LOADING
Our iGEM project was conceived through a combination of academic inspiration, real-world clinical observation and emerging pharmaceutical advances. The initial idea emerged when a team member encountered a Nature report on next-generation glucose-lowering strategies during a specialized course, which highlighted the need for innovative therapeutic approaches. This academic insight was reinforced by observing the practical challenges faced by diabetic patients such as the frequent injections and resistance issues encountered by a PhD student in our lab, revealing the clear limitations of existing treatments. The project direction was further validated when Novo Nordisk launched China's first oral GLP-1 receptor agonist, confirming both the feasibility and clinical relevance of oral delivery routes for metabolic disease treatment.
Building on these insights, we designed an engineered probiotic system that integrates biosensing, protein engineering and biosafety circuits to enable smart, ultra-long-acting GLP-1 production directly in the gut. By combining biological innovation with patient-centered design, our work not only offers a potential alternative to conventional injection-based therapies but also provides the iGEM community with a modular framework for developing safe, responsive and orally delivered living therapeutics.
To establish an end-to-end pipeline for the design and preparation of oral GLP-1-based bacterial therapeutics, we have introduced a variety of genetic circuit components and protein engineering elements into the Escherichia. coli (E. coli) host chassis. These parts support GLP-1 synthesis and signal transduction, providing a rich toolkit for molecular sensing, amino acid sequence modification and biosafety assurance within the gut environment.
We have newly introduced a protocatechuic acid (PCA)-responsive operon (BBa_25N5FWAK) to sense PCA, a metabolite derived from green tea. Furthermore, we employed orthogonal components based on the split intein trans-splicing principle (BBa_2581ND93 and BBa_25MR74ON) to construct an AND-gate sensing circuit. This design creates an intestinal engineered probiotic with a spatiotemporally regulated production switch, activated by the dual input signals of PCA as the "inducer" and sodium cholate as the "localizer". Our work demonstrates a functional framework for constructing AND-gate circuits that respond to gut molecular signals, paving the way for other teams to develop similar sensing pathways.
Additionally, integrating the upstream AND-gate signal to orthogonal amplification and feedback regulation systems can further optimize the final output expression level. While the Amp30E composite part has been well-developed and applied, previous teams have seldom integrated the hrpV-based feedback component into circuits or validated the combined amplification and feedback effects. Concurrently, we incorporated the glucose uptake negative-response promoter GURB3-2 (BBa_25TRS3OS) to conducte preliminary characterization of its integrated performance, enabling feedback regulation aligned with feeding and fasting states.
| Part Number | Name | Type | Brief Introductions |
|---|---|---|---|
| BBa_25Y8JP8F | PLV | Promoter | A protocatechuic acid (PCA)-inducible promoter PLV. |
| BBa_25OBUNJP | PcaV | Coding | The protocatechuic acid (PCA)-responsive transcriptional repressor PcaV. |
| BBa_25YZWH1A | SspGyrBC-ECF16C | Coding | A fusion protein comprising the C-terminal halves of split intein SspGyrB and split ECF16. |
| BBa_259H9ZPU | SspGyrBN-ECF16N | Coding | A fusion protein comprising the N-terminal halves of split intein SspGyrB and split ECF16. |
| BBa_25RT9PC8 | P16 | Promoter | The ECF16-cognate promoter P16. |
| BBa_25TRS3OS | GURB3-2 promoter | Promoter | A glucose uptake negative-response promoter GURB3-2. |
| BBa_25N5FWAK | J23105-PcaV-PLV | Device | A protocatechuic acid (PCA)-responsive operon. |
| BBa_25MR74ON | P16-GFP | Reporter | A GFP reporting circuit for ECF16 expression. |
| BBa_2581ND93 | J23105-PcaV-PLV-SspGyrBC-ECF16C-J23101-RamA-PacrRA-SspGyrBN-ECF16N | Device | An upstream AND-gate circuit based on split intein trans-splicing with orthogonal SspGyrB-ECF16. |
In this module, we have introduced multiple human-derived GLP-1(7-36) mutants suitable for protein purification and secretory expression. Additionally, we incorporated two orthogonal aminoacyl-tRNA synthetase/tRNA (aaRS/tRNA) pairs for the in vivo biosynthesis and site-specific incorporation of non-canonical amino acids (5-HTP and sTyr), along with the superfolder GFP (sfGFP) reporter protein (BBa_25R2L2YD). The module also includes the plasmid pUltra (BBa_25A3A634) for better aaRS/tRNA expression. These parts enable future teams to design other genetic code expansion systems and evaluate the effects of ncAA incorporation on protein properties.
| Part Number | Name | Type | Brief Introductions |
|---|---|---|---|
| BBa_25DRMK5X | SUMO-GLP-1 | Coding | A fusion peptide with N-terminal His-tag, SUMO-tag, and codon-optimized human GLP-1 (7-36) sequence. |
| BBa_25FJQXDY | S-GLP-1Z-CPP | Coding | A fusion peptide with a signal peptide, GLP-1Z, a cell-penetrating peptide and a C-terminal His-tag. |
| BBa_2504KELH | SUMO-GLP-1Z-CPP | Coding | A fusion peptide with N-terminal His-tag, SUMO-tag, GLP-1Z and a cell-penetrating peptide(CPP). |
| BBa_25QIIBEZ | 5-HTP aaRS | Coding | Orthogonal 5-hydroxytryptophan (5-HTP) aminoacyl-tRNA synthetase. |
| BBa_25LSICU8 | MjtRNAsTyr | Coding | Orthogonal tRNA for transferring sulfotyrosine (sTyr). |
| BBa_255G11B4 | SctRNA5-HTP | Coding | Orthogonal tRNA for transferring 5-hydroxytryptophan (5-HTP). |
| BBa_259WRPM0 | sTyr aaRS | Coding | Orthogonal sulfotyrosine (sTyr) aminoacyl-tRNA synthetase. |
| BBa_25R2L2YD | sfGFP reporter protein harboring amber codons | Coding | sfGFP reporter protein harboring two amber codons. |
| BBa_252MVYLF | proK promoter | Promoter | The proK promoter is a natural strong promoter in E. coli, primarily driving efficient tRNA gene transcription and used for tRNA expression cassettes. |
| BBa_25A3A634 | pUltra | Plasmid | The pUltra plasmid is a high-efficiency tool for site-specific incorporation of non-canonical amino acids in E. coli. |
| BBa_2565RBMY | Orthogonal 5-HTP aaRS and tRNA | Coding | 5-HTP aaRS can pair with SctRNA5-HTP to form orthogonal aminoacyl-tRNA synthetase/tRNA pairs. |
| BBa_2543FT42 | Orthogonal sTyr aaRS and tRNA | Coding | sTyr aaRS can pair with MjtRNAsTyr to form orthogonal aminoacyl-tRNA synthetase/tRNA pairs. |
We constructed a well-characterized thermoswitch system: the PtlpA-tlpA* self-repressive system (BBa_2575N2V6) based on a temperature-responsive PtlpA promoter (BBa_25RFWS4Y) and the corresponding sequence-optimized repressor TlpA* (BBa_25XF4HZ5). To enhance temperature sensitivity under repressor-free conditions, we further designed a tandem dual-promoter construct (BBa_2552PNC9). By integrating these components, we built a final temperature-arabinose dual-input OR-gate biosafety circuit (BBa_25NKC5AV). This system provides a reference strategy and practical insights for constructing OR-gate logic in a relatively simple form, particularly suitable for the design and development of biosafety circuits requiring precise in vitro and in vivo temperature control.
| Part Number | Name | Type | Brief Introductions |
|---|---|---|---|
| BBa_25XF4HZ5 | TlpA* | Coding | The temprature-responsive transcriptional repressor TlpA* |
| BBa_25RFWS4Y | PtlpA with RBS | Promoter | The temprature-responsive promoter PtlpA with its cognate RBS |
| BBa_25XUNXCE | RiboJ | Coding | Self-cleaving ribozyme RiboJ as an insulator in bacterial systems |
| BBa_25NOQJ8F | PBAD | Promoter | Minimal arabinose-inducible promoter pBAD |
| BBa_2575N2V6 | PtlpA-tlpA* | Device | The temperature-responsive ptlpA-tlpA*self-repressive system |
| BBa_2552PNC9 | J23101-PBAD-RiboJ | Promoter | Tandem dual-promoter element with the insulator RiboJ |
| BBa_25NKC5AV | PtlpA-tlpA*-araC-J23101-PBAD-RiboJ | Device | A dual-input OR-gate circuit responsive to arabinose and temperature |
After multiple rounds of experimental validation and characterization, we have selected our high-performance OR-gate biosafety circuit (BBa_25NKC5AV) as a candidate for the Best New Composite Part special prize, based on its optimal performance, adaptability and modular transferability.
Safety is both the foundation and ultimate goal of our project, especially when developing intestinal microbial therapeutics that directly impact human health. To this end, we have innovatively established a systematic safety practice framework: "HELPPS" model which integrates safety principles across the entire project lifecycle. For details, please visit our Safety page.
• H (Human Practices): We engage with diverse stakeholders to identify real-world safety concerns, ensuring our design reflects actual needs and avoids oversight.
• E (Education): We implement rigorous safety training for all members and promote awareness beyond the lab, building a culture of responsibility from the start.
• L (Laboratory Safety): Through zoning management, clear protocols, and collective accountability, we maintain a safe experimental environment and minimize risks.
• P (Policy Compliance): We strictly follow institutional and national regulations, use only approved biological materials, and ensure full legal and ethical compliance.
• P (Project Design): Safety is integrated at the design stage from chassis selection and multi-layer genetic safeguards to hardware delivery and AI risk assessment.
• S (Safety as Core Goal): Safety always comes first, even at the potential cost of experimental efficiency, and extends from the lab to public outreach.
In summary, HiZJU-China has built a safety ecosystem that spans from concept to implementation, lab to society and technology to governance. We have demonstrated how safety can be systematically designed, implemented and communicated, making synthetic biology applications such as live therapeutics more controllable, trustworthy and feasible.
Based on this, we decided to compete for the Safety and Security Award.
To support the development of a long-acting GLP-1-based oral microbial therapeutic, our modeling efforts span multiple scales from GLP-1 stability engineering to intestinal targeted delivery and bacterial growth kinetics. These models provide crucial data and theoretical support for the functional realization, industrial production and safe application of the engineered bacteria. Complete model details, parameters and interactive plots are available on our team’s Model page.
We have established a reproducible and adaptable workflow for molecular modeling. For systems with unknown protein–protein spatial interactions, we integrated molecular docking and 300 ns molecular dynamics (MD) simulations, confirming that the rationally designed GLP-1Z mutant (A8H, E27K, K34R) exhibits reduced binding stability with the degrading enzyme NEP, clarifying a previously unexplored aspect of their interaction. Subsequently, we performed non-canonical amino acid (5-HTP) scanning using Rosetta, identifying potential sites that enhance both stability and binding activity. This pipeline offers precise guidance for peptide optimization and is generally applicable to systems where key influencing factors are clarified. It can even be adapted for use with AlphaFold3 predictions, reducing dependency on experimentally resolved protein structures.
After obtaining the engineered E. coli, we focused on modeling probiotic release kinetics of hydrogel delivery carriers. We adapted several hydrogel microsphere release models that account for both diffusion and swelling mechanisms. Based on Fick’s laws, we derived a drug release kinetic model for spherical carriers and fitted experimental data using the Peppas empirical model. This helped us to identify key controlling factors in the release kinetics of sodium alginate–calcium chloride hydrogels, laying a foundation for tuning cross-linker concentration and polymer ratios to design hydrogels suitable for in vivo applications. This model also provides a practical theoretical tool and parametric design strategy for the iGEM community working on gut-targeted delivery systems.
Since the modified probiotics are intended for in vivo use, we modeled the growth kinetic parameters of the engineered bacteria to understand their competitive advantages relative to other gut microbiota. This aids in optimizing embedding density and storage parameters. Based on microbial fermentation kinetics and material balancing, we also evaluated the metabolic response of the strain to inducer addition, providing insights for maintaining bacterial density in the gut environment. Apart from in vitro GLP-1 production, this model supplies essential parameters for industrial scale-up as well. Together with our in vitro turbidostat platform, it offers a standardized system for safely and controllably testing genetic circuit performance.
Although these modeling efforts are distributed across different modules, they collectively provide the necessary parametric foundation for realizing an effective oral bacterial therapeutic. They guide subsequent experimental optimization and reduce unnecessary experimental iterations. Our modeling contributions are not merely theoretical but are deeply integrated with wet-lab work.
Among them, the systematic protein stability engineering employed in our project not only extends protein engineering from natural to non-canonical amino acids but also represents a practical application of AI in protein design, serving as a reproducible and transferable exemplar. Based on this, we finally decided to compete for the Best Model special prize.
The primary objective of our in vitro turbidostat design was to address the limitations of direct in vivo experimentation. This system aims to simulate the dynamic balance of bacterial density in the intestinal tract, thereby providing a stable initial bacterial load for genetic circuit characterization and enabling accurate measurement of key parameters such as induction time and inducer dosage.
To construct this simplified turbidostat, we assembled a system comprising feed bottles, a three-neck flask, and peristaltic pumps. Bacterial density (OD600) was monitored in real-time to determine when the culture reached the turbidostatic state. By precisely adjusting the inflow and outflow rates, we maintained the engineered E. coli concentration at approximately OD600 = 1.4 for over three hours, simulating typical intestinal E. coli abundance levels.
Once turbidostatic conditions were established, inducers specific to the AND-gate circuit including protocatechuic acid (PCA) and sodium cholate were introduced via syringes. We then collected samples at regular intervals and assessed target gene expression by measuring GFP fluorescence intensity, allowing us to optimize the induction conditions effectively.
Using this customized turbidostat, we successfully maintained the bacterial culture at a constant density and determined that the AND-gate circuit requires approximately 3 hours to respond fully to induction (read our Results page for details). This provides a preliminary estimate of the time needed for engineered EcN to produce GLP-1 in a simulated gut environment, offering a realistic perspective on its functional potential before advancing to in vivo studies.
As an innovative attempt leveraging our major background in bioengineering, this hardware setup offers a valuable reference for other teams, particularly those in medically oriented tracks lacking immediate access to in vivo testing facilities, to conduct reproducible, controlled circuit characterization in vitro.
For safety and applicability reasons, we selected Escherichia coli Nissle 1917 (EcN) as our final chassis strain and successfully expressed parts of the protein module in this host. However, due to the presence of endogenous plasmids and its unique membrane structure and composition, conventional chemical transformation proved inefficient in EcN. We therefore adopted electroporation as the preferred method for plasmid introduction.
During the testing phase for non-canonical amino acid (ncAA) incorporation, we aimed to co-transform EcN with two plasmids: the pUltra plasmid (StrR) carrying the orthogonal tRNA/aminoacyl-tRNA synthetase pairs, and a reporter plasmid (AmpR) encoding an sfGFP reporter protein with two amber codons. While we successfully transformed the sfGFP reporter plasmid initially, repeated attempts to subsequently introduce the pUltra plasmid into the same EcN strain failed. Extensive optimization of cell concentration and incubation conditions over two weeks did not resolve the issue.
We hypothesized that the order of plasmid introduction and the specific antibiotic resistance markers could critically affect transformation efficiency. By reversing the transformation sequence: first introducing the streptromycin-resistant pUltra plasmid, followed by the ampicillin-resistant reporter plasmid, and further optimizing the incubation conditions, we ultimately obtained EcN colonies harboring both target plasmids. These strains were subsequently used for ncAA incorporation testing.
Detailed procedures and test results are available on our Results page. We believe this practical experience provides valuable insights and a potential workflow for other teams engineering EcN-based host systems.
Our project involved the design and optimization of multiple genetic circuits, requiring consistent characterization of components with varying expression strengths under standardized culture conditions using a GFP reporter system. This includes comparing different constitutive promoters and RBS combinations. Without access to integrated high-throughput mutagenesis and characterization instruments, such optimization typically demands extensive experimental work and sequencing costs.
To address this, we adopted a practical strategy of selecting several well-characterized constitutive promoters and RBS sequences with predefined expression strengths from the iGEM Parts Registry for parallel plasmid construction and characterization. This approach not only shortened the experimental timeline but also enabled systematic comparison of how different expression levels affect circuit performance.
In optimizing the PCA-responsive operon, we initially selected the weak constitutive promoter J23117 to drive PcaV repressor expression, aiming to enhance circuit sensitivity. However, this configuration resulted in significantly reduced saturation induction levels and noticeable leakage in the absence of inducer. We subsequently tested medium- and high-strength promoters J23105 and J23101, finding that J23105 effectively balanced low leakage with strong output signal at relatively low PCA concentrations (400 μM), ensuring a robust response dynamic.
Similarly, we systematically tested RBS sequences with varying strengths (RBS30, RBS32 and RBS34) to fine-tune regulatory component expression. Through RBS optimization in our biosafety OR-gate circuit, we achieved approximately 5-fold enhancement in output signal intensity and significantly improved temperature response discrimination under arabinose-free conditions compared to earlier circuit versions. These results demonstrate that successful circuit performance depends not only on component arrangement but also critically on precise expression level matching.
For uncharacterized genetic parts, we recommend parallel construction of "component sets" with varying expression strengths from the outset to efficiently identify optimal configurations. Additional circuit optimization examples are documented on our Engineering page, which we believe will provide valuable reference for other teams designing genetic circuits.
[1] Prillaman M. (2024). Obesity drugs aren't always forever. What happens when you quit?. Nature, 628(8008), 488–490.
[2] Medbelove. (2025, January 20). Novo Nordisk: The world's first "oral version for blood sugar reduction"—semaglutide Rybelsus®.
[3] Jiang, S., Chen, H., Chen, S., Chen, N., Yang, H., Duan, Y., Ao, S., Wang, R., Wang, X., Zhang, Y., & Yuan, J. (2025). Genetically Encoded Biosensors for Constrained Biological Functions in Probiotic Escherichia coli Nissle. ACS synthetic biology, 14(1), 296–303.
[4] Pinto, F., Thornton, E. L., & Wang, B. (2020). An expanded library of orthogonal split inteins enables modular multi-peptide assemblies. Nature communications, 11(1), 1529.
[5] Wang, B., Barahona, M., & Buck, M. (2014). Engineering modular and tunable genetic amplifiers for scaling transcriptional signals in cascaded gene networks. Nucleic acids research, 42(14), 9484–9492.
[6] Ding, D., Zhu, Y., Bai, D., et al. (2025). Monitoring and dynamically controlling glucose uptake rate and central metabolism. Nature Chemical Engineering, 2, 50–62.
[7] Ma, Y. P., Gu, J. W., & Ma, J. (2019). A GLP-1 mutant and its preparation method and use [Chinese Patent Publication No. CN 110256553 A]. National Intellectual Property Administration of the People's Republic of China (CNIPA).
[8] Chen, Y., Tang, J., Wang, L., Tian, Z., Cardenas, A., Fang, X., Chatterjee, A., & Xiao, H. (2020). Creation of Bacterial cells with 5-Hydroxytryptophan as a 21st Amino Acid Building Block. Chem, 6(10), 2717–2727.
[9] Chen, Y., Jin, S., Zhang, M., Hu, Y., Wu, K. L., Chung, A., Wang, S., Tian, Z., Wang, Y., Wolynes, P. G., & Xiao, H. (2022). Unleashing the potential of noncanonical amino acid biosynthesis to create cells with precision tyrosine sulfation. Nature communications, 13(1), 5434.
[10] Rottinghaus, A. G., Ferreiro, A., Fishbein, S. R. S., Dantas, G., & Moon, T. S. (2022). Genetically stable CRISPR-based kill switches for engineered microbes. Nature communications, 13(1), 672.
[11] Nanda, B., Bhowmick, J., Varadarajan, R., & Sarma, S. P. (2024). Backbone assignment of CcdB_G100T toxin from E.coli in complex with the toxin binding C-terminal domain of its cognate antitoxin CcdA. Biomolecular NMR assignments, 18(2), 285–292.
[12] Din, M. O., Danino, T., Prindle, A., Skalak, M., Selimkhanov, J., Allen, K., Julio, E., Atolia, E., Tsimring, L. S., Bhatia, S. N., & Hasty, J. (2016). Synchronized cycles of bacterial lysis for in vivo delivery. Nature, 536(7614), 81–85.
[13] You, L., Cox, R. S., 3rd, Weiss, R., & Arnold, F. H. (2004). Programmed population control by cell-cell communication and regulated killing. Nature, 428(6985), 868–871.
[14] Lee, S., Lewis, D. E. A., & Adhya, S. (2018). The Developmental Switch in Bacteriophage λ: A Critical Role of the Cro Protein. Journal of molecular biology, 430(1), 58–68.
[15] Richbourg, N. R., & Peppas, N. A. (2020). The swollen polymer network hypothesis: Quantitative models of hydrogel swelling, stiffness, and solute transport. Progress in Polymer Science, 105, 101243.
[16] Aleš Ručigaj & Tilen Kopač.(2025).Mathematical modeling of swelling and shrinking dynamics in pH-sensitive hydrogels composed of alginate, anionic cellulose, and chitosan.Polymer,328,128452-128452.
[17] Takahashi, C. N., Miller, A. W., Ekness, F., Dunham, M. J., & Klavins, E. (2015). A low cost, customizable turbidostat for use in synthetic circuit characterization. ACS synthetic biology, 4(1), 32–38.
[18] Wang, B., Barahona, M., & Buck, M. (2015). Amplification of small molecule-inducible gene expression via tuning of intracellular receptor densities. Nucleic acids research, 43(3), 1955–1964.