Team:NTU-Singapore/RNA Editing

Mutation-internal

Strategy to generate a dCasRx-InternalADAR2DD protein with high specificity and efficiency

Off-target activity of REPAIR can be largely attributed to the hyperactivating E488Q mutation in ADAR2DD. ADAR2DD is attached to the C-terminus of dPspCas13b by a flexible linker, which allows it to freely bind any dsRNA that is not targeted for RNA editing by gRNA-directed dPspCas13b. Hence, we hypothesized that ADAR2DD off-target RNA editing activity can be reduced by limiting its activity only to the double-stranded gRNA-target duplex. Guided by ADAR2 crystal structure data, we identified 12 positively-charged residues located on the ADAR2 RNA-binding loop that contributes to dsRNA binding affinity. We performed site-directed mutagenesis via primer design to mutate these residues into negative, neutral or hydrophobic residues. This would reduce ADAR2DD binding to non-targeted dsRNA and transfer all targeting control to the gRNA, thereby increasing specificity. We fused our 12 ADAR2DD mutants to dCasRx and characterised their RNA editing efficiency and specificity by amplicon sequencing. Previous preliminary data generated by our advisors at GIS tentatively show that the ADAR2DD mutants V351L, A454C, F457Y, H640D and Q479L+K475D have the highest RNA editing activity out of the 12 mutants. We characterized the 5 previously identified ADAR2DD mutants by performing luciferase assay and Sanger sequencing. To confirm their data, we performed amplicon sequencing and generated Figure 8. Figure 8 was generated by obtaining the average editing rate of the 12 ADAR2DD mutants for all 4 on-target genes (GAPDH, KRAS, PPIB, RAB7A) and 3 off-target genes (F11R, APOOL, XIAP). Values were all normalised to CasRx V1. The values were plotted with on-target editing on the x axis and off-target editing on the y axis.

To increase fidelity of ADAR2DD-mediated RNA editing, we also identified internal sites within dCasRx loop regions in which to insert our previously identified ADAR2DD mutants into. These sites were guided by CasRx crystal structure information and selected on the basis that they should minimally impact dCasRx structure and function when disrupted by ADAR2DD insertion. To quickly identify the optimal distance for ADAR2DD-mediated RNA editing, we inserted ADAR2DD into 7 dCasRx internal sites at varying distances from the double-stranded gRNA-target duplex, and characterised their RNA editing efficiency and specificity by luciferase reporter assay. Based on our results, ADAR2DD insertion into dCasRx internal site D338 demonstrates the highest RNA editing efficiency and specificity out of the 7 internal sites.

From our previous experiments, we have established that internal site D338 in dCasRx is an optimal site for ADAR2DD insertion. Following that, we inserted our 5 ADAR2DD mutants into internal site D338 in dCasRx and assessed RNA editing efficiency and specificity by luciferase reporter assay and Sanger sequencing.

Methodology

Results and discussion

7 Internal Sites - Luciferase assay

Figure 1. On-target RNA editing of ADAR2DD inserted into different dCasRx internal sites. The higher the relative luminescence unit (RLU), the higher the RNA editing activity.

Figure 2. Off-target RNA editing of ADAR2DD inserted into different dCasRx internal sites. The higher the relative luminescence unit (RLU), the lower the RNA editing specificity.

Based on our luciferase reporter assay, we identified internal sites D338, G655 and E689 as optimal sites for ADAR2DD insertion as they have relatively high on-target RNA editing. Out of the 3 shortlisted internal sites, D338 appears to be the best site for ADAR2DD insertion as it has the highest and lowest on-target and off-target activity respectively.

5 ADAR2DD mutant insertion into D338 - Luciferase assay

Figure 3. On-target RNA editing of 5 previously identified ADAR2DD mutants inserted into internal site D338.

Figure 4. Off-target RNA editing of 5 previously identified ADAR2DD mutants inserted into internal site D338.

After shortlisting D338, we inserted our 5 previously identified ADAR2DD mutants (V351L, A454C, F457Y, H640D, Q479L+K475D) into internal site D338. Out of the 5 previously identified ADAR2DD mutants, H460D appears to have the highest RNA editing activity. However, its editing activity is similar to that of ADAR2DD inserted into D338. Insertion of other ADAR2DD mutants also led to an overall decrease in RNA editing activity across mutants.

5 ADAR2DD mutant insertion into D338 - On-target editing

Figure 5. Sanger sequencing of on-target RNA editing of 5 previously identified ADAR2DD mutants inserted into dCasRx internal sites D338.

Next, we assayed the on-target endogenous RNA editing activity of our 5 previously identified ADAR2DD mutants inserted into internal site D338. As per data obtained from the luciferase reporter assay, H460D appears to have the highest editing activity. However, its editing activity is also similar to that of ADAR2DD inserted into D338. Based on these results, it appears that variation of mutant ADAR2DD insertion does not seem to significantly alter RNA editing activity. However, more genes should be tested to confirm whether this is true. This set of mutants were not used for Amplicon sequencing due to time constraints.

12 ADAR2DD - Amplicon Sequencing for on-target

Figure 6. Amplicon sequencing reads for on-target genes (GAPDH, KRAS, PPIB, RAB7A) of 12 dCasRx-ADAR2DD mutants.

12 ADAR2DD - Amplicon Sequencing for off-target

Figure 7. Amplicon sequencing reads for off-target genes (F11R, APOOL, XIAP) of 12 dCasRx-ADAR2DD mutants. Greyed out boxes indicate no reads at that region.

Figure 8. Scatter plot of average editing rate of the 12 ADAR2DD mutants from amplicon sequencing data. Average editing rate calculated for all 4 on-target genes (GAPDH, KRAS, PPIB, RAB7A) and 3 off-target genes (F11R, APOOL, XIAP).

Data generated from amplicon sequencing was used to calculate editing rates for the 12 ADAR2DD mutants for 4 on-target genes (GAPDH, KRAS, PPIB, RAB7A) and 3 off-target genes (F11R, APOOL, XIAP). The average editing rates were calculated and plotted. ADAR2DD mutants that occur at the bottom right of the plot are ideal as they have high on-target and low off-target editing activity. Based on the plot, ADAR2DD inserted into D338 (In-D338) appears to be an improvement over CasRx v1 as it has a lower off-target editing rate while still retaining its relatively high on-target editing rate. This means In-D338 is more specific while being almost equally efficient as CasRx v1.