---

Summary

Summary: With usability at the forefront, we designed a set of integral controller to facilitate robustness in a synthetic genetic circuit with applications in metabolic engineering (see design and/or glossary).

Problem: Synthetic biology is incredible and allows us to engineer biology with novel abilities, and use it in many different applications. However, synthetic genetic circuits are fragile and often fail, especially when scaled to industry (Human Practices).

Solution: To help tackle this fundamental challenge we are applying engineering principles of control theory to synthetic biology to create an integral controller. An integral controller acts to hold production of a protein at a reference level in response to perturbation (see design)

Build design: Throughout our project we implemented the core engineering basis of the design-build-test-learn cycle. Our project was split into three prongs: (1) model, (2) through factorial design build circuits using JUMP cloning, (3) test capacity for robustness (see results)

Importance of human practices: To ensure that our project has real-world value and will be used with consulted academics, industrial experts and potential users to design a set of standardised controllers tailored for different users needs. Human practices has been a core principle of our design process (see human practices)

Applications: Our circuit can be directly applied to synthetic biology and a variety of applications including metabolic engineering, biomanufacturing and microbial drug delivery (see implementation)

Implementation: Designed a use-friendly calculator to determine which circuit to use and how it can be implemented see implementation).

The Problem

Synthetic biology gives us the ability to manipulate living systems. Through genetic and metabolic engineering it also us the ability to provide novel abilities with many real-world applications. However the field is relatively new and there are still many challenges before its full potential is to be reached. In particular molecular interactions are random which results in system stochasticity. The result of this is that designed synthetic circuits are considered “fragile” to disturbances both intrinsically and extrinsically. This is particularly relevant when building larger, more complex circuits and when applying these circuits outside the lab and in industry (see human practices for more detail). Therefore to ensure that synthetic biology can grow as a discipline the challenge of circuit robustness must be tackled head on.

Our Motivation

Throughout our ideation process we used a framework based in effective altruism (see human practices). This philosophy is founded on the principle of using evidence to maximise the amount of good with the minimal amount of action. This was important to us as we wanted to ensure that synthetic biology and our project could make a real-world change. Furthermore, we reasoned, to ensure we have a large impact we wanted to tackle a fundamental problem in synthetic biology. Solving these problems would result in the most good. This led us to the path of circuit robustness and integral controllers.

What is our solution

One field of research that is directly targeting this challenge of circuit robustness is cybergenetics. Cybergenetics describes the combination of control field and synthetic biology. Our project aimed to implement an integral controller in a genetic circuit (see below). An integral controller acts to ensure that the bacteria produce a desired product consistently at a desired setpoint i.e. a controlled quantity of molecules. An integral controller reduces the impact of perturbation to the system through a property called robust perfect adaptation (RPA). RPA describes a molecular homeostasis which expression is kept at a set point.

Control theory provides the toolkit to design our circuit

Control theory is integral helping solving the challenge of circuit robustness. The engineering field of control theory describes the approaches to develop a system that ensures the output is at a desired state through the action of a controller. Control theory is found through biology as feedback circuits. In a feedback circuit, the output influences the input which in turn alters the output levels. We have build an integral controller which can achieve a ‘zero steady state error’ because it has a ‘memory’ of the error. How far away the output is from the set point is called the steady state error. Integral controllers are found in every-day life. For example, a cruise controller in a car will be an example. The speed at which you would want to drive is called the set point. If the car is slower than the set speed, then the controller would increase acceleration to match set speed. Throughout, the controllers constantly evaluate the speed the car is at and the magnitude of how much you are below the setpoint by. This is the ‘memory’ of the error.

Below is a short animation showing the basic operation of the antithetic integral control motif and how it can ensure robust adaptation to perturbations.

We have designed an antithetic integral controller based on Aoki et al., 2019. An antithetic motif describes the process in which two controllers permanently annihilate (cancel) each other. A perturbation may result in an increase expression of X. This results in more Z₂ production (since X increases Z₂ expression). More Z₂ annihilates the Z₁ controller. Z₁ promotes expression of X, resulting in less X produced, and therefore compensates for an increase in X. This ensures a steady state of zero. The difference of controller protein concentration accumulation depends on how far away the system is from the desired output. This is the ‘memory’ of the error.

Within our project we aim to compare three different circuits: constitutive expression (1), autoregulatory negative feedback loop (2) and an antithetic integral controller (3). We tested the capacity for circuit robustness or upon perturbation, whether expression level returns to the desired set point. In silico, we show that an integral controller is the only method tested which enables robust perfection adaptation.

The Project

To successfully build the circuits we implemented a three step approach: (1) modelling, (2) build and (3) test.

First, we wanted to model our circuits to help characterise our circuits and aid in our design process. A variety of different modelling approaches were used to optimally characterise the system. Firstly, deterministic ordinary differential equations (ODEs) were used to verify our circuit and to fit our data to better characterise the circuit. Secondly, we used stochastic simulations to determine noise in our system and then global sensitivity analysis to identify the best ways to perturb our circuits. Finally, we used control theory to evaluate the stability of our system. Throughout there was tight integration between our modelling and experimental designs. Extensive parameter optimisation enabled pre-experimental analysis which allowed predictions of which circuits would work best in our combinatorial design.

Once we had optimised our design process through modelling we built these circuits. Throughout our project we used joint universal modular plasmids (JUMP) assembly. This is a novel golden gate assembly which took advantage of the iGEM distribution kit. JUMP assembly enables modular and quick assembly of highly complex genetic systems. First, we cloning all the individual genetic parts into level 0 plasmids. From this, we assembled level 1 transcriptional units (devices) and later level 2 systems. We also took advantage of a downstream cloning site to make a 5 transcriptional unit system (see design page for more information). Whilst we made 39 level 0 plasmids and 37 level 1 plasmids we were unable to complete any level 2 circuits.

However, our combinatorial design should act to demonstrate the incredible value of JUMP assembly in assembly of complex circuits whilst relying on engineering factorial principles. We hope that future iGEM teams would consider similar design principles.

Finally, we planned to test our assembled circuits. In order to demonstrate that we created an integral controller we first planned test its capacity using positive perturbation of ‘X’. Secondly, we planned to test the circuits against alternative strategies to achieve RPA. Finally, to illustrate the incredible utility of these circuits to adapt to perturbations we planned to test our circuits against global perturbations of temperature, nutrients/toxins (ammonia) and antibiotics (chloramphenicol).

Human practices significantly influenced our values and design process

Using effective altruism as a framework for our ideation we wanted to target a fundamental challenge across synthetic biology (see human practices). However to establish whether our project could have value in the real world we spoke to a variety of different professionals and potential users. This led us to understanding that our project has value in scientific, industrial, environment and medical applications. We also wanted to integrate our human practices to optimise usability of our design circuit. In particular these led to changes in how the circuit would be perturbed, the model organism, plasmid type, modelling approaches, combinatorial design and enhanced modularity.

Applications and Implementations

Cybergenetics has significant value to synthetic biology to facilitate circuit robustness. Principally an integral controller could have the ability to enhance metabolic engineering. Metabolic engineering describes the ability to enhance the cells ability to produce a certain substance and as arguably a basis for all applications of synthetic biology. See figure below

We have tried to ensure that our circuit is as user friendly as possible. Firstly, our combinatorial design allows users to choose a circuit which best matches what features of expression would be important. These include protein expression, metabolic load or steady state recovery rate. Secondly, we have designed the implementation to be as simple as possible. In a one-step restriction enzyme knock-in reaction, the user can assemble their circuit.

References:

1. Kwok, R. Five hard truths for synthetic biology. Nature 463, 288–290 (2010). https://doi.org/10.1038/463288a


2. Aoki SK, Lillacci G, Gupta A, Baumschlager A, Schweingruber D, Khammash M. A universal biomolecular integral feedback controller for robust perfect adaptation. Nature. 2019 Jun;570(7762):533-537. doi: 10.1038/s41586-019-1321-1. Epub 2019 Jun 19. PMID: 31217585.


3. Del Vecchio D, Dy AJ, Qian Y. Control theory meets synthetic biology. J R Soc Interface. 2016 Jul;13(120):20160380. doi: 10.1098/rsif.2016.0380. Epub 2016 Jul 20. PMID: 27440256; PMCID: PMC4971224.


4. Aoki SK, Lillacci G, Gupta A, Baumschlager A, Schweingruber D, Khammash M. A universal biomolecular integral feedback controller for robust perfect adaptation. Nature. 2019 Jun;570(7762):533-537. doi: 10.1038/s41586-019-1321-1. Epub 2019 Jun 19. PMID: 31217585.