Engineering
Background
Inflammatory bowel disease (IBD) is a chronic intestinal inflammatory disease of unknown etiology, including ulcerative colitis (CD) and Crohn’s disease (CD). This chronic disease, which is prone to repeated deterioration, currently lacks unified diagnostic and treatment standards, and is posing a great threat to public health. Drug therapy (anti-inflammatory drugs) is the preferred treatment for IBD. However, studies in the past 10 years have found that 30-50% of IBD patients do not respond to anti-TNF treatment. In addition, after long-term use of anti-inflammatory drugs, the patient's intestinal microbial status changes over time, and the effect may be lost due to drug resistance. Therefore, we need to seek help from other treatments for IBD.
At present, some researchers have developed engineered bacteria that respond to biomarkers of intestinal inflammation, demonstrating the feasibility of using synthetic biology to design engineered bacteria in the field of IBD treatment.
Design
Colonic bacteria can produce a large amount of genetically toxic hydrogen sulfide (H2S) during metabolism, and the lumen mucosa can protect itself from H2S by converting H2S into thiosulfate (S2O32-). The sulfate-reducing bacteria in the patient's intestines produce ROS (reactive oxygen species) to convert S2O32- into tetrathionate/tetrathionate (Tetrathionate, S4O62-). In addition, inflammation of the intestine can induce high-level expression of nitric oxide synthase (iNOS) and increase the production of NO. NO is converted to nitrate (NO3-). Therefore, both S2O32-/ S4O62- and NO3- can be used as biomarkers for the development of biosensors for intestinal inflammation.
We can build Boolean logic circuits to combine thiosulfate/tetrasulfate and nitrate biosensors to optimize performance, improve sensor accuracy and enhance robustness, and optimize intestinal inflammation detection (Figure 1). When and only when S2O32-/S4O62- and NO3- are present at the same time, the fluorescent gene is expressed (Figure 1). The fluorescent gene can be replaced with a gene that produce the interleukin-10 (anti-inflammatory factor) for the treatment of IBD.
Figure 1. Boolean gate logic circuit/AND circuit construction.
Build
There are three composite parts were constructed using pUC57 or pSU2718 vector as backbone (Figure 2).
Figure 2. A. PyeaR_HrpR, sensing NO3- to release the substance R; B. PttB344_HrpS_PJ23105_ttrR, sensing S4O62- to release the substance S; C. PhrpL_amilGFP_IL10_PJ23104_ttrS.
Test
(1) Function of S2O32-/S4O62-和NO3- biosensor
Escherichia coli was transformed with pUC57-PhrpL_amilGFP_IL10_PJ23104_ttrS, and or pSU2718- PyeaR_HrpR_ PttB344_HrpS_PJ23105_ttrR. We can see that there was no fluorescence in Sample 3 (the report part) (Figure 3). According to the result of Sample 3 and Sample 4 (Figure 3), we could infer that the design works, since the pUC57-PhrpL_amilGFP_IL10_PJ23104_ttrS didn’t respond when PyeaR_HrpR and PttB344_HrpS_PJ23105_ttrR were not present.
Figure 3. Sample 1: Negative Control (E. coli); Sample 2: Positive Control (E. coli/amilGFP); Sample 3: E. coli/ pUC57-PhrpL_amilGFP_IL10_PJ23104_ttrS; Sample 4: E.coli/ pSU2718- PyeaR_HrpR-PttB344_HrpS_PJ23105_ttrR + pUC57-PhrpL_amilGFP_IL10_PJ23104_ttrS.
(2) Model build
In order analyze the relationship between OD600 and the fluorescence intensity, we built a model by MATLAB (Table 1).
After roughly analyze the scatter plots, we chose to use the quadratic polynomial equation to adapt to our data:
Coding we used in MATLAB is given below:
clear;clc;
od0=[0.6 0.8 1];
intensity1=[1356 5822 6611];
intensity2=[1221 4709 6189];
intensity3=[1578 6700 5927];
intensity0=(intensity1+intensity2+intensity3)/3;
p=polyfit(od0,intensity0,2);
od=[0.6:0.01:1];
intensity=polyval(p,od);
plot(od,intensity,'LineWidth',2)
ymax=max(intensity)
i=find(intensity==max(intensity));
xmax=od(i)
hold on
plot(od0,intensity0,'r*','LineWidth',2)
plot(xmax,ymax,'g*','LineWidth',2)
hold off
After calculation, the constants of this model are listed below:
Therefore the final model would be:
Figure 4. The fitting curve of the model f(x)=-48250x2+89343x-34851
According to the fitting curve (Figure 4), it indicates that the optimal condition for our biosensor to detect IBD related compounds, would be the time when OD600 reaches 0.93 around. This result could be also served as a reference for more performance tests in the future after we complete the improvement of our biosensor genetic design.