CRISPaint allows modular base-specific gene tagging using a ligase-4-dependent mechanism

Abstract

The site-specific insertion of heterologous genetic material into genomes provides a powerful means to study gene function. Here we describe a modular system entitled CRISPaint (CRISPR-assisted insertion tagging) that allows precise and efficient integration of large heterologous DNA cassettes into eukaryotic genomes. CRISPaint makes use of the CRISPR-Cas9 system to introduce a double-strand break (DSB) at a user-defined genomic location. A universal donor DNA, optionally provided as minicircle DNA, is cleaved simultaneously to be integrated at the genomic DSB, while processing the donor plasmid at three possible positions allows flexible reading-frame selection. Applying this system allows to create C-terminal tag fusions of endogenously encoded proteins in human cells with high efficiencies. Knocking out known DSB repair components reveals that site-specific insertion is completely dependent on canonical NHEJ (DNA-PKcs, XLF and ligase-4). A large repertoire of modular donor vectors renders CRISPaint compatible with a wide array of applications.

Figure 1. Self-cleaving sgRNA plasmid integration.

(a) Three steps of a self-cleaving plasmid integration system. I, an sgRNA is expressed from an sgRNA plasmid under the control of a U6 promoter. II, in conjunction with Cas9 protein, the sgRNA cleaves the coding region of a target gene as well as the sgRNA plasmid itself. III, the generated DNA ends of genomic and plasmid DNA are intended to be ligated by the cellular NHEJ or HR machinery, resulting in the genomic integration of the full-length plasmid. (b) To allow tag expression after integration of a plasmid according to a, stop codons in the constant part of the sgRNA sequence have to be removed. Left panel, normal sgRNA sequence with stop codons in red. Middle panel, stop codons have been removed, base changes in green. Right panel, a PAM motif (blue) has been introduced into the sgRNA sequence to allow efficient cleavage of the sgRNA plasmid. Lower panel, exemplary sequences for tagging the human ACTG1 gene using a −stop+PAM construct. (c) Microscopic images of HEK 293 cells transfected with different integration plasmids. (d) Image quantification of ACTG1-mNeon-positive cells. Shown are mean values+s.e.m. from three independent biological replicates. (e) Deep sequencing analysis of random genomic editing events using primer pairs spanning the targeting region (upper pie charts) and mNeon integration events using primer pairs upstream of the targeting region and within the mNeon gene (lower pie charts). Frame shifts are colour-coded as indicated. The shades of each colour allow to distinguish individual indel events (see legend). Shown are representative results of one out of three independent biological replicates. The percentages indicated are mean values+s.e.m. from three independent biological replicates. ND, not determined. (f) Confocal images of HEK 293 cells transfected with targeting constructs for the human HIST1H4C and TUBB genes. Red arrows indicate an individual chromosome (left panel) or the microtubule-organizing center (right panel).

Figure 2. Modular three-plasmid gene tagging system.

(a) Three-plasmid tagging system. A target selector plasmid expresses an sgRNA targeting a gene of interest. A frame selector plasmid expresses an sgRNA targeting the donor plasmid. A universal donor plasmid contains the tag gene. (b) Sequence details of the universal donor plasmid when integrating into the human ACTG1 gene. (c) Due to a poly-G stretch within the target site of the universal donor plasmid it can be cleaved at three adjacent nucleotide positions, which allows specifying the frame of integration at the time of transfection. (d) Fluorescence imaging and deep sequencing analysis of ACTG1-mNeon gene tagging using a three-plasmid system and different plasmid combinations. ND, not determined. (e) Image quantification of ACTG1-mNeon-positive cells. Shown are mean values+s.e.m. from three independent biological replicates. (f) Immunoblotting of ACTG1-Flag gene tagging using a three-plasmid system and different plasmid combinations.

Figure 3. Flexible frame selection of CRISPaint gene tagging.

(a–c, left panels) Scheme of the exon structure of three genes targeted by CRISPaint. (a–c, right panels) Fluorescence imaging and deep sequencing analysis of ACTG1-mNeon (a) HIST1H4C-mNeon (b) and TUBB-mNeon (c) gene tagging using different frame selector plasmids. The frame selector predicted for in-frame tagging is marked in green. (d) Image quantification of mNeon tagging-positive cells. Shown are mean values+s.e.m. from three independent biological replicates. (e) Immunoblotting result of CRISPaint-mediated tagging of 16 human genes with a 3 × FLAG tag. Indicated below are the predicted protein sizes and frame selectors used.

Figure 4. Efficient enrichment of in-frame tagged cells by antibiotic selection.

(a) Scheme of a three-plasmid tagging system that allows selection for positively tagged cells by expressing a puromycin resistance gene separated from the tag gene by a T2A peptide. (b) Experimental set-up and timeline for selection-based gene tagging in HEK 293T cells. (c) The endogenous gene loci of seven genes were C-terminally tagged in HEK 293T cells according to the strategy outlined in b. After selection, cells were analysed for mNeon fluorescence by fluorescence-activated cell sorting (FACS) and data are depicted as histogram plots. Tagged cells are shown in black, whereas a reference histogram of mock-treated cells is depicted in grey. In addition, cells were subjected to fluorescence imaging of tag gene expression. (d) Experimental set-up and timeline for selection-based gene tagging in the difficult-to-transfect cell line THP1. (e) FACS-based assessment of mNeon fusion-gene expression in THP1 cells after tagging of the endogenous ACTG1 gene and subsequent puromycin selection according to d. (f) Fluorescence imaging of tag gene expression after selection according to d and after additional PMA treatment to induce differentiation. (g) Image quantification of mNeon tagging-positive cells generated according to f. Shown are mean values+s.e.m. from two independent visual fields. Results are representative for two independent experiments.

Figure 5. Minicircle DNA-based insertion tagging.

(a) Inserts of the universal donor 2A-PuroR construct library are subcloned into pMC.BESPX-MCS1 via NheI or SpeI and Bsp120I. The resulting vector can be used to generate a minicircle universal donor construct in an E. coli strain that expresses inducible ϕC31 integrase and I-SceI endonuclease. The resulting minicircle DNA contains the universal donor PuroR cassette and is devoid of bacterial plasmid backbone sequences. On delivery of the minicircle donor construct in conjunction with the CRISPaint plasmid mix, the donor plasmid is cut and integrated into the DSB at the genomic target region. (b) Agarose gel confirming the elimination of plasmid backbone sequences from donor DNA by cultivation of transformed E. coli strain ZYCY10P3S2T with arabinose induction solution for 5 hours before DNA preparation. DNA was linearized with BamHI before loading on the gel. (c) Fluorescence microscopy of HEK 293T cells with an mNeon-2A-PuroR-tagged TUBB gene using minicircle DNA as a donor and selected with puromycin for 4 days. Shown is a representative result from two biological replicates. (d) Quantification of TUBB-mNeon-positive cells using minicircle DNA as a donor and selected with puromycin for four days. Shown are mean values+s.e.m. from two independent biological replicates.

Figure 6. Involvement of cNHEJ repair components in CRISPaint-mediated gene tagging.

(a,d,g) Fluorescence imaging of GFP expression from a control plasmid (upper panels) or ACTG1-mNeon gene tagging (lower panels) in indicated gene-deficient cell lines. (b,e,h) Immunoblot validation of CRISPR-Cas9 generated knockout cell lines that were pre-validated by deep sequencing to bear all-allelic frame shift mutations (knockout (KO) A, B) or heterozygous mutations (−/+). Size marker bands are indicated in the first column; expected protein sizes are given in the last column. (c,f,j) Deep sequencing analysis of the fusion junctions created by NHEJ-mediated gene tagging. (i) Image quantification of ACTG1-mNeon-positive cells. Shown are mean values+s.e.m. from three independent biological replicates.