Introduction

Our kill switch is able to be “off” for a certain time for the bacteria to perform its function and then trigger the suicide progress spontaneously at a certain time.

In the medium, the bacteria are easily controlled by adding or removing regulatory factors. However, when the bacteria perform its function in an unregulated environment, the suicide progress needs to be activated spontaneously. Moreover, the kill switch is supposed to be “off” for a certain time, so the bacteria will gain enough time to perform its function.

For these reasons, toxin protein MazF is the best candidate to kill the bacteria for us.

MazEF is a toxin-antitoxin module composed of mazE and mazF locating on the chromosome of E. Coli and other pathogens (Hanna et al, 2005). The expression product of mazF is a stable toxin, while that of mazE is a labile antitoxin of MazF (Hazan et al, 2004; Schusteret al., 2013). MazF is a sequence-specific mRNA endoribonuclease that initiates a programmed cell death pathway in response to various stresses. The mazEF-mediated death pathway can act as a defense mechanism that prevents the spread of bacterial phage infection, allowing bacterial populations to behave like multicellular organisms.

Sequence and Features

Fig 1. Stable State

Fig 2. Killing State

Fig 3. Node Figure

Fig 4. Truth Table

We plan to establish a new E.Coli strain without lacI and MazEF system gene to avoid the inference of background expression. We will test cI background expression level to see if we need to knock out cI.

As is shown in Fig.1, Fig.2, Fig.3 and Fig.4, in the “Stable State”, before poured into the soil, the E. Coli fertilizer is cultured in the medium with IPTG. Promoter 1 is a weak constitutive promoter, so lacI is transcribed and translated at a considerate low level. Therefore, high concentration of IPTG, which binds LacI and changes LacI’s conformation, is able to inactive LacI and open the promoter 2. Then, CI can be highly expressed to repress promoter3. Finally, mazF is inhibited.

After fertilizing, no more IPTG exists in the soil to bind LacI. As a result, LacI with a tag which stabilize the protein takes time to accumulate to a certain high concentration, which is the repression threshold of promoter 2, aiming to represses promoter 2 as well as the expression of cI. The time needed to finish this progress is the designed “memory time”. During this process, E. coli takes its own responsibility to deliver Mo. And after that, it can be killed to reduce the pollution to environment.

As is shown in table 1, all the biobricks in the design are from the top 10 most used parts of iGEM to guarantee the feasibility.

Table 1 Biobricks used in the design

In the future, we plan to model the “memory time” and do more wet-experiments to confirm its feasibility. As for the modeling, we will test the expression level of promoter 1, the minimal stand of LacI to repress promoter 2 and the degradation speed of LacI. On the other hand, we will test the background expression of CI after removing IPTG to check if the CI concentration is low enough to open the promoter 3. Then we will strive to make this system more effective.

For more information, please go to our wiki:

http://2014.igem.org/Team:BNU-China/home.html

Modeling

The iGEM Team Marburg 2016 worked on a evolutionary stability analysis to determine a kill switch's network topology guideline. For that, BNU China's kill switch of 2014 has been modeled and subsequently been evaluated with regard to its evolutionary stability analysis.

This kill switch has been chosen to be modeled due to its simplicity. It is a cascade of negative regulations. Even though the network topology by BNU China 2014 can be found above, we studied the biological processes in more detailed and extended their provided network topology with these relevant processes which, for instance, involve dimerization processes. These again lead to higher order reactions that introduce complexity (see Fig. 1).

Figure 1: Shows the (a) topology by BNU China 2014 in comparison to the (b) biologically relevant topology introduced by iGEM Marburg 2016's team.

Dynamics

First, the kill switch has been modeled as a genetic regulatory network by means of a continuous ODE model. The details of that modeling can be found on our wiki: The detailed ODE system and the used biologically relevant parameters.

The 10 dimensional concentration vector has been computed over the time (see Fig. 2). After the inducer IPTG is reduced due to an escape into wild life, the toxin concentration MazF rises to an lethal value. A toxin threshold \(\theta\) has been introduced.

Figure 2: The all concentrations against time.

Evolutionary Stability

Based on the BNU China 2014 kill switch, we designed different comparable topologies: Parallel (OR), serial and parallel (AND). These can be found in figures 3 to 5.

Figure 3: Parallel (OR) topology.

Figure 4: Serial topology.

Figure 5: Parallel (AND) topology.

As one would expect, the parallel (OR) topology is the most stable: Mutations might destroy one of the branches but the second one still leads to cell death. The serial or parallel (AND) topologies are prone to destruction through mutations since there is only the regulation is not redundant. This result has been quantified and can be found in figure 6.

Figure 6: The escape probabilities of organism implementing the three presented kill switch topologies (figure 3 to 5). The parallel (OR) - and thus, the one with redundant regulation - has the smallest escape probabilities. It is the most stable one.

Conclusion

The fact that a genetically modified organism dies once it unintentionally escapes into wild life is crucial. We suggest a kill switch topology that utilizes a redundant regulation. Such a design shows resistivity against destruction through mutations.

As an example, the BNU China 2014's kill switch should be redesigned from the given long serial cascade to a redundant regulation of MazF.

Kill Switch Database

We performed an exhaustive literature research: We collected all kill switches ever implemented in iGEM and drew according topologies. This database can be found on our wiki.

The result is that many teams' kill switches lack a proper topology design. Our result of a parallel regulation should be applied.

Furthermore, this database can be used to study different kill switch designs in order to find the most appropriate one for an own design.

Find the kill switch database here: click.

Part Improvement

Team 2019 UFRGS_Brazilmade an improved version of this part (BBa_K3215016), in which the cAMP interference was removed, together with one lac Repressor binding site (operator region 3), enhancing the overall performance of this part.