Transposition and site-specific recombination: adapting DNA cut-and-paste mechanisms to a variety of genetic rearrangements (original) (raw)

Abstract

In bacteria, two categories of specialised recombination promote a variety of DNA rearrangements. Transposition is the process by which genetic elements move between different locations of the genome, whereas site-specific recombination is a reaction in which DNA strands are broken and exchanged at precise positions of two target DNA loci to achieve determined biological function. Both types of recombination are represented by diverse genetic systems which generally encode their own recombination enzymes. These enzymes, generically called transposases and site-specific recombinases, can be grouped into several families on the basis of amino acid sequence similarities, which, in some cases, are limited to a signature of a few residues involved in catalysis. The well characterised site-specific recombinases are found to belong to two distinct groups, whereas the transposases form a large super-family of enzymes encompassing recombinases from both prokaryotes and eukaryotes. In spite of important differences in the catalytic mechanisms used by these three classes of enzymes to cut and rejoin DNA molecules, similar strategies are used to coordinate the biochemical steps of the recombination reaction and to control its outcome. This review summarises our current understanding of transposition and site-specific recombination, attempting to illustrate how relatively conserved DNA cut-and-paste mechanisms can be used to bring about a variety of complex DNA rearrangements.

1 Introduction

It is now well documented that genetic information can be reshuffled by inversion, insertion, duplication, deletion, or translocation of DNA segments. Whereas some rearrangements fulfil specific physiological functions or are involved in programmed processes, others occur ‘spontaneously’ with respect to time and position in the genome and contribute to the genetic diversity of a population. In bacteria, as in most organisms, DNA rearrangements can be mediated by general recombination between homologous DNA sequences, but also by a variety of specialised recombination mechanisms commonly grouped into two categories: genetic transposition and conservative site-specific recombination.

Although transposition and site-specific recombination seem to be fundamentally distinct processes that often have very different biological outcomes, increasing evidence indicates that they are related in many ways. In general recombination, genetic material is exchanged through a cascade of events involving multiple proteins assembled in different enzymatic complexes [1]. In contrast, both categories of specialised recombination systems utilise relatively simple recombination machineries in which one (or sometimes two) enzyme, generically referred to as the transposase and site-specific recombinase, catalyses the essential DNA breakage and joining reactions. These proteins act at specific DNA sequences which are characteristic for each genetic element.

For all systems analysed, coordination of the recombination reactions requires the formation of an enzymatically competent nucleoprotein complex in which the recombinase and at least two distant DNA sites are brought together. Whereas the molecular transactions controlling the assembly of this active complex can vary from one system to another within the same family, unrelated systems also appear to use convergent strategies. As a consequence, a single recombination mechanism can be adapted to many different recombination functions. Conversely, recombination by distinct biochemical mechanisms may show very similar outcomes.

The present review attempts to give a general survey of the specialised recombination repertoire in bacteria, trying to outline the differences, but also discussing some aspects which in the light of recent studies, sound very much like ‘variations on a theme’. Although this review is focused on bacterial recombination systems, those in eukaryotes show many similarities. Other discussion on site-specific recombination and transposition can be found in recent reviews [2–10].

2 Transpositional and site-specific recombination: overview

2.1 A plethora of transposable elements with common transposition motifs

Transposable elements or transposons can be described as discrete DNA segments that are able to move between different, non homologous, genomic loci. Transposition of an element is thus a recombination reaction involving three separate sites: the two transposon ends and the new target locus. In autonomous transposable elements, the transposase is encoded by a gene located within the element. Transposon ends often contain inverted repeated sequences on which the transposase binds specifically. Although most transposons can transpose into many different places, insertion is never totally random. While a specific target consensus sequence has been found for some elements, the insertion of many transposons is also influenced by regional features of the target locus such as a local DNA structure, the presence host protein binding sites, or DNA supercoiling (reviewed in Ref. [10]). The transposon Tn_7_ is distinguished by its ability to transpose either to a specific site of the bacterial chromosome (attTn7) at a high frequency or to many different sites at a low frequency (see below) [11].

Bacterial transposons utilise two major modes of transposition. In a non-replicative or ‘cut-and-paste’ mechanism, the element is excised from its original location and inserted into the new target locus. This mode of transposition results in the simple insertion of the element in the target DNA (Fig. 1). The gapped donor molecule resulting from the transposon excision is either degraded or repaired by homologous recombination using a second copy of the donor as a template [12, 13]. Rarely, the ends of the donor molecule may be resealed. In replicative transposition, the element is copied so that insertion of the element into the same DNA molecule leads to deletion and/or replicative inversion, while transposition from one circular molecule to another generates a cointegrate structure in which the donor and target backbones are joined by directly repeated copies of the transposon at each junction. This cointegrate may be subsequently resolved by recombination between the two copies of the element, giving rise to a restored donor molecule and a target molecule in which one copy of the transposon has inserted (Fig. 1) [9]. Both modes of transposition generate short target sequence duplications of a characteristic length (2–14 bp) that form direct repeats flanking the element in its new locus. These duplications are created by repair of the gaps arising from staggered cuts of the target DNA (see below).

1

Two major modes of transposition. In non-replicative transposition, the element (rectangle) is excised from the donor locus by double-strand breaks at both ends (solid triangles) and transferred to the target site (cut-and-paste mechanism). The product of intermolecular transposition by the replicative mechanism is a cointegrate molecule in which the donor and target backbones are fused. The element is duplicated during the process and each copy remains attached to the donor backbone at one end, whereas the other end is joined to the target. The cointegrate may be resolved by homologous or site-specific recombination between the two transposon copies to restore the initial donor molecule and yield a target copy in which the element is inserted. Both modes of transposition usually generate short target duplications flanking the element in the new target locus (open triangles).

Beyond this overall description hides a broad diversity of transposable elements in bacteria. More than 400 different elements have been identified and characterised from a wide range of archaebacteria and eubacteria (J. Mahillon and M. Chandler, personal communication). The size and genetic organisation of bacterial transposons is highly variable. They range from relatively small and compact (0.8–2 kb) insertion sequences (IS), which in the simplest form contain a single transposase gene [14], to larger (and more complex) multi-drug resistance transposons such as Tn_7_ (14 kb) [11], or the bacteriophage Mu (37.5 kb) which uses transposition as part of its life cycle strategy [6, 15].

Historically, bacterial transposable elements have been grouped into several classes on the basis of different criteria. However, it is quite clear that the largest class, the IS elements, defined by the sole feature that they only encode functions required for transposition, actually represents a very heterogeneous group of genetically and functionally distinct transposable elements. The identification and comparison of an increased number of transposons has made an alternative classification possible based on sequence similarities ([14]; J. Mahillon and M. Chandler, personal communication). These studies not only revealed the existence of several bacterial families of related elements, but more importantly, they also identified conserved transposase domains amongst a broad diversity of transposons from prokaryotes and eukaryotes. This was first shown for the IS_3_ family of transposases which exhibit striking similarities with the integrase proteins of retroviruses and retrotransposons, notably by the presence of a invariant triad of acidic residues termed the DDE motif [16–18]. Subsequently, a DDE-like signature has been identified within the transposase of many different transposons including IS_10_, Tn_7_ and the bacteriophage Mu, which are amongst the best characterised bacterial elements [19–22]. The widespread occurrence of this signature indicates that transposases and retroviral integrases from very different sources form a super-family of recombination enzymes which share a common catalytic mechanism.

It is worth noting that not all the proteins involved in transposition contain a DDE motif, and that some of them clearly belong to completely distinct classes of enzymes. For example, the recombinases involved in the translocation of different kinds of conjugative transposons and members of the IS_110_/IS_492_ family, belong to three different groups of site-specific recombinases [23–25]. IS_1_ transposase also shares similarities with the restriction endonuclease Eco_RII [26]. In contrast, the IS_91 family transposase is related to proteins involved in replication by rolling-circle mechanism [27, 28]. Therefore, it is not surprising that these elements appear to use quite different recombination mechanisms.

Transposable elements are often regarded as selfish genetic entities that have evolved a set of mechanisms to replicate faster than their host [29]. An extreme case is bacteriophage Mu which transposes into the host chromosome for the establishment of lysogeny and, during lytic growth, uses the host machinery to replicate its own DNA by undergoing multiple rounds of replicative transposition [6, 15]. By promoting various DNA rearrangements such as deletions, inversions, and replicon fusions, transposable elements can also be viewed as naturally occurring ‘genetic engineers’ which by modifying the genome of their host may contribute to the genetic diversity of a population [30]. Through their ability to translocate genes, transposons are also potent agents for horizontal transfer and in that respect, they play a crucial role in the spread of antibiotic resistance factors amongst bacteria.

2.2 Site-specific recombination: integration, resolution, inversion and transposition

In DNA rearrangements mediated by site-specific recombination, four DNA strands are broken, exchanged and resealed at specific positions of two separate recombination sites [3–5, 31]. The outcome of a recombination event depends on the relative disposition of the two sites. Intramolecular recombination between inverted or directly repeated sites will invert or excise respectively the intervening DNA segment. Recombination between sites on separate DNA molecules will integrate one molecule into the other (Fig. 2). Although these DNA rearrangements are reminiscent of those promoted by transposable elements, one main difference is that site-specific recombination occurs between predetermined loci, whereas rearrangements mediated by transposons largely rely on their target-site specificity. A second major difference lies in the conservative nature of site-specific recombination, i.e., DNA strand exchange is completed without involving any DNA synthesis or degradation. That allows site-specific recombination to be reciprocal, i.e., a second round of recombination can restore the initial DNA configuration.

2

Outcomes from site-specific recombination. Triangles show the orientation of the recombination sites. a and b indicate the position of distinct genetic markers around the recombination loci. ‘Excision’ and ‘integration’ refer to recombination events involving genetic entities of different size and/or function (e.g., the bacterial chromosome and a phage genome), whereas ‘resolution’ and ‘fusion’ apply to equivalent DNA molecules, (e.g., two plasmids).

As a consequence of these distinguishing features, site-specific recombination is exploited for a range of biological functions [5]. In a number of examples, site-specific inversion is used to switch the expression of genes between alternative patterns [32]. As the genes controlled by these systems often code for surface structures proteins, e.g., bacteriophage tail fibres, flagellar antigens or pilin, recombination provides a source of genetic variability which helps the host organism to adapt to environmental fluctuations. For these different systems, alternative gene expression is achieved, either by flipping the orientation of a promoter between two structural genes or by connecting variable C-terminal portions of a gene to the constant N-terminal region located outside the invertible segment [32]. An extreme example of the latter is provided by the shufflon system of R64 plasmid in which a set of four invertible fragments controls the selection of seven C-terminal segments for the _pil_V gene [33]. DNA splicing reactions by site-specific recombination between directly repeated sites has also been reported as a mechanism of gene activation in developmentally regulated prokaryotic processes, such as Bacillus subtilis sporulation [34] and Anabaena sp. heterocyst differentiation [35].

Transposons of the Tn_3_ family encode a site-specific resolution system to reduce the cointegrate intermediates generated by their replicative mode of transposition [36]. Another widespread function of site-specific recombination is to resolve replicon dimers arising from homologous recombination in order to facilitate their segregation to daughter cells at the time of cell division. Whereas several different replicons encode their own resolution systems (e.g., Refs. [37–39]), plasmids of the ColE1 family utilise the multifunctional recombination system Xer encoded by the Escherichia coli chromosome [40]. By acting at a specific recombination site (dif) located in the terminus region of replication, Xer recombination also appears to play an active role in the partitioning of the bacterial chromosome itself. Bacteria deficient in Xer recombination form filaments with aberrant nucleoids, presumably as a result the accumulation of chromosome dimers [41]. This seems to be an important physiological role of site-specific recombination that is highly conserved among bacteria.

Finally, the paradigm of site-specific recombination is the reaction by which bacteriophage λ and related phages choose whether they prefer to remain dormant as a prophage or to embark on the lytic cycle of phage growth by integrating into and excising from the host genome [4, 42]. A similar excision/integration reaction is used by different systems to translocate non-replicating genetic material, e.g., antibiotic resistance genes, between different genomic loci. Integrons are composite recombination systems that exchange mobile gene cassettes between specific expression units located on various replicons or within transposons [43]. Conjugative transposons have combined site-specific recombination and conjugation mechanisms to transfer a circular excised form of the element between a donor and a recipient cell [23]. A remarkable feature of integrons and conjugative transposons is their low level of target site selectivity when compared to other site-specific recombination systems, illustrating the potential for a recombination mechanism to accommodate distinct functions.

The functional flexibility of site-specific recombination is also demonstrated by the fact that the various biological functions listed above are fulfilled by recombinases belonging to two major families of enzymes using distinct biochemical mechanisms. The resolvase/invertase family, of which there are currently approximately 40 different members, forms a rather homogenous group of related proteins in which a conserved serine residue plays a key catalytic role [44, 45]. Recombinases of the λ integrase (λ Int) family (∼60 members) are much more divergent, with only four completely invariant residues intimately involved in catalysis: the RHRY tetrad [46–48]. There is no functional segregation between the two families and systems of both groups are used to perform virtually all kinds of rearrangements [5]. As for DDE transposases, these two families are not exclusive and there are a few recombinases that appear to belong to neither group. An interesting example is the Piv invertase of Moraxella sp. which together with the transposases of unusual IS elements (the IS_110_/IS_492_ family) form what can be viewed as a third family of related recombinases [25]. Like conjugative transposons, several IS elements of this family fail to generate target duplications at the insertion site and/or produce an excised circular form as a transposition intermediate.

3 Three mechanisms of DNA cut-and-paste

The establishment of a cell-free system for different types of specialised recombination has allowed elucidation of the biochemical steps of the DNA strand exchange reactions. These in vitro studies have identified three different mechanisms for cutting and rejoining DNA molecules, corresponding to the three major classes of proteins described above: the DDE transposases, the resolvase/invertase, and the λ Int families of site-specific recombinases. A common feature of the three mechanisms is that they proceed by transesterification reactions without requiring high-energy cofactors such as ATP (Fig. 3) [3, 31, 49]. The energy of the broken phosphodiester bonds is conserved for the formation of new bonds. The DDE transposases use a one-step transesterification mechanism, whereas the two distinct families of site-specific recombinases use contrasting two-step transesterification mechanisms involving different amino acid residues in the formation of a covalent DNA-enzyme intermediate (Fig. 3). Understanding of these mechanisms has now reached the atomic level of resolution by having obtained crystal structure data for each of the three families of proteins.

3

Chemistry of transposition and site-specific recombination reactions. Recombination DNA strand breakage and joining occur by transesterification reactions in which the phosphate of the scissile phosphodiester bond is subject to nucleophile attack by a hydroxyl group (arrows). Endonucleolytic cleavage at the transposon ends (A) and the strand-transfer reaction that join the ends to the target DNA (B) are one-step transesterifications in which the nucleophile is a water molecule and the 3′-OH end of the element, respectively. Strand exchange catalysed by site specific recombinases (C and D) occurs by two steps of transesterification (cleavage and rejoining) involving a covalent protein-DNA intermediate. The nature of the catalytic residue and the line of entry of the nucleophile is different between the two recombinase families. For cleavage catalysed by the invertase/resolvase family (C), the nucleophile hydroxyl is derived from a serine and the leaving group is the 3′-OH of the deoxyribose. For the λ integrase family (D), the catalytic residue is a tyrosine and the leaving group is the 5′-OH. For both recombinase families, the rejoining step is the reverse of the cleavage step. Phosphate backbones are drawn in thick and thin lines to distinguish the donor and target DNA (panel B) or the two recombination partner DNA strands (panels C and D).

3.1 The DDE transposases: a universal one-step transesterification reaction

The biochemistry of reactions catalysed by the DDE recombinases has been examined in detail for several different systems including three bacterial transposable elements: the bacteriophage Mu, IS_10_ and Tn_7_ (for recent reviews, see [6, 11, 15, 50]. In all cases, the transposase executes a set of critical reactions that join the 3′ ends of the element to the target DNA, whereas connection of the 5′ ends and completion of transposition requires processing events performed by host repair and replication functions (Fig. 4). An important variation between the three bacterial systems is in the number and nature of cuts that sever the transposon from the flanking donor DNA, resulting in different transposition end-products. For the three transposons, transposition initiates with a pair of specific single strand cleavages exposing the 3′-OH ends of the element (Fig. 4; see also Fig. 3A). For IS_10_ and Tn_7_ which use a non-replicative cut-and-paste mechanism, the other strand is also cleaved to reveal the transposon 5′ ends and to excise the element from its initial genomic locus. IS_10_ excision yields flush transposons ends [51], whereas the double-strand breaks at the ends of Tn_7_ are staggered by three nucleotides, with the 5′ end cleavages occurring in the adjacent donor backbone [52]. In sharp contrast, the second strand cleavage does not occur at this stage of bacteriophage Mu transposition which remains connected at its 5′ ends to the flanking DNA sequences. In a second step, the 3′-OH ends of the excised transposon (IS_10_ and Tn_7_) or the nicks in the donor molecule (Mu) participate in a concerted strand transfer reaction that join both ends of the element to staggered phosphates of the two target DNA strands (Fig. 4; see also Fig. 3B).

4

Biochemical steps underlying the non-replicative or replicative transposition of three bacterial elements. The shaded rectangles represent the DNA strands of the transposable elements ends. For IS_10_ and Tn_7_, reactions occurring at a single end are shown. Black and white rectangles are the flanking donor sequences and the target DNA, respectively. The black and white arrowheads show the 3′-end and 5′-end cleavage, respectively. Curved arrows indicate the nucleophilic attack transferring the 3′-OH ends on staggered phosphates of the target DNA (black dots). Crenellated lines represent the few target nucleotides that are duplicated during the transposition process. For the three elements, the biochemical steps are catalysed by the transposase in a complex where the transposon ends are in synapsis. For IS_10_, the target is captured after completion of the double-strand breaks at the transposon ends, whereas for Mu and Tn_7_, the presence of the target within the complex is required to activate the cleavage reactions. The cross-hatching represents replication events that complete transposition after complex dissociation.

In the case of IS_10_ and Tn_7_, host processing of the resulting strand transfer intermediate results in the filling-in of the short single-strand gaps lying at each transposon-target DNA junction. This repair, which in the case of Tn_7_ is also presumed to remove the overhanging three nucleotides at the 5′ ends, generates the short target duplications that flank the element in its new locus (Fig. 4). After eventual cleavage of the donor backbone by an undetermined mechanism, the strand transfer product of bacteriophage Mu can be processed in a similar way during the non-replicative integration of the phage. Alternatively, in the absence of donor backbone dissociation, complete replication of the element gives rise to a cointegrate, which is the final product of replicative transposition (Fig. 4).

Variations of the transposition pathway have been reported for both prokaryotic and eukaryotic transposons. In particular, the transposase of IS_911_, a member of the IS_3_ family, has been found to carry out a distinctive single-strand circularisation reaction in which one transposon end is transferred to a target site three nucleotides distant from the other end [53, 54]. In vivo, the resulting ‘figure-eight’ molecule is processed, giving rise to a circular excised form of the transposon thought to be a transposition intermediate that can efficiently insert in a new locus. Although mechanistically distinct, this ‘site-specific’ circularisation of IS_911_ is reminiscent of the excision reaction performed by site-specific recombinases. Interestingly, the formation of excised transposon circles has been observed for other members of the IS_3_ family [54, 55], and also for IS_1_[56] and for diverse eukaryotic transposons, suggesting that it may represent a major mode of transposition [54].

As mentioned above, the initial cleavage that generates the 3′-OH ends and the strand transfer reaction catalysed by bacterial DDE transposases are chemically equivalent to the reactions catalysed by the retroviral integrase proteins (IN) for the integration of the retrovirus cDNA into the genome of an infected cell [6, 7, 57]. IN-mediated cleavages at the ends of HIV cDNA and the subsequent strand transfer step proceed with inversion of chirality at the target DNA phosphates. The same result was obtained for the strand transfer reaction catalysed by the bacteriophage MuA transposase [58, 59]. This analysis indicates that the reactions catalysed by the DDE recombinases occur by a one-step transesterification mechanism in which the scissile phosphodiester bond is directly attacked, either by a water molecule (end cleavage), or by the 3′-OH end of the element (strand transfer) without the formation of a protein-DNA covalent intermediate (Fig. 3). Both transesterification steps require divalent cations (Mg2+ or Mn2+) and it was postulated that the role of the DDE residues was to form a metal ion binding pocket in the active site [49]. This suggestion is strongly supported by the recent finding that the catalytic domain of IN, MuA and other enzymes involved in metal ion-dependent phosphoryl transfers exhibit very similar structures.

The crystal structure of the DDE-containing catalytic core of MuA shows that this region of the protein contains two sub-domains [60]. The 70 C-terminal residues form a β-barrel, on one face of which is a large region of positive potential that could play a role in the non specific DNA binding activity associated with the whole catalytic core. It was proposed that this activity might be important for interactions either with the DNA sequences surrounding the cleavage points at the Mu ends or with the target DNA, in order to position the substrate into the enzyme active site [60]. Remarkably, in spite of a very low level of sequence homology, the structure of the N-terminal sub-domain containing the DDE triad shows striking similarities, not only with the catalytic domain of the IN proteins of HIV and ASV retroviruses, but also with the structure of more functionally distant enzymes such as the E. coli and HIV ribonucleases H (RNase H) and the E. coli Holliday junction resolving enzyme RuvC (reviewed in Refs. [61–63]). In the structure shared by these different proteins (a central five-stranded β-sheet surrounded by α-helices on either side) three or four catalytically important acidic residues, i.e., the DDE motif in the case of the MuA and IN proteins, are clustered to form a possible two metal ion binding pocket as actually observed in the structure of HIV RNase H active site [64]. This structural similarity between unrelated proteins strongly supports the view that the DDE recombinases belong to a larger group of enzymes, including polymerases and ribozymes, that catalyse transesterification reactions using a common ‘carboxylate-chelated two-metal ion’ catalytic mechanism [61, 63]. In this mechanism, initially proposed for the 3′-5′ exonuclease activity of DNA polymerase I, the two chelated metal ions participate in the activation of the nucleophile hydroxyl and the scissile phosphodiester bond by stabilising the phosphate in a penta-coordinated form [65].

However, DDE recombinases are not simple nucleases or polynucleotidyl transferases. These enzymes are distinguished by the fact that they consecutively catalyse two (or three) transesterification reactions at both transposons ends. Successful transposition requires that these two sets of reactions be temporally and spatially coordinated in order to prevent incomplete recombination events which are likely to be deleterious to the host and to the transposable element itself. Recent work from different laboratories has provided new interesting insight into how this control is effected in the case of Mu (reviewed in Refs. [66, 67]). A key step in Mu transposition is the assembly of a synaptic complex into which catalytically inert monomers of MuA become activated by the formation of a tetrameric core tightly bound to the two transposon ends. Assembly of the tetramer thus implies a structural transition by which the enzyme and the DNA cleavage sites become engaged for catalysis ([6, 15]; see below). Complementation experiments performed by mixing distinct catalytic mutants of MuA have shown that the tetramer active sites are built by interlocking separate domains of distinct transposase subunits. In each active site, one monomer provides the DDE catalytic core (domain II), whereas a different monomer donates another catalytic region located in the C-terminal part of the protein [68]. Although the exact function of this second catalytic domain (domain IIIA) is not known, it appears to play a role in the activation and/or positioning of the DNA [69]. For the strand transfer reactions, the active site domain IIIA is provided by MuA monomers occupying the inner part of the Mu ends whereas the DEE-containing domain II is provided by transposase subunits bound on the external sites [70, 71]. By reciprocality of domain sharing, this spatial arrangement appears to be reversed for the ends cleavage reactions [71]. Furthermore, in both reaction steps, the DDE domain II operates in trans, i.e., the subunit bound to one end catalyses the cleavage and joining of the opposite end [70, 72]. This intricate architecture of the MuA enzymatic complex is viewed as a mechanism which ensures that substrate synapsis and catalysis are tightly coupled. The involvement of a structural transition upon tetramerisation and substrate binding is also consistent with the fact that the DDE residues of the MuA catalytic core appear to be in an inactive configuration [60]. Taking these data into account, Yang et al. propose a pathway in which the Mu ends are first nicked within the two cleavage active sites and then swing onto two other catalytic pockets to be transferred on the target DNA [71].

As for Mu, a single transposase protein carries out the three reaction steps underlying IS_10_ transposition and the assembly of a precleaved synaptic complex is required to ensure the coordination of the processing events occurring at both ends [50, 73, 74]. IS_10_ transposition pathway is also highly ordered. Cleavage of the transferred strand (3′ end) always precedes cleavage of the non-transferred strand (5′ end), and complete excision of the transposon is required before a target DNA molecule is captured by the complex for the strand transfer reactions [75, 76]. In sharp contrast to Mu, a single transposase monomer catalyses the three chemical steps of IS_10_ transposition at each of the two transposon ends [77]. Thus, two monomers of transposase appear to carry out the entire transposition reaction without sharing domains. The mechanism which allows the repeated use of a single DDE-containing catalytic pocket is not known. Current models propose that consecutive structural rearrangements must occur between each reaction step to re-adjust the active site to novel substrate configurations [50, 77].

A third and distinctive example of transposase architecture is provided by Tn_7_[11]. In contrast to Mu and IS_10_, Tn_7_ transposition reactions are catalysed by two DDE-containing enzymes with clearly separate activities. TnsA mediates cleavage at the 5′ ends of the transposon whereas TnsB, which is also responsible for the recognition and binding of Tn_7_ ends, promotes the 3′ end cleavage and carries out the strand transfer reaction [22]. Although the respective functions of TnsA and TnsB can be selectively blocked by mutation of their DDE residues, the presence of both proteins is required for the formation of an active complex. Tn_7_ transposase is thus a heteromeric enzyme containing two different catalytic subunits [22]. The possibility that the active sites of the TnsA+B transposase core could be assembled by contribution of distinct TnsA or TnsB protomers has not been examined thus far. As with IS_10_, there is currently no evidence for cis or trans activity by TnsA and/or TnsB. Nevertheless, the biochemical separation of the 5′ and 3′ ends processing reactions has a remarkable implication. Inactivation of TnsA converts the normal cut-and-paste transposition pathway of Tn_7_ into a Mu-like replicative mechanism, the 3′ ends breakage and joining reactions catalysed by TnsB within the transposase core being functionally equivalent to those performed by the MuA tetramer in Mu transposition [78].

These three different examples of ‘division of labour’ within a transposition complex illustrate the remarkable flexibility by which a common chemical mechanism can be adapted in different ways to accomplish complex and highly controlled recombination reactions.

3.2 Site-specific recombinases: two-step transesterifications by distinct mechanisms

Unlike transposases, site-specific recombinases, at least in principle, execute all recombination DNA breakage and joining reactions without involving host repair or replication functions [3–5, 31]. These reactions are biochemically related to those catalysed by the topoisomerase enzymes that regulate the intracellular level of DNA supercoiling [79]. Indeed, site-specific recombinases often exhibit type I topoisomerase activity by which they can relax supercoiled DNA substrates. However, in a recombination reaction between two sites, DNA strands are not only broken and rejoined, they are also exchanged. Therefore, as discussed above for the DDE recombinases, site-specific recombination requires the assembly of a synaptic complex containing multiple recombinase subunits and the relevant DNA recombining partners.

For most of the systems of the two major families (i.e., the resolvase/invertase and the λ Int families), recombination takes place within a short (∼30 bp) DNA segment called the ‘core’ or ‘crossover’ site onto which two recombinase subunits bind, usually by recognising specific sequences with dyad symmetry [3, 80, 81]. In the synaptic complex, the two core sites are brought in close proximity. It is generally admitted that the four recombinase subunits bound on the two duplexes participate in the recombination reaction (Figs. 5 and 6). A few systems, all belonging to the λ Int family, involve two distinct recombinase proteins. In the E. coli inversion system Fim, the two recombinases, FimB and FimE, act independently on the same recombination sites to control the ‘on’ or ‘off’ position of this particular genetic switch [82]. By contrast, the XerC and XerD recombinases cooperate in all Xer-mediated DNA rearrangements and the two proteins bind with distinct specificities to separate regions of the different recombination sites of this system [83, 84].

5

Concerted DNA breakage and rejoining reactions catalysed by resolvase/invertase family enzymes. The subunit rotation model is shown. The ovals represent recombinase subunits with the conserved catalytic serine ‘S’. Thick and thin lines are the top and bottom strands of the recombination sites, respectively. The short vertical bars are the 2 bp of the overlap region between the two cleavage points. Black arrows represent the nucleophilic attacks of phosphates (black dots) by hydroxyl groups (arrowheads). The four DNA strands are cleaved (a), exchanged by 180° rotation of the half-site bound subunits (b) and religated in the recombinant configuration (c).

6

Sequential strand exchange by the λ Int family site-specific recombinases. The DNA strand swapping/isomerisation model is presented. The letter ‘Y’ refers to the conserved catalytic tyrosine. Other symbols are as in Fig. 5. The top strands (thick lines) are cleaved first (a), swapped between the two partners (b), and then religated (c). The branch point of the generated Holliday junction intermediate is positioned at the middle of the (6-bp) overlap region and the top strands are crossed. Isomerisation of the Holliday junction to a recombinant configuration in which the bottom strands are crossed requires the reorganisation of the DNA helices and the four half-sites-bound recombinase subunits within the complex (d). The resulting Holliday junction isoform is resolved by repeating steps a to c in order to exchange the bottom strands (e).

The recombination locus of all systems, even those working with a single recombinase, contain a certain degree of asymmetry so that ‘top’ and ‘bottom’ strands can be distinguished. In the simplest cases, this asymmetry is entirely encoded within the core site whereas in other systems, external elements may contribute to the recombination site polarity by imposing a specific geometry on the synaptic complex ([80, 81]; see also below). The catalytic mechanism used by the two families of site-specific recombinases is different as is the structural organisation of the enzymes.

3.2.1 The resolvase/invertase family: concerted breakage and rejoining of four DNA strands

The best characterised recombinases of this family are the invertases Gin from bacteriophage Mu and Hin from Salmonella sp. and the resolvases of Tn_3_ and γδ transposons [3, 32, 36, 80, 81, 85]. Although the chemistry of strand exchange used by these recombinases appears to be highly conserved, important variation is found in the assembly of the recombination synapse determining the selectivity of the reaction, i.e., inversion or resolution (see below).

In a recombination reaction catalysed by resolvases or invertases, double strand breaks staggered by 2 bp occur at the middle of the two paired core sites, giving rise to recessed 5′ ends and 3′-OH overhangs (Fig. 5; see also Fig. 3C). One recombinase subunit is linked to each of the 5′ ends through the conserved serine residue of the family [86, 87]. This serine presumably provides the primary nucleophile hydroxyl group in the cleavage reaction [45]. The ligation step that follows strand exchange can be viewed as the converse of the cleavage: the protein-DNA phosphoseryl bond of one strand is attacked by the 3′-OH end of the partner to release the enzyme and reseal the DNA backbone in the recombinant configuration (Fig. 5; see also Fig. 3C).

Although cleavages of the four DNA strands or the religation steps can be experimentally uncoupled by using recombination site variants or mutated recombinases, both types of reaction are normally highly coordinated [88, 89]. The cleaved complex in which four enzyme-linked recombination half-sites are held together by recombinase subunit interactions seems to be an obligate intermediate [88]. Thus, recombination by the resolvase/invertase family occurs by a mechanism in which four DNA strands are broken and rejoined in a concerted manner.

The standard substrates for invertases and resolvases are supercoiled molecules (see below) and the topological change induced by recombination has been found to be equivalent to a right-handed 180° rotation of one pair of cleaved half-sites relative to the other prior the rejoining step (Fig. 5) [90–92]. In this mechanism, the 2 bp ‘overlap’ regions that separate the top and bottom strand cleavage positions need to be identical in the two core sites to stably reconnect the recombinant DNA duplexes [93, 94]. It has been shown recently that in reactions involving non-homologous overlap regions, strand exchange mediated by Tn_3_ resolvase proceeds through apparent 360° (2×180°) rotation of the half-sites without rejoining the mis-paired strands in the recombinant (180° rotation) configuration [89].

The crystal structure of the γδ resolvase dimer complexed to the core recombination site has been determined recently [95]. The DNA-bound resolvase monomer contains two globular domains lying on opposite faces of the DNA helix and an extended arm region that connects the two domains. The small C-terminal DNA binding domain is involved in the recognition of the outer part of the core site consensus sequence by making specific contacts in both the major and minor grooves. Its structure, containing a helix-turn-helix DNA binding motif, is similar to that previously found for the DNA binding domain of Hin invertase [96]. The arm region that joins the two globular domains also contributes to DNA binding through interactions in the minor groove. The large N-terminal catalytic domain contains the active site serine and other catalytic residues, as well as a set of residues forming a hydrophobic core at the dimer interface. This domain also appears to be important for higher order interactions between resolvase dimers in the recombination complex [97–99]. The DNA in the co-crystal is bent by 60° away from the enzyme catalytic domain.

The γδ resolvase-DNA complex appears to be in an inactive configuration, the catalytic serines of the two resolvase subunits being too far away from the scissile phosphates to cleave the DNA [95]. However, the active serine of either monomer is closest to the DNA cleavage position proximal to the half-site bound by the recombinase subunit. This correlates with the fact that the active serine and several other catalytic residues of γδ resolvase act in cis on the nearest scissile phosphodiester bond [88]. The crystal structure also suggests that, as in MuA, catalytic residues may be shared between the two monomers to form the active site. While there is currently no experimental evidence to support the possibility for a composite catalytic site, mutations in one resolvase subunit seem to activate the adjacent subunit in trans, presumably by altering inter-subunit interactions at the dimer interface ([88]; M.R. Boocock and N.D.F. Grindley, unpublished). Likewise, mutations in the equivalent dimerisation domain of the DNA invertases Gin, Hin and Cin, make the recombinases more reactive and independent of structural elements controlling the formation of the recombination synaptic complex ([92, 100–103]; see also below). These observations indicate that activation of the active site serine recombinases appears to require conformational changes during the assembly of an enzymatically competent recombination structure as discussed above for the DDE transposases. Indeed, some level of structural flexibility as revealed by the γδ resolvase-DNA co-complex could be compatible with limited distortions [95].

The conformational changes that take place during strand exchange appear less straightforward. Since resolvase cleaves the DNA in cis and seems not to dissociate from its binding site during recombination [104], two models have been proposed to account for the apparent 180° half-site rotation in the reaction. In the ‘subunit rotation’ model, strand exchange is coupled to a rotational rearrangement of the DNA-linked recombinase subunits within the tetramer [91]. A major difficulty with this mechanism is that the recombinase dimer interface holding the cleaved ends in the complex must be disrupted transiently during the dissociation/reassociation process (Fig. 5). An alternative ‘static subunit’ model postulates that the recombinase tetramer does not dissociate and that recombination occurs by localised conformational and topological rearrangement of the DNA molecules within the complex [95, 99]. This second model is difficult to reconcile with the recent observation that Tn_3_ resolvase brings about multiple rounds of rotations without rejoining the single 180° rotation intermediate. It is argued that in a static subunit mechanism, such an iteration of the reaction would entangle the DNA within the catalytic complex to an unacceptable level [89, 105].

3.2.2 The λ Int family: sequential pairs of DNA strand exchange

Unlike the recombinases of the resolvase/invertase family, site-specific recombinase related to λ Int such as the Cre recombinase of phage P1, E. coli XerC and XerD and the Flp protein from the yeast 2μ plasmid, exchange the two pairs of DNA strands separately and sequentially (Fig. 6) [3, 4, 31, 106].

To initiate the first strand exchange, the tyrosine residue of the conserved catalytic motif RHRY attacks a specific scissile phosphate in one strand (defined here after as the top strand) of each recombination core sites, thereby forming a 3′ phosphotyrosyl-linked recombinase-DNA complex and generating a free 5′-OH end (Fig. 6). The polarity of this cleavage reaction is thus reversed when compared to that of the resolvase/invertase-mediated cleavages (compare also Fig. 3C and Fig. 3D). In a second step, the recombinase-DNA phosphotyrosyl bond is attacked by the 5′-OH end from the partner duplex to generate a four-way branched structure, or ‘Holliday junction’ intermediate, in which only two DNA strands have recombined. To resolve this intermediate and complete the recombination reaction, the two other (bottom) strands are exchanged by repeating the cleavage/religation process 6–8 bp downstream of the first strand cleavage position (Fig. 6).

As in the resolvase/invertase-mediated recombination, sequence homology in the 6–8 bp overlap region that separates the top and bottom strand cleavage positions in the two partner recombination core sites is also essential for most (but not all) of the systems belonging to the λ Int family. Although the homology appears to play a role after the synapsis of the two core sites, the exact mechanism by which it influences the reaction remains unclear. The classical model supposes that sequence identity between recombination sites is required for a reversible process called ‘branch migration’ that moves the branch point of the Holliday junction from its site of formation at one end of the overlap region to the site of resolution at the opposite end. This is achieved by stepwise melting and reannealing of the complementary strands of the parental and partner DNA duplexes, [5, 107]. This view is now challenged by an alternative ‘strand swapping/isomerisation’ mechanism (Fig. 6) [108]. The model proposes that, after cleavage, two or three nucleotides from the parental overlap sequences are melted and then swapped between the partner duplexes. In this mechanism, sequence homology is required in the reannealing reaction that orients the 5′-OH end of the invading strand for ligation. Movement of the Holliday junction is limited to the 1–3 central bp of the overlap region. This movement would simply be unstacking-restacking events whereby the Holliday junction DNA helices are reorganised from a ‘parental’ (top strands crossed) configuration to a ‘recombinant’ (bottom strands crossed) configuration in which the bottom strand cleavage sites are adequately positioned for strand exchange (Fig. 6) [108–110]. This view is supported by recent work showing that one Holliday junction isomer preferentially undergoes top strand exchange, whereas resolution of the other iso-form predominantly occurs by bottom strand exchange [111, 112]. Also, in the recombinase-Holliday junction complex, the centre of the overlap region is partially unstacked [113].

Although variations of both models can be envisioned, the strand swapping mechanism seems more suitable for other recombination systems of the λ Int family, such as conjugative transposons and integrons, which are less strict in their requirement for sequence homology between partner recombination sites [23, 43]. These systems may be less sensitive to DNA mispairing by catalysing both strand exchanges without annealing recombinant strands. Alternatively, the Holliday junction iso-form generated by a single strand exchange may be a stable recombination product that is not processed back by the recombinases but could be well resolved by some other host function (as discussed in Ref. [113]).

Another matter of divergence within the λ Int family concerns the catalytic role of the different recombinase subunits within the synaptic complex. A wealth of data indicate that the yeast recombinase Flp uses a trans cleavage mechanism in which the catalytic RHR triad of one monomer activates the scissile phosphodiester adjacent to its binding site, while the tyrosine nucleophile is contributed by a different Flp monomer bound on a different half-site. As for MuA, this active site sharing observed for Flp was proposed as a mechanism for coupling catalysis and recombination site pairing [31, 114]. However, this view was weakened by experiments showing that the two collaborating monomers are bound on the same core site and that cleavage may actually precede synapsis [115, 116]. An alternative model based on an asymmetrical (head to tail) assembly of Flp monomers has been proposed, again suggesting that active site-dependent molecular bridges may be required [117]. In contrast, all current data indicate that XerC and XerD recombinases act on the closest cleavage site by providing all catalytic residues in cis[118]. For λ Int, experiments supporting both cis and trans mechanisms have been reported [119, 120]. These apparently conflicting results have raised a number questions that have been debated in the recent literature [48, 121–123].

Although the details of the molecular interactions between recombinases subunits are not known, some new insight has been provided by the recent acquisition of structural data for the C-terminal catalytic moieties of λ Int and the Haemophilus phage 1 integrase (HP1 Int) and for the complete 298 aa XerD recombinase [124–126]. The C-terminal catalytic domain of these three recombinases exhibits a similar fold where the conserved residues RHR are exposed in a basic groove likely to contact the DNA for the activation of the scissile phosphodiester bond. However, the structure of the region containing the active tyrosine (the extreme C-terminal part of the domain) is very different in the three structures. This region also seems to be involved in recombinase inter-subunit interactions. In λ Int, the tyrosine is positioned on a disordered flexible loop in a configuration which, by allowing conformational changes, could be consistent with either a cis or trans cleavage [124]. In contrast, the structures of both HP1 Int and XerD support a cis mechanism, the tyrosine being docked at a fixed position within the catalytic pocket. In the HP1 Int catalytic domain, which crystallised as a dimer, the tyrosine is in a position ready for in line nucleophilic attack of the scissile phosphate [125], whereas in the model proposed for XerD-DNA complex, some local alteration of the structure would be required to bring the residue in an active configuration [126]. As for γδ resolvase, it is thought that this conformational activation is induced by protein interactions between XerD and the partner recombinase XerC (our unpublished results). Another major conformational change would be required to move the globular N-terminal domain of XerD (absent from the λ and HP1 Int structures) that blocks the access to the active site. The fact that the peptide connecting the two XerD domains is disordered in the crystal is again indicative of structural flexibility [126].

4 Higher order nucleoprotein complexes: an additional level of control

The strategies used by transposable elements and site-specific recombination systems to assemble their recombination machinery share many similarities. In particular, for the majority of elements, the minimal number of recombinase molecules bound to their binding sites on DNA dictated by the chemistry of the reaction is usually insufficient to mediate a ‘normal’ recombination reaction at a ‘normal’ frequency. These systems contain additional accessory DNA sequences, to which further recombinase subunits and/or other proteins bind thereby forming part of the functional nucleoprotein complex. In many cases, DNA supercoiling is also required for the assembly and/or activity of the recombination complex. The precise architecture of the complex provides a mechanism for controlling the outcome of a recombination reaction, while a requirement for cellular proteins may provide a means of relating the frequency of a recombination reaction to host metabolism. Thus, the requirement for higher order interactions represents an additional level of regulation ensuring that recombination only occurs at the correct time, between the correct sites [5, 9].

For example, bacteriophage λ integration and excision need to be irreversible and separate steps of phage development to permit a committed decision between lysogeny and lytic growth [42]. Both recombination reactions are mediated by the formation of nucleoprotein structures that arrange the recombination partners in a complementary configuration for synapsis [4, 42, 127]. In addition to the core site DNA binding activity, λ Int has a second DNA binding domain that recognises accessory ‘arm sites’ in the flanking regions of the recombination loci. This enables a single recombinase protomer to form a molecular bridge between adjacent sites on the same DNA molecules or between the two recombining duplexes in the synaptic complex. For both integration and excision, Int subunits bound to the arms of one partner are delivered in trans to the core site of the other partner.

However, λ Int-mediated integration and excision employ distinct subsets of proteins acting on different DNA binding sites and, therefore, the two reactions are not the simple reverse of each other. For integration, the 240-bp phage recombination locus attP is wrapped in an ‘intasome’ complex in which the relevant Int subunits can capture the naked 25-bp core region of the bacterial attachment site, attB[127, 128]. Formation of the intasome involves precisely positioned DNA bends induced by the host protein IHF and a set of Int-mediated bridges between the core site and specific arm binding sites in the flanking DNA segments of attP. Excision utilises two other proteins in addition to Int and IHF, namely the phage-encoded Xis and the host factor FIS, as well as a different set of arm site/core sites interactions, to assemble a synaptic complex between the ends of the integrated prophage attL and attR. In this complex, the four Int subunits are now contributed by both partners (three from attL and one from attR) [129]. Thus, λ integration and excision use distinct and mutually exclusive ‘lock and key’ interactions which allow the phage to become committed to either lysogeny or lytic growth [5, 42]. Furthermore, the higher order interactions responsible for the assembly of the integrative and excisive synaptic structures also influence the resolution of the Holliday junction intermediates, thereby reinforcing the directionality of both recombination reactions [130].

During λ integration and excision, DNA supercoiling acts as an additional architectural element facilitating directional wrapping of the DNA around the complex, but recombination then occurs by random collision between sites in various configurations. In contrast, it is crucial for the in vivo function of site-specific resolution and DNA inversion systems to recombine selectively sites which are in direct or inverted repeat on the same DNA molecule, respectively. For these systems, such selectivity is achieved through a recombinational synapse of a precise local geometry [80].

The 120-bp resolvase recombination site res contains three subsites, each of which binds two resolvase subunits. In the recombination complex, strand exchange occurs at subsite I, while the eight resolvase molecules bound to the other two subsites of the two synapsed res sites form a protein scaffold around which three plectonemic DNA supercoils are interwrapped (Fig. 7a) [36]. Such a −3 synapse, in which three negative supercoils are trapped, can only form readily on directly repeated sites on the same DNA molecule, thereby limiting recombination to intramolecular resolution. To form a −3 synapse between inverted repeated sites, or between sites on separate circular molecules, unfavourable compensating positive supercoils would have to be introduced elsewhere. This system therefore constitutes a so-called ‘topological filter’ [80]. A similar topological filter is used in the inversion reaction catalysed by the DNA invertases like Gin and Hin. In this case however, two negative supercoils are trapped in the three-branched ‘invertasome’ complex in which the two inverted repeated recombination sites bound by the recombinase are aligned on a third and distant element, the recombinational enhancer, which is bound by the host factor FIS (Fig. 7b) [32, 80, 85].

7

Models for the assembly of a synaptic complex for different specialised recombination systems showing topological selectivity. a: Synaptosome in resolvase/res system. b: Invertasome of DNA inversion systems. c: Synaptic complex assembled by Xer recombination system at the ColE1 plasmid resolution site cer. d: Possible structure of the LER synaptic complex intermediate during bacteriophage Mu transpososome formation. Grey arrows represent the directly repeated (DR) or inverted repeated (IR) recombination core sites, or the bacteriophage Mu left (L) and right (R) ends. Open ribbons are the res site accessory subsites II and III, the cer site accessory sequences (AS), or the recombinational and transpositional enhancers (E). In the Mu LER complex, the thick and thin lines represent the phage genome and the donor DNA molecule, respectively. The wavy line is the target DNA. Only the MuA subunits that will form the active tetramer after conversion of the LER complex are represented. Additional molecules of MuA and of the proteins HU and IHF which participate in the formation of the LER complex are not shown.

The FIS-bound enhancer is required for both for the synaptic complex assembly and the activation of the recombinase subunits. The latter may be achieved by inducing a conformational change at the dimer interface [92, 103, 131]. Likewise, in both resolvase and invertase-mediated recombination, DNA supercoiling energy is used after initial synapsis to provide the driving force for the reaction and to determine the right-handed rotational direction of strand exchange [90–92, 132]. As mentioned above, single amino acid changes in resolvase and invertase enzymes allow the requirement for supercoiling and higher order nucleoprotein interactions to be overcome, underlining the regulatory function of these interactions ([92, 100, 101, 103]; M.R. Boockock and N.D.F. Grindley, personal communication). By promoting deletions and inversions as well as fusion products, the behaviour of these relaxed recombinase variants is similar to that of recombinases such as Cre and Flp which mediate recombination with little selectivity for a particular site organisation and with no requirement for a specific DNA topology [4, 106].

The Xer recombination system provides an intriguing example of functional flexibility. Recombination at the E. coli chromosome site dif which can occur intermolecularly and intramolecularly appears to be achieved solely by the action of the XerC and XerD recombinases on a 28-bp core recombination site. In contrast, Xer recombination at sites present in natural multicopy plasmids (e.g., cer in ColE1 plasmid) is preferentially intramolecular and has the requirement for ∼200 bp of accessory DNA sequence and for additional host proteins (ArgR, the repressor of the arginine regulon; and PepA, the aminopeptidase A) [40]. These are involved in forming a synapse of defined geometry in which, as in the resolvase synaptosome, three negative supercoils are trapped (Fig. 7c) [133]. It is likely that functionally equivalent topological filters have evolved independently in a number of different recombination systems by multiplying the recombinase binding sites within the recombination locus (as in the resolvase system) and/or by recruiting different host DNA binding proteins (as in Xer recombination or in Gin and Hin inversion systems).

In transposition, the higher order interactions controlling the assembly of the active complex, or ‘transpososome’, not only dictate the configuration with which the transposon ends can interact, but they also influence target DNA site selection. Multiple checkpoints in the assembly and progression of bacteriophage Mu transpososome ensure that each cycle of replicative transposition will be successfully executed and that the integrity of the phage genome will be maintained (reviewed in Refs. [6, 15, 66, 67]). Like λ Int, MuA has two distinct DNA binding domains enabling a single transposase subunit to interact with specific sequence elements at the ends of Mu, as well as with a transposition enhancer located within the Mu genome, about 1kb from the left end. At an early stage of the transposition reaction, a complex circuit of interactions between MuA subunits bound to the left end (L), the enhancer (E), and the right end (R) promotes the formation of a transient three-site synaptic complex called ‘LER’ (Fig. 7d) [134]. Assembly of this complex requires DNA supercoiling as well as precisely-positioned DNA bends induced by the host proteins HU and IHF. Upon assembly, the catalytically inert components of the LER complex undergo a conformational transition converting the inactive pre-transpososome into the stable synaptic complex in which the two Mu ends and the active sites of the MuA tetramer are engaged for catalysis.

As for other recombination systems, the geometry of the LER complex may act as a topological filter restricting synapsis to two adjacent inverted ends [135]. DNA supercoiling may also provide the free energy required for the DNA and/or protein conformational changes likely to occur during conversion of the LER complex [136]. In transpososome assembly, the enhancer acts as a platform from which MuA monomers are delivered to form the active tetramer at the Mu ends [68, 137, 138]. These MuA-promoted bridges between the enhancer and the two ends are reminiscent of those formed by λ Int for the synapse of two recombination sites.

A similar role as a scaffold has been proposed for the Mu transposition cofactor protein, MuB [137]. MuB is an allosteric activator of MuA that is also involved in target capture and in the mechanism of target immunity whereby self-integration by Mu is prevented. Such immunity is conferred by a set of interactions between MuA and MuB that precludes the formation of a MuB-DNA complex in the vicinity of a transposase-bound end. Although association of a MuB-bound target DNA to the transpososome is not absolutely required to initiate transposition, its presence strongly stimulates both the end cleavage and strand transfer reactions [6, 15]. This ensures that Mu transposition will only occur in an appropriate target, thus avoiding self-destruction.

The coordination of the transposition reactions seems even tighter in the case of Tn_7_. In addition to the transposase TnsA+B core, three other proteins (TnsC, D and E) participate in Tn_7_ transposition [11]. The use of either TnsD or TnsE in the complex directs the transposon to different types of target sites. TnsD is responsible for the recognition of the specific chromosome integration site attTn7, whereas TnsE allows Tn_7_ to transpose with little target specificity into different sites. TnsC mediates communication between the transposon ends and the target DNA complex, and has an immunity function analogous to that of MuB. In the TnsD-dependent pathway, no cleavage is observed in the absence of the target DNA, suggesting that all participants of the reaction, i.e., TnsA, B, C and D, the transposon ends and the target site attTn7, must be assembled in the transpososome to initiate recombination [22, 139].

By contrast, in IS_10_ transposition, the double-strand breaks severing the element from the donor locus must occur before the transpososome can interact with the target DNA [76]. IS_10_ transposase is also able to use transposon ends located on distinct molecules or in directly repeated configuration rather than the canonical inverted orientation, to carry-out transposition [140]. Finally, when compared to Mu and Tn_7_, IS_10_ is also more promiscuous with respect to target site selection. Intra-transposon insertion events are observed and may be favoured through a regulation pathway involving the host factor IHF ([141]; R. Chalmers and N. Kleckner, personal communication). This relative flexibility of IS_10_ transposition accounts for how IS_10_-promoted adjacent deletions and replicon fusions occur, as well as how new transposable element based on IS_10_ can form [50].

5 Conclusion

Recent advances in our understanding of transposition and site-specific recombination have provided new pictures of specialised recombination machines. Some of these snapshots must now be viewed with stereoglasses! The emerging theme is that recombinases are built by combining simple and conserved chemical mechanisms of DNA cut and paste with specific DNA binding activities and the ability to form a complex in which the DNA substrates are brought together in order to trigger the recombination reactions. Adding more requirements to the system allows the outcome of recombination to be controlled at a higher level to prevent undesired DNA rearrangements.

Although the flexibility of the conservative recombination mechanism used by site-specific recombinases to achieve different outcomes is obvious, the mechanism of one-step transesterification used by the DDE transposases seems restricted to the movement of transposable elements. Nevertheless, the DNA double-strand break reactions that initiate V(D)J recombination, which in vertebrates serves to assemble the immunoglobulin and T-cell receptor genes, are chemical equivalents of those catalysed by the DDE recombinases [142]. Further characterisation of the bacterial systems that are not related to those reviewed here, such as IS_1_, IS_91_ and the Piv family of recombinases, will undoubtedly provide new examples on how different DNA breakage and joining mechanisms are adapted to promote specific DNA rearrangements [25, 26, 28].

For the different systems, the next challenge is to obtain a dynamic view describing how the recombination partners come together and then, how the structure of the recombination complex changes during the different reaction steps. These studies are not only crucial for the understanding of fundamental recombination mechanisms, but should also assist in the improvement or development of new tools based on transposition and site-specific recombination systems for use in a variety of in vivo and in vitro applications.

Acknowledgements

We are grateful to Sean Colloms, Francois Cornet and Finbarr Hayes for critical reading of the manuscript and for helpful comments. Work in D.J.S.’s laboratory was supported by grants from the Wellcome Trust and the Medical Research Council. B.H. was the recipient of a postdoctoral fellowship from the European Molecular Biology Organization and from the European Communities BIOTECH program.

References

[1]

(

1996

)

Exchanging partners: recombination in E. coli

.

Trends Genet.

12

,

20

–

26

.

[2]

(

1993

)

Site-specific genetic recombination: hops, flips, and flops

.

FASEB J.

7

,

760

–

767

.

[3]

(

1992

)

Catalysis by site-specific recombinases

.

Trends Genet.

8

,

432

–

439

.

[4]

(

1993

)

Mechanistic and structural complexity in the site-specific recombination pathways of Int and FLP

.

Curr. Opin. Genet. Dev.

3

,

699

–

707

.

[5]

(

1996

)

Site-specific recombination: integration, excision, resolution, and inversion of defined DNA segments

. In:

Escherichia coli and Salmonella typhimurium: Cellular and Molecular Biology

( et al., Eds.), 2nd edn., Vol.

2

, pp.

2363

–

2376

.

American Society for Microbiology

,

Washington, DC

.

[6]

(

1992

)

Transpositional recombination: mechanistic insights from studies of Mu and other elements

.

Annu. Rev. Biochem.

61

,

1011

–

1051

.

[7]

(

1995

)

Bacterial transposases and retroviral integrases

.

Mol. Microbiol.

15

,

13

–

23

.

[8]

(

1995

)

Mechanisms of DNA transposition

. In:

Mobile Genetic Elements

(, Ed.), pp.

18

–

37

.

IRL

,

Oxford

.

[9]

(

1996

)

Transposition

. In:

Escherichia coli and Salmonella typhimurium: Cellular and Molecular Biology

( et al., Eds.), 2nd edn., Vol.

2

, pp.

2339

–

2362

.

American Society for Microbiology

,

Washington, DC

.

[10]

(

1997

)

Target site selection in transposition

.

Annu. Rev. Biochem.

66

,

437

–

474

.

[11]

(

1996

)

Transposon Tn_7_

.

Curr. Top. Microbiol. Immunol.

204

,

27

–

48

.

[12]

(

1991

)

Genetic evidence against intramolecular rejoining of the donor DNA molecule following IS_10_ transposition

.

Genetics

128

,

687

–

694

.

[13]

and Craig, N.L. (

1993

)

Tn_7_ transposition creates a hotspot for homologous recombination at the transposon donor site

.

Genetics

133

,

9

–

16

.

[14]

(

1996

)

Bacterial insertion sequences

.

Curr. Top. Microbiol. Immunol.

204

,

1

–

26

.

[15]

(

1996

)

Transposition of phage Mu DNA

.

Curr. Top. Microbiol. Immunol.

204

,

83

–

102

.

[16]

(

1990

)

Functional similarities between retroviruses and the IS_3_ family of bacterial insertion sequences? Mol

.

Microbiol.

4

,

1771

–

1777

.

[17]

(

1991

)

Retroviral integrase domains: DNA binding and the recognition of LTR sequences

.

Nucleic Acids Res.

19

,

851

–

860

.

[18]

(

1992

)

Residues critical for retroviral integrative recombination in a region that is highly conserved among retroviral/retrotransposon integrases and bacterial insertion sequence transposases

.

Mol. Cell. Biol.

12

,

2331

–

2338

.

[19]

(

1993

)

The IS_4_ family of insertion sequences: evidence for a conserved transposase motif

.

Mol. Microbiol.

9

,

1283

–

1295

.

[20]

(

1994

)

A proposed superfamily of transposase genes: transposon-like elements in ciliated protozoa and a common ‘D35E’ motif

.

Proc. Natl. Acad. Sci. USA

91

,

942

–

946

.

[21]

(

1994

)

Identification of residues in the Mu transposase essential for catalysis

.

Proc. Natl. Acad. Sci. USA

91

,

6654

–

6658

.

[22]

(

1996

)

The Tn_7_ transposase is a heteromeric complex in which DNA breakage and joining activities are distributed between different gene products

.

EMBO J.

15

,

6348

–

6361

.

[23]

(

1995

)

Conjugative transposition

.

Annu. Rev. Microbiol.

49

,

367

–

397

.

[24]

(

1995

)

Molecular genetics of the chloramphenicol-resistance transposon Tn_4451_ from Clostridium perfringens: the TnpX site-specific recombinase excises a circular transposon molecule

.

Mol. Microbiol.

16

,

535

–

551

.

[25]

(

1994

)

Amino acid sequence homology between Piv, an essential protein in site-specific DNA inversion in Moraxella lacunata, and transposases of an unusual family of insertion elements

.

J. Bacteriol.

176

,

4160

–

4164

.

[26]

(

1995

)

Mutagenesis of the IS_1_ transposase: importance of a His-Arg-Tyr triad for activity

.

J. Bacteriol.

177

,

5070

–

5077

.

[27]

(

1992

)

IS_91_ transposase is related to the rolling-circle-type replication proteins of the pUB110 family of plasmids

.

Nucleic Acids Res

.

20

,

3521

.

[28]

(

1994

)

Differential roles of the transposon termini in IS91 transposition

.

Proc. Natl. Acad. Sci. USA

91

,

1922

–

1926

.

[29]

(

1995

)

Transposable elements as selfish DNA

. In:

Mobile Genetic Elements

(, Ed.),

IRL

,

Oxford

.

[30]

(

1993

)

Natural genetic engineering of the bacterial genome

.

Curr. Opin. Genet. Dev.

3

,

845

–

848

.

[31]

(

1994

)

Phosphoryl transfer in Flp recombination: a template for strand transfer mechanisms

.

Trends Biochem. Sci.

19

,

78

–

82

.

[32]

(

1992

)

DNA inversions in phages and bacteria

.

Trends Genet.

8

,

457

–

462

.

[33]

(

1994

)

DNA rearrangement of the shufflon determines recipient specificity in liquid mating of IncI1 plasmid R64

.

J. Mol. Biol.

243

,

6

–

9

.

[34]

(

1990

)

The cisA cistron of Bacillus subtilis sporulation gene spoIVC encodes a protein homologous to a site-specific recombinase

.

J. Bacteriol.

172

,

1092

–

1098

.

[35]

(

1995

)

Programmed DNA rearrangement of a cyanobacterial hupL gene in heterocysts

.

Proc. Natl. Acad. Sci. USA

92

,

791

–

795

.

[36]

(

1994

)

Resolvase-mediated site-specific recombination

. In:

Nucleic Acids and Molecular Biology

( , Eds.), pp.

236

–

267

.

Springer-Verlag

,

Berlin

.

[37]

(

1981

)

A novel role for site-specific recombination in maintenance of bacterial replicons

.

Cell

25

,

729

–

736

.

[38]

(

1986

)

D protein of miniF plasmid acts as a repressor of transcription and as a site-specific resolvase

.

Nucleic Acids Res.

14

,

9713

–

9728

.

[39]

(

1996

)

Site-specific recombination in gram-positive theta-replicating plasmids

.

FEMS Microbiol. Lett.

142

,

1

–

10

.

[40]

(

1995

)

Site-specific recombination and circular chromosome segregation

.

Phil. Trans. R. Soc. Lond. B. Biol. Sci.

347

,

37

–

42

.

[41]

(

1992

)

Chromosome partitioning in Escherichia coli

.

J. Bacteriol.

174

,

7883

–

7889

.

[42]

(

1989

)

Dynamic, structural, and regulatory aspects of lambda site-specific recombination

.

Annu. Rev. Biochem.

58

,

913

–

949

.

[43]

(

1995

)

Mobile gene cassettes and integrons: capture and spread of genes by site-specific recombination

.

Mol. Microbiol.

15

,

593

–

600

.

[44]

(

1988

)

Resolvases and DNA-invertases: a family of enzymes active in site-specific recombination

. In:

Genetic Recombination

( , Eds.), pp.

357

–

396

.

American Society for Microbiology

,

Washington, DC

.

[45]

(

1995

)

The tyrosine-6 hydroxyl of gamma delta resolvase is not required for the DNA cleavage and rejoining reactions

.

Mol. Microbiol.

15

,

865

–

870

.

[46]

(

1986

)

The integrase family of site-specific recombinases: regional similarities and global diversity

.

EMBO J.

5

,

433

–

440

.

[47]

(

1992

)

Evidence for a second conserved arginine residue in the integrase family of recombination proteins

.

Protein Eng.

5

,

87

–

91

.

[48]

(

1996

)

Cis and trans in site-specific recombination

.

Mol. Microbiol.

20

,

233

–

238

.

[49]

(

1992

)

Polynucleotidyl transfer reactions in transpositional DNA recombination

.

J. Biol. Chem.

267

,

21273

–

21276

.

[50]

(

1996

)

Tn_10_ and IS_10_ transposition and chromosome rearrangements: mechanism and regulation in vivo and in vitro

.

Curr. Top. Microbiol. Immunol.

204

,

49

–

82

.

[51]

(

1992

)

Excision of Tn_10_ from the donor site during transposition occurs by flush double-strand cleavages at the transposon termini

.

Proc. Natl. Acad. Sci. USA

89

,

4648

–

4652

.

[52]

(

1991

)

Tn_7_ transposition in vitro proceeds through an excised transposon intermediate generated by staggered breaks in DNA

.

Cell

65

,

805

–

816

.

[53]

(

1995

)

An in vivo transposase-catalyzed single-stranded DNA circularization reaction

.

Genes Dev.

9

,

2846

–

2858

.

[54]

(

1996

)

IS_911_-mediated transpositional recombination in vitro

.

J. Mol. Biol.

264

,

68

–

81

.

[55]

(

1996

)

Identification and characterization of the linear IS_3_ molecules generated by staggered breaks

.

J. Biol. Chem.

271

,

197

–

202

.

[56]

(

1995

)

IS_1_-mediated intramolecular rearrangements: formation of excised transposon circles and replicative deletions

.

EMBO J.

14

,

5410

–

5421

.

[57]

(

1995

)

Unity in transposition reactions

.

Science

270

,

253

–

254

.

[58]

(

1991

)

Inversion of the phosphate chirality at the target site of Mu DNA strand transfer: evidence for a one-step transesterification mechanism

.

Cell

66

,

129

–

140

.

[59]

(

1991

)

HIV-1 DNA integration: mechanism of viral DNA cleavage and DNA strand transfer

.

Cell

67

,

1211

–

1221

.

[60]

(

1995

)

Structure of the bacteriophage Mu transposase core: a common structural motif for DNA transposition and retroviral integration

.

Cell

82

,

209

–

220

.

[61]

(

1995

)

Recombining the structures of HIV integrase, RuvC and RNase H

.

Structure

3

,

131

–

134

.

[62]

(

1995

)

DNA transposition: from a black box to a color monitor

.

Cell

83

,

1063

–

1066

.

[63]

(

1996

)

Retroviral integrases and their cousins

.

Curr. Opin. Struct. Biol.

6

,

76

–

83

.

[64]

(

1991

)

Crystal structure of the ribonuclease H domain of HIV-1 reverse transcriptase

.

Science

252

,

88

–

95

.

[65]

(

1991

)

Structural basis for the 3′-5′ exonuclease activity of Escherichia coli DNA polymerase I: a two metal ion mechanism

.

EMBO J.

10

,

25

–

33

.

[66]

(

1996

)

Quality control in Mu DNA transposition

.

Cell

85

,

137

–

140

.

[67]

(

1996

)

DNA transposition: jumping gene machine, some assembly required

.

Curr. Biol.

6

,

817

–

820

.

[68]

(

1995

)

A domain sharing model for active site assembly within the Mu A tetramer during transposition: the enhancer may specify domain contributions

.

EMBO J.

14

,

2374

–

2384

.

[69]

(

1995

)

A novel DNA binding and nuclease activity in domain III of Mu transposase: evidence for a catalytic region involved in donor cleavage

.

EMBO J.

14

,

3835

–

3843

.

[70]

(

1996

)

The interwoven architecture of the Mu transposase couples DNA synapsis to catalysis

.

Cell

85

,

257

–

269

.

[71]

(

1996

)

Positional information within the Mu transposase tetramer: catalytic contributions of individual monomers

.

Cell

85

,

447

–

455

.

[72]

(

1996

)

Mu transpositional recombination: donor DNA cleavage and strand transfer in trans by the Mu transposase

.

Cell

85

,

271

–

280

.

[73]

(

1994

)

Tn_10_ transposition in vivo: temporal separation of cleavages at the two transposon ends and roles of terminal basepairs subsequent to interaction of ends

.

EMBO J.

13

,

3401

–

3411

.

[74]

(

1995

)

Identification and characterization of a pre-cleavage synaptic complex that is an early intermediate in Tn_10_ transposition

.

EMBO J.

14

,

4374

–

4383

.

[75]

(

1995

)

The two single-strand cleavages at each end of Tn_10_ occur in a specific order during transposition

.

Proc. Natl. Acad. Sci. USA

92

,

7814

–

7818

.

[76]

(

1997

)

The Tn_10_ synaptic complex can capture a target DNA only after transposon excision

.

Cell

89

,

205

–

214

.

[77]

(

1996

)

The three chemical steps of Tn_10_/IS_10_ transposition involve repeated utilization of a single active site

.

Cell

84

,

223

–

233

.

[78]

(

1996

)

Switching from cut-and-paste to replicative Tn_7_ transposition

.

Science

272

,

401

–

404

.

[79]

(

1996

)

DNA topoisomerases

.

Annu. Rev. Biochem.

65

,

635

–

692

.

[80]

(

1995

)

Topological selectivity in site-specific recombination

. In:

Mobile Genetic Elements

(, Ed.), pp.

101

–

129

.

IRL

,

Oxford

.

[81]

(

1995

)

Site-specific recombinases an their interaction with DNA

. In:

DNA-Protein: Structural Interactions

(, Ed.), pp.

141

–

176

.

IRL

,

Oxford

.

[82]

(

1996

)

Interaction of FimB and FimE with the fim switch that controls the phase variation of type 1 fimbriae in Escherichia coli K-12

.

Mol. Microbiol.

21

,

725

–

738

.

[83]

(

1993

)

Two related recombinases are required for site-specific recombination at dif and cer in E. coli K12

.

Cell

75

,

351

–

361

.

[84]

(

1996

)

Xer-mediated site-specific recombination in vitro

.

EMBO J.

15

,

1172

–

1181

.

[85]

(

1991

)

Mechanism of site-specific DNA inversion in bacteria

.

Curr. Opin. Genet. Dev.

1

,

404

–

411

.

[86]

(

1984

)

Resolvase-mediated recombination intermediates contain a serine residue covalently linked to the DNA

.

Cold Spring Harbor Symp. Quant. Biol.

49

,

245

–

249

.

[87]

(

1988

)

The DNA invertase Gin of phage Mu: formation of a covalent complex with DNA via a phosphoserine at amino acid position 9

.

EMBO J.

7

,

1229

–

1237

.

[88]

(

1995

)

Catalytic residues of gamma delta resolvase act in cis

.

EMBO J.

14

,

5129

–

5140

.

[89]

(

1997

)

Tn_3_ resolvase catalyses multiple recombination events without intermediate rejoining of DNA ends

.

J. Mol. Biol.

266

,

108

–

121

.

[90]

(

1988

)

Gin-mediated DNA inversion: product structure and the mechanism of strand exchange

.

Proc. Natl. Acad. Sci. USA

85

,

752

–

756

.

[91]

(

1989

)

Site-specific recombination by Tn_3_ resolvase: topological changes in the forward and reverse reactions

.

Cell

58

,

779

–

790

.

[92]

(

1993

)

Analysis of strand exchange and DNA binding of enhancer-independent Gin recombinase mutants

.

EMBO J.

12

,

1047

–

1057

.

[93]

(

1991

)

Alignment of recombination sites in Hin-mediated site-specific DNA recombination

.

Genes Dev.

5

,

1635

–

1645

.

[94]

(

1994

)

Stereoselectivity of DNA catenane fusion by resolvase

.

Nature

368

,

76

–

78

.

[95]

(

1995

)

Crystal structure of the site-specific recombinase gamma delta resolvase complexed with a 34 bp cleavage site

.

Cell

82

,

193

–

207

.

[96]

(

1994

)

Hin recombinase bound to DNA: the origin of specificity in major and minor groove interactions

.

Science

263

,

348

–

355

.

[97]

(

1993

)

Protein-protein interactions directing resolvase site-specific recombination: a structure-function analysis

.

EMBO J.

12

,

1447

–

1458

.

[98]

(

1994

)

Refinement of gamma delta resolvase reveals a strikingly flexible molecule

.

Structure

2

,

371

–

384

.

[99]

(

1994

)

Model for a DNA-mediated synaptic complex suggested by crystal packing of gamma delta resolvase subunits

.

EMBO J.

13

,

1514

–

1524

.

[100]

(

1988

)

Enhancer-independent mutants of the Cin recombinase have a relaxed topological specificity

.

EMBO J.

7

,

3991

–

3996

.

[101]

(

1988

)

Isolation and characterization of unusual Gin mutants

.

EMBO J.

7

,

3983

–

3989

.

[102]

(

1994

)

Processive recombination by wild-type Gin and an enhancer-independent mutant

.

Insight into the mechanisms of recombination selectivity and strand exchange. J. Mol. Biol.

243

,

437

–

457

.

[103]

(

1996

)

The Hin dimer interface is critical for Fis-mediated activation of the catalytic steps of site-specific DNA inversion

.

Curr. Biol.

6

,

163

–

177

.

[104]

(

1996

)

Site-specific recombination by Tn_3_ resolvase, photocrosslinked to its supercoiled DNA substrate

.

J. Mol. Biol.

260

,

299

–

303

.

[105]

(

1994

)

The linkage change of a knotting reaction catalysed by Tn_3_ resolvase

.

J. Mol. Biol.

239

,

25

–

36

.

[106]

(

1995

)

The Flp recombinase of the 2-microns plasmid of Saccharomyces cerevisiae

.

Prog. Nucleic Acid Res. Mol. Biol.

51

,

53

–

91

.

[107]

(

1995

)

Suicide substrates reveal properties of the homology-dependent steps during integrative recombination of bacteriophage lambda

.

Curr. Biol.

5

,

1312

–

1321

.

[108]

(

1995

)

Swapping DNA strands and sensing homology without branch migration in lambda site-specific recombination

.

Curr. Biol.

5

,

139

–

148

.

[109]

(

1994

)

Resolution of immobile chi structures by the FLP recombinase of 2 microns plasmid

.

J. Mol. Biol.

243

,

199

–

207

.

[110]

(

1995

)

Effects of Holliday junction position on Xer-mediated recombination in vitro

.

EMBO J.

14

,

2651

–

2660

.

[111]

(

1997

)

The isomeric preference of holliday junctions influence resolution bias by lambda integrase and also shows a predictable dependence on the branch point sequence

.

EMBO J.

16

,

3744

–

3755

.

[112]

(

1997

)

Action of site-specific recombinases XerC and XerD on tethered holliday junctions

.

EMBO J.

16

,

3731

–

3743

.

[113]

(

1994

)

Xer-mediated site-specific recombination at cer generates Holliday junctions in vivo

.

EMBO J.

13

,

1844

–

1855

.

[114]

(

1992

)

DNA cleavage in trans by the active site tyrosine during Flp recombination: switching protein partners before exchanging strands

.

Cell

69

,

647

–

658

.

[115]

(

1994

)

Directed protein replacement in recombination full sites reveals trans-horizontal DNA cleavage by Flp recombinase

.

EMBO J.

13

,

5346

–

5354

.

[116]

(

1996

)

Analyses of the first chemical step in Flp site-specific recombination: Synapsis may not be a pre-requisite for strand cleavage

.

J. Mol. Biol.

256

,

720

–

735

.

[117]

(

1995

)

Asymmetry in active complexes of FLP recombinase

.

Genes Dev.

9

,

2053

–

2064

.

[118]

(

1997

)

Binding and cleavage of nicked substrates by site-specific recombinases XerC and XerD

.

J. Mol. Biol.

265

,

30

–

39

.

[119]

(

1993

)

Complementation of bacteriophage lambda integrase mutants: evidence for an intersubunit active site

.

EMBO J.

12

,

4577

–

4584

.

[120]

(

1994

)

Lambda integrase cleaves DNA in cis

.

EMBO J.

13

,

4421

–

4430

.

[121]

(

1995

)

Gatecrashers at the catalytic party

.

Trends Genet.

11

,

121

–

123

.

[122]

(

1995

)

Return to sobriety after the catalytic party

.

Trends Genet.

11

,

432

–

433

.

[123]

(

1997

)

The Cis-trans paradox of integrase

.

Science

276

,

49

–

50

.

[124]

(

1997

)

Flexibility in DNA recombination: structure of the lambda integrase catalytic core

.

Science

276

,

126

–

131

.

[125]

(

1997

)

Molecular organization in site-specific recombination: the catalytic domain of bacteriophage HP1 integrase at 2.7 Å resolution

.

Cell

89

,

205

–

214

.

[126]

Crystal structure of the site-specific recombinase

, XerD. Submitted.

[127]

(

1990

)

Bending and supercoiling of DNA at the attachment site of bacteriophage lambda

.

Trends Biochem. Sci.

15

,

222

–

227

.

[128]

(

1988

)

Synapsis of attachment sites during lambda integrative recombination involves capture of a naked DNA by a protein-DNA complex

.

Cell

52

,

9

–

17

.

[129]

(

1992

)

Lambda Int protein bridges between higher order complexes at two distant chromosomal loci attL and attR

.

Science

256

,

198

–

203

.

[130]

(

1995

)

The Holliday junction intermediates of lambda integrative and excisive recombination respond differently to the bending proteins integration host factor and excisionase

.

EMBO J.

14

,

397

–

406

.

[131]

(

1994

)

Analysis of subunit interaction by introducing disulfide bonds at the dimerization domain of Hin recombinase

.

J. Biol. Chem.

269

,

31134

–

31142

.

[132]

(

1996

)

Contributions of supercoiling to Tn_3_ resolvase and phage Mu Gin site-specific recombination

.

J. Mol. Biol.

256

,

50

–

65

.

[133]

(

1997

)

Topological selectivity in Xer site-specific recombination

.

Cell

88

,

855

–

864

.

[134]

(

1996

)

Three-site synapsis during Mu DNA transposition: a critical intermediate preceding engagement of the active site

.

Cell

85

,

435

–

445

.

[135]

(

1986

)

Role of DNA topology in Mu transposition: mechanism of sensing the relative orientation of two DNA segments

.

Cell

45

,

793

–

800

.

[136]

(

1994

)

Crucial role for DNA supercoiling in Mu transposition: a kinetic study

.

Proc. Natl. Acad. Sci. USA

91

,

699

–

703

.

[137]

(

1995

)

Assembly of phage Mu transpososomes: cooperative transitions assisted by protein and DNA scaffolds

.

Cell

83

,

375

–

385

.

[138]

(

1995

)

Enhancer-independent variants of phage Mu transposase: enhancer-specific stimulation of catalytic activity by a partner transposase

.

Genes Dev.

9

,

2545

–

2555

.

[139]

(

1993

)

Tn_7_ transposition: target DNA recognition is mediated by multiple Tn_7_-encoded proteins in a purified in vitro system

.

Cell

72

,

931

–

943

.

[140]

(

1996

)

IS_10_/Tn_10_ transposition efficiently accommodates diverse transposon end configurations

.

EMBO J.

15

,

5112

–

5122

.

[141]

(

1995

)

Negative and positive regulation of Tn_10_/IS_10_-promoted recombination by IHF: two distinguishable processes inhibit transposition off of multicopy plasmid replicons and activate chromosomal events that favor evolution of new transposons

.

Genes Dev.

9

,

1123

–

1136

.

[142]

(

1996

)

Similarities between initiation of V(D)J recombination and retroviral integration

.

Science

271

,

1592

–

1594

.

© 1997 Federation of European Microbiological Societies. Published by Elsevier Science B.V.