The human genome contains many types of chimeric retrogenes generated through in vivo RNA recombination

Abstract

L1 retrotransposons play an important role in mammalian genome shaping. In particular, they can transduce their 3'-flanking regions to new genomic loci or produce pseudogenes or retrotranscripts through reverse transcription of different kinds of cellular RNAs. Recently, we found in the human genome an unusual family of chimeric retrotranscripts composed of full-sized copies of U6 small nuclear RNAs fused at their 3' termini with 5'-truncated, 3'-poly(A)-tailed L1s. The chimeras were flanked by 11-21 bp long direct repeats, and contained near their 5' ends T2A4 hexanucleotide motifs, preferably recognized by L1 nicking endonuclease. These features suggest that the chimeras were formed using the L1 integration machinery. Here we report the identification of 81 chimeras consisting of fused DNA copies of different RNAs, including mRNAs of known human genes. Based on their structural features, the chimeras were subdivided into nine distinct families. 5' Parts of the chimeras usually originated from different nuclear RNAs, whereas their 3' parts represented cytoplasmic RNAs: mRNAs, including L1 mRNA and Alu RNA. Some of these chimeric retrotranscripts are expressed in a variety of human tissues. These findings suggest that RNA-RNA recombination during L1 reverse transcription followed by the integration of the recombinants into the host genome is a general event in genome evolution.

Figure 1

Schematic representation of the chimeric retrogenes identified in public databases. The chimeras’ genomic locations and GenBank accession numbers are listed in Supplementary Material, Table S1.

Figure 2

Two examples of chimeric insertion locus-specific PCR analysis. The PCR products obtained with primate genomic DNA templates and unique primers flanking the chimeras’ insertions were transferred to membranes and separately hybridized with probes specific to the 5′ and 3′ parts of the chimeras and with probes specific to pre-integration sequences. (A) Analysis of U6-mRNA for Keratin 19 chimera insertion from the human 16q22 genomic locus (GenBank accession no. AC009131). (B) Analysis of 5S-AluY chimera insertion from the Xq13 locus (GenBank accession no. AL158069).

Figure 3

Results of the 12 chimeric retrogenes insertional polymorphism study. The chimeras’ integration times were estimated according to the presence/ absence of the inserts in genomic DNAs of different primate species.

Figure 4

A probable mechanism for the chimeras’ formation. (Step 1) An L1 pre-integration complex binds L1, Alu or the host mRNA in the cytoplasm. (Step 2) The ribonucleoprotein formed is transferred to the nucleus. (Step 3) Reverse transcription of the bound mRNA primed by a genomic DNA single-stranded break within the TTTTAA sequence. (Step 4) Another (nuclear) RNA binds to the L1 reverse transcription/integration complex. (Step 5) Switch of templates for the reverse transcription. (Step 6) The DNA reparation mediated formation of a new chimeric retrogene insertion flanked by short direct repeats and carrying a poly(A) sequence at the 3′ terminus.