Background

By creating targeted chromosomal DNA double strand breaks (DSBs) or single strand breaks (nicks) or serving as a DNA binding module for other DNA modification effectors, programmable endonucleases have become an important tool for genome modification in eukaryotic cells (Gaj et al., 2013). In response to targeted DNA breaks, host cells can invoke various repair pathways to mend the damages to maintain the genome integrity. Insertions and/or deletions derived from NHEJ repair errors can be capitalized for gene knockout and homologous recombination can be exploited for introducing pre-determined changes on gene of interest by providing a DNA donor. In addition to these more traditional gene editing applications, catalytically inactive forms of programmable endonucleases are increasingly used as DNA binding modules for other DNA modification effectors, such as cytidine deamination enzymes (Komor et al., 2016). However, no matter which forms of programmable nucleases are utilized, target site binding is the prerequisite step and local chromatin structure can determine whether or not or how efficiently a programmable nuclease can bind the target site (Knight et al., 2015; Hinz et al., 2015; Horlbeck et al., 2016; Isaac et al., 2016, Chen et al., 2017). We hypothesize that binding at proximal locations by a programmable DNA binding protein could change the local chromatin structure and render an otherwise inaccessible target site accessible for binding.

Previous generations of programmable nucleases, such as meganucleases, zinc finger nucleases (ZFNs), and transcription activator-like effector nucleases (TALENs), solely rely on protein structure to recognize target sites, and thus re-targeting of these nucleases requires rather laborious protein structural change. In contrast, class 2 CRISPR-Cas effector nucleases use protein structure to recognize a protospacer adjacent motif (PAM) and employ CRISPR RNA (crRNA) to bind the target site adjacent to the PAM. Because PAM is typically a short DNA sequence, such as 5’-NGG-3’ for SpCas9, and thus occurs frequently in a genome, re-targeting of CRISPR-Cas nucleases is a simple process of changing the crRNA sequence by molecular cloning or chemical synthesis. This targeting modality makes CRISPR-Cas systems very suitable for use as nucleases or as DNA binding proteins. This protocol combines these two utilities together to expand CRISPR gene editing capability. The CRISPR-Cas system used as nuclease must be orthogonal to the CRISPR-Cas system used as DNA binding protein to avoid binding site sharing. In general, different subtypes of class 2 CRISPR-Cas systems (e.g., type II-A, type II-B, type II-C, and type V) are orthogonal to one another. Within each subtype, some systems are highly divergent and could be also orthogonal to one another, but they need to be experimentally verified. Currently, it is highly recommended to use SpdCas9 as proximal DNA binding protein, for SpCas9 is the most robust system in mammalian cells to date, although it can also be inactive at certain genomic sites. However, it is anticipated that more robust Cas9 systems will be developed for use as DNA binding proteins.