Results

1 Overview

The spoilage and deterioration of food, especially aquatic products, pose significant challenges to global food safety. Shortly after harvest, seafood often experiences a decline in quality, which not only decreases its shelf life but also cause symptoms such as diarrhea, vomiting, and acute poisoning in consumers, primarily due to the accumulation of histamine toxins and proliferation of pathogens. Although the industry predominantly employs chemical preservatives and cold chain systems to preserve product freshness, the carcinogenic risks associated with additives like sulfites and the environmental impact from carbon emissions of cold chain systems reliant on fossil fuels necessitate a trade-off between economic expense and environmental health. Hence, the development of safe and efficient natural preservation techniques is of paramount importance.

Chitosan is the second most abundant natural polysaccharide in nature, predominantly found in the exoskeletons of crustaceans such as shrimp and crab. However, it is often discarded due to the high costs associated with its processing. Chitooligosaccharides (COS), a degradation product of chitosan, demonstrate enhanced water solubility and antibacterial properties, which can effectively damage bacterial cell membranes and offering potential preservatives for seafood and fruits (1). According to the HP survey (see IHP for more details), challenges include the complexity of the process, high costs, environmental concerns, low efficiency, and difficulties in product control, particularly in the biology, chemical and physical conversion of chitosan into chitosan oligosaccharide. Hence, our CRUSTA project seeks to produce COS using a route developed through synthetic biology, which can serve as alternative to chemical preservatives.

➤ Initiated by the discovery that COS with higher degrees of polymerization possess enhanced antibacterial properties, we established a chitosanase screening method that relies on the antibacterial efficiency of COS degraded by mutants. At last, we obtained the SaCsn46A mutant 2 (SaCsn46A E151G D222G) from the various mutants whose degradation products exhibited the best antibacterial properties.

➤ To reduce the subsequent enzyme extraction cost, we semi-rationally designed the signal peptide LMT and obtained the mutant LMT C14A with higher secretion efficiency.

➤ We achieved chitosan degradation via a secretion system, enhancing the antibacterial performance of the degradation products by 7.96-fold compared to that of wild-type.

➤ To prevent engineered bacteria in the chitosanase production chain from contaminating other processes and posing biosafety risks, we designed two independent suicide switches. The first is a ccdB/ccdA toxin-antitoxin system regulated by cuminic acid, and the second is an IPTG-inducible system that triggers overexpression of the tyrS gene. Together, we integrated these two systems into a same plasmid, they can operate independently, thus guarantee rapid cell death, keeping the engineered bacteria in a controllable state.

When the viable bacterial count reaches 10⁸ CFU/g or higher, the food is considered to be in the initial stage of spoilage. The results showed that the degradation products had an excellent fresh-keeping effect and could effectively inhibit microbial growth. This outcome confirms the rationality and feasibility of our designed solution, paving the way for its application in practical production.

2 Highlights

We established a chitosanase screening method that relies on the antibacterial efficiency of COS degraded by mutants, by which obtained the SaCsn46A mutant 2 (SaCsn46A E151G D222G) from the various mutants whose degradation products exhibited the best antibacterial properties.

We semi-rationally designed the signal peptide LMT and obtained the mutant LMT C14A with higher secretion efficiency.

We achieved extracellular chitosan degradation via a secretion system, enhancing the antibacterial performance of the degradation products by 7.96-fold compared to that of wild type.

Our route saves about 61.0% of CO₂ emissions and 71.7% of cost compared to the chemical production of traditional preservative, potassium sorbate, in the same functional performance.

We integrated ccdB/ccdA toxin-antitoxin with tyrS system into a kill switch circuit, which can operate independently, keeping the engineered bacteria in a controllable state.

3 Determination of the Antibacterial Activity and Preservation Performance of commercialized COS

To investigate the potential of COS with different average molecular weights as preservatives, we purchased COS standards with different average molecular weights, and conducted a series of experiments to test their antibacterial activity and preservation performance at different concentrations (3).

3.1 The Antibacterial Activity of commercialized COS

We purchased three varieties of commercial COS products. These are mixtures comprising COS of varying polymerization degrees, with average molecular weights under 1 kDa (Beyotime, Y261047), 2 kDa (Beyotime, Y261046), and 3 kDa (Beyotime, Y261045), respectively. We employed Escherichia coli BL21(DE3) as the indicator strain, culturing it until the optical density (OD₆₀₀) reached 0.6. Subsequently, we introduced three types of commercial COS solutions at varying concentrations and measured the 24-hour growth curves of the strains, as shown in Figures 1 A-C, to evaluate their antibacterial efficacy. The experimental results showed that all three types of COS exhibited significant antibacterial effects when their concentration exceeded 1.2% (w/v). Furthermore, higher concentrations of COS resulted in bacterial cell death and flocculation. Additionally, we prepared LB solid media containing COS with different type and concentrations, onto which we inoculated single colonies of E. coli BL21(DE3). After 14 hours, the survival rates were calculated to assess the antibacterial properties (Figure 1 D). The results showed a general decline in survival rates as COS concentrations increased; when the concentration exceeded 1.5% (w/v), it completely inhibited the growth of the strain.

Figure 1 Antibacterial activity of COS with different average molecular weights and concentrations. (A-C) OD₆₀₀ values of E. coli BL21(DE3) cultures containing COS with average molecular weights < 1 kDa (A), < 2 kDa (B), and < 3 kDa (C) at indicated concentrations over 24 h. (D) Bacterial survival rates on LB solid media containing COS with different average molecular weights and concentrations.

Therefore, our experiments demonstrate that COS with varying average molecular weights exhibit significant antibacterial effects at concentrations exceeding 1.5% (w/v). Notably, comparative analysis under identical conditions revealed that COS with average molecular weights below 3 kDa exhibits the strongest antibacterial activity.

3.2 The Preservation Performance of COS

Based on the results of the antibacterial activities of COS, we have preliminarily confirmed that COS effectively inhibit microbial reproduction and possesses considerable potential for food preservation. To further validate the COS's efficacy in food preservation, subsequent experiments will focus on evaluating its preservation performance. After approval by the Safety and Security Committee (Safety Form), we will measure the total bacterial colony count to assess the ability of COS to delay the spoilage in aquatic products and fruits, thus providing a comprehensively evaluation of its utility in food preservation.

3.2.1 Aquatic Product Preservation

3.2.1.1 Effects of COS Concentration and Average Molecular weights on Aquatic Product Preservation

To evaluate the preservation performance of COS with various average molecular weights under 1 kDa, 2 kDa, and 3 kDa on aquatic products, we immersed fish samples in sterile solutions of the commercial COS. These treated fish samples were then sealed in sterile bags and stored at 4°C for the number of days. After storage period, the fish samples were mashed within these sterile bags. We added sterile water to the mashed samples, collected the resultant leachate, diluted it at a specified ratio, and then plated it on plate count agar (PCA) medium. The medium was then incubated at 37°C for 48 hours, after which colonies were counted. Colony counts were normalized to colony-forming units per gram of fish meat (CFU/g) (Figure 2A).

The results indicated that by the second day of the experiment, all groups treated with COS of various average molecular weights exhibited significant preservation effects on aquatic product, with marked differences from the sterile water control group (Figure 2B). However, no statistically significant differences in preservation efficacy were observed among the COS-treated groups with different average molecular weights (Figure 2C).

Figure 2 Aquatic product preservation performance of COS with different average molecular weights in fish samples. (A) Treatment of fish samples with COS of different average molecular weights (< 1 kDa, < 2 kDa, < 3 kDa). (B) On Day 2 post-treatment, all COS-treated groups showed significant aquatic product preservation performance, with obvious differences from the sterile water control group. (C) No significant differences in aquatic product preservation performance between COS-treated groups with average molecular weights.

We evaluated the preservation performance of COS with average molecular weights under 3 kDa on fish samples using three concentrations: 15, 20, and 25 mg/mL. This experimental design allowed us to systematically assess the impact of varying COS concentrations on aquatic product preservation. The results indicated a decreasing trend in colony count across all COS-treated samples, in contrast to a marked increase observed in the sterile water control group (Figure 3A). Although a dose-dependent trend was apparent, there was no significant difference in preservation performance among the COS-treated groups (Figure 3B). By the third day of the experiment, all COS-treated groups exhibited statistically significant differences compared to the control group (Figure 3B). We observed an interesting phenomenon that the initial bacterial count in fish purchased at varying times shows significant variation, with differences reaching an 1-2 of magnitude. At an initial CFU of 10⁵, a concentration of 20 mg/mL of COS demonstrates a robust inhibitory effect lasting three days or longer. However, when the initial CFU increases to 10⁶, this inhibitory effect diminishes.

Figure 3 The impact of varying COS concentrations on fish samples preservation. (A) Treatment of fish samples with COS (< 3 kDa) of different concentrations (25 mg/mL, 20 mg/mL, 15 mg/mL). (B) Significant differences between COS-treated groups and the sterile water control group on Day 3 post-treatment. (C) Comparison of preservation performance between COS (< 3 kDa) and potassium sorbate in fish samples on Day 1 post-treatment.

3.2.1.2 Comparison of Aquatic Product Preservation Performance Between COS and Potassium Sorbate

To evaluate the practicality of COS, its preservation efficacy was compared with that of potassium sorbate, a widely used chemical preservative. Preservation experiments were conducted on fish samples using 20 mg/mL of COS (average molecular weights < 3 kDa) and 0.2% (w/v) potassium sorbate. At these concentrations, COS exhibited superior preservation efficacy to potassium sorbate on the first day post-treatment (Figure 3C).

The findings from both methods showed relative consistency. On the second day, the < 3 kDa COS group exhibited significant fruit preservation efficacy, whereas the < 1 kDa and < 2 kDa COS groups demonstrated lesser effectiveness. Additionally, experimental outcomes suggested a correlation between the preservation efficacy of COS on tomatoes and its average molecular weights (Figure 4C, D).

Figure 4 Fruit preservation performance of COS with different average molecular weights (< 1 kDa, < 2 kDa, and < 3 kDa) on tomatoes. (A) Colony counts after direct treatment. (B) Colony counts after integral crushing treatment. (C) Comparison between the COS-treated group and the sterile water control group on Day 2 post-direct-treatment. (D) Comparison among COS-treated groups on Day 2 post-direct-treatment. (E) Comparison of fruit preservation performance between COS (< 3 kDa) and potassium sorbate in the tomato sample on Day 3 post-treatment.

3.2.2 Fruit Preservation

3.2.2.1 Preservation Effects of COS on Fruit Preservation

To evaluate the preservation efficacy of COS with various average molecular weights (< 3 kDa, < 2 kDa, < 1 kDa) on fruits, tomatoes were selected as the test subject. The initial method entailed a direct approach to isolate microorganisms present on the surface, while the second method, known as integral crushing treatment, was designed to gather both surface and internal microorganisms (see Experiments for details). By integrating these two methods, a comprehensive assessment of the overall microbial contamination on tomatoes was achieved.

The findings from both methods showed relative consistency. On the second day, the < 3 kDa COS group exhibited significant fruit preservation efficacy, whereas the < 1 kDa and < 2 kDa COS groups demonstrated lesser effectiveness. Additionally, experimental outcomes suggested a correlation between the preservation efficacy of COS on tomatoes and its average molecular weights (Figure 4C, D).

3.2.2.2 Comparison of Fruit Preservation Performance Between COS and Potassium Sorbate

Similarly, we compared the efficacy of COS in preserving fruits with that of potassium sorbate. Preservation trials were performed on tomato samples using 20 mg/mL of COS (average molecular weights < 3 kDa) and 0.2% (w/v) potassium sorbate. By the third day post-treatment, both COS and potassium sorbate demonstrated significant and comparable antibacterial properties, showing significant differences from the sterile water control group (Figure 4E).

4 Screening for superior chitosanase

Current production methods for COS involve the degradation of chitosan, a deacetylated derivative of crustacean waste, and fall into three primary categories: chemical, physical, and enzymatic degradation (3). Chemical degradation utilizes acids, alkalis, or oxidizers, providing low cost and high efficiency but resulting in significant environmental pollution (4). In contrast, physical degradation methods, such as ultrasound or microwave-assisted techniques, are eco-friendly but less efficient (5).

Enzymatic degradation, which employs chitosanase and other enzymes to specifically cleave glycosidic bonds under mild conditions (ambient temperature and near-neutral pH), is the preferred method due to its eco-friendliness and efficient. This method produces products with uniform degree of polymerization (DP) and preserved biological activity, which are ideal for high-value industries like food and pharmaceuticals. Nonetheless, the high cost of enzymes limits its scalability (3).

In the CRUSTA project, aiming for rapid and sustainable production of COS, enzymatic degradation has been selected. After reviewing the literature, we chose 6 kinds of chitosanases and constructed a mutant library, targeting the identification of a chitosanase mutant that optimally degrades chitosan into COS with enhanced antibacterial properties (see our Design page for details).

4.1 Preliminary Screening of Chitosanases

The targeted sequences BBa_25E8EHJF, BBa_25JJYAF6, BBa_25JA5N69, BBa_25RYX5T9, BBa_25WTOTOC, BBa_252ULD2I were individually inserted in pET-28a(+) to construct circuits (Table 1), which were transformed into E. coli BL21(DE3) to express each enzyme. The positive transformants were confirmed by kanamycin, colony PCR (Figure 5), and sequencing.

Figure 5 DNA gel electrophoresis of the colony PCR products of BamCsn (A), CsnMY002 (B), Glm_Tk (C), CsnCA WT (D), McChoA WT (E) and SaCsn46A WT (F)_pET-28a(+) in E. coli BL21(DE3). Target bands of BBa_25JA5N69 can be observed at the position between 2000 bp and 3000 bp, the others target can be observed at the position between 1000 bp and 2000 bp.

After being cultivated and induced under responsive conditions (see our Experiments page for details), we lysed the cells via sonication and collected the supernatant of the lysate. SDS-PAGE analysis and Coomassie blue staining were used to verify the expression of the target protein (Figure 6).

Figure 6 SDS-PAGE analysis. BamCsn (A), CsnMY002 (B), Glm_Tk (C), CsnCA (D), McChoA (E), and SaCsn46A (F). Target bands can be observed between 25 kDa and 35 kDa, and the Glm_Tk can be observed between 80 kDa and 150 kDa.

Chitosan powder was added to the supernatant of lysate, followed by incubation at 37°C for 18 h. The supernatant obtained thereafter was designated as the degradation product. The degradation product was then co-cultured with the test strain (E. coli BL21(DE3), OD₆₀₀ ~0.6) in a 1:1 ratio. The OD₆₀₀ of the co-culture was measured at 2 h and 8 h. The antibacterial activity of the enzymatic degradation product assessed by calculating the relative ΔOD₆₀₀ value as (OD₆₀₀ at 8 h - OD₆₀₀ at 2 h) / OD₆₀₀ at 2 h. The results were subsequently plotted (Figure 7).

Figure 7 Antibacterial activity of degradation products of BamCsn, CsnMY002, GlmA_TK, CsnCA WT, McChoA WT, SaCsn46A WT.

Based on the preliminary experiment results, we evaluated the antibacterial efficacy of the enzymatic degradation products from six chitosanases. Notably, the enzymatic degradation products of CsnMY002, CsnCA, McChoA, and SaCsn46A displayed significant antibacterial effects. Nevertheless, the antibacterial activity of their degradation products was inadequate for optimal preservation that we had to conducted random mutagenesis on these four chitosanases to generate mutants for further screening.

4.2 Secondary Screening of Chitosanases

Detailed steps for circuit construction and induced expression are outlined on our Experiments page. Ultimately, we successfully obtained 103 mutants of SaCsn46A, 18 mutants of CsnCA, and 19 mutants of McChoA (Table 2). Unfortunately, despite multiple attempts, we were unable to successfully construct genetic circuits containing fragments of CsnMY002 mutant. Therefore, we try to focuses on evaluating the antibacterial activity of the degradation products from the SaCsn46A, CsnCA, and McChoA.

The targeted sequences were inserted in pET-28a(+) to construct circuits, which were transformed into E. coli JM109(DE3) to express each enzyme. The positive transformants were confirmed by kanamycin, colony PCR (see our Parts page for details), and sequencing. After being cultivated and induced under responsive conditions (see our Experiments page for details), we lysed the cells vis sonication and collected the lysate supernatant. Chitosan powder was added to the lysate supernatant, followed by incubation at 37°C for 18 h. The resulting supernatant was collected as the degradation product.

The degradation products of SaCsn46A, McChoA, and CsnCA mutants were separately co-cultured with E. coli BL21(DE3) at a 1:1 ratio. The initial optical density (OD₆₀₀) of the test strains used for SaCsn46A mutants, McChoA mutants and CsnCA mutants was 0.8, 0.8, and 0.6, respectively. OD₆₀₀ measurements in co-culture systems were recorded at 1.5 h and 8 h for SaCsn46A and CsnCA mutants, and at 2 h and 8 h for McChoA mutants. Antibacterial activity of degradation products from SaCsn46A mutants was evaluated using the relative ΔOD₆₀₀ value, and calculated by the following formula:

Relative ΔOD₆₀₀ value = (OD₆₀₀ at 8 h - OD₆₀₀ at t₀) / OD₆₀₀ at t₀

where t₀ refers to the earlier measurement time point.

Through the exploration of the antibacterial effects of SaCsn46A WT, McChoA WT, CsnCA WT and their mutants, we successfully obtained mutants with different antibacterial efficiencies of degradation products. The relative OD₆₀₀ values of most McChoA mutants fluctuates slightly around the EV baseline (Figure 8B), with no mutants showing extremely enhanced or weakened antibacterial activity of their enzymatic degradation products, and no mutants exhibit extreme changes in the antibacterial activity of their enzymatic degradation products. The overall antibacterial performance of CsnCA mutants is stable (Figure 8C), and mutations do not induce substantial alterations in their antibacterial efficacy. Notably, SaCsn46A mutants demonstrated two key advantages. Firstly, we obtained more SaCsn46A mutants, and these mutants exhibited greater diversity in mutation sites and a broader range of antibacterial activity in their enzymatic degradation products. Secondly, they exhibited better antibacterial activities (Figure 8A). Therefore, we selected the six SaCsn46A mutants that showed the most promising results for further analysis. These included: SaCsn46A mutant 2 (SaCsn46A E151G D222G), SaCsn46A mutant 14 (SaCsn46A N29S K50R D150G D189G F247L), SaCsn46A mutant 16 (SaCsn46A V52A D134G K165I F175L S190C G210D K250E N252D), SaCsn46A mutant 60 (SaCsn46A D40G Q125R H234R), and SaCsn46A mutant 94 (SaCsn46A T249A L259P).

Figure 8 Antibacterial activity of enzymatic degradation products from wild-type (WT) and mutant chitosanases. (A) SaCsn46A mutants. (B) McChoA mutants. (C) CsnCA mutants. BL: sterile water, EV: empty vector pET-28a(+), WT: SaCsn46A WT, McChoA WT and CsnCA WT respectively. Relative OD₆₀₀ values were normalized using empty vector pET-28a(+) as the baseline value of 1.

4.3 Tertiary Screening of Chitosanases

At this stage, following cultivation and induction, the GE AKTA Prime Plus FPLC System was employed to purify proteins from the lysate supernatant of these six SaCsn46A mutants. Ultimately, these six SaCsn46A mutants were purified and normalized to a uniform concentration of 0.2207 mg/mL. In contrast, the protein concentration of SaCsn46A mutant 2 (SaCsn46A E151G D222G) was 0.0167 mg/mL, and that of SaCsn46A WT was 0.1607 mg/mL. The expression of the target protein was confirmed using SDS-PAGE and Coomassie blue staining (Figure 9).

Figure 9 SDS-PAGE analysis of SaCsn46A and its mutants. Target bands (about 30.5 kDa) can be observed between 25 kDa and 35 kDa.

Employing the same screening method as in the previous section, we evaluated the antibacterial activity of their degradation products. As shown in Figure 10, SaCsn46A mutant 2 (E151G D222G) and SaCsn46A mutant 60 (D40G Q125R H234R) exhibited superior performance, with their enzymatic degradation products demonstrating superior antibacterial activity. Consequently, these two mutants were therefore selected for subsequent analysis of fundamental enzymatic properties.

Figure 10 Antibacterial activity of enzymatic degradation products of SaCsn46A WT and its mutants under fixed protein concentration conditions.

5 Enzymatic Characterization of Chitosanase Mutants

5.1 Enzymatic property characterization experiments of Purified Chitosanase

Following the selection of SaCsn46A mutant 2 (E151G D222G) and SaCsn46A mutant 60 (D40G Q125R H234R), further analysis was carried out to assess their optimal temperature, pH and stability.

Chitosanase activity was measured using the 3,5-dinitrosalicylic acid (DNS) method. The quantity of reducing sugar was determined by measuring optical density (OD) at 540 nm using D-glucose as standard. The reaction mixture was incubated at 100°C for 5 min. One unit (U) of chitosanase activity is defined as the amount of enzyme that liberates 1 mmol of D-glucose-equivalent reducing sugars per minute under these conditions.

Details on the characterization methods for chitosanase properties can be found in the Experiments section.

The optimal reaction temperature of SaCsn46A mutant 60 (D40G Q125R H234R) is approximately 40°C, where it maintains relatively high activity even at 45°C, albeit with a slight decrease. In contrast, SaCsn46A mutant 2 (E151G D222G) displays a different temperature profile, experiencing a significant decline in activity at 40°C, which unexpectedly recovers at 45°C, returning to levels observed at 35°C (Figure 11A). The peak chitosanase activity for SaCsn46A mutant 2 (E151G D222G) and SaCsn46A mutant 60 (D40G Q125R H234R) occur at pH 5.0 and pH 4.0, respectively (Figure 11B). Thermostability analysis indicates that both mutants exhibited greater stability below 50°C. A minor increase in activity was observed around 40°C, which generally remains stable across this temperature spectrum (Figure 11C). Regarding pH stability, SaCsn46A mutant 2 (E151G D222G) shows robust stability between pH 6 and 8, whereas SaCsn46A mutant 60 (D40G Q125R H234R) is notably stable at pH 5 (Figure 11D).

Figure 11 Enzymatic properties of SaCsn46A mutant 2 and 60. (A) Effects of temperature conditions on enzyme activity. (B) Effects of pH conditions on enzyme activity. (C) Thermostability. (D) pH stability. The 100% relative activity of mutant 2 was 2.385 U/mg, while the 100% relative activity of mutant 60 was 2.124 U/mg.

5.2 hydrolytic properties of chitosanase

5.2.1 Characterization Results of the DP for COS Mixture Standards

To facilitate subsequent characterization, we utilized COS standards for High Performance Liquid Chromatography (HPLC) and Thin-Layer Chromatography (TLC) analysis. The objective was to pinpoint the standard peak positions of COS with various DPs, thereby establishing a basis for characterizing the degradation products of chitosanase in subsequent experiments.

We purchased COS mixture standards which consists of a mixture of COS with DPs ranging from 3 to 7. Subsequently, we prepared a series of COS standard solutions at concentrations of 0.62, 1.24, 2.49, 4.98 and 9.95 mg/mL, which were then filtered through a 0.22 μm microporous membrane. Then, a Shodex Asahipak NH₂P-50 4E column (4.6 mm × 250 mm, 5 μm) was used to quantitative analyze the content of COS with different DPs in the standard solutions (see Experiments for specific parameters). This procedure generated standard curves plotted the peak area of COS (with varied DPs) against mass concentration (Figure 12), establishing a foundational basis for the subsequent determination of COS with different DPs in degradation products.

Figure 12 Standard curves of COS with different DPs. Relationship between peak area and mass concentration for (A) Glucosamine (DP1), (B) Chitobiose (DP2), (C) Chitotriose (DP3), (D) Chitotetraose (DP4), (E) Chitopentaose (DP5), (F) Chitohexaose (DP6), and (G) Chitoheptaose (DP7).

A series of COS mixture standard solutions with concentrations of 1.24, 2.49, 4.98, and 9.95 mg/mL were prepared for TLC analysis, using a developing solvent system comprising isopropanol: water: ammonia in a 60: 30: 4 volume ratio. The COS mixture standard solutions with different concentrations were spotted on a silica gel plate for separation. After spraying with ninhydrin solution and heating for color development, fuchsin spots were observed at different Rf (retardation factor) values (Figure 13), indicating the presence of amino-containing carbohydrates in all detected components. The standards showed uniform color development, with clear spots, reasonable distribution, and no obvious tailing or impurity bands. This confirms that the used developing system is suitable for the separation and detection of COS with different DPs.

Figure 13 TLC analysis of COS standards. Lanes A-F correspond to COS concentrations of 1.24, 2.49, 4.98 (duplicate), and 9.95 (duplicate) mg/mL, respectively.

5.2.2 Characterization Results of the DP for Commercial COS

After preliminary protein purification, we obtained six wild-type chitosanases. Subsequently, we conducted degradation experiments and analyzed the DP of the degradation products from these enzymes. This analysis aimed to identify the product spectrum of each enzyme, which will provide essential guidance for subsequent research.

During conducting antimicrobial experiments using commercial COS with average molecular weights under 1 kDa, 2 kDa, and 3 kDa, we performed HPLC analysis on these COS solutions to investigate the effect of COS composition on their antimicrobial activity. Solutions of 50 mg/mL for each COS type were prepared, filtered through a 0.22 μm microporous membrane to yield samples, and then analyzed using HPLC with the same method as that for the standard curves (Figure 14).

Figure 14 HPLC analysis of commercial COS with different DPs. Chromatograms of COS with average molecular weights of (A) < 1 kDa, (B) < 2 kDa, and (C) < 3 kDa.

Among the commercial COS with average molecular weights < 1, < 2, and < 3 kDa, concentration of COS with DPs ranging from 1 to 3 progressively decreased, while those with DPs ranging from 4 to 6 progressively increased. Moreover, the commercial COS with average molecular weights < 3 kDa had the lowest content of low DPs COS.

Meanwhile, we performed TLC analysis on commercial COS samples with average molecular weights of < 1 kDa, < 2 kDa, and < 3 kDa to assess variations in COS composition across different DPs. Each sample were developed using the same solvent system and subsequently visualized by spraying with ninhydrin solution (Figure 15).

Results showed that multiple spots with different mobilities corresponding to COS components of varying polymerization degrees Notably, as average molecular weights increased, the intensity of spots associated with lower DPs diminished or disappeared altogether in the TLC profiles. This phenomenon indicates that the sample with average molecular weights < 3 kDa encompasses a larger proportion of high DPs COS, whereas the sample with average molecular weights < 1 kDa predominantly contains lower polymerization components. These findings validate significant disparities in COS composition ratios and imply a potential link to the observed variations in biological activity in later antimicrobial experiments.

Figure 15 TLC separation of COS with different average molecular weights. Lanes A, B, and C correspond to < 1 kDa, < 2 kDa, and < 3 kDa fractions.

5.2.3 Characterization Results of the DP of Degradation Products

After determining the optimal temperature and pH for SaCsn46A mutant 2 (SaCsn46A E151G D222G) and SaCsn46A mutant 60 (SaCsn46A D40G Q125R H234R), we purified the enzymes and obtained 0.1392 mg/mL of SaCsn46A mutant 2 (SaCsn46A E151G D222G) and 0.121 mg/mL of SaCsn46A mutant 60 (SaCsn46A D40G Q125R H234R). Degradation was then carried out for 30 minutes under optimal conditions using pretreated chitosan solution as the substrate. We first performed TLC, observing sample spots near the application point with minimal visibility elsewhere, leading us to infer the presence of high-degree polymerization COS in the products (Figure 16). To further characterize the distribution of oligomeric COS, we conducted HPLC analysis, obtaining the chromatograms shown (Figure 17). Based on retention times, the product peaks of SaCsn46A mutant 2 (SaCsn46A E151G D222G) included glucosamine, chitobiose, chitotriose, among others, while SaCsn46A mutant 60 (SaCsn46A D40G Q125R H234R) product peaks consisted of glucosamine, chitobiose, etc. Analysis of peak areas indicated that SaCsn46A mutant 2 (SaCsn46A E151G D222G) exhibits higher degradation efficiency.

Figure 16 TLC separation of products from the degradation experiment under optimal conditions. Lanes A, B, C, D correspond to glucosamine, COS standards (DP 3-7) of equal mass at 50 mg/mL, degradation products of SaCsn46A mutant 2 (SaCsn46A E151G D222G), degradation products of SaCsn46A mutant 60 (SaCsn46A D40G Q125R H234R).

Figure 17 HPLC analysis of degradation products of SaCsn46A mutant 2 (SaCsn46A E151G D222G) (A), SaCsn46A mutant 60 (SaCsn46A D40G Q125R H234R) (B).

6 Improvement of the Extracellular Secretion System

LMT, a Sec pathway signal peptide (BBa_K5136066), was first identified by our team from the upstream region of the lytic murein transglycosylase gene of Vibrio natriegens in 2021 (SALVAGE). Then in the following 2022 (OMEGA), 2023 (NAIADS), 2024 (REPARO) projects, LMT was successfully utilized to enable the extracellular secretion of functional proteins. Further characterization of LMT alongside several other signal peptides in 2024 (REPARO) revealed that LMT achieved our secretion requirements while minimizing metabolic burden, thus demonstrating its superior performance in comparison to other signal peptides.

Compared to intracellular expression systems that might form inclusion bodies, secretion-expression systems offer potential advantages for the production of recombinant proteins . Therefore, the signal peptide we identified holds significant application values. In this year's work, we aim to improve the secretion performance of LMT signal peptide to produce more recombinant proteins extracellularly, thus providing more robust tools and deeper insights for future iGEM teams and even SynBio community.

We fused sfGFP (BBa_K5136028) as a reporter to the LMT signal peptide (variants) and tested it in E. coli DH5α (Figure 18). The fluorescence intensity (λ_ex = 475 nm, λ_em = 515 nm) of bacterial culture and supernatant as well as OD₆₀₀ were monitored as time progressed during characterizations (see our Experiments for details).

Figure 18 Characterization device for LMT signal peptide or variants in E. coli DH5α.

6.1 Basic of Sec secretion pathway and sequence features of LMT signal peptide

Figure 19 Sec pathway-dependent secretion process and the corresponding sequence structures of signal peptide. (A) Secretion of recombinant proteins through Sec pathway. (B) The typical sequence features of signal peptide can be identified in the sequence of LMT.

The Sec pathway facilitates the translocation of many proteins in bacteria, archaea, and eukaryotes across the cytoplasmic or endoplasmic reticulum membrane (2). A classical Sec pathway signal peptide comprises three distinct regions: an N-terminal region enriched in positive charges (N-region), a central hydrophobic core (H-region), and a C-terminal region exhibiting strong polarity (C-region). In Sec pathway, proteins first bind to SecB via the positively charged N-region of their signal peptide while SecB inhibits premature folding of the protein. Subsequently, SecA recognizes the hydrophobic H-region of the signal peptide and translocates the protein to the cell membrane, in which the SecYEG translocon acts as a channel, guiding the protein to the periplasmic space. After the signal peptidase (SPase) recognizes and cleaves at the cleavage site (AXA) in the C-region of signal peptide, the protein then correctly folds in the periplasm (Figure 19A) (3-5).

Based on our previous work, to better understand the sequence features of LMT, we used ProtParam to calculate the positive charge intensity, hydrophobicity, and polarity of the LMT sequence, thereby dividing the LMT signal peptide into N-region, H-region, and C-region (Figure 19B). It should be noted that these region partitions are not absolute but are based on a qualitative analysis of the LMT sequence features.

Figure 20 Probe the key secA and secB genes in E. coli DH5α via colony PCR. DNA gel electrophoresis of colony PCR products to probe secA (A) and secB (B). Target bands, 2706 bp and 468 bp, can be observed at the position around 3000 bp and 500 bp, respectively. Specific primers were designed according to the genome sequence of DH5α (GCF_002899475.1).

Before investigating the influence of some specific residues in this signal peptide on protein (here, sfGFP) secretion, we firstly confirmed the presence of the key genes secA and secB of the Sec pathway in the chassis E. coli DH5α via colony PCR (Figure 20), which set the foundation of downstream works and characterizations. Two specific variants of the N-terminal positively charged region (presumed to be the SecB binding region) and the C-terminal AXA cleavage site (SPase recognition site) were then constructed, namely LMT Δ2-4 and LMT A19V, respectively (Figure 21A). Gibson Assembly was performed to construct the expression module (BBa_25G6VM6I and BBa_25CWGZSX, respectively) of these variants at pSB1C3 vector and positive transformants were selected and confirmed by colony PCR (Figure 21B) and sequencing.

Figure 21 Construction of the expression modules of LMT Δ2-4 and A19V. (A) Sequence alignments of LMT and two variants. Mutated sites were colored in black. (B) DNA gel electrophoresis of the colony PCR products of expression modules of LMT Δ2-4 and A19V fused with sfgfp at pSB1C3 in DH5α. Target bands, 1383 bp and 1374 bp, can be observed at the position between 1500 bp and 1000 bp, respectively.

The fluorescence intensity of the two variants, no matter for the culture or supernatant, was much lower than that of the wild-type (WT) LMT group (Figure 22B and 22C), although no significant growth defects due to the modifications were observed (Figure 22A). Characterization experiments revealed that both the reduction of positive charges in N-region (LMT Δ2-4) and mutation on SPase cleavage site (A19V) are “lethal” to the function of LMT signal peptide, which highlights the importance of critical interactions between transportation machinery of host and signal peptide during secretion process. Notably, the LMT Δ2-4 variant exhibited slightly higher culture and supernatant fluorescence intensity compared to A19V. Although the Δ2-4 variant deletes a significant portion of the positively charged amino acids in N-region, it only reduces, rather than completely blocks, the binding of SecB to the signal peptide. Therefore, the signal peptide may retain some degree of Sec pathway transport capability. However, the A19V variant disrupts the crucial AXA cleavage site, significantly impairing the specific recognition and cleavage by the SPase. This could even block of the entire Sec secretion pathway′s function, thereby greatly inhibiting the secretion of the cargo protein. Additionally, the folding of cargo sfGFP might be interfered due to these two modifications, which finally resulted in the lower fluorescence intensity of bacterial culture.

Figure 22 Characterization of secretion performance of LMT and two variants. The OD₆₀₀ (A), fluorescence intensity of culture (B) and supernatant (C) was measured as time progressed. Relative fluorescence units (RFU) was defined as the fluorescence of bacterial culture subtracted the background fluorescence of growth medium.

6.2 Targeted mutations and screening of high-performance variants

The importance of the N-terminal positively charged region and the AXA cleavage site of LMT signal peptide was validated. Based on the mechanistic understanding, and inspired by the high-throughput experiments and analysis by Gresso et al. (3), we planned to design and construct a series of sequence mutants (especially mutations in H-region and C-region) for the purpose of screening variant(s) with higher-performance on protein secretion.

6.2.1 Enhancing the hydrophobicity of the H-region of LMT signal peptide

Extensive literature reviews indicated that one of the crucial factors for signal peptides to successfully mediate protein secretion is their overall hydrophobicity, especially the hydrophobicity of the H-region core (3-5). This hydrophobicity is directly correlated with the efficiency of the signal peptide in traversing the hydrophobic biological membrane. Aiming to strengthen the interactions between the cell membrane and signal peptide to promote protein secretion, we focused on enhancing the hydrophobicity of the LMT signal peptide's H-region.

Figure 23 Targeted mutations in H-region of LMT signal peptide. (A) Sequence alignments of LMT and 4 variants. Mutated sites were colored in black. (B) DNA gel electrophoresis of the colony PCR products of expression modules of LMT variants fused with sfgfp at pSB1C3 in DH5α. Target bands (1383 bp) can be observed at the position between 2000 bp and 1000 bp. (C) Each variant's fluorescence intensity of supernatant of 10 h was normalized to LMT WT, for the first characterization. (D) Fluorescence intensity of supernatant of M13L variant with LMT WT was measured as time progressed, for the second characterization.

Thus, we designed 4 LMT variants: G12L, M13L, C14L, and A15L, which replaced the residue in H-region with the hydrophobic leucine (L) residue (Figure 23A) and constructed corresponding expression modules (Figure 23B) using the same method mentioned above. For the first time of characterization, two variants, M13L and C14L showed higher fluorescence intensity of supernatant than LMT (Figure 23C). Given the poor reproducibility observed in C14L group, we evaluated M13L only for the second time of characterization with LMT as control. However, as time progressed, the M13L variant presented a little bit lower fluorescence intensity of supernatant than the wild-type (Figure 23D), inconsistent with the previous observation.

It is evident that not all mutations that enhance hydrophobicity will certainly improve secretion efficiency. In addition, there are other hydrophobic residues to be tested. An optimal range of hydrophobicity might exist, and exceeding this range may negatively affect cargo protein's folding, transmembrane transportation, or other crucial steps. Taking together, the strategy of enhancing the hydrophobicity of the H-region core might not be so feasible and robust for the case of LMT signal peptide. Therefore, we turned to consider the mutations in the C-region of LMT.

6.2.2 Increasing the number of Q residues in C-region of LMT signal peptide

Figure 24 Construction of the expression modules of the variants increasing the number of glutamine residues in the C-region. (A) Sequence alignments of LMT and four variants. Mutated sites were colored in black. (B) DNA gel electrophoresis of the colony PCR products of four variants' expression modules fused with sfgfp at pSB1C3 in DH5α. Target bands (1383 bp) can be observed at the position between 1500 bp and 1000 bp.

Demonstrated through high-throughput experiments, the number of glutamine residues (Q) in the C-region of most signal peptides is positively correlated with secretion efficiency (3). Therefore, we designed to increase the number of glutamine residues in the C-region of LMT and constructed four C-terminal point mutants (Figure 24), named LMT A15Q, LMT S16Q, LMT A17Q, and LMT F18Q. In characterization, none of the variants impaired the cell growth (Figure 25A). However, even though some variants presented higher culture fluorescence intensity than the wild-type LMT (Figure 25B), all variants showed much lower fluorescence intensity of supernatant than LMT (Figure 25C), which means no improvements of these variants in secretion performance. Especially, A17Q mutation disrupts the AXA cleavage site like A19V variant, thus resulting in the same “abortive” phenomenon of fluorescence intensity. With the failure experience of increasing the number of glutamine residues in the C-region, we then turned to focus on the key factor for SPase cleavage efficiency-the composition of residues in C-region (3).

Figure 25 Characterization of secretion performance of LMT and the variants increasing the number of glutamine residues in the C-region. The OD₆₀₀ (A), fluorescence intensity of culture (B) and supernatant (C) was measured as time progressed.

6.2.3 Increasing the number of A residues in LMT signal peptide

It was reported that the number of alanine (A) residues at the C-terminus of the signal peptide might be closely related to the cleavage efficiency of SPase, thus influencing secretion efficiency (3). Therefore, we hypothesized that increasing the number of C-terminal alanine residues could effectively enhance SPase cleavage efficiency of the signal peptide, resulting in improved secretion performance. This time, we designed to increase the number of alanine residues in LMT and constructed three C-terminal point mutants, named LMT M13A, LMT C14A, and LMT F18A (Figure 26A and 26B). As time progressed, the M13A and F18A variants showed a little bit higher fluorescence intensity of supernatant than wild-type LMT, while it can be easily observed that the C14A variant's performance was superior to other variants and even the wild-type LMT (Figure 26C). To confirm whether the improvement was true or not, we conducted the characterization experiments of LMT C14A for the second time with LMT as control. Stronger fluorescence intensity of both bacterial culture and supernatant was observed for the C14A variant when compared to the wild-type LMT signal peptide (Figure 26D), validating the effectiveness of this mutation in improving secretion performance.

Figure 26 Increasing the number of alanine residues in LMT. (A) Sequence alignments of LMT and 3 variants. Mutated sites were colored in black. (B) DNA gel electrophoresis of the colony PCR products of expression modules of LMT variants fused with sfgfp at pSB1C3 in DH5α. Target bands (1383 bp) can be observed at the position between 1500 bp and 1000 bp. (C) Fluorescence intensity of the supernatant was measured as time progressed. (D) Fluorescence intensity (at 10 h) of culture and supernatant of C14A and LMT WT were compared, for the second time of characterization. p-value: 0.0008 (***) and 0.0098 (**).

The successful construction and characterization of the high-performance C14A variant confirmed the hypothesis that increasing the number of C-terminal A residues might effectively enhance the function of signal peptide to direct proteins out of the bacterial cell. The results (Figure 26C and 26D) suggested that the C14A mutation might promote protein secretion through mechanisms such as boosting SPase cleavage efficiency or optimizing the signal peptide conformation. Thus, C14A represents a promising optimization site for the LMT signal peptide with potential application values. Future studies should focus on an in-depth exploration of the SPase mechanism and the structure-activity relationship of C14A.

6.3 Modeling and analysis of the secretion mechanism

With the help of the model implemented by XMU-China 2024 that describes the relationship between the amount of secreted recombinant protein and time in the natural state of our engineered bacteria, rather than the growth arrest state in pulse-chase experiments, the performance difference of C14A and LMT can be revealed at the microlevel quantitatively. By fitting the model with our experimental data, two important parameters, the translocation rate from cytoplasm space to periplasm space (α, 1/h) and the translocation rate from periplasm space to medium space (β, 1/h), are shown in Figure 27. A 24% increase in the translocation rate from cytoplasmic space to periplasmic space by C14A mutation proves that the improvement is attributed to the interaction between secretion machine in the Sec pathway and mutant, with the translocation rate from periplasm space to medium space showing non-significant changes. With the help of the model implemented by XMU-China 2024 that describes the relationship between the amount of secreted recombinant protein and time in the natural state of our engineered bacteria, rather than the growth arrest state in pulse-chase experiments, the performance difference of C14A and LMT can be revealed at the microlevel quantitatively. By fitting the model with our experimental data, two important parameters, the translocation rate from cytoplasm space to periplasm space (α, 1/h) and the translocation rate from periplasm space to medium space (β, 1/h), are shown in Figure 27. A 24% increase in the translocation rate from cytoplasmic space to periplasmic space by C14A mutation proves that the improvement is attributed to the interaction between secretion machine in the Sec pathway and mutant, with the translocation rate from periplasm space to medium space showing non-significant changes.

Figure 27 Model analysis for C14A and LMT. (A) The translocation rate from cytoplasmic space to periplasmic space. (B) The translocation rate from periplasmic space to medium space.

In summary, this mutant (LMT C14A, BBa_25LUSMUT) significantly improved the secreted amount of the target protein. This finding provides new strategies and tools for signal peptide engineering and research on the extracellular secretion mechanism of enzymes. In our subsequent work, we will apply it to the secretion process of chitosanase, aiming to further increase the production yield of COS.

7 Enzymatic Hydrolysis Products Used in Preservation Experiments

7.1 Circuit Construction and Characterization

We demonstrated that the chitosanase mutants SaCsn46A mutant 2 (SaCsn46A E151G D222G) and SaCsn46A mutant 60 (SaCsn46A D40G Q125R H234R) exhibit excellent performance, whereas the LMT C14A shows high secretion efficiency. Consequently, we attempted to validate the degradation capabilities of these chitosanase mutants when secreted into the supernatant by LMT C14A.

The targeted sequences BBa_2507E4TO, BBa_25B08G08, BBa_25PUHM7F were individually inserted in pET-28a(+) to construct circuits (Table 4), which were transformed into E. coli BL21(DE3) to express each enzyme. The positive transformants were confirmed by kanamycin, colony PCR (Figure 28), and sequencing.

Figure 28 DNA gel electrophoresis of the colony PCR products of BBa_2507E4TO (A), BBa_25B08G08 (B) and BBa_25PUHM7F (C)_pET-28a(+) in E. coli BL21(DE3). Target bands (1143 bp) can be observed at the position between 1000 bp and 2000 bp.

The engineered bacteria were cultured at 16°C, and supernatants were collected at 0, 5, 8, 12, 16 and 20 hours. SDS-PAGE analysis confirmed the successful secretion of fusion proteins into the supernatant (Figure 29 A-C).

Figure 29 SDS-PAGE analysis of LMT C14A-linker-SaCsn46A WT (A), LMT C14A-linker-SaCsn46A mutant 2 (B), and LMT C14A-linker-SaCsn46A mutant 60 (C). Target bands can be observed about 30 kDa.

7.2 Characterization of the DP of Degradation Products from SaCsn46A Mutants with LMT C14A

Meanwhile, the supernatant collected at 20 h was used for degradation experiments (see Experiments for details). After degradation, the supernatant was collected for TLC and HPLC analyses, following the same methods as described above. The TLC and HPLC results of the degradation products are shown in Figures 30 and 31, respectively.

The TLC results indicated that the chitosan degradation products catalyzed by the wild-type enzyme and its mutants produced apparent magenta spots. The COS in these degradation products exhibited relatively high degrees of polymerization (DP), making clear separation challenging (Figure 30).

Figure 30 TLC analysis of degradation products of SaCsn46A WT and its mutants. Lanes A-E correspond to: glucosamine (A) 9.95 mg/mL COS products (B) Degradation products of LMT C14A-linker-SaCsn46A WT (C) Degradation products of LMT C14A-linker-SaCsn46A mutant 2 (D) Degradation products of LMT C14A-linker-SaCsn46A mutant 60 (E).

To further verify the results, HPLC analysis was subsequently performed on the same samples. The HPLC results revealed that the degradation products were primarily concentrated as glucosamine and chitobiose (Figure 31), with concentration trends consistent with the earlier TLC findings. Therefore, the HPLC results provided more direct evidence for the functional changes of the mutants.

Figure 31 HPLC analysis of degradation products of SaCsn46A WT and its mutants. (A) Degradation products of SaCsn46A WT. (B) Degradation products of LMT C14A-linker-SaCsn46A mutant 2. (C) Degradation products of LMT C14A-linker-SaCsn46A mutant 60.

7.3 Preservation Experiments

After obtaining the degradation products of the SaCsn46A and its mutants, we subjected these products to membrane filtration sterilization. Subsequently, we treated fish samples with the sterile degradation products, stored the samples at 4°C, and counted daily counts of bacterial colonies. Figure 32 shows a comparison of bacterial colony counts among different groups on the second day, along with a significance analysis. Notably, the degradation products of the LMT C14A-linker-SaCsn46A mutant 2 exhibited more excellent antimicrobial activity.

Figure 32 Fish preservation performance of degradation products from LMT C14A-linker-SaCsn46A WT, LMT C14A-linker-SaCsn46A mutant 2, and LMT C14A-linker-SaCsn46A mutant 60. (A) Colony counts after direct treatment. (B) Comparison between the degradation products group and the LB control group on Day 2 post-direct-treatment.

8 Kill Switch

8.1 tyrS Overexpression System

tyrS is an essential gene in E. coli that encodes tyrosyl-tRNA synthetase. We have developed a regulatory system using the T5 promoter and tyrS, resulting in the composite component BBa_2567UPRQ, which was subsequently assembled into the expression vector pCA24N. In the uninduced state (OFF state), tyrS maintains cell viability by providing a low level of enzymatic activity through leaky expression. In the induced state (ON state), excessive expression of tyrS leads to an enzymatic overload, which results in cell death. The killing effect was characterized via a spot assay (see our Experiments for details), and cell viability was measured by colony-forming unit (CFU) counting—expressed as the ratio of the number of cells in the IPTG-treated group to that in the non-IPTG-treated group (survival rate).

We characterized this IPTG-inducible system using E. coli BL21-AI (Figure 33A) and selected positive transformants verified by colony PCR and sequencing (Figure 33B). After approximately 6 h of inoculation with or without IPTG (final concentration: 1 mM), CFUs were counted and survival rates were calculated. The CFU in the IPTG-treated group was significantly lower than that in the untreated group, which implied that IPTG-induced tyrS overexpression exhibited cytotoxicity (Figure 33C).

Figure 33 Characterization of IPTG-induced tyrS overexpression. (A) The gene circuit of IPTG-induced kill switch (BBa_2567UPRQ) on pCA24N vector. (B) Agarose gel electrophoresis of the colony PCR products of BBa_2567UPRQ in E. coli BL21-AI. (C) Cell viability was measured by CFU count and is displayed as a ratio of cells treated with IPTG to cells maintained without IPTG. p-value: 0.0005 (***).

Although we have confirmed the lethal effect of the tyrS overexpression system, we continue to optimize the induction conditions to enhance this effect further. A spot assay was conducted at 0 h, 2 h, 4 h, 6 h, and 8 h, with IPTG concentrations ranging from 0 mM to 1 mM, to assess tyrS cytotoxicity. The survival rate decreased significantly at 4 h and reached its nadir at 6 h. The relationship between survival rate and IPTG concentration exhibited a Hill curve pattern; notably, the survival rates at 0.5 mM and 1 mM IPTG were nearly identical.

Figure 34 Characterization of the tyrS overexpression system under different induction conditions. Cell viability was measured by CFU count and is displayed as a ratio of cells treated with IPTG to cells maintained without IPTG. (A) Survival rate over time. (B) Survival rate over IPTG concentration.

Meanwhile, the tyrS overexpression system was applied to and characterized in E. coli strains including BL21-AI, Top10, MG1655 and DH5α. IPTG-induced tyrS expression demonstrated significant cytotoxic effects across all tested strains, indicating the broad applicability of the tyrS overexpression system.

Figure 35 Characterization of IPTG-induced tyrS overexpression in different bacterial strains. p-values: 0.0002 (***) for ,.E.coil BL21-AI, 0.0024 (**) for E.coli Top10, 0.0001 (****) for E.coli DH5α, and 0.0002 (***) for E.coli MG1655.

8.2 CcdB/CcdA Toxin-Antitoxin System

The second kill switch system employs the ccdB gene, encoding a toxin that inhibits DNA gyrase (topoisomerase II) and consequently induces cell death (6). The expression of ccdB is regulated by the pCymR promoter and is induced by cuminic acid (7). To reduce cellular burden caused by leaky toxin expression, the antitoxin gene ccdA is expressed at a low level via a weak constitutive promoter. When cuminic acid is added, the expression of CcdB surpasses the neutralizing capacity of CcdA, effectively triggering cell death.

We characterized this cuminic acid-inducible system using E. coli Top10 (Figure 36A) and selected positive transformants, which were verified by colony PCR and sequencing (Figure 36B). Approximately 4 hours after inoculation with or without cuminic acid (final concentration: 0.1 mM), CFUs were counted and survival rates were calculated. The CFU in the cuminic acid-treated group was significantly lower than that in the untreated group, suggesting that cuminic acid-induced CcdB toxin exhibited cytotoxicity (Figure 36C).

Figure 36 Characterization of cuminic acid-induced CcdB toxin. (A) The gene circuit of cuminic acid-induced kill switch (BBa_25464RFQ) on pSB3C5 vector. (B) Agarose gel electrophoresis of the colony PCR products of BBa_25464RFQ in E. coli Top10. (C) Cell viability was measured by CFU count and is displayed as a ratio of cells treated with cuminic acid to cells maintained without cuminic acid. 0.0114 (*). (D) Bacterial culture at 4 h with and without the addition of cuminic acid.

8.3 Integrated Kill Switch System

We integrated these two IPTG and cuminic acid-inducible systems into the same plasmid (Figure 37A) and transformed into E. coli Trans10. Then, the positive transformants were selected and verified by colony PCR and sequencing (Figure 37B). After approximately 4 h of induction, the CFU count in the inducer-treated group was significantly lower than in the untreated group, suggesting that the two independent suicide switches effectively enhance biosafety (Figure 37C).

Figure 37 Characterization of IPTG and cuminic acid-induced kill switch. (A) The gene circuit of IPTG and cuminic acid-induced kill switch (BBa_25V3X0FU) on pSB3C5 vector. (B) Agarose gel electrophoresis of the colony PCR products of BBa_25V3X0FU in E. coli Trans10. (C) Cell viability was measured by CFU count.

8.4 Conclusion

To prevent engineered bacteria in the chitosanase production chain from contaminating other processes and posing biosafety risks, we designed two independent suicide switches. The first is a ccdB/ccdA toxin-antitoxin system regulated by cuminic acid, and the second is an IPTG-inducible system that triggers overexpression of the tyrS gene. Together, we integrated these two systems into a same plasmid, they can operate independently and provide mutual backup for each other, thus guarantee rapid cell death, keeping the engineered bacteria in a controllable state.

In summary, our modular verification and engineering iteration of the IPTG-inducible killing switch have contributed to the biosafety of our CRUSTA project (see our Proposed Implementation for details). We hope this design of biocontainment system will provide valuable experience to other iGEM teams and the synthetic biology community.

9 Conclusion

In our CRUSTA project, we designed and developed an innovative COS-based bio-preservation technology using principles of synthetic biology. This technology provides a green and efficient alternative to traditional chemical preservatives in the food industry. In our project, we achieved chitosan degradation via a secretion system, enhancing the antibacterial performance of the degradation products by 7.96-fold compared to the wild-type. In the aspect of production of chitosan oligosaccharides, the enzymatic approach decreases about 99.1% of CO₂ emissions and 96.4% of cost compared to the chemical one. Additionally, our method saves about 61.0% of CO₂ emissions and 71.7% of cost compared to the chemical production of traditional preservative, potassium sorbate, in the same functional performance. This project not only has the potential to propel technological innovation and catalyze industrial advancement in food preservation but also aims to introduce advanced concepts of synthetic biology and eco-friendly preservation. These initiatives are expected to improve preservation efficiency, elevate food safety standards, and enhance environmental sustainability, thus bolstering market competitiveness and fostering the sustainable growth of the green food industry and the circular economy.

10 Future Prospects

In the future, we aim to leverage AI conduct the learning step in the final stage of the DBTL cycle to obtain new LMT and chitosanase with enhanced efficacy. We aim to comprehensively characterize the synergistic efficiency of high-performance chitosanase mutants and the optimized LMT C14A secretion system, and explore the optimal reaction conditions (such as temperature, pH, and substrate concentration) that maximize the yield of bioactive COS. Concurrently, we aim to validate the model-predicted properties of novel chitosanase sequences (e.g., catalytic activity, thermostability) through systematic experimental testing, and conduct in-depth analysis of degradation products to clarify the correlation between COS molecular weight, polymerization degree, and antibacterial/preservation efficacy. Importantly, we are committed to expanding the application of COS-based preservation technology to more food scenarios (such as aquatic products, fruits, and vegetables) and optimizing the formulation of COS preservatives to enhance their stability and applicability. Ultimately, we dedicate ourselves to advancing the development of a convenient, efficient, and eco-friendly green preservative product, providing practical solutions for the industrialization of natural biological preservation technology.

References

1. S. Jia, H. Hong, Q. Yang, X. Liu, S. Zhuang, Y. Li, J. Liu, Y. Luo, TMT-Based Proteomic Analysis of the Fish-Borne Spoiler Pseudomonas psychrophila Subjected to Chitosan Oligosaccharides in Fish Juice System, Food Microbiol 90, 103494 (2020).
2. X.Y. Wei, W. Xia, T. Zhou, Antibacterial activity and action mechanism of a novel chitosan oligosaccharide derivative against dominant spoilage bacteria isolated from shrimp Penaeus vannamei, Lett Appl Microbiol 74, 268-276 (2022).
3. S. Grasso, V. Dabene, M.M.W.B. Hendriks, P. Zwartjens, R. Pellaux, M. Held, S. Panke, J.M. van Dijl, A. Meyer, T. van Rij, Signal Peptide Efficiency: From High-Throughput Data to Prediction and Explanation, ACS Synthetic Biology 12, 390-404 (2023).
4. S. Kaushik, H. He, R.E. Dalbey, Bacterial Signal Peptides- Navigating the Journey of Proteins, Front Physiol 13, 933153 (2022).
5. H. Owji, N. Nezafat, M. Negahdaripour, A. Hajiebrahimi, Y. Ghasemi, A comprehensive review of signal peptides: Structure, roles, and applications, Eur J Cell Biol 97, 422-441 (2018).
6. P. Bernard, M. Couturier, Cell killing by the F plasmid CcdB protein involves poisoning of DNA-topoisomerase II complexes, Journal of Molecular Biology 226, 735-745 (1992).
7. J. Choi Young, L. Morel, T. Le François, D. Bourque, L. Bourget, D. Groleau, B. Massie, B. Míguez Carlos, Novel, Versatile, and Tightly Regulated Expression System for Escherichia coli Strains, Applied and Environmental Microbiology 76, 5058-5066 (2010).