Our team has successfully developed an E. coli-based system for monitoring strawberry spoilage and implementing sterilization, demonstrating the innovative application of synthetic biology in the field of agricultural product preservation. This system was optimized through four rounds of the DBTL (Design-Build-Test-Learn) cycle, ensuring the reliability and practicality of the engineering outcome.

Initially, within the VOC-responsive system module, we constructed reporter gene circuits driven by various promoters. This enables the engineered bacteria to detect characteristic volatile organic compounds (VOCs)—such as 1-octanol, 1-octen-3-ol, and phenylethyl alcohol—released during strawberry spoilage. Detection triggers reporter gene expression, leading to the synthesis of visible violacein, thereby achieving early warning of spoilage. This DBTL iteration focused on optimizing promoter specificity and sensitivity.

Subsequently, in the sterilization and control module, we successfully expressed and purified active preparations of chitinase and glucanase in E. coli. When sprayed onto strawberry surfaces in vitro, these enzyme preparations effectively inhibit the growth of common spoilage fungi, such as Botrytis cinerea, demonstrating a green and safe antifungal effect. This DBTL cycle concentrated on the engineering optimization of enzyme expression and application validation.

The successful construction and validation of these modules not only provide an innovative solution for addressing postharvest strawberry spoilage but also establish a solid foundation for expanding the application of synthetic biology in agricultural product preservation and food safety. Through multiple DBTL iterations, we ensured the engineering success of the system, highlighting a closed-loop optimization process from design to practical application.

Figure 1 Engineering: Design - Build - Test - Learn

Design

This experiment aimed to detect the specificity of responses to Volatile Organic Compounds (VOCs). We designed genetic circuits responsive to VOCs, based on four different promoters, using the plasmid pSB1A3 as the vector. The recA promoter (PrecA) was designated to sense VOCs associated with DNA damage, the grpE promoter (PgrpE) to respond to cytotoxic VOCs, the soxS promoter (PsoxS) to detect VOCs related to oxidative stress, and the lasI promoter (PlasI) to recognize the quorum-sensing signal molecule 3OC12-HSL. All promoters were ligated upstream of the mRFP gene via restriction enzyme digestion and ligation, forming "promoter-mRFP" expression units. Consequently, upon VOC exposure, promoter activation drives mRFP expression, enabling the visual detection of VOCs through changes in fluorescence signal. Our design leverages the specific response mechanisms of these promoters, combined with the fluorescent reporting function of mRFP, to construct an engineered bacterial detection system capable of rapidly sensing multiple types of VOCs.

Figure 2 Construction and Detection of VOCs Response System

Build

The nucleotide sequences of the PrecA, PgrpE, PsoxS, and PlasI promoters were obtained via gene synthesis. Using the pSB1A3 plasmid as the backbone, the four promoters were individually cloned upstream of the mRFP gene using the Beyotime Seamless Cloning Kit to construct the recombinant plasmids. These recombinant plasmids were then transformed into E. coli DH5α and BL21(DE3) competent cells. Positive clones were selected on LB agar plates containing 100 μg/mL ampicillin. Correct construction was confirmed by sequencing (performed by Tsingke), resulting in the final engineered bacterial strains.

Figure 3 Agarose gel electrophoresis verification of PCR results

Test

Response of the Four Promoters to VOCs

To evaluate the ability of the engineered strains to recognize various VOCs, the bacteria were first cultured in LB medium at 37°C with shaking. The OD₆₀₀ was monitored every 30 minutes using a FlexStation 3 microplate reader to track growth. Once the culture reached an OD₆₀₀ of 0.6, 200 µL aliquots were transferred to a 96-well plate and co-incubated with a series of VOCs, including gradient dilutions of 1-octanol. After 3 hours of incubation, fluorescence intensity (Excitation 584 nm / Emission 607 nm) and OD₆₀₀ were measured simultaneously. The Fluorescence/OD₆₀₀ ratio was calculated to normalize for differences in cell density.

The results demonstrated that each engineered strain produced a significant fluorescent signal specifically in response to its corresponding VOC. The signal intensity increased with both VOC dosage and incubation time, indicating that the promoters were not readily induced by non-cognate substances. The growth curves of the induced cultures nearly overlapped with those of the blank control groups, confirming that the tested VOCs did not affect bacterial proliferation at the concentrations used. Based on these data, a dose-response heatmap was generated, which visually represents the specific response fingerprint of each promoter across the VOC matrix. This quantification provides a foundational basis for subsequent development of multiplexed detection systems.

Figure 4 Response of different reporter strains to various VOCs

Comparison of Response Fold-Changes Across Different Promoters to VOCs

The objective was to compare the response intensity of the four VOC-responsive systems, each constructed with a different promoter, in order to identify the most sensitive promoter to the VOCs released during the spoilage process. Through this experiment, we aimed to determine the optimal responsive element, thereby establishing a foundation for constructing the subsequent colorimetric reporter system. This will enable efficient detection and visualization of strawberry spoilage signals.

Figure 5 Characterization of the response characteristics of four promoters to VOCs at the 4-hour time point

(A) Comparison of induction fold changes for VOCs across groups (4h) relative to the control group without VOCs. Error bars represent ± standard deviation (n=3 biological replicates). Statistical significance was determined by one-way ANOVA followed by Tukey's post hoc test. (B) Evaluation of promoter response to VOCs. The number of signals is proportional to the induction fold change.

As the results show, the fluorescence fold-change measured after 4 hours of induction demonstrated that the grpE promoter exhibited the most outstanding performance in detecting the three target VOCs: 1-octanol, 1-octen-3-ol, and phenylethyl alcohol.

• When exposed to 1-octanol, the BL21-grpE-mRFP strain showed a significantly higher Fluorescence/OD₆₀₀ ratio compared to the control (CK) and strains with other promoters. It also achieved the highest response fold-change, indicating the strongest activation effect of grpE by this VOC.
• Under treatment with 1-octen-3-ol, grpE similarly demonstrated the strongest response, with both its fluorescence output and response fold-change significantly leading those of other promoters, confirming its high sensitivity to this VOC.
• For phenylethyl alcohol, while variations existed in the responses among the different promoters, grpE still maintained a high fluorescence level, demonstrating good response stability.

In contrast, the soxS promoter showed a weak response. Its Fluorescence/OD₆₀₀ ratio increased only marginally after treatment with most VOCs, indicating a low association between soxS and these compounds, making effective activation difficult. The lasI and recA promoters showed some response under specific VOC conditions, but their overall performance levels were markedly lower than that of grpE, indicating limited sensing capability.

In summary, the grpE promoter demonstrated the highest sensitivity and stability in responding to all three target VOCs, identifying it as the most suitable candidate to serve as the core sensing element for constructing the subsequent biosensor.

Response of the PgrpE Reporter Strain to a Mixed VOC

Based on the previous screening which identified the PgrpE promoter as having the best response intensity to individual VOCs, this experiment aimed to further evaluate the response characteristics of the PgrpE reporter strain to a mixed VOC environment simulating strawberry spoilage (with a phenylethyl alcohol : 1-octanol : 1-octen-3-ol ratio of 6:1:1). By measuring the response fold-change, we sought to validate the sensitivity and applicability of the PgrpE promoter within a complex VOC milieu, thereby providing experimental support for constructing an early detection system for strawberry spoilage.

Figure 6 Response fold of PgrpE reporter strains to the mixed VOCs at different time points. According to the composition distribution of VOCs, the ratio of phenylethyl alcohol:1-octanol:1-octen-3-ol in the VOC mixture is 6:1:1.

The results demonstrated that under treatment with the mixed VOCs, the induction fold-change (FC) of the PgrpE-mRFP reporter strain showed a sustained increasing trend over time (Figure 6). At 2 hours post-treatment initiation, the FC was approximately 1.3-fold, indicating the system's capability to detect VOC signals within a short timeframe. The FC increased to 2.6-fold at 4 hours, reached 3.2-fold at 6 hours, and stabilized at a high level of around 3.5-fold after 10 hours. This demonstrates that PgrpE exhibits a rapid response and a significant signal amplification effect to the mixed VOC signal, further supporting its selection as the superior promoter.

Therefore, based on testing with individual VOCs like 1-octanol and their mixtures, the fluorescence signal mediated by the grpE promoter was significantly stronger than those mediated by lasI, recA, and soxS, and it produced a sustained response to the mixture, confirming its status as a superior VOC-responsive promoter.

Learn

The experimental results confirm that our engineered bacterial system can rapidly and visually detect various Volatile Organic Compounds (VOCs). However, in practical applications, VOC concentrations and combinations are often more complex, and a single promoter may suffer from issues like cross-reactivity or insufficient sensitivity. Therefore, future work could involve collecting comprehensive fluorescence/OD₆₀₀ response matrices to a wider array of VOC stimuli. Integrating these datasets with machine learning algorithms could enable the development of predictive models for the accurate identification of VOC types and concentrations. Simultaneously, model feedback could guide the optimization of promoter combinations and expression systems, continuously enhancing the accuracy, sensitivity, and practicality of the detection system to provide a more reliable biosensing solution for VOC monitoring in complex environments.

Design

This experiment aimed to construct an engineered bacterium capable of producing violacein, providing a reliable color-based visual reporting system for the strawberry spoilage biosensor. We used E. coli BL21(DE3) as the chassis. We designed a plasmid containing the constitutive strong promoter PJ23100, the medium-strength Ribosome Binding Site (RBS) B0034, and the violacein biosynthetic gene cluster (vioAB and vioCDE). The constitutive expression of violacein leads to substantial accumulation of the purple pigment in the engineered bacteria. The effectiveness of the reporting system was confirmed via absorbance readings.

Build

During the construction process, we first performed codon optimization of the complete violacein biosynthetic gene cluster (vioABCDE) for E. coli and completed the full gene synthesis according to BioBrick standards. Subsequently, this optimized fragment was cloned into the high-copy-number plasmid vector pSB1A3 via EcoRI/XbaI restriction sites, successfully yielding the recombinant plasmid p1A3-vioABCDE. Correct construction was confirmed by restriction digestion verification and sequencing. This plasmid was transformed into E. coli BL21 competent cells, and positive clones were selected using ampicillin resistance. Initial screening was facilitated by the obvious purple colony phenotype. Finally, the verified correct engineered strain was prepared as a glycerol stock and stored at -80°C for future use.

Figure 7 Construction of Violette Plasmid

Test

We established a reliable method for the quantitative detection of violacein by plotting a standard absorbance curve at a wavelength of 570 nm. The production strain was cultured in M9 minimal medium supplemented with either glucose or glycerol as the carbon source. After a 24-hour incubation period, violacein was extracted separately from both the cell pellet and the supernatant using an ethanol-based extraction method combined with sonication. The yields under different culture conditions were then quantitatively compared. The results demonstrated that detection via absorbance at 570 nm effectively ensured its visualization capability.

Figure 8 Violet compound scan spectrum and standard curve preparation

Figure 9 Detection of violacein content in the produced bacteria

Learn

The genetic construct successfully synthesized violacein, thereby validating the functional effectiveness of the circuit. Notably, supplementing the culture medium with glycerol significantly enhanced violacein production compared to glucose, demonstrating superior performance and indicating that carbon source optimization is crucial for yield maximization. This engineered strain not only provides a functional visual reporting system but also establishes an optimized platform for product synthesis. Subsequent work will focus on replacing the constitutive promoter with a promoter activated by strawberry VOCs and, under glycerol-supplemented conditions, further enhancing the response amplitude of the sensor.

Design

This project aims to develop a biosensor for the rapid, sensitive, and visual detection of early-stage strawberry spoilage. We selected E. coli DH5α (Beyotime, D0351) as the chassis organism and constructed a genetic circuit comprising a spoilage-related Volatile Organic Compound (VOC) response promoter (PgrpE) and the violacein biosynthetic operon (vioABCDE). The target VOC mixture simulating spoilage consists of phenylethyl alcohol, 1-octanol, and 1-octen-3-ol in a 6:1:1 ratio. When the engineered bacteria sense the characteristic VOCs released during strawberry spoilage, the PgrpE promoter is activated, driving the expression of the downstream violacein operon. This results in a visible purple color change, reporting the spoilage signal and enabling rapid, on-site detection without the need for complex instruments.

Build

During the construction phase, we first successfully amplified the optimally screened VOC-responsive promoter PgrpE and all five genes required for violacein biosynthesis (vioABCDE) using PCR. Subsequently, employing molecular cloning techniques such as restriction digestion and ligation, we sequentially linked PgrpE with the violacein operon and inserted the assembly into the standard vector pSB1C3, yielding the recombinant plasmid pSB1C3-PgrpE-vioABCDE. This created the complete "sense-and-report" genetic circuit. The correctness of the recombinant plasmid was verified through colony PCR, double restriction enzyme digestion, and sequencing. The verified plasmid was then introduced into E. coli BL21(DE3) via heat shock transformation. Positive clones were selected using antibiotic resistance and prepared as glycerol stocks stored at -80°C. During expression condition optimization, it was discovered that replacing 0.4% glucose with 0.4% glycerol as the carbon source increased violacein production by approximately 1.9-fold. Therefore, M9 medium supplemented with glycerol was established as the optimal working medium for the sensor to enhance the output signal intensity.

Figure 10 Construction of PgrpE-volacein engineered bacteria

Test

Our team designed a systematic testing protocol: First, strawberry samples with varying degrees of spoilage (fresh, slight, moderate, severe) were prepared either by artificial inoculation with Botrytis cinerea or through natural spoilage. Cultures of the engineered strain, grown to the logarithmic phase, were co-incubated with the individual samples in a sealed container for 4-12 hours. A negative control was established using bacterial culture not exposed to any sample, and a positive control used culture exposed to a high concentration of a single VOC.

For signal readout, a qualitative assessment was performed by visually observing whether the bacterial culture developed a purple color. For quantitative analysis (planned), violacein will be extracted using ethanol and its absorbance measured at 575 nm. This data will be used to plot a relationship curve between the signal response and the degree of spoilage, thereby evaluating the sensor's detection range, sensitivity, and quantitative potential. Additionally, the sensor's specificity for strawberry spoilage VOCs will be validated by testing its response to volatile compounds from other common fruits and vegetables.

Figure 11 Strawberry Spoilage Detector Response Flowchart

Learn

Based on preliminary experimental results, this sensor is expected to effectively respond to strawberry spoilage markers. However, the current system has several limitations, and future improvements will focus on the following areas:

• Performance Enhancement: To improve response sensitivity and kinetic characteristics, we plan to employ directed evolution of the PgrpE promoter and introduce signal amplification circuits to shorten the detection time and lower the detection limit.
• Signal Readout: We will develop a portable colorimetric device or a smartphone image analysis algorithm to achieve quantification and standardization of the output signal, overcoming the subjectivity of visual interpretation.
• Application Robustness: Efforts will be made to enhance the environmental robustness of the engineered strain. We intend to use metabolic engineering or microencapsulation techniques to improve the stability of E. coli DH5α under real-world conditions.
• Specificity: Through transcriptomic analysis and the design of logic gate circuits capable of responding to two or more characteristic VOCs simultaneously, we aim to further enhance system specificity and reduce cross-reactivity.

Looking forward, we will advance the integration of the sensor into portable detection devices and explore its coupling with a bactericidal module to create a closed-loop "sense-report-act" system. Furthermore, leveraging the platform's extensibility, the sensing module can be readily replaced to quickly adapt the system for spoilage monitoring of other perishable fruits and vegetables, thereby broadening its application scenarios.

Design

While our detector successfully identifies strawberry spoilage, interviews with stakeholders (HP) revealed that detection alone is not a complete solution. To further reduce losses, we proposed a proactive method: spraying chitinase and glucanase onto strawberries before transport. These enzymes target and degrade the core structural polymers (chitin and glucan) of the fungal cell wall at the molecular level. This disruption causes structural collapse and functional loss of the cell wall, ultimately leading to fungal inhibition or death, thereby reducing strawberry spoilage caused by fungi.

Build

• The optimized chiA gene was amplified by PCR, digested with NdeI and XhoI restriction enzymes, and inserted into the corresponding site of the pET-28a(+) plasmid. This vector drives target gene expression via the T7/lac promoter, incorporates an N-terminal 6×His tag to facilitate subsequent protein purification and detection, and carries a kanamycin resistance (Kan^R) marker.
• Similarly, the optimized gluA gene was amplified by PCR, digested with BamHI and NotI restriction enzymes, and cloned into the corresponding site of the pCDFDuet-1 plasmid. This vector also regulates expression with a T7/lac promoter and carries a spectinomycin resistance (Spec^R) marker.
• Following digestion, the target gene fragments and plasmid backbones were ligated to construct the recombinant plasmids, which were then transformed into E. coli DH5α as the cloning host. The resulting clones were verified by double restriction enzyme digestion. Positive clones were selected for plasmid extraction and sequencing to confirm the correctness and integrity of the inserted fragments.

Figure 12 Construction of pET28a-Chitinase and pET28a-Glucanase vectors

Test

Validation of Chitinase and Glucanase Expression

To confirm the successful expression of the target proteins, our team conducted induced cultivation of the recombinant strains and performed qualitative protein analysis using SDS-PAGE (Sodium Dodecyl Sulfate-Polyacrylamide Gel Electrophoresis). Bacterial cells were harvested under the respective induction conditions, lysed, and soluble proteins were extracted for electrophoretic separation on a gel.

The results showed that specific protein bands corresponding to the theoretical molecular weights were observed in the experimental samples: PelB-ChiA exhibited a band at approximately 70 kDa, and PelB-GluB showed a band at around 65 kDa, both matching the expected sizes. No corresponding bands were detected in the empty vector control group, confirming that these bands originated from the expression of the target genes.

Figure 13 Verification of Chitinase and Glucanase Expression

Functional Validation of Chitinase and Glucanase Expression

This experiment aimed to validate the in vitro broad-spectrum antifungal activity of the purified Chitinase and β-1,3-Glucanase. Using the model fungus—Saccharomyces cerevisiae—as the target organism, we assessed the inhibitory effects on fungal growth for the individual enzymes and their combination. Since chitin and β-1,3-glucan are core structural components of the cell walls in the vast majority of fungi—including those causing strawberry spoilage such as Botrytis cinerea, Rhizopus spp., and Penicillium spp.—this experiment provides crucial functional evidence to support the subsequent application of these two enzymes on strawberry surfaces for the broad-spectrum control of various fungi responsible for postharvest spoilage.

Figure 14 Synergistic inhibition of S. cerevisiae by a combination of chitinase and glucanase.

Table 1. Fungistatic activity of chitinase and glucanase against S. cerevisiae.

(Note: CFU: Colony Forming Unit. The relative survival rate was calculated based on the viable count of Group 2. Statistical significance was determined by one-way ANOVA; p < 0.001 compared to Group 2.)

The quantitative results from the antifungal assay clearly characterized the inhibitory effects of the two enzymes individually and in combination (Figure 14, Table 1). Data from colony counting revealed:

• Negative Control (Group 2, Heat-inactivated enzymes): The yeast colony count was 7×10⁷ CFU/mL, set as the baseline survival rate (100%). This confirmed that non-active components in the enzyme preparation did not affect yeast growth.
• Chitinase Alone (Group 3): Reduced the viable yeast count to 1×10⁷ CFU/mL, corresponding to a relative survival rate of 14.3%. This demonstrates that chitinase alone can effectively degrade fungal cell walls and inhibit cell survival.
• Glucanase Alone (Group 4): Showed stronger antifungal activity, reducing the yeast colony count to 3×10⁶ CFU/mL, with a relative survival rate of only 4.3%.
• Enzyme Combination (Group 5): The most significant effect was observed with the dual-enzyme combination, which drastically reduced the viable yeast count to 1.4×10⁶ CFU/mL, yielding a very low relative survival rate of 2.0%. This result was significantly superior to either single-enzyme treatment group (p < 0.001), revealing a powerful synergistic antifungal effect between the two enzymes.

Learn

Through this experimental cycle, we successfully achieved high-yield expression in E. coli and obtained biologically active Chitinase and β-1,3-Glucanase. Functional validation using Saccharomyces cerevisiae as a model fungus demonstrated that the two enzymes, working synergistically, significantly reduced cell viability, confirming their excellent activity and application potential in degrading fungal cell walls.

These results provide crucial support for our subsequent application research. In the future, we plan to further evaluate the performance of the dual-enzyme system in practical strawberry preservation. This will involve testing its efficacy in inhibiting fungal infection through methods like spraying or coating on fruit surfaces. Concurrently, we will explore optimizations for enzyme stability and adjustments to the formulation ratio to achieve more durable and efficient protective effects. Through these refinements, we aim to integrate this module with our previously developed VOC detection system, ultimately constructing an intelligent, closed-loop "detection-response-inhibition" bioprotection system.