Cas9 is the molecular scissors of the CRISPR system. It’s a protein that cuts both strands of a DNA molecule at a precise location, guided by a short piece of RNA that matches the target sequence. This ability to make programmable, site-specific cuts in DNA is what transformed CRISPR from a curiosity of bacterial biology into the most widely used gene-editing tool in history, earning Emmanuelle Charpentier and Jennifer Doudna the 2020 Nobel Prize in Chemistry.
How Cas9 Finds Its Target
Cas9 doesn’t work alone. Before it can do anything, it needs to load a piece of guide RNA, a short molecule whose sequence matches the stretch of DNA you want to edit. In nature, bacteria produce two separate RNA molecules (called crRNA and tracrRNA) that pair together and slot into Cas9. In the lab, scientists typically fuse these into a single synthetic guide RNA, which simplifies the process without changing how Cas9 functions.
Once loaded with guide RNA, the Cas9 protein changes shape and begins scanning along DNA, looking for a match. But it doesn’t check every base pair it encounters. Instead, it first looks for a short signature sequence right next to the target site called a PAM (protospacer adjacent motif). For the most commonly used version of Cas9, from the bacterium Streptococcus pyogenes, the PAM is just three letters: NGG, where N can be any base. Only after recognizing this motif does Cas9 pry open the double helix and check whether the adjacent DNA matches the guide RNA. This two-step process, PAM recognition followed by sequence matching, is what gives Cas9 both its speed and its specificity.
How Cas9 Cuts DNA
The Cas9 protein is large, composed of 1,368 amino acids, and contains two distinct cutting domains that work together to sever both strands of the DNA double helix. The HNH domain cuts the strand that pairs with the guide RNA, always at a fixed position. The RuvC domain cuts the opposite strand using a different strategy: it measures its cut site from the PAM, acting like a molecular ruler. Together, these two cuts produce a clean double-strand break.
This double-strand break is the critical event. DNA with a broken backbone triggers the cell’s emergency repair systems, and the type of repair that follows determines the outcome of the edit.
What Happens After the Cut
Cells have two main options for repairing a double-strand break. The first, and more common, is a quick-and-dirty process that simply glues the broken ends back together. This repair is error-prone: the cell often adds or removes a few DNA bases at the break site, creating small insertions or deletions (called indels). If this happens inside a gene, the indels usually scramble the gene’s reading frame and effectively disable it. This is how researchers use Cas9 to “knock out” a gene.
The second repair path is slower and more precise. The cell uses a nearby template, normally its second copy of the chromosome, to rebuild the broken region accurately. Scientists exploit this by supplying an artificial DNA template alongside Cas9. The cell copies the template into the break site, allowing researchers to insert a new gene, correct a mutation, or make any other specific change they want. This precise editing is harder to achieve because cells strongly prefer the fast, error-prone repair, especially in non-dividing cells.
Modified Versions of Cas9
One of Cas9’s most versatile spin-offs doesn’t cut DNA at all. By disabling both the HNH and RuvC cutting domains, scientists created a “dead” version called dCas9. This deactivated protein still binds to its target with the same precision, but it just sits there without making a cut. That turns it into a programmable DNA-binding platform that can carry other molecular tools to any spot in the genome.
Fusing dCas9 to a gene-silencing protein creates a system called CRISPRi, which blocks a gene’s activity by physically preventing the cell’s transcription machinery from reading it. Fusing it to a gene-activating protein creates CRISPRa, which does the opposite, boosting a gene’s output. Neither system alters the DNA sequence itself, making them useful for studying gene function without permanent changes.
Another important variant is the Cas9 “nickase,” where only one of the two cutting domains is disabled. A nickase cuts just one strand of DNA instead of both, which dramatically reduces unintended edits at the wrong locations. Pairing two nickases that target nearby sites on opposite strands can produce an effective double-strand break with far greater accuracy than a single standard Cas9.
Off-Target Cutting and Accuracy
Cas9’s biggest limitation is that it sometimes cuts DNA at sites that closely resemble, but don’t exactly match, its target. Sequences with as few as three to five mismatched bases in the region farthest from the PAM can still be cut, though the frequency drops as mismatches increase. Three or more mismatches generally block the HNH domain from reaching its active shape, which inhibits cleavage, but doesn’t always prevent it.
Several strategies reduce these off-target effects. Designing guide RNAs with a balanced GC content between 40% and 60% stabilizes binding at the correct site while discouraging binding elsewhere. Shortening the guide RNA to fewer than 20 nucleotides also helps, trimming off-target activity without sacrificing on-target performance. Chemical modifications to the guide RNA backbone can further suppress unwanted cuts while maintaining editing efficiency at the intended site.
On the protein side, engineered high-fidelity variants of Cas9 have been developed. One such variant retains on-target activity comparable to the original Cas9 with more than 85% of guide RNAs tested in human cells, while sharply reducing off-target events. Using Cas9 versions that require a longer or rarer PAM sequence also helps, since fewer sites in the genome will match by chance. A newer approach called prime editing avoids double-strand breaks entirely, using a modified Cas9 fused to a reverse transcriptase to directly write new genetic information into the target site. Studies in human cells have found no detectable off-target mutations from prime editing.
Cas9 in Approved Therapies
In December 2023, the FDA approved Casgevy, the first therapy built on CRISPR/Cas9 technology, for treating sickle cell disease in patients 12 and older. The treatment works by removing a patient’s blood stem cells, using Cas9 to edit them outside the body, and transplanting the modified cells back. The edit boosts production of fetal hemoglobin, a form of hemoglobin that carries oxygen effectively and compensates for the defective hemoglobin that causes sickle cell symptoms. Once the edited stem cells engraft in the bone marrow, they multiply and sustain higher fetal hemoglobin levels long-term.
Casgevy’s approval marked a concrete milestone for Cas9: the transition from lab tool to clinical treatment. The same basic mechanism, a guide RNA directing Cas9 to a specific genomic location to make a targeted cut, underlies the therapy, just applied inside patient cells rather than in a research dish.