Abstract

Considering the relevance of carbon source on PHA production cost, it is important to explore the full metabolic potential of microbial cells. Knowledge of the metabolic network operation under PHA-producing conditions will enable the rational streamlining of catabolic pathways to harness the greatest possible amount of carbon source for polymer synthesis. Knowing the distribution of fluxes is an important way to improve PHA production process towards efficient (and sustainable) polymer accumulation (Gomez et al., 2012). Metabolic modeling of Burkholderia sacchari was performed in this context. B. sacchari had its metabolic pathways proposed and utilized for flux analysis by employing measurements of biomass and extracellular products during the pseudo-stationary phases of a fed-batch cultivation. The resulted elementary fluxes were translated into a set of overall reactions and a dynamical mass model were then established. Such methodology allowed an unstructured model to be oriented by the main metabolic reactions. From this information and from the catabolic routes suggested in literature, a set of metabolic reactions were proposed and employed for a dynamical mathematical model and fluxes analysis, using the methodology proposed by Provost & Bastin (2004). Here, this methodology is applied to B. sacchari, cultivated in mineral media with glucose as carbon source. The proposed model is applied to the pseudo-stationary phases of the cultivation. Bioreactor assays were performed in a Biostat MD fermenter (B. Braun), 2 L total volume (4 L working volume). Dissolved oxygen was controlled over 20% (air saturation), pH at 7,0 and temperature at 30ºC. The model development steps were the following: 1) Proposition of a metabolic network; 2) Elementary flux modes by Metatool software; 3) Set of specific uptake and formation rates from experimental data; 4) Determination of flux in the elementary mode and 5) Mathematical model simulation by Matlab® tools. This methodology resulted on an unstructured model to be orientated by the main metabolic reactions. That approach allows, beyond the comprehension of the metabolic reactions, to estimate important state variables in process monitoring. This cannot substitute the phenomenological modeling, be it structured or unstructured, but certainly represents the total observed phenomena.