Prime editing is a form of genome editing that introduces specified changes into DNA by copying genetic information from an engineered RNA molecule. It combines a modified CRISPR-associated protein with a reverse transcriptase and a prime-editing guide RNA, or pegRNA. The method can produce all twelve possible single-base substitutions, as well as small insertions, deletions, and combinations of these changes, without requiring an intentionally introduced DNA double-strand break or a separate donor DNA template. Its efficiency and precision vary substantially with the target sequence, intended edit, delivery method, and cellular context.
Origin and development
Prime editing was first reported in 2019 by Andrew Anzalone and colleagues in David Liu’s laboratory. The initial study demonstrated targeted sequence changes in human cells and introduced several editor configurations, including PE1, PE2, and PE3.
The method addressed a limitation of existing editing approaches. Conventional CRISPR–Cas9 editing commonly begins with a double-strand break. Cellular repair can then disrupt a target sequence or incorporate information supplied by an external DNA template. However, repair outcomes are often heterogeneous, and precise changes may be inefficient in some cell types. Base editing avoids a programmed double-strand break but generally supports particular classes of base conversion, with capabilities determined by the editor used.
Prime editing expanded the range of changes that could be directly specified by an RNA template. Subsequent developments have improved editor activity, guide-RNA stability, target accessibility, and the handling of DNA repair intermediates. Paired-guide strategies have also extended the approach beyond the small edits emphasized in its original description.
Molecular components
Prime-editor protein
A conventional prime editor contains two connected functional components:
- A Cas9 nickase: a modified Cas9 enzyme that cuts one DNA strand rather than both strands.
- A reverse transcriptase: an enzyme that synthesizes DNA using RNA as a template.
The original system used a Cas9 nickase derived from Streptococcus pyogenes and an engineered reverse transcriptase derived from Moloney murine leukemia virus. Later systems have altered these components to improve activity, size, expression, or targeting range.
The nickase positions the editing machinery at a selected genomic site and creates a free DNA end that can initiate synthesis. The reverse transcriptase then extends that end using instructions encoded in the pegRNA.
Prime-editing guide RNA
The prime-editing guide RNA combines target recognition with an editing template. Its main functional regions are:
- Spacer: directs the Cas9 complex toward a complementary DNA sequence.
- Scaffold: provides the RNA structure needed for association with Cas9.
- Primer-binding site: pairs with the exposed end of the nicked DNA strand.
- Reverse-transcription template: specifies the sequence to be copied into DNA, including the intended alteration.
These regions perform different roles. The spacer identifies the genomic neighborhood, while the reverse-transcription template determines the sequence change. Consequently, prime editing is not restricted to modifying a particular base within a narrow enzyme-activity window.
Target recognition nevertheless depends on an appropriate protospacer-adjacent motif, or PAM, and on a workable arrangement of the nick, primer-binding sequence, and intended edit.
Editing mechanism
Prime editing proceeds through a series of molecular steps.
- Target recognition and nicking. The editor–pegRNA complex binds a compatible DNA site and nicks one strand, exposing a free end.
- Primer binding. The newly exposed DNA end pairs with the pegRNA’s primer-binding site.
- Template-directed synthesis. Reverse transcriptase extends the DNA end by copying the pegRNA template. The resulting DNA extension contains the intended edit.
- Flap processing. The newly synthesized sequence forms an edited DNA flap that competes with an unedited flap. Cellular processing can remove the displaced sequence and incorporate the edited strand.
- Resolution of the opposite strand. The intermediate may contain one edited strand and one unedited strand. Further repair or DNA replication can resolve this mismatch into either the edited or original sequence.
The final outcome therefore depends on both the engineered editor and endogenous DNA repair pathways. Copying the RNA template does not, by itself, guarantee stable incorporation of the edit.
DNA mismatch repair is especially important. Depending on the intermediate, it can oppose incorporation of the newly written sequence. Other repair processes influence flap removal, strand sealing, and the production of unintended insertions or deletions.
Major configurations
Prime-editor names describe particular protein designs or combinations of editing components; they are not a universal ranking of performance.
| Configuration or approach | Main feature |
|---|---|
| PE1 | Initial Cas9-nickase–reverse-transcriptase fusion. |
| PE2 | Uses an engineered reverse transcriptase with improved editing activity relative to PE1. |
| PE3 | Adds a second guide RNA to nick the opposite, initially unedited strand, often increasing edit incorporation. |
| PE3b | Uses a second guide designed to recognize the edited sequence preferentially, helping coordinate opposite-strand nicking with edit formation. |
| PE4 and PE5 | Combine PE2- or PE3-type systems with transient inhibition of mismatch repair, commonly through a dominant-negative MLH1 protein. |
| Engineered pegRNAs | Add stabilizing RNA structures that protect the pegRNA extension and can improve editing efficiency. |
| Paired-pegRNA approaches | Use two pegRNAs to generate coordinated editing intermediates, enabling certain larger replacements, deletions, or insertions. |
PE3 can improve desired editing but may also increase unwanted repair products. Although prime editing is designed around single-strand nicks, nearby nicks on opposite strands can sometimes produce double-strand breaks.
Paired-prime-editing systems can replace stretches of DNA using complementary newly synthesized sequences. Some expanded platforms combine prime editing with other enzymes to enable larger DNA integration; these are extensions of the basic method rather than properties of every prime editor.
Comparison with other editing methods
Nuclease-based editing
Conventional nuclease editing uses a double-strand break as its central intermediate. Repair through non-homologous end joining often produces insertions or deletions useful for disrupting a gene. Precise replacement can instead rely on template-directed repair, including pathways associated with homologous recombination.
Prime editing supplies sequence information through RNA and does not require an intentionally created double-strand break. This can reduce some break-associated outcomes, but it does not eliminate unintended sequence changes or structural alterations.
Base editing
Base editors typically couple a DNA-targeting protein to a base-modifying enzyme. Their capabilities depend on the chemical conversions supported by that enzyme, the editing window, and which nearby bases are exposed.
Prime editing can specify substitutions outside the conversion classes supported by a given base editor and can introduce insertions or deletions. It can also avoid some unwanted “bystander” conversions of neighboring bases. In exchange, prime editing often requires more complex guide design and may be less efficient than a well-matched base editor at a particular target.
Neither approach is uniformly superior: suitability depends on the desired sequence change and biological system.
Applications
Prime editing is used to introduce or correct defined mutations in experimental systems. Creating matched cells that differ at a selected sequence can help test whether a variant changes gene expression, RNA splicing, protein function, or another measurable phenotype.
Research applications include:
- constructing cellular and animal disease models;
- investigating the effects of individual variants or combinations of variants;
- modifying regulatory sequences without deleting an entire region;
- studying genetic interactions through targeted variant libraries;
- introducing selected sequence changes in plants and other organisms.
Prime editing has also been investigated for gene therapy, including correction of disease-associated variants in cultured cells and preclinical models. Results in a cell line or animal model do not establish clinical safety, durability, or benefit in humans.
Limitations and assessment
Efficiency and target dependence
Editing efficiency can vary sharply between genomic sites and between different edits at the same site. Important influences include pegRNA stability, primer-binding and template sequences, local chromatin, editor expression, and cellular repair activity. A high-performing design in one cell type may not perform similarly in another.
Unintended products
Possible outcomes include:
- retention of the original sequence;
- partial incorporation of the intended change;
- unwanted insertions or deletions;
- incorporation of sequences derived from the pegRNA scaffold;
- unintended changes at the target or at other genomic sites.
Assessment therefore requires more than measuring the percentage of DNA molecules carrying the desired edit. DNA sequencing and complementary assays are used to examine product purity, off-target changes, and larger structural events. Bulk measurements may also conceal differences among individual cells or between copies of a chromosome.
Delivery
Prime-editor proteins and their encoding sequences are relatively large. Delivering sufficient editor and guide RNA to the appropriate cells is a major constraint, particularly inside an organism. Delivery as DNA, messenger RNA, protein-containing complexes, or viral-vector systems presents different trade-offs involving payload capacity, expression duration, tissue access, and immune responses.
Repair manipulation and long-term safety
Suppressing mismatch repair can improve editing in some experimental settings, but mismatch repair normally helps preserve genome stability. Its manipulation introduces additional safety considerations. Long-term assessment must also distinguish effects of the intended edit from those of delivery, transient editor expression, unintended genomic changes, and selection of altered cell populations.
Use in reproductive cells or embryos raises additional questions because changes may become heritable. These issues differ from those associated with editing non-reproductive cells and are subject to jurisdiction-specific ethical and regulatory frameworks.