Prime assembly inserts DNA without severing chromosomes

Genome editing has long been better at making small corrections than inserting large stretches of new DNA. Technologies like prime editing can rewrite individual genetic letters with surgical accuracy, but they struggle as the desired change grows longer. Meanwhile, many methods for integrating bigger DNA cargo rely on deliberately cutting both strands of a chromosome, a powerful but blunt approach that can cause unwanted deletions, rearrangements, or misjoined DNA ends. A Nature paper published 16 September describes a method called prime assembly that bridges this gap: it uses prime-editing machinery to integrate externally supplied DNA and produce large chromosomal rearrangements, without deliberately creating a nuclease-driven double-strand break.

Teaching an old editor new tricks

Standard prime editing works like a biological search-and-replace. A modified Cas9 protein nicks one strand of DNA at a programmed location, and an enzyme called reverse transcriptase copies a short template from a guide RNA directly into the chromosome. The approach is elegant for single-letter corrections and small insertions, but it falters with larger cargo because the guide RNA must encode the entire template, and longer extensions become unstable (original prime-editing paper).

Prime assembly splits the job differently. The editor still nicks one strand and writes short sequences into the chromosome, but those sequences now serve as connection points for a separately supplied, larger piece of donor DNA. The guide RNA directs where the integration happens; the donor carries what gets inserted. Because Cas9 nicks only one strand rather than cutting through both, the method avoids the risks that come with fully severing a chromosome (Nature).

This distinction matters for another reason. Many large-insertion methods depend on homology-directed repair, a cellular process largely confined to actively dividing cells. The team demonstrated prime assembly working in non-dividing cells and in primary human cells, the kind researchers harvest directly from patients rather than growing indefinitely in a lab. Many of the cell types clinicians most want to edit, certain immune cells or tissues that rarely divide, have been difficult targets for integration-based editing (Nature).

What they showed it can do

The paper reports three categories of genome manipulation: exon recoding, transgene insertion, and megabase-scale genomic rearrangements. Exon recoding is worth pausing on. An exon is one segment of a gene that encodes part of a protein. Many genetic diseases arise from mutations scattered across the same exon. Rather than designing a custom fix for each variant, replacing the entire exon with a healthy reference copy could allow one molecular design to address multiple mutations within that stretch (Nature). This was demonstrated as a capability in cells, not as a treatment for any disease.

Several other break-free integration strategies already exist. TwinPE coordinates two guide RNAs on opposite strands (Nature Biotechnology). PASTE installs a recombination site and recruits an integrase (Nature Biotechnology). Bridge recombinases use an entirely different RNA-guided chemistry (Nature). A separate team published a method also called "prime assembly" in Nature Biotechnology in August, using a related but distinct strategy (Nature Biotechnology). None of these approaches have been compared under matched conditions, so ranking them on efficiency or product quality is premature.

What sets the September Nature result apart is its reported combination: programmable targeting, medium-to-large external cargo, operation in non-dividing cells, and demonstrations in primary human cells.

What it means, and what it doesn't

This is a peer-reviewed laboratory result. No patient has been treated and no human safety data exist. Product precision at the editing site, off-target integration, and long-term risks remain unquantified in the accessible record. Avoiding a double-strand break reduces one known hazard; it does not make the process error-free.

Prime assembly has integrated externally supplied DNA and produced large chromosomal rearrangements in non-dividing and primary human cells without a designed nuclease-driven double-strand break. Prime editing began as a tool for small corrections. This work shows it can handle construction jobs that previously required cruder cuts. Translating that into medicine is a separate, longer project, but the demonstrated capability is real and new.