Team:Exeter/Kill Switches


Exeter | ERASE


Kill Switches


Synthetic Biology has a vast array of potential applications beyond the limits of a sterile laboratory; from the degradation of explosive residue left in soil to enhancing the nutritional value of crop plants. As a result, it is necessary to implement control mechanisms over these synthetic organisms to ensure that they continue to function correctly and can remain contained to their intended environment.

Kill Switches: What are they?

Kill switches are devices incorporated into synthetic organisms which offer a potential mechanism of regulating the life cycles of the organisms. They are often designed such that they operate to terminate the organism when its intended function has been fulfilled. This concept is termed induced lethality.

Kill switches will often comprise a toxic gene product under the control of an inducible promoter. For the purposes of laboratory testing this promoter will likely be responsive to a simple common molecule which can be added to growth media, for example a sugar such as arabinose. However, this would ideally be responsive to a molecule indicative of the progression or completion of the cell's intended function. This could be concentration dependant or even just detecting the presence of the molecule such as in a biosensor.

Why we wanted one

As our synthetic E. coli are designed to degrade the residual explosives found in the ground at munition waste sites they would in theory be used actively in the soil. It would not be possible to simply release synthetic organisms into the environment without implementing essential biosafety mechanisms and it was therefore necessary for our project to include a kill switch.

Our initial ideas

  • Looking into Epigenetic modifications and how these could influence gene expression over time.
  • Long term genetic stability of the population or individual parts – Biobrick stability
  • A device which could activate a lytic cycle of bacteriophage, their DNA incorporated into our bacteria
  • A method of interfering with quorum sensing e.g. the absence of TNT terminates QS. A second signal then initiates killing of the cells.
  • Using our synthetic Promoter (indirectly sensitive to TNT) to activate the kill switch device when the levels of TNT are low enough
  • RNA aptamers to TNT
  • Then having a timer so that it would continue to degrade the explosives even below the low triggering concentration
  • Timer using small cis/trans acting RNA molecules

Initial Research

It was decided that we would be investigating the retention of function of the kill switch over time. The notion that kill switches are inherently unstable was discussed as bacteria containing a device designed to kill it are at a significant competitive disadvantage compared to those without or those acquiring even a small mutation if it were enough to render the device functionally less active or completely inactive therefore allowing the cells to live.

In order to test the kill switch part over time, research was done into the E. coli long term evolution experiments in which the separate populations of E. coli are kept growing and each day 1% of the culture is transferred to the next container of minimal media. This set up would have been possible in our lab though may have required a lot of space in the shaking incubators and unfortunately there were also serious issues with contamination.

Another source of inspiration for the work came from the Jove set up of multiplex chemostat arrays. The idea that it would not be necessary to physically move cultures by hand was appealing although it proved difficult to source all of the necessary equipment.

It was decided instead to monitor the cells as they grow under different conditions in a high throughput manner. Using a 96 well microplate and plate reader it was possible to obtain a lot of data about how growth is affected. The online registry was searched and many different parts used in previous kill switch designs were found including toxin-antitoxin systems for example, holin/antiholin or MazF/E in Bacillus, there were also RNA exonucleases and mechanisms of inhibiting of DNA gyrase on the registry. After assessing multiple parts it was decided that part BBa_K628806 was to be selected for testing.

Part BBa_K628806 comprises a gene encoding an antimicrobial peptide under the control of a pBad arabinose inducible promoter. It was selected from a list of many potential kill switch parts as it had a simple mechanism of activation and results recorded on the parts registry showed successful killing over a time period of a few hours. It was decided that this part could be tested to determine whether this killing function was retained over a longer period of time.

Our Experiments

  1. Overnight growth at different concentrations of arabinose (29/07/14)
  2. 6 hour monitored growth (30/07/14)
    • 29/07/14 - Cultures were grown over-night.
    • 30/07/14 - 96 well microplate was set up. Each well held 200µl and contained overnight culture and chloramphenicol and varying volumes of LB media and arabinose stock solution to create different concentrations.
    • The cells were grown in the microplate for 6 hours and readings of optical density were taken every 30 minutes at five points within the well and an average was recorded.
  3. 10 hour monitored growth (31/07/14)
    • 30/07/14 – 3 colonies were picked and grown in a shaking incubator overnight at 37°C in LB media containing Chloramphenicol.
    • 31/07/14 – a 96 well microplate was set up. Each microplate well contained 200µl of liquid. This included 20µl of chloramphenicol and varying amounts of LB and arabinose stock solution to create different concentrations of arabinose.
    • Plain LB media and Top10 cells growing in LB and 7400µM chloramphenicol were used as controls.
    • This was grown over a period of 10 hours in a micro plate reader which maintained the temperature at 37°C and shook the plate.
    • Readings of optical density were taken every 30 minutes at five points in the well and an average of these values was recorded.

Results and Conclusions from Experiments

Figure 1: A graph comparing the growth rates of transformed and untransformed cells in media at a concentration of 7400µM arabinose

The graph above depicts the growth rates of untransformed Top10 cells and culture A, transformed with the kill switch device, BBa_K628806 and allows for comparison. As can be seen, the part does not completely inhibit growth. Even in the presence of arabinose, the molecule which should activate the kill switch, it can be seen that the cells have grown. However the shape of the growth curve produced is altered from typical sigmoidal or ‘S’ shape to something more linear with a peak in optical density after around 7 hours before appearing to begin a plateau or slight decline.

Interestingly, there seems to be no lag phase in the transformed culture and the population begins to grow straight away. This seems to suggest it was taken from an exponentially growing population. However all of the samples in this experiment were taken from overnight cultures which should have reached stationary phase. It could potentially be the case that what we are seeing is just the top of a sigmoidal curve. This would suggest that the population of transformed cells grows in a normal manner but just much more slowly than the untransformed cells.

This raises the question of stability and retention of function. If the cells are able to grow even in the presence of high concentrations of the inducer molecule it may be indicative that the part is not functioning correctly and failing to kill the cells as intended.

Future Plans

If Exeter were to design and make our own kill switch for this project this year, it would be desirable to incorporate the synthetic promoter construct capable of indirectly sensing TNT. The absence of which would activate the killing function.

For Exeter 2014 our work on kill switches is only a small side part of the project though it would be interesting to investigate in more depth in future. Other teams have in the past looked at preparing synthetic biology for the world and whole iGEM projects could be built around this idea, for for example Imperial’s development of the geneguard system in 2011.

  • Stop using antibiotics as markers – Engineered Auxotrophs
  • No reliance on kill switches
  • No genetic pollution – prevention of cell lysis
  • Modularity – plug-in parts
  • Engineering dependency of plasmids on their host strains
  • Making plasmids disadvantageous for the hosts
  • Conditional origins of replication

It would be desirable to look at the long term function of these sorts of devices and use this information to make a much more informed decision on how trustworthy these parts are. If synthetic biology is to be applied to the real world we must be able to correctly predict whether, and if so how, these organisms will change over time.


Previous: HPLC

Next: E. coli. Xenobiotic Tolerance

Exeter | ERASE