Nitrogen-fixing cyanobacteria for protein production: experimental and computational approach

To convey the food and feed industry towards a more sustainable approach, the production of vegetal-derived compounds, as proteins or pigments, represents a relevant topic of research. In that regard, nitrogen-fixing cyanobacteria are of interest thanks to their capability to fix atmospheric nitrogen into valuable N-based macromolecules. Looking for a potential industrial exploitation, laboratory experiments were conducted to assess Nostoc PCC 7120 cultivation features, both in batch and continuous configurations. Discontinuous cultivation was conducted in a 275-L pilot scale bioreactor in order to assess species performances in a larger operational scale, while 200-mL continuous photobioreactors were used to evaluate the effects of operating variables on harvested biomass and production of valuable compounds. With respect to enlightenment conditions, an optimal light intensity of 550 µmol m-2 s-1 was found, as the best compromise between biomass productivity and embedded valuable compounds. Secondly, a detailed study on diazotrophic growth performances was conducted, by varying the amount and source of nitrogen within the culture environment. The exploitation of such peculiar metabolic path seems to be a promising choice, since it contributes to lower nutrients-related capital costs as well as improve cultivation outcomes. Moreover, successful industrial exploitation relies also on the possibility of predicting such outcomes. Therefore, the development of a simulation model for two cyanobacterial species, namely Anabaena PCC 7122 and Nostoc PCC 7120, was considered. In detail, the model embeds the Droop theory based on the nutrient quota, accounting for the variation of biomass growth rate based on the amount of nutrients within the biomass itself. Compared to the various predictive models already available in literature, focusing only on overall biomass, the introduction of such peculiar kinetics allows then to simultaneously account for both the two typologies of cells characterizing the species in question, namely vegetative cells and heterocysts. According to the experimental background of this work, the model was developed for continuous operations, predicting cultivation outcomes with respect to the operating residence time. The derived computational outcomes seem in line with data collected in previous experimental campaigns, which were also used for the fitting of additional unknown model parameters.