We conducted a genome-wide survey of Saccharomyces cerevisiae retrotransposons and identified a total of 331 insertions, including 217 Ty1, 34 Ty2, 41 Ty3, 32 Ty4, and 7 Ty5 elements. Eighty-five percent of insertions were solo long terminal repeats (LTRs) or LTR fragments. Overall, retrotransposon sequences constitute >377 kb or 3.1% of the genome. Independent evolution of retrotransposon sequences was evidenced by the identification of a single-base pair insertion/deletion that distinguishes the highly similar Ty1 and Ty2 LTRs and the identification of a distinct Ty1 subfamily (Ty1Ј). Whereas Ty1, Ty2, and Ty5 LTRs displayed a broad range of sequence diversity (typically ranging from 70%-99% identity), Ty3 and Ty4 LTRs were highly similar within each element family (most sharing >96% nucleotide identity). Therefore, Ty3 and Ty4 may be more recent additions to the S. cerevisiae genome and perhaps entered through horizontal transfer or past polyploidization events. Distribution of Ty elements is distinctly nonrandom: 90% of Ty1, 82% of Ty2, 95% of Ty3, and 88% of Ty4 insertions were found within 750 bases of tRNA genes or other genes transcribed by RNA polymerase III. tRNA genes are the principle determinant of retrotransposon distribution, and there is, on average, 1.2 insertions per tRNA gene. Evidence for recombination was found near many Ty elements, particularly those not associated with tRNA gene targets. For these insertions, 5Ј-and 3Ј-flanking sequences were often duplicated and rearranged among multiple chromosomes, indicating that recombination between retrotransposons can influence genome organization. S. cerevisiae offers the first opportunity to view organizational and evolutionary trends among retrotransposons at the genome level, and we hope our compiled data will serve as a starting point for further investigation and for comparison to other, more complex genomes.

L1 elements are highly repeated mammalian DNA sequences whose structure suggests dispersal by retrotransposition. A consensus L1 element encodes a protein with sequence similarity to known reverse transcriptases. The second open reading frame from the human L1 element L1.2A was expressed as a fusion protein targeted to Ty1 virus-like particles in Saccharomyces cerevisiae and shown to have reverse transcriptase activity. This activity was eliminated by a missense mutation in the highly conserved amino acid motif Y/F-X-D-D. Thus, L1 represents a potential source of the reverse transcriptase activity necessary for dispersion of the many classes of mammalian retroelements.

More than one million copies of the ∼300-bp Alu element are interspersed throughout the human genome, with up to 75% of all known genes having Alu insertions within their introns and/or UTRs. Transcribed Alu sequences can alter splicing patterns by generating new exons, but other impacts of intragenic Alu elements on their host RNA are largely unexplored. Recently, repeat elements present in the introns or 3′-UTRs of 15 human brain RNAs have been shown to be targets for multiple adenosine to inosine (A-to-I) editing. Using a statistical approach, we find that editing of transcripts with embedded Alu sequences is a global phenomenon in the human transcriptome, observed in 2674 (∼2%) of all publicly available full-length human cDNAs (n = 128,406), from >250 libraries and >30 tissue sources. In the vast majority of edited RNAs, A-to-I substitutions are clustered within transcribed sense or antisense Alu sequences. Edited bases are primarily associated with retained introns, extended UTRs, or with transcripts that have no corresponding known gene. Therefore, Alu-associated RNA editing may be a mechanism for marking nonstandard transcripts, not destined for translation

Using a selective screening strategy to enrich for active L1 elements, we isolated 13 full-length elements from a human genomic library. We tested these and two previously-isolated L1s (L1.3 and L1.4) for reverse transcriptase (RT) activity and the ability to retrotranspose in HeLa cells. Of the 13 newly-isolated L1s, eight had RT activity and three were able to retrotranspose. L1.3 and L1.4 possessed RT activity and retrotransposed at remarkably high frequencies. These studies bring the number of characterized active human L1 elements to seven. Based on these and other data, we estimate that 30-60 active L1 elements reside in the average diploid genome.

The abundance of short and long interspersed nuclear sequences (SINEs and LINEs) and pseudogenes in eukaryotic genomes indicates that reverse transcriptase (RT)-mediated phenomena are important in genome evolution. However, the mechanisms involved in their spread are largely unknown. We have developed a selection system in the yeast Saccharomyces cerevisiae to test whether RT-mediated events could be linked to the repair of double-strand breaks (DSBs). Here we show that DSBs can be fixed by the insertion of complementary DNAs at the break site. In the presence of functional RT (from human L1, yeast Tyl or Crithidia CRE1), and in the absence of homologous recombination, an HO endonuclease-induced DSB at the mating type (MAT) locus is the primary site at which a marked cDNA is observed among surviving cells. The structure and junctional sequences of these insertions suggest that repair occurs primarily by non-homologous recombination. Our data support a role for endogenous retroelements in the repair of chromosomal breaks.