Cas13 can be harnessed for RNA knockdown and editing in mammalian cells, opening up a range of applications in RNA biology.
CRISPR–Cas proteins are established and powerful tools for DNA editing. Now a pair of papers begins to explore their potential to edit another nucleic acid, RNA, opening up numerous applications in RNA biology. In October, Abudayyeh et al.1 showed in Nature that Cas13a (previously known as C2C2) can be programmed to induce decay of specific mRNA and that its catalytically inactive mutant (dCas13a) enabled mRNA imaging and RNA purification in vitro. More recently, a paper in Science by Cox et al.2 describes the engineering of yet another homolog of Cas13. Cas13b can mediate RNA silencing with surprisingly high specificity and efficiency. By fusing a catalytically dead Cas13b mutant (dCas13b) to a hyperactive adenosine deaminase (ADAR) mutant, Cox et al.2 created a programmable RNA editing enzyme that mediates A-to-I editing (Fig. 1). These papers provide new tools for investigating RNA biology and provide new therapeutic options for diseases caused by nonsense or other point mutations.
The Cas13 proteins (Cas13a, Cas13b, and Cas13c) are found in class 2 subtype VI CRISPR–Cas systems that use a single effector protein3. The active endoRNase comprises just two components: the Cas13 protein and a 64- to 66-nt CRISPR RNA (crRNA; which functions alone as the guide RNA) that recognizes a 28- to 30-nt sequence on the target RNA via a so-called spacer sequence. Structural studies on Cas13a revealed two catalytic activities4,5. First, Cas13a catalyzes the processing of pre-crRNAs to generate mature crRNAs. A second and separate catalytic activity mediates the cleavage of the target RNA. This catalytic site is formed between two HEPN domains of Cas13 when the guide RNA base-pairs with the target RNA.
Several properties of Cas13 make it an attractive candidate for an RNA-modulating tool. First, Cas13 can be programmed to target potentially any RNA by changing the crRNA spacer sequence. Second, Cas13 (particularly Cas13b) does not require a specific sequence element in the target site, which is different from Cas9 and Cpf1 that require a specific PAM (protospacer adjacent motif) to be present in the DNA. Third, the effector complex is one of the simplest among the CRISPR–Cas systems. This two component system (Cas13–crRNA) is easier to manipulate and deliver than trimeric or multimeric systems. Fourth, because Cas13 carries out both crRNA biogenesis and silencing, multiple crRNAs can be loaded at the same time. Lastly, the catalytic mutant of Cas13 retains a high affinity to the target so the mutant can be used as a sequence-specific RNA-binding module, which can be fused to other functional modules.
In previous studies using Cas13a from Leptotrichia, Cas13a was shown to have a 'collateral' cleavage activity, which means that Cas13a non-specifically cleaves other unbound RNAs once it is activated by base pairing with the on-target sequence6,7. Based on this result, it was proposed that the function of the Cas13 system is to induce dormancy or programmed cell death in response to phage infection. However, this collateral activity does not seem to exist with other Cas13 proteins expressed in mammalian cells. Cox et al.2 screened various Cas13 proteins and found that PspCas13b is highly specific to on-target sequences in knockdown experiments. Preliminary transcriptomic and gene-level analyses indicated that PspCas13b may be generally more specific in knockdown experiments than are short-hairpin RNAs (shRNAs). Silencing efficiency was also impressive, with 60–90% suppression. However, care might be needed in designing guide RNAs to avoid off-target effects, as the central seed tolerates single mismatches to some extent. It remains to be determined how some Cas13 achieve such high selectivity, and if the specificity can be ensured under all cellular and molecular contexts.
Separate from applications that require RNA cleavage, Cas13a's RNA binding activity may also be used in many other applications. Site-directed RNA editing has been a difficult technical challenge. ADAR enzymes naturally target adenosine in RNA duplex with some sequence specificity8. Cox et al.2 tackle this problem by fusing dCas13b with an hyperactive ADAR mutant that has a reduced sequence constraint and an enhanced editing rate2,9. The authors introduced further mutations to ADAR to reduce non-specific RNA-binding activity so as to avoid off-target editing. One such mutant (ADARDD) fused to dCas13b achieved high specificity toward the intended sites as well as relatively high frequency (10–40%) of editing. ADAR mutants preferentially deaminate adenosine mispaired with cytidine in RNA duplexes, which is exploited to enhance precise base editing by introducing A-C mismatch to the guide-target duplex region. ADAR induces the conversion of adenosine to inosine which is equivalent to guanosine in translation and splicing. This technique, referred to as “RNA editing for programmable A-to-I replacement” (REPAIR) can indeed repair disease-relevant G>A nonsense mutations in cell culture, including AVPR2 W293X (878G>A) found in X-linked nephrogenic diabetes insipidus and FANCC W506X (1517G>A) associated with Fanconi anemia. Thus, this study shows a possibility for programmable RNA editing as a new therapeutic avenue, although any therapeutic applications will clearly warrant further improvements of the technique.
Moreover, this dCas13b platform is likely to offer ample opportunities to create new tools for RNA research. It will be interesting to test in the future if C-to-U editing is plausible when APOBEC is used instead of ADAR. The emerging field of epitranscriptomics will also benefit from versions of Cas13 that allow manipulating epitranscriptome marks directly by fusing Cas13b to RNA-modifying enzymes such as methyltransferases.
What's more, many other regulatory factors, such as splicing factors, translation factors, or localization factors, could also be studied with appropriately designed Cas13 variants. Thus, fusion proteins may be useful in correcting pathological misregulation as well as in investigating the biological functions of the regulatory factors.
Live RNA imaging is another area to which dCas13 can contribute greatly. The same group previously demonstrated that fluorescence protein fusion to dCas13 can be used to track cellular RNAs in real time1. Moreover, dCas13 may be engineered to harbor tandem tags for purification of RNA–protein complexes and identify the RNA-binding proteins bound to a specific endogenous RNA. In-depth understanding of the Cas13 complex, particularly as gained through structural studies, will be crucial in engineering this new arsenal.