Tetrathionate, a key biomarker for inflammatory bowel disease (IBD). However, due to tetrathionate's rapid degradation in the gut, we shifted focus to thiosulfate, creating a thiosulfate-specific sensor using the ThsS/R system. This biosensor was further optimized to enhance its sensitivity and accuracy. We tested its dose-dependence and specificity, demonstrating strong responses at higher thiosulfate concentrations and minimal interference from other sulfate-based compounds. Next, we incorporated lacZ downstream of the sensing system to test β-galactosidase activity using the ONPG method. This enabled us to visually detect the sensor’s activity, offering a clearer readout for thiosulfate detection.
For the therapeutic system, we engineered bacteria to secrete human epidermal growth factor (hEGF) . We measured the differences in intracellular and extracellular hEGF levels. Additionally, we assessed the impact of these engineered strains on 293T cell viability.
For the suicide system, we developed a mazE/mazF toxin-antitoxin system controlled by rhamnose. Growth tests showed normal growth in the presence of rhamnose, while its absence triggered cell death, confirming the system’s functionality. This ensures controlled regulation of engineered bacteria for future applications
The construction of the three systems—the sensing system, the therapeutic system, and the suicide system—follows engineering design principles.
Tetrathionate is an important biomarker for inflammatory bowel disease (IBD). To develop a biosensor for detecting tetrathionate, we designed and synthesized the TtrS (BBa_K5084000) and TtrR(BBa_K5084001) operon system, incorporating the red fluorescent protein mRFP1 as a reporter gene. This construct was cloned into the pSB1A3 plasmid. After plasmid extraction, it was successfully transformed into E. coli DH5α. The engineered strain, referred to as the tetrathionate biosensor(BBa_K5084003), was cultured in LB medium with ampicillin (Amp, 100 μg/mL) at 37℃ (Figure 1).
To assess the Tetrathiosulfate biosensor’s ability to detect different tetrathionate concentrations, we performed functional tests by inoculating the E. coli strain into LB media containing varying concentrations of tetrathionate (0 µM, 10 µM, 50 µM, 100 µM, 500 µM, and 1000 µM). After 8 hours of incubation at 37℃, we measured the fluorescence intensity (excitation at 584 nm and emission at 607 nm) and OD600 using a microplate reader. The normalized fluorescence ratio (Fluorescence/OD600) was calculated to evaluate the Tetrathiosulfate biosensor’s response.
The results demonstrated a clear increase in the normalized fluorescence ratio as tetrathionate concentrations rose. At lower concentrations (0 µM to 50 µM), the fluorescence ratio increased rapidly, indicating the Tetrathiosulfate biosensor’s high sensitivity in this range. Between 10 µM and 50 µM, the fluorescence ratio more than doubled, reflecting a significant response (p < 0.01). However, as the concentration increased beyond 100 µM, the fluorescence signal continued to rise but began to plateau between 500 µM and 1000 µM, suggesting that the Tetrathiosulfate biosensor’s response was nearing saturation at higher concentrations (Figure 2).
Despite strong in vitro results, literature indicates a key limitation: tetrathionate is rapidly metabolized by gut microbiota, reducing its stability as an IBD biomarker in vivo. This compromises the biosensor’s ability for long-term inflammation monitoring. To overcome this, we will shift to thiosulfate, a more stable biomarker, and redesign the biosensor for thiosulfate detection, aiming for improved performance in IBD monitoring applications.
Thiosulfate is an important biomarker for IBD. To develop a biosensor capable of sensitively detecting changes in thiosulfate concentration, we constructed a system that incorporates two key components, ThsS(BBa_K5084010) and ThsR(BBa_K5084011), designed to sense thiosulfate. The system uses the PphsA(BBa_K5084002) promoter to drive the expression of the red fluorescent protein mRFP as a reporter for thiosulfate detection. The construct was cloned into the pSB1A3 plasmid, forming the recombinant plasmid pSB-PphsA-mRFP. This plasmid was then transformed into E. coli DH5α (purchased from Takara), and the engineered strain was cultured in LB medium containing ampicillin (Amp, 100 μg/mL) at 37℃. The successfully engineered strain was named Thiosulfate biosensor-A(BBa_K5084012) (Figure 3).
To validate the function of Thiosulfate biosensor-A, we inoculated the strain into LB medium containing varying concentrations of sodium thiosulfate and incubated it at 37℃ for 8 hours. Fluorescence signals were measured at concentrations of 0 mM, 0.2 mM, and 1 mM sodium thiosulfate. After incubation, fluorescence intensity (excitation wavelength at 584 nm, emission wavelength at 607 nm) and OD600 values were measured using a microplate reader, and the normalized fluorescence ratio (Fluorescence/OD600) was calculated.
In the absence of sodium thiosulfate, Thiosulfate biosensor-A displayed very low fluorescence, indicating that the sensor was not activated without thiosulfate. This also suggests that the background noise is minimal, which reduces the likelihood of false positives in future applications. At a concentration of 0.2 mM, the fluorescence signal increased significantly, indicating that the sensor was activated and responded clearly to the presence of thiosulfate. At 1 mM sodium thiosulfate, the fluorescence signal further increased, demonstrating a strong response to higher concentrations of thiosulfate (Figure 4).
The data show that Thiosulfate biosensor-A exhibits a concentration-dependent response to sodium thiosulfate, with significant fluorescence increases at 0.2 mM and 1 mM, particularly at the higher concentration. This indicates high sensitivity to thiosulfate. Future efforts will focus on optimizing the sensor for lower concentrations and testing its performance in more complex biological environments to evaluate practical applications.
To further enhance the sensitivity and detection accuracy of the thiosulfate biosensor, we optimized the original Thiosulfate biosensor-A and developed a new version, Thiosulfate biosensor-B(BBa_K5084013). The primary strategy involved modifying the promoter of ThsR and adjusting the ribosome binding site (RBS) sequences of both ThsS and ThsR. These modifications were aimed at balancing the expression of the sensing system, thereby improving its response speed and sensitivity to thiosulfate (Figure 5).
During the optimization process, we replaced the promoter of ThsR and adjusted the RBS sequences of ThsS and ThsR to enhance their translation efficiency and sensing capability. By fine-tuning the regulatory mechanisms of ThsR and ThsS, Thiosulfate biosensor-B was expected to detect changes in thiosulfate concentration more efficiently and sensitively than the original version (Figure 6).
To evaluate the performance of Thiosulfate biosensor-B, we conducted multiple rounds of screening and functional tests. The testing method was identical to that used for Thiosulfate biosensor-A: the engineered bacteria were inoculated into LB medium containing varying concentrations of sodium thiosulfate (0 mM and 1 mM) and incubated at 37℃ for 8 hours. After incubation, fluorescence intensity (excitation at 584 nm, emission at 607 nm) and OD600 were measured using a microplate reader, and the normalized fluorescence ratio (Fluorescence/OD600) was calculated.
The results demonstrated that both biosensors, A and B, exhibited low and comparable fluorescence signals in the absence of thiosulfate, indicating minimal background noise without the target molecule. However, Thiosulfate biosensor-B showed a significantly higher response to 1 mM sodium thiosulfate compared to Thiosulfate biosensor-A, with markedly increased fluorescence intensity. This indicates that the adjustments to the promoter and RBS sequences successfully enhanced the sensitivity of Thiosulfate biosensor-B (Figure 7).
In conclusion, by optimizing the promoter and RBS, we developed Thiosulfate biosensor-B with improved sensitivity. Further testing will assess its performance in more complex environments and conditions.
To evaluate the dose-dependence of the constructed Thiosulfate biosensor-B, we measured the fluorescence response at different concentrations of sodium thiosulfate. We used the Thiosulfate biosensor-B and inoculated the engineered strain into LB medium containing various concentrations of sodium thiosulfate (0, 0.125, 0.25, 0.5, and 1 mM). After 8 hours of incubation at 37℃, the fluorescence intensity of the cultures was measured using a microplate reader. The fluorescence values were normalized to the OD600 (Fluorescence/OD600) to account for cell growth differences.
As shown in the Figure 8, the normalized fluorescence ratio increased as the concentration of thiosulfate increased. At 0 mM thiosulfate, the fluorescence signal is very low, with an average of 15.79, indicating that the sensor is not activated. At a concentration of 0.125 mM, the fluorescence ratio increases sharply to 3449.43, showing a clear response to low concentrations of thiosulfate. At 0.25 mM, the fluorescence signal continues to increase, reaching an average of 4900.49, demonstrating a stronger dose-dependence. At 0.5 mM, the fluorescence signal nears saturation, with an average of 6079.88, indicating that the sensor’s response is close to its maximum. At 1 mM thiosulfate, the fluorescence signal is 6236.64, and the curve begins to plateau, suggesting that the sensor’s response has stabilized at high concentrations.
The data show a strong dose-dependent response in biosensor B, with fluorescence increasing as thiosulfate concentration rises. The sensor reaches optimal performance at 0.5 mM, with signal growth plateauing around 1 mM.
To assess the specificity of the thiosulfate biosensor and determine whether it could be falsely triggered by structurally similar compounds, we conducted a specificity test using various sulfate-based substrates. The goal was to ensure the biosensor would not produce false-positive results in the presence of these compounds.
We inoculated the engineered E. coli strain into LB medium containing 1 mM of different sulfate compounds, including sodium sulfate (MACKLIN, 7757-82-6), sodium sulfite (MACKLIN, 7757-83-7), sodium tetrathionate (MACKLIN, 13721-29-4), and sodium thiosulfate. After 8 hours of incubation at 37℃, OD600 values were measured and adjusted to approximately 1. The fluorescence intensity (excitation at 584 nm, emission at 607 nm) and OD600 values were then measured using a microplate reader. The normalized fluorescence ratio (Fluorescence/OD600) was calculated for each compound.
The specificity analysis shows distinct differences in the biosensor’s response to various sulfate-based compounds: The biosensor exhibited very low responses to sulfate and sulfite, with average normalized fluorescence ratios of 22.23 ± 3.61 and 25.08 ± 7.28, respectively. These results confirm that the biosensor is not activated by these compounds, demonstrating high specificity and minimal false-positive potential. The biosensor showed a moderate response to tetrathionate, with an average fluorescence ratio of 1027.88 ± 96.92. This response indicates some cross-reactivity, but given that tetrathionate is also an important biomarker for inflammatory conditions, this level of response is acceptable. The biosensor exhibited a strong response to thiosulfate, with an average normalized fluorescence ratio of 5853.38 ± 360.09, significantly higher than the responses to other compounds. This confirms that the biosensor is highly sensitive and specific to thiosulfate, with minimal interference from structurally similar compounds (Figure 9).
Specificity tests show that the thiosulfate biosensor accurately detects thiosulfate with minimal or no response to sulfate and sulfite. A slight response to tetrathionate is acceptable due to its relevance as an inflammation marker. Overall, the biosensor demonstrates high specificity and reliability for thiosulfate detection.
In order to colourise human stool, we added LacZ downstream of the sensory system (BBa_K5084024). We first synthesized the thiosulfate operon, including the ThsS and ThsR genes, and coupled them with the LacZ gene (encoding β-galactosidase) using the PphsA promoter. These gene fragments were cloned into the pSB1A3 plasmid. The plasmid construction was verified through sequencing (Qingke, China). After successful validation, the plasmid was transformed into E. coli DH5α (Figure 10).
Appropriate amount of overnight cultured bacterial solution was inoculated into fresh LB medium at 1:100 ratio, and different concentrations of thiosulfate were added. β-galactosidase activity was measured using a β-galactosidase reporter gene assay kit (AKSU042M, boxbio).The yellow product (ortho-nitrophenol, ONP) from the hydrolysis of ONPG was used to measure the fluorescence value (OD400), allowing the calculation of enzyme activity. A standard curve (3 repetitions) was established using a standard material, and the unit of enzyme activity was defined as OD600=1 1 nM ONPG produced by bacteria per hour.
The Figure 11A shows a positive correlation between OD400 and the amount of ONPG substrate hydrolyzed. The linear regression equation is Y=0.004543X-0.06551, which means we can caculate the concentration of ONPG by mesure the OD400. The Figure 11B shows the OD400 values at different concentrations of thiosulfate. By bringing it into the linear regression equation, we can get the ONPG concentration corresponding to each sample. According to the formula of enzyme activity unit, the enzyme activity under different concentration of thiosulfate was calculated(Figure 11C).With the increase of thiosulfate concentration, the enzyme activity also increased.Through the addition of X-gal, we successfully observed the blue color reaction produced by β-galactosidase.(Figure 11E).
We successfully expressed β-galactosidase in engineered E. coli DH5α and confirmed its ability to degrade X-gal. Future work will focus on expressing it in E. coli Nissle 1917 to optimize its use as an oral product.
To verify the expression and secretion of hEGF in the constructed smart probiotic strains under simulated colitis conditions, we developed two recombinant plasmids: pSB-PE and pSB-PEA.
Plasmid pSB-PE(BBa_K5084018): Contains the ThsS/R sensing system (Pj23104/Pj23100 promoters) and the Hly-EGF(BBa_K5084017) expression module (PphsA promoter encoding hEGF), but lacks the α-hemolysin secretion system. This was used as a control to measure hEGF secretion efficiency.
Plasmid pSB-PEA(BBa_K5084019): In addition to the ThsS/R system, it includes the Hly-EGF expression module and a complete α-hemolysin secretion system (hlyB BBa_K5084015 and hlyD BBa_K5084016 genes, Pj23104 promoter) to evaluate the role of the α-hemolysin secretion system in promoting hEGF secretion.
The recombinant plasmids were verified by sequencing (Qingke, China) and then transformed into E. coli DH5α and E. coli Nissle 1917 (EcN) to generate the engineered strains expressing and secreting hEGF (Figure 12).
To assess the secretion efficiency of hEGF in different cellular environments, E. coli Nissle 1917 (EcN) expressing hEGF was inoculated into 50 mL LB medium containing different concentrations of sodium thiosulfate (0 mM and 1 mM) and cultured overnight. The following day, 5 mL of the bacterial culture was collected and adjusted to OD600 = 1, followed by centrifugation at 10,000 x g for 1 minute. The supernatant was taken as the extracellular sample, while the bacterial pellet was resuspended in 5 mL PBS and sonicated to lyse the cells. The resulting cell lysate was used as the intracellular sample. The hEGF content was measured using an ELISA kit (mlbio, China).
At 1 mM thiosulfate, the extracellular hEGF content in the EcN/pSB-PE strain was 2.37 ± 0.44 μg/L, while the EcN/pSB-PEA strain showed a significant increase, with hEGF levels reaching 19.92 ± 1.67 μg/L. These results demonstrate that the α-hemolysin secretion system greatly enhanced the extracellular secretion of hEGF, improving secretion efficiency (Figure 13A).
At 1 mM thiosulfate, the intracellular hEGF content in the EcN/pSB-PE strain was 17.15 ± 1.18 μg/L, while in the EcN/pSB-PEA strain it was 8.16 ± 1.19 μg/L. This indicates that the introduction of the α-hemolysin secretion system in the EcN/pSB-PEA strain significantly reduced intracellular hEGF levels compared to the EcN/pSB-PE strain, suggesting that more hEGF was secreted to the extracellular environment (Figure 13B).
The α-hemolysin secretion system in the EcN/pSB-PEA strain significantly enhances extracellular hEGF secretion while reducing intracellular levels, indicating its key role in improving secretion efficiency, especially in a simulated colitis environment.
The E. coli Nissle 1917 (EcN) strain expressing hEGF(pSB-PE and pSB-PEA) was inoculated into 50 mL LB medium containing 1 mM sodium thiosulfate and cultured overnight at 37℃. The next day, 5 mL of the bacterial culture was collected and adjusted to OD600 = 1. The supernatant was collected as the extracellular sample by centrifugation at 10,000 x g for 1 minute. The bacterial pellet was subjected to ultrasonic disruption (75 W, 1s sonication, 3s intervals, for 20 min) to obtain bacterial content samples. The same procedure was applied to the EcN strain containing the empty vector as a control.
Human embryonic kidney 293T cells were seeded in 24-well plates at a density of 5×10^5 cells per well. Cells were cultured in DMEM containing 10% fetal bovine serum (FBS) and 1% penicillin-streptomycin at 37℃ in 5% CO2 until 70% confluency was reached. Once cells reached 70% confluency, 100 μL of either the bacterial content or filtered supernatant from the engineered strains was added to the DMEM medium in each well. After 8 hours of incubation, cell viability was assessed using the CCK8 kit. 100 μL of CCK8 solution was added to each well and incubated for 3 hours at 37℃. Absorbance at 450 nm (A450) was measured using a microplate reader to assess cell viability.
The impact of the supernatant and bacterial content from different strains on 293T cell viability was measured, with the absorbance (A450) values.The supernatant group showed that the EcN/pSB-PEA strain, containing the α-hemolysin secretion system, exhibited the highest A450 value (2.87 ± 0.41), significantly higher than the EcN/pSB-PE strain (2.24 ± 0.24), indicating that the secretion system effectively enhanced hEGF secretion, thereby improving 293T cell viability. In the bacterial content group, the EcN/pSB-PE strain exhibited a higher A450 value (4.75 ± 0.49) compared to the EcN/pSB-PEA strain (3.16 ± 0.40), suggesting that without the secretion system, hEGF predominantly accumulated intracellularly. In contrast, in the EcN/pSB-PEA strain, hEGF was efficiently secreted extracellularly.
The results show that the EcN/pSB-PEA strain, which includes the α-hemolysin secretion system, significantly increased extracellular hEGF secretion, enhancing 293T cell viability. In contrast, in the EcN/pSB-PE strain, hEGF largely remained intracellular, resulting in lower extracellular activity.
The rhamnose-inducible promoter PRha(BBa_K914003), along with the ribosome binding site (RBS) B0034, was synthesized and cloned into the pSB1A3 plasmid using the XbaI and SpeI restriction sites. This construct was placed upstream of the coding sequence for the red fluorescent protein (mRFP). The recombinant plasmid was then transformed into E. coli BL21 cells using the calcium chloride heat-shock method.
A 1:100 dilution of the rhamnose-inducible reporter strain was inoculated into LB medium and cultured overnight at 37°C. The following day, the overnight culture was further diluted 1:50 into fresh M9 medium containing 50 μg/mL ampicillin and supplemented with 0.4% glucose. Simultaneously, a rhamnose solution (final concentration ranging from 0.1% to 1%) was added to the medium to induce the reporter strain. The cultures were incubated at 37°C for 5 hours. After induction, 1 mL of the culture was sampled, and the OD600 value was measured using a microplate reader. Fluorescence was also measured at an excitation wavelength of 584 nm and an emission wavelength of 607 nm. The normalized fluorescence intensity (Fluorescence/OD600) was then calculated to assess reporter strain activity.
In the absence of L-rhamnose, Rhamnose-inducible reporter strain displayed very low fluorescence, indicating that the reporter strain was not activated without L-rhamnose. At a concentration of 0.1%, the fluorescence signal increased significantly, indicating that the reporter strain was activated and responded clearly to the presence of L-rhamnose. At 0.5% and 1% L-rhamnose, the fluorescence signal further increased, demonstrating a strong response to higher concentrations of L-rhamnose (Figure 16).
The rhamnose-inducible system shows a clear, dose-dependent response to L-rhamnose, making it a reliable tool for controlled gene expression.
To construct a controllable suicide system, the antitoxin gene mazE and toxin gene mazF were synthesized and cloned into the pSB1A3 plasmid. The antitoxin gene mazE was placed downstream of the rhamnose-inducible promoter PRha and RBS B0034, regulated by rhamnose. The toxin gene mazF was positioned downstream of the arabinose-inducible promoter PBAD and regulated by arabinose. After sequencing verification, the recombinant plasmid Rha-MazTox (BBa_K5084020) was transformed into E. coli BL21 to validate the functionality of the suicide system.
To evaluate the effectiveness of the suicide system, the engineered E. coli BL21 strain was inoculated at a 1:100 ratio into LB medium and cultured overnight at 37°C. The next day, the culture was diluted 1:50 into fresh M9 medium containing 50 μg/mL ampicillin, 1% rhamnose, 0.2% arabinose, and 0.4% glucose. The OD600 values were monitored over time using a microplate reader to assess the impact of the suicide system.
The results show clear differences in the growth of the tested strains. Both the BL21 strain and the BL21/PBAD-mazE-PRha strain (with 1% L-rhamnose) displayed normal growth over the 8-hour period, with OD600 values reaching approximately 0.7 by the end of the experiment, indicating healthy bacterial growth. In contrast, the BL21/PBAD-mazE-PRha strain (without 1% L-rhamnose) exhibited a marked reduction in OD600 over time. By 3 hours, its growth had already begun to decline compared to the control strains (Figure 18). This suggests that the rhamnose-induced suicide system was successfully activated in the absence of L-rhamnose, leading to bacterial cell death.
The rhamnose-induced suicide system worked as expected. With 1% L-rhamnose, the strain grew normally, but without it, growth declined by 3 hours, confirming activation of the system and cell death.
[1] Wang C, Zeng H S, Liu K X, et al. Biosensor-based therapy powered by synthetic biology[J]. Smart Materials in Medicine, 2023, 4: 212-224.
[2] Wang X G, Zou Z P, Du Y, et al. Construction of an Engineered Escherichia coli with Efficient Chemotactic and Metabolizing Ability toward Tetrathionate[J]. ACS Synthetic Biology, 2023, 12(11): 3414-3423.
[3] Daeffler K N M, Galley J D, Sheth R U, et al. Engineering bacterial thiosulfate and tetrathionate sensors for detecting gut inflammation[J]. Molecular systems biology, 2017, 13(4): 923.
[4] Zammit C M, Quaranta D, Gibson S, et al. A whole-cell biosensor for the detection of gold[J]. PloS one, 2013, 8(8): e69292.
[5] Woo S G, Moon S J, Kim S K, et al. A designed whole-cell biosensor for live diagnosis of gut inflammation through nitrate sensing[J]. Biosensors and Bioelectronics, 2020, 168: 112523.
[6] Pardee A B, Jacob F, Monod J. The genetic control and cytoplasmic expression of “inducibility” in the synthesis of β-galactosidase by E. coli [J]. Journal of Molecular Biology, 1959, 1(2): 165-178.
[7] Matthews B W. The structure of E. coli β-galactosidase[J]. Comptes rendus. Biologies, 2005, 328(6): 549-556.
[8] Wiles T J, Mulvey M A. The RTX pore-forming toxin α-hemolysin of uropathogenic Escherichia coli: progress and perspectives[J]. Future microbiology, 2013, 8(1): 73-84.
[9] Li C, Li Y, Zhuang M, et al. Long noncoding RNA H19 act as a competing endogenous RNA of Let‐7g to facilitate IEC‐6 cell migration and proliferation via regulating EGF[J]. Journal of cellular physiology, 2021, 236(4): 2881-2892.
[10] Hu W, Zheng S, Guo H, et al. PLAGL2‐EGFR‐HIF‐1/2α signaling loop promotes HCC progression and erlotinib insensitivity[J]. Hepatology, 2021, 73(2): 674-691.
[11] Engelberg-Kulka H, Hazan R, Amitai S. mazEF: a chromosomal toxin-antitoxin module that triggers programmed cell death in bacteria[J]. Journal of cell science, 2005, 118(19): 4327-4332.
[12] Guzman L M, Belin D, Carson M J, et al. Tight regulation, modulation, and high-level expression by vectors containing the arabinose PBAD promoter[J]. Journal of bacteriology, 1995, 177(14): 4121-4130.
[13] Giacalone M J, Gentile A M, Lovitt B T, et al. Toxic protein expression in Escherichia coli using a rhamnose-based tightly regulated and tunable promoter system[J]. Biotechniques, 2006, 40(3): 355-364.