Résumé
Non-long terminal repeat retrotransposons, widespread among eukaryotic genomes, transpose by reverse transcription of an RNA intermediate. Some of them, like L1 in the human, terminate at the 3'-end with a poly(dA) stretch whereas others, like the I factor in Drosophila melanogaster, have instead a short sequence repeated in tandem. This suggests different requirements for the initiation of reverse transcription. Here, we have used an RNA circularization/reverse transcription-PCR technique to analyze the 5'- and 3'-ends of the full-length transcripts produced by the I factor at the time of active retrotransposition. These transcripts are capped and polyadenylated similar to conventional messenger RNAs. We have analyzed the 3'-ends of transcripts and transposed copies produced by I elements mutated at the 3'-ends. Transcripts devoid of tandem UAA repeats, although capable of building the components of the retrotransposition machinery, are inefficiently used as retrotransposition intermediates. Such transcripts produce rare new integrated copies issued from the inaccurate initiation of reverse transcription near the 3'-end of the element. The tandem UAA repeats at the 3'-end of the transcripts of I are required for the efficient and precise initiation of reverse transcription. This strong specificity of the I factor reverse transcriptase for its own transcript has implications for the impact of I factor retrotransposition on the host genome.Non-long terminal repeat retrotransposons, widespread among eukaryotic genomes, transpose by reverse transcription of an RNA intermediate. Some of them, like L1 in the human, terminate at the 3'-end with a poly(dA) stretch whereas others, like the I factor in Drosophila melanogaster, have instead a short sequence repeated in tandem. This suggests different requirements for the initiation of reverse transcription. Here, we have used an RNA circularization/reverse transcription-PCR technique to analyze the 5'- and 3'-ends of the full-length transcripts produced by the I factor at the time of active retrotransposition. These transcripts are capped and polyadenylated similar to conventional messenger RNAs. We have analyzed the 3'-ends of transcripts and transposed copies produced by I elements mutated at the 3'-ends. Transcripts devoid of tandem UAA repeats, although capable of building the components of the retrotransposition machinery, are inefficiently used as retrotransposition intermediates. Such transcripts produce rare new integrated copies issued from the inaccurate initiation of reverse transcription near the 3'-end of the element. The tandem UAA repeats at the 3'-end of the transcripts of I are required for the efficient and precise initiation of reverse transcription. This strong specificity of the I factor reverse transcriptase for its own transcript has implications for the impact of I factor retrotransposition on the host genome.