Structure-guided engineering of CRISPR-based tools

Wenxia Yu, Yunbo Qiao

Like Comment
Read the paper

Along with the development of CRISPR technology, a series of CRISPR-based genetic and epigenetic tools, such as base editing, RNA editing, DNA methylation editing, priming editing, etc., have been developed with high programmability and precision (Figure 1). Under the guidance of Cas proteins with catalytic inactive or nickase activities, functional proteins (i.e., deaminases, activators, repressors, epigenetic modifiers) are carried to the target sites for genome or epigenome editing. These CRISPR-based tools facilitate convenient gene targeting, functional analysis of regulatory elements and epigenetic modifications and CRISPR imaging; they also provide promising strategies for gene therapy.

CRISPR-guided base editors, including adenine base editors (ABEs) and cytosine base editors (CBEs) [1, 2], facilitates the precise installation of point mutations in genomic DNA in multiple cell types and organisms, to mimic or correct pathogenic mutations, to regulate mRNA splicing or coding proteins, or to interfere post-translational modifications, as we reported previously [3]. In spite of the great progress in base editing tools, the major concerns on off-target base editing on DNA or RNA are intensively highlighted in recent studies [4-8]. The guide RNA-dependent off-target base editing on DNA resulted from mis-match pairing can be improved by optimizing Cas9 specificity, sgRNA structures, or delivery methods. However, both ABEs and CBEs can induce tens of thousands of A-to-inosine (I) or C-to-uracil (U) RNA edits transcriptome-widely in guide RNA-independent manners in human cells [8, 9]. The RNA off-target activities of base editors can be reduced by an engineered CBE variant called SECURE-BE bearing rAPOBEC1 mutations [9], as well as engineered ABE variants [8, 10, 11]. GATK HaplotypeCaller, a tool for evaluating germline single nucleotide polymorphisms (SNPs) and indels, is employed as the main tool for analyzing RNA A-to-I edits. Interestingly, rare RNA edits with 0-10% efficiency was recovered in these studies, which leads us to reanalyze ABE-induced off-target editing of cellular RNAs in our recent Nature Communications paper.

Thus, we reanalyzed the RNA-seq data regarding ABE-induced RNA off-targets using another tool GATK MuTect2 that is adaptive for calling somatic mutations, recovering 2.7-11-fold of HaplotypeCaller-revealed RNA edits. Considering the native affinity of TadA with RNA, we also analyzed the preference of RNA edits with high editing rates, revealing that adenines were efficiently edited at the classical tRNA-like structures with a relatively conserved “UACGA” motif. To destruct the RNA editing activities without affecting DNA editing events, we started to do literature retrievals regarding RNA and TadA and designed the strategy for mutating key residues responsible for RNA binding within TadA. To facilitate detection of RNA-offtargets, we constructed a reporter containing a TadA substrate. After the first round of engineering of ABE, we obtained a variant containing TadA/R153P mutation, which showed perfect reduction of RNA off-targeting activity but retained the DNA on-targeting activity, as we expected. We were excited to see this result, and though this variant to be the final optimized version. However, transcriptome analysis demonstrated that the total number of R153P variant-induced RNA edits was not reduced, while the editing rates with “UACGA” motif were significantly decreased. We speculated that the affinity preference of R153P variant was changed from tRNA-like structure to a more flexible structure. Anyway, we can conclude that R153 of TadA is effectual for deaminating RNA substrate.

Based on this observation, we performed the second round of engineering by mutating R153 into A or E or directly deleting R153. At the beginning, the deletion strategy was not considered as a prior mode for engineering ABE or mini ABE, an optimized version with only an evolved TadA. Surprisingly, the variant with deletion of R153 (del153/del153* and mini del153) turned out to be the best optimized version, with minimized RNA off-targets and comparable or lower DNA on-targeting activities, even calculated with MuTect2. Thus, we designated del153/del153* and mini del153 as the final engineered ABEs (eABEs).

Intriguingly, deletion of R153 decreased RNA off-targeting activities with a limited extent from ABE8e or ABE8s variants [12, 13], but not to a minimal level as shown in del153/del153* and mini del153 variants. We summarized all engineered ABE variants (Figure 2) and found that the mutation sites within ABE8e (Y147D and F149Y) were quite close to R153. Considering the especially high DNA on-targeting activity of ABE8e, we postulate that engineered ABE8e may have changed the characteristics of native TadA for DNA and RNA deaminase activities, such as DNA/RNA binding pockets and interacting affinities. Then, R153 dose not act as a guard amino acid for RNA substrate anymore in ABE8e.

To rethink the RNA off-targeting activity of base editors, it is possibly elicited from the overexpression effects of the deaminase-fusion proteins. When overexpressed deaminase-fusion proteins are overloaded for the nCas9-served trucks and sgRNA-guided drivers, these deaminase-fusion proteins will serve as a free deaminase to meet its RNA substrates in cellular cytoplasm. Therefore, it leads us to think about the off-targeting activity of other CRISPR-based tools, such as P300-dCas9 and HDAC1-dCas9.

In sum, combining our report here, the successful engineering of CBEs and ABEs variants expands our understanding of desired and undesired features of DNA and RNA editing activities of base editors and provides a feasible pathway available to engineer CRISPR tools based on structural analysis to minimize the unwanted properties while retaining the desired on-target editing ability for CBEs or ABEs.

Figure 1. Summary for CRISPR-based tools.


Figure 2. Summary for reported variants for ABEs. 


  1. Gaudelli, N.M., et al., Programmable base editing of A*T to G*C in genomic DNA without DNA cleavage. Nature, 2017. 551(7681): p. 464-471.
  2. Komor, A.C., et al., Programmable editing of a target base in genomic DNA without double-stranded DNA cleavage. Nature, 2016. 533(7603): p. 420-4.
  3. Yang, G., et al., Base-Editing-Mediated R17H Substitution in Histone H3 Reveals Methylation-Dependent Regulation of Yap Signaling and Early Mouse Embryo Development. Cell Rep, 2019. 26(2): p. 302-312 e4.
  4. Jin, S., et al., Cytosine, but not adenine, base editors induce genome-wide off-target mutations in rice. Science, 2019. 364(6437): p. 292-295.
  5. Kim, D., et al., Genome-wide target specificity of CRISPR RNA-guided adenine base editors. Nat Biotechnol, 2019. 37(4): p. 430-435.
  6. Liang, P., et al., Genome-wide profiling of adenine base editor specificity by EndoV-seq. Nat Commun, 2019. 10(1): p. 67.
  7. Zuo, E., et al., Cytosine base editor generates substantial off-target single-nucleotide variants in mouse embryos. Science, 2019. 364(6437): p. 289-292.
  8. Rees, H.A., et al., Analysis and minimization of cellular RNA editing by DNA adenine base editors. Sci Adv, 2019. 5(5): p. eaax5717.
  9. Grunewald, J., et al., Transcriptome-wide off-target RNA editing induced by CRISPR-guided DNA base editors. Nature, 2019. 569(7756): p. 433-437.
  10. Zhou, C., et al., Off-target RNA mutation induced by DNA base editing and its elimination by mutagenesis. Nature, 2019. 571(7764): p. 275-278.
  11. Grunewald, J., et al., CRISPR DNA base editors with reduced RNA off-target and self-editing activities. Nat Biotechnol, 2019. 37(9): p. 1041-1048.
  12. Doman, J.L., et al., Evaluation and minimization of Cas9-independent off-target DNA editing by cytosine base editors. Nat Biotechnol, 2020. 38(5): p. 620-628.
  13. Gaudelli, N.M., et al., Directed evolution of adenine base editors with increased activity and therapeutic application. Nat Biotechnol, 2020. 38(7): p. 892-900.


Yunbo Qiao

Professor, Guangzhou University