A genome-wide CRISPR library for high-throughput genetic screening in Drosophila cells

Abstract

The simplicity of the CRISPR/Cas9 system of genome engineering has opened up the possibility of performing genome-wide targeted mutagenesis in cell lines, enabling screening for cellular phenotypes resulting from genetic aberrations. Drosophila cells have proven to be highly effective in identifying genes involved in cellular processes through similar screens using partial knockdown by RNAi. This is in part due to the lower degree of redundancy between genes in this organism, whilst still maintaining highly conserved gene networks and orthologs of many human disease-causing genes. The ability of CRISPR to generate genetic loss of function mutations not only increases the magnitude of any effect over currently employed RNAi techniques, but allows analysis over longer periods of time which can be critical for certain phenotypes. In this study, we have designed and built a genome-wide CRISPR library covering 13,501 genes, among which 8989 genes are targeted by three or more independent single guide RNAs (sgRNAs). Moreover, we describe strategies to monitor the population of guide RNAs by high throughput sequencing (HTS). We hope that this library will provide an invaluable resource for the community to screen loss of function mutations for cellular phenotypes, and as a source of guide RNA designs for future studies.

Keywords: CRISPR/Cas9; Drosophila; Genome-wide library.

Fig. 1

Design of a genome-wide sgRNA library. A: Strategy for library design. Fragments of coding exons shared between the maximum number of transcriptional isoforms were selected, and all possible sgRNAs of the format N₂₀NGG were designed on both strands. Exons (blue boxes), transcriptional start sites (arrows) and untranslated regions (thick blue lines) are indicated. sgRNAs were selected based on the absence of any off-targets with less than three mismatches, and their position early in the coding sequence. Five non-overlapping sequences were selected. B: Example of designs. A screenshot from the UCSC browser shows designs for a typical gene.

Fig. 2

Cloning of sgRNA library. A: sgRNA expression vector. sgRNAs (blue) are expressed from a Drosophila U6:2 promoter, along with the Cas9 protein from an Actin-5C promoter. Cas9 (red box) contains N- and C-terminal nuclear localisation sequences (NLS, grey oval), and is expressed as a bicistronic transcript with a puromycin N-acetyltransferase gene (purple oval) separated by a viral 2A peptide (black oval). An SV40 transcriptional terminator is also included (SV40 term). B: Oligo synthesis. sgRNA sequences were synthesised with common 5′ and 3′adaptors, and amplified by PCR followed by digestion with restriction enzymes and cloned into the expression vector. C: Cloning strategy for sgRNAs. The synthesised oligonucleotides were amplified by PCR using common adaptor sequences, and digested with the BspQ I restriction enzyme, followed by ligation into a similarly digested expression vector. The first base transcribed by the dU6:2 promoter (G) is indicated by an arrow.

Fig. 3

Analysis of cloned sgRNA library. A: Pie chart of the sgRNA sequences represented in the cloned library. Those sequences represented by 0 read (purple), 1 read (dark blue), 2–5 reads (mid blue) and more than 5 reads (light blue) are indicated. B: Histogram of the number of sgRNAs per gene in the cloned library. Total number of genes with designs is 13,668, and the number of genes with at least one sgRNA is 13,501. C: Example of cloned sgRNA. Screenshot from UCSC browser shows designed sgRNAs and cloned sgRNAs at a typical gene (CG2219).

Fig. 4

Screening and quantification strategy. A: Screening strategy. Cells were transfected with the library (coloured circles), and selected in puromycin to enrich for transfected cells followed by growth for 1, 4 and 10 days (d). sgRNAs were quantified by PCR and high throughput sequencing (HTS). B: Amplification of sgRNAs from cells. sgRNA sequences were amplified by PCR using common flanking sequences to obtain an 84 bp product. A second round of PCR was performed to add adaptors. These included the sequences required for amplification prior to sequencing (P5, light red, P3, green) and sequencing primer binding sites (seq1, dark red, seq2, orange) and included two barcodes (BC1, light blue, BC2, dark blue). BC1 is of variable length to increase sequencing library diversity.

Fig. 5

Optimisation of library screening conditions and pilot screen. A: Principal component analysis (PCA). sgRNA distributions in each condition were anlaysed by PCA. PC1 explained 98.6% and PC2 1.8% of the variance in the samples. B: Heat map of gene enrichment and depletion. Unsupervised hierarchical clustering of log₂ fold change in sgRNA abundance for genes across different conditions. a and b correspond to biological replicates. Note that this does not include the entire gene set analysed, only the region that shows enrichment or depletion. C: Differential sgRNA abundance analysis. DESeq2 was used to identify statistically significant changes in sgRNA counts for all genes with ≥3 sgRNAs between samples at 1 day and 10 days (d) at a 1:100 dilution. MA plot (left panel) shows log₂ fold change against sgRNA counts, with significant changes highlighted in red. Right panel shows the significantly enriched (red) or depleted (purple) genes, and functional enrichment as determined by DAVID. Lysosomal genes are indicated in bold red type.