Summary: | <p>Abstract</p> <p>Background</p> <p>Long interspersed nuclear element-1 (LINE-1 or L1) is a dominant repetitive sequence in the human genome. Besides mediating its own retrotransposition, L1 can mobilize <it>Alu </it>and messenger RNA (mRNA) in <it>trans</it>, and probably also SVA and non-coding RNA. The structures of L1 copies and <it>trans</it>-mobilized retrocopies are variable and can be classified into three categories: full-length; 5'-truncated; and 5'-inverted insertions. These structures may be generated by different 5' integration mechanisms.</p> <p>Results</p> <p>In this study, a method to correctly characterize insertions with short target site duplications (TSDs) is developed and extranucleotides, TSDs and microhomologies (MHs) at junctions were analysed for the three types of insertions. Only 5'-truncated L1 insertions were found to be associated with short TSDs. Both full-length and 5'-truncated retrotransposed sequences in <it>trans</it>, including <it>Alu</it>, SVA and mRNA retrocopies and also full-length and 5'-inverted L1, were not associated with short TSDs, indicating the difference of 5' attachment between retrotransposition in <it>cis </it>and retrotransposition in <it>trans</it>. Target sequence analysis suggested that short TSDs were generated in an L1 endonuclease-dependent manner. The MHs were longer for 5'-inverted L1 than for 5'-truncated L1, indicating less dependence on annealing in 5'-truncated L1 insertions.</p> <p>Conclusions</p> <p>The results suggest that insertions flanked by short TSDs occur more often coupled with the insertion of 5'-truncated L1 than with those of other types of insertions <it>in vivo</it>. The method used in this study can be used to characterize elements without any apparent boundary structures.</p>
|