Design, Build, Test, and Learn. This is the engineering cycle when we are designing new products. We first design new modules that fit the function we want. Then we build, building the module, creating what we design. Next, we test the module to see if it runs as expected. At last, we learn, that since the design could succeed or fail, learning from the result would give us knowledge that we could use in the next round of the engineering cycle. In areas of engineering, the engineering cycle is the most fundamental and efficient process we must go through for better products.
Synthetic biology is the subject of engineering in biology. We discover new modules, generating new genetic circuits, aiming to design new “genetic machines” with new or better functions. Our product is bacterial therapy, using Escherichia coli Nissle 1917 (EcN) as the carrier, displaying tetrazine on the surface to release the prodrug at the tumor microenvironment (TME). The design includes construction of the EcN strain; display of tetrazine on the surface of CsgA; control of click-to-release reaction; and safety module. Here is the engineering cycle we've done in our experiments:
Cycle 1: Acquisition of EcNΔCsg
Design
To guarantee the effectiveness and safety of our design, we need to change several properties of the EcN strain.
For effectiveness, the goal is to display more tetrazine on the bacterial surface. The two things that we can do so far according to the previous works1are displaying more engineered CsgA on the surface and displaying tetrazine on CsgA better. So, We started our first cycle on displaying more engineered protein: constructing a new fundamental EcN strain, EcNΔCsg.
Build
We have to knock out all the genes associated with Csg from the genome of EcN. We designed the homologous fragment, guide RNA and successfully constructed pgRNA-Csg-1. After rounds of induction and selection, we obtained bacterial strains with the target gene deleted and the plasmid responsible for deletion removed through screening.
Test
The test was conducted in two dimensions: Gene and Protein. We wanted to make sure that, the gene deletion was successful, and no curli fiber could be generated by EcNΔCsg.
1. Gene dimension
We tested in gene dimension by colony PCR. By designing two primers before and after the homologous fragments, we could see the length of homologous fragments in between, which should be 1312 bp if the deletion was successful. The Agarose Gel Electrophoresis image validated the result.
2. Protein dimension
On dimension of Protein, we performed Scanning Electron Microscope (SEM). By directly looking at the surface of the bacteria, we could check if there were curli fiber. Comparing the pictures from EcNΔCsg (Left) and EcNΔCsg + pBbB8k-CsgAWT (Right), which served as positive control, no fibers were generated on EcNΔCsg. The result validated that Csg gene was deleted successfully.
Learn
In this cycle of engineering, we learnt:
- Techniques of gene editing and Scanning Electron Microscope. They could be later used for other cycle of design.
- The displaying of Curli Fiber on the surface of EcN was abundant. If all of these fibers could carry a tetrazine per protein, the concentration should be enough for click reaction and click-to-release reaction.
Cycle 2: EcNΔCsgΔThyA
Design
To guarantee the safety of EcN, we tend to engineer upon the basis of EcNΔCsg. One design that we found might be effective is the deletion of ThyA, which was part of safety design of bacterial therapy2 that had passed Phase I trials. ThyA is responsible for the synthesis of thymidine. Knocking out ThyA is thought to enhance the bacteria's safety, proliferating in thymidine-rich area such as TME or other places we expected.
Build
It was another round of gene editing, which we had to design new homologous fragments and guide RNA. However, according to the reference3, we couldn't knock out the complete gene as what we did in Csg, since umpA also used some of the genes, knocking out both would cause death of the bacteria. So, we chose to conserve the first 149 bp of ThyA to construct pgRNA-ThyA. After rounds of induction and selection, we obtained some bacteria to be tested.
Test
We tested the bacteria in dimension of gene. As we did in last cycle, we performed colony PCR on the bacteria on plates. However, we did not find bacteria whose ThyA gene was successfully deleted.
Learn
We estimated that our failure was because inappropriate site selection of gene deletion, maybe still we delete base pair in umpA. As the experiment was conducted in September, time was very limited and we failed to get our expected results; but later we'll continue our research on ThyA deletion.
Cycle 1: Initial selection
Design
To maximum the effectiveness of tetrazine, we have to find a better site to insert Tet v2.0 into the protein CsgA. Previous work reported S89TAG (means inserting the unnatural amino acid at site 89 of the protein, which is originally the place for serine) for the insertion of pAzF4. Whether this site is still the best for Tet v2.0 needs to be tested. Also, after simulation and selection of site from the Dry Lab, we find another site F97TAG might have a better displaying property. So, we decide to have a initial selection and investigation between the two site, S89TAG and F97TAG.
Build
Based on the plasmid we bought from addgene, we successfully constructed pBbB8k-mutCsgA-F97TAG and pBbB8k-mutCsgA-S89TAG, which both the protein mutCsgA had the correct insertion site and His-tag at C-terminal. Then, we transferred each of the plasmid with pUltra-Ambrx, which was responsible for the expression of RS-tRNA pair of Tet v2.0, into EcNΔCsg. As the result, we successfully constructed the bacteria: EcNΔCsg-S89TAG and EcNΔCsg-F97TAG.
Test
We had abundant test on these two sites. However, for this cycle, we began our testing on the generation of Curli Fiber. We wanted to make sure that the insertion of Tet v2.0 wouldn't affect too much on the expression of the protein. The verification was conducted by two-steps.
1. Congo Red quantitative binding assay
Congo Red has the property of characterizing the generation of Curli Fiber. It will be absorbed to the fiber, so calculating the Congo Red left on supernatant after centrifugation could partially reflect the generation of Curli Fiber. We set EcNΔCsg with only pUltra-Ambrx and with pBbB8k-CsgAWT as control.
2. Scanning Electron Microscope (SEM)
Another test we performed was SEM, to directly find the fiber on the bacterial surface. It could also serve as supporting evidence to the results of CR quantitative binding assay.
Learn
For this cycle, we validated the formation of almost same amount of curli fiber of S89TAG and F97TAG. However, some interesting results drove to conduct the next cycle of engineering, adding new sites into our consideration. We found that, after induction, even without Tet v2.0, the protein could still be expressed as the same level of expressing with Tet v2.0. The result was validated by the pictures of SEM. We thought that it might be because truncated CsgA expression with only 2 repeats of β sheets could still be piled and formed curli fibre. Thus, we decided to select two more sites which was at the beginning of CsgA to see if we could tackle the problem.
Cycle 2: Continue selection of insertion sites
Design
After the last cycle of engineering, we decided to find two more sites at the beginning of CsgA. We selected two sites Y48TAG and Y50TAG which expressed tyrosine naturally, since the structure of Tet v2.0 is similar to tyrosine, the insertion might cause the least side effect to the folding of CsgA. For this cycle of selection, we didn't consider the display of tetrazine. We just wanted to make clear the function of truncated CsgA that we thought we observed in the last engineering cycle.
Build
As we did in the last cycle, we constructed the plasmids pBbB8k-mutCsgA-Y48TAG and obtained the bacteria EcNΔCsg-Y48TAG and EcNΔCsg-Y50TAG.
Test
In this cycle, we tested the generation of curli fiber primarily on CR quantitative binding assay. Altogether, we compared the results of the previous cycle and this cycle to verify our hypothesis.
After the experiment, we simulated the ability of each insertion site to display tetrazines. According to the simulation results from the dry lab, the F97TAG site has the strongest hydrophobicity in tetrazine display. This indicates that this site is more inclined to react rather than be exposed in the aqueous phase, and it has stronger reactivity.
Learn
According to the result, sites Y48TAG and Y50TAG didn't affect the formation of curli fiber with or without Tet v2.0, which meant our previous hypothesis was wrong. The formation of curli fiber even decreased for Y50TAG. After discussing with our PI, we thought another reason might be the false insertion of natural amino acids such as phenylalanine or tyrosine since the RS we used would have such a false positive effect. We planned to test such a hypothesis after the competition using the GMML culture medium which lacks nutrition and natural amino acids. As a result of this engineering cycle, we decided to keep our selection range within S89TAG and F97TAG, since they were simulated to display tetrazine better, and the expression of CsgA was better than Y48TAG and Y50TAG.
Cycle 3: Final selection of insertion sites
Design
After the last engineering cycle, we realized that the CR quantitative binding assay couldn't help us decide on the final site, so we continued our experiment for the next cycle. We still chose S89TAG and F97TAG to investigate. This cycle of engineering mainly continued our test in cycle 3.
Build
The plasmids and bacteria have been already built in cycle 3.TCO-Coumarin was synthesized through a two-step reaction and verified by MS.
Test
We tested S89TAG and F97TAG this engineering cycle mainly on performance on displaying tetrazine. In total, we conducted 3 sets of experiments:
1. TCO-Cy5 assay
Cy5 is a type of Cyanine dye which would generate fluorescence after excitement. TCO-Cy5 would react with tetrazine, the more tetrazine displayed on the surface, the more fluorescence signal would be generated. Using TCO-Cy5, we tested the 2 sites from 3 dimensions:
a. Bacterial dimension
In this dimension, we tend to compare the 2 sites overall. We first directly incubated TCO-Cy5 with the bacteria which were overnight induced the expression of CsgA. After rounds of washing, S89TAG displayed a better Cy5 fluorescence signal compared to F97TAG, indicating that S89TAG might express more CsgA with Tet v2.0 inserted.
b. Protein dimension
In the dimension of protein, we wanted to have a deeper insight into the expression of CsgA and the insertion of Tet v2.0. We successfully purified the Tetv2.0-incorporated CsgA variants, which we applied TCO-Cy5 for labeling, observing that only the Tet v2.0-incorporated CsgA proteins exhibited a strong fluorescent signal. Furthermore, CsgAS89TAG and CsgAF97TAG were validated by protein molecular weight mass spectrometry. Collectively, these results confirm the successful expression and purification of Tetv2.0-incorporated CsgA proteins. S89TAG exhibited a better amount of protein expression with Tet v2.0 inserted.
c. Single bacterial dimension
Though we tested in both dimensions of bacterial and protein, we still wanted to see the performance of Tet v2.0 insertion per bacteria. We tested this through Flow Cytometry and Confocal.
For confocal experiment, we also used TCO-Cy5 to stain Tetv2.0-incorporated CsgA on the bacterial surface; moreover, to visualize the bacteria, we chose DMAO and PI to distinguish between the live(only DMAO could stain) and the dead(both DMAO and PI could stain) bacteria. After appropriate adjustment of excitation and emission fluorescence channel under the CsgAS89TAG expressing with Tetv2.0, we successfully obtained the fluorescent co-localization signal of DMAO and TCO-Cy5 in CsgAS89TAG and CsgAF97TAG, while there was no Cy5 signal in CsgAWT either under the same expression condition or without induction, which confirmed that Tetv2.0 was site-specifically inserted into CsgA and well displayed on bacterial surface. Moreover, from our fluorescence co-localization results, CsgAF97TAG performed better than CsgAS89TAG, no matter considering corresponding overall fluorescence intensity or the co-localization performance.
For flow cytometry, S89TAG and F97TAG displayed a similar ratio of about 80%, much higher than that of CsgAWT which is about 10%. After analyzing 10,000 living cells, about 65% of F97TAG were detected to be labeled by TCO-Cy5, and was slightly better than the 63% of S89TAG, indicating that tetrazine on F97TAG may have a better displaying.
2. Growth Curve
While analyzing with Confocal and Flow Cytometry, we found that a great number of bacteria were dead after induction. We wanted to check if it was the problem of Tet v2.0's toxicity. So, we experimented with the growth curve, letting them grow in different conditions. After induction, S89TAG and F97TAG both with or without Tet v2.0 grew in a much worse condition.
3. TCO-Coumarin assay
Testing only with TCO-Cy5 isn't adequate, since reacting with TCO for ligation and click-to-release have slight differences. So before reacting with the prodrug, we decided to test the de-caging property by TCO-Coumarin.
TCO-Coumarin is the Coumarin whose fluorescence was partially caged by TCO. After the click-to-release reaction with tetrazine, the fluorescence intensity would increase. So, we incubated TCO-Coumarin with the bacteria and normalized by it OD600. S89TAG and F97TAG all displayed a similar de-caging effect, while S89TAG performed a faster reaction rate, and F97TAG exhibited a better reaction percentage at 30 minutes.
Learn
With all the tests above, several conclusions can help us make the decision:
- Both S89TAG and F97TAG wouldn't affect the generation of curli fiber.
- S89TAG expressed more CsgA with Tet v2.0 inserted.
- F97TAG displayed tetrazine better than S89TAG.
Because of the need of capability of reaction, we finally selected F97TAG as our ultimate site to incorporate Tet v2.0.
Cycle 1: Promoter Verification
Design
In order to improve the safety and targeting of our engineered bacteria, as well as to meet the actual needs of subsequent animal experiments and clinical applications, we replaced the promoter of the expression system with the hypoxic promoter.
To determine the expression efficiency of the hypoxic promoter and to qualitatively and quantitatively analyze the difference in expression efficiency between it and the arabinose promoter, we carried out the verification and characterization of promoter expression efficiency.
Build
To verify and analyze the difference in promoter expression efficiency, we successfully constructed three plasmids—pBbB8k araBAD+GFP, pBbB8k JWW+GFP, pBbB8k CsgA. The three plasmids were used as positive control, experimental group, and negative control, respectively. They were transferred into EcN to complete the construction of the engineered bacteria.
Test
We performed the same treatment under arabinose induction and hypoxic environment induction, and used PBS to wash the culture medium, finally obtaining a PBS solution of bacteria. In the enzyme marker, the OD600 and fluorescence value were measured, and the average was taken from three repetitions of each group. The results were plotted as a bar graph with the ratio of fluorescence value to OD600 as the Y-axis.
Using the same culture and treatment methods, but employing different induction conditions: arabinose induction or hypoxic environment induction, and reflecting promoter efficiency by measuring the ratio of fluorescence value to OD600.
We found that: compared to the arabinose promoter, the expression efficiency of the hypoxic promoter is very low. And throughout the verification process, it is difficult to control a single variable.
Learn
In response to the above situation, we proposed a new idea to improve the expression efficiency of the hypoxic promoter, introducing the T7 Promoter expression system into the EcN strain. The hypoxic promoter is used to express T7 polymerase, and then the T7 promoter is used to express the target protein, thereby achieving an apparent increase in the expression efficiency of the hypoxic promoter through a "cascade reaction".
1. Praveschotinunt, Pichet et al. “Tracking of Engineered Bacteria In Vivo Using Nonstandard Amino Acid Incorporation.” ACS synthetic biology vol. 7,6 (2018): 1640-1650.
2. Leventhal, Daniel S et al. “Immunotherapy with engineered bacteria by targeting the STING pathway for anti-tumor immunity.” Nature communications vol. 11,1 2739. 1 Jun. 2020,
3. 黄维,钟辉,李平,等.大肠杆菌胸腺嘧啶合成酶thyA缺失突变菌株的构建[J].微生物学报,2004,(01):62-66.
4. Marcus, Alan et al. “Fluorescence microscopy is superior to polarized microscopy for detecting amyloid deposits in Congo red-stained trephine bone marrow biopsy specimens.” American journal of clinical pathology vol. 138,4 (2012): 590-3.
