Transposable elements are discrete DNA segments that can move between nonhomologous positions within a genome and have been found in virtually all organisms examined. The recombination pathway by which such elements move is called transposition. Most elements encode a transposase, that is, the recombinase that executes the DNA breakage and joining reactions that underlie transposition, as well as special recombination sequences at the ends of the transposon arranged as inverted terminal repeats that include transposase binding sites; elements lacking a transposase can often be mobilized by the transposase from another cognate element. The insertion of a transposable element into a new insertion site alters the host DNA at that point and often results in a mutation through gene disruption.
Transposition requires the binding of transposase to both ends of the transposable element and subsequent synapsis of the ends. This requirement for synapsis prior to DNA breakage and joining ensures that an intact two-ended element is present. Many elements—from bacteria to Drosophila to fish—transpose by a "cut and paste" mechanism, in which Mg -dependent double-strand breaks at the ends of the element separate the transposon from the donor backbone; that is, the transposable element is excised from the donor site (1) (see Fig. 2 in Transposable Elements). The cleavages that expose the 3′OH ends are key, because these reactions expose the transposon ends that transposase will join to the target DNA. Cleavage of the other strand at the 5′ ends of the element can occur by a variety of means, including simple endonucleolytic cleavage by the transposase or using an alternative transposase subunit specific for 5′ ends cleavage (2).
In the subsequent target-joining step, the exposed 3′OH ends of the transposon directly attack the target DNA. It is important to appreciate that this excision and insertion cycle occurs in the absence of any covalent protein-DNA intermediates; instead, these steps occur via simple one-step transesterification reactions in which H2O attacks the transposon end to disconnect it from the donor DNA, and the transposon end joins to the target DNA by the direct attack of the 3′OH transposon end (1). This attack of the transposon ends occurs at staggered positions on the target DNA—that is, the positions of end joining displaced from each other by several nucleotides. Because of these staggered positions of attack, a small gap the size of the stagger flanks each end of the newly inserted element. Repair of this gap occurs by the host DNA repair functions to generate intact duplex DNA, in which the newly inserted transposon is flanked by a target sequence duplication.
The transposition of retroviruses and retroviral-like transposons occurs by making a messenger RNA copy from the DNA provirus present in a genome, which is then turned into a DNA copy of the element by reverse transcriptase (3). The actual DNA breakage and joining steps that underlie the insertion of this DNA copy into a target DNA are very similar to the mechanisms described above for bacterial elements and for elements from C. elegans and Drosophila. The ends of the viral DNA are often trimmed to expose the actual 3′OH ends of the transposon, and the resulting 3′OH ends the attack the target DNA at staggered positions. As with the excision-insertion elements described above, the newly inserted transposon is flanked by target-sequence duplications resulting from the repair of these gaps.
For some bacterial elements, in particular Mu phage and probably also Tn3, transposition involves 3′ end cleavage and joining of the ends to the target DNA, but no 5′ end cleavage. The products of these transposition reactions are different than the simple insertions that arise by the cut-and-paste pathway (Fig. 1). Transposon end cleavage occurs only at the 3′ ends of Mu; the 5′ ends of Mu remain attached to the donor DNA. Thus, when the 3′ ends attack the target DNA, the transposon is still linked to the donor DNA via its 5′ ends and is now also linked to the target DNA. This structure is variably called a fusion product, a strand transfer product, or a Shapiro intermediate. The exposed target ends that flank the newly inserted transposon have 3′OH ends that can serve as primers for DNA replication. Such replication results in a structure called a cointegrate, which contains two transposon copies linked by the donor backbone and the target DNA. In this reaction, as opposed to a simple cut-and-paste reaction, the transposon is copied during recombination; thus this type of reaction is called replicative transposition.
Figure 1. Simple insertion and cointegrate products of transposition. Some elements transpose through a cut-and-paste mechanism in which the transposon is completely excised from the donor site and is then inserted into the target DNA. Other elements, such as phage Mu and Tn3-like elements, carry out replicative transposition in which an additional copy of the element is made by DNA replication. In this pathway, recombination begins with cleavage to expose the 3′ ends of the element and the resulting 3′OH termini attack the target DNA. Replication of the resulting joint structure from 3′OHs in the target DNA that flank the newly inserted element results in a molecule called a cointegrate in which two copies of the transposon link the donor and target replicons. Recombination between the directly repeated copies of the element can generate two species, one which looks like an intact donor molecule and the other which looks like a simple insertion. Tn3-like elements encode a special resolvase enzyme and recombination site within the element that very efficiently promote resolution.
For Mu, the final product is a cointegrate; Tn3 cointegrates are, however, processed further. Tn3 encodes both a transposase and a resolvase, an enzyme that can act at special sites to exchange DNA duplexes. Resolvase action on the cointegrates results in a target molecule that contains a copy of the transposon and regeneration of a donor molecule containing the transposon. It should be noted that different transposition reactions can both give rise to simple insertion products; for example, both cut-and-paste elements and Tn3-like elements yield similar insertions as their final product, but are generated by distinct mechanisms. It is very difficult to infer the recombination mechanism from in vivo studies, because the products observed may have undergone other recombination reactions prior to or after the actual transposition event.
Another feature that affects whether transposition appears replicative or not is how the gapped donor DNA is dealt with after translocation of an element to a new site. In some cases, the gapped backbone is repaired by double-strand gap repair using a sister chromosome as a template (4-6); in this case, the donor can be restored to its transposon-containing state. When such repair occurs, transposition appears replicative; that is, there is one transposon copy at the donor site and one at the target site, although transposition itself occurred by a nonreplicative cut-and-paste mechanism, and the transposition copy at the donor site was generated by homologous recombination. In other cases, the gapped donor is repaired by an end-joining reaction, generally without restoration of the donor to the original pre-transposon state. Thus transposition can leave "footprints" that may alter gene expression at the insertion site, even though the transposable element itself is no longer present (4, 6, 7). With retroviruses and retrotransposons, the "donor" provirus site is not altered, because the translocation substrate is a DNA copy of the element made by reverse transcription of an RNA copy.
