Author(s): Ferreira, Ana Raquel Santos
Date: 2012
Persistent ID: http://hdl.handle.net/10362/9369
Origin: Repositório Institucional da UNL
Subject(s): Pichia pastoris; Single-chain Fragment variable antibody; Cell functional enviromics; Chemically defined medium optimization; Process scaleup; Adaptive process control
Dissertação para obtenção do Grau de Doutor em Engenharia Química e Bioquímica
Culture media (CM) formulations contain hundreds of ingredients in aqueous solutions that may be involved in complex interactions in the same or competing pathways within the cell. This thesis proposes a new methodology for determining the optimal composition of CM that migrates from an empirical to a mechanistic or hybrid mechanistic CM development approach. A framework consisting in the execution of an array of cell cultures, endpoint exometabolomic assays and bioinformatics algorithm were brought together into a platform for CM engineering called Cell Functional Enviromics. This technology consists of a largescale reverse engineering approach that reconstructs cellular function on the basis of measured dynamic exometabolome data. To support this concept, a computational algorithm, called “envirome-guided Projection to Latent Pathways”, was developed. This method yields envirome-wide Functional Enviromics Maps (FEM), with rows representing medium factors, columns representing elementary (orthogonal) cellular functions and color intensity values, the strength of up-/down- regulation of cellular functions by medium factors. This method was applied to optimize Pichia Trace Metal salts for the yeast Pichia pastoris to improve the expression of heterologous proteins. An array of shake flasks experiments of the P. pastoris X33 strain were performed and used to build a FEM. Then, optimized CM formulations were calculated targeting predefined single-chain Fragment variable antibody (scFv) production improvements. Experimental validation shows a scFv productivity increase of approximately twofold, in relation to the control BSM recipe proposed by Invitrogen. These results were further validated in 2 L bioreactor experiments. Thereafter, scale-up to 50 L bioreactors was developed a mathematical model for further optimization of BSM salts in experiments of P. pastoris GS115. Direct adaptive (DO)-stat feeding controller that maximizes glycerol feeding through the regulation of DO concentration at 5% of saturation was developed and applied to the 50 L bioreactor, with the fully optimized CM composition.
Fundação para a Ciência e Tecnologia - bolsa de doutoramento SFRH/BD/36285/2007
