We engineered a multi-target probiotic chassis with three functional systems to enhance efficacy and ensure biosafety in Alzheimer’s disease intervention. System 1 (Sakuranetin production): Expression of F7-OMT boosted sakuranetin to 50.21 ± 2.67 mg/L, further enhanced to 45.21 ± 3.75 mg/L through YdaO and CysE co-expression, with BL21-Nar supplying 0.42 g/L naringenin as precursor. System 2 (LPC production): Overexpression of PLA2 yielded a clear protein band (~30–42 kDa) and increased LPC from 2.43 ± 1.28 μg/mL in controls to 74.23 ± 15.91 μg/mL. System 3 (Dual-layer kill switches): The arabinose-inducible pBAD–T4 lysis reduced OD600 to 0.19 at 20 h, while the cold-inducible pCspA–T4 lysis suppressed growth at 16 ℃ (OD600 0.34 vs. 0.83 in control). Together, these three systems establish a probiotic platform capable of high-level metabolite production with robust “user-triggered + environment-triggered” containment, balancing efficacy and biosafety.

Figure: three functional systems to enhance efficacy and ensure biosafety in Alzheimer’s disease intervention

We established a sakuranetin production system using naringenin as the substrate and progressively optimized it through three iterative generations. In the first-generation strain BL21-OMT, the introduction of F7-OMT significantly increased sakuranetin production from 0.65 ± 0.05 mg/L in the control to 49.46 ± 9.53 mg/L. Substrate optimization revealed that 150 mg/L naringenin yielded the highest production (50.21 ± 2.67 mg/L), and supplementation with 1 g/L methionine further enhanced production by approximately 55%. In the second-generation strain BL21-OMT-YdaO, co-expression of ydaO strengthened energy supply and the Met–SAM cycle, increasing production from 18.92 ± 2.58 mg/L to 31.07 ± 2.77 mg/L . The third-generation strain BL21-OMT-YdaO-CysE further introduced CysE to reinforce the cysteine-to-methionine pathway and boost SAM availability, resulting in an additional increase to 45.21 ± 3.75 mg/L. In parallel, BL21-Nar accumulated 0.42 ± 0.05 g/L naringenin during 48 h fermentation, providing a stable substrate supply for sakuranetin synthesis. Overall, through stepwise optimization across the first, second, and third generations (F7-OMT → ydaO → CysE), the system achieved more stable and efficient sakuranetin production, laying the foundation for future integration of endogenous naringenin supply with simultaneous conversion into sakuranetin in a single streamlined system.

Objective：

To establish and validate a high-performance liquid chromatography (HPLC) method for the qualitative and quantitative analysis of the substrate (2S)-naringenin and the product (2S)-Sakuranetin, providing a reliable analytical approach for subsequent yield optimization(Sun et al., 2022).

Method：

The F7-OMT gene (BBa_259M31WV, Generalbiol, China) was synthesized and codon-optimized for expression in E. coli. EcoRI, XbaI, SpeI, and PstI restriction sites were removed to meet the RFC#10 standard. The promoter J23100 and RBS B0034 were used as cis-regulatory elements, and the gene fragment was cloned into the pSB1A3 vector via XbaI and SpeI restriction sites. Competent E. coli BL21 (DE3) (D1009S) and DH5α (D0351) cells were purchased from Beyotime (China). The recombinant plasmid was first transformed into E. coli DH5α. Positive clones were selected on LB agar plates (1.5% agar) containing 100 μg/mL ampicillin (Amp) and verified by sequencing (Qingke, Beijing), generating the recombinant strain DH5α-OMT. The recombinant plasmid was then extracted using a plasmid extraction kit (DP103, Tiangen) and transformed into E. coli BL21 (DE3) to obtain the engineered strain BL21-OMT. Strains were stored at –80 ℃ in 25% (v/v) glycerol. The engineered strains were cultured at 37 ℃ and 150 rpm in LB broth (G3102, Servicebio, China) supplemented with 100 μg/mL ampicillin for inoculation and amplification.

Figure: Construction of BL21-OMT. (A) The plasmid map of pSB1A3-OMT. (B) Agarose gel electrophoresis of F7-OMT (1056 bp). (C) The gene circuit of BL21-OMT.

A 1:100 inoculum of the engineered strain BL21-OMT was transferred into LB broth containing 100 μg/mL ampicillin and cultured overnight at 37 ℃, 150 rpm for activation. Activated cells were inoculated into TB medium (BL1176A, Biosharp) supplemented with Amp. The culture was incubated at 37 ℃, 180 rpm for 3–4 h until OD600 = 0.6, followed by the addition of 150 mg/L naringenin (A06896, Inonochem). The culture temperature was then shifted to 25 ℃ and fermentation continued for 24 h. After fermentation, 1 mL of the culture broth was mixed with 1 mL methanol, centrifuged at 12,000 × g for 5 min, and the supernatant filtered through a 0.22 μm nylon organic-phase filter (110414017, BKMAN). Samples were analyzed by HPLC (Agilent 2100) equipped with an Agilent C18 column (4.6 mm × 250 mm, 5 μm), with detection at 290 nm. The mobile phases were water containing 0.1% TFA (phase A) and acetonitrile containing 0.1% TFA (phase B), with a flow rate of 1 mL/min, using the following gradient: 0–10 min: 10% → 60% B; 10–20 min: 40% → 80% B; 20–25 min: 80% → 10% B. At this stage, phase B was initiated at 40% (instead of 60%) to optimize the resolution of the target component and avoid peak overlap.

Figure：High performance liquid chromatography（HPLC）to detect naringenin and sakuranetin

Result：

HPLC analysis revealed two distinct and stable peaks at 12.4 min and 24.5 min, corresponding to the substrate naringenin and the product sakuranetin, respectively. Comparison with standards confirmed the peak identities, validating the accuracy of substrate conversion and product formation. Both peaks were sharp and symmetrical with stable baselines, indicating excellent separation performance, with no significant tailing or co-elution interference, ensuring reliable and reproducible detection results.

Based on the standard curve fitting formula, peak areas can be converted into concentration values, enabling quantitative analysis of naringenin consumption and sakuranetin production under various conditions.

Figure: HPLC chromatograms of naringenin and sakuranetin

Conclusion：

This HPLC method enables sensitive and stable detection of both naringenin and sakuranetin, accurately reflecting the dynamic changes in substrate consumption and product generation. It provides a reliable analytical tool for comparing sakuranetin production across different strains, substrate concentrations, and fermentation conditions.

Objective：

After a stable and reliable HPLC method was established, the effect of flavonoid 7-O-methyltransferase (F7-OMT) gene expression on sakuranetin production was evaluated. In this reaction, naringenin and S-adenosylmethionine (SAM) were converted into sakuranetin and S-adenosylhomocysteine (SAH) under the catalysis of F7-OMT.

Method：

The control strain BL21 and the engineered strain BL21-OMT carrying the F7-OMT gene were inoculated into TB medium supplemented with 150 mg/L naringenin and fermented at 25 ℃ for 24 h. After fermentation, culture samples were treated with methanol, centrifuged, and filtered through a 0.22 μm organic membrane filter. Product peak areas were then detected at 290 nm using the previously established HPLC method, and the sakuranetin concentration was calculated.

Result：

Sakuranetin production was compared between the engineered strain BL21-OMT and the control strain BL21. The control strain produced only a very low level of sakuranetin (0.65 ± 0.05 mg/L), whereas the BL21-OMT strain exhibited a markedly higher yield, reaching 49.46 ± 9.53 mg/L—an approximately 76-fold increase on average.

Figure：Sakuranetin production in BL21 and BL21-OMT strains after 24 h fermentation

Conclusion：

The expression of F7-OMT greatly enhanced sakuranetin production in the engineered strain, confirming its key role in sakuranetin biosynthesis. However, the current yield (49.46 ± 9.53 mg/L) is still insufficient for our needs; therefore, subsequent work will focus on optimizing experimental conditions to further improve sakuranetin production.

Objective：

Sakuranetin can be produced from naringenin under the catalysis of flavonoid F7-OMT. Since naringenin is an essential precursor, we aimed to investigate the effect of its concentration on sakuranetin production in the engineered strain BL21-OMT.

Method：

The engineered strain BL21-OMT was inoculated into TB medium supplemented with 100, 150, and 300 mg/L naringenin and fermented at 25 ℃ for 24 h. Sakuranetin concentration was measured using the same analytical procedure as described above.

Result：

Sakuranetin production by the engineered strain varied significantly under different naringenin concentrations (ANOVA, F = 21.09, p = 0.0019 < 0.05). The highest yield was obtained at 150 mg/L naringenin (50.21 ± 2.67 mg/L). At 100 mg/L, the yield was slightly lower (39.19 ± 4.84 mg/L) but not significantly different from the 150 mg/L group. In contrast, at the high concentration of 300 mg/L naringenin, sakuranetin production decreased significantly to 26.47 ± 5.45 mg/L.

Figure: Sakuranetin production of BL21-OMT under different naringenin concentrations

Conclusion ：

Substrate concentration had a significant effect on sakuranetin production in the F7-OMT engineered strain, with 150 mg/L naringenin giving the highest yield. Although no significant difference was observed between the 100 mg/L and 150 mg/L groups, this may be due to large variation within the group or the small sample size, which reduced statistical power. At 300 mg/L, the yield dropped significantly, likely because of substrate inhibition or toxic stress affecting cell metabolism and enzymatic conversion. Therefore, 150 mg/L naringenin was used as the standard culture condition for subsequent experiments on sakuranetin production.

Objective：

Building on the investigation of substrate concentration, we further evaluated the effect of exogenous methionine (Met) supplementation on sakuranetin production in the engineered strain BL21-OMT. Since Met is converted into S-adenosylmethionine (SAM) under the catalysis of MetK, and SAM serves as the methyl donor in the F7-OMT–mediated reaction, this experiment aimed to verify the promoting role of Met in sakuranetin biosynthesis.

Figure: Met-methionine is converted into S-adenosylmethionine (SAM) under the catalysis of MetK

Method：

The engineered strain BL21-OMT was inoculated into TB medium without methionine (control group) and with 1 g/L methionine supplementation (Met group), with the substrate concentration fixed at 150 mg/L naringenin. Sakuranetin concentration was determined using the same analytical method as described above.

Result：

Sakuranetin production was compared between the control group (without methionine) and the Met group (with 1 g/L methionine supplementation). With methionine supplementation, sakuranetin production in the engineered strain BL21-OMT increased significantly from 17.16 ± 1.67 mg/L to 26.56 ± 0.80 mg/L, representing an average improvement of approximately 55%.

Figure：Effect of 1 g/L methionine supplementation on sakuranetin production in BL21-OMT

Conclusion：

Methionine supplementation significantly increased sakuranetin production in the BL21-OMT strain, which is likely due to enhanced intracellular SAM availability that promotes F7-OMT–mediated methylation. Nonetheless, the extent of improvement may be constrained by regulatory factors in the methionine–SAM cycle or by competing pathways consuming SAM.

Objective：

Building on the first-generation F7-OMT system, a second-generation engineered strain carrying the ydaO gene was constructed to strengthen the Met–SAM metabolic cycle, thereby enhancing the synthesis and recycling efficiency of the methyl donor SAM and further improving sakuranetin production. The ydaO element is an ATP-sensing riboswitch from Bacillus subtilis that binds ATP and dynamically regulates downstream gene transcription, thus enabling feedback control of cellular energy status.

Figure: Role of the YdaO riboswitch in dynamic ATP homeostasis control.

Method：

The YdaO coding gene was synthesized (BBa_25Z8A0VX, Generalbiol, China) and codon-optimized for E. coli, with restriction sites incompatible with RFC#10 removed. The constitutive elements J23100-B0034 were used in cis, and the construct was cloned into the pSB1A3 vector via XbaI and SpeI restriction sites. In parallel, PCR and overlap extension were employed to place B0034-YdaO downstream of F7-OMT, generating a co-expression sequence of F7-OMT and YdaO, which was similarly cloned into pSB1A3 using XbaI and SpeI (Note:The YdaO operon that containing the ydaO riboswitch, vhb, and ptxD was introduced into the system. Hereafter, it is referred to simply as YdaO). As described above, recombinant plasmids were transformed into E. coli DH5α and BL21. Positive clones were screened on LB agar plates (supplemented with 100 μg/mL ampicillin and 1.5% agar) and verified by sequencing (Qingke, Beijing), resulting in the recombinant engineered strain BL21-OMT-YdaO. Strains were stored at –80 ℃ in 25% (v/v) glycerol as cryoprotectant. The culture conditions for the engineered strain were 37 ℃ and 150 rpm in LB broth (G3102, Servicebio, China) containing 100 μg/mL ampicillin for inoculation and expansion.

Subsequently, BL21-OMT and BL21-OMT-YdaO engineered strains were inoculated into TB medium containing 150 mg/L naringenin and fermented at 25 ℃ for 12 h. Sakuranetin concentration was measured using the same method as described above.

Figure: Construction of BL21-OMT-YdaO. (A) The plasmid map of pSB1A3-OMT-YdaO. (B) Agarose gel electrophoresis of YadO (125 bp). We hypothesize that the fragment may have migrated out of the expected range due to excessively high voltage during electrophoresis.(C) The gene circuit of BL21-OMT-YdaO.

Result：

This comparison evaluated sakuranetin production between the first-generation engineered strain BL21-OMT and the second-generation strain BL21-OMT-YdaO. The first-generation strain produced 18.96 ± 3.06 mg/L of sakuranetin, whereas the co-expression of ydaO in the second-generation strain significantly increased the yield to 36.10 ± 2.19 mg/L, representing an overall improvement of nearly 90%.

Figure: Enhanced sakuranetin production in BL21-OMT-YdaO compared with BL21-OMT.

Conclusion：

Co-expression of ydaO in the engineered strain BL21-OMT significantly enhanced sakuranetin production, nearly doubling the yield compared with the first-generation system. This demonstrates that strengthening the Met–SAM cycle through ydaO-mediated regulation is an effective strategy to improve methyl donor availability and optimize sakuranetin biosynthesis.

Objective：

Although the introduction of F7-OMT and ydaO improved sakuranetin production, the yield still did not meet our expectations. Therefore, we aimed to develop a third-generation optimization system. Given that reproducing the entire design of the third-generation system would be overly complex, we focused on the most critical component, the CysE gene, whose circuit design has been described in the project overview. CysE regulates the cysteine → methionine metabolic pathway and thereby indirectly influences the synthesis of the methyl donor SAM, offering a potential route to further enhance sakuranetin production.

Method：

To enhance methionine production in E. coli, the l-serine O-acetyltransferase coding gene CysE (BBa_25MTY54W, Generalbiol, China) was synthesized and codon-optimized for E. coli, with restriction sites incompatible with RFC#10 removed. The constitutive elements J23100-B0034 were placed upstream of CysE as cis-acting regulatory elements. Using PCR and overlap extension, the J23100-B0034-CysE fragment was positioned downstream of OMT-YdaO, generating a co-expression sequence of OMT, YdaO, and CysE, which was cloned into the pSB1A3 vector via XbaI and SpeI restriction sites.

Figure: Construction of BL21-OMT-YdaO-CysE. (A) The plasmid map of pSB1A3-OMT-YdaO-CysE (B) Agarose gel electrophoresis of cysE (819 bp).

As described above, the recombinant engineered strain BL21-OMT-YdaO-CysE was obtained. The strain was stored at –80 ℃ in 25% (v/v) glycerol as a cryoprotectant. The culture conditions for the engineered strain were 37 ℃ and 150 rpm. Inoculation and expansion were carried out in LB broth (G3102, Servicebio, China) supplemented with 100 μg/mL ampicillin.

Figure: The gene circuit of BL21-OMT-YdaO-CysE(From version 1 to version 3).

Subsequently, BL21-OMT, BL21-OMT-YdaO and BL21-OMT-YdaO -CysE were inoculated into TB medium containing 150 mg/L naringenin and fermented at 25 ℃ for 12 h. Sakuranetin concentration was measured using the same method as described above.

Result：

With the optimization of the metabolic pathway, sakuranetin production was progressively enhanced. The first-generation BL21-OMT strain yielded 18.92 ± 2.58 mg/L; the second-generation BL21-OMT-YdaO system showed a significant increase to 31.07 ± 2.77 mg/L (p < 0.01); and the third-generation BL21-OMT-YdaO-CysE system further improved the yield to 45.21 ± 3.75 mg/L (p < 0.001), representing an increase of approximately 139% and 45% compared with the first- and second-generation systems, respectively.

Figure: Comparison of sakuranetin production among different engineered strains (BL21-OMT, BL21-OMT-YdaO, and BL21-OMT-YdaO-CysE)

Conclusion：

Introduction of the CysE gene in the third-generation system significantly enhanced endogenous methionine biosynthesis, thereby increasing the intracellular supply of SAM. As a result, the engineered strain achieved its highest sakuranetin production in the third-generation system, reaching 45.21 ± 3.75 mg/L. Co-expression of F7-OMT, YdaO, and CysE effectively enhanced sakuranetin production, providing a valuable strategy for the future optimization of sakuranetin biosynthetic pathways.

Objective：

In the previous experiments, exogenous naringenin supplementation was required for sakuranetin production. To overcome this limitation, we aimed to engineer E. coli to endogenously produce naringenin, thereby reducing dependence on external substrate supply and supporting more efficient sakuranetin biosynthesis.

Figure: Endogenous biosynthetic pathway for naringenin production.

Method：

The coding genes TAL, 4CL, CHS, and CHI (BBa_25886WFT, Generalbiol, China) were synthesized and codon-optimized for E. coli, with restriction sites EcoRI, XbaI, SpeI, and PstI removed to comply with RFC#10 standards. The constitutive promoter–RBS element J23100-B0034 was used to drive expression of TAL/4CL and CHS/CHI separately to avoid interference between dual promoters. A B0015 terminator sequence was added to both operons. The synthetic sequences were cloned into the pSB1A3 vector via XbaI and SpeI restriction sites. As described above, the recombinant engineered strain BL21-Nar was obtained. The strain was stored at –20 ℃ in 25% (v/v) glycerol as a cryoprotectant. Culture conditions for the engineered strain were 37 ℃ and 150 rpm. Inoculation and expansion were carried out in LB broth (G3102, Servicebio, China) supplemented with 100 μg/mL ampicillin.

Figure: Construction of BL21-Nar. (A) The plasmid map of pSB1A3-Nar. (B) Agarose gel electrophoresis of TAL (654 bp). (C) Agarose gel electrophoresis of 4CL (1710 bp). (D) Agarose gel electrophoresis of CHS (1167 bp). (E) Agarose gel electrophoresis of CHI (666 bp). （F）The gene circuit of BL21-Nar

A 100 μL aliquot of frozen BL21-Nar recombinant strain stock was inoculated into 5 mL LB medium supplemented with ampicillin (Amp+) and incubated overnight at 37 ℃, 150 rpm for activation. The overnight culture was transferred into 250 mL Erlenmeyer flasks containing 50 mL Amp+ LB medium, adjusted to an initial OD₆₀₀ of 0.1, and grown at 37 ℃, 180 rpm until OD₆₀₀ reached 1. Cells were harvested by centrifugation at 8000 × g for 5 min, then resuspended in M9 medium (A510881, Sangon Biotech) supplemented with 0.4% glucose (G6500, Innochem), 1 mM MgSO₄ (A40004, Innochem), 50 μM CaCl₂ (A01289, Innochem), 340 mg/L thiamine (BP892, Fisher), and 5 g/L CaCO₃ (A05990, Innochem). Cultures were incubated at 25 ℃, 180 rpm for 48 h.

For naringenin quantification, 1 mL of fermentation broth was mixed with 1 mL ethyl acetate, vortexed for 2 min, and centrifuged at 10,000 × g for 5 min. The upper organic phase was collected and dried overnight at 37 ℃. The residue was resuspended in 1 mL acetonitrile and analyzed at 290 nm using a microplate reader. Naringenin concentration was calculated from a standard curve (0–800 mg/L) prepared with authentic naringenin (A06896, Innochem).

For glucose quantification, 1 mL of fermentation broth was centrifuged at 10,000 × g for 5 min, and the supernatant was collected. Glucose levels were determined using a glucose assay kit (60408ES60, Yeasen, China). Samples were homogenized by ultrasonic disruption under ice-water conditions (power 300 W, 3–5 s per cycle, 30 s intervals, repeated 3–5 times) with 0.2–0.3 mL homogenization medium added. In a 96-well plate, 2.5 μL distilled water and 250 μL working solution were added to blank wells, 2.5 μL standard solution and 250 μL working solution to standard wells, and 2.5 μL sample with 250 μL working solution to sample wells. The plate was gently mixed and incubated at 37 ℃ for 10 min, after which absorbance was measured at 505 nm using a microplate reader. Glucose concentrations were calculated based on the standard curve.

Figure: Principle of glucose assay based on the GOD–POD coupled colorimetric reaction

Result：

Under fermentation conditions of 25 ℃ and 180 rpm, both naringenin concentration and glucose concentration were monitored over a 48 h period. In the recombinant strain BL21-Nar, naringenin production increased continuously from 0.0045 ± 0.0018 g/L at 0 h to 0.4214 ± 0.0474 g/L at 48 h. Meanwhile, glucose concentration decreased from an initial 39.88 ± 1.70 g/L to 15.48 ± 2.68 g/L, indicating a clear negative correlation between substrate consumption and product accumulation during the fermentation process.

Figure: Time-course of naringenin production and glucose depletion during fermentation

Conclusion：

In this study, a recombinant production strain capable of synthesizing naringenin was successfully constructed, demonstrating the feasibility of microbial naringenin biosynthesis. At present, this strain was developed as an independent system. Future work will focus on integrating it with the previously established sakuranetin production pathway, enabling a coupled system in which naringenin is both produced and simultaneously converted into sakuranetin.

We constructed an engineered strain expressing B. ovatus-derived PLA2 to promote the conversion of PC to LPC. Western blot analysis confirmed the successful expression of PLA2 in the engineered strain BL21-PLA2, with a distinct band observed at the expected molecular weight of PLA2 (≈30–42 kDa). ELISA results further showed that the control strain BL21 produced only 2.43 ± 1.28 μg/mL LPC, whereas the engineered strain BL21-PLA2 exhibited a significantly elevated LPC level of 74.23 ± 15.91 μg/mL. These findings demonstrate that overexpression of PLA2 was achieved at the protein level and effectively enhanced the strain’s capacity for substrate degradation and LPC biosynthesis, providing a novel strategy for targeting AD-related metabolic pathways.

Objective：

Building on the system in which sakuranetin alone served as the intervention molecule, we further explored whether synthetic biology could be leveraged to construct a multi-target intervention system, thereby providing more diverse and precise molecular tools for the prevention and treatment of AD. Mechanistic studies have shown that lysophosphatidylcholine (LPC) is a key metabolite generated from phosphatidylcholine (PC) through the catalytic action of phospholipase A2 (PLA2), and this pathway is particularly active in Bacteroides ovatus. Clinical data likewise indicate that both the abundance of B. ovatus in the gut and serum LPC levels are significantly reduced in AD patients compared with healthy controls, suggesting that the B. ovatus–PLA2–LPC metabolic axis plays an important role in AD pathogenesis.

Figure: Microbiota - derived lysophosphatidylcholine alleviates Alzheimer's disease pathology via suppressing ferroptosis(Zha et al., 2025).

Based on these findings, we designed an innovative strategy: engineering E. coli BL21 to express the PLA2 gene derived from B. ovatus (BL21-PLA2). Leveraging the strain’s high cultivability and biosafety, this probiotic chassis is intended to efficiently catalyze the conversion of PC to LPC in the intestinal environment, thereby elevating LPC levels, mitigating Aβ deposition, and ultimately delaying AD progression.

Figure：the conversion of phosphatidylcholine (PC) to lysophosphatidylcholine (LPC)

Method：

The phospholipase A2 (PLA2) coding gene (BBa_25ST8FLD, Generalbiol, China) was synthesized and codon-optimized for E. coli, with restriction sites EcoRI, XbaI, SpeI, PstI, NdeI, and XhoI removed to comply with RFC#10 standards and the cloning requirements of the pET28a(m) vector. The gene was cloned into pET28a(m) via the NdeI and XhoI restriction sites. The recombinant plasmid was then transformed into E. coli DH5α. Positive clones were selected on LB agar plates (supplemented with 1.5% agar) containing 100 μg/mL kanamycin (Kana) and verified by sequencing (Qingke, Beijing), generating the recombinant strain DH5α-PLA2. Recombinant plasmids were extracted using a plasmid extraction kit (DP103, Tiangen) and transformed into E. coli BL21 (DE3) to obtain the engineered strain BL21-PLA2. The strain was stored at –80 ℃ in 25% (v/v) glycerol as a cryoprotectant. Culture conditions for the engineered strain were 37 ℃ and 150 rpm, with inoculation and expansion carried out in LB broth (G3102, Servicebio, China) supplemented with 100 μg/mL kanamycin.

Figure: Construction of BL21-PLA2. (A) The plasmid map of pET-28a(m)-PLA2. (B) Agarose gel electrophoresis of PLA2 (843 bp). (C) The gene circuit of BL21-PLA2.

The PLA2-overexpressing engineered strain was inoculated into 50 mL LB medium supplemented with Kana and cultured overnight. On the following day, the bacterial culture was harvested and centrifuged at 10,000 × g for 1 min. The cell pellet was resuspended in PBS and disrupted by ultrasonication (150 W, 1 s on/3 s off cycles, for a total of 20 min). The lysate was transferred into a fresh centrifuge tube and used as the intracellular protein sample. Protein solutions were mixed with 5× reducing sample buffer at a ratio of 4:1 and denatured in a boiling water bath for 15 min. Proteins were separated on 15% SDS-PAGE gels at 120 V. Subsequently, proteins were transferred from the SDS-PAGE gel to a PVDF membrane under cold conditions. The membrane was blocked at room temperature with 5% non-fat milk in TBST for 1 h, followed by incubation at 4 ℃ overnight with a rabbit anti-PLA2 primary antibody (1:1000, e.g., Abcam or Cayman). After three washes with TBST (10 min each), the membrane was incubated at room temperature for 1 h with HRP-conjugated goat anti-rabbit IgG (H+L) secondary antibody (1:1000 dilution, A0208, Beyotime). The membrane was then washed three times with TBST (10 min each). ECL reagents A and B were mixed at a 1:1 ratio and applied to the washed PVDF membrane placed on absorbent paper to remove excess liquid. The membrane was then covered with the prepared ECL solution for 1 min to ensure complete coverage. Excess liquid was removed with absorbent paper, and the membrane was exposed using a chemiluminescence imaging system under preset parameters.

The BL21-PLA2 and BL21 were inoculated at a 1:100 ratio into 5 mL LB medium containing 100 μg/mL kanamycin and cultured overnight at 37 ℃, 150 rpm for activation. Approximately 1 mL of the activated E. coli culture was then transferred into 50 mL M9 medium and adjusted to an initial OD600 of 0.1. The M9 medium contained 1× M9 salts (A510881, Sangon Biotech), 0.2% (w/v) glucose, 1 mM MgSO₄, and 50 μM CaCl₂, with kanamycin maintained at a final concentration of 100 μg/mL. When the culture reached OD₆₀₀ ≈ 0.6 at 37 ℃ and 150 rpm, 0.5 mM IPTG and 1 mg/mL egg yolk lecithin (L305002, Aladdin) were added, and incubation was continued for 24 h.

After fermentation, 1 mL of culture broth was centrifuged at 10,000 × g for 10 min, and the supernatant was collected and filtered through a 0.22 μm hydrophilic PES membrane (BS-PES-45, Biosharp). The filtrate was diluted 100-fold with PBS. The concentration of lysophosphatidylcholine (LPC) was then determined using an LPC ELISA kit based on a competitive binding assay. A standard curve was generated by plotting absorbance values against the log₁₀-transformed concentrations of LPC standards. Absorbance was measured at 450 nm using a microplate reader, and LPC concentrations in the samples were calculated according to the standard curve.

Result：

Western blot analysis confirmed the expression of PLA2 in the engineered strain. A distinct band corresponding to the expected molecular weight of PLA2 (≈30-42 kDa) was detected in the PLA2-overexpressing strain. The intensity of the band indicated successful overexpression of PLA2, validating the efficiency of the expression system.

Figure: Western blot analysis of PLA2 expression in engineered strain

A standard curve was established using LPC standard solutions at different concentrations, yielding the following fitted equation:

A450 = –0.915 × log₁₀(concentration) + 4.051, R² = 0.98.

The results demonstrated a strong linear relationship between A450 and log₁₀(concentration) with a high degree of fit (R² = 0.98), indicating that the assay is accurate and reproducible within the experimental concentration range. As LPC concentration increased, the A450 value decreased linearly, consistent with the principle of competitive ELISA: higher LPC levels in the sample result in fewer enzyme-labeled molecules bound to the antibody, weaker color development, and thus lower A450 readings.

Comparative analysis further showed that the unmodified BL21 control strain produced only a very low level of LPC (2.43 ± 1.28 μg/mL), whereas the recombinant strain BL21-PLA2 exhibited a markedly higher LPC level of 74.23 ± 15.91 μg/mL (p < 0.01). These findings indicate that overexpression of the PLA2 gene greatly enhanced the engineered strain’s capacity for substrate degradation and LPC generation.

Figure: LPC detection and analysis of engineered strains. (A) LPC standard calibration curve; (B) Comparison of LPC production (μg/mL) between control strain BL21 and engineered strain BL21-PLA2.

Conclusion：

Western blot analysis confirmed successful expression of PLA2 in the engineered strain BL21-PLA2, which showed markedly increased LPC production in vitro compared with the control strain. These findings validate the feasibility of the B. ovatus–PLA2–LPC metabolic pathway and provide an engineering basis for enhancing LPC levels in the gut. Combined with our previous optimization of sakuranetin biosynthesis, this multi-target strategy offers a new approach for intervening in the progression of Alzheimer’s disease.

To improve biosafety, we designed a dual-layer suicide system with both artificial control and environmental triggering. In the pBAD–T4 lysis module, arabinose addition induced cell lysis within 4 h, and cell density dropped to 0.19 ± 0.16 at 20 h, showing almost no growth. In the pCspA–T4 lysis module, cell growth was strongly inhibited at 16 ℃, with OD600 reaching only 0.34 ± 0.08 at 8 h versus 0.83 ± 0.06 in the control. Together, these results show that the system allows artificial clearance when needed and automatic lysis at low temperature, effectively reducing the risk of engineered strain spread.

Objective：

To verify the effectiveness of the arabinose-inducible promoter pBAD, we linked mRFP downstream as a reporter gene and tested its induction under different L-arabinose concentrations. The experiment aimed to assess whether pBAD can achieve controllable and dose-dependent gene expression, thereby providing evidence for the reliability of the pBAD–T4 lysis suicide system.

Method：

The arabinose-inducible promoter pBAD (K808000), RBS B0034, and the red fluorescent protein mRFP were synthesized. The mRFP gene was placed downstream of pBAD-B0034. The mRFP sequence was codon-optimized for E. coli and modified to remove EcoRI, XbaI, SpeI, and PstI restriction sites to comply with RFC#10 standards (Genewiz, USA). The construct pBAD-mRFP was cloned into the pSB1A3 vector between the XbaI and SpeI restriction sites. The recombinant plasmid was transformed into E. coli DH5α, and positive clones were screened on LB agar plates containing 50 μg/mL ampicillin (Amp) and verified by sequencing (Qingke, Beijing), yielding the engineered strain. The engineered strain was cultured at 37 ℃ in LB broth for inoculation and expansion. Bacterial growth was monitored using a spectrophotometer (GenStar, China) by measuring optical density at 600 nm (OD600).

Figure: Construction of DH5α-pBAD-mRFP. (A) The plasmid map of pSB1A3-pBAD-mRFP. (B) Agarose gel electrophoresis of araC-pBAD (1210 bp). Since the pBAD fragment is relatively short, it was linked with the araC regulatory sequence. (C) The gene circuit of DH5α-pBAD-mRFP.

The arabinose-inducible reporter strain was inoculated at a 1:100 ratio into 5 mL LB medium containing 50 μg/mL ampicillin, with different concentrations of L-arabinose added. Cultures were incubated at 37 ℃ and 180 rpm in a shaking incubator. At designated time points, 200 μL of culture was sampled. Fluorescence intensity (excitation: 584 nm; emission: 607 nm) and OD600 were measured using a microplate reader (Multiskan GO). Normalized fluorescence was calculated as the ratio of fluorescence to OD600（Fluorescence/OD600）.

Result：

The experimental results were expressed as normalized fluorescence (Fluorescence/OD600) in arbitrary units (AU). The data showed that the activity of the pBAD promoter exhibited a clear dose-dependent response to exogenous L-arabinose concentration.

Under 0% arabinose, fluorescence remained at a low background level across all time points (0 h: 11.19 ± 0.56 AU; 2 h: 7.30 ± 0.52 AU; 4 h: 11.06 ± 0.63 AU; 8 h: 9.90 ± 1.98 AU), indicating that the pBAD promoter was inactive.

At 0.05% arabinose, fluorescence increased rapidly to 69.80 ± 5.25 AU at 2 h, reached 274.62 ± 77.33 AU at 4 h, and further rose to 451.47 ± 139.93 AU at 8 h, suggesting that even a low concentration could trigger significant induction.

At 0.2% arabinose, induction was more pronounced, with fluorescence reaching 524.11 ± 56.77 AU at 4 h and 829.13 ± 62.16 AU at 8 h.

At 0.5% arabinose, the induction was strongest, with fluorescence rising to 630.96 ± 44.09 AU at 4 h and exceeding 1089.19 ± 108.40 AU at 8 h, demonstrating the strong driving capacity of pBAD.

In summary, these results clearly indicate that the expression strength of the pBAD promoter increases with rising arabinose concentrations, showing rapid response and good reproducibility.

Figure: Dose- and time-dependent induction of pBAD promoter activity by L-arabinose

Conclusion：

This experiment confirmed the dose-dependent and strong inducible properties of the pBAD promoter. It maintained low basal expression in the absence of arabinose, while rapidly and robustly driving downstream gene expression upon arabinose supplementation.

Objective：

To enhance the biosafety of engineered strains, we designed an arabinose-inducible suicide system (pBAD–T4 lysis). In this system, T4 holin and T4 lysozyme were placed downstream of the pBAD promoter, enabling engineered bacteria to undergo self-lysis in the presence of exogenous L-arabinose, thereby allowing artificially controlled clearance when needed(Couse, 1968).

Figure: Design principle of the arabinose-inducible suicide system (replacement of mRFP with T4-lysis downstream of pBAD)

Method：

The T4 holin (BBa_254YB8C8) and T4 lysozyme (BBa_25L8X41M) (T4 lysis) coding genes were inserted downstream of pBAD-B0034. The T4 lysis sequence was codon-optimized for E. coli and restriction sites EcoRI, XbaI, SpeI, and PstI were removed to comply with RFC#10 standards (Genewiz, USA). The pBAD-B0034–T4 lysis construct was cloned into the pSB1A3 plasmid at the XbaI and SpeI restriction sites and subsequently transformed into E. coli DH5α.

Figure: Construction of DH5α-pBAD-T4 lysis. (A) The plasmid map of pSB1A3-pBAD-T4 lysis. (B) Agarose gel electrophoresis of T4 Holin (657 bp). (C) Agarose gel electrophoresis of T4 Lysozyme (492 bp). (D) The gene circuit of DH5α-pBAD-T4 lysis.

The arabinose-inducible suicide strain was cultured overnight, and 100 μL of the culture was inoculated into 5 mL LB medium containing 50 μg/mL ampicillin. After 4 h of cultivation, 0.5 mM L-arabinose was added, and the culture was incubated at 37 ℃ with shaking at 180 rpm. At different time points, 200 μL samples were taken to measure the OD600 of the cells.

Result：

The OD600 changes of the control group (DH5α) and the experimental group (DH5α/pBAD-T4 lysis) were as follows: at 0 h, both strains had similar initial densities of 0.36 ± 0.11 and 0.38 ± 0.08, respectively. At 2 h, the OD600 values increased to 0.67 ± 0.08 and 0.65 ± 0.11, both showing normal exponential growth. After the addition of L-arabinose at 4 h, the control group continued to grow to 1.11 ± 0.17, while the experimental group reached 0.86 ± 0.12, with no significant difference observed. Subsequently, at 8 h, the OD600 of the control group rose to 1.60 ± 0.15, whereas the experimental group dropped sharply to 0.29 ± 0.16. By 20 h, the control group further increased to 2.31 ± 0.22, while the experimental group remained at only 0.19 ± 0.16, indicating almost no growth. These results demonstrate that after induction at 4 h, the pBAD-T4 lysis engineered strain underwent significant cell lysis in response to L-arabinose, whereas the control strain maintained normal growth.

Figure：Growth dynamics of DH5α and pBAD–T4 lysis strains after L-arabinose induction at 4 h

Conclusion：

This study confirmed the effectiveness of the arabinose-inducible suicide system (pBAD–T4 lysis). Without induction, engineered and control strains grew normally. After L-arabinose addition at 4 h, the engineered strain underwent rapid self-lysis, with growth almost completely lost by 8–20 h. These results demonstrate that the pBAD–T4 lysis module functions as a reliable, controllable safety switch. Importantly, it allows the clearance of residual engineered bacteria in vivo through oral administration of L-arabinose, thereby reducing potential environmental and biosafety risks.

Objective：

In the previous experiment, we confirmed the controllability of the arabinose-inducible pBAD system and demonstrated its artificial killing function through the pBAD–T4 lysis module. However, artificial induction alone is not sufficient to ensure biosafety. Therefore, in this experiment, we tested the cold-inducible promoter pCspA at different temperatures to evaluate its expression activity and verify its potential as an environment-triggered safety switch for a cold-inducible suicide system.

Method：

The cold-inducible promoter pCspA (K4987003) was synthesized. The mRFP gene was placed downstream of pCspA. The mRFP sequence was codon-optimized for E. coli, and the EcoRI, XbaI, SpeI, and PstI restriction sites were removed to comply with RFC#10 standards (Genewiz, USA). Using pSB1A3 as the vector, the pCspA-mRFP construct was cloned between the XbaI and SpeI restriction sites. The recombinant plasmid was transformed into E. coli DH5α, and positive clones were screened on LB (Luria Bertani) agar plates containing 50 μg/mL ampicillin (Amp) and verified by sequencing (Qingke, Beijing), generating the engineered strain. The engineered strain was cultured at 37 ℃ in LB broth for inoculation and amplification. Bacterial growth was monitored by measuring optical density (OD) at 600 nm using a spectrophotometer (GenStar, China).

Figure: Construction of DH5α-pCspA-mRFP. (A) The plasmid map of pSB1A3-pCspA-mRFP. (B) Agarose gel electrophoresis of pCspA-mRFP (738 bp); since the pCspA promoter is too short, the mRFP reporter was placed downstream for validation. (C)The gene circuit of DH5α-pCspA-mRFP.

The overnight-activated cold-inducible reporter strain was inoculated into LB medium at a 1:100 ratio and cultured at different temperatures for 12 h under shaking conditions (180 rpm). A 200 μL aliquot of culture was collected, and fluorescence (excitation 584 nm, emission 607 nm) together with OD600 was measured using a microplate reader (Thermo Fisher Scientific, USA). The normalized fluorescence ratio (Fluorescence/OD600) was then calculated.

Result：

Under low-temperature conditions, the pCspA promoter strongly drove fluorescent expression. At 16 ℃, fluorescence rapidly increased from 32.77 ± 14.58 AU at 2 h to 168.20 ± 20.80 AU at 6 h, reaching 270.84 ± 19.36 AU at 12 h. At 25 ℃, induction was moderate, with fluorescence reaching 240.04 ± 14.51 AU at 12 h, slightly lower than at 16 ℃. In contrast, at 37 ℃, fluorescence remained at a low background level, reaching only 67.92 ± 22.24 AU at 12 h. These results clearly demonstrate that pCspA activity is significantly enhanced at lower temperatures, enabling strong induction under cold conditions while maintaining minimal basal expression at higher temperatures.

Figure: Temperature-dependent activation of pCspA promoter driving mRFP expression

Conclusion：

The cold-inducible promoter pCspA can strongly drive downstream gene expression under low-temperature conditions while maintaining minimal activity at 37 ℃. This “on at low temperature, off at high temperature” feature makes it an ideal environmental sensor, capable of triggering suicide modules when engineered bacteria accidentally enter external low-temperature environments to prevent dissemination, while remaining inactive in the gut (37 ℃) to preserve normal function. This provides a solid foundation for subsequent replacement of mRFP with T4 lysis.

Objective：

We previously confirmed that the cold-inducible promoter pCspA can effectively drive gene expression at low temperatures. To further assess its biosafety potential, we constructed the pCspA–T4 lysis suicide system and tested the growth of engineered strains at 16 ℃ to determine whether the system could trigger cell lysis under low-temperature conditions and achieve automatic clearance.

Figure: Design principle of the cold-inducible suicide system (replacement of mRFP with T4-lysis downstream of pCspA)

Method：

The T4 holin and T4 lysozyme (T4 lysis) genes were placed downstream of the pCspA promoter, codon-optimized, and modified to remove EcoRI, XbaI, SpeI, and PstI restriction sites to comply with RFC#10 standards. The recombinant fragment was cloned into the pSB1A3 plasmid and transformed into E. coli DH5α, generating the pCspA–T4 lysis engineered strain.

Figure: The plasmid map of pSB1A3-pCspA-T4 lysis.

The cold-inducible suicide strain and control strain were cultured overnight, and 100 μL of each was inoculated into 5 mL of LB medium supplemented with ampicillin (50 μg/mL). Cultures were incubated at 16 ℃ with shaking at 180 rpm. At designated time points, 200 μL samples were collected to measure bacterial OD600.

Result：

The OD600 changes of the control strain (DH5α) and the experimental strain (DH5α/pCspA–T4 lysis) were as follows : at 0 h, values were 0.21 ± 0.03 and 0.19 ± 0.05, respectively, with no significant difference between the two groups. At 2 h, the OD600 of the control strain increased to 0.36 ± 0.04, while the experimental strain remained at 0.22 ± 0.06. At 4 h, the control strain further increased to 0.60 ± 0.07, whereas the experimental strain reached only 0.31 ± 0.06. By 8 h, the control strain grew to 0.83 ± 0.06, while the experimental strain remained at 0.34 ± 0.08. These results indicate that under low-temperature conditions, the growth of the pCspA–T4 lysis engineered strain was significantly inhibited, with OD600 consistently maintained at low levels, while the control strain proliferated normally.

Figure：Growth inhibition of DH5α/pCspA–T4 lysis at 16 ℃ compared with wild-type DH5α

Conclusion：

The cold-inducible suicide system pCspA–T4 lysis effectively restricted bacterial growth at low temperatures, confirming its ability to trigger efficient self-lysis. Together with the arabinose-inducible pBAD–T4 lysis system, they form a complementary “dual-safeguard” design. The first layer, pCspA–T4 lysis, is activated by environmental signals and triggers cell lysis when engineered bacteria are exposed to low temperatures outside the host, preventing their survival and spread. The second layer, pBAD–T4 lysis, can be activated by adding arabinose, enabling researchers or users to deliberately eliminate engineered strains inside the host or during experiments. This combined strategy of “environment-triggered + manually controlled” ensures both stable metabolic function within the host and rapid clearance when needed, greatly reducing biosafety risks.

[1] Sun, Q., Gao, S., Yu, S., Zheng, P., & Zhou, J. (2022). Production of (2S)-sakuranetin from (2S)-naringenin in Escherichia coli by strengthening methylation process and cell resistance. Synthetic and systems biotechnology, 7(4), 1117-1125.

[2] Zha, X., Liu, X., Wei, M., Huang, H., Cao, J., Liu, S., ... & Zhang, C. (2025). Microbiota-derived lysophosphatidylcholine alleviates Alzheimer’s disease pathology via suppressing ferroptosis. Cell metabolism, 37(1), 169-186.

[3] Couse, N. L. (1968). Control of lysis of T4-infected Escherichia coli. Journal of Virology, 2(3), 198-207.

[4] https://registry.igem.org/parts/bba-k808000

[5] https://parts.igem.org/Part:BBa_K4987003