Fig 2: Properties of modified Cas13 variants. a Schematic of nuclease-dead CasFX (dCasFX) activity. dCasFX carries quadrupl e point mutations that abolish its nuclease activity. As a result, the dCasFX/crRNA complex can be recruited and bind to target transcripts, but it cannot cleave the RNA. b Evaluation of Cas13 cleavage efficiency of dCasFX compared to wild-type CasFX. qPCR data represent expression levels of eCFP. Data were normalized to samples treated with blank crRNA (control). * = p value < 0.05, ** = p value < 0.01, *** = p value < 0.001, p values based on Student’s t test, error bars represent 95% confidence intervals. c eCFP fluorescence when targeted by either CasFX or dCasFX. Nuclei were stained with nuclear green DCS1 (Abcam ab138904). Color was adjusted for color-blind-friendly purpose. eCFP and DsRed fluorescence were measured using their native fluorescence property without using antibody staining. Scale bar = 50 μm. d Schematic of dCasFX for the validation of RNA-protein interactions. dCasFX and crRNA targeting Fer1HCH-RA mRNA were transfected together in one sample. Fer1HCH-RA and IRP1AC450S, a constitutively RNA-binding form of IRP1A that interacts with the iron-responsive element (IRE) in the Fer1HCH-RA mRNA, were transformed together in a different sample. The two samples were each lysed and combined, followed by immunoprecipitation (IP) of dCasFX (utilizing the attached HA tag) to test for the presence of IRP1A in the pull-down assay. e Western blot showing the IP of dCasFX combined with different crRNAs along Fer1HCH-RA mRNA and the detection of IRP1A in corresponding samples. f Functional schematic of CasFX that carries a mitochondrial localization signal (CasFXmt). At the N terminus, CasFXmt is fused with the tim23 mitochondrial signal sequence. Upon binding with crRNA, the complex will localize into mitochondria and target mitochondrial-encoded transcripts. g Mitochondrial localization of CasFXmt. Nuclei were stained with DAPI (blue) while mitochondria were stained with mitotracker green (Cell signaling 9074S) and CasFX polypeptide was stained with anti-HA antibody (magenta). Scale bar = 25 μm. Color was adjusted for color-blind-friendly purpose. h The relative expression level of mitochondrial-encoded transcripts, COXI and COXII, targeted by RNAi, CasFXO, and CasFXmt. Data were normalized to samples treated with no transfected plasmid (control). * = p value < 0.05, ** = p value < 0.01, *** = p value < 0.001, ns = not significant, p values based on Dunnett’s post hoc test, error bars represent 95% confidence intervals. i Western blotting of COXI and COXII when being targeted by RNAi, CasFXO, and CasFXmt

Fig 3: Adaptation of the REPAIRv2 system to modify RNA in Drosophila Sg4 cell culture. a Schematic for the Drosophila-modified REPAIRv2 system (FREPAIRv2), to modify a mutant eCFP transcript. Mutant eCFP carries an early stop codon that normally encodes Tryptophan at residue 57 (W57*). By generating an A to C mismatch in the crRNA spacer that corresponds to the stop codon, the ADAR2DD domain will change the equivalent adenosine (A) to inosine (I). Inosine will be treated as guanosine by the translation machinery. b Schematic of FREPAIRv2 outcome. Originally, the mutant eCFP transcript harbors a stop codon at position 57, which will generate a short polypeptide with 56 amino acids. However, once modified by FREPAIRv2, codon 57 will be reverted to wild-type tryptophan and restore the production of a full-length polypeptide. c Western blotting monitoring eCFP productions relative to transfection time. d Fluorescence emitted by eCFP relative to transfection time. Nuclei were stained with nuclear green DCS1 (Abcam ab138905). Color was adjusted for color-blind-friendly purpose. eCFP and DsRed fluorescence were measured based on their natively emitted fluorescence. Scale bar = 50 μm. e Schematic of crRNAs that we used for FREPAIRv2. We considered two criteria for the crRNA design: (i) mismatch distance from the first nucleotide and (ii) spacer length. f Editing rate and off-target rate of FREPAIRv2 concerning mismatch distance when spacer length was kept at a constant 50 nucleotides. Error bars represent standard deviation. g Editing rate and off-target rates of FREPAIRv2 in relation to spacer lengths when the mismatch distance was kept at the constant position 26. Error bars represent standard deviation

Fig 4: Properties of modified Cas13 variants. a Schematic of nuclease-dead CasFX (dCasFX) activity. dCasFX carries quadrupl e point mutations that abolish its nuclease activity. As a result, the dCasFX/crRNA complex can be recruited and bind to target transcripts, but it cannot cleave the RNA. b Evaluation of Cas13 cleavage efficiency of dCasFX compared to wild-type CasFX. qPCR data represent expression levels of eCFP. Data were normalized to samples treated with blank crRNA (control). * = p value < 0.05, ** = p value < 0.01, *** = p value < 0.001, p values based on Student’s t test, error bars represent 95% confidence intervals. c eCFP fluorescence when targeted by either CasFX or dCasFX. Nuclei were stained with nuclear green DCS1 (Abcam ab138904). Color was adjusted for color-blind-friendly purpose. eCFP and DsRed fluorescence were measured using their native fluorescence property without using antibody staining. Scale bar = 50 μm. d Schematic of dCasFX for the validation of RNA-protein interactions. dCasFX and crRNA targeting Fer1HCH-RA mRNA were transfected together in one sample. Fer1HCH-RA and IRP1AC450S, a constitutively RNA-binding form of IRP1A that interacts with the iron-responsive element (IRE) in the Fer1HCH-RA mRNA, were transformed together in a different sample. The two samples were each lysed and combined, followed by immunoprecipitation (IP) of dCasFX (utilizing the attached HA tag) to test for the presence of IRP1A in the pull-down assay. e Western blot showing the IP of dCasFX combined with different crRNAs along Fer1HCH-RA mRNA and the detection of IRP1A in corresponding samples. f Functional schematic of CasFX that carries a mitochondrial localization signal (CasFXmt). At the N terminus, CasFXmt is fused with the tim23 mitochondrial signal sequence. Upon binding with crRNA, the complex will localize into mitochondria and target mitochondrial-encoded transcripts. g Mitochondrial localization of CasFXmt. Nuclei were stained with DAPI (blue) while mitochondria were stained with mitotracker green (Cell signaling 9074S) and CasFX polypeptide was stained with anti-HA antibody (magenta). Scale bar = 25 μm. Color was adjusted for color-blind-friendly purpose. h The relative expression level of mitochondrial-encoded transcripts, COXI and COXII, targeted by RNAi, CasFXO, and CasFXmt. Data were normalized to samples treated with no transfected plasmid (control). * = p value < 0.05, ** = p value < 0.01, *** = p value < 0.001, ns = not significant, p values based on Dunnett’s post hoc test, error bars represent 95% confidence intervals. i Western blotting of COXI and COXII when being targeted by RNAi, CasFXO, and CasFXmt

Fig 5: Properties of modified Cas13 variants. a Schematic of nuclease-dead CasFX (dCasFX) activity. dCasFX carries quadrupl e point mutations that abolish its nuclease activity. As a result, the dCasFX/crRNA complex can be recruited and bind to target transcripts, but it cannot cleave the RNA. b Evaluation of Cas13 cleavage efficiency of dCasFX compared to wild-type CasFX. qPCR data represent expression levels of eCFP. Data were normalized to samples treated with blank crRNA (control). * = p value < 0.05, ** = p value < 0.01, *** = p value < 0.001, p values based on Student’s t test, error bars represent 95% confidence intervals. c eCFP fluorescence when targeted by either CasFX or dCasFX. Nuclei were stained with nuclear green DCS1 (Abcam ab138904). Color was adjusted for color-blind-friendly purpose. eCFP and DsRed fluorescence were measured using their native fluorescence property without using antibody staining. Scale bar = 50 μm. d Schematic of dCasFX for the validation of RNA-protein interactions. dCasFX and crRNA targeting Fer1HCH-RA mRNA were transfected together in one sample. Fer1HCH-RA and IRP1AC450S, a constitutively RNA-binding form of IRP1A that interacts with the iron-responsive element (IRE) in the Fer1HCH-RA mRNA, were transformed together in a different sample. The two samples were each lysed and combined, followed by immunoprecipitation (IP) of dCasFX (utilizing the attached HA tag) to test for the presence of IRP1A in the pull-down assay. e Western blot showing the IP of dCasFX combined with different crRNAs along Fer1HCH-RA mRNA and the detection of IRP1A in corresponding samples. f Functional schematic of CasFX that carries a mitochondrial localization signal (CasFXmt). At the N terminus, CasFXmt is fused with the tim23 mitochondrial signal sequence. Upon binding with crRNA, the complex will localize into mitochondria and target mitochondrial-encoded transcripts. g Mitochondrial localization of CasFXmt. Nuclei were stained with DAPI (blue) while mitochondria were stained with mitotracker green (Cell signaling 9074S) and CasFX polypeptide was stained with anti-HA antibody (magenta). Scale bar = 25 μm. Color was adjusted for color-blind-friendly purpose. h The relative expression level of mitochondrial-encoded transcripts, COXI and COXII, targeted by RNAi, CasFXO, and CasFXmt. Data were normalized to samples treated with no transfected plasmid (control). * = p value < 0.05, ** = p value < 0.01, *** = p value < 0.001, ns = not significant, p values based on Dunnett’s post hoc test, error bars represent 95% confidence intervals. i Western blotting of COXI and COXII when being targeted by RNAi, CasFXO, and CasFXmt