(820g) Understanding Single-Chain Antibody Fragment Production in Pichia Pastoris

Pichia pastoris is a commonly used expression host for heterologous protein production; considerable yields can be achieved from high cell densities in inexpensive culture media. Despite this, the specific productivity is relatively low. Consequently, the full impact of this host on industrial biotechnology has not yet been realised.

The majority of bioengineering studies to increase specific productivity have targeted one factor in isolation, and despite using comparable strategies have variable outcomes. Here, an integrated experimental and computational modelling approach has been taken to understand how the factors interact and develop a global optimisation strategy. We have selected single-chain antibody fragments (scFvs) as a model heterologous protein for this work as they are industrially relevant.

Initially, a deterministic single cell model of protein production in P. pastoris was devised incorporating the essential aspects of transcription, translation and folding in the endoplasmic reticulum (ER). Additionally, as heterologous protein production can induce stress responses, the unfolded protein response (UPR) and ER associated degradation pathway (ERAD) have been accounted for.

Preliminary simulations highlighted key regulators of capacity; the experimental picture, however, is much more complex. P. pastoris exhibits clonal variation – whereby a culture, derived from a single cell, shows a range of phenotypic behaviour despite being genetically consistent. We aimed to understand how differences in yield arise from genetically identical strains by systematically characterising those key regulators in strains which express a scFv to different degrees.

Consequently, two single chain antibody fragments were cloned into P. pastoris and a high and a low secretor of each identified. Time courses of RT Q-PCR carried out targeting scFV transcript levels, the UPR and the ERAD response. LC-MS/MS was used for absolute quantification of the key regulators Kar2 and PDI, both under normal and conditions of over-expression. The latter provided us with a means to validate the model, and derive strains with a higher level of productivity.

The experimental data has been used to expand the preliminary model, revealing an interesting competition between the degradation and folding pathways. Indeed, high yield can be derived from a number of different pathways across the cellular landscape. Despite this, the knowledge of the upper and lower bounds for parameters from the low and high secreting strains has allowed for their optimal values to be predicted, deriving an in silico design strategy. This constitutes an integrated modelling and experimental approach to improving productivity in P. pastoris.