Origin: bacterial immune memory
CRISPR stands for Clustered Regularly Interspaced Short Palindromic Repeats — short DNA sequences in bacteria that store fragments of past viral invaders. When the same virus attacks again, the bacterium transcribes the stored fragment into a guide RNA that leads the Cas9 protein straight to the matching viral DNA and cuts it: a genuine adaptive immune system in a single-celled organism. It was discovered in E. coli in 1987, but its function wasn't understood until Francisco Mojica (2005) showed it matched viral sequences. In 2012, Jennifer Doudna and Emmanuelle Charpentier demonstrated that a simplified single guide RNA could direct Cas9 to cut any DNA sequence in a test tube — the era of programmable CRISPR gene editing began, and the discovery won them the 2020 Nobel Prize in Chemistry.
How Cas9 cuts DNA
Cas9 first scans DNA for a PAM (protospacer adjacent motif) — for the common SpCas9 enzyme, the sequence 5'-NGG-3'. If the 20-nucleotide guide RNA matches the DNA strand next to a PAM, it base-pairs and unwinds the double helix. Two nuclease domains, HNH and RuvC, then each cut one strand, producing a clean double-strand break. What happens next is entirely up to the cell's own repair machinery.
Guide RNA: 5'- AGCUAGCUAGGCUAUGCAUC -3' (20 nt)
||||||||||||||||||||
Target DNA: 3'- TCGATCGATCCGATACGTAG -5'
NGG ← PAM
5'- AGCTAGCTAGGCTATGCATC NGG -3'
↑ Cas9 cuts here, 3 bp upstream of PAM
NHEJ vs HDR: two repair paths
NHEJ (non-homologous end joining) is fast but error-prone: the cell glues the broken ends back together, often inserting or deleting a few base pairs, usually disrupting the gene — a knockout. It handles roughly 70-90% of repairs in most cell types. HDR (homology-directed repair) instead uses a donor DNA template supplied by researchers, with homology arms matching the target site, to precisely insert or replace sequence — a knock-in — at a lower efficiency of typically 5-50%. Most lab research uses knockouts; therapies that need to correct a specific mutation increasingly require the more difficult HDR route.
Guide design, off-targets, and beyond Cas9
Because the human genome has an NGG PAM roughly every 8 base pairs, most genes offer many possible target sites, scored by GC content (best between 40-70%) and by predicted off-target similarity elsewhere in the genome — mismatches closest to the PAM matter most. To cut off-target risk, researchers use high-fidelity Cas9 variants, shorter truncated guides, paired nickases that cut only where both partners bind, or deliver Cas9 as a short-lived protein rather than DNA. Since 2016, base editors fuse a "dead" Cas9 to a deaminase enzyme to flip a single DNA letter without cutting both strands at all, and since 2019 prime editing fuses a nickase Cas9 to a reverse transcriptase, enabling "search-and-replace" edits with no donor template needed.
Real-world applications
In 2023, Casgevy became the first FDA-approved CRISPR therapy, editing a patient's own stem cells to reactivate fetal haemoglobin and functionally cure sickle cell disease. CAR-T cancer immunotherapies use CRISPR to knock out immune checkpoints and insert tumour-targeting receptors. In agriculture, non-transgenic CRISPR crops — drought-resistant wheat, non-browning mushrooms — carry no foreign DNA. Diagnostic tools like SHERLOCK use related Cas enzymes to detect pathogens such as COVID-19 from a paper strip in 30 minutes. Germline editing of human embryos, demonstrated controversially in 2018, remains illegal in most countries; the field draws a hard line between treating a patient's own cells and making heritable changes to future generations.
Frequently asked questions
Where does CRISPR come from, and who turned it into a gene-editing tool?
CRISPR is a bacterial immune system that stores fragments of past viral DNA and uses them to guide the Cas9 protein to cut matching viral DNA on a repeat attack. In 2012, Jennifer Doudna and Emmanuelle Charpentier showed that a simplified single guide RNA could direct Cas9 to cut any chosen DNA sequence in a test tube, work that won them the 2020 Nobel Prize in Chemistry.
What is the difference between NHEJ and HDR repair?
After Cas9 cuts DNA, the cell repairs the break through one of two paths. NHEJ (non-homologous end joining) is fast but error-prone, usually inserting or deleting a few bases and disabling the gene — a knockout. HDR (homology-directed repair) uses a supplied DNA template to make a precise edit, such as correcting a disease mutation — a knock-in, though it is less efficient than NHEJ.
What are off-target effects, and how do scientists reduce them?
Cas9 tolerates some mismatches between the guide RNA and DNA, so it can occasionally cut similar but unintended sites, risking disrupted genes or activated oncogenes. Researchers reduce this risk with high-fidelity Cas9 variants, shorter truncated guides, paired nickases that only cut when both bind correctly, and delivering Cas9 as a short-lived protein-RNA complex rather than as DNA.
Try it live
Everything above runs in your browser — open CRISPR-Cas9 Genome Editing and watch the guide RNA scan DNA, recognise its PAM, cut both strands, and follow either the NHEJ or HDR repair path. Nothing is installed, nothing is uploaded.
▶ Open CRISPR-Cas9 Genome Editing simulation