Variation in zygotic CRISPR/Cas9 gene editing outcomes generates novel reporter and deletion alleles at the Gdf11 locus

Abstract

Recent advances in CRISPR/Cas gene editing technology have significantly expanded the possibilities and accelerated the pace of creating genetically engineered animal models. However, CRISPR/Cas-based strategies designed to precisely edit the genome can often yield unintended outcomes. Here, we report the use of zygotic CRISPR/Cas9 injections to generate a knock-in GFP reporter mouse at the Gdf11 locus. Phenotypic and genomic characterization of founder animals from these injections revealed a subset that contained the correct targeting event and exhibited GFP expression that, within the hematopoietic system, was restricted predominantly to lymphoid cells. Yet, in another subset of founder mice, we detected aberrant integration events at the target site that dramatically and inaccurately shifted hematopoietic GFP expression from the lymphoid to the myeloid lineage. Additionally, we recovered multiple Gdf11 deletion alleles that modified the C-terminus of the GDF11 protein. When bred to homozygosity, most of these alleles recapitulated skeletal phenotypes reported previously for Gdf11 knockout mice, suggesting that these represent null alleles. However, we also recovered one Gdf11 deletion allele that encodes a novel GDF11 variant protein (“GDF11-WE”) predicted to contain two additional amino acids (tryptophan (W) and glutamic acid (E)) at the C-terminus of the mature ligand. Unlike the other Gdf11 deletion alleles recovered in this study, homozygosity for the Gdf11WE allele did not phenocopy Gdf11 knockout skeletal phenotypes. Further investigation using in vivo and in vitro approaches demonstrated that GDF11-WE retains substantial physiological function, indicating that GDF11 can tolerate at least some modifications of its C-terminus and providing unexpected insights into its biochemical activities. Altogether, our study confirms that one-step zygotic injections of CRISPR/Cas gene editing complexes provide a quick and powerful tool to generate gene-modified mouse models. Moreover, our findings underscore the critical importance of thorough characterization and validation of any modified alleles generated by CRISPR, as unintended on-target effects that fail to be detected by simple PCR screening can produce substantially altered phenotypic readouts.