Safety of Individuals

Before we were allowed into the laboratory, we were given a safety induction by the safety officer of the department. This included the hazards and risks associated with different chemicals and pieces of equipment, as well as how to minimise risk. Having completed this safety assessment, we were given a laboratory induction, in which our supervisor demonstrated each of the protocols, highlighting any points where there were safety risks associated, and, again, how to minimise these. Throughout the project, we have been following the safety protocols outlined by our supervisor. These range from the basics of wearing lab coats, goggles and gloves (which we change regularly), to dealing with chemicals such as ethidium bromide, to using equipment such as centrifuges and UV transilluminators. Details of the hazards and risks, and safety precautions we took to avoid these, can be found in the lab notebook. Thorough details of all the chemicals we've used, including all the safety aspects can be found here.

Project Safety


We have a variety of different options when it comes to how our bacteria will be used. Each of these has their own set of safety considerations. For example, for medical applications, we need to consider strategies that are suitable for the welfare of the patient; for instance, if our bacteria are inside the patient and need to be killed, we need to trigger a method of self-destruction which will not release molecules which are harmful to the patient. Similarly, for industrial applications, we need to make sure that bacteria aren’t going to be spread to a very wide area (e.g. through the water supply), so we need a mechanism of killing them as soon as they become a threat. For each situation, we have a strategy for physical and molecular containment to ensure that our bacteria will never become a health or environmental risk. For each option, there is a list of good points and problems. Within the problems list, the identations go as follows: problem, then solution to that problem indented, then problems associated with that solution indented within that.

Each of our kill switches is based on the lac operon. AI2 is a quorum sensing molecule for Escherichia coli. We would use a promoter which senses AI2 (e.g. LsrK), which would then stimulate the production of LacI. LacI is the repressor for the lac operon. We would then have an apoptosis gene (CASP8) downstream of the lac promoter and operator. When there are pathogenic E. coli, AI2 will be sensed, so the repressor will be expressed, and will bind to the operator upstream of the apoptosis gene, preventing the expression of CASP8, and hence apoptosis will not occur. However, when there is no AI2 (e.g. all of the pathogenic bacteria have been killed, and there is no need for our bacteria to be around any longer), LacI will not be expressed. This means that it will not be bound to the operator, so CASP8 can be expressed, triggering apoptosis. Additionally, if we want to kill the engineered bacteria before all the pathogenic bacteria have been removed (for example, if there was a problem with our bacteria which we had been unable to test), we can trigger apoptosis via the provision of IPTG. IPTG binds to LacI, causing a conformational change which prevents it binding to the operator. Hence, even if AI2 is being sensed, CASP8 will still be expressed, triggering apoptosis. IPTG is not harmful to humans, so this will have no adverse effects, and can be easily introduced through the catheter. These kill switches can be modified by adding extra layers of sensing to the promoters, further narrowing the range of conditions that our bacteria will live in.

Sealed within a section of catheter

Our first option is to keep the bacteria within a sealed section of the catheter. We would use semipermeable membranes to create a section within the catheter, with additional membranes further up the catheter in case of perforation which would allow bacteria to escape. The quorum sensing molecules would be able to diffuse through the membranes, and stimulate a response in the trapped bacteria, and the secretion products would be able to diffuse out to interact with the pathogenic bacteria outside.

For this situation, we would use a kill switch that is also sensitive to valine. Valine would be provided in the medium which would be present in the catheter compartment. This would mean that if the bacteria were to escape the compartment, the valine concentration would be insufficient to allow expression of LacI, so CASP8 would be expressed, resulting in apoptosis.

Good points:

  • Bacteria carefully contained, no chance of affecting patient directly
  • Constant supply of constitutively produced anti-biofilm agents will allow consistent prevention of biofilm formation as well as treatment of existing biofilms when catheter is fitted.
  • Any living bacteria still present in the compartment when catheter is removed will be destroyed through autoclaving.


  • The distance between the engineered bacteria and the pathogenic bacteria may to be too great to stimulate a response in the engineered bacteria/attack the pathogenic bacteria, because the quorum sensing molecules would effectively be too diffuse, giving a false reading of the population density of the pathogenic population. This could also result in the kill switch being triggered, because if there is not enough AI2 to keep LacI being expressed, apoptosis will be activated.
  • We’d need to find a way to sterilise the catheter either with the engineered bacteria inside (this is usually done using heat, which would denature normal E. coli) or adding the bacteria later, which would likely involve puncturing the membranes, which would compromise their integrity.
    • A possible way round this is to puncture the wall of the catheter itself to insert the bacteria, then melting the edges together to seal it in. This would have to be performed under sterile conditions to prevent the sterility of the rest of the catheter being compromised, which would be tricky, as the sterilisation of the catheter occurs within a sterile bag, which you’d have to puncture and reseal as well.
    • Alternatively, catheters have a balloon which is filled with water to hold them in place. This is filled by pumping water between an inner and outer tube from an external junction. This suggests we may be able to add a second tube (so you have an outer layer for pumping water, a middle layer for pumping bacteria, and a tube through the centre for urine drainage), so that we can fill a section of our catheter with bacteria after it has been sterilised.
  • We can’t increase the diameter of the catheter too much, as this would make the catheter fit worse, but we also can’t decrease the inner diameter too much either, because this would reduce urine flow through the catheter, and increase the risk of blockage, which would create more problems for the patient. Therefore, it could be tricky to design a suitably sized compartment for the bacteria, which will still contain enough bacteria to be effective.

Bacteria enclosed in beads, sealed within a section of catheter

To add an additional layer of protection into this system, we could encase the bacteria in sodium alginate beads, which would then reside within the compartment of the catheter. This has the advantage of preventing the bacteria colonising the urinary tract even if the membranes rupture, but there is the disadvantage that the treatment would be less effective due to diffusion speeds. Additionally, the quorum sensing molecules might not get into the bead in sufficient quantity to prevent the kill switch being activated. For more information on the AlgiBeads, click here.

Attached to the end of catheter

Bacteria can be fixed to surfaces using their pili or flagella. We could use this option to attach the bacteria to the outside of the catheter, reducing the diffusion distance of the quorum sensing and increasing the efficiency of this system.

Since the bacteria wouldn’t be in media for this option, they would need a different kill switch. This would only be sensitive to AI2, but we would use IPTG to induce apoptosis if problems arose. The IPTG could be easily supplied through the catheter.

Good points:

  • The diffusion distance for quorum sensing and secreted molecules is much smaller, increasing the efficiency and efficacy of the system.
  • Again, any bacteria still alive when catheter is removed can be destroyed through autoclaving.


  • This method has the same problem with sterility as the first option of containment.
  • There is increased risk of our bacteria escaping, as the pili or flagella could break, releasing our bacteria into the environment.
    • Using multiple pili will decrease the probability of the bacteria being released, but presents more problems.
      • You would need to use pili which are close together so that the bacteria isn’t under a large amount of torsional stress which would cause it to function in an unusual and unpredictable manner. This is tricky, since you would need to select several for individual bacteria, which would be technically difficult and very time consuming.
      • Since one of the secretion systems we are using for our artilysins and other attacking molecules is through these structures, it would reduce the efficiency of secretion, since some structures would be taken up through attachment.
      • The inclusion of flagella in our bacteria would increase the likelihood of pathogenicity, since pathogenic E. coli use their flagella to secrete toxins into the environment, and lab strains have this quality removed for safety.
    • You could coat the catheter in the sodium alginate/media which we used to make the AlgiBeads. This would mean that the pili are less likely to break and provides an additional protective layer between the bacteria and the patient.
      • This would decrease efficiency slightly, but not as much as having the bacteria within the catheter.
      • In this instance, we would use the “media” kill switch, which is sensitive to valine, so that if the bacteria escape from the protective coating, they would die.
  • We can’t increase the diameter of the catheter too much, as this would make the catheter fit worse, but we also can’t decrease the inner diameter too much either, because this would reduce urine flow through the catheter, and increase the risk of blockage, which would create more problems for the patient. Therefore, it could be tricky to design a suitably sized compartment for the bacteria, which will still contain enough bacteria to be effective.

Free-living Bacteria

Our surveys have suggested that the general public are not against altering their microbiome, and having our engineered bacteria living freely inside their urinary tracts. In this situation, they would be either injected into the urinary tract or pumped in through the catheter, with no physical containment.

Engineered bacteria are introduced to the urinary tract so that they can colonise and outcompete the pathogenic bacteria. If the population of pathogenic bacteria is above the threshold value that the engineered bacteria can recognise, dispersins and artilysins will be produced to break up the pathogenic biofilms and kill the pathogenic bacteria. Some bacteria will also sacrifice themselves to take out more pathogenic bacteria in one go, using holins to break the engineered bacterial membranes. This helps to keep the population of engineered bacteria down. For these “self-sacrificing” cells, the holin system acts as its own kill switch, since when AI2 is sensed, the cell will be destroyed.

Since there would be no physical containment, we would use the free-living kill switch detailed above. Once again, IPTG would play an important role in artificially maintaining bacterial populations at a safe level, or ridding the urinary tract of our engineered bacteria entirely.

Good points:

  • In addition to the constitutively produced anti-biofilm agents, using this set-up, we’ll be able to implicate the lysis system, targeting a substantial burst of agents in one go, resulting in a more comprehensive solution, which would theoretically work faster than just the constitutively produced agents.
  • This prevents the probability of reinfection, because our engineered bacteria would colonise and out-compete the pathogenic bacteria.
  • These bacteria would also help deal with biofilms even after the catheter has been removed.
  • The engineered bacteria have no harmful qualities (except for those few which have the additional holin release system), so even if the population density did get higher than we would ideally specify, there shouldn’t be any adverse effects, as they would still be outcompeting the pathogenic bacteria.


  • One concern of the “sacrificing cells” approach is that, unlike during apoptosis, everything will be released from the dying cell. This includes enzymes such as lysozymes and peptidases which, when released into the cell’s environment, could attack not only the pathogenic cells being targeted by the artilysins as well, but also the patient’s body cells, resulting in tissue damage. In short, that system may create more problems for the patient than it solves, but it is impossible to know the effects of such a treatment without testing within a system comparable to a human body.

Cartridge in an Industrial Pipe

In this setting, we would house the bacteria within a cartridge, which will be stuck to the inside of the pipe. The cartridge will be separated into at least 2 compartments, one containing media and bacteria, and the other which allows water to flow past the bacterial compartment. The two compartments would be separated by a semipermeable membrane, which would have pores big enough to let proteins and signalling molecules through, but small enough to prevent the bacteria escaping.

We would include a modified version of the kill switch, which has a promoter sensitive to AI2, valine and light. This means that only when valine is supplied in high enough quantities (e.g. when the bacteria are within the media containing compartment), when there is no light (e.g. when within the pipe), and when other bacteria are being sensed, LacI will be expressed, meaning that it will bind to the lac operator, preventing the expression of CASP8, allowing the cell to live. If any one of these criteria do not hold, then the bacteria will be killed, since LacI will not be expressed.

Good points:

  • Multiple levels of containment to prevent bacteria getting into the water pipe, and still more molecular criteria to prevent bacteria living in an undesired environment.
  • We can design the cartridge to allow maximum expression of biofilm degrading agents without having to alter the bacterial system.


  • If the bacteria get out of the compartment through perforation of the membrane, the public may not want bacteria, even if they are dead, being transported through the water system.
    • This would need to be combated by having extra levels of filtration during water purification, but it is relatively simple to centrifuge out bacteria from water, as they pellet easily.