NEW YORK (GenomeWeb) – Long interspersed nuclear elements are present in about half of healthy neurons, where they contribute to genomic diversity in the brain, according to a new study.
Researchers from the Salk Institute developed a new sequencing-based approach to characterize such L1s in the human brain. As they reported today in Nature Neuroscience, their single-cell sequencing and machine-learning method revealed that somatic L1-associated variants (SLAVs) affected about 50 percent of cells in a healthy human brain and are present in key neural genes like DLG2. Further, these L1s could not only insert DNA, but also delete it, the researchers reported.
"The surprising part was that we thought all L1s could do was insert into new places," first author Jennifer Erwin, a staff scientist in Fred Gage's group at the Salk, said in a statement. "But the fact that they're causing deletions means that they're affecting the genome in a more significant way."
Gage's group sequenced the whole genomes of 89 single nuclei and bulk samples obtained from the frontal cortex and hippocampus of three healthy people. To generate genome-wide profiles of L1 insertions, the researchers developed an approach they dubbed SLAV-seq. It relies on a non-PCR-based method for fragmentation and adaptor ligation, and on pair-end sequencing in which one read spans the L1 and flanking sequence junction, among other modifications. This way, the researchers said, they could better identify unique molecules and more confidently detect novel insertions.
They then applied a random forest classifier, comparison to known non-reference germline loci, and other steps to identify 46 putative SLAVs in the single-nucleus samples obtained from brain cells.
Gage and his colleagues estimated that SLAVs appeared at a rate of about 0.58 events to one event per cell — in both glia and neurons — and affected between 44 percent and 63 percent of cells in the healthy brain. About half of these events, though, appeared to be independent of retrostransposition, they noted.
These retrotransposon-independent structural variants — typically deletions — likely arise because L1 sequences are sensitive to the double-stranded DNA damage that occurs when L1 is upregulated during neural differentiation, the researchers said. L1 endonuclease cutting, followed by microhomology-mediated repair, leads to such deletions, they said
Gage and his colleagues previously reported that when an engineered L1 was inserted in the DLG2 gene, it became overexpressed. In this cohort, they reported that DLG2 housed a SLAV. DLG2, they noted, has a key role in complex learning, cognitive flexibility, and attention, but variants in the gene have been linked to schizophrenia.
This, they noted, indicates that SLAVs can appear in important neural genes and suggests that DLG2 could be a SLAV hotspot. But, the researchers added that it also indicates that while genomic diversity in the brain can be beneficial, too much could lead to disease. Indeed, they pointed out that other studies have suggested that people with schizophrenia or Rett's syndrome have more than the usual amount of L1 variation.
The team added that they plan to examine the role of L1 variation in other genes and diseases.