Enzymatically Generated CRISPR Libraries for Genome Labeling and Screening

Abstract

CRISPR-based technologies have emerged as powerful tools to alter genomes and mark chromosomal loci, but an inexpensive method for generating large numbers of RNA guides for whole genome screening and labeling is lacking. Using a method that permits library construction from any source of DNA, we generated guide libraries that label repetitive loci or a single chromosomal locus in Xenopus egg extracts and show that a complex library can target the E. coli genome at high frequency.

Figure 1. Repetitive genomic loci can be visualized using dCas9-Neon in Xenopus egg extracts

A: dCas9-Neon is programmed to label specific genomic loci by conjugation to an sgRNA molecule containing a complementary target sequence. See also Figure S1. B: dCas9-Neon programmed using RHM2 sgRNA (black) localizes rapidly to loci in sperm nuclei (Sytox Orange dye, magenta). Time (min) after imaging started is indicated in the top left of each image. See also Supplementary Movie 1 and Figure S2. C: Labeled RHM2 loci (green) are maintained following formation of a mitotic spindle (red). D: Three examples of repeat classes labeled on sperm nuclei in Xenopus egg extract (1n = 18). Left: RHM2 is a centromere-proximal locus on ~65% of chromosomes (Freeman and Rayburn, 2005). Middle: Telomere repeats target chromosome termini. Right: REM3 is reported to target a single centromere-proximal locus on chromosome 1, appearing here as two spots (Hummel et al., 1984). E: Left: Sperm nuclei driven into interphase in the presence of dCas9-tdTomato Telomere sgRNA and dCas9-Neon RHM2 sgRNA demonstrate simultaneous dual-color labeling (scale bar, 5 µm). Right: A subset of RHM2 and telomere loci appear to co-localize, while others do not (scale bars 10 µm, except magnification in panel E, 1 µm).

Figure 2. An enzymatically generated guide library can program dCas9-Neon labeling of a repetitive locus

A: Outline of enzymatic library generation approach. B: dCas9-Neon programmed using an RHM2 repeat unit processed with this method localizes in a labeling pattern similar to that seen for RHM2 in Figure 1B and 1D (scale bar, 5 µm).

Figure 3. A single 3.4 MB locus can be labeled using an enzymatically generated guide library

A: Specificity score distribution for all guides predicted to be generated by subjecting 3.4 MB region to procedure outlined in Figure 2A. Only sub-regions predicted to generate guides with a score of ≥95 were used as PCR templates for library construction. B: Processing of 100 PCR products (See Figure S3) spanning regions within a 3.4MB region of X. laevis chromosome 4 generates a single labeled spot in haploid sperm nuclei (scale bar, 5 µm). C: Count of fluorescent foci per sperm nucleus when incubated with 3.4 MB library. n = 3 experiments, 11–13 nuclei scored per experiment. Bars are ± standard deviation. See also Figure S3, Table S3, and Supplemental Data S1–S3.

Figure 4. A complex guide library targeting sequences within the E. coli genome

A: Comparison of theoretical maximum number of guides generated by E. coli genome digestion with guides identified by sequencing (black text) and of sequencing reads that represent expected guides versus those reads that do not correctly target E. coli PAM-adjacent 20mers (blue text). B: Length distribution of variable spacers (region between T7 promoter and sgRNA guide body) in library as determined by high-throughput sequencing. C: Distribution of abundance of unique guides within library. D: Coverage of selected GO-term gene groups by library sgRNAs compared to the total number of genes annotated by those GO terms. E: Analysis of genes targeted by guides in sequenced library as binned by gene length. F: In silico analysis of guide specificity as predicted to be produced by digestion/ligation of E. coli genomic DNA. A score of 100 indicates no predicted off-target effects.