GOTI, a method to identify genome-wide off-target effects of genome editing in mouse embryos

Abstract

Genome editing holds great potential for correcting pathogenic mutations. We developed a method named GOTI (genome-wide off-target analysis by two-cell embryo injection) to detect off-target mutations by editing one blastomere of two-cell mouse embryos using either CRISPR-Cas9 or base editors. GOTI directly compares edited and non-edited cells without the interference of genetic background and thus could detect potential off-target variants with high sensitivity. Importantly, the GOTI method was designed to detect potential off-target variants of any genome editing tools by the combination of experimental and computational approaches, which is critical for accurate evaluation of the safety of genome editing tools. Here we provide a detailed protocol for GOTI including mice mating, 2-cell embryo injection, E14.5 embryo digestion, FACS, whole genome sequencing and data analysis. To enhance the utility of GOTI, we also include a computational workflow named GOTI-seq (https://github.com/sydaileen/GOTI-seq) for the sequencing data analysis, which can generate the final genome-wide off-target variants from raw sequencing data directly. The protocol typically takes 20 days from the mice mating to sequencing and 7 days for sequencing data analysis.

EDITORIAL SUMMARY

Here the Authors describe GOTI (genome-wide off-target analysis by two-cell embryo injection), a method to detect off-target mutations of either CRISPR-Cas9-based genome editing, or base editors. This is achieved by editing one blastomere of a two-cell mouse embryo so that edited and unedited cells from the same genetic background can be compared.

Introduction

The recent development of genome editing tools holds great promise in diverse fields such as animal disease modeling, gene therapy, drug development, genetically modified plants, and biofuel technology¹. In addition, gene editing technology has accelerated the study of the functional organization of the genome and the causal links between genetic variations and biological phenotypes^2–5. A booming genetic editing technology within the life sciences field, CRISPR holds great hope for the treatment of genetic diseases^6,7. A majority of CRISPR-Cas9 edited products usually contain small indels at the target site due to the non-homologous end joining (NHEJ) in response to double strand breaks (DSBs)^8,9. Base editors, on the other hand, induce base pair substitutions using deaminases at the target loci without generating DSBs^11–14. There are two classes of DNA base editors. Cytosine base editors (CBEs) convert C•G base pairs to A•G base pairs¹¹ and adenine base editors (ABEs) conversely convert A•G base pairs to G•C base pairs¹³. However, the issue of off-target mutations, which may cause genetic instability and dysfunction, has been a major concern in the application of both these methods^15–19. Specifically, the potential for off-target effects remains a major barrier to the applications of genome editing for human gene therapy. Several techniques have been developed to detect genome-wide gene-editing off-target activity in cells, including selective enrichment and identification of adapter-tagged DNA ends by sequencing (SITE-Seq)²⁰, high-throughput genome-wide translocation sequencing (HTGTS)²¹, genome-wide, unbiased identification of double-strand breaks (GUIDE-seq)²² and circularization for in vitro reporting of cleavage effects by sequencing (CIRCLE-seq)²³. However, these approaches can only detect off-targets from double-strand breaks (DSBs) generated by genome editing, thus are not applicable to detect single-nucleotide variants (SNVs) in vivo. Given the rapidly increased application of base editors, it has become desirable to develop technologies for the comprehensive evaluation of off-target effects in an unbiased manner.

Recently, we have developed a genome-wide off-target effects detection method, GOTI, for the de novo identification of off-target mutations in mouse embryos^24,25. This technique examines off-target effects in a cell population derived from a single gene-edited blastomere, whereas previous studies used large pools of cells within which gene editing outcomes are variable, resulting in loss of signal for random off-target effects due to population averaging. In addition, as the edited and non-edited cells are from one single-ancestor cell, GOTI can minimize any confounding effects of genetic background and somatic mutations. GOTI could thus be generalized to be used for genome editing tools that do not introduce DSBs, including triplex-forming oligonucleotides, base editors, and potentially any other strategy that produces indels or SNVs.

In this article, we systemically describe the experimental procedures of GOTI and bioinformatic pipeline GOTI-seq (https://github.com/sydaileen/GOTI-seq) in detail. We also emphasize the specific steps and small tricks that required great care to ensure the data quality generated. The GOTI method could be broadly applied to evaluate the specificity of genome editing tools in animal models.

Overview of GOTI

The overall workflow of GOTI is illustrated in Fig. 1. A mixture of Cre, gene editing mRNA (Cas9/BE3/ABE7.10) and sgRNA is injected into one blastomere of a 2-cell mouse embryo, derived from Ai9 male mice mating with wild-type female mice²⁶. The action of Cre, injected into only one of the 2 cells of the embryo, is expected to generate a chimeric embryo with half of the cells labeled with tdTomato (colored red) with the Cre-LoxP system^26,27 so that the cells that received editing reagents cells can be distinguished from those that did not. When the chimeric embryo reaches E14.5, it is minced into small pieces and digested into a single cell suspension (Fig. 2). The tdTomato⁺ cells and tdTomato⁻ cells are then separated by fluorescence-activated cell sorting (FACS). Next, the two populations of cells are independently processed for high-throughput whole genome sequencing (WGS). Typically, the process from the 2-cell injection to the completion of the final sequencing takes 3 weeks for an experienced operator (Fig. 1a).

Fig. 1. Schematics of GOTI method.

a, Overall workflow of GOTI. b, Flowchart of GOTI-seq data processing pipeline. File formats and software used for each step are denoted in parentheses.

Fig. 2. Isolation of embryonic cells.

Pictures of key steps 59–61. a, The embryos with the placenta. b, Dissect mouse embryos and wash the embryos with PBS twice briefly. c, Add 5 mL of pre-warmed 37°C Trypsin-EDTA (0.05%) and transfer to a 50mL centrifuge tube.

The WGS reads are subsequently processed by a customized pipeline illustrated in Fig. 1b. Raw sequencing reads are first quality checked and the adapters are further trimmed out. Qualified reads are then mapped to the reference genome. The mapped alignment (BAM format) files are sorted and duplicates marked before downstream analysis. The off-target SNVs and indels are identified by comparing the tdTomato⁺ cells with tdTomato⁻ cells using three variant calling algorithms as indicated (Fig. 1b). Several previous studies have reported the discordance among variant calling methods due to different mathematical models and algorithms^28–31. Mutect2 applies a Bayesian classifier to detect somatic mutations with even low allele fractions³². Strelka2 detects both SNVs and indels by a mixture-model-based estimation with high speed³³. Lofreq is an ultra-sensitive variant caller to call somatic variants³⁴. Meanwhile, Scalpel performs well for the discovery of insertions and deletions³⁵. Different algorithms apply different models and have their own preferences to variants, so to reduce the false positives, we considered the overlap of three algorithms of SNVs or indels as the true variants and annotated these for functional analysis.

Applications of GOTI

We have demonstrated that GOTI is applicable to detect the off-target effects of various genome editing tools including CRISPR/Cas9, BE3 and ABE7.10. We have found numerous de novo SNVs induced by BE3, while CRISPR/Cas9 and ABE7.10 generated no unwanted mutations for the sgRNAs that were examined. Notably, the specificity of CRISPR/Cas9 depends heavily on the choice of sgRNA, so we cannot rule out the possibilities that some other sgRNAs may have significant off-target effects even though we have examined several. GOTI is a valuable method that is expected to help improve the current gene editors and facilitate the generation of new genome editing tools with higher specificity. This method could be applied in the preclinical evaluation of genome editing reagents before their approval for clinical trials in gene therapy. GOTI depends on Cre/LoxP recombination system to distinguish edited cells from unedited cells, so it could theoretically be used in other transgenic animal models carrying the reporting system like rat^36,37 or pig^38,39.

Alternative methods and advantages of GOTI

There have been several genome-wide methods previously developed for the detection of off-target effects. IDGV characterizes the genome-wide specificity by capturing DSBs in vivo⁴⁰, HTGTS detects genome-wide translocations generated from DSBs in vitro²¹, GUIDE-seq relies on the capture of double-stranded oligodeoxynucleotides integrated into the DSBs in vivo²². EndoV-seq⁴¹ and Digenome-seq^42,43 are both in vitro assays to investigate the sgRNA-dependent off-target effects of genome editing tools. CIRCLE-seq²³ examines the in vitro DSBs of naked DNA for the prediction of off-target sites. In addition, DISCOVER-seq tracks the recruitment of MRE11 to DSBs in vivo⁴⁴. These methods mostly rely on the generation of DSBs or nicks on the genome caused by the cleavage of Cas9, but base editors do not generate DSBs because they use catalytically deficient Cas9. Anderson K.R. et al.⁴⁵ and Lyer V et al.⁴⁶ also applied deep sequencing for the detection of sgRNA-dependent off-target effects in genomic specific regions based on sequence similarity. The sgRNA-dependent off-target effects could be solved by using different sgRNAs. However, none of these methods were applicable to detect random off-target effects which are independent of sgRNA or Cas9. Several previous studies also applied WGS to detect potential off-target effects of CRIPSR/Cas9 comparing edited and non-edited animals^44–50. However, the true off-target variants could not be distinguished from the single nucleotide polymorphisms (SNPs) by the genome comparison of two different individuals. One of the major advantages of GOTI is to directly compare edited and non-edited cells with identical genetic backgrounds. Therefore, potentially random off-target variants could be detected without bias even when they occur at low frequency.

The other main advantage of GOTI is its single-cell nature. The approach detects the off-target effects generated in a single ancestor cell and inherited by all descendant cells, which can be easily detected by sequencing, while previous work has been based on large pools of cells in which random off-target activity would be lost in the population average.

Limitations of GOTI

The embryo has to develop to E14.5 when the whole embryo could be readily digested to obtain enough single cells, so the duration of whole experiment takes up to one month. Also, GOTI relies on mice as a model system and requires regulatory approval with respect to animal welfare. However, GOTI can be theoretically conducted in other animal models with Cre/LoxP recombination system integrated. Another limitation is that GOTI is specific to the species in which it is performed. There are major ethical and legal considerations that prevent its use to assess off-target effects in the context of the human genome. In addition, the accomplishment of some procedures such as two-cell embryo injection requires microinjection apparatus, professional training and technical skill. GOTI is much more expensive compared to other methods based on the enrichment of different off-target sites, which could be used to detect sgRNA-dependent off-target effects.

Experimental design

The generation of GOTI reagents (Steps 1–13 and Box 1)

GOTI reagents consist of single-guide RNA (sgRNA), gene editor (Cas9 or base editor) mRNA and Cre mRNA. These components are generally transcribed in vitro from a T7 bacteriophage promoter (Supplementary Fig. 1). For sgRNA, a DNA template that contains T7 promotor, the designed sequence of sgRNA and the sgRNA scaffold is generated (as described in Steps 11–13) by performing a PCR reaction with the scaffold template (px330) and PCR primers (see Table 1). The forward primer must contain a T7 promoter sequence (20 bp), a sgRNA target sequence (20 bp) and a scaffold template-specific sequence (19 bp). The T7 promoter sequence and the scaffold template-specific sequence are fixed. The reverse primer is a short complementary sequence only targeting the end of the gRNA scaffold in the reverse direction. Then an sgRNA containing the sgRNA target sequence is created by in vitro transcription of the DNA template with the MEGA shortscript T7 kit. For gene editor mRNA, a DNA template that contains the T7 promotor, gene editor coding sequence and polyA sequence is generated (as described in Steps 1–8 for Cas9 or Box 1 for BE mRNA) by performing a PCR reaction with plasmid template and PCR primers (see Table 1). The forward primer includes the T7 promoter and 20–30 bp upstream sequence of start codon (ATG). The Reverse primer is located downstream of the polyA sequence. The mRNA encoding the gene editor is created by in vitro transcription of the DNA template with the mMESSAGE mMACHINE T7 ULTRA kit. For Cre mRNA, a DNA template that contains the T7 promotor, Cre coding sequence and polyA sequence is generated (as described in Steps 9–10) by performing a PCR reaction with pCAG-Cre plasmid template and PCR primers (see Table 1). The forward primer includes the T7 promoter and a sequence upstream of the start codon (ATG). Reverse primers are located downstream of the polyA sequence. The mRNA encoding Cre is created by in vitro transcription of the DNA template with the mMESSAGE mMACHINE T7 ULTRA kit.

Table 1.

Primers for in vitro transcription and genotyping

Primer	Sequence (5′−3′)	Purpose	Step
Cas9 IVT F	TAATACGACTCACTATAGGGAGATT TCAGGTTGGACCGGTG	Generation of Cas9 mRNA	1
Cas9 IVT R	GACGTCAGCGTTCGAATTGC	Generation of Cas9 mRNA	1
BE3 IVT F	TCCGCGGCCGCTAATACGACT	Generation of BE3 mRNA	1
BE3 IVT R	TGGTTCTTTCCGCCTCAGAAGCC	Generation of BE3 mRNA	1
Cre IVT F	TAATACGACTCACTATAGGGAGACA GATCACCTTTCCTATCAACC	Generation of Cre mRNA	2
Cre IVT R	TCGGTATTTCCAGCACACTGGA	Generation of Cre mRNA	2
sgRNA IVT F	TAATACGACTCACTATAGGGNNNNN NNNNNNNNNNNNNNNGTTTTAGAG CTAGAAATAG	Generation of sgRNA	9
sgRNA IVT R	AAAAGCACCGACTCGGTGCC	Generation of sgRNA	9
Tyr-OF	GTTATCCTCACACTACTTCTG	On-target genotyping	44
Tyr-OR	GTAATCCTACCAAGAGTCTCA	On-target genotyping	44
Tyr-IF	TCCTCACACTACTTCTGATG	On-target genotyping	44
Tyr-IR	GTCTCAAGATGGAAGATCAC	On-target genotyping	44

Validation of on-target editing efficiency in blastocyst (Steps 44–57)

To evaluate the efficiency of on-target editing in the GOTI system (to ensure that the injected embryos were edited by genome editing tools, as if not, downstream experiments would not be performed) in addition to transferring 80% of the injected embryos to the pseudopregnant mother, we culture the remaining 20% to the blastocyst stage in vitro (Step 34). The single blastocyst is then lysed and PCR amplification products of the targeting sites are ligated into T-Vectors. The ligation mixture is transformed into E. coli DH5α and cloned into E. coli (TA clone). Thirty TA clones are picked for Sanger sequencing to estimate the on-target editing efficiency. If at least 6 TA clones (20%) show on-target editing, further in vivo experiments (Steps 35–43) are performed.

FACS gating strategy for isolation of embryonic cells (Steps 58–67)

We describe a step-by-step protocol to digest embryos and isolate embryonic cells by flow cytometry. The procedure is simple and a cell suspension can be prepared from a single mouse embryo using enzymatic digestion and mechanical disaggregation in less than 35 minutes. To maximize cell viability, FACS sorting must be performed as soon as possible after preparation of the sample. We aim to finish sorting within 4 h after digesting embryos. We show the FACS gating strategy for isolation of embryo cells in Supplementary Fig. 2.

Sorting accuracy validation (Step 68)

After FACS, DNA of tdTomato⁺ and tdTomato⁻ cells are extracted separately and PCR amplified. These PCR products are ligated into T-Vectors, and 20 TA clones are picked for Sanger sequencing. Further experiments (Steps 69–76) will be conducted only if less than 10% (2/20) of clones show on-target editing in tdTomato⁻ cells and more than 20% (4/20) of clones show on-target editing in tdTomato⁺ cells. The sensitivity and specificity are used to validate the sorting accuracy of FACS. Some tdTomato+ cells may not be edited and tdTomato- cells might be edited as tdTomato+ and tdTomato- cells may be mis-sorted sometimes. So in order to maximize the sorting accuracy, we set sensitivity and specificity thresholds here to guarantee that tdTomato+ and tdTomato- cells for downstream analyses are well separated. Sensitivity is indicated by the percentage of tdTomato⁺ cells edited, and specificity is represented by the percentage of tdTomato⁻ cells edited (which did not receive editing reagents at the 1-cell stage).

Controls

Two kinds of controls are used in GOTI. The tdTomato⁻ cells are sequenced and analyzed together with tdTomato⁺ cells to eliminate the influence of genetic background. Plus, a control group with only the injection of Cre (Cre-only group) is necessary for the experiment to control the background noise in GOTI. In addition, the survival rate of embryos to blastocysts in the Cre-only group (usually more than 90%) acts as a control for evaluating the toxicity of genome editing tools. Genome editing tools with less than 80% blastocyst rate are potentially toxic to embryos as the normal blastocyst rates are above 80%¹⁶. Groups with different sgRNAs and no sgRNAs are optional to control for the sgRNA-dependent and independent off-target effects.

Materials

BIOLOGICAL MATERIALS

• C57BL/6 female mice; Shanghai Laboratory Animal Center (SLAC) laboratory Animal Co., Ltd (Shanghai SLAC Laboratory)

• Ai9 male mice²⁶ (full name B6.Cg-Gt (ROSA) 26Sortm9 (CAG-td-Tomato) Hze/J), JAX strain 007909; The Jackson Laboratory.

• ICR female mice; Shanghai SLAC Laboratory.

• Vasectomized male mice (ICR strain mice); Shanghai SLAC Laboratory. !CAUTION Experimental procedures involving animals must be carried out according to all relevant institutional and governmental regulations.

• DH5α competent cells (Tiangen, cat.no CB101)

Reagents

• KOD-Plus-Neo (toyobo cat. no. KOD-401)

• Premix Taq^™ (Ex Taq^™ Version 2.0 plus dye (Takara,cat. no.RR902A)

• DL10,000 DNA Marker (Takara, cat.no 3584A)

• DL15,000 DNA Marker (Takara, cat.no 3582B)

• 6 × Loading Buffer (Takara, cat.no 9156)

• EDTA (Sangon Biotech, cat. no. A500838-0500)

• Tris (Sangon Biotech, cat.no A100826-0500)

• Glacial acetic acid (Sangon Biotech, cat.no A501931-0500)

• PCR DNA primers oligomers (Shanghai HuaGen Biotech Co., Ltd.)

• mMESSAGE mMACHINE T7 ULTRA kit (Life Technologies, cat.no AM1345)

• MEGA shortscript T7 kit (Life Technologies, cat.no AM1354)

• MEGA clear kit (Life Technologies, cat.no AM1908)

• HEPES-CZB medium (see REAGENT SETUP)

• Pregnant mare serum gonadotropin (PMSG; Sigma, cat. no. G4527)

• Human chorionic gonadotropin (hCG; Sigma, cat. no. C8554)

• KSOM medium (Millipore, cat. no. MR-106-D) ! CAUTION Store the medium at −20°C.

• After thawing, keep it at 4°C and use it within 2 weeks.

• Hyaluronidase (Sigma, cat. no. H-3884)

• M2 medium (Millipore, cat. no. MR-015-D)

• Cytochalasin B (Sigma, cat. no. C6762)

• DMSO (Sigma, cat. no. D2650) ! CAUTION It is flammable; harmful if swallowed; toxic when in contact with skin and eye; and use protective gloves and safety glasses when handling.

• Mineral oil (Sigma, cat. no. M8410)

• Nuclease-free water (Life Technologies, cat. no. AM9932)

• Trypsin-EDTA (0.05%) (Gibco, cat. no. 25300062)

• Dulbecco’s Modified Eagle Medium (DMEM) (GIBCO, cat. no.11965092)

• Fetal Bovine Serum (FBS) (GIBCO, cat. no. 16000044)

• Tween 20 (Sangon Biotech, cat.no A100777-0500)

• Triton X-100 (Sangon Biotech, cat.no A110694-0500)

• Proteinase K (Tiangen, cat.no RT403-02)

• DNeasy blood and tissue kit (Qiagen; cat.no 69504)

• Universal DNA Purification Kit (TIANGEN, cat.no DP214)

• TIANprep Rapid Mini Plasmid Kit (TIANGEN, cat.no DP105)

• px330 (Addgene plasmid #42230)

• px260 (Addgene plasmid #42229)

• pCMV-BE3 (Addgene plasmid #73021)

• pCAG-Cre (Yang lab #ZP156)

• FastDigest BbsI (Thermo Scientific, cat.no FD1014)

• FastDigest PvuI (Thermo Scientific, cat.no FD0624)

• FastDigest NotI (Thermo Scientific, cat.no FD0595)

• Nucleic Acid Dye (TIANGEN, cat.no RT210)

• NaCl (Sigma, cat.no S-5886)

• KCl (Sigma, cat.no P-5405)

• CaCl₂ 2H₂O (Sigma, cat.no C-7902)

• MgSO₄ 7H₂O (Sigma, cat.no M-5921)

• KH₂PO₄ (Sigma, cat.no P-5655)

• EDTA·2Na·2H₂O (Sigma, cat.no E-4884)

• NaHCO₃ (Sigma, cat.no S-5761)

• L-glutamine (Sigma, cat.no G-8540)

• Na-lactate (Sigma, cat.no L7900)

• Sodium Pyouvate (Sigma, cat.no P-8574)

• Sodium Penicillin (Sigma, cat.no P-3032)

• Streptomycin (Sigma, cat.no S1277)

• BSA (Sigma, cat.no A-3311)

• HEPES (Sigma, cat.no H-7006)

• H₂O (Merck Millipore, cat.no TMS-006-B)

• 2,2,2-tribromoethanol (Sigma, cat.no T48402)

• 2-methyl-2-butanol (Sigma, cat.no 240486)

REAGENT SETUP

Ampicillin

Prepare a stock solution of 100 mg/ml in H₂O and filter-sterilize. Store at −20°C for up to 12 months.

Cell lysis buffer

Component	Volume	Final Concentration
Tween (1% (wt/vol) Tween in ddH2O)	1ml	0.1% (wt/vol)
Triton X-100 (1% (wt/vol) Triton X-100 in ddH2O)	1ml	0.1% (wt/vol)
Proteinase K (20 mg/ml)	2ml	4 mg/ml)
ddH2O	6ml
Total	10ml

Store at 4°C for up to 3 months.

LB medium

Add 10g bacto-tryptone, 5g yeast extract and 5g NaCl to 1L ddH₂O and sterilize by autoclaving. Store at 4°C for up to 1 month.

LB agar plates

Add 12 g agar to 1l LB medium before autoclaving. To prepare plates, allow medium to cool to 50°C and then add antibiotic stock to achieve a final concentration of 100mg/L, mix by gentle swirling and pour or pipette approximately 10 ml into each sterile Petri dish (60 mm diameter). Note that plates should be covered to prevent evaporation and stored agar side up at 4°C for up to 1 month until use.

50 × TAE buffer

242 g of Tris base, 18.61 g of EDTA and 57.1 ml of glacial acetic acid (100%) per liter of water (final pH 8.5). The stock can be stored at room temperature (25–30 °C) for up to 3 months.

HEPES-CZB medium

(1 liter) Store at 4°C for up to 2 month.

Component	Final concentration (g/L)	Final Concentration (mM)
NaCl	4.76	81.62
KCl	0.36	4.83
CaCl2 2H2O	0.25	1.70
MgSO4 7H2O	0.29	1.18
KH2PO4	0.16	1.18
EDTA·2Na2·H2O	0.04	0.11
NaHCO3	0.42	5.00
L-glutamine	0.15	1.00
Na-lactate	0.35	31.30
Sodium Pyouvate	0.03	0.27
Sodium Penicillin	3.56	100.00
Streptomycin	0.51	0.70
BSA	5	-
HEPES	0.125	0.48
H2O	1	-

Hyaluronidase (Hy), 10×stock, 10 mg/ml

Add 100 mg of Hy to 10 ml of M2 medium, and then divide it into 100 × 100μl tubes. Store at −20°C for up to 12 months.

M2 + Hy, 10 mg/ml

Add 100μl of Hy stock solution to 900μl M2 medium. Mix it just before use.

Cytochalasin B (CB), 50×stock, 500μg/ml

Add 2 mg of CB to 4 ml DMSO, and then divide it into 100 × 20μl tubes. Store at −20°C for up to 12 months.

HEPES-CZB + CB, 10μg/ml

Add 20μl of CB stock solution to 1ml HEPES-CZB. Mix it just before use.

Avertin A stock solution

Prepareby dissolving 10g of 2,2,2-tribromoethanol in 10 ml of 2-methyl-2-butanol in 50 °C water bath until it is fully dissolved. Prepare a working solution of 2.5% avertin by mixing 2.5ml stock solution with 97.5 ml PBS. Sterilize by passing solution through a 0.22 mm bottle top vacuum filter. Store the solution at 4 °C in the dark for up to 3 months.

DMEM medium

Add 50 ml of FBS to 450 ml of DMEM. Store at 4 °C for up to 2 months.

Equipment

• Cell strainer (Falcon, 40 μm Cell Strainer, cat.no 352340)

• Polystyrene round-bottom tube (5 ml, with 35-μm cell strainer cap, 12×75 mm2; BD

• Biosciences, cat. no. 352235)Falcon 10-cm (100 × 20 mm) dishes; bottoms are suitable for oocyte/embryo collection and lids that are suitable for micromanipulation (Becton Dickinson, cat. no. 353003)

• Falcon 6-cm (60 × 15 mm) dishes (Becton Dickinson, cat. no. 351007)

• Microloader tips (Eppendorf, 5242 956.003)

• Thin Wall Borosilicate glass with filament (Borosilicate, BF100-78-10)

• Thin Wall Borosilicate glass without filament (Borosilicate, B100-75-10)

• 0.2 mL PCR Tubes (Axygen cat. no.14-222-262)

• FemtoJet microinjector (Eppendorf)

• Inverted microscope with Hoffman optics (Olympus, IX73)

• Micromanipulator set (Narishige, MMO-202ND)

• CO₂ incubator (Thermo, BB15)

• Stereo microscope (Olympus, SZ61)

• Cell Sorter (Beckman, MoFlo XDP)

• Centrifuge (Eppendorf, 5424R)

• Micropipette Puller (Sutter Instrumen, P97)

• Micropipette Microforge (Narishige, MF-900)

• Fluorescence microscopy (Olympus, BX51)

Procedure

Generation of Cas9 or BE3 mRNA ~10h

• 1
  For vectors without T7 promoter, firstly provide a template for in vitro transcription of Cas9 or BE3 mRNA, add the T7 promoter sequence to the Cas9 or BE3 coding region by PCR amplification using the appropriate primer pair listed in Table 1 and the following reaction mix (see Box 1 instead for details of how to generate Cas9 or BE3 mRNA with T7 promoters):

  Component	Volume	Final Concentration
  10× PCR Buffer for KOD -Plus- Neo	5μl	1x
  2 mM dNTPs	5μl	0.2 mM each
  Cas9 IVT F (10μM) for Cas9	2μl	0.4μM
  Cas9 IVT R (10μM) for Cas9	2μl	0.4μM
  KOD -Plus- Neo (1U/μl)	1μl	1.0 U/50 μl
  px260 (100ng/pl) for Cas9	0.5 μl	1ng/μl
  25 mM MgSO4	3μl	1.5 mM
  ddH2O	31.5 μl
  Total	50μl

• 2
  Perform PCR using the following cycling conditions:

  Cycle number	Denature	Extend
  1	95°C, 5min
  2–34	95°C, 30s	68°C, 5min
  35		72°C, 15min

Box 1. Linearized plasmid is used as a template for in vitro transcription

For vectors including T7 promoters ~20 bp upstream of coding sequence of gene editing protein, plasmid is linearized by restriction enzyme cutting. Restriction site is selected downstream of polyA. Here we provide a brief protocol for linearization of plasmid pCMV-BE3 and in vitro transcription of BE3 mRNA.

1. Purify plasmid pcmv-be3, using the TIANprep Rapid Mini Plasmid Kit according to the manufacturer’s instructions.

2. Prepare the following reaction mixture at room temperature:

   Component	Volume	Final Concentration
   10× FastDigest Green buffer	5 μl	1x
   		0.1 μg/μl
   Plasmid pCMV-BE3 (0.5 μg/μl)	10 μl	5 μl / 50μl
   FastDigest BbsI	5 μl	5 μl / 50μl
   FastDigest PvuI	5 μl	5 μl / 50μl
   FastDigest NotI	5 μl	-
   ddH2O	20 μl	-
   Total	50μl

3. Mix gently and spin down for a few seconds.

4. Incubate at 37°C in water thermostat for 30 min.

5. Prepare 1x TAE by diluting appropriate amount of 50 x TAE with deionized water.

6. Mix 0.5g agarose powder with 50 mL 1xTAE in a 250ml conical flask.

7. Microwave for 1–3min until the agarose is completely dissolved. If not, boil the solution again.

8. Cool down the agarose solution to about 60 °C.

9. Add 5 μL of Nucleic Acid Dye to the agarose solution and mix gently.

10. Pour the agarose into a gel tray at room temperature and rest for 30 mins, until it has completely solidified.

11. Remove gel dams and place agarose gel into the gel box. Add 1xTAE into the gel box until the gel is covered.

12. Load 5 μL DNA Marker into the first lane of the gel.

13. Load 50 μL digested plasmid DNA into the additional well of the gel.

14. Run the gel at 110 V for 30 min. Three digested DNA fragments are supposed to have 5412, 1170 and 1410 bp in length, respectively, as shown in the Supplementary Fig. 3:.

    CAUTION! Examine the linearized template DNA on a gel to confirm that cleavage is complete.

15. Excise the 5412 bp DNA fragment band under long wavelength UV light.

16. Purify DNA fragment from agarose gels using the Universal DNA Purification Kit according to the manufacturer’s instructions.

17. Elute the DNA fragment with 40 μl nuclease-free water.

18. Use 1μg purified DNA fragment as the template for in vitro transcription of BE3 mRNA using the mMESSAGE mMACHINE T7 kit according to the protocol.

19. After purification, dilute the BE3 mRNA to 500 ng/μL with nuclease-free water and check its quality on a 1% (wt/vol) agarose gel in TAE buffer. The in vitro transcribed BE mRNA should have a band at ~1000bp (Supplementary Fig. 3). Failed in vitro transcription products show no clear bands at these positions as also shown in Supplementary Fig. 3.

20. Dispense 1 μl of the purified mRNA into 0.2 mL PCR Tubes.

    PAUSE POINT The samples could be stored at −80°C for up to 1 year.

• 3Run 5 μL of PCR products on a 1% (wt/vol) agarose gel at 110V for 30 min with 6 × loading buffer in TAE buffer to validate that the DNA fragment is unique and of the expected size (~4.5 kb for Cas9; ~kb for BE3; Supplementary Fig. 3).
• 4Purify PCR products, using the Universal DNA Purification Kit according to the manufacturer’s instructions.
• 5Use 1μg purified PCR product as the template for in vitro transcription of Cas9 mRNA using the mMESSAGE mMACHINE T7 kit according to the kit protocol.
• 6Purify the mRNA using the MEGAclear kit following the manufacturer’s instructions and elute the RNA in 100 μl of TB buffer. Determine the RNA concentration using the NanoDrop 2000 spectrophotometer following the manufacturer’s instructions.
• 7Dilute the purified mRNA to 500 ng/μl in 0.1 mM RNase-free ddH₂O and check its quality on a 1% (wt/vol) agarose gel in TAE buffer. The in vitro transcribed Cas9 mRNA should have a band at ~1000bp (Supplementary Fig. 3). Failed in vitro transcription products show no clear bands at these positions as also shown in Supplementary Fig. 3.
• 8Dispense 1 μl of the purified mRNA into 0.2 mL PCR Tubes.

  PAUSE POINT The samples can be stored at −80°C for up to 1 year.

Generation of Cre mRNA~10h

• 9
  To provide a template for in vitro transcription of Cre mRNA, add the T7 promoter sequence to the Cre coding region by PCR amplification using the appropriate primer pair listed in Table 1, and the following reaction mix:

  Component	Volume	Final Concentration
  10× PCR Buffer for KOD -Plus- Neo	5μl	1x
  2 mM dNTPs	5μl	0.2 mM each
  Cre IVT F (10μM) for Cre	2μl	0.4μM
  Cre IVT F (10μM) for Cre	2μl	0.4μM
  KOD -Plus- Neo (1U/pl)	1μl	1.0 U/50 μl
  pCAG-Cre (100ng/μl) for Cre	0.5μl	1ng/μl
  25 mM MgSO4	3μl	1.5 mM
  ddH2O	31.5μl
  Total	50μl

• 10Repeat Steps 2–8, diluting the purified sgRNA to 500 ng/μl in 0.1 mM EDTA, the band in Step 3 should be of the expected size (~2kb; Supplementary Fig. 3).

  The in vitro transcribed Cre mRNA should have a band at ~400bp (Supplementary Fig. 3).

  PAUSE POINT The samples can be stored at −80°C for up to 1 year.

Generation of sgRNA ~8h

• 11
  For sgRNA preparation, prepare the following PCR reaction mix to add the T7 promoter sequence to the sgRNA template by PCR amplification using the appropriate primer pair listed in Table 1:

  Component	Volume	Final Concentration
  10× PCR Buffer for KOD -Plus- Neo	5μl	1x
  2 mM dNTPs	5μl	0.2 mM each
  sgRNA IVT F (10μM)	2μl	0.4μM
  sgRNA IVT R (10μM)	2μl	0.4μM
  KOD -Plus- Neo (1.0 U/μl)	1μl	1.0 U/50 μl
  px330(100ng/μl)	0.5μl	1ng/μl
  25 mM MgSO4	3μl	1.5 mM
  ddH2O	31.5μl
  Total	50μl

• 12
  Perform PCR using the following cycling conditions:

  Cycle number	Denature	Anneal	Extend
  1	95°C, 5min
  2–34	95°C, 30s	60°C, 30s	72°C, 30s
  35			72°C, 5min

• 13Repeat Step 10 (the band in Step 3 should be of the expected size (120 bp; Supplementary Fig. 3) and the in vitro transcribed sgRNA should have a clear band at ~100bp (Supplementary Fig. 3).

  PAUSE POINT The samples can be stored at −80°C for up to 1 year.

Superovulation and mating TIMING ~3 d; ~2 h hands-on

• 14Inject 10 females C57BL/6 (3–4 weeks old) with 5 IU PMSG through intraperitoneal (i.p.) at 2:00 p.m. on day 1.

  ? TROUBLESHOOTING

• 15Approximately 47–49 h after PMSG injection (1:00–3:00 p.m. on day 3), inject the females i.p. with 5 IU hCG to induce ovulation.
• 16Put each hormone-stimulated female together with Ai9 males in a 1:1 ratio in a mating cage overnight.

  ? TROUBLESHOOTING

Zygote collection and processing ~3h

• 17Prepare two 35-mm dishes each containing 14 drops (30 μl for each drop) of KSOM medium covered with mineral oil for embryo culture. Transfer dishes into a 37°C incubator for at least 20 min before use. Prepare one 100-mm dish with 8 drops (200 μl for each drop) M2 medium.
• 18Euthanize females by CO₂ asphyxiation at 22–24 h after hCG injection.
• 19Isolate oviducts and place all the oviducts into one 200 μl drop of M2 medium in the 100 mm dishes (Supplementary Fig. 4).
• 20Under the stereoscopic microscope, transfer one oviduct at a time into the second 200 μl drop of M2 medium (prepared in Step 17). Tear the oviduct where it is most swollen using a 1 ml syringe attached to a 26G needle, releasing the zygote-cumulus complexes (ZCCs). Repeat this step for every oviduct until all the zygotes are isolated.
• 21Add 200 μl of hyaluronidase to the ZCCs in the droplet, and then place the dish into a 37°C incubator for 3 min.

  Critical Step: No longer than 3 min.

• 22Pipette the droplet up and down several times with a yellow tip until the cumulus cells are completely removed from the zygotes. Successively transfer the zygote clockwise through five wash drops of M2 medium (prepared in Step 17).
• 23Finally, transfer the zygotes to one drop of KSOM medium in the pre-equilibrated 35-mm dish from Step 17 using a hand pipette, pass the zygotes through six to ten additional KSOM droplets to wash and place the dish at 37°C in a 5% CO₂ incubator up to 24h until ready for injection.

Microinjection Preparation ~20min

• 24Approximately 48–50 h after hCG injection, the zygotes develop to the end of the two-cell stage and the cytoplasms of the two blastomeres should be completely separated.
• 25Prepare the appropriate injection mix depending on the aim of the experiment, as outlined in the table below. Cas9 or BE3 mRNA and Cre mRNA stocks from Step 8 Box 1, and Step 10 and sgRNA stock from Step 13 are both at concentrations of 500 ng/μl. Cre mRNA stock should be diluted down to 20 ng/μl with Nuclease-free water.
  Combine the following reagents at room temperature in a sterile, RNase-free eight-well PCR strip.

  Component	Volume	Final concentration
  Cas9/BE3 mRNA (500 ng/μl)	1μl	50 ng/μl
  sgRNA (500 ng/μl)	1μl	50 ng/μl
  Cre mRNA (20 ng/μl)	1μl	2 ng/μl
  Nuclease-free water	7μl	-
  Total	10μl	-

  Mix all the injection components just before use.
• 26Centrifuge components at 15,000 × g for 5 min at 4 °C and and store on ice ready for microinjection.
• 27Load 3μl of injection mix into the injection needle using a microloader tip. Keep the injection needle vertically for 5 minutes by sticking it to a piece of plasticine to remove small air bubbles in the needle.
• 28Place a droplet of HEPES-CZB + CB on top of a 10-cm dish, and then cover it with mineral oil. Place the dish under a stereoscopic microscope.
• 29Attach the loaded injection needle to the instrument holder connected to the Femtojet, lower the capillary into the medium drop on the stage such that it is positioned in the center just below the embryos and switch on the Femtojet and allow it to reach pressure.
• 30Attach the holding pipette to the other side of the micromanipulator.

Injection of embryos ~1h (Supplementary Video 1)

• 31Transfer 100 2-cell stage embryos into a large drop of HEPES-CZB + CB (prepared in Step 28) using a hand pipette. The number of zygotes to be moved into the microinjection drop should be determined by the skills of the injector and quality of the setup.

  Critical Step: Do not attempt to work with more zygotes than can be injected within 30 minutes.

• 32Determine whether the microinjection capillary is open and unclogged by placing the tip of the microinjection capillary close to a zygote in the same horizontal plane under a continuous flow stream. If the microinjection capillary is open, a stream of DNA will move the zygote away from the tip of the microinjection capillary. If the injection pipette is still not open, tip it carefully on the holding pipette until it breaks at a larger dimension.
• 33Hold an embryo using a holding pipette. Insert the injection tip into one blastomere and pause briefly halfway inside the blastomere to see the formation of a small droplet around the injection tip.
• 34When all the embryos in the chamber have been injected, they should immediately be moved back into KSOM medium from Step 17, wash 6–8 times and incubated at 37°C in a 5% CO₂ for 30 min. At this point, 80% of the embryos will be transferred to a foster mother oviduct for development to E14.5 (Steps 35–43), while the other 20% will be cultured in vitro to blastocyst stage for genotyping (Steps 44–57).

Reimplantation of injected embryos ~2h

• 35Prepare 4 pseudopregnant foster mothers by mating estrous ICR female mice with vasectomized male mice on the same day as zygote collection.
• 36On the morning of embryo transfer, identify 0.5 days post coitum (dpc) foster mothers with visible copulatory plugs.
• 37Anesthetize foster mother by peritoneal injection of Avertin using a 1-ml syringe attached to a 26-gauge needle, after having weighed foster mothers to calculate the injection dose of Avertin (2.5% Avertin solution at a dose of 0.01 ml/g of body weight).
• 38After 1 min, check that the mouse is fully anesthetized by lightly pinching the most medial toe. Once the mouse is unresponsive, place the anesthetized mouse under the stereomicroscope and disinfect with 75% (vol/vol) ethanol.
• 39Make a small longitudinal incision (≤1 cm) parallel to the midline at the level of the last rib and slide the skin to expose the body wall. Pick up the body wall with forceps over the site of the ovary and make a small incision through the body wall. Expose the ovary, oviduct, and part of the uterus through the incision.
• 40Load the glass transfer pipette by drawing KSOM medium ~1 cm up the pipette, then one small air bubble, then 18–22 embryos from Step 34 within a minimal amount of medium, and finally another small air bubble.
• 41Make a small hole in the upper ampulla using a 30G needle. Insert glass transfer pipette into the hole and transfer the embryos with air bubbles.
• 42Remove the transfer pipette, gently return the ovary, oviduct, and uterus back inside the body and seal the incision with absorbable sutures.
• 43Keep the mice warm on a 37°C warming plate until the mouse recovers from the effects of the anesthesia.

  ? TROUBLESHOOTING

Nested PCR detection of targeted embryos ~6h

• 44Incubate the remaining 20% injected embryos from Step 34 at 37°C, 5% CO₂ for 3 days until blastocyst stage.
• 45Wash single blastocysts 3–6 times with KSOM.
• 46Transfer blastocysts directly into PCR tubes (one blastocyst per tube), and add1.5 μl embryo lysis buffer.

  Critical Step: Using stereo microscope to confirm the blastocysts are indeed put into PCR tubes.

• 47Incubate at 56°C for 30 min, heat inactivate at 95°C for 10 min.
• 48
  Prepare nested PCR amplification reactions to amplify the crude DNA solutions, using the following reaction mix.

  Component	Volume	Final Concentration
  Premix Taq™ (Ex Taq™ Version 2.0 plus dye)	25μl	1x
  Tyr-OF/IF (10μM)	2μl	0.4μM
  Tyr-OR/IR (10μM)	2μl	0.4μM
  Crude DNA solution	1.5μl
  Nuclease-free water	Up to 50μl

• 49
  Perform PCR using the following cycling conditions:

  Cycle number	Denature	Anneal	Extend
  1	95°C, 5min
  2–34	95°C, 30s	60°C, 30s	72C, 30s
  35			72°C, 5min

• 50Purify PCR products from Step 49, using the Universal DNA Purification Kit according to the manufacturer’s instructions.
• 51
  Clone purified PCR products into pMD-19T using the pMD-19T cloning vector kit, according to the manufacturer’s instructions.

  Component	Volume	Final Concentration
  pMD-19T (50 pg/μl)	0.5 μl	2.5 μg/μl
  PCR product (50 μg/μl)	1μl	50 μg/μl
  Solution I	5μl	1x
  Nuclease-free water	Up to 10μl

  Mix all the components and incubate at 56°C for 30 min.
• 52Transform the above reaction mix into E. coli DH5α and cloned into E. coli (TA clone).
• 53Plate all of the transformation onto a 10 cm LB agar plate with 100ug/ml of ampicillin and incubate plates at 37°C overnight.
• 54Pick a single colony from an agar plate using a pipet tip and drop the pipet tip in a tube containing 5 mL LB medium with 100ug/ml of ampicillin.
• 55Incubate the bacteria in a shaking incubator for 15h at 37°C, 200 rpm.
• 56Extract plasmid from E. coli using the TIANprep Rapid Mini Plasmid Kit according to the manufacturer’s instructions.
• 57Determine the number of clones with indels or SNVs by Sanger Sequencing. 20 TA clones are picked for Sanger sequencing, further experiments will be conducted only if the on-target editing was more than 20% (4/20 clones).

Isolation of embryonic cells and fluorescent-activated cell sorting (FACS) ~10h

• 58Euthanize 14.5-d pregnant mice from Step 43 by cervical dislocation or anesthesia and separate the embryos with the placenta, then aseptically dissect mouse embryos. Wash the embryos with PBS twice briefly.
• 59Place each embryo into a 100-mm petri dish and mince them into the smallest possible pieces less than 1mm.
• 60Add 5 mL of pre-warmed 37°C Trypsin-EDTA (0.05%) and transfer to a 50mL centrifuge tube and incubate at 37°C for 30 minutes.
• 61Add 5 ml DMEM medium and Homogenize fetal tissues by passing 30–40 times through a 1ml pipette tip.
• 62Centrifuge the cell suspension at 200g for 6 min at room tempetature, and resuspend the pellet in 5 mL of DMEM medium.
• 63Filter the cell suspension through a 40-μm cell strainer.
• 64Centrifuge the cell suspension at 200g for 6 min at room temperature, aspirate the medium gently without disturbing the pellet and resuspend the pellet in 5 mL of DMEM medium.
• 65Pass cell suspension through a 35-μm cell strainer cap into a 5-ml polystyrene round-bottom tube.
• 66Sort the tdTomato⁺ cells and tdTomato⁻ cells into two separate tubes (~4 million cells for each) using a Moflo XDP fluorescence-activated cell sorter (Beckman Coulter), respectively. The FSC/SSC gates of the starting cell population were set in order to include all cells. Then doublet cells were excluded by SSC-H vs SSC-A. Positive and negative boundaries were defined by control progeny cells of nonedited blastomeres (see Supplementary Figure 2.
• 67Extract genomic DNA from sorted cells using the DNeasy blood and tissue kit according to the manufacturer’s instructions.

Sample Quality control (QC) ~36h

• 68Perform Nested PCR amplification reactions on the extracted purified genomic DNA from tdTomato⁺ cells and tdTomato⁻ cells, respectively, as described in Steps 48–57. Ligate the PCR products into T-Vectors, and pick 20 TA clones for Sanger sequencing. Conduct further experiments (Steps 69–78) only if less than 10% (2/20) of clones show on-target editing in tdTomato⁻ cells and more than 20% (4/20) of clones show on-target editing in tdTomato⁺ cells.

WGS .Timing ~3 d

• 69Perform whole genome sequencing on the extracted genomic DNA from Step 67 at mean coverage of 50x by Illumina HiSeq X Ten.

Processing of raw reads ~20 h

• 70
  Quality control raw sequencing reads by fastQC, using the following commands:

  fastqc raw/cre_neg_R1.fastq.gz raw/cre_neg_R2.fastq.gz -o fastQC/pretrim
  fastqc raw/cre_pos_R1.fastq.gz raw/cre_pos_R2.fastq.gz -o fastQC/pretrim
  

  If the quality of sequencing reads is good (Quality score per base > 20; Supplementary Fig. 5), just skip to Step 73. Otherwise, apply Step 72 to trim low quality reads and remove adapter sequences (Supplementary Fig. 5).

• 71
  (Optional) Use Trimmomatic to trim the low quality reads and adapter sequences in the FASTQ files, using the following commands:

  java -jar trimmomatic-0.36.jar PE -threads 8 cre_neg_R1.fastq.gz cre_neg_R2.fastq.gz
  fastQC/trim/cre_neg_R1_paired.fastq.gz fastQC/trim/cre_neg_R1_unpaired.fastq.gz
  fastQC/trim/cre_neg_R2_paired.fastq.gz fastQC/trim/cre_neg_R2_unpaired.fastq.gz
  ILLUMINACLIP:TruSeq3-PE.fa:2:30:10 LEADING:5 TRAILING:5 MINLEN:70
  java -jar trimmomatic-0.36.jar PE -threads 8 cre_neg_R1.fastq.gz cre_neg_R2.fastq.gz
  fastQC/trim/cre_pos_R1_paired.fastq.gz fastQC/trim/cre_pos_R1_unpaired.fastq.gz
  fastQC/trim/cre_pos_R2_paired.fastq.gz fastQC/trim/cre_pos_R2_unpaired.fastq.gz
  ILLUMINACLIP:TruSeq3-PE.fa:2:30:10 LEADING:5 TRAILING:5 MINLEN:70
  fastqc fastQC/trim/cre_neg_R1_paired.fastq.gz fastQC/trim/cre_neg_R2_paired.fastq.gz -o fastQC/posttrim
  fastqc fastQC/trim/cre_pos_R1_paired.fastq.gz fastQC/trim/cre_pos_R2_paired.fastq.gz -o fastQC/posttrim
  

• 72
  Construct the BWA index file of the reference genome mm10.fa, using the following command:

  bwa index mm10.fa
  

• 73
  Align clean reads to the reference genome using BWA mem and use Picard-tools to reorder, sort, add read groups and mark duplicates of the aligned BAM files as follows:

  bwa mem -t 8 -M mm10.fa fastQC/trim/cre_neg_R1_paired.fastq.gz fastQC/trim/cre_neg_R2_paired.fastq.gz |
  samtools view -bS -o cre_neg.bam
  bwa mem -t 8 -M mm10.fa fastQC/trim/cre_pos_R1_paired.fastq.gz fastQC/trim/cre_pos_R2_paired.fastq.gz |
  samtools view -bS -o cre_pos.bam
  java -Xmx20g -jar picard.jar ReorderSam INPUT=cre_neg.bam OUTPUT=cre_neg.reorder.bam
  REFERENCE=mm10.fa
  java -Xmx20g -jar picard.jar ReorderSam INPUT=cre_pos.bam OUTPUT=cre_pos.reorder.bam
  REFERENCE=mm10.fa
  java -Xmx20g -jar picard.jar SortSam INPUT=cre_neg.reorder.bam OUTPUT=cre_neg.reorder.sort.bam
  SORT_ORDER=coordinate
  java -Xmx20g -jar picard.jar SortSam INPUT=cre_pos.reorder.bam OUTPUT=cre_pos.reorder.sort.bam
  SORT_ORDER=coordinate
  java -Xmx20g -jar picard.jar AddOrReplaceReadGroups VALIDATION_STRINGENCY=SILENT
  INPUT=cre_neg.reorder.sort.bam OUTPUT=cre_neg.reorder.sort.add.bam RGLB=WES RGPL=Illumina
  RGPU=HiSeq RGSM=cre_neg
  java -Xmx20g -jar picard.jar AddOrReplaceReadGroups VALIDATION_STRINGENCY=SILENT
  INPUT=cre_pos.reorder.sort.bam OUTPUT=cre_pos.reorder.sort.add.bam RGLB=WES RGPL=Illumina
  RGPU=HiSeq RGSM=cre_neg
  java -Xmx20g -jar picard.jar MarkDuplicates VALIDATION_STRINGENCY=SILENT
  INPUT=cre_neg.reorder.sort.add.bam OUTPUT=cre_neg.reorder.sort.add.mkdup.bam
  METRICS_FILE=cre_neg.reorder.sort.add.mkdup.metrics
  java -Xmx20g -jar picard.jar MarkDuplicates VALIDATION_STRINGENCY=SILENT
  INPUT=cre_pos.reorder.sort.add.bam OUTPUT=cre_pos.reorder.sort.add.mkdup.bam
  METRICS_FILE=cre_pos.reorder.sort.add.mkdup.metrics
  java -Xmx20g -jar picard.jar BuildBamIndex VALIDATION_STRINGENCY=SILENT
  INPUT=cre_neg.reorder.sort.add.mkdup.bam
  java -Xmx20g -jar picard.jar BuildBamIndex VALIDATION_STRINGENCY=SILENT
  INPUT=cre_pos.reorder.sort.add.mkdup.bam
  

Detection of off-target effects ~30h

• 74
  To detect whether off-target variants specifically existed in tdTomato⁺ cells, apply three variant calling tools to identify SNVs and indels by comparing tdTomato⁺ cell with tdTomato⁻ cells from the same embryo, using the following commands:

  java -Xmx20g -jar GenomeAnalysisTK.jar -R mm10.fa -T MuTect2 -I:tumor
  cre_pos.reorder.sort.add.mkdup.bam -I:normal cre_neg.reorder.sort.add.mkdup.bam -o
  mutect/cre_pos/output.vcf
  lofreq somatic -n cre_neg.reorder.sort.add.mkdup.bam -t cre_pos.reorder.sort.add.mkdup.bam -f mm10.fa –
  threads 8 -o lofreq/cre_pos
  strelka-2.7.1.centos5_x86_64/bin/configureStrelkaSomaticWorkflow.py --normalBam
  cre_neg.reorder.sort.add.mkdup.bam --tumorBam cre_pos.reorder.sort.add.mkdup.bam --referenceFasta
  mm10.fa --runDir strelka/cre_pos
  scalpel-0.5.3/scalpel-discovery --somatic –normal cre_neg.reorder.sort.add.mkdup.bam –tumor
  cre_pos.reorder.sort.add.mkdup.bam –bed mm10.bed --window 600 --numprocs 8 --ref mm10.fa --dir
  scalpel/cre_pos
  awk ‘$7==”PASS” {print $0}’ strelka/cre_pos/results/variants/somatic.snvs.vcf >
  strelka/cre_pos/results/variants/somatic.snvs.pass.vcf
  awk ‘$7==”PASS” {print $0}’ strelka/cre_pos/results/variants/somatic.indels.vcf >
  strelka/cre_pos/results/variants/somatic.indels.pass.vcf
  awk ‘$7==”PASS” {print $0}’ mutect/cre_pos/output.vcf > mutect/cre_pos/output.pass.vcf
  awk ‘length($4)==1 && length($5)==1 {print $0}’ mutect/cre_pos/output.pass.vcf >
  mutect/cre_pos/output.pass.snv.vcf
  awk ‘length($4)>1 ∥ length($5)>1 {print $0}’ mutect/cre_pos/output.pass.vcf >
  mutect/cre_pos/output.pass.indel.vcf
  

  The SNVs identified by all three algorithms Mutect2, Strelka2 and Lofreq are considered true SNVs, and the overlap of Mutect2, Strelka2 and Scalpel is identified as true indels. This step can be performed using the following commands:

  perl filter_overlap.pl mutect/cre_pos/output.pass.snv.vcf lofreq/cre_pos/somatic_final.snvs.vcf
  strelka/cre_pos/results/variants/somatic.snvs.pass.vcf cre_pos.snv.overlap.vcf
  perl filter_overlap.pl mutect/cre_pos/output.pass.indel.vcf scalpel/cre_pos/somatic_final.indels.vcf
  strelka/cre_pos/results/variants/somatic.indels.pass.vcf cre_pos.indel.overlap.vcf
  

Validation of the off-target variants using Sanger sequencing ~2d

• 75Design primers for a dozen of the detected SNVs for Sanger sequencing. The primers should have ~50% GC content.
• 76Perform PCR amplification as described in Steps 48–49 and Sanger sequence the purified PCR products.

Filtration and functional annotation ~10min

• 77
  Use Annovar as follows to annotate the identified variants from Step 75 and only use those with allele frequency more than 10% for the final results.

  annotate_variation.pl -buildver mm10 -downdb -webfrom annovar refGene mousedb/
  table_annovar.pl cre_pos.snv.overlap.vcf mousedb -buildver mm10 -out cre_pos.snv.overlap.vcf.anno -remove
  -protocol refGene -operation g -nastring . -vcfinput
  perl awk_anno.pl cre_pos.snv.overlap.vcf.anno.mm10_multianno.txt cre_pos.anno.tsv
  awk ‘$10>0.1 {print $0}’ cre_pos.anno.tsv > cre_pos.anno.0.1.tsv
  

Sequence comparison between off-target and on-target variants ~20min

• 78
  Retrieve the adjacent 23bp sequences of the off-target variants (5bp upstream and 17bp downstream as the target base of on-target site in this study is the 6^th nucleotide of the sgRNA sequence) from the marked BAM file (output from Step 74) and then blasted with the on-target sequence sgRNA.fasta (23bp; 20 bp sgRNA target sequence + 3bp PAM). Meanwhile, blast the predicted off-target sites from Cas-OFFinder⁵¹ based on sequence similarity with the on-target sequences using National Center for Biotechnology Information (NCBI) Basic Local Alignment Search Tool (BLAST) as follows:

  cat cre_pos.anno.0.1.tsv | while read line
  do
   chr=$(echo $line | cut -d” “ -f 1)
   pos=$(echo $line | cut -d” “ -f 2)
   start=`expr $pos - 5`
  end=`expr $pos + 17`
   samtools faidx mm10.fa $chr:${start}-${end} >> cre_pos.anno.0.1.tsv.fasta
  done
  makeblastdb -in sgRNA.fasta -dbtype nucl -parse_seqids
  blastn -db sgRNA.fasta -query cre_pos.anno.0.1.tsv.fasta -dust no -outfmt 0 -word_size=7 -out
  cre_pos.sgRNA.blast.out
  

  This step reveals the sequence similarity between off-target variants and on-target edits. High sequence similarities indicate the off-target variants are associated with sgRNAs, while low sequence similarities suggest the off-target effects are sgRNA-independent.

Timing

• Steps 1–8, generation of Cas9 mRNA: ~10h

• Steps 9–11, generation of Cre mRNA: ~10h

• Steps 12–13, generation of sgRNA: ~8h

• Steps 14–16, superovulation and mating: ~3d

• Steps 17–23, zygote collection and processing: ~3h

• Steps 24–30, microinjection preparation: ~20min

• Steps 31–34, injection of embryos: ~1h

• Steps 35–43, reimplantation of injected embryos: ~2h

• Steps 44–57, nested PCR detection of targeted embryos: ~6h

• Steps 58–67, isolation of embryonic cells and fluorescent-activated cell sorting (FACS): ~10h

• Steps 68, sample quality control (QC): ~36h

• Steps 69, WGS: ~3d

• Steps 70–73, processing of raw reads: ~20h

• Step 74, detection of off-target effects: ~30h

• Steps 75–76, validation of the off-target variations using Sanger sequencing: ~2d

• Step 77, filtration and functional annotation: ~10 min

• Step 78, sequence comparison between off-target and on-target sequences ~20 min

Troubleshooting

Troubleshooting advice can be found in Table 2.

Table 2.

Troubleshooting.

Step	Problem	Possible reason	Solution
15	Too many or two few eggs	Incorrect age of female mice	Select female mice with 3–4 weeks old
16	Low fertilization ratio	Frequent use of male mice	Mate at most twice a week for the male mice, with at least two days between
44–46	Low development rate	The injected mRNA was not pure and is toxic	Re-generate the mRNAs from Steps 1–16
58–66	Too low a proportion of tdTomato+ cells after FACS	Gene editing tools may be toxic for the development of tdTomato+ cells The targeting gene is essential or important to the development of cells The quality of mRNA is inadequate	Choose high fidelity version of gene editing tools Select non essential genes as target genes Re-prepare mRNA for injection
74	Mutect2 takes too much time to complete	Mutect2 takes a lot of memory when running WGS data	Split the WGS into separate chromosomes to run each time

Anticipated results

A successful application of the experimental procedure will generate two equal populations of cells with tdTomato or not (Fig. 3a). Embryos with obviously unequal proportions of tdTomato⁺ cells should not be sequenced for further analysis (Supplementary Fig. 6). Examples of the on-target efficiency for CRISPR/Cas9 (Cas9-Tyr-C¹⁶) and BE3 (BE3-Tyr-C-#1¹⁶) editing following injection of sgRNA are shown in Fig. 3b. The bioinformatic pipeline will reveal the number of off-target variants in each embryo (Supplementary Fig. 7). The off-target mutations induced by CRISPR/Cas9 or BE3 editing were compared and BE3 induced many more off-target edits than CRISPR/Cas9 (Fig. 3c). The off-target variants can be used for downstream analysis by group comparison, mutation bias and functional enrichment¹⁶. The adjacent sequences of identified variants are compared with the on-target sites (Fig. 3d), but poor sequence similarity is observed (mean Bit-score = 10.4 for Cas9-Tyr-C; mean Bit-score = 12.2 for BE3-Tyr-C-#1). By contrast, potential off-target sequences predicted by Cas-OFFinder show high similarity with the on-target sequence (mean Bit-score = 29.4; Fig. 3d). These results suggest that the off-target variants identified by GOTI in CRISPR/Cas9 or BE3 editing are sgRNA-independent.

Fig. 3. Anticipated results from GOTI.

a, The separation of tdTomato⁺ and tdTomato⁻ cells. FACS analysis for E14.5 embryo of Cas9-Tyr-C¹⁶. b, On-target efficiency for tdTomato⁺ and tdTomato⁻ cells based on WGS for Cas9-Tyr-C¹⁶ and BE3-Tyr-C-#1¹⁶-treated embryos. The target site was edited with high efficiency in tdTomato⁺ cells, but not in tdTomato⁻ cells. The deletions induced by Cas9-Tyr-C and base substitutions in BE3-Tyr-C-#1 were highlighted in red rectangles. Dark lines represent deletions in the region and green rectangles represent substitutions. c, The number of off-target SNVs detected in Cas9-Tyr-C and BE3-Tyr-C-#1-treated embryos. BE3-Tyr-C-#1 induced much higher number of off-target variants than Cas9-Tyr-C-#1. d, The sequence similarity (Bit-score) between on- and off-target sequences identified by GOTI in Cas9-Tyr-C¹⁶ and BE3-Tyr-C-#1¹⁶-treated embryos or predicted by Cas-OFFinder. The off-target SNVs are sgRNA independent. n = 1809 for Cas-OFFinder, n = 18 for Cas9-Tyr-C and n = 247 for BE3-Tyr-C-#1. Box-and-whisker plots: center line indicates median, the bottom and top lines of the box represents the first quartile and third quartile of the values, respectively. The bottom and top lines represent the minimum and maximum value. P values are calculated with two-sided Wilcoxon rank sum test.

Supplementary Material

Supplementary video

Supplementary Video Gene editing system are introduced into one blastomere of two-cell mouse embryos by microinjection.