The NF1 Gene Contains Hotspots for L1 Endonuclease-Dependent De Novo Insertion

Long interspersed (L1) and Alu elements are actively amplified in the human genome through retrotransposition of their RNA intermediates by the,100 still retrotranspositionally fully competent L1 elements. Retrotransposition can cause inherited disease if such an element is inserted near or within a functional gene. Using direct cDNA sequencing as the primary assay for comprehensive NF1 mutation analysis, we uncovered in 18 unrelated index patients splicing alterations not readily explained at the genomic level by an underlying point-mutation or deletion. Improved PCR protocols avoiding allelic drop-out of the mutant alleles uncovered insertions of fourteen Alu elements, three L1 elements, and one poly(T) stretch to cause these splicing defects. Taken together, the 18 pathogenic L1 endonuclease-mediated de novo insertions represent the largest number of this type of mutations characterized in a single human gene. Our findings show that retrotransposon insertions account for as many as similar to 0.4% of all NF1 mutations. Since altered splicing was the main effect of the inserted elements, the current finding was facilitated by the use of RNA-based mutation analysis protocols, resulting in improved detection compared to gDNA-based approaches. Six different insertions clustered in a relatively small 1.5-kb region (NF1 exons 21(16)-23(18)) within the 280-kb NF1 gene. Furthermore, three different specific integration sites, one of them located in this cluster region, were each used twice, i.e. NM_000267.3(NF1):c.1642-1_1642 in intron 14(10c), NM_000267.3(NF1):c.2835_2836 in exon 21(16), and NM_000267.3(NF1):c.4319_4320 in exon 33(25). Identification of three loci that each served twice as integration site for independent retrotransposition events as well as 1.5-kb cluster region harboring six independent insertions supports the notion of non-random insertion of retrotransposons in the human genome. Currently, little is known about which features make sites particularly vulnerable to L1 EN-mediated insertions. The here identified integration sites may serve to elucidate these features in future studies.