Recombinant adeno associated virus (rAAV) is a precise and effective method to introduce defined changes into endogenous genes and rAAV vectors can stimulate homologous recombination (HR) up to 1000-fold over that seen using plasmids.
While rAAV incorporates targeted changes without the requirement of a double strand DNA-break (DSB) it has been demonstrated that introduction of a DSB by nucleases further increases the rates of targeted incorporation of the rAAV donor1.
Below we detail the approaches we've taking to improving rAAV editing efficiency still futher, including:
- DNA mismatch repair pathway inhibition by RNA interference
- incorporation of negative selection into rAAV donors
- combination of rAAV and CRISPR with the above approaches
Measuring homologous recombination rates
To measure rates of HR and targeting frequencies, we generated cell lines with an integrated GFP reporter gene inactivated by a nonsense mutation to prevent fluorescence. Donor molecules could then be used to integrate the correct sequence to revert the mutation and switch-on GFP, enabling rates of HR to be measured by flow cytometry.
Digital droplet PCR (ddPCR) was also used to measure the abundance of a mutation introduced by a donor via HR in a pool of cells using locus specific primers and probes.
Figure 1. GFP reporter for measuring homologous recombination. Schematic representation of the GFP reporter inserted into the endogenous HPRT allele in HCT116 and HEK293 cells. GFP is split into exons and a nonsense mutation is introduced in exon 3 to render the protein inactive. gRNA sites close to the mutated STOP codon enable targeted cleavage of the exon by Cas9. Plasmid, ssODN and rAAV donors are designed to switch on GFP by correcting the nonsense mutation and allow identification of cells undergoing correct HR by FACS.
DNA mismatch repair (MMR) suppression during homologous recombination
The DNA MMR pathway has been implicated in the inhibition of gene editing, potentially through the suppression of homologous recombination between genetically different sequences2.
We found an increase in targeting frequency of rAAV donors in cells with intact MMR pathways in which expression of the important MMR gene MSH2 had been down regulated by RNAi (Figure 2)
Figure 2. MMR-proficient cells show increased rates of rAAV gene editing when MSH2 is suppressed by RNAi. The targeting frequency of the BRAF V600E mutation was measured by ddPCR 72 hours after infection. HCT116 and DLD1 cells known to be already MMR-deficient showed marginal or no improvement in the frequency of gene editing whereas siRNA knockdown of MSH2 in MMR-proficient MCF10a cells gave a more significant increase.
Gene editing efficiency improvements with negative selection
shRNA expressions cassettes to knockdown genes required for cell survival, or the donors own positive selection gene, are placed at one or both ends of the vector outside of the region homologous to the target gene. Cells that have undergone HR will not integrate the shRNAs, whereas cells in which the construct integrates randomly will retain them and be eliminated via its expression.
We have found that incorporation of these negative selection elements in combination with positive selection using a drug resistance marker increases the targeting frequency of rAAV (Figure 3)
Figure 3. rAAV vectors incorporating shRNA negative selection increase rates of gene editing. (A) Typical rAAV vector incorporating shRNA cassettes (negative selection) outside the region of homology. (B) Single or dual shRNA cassettes increase targeting frequencies of the GFP reporter allele in HCT116 cells. (C) and (D) Targeting efficiency of the CDK2 D145N kinase-dead mutation in HCT116 cell line is increased using vectors with shRNA cassettes against endogenous HPRT (selected against using HAT containing media).
Combining MMR suppression and negative selection vectors
MSH2 suppression using siRNA followed by targeting with rAAV vectors incorporating shRNA negative selection gave an increased rate of gene targeting in the MMR-proficient cell lines MCF10a and NCI-H838. The improved rate of editing was larger than seen with either technique when applied alone (Figure 4).
Figure 4. Increase in gene targeting frequency by combination of MSH2 suppression and negative selection. Cells were transfected with MSH2 siRNA, or left untransfected (NT) and cultured for 72 hours before infection with either BRAF V600E (left) or EGFR T790M (right) virus. No shRNA vectors (white bars) and single HPRT shRNA negative selection vectors (black bars) were tested. Cells were selected in G418 and HAT in combination for 2-weeks, then analysed for targeted allele frequency by locus specific ddPCR.
The combination of MSH2 suppression and rAAV vectors incorporating shRNA negative selection has been used to engineer mutant isogenic cell lines which could not be isolated without these improvements.
Figure 5. Successful engineering of the AKT E17K mutation in MCF10a cells using rAAV vectors with single shRNA negative selection cassettes plus MSH2 suppression
Gene editing with CRISPR and improved rAAV donors
The use of CRISPR with rAAV donors enables highly efficient gene editing. This combination can be further improved using rAAV donors with shRNA negative selection.
Figure 6. rAAV targeting vectors incorporating shRNA negative selection increase targeting frequencies of the GFP reporter allele in HCT116 cells when combined with CRISPR.
Vectors with single or dual shRNA cassettes targeting the Neomycin resistance marker were infected in cells that had been transfected with plasmids expressing Cas9 and gRNA to introduce double stand breaks in the GFP exon. Cells were selected for 2 weeks in the presence of G418 and then analysed by FACS
Note: Transfection efficiencies are also being measured to determine if the increase might also be related to expression levels of Cas9/gRNA
The use of MSH2 suppression during CRISPR + rAAV targeting did not show any significant increase in targeting frequencies in either the MMR-deficient HCT116 line or the MMR-proficient NCI-H838 lung carcinoma line. The same was observed when using both ssODN and dsDNA plasmids as donors (Figure 7).
Figure 7. MSH2 suppression gives no significant benefit to targeting frequencies using CRISPR combined with different donors for HR. HCT116 (left) and NCI-H838 (right) GFP reporter cells in which MSH2 expression had been knockdown by siRNA were analysed for GFP positive cells 72 hours after transfection/infection with CRISPR and indicated donor
We have shown the increase in gene editing efficiency of rAAV donors when incorporating shRNA negative selection and supressing MSH2 gene expression during HR. When these improvements are used in combination, the boost in targeting frequency has improved the success of generating isogenic cell lines with defined genetic mutations.
The use of rAAV donors in combination with CRISPR further enhances gene targeting rates and while MSH2 suppression of CRISPR/rAAV treated cells did not further elevate this, donors incorporating shRNA negative selection did increase the proportion of correctly targeted cells further
1. Miller et al Human Gene Targeting by Adeno-Associated Virus Vectors Is Enhanced by DNA Double-Strand Breaks. Mol. Cell. Biol. May 2003 vol. 23 no. 10 3550-3557
2. Evans and Alani. Roles for Mismatch Repair Factors in Regulating Genetic Recombination. Mol Cell Biol. 2000 Nov; 20(21): 7839–7844