CRISPR-Cas9 Explained: How It Works, Guide RNA and PAM

 

CRISPR-Cas9 Explained: From Bacterial Immunity to Genome Editing

CRISPR-Cas9 has gone from a curiosity in bacterial genomes to one of the most powerful tools in molecular biology. In this post, we look at where it came from, how the system works, and the key terms you will meet when you start designing guide RNAs, using sickle cell disease as our worked example.

Where CRISPR came from

The story starts with an odd observation. Researchers sequencing microbial genomes (bacteria and archaea) kept finding clusters of short, repeated sequences, first noticed in E. coli in 1987. The repeats were separated by short, unique stretches of DNA called spacers. The name CRISPR, short for clustered regularly interspaced short palindromic repeats, was coined in 2002. (The repeats are only partly palindromic, but the name stuck.)

The breakthrough came in the mid-2000s, when several groups realised that the spacers match the genomes of bacteriophages (viruses that infect bacteria) and plasmids. That led to a striking idea: spacers are a molecular memory. When a bacterium survives an infection, it stores a snippet of the invader's DNA as a new spacer, flanked by repeats. In 2007, experiments confirmed that CRISPR is an adaptive immune system in bacteria.

How the bacterial system works

The CRISPR array is transcribed and processed into short crRNAs, each carrying one spacer plus part of a repeat. In the type II system, each crRNA pairs with a second RNA, the tracrRNA, and the RNA pair loads onto a Cas (CRISPR-associated) protein. If the same virus returns, the crRNA recognises its matching DNA and the Cas protein cuts it, stopping the infection.

There are many Cas proteins, and they are numbered. Not all of them cut DNA: Cas13, for example, targets RNA, and Cas1 and Cas2 insert new spacers into the array. The best known and most widely used is Cas9, a DNA-cutting enzyme. In 2012, Jinek, Charpentier, Doudna and colleagues showed that Cas9 could be programmed to cut any DNA sequence of choice, and Charpentier and Doudna shared the 2020 Nobel Prize in Chemistry for it.

Key terms

Cas9. An RNA-guided DNA endonuclease. On its own it cannot choose where to cut. The guide RNA tells it.

Guide RNA (gRNA). In bacteria, the guide is two separate RNAs (crRNA and tracrRNA). In the lab, they are fused into a single sgRNA of about 100 nucleotides (nt). Because RNA is single-stranded, we count it in nucleotides, not base pairs.

The 20-nt targeting sequence. The 5' end of the sgRNA, 20 nt long, is the part you design. It pairs by complementarity with one strand of the target DNA, the target strand. Change those 20 nt and you re-aim Cas9 at a different place in the genome.

PAM (protospacer adjacent motif). A short sequence next to the target that Cas9 must see before it will cut. For the Cas9 from Streptococcus pyogenes, the PAM is NGG (N is any base), located immediately 3' of the 20-nt target on the non-target strand. Cas9 recognises the PAM first, then unwinds the DNA and checks whether the guide matches. This is also why the PAM limits where you can edit: you can only target sites that have an NGG next to them.

What Cas9 actually does

Once Cas9 is bound to the guide and the target, it makes a blunt double-strand break about 3 bp upstream of the PAM. It is important to understand that Cas9 only cuts. The edit itself comes from the cell's own repair machinery:

• NHEJ (non-homologous end joining) is fast and error-prone, and often leaves small insertions or deletions (indels) that disrupt a gene.
• HDR (homology-directed repair) can copy a supplied donor template to make a precise change, but it is much less efficient, particularly in non-dividing cells.

Which pathway wins will shape what kind of edit you can realistically make.

Worked example: sickle cell disease

Sickle cell disease is a good model for thinking about guide design because the cause is a single, well-characterised mutation. It lies in the gene for the haemoglobin subunit beta, HBB. A single adenine is replaced by a thymine (A>T; c.20A>T), which changes the codon GAG to GTG on the coding strand. As a result, glutamic acid is replaced by valine at position 6 of the beta chain (Glu6Val, or p.Glu7Val in current HGVS nomenclature, which counts the initiator methionine). That one amino acid change makes haemoglobin S polymerise when deoxygenated, distorting red blood cells.

The first step in designing a guide is to get the sequence:

1. Go to NCBI Entrez and search for the gene responsible for the disease, in this case HBB.
2. Open the gene record and retrieve the reference sequence around the mutation site.
3. Look for NGG PAM sites close to the mutation, then choose a 20-nt sequence next to each one as a candidate guide.

In the next post, we will go through choosing and checking candidate guides.

A reality check on therapy

A word of caution before anyone imagines Cas9 simply "fixing" the sickle cell mutation. Correcting a point mutation by HDR is inefficient, and that is why the first approved CRISPR therapy for sickle cell disease (Casgevy, approved in 2023) takes a different route. Rather than correcting HBB, it cuts the erythroid enhancer of BCL11A to switch fetal haemoglobin back on, which compensates for the faulty adult haemoglobin. Base editors cannot directly revert T to A, but adenine base editors have been used to convert the sickle codon into a benign variant (Hb G-Makassar), and prime editing offers another route to the precise correction.

Key takeaways

• CRISPR is a bacterial adaptive immune system: spacers store a memory of past infections.
• Cas9 is guided by an RNA; in the lab, the two bacterial RNAs are fused into a single sgRNA of about 100 nt.
• The first 20 nt of the sgRNA are the part you program, and the target must sit next to an NGG PAM.
• Cas9 cuts about 3 bp from the PAM; the cell's repair pathways, NHEJ or HDR, decide the final edit.
• Sickle cell disease comes from a single A>T change in HBB, which makes it an excellent worked example.