Characterization of the Tn5 transposase and inhibitor proteins: a model for the inhibition of transposition.

Tn5 is a composite transposon consisting of two IS50 sequences in inverted orientation with respect to a unique, central region encoding several antibiotic resistances. The IS50R element encodes two proteins in the same reading frame which regulate the transposition reaction: the transposase (Tnp), which is required for transposition, and an inhibitor of transposition (Inh). The inhibitor is a naturally occurring deletion variant of Tnp which lacks the N-terminal 55 amino acids. In this report, we present the purification of both the Tnp and Inh proteins and an analysis of their DNA binding properties. Purified Tnp, but not Inh, was found to bind specifically to the outside end of Tn5. Inh, however, stimulated the binding activity of Tnp to outside-end DNA and was shown to be present with Tnp in these bound complexes. Inh was also found to exist as a dimer in solution. These results indicate that the N-terminal 55 amino acids of Tnp are required for sequence-specific binding. They also suggest that Inh inhibits transposition by forming mixed oligomers with Tnp which still bind to the ends of the transposon but are defective for later stages of the transposition reaction.     

A high-throughput assay for Tn5 Tnp-induced DNA cleavage

Transposition causes genomic instability by mobilizing DNA elements. This phenomenon is mechanistically related to other DNA rearrangements, such as V(D)J recombination and retroviral DNA integration. A conserved active site architecture within the transposase/integrase superfamily catalyzes these distinct phenomena. The Tn5 transposase (Tnp) falls within this protein class, and many intermediates of the Tn5 transposition reaction have been characterized. Here, we describe a method for the rapid identification of Tn5 Tnp small molecule effectors. This high-throughput screening strategy will aid in the identification of compounds that perturb Tnp-induced DNA cleavage. This method is advantageous, since it identifies effectors that specifically inhibit catalysis without inhibiting Tnp-DNA binding interactions. Effectors identified using this method will serve as a valuable aid both in the isolation and characterization of metal-bound reaction intermediates and in co-crystallization studies involving the effector, Tnp and DNA, to identify the structural basis of the interaction. Furthermore, since Tn5 Tnp shares a similar active site architecture to other transposase/integrase superfamily members, this strategy and any effectors identified using this method will be readily applicable to these other systems.     

DNA sequence bias during Tn5 transposition

Transposition is one of the primary mechanisms causing genome instability. This phenomenon is mechanistically related to other DNA rearrangements such as V(D)J recombination and retroviral DNA integration. In the Tn5 system, only one protein, the transposase (Tnp), is required for all of the catalytic steps involved in transposon movement. The complexity involved in moving multiple DNA strands within one active site suggests that, in addition to the specific contacts maintained between Tnp and its recognition sequence, Tnp also interacts with the flanking DNA sequence. Here, we demonstrate that Tnp interacts with the donor DNA region. Tnp protects the donor DNA from DNase I digestion, suggesting that Tnp is in contact with, or otherwise distorts, the donor DNA during synapsis. In addition, changes in the donor DNA sequence within this region alter the affinity of Tnp for DNA by eightfold during synapsis. In vitro selection for more stable synaptic complexes reveals an A/T sequence bias for this region. We further show that certain donor DNA sequences, which favor synapsis, also appear to serve as hot spots for strand transfer. The TTATA donor sequence represents the best site. Most surprising is the fact that this sequence is found within the Tnp recognition sequence. Preference for insertion into a site within the Tnp recognition sequence would effectively inactivate one copy of the element and form clusters of the Tn5 transposon. In addition, the fact that several donor DNA sequences, which favor synapsis, appear to serve as hot spots for transposon insertion suggest that similar criteria may exist for Tnp-donor DNA and Tnp-target DNA interactions.     

Purification and biochemical analyses of a monomeric form of Tn5 transposase.

The binding of transposase (Tnp) to the specific Tn5 end sequences is the first dedicated reaction during transposition. In this study, comparative DNA-binding analyses were performed using purified full-length Tnp and a C-terminal deletion variant (delta369) that lacks the putative dimerization domain. The shape of the binding curve of full-length Tnp is sigmoidal in contrast to the hyperbolic-shaped binding curve of delta369. This observation is consistent with previous observations as well as a rate of binding study presented here, which suggest that the full-length Tnp-end interaction, unlike that of the truncated protein, is a complex time-dependent reaction possibly involving a subunit exchange. Circular permutation assay results indicate that both proteins are capable of distorting the Tn5end sequences upon binding. Molecular weight determinations based on the migratory patterns of complexed DNA in polyacrylamide gels has shown that delta369 specifically binds the Tn5 end sequences as a monomer while full-length Tnp in complex represents a heterodimer.     

Mutational analysis of the base flipping event found in Tn5 transposition.

This work identifies novel structure-function relationships between Tn5 transposase (Tnp) and its DNA recognition sequence. The Tn5 Tnp-DNA co-crystal structure revealed the protein-DNA contacts of the post-cleavage complex (Davies, D. R., Goryshin, I. Y., Reznikoff, W. S., and Rayment, I. (2000) Science 289, 77-85). One of the most striking features of this complex is the rotation of thymine 2 (T2) away from the DNA helix and into a pocket within the Tnp. This interaction appears similar to the "base flipping" phenomenon found in many DNA repair enzymes such as T4 endonuclease V and uracil DNA glycosylase (Roberts, R. J., and Cheng, X. (1998) Annu. Rev. Biochem. 67, 181-198). To study the biochemical significance of this phenomenon, we mutated the Tnp residues proposed to be involved in stabilizing this interaction and removed the T2 nucleotide to examine which steps in the transposition reaction require T2-Tnp interactions. From this work, we have determined that stacking interactions between T2 and Tnp are critical for efficient transposition in vitro. In addition, our results suggested that T2-Tnp interactions facilitate hairpin formation and hairpin resolution primarily through base stacking and that T2 plays a role in the alignment of the transposon DNA for strand transfer.     

Mutation of Tn5 transposase beta-loop residues affects all steps of Tn5 transposition: the role of conformational changes in Tn5 transposition

X-ray cocrystal structures of Tn5 transposase (Tnp) bound to its 19 base pair (bp) recognition end sequence (ES) reveal contacts between a beta-loop (amino acids 240-260) and positions 3, 4, 5, and 6 of the ES. Here, we show that mutations of residues in this loop affect both in vivo and in vitro transposition. Most mutations are detrimental, whereas some mutations at position 242 cause hyperactivity. More specifically, mutations to the beta-loop affect every individual step of transposition tested. Mutants performing in vivo and in vitro transposition less efficiently also form fewer synaptic complexes, whereas hyperactive Tnps form more synaptic complexes. Surprisingly, two hypoactive mutations, K244R and R253L, also affect the cleavage steps of transposition with a much more dramatic effect on the second double end break (DEB) complex formation step, indicating that the beta-loop likely plays an important roll in positioning the substrate DNA within the catalytic site. Finally, all mutants tested decrease efficiency of the final transposition step, strand transfer. A disparity in cleavage rate constants in vitro for mutants with changes to the proline at position 242 on transposons flanked by ESs differing in the orientation of the A-T base pair at position 4 allows us to postulate that P242 contacts the position 4 nucleotide pair. On the basis of these data, we propose a sequential model for end cleavage in Tn5 transposition in which the uncleaved PEC is not symmetrical, and conformational changes are necessary between the first and second cleavage events and also for the final strand transfer step of transposition.     

Tn5 Transposase with an Altered Specificity for Transposon Ends

Tn5 is a composite bacterial transposon that encodes a protein, transposase (Tnp), required for movement of the transposon. The initial step in the transposition pathway involves specific binding of Tnp to 19-bp end recognition sequences. Tn5 contains two different specific end sequences, termed outside end (OE) and inside end (IE). In Escherichia coli, IE is methylated by Dam methylase (IE(ME)). This methylation greatly inhibits recognition by Tnp and greatly reduces the ability of transposase to facilitate movement of IE defined transposons. Through use of a combinatorial random mutagenesis technique (DNA shuffling), we have isolated an IE(ME)-specific hyperactive form of Tnp, Tnp sC7v.2.0, that is able to promote high levels of transposition of IE(ME) defined transposons in vivo and in vitro while functioning at wild-type levels with OE transposons. This protein contains a critical glutamate-to-valine mutation at amino acid 58 that is responsible for this change in end specificity.     

Effect of ethidium on the torsion constants of linear and supercoiled DNAs.

The torsion elastic constants (alpha) of linear pBR322 (4363 bp) and pUC8 (2717 bp) DNAs and supercoiled pBR322 and pJMSII (4375 bp) DNAs are measured in 0.1 M NaCl as a function of added ethidium/base-pair (EB/BP) ratio by studying the fluorescence polarization anisotropy (FPA) of the intercalated ethidium. The time-resolved FPA is measured by using a picosecond dye laser for excitation and time-correlated single photon counting detection. Previously developed theory for the emission anisotropy is generalized to incorporate rotations of the transition dipole due to excitation transfer. The excitation transfers are simulated by a Monte Carlo procedure (Genest et al., Biophys. Chem. 1 (1974) 266-278) and the consequent rotations of the transition dipole are superposed on the Brownian rotations. After accounting for excitation transfer, the torsion constants of the linear DNAs are found to be essentially independent of intercalated ethidium up to a binding ratio r = 0.10 dye/bp. Dynamic light scattering measurements on linear pUC8 DNA confirm that the torsion constant is independent of binding ratio up to r = 0.20 dye/bp. If alpha d denotes the torsion constant between ethidium and a base-pair, and alpha 0 that between two base-pairs, then our data imply that alpha d/alpha 0 lies in the range 0.65 to 1.64 with a most probable value of 1.0. The torsion constants of supercoiled DNAs decrease substantially with increasing binding ratio even after accounting for excitation transfer. At the binding ratio r* = 0.064, where the superhelix density vanishes and superhelical strain is completely relaxed, the torsion constant of the supercoiled pBR322 DNA/dye complex lies below that of the corresponding linear DNA/dye complex by about 30%. This contradicts the conventional view according to which linear, nicked circular, and supercoiled DNA/dye complexes with r = r* should coexist with the same concentration of free dye, display the same distribution of bound dye, and exhibit identical secondary structures, twisting and bending rigidities, and FPA dynamics. These and other observations imply the existence of metastable secondary structure in freshly relaxed supercoiled DNAs. A tentative explanation is presented for these and other unexpected observations on supercoiled DNAs.     

Effects of chloroquine on the torsional dynamics and rigidities of linear and supercoiled DNAs at low ionic strength

The magnitude and uniformity of the torsion elastic constant (alpha) of linear and supercoiled pBR322 DNAs are measured in 3 mM Tris as a function of added chloroquine/basepair ratio (chl/bp) by studying the fluorescence polarization anisotropy of intercalated ethidium dye. The time-resolved FPA is measured using a picosecond dye-laser for excitation and time-correlated single-photon counting detection. For both linear and supercoiled DNAs, alpha remains uniform except at the very highest chl/bp ratio examined. For the linear DNA, alpha decreases from 5.0 x 10(-12) dyne-cm at chl/bp = 0 to about 3.5 x 10(-12) dyne-cm at chl/bp = 0.5, and remains at that value up to chl/bp = 5, whereupon it increases back up to its original value. For the supercoiled DNA, alpha remains constant at about 5.2 x 10(-12) dyne-cm from chl/bp = 0 up to chl/bp = 5, whereupon it increases in parallel with the linear DNA. The effect of chloroquine on the secondary structure, torsion constant, and torsional dynamics evidently differs substantially between linear and supercoiled DNAs, even under conditions where the supercoiled DNA is completely relaxed and both DNAs bind the same amount of dye. This strongly contradicts any notion that the local structures of linear and relaxed supercoiled DNA/dye complexes with the same binding ratio are identical. The increase in apparent alpha at chl/bp = 5 for both DNAs may be due to stacking of the chloroquine in the major groove and consequent stiffening of the filament.     

Escherichia coli DNA Topoisomerase I Copurifies with Tn5 Transposase, and Tn5 Transposase Inhibits Topoisomerase I

Tn5 transposase (Tnp) overproduction is lethal to Escherichia coli. Genetic evidence suggested that this killing involves titration of E. coli topoisomerase I (Topo I). Here, we present biochemical evidence that supports this model. Tn5 Tnp copurifies with Topo I while nonkilling derivatives of Tnp, Delta37Tnp and Delta55Tnp (Inhibitor [Inh]), show reduced affinity or no affinity, respectively, for Topo I. In agreement with these results, the presence of Tnp, but not Delta37 or Inh derivatives of Tnp, inhibits the DNA relaxation activity of Topo I in vivo as well as in vitro. Other proteins, including RNA polymerase, are also found to copurify with Tnp. For RNA polymerase, reduced copurification with Tnp is observed in extracts from a topA mutant strain, suggesting that RNA polymerase interacts with Topo I and not Tnp.     

Tn5 transposase active site mutants

Tn5 transposase (Tnp) is a 53.3-kDa protein that is encoded by and facilitates movement of transposon Tn5. Tnp monomers contain a single active site that is responsible for catalyzing a series of four DNA breaking/joining reactions at one transposon end. Based on primary sequence homology and protein structural information, we designed and constructed a series of plasmids that encode for Tnps containing active site mutations. Following Tnp expression and purification, the active site mutants were tested for their ability to form protein-DNA complexes and perform each of the four catalytic steps in the transposition pathway in vitro. The results demonstrate that Asp-97, Asp-188, and Glu-326, visible in the active site of Tn5 crystal structures, are absolutely required for all catalytic steps. Mutations within a series of amino acid residues that are conserved in the IS4 family of transposases and retroviral integrases also impair Tnp catalytic activity. Mutations at either Tyr-319 or Arg-322 reduce both hairpin resolution and strand transfer activity within protein-DNA complexes. Mutations at Lys-333 reduce the ability of Tnps to form protein-DNA complexes, whereas mutations at the less strongly conserved Lys-330 have less of an effect on both synaptic complex formation and catalytic activity.     

Overexpression of the Tn5 transposase in Escherichia coli results in filamentation, aberrant nucleoid segregation, and cell death: analysis of E. coli and transposase suppressor mutations.

Overexpression of the Tn5 transposase (Tnp) was found to be lethal to Escherichia coli. This killing was not caused by transposition or dependent on the transpositional or DNA binding competence of Tnp. Instead, it was strictly correlated with the presence of a wild-type N terminus. Deletions removing just two N-terminal amino acids of Tnp resulted in partial suppression of this effect, and deletions of Tnp removing 3 or 11 N-terminal amino acids abolished the killing effect. This cytotoxic effect of Tnp overexpression is accompanied by extensive filament formation (i.e., a defect in cell division) and aberrant nucleoid segregation. Four E. coli mutants were isolated which allow survival upon Tnp overexpression, and the mutations are located at four discrete loci. These suppressor mutations map near essential genes involved in cell division and DNA segregation. One of these mutations maps to a 4.5-kb HindIII region containing the ftsYEX (cell division) locus at 76 min. A simple proposition which accounts for all of these observations is that Tnp interacts with an essential E. coli factor affecting cell division and/or chromosome segregation and that overexpression of Tnp titrates this factor below a level required for viability of the cell. Furthermore, the N terminus of Tnp is necessary for this interaction. The possible significance of this phenomenon for the transposition process is discussed.     

StpA protein from Escherichia coli condenses supercoiled DNA in preference to linear DNA and protects it from digestion by DNase I and EcoKI

The nucleoid-associated protein, StpA, of Escherichia coli binds non-specifically to double-stranded DNA (dsDNA) and apparently forms bridges between adjacent segments of the DNA. Such a coating of protein on the DNA would be expected to hinder the action of nucleases. We demonstrate that StpA binding hinders dsDNA cleavage by both the non-specific endonuclease, DNase I, and by the site-specific type I restriction endonuclease, EcoKI. It requires approximately one StpA molecule per 250–300 bp of supercoiled DNA and approximately one StpA molecule per 60–100 bp on linear DNA for strong inhibition of the nucleases. These results support the role of StpA as a nucleoid-structuring protein which binds DNA segments together. The inhibition of EcoKI, which cleaves DNA at a site remote from its initial target sequence after extensive DNA translocation driven by ATP hydrolysis, suggests that these enzymes would be unable to function on chromosomal DNA even during times of DNA damage when potentially lethal, unmodified target sites occur on the chromosome. This supports a role for nucleoid-associated proteins in restriction alleviation during times of cell stress.     

Substrate recognition and induced DNA deformation by transposase at the target-capture stage of Tn10 transposition

The bacterial transposon Tn10 inserts preferentially into sites that conform to a 9 bp consensus sequence: 5' NGCTNAGCN 3'. However, this sequence is not on its own sufficient to confer target specificity as the base-pairs flanking this sequence also contribute significantly to target-site selection. We have performed a series of "contact-probing experiments" to define directly the protein-DNA interactions that govern target-site selection in the Tn10 system. The HisG1 hotspot for Tn10 insertion was the main focus here. We infer that there is a rather broad zone ( approximately 24 bp) of contact between transposase and target DNA in the target-capture complex. This includes base-specific contacts at all of the purine residues in the consensus positions of the target core and primarily backbone contacts out to 7-8 bp in the two flanking regions immediately adjacent to the core. Also, highly localized sites of chemical hypersensitivity are identified that reveal symmetrically disposed deformations in DNA structure in the target-capture complex. Furthermore, the level of strand transfer is shown to be reduced by phosphorothioate substitution of phosphate groups at or close to the sites of target DNA deformation. Interestingly, for one particular target DNA, a mutant form of HisG1 called MutF, the above phosphorothioate inhibition of strand transfer is suppressed by replacing Mg(2+) with Mn(2+). Based on these results a model for sequence-specific target capture is proposed which attempts to define possible relationships between transposase interactions with the target core and flanking sequences, transposase-induced DNA deformation of the target site and divalent metal ion binding to the target-capture complex.     

Dissecting Tn5 transposition using HIV-1 integrase diketoacid inhibitors

Diketoacid (DKA) compounds have been shown to inhibit HIV-1 integrase by a mechanism that involves sequestration of the active site metals. Because HIV-1 integrase and Tn5 transposase have similar active site architectures and catalytic mechanisms, we investigated whether DKA analogues would inhibit Tn5 transposase activity and provide a model system to explore the mechanisms of action of these inhibitors. A screen of several hundred DKA analogues identified several with activity against Tn5 Tnp. Six DKA inhibitors used in this study manifested a variety of effects on different transposition steps suggesting that different analogues may have different binding contacts with transposase. All DKA compounds inhibited paired end complex (PEC) formation in which the nucleoprotein complex required for catalysis is assembled. Dissociation of PECs by some DKA compounds indicates that these inhibitors can decrease PEC stability. Four DKA compounds inhibited the two cleavage steps releasing transposon DNA from flanking DNA, and one of these four compounds preferentially inhibited the second cleavage step. The differential effect of this inhibitor on the second cleavage event indicates that cleavage of the two transposon-donor DNA boundaries is a sequential process requiring a conformational change. The requirement for a conformational change between cleavage events was also demonstrated by the inability of transposase to perform second cleavage at 25 degrees C. Finally, all six compounds inhibit strand transfer, the final step of Tn5 transposition. Two of the compounds that inhibited strand transfer have no effect on DNA cleavage. The strand transfer inhibition properties of various DKA compounds was sensitive to the structure of the 5'-non-transferred strand, suggesting that these compounds bind in or near the transposase active site. Other results that probe compound binding sites include the effects of active site mutations and donor DNA on DKA compound inhibition activities. Thus, DKA inhibitors will provide an important set of tools to investigate the mechanism of action of transposases and integrases.     

Tn5 transposase active site mutations suggest position of donor backbone DNA in synaptic complex

Tn5 transposase (Tnp), a 53.3-kDa protein, enables the movement of transposon Tn5 by a conservative mechanism. Within the context of a protein and DNA synaptic complex, a single Tnp molecule catalyzes four sequential DNA breaking and joining reactions at the end of a single transposon. The three amino acids of the DDE motif (Asp-97, Asp-188, and Glu-326), which are conserved among transposases and retroviral integrases, have been shown previously to be absolutely required for all catalytic steps. To probe the effect of active site geometry on the ability to form synaptic complexes and perform catalysis, single mutations at each position of the DDE motif were constructed. The aspartates were changed to glutamates, and the glutamate was changed to an aspartate. These mutants were studied by performing in vitro binding assays using short oligonucleotide substrates simulating the natural substrates for the synaptic complex formation and subsequent transposition steps. The results indicate that the aspartate to glutamate mutations restrict synaptic complex formation with substrates resembling the natural transposon prior to transferred strand nicking. This suggests a structural model in which the donor backbone DNA, prior to nicking, occupies the same space that is invaded by the longer side chains present in the aspartate to glutamate mutants. Additionally, catalytic assays support the previous proposal that the active site coordinates two divalent metal ions.     

Evidence for "unseen" transposase--DNA contacts

In this study, evidence of novel, important interactions between a hyperactive Tn5 transposon recognition end sequence and hyperactive Tn5 transposase (Tnp) are presented. A hyperactive Tn5 end sequence, the mosaic end (ME), was isolated previously. The ME and a wild-type end sequence, the outside end (OE), differ at only three positions, yet transposition on the ME is tenfold higher than on the OE in vivo. Also, transposition on the ME is much more efficient than transposition on the OE in vitro. Here, we show that the decreased activity observed for the OE is caused by a defect in paired ends complex (PEC) formation resulting from the orientation of the A-T base-pair at position 4 of this end. Efficient PEC formation requires an interaction between the C5-methyl group (C5-Me) on the non-transferred strand thymine base at position 4 (T4) and Tnp. PEC formation on nicked substrates is much less affected by the orientation of the A-T base-pair at position 4, indicating that the C5-Me group is important only for steps preceding nicking. A recently determined co-crystal structure of Tn5 Tnp with the ME is discussed and a model explaining possible roles for the base-pair at position 4 is explored.     

Macroevolution by transposition: drastic modification of DNA recognition by a type I restriction enzyme following Tn5 transposition.

We have characterized a novel mutant of EcoDXXI, a type IC DNA restriction and modification (R-M) system, in which the specificity has been altered due to a Tn5 insertion into the middle of hsdS, the gene which encodes the polypeptide that confers DNA sequence specificity to both the restriction and the modification reactions. Like other type I enzymes, the wild type EcoDXXI recognizes a sequence composed of two asymmetrical half sites separated by a spacer region: TCA(N7)RTTC. Purification of the EcoDXXI mutant methylase and subsequent in vitro DNA methylation assays identified the mutant recognition sequence as an interrupted palindrome, TCA(N8)TGA, in which the 5' half site of the wild type site is repeated in inverse orientation. The additional base pair in the non-specific spacer of the mutant recognition sequence maintains the proper spacing between the two methylatable adenine groups. Sequencing of both the wild type and mutant EcoDXXI hsdS genes showed that the Tn5 insertion occurred at nucleotide 673 of the 1221 bp gene. This effectively deletes the entire carboxyl-terminal DNA binding domain which recognizes the 3' half of the EcoDXXI binding site. The truncated hsdS gene still encodes both the amino-terminal DNA binding domain and the conserved repeated sequence that defines the length of the recognition site spacer region. We propose that the EcoDXXI mutant methylase utilizes two truncated hsdS subunits to recognize its binding site. The implications of this finding in terms of subunit interactions and the malleability of the type I R-M systems will be discussed.     

Occurrence of potential cruciform and H-DNA forming sequences in genomic DNA.

We have used computer-assisted methods to search large amounts of the human, yeast and Escherichia coli genomes for inverted repeat (IR) and mirror repeat (MR) DNA sequence patterns. In highly supercoiled DNA some IRs can form cruciforms, while some MRs can form intramolecular triplexes, or H-DNA. We find that total IR and MR sequences are highly enriched in both eukaryotic genomes. In E. coli, however, only total IRs are enriched, while total MRs only occur as frequently as in random sequence DNA. We then used a set of experimentally derived criteria to predict which of the total IRs and MRs are most likely to form cruciforms or H-DNA in supercoiled DNA. We show that strong cruciform forming sequences occur at a relatively high frequency in yeast (1/19 700 bp) and humans (1/41 800 bp), but that H-DNA forming sequences are abundant only in humans (1/49 400 bp). Strong cruciform and H-DNA forming sequences are not abundant in the E.coli genome. These results suggest that cruciforms and H-DNA may have a functional role in eukaryotes, but probably not prokaryotes.