After completing the first version of the Prometheus build, it is capable of processing natural language, searching for parts, and assembling plasmids. The best way to demonstrate the success of the Prometheus model is by using it to successfully direct an experiment.

In 2024, NJU China utilized the Prometheus model to conduct a small yet significant synthetic biology experiment with potential real-world applications. This not only showcased the model's ability to guide synthetic biology experiments but also demonstrated its potential to make a real contribution to the world.

During the course of wet experiments, the feedback from the wet experiment helped to create new inspiration for the dry experiment.

Trehalose is a non-reducing disaccharide composed of two glucose molecules linked by a 1,1-glycosidic bond. Its chemical name is α,D-glucopyranosyl-α,D-glucopyranoside, with the molecular formula C12H22O11·2H2O and a molecular weight of 378.33. Trehalose can be endogenously expressed in Saccharomyces cerevisiae. Research shows that trehalose serves not only as a microbial carbon reserve but also as a natural protectant, playing a crucial role in protecting cells from various stresses such as high temperatures and osmotic pressure by stabilizing cell membranes, protecting proteins, and maintaining osmotic balance.

Saccharomyces cerevisiae is widely used in various production settings, ranging from food and pharmaceuticals to biofuels. However, not all production environments are conducive to its survival; factors such as unsuitable temperatures, toxicity of byproducts, excessive osmotic pressure, and chemical toxicity can limit yeast productivity. Therefore, yeast strains with high trehalose expression exhibit enhanced resilience in adverse production conditions, resulting in improved survival and higher productivity. These strains can even thrive in previously inhospitable environments, opening up new possibilities for production applications.

This broader application potential enables biological production to replace chemical processes, significantly reducing the use of chemical additives and the generation of hazardous waste. Additionally, it conserves energy, enhances safety, and positively impacts both the natural environment and human health.

Therefore, we aim to leverage the Prometheus model to create yeast with high stress resistance, demonstrating its capability to make genuine contributions to synthetic biology in real-world applications.

Through literature review, we identified the key gene TPS1 in the endogenous synthesis pathway of trehalose, confirming that high expression of TPS1 can increase the content of endogenous trehalose.

We aim to achieve high expression of the TPS1 gene in yeast using a strong promoter. After discussions with Professor Shi Jing from the School of Life Sciences at Nanjing University and Professor Shi Junfeng from Fudan University Affiliated Hospital, we defined the inputs for synthetic biology standards. Consequently, we provided the Prometheus model with the following input: "I want to highly express the TPS1 gene in yeast. Please help me find a strong constitutive promoter that can initiate expression in yeast."

We selected three strong promoters for yeast from the options recommended by the Prometheus model: TDH3 (BBa_K124002), TEF2 (BBa_K3753003), and PGK1 (BBa_K2308011). Subsequently, the Prometheus model assisted us in constructing plasmids, and we successfully built plasmids using the TDH3 and TEF2 promoters in the laboratory.

The Construction of Plasmids pYET-TPS1-TDH3 and pYET-TPS1-TEF2

We cultured four types of yeast: one overexpressing the TPS1 gene using the TDH3 promoter, another using the TEF2 promoter, a third with the pYET plasmid overexpressing TPS1, and a control yeast transformed with an empty pYET vector. After four days of fermentation, we collected samples. We measured the mRNA concentration of each promoter through total RNA extraction, reverse transcription, and quantitative PCR. This allowed us to verify whether TPS1 was overexpressed, confirm the correct construction of the plasmids by the Prometheus model, and compare the effectiveness of the three promoters to validate the model's part selection.

Amplification curve of TDH3, TEF2, ADH2 and pYET
Legend:
Light blue: ADH2-1; Gray: pYET-1; Orange: ADH2-2; Yellow: pYET-2;
Green: TEF2-1; Pink: TEF2-2; Dark blue: TDH3-1; Red: TDH3-2.
Relative expression of TDH3 relative to blank group: 2^{(31.48-19.92)}= 3019
Relative expression of TEF22 relative to blank group: 2^{(31.48-23.32)} = 286
Relative expression of ADH2 relative to blank group: 2^{(31.48-20.49)} = 2033
Relative expression of TDH3 relative to ADH2 group: 2^{(20.49-19.92)} = 1.484

Relative expression to blank group

From the amplification curves, it can be seen that all six experimental groups of the three promoters greatly increased the expression of TPS1, and the Ct values were all in the range of 20-25. The Ct values of the blank group were all after 30. The successful expression of the promoters TDH3 and TEF2 provided by Prometheus was demonstrated, proving the usefulness of the model. Comparing the expression of TDH3 promoter and the original ADH2 of pYET plasmid, the TDH3 promoter increased the mRNA concentration of TPS1 by nearly 50% compared to ADH2, proving once again the superiority of Prometheus in part selection.

Subsequently, we conducted stress resistance tests on the four yeast strains. We placed them in conditions unsuitable for yeast growth at 37 degrees Celsius. OD values were measured every two hours, and we plotted the curve of OD values over time. The results showed that the three yeast strains overexpressing TPS1 grew better at 37 degrees Celsius than the control strain. Among them, the yeast overexpressing TPS1 using TDH3 exhibited the best growth, with TDH3 provided by the Prometheus model. The figure below effectively illustrates the differences among the four yeast strains.

OD value versus time curve under cultivation at 37°C.

Photos of cultivation at 37°C.
Legend: From left to right: TDH3, TEF2, pYET, ADH2.

During the experimental process, we realized that the parts provided by the Prometheus model may not always be the optimal choice, and the effectiveness of output parts corresponding to the same input can vary. We propose that the model should include a feature that allows for real-time user feedback on the performance of parts and experimental data from experiments guided by the Prometheus model. This feedback can be used for reinforcement learning and model optimization, further enhancing its capabilities.

Moreover, we recognized that achieving good experimental results with this model still requires a solid biological background. This indicates that further improvements are necessary to broaden the model's audience and truly unleash the potential of synthetic biology. After the wet lab phase, our model embarked on a new journey of advancement.

[1] Dunwu Chen, Xiong Chen, Xin Li. Advances in regulation of endogenous trehalose metabolism in yeast. Acta Microbiologica Sinica, 2019, 59(12): 2276-2284.

[2] Paul MJ, Primavesi LF, Jhurreea D, Zhang YH. Trehalose metabolism and signaling. Annual Review of Plant Biology, 2008, 59: 417-441.

[3] Petitjean M, Teste MA, François JM, Parrou JL. Yeast tolerance to various stresses relies on the trehalose-6P synthase (Tps1) protein, not on trehalose. Journal of Biological Chemistry, 2015, 290(26): 16177-16190.

[4] Petitjean M, Teste MA, Léger-Silvestre I, François JM, Parrou JL. A new function for the yeast trehalose-6P synthase (Tps1) protein, as key pro-survival factor during growth, chronological ageing, and apoptotic stress. Mechanisms of Ageing and Development, 2017, 161: 234-246

[5] Saini P, Beniwal A, Kokkiligadda A, Vij S. Response and tolerance of yeast to changing environmental stress during ethanol fermentation. Process Biochemistry, 2018, 72: 1-12.

[6] Xu HX, Zhou P, Wang X, Zhang XP, Duan G. Effects of stress factors on trehalose accumulation in the process of ethanol fermentation. Liquor-Making Science & Technology, 2016(5): 65-69.

[7] Li CS, Yang XQ, Xu Y, Li LH, Wang YQ. Cadmium detoxification induced by salt stress improves cadmium tolerance of multi-stress-tolerant Pichia kudriavzevii. Environmental Pollution, 2018, 242: 845-854.

[8] Tan HG, Dong J, Wang GL, Xu HY, Zhang CY, Xiao DG. Enhanced freeze tolerance of baker's yeast by overexpressed trehalose-6-phosphate synthase gene (TPS1) and deleted trehalose genes in frozen dough.

[9] Divate NR, Chen GH, Wang PM, Ou BR, Chung YC. Engineering Saccharomyces cerevisiae for improvement in ethanol tolerance by accumulation of trehalose. Bioengineered, 2016, 7(6): 445-458.