Borrowed from Bacteria: How CRISPR Became a Gene-Editing Tool

A defense system, not an invention

The thing people forget about CRISPR is that nobody invented it. Bacteria did, billions of years ago, as a defense against the viruses that prey on them. The acronym stands for a mouthful, clustered regularly interspaced short palindromic repeats, which describes an odd pattern that microbiologists kept noticing in bacterial genomes: short repeated sequences with unique spacers in between. For years it was a curiosity with no known function. Then researchers realized the spacers were trophies. When a virus infects a bacterium and the bacterium survives, it keeps a snippet of the virus’s DNA filed away between those repeats, like a mugshot. If the same virus attacks again, the bacterium reads out that stored snippet as a piece of RNA and uses it to recognize and destroy the matching viral DNA. It is, in effect, an immune system with a memory, written in genetic code.

The molecular scissors in this system, in the version that became famous, is a protein called Cas9. On its own Cas9 does nothing useful. It needs a guide, a short piece of RNA whose sequence matches the DNA to be cut. The guide leads Cas9 to the matching spot in the genome, and Cas9 clamps on and slices straight through both strands of the DNA. The bacterium uses this to shred invading viruses. The guide is the address, and the protein is the blade.

The leap

Here is the insight that changed biology. If Cas9 can be aimed at any DNA sequence simply by giving it a guide RNA that matches, then you are not limited to viral DNA. You can write a guide for any sequence you like, anywhere in any genome, and Cas9 will go cut it there. In 2012, Jennifer Doudna and Emmanuelle Charpentier showed exactly this. They simplified the natural system into a two-part tool, a single guide RNA plus the Cas9 protein, and demonstrated that they could program it to cut a chosen DNA sequence in a test tube. Change the twenty-letter guide, and you change the target. That was the whole game. They shared the Nobel Prize in Chemistry in 2020 for it, one of the fastest such recognitions in memory.

To appreciate why this was revolutionary, you have to know what came before. Editing a specific gene was already possible using earlier tools with names like zinc-finger nucleases and TALENs, but those required engineering a whole new custom protein for every target, a slow and expensive process that only well-funded labs could manage. CRISPR replaced protein engineering with RNA design. To aim it somewhere new, you just order a different short strand of RNA, which is cheap and quick. A technique that had been the province of specialists became something a graduate student could do in an afternoon. That accessibility, more than any single capability, is why CRISPR spread through biology so fast.

What the cut actually does

Cutting DNA is only half the story; what matters is how the cell repairs the cut. When Cas9 makes its double-strand break, the cell rushes to fix it, and it has two main ways to do so. The quick and sloppy method stitches the ends back together but often loses or adds a few letters in the process, which garbles the gene and switches it off. This is how researchers knock out a gene to see what it does. The second method is more careful: if you supply a DNA template alongside the cut, the cell can use it as a guide to repair the break, copying in whatever new sequence you provided. This is how you correct a mutation or insert a new piece of code, though it works less efficiently than the knockout route.

There is a safety detail worth naming. Cas9 will only cut next to a short signal sequence in the DNA called a PAM, and the guide has to match closely, but the matching is not perfect. Sometimes Cas9 cuts at sites that resemble the target but are not it, so-called off-target edits, and a great deal of the work in turning CRISPR into medicine has been about predicting and minimizing those stray cuts.

From tool to treatment

For the first decade CRISPR was mostly a laboratory technique, a faster way to do experiments. That changed recently. In late 2023, the first CRISPR-based therapy was approved by regulators in the United Kingdom and the United States, a treatment for sickle cell disease and a related blood disorder. The approach edits a patient’s own blood stem cells to switch back on a fetal form of hemoglobin that the body normally shuts off after birth, compensating for the defective adult hemoglobin. It is a genuine cure-like therapy built directly on the bacterial defense system Doudna and Charpentier repurposed. More trials, for other diseases, are underway.

The power comes with hard questions, and one episode made them unavoidable. In 2018 a researcher in China announced he had used CRISPR to edit human embryos that were brought to term, altering a gene in twin girls. The scientific community condemned it almost universally, both because the editing was medically unjustified and risky and because changes made to an embryo are heritable, passed to all future generations, crossing a line most researchers had agreed not to cross. He was later imprisoned. The case sharpened a distinction that still governs the field: editing the cells of a consenting patient to treat their disease is one thing, and editing embryos to make permanent, inheritable changes to the human line is another, subject to a near-total moratorium.

CRISPR is a strange gift. A pattern in bacterial genomes that looked like junk turned out to be an immune system, and that immune system turned out to be programmable, and the programmable version turned out to edit the code of life itself. We did not design any of the underlying machinery. We found it, understood it, and pointed it somewhere new. That is closer to how a lot of science actually works than the story of lone invention usually admits.