Mechanism of IS1 transposition in E. coli: choice between simple insertion and cointegration

Insertion element IS1 and IS1-based transposon Tn9 generate cointegrates (containing vector and target DNAs joined by duplicate copies of IS1 or Tn9) and simple insertions (containing IS1 or Tn9 detached from vector sequences). Based on studies of transposon Tn5 we had proposed a conservative (non-replicative) model for simple insertion. Others had proposed that all transposition is replicative, occurring in a rolling circle structure, and that the way DNA strands are joined when replication terminates determines whether a simple insertion or a cointegrate is formed.--We selected for the transposition of amp and cam resistance markers from pBR322::Tn9 plasmids to an F factor in recA-E. coli and identified products containing three and four copies of IS1, corresponding to true cointegrates (from monomeric plasmids), and simple insertions (from dimeric plasmids). The simple insertions with four copies of IS1 outnumbered those with three by a ratio of about 3:1, whereas true cointegrates containing three copies of IS1 were more numerous than those with four.--A straightforward rolling circle model had predicted that the simple insertions containing three copies of IS1 should be more frequent than those with four. Because we obtained the opposite result we propose that simple insertions only arise when the element fails to replicate or if replication starts but then terminates prematurely. The two classes of products, simple insertions and cointegrates, reflect alternative conservative and replicative fates, respectively, of an early intermediate in transposition.