IS Families/IS6 family

General

There are at present nearly 160 family members in ISfinder from nearly 80 bacterial and archaeal species but this represents only a fraction of those present in the public databases. The family was named after the directly repeated insertion sequences in transposon Tn6 to standardize the various names that had been attributed to identical elements (e.g. IS15, IS26, IS46, IS140, IS160, IS176) , although more recently there has been some attempt to rename the family (see ), presumably because of accumulating experimental data from IS26 itself and the importance of this IS in accumulation and transmission of multiple anti biotic resistance, although this might potentially introduce confusion in the literature. IS6 family members have a simple organization (Fig. IS6.1) and generate 8bp direct target repeats on insertion. This family is very homogenous with an average length of about 800 bp and highly conserved short, generally perfect, IRs (Fig. IS6.2). Many are found as part of compound transposons (called pseudo-compound transposons described below) invariably as flanking direct repeats (Fig. IS6.1) a consequence of their transposition mechanism .

Distribution and Phylogenetic Transposase Tree

A phylogenetic tree based on the transposase amino acid sequence of the ISfinder collection (Fig. IS6.3) shows that the IS6 family members fall into a number of well-defined clades. This slightly more extensive set of IS corresponds well to the results of another wide-ranging phylogenetic analysis . These clades include one which groups all archaeal IS6 family members (Fig. IS6.3a) composed mainly of Euryarchaeota (Halobacteria ; Fig. IS6.3ai-iii). Group aiv includes both Euryarchaeota (Thermococcales and Methanococcales) and Crenarchaeota (Sulfolobales). Of the 10 clades containing bacterial IS: clade b includes examples from the Alpha-, Beta-, and Gamma-proteobacteria, Firmicutes, Cyanobacteria, Acidobacteria and Bacteroidetes ; clade c is more homogenous and is composed of Alphaproteobacteria (Rhizobiaceae and Methylobacteriaceae); clade d includes some Actinobacteria, Alpha-, Beta-, and Gamma-proteobacteria ; while clades e, f, g and h are composed exclusively of Firmicutes (almost exclusively Lactococci in the case of clades e and f). Clades I and j are more mixed. Clearly, the ISfinder collection does not necessarily reflect the true IS6 family distribution and these grouping should be interpreted with care. For example, although many do not form part of the ISfinderdatabase, IS6 family elements are abundant in archaea and cover almost all of the traditionally recognized archaeal lineages (methanogens, halophiles, thermoacidophiles, and hyperthermophiles (Fig 1.6.3).

Terminal Inverted Repeats.

The division into clades is also underlined to some extent by the IR sequences. As shown in Fig. IS6.2 (bottom), in spite of the wide range of bacterial and archaeal species in which family members are found, there is a surprising sequence conservation. In particular, the presence of a G dinucleotide at the IS tips and cTGTt and caaa internal motifs. Sequence motifs are more pronounced when each clade is considered separately (Fig. IS6.4).

Clade b

(n=24; with Alpha-, Beta-, and Gamma-proteobacteria, Firmicutes, Cyanobacteria, Acidobacteria and Bacteroidetes) maintains stronger traces of parts of these motifs (GG.. tcTGtt and CAaa).

Clade c

(n= 14; Alphaproteobacteria: Rhizobiaceae and Methylobacteriaceae) shows considerable conservation of an extended motif (GGG... TGTCGCAAA) and some conservation further into both IRL and IRR, although these are different for each end.

Clade d

(n=16; Actinobacteria, Alpha-, Beta-, and Gamma-proteobacteria) includes a well conserved GG..cTGTTGCAAA signature with little conservation further into each end.

Clade e

(n=7) is composed entirely of IS from Lactococcus and, as might be expected, exhibits fully conserved IR (GGTTCTGTTGCAAAGTTT) with significant conservation towards the IS interior, as would be expected in very closely related IS. However, some of the sequence conservation is present in both ends possibly indicating some functional role.

Clade f

(n = 8; Lactococcus and a single Leuconostoc) also exhibits a highly conserved ggTTCTGTTGCAAAGTTT signature at IRL (with only a partially conserved GG at the tip due to just 2 members) which is less conserved in IRR. Again, there is some internal conservation.

Clade g

(n = 8; composed of bacilli: Lysteria, Lactococcus, Enterococcus) also exhibits a similar signature (GGtTctgtTgcaAAgtTt) albeit less conserved. The left and has significant internal conservation.

Clade h

(n = 11; largely Staphylococci with 2 B. thuringiensis) also exhibit the typical GGTTCTGTTGCAAAGTTt signature and some internal conservation in IRL.

Clade i

(n = 1) is more heterogenous (Alpha proteobacteria: Methylobacterium, Paracoccus, Roseovarius, Rhizobium, Bradyrhizobium ; Deinococci and Halobacteria). It contains a poorly conserved IR sequence but does include a prominent gG dinucleotide tip and a poorly pronounced tgtcaagtt signature.

Clade j

Finally, bacterial Clade j (n= 5) composed entirely of Firmicutes (Natranaerobius, Clostridium and Thermoanaerobacter) exhibits a moderately well-defined internal signature TcTgTtAAgTt.

The archaeal-specific clades also generally exhibit well-defined consensus sequences.

Clade Ai

(n = 12) is composed uniquely of Halobacterial Euryarchaeota with a ggtaGTGTTcagatAaG signature and significant internal conservation which is different for each end.

Clade Aii

(n = 5), again, composed entirely of Halobacterial Euryarchaeota (Haloarcula, Halomicrobium, Natronomonas, Natronobacterium, Natrinema) also has well conserved ends, ggtcgTGTTTaGTT, and significant internal conservation which is different for each end.

Clade Aiii

This is also the case for Clade Aiii, also composed of diverse Halobacterial species (Halohasta, Haloferax, Natrinema, Natrialba, Halogeometricum, Natronomonas, Natronococcus, and Haloarcula): GgcACtGTCTAGtT.

Clade Aiv

However, Clade Aiv (n = 9) which includes both Euryarchaeota and Crenarchaeota, has poor conservation although on further analysis, an alignment shows significant conservation in the Sulfolobus and in the Pyrococcus groups with good interior conservation also in the 3 Pyrococcal members. It is possible that the IS ends in the Sulfolobus members have not been accurately identified.

The answer to the recent question:

“An analysis of the IS6/IS26 family of insertion sequences: is it a single family ?” is therefore “Probably, yes”.

Genomic Impact

Activity resulting in horizontal dissemination is suggested, for example, by the observation that copies identical to Mycobacterium fortuitum IS6100 ��[32]�(Clade d) occur in other bacteria: as part of a plasmid-associated catabolic transposon carrying genes for nylon degradation in Arthrobacter sp. ��[33]�, from the Pseudomonas aeruginosa plasmid R1003 ��[34]�, and within the Xanthomonas campestris transposon Tn5393b ��[35]�. Similar copies have also been reported in Salmonella enterica (typhimurium) ��[36]�, and on plasmid pACM1 from Klebsiella oxytoca (AF107205)��[37]�.

A single member, ISDsp3, present in single copy in Dehalococcoides sp. BAV1 carries a passenger gene annotated as hypothetical protein.

IS257 ��[38]�(Clade h) (also known as IS431) has played an important role in sequestering a variety of antibiotic resistance genes in clinical isolates of methicillin resistant Staphylococcus aureus (MRSA) (e.g. ��[39–42]�. It provides an outward oriented promoter which drives expression of genes located proximal to the left end. Moreover, both left and right ends appear to carry a –35 promoter component which would permit formation of hybrid promoters on insertion next to a resident –10 element ��[41]�, ��[43]�. Insertion of can result in activation of a neighboring gene using both a hybrid promoter and an indigenous promoter ��[41]�. IS257 is also involved in expression of tetA ��[44]� and dfrA ��[42]� in S. aureus.

IS26 ��[6–8]�(clade d) is encountered with increasing frequency in plasmids of clinical importance where it is involved in expression of antibiotic resistance genes and plasmid rearrangements (see ��[27,45–49]�). Its transposition mechanism contributes to its ability to assemble anti-bacterial resistance genes into clusters (e.g. ��[50]�). It can also form hybrid promoters capable of driving different antibiotic resistance genes: aphA7, bla_S2A (Klebsiella pneumoniae ��[21]�), bla_SHV-2a (Pseudomonas aeruginosa ��[51]�) and aphA7 (Pasteurella piscicida ��[52]�) as well as the wide spectrum beta-lactam resistance gene bla_KPC (Table IS and Gene Expression).

The formation of hybrid promoters on insertion (Table IS and gene expression) is clearly a general property of members of the IS6 family ��[21,41,42,53–55]�.

Another member, IS6100 ��[32]� (Clade d), often used as an aid in classifying mycobacterial isolates ��[56–58]� has been found to drive strA strB expression in X. campestris pv. vesicatoria, ��[35]�.

This IS family is able to form transposons which resemble compound transposons with the flanking IS in direct repeat but, because of the particular transposition mechanism of IS6 family members (see below), were called pseudo-compound transposons ��[1]�. These include Tn610 (flanked by IS6100 ��[32]�), Tn4003 and others (flanked by IS257 ��[39,59]�) and Tn6023 (flanked by IS26 ��[60]�)

Clinical Importance of IS26.

In view of the particular importance of IS26 in sequestering antibiotic resistance genes and generating arrays of these genes in clinically important conjugative plasmids and in the host chromosome (see ��[27,50]�), it is worthwhile devoting a separate section to the contribution of this IS to the clinical landscape. Recognition of its place as an important player has derived from the large number of sequences now available of multiple antibiotic resistance plasmids and chromosomal segments such as Genomic Resistance Islands (GRI). It is now no longer practical to provide a complete analysis of the literature. At present (19^th November 2020) a PubMed search using IS26 as the search term yielded nearly 450 citations. The references in the following are not exhaustive but simply provide examples.

Arrays

IS6 family members are often found in arrays (Fig. IS6.5 and Fig. IS6.6) in direct and inverted repeat in multiple drug resistant plasmids (e.g. S. typhimurium ��[27,60,61]�, Klebsiella quasipneumoniae ��[62]�, Acinetobacter baumannii ��[48,63]�, Proteus mirabilis ��[64]� and uncultured sewage bacteria ��[65]� among many others). These are often intercalated in or next to other transposable elements rather than neatly flanking ABR genes and can form units able to undergo tandem amplification.

Amplification

Shropshire et al ��[66]�, studying clinical isolates of non-carbapenemase-producing Carbapenem-Resistant Enterobacteria, non-CP-CRE, isolated from several patients with recurrent bacteraemia, observed an increase in carbapenem resistance partially due to IS26-mediated amplification of up to 10 fold of a cassette blaOXA-1 and blaCTX-M-1 which forms part of a larger chromosomal structure of IS26 arrays which they call TnMB1860 (Fig. IS6.6). It was unclear whether this cassette amplification was due to transposition activity or, as had been observed in similar, IS1-mediated, gene amplifications ��[67–72]�. Another example has been revealed by Hastak et al ��[73]� who analysed a multi resistant derivative of the clinically important, globally dispersed pathogenic, Escherichia coli ST131 subclade H30Rx, isolated from a number of bacteraemic patients and revealed that increased piperacillin/tazobactam resistance was due to IS26-mediated amplification of blaTEM-1B. A similar type of limited (tandem dimer) amplification of an IS26-flanked blaSHV-5-carrying cassette found in plasmids from a number of geographically diverse enteric species was identified in a nosocomial E. cloacae strain ��[74]�. A more extensive amplification (>10 fold) was observed with the same cassette located in a different plasmid in a well-characterised laboratory strain of Escherichia coli and occurred in a recA-independent manner ��[46]� and even higher levels of tandem amplification (~65 fold) of the aphA1 gene in the IS26-based Tn6020 were identified in Acinetobacter baumannii ��[75]�.

Cointegrating plasmids.

The earliest studies on this family of IS demonstrated that they could generate cointegrates as part of the transposition mechanism ��[5,7,9,12,29,76]�.

Several studies have now demonstrated that this can occur in a clinical setting. For example, plasmid pBK32533 (KP345882)��[77]�, carried by E. coli BK32533 isolated from a patient with a urinary tract infection is an IS26-mediated cointegrate between Klebsiella pneumoniae BK30661 plasmid pBK30661 (KF954759)��[78]� and a relative of Salmonella enterica p1643_10 (KF056330)��[79]�. Interestingly, the flanks of the IS26 copies at the junction of the two plasmids are TGTTTTTT-IS-TTATTAAT and TTATTAAT-IS-TGTTTTTT. This pattern of flanking sequences is consistent with an intramolecular inversion event (Fig. IS6.8). Plasmid pBK32533 was therefore probably generated in a multi-step process involving both inter- and intra-molecular transposition/recombination events. Additional examples have been identified in KPC-producing Proteus mirabilis ��[64]� and in Klebsiella pneumoniae also involving inversions ��[50,80]�

Organization

The putative Tpases are very closely related and show identity levels ranging from 40 to 94%. They generally range in length from 789 bp (IS257) to 880 bp (IS6100). However, a separate group represented by 7 members are somewhat larger (approximately 1200bp) as a consequence of an N-terminal extension with a predicted ZF. Several members (e.g. ISRle39a, ISRle39b and ISEnfa1) apparently require a frameshift for Tpase expression. It is at present unclear whether this is biologically relevant. However, alignment with similar sequences in the public databases suggests that ISEnfa1 itself has an insertion of 10 nucleotides and is therefore unlikely to be active.

Note that one isolate however, IS15, corresponds to an insertion of one iso-IS6 (IS15D) into another ��[4]�.

All carry short related (15-20 bp) terminal IR and generally create 8 bp DR. ��[6]�. A single orf is transcribed from a promoter at the left end and stretches across almost the entire IS. In the case of IS26 this is located within the first 82 bp of the left end and the intact orf is required for transposition activity.

In the case of IS26, an open reading frame is transcribed from a promoter located within the first 82 bp of the left end. It stretches across almost the entire IS, is required for transposition activity ��[8]�, and the predicted amino acid sequence of the corresponding protein exhibits a strong DDE motif Translation products of this frame have been demonstrated for IS240 ��[25]�. Little is known concerning the control of Tpase expression although transposition activity of IS6100 in Streptomyces lividans ��[81]� is significantly increased when the element is placed downstream from a strong promoter. This is surprising since IS generally incorporate mechanisms to restrict transposition induced by insertion into highly transcribed genes (see Fig 1.32.1).

Mechanism: the state of play

Early studies suggested that IS6 family members give rise exclusively to replicon fusions (cointegrates) in which the donor and target replicons are separated by two directly repeated IS copies (e.g. IS15D, IS26, IS257, IS1936) ��[5,7,9,13,82]�. More recent results principally with IS26 have suggested that, perhaps like IS1 (IS1 family) ��[83]� and IS903 (IS5 family) ��[84,85]� this IS may be able to transpose using several different pathways ��[16,86–88]�.

Cointegrate formation

Transposition of IS6 family elements to generate cointegrates ��[5,9,11,12]� presumably occurs in a replicative manner by Target Primed Transposon Replication (For a discussion see “Influence of transposition mechanisms on genome impact”; Fig 17.1 and Fig. 17.2). As shown in Fig. IS6.7 (top), intermolecular replicative transposition of this type generates fused donor and target replicons which are separated by two copies of the IS in direct repeat at the replicon boundaries. The initial direct repeats (DR) flanking the donor IS are distributed between each daughter IS in the cointegrate as is the DR generated in the target site. Recombination between the two IS then regenerates the donor molecule with the original DRs and a target molecule in which the IS is flanked by new DR. No known specific resolvase system such as that found in Tn3-related elements (see IS Derivatives of Tn3 family transposons) has been identified in this family but “Resolution” of IS6-mediated cointegrates was observed to depend on a functional recA gene in several cases and therefore occurs using the host homologous recombination pathway ��[5,9]�.

In addition to the intermolecular cointegrate pathway, intramolecular transposition using the replicative mechanism can give rise to deletion or inversion of DNA located between the IS and its target site. The outcome depends on the orientation of the two attacking IS ends (Fig. IS6.8)

IS6 family members are known to generate structures that resemble composite transposons in which a passenger gene (such as a gene specifying antibiotic resistance) is flanked by two IS copies. Generally, the flanking IS in these compound structures can occur as direct or inverted repeat copies (IS history; Fig 2.3, Fig. 2.4). However, in the case of IS6 functional “compound transposons”, the flanking IS are always found as direct repeats. This is a direct consequence of the (homologous) recombination event required to resolve the cointegrate structure. As shown in Fig. IS6.7 (bottom) ��[89]�, transposition is initiated by one of the flanking IS to generate a cointegrate structure with three IS copies. “Resolution” resulting in transfer of the transposon passenger gene requires recombination between the “new” IS copy and the copy which was not involved in generating the cointegrate. The implications of this model ��[1]���[89]� are that the transposon passenger gene(s) are simply transferred from donor to target molecules in the “resolution” event and are therefore lost from the donor “transposon”. Clearly this pathway could initiate from a donor in which the flanking IS6 family members were inverted with respect to each other. However, transposition would be arrested at the cointegrate stage because there is no suitable second IS to participate in recombination. It is for this reason that compound IS6-based transposons carry directly repeated flanking IS copies. These previously published models (e.g. ��[1,50,80,89]� have recently been revisited and it has been recently proposed ��[16]� that the term pseudo-compound transposons first used over 30 years ago ��[1]� should be resurrected to describe these IS6 family structures.

Circular transposon molecules: translocatable units (TU)

Although IS26 transposition appears to be replicative with formation of cointegrate molecules, results from in vivo experiments suggest that its transposition may be more complex ��[88]�. The idea that IS26 might mobilize DNA in an unusual way arose from the observation that IS6 family members can often be found in the form of arrays ��[87,88]� which could be interpreted as overlapping pseudo-compound transposons ��[1]� (Fig. IS6.5). Note that IS26 and potential IS26-based transposons do not necessarily carry flanking direct target repeats but, as is the case for other TE which transpose by replicative transposition such as members of the Tn3 family, intramolecular transposition would lead to loss of the flanking repeats (Fig. IS6.8). This led to the suggestion that IS26 might be able to transpose via a novel circular form called translocatable units (TU) ��[87,88]� such as those shown in Fig. IS6.9. These potential circular transposition intermediates which were proposed to include a single IS26 copy along with neighboring DNA are structurally similar to IS1-based circles observed in the 1970s (e.g. ��[67,70]�). Translocatable units differ from the transposon circles identified during copy-out-paste-in transposition by IS of the IS3 (Fig. IS3. 9A; IS3 family transposition pathway), IS21 (Fig. IS21.7), IS30, IS256 and ISL3 families where the circular IS transposition intermediate has abutted left and right ends separated by a few base pairs and is extremely reactive to the cognate transposase.

In stark contrast, for IS26, the IS ends would be separated by the neighboring DNA sequence rather than by a few base pairs (Fig. IS6.10).

Evidence for the excision step of translocatable units was obtained ��[87]� from the study of the stability of two IS26-based pseudo-compound transposons, Tn4352 ��[24]� and Tn4352B ��[90]� which carry the aphA1 gene specifying resistance to kanamycin. Tn4352B is a special mutant derivative of Tn4352 including an additional GG dinucleotide at the left internal end of one of the component IS26 copies to generate a GGGGG pentanucleotide at the IS tip which appears to render the transposon unstable. Cells carrying the plasmid lose the resistance gene from the mutant Tn4352B at an appreciable rate in the absence of selection. This generates a “donor” plasmid with one copy of IS26 flanked by the original Tn4352B-associated 8bp direct repeats and an excision product with the size expected for a TU containing the second IS flanked by the sequences of the original central segment presumably including the additional GG dinucleotide together with the aphA1 gene. TU formation appeared to be dependent on the GG insertion since no TU could be detected from the wildtype Tn4352 but not on the surrounding sequence environment. Excision required an active transposase. In a test in which the target plasmid also carried an IS26 copy (a targeted integration reaction – see below), there appeared to be no difference in cointegrate formation frequencies between single IS26 copies with or without the additional GG dinucleotide. However, results from a standard integration test into a plasmid without a resident IS26 copy were not reported. The excision process occurs in a recA background and frameshift mutations in both IS, which should produce severely truncated transposase, eliminated activity. However, excision was observed if the transposase of the GG-IS copy was inactivated but was eliminated when the same transposase mutation was introduced into the ”wildtype” IS copy. This is curious since it implies that the IS26 transposase must act exclusively in cis on the IS from which it is expressed (see Co-translational binding and multimerization). A summary of these results is shown in Fig. IS6.9. These data suggest that excision is driven by the wildtype IS26 (L), leaving the right hand IS in the excisant. At present, there is no obvious mechanistic explanation for this phenomenon. It should be noted that recombination between directly repeated copies of IS1 which flank the majority of ABR genes in the plasmid R100.1 (NR1) generates a non-replicative circular molecule, the r-determinant (r-det), with a single IS1 copy. In this case too, this “constitutive circle production is due to a (uncharacterized) mutation in the plasmid, although in this case, circle production requires recA ��[91]�.

However, the results appear to rule out two obvious models (Fig. IS6.10): since although both would generate the correct TU and “excisant”, the first (top panel) requires homologous recombination between two directly repeated IS26 copies (mechanistically equivalent to the “resolution” step in intermolecular IS6 transposition) and the second (bottom panel), which requires a functional transposase as observed ��[87,88]�, would not generate the correct flanking sequences. Modification of the transposition model to take into account the entire transposon (Fig. IS6.11) in which the active IS26L uses either of flanking sequences of IS26R does not generate the correct structures. Thus the observed structures must be generated by another, and at present unknown, pathway. One possibility is that TU are generated by reversing the non-replicative targeted insertion mechanism presented in Fig. IS6.13 (see Targeted Transposition below).

To summarize: it has been clearly demonstrated that circular DNA species carrying a single IS26 copy together with flanking “passenger” DNA can be generated efficiently in vivo from a variant plasmid replicon ��[90]� and also that replicons carrying a single IS26 copy are capable of integrating into a second replicon to form a cointegrate. This occurs at a frequency 10²-fold higher if the target plasmid contains a single IS copy and in a targeted manner not involving IS duplication.

The TU insertion pathway was addressed by transforming TU, constructed in vitro taking advantage of a unique IS26 restriction site, into recombination deficient cells carrying an appropriate target plasmid ��[86]�. Establishment of the aphA1-carrying TU was dependent on the presence of a resident plasmid carrying an IS26 copy and occurred next to the resident IS26 copy. The DNA of two TU each with a different antibiotic resistance gene was shown to undergo this type of targeted integration and, moreover, were able to consecutively insert to generate a typical IS26 array.

Targeted transposition.

Targeted IS26 transposition, was also observed in intermolecular cointegrate formation where cointegrate formation frequency was significantly increased about 100 fold if the target replicon also contained an IS26 copy ��[88]�. A similar result was obtained in Escherichia coli with a related IS, IS1216 ��[92]� whereas a third member of the family, IS257 (IS431) showed a much lower level of activity using the same assay. As for TU integration, this phenomenon does not appear to be the result of homologous recombination between the IS copies carried by donor and target molecules since the reaction was independent of RecA. Using a PCR-based assay, to identify the replicon fusions between IS26-containing donor and target plasmids, it was observed that the resulting cointegrate (Fig. IS6.12) did not contain an additional copy of IS26 which would be expected for replicative transposition (Fig. IS6.11). This suggests that the phenomenon results from a conservative recombination mechanism. Despite the absence of RecA, the observed cointegrate is structurally equivalent to the recombination product between the two IS26 copies in the donor and target plasmids. However, it indeed appears to be transposition related since the phenomenon requires an active transposase in both donor and target replicons ��[88]�. When each of the triad of conserved DDE residues were mutated individually in the donor plasmid, the targeted insertion frequency decreased significantly. It should be pointed out that the dominant negative effect of these mutant transposases may be due to interference with the multimeric transpososome presumably implicated in the cointegration reaction, a possibility that may also explain the negative effect of the N-terminal fragment of transposase which may be produced from the frameshift mutant.

Another characteristic of the products was that the 8 bp direct target repeats carried by the donor and recipient IS26 copies are in some way exchanged ��[88]� (Fig. IS6.12). This suggests a model in which transposase might catalyze an exchange of flanking DNA during the fusion process.

One possibility (Fig. IS6.13) is that two IS ends from different IS copies in separate replicons are synapsed intermolecularly in the same transpososome (Fig. IS6.13i). Strand exchange would then couple the donor and target replicons (Fig. IS6.13ii). A similar mechanism has been invoked to explain “targeted” insertion of IS3 and IS30 family members into TIRs (Fig. IS3.14) ��[93,94]�. Branch migration (Fig. IS6.13iii) would lead to exchange of an entire IS strand (Fig. IS6.13iv) and cleavage at the distal IS end and strand transfer (Fig. IS6.13v) would result in the observed cointegrate (Fig. IS6.13vi) containing a single strand nick on opposite strands at each end of the donor DNA molecule. These could simply be repaired or eliminated by plasmid replication. Each IS would be composed of complementary DNA strands from each of the original donor and target IS copies. This proposed mechanism would retain the DNA flanks of the IS in the original target replicon, be dependent on an active transposase and independent of the host recA system. It seems probable that mismatches between the two participant IS would inhibit the strand migration reaction. This may be the reason for the observation that introducing a frameshift mutation by insertion of additional bases into the transposase gene of either participating IS26 copy reduces the frequency of targeted cointegration ��[88]� since, not only does this produce a truncated transposase but also introduces a mismatch. As in the case of intermolecular targeting of the IS3 family member, IS911 ��[95]�, might require the RegG helicase to promote strand migration.

The model shown in Fig. IS6.13 presents the transposition process as a progression involving two consecutive temporally separated strand cleavages separated by a strand migration. However, it seems equally probable that both cleavage reactions are coordinated within a single transpososome (TnPedia section: The Transpososome) including both donor IS ends and the target IS ends. This would be compatible with the known properties of trans cleavage of several transposases in which a transposase molecule bound to one transposon end catalyses cleavage of the opposite end (TnPedia section: Cleavage in Trans: A Committed Complex).

The answers to many of these fascinating outstanding questions will be provided when the biochemistry of the reactions is known.

Acknowledgements

We would like to thank Susu He (Nanjing University) for stimulating discussions concerning the transposition models.