Prime Editing: CRISPR's More Precise Successor

Prime Editing Can Fix 89% of Known Pathogenic Human Genetic Variants

David Liu's lab at the Broad Institute published the original prime editing paper in Nature in October 2019, estimating that the technology could theoretically address 89% of the approximately 75,000 known pathogenic human genetic variants catalogued in ClinVar. Traditional CRISPR-Cas9 cuts DNA but relies on the cell's own imprecise repair machinery — introducing insertions and deletions (indels) that disrupt function but cannot install a specific correct sequence. Prime editing writes new genetic information directly into a specific genomic location without requiring a double-strand DNA break and without depending on a donor template delivered separately.

The Three CRISPR-Cas9 Limitations Prime Editing Addresses

Understanding prime editing requires first understanding why CRISPR-Cas9 alone is insufficient for precision medicine applications:

• Double-strand breaks (DSBs) — Cas9 cuts both strands of DNA. DSBs trigger the cell's non-homologous end joining (NHEJ) repair pathway, which is fast but error-prone, producing random indels at the cut site. Homology-directed repair (HDR) can introduce precise changes but occurs only in dividing cells during S/G2 phase, limiting therapeutic applications.
• Off-target cleavage — Cas9 can bind and cut DNA at sites with partial sequence complementarity to the guide RNA, causing unintended mutations elsewhere in the genome. Off-target rates vary from 0.01% to over 1% depending on the guide sequence and delivery context.
• Limited edit types — Cas9 alone cannot install a specific base change, small insertion, or small deletion at a precise location without HDR, which requires a separate donor DNA template and is inefficient in most cell types.

Base Editing: The Intermediate Step

Base editing, also developed by Liu's lab (2016–2017), couples a catalytically impaired Cas9 (nickase or dead Cas9) to a chemical enzyme that directly converts one DNA base to another without cutting the double strand. Two classes exist:

Base editing advanced therapeutic gene correction significantly — ABE8e versions achieve up to 99.9% base conversion efficiency in some contexts. But bystander editing (off-target base conversions at other C or A bases within the editing window) and the inability to install transversions or make insertions/deletions limited the fraction of diseases addressable.

Prime Editing: Mechanics of pegRNA

Prime editing uses two engineered components working together:

Prime Editor (PE) protein — a fusion of a Cas9 nickase (which cuts only one DNA strand) fused to a reverse transcriptase (RT) enzyme. The most therapeutically advanced version, PE3, couples the RT to a Cas9(H840A) nickase variant.

Prime editing guide RNA (pegRNA) — a specialized RNA molecule with two functional regions: (1) the spacer sequence, identical to a standard Cas9 guide RNA, which directs the PE protein to the target genomic locus; and (2) a 3' extension called the primer binding site (PBS) and RT template, which encodes the desired edit sequence and serves as the template for reverse transcription.

The mechanism proceeds in five steps:

• The pegRNA's spacer directs the PE-pegRNA complex to the target DNA site via Watson-Crick base pairing
• The Cas9 nickase cuts the non-template strand, creating a nick (single-strand break)
• The nicked DNA's 3' flap hybridizes to the PBS region of the pegRNA
• The reverse transcriptase uses the RT template in the pegRNA to synthesize a new DNA strand encoding the desired edit
• The cell's flap equilibration and DNA repair processes incorporate the edited strand, completing the edit

A second nick is often introduced on the opposite strand (PE3 strategy) to bias repair toward the edited sequence, improving installation efficiency.

12 Mutation Types Addressable

Prime editing can install all 12 types of point mutations — all four transitions (C→T, T→C, A→G, G→A) and all eight transversions (C→A, C→G, A→T, A→C, etc.) — as well as small insertions up to approximately 44 base pairs and small deletions up to approximately 80 base pairs. This coverage dramatically expands the addressable genetic disease space compared to base editing alone.

Off-Target Profile vs. Cas9

Multiple independent comparisons have found prime editing's off-target activity to be lower than standard Cas9 in equivalent delivery contexts. A 2021 study in Nature Biotechnology (Kim et al.) using unbiased off-target detection methods (GUIDE-seq and CIRCLE-seq) found PE3 produced fewer detectable off-target edits than Cas9 at the same genomic loci. The mechanistic reason: prime editing requires three sequential molecular recognition events (spacer binding, PBS hybridization, RT template encoding) rather than Cas9's single guide RNA recognition, making productive off-target editing less probable. However, the reverse transcriptase may also introduce unintended edits at the target site itself (indels from flap processing), requiring careful pegRNA design.

Clinical Pipeline

Prime Medicine, founded by David Liu in 2019, raised $315 million in Series B financing in 2022 and filed for an IPO in late 2022. The company's therapeutic programs use prime editing delivered via lipid nanoparticles (LNPs) for in vivo liver targets and via ex vivo stem cell modification for hematopoietic diseases. The delivery challenge remains the primary barrier to clinical translation: getting pegRNA and PE protein efficiently into non-dividing target cells (neurons, muscle fibers, photoreceptors) in vivo requires adeno-associated virus (AAV) vectors or LNPs capable of reaching the target tissue without triggering immune responses. AAV's 4.7 kb packaging capacity is tight for the PE protein (~6.3 kb coding sequence), requiring split-intein delivery strategies that add complexity. Solving these delivery problems is the central research focus of the field entering 2025.