Gene knockout in cultured cells with CRISPR–Cas9

CRISPR–Cas9 is a way to make a cut at a chosen place in the genome. A short guide RNA carries about 20 bases that match the target, and it directs the Cas9 protein there; Cas9 then cuts both strands of the DNA. The protein will only cut next to a short signal in the DNA called the protospacer-adjacent motif, or PAM, which for the common Cas9 of Streptococcus pyogenes (SpCas9) is the three-letter pattern NGG. The cut is made about three base pairs away from the PAM. Jinek and colleagues showed that a dual RNA guides Cas9 to cleave matching DNA and that the two RNAs can be fused into one single guide RNA (Jinek et al., 2012), and the system was soon shown to cut genes inside human and mouse cells (Cong et al., 2013).

The cell repairs the cut in one of two main ways. Non-homologous end joining (NHEJ) rejoins the broken ends quickly but imprecisely, often adding or losing a few bases. Such a small insertion or deletion, an indel, whose length is not a multiple of three shifts the reading frame of the gene, which usually scrambles everything downstream and introduces an early stop codon. That is how a gene is knocked out. If instead a matching DNA template is supplied, the cell can sometimes use homology-directed repair (HDR) to copy a precise change, such as a single point mutation, from the template; HDR is much less frequent than NHEJ.

Choosing the guide is the step that most affects the result. A guide is preferred that cuts in an early coding exon, so that a frameshift disrupts most of the protein; that has a middling GC content and no run of four or more T bases (which would stop the RNA polymerase that makes the guide); and that has few close matches elsewhere in the genome, because Cas9 can also cut at similar sequences. These off-target sites matter most when the mismatches lie far from the PAM, where Cas9 tolerates them more readily (Hsu et al., 2013).

This session uses a clearly labelled artificial gene and genome, not a real sequence, and works only in cultured cells. After editing, the DNA is harvested and the cut site is amplified by the polymerase chain reaction. A quick T7 endonuclease I (T7E1) assay gives a rough editing rate, amplicon sequencing gives the exact spectrum of indels and the reading frame, single-cell clones are isolated to obtain a defined line, and loss of the protein is confirmed by Western blot. A non-targeting guide and an untreated sample are run alongside as controls throughout.

Learning objectives

• Read a gene map and explain why an early coding exon is the usual target for a knockout
• Find NGG PAM sites, choose an SpCas9 guide RNA, and weigh its on-target features against its predicted off-target sites
• Set up a cultured-cell editing experiment with a non-targeting guide and an untreated control, and say what each control rules out
• Distinguish repair by NHEJ, which leaves indels and can shift the reading frame, from HDR with a donor template, which installs a precise change
• Measure editing by the T7E1 assay and by amplicon sequencing, and explain when the T7E1 estimate misleads
• Isolate and genotype a knockout clone, confirm loss of protein by Western blot, and explain why a frameshift does not always abolish the protein

Ethics and safety

Making and using genetically modified cells is regulated work. It requires approval from an Institutional Biosafety Committee (in India, under the Rules, 1989 of the Environment (Protection) Act 1986, with comparable national rules elsewhere) and is carried out at the containment level set for the cells and the reagents; viral delivery vectors, where used, have their own approvals and handling rules. Human cell lines carry their own requirements: they should come from a recognised source with evidence of donor consent and appropriate ethical approval, and work with them follows the national rules (in India, the ICMR National Ethical Guidelines of 2017). Good practice includes authenticating the cell line and testing it for mycoplasma, keeping accurate records of every edited line, and disposing of genetically modified material as the local rules require. This simulation does not involve animals, human germline or embryo editing, pathogens, or gene drives, and none of those is covered here.

Key terms

• Guide RNA (sgRNA): The short RNA that directs Cas9 to its target. About 20 of its bases match the DNA to be cut; the single guide RNA (sgRNA) fuses the two natural RNAs into one molecule.
• SpCas9: The Cas9 nuclease of Streptococcus pyogenes, the most commonly used Cas9. It cuts both strands of DNA next to an NGG PAM.
• PAM (NGG): The protospacer-adjacent motif: a short sequence that must lie right next to the target for Cas9 to cut. For SpCas9 it is NGG, where N is any base.
• Protospacer: The 20-base stretch of genomic DNA that matches the guide and sits immediately next to the PAM.
• Double-strand break: A cut through both strands of the DNA, made by Cas9 about three base pairs from the PAM; the cell must repair it.
• NHEJ: Non-homologous end joining: the fast, error-prone repair that rejoins the broken ends and often leaves a small indel.
• HDR: Homology-directed repair: precise repair that copies a change from a supplied template; much rarer than NHEJ.
• Indel: A small insertion or deletion. If its length is not a multiple of three it shifts the reading frame.
• Frameshift: A change in the reading frame caused by an indel that is not a multiple of three, which usually introduces an early stop codon and disrupts the protein.
• ssODN donor: A single-stranded oligonucleotide carrying the wanted change, supplied as the template for HDR.
• T7E1 assay: T7 endonuclease I assay: the cut region is amplified, melted and re-annealed, and the enzyme cuts where edited and unedited strands pair imperfectly; the cut fragments are seen on a gel as a rough editing rate.
• Off-target site: A place in the genome similar enough to the target that the guide binds and Cas9 may cut there too; most tolerated when the mismatches lie far from the PAM.
• Clonal isolation: Growing a line from a single cell, so that every cell shares the same edit, and genotyping the clone before using it.
• Loading control: A protein expected to be equal in every lane of a Western blot (such as actin), used to check that equal amounts were loaded.