Part:BBa_K5242001
ADAR2_MCP
1.Biology and Usage
A-to-I editing exists in many metazoan organisms, and ADAR (Double-Stranded RNA-Specific Adenosine Deaminase) is the crucial enzyme that catalyzes the process. The three human ADAR genes give rise to four known isoforms: ADAR1_p150, ADAR1_p110, ADAR2 and ADAR3. Among these isoforms, ADAR1_p150 can shuttle in and out of the nucleus, while ADAR1_p110 and ADAR2 exist in the nucleus. Because mature RNAs exist in the cytoplasm, we choose ADAR_p150 and engineered MCP-ADAR2 to be introduced to yeast cells.
This part is codon-optimized ADAR2-MCP gene, with a Flag-tag added at the N-terminal for Western blot. What's more, we altered the subcellular localization of ADAR2-MCP by adding NES (Nuclear Export Signal). The hADAR2 protein is composed of 2 dsRNA binding domains and a deaminase domain. To improve the editing efficiency, we applied the method of previous work. We replaced the dsRNA binding domain of the ADAR2 gene with an MCP domain, which can improve the editing efficiency largely. Compared to the dCas13 protein, which was applied in the project of NTU-Singapore-2018-and-2019 and the dCas9 protein, which was applied in the project of WPI-Worcester-2016, our ADAR2-MCP is much smaller.
Usually, ADAR2_MCP doesn't work when the double-stranded RNA is well matched, while this deamination process can be accelerated by secondary structure elements like terminal loops, internal loops, bulges and mismatches. Especially, the ADARs can edit the A-C mismatch point in Double-stranded RNA with high specificity, which can be used for RNA editing and RNA sensing. In our design, ADAR_p150 is an exogenous protein introduced into yeast cells to edit our sensor, which contains an A-C mismatch site. The design is shown in Figure 1.
Figure 1.The design of our senser RNA.
2.Experimental Characterization
2.1 Preliminary Experiment
2.1.1 Gibson Assembly and Yeast transformation
Initially, given that ADAR2_MCP is an exogenous protein, we selected weaker promoters PDC1 (for homologous recombination) and HXT7 (for free plasmid expression), along with the HHF2 terminator and the ADH1 terminator, respectively, to construct the expression cassette for ADAR2-MCP.
We used Gibson Assembly to build our plasmid and we cultured transformed E. coli to get more plasmids. We simultaneously tried two methods to express our proteins: Homologous Recombination (pUC) and Free Plasmid Expression (pCEV).Then, we sent the 2 plasmids to the company for sequencing. The result is correct. It indicates that our assembly was successful.
Figure 2 . The sequencing results indicated that our plasmid construction for Free Plasmid Expression was successful.
Figure 3 . The sequencing results indicated that our plasmid construction for Homologous Recombination was successful.
After that, we performed a successful yeast transformation.
2.1.2 Gene expression
After transformation, qPCR and Western blot were done to check whether the transcription and translation processes worked well. Because ADARs may be toxic to yeast cells, we found that the yeast strains with free plasmid expression grew quite slowly. We only conducted the experiments on the strains of Homologous Recombination.
2.1.2.1 qPCR
We used ACT1 as an internal reference for the qPCR experiment. What's more, the transformation of 4 sensor plasmids is done before qPCR. The result is shown below. The mRNA level of ADAR2_MCP is close to ACT1. Because ACT1 is expressed in yeast cells at a low level, we need to use a strong promoter in our future improvement.
Figure 4: The qPCR rusult of ADAR2_MCP.
2.1.2.2 Western blot
Figure 5: The Western Blot rusult of ADAR2_MCP, we did not observe a band for ADAR2_MCP. Naming rule: '- T' represents total, '- S' represents supernatant, and' - P 'represents precipitate
Interestingly, we did not observe a band for ADAR2_MCP. However, in other experiments, we found that systems involving ADAR2_MCP exhibited high editing efficiency (compared to the blank control), which confirms that ADAR2_MCP was indeed expressed in the yeast. It was simply not detected in the Western blot results. To explain this, we propose the following hypotheses:
1. ADAR2_MCP may have high intracellular activity but is rapidly degraded outside the cell, and the prolonged duration of the Western blot procedure may not favor the long-term stability of ADAR2_MCP.
2. Another possibility is that the spatial structure of ADAR2_MCP is such that the FLAG tag at the protein's terminus may be folded into the interior of the protein, making it difficult for the primary antibody to bind.
2.1.3 Editing efficiency analysis
Because our sensor RNAs contain eGFP and eBFP, we introduced our sensor plasmids into transformed cells to indicate editing efficiency.
Our sensor RNA is composed of eGFP, E2A peptide and eBFP. A termination codon is at the end of eGFP. If the sensor RNA is edited by ADAR, the termination codon UAG will be transformed to UGG, which can be translated into tryptophan. E2A is used to cut the long peptide chain into two fluorescent proteins, so that the two fluorescent proteins would not interfere with each other. By measuring the fluorescence intensity and calculating eBFP/eGFP, we can quantify the editing efficiency of ADARs. Our gRNA can complementarily pair with the sensor RNA and become an ADAR-editable A-C mismatch at the termination codon of the sensor RNA.
To make sure our sensor RNA expresses with gRNA at the same time, we loaded them into the same plasmid, which is called pSensor1. Also, we set another 2 groups as negative control and 1 group as positive control. In pSensor2, gRNA is mutated; it can bind with sensor RNA perfectly, without a A-C mismatch. We set up this control group to verify the effect of A-C mismatch on ADAR editing. In sensor 3, gRNA is replaced by another gRNA, which will not form dsRNA with our sensor RNA. We set up this control group to verify the effect of dsRNA on ADAR editing. In pSensor4, sensor RNA is mutated; the terminal codon UAG was already edited to UGG. We set this control group to check if the 2A peptide and downstream eBFP could work normally.
According to our experimental design, the predicted results are as follows: The pSensor1 group, which serves as the experimental group, is expected to exhibit blue fluorescence; the pSensor2 group will not display blue fluorescence because there are no A-C mismatch sites available for editing; the pSensor3 group will not show blue fluorescence due to the lack of double-stranded RNA formation; and the pSensor4 group, because the UAG codon has been edited to UGG, will display blue fluorescence with the highest intensity.
2.1.3.1 LSCM (laser scanning confocal microscope) Analysis
After going through the procedures of transformation, incubation and sample preparation, we put the four groups of samples under the microscope and observed the results as shown in Fig7.
We observed blue fluorescence in both pSensor1 and pSensor4. The result indicated that both our ADAR2_MCP and our sensor RNA could work in yeast cells. However, the blue fluorescent in pSensor1 is much weaker than pSensor4. Therefore, we need to improve our sensing system to enhance its editing efficiency.
Figure 6: The LSCM rusult of ADAR2_MCP.
2.1.3.2 Flow cytometry analysis
Figure.12 has shown the result. Unfortunately, for both ADAR1_p150 and ADAR2_MCP, between pSensor1, psensor2 and pSensor3, we did not observe significant differences. While the eBFP/eGFP value of pSensor4 is close to 100%. The result showed that we need to improve the editing efficiency.
Figure 7: The result of Flow cytometry analysis
2.1.4 Subcellular Localization of ADAR2_MCP
The dsRNA is mainly formed in the cytoplasm, and the subcellular location of ADARs may have a negative impact on the editing process. Therefore, we set out to investigate the subcellular location of ADARs. We used the GGGGS linker to fuse mScarlet to the C-terminus of ADARs, allowing us to observe the subcellular localization of ADARs under a microscope.
As the Figure 8 shown, ADAR_MCP exhibited a clear exclusion of fluorescence from the nuclear region, indicating predominant cytoplasmic localization. This may be attributed to the presence of a nuclear export signal (NES) added to ADAR_MCP.
Figure 8: The Subcellular Localization of ADAR2_MCP
2.1.5 ADAR toxicity verification
During pre-experimental cultures, we observed that yeast introduced with ADARs grew more slowly than normal yeast. We suspected that this might be due to the RNA editing activity of ADARs that allowed the modification of key RNAs during yeast growth resulting in the inhibition of yeast growth. To better control the copy number of our genes, we utilized homologous recombination to express two ADAR proteins in yeast using two promoters of different strengths. The groupings in the experiment are shown in Table 1.
| Group | ADAR | Promoter | Expression intensity |
|---|---|---|---|
| B | \ | \ | \ |
| A1 | ADAR1_p150 | PDC1 | Low |
| A2 | ADAR2_MCP | PDC1 | Low |
| T1 | ADAR1_p150 | TEF1 | High |
| T2 | ADAR2_MCP | TEF1 | High |
To determine the optimal initial concentration , we conducted multiple sets of preliminary experiments. The final determined initial concentration was OD600=0.001.The results of the incubation of OD600=0.001 in the pre-experiment are shown in Figure 9, and the final experimental results are shown in Figure 11.
Figure 9.The results of the incubation of OD600=0.001 in the pre-experiment.
Figure 10.The final experimental results.
Interestingly, in the preliminary experiment, the ADAR2_MCP strain with the TEF 1promoter exhibited a severe inhibitory effect on yeast growth, whereas in the current experiment, the TEF1 promoter promoted faster growth compared to the weak promoter. Referring to other team members' results, T2 yeast does indeed grow slower under similar conditions, aligning more closely with the pre-experiment results. This suggests that there may have been an error in the preparation of the T2 strain during this toxicity experiment, potentially leading to an over-concentration of the T2 solution.
2.1.6 Conclusion of Preliminary Experiment
In summary, the preliminary experiment basically proved that our RNAssay is feasible. ADAR2_MCP is correctly expressed in yeast and play certain functions, but the editing efficiency is relatively low. Subcellular localization experiments show that ADAR2_MCP is mostly located in the cytoplasm. The toxicity test of ADAR2_MCP proves that ADAR2_MCP does have certain cytotoxicity. In formal experiments, we had a total of two groups involved in testing ADAR2_MCP itself, with one group testing how the editing efficiency of ADAR2_MCP could be improved, and the other group testing the specificity of ADAR2_MCP in recognizing similar sequences.
2.2 Editing System Optimation
After obtaining preliminary experimental results, we realized that the editing efficiency of ADAR was relatively low. Therefore, we attempted the following two improvements: 1) replacing the promoter with the stronger TEF1 promoter, and 2) increasing the number of MS2 sites around the binding region, with quantities of 0, 2, and 4 respectively. After optimizing the ADAR editing efficiency, we designed a plasmid induced by xylose expression. By measuring the editing efficiency and the expression level of the target transcript, we can semi-quantitatively detect the expression level of the target transcript.
2.2.1 Promoter Replacement
2.2.1.1 Plasmind Construction and Yeast Transformation
The plasmids are basically the same as pUC19-C1and pUC19-ADAR1-C2. We only replaced the PDC1 promoter with the TEF1 promoter. As before, we used the Gibson Assembly to build our plasmids. The sequencing results of the two plasmids are shown in the Figure11, proving that the plasmid construction was successful.
Figure 11.The Sequencing results
After that, we performed a successful yeast transformation.
To reduce the experimental burden, the effect of the ADAR promoter on the fluorescence intensity will be characterized later together with the pair fluorescence intensity of MS2. This section only shows the results of qPCR.
2.2.1.2 qPCR
To verify whether there was an increase in transcription after switching to a different promoter, we performed qPCR experiments, the results of which are shown in Figure 14. The relative transcript level of ADAR2_MCP relative to ACT1 increased significantly.
Figure 12. The relative transcript level of ADAR2_MCP relative to ACT1 increased significantly.
2.2.2 Add MS2
2.2.2.1 Plasmid Construction
Literature suggests that adding MS2 sequences on both sides of the binding region can improve ADAR editing efficiency. Therefore, we designed three types of sensors with 0, 2, and 4 MS2 sequences around the binding region. As before, we used the Gibson Assembly to build our plasmids. The sequencing results of the two plasmids are shown in the Figure15, proving that the plasmid construction was successful.
Figure 13.The Sequencing results
After that, we performed a successful yeast transformation.
2.2.2.2 Edition efficiency analysis
We replaced the fluorescent protein due to the relatively weak intensity of the blue fluorescence.We placed the red fluorescence before the stop codon and the green fluorescence after the stop codon. In this way, when sensor RNA is edited, there will be green fluorescence.
We planed to include a side-by-side comparison of the two ADAR proteins at the same time. In our experimental design, we introduced two parts into yeast.The first part is the ADAR protein expression, with four gene types: PDC promoter + ADAR1, PDC promoter + ADAR_MCP, TEF promoter + ADAR1, and TEF promoter + ADAR_MCP. The second part is the sensor, with three types containing 0, 2, or 4 MS2 sequences. The overall design is shown in Table 2. This method allows us to determine the combination with the highest editing efficiency. After construction, we observed these 12 engineered yeast strains using confocal microscopy and flow cytometry to calculate ADAR editing efficiency.
| Group | Promoter | ADARs | the Number of MS2 |
|---|---|---|---|
| 1 | TEF1 | ADAR1_p150 | 0 |
| 2 | TEF1 | ADAR1_p150 | 2 |
| 3 | TEF1 | ADAR1_p150 | 4 |
| 4 | TEF1 | ADAR2_MCP | 0 |
| 5 | TEF1 | ADAR2_MCP | 2 |
| 6 | TEF1 | ADAR2_MCP | 4 |
| 7 | PDC1 | ADAR1_p150 | 0 |
| 8 | PDC1 | ADAR1_p150 | 2 |
| 9 | PDC1 | ADAR1_p150 | 4 |
| 10 | PDC1 | ADAR2_MCP | 0 |
| 11 | PDC1 | ADAR2_MCP | 2 |
| 12 | PDC1 | ADAR2_MCP | 4 |
Figures 14 and 15 clearly show that the addition of MS2 sequences effectively enhances the editing efficiency of the entire system. However, an unexpected observation is that increasing the number of MS2 sequences leads to a decrease in editing efficiency for both TEF-ADAR1 and TEF-ADAR2. This outcome contrasts with the results observed for PDC-ADAR2. We hypothesize that the MS2 sequences may interfere with the transcription process, potentially causing premature termination of the mScarlet-EGFP fusion protein. When ADAR1/2 is driven by the PDC promoter, the number of MS2 sequences correlates with increased binding affinity, leading to enhanced editing efficiency. In contrast, under the control of the TEF promoter, the high expression levels of ADAR1/2 may lead to saturation, making the concentration of sensor RNA a limiting factor. Given that MS2 sequences might increase the likelihood of abortive transcription, a higher number of MS2 sequences could reduce the available concentration of sensor RNA, thereby decreasing editing efficiency.
Figure 14.Confocal Results of Editing Systems Containing Different MS2 sequences
Figure 15.FACs Results of Editing Systems Containing Different MS2 sequences
Furthermore, Figures 16 and 17 highlight that TEF-ADAR2 exhibits the highest editing efficiency, likely attributable to its elevated expression level.
Figure 16.Confocal Results of Editing Systems Containing Different ADAR Types and Promoters
Figure 17.FACs Results of Editing Systems Containing Different ADAR Types and Promoters
2.2.3 Quantitative characterization of editing intensity
We designed a xylose-inducible promoter upstream of the target transcript. Under low xylose concentrations, xylR can inhibit the transcription of the xylose promoter. However, when the xylose concentration increases, xylose binds to xylR, causing it to lose its repressive activity and thus promoting the expression of downstream genes. The other parts of the plasmid remain essentially the same as before, so they are not elaborated here. Figure 18 shows the structure and principle of our xylose-inducible promoter.
Figure 18.The Structure and Priciple of our Xylose-Inducible Promoter
2.2.3.1 Plasmid Construction and Yeast Transformation
As before, we used the Gibson Assembly to build our plasmids. The sequencing results proved that the plasmid construction was successful.
2.2.3.2 Results of qPCR and FACs.
As Figure 19 shown, no clear relationship was observed between the target transcript concentration and editing efficiency. We hypothesized that this could be attributed to an insufficient induction period. Consequently, we planned to extend the induction time from 12 hours to 24 hours in subsequent experiments.
Figure 19. No clear relationship was observed between the target transcript concentration and editing efficiency.
However, in our next experiment, we were unable to detect red fluorescence under the microscope, suggesting a potential issue with homologous recombination, despite obtaining correct results in the colony PCR. Due to the extended duration of the experiment and the time constraints imposed by the start of the new semester, we were unable to repeat the experiment at this stage. Moving forward, we aim to obtain a more robust dataset in order to enhance the overall completeness of our project.
2.2.4 Conclusions of System Optimization
Through the above experiments, we demonstrated that switching promoters did increase the expression of the two ADAR proteins, however, although for ADAR2_MCP there was a significant increase in the editing efficiency after switching promoters, for ADAR1_p150 the increase in the editing efficiency by switching promoters was not significant. Therefore, we abandoned the idea of using ADAR1_p150 during the subsequent determination of the change curve of fluorescence intensity with RNA level. In addition, although the addition of the MS2 loop did improve the editing efficiency of the two ADARs, the effect of the amount of MS2 on the editing efficiency was more complex, which may imply that there is still an unknown mechanism in the process of ADAR and RNA action.
In addition, we also tried to find the quantitative relationship of fluorescence intensity with target RNA transcription level, although the results were not very satisfactory due to the limitation of time.
2.3 Application for in vivo dynamic monitoring of splice variant
Due to the presence of half identical sequences in the vicinity of the splice heterodimer's junction, a proof-of-concept for the splice heterodimer is somehow equivalent to a characterization of ADAR2_MCP's ability to recognize and edit similar sequences.
Knowing that Chk1 and Chk1s are a pair of splice heterodimers, we set up four sets of experiments to examine the potential of ADAR2_MCP to detect splice heterodimers. In the experiments, we applied the same plasmid importing sensor RNA and target RNA. And we used the TEF1 promoter and 4 MS2 loops. The groupings and predicted results are shown in Table 3.
As before, we have also included in this section a side-by-side comparison of the two ADAR proteins
| Group | the Target of Sensor RNA | the Transcribed RNA | Green Fluorescence |
|---|---|---|---|
| kk | Chk1 | Chk1 | Yes |
| ks | Chk1 | Chk1s | No |
| sk | Chk1 | Chk1s | No |
| ss | Chk1s | Chk1s | Yes |
2.3.1 Plasmid Construction and Yeast Transformation
As before, we used the Gibson Assembly to build our plasmids. The sequencing results of the two plasmids are shown in the Figure 20,Figure 21, Figure 22,Figure 23, proving that the plasmid construction was successful.
Figure 20.The sequencing results
Figure 21.The sequencing results
Figure 22.The sequencing results
Figure 23.The sequencing results
After that, we performed a successful yeast transformation.
2.3.2 Fluoresence Observasion
2.3.2.1 LCFM Analysis
Figures 24 and 25 show the fluorescence microscopy observations. However, contrary to the expected results, for ADAR1, all four groups showed green fluorescence, while for ADAR2, all but the last group showed green fluorescence. We decided to use flow cytometry to count the percentage of each group that showed green fluorescence.
Figure 24.The LCFM results of ADAR1_p150
Figure 25.The LCFM results of ADAR2_MCP
2.3.2.2 FACs Analysis
We got nice data from Flow cytometry analysis, though the results were not exactly as what we expected. The editing efficiency are shown in Figure 26 and Figure 27.
Figure 26.The FACs results of ADAR1_p150
Figure 27.The FACs results of ADAR2_MCP
2.3.3 Conclusions of in vivo dynamic monitoring of splice variant
According to the datas, we came to some conclusions:
①ADAR2_MCP has the higher edit efficiency at about 80-95%, but this causes no evidence of selectivity. So we couldn't evaluate whether our sensor function well or not.
②The extremely high efficiency of TA2 in yeast shows the potential of becoming a powerful gene edit tool in yeast.
③ADAR1_p150 has the edit efficiency >15% according to the datas from our Optimization group, which is significantly higher than the stop codon readthrough or off target effect (edit efficiency is around 5%). So we could infer that TA1 has a strong selective edit effect for pSensor-Chk1 with Target-Chk1. But TA1 only has the edit effect on pSensor-target_Chk1-Chk1, that was beyond our expectations.
④Design of Sensor is important for the selectivity of ADAR. After the experiment, we put the sequences into IntaRNA and predict the binding energy and binding area of sensor and its corresponding target. According to the results differences, we inferred that the sensor might not have more than one potential binding area and the base around the A-C mismatch might be strictly complementary pairing (not pairing like CAU-AUC).
Figure 28.IntaRNA Simulation of Binding Energy of Chk1(s)-Chk1(s)
Both ADAR1 and ADAR2_MCP demonstrated editing ability in yeast, but the underlying mechanisms are complex. We observed some positive results regarding selective ability, though the outcomes were somewhat unexpected. More experiments are needed. We are also eager to find out more potential of this monitoring system.
3.Reference
Kaseniit, K. E. et al. Modular, programmable RNA sensing using ADAR editing in living cells. Nature Biotechnology 41, 482-487 (2023).
Savva, Y. A., Rieder, L. E. & Reenan, R. A. The ADAR protein family. Genome Biology 13, 252 (2012).
Sun, J. et al. Cloning and characterization of a panel of constitutive promoters for applications in pathway engineering in Saccharomyces cerevisiae. Biotechnology and bioengineering 109, 2082-2092 (2012).
Martin Mann, Patrick R. Wright, and Rolf Backofen IntaRNA 2.0: enhanced and customizable prediction of RNA–RNA interactions Nucleic Acids Research, 2017, 45 (W1), W435–W439.
Patrick R. Wright, Jens Georg, Martin Mann, Dragos A. Sorescu, Andreas S. Richter, Steffen Lott, Robert Kleinkauf, Wolfgang R. Hess, and Rolf Backofen CopraRNA and IntaRNA: predicting small RNA targets, networks and interaction domains Nucleic Acids Research, 2014, 42 (W1), W119-W123.
Anke Busch, Andreas S. Richter, and Rolf Backofen IntaRNA: efficient prediction of bacterial sRNA targets incorporating target site accessibility and seed regions Bioinformatics, 2008, 24 (24), 2849-56.
Martin Raden, Syed M Ali, Omer S Alkhnbashi, Anke Busch, Fabrizio Costa, Jason A Davis, Florian Eggenhofer, Rick Gelhausen, Jens Georg, Steffen Heyne, Michael Hiller, Kousik Kundu, Robert Kleinkauf, Steffen C Lott, Mostafa M Mohamed, Alexander Mattheis, Milad Miladi, Andreas S Richter, Sebastian Will, Joachim Wolff, Patrick R Wright, and Rolf Backofen Freiburg RNA tools: a central online resource for RNA-focused research and teaching Nucleic Acids Research, 46(W1), W25-W29, 2018.
4.Sequence and Features
- 10INCOMPATIBLE WITH RFC[10]Illegal XbaI site found at 190
Illegal XbaI site found at 538
Illegal XbaI site found at 1429
Illegal PstI site found at 1455 - 12INCOMPATIBLE WITH RFC[12]Illegal PstI site found at 1455
- 21INCOMPATIBLE WITH RFC[21]Illegal BglII site found at 193
Illegal BglII site found at 289
Illegal BglII site found at 757
Illegal BglII site found at 823
Illegal BglII site found at 1036
Illegal BglII site found at 1501 - 23INCOMPATIBLE WITH RFC[23]Illegal XbaI site found at 190
Illegal XbaI site found at 538
Illegal XbaI site found at 1429
Illegal PstI site found at 1455 - 25INCOMPATIBLE WITH RFC[25]Illegal XbaI site found at 190
Illegal XbaI site found at 538
Illegal XbaI site found at 1429
Illegal PstI site found at 1455 - 1000COMPATIBLE WITH RFC[1000]
| None |