A large chunk of effort in synthetic biology has been aimed at attempting to view genes as parts of a circuit. Thus, a lot of focus has been directed toward creating biological analogues of digital logic gates, such as an AND or a NOT gate, which give a digital 1 or 0 response, depending on the truth table of the gate.
Figure – A typical example of a biological NOT Gate, considered to be ON when the concentration of inducer is below 10-6, and OFF when it is greater than 10-4. Such a response gives a corresponding gene expression level (au) of ~1 in the ON state, and ~0.4 in the OFF state (Source: “Generation of Pulse of a Bacterial Species in E.coli”, Kshitij Rai, Department of Biochemical Engineering and Biotechnology IIT Delhi, Master’s Thesis, 2017)
However, there is a serious issue in the scale up of these circuits. While that can be attributed to several reasons, one of the major reasons is this simplification under which the systems work. We can see that the output response from the gate is not close to the actual digital “1 or 0” kind that one would ideally want from a logic gate, and in the range that is neither in the ON nor the OFF regime, the response is really graded.
Thus, when a combination of gates is used in conjunction, one could not expect them to remain digital. Just imagine using a 10 input AND gate as shown below, that could possibly be used as an environmental biosensor. If the concentration of a few inputs is in the ON range, while the others are in the middle range, should the device show an ON or an OFF state? It may show neither, as we have seen above, leaving the researcher confused as to how the results must be perceived.
Figure: input AND gate which can act as a biosensor, sensing a particular environmental condition, and generating an output such as fluorescence through GFP.
High Cooperativity Repressors – A Possible Solution
The solution to the problem lies in the cooperativity of the parts that are used as actuators in these digital devices. These are mostly created using repressors or activators, which can repress or activate their respective promoters. Cooperativity is basically the phenomenon where the repressor (or activator) molecules do not act independently of each other, and two or more molecules of the same are needed to first bind to each other, before binding to the promoter that they are to repress.
What this means is that if the cooperativity of the repressor is “n”, then unless n molecules of the repressor combine, they would not be able to repress the promoter and control gene expression. This kind of behavior becomes desirable, since when low quantities of the repressor are present, we would expect lesser repression in the case of a high cooperativity repressor. This would generate a response that would be closer to the digital output, which is desired. This is further explained in the modeling section, but just for a brief idea, here is what the picture looks like, for increasing cooperativity.
Figure – Variation of output with repressor concentration. As can be seen, the response for n = 1 is the furthest away from a digital response, while as n increases, the output moves from 1 to 0 in a more digital fashion.
The Square Wave Generator
With the goal of searching and employing high cooperativity repressors, we looked at a paper by Voigt et al (Nature Chem Biol. 2013), in which they reported 73 analogs of the TetR repressor, of which 16 of them were found to be perfectly orthogonal to each other. A few of these repressors had really high cooperativities, with the highest n value being 6.1 for the Orf2 repressor.
We therefore decided to use repressors which were orthogonal and had the highest possible cooperativity, to demonstrate computationally, how these could be used in simple devices such as NOT gates and novel collapsible AND gates (where once the output switches from 1 to 0, it cannot be switched back. Think of it as a fuse box, which melts if the input voltage goes higher than a certain point). This entire module was called the Basic Logic Assessment and Signaling Toolbox, or the BLAST Toolbox.
Having demonstrated these successfully, we moved ahead to our main aim, which was to engineer and demonstrate a square wave oscillator in E.coli. Square waves are commonly used in electrical circuits, and can have a wide array of applications in various areas in biology such as clock inputs for timing events, time dependent drug delivery, switching of metabolic pathways and shunt activation, and would also help understand variations of biological clocks such as the circadian clock, whose gene regulatory network still remains largely unknown. This has been discussed at length in the next section.
Figure – Comparison of a sine wave and a square wave. While both sinusoidal and square inputs are used in electrical engineering, the analogue of square waves in biological systems has not yet been reported. This largely limits the applications of oscillators to regions where a sine wave is required, since a proper, well characterized square wave oscillator has not been reported.