RNA editing restricts hyperactive ciliary kinases
RNA editing restricts ciliary kinases . Ciliary kinases are essential for cilia formation and function but it remains unknown how their activities are regulated in vivo. Li et al. created roundworm animal models carrying hyperactive ciliary kinases that disrupt cilia. Their genetic suppressor screens revealed that loss of an RNA adenosine deaminase that catalyzes adenosine-to-inosine (A-to-I) RNA editing rescued ciliary abnormalities. They found that kinase hyperactivation caused this RNA adenosine deaminase to edit kinase RNA and impair kinase RNA splicing and translation, thereby downregulating ciliary kinases from nuclei. These results suggest that ciliopathies may be treated by targeting the pathways outside of cilia. Science , abd8971, this issue p. 984 Abstract . Protein kinase activity must be precisely regulated, but how a cell governs hyperactive kinases remains unclear. In this study, we generated a constitutively active mitogen-activated protein kinase DYF-5 (DYF-5CA) in Caenorhabditis elegans that disrupted sensory cilia. Genetic suppressor screens identified that mutations of ADR-2, an RNA adenosine deaminase, rescued ciliary phenotypes of dyf-5CA . We found that dyf-5CA animals abnormally transcribed antisense RNAs that pair with dyf-5CA messenger RNA (mRNA) to form double-stranded RNA, recruiting ADR-2 to edit the region ectopically. RNA editing impaired dyf-5CA mRNA splicing, and the resultant intron retentions blocked DYF-5CA protein translation and activated nonsense-mediated dyf-5CA mRNA decay. The kinase RNA editing requires kinase hyperactivity. The similar RNA editing–dependent feedback regulation restricted the other ciliary kinases NEKL-4/NEK10 and DYF-18/CCRK, which suggests a widespread mechanism that underlies kinase regulation. Protein kinases mediate signal transduction and regulate fundamental cellular processes in eukaryotic cells. The catalytic activity of kinases must be stringently controlled, and its dysregulation leads to developmental defects, metabolic disorders, and cancer ( 1 , 2 ). Although loss-of-function kinase mutations have been extensively studied, gain-of-function mutations are less understood. Kinase hyperactivities have been described in many disease conditions, such as the constitutively active BCR-ABL1 tyrosine kinase in chronic myelogenous leukemia and BRAF serine/threonine kinase in malignant melanomas ( 1 ). Potent kinase inhibitors treat cancer effectively; however, tumors inevitably develop drug resistance and circumvent kinase inhibition by activating alternative pathways that perpetuate cell proliferation ( 3 ). Therefore, it is crucial to understand how an organism regulates hyperactive kinases. We studied male germ cell–associated kinase (Mak), an evolutionarily conserved mitogen-activated protein kinase (MAPK). Mak localizes at the ciliary tip to restrict cilia elongation ( 2 , 4 , 5 ). Cilia are microtubule-based organelles essential for cell motility and sensory signaling, and ciliary defects cause more than 35 ciliopathies ( 6 ). The loss of Mak orthologs results in abnormally long cilia and leads to retinitis pigmentosa ( 2 , 4 , 5 , 7 , 8 ). By contrast, overexpression of mammalian Mak or Caenorhabditis elegans genomic DNA (gDNA) encoding the Mak ortholog DYF-5 inhibits ciliogenesis ( 2 , 5 ), which indicates that an optimal Mak/DYF-5 activity is essential. The MAPK superfamily contains a canonical TxY motif in the activation T-loop, phosphorylation of the threonine (pThr) residue activates the kinase, and glutamate can be a phosphomimetic pThr ( 9 , 10 ). Results . A constitutively active DYF-5 kinase disrupted cilia . We first established that substitution of Thr 164 with glutamate (T164E) in the TxY motif of DYF-5 leads to constitutive kinase activation in vitro. DYF-5(T164E) phosphorylated an intraflagellar transport (IFT) protein IFT-74, whereas wild-type (WT) DYF-5 did not (fig. S1, A to D), demonstrating that the substitution produced a constitutively active DYF-5 kinase (hereafter DYF-5CA). We then generated a T164E dyf-5CA knock-in animal in C. elegans ( Fig. 1A and fig. S2, A and B). The dyf-5CA mutant ( n > 200) exhibited normal morphology, viability, and growth rates comparable with those of the WT N2 strain. However, 87% of dyf-5CA worms ( n = 150) could not use their sensory cilia to uptake fluorescence dye 1,1′-dioctadecyl-3,3,3′,3′-tetramethylindocarbocyanine perchlorate from the medium (Dyf) ( Fig. 1B and fig. S2C), indicating an impaired ciliary function and integrity ( 11 ). To visualize IFT and cilia, we introduced a green fluorescence protein (GFP)–tagged IFT52/OSM-6 into dyf-5CA . In contrast to the bidirectional motility of OSM-6::GFP in WT cilia ( 12 ), no IFT was detectable from the distal ciliary segments of dyf-5CA ( n > 20 animals), in which the OSM-6::GFP fluorescence formed aggregates ( Fig. 1, C and D , and fig. S3, A and C to E). Download high-res image . Open in new tab . Download Powerpoint . Fig. 1 ADR-2 mutations rescued ciliary defects in dyf-5CA . ( A ) Multisequence alignment of the C. elegans (Ce) DYF-5 kinase and its homologs in human (Hs), mouse (Mm), zebrafish (Dr), and Chlamydomonas (Cr). E, Glu; T, Thr. ( B ) Percentage of dye-filling positive animals in the indicated strains (mean ± SD) (left). N = 100 to 200. The cilium length of phasmid cilia (right, measured using the OSM-6::GFP fluorescence) is shown. N = 30 to 60. Statistical significance, compared with the control with a matching color code, is based on Student’s t test; n.s., not significant, P < 0.001. ( C ) Schematic of the amphid and phasmid cilia. Dashed boxes are enlarged on the right. d.s., distal ciliary segment; m.s., middle segment. ( D ) Cilia in WT and mutants. Arrowheads indicate the ciliary base and transition zone, and arrows indicate junctions between the middle and distal segments. Scale bar, 5 μm. ( E ) ADR-2 protein domains. Amino acid changes and the deletion in adr-2 mutant alleles are indicated. dsRBD, dsRNA binding domain. F, Phe; G, Gly; I, Ile; Q, Gln; S, Ser. Scale bar, 50 amino acids (aa). The dyf-5CA animals developed abnormally elongated cilia, resembling the long cilia defect in the dyf-5(mn400) null allele ( Fig. 1, B and D , and fig. S3B) but opposite to the truncated cilia phenotype when dyf-5 gDNA was overexpressed ( 2 ). None of the dyf-5CA heterozygotes showed ciliary defects ( n > 50 examined heterozygotes), and expression of the dyf-5 cDNA in ciliated neurons fully rescued ciliary phenotypes in dyf-5CA ( Fig. 1, B and D ), indicating that dyf-5CA is a recessive loss-of-function mutant. Expression of dyf-5CA cDNA did not disrupt cilia in WT but rescued ciliary defects in dyf-5CA mutants ( Fig. 1B and fig. S2D). Thus, the DYF-5CA protein did not have a dominant-negative effect. The discrepancy between the in vitro kinase hyperactivity and the in vivo kinase-null phenotype suggests that a previously unrecognized mechanism may control DYF-5CA in a living organism. RNA adenosine deaminase mutations rescued ciliary defects by DYF-5CA kinase . We performed a forward genetic suppressor screen for mutations that restore dyf-5CA cilia (fig. S4A) and isolated four suppressors, cas506 , cas517 , cas518 , and cas519 , which exhibited uptake of fluorescence dye like the WT ( n = 100) (fig. S4, B and C). All four mutations were mapped to the deaminase domain of ADR-2, with cas517 carrying a premature stop codon and the other three mutants carrying missense mutations ( Fig. 1E and fig. S4, B and D). A deletion mutant of adr-2(ok735) that removes the entire double-stranded RNA (dsRNA) binding domain and most of the deaminase domain ( Fig. 1E and fig. S4E) showed the same ciliary rescue effects on dyf-5CA mutant cilia ( Fig. 1, B and D , and fig. S3B). ADR-2 is an ortholog of human ADARB1 (adenosine deaminase acting on RNA B1, or ADAR2). It exhibits dsRNA adenosine deaminase activity, catalyzing the adenosine-to-inosine (A-to-I) RNA editing, which is interpreted by the ribosome as guanosine ( 13 , 14 ). The C. elegans genome encodes two ADARs: ADR-1 is catalytically inactive, and ADR-2 is the only active RNA deaminase. The adr-2 deletion completely abolishes A-to-I editing ( 15 – 17 ). Although ADR-1 regulates the activity of ADR-2 ( 15 ), the adr-1 deletion did not recover dyf-5CA cilia (fig. S4C). By contrast, the adr-2 deletion reduced the ciliary aggregates of OSM-6::GFP and improved IFT in both amphid and phasmid cilia of dyf-5CA animals, albeit at a slower speed ( Fig. 1D and fig. S3, A and C to E). None of the adr-2 mutant alleles displayed ciliary defects ( n > 100, for each genotype) ( Fig. 1B and fig. S4C). We used a ciliated neuron-specific promoter, P dyf-1 , to express adr-2 cDNA in the dyf-5CA; adr-2 double mutant, which caused 75% of animals to develop the same ciliary defects that were observed in dyf-5CA animals ( n = 300 transgenic animals) (fig. S4C), indicating that ADR-2 functions cell autonomously. DYF-5CA caused aberrant RNA editing of the dyf-5CA pre-mRNA . We sought a mechanistic understanding between ADR-2 and DYF-5CA. An ADR-2::GFP reporter confirmed its nuclear localization ( 13 ) but did not reveal the ciliary localization of ADR-2 (fig. S5, A and B). Using GFP-affinity purification and mass spectrometry, we identified a known ADR-2 binding protein, ADBP-1, required for its deaminase activity (fig. S5C and data file S1) ( 18 ). The adbp-1 ( qj1 ) mutant recovered dyf-5CA ’s ciliary defects ( Fig. 1, B and D , and fig. S3, A to C), and GFP::ADBP-1 localized within nuclei, but not cilia (fig. S5D), suggesting that ADR-2 and ADBP-1 do not directly function in cilia but regulate DYF-5CA by their nuclear actions. Because ADR-2 and ADBP-1 are involved in small interfering RNA– and microRNA-associated pathways ( 18 – 20 ), we introduced RNA interference (RNAi) mutants into dyf-5CA but did not observe rescue effects (fig. S5E), indicating that the function of adr-2 or adbp-1 is independent of the RNAi machinery. To understand ADR-2–based RNA editing in dyf-5CA , we implemented a comparative RNA-sequencing (RNA-seq) screen on WT, dyf-5CA , adr-2 , and dyf-5CA; adr-2 animals by using an established computational pipeline ( 21 ). WT animals had a total of 18,097 A-to-I editing sites, but RNA editing events were almost undetectable in adr-2 or adbp-1 ( Fig. 2A , fig. S6A, and data file S2), which confirmed previous findings ( 15 – 18 ). In dyf-5CA animals, 12,871 sites that were edited in WT became unedited ( Fig. 2A and data file S2). Because removing all the editing sites in the dyf-5CA;adr-2 double mutant rescued dyf-5CA ’s ciliary defects, we argue that these editing sites lost in dyf-5CA are not responsible for the phenotypes, although it is unclear why and how these sites were lost. The dyf-5CA animals unexpectedly acquired 1537 RNA editing sites, which were not detected in the WT and were lost in the dyf-5CA; adr-2 ( Fig. 2A and data file S2). Most of the identified RNA editing sites are located on chromosomes as clusters ( 17 ). We searched for the groups of ectopic editing sites and found that clustered sites were present in seven genes (fig. S6B and data file S2), including the dyf-5CA gene itself ( Fig. 2B ). The 269 editing sites were distributed from the sixth exon to the seventh intron of dyf-5CA and occurred at frequencies of 10 to ~60% ( Fig. 2C ). By contrast, we did not detect any RNA editing sites in the dyf-5 locus in other strains ( Fig. 2B ). Except for dyf-5 , none of the other six genes were known to have ciliary functions. Download high-res image . Open in new tab . Download Powerpoint . Fig. 2 ADR-2–dependent RNA editing clusters in dyf-5CA . ( A ) Heatmap and dendrogram of hierarchical clustering of editing sites using editing ratios among WT and mutants in different genomic bins (1-Mb tile) from RNA-seq data. Relative editing levels were used in hierarchical clustering. Each row corresponds to a sample. Each column represents the relative editing ratio in a genomic bin. Experiments were repeated three times with the similar results. The red box highlights the bin with the abnormally gained RNA editing sites of the dyf-5 locus in dyf-5CA . ( B ) Representative RNA editing analysis at the dyf-5 locus. Each row corresponds to a sample. Red lines represent RNA editing sites in the genomic locations of dyf-5 , and the bar for editing level is shown on the right. Gene bodies are above; blue boxes show exons encoding the kinase domain. A, Ala; D, Asp. ( C ) Ectopic gain of RNA editing sites at the dyf-5 locus in dyf-5CA . ( D ) Percentage of dye-filling (mean ± SD) positive animals in the indicated strains, n = 100 to 200 (left). Cilium length, n = 30 to 60 (right), is shown. We next investigated whether and to what extent the aberrant RNA editing clusters in dyf-5CA affect cilia. The sixth and seventh introns contain 130 editing sites, and the deletion of either intron or both in dyf-5CA animals rescued ciliary defects ( Fig. 2D and figs. S2A and S7, A and B). The double intron deletion removed all the ectopic editing clusters in dyf-5 exons ( Fig. 2B ). To rule out the effects of intron deletion on pre-mRNA, we synthesized the sixth and seventh introns by placing a noneditable T at all detected A-to-I editing sites. Replacing introns with noneditable sites rescued the ciliary phenotypes in dyf-5CA ( Fig. 2D and figs. S2A and S7, B and C). Although other aberrantly gained or lost RNA editing sites may have biological functions under certain conditions, our data show that RNA editing sites in dyf-5CA are the primary cause for ciliary defects. dyf-5CA animals generated antisense transcripts in dyf-5 . Why does RNA editing occur specifically in this region? Because substrates of RNA editing enzymes are dsRNA ( 13 ), we examined whether long dsRNAs are formed intramolecularly at this locus but failed to identify any, which is consistent with lacking RNA editing at this locus in the WT. We then investigated whether an antisense RNA was generated to form dsRNA with the dyf-5CA transcript subject to ADR-2 editing. Our strand-specific RNA-seq detected abundant antisense transcripts in dyf-5CA animals, starting from the seventh intron to the fourth exon of dyf-5CA and aligning with the region where aberrant RNA editing occurred ( Fig. 3, A and B ). Next, we examined whether the eighth intron of dyf-5 , which is immediately upstream of the antisense transcripts, regulates antisense transcription. Although dyf-5 cDNA expression did not cause ciliary abnormalities ( Fig. 3C and figs. S2D and S8A), expression of dyf-5 cDNA containing the sixth, seventh, and eighth introns, but not any individual intron, disrupted cilia ( Fig. 3C ). Conversely, dyf-5 gDNA expression impaired cilia in an ADR-2–dependent manner ( Fig. 3C and fig. S8A); however, expression of dyf-5 gDNA that lacked the sixth, seventh, or eighth intron reduced the penetrance of ciliary defects ( Fig. 3C ). Consistently, dyf-5 gDNA expression produced ectopic RNA editing clusters and antisense RNA transcription of dyf-5 , which were not limited to the regions in dyf-5CA animals but spread to additional introns and exons of dyf-5 ( Figs. 2B and 3A ), generating more profound effects than dyf-5CA . Deleting the eighth intron reduced the ciliary defects of dyf-5CA animals ( Fig. 3C and fig. S7B), which demonstrates its contribution to ciliary defects. The incomplete penetrance of the eighth intron deletion suggests that other genomic region(s) might be involved. This intron may serve as a cryptic promoter for antisense transcription, and future studies will need to identify the additional cis-regulatory elements and the corresponding trans factor(s). Download high-res image . Open in new tab . Download Powerpoint . Fig. 3 Antisense RNA generation in dyf-5CA . ( A and B ) RNA expression profiles at the dyf-5 locus in different strains. Sense RNA (top) is in blue, and antisense RNA (bottom) is in pink. (B) Percentage of the normalized antisense/sense ratio in the indicated strains. Mean ± SEM. Experiments in (A) and (B) were performed three times with similar results. RNA-seq signals are shown in the labeled range. ( C ) Percentage of Dyf phenotypes in the indicated strains. n = 100 to 200, mean ± SD. ( D ) Distributions of normalized stranded RNA-seq reads for a portion of the dyf-5 gene in (A) show IRs. Gray boxes show the location of retained introns. RNA-seq signals are shown in the labeled range (left). Data are from three independent experiments and show mean ± SEM (right). Impaired dyf-5CA mRNA splicing blocked DYF-5CA kinase translation . Whereas antisense transcripts resulted in promiscuous RNA editing in dyf-5CA , it is the RNA editing, but not the antisense transcripts, that caused ciliary defects because dyf-5CA;adr-2 double mutants that retained antisense transcripts but lacked RNA editing restored their cilia ( Fig. 3, A and B ). Thus, we studied the molecular consequence that RNA editing conferred in dyf-5CA . The hyperediting of the sixth and the seventh exons changed 40.5% of the amino acids in this region. The sixth exon is located in the kinase domain, implying a possible down-regulation of kinase activity (fig. S7D). The sixth and the seventh introns in the dyf-5CA pre-mRNA were not sufficiently spliced out, causing intron retention (IR) ( Fig. 3, A and D ). Using IRFinder ( 22 ), we found that the sixth and seventh IR ratios are 5.7 and 34.6% in dyf-5CA , respectively ( Fig. 3D ). By contrast, IR ratios were reduced to 0.6 and 16% in the dyf-5CA;adr-2 double mutant, indicating that RNA editing of the intron sequence may promote IR. The incomplete elimination of IR in dyf-5CA;adr-2 double mutants suggests that other mechanisms may be involved. For example, RNA editing is known to occur cotranscriptionally in nascent RNAs ( 23 ), and the ectopic antisense RNA transcripts can form dsRNA to prevent the spliceosome from accessing the dyf-5CA pre-mRNA. To examine kinase production, we fused GFP to DYF-5 at its N terminus in WT or dyf-5CA animals. DYF-5 localized in the WT or adr-2 mutant cilia (100%, n > 30 animals) ( Fig. 4A and fig. S8B); however, GFP fluorescence was barely detected from 89.6% of gfp::dyf-5CA animals ( Fig. 4A and fig. S8C). In gfp::dyf-5CA;adr-2 double mutants, the GFP::DYF-5CA signal increased, albeit at a lower level than the WT ( Fig. 4A ), indicating that a small amount of hyperactive kinase restored cilia. To further evaluate whether the loss of GFP::DYF-5CA resulted from translation blockade, we fused a red fluorescent protein gene (wrmScarlet, the Scarlet protein with worm codon optimization) to the C terminus of GFP::DYF-5 or GFP::DYF-5CA ( Fig. 4B ). The green and red fluorescence colocalized in all the gfp::dyf-5::wrmScarlet cilia ( n = 60 animals) ( Fig. 4C ); however, only the faint green but no red fluorescence was visible from all the examined gfp::dyf-5CA::wrmScarlet animals ( n = 43) ( Fig. 4C ), which implies premature translation termination. Furthermore, the GFP-tagged DYF-5 kinase domain (11 to 291 amino acids) was localized in the cell body and dendrite but not in cilia, whereas the C terminus of DYF-5 (292 to 489 amino acids) entered cilia (fig. S8, D to F), indicating that its C terminus targeted the kinase domain to cilia. Thus, the dyf-5CA mRNA confers substantial defects in DYF-5 protein synthesis and, at least, its C-terminal ciliary localization domain, which prevents the kinase from phosphorylating its ciliary substrates. Download high-res image . Open in new tab . Download Powerpoint . Fig. 4 DYF-5CA protein production and a working model. ( A ) Localization of GFP::DYF-5 or GFP::DYF-5CA with different genotypes. OSM-6::mCherry marks cilia. Percentages of DYF-5 localization patterns are indicated. ( B ) Schematic of gfp::dyf-5::wrmScarlet knock-in allele. Scale bar, 50 amino acids. ( C ) Localization of GFP::DYF-5::wrmScarlet (left) or GFP::DYF-5CA::wrmScarlet (right) in phasmid cilia. The blue fluorescence protein (BFP)–tagged IFT-81 labels cilia. Scale bars, (A) and (C), 2.5 μm. ( D ) Ciliary defects in the nekl-4 gDNA overexpressed animals. Scale bar, 5 μm. ( E ) Percentage of dye filling (mean ± SD)–positive animals, n = 100 to 200 (left). Cilium length, n = 20 to 40 (right). ( F ) Mapping profiles of sense and antisense RNA at nekl-4 in the WT nekl-4 (upper) or kinase-dead nekl-4(D609A) (lower) gDNA overexpression animals (sense reads, blue; antisense reads, pink). ( G ) RNA editing clusters at nekl-4 . The bar for editing level is shown on the top. ( H ) Model summarizes how RNA editing restricts hyperactive ciliary kinases. Nonsense-mediated dyf-5CA mRNA decay . We examined the fate of dyf-5CA mRNAs. The A-to-I RNA editing did not create a stop codon; however, the intron-retaining transcripts often contained a premature termination codon (PTC) (fig. S7E) and might be targeted for degradation by the nonsense-mediated mRNA decay (NMD) pathway ( 24 ). Therefore, we crossed dyf-5CA animals with NMD mutants, including smg-1 , a critical kinase that mediates the phosphorylation of an RNA helicase UPF1, and smg-3 , an ortholog of human UPF2 ( 25 , 26 ). The dyf-5CA transcripts in the dyf-5CA; smg-1 double mutants became ~1.8-fold more abundant than in dyf-5CA , which was confirmed by using quantitative polymerase chain reaction (PCR) experiments ( Fig. 3A and fig. S9A). smg-1 or smg-3 mutation, which did not affect WT cilia ( n > 100), partially rescued the ciliary dye-filling defects of dyf-5CA (fig. S9B) and reduced IFT-particle aggregation (fig. S9, C to E), although smg-1 mutation did not shorten the elongated cilia (fig. S9F). Blocking of NMD rescued the phenotype in Ullrich’s disease by up-regulating the mRNA level of the mutant collagen subunit and forming a partially functional extracellular matrix ( 27 ). Likewise, the inhibition of NMD increases the dyf-5CA mRNA level, a small portion of which may translate into functional kinase, partially restoring dyf-5CA cilia. RNA editing restricted two other ciliary kinases . To explore whether RNA editing restricts other kinases, we screened five kinases by overexpressing the kinase gDNA in WT. Two ciliary kinases, NEKL-4/NEK10 (never in mitosis kinase like) and DYF-18/CCRK (ortholog of human cyclin-dependent kinase 20) ( 28 – 30 ), when overexpressed, resulted in ciliary defects, aberrant antisense transcripts, and RNA editing on kinase pre-mRNA ( Fig. 4, D to G , and fig. S10, A to D). adr-2 deletion rescued the ciliary phenotypes in kinase overexpression (OE) animals ( Fig. 4, D and E , and fig. S10, A and B), indicating that ciliary defects require RNA editing. However, we did not detect abnormalities in animal phenotypes, antisense transcripts, or RNA editing of the rest of the three cytoplasmic kinase OE strains, including the POLO kinase PLK-1, the AKT kinase family kinase AKT-1, and the adenosine monophosphate–activated protein kinase AAK-2 (fig. S11), suggesting that RNA editing may specifically regulate ciliary kinases. Kinase hyperactivation caused aberrant RNA editing . We determined whether kinase hyperactivity or transgene overexpression caused ectopic RNA editing. Expression of a catalytic dead D609A mutation of NEKL-4 ( 31 ) did not disrupt cilia ( Fig. 4, D and E ) and did not generate aberrant antisense transcription or RNA editing ( Fig. 4, F and G ). Similarly, the catalytically dead DYF-18(D145A) OE did not show ciliary phenotypes (fig. S10, A and B). We detected antisense transcription and RNA editing at levels similar to those of DYF-18 OE animals (fig. S10, C and D). We speculate that DYF-18(D145A) overexpression might have a residual activity in vivo, triggering RNA editing, but could not impair cilia. Consistently, the same amount of gDNA transformation disrupted cilia in 82% of NEKL-4OE but in only 23% of DYF-18OE animals ( Fig. 4, D and E , and fig. S10, A and B), which suggests that the ciliary structure might be more sensitive to up-regulated NEKL-4. Furthermore, the catalytically dead DYF-5(D150A) or DYF-5(K40A) (K40A, Lys 40 →Ala) ( 32 ) OE reduced the ciliary phenotype penetrance (fig. S12A), accompanying a reduction of aberrant antisense transcripts and RNA editing (fig. S12, B and C). Thus, kinase activity is a substantial cause for ciliary defects and abnormal RNA editing. To overcome the limitation of using kinase overexpression as a proxy kinase hyperactivity, we introduced D150A in dyf-5CA animals (fig. S2A). DYF-5CA(D150A) protein was translated and correctly localized to cilia (fig. S12D), and dyf-5CA(D150A) animals did not generate any antisense transcript or RNA editing on dyf-5 ( Figs. 2B and 3, A and B ), which indicated that aberrant antisense expression and RNA editing resulted from DYF-5CA hyperactivity rather than overexpression. Because of catalytic null, dyf-5CA(D150A) animals developed ectopically long cilia like dyf-5 null (fig. S12D). These results support a direct contribution of kinase hyperactivity to aberrant antisense transcription and RNA editing. We suggest that transgene overexpression increased WT kinase levels, which up-regulates kinase activity, which in turn triggers antisense transcription and RNA editing of the native kinase locus. Considering that ADR-2 is known to suppress transgene silencing through editing dsRNA that is produced from aberrant overlapping antisense transcription ( 33 ), ciliary phenotypes from kinase gDNA overexpression may involve transgene silencing. We introduced the rde-1(ne219) and rde-4(ne301) mutant alleles defective in transgene silencing ( 34 ) into dyf-5 , nekl-4 , and dyf-18 gDNA OE animals, but their ciliary phenotypes were not changed (fig. S12E), suggesting that transgene silencing may not be involved. Transcription regulation of dyf-5CA mRNA . dyf-5CA animals have more RNA-seq reads in the first seven exons but fewer reads in the final four exons than that of WT ( Fig. 3A ). We compared the stability of the dyf-5 transcripts in WT and dyf-5CA animals after treatment with actinomycin D, which blocks transcription. dyf-5 pre-mRNAs were degraded at a similar rate in both strains (fig. S13A), which suggests that DYF-5CA may not affect transcript stability. dyf-5CA animals abnormally transcribe antisense RNA from the fourth exon to the seventh intron of dyf-5 ; the resultant dsRNA may inhibit RNA polymerase II (Pol II) elongation ( 35 ). To assess the potential transcriptional changes, we performed RNA Pol II chromatin immunoprecipitation followed by sequencing chromatin immunoprecipitation–sequencing (ChIP-seq). Compared with WT, dyf-5CA animals increased Pol II occupancy in the third to eighth exon of dyf-5 (fig. S13B). Similar increases were observed for the elongation-specific Pol II phosphorylated at serine 2 (fig. S13B). ChIP-seq of histone H3 trimethylation at Lys 36 (H3K36me3) revealed increased H3K36me3 modifications in this region due to Pol II elongation and cotranscriptional deposition of H3K36me3 (fig. S13B). The increased densities indicate aberrantly high transcript elongation in the region and suggest that the lack of transcripts in the final few exons may result from transcription regulation. The adr-2 deletion did not affect these densities in dyf-5CA (fig. S13B), which suggests that the increased occupancy may occur upstream or independent of RNA editing. By contrast, introducing D150A to dyf-5CA eliminates all the increased occupancy (fig. S13B), indicating that kinase hyperactivity causes abnormal transcription regulation. Ciliary phenotypes in dyf-5CA and dyf-5 gOE animals . Why do dyf-5CA animals aberrantly elongate cilia but dyf-5 gOE worms shorten cilia? The absence of GFP fluorescence from gfp::dyf-5CA animals indicates that DYF-5CA protein is not produced, phenocopying the long cilia phenotype in dyf-5 null. The wrmScarlet::dyf-5 gOE animals produced the wrmScarlet::DYF-5 fluorescence that was incorrectly distributed around the ciliary base (fig. S14, A, B, and G). Introducing adr-2 null into wrmScarlet::dyf-5 gOE animals restored DYF-5 localization to the distal ciliary segments (fig. S14B) where DYF-5 is normally found, indicating that DYF-5 mislocalization owing to dyf-5 gOE is suppressed by the loss of ADR-2. Consistently, deletion of adr-2 rescued the short cilia defects in dyf-5 gOE (fig. S14, C and G). Using quantitative fluorescence and quantitative real-time PCR experiments, we showed that the loss of ADR-2 does not affect the transgene expression level of wrmScarlet::dyf-5 gOE or osm-6::gfpOE (fig. S14, D and E), which indicates that ADR-2 does not regulate dyf-5 gOE phenotypes through promoting transgene expression. Because DYF-5 prevents cilia overgrowth ( 2 ), mislocalized DYF-5 at the ciliary base might have blocked cilia elongation in dyf-5 gOE . To test this possibility, we artificially targeted DYF-5 to the ciliary base by fusing dyf-5 cDNA with a ciliary transition zone gene mksr-2 . Indeed, mksr-2::dyf-5cDNA OE animals developed short cilia like that of dyf-5 gOE (fig. S14, C, F, and G); however, overexpressed DYF-5 protein from dyf-5 cDNA correctly localized at the ciliary distal and did not impair cilia (fig. S14, B, C, and G). We suggest that as long as DYF-5 localizes at the correct place, it will not generate deleterious effects, even though its amount or activity might be higher than in WT animals. In adr-2 mutants that have increased dyf-5CA levels but lose RNA editing, the full-length DYF-5CA proteins with the C-terminal localization domain can be produced and properly localize at the distal ciliary segments; therefore, no shortening of cilia was observed. We speculate that dyf-5 gOE animals expressed abundant dyf-5 pre-mRNA, a portion of which, after RNA editing, may translate the DYF-5 kinase domain, incorrectly localizing around the ciliary base to inhibit cilia growth. By contrast, dyf-5CA animals expressed a much lower level of dyf-5CA pre-mRNA than dyf-5 gOE , and RNA editing may severely block DYF-5CA protein production, causing aberrant cilia elongation. If the dyf-5CA pre-mRNA increases under certain conditions, the animal may produce truncated DYF-5 protein that localizes at the ciliary base to inhibit cilia elongation. Cilia-to-nuclei signaling . dyf-5CA animals exhibited higher levels of dyf-5 mRNA than WT ( Fig. 3A ), which suggests negative transcriptional feedback so that the loss of DYF-5 up-regulates dyf-5 expression. We found an up-regulation of other ciliary gene expressions in the dyf-5CA and dyf-5(mn400) null allele and ciliary mutants that were defective in IFT-dynein CHE-3 and IFT-particle subunit DAF-10 (fig. S15 and data files S3 and S4). Flagella removal in Chlamydomonas caused a rapid transcription up-regulation of hundreds of flagellar-associated genes ( 36 ). Thus, a cell appears to sense whether cilia function or exist, and the “stressed” cilia may signal to the nucleus, requesting more ciliary proteins, which resembles the unfolded protein response (UPR) between the endoplasmic reticulum and the nucleus ( 37 ). Analogous to UPR regulation, an overphosphorylated ciliary protein of yet unknown identity may travel from dyf-5CA cilia into the nucleus to activate antisense transcription. This study describes a negative feedback loop in dyf-5CA animals. Given that cilia-to-nucleus signaling occurs in other ciliary mutants and deflagellated conditions, we speculate that the feedback loop elements may play roles in conditions beyond dyf-5CA . For example, WT worms that carry a mutation disabling the feedback loop (such as the intron substitution dyf-5 allele in fig. S2A) might display phenotypes under certain stressful conditions. Discussion . We propose a model through which RNA editing restricts the hyperactive DYF-5CA ciliary kinase ( Fig. 4H ): (i) In response to kinase hyperactivity, antisense RNAs are transcribed in nuclei and pair with the kinase pre-mRNA. (ii) The dsRNA recruits RNA deaminase to edit kinase pre-mRNA, which changes the kinase protein sequence and impairs kinase pre-mRNA splicing. (iii) The resultant IRs generate premature stop codons, which inhibit kinase translation and activate nonsense-mediated kinase mRNA decay. (iv) The loss of kinase production converts a biochemically hyperactive kinase to loss of function in vivo, phenocopying the kinase-null cilia. Considering that dyf-5 gOE generated more RNA editing sites on dyf-5 locus than dyf-5CA ( Fig. 2B ), we suggest that kinase gRNA OE might involve additional regulation. However, the loss of RNA editing rescued ciliary defects under kinase gDNA OE conditions ( Figs. 3C and 4, D and E , and fig. S10, A and B), underscoring the crucial role of RNA editing in safeguarding against the overactivation of kinases. In addition to ciliary kinases, an increasing number of kinase mRNAs are hyperedited during disease progression. Examples include the protein kinase R during type I interferon responses and the CDK12 kinase mRNA from ovarian cancer cells ( 38 – 40 ). Although the mechanisms and impacts may differ from each of these RNAs, a parallel does emerge: RNA editing is linked to kinase hyperactivity. Recognizing this common linkage, we suggest that in response to physiological or pathological stimuli, RNA editing may target the kinase pre-mRNA, restricting kinase production and down-regulating its activity in vivo. Our demonstration, in which suppression of RNA editing rescues ciliary defects caused by hyperactive kinases, connects cilia with other processes. Most of the ciliopathy genes are not ideal for drug development ( 6 ). We anticipate that the development of small molecules that target the pathways outside of cilia holds promise to devise strategies for the treatment of ciliopathies. Supplementary Materials . science.sciencemag.org/content/373/6558/984/suppl/DC1 Materials and Methods Figs. S1 to S15 Tables S1 to S5 References ( 41 – 47 ) MDAR Reproducibility Checklist Data S1 to S4 https://www.sciencemag.org/about/science-licenses-journal-article-reuse This is an article distributed under the terms of the Science Journals Default License . References and Notes . ? P. Lahiry , . A. Torkamani , . N. J. Schork , . R. A. Hegele . , Kinase mutations in human disease: Interpreting genotype-phenotype relationships . Nat. Rev. Genet. 11 , 60 – 74 ( 2010 ). doi: 10.1038/nrg2707 pmid: 20019687 OpenUrl CrossRef PubMed ? J. Burghoorn , . M. P. J. Dekkers , . S. Rademakers , . T. de Jong , . R. Willemsen , . G. Jansen . , Mutation of the MAP kinase DYF-5 affects docking and undocking of kinesin-2 motors and reduces their speed in the cilia of Caenorhabditis elegans . Proc. Natl. Acad. Sci. U.S.A. 104 , 7157 – 7162 ( 2007 ). doi: 10.1073/pnas.0606974104 pmid: 17420466 OpenUrl Abstract / FREE Full Text ? N. Vasan , . J. Baselga , . D. M. Hyman . , A view on drug resistance in cancer . Nature 575 , 299 – 309 ( 2019 ). doi: 10.1038/s41586-019-1730-1 pmid: 31723286 OpenUrl CrossRef PubMed ? S. A. Berman , . N. F. Wilson , . N. A. Haas , . P. A. Lefebvre . , A novel MAP kinase regulates flagellar length in Chlamydomonas . Curr. Biol. 13 , 1145 – 1149 ( 2003 ). doi: 10.1016/S0960-9822(03)00415-9 pmid: 12842015 OpenUrl CrossRef PubMed Web of Science ? Y. Omori , . T. Chaya , . K. Katoh , . N. Kajimura , . S. Sato , . K. Muraoka , . S. Ueno , . T. Koyasu , . M. Kondo , . T. Furukawa . , Negative regulation of ciliary length by ciliary male germ cell-associated kinase (Mak) is required for retinal photoreceptor survival . Proc. Natl. Acad. Sci. U.S.A. 107 , 22671 – 22676 ( 2010 ). doi: 10.1073/pnas.1009437108 pmid: 21148103 OpenUrl Abstract / FREE Full Text ? J. F. Reiter , . M. R. Leroux . , Genes and molecular pathways underpinning ciliopathies . Nat. Rev. Mol. Cell Biol. 18 , 533 – 547 ( 2017 ). doi: 10.1038/nrm.2017.60 pmid: 28698599 OpenUrl CrossRef PubMed ? R. K. Ozgül , . A. M. Siemiatkowska , . D. Yücel , . C. A. Myers , . R. W. Collin , . M. N. Zonneveld , . A. Beryozkin , . E. Banin , . C. B. Hoyng , . L. I. van den Born , . R. Bose , . W. Shen , . D. Sharon , . F. P. Cremers , . B. J. Klevering , . A. I. den Hollander , . J. C. Corbo ; European Retinal Disease Consortium . , Exome sequencing and cis-regulatory mapping identify mutations in MAK, a gene encoding a regulator of ciliary length, as a cause of retinitis pigmentosa . Am. J. Hum. Genet. 89 , 253 – 264 ( 2011 ). doi: 10.1016/j.ajhg.2011.07.005 pmid: 21835304 OpenUrl CrossRef PubMed ? B. A. Tucker , . T. E. Scheetz , . R. F. Mullins , . A. P. DeLuca , . J. M. Hoffmann , . R. M. Johnston , . S. G. Jacobson , . V. C. Sheffield , . E. M. Stone . , Exome sequencing and analysis of induced pluripotent stem cells identify the cilia-related gene male germ cell-associated kinase (MAK) as a cause of retinitis pigmentosa . Proc. Natl. Acad. Sci. U.S.A. 108 , E569 – E576 ( 2011 ). doi: 10.1073/pnas.1108918108 pmid: 21825139 OpenUrl Abstract / FREE Full Text ? P. D. Mace , . Y. Wallez , . M. F. Egger , . M. K. Dobaczewska , . H. Robinson , . E. B. Pasquale , . S. J. Riedl . , Structure of ERK2 bound to PEA-15 reveals a mechanism for rapid release of activated MAPK . Nat. Commun. 4 , 1681 ( 2013 ). doi: 10.1038/ncomms2687 pmid: 23575685 OpenUrl CrossRef PubMed ? Z. Fu , . M. J. Schroeder , . J. Shabanowitz , . P. Kaldis , . K. Togawa , . A. K. Rustgi , . D. F. Hunt , . T. W. Sturgill . , Activation of a nuclear Cdc2-related kinase within a mitogen-activated protein kinase-like TDY motif by autophosphorylation and cyclin-dependent protein kinase-activating kinase . Mol. Cell. Biol. 25 , 6047 – 6064 ( 2005 ). doi: 10.1128/MCB.25.14.6047-6064.2005 pmid: 15988018 OpenUrl Abstract / FREE Full Text ? L. A. Perkins , . E. M. Hedgecock , . J. N. Thomson , . J. G. Culotti . , Mutant sensory cilia in the nematode Caenorhabditis elegans . Dev. Biol. 117 , 456 – 487 ( 1986 ). doi: 10.1016/0012-1606(86)90314-3 pmid: 2428682 OpenUrl CrossRef PubMed Web of Science ? G. Ou , . O. E. Blacque , . J. J. Snow , . M. R. Leroux , . J. M. Scholey . , Functional coordination of intraflagellar transport motors . Nature 436 , 583 – 587 ( 2005 ). doi: 10.1038/nature03818 pmid: 16049494 OpenUrl CrossRef PubMed Web of Science ? K. Nishikura . , A-to-I editing of coding and non-coding RNAs by ADARs . Nat. Rev. Mol. Cell Biol. 17 , 83 – 96 ( 2016 ). doi: 10.1038/nrm.2015.4 pmid: 26648264 OpenUrl CrossRef PubMed ? L. A. Tonkin , . L. Saccomanno , . D. P. Morse , . T. Brodigan , . M. Krause , . B. L. Bass . , RNA editing by ADARs is important for normal behavior in Caenorhabditis elegans . EMBO J. 21 , 6025 – 6035 ( 2002 ). doi: 10.1093/emboj/cdf607 pmid: 12426375 OpenUrl Abstract / FREE Full Text ? M. C. Washburn , . B. Kakaradov , . B. Sundararaman , . E. Wheeler , . S. Hoon , . G. W. Yeo , . H. A. Hundley . , The dsRBP and inactive editor ADR-1 utilizes dsRNA binding to regulate A-to-I RNA editing across the C. elegans transcriptome . Cell Rep. 6 , 599 – 607 ( 2014 ). doi: 10.1016/j.celrep.2014.01.011 pmid: 24508457 OpenUrl CrossRef PubMed J. A. Arribere , . H. Kuroyanagi , . H. A. Hundley . , mRNA Editing, Processing and Quality Control in Caenorhabditis elegans . Genetics 215 , 531 – 568 ( 2020 ). doi: 10.1534/genetics.119.301807 pmid: 32632025 OpenUrl Abstract / FREE Full Text ? H. Q. Zhao , . P. Zhang , . H. Gao , . X. He , . Y. Dou , . A. Y. Huang , . X.-M. Liu , . A. Y. Ye , . M.-Q. Dong , . L. Wei . , Profiling the RNA editomes of wild-type C. elegans and ADAR mutants . Genome Res. 25 , 66 – 75 ( 2015 ). doi: 10.1101/gr.176107.114 pmid: 25373143 OpenUrl Abstract / FREE Full Text ? H. Ohta , . M. Fujiwara , . Y. Ohshima , . T. Ishihara . , ADBP-1 regulates an ADAR RNA-editing enzyme to antagonize RNA-interference-mediated gene silencing in Caenorhabditis elegans . Genetics 180 , 785 – 796 ( 2008 ). doi: 10.1534/genetics.108.093310 pmid: 18780728 OpenUrl Abstract / FREE Full Text D. P. Reich , . K. M. Tyc , . B. L. Bass . , C. elegans ADARs antagonize silencing of cellular dsRNAs by the antiviral RNAi pathway . Genes Dev. 32 , 271 – 282 ( 2018 ). doi: 10.1101/gad.310672.117 pmid: 29483152 OpenUrl Abstract / FREE Full Text ? M. B. Warf , . B. A. Shepherd , . W. E. Johnson , . B. L. Bass . , Effects of ADARs on small RNA processing pathways in C. elegans . Genome Res. 22 , 1488 – 1498 ( 2012 ). doi: 10.1101/gr.134841.111 pmid: 22673872 OpenUrl Abstract / FREE Full Text ? F. Zhang , . Y. Lu , . S. Yan , . Q. Xing , . W. Tian . , SPRINT: An SNP-free toolkit for identifying RNA editing sites . Bioinformatics 33 , 3538 – 3548 ( 2017 ). doi: 10.1093/bioinformatics/btx473 pmid: 29036410 OpenUrl CrossRef PubMed ? R. Middleton , . D. Gao , . A. Thomas , . B. Singh , . A. Au , . J. J.-L. Wong , . A. Bomane , . B. Cosson , . E. Eyras , . J. E. J. Rasko , . W. Ritchie . , IRFinder: Assessing the impact of intron retention on mammalian gene expression . Genome Biol. 18 , 51 ( 2017 ). doi: 10.1186/s13059-017-1184-4 pmid: 28298237 OpenUrl CrossRef PubMed ? J. Rodriguez , . J. S. Menet , . M. Rosbash . , Nascent-seq indicates widespread cotranscriptional RNA editing in Drosophila . Mol. Cell 47 , 27 – 37 ( 2012 ). doi: 10.1016/j.molcel.2012.05.002 pmid: 22658416 OpenUrl CrossRef PubMed Web of Science ? T. Kurosaki , . M. W. Popp , . L. E. Maquat . , Quality and quantity control of gene expression by nonsense-mediated mRNA decay . Nat. Rev. Mol. Cell Biol. 20 , 406 – 420 ( 2019 ). doi: 10.1038/s41580-019-0126-2 pmid: 30992545 OpenUrl CrossRef PubMed ? A. Grimson , . S. O’Connor , . C. L. Newman , . P. Anderson . , SMG-1 is a phosphatidylinositol kinase-related protein kinase required for nonsense-mediated mRNA Decay in Caenorhabditis elegans . Mol. Cell. Biol. 24 , 7483 – 7490 ( 2004 ). doi: 10.1128/MCB.24.17.7483-7490.2004 pmid: 15314158 OpenUrl Abstract / FREE Full Text ? L. Johns , . A. Grimson , . S. L. Kuchma , . C. L. Newman , . P. Anderson . , Caenorhabditis elegans SMG-2 selectively marks mRNAs containing premature translation termination codons . Mol. Cell. Biol. 27 , 5630 – 5638 ( 2007 ). doi: 10.1128/MCB.00410-07 pmid: 17562857 OpenUrl Abstract / FREE Full Text ? F. Usuki , . A. Yamashita , . T. Shiraishi , . A. Shiga , . O. Onodera , . I. Higuchi , . S. Ohno . , Inhibition of SMG-8, a subunit of SMG-1 kinase, ameliorates nonsense-mediated mRNA decay-exacerbated mutant phenotypes without cytotoxicity . Proc. Natl. Acad. Sci. U.S.A. 110 , 15037 – 15042 ( 2013 ). doi: 10.1073/pnas.1300654110 pmid: 23983263 OpenUrl Abstract / FREE Full Text ? R. R. Chivukula , . D. T. Montoro , . H. M. Leung , . J. Yang , . H. E. Shamseldin , . M. S. Taylor , . G. W. Dougherty , . M. A. Zariwala , . J. Carson , . M. L. A. Daniels , . P. R. Sears , . K. E. Black , . L. P. Hariri , . I. Almogarri , . E. M. Frenkel , . V. Vinarsky , . H. Omran , . M. R. Knowles , . G. J. Tearney , . F. S. Alkuraya , . D. M. Sabatini . , A human ciliopathy reveals essential functions for NEK10 in airway mucociliary clearance . Nat. Med. 26 , 244 – 251 ( 2020 ). doi: 10.1038/s41591-019-0730-x pmid: 31959991 OpenUrl CrossRef PubMed A. K. Maurya , . T. Rogers , . P. Sengupta . , A CCRK and a MAK Kinase Modulate Cilia Branching and Length via Regulation of Axonemal Microtubule Dynamics in Caenorhabditis elegans . Curr. Biol. 29 , 1286 – 1300.e4 ( 2019 ). doi: 10.1016/j.cub.2019.02.062 pmid: 30955935 OpenUrl CrossRef PubMed ? K. M. Power , . J. S. Akella , . A. Gu , . J. D. Walsh , . S. Bellotti , . M. Morash , . W. Zhang , . Y. H. Ramadan , . N. Ross , . A. Golden , . H. E. Smith , . M. M. Barr , . R. O’Hagan . , Mutation of NEKL-4/NEK10 and TTLL genes suppress neuronal ciliary degeneration caused by loss of CCPP-1 deglutamylase function . PLOS Genet. 16 , e1009052 ( 2020 ). doi: 10.1371/journal.pgen.1009052 pmid: 33064774 OpenUrl CrossRef PubMed ? S. H. Schmidt , . M. J. Knape , . D. Boassa , . N. Mumdey , . A. P. Kornev , . M. H. Ellisman , . S. S. Taylor , . F. W. Herberg . , The dynamic switch mechanism that leads to activation of LRRK2 is embedded in the DFGψ motif in the kinase domain . Proc. Natl. Acad. Sci. U.S.A. 116 , 14979 – 14988 ( 2019 ). doi: 10.1073/pnas.1900289116 pmid: 31292254 OpenUrl Abstract / FREE Full Text ? L. Xia , . D. Robinson , . A.-H. Ma , . H.-C. Chen , . F. Wu , . Y. Qiu , . H.-J. Kung . , Identification of human male germ cell-associated kinase, a kinase transcriptionally activated by androgen in prostate cancer cells . J. Biol. Chem. 277 , 35422 – 35433 ( 2002 ). doi: 10.1074/jbc.M203940200 pmid: 12084720 OpenUrl Abstract / FREE Full Text ? S. W. Knight , . B. L. Bass . , The role of RNA editing by ADARs in RNAi . Mol. Cell 10 , 809 – 817 ( 2002 ). doi: 10.1016/S1097-2765(02)00649-4 pmid: 12419225 OpenUrl CrossRef PubMed Web of Science ? H. Tabara , . E. Yigit , . H. Siomi , . C. C. Mello . , The dsRNA binding protein RDE-4 interacts with RDE-1, DCR-1, and a DExH-box helicase to direct RNAi in C. elegans . Cell 109 , 861 – 871 ( 2002 ). doi: 10.1016/S0092-8674(02)00793-6 pmid: 12110183 OpenUrl CrossRef PubMed Web of Science ? V. Pelechano , . L. M. Steinmetz . , Gene regulation by antisense transcription . Nat. Rev. Genet. 14 , 880 – 893 ( 2013 ). doi: 10.1038/nrg3594 pmid: 24217315 OpenUrl CrossRef PubMed ? A. J. Albee , . A. L. Kwan , . H. Lin , . D. Granas , . G. D. Stormo , . S. K. Dutcher . , Identification of cilia genes that affect cell-cycle progression using whole-genome transcriptome analysis in Chlamydomonas reinhardtti . G3 3 , 979 – 991 ( 2013 ). doi: 10.1534/g3.113.006338 pmid: 23604077 OpenUrl Abstract / FREE Full Text ? P. Walter , . D. Ron . , The unfolded protein response: From stress pathway to homeostatic regulation . Science 334 , 1081 – 1086 ( 2011 ). doi: 10.1126/science.1209038 pmid: 22116877 OpenUrl Abstract / FREE Full Text ? C. Calabrese , . N. R. Davidson , . D. Demircio?lu , . N. A. Fonseca , . Y. He , . A. Kahles , . K.-V. Lehmann , . F. Liu , . Y. Shiraishi , . C. M. Soulette , . L. Urban , . L. Greger , . S. Li , . D. Liu , . M. D. Perry , . Q. Xiang , . F. Zhang , . J. Zhang , . P. Bailey , . S. Erkek , . K. A. Hoadley , . Y. Hou , . M. R. Huska , . H. Kilpinen , . J. O. Korbel , . M. G. Marin , . J. Markowski , . T. Nandi , . Q. Pan-Hammarstr?m , . C. S. Pedamallu , . R. Siebert , . S. G. Stark , . H. Su , . P. Tan , . S. M. Waszak , . C. Yung , . S. Zhu , . P. Awadalla , . C. J. Creighton , . M. Meyerson , . B. F. F. Ouellette , . K. Wu , . H. Yang , . A. Brazma , . A. N. Brooks , . J. G?ke , . G. R?tsch , . R. F. Schwarz , . O. Stegle , . Z. Zhang ; PCAWG Transcriptome Core Group ; PCAWG Transcriptome Working Group ; PCAWG Consortium . , Genomic basis for RNA alterations in cancer . Nature 578 , 129 – 136 ( 2020 ). doi: 10.1038/s41586-020-1970-0 pmid: 32025019 OpenUrl CrossRef PubMed H. Chung , . J. J. A. Calis , . X. Wu , . T. Sun , . Y. Yu , . S. L. Sarbanes , . V. L. Dao Thi , . A. R. Shilvock , . H.-H. Hoffmann , . B. R. Rosenberg , . C. M. Rice . , Human ADAR1 Prevents Endogenous RNA from Triggering Translational Shutdown . Cell 172 , 811 – 824.e14 ( 2018 ). doi: 10.1016/j.cell.2017.12.038 pmid: 29395325 OpenUrl CrossRef PubMed ? J. J. Ishizuka , . R. T. Manguso , . C. K. Cheruiyot , . K. Bi , . A. Panda , . A. Iracheta-Vellve , . B. C. Miller , . P. P. Du , . K. B. Yates , . J. Dubrot , . I. Buchumenski , . D. E. Comstock , . F. D. Brown , . A. Ayer , . I. C. Kohnle , . H. W. Pope , . M. D. Zimmer , . D. R. Sen , . S. K. Lane-Reticker , . E. J. Robitschek , . G. K. Griffin , . N. B. Collins , . A. H. Long , . J. G. Doench , . D. Kozono , . E. Y. Levanon , . W. N. Haining . , Loss of ADAR1 in tumours overcomes resistance to immune checkpoint blockade . Nature 565 , 43 – 48 ( 2019 ). doi: 10.1038/s41586-018-0768-9 pmid: 30559380 OpenUrl CrossRef PubMed ? S. Brenner . , The genetics of Caenorhabditis elegans . Genetics 77 , 71 – 94 ( 1974 ). doi: 10.1093/genetics/77.1.71 pmid: 4366476 OpenUrl Abstract / FREE Full Text D. J. Dickinson , . J. D. Ward , . D. J. Reiner , . B. Goldstein . , Engineering the Caenorhabditis elegans genome using Cas9-triggered homologous recombination . Nat. Methods 10 , 1028 – 1034 ( 2013 ). doi: 10.1038/nmeth.2641 pmid: 23995389 OpenUrl CrossRef PubMed Web of Science A. E. Friedland , . Y. B. Tzur , . K. M. Esvelt , . M. P. Colaiácovo , . G. M. Church , . J. A. Calarco . , Heritable genome editing in C. elegans via a CRISPR-Cas9 system . Nat. Methods 10 , 741 – 743 ( 2013 ). doi: 10.1038/nmeth.2532 pmid: 23817069 OpenUrl CrossRef PubMed Web of Science E. M. Hedgecock , . J. G. Culotti , . J. N. Thomson , . L. A. Perkins . , Axonal guidance mutants of Caenorhabditis elegans identified by filling sensory neurons with fluorescein dyes . Dev. Biol. 111 , 158 – 170 ( 1985 ). doi: 10.1016/0012-1606(85)90443-9 pmid: 3928418 OpenUrl CrossRef PubMed Web of Science T. A. Starich , . R. K. Herman , . C. K. Kari , . W. H. Yeh , . W. S. Schackwitz , . M. W. Schuyler , . J. Collet , . J. H. Thomas , . D. L. Riddle . , Mutations affecting the chemosensory neurons of Caenorhabditis elegans . Genetics 139 , 171 – 188 ( 1995 ). doi: 10.1093/genetics/139.1.171 pmid: 7705621 OpenUrl Abstract / FREE Full Text P. N. Inglis , . G. Ou , . M. R. Leroux , . J. M. Scholey . , The sensory cilia of Caenorhabditis elegans . WormBook 1 – 22 ( 2007 ). doi: 10.1895/wormbook.1.126.2 pmid: 18050505 OpenUrl CrossRef PubMed ? A. Mukhopadhyay , . B. Deplancke , . A. J. Walhout , . H. A. Tissenbaum . , Chromatin immunoprecipitation (ChIP) coupled to detection by quantitative real-time PCR to study transcription factor binding to DNA in Caenorhabditis elegans . Nat. Protoc. 3 , 698 – 709 ( 2008 ). doi: 10.1038/nprot.2008.38 pmid: 18388953 OpenUrl CrossRef PubMed Web of Science Acknowledgments: We thank J. Reiter, X. Yang, V. Ambros, D. Xue, Z. Lu, X. Shen, Y. Qi, S. Cai, and Y. Wang for discussion. Funding: This work was supported by the following funding programs: National Natural Science Foundation of China (grants 31991190, 31730052, 31525015, 31861143042, 31561130153, 31671444, and 31871352) and National Key R&D Program of China (2017YFA0503501, 2019YFA0508401, and 2017YFA0102900). Author contributions: D.L., P.Y., Z.Z., W.L., Q.C.Z., J.B.L., and G.O. conceived and designed the project. D.L., Z.Z., and P.Y. performed research; D.L. and Y.L. analyzed data; D.L., Z.Z., W.L., Q.C.Z., J.B.L., and G.O. wrote the paper. Competing interests: G.O., W.L., D.L., Z.Z., and P.Y. are inventors on patent application no. CN202110893049.4, submitted by Tsinghua University, that covers applications of RNA editing on genetic disorders that are associated with gain-of-function proteins. Data and materials availability: All data are available in the main text or supplementary materials. All sequencing data from this study have been submitted to the NCBI Sequence Read Archive (SRA; www.ncbi.nlm.nih.gov/sra ) with BioProject accession number PRJNA655606. .