/Products/CRISPR Genome Editing/transEDIT-dual CRISPR/Supporting Data
Here we report the construction of a sequence verified CRISPR collection, whose design integrates both existing and novel effector-selection and effector-expression strategies. This approach is validated through the re-analysis of existing efficacy data and through novel loss-of-function screens. Together these studies demonstrate the significant superiority of this library relative to currently available reagents.

Highlights
  • gRNAs are selected to target conserved coding regions, as well as to induce frame-shift mutations (FSM)
  • Two gRNAs are expressed from each construct to increase functional target ablation
  • Design strategies allow for combinatorial screening
  • Library format allows for multiplexed and array format screening
   

gRNA Selection Strategy

gRNA efficacy is dependent on conservation at the target site and the likelihood of homologous repair driven frame shift mutations.

Figure 1: A) Efficacy of gRNAs when stratified by conservation; B) Efficacy of gRNAs when stratified by FSM likelihood.

Combining these strategies with machine-learning approaches yields a predictive scoring algorithm.

Figure 2: A) Doench et al 2014, gRNAs stratified by CRoatan score; B) Distribution of CRoatan scores in 10X human library.

Negative selection screens validate, empirically the utility of combining conservation and frame-shift mutation estimates for gRNA selection.

Figure 3: A) Depletion rate of essential gene targeting gRNAs in a negative selection CRISPR screen, stratified by conservation pass/fail (Y/N) status; B) Depletion rate of gRNAs in A, stratified by frame-shift likelihood; C) Depletion rate of gRNAs in A, stratified by CRoatan score.

gRNA Expression Strategy

 
CRISPR/Cas9 screens were performed in melanoma A-375 and chronic myelogenous leukemia K-562 cell lines stably expressing Cas9.
gRNAs were screened via viral infection (transduction) at an MOI of 0.3, maintained at a 1000-fold representation of each gRNA. All screens were performed in triplicate. Genomic analysis was carried out using single read NGS (Next Generation Sequencing) and depletion of cells containing knockout of essential genes (EG). Genomic CRISPR/Cas9 cut-type was determined via NGS using paired-end reads.

Single-gRNA screening libraries were created for purposes of comparison of the CRoatan designs to those published by Doench et al. (Harvard, sgRNAscorer), Chari et al. (Broad, GPP Portal) and those used by Edit-R (Dharmacon), (see Figure 4). The single gRNAs were cloned into a 3rd generation lentiviral vector harboring a U6 promoter, a gRNA backbone, and a ZsGreen-P2A-puromycinR transcript driven by an SFFV promoter.
 
For dual gRNA libraries, gRNA sequences were predicted using the CRoatan Algorithm and cloned into a 3rd generation lentiviral vector depicted in Figure 5. The 2 gRNAS were expressed from a single lentiviral vector that harbored two divergent U6 promoters, a 25bp identification barcode, Illumina-adapters, an SFFV promoter and a bicistronic ZsGreen-P2A-PuromycinR transcript. It was reasoned that a higher frequency of deleterious mutations could be inflicted by simultaneously expressing two independent sgRNAs to the same gene target. In gRNA pairings overlapping target regions were avoided to minimize interference, and CRoatan scores were balanced to maximize the number of couples with at least one strong guide (see Figure 6).
Figure 4: Depletion rate of EG- and NEG-targeting constructs was established using the Edit-R (Dharmacon), sgRNAScorer (Harvard), GPP web portal (Broad Institute) versus CRoatan selection strategies.

Figure 6: CRoatan gRNA selection when designing 5 targeting constructs (paired) for each gene target. CRoatan scores were balanced to maximize the number of couples with at least one strong guide and minimal off-target effects.

Figure 5: Vector map of the dual gRNA expression vector with the relevant features highlighted.












































The CRoatan dual-gRNA expression strategy was analyzed using the genomic DNA of cells containing 0, 1 or 2 essential genes (EG). This was achieved by infecting the cells with transEDIT-dual vectors with 2 gRNAs against Non-Essential Genes (NEG) only; one NEG and one EG, or two EG; (thus 0, 1 and 2 EG respectively). Resultant depletion of cells where the essential genes were sufficiently knocked out was established via NGS (see Figures 7 and 8).
 
Dual-gRNA strategy analysis to analyze the cut-types (genomic scars), was carried out via deep sequencing after infection with three NEG targeting constructs. This set was chosen for the short distance between their corresponding targets for NGS proficiency (see Figure 9).

Figure 7: Depletion rate of constructs harboring a 0, 1 or 2 EG (Essential Genes) -gRNAs. gRNAs against olfactory genes were used as NEG (Non-Essential Genes) controls as these genes are considered non-functional post-development.


Figure 8: A heatmap representation of EG-gRNA depletion rates, when paired with NEG-gRNAs (alone) and with 2 gRNAs targeting the same gene (paired). Significant increases in depletion were observed in cells expressing two EG-gRNAs.


Figure 9: Deep-sequencing analysis of genomic scars left by NEG-targeting constructs. hsgRNA (expressed from human U6 promoter) and csgRNA (expressed from chicken U6 promoter); depict insertions/deletions observed with either one or both paired constructs. Mutations were observed for all constructs, with fragment deletions being the predominant scar type.

 

Figure 10: Once genes have been identified as relevant in a research scenario; gRNAs can be re-shuffled for 2
or more genes of interest, to enable multiplexed analysis of multiple gene function via transEDIT-dual.
 
CRoatan algorithm gRNA designs - outstanding knockout efficiency
  • Two independent gRNAs in the same lentiviral vector backbone for potent knockout
  • Unique barcode sequence identifier (25bp) and Illumina Indexes – enable downstream analysis
  • Sequence verified gRNAs for confidence in results
  • Higher frequency of deleterious mutations and minimal off-target effects
  • Efficient delivery: transfect or transduce
  • Combinatorial re-shuffling on finalized gene sets of interest

References:
 
Erard, N., Knott S., & Hannon, G. Submitted for publication.
Doench, J. G. et al. Nature Biotechnology, 34, 184-191, 2016.
Doench, J. G. et al. Nature Biotechnology, 32, 1262-1267, 2014.
Chari, R., Mali, P., Moosburner, M. & Church, G. M. Nature Methods, 12, 823-826, 2015.
Bae, S., Kweon, J., Kim, H. S. & Kim, J. S. Nature Methods, 11, 705-706, 2014.
Shi, J. et al. Nature Biotechnology, 33, 661-667, 2015.