A perfusion bioreactor allowing continuous extraction of secondary metabolites was designed and challenged for Eschscholtzia californica plant cell suspensions. Four sedimentation columns mounted inside a 2.5-L bioreactor separated single cells and cell aggregates from the culture medium. Cells were elicited with chitin at day 4 and the liquid medium free of cells and debris was then continuously pumped to the extraction columns containing fluidized XAD-7 resins, and then recirculated back to the cell suspension. A medium upward velocity corresponding to cell sedimentation velocity maintained a stable cell/medium separation front in the columns for sedimented cell volume (SCV) of 90% (70% packed cell volume, PCV). Two perfusion bioreactor cultures of 10 and 14 days were performed. A maximum dilution rate of 20.4/day was reached from day 4 to day 6, and was then reduced to 5/day at day 9 for 55% SCV. Control cultures were performed without and with free extraction resins into the cell suspension. Perfusion cultures showed similar specific growth rates of 0.24 +/- 0.04/day before and after elicitation. However, production level in the perfusion cultures was similar to that from the culture without resins with a maximum of 2.06 micromole/gDW total alkaloids, with 1.54 micromole/gDW in the resins. Cultures with free resins resulted in 30.94 micromole/gDW with 28.4 +/- 8.8 micromole/gDW in the resins. Difference in the cells nutritional state from elicitation was identified as a major cause in the production reduction. However, pathway to chelilutine was favored in the continuous extraction culture.

SummaryA dynamic model for plant cell metabolism was used as a basis for a rational analysis of plant production potential in in vitro cultures. The model was calibrated with data from 3-L bioreactor cultures. A dynamic sensitivity analysis framework was developed to analyse the response curves of secondary metabolite production to metabolic and medium perturbations. Simulation results suggest that a straightforward engineering of cell metabolism or medium composition might only have a limited effect on productivity. To circumvent the problem of the dynamic allocation of resources between growth and production pathways, the sensitivity analysis framework was used to assess the effect of stabilizing intracellular nutrient concentrations. Simulations showed that a stabilization of intracellular glucose and nitrogen reserves could lead to a 116% increase in the specific production of secondary metabolites compared with standard culture protocol. This culture strategy was implemented experimentally using a perfusion bioreactor. To stabilize intracellular concentrations, adaptive medium feeding was performed using model mass balances and estimations. This allowed for a completely automated culture, with controlled conditions and predefined decision making algorithm. The proposed culture strategy leads to a 73% increase in specific production and a 129% increase in total production, as compared with a standard batch culture protocol. The sensitivity analysis on a mathematical model of plant metabolism thus allowed producing new insights on the links between intracellular nutritional management and cell productivity. The experimental implementation was also a significant improvement on current plant bioprocess strategies.