2×2 transfer switch

Cycle I

As mentioned in the design page, we decided to construct a bistable switch as a basic frame because this kind of switch is efficiently manipulated and the boundaries between each state of these switches are clear. Given food safety and popularization, our bistable switch should be regulated in a nontoxic and accessible way. The two typical bistable switches - cl&cro type is relatively cumbersome in design; tetR&lacl type uses IPTG as a regulator, which cannot be used as food. Thus, we decided to create a brand new composite bistable switch modulated by natural light and temperature.

We designed two plasmids (pET-28a) containing our switch sequences to verify their practicability. After managing to transform our plasmids into the competent cell (DH5α), We cultured them in four circumstances respectively. To maintain the temperature, we placed the conical flasks filled with bacterial solution (2/5 of the volume) in our incubation device designed in our Hardware part.

To guarantee the light variation, we used tinfoil to cover one of the conical flasks, and used blue-colored transparent wrapping up the white illuminant and the euphotic form of the shaker. After 20h-incubating, we took out 15μL bacterial fluid to make a temporary slide and observed it through the fluorescence microscope in the excitation wavelength corresponding to the protein.

During the first cycle, we overcome several problems

1. We were unfamiliar with the use of hardware and software for fluorescence microscopy, failing to take clear results. The problem was well solved after the advisor's guidance. Meanwhile, there were unidentified blocks that reflected intense fluorescent light. After comparison with the blank control, we speculated that these substances may be internal impurities of the slides or accumulations of protein after bacterial fragmentation.

2. We did not quantitate the expression of the protein and lacked pre-experiments to explore its optimal incubation time. As a result, the expression of fluorescent proteins we observed appears to be relatively low after incubating for12h, 16h, and even 20h. Therefore, we readjusted the ratio of bacterial solution to liquid medium (from 1:200 to 1:1000 in volume) and extended the incubation time to 36 hours in the next cycle.

Cycle II

After Redesigning and rearranging our experiment according to those problems listed above in Cycle I, we managed to obtain a clear result of our bistable switch. For results and more information, please visit the proof of concept page.

However, problems still existed.

1. In terms of the whole project, when choosing our reporters, we focused on fluorescent proteins (For example EYFP) instead of Chromoproteins, which caused a lot of difficulty in utilizing the fluorescence microscope to obtain results. Additionally, we did not consider the quenching of fluorescent proteins and we neglected the damage to the cells done by these proteins. We plan to change our reporter proteins with Chromoproteins in our future design.

2. Due to the coronavirus pandemic and inadequate staffing, we only validate the practicability of each pathway in our switch. The changes between different statuses of our switch have not been completely texted. Within our design, the switch should be changed along with the existence of the blue light. Though being able to express in E. coli, our switch is not sure if it can operate properly in lactobacillus. The further experiment should be conducted on food-grade bacteria like lactobacillus

Flavor Substance

Vanillin

Cycle I

It is mentioned in the contribution part that nine enzymes are involved in the pathway from carbon sources to vanillin. Considering the high price of ferulic acid, we refused to use it in the mass production of meal replacement. Biosynthesis of vanillin from simple carbon sources is much more attractive because they are much cheaper and more readily available. As a result, we chose carbon sources, such as glucose and glycerol, as the ingredient.

The production of vanillin involves three plasmids, but it is difficult to transfect three plasmids into one strain because bacteria are easy to die under that condition. So we planned to test the functions of three plasmids respectively before merging them. E.coli strains with one of these plasmids were cultured in LB medium. The overnight LB culture was inoculated to LB medium with the corresponding ingredient(glucose, glycerol, L-tyrosine, and ferulic acid), and IPTG was added after it grew for several hours. Culture samples were collected at an interval of 3 or 6 hours and analyzed by High Performance Liquid Chromatography (HPLC).

During the first experiments, we found many problems and came up with some improvements.

We added a preset concentration of 1g/L, 2g/L, and 4g/L of ferulic acid as the ingredient to LB medium to culture E.coli with the third plasmid(BBa_K4256013). If bacteria had grown vigorously, the culture should have been cloudy liquid. On the contrary, it was out of expectation that the solution with 1g/L of ferulic acid became cloudy slowly, and the latter two solutions kept clear and transparent for over 10 hours, which might indicate that the concentration was too high for E.coli. Thus, in the second cycle, we decreased the concentration of ferulic acid to 0.05g/L.

The bacteria were not broken before analyzing the composition of samples containing strains with the second plasmid(BBa_K4256012) in the first cycle, and we didn’t obtain the peak signal for ferulic acid from HPLC. We supposed that ferulic acid might remain in bacteria instead of being exported across the membrane, which required fragmentation of bacteria. We decided to add this step to the protocol in the second cycle.

In a gesture to validate the functions of the first plasmid(BBa_K4256011), we monitored the production of L-tyrosine by measuring the absorbance at 274nm during HPLC analysis. However, many metabolites of bacteria also have absorbance at 200-300nm. Although we set up an experimental group and a control group, there was still a lot of overlap in the peak profiles from the HPLC analysis. Further research may focus on a more selective test method for tyrosine.

Cycle II

Besides the first two improvements mentioned above, we changed the intervals of sample collection to 12 hours. Since we decreased the concentration of the ingredient, a prolonged induced expression might bring more vanillin, making the detection results more obvious. Due to the time limit, we just transfected the third plasmid to verify its functions.

Problems still existed, most of which were in the HPLC analysis. We followed the gradient elution program from the reference, but the chromatographic column was of a different type and brand from that in the original literature. When we tested the standard samples, signals of main compositions came out just before the end of gradient elution, indicating that this program was not adapted to our column. Our next step would be to consult professional engineers and adjust the gradient elution program. Although encountered many problems, we were pleased that we learned new knowledge and skills in this process, given that HPLC was a new instrument for us and no one had used it before.

2-Phenylethanol & 3-Methylbutyraldehyde

The initial design in the synthesis of 2-phenylethanol was prepared by using three plasmids to bear three protein genes each, and then we found that such a design faced cost and plasmid incompatibility problems. By studying the structure of the bacterial genome, we found that multicistron can be constructed, so only one plasmid was needed to carry the three genes. In order to verify the experiment, I added GFP to the last part of many cis-trans substrates. Subsequent experiments showed that the protein was successfully expressed (the molecular weight measured by SDS-PAGE was consistent with the theoretical value), but the phage and the broken supernatant showed no fluorescence. All the gene sequences we used in this series are from NCBI, so we can assume that the gene sequences are correct. Finally, we found that no linker amino acids was added during the process of constructing ADH-GFP fusion protein, which may lead to the fact that the GFP part of the fusion protein could not be folded correctly after expression, so it did not emit fluorescence. The linker (oligo glycine) was added in the subsequent design, but further experiments were not performed to verify it due to time constraints. A similar plasmid construction method and validation step design was used in the synthesis of 3-Methylbutyraldehyde.

We used HPLC/MS to detect 2-Phenylethanol. We used the protocol 1 to first test our sample supernatant. But it doesn’t seem to go well (Figure 1). Reflecting on our experimental design, we decided to focus on several significant flaws, including the processing of samples, the quality of the supernatant, and the sensitivity of the detection method. Then, we restarted our experiment by utilizing the HPLC/MS, changing the dilution solution from water to methanol in order to filter the insoluble impurity. Then changing the protocol to possibly distinguish 2-phenylethanol in different wavelengths. After HPLC/MS result demonstrated that in such method, 2-phenylethanol could be combining with Na+ , which gives additional difficulty to the test. Moreover, water-soluble impurities disrupted the result. By changing the protocol again, the HPLC/MS result appeared to be much better than before, giving a preliminary conclusion of the existence of 2-phenylethanol.

Suicide switch

Cycle I

Design:

In the original design, our suicide switch is regulated by temperature conditions. In this suicide system, the killing task is achieved by toxin Doc, and its C-terminal structural domain is fused with tevS/ssrA tag. SsrA tag can help degrade Doc toxin, while tevS is used for protein hydrolysis to remove the ssrA tag. When the temperature is below 34 ℃, a temperature-regulated promoter controls tevS expression and excises the ssrA tag, then the toxin will kill engineered bacteria; at 37 ℃, tevS expression is inhibited and the ssrA tag cannot be removed, so the Doc toxin is degraded and the engineered bacteria can survive.

Learn:

However, the problem with this design is that the temperature in the storage conditions of the engineered bacteria is often lower than 34 ℃, so our engineered bacteria may be killed before they can function.

Cycle II

Design:

After preliminary literature research, we decide to use the difference in oxygen concentration between the engineered bacteria's storage environment, working environment, and atmospheric environment to implement the suicide switch function. Our initial idea is to modify the previous temperature-sensitive suicide switch by replacing the temperature-sensitive promoter with a PhoPR promoter that responds to changes in dissolved oxygen levels in the PhoPR two-component system (TCS) of Corynebacterium glutamicum. Under atmospheric oxygen concentrations, the expression level of genes downstream of the PhoPR promoter is low; under anaerobic conditions, the expression level of genes downstream of the PhoPR promoter will increase significantly. Therefore, we could remove the tevS tag in the C-terminal structural domain of Doc toxin and fuse only the ssrA tag, designing the expression of the ssrA tag to be initiated by the regulation of the PhoPR promoter. In the anaerobic conditions, the ssrA tag plays its role, the Doc is degraded so the engineered bacteria will survive; in the normoxic escape environment, the ssrA tag does not function and Doc kills the escaped engineered bacteria.

Learn:

But this system may not be mature enough. After discussion, we concluded that using this TA system might pose a higher risk of failure. Finally we agreed to choose a more mature TA system.

Cycle III

Design:

In the subsequent design process, we decide to use the more mature ccdA-ccdB toxin-antitoxin system to replace the Doc toxin-ssrA tag system, using the T7 promoter to regulate the toxin ccdB expression and the oxygen-sensitive promoter to regulate the antitoxin ccdA expression. In addition, because the two-component system of Corynebacterium glutamicum uses a two-component signaling cascade for oxygen concentration sensing, which involves a complex mechanism, we change the oxygen-controlled promoter used from the PhoPR promoter to the HIP-1 promoter regulated by FNR (a transcriptional regulator that is a sensing protein containing the [Fe-S] cluster). At this point, the design of our suicide switch has been primarily finalized.