Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2017 Oct 12;550(7675):280-284.
doi: 10.1038/nature24049. Epub 2017 Oct 4.

RNA Targeting With CRISPR-Cas13

Affiliations
Free PMC article

RNA Targeting With CRISPR-Cas13

Omar O Abudayyeh et al. Nature. .
Free PMC article

Abstract

RNA has important and diverse roles in biology, but molecular tools to manipulate and measure it are limited. For example, RNA interference can efficiently knockdown RNAs, but it is prone to off-target effects, and visualizing RNAs typically relies on the introduction of exogenous tags. Here we demonstrate that the class 2 type VI RNA-guided RNA-targeting CRISPR-Cas effector Cas13a (previously known as C2c2) can be engineered for mammalian cell RNA knockdown and binding. After initial screening of 15 orthologues, we identified Cas13a from Leptotrichia wadei (LwaCas13a) as the most effective in an interference assay in Escherichia coli. LwaCas13a can be heterologously expressed in mammalian and plant cells for targeted knockdown of either reporter or endogenous transcripts with comparable levels of knockdown as RNA interference and improved specificity. Catalytically inactive LwaCas13a maintains targeted RNA binding activity, which we leveraged for programmable tracking of transcripts in live cells. Our results establish CRISPR-Cas13a as a flexible platform for studying RNA in mammalian cells and therapeutic development.

Figures

Extended Data Fig. 1
Extended Data Fig. 1. Evaluation of LwaCas13a PFS preferences and comparisons to LshCas13a
a, Sequence comparison tree of the fifteen Cas13a orthologs evaluated in this study. b, Ratios of in vivo activity from Fig. 1B. c, Distributions of PFS enrichment for LshCas13a and LwaCas13a in targeting and non-targeting samples. d, Number of LshCas13a and LwaCas13a PFS sequences above depletion threshold for varying depletion thresholds. e, Distributions of PFS enrichment for LshCas13a and LwaCas13a in targeting samples, normalized to non-targeting samples. f, Sequence logos and counts for remaining PFS sequences after LshCas13a cleavage at varying enrichment cutoff thresholds. g, Sequence logos and counts for remaining PFS sequences after LwaCas13a cleavage at varying enrichment cutoff thresholds.
Extended Data Fig. 2
Extended Data Fig. 2. Biochemical characterization of LwaCas13a RNA cleavage activity
a, LwaCas13a has more active RNAse activity than LshCas13a. b, Gel electrophoresis of ssRNA1 after incubation with LwaCas13a and with and without crRNA 1 for varying amounts of times. c, Gel electrophoresis of ssRNA1 after incubation with varying amounts of LwaCas13a-crRNA complex. d, Sequence and structure of ssRNA 4 and ssRNA 5. crRNA spacer sequence is highlighted in blue. e, Gel electrophoresis of ssRNA 4 and ssRNA 5 after incubation with LwaCas13a and crRNA 1. f, Sequence and structure of ssRNA 4 with sites of poly-x modifications highlighted in red. crRNA spacer sequence is highlighted in blue. g, Gel electrophoresis of ssRNA 4 with each of 4 possible poly-x modifications incubated with LwaCas13a and crRNA 1. h, LwaCas13a can process pre-crRNA from the L. wadei CRISPR-Cas locus. i, Cleavage efficiency of ssRNA 1 for crRNA spacer truncations after incubation with LwaCas13a.
Extended Data Fig. 3
Extended Data Fig. 3. Engineering and optimization of LwaCas13a for mammalian knockdown
a, Knockdown of Gluc transcript by LwaCas13a and Gluc guide 1 spacers of varying length. b, Knockdown of Gluc transcript with Gluc guide 1 and varying amounts of transfected LwaCas13a plasmid. c, Knockdown of Gluc transcript by LwaCas13a and varying amounts of transfected Gluc guide 1 and 2 plasmid. d, Knockdown of Gluc transcript using guides expressed from either U6 or tRNAVal promoters. e, Knockdown of KRAS transcript using guides expressed from either U6 or tRNAVal promoters. f, Knockdown of KRAS and CXC4 transcripts by LwaCas13a using guides transfected in A375 cells with shRNA comparisons. g, Knockdown of Gluc transcript and endogenous transcripts PPIB, KRAS,and CXCR4 with active and catalytically dead Cas13a. h, Validation of the top three guides from the arrayed knockdown Gluc and Cluc screens with shRNA comparisons. i, Arrayed knockdown screen of 93 guides evenly tiled across the XIST transcript. All values are mean ± SEM with n = 3. **p< 0.01; *p< 0.05. ns = not significant. A two-tailed student's T-test was used for comparisons.
Extended Data Fig. 4
Extended Data Fig. 4. LwaCas13a targeting efficiency is influenced by accessibility along the transcript
a, First row: Top knockdown guides are plotted by position along target transcript. The top 20% of guides are chosen for Gluc and top 30% of guides for Cluc, KRAS, and PPIB. Second row: Histograms for the pairwise distance between adjacent top guides for each transcript (blue) compared to a random null-distribution (red). Inset shows the cumulative frequency curves for these histograms. A shift of the blue curve (actual measured distances) to the left of the red curve (null distribution of distances) indicates that guides are closer together than expected by chance. b, Gluc, Cluc, PPIB, and KRAS knockdown partially correlates with target accessibility as measured by predicted folding of the transcript. c,Kernel density estimation plots depicting the correlation between target accessibility (probability of a region being base-paired) and target expression after knockdown by LwaCas13a. d, First row: Correlations between target expression and target accessibility (probability of a region being base-paired) measured at different window sizes (W) and for different k-mer lengths. Second row: P-values for the correlations between target expression and target accessibility (probability of a region being base-paired) measured at different window sizes (W) and for different k-mer lengths. The color scale is designed such that p-values > 0.05 are shades of red and p-values < 0.05 are shades of blue.
Extended Data Fig. 5
Extended Data Fig. 5. Detailed evaluation of LwaCas13a sensitivity to mismatches in the guide:target duplex at varying spacer lengths
a, Knockdown of KRAS evaluated with guides containing single mismatches at varying positions across the spacer sequence. b, Knockdown of PPIB evaluated with guides containing single mismatches at varying positions across the spacer sequence. c, Knockdown of Gluc evaluated with guides containing non-consecutive double mismatches at varying positions across the spacer sequence. The wild-type sequence is shown at the top with mismatch identities shown below. d, Collateral cleavage activity on ssRNA 1 and 2 for varying spacer lengths. (n=4 technical replicates; bars represent mean ± s.e.m.) e, Specificity ratios of guide tested in (d). Specificity ratios are calculated as the ratio of the on-target RNA (ssRNA 1) collateral cleavage to the off-target RNA (ssRNA 2) collateral cleavage. (n=4 technical replicates; bars represent mean ± s.e.m.) f, Collateral cleavage activity on ssRNA 1 and 2 for 28 nt spacer crRNA with synthetic mismatches tiled along the spacer. (n=4 technical replicates; bars represent mean ± s.e.m.) g, Specificity ratios, as defined in (e), of crRNA tested in (f). (n=4 technical replicates; bars represent mean ± s.e.m.) h, Collateral cleavage activity on ssRNA 1 and 2 for 23 nt spacer crRNA with synthetic mismatches tiled along the spacer. (n=4 technical replicates; bars represent mean ± s.e.m.) i, Specificity ratios, as defined in (e), of crRNA tested in (h). (n=4 technical replicates; bars represent mean ± s.e.m.) j, Collateral cleavage activity on ssRNA 1 and 2 for 20 nt spacer crRNA with synthetic mismatches tiled along the spacer. (n=4 technical replicates; bars represent mean ± s.e.m.) k, Specificity ratios, as defined in (e), of crRNA tested in (j). (n=4 technical replicates; bars represent mean ± s.e.m.).
Extended Data Fig. 6
Extended Data Fig. 6. LwaCas13a is more specific than shRNA knockdown on endogenous targets
a, Left: Expression levels in log2(transcripts per million (TPM)) values of all genes detected in RNA-seq libraries of non-targeting shRNA-transfected control (x-axis) compared to KRAS-targeting shRNA (y-axis). Shown is the mean of three biological replicates. The KRAS transcript data point is colored in red. Right: Expression levels in log2(transcripts per million (TPM)) values of all genes detected in RNA-seq libraries of non-targeting LwaCas13a-guide-transfected control (x-axis) compared to KRAS-targeting LwaCas13a-guide (y-axis). Shown is the mean of three biological replicates. The KRAS transcript data point is colored in red. b, Left: Expression levels in log2(transcripts per million (TPM)) values of all genes detected in RNA-seq libraries of non-targeting shRNA-transfected control (x-axis) compared to PPIB-targeting shRNA (y-axis). Shown is the mean of three biological replicates. The PPIB transcript data point is colored in red. Right: Expression levels in log2(transcripts per million (TPM)) values of all genes detected in RNA-seq libraries of non-targeting LwaCas13a-guide-transfected control (x-axis) compared to PPIB-targeting LwaCas13a-guide (y-axis). Shown is the mean of three biological replicates. The PPIB transcript data point is colored in red. c, Comparisons of individual replicates of non-targeting shRNA conditions (first row) and Gluc-targeting shRNA conditions (second row). d, Comparisons of individual replicates of non-targeting guide conditions (first row) and Gluc-targeting guide conditions (second row). e, Pairwise comparisons of individual replicates of non-targeting shRNA conditions against the Gluc-targeting shRNA conditions. f, Pairwise comparisons of individual replicates of non-targeting guide conditions against the Gluc-targeting guide conditions.
Extended Data Fig. 7
Extended Data Fig. 7. Detailed analysis of LwaCas13a and RNAi knockdown variability (standard deviation) across all samples
a, Heatmap of correlations (Kendall's tau) for log2(transcripts per million (TPM+1)) values of all genes detected in RNA-seq libraries between targeting and non-targeting replicates for shRNA or guide targeting either luciferase reporters or endogenous genes. b, Heatmap of correlations (Kendall's tau) for log2(transcripts per million (TPM+1)) values of all genes detected in RNA-seq libraries between all replicates and perturbations. c,Distributions of standard deviations for log2(transcripts per million (TPM+1)) values of all genes detected in RNA-seq libraries among targeting and non-targeting replicates for each gene targeted for either shRNA or guide.
Extended Data Fig. 8
Extended Data Fig. 8. LwaCas13a knockdown is specific to the targeted transcript with no activity on a measured off-target transcript
a, Heatmap of absolute Gluc signal for first 96 spacers tiling Gluc. b, Heatmap of absolute Cluc signal for first 96 spacers tiling Gluc. c, Relationship between absolute Gluc signal and normalized luciferase for Gluc tiling guides. d, Relationship between absolute Cluc signal and normalized luciferase for Gluc tiling guides. e, Relationship between PPIB 2-Ct levels and PPIB knockdown for PPIB tiling guides. f, Relationship between GAPDH 2-Ct levels and PPIB knockdown for PPIB tiling guides. g, Relationship between KRAS 2-Ct levels and KRAS knockdown for KRAS guides. h, Relationship between GAPDH 2-Ct levels and KRAS knockdown for KRAS guides. i, Bioanalyzer traces of total RNA isolated from cells transfected with Gluc-targeting guides 1 and 2 or non-targeting guide from the experiment with active Cas13a in Extended Data Fig. 3g. The RNA-integrity number (RIN) is shown and 18S rRNA and 28S rRNA peaks are labeled above. A student's t-test shows no significant difference for the RIN between either of the targeting conditions and the non-targeting condition. The curves are shown as a mean of three replicates and the shaded area in red around the curves show the s.e.m. j, The Bioanalyzer trace for the RNA ladder with peak sizes labeled above.
Extended Data Fig. 9
Extended Data Fig. 9. dCas13a-NF can be used for ACTB imaging
a, Comparison between localization of dCas13-GFP and dCas13a-GFP-KRAB constructs for imaging ACTB. Scale bars, 10μm b, Additional fields of view of the dCas13a-NLS-msfGFP negative-feedback construct delivered with a non-targeting guide. Scale bars, 10μm. c, Additional fields of view of the dCas13a-NLS-msfGFP negative-feedback construct delivered with ACTB guide 3. Scale bars, 10μm. d, Additional fields of view of the dCas13a-NLS-msfGFP negative-feedback construct delivered with ACTB guide 4. Scale bars, 10μm.
Extended Data Fig. 10
Extended Data Fig. 10. dCas13a-NF can image stress granule formation in living cells
a, Representative images from RNA FISH of the ACTB transcript in dCas13a-NF-expressing cells with corresponding ACTB-targeting and non-targeting guides. Cell outline is shown with a dashed line. Scale bars, 10μm b, Overall signal overlap between ACTB RNA FISH signal and dCas13a-NF quantified by the Mander's overlap coefficient (left) and Pearson's correlation (right). Correlations and signal overlap are calculated pixel-by-pixel on a per cell basis. All values are mean ± SEM with n = 3. ****p< 0.0001; ***p< 0.001; **p < 0.01. A two-tailed student's T-test was used for comparisons. c, Representative images from live-cell analysis of stress granule formation in response to 400 uM sodium arsenite treatment. Scale bars, 20μm d, Quantitation of stress granule formation in response to sodium arsenite treatment. Quantitation is based on overlapping dCas13a-NF and G3BP1 puncta. All values are mean ± SEM with n = 3. ****p< 0.0001; ***p< 0.001; **p< 0.01; *p< 0.05. ns = not significant. A two-tailed student's T-test was used for comparisons.
Figure 1
Figure 1. Cas13a from Leptotrichia wadei (LwaCas13a) is capable of eukaryotic transcript knockdown
a, Schematic of PFS characterization screen on Cas13a orthologs. b, Quantitation of Cas13a activity in E. coli measured by colony survival from PFS screen. c, in vivo PFS screening shows LwaCas13a has a minimal PFS preference. Error bars indicate an approximate Bayesian 95% confidence interval. d, Imaging showing localization and expression of each of the mammalian constructs. Scale bars, 10μm. e, Schematic of the mammalian luciferase reporter system used to evaluate knockdown. f, Knockdown of Gaussia luciferase (Gluc) using engineered variants of LwaCas13a. Sequences for guides and shRNAs are shown above. g, Knockdown of three different endogenous transcripts with LwaCas13a compared against corresponding RNAi constructs. h, Schematic for LwaCas13a knockdown of transcripts in rice (Oryza sativa) protoplasts. i, LwaCas13a knockdown of three transcripts in O. sativa protoplasts using three targeting guides per transcript. All values are mean ± SEM with n = 3, unless otherwise noted.
Figure 2
Figure 2. LwaCas13a arrayed screening of mammalian coding and non-coding RNA targets and multiplexed guide delivery
a, Schematic of LwaCas13a arrayed screening. b-e, Arrayed knockdown screen of 186 guides evenly tiled across the Gluc transcript (b) or 93 guides evenly tiled across each of the Cluc (c), KRAS (d), and PPIB (e) transcripts. f, Validation of the top three guides from the endogenous arrayed knockdown screens with shRNA comparisons. All values are mean ± SEM with n = 3. ***p< 0.001; **p < 0.01; two-tailed student's T-test). g, Arrayed knockdown screen of 93 guides evenly tiled across the MALAT1 transcript. h, Validation of top three guides from the endogenous arrayed MALAT1 knockdown screen with shRNA comparisons. i, Multiplexed delivery of five guides in a CRISPR array against five different endogenous genes under the expression of a single promoter is capable of robust knockdown. j, Multiplexed delivery of three guides against three different endogenous genes or with constructs replacing each of the guides with a non-targeting sequence shows specific knockdown of the genes targeted. All values are mean ± SEM with n = 3.
Figure 3
Figure 3. Evaluation of LwaCas13a knockdown specificity and comparisons to RNA interference
a, b Knockdown of Gluc (a) or CXCR4 (b) evaluated with guides containing single mismatches at varying positions across the spacer sequence (shown above). c, Knockdown of Gluc evaluated with guide 3 containing single or double mismatches at varying positions across the spacer sequence (shown above). d, e Expression levels in log2(transcripts per million (TPM)) values of all genes detected in RNA-seq libraries of non-targeting control (x-axis) compared to Gluc-targeting condition (y-axis) for shRNA (d) and LwaCas13a (e). Shown is the mean of three biological replicates. The Gluc transcript data point is colored in red. The guide sequence used is shown above. f, Differential gene expression analysis of six RNA-seq libraries (each with three biological replicates) comparing LwaCas13a knockdown to shRNA knockdown at three different genes. g, Quantified mean knockdown levels for the targeted genes from the RNA seq libraries. h, Luciferase knockdown (left), cell viability (middle), and LwaCas13a-GFP expression (right) for cells transfected with LwaCas13a for 72 hours with and without selection. All values are mean ± SEM with n = 3.
Figure 4
Figure 4. Catalytically-inactive LwaCas13a (dCas13a) is capable of binding transcripts and tracking stress granule formation
a, Schematic of RNA immunoprecipitation for quantitation of dCas13a binding. b, dCas13a targeting Gluc and ACTB transcripts is significantly enriched compared to non-targeting controls. c, Schematic of dCas13a-GFP-KRAB construct used for negative-feedback imaging. d, Representative images for dCas13a-GFP-KRAB imaging with multiple guides targeting ACTB. Scale bars, 10μm. e, Quantitation of translocation of dCas13a-GFP-KRAB. f, Representative immunofluorescence images of HEK293FT cells treated with 400 uM sodium arsenite. Stress granules are indicated by G3BP1 staining. Scale bars, 5μm. g, G3BP1 and dCas13a-GFP-KRAB co-localization quantified per cell by Pearson's correlation. All values are mean ± SEM with n = 3. ****p< 0.0001; ***p< 0.001; **p< 0.01; *p< 0.05. ns = not significant. A one-tailed student's t-test was used for comparisons in (b) and a two-tailed student's t-test was used for comparisons in (e) and (g).

Similar articles

See all similar articles

Cited by 184 articles

See all "Cited by" articles

Publication types

MeSH terms

Feedback