LOOP 1（June）
Design＆Learn
Chronic hepatitis is a major global public health problem. Almost 1 million people are dying from hepatitis infection every year. However, in many cases, there are no obvious symptoms of liver dysfunction in an early stage of chronic hepatitis. For many patients, they don't even realize they need medical attention until it's too late. Current diagnostic methods of liver disease focus on accurate detection in hospital environments, but few convenient home-based tests are available. However, blood sampling is required in most of the sophisticated tests in hospital, which is time-consuming and expensive.
Figure 1. Global Burden of Cirrhosis Mortality
By investigating literatures, we found that bile acids can be used as an indicator for liver disease. So, we focus on receptors that can detect bile acids. Farnesoid X receptor (FXR) and Retinoid X receptor (RXR) are two important receptors for bile acid sensing in human. FXR can efficiently identify and bind with Chenodeoxycholic acid (CDCA), Lithocholic acid (LCA) and Deoxycholic acid (DCA). Once activated by bile acids, it undergoes a conformational change and be activited. Then, it can dimerize RXR and turn on the downstream related gene expression. Among many kinds of bile acids, CDCA is relatively higher in human urine and has significantly difference between patients with liver disease and normal subjects. So, it can be used as a sensitive indicator for the diagnosis of liver disease. In addition, most of the bile acids in urine are dominated by 3α-sulfated bile acids. Therefore, we used bile acid sulfate sulfatase (BSS) from Pseudomonas testosteroni to remove the 3-sulphate sulfate groups of bile acids. Subsequently, FXR can precisely recognize the substrates, which were produced in the previous reaction.
With the system above, we selected CadC, promoter CadBA and GFP as our report device. When FXR and RXR dimerize, the linked CadC will be in close proximity and the dimerization occurs, activating promoter CadBA to express GFP.
Finally, we chose FXR and RXR as bile acid sensing receptors, GFP as a signaling device and E. coli BL21 as our chassis organism, hoping to demonstrate the feasibility of our project.
Build
In the construction of the expression system, we followed the RFC10 rules and constructed composite parts to facilitate the construction of the final expression devices as bellow. We obtained the genes and the promoters through complete synthesis and chose E. coli BL21 as the chassis organism which will not cause potential biological risk technically. If you want to know the detailed experiment protocols, you can visit our protocol page.
Figure 2.Gene circuit based on GFP
Test
Firstly, we tested parts we constructed through sequencing verification. The results showed that we had successfully constructed the following components: J23105-BSS (BBa_K4164008), J23105-FXR-CadC(BBa_K4164023), J23105-RXR-CadC(BBa_K4164024), PCadBA-GFP(BBa_K4164028).
Secondly, we conducted a set of CDCA concentration gradient experiments. We tried to perform a mathematical model, but there was no more significant correlation between the data.
Figure 3. Fluorescence intensity/OD using a gradient of CDCA concentrations based on GFP.
Learn
The results have shown that there may be something wrong with our genetic pathways. After analysis and discussion, we speculated that perhaps the gene pathway was too complex. The system contained a cascade reaction of multiple protein-protein interactions and protein-nucleic acid interactions, which made it difficult to achieve the expected results. Moreover, since CDCA is bacteriostatic and its entry into cells is restricted, it is inappropriate for our application to use cells as our chassis.
After communication and discussion with NJMU-China, we decided to use the cell-free system to solve the problems above. In the meantime, when exchanging ideas with Professor Xin Yan , he advised that we could apply other approaches instead of complex pathways that depend on CadC proteins, avoiding complex reactions by simplifying gene circuits and obtaining more reliable results.
LOOP 2（July）
Design
By investigating literatures, we chose a specific fluorescent protein：dimerization-dependent fluorescent protein (ddRFP). This fluorescent protein is able to fluoresce strongly upon heterodimerization, which is a formal requirement for our project. Therefore, we constructed a streamlined pathway with ddRFP directly connected to the receptor.
Next, we linked FXR and RXR with two monomeric proteins respectively. After FXR is activated by CDCA and dimerizes with RXR, the downstream ddRFP-A and ddRFP-B are correspondingly driven to dimerize and emit fluorescence. We also added 6xHis tags to them to perform Ni-NTA affinity chromatography for protein extraction and identification.
Build
During the construction of our new system, we followed the RFC10 rules and built modular parts to meet the needs of our expression system. We selected E. coli BL21 as the chassis organism for initial testing. We eventually extracted plasmids for application in our cell-free system.
Test
Firstly, we tested the parts we constructed through sequencing verification. The results showed that we had successfully constructed the following components: J23105-FXR-ddRFPA、J23105-RXR-ddRFPB.
Then, on the basis of these plasmids, we conducted a set of CDCA concentration gradient experiments to measure its fluorescence intensity. The data showed that the fluorescence intensity changed very little under activation of different concentrations of CDCA. Consistently, no significant differences were observed.
Figure 4. Fluorescence intensity/OD using a gradient of CDCA concentrations based on ddRFP.
Learn
After above experiments, we found that the fluorescence intensity increased slightly with the increase of CDCA concentration, but no significant difference. The difference in the fluorescence produced by our exposure to corresponding wavelength of light was minimal. Therefore, we discussed and offered 2 speculations as follows:
1. Under CDCA activation, FXR binds weakly to RXR, thus preventing ddRFP from fluorescing to the naked eye.
2. There is a problem with the ddRFP, which does not have enough luminescence to produce fluorescence visible to the naked eye.
When talking with our PI Professor Chen, he suggested that we could use a linker to link ddRFP monomers for the experiment. In this way, we tested whether the ddRFP function is sufficient to emit fluorescence visible to the naked eye.
LOOP3（July）
Design＆Build
In response to the problems of the preliminary experiments and the suggestions given by Professor Chen, we decided to test the ddRFP fluorescence intensity. We linked the ddRFP-A and ddRFP-B respectively to narrow the distance so that they could dimerize efficiently.
Meanwhile, we also predicted the structure of two fusion proteins, FXR-ddRFPA and RXR-ddRFPB. And we conducted molecular dynamics simulations to seek the possible problems that the two proteins might occur during their interaction. As the famous idea illustrated, "Model directs experiment, experiment reflects model." By simulating and analyzing protein structures, we could obtain valid information about the interactions between the proteins, telling us whether the proteins would bind as we expected.
Test
In the test of the ddRFP, we found that ddRFP had weak demonization ability and only emit weak fluorescence that is not visible to the naked eye which coincided with molecular dynamics simulations. Our fluorescence intensity was low due to the weak dimerization ability between the two receptors. At the same time, we also found that the spatial structure of the fluorescent protein was different from our expectation and could not dimerize effectively.
Figure 5. Prediction of CDCA+FXR+ddRFPA1 with RXR+ddRFPB1.
Learn
We searched relevant literature to interpret the experimental results above. Compared with the literature, the structure of ddRFP they offered was different from what we obtained, and its original sequence had a slight difference from the one we use. We believed that it was these sequence differences that affected the folding of the protein and resulted in the low fluorescence intensity. Fortunately, FXR and RXR have long been well-studied and characterized. Their gene sequences and expression conditions were consistent with literature, and they could formulate correct folding structures.
Next, during the exchange with Professor Xi Chen, he learned about our current problems and introduced Professor Josef Voglmeir from Nanjing Agricultural University, who was experienced in the study of ddRFP. What's more, Professor Chen suggested that we could replaced the promoter with a higher expression intensity to express more proteins to get high-quality fluorescence when solving the problem of ddRFP.
After an in-depth conversation about our situation with Professor Josef Voglmeir, he provided us with his strains that could emit visible fluorescence to help us continue our project.
LOOP4 (August)
Design
The results of the pre-experiment showed that J23105 was a low-intensity promoter that could not emit visible fluorescence, even it was fully expressed. Therefore, we decided to change the promoter with higher expression strength to increase protein expression for stronger fluorescence intensity. The high-intensity expression of the protein may lead to the formation of inclusion bodies, affecting the interaction of the protein. So we planned to use affinity chromatography to extract the protein to confirm whether an inclusion body structure was formed.
Next, we would like to test the fluorescence intensity of the ddRFP from Professor Josef Voglmeir and compare it with the former ddRFP we used to obtain more significant results. Then we modified the gene sequence and ran the molecular dynamics simulations again.
Build
We linked the ddRFP-A1 with the ddRFP-B1 given by Professor Josef Voglmeir to different expression strength. We replaced the J23105 promoter with the T7 promoter by homologous recombination. In the construction of the expression system, we followed the RFC10 rules strictly.
Test
Firstly, we tested parts we constructed through sequencing verification. The results showed that we have successfully constructed the following components: T7-ddRFPA-ddRFPB, J23105-ddRFPA1-ddRFPB1(BBa_K4164012), T7-BSS(BBa_K4164013), T7-FXR-ddRFPA1(BBa_K4164014), T7-RXR-ddRFPB1(BBa_K41640150, T7-ddRFPA1-ddRFPB1(BBa_K4164016), T7-FXR-ddRFPA1-T7-RXR-ddRFPB1(BBa_K4164017).
Figure 6. a. Inductive expression of FXR-ddRFPA1, BBa_K4164014; b.Inductive expression of RXR-ddRFPB1, BBa_K4164015; c. Inductive expression of FXR-ddRFPA1-RXR-ddRFPB1, BBa_K4164017.
Next, we compared the fluorescence intensities of T7-ddRFPA-ddRFPB, J23105-ddRFPA1-ddRFPB1, and T7-ddRFPA1-ddRFPB1 with J23105-ddRFPA-ddRFPB we previously constructed.The result showed that T7-ddRFPA1-ddRFPB1 was capable of emitting visible fluorescence with very high intensity. On the basis of T7-FXR-ddRFPA1 and T7-RXR-ddRFPB1, we conducted a set of CDCA concentration gradient experiments to measure its fluorescence intensity again.
Figure 7.Comparison of BBa_K4164012 and BBa_K4164016. Left:BBa_K4164012; Right:BBa_K4164016.
Figure 8. Comparison of Fluorescence intensity/OD using PT7 and PJ23105 in different concentrations of CDCA
Then, we simulated the binding efficiency of FXR and RXR in different CDCA concentrations and determined that the optimum concentration is 25 μM.
Figure 9. The curve of binding efficiency in different CDCA concentrations
Finally, we extracted T7-RXR-ddRFPB1 and T7-FXR-ddRFPA1 by Ni-NTA affinity chromatography and determined the amount of expression by SDS-PAGE. The results of SDS-PAGE showed that most of the proteins we expressed were presented in the cell precipitate, and only a small fraction of the proteins was extracted successfully by affinity chromatography.
Figure 10. Purification of FXR-ddRFPA1 and RXR-ddRFPB1. Lane 1: protein contained in the pellet after bacterial disruption. Lane 2: RXR-ddRFPB1 purified from supernatant of bacteria liquid. Lane 3: FXR-ddRFPA1 purified from supernatant of bacteria liquid in optimized expression conditon.
Learn
The above experimental results indicated that we successfully proved our assumption that variation in the ddRFP sequence accounted for the phenomenon of reduced fluorescence intensity. Furthermore, the ability of the ddRFP to emit high-intensity fluorescence indicated that our project was very promising.
However, the SDS-PAGE results showed that our protein had formulated inclusion body. To solve this problem, we asked Professor Weiwu Wang from the College of Life Sciences of Nanjing Agricultural University. Professor Wang pointed out that we should first repeat the experiment according to the existing protocol in the literature, and suggested reducing the concentration of IPTG to lower the induced expression intensity. In the subsequent experiment, we successfully overcame the problem of inclusion body and successfully expressed the protein.
LOOP5（September）
Design
Having selected the appropriate part, we proceeded to evaluate the ability of our cell-free system and the protein to interact. Thus, we test the cell-free system firstly.
The test results showed that our cell-free system was unable to emit fluorescence visible to the naked eye, but was able to show a large difference under specific wavelength excitation. So, we need irradiation with a particular wavelength of light to make our fluorescence difference more significant. To the end, we wanted to develop a three-dimensional (3D)-printed, handheld fluorescence illuminator that could visualize fluorescent output with the naked eye. (More details can be found on Hardware )
Figure 11. Impact of different concentrations of CDCA on report device fluorescence intensity. blank: cell-free system with no plasmid a. Direct observation of fluorescence. b. Fluorescence under specific wavelength excitation.
To make our testing method more convenient, we did not design criteria, but rather tested by qualitative methods. We explored the optimal response time for the cell-free system in order to obtain the best results. In addition, we would like to explore a suitable dilution multiple by examining the fluorescence intensity of the cell-free system at different concentrations of CDCA activation.In this way, diluted CDCA samples at normal person urine levels do not fluoresce. We were then able to achieve qualitative detection.Through an extensive review of the literatures, we chose dilution multiple of 5x.
Finally, since the external environment may affect the normal function of the cell-free system, we simulated the cell-free system in a perturbed state of the environment. We wanted to explore whether the functionality of our system would be affected under different environmental conditions.
Build＆Test
Based on literatures, we have developed a three-dimensional (3D)-printed, handheld fluorescence illuminator. It is equipped with an internal LED light source and filter to enable the observation of fluorescent proteins, and it includes a biodegradable material 3D-printed housing to ensure the device is portable, low-cost, recyclable and will not increase the burden of envirnment.
We placed the cell-free system in our testing facility and added a certain concentration of CDCA as the sample to be tested. The excitation by the light source in the device is able to produce a more pronounced fluorescence.
We have explored the optimal reaction time for our cell-free system. After the addition of IPTG overnight induction, we conducted a set of time gradient experiments. The experimental results showed that we were able to obtain a strong fluorescence intensity at 8h. We simulated our cell-free system under different environmental conditions, such as under various temperatures. The results showed that at different temperatures has little interference on our systems.
Figure 12.The curve of our cell-free system in disturbance
Finally, we explored the most reasonable sample dilution multiple. We compared activation effects using a 5-fold dilution of normal human CDCA concentrations with patient CDCA concentrations, and found that our design was effective. The normal human CDCA concentration failed to show significant fluorescence under light excitation, while the patient CDCA concentration did the opposite. This also justified that a 5-fold dilution was reasonable.
Figure 13. Handheld fluorescence illuminator a.Handheld illuminator b.A handheld illuminator in use
Learn
We successfully detected bile acid levels in patients with chronic hepatitis by cell-free system and demonstrated significant differences when compared to bile acid levels in people without the disease. Ultimately, our project is designed to provide an initial diagnosis to warn chronic hepatitis patients of their risk and to alert them to seek medical attention in time.
Reference
Asrani, Sumeet K., et al. "Burden of liver diseases in the world." Journal of hepatology 70.1 (2019): 151-171.
Thomas, David L. "Global elimination of chronic hepatitis." New England Journal of Medicine 380.21 (2019): 2041-2050.
Tang, Ying-Mei, et al. "Urine and serum metabolomic profiling reveals that bile acids and carnitine may be potential biomarkers of primary biliary cirrhosis." International journal of molecular medicine 36.2 (2015): 377-385.
Küper, Christoph, and Kirsten Jung. "CadC-mediated activation of the cadBA promoter in Escherichia coli." Microbial Physiology 10.1 (2005): 26-39.
Alnouti, Yazen. "Bile acid sulfation: a pathway of bile acid elimination and detoxification." Toxicological Sciences 108.2 (2009): 225-246.
Chang, Hung-Ju, et al. "Programmable receptors enable bacterial biosensors to detect pathological biomarkers in clinical samples." Nature communications 12.1 (2021): 1-12.
Fiorucci, Stefano, et al. "Targeting farnesoid X receptor for liver and metabolic disorders." Trends in molecular medicine 13.7 (2007): 298-309.
Wang, Hong, et al. "FXR modulators for enterohepatic and metabolic diseases." Expert Opinion on Therapeutic Patents 28.11 (2018): 765-782.
Nakashima, Toshiaki, et al. "Unusual trihydroxy bile acids in the urine of patients treated with chenodeoxycholate, ursodeoxycholate or rifampicin and those with cirrhosis." Hepatology 11.2 (1990): 255-260.
Thakare, Rhishikesh, et al. "Species differences in bile acids I. Plasma and urine bile acid composition." Journal of Applied Toxicology 38.10 (2018): 1323-1335.
Tu, Hua, Arthur Y. Okamoto, and Bei Shan. "FXR, a bile acid receptor and biological sensor." Trends in cardiovascular medicine 10.1 (2000): 30-35.
Makishima, Makoto, et al. "Identification of a nuclear receptor for bile acids." Science 284.5418 (1999): 1362-1365.
Obatake, Masayuki, et al. "Uninary sulfated bile acids: A new simple urine test for cholestasis in infants and children." Journal of pediatric surgery 37.12 (2002): 1707-1708.
Alford, Spencer C., et al. "A fluorogenic red fluorescent protein heterodimer." Chemistry & biology 19.3 (2012): 353-360.
Jung, Jaeyoung K., et al. "Cell-free biosensors for rapid detection of water contaminants." Nature biotechnology 38.12 (2020): 1451-1459.
Dan Li.Study on Detection Method of N-Glycosylase Activity Based on Red Fluorescent Protein Dimer.2018.Nanjing Agricultural University,MA thesis.
Lavickova, Barbora, and Sebastian J. Maerkl. "A simple, robust, and low-cost method to produce the PURE cell-free system." ACS Synthetic Biology 8.2 (2019): 455-462.
Tazuke, Yasuhiko, et al. "A new enzymatic assay method of sulfated bile acids in urine." Japanese Journal of Clinical Chemistry 21.4 (1992): 249-258.
