Engineering Success
On this page, we show the results of our many iterations and the progress of our project
In order to successfully construct engineered bacteria that can effectively remove biofilm, we need to iterate the experiment according to the experimental results during the whole project. We follow the cycle of "Design → Build → Test → Learn" and continuously improve the experimental conditions to achieve better experimental results.
We designed the experiments in four aspects: target protein purification and function test, cytotoxicity test, quorum sensing system test, lysis system test and secretion system test. The target proteins were first purified in vitro to test their ability to disrupt biofilm formation. Secondly, cytotoxicity tests were performed with purified proteins to observe in what concentration range our target proteins were harmless to humans. After determining the concentration of the protein that would be safe, we targeted the concentration of the protein by measuring the expression potency of the quorum-sensing system. Finally, we will release the target proteins through the lysis system and secretion system to destroy the biofilm.
In order to verify the disassembly ability of PslG to biofilm and to find out the protein with the strongest biofilm disassembly ability among the four HMGB1-based modified proteins (including HMGB1_Full Length, HMGB1_AB box, HMGB1_A box and HMGB1_B box), we use crystal violet staining method.
Five kinds of plasmids were constructed to express PslG, HMGB1_FL, HMGB1_AB box, HMGB1_A box and HMGB1_B box, respectively. After transferring the plasmid into DH5α for amplification, transferring into BL21 to express the target protein, cleaving BL21 and purifying using nickel column and molecular sieve, we obtained high purity of the target protein. The experiment was divided into two parts: inhibition of biofilm growth and disassembly of biofilm. In the experiment of inhibiting biofilm growth, we inoculated PAO1 and added the proteins to the mediums at the same time, and the ability of the proteins to disassemble the biofilm is reflected by measuring the growth status of the biofilm; In the biofilm disassembly experiment, we inoculated PAO1 and cultured it first, and then added proteins after it formed a sufficient amount of biofilm, so as to reflect the ability of protein disassembly of biofilm by measuring the biofilm disassembly status. The biomass of biofilm in both experiments was measured by crystal violet staining assay.
In both experiments, we treated the samples with PslG, HMGB1_FL, HMGB1_AB box, HMGB1_A box and HMGB1_B box, respectively. The group of control is protein free.
PslG indeed have an effective ability to disassemble biofilm. We illustrated that the HMGB1_FL box is the most efficient one among four HMGB1 modified proteins to disassemble biofilm. Those two proteins will be chosen to be our functional protein, and we also need a more visualized method to prove that our idea is successful. Confocal microscopy is a great choice.
The HMGB1_FL was found to be the most effective part of HMGB1 for biofilm disassembling. And also PslG is effective. We have confirmed that our target protein can effectively remove biofilm through the previous experiment, but we wanted to observe the biofilm state before and after treatment more intuitively, so we designed a more detailed observation using confocal microscopy
The HMGB1_FL was mixed with PslG to co-treat the biofilm. After co-treatment with HMGB1_FL and PSLG, the grid structure of biofilm was observed by confocal microscopy after staining with SYTO9 and HHL.
Imaging was performed using confocal microscopy after treatment with protein for 24 h and co-staining with the staining solution for 4 h.
Compared with the control group, most of the eDNA and polysaccharides in the experimental group were digested. At the same time, the biofilm thickness was significantly reduced and the grid structure was completely destroyed, indicating that our idea was successful. The co-treatment of PslG and HMGB1 indeed had a strong ring breaking effect on biofilm .
After the successful expression of these effector proteins and examination of their effect on eliminating the biofilm, it is necessary that they should be validated for safety, given that we want these proteins to function in the intestines. There are two main approaches to verifying the safety of the protein: cytotoxicity and pro-inflammatory activity, that is, the protein will not cause a large number of cells to undergo cell death within a certain range of concentration, and will not stimulate the relevant pathways of intestinal epithelial cells to cause inflammation. To this end, we designed cytotoxicity tests and inflammation verification experiments. We first added different proteins to the culture medium for a certain period of time and then calculated the cell survival rate to characterize the toxic and side effects of the protein on the cells. In addition, we also examined the effect of different proteins on the expression of cytokines in cells. If there is an up-regulation, it indicates that the pro-inflammatory activity is obvious. If there is no significant difference from the control group, it means that there is almost no obvious pro-inflammatory activity.
To accomplish our objective, we cultured HT-29 cells to mimic the intestinal epithelium and added different proteins into the medium fora certain period of time. For example, in terms of validation of the cytotoxicity, proteins were added to the culture medium and incubated for 24 hours. After that, we employed trypan blue staining to calculate the cell survival rate because trypan blue only stains dead cells blue but not living cells. For experiments related to inflammation verification, it turned out that one hour or two hours of incubation was sufficient, as this is about how long it takes from the activation of the relevant receptor by the protein to the transcription of the gene of the cytokine, which is what we want to measure. Therefore, we first lysed the cells, extracted total RNA, reverse transcribed them, and then subjected them to real-time PCR by adding primers for the specified cytokine genes. Thus, we could examine the pro-inflammatory activity of different proteins in terms of the amount of transcription.
Different concentrations of engineered HMGB1 (full length) and PslG were added into the cell culture medium and we incubated the cells for a certain period of time. After several attempts, we found that those proteins would not impose a fatal influence on intestinal cells in humans considering they would only cause strong toxic effects at very high concentrations (over 2 μg/mL). What’s more, they are unlikely to exert a lethal impact concerning inflammation of human intestinal cells according to the results.
The safe concentration range of the effector proteins was explored by cytotoxicity experiments, and it was verified that the concentration range did not cause inflammatory responses in intestinal cells. After discussion, we believed that it was necessary to explore the secretion potency of the secretion system of the engineered bacteria, so as to ensure that the concentration of secreted proteins would not have a serious effect on human intestinal cells.
The sensing device was designed based on the Type I quorum sensing mechanism of P. aeruginosa to effectively produce PslG and HMGB1 only in response to the presence of P. aeruginosa. It is important to figure the relationship between target protein expression rate and induction concentration of 3OC12HSL. However, production of target protein is hard to measure. In this work, we use green fluorescent protein (GFP) signal as a reporter to quantificationally characterize the quorum sensing device by flux gene that encoding GFP to sensing device to build a test circuit.
As described above, we build the plasmid: ptetR-LasR-pluxR-GFP. The tetR promoter constitutively enhances expression of transcriptional factor LasR, which can bind to AHL 3OC12HSL.Then, LasR-3OC12HSL activator complex can open luxR promoter, leading to production of GFP which can be quantificationally measured by microplate reader.
Single colonies of ptetR-LasR-pluxR-GFP (BL21) were induced with 3OC12HSL at varying molar concentration (5e-9,1e-8, 2.5e-8, 5e-8, 1e-7, 5e-7, 1e-6, 5e-6, 1e-5). Time-series fluorescence and OD600 data were obtained at intervals of 10m for a total run time of 3h. We observed a basal expression level of 0.427 RFU per OD per minute with lower concentration (< 5.0E-9M) of 3OC12HSL induction, followed by a sharp increase in GFP production rate as the concentration of 3OC12HSL was increased beyond1.0E-8M. GFP expression achieved a peak at 1.0E-7M. These results announced that the optimal detection range of the sensing device was higher than 1.0E-7M3OC12HSL. At the same time, we also used Hill equation to curve fit the relationship between the time-average GFP expression rate and 3OC12HSL concentration.
To verify that the sensing device would be able to sense the natively produced 3OC12HSL by P. aeruginosa, the sensing device was induced by the filtered culture of P. aeruginosa. And the result show that GFP expression rate was 1.159 RFU per OD per minute so that we can estimate the concentration of natively produced 3OC12HSL was 5.0E-7M.This value is higher than 1.0E-7 which confirmed that the sensing device was able to detect the natively produced 3OC12HSL.
We confirmed that the sensing device have a lower leakage expression and a sharp induced response curve that can effective induce production of target protein with natively produced 3OC12HSL. It is excited that GFP expression rate reach the peak in 40min (> 2.5 RFU per OD per minute) and then become lower with a rapid decline, which means the quorum sensing device has a very fast response rate and the regulation of 3OC12HSL is instantaneous. However, the overall induced expression of GFP remained unsatisfactory, increasing the number of copies of pluxR may be absolution. At the same time, how to secrete the protein we induce to extracellular expression is also a problem we need to consider next.
Our previous experiments demonstrated that the quorum sensing factor of Pseudomonas aeruginosa can successfully activate the expression of downstream proteins. In a practical application scenario, we need the engineered bacteria to express two proteins and release them to the outside world to act on the biofilm. We tried to let bacteria lyse itself by expressing lysis E7 protein, in order to release a large number of disassembly proteins in a short time near the biofilm like a "demolition truck". We also tried to insert lysis E7 into arabinose operon, so as to achieve artificial regulation of cell lysis. Here we need to verify whether this lysing system can function properly.
We first construct the plasmid in which lysis E7 is inserted into Arabinose operon by homologous recombination and transfer it to BL21. Next, we designed two detection methods: First, 1%, 0.1% and 0.01% arabinose were used for induction, while a control group inhibited by 0.2% glucose was added (theoretically, the function of the arabinose operon could be effectively inhibited), and growth curves were drawn to detect the induction effect and lysis status. Second, arabinose induced group, non-induced group and non-induced physical lysing group were set up. After the bacterial solution of the three groups was cultured for a period of time, the supernatant was centrifuged to detect the DNA content, and the lysing effect was measured by comparing the DNA concentration with each other.
Arabinose-induced growth curve: The bacterial solution with an OD600 absorbance of 2 was diluted 1:100 and cultured until the OD600 absorbance reached 0.5. 1%, 0.1% or 0.01% arabinose was added for induction, while the control group inhibited by 0.2% glucose was added. The OD600 absorbance was detected every 15min and a curve was drawn.
DNA concentration detection of supernatant: The bacterial solution with an OD600 absorption value of 2 was diluted 1:100 and cultured until the OD600 absorbance reached 0.5. The treatment group was induced by adding 1% arabinose, while the two control groups were arabinose free. After culturing for 5h, we collect the supernatant of three groups by centrifuging (one of the two control groups was cracked by freezing and thawing repeatedly before centrifuging). Test their DNA concentrations.
After adding arabinose of different concentrations, the growth of bacterial solution concentration slowed down and began to decline at 50-100 minutes. The induction effect of 1% arabinose and 0.1% arabinose is basically the same, and both can greatly reduce the concentration of fine liquid. However, the induction effect of 0.01% arabinose was relatively weak, and the concentration of bacterial solution showed a rising trend when it was about 4h20min. The curve of the control group treated with 0.2% glucose reached a plateau at about 3h, and there was no downward trend. The DNA content of the bacterial supernatant after 1% arabinose induction was basically the same as that of the physical cracking group, both of which were significantly higher than that of the non-induction group.
The results of these two sets of experiments indicate that arabinose operon can function normally, and the cracking system we constructed can also be artificially regulated. It means we have at least one way to ensure a relatively sufficient proteins can reach the biofilm to serve the function of disassembly. Based on this, we can add some signal peptides to help the functional proteins be secreted out of the cell. Under this condition, lysis E7 can become a backup solution which will only used when there is a large amount of biofilm.