Since our intention was to individually and independently monitor the fluctuating levels of each metabolite, along with an easily readable readout for overall risk that accounted for all three, we had to design a system that allowed for a certain level of cross-talk. Thus, taking into account all that has been rationalized and justified, we designed a biological system composed of two cells working in tandem with each other.
This two-cell logic gate design was heavily influenced by a paper titled “A modular cell-based biosensor using engineered genetic logic circuits to detect and integrate multiple environmental signals" by Wang et al [1]. In this study they developed a bacterial biosensor that could detect and integrate three environmental signals - arsenic, mercury, and copper ions - to produce a visible signal using a two-cell system [1]. We felt that this design was useful because the complex nature of the MDD-metabolite relationships made it critical that we had a system that could consider multiple metabolite inputs to improve the validity of our biosensor. This paper significantly influenced our choice of parts for our proposed design, including our use of the plant pathogen P. syringae-derived Hrp (hypersensitive response and pathogenicity) regulatory system [1]. This system encodes two co-regulatory proteins (HrpR and HrpS) that form a heteromeric protein complex to activate the HrpL promoter which in turn induces LuxI expression. LuxI is synthase for the quorum signalling molecule, 3OC6HSL, which diffuses out of the first cell into the second cell where it binds and activates LuxR, the activator protein for the PluxI promoter. This two-cell consortium system was the basis for the AND logic gate; we used a modified version of this setup in our own design. We additionally included the transcription of HrpV into our system, an inhibitor that binds to HrpS to prevent the formation of the activation complex with HrpR [2]. This overall design allows for more flexibility in the design of a genetic logic gate and was crucial to enabling a tri-input detection system which could not be expressed in a single E. coli cell.
Each metabolite is linked to a unique fluorescent protein for individual read-outs, however, in order for RFP to be produced, there must be high [indole], low [GABA], and low [butyrate], therefore, indicating a greater risk for MDD. Specifically, in the presence of high [indole] and low [GABA], a greater amount of HrpS/HrpR will be transcribed than HrpV, allowing for P_HrpL to be successfully activated and thereby LuxI to be transcribed. If there is low [butyrate] detected in the second cell, then via LuxI, 3OC6HSL, and LuxR, additional HrpS/HrpR will be freely formed without inhibition from HrpV, leading to the final activation of RFP (or any other fluorescent marker that is chosen).