Correspondence: Ariel Imre, Veterinary Medical Research Institute of the Hungarian Academy of Sciences, Budapest, Hungary. Tel.: +36 1 467 4060; fax: +36 1 467 4076; e-mail: firstname.lastname@example.org
Site-directed integration/mutagenesis systems are used to carry out targeted transpositions on DNA. The well-characterized IS30-element and its transposase have numerous advantages that predestine it to be a good candidate for such applications. In order to generate nonflagellated mutants of Salmonella Enteritidis, a new site-directed mutagenesis system has been developed and applied. The system was constructed based on the assumption that the DNA-binding FljA component of the fusion transposase would bind to its target (the operator of fliC), and as a consequence, insertions could be concentrated in the flagellin operon. The system consists of two components: one expresses the fusion transposase and the other is an integration donor plasmid harbouring the (IS30)2 reactive structure. The application of this site-directed mutagenesis system on a strain of S. Enteritidis 11 (SE11) resulted in several nonmotile mutants with fliD insertion that could serve as negatively markered vaccine candidates. Analysis of less motile mutants generated by the fusion transposase revealed further hot spot sequences preferred by the fusion construct.
Insertional transposon mutagenesis is a frequently used technique with the enormous advantage not only of the generation of new phenotypes, but the identification of the mutated gene directly. Transposon mutagenesis can be achieved by several means including both random and site-directed methods. Site-directed or targeted mutagenesis mediated by insertion sequence (IS) elements and transposons relates to the use of a novel recombinant DNA technology for the targeted modification of DNA. Because of their ability to generate insertions, IS elements and transposons represent a useful and efficient tool in biotechnology by introducing ‘foreign’ DNA into the genome of various plants, animals or bacteria (for a review, see Coates et al., 2005; Kolb et al., 2005; Voigt et al., 2008).
There are two major ways of modifying the mobile element enable it able to carry out targeted transposition. One can alter the characteristics of the transposition itself by modifying the specificity of the transposase and/or its target sites. A good example is the use of modified IS30 transposase as a site-specific recombinase (Kiss et al., 2003). On the other hand, it is more frequent to relay new DNA-binding specificity to transposases by adding/replacing their DNA-binding domains with that of heterologous DNA-binding proteins (Bushman, 1994; Szabo et al., 2003; Feng et al., 2010). This technology allows the delivery of DNA fragments into a single integration site or into a series of integration sites in the chromosome of prokaryotes and eukaryotes. In this targeting technique, a chimeric protein generally consisting of a recombinase (site-specific recombinase, transposase) and a DNA-binding domain of DNA-recognition enzymes (repressors, activators, etc.) is used to mediate integration into the neighbourhood of a specific DNA sequence.
The well-characterized IS30-element (Olasz et al., 1993, 1998; Kiss & Olasz, 1999; Szabo et al., 2003; Nagy et al., 2004) and its transposase have numerous advantages that predestine it to a promising candidate for applications in site-directed systems. Based on the favourable properties of IS30, we developed the first transposon-based targeting system (Szabo et al., 2003). The modification of IS30 transposase by fusion resulted in the recognition of the binding site of the unrelated DNA-binding domains both in Escherichia coli and in zebrafish. The insertions occurred in the close vicinity of the binding site: a few hundred base pairs from the binding site in E. coli and within 100 bp in zebrafish. This kind of target specificity can be explained by tethering the transposase to a specific DNA sequence.
A specific property of the biphasic Salmonellae is the presence of the flagellin genes (fliC and fljB) at different locations on the chromosome, expressing different flagellins, that could help the bacteria to evade the host's immune reactions (Macnab, 1996). The genes encoding for the different flagellar phases (H1, H2) are highly similar, although not identical (Okazaki et al., 1993). The flagellin gene fliC codes for phase H1, while fljB is responsible for the production of flagellin phase H2 (Fig. 1a). Besides the typical, biphasic Salmonella serovars described above, there are several monophasic serovars lacking the phase variation system or carrying mutations in some of those elements. A classical example is Salmonella Enteritidis in which neither the phase variation system nor the fljAB genes can be found; therefore, only phase H1 flagellin is produced (Fig. 1b). Earlier studies reported that fliC mutants of S. Enteritidis can be attenuated (Parker & Guard-Petter, 2001), and as such, could be used as potential vaccine strains.
Here, we aimed to provide a new site-directed mutagenesis system using IS30 transposase fused to a specific DNA-binding protein, the flagellin repressor FljA, to insert the transpositionally active (IS30)2 intermediate (Olasz et al., 1993; Kiss & Olasz, 1999) close to the operator of the fli operon. This new system was used to generate nonflagellated (negatively markered) mutants of S. Enteritidis for the subsequent development of potential live oral vaccines.
Materials and methods
Escherichia coli laboratory strains TG2 (Gibson, 1984) and E. coli S17-1λpir (Simon et al., 1983) were used for molecular cloning. Salmonella enterica serovar Enteritidis 11 PT1 (SE11) is a wild-type (wt) strain, isolated from poultry and designated as E296 in an earlier study on flagellar systems (Imre et al., 2005). Chromosomal DNA of S. enterica serovar Typhimurium 1868 (a gift from M. Susskind) was used as a template for amplifying and cloning the fljA gene.
Microbiological and molecular techniques
For culturing bacteria, the following media were used: Luria–Bertani broth and agar (Sambrook et al., 1989), for molecular biological techniques. SOC medium (Sambrook et al., 1989) was used for the resuspension of bacterial cells after electrotransformation. Antibiotics (Sigma) were used in the following final concentrations: ampicillin (Ap), 150 μg mL−, and chloramphenicol (Cm), 20 μg mL−.
Standard molecular methods were applied according to Sambrook et al. (1989). Restriction endonucleases, Taq polymerase, T4 DNA ligase, RNaseA, proteinase-K and chemicals were purchased from Fermentas, New England Biolabs, Amersham, Sigma, Roche and Roth.
The wt IS30 transposase producer pJKI132 plasmid was described previously (Farkas et al., 1996). For the construction of the IS30–FljA transposase producer and integration donor pFOL1069, see Fig. 2.
For the mutagenesis process, the IS30–FljA fusion transposase producer pFOL1111 plasmid was electroporated into the SE11 strain and transformants were selected for the ApR marker of the plasmid. This was followed by the conjugal transfer of the pFOL1069 insertion donor plasmid into the SE11(pFOL1111) strain and the transposon mutants were selected according to their prototroph, CmR phenotype (Fig. 2). In the control experiment, the pJKI132 plasmid was used instead of pFOL1111, expressing the wt IS30 transposase.
Determination of insertion sites
For the determination of the insertion sites of pFOL1069, genomic DNA was isolated from the bacteria and digested with the ClaI enzyme. The resulting genomic fragments were self-ligated using T4 DNA ligase and transformed into E. coli S17-1 λpir bacteria. In the S17-1 λpir strain, only the recircularized pFOL1069 insertion derivatives were able to replicate as recombinant plasmids carrying the flanking regions of the insertion site. The exact site of pFOL1069 insertion was determined by sequencing of purified plasmid DNA (ABI Prism 310).
Results and discussion
Construction of a site-directed IS30–FljA transposon mutagenesis system to produce a nonflagellated mutant of SE11
The fact that in S. Enteritidis the elements of the phase variation system are absent and the fliC operon is present at the same time (Imre et al., 2005) made this serovar an excellent target for the directed transposon mutagenesis based on the FljA–IS30 fusion. Our aim with this new mutagenesis system was to modify the original target specificity of IS30 transposase by joining it to the DNA-binding FljA repressor protein. With this fusion protein, we established a directed transposon mutagenesis system that is expected to directly integrate close to the fliC operator. The system is composed of three main elements: (1) the target fliC operator flanked by fliC and fliD genes; (2) the IS30–FljA fusion transposase; (3) and the integration donor sequence containing the (IS30)2 intermediate together with the CmR marker gene. Two essential components of the mutagenesis system required to be constructed: the fusion transposase producer plasmid pFOL1111 and the integration donor pFOL1069 (Fig. 2).
The insertion donor plasmid pFOL1069 containing the (IS30)2 intermediate (Olasz et al., 1993; Kiss & Olasz, 1999) represented a highly reactive DNA segment in the presence of the IS30 transposase. The pFOL1069 additionally contained the CmR marker gene, the mob region necessary for bacterial conjugation and the defective replication origin R6K. Because of the R6K replication origin, the donor plasmid is unable to replicate in Salmonella lacking the pir gene. As a consequence, Salmonella bacteria possessing CmR after the conjugation of pFOL1069 were the ones in which the donor plasmid was integrated into the chromosome. The integration ability of pFOL1069 was verified earlier in E. coli (data not shown).
The FljA–IS30 fusion transposase producer plasmid pFOL1111 was constructed by the fusion of the fljA flagellin repressor gene to the C-terminal end of the IS30 transposase gene. The resulting isopropyl-β-d-thiogalactopyranoside-inducible FljA-transposase producer plasmid also contains the ApR bacterial marker. Because this plasmid codes only for the fusion transposase, but lacking the IS30 inverted repeat ends necessary for transposition, it is not capable of integrating into any target DNA. The inducible expression of the FljA–IS30 fusion protein was verified by sodium dodecyl sulphate polyacrylamide gel electrophoresis (Fig. 3a). No alteration was detected in the amount of the transposase compared with that of the wild type produced by the plasmid pJKI132 (Fig. 3a). The functionality of the FljA part of the fusion was tested by introducing pFOL1111 into the wt S. Enteritidis strain 11 and the motility of the transformants was investigated. The pFOL1111 plasmid-harbouring strains (Fig 3b, column 2) showed reduced motility as compared with the plasmid-free bacteria (Fig 3b, column 1). However, a complete elimination of motility never occurred due to the presence of exogenous FljA, and it was always reversible, as the partly motile strains regained their full motility after the plasmid pFOL1111 was eliminated (results not shown). The transposition activity of pFOL1111 was verified similarly as described by Szabo et al. (2008) (data not shown). In summary, it can be stated that all components of the targeting system have proven their expected activity for subsequent immobilization.
Targeted mutagenesis in S. Enteritidis strain 11
To produce insertion mutants, the fusion transposase producer plasmid pFOL1111 was electrotransformed into the wt S. Enteritidis 11 (SE11) strain. After selecting for the ApR marker of the plasmid, the presence of pFOL1111 and the expression of IS30–FljA fusion transposase were confirmed. Subsequently, the insertion donor pFOL1069 from E. coli S17-1 λpir bacteria was conjugated to SE11(pFOL1111)ApR and the transconjugant bacteria were selected for CmR of pFOL1069 and the auxotrophy of the wt S. Enteritidis strain (Fig. 2). In the control experiment, the wt IS30 transposase producer plasmid pJKI132 was used instead of pFOL1111, where only the IS30 transposase was expressed without the FljA domain. In this case, the insertion pattern of wt IS30 was expected due to the lack of the FljA-specific DNA-binding ability. Performing the transposon mutagenesis on the wt SE11 strain using both the IS30–FljA fusion or the wt IS30 transposase, the results of three independent experiments (Supporting Information, Table S1) showed that the transpositional frequency mediated by the IS30–FljA fusion transposase (1.78E-04–1.62E-04) was as high as that of the wt IS30 transposase (1.45E-04–8.35E-05). The data indicated that the fusion transposase maintained full activity compared with the wild type. The CmR transposon mutant Salmonella bacteria carrying pFOL1069 insertion in their genome were selected and tested for motility.
Nonmotile mutants generated by the IS30–FljA fusion transposase.
As a result of the mutagenesis experiments, altogether 1200 randomly selected ApRCmR SE11 transposon mutants were isolated and investigated: 600 were generated by the IS30–FljA fusion transposase and 600 by the wt IS30 transposase, respectively. The motility of the mutants was tested individually using the motility agar tube test. Four out of 600 mutants (0.67%) generated by the site-directed system proved to be completely nonmotile. In contrast, no nonmotile mutants were detected among the 600 mutants (<0.16%) generated by the wt IS30 transposase.
At least three of the four nonmotile insertional mutants could be considered as independent mutants, originating from three independent experiments (Fig. 3b, column 3). These insertional mutants were confirmed as nonflagellated phenotypes using S. Enteritidis-specific Hg,m antiserum. At the same time, all of the four investigated mutants retained their agglutinability in group D antiserum. Thus, they were confirmed as flagella-free derivatives of SE11.
Analysis of the target sites in nonmotile and less motile mutants
In order to determine the target specificity of the IS30–FljA fusion transposase, altogether 40 different pFOL1069 insertions were cloned (see Materials and methods) and the integration sequences were identified. On analysing the target sequences (Table 1a), it was found that the IS30–FljA fusion transposase show pronounced target specificity. The consensus sequence derived from 24 insertion sites (Table 1b) showed high similarity to the previously determined CIG consensus of insertions of the wt IS30 in the genome of E. coli. The consensus sequence was generated according to the following criteria: (1) a single base was accepted at a given position if it occurred there with at least 40% frequency and (2) alternative bases were used if two bases were found in a certain position with about equal frequency and they represented at least 70% of the cases (Table 1c, Olasz et al., 1998). At the same time, out of the 22 conserved nucleotide positions of the CIG, 17 positions were identical to the 40C consensus generated by the IS30–FljA fusion transposase. The 40C consensus was generated similar to the CIG consensus, i.e. a single base at a given position was accepted if it occurred there with at least 40% frequency. These results allow us to conclude that the fusion transposase retained its IS30-like target specificity.
Table 1. Summary of insertion sites targeted by the IS30–FljA fusion protein
Another important attribute of the IS30 transposase is the multiple usage of a preferred – so-called hot spot – target sequence. Having analysed the insertion sites, the fusion transposase chose the same sites several times. We identified four preferred target sequences that were chosen at least three times by the fusion transposase (Table 1). These sequences showed pronounced similarity to both the 40C consensus of the IS30–FljA and the CIG consensus of IS30 (Table 1). One of the four hot spots was located in the fliD gene mentioned. Three mutants (i115, i116, i118) out of the four nonmotile mutants proved to carry insertions in the fliD gene (NP_460913 in S. Typhimurium LT2 strain) exactly at the same location (Table 1a and Fig. 3c). This result indicated that in these nonmotile isolates, the insertion occurred close to the recognition site of the FljA protein. It should be noted that based on alignments with 40C consensus insertions in fliC were also expected. However, further analysis using more stringent consensus sequences indicated that the hotspot in fliD could be more attractive (results not shown). Determination of the insertion site in the fourth mutant indicated that pFOL1069 insertion occurred in the putative yjjY gene (assigned as NP_463455 in S. Typhimurium LT2 strain). The yjjY gene is located on a different segment of the Salmonella chromosome as a putative inner membrane protein gene without any functional description. The second hot spot (18i2 – three isolates) was found in the terminator sequence of the transposase producer plasmid itself, while the third (136i1 – three isolates) was in an intergenic region of the Salmonella chromosome.
The fourth, and the most preferred, hot spot (17i1) was located in the putative gene yjjY where 11 insertions from three independent experiments were identified exactly in the same position. The inserted pFOL1069 was found in both orientations. In order to verify whether this site was a very frequent hot spot, 278 mutants were tested by PCR (see Fig. S1). We found that pFOL1069 integrated into the putative yjjY gene in 48/278 cases. Regarding the phenotype, most of the yjjY mutants (23/48) showed strongly reduced motility. However, only 24 mutants in total were found to be less motile out of the 278 tested. To eliminate the disturbing effect of the fusion protein (Fig. 3b), the fusion transposase producer plasmid was eliminated from five yjjY mutants and the motility of these strains was tested again. Reduced motility was observed in all cases, indicating that in (or close to) the yjjY gene, a DNA segment is located that affects motility.
Because the sequence of the yjjY insertion site showed high similarity to the consensus used by the wt IS30 transposase, we tested whether the wt IS30 uses this target sequence as a hot spot. Only seven yjjY mutants were found to be generated by the wt IS30 out of the 222 mutants tested. These data demonstrate that the fusion transposase has a much more pronounced target preference for the yjjY hot spot (17.3%) compared with that of the wt transposase (3.2%).
In this study, we have worked out and successfully applied a novel method based on IS30-mediated site-directed mutagenesis in order to produce nonflagellated S. Enteritidis mutants. The system was constructed based on the assumption that the FljA repressor component of the fusion transposase – as a DNA-binding protein – would bind to its target (the operator of fliC), and as a consequence, insertions could be concentrated with a relatively high frequency in the flagellin operon. The system constructed on the above basis worked well and generated insertions. It turned out that the sequenced insertion sites showed pronounced similarity to the IS30 consensus sequence of insertions (Table 1; Olasz et al., 1998). This indicated that the fusion transposase retained the target recognition ability of the wt IS30 transposase. Another feature of the insertions was that four target sites – called hot spots – were utilized several times. One of these hot spots was the target sequence in the fliD gene and these insertions resulted in nonmotile phenotypes. This fact could be considered as a proof of FljA-targeted transposition, because fliD is located in close proximity to the fliC operator sequence, which is the binding site of the native FljA repressor protein. These data suggested that the fusion of the FljA repressor protein modulated the target preference of the IS30 transposase and increased the frequency of integration into a new target site not preferred by the wt transposase. This result is in good agreement with earlier observations that the target preference of IS30 transposase can be modified by fusing the enzyme to unrelated DNA-binding proteins (Szabo et al., 2003 and unpublished data).
Unexpectedly, another highly preferred hot spot was identified in the putative gene yjjY. Although this target site was recognized by both the wt and the fusion transposase, the frequency of the mutations generated by the IS30–FljA transposase was almost six times higher than that of the wild type (17.3% vs. 3.2%). The fact that these insertions resulted in a less motile phenotype in Salmonella implied that yjjY or its neighbouring sequences/genes may influence motility indirectly. This might also indicate that this unknown function could be under the control of the FljA protein. It is tempting to speculate that a site-directed integration event also occurred in the case of the yjjY mutants.
An IS30-based site-directed integration system could be utilized in several ways, for example to search for and to tag the targets of DNA-binding proteins in vivo. The IS30 transposase has a number of features that make the further development of the IS30-based site-directed integration system as a tool for functional genomics worthwhile. These advantages include the high activity of the (IS30)2 intermediate structure (Olasz et al., 1993; Kiss & Olasz, 1999; Table S1), the lack of size limitations (high-molecular-weight plasmids can be integrated as well – unpublished data), the integrated product is stable in the absence of the IS30 transposase because IS30 is not present in the vast majority of bacteria and IS30 is active both in bacteria and in eukaryotes (Szabo et al., 2003). Fusion of the IS30 transposase with transcription factors, repressors, DNA methylases or with any other kind of DNA-binding proteins may establish a vast array of potential integration sites. A further advantage of this mutagenesis system is that it might be useful in such cases when the sequence of the target gene is not known (e.g. new isolates of pathogenic bacteria), and the traditional molecular methods (e.g. Datsenko & Wanner, 2000) cannot be applied. In such a situation, the adaptation of this technique is more promising.
Because of the absence of flagellae, the lack of antiflagellar antibodies can be used as a negative marker in the serological differentiation of vaccinated chicks from those infected by wild strains of S. Enteritidis (Adriaensen et al., 2007). It is believed that nonmotile mutants produced by our site-directed mutagenesis method could also aid the development of a negatively marked vaccine against S. Enteritidis infection of chicks.
This study was supported by the Hungarian Grant NKFP 4/040/2001 and in part by the EU FP6 SUPASALVAC and CRAB (LSH-2004-2.1.2-4) Program. We thank M. Szabó, J. Kiss and Z. Nagy for their helpful advice and fruitful discussions on molecular techniques. We also thank I. Könczöl, E. Keresztúri and M. Turai for their skilful help with the bacterial techniques.