Optimising recombinant protein secretion in wild strains of Saccharomyces cerevisiae

Yeasts such as Saccharomyces cerevisiae and Pichia pastoris, are challenging the dominance

of Escherichia coli as the preferred unicellular system for the manufacture of

biopharmaceuticals, primarily because of their ability to produce and secrete complex

proteins containing disulphide bonds. The most significant bottleneck in the secretion of

recombinant proteins from yeast cells is the relatively low abundance of key processing

enzymes (such as Sil1 or Lhs1), and chaperones in the secretion pathway yet nevertheless

laboratory strains of yeasts have been successfully engineered to secrete a wide range of

complex recombinant proteins. In this project I am asking whether a group of unmodified

wild strains are efficient at secreting large amounts of recombinant protein. To do this I am

examining the efficacy of protein secretion in a range of novel wild strains some of which

have been discovered in remote uninhabited regions around the globe. To achieve this aim, I

have first characterised the ability of various wild strains to secrete the yeast dsRNA virusencoded

'killer toxin' (KT). A 'halo' assay has been used to quantify the amount of KT

secreted by each strain. To this end technology to insert the KT gene into strains lacking

auxotrophic makers was developed. This analysis has also included characterising the

secretome of the strains. 8 strains were identified as possible hosts for recombinant protein

production, with 3 of these strains showing a high level of secretion. A set of industrial used

'domesticated 'strains were also used as a control. These strains had KT secretion tested as

well as gaussia luciferase secretion. There was a marked difference between the strains and

between the two different proteins used, in terms of secretion levels