Overview

Our project aims to design a more environmentally friendly and interesting biological dyeing method.

Our wet laboratory work focuses on controlling pigment expression and increasing pigment production. We have confirmed the reliability of upstream gene editing control expression based on the cre-loxp system and midstream information transmission based on the quorum sensing system. Since it is very complex to improve pigment production by reengineer the metabolic network of E. coli, we need accurate models and software to combine wet experiments to improve pigment production. Our model predicted several genes that are strongly associated with pigment production in the metabolic network of E. coli K12. Our software screened the vioE random point mutation library through self-adapted multi-model ensemble learning to help us select the most efficient vioE mutant. Our hardware is based on the actual needs of our project's coloring to design and develop the device Suncraft, which can reliably control the lighting of the specified area and provide a suitable environment for the sample reaction. In addition, our implementation and entrepreneurship demonstrate the market potential of our products. This is a rigorous assessment of the commercial value of the project, indicating a bright future for our products.

Figure1. An overview to the proof of concept. The picture has shown the interation between each part.

Yield improvement module

Our wetlab work successfully screened a batch of amino acid sites with high transformation potential through gene family analysis and molecular docking. And we selected two sites to construct a dual-site mutation library. We chose to use the crude yield of violacein to characterize the activity of vioE, and thus software obtained the first batch of enzyme activity-protein structure characterization data. Based on the obtained causal relationship of specific enzyme activity contributions, we were able to screen the samples with the greatest enzyme activity in the larger virtual mutant library.

Figure 2. vioE semi-rational design
a. Amino acid site libraries with high design potential; b. Optimized integrated sampling shows the expected range of the highest enzyme activity; c. Virtual mutant library screening sample enzyme activity assay

We successfully screened several target genes that have obvious "push-pull" effects on violacein synthesis, genetically modified these genes in chassis microorganisms, and successfully obtained relevant strains. We also determined the growth curves of the modified strains to evaluate the impact of these genetic modifications on host organism performance. Finally, we verified their effectiveness in improving product yield.

Figure 3 Growth curve of metabolically engineered strain measured for yield verification
a. Growth curves of two gene knockout types; b. Determination of violacein production of two gene knockout types.

Communication module

We tried to verify that AHL molecules generated by heteroexpression of pagI gene from Pantoea agglomerans in E.coli K12 can bind to RhlR in Pseudomonas aeruginosa which can bind with C4-HSL and induce the corresponding promoter expression. Therefore, we demonstrated by mass spectrometry that the type of AHL produced by heterogeneously expressed pagI gene in E.coli is still C4-HSL.

PagI generates AHL molecular type analysis and functional verification of PagI/RhlR quorum sensing system. A)AHL molecular type detection by mass spectrometry. B)PagI/RhlR quorum sensing system function verification.

Further, we also verified that salicylic acid synthesis genes PagI/RhlR and pchBA/nahR can constitute quorum sensing systems respectively, and they are independent of each other, so as to control the expression of different genes.

Figure 5. PchBA/nahR function verification and two quorum sensing systems’ independence verification. A) PchBA/nahR function verification. B) Verification that the substance produced by pchBA can’t influent genes controled by RhlR. C)Verification that the substance produced by pagI can’t influent genes controled by nahR.

painting module

Through literature investigation and algorithm prediction, we successfully selected and constructed a painting module based on blue-light-Inducible recombinases. In the experiment, we verified that the Escherichia coli that co-transferred the corresponding plasmid underwent protease induced expression, blue-light induced polymerization and other steps, and considerable gene editing occurred, which indicated that we could effectively use this system to control the write in of DNA stored information.

Figure 6 the edit situation of “disk unit”

Software

In pigment synthesis pathways of our wet lab, a key enzyme we identified, vioE, is the limiting factors that determine yield. In current computational studies of protein rational design, integrating various molecular properties and shaping the energy landscape to extract precise features is a significant challenge. Traditional multi-objective optimization methods are easily affected by confounding variables, leading to poor interpretability of the optimization process and making it difficult to extract principles from successful designs. To address this issue, the software developed a Bayesian optimization method that eliminates confounding factors by causal discovery and test, successfully identifying a mutant with enzyme activity at 132% of the wild type in regions with maximum predictive expectation. 

Model

We first used the CobratToolBox package to construct the metabolic network model of Escherichia coli and integrated the required reaction pathways into the network. By utilizing the optgene function, we calculated the flux of pigment production reactions after knocking out certain genes. Subsequently, we adopted a comprehensive modeling approach that combines cellular automata (CA) and a nonlinear convection-diffusion model to validate the experimental observations of bacterial growth and pigment deposition on fabric surfaces. This model not only considers the nonlinear effects of bacterial density but also incorporates Darcy’s law to describe the interactions between pressure gradients and fluid dynamics, successfully simulating the diffusion patterns of pigment concentration on complex surface structures. The model makes the visualization of the knockout results obtained by the optgene function more distinct. Conducting gene knockout experiments in the wet lab would require an enormous amount of effort. By using the metabolic network model, we were able to provide more promising experimental options, thus reducing the workload of the wet lab and offering a quantitative basis for further experiments. Additionally, the model verified the robustness and reliability of experimental parameters under various conditions.