Selective constraints on P-element evolution.

Abstract

P elements, like mariners, inhabit eukaryotic genomes and transpose via a DNA intermediate. Mutant and wild-type elements in the same genome should be transposed with equal probability by trans-acting transposase, and so no selection should counteract the accumulation of inactivating mutations in transposase genes. Thus, copies of mariner elements diverge within a host species under no selection (Robertson and Lampe 1995). It is unknown whether or not this pattern holds for P elements, which are unrelated to mariner elements but share the same life history. Publicly available P-element sequences were analyzed for evidence of conservative selection for the function of P-element-encoded proteins. Results were compared to predictions derived from several hypotheses that could explain selection, or the lack of it. P-element protein-coding sequences do evolve under conservative selection but apparently because of more than one selective force. Of the four exons in the P-element transposase, the first three (exons 0, 1, and 2) can be translated alone into a repressor of transposition, while the last (exon 3) is only expressed as part of the full-length transposase and probably serves a transposition-specific role. As full-length P-element copies diverge from each other within a host population, selection maintains exons 0-2 but apparently not exon 3. The selection acting on exons 0-2 may act at the host level for repression of transposition (since host level selection does act on orthologous truncated elements that contain only exons 0-2). Evidence of selection on exon 3 is only found in comparisons of more diverged elements from different species, suggesting that selection for transposition acts primarily at horizontal transfer events. Thus, horizontal transfer events may be the sole source of the selection that is crucial to the maintenance of autonomous P elements in the face of mutation (as suggested by Robertson and Lampe 1995). The predictions derived here suggest a strategy for collecting sequence data that could definitively answer these questions.