Insertion sequence distribution bias in Archaea.

Insertion sequences (IS) are common transposable elements in Archaea. Intergenic IS elements are usually less harmful than intragenic ISs, simply because they are less likely to disrupt host gene function. However, because regulatory sequences are intergenic and upstream of genes, we hypothesized that not all intergenic regions are selectively equivalent for IS insertion. We tested this hypothesis by analyzing the distributions of intergenic IS elements within 155 fully sequenced archaeal genomes. Of the 22 genomes with enough IS elements for statistical analysis, five have significantly fewer ISs between divergently oriented neighboring genes than expected by chance, and seven have significantly more ISs between convergently oriented genes. Furthermore, of the 85 genomes with at least one expected IS within each of the three possible neighboring gene orientations (i.e., divergent, convergent, and tandem), 73 genomes have fewer ISs between divergently oriented genes than expected, and 60 have more ISs between convergently oriented genes than expected (both values deviate significantly from binomial probabilities of random distribution). We suspect that these non-random IS distributions are molded by natural selection resulting from differential disruption of neighboring gene regulation, and that this selective pressure has affected transposable element distributions in prokaryotes for billions of years.