Hyperactive transposase mutants of the Himar1 mariner transposon

Strains were grown at the temperatures indicated in LB broth or on agar plates prepared as described (24). Papillation assays were performed on thick MacConkey lactose agar plates. Antibiotics concentrations were ampicillin (Amp), 100 μg; gentamicin (Gen), 10 μg; kanamycin (Kan), 40 μg; tetracycline (Tet), 15 μg; naladixic acid (Nal), 20 μg; chloramphenicol (Cam), 34 μg, per ml, respectively, except where otherwise noted.

pMarNco was constructed via PCR using the primers 5′-CCCCTCGAGCCATGGAAAAAAAGGAATTTCGTG-3′ and 5′-CCGCTCAGAATCATCAACACGTT-3′ and pMarNde18 (10) as a template. The resulting PCR fragment and pMarNde18 were cut with XhoI and EcoRV and were ligated to created pMarNco, which contains the Himar1 transposase coding sequence with an NcoI site at the start codon. pBCMar was constructed by cleaving pMarNde18 with PstI, creating a blunt end using T4 DNA polymerase, then digesting with XhoI. The fragment that encodes the transposase was gel purified and cloned into pBCKS+ that had been digested with XhoI and Ecl136I.

A miniHimar1 transposon containing an ORF through the 3′ inverted terminal repeat (ITR) was constructed by PCR using a PCR-ligation-PCR method (25). The 5′ ITR was amplified with primers 5′ TACCCGGGAATCATTTGAAGGTTGGTAC (76rSma) and 5′ TAATACGACTCACTATAGGG (T7), and the 3′ ITR was amplified with the primers 5′AACGAATTTTAACAAAAAAATGTG (Mar3′r) and 5′CGATTTAGGTGACACTATAG (SP6), both using pMinimariner (13) as a template and Pfu polymerase. The two separate PCR reactions were treated with T4 polynucleotide kinase and ATP. Five microliters of each kinase reaction were ligated with T4 DNA ligase for 15 min at room temperature. Another PCR reaction then was performed by using the SP6 and T7 primers and 1 μl of the ligation as a template, using Taq DNA polymerase. The resulting product was cloned as a t-tailed fragment into pTAdv1 (CLONETECH), producing pTAdvMMOrf. This clone was cut with BamHI, and the fragment containing the minimariner was isolated and ligated to the BamHI site of pCDNAII (Invitrogen) to produced pMMOrf, which contains a unique BglII site in the middle of the element.

A papillation construct was produced by cleaving pMMOrf with BglII and ligating it to the BamHI/BglII fragment of pRZ1495 (26) to produce pMMLacTet. The lacZ gene of this insert has no transcriptional or translational controls (27).

An F-plasmid containing the papillation transposon from pMMLacTet was produced by transforming E. coli β2155 (28) by electroporation with pBCMar and pMMLacTet with selection on Amp, Cam (20 μg/ml), and diaminopimelic acid. Diaminopimelic acid is required for growth of β2155, which is a dapA mutant. Approximately 5,000 colonies were pooled and mixed with HB101 for a 6 h of mating on LB agar. Exconjugants in which the F′ from β2155 was mated out into HB101 were selected on Tet and Xgal (20 μg/ml) in the absence of diaminopimelic acid, resulting in colonies that were dark blue, light blue, white, or mosaic. White colonies were picked and colony purified on a second LB-Tet-Xgal plate. Resulting clones were patched to plates containing either Amp, Cam, Kan, or Tet, and clones confirmed to be Amp^s, Cam^s, Kan^s, and Tet^R were named HBFlac.

To verify that the HBFlac strains contain functional Himar1 ′lacZ elements, they were transformed with pBCMar and were selected on Cam and Xgal. The resulting colonies were dark blue, light blue, white and mixed (i.e., exhibited blue papillae on white colonies). The F′ from HBFlac3 (the lightest blue strain) was transferred to DH5α by conjugation with selection on Nal and Tet, yielding the strain DL1.

A Himar1 transposon for use in the mating-out assay was constructed by ligating the BamHI/EagI fragment of pMarKan (10) containing a KanR-marked Himar1 transposon to the BamHI/EagI fragment of pACYC184, resulting in the plasmid pACMarKan.

E. coli protein expression plasmids were made for each of the hyperactive transposase mutants by cutting the pBAD24 vectors containing the hyperactive inserts with NcoI, making this site blunt with Klenow, and cutting again with KpnI. The coding sequence fragments were purified from a 0.5% 1× TAE (40 mM Tris-acetate/1 mM EDTA) gel and were ligated to the NdeI (made blunt with Klenow as above)/KpnI sites of pET29b+ (Novagen), yielding pET29A7 and pET29C9 for the pBADA7- and pBADC9-containing transposase mutants, respectively.

Mutations were introduced into the coding region of Himar1 transposase by error-prone PCR using pMarNco as a template (29). The reactions contained ≈2 ng of template DNA, 50 mM KCl, 10 mM Tris⋅HCl (pH 9.0 at 25°C), 0.1% Triton X-100, 1.5 mM MgCl2, and either 200 or 100 μM MnCl₂ in a volume of 25 μl and were run for 30 cycles at 95°C for 1 min, 52°C for 1 min, and 75°C for 1.5 min. PCR products were cut with NcoI and PstI at 37°C for 45 min. The cleaved products were isolated from a 0.5% agarose gel in 1× TAE buffer by using a Qiagen (Chatsworth, CA) gel purification kit. Purified products were ligated into the NcoI/PstI sites of pBAD24 (30). These ligation reactions were used as the source of transposase mutants in the papillation assay.

A papillation assay to detect mutants of Himar1 was performed by transforming 1 μl of a ligation of mutated Himar1 coding sequences in pBAD24 into electrocompetent DL1 cells (see Fig. 2A). Screens of similar design have been used for bacterial transposons, including in Tn5 and Tn10 transposases (31–33). Cells also were transformed with pBADMar1 as a wild-type control. Transformed cells were resuspended in 1 ml of cold LB medium and were shaken at 37°C for 1 h. Dilutions of these cells were plated onto thick (50 ml in 100- × 15-mm dishes) MacConkey lactose agar plates containing Amp and Tet so that there were ≈100–150 colonies per plate, and the plates were incubated at 32°C for 2–3 days. Typically, papillae could be detected by using the wild-type transposase source at ≈50 h after plating. Potential hypertransposers were picked and grown overnight in LB with Amp, and the mutant transpose source DNA was purified. Putative hypertransposers were examined again by using the papillation assay to confirm hyperactivity.

A mating-out assay, which measures the frequency of transposition of a KanR Himar1 minitransposon from a plasmid to an F factor, was carried out to quantify the activity of the putative hyperactive mutants (see Fig. 2B). RZ212(MK) cells were transformed with individual mutant transpose sources isolated in the papillation assay, and the cells were grown as above. Cells were plated on LB agar containing Amp, Gen, and Cam and were grown overnight at 37°C. Five colonies were picked the following day and were grown for 16 h in LB containing Amp, Gen, and Kan at 37°C. These cells were mated to RZ221 cells by mixing 10 μl of the donor strain and 30 μl of recipient cells from overnight cultures in 1 ml of LB medium. The mating mixture was shaken gently at 37°C for 6–10 h. Mating cultures were vortexed vigorously, and suitable dilutions were plated on LB agar plates containing Nal and Gen to detect total numbers of exconjugates and on Nal and Kan to measure the number of exconjugates that contained a Himar1 insertion.

Transposases were purified as described in ref. 10. Protein purity was determined by Coomassie blue-stained 10–20% polyacrylamide gradient gels. Protein concentrations were determined spectrophotometrically as described in ref. 13 and were confirmed visually on Coomassie blue-stained 4–20% SDS/PAGE gels.

Comparative rates of transposition were determined by measuring the relative ability of the transposases to incorporate a radiolabeled DNA fragment containing the left ITR of Himar1 into an unlabeled supercoiled plasmid target in a reaction similar to that for Tn10 in vitro (Fig. 4) (34). The fragment containing the ITR was labeled by cutting p27fH5′ (10) with EcoRI and isolating the 111-bp fragment on a 1.5%, 1× TAE agarose gel. The DNA was purified from the agarose as described above and then was radiolabeled by filling in the overhanging ends with ³²P-α-dATP using Klenow under standard conditions. The reaction was stopped by heating to 70°C for 20 min, and the labeled DNA was purified by passing the reaction over a G50 spin column.

Transposition reactions contained 10% glycerol (vol/vol), 25 mM Hepes (pH 7.9 at room temperature), 250 of μg acetylated BSA, 2 mM DTT, 100 mM NaCl, 5 mM MgCl₂, 450 ng of target plasmid DNA, ≈10,000 cpm labeled ITR DNA, and a 10 nM concentration of one of the purified transposases. Reactions were performed at 28°C, the optimal temperature for wild-type Himar1 transposase (13). Ten-microliter aliquots were removed at 1-h intervals for 6 h, and the reaction was stopped by adding 2 μl of stop solution (60 mM EDTA/0.25% bromophenol blue/0.25% xylene cyanol/15% ficoll). Reaction products were separated on a 0.5% 1× TAE agarose gel. The gel was photographed, was placed on a piece of exposed x-ray film as a support, and then was dried until completely flat in a forced-air oven set at 55°C for 5–6 h. Reaction products in dried gels were analyzed by using a Molecular Dynamics PhosphorImager and imagequant software.

In vitro reactions for H. influenzae mutagenesis were conducted as above except that 100 nM transposase was added to reactions containing 500 ng of target PCR product and 200 ng of transposon donor plasmid pENT3 carrying Tn-magellan1. Independent reactions and transformations were performed in triplicate. Repair of transposon junctions and transformation of H. influenzae was as described (18).

A papillation screen was used to detect altered levels of transposition frequency for a pool of transposase mutants constructed by error-prone PCR (Fig. 2A). The Himar1 papillation screen is based on the ability of Himar1 to mobilize a nonautonomous Himar1 transposon carrying an in-frame fusion of a lacZ gene off an F plasmid and into the E. coli chromosome. If the transposon insertion fuses in frame with an expressed E. coli gene, a protein fusion can be produced that contains β-gal activity. The Lac(+) subpopulation of cells in an otherwise Lac(−) E. coli colony can metabolize lactose if plated on MacConkey agar. The Lac(+) cells grow faster than the surrounding Lac(−) cells and thus will produce bumps, or papillae, on the colony. Moreover, these papillae turn red on MacConkey lactose agar because of the production of lactic acid and its detection by the neutral red in the media. The greater the number of papillae produced after a given period of time, the greater is the frequency of transposition.

Transposition of Himar1 in E. coli was readily detected by using the papillation screen. Papillation only occurred when using Himar1 transposase constructs. The proportion of colonies showing any transpositional activity depended strongly on the concentration of MnCl₂ used in the error-prone PCR. Using 250 μM MnCl₂, we recovered only one hypertransposer (A7) of 2,500 colonies screened, the vast majority of which were nulls or hypomorphs, presumably through introduction of multiple mutations. Using 100 μM MnCl₂, we were able to increase the number of colonies showing papillation of some degree to ≈30% and recovered 10 potential hypertransposers from 2,300 colonies screened.

A mating-out assay is used to measure quantitatively the relative frequency of transposition produced by individual transposases (31, 35). This assay measures the frequency with which an F plasmid is used as a target by a nonautonomous Himar1 transposon mobilized by a given transposase source. After mating the target F plasmids to a recipient strain, the transposition frequency is determined by measuring the ratio of F plasmids carrying KanR (the marker in Himar1) to all exconjugates (Fig. 2B).

Mutants A7 and C9 were particularly active in the papillation screen and so were chosen for further analysis in the mating-out assay. The A7 mutant was ≈10-fold more active in E. coli than the wild type whereas the C9 mutant was ≈50-fold more active (Fig. 3). Sequencing showed that both mutants contained multiple amino acid changes (two each in A7 and C9). By substituting wild-type sequences for mutated ones and testing the isolated mutant amino acid changes again in the papillation assay, it was possible to determine which amino acid changes actually conferred the hyperactivity. Doing this, we found that a H267R change in mutant A7 and both the Q131R and E137K changes in mutant C9 conferred hyperactivity (Fig. 1).Testing the individual C9 mutations in the mating-out assay showed that the Q131R mutation alone was ≈4-fold more hyperactive whereas the E137K alone was ≈20-fold more hyperactive. The combination of these mutations is ≈50-fold more active, and, thus, it would appear that these mutations act synergistically, and not simply additively.

Most of the seemingly hyperactive mutants isolated in the papillation assay were not hyperactive in the mating-out assay (data not shown). The reason for this is unclear, but false positive results have been reported when using similar assays with both Tn10 and Tn5 (31, 32). Wild-type Himar1 is most active at 30°C (13), so the fact that we isolated the mutants at 32°C and quantified their activity at 37°C may mean that A7 and C9 are more stable at the higher temperature used for the mating-out assay.

It is possible that the hyperactive mutations we recovered were attributable to some novel interaction with the E. coli host and not to some property in the transposase itself. To address this question, we examined purified A7 and C9 transposases in an in vitro transposition assay. This assay measures the relative ability of purified transposase to process and insert two ³²P end-labeled ITR DNA fragments into an unlabeled supercoiled DNA target, thus producing a labeled linear transposition product that is easily quantified (Fig. 4A). The rate at which this labeled product accumulates is a measure of transposition frequency. Fig. B and C shows that the A7 and C9 transposases were both hyperactive compared with the wild type. By measuring the slope of the linear portions of the curves in Fig. 4C (between 1 and 5 h), it is possible to compare the rate of product accumulation. Doing this, we found that the A7 transposase was 4.8-fold more active than the wild type whereas C9 was 7-fold more active. Thus, neither of the purified transposases was as active in the in vitro assay as in the mating-out assay. These results, however, do confirm that hyperactivity is intrinsic to the transposase protein and not wholly the result of some novel interaction with E. coli.

In vitro mutagenesis is rapidly becoming an important tool for studies of gene function. We have developed a technique using the Himar1 mariner system to mutagenize targeted genomic regions of the chromosome of a human respiratory pathogen, Haemophilus influenzae. (18). Analysis of such regions is enhanced by using large pools of mutants, which requires high efficiency of transposition. We examined whether the hyperactive transposases could improve mutagenesis frequencies in this system. Wild-type, A7, and C9 transposases were tested for the ability to move the minimariner elementTn-magellan1 carrying the gene for kanamycin resistance into a PCR-amplified chromosomal segment of H. influenzae. After repair of single-strand gaps introduced into the target DNA by the transposition reaction, DNA was transformed into competent H. influenzae cultures as described (18). The number of Kan-resistant H. influenzae colonies obtained with the wild-type, A7, and C9 transposases were 217 ± 122, 633 ± 50, and 733 ± 49, respectively. These results indicate that mutagenesis of H. influenzae with Himar1 could be significantly improved by the use of hyperactive transposases.

We were able to isolate two hyperactive mutants in an E. coli papillation screen that produced significant levels of hyperactivity in a quantitative mating-out assay. Both were double mutants, and, in one case (C9), both amino acid changes contributed synergistically to the overall hyperactivity, suggesting that additional combinations of mutants constructed directly, or by shuffling during the mutagenesis, might be even more hyperactive.

Although there is no structural data available for the Himar1 transposase, analysis of the mutant sequence along with comparison to other known transposases suggests the locations of functional domains. The only structural information for any of the Tc1-mariner superfamily of transposons is that for the Caenorhabditis elegans transposase Tc3, and then only for the specific DNA binding domain, a region that does not include the Himar1 mutations (37). Functional studies have been performed for both Tc1 and Tc3 transposases that demonstrated the existence of a separate, nonspecific DNA binding domain in each (38, 39). By comparing Tc1 and Tc3 transposases and computer models (40) with Himar1 transposase, we can assume that specific DNA binding is encoded by the first ≈113 amino acids of Himar1, nonspecific DNA binding by approximately amino acids 114–173, and catalysis by at least amino acids 158 and 287, the first and last amino acids of the D, D34D catalytic triad (Fig. 1). The C-terminal-most region is of unknown function. Given the fact that the Q131R and E137K mutations occur in a region of the transposase implicated in nonspecific DNA binding, the enhanced activity in these mutants may be attributable to increased affinity for DNA in general (Fig. 1). Similarly, the H267R mutation, which occurs in the putative catalytic domain, may be attributable to increased or altered catalysis. In Tn5 transposase, these various regions are known to overlap extensively, so a mutation in one region may effect a completely different property of the transposase (41). Indeed, the ways in which a transposase can become hyperactive are diverse. For example, at least three different classes of hyperactive mutations have been uncovered for Tn5. These affect the production of cotranslated inhibitor protein (42), an increase in the affinity of Tn5 transposase for ITR DNA (29), and a decrease in the self-inhibitory activity of intact Tn5 transposase (43). The combination of these three classes of hyperactive mutants are synergistic, leading to an extraordinarily active transposase (44).

Abundant sequence data is available for members of the Tc1-mariner family of transposons. Alignment of the available transposase sequences (3) allowed us to determine whether any of the amino acid replacements we uncovered are present in related transposases. Interestingly, one of the amino acid changes is present in a homologous position in the highly active Mos1 mariner. This transposase contains an arginine residue at the position of the Q131R mutation we isolated. This is not a highly conserved position in mariner transposases generally, although Tc1-like elements are biased toward basic residues at this position. The E137R mutation is not present in most other mariner family elements because this region is a unique small insertion in the irritans subfamily of mariner transposases to which Himar1 belongs. Finally, the H267R replacement of A7 is shared with one other member of the irritans subfamily (Hsmar2) and two members of the mellifera subfamily (Gpmar1 and Demar1), but, again, this is not a widely conserved position in mariner transposases.

Reznikoff and coworkers have stressed that the ability to isolate hyperactive transposases of Tn5 strongly suggests it has not evolved for maximal activity (41). Tn5 transposase mutants have been isolated that can increase the intrinsic activity of the transposase or eliminate regulatory mechanisms (29, 45). Low intrinsic activity and self-regulation appear to allow Tn5 to persist in E. coli without producing serious levels of genetic damage. Our ability to isolate hyperactive mutants for Himar1 may be attributable to similar evolutionary forces at work on mariner−family transposons. Horizontal transfer is a major feature in the evolutionary history of these mobile genes (4, 7, 46, 47). Clearly, the elements must be active enough to make copies of themselves when they transfer to a new host to persist. If not, they will be eliminated because of stochastic mechanisms (47). Their activity, however, cannot be so high as to significantly reduce the fitness of the host. Unregulated transposition can be highly deleterious to a host organism (48). Thus, the mariners that persist are not likely to be present in their most active forms. From the standpoint of copy number in the host organism, Himar1 is very successful, being present in 17,000 copies in the H. irritans genome (7). It may be that this success is attributable to a fairly benign level of activity in that genome because of either a comparatively low intrinsic activity level, some self-regulatory mechanism (13), or both. Our ability to isolate hyperactive forms of this transposase lends credence to this view, assuming that such mutants retain their hyperactivity in the natural host as well as in E. coli.

The mariner-family transposons Himar1 and Mos1 are being developed in a number of organisms for a variety of uses and hold great promise to extend the techniques of transposon-based genetic analyses available for E. coli and Drosophila to other organisms (12, 15, 18, 19, 49). For these elements to be used easily and efficiently, especially Himar1, their activity needs to be improved (13, 15). One way to do this is to establish optimum conditions under which transposition can take place. This method is most suitable for in vitro work by using purified transposases (13).

Alternatively, hyperactive transposase mutants such as those described here can be used and are well suited for both in vivo and in vitro work. The two mutants we recovered appear to be somewhat tolerant of high temperatures, which would be particularly useful in E. coli and human cells. Functional genomic applications of these transposons, like gambit (18), are improved by using these enhanced transposases. Clearly, labor-intensive methods such as germline transformation could be eased by more active transposases. A better understanding of the biochemistry of mariner transposases provided by mutagenesis methods described herein can be expected to provide transposases that have differing target sites, improved thermal stability, and enhanced activity in specific hosts. Finally, mutants may be uncovered that are amenable to crystallization and structural study, a prospect that is not currently possible because of the poor solubility of mariner transposases.