iCFREE

We aim to optimize and develop optimized cell-free protein synthesis systems (CFPS) and protocols for the production of therapeutics difficult to be synthesized in vivo. Three CFPS will be considered, one based on E. coli cell-extract (proCFPS), one composed of well-defined purified bacterial components (pureCFPS) enabling the integration of non-canonical amino acids (ncAAs), and one based on hybrid mammalian cell or tissue lysate (euCFPS) to express therapeutic proteins that cannot be produced in prokaryotic systems. 

Protein production in vivo is well established, but it can be costly, and some proteins remain difficult to produce due to toxicity, inclusion body formation, and protein inactivity and instability. CFPS allows expressing proteins in a few hours disregarding toxicity and grants access to large libraries of mutants. Optimization of CFPS has been tackled before varying one component after another, recently, we proposed a strategy based on machine learning to search large combinatorial spaces varying several components at one time to efficiently find an optimum for cell-free productivity. 

Here we will expand upon our strategy optimizing for the multi-objective of maximizing productivity and minimizing cost while maintaining functionality. For any of our three proposed CFPS systems, an optimization workflow based on machine learning will determine the stoichiometry of specific supplementary components be they either chemicals, enzymes, DNA, tRNAs, or riboproteome derived. Provided a gene sequence, within less than 10 days our systems will enable on-the-fly synthesis, transcription-translation of the gene, quantification - purification of the corresponding protein for further scale up or small-scale on-site and on-demand production at for instance point-of-care facilities.To achieve this goal, we will develop and deliver a toolbox containing – active and reinforcement machine learning software products to design standardized experiments (ML-DoE) and drive robotized workstations – transcription/translation prokaryotic components (cell extract with transcription and translation components) and buffer components – transcription/translation purified components including specific tRNAs loaded with ncAAs and components for PTMs – transcription components (RNA polymerase) and translation (full riboproteome isolation containing post-translational modification activities) for expression eukaryotic systems – a DNA synthesizer switching optimized prokaryotic to eukaryotic well adapted constructs – standardized and robotized protocols for individual and merged transcription and translation – standardized high-throughput online assays to quantify targeted proteins and measure their functionalities such as activities and – standardized technology transfer protocols to bioproduction facilities including protein purification.While all together the three systems should enable the production of a priori any therapeutic proteins amenable to in vitro production, they will be benchmarked for the production of antimicrobial proteins difficult-to-express in vivo. 

Our consortium is composed of four partners with complementary expertise in cell-free systems and protein production. The MICALIS institute has several years of experience and published records in optimizing with machine learning prokaryotic cell-free systems for bioproduction and biosensing. The CarMeN institute (UCBL1/INSERM) has a strong background with in vitro transcription and eukaryotic translation mechanism required for the project. I2BC (CNRS/CEA) brings know-how in the recombinant production of peptides and proteins and in the biological incorporation of ncAAs into peptides. SYNTHELIS is a French biotech company offering more than 10 years of expertise in protein production using cell-free expression technology. To date, SYNTHELIS team has successfully produced over 230 proteins using its cell-free system and most of them were difficult-to-express targets.