Internal priming: an opportunistic pathway for L1 and Alu retrotransposition in hominins

Abstract

Retrotransposons, specifically Alu and L1 elements, have been especially successful in their expansion throughout primate genomes. While most of these elements integrate through an endonuclease-mediated process termed target primed reverse transcription, a minority integrate using alternative methods. Here we present evidence for one such mechanism, which we term internal priming and demonstrate that loci integrating through this mechanism are qualitatively different from "classical" insertions. Previous examples of this mechanism are limited to cell culture assays, which show that reverse transcription can initiate upstream of the 3' poly-A tail during retrotransposon integration. To detect whether this mechanism occurs in vivo as well as in cell culture, we have analyzed the human genome for internal priming events using recently integrated L1 and Alu elements. Our examination of the human genome resulted in the recovery of twenty events involving internal priming insertions, which are structurally distinct from both classical TPRT-mediated insertions and non-classical insertions. We suggest two possible mechanisms by which these internal priming loci are created and provide evidence supporting a role in staggered DNA double-strand break repair. Also, we demonstrate that the internal priming process is associated with inter-chromosomal duplications and the insertion of filler DNA.

Figure 1. Alternative mechanisms of retrotransposon integration

(a) Classical TPRT-mediated L1 or Alu insertion into the host primate genome. L1 EN creates a nick in the first strand (orange arrow) at the 5′-TTTT/A-3′ consensus and the retrotransposon mRNA (purple line) anneals to the genomic DNA (blue line) using its polyA tail (purple outline). L1 RT (pink oval) synthesizes the retrotransposon mRNA to complete insertion and the TSDs (grey) are filled in. (b) TPRT variant-mediated retrotransposon insertion. L1 RT internally primes on the L1 or Alu mRNA and the break is filled using classical TPRT machinery. (c) Staggered DSB repair with 5′ overhangs. A staggered DSB (lightning bolt) occurs and RT (pink oval) internally primes on the mobile element mRNA (purple line) that bridges the gap by binding to either end. Subsequent cDNA synthesis fills the break with a copy of the truncated element.

Figure 2. A schematic detailing IP locus investigation

All computationally derived candidate loci were triple-aligned (human, chimpanzee, and rhesus macaque), and those loci found to be human- or hominin-specific were kept for wet bench verification. Gel chromatograph of PCR products from a phylogenetic analysis of a hominin-specific AIP locus (AIP 9). The numbers indicate the DNA template used: 1 & 9, 100bp ladder; 2, negative control (H2O); 3, human; 4, chimpanzee; 5, gorilla; 6, orangutan; 7, rhesus macaque; 8, green monkey.

Figure 3. Alignment of IP loci to their respective consensus sequences

(a) AIP fragments juxtaposed with a representation of a full-length Alu element consensus sequence. The Alu fragments are pink and the consensus sequence is light blue. Two AIP loci are 5′ intact and overall AIP loci align to the consensus sequence with no bias. (b) L1IP fragments juxtaposed with a representation of a full-length L1 element consensus sequence The L1 fragments are. dark blue and the consensus sequence is green. L1IP loci show an alignment bias for the 3′ end of the consensus sequence.

Figure 4. Combined AIP & L1IP microhomology analysis

Complementary nucleotide positions are counted in opposite directions at the 5′ and 3′ ends of the respective consensus sequences. Bases are highlighted in grey if they are complementary to the corresponding nucleotide on the L1 or Alu RNA. The number of matches at each position (r) and the corresponding p-values indicate the likelihood of obtaining the observed numbers of matches by chance alone. Using a binomial probability distribution, we calculated p-values assuming the chance of success (i.e. complimentary pairing) was 1/4 and the chance of failure was 3/4 at each position.

Figure 5. IP insertion site divergence from the preferred L1 endonuclease cleavage site sequence

Loci generated by three different insertion studies (L1IMD, NCI and IP) were analyzed for presence or absence of the preferred L1 EN cleavage site motif. The red line indicates loci analyzed for L1IMD events, which occur via classical TPRT; the blue line indicates NCI events, which are L1 EN-independent; and the green line indicates IP events. The results indicate increased divergence from the preferred motif used by L1 EN-mediated classical TPRT, suggesting that IP events use a mechanism more similar to NCI than L1IMD. These findings are consistent with an opportunistic mechanism.