Towards open-source optimisation for bespoke cell-free protein synthesis systems

Abstract

Cell-free protein synthesis (CFPS) enables transcription and translation outside the con- fines of a cell. CFPS systems are based on either crude cell lysates, which contain most of the soluble cellular machinery, or the PURE system, a minimal system reconstituted from purified components. CFPS has been essential in molecular biology discovery, biosensing, and bioproduction and is emerging as a useful tool in acquisition of large datasets via high- throughput expression e.g. de novo protein design prototyping. It has the potential to play a significant role in future biotechnology including in industrial biologics manufacturing and as the transcription translation (TXTL) chassis on which to base bottom-up synthetic cells capable of self-replication. However CFPS reactions exhibit lower titers than their in vivo equivalents and still use expensive substrates limiting their current utility meaning for CFPS to fully realise its potential, further system engineering is required to improve protein yields and enable the use of cost-effective substrates. Design of Experiments (DoE) is a powerful statistical method for efficiently designing and analysing candidate screens and optimisation experiments, where it can produce insights with relatively few experimental trials. However, its accessibility is hindered by the need for specialised knowledge and the cost of proprietary software.

This thesis presents work addressing these challenges. DoEasy, a free, open-source DoE platform was developed and its efficacy proven via the optimisation of deterministic model. E. coli Lysate CFPS was established and the platform was applied to modulating key buffer concentrations of lysate reactions expressing different GFP-tagged proteins relevant to medicine and biotechnology. The PURE system was implemented and advances towards fine-grained optimisation of its components were made through the development of a multi-pot PURE configuration. The substrates of a de novo synthetic metabolism system were optimised resulting in a 60.2% improvement in yield. The platform was demonstrated in multiple CFPS optimisations and democratises advanced statistical tools, making them accessible to a broader scientific community and enabling CFPS to reach its full potential.