Transposons families/Tn3 family

Historical

Members of the Tn3 family were among the earliest transposons to be identified. In fact, the word “transposon” was used for the first time in 1974 by Hedges and Jacob in a seminal article in which they showed that ampicillin resistance could be transmitted between a number of different plasmids :

“We designate DNA sequences with transposition potential as transposons (units of transposition) and the transposon marked by the ampicillin resistance gene(s) as transposon A “.

TnA, later called Tn1, was isolated from the plasmid RP4 while the closely related TnB and TnC (later called Tn2 and Tn3 respectively) were isolated from plasmids RSF1010 and R1 . Tn3 proved to be inserted into another, larger Tn3 family transposon, Tn4 . A number of early studies using electron microscope DNA heteroduplex analysis (e.g. Fig. Tn3.1) demonstrated that movement of ampicillin resistance was accompanied by insertion of a DNA segment of about 4-5 kilobases (kb).

Fig. Tn3.1. Early Electron microscope heteroduplex study. One of many studies at this time. This example is from Rubens et al. 1976 . Examples of plasmid RSF1010 with Tn1 (TnA) insertions. Left. EcoRI linearized heteroduplex between two RSF1010 plasmids with different  Tn1 insertions at the same position but in opposite orientations, showing 2 single strand Tn1 loops (SS TnA). Right. Two RSF1010 plasmids with different Tn1 insertions at 150 bp distant, showing 2 single strands Tn1 loops. A scale bar of 1 kilobase is shown. Note the short stem, which could represent the ~40bp IRs.

The DNA sequence of the 4957 base pair (bp) Tn3 was obtained in 1979 and shown to be bordered by two inverted repeat sequences of 38 bp and included 2 genes in addition to the ampicillin resistance (beta-lactamase, bla) gene: a transposase gene, tnpA, and a gene involved in regulating tnpA and its own expression, tnpR (R for repressor). TnpR was subsequently shown to be a site-specific recombinase intimately involved in the transposition pathway which acts on a specific site, IRS (Internal Resolution Site) (Fig. Tn3.2 i). In its absence, insertion of two complete, directly repeated, Tn3 copies occurred .

It was suggested that this type of structure was an intermediate in Tn3 transposition and that the IRS site was required for recombination and subsequent segregation of the direct repeats to leave a single copy of Tn3 according to the Shapiro cointegrate model of replicative transposition (Fig. Tn3.2 ii; Fig. 2.7 Early models).

Indeed, Tn3 was shown to be instrumental in permitting transfer of a non-transmissible plasmid by a co-resident conjugative plasmid resulting in fusion of the two plasmids which were separated at their junctions by two directly repeated Tn copies .

Fig. Tn3.2. Organization and transposition mechanism of TnA. i) Organization.  Early information describing the structure of the 4957 base pair Tn1 (now called Tn1 and similar to Tn2 and Tn3). The figure shows the length (in amino acids, aa) and relative position and orientation of the transposase, tnpA, repressor (resolvase), tnpR and Beta-lactamase passenger gene, blaTEM. The genes are indicated by horizontal unfilled arrows. The long terminal inverted repeats (IRL for left, IRR for right) are shown as black triangles and their sequence is written below included in a blue oval. The TnpR binding site, res, is indicated as a black box . ii) The transposition cycle. The pathway involves replicative transposition resulting in fusion of the donor replicon (thick black circle) carrying the transposon (unfilled box) whose orientation is indicated by an arrow., and a target replicon (thin black circle). Replicon fusion requires TnpA and produces a cointegrated in which both donor and target replicons are fused and separated at each junction by a copy of the transposon in direct repeat. In a second step, TnpR binds to each res site and catalyses inter-res recombination, to regenerate the donor replicon and a target replicon which carries a single copy of the transposon .

A related TE, or Tn1000, was identified as part of the plasmid F and appeared as an insertion loop in heteroduplex analysis . It was also implicated in the integration of the F plasmid into the Escherichia coli host chromosome and deletion of chromosomal DNA in F’ plasmids derived from F-excision with flanking chromosomal DNA . It generates 5bp direct target repeat (DR) on insertion and carries similar ends to those of Tn3 and to IS101, a small 200bp sequence carried by the pSC101 plasmid .

Many other related transposons have since been identified with a highly diverse range of passenger genes (see and Fig. Tn3.4B). The tetracycline resistance transposon, Tn1721 from plasmid pRSD1 and the multi-resistance transposons, Tn4 from R6-5 and Tn21, a component of the 25 kb resistance determinant (r-det) of the plasmid NR1 (R100) are two of many early examples.

General Organization

Fig. Tn3.3. Tn3 family: diverse and variable. A number of examples of Tn3 family transposons are shown to illustrate the extent of diversity of the family. The Tn are indicated by pale-yellow horizontal boxes. Open reading frames are shown as horizontal arrows with the arrowheads indicating the direction of expression: purple, transposition-associated genes; blue, integron integrase; red, antibiotic resistance genes; bright yellow, plant virulence genes; chrome, heavy metal resistance genes; orange, toxin/antitoxin genes; pale salmon, other passenger genes. The inverted terminal repeats are shown as grey arrows. The names of each transposon are shown on the left and their length in base pairs on the right. Accession numbers: IS1071 (M65135); Tn3 (V00613); Tn4330 (X07651.1); TnXax1 (Tn7206) (AE008925); TnXc4 (Tn7210) (CP009039); Tn21, (AF071413); Tn4651.

Members of the Tn3 transposon family form a tightly knit group with related transposases and DNA sequences at their ends. The basic Tn3 family transposition module is composed of transposase and resolvase genes and two ends with related terminal inverted repeat DNA sequences, the IRs, of 38-40bp or sometimes even longer (Fig. 3.2 i) .

There is a large (~1000 aa) DDE transposase, TnpA, significantly longer than the DDE transposases normally associated with Insertion Sequences (IS) (see ). TnpA catalyzes the DNA cleavage and strand transfer reactions necessary for formation of a cointegrate transposition intermediate during replicative transposition.

A second feature of members of this transposon family is that they carry short (~100-150bp) DNA segments, res (for resolution) or rst (for resolution site tnpS tnpT – see below; ) at which site-specific recombination between each of the two Tn copies occurs to “resolve” the cointegrate into individual copies of the transposon donor and the target molecules each containing a single transposon copy (Fig. Tn3.2 ii)(see ). This highly efficient recombination system is assured by a transposon-specified sequence-specific recombinase enzyme: the resolvase.

There are at present three known major resolvase types: TnpR (which includes two subgroups, long and short with and without a C-terminal extension; Resolution), TnpI, and TnpS+TnpT, distinguished, among other things, by the catalytic nucleophile involved in DNA phosphate bond cleavage and rejoining during recombination: TnpR, a classic serine (S)-site-specific recombinase (e.g. ); TnpI, a tyrosine (Y) recombinase similar to phage integrases (see ); and a heteromeric resolvase combining a tyrosine recombinase, TnpS, and a divergently expressed helper protein, TnpT, with no apparent homology to other proteins .

The resolvase genes can be either co-linear, generally upstream of tnpA or divergent. In the former case the res site lies upstream of tnpR and in the latter case, between the divergent tnpR and tnpA genes. For relatives encoding TnpS and TnpT, the corresponding genes are divergent and the res (rst) site lies between tnpS and tnpT.

Examples of these architectures are shown in Fig. Tn3.3. Each res includes a number of short DNA sub-sequences which are recognized and bound by the cognate resolvases. These are different for different resolvase systems. But where analyzed, res sites also include promoters which drive both transposase and resolvase expression. Indeed, TnpR from Tn3 was originally named for its ability to repress transposase expression by binding to these sites (see later: Tn3 family resolution systems).

Diversity: TnpA Tree

The complexity of these Tn resides in the diversity of other mobile elements incorporated into their structures (such as IS and integrons as well as other Tn3 family members – see - and other passenger genes). The most notorious of these genes are those for antibiotic and heavy metal resistance although other genes involved in organic catabolite degradation and virulence functions for both animals and plants (Fig. Tn3.3) also form part of the Tn3 family arsenal of passenger genes.

The diversity of Tn3 family members was investigated using a library of carefully annotated examples in the ISfinder database , those listed in Nicolas et al. , those resulting from a search of NCBI for previously annotated Tn3 family members (March 2018) and those obtained using a script, Tn3_TA_finder, which can searched for tnpA, tnpR, genes located in proximity to each other (Tn3finder, https://tncentral.proteininformationresource.org/TnFinder.html; Tn3_TA_finder, https://github.com/danillo-alvarenga/tn3-ta_finder) in complete bacterial genomes in the RefSeq database at NCBI. This yielded 190 Tn3 family transposons for which relatively complete sequence data (transposase, resolvase, and generally both IRs) were available.

Full annotations can be found at TnCentral (https://tncentral.proteininformationresource.org/index.html). A tree based on the transposases of these transposons is shown in Fig. Tn3.4A .

Fig. Tn3.4A. A phylogenetic tree of 190 Tn3-family members based on their TnpA sequences. Lima Mendez et al. Tn3 family members were extracted from the ISfinder database, which served to generate the subgroups defined in Nicolas et al. . Many others were drawn from the literature and have been given official names (Tn followed by digits, e.g. Tn1234: https://transposon.lstmed.ac.uk/tn-registry) while others were identified using Tn3_finder software (TnCentral: https://tncentral.ncc.unesp.br/TnFinder.html) and given temporary names (these have now been registered at the Transposon registry and have been assigned Tn numbers which can be found as synonyms in TnCentral). Each is associated with its GenBank accession number; the GenBank file contains either the extracted transposon or the DNA sequence from which it was extracted (e.g., DNA fragment, plasmid or chromosome). Numbers above the lines of each clade indicate the maximum likelihood bootstrap values. The sub-groups adhere closely to those defined by Nicolas et al. with some minor variations resulting from the significantly larger Tn sample. The majority of members carry tnpR, serine resolvases (purple circles) although one group in the Tn3 clade encodes “long’ serine resolvases. Those that include tnpI or tnpT/tnpS are indicated by salmon and pink circles, respectively. The TA gene pairs are indicated by coloured squares. Note that Tn5501.5 carries a mutation which truncates its toxin gene, leaving the antitoxin intact. The outer squares represent the toxin and the inner squares, the antitoxin. The five toxin types are: Gp49 (PF05973), purple; PIN_3 (PF13470), dark green; PIN (PF01850), bright blue; ParE (PF05016), yellow; and HEPN, black. The antitoxins are: HTH_37 (PF13744), orange; RHH_6 (PF16762), blue; [1]Phd/YefM (PF02604), magenta; RelB/ParD/CcdA/DinJ, dark grey; AbrB/MazE, light grey; MNT, bright green. The corresponding Tn names and accession numbers are highlighted in bold type for clarity. Note that the branches have been extended for clarity.

The tree defines 7 deeply branching clades which supports the divisions proposed by Nicolas et al., . They were named after a representative Tn from each clade: Tn3; Tn4651; Tn3000; Tn1071; Tn21; Tn163; and Tn4330. As can be seen from Fig. Tn3.4A, the vast majority of Tn3 family members encode a tnpR/res resolution system and encode a TnpR without the C-terminal extension (shown by blue circles) and a small group which encodes a TnpR derivative with the C-terminal extension (Fig. Tn3.4A). However, a significant sub-group of the Tn4651 clade encodes the tnpS/tnpT/rst resolution system (pink circles) while the tnpI/irs is represented in only three cases.

An overview, extracted from TnCentral, of the diversity and distribution of different passenger genes within the Tn3 family and their presence in different bacterial hosts is shown in Fig. Tn3.4B.

Fig. Tn3.4B. Distribution of Different Passenger Genes within the Tn3 Family. A key to the types of passenger genes is shown on the left of the figure. This figure was kindly provided by Nicolas Aryanpour based on information extracted and collated from the TnCentral database by Gipsi Lima-Mendez.

Tn3 family complementation groups

Early studies on the relationship between different Tn3 family members revealed that they could be divided into different functional groups by genetic complementation of their tnpA and tnpR genes .

Transposition-deficient tnpA mutants of Tn1721 (Tn21 clade; Fig. Tn3.4A) and the mercury resistance transposon Tn501 (close to Tn1721 in the Tn21 clade;) could be complemented in trans by co-resident wild type copies of either Tn21, Tn501, or Tn1721, while transposition of a Tn21 tnpA mutant could only be restored by Tn21. Moreover, Tn3 was unable to complement either Tn21, Tn501, or Tn1721, and vice versa . Similarly, a Tn21 tnpR mutant could be complemented by Tn21, Tn501 or Tn1721, but not by Tn3. Moreover, mutations in the Tn2603 tnpA and tnpR genes could be complemented by mercury resistance transposons Tn2613 and Tn501 (although Tn501 was much less efficient in complementation than Tn2613) but not by gamma delta, Tn2601 or Tn2602 (both of which resemble the Tn3 group – see Fig. Tn3.7 A) . In this context, it is perhaps useful to note the Tn501 and Tn1721 are located at some distance from Tn21 in the tnpA phylogenetic tree. This reinforced the idea, based principally on the direction of transcription of their tnpA and tnpR genes, that the Tn3 family could be divided into 2 major groups: Tn3 and Tn21 .

Tn3 and Tn21 groups

Grinsted et al. identified at least five Tn3 family subgroups which correspond to those shown in Fig. Tn3.4A. In addition to the Tn3 and Tn21 subgroups, the others included Tn2501 (Tn163 subgroup), Tn917/Tn551 (Tn4430 subgroup) and Tn4556 (Tn3000 subgroup). Tn917 and Tn551 are quasi-identical and Tn4430 was included in a separate subgroup because it encodes a resI/tnpI resolution system.

These divisions were based on the observations that: transposition proteins within each group were at least 70% similar or identical whereas this value was only about 30% between groups and that the IR sequences were less than 26/38 identical. The authors propose a model for the evolution of the Tn3 family transposition modules (Fig. Tn3.5) in which two ancestral modules were assembled: the first included a tnpR gene (which they suggest was flanked by an invertible DNA segment incorporating the res site) and a tnpA gene. This subsequently gave rise to each of the Tn3 subgroups by tnpR/res inversion and sequence divergence. For Tn such as Tn4430, the assembly involved tnpI/res and tnpA components. The tnpS/tnpT/rsc resolution system was not included since it had not been identified at that date but could easily be incorporated into this scheme. To our knowledge, the proposed ancestral components in this scheme have not yet been identified.

Fig. Tn3.5. Hypothetical pathways Tn3 family evolution (adapted from ). The transposons are shown as yellow filled boxes, resolvase (R) and transposase (A) genes as horizontal purple arrows. Res sites are shown as green boxes. The figure proposes that present day Tn3 family members arose from three ancestors: one in which the tnpR gene lies downstream from tnpA with res between them (left: Tn4556), a second which encodes a tnpI resolvase and an alternative res site (right: Tn4330) and one which is a precursor to the majority of family members (middle). This is proposed to include tnpR and tnpA genes expressed in the same direction and two sequences, one at the left end and another between tnpR and tnpA which gave rise to res. Inversion of tnpR generated one Tn3 class which, on acquisition of the bla (penicillin resistance) passenger gene, became Tn3 itself, while in the other branch, the internal res sequences were lost to generate the Tn21 class and acquisition of an erm (erythromycin resistance) passenger gene gave rise to Tn917/Tn551. The branches also correspond to the sequence relationships of tnpR and tnpA in these groups.

The diversification of different Tn21 clade members was also examined (Fig. Tn3.6) and forms two subclades. One includes Tn21, Tn2613 (whose sequence is not available but which may be identical to Tn5060-AJ551280.1) and Tn3926 (with only a partial sequence available but which complements a tnpA-defective Tn21 but not Tn1721 or Tn501 mutants ). The other includes Tn501, Tn1722, Tn1721 and Tn4653. Tn501 and Tn1721 are located in a sub-clade distinct from Tn501 and Tn5060 (Fig. Tn3.4A). In this scheme, mercury resistance was proposed to have been acquired twice independently in each subclade, early in the Tn21 subclade lineage and later in the line leading to Tn501. The ancestor of Tn21 had acquired an integron platform transported by a Tn402 family transposon and Tn1721 was derived from Tn1722 by acquisition of a tet resistance gene.

Fig. Tn3.6. Proposed Evolution of elements of the Tn21 subgroup (adapted from ). This is based on genetic complementation and sequence similarities. The ancestral Tn of the Tn21 group was proposed to have acquired a mercury (mer) resistance operon early in the pathway (left) with the Tn21 lineage acquiring sulfonamide resistance by insertion of an integron platform. In the right-hand pathway, acquisition of a mercury (mer) resistance operon was proposed to give rise to Tn501, while tetracycline was proposed to have been acquired leading to formation of Tn1721 and the ability to degrade toluene (xyl) was acquired to generate Tn4651 (see Fig. Tn3.3).

The Tn21 Clade

The Tn21 is a large group with 49 members at present in TnCentral (most of these are shown in Fig. Tn3.7 A). Like the entire Tn3 family, Tn21 clade members possess highly conserved IRL and IRR (Fig. Tn3.7 B, C and D).

Fig. Tn3.7A. The relationship between the transposases of different members of the Tn21 clade expanded from Fig.Tn3.4A. Each transposon is shown together with its Genbank accession number. The small purple circles indicate that all members encode a tnpR resolution system. One member, TnSod9, encodes a toxin/antitoxin gene pair (black and green squares) and is shown in bold type. Those Tn encircled by blue boxes are treated in detail in the text. The lozenges on the outer circle indicate the presence and type of passenger gene carried by the corresponding Tn. mer: mercury resistance operon; ars: arsenic resistance; chr: chromate resistance; red: integron plataform; pale-yellow: other passenger genes; pale-yellow/red: other passenger genes + non integron; white: no passenger genes; black line: not in TnCentral.

Many clade members encode tnpR with a res site immediately upstream and, in a majority (but not all), tnpA is located downstream and in the same orientation. The res sites of this class (Fig. Tn3.7 E) show a high degree of identity (Fig. Tn3.7 F). However other tnpR/tnpA configurations also occur (Fig. Tn3.3; Fig. Tn3.7 E) and their res sites (see below: The Tn1721, Tn21 and Tn501 res) show relatively good conservation (Fig. Tn3.7 F)

• Fig. Tn3.7. Panels B, C and D (use the arrows to scroll the figures).
• 

  Fig. Tn3.7B. The Ends of Members of the Tn21 Clade. Left (IRL) and right IRR inverted terminal repeats are shown in WebLogo format. Three well conserved regions are interested in horizontal double-headed arrows.

• 

  Tn3.7C and D. Alignments of IRL and IRR of Tn21 family members. Transposon names are shown on the left. Those with non-conforming names have all now been attributed names in the transposon registry

• 

  Tn3.7C and D. Alignments of IRL and IRR of Tn21 family members. Transposon names are shown on the left. Those with non-conforming names have all now been attributed names in the transposon registry .

Tn3.7E. Tn21 clade tnpR/tnpA configurations. The right-hand column shows members with the “classical” Tn21 configuration in which the passenger genes are located upstream from tnpR and the res site; the next column indicates those members with the same tnpR/tnpA configuration but in which the res site has been split by insertion of a Tn401-like integron structure; the third column includes those clade members with divergent tnpR/tnpA and a res site between the two genes; the fourth column lists members with the divergent tnpR/tnpA but in which passenger genes have been inserted in between.  Where appropriate, registered transposon names are shown in brackets

Fig. Tn3.7F. res site alignment of Tn21 clade members with co-linear tnpR/tnpA genes. The Tn names are shown to the left of the figure. Sequence conservation is shown by the depth of the blue background. Sites I, II and III are indicated.

Derivatives with a simple mercury operon.

In general, passenger genes in this clade are located upstream of tnpR and the res site (Figs. Tn3.7 G-N). Ten carry only genes for resistance to mercury salts.

Two of these, Tn5060 (AJ551280.1) (Tn3.7 G), the proposed ancestor of the Tn21 integron group (Tn3.7 I) , and Tn20 (AF457211.1) are nearly identical except for a few SNP and a deletion of a few base pairs in ufrM (Tn20).

Fig. Tn3.7G. Integron platform insertion sites. The top of the figure shows a probable generic ancestor of Tn21 clade Transposons which carry Tn402-based class I integrons. Horizontal filled arrows represent the open reading frames: chrome color, mercury resistance genes (merR, merT,merP,merC,merA,merD,merE); pink, reading frame of unknown function (urfM); purple, transposition- related genes tnpR (resolvase) and tnpA (transposase). The resolution site, res, is shown in green. The points of integration (into an ancestor, such as Tn5060 or Tn20) are shown by vertical red and blue arrows.  Integration into urfM (blue) gave rise to Tn21 and its related transposons (blue box, left). Those which for which the DNA sequence is available (blue) share an identical insertion site together with a 5 base pair flanking target duplication expected to be generated by Tn402 family integration. Those for which no sequence is available also appear to carry  the inserted integron in an identical position as judged by restriction mapping. Integration into the res site of an ancestor such as Tn1696.1 or Tn5036 (right) gave rise to Tn1696 and Tn6005 (red box, right). The transposons in the middle  column are referred to in the text.

These are quite different in sequence both in the mer operon and in tnpR/tnpA segments from the other transposons of similar organization. Tn1696.1 (CP047309) and Tn5036 (Y09025) differ by only a few SNPs while Tn4378 (CP000355), Tn6203 (CP065412) and Tn6346 (KM659090) are also quite different from the these. Tn4378 and Tn6203 show many sequence differences along their entire length as does TnAs2 (JN106175.1) while clearly, Tn6346 shares identity with Tn4378 over the entire length of the mer operon up to res but shows variability in the tnpR/tnpA region.

This clearly indicated that there has been an exchange by inter res recombination between two different transposons (Tn3.7H). A similar recombination has occurred with Tn501. In addition, Tn4380 appears to have been derived from Tn6346 by deletion of the entire res site. Thus Tn4378, Tn6436 (Tn4380) and Tn501 share highly related mer operons but vary in the sequences of tnpR and tnpA.

Fig. Tn3.7H. Relationship between mercury resistant transposons without integron insertions and of similar organization.  i) The features are identical to those in Fig. Tn3.7G. Low resolution alignments against Tn4378 are shown as horizontal red lines above. While TnAs2 (Tn7144; JN106175.1) and Tn6203(CP065412) show significant divergence along their entire length, Tn6346 (KM659090), Tn4380 (CP000354), and Tn501 (Z00027) all share their mercury resistance genes with those of Tn4378. However, Tn6346 (KM659090), Tn4380 (CP000354) exhibit significant divergence in tnpR and tnpA and Tn501 shows even a higher level of divergence. The changes in levels of identity occur in or close to the res site. The Tn is indicated by a pale-yellow horizontal box. Open reading frames are shown as horizontal arrows with the arrowheads indicating the direction of expression: purple, transposition-associated genes; chrome, mercury resistance genes; pale salmon, other passenger genes. The inverted terminal repeats are shown as grey arrows. ii) The DNA sequences in the res site region showing the three subsites, resI, resII and resIII of Tn6346 and Tn4378 and Tn501. The non-identical bases are shown in red, conserved bases in two of the three sequences are in bold and underlined and the probable recombination point ATA is shown in red/bold.

Derivatives with class I integrons: 2 events leading to multiple antibiotic resistance

At least 22 Tn21 clade members carry class I integrons (Fig. Tn3.7 A and Fig. Tn3.7 I) although the DNA sequence of some of these is not available. These are transmitted by Tn402 derivative transposons which exhibit pronounced target specificity (more details at: Tn402 family) and show a preference for insertion into or close to Tn3 family res sites or into plasmid res sites. A major pathway for the acquisition of passenger genes was the initial integration of a Tn402-like transposon which carried a class I integron platform. The integron insertions have occurred at one of two positions in the Tn5060 /Tn20 related examples (Fig. Tn3.7 G). In one group, which all encode an identical mer operon, insertion occurred in a precise position in a gene of unknown function, ufrM (see: The Tn21 Lineage) (Fig. Tn3.7 I). Since these occur at the same nucleotide, it seems possible that all diverged from a single insertion event.

Fig. Tn3.7I. Integron insertion into ufrM. An alignment of various Tn21 group transposons (red horizontal lines) against Tn5060 (AJ551280.1; features identical to those described in Fig. Tn3.7G) show that insertion of the integron platform (small red triangles) occurs at the same position in ufrM. A number of small sequence differences with Tn5060 are shared with Tn20, suggesting that Tn20 was perhaps the ancestor of this group. Tn21 (AF071413); Tn2411 (FN554766); Tn2424 (UGCJ01000005); TnAs3 (Tn7145; CP000645.1); Tn5086 (CP054343); Tn4 (KY749247.1); Tn21.1 (MH257753); Tn21.2 (MH626558); Tn20 (AF457211.1). The Tn is indicated by a pale-yellow horizontal box. Open reading frames are shown as horizontal arrows with the arrowheads indicating the direction of expression: purple, transposition-associated genes; chrome, Mercury resistance genes; pale salmon, other passenger genes. The inverted terminal repeats are shown as grey arrows.

In the others, the res site itself has been targeted: at two slightly different positions both in the Tn1696 (Fig. Tn3.7 J) (also carrying a mer operon) and Tn1721 (with an mcp gene) groups (Fig. Tn3.7 K) while a third example can be observed in Tn5045.1 carrying the tao gene cluster (Fig. Tn3.7 L). The fact that integrons In2 and In4 are located in different sequence environments in two distinct mercury resistance transposons, Tn21 and Tn1696 has previously been noted . Thus, although widespread in nature, class 1 integrons appear to have inserted in only six target sequences in the entire Tn21 clade in TnCentral. The significant variability therefore arises principally by acquisition and loss of integron cassettes and by frequent various degrees of loss by deletion/inactivation (see: Tn21 lineage) of the Tn401 transposition genes tniA,B,Q and its resolvase tniR (see: Tn402 family).

Fig. Tn3.7J. Tn1696 and relatives. The transposon features are identical to those in previous figures. The Tn is indicated by a pale-yellow horizontal box. Open reading frames are shown as horizontal arrows with the arrowheads indicating the direction of expression: purple, transposition-associated genes; chrome, Mercury resistance genes; pale salmon, other passenger genes. The inverted terminal repeats are shown as grey arrows. The presumed ancestor, Tn16961.1 (CP047309), is shown below and alignments of Tn5036 (Y09025) a similar transposon without the integron insertion), Tn1696 (U12338.3) and Tn6005 (EU591509.1; both with integron insertions; small vertical triangles) are shown as horizontal red lines above. The top of the figure shows the sequence of the Tn1696.1 res site, indicating the position of integron Insertions (vertical blue arrows). The target sequences which are the flanking DRs are shown in red.

Fig. Tn3.7K. Tn1721 and relatives. The bottom panel shows the tetracycline resistance transposon Tn1721 (X61367.1) in which the passenger genes are located both upstream and downstream of tnpR and tnpA. The Tn are indicated by pale-yellow horizontal boxes. Open reading frames are shown as horizontal arrows with the arrowheads indicating the direction of expression: purple, transposition-associated genes; red, antibiotic resistance genes; pale salmon, other passenger genes. The inverted terminal repeats are shown as grey arrows. A potential ancestor, TnpCTXM9 (Tn7181; probably identical to Tn1722; CP031724) is shown above. It carries the upstream passenger gene, but not the downstream tet genes. Alignments of  Tn1721.1 (HQ730118.1) and TnCfrpOZ172 (Tn7154; CP016763.1) both carrying integron insertions (small vertical triangles) are shown as red lines above. The DNA sequence of the res site is shown with the positions of resI, resII and resIII and the points of integration. The target sequences which are the flanking DRs are shown in red.

Fig. Tn3.7L. Tn5045.1 and relatives. An integron insertion into Tn5045.1 (NC_008357.1) generated Tn5045 (FN821089.1). The DNA sequence of the Tn5045.1  res site is shown with the positions of resI, resII and resIII and the point of integration. The target sequences, which are the flanking DRs, are shown in red.

Derivatives with upstream passenger genes: colistin resistance.

Of the four colistin resistant examples (Fig. Tn3.7 M): TnSen1.1 [Tn7191] and TnSen1.2 [Tn7192] are nearly identical except that TnSen1.2 carries an ISPa96 insertion; both TnEc026 [Tn7159] and TnMCR5ECO26H11 [Tn7163] are identical but TnEcO26 has two right ends.

Moreover, while the left segment of all 4 are closely related, there appears to have been a recombination event in the region of the res site two right ends and TnSen11.2/TnSen1.2 and TnEcO26/ TnMCR5ECO26H11 carry divergent tnpR and tnpA.

Fig. Tn3.7M. Tn21 Colistin resistance. Top panel: the colistin (mcr5) resistance transposon TnSen1.1 (Tn7191; KY807921) in which the passenger genes are located upstream of tnpR and tnpA. Alignments of  TnSen1.2 (Tn7192; CP028162), TnMCR5ECO26H11 (Tn7163; BEPM01000040) and TnEc026 (Tn7159; BDIH01000107.1) are shown as red lines above. Apart from three small sequence differences, the major difference with TnSen1.1 is that TnSen1.2 carries an insertion of ISPa96 within the transporter gene(s) at its left end (red vertical triangle). TnMCR5ECO26H11 and TnEc026 are identical except that TnEc026 includes an extension at its right end which duplicates an IRR sequence. Bottom panel: The DNA sequence of the res site is shown below with the proposed positions of resI, resII and resIII and the points of recombination. The non-identical bases are shown in red, conserved bases are in bold and underlined and the probable recombination point ATA is shown in red/bold.

Derivatives with upstream passenger genes: other passengers.

There are a number of other Tn21 clade members with different upstream passenger genes. Analysis of these reveals that, although there has been some diversification of the tnpR and tnpA genes (Fig. Tn3.7 N), there is a clear breakpoint in identity which occurs at the res site. Sequence analysis (Fig. Tn3.7 N) indicates that the break in identity occurs at the potential AT recombination dinucleotide (see: Resolution topic below) strongly suggesting that acquisition of various passenger genes frequently occurs by modular exchange via inter-res recombination.

Fig. Tn3.7N. Tn21 other passenger genes. Top panel: TnPa40 carrying chromate resistance genes (Tn7173; CP003149) is shown together with alignments of TnPa40.1 (Tn7174; CP020704.1; carrying a miscellaneous set of passenger genes and an insertion of IS1411), Tn5045.1 (Tn1013; NC_008357.1; carrying a tao operon) and Tn4656 (NC_008275.1; carrying mcp genes). Bottom panel: The DNA sequence of the res site is shown below with the proposed positions of resI, resII and resIII and the points of recombination.

Derivatives with divergent tnpR and tnpA

There are a number of Tn21 clade members in which the tnpR and tnpA genes are expressed divergently. Several of these (e.g. Tn4659, TnAcsp1 [Tn7133], TnEc1 [Tn7158] and TnSba14 [Tn7190]) (Fig. Tn3.7 O) do not encode passenger genes and are not closely related, while others encode heavy metal resistance operons located between tnpR and tnpA (e.g. TnLfArs [Tn7162], TnOtChr [Tn7169]) while TnPa38 [Tn7172] encodes genes of unknown function and TnSod9 [Tn7199] is the only example in the Tn21 clade to encode a Toxin/Antitoxin gene pair. These are not closely related.

Fig. Tn3.7O. Tn21 clade Members with Divergent tnpR and tnpA. The Tn are indicated by pale-yellow horizontal boxes. Open reading frames are shown as horizontal arrows with the arrowheads indicating the direction of expression: purple, transposition-associated genes; chrome, heavy metal resistance genes; pale salmon, other passenger genes; bordeaux, hypothetical. The inverted terminal repeats are shown as grey arrows. TnLfArs (Tn7162, DQ057986.1) ; TnOtChr (Tn7169, EF469735.1); TnPa38 (Tn7172, CP003149).

The Tn21 Lineage.

The Tn21 lineage is an example of the plasticity of Tn3 family transposons. Tn21 was originally identified in the multiple antibiotic resistance plasmid NR1/R100 , as part of the IS1-flanked r-determinant and its component antibiotic resistance genes were first mapped by restriction enzyme digestion and cloning . The Tn21 group of transposons appear to be very successful as judged by their distribution.

Fig. Tn3.7P. Early Restriction site maps of related antibiotic resistance Tn21 clade transposons. Redrawn from . Transposons are indicated as horizontal yellow boxes. Retriction sites: EcoRI (down-arrow); HindIII (up-arrow); BamHI (down circle-arrow)); PstI (up circle-arrow). Passenger genes: mer: mercury; Su: sulphonamide; Sm: streptomycin; bla: beta-lactamase. There is no available sequence for many of these transposons: Tn2613, Tn2608, Tn21 (AF071413), Tn2603, Tn2607, Tn2601, Tn4 (KY749247.1), Tn3 (V00613).

This is arguably the result of acquisition of an integron platform permitting incorporation of various resistance genes as integron cassettes (Fig. Tn3.7 A and Fig. Tn3.7 G). Tanaka and collaborators proposed in the early 1980s that Tn21-like transposons which carry a variety of antibiotic resistance genes are related and evolved from an ancestor carrying a mercury resistance operon (Fig. Tn3.5; Fig. Tn3.7 P).

Tn21 itself is a complex collection of intercalated TE and a comprehensive and detailed schemes for its formation has been proposed (see Fig. Tn3.6.; Fig. Tn3.7 P). Unfortunately, although the DNA sequences of some of the component transposons are now available (e.g. Tn4, Tn21, Tn2411), many are not and comparison was based on physical and functional maps (restriction, genetic features) .

This scheme was later expanded with the addition of more up-to-date information to include a number of potential Tn21 descendants (see ) (Fig. Tn3.7 Q). It was proposed that a Tn21 precursor (Tn21) acquired an integron platform such as is found in Tn4 (for convenience, called In_Tn4 here) which then received an insertion of IS1353 into a resident IS1326 copy to generate In2 found in Tn21 .

Fig. Tn3.7Q. The proposed lineage of the Tn21 integron-carrying group of transposons. Redrawn and amended from . The steps leading to each transposon (within a red box) are shown in blue boxes. Filled red boxes show transposons for which the DNA sequence is available. The sequences of Tn2608 and Tn5086 were reconstructed from known sequence elements. Tn2411 (FN554766); Tn4 (KY749247.1); Tn21  (AF071413); Tn2424 (UGCJ01000005).

Although the Tn21 group ancestor prior to acquisition of the mercury resistance genes is at present unknown, the later identification of a mercury resistance transposon, Tn5060 (AJ551280.1), isolated from the Siberian permafrost (Fig. Tn3.7 R) provided a possible candidate for the hypothetical Tn21 precursor, Tn21°.

Fig. Tn3.7R. Proposed lineage of the Tn21 integron-carrying group of transposons. Redrawn from . Integration of a type 1 Integron was proposed to occur into an ancestral transposon Tn21D (Top) into the ufrM gene of unknown function, giving rise to a characteristic five base pair DR (ATGGA) and to Tn2411 (FN554766) (Middle).  Insertion of a copy of IS1353 into the resident IS1536 then gave rise to Tn21 (Bottom).

Other examples of this Tn such as Tn20 (AF457211) (Fig. Tn3.7 I) can be identified which share a number snips with other members of the group compared to Tn5060 and therefore is perhaps a better candidate as an ancestor. An alternate view of the path from Tn5060 to Tn21 is that evolution of the integron platform occurred “in situ” by the gradual loss/accumulation of component TE. In this scheme (Fig. Tn3.7 S and Fig. Tn3.7 P), a first step would be insertion into the ufrM (unknown function) gene of a Tn402 family transposon to provide the integron platform (Fig. Tn3.7 S).

Fig. Tn3.7S. Modified proposed lineage of the Tn21 integron-carrying group of transposons.  The identification of Tn5060 (Top) and Tn20 (Fig. Tn3.7G) revealed possible Tn21 ancestors. The figure shows (Bottom) insertion of a simple active copy of a Tn402 derivative, the  vector of type 1 integrons, carrying a complete set of transposition genes at position  4626 – 4633 to generate the 5 bp DR ATGGA.

Although it has been shown that transposition of defective Tn402 transposons (e.g. In0 and In2) can be complemented by a related, wildtype copy , it seems simpler to hypothesize that an initial insertion involved a Tn402 derivative with a complete functional set of Tn402 transposition genes. We have chosen a simple integron platform, In_Tn1721.1 from Tn1721.1 (HQ730118.1), for convenience.

This carries tniA,B,Q, the resolvase tniR together with the Tn402 res site, both ends (IRt and IRi), the integron integrase int and a common qac gene cassette. Insertion into the Tn5060 urfM gene generates a 5 bp DR (Fig. Tn3.7 S) and leads to the formation of tnpM from the 3’ end of ufrM (serendipitously generating an ATG initiation codon) . TnpM has been suggested to be a transposition regulatory gene (but see Resolution below). Subsequent steps in the Tn21 lineage (Fig. Tn3.7 T) would then involve modification of the integron platform by acquisition of the typical GNAT (previously known as orf5) and sul genes, decay of the Tn402 transposition genes and insertion, first of IS1326 (resulting in In0) followed by acquisition of the aadA integron cassette (generating In_Tn4) and, finally, insertion of IS1353 into IS1326 (IS1326::IS1353) between IRL and the start of the istA gene presumably not affecting IS1326 transposition functions (generating In2).

Fig. Tn3.7T. A plausible lineage of the Tn21-related integrons. The insertion point of all transposons of the Tn21 group is identical, suggesting that insertion occurred only once and subsequence variation to occurred from this common ancestor. The figure shows a plausible pathway for the changes in the integron structure leading to that found in Tn21. The relevant transposons are indicated at the left of the figure. The ancestral integron structure chosen as the simplest known example of a functional Tn401 derivative transposon is that found at present in Tn1721.1 (In_Tn1721.1; HQ730118.1) (first panel) although Tn1721.1 cannot be a direct source since the integron is inserted into another target in this transposon (Fig. Tn3.7K). In_Tn1721.1 could then lead to In0 (U49101) (second panel) by acquisition of two integron cassettes, loss of the Tn402 tniQ and tniR genes and part of tniB  and insertion of IS1356. Acquisition of an additional integron cassette would then generate In_Tn4 (KY749247.1) (third panel) found in Tn4 and  Tn2411 (FN554766) and insertion of IS1356 would generate In2 (AF071413) in Tn21 (fourth panel).

Due to their conservation in a large number of class I integron platforms, the DNA region including the sul, qac and GNAT family (previously called orf5) genes has been called the 3’CS (conserved segment) while that including the attI site and intI gene has been called the 5’CS (however, using a more extended data set it was noted that, while the 5’CS was highly conserved across a number of integrons, the 3’CS proved to be somewhat more variable ). Tn2411 is not only the precursor of Tn21. It was proposed to give rise to additional transposons (Fig. Tn3.7 Q): to Tn4 by insertion of a Tn3 transposon copy into the merP gene (Fig. Tn3.7 U); to Tn5086 by deletion of the In_Tn4 IS1326 copy to generate Tn2608 and replacement of the aadA cassette and acquisition of dfrA7 (Fig. Tn3.7 V); and to Tn2410 by replacement of the aadA cassette by an oxa cassette .

Fig. Tn3.7U. Relationship between Tn2411, Tn4 and Tn5060. Tn2411 carries a copy of In_Tn4 (right) while Tn4 also includes a copy of Tn3 inserted into merP.

Fig. Tn3.7V. Relationship between Tn2411, Tn2608 and Tn5086. Tn2411 (Top panel) carries a copy of In_Tn4 which includes a copy of IS1326 while this insertion is absent in In_Tn2608 (which carries the same integron cassette as In_Tn4) in Tn2608 (FN554766) (Second panel) and In22 in Tn5086 (CP054343) (Third panel). Moreover, in In22, the aad integron cassette of In_Tn2608 has been exchanged for a dfr cassette. It seems probable, from the DNA sequence (Fourth panel) that In_Tn2608 and In22 were derived by deletion from a structure similar to In_Tn4 because neither carry an IS1326 copy although they both retain the tip of the IRL (4 bp for In_Tn2608 and 3bp for In22) at one end and are missing 5bp of In_Tn4 DNA flanking the right IS1326 end. IS1326 IR are shown in red and contained within a blue horizontal arrow. Identical nucleotides are underlined.

The complete DNA sequences of many of these Tn are not available but Tn5086 or Tn2608 could be reconstructed from Tn21 using the limited sequence data in refeference . Moreover, using the reconstructed Tn5086 sequence in a BLAST search revealed an identical sequence in the E. coli SCU-164 chromosome (CP054343) and a nearly identical copy, in which the IRL had been interrupted by an insertion of IS4321, in E. coli plasmid pSCU-397-2 (CP054830) in addition to many closely related copies.

This analysis suggests that deletion of IS1326 had occurred by nearly-precise excision since the deletion junction observed in Tn5086 is not the original sequence identified in Tn2411. Indeed, the DNA sequences of Tn2411, Tn2608 and Tn5086, (Fig. Tn3.7 V) suggest that In_Tn2608 and In22 were derived by deletion from a structure similar to In_Tn4 because neither carry an IS1326 copy although they both retain the tip of the IRL (4 bp for In_Tn2608 and 3bp for In22) at one end and are missing 5bp of In_Tn4 DNA flanking the right IS1326 end.

Fig. Tn3.7W. Relationship between Tn21 and Tn1831. Tn1831 was generated from Tn21 by deletion mediated by IS1326::IS1353. Since no sequence is available for Tn1831, its restriction map (bottom) was mapped against the Tn21 restriction map derived from its DNA sequence (top).  The segment of DNA missing in Tn1831 compared to Tn21 is shown in red.

Tn21 was also proposed to give rise to a number of different transposons : to Tn1831 by IS1326-mediated deletion (IS1326 in IS1326::IS1353 is almost certainly functional) rightwards towards or past the IRt end of the integron while retaining the IS (Fig. Tn3.7Q and Fig. Tn3.7 W); to Tn2607 by insertion of Tn2601 (probably similar to Tn3) into the mer genes; to Tn2424 by insertion of IS161 to first generate Tn2425 and subsequent acquisition of two integron cassettes aacA1 and catB2 (Fig. Tn3.7Q and Fig. Tn3.7 X); and to Tn2603 by insertion of an oxaA1 cassette.

Fig. Tn3.7X. Tn21 to Tn2425 and Tn2424. The figure shows how Tn2424 is thought to be generated from Tn21 (Top) by insertion of IS161 to first generate Tn2425 (Bottom) and subsequent acquisition of two integron cassettes aacA1 and catB2 to generate Tn2424 (Middle). Since no sequence is available for Tn2424 or Tn2425, their restriction maps were mapped against the Tn21 restriction map (Top) derived from its DNA sequence. The relevant restriction fragments are shown as blue lines.

Tn1721 and (tandem) amplification of the tet genes

Tn1721 (Fig. Tn3.7 K) carries resistance to tetracycline (tet), is present on plasmid pRSD1 and is capable of undergoing amplification to generate tandem repeats . It was isolated by transposition to a lambda phage followed by a further transposition event onto plasmid R388 where it retained the ability to amplify .

Amplification was identified using restriction enzyme mapping (Fig. Tn3.7 Y) which showed a duplication of an EcoRI fragment and presumably occurs via replication slippage or unequal crossing over during replication between the full tnpA gene and the 5’-end tnpA segment at the right end of Tn1721. Indeed, amplification was shown to depend on the host recA gene .

Fig. Tn3.7Y. Tn1721. 1 amplification of tet. Tet module duplication following acquisition by Tn1722 of the tet gene module (Fig. Tn3.7K) .

The Tn163 Clade

There are 39 members of this clade (May 2021). Two (TnSku1 [Tn7197] (CP002358.1) and TnAmu_p (NC_015188.1) have acquired toxin/antitoxin gene pairs and most members (Fig. Tn3.8 A; Fig. Tn3.8 B) encode divergent tnpR and tnpA genes. There are a number of members without passenger genes as in the Tn21 clade (e.g. Tn6137, TnMex22[ Tn7165], TnMex38 [Tn7166], TnChe1, [Tn7155], TnAmu1 [Tn7138 ], TnAli20 [Tn7136], Tn6122, Tn3434).

Fig. Tn3.8A. The Tn163 Clade. The relationship between the transposases of different members of the Tn163 clade expanded from Fig.Tn3.4. Each transposon is shown together with its Genbank accession number. The small purple circles indicate that all members encode a tnpR resolution system. Those carrying a toxin/antitoxin gene pair (Bordeaux, orange and turquoise squares) are shown in bold type. Those Tn encircled by blue boxes are treated in detail in the text. The losenges on the outer circle indicate the presence and type of passenger gene carried by the corresponding Tn:  mer: mercury resistance operon; ars: arsenic resistance; red: antibiotic resistance; pale-yellow: other passenger genes; purple: hypothetical; white: no passenger genes; orange: toxin/antitoxin; black-line: not in TnCentral.

One small related group (Tn6137, Tn6136, Tn6134, Tn6138) all identified within the hexachlorocyclohexane-degrading bacterium Sphingobium japonicum UT26 genome show evidence at the DNA sequence level of several recombination events including acquisition of an sdr passenger gene and exchange of tnpR and tnpA by exchange at a location at which res should occur (Fig. Tn3.8 C).

Fig. Tn3.8B. The Tn163 Clade: Organisation of TnpA, TnpR and res. The configuration of the various genes is shown above each column, with the direction of expression shown by arrow heads. Where appropriate, registered transposon names are shown in brackets. Note that several Tn5393 derivatives are not included, since they have undergone extensive insertion and rearrangement.

Alignment against Tn6136 (Fig. Tn3.8 Ci) shows that Tn6137 carries the left half while Tn6134 carries the right section while Tn6137 carries the right while Tn6134 carries the left segments of Tn6138 (excluding the passenger gene insertion). Although the res sites have yet to be defined in detail, comparisons clearly show sequence divergence in this region (Fig. Tn3.8 Cii). Both Tn6134 and Tn6138 carry the same passenger gene (Fig. Tn3.8 Ciii) whose insertion has occurred proximal to IRL (Fig. Tn3.8 Civ).

Fig. Tn3.8C. Acquisition of passenger genes and inter-res site exchange in Tn6134 and relatives. i) The map of Tn6134 (AB610645.1) is shown with alignments of Tn6136 (AB610647.1), Tn6137 (AB610648.1) and Tn6138 (AB610649.1) above showing that Tn6134 and Tn6138 carry a sdr passenger gene while Tn6136 and Tn6137 do not. ii) The DNA sequences in the passenger gene region show that Tn6136 and Tn6137 have identical IRL sequences (blue filled box) while Tn6134 and Tn6138 (unfilled box) are also identical but have a number of SNPs compared to Tn6136 and Tn6137. Both passenger gene insertions are identical, with a 39 bp non-coding region downstream (shown in red) of the sdr gene (shown as a pink horizontal arrow). All four transposons have quasi identical sequences upstream of sdr but those of Tn6134/Tn6138 (red filled box) have a number of SNPs compared to Tn6136/Tn6137 (blue filled box).

Fig. Tn3.8C (continuation). Acquisition of passenger genes and inter-res site exchange in Tn6134 and relatives. iii) This shows an alignment which emphasizes the probable recombination at the res site. iv) DNA sequence around the probable res site (although res site sub-sites have not been identified). To the left Tn6134/Tn6138 (red filled box) are identical as are Tn6136/Tn6137 (blue filled box) but differ from Tn6134/Tn6138 by a number of SNPs. To the right, the Tn6134 and Tn6136 sequences are identical (blue filled box) as are those of Tn6137 and Tn6138 (red filled box) but differ from Tn6134 and Tn6136 by a number of SNPs. This indicates that recombination has occurred somewhere within the 29 identical base pairs in the open box.

The ancestor of another group of related transposons, the Tn5393 group (Fig. Tn3.8 D), appears to be Tn5393c (AY342395.1; Pseudomonas syringae pv. syringae plasmid pPSR1) which underwent an insertion of Tn5501.6 to generate Tn5393.1 (MF487840.1; Pseudomonas aeruginosa PA34), of IS1133 to generate Tn5393 (M95402; Erwinia amylovora plasmid pEa34) (Fig. Tn3.8 E) and of a complex set of mobile elements to generate Tn5393.4 (AJ627643; Alcaligenes faecalis).

Fig. Tn3.8D. The Tn5393 group: Organisation of TnpA, TnpR and res. Different members of this group are shown within red boxes on a pale red background. Insertions, deletions and rearrangements leading from one to the other are shown in blue boxes.

Tn5393 also gave rise to a number of other derivatives: Insertion of Tn3 into its transposase gene generated Tn5393.7 (LT827129; Escherichia coli strain K12 J53); insertion of Tn10 into IS1133 to generate Tn5393.2 (CP030921; Escherichia coli KL53 plasmid pKL53-M) (Fig. Tn3.8 F) followed by insertion of IS903 to generate Tn5393.11 (CP000602; Yersinia ruckeri YR71 plasmid pYR1); insertion of Tn10 in res to generate Tn5393.8 (CP002090; Salmonella enterica subsp. enterica plasmid pCS0010A).

Fig. Tn3.8E.  Detailed map of the relationship between Tn5393c, Tn5393 and Tn5393.1. Target Tn5393c sequences of the insertion of  Tn5501.6 to generate Tn5393.1 and of IS1133 to generate Tn5393 are shown in red.

There are also 4 examples carrying derivatives of Tn5 inserted into tnpA. They have an identical 3’ junction. In Tn5393.12 (KM409652; Escherichia coli REL5382 plasmid pB15), carries a complete Tn5. A second, Tn5393.13 (AB366441; Salmonella enterica subsp. enterica serovar Dublin plasmid pMAK2) is derived from Tn5393.12 by insertion of Tn2 into the IS1133 copy. In Tn5393.3 (LT985287; Escherichia coli strain RPC3 plasmid: RCS69_pI) the Tn5 insertion is a partial head-to-head Tn5 dimer, and in the other, Tn5393.10 (CP019905; Escherichia coli MDR_56 plasmid unnamed 6), insertion(s) and deletion(s) have occurred leaving only a partial Tn5 sequence. Finally, Tn5393 also gave rise to Tn5393.9 (KU987453; Klebsiella pneumoniae 05K0261 plasmid F5111) by multiple insertion including a type II intron, IS5708, ISCR1, ISEc28, ISEc29 and Tn2. A number of intermediate structures have yet to be identified but can probably be found in the large number of Tn5393 derivatives in the public databases. This group of Tn163 clade members have undergone a large number of modifications and constitute a broad network of related elements.

Fig. Tn3.8F. Detailed map of the relationship between Tn5393, Tn5393.7 and Tn5393.8. Target Tn5393 sequences of the insertion of  Tn10 to generate Tn5393.8 and of Tn3 to generate Tn5393.7 are shown in red.

The Tn4430 Clade

At present (May 2021) this clade is composed of only 11 examples (Fig. Tn3.9 A). One example, Tn4430 (X07651.1), encodes a tnpI gene and a res site with its associated organization but no passenger genes. The others encode a tnpR gene (Fig. Tn3.9 B). There are two small groups: Tn1546 which carry vancomycin resistance genes, and Tn6332 which carry mercury resistance genes.

Fig. Tn3.9A. The Tn4430 Clade. The relationship between the transposases of different members of the Tn4430 clade expanded from Fig.Tn3.4. Each transposon is shown together with its Genbank accession number. The small purple circles indicate that all members encode a tnpR  resolution system. The lozenges on the outer circle indicate the presence and type of passenger  gene carried by the corresponding Tn:  mer: mercury resistance operon; ars: arsenic resistance; red: antibiotic resistance; pale-yellow: other passenger genes; purple: hypothetical; white: no passenger genes; black-line: not in TnCentral.

ig. Tn3.9B. The Tn4430 Clade: Organisation of TnpA, TnpR and res. The configuration of the various genes is shown above each column, with the direction of expression shown by arrow heads. Where appropriate, registered transposon names are shown in brackets.

The Tn1564 Vancomycin Resistance Group

Resistance to Vancomycin in Enterococci appeared in 1988 , was shown to be transmissible and carried by a transposon, Tn1546 (M97297.1) . The relationship within the Tn1546 vancomycin resistant transposons is relatively simple and the result of insertions/deletions mediated by several different insertion sequences: Tn1546.2 (AB247327) is derived from Tn1546 by insertion of IS1216E between vanYA and vanXA and Tn1546.1_p (KR349520.1) appears to be derived from Tn1546.2 by insertion of IS1251 between vanHA and vanSA and a neighboring deletion to the right of IS1216E bringing vanYA and vanXA closer to each other. Other examples identified in surveys of vancomycin-resistant Enterococci from human and other animal sources also include insertions of ISEf1, IS1542 and IS19 , in addition to a number of other IS1216 insertions (often in multiple copies and accompanied by neighboring deletions) . A number of these insertion/deletion derivatives have been identified from several sources and different geographical locations (Fig. Tn3.9 C).

Fig. Tn3.9C.  Examples of  Tn1546 derivatives showing the position of various insertions. A map of Tn1546 without insertions is shown below, with the vancomycin resistance genes in red. Data from .

The Mercury Resistance Group

Within the mercury resistance group (Tn6294-LC015492.1, Tn5084-AB066362.1, Tn6332-LC155216.1 and TnMERI1-LC152290 – note that we have reconstituted the left end by comparison with Y08064; Fig. Tn3.9 D), the mercury resistance genes are expressed to the left while TnpR and TnpA are expressed to the right. All four carry additional copies of merB and merR. Huang et al have shown that expression of the mercury resistance genes of TnMERI1 is driven by three promoters (Fig. Tn3.9 E). Comparison with Tn6294 suggests that the mercury gene set has been exchanged by recombination at the level of the res site (Fig. Tn3.9 D). The sequences of two closely related members of the same group, Tn5083 and Tn5085, are incomplete .

Fig. Tn3.9D. The Tn4430 Clade: Tn6332 group i) Alignment against Tn6294 to illustrate probable inter transposon recombination at the res site. ii) Alignment against Tn6332.

Fig. Tn3.9E. Tn6332 group Expression of Mercury Resistance. A map of TnMERI1_p showing the position of three operator/promoter sequences identified by . These are shown as blue vertical lines, with the horizontal lines indicating the direction of expression. Although this transposon is only partial because it lacks at least part of the left end, it is similar to Tn6332. 

The Tn3 Clade

This clade includes the classical Tn1, 2 and 3 (see Historical) as well as Tn1000. There are 29 examples of the Tn3 clade (of which 26 can be found in TnCentral) (Fig. Tn3.10 A) which fall into two subgroups. The majority have divergently expressed tnpR and tnpA and most carry passenger genes (Fig. Tn3.10 B). The res sites of each sub-group show significant similarity (Fig. Tn3.10 C). A number carry toxin-antitoxin genes (TA) generally located between the divergent tnpR and tnpA. These are of two types (Fig. Tn3.10 A) and appear to be specific for each subgroup. Passenger genes can be located upstream of downstream of the tnpR/tnpA transposition module (Fig. Tn3.10 B). All except two encode tnpR type resolvases. The two which do not, TnBth4 and Tn5401, also encode a TA module.

Fig. Tn3.10A. The Tn3 Clade. The relationship between the transposases of different members of the Tn3 clade expanded from Fig.Tn3.4. Each transposon is shown together with its Genbank accession number. The small purple circles indicate that all members encode a tnpR resolution system. Members with toxin/antitoxin gene pairs are shown in bold type. The lozenges on the outer circle indicate the presence and type of passenger gene carried by the corresponding Tn: mer: mercury resistance operon; small-red: intron-associated antibiotic resistance; large-red: non-intron associated antibiotic resistance; pale-yellow: other passenger genes; yellow: plant pathogenicity genes; orange: toxin/antitoxinf gene pairs; white:no passenger genes

Fig. Tn3.10B. The Tn3 Clade: Organisation of TnpA, TnpR and res. The configuration of the various genes is shown above each column, with the direction of expression shown by arrow heads. Where appropriate, registered transposon names are shown in brackets.

Fig. Tn3.10C. Res site alignment of Tn3 clade members with divergent tnpR/tnpA genes. The Tn names are shown to the left of the figure. Sequence conservation is shown by the depth of the blue background. Sites I, II and III are indicated. Top: Tn3-like without passenger genes between tnpR and tnpA. Bottom: TnXc4-like with TA passengers between tnpR and tnpA

Importance of ISEcp1 in bla CTX-M-expression

There are examples of members of the Tn3 clade which carry insertions of ISEcp1-like sequences (see: IS1380 family) closely upstream of a bla-CTX-M gene. Indeed, upstream insertion of ISEcp1 derivatives have been identified associated with a number of different bla-CTX-M variants in both Tn3 and other groups . In some examples, this is limited to an isolated right end which is responsible for expression of the bla-CTX-M gene by providing a mobile promoter .

The Tn3 group

Tn3, Tn1, Tn1MER, Tn2, Tn2.1 and Tn3.1. all carry a probable internal IR upstream of the bla gene (Fig. Tn3.10 D) which acts as a hotspot for IS231A insertion and was initially observed in the bla gene of plasmid pBR322 . Tn2 and Tn2.1 are identical except for the ISEcp1 insertion which also carries an internal IS1 insertion (Fig. Tn3.10 E). Note that an ISEcp1 promoter drives bla CTX-M-expression. There are a number of closely related derivatives (e.g. Tn6339-MF344565) in which the IS1 copy appears to have been involved in small rearrangements of the ISEcp1 copy while maintaining the ISEcp1 promoter. Three examples carry a number of integron cassettes without either the integrase gene, the Tn402 ends or the Tn402 transposition genes that are often associated with integrons in the Tn21 clade.

Fig. Tn3.10D. Tn3 showing the potential internal IR. (Top). Potential internal IR sequence. (Bottom left) Map of Tn3. (Bottom right) plasmid pBR322. The position of the potential internal IR, a hotspot for Tn4430 insertion, is shown as a blue arrow. From .

Fig. Tn3.10E. Tn3 group alignment against Tn3. This shows the variation in tnpR, the res site and the 5’ end of the tnpA gene

Inspection of the alignment (Fig. Tn3.10 E) shows that apart from insertion of different mobile elements, the major sequence variations occur in the region of the res sites, the 5’ ends of tnpA and tnpR as had been previously noted for Tn1, 2 and 3 (for res, see Fig. Tn3.10 C) and an evolutionary pathway involving a combination of homologous and resolvase-mediated recombination has been proposed.

This can be detected by the distribution of SNPs on each side of the res site (e.g. Tn1331 and Tn1332). In this respect, the integron carrying Tn6238 is more similar to Tn3 while Tn1MER, Tn1331, and Tn1332 are more similar to Tn1 and Tn2.1 resembles Tn2.

The Xanthomonas group

This group except for TnPsy39 (Tn7187), all members of this group in the tree carry the same TA pair and the passenger genes are located to the right of the transposition module. The Xanthomonas transposon cluster (Fig. Tn3.10 F) are closely related and differ essentially by insertion of ISXac1 and ISXac5 (Fig. Tn3.10 G) as well as deletions (in particular of the res site in TnXc4.2 [Tn7212]). TnXc4.1 [Tn7211], although having an organisation identical to that of TnXc4 [Tn7210] has undergone significant sequence divergence along its entire length. TnThsp9 [Tn7202] also shows sequence variation within the region carrying transposition and TA functions (but includes mercury genes instead of plant pathogenicity functions while TnPsy39 [Tn7187] only exhibits similarity in the TnpA gene.

Fig. Tn3.10F. Variation in the Xanthomonas Tn3 clade transposons. An alignment against TnXc4 (Tn7210). 

Fig. Tn3.10G. Relationship between Xanthomonas Tn3 derivative transposons. The figure shows that sequential IS insertions are involved in the decay of the TnXc4 (Tn7210) secC gene.

All members of the second cluster, which encode for the same TA gene pair as the Tn3 group (Fig. Tn3.33A), also carry mercury resistance genes although these have undergone some rearrangements and sequence divergence (Fig. Tn3.10 H) and are also divergent from those present in TnThsp9 (Tn7202).

The Tn3000 Clade

This clade is composed of nearly 30 members (25 in TnCentral) all of which encode TnpR resolvases and carry tnpR-related res sites. Most also encode TA gene pairs and these are of three types (Fig. Tn3.11 A).

Fig. Tn3.11A. The Tn3000 Clade. The relationship between the transposases of different members of the Tn3000 clade expanded from Fig.Tn3.4. Each transposon is shown together with its Genbank accession number. The small purple circles indicate that all members encode a tnpR resolution system. Members with toxin/antitoxin gene pairs are shown in bold type. The lozenges on the outer circle indicate the presence and type of passenger gene carried by the corresponding Tn: red: non-intron associated antibiotic resistance; pale-yellow:other passenger genes; yellow: plant pathogenicity genes; orange: toxin/antitoxin gene pairs; white: no passenger genes; purple: hypothetical.

The Tn5501 cluster.

There are a number of Tn5501 examples (Fig. Tn3.11 B). All have their passenger genes located upstream of the transposition module and all except TnPysy42 [Tn7188] and Tn5501.12 encode the same parE/parD TA genes (Fig. Tn3.11 A). Tn5501.12 appears to have acquired different TA genes (HTH_37, GP49) by recombination at the res site (Fig. Tn3.11 C).

Fig. Tn3.11B. The Tn3000 Clade: Organisation of TnpA, TnpR and res. The configuration of the various genes is shown above each column, with the direction of expression shown by arrow heads. Where appropriate, registered transposon names are shown in brackets.

The relationship between members of the cluster is shown in Fig. Tn3.11 C. Most have retained the same transposition and TA modules but vary in the type of passenger genes they carry. They all carry deletions with respect to Tn5051.3. For 8 of these, the right junctions of the deletions are close but not identical (Fig. Tn3.11 Di and Dii). All leave the TA module intact. In only one example, the toxin gene has undergone deletion leaving the antitoxin intact (Fig. Tn3.11 Diii). The left junction is less clear and difficult to interpret. A number of Tn5501 derivatives are related by IS insertions and deletion (Fig. Tn3.11 E).

Fig. Tn3.11C. The Tn5501 cluster, showing the acquisition of different set of passenger genes, including toxin-antitoxin gene pairs.

Finally, a small group of Tns which, like Tn5501.12, all carry the HTH_37/GP49 TA pair is shown in Fig. Tn3.11 F. It appears that there has been an exchange between a Tn5501.5-like transposon and a derivative of Tn4662a (lacking the ISAs20 insertion) by recombination at the res site to generate Tn5501.12.

Fig. Tn3.11D. The Tn5501 cluster: right deletion junction The DNA sequence of the deletion junctions compared to Tn5501.3. i) Junction of Tn5501 ii) Tn5501.7, Tn5501.9, Tn5501.6, Tn5501.10, Tn5501.1, Tn5501.8, Tn5501.4, Tn5501.11. iii) toxin deletion endpoint in Tn5501.5.

Fig. Tn3.11E. The Tn5501 derivatives with IS insertions.

Fig. Tn3.11F. The Tn4662 cluster.

Clinical Importance of Tn4401

In the past decades, carbapenemase-producing Enterobacteriaceae (CPE) have appeared that are resistant to most or all clinically available antibiotics, including carbapenems, which are often considered the antibiotics of last resort . The 10kb transposon, Tn4401 has been instrumental in the spread of the carbapenem resistance gene bla_KPC. It was described in 2008 in a number of clinical isolates of Klebsiella pneumoniae and Pseudomonas aeruginosa from the United States, Colombia and Greece .

Members of this small group have divergently expressed tnpR and tnpA genes located towards the left end and bla_KPC towards the right end downstream from tnpA (Fig. Tn3.11 B) flanked by two different insertion sequences, ISKpn6 and ISKpn7 (Fig. Tn3.11 G). The ISKpn7 insertion had occurred within an additional Tn4401 IR. It was further observed that there were two “isoforms” of Tn4401: Tn4401a and Tn4401b. Tn4401a, isolated in the United States and Greece carried a 100bp deletion upstream of the bla gene compared to Tn4401b from Colombia. The Tn4401 backbone appears to have undergone a number of recombination events. A third derivative, Tn4401c , was found to carry a deletion of about 200 bp upstream of bla while in a fourth, Tn4401d , the ISKpn7 copy along with flanking DNA has undergone deletion to leave a 3’ segment of bla_KPC and a 5’ segment of tnpA and therefore would not be capable of autonomous transposition.

Fig. Tn3.11G. Tn4401 an Important Vector in the Spread of blaKCP (Top) A map of Tn4401 indicating the large deletion carried by Tn4401d. The region representing the DNA sequences below is circled. (Bottom) Nucleotide sequences located upstream of blaKCP show the location of three potential promoters. The transcription start sites (+1), the ribosome binding site (RBS), and the translation start site (ATG) are also indicated. The IRR end of ISKpn7 is shown in yellow, and the disrupted Tn4401 IR is shown in pink. ISKpn7 insertion creates a hybrid promoter using a -35 box in ISKpn7 and a -10 sequence in the disrupted Tn4401 IR.

Furthermore, analysis of a number of clinical isolates from different regions of the United States which exhibited various levels of carbapenem resistance, revealed deletions of different extent in the region upstream of bla_KPC . Closer analysis using RACE (Rapid amplification of cDNA ends) to locate transcriptional start points revealed 3 (possibly 4) promoters, one of which had been generated from the -35 element located in the IR of the inserted ISKpn7 (as is characteristic for a member of the IS21 family (see IS21 chapter; formation of hybrid promoters figure IS21.1).

The Tn4651 Clade

Fig. Tn3.12A. The Tn4651 Clade. The relationship between the transposases of different members of the Tn4651 clade expanded from Fig.Tn3.4. Each transposon is shown together with its Genbank accession number. The small purple circles indicate that all members encode a tnpR resolution system. The lozenges on the outer circle indicate the presence and type of passenger gene carried by the corresponding Tn.

The Tn4651 mix of radically different structures

This Tn3 family clade (Fig. Tn3.12 A) contains members with very diverse structures (Fig. Tn3.12 B). They fall into three major clusters. Two encode the tnpT/S/rst while the third encodes the tnpR/res system.

Fig. Tn3.12B. The Tn4651 Clade: Organisation of TnpA, TnpR and res. The configuration of the various genes is shown above each column, with the direction of expression shown by arrow heads. Where appropriate, registered transposon names are shown in brackets.

The tnpT/S/rst clusters

In the first tnpT/S/rst cluster, mostly from the plant pathogen Xanthomonas (Fig. Tn3.12 C), TnXax1.1 [Tn7207] appears to have undergone res-recombination in which the upstream passenger genes and tnpT have been exchanged. TnpT is significantly different from the other four. TnXax1.3 [Tn7209] differs from the others (TnXax1 [Tn7206]; TnXax1.2 [Tn7208]; TnXax1.3 [Tn7209] in the 3’ region of tnpA and there is some variation in tnpS and tnpT.

Fig. Tn3.12C. The Tn4651 Clade: Organisation of TnpA, TnpR, and res of the TnXax1 group. Alignment of different TnXax1 derivatives against TnXax1.3. The blue boxes indicate significant sequence variation between TnXax1.3 and the other derivatives.

TnXax1 derivatives are generally vehicles for pathogenicity genes such as Transcriptional Activator Like Effectors (TALE genes), lytic transglycosilases (mtlB2) and genes (xop) involved in type III secretion system (TTSS) translocation of effector proteins into host plant cells (Fig. Tn3.12 C). TnXax1 derivatives can include IR which are significantly longer (72/92 bp) than the 38-40bp characteristic of the Tn3 family (Fig. Tn3.12 D) although the functional significance of this has not been investigated. The IR also terminate in a GAGGG pentanucleotide. The left end of group members is quite variable (Fig. Tn3.12 E) while their right ends appear more homogeneous (Fig. Tn3.12 F).

Fig. Tn3.12D. i) Genetic organization of the canonical TnXax1 from X. citri strain 306 plasmid pXAC64. Genes are indicated by colored boxes, with the direction of transcription shown by the arrow heads. Transposition-related genes are shown in purple, passenger genes in yellow. The presumed resolution site is located between the tnpT and tnpS genes, as observed in the transposable element Tn4651. This includes two palindromes: IR1 and IR2 are probably part of the core site at which recombination occurs, recognized by the TnpS recombinase whereas IRa and IRb are potential binding sites for TnpT. The terminal inverted repeats (IRL and IRR) are shown as grey triangles. The convention used for orientation of the transposon is that the transposase (TnpA) is transcribed from left to right. ii) Sequence of the long IRs; iii) IRs identified from other Tn3-like Transposable Elements. All include the GAGGG tips, sharing sequence similarities to TnXax1 IRs. These are: TnPa43 from Pseudomonas aeruginosa; TnStma1 from Stenotrophomonas maltophilia strain D457; TnTin1 from Thiomonas intermedia strain K12 and TnXca1 from X. campestris pv. vesicatoria plasmid pXCV183.

Fig. Tn3.12E.  Genetic organization of the Left end of the TnXax1 related structures found in other Xanthomonas species. Abbreviations of the TnXax1 related structures: (1) X. fuscans subsp. aurantifolii: XauB ctg 621_0 (2) X. axonopodis strain 29-1, chromosome copy: XccA-29-1 (3) X. citri subsp. citri strain Aw 12879  chromosome copy: XccA(w) (4) TnXax1 (5) X. axonopodis strain 29-1, chromosome copy: XccA-29-1 (7) plasmid copy: XccA-29-1 pXac64 (8) X. campestris pv. vesicatoria 85-10, chromosomal copy: Xcv B (9) X. axonopodis pv. citrumelo F1, chromosomal copy: Xac F1 A (10) X. axonopodis pv. citrumelo F1, chromosomal copy Xac F1 B (11) X. campestris pv. vesicatoria 85-10, chromosomal copy: XcvA (12) X. arboricola pv. pruni plasmid pXap41: Xap pXap41 (13) X. fuscans subsp. aurantifolii: XauC ctg 1147_0.

Fig. Tn3.12F.  Genetic organization of the Right end of the TnXax1 related structures found in other Xanthomonas species. Abbreviations as for Fig. Tn3.12E.

The second tnpT/S/rst cluster is characterized by Tn4651, a toluene-catabolic transposon identified in from Pseudomonas putida plasmid pWW0 . In addition to the tnpS/T resolution system, it encodes an additional small transposition-related gene, tnpC which impacts cointegrate formation. Using, Tn4652, a Tn4651 deletion derivative lacking the toluene-catabolic genes , TnpC was shown to regulate TnpA expression post-transcriptionally . Moreover, the host protein IHF binds to sites in both Tn4652 ends (Fig. Tn3.12 G) . These overlap the region protected by TnpA binding and binding positively regulates both tnpA transcription and TnpA binding to the terminal IRs. Indeed, transposase binding to the IRs in vitro was shown to occur only after binding of IHF . TnpA protects an extensive region encompassing the IRs and 8-9 bp of flanking DNA (Fig. Tn3.12 G). Tn4652 transposition appears to be elevated in stationary phase, involves the stationary phase sigma factor, sigma S , and is limited by the levels of IHF whose level is increased in stationary phase. Another DNA chaperone host factor, FIS, has a negative effect on transposition, apparently by competing for IHF binding .

Fig. Tn3.12G. Integration host Factor (IHF) and the Ends of Tn4652  (Top) Map of Tn4652. (Bottom) Left and right Tn4652 ends. The position of the IHF binding sites at each end of Tn4652 is shown in red. The bracketed regions indicate the extent of IHF protection. The IR are indicated within the blue boxes .

IHF and FIS have been implicated in other transposition systems such as IS10 (see: IS4 family). Moreover, Tn1000 (Tn3 clade) carries an IHF binding site proximal to each IR which acts copoperatively to increase TnpA binding and immunity . One additional interface with host physiology is the observation that the CorR/CorS two component system regulates transposition positively .

Other members of the cluster include: Pseudomonas sp. mercury resistance transposon Tn5041  ; Tn4676, a long (72,752bp) and complex Pseudomonas resinovorans carbazole-catabolic transposon from plasmid pCAR1 ; and Tn4661, a Pseudomonas aeruginosa cryptic transposon . All include tnpA, tnpC and the tnpS/T resolution system.

Tn5041 transposition has also been addressed experimentally and was observed to be host-dependent : it occurred in the original Pseudomonas sp. KHP41 host but not in P. aeruginosa PAO-R or in Escherichia coli K12. Interestingly, transposition in these strains was found to be complemented by the Tn4651 transposase gene (tnpA) and the region which determines this host dependence was mapped to a 5’ tnpA gene segment by construction of hybrid Tn5041-Tn4651 tnpA genes. Tn5041 apparently acquired its mer operon from a derivative of Tn21 or Tn501 . It is reported to be preceded by a 24 bp element with 75% sequence similarity to the outermost part of IRs typical for Tn21-like transposons.

The Tn1071 Clade

The Tn1071 group.

Members of this small group are often associated with xenobiotic catabolism and other “exotic” functions (Fig. Tn3.13 A).

Fig. Tn3.13A. The Tn1071 Clade.

Tn1071 itself (Fig. Tn3.13 Bi), the founding member, was identified as part of a compound transposon, Tn5271, in Comamonas testosteroni where it flanks a chlorobenzoate catabolic operon in . It is unusual since it carries only tnpA and not tnpR, has unusually long (110bp) IR (Fig. Tn3.13 Bii) and was first described as IS1071. Two other members of this small group, IS882 from Ralstonia eutropha H16 megaplasmid pHG1 encoding key enzymes for H2-based lithoautotrophy and anaerobiosis and ISBusp1 (aka ISBmu13; NC_007509.1) from the Burkholderia multivorans ATCC 17616 genome , were also originally identified as IS. Their structure fits the definition of an IS since they all contain a single transposase open reading frame located between two IR.

Fig. Tn3.13B. Tn1071 organisation. (Top panel) The structure of Tn1071 (M65135) showing the single tnpA reading frame and the long terminal IRs. (Bottom panel) Sequence of the long IRs. Boxed sequence indicate the length of normal Tn3 family IRs. Bold and underlined bases indicate identity, red bases indicate non identity.

A limited functional analysis of Tn1071 transposition is available . It was only able to transpose at high frequencies in two environmental β-proteobacteria Comamonas testosteroni and Delftia acidovorans but not in Agrobacterium tumefaciens (α-proteobacteria) or Escherichia coli, Pseudomonas alcaligenes and Pseudomonas putida (all γ-proteobacteria). These studies showed that Tn1071 generates cointegrates as a final transposition product since it has no resolution functions, produces 5bp DR on insertion and requires the entire 110bp IRs for activity. This is therefore in contrast to many other Tn3 family members which only require the 38 bp IR.

The absence of a resolution system implies that, like IS26(see: IS6 family) , Tn1071 probably forms “pseudo-compound transposons” . In these structures the flanking Tn1071 copies must be in direct orientation as a consequence of the homologous recombination event required to resolve the cointegrates structure. Transposition is initiated by one of the flanking IS to generate a cointegrates structure with three Tn1071 copies (similar to those generated by the IS6 family of insertion sequences; Fig. IS6.8B). “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 as for IS6 family members 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” leaving a single Tn1071 copy in the donor plasmid. However, it is possible that both Tn1071 copies are used in transposition in which case the cointegrates would be expected to contain two directly repeated copies of the entire transposon sat the donor/target junctions.

A significant number of Tn1071-associated xenobiotic-degrading genes on many catabolic plasmids have been documented by population-based PCR and genetic studies .

Tn5271 itself is widely distributed in bacteria isolated from a large ground water bioremediation site and plasmid derivatives carrying the transposon together with a third Tn1071 copy in an inverted orientation were also identified. The interstitial DNA segment between the old and new copy in these derivatives was also inverted as expected from intra-molecular transposition events (Fig. for intramol transposition).

A number of additional potential compound transposons have been identified although these may be inactive: a >28kb transposon, Tn5330 (AF029344), from Delftia acidicorans carries the entire 2,4-dichlorophenoxyacetic acid degradation pathway and, although the sequence data for the flanking Tn1071 copies is not complete, both carry inactivating insertions of IS1471; a similar ~48 kb transposon (NC_005793) with 5bp flanking DR from Achromobacter xylosoxidans plasmid, pEST4011, also carries identical IS1471 inactivating insertions in each flanking Tn1071 copy and a 7kb internal tandem duplication compared to the Delftia acidovorans transposon.

When analyzed in more detail, these genes are sometimes flanked by Tn1071 copies in direct repeat as in the original Tn5271 but are found in more complex Tn1071-based structures.

TnHad2 (Fig. Tn3.13 Ci), for example, from a Delftia acidovorans haloacetate-catabolic plasmid, pUO1, carries a nested copy of a potential Tn1071-based compound transposon, TnHad1 which does not carry flanking DR. TnHad1 is inserted into a larger structure, TnHad2 with flanking 5bp DR, typical Tn3 family ends related to those of Tn21 but no apparent dedicated transposase except that of the Tn1071 copies.

The authors state that TnHad2 was unable to transpose as judged by a “mating out” assay using the plasmid R388 as a target. However, The TnHad2 Tn21-like IRs were found to be active in transposition if supplied with Tn21 but not with Tn1722 transposition functions . TnHad2 also appeared to carry a functional res site.

Tn1071 also flanks atrazine degrading genes in plasmid Pseudomonas pADP-1 (U66917) in a structure with three directly repeated Tn1071 copies intercalated with three copies of an IS91 family member, ISPps1. These are apparently generated by duplication events since regions with identical sequence stretch from the oriIS end of ISPps1 through Tn1071 and terminate just before the atz genes (Fig. Tn3.13 Cii). The repeated regions also includes the DR sequences at each Tn1071 except for that at the far right.

Fig. Tn3.13C. Tn1071 Involvement in Compound transposons. i) TnHad1, a composite transposon inserted into TnHad2 . ii) Amplified Tn1071 copies flanking atrazine degrading genes

There are a number examples of other structures with multiple Tn1071 copies and in a large proportion of these cases, the multiple copies occur in direct repeat. They are associated with plasmids which degrade the phenylurea herbicide linuron e.g. pBPS33-2 (CP044551) and have been isolated from a variety of bacteria with the capacity to degrade a wide range of chlorinated aromatics and pesticides or p-toluene sulfonate (PTSA) where they flank the PTSA genes in plasmid pTSA (AH010657) .

MITES, MICs and TALES

Many TE families also include non-autonomous transposable derivatives with no transposition related genes. These are simple and composed of two correctly oriented ends with or without an intervening passenger gene and are called MICs (Minimal Insertion Cassette) and MITEs (Miniature inverted-repeat transposable elements) respectively. For Tn3, related MITEs are known as TIMEs (Tn3-Derived Inverted-Repeat Miniature Elements) .

Fig. Tn3.14A. Map of Xanthomonas plasmid pXAC64 This identifies a number of Tn3 family transposable elements including TnXc4 (Tn7210) and TnXax1 (Tn7206) and a TAL-carrying MITE. Our standard labelling convention and colour coding is used.

Studies have shown that Xanthomonas genomes are often havens for MICs carrying genes involved in pathogenicity towards their host plants . A number of Tn3 family structures were identified in a conjugative plasmid, pXac64 (CP024030), of the principal pathogen of citrus trees, Xanthomonas citri, an important economic problem (e.g., reference ) (Fig. Tn3.14 A). The plasmid includes two Tn3 family transposons, TnXc4 (Tn7210) and TnXac1.4 (Tn7206) and a MIC (MIC XAC64.T1; 3948bp) which carries a TAL effector gene. Other TAL effector-carrying MICs can be identified in other Xanthomonas plasmids such as pXac33 (CP008996) (two TAL-carrying MIC: MIC XAC33.T1, 3739bp, and MIC XAC33.T2, 3538bp; and the Tn3 family transposon TnXc5) and from the Xanthomonas fuscans plasmid pplc XAF (FO681497) (a single MIC, MIC XAF.T1 ,3768bp, and a 10kb MIC with a number of virulence genes. Some MICs, e.g. MIC XAC33.T1 (Fig. Tn3.14 B right), are flanked by 5bp DR, a hallmark of Tn3 family transposition.

A global analysis of TAL effector genes in (Fig. Tn3.14 C) within the Xanthomonas genus (available in 2014) identified a large number which were flanked by Tn3-like IR although a some carried a single identifiable IR while others failed to exhibit clear IRs .

Fig. Tn3.14B. Plasmids pXAC33 and plc XAF This figure shows additional examples of both transposons and MICs.

Fig. Tn3.14C. Molecular Phylogenetic analysis by Maximum Likelihood method of the TALEs genes found in completely sequenced Xanthomonas genomes. Each TALEs gene is shown with its respective genomic coordinates (for TALEs extracted from complete genomes) or TALEs gene Genbank Accession Number inside the brackets. Red diamonds indicate TALEs genes associated with MIC structures, yellow triangles TALEs genes associated with a solo IR flanking one of their extremities, blue circles genomes which carry TnXax1 or related structures and asterisks indicate MICs with the direct TACTC(G) target repeat . The blue arrow indicates the TALEs gene from X. fuscans subsp. fuscans strain 4834-R plasmid pla and plc. TALEs from different Xo pathovars are highlighted as follows: XooP, blue; XooK, green; XooM, purple; and XocB, pink. TA TALEsLEs from X. citri, X. axonopodi, X. vesicatoria and X. campestris are highlighted in gray .

• Fig. Tn3.14. Panels D1,2 and 3 (use the arrows to scroll the figures).
• 

  Fig. Tn3.14D1. List of MICs present in the Genomes of X. oryzae PXO99A, MAFF and KACC For each strain, the columns show the relative positions of each MIC along the respective genome from the point of origin together with the type of MIC. Those carrying genes other than TAL effectors are marked as such, with their passenger gene noted. The Tale containing MICs are marked MIC P.T, MIC P.M and MIC P.K for PXO99A, MAFF and KACC respectively. The coloured boxes indicate MICs that are identical from one strain to another.

• 

  Fig. Tn3.14D2 (continuation). List of MICs present in the Genomes of X. oryzae PXO99A, MAFF and KACC For each strain, the columns show the relative positions of each MIC along the respective genome from the point of origin together with the type of MIC. Those carrying genes other than TAL effectors are marked as such, with their passenger gene noted. The Tale containing MICs are marked MIC P.T, MIC P.M and MIC P.K for PXO99A, MAFF and KACC respectively. The coloured boxes indicate MICs that are identical from one strain to another.

• 

  Fig. Tn3.14D3 (continuation). List of MICs present in the Genomes of X. oryzae PXO99A, MAFF and KACC For each strain, the columns show the relative positions of each MIC along the respective genome from the point of origin together with the type of MIC. Those carrying genes other than TAL effectors are marked as such, with their passenger gene noted. The Tale containing MICs are marked MIC P.T, MIC P.M and MIC P.K for PXO99A, MAFF and KACC respectively. The coloured boxes indicate MICs that are identical from one strain to another.

Fig. Tn3.14E. MIC clusters in PX099 Panels i) and ii) show two tandem repeats at different genome positions which share one identical MIC. These have presumably arisen by transposition and diversification of the second MIC which is missing the right-hand IR. In panels v) and vi) we have not been able to identify IRs. Note that in i), ii), iii), iv), vii) and viii) the interstitial DNA sequence between each MIC is identical.

Inspection showed that the chromosome of Xanthomonas citri strains do not carry identifiable TAL-carrying MICs but those of X. oryzae carry relatively high numbers . A smaller number of MICs carrying other pathogenicity-related genes are also observed (e.g. Type III Xop genes) It is notable that the majority of the TAL-associated MICs occur as two or more tandem copies. These are listed for three example genomes, X. oryzae PXO99A, MAFF and KACC in Fig. Tn3.14 D.1-3.

Fig. Tn3.14F. A Model for generating variability in the number of TALE genes repeats: Expansion and Contraction by Replication Slippage Replication forks traversing a MIC are shown. The TAL repeat sequences are shown in different shades of green, the IRs in either red or green boxes. These structures could arise from normal replication or during the Tn3-family transposition process. The intermediate structure shows leading strand synthesis (bottom branch of the fork) and lagging strand synthesis (top branch of the fork); replication fork slippage on the leading strand template (left panel) results in removal of repeat sequences, while slippage on the nascent strand (right panel) results in expansion of repeat sequences. These are resolved in a second round of replication.

Those where no IRs could be detected at either end are shown simply as open reading frames. In each case, the DNA segment between tandemly repeated MICs is identical (Fig. Tn3.14 E), suggesting that the tandem dimers and multimers arose by amplification possibly via replication slippage and unequal crossing over (Fig. Tn3.14 F). Another characteristic is that they are often flanked by transposase genes raising the possibility that their appearance at different chromosome locations (“radiation”) has occurred by transposition of a single ancestral MIC. This might have been mediated either by flanking transposable elements or by complementation from a Tn3 family transposase. In many cases, one of the terminal MICs is truncated and does not exhibit an IR and could often be attributed to insertion of an IS.

Fig. Tn3.14G. MIC Duplication and Diversification The figure shows an alignment of the MIC clusters i) and ii) from Fig. Tn3.14E clearly indicating the occurrence of a deletion in i) compared to ii) or an insertion of repeats in ii) compared to i).

It is clear that this “radiation” of TAL-associated MICs does not only occur by transposition. In one case (Fig. Tn3.14 Ei, Eii and Fig. Tn3.14 G) an entire DNA segment containing a tandem MIC dimer (MIC P.T11-MIC P.T13) appears to have been translocated together with surrounding genomic sequences with MIC P.T13 undergoing deletion to generate MIC P.T11-MIC P.T12.

Fig. Tn3.14H. TALE protein Architecture. (Top) Structure of a 34 amino acid repeat. (Bottom) Organisation of a TAL protein showing the N- and C-terminal domains separated by a series of  peptide repeats of different colours, each with its recognition dipeptide. Below is shown the DNA base recognition code. From Amanda Zucoloto   Group: iGEM13_Calgary (2013-09-17).

This variability in MIC sequence can be observed within the longer arrays (e.g. Fig. Tn3.14 E viii) suggesting that diversification follows amplification. This is due to changes in the TAL genes. TAL proteins are composed of conserved N-terminaland C-terminal regions separated by a variable number of 34 amino acid repeats (Fig. Tn3.14H) which can number between 1.5 and 35.5 tandem copies. Each repeat includes a pair of adjacent amino acids capable of recognizing a single base in a DNA sequence (Fig. Tn3.14 H; Fig. Tn3.14 I) e.g. . A tandem array of repeats therefore enables the TAL protein to recognize specific sequences within the target plant genome. This is illustrated by the TAL effector carried by MIC P.T14 (Fig. Tn3.14 J) which includes 19.5 such repeats. The TAL effectors encoded by other members of this cluster (Fig. Tn3.14 Eviii), MIC P.T15, MIC P.T16, MIC P.T17, MIC P.T18, which have presumably all arisen by amplification of a single ancestral MIC, each carry a different number of repeats and vary in their sequence recognition properties.

• Fig. Tn3.14. Panels I and K (use the arrows to scroll the figures).
• 

  Fig. Tn3.14I. TALE Protein Specificity. Left column show the recognition dipeptide. The right column shows the nucleotide which is recognised by the amino acid pair. Data from .

• 

  Fig. Tn3.14J. Amino acid Sequence of the TAL expressed from MIC P.T14 This shows the 19.5 34 amino acid repeats together with the recognition dipeptides (boxed).

• 

  Fig. Tn3.14Ki. Comparison of the MIC P.T14 Clusters in  PXO99A and MAFF Strains The Sequence of the 34 amino acid repeats of each TAL protein in consecutive MICs is shown in equivalent positions in the clusters. Recognition dipeptides are boxed. Right: PXO99A, left, MAFF. i) The recognition sequences are presented between the two panels and show some variation; ii) both PXO99A and MAFF include a single repeat with an insertion while MAFF also has a repeat which is deleted for a single amino acid in the recognition dipeptide; iii) PXO99A MIC P.T16 carries a repeat with a single amino acid removed, while both MIC P.T17 are identical; iv) shows the final MIC in the PXO99A cluster with no equivalent in the MAFF cluster.

• 

  Fig. Tn3.14Kii. Comparison of the MIC P.T14 Clusters in  PXO99A and MAFF Strains The Sequence of the 34 amino acid repeats of each TAL protein in consecutive MICs is shown in equivalent positions in the clusters. Recognition dipeptides are boxed. Right: PXO99A, left, MAFF. i) The recognition sequences are presented between the two panels and show some variation; ii) both PXO99A and MAFF include a single repeat with an insertion while MAFF also has a repeat which is deleted for a single amino acid in the recognition dipeptide; iii) PXO99A MIC P.T16 carries a repeat with a single amino acid removed, while both MIC P.T17 are identical; iv) shows the final MIC in the PXO99A cluster with no equivalent in the MAFF cluster.

• 

  Fig. Tn3.14Kiii. Comparison of the MIC P.T14 Clusters in  PXO99A and MAFF Strains The Sequence of the 34 amino acid repeats of each TAL protein in consecutive MICs is shown in equivalent positions in the clusters. Recognition dipeptides are boxed. Right: PXO99A, left, MAFF. i) The recognition sequences are presented between the two panels and show some variation; ii) both PXO99A and MAFF include a single repeat with an insertion while MAFF also has a repeat which is deleted for a single amino acid in the recognition dipeptide; iii) PXO99A MIC P.T16 carries a repeat with a single amino acid removed, while both MIC P.T17 are identical; iv) shows the final MIC in the PXO99A cluster with no equivalent in the MAFF cluster.

• 

  Fig. Tn3.14Kiv. Comparison of the MIC P.T14 Clusters in  PXO99A and MAFF Strains The Sequence of the 34 amino acid repeats of each TAL protein in consecutive MICs is shown in equivalent positions in the clusters. Recognition dipeptides are boxed. Right: PXO99A, left, MAFF. i) The recognition sequences are presented between the two panels and show some variation; ii) both PXO99A and MAFF include a single repeat with an insertion while MAFF also has a repeat which is deleted for a single amino acid in the recognition dipeptide; iii) PXO99A MIC P.T16 carries a repeat with a single amino acid removed, while both MIC P.T17 are identical; iv) shows the final MIC in the PXO99A cluster with no equivalent in the MAFF cluster.

It is interesting to note that while the amino acid repeats are always maintained in phase, certain TAL effectors have undergone removal of a single amino acid while another has acquired a short insertion. These changes might be expected to influence the capacity of the proteins to recognise their cognitive DNA sequence.

Diversification can also be observed between clusters in related X. oryzae strains such as MAFF and KACC.

Strain PXO99A and MAFF share the cluster MIC P.T15, MIC P.T16, MIC P.T17 (Fig. Tn3.14 Ki; Fig. Tn3.14 L). Both clusters have identical genomic environments (with some sequence variation) and the inter MIC sequences are identical. Not only has there been a large deletion of MIC P.T18 in the MAFF cluster, but sequence variations are apparent along the entire cluster length both within and between the clusters potentially modifying the DNA sequence recognition properties. Strain MAFF and KACC also share a cluster (MIC M.T2 and MIC M.T3).

Further analyses and experimental approaches are necessary to fully understand the role of MICs in the dispersal and diversification of these important instruments of Xanthomonad virulence, the TAL effectors.

Fig. Tn3.14L. Alignment of the MIC P.T14 Clusters in  PXO99A and MAFF Strains. The figure shows an alignment of the MAFF (top) and PXO99A (bottom) MIC clusters, indicating a deletion in MAFF in MIC P.T18 or an insertion in PXO99A of MIC P.T18.

Acquisition of Passenger Genes.

Tn3-family transposons carry large and diverse and diverse sets of passenger genes (e.g. Fig. Tn3.3). These have been acquired by a number of different processes.

Tn402 and integron platforms.

One major source of antibiotic passenger genes has been by ancestral insertions of Tn402 derivatives which have often “decayed” to lose their transposition properties but have retained their abilities to acquire (and lose) integron gene cassettes (Fig. Tn3.7 G; Fig. Tn3.7 I; Fig. Tn3.7 J; Fig. Tn3.7 R; Fig. Tn3.7 S; Fig. Tn3.7 U; Fig. Tn3.7 V; Fig. Tn3.18 C).

Additional TE

A second pathway to acquisition is by insertion of additional transposable elements with, or without rearrangement (Fig. Tn3.7 U; Fig. Tn3.8 E; Fig. Tn3.8 F). It is also interesting to note that there are a number of cases in which additional IR appear within certain structures (e.g. Fig. Tn3.10 D; Fig. Tn3.11 G) such as Tn3 and Tn501 raising the possibility that these have been involved in generating the host transposon.

Recombination at res.

A third major pathway to passenger gene acquisition is by inter-transposon exchange via res sites (see: Resolution). This was first suggested to explain the formation of Tn501, by exchange of a transposition module with a Tn1721-related transposon . It was later observed by Kholodii and coworkers and called “shuffling”, by Yano et al., and by others .

As judged by the analyses included here, this seems to be a recurring type of event and can be found in members of most clades (Tn21: Fig. Tn3.7 H; Fig. Tn3.7 M; Fig. Tn3.7 N; Tn163: Fig. Tn3.8 Ci and Fig. Tn3.8 Cii; Tn4330: Fig. Tn3.9 D; Tn3000: Fig. Tn3.11 C; Fig. Tn3.11 F; and Tn4561: Fig. Tn3.12 C). This type of behavior can also lead to “suicide” of a transposon in which the transposition module is removed by res recombination with a site outside the transposon .

Mercury Resistance: a Major Passenger Gene Group

The Mercury Operon and the Tn3 family

Fig. Tn3.15A. The mercury resistance operon. A typical mercury operon redrawn from Liebert et al 1997. Each gene is represented by a yellow horizontal arrow indicating the direction of expression. merO is the divergent operator site to which the merR regulator binds. The mercury transport genes, merC and merF, shown bracketed by dotted lines, are not always present and only one or the other is ever included in a given operon. The organo-mercurial lyase, merB, is not always present. Both merC and merT encode transmembrane segments (grey boxes).

Not surprisingly, bacteria carrying mer operons are particularly abundant in areas with increased mercury concentrations such as mercury mines and contaminated soil or water and it was suggested that mercury resistance is an ancient system as reflected by a wide geographical, environment and species range and that it evolved as a response to increased levels of mercury in natural environments resulting, for example, from volcanic activity .

It is certainly present in the Murray collection , a collection of Enterobacteriaceae isolated in the pre-antibiotic era, as part of transposons Tn5073 and Tn5074 which show high similarity to present day examples such as Tn5036 and Tn1696 (Tn3 family members of the Tn21 clade) and Tn5053 (a Tn402 family member of the Tn5053 clade) and Tn5075 respectively .

Although Tn3 family members carry a large variety of passenger genes, mercury resistance is found repeatedly within the family and is thought to be one of the first sets of passenger genes to be acquired (Fig. Tn3.6) and appears in precursors of the major groups of antibiotic resistance carrying Tn3 family members (Fig. Tn3.7 G). Mercury resistance operons were proposed to have been acquired at least twice (Fig. Tn3.6): once by an ancestor of Tn21 and once by an ancestor of Tn501. Their acquisition presumably predates the acquisition of antibiotic resistance integron platforms since a number of mercury resistance Tn3 family transposons have been identified and, in at least two cases, Tn21 and Tn1696 (whose mer genes appear to fall largely into different groups; Fig. Tn3Bi-vii), clear precursors devoid of integrons (Tn5060 and Tn20 and Tn1696.1 respectively) have been identified. Mercury resistance genes are found in a number of Tn3 family clades (Fig. Tn3.15 B and Fig. Tn3.15 Bi-vii). These include Tn3, Tn21, Tn163, Tn4430 and Tn4651. Those associated with the Tn21 clade occur upstream of, and are generally expressed towards, tnpR (Fig. Tn3.7 G); those of the Tn3 clade are located downstream of tnpA (Fig. Tn3.15 C) and in those carrying the tnpS/T genes, they are between the transposase module and the tnpS/tnpT module (Fig. Tn3.15 D).

Fig. Tn3.15B. Table showing the distribution of various mer resistance gens in Tn3 family transposons. The mer genes R through E are indicated at the top. The various Tn3 family transposons are cited in the left-hand column. Each line is coloured according to the Tn3 family clade. Transposons highlighted in yellow have mer gene duplications.

• Fig. Tn3.15. Panels C mer phylogenetic trees (use the arrows to scroll the figures).
• 

  Fig. Tn3.15Ci. MerR phylogenetic tree.

• 

  Fig. Tn3.15Ciii. MerT phylogenetic tree.

• 

  Fig. Tn3.15Civ. MerP phylogenetic tree.

• 

  Fig. Tn3.15Cv. MerC phylogenetic tree.

• 

  Fig. Tn3.15Cvi. MerA phylogenetic tree.

• 

  Fig. Tn3.15Cvii. MerD phylogenetic tree.

• 

  Fig. Tn3.15Cviii. MerE phylogenetic tree.

A survey of 29 functional mercury resistance transposons isolated from Gram negative bacteria in environmental isolates revealed that the most widespread of transposons belong to two types: transposons of the Tn21 clade of the Tn3 family and relatives of Tn5053, a member of the Tn402 family . In addition, Yurieva et al identified a third group, related to Tn5041, a member of the Tn4651 clade They also identify “mosaic” mer operonswhich, they suggest, are generated by homologous recombination between short DNA sequences. While MerR appears to be very similar between different mer operons, while MerA showed a higher degree of mosaicism as did MerT and MerP to some extent .

The Mercury Operon: Organization, Regulation, and Resistance Mechanism

The mechanism underlying mercury resistance has been extensively reviewed a number of times . Briefly, mercury resistance in gram-negative bacteria results in the release of gaseous mercury Hg₀. Mercury salts (HgII) are captured by the periplasmic MerP, transferred across the periplasm to the inner membrane proteins MerC or MerT and then across the cytoplasmic membrane to the mercuric reductase, MerA which converts it to the volatile Hg₀. The operon is regulated by two genes, merR and merD (Fig. Tn3.15 A). The order of these genes is generally merT, merP, merC, merA, merD and merE. merR is located upstream and is transcribed in the opposite direction with overlapping promoters. Binding of MerR represses expression of the operon and of itself. Interaction with Hg(II) releases MerR repression of the mer structural genes permitting their expression without significantly impacting on its autorepression and its interaction with RNA polymerase creates a pre-transcription initiation complex .

The product of the secondary regulator gene, merD , appears to play a role in down-regulating the mer operon . It binds weakly but specifically to the merOP region and DNase I footprinting identified a common operator binding sequence for both MerR and MerD .

The genes essential for mercury resistance were identified as merR, merT, merP and merA . An additional mercury ion transmembrane transporter gene, merE (UniProtKB - D4N5J4) involved in the accumulation of methyl-mercury is often present. Not all mercury operons include merC and some have a gene, merF , an alternative mercury ion transmembrane transporter (UniProtKB - Q1H9Y3). Some also include a mercury lyase gene, merB, involved in resistance to organo-mercury .

The Mercury Operon: Diversity in various Tn3 family clades.

The mer carrying Tn3 family members (Fig. Tn3.15 B) all lack merF. Most examples carry a full mer gene complement although a small group (Tn501, Tn511, Tn1412, Tn4378 and Tn4380) lack the merC gene and only 3 (Tn5084, Tn6294, Tn6332), all members of the Tn4430 clade) carry a merB gene and have a duplicated or partially duplicated mer operon.

Phylogenetic trees generated for MerR (Fig. Tn3.15 Ci and Cii), MerT (Fig. Tn3.15 Ciii), MerP (Fig. Tn3.15 Civ), MerA (Fig. Tn3.15 Cvi), MerD (Fig. Tn3.15 Cvii) and MerE (Fig. Tn3.15 Cviii) reveal that, in general, Tn501-related mer genes group separately from those of Tn21 relatives. This provides some support for the hypothesis that the mer operon had been acquired at least twice. These groups are separated by mer genes from Tn402 family relatives.

Within the Tn21 clade, all members carry the mer operon upstream of tnpR with the direction of transcription to the right (Fig. Tn3.15 D top). merR, on the other hand, is transcribed in the opposite direction and terminates with a TAG codon within the IRL sequence (Fig. Tn3.15 D bottom) with one exception, Tn6023.

Fig. Tn3.15D. MerR and the Tn21 clade. The top of the figure shows a generic Tn21 clade transposon with a typical mer operon. merR is shown at the left. An alignment of the DNA sequences of members of the Tn21 clade is shown underneath indicating the transposon name and accession number, IRL (enclosed in a blue filled box emphasized with a blue arrow), and the merR termination codon, TAG (enclosed in the red filled box). Note that in the case of Tn6203, there is an upstream termination codon. The termination codon is located within the IR.

On the other hand, for the few members of the Tn3 clade, the mer genes are located downstream of tnpA and are transcribed to the left (Fig. Tn3.15 E top) except for merR which is transcribed towards and terminates some distance from IRR (Fig. Tn3.15 E bottom), while for the unique Tn4651 member, the mer operons is located between tnpA and tnpS/T (Fig. Tn3.15 F).

Fig. Tn3.15E. MerR and the mer operon Tn3 clade. The top of the figure shows a generic Tn3 clade transposon with a typical mer operon. Note that here, the mer operon is not complete and the mercury genes are located downstream. merR is shown at the right. An alignment of the DNA sequences of members of the Tn3 clade are shown underneath to indicating the transposon name and accession number, IRR (enclosed in a blue filled box emphasised with a blue arrow.

Fig. Tn3.15F. MerR and mer operon in a tnpS/tnpT-carrying Tn3 family members  A map of Tn5041 (X98999.3) is shown. Note that the mercury operon is located between the tnpA and tnpS/tnpT genes.

The Mercury Operon: Tn21 in mer acquisition by Tn402?

It is worth noting that members of the Tn402 family Tn5053 mercury resistance subgroup carry a single copy of a sequence closely related to Tn21 IRL (Fig. Tn3.15 G top) located in a similar position with respect to the mercury operon as the resident IRL in the Tn21 group (Fig. Tn3.15 D and Fig. Tn3.15 G top and middle) (see ). There is some variability in the 10 C-terminal amino acid tail of the neighboring MerR protein (Fig. Tn3.15 G bottom) although the major part of MerR amino acid sequence is highly conserved. This raises the possibility that the mercury resistance genes carried by the Tn402 family elements was derived from an ancestral Tn21 group transposon.

Fig. Tn3.15G. Tn3 family Mercury resistance and its relationship with the Tn402 family. Members of the Tn402 family in the Tn5053 mercury resistance subgroup carry a single copy of a sequence closely related to Tn21 IRL (Top panel) located in a similar position with respect to the mercury operon as the resident IRL in the Tn21 group (Middle panel). There is some variability in the 10 C-terminal amino acid tail of the neighboring MerR protein (Bottom panel) although the major part of MerR amino acid sequence is highly conserved.

Transposition Mechanism Overview

Early Studies

In early studies of Tn3 (Tn1 and 2) , Tn4651 and Tn4430 it was clearly demonstrated that Tn3 family transposition occurs in a two-step process involving a replicative step in which the transposon first couples the donor and target replicons by single strand transfer to create a forked allowing replication to generate a fully double stranded cointegrates structure followed by a site-specific recombination step, resolution, catalyzed by a dedicated enzyme, the resolvase (Fig. Tn3.2). While the resolution step for a number of Tn3 family members has been studied in exquisite detail (see Resolution below), study of the initial strand transfer and replication steps have proved problematic.

Fig. Tn3.16A. The Consequences of Replicative Transposition. i) intermolecular transposition. a) donor and target molecules; b) cleavage at both terminal inverted repeats (TIRs) of the Insertion Sequence (IS; yellow double-headed arrow) results in nicks on both strands generating 3’-OH groups (grey circles) that attack the target site (red arrow); c) DNA replication generates a cointegrates containing a duplication of the IS and the target site; d) this can be subsequently resolved into a plasmid identical to the original donor plasmid and a modified target plasmid carrying an IS copy flanked by target site duplications arranged as direct repeats(DRs). ii) intramolecular transposition. a) molecule showing target sequence and two genes, a and b; b) the 3’-OH groups generated by cleavage at both TIRs can either attack the target site on the same (cis) or c) the opposite (trans) strand; d) When in cis, DNA between the IS and target site becomes circularized and contains one IS copy and target site; e) In trans, DNA between IS and the target site is instead inverted ("a  b" becomes "b  a"), bracketed by the original IS and a new copy in an inverted orientation. The target site is also duplicated but in inverted orientation and each TSD is associated with one IS copy. Black arrows indicate potential DRs from previous transposition events; different numbers represent different sequences. Oval and square represent the origins of replication on the donor and target molecules respectively. The red arrow marked “0” represents the target sequence which becomes the newly duplicated sequence following transposition. Black arrows marked “1” and “2” represent the original DR sequences. Figure from .

The consequences of these pathways are shown in greater detail in Fig. Tn3.16A. This underlines why not all Tn3 family transposition events yield transposons flanked by 5bp direct repeats. Figure Tn3.16Ai shows intermolecular transposition generating a cointegrates which, following resolution yields donor and target each with a single copy of the transposon in flanked by two DR copies. In intramolecular transposition, one pathway leads to a deletion while the other to an inversion. In neither case is the transposon flanked by direct target repeats. Tn3-mediated inversions and deletions of this type have been described a number of times with Tn3, Tn1, Tn2660 and Tn1721 .

Early studies also demonstrated that, like a number of transposons, the transposition frequency of Tn3 family transposons appears to decrease exponentially with increasing length (Fig. Tn3.16B). Tanaka and colleagues investigated Tn2603 and various derivatives ranging in length from approximately 5kb to 22.5kb from a number of different donor plasmids to both R386 and R388 target plasmids and noted a steep exponential reduction in transposition frequencies of over 1000-fold with increasing length. This observation would be more robust if transposition frequencies had been measured from the same donor plasmid and the transposons all had identical genetic contexts.

Fig. Tn3.16B. Transposition frequencies (relative) of Tn2603-related Tn3 family transposons of different lengths. Two target plasmids were used : R386 (blue squares) and R388 (blue circles). The donors were A: pMK1::Tn2613#1; B: pTY61 (Tn2603del-61); C: ColE1::Tn2608#1; D: pMK1::Tn21#1; E: pMK1::Tn2603#1; F: pSC101::Tn4.

Replicative transposition

One of the major problems in studying transposition of Tn3 members is that their transposases, TnpA, are long (~1000 amino acids) and difficult to solubilize.

Interaction of transposase and transposon ends

The Tn3 transposase, TnpA_Tn3, was first purified in 1981 and shown to bind DNA in a salt resistant way and one of the first attempts to investigate TnpA_Tn3 activity in vitro concluded that addition of ATP was necessary to obtain TnpA_Tn3 binding to the Tn3 ends. However, in a subsequent article this was shown to be erroneous and probably due to a pH effect of the added ATP solution . Purified TnpA_Tn3 was observed to bind specifically to both IR_Tn3 and protect a sub-terminal DNA region within the IR (Fig. Tn3.16 Ci) in a heparin resistant manner a measure of its strong and highly sequence-specific DNA binding activity while another study using a different TnpA_Tn3 purification scheme and DNA binding conditions showed a much less sequence-specific protection which included the entire IR_Tn3 and a significant region of flanking DNA. Further functional analysis of the Tn3 ends demonstrated that mutations in the first 10 IR_Tn3 base pairs (domain A) did not influence TnpA_Tn3 binding while mutations in the 13-38 base pair region (domain B) inhibited binding (Fig. Tn3.16 Ci), behavior confirmed in a second study .

This is a similar functional architecture to the ends of other transposable elements (see: General Information/IS Organization/Terminal Inverted Repeats). In addition, the effects of mutations in the Tn3 ends on transposition in vivo indicated that mutations in the TnpA_Tn3 binding site have a stronger effect when present at both transposon ends than when located at only one end.

Similar binding studies have been undertaken for Tn1000 (γΔ) (Fig. Tn3.16 Cii). Protection against DNAse is more extensive than for Tn3 although this depends critically on the binding and digestion conditions . The protection pattern is broadly similar with the tip of the terminal IR_Tn1000 remaining unprotected and protection extended to the inner end of the IR. Some weak protection occurred on the DNA region flanking the IR tip. In addition, however, the Tn1000 ends include a binding site for the host DNA architectural protein, IHF, and both proteins were found to bind cooperatively . However, IHF appeared to downregulate Tn1000 transposition . The juxtaposition of IHF sites and transposon ends has been observed in several other TE (see ).

Binding studies have also been carried out with the transposase of Tn4430, TnpA_Tn4430, a Tn3 derivative which encodes a TnpI resolvase . Here, it was necessary to use a mutant transposase (Fig. Tn3.16 Ei) which had been selected for a reduction in its transposition immunity (see Transposition immunity below) and which concomitantly showed an increase in transposition activity. Similar protection patterns (Fig. Tn3.16 Eii) were observed as with TnpA_Tn3 and TnpA_Tn1000: transposase binding protects the distal IR_Tn44300 internal region. The IR was divided into three regions (A, B1 and B2) based on sequence conservation, which largely correspond to the A and B regions of IR_Tn3 (Fig. Tn3.16 Ci).

Fig. Tn3.16C. Organisation of Tn3 and Tn1000 IRs. A) Nucleotide sequence of Tn3 terminal IRs. The IRL and IRR, which are perfect inverted repeats, are boxed and flanking DNA is shown as XXXXXXX. The horizontal striped boxes marked A and B indicate the two functional IR domains. The horizontal green lines represent the extent of protection by transposase binding to Dnase I digestion (redrawn from Ichikawa et al., 1987). B) Nucleotide sequence of Tn1000 terminal IRs. IRL (g) and IRR (D) sequences are boxed. The internal abutting IHF binding site is shown in red. The horizontal green and blue lines represent the extent of protection by transposase anf IHF binding respectively to Dnase I digestion (redrawn from Wiater and Grindley, 1988).

TnpA functional domains

The TnpA_Tn3 is 1004 amino acid residues long. Like many other transposases, it carries a DDE catalytic motif (General Information/Reaction mechanisms). Characterization of a series of fusions of TnpA_Tn3 segments to β-galactosidase (Fig. Tn3.16 Di) revealed that the N-terminal segment (residues 1-242) exhibited sequence-specific binding to the 38 base pair IR and that this region could be dissected into two sub-regions, amino acids 1-86 and 87-242, which showed non-specific DNA binding activity, implying that both were involved in sequence-specific end binding. The large central region also included two regions with non-specific DNA binding properties while the C-terminal region encodes the DDE catalytic site.

Fig. Tn3.16D. Tn3 and Tn1000 Transposases. i) A map of the 1004 amino acid residues long TnpATn3. The length of each of the three domains is shown above in amino acids while the regions with DNA binding functions are shown as orange boxes with the amino acid residues within; ii) N-terminal sequences of Tn3 (1-242) and Tn1000 (1-243). * and black residues indicate identical amino acids, red residues indicate differences. The boxed region is responsible for the different DNA sequence-specific binding activities of the two proteins. iii) A dot-plot of the two proteins showing their shared C-terminal end sequence and the more variable N-terminal regions. Data from .

The region of TnpA involved in DNA sequence recognition for binding to the transposon IRs was further investigated using a series of hybrid TnpA genes carrying the N-terminal IR-binding region constructed between TnpA_Tn3 and TnpA_Tn1000 . TnpA_Tn3 and TnpA_Tn1000 were found to share over 64% identity (Fig. Tn3.16 Dii). This enabled the definition of a region of TnpA which permits distinction between binding to an IR_Tn3 and an IR_Tn1000 (Fig. Tn3.16 Dii). A dotplot comparison of tnpA_Tn3 and tnpA_Tn1000 nucleotide sequences indicated that the 3’ ends of both genes were conserved whereas the 5’ ends showed some variation (Fig. Tn3.16 Diii) .

A functional map of the Tn4330 transposase, TnpA_Tn4430, was obtained by partial proteolysis with trypsin and chymotrypsin (Fig. Tn3.16 Di) . This treatment indicated that, like TnpA_Tn3, TnpA_Tn4430 has three major domains: an N-terminal domain (amino acids 1-152) similar to a CENP-B DNA binding domain ; a central region (amino acids 153-682); and a C-terminal domain (amino acids 683-980) with an RNase H fold-like domain including the catalytic DDE triad. Like other members of the family, the distance between the second D and E residues is somewhat longer than in typical DDE transposases and has been called an insertion domain and is likely composed of alpha-helical structures . The presence of insertion domains between the D and E residues observed in other transposases does not disturb the catalytic RNAse fold and, in both cases studied in detail , performs crucial functions in the transposition chemistry specific for each element.

Cleavage and Strand transfer.

In spite of the extensive DNA binding studies, the biochemistry of Tn3 family transposition has proved refractory to detailed analysis. A single study with Tn3 in vitro used a cell extract with high TnpA levels, a donor minimal plasmid replicon containing a mini transposon with Tn3 ends and a target molecule composed of concatemeric phage lambda DNA. Following the reaction, the phage DNA was packaged in an in vitro system and used to infect suitable recipient cells. The process yielded cells which appeared to carry large plasmids consistent with the formation of cointegrates. However, these were not physically characterized and the approach does not seem to have been developed further. Additionally, sequence-specific 3’ cleavage at the ends of a plasmid carried mini Tn3 derivative was observed with a cell-free extract containing TnpA_Tn3 in a reaction which required Mg2+ and was stimulated by a host factor determined to be acyl carrier protein (ACP) . A similar observation had been made for the Tn7 transposition reaction .

In a more recent a study using the mutant TnpA_Tn4430 an in vitro system including both strand cleavage and strand transfer was developed. The mutant TnpA_Tn4430 carried 3 mutations (Fig. Tn3.16 Ei) selected for a reduced level of transposition immunity but exhibiting a hyper transposition efficiency . It was shown, using a gel shift assay and differentially fluorescently labeled IR, that this TnpA derivative formed two types of complex which appeared to be single end and paired end (SEC and PEC) species containing one or two IR_Tn4430 molecules bridged by the transposase.

Fig. Tn3.16E. Tn4430 Transposase. i) A map of the 980 amino acid residues long TnpATn4430. The length of each of the three domains identified by partial trypsin and chymotrypsin digestion is shown above in amino acids. Mutations are shown as orange vertical lines marked with their positions in black below. Those mutations conferring hyper-transposition and reduced immunity are indicated by red vertical arrows below with the mutant residues marked. ii) DNA footprint analysis. SEC and PEC protection against DNAse on both strands (green horizontal lines). Small red dots in flanking DNA indicate additional protection of flanking DNA in the PEC. [(OP)2-Cu+] DNA enhanced cleavage in the PEC is indicated by orange triangles and in the SEC by purple triangles. iii) Consequences of strand transfer into a circular target. Left: a single strand transfer event; right a concerted strand transfer reaction which linearises the target DNA. Data from .

Footprinting both types of complex revealed an identical pattern of DNase protection (Fig. Tn3.16 Eii) except for some additional weak protection of flanking DNA in the PEC. When probed with the 1,10-phenanthrolinecopper [(OP)2-Cu+] nuclease, the PEC showed significantly enhanced cleavage at the IR tip and in the DNA flank, particularly on the lower strand indicating a change of DNA conformation (Fig. Tn3.16 Eii).

Correct single strand cleavage at the 3’ end of the IR tip was observed in typical cleavage conditions as well as some double strand cleavage (3’ and 5’). This was examined using both wildtype TnpA_Tn4430 and mutant derivatives with different transposition activities. The unexpected 5’ cleavage increase with increasing TnpA_Tn4430 activity and when Mn²⁺ was used instead of Mg²⁺ indicating that this is an aberrant activity.

Furthermore, precleaved IR substrates were able to form a more stable PEC as observed in other in vitro transposition systems such as those of transposon Tn10 and bacteriophage Mu. The system was also shown to support strand transfer of a precleaved IR into a supercoiled target plasmid. Integration of both single and to a lower extent concerted integration of two IR was observed (Fig. Tn3.16 Eiii). Initial data have also suggested that TnpA_Tn4430 binds preferential to DNA structures which resemble replication forks in vitro (cited in ) and insertion appears to be influenced by replication of the target molecule in vivo (cited in ).

Some initial evidence was also presented suggesting that the PEC was composed of a pair of IRs and a single TnpA_Tn4430 molecule. This has proved to be a misinterpretation of the data. In all other transposition systems, PEC complexes include two (or more) transposase molecules (e.g.). Recent data both from Atomic Force Microscopy (AFM) and Cryoelectron microscopy demonstrates that the TnpA_Tn4430 is indeed a dimer (B. Hallet personal communication; ).

Mechanism in the Light of Structure

A 3.6 Å average resolution cryoelectron microscopy structure has demonstrated that TnpA_Tn4430 is indeed dimeric and has provided some insight into how it might function in transposition . Moreover, using the hyperactive immunity deficient TnpA mutant it was possible to resolve a structure for the PEC which was composed of the transposase dimer and two double strand Tn4330 ends.

The structural model permitted a refinement of the TnpA_Tn4430 functional modules obtained from partial proteolysis and footprinting (Fig. Tn3.16 Ei and Eii). Four DNA binding domains were identified (DBD1-4; Fig. Tn3.16 F top). DBD1,2 and 4 bind the IR in a sequence-specific manner. The first (N-terminal proximal) DBD1 establishes both base and phosphate contacts largely with the internal region of the IR previously defined as B2 while DBD2 and DBD4 interactions are located towards the external end of B2 and into A. DBD3 interacts principally with the DNA flank in a non-sequence-specific manner (Fig. Tn3.16 F bottom). There are also phosphate contacts across the IR/flank junction by residues in the catalytic RNH domain. When bound, there flank is bent from the IR axis, an observation which was expected from the enhanced [(OP)2-Cu+] cleavage sites in this region. Note the similarities with the Tn3/Tn1000 transposase organization (Fig. Tn3.16 Di).

Fig. Tn3.16F. Tn4430 Transposase Functional Modules i) Tn4430 Transposase organization from (Fig. Tn3.16D); ii) refined functional module organization. The 10 functional modules are DBD1 and DBD2 separated by an alpha-helical arm domain; a dimerization domain, DD; DBD3 and DBD4 separated by a linker domain, LN; the catalytic domain containing the RNaseH fold, RNH, interrupted by an insertion domain called a scaffold domain, SFD; and a c-terminal domain, CT. iii) a combination of the footprinting data of , and structural data from Shkumatov et al 2022. SEC and PEC protection against DNAse on both strands (green horizontal lines). Small red dots in flanking DNA indicate additional protection of flanking DNA in the PEC. [(OP)2-Cu+] DNA enhanced cleavage in the PEC is indicated by orange triangles and in the SEC by purple triangles. Phosphate and base contacts are indicated by orange letters in the sequence which are boxed with the appropriate color corresponding to the interacting protein domain. The IR and the DNA flank are boxed in yellow and pale yellow respectively to correspond to the color in the structural figure.

The apo-protein appears relatively compact (Fig. Tn3.16 Gi). The dimer is held together at the bottom by the DD domains and at the top by the C-terminal domain which docks onto the surface of the adjacent monomer. The CT interaction appears to be further stabilized by DNA binding (Fig. Tn3.16 Gi). The authors point out that this is an unusual dimer interface. IR binding is accompanied by large conformational change (Fig. Tn3.16 Gi). In this pre-cleavage complex, the protein “arms” align the 4 DBD along IR, bend the DNA at Site A (Fig. Tn3.16 Fiii) which moves the flank with respect to the IR tip and places the scissile phosphate bond at the catalytic site of the opposite monomer both LN and RNH residues are involved.

Fig. Tn3.16G. A sketch of the Tn4430 Cryo-em Structure. i) The dimeric apo-protein (green and purple); ii) PEC structure. The bound IRs (yellow) and flank (pale yellow) are shown; iii) 90 degree rotation of i) (view from the top); iv) 90 degree rotation of ii) (view from the top). The dimerization domain (DD; bottom) is ringed in red as is the CT (top) domain.  

Like other transposition systems cleavage appears to be “in trans” (see: Cleavage in Trans: A Committed Complex), a constraint which ensures that the transpososome complex has been assembled before cleavages occur and prevents adventitious initiation of transposition. The two scissile phosphate bonds are correctly positioned to generate the expected 5bp DR. The S911 mutation which leads to hyper transposition and decreased immunity (T⁺/I^-) would appear to assist the apo-PEC transition, as indeed would the other T⁺/I^- mutations.

Another consequence of the transition is that, while the RNaseH fold is poorly defined in the apo-protein, it becomes more easily recognizable in the rearranged PEC. However, in this conformation only E881 (Fig. Tn3.16 Ei) is stably positioned while the other two members of the triad D679 and D751 are mobile.

The authors suggest that this is part of a regulatory process, protein metamorphism, and that additional factor(s) are involved in stabilising the catalytic pocket. It seems possibly that this may be regulated by correct docking of the target DNA. Which, they propose, could enter by opening of the DD interaction domains, a suggestion from studies with a branched DNA substrate (Fig. Tn3.16 Hi and Fig. Tn3.16 Hiii) representing a strand transfer product. The low-resolution structure suggests that the target segment of the branched molecule is located at the base (Fig. Tn3.16 Hii).

Fig. Tn3.16H. Proposed Target Engagement. The figure shows how a branched DNA molecule appears to dock with the Tn4430 transposase. i) and iii) representation of the branched molecule used in the study: These include the IR (yellow box) attached by a single DNA strand to the double-strand flank (pale yellow box) and by the other strand to a double-strand target (pink/red box). ii) representation of the cryo-em structure showing the pathway of the branched molecule with the « exit » position of the target and flank. iv) a reminder of the uncleaved target molecule.

These are proposed to be the position at which the target (Fig. Tn3.16 Hiv) may dock. This led to a model of stepwise transpososome assembly in which the apo-protein first engages a target molecule which opens a “cavity” between the two protomers and subsequently allows engagement of the IR.

Tn3 Transposition immunity, a poorly understood phenomenon.

In some of the earliest studies on TnA (Tn1) it was observed that transposition into a plasmid already carrying a TnA copy was severely inhibited, a phenomenon known as Transposition Immunity. The effect, identified by transposition of TnA from the E. coli chromosome to plasmid R388 or a derivative already carrying TnA was pronounced (a 10⁵ fold reduction in the immune target).

Two other Tn3 family transposons, Tn501 and Tn1721, also exhibited this inhibition phenomenon (cited as personal communication in ). However, other studies have identified plasmids having received two copies of TnA but these probably occurred at the same time rather than consecutively .

Transposition Immunity is a poorly understood phenomenon and some of the early studies gave a number of conflicting results. Immunity has since been observed for bacteriophage Mu and for transposon Tn7 (e.g. ) where it involves proteins with ATPase activity, MuB and TnsC respectively. However, Tn3 and its relatives do not encode this type of protein and only a single large transposase with no demonstrated ATPase activity is involved in transposition. It is therefore possible that immunity here is mechanistically distinct from that of both phage Mu and Tn7.

Immunity Requires a Transposon End

Further analyses of TnA demonstrated that between 290 bp and 470 base pairs at the right end (Fig. Tn3.16 Ii) were sufficient to confer immunity . These measurements were made either by accumulation of transposition events in bacteria grown on agar “slopes” or transpositions from the chromosome into a plasmid target in stationary phase cell . While plasmids carrying the right end showed immunity, those carrying the left end showed no immunity or only “partial-immunity”.

Fig. Tn3.16I. Immunity determinant in Tn3 and Tn1000. i) Tn3 Map (top) showing the region involved in immunity as a red dotted line and (bottom) IRR sequence. Differences with the Tn1000 IRL are shown in red .  ii) Tn3 IRR-IRL junction and deletion derivatives showing the level of immunity (right column) . Differences with Tn3 IR are shown in red.

Unfortunately, the quantitative effects are not clear from this publication. However, the conclusions are generally supported by another study which uses a different assay system involving a temperature sensitive replication mutant of plasmid pSC101 carrying a Tn3 derivative in which tnpR was inactivated by linker insertion.

In this system , cointegrates are not resolved and were isolated by “rescue” of the temperature sensitive donor plasmid by a coresident target plasmid following a shift to high temperature . Here, plasmids carrying restriction fragments containing one or other Tn3 ends conferred immunity; inclusion of both ends did not enhance immunity; and immunity was observed regardless of the orientation of the 38 bp IR end. Intriguingly, the distribution of insertions into an immune and non-immune targets appeared to be different . However, the study also indicated in some cases that the orientation of the Tn3 DNA fragment in the target affected the immunity level.

Furthermore, it was observed that deletions within the IRs which eliminated transposition, also eliminated immunity (Fig. Tn3.16 Iii) . However, studies comparing TnpA_Tn3 binding and immunity suggested that some mutants which do not affect transposase binding capacity do impact on transposition immunity. Moreover, a study which implicated TnpR_Tn3 in immunity was not supported by subsequent studies .

A finer scale analysis of the extent of the Tn3 IR sequence required for immunity was obtained by sequential deletion analysis of one IR (Fig. Tn3.16 Iiii). While a number of the deletions resulted in retention of certain internal IR nucleotides, a clear pattern is that the distal end of the IR segment rather than the tip of the IR is important (sequences in Box B; Fig. Tn3.16Ci). This is also largely in agreement with the results from Huang et al..

Fig. Tn3.16Iiii. Immunity determinant in Tn3 and Tn1000. iii) Effect of IHF on the immunity of Tn1000. The left column shows the relative transposition immunity levels in wild-type and IHF-cells . iv) Effect of IHF on immunity of Tn1000. The left column shows the relative transposition immunity levels in wild type and IHF-cells .

Interestingly, Bishop and Sherratt , using a plasmid system which allows identification of both inter- and intra-molecular Tn1 transposition Inversions and deletions were found to occur at frequencies similar to insertion suggesting that insertion into its own vector plasmid is not significantly subject to immunity. However, when Tn3 sequences, such as those present in pBR322, were also present in the transposon donor plasmid, inversions and deletions occurred at significantly lower frequencies.

For Tn1000, it was observed that 200 base pairs of the IRL (Gamma end) or 400 base pairs of the IRR (delta end) showed immunity to Tn1000 insertion while no other segment of Tn1000 conferred immunity. This was further refined to the terminal 38-base-pairs of IRR which were sufficient to confer immunity, whereas the 38-bp sequence of IRL conferred only moderate immunity (note that we use the standard nomenclature for IRL and IRR: viz IRR is defined as the IR towards which the transposase is expressed. This is the opposite of the nomenclature originally used for Tn1000). The IR sequence of both ends is identical for the first 35 base pairs and it was observed that this common sequence alone was not able to confer immunity .

Like Tn4652 (Fig. Tn3.12 G) in which IHF binding to sites located close to the ends positively regulates TnpA binding to the terminal IRs, Tn1000 also carries IHF sites proximal to the IRs. A more detailed analysis of the related Tn1000 IRR using a mating-out assay to measure transposition frequencies, showed that while the 38 base pair end was capable of conferring immunity on a target replicon, the neighboring IHF site (which is not present in TnA/Tn1,Tn2,Tn3) conferred a significantly higher level of immunity in the presence of IHF (Fig. Tn3.16Iiv) while removal of the terminal 2 GC base pairs at the tip had no real effect. IHF has been shown to bind cooperatively with TnpA_Tn1000 . This result strongly suggested that it is the IHF-enhanced binding strength TnpA_Tn1000 which determines the level of immunity .

The available data is relatively old and restricted by the experimental approaches available at that time. Since every assay system is different, it is not possible to directly compare results. However, in spite of the apparently conflicting detailed data, it appears likely that TnpA_Tn3 and TnpA_Tn1000 binding to an IR in the immune target is necessary for immunity.

Immunity in Tn4430

More recent studies on immunity of Tn4430 have involved isolation of TnpA_Tn4430 mutants which escape immunity . The mutants were screened for both transposition and loss of immunity (T⁺/I^-) using a papillation test. Surprisingly, these were not localized to a specific region of the protein but occurred over its entire length (Fig. Tn3.16 Ei).

The frequency of transposition into the permissive (non-immune) target of most mutants was similar to that of wild-type TnpA_Tn4430. However, immune-deficient mutations in the N-terminal region appeared to have a slightly increased transposition frequency whereas those clustering in the C-terminus exhibited a slightly decreased transposition frequency. Based on the cryo-em structure, these T⁺/I^- mutants are expected to positively affect the apo-PEC transition .

Although some data suggested that immunity could be observed in a relatively crude cell-free system , the establishment of a more defined and robust in vitro transposition system might permit further experimental investigation into the molecular basis of Tn3 family transposition immunity.

On Ended Transposition.

Early in the study of Tn21 and Tn1721, it was observed that, In the presence of the cognate transposase, plasmids containing a single inverted repeat (IR) can fuse efficiently with other plasmids in a reaction that requires neither the resolution system nor a functional host recA gene.

Insertion occurred at different sites in the target plasmid and the products contained a complete copy of the IR-carrying donor plasmid often with a duplication of various lengths of donor DNA. The sequence across the junction showed that the segment of donor DNA started precisely at the IR at one end, was variable at the other and the insertion was generally flanked by a 5bp DR generated in the target plasmid . Some recombinants were observed to contain only short segments of the donor plasmid .

Models involving asymmetric (rolling circle or processive) replicative transposition or simple insertion have been proposed for this type of transposition and it seems possible that this in some way results from insertion into an extant replication fork in the target DNA.

Resolution

The serine recombinases.

Efficient resolvase-catalyzed recombination between two directly repeated res sites is instrumental in completing transposition by physically separating donor and target molecules. This was first recognized in studies on complementation of transposition deficient Tn1 and Tn3 mutants where mutation of tnpR resulted in accumulation of cointegrates (Fig. Tn3.2 ii). It therefore showed that TnpR functions not only as a repressor of TnpA and TnpR expression by binding to the res site and blocking the promoters for both genes (see Fig. Tn3.17 Ci), but that it has an active function in the transposition process.

A number of resolvase enzymes have since been recognized (for a comprehensive review see (Fig. Tn3.17 Ai-iv).

Fig.Tn3.17A. Tn3 family res sites. Transposons are shown as pale yellow boxes ending in arrowheads. The transposon length in base pairs is indicated. Terminal inverted repeats (IRs) are indicated by gray arrowheads (IRL and IRR, respectively, labeled by convention with respect to the direction of tnpA transcription from left to right). Recombination sites (res, irs, and rst) are shown in green, transposition genes in purple, passenger genes in red (antibiotic resistance genes), orange-yellow (heavy metal resistance genes), and bright yellow (plant pathogenicity genes). (i) Tn3. Accession number V00613 (Tn3 clade). Carries the blaTEM-1a beta-lactamase gene and divergent serine recombinase/resolvase (tnpR) and transposase (tnpA) genes. The recombination site, res, composed of three subsequences, I, II, and III, is located between tnpR and tnpA, with site III proximal to tnpR. Recombination occurs within site I. (ii) Tn501. Accession number Z00027 (Tn21 clade). Carries an operon containing mercury resistance genes (mer) and colinear serine recombinase/resolvase (tnpR) and transposase (tnpA) genes. The res site is located upstream of tnpR. It has a similar organization as that of Tn3 with site III proximal to tnpR. Recombination occurs within site I. (iii) Tn4430. Accession number X07651.1 (Tn4430 clade). Carries no known passenger genes. Tyrosine recombinase/resolvase (tnpI) and transposase (tnpA) genes are colinear, and the recombination site, irs, is located upstream of and proximal to the resolvase gene with four subsites: inverted repeats IR1 and IR2 and direct repeats DR1 and DR2. Recombination occurs at the recombination core site IR1-IR2. (iv) TnXax1. Accession number AE008925 (Tn4651 clade). Carries two passenger genes involved in plant pathogenicity located at the left (xopE) and right (mlt) ends of the transposon. The resolvase has two components: a tyrosine recombinase (tnpT) and a helper protein (tnpS) expressed divergently. The res site, rst, is located between tnpT and tnpS and is composed of two pairs of inverted repeats, IR1 and IR2 and IRa and IRb. Recombination occurs at the IR1-IR2 inverted repeat. From .

The majority so far identified appear to be recombinases which use a serine residue as the nucleophile during recombination (Fig. Tn3.17 Ai and ii). These serine recombinases can be divided into two major groups (Fig. Tn3. 17A): the “classical” recombinases, TnpR encoded by Tn3, Tn21 and their relatives (~185 aa); and “long” serine recombinases (~300aa) (Fig. Tn3.17 B) (see . In both types, the catalytic center is located at the N-terminal end in a large catalytic domain which is followed by a smaller helix-turn-helix DNA binding domain. In the case of the “long” recombinases, there is a C-terminal extension compared to the “classic” resolvases. These fall largely within a small subclade in the Tn3 subgroup which includes Tn5044, the Xanthomonas transposons TnXc4 and TnXc5 and Tn1412 (Fig. Tn3.4A). It is worth noting that all members of this Tn group also encode a toxin/antitoxin system located between the divergent tnpA and tnpR genes (Fig. Tn3.4).

Studies with Tn1000 (γδ) and Tn3 res.

Early studies using the resolvase of Tn1000 (aka γδ) in vitro demonstrated that the enzyme could introduce double strand breaks in a res site and, in the absence of the divalent cation Mg²⁺, formed covalent TnpR-DNA intermediates .

Cleavage occurred at a crossover point in a palindromic sequence to generate a cleavage product with a free 3’OH group and a 2 base 3’ overhang (Fig. Tn3.17 Ci). Furthermore, formation of a free 3’OH implied that the covalent protein-DNA linkage occurred at the 5’ end and was more efficient if the substrate carried 2 directly repeated res copies. This led to the hypothesis that although TnpR acts as a repressor at res, binding simultaneously to two res copies in some way changes the protein conformation allowing recombination to proceed . It was further shown using DNase and footprinting that res_Tn3 and res_Tn1000 carry three TnpR binding sites , I, II and III (where sites II and III, known as accessory sites, are closely spaced and site I known as the core site, is very slightly distanced) (Fig. Tn3.17 Ci) and that the recombination point (the dinucleotide TA) is included within site I. Each site shows some degree of two-fold symmetry (Fig. Tn3.17 Ci). The res_Tn3 has an identical organization and almost identical sequence and the Tn3 and Tn1000 TnpR products are interchangeable . These similarities were exploited to determine the crossover point using Tn1000 TnpR-mediated resolution between res_Tn3 and res_Tn1000 carried by a single plasmid .

Fig. Tn3.17B. Organisation of Serine resolvases. The short (top) and long (bottom) serine resolvases are shown. Both contain the catalytic site near their N-terminal ends. The active site serine is indicated in red.  The catalytic domain is followed by a short helix which acts as an oligomerization helix and a helix-turn-helix DNA binding domain. In the case of the long serine resolvase derivatives, a C-terminal extension of approximately 100 amino acids is present .

Tn3 res, tnpR and tnpA gene expression.

In both Tn3 and Tn1000, tnpA and tnpR are divergent and the res site is located in the intergenic space with subsite III proximal to tnpR (Fig. Tn3.17 Ci). Promoters for both tnpA and tnpR, were located by footprinting of RNA polymerase and lie within res (Fig. Tn3.17 Ci). The -35 promoter elements of both gene are only 10 bp distant from each other and the -10 element of tnpA is located within site I straddling the point of recombination crossover (Fig. Tn3.17 Ci). The transcription start point for both genes has been mapped. Clearly, tnpA and tnpR expression would be regulated by TnpR binding.

Variant res sites with this configuration have been observed in which the center of sites I and II are separated by 4, 5, 6 and seven helical turns (see ).

The Mechanics of Resolution.

TnpR binding to res generates a highly compact protein-DNA complex as judged by electron microscopy . This was explained by the observation that TnpR binding to res-containing linear DNA fragments results in significant bending of the DNA although it was noted that the complex contains a single DNA molecule under the conditions use rather than two res sites .

Gentle proteolysis of purified Tn1000 TnpR was observed to generate two fragments: a large N-terminal fragment which includes the catalytic center and a smaller C-terminal fragment which binds to each of the three res sites (Fig. Tn3.17 B). Unlike full length TnpR which binds the res sub-sites with equal affinity, the C-terminal fragment binds to each of the half-sites but with different affinities suggesting that the N-terminal part of TnpR is involved in protein-protein interactions within the TnpR-res complex . Footprinting of small fragment binding indicated that the protection was centered on the 9bp half-sites (Fig. Tn3.17 Ci). Further studies using saturated mutagenesis of a halfsite from subsite I and chemical probing identified how the protein contacts DNA in both the major and minor grooves .

A model of the overall architecture of single TnpR-res complexes was proposed based on results using a number of footprinting agents to reveal sensitive sites on the DNA and permutation experiments to identify DNA curvature in which each subsite binds a TnpR dimer (with one monomer recognizing each partial diad symmetry element called “half-sites”) and introduces an “intra-site” bend in the DNA at each site while at another level, protein-protein interactions introduce inter-site bends (Fig. Tn3.17 D). Experimentally, this conformation requires all 3 sites and a correct spacing between sites I and II.

In vitro resolution systems have been developed and require supercoiled DNA together with a divalent cation, Mg²⁺ . A number of laboratories have contributed to an understanding of how the complex site-specific resolution recombination reaction takes place. These studies have used extremely clever techniques to understand the mechanics of this process including topology, mutagenesis and structural biology.

In vitro resolution requires a supercoiled DNA substrate carrying two directly repeated res sites and results in a simple concatenated recombination product with a specific change in linking number (the number of time one DNA strand crosses another) (Fig. Tn3.17 E) indicating that the synaptic complex must have a very precise type of protein-DNA architecture. The in vitro reaction is very inefficient when the res sites are in an inverted orientation raising the question of how the two res sites are aligned for recombination (for review see ). Random collision between res sites on a supercoiled molecule was ruled out since this would generate a complex concatenated product with a variable number of supercoils trapped between the recombined product (Fig. Tn3.17 Eii). Alignment of the two res sites was first proposed to occur when TnpR recognizes one site and tracks along the DNA molecule until encountering the second site. However, present evidence suggests that this is not the case [123]. In particular the observation that res site recombination can occur intermolecularly. A second hypothesis was that res sites meet via “slithering” i.e. continuous one-dimensional diffusion of supercoils in plectonemically (Fig. Tn3.17 Eii) wound DNA molecules (for review see ).

Fig. Tn3.17C. The Sequence and organization of res sites used by short serine resolvases. Short sequences required for resolution of Tn3 and Tn1000 (i) and Tn21, Tn501, Tn1721 (ii) are shown. The regions of resolvase binding defined by footprinting are shown as horizontal green lines. Sites I (core site) and sites II and III (accessory sites) are indicated and the generally accepted functional limits are boxed in blue were determined. The conserved inverted repeats defining “half-sites” are shown as horizontal blue arrows for Tn1000. -35 and -10 promoter elements are shown in red. And the direction of expression of the flanking genes is indicated and represented by thick horizontal blue arrows. The TA dinucleotide at which recombination occurs is shown in red and the cleavage is indicated at the bottom of the panel (i). Data for Tn3 and Tn1000  ; for Tn21, Tn501, Tn1721.

Intensive studies using both gel electrophoresis and electron microscopy to visualize TnpR recombination activities led to a model in which the two res Fig. Tn3.17 Eiii) sites are constrained in a configuration which entraps 3 supercoils (Fig. Tn3.17 Eiiib) and which takes into account the observation that Tn3 resolution (Fig. Tn3.17 Eiiic) removes four negative supercoils on recombination (Fig. Tn3.17 Eiiid) . The resulting energy change probably drives the reaction. In this model, it is TnpR interactions at the accessory sites II and III which are important for this allowing the recombining site I to finalize the recombination event (Fig. Tn3.17F).

Fig. Tn3.17D. Model for the structure of a single resolvase-bound res site. The larger N-terminal Catalytic domain has been eliminated for clarity. The DNA binding domain is indicated as blue circle. The DNA is shown as black parallel lines. The resolvase-induced bends Predicted by circular permutation experiments are indicated, together with the large kink at the recombination site I. Half sites containing individual repeated sequences are shown as orange arrows.

This occurs by simple rotation at site I on the flat hydrophobic surface after simultaneous cleavage of all four strands in the synaptic complex. The TnpR monomers remain attached to the 5’ ends (Fig. Tn3.17 F left) and the serine-DNA bond is then broken by attack by the 3’OH of the recombining site (Fig. Tn3.17 F right) to complete recombination. More than a single round of recombination can occur and this results in the generation of knots of increasing complexity with increasing numbers of recombination events (not shown).

This model is supported by the structure of a TnpR tetramer bound to two site I DNA molecules in a synaptic complex which shows that each TnpR dimer bound to its DNA presents an unusual flat, hydrophobic surface to the other member of the pair (Fig. Tn3.17 G) with the suggestion that strand exchange indeed occurs by rotation around this interface.

The Tn3 synaptosome

An understanding the structure of the Tn1000/Tn3 synaptosome is important to validate the proposed resolution reaction pathway (Fig. Tn3.17 E). Visualization of the entire synaptosome has proved challenging due to its complexity: with two res sites each composed of three TnpR binding sites, the catalytic site I and the “architectural” sites II and III, each with its six TnpR dimers. The first synaptosome structure obtained was that of Sin, a simpler organization with only two TnpR binding sites . A long awaited Tn3 synaptosome complex has now been described at the structural level (Fig. Tn3.17 H) and provides convincing evidence for the previous models (Fig. Tn3.17 Eiii).

The Tn1721, Tn21 and Tn501 res.

In contrast to those of Tn3 and Tn1000, the tnpA and tnpR genes of Tn1721, Tn21 and Tn501 are transcribed in the same orientation, with tnpR upstream of tnpA and their res sites located upstream of tnpR (Fig. Tn3.17 Aii and Fig. Tn3.17 Cii). They are relatively well conserved within the Tn21 clade (Fig. Tn3.7 F). Early experiments with Tn501 showed that it too underwent transposition using a cointegrate intermediate .

Like res_Tn3 and res_Tn1000, res_Tn1721 and res_Tn21 are composed of three TnpR binding sites (I, II and III) as determined by footprinting with site III proximal to tnpR (Fig. Tn3.17 Aii and Fig. Tn3.17 Cii) and each site exhibits some degree of dyad symmetry. Moreover, there is considerable identity observed the Tn21, Tn501 and Tn1721 tnpR genes and also between the res_Tn21 , res_Tn501 and res_Tn1721 sites .

All three elements complement a tnpR mutant of Tn21 whereas Tn3 does not . This is perhaps not surprising since the res_Tn3 sequence appeared to be quite different from those of this Tn group (Fig. Tn3.17 C) and the authors were unable to identify a res site homologous to that of Tn3. In addition, the TnpR amino acid sequence of Tn3 is somewhat distant from those of Tn21, Tn501 and Tn1721.

These observations were reinforced by additional studies demonstrating that purified Tn1721 TnpR can resolve cointegrate substrates containing repeat copies of res_Tn1721, of res_Tn21, and of a substrate carrying both res_Tn21 and res_Tn1721 copies, but not of res_Tn3 while Tn21 TnpR catalyzed site-specific recombination between directly repeated res_Tn21 and res_Tn1721 but not res_Tn3 . The reaction required a supercoiled substrate with two directly oriented res sites.

Fig. Tn3.17E. The serine recombinase resolution mechanism. A) The main product of the resolution reaction. The cointegrate (left) gives rise to a recombination product, a linked catenane. The entire res site is shown as a single filled arrow and the cointegrate as a circle with two replicons in blue and orange. Recombination results in two interlinked molecules. B) Recombination following random collision of two res sites. This would trap a variable number of negative supercoils depending on the position of the two res sites relative to the supercoils. C) In the two res sites, shown as three filled arrows (Ci) are synapsed to bring the crossover sites I into apposition (Cii) while trapping three supercoils. Resolvase tetramers at each site are shown as filled circles. Strand exchange between both sites I (Ciii) gives rise to the simple catenane (Civ) with two crossovers.

Several studies explored the DNA sequence binding and recombination specificities between Tn3 and Tn21 using hybrid TnpR containing the DNA binding domain of one and the catalytic domain of the other . These studies showed that, while a Tn21 TnpR catalytic DNA domain spliced to the Tn3 DNA binding domain has a somewhat lower affinity for res_Tn21, it retained some ability to mediate recombination between res_Tn21 but was unable to recombine res_Tn3 sites in spite of the fact that the hybrid protein was able to bind res_Tn3. This led to the conclusion that although

“alterations in amino acid sequence of resolvase within the helix-turn-helix DNA binding domain modulate the affinity of the protein for its DNA target sequence, the specificity of resolvase for recombination at its cognate res sites is determined by the resultant organization of the DNA-protein complex” .

Tn res activity tnpR and tnpA gene expression.

It was proposed that in all three elements, tnpA may be transcribed independently of tnpR and that its promoter is located within tnpR. Moreover, no Tn501 tnpR promoter could be found in vitro. This is consistent with the observation that in interreplicon Tn501 transposition into plasmid R388, resolution could be induced in the recipient by mercury selection suggesting that tnpR may be expressed at least partially as part of the mercury resistance operon located upstream of tnpR.

Interestingly, a study using Tn21 revealed a gene, tnpM (for modulator), whose expression appeared to enhance transposition and suppress resolution . TnpM results from the insertion of the Tn402 derivative, Tn5060 which led to the formation of Tn21 (Fig. Tn3.7 G). This event interrupted the urfM gene, of unknown function but possibly part of the mercury operon, generating the C-terminal fragment with a fortuitous translation initiation codon. Removal of the region in Tn21 resulted in a reduced transposition frequency and increased resolution activity and these activities were restored when the tnpM “gene” cloned into another compatible plasmid was provided in trans. Moreover, transposition of Tn501, which like Tn21, also includes a complete ufrM gene, was also affected. The mechanism by which the UfrM fragment, TnpM, functions is unclear and has not been addressed since its initial description .

The long serine recombinases

TnpR proteins carrying an extended C-terminus (TnpR_L) (Fig. Tn3.17 B) have been studied in only a single case, TnXca5 (ISXca5) from Xanthomonas campestris pv. citri XAS450 . Establishment of an in vitro system has shown that, as for the short forms of TnpR, recombination requires two directly repeated res_TnXc5 copies in a supercoiled plasmid substrate and Mg²⁺ as a divalent cation. Footprinting reveal three TnpR-binding subsites with a relative spacing and similar position with respect to tnpR_L as Tn3 and Tn1000 (Fig. Tn3.17 I).

Topological analysis of the recombination products suggests that the res_TnXc5 synaptic complex must be very similar to those of res_Tn3 and res_Tn1000 since 4 supercoils are lost on recombination and the directionality of strand exchange is the same . No structural studies are at present available.

Fig. Tn3.17F. Mechanism of cointegrate resolution by Serine-recombinases by rotational strand exchange (Tn3 and Tn1000). The two res sites at which strand crossover occurs, sites I, are shown as filled arrows interrupted at the point of recombination by the TA dinucleotide (in red) at which cleavage occurs. The repeated sequences to which resolvase binds are shown as blue arrows within the res sites. The recombination sites are aligned in parallel and their strand polarities are shown. Colors are as for those in Fig. Tn3.17E. The resolvase tetramer formed of two dimers is shown in dark and light green, in which the monomer DNA binding (B) and catalytic (C) domains are separated by a linker. The 5’ phospho-serinebond at the recombination point is indicated by a small blue circle and the free 3’OH as a thin blue arrowhead. The resolvase dimers at the left and right part of site I interact through a flat hydrophobic surface. Note that all 4 strands are interrupted simultaneously. Left: In the pre-strand exchange complex, the catalytic domains are shown on the inside of the synapse and the DNA-binding domain on the outside. Right: Strand exchange occurs by a 180° rotation of one partner dimer pair with respect to the other along the flat hydrophobic tetramer interaction surface. Rejoining occurs when the free 3’OH attack the 5’ phospho-serine bond.

Fig. Tn3.17G. Structure of the cleaved Site I synaptic complex: The figure shows the crystallographic structure of the Tn1000-resolvase tetramer covalently linked to cleaved DNA. This corresponds to the left panel in Fig. Tn3.17F. The two site I DNAs are shown in magenta and gold and blue and green. Each is bent at the point of cleavage. The resolvase tetramer is shown in grey (top) and brown (bottom) intersected by the dimer interface. The DNA binding and catalytic domains are indicated. The resolvase DNA binding domains are shown as small helix-turn-helix structures tucked into the major groove. 5’ phospho-serine bonds are shown at serine 10 – top green, bottom red. The flat hydrophobic interface at while strand exchange will occur by rotation is also indicated. This figure was taken from the PDB file: 1ZR2.

Serine-recombinases which use IHF/Hu: the Sin Synaptosome.

It is worth noting that certain serine recombinases, such as Gin and Hin , involved in inversion switches or Sin which is involved in plasmid recombination use “simpler” recombination sites but depend on DNA bending proteins such as IHF, Fis, HU and HUB to achieve the correct architecture. These are not known to act in the resolution process of Tn3 family transposons.

Fig. Tn3.17H. The Tn3 Synaptosome. i) The Tn3 res site (see Fig. Tn3.17 Ciii). Arrangement of TnpR dimers at sites II and III and the dimer of dimers between the two res sites at site I. Double headed arrows indicate monomer/monomer interactions in the TnpR dimers and dimer/dimer interactions between the two res sites at site I . iii) The synaptosome structure showing both DNA and TnpR molecules, from which the cartoon was drawn. Red bonds indicate hypersensitivity to Dnase in footprinting.

The Staphylococcal Sin resolvase was identified as part of a multi-resistance staphylococcal plasmid, pI9789 . While Sin is not part of a transposon or an invertible DNA segment, DNA sequences immediately upstream of the Sin binding site, sin, are favored integration sites for Tn554-family transposons. The sin site is composed of only two Sin binding sites I and II and an intervening region bound by an HU-like DNA bending protein such as the Bacillus subtilis Hbsu (Fig. Tn3.17 Ii). The functional site includes four imperfect: two inverted copies of a 12-bp motif repeat (sites I-L and I-R) and two in direct repeat (sites IIa and IIb) . While Sin on its own binds to sites I and II to generate DNase-protected regions, Hsbu on its own showed no protection. However, together, the two proteins result in protection of the entire site (note that protection is not even and additional sites of hypersensitivity are generated .

Fig. Tn3.17I. The Sequence and organization of res sites used by long serine resolvases. The regions of resolvase binding defined by footprinting are shown as horizontal green lines. Sites I (core site) and sites II and III (accessory sites) are indicated and the generally accepted functional limits are boxed in blue where determined .

Similar biochemical and topological and genetic approaches to those used in the analysis of Tn3/Tn1000 TnpR/res function were used to investigate the Sin/sin system . This showed that the synaptosome traps three supercoils between the points of strand exchange and requires both site I, where strand exchange occurs, and site II together with Hbsu. A synaptosome structure, the first determined foe a resolution system, has been solved and a cartoon showing the arrangement of Sin dimers and a dimer of dimers bridging the two sin sites is shown in Fig. Tn3.17 J. This illustrates how bound non sequence-specific DNA bending proteins assist in synaptosome formation.

The irs/TnpI system

A small group of known Tn3 family members which include the Bacillus thuringiensis transposons Tn4430 and Tn5401 encode a resolvase, TnpI, carrying a tyrosine residue at the active-site nucleophile (Fig. Tn3.17 Ki). The tnpI gene lies upstream of tnpA and both genes are transcribed in the same direction (Fig. Tn3.17 Aiii). Insertion mutagenesis showed that interruption of tnpI resulted in an increased level of cointegrate intermediates in Escherichia coli .

The Tn4430 sequence revealed a series of small sequence repeats directly upstream of tnpI as well as two smaller repeats abutting the inside border of IRL. The tnpI proximal repeat sequences include two 14bp inverted repeats, IR1 and IR2, together with two longer direct repeats, DR1 and DR2, related in sequence to IR1 and IR2 (Fig. Tn3.17 Kii).

DNase footprinting revealed that TnpI bound to all four sites together called the internal resolution site (irs) but not to the (unrelated) IRL proximal repeats (Fig. Tn3.17 Kii). Using a suicide substrate which contains a nick close to the point of recombination and which traps intermediates in the cleavage reaction, in an in vitro reaction TnpI was found to be able to bind to a linear DNA fragment containing IR1-IR2 and did not require assistance from the two DR repeats.

DNA cleavage is staggered occurring six base pairs apart (Fig. Tn3.17 Kii) forming a transient 3′-phosphotyrosyl bond leading to 3’OH in an identical way to other tyrosine recombinases (e.g. . A complete in vitro resolution reaction requires supercoiled DNA substrate .

A similar overall sequence architecture was observed upstream of tnpI in Tn5401 (Fig. Tn3.17 Kii). Here, the repeated sequences are 12bp long with identical repeats abutting the inside border of IRL and of IRR. Footprinting also identified the TnpI irs binding sites but, in addition showed TnpI binding to the IR proximal site .

Fig. Tn3.17J. Serine-recombinases which use IHF/Hu: the Sin synaptosome i) the sin site showing the sequence repeats horizontal arrows indicate repeat orientation, blue Sin binding sites . General region of protection against Dnase digestion are shown as horizontal green lines. ii) The Sin synaptosome. The blue elipses represent the location of the DNA bending protein. The orange and pink elipses are Sin monomers. Double-headed arrows indicate monomer/monomer interactions in the Sin dimers and dimer/dimer interactions which occur between the two res sites at site I. iii) The sin synaptosome structure. Note that the DNA bending proteins are dimeric (pink and pale pink) .

The Mechanics of Resolution.

In contrast to the requirements for the accessory sites I and II in serine recombinase-catalyzed resolution , there is no absolute requirement for the DR1 and DR2 accessory sites for activity in TnpI-catalyzed recombination. Instead, in their absence IR1-IR2 core site recombination can give rise to different recombination products such as deletions, inversions and intermolecular recombination in vivo and topologically complex products in vitro instead of the simple catenanes . In other words, the accessory sites channel recombination to generate resolutive recombination between two directly repeated irs sites on the same DNA molecule. This gave rise to the model shown in Fig. Tn3.17 K.

Fig. Tn3.17Ka. Y recombinases: The First Strand exchange. The model of ordered strand exchange is based on that proposed for the Cre recombinase (Van Duyne, 2001). The two core irs sites at which strand crossover occurs (IR1/IR2), are shown as filled arrows interrupted at the point of recombination by the six base pair core site at which recombination occurs (red and blue). The repeated sequences to which resolvase binds are shown as blue arrows within the irs sites. The recombination sites are aligned in parallel (i) and their strand polarities are shown. Only the C-terminal catalytic domain with its active site tyrosine (purple arrow) is shown for simplicity. In this system, each monomer in the tetramer undergoes activation by allosteric connections to its neighbour. (ii) The C-terminal tyrosine nucleophile (dark green monomers) attacks the appropriate strand in each IR1 site to generate a 3’ phosphotyrosine bond (small yellow circle) generating a 5’ OH (small red and blue circles). (iii) The 5’OH subsequently attacks the phospho-tyrosine bond (small yellow circle) from the opposite (dark green) monomer to complete strand transfer (iv) and generate a Holliday junction (v next figure).

Fig. Tn3.17Kb. Y recombinases: The Second Strand exchange. The first strand transfer product  (iii, previous figure) in the form of a Holliday junction then isomerises (v) and the second pair of recombinases (light green) are activated, the tyrosine (purple arrows) nucleophile (purple arrows) attacks the opposite strand (vi) at IR2, creating a phospho-tyrosine bond (small yellow circle) and a 3’OH (small red and blue circles). The 3’OH then attacks the opposing 5’ phosphotyrosine bond (vii) to complete strand transfer.

More specifically, formation of synapses including DR1 and DR2 was found to stabilize recombination intermediates favoring the forward recombination reaction and to impose an order of cleavage at the IR1-IR2 core sites: activation of the IR1-bound TnpI subunits (those furthest from the accessory sites) occurs resulting in IR1 cleavage (Fig. Tn3.17 Lii) and first strand exchange Fig. Tn3.17 Liii to form a Holliday junction (Fig. Tn3.17 Liv) while the second pair, the IR2-bound subunits, are then activated to resolve the holliday junction Fig. Tn3.17 Lv) by cleavage and exchange of the second strand (Fig. Tn3.17 Lvi) to resolve the cointegrate Fig. Tn3.17 Lvii) .

Although the exact topology of the synaptic complex is unknown, two alternative models lead to the conclusion formation of the synaptic complex induces the same net change in substrate topology, trapping two negative supercoils between the crossover sites and converting them into catenation nodes in the product (see Fig. Tn3.17 L).

Fig. Tn3.17L. The S/T System i) Map of Tn5041 (top) and proposed res site (bottom); ii) TnXO19 (Tn7217) showing an insertion of  two genes with potential ATPase activity between TnpT and TnpS.

Irs, tnpR and tnpA and gene expression.

Transcriptional start sites within Tn5401 were mapped by primer extension analysis and the -35 and -10 promoter elements were identified (Fig. Tn3.17 Kiii) . Two overlapping and divergent promoters were identified: one which would drive expression of tnpI and tnpA and the other which could drive the upstream but divergent toxin antitoxin genes (see: Tn3 family-associated TA passenger genes are located in a unique position).

The rst/TnpS/T system.

The third major type of resolution system encoded by Tn3 family members is the TnpT-TnpS system which uses a resolution site, rst (res site for TnpS and TnpT) encoded by the catabolic transposons Tn4651 . Tn4651, isolated from a Pseudomonas plasmid carries a set of toluene degrading (xyl) passenger genes (Fig. Tn3.3) and is similar to the mercury resistance transposon Tn5041 (Fig. Tn3.17Ki) . The tnpS and T genes are expressed divergently with the res site between the two. In some cases, tnpT and tnpS are separated by insertion of passenger genes (Fig. Tn3.17 Lii). Resolution of cointegrates generated by Tn4651 was shown to require three Tn4651-encoded factors: the res site (now called rst) and the tnpS and tnpT gene products which are located at a significant distance (48kb) away from the tnpA transposase gene. A similar long distance between transposase and resolvase is found in Tn5041 Fig. Tn3.17 K). Here, tnpS and tnpT are referred to as orfQ and orfI respectively and rst as att_Tn5041.

Fig. Tn3.17M. The TnpS Y Recombinase Domains I and II. The conserved catalytic motif of typical tyrosine recombinases is shown by vertical blue arrow heads. Data initially from .  

The 323 aa TnpS protein is a tyrosine recombinase (Fig. Tn3.17 O) with similarity to the Cre resolvase while the 332 aa TnpT appears to enhance TnpS-mediated recombination .

The sequence of the tnpS/T intergenic region is very similar in Tn4651 , Tn4652 (a Tn4651 deletion derivative lacking the toluene-catabolic genes) , Tn4661 and Tn4676 . It is composed of a 203 bp sequence which includes two pairs on inverted repeats, IRL and IRR and IR1 and IR2 (Fig. Tn3.17 Pi) with overlapping promoters which drive TnpS and TnpT expression. The mRNA start point was identified by primer extension .

The length of functional rst site, 136 bp, was defined by the recombination activities of sequential deletion derivatives in an in vivo resolution system . This involved the construction of an artificial cointegrate containing one complete rst copy and a second copy which carries the deletions. IR1 and IR2 are indispensable for the full resolution activity and cointegrate resolution was shown to require both TnpS and TnpT.

Moreover, the resolution reaction could be reversed to obtain site-specific integration (recombination between rst sites on different DNA molecules) in a reaction which requires TnpS but not TnpT. Suggesting that TnpT is a factor which determines the direction of recombination .

Although the TnpS/T proteins of Tn4651 and Tn4661 are highly similar and the Tn share highly similar sequences in the inverted repeat motif, IRL and IRR, of the rst core site, the 7bp spacer separating the repeats are somewhat different (Fig. Tn3.17 Pi). An artificial cointegrate composed of an rst_Tn4651 and an rst_Tn4661 site could not be resolved using TnpS_Tn4651 and TnpT_Tn4651 . The mismatches in the IRL-IRL region concern principally the spacer region between IRL and IRR. rst_Tn4651 and rst_Tn4661 have six mismatches, with five located in the spacer (Fig. Tn3.17 Pii). In other tyrosine recombinase systems such as xerC/D or phage P1 Cre protein (see ) this is the region where strand cleavages occur and sequence differences have a strong influence on the ability for two sites to recombine.

The effect of these sequence differences was investigated using a cointegrate, in which the spacer sequence of rst_Tn4661 was replaced with that from rst_Tn4651, (rst_Tn4661v1) (Fig. Tn3.17 Nii). In contrast to the cointegrate carrying both rst_Tn4651 and rst_Tn4661, that carrying rst_Tn4651 and rst_Tn4661v1 underwent TnpS_Tn4651 and TnpT_Tn4651-mediated resolution, demonstrating that it is the differences spacer sequence which prevents recombination . Moreover, the sequence of this region in the resolved products confirmed that recombination occurred within the IRL-IRR region. It should be noted that neither the tnpS/T intergenic sequence nor the proposed core recombination site of Tn5041 (Fig. Tn3.17 K) shows significant similarity to those of Tn4661 . Moreover, in depth analysis of the Tn5041 resolution reaction has not been reported.

No footprinting, binding stoichiometry or topology studies are yet available for the TnpS/T system and the exact role of TnpT in the resolution reaction is not known although it has been demonstrated to bind the DNA region containing IR1 and IR2 .

Fig.Tn3.17N. S/T Recombinase res Sites. i) Tn4651/Tn4652 res site showing the expression signals for the tnpS and tnpT genes with convergent (overlapping) promoter elements. mRNA initiation and the IRL-IRR (blue horizontal arrows) and core recombination sequence. The equivalent sequences of Tn4661 and Tn4676 are shown below, with the variant bases marked in red. ii) res site sequences used in inter res recombination analyses. Data from .

Toxin-Antitoxin genes: Special Passengers linked to the transposition process?

Several studies had identified individual Tn3 family members with type II toxin-antitoxin (TA) passenger genes (see reference ). Unusually for passenger genes of this Tn family, the T/A genes are consistently found adjacent to a resolvase gene.

Some type II TA systems are involved in plasmid maintenance in growing bacterial populations by a mechanism known as post segregational killing. Upon plasmid loss, degradation of the labile antitoxin liberates the toxin from the inactive complex, which in turn is free to interact with its target and cause cell death. They were first identified in the mid-1980s in plasmids F and R1 and it was recently shown that acquisition of a Tn3 family transposon Tn6231 carrying a type II TA gene pair was indeed able to “stabilize” an unstable target plasmid .

Many different type II TA gene pairs have now been identified in bacterial chromosomes as well as plasmids . They are generally composed of 2 relatively short proteins: a stable toxin and a labile antitoxin that binds the toxin and inhibits its lethal activity (see reference ). The antitoxin includes a DNA binding domain involved in promoter binding and negative regulation of TA expression.

Identification of TA gene pairs in Tn3 family members.

Among nearly 200 Tn3 family members, 39 were observed to carry type II T/A genes (colored squares; Fig. Tn3.4A, Fig. Tn3.18 B) . The host transposons included examples from all known combinations and orientations of transposase and resolvase genes (Fig. Tn3.18 A) and were almost all located adjacent to the resolvase genes in family members with TnpR, with long serine TnpR, with TnpI (e.g. Tn5401 and TnBth4) and with TnpS/T (e.g. TnHdN1.1) (Fig. Tn3.4A, Fig. Tn3.18 B). Illustrative examples are shown in (Fig. Tn3.18 A).

Fig.Tn3.18A. Examples of TA gene pair location in a variety of Tn3 family transposons. Toxin-antitoxin gene pairs shown in bright orange, genes of unknown function in magenta, transposition-related genes in purple, heavy metal resistance genes in chrome, and antibiotic resistance genes in red. i) Tn5046, accession Y18360.1, has an unusual structure with the mer passenger genes located downstream from the transposase gene. It carries a typical tnpR cognate res site. ii) Tn5501.6, accession MF487840.1, carries a blaNPS-1 passenger gene. It carries a typical tnpR cognate res site. iii) Tn5401, accession U03554.1, there are no known passenger genes apart from the TA gene pair. It carries a typical tnpI cognate irs site with, in addition, a copy of the DR2 TnpI binding site close to each end. iv) TnHdN1.1, accession FP929140.1, is treated as a partial copy since the ends of the transposon have not yet been identified. Consequently, no passenger genes except the TA gene pair have been identified. However, TnHdN1.1 carries a typical tnpT/tpnS resolvase pair and the toxin/antitoxin genes are located between the resolvase and the tnpA gene. The rst site has not yet been defined, but for other transposons with a TnpS/T/rst resolution system, it is located between the divergent tnpS and tnpT genes. v) Tn4662a: accession NC_014124.1. This transposon carries a potential metal-dependent phosphohydrolase passenger gene and a tnpR cognate res site. In this case, in contrast to the vast majority of cases, the toxin gene is located upstream of the antitoxin gene.

Fig.Tn3.18B. Tn3 family TA systems and configuration of the TA operon with respect to the resolvase and TnpA genes (extracted from ).

TA diversity in Tn3 family members.

The Tn3-associated TA modules include a number of different types of TA module: 5 toxin (RelE/ParE, Gp49, PIN_3, PIN,and HEPN) and 6 antitoxin families (ParD, HTH_37, RHH_6, Phd/YefM, AbrB/MazE, and MNT) (Fig. Tn3.4A; Fig. Tn3.18 B). All, except ParE, are associated with RNase activity , while ParE inhibits DNA gyrase activity by an unknown molecular mechanism .

TA distribution and organization within the Tn3 family

The majority of examples occurred in two Tn3 subgroups: Tn3 (2 toxin families) and Tn3000 (3 toxin families), but other subgroups also included members with T/A modules (6 members of 5 different toxin families (ParE, Gp49, PIN_3, PIN, and HEPN). The majority of T/A-containing members of the Tn3 subgroup also encode a long serine recombinase, TnpR_L as their resolvase and two (Tn5401 and TnBth4) encode the tyrosine TnpI resolvase, while those in the Tn3000 subgroup all encode a short serine resolvase, TnpR_S (Fig. Tn3.4A; Fig. Tn3.18 B). It is also noteworthy that, a given toxin gene can be paired with different antitoxins forming 7 different toxin-antitoxin pairs: ParE-ParD, ParE-PhD, PIN_3-RHH_6 (??), Gp49-HTH_37, PIN-Phd, PIN-AbrB, and HEPN-MNT (Fig. Tn3.18 B).

Although TA genes are generally arranged with the antitoxin upstream of the toxin gene, TA systems of reverse order have been identified . Among the Tn3 family-associated TA systems, in five Tn3000 subgroup members (Fig. Tn3.18 B and Fig. Tn3.4A) the RelE/ParE superfamily toxin Gp49 (PF05973) toxin gene precedes that of a HigA superfamily antitoxin, HTH_37 (PF13744) . A similar situation is found in the unrelated Tn4651 subgroup member TnPosp1_p.

In addition to encoding a TnpI resolvase, Tn5401 and TnBth4, both encode a ParD antitoxin, which appears to lack the DNA-binding domain.

Acquisition and exchange of TA modules.

An important question is whether these systems have been repeatedly recruited or have evolved from a common ancestor. Putting aside the fact that several groups of T/A encoding Tn (e.g. Tn5051 and its derivatives which differ essentially by their other passenger genes), clearly the fact the Tn collection also includes examples of different combinations of T/A genes and examples in which the gene order has been inverted argue for a certain level of repeated acquisition.

In cases where the TA module is found in related transposons (with similar tnpA and/or resolvase genes), it is likely that it was first acquired by a transposon that subsequently diverged. Alternatively, for transposons which are generally not related (different tnpA family group, different resolvase) but which harbor TA modules that are similar at the DNA level, it is likely that the TA module was acquired by recombination with another transposon.

Phylogenetic analysis suggested that ParE had been acquired three times, Gp49 together with an HTH antitoxin on three occasions and PIN on two occasions .

Although it is unclear how most of the T/A modules were initially acquired, it is important to underline that the res/rst/irs sites are highly recombinogenic in the presence of their cognitive resolvases producing transitory single (Y-recombinases) or double (s-recombinases) breaks. It seems possible that the modules were recruited via non-productive resolution events.

Moreover, this recombination activity has clearly led to spread of T/A modules to different Tn3 family members by inter-res recombination (Fig. Tn3.18 B). There are two cases in the Tn3 library, both involving Tn5051 and its derivatives, which demonstrate this capacity (Fig. Tn3.18 C). In the first case (Fig. Tn3.18 Ci), comparison between Tn5051 and TnTsp1 shows a clear break in the homology between the two Tn which occurs at the res site (Fig. Tn3.18 Ci). The identities towards the right of the res site III decrease rapidly within a short distance. In the second case (Fig. Tn3.18 Cii), comparison of Tn5051 and Tn4662a shows a clear break in identity at res site I suggesting that they have previously exchanged left and right ends via recombination at res site I. An additional transposon, Tn5051.12 is clearly a hybrid of these two since it carries the left end of Tn4662a and the right end of Tn5051.

Fig.Tn3.18C. Inter-transposon recombination at the res site exchanges TA modules. i) Comparison of Tn5501 accession JN648090.1 and TnTsp1 accession NC_014154 showing a possible recombination point between the two Tn where exchange at the TA gene pair may have occurred. The bottom section shows the region of TnTsp1 including the TA gene module (orange), the res site (green), tnpR and part of tnpA (purple). The top segment shows the equivalent map of Tn5501. Below is shown a DNA sequence alignment (magenta) with the equivalent region of Tn5501. Both transposons have similar DNA sequences to the left of res site I. The level of sequence identity is reduced in tnpR and is insignificant in tnpA. The res site I sequences (green) are shown between the two panels and the AT dinucleotide at which recombination probably occurs is indicated in red. Sequence non-identities are underlined. The two sequences are identical up to the probable recombination site and show some diversity to its right. ii) The region of Tn5501.12 accession CP017294.1 showing the 5’ end of the tnpA gene, the tnpR gene, a res site typical of the tnpR res sites, and toxin/antitoxin gene pair (note that the toxin gene is upstream of the antitoxin gene. The horizontal magenta lines at the bottom show the alignment of Tn5501.12 with Tn5501 and Tn4662a (NC_014124.1). The right half of Tn5501 is clearly highly homologous to the right side of Tn5501.12 whereas the left side of Tn4662a is homologous to the left side of Tn5501.12. The DNA sequences at the top show the res subsite I (green) with the dinucleotide at which recombination should occur in red together with flanking sequences. Underlined bases indicate regions of nucleotide identity. This suggests a scenario in which Tn5501.12 was generated by recombination at res I between transposons similar to Tn5501 and Tn4662a.

Tn3 family-associated TA passenger gene are located in a unique position.

In most of the cases identified, the T/A modules are embedded within the transposition module comprising transposase and resolvase genes and the res site at a position very close to the res sites (Fig. Tn3.18 A). This is in sharp contrast to all other Tn3 family passenger genes, which are generally located away from the resolution and transposon genes and, where known, have often been acquired as integron cassettes or by insertion of other transposons.

Indeed, several TA-carrying transposons represent closely related derivatives with identical transposase, resolvase, and TA modules but contain different sets of passenger genes (e.g., Tn5501.1 and derivatives 5501.2, 5501.3, 5501.4, etc.). Most T/A modules are located directly upstream of the resolvase genes (tnpR, tnpR_L or tnpI) (Fig. Tn3.18 A) with only three exceptions: a single example of a derivative with the TnpS/TnpT resolvase, TnHdN1.1 (Fig. Tn3.18 Aiv), where they are located between the resolvase tnpS and transposase genes; TnSku1 [Tn7197], where they are located downstream of and transcribed towards tnpR; and a partial transposon copy, TnAmu2_p with a short open reading frame (ORF) of unknown function between the divergently transcribed antitoxin and tnpR genes.

Regulation of Tn3 family TA gene expression.

An as yet unanswered question is how expression of the identified Tn3-associated T/A genes is regulated. It is possible that it occurs from their own promoters although it has not yet been demonstrated any of the T/A modules carry their own promoters. Alternatively, the fact that the genes are embedded in the transposition modules, it is tempting to speculate that they may be regulated in a similar way to tnpR and tnpA expression.

Tn3 family with TnpR and TnpR_L

In Tn3 itself, which has been examined in detail, transposase and resolvase gene expression is controlled by promoters found within the res site located between the two divergent genes (Fig. Tn3.17C) which are regulated by resolvase binding. The location of the TA genes in proximity to the res sites raises the possibility that their expression is also controlled by these promoters (Fig. Tn3.18 Di). Although few of the res sites in the collection of TA-associated Tn3 family members have been defined either experimentally or by sequence comparison, 27 potential sites were identified using the canonical tnpR-associated res-site organization schematized in as a guide, a res site library (kindly provided by Martin Boocock), and RSAT tools (Regulatory Sequence Analysis Tools) . For transposons with a TnpR or TnpR_L resolvase, the TA genes are always located just downstream from res site I, whereas tnpR is located next to res site III (Fig. Tn3.18 A i, ii and v; Fig. Tn3.18 Di).

In transposons with divergent tnpA and tnpR such as TnXc5 and Tn5563a (Fig. Tn3.18 Di), tnpA, tnpR and res are organized similarly to those of Tn3, which itself does not carry the TA module, except that tnpA is separated from res by the intervening TA genes. This organization is also similar in Tn3 members in which tnpA is downstream of tnpR and in the same orientation (e.g., Tn5501 and Tn4662a; Fig. Tn3.18 A i and v).

Tn3 family with TnpI

Promoters have also been defined in the res (irs) site of the tnpI-carrying Tn5401 , and tnpI and tnpA expression is modulated by TnpI binding to the irs site (Fig. Tn3.17I). The other tnpI-carrying transposon with TA genes, TnBth4, has an identical irs site, and therefore expression is probably regulated in the same way. Again, the potential promoters are pertinently located for driving expression of the TA module (Fig. Tn3.18 Dii).

Fig.Tn3.18D. Relationship between the res site, known promoter elements, and TA gene pairs. In addition, potential or proven minus-10 and minus-35 promoter elements are shown as red arrows. i) Res sites (green) with a structure related to Tn3. 300 bp including flanking DNA is shown. TnpR (purple) is expressed to the left, and TnpA (purple, Tn3) and the toxin/antitoxin genes (orange, TnXc5, and Tn5044) to the right. In this type of organization, the res III subsite is proximal to tnpR. Recombination leading to cointegrate resolution occurs at a TA dinucleotide within res site I. a) Tn3 res site. The res site was defined by footprinting using TnpR and by functional deletion analysis. The promoter elements are predicted. b) TnXc5 res site, also called ISXc5. The res site was defined by footprinting with TnpR. c) Tn5044 . The res site was defined by comparison with TnXc5 and as described here. ii) Res site organization for transposons carrying res sites for the TnpI resolvase. a) Tn5401 res site. This was identified by footprinting with TnpI and by deletion analysis . b) TnBth1 res site (NZ_CP010092.1). TnBth1 is similar but not identical to Tn5401 over the res site but varies considerably in the tnpI and tnpA genes. It maintains the promoter elements (red arrows) identified in Tn5401. tnpI and tnpA are expressed to the right. The toxin-antitoxin pair is expressed to the left. iii) Gene organization for transposon TnHdN1.1 carrying the TnpS/T resolvase.

Tn3 family with TnpS/T

Finally, transposon TnHdN1.1 (Fig. Tn3.18 Aiv) is the only example in our collection of a tnpS/tnpT transposon carrying a TA module. The res (rst) site and relevant promoter elements for the divergently expressed tnpS and tnpT have been identified between the two genes in transposon Tn4651 (Fig. Tn3.18 Diii). In TnHdN1.1, the TA gene pair is to the right of tnpS, between tnpS and tnpA, and all three genes are oriented in the same direction. Although the exact regulatory arrangement remains to be determined, it seems possible that the promoters in the rst site regulate expression of the TA gene pair.

Thus, for all three types of resolvase-carrying Tn3 family members, the T/A gene module is strategically placed so that it could be place under control of the resolvase/transposase transcriptional expression signals except for the  two exceptions TnSku1 and TnAmu2_p. T/A activity could therefore be intimately linked to the transposition process itself rather than, or in addition to, simply providing a general addiction system that stabilizes the host replicon, generally a plasmid, carrying the transposon.

A model for T/A activity in transposon transposition.

Type II TA expression, like that of tnpA and tnpR, is tightly regulated at the transcriptional level (see ). Where analyzed, the toxin-antitoxin complex binds via the antitoxin DNA-binding domain to palindromic sequences located in the operon promoter and acts as a negative transcriptional regulator. This regulation depends critically on the relative levels of toxin and antitoxin in a process known as conditional cooperativity, a common mechanism of transcriptional regulation of prokaryotic type II toxin-antitoxin operons in which, at low toxin/antitoxin ratios, the toxin acts as a corepressor together with the antitoxin. At higher ratios, the toxin behaves as a derepressor. It will be important to determine whether the Tn-associated TA genes include their indigenous promoters .

Transposon Tn6231 (99% identical to Tn4662) clearly provides a level of stabilization of its host plasmid implying that TA expression occurs in the absence of transposition. There are a number of ways in which this could take place (Fig. Tn3.18 E). Expression could occur from a resident TA promoter (Fig. Tn3.18 Ei) if present. However, this might lead to expression of the downstream tnpA gene by readthrough transcription. Alternatively, in the absence of a TA promoter, TA expression could occur stochastically from the res promoter (Fig. Tn3.18 Eii). However, this does not rule out the possibility that TA expression is regulated at two levels with a low-level “maintenance” expression, resulting in the plasmid stabilization properties described by Loftie-Eaton et al. together with additional expression linked to derepression of the tnpA (and tnpR) promoters that must occur during transposition (Fig. Tn3.18 Eiii). Regulation of tnpR and tnpA by TnpR is a mechanism allowing a burst of TnpA (and TnpR) synthesis, transitorily promoting transposition as the transposon invades a new host. Subsequent repression by newly synthesized TnpR would reduce transposition activity, reinstalling homeostasis once the transposon has been established, a process similar to zygotic induction or plasmid transfer derepression as originally observed for the plasmid ColI and subsequently for plasmids R100 and R1 . An alternative but nonexclusive explanation stems from the observed enhanced plasmid stability afforded by Tn6231 TnpR, in addition to that afforded by the neighboring TA system . Resolvase systems are known to promote resolution of plasmid dimers (see reference ), and it was suggested that integration of the TA system into Tn6231 “such that all the transposon genes shared a single promoter region” permits coordinated TA and TnpR expression and may facilitate temporary inhibition of cell division while resolving the multimers, promoting plasmid persistence. In this light, it is interesting that the ccd TA system of Escherichia coli plasmid F is in an operon with a resolvase-encoding gene .

Expression of the TA module from the tnpA/tnpR promoter at the time of the transposition burst could transiently increase invasion efficiency (“addiction”) over and above that provided by the endogenous TA regulation system. If the transposon is on a molecule (e.g. a conjugative plasmid) that is unable to replicate vegetatively in the new host, expression of the TA module without transposition to a stable replicon would lead to loss of the transposon and consequent cell death, whereas cells in which transposition had occurred would survive and give rise to a new population in which all cells would contain the Tn. This might be seen as a “take me or die” mechanism , a notion which could be explored experimentally.

Fig. Tn3.18E. Working Model for the Integration of TA Activity into the Transposition Process. A hypothetical Tn3 family transposon carrying a TA gene pair is shown. i) Homeostasis on a plasmid stably established in the cell. Transcription (orange and blue dotted wavy line) occurs from a putative endogenous TA promoter (P, proximal to TA) and maintains low toxin (T) and antitoxin (A) levels to maintain the vector plasmid in the host cell population. Expression of tnpA and tnpR from the res site promoters is largely repressed by TnpR binding. However, readthrough transcription from the TA gene pair into tnpA would be expected to result in a level of background TnpA expression. ii) Stochastic expression If the TA genes do not have an endogenous promoter, stochastic expression (blue and orange dotted wavy lines) from the divergent res promoters (P, within the res site) would result in low TnpA and TnpR levels as well as low level TA expression. iii) Plasmid conjugation into a recipient cell resulting in derepression of the res promoters results in higher levels of tnpA, tnpR, and TA transcription (blue and orange wavy lines) and expression of TA proteins resulting in an increased level of “addiction”.

Conclusion and Future.

The Tn3 family is widely spread and diverse as we have underlined and illustrated here. There is some understanding of the different evolutionary pathways and mechanisms which have permitted family members to sequester a large set of passenger genes widely variable functions and to shuffle them between and within both plasmids and chromosomes. Although much is known about a number of model Tn3 family members, there remain a number of open questions. In particular, historically this has proved recalcitrant to analysis in spite of much effort from their discovery in the 1970s to the present day.

Recent studies with Tn4330 however may have unlocked a door to understanding Tn3 family transposition in molecular detail. The structural studies using cryo-em have provided precious information as to the location and function of a large number of domains in the exceptionally long transposases. The studies point to the way in which the transpososome may be assembled although additional analyses are essential to a full understanding of docking of target DNA and its place in the assembly pathway. In addition, it is at present unclear how duplication of this family occurs during transposition: how it may recruit replication enzymes, whether replication initiates from one particular end, or, indeed whether it involves parasitizing existing replication forks in the target. The phenomenon of immunity is also not understood although it is clear that, mutationally, it is linked to transposition activity. In the absence of an ATPase activity, it seems unlikely that it occurs with the same mechanism as does that of bacteriophage Mu or transposon Tn7.

---