We successfully constructed engineered strains for genistein production, bacterial cellulose synthesis, and a blue light-induced suicide system. The genistein-producing strain, optimized through N-terminal truncation of LjIFS and adding OmpAL tag to CrCPR, achieved a yield of 176.97 ± 15.30 μM. The bacterial cellulose strain, expressing acsAB and acsCD, produced 177.70 ± 31.62 mg/L of extracellular cellulose, demonstrating excellent water retention and tensile strength in mask applications. Additionally, the blue light-induced suicide system, using the MazF gene, effectively triggered bacterial cell death under blue light. These results validate our systems’ functionality, with future improvements focused on optimizing efficiency and exploring broader applications.

Genistein is a natural phytoestrogen. It not only regulates skin moisture balance but also participates in the synthesis of collagen and elastin, helping to delay skin aging in menopausal women. To achieve the biosynthesis of genistein, we selected three key enzymes: cytochrome P450 reductase (CrCPR) from Catharanthus roseus, 2-hydroxyisoflavone synthase (LjIFS) from Lotus japonicus, and 2-hydroxyisoflavanone dehydratase (GmHID) from Glycine max. These three enzymes are essential for the conversion of naringenin to genistein.

In the first engineering phase, our main goal is to extract and test genistein production from the engineered E. coli BL21 strain. To this end, we designed and constructed the plasmid pSB-LjIFS-CrCPR-GmHID, integrating the genes of these three key enzymes into the plasmid and transforming it into E. coli BL21 for expression. In the presence of the substrate naringenin, the engineered strain will convert naringenin into genistein. Through this design, we aim to achieve efficient biosynthesis of genistein, verify the functionality of the engineered strain, and lay the foundation for future optimization of production yields.

We optimized the codons of the three key enzyme genes (LjIFS, CrCPR, GmHID) and outsourced their synthesis to a biotech company. The plasmid vector pSB1A3 was selected, and a polycistronic sequence was constructed. A dual promoter system was introduced, with one promoter driving the expression of IFS and CPR, while the other promoter specifically drives the expression of HID. Finally, the constructed plasmid pSB-LjIFS-CrCPR-GmHID was successfully transformed into E. coli BL21.

The results of the agarose gel electrophoresis clearly show that we obtained the gene fragments of the expected size (CrCPR~2142 bp; LjIFS~1554 bp; GmHID 957 bp).

Inoculate the engineered bacteria at a 1:100 ratio into 5 mL of LB medium (supplemented with 50 µg/mL ampicillin) and incubate overnight at 37°C. Wash once with PBS. Transfer the culture into 50 mL of M9 medium (supplemented with 10 g/L glucose and 0.5 mM naringenin). Adjust the initial OD600 to 1 and incubate at 30°C for 24 hours.

After fermentation, collect 1 mL of the bacterial culture and add an equal volume of ethyl acetate. Vortex for 2 hours. Centrifuge at 15,000 rpm for 20 minutes, let it stand for 5 minutes, and collect the supernatant organic phase. Transfer 200 µL of the supernatant into an HPLC vial and evaporate it to dryness in a 40°C oven for 3-5 hours. Resuspend the residue in 200 µL of methanol. Measure the absorbance at 250 nm using a microplate reader.

Figure 4 presents the standard curve for genistein, illustrating the linear relationship between absorbance and genistein concentration. The regression equation obtained is Y = 0.02631X + 0.3105 , with a correlation coefficient (R) of 0.9831, indicating a strong linear correlation. This standard curve can be used to accurately determine the concentration of genistein in unknown samples by measuring their absorbance.

Figure 5 shows the genistein concentrations in different bacterial strains. BL21 and BL21/pSB1A3 were used as blank and negative controls, respectively, while the experimental group consisted of the engineered genistein-producing strain BL21/LjIFS-CrCPR-GmHID. The results indicate that there is no significant difference in genistein concentration between the two control strains, BL21 and BL21/pSB1A3. However, the genistein concentration in the experimental group was significantly higher, reaching 1.97 ± 0.38 μM. This confirms that the engineered strain is capable of producing genistein. The detection of genistein in the control groups is likely due to experimental error or the absorbance of other compounds at 250 nm, which may have influenced the measurements.

We successfully constructed an engineered strain capable of producing genistein; however, the concentration of genistein synthesized by the strain was lower than expected. We conducted further research and exploration on how to increase the yield of genistein.

We discovered that the hydrophobic structure at the N-terminus of the IFS protein could cause protein aggregation, significantly reducing the yield of genistein. To mitigate the interactions between these hydrophobic structures, we decided to alter a segment of the DNA sequence. Our research identified that the first 21 amino acids at the N-terminus of LjIFS were the primary factors contributing to aggregation. To avoid affecting the expression of other genes while reducing this impact, we employed a substitution method, replacing these 21 amino acids with alanine. Alanine has minimal impact on protein structure and function, thus achieving our goal.

We optimized the codons of the three key enzyme genes (LjIFS, CrCPR, GmHID) and outsourced their synthesis to a biotech company. The N-terminal 21 amino acids of LjIFS were truncated and replaced with alanine, resulting in a truncated version of LjIFS(BBa_K5481004). In this modified system, segments of LjIFS were replaced while retaining the GmHID and CrCPR genes to enhance genistein biosynthesis. The constructed plasmid pSB-LjtIFS-CrCPR-GmHID was successfully transformed into E. coli BL21.

.

After cultivating the engineered strain at 30°C for 24 hours, we measured the concentration of genistein by detecting the absorbance at 250 nm.The bacterial culture was collected and treated with ethyl acetate to dissolve the genistein. After centrifugation, the organic phase supernatant was collected, concentrated, and resuspended. The absorbance at 250 nm was then measured using a microplate reader. Finally, we converted the results into the two sets of data shown in the figure below using a standard curve.

Experimental results showed that the genistein concentration of truncated LjIFS strain (LjtIFS) was significantly higher than the wild-type LjIFS strain, reaching 17.33 ± 2.37 μM. This indicates that truncating the N-terminal sequence of LjIFS and replacing it with alanine can significantly increase genistein production.

By truncating the N-terminal sequence of LjIFS, we successfully resolved the issue of IFS enzyme activity and increased the production of genistein.This method demonstrates the potential of optimizing biosynthetic yields through fine adjustments in protein structure.

The hydrophobic structure at the N-terminus of IFS may cause protein aggregation, which significantly affects IFS activity. To enhance the solubility and functionality of LjIFS, the N-terminus was modified by adding hydrophilic (KKK, HHHH) or hydrophobic tags.

We tested four variants of LjIFS: a truncated version without a tag, and three modified versions where the N-terminal 21 amino acids were replaced with hydrophilic tags—17A (BBa_K5481006), KKK (BBa_K5481007), and HHHH (BBa_K5481008). These gene variants were synthesized and used to construct recombinant plasmids. The resulting plasmids were transformed into E. coli BL21, and positive clones were selected on LB (Luria Bertani) agar plates containing 50 μg/mL ampicillin (Amp). The selected clones were then verified through DNA sequencing to confirm the successful construction of the engineered strains.

These engineered strains were cultured at 30°C for 24 hours, and the genistein concentration produced by each strain was measured.

The results demonstrate varying effects on genistein production among the different LjIFS variants. The truncated LjIFS with the 17A tag showed a decrease in genistein concentration, producing 11.35 ± 1.28 μM of genistein, compared to the truncated LjIFS(LjtIFS) baseline of 17.44 ± 5.62 μM. In contrast, the variants with KKK and HHHH tags exhibited significant increases in genistein production. Notably, the LjIFS strain with the KKK tag produced the highest concentration, reaching 130.66 ± 14.68 μM, representing a substantial improvement.These findings suggest that adding hydrophilic tags such as KKK and HHHH to the N-terminus of LjIFS enhances its solubility and activity. Among these, the KKK tag proved to be the most effective.

Based on these results, we selected the KKK-tagged LjIFS for further optimization and genistein production studies.By modifying the structure of LjIFS to improve its solubility and protein functionality, we successfully increased the concentration of genistein produced. A comparison of LjIFS with different tags revealed varying effects on protein modification. Based on these findings, we aim to apply a similar approach to CPR to enhance its stability.

Initially, we designed a project to use CrCPR in E. coli to convert naringenin into genistein. However, we discovered that this gene is difficult to stabilize and maintain activity in E. coli, potentially due to its hydrophobic N-terminal tail, which functions as a membrane anchor in plant cells. In E. coli, the absence of an endoplasmic reticulum may reduce protein solubility and activity. The specific method we decided to use is to employ OmpAL (bacterial outer membrane protein A) as the signal peptide to enhance the activity and stability of CrCPR, so that CrCPR can play its role more completely.

In this modified system, CrCPR were ligated to OmpAL tag while retaining the GmHID and LjIFS genes to enhance genistein biosynthesis. Finally, the constructed plasmid was successfully transformed into E. coli BL21.

Using the same method, we tested two versions of CPR—one with the OmpAL sequence and one without—under identical cultivation conditions. The final results demonstrated that the CPR with the OmpAL sequence significantly increased genistein production, reaching 176.97 ± 15.30 μM, compared to the lower yield from the CPR without the OmpAL sequence (Figure 11).

By introducing the OmpAL sequence , the efficiency of CPR in converting naringenin in E. coli has significantly improved. The increased output was in line with expectations, so we started building a second system

Bacterial cellulose has excellent water retention capacity, biocompatibility, and biodegradability, making it an ideal primary ingredient for facial masks. Therefore, we designed a second system specifically for the production of bacterial cellulose. acsAB and acsCD are key genes involved in bacterial cellulose synthesis. acsAB encodes the core protein of the cellulose synthase complex, responsible for polymerizing glucose molecules into cellulose. acsC forms pores to secrete cellulose, while acsD controls the crystallization of cellulose into nanofibers. By constructing a cellulose-producing strain containing acsAB and acsCD, we will test its cellulose yield. At the same time, we will use the cellulose produced to create facial masks and evaluate their water content and tensile strength.

We selected the synthetic cellulose synthase-encoding genes acsAB and acsCD and added them to the bacterium, and codon optimization was performed for E. coli. Multiple cis-trans sequences were constructed using pSB1A3 as the vector and J23100-B0034 as the promoter-RBS sequence. Transform the recombinant plasmid into E. coli BL21.

To assess cellulose synthesis, the engineered BL21 strain was inoculated into 100 mL of fresh LB medium and cultured at 30°C and 180 rpm for 48 hours. The 100 mL culture was collected and centrifuged at 8,000 rpm for 5 minutes. The supernatant (20 mL) was collected as the extracellular sample. The bacterial pellet was resuspended in an equal volume of pre-chilled PBS (pH 7.4), and then sonicated on ice to lyse the bacteria. Next, the lysate was centrifuged at 15,000 × g for 20 minutes to remove cell membranes and debris, and the supernatant was collected as the intracellular sample.

Both the 20 mL intracellular and extracellular samples were mixed with 20 mL of 4M NaOH and incubated at 80°C for 2 hours to dissolve non-cellulose components and precipitate the cellulose. After incubation, the samples were centrifuged at 15,000 × g at 4°C for 30 minutes to collect the cellulose pellet. The pellet was washed repeatedly with distilled water until the pH reached 7.0. The washed cellulose was dried at 60°C to a constant weight, and the dry weight was measured to determine the cellulose content.

The results demonstrated a clear difference in cellulose production between the strains. The strain expressing only acsAB produced 9.99 ± 3.25 mg/L of extracellular cellulose, while the strain BL21/acsAB-acsCD significantly increased extracellular production to 177.70 ± 31.62 mg/L. This confirms the critical role of acsCD in exporting cellulose out of the cell. In terms of intracellular cellulose, the strain expressing acsAB alone accumulated a higher amount (524.61 ± 96.39 mg/L), whereas BL21/acsAB-acsCD reduced intracellular cellulose to 208.97 ± 40.40 mg/L, further supporting the role of acsCD in facilitating cellulose export.

To test the moisture content of our final mask product and compare it with common masks on the market:First, the mask was folded and soaked in serum for 0.5 hours to allow it to fully absorb the serum. Then, it was weighed using a balance and the weight was recorded. After drying the mask until no surface moisture remained, a second weight was measured. The difference between the two readings was used to determine the mask’s moisture content. This method was repeated to measure the moisture content of conventional market masks. According to the results, the maximum water content of our generated masks is much higher than that of the common masks in the market, which is 5 times of their moisture content. This means that our masks do not need to add the water retention agents that are added to regular masks to achieve better water retention than they do.

After successfully producing the bacterial cellulose material required for our project, we conducted a tensile strength test to evaluate the mechanical properties of the bacterial cellulose-based mask. The testing method involved suspending a foam box filled with water, weighing approximately 2 kg, using multiple layers of tape wrapped around the box and then hanging the tape from the bacterial cellulose sheet. One end of the bacterial cellulose sheet was tied and secured to a wooden stick to hold the weight. A single sheet of bacterial cellulose, stretched to approximately the thickness of a 1-yuan coin (about 2 mm), was able to support the weight of the foam box filled with water for an extended period of time. This indicates that bacterial cellulose exhibits excellent tensile strength, making it a durable material suitable for mask production.

We successfully constructed a bacterial cellulose-producing strain. The acsAB gene effectively produces cellulose intracellularly, while the acsCD gene enables the secretion of cellulose to the extracellular space. The final cellulose yield met our expectations. In addtion,these findings confirm that the bacterial cellulose mask not only provides effective hydration but also possesses strong mechanical properties, further validating its potential as a robust and practical product in the skincare industry. Considering safety concerns, we need to design a suicide system next.

To test whether the established pDawn promoter and Pr promoter can activate the blue light-inducible system properly under blue light induction. To achieve this goal, we constructed a test system that places the red fluorescent protein (mRFP) gene downstream of the pDawn promoter with the aim of verifying the functionality of the blue light-inducible system.

To ensure the success of our genetic construct, we selected the pSB1A3 plasmid as the vector for cloning the pDawn promoter and the mRFP gene between the XbaI and SpeI restriction sites. The pDawn promoter is light-sensitive and can be activated under blue light to drive the expression of the mRFP gene. After successfully constructing the recombinant plasmid pDawn-mRFP, we transformed it into E. coli DH5α.

The transformed blue light-inducible reporter strain was inoculated at a 1:100 ratio into 5 mL of LB medium containing 50 μg/mL ampicillin. The test tubes were wrapped in aluminum foil to prevent light exposure, and the culture was incubated overnight in a shaker at 180 rpm and 37°C. The next day, 1 mL of the culture was collected and centrifuged at 10,000 rpm for 1 minute to pellet the cells, followed by one wash with PBS (pH 7.4). The cells were then transferred into 50 mL of fresh M9 medium (supplemented with 10 g/L glucose and ampicillin). A 5 mL aliquot of the culture was placed into 33 mm petri dishes, with the experimental group exposed to blue light in a light box, while the control group was shielded from light with aluminum foil. The cultures were incubated statically at 30°C for 12 hours. After incubation, 200 μL of the culture was sampled, and fluorescence (excitation wavelength 584 nm, emission wavelength 607 nm) along with OD600 values were measured using a microplate reader. The normalized fluorescence ratio (Fluorescence/OD600) was then calculated.

The results showed minimal fluorescence in both the BL21/pSB1A3 and BL21/pDawn strains, indicating low background signal. However, the BL21/pDawn-mRFP strain displayed a significant increase in fluorescence under blue light, with an average standardized fluorescence ratio of 702.03 ± 140.88, confirming successful activation of the pDawn promoter and strong mRFP expression. These findings demonstrate that the blue light induction system functions effectively.

During the experiment, we successfully constructed and validated the pDawn-mRFP system. The results showed that the blue light receptor system could activate the expression of the mRFP gene under appropriate conditions, proving the functionality of the blue light-induced system. Through this experiment, we verified the effectiveness of the blue light-induced system and provided a solid foundation for the subsequent development of the blue light-induced suicide system. In subsequent experiments, we replaced the red fluorescent protein mRFP with MazF, the main body of the suicide system, for further validation and application development.In the future, we plan to optimise the blue light induced suicide system and further test its stability and effectiveness under different conditions. Through further research, we hope to develop an efficient and reliable blue light-induced suicide system that will provide new solutions for biosafety and control applications.

the mRNA synthesis disrupting enzyme mazF gene was replaced with the red fluorescent protein mRFP and inserted into the blue light sensing promoter pDawn. The effect of bacterial suicide was detected by testing the OD600 value of bacteria.

The BL21/pDawn-MazF and BL21/pDawn-mRFP strains were constructed by transforming pDawn-mazF and pDawn-mRFP vectors into BL21 strains, respectively, by molecular cloning. The transformed strains were screened and validated to ensure successful transformation.

100 μL of the overnight culture was inoculated into 5 ml of LB medium containing ampicillin (50 μg/ml) and incubated for 2 hours at 37°C and 220 rpm with oscillation until the OD600 value reached 0.4 .The culture solution was poured into 33 mm Petri dishes and the dishes were irradiated with 488 nm blue light at 30°C. A 200 μL sample was taken at 2 hours, 4 hours, 6 hours and 12 hours to determine the OD600 value.

The results demonstrate distinct differences in the growth of the strains: BL21/pSB1A3 and BL21/pDawn-mRFP strains exhibited normal growth throughout the 12-hour period. At 12 hours, their OD600 values reached approximately 1.5, indicating healthy bacterial growth. In contrast, the BL21/pDawn-MazF strain showed a significant reduction in OD600 over time. At 2 hours, there was a slight decrease in growth compared to the control strains. By 6 hours, the OD600 values had dropped considerably, and at 12 hours, the OD600 was near zero, indicating almost complete bacterial death. The results demonstrated that the blue light-induced system connected to mazF can be induced by blue light, leading to bacterial cell death.

In the future, we will improve our light-induced suicide system in the following aspects：

Optimisation of blue light irradiation time and intensity: Determine the optimal irradiation time and intensity to ensure maximum expression of suicide genes and efficient killing of bacteria.

Stability of gene components: adjust the expression vector to improve the stability and expression efficiency of the gene in bacteria.

Cell protection mechanism: introduce protective genes to prevent non-target bacteria from being affected by blue light irradiation.

Medical field: for precise control of drug-resistant bacterial infections and reducing the use of antibiotics.

Agricultural field: used to control crop diseases and reduce the use of chemical pesticides.

Environmental protection: used to deal with harmful bacterial pollution in the environment and safeguard ecological safety.

Integration of multiple control elements: Combine other light-sensitive elements to develop a suicide system with multiple light controls to improve the flexibility and application range of the system.

Large-scale application testing: Conduct large-scale testing in real environments to evaluate the stability and practicality of the system.

Commercialisation and promotion: Collaborate with relevant companies to promote the commercialisation of blue light-induced suicide systems, providing a new approach to solving the problem of drug-resistant bacteria.

[1]Liu X, Li L, Zhao G R. Systems metabolic engineering of Escherichia coli coculture for de novo production of genistein[J]. ACS Synthetic Biology, 2022, 11(5): 1746-1757.

[2]Scott E E, Spatzenegger M, Halpert J R. A truncation of 2B subfamily cytochromes P450 yields increased expression levels, increased solubility, and decreased aggregation while retaining function[J]. Archives of Biochemistry and Biophysics, 2001, 395(1): 57-68.

[3]Hwang Y, Noh M H, Jung G Y. Recent advancements in flavonoid production through engineering microbial systems[J]. Biotechnology and Bioprocess Engineering, 2024: 1-14.

[4]Quattrone A, Czajka A, Sibilla S. Thermosensitive hydrogel mask significantly improves skin moisture and skin tone; bilateral clinical trial[J]. Cosmetics, 2017, 4(2): 17.

[5]Marbun N, Fhitriana S, Leny L, et al. Preparation and evaluation of moisturizing sheet mask of Lady Finger (Abelmoschus esculentus (L.) Moench) extract[J]. Journal of Drug Delivery and Therapeutics, 2022, 12(3): 14-19.

[6]Wang H, Zhou B. Development and performance study of a natural silk fiber facial mask paper[J]. Journal of Engineered Fibers and Fabrics, 2020, 15: 1558925020975756.

[7]Fan X, Bai X, Tan L, et al. A Transparent Additive‐Free Fibrous Facial Mask for Skin Moisturizing[J]. Macromolecular Rapid Communications, 2023, 44(17): 2300180

[8]Mazraeh D, Di Ventura B. Synthetic microbiology applications powered by light[J]. Current Opinion in Microbiology, 2022, 68: 102158.

[9]Chattopadhyay G, Bhasin M, Ahmed S, et al. Functional and biochemical characterization of the MazEF6 toxin-antitoxin system of Mycobacterium tuberculosis[J]. Journal of Bacteriology, 2022, 204(4): e00058-22.