|Home | About | Journals | Submit | Contact Us | Français|
Four transposition proteins encoded by the bacterial transposon Tn7, TnsA, TnsB, TnsC, and TnsD, mediate its site- and orientation-specific insertion into the chromosomal site attTn7. To establish which Tns proteins are actually present in the transpososome that executes DNA breakage and joining, we have determined the proteins present in the nucleoprotein product of transposition, the Post-Transposition Complex (PTC) using fluorescently labeled Tns proteins. All four required Tns proteins are present in the PTC in which we also find that the Tn7 ends are paired by protein-protein contacts between Tns proteins bound to the ends. Quantification of the relative amounts of the fluorescent Tns proteins in the PTC indicates that oligomers of TnsA, TnsB, and TnsC mediate Tn7 transposition. High-resolution DNA footprinting of the DNA product of transposition attTn7Tn7 revealed that about 350 bp of DNA on the transposon ends and on attTn7 contact the Tns proteins. All seven binding sites for TnsB, the component of the transposase that specifically binds the ends and mediates 3’ end breakage and joining, are occupied in the PTC. However, the protection pattern of the sites closest to the Tn7 ends in the PTC are different from that observed with TnsB alone, likely reflecting the pairing of the ends and their interaction with the target nucleoprotein complex necessary for activation of the breakage and joining steps. We also observe extensive protection of the attTn7 sequences in the PTC and that alternative DNA structures in substrate attTn7 that are imposed by TnsD are maintained in the PTC.
Transposable elements are discrete DNA segments that can move between non-homologous positions within genomes. They are present in virtually every genome that has been examined and in some instances form a substantial fraction of the genome; for example, at least 45% of the human genome is derived from transposable elements1. Transposable elements also have substantial impact on bacterial genomes2.
Element-encoded transposases bind specifically to transposon ends and mediate the catalytic steps in transposition, cleaving the phosphodiester bonds that link the transposon to the donor site and joining the transposon ends to the insertion site3; 4; 5. The transposase often acts in concert with other element- and/or host-encoded proteins to assemble the nucleoprotein complexes called transpososomes that mediate key steps in transposition such as pairing of the transposon ends and selection of a target site6; 7; 8. The proper nucleoprotein architecture of these complexes is central to the activation and coordination of the DNA breakage and joining events that underlie transposition.
Tn7 transposition occurs by a “cut & paste” mechanism in which the transposon is excised from the donor site and integrated into the target site9; 10. This process is tightly regulated by the recognition of the target site and assembly of a TnsABC+D transpososome: no excision of Tn7 from the donor site occurs unless attTn7 and all the Tns proteins are present11. Using DNA crosslinking, we have previously identified nucleoprotein complexes containing both a donor DNA from which Tn7 has not yet been excised and a target DNA whose formation requires all the Tns proteins12. The architecture of the transpososome that mediates DNA breakage and joining must result in the juxtaposition of three DNA sites, the Left (Tn7.L) and Right (Tn7.R) ends of the transposon and a target DNA. We show here that all the Tns proteins required for insertion into attTn7 are still present in the nucleoprotein complex in which breakage and joining have occurred.
Tn7’s transposition machinery is especially elaborate9; 10. Unlike simpler systems that utilize one or two transposition proteins13, four Tn7-encoded proteins, TnsA, TnsB, TnsC, and TnsD, are required to insert Tn7 into a specific site in bacterial chromosomes called attTn711. Tn7.L and Tn7.R both contain multiple specific binding sites for the transposase subunit TnsB but these sites are differently positioned on each end, making the ends structurally asymmetric14; 15. The ends of Tn7 are also functionally asymmetric15; 16 elements with two Tn7.Rs transpose whereas those with two Tn7.Ls do not. Furthermore, Tn7 insertion into attTn7 downstream of the highly conserved glmS gene is orientation-specific as well as site-specific17.
attTn7 is chosen as a specific site for Tn7 insertion by the binding of TnsD to a particular sequence at the end of the highly conserved glmS gene11; 18; 19. The exceptional conservation of the TnsD binding site within the region of the glmS encoding the GlmS active site provides a specific insertion site for Tn7 in virtually all prokaryotic genomes that have been examined20. Similar sequences in the human glmS homologs provide specific sites for the high frequency insertion of Tn7 insertion in vitro19; 21.
TnsA and TnsB together form the transposase that carries out Tn7 DNA breakage and joining22; 23; 24. The TnsAB transposase is not, however, constitutively active: transposase activity is controlled by interaction of TnsAB with TnsCD bound to attTn711. Selection of attTn7 as the target DNA is initiated by TnsD binding to its 30 bp binding site, which generates distortions in attTn711; 18; TnsC is recruited to this distorted DNA25. The interaction of TnsC with TnsD-attTn7 is ATP-dependent11 and leads to the regulation of transposase activity via TnsC interactions with both TnsA and TnsB12; 26; 27; 28.
At what stages of transposition are the various Tns proteins required? For example, does TnsD simply serve as an "assembly" factor that dissociates from the rest of the transposition machinery once TnsC is loaded onto the attTn7 target DNA as is seen for the targeting protein UvrA protein within the Nucleotide Excision Repair pathway29. We report here our analysis of the architecture of the Tn7 transpososome in which transposition has occurred, the Post-Transposition Complex (PTC). We demonstrate that the PTC is an elaborate nucleoprotein complex in which the Tn7 ends are paired and contains all four Tns proteins that are required to execute transposition, TnsABCD, in addition to the DNA product of transposition, the Simple Insertion DNA attTn7Tn7. We also characterize the extensive Tns protein-DNA contacts within the PTC.
After Tns-mediated DNA breakage and joining, conversion of the attTn7Tn7 transposition product to intact duplex DNA requires repair by host proteins of the single-strand gaps that flank the 5’ ends of the newly inserted transposon. These single-strand gaps result from the attack of the transposon’s 3’ ends on staggered positions in the target DNA. We find that such repair cannot occur on the PTC in vitro. Thus host proteins are also likely involved in disassembly of the PTC to remove the Tns proteins such that host DNA repair machinery can access the newly inserted transposon. Host proteins have been shown to remodel the Mu transpososome after transposition to allow DNA replication and repair30; 31; 32.
We carried out transposition reactions using a supercoiled Tn7 donor plasmid and a small linear attTn7 target fragment as DNA substrates and then used multiple cycles of size filtration and washing to isolate the stable PTCs from unincorporated Tns proteins (Fig 1a). In contrast to pre-transposition complexes that are detected only upon crosslinking12, the PTC is stable in the absence of crosslinking, indicating that an increase in stability of Tns-DNA complexes accompanies Tn7 transposition. We used Atomic Force Microscopy (AFM) to directly examine the architecture of PTCs (Fig 1b). Two types of PTCs were present that contained circular DNAs with measured lengths appropriate to be the Tn7 element with a single Tns protein complex holding the attTn7Tn7.L and Tn7.RattTn7 ends together. One type of PTC had open circular, i.e. relaxed, DNA, and the other had twisted, i.e. supercoiled, DNA. The flanking donor DNA was not contained in the PTCs.
We also visualized PTCs isolated by size filtration by gel electrophoresis and directly characterized the PTC Tns and DNA components. The excision of Tn7 from the donor site and its insertion into attTn7 requires TnsABCD9; 10; 33. To identify the protein components of the PTC, we generated fluorescently labeled derivatives of all four Tns proteins by Expressed Protein Ligation34 that added a short peptide containing fluorescein (green) to the carboxylterminus of each protein (Tnsf); the specific activities of the labeled proteins in transposition were equivalent to those of unlabeled proteins (data not shown). We carried out transposition reactions using the Tnsf proteins, a supercoiled donor plasmid containing a Tn7 element and an attTn7 fragment, either unlabeled or end-labeled with a different fluorophore, Alexa Fluor 555 (red). After transposition, reactions were digested with SmaI that cuts the Tn7 element between the Tn7.L and Tn7.R ends, the PTCs purified by filtration, and then displayed on non-denaturing gels. Schematic diagrams of the substrates and products of transposition including the χ–PTC that results from SmaI digestion are shown in Fig 2a.
The four panels in Fig 2b display reactions in which one of the Tns proteins is fluorescently labeled. In each panel, lanes 1–3 display reactions with unlabeled attTn7 and lanes 4–7 display reactions with attTn7f; the reactions in lanes 2 and 5 were deproteinized while those in lanes 1 and 4 were not as indicated; Tns proteins were omitted in lane 3 and 9. Fluorescence was used to track the Tnsf proteins (green) and attTn7f (red); merging of coincident Tnsf and attTn7f signals results in a yellow signal. Tn7 end-containing fragments were identified by hybridization to 32P-labeled Tn7.L (lanes 4–6) or Tn7.R (lanes 7–9) probes following southern blotting of the gels (Fig 3).
The χ–PTC isolated by size filtration and gel electrophoresis (Fig 2,,3),3), is observed with each Tnsf protein and thus contains all four Tns proteins, TnsABCD. The χ–PTC also contains attTn7Tn7 as indicated by the merged (yellow) signal of the Tnsf proteins (green) and attTn7f (red) (Fig 2 lane 4). The χ–PTC seen without deproteinization also contains Tn7.L (Fig 3 lane 5) and Tn7.R (lane 8), although the ends are no longer covalently linked because of cleavage of the Tn7 element by SmaI. Thus, consistent with the AFM analysis shown above, the ends of Tn7 are paired within the χ-PTC. In the absence of SmaI digestion, relaxed and supercoiled PTCs were also observed by gel electrophoresis (data not shown). Equivalent amounts of recombinant attTn7Tn7.L and Tn7.RattTn7 products are present in the PTC when reactions are deproteinized. (Fig 2 lane 5 and Fig 3). Several conditions contribute to the smears below the χ-PTC. In reactions with TnsAf and TnsCf (Fig 2,,3),3), a TnsACD-attTn7 complex 26; 35 is detectable. Also, some dissociation of the Tn7 ends does occur as evidenced by the Tn7.L (Fig 2, all panels, lane 4) and Tn7.R (lane 5) smears and these ends are bound by TnsBf (Fig 2, TnsBf panel, lane 1).
We estimated the relative amounts of each of the Tns proteins in the PTC by comparing Tnsf fluorescence to the amount of end-labeled 32P attTn7Tn7. After in vitro transposition, we filter-purified the PTCs, crosslinked them to decrease protein dissociation during electrophoresis and displayed them by native agarose electrophoresis. The amount of fluorescence and 32P was quantified for PTCs formed with each of the Tnsf proteins and ratios of the Tnsf/ attTn7 measurements compared (Table 1).
We believe that the observed TnsDf/32P attTn7 ratio of 0.12 reflects the presence of 1TnsD in each PTC. TnsD binds specifically to a site in attTn7 that lies only to one side of the point of Tn7 insertion18 and DNA footprinting shows that this site is occupied in the PTC (see below). Moreover, there are no symmetrical features within the TnsD binding site. These observations are consistent with the suggestion that one TnsD protein binds to attTn7.
We then compared the TnsDf/ 32P attTn7 value to the Tnsf/ 32P attTn7 values for all the other Tnsf proteins. The value of the TnsDf/ 32P attTn7 was consistently much lower than that for any other Tns protein in the PTC, suggesting that there are multiple copies of the other Tns proteins in the PTC. There are 7 TnsB binding sites on the ends of Tn7. The PTC TnsBf/ 32P attTn7 value was 6.3 fold higher than that for TnsDf/ 32P attTn7, a value consistent with the presence of 7 TnsBs in each PTC. All 7 TnsB binding sites in the PTC are protected in DNA footprinting experiments (see below).
The ratios of TnsAf and TnsCf to 32P attTn7 in the PTC were much higher than for TnsD; the TnsAf/ 32P attTn7 value was 18 fold-higher than TnsDf/ 32P attTn7 and the TnsCf/ 32P attTn7 value was 25 fold-higher than TnsDf/ 32P attTn7. Thus the PTC contains more TnsA and TnsC than the 1 TnsD and 7 TnsBs suggested above. Previous work has shown that TnsA and TnsC form a TnsA2C2 complex36. DNA footprinting data is consistent with the suggestion that multiple TnsA2C2 complexes are present in the PTC (see below). The lower value for Tnsf/ 32P attTn7 DNA for TnsA (18) compared to TnsC (25) likely reflects preferential loss of TnsA from the PTC during filter purification35.
Where do the multiple Tns proteins in the PTC contact the Simple Insertion product attTn7Tn7? To examine these protein-DNA contacts at base pair resolution, we carried out DNA-footprinting with line graph analysis using hydroxyl radicals, KMnO4, and DMS as attack reagents. PTC footprinting was compared to that of naked Simple Insertion DNA of identical sequence but lacking the single-strand gaps at the transposon ends that are present in the PTC. By using attTn7 substrates radiolabeled at either 3' end of attTn7 DNA, we evaluated protection of the Tn7 L and R 3’ ends and attTn7 strands that are covalently joined in the Simple Insertion DNA. Fig 4 shows hydroxyl radical footprinting gels and line traces and Fig 5 shows KMnO4 and DMS footprinting gels for analysis of protection the Tn7 ends. The results of these footprinting experiments are summarized at the nucleotide level in Fig 6.
Large segments of Tn7.L (Fig 4a) and Tn7.R (Fig 4b) end sequences within the PTC are protected from hydroxyl radical attack, indicating extensive contacts between the Tn7 ends and the Tns proteins. Most of the end protection in the PTC reflects the binding of the transposase subunit TnsB to its specific binding sites in the transposon ends. When examined alone with individual Tn7 ends, TnsB (purple circles) binds to seven sites, three in Tn7.L including the terminal α site, and the internal β and γ sites, and four in Tn7.R including the internal sites , χ, ψ, and the terminal ω site14; 15. The presence of all these sites in a Tn7 element is required for wild-type transposon activity16. In the PTC, the attTn7Tn7 DNA is protected at all seven TnsB binding sites, indicating that transposition does not involve the loss of TnsB from these sites.
There are, however, several striking differences between the hydroxyl radical protection patterns observed in the terminal TnsB binding sites of the PTC and the previously observed TnsB-only protection of the individual Tn7 ends that likely reflect changes in the transpososome when it is activated for breakage and joining. One PTC-dependent difference is that the terminal Tn7.L α and the terminal Tn7.R ψ and ω sites are more completely protected in the PTC (Figs. 4, dark blue brackets, yellow highlights). On the transferred strands in which the 3’ ends of Tn7 are covalently linked to attTn7, TnsB-only binding on individual Tn7 ends results in characteristic regions of protection designated I, III and V within the TnsB binding sites, separated by regions of susceptibility to hydroxyl radical attack. In the PTC, however, in the terminal Tn7.L α and the terminal Tn7.R ψ and ω sites, the susceptible regions that are between the TnsB domains I, III and V are not cleaved by hydroxyl radicals, indicating a more extensive nucleoprotein complex at the termini. Another PTC-dependent difference is that the -ACAs at the 3’ ends of Tn7 that join to the target DNA are more completely protected, likely reflecting the positioning of the active transposase at the Tn7 ends.
In addition to increased protection at Tn7 termini within the PTC, we also find regions of strong protection immediately internal to the terminal Tn7L TnsB α site on the internal edge of the site between the α and β sites (Fig. 4, dark blue bracket and yellow highlight) and also weaker protections between the internal Tn7L β and γ sites (Fig. 4, light blue brackets and yellow highlight).
We also used KMnO4 to probe the structure of Tn7.R within the PTC (Fig. 5a and summarized at the nucleotide level in Fig. 6). Hyper-reactivity to KMnO4 results from distortions in DNA structure that lead to an increased accessibility to KMnO418. Several positions of KMnO4 hyper-reactivity are observed in the internal Tn7.R TnsB sites and χ at thymidine nucleotides flanking binding region III of these sites at R51, R58, and R73 (Fig 5a, red triangles). Despite similarly positioned thymidine nucleotides in the Tn7.R ψ and ω sites, no such KMnO4 hyper-reactivity is observed at these sites, thus providing additional support for the view that the protein environment of the terminus of Tn7.R is distinct from that of the interior regions.
Tn7 insertion occurs by the attack of the 3’OH ends of Tn7 on staggered positions on the top and bottom strands in attTn7, resulting in single strand gaps at the ends of Tn7. It is notable that in the PTC, the single-stranded gap that flanks the 5’ end of Tn7 at the Tn7.RattTn7 junction is also protected from hypersensitivity to KMnO4 whereas in the deproteinized Tn7.RattTn7 transposition product, the single-strand gap was hyper-cleaved in the presence of KMnO4 (Fig 5a, blue triangles).
The middle bp of the 5bp attTn7 target sequence duplicated upon Tn7 insertion is designated "0", sequences that lie rightward towards glmS are "+" and those leftward are "−"; thus the Tn7 target site duplication is attTn7 −2 to +2. The target information required to promote Tn7 insertion at attTn7 −2 to +2 is contained within the TnsD binding site, which is located at about attTn7 +25 to +55. There is no evidence that any particular DNA sequence is required to the left of the TnsD binding site in attTn7 including that at the actual point of insertion11; 19; 37.
DNA-footprinting analysis using hydroxyl radicals of the PTC revealed strong and extensive protection of the attTn7 sequences from attTn7 −58 to Tn7.L (Fig 7) and from Tn7.R to attTn7 +55 (Fig 5b), including the single-strand gaps at the transposon ends. Thus the proteins within the PTC strongly interact with 114 bp of attTn7 DNA sequences. As no particular sequence leftwards of the TnsD binding site at attTn7 +25 to +55 is required for transposition11; 18; 19; 37, it is unclear why the left border of protection in the PTC stops sharply at attTn7 −58,
Our DNA-footprinting analysis with DNA distortion-sensitive probes (Fig 5) also supports the view that TnsD remains bound to attTn7 in the PTC. TnsD binding to the pre-transposition substrate attTn7 results in DNA distortions at attTn7 +27 as evidenced by KMnO4 hyper-reactivity and DMS hyper-reactivity at attTn7 +4518. These distortions are still present in the PTC, confirming the presence of TnsD and the TnsD-induced attTn7 distortions.
The 5’ ends of Tn7 in the Simple Insertion product attTn7Tn7 are flanked by single-strand gaps that require repair by host factors to generate intact duplex DNA. We find that when the PTC is deproteinized, repair of these gaps is readily observed upon the addition of DNA polymerase I and DNA ligase to attTn7Tn7 DNA as evaluated by denaturing gel electrophoresis showing that the top 5’ strand of Tn7.L becomes covalently linked to the flanking attTn7 target (Fig 8, lanes 1). When Tn7 inserts into attTn7 to form attTn7Tn7, the gap sequence adjacent to the top 5’ strand of Tn7.L which must be synthesized is 5’-GGGCG. If DNA synthesis proceeds into the 5’ end of Tn7.L, the newly synthesized DNA extending from the gap into the end would include 5’ TGT. However, efficient repair is still observed when dA or dT is omitted from the repair reaction (lanes 2 and 3), indicating that repair synthesis need not proceed into the end of Tn7 to repair the transposition-generated gap.
Repair of the gap flanking Tn7.L of the Simple Insertion DNA in the PTC is not observed without deproteinization (lanes 4–6), revealing that the Tns proteins block access of the repair enzymes to the flanking gap. The fact that this gap remains refractory to repair even in PTCs that have been incubated for several days at 4°C indicates the stability of the PTC (data not shown). Thus a key step in Tn7 transposition must be the disassembly of the PTC to allow repair. Incubation of the Tn7 PTC in E. coli crude extracts can remove the block to DNA repair observed with PTCs, suggesting that host factors can actively promote disassembly of the PTC to allow repair (J. Holder, C. Johnson, and NLC unpubl. observations).
Successful transposition requires specific recognition and juxtaposition of the transposon ends and the target DNA, and careful coordination of the DNA breakage and joining events, avoiding non-productive DNA breaks at the transposon ends and promoting equivalent joining of the transposon ends to the target DNA. These reactions occur in nucleoprotein complexes called transpososomes6; 8. We have defined the components and architecture of the transpososome product of Tn7 transposition, the Post-Transposition Complex (PTC), which mediated the site-and orientation-specific insertion of Tn7 into the attTn7 target DNA to form the Simple Insertion product attTn7Tn7.
We have found that the Tn7 ends are stably paired within the PTC such that supercoiling of the Tn7 substrate can be maintained within the PTC. Such stable end pairing must be key to the coordination of Tns protein activity on both transposon ends and has also been observed in other transposition systems38; 39 The fact that PTCs including those in which the Tn7 segment has been cut by a restriction enzyme (χ-PTCs) can be observed after filter purification and gel electrophoresis in the absence of stabilization by crosslinking attests to their stability. By contrast, pre-transposition Tn7 donor/ target complexes can be observed only in the presence of crosslinking12. Increasing transpososome stability provides a mechanism for guiding transposition to convert the Tn7 donor and attTn7 target substrate DNAs to the attTn7Tn7 product as has been suggested for other transposition systems8; 40.
In addition to promoting breakage at the transposon ends and joining to the target DNA, another key role of transpososomes is to prevent side reactions that would convert the transposition product to a form that cannot be repaired to form an simple insertion containing intact duplex DNA41; 42; 43 .
The production of a highly stable PTC presents a challenge to the cell. We have found that the PTC is refractive to the repair reactions necessary to convert the single strand gaps at the Tn7 ends in the actual transposition product to duplex DNA, suggesting that active disassembly of the PTC occurs in vivo. Active disassembly of the highly stable MuA tetramer that is the core of the Mu transposition machinery has been demonstrated to play a key role in the completion of Mu transposition44.
A key finding of this work is that the Tn7 PTC contains the four Tns proteins that are essential for transposition into attTn7, TnsABCD, and that multiple copies of TnsA, TnsB and TnsC are present in the PTC. We have also defined the extensive, nearly 360 bp, interactions between the Tns proteins and the DNA product of transposition attTn7Tn7. The extensive protein-DNA interactions both at the ends of Tn7 and on the attTn7 sequences reflect the complexity of the Tn7 DNA substrates and their interaction with many Tns proteins required to promote high-frequency site- and orientation-specific Tn7 insertion into attTn7.
TnsB is the transposase subunit that mediates DNA breakage and joining at the 3’ ends of Tn723 and transposition requires multiple TnsB binding sites in both ends of Tn714; 15; 16. Using DNA footprinting, we have found that all 7 of these TnsB binding sites are occupied in the PTC. The stoichiometry of TnsB in the PTC that we have determined, 6.3 TnsBs/ attTn7, is consistent with these protection results.
There are, however, significant differences in the Tns protein interactions with the ends of Tn7 in the PTC in which both pairing of the Tn7 ends and their juxtaposition to attTn7 has occurred as compared to the interaction of TnsB alone: notably, protection is increased within the terminal Tn7L TnsB site α and on the Tn7.R TnsB sites ψ and ω. Moreover, in the PTC, DNA protection extends to the 3’ Tn7 ends where TnsB-mediated breakage and joining occurs. These changes in the interaction of the Tns proteins with the ends of Tn7 must reflect the TnsB-mediated pairing of the Tn7 ends12, the positioning of the TnsBs at the 3’ ends for cleavage, and likely additional end interaction by TnsA, the other subunit of the transposase that promotes cleavage at the 5’ termini23; 24.
TnsB is a member of the Retroviral Integrase superfamily of transposases which includes the well-characterized Tn5/ Tn10 transposases whose catalytic centers are RNaseH-like domains4; 5. In the Tn5 system, one transposase protomer and hence one active site is positioned at each transposon end and mediates all the DNA breakage and joining events at that end45; 46; 47. An attractive hypothesis is that, similarly, one TnsB bound at each 3’ end mediates the DNA breakage and joining events at that end and that one TnsA mediates the breakage events at each 5’ end. In this view, TnsBs bound at the internal TnsB binding sites are thus likely involved in positioning the catalytic machinery at the point of insertion and imposing the orientation-specific insertion of Tn7, presumably by interaction with an asymmetric component of a target complex (see below).
Tn7 is recruited specifically to attTn7 by the specific binding of TnsD to a sequence within attTn711; 18. A reasonable question is whether TnsD remains in the pre-transposition complex once the TnsABC machinery has been positioned on attTn7 or does it simply serve as an "assembly" factor. We have found that TnsD remains bound to its recognition site on attTn7 in the PTC. The continued presence of TnsD and the distortions in attTn7 DNA structure it provokes may be necessary for asymmetric activation of TnsABC to promote orientation-specific insertion of Tn7 into a unique position in attTn7 (below).
TnsC is positioned on the insertion site by interaction with DNA distortions imposed on attTn7 by TnsD18 and perhaps through protein-protein interactions between TnsC and TnsD (R. Mitra and N.L. Craig unpubl.). TnsC also interacts with both transposase subunits, TnsA and TnsB,12; 26; 27 (see below).
A striking feature of the Tn7 PTC is the extensive protection of attTn7 sequences, extending from its leftward border at attTn7 −58 to its rightward border at attTn7 +55. TnsD binding to attTn7 +25 to +55 accounts for some of this protection. We suggest that the extended region of protection leftward of the TnsD binding site in the PTC reflects the interaction of TnsA and TnsC, likely as multiple TnsA2C2 complexes which have been previously identified36. This would account for the much higher amounts of TnsA and TnsC in the PTC compared to TnsD (1 TnsD/ PTC) and TnsB (7 TnsB/ PTC), being (at least) 18 TnsA/PTC and 25 TnsC/ PTC. It remains to be determined why there is a sharp border of protection at attTn7 −58 as no specific sequences between attTn7 −58 and attTn7 +25 are required for transposition.
We have previously observed the formation of distinct TnsD-attTn711. TnsCD-attTn718 and TnsACD-attTn7 complexes26; 35 and have argued that TnsC interacts with both the target DNA11; 25 and the TnsAB transposase48. We suggest here that a TnsA2C2 complex is positioned at the point of Tn7 insertion in attTn7 by interaction with TnsD-attTn7 to form a target complex that recruits TnsB bound to the Tn7 ends. There are multiple protein-protein interactions between TnsA, TnsB and TnsC: TnsA interacts with TnsC36; 49. TnsB interacts with TnsC28; 49 and TnsA and TnsB likely interact because TnsA is required to provoke TnsB-mediated 3’ end breakage and joining in a minimal TnsAB-only recombination system22. TnsA and TnsB form the transposase, TnsA mediating DNA breakage at the transposon 5’ ends and TnsB mediated breakage and joining at the 3’ transposon ends23; 24. Notably, breakage and joining is seen only in the presence of both TnsA and TnsB11. Thus only when a TnsACD-attTn7 target complex interacts with a TnsB-end complex is the TnsAB transposase formed and activated.
The molecular mechanism by which apparently only the TnsA2C2 complex at the insertion site can capture the TnsB-bound ends for insertion in highly site- and orientation-specific insertion remains to be determined but it seems likely that its asymmetric interaction with TnsD-attTn7 including the distortions introduced by TnsD into attTn7 will be key in this process18; 25. Perhaps the interaction of TnsD with one of the C subunits of a TnsA2C2 complex at the point of insertion defines an asymmetric target site.
A striking observation from this work is that almost 60 bp of target DNA leftward to Tn7L in attTn7Tn7 are protected within the PTC. This is surprising because we have demonstrated that no specific DNA sequences are required in this region of a target DNA18. We suggest that this region is also bound by role of TnsA2C2 complexes beyond the point of insertion but their role remains to be determined. Perhaps they mediate interactions with TnsBs that determine including orientation- and site-specific insertion. If multiple TnsA2C2s are present in pre-transposition complexes, they may provide for a multi-valent landing site for the TnsB-Tn7 ends that can channel the ends to the point of insertion.
The simplest model that accounts for the chemical steps of Tn7 transposition is that one TnsB bound at each end of Tn7 mediates 3’ end breakage and joining and one TnsA bound at each end mediates 5’ end breakage and that this TnsAB transposase is part of a TnsA2B2C2 complex that can be positioned at a specific site by interaction with TnsD. However, our studies of the Tn7 PTC have revealed a far more complex and elaborate transposition machine. It seems likely the additional TnsBs and TnsA2C2 complexes that we have established are in the PTC play important roles in Tn7 end identification, target site selection and controlling and directing the highly site- and orientation-specific insertion of Tn7 insertion into attTn7. The elaborate architecture of the PTC we have described here has deepened our understanding of the control of Tn7 transposition and raised new questions for future investigations.
The tnsA, tnsB, tnsC and tnsD genes were cloned into pCYB1 (New England Biolabs = NEB) to express Tns-Intein-Chitin Binding Domain fusion proteins for affinity purification on chitin beads. 1 liter cultures of CAG45623 containing Tns-Intein-chitin binding domain fusion proteins in the pCYB1 vector were grown at 30°C to an OD of 0.5, and cooled on ice for 20 minutes prior to induction with 400 µM IPTG for 18 hours at 16°C. Cells were harvested by centrifugation and resuspended in 10 ml Lysis BufferTnsABD = 25 mM HEPES pH 8.0, 10 % glycerol, 0.1 mM EDTA, 500 mM NaCl, 0.1% Tween 20 or Lysis BufferTnsC = 25 mM HEPES pH 8.0, 10 % glycerol, 1 mM EDTA, 1 M NaCl, 1 mM ATP. Cells were lysed in a French Press with pressure in excess of 1000 psi and then cell extracts were spun at 100,000 × G for 20 minutes. The resulting supernatant was filtered through a 0.45 µM filter (Millipore) prior to mixing with 1 ml Lysis Buffer-equilibrated chitin beads (NEB) in a capped and sealed 20 ml column (BioRad), followed by gentle shaking at 4°C for 20 minutes. The chitin beads were allowed to settle, the column allowed to drain and washed with 4 × 20 ml Lysis Buffer. For TnsA, TnsB, and TnsD, the beads were resuspended with shaking for 60 minutes at room temperature with 10 ml Release Buffer = 20 mM HEPES pH 8.0, 0.1% Tween 20, 10 % glycerol, 500 mM NaCl, 10 mM MgCl2, 10 mM ATP to remove E. coli chaperone proteins. The chitin beads were then allowed to settle, the column allowed to drain and washed with 10 ml Lysis Buffer.
To generate fluorescently labeled Tns proteins, a fluorescent peptide CKH6 labeled on the ε amino group of the lysine with fluorescein and HPLC purified (Sigma-Genosys) was ligated to each Tns protein by Expressed Protein Ligation34. Intein-mediated cleavage of the Tns fusion protein is coupled to peptide incorporation at the C-terminus of the Tns proteins. A 1 ml bed of chitin beads bound by Tns fusion protein was incubated with 500 µM peptide and 200 mM sodium 2-mercaptoethanesulfonate (MESNA) pH 8.0 (Sigma-Aldrich) in Lysis Buffer overnight at 4°C in the dark. The beads were then allowed to settle, the column drained and washed with 3 ml Lysis Buffer to yield a final eluate of 4 ml containing Tnsf and unincorporated peptide. Heparin chromatography was used to separate TnsAf, TnsBf, and TnsDf from the peptide. Peptide was removed from TnsCf by ammonium sulfate precipitation of TnsC (30 % saturation = 178 mg/ml) and washes in TnsC Low Salt buffer as previously published50. The Tnsf protein was at least 70% of each preparation and had the same specific activity as Tns protein (data not shown).
TnsAf is stored in 25 mM HEPES pH 8.0, 0.1% Tween 20, 10 % glycerol, 150 mM NaCl, 1 mM DTT. TnsBf is stored in 25 mM HEPES pH 8.0, 0.1% Tween 20, 25 % glycerol, 500 mM NaCl, 1.0 mM EDTA, 1 mM DTT. TnsCf is stored in 25mM HEPES pH 8.0, 0.1 mM EDTA, 1.0 M NaCl, 2.5 mM DTT, 10 % glycerol, 10 mM CHAPs, 10 mM MgCl2, 1 mM ATP. TnsDf is stored in 25 mM HEPES pH 8.0, 500 mM NaCl, 1 mM EDTA, 2 mM DTT, 25 % glycerol.
TnsA, TnsB, TnsC, and TnsD proteins were purified as previously described19 or TnsC, and TnsD were derived from purified Intein fusion proteins as for Tnsf proteins except that 40 mM DTT was used for cleavage.
Tn7 donor plasmids used as indicated in the Figure legends were pEMΔ (5.9 kB) that contains a 1,625 bp Tn7 element with 166 bp Tn7.L and 199 bp Tn7.R in a donor backbone of 4301 bp and pPK21 (3.2 kB) that contains a 373 bp Tn7 element with 166 bp Tn7.L and 199 bp Tn7.R in a donor backbone of 2785 bp19. attTn7 target fragments used as indicated in the Figure legends were generated as described below. The target attTn7 plasmid used in Fig 8 was pRM2.
The 244 bp attTn7 −123 to +120 target fragment used in the AFM analysis of Figure 1 was generated by PCR using pRM2 as a template with oligonucleotides NLC1193 5’-agccggaattcgacagaaaattttcattctg and NLC1165 5’-ctttgatcagcgcgacatggtaaggcggccgcattcttat, followed by gel purification.
The 107 bp attTn7 −38 to +68 used in the experiments of Figures 2 and and33 was generated by PCR using pRM2 as a template with oligonucleotides NLC 1456 5’-tccaaagccggaattcttgattaaa-aacataacaggaag and NLC 1667 5’-gccaggttacgcggctggtgaat. The PCR fragment was then digested with EcoRI at the two sites that flank the attTn7 sequences and the resulting 3’ ends filled in with dA and dU-Alexa Fluor 555 using the ARES DNA labeling Kit (Molecular Probes) followed by gel purification.
The target used in the experiment of Table 1 was the 107 bp attTn7 −39 to +68 fragment generated by PCR with primers 1456, and 1667 using pRM2 as a substrate and was labeled by a 32P dATP incorporation following EcoRI digestion.
In the footprinting experiments of Figures 4, ,55 and and7,7, the 204 bp attTn7 −83 to +120 fragment bounded by HindIII and XbaI sites in pPK1318 was used as the target DNA for the generation of PTCs. For analysis of attTn7Tn7.R, the fragment was labeled at the 3’ end of the attTn7.R sequences (attTn7.R +120, XbaI) and for analysis of attTn7Tn7.L, the fragment was labeled at the 3’ end of the attTn7.L sequences (attTn7.L −81; HindIII). End-specific labeling was achieved by digestion of the plasmid at the end to be labeled, filling in with α-32P dATP using Klenow exo− DNA polymerase (NEB), the plasmid digested at the other end and then gel purified. In each case, the 3’ labeled attTn7 strand becomes covalently linked to the 3’ end of Tn7.
In the footprinting experiment of Figure 7 where the left boundary of PTC protection was determined, the 424 bp attTn7 −300 to +120 fragment bounded by NdeI and XbaI sites in pPK1318 was used as the target DNA for the generation of PTCs. The 424 bp fragment was labeled at the 3’ end of the attTn7.L sequences by digesting with NdeI and filling in with dTTP and 32P α-dATP using Klenow exo− DNA polymerase (NEB), the plasmid digested with XbaI and the labeled attTn7 fragment purified using a PCR clean up kit (Qiagen).
A repaired Simple Insertion DNA containing the same sequences as the gapped simple insertion DNA product in the PTC that was used as the control in the DNA footprinting experiments was generated by in vitro transposition and cloning. The 204 bp fragment containing attTn7 −83 to +120 was isolated by digestion of pPK13 with HindIII and XbaI and used as a target in an in vitro transposition reaction using as a donor plasmid pPK21 which contains the 373 bp Tn7 element. The Simple Insertion product attTn7 HindIII-XbaITn7-373 bp was isolated from a gel, purified by Qiagen treatment, ligated to pPK13 digested with HindIII and XbaI and the attTn7Tn7 products recovered by transformation creating plasmid pJH375. Digestion of the resulting plasmid with HindIII and XbaI yields a double-stranded, i.e. repaired, attTn7Tn7 fragment that contains the same sequences as are contained in the gapped Simple insertion DNA in the PTC as in Figure 4, ,55 and and77.
Transposition reactions were carried out as previously described19 except where indicated. Reactions were performed in 24 µl with 22 mM HEPES pH 7.6, 3.3 mM DTT, 375 ng herring sperm DNA (Roche), 1.7 mM ATP, 1.6 µl TnsA, 1 µl TnsB, 1 µl TnsC, and 3 µl TnsD. In Figure 2, the donor was 0.8 nM pEMΔ, the target was 11 nM Alexa Fluor 555-labeled attTn7 −38 to +68 and Tns proteins including Tnsf were 35 nM TnsA, 7 nM TnsB, 25 nM TnsC, and 3 nM TnsD. In Figures 4 and and5,5, the donor was 1.5 nM pPK21 and the 1.7 nM 32P-labeled targets were either attTn7 −83 to +120 for analysis of attTn7Tn7.L (Fig 4a) and Tn7.RattTn7 (Figs. 4b,,5)5) and attTn7 −300 to +120 for analysis attTn7Tn7.L (Fig 7) and the Tns proteins were 52 nM TnsA, 7 nM TnsB, 50 nM TnsC, and 28 nM TnsD.
Reaction mixtures were added after transposition to 400 µl Wash Buffer (25 mM HEPES 8.0, 150 mM NaCl, 1mM DTT, 1 mM ATP pH 8.0) in an equilibrated 100 kDa cut off microcon filter (Millipore). Samples were concentrated by spinning as directed by manufacturer, 400 µl of buffer added with mild agitation and respun. This cycle was repeated 5 times to a 20–30 µl final retentate that contains PTCs, donor, and flanking donor DNA that were separated during gel electrophoresis or on a mica grid for AFM.
Filter purified PTCs were resuspended in AFM buffer (25 mM HEPES pH 7.6, 25 mM KOAc pH 7.2, 5 mM DTT, 1 mM ATP, 2 mM MgOAc). 10–20 µl of sample was adsorbed onto a 12 mm disk of freshly cleaved mica for 5–10 minutes and rinsed with water. The liquid was blown off the mica with compressed gas and the sample allowed to dry. AFM imaging was performed in ambient Tapping Mode using a Nanoscope IIIa controller with a multimode AFM (Digital Instruments, Santa Barbara, CA). Conventional silicon cantilevers type TESP (Veeco Instruments, Santa Barbara, CA) were used.
In the experiments in Figure 2 and and3,3, filter-purified χ-PTCs were loaded in 10% glycerol onto a 19 cm long 1.0% 0.5 × TBE agarose gel (Life Tech) and run for 714 V-Hr. Gels to be blotted were transferred to Gene Screen (NEN) and southern blotted with oligonucleotide probes specific for the Tn7 ends, NLC 272 5’-attttcgtattagcttacgacgctacaccc for Tn7.L and NLC 925 5’-gttcagtttaagactttattgtc for Tn7.R, that had been phosphorylated at their 5’ends by T4 polynucleotide kinase (NEB) and γ 32P ATP.
Hydroxyl radical footprinting reactions were done by mixing 3 µl 3 mM Fe(NH4)2(SO4)2, 3 µl 6 mM EDTA, and 3 µl 30 mM Na-Ascorbate with 20 µl retentate containing filter-purified PTCs and allowed to react at room temperature. The reactions were quenched after 1 minute by addition of 1 µl βME; protein was stripped off the DNA by addition of EDTA and SDS to 10 mM EDTA/ 0.08% SDS and the samples incubated at 65°C for 10 minutes. Simple Insertion DNA was then purified by agarose gel electrophoresis and gel extraction (Qiagen). For KMnO4 footprinting, filter-purified PTCs, repaired SI DNA and gapped SI DNA were treated with 2.5 mM KMnO4 for 1 minute at 24°C in 20 µl total volume18; the DNAs were then stripped of protein, purified as described above and then treated with 10% piperidine for 20 minutes at 90°C. For DMS footprinting, filtered-purified PTCs and SI DNA in 20 µl reactions were treated with 0.2% dimethyl sulfate (DMS) for 30 seconds and then stopped by the addition of 50 µl 25 mM HEPES pH 7.6, 125 mM βME, 1 mM EDTA18. Proteins were stripped from the DNA, the DNA purified as described above and described and then treated with 0.1 M NaOH for 20 minutes at 90°C.
In the footprinting experiments of Figure 4, ,6,6, and and7b,7b, the DNA was displayed on an 8M urea 7% 29:1 acrylamide: bis-acrylamide gel (Explorer, JT Baker) run for 9375 V-Hr. In Figures 7a and supplementary figures 3 and 5 the DNAs were displayed on 8M urea 8% 29:1 acrylamide: bis-acrylamide gels run for 6868V-Hr. Footprinting gels were dried down on 3mM paper, exposed to a phosphor screen and visualized by phospho-imaging.
Gels were imaged on a Typhoon 9410 (Molecular Dynamics). 32P was evaluated by phosphorimaging. Quantification of DNA fragments in denaturing gels was done using line graph analysis performed with Image Quant 5.0. The intensity values for each lane of the gel were transferred into Microsoft Excel for graphing DNA-cleavage intensity as a function of distance from the well. Scans for attTn7f DNA labeled with Alexa Fluor 555 were done using a 532 nm laser for excitation and a 580BP30 emission filter; the emitted light passed through a photomultiplier tube PMT at 800V. The fluorescein dye contained on the Tnsf proteins using a 488 nm laser and a 520BP40 emission filter.
For the experiment in Table 1, after transposition incubation, PTCs were filter purified and then crosslinked with 0.01 % gluteraldehyde for 10 minutes and then quenched. They were then loaded onto a 1 % agarose gel and run for 714V-Hr at room temperature. TnsA, TnsB and TnsC, the amount of fluorescence was determined for supercoiled PTC only. The amount of fluorescence in both the relaxed and supercoiled PTCs was determined for TnsD. The DNA gel was imaged on a Typhoon 9410 for the fluorescein dye contained on the Tnsf proteins using a 488 nm laser and a 520BP40 emission filter; the amount of fluorescence for each protein was determined by subtraction of background fluorescence. After fluorescent imaging was performed, the agarose gel was dried down onto DEAE paper and exposed for phophorimaging. It should be noted that some TnsA can be lost from the PTC during purification of PTCs35 and thus the relative amount of TnsA is likely closer to the TnsC value. This measurement does not give absolute stoichiometry because of local quenching effects on the fluorescent emission that may vary for each fluorophores location within the PTC.
Tn7 donor plasmid was pEMΔ which contains a 1.6 kb Tn7 element with 166 bp Tn7.L and 199 bp Tn7.R51. The attTn7 target plasmid pRM2 contains attTn7 −342 to +16551. Gap repair was performed by addition of 2.5 units of DNA Polymerase I (NEB), 100 units of T4 DNA ligase (NEB), 200 µM dNTPs, and 10 mM MgOAc to 25 µl of Wash Buffer containing filter purified PTCs or deproteinized transposition reactions. Following repair, proteins were removed (Qiagen PCR clean-up kit) and DNA was then digested PstI (NEB). Digested repair reactions were displayed on a denaturing gel that was then Southern blotted with a Tn7.L-specific oligonucleotide.
We thank Robert Sarnovsky cloning of TnsC into pCYB1, Prasad Kuduvalli, Gregory McKenzie, Chuck Merryman, and Rupak Mitra for helpful discussions, Jan Hoh for assistance with AFM, Iva Ivanovska, Helen McComas and Patti Kodeck for assistance with manuscript preparation. This work was supported by NIH grants P01 CA16519 and RO1 GM076425 to NLC. NLC is an Investigator of the Howard Hughes Medical Institute.
Role of the Funding Source:
JWH and NLC conceived of and designed the experiments, JH performed the experiments and JWH and NLC wrote the paper. This work was supported by NIH grants P01 CA16519 and RO1 GM076425 to NLC and NIH Training Grant T32-GM07445 to the Biological Chemistry Molecular Biology Program at Johns Hopkins University. NLC is an Investigator of the Howard Hughes Medical Institute.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
All authors declare no actual or potential conflict of interest.