As stated in the Orthogonality test in experiment part, we failed to find the orthogonal gene circuit of the two warriors. We can just determine part of the two warriors as follows (Fig.1). From the experimental results(See details at Orthogonality test), we know that warrior II just response to C4HSL secreted by warrior 1, but not 3OC6HSL secreted by himself.

However, we cannot design a warrior I that can be killed by warrior II but not by itself regardless of whatever receptor-promoter we put inside warrior I.what can we do to construct an orthogonal gene circuit?

    The problem now becomes how to block warrior II’s response to C4HSL secreted by himself. An intuitive answer to this question is that we could let C4HSL inside warrior I specifically stimulate the expression of an inhibitor, and then the inhibitor can inhibit the promoter of LacI.
Considering this design principle, we design the gene circuit of warrior I and warrior II as follows.(Fig.2)

    Combining with the experiment data(See details at Orthogonality test), we can find C4HSL can specifically bind to RhlR and then the complex bind to Prhl so that TetR can be expressed. Then TetR can bind to Plux/tet and inhibit the expression of LacI. By this circuit, we can see that even though C4HSL-RhlR can stimulate expression of LacI by binding to Plux/tet directly, its stimulation of TetR at the same time can counteract this effect and so warrior I won’t response to C4HSL secreted by himself.

    Like in the orthogonality test of experiment part, gene circuit of two warriors is orthogonal when the AHL molecules secreted by warrior I can just activate the promoter inside warrior II, but cannot activate the promoter inside itself and vice versa.Also, we use RFP intensity to indicate the level of LacI inside cell and thus its intensity inside each bacteria can be used as a relative measurement of warrior’s response to AHL.
    Therefore, we can know the orthogonality in real time by looking at the change of warrior I’s RFP intensity as time with or without warrior II. (Because warrior II can just respond to C4HSL according to the experiment results, so it is accord with the orthogonality requirement and we do not consider response of warrior II)

Note: This is part of fluid model to help us solve this specific problem. There are also small changes in part 6 to adapt to our experiment results. Detailed explanation can be seen at Fluid Model. Subscript 1 indicate warrior I, subscript 2 indicate warrior II. AHL1 is C4HSL and AHL2 is 3OC6HSL.

5. AHL2 (Produced by warrior II) molecule concentration inside environment (molecule/medium)

7. Amount of TetR inside warrior I (molecules/cell)

1. Results under original parameter sets:

    As stated above, RFP intensity is used to indicate the response of the warrior. So from this original parameter set and result, we can see that there is no distinguish between the response of warrior I without with warrior II, so orthogonality is not good. In the real situation, this means that the warrior I has high amount of LacI even without warrior II and thus almost no CmR. So it has no resistance to chloramphenicol and will be killed by itself.
    But how could this happen? Intuitively, we may hypothesize that TetR actually not inhibit the expression successfully when its expression is activated by C4HSL-RhlR. So we may solve this problem by increasing the maximal expression rate of TetR. In the experiment, this can be achieved by changing the RBS afterPrhlof TetR in warrior I.

2. Increasing the maximal expression rate of TetR in warrior I

    From the results of the model, we are delighted to see that warrior I just respond when there is warrior II. So the orthogonality of gene circuit is satisfied!     However, if the level of TetR is too high, we will see thatwarrior I cannot respond with or without warrior II.

    In conclusion, the core is that we can make the gene circuit orthogonal by regulating the TetR level to a moderate level. This can be easily achieved by change the RBS after Prhl of TetR in warrior I.
    According to the results obtained from our model, it is clear what we should do in the experiment to achieve the goal.

[1] Pai A, You L. Optimal tuning of bacterial sensing potential. Mol Syst Biol, 2009, 5:286
[2] Ron M, Rob P. Cell Biology by the numbers. United States: Garland Science, 2015
[3] ETH_Zurich 2014: https://2014.igem.org/Team:ETH_Zurich/expresults
[4] Ye C, Jae K. K, Andrew J. H. Emergent genetic oscillations in asynthetic microbial consortium Science, 2015, 349:6215