Team:QHFZ-China/Demonstrate
_DEMONSTRATE_
Demonstration of uric acid (UA) removal system in HeLa cells: Figure1 shows our design of the UA removal system. mUTS constitutively expresses and binds to hucO8 module in the absence of UA, which inhibits the expression of smUOX. When the UA level is high, the small molecules can be transported into cytoplasm by URAT1. High-level UA makes HucR release from hucO8, therefore smUOX expresses and degrades UA to allantoin. Here are our demonstration experiments.
Figure 1. Gene circuits designed for uric acid removal. (A) Schematic diagram of main parts. (B) Schematic diagram of the mechanism of the gene circuit.
Firstly, we tested if we could measure UA concentration in samples quantitatively. We prepared a series of standard UA solution sample and measured them by the uric acid detection kit (Nanjing Jiancheng Bioengineering Institute, C012-2). Fig. 2A showed the results which were linearly related to the UA concentration. It indicated we could estimate the UA concentration in the sample based on the absorbance in the following experiments.
Next, we introduced one plasmid (expressing hucO8-smUOX) or two plasmids (expressing URAT1 and hucO8-smUOX, respectively) into human HeLa cells, while using eGFP as a negative control, which was placed on plasmid pEGFP-N1. Because there was no mUTS in HeLa cell line, we considered the expression of smUOX was constitutive. The total amount of plasmids in every group was balanced by pEGFP-N1. After transfection, cells were cultured in different initial concentrations of UA. We measured remaining UA concentration of each sample, and found that smUOX could degrade UA successfully, while URAT1 could enhance the clearance efficiency of smUOX (Fig. 2B). After culturing 2 days, the cellular morphology remained stable in all cells (Fig. 2C), which meant HeLa cells could grow normally even in high-level UA addition. These results indicated that URAT1 and smUOX could transport and degrade UA in HeLa cells.
Figure 2. Function test of smUOX and URAT1. (A) A standard curve of the relationship between UA concentration in cell culture medium and OD510 after being mixed with uric acid detection reagent. Data was shown as mean ± SD. N = 3 technical repetitions. (B) Remaining UA concentration of samples from 0 to 48 h. The time when UA was added was defined as 0 h. UA concentration was converted from the absorbance data by the linear correlation equation shown in Fig. 2A. “No cell” group was blank control. Data were shown as Mean ± SD. For “No cell” and “GFP” groups, N = 2 biological repetitions; for the other groups, N = 3 biological repetitions. NS, not significant, P ≥ 0.05; **, P < 0.01; ***, P < 0.001 (two-tailed T test). (C) Photos of different HeLa cells groups. BF, bright field.
Then we demonstrated that whether the mUTS-hucO8 system could control the expression of downstream gene according to extracellular UA concentration. Using hucO8-eGFP as a reporter, we tested the system directly. We transfected 0.3 μg plasmid containing mUTS per well to experimental groups. After culturing 2 days of cultivation, the cells in low UA level didn’t get green, which indicated the mUTS showed inhibition function. However, comparing with the positive control, even if the concentration of UA was 800 μM, the production of eGFP was still rare (Fig. 3A). That meant the interaction of mUTS to hucO8 was over strong, even high-level UA could not release the inhibitor from DNA. A hypothesis was the expression of mUTS was too much.
To optimize the behavior of our UA removal system, changing promoter into a weaker one was a straightforward way. However, the promoters of eukaryotes are more complex than the ones of prokaryotes. As a result, we reduced the amount of plasmid expressing mUTS during transfection, from 0.3 μg / well to 0.1 μg / well. In this situation, hucO8-eGFP could express under the induction of high-level UA (Fig. 3B). The quantitative result indicated mUTS-hucO8 system worked as we predicted: if UA concentration is low, mUTS powerfully suppress the eGFP expression, while under the induction of high UA concentration, the inhibition would be resumed (Fig. 3C, 3D).
Figure 3. Function tests of mUTS-hucO8 system. (A) Photos of the cells which were transfected with 0.3 μg / well mUTS expression plasmids at 48 h after uric acid addition. (B) Photos of the cells which were transfected with 0.1 μg / well mUTS expression plasmids at 48 h after uric acid addition. (C) Fluorescence intensity distribution of the cells measured by flow cytometry. NTC, negative control. Population 1 (P1) represented cells that did not successfully intake hucO8-GFP plasmid. (D) Geo. mean of the fluorescence intensity of P2 in Fig. 3C. Data were shown as mean ± SD. N = 3 technical repetitions.
At last, using the optimized system, we changed the hucO8-eGFP plasmid to hucO8-smUOX, and tested the clearance efficiency of UA by our final system. Excitingly, transfection of the combination of URAT1, mUTS and hucO8-smUOX into one cell could reduce the UA concentration in cell culture medium (Fig. 4). What’s more, as the initial UA concentration increased, the clearance rate to UA of our final system became more and more similar to the positive control group, which cells were only transfected hucO8-smUOX.
All the results above indicated that at low UA concentration (such as 400 μM, which was an acceptable concentration for human body), mUTS in UA removal system would suppressed the function of smUOX, while at high UA concentration (such as 1600 μM, which was much higher than the standard for human body), smUOX would express and clear uric acid.
Figure 4. Final performance of UA removal system under treatment with different initial concentration of UA. (A) Measurement of the remaining UA concentration in cell culture medium from 0 to 48 h. The initial UA concentrations were 400, 800 and 1600 μM, respectively. The time when uric acid was added was defined as 0 h. The absorbance data was converted into UA concentration by linear correlation equation shown in Fig. 2A. Data were shown as Mean ± SD. N = 3 biological repetitions. (B) Comparation of clearance efficiency between the final system and the positive control. Percentage number referred to the ratio of reduced UA by the final system to the one by positive control. Data were shown as Mean ± SD. N = 3 biological repetitions.
These data support that our smart cells could detect and degrade UA molecules in the culture medium. However, the efficiency of UA clearance is not high. There might be several reasons. On the one hand, it may be related to the strong inhibition of downstream genes by mUTS. On the other hand, the transient transfection efficiency of plasmids was not 100%, which indicated not all cells contain the system. In addition, all experiments above were achieved by transient transfection of plasmids. So, we do not know whether the UA removal system could work for a long time, or whether the UA concentration in patients’ body fluid can be controlled within a certain range. In the future, we will further optimize the system by constructing a stable cell line, improving the genetic circuit, demonstrating to construct a stable cell lines, to improve the genetic circuit, to demonstrate a long-term cell culture and so on.
Demonstration of the uric acid (UA) detection system in E. coli cells: We designed a uric acid (UA) detection system in E. coli (Fig. 1). Pc is a constitutive promoter, Pcp6 promoter, and it promotes the expression of HucR and YgfU. If the concentration of UA in environment is low, HucR will bind to PhucR, which inhibits the expression of downstream reporter, dsRed or sfGFP. When extracellular UA is present, YgfU can transport UA into the cytoplasm, which leads HucR dissociates from PhucR, and induces the fluorescent protein expression.
Figure 1. Working mechanism of the uric acid detection system in E. coli.
To demonstrate the design, firstly, we prepared a solution of 1.1mM uric acid (UA) in PBS buffer. By using our uric acid detection kit and professional instruments in hospital, we confirmed our solution was correct (Fig. 2), and this 1.1mM UA solution was used in all subsequent experiments.
Figure 2. A standard curve about the relationship between different UA concentration and OD510 after being mixed with the uric acid detection kit. Negative control (NTC) stood for only PBS buffer. Data were shown as mean ± SD. N = 2 technical repetitions.
Then, we cultured E. coli trans5α strain (equals to DH5α) with or without dsRed gene at different concentrations of UA to test if UA affects the physiological activities of E. coli. Fig. 3. showed there were not significant differences of the growth curve and dsRed expression in a range of 0 to 200 μM UA. That means E. coli is a stable chassis for UA detection.
Figure 3. Effects of UA on E. coli growth and exogenous gene expression. (A) Growth curve of E. coli in different UA concentration. All OD600 data were normalized by taking the OD600 of [UA] = 0 group at 0 h as standard. (B) Responding curve about the dsRed fluorescence / OD600 in a range of 0 to 200 μM UA. NTC was a colony without dsRed gene, and PC was a colony that constructively expressed dsRed. Data in both figures were shown as mean ± SD. N = 3 technical repetitions.
We designed an experiment process based on the literature [1], shown in Fig. 4, to test if our UA detection system can respond to extracellular UA and express the reporter gene.
Figure 4. Experiment process to verify whether our system works as predicted.
Two clones with UA detection system which used dsRed as a reporter were tested by the process mentioned above in Fig. 4. The original gene circuit was able to response to UA (Fig. 5A), and the clone 1 showed much better dynamics than the other (Fig. 5B) in a range of 0 to 200 μM UA, which fitted the user scenario we predicted. (During the Jamboree, we disscussed it with other teams. They suggested that clone 2 might hold mutation, which should be tested in the future.) Then clone 1 was used in the next experiments. Time course experiments showed that the fluorescence intensity became quite strong at 4 to 6 hours after UA induction, and became stable at 10 to 12 hours (Fig. 5C). If we removed UA induction by replacing fresh LB medium, after 48 hours shaking, the fluorescence was still notable (Fig. 5D) and there was not significant difference of dsRed fluorescence / OD600 between before and after UA removing (Fig. 5E). All the data meant our design could detect high UA concentration quickly and stably, which may have a great applying value.
Figure 5. Response of UA detection system after different concentration of UA induction. (A) A photo to visualize the fluorescence induced by UA under a blue light. (B) Responding curve about the dsRed fluorescence / OD600 to different UA concentration of two E. coli clones. Data were shown as mean ± SD. N = 3 technical repetitions. (C) Time course experiments about the dsRed fluorescence / OD600 of E. coli after 0, 20 or 100 μM UA addition. Data were shown as mean ± SD. N = 3 technical repetitions. (D) A photo to visualize the fluorescence after UA removal under a blue light. (E) Quantitative measurement of dsRed fluorescence / OD600 before and after UA removal.
We also tried more conditions to test if this system could work well in different environment. In the range of pH 6.0 to 8.0, response of the gene circuit was relatively stable (Fig. 6A). However, the volume of the reaction system would influence the response to UA (Fig. 6B). A possible explanation was the relative surface area of the liquid level changed and consequently the dissolved oxygen changed. This result meant the experiments for UA detection should be done at the same reaction system volume. In other experiments, 1 mL reaction volume was used.
Figure 6. Impact of different conditions on the UA detection system. (A) Impact of pH value on dsRed fluorescence / OD600 after 20 μM UA addition. Data were shown as mean ± SD. N = 3 technical repetitions. (B) Impact of reaction volume on dsRed fluorescence / OD600 after 0 and 20 μM UA addition. Data were shown as mean + SD. N = 3 technical repetitions.
Because the detector was designed to sense UA in blood or saliva sample, which contained serum, we tested if serum affects the detection efficiency. In view of safety, commercial fetal bovine serum (FBS) (EVERY GREEN, 11011-8611) was used here. The growth of the bacteria was obviously suppressed when the volume of FBS fraction was more than 1/1000 (Fig. 7A, 7B), which meant 1 μL fetal bovine serum was added to 1000 μL final reaction system. When the volume of FBS fraction was 1/1000, the UA detection efficacy was unaffected by serum (Fig. 7C).
Figure 7. Impact of commercial FBS on the UA detection system. (A) Pictures of E. coli cultured with or without FBS. (B) Quantitative measurement of sample shown in Fig 7A. (C) Quantitative measurement of dsRed fluorescence / OD600 after 20 μM UA addition with or without FBS. Data were shown as mean + SD. N = 3 technical repetitions.
We verified the designed system did respond to UA in different environments. However, during our human practices, some of the interviewees were worried about the responding time, which needed 4 to 6 hours after UA induction to express strong fluorescence intensity. In their opinions, users would not wait for a 4-hour reaction. And there was another feedback that the standard of hyperuricemia is ≥ 7 mg/dL (about 400 μM) UA for men and ≥ 6.0 mg/dL (about 350 μM) UA for women [2]. If we want to use our system to detect a clinical sample directly, the sample should be diluted to 1/1000 before start, which means the gene circuit is required to detect 400× 1/1000 = 0.4 μM UA as a threshold. In one word, we need a modification to shorten the responding time and increase the sensitivity of the UA detector.
For the first question, we interviewed Dr. Xiaoyu Chen, a scientist who majors in biosensors. He suggested us that changing the fluorescent protein was a means to optimize the response speed. We blasted the sequence of dsRed we used in the database, and found the maturation half-time of dsRed is about 40 minutes in E. coli [3]. To shorten the maturation time, we decided to change dsRed to superfolder GFP, whose maturation half-time was only about 13 minutes [4]. To solve the second problem, we referenced a modular, cascaded signal amplifying methodology, which induces a module named amplifier, and it may increase sensitivity of the biosensor and boost the output expression [5]. We ordered two sequences of ultrasensitive phage activator RinA_p80α (from Staphylococcal aureus phage 80α) and a promoter PRinA_p80α. We introduced these new parts to design a new version of the UA detection system, called Version 2, shown in Fig. 8.
The processes in Version 2 were almost equal to the old version, except that the downstream of PhucR was RinA_p80α. This meant if UA presented, RinA_p80α would express and active transcription of sfGFP which was under control of PrinA_p80α. Theoretically, the new design would sense UA with much higher sensitivity than the old one. In the same time, the fluorescence production of Version 2 would get faster because that sfGFP had a shorter maturation time than dsRed in old version.
Figure 8. Working mechanism of Version 2 in E. coli.
We tested the sfGFP production of Version 2 under different concentration of extracellular UA. The curve in Fig. 9A showed the fluorescence was saturated under only 15 μM UA induction, while the old version needed about 100 μM UA to get saturated (Fig. 5B). This result verified the Version 2 had higher sensitivity than the old one. To test if sfGFP could shorten the reaction time, we used the same construct only except reporter genes, called PrinA_p80α – sfGFP and PrinA_p80α – dsRed, respectively. After adding 20 μM UA into the reaction system, the curve of PrinA_p80α – sfGFP climbed much faster than PrinA_p80α – dsRed, which suggested our new design had a great induction performance, and fitted our predictions very well.
Figure 9. The induction performances of the Version 2. (A) Induction curve of Version 2 under 0 to 200 μM UA treatment by measuring the sfGFP fluorescence / OD600. Data were shown as mean ± SD. N = 3 technical repetitions. (B) Time course experiment of sfGFP Version 2 and dsRed Version 2. Data were normalized by taking the fluorescence / OD600 of two groups at 0 h as standard, respectively. Data were shown as mean ± SD. N = 3 technical repetitions.
However, the induction time and sensitivity of Version 2 were still not reach the clinic required. In the future, there are many places need to be optimized. We can change the RBS sequence of RinA_p80α and the RBS of sfGFP. We can also change the pSB1C3 plasmid which carries PrinA_p80α – sfGFP in Version 2, to plasmids with higher or lower copy number. What's more, we can try to introduce more layers of amplifier to get high-gain amplification of output, which may make our system more sensitive.
[1] Liang, C., Xiong, D., Zhang, Y., Mu, S., & Tang, S. Y. (2015). Development of a novel uric-acid-responsive regulatory system in Escherichia coli. Applied microbiology and biotechnology, 99(5), 2267-2275.
[2] de Oliveira, E. P., & Burini, R. C. (2012). High plasma uric acid concentration: causes and consequences. Diabetology & metabolic syndrome, 4(1), 12.
[3] Bevis, B. J., & Glick, B. S. (2002). Rapidly maturing variants of the Discosoma red fluorescent protein (DsRed). Nature biotechnology, 20(1), 83.
[4] Pédelacq, J. D., Cabantous, S., Tran, T., Terwilliger, T. C., & Waldo, G. S. (2006). Engineering and characterization of a superfolder green fluorescent protein. Nature biotechnology, 24(1), 79.
[5] Wan, X., Volpetti, F., Petrova, E., French, C., Maerkl, S. J., & Wang, B. (2019). Cascaded amplifying circuits enable ultrasensitive cellular sensors for toxic metals. Nature chemical biology, 15(5), 540.
