How does gene editing work?

Published · Updated

Every method follows the same three steps: find the right place in the DNA, make a change there, and let the cell's own repair machinery finish the job.

How does gene editing work?

A gene-editing tool is steered to one DNA sequence, where it either cuts the DNA or changes a letter directly. The cell's own repair systems then complete the change. In plants, the editing tool is removed and the result is checked.

The 5 gene-editing methods share 4 stages: find one DNA sequence, cut or change it, let the cell's repair complete the change, and, for precision-bred status in England, confirm no transgenic material remains.
Animation: 4 stages shared by the 5 gene-editing methods: find the target, change the DNA, cell repair completes it, then checks for precision-bred status in England.

Animation: Precision Plants. Sources: Science (2012); Genetics, via PubMed (2010); Proceedings of the National Academy of Sciences (1996); Nature (2016); Nature (2019); National Academies of Sciences, Engineering, and Medicine (2017); Advisory Committee on Releases to the Environment, GOV.UK (2025); Department for Environment, Food and Rural Affairs, GOV.UK (2026)

Step 1: find the target

Each tool needs to recognise one sequence among millions. Three methods use a guide RNA that matches the target: CRISPR/Cas9, base editing and prime editing [1][4][5]. Two use a protein built to bind the target: TALENs and zinc finger nucleases [2][3].

Step 2: make the change

Nucleases cut both strands of the DNA at the target [1][2][3]. Base editors and prime editors instead change the sequence without a double-strand break or a donor template; prime editors, and later base editors, nick a single strand [4][5].

Five gene-editing methods, as first described
MethodHow it finds its targetWhat it does to the DNAFirst described
CRISPR/Cas9A guide RNA pairs with the target sequence [1]Cuts both DNA strands at the chosen site [1]2012 [1]
TALENsA custom TAL effector protein that binds the target sequence [2]A FokI cutting domain breaks both strands at the site [2]2010 [2]
Zinc finger nucleasesZinc finger proteins built to match the target [3]A FokI cutting domain cuts near the site [3]1996 [3]
Base editingA guide RNA, as with CRISPR [4]Converts one DNA base into another without cutting both strands [4]2016 [4]
Prime editingA prime editing guide RNA that also carries the edit [5]Writes a short new sequence into the site without cutting both strands [5]2019 [5]

Step 3: the cell repairs the DNA

Cells have their own ways of repairing breaks in DNA, and these can be used to change the sequence [6]. After a double-strand break there are two main routes. With base and prime editors, the cell's repair of the edited strand completes the change instead [4][5].

  • Rejoining the ends, called non-homologous end joining. The broken ends are joined back together, sometimes with a small change at the join [6][7].
  • Copying a template, called homology-directed repair. If a matching stretch of DNA is supplied, the cell can copy it in, making a precise change [6][7].

Step 4: remove the tool and check

In England, a plant is not precision bred if it contains transgenic material, including editing cassettes, selectable markers and vector genes [8]. Transgenic intermediates used during editing are allowed only if all transgenic sequences are removed before release or marketing [8].

Before a plant is released, it must be analysed to confirm that no transgenes remain [9]. Developers also look for unintended changes elsewhere in the genome, called off-target changes [10].

Frequently asked questions

Does gene editing always cut DNA?

No. Nucleases cut both strands, but base editors and prime editors change the sequence without a double-strand break. Prime editors, and later versions of base editors, nick only one strand.

What does the cell do after DNA is cut?

After a nuclease cuts both strands, the cell repairs the break, either by rejoining the ends or by copying a template if one is supplied. Base and prime editors change the DNA without a double-strand break, and the cell's repair completes the change.

Is the editing tool left in the plant?

Not in a precision bred plant in England. Any transgenic material, including the editing cassette, must be removed before release or marketing.

References

  1. Science (2012). A programmable dual-RNA-guided DNA endonuclease in adaptive bacterial immunity (Jinek and others)
  2. Genetics, via PubMed (2010). Targeting DNA double-strand breaks with TAL effector nucleases (Christian and others)
  3. Proceedings of the National Academy of Sciences (1996). Hybrid restriction enzymes: zinc finger fusions to Fok I cleavage domain (Kim, Cha and Chandrasegaran)
  4. Nature (2016). Programmable editing of a target base in genomic DNA without double-stranded DNA cleavage (Komor and others)
  5. Nature (2019). Search-and-replace genome editing without double-strand breaks or donor DNA (Anzalone and others)
  6. National Academies of Sciences, Engineering, and Medicine (2017). Human genome editing: science, ethics, and governance. Appendix A: the basic science of genome editing
  7. EFSA Panel on Genetically Modified Organisms, EFSA Journal (2012). Scientific opinion addressing the safety assessment of plants developed using Zinc Finger Nuclease 3 and other Site-Directed Nucleases with similar function
  8. Advisory Committee on Releases to the Environment, GOV.UK (2025). ACRE guidance on producing precision bred plants
  9. Department for Environment, Food and Rural Affairs, GOV.UK (2026). Releasing precision bred plants into the environment
  10. Plant Physiology (2020). Plant genome editing and the relevance of off-target changes (Graham and others)

Last reviewed 2026-09-26. Edited by Mark Turner.