To achieve
The system uses Escherichia coli Nissle 1917 (EcN) as the chassis strain and integrates six functional modules, iteratively optimized through a
Figure 1. Schematic of the Design-Build-Test-Learn (DBTL) cycle
Neohesperidin is a natural flavonoid glycoside with significant antioxidant, anti-inflammatory, and anti-colorectal cancer activities. Its structure is formed from hesperetin through a three-step glycosylation pathway. We designed reconstruction of this pathway in E. coli by introducing three key enzymes:
This design aims to enable efficient and highly selective conversion from hesperetin to neohesperidin, overcoming limitations such as low purity and limited yield inherent in traditional extraction methods.
Figure 2. Biosynthetic pathway from hesperetin to neohesperidin
Gene sequences for VvRHM, UGT73B2, and Cm1,2RhaT were obtained from NCBI, codon-optimized for E. coli, and synthesized by GenScript. The pET28a(+) vector was used for expression. A polycistronic construct was assembled using the B0034 RBS and cloned into the plasmid via NdeI/XhoI double digestion. After sequence verification, the recombinant plasmid was transformed into E. coli BL21(DE3) competent cells via heat shock, yielding the engineered strain BL21-Neohes. Positive clones were selected on LB agar plates containing kanamycin.
Figure 3. Neohesperidin biosynthesis system. Panel A: plasmid map; Panels B–D: agarose gel electrophoresis of UGT73B2, Cm1,2RhaT, and VvRHM; Panel E: electrophoresis of the genetic circuit
When BL21-Neohes reached OD₆₀₀ ≈ 0.6, protein expression was induced with 0.5 mM IPTG. The culture temperature was lowered to 25°C, and 3 g/L hesperetin was added as substrate. After 48 hours of fermentation,
Further temperature gradient experiments revealed the highest yield at 30°C, outperforming 25°C and 37°C, suggesting an optimal balance between metabolic flux and protein folding.
Figure 4. Neohesperidin production by engineered bacteria induced with 0.5 mM IPTG and supplemented with 3 g/L hesperetin. Panel A: HPLC chromatogram of neohesperidin production; Panel B: neohesperidin levels in wild-type BL21 and engineered strain after 48 h fermentation at 25°C; Panel C: time-course profile of fermentation at 25°C for 48 h (Hes: hesperetin; Hes7Og: hesperetin-7-O-glucoside; Neohes: neohesperetin); Panel D: comparison of neohesperidin production after 24 h fermentation at different temperatures.
The experiment demonstrated efficient neohesperidin synthesis from hesperetin in E. coli, with good coordination among the three enzymes and a smooth biosynthetic pathway. However, hesperetin is expensive, and using it directly as a substrate would significantly increase therapeutic costs, hindering scalability. Therefore, we need to construct a biosynthetic pathway from
The tumor microenvironment in colorectal cancer is significantly acidic (pH ≈ 5.8–6.5), whereas the normal intestinal environment is near neutral (pH ≈ 7.0–7.4). We leveraged the endogenous E. coli acid-sensing promoter pcadBA, which is activated under low pH by the regulator CadC to drive downstream gene expression. We applied this system to control the expression of Cm1,2RhaT, enabling the final step of neohesperidin synthesis only under acidic conditions, thereby enhancing targeting and safety. Based on this concept, we first constructed an acid-responsive biosensor to verify the function of the pcadBA promoter, then placed pcadBA upstream of Cm1,2RhaT to regulate neohesperidin synthesis.
Figure 5. Acid-responsive control of neohesperidin synthesis. Panel A: schematic of the acid-responsive biosensor; Panel B: schematic of acid-induced neohesperidin synthesis.
The pcadBA promoter sequence was synthesized and inserted upstream of the red fluorescent protein (mRFP) via seamless cloning (In-Fusion HD Kit), constructing a reporter system pcadBA-mRFP. This was transformed into BL21 to generate an acid-responsive biosensor strain. Subsequently, pcadBA was cloned upstream of Cm1,2RhaT via seamless cloning, and VvRHM-UGT73B2 was inserted upstream, resulting in the construct pET28a-VvRHM-UGT73B2-pcadBA-Cm1,2RhaT.
Figure 6. Acid-responsive biosensor. Panel A: plasmid map; Panels B–D: agarose gel electrophoresis of CadC+RFP, pCadBA, and CadC; Panel E: genetic circuit diagram.
Figure 7. Acid-responsive neohesperidin synthesis system. Panel A: plasmid map; Panels B–C: agarose gel electrophoresis of pcadBA and CadC; Panel D: genetic circuit diagram.
The engineered strain was cultured in LB medium at pH 5.8 and pH 7.3, with mRFP fluorescence and OD₆₀₀ measured over time. Results showed a significant increase in fluorescence signal over time at pH 5.8, while expression was nearly undetectable at pH 7.3. Normalized fluorescence values showed significant differences (p < 0.01), confirming the high acid responsiveness of pcadBA. Additionally, low pH had limited impact on cell growth, sufficient to support product synthesis. We then applied the validated system to neohesperidin production.
Figure 8. Functional validation of the acid-responsive promoter pCadBA. Panel A: relative fluorescence intensity of pcadBA-mRFP strain under different pH conditions; Panel B: OD600 comparison under different pH conditions.
Figure 9. Neohesperidin production under different pH conditions induced by the low-pH promoter. Legend: cultures with OD600 = 1.2 were transferred to media of varying pH, supplemented with 3 g/L hesperetin, and neohesperidin production was measured every 12 hours over 48 hours.
We successfully constructed and validated an acid-responsive biosensor, confirming that pcadBA is specifically activated under acidic conditions with minimal background expression, demonstrating strong application potential. While this system enables tumor targeting, there remains a risk of patient discomfort or environmental gene transfer if probiotics are excreted in feces. Therefore, a biosecurity system is essential—engineered bacteria must be eliminable upon patient discomfort and must initiate a self-destruction program upon exposure to low temperatures after excretion to prevent ecological spread.
Hesperetin is a key precursor of neohesperidin, but its natural sources are limited. We designed a two-step enzymatic pathway from naringenin to hesperetin:
The S142V mutant was selected to significantly enhance selectivity for eriodictyol, suppressing non-specific methylation side reactions (e.g., formation of isosakuranetin) and ensuring product purity.
Figure 10. Biosynthetic pathway from naringenin to hesperetin
Genes for ThF3'H, CPR, and MpOMT S142V were codon-optimized, synthesized, and assembled into a polycistronic expression system on the pET28a(+) vector. The construct was ligated via NdeI/XhoI digestion and transformed into BL21(DE3), generating the engineered strain BL21-Hes. Correct assembly was confirmed by antibiotic selection and sequencing.
Figure 11. Hesperetin biosynthesis system. Panel A: plasmid map; Panels B–D: agarose gel electrophoresis of ThF3'H, CPR, and MpOMT; Panel E: genetic circuit diagram.
After induction, 20 g/L naringenin was added to the culture, followed by 24 hours of fermentation at 25°C. HPLC analysis showed that the engineered strain efficiently converted naringenin to hesperetin, reaching a concentration of
Figure 12. Hesperetin production by engineered bacteria induced with 0.5 mM IPTG and supplemented with 20 g/L naringenin. Panel A: HPLC chromatogram; Panel B: hesperetin levels in wild-type BL21 and engineered strain after 24 h fermentation at 25°C.
We successfully constructed a complete pathway from naringenin to hesperetin, validating its functionality in E. coli and enabling autonomous supply of a key intermediate. Currently, this pathway operates independently and has not yet been coupled with downstream glycosylation modules. Therefore, integration with the glycosylation system from cycle 1.1 is required to achieve integrated synthesis from
To enable neohesperidin synthesis from a single precursor (naringenin), we integrated the two pathways above: first converting naringenin to hesperetin via the ThF3'H/CPR/MpOMT system, followed by three-step glycosylation via the VvRHM/UGT73B2/Cm1,2RhaT system. This theoretically enables full-pathway "naringenin → neohesperidin" synthesis in a single strain, improving autonomy and production efficiency.
We constructed the plasmid pET28a-ThF3'H-CPR-MpOMT-VvRHM-UGT73B2-Cm1,2RhaT and transformed it into BL21(DE3) to generate the engineered strain. Antibiotic selection (kanamycin) ensured plasmid stability.
Figure 13. Naringenin to neohesperidin conversion. Panel A: plasmid map; Panels B–D: agarose gel electrophoresis of ThF3'H, CPR, and MpOMT; Panel E: genetic circuit diagram.
Under identical induction conditions, 20 g/L naringenin was added to the fermentation system. After 48 hours, HPLC analysis detected successful neohesperidin production at a concentration of approximately
Figure 14. Neohesperidin production by naringenin-converting engineered bacteria. Legend: comparison of neohesperidin levels in wild-type BL21 and engineered strains after 48 h fermentation at 25°C. Protein expression was induced with 0.5 mM IPTG, with 20 g/L naringenin added as substrate. BL21: wild-type E. coli BL21 (negative control); BL21-Hes: BL21 carrying the hesperetin production plasmid; BL21-Neohes: BL21 carrying the neohesperidin production plasmid; BL21-Hes-Neohes: BL21 carrying both plasmids.
We achieved, for the first time in E. coli, the complete biosynthetic pathway from naringenin to neohesperidin, validating the integrability of multi-step metabolic pathways. However, since neohesperidin is intended for CRC therapy, it should ideally be
To ensure patient-controllable clearance of engineered bacteria, we designed an in vivo inducible suicide system. The arabinose-inducible promoter pBAD (BBa_I13453) drives gene expression in the presence of arabinose. We placed the toxin gene MazF under pBAD control; MazF specifically cleaves mRNA, halting protein synthesis and inducing programmed cell death. Oral administration of arabinose by the patient would trigger bacterial suicide, enabling clinical intervention. First, we constructed an arabinose-responsive biosensor to validate pBAD function, then placed pBAD upstream of MazF to control bacterial survival.
Figure 15. Schematic of the in vivo suicide system
We first constructed a pBAD-RFP expression cassette, cloned into the pSB1A3 vector, and transformed into BL21 to obtain the arabinose-responsive biosensor strain. Correct assembly was confirmed by sequencing. Next, we synthesized the pBAD-MazF expression cassette, cloned it into pSB1A3, and transformed into BL21 to generate the in vivo suicide system strain. Sequence verification confirmed correct construction.
Figure 16. Arabinose-responsive biosensor. Panel A: plasmid map; Panel B: agarose gel electrophoresis of araC-pBAD; Panel C: genetic circuit diagram.
Figure 17. Arabinose-inducible suicide system. Panel A: plasmid map; Panel B: genetic circuit diagram; Panels C–D: agarose gel electrophoresis of araC-pBAD and MazF.
The arabinose-responsive biosensor strain was cultured in media containing varying arabinose concentrations. Results showed strong fluorescence at 0.05% arabinose, with signal intensity increasing with arabinose concentration, confirming effective pBAD responsiveness. Additionally, the in vivo suicide system strain was cultured without arabinose. Growth was nearly identical to wild-type E. coli, indicating no activation or growth impact in the absence of arabinose.
Figure 18. Fluorescent protein expression in engineered strains under different arabinose concentrations.
Figure 19. Arabinose-inducible suicide system. Panel A: growth comparison of engineered and control strains without arabinose; Panel B: growth of engineered strain under different arabinose concentrations.
The in vivo suicide system was successfully established under in vitro simulation conditions, showing rapid response and precise regulation, with promising clinical intervention potential.
However, bacterial excretion by patients could lead to environmental leakage, posing risks of biocontamination, potential immunogenicity, or horizontal gene transfer. Therefore, an
To prevent environmental gene pollution due to escape of engineered bacteria, we constructed an environmentally responsive in vitro suicide system. We used the cold-shock protein promoter pCspA, which is strongly activated at low temperatures (e.g16°C). The toxin gene MazF was placed under the control of pCspA, such that when engineered bacteria are excreted in feces and enter a cold external environment, the suicide program is automatically triggered, preventing ecological dissemination. To validate this concept, we first constructed a temperature-responsive biosensor to test the functionality of the cold-shock protein promoter pCspA, then placed pCspA upstream of MazF to control bacterial survival.
Figure 20. Schematic of the in vitro suicide system
The pCspA-RFP expression module was synthesized, cloned into the pSB1A3 vector, and transformed into DH5α to generate the temperature-responsive biosensor strain. Correct assembly was confirmed by sequencing. Subsequently, the pCspA-MazF expression module was synthesized, cloned into pSB1A3, and transformed into DH5α to generate the in vitro suicide system strain. Sequence verification confirmed successful construction.
Figure 21. Temperature-responsive biosensor. Panel A: plasmid map; Panel B: agarose gel electrophoresis of pCspA-mRFP; Panel C: genetic circuit diagram.
Figure 22. Cold-inducible suicide system. Panel A: plasmid map; Panel B: agarose gel electrophoresis of MazF; Panel C: genetic circuit diagram.
The temperature-responsive biosensor strain was cultured at different temperatures. Results showed strong fluorescence at 16°C, with signal intensity decreasing as temperature increased, confirming effective responsiveness of pCspA to environmental temperature.
Furthermore, the cold-inducible suicide strain was cultured under different temperature conditions. Results showed that at 16°C (compared to 37°C), the OD₆₀₀ of the strain was significantly lower, with almost no growth, indicating effective activation of the suicide system under low temperature. Additionally, when the cold-inducible suicide strain and wild-type strain were both cultured at 16°C, the OD₆₀₀ of the engineered strain showed almost no increase within 8 hours, while the control strain grew normally. This confirms the system’s efficacy in inducing bacterial suicide under cold conditions.
Figure 23. Fluorescent protein expression in the temperature-responsive biosensor strain at different temperatures.
Figure 24. Cold-inducible suicide system. Panel A: growth of the cold-inducible suicide strain over 8 hours at different temperatures; Panel B: growth curve comparison between the cold-inducible suicide strain and wild-type strain at 16°C.
We successfully constructed a temperature-dependent biocontainment system, endowing engineered bacteria with “environment-sensing and self-elimination” capability. The pCspA system requires no exogenous inducer and is suitable for environmental protection.
Next, this system will be integrated with the in vivo suicide system into the same chassis strain (EcN), forming a dual safety assurance system of “
Through systematic DBTL cycles, this project successfully constructed and validated functional modules, marking the development of a fully functional prototype of an intelligent engineered bacterial system for colorectal cancer therapy. The
In summary, this project has completed full-chain engineering validation from
