Optimization of cell-free biosensors for synthetic biology

By Amir Pandi and Olivier Borkowski

Cell-free synthetic biology recently became a branch of synthetic biology with dedicated research groups and conferences. Cell-free systems present a great potential for synthetic biology, allowing for quick in vitro transcription-translation from circular or linear DNA. The most common cell-free systems nowadays are lysate-based cell-free systems which are made by combining cell extract plus reaction buffers. These systems were initially used for fundamental discoveries in molecular biology (study of the genetic code and translation process) and later on to produce recombinant protein. In the past few years, cell-free attracted synthetic biologists’ attention as a platform for high-throughput characterization and prototyping of natural and synthetic biological circuitry. As advantages of cell-free systems we can be list: Non-GMO hosts, absence of growth dependent challenges, lower level of noise, less susceptibility to toxicity, simple cloning as genes can be cloned separately or possibility of using linear DNA (PCR product), high adjustability by varying the concentration of DNA parts or buffer elements.

However, there are still obstacles to use cell-free systems in synthetic biology. A major challenge is inefficient repression behavior. Since many bacterial regulatory elements rely on repression (i.e. most of transcription factors building blocks used in synthetic gene circuits), cell-free synthetic biology has been further developed for metabolic engineering applications than gene circuits development.

Composition and functioning of biosensors in transcription-translation cell-free systems. (a) A Cell-free biosensor is composed of the cell-free reaction mix (cell lysate and reaction buffers) plus the DNA. (b) The addition of the chemical (inducer) produces GFP. In this case, the inducer de-represses the promoter.

In a recent study published in ACS synthetic biology, we explored different optimization strategies to improve repression in a cell-free system. We designed a simple biosensor responding to D-psicose: psiR, a transcription factor (TF) actuates the expression of gfp from ppsiA promoter.

Sampling a wide range of concentrations for both plasmids expressing TF and GFP reporter is crucial. By trying random concentrations, you will likely not be able to see any GFP production and give up on the experiment. Initially, we only measured a very weak signal with the maximum concentration of the TF and low concentration of reporter DNA. At its best, our first experiment, based on variation of the 2 plasmid concentrations, led to an inefficient cell-free biosensor (very low fold change in the signal).

Optimization strategies applied to improve the fold change of a cell-free biosensor functioning through a transcriptional repressor. (a) Doping, (b) Preincubation, and (c) reinitiation of (two-step) reaction. Adapted from Pandi et al. 2019, ACS synthetic biology.

Then, we applied three strategies to overcome the issue of our low fold change.

The first strategy is using a TF-doped extract: the lysate is prepared from cells harboring a constitutive TF-expressing vector so the lysate already contains TF proteins. The cell-free reaction starts with the TF ready to repress its cognate promoter in the absence of inducer. Adding the inducer derepresses the promoter and produces GFP.

The second strategy is using preincubation: first, the cell-free reaction is performed only with the TF plasmid to produce the TF protein (preincubation). Then the reporter plasmid and the inducer are added to the mix before the reaction runs out resources to produce protein. The biosensor efficiency depends on the preincubation time modifying the balance between the amount of expressed TF (increases over time) and the available resources for GFP production (decreases over time). After 8 hours of preincubation, the repression of the promoter is at its highest level but there are not enough resources left for GFP production. Gene expression in cell-free drastically diminishes after 8-10 hours.

The third strategy is using the reinitiation of the cell-free reaction (two-step reaction): first, we preincubated the TF for 8 hours. Then we added the reporter plasmid plus fresh cell-free mix (lysate plus buffers) to reinitiate the cell-free reaction. We saw an improvement in the biosensor efficiency when either 15 or 30 µl were added with the reporter DNA.

Eventually, we compared the unoptimized biosensor as well as two different optimized biosensors to monitor the enzymatic production of D-psicose from fructose. With the optimized biosensors, we were able to quantify D-psicose production. The same preincubation or reinitiation approaches can be used to monitor the prototyping of multi-enzyme pathways in a faster and more efficient design-build-test cycle. Our strategies can be applied to optimize cell-free biosensors and gene circuits that mostly function through repressors and so generalize the use of cell-free systems in synthetic biology. 

Short bios

Amir Pandi:

I am a PhD student in synthetic biology at Micalis institute, INRA, University of Paris-Saclay. With a bachelor's in the cell and molecular biology from the University of Tehran and a master of systems and synthetic biology from Paris-Saclay, I also participated in iGEM competition as a member (2016), as an advisor (2017), and a mentor (2019). In my PhD, I have been working on the development of biosensors and analog metabolic circuits in whole-cell and cell-free systems.

Olivier Borkowski:

I am a Research Associate at Genoscope, located in Paris area. My research focus is on the relationship between protein production and host physiology. I work both with living cells and cell-free to understand the mechanisms behind the optimization of protein production and resource competition. Currently I am using approaches coupling cell-free technology and machine learning to optimize metabolic pathways.