We aim to develop an engineered bacterium capable of normal growth and proliferation in a medium simulating seawater salinity (approximately 600 mM NaCl). Based on the in-depth research by Raj Kishor Kapardar^[1] et al., we confirmed from the metagenome of the pond that the salt-tolerant genes gspM and echM can enable bacteria to survive at 750mM sodium chloride concentration. They respectively express salt stress tolerance proteins and proteins with enyl-coA hydratase activity. We plan to introduce them into Escherichia coli to enable them to survive in seawater. For details, please refer to the Design page.

After confirming the sequence, we designed the [echM-gspM]-pET28a(+) plasmid, which can be used to construct salt-tolerant survival systems in Escherichia coli. The plasmid was successfully constructed, and the following figure shows detailed plasmid information.

Fig1. echM-gspM Genetic Circuit

We aim to quantitatively compare the growth curve differences between the engineered bacteria and wild-type strains introduced with the [echM-gspM]-pET28a(+) plasmid in high-salt (600mM NaCl) and salt-free media to verify the function of the salt-tolerant gene. We measured the OD600 absorbance values of each group of strains in different culture media at regular intervals within 0 to 27 hours and conducted fitting analysis using the Logistic model.

Our design observation shows that the engineered bacteria in 600 mM NaCl NaCl medium exhibited normal growth patterns comparable to those in NaCl-free LB medium. In contrast, growth of negative control E. coli was significantly inhibited under high-salt conditions, confirming that salt-tolerant genes enhance engineered bacteria's adaptability to high-salt environments. However, further analysis revealed slightly NaCl-free LB medium supported slightly better growth of engineered bacteria than 600 mM NaCl-containing LB medium, indicating residual growth inhibition by high-salt stress.

Fig2&3. Growth Curves of E.coli Strains with Logistic Model

This DBTL cycle verified that gspM and echM expression enables engineered bacteria to achieve near-normal growth in 600 mM NaCl (high-salt) environments, comparable to salt-free conditions. However, the K-value of their growth curve remained slightly lower than the salt-free control, indicating potential metabolic burden from high-salt osmotic pressure or exogenous protein expression. This "slight gap" discovery directly gave rise to the problems in our next round of engineering cycle. In the future, we will conduct metabolic flux analysis at the modeling level to determine whether the slight K-value reduction primarily stems from salinity-induced osmotic pressure or protein expression burden.

Although we already know that engineered bacteria can survive normally in a high-salt environment. However, the seawater environment is very complex. It not only contains sodium chloride but also a large amount of other minerals. To test the survival ability of engineered bacteria in natural seawater environments, we redesigned the experiment for the future:

1. The engineered bacteria were inoculated in naturally sterilized seawater to test their response ability to environmental stress in the natural environment.

2. Explore the size of the dominant ecological niche of engineered bacteria in the natural seawater environment.

Our design specification is: to achieve the controllable formation of engineered bacterial biofilms by regulating the expression of key proteins. Literature analysis identified Ag43 and OmpA as core regulators of biofilm formation: Ag43 specifically mediates cell aggregation, adhesion, and biofilm development; OmpA, a key determinant of bacterial colonization and biofilm establishment, significantly enhances biofilm formation on hydrophobic surfaces^[6][7]. The design employs lactose operon-mediated regulation of these exogenous genes. Optimization of IPTG induction concentration ensures sufficient target protein expression while minimizing host cytotoxicity. To streamline experimental workflow, a predictive model adapted from XJTLU-CHINA (2023) was utilized to estimate the optimal IPTG concentration range, facilitating experimental design and improving efficiency^[5].

We use ordinary differential equations (ODEs) to describe the mRNA expression processes of ag43 and ompA. These ODEs are numerically solved using Python. Given that ag43 and ompA are driven by the same operon and co-transcribed into a single polycistronic mRNA, their mRNA expression levels are collectively represented by a single variable, mRNA_AG43·ompA.

For details, please refer to the Model page.

We drew the prediction graphs of mRNA synthesis amounts for ag43 and ompA under IPTG gradient induction by solving the ODE equations using Python.

Fig4. Effect of IPTG Concentration on mRNA_AG43_ompA Expression

Our model prediction shows that the mRNA expression levels of ag43 and ompA tend to saturate when the IPTG concentration reaches approximately 400-600μM, with diminishing returns from further concentration increases.

This round of modeling provides guidance for the subsequent DBTL cycle: the IPTG concentration gradient in wet experiments will be focused on the 0.1–0.8 mM range for testing. This design avoids blind large-scale screening and conserves valuable time and resources. While the model cannot fully replicate real-world conditions, it has effectively delineated high-value regions for experimental exploration.

To actually construct a strong biofilm, we designed a plasmid that overexpresses Ag43-OmpA using a strong promoter and strong ribosome binding site (RBS)and used the data obtained from cycle 1 to induce expression at 400-600μM IPTG.

We designed recombinant plasmids containing ag43 and ompA genes to regulate gene overexpression by assembling strong promoters and strong RBS sequences. The plasmid was successfully constructed, and the following figure shows detailed plasmid information.

Fig5. ag43-ompA Genetic Circuit

The wet experiment verified the expression of the target protein by Western blot and detected the biofilm formation ability by crystal violet staining. According to Fig6, it was found that the biofilm formation ability of the engineered bacteria in the experimental group introduced with [ag43-ompA]-pET28a(+) plasmid was significantly enhanced compared with that of the control group introduced with pET28a(+) empty plasmid. During the experiment of inducing protein expression, we used the prediction graph of the optimal concentration of mRNA synthesis induced by IPTG in cycle 1 and conducted gradient attempts near the 400μM IPTG concentration. The final experiment found that the optimal induction concentration of IPTG was approximately 600 μM.

Fig6. The Western Blot(WB) result induced by different IPTG concentrations.

Fig6 & 7 & 8. Crystal Violet Staining Results of 24-hour Growth Biofilm

We successfully utilized strong promoters and strong RBS to regulate the overexpression of ag43 and ompA to achieve the purpose of enhancing biofilms, and ensured that the engineered bacteria could survive normally and function properly. However, Professor Quanfeng Liang from Shandong University mentioned that excessive expression of membrane proteins can impose an excessive burden on bacteria, and overexpression of OmpA poses a risk of bacterial self-lysis. Subsequently, we plan to seek better ways to reduce the expression burden of engineered bacteria and ensure the enhancement capacity of biofilms. We find bcsQ is another key gene to promote the formation of biofilms.

Considering Excessive expression of membrane proteins can impose an excessive burden on bacteria, and overexpression of OmpA poses a risk of bacterial self-lysis. We find bcsQ is another key gene to promote the formation of biofilms.

Exopolysaccharide (EPS) is an ideal abiotic component in biofilm by exerting the following functions:

1. Enhancing surface adhesion through interactions among functional groups.

2. Defensing against abiotic stressors, granting the ecological biofilm greater survival capabilities.

The method of enhancing E. coli biofilms by repairing the bcsQ mutation[14] works based on the following mechanism: In E. coli K-12 strains, the 6th codon of bcsQ is a premature termination codon (TAG), which truncates BcsQ translation early. This triggers "operon polarity," inhibiting the expression of downstream genes in the bcs operon (e.g., bcsA, bcsB) and thus blocking cellulose biosynthesis.

By prime editing repairing TAG to TTG (encoding leucine), functional BcsQ is produced, and the continuous ribosome binding during translation prevents Rho-dependent transcriptional termination, restoring the expression of downstream bcs genes.

This dual restoration of the bcs pathway promotes cellulose synthesis, which, together with curli, forms an elastic, cohesive matrix. This enhances biofilm structure (e.g., larger colonies with complex folds) and mechanical properties, thereby strengthening the biofilm. For a more comprehensive explanation, refer to the Design section.

Fig9. Prediction Effect of bcsQ Repair

Based on the guidance of senior researchers and professors as well as literature research, and in accordance with the principle of guided editing, the highly efficient site-directed mutagenesis binary plasmid system PE-[bcsQ] -PacyCDUet-1 plasmid was successfully constructed[15]. Compared with λ-Red and traditional CRISPR, guided editing has significant advantages: it does not require double-strand breaks in the genome, has no "scar" problem of λ-Red editing, and has higher targeting accuracy and editing efficiency, which can solve the problems of low efficiency of λ-Red and potential off-target effects of traditional CRISPR.

Fig10. Mechanism of Prime Editing

Under the guidance of senior researchers and professors, and through reviewing relevant literature, we successfully constructed two-in-one plasmid prime editing system for efficient site-directed mutagenesis based on the principle of prime editing. The evopreq1-1 trimmed motif was added to protect pegRNA from degradation[16].

Fig11. The Synthesis Process of PE-[bcsQ]-pACYCDuet-1 Plasmid

We have successfully made the purchase, but due to time constraints, the plasmid has not arrived yet. We look forward to achieving successful results before jamboree.

Our design specification is: to construct a "signal capture - spore anchoring system", enabling the engineered bacteria to effectively settle and anchor spores by diffusing and releasing specific functional molecules in the water body and forming a stable concentration gradient, thus preventing them from detaching from the biofilm. Acyl-Homoserine Lactone (AHL) is a signaling molecule used by bacteria for quorum sensing and is widely present near biological membranes. We chose it as our first functional molecule. According to reference ^[8], to induce irreversible spore deposition, the concentration of AHL must be at least 25μmol · L⁻¹. Based on this, the functional molecule release and gradient formation mechanism of the engineered bacteria is designed to achieve spore deposition through molecular action and anchor on the biofilm.

During the model-building process, we assume:

1. The diffusion coefficient of AHL can be calculated by the Einstein–Smolukhovsky equation.
   $$D = \frac{\mu_q k_B T}{q} \approx 4.65 \times 10^{-10} \text{ m}^2/\text{s}$$
2. Only consider two environments:
   1. Still water environment: Only consider diffusion.
   2. Marine environment: Consider convection and diffusion.

Diffusion models are mainly divided into the following two types:

1. In a still water environment, Fick's Second Law was used to simulate the variation of AHL concentration with time and space.
2. In the marine environment, the Convection–diffusion Equation was adopted and integrated with the seawater harmonic diffusion model to simulate the ocean flow velocity in both the horizontal and vertical directions.

For details, please refer to the Model page.

1. Simulation results of still water environment

Although AHL molecules theoretically continue to diffuse, due to their extremely low diffusion coefficient \( \sim 10^{-10} \, \text{m}^2/\text{s} \), their effective diffusion range is very limited. Even if the initial concentration of AHL is as high as 600 μM, it is difficult for AHL to spread beyond 5 cm with the injection point as the center within 24 hours. For details, please refer to the Fig13 below.

Fig13. AHL Simulation Results of Still Water Environment

2. Marine environment simulation results:

In the presence of convection, the spatial diffusion rate of AHL increases, but it is still limited by an extremely low generation rate.

The calculation results show that the diffusion rate of AHL is much higher than the physiological reasonable production rate of bacteria \( \sim 10^{-7} \, \text{m}^2/\text{s} \), making it difficult for its concentration to accumulate effectively. For details, please refer to the Fig14 below.

Fig14. AHL Simulation Results in the Presence of Convection

Even under convective conditions, the diffusion rate of AHL is far higher than its generation rate, failing to reach the minimum concentration required for spore sedimentation. This indicates that the AHL-based spore capture strategy is not feasible in real aquatic environments.

Thus, the scheme of using AHL for spore sedimentation was abandoned. Notably, an alternative substance capable of inducing Ulva prolifera spore sedimentation—arginine short peptide—was identified, which can accelerate the natural sedimentation process of spores^[4]. To avoid the recurrence of diffusion-related issues, the design is adjusted to immobilize the arginine short peptide onto Escherichia coli.

Under the modeling guidance of Cycle 1, we aimed to identify a substance that can be immobilized on engineered bacteria to induce premature spore sedimentation. Notably, T. Ederth demonstrated that arginine oligopeptides induce Enteromorpha margin spores to prematurely release adherent vesicle contents, shortening spore formation time from 2 hours to 5 minutes^[4]. Accordingly, we designed to fuse arginine oligopeptides with the csgA gene (derived from Aspergillus niger fimbriae) in Escherichia coli, enabling spores to anchor more firmly to the biofilm matrix.

Literature indicates that in the experimental system of Ederth et al., peptides with similar length and terminal exposure to Arg can significantly increase the adhesion rate of Ulva swimming spores (pseudo-colonization ratio >70%).

We fused short peptide sequences containing arginine into the outer membrane protein CsgA of Escherichia coli to construct a functional module capable of extracellular display and mediating spore deposition. AlphaFold3 simulations predicted the three-dimensional structure of the fusion protein, revealing stable extracellular exposure of the short peptide—this confirms the module's potential to endow host cells with spore deposition-inducing functionality. For the specific 3D structure, please refer to Fig15.

Fig15. CsgA-Arginine Oligopeptide 3D Structure

AlphaFold3 predictions show high confidence in the overall structure of our designed CsgA-Arg fusion protein. The β-sheet region of the CsgA backbone remains highly reliable, confirming the fusion peptide does not disrupt the core scaffold. However, as amino acid count increases, the arginine short peptide terminus exhibits lower prediction confidence (evidenced by lighter areas in the Fig16), indicating it is an intrinsically disordered region (IDR). This design avoids stable folding domains, allowing the short peptide to act as an "oscillating pendulum" to maximize contact between terminal arginines and external targets, such as Ulva spore membranes.

Fig16. CsgA-Arginine Oligopeptide PAE Picture

The results demonstrate that short peptides can be stably exposed outside the membrane, promising to endow host cells with a novel function of inducing spore deposition. With this structural basis verified, we proceeded to the next phase of Cycle 2.2, concentrating on the practical construction of arginine fusion short peptides.

We select the tetracycline operon as the main body of the suicide switch. We fused the LOVdeg tag into the C-terminal of the galactose repressor TetR, enabling TetR to be rapidly degraded under blue light irradiation and regulating the expression of the downstream suicide gene mazF. According to the literature^[13], we know that the C-terminal of this tag will expand and expose the degradation signal under the irradiation of blue light with a wavelength of 465nm, and then be recognized by the ClpXP protease system of Escherichia coli, thereby achieving light-dependent protein degradation^[11]. This new type of blue light-controlled suicide system will be more sensitive and controllable, and have stronger mobility. For details, please refer to the Model page.

1. No light

When light (about 465 nm light irradiation) is absent, the tetR repressor binds tightly to the operator, preventing transcription by RNA polymerase.



Fig32. The gene circuit has no transcription without light.

2. With light

When cells are under about 465 nm light irradiation, the LOVdeg tag promotes the degradation of the tetR repressor protein. This allows transcription by RNA polymerase.



Fig33. TetR-LOVdeg is degraded irradiated by blue light.

This initiates the downstream expression of the mazF toxin gene, triggering programmed cell death.



Fig34. The gene circuit finally initiates because of the removement of the repressor.

According to the content of the paper[11], we adopted eight dark state stability mutations in the iLoD system to enhance the effect of the light-controlled switch. We used AlphaFold3 to predict and verify the structural rationality of the fusion protein to ensure that the binding of TetR and LOVdeg does not affect the core function. For details, please refer to the following figure.

Fig35. TetR-LOV 3D Structure

Fig36. TetR-LOV PAE Picture

We used molecular docking and molecular dynamics simulations to verify the binding ability of fusion proteins and DNA operon regions. The following figures show part of the analysis results. For details, please refer to the Model page.

Fig37 &38. Partial Images of Molecular Docking and Molecular Dynamics Simulation

After comparative analysis of docking and simulation results for multiple protein-DNA groups, the Lucia fusion protein exhibited remarkably high confidence. This confirms that the LOVdeg can bind to the TetR protein effectively and stably, thereby providing theoretical basis and support for subsequent experiments.