Methodology in synthetic biology reflects basic principles in engineering, and the engineering cycle reigns amongst one of the most significant ones. The engineering cycle offers comprehensive and systematic guidance on the design and test of biological systems and consists of four stages: design, build, test, and learn.
We propose a trace chemical detection platform combining modeling and hardware design. We are fully aware that a standardized engineering cycle is even more critical to our cross-disciplinary work, and we followed it strictly and explored it innovatively.
We have reached decent results with the implementation of it, and a summary is as follows.
- In the first stage, we designed the genetic circuits, possible optimal strategies, and models.
- In the second stage, we constructed a designed circuit with modeling assistance.
- In the third stage, we implemented as desired and collected the data carefully.
- In the last stage, we feed the result to modeling, hardware design, and experiment, gaining precious experience to initiate a new cycle.
We have also done more beside the key engineering stages and parts listed here. For more information regarding the experiment and modeling, please refer to our Results and Model page.
Cycle 1: The Threshold Guard Switch
Iteration 1
Design
We first decided to design the basic circuit with typical and representative inducers, benzoic acid and 3-methyl benzoate (3MBz), to identify the basic features and conditions required for the best performance.
Moreover, we are aware of the fabulous mathematical properties of the circuit and decided to develop models based on it.
Build
To modify the function of the threshold guard, we initially inserted it downstream of the XylS/Pm expression module in plasmid pET28a(+) after codon optimization. The plasmid was then transformed and expressed in E. coli BL21(DE3).
Moreover, we built a model upon the digitalizer's threshold and the repressor's strength, giving guidance to the experiment threshold control (see page Model), mainly using numerical simulation on differential equations on Simulink. We calculated the rough interval of the threshold of our digitalizer and designed a series of experiments for our wet lab to locate the absolute entry.
Test
To evaluate the performance of the threshold guard, a series of tests were carried out on the gradient concentration of the inducer to find the turning-on threshold of the guard switch. We used benzoic acid and 3-methyl benzoate (3MBz) as inducers (input), recorded the time span, and fluorescence intensity as output. All the measurement was done by synergy HTX microplate reader.
After collecting the data, we fit it and compared it with the model.
Learn
We found that a small amount of inducer will turn on the guard switch (between 100~150 μM benzoic acid), allowing us to detect small molecules at a precise degree. Therefore, we replaced the downstream to the PobR system (described in the next part of this cycle). The time required to ultimately peak was not that promising, with the minimum time 99 minutes under the induction of benzoic acid for 900 μM.However, this is correlated to the category and the concentration of the inducer. We hope to improve this by promoting the affinity of the capturing process, as described in the Design page.
We were also delighted to find out that the curve found the model very well, surprisingly found the actual threshold, determined its strength to be 680 \mu, and further estimated the repressor strength to be more than 10 in terms of the relative strength intervals we divided. Then, we improved the model on the threshold of the digitalizer and the strength of the repressor, giving prediction to the PobR experiment threshold control (see page Model).
Iteration 2
Design
Once we understood how the guard switch worked and confirmed that it was functional, we decided to modify the switch to better suit our requirements and assisted with modeling. By merging the pobR and digitizer components, we were able to create the pobR-digitalizer using the process dewtailed in our design section (See page Design) , which included removing certain sections of two strands of DNA and joining together the remaining pieces.
Build
The pobR-digitizer switch was established in pET-28a (+) with a Kanamycin resistance gene and cloned and expressed in E. coli Fast-T1 after codon optimization.
Test
We measured the intensity of GFP expression at a gradient concentration of 4-HBA over time. All the measurement was done by synergy HTX microplate reader. We normalized the data to better convey the actual expression status.
Learn
We were thrilled that the threshold corresponded as we predicted, and the overall fluorescence change obeyed the model.
We plan to replace GFP with luxI, producing VAI to trigger the downstream Quorum sensing system at a meager level. If LuxI is efficiently expressed, the reaction time and intensity could be promoted.
Lastly, we narrowed down the threshold range by feeding it back to the model, trying to further improve the performance of both the model and the biosensor.
Cycle 2: NOX Kit Optimization
Iteration 1
Design
Low cost and high sensitivity are not contradictory goals. To achieve this, we opted not to depend on existing development boards but instead started from the chip level, designing from basic detection components.
Build
We chose the STC8G1K08A chip as the main controller for the NOX Kit. For rapid iteration, we used a breadboard to quickly prototype the circuit and 3D printing to construct a basic reaction chamber shell.
Developing the Prototype Shell Design through team discussion
Test
Through a portable host computer program reading data from the NOX Kit's sensors, we verified that the device can capture variations in temperature and luminosity.
A Single Detector Prototype in Standalone Mode
Learn
Through testing the prototype and modeling simulations, we confirmed the feasibility of the current technical approach and began a second iteration based on requirements from biological experiments.
Iteration 2
Design
In the design process, we drew on lessons from the first prototype and adjusted the reaction chamber volume based on experiment and modeling feedback. We also chose more sensitive detection resistors around the normal operating point. Finally, we optimized the NOX Kit user interface for greater usability.
Build
We redesigned the circuit with a PCB for greater compactness and reduced component exposure for safety and reliability. For the reaction chamber, we reused the original shell to reduce 3D printing waste and only replaced the core reaction vessel. For the user interface, we introduced more control buttons and encapsulated the Python calling interface so any developer can build their own detection programs based on the NOX Kit. We also introduced the I2C bus protocol, enabling developers to build high-throughput detector arrays based on the bus.
PCB Design of the circuit
Test
In the circuit prototype design, we used EDA built-in trace checking to ensure design validity. For developers, we provided Python encapsulation with dummy_nox for debugging and nox classes for actual development. Through repeated debugging, we ensured the user control logic was correct with no bugs during use.
Learn
During development, we found the NOX Kit could not only be applied in the current project but also be modularly expanded for other detection capabilities in the future, with great extensibility. With accumulating experience, the NOX Kit shows promise to evolve into a versatile, low-cost bio detection platform.
References
[1] Calles, B., Goñi-Moreno, Á., & de Lorenzo, V. (2019). Digitalizing heterologous gene expression in Gram-negative bacteria with a portable ON/OFF module. Molecular systems biology, 15(12), e8777. https://doi.org/10.15252/msb.20188777
[2] Arrow icons from loading.io.
