How LINE-1 causes chromosome rearrangements

LINE-1 causes chromosome rearrangements when unfinished copies of this mobile element recombine with DNA breaks or with one another. In cultured human cells, these intermediates brought together genomic regions that would normally remain far apart, producing translocations, inversions and large deletions.

LINE-1 is a mobile genetic element that can copy itself and insert a new copy elsewhere in the genome. LINE-1-derived sequences account for roughly 17% of human DNA, although only about 150 full-length copies in an individual genome are potentially capable of moving. They are usually silenced in normal somatic cells but are reactivated in many cancers.

When an unfinished insertion becomes a dangerous DNA end

LINE-1 does not move as a ready-made piece of DNA. It first produces an RNA copy. A protein encoded by LINE-1, ORF2p, then nicks chromosomal DNA and uses the RNA as a template to synthesize a new DNA copy directly at the cut site. The usual outcome is a completed LINE-1 insertion.

The process can also stall before the copy is fully integrated. This leaves an exposed chromosomal end beside an unresolved LINE-1 DNA intermediate. The researchers proposed that this intermediate might enter an abnormal recombination reaction, joining a distant chromosome break instead of completing a conventional insertion.

To capture these rare events, the team divided the gene for green fluorescent protein, GFP, into two incomplete parts. LINE-1 carried one part, while the other was placed beside a chromosome break generated with Cas9. A cell became green only if the LINE-1 intermediate recombined with the broken chromosome and restored a functional GFP gene.

The experiments used cultured human U2OS and HEK293 cells. The researchers then analyzed green cells with Oxford Nanopore long-read sequencing, which can read across an entire rearrangement junction and reveal which chromosome segments have been joined.

In the first reporter system, the team resolved six rearrangements in detail: five translocations between different chromosomes and one inversion within chromosome 11. The junctions retained characteristic signs of LINE-1 activity, including processed RNA-derived sequence, poly-A tracts and evidence of cutting by ORF2p. These features showed that the rearrangements had formed through a LINE-1 DNA intermediate rather than through accidental recombination between experimental plasmids.

Two unfinished LINE-1 copies can join each other

In a second system, two different LINE-1 molecules carried complementary halves of GFP. Green cells appeared when two DNA copies being synthesized at the same time recombined with one another. When the two elements were introduced sequentially, no green signal was detected. This indicated that the recombining substrates were unresolved intermediates rather than completed insertions.

Long-read sequencing revealed segmental deletions and translocations among these events, including a rearrangement predicted to create an unstable chromosome with two centromeres. Across the analyses, the authors identified 12 LINE-1-associated rearrangements and 75 conventional insertions in U2OS cells, and 6 rearrangements and 17 insertions in HEK293 cells. These counts are not natural event rates because the experimental systems were deliberately designed to enrich for rare rearrangements.

Sequence similarity strongly affected whether the abnormal joining occurred. Shortening the shared region from 150 to 50 base pairs reduced recombination between a LINE-1 intermediate and a chromosome break by about fourfold. With no shared sequence, this reporter produced no signal. For recombination between two LINE-1 intermediates, at least 25 base pairs of matching sequence supported efficient joining, whereas 6 base pairs or no match did not.

DNA repair systems can play opposing roles

BRCA1 produced a particularly counterintuitive result. The protein normally helps preserve genome stability and also restricts conventional LINE-1 insertions. Yet in these reporters, BRCA1 was required for rearrangements that relied on long matching sequences. Reducing BRCA1 increased ordinary LINE-1 insertions but decreased homology-dependent rearrangements. One possible explanation is that BRCA1 helps process and align similar DNA ends, which can then be joined inappropriately.

The mismatch-repair protein MSH2 acted as a quality-control barrier. Introducing only three mismatches into a shared 150-base-pair region substantially reduced recombination. When MSH2 was depleted, recombination between the mismatched copies increased by about 1.8-fold. The repair system therefore prevented sufficiently divergent repetitive sequences from pairing too readily.

The study does not establish how often this mechanism operates in the human body or in any particular tumor. It used cell lines, forced LINE-1 expression, engineered DNA breaks and reporters that preferentially detect rearrangements involving shared sequence. Simpler end-joining events without extensive homology may be more common, but this system was not designed to capture them efficiently.

Even with those limitations, the work provides a direct mechanistic model for how an active mobile element can do more than add a new copy to the genome. An unfinished LINE-1 intermediate can physically connect distant DNA breaks and alter chromosome architecture. This may help explain some complex rearrangements found in cancer genomes and highlights a potential safety consideration for genome-editing approaches that use engineered retrotransposons.