When transposition is from one locus to a second locus for certain transposons, a copy of the transposable element is left behind at the first locus. An analysis of transposon mutants revealed an interesting fact about the mechanism of transposition. Using the transposon Tn3 (Figure 20-22), researchers grouped the mutations that prevent transposition into two categories. A trans-recessive class maps in the gene that encodes the transposase enzyme, a catalyst of transposition. A second class of cis-dominant mutations results in the buildup of an intermediate in the transposition process.Figure 20-23 diagrams the transposition pathway in the Tn3 transposition from one plasmid to another. The intermediate is a double plasmid, with both donor and recipient plasmid being fused together. The combined circle resulting from the fusion of two circular elements is termed acointegrate. Apparently, the mutations in this second class delete a region on the transposon at which a recombination event takes place that resolves cointegrates into two smaller circles. This region, called the internal resolution site (IRS), appears in Figure 20-22.

The finding of a cointegrate structure as an intermediate in transposition helped establish a replicativemode of transposition for certain elements. In Figure 20-23, note how the transposable element is duplicated in the fusion event and how the recombination event that resolves the cointegrate into two smaller circles leaves one copy of the transposable element in each plasmid.

Some transposons, such as Tn10, excise from the chromosome and integrate into the target DNA. In these cases, DNA replication of the element does not occur, and the element is lost from the site of the original chromosome. Researchers demonstrated this lack of replication by constructing heteroduplexes of λTn10 derivatives containing the lac region of E. coli. The researchers used DNA from Tn10-lacZ+ and Tn10-lacZ− derivatives. The heteroduplexes, therefore, contain one strand with the wild-type lac region and a second strand with the mutated (Z−) lac region. Figure 20-24 diagrams this part of the experiment. The heteroduplex DNA is used to infect cells that have no lac genes, and transpositions of the TetR Tn10 are selected. Different types of colonies arise from the transpositionof a heteroduplex Z−/Z+ carrying transposon (Figure 20-25). If replication takes place (the replicativemode of transposition), all colonies are either completely Lac+ or completely Lac−, because the replication will convert the heteroduplex DNA into two homoduplex daughter molecules. The mechanism by which this conversion takes place will be examined in detail in the next section. However, if the transposition is conservative and does not include replication, each colony arises from a lacZ+/lacZ− heteroduplex. Such colonies are partly Lac+ and partly Lac−. By using media that stain Lac+ and Lac− cells different colors, researchers can observe the Lac+ and Lac− sectors in colonies.

Therefore, the determination of whether Tn10 undergoes replicative or conservative transposition can be made by observing whether differently colored sectors exist within the same colony resulting from the transposition. Sectored colonies are observed in a majority of cases (Figure 20-26). Thus, Tn10—and perhaps other transposable elements in E. coli—transpose by excising themselves from the donor DNA and integrating directly into the recipient DNA.

The molecular consequences of transposition reveal an additional piece of evidence concerning the mechanism of transposition: on integration into a new target site, transposable elements generate a repeated sequence of the target DNA in both replicative and conservative transposition. Figure 20-27depicts the integration of IS1 into a gene. In the example shown, the integration event results in the repetition of a 9-bp target sequence. Analysis of many integration events reveals that the repeated sequence does not result from reciprocal site-specific recombination (as is the case in phage λ integration; see page 229); rather, it is generated in the process of integration itself. The number of base pairs is a characteristic of each element. In bacteria, 9-bp and 5-bp repeats are most common.

The preceding observations have been incorporated into somewhat complicated models oftransposition. Most models postulate that staggered cleavages are made at the target site and at the ends of the transposable element by a transposase enzyme that is encoded by the element. One end of the transposable element is then attached by a single strand to each protruding end of the staggered cut. Subsequent steps depend on the mode of transposition (replicative or conservative).

Transposable elements generate a high incidence of deletions in their vicinity. These deletions emanate from one end of the element into the surrounding DNA (Figure 20-28). Such events, as well as element-induced inversions, can be viewed as aberrant transposition events. Transposons also give rise to readily detectable deletions in which part of the element is deleted together with varying lengths of the surrounding DNA. This process of imprecise excision is now recognized as deletions or inversions emanating from the internal ends of the IR segments of the transposon. The process ofprecise excision—the loss of the transposable element and the restoration of the gene that was disrupted by the insertion—also occurs, although at very low rates compared with the frequencies of the events just described.