Precise transcript targeting by CRISPR-Csm complexes

Abstract

Robust and precise transcript targeting in mammalian cells remains a difficult challenge using existing approaches due to inefficiency, imprecision and subcellular compartmentalization. Here we show that the clustered regularly interspaced short palindromic repeats (CRISPR)-Csm complex, a multiprotein effector from type III CRISPR immune systems in prokaryotes, provides surgical RNA ablation of both nuclear and cytoplasmic transcripts. As part of the most widely occurring CRISPR adaptive immune pathway, CRISPR-Csm uses a programmable RNA-guided mechanism to find and degrade target RNA molecules without inducing indiscriminate trans-cleavage of cellular RNAs, giving it an important advantage over the CRISPR-Cas13 family of enzymes. Using single-vector delivery of the Streptococcus thermophilus Csm complex, we observe high-efficiency RNA knockdown (90-99%) and minimal off-target effects in human cells, outperforming existing technologies including short hairpin RNA- and Cas13-mediated knockdown. We also find that catalytically inactivated Csm achieves specific and durable RNA binding, a property we harness for live-cell RNA imaging. These results establish the feasibility and efficacy of multiprotein CRISPR-Cas effector complexes as RNA-targeting tools in eukaryotes.

Fig. 1. An all-in-one type III CRISPR-Cas system in human cells.

a, Diagram showing cis- and trans-cleavage of Cas13. b, Diagram showing S. thermophilus type III-A CRISPR-Cas locus. crRNAs are transcribed from the CRISPR array, processed by Cas6 and assemble with Csm proteins. c, Close-up of crRNA:target binding, showing the 6-nt cleavage pattern. d, Western blot showing proper size and expression of Cas/Csm proteins (red) in HEK293T cells. Csm1 and Csm4 are less stable when expressed separately. GAPDH (glyceraldehyde-3-phosphate dehydrogenase) shown as loading control (green). Arrows indicate faint bands. L, ladder; U, untransfected. One of two replicates with similar results is shown. e, Immunofluorescence showing expression and nuclear localization of Cas/Csm proteins in HEK293T cells. Scale bar, 10 μm. One of two replicates with similar results is shown. f, Relative GFP fluorescence (= MFI targeting crRNA/MFI nontargeting crRNA) of HEK293T-GFP cells transfected with plasmids expressing Cas6, Csm1-5 and the indicated GFP-targeting crRNA, measured by flow cytometry. Error bars indicate mean ± s.d. of three biological replicates. g, Same as f, but with the indicated Csm mutants (or crRNA + Cas6 only). GFP crRNA 1 was used to target GFP. Error bars indicate mean ± s.d. of three biological replicates. h, Same as f, but with GFP crRNA 1 adjusted to the indicated spacer length. Error bars indicate mean ± s.d. of three biological replicates. i, Relative GFP and RFP fluorescence of HEK293T-GFP/RFP cells transfected with plasmids expressing Cas6, Csm1-5 and the indicated crRNAs (individual or multiplexed), measured by flow cytometry. GFP crRNA 1 was used to target GFP. RFP-targeting crRNA is listed in Supplementary Table 1. Error bars indicate mean ± s.d. of three biological replicates. j, Diagram showing all-in-one delivery vector designs. k, Western blot showing proper size and expression of Cas/Csm proteins (red) in HEK293T cells. GAPDH is shown as loading control (green). Arrows indicate each subunit. One of two replicates with similar results is shown. l, Relative GFP fluorescence of HEK293T-GFP cells transfected with the indicated delivery vectors and expressing the indicated GFP-targeting crRNAs, measured by flow cytometry. Error bars indicate mean ± s.d. of three biological replicates. Source data

Fig. 2. Robust KD of endogenous nuclear and cytoplasmic RNAs.

a, Relative RNA abundance (normalized to nontargeting crRNA) of the indicated targets in HEK293T cells transfected with all-in-one plasmid expressing Cas/Csm proteins and the indicated crRNAs, measured by RT-qPCR. Error bars indicate mean ± s.d. of three biological replicates. b, Relative RNA abundance (normalized to GAPDH) of the indicated targets in untransfected HEK293T cells, measured by RT-qPCR. Error bars indicate mean ± s.d. of three biological replicates. c, Relative RNA abundance (normalized to nontargeting crRNA) of the indicated targets in HEK293T cells transfected with all-in-one plasmid expressing Cas/Csm proteins and the indicated crRNAs (multiplexed), measured by RT-qPCR. XIST crRNA 1, MALAT1 crRNA 1 and NEAT1 crRNA 2 were used to target XIST, MALAT1 and NEAT1, respectively. Error bars indicate mean ± s.d. of three biological replicates. d, Relative RNA abundance (normalized to nontargeting crRNA) of XIST and BRCA1 in HEK293T cells at the indicated times post transfection with all-in-one plasmid, measured by RT-qPCR. XIST crRNA 1 and BRCA1 crRNA 2 were used to target XIST and BRCA1, respectively. Error bars indicate mean ± s.d. of three biological replicates. e, Relative RNA abundance (normalized to nontargeting crRNA) of XIST and BRCA1 in HEK293T cells transfected with all-in-one plasmid expressing Cas/Csm proteins and intron- or exon-targeting crRNAs, measured by RT-qPCR. XIST crRNA 1 and BRCA1 crRNA 2 were used to target XIST and BRCA1 exons, respectively. Intron-targeting crRNAs are listed in Supplementary Table 1. Error bars indicate mean ± s.d. of three biological replicates. f, RNA FISH (red) for the indicated targets in HEK293T cells transfected with all-in-one plasmid expressing targeting (T) or nontargeting (NT) crRNA and RNase-active or -inactive (Mut) Cas/Csm proteins. Untransfected cells serve as internal control for transfected (green) cells. XIST crRNA 1, MALAT1 crRNA 1 and NEAT1 crRNA 2 were used to target XIST, MALAT1 and NEAT1, respectively. Scale bar, 10 μm. g, Quantification of f. One hundred transfected cells were counted for each condition. Error bars indicate mean ± s.d. of three biological replicates.

Fig. 3. RNA KD with minimal off-targets or cytotoxicity.

a,b, Scatterplots showing differential transcript levels between HEK293T cells transfected with plasmid expressing Csm, Cas13 or shRNA targeting CKB (a) or MALAT1 (b) versus EV control. Target transcript indicated in black; off-targets (≥2-fold change) indicated in red. c, Quantification of upregulated or downregulated transcripts (≥2-fold change) for each sample. CKB crRNA 1, MALAT1 crRNA 2, SMARCA1 crRNA 1 and XIST crRNA 1 were used to target CKB, MALAT1, SMARCA1 and XIST, respectively. d,e, RNA-seq read coverage across target transcripts CKB (d) or MALAT1 (e). Red arrow indicates location of crRNA/shRNA target site. f, Relative cell viability and proliferation (normalized to EV control) of HEK293T cells at the indicated times post transfection with the indicated targeting (T) or nontargeting (NT) plasmids, measured by WST-1 assay. CKB crRNA 1 was used for targeting. Error bars indicate mean ± s.d. of three biological replicates. g, Relative abundance of RFP-positive (transfected) HEK293T cells at the indicated times post transfection with the indicated targeting (T) or nontargeting (NT) plasmids, measured by flow cytometry. CKB crRNA 1 was used for targeting. Error bars indicate mean ± s.d. of three biological replicates.

Fig. 4. Live-cell RNA imaging without genetic manipulation.

a, Diagram showing Csm3-GFP fusion complex used for live-cell imaging. b, Live-cell fluorescence imaging of HEK293T cells transfected with plasmid expressing Csm3-GFP fusion complex and the indicated crRNAs (Supplementary Table 1). NT, nontargeting. Scale bar, 10 μm. c, Quantification of b. One hundred transfected cells were counted for each condition. Error bars indicate mean ± s.d. of three biological replicates.