Tom Ellis Lab
Engineering refactored and tuneable GPCR signalling
Will Shaw's paper on engineering yeast for tuneable GPCR signalling in now published in Cell. Will did this amazing body of work during his PhD in our group, co-supervised by Mark Wigglesworth at AstraZeneca. We set out to make yeast cells that could sense and respond to medically-relevant hormones, metabolites and proteins that are sensed in nature by G-protein coupled receptors (GPCRs). This would give us new biosensors and pave the way to using engineered cells as therapeutics that sense and act.
Early on Will realised that to get this technology to work we'd need to be able to reliably tune the cells to respond to their input molecules, so that we can get the right outputs for the relevant input levels we need to detect for each case. This is far from straightforward as sensing via GPCRs goes through a cascade of proteins before activating gene expression. So we set ourselves the challenge - can we work out how to predictably tune all aspects of the dose-response curve for GPCR signalling from ligand to gene expression.
To tackle this we joined forces with Graham Ladds (Cambridge) who developed a mathematical model of GPCR signalling in yeast. This showed us that by varying the levels of just a few key components we could achieve what we wanted. However, matching the model in a living cell is never straightforward. Thankfully genome engineering and synthetic biology tools in yeast make it now much more feasible. Will used these tools to turn Baker's yeast into a living model for GPCR signalling - deleting unwanted genes that complicate the signalling, and refactoring all the key genes into a cluster where we can quickly change their expression levels to desired amounts.
The result of the genome engineering was striking - Will now had robust, predictable full control over how cells relay they ligand input to expression output via GPCRs. He'd made the cell match the model and could reap the benefits. This allowed him to quickly turn yeast into tuneable biosensors for human metabolites, hormones and proteins. And by adding cell-to-cell communication into the mix, he could also alter a graded response to an input into a switch-like response. As demonstrations of the technology we used our sensor cells to detect which designs of engineered yeast produce the most melatonin, and made sensors cells that give a digital-like response to nanomolar levels of a brain pathogen biomarker.
We're very proud of this work which represents the best kind of collaboration possible and shows the fruits of labour of a remarkably talented and productive researcher during his PhD. We plan to take this work into many different territories in the next few years as we ca
n use it to research genome biology, signalling and multicellularity and to make biosensors for commercial use and as a way to improve cell-based therapies. Thanks to everyone who helped with this work and especially to those who commented on our preprint of this on bioRxiv back in summer 2018. The preprint community peer review really helped us to turn this into a great final paper. Art by Sigrid Knemeyer, SciStories LLC, scistories.com