Prime Editing: The Complete Guide to DNA's Search-and-Replace Revolution

The Promise of Precision

Imagine you could open a book, find a single misspelled word among billions of letters, and correct it without tearing any pages. That is what prime editing does to the human genome.

Standard CRISPR-Cas9 gene editing works by cutting both strands of DNA at a target site and relying on the cell's natural repair machinery to fix the break. This is powerful but imprecise — like using scissors to fix a typo. The repair process can introduce unwanted insertions or deletions, and outcomes are sometimes unpredictable. Base editing, developed in 2016, improved on this by chemically converting one DNA letter to another without cutting, but it can only make 4 of 12 possible single-letter changes and suffers from bystander editing at neighboring bases.

In October 2019, David Liu's laboratory at the Broad Institute of MIT and Harvard published a landmark paper in Nature describing an approach that sidesteps all these problems. They called it prime editing — and it has since proven to be the most versatile and precise gene editing tool ever created, capable of correcting approximately 89% of all known disease-causing mutations.

By December 2025, prime editing had reached a milestone that even its creators might not have predicted so soon: the first clinical results published in the New England Journal of Medicine, showing 69-83% gene correction in patients with chronic granulomatous disease. This is the story of how we got there — and where prime editing is going next.

How Prime Editing Works: The Molecular Mechanism

Prime editing uses three molecular components working in precise coordination:

1. The Cas9 Nickase (nCas9 H840A)

Unlike standard CRISPR-Cas9, which cuts both strands of DNA, prime editing uses a modified Cas9 where one catalytic domain (HNH) is inactivated. This "nickase" can only cut one strand — like carefully slicing one side of a zipper instead of cutting through both. Specifically, the RuvC domain remains active and nicks the non-target strand (the PAM-containing strand), three base pairs upstream of the PAM sequence.

2. The Reverse Transcriptase (RT)

An engineered Moloney murine leukemia virus (M-MLV) reverse transcriptase is fused to the C-terminus of the Cas9 nickase. This enzyme does something remarkable: it reads an RNA template and writes a complementary DNA strand — the reverse of normal cellular transcription. The engineering is critical: five point mutations (D200N, L603W, T306K, W313F, T330P) improve its thermostability, processivity, and DNA-RNA binding, making it work efficiently at human body temperature.

3. The Prime Editing Guide RNA (pegRNA)

This is what makes prime editing truly special. A pegRNA is a single RNA molecule that performs dual duty:

• Its 5' spacer (~20 nucleotides) guides the complex to the target DNA site, exactly like a standard CRISPR guide RNA
• Its 3' extension contains two critical elements: a primer binding site (PBS) that hybridizes with the nicked DNA strand, and a reverse transcriptase template (RTT) that encodes the desired edit

The pegRNA literally carries the corrected genetic information to the target site.

The Six-Step Process

Step 1 — Target Recognition: The pegRNA spacer directs the prime editor to the target DNA, forming an R-loop where the spacer base-pairs with the target strand.

Step 2 — Nicking: The Cas9 nickase cuts only the non-target strand, exposing a free 3' DNA end. No double-strand break occurs.

Step 3 — PBS Hybridization: The primer binding site on the pegRNA hybridizes with the exposed 3' end of the nicked strand, creating an RNA-DNA hybrid that serves as a starting point for the reverse transcriptase.

Step 4 — Reverse Transcription: The fused RT enzyme extends the 3' end of the nicked strand, using the RTT portion of the pegRNA as a template. This directly writes the desired edit into new DNA at the target site.

Step 5 — Flap Resolution: The cell now has two competing sequences: the newly synthesized 3' flap containing the edit, and the original unedited 5' flap. Cellular enzymes (like FEN1) preferentially remove the unedited 5' flap, and ligases seal the nick, incorporating the edit into one strand.

Step 6 — Mismatch Resolution: The DNA now contains a heteroduplex — one strand has the edit, the other retains the original sequence. The cell's mismatch repair machinery resolves this, permanently installing the edit.

The Word Processor Analogy

If CRISPR-Cas9 is like scissors cutting a sentence out of a book and hoping someone pastes in the right replacement, prime editing is like a word processor: it finds the exact word (spacer), opens a cursor at the right position (nick), and types in the correction directly (reverse transcription) — all without ever tearing the page (no double-strand break).

What Prime Editing Can Do

• All 12 possible point mutations (4 transitions + 8 transversions)
• Small insertions up to ~50 base pairs
• Small deletions up to ~80 base pairs
• Combination edits — simultaneous substitution plus insertion or deletion
• With twin prime editing and PASTE technology: insertions of up to 36 kilobases

The Inventor: David R. Liu

David R. Liu's path to revolutionizing gene editing came through chemistry, not molecular biology — and that perspective made all the difference.

Born in 1973 in Riverside, California, Liu graduated summa cum laude from Harvard College in 1994 with a degree in chemistry, then earned his PhD at UC Berkeley under Peter Schultz. He joined Harvard's faculty at age 25 — one of the youngest professors in the university's history.

Where biologists saw CRISPR as a cutting tool, Liu's chemist's eye saw opportunities for chemical transformation at specific genomic sites. His directed evolution expertise — particularly PACE (phage-assisted continuous evolution), a system he developed for rapidly evolving proteins — gave him unique tools to engineer new enzyme variants that no one else could create.

The Invention Timeline

• 2016: Liu's lab created the first cytosine base editor (BE3, with Alexis Komor), published in Nature
• 2017: First adenine base editor (ABE, with Nicole Gaudelli), also in Nature — remarkably, the TadA* deaminase was evolved in the lab using PACE; it does not exist in nature
• 2019: Prime editing (with Andrew Anzalone), published in Nature on October 21, 2019

Andrew Anzalone, a Jane Coffin Childs postdoctoral fellow in Liu's lab, conceived the prime editing concept and was first author on the landmark paper "Search-and-replace genome editing without double-strand breaks or donor DNA." The paper was named one of Nature's top 10 papers of 2019.

Recognition

In April 2025, Liu received the Breakthrough Prize in Life Sciences — often called the "Oscars of Science" — for the development of base editing and prime editing. As of 2026, at least 19 clinical trials using base editing or prime editing are underway in five countries to treat diseases including leukemias, hypercholesterolemia, alpha-1 antitrypsin deficiency, sickle cell disease, beta-thalassemia, and chronic granulomatous disease.

Liu currently serves as Richard Merkin Professor at Harvard, Director of the Merkin Institute of Transformative Technologies in Healthcare at the Broad Institute, and HHMI Investigator. He has co-founded four companies based on his research: Beam Therapeutics (base editing), Prime Medicine (prime editing), Chroma Medicine (epigenome editing), and Exo Therapeutics.

The Evolution: PE1 Through PE7

One of the most remarkable aspects of prime editing is how rapidly it has improved. Each generation addressed specific limitations of its predecessors.

PE1 and PE2 (2019) — The Foundation

The original prime editor (PE1) used wild-type M-MLV reverse transcriptase fused to Cas9 nickase. It worked, but editing efficiencies were low: typically 0.7-5.5% in human cells. PE2, described in the same paper, incorporated five mutations into the RT domain that improved thermostability and processivity, boosting efficiency 1.6- to 5.1-fold over PE1.

PE3 and PE3b (2019) — The Second Nick

PE3 added a second guide RNA that nicks the non-edited strand 14-116 base pairs from the prime editing site. This biases the cell's mismatch repair to copy from the edited strand, improving efficiency an additional 1.5- to 4.2-fold. PE3b refined this by designing the second guide RNA to only match the sequence after the edit is installed, reducing unwanted indels while maintaining PE3-level efficiency.

PEmax and PE4/PE5 (2022) — Breaking Through Mismatch Repair

Published by Chen et al. in Nature Biotechnology, PEmax was a substantially re-engineered prime editor with codon-optimized RT, additional nuclear localization signals, and Cas9 mutations that enhanced activity — roughly 2-3x improvement over PE2.

The breakthrough insight: the cell's mismatch repair (MMR) system often "corrects" the edit back to the original sequence. PE4 and PE5 addressed this by transiently co-expressing a dominant-negative MLH1 protein (MLH1dn) to inhibit endogenous MMR. PE4 improved efficiency ~7.7-fold over PE2 for certain edits. This was a paradigm shift — working with the cell's biology rather than fighting it.

PE6 (2023-2024) — Solving the Delivery Problem

The PE6 family, developed through Liu's PACE directed evolution system, addressed a critical bottleneck: the prime editor protein (~6.3 kb coding sequence) is too large for a single AAV vector (~4.7 kb capacity).

• PE6a and PE6b use compact RT domains from bacterial retron and yeast retrotransposon RTs, small enough for potential single-AAV delivery
• PE6c and PE6d are further evolved variants that balance compactness with efficiency for complex edits
• The PE6 family achieves 2-20x greater efficiency than previous versions while being significantly smaller

PE7 (2024) — The La Protein Revolution

Published in Nature in 2024, PE7 represents a conceptual leap. Through genome-scale CRISPR-interference screens, Liu's lab identified the La protein — a ubiquitous RNA-binding protein — as the strongest mediator of prime editing efficiency.

La naturally protects RNA polymerase III transcripts from exonuclease degradation. By fusing La's N-terminal domain to the prime editor, PE7 stabilizes pegRNAs inside the cell, dramatically extending their functional lifetime.

The results were stunning: 21.2-fold improvement over PE2 with standard pegRNAs, and 10.8-fold improvement at disease-related loci. PE7 fundamentally closed the efficiency gap between prime editing and base editing at many genomic targets.

PE7 + AI-Designed MLH1-SB (2025) — The Cutting Edge

Published in Cell in 2025, researchers used RFdiffusion (an AI protein design tool) to create MLH1 small binders (MLH1-SBs) — just 82 amino acids — that disrupt the MLH1-PMS2 mismatch repair complex more efficiently than the much larger MLH1dn protein.

Fusing MLH1-SB with PE7 achieved editing efficiencies 18.8 times higher than PEmax and 2.5 times higher than PE7 alone in human cells. This represents the convergence of two revolutionary technologies: AI protein design and prime editing.

Prime Editing vs. CRISPR-Cas9 vs. Base Editing

Understanding prime editing requires placing it in context with its predecessors.

Feature	CRISPR-Cas9	Base Editing	Prime Editing
Mechanism	Double-strand break	Chemical deamination at nick	Reverse transcription at nick
Double-strand break?	Yes	No	No
Edit types	Knockouts, large indels	C→T, A→G only (4 of 12)	All 12 point mutations + small indels
Bystander editing	N/A	Yes (30-60% at nearby bases)	None
RNA off-targets	None	Significant (ABE8e)	None
Efficiency	High for knockout (80-95%)	High for supported edits (30-80%)	High with PE7 (40-70%+)
Works in non-dividing cells	Knockout yes; HDR no	Yes	Yes
Disease mutations addressable	~10% (via knockout)	~30% (transitions only)	~89% (theoretically)
Clinical status	FDA-approved (Casgevy)	Phase 1/2 (Beam)	NEJM-published (PM359)

Why No Double-Strand Break Matters

The absence of double-strand breaks is not merely a technical detail — it has profound safety implications. DSBs are associated with:

• Large deletions — sometimes megabase-scale, far from the cut site
• Chromosomal translocations — linking two different chromosomes
• Chromothripsis — chromosome shattering events
• p53 pathway activation — selecting for cells with defective tumor suppressor function

Prime editing avoids all of these risks entirely, making it inherently safer for therapeutic applications.

The Bystander Problem

Base editing's bystander editing problem is often understated. Cytosine and adenine base editors convert ALL target bases within their ~4-8 nucleotide editing window. If there are multiple cytosines (for CBE) or adenines (for ABE) in the window, they all get edited — potentially creating missense or nonsense mutations. Studies show bystander rates of 30-60% at neighboring bases.

Prime editing has zero bystander editing because the pegRNA template specifies the exact edit to be made at the exact position.

Engineered pegRNAs: The Unsung Innovation

One of the most impactful yet under-appreciated advances in prime editing came not from improving the protein, but from protecting the RNA.

Standard pegRNAs have a fundamental vulnerability: their 3' extension (containing the PBS and RTT) is exposed to cellular exonucleases — enzymes that chew up unprotected RNA ends. Inside a living cell, pegRNAs are rapidly degraded before they can complete their job, severely limiting editing efficiency.

In 2021, Nelson, Randolph, and colleagues in the Liu lab published a solution in Nature Biotechnology: engineered pegRNAs (epegRNAs). By adding structured RNA motifs to the 3' end of pegRNAs — particularly tevopreQ1, a trimmed version of a pseudoknot from a natural riboswitch — they created pegRNAs with self-protecting "caps" that fold into stable structures resistant to exonuclease degradation.

The results were transformative: 3-7 fold improvement in editing efficiency across multiple human cell types (HeLa, U2OS, K562, primary fibroblasts) without increasing off-target activity. Importantly, epegRNAs and PE7 solve the same problem (pegRNA degradation) through different mechanisms — epegRNAs protect the 3' end structurally, while PE7's La protein protects it through binding. Combining both provides diminishing returns, but either alone is a major advance.

This innovation also had practical implications: epegRNA design could be applied to any prime editing experiment immediately, without changing the protein component. It was a "software update" rather than a "hardware upgrade."

The Safety Advantage: What the Data Shows

Prime editing's safety profile deserves special attention because it directly addresses the most concerning risks of other gene editing approaches.

No Double-Strand Breaks = No Genotoxicity

Double-strand breaks (DSBs) are the most dangerous type of DNA damage a cell can experience. When CRISPR-Cas9 creates a DSB, several adverse outcomes can occur:

• Large deletions: Studies have found that CRISPR-Cas9 can cause deletions of thousands or even millions of base pairs around the cut site — far larger than intended
• Chromosomal translocations: If two DSBs occur simultaneously on different chromosomes, the broken ends can be joined incorrectly, creating fusion chromosomes associated with cancer
• p53 selection: DSBs activate the p53 tumor suppressor pathway, causing cells with functional p53 to undergo growth arrest or apoptosis. This means CRISPR editing preferentially selects for cells with defective p53 — the very cells most likely to become cancerous

Prime editing avoids all of these risks because it never creates a DSB. The single-strand nick it creates is orders of magnitude less genotoxic and is routinely repaired by the cell without drama.

No RNA Off-Targets

Adenine base editors (ABEs), particularly the widely used ABE8e, cause tens of thousands of A-to-I RNA edits across the transcriptome. While these are transient (they don't alter the genome permanently), their clinical significance remains debated. CBEs similarly cause off-target C-to-U RNA deamination.

Prime editing has no deaminase component and therefore causes zero transcriptome-wide RNA off-targets. This is a clean sheet.

Dual Recognition Specificity

Prime editing has an inherent specificity advantage: for editing to occur, BOTH the spacer must correctly bind the target DNA AND the PBS must correctly hybridize with the nicked strand. This dual-recognition requirement means off-target prime editing is far less likely than off-target CRISPR cutting, which requires only spacer recognition.

First-in-Human: The PM359 Story

On December 7, 2025, the New England Journal of Medicine published the first peer-reviewed clinical data for prime editing in human patients — a landmark for the entire field.

The Disease: Chronic Granulomatous Disease (CGD)

CGD affects approximately 1 in 200,000-250,000 people. Patients lack functional NADPH oxidase in their neutrophils (a type of white blood cell), leaving them unable to kill certain bacteria and fungi. Without treatment, patients face life-threatening infections and chronic inflammation. Current options — prophylactic antibiotics, interferon-gamma injections, or allogeneic bone marrow transplant (which risks graft-versus-host disease) — are all imperfect.

The Treatment: PM359

Prime Medicine's PM359 is an autologous therapy: a patient's own CD34+ hematopoietic stem cells are collected, prime-edited ex vivo to correct the mutation in the NCF1 gene (which encodes p47phox), and infused back into the patient after myeloablative busulfan conditioning.

The Results

Two patients were treated in the Phase 1/2 trial:

• Patient 1: Achieved 69% DHR-positive neutrophils by Day 30
• Patient 2: Achieved 83% DHR-positive neutrophils by Day 30

Both far exceeded the 20% minimum threshold projected for clinical benefit. DHR (dihydrorhodamine) positivity measures functional NADPH oxidase activity — the very enzyme CGD patients lack.

Critically, DHR activity remained stable over time in both patients, suggesting that gene correction occurred in the long-term repopulating hematopoietic stem cells, not just short-lived progenitors. Both patients remained free of new CGD-related complications, and Patient 1 was able to discontinue mesalamine treatment for CGD-related colitis.

No clinically significant adverse events were attributed to PM359 itself — all observed toxicities were consistent with standard busulfan conditioning.

The Paradox

Despite these extraordinary results, Prime Medicine announced it would not independently advance PM359 further. The reason: CGD's small patient population makes it commercially unviable for a company of Prime Medicine's scale. The company is exploring partnerships for continued development and believes the clinical data may support an accelerated FDA approval. CEO Keith Gottesdiener noted: "We may have effectively cured two patients, but the market is simply too small to sustain a company."

This highlights a broader challenge in gene editing: the therapies that demonstrate the most dramatic proof-of-concept are often for ultra-rare diseases with populations too small to generate commercial returns. The same challenge has plagued gene therapy — bluebird bio had three FDA-approved gene therapies but is going private for less than $30 million due to commercial failure. Prime Medicine's strategic pivot to larger-market liver diseases reflects a hard-learned industry lesson: scientific triumph and commercial viability are not the same thing.

What PM359 Proved

The significance of PM359 extends far beyond CGD. It established several critical proof points for the entire prime editing field:

1. Prime editing works in human patients — not just in cell lines and animal models, but in actual people
2. The editing is durable — correction occurred in long-term repopulating HSCs, meaning the fix is permanent
3. Efficiency is therapeutically relevant — 69-83% correction rates exceed what is needed for clinical benefit by a wide margin
4. Safety is manageable — no treatment-attributable adverse events in either patient
5. The platform is ready for broader application — if prime editing works this well ex vivo in HSCs, the technology is ready for in vivo applications in liver and beyond

Prime Medicine's Pipeline: The Liver Franchise

Following the PM359 proof of concept, Prime Medicine has pivoted its focus to in vivo liver diseases — larger patient populations with clear commercial potential.

PM577 — Wilson's Disease

Wilson's disease is caused by mutations in the ATP7B gene, leading to toxic copper accumulation in the liver and brain. The most common mutation in the US is H1069Q. PM577 uses in vivo LNP delivery to prime edit this mutation directly in liver cells.

At the AASLD Liver Meeting in November 2025, Prime Medicine presented preclinical data showing normalization of hepatic copper levels detected by 64Cu PET imaging in mice treated with PM577. IND/CTA filing is planned for H1 2026, with initial clinical data expected in 2027.

PM647 — Alpha-1 Antitrypsin Deficiency (AATD)

AATD, caused primarily by the PiZ mutation (E342K in SERPINA1), leads to liver disease and emphysema. PM647 demonstrated high editing efficiency and restoration of corrected M-AAT protein to healthy human range at clinically relevant doses in fully humanized mouse models. IND filing is planned for mid-2026.

Cystic Fibrosis Program

Funded by $39 million from the Cystic Fibrosis Foundation, Prime Medicine's CF program targets the F508del mutation found in 85% of CF patients — a three-nucleotide deletion that base editing simply cannot correct.

In July 2024, researchers from the Broad Institute and University of Iowa published a breakthrough in Nature Biomedical Engineering: by combining six optimizations for prime editing (engineered pegRNAs, PEmax architecture, transient MMR inhibition, strategic silent edits, PE6 variants, and proximal dead sgRNAs), they achieved 58% correction of CFTR F508del in bronchial epithelial cells and 25% correction in patient-derived airway cells — restoring function to levels comparable to Trikafta, the current standard of care.

The challenge remains lung delivery: thick mucus barriers and immune clearance make LNP delivery to airway cells far harder than liver targeting. Pulmonary-targeted LNPs with peptide ionizable lipids are showing promise in preclinical models.

Beyond Single Edits: Twin Prime Editing, PASTE, and PERT

Prime editing's versatility extends far beyond single-base corrections.

Twin Prime Editing (2021)

Uses two pegRNAs targeting opposite strands at adjacent sites. Each pegRNA directs synthesis of a complementary DNA flap; the two flaps hybridize and replace the intervening sequence. This enables larger deletions, replacements, inversions, and installation of recombinase recognition sites.

PASTE and eePASSIGE (2022-2024)

PASTE (Programmable Addition via Site-specific Targeting Elements) combines twin prime editing with site-specific integrases. First, twin prime editing installs a landing pad (attB site) at the target. Then a serine integrase (like Bxb1) catalyzes recombination, inserting cargoes of up to 36 kilobases with efficiencies up to 60%.

eePASSIGE (2024) improved this further with evolved recombinase enzymes, making gene-sized insertions several times more efficiently — a critical advance for diseases requiring full gene replacement.

PERT — Disease-Agnostic Editing (2025)

Published in Nature in November 2025, PERT (Prime Editing-installed suppressor tRNAs for Readthrough of premature Termination codons) represents a paradigm shift in thinking about genetic therapies.

Approximately 24% of all disease-causing mutations in ClinVar create premature stop codons — nonsense mutations that halt protein production too early. Rather than correcting each mutation individually (which would require developing a different editing agent for each), PERT uses prime editing to convert a dispensable endogenous tRNA into an optimized suppressor tRNA that reads through premature stop codons.

The researchers screened thousands of variants of all 418 human tRNAs and identified those with the strongest suppressor potential. In cell models of Batten disease, Tay-Sachs disease, and Niemann-Pick disease type C1, PERT restored enzyme activity to 20-70% of normal levels.

The transformative implication: a single PERT composition could potentially treat patients across multiple different diseases — including approximately 8,000 people with cystic fibrosis, 252,000 with Stargardt disease, and 43,500 with Duchenne muscular dystrophy who carry nonsense mutations.

This represents a fundamental shift in how we think about genetic medicine. Instead of developing a unique therapy for each of the thousands of known genetic diseases — an approach that is economically unsustainable for rare conditions — PERT offers a "one-to-many" model where a single therapeutic composition addresses an entire class of mutations across multiple diseases. If validated clinically, this could solve the commercial viability problem that has haunted the gene therapy field.

Prime Assembly (2025)

The newest addition to the prime editing toolkit, Prime Assembly (PA) enables insertion of large DNA donor fragments by designing donor ends that overlap with flaps generated by twin prime editing. This approach can insert one, two, or three overlapping DNA fragments with total sizes ranging from 0.1 to 11 kilobase pairs — bridging the gap between small prime edits and full gene replacement.

Delivery: The Remaining Challenge

The biggest barrier between prime editing and widespread clinical use is delivery — getting the molecular machinery into the right cells.

The Size Problem

The PEmax fusion protein coding sequence is approximately 6.3 kb. AAV vectors, the most clinically validated gene delivery vehicles, can only package about 4.7 kb. This means the standard prime editor does not fit in a single AAV.

Current Solutions

Lipid Nanoparticles (LNPs) have emerged as the leading delivery platform for prime editing. Advantages include no packaging size limit, transient expression (reducing off-target accumulation), re-dosing capability (no anti-capsid immune response), and strong clinical precedent from COVID-19 mRNA vaccines. LNPs naturally accumulate in the liver, making liver-targeted diseases the first in vivo applications.

Dual AAV with Split Inteins splits the prime editor into two halves, each packaged in a separate AAV. Upon co-infection, split intein domains mediate protein trans-splicing to reconstitute the full-length editor. Limitation: both vectors must infect the same cell.

PE6 Compact Editors use smaller RT domains that bring total size closer to single-AAV capacity, potentially eliminating the need for dual-vector approaches.

Ribonucleoprotein (RNP) Delivery uses pre-assembled protein-RNA complexes delivered via electroporation (ex vivo) — this is how PM359 is manufactured for the CGD program.

Tissue-Specific Challenges

• Liver: Most tractable. LNPs naturally accumulate there. Both PM577 and PM647 target liver.
• Lung: Major challenge for CF. Mucus barrier and immune clearance. Peptide ionizable lipids showing promise in mouse models.
• Brain: Blood-brain barrier limits access. Intrathecal injection of AAVs or LNPs under investigation. Liu lab demonstrated prime editing treatment of alternating hemiplegia of childhood (AHC) in mice — correcting five different mutations.
• Bone marrow/HSCs: Ex vivo electroporation (PM359 approach). Well-established from CRISPR HSC therapies like Casgevy.

Prime Editing in Agriculture

While therapeutic applications capture the most attention, prime editing is quietly transforming crop science as well.

Demonstrated Applications

• Rice: PE systems optimized with conditional knockdown of OsMLH1 (mismatch repair gene) improved efficiency while maintaining plant fertility. Co-expression of T5 exonuclease with PE2 gave a 5-fold increase in homozygous mutants. Disease resistance has been enhanced by correcting specific point mutations in susceptibility genes.
• Wheat: The ePPEplus system boosted editing efficiency 33-fold, enabling simultaneous editing of up to 8 genes in hexaploid wheat — a remarkable feat given wheat's complex polyploid genome.
• Maize, tomato, soybean: Early demonstrations of prime editing feasibility with crop-optimized PE variants.

The Regulatory Advantage

Prime editing creates precise changes identical to natural genetic variation — no foreign DNA is inserted. In many jurisdictions (including the US under the USDA SECURE rule), gene-edited crops that could have arisen through conventional breeding are exempt from GMO regulation. This gives prime editing a significant regulatory advantage over transgenic approaches.

Remaining Challenges in Plants

Editing efficiency in plant cells remains lower than in mammalian cells, partly due to differences in DNA repair machinery and chromatin structure. Tissue culture and regeneration requirements add complexity. Researchers are addressing these through plant-specific codon optimization, overexpression of chromatin remodeling factors (like hFTO, which increased average efficiency from 33% to 52%), and repeated high-temperature treatment to boost editing in dicot species.

Market Outlook and the Future

The prime editing technology market is estimated at $1.7 billion (2024) and projected to reach $7.8 billion by 2033 (CAGR 18.5%). The broader genome editing market is projected at $36.4 billion by 2033.

Key Clinical Milestones to Watch

• H1 2026: PM577 IND filing (Wilson's disease)
• Mid-2026: PM647 IND filing (AATD)
• 2027: Initial clinical data from PM577 and PM647 — the first in vivo prime editing data in humans
• Ongoing: PM359 partnership discussions for continued CGD development; potential accelerated BLA
• 2026-2027: CF program preclinical proof of concept with pulmonary delivery

The Competitive Landscape

Prime editing exists within a rapidly expanding toolkit:

• CRISPR-Cas9: FDA-approved (Casgevy), dominant for gene knockout
• Base editing: More clinical programs, BLA-track for BEAM-101 in SCD
• Epigenetic editing: Reversible gene regulation without DNA changes (Tune Therapeutics, Chroma Medicine)
• RNA editing: Transient corrections using endogenous ADAR (Wave Life Sciences)

Prime editing's unique position: it is the only tool that can make virtually any small precise DNA edit without double-strand breaks, bystander mutations, or RNA off-targets. As PE7 and AI-enhanced variants close the efficiency gap, and as delivery technology matures beyond the liver, prime editing is positioned to become the default precision editing tool for the majority of genetic diseases.

The David Liu Vision

In his 2025 Breakthrough Prize acceptance speech, Liu articulated a vision for the future of genetic medicine: a world where a patient's specific mutation is identified, the appropriate prime editing guide RNA is designed, and a one-time treatment permanently corrects the root cause of their disease. With PERT technology extending this vision to mutation-agnostic approaches, and with AI accelerating both protein engineering and guide RNA design, that future is closer than most people realize.

The journey from PE1's 0.7% editing efficiency in 2019 to PE7+MLH1-SB's 18.8-fold improvement over PEmax in 2025, and from theoretical promise to NEJM-published 83% gene correction in patients, represents one of the fastest translational timelines in modern biotechnology.

Prime editing is no longer a promising concept. It is a clinically validated technology, rapidly expanding in capability and scope, that may ultimately prove to be the most important biomedical invention of the 21st century.

Sources & Further Reading

• Anzalone, A.V. et al. "Search-and-replace genome editing without double-strand breaks or donor DNA." Nature 576, 149-157 (2019). — The original prime editing paper.
• PM359 NEJM Publication — First-in-human prime editing. New England Journal of Medicine, December 7, 2025.
• Prime Medicine PM359 Breakthrough Clinical Data. May 2025.
• PE7 — Improving prime editing with La protein. Nature (2024). 21.2-fold efficiency improvement.
• AI-designed MLH1-SB enhances prime editing. Cell (2025). 18.8x improvement over PEmax.
• PE6 evolved prime editors — smaller and more efficient. Broad Institute.
• PERT: Disease-agnostic prime editing for nonsense mutations. Nature, November 19, 2025.
• Prime editing corrects cystic fibrosis F508del mutation. Nature Biomedical Engineering, July 2024.
• David Liu receives 2025 Breakthrough Prize. Broad Institute.
• Prime Medicine FY2025 Financial Results & Pipeline. March 2026.
• Prime Medicine Pipeline. Official pipeline page.
• Prime Medicine Strategic Restructuring. Focus on liver diseases, CF, and partnered programs.
• Chen, P.J. et al. "Enhanced prime editing systems by manipulating cellular determinants of editing outcomes." Cell 184, 5635-5652 (2021). — PEmax, PE4, PE5.
• Nelson, J.W. et al. "Engineered pegRNAs improve prime editing efficiency." Nature Biotechnology 40, 402-410 (2022). — epegRNAs.
• Yu et al. "Evolution of Prime Editing: Enhancing Efficiency and Expanding Capacity." Advanced Science (2026). — Comprehensive review.
• Liu Group — Genome Editing Research. Broad Institute / Harvard.

Last updated: March 2026.