Programmable Editing of a Target Base in Genomic DNA without Double-Stranded DNA Cleavage
Nature, 2016 · paper · Biomedicine
Develops base editing, which fuses CRISPR/Cas9 to a cytidine deaminase to convert one DNA base to another without double-strand breaks or donor templates. In human and mouse cell lines, later versions corrected about 15–75% of DNA with typically 1% or fewer indels.
7 findings
Substance p005: "We also monitored C→T mutations at 3,200 cytosines surrounding the six on-target and 44 off-target loci tested ... (Extended Data Fig. 7)", while the legend of that figure (p016) states "2,500 distinct cytosines and guanines surrounding the six on-target and 34 off-target loci tested" and again "all 2,500 cytosines/guanines".
- p. 5 Results, off-target paragraph
We also monitored C→T mutations at 3,200 cytosines surrounding the six on-target and 44 off-target loci tested
- p. 16 Extended Data Figure 7 legend
the C to T and G to A mutation rates at 2,500 distinct cytosines and guanines surrounding the six on-target and 34 off-target loci tested
- p. 16 Extended Data Figure 7a legend
plotted individually against their positions relative to a protospacer for all 2,500 cytosines/guanines
- p. 5 Results, off-target paragraph
We also monitored C→T mutations at 3,200 cytosines surrounding the six on-
Only the authors can say what is right here.
Judge's note: Read the Results sentence on the main p005 page image (crops C009-a, C009-d, C009-f, C009-h) and the Extended Data Figure 7 legend on the main p016 page image (crops C009-b, C009-c, C009-e, C009-g, C009-i). The text cites Extended Data Fig. 7 for exactly this analysis; the legend gives 2,500 twice, for the broader category (cytosines and guanines), so a cytosine-only count could not legitimately be the larger 3,200. The legend adds a total of 14,700,000 reads from approximately 1.8x10^6 cells; no sentence in the article or in the Methods (main p010, Data analysis) defines a second, larger set of surveyed positions. One of the two counts of the surveyed positions is wrong and only the authors can say which.
Second check: another model agreed
Report a problem with this finding
Substance p005: "the six on-target and 44 off-target loci tested" agrees neither with the same paragraph's enumeration ("the top 34 known Cas9 off-target sites in human cells15, and the top 12 known dCas9 off-target binding sites", i.e. 46) nor with the Extended Data Fig. 7 legend (p016), "the six on-target and 34 off-target loci tested".
- p. 5 Results, off-target paragraph
surrounding the six on-target and 44 off-target loci tested
- p. 16 Extended Data Figure 7 legend
surrounding the six on-target and 34 off-target loci tested
- p. 5 Results, off-target paragraph
we sequenced the top 34 known Cas9 off-target sites in human cells15, and the top 12 known dCas9 off-target binding sites
- p. 5 Results, off-target paragraph
Therefore we sequenced the top 34 known Cas9 off-target sites in human cells
Only the authors can say what is right here.
Judge's note: Read the whole off-target paragraph on the main p005 page image (crops C010-c, C010-d, C010-e, C010-i, C010-j, C010-k, C010-l) and the Extended Data Fig. 7 legend on main p016 (crops C010-b, C010-h). The paragraph states that 34 Cas9 off-target sites and 12 dCas9 off-target binding sites were sequenced (Supplementary Tables 1-5) and reports detectable editing at "16/34 for BE1 and BE2; and 17/34 for BE3" Cas9 off-target sites with none at the dCas9 sites. The printed 44 matches neither 34+12=46 nor the figure legend's 34, and no statement in the article (including the Methods on main p010) excludes two amplicons or otherwise defines a 44-locus set. Distinct from C009, which concerns the count of surveyed cytosines.
Second check: another model agreed
Report a problem with this finding
Substance Methods, Cell culture (p. 9) describes "Immortalized rat astrocytes containing the ApoE4 isoform of the APOE gene (Taconic Biosciences)", while the Results (p. 5), the Extended Data Fig. 8a legend (p. 17), the Table 1 legend (p. 25) and the Acknowledgments (p. 19) all call the same cells mouse astrocytes.
- p. 9 Methods, Cell culture
Immortalized rat astrocytes containing the ApoE4 isoform of the APOE gene (Taconic Biosciences)
- p. 5 Results, APOE paragraph
in immortalized mouse astrocytes in which the endogenous APOE gene was replaced by human APOE4
- p. 17 Extended Data Figure 8 legend, panel a
allele is converted to APOE3r in mouse astrocytes by BE3 in 58.3% of total reads
- p. 25 Table 1 legend, panel a
allele was converted to APOE3r in mouse astrocytes by BE3 in 74.9% of total reads
Only the authors can say what is right here.
Judge's note: Checked each location on the page images: crop C020-a (p. 9 Cell culture) prints "Immortalized rat astrocytes containing the ApoE4 isoform of the APOE gene (Taconic Biosciences)"; crop C020-b (p. 5) prints "in immortalized mouse astrocytes in which the endogenous APOE gene was replaced by human APOE4"; page image p017 shows the Extended Data Fig. 8a legend "converted to APOE3r in mouse astrocytes by BE3 in 58.3% of total reads"; page image p025 shows the Table 1a legend "converted to APOE3r in mouse astrocytes by BE3 in 74.9% of total reads"; crop C020-e (p. 19) prints "providing immortalized mouse astrocytes containing APOE4". Only one astrocyte culture is described in the Methods (one APOE4 genotype, one Geneticin selection, one nucleofection program T-020), and only one astrocyte experiment is reported, so the two species statements cannot both describe it. Which species is correct can only be settled by the authors; the Methods also attribute the cells to a commercial supplier while the Acknowledgments thank colleagues for providing them.
Second check: another model agreed
Report a problem with this finding
Substance The Extended Data Fig. 8 legend (p. 17) states for panel a "Identical treatment of these cells with wt Cas9 and donor ssDNA results in 0.2% correction, with 26.7% indel formation", but the indel % column of the Cas9 + HDR row in the panel itself reads 40.1.
- p. 17 Extended Data Figure 8a, "Cas9 + HDR" row, "indel %" column
40.1
- p. 17 Extended Data Figure 8 legend, panel a
Identical treatment of these cells with wt Cas9 and donor ssDNA results in 0.2% correction, with 26.7% indel formation.
- p. 6 Results, APOE paragraph
0.1-0.3% APOE4 correction and 26-40% indels at the targeted locus
- p. 17 Extended Data Figure 8a legend
wt Cas9 and donor ssDNA results in 0.2% correction, with 26.7% indel formation
Judge's note: Read the full page image main-p017.png: in panel a the four rows carry indel % values 0.0 (untreated), 6.1 (BE3 + on-target sgRNA), 0.0 (BE3 + off-target sgRNA) and 40.1 (Cas9 + HDR); the legend for the same panel prints 26.7%. The legend's correction value (0.2%) and its panel-b statement (8.0% indel, matching the 8.0 printed in panel b) are consistent, so only the panel-a indel value disagrees. The Table 1a legend (page image main-p025.png) gives 26.1% indels for the corresponding Cas9 + HDR astrocyte treatment, matching the 26.1 printed in Table 1a, and the Results on p. 6 (crop C023-c) give the range "26-40% indels", which spans 26.1 and 40.1 and not 26.7. A reader taking the legend at face value would carry away a wrong indel frequency.
Second check: another model agreed
Report a problem with this finding
Reference or label Introduction, p002: "Recent studies on the dCas9-target DNA complex reveal that at least nine nt of the displaced DNA strand are unpaired upon formation of the Cas9:guide RNA:DNA "R-loop" complex.10" is supported by reference 10, which the reference list gives as "Jore MM, et al. Structural basis for CRISPR RNA-guided DNA recognition by Cascade" (a type I Cascade complex, not Cas9/dCas9).
- p. 2 Introduction, final paragraph
Recent studies on the dCas9-target DNA complex reveal that at least nine nt of the displaced DNA strand are unpaired upon formation of the Cas9:guide RNA:DNA "R-loop" complex.10
- p. 19 Reference list, ref. 10
Jore MM, et al. Structural basis for CRISPR RNA-guided DNA recognition by Cascade.
- p. 2 Introduction, final paragraph
Recent studies on the dCas9-target DNA complex reveal that at least nine nt of the displaced DNA strand are unpaired upon formation of the Cas9:guide RNA:DNA “R-loop” complex.
Only the authors can say what is right here.
Judge's note: Read the sentence on the main p002 page image (crops C005-a, C005-c) and the reference list on main p019 (crop C005-b): entry 10 is Jore MM et al., Nature structural & molecular biology 2011;18:529-536, titled "Structural basis for CRISPR RNA-guided DNA recognition by Cascade". The next sentence in the text cites reference 11 for the Cas9 R-loop, and entry 11 is "Jiang F, et al. Structures of a CRISPR-Cas9 R-loop complex primed for DNA cleavage", so the Cas9 structural work is cited separately. The sentence carrying reference 10 names the dCas9-target DNA complex specifically, while the cited title is a different complex; the mismatch is visible entirely within the files (title in the reference list). The correct replacement reference cannot be inferred from the article, hence author_query.
Second check: another model agreed
Report a problem with this finding
Reference or label Reference list (p. 20) prints references 31 and 32 with volume and page range fused: "Nat Biotech. 2013; 31838:833." and "Nat Biotech. 2013; 31843:839."
- p. 20 Reference list, ref. 31
Nat Biotech. 2013; 31838:833.
- p. 20 Reference list, ref. 32
Nat Biotech. 2013; 31843:839.
Judge's note: Viewed crops C027-a and C027-b on the page image. Entries 31 (Mali et al.) and 32 (Pattanayak et al.) are typeset as "31838:833." and "31843:839.", whereas every neighbouring entry uses volume:first-last with an en dash retained in the image (ref. 29 "33:73-80", ref. 30 "523:481-485", ref. 33 "10:751-754"), so the dash has not been lost in extraction. As printed, the volume fields are not parsable citations; the reader cannot identify the pages. The internally evident reading is volume 31, pages 833-838 and 839-843.
Second check: another model agreed
Report a problem with this finding
Wording Concluding paragraph, p006: "or more efficient editing with ≤1% indel formation (BE3)", whereas the paper's own BE3 indel data exceed 1% (Extended Data Table 1: EMX1 1.34±0.35, FANCF 1.47±0.93, RNF2 1.39±0.72; Results p004 "averaging 1.1% for the six tested loci"; astrocytes "4.6-6.1% indels").
- p. 6 Discussion, first paragraph
or more efficient editing with ≤1% indel formation (BE3)
- p. 18 Extended Data Table 1, BE3 row
1.34±0.35
- p. 18 Extended Data Table 1, BE3 row
1.47±0.93
- p. 4 Results, BE3 paragraph
only a small frequency of indels, averaging 1.1% for the six tested loci, was observed from BE3 treatment
Only the authors can say what is right here.
Judge's note: Read the concluding sentence on the main p006 page image (crops C012-a, C012-g, C012-i), the BE3 row of Extended Data Table 1 on main p018 (crops C012-b, C012-c, C012-f: 1.34±0.35, 1.47±0.93, 0.62±0.35, 0.91±1.07, 0.95±1.64, 1.39±0.72), the six-locus average on main p004 (crops C012-d, C012-h), and the astrocyte indel range on main p006 (crop C012-e). I recomputed the mean of the six tabulated BE3 values as 1.11%, consistent with the printed "1.1%" but above the "≤1%" ceiling; three of six loci individually exceed 1%, and the astrocyte experiment gives 4.6-6.1%. The abstract hedges the same figure as "typically ≤1%", but the concluding sentence states it without qualification, so the text overstates the data; the underlying values are reported correctly, hence S3.
Second check: another model agreed
Report a problem with this finding
11 minor (typos and the like; most checked by the judge only)
Wording Abstract: "In four transformed human and murine cell lines, second- and third-generation base editors ... resulting in permanent correction of ∼15-75% of total cellular DNA" — the lower bound of the stated range is not reached in two of the four lines (HCC1954: "3.3-7.6%"; U2OS BE2 in Extended Data Fig. 5, all bars below ~10%).
- p. 1 Abstract
In four transformed human and murine cell lines, second- and third-generation base editors ... resulting in permanent correction of ∼15-75% of total cellular DNA
- p. 6 Results, p53 paragraph
We observed correction of the Tyr163Cys mutation in 3.3-7.6% of nucleofected HCC1954 cells
- p. 14 Extended Data Figure 5
Base editing efficiencies of BE2 in U2OS and HEK293T cells
- p. 6 Results, TP53 correction
correction of the Tyr163Cys mutation in 3.3-7.6% of nucleofected HCC1954 cells
Only the authors can say what is right here.
Judge's note: Read the abstract on main p001 (crop C001-a), the TP53/HCC1954 sentence on main p006 (crops C001-b, C001-d: "correction of the Tyr163Cys mutation in 3.3-7.6% of nucleofected HCC1954 cells"), and Extended Data Fig. 5 on main p014 (crop C001-c and full page): the U2OS BE2 bars peak at roughly 8-9% and the HEK293T BE2 bars at roughly 20%. The highest values in the paper are the astrocyte APOE4 results (74.9% in Table 1a, 58.3% in Extended Data Fig. 8a) and BE3 in HEK293T (up to 37%, p004). No sentence in the article restricts the "~15-75%" range to selected lines or loci, so as printed the abstract attaches a range to four named cell lines that two of them do not reach. Wording/overstatement only; the per-experiment values are reported correctly elsewhere, hence S3.
Second check: another model agreed
Report a problem with this finding
Wording Abstract, p001: "second- and third-generation base editors that fuse uracil glycosylase inhibitor (UGI), and that use a Cas9 nickase targeting the non-edited strand" ascribes nickase activity to the second-generation editor as well, although BE2 is "APOBEC–XTEN–dCas9–UGI" (catalytically dead Cas9) and only BE3 "nicks the non-edited strand".
- p. 1 Abstract
second- and third-generation base editors that fuse uracil glycosylase inhibitor (UGI), and that use a Cas9 nickase targeting the non-edited strand
- p. 4 Results, BE2 construction
second-generation base editor (BE2, APOBEC–XTEN–dCas9–UGI)
- p. 4 Results, BE3 construction
third-generation base editor (BE3, APOBEC–XTEN–dCas9(A840H)–UGI) that nicks the non-edited strand
Only the authors can say what is right here.
Judge's note: Compared the abstract sentence on main p001 (crop C004-a) with the Results on main p004 (crop C004-b and full page): BE2 is defined as "a second-generation base editor (BE2, APOBEC–XTEN–dCas9–UGI)", and the nicking activity is introduced only for BE3, "We therefore restored the catalytic His residue at position 840 in the Cas9 HNH domain of BE2 ... (BE3, APOBEC–XTEN–dCas9(A840H)–UGI) that nicks the non-edited strand". As printed, both relative clauses attach to both generations, so the abstract describes BE2 inaccurately. The body is correct, so this is a wording ambiguity in the abstract rather than a wrong datum.
Second check: another model agreed
Report a problem with this finding
Typo UGI is expanded two different ways: "uracil glycosylase inhibitor (UGI)" (abstract, p001) and "Uracil DNA glycosylase inhibitor (UGI)" (Results, p004).
- p. 1 Abstract
base editors that fuse uracil glycosylase inhibitor (UGI)
- p. 4 Results, BE2 paragraph
Uracil DNA glycosylase inhibitor (UGI), an 83-residue protein from B. subtilis
Judge's note: Checked both renderings on the page images (main p001 abstract, crop C002-a; main p004 second paragraph, crop C002-b). The two expansions are as quoted, and on p004 the inhibited enzyme is defined as "Uracil DNA glycosylase (UDG)" in the preceding sentence. Both expansions are in ordinary use, so this is a cosmetic inconsistency of abbreviation definition rather than a substantive error.
Second check: none (judge only)
Report a problem with this finding
Typo p006: "The dominant-negative p53 mutation Tyr163Cysis strongly associated with several types ofcancer" — two missing word spaces ("Tyr163Cys is", "of cancer").
- p. 6 Results, p53 paragraph, first sentence
The dominant-negative p53 mutation Tyr163Cysis strongly associated with several types ofcancer
- p. 6 p53 paragraph
Tyr163Cysis strongly associated with several types ofcancer
Judge's note: Checked the rendered page image, not only the text layer: on main p006 (crops C015-a, C015-b) the glyphs are set without spaces, "Tyr163Cysis" at the end of the first line of that paragraph and "ofcancer25" at the start of the next line. Not an extraction artefact; a typesetting slip with an evident fix.
Second check: none (judge only)
Report a problem with this finding
Typo Methods, Cloning (p. 7) uses the abbreviation "NBE plasmids were constructed using USER cloning (New England Biolabs)" although "NBE" is never defined or used anywhere else in the paper.
- p. 7 Methods, Cloning
NBE plasmids were constructed using USER cloning (New England Biolabs).
- p. 9 Methods, Transfections
750 ng of BE and 250 ng of sgRNA expression plasmids were transfected
Only the authors can say what is right here.
Judge's note: Read crop C016-a (p. 7 Cloning) and C016-b (p. 9 Transfections) on the page images: the Cloning paragraph prints "NBE plasmids", while the Transfections paragraph and the rest of the paper use "BE"/"BE1, BE2, BE3". A grep of the whole extracted text finds "NBE" exactly once (line 305), with no expansion anywhere, so the abbreviation is undefined. Minor/cosmetic; only the authors can say whether "nucleobase editor" or simply "BE" was intended.
Second check: none (judge only)
Report a problem with this finding
Typo The same reagent is spelled two ways: "VeraSeq ULtra DNA polymerase (Enzymatics)" (p. 7, Cloning) and "VeraSeq Ultra (Enzymatics)" (p. 9, Preparation of in vitro-edited dsDNA for HTS).
- p. 7 Methods, Cloning
PCR was performed using VeraSeq ULtra DNA polymerase (Enzymatics)
- p. 9 Methods, Preparation of in vitro-edited dsDNA for HTS
VeraSeq Ultra (Enzymatics)
- p. 7 Cloning
VeraSeq ULtra DNA polymerase
Judge's note: Viewed crop C017-a/C016-a (p. 7) where the typeset text clearly shows a mid-word capital L in "ULtra", and crops C017-b and C019-b (p. 9) where the same product from the same supplier is set as "Ultra". The capital is rendered in the PDF image, not an extraction artefact. Cosmetic inconsistency in a product name.
Second check: none (judge only)
Report a problem with this finding
Typo Methods, Deaminase assay on dsDNA (p. 8): "purified on aEconoSpin micro spin column (Epoch Life Science)" — the article and product name are run together.
- p. 8 Methods, Deaminase assay on dsDNA
purified on aEconoSpin micro spin column (Epoch Life Science)
- p. 9 Methods, Preparation of in vitro-edited dsDNA for HTS
purified on a EconoSpin micro spin column (Epoch Life Science)
- p. 8 Deaminase assay on dsDNA
purified on aEconoSpin
Judge's note: Viewed crop C019-a (p. 8) on the page image: the words are typeset joined as "aEconoSpin", with no space, so this is not a lost space in text extraction. Crop C019-b (p. 9) shows the parallel sentence correctly set as "purified on a EconoSpin micro spin column (Epoch Life Science)". Cosmetic.
Second check: none (judge only)
Report a problem with this finding
Typo Extended Data Fig. 2b (p. 12) labels the second-to-last gel lane "53 N239D +", missing the leading P, while the adjacent lane is labelled "P53 N239D –".
- p. 12 Extended Data Figure 2b, second-to-last lane label
53 N239D +
- p. 12 Extended Data Figure 2b, last lane label
P53 N239D –
Judge's note: Read the full page image main-p012.png. The rotated lane labels read, in order, ... "P53 Y236C –", "P53 Y236C +", "pos. control U4", "53 N239D +", "P53 N239D –". Every other p53 lane on the same gel carries the "P53" prefix, so the missing P in one label is a typesetting slip, not a style. Cosmetic.
Second check: none (judge only)
Report a problem with this finding
Typo Extended Data Table 1 title (p. 18): "Indel formation following treatment of HEK293T cells with BE1, BE2, BE3, or wt Cas9 + assDNA template for HDR" — "a ssDNA" is run together as "assDNA".
- p. 18 Extended Data Table 1 title
with BE1, BE2, BE3, or wt Cas9 + assDNA template for HDR
- p. 18 Extended Data Table 1 title
wt Cas9 + assDNA template for HDR
- p. 18 Extended Data Fig. 9 title
base editingg
- p. 18 Extended Data Table 1 title
Cas9 + assDNA template
Judge's note: Viewed crop C024-a on the page image: the bold title is typeset "wt Cas9 + assDNA template for HDR", while the descriptive paragraph immediately below reads "with wt Cas9 and a ssDNA template for HDR". The words are joined in the rendered PDF, so this is not a lost space in extraction. Cosmetic. (The "base editingg" typo also mentioned by one reviewer here is adjudicated under C025.)
Second check: none (judge only)
Report a problem with this finding
Typo Extended Data Figure 9 title (p. 18): "Genetic variants from ClinVar that in principle can be corrected by base editingg" — doubled final g.
- p. 18 Extended Data Figure 9 title
that in principle can be corrected by base editingg
- p. 18 Extended Data Figure 9 title
base editingg
Judge's note: Viewed crop C025-a on the page image: the bold figure title ends "base editingg" with a doubled g; the term is "base editing" throughout the article. Visible in the typeset title, so not an extraction duplication. Cosmetic.
Second check: none (judge only)
Report a problem with this finding
Typo Acknowledgments (p. 19) spell the same NIH fellowship two ways: "A.C.K. is a Ruth L. Kirchstein National Research Service Awards Postdoctoral Fellow" and, three sentences later, "J.A.Z. was a Ruth L. Kirschstein National Research Service Award Postdoctoral Fellow".
- p. 19 Acknowledgments
A.C.K. is a Ruth L. Kirchstein National Research Service Awards Postdoctoral Fellow
- p. 19 Acknowledgments
J.A.Z. was a Ruth L. Kirschstein National Research Service Award Postdoctoral Fellow
- p. 19 Acknowledgments
Ruth L. Kirchstein National Research Service Awards Postdoctoral Fellow
- p. 19
Ruth L. Kirchstein National Research Service Awards
Judge's note: Viewed crop C020-e / C026-a, which shows the whole Acknowledgments paragraph on the page image: the first mention reads "Ruth L. Kirchstein National Research Service Awards Postdoctoral Fellow (F32 GM 112366-2)" and the second "Ruth L. Kirschstein National Research Service Award Postdoctoral Fellow (F32 GM 106601-2)". Both are F32 NRSA fellowships, so the two spellings (and the Awards/Award mismatch) cannot both be intended. Cosmetic.
Second check: none (judge only)