The present thesis is focused on the characterization of alternative expression systems used for recombinant protein production (RPP) in the methylotrophic yeast Pichia pastoris. Over this whole work, the integration of results from different fields -bioprocess engineering and gene regulation- is attempted in order to fill the gaps that frequently come up when designing a RPP bioprocess. In the first chapter of the thesis, the classical PAOX1-based expression system is thoroughly studied. Specifically, through a set of chemostat cultivations of two clones expressing the Candida rugosa lipase 1 (Cr1l) with different gene dosage, it was determined the interrelation of three important factors in RPP processes such as specific growth rate (µ), heterologous gene relative transcript levels (RTL) and specific product generation rate (qp). Moreover, the expression of a crucial transcription factor of the methanol utilization pathway (MIT1) was also determined in chemostat cultivations. Once the optimal operation conditions were identified in steady state conditions, fed-batch cultivations were conducted to validate the clones' behaviour observed previously in terms of both physiological state and Crl1 production kinetics. The inherent drawbacks of using the powerful PAOX1-based expression system -mainly derived from the use of methanol as carbon source, electron source and RPP inducer- has forced the Pichia community to investigate and develop alternative methanol-free expression systems. Due to that, in the second chapter of the thesis, two novel expression systems, based on the PPDF and PUPP, were similarly characterized in order to determine whether they can compete in terms of protein production with the widely used PGAP, usually considered the methanol-free reference, for the production of the Lipase B from Candida antarctica (CalB). All the three expression systems performance was firstly compared in chemostat cultivations, which enabled to shed light on the influence of µ and CALB expression on the CalB production kinetics. Additionally, the clones harboring the novel expression system were cultivated in fed-batch mode at the optimal conditions observed in chemostat in order to test their potential bioprocess scalability. Finally, in the last chapter of the thesis, the production of active whole cell biocatalyst based on the human cytochrome P450 2C9 (CYP2C9) in P. pastoris was afforded. The coexpression of the protein along with its redox partner (cytochrome P450 reductase, CPR) was achieved by means of a bidirectional promoter system. After a screening phase in which up to 8 promoters were tested for CYP2C9/CPR production, the combination that provided the best balance was selected for subsequent bioprocess optimization experiments. In this way, the influence of important bioprocess parameters - pH, µ and methanol addition- on active CYP2C9/CPR whole cell biocatalyst production was determined. Finally, the efficiency of P. pastoris whole cell biocatalyst based on CYP2C9/CPR was tested in a proof of concept reaction of interest, in which ibuprofen is hydroxylated into its oxidized derivatives.