxes were quantified by NMR and reported as mean {+/-} SD, providing an experimental basis for model comparison. PMSR7 was benchmarked using MEMOTE against representative reference reconstructions, supporting structural consistency suitable for constraint-based analyses. Under growth-resolved simulations, ATP demand scaled linearly with growth rate, enabling inference of maintenance-energy behavior across the regime. Growth-dependent feasibility and magnitude of overflow secretion were evaluated for acetate, lactate, and formate using feasible-space analyses, highlighting both agreement and regime-sensitive limitations. Finally, growth-resolved leucine and valine production was assessed in both raw and fold-change space, with experimental means compared against median-based summaries of sampled model distributions to account for feasible-space skew. Together, these results establish PMSR7 as a reproducible, quality-benchmarked platform for SR7 chassis development and provide a framework for iterative experimental integration in non-model organisms, where the dominant challenge is achieving congruence between measured physiology and model-feasible behavior.
Matching journals
The top 5 journals account for 50% of the predicted probability mass.