(555c) Multiscale Metabolic Modeling of the Oleaginous Microalga Chlorella vulgaris in a Photobioreactor | AIChE

(555c) Multiscale Metabolic Modeling of the Oleaginous Microalga Chlorella vulgaris in a Photobioreactor


Tibocha-Bonilla, J. D. - Presenter, Universidad Nacional de Colombia
Zuniga, C., University of California
Broddrick, J. T., University of California, San Diego
Zengler, K., University of California
Godoy-Silva, R. D., Chemical and Biochemical Processes Research Group. Universidad Nacional de Colombia
Metabolic modeling of microalgae has been a resourceful tool to predict and analyze metabolic behavior of organisms for almost one decade. In the case of oleaginous microalgae, some genome-scale metabolic models have been generated and improved to study their metabolism [1]. However, little effort has been made on applying metabolic models to control large-scale cultures of industrial interest. For such a purpose, nutrients starvation as well as light uptake and attenuation have been identified as drivers of the process performance. In this work, we combined the latest genome-scale metabolic model of Chlorella vulgaris [2] with a kinetic model that considers light uptake, photoinhibition, nitrogen and carbon uptake, metabolite-specific carbon allocation (carbohydrates, lipids, and nucleotides) and reactor geometry. We successfully predicted growth under different conditions, including various light intensities. The developed model is highly robust, which enables to design strategies based upon different light sources and arrangements as well as culture timing for increased lipid productivity.

The metabolic model of C. vulgaris, iCZ843, was combined with mass and energy phenomena according with previously identified physical and physiological mechanisms affecting growth. Included mechanisms were light attenuation, light uptake, photoinhibition, nitrogen and carbon uptake kinetics, and carbon allocation (carbohydrates, lipids, and nucleotides). Genome-scale model simulations were performed using the Gurobi Optimizer Version 5.6.3 (Gurobi Optimization Inc.) solver in MATLAB (The MathWorks Inc.).

First, we assessed the ability of our model to predict intracellular concentrations at different nitrogen availability conditions. In their study, Adesanya et al. [3] cultivated Chlorella vulgaris at two different initial nitrogen concentrations and tracked macromolecular content of the cells through the culture time. Initial nitrate concentrations of 0.35 and 1.89 mM were employed to evaluate the impact of nitrogen availability on carbon allocation and growth. This allowed us to test our model for its capability of predicting intracellular concentrations of molecules of interest, specifically triacylglycerols (TAGs). Our model includes a mechanistic approach to the modeling of carbon distribution across the cell, which is bound to provide it with the ability of predicting the microscopic and macroscopic effect of nitrogen concentration.

We used one set of reported data for the regression of strain-specific parameters and simulated a second scenario to test for predictive capability. In the first scenario, a relatively low initial nitrate concentration in the media (roughly half that of standard BBM medium) caused the size of the internal nitrogen pool to decrease steadily throughout the time of culture. Although the microalga was not able to replenish its nitrogen reserves, lipid accumulation was triggered only 200 h after nitrogen was depleted in the medium. More accurately, nitrogen depletion from the medium signifies the beginning of the end of exponential growth, rather than the end itself. A similar behavior was obtained by Mansouri et al. [5] under a comparable setup. They reported that exponential growth was maintained for the first 96 h of growth, time after which growth gradually stopped until their last recorded instance at 168 h.

To test for predictive capability at different irradiances, data reported by Kim et al. [4] at an irradiance of 848 μE/m2 s was used for the regression of parameters (see Methods), while data at 30, 55, 80, 197 and 476 μE/m2 s were employed for model validation. Simulated results about growth and nitrate uptake at irradiance of 848 μE/m2 s followed closely the reported data of total biomass concentrations, and gave insight into both internal and external concentrations of important nutrients and macromolecules, along with circadian oscillations and global growth after nitrogen depletion.

To date, metabolic modeling had been focused on intracellular metabolic functions, but none effort towards modeling of large-scale photobioreactors had been made. In this work we generated a multiscale metabolic model that can predict global reactor growth behavior after obtaining a few strain-specific parameters, which allowed to analyze differential kinetics at various irradiances. Such predictions will allow the bioengineer to a priori perform sensitivity analyses of reactor variables to decide upon the best light strategy for certain geometries.


[1] Tibocha-Bonilla JD, Zuñiga C, Godoy-Silva RD, Zengler K. Advances in metabolic modeling of oleaginous microalgae. Biotechnol Biofuels. 2018;11:241.

[2] Zuñiga C, Levering J, Antoniewicz MR, Guarnieri MT, Betenbaugh MJ, Zengler K, et al. Predicting dynamic metabolic demands in the photosynthetic eukaryote Chlorella vulgaris. Plant Physiol. 2017;176:450–62.

[3] Adesanya VO, Davey MP, Scott SA, Smith AG. Kinetic modelling of growth and storage molecule production in microalgae under mixotrophic and autotrophic conditions. Bioresour Technol. 2014;157:293–304.

[4] Kim J, Lee JY, Lu T. A model for autotrophic growth of Chlorella vulgaris under photolimitation and photoinhibition in cylindrical photobioreactor. Biochem Eng J. 2015;99:55–60. W. Black, E.B. White, The Elements of Science, third ed., MacCluski, New York, 1987.

[5] Mansouri M. Predictive modeling of biomass production by Chlorella vulgaris in a draft- tube airlift photobioreactor. Adv Environ Technol. 2017;3:119–26.