CRISPR/Cas9 Methodology for the Generation of Knockout Deletions in Caenorhabditis elegans

Abstract

The Caenorhabditis elegans Gene Knockout Consortium is tasked with obtaining null mutations in each of the more than 20,000 open reading frames (ORFs) of this organism. To date, approximately 15,000 ORFs have associated putative null alleles. As there has been substantial success in using CRISPR/Cas9 in C. elegans, this appears to be the most promising technique to complete the task. To enhance the efficiency of using CRISPR/Cas9 to generate gene deletions in C. elegans we provide a web-based interface to access our database of guide RNAs (http://genome.sfu.ca/crispr). When coupled with previously developed selection vectors, optimization for homology arm length, and the use of purified Cas9 protein, we demonstrate a robust and effective protocol for generating deletions for this large-scale project. Debate and speculation in the larger scientific community concerning off-target effects due to non-specific Cas9 cutting has prompted us to investigate through whole genome sequencing the occurrence of single nucleotide variants and indels accompanying targeted deletions. We did not detect any off-site variants above the natural spontaneous mutation rate and therefore conclude that this modified protocol does not generate off-target events to any significant degree in C. elegans We did, however, observe a number of non-specific alterations at the target site itself following the Cas9-induced double-strand break and offer a protocol for best practice quality control for such events.

Keywords: C. elegans; CRISPR/Cas9; homology dependent repair; mutagenesis.

Figure 1

Generation of a deletion using the CRISPR/Cas9 protocol. (1) Guide RNAs direct Cas9 to create targeted DSBs in the gene of interest. (2) Through HDR, a portion of the ORF is replaced with the selection cassette containing pharyngeal GFP and G418 resistance (neoR) markers. (3) The dual-marker cassette, flanked by loxP sites, can be excised from the genome by injecting Cre recombinase. (Adapted from Norris et al. 2015.)

Figure 2

Assessing editing efficiency with homology arms of varying lengths. A 536 bp deletion was generated in rap-3 using a single guide RNA and various lengths of homology arms (Table S1). The proportion of individual injected P₀s giving rise to animals with the selection cassette integrated at the desired location in the genome was determined by PCR validation. Error bars indicate standard error of the mean. Statistical significance between groups was determined using the Chi-squared test, without correction for multiple testing.

Figure 3

Comparison of Cas9 delivery methods. A direct comparison of editing efficiency between Cas9 plasmid and protein was done across three genes, using two guide RNAs for each gene (Table S1). Each P₀ plate contained four injected animals, and F2 progeny were screened for selection cassette integration at the desired location using PCR. Overall, the average editing efficiency for the three genes targeted with Cas9 protein was significantly higher (Chi-squared test, P = 2e-5) than that of the same three genes targeted with Cas9 plasmid by more than a factor of three.

Figure 4

Rearrangements at the CRISPR/Cas9 target site. Two examples of removing the target gene through CRISPR/Cas9 HDR visualized using IGV. In strain VC3671, there is a perfect deletion removing the first two exons of gene F01D4.9. In strain VC3674, there is evidence of a complex rearrangement. A large deletion encompasses the same two exons as well as the downstream region of F01D4.5 and a pseudogene, F01D4.3. This region likely harbors multiple copies of the repair template, since the average coverage of the homology arms are relatively high compared to the adjacent genomic sequence. At the bottom of the figure, the exons for the various genes are shown in blue, the homology arms chosen for F01D4.9 are shown in green, and the guide RNA is in red.

Figure 5

Complex rearrangements at the CRISPR/Cas9 target site. Whole genome sequencing of VC3743, a CRISPR-generated strain, revealed an unintended 16 kb deletion that disrupts the target gene as well as eight other genes in the surrounding area. This deletion is also accompanied by a complex rearrangement that consists of fragments of genomic sequence from the local region and the repair template. The rearrangement joins (1) the 5′ portion of srh-281 to a duplicated fragment of oac-15, followed by (2) a duplicated inverted portion of oac-15, (3) an inverted intergenic region, and (4,5) the selection cassette flanked by homology arms. In blue are the exons for the various genes and in pink are the homology arms chosen for F11A5.3. The marker for the guide RNA is not visible at this scale.

Figure 6

Homology-directed repair resolves CRISPR/Cas9 double-strand breaks in unpredictable ways. Using the PCR validation scheme described in Table 2, analysis of 330 CRISPR/Cas9-derived mutants (Table S3) reveals that integration of the selection cassette via HDR does not always occur as predicted. These mutant strains were generated using up to two guide RNAs and either Cas9 protein or a Cas9 expression plasmid. The proportion of integrant strains belonging to each mutant class within each gene target was determined and the average proportions for 81 gene targets were calculated.