Physiological response of the methylotrophic yeast Pichia pastoris to recombinant protein expression

The production of recombinant proteins for use as human or animal therapeutic purposes requires a manufacturing process that is both in control and rugged as mandated by the US Food and Drug Administration. When expressing the cytokine roIFN-τ in the methylotrophic yeast Pichia pastoris the yield of product was highly variable between production runs resulting in a process that was neither in control nor rugged. The purpose of the work presented herein was to develop a stable and predictable fermentation process which meets requirements of the US Food and Drug Administration, the governing body for biologic production and distribution. The central hypothesis guiding this work was: the expression of roIFN-τ places a metabolic burden on P. pastoris. To test this hypothesis a stable fermentation process was developed based upon a quasi-steady state fed batch fermentation method that assumed the production of product was growth associated. We observed that the yield of product dropped after what was coined the decoupling point. Therefore, we tested the working hypothesis: protease levels increase after the decoupling point and the enzymes degrade product at a greater rate than synthesis. It was determined from this work that metallo, aspartyl and serinyl proteases are active against roIFN-τ but that no significant increase in their rate of synthesis occurs after the decoupling point. The final hypothesis tested was: the energy balance of the cell regulates production of recombinant proteins. The adenylate pools were measured for cultures growing on glycerol or methanol, for cultures starved after growth on either glycerol or methanol, and for cultures expressing β-galactosidase. The adenylate energy charge (AEC) was found to be different for the two growth conditions and found to vary over the growth curve of each culture. These results show that P. pastoris was not able to sustain it\u27s adenylate energy charge in the same manner as Saccharomyces cerevisiae . When starving, the AEC dropped to near 0.15 but the cells remained viable. These results are very similar to S. cerevisiae. When expressing β-galactosidase the yield increased sharply during the lag phase after transition from glycerol to methanol, reached a peak during mid-log growth and then declined. A small drop in the AEC occurred just prior to the shift from β-galactosidase accumulation to loss, and there was a concomitant drop in the total size of the adenylate pool