Unexpected genomic rearrangements at targeted loci associated with CRISPR/Cas9-mediated knock-in

Abstract

The CRISPR/Cas9 gene editing tool enables accessible and efficient modifications which (re)ignited molecular research in certain species. However, targeted integration of large DNA fragments using CRISPR/Cas9 can still be challenging in numerous models. To systematically compare CRISPR/Cas9's efficiency to classical homologous recombination (cHR) for insertion of large DNA fragments, we thoroughly performed and analyzed 221 experiments targeting 128 loci in mouse ES cells. Although both technologies proved efficient, CRISPR/Cas9 yielded significantly more positive clones as detected by overlapping PCRs. It also induced unexpected rearrangements around the targeted site, ultimately rendering CRISPR/Cas9 less efficient than cHR for the production of fully validated clones. These data show that CRISPR/Cas9-mediated recombination can induce complex long-range modifications at targeted loci, thus emphasizing the need for thorough characterization of any genetically modified material obtained through CRISPR-mediated gene editing before further functional studies or therapeutic use.

Figure 1

Comparison of cHR and Cas9n knock-in performance in mouse embryonic stem cells. (A) Number of experiments and loci targeted in mouse ES cells using Cas9n technology (Cas9n) and cHR (see Sup Table for further details). (B) Box plots of percentages of positive clones detected by 5′PCR (left) or 3′PCR (right) for all experiments at all loci using Cas9n or cHR. (C) Box plots of percentages of clones validated by Southern Blot (SB; left) or sequencing (right) of the targeted locus for all experiments at all loci using Cas9n or cHR. (D) Box pots of Discovery Index (DI, left) and Validation Index (VI, right) for all experiments at all loci using Cas9n or cHR. DI and VI represents the technology’s efficiency to produce PCR-positive clones and fully validated clones, respectively.

Figure 2

Knock-In efficiency for inserts of different size with Cas9n and cHR. (A) Number of experiments and loci targeted in mouse ES cells using Cas9n technology or cHR for inserts of various size (see Sup Table for further details). (B) Box plots of Discovery Index (DI, top) and Validation Index (VI, bottom) for small (<2 kb) and bigger inserts (>2 kb). (C) Box plots of Discovery Index (DI, top) and Validation Index (VI, bottom) for large (>8 kb) and smaller inserts (<8 kb). Note that Cas9n’s DI is superior to cHR’s for all insert sizes whereas cHR’s VI is higher than Cas9n’s in all conditions except for inserts >8 kb.

Figure 3

Homology arm length and Cas9n enzyme concentration effect on Cas9n efficiency. (A) Number of experiments and loci targeted in mouse ES cells using Cas9n technology with short (sHA) or long homology arms (lHA), low (LC) and high concentration (HC) of enzyme, and cHR (see Sup Table for further details). (B) Box plots of Discovery Index (DI, top) and Validation Index (VI, bottom) for sHA and lHA. (C) Box plots of Discovery Index (DI, left) and Validation Index (VI, right) for LC and HC. Note that HA length does not affect Cas9n’s efficiency. HC shows higher DI than LC but the VI is still inferior to cHR’s. R: μg Cas9n/million cells.

Figure 4

Cas9n technology induces unexpected rearrangements at the targeted locus. (A) Number of experiments and loci targeted in mouse ES cells using Cas9n technology and cHR, tested by Southern Blot (SB) with an external probe (see Sup Table for further details). (B) Blot of Cas9n experiment #34 targeting locus E showing profiles obtained with an external probe after EcoNI digestion (WT: 4011; Rec: 7108). Note that clones 34b, 34 f, 34i, 34 l and 34n show bands of unexpected size. (C) Box plots of percentages of clones showing bands of unexpected sizes by SB with Cas9n and cHR for all loci. L: ladder.