Team:EPF Lausanne/Envelope stress responsive bacteria


Envelope Stress Responsive Bacteria

The EPF Lausanne iGEM team has been working on engineering E. coli responding to stress - in particular mechanical stress – quickly and efficiently. Taking in consideration that bacteria naturally respond to various envelope stresses through the Cpx pathway, we combined protein complementation techniques with biosensors to achieve fast spatiotemporal analysis of bacteria response to stimuli.

To help you understand how our engineered bacteria work, we would like to start out by introducing two major components of our project: the Cpx pathway and the split reporter proteins.

The Cpx pathway

How can bacteria sense and adapt themselves to environmental stress? The key lies in a fairly unknown two component regulatory system: the Cpx pathway. Its natural function is to control the expression of "survival" genes whose products act in the periplasm to maintain membrane integrity. The three main members of the Cpx pathway are CpxA, CpxR, and CpxP.


Split reporter proteins: Infrared Fluorescent Protein

Protein complementation techniques consist in splitting reporter proteins with fluorescent, bioluminescent, or colorimetric properties at specific locations and fusing them to proteins of interest. The fused split fragments remain inactive until physical interaction of the proteins of interest leads to the reconstitution of the chemical properties and the proper folding of the split reporter proteins. This technique is widely used to characterise the interactions between proteins of interest in a given pathway.

The EPFL iGEM team however chose to implement protein complementation techniques in a new way. Our team hijacked the Cpx pathway in E. coli by fusing split protein fragments to CpxR to develop fast and precise stress biosensors. On top of being used in basic science as a stress sensor and a proof of concept of our novel way of viewing biosensors, we aimed to integrate our engineered bacteria in the BioPad to allow fast and precise signal emission upon touch.

The split protein we chose to fuse to CpxR was the Infrared Fluorescent Protein (IFP 1.4). The IFP1.4 is an engineered chromophore-binding domain of a bacteriophytochrome from Deinococcus radiodurans. Its split was an important landmark in the field of protein complementation techniques, as the split IFP1.4 is the first fluorescent split protein to be reversible. Moreover, due to its emission in the infrared wavelengths, the IFP1.4 benefits of high signal-to-noise ratios allowing precise analysis of spatial dynamics. The split IFP was developped in 2014 by the Michnick lab 1.

Experiment planification

Having chosen a protein complementation candidate - IFP1.4 - and a stimuli responsive dimerizing protein (CpxR), we planned out a set of experiments that would allow us to prove that fast and specific spatiotemporal analysis of stimuli is possible.
We therefore set ourselves three intermediate objectives:

  • Prove that CpxR dimerizes and find out its dimerization orientation.
  • Prove that the system can be activated and turned off quickly.
  • Prove that the system can provide spatial information and that sub-localisation of the protein within bacteria is detectable.
  • Estimate the time elapsed between a transcription factor activation and its associated reporter protein synthesis

Moreover, to assert that our system could be used to build a BioPad, we also had two extra objectives:

  • Show that our system can be activated by envelope deformation - particularily by mechanically stress in a microfluidic chip.
  • Prove that after envelope deformation the system can be shut off.

We planned out various experiments to accomplish these objectives.

Demonstration of CpxR's dimerization & elucidation of its dimerization orientation

To prove that CpxR dimerizes and to find out its dimerization orientation, we synthesized four constructs with combinations of the split IFP1.4 fragments fused to the C or N terminal of CpxR. We amplified CpxR from the E.coli K-12 MG1655 genome. The IFP1.4 fragments - IFP1 and IFP2 - were obtained from vectors sent to us from the Michnick lab. After many PCR's and Gibson assemblies, we finally obtained our four constructs:

The four resulting strains were stressed with KCl (a known stressor for the Cpx pathway) and analysed on a plate reader to understand how CpxR dimerized. The experiment planned an initial period of 24 minutes without stress, addition of KCl at the 24th minute, followed by 2 hours of reading. To get more information about this experiment and its results check out this page.

Activation and deactivation of the signal

After having shown that we were able to activate the Cpx pathway, we set forth to accomplish our second objective: prove that the system can be activated and shut down quickly.
The experiment plan was to read IFP emission for 20 minutes without stress, and then stress our functional strain (BBa_K1486056) with various concentrations of KCl. After the two hours of stress, we planned to centrifuge our cells for 10 min and put them into non-stressful media to continue signal reading for two more hours.
Check out the results for this experiment! The signal drops in an impressive way.

Localisation of CpxR - split IFP1.4 within E.coli

Having characterised the temporal dynamics of the CpxR dimerization, we proceeded to spatial analysis of our construct. In this experiment we aimed to address our third objective: prove that our system can provide spatial information on the sub-localisation of the protein within bacteria. Samples of stressed cells (80 mM of KCl) were put on a glass slide and observed under a microscope with appropriate filter. On top of giving us precise information on the localization of our proteins, this experiment allowed us to see a difference between stressed and non-stressed cells in terms of signal activation. Check out the results for this experiment here!

Image of IFP signal in stressed cells with 80mL of KCl

Activation of CpxR on microfluidic chip by chamber crushing

To prove that our strain of engineered bacteria is suitable for the development of BioPad we aimed to verify that we were able to activate the system by mechanical stress on a microfluidic chip. Thanks to our SmashColi chip, we applied stress by activating buttons at a pressure of 25 psi. Check out the results for this experiment here!

non stressed

Investigation of time delay between CpxR activation and RFP expression under CpxR responsive promoter

We aim to estimate the time elapsed between the activation of a given transcription factor and the synthesis of an associated reporter protein by comparing these two signals. We want to explore this idea with our engineered CpxR transcription factor (leading to IFP signal upon activation) and its regulatory sequence promoting the transcription of RFP.

We obtained a RFP sequence under CpxR responsive promoter from the Calgary 2010 team (BBa_K339007) and noticed that the regulatory sequence was missing. After obtaining their theoretical sequence by cloning strategies (BBa_K1486048), the promoter was still both non-responsive to stress and leaky, as RFP production was not correlated to the activation of the transcription factor (same signal in both stressed and non-stressed cells).
We used an other strategy and replaced the regulatory sequence by the genomic region between CpxR and CpxP (which is thought to contain the CpxR responsive promoter) in both orientations (BBa_K1486049 & BBa_K1486050). This time the RFP signal was very low, most probably due to the fact that the sequence we inserted was not fully specific to CpxR thus being unoptimal for such an experiment. An experiment containing only the promoter before the CpxP coding region should be tried. The issue is still of great interest and should be studied.

Split Luciferase & Stress Responsive bacteria

The initial plan for the development of stress responsive bacteria was to evaluate signal dynamics of CpxR with both split IFP1.4 and split luciferases (firefly and renilla).

The reasons for which we initially wanted to make use of split luciferases in our project were the following:

  • Fast and reversible signal (quasi-instant)
  • Light excitation not needed (minimal autofluorescence) => simplification of the signal detection and processing mechanisms
  • Signal highly correlated with concentration of substrate => very modulable
  • Small concentrations of substrate are needed

To explore how split luciferases worked, we aimed to reproduce past experiments making use of this technology. We thus set ourselves as objective to study the signal dynamics of CheY/CheZ in E.coli 2 before fusing split luciferases with CpxR.

CheY and CheZ are two proteins involved in the bacterial chemotaxis pathway. It has been shown by split luciferase complementation assay that these two proteins do not interact in presence of chemoattractant, but do interact (CheZ being the phosphatase of CheY) in absence of chemoattractant or presence of chemorepellent. Based on the work of Waldor Laboratory, we wanted to reproduce and adapt the experiment to test our own splits.


We created two different constructs, one with split Firefly luciferase and one with split Renilla luciferase, to test different possibilities with different substrates (Firefly luciferase oxidizes d-luciferin and Renilla luciferase oxidizes coelenterazine). Check out our results page to see how our experiments went!


  1. An infrared reporter to detect spatiotemporal dynamics of protein-protein interactions ; Emmanuelle Tchekanda, Durga Sivanesan & Stephen W Michnick (Nature, 2014)
  2. Studies of Dynamic Protein-Protein Interactions in Bacteria Using Renilla Luciferase Complementation Are Undermined by Nonspecific Enzyme Inhibition; Stavroula K. Hatzios, Simon Ringgaard, Brigid M. Davis, Matthew K. Waldor (PLOS 2012)