An artificial neural network for membrane-bound catechol-O-methyltransferase biosynthesis with Pichia pastoris methanol-induced cultures

Abstract Background Membrane proteins are important drug targets in many human diseases and gathering structural information regarding these proteins encourages the pharmaceutical industry to develop new molecules using structure-based drug design studies. Specifically, membrane-bound catechol-O-methyltransferase (MBCOMT) is an integral membrane protein that catalyzes the methylation of catechol substrates and has been linked to several diseases such as Parkinson’s disease and Schizophrenia. Thereby, improvements in the clinical outcome of the therapy to these diseases may come from structure-based drug design where reaching MBCOMT samples in milligram quantities are crucial for acquiring structural information regarding this target protein. Therefore, the main aim of this work was to optimize the temperature, dimethylsulfoxide (DMSO) concentration and the methanol flow-rate for the biosynthesis of recombinant MBCOMT by Pichia pastoris bioreactor methanol-induced cultures using artificial neural networks (ANN). Results The optimization trials intended to evaluate MBCOMT expression by P. pastoris bioreactor cultures led to the development of a first standard strategy for MBCOMT bioreactor biosynthesis with a batch growth on glycerol until the dissolved oxygen spike, 3 h of glycerol feeding and 12 h of methanol induction. The ANN modeling of the aforementioned fermentation parameters predicted a maximum MBCOMT specific activity of 384.8 nmol/h/mg of protein at 30°C, 2.9 mL/L/H methanol constant flow-rate and with the addition of 6% (v/v) DMSO with almost 90% of healthy cells at the end of the induction phase. These results allowed an improvement of MBCOMT specific activity of 6.4-fold in comparison to that from the small-scale biosynthesis in baffled shake-flasks. Conclusions The ANN model was able to describe the effects of temperature, DMSO concentration and methanol flow-rate on MBCOMT specific activity, as shown by the good fitness between predicted and observed values. This experimental procedure highlights the potential role of chemical chaperones such as DMSO in improving yields of recombinant membrane proteins with a different topology than G-coupled receptors. Finally, the proposed ANN shows that the manipulation of classic fermentation parameters coupled with the addition of specific molecules can open and reinforce new perspectives in the optimization of P. pastoris bioprocesses for membrane proteins biosynthesis.