Quantifying genome-editing outcomes at endogenous loci with SMRT sequencing

Abstract

Targeted genome editing with engineered nucleases has transformed the ability to introduce precise sequence modifications at almost any site within the genome. A major obstacle to probing the efficiency and consequences of genome editing is that no existing method enables the frequency of different editing events to be simultaneously measured across a cell population at any endogenous genomic locus. We have developed a method for quantifying individual genome-editing outcomes at any site of interest with single-molecule real-time (SMRT) DNA sequencing. We show that this approach can be applied at various loci using multiple engineered nuclease platforms, including transcription-activator-like effector nucleases (TALENs), RNA-guided endonucleases (CRISPR/Cas9), and zinc finger nucleases (ZFNs), and in different cell lines to identify conditions and strategies in which the desired engineering outcome has occurred. This approach offers a technique for studying double-strand break repair, facilitates the evaluation of gene-editing technologies, and permits sensitive quantification of editing outcomes in almost every experimental system used.

Figure 1. Measuring gene editing at an endogenous locus with SMRT sequencing

(A) Sequence of the TALEN target site at the IL2RG locus and the IL2RG donor template. The donor template harbors seven point mutations that, when integrated into IL2RG, create silent mutations and a novel AflII restriction site. These substitutions alter the right TALEN binding site and provide a signature for alleles precisely modified by HDR. (B) Diagram of gene editing at an endogenous locus. TALENs create a double strand break (DSB), which can lead to no modification, insertion or deletion mutations, or integration of point mutations from the donor template. (C) Schematic of SMRT DNA sequencing analysis. The endogenous locus is amplified by PCR, with at least one primer outside the arms of homology of the donor template, and SMRT adapters are added to PCR amplicons. Individual DNA molecules are sequenced by SMRT sequencing, with read lengths averaging ~3kb in length and approaching ~15kb. (D) Measurement of gene editing outcomes at the IL2RG locus in K562 cells. Modification frequencies are normalized to transfection efficiency. Data for graph is from Table S1. Bars represent three independent biological replicates; error bars, s.d.

Figure 2. Reliability of SMRT sequencing analysis for measuring gene editing outcomes at an endogenous locus

(A) RFLP analysis of K562 cells targeted with 1μg of each TALEN and 5μg donor in triplicate. The frequency of HDR in each sample as measured by RFLP and SMRT sequencing analysis is shown. (B) Quantification of NHEJ and HDR frequencies in single cell clones grown from a representative population of K562 cells. Error bars represent 90% confidence intervals. (C) A representative sample of targeted K562 cells was analyzed by SMRT sequencing 8 separate times to determine the variability introduced by PCR, SMRT library synthesis, and sequencing. Error bars represent 90% confidence intervals. (D) Quantification of the observed experimental variation compared to the expected statistical variation for the number of sequences analyzed for the 8 replicates. Error bars for experimental variation represent standard deviation. Error bars for statistical variation represent 68% confidence intervals (corresponding to the fraction of the normal distribution covered by ± 1 standard deviation).

Figure 3. Measurement of genome editing at an endogenous locus in human primary cells

(A) Measurement of gene editing outcomes at IL2RG in CD34+ HSPCs using the high-expression TALEN plasmids (Supplemental Experimental Procedures: Plasmid Sequences). Data for graph is from Table S2. Bars represent three independent biological replicates; error bars, s.d. (B) Measurement of gene editing outcomes at IL2RG in hESCs using the high-expression TALEN plasmids (Supplemental Experimental Procedures: Plasmid Sequences). Data for graph is from Table S3. Bars represent three independent biological replicates; error bars, s.d.

Figure 4. Measuring gene editing with different engineered nuclease platforms at different genomic targets

(A) Left; IL2RG target site for TALENs and RGEN guide sequence. The IL2RG start codon is shown in cyan. Right; Modification of the IL2RG locus in K562 cells. Data for graph is from Table S4. (B) Left; HBB target site for TALENs and RGEN guide sequence. The HBB start codon is shown in cyan. Right; Modification of the HBB locus in K562 cells. Data for graph is from Table S5. (C) Left; CCR5 target site for TALENs and ZFNs in exon 3. Right; Modification of the CCR5 locus in K562 cells. Data for graph is from Table S6. For CelI assay results see also Figure S1. Bars represent three independent biological replicates; error bars, s.d.

Figure 5. Optimization of gene editing parameters at IL2RG with SMRT sequencing

(A) Titration of TALEN amount in K562 cells with amount of donor DNA held constant at 5μg. Data for graph is from Table S7. (B) Titration of donor DNA amount in K562 cells with TALEN DNA amount held constant at 1μg each TALEN. Data for graph is from Table S8. (C) Ratio of HDR to NHEJ for samples in (B). (D) Left; Schematic of donor templates with varying arm of homology lengths. Right; Quantification of effect of homology arm length on gene editing frequencies in K562 cells. Data for graph is from Table S9. Bars represent three independent biological replicates; error bars, s.d.

Figure 6. DNA repair by insertion of large sequences from various sources

(A) An event from the IL2RG RGEN-only transfections where the insert is an exact repeat—in the inverse orientation—of a sequence near the cleavage site. (B) Reads from the IL2RG RGEN and Donor transfections containing inserts derived from the E. coli genome, an intronic sequence in chromosome 12, and a Donor plasmid. (C) Several representative SMRT reads from the HBB TALENs-only transfections were recovered containing inserts derived from a repetitive genomic element, one of the TALEN plasmids, and a region in chromosome 11 on the opposite arm from the HBB gene. (D) Reads from the HBB RGEN and Donor transfections containing inserts derived from the E. coli genome and a region in chromosome 11 on the opposite arm from the HBB gene. (E) Reads from the CCR5 TALENs-only transfections containing inserts derived from a TALEN plasmid and an unknown plasmid. Wild-type (WT) sequences are shown with nuclease binding sites highlighted in yellow. Inserted sequences in the SMRT reads are lowercase and highlighted in blue with the size and source listed below. Deleted bases, likely resulting from sequencing errors, are represented by hyphens highlighted in red. Sequence identifiers are provided to cross reference the Supplemental Results section for further information: IL2RG_R_4, IL2RG_R+D_4, IL2RG_R+D_7, IL2RG_R+D_8, HBB_T_1, HBB_T_3, HBB_T_5, HBB_R+D_1, HBB_R+D_3, CCR5_T_1, CCR5_T_2.