Overview

We continued our literature search and found that adsorbing PET has the potential to further enhance degradation. Current research has utilized hydrophobins to enhance the adsorption effect. The methods described in the literature all involve the co-secretion or co-display on the surface with degradation enzymes^[1].

We believe that by using surface display technology combined with excreted degrading enzymes, we can also achieve an enhanced degradation effect. This is our unique innovation compared to other methods!

We employed the traditional yeast surface display technology - using the α-agglutinin of Saccharomyces cerevisiae - to construct a fusion protein that can adsorb PET and has the function of surface display technology. Saccharomyces cerevisiae constitutively expresses Aga1, which can bind to the Aga2 functional domain in the fusion protein and secrete it as cell display. Initially, we controlled this process through the high expression of the constitutive promoter PTEF. Additionally, we selected a flexible linker (GGGGS)n to connect the key protein domains, and we tested how to determine the length of the linker.

Fig 1. HFBI fusion protein gene map

Fig 2. HFBI fusion protein expressing process

We conducted a preliminary prediction on the Alphafold Server. We found that the length of Linker2 had little impact on the prediction of V5 epitope, while the length of Linker1 had a significant effect on the structure of the HFBI and Aga2 functional domains. Therefore, we chose to test with different Linker1 lengths: (GGGGS)3, (GGGGS)5, and (GGGGS)8. They were numbered as 3_1, 5_1, and 8_1 respectively.

Fig 3. HFBI fusion proteins with different linker lengths

We conducted transmembrane domain and hydrophobicity score analyses on the three fusion protein structures. The results initially demonstrated that the fusion proteins could be successfully displayed on the surface and that the hydrophobicity of the HFBI functional domain was good.

Fig 4. HFBI fusion protein transmembrane and hydrophobicity analysis

Based on this, we compared the structures of the active domains to preliminarily verify the structural fidelity. We used two key monomeric proteins and fusion proteins for docking. The smaller the RMSD result obtained, the better.

Fig 5. The structure of key functional domains of fusion proteins

Fig 6. Verification of the structural fidelity of fusion proteins

Next, we conducted a docking between the fusion protein and PET. The binding energies obtained were all within the expected range. The larger the absolute value of the binding energy, the better the affinity.

Fig 7. Fusion protein and PET docking site and binding energy

After an initial analysis, we concluded that all three designs were feasible. To highlight the differences, we chose 3_1 and 8_1 for constructing plasmids for the experiments.

Before conducting the route design, we had already taken into account the impact of the metabolic toxicity of the chassis bacteria. We plan to reduce the metabolic toxicity by adjusting different control conditions of the engineering route and minimizing the constitutive expression.

Is there a dynamic regulatory system that can control the expression of the adsorption module?

In the experiment, we learned about the "Pgal1+Δgal80" strategy. Under high glucose conditions, the expression of downstream genes controlled by this dynamic regulatory system is inhibited, and at this time, yeast energy is used for growth. While under low glucose conditions, the downstream genes are induced to express, and this process has a temporal sequence^[2].Furthermore, we discovered that this dynamic regulatory system forms an orthogonal relationship with the PHXT-1 promoter of our degradation module. PHXT-1 responds to feeding signals and induces the expression of downstream genes under high glucose conditions.

Fig 8. Pgal1/Δgal80 dynamic regulation system

So, based on our imagination, we constructed the following plasmid transformation yeast, and conducted an adsorption verification experiment after co-fermentation with microplastics. In both immunofluorescence and transmission electron microscopy, the desired results were observed.

Fig 9. Pgal1/Δgal80 dynamic regulation system

Since the Saccharomyces cerevisiae itself can express Aga1, but its basal expression level may not be high, resulting in poor surface display effect, we decided to enhance the expression by co-integrating Aga1 into a plasmid to improve the expected surface display effect. We used the Pgal1,10 bidirectional promoter to control Aga1 and the fusion protein part separately. Moreover, in the laboratory, we successfully obtained the Aga1 fragment in the genome through PCR, linearized the plasmid containing the HFBI-Aga2 fusion protein,and will proceed with a further cloning step to enhance the expression of Aga1.

In addition to the optimization strategy of enhancing Aga1 expression by using the Pgal1,10 bidirectional promoter, we also learned that the surface display system applied to Saccharomyces cerevisiae is very diverse. Although only α-agglutinin has been applied to HFBI, we still want to enrich the application of the surface display system in the future.

Although we have achieved good surface display results using α-agglutinin, can we achieve even better results through other surface display systems, and can the process be presented more concisely?

Our Aga1 is not the entire protein that is used for attaching to the surface of the yeast cell membrane, as shown in our schematic diagram. The amino acids from 150 to 725 act as the anchor protein. We hypothesize that directly displaying HFBI through the anchor protein might be a better approach. Therefore, we designed and constructed a new surface display system.

Fig 10. The surface display system of utilizing GPI

We conducted a literature search to preliminarily explore the feasibility. We found that this surface display method was more effective than α-agglutinin for other target proteins. The article also compared various types of GPI and demonstrated the great potential of this display method.

Fig 11. The surface display system of utilizing GPI^[3]

The results in the literature have greatly boosted our confidence for future experiments. Additionally, the literature also proposed a method applicable to the surface display system. We plan to conduct preliminary dry experiments to narrow down the range of GPI types first. At the same time, we will use online software such as Signalp to screen different signal peptides. Finally, we will construct plasmids for experiments, thus forming another engineering cycle.

Fig 12. Surface display of the system's signal peptide selection^[4]

After optimizing the adsorption capacity of the engineered yeast, we encountered a core challenge: how to ensure that it can function stably and persistently in the complex intestinal environment. Through literature research, we discovered that microorganisms often form multi-species biofilms through co-aggregation in nature to enhance their survival advantages. This strategy not only enhances their resistance to environmental stress but also facilitates colonization in specific ecological niches (such as intestinal mucosa).

We envision that an ideal symbiotic partner for engineered yeast can be identified. This strain would be able to aggregate with the yeast to enhance intestinal colonization, and also work in tandem with us to accomplish the task of adsorbing microplastics.

By reviewing the literature, we focused our attention on Lactobacillus bacteria. A large number of studies have confirmed that specific Lactobacillus strains can form co-aggregation with Saccharomyces cerevisiae. Among them, the mannose-specific adhesin (Msa) on the surface of Lactiplantibacillus plantarum can specifically bind to the mannoprotein on the yeast cell wall, forming a stable physical connection.

Fig 13. Synergistic adsorption in a symbiotic system

The adhesive ability of Lactobacillus plantarum to intestinal mucosal cells is a key aspect in exerting its probiotic effects. Probiotics attach to the intestinal mucosal cells, thereby preventing the penetration of harmful microorganisms and toxins, and play an important role in enhancing the intestinal barrier function. Moreover, they can compete with potential pathogens for nutrients and living space, disrupt the biofilm of pathogenic bacteria, and hinder the colonization of harmful bacteria^[5].Researchers identified a collagen-binding protein from Lactobacillus plantarum. This protein not only promotes adhesion but also inhibits the adhesion of Escherichia coli O157:H7 to extracellular matrix components^[6]. Furthermore, through adhesion and interaction with intestinal mucosal cells, probiotics help maintain the immune balance within the intestinal tract^[7]. The elongation factor Tu is a cell surface-related protein that has been shown to interact with fibronectin on intestinal epithelial cells. It may play a role in intestinal immunity and homeostasis by inhibiting cell apoptosis^[8].Lactobacillus plantarum can also enhance its adhesion ability to the intestinal mucosa by sensing unknown intestinal signals and relieving the inhibition of transcription factor MbpR on the adhesion protein genes.

Fig 14. The key regulatory mechanism for achieving intestinal adhesion

More importantly, we discovered a strain named Lactobacillus plantarum DT88. This strain not only has the potential to aggregate with yeast, but also demonstrates a strong ability to adsorb microplastics on its own^[9]. Therefore, we constructed a dual-bacterial symbiotic system consisting of engineered Saccharomyces cerevisiae and Lactobacillus plantarum DT88, aiming to leverage the synergistic effects of the two organisms to achieve a "1 + 1 > 2" colonization and adsorption outcome.

After envisioning the dual-bacteria symbiotic system, we conducted further literature research to assess the potential shortcomings of this strategy. We found that although co-culture theoretically has advantages, it faces challenges in practical applications:
Internal stability and competitive relationship: Under ideal laboratory conditions, yeast and lactobacillus can exhibit mutualistic symbiosis. However, in the dynamic and complex intestinal microenvironment, the relationship between the two may become unstable and even turn into a competitive relationship, causing one of the organisms to be eliminated, thereby rendering the entire system ineffective.

In response to these challenges, we once again turned to the literature to seek innovative solutions to optimize our symbiotic system:
Utilize "postbiotics": To avoid the instability risk of live bacteria in the body, a more advanced strategy is to use "postbiotics" or "paraprobiotics", which involve the use of inactivated probiotic cells or their metabolites. Studies have shown that even inactivated bacteria or their cellular components retain most of their biological activity, such as immune regulation and adsorption^[10].This method not only ensures the stability and safety of the product, but also avoids the uncertainty of viable bacteria colonization. It provides another highly promising optimization path for our project.

Initially, we assumed that microplastics mainly caused physical damage. However, through research and preliminary experiments, it was discovered that oxidative stress and inflammatory responses were the key pathogenic mechanisms. This refuted our initial assumption of only considering physical damage. Therefore, we turned to intervention at the molecular mechanism level. We investigated the most common harmful diseases caused by microplastics in the human intestinal tract. Currently, PET has not found any diseases that are significantly related to specific diseases. Among them, inflammatory bowel disease (IBD) has the most active research on the association with microplastic exposure and has been proven to have a certain correlation. Additionally, microplastics can also cause dysregulation of the intestinal microbiota and damage to the intestinal barrier function^[1-3].

In addition to literature review, we aim to identify the targets of microplastics causing intestinal diseases through bioinformatics methods for the purpose of designing treatments. We plan to start with IBD, a common disease.

The students in the modeling group came up with specific plans for target identification:
a. Data collection and intersection with targets.
b. Construction of protein networks and screening of core nodes.
c. Multi-dimensional functional enrichment and verification.
d. Final determination through machine learning algorithms.
The specific methods are presented in the Test section.

For details, see the Model section.

Firstly, the researchers established two independent gene databases. One was for PET microplastics, where potential targets that might interact with it were comprehensively collected from multiple databases such as PubChem, CHEMBL, and STITCH based on the SMILES chemical encoding of its main degradation products. The other was for inflammatory bowel disease (IBD), where a group of genes with the highest association with IBD were selected from disease databases such as GeneCards and OMIM. By taking the intersection of these two large databases, the "common targets" that were related to both PET microplastics and IBD were finally identified. This step ensured that the targets for the subsequent analysis had dual correlations.

Fig 1. The search for cross-targets between PET and IBD

Just having a list of genes is not enough, because genes usually work together through complex networks. Researchers used the STRING database to construct a protein-protein interaction (PPI) network based on the common targets from the previous step. Subsequently, this network was imported into the Cytoscape software for visual analysis.
The researchers employed two complementary algorithms to identify the key genes in the network:
a. CytoHubba_MCC algorithm: This is used to find the nodes with the most connections and located in the "central hub" position. This method can directly identify the proteins that play a key role in information transmission.
b. MCODE plugin: This is used to discover the most densely connected "functional modules" or protein clusters in the network. This helps to identify the protein groups that collaborate to perform specific functions in the cell.

Fig 2. Protein interaction network results

To understand the specific biological functions of these core genes, the researchers conducted Gene Ontology (GO) and KEGG pathway enrichment analyses. The results showed that these genes were highly enriched in biological processes such as "inflammatory response" and "immune response", which are highly relevant to the pathology of IBD, and were closely related to important signaling pathways such as PI3K-Akt and Jak-STAT.

Fig 3. Metabolic pathway analysis

To verify whether PET can directly act on these targets at the physical level, we conducted molecular docking and molecular dynamics simulation tests. The results showed that the binding energies of PET to proteins such as JAK2, IL2, and TGFBR2 were negative, which proved in physics that there is a stable and spontaneous binding between them.

Fig 4. PET and three target molecules docking

After going through the above multiple rounds of screening and verification, a series of machine learning algorithms (including LASSO regression, SVM-RFE, boruta and XGBoost) were finally employed for the final and most rigorous screening. These algorithms further evaluated the importance of each gene from a statistical perspective, and ultimately precisely identified the three core targets: JAK2, IL2 and TGFBR2.

Fig 5. Machine learning screening

JAK2: It belongs to non-receptor tyrosine kinases. The accumulation of ROS can directly activate JAK2, making it an important upstream trigger point in the inflammatory cascade.
IL2: T-cell growth factor, over-activation is associated with immune dysregulation and abnormal T-cell proliferation in IBD.
TGFBR2: A key receptor of the TGF-β signaling pathway, regulating cell proliferation, differentiation, and tissue repair. Its abnormality may lead to failure of intestinal epithelial repair or a tendency towards fibrosis.

We have identified three therapeutic targets for the harm caused by microplastics to the human body. In the future, when we are searching for therapeutic components, we will pay particular attention to their relationship with the target sites.

Firstly, we address the issue from the first aspect of microplastics causing disease - oxidative stress. Research indicates that reactive oxygen species (ROS) play a significant role in the pathogenesis of inflammatory bowel disease. Superoxide anion (O₂⁻) and hydrogen peroxide (H₂O₂) are two core types of ROS. The accumulation of ROS can directly activate the JAK2 pathway and trigger an inflammatory cascade reaction. Therefore, eliminating ROS is a key strategy for inhibiting inflammation induced by microplastics.

Can we establish a "cohesive" antioxidant regulatory system to rapidly eliminate ROS at the source and achieve antioxidant stress response? Does the molecular mechanism layer have any effect on the targets we are exploring?

There are two key enzymes that play a dominant role in the removal of ROS:
a. SOD1: Converts O₂⁻ into H₂O₂;
b. CTT1: Further decomposes H₂O₂ into water and oxygen.
Theoretically, the cascade reaction formed by these two enzymes can efficiently remove ROS.

Fig 6. The function of the antioxidant module

After reviewing the literature, SOD1 and CTT1 were found to be capable of performing their intended functions in our chassis - Saccharomyces cerevisiae. Therefore, we selected these two genes as the effect modules and designed a dual-enzyme expression pathway to enable the engineered bacteria to possess systematic antioxidant capabilities.
In terms of the molecular mechanism, this antioxidant component directly eliminates reactive oxygen species (ROS) to block the activation of JAK2 at the source, achieving direct intervention in the downstream inflammation^[22][23].

The team realized that the antioxidant module can precisely address the "downstream trigger point" of ROS → JAK2, but the inflammatory network is complex, and relying on a single module is not sufficient to completely inhibit the pathological process. Oxidative stress is merely a "trigger", not the sole effector. Once the inflammatory pathway is activated, merely removing the trigger factor is not enough to stop the pathological progression. This discovery prompted the project to shift to a more comprehensive treatment approach - in addition to the antioxidant module, functional modules such as immune regulation and tissue repair need to be added.
Furthermore, model analysis and literature comparison indicate that this system can indeed significantly clear ROS and directly block the activation of JAK2 in theory. However, there is a fundamental defect: it can only eliminate the initial oxidative signal, but cannot inhibit the downstream inflammatory cascade reaction triggered by ROS. This deserves our team to continue to think about optimization strategies.

The team first realized that direct intervention in the inflammation itself was necessary to solve the problem. Thus, they embarked on in-depth research on the core signaling pathways of the host immune system (such as JAK2, IL-2, and TGFBR2). By analyzing the mechanisms of these pathways, the team realized that they were key nodes regulating the immune system, and theoretically, the inflammation could be controlled through external factors.

Based on this understanding, the team proposed an ambitious idea:
Utilizing the synthetic biology pathways of engineered bacteria, to secrete inhibitors or agonists, to "remote control" the functions of the host immune cells from the outside, directly acting on the JAK2, IL-2 and TGFBR2 pathways, in order to achieve precise control of inflammation.

During the specific design and analysis phase, the team decisively rejected this approach. The reason was as follows: These host immune pathways are extremely complex and mainly function within the host immune cells; engineered microorganisms (such as yeast) are difficult to stably, safely and precisely directly regulate them; directly interfering with key kinases such as JAK2 might cause unforeseeable off-target effects on systemic functions such as hematopoiesis and immune surveillance; and there are significant risks in terms of biological safety.
Therefore, the team shifted its focus from "Is it technically possible?" to "Is it functionally safe and controllable?". This strategic retreat avoided the high-risk dead end. Subsequently, the team shifted to designing new solutions, seeking molecules that could the organization's repair and stabilization recovery. Eventually, Trefoil Factor 3 (TFF3) was identified as the ideal effector factor.

The team further analyzed the mechanism of action of TFF3:
Promoting epithelial repair: Activating the JAK2/STAT3 pathway in intestinal epithelial cells in a local manner, promoting mucosal healing and recovery of barrier function.
Actively inhibiting inflammation: Reducing key pro-inflammatory cytokines (such as TNF-α, IL-1β) in the inflammatory microenvironment, and weakening the excessive activation of IL-2 and TGFBR2-related signals.
This dual function ensures both local and controllable efficacy, meeting safety requirements^[6].

Through this cycle, the team learned that:
Direct intervention in the core host immune pathways is an unbridgeable red line because it is unpredictable and high-risk;
Strategically giving up high-risk solutions and instead adopting more indirect and safer mechanisms reflects engineering wisdom and a high sense of responsibility for biological safety;
Ultimately, the team established the "antioxidation + mucosal repair" strategy to replace the direct blocking of inflammation, thereby achieving the anti-inflammatory goal.

In the inflammatory environment of the human intestinal tract, a large amount of reactive oxygen species (ROS) accumulate, driving epithelial damage and inflammatory cascade reactions. We need a regulatory element that can precisely sense ROS signals as a molecular switch for the therapeutic pathway, so as to achieve the expression of conditionally induced anti-inflammatory factors.

The literature indicates that the transcription factors Yap1 and Skn7 in Saccharomyces cerevisiae are key regulatory elements for sensing ROS:
Yap1: Highly sensitive to ROS such as hydrogen peroxide, and can serve as the main sensing module;
Skn7: Provides an auxiliary function, and can enhance the stability and intensity of the response;Synergistic effect: When Yap1 and Skn7 synergistically bind to the promoter, they can significantly increase the gene expression level and accelerate the response speed.
Therefore, we hypothesize that the natural ROS-responsive promoter regulated by the synergistic control of Yap1 and Skn7 can be used as a candidate regulatory tool.

Based on the regulatory mechanism of Yap1/Skn7, we have identified three types of naturally existing promoters in yeast:
pGSH1 (glutathione synthetase promoter): contains multiple Yap1 binding sites;
pTRX2 (thioredoxin reductase promoter): contains Yap1 sites and Skn7 elements;
pGPX2 (glutathione peroxidase promoter): contains Yap1 sites and Skn7 elements, and can simultaneously sense ROS and inflammation-related oxidative signals.
These candidate promoters can all respond to oxidative stress signals, but there are differences in their activation intensity, response speed, and background leakage.

We conducted a literature review and comparative analysis of the three types of promoters^[25]：

Table 1. Oxidative stress promoter selection

After comprehensive analysis, pGPX2 stood out:
Low background expression → Avoid energy waste and side effects during non-inflammatory states;
High induction intensity → Capable of rapidly producing sufficient therapeutic proteins during inflammatory outbreaks;
Fast response → Exhibits a faster response speed to ROS and inflammatory signals compared to other promoters;
Broad response spectrum → Covers ROS and a wider range of inflammatory-related oxidative signals.
Therefore, the pGPX2 promoter has been confirmed as the optimal ROS-responsive regulatory tool, constituting the core component of our treatment line design.
Based on the literature, we found the pGPX2 promoter from the yeast genome, located upstream of the GPX2 CDS fragment (-709/+7).

Fig 7. GPX2 promoter gene map

We confirmed that pGPX2 is the optimal ROS-responsive promoter, capable of achieving low background, high intensity, and rapid response characteristics in an inflammatory environment, providing a reliable "molecular switch" for the therapeutic pathway. However, in the subsequent design process, when we attempted to place SOD1, CTT1, and TFF3 under three separate pGPX2 promoters respectively, new challenges emerged:
Metabolic burden: Multiple parallel pathways would significantly increase energy consumption;
Expression incoordination: Even relying on the same promoters, the dynamics of the three pathways are difficult to be completely consistent;
Uncontrollable dose ratio: The production of the three proteins is difficult to maintain a stable stoichiometric ratio, thereby weakening the synergistic effect.
Therefore, the team realized that the selection of a single high-quality promoter is only the first step. The more crucial issue lies in how to achieve safe, efficient, and coordinated expression of multiple proteins within the same pathway. This thinking directly led to the next round of iteration (Cycle 5): how to achieve the joint control of the anti-inflammatory and antioxidant modules.

After identifying the final three effect proteins, SOD1, CTT1 and TFF3, the team needed to address a key issue:
How to rationally express these proteins in the engineered bacteria to maximize their synergistic effect?
The research found that if each protein was driven by an independent expression pathway, it would cause excessive metabolic burden; if the expression was not synchronized, it would be difficult to ensure the precise stoichiometric ratio; and this would further weaken the synergistic therapeutic effect of the three proteins.

Based on this, the team envisioned giving up the "parallel scheme" of multiple independent lines and instead attempting a more efficient and integrated approach:
Integrate multiple effector proteins into a single line to ensure their synchronous expression, thereby enhancing the robustness and safety of the system.

The team finally constructed a single-line multi-module design: GPX2 → NCW2-SOD1 → P2A → NCW2-TFF3 → T2A → NCW2-CTT1.
The key design points include:
Promoter (GPX2): Only drives expression when inflammatory signals are present, ensuring the synchrony and spatiotemporal consistency of the three proteins.
2A self-cleavage peptide (P2A, T2A): Initially, an attempt was made to fuse the three proteins through a flexible linker, but analysis indicated that it would disrupt their three-dimensional structure and lead to functional loss. The team conducted literature research and adopted 2A self-cleavage peptides, enabling a single mRNA to generate three functionally independent proteins in a 1:1:1 ratio during translation, maintaining structural integrity while reducing metabolic burden.

Fig 8. The mechanism of action of 2A peptide^[8]

Secretion signal peptide (NCW2): Add the NCW2 sequence before each protein, directing the product to be secreted outside the cell and ensuring its release as needed in the intestinal lumen^[9].
By combining various elements, we have developed our treatment plan.

Fig 9. Treatment Module Diagram

In the scheme simulation and literature comparison, the team verified that the single promoter + 2A peptide strategy is feasible and has been widely applied in multi-protein systems; the 2A peptide can ensure a stable stoichiometric ratio and avoid functional loss caused by the fusion protein; the application of the secretion signal peptide NCW2 has been well reported in eukaryotic expression systems, and can reliably guide protein secretion.

Fig 10. Comparison of signal peptide efficiency^[9]

Influences on heterologous protein secretion by accessory proteins linked to SPs with different properties. a) α-amylase secretion titer in the ref, △SPC1 and ΔSEC72 strains when using different SPs. The α-factor SP was used as a reference to compared with other 11 super-secreted SPs. b) Hydrophobicity of SPs determined α-amylase secretion change tendency (increase or decrease) in the ΔSEC72 strain. The secretion change was calculated as follow: (amylase secretion by the ΔSEC72 strain – amylase secretion by the ref strain)/(amylase secretion by the ref strain) × 100%. was used as the control strain. Data shown are mean values ±SDs of duplicates^[27-28].

After the information was confirmed, we began the wet experiment work, which was divided into two parts: the construction of ΔSEC72 Saccharomyces cerevisiae and the construction and introduction of therapeutic genes. Preliminary laboratory results have been obtained at present.

Fig 11. Treatment module-related gene map

Through this round of cycle, the team realized that a single-line multi-module design can significantly reduce genetic complexity and metabolic burden; using 2A self-cleaving peptides is the most elegant way to solve the problem of multi-protein co-expression, which is both efficient and reliable; in the iGEM project, engineering wisdom lies not only in "getting the proteins expressed", but also in how to express them safely, efficiently and coordinately to maximize the therapeutic synergy.

After modeling and testing the engineered yeast with the integrated SOD1–TFF3–CTT1 gene circuit, we found that the performance of the single yeast treatment module still had limitations: although it had significant effects on ROS clearance (targeting the upstream signal of JAK2) and intestinal barrier repair (affecting the sustained activation of IL2), it was still insufficient in terms of the breadth of immune regulation and long-term stability maintenance. Moreover, from the literature comparison, it was also clear that there was a knowledge gap: this module had no known direct connection with the TGFBR2 signaling pathway.

In response to this performance gap, the team did not choose to continue stacking more complex circuits in yeast. Instead, they proposed a new idea: introducing a naturally immune-regulating partner that can complement the yeast; expanding the engineering goal from a single cell to a multi-microbial system, achieving a transition from molecular engineering to micro-ecosystem engineering.

The design concept is to combine the precise elimination ability of engineered yeast with the broad-spectrum regulatory ability of probiotics to achieve coordinated intervention at both the upstream and downstream of the inflammatory signals.
a. The molecular circuitry of the yeast treatment module (precise elimination and repair)
The core of the engineered yeast is the condition-inducible therapeutic pathway, as described above: pGPX2 → NCW2-SOD1 → P2A → NCW2-TFF3 → T2A → NCW2-CTT1.
b. The natural immune regulation advantages of DT88 (broad-spectrum regulation and signal weakening)
The Lactobacillus plantarum DT88, as an auxiliary strain, compensates for the deficiencies of the yeast module at the ecological and signaling levels through its natural metabolic and cytokine regulatory capabilities:
Inhibiting pro-inflammatory factors:DT88 can significantly reduce the elevated levels of pro-inflammatory cytokines TNF-α, IL-6, and IL-1β in the microplastic-exposed mouse model, while simultaneously synergistically enhancing the production of anti-inflammatory factor IL-10.
Inhibiting co-stimulatory molecules: The symbiotic system has been proven to be able to reduce the expression of activation markers CD86 on the surface of monocytes.
Metabolic products anti-inflammatory: The short-chain fatty acids (SCFAs) produced by DT88, especially butyric acid, can inhibit the NF-κB signaling pathway in immune cells and epithelial cells, reducing the transcription of inflammatory factors at the nuclear level.

The objective of this section is to combine the evidence from the literature and the modeling results to verify whether the dual intervention of the symbiotic system can precisely act on the core target, and to focus on evaluating the micro-regulatory effect of DT88.
a. Immunomodulatory effects:
The co-culture system of Saccharomyces cerevisiae and Lactobacillus exhibits unique immunomodulatory properties, and its anti-inflammatory effect is significantly superior to that of a single strain^[10]. This synergy is not simply additive; rather, it stems from the host immune system's specific recognition and response to the combined patterns of microbial molecules (MAMPs). In mice exposed to microplastics, DT88 can restore elevated levels of TNF-α, IL-6, and IL-1β to normal, and promote an increase in IL-10, thereby guiding the immune response towards a tolerative state^[11].
b. Inhibition of immune cell activation:
This system can significantly reduce the expression of activation markers CD14 and CD86 on the surface of monocytes, demonstrating its stable effect in inhibiting excessive inflammatory signals^[10].
c. Contribution to metabolic homeostasis:
Microbial metabolites, short-chain fatty acids (SCFAs), are the main energy source for colonic epithelial cells, promoting their proliferation and repair. At the same time, SCFAs can inhibit the NF-κB pathway, reduce the release of inflammatory factors, and stimulate goblet cells to secrete mucin to strengthen the intestinal mucus barrier^[12].
d. Molecular targeting mechanism:
JAK2 pathway (upstream weakening): DT88 reduces the levels of TNF-α and IL-6, thereby indirectly inhibiting the activation of the JAK/STAT pathway and synergistically promoting the production of IL-10.
IL-2 pathway (co-stimulation inhibition): This system can down-regulate the expression of CD86 on monocytes, which is equivalent to adding a "molecular brake" in the T cell activation circuit, alleviating the excessive immune activation induced by microplastics.
TGFBR2 pathway: Current research shows that there is no known direct molecular association between this system and TGFBR2.

Table 2. The effect of the symbiotic system on the target

In conclusion, the therapeutic effect of this symbiotic system is multi-level and three-dimensional. It has established a "deep defense" strategy. Additionally, we later discovered that the symbiotic system has a reinforcing effect on the barrier structure: The symbiotic system achieves structural reinforcement by regulating tight junction proteins (TJPs). Exposure to microplastics can damage the integrity of the intestinal barrier. Probiotics interact with epithelial cells and upregulate the expression of "sealing-type" proteins such as ZO-1, Occludin, and Claudin-1, reducing the permeability of the paracellular pathway and preventing harmful substances from leaking. Thus, this module has initially solved the problem of the harm caused by microplastics to the human body. In the next step, we need more experiments for verification^[32].
In conclusion, the therapeutic effect of this symbiotic system is multi-level and three-dimensional. It has established a "deep defense" strategy. Additionally, we later discovered that the symbiotic system has a reinforcing effect on the barrier structure: The symbiotic system achieves structural reinforcement by regulating tight junction proteins (TJPs). Exposure to microplastics can damage the integrity of the intestinal barrier. Probiotics interact with epithelial cells and upregulate the expression of "sealing-type" proteins such as ZO-1, Occludin, and Claudin-1, reducing the permeability of the paracellular pathway and preventing harmful substances from leaking. Thus, this module has initially solved the problem of the harm caused by microplastics to the human body. In the next step, we need more experiments for verification^[13].

The chassis bacteria must meet the following conditions in order to exert therapeutic effects in the intestinal tract:
a. Safety and modifiability: They must be safe for human use and have mature modification tools;
b. Ability to colonize the intestine: They must be able to survive in the complex intestinal environment for a long time;
c. Expression ability: They must ensure the effective operation of the synthetic biological circuit;
d. Natural therapeutic properties (plus points): such as anti-inflammatory and adsorption capabilities.

Based on this, the team compared several candidate strains:
a. E. coli (K-12, Nissle 1917): The tools are mature, but the safety and colonization ability are insufficient;
b. Bifidobacteria, Bacteroides, Clostridium: They have colonization or metabolic advantages, but the tools are not mature or their safety is questionable;
c. Lactobacillus: Food-grade probiotics, resistant to acid and bile, able to survive for a long time, but the tools and expression efficiency are limited;
d. Saccharomyces cerevisiae: GRAS certified, mature tools, capable of secreting complex proteins, but limited colonization.

Table 3. Comparison of adaptability of candidate chassis strains

Conclusion: No single strain can fully meet the requirements. However, E. coli Nissle 1917 and Saccharomyces cerevisiae demonstrate excellent safety and engineering capabilities, while Lactobacillus plantarum has unique anti-inflammatory and adsorption advantages.

After a preliminary comparison, we realized that no single strain could simultaneously meet all the requirements of safety, engineering feasibility, intestinal colonization ability and expression efficiency. This led us to propose a new idea - to construct a "symbiotic system", hoping to select two food-grade strains to make up for their respective deficiencies.

We hypothesized that the symbiotic system might be a more ideal choice than a single chassis strain, as it not only enhances the overall safety and stability of the system, but also takes advantage of the intestinal therapeutic properties of Lactobacillus plantarum itself.

During further literature research, we discovered a unique strain of Lactobacillus plantarum, L. plantarum DT88. Previous studies have shown that it can survive in the intestinal tract for an extended period and has the ability to adsorb microplastics, which perfectly aligns with the goals of our project^[1].
By reviewing the relevant literature on symbiotic systems, we found that yeast and Lactobacillus plantarum often coexist and mutually benefit, exchanging nutrients to promote each other's growth and enhance the stability of the community^[2].However, in most environments, Escherichia coli is inhibited by Lactobacillus plantarum - because it secretes organic acids, lowers the pH, competes for resources and space - making it difficult for it to maintain stability and functionality as a symbiotic system^[3].
So, our team members immediately agreed and we designed a system of division of labor and cooperation:
Saccharomyces cerevisiae: The main base strain, responsible for logic gate control, multi-enzyme co-secretion and treatment modules;
Lactobacillus DT88: The auxiliary strain, utilizing its natural microplastic adsorption ability and probiotic characteristics to enhance overall safety and application potential.
Eventually, we selected the "yeast + DT88" symbiotic consortium as the engineering bacterial chassis.

We verified through literature and existing studies: 1. Lactobacillus DT88 demonstrated excellent microplastic adsorption ability in vitro and could survive stably in the intestinal tract.

Fig 1. Scanning electron microscopy (SEM) images of bacteria aggregated with microplastics^[1]

a. Saccharomyces cerevisiae has a long-term application history in the food and pharmaceutical industries (GRAS certification), and can efficiently secrete complex proteins.
b. The two can support each other under a symbiotic condition, enhancing the stability and functionality of the community. The rationality of the symbiotic system has been further confirmed at this stage.

Each single strain has its own drawbacks: it cannot simultaneously achieve safety, tool maturity or functionality; the symbiotic system, which combines safety, functionality and scalability, better meets the application requirements of intestinal treatment.
At this stage, the team realized that the transition from single-bacterial engineering to symbiotic system engineering is the key to improving the performance and safety of the project. In the future, we will further explore the functional potential of the symbiotic system to make it a more powerful and stable treatment platform.

In the first safety cycle, we identified the limitations of a single strain in terms of safety, functionality, and intestinal colonization ability. Although engineering modifications can confer strong functions to the strain, they also bring potential biosafety risks, such as stability and controllability issues in a complex intestinal environment. Therefore, we studied how to enhance overall safety through systematic design. Literature indicates that introducing an unmodified probiotic at the GRAS (Generally Recognized as Safe) level as a companion not only utilizes its natural probiotic properties but also enhances the stability and predictability of the entire system through a mutually beneficial symbiotic relationship.

We envision that the symbiotic system itself can serve as a "security barrier", using natural probiotics to "support" and "restrict" the engineered yeast, thereby enhancing the biological safety of the entire product. This is the original intention behind our construction of the symbiotic system.

We designed and constructed a symbiotic system consisting of engineered Saccharomyces cerevisiae (GRAS-certified) and the natural probiotic Lactobacillus plantarum DT88. We conducted an in-depth analysis of the advantages in the previous strain selection process. The core safety logic of this design is "responsibility separation":
a. Function and modification separation: We concentrated the complex genetic circuits (such as adsorption, anti-inflammatory modules) in the modified Saccharomyces cerevisiae, while Lactobacillus plantarum DT88 was completely retained in its wild type and underwent no genetic modification. This enabled most risks to be concentrated in a chassis.
b. Utilizing natural probiotic properties: Lactobacillus plantarum DT88 is itself a verified probiotic with the abilities to resist inflammation, regulate the intestinal microbiota, and adsorb microplastics. Its addition not only does not introduce new safety risks but also adds additional therapeutic benefits and safety guarantees to the system.
c. Enhancing system stability: The two organisms form a more stable micro-ecosystem through nutritional complementation and stress sharing, reducing the risk of the engineered bacteria losing control or failing in the intestinal tract.

Fig 2. Detailed Description of Symbiotic Systems

The symbiotic system composed of Saccharomyces cerevisiae and Lactobacillus plantarum can establish metabolic interdependence and mutual provision of nutrients. In an environment rich in nitrogen sources such as the intestinal tract, Saccharomyces cerevisiae will, through an active metabolic strategy regulated by the TORC1 signaling pathway, secrete excessive amino acids. The amino acids secreted by the yeast provide essential growth factors for many amino acid-deficient lactic acid bacteria (such as Lactobacillus plantarum). This targeted nutrient supply significantly promotes the growth and survival of Lactobacillus plantarum in the co-culture system, with its biomass being much higher than that in a single culture condition. Although the main nutrient flow is from yeast to bacteria, this relationship can develop into mutualism. For example, lactic acid bacteria can hydrolyze some complex carbohydrates that yeast cannot directly utilize (such as lactose), breaking them down into monosaccharides like glucose and galactose, which can be absorbed and utilized by yeast.
Saccharomyces cerevisiae can secrete some extracellular peptide substances known as "reactivation factors". These substances have been proven to protect the coexisting Lactobacillus from the stress damage caused by bile salts, which is crucial for the survival of probiotics in the upper digestive tract.The deep cavity of the intestine is an anaerobic or micro-aerobic environment. Saccharomyces cerevisiae mainly undergoes alcohol fermentation here, producing ethanol. High concentrations of ethanol are also toxic to the yeast itself. Studies have found that co-culture with Lactobacillus plantarum can significantly improve the ethanol tolerance of Saccharomyces cerevisiae, enabling it to persist longer and function more effectively in the intestinal environment^[4-7].

Although the symbiotic system can provide mutual support and enhance stability in the intestinal environment, the team found that it still faces severe challenges when passing through the gastrointestinal tract. Saccharomyces cerevisiae must first withstand the harsh acidic gastric juice and bile salts, but its survival rate will significantly decrease. This directly affects the number of effective bacteria that can reach the intestine and function properly. The team realized that if most of the engineered yeast were inactivated before reaching the target site, then no matter how sophisticated the symbiotic system design was, it could not achieve the expected therapeutic effect. Therefore, although the symbiotic system provided important safety support for the project, it itself could not completely solve the physical and chemical challenges brought by the digestive tract environment, and additional protective measures were urgently needed.

Through this round of safety cycle, the team has clearly recognized the value of the symbiotic system as an endogenous security strategy. However, we also realize that relying solely on the biological symbiosis and mutual benefit cannot completely address the stringent challenges of the digestive tract environment. To ensure that the engineered yeast can reach the intestine stably and efficiently and exert its therapeutic effect, we must develop additional protective mechanisms. Our next step will be to focus on researching and comparing various protection strategies against gastric acid and bile salts, in order to find the safest, most effective and feasible solution.

Before the Saccharomyces cerevisiae enters the intestine as the host strain, it must overcome the challenges of strong acidic gastric juice (pH 1.5–3.5) and bile salts. To this end, we systematically investigated various protective strategies and classified them into three categories:
a. Intrinsic biological strategies: enhancing tolerance through genetic engineering (PMA1 overexpression, strengthening the CWI pathway, ALE evolution).
b. Extrinsic physical strategies: providing a barrier through microcapsule encapsulation (alginate, SA-chitosan complex, functional composite materials).
c. Exogenous and symbiotic strategies: buffering acid and bile salts by using food matrices or co-cultured microorganisms.
After comparison, we found that the sodium alginate-chitosan composite microcapsule achieved the best balance point in terms of protective effect, feasibility, cost and safety. Therefore, it was used as the anti-gastric acid and anti-bile acid module.

Table 4. Comparison of yeast protection strategies

We envision using the alginate-chitosan composite microcapsules to provide an external "armor" for the Saccharomyces cerevisiae:
The yeast cells are first encapsulated in the alginate gel matrix; through Ca²⁺ cross-linking,a stable gel network is formed；an outer layer of chitosan is then coated to form a dense poly-electrolyte composite membrane；after entering the small intestine,the microcapsules gradually disintegrate, releasing the active yeast.

Based on the literature and experimental feasibility, we have formulated the following plan:
Core matrix: 3% (w/v) sodium alginate solution;
Crosslinking agent: 2% (w/v) CaCl₂ solution, cured for 30 minutes;
Protective shell: 0.38% (w/v) chitosan solution, coated for 60 minutes;
Post-treatment: Sodium citrate treatment is carried out to obtain a "liquid core", avoiding physical compression and improving survival rate.
Preparation method: Gel beads are formed by syringe drop method, with aseptic operation throughout the process

We verified the protective effect of microcapsules through in vitro simulation experiments: free yeast was basically inactivated; single-layer SA embedding provided partial protection; SA-chitosan composite embedding showed a significantly higher survival rate.

Fig 3. Microcapsule coating experiment

Through this round of cycle, we have gained the following experiences:
a. The SA-chitosan composite system effectively enhanced the survival rate of yeast.
b. The liquid core design prevented cells from being compressed and improved the long-term survival rate.
c. The subsequent optimization directions are:
adjusting the concentrations of SA and chitosan to improve acid tolerance;
adding prebiotics (such as inulin) to the matrix to achieve a "synbiotic" effect;
combining genetic engineering strategies (such as overexpression of PMA1) with physical encapsulation to construct a dual-layer defense system^[8-13].

Just like most oral medications and health supplements, when addressing a certain issue in the human body, we should minimize the harm to the body as much as possible. In our previous design and construction, we took this into consideration. However, we felt that a safe and effective way was still needed to directly eliminate the harm of the product to the human body. We conducted extensive literature research and reviewed the safety control systems of past iGEM teams, from which we gained inspiration.

We believe that the safest and most effective method is through a kind of "signal-controlled sterilization". It consists of sensors and effectors, just like the reflexes in the human body. It is simple yet highly effective.

Fig 4. Our security strategy direction

Since this product is used through the human body, we initially selected the sensors of the applicable safety lines for this project: 1. Input signal from the front of the body 2. Changes in internal signals 3. Input signal from the rear of the body. The input signal from the front of the body refers to when the product is inside the body, it activates the safety line by receiving signals from certain substances taken orally; the change in internal signals means that some inherent conditions within the human body act as signals, and this signal change is activated by similar "0-1" input changes in the engineering process; the input signal from the rear of the body refers to the need to apply external factors to activate the safety line when a small number of engineering bacteria are excreted outside the body and there may be a risk of leakage.
As for the effectors, we designed them from the very beginning using the anti-toxin-toxin system and nucleic acid metabolism perspectives. This part of the review is mainly conducted from these two perspectives, and in the future, our method will also be referred to by other teams' methods for further optimization.

After conducting a systematic review, we verified that our idea was feasible. Moreover, we also applied this perception solution in other modules. This part reflects our contribution to the engineering cycle in our safety review. For the specific review, please refer to the safety section.

Through a systematic review of the sensing signals and in combination with the characteristics of our own project, we have initially understood how to select the conditions for regulating the safe circuit. However, most of our review was focused on prokaryotes. The activation expression elements for the corresponding conditions in Saccharomyces cerevisiae still require further exploration by us.

The first type of chassis bacteria we consider to be in a "suicide scenario" is the issue of preventing leakage when they are excreted out of the body.
1. Environmental persistence: The engineered bacteria survive unexpectedly in the soil or water.
2. HGT risk: The genetic circuits of the engineered bacteria transfer to the natural bacterial community, causing unpredictable ecological impacts.
So, how can we ensure that the engineered bacteria do not spread to the environment when they are in the action environment? How can we achieve "elimination upon exhaustion"?

We reviewed the results of the previous engineering cycle and wondered: Isn't it precisely by sensing the changes within the body that we can achieve the effect of "exhaustion leads to extinction" in a specific environment?

The most reliable external safety switch must utilize the inherent and inevitable physical or chemical differences between the internal environment (the intestinal tract) and the external environment (waste). Research has found that the temperature difference is one of the simplest and most reliable signals.
Internal safety state: The core body temperature of the human body remains at 37 ℃ ± 1 ℃.
External clearance state: After excretions enter the environment (such as a laboratory or environmental water bodies), the temperature usually drops rapidly to the environmental temperature (usually below 30 ℃).
This stable temperature gradient can meet the ideal control logic of "OFF in the body, ON outside the body". Based on this, we chose to construct a cold-triggered kill switch.
To achieve the logic of "surviving at 37℃ (OFF), killing at low temperatures (ON)", we cannot simply use cold-inducing promoters because many cold promoters have the defect of low background expression, which may lead to leakage at 37℃. We conceived a reverse logic circuit based on heat shock promoters to ensure the highest control accuracy:
Utilizing the heat-activated promoter (PHSP) to drive a transcriptional repressor.
At 37℃ (inside the body): PHSP activity is high → A large number of inhibitory factors are expressed → The killing system is completely inhibited → Safe State (OFF).
At low temperature (outside the body): PHSP activity is low → The expression of inhibitory factors decreases → The killing system is un-inhibited and activated → Kill State (ON)

Fig 5. Temperature-killing design direction

Regarding the comparison and selection of promoters, Saccharomyces cerevisiae possesses multiple heat shock protein promoters (such as PHSP104, PHSP82, PHSP26). After comparing the characteristics of the promoters in the literature, we chose PHSP26. The PHSP26 promoter is renowned for its high sensitivity to heat stress and fast response. We hypothesized (and will verify in the tests) that: PHSP26 can produce sufficient high concentrations of inhibitory protein (TetR) at 37°C (equivalent to a mild heat shock for yeast), ensuring the inhibition and suppression of the downstream killing pathway (with extremely low background leakage). When the temperature drops, the transcriptional activity of PHSP26 will rapidly decline, releasing the inhibition and triggering the killing process.
In terms of the effectors, we start from the genetic material DNA to design the killing mechanism. We have divided it into two schemes: inhibiting and blocking DNA synthesis, and directly destroying the DNA structure. First, we identified the key gene ADE2 in the nucleic acid metabolism of Saccharomy cescerevisiae.The ADE2 gene encodes phosphoribosylaminoimidazole carboxylase, which is a key enzyme in the de novo biosynthesis pathway of purines (such as adenine)^[14].
Then we identified the nucleases nucA, which, as an heterologous protein, has been experimentally demonstrated in the literature to be capable of exerting a lethal effect on yeast cells^[15].
Regarding this effector aspect, we proposed two schemes:
Scheme One (ADE2): (Fig 6. A)
Design the lox66 and lox71 sequences, and insert them into the upstream and downstream of the ADE2 gene using the CRISPR/Cas9 technology, so that ADE2 is in a "ready for deletion" state that Cre can recognize.
Construct the PtetR→Cre Recombinase expression plasmid → express the knockout of ADE2
Scheme Two (NucA): (Fig 6. B)
Construct the PtetR→nucA expression plasmid
By combining the sensor and the effector in this way, we obtained the in vitro safe circuit that we designed.

Fig 6. Construction of external safety circuits

In this part, we conducted an intensity experiment to verify the important promoter HSP26. We placed EGFP downstream of the HSP26 promoter and performed an enzyme reader analysis. The final results showed that HSP26 is indeed an excellent temperature promoter.
Additionally, due to the heterogeneity of nucA, we needed to be cautious. We discovered that nucA without a signal peptide segment in the literature still has functionality. Therefore, we predicted the site of signal peptide cleavage and finally removed the signal peptide segment to obtain the target protein. The predicted 3D structure has high confidence and is close to the real structure. This provided a strong foundation for our laboratory's construction.

Fig 7. Protein acquisition of nucA

In this part, we conducted an intensity experiment to verify the important promoter HSP26. We placed EGFP downstream of the HSP26 promoter and performed an enzyme reader analysis. The final results showed that HSP26 is indeed an excellent temperature promoter.
Additionally, due to the heterogeneity of nucA, we needed to be cautious. We discovered that nucA without a signal peptide segment in the literature still has functionality. Therefore, we predicted the site of signal peptide cleavage and finally removed the signal peptide segment to obtain the target protein. The predicted 3D structure has high confidence and is close to the real structure. This provided a strong foundation for our laboratory's construction.

Following the previous cycle, how can we ensure that the engineered bacteria can be effectively removed after completing their tasks, avoiding their prolonged retention in the intestinal tract and thereby reducing the risk of entering the bloodstream?
Referring to our review of the sensors, we discussed that the "drug" control system has great potential for regulating the safe pathways within the body. When the product is put into use, this sensing method can fully exert its flexibility to ensure personal safety after consumption. For the effector part, we investigated various biological safety control strategies, including traditional antibiotic killing, nutritional defect types, and inducible suicide switches. Among them, the exogenous small molecule-induced toxin-antitoxin system is highly regarded due to its strong controllability, rapid response, and independence from host endogenous signals.

We envision that a highly controllable "suicide switch" can be designed, which will be activated when the engineered bacteria reach a specific stage within the host body. This switch should be like a secure "key", only activating the clearance program upon receiving a clear external instruction, thereby minimizing the potential risks to the host to the greatest extent.

Based on the above concept, we learned from the GreatBay-SCIE 2023 team and designed a set of an Araboside-induced K1 toxin expression system as a safe clearance pathway in vivo:
Sensor: We use Araboside to induce the safety system. When oral Araboside is taken, the engineered bacteria will sense this safety signal.
Effectors: Use the expression system of K1 killer toxin. The precursor of K1 toxin is processed in the cells to form a mature αβ heterodimer, which can specifically bind and lyse the cell wall of sensitive yeast cells, causing cell death.
Working principle:
a. During normal operation, the repressive promoter pNOTAra inhibits toxin expression, and the system is in a safe state;
b. When it is necessary to clear the engineered bacteria, the patient orally takes Araboside. The Araboside enters the intestine and is perceived by the engineered bacteria;
c. The inducible promoter pAra is activated, driving the expression of the K1 toxin precursor protein;
d. The toxin precursor is processed and cut in the endoplasmic reticulum and Golgi apparatus, and finally secreted to the extracellular space;
e. The mature toxin binds to the β-1,6-glucan of the yeast cell wall through its α subunit, and the β subunit forms a channel, causing the engineered yeast cells to lyse and die.

Fig 8. Construction of internal safety circuits

We verified the feasibility of this safety circuit through literature and existing iGEM projects, and also identified its potential shortcomings: Feasibility verification: The arabinose induction system has been successfully applied in both Escherichia coli and yeast, with rapid response and high sensitivity. The K1 toxin has a highly specific killing effect on Saccharomyces cerevisiae. We have conducted preliminary experiments so far, but still need optimization.

Fig 9. Arabinose-induced preliminary experiment

At present, we have also identified some potential deficiencies in the system: 1. Leakage expression risk: There is background leakage expression of the two promoters in the arabinose regulatory system, and the dynamic relationship between them needs further study. 2. Individual differences impact: Different individuals have different rates of absorption and metabolism of arabinose in the intestine, which may lead to unstable induction efficiency. In the future, many aspects need to be considered and optimized.

Firstly, we conducted a thorough literature review on the spore formation mechanism of Saccharomyces cerevisiae, and focused on referring to the official safety guidelines of iGEM. According to the White List (White List | iGEM Responsibility)published on the iGEM website, although Saccharomyces cerevisiae is classified as a microorganism with biosafety level 1, since it has the potential to form spores, the team must formulate measures to reduce the associated risks. The official clearly stated that spore formation, due to its difficulty to be controlled, may lead to "Release Beyond Containment", which would directly violate the core safety policy of iGEM.
Studies have shown that spore formation is a complex process that is strictly regulated, and its initiation depends on specific environmental signals - extreme nitrogen deficiency and the presence of non-fermentable carbon sources. Among the key regulatory genes, the transcriptional activator Ime1p encoded by the IME1 gene is the "master switch" for meiosis and spore formation; the IME4 gene determines the accumulation of IME1 transcripts; and the SPO11 gene is responsible for the DNA double-strand breaks required for meiosis^[16].

Based on the above research, we envision using genetic engineering techniques to fundamentally eliminate the spore-forming ability of Saccharomyces cerevisiae. Our goal is to establish a "spore-free" chassis that permanently resists the environmental signals that induce spore formation. Compared to relying solely on external environmental control (which is unreliable outside the laboratory), directly disrupting the genetic pathway is a more thorough and controllable safety strategy.

Among all the candidate genes, we finally chose IME1 as the knockout target. As the highest regulator of spore formation, the deletion of IME1 can completely block the entire cascade reaction from its source, and has the least impact on the normal growth of yeast.

Literature verification
Multiple authoritative studies have shown that the deletion of IME1 can specifically and stably eliminate the ability to form spores. IME1 has been confirmed to be the "master switch" for meiosis and spore formation. After deleting this gene, cells are completely insensitive to induction signals such as nitrogen source deficiency. Relevant experimental evidence indicates that the diploid strain ime1Δ/ime1Δ is completely unable to form spores in the spore formation medium. Additionally, the same study also proved that the ime1Δ mutation does not affect mitotic reproduction and basic metabolism. For example, some related applications have successfully constructed sporeless ime1Δ strains while maintaining excellent biotechnological characteristics^[15].
Experimental Design
Knockout of the IME1 gene: We used the CRISPR/Cas9 technology to design and implement the precise knockout of the IME1 gene.
Spore formation experiment: The modified yeast and the wild-type yeast were placed together in standard spore formation media (such as potassium acetate medium), and observed under a microscope for several days. The wild-type should show typical tetrad spores, while the modified strain would completely fail to form any spore structures.

Fig 10. The experimental design for the verification after knocking out IME1

Through this round of cycle, we envision having constructed a safe platform that cannot form spores, effectively addressing the significant biological safety risks posed by spore dispersion in engineered bacteria.
However, we also recognize that merely achieving "spore-free" status is not sufficient to establish a comprehensive safety system. Although the spore risk has been eliminated, the nutrient cells may still survive for a short period after being released into the environment, and there is a very low probability of horizontal gene transfer (HGT) risk.
In the future, we will conduct more exploration of target genes and carry out related wet experiments to select the most suitable knockout genes for this project.