IS50-mediated inverse transposition: specificity and precision

The IS50 elements, which are present as inverted repeats in the kanamycin-resistance transposon, Tn5, can move in unison carrying with them any interstitial DNA segment. In consequence, DNA molecules such as a lambda::Tn5 phage genome are composed of two overlapping transposons - the kan segment bracketed by IS50 elements (Tn5), and lambda bracketed by IS50 elements. During direct transposition, mediated by IS50 "O" (outside) ends, the kan gene is moved and the lambda vector is left behind. During inverse transposition, mediated by the "I" (inside) ends of the IS50 elements, the lambda vector segment is moved and the kan gene is left behind. Direct transposition is several orders of magnitude more frequent than inverse transposition (Isberg and Syvanen, 1981; Sasakawa and Berg, 1982). We assessed the specificity and precision of the rare events mediated by pairs of I ends by mapping and sequencing independent inverse transpositions from a lambda::Tn5 phage into the amp and tet genes of plasmid pBR322. Using restriction analyses, 32 and 40 distinct sites of insertion were found among 46 and 72 independent inverse transpositions into the amp and tet genes, respectively. Eleven sites were used in two or more insertion events, and the two sites in tet used most frequently corresponded to major hotspots for the insertion of the Tn5 (by direct transposition). The sequences of 22 sites of inverse transposition (including each of the sites used more than once) were determined, in eleven cases by analyzing both pBR322-IS50 junctions, and in eleven others by sequencing one junction. The sequence of the "I" end of IS50 was preserved and 9-bp target sequence duplications were present in every case analyzed. GC pairs were found at each end of the target sequence duplication in ten of the eleven sites used more than once, and also in seven of the other eleven sites. Our data indicate that transposition mediated by pairs of "I" ends is similar in its specificity and precision to the more frequent transposition mediated by IS50 "O" ends.     

A sequence at the inside end of IS50 down regulates transposition.

Deletion analysis has shown that the segment at the IS50 inside (I) end that is needed for efficient transposition is approximately 19 bp long. Dam methylation at two 5' GATC sequences within this segment decreases I-end transposition activity. A third 5' GATC sequence is present at bp 21-24 of the I end. The comparisons presented here show that extension of the I end from 19 to 24 bp decreases its transposition activity in dam cells 5- to 50-fold, depending on the overall transposon structure.     

Fis plays a role in Tn5 and IS50 transposition.

The Fis (factor for inversion stimulation) protein of Escherichia coli was found to influence the frequency of transposon Tn5 and insertion sequence IS50 transposition. Fis stimulated both Tn5 and IS50 transposition events and also inhibited IS50 transposition in Dam-bacteria. This influence was not due to regulation by Fis of the expression of the Tn5 transposition proteins. We localized, by DNase I footprinting, one Fis site overlapping the inside end of IS50 and give evidence to strongly suggest that when Fis binds to this site, IS50 transposition is inhibited. The Fis site at the inside end overlaps three Dam GATC sites, and Fis bound efficiently only to the unmethylated substrate. Using a mobility shift assay, we also identified another potential Fis site within IS50. Given the growth phase-dependent expression of Fis and its differential effect on Tn5 versus IS50 transposition in Dam-bacteria, we propose that the high levels of Fis present during exponential growth stimulate transposition events and might bias those events toward Tn5 and away from IS50 transposition.     

Orientation of IS50 transposase gene and IS50 transposition.

Reversal of transposase gene orientation with respect to the nonidentical ends of IS50 strongly decreased IS50 transposition in both Dam- and Dam+ hosts. In either orientation, IS50 transposase expression was unaffected. These effects were independent of the surrounding DNA context. This shows that the efficiency of IS50 transposition is dependent on transposase gene orientation. The transposition frequencies of transposons utilizing inverted IS50 inside ends (IE), IE-IE transposons, were lower than either outside end (OE)-IE or OE-OE transposons.     

Saturation mutagenesis of the inside end of insertion sequence IS50

A 19-bp segment at the inside (I) end of IS50 (Tn5) is needed for efficient transposition. The importance of each position was assayed by making at least one base substitution at each position by either chemical-or oligodeoxyribonucleotide-directed mutagenesis. Mutant I ends were paired with a wild-type (wt) segment from the outside (O) end of IS50 and the transposase (tnp) gene was placed either between the ends or 1200 bp from the O end. The frequency of transposition of the resultant elements to bacteriophage lambda was measured. At least one substitution at each of the 19 I-end positions decreased transposition activity to less than 25% of wt, and most substitutions (25 of 28) decreased it to less than 5% of wt from one or both donor plasmids. These results show that each position in the I end is important during transposition.     

Localization of Action of the Is50-Encoded Transposase Protein

The movement of the bacterial insertion sequence IS50 and of composite elements containing direct terminal repeats of IS50 involves the two ends of IS50, designated O (outside) and I (inside), which are weakly matched in DNA sequence, and an IS50 encoded protein, transposase, which recognizes the O and I ends and acts preferentially in cis. Previous data had suggested that, initially, transposase interacts preferentially with the O end sequence and then, in a second step, with either an O or an I end. To better understand the cis action of transposase and how IS50 ends are selected, we generated a series of composite transposons which contain direct repeats of IS50 elements. In each transposon, one IS50 element encoded transposase (tnp+), and the other contained a null (tnp-) allele. In each of the five sets of composite transposons studied, the transposon for which the tnp+ IS50 element contained its O end was more active than a complementary transposon for which the tnp- IS50 element contained its O end. This pattern of O end use suggests models in which the cis action of transposase and its choice of ends is determined by protein tracking along DNA molecules.     

Functional organization of the ends of IS 1: specific binding site for an IS1-encoded protein

The IS 1-encoded protein InsA binds specifically to both ends of IS1, and acts as a repressor of IS1 gene expression and may be a direct inhibitor of the transposition process. We show here, using DNasel 'foot-printing' and gel retardation, that the InsA binding sites are located within the 24/25 bp minimal active ends of IS1 and that InsA induces DNA bending upon binding. Conformational modification of the ends of IS1 as a result of binding of the host protein integration host factor (IHF) to its site within the minimal ends has been previously observed. Using a collection of synthetic mutant ends we have mapped some of the nucleotide sequence requirements for InsA binding and for transposition activity. We show that sequences necessary for InsA binding are also essential for transposition activity. We demonstrate that InsA and IHF binding sites overlap since some sequence determinants are shared by both InsA and IHF. The data suggest that these ends contain two functional domains: one for binding of InsA and IHF, and the other for transposition activity. A third region, when present, may enhance transposition activity with an intact right end. This 'architecture' of the ends of IS1 is remarkably similar to that of IS elements IS10, IS50 and IS903.     

Tn5 insertion specificity is not influenced by IS50 end sequences in target DNA.

Summary The bacterial transposon Tn5 inserts into dozens of sites in a gene, some of which are used preferentially (hotspots). Features of certain sites and precedents provided by several other transposons had suggested that sequences in target DNA corresponding to the ends of Tn5 or of its component IS50 elements might facilitate transposition to these sites. We tested this possibility using derivatives of plasmid pBR322 carrying IS50 I or O end sequences. Tn5 inserted frequently into an IS50 I end at the major hotspot in pBR322, but not into either an I end or an O end 230 by away from this hotspot. Adenine (dam) methylation at GATC sequences in the I end segment interferes with its use as the end of a transposon, but a dam ^– mutation did not affect Tn5 insertion relative to an I end sequence in target DNA. These results support models in which the ability of Tn5 to find its preferred sites depends on several features of DNA sequence and conformation, and in which target selection is distinct from recognition of the element ends during transposition.     

Artificial transposable elements in the study of the ends of IS1

We have constructed artificial IS1-based transposons by attaching synthetic oligodeoxynucleotides, corresponding to the sequence of the ends of IS1, to a selectable DNA segment ['omega' fragment; Prentki and Krisch, Gene 29 (1984) 303-313]. These transposons were used to examine the sequence requirements at the ends for IS1 transposition. We show here that a 24- to 28-bp sequence from the left or right ends of IS1 is capable of transposition when present at both ends of the omega fragment in the correct orientation. Transposition activity requires the presence of an intact IS1 in cis on the same plasmid molecule. In trans, however, neither resident genomic copies of IS1, nor copies carried by a compatible, high-copy-number plasmid present in the same cell, complement the artificial transposons efficiently. Transposition frequencies in the presence of a cis-complementing IS1 are, however, similar to those of the naturally occurring IS1-based transposon, Tn9. In addition, transposition results in a 9-bp duplication in the target DNA molecule as is usually the case for insertion of the intact IS1. Using this system, we have obtained evidence indicating that the activity of a synthetic IS1 end is not determined exclusively by its sequence, but can be strongly enhanced by a second, wild-type end used in the transposition event. The data also show that single base pair mutations can exhibit a cumulative effect in reducing transposition activity.     

Integration host factor plays a role in IS50 and Tn5 transposition.

In Escherichia coli, the frequencies of IS50 and Tn5 transposition are greater in Dam- cells than in isogenic Dam+ cells. IS50 transposition is increased approximately 1,000-fold and Tn5 transposition frequencies are increased about 5- to 10-fold in the absence of Dam methylation. However, in cells that are deficient for both integration host factor (IHF) and Dam methylase, the transposition frequencies of IS50 and Tn5 approximate those found in wild-type cells. The absence of IHF alone has no effect on either IS50 or Tn5 transposition. These results suggest that IHF is required for the increased transposition frequencies of IS50 and Tn5 that are observed in Dam- cells. It is also shown that the level of expression of IS50-encoded proteins, P1 and P2, required for IS50 and Tn5 transposition and its regulation does not decrease in IHF- or in IHF- Dam- cells. This result suggests that the effects of IHF on IS50 and Tn5 transposition are not at the level of IS50 gene expression. Finally, IHF is demonstrated to significantly retard the electrophoretic mobility of a 289-base-pair segment of IS50 DNA that contains a putative IHF protein-binding site. The physiological role of this IHF binding site remains to be determined.     

Sequences essential for IS50 transposition. The first base-pair

Sequences near the ends of the insertion element IS50 are essential for its transposition, probably because they serve as sites upon which the IS50-encoded transposase protein acts. To determine if these essential sequences include the first base-pair at each end of IS50 we generated 5'C to 5'G transversions at these positions. Each mutation reduced the transposition frequency to 1% to 2% of wild-type. DNA sequence analyses showed that the mutant 5'G is preserved during transposition.     

Escherichia coli integration host factor binds specifically to the ends of the insertion sequence IS1 and to its major insertion hot-spot in pBR322

We report here that the ends of IS1 are bound and protected in vitro by the heterodimeric protein integration host factor (IHF). Under identical conditions, RNA polymerase binds to one of these ends (IRL) and protects a region that includes the sequences protected by IHF. Other potential sites within IS1, identified by their homology to the apparent consensus sequence, are not protected. Footprinting analysis of deletion derivatives of the ends demonstrates a correspondence between the ability of the end sequence to bind IHF and its ability to function as an end in transposition. Nonetheless, some transposition occurs in IHF- cells, indicating that IHF is not an essential component of the transposition apparatus. IHF also binds and protects four closely spaced regions within the major hot-spot for insertion of IS1 in the plasmid pBR322. This striking correlation of hot-spot and IHF-binding sites suggests a possible role for IHF in IS1 insertion specificity.     

Genetic organization of transposon Tn10

Transposon Tn10 is 9300 bp in length, with 1400 bp inverted repeats at its ends. The inverted repeats are structurally intact IS-like sequences (Ross et al., 1979). Analysis of deletion mutants and structural variants of Tn10, reported below, shows that the two IS10 segments contain all of the Tn10-encoded genetic determinants, both sites and functions, that are required for transposition. Furthermore, the two repeats (IS10-Right and IS10-Left) are not functionally equivalent: IS10-Right is fully functional and is capable by itself of promoting normal levels of Tn10 transposition; IS10-Left functions only poorly by itself, promoting transposition at a very low level when IS10-Right is inactivated. Complementation analysis shows that IS10-Right encodes at least one function, required for Tn10 transposition, which can act in trans and which works at the ends of the element. Also, all of the sites specifically required for normal Tn10 transposition have been localized to the outermost 70 bp at each end of the element; there is no evidence that specific sites internal to the element play an essential role. Finally, Tn10 modulates its own transposition in such a way that transposition-defective point mutants, unlike deletion mutants, are not complemented by functions provided in trans; and wild-type Tn10, unlike deletion mutants, is not affected by functions provided in trans from a "high hopper" Tn10 element.     

Role of the IS50 R proteins in the promotion and control of Tn5 transposition

IS50R, the inverted repeat sequence of Tn5 which is responsible for supplying functions that promote and control Tn5 transposition, encodes two polypeptides that differ at their N terminus. Frameshift, in-frame deletion, nonsense, and missense mutations within the N terminus of protein 1 (which is not present in protein 2) were isolated and characterized. The properties of these mutations demonstrate that protein 1 is absolutely required for Tn5 transposition. None of these mutations affected the inhibitory activity of IS50, confirming that protein 2 is sufficient to mediate inhibition of Tn5 transposition. The effects on transposition of increasing the amount of protein 2 (the inhibitor) relative to protein 1 (the transposase) were also analyzed. Relatively large amounts of protein 2 were required to see a significant decrease in the transposition frequency of an element. In addition, varying the co-ordinate synthesis of the IS50 R proteins over a 30-fold range had little effect on the transposition frequency. These studies suggest that neither the wild-type synthesis rate of protein 2 relative to protein 1 nor the amount of synthesis of both IS50 R proteins is the only factor responsible for controlling the transposition frequency of a wild-type Tn5 element in Escherichia coli.     

Intramolecular transposition by a synthetic IS50 (Tn5) derivative.

We report the formation of deletions and inversions by intramolecular transposition of Tn5-derived mobile elements. The synthetic transposons used contained the IS50 O and I end segments and the transposase gene, a contraselectable gene encoding sucrose sensitivity (sacB), antibiotic resistance genes, and a plasmid replication origin. Both deletions and inversions were associated with loss of a 300-bp segment that is designated the vector because it is outside of the transposon. Deletions were severalfold more frequent than inversions, perhaps reflecting constraints on DNA twisting or abortive transposition. Restriction and DNA sequence analyses showed that both types of rearrangements extended from one transposon end to many different sites in target DNA. In the case of inversions, transposition generated 9-bp direct repeats of target sequences.     

Two Classes of Tn10 Transposase Mutants That Suppress Mutations in the Tn10 Terminal Inverted Repeat

Tn10 transposition requires IS10 transposase and essential sequences at the two ends of the element. Mutations in terminal basepairs 6-13 confer particularly strong transposition defects. We describe here the identification of transposase mutations that suppress the transposition defects of such terminus mutations. These mutations are named ``SEM'''' for suppression of ends mutations. All of the SEM mutations suppress more than a single terminus mutation and thus are not simple alterations of transposase/end recognition specificity. The mutations identified fall into two classes on the basis of genetic tests, location within the protein and nature of the amino acid substitution. Class I mutations, which are somewhat allele specific, appear to define a small structural and functional domain of transposase in which hydrophobic interactions are important at an intermediate stage of the transposition reaction, after an effective interaction between the ends but before transposon excision. Class II mutations, which are more general in their effects, occur at a single residue in a small noncritical amino-terminal proteolytic domain of transposase and exert their affects by altering a charge interaction; these mutations may affect act early in the reaction, before or during establishment of an effective interaction between the ends.     

The organization of the outside end of transposon Tn5.

The end sequences of the IS50 insertion sequence are known as the outside end (OE) and inside end. These complex ends are related but nonidentical 19-bp sequences that serve as substrates for the activity of the Tn5 transposase. Besides providing the binding site of the transposase, the end sequences of a transposon contain additional types of information necessary for transposition. These additional properties include but are not limited to host protein interaction sites and sites that program synapsis and cleavage events. In order to delineate the properties of the IS50 ends,the base pairs involved in the transposase binding site have been defined. This has been approached through performing a variety of in vitro analyses: a ++hydroxyl radical missing-nucleoside interference experiment, a dimethyl sulfate interference experiment, and an examination of the relative binding affinities of single-site end substitutions. These approaches have led to the conclusion that the transposase binds to two nonsymmetrical regions of the OE, including positions 6 to 9 and 13 to 19. Proper binding occurs along one face of the helix, over two major and minor grooves, and appears to result in a significant bending of the DNA centered approximately 3 bp from the donor DNA-OE junction.     

IS10 transposase mutations that specifically alter target site recognition.

IS10 inserts preferentially into particular hotspots. We describe here mutations of IS10 transposase, called 'ATS' that confer Altered Target Specificity. These mutations yield a general relaxation in target specificity but do not affect other aspects of transposition. Thus, the preference for specific nucleotide sequences at the target site can be cleanly separated from other steps of the transposition reaction. Eleven ATS mutations identified in a genetic screen occur at only two codons in transposase, one in each of two regions of the protein previously implicated in target site interactions (Patch I and Patch II). Genetic analysis suggests that mutations at the two ATS codons affect the same specific function of transposase, thus raising the possibility that Patch I and Patch II interact. For wild-type IS10, insertion specificity is determined in part by a specific 6 bp consensus sequence and in part by the immediately adjacent sequence context of the target DNA. The ATS mutations do not qualitatively alter the hierarchy with which base pairs are recognized in the consensus sequence; instead, sites selected by ATS transposase exhibit a reduction in the degree to which certain base pairs are preferred over others. Models for the basis of this phenotype are discussed.     

IS10/Tn10 transposition efficiently accommodates diverse transposon end configurations.

Transposon Tn10 and its component insertion sequence IS10 move by non-replicative transposition. We have studied the array of reaction intermediates and products in a high efficiency in vitro IS10/Tn10 transposition reaction. Synapsis of two transposon ends, followed by cleavage and strand transfer, can occur very efficiently irrespective of the relative locations and orientations of the two ends. The two participating ends can occur in inverted or direct orientation on the same molecule or, most importantly, on two different molecules. This behavior contrasts sharply with that of Mu, in which transposition is strongly biased in favor of inverted repeat synapsis. Mechanistically, the absence of discrimination amongst various end configurations implies that the architecture within the IS10/Tn10 synaptic complex is relatively simple, i.e. lacking any significant intertwining of component DNA strands. Biologically these observations are important because they suggest that the IS10 insertion sequence module has considerable flexibility in the types of DNA rearrangements that it can promote. Most importantly, it now seems highly probable that a single non-replicative IS10 element can promote DNA rearrangements usually attributed to replicative transposition, i.e. adjacent deletions and cointegrates, by utilizing transposon ends on two sister chromosomes. Other events which probably also contribute to the diversity of IS10/Tn10-promoted rearrangements are discussed.