Skip to main content

Thank you for visiting You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Genome-wide target specificities of CRISPR RNA-guided programmable deaminases

An Erratum to this article was published on 01 August 2017

This article has been updated


Cas9-linked deaminases, also called base editors, enable targeted mutation of single nucleotides in eukaryotic genomes. However, their off-target activity is largely unknown. Here we modify digested-genome sequencing (Digenome-seq) to assess the specificity of a programmable deaminase composed of a Cas9 nickase (nCas9) and the deaminase APOBEC1 in the human genome. Genomic DNA is treated with the base editor and a mixture of DNA-modifying enzymes in vitro to produce DNA double-strand breaks (DSBs) at uracil-containing sites. Off-target sites are then computationally identified from whole genome sequencing data. Testing seven different single guide RNAs (sgRNAs), we find that the rAPOBEC1–nCas9 base editor is highly specific, inducing cytosine-to-uracil conversions at only 18 ± 9 sites in the human genome for each sgRNA. Digenome-seq is sensitive enough to capture off-target sites with a substitution frequency of 0.1%. Notably, off-target sites of the base editors are often different from those of Cas9 alone, calling for independent assessment of their genome-wide specificities.

Access options

Rent or Buy article

Get time limited or full article access on ReadCube.


All prices are NET prices.

Figure 1: Tolerance of BE3 and Cas9 for mismatched sgRNAs.
Figure 2: Digenome-seq to identify BE3 off-target sites in the human genome.
Figure 3: Genome-wide BE3 off-target sites revealed by Digenome-seq.
Figure 4: Reducing BE3 off-target effects via modified sgRNAs.

Accession codes

Primary accessions

Sequence Read Archive

Change history

  • 06 July 2017

    In the version of this article initially published, in the HTML only, Daesik Kim should have been the second corresponding author rather than Seuk-Min Ryu. In Figure 4b, in all versions, the bar graphs were misaligned with the specificity ratios, so that the first row of bar graphs were above the specificity ratios, rather than aligned with 3.5, 1.0, etc. The errors have been corrected in the HTML and PDF versions of the article.


  1. 1

    Komor, A.C., Kim, Y.B., Packer, M.S., Zuris, J.A. & Liu, D.R. Programmable editing of a target base in genomic DNA without double-stranded DNA cleavage. Nature 533, 420–424 (2016).

    CAS  Article  Google Scholar 

  2. 2

    Nishida, K. et al. Targeted nucleotide editing using hybrid prokaryotic and vertebrate adaptive immune systems. Science 353, 1248–1257 (2016).

    CAS  Article  Google Scholar 

  3. 3

    Ma, Y. et al. Targeted AID-mediated mutagenesis (TAM) enables efficient genomic diversification in mammalian cells. Nat. Methods 13, 1029–1035 (2016).

    CAS  Article  Google Scholar 

  4. 4

    Hess, G.T. et al. Directed evolution using dCas9-targeted somatic hypermutation in mammalian cells. Nat. Methods 13, 1036–1042 (2016).

    CAS  Article  Google Scholar 

  5. 5

    Yang, L. et al. Engineering and optimising deaminase fusions for genome editing. Nat. Commun. 7, 13330 (2016).

    CAS  Article  Google Scholar 

  6. 6

    Kim, H. & Kim, J.S. A guide to genome engineering with programmable nucleases. Nat. Rev. Genet. 15, 321–334 (2014).

    CAS  Article  Google Scholar 

  7. 7

    Tsai, S.Q. et al. GUIDE-seq enables genome-wide profiling of off-target cleavage by CRISPR-Cas nucleases. Nat. Biotechnol. 33, 187–197 (2015).

    CAS  Article  Google Scholar 

  8. 8

    Frock, R.L. et al. Genome-wide detection of DNA double-stranded breaks induced by engineered nucleases. Nat. Biotechnol. 33, 179–186 (2015).

    CAS  Article  Google Scholar 

  9. 9

    Ran, F.A. et al. In vivo genome editing using Staphylococcus aureus Cas9. Nature 520, 186–191 (2015).

    CAS  Article  Google Scholar 

  10. 10

    Crosetto, N. et al. Nucleotide-resolution DNA double-strand break mapping by next-generation sequencing. Nat. Methods 10, 361–365 (2013).

    CAS  Article  Google Scholar 

  11. 11

    Wang, X. et al. Unbiased detection of off-target cleavage by CRISPR-Cas9 and TALENs using integrase-defective lentiviral vectors. Nat. Biotechnol. 33, 175–178 (2015).

    CAS  Article  Google Scholar 

  12. 12

    Gabriel, R. et al. An unbiased genome-wide analysis of zinc-finger nuclease specificity. Nat. Biotechnol. 29, 816–823 (2011).

    CAS  Article  Google Scholar 

  13. 13

    Kim, D. et al. Digenome-seq: genome-wide profiling of CRISPR-Cas9 off-target effects in human cells. Nat. Methods 12, 237–243 (2015).

    CAS  Article  Google Scholar 

  14. 14

    Kim, D., Kim, S., Kim, S., Park, J. & Kim, J.S. Genome-wide target specificities of CRISPR-Cas9 nucleases revealed by multiplex Digenome-seq. Genome Res. 26, 406–415 (2016).

    CAS  Article  Google Scholar 

  15. 15

    Kim, D. et al. Genome-wide analysis reveals specificities of Cpf1 endonucleases in human cells. Nat. Biotechnol. 34, 863–868 (2016).

    CAS  Article  Google Scholar 

  16. 16

    Lin, Y. et al. CRISPR/Cas9 systems have off-target activity with insertions or deletions between target DNA and guide RNA sequences. Nucleic Acids Res. 42, 7473–7485 (2014).

    CAS  Article  Google Scholar 

  17. 17

    Bae, S., Park, J. & Kim, J.S. Cas-OFFinder: a fast and versatile algorithm that searches for potential off-target sites of Cas9 RNA-guided endonucleases. Bioinformatics 30, 1473–1475 (2014).

    CAS  Article  Google Scholar 

  18. 18

    Kyoungmi Kim, K. et al. Highly efficient RNA-guided base editing in mouse embryos. Nat. Biotechnol. (2017).

  19. 19

    Fu, Y., Sander, J.D., Reyon, D., Cascio, V.M. & Joung, J.K. Improving CRISPR-Cas nuclease specificity using truncated guide RNAs. Nat. Biotechnol. 32, 279–284 (2014).

    CAS  Article  Google Scholar 

  20. 20

    Cho, S.W. et al. Analysis of off-target effects of CRISPR/Cas-derived RNA-guided endonucleases and nickases. Genome Res. 24, 132–141 (2014).

    CAS  Article  Google Scholar 

  21. 21

    Slaymaker, I.M. et al. Rationally engineered Cas9 nucleases with improved specificity. Science 351, 84–88 (2016).

    CAS  Article  Google Scholar 

  22. 22

    Kleinstiver, B.P. et al. High-fidelity CRISPR-Cas9 nucleases with no detectable genome-wide off-target effects. Nature 529, 490–495 (2016).

    CAS  Article  Google Scholar 

  23. 23

    Cho, S.W., Lee, J., Carroll, D., Kim, J.S. & Lee, J. Heritable gene knockout in Caenorhabditis elegans by direct injection of Cas9-sgRNA ribonucleoproteins. Genetics 195, 1177–1180 (2013).

    CAS  Article  Google Scholar 

  24. 24

    Kim, S., Kim, D., Cho, S.W., Kim, J. & Kim, J.S. Highly efficient RNA-guided genome editing in human cells via delivery of purified Cas9 ribonucleoproteins. Genome Res. 24, 1012–1019 (2014).

    CAS  Article  Google Scholar 

Download references


This research was supported by grants from the Institute for Basic Science (IBS-R021-D1) to J.-S.K. and ToolGen, Inc. (0409-20160107) to D.K. The plasmid encoding the His6-rAPOBEC1-XTEN-dCas9 protein (pET28b-BE1) was a gift from David Liu.

Author information




J.-S.K. and D.K. supervised the research. D.K., K.L., S.Y., K.K., and S.-M.R. performed the experiments. D.K., K.L., and S.-T.K. carried out bioinformatics analyses.

Corresponding authors

Correspondence to Daesik Kim or Jin-Soo Kim.

Ethics declarations

Competing interests

J.-S.K. is a founder of and shareholder in ToolGen.

Integrated supplementary information

Supplementary Figure 1 Comparison of BE3-associateded base editing efficiencies and Cas9-associated indel frequencies in human cells

(a) Base editing efficiencies obtained with BE1 (rAPOBEC1–dCas9), BE2 (rAPOBEC1–dCas9–UGI), and BE3 (rAPOBEC1–nCas9–UGI) at seven endogenous target sites in HEK293T cells. Base editing efficiencies were measured by targeted deep sequencing. Error bars indicate s.e.m. (b) Cas9 nuclease-driven mutation frequencies were measured by targeted deep sequencing at seven endogenous target sites in HEK293T cells. (c) A table showing target DNA sequences and mutation frequencies. The PAM is shown in blue. (d) A graph showing the rank order of indel frequencies or base editing efficiencies at seven endogenous target sites.

Supplementary Figure 2 Tolerance of BE3 and Cas9 for mismatched sgRNAs.

Specificities of BE3 and Cas9 examined using mismatched sgRNAs at the RNF2 site. Base editing efficiencies and indel frequencies obtained with mismatched sgRNAs were measured by targeted deep sequencing. The PAM is shown in blue. Red or black asterisks indicate mismatched sgRNAs that were highly active with BE3 but poorly active with Cas9 or vice versa, respectively. Error bars indicate s.e.m. (n = 3).

Supplementary Figure 3 Correlation between indel frequencies associated with Cas9 nucleases and base editing frequencies associated with BE3 using mismatched sgRNAs at the EMX1 (a), HBB (b), and RNF2 (c) sites.

The red dots indicate mismatched sgRNAs with which the relative frequency of BE3-associated base editing was more than three times higher than the relative frequency of Cas9 nuclease-associated indels and the blue dots indicate sgRNAs with which the relative frequency of Cas9 nuclease-associated indels was more than three times higher than the relative frequency of BE3-associated base editing.

Supplementary Figure 4 SDS-Polyacrylamide gel showing the integrity of Cas9 and BE3ΔUGI proteins before and after incubation with genomic DNA.

Supplementary Figure 5 Quantitative real-time PCR showing almost complete cleavage of genomic DNA at the on-target site by Cas9 or BE3ΔUGI plus USER.

Supplementary Figure 6 IGV images showing straight alignments of sequence reads at the 6 different on-target sites.

Supplementary Figure 7 In vitro DNA cleavage scoring system for Digenome-seq analysis of BE3ΔUGI

Supplementary Figure 8 Venn diagrams showing the number of sites with DNA cleavage scores over 2.5 identified by Digenome-seq of Cas9 nuclease- and BE3ΔUGI-treated genomic DNA.

Supplementary Figure 9 The number of total sites (red) and the number of PAM-containing sites with ten or fewer mismatches (blue) for a range of DNA cleavage scores.

Intact human genomic DNA (left) and genomic DNA digested by BE3ΔUGI and USER (right) were subjected to whole genome sequencing.

Supplementary Figure 10 Venn diagrams showing the number of PAM-containing homologous sites with DNA cleavage scores over 0.1 identified by Digenome-seq of Cas9- and BE3ΔUGI-treated genomic DNA.

Supplementary Figure 11 Fraction of homologous sites captured by Digenome-seq.

Blue bars represent the number of homologous sites that differ from on-target sites by up to 6 nt. Red squares (BE3ΔUGI) and green triangles (Cas9) represent the fraction of Digenome-identified sites for a range of mismatch numbers.

Supplementary Figure 12 Significant correlation between the number of BE3ΔUGI- and Cas9-associated sites identified by Digenome 1.0 (a) and Digenome 2.0 (b).

Supplementary Figure 13 Significant correlation between the number of BE3ΔUGI-associated sites identified by Digenome 1.0 (a) or Digenome 2.0 (b) and the number of sites with 6 or fewer mismatches.

Supplementary Figure 14 Examples of Digenome-captured off-target sites associated only with Cas9, which contain no cytosines at positions 4-8.

Supplementary Figure 15 Base editing efficiencies at Digenome-captured sites associated only with 3 different Cas9 nucleases.

No substitutions were detectably induced by BE3 at these Cas9-associated sites. On-target sequences (EXM1_On, HBB_On, and RNF2_ON) are also shown.

Supplementary Figure 16 Base editing efficiencies of 6 different BE3 deaminases at Digenome-negative sites with ≤ 3 mismatches with respective on-target sequences.

No substitutions were detectably induced by BE3 at these sites. On-target sequences are also shown.

Supplementary Figure 17 Digenome-seq to identify off-target site of BE3 in the mouse genome.

(a) IGV image showing straight alignments of sequence reads at the Dmd on-target site. (b) Three sites, including the on-target site, were identified by Digenome 2.0. (c) No off-target substitutions were detectably induced at the two candidate sites identified by Digenome-seq in NIH3T3 cells.

Supplementary Figure 18 Reducing BE3 off-target effects using modified sgRNAs.

(a) Sequences of sgRNAs at the 5’ terminus. (b) Base editing efficiencies were measured at the EMX1 on- and off-target sites by targeted deep sequencing in HEK293T cells. The heatmap represents relative specificities of modified sgRNAs, compared to that of gX19 sgRNA. The specificity ratio was calculated by dividing (on-target frequency of modified sgRNA/off-target frequency of modified sgRNA) by (on-target frequency of gX19 sgRNA/off-target frequency of gX19 sgRNA).

Supplementary information

Supplementary Text and Figures

Supplementary Figures 1–18 and Supplementary Tables 1–4. (PDF 6335 kb)

Supplementary Code

Digenome-toolkit2-hotfix. (ZIP 39 kb)

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Kim, D., Lim, K., Kim, ST. et al. Genome-wide target specificities of CRISPR RNA-guided programmable deaminases. Nat Biotechnol 35, 475–480 (2017).

Download citation

Further reading


Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing