In Vitro Transposition System for Efficient Generation of Random Mutants of Campylobacter jejuni

Abstract

Campylobacter jejuni is the most common cause of food-borne illnesses in the United States. Despite the fact that the entire nucleotide sequence of its genome has recently become available, its mechanisms of pathogenicity are poorly understood. This is in part due to the lack of an efficient mutagenesis system. Here we describe an in vitro transposon mutagenesis system based on the Staphylococcus aureus transposable element Tn552 that allows the efficient generation of insertion mutants of C. jejuni. Insertions occur randomly and throughout the entire bacterial genome. We have tested this system in the isolation of nonmotile mutants of C. jejuni. Demonstrating the utility of the system, six nonmotile mutants from a total of nine exhibited insertions in genes known to be associated with motility. An additional mutant had an inactivating insertion in sigma 54, implicating this transcription factor in flagellum regulation. The availability of this efficient system will greatly facilitate the study of the mechanisms of pathogenesis of this important pathogen.

Campylobacter jejuni is the most common cause of food-borne illnesses in the United States (11). Despite its importance as a human pathogen and the recent completion of the determination of the nucleotide sequence of its genome (13), little is known about its mechanisms of pathogenesis. This lack of knowledge is in part due to the lack of suitable tools for the efficient generation of random mutants that can be tested in relevant biological assays. Previous attempts to generate mutants have relied on shuttle mutagenesis and homologous recombination (2, 9, 19). More recently, an in vivo transposition system based on the Himar1 transposable element has also been reported (4). The weakness of this system, however, is that restriction of the suicide vector is unavoidable, severely affecting its efficiency. Although suitable for some applications, these low-efficiency systems are less optimal when the isolation of a large number of mutants is required. Difficulties in generating a high-efficiency mutagenesis system in C. jejuni are most likely due to the existence of powerful restriction barriers, inefficient expression of the appropriate transposase enzymes in vivo, or a combination of these and other factors. To overcome these limitations, we have designed an in vitro mutagenesis strategy based on the Staphylococcus aureus transposon Tn552 (14, 15). This transposon exhibits features that resemble those of Mu (12), such as the coding of a single-subunit transposase and the requirement of a single accessory protein for in vivo transposition (15). However, the ends of Tn552, consisting of only 48-bp terminal inverted repeats (10, 16), are much simpler than those of Mu and Tn7. These are the only sequences required for transposition, which facilitates the engineering of Tn552 for different applications. In contrast to Tn5 (5), Tn552 displays virtually no target preference (3, 6). A previous study has demonstrated that Tn552 can insert efficiently and randomly into target DNA molecules after in vitro transposition reactions (10). These properties have allowed the use of this element as a tool for nucleotide sequencing and mutagenesis (3, 6). We have designed a derivative of Tn552 (Tn552kan-Campy) that can be used in C. jejuni for the efficient generation of random mutants. We have tested its performance in an in vitro mutagenesis protocol to generate nonmotile mutants.

Construction of Tn552kan-Campy.

To construct a derivative of Tn552 with an antibiotic resistance gene that can confer resistance in C. jejuni, we replaced the chloramphenicol resistance gene (cat) present in Tn552cat (6) with the kanamycin resistance gene aphA-3 from Campylobacter coli (18) (Fig. 1). The Tn552cat-bearing plasmid pTG426 (6) was digested with BsrGI, which releases the cat cassette, and ligated to the aphA-3 gene of pILL600 (8), yielding the plasmid pSB1698, which carries Tn552kan-Campy. Since the aphA-3 gene can also confer kanamycin resistance in Escherichia coli, we compared the efficiency of transposition of Tn552kan-Campy with that of Tn552cat. One hundred nanograms of pBluescript SKII plasmid (Stratagene) DNA (as a target) was mixed with 100 ng (0.1 pmol) of Tn552kan-Campy or Tn552cat DNA obtained from pSB1698 or pTG426, respectively, by digestion with BglII and subsequent agarose gel purification. The transposition reaction was initiated by the addition of 30 ng of purified His-tagged TnpA transposase and carried out for 1 h at 37°C in a total volume of 20 μl. The transposition reactions and the purification of TnpA were carried out as previously described (6, 10). The reaction mixtures were ethanol precipitated and electroporated into E. coli DH5α, and transposon insertions were selected by plating the transformants on Luria broth plates containing either 50 μg of kanamycin or 30 μg of chloramphenicol per ml to select for Tn552kan-Campy or Tn552cat, respectively. A total of 1.38 × 10³ kanamycin-resistant colonies and 9.46 × 10² chloramphenicol-resistant colonies were obtained, indicating that Tn552kan-Campy can undergo transposition in vitro with an efficiency similar to that of the parent element Tn552cat (equivalent results were obtained in several repetitions of this experiment). In all cases, no antibiotic-resistant colonies were obtained in the absence of the TnpA transposase in the reaction mixture.

FIG. 1.

Diagram of Tn552kan-Campy. Only relevant restriction enzyme sites are shown. BglII cleavage creates the CA_OH-3′ ends needed for transposition and leaves a 5′-GATC single-stranded extension. L, left; R, right.

Generation of Tn552kan-Campy insertion mutants of C. jejuni.

We next tested the use of Tn552kan-Campy in the generation of random insertion mutants in C. jejuni. Chromosomal DNA was isolated from the C. jejuni strain 81–176 by using Qiagen (Qiagen, Inc., Valencia, Calif.) according to the manufacturer's instructions. One hundred nanograms of C. jejuni DNA was mixed with 100 ng (0.1 pmol) of agarose gel-purified Tn552kan-Campy DNA, and the transposition reaction was carried out as described above. The reaction mixture was ethanol precipitated and electroporated into C. jejuni 81–176 as previously described (8), and kanamycin-resistant colonies were selected on brain heart infusion (BHI) agar plates containing 30 μg of kanamycin per ml. Approximately 100 kanamycin-resistant colonies were recovered in several repetitions of this experiment, which represents a significant drop from the number of transposition events obtained using an E. coli plasmid as a target (see above). The reduction in the number of Tn552kan-Campy insertion events obtained with C. jejuni can be partially attributed to the requirement of homologous recombination for the rescue of kanamycin-resistant colonies (see below). However, we hypothesized that, most likely, DNA restriction of the E. coli-grown Tn552kan-Campy DNA was responsible for a significant portion of the reduction in the number of insertion events recovered in C. jejuni. We tested this hypothesis by constructing a C. jejuni shuttle plasmid carrying Tn552kan-Campy. This plasmid was constructed by cloning the 1.5-kb SpeI fragment from pSB1698 carrying Tn552kan-Campy into the unique SpeI site of the C. jejuni shuttle plasmid pRY112 (8). The resulting plasmid, pSB1699, was introduced into C. jejuni strain 81–176 by conjugation. Plasmid pSB1699 was isolated from C. jejuni 81–176 and digested with BglII to liberate Tn552kan-Campy. One hundred nanograms of C. jejuni 81–176 genomic DNA was mixed with 100 ng of Tn552kan-Campy DNA obtained from pSB1699 grown in C. jejuni 81–176, and the transposition reaction was carried out as described above. After electroporation of C. jejuni 81–176, between 2.85 × 10³ and 7.68 × 10³ kanamycin-resistant colonies were recovered per reaction in different independent experiments. In all cases, the absence of the TnpA transposase from the reaction mixture resulted in no antibiotic-resistant colonies after electroporation. The significant increase in the number of kanamycin-resistant colonies obtained using Tn552kan-Campy isolated from C. jejuni indicates that the low yield of insertions obtained with the E. coli-grown transposon DNA is probably the consequence of DNA restriction barriers between these two microorganisms.

Tn552kan-Campy insertions occur throughout the C. jejuni chromosome.

Tn552 has been reported to insert nearly randomly in plasmid targets (3, 6). To determine if Tn552kan-Campy inserted throughout the C. jejuni genome, Southern blot analysis of 16 randomly chosen kanamycin insertion mutants was carried out using purified Tn552kan-Campy DNA labeled with a randomly primed nonradioactive labeling kit (ECL RPN 3040; Amersham) as a probe. Chromosomal DNAs from the different C. jejuni strains were isolated as indicated above, digested with BspHI, separated on a 0.9% agarose gel, and transferred to a GeneScreen Plus membrane (NEN Life Science Products) according to standard procedures (17). BspHI was chosen because it cuts within the transposon. Therefore, when probed with Tn552kan-Campy DNA, the digested chromosomes were expected to produce two unique bands of different sizes if the different mutants were the result of a single insertion event. As shown in Fig. 2, all but one sample showed two unique bands representing unique transposon insertions. One sample produced a single band that additional experiments demonstrated to be the result of the overlap of two bands of similar sizes (data not shown). As expected, the wild-type C. jejuni did not show hybridization to the Tn552kan-Campy probe. These results indicate that the Tn552kan-Campy insertions are distributed throughout different sites of the C. jejuni chromosome.

FIG. 2.

Southern blot analysis of Tn552kan-Campy insertions in C. jejuni. Chromosomal DNAs from randomly chosen insertion mutant strains were digested with BspHI, separated on an agarose gel, transferred to a GeneScreen Plus membrane, and probed with labeled Tn552kan-Campy DNA. R1 to R16, C. jejuni Tn552kan-Campy insertion mutants; wt, wild-type C. jejuni 81–176.

To investigate if Tn552kan-Campy exhibits any preference towards certain regions of the C. jejuni chromosome, we determined the precise sites of insertion of several randomly chosen mutations by nucleotide sequencing. Genomic DNAs from the different C. jejuni insertion mutant strains were prepared as follows. Approximately 10¹⁰ C. jejuni cells were scraped from BHI plates and resuspended in 567 μl of TE (10 mM Tris [pH 8.0], 1 mM EDTA). Bacterial cells were lysed by adding 30 μl of 10% sodium dodecyl sulfate and 3 μl of 20-mg/ml proteinase K and incubating the cell suspension at 37°C for 1 h. Cell lysates were treated with 100 μl of 5 M NaCl and 80 μl of CTAB-NaCl solution (5% hexadecyltrimethylammonium bromide [CTAB], 0.5 M NaCl) and incubated for 10 min at 65°C. Cell lysates were then extracted once with chloroform, six times with phenol-choloform-isoamyl alcohol (25:24:1), once with chloroform, and twice with ether. DNA was ethanol precipitated and resuspended in 150 μl of TE. Ten micrograms of genomic DNA from each mutant was digested overnight with BspHI, ethanol precipitated, and resuspended in 11 μl of TE. Genomic DNA sequences were then determined using an ABI PRISM BigDye Terminator Cycle Sequencing Ready Reaction Kit (Applied Biosystems, Perkin-Elmer Corporation) according to the recommendations of the manufacturer and with a primer directed outward from the 5′ end of the aphA-3 kanamycin resistance gene (5′-CCAATTCTCGTTTTCATACC-3′). The sequence obtained from each of the mutants was compared to the nucleotide sequence of the C. jejuni NCTC 11168 chromosome (14). As shown in Table 1 and Fig. 3, all insertions analyzed occurred at different sites of the C. jejuni genome and showed no clustering in any particular area of the chromosome.

TABLE 1.

Location of Tn552kan-Campy insertions in the C. jejuni chromosome

Mutant	Locationa	Locus	Function
R1	21,949	CjaE	Putative membrane protein
R2	1,032,893	Intergenic region
R3	842,931	Cj0903c	Putative amino acid transport protein
R4	5,188	Intergenic region
R5	1,489,676	Cj1555c	Hypothetical protein
R6	1,043,086	Cj1109 (aat)	Putative leucyl- or phenylalanyl-tRNA, protein transferase
R7	1,100,770	Cj1171c (ppi)	Peptidyl-prolyl cis-trans isomerase
R8	843,422	Cj0903c	Putative amino acid transport protein
R9	584,529	hypB	Hydrogenase isoenzymes, formation protein
R10	1,070,352	Cj1135	Putative two-domain glycosyltransferase
R12	378,598	Cj0412	Putative ATP or GTP binding protein
R13	1,097,675	Cj1167 (ldh)	Putative l-lactate dehydrogenase
R14	552,826	Cj0595c (nth)	Probable endonuclease III
R15	309204	Cj0339	Putative transmembrane transport protein
R16	1,083,498	Cj1150c (waaE)	Putative ADP-heptose synthase
R17	343,343	Intergenic region
R18	1,077,395	neuC	N-Acetylglusomine-6-phosphatase

^aThe location represents the nucleotide position of the C. jejuni NCTC 11168 genome (13). 

FIG. 3.

Diagram of the C. jejuni chromosome showing the positions of the different Tn552kan-Campy insertions.

Isolation of C. jejuni nonmotile mutants using Tn552kan-Campy.

To test the utility of the Tn552kan-Campy in vitro transposition system, we screened a panel of in vitro-generated transposon insertion mutants for motility defects. We chose motility as a testing phenotype because several genes involved in this phenotype are well characterized and known to be located throughout the C. jejuni genome (7, 19). Nonmotile Tn552kan-Campy insertion mutants of C. jejuni strain 81–176 were identified in semisolid BHI agar (containing 37 g of BHI and 4 g of agar per liter of medium). Of 205 kanamycin-resistant mutants screened, 9 exhibited defects in motility. To identify the mutated genes, the transposon insertion sites were determined by sequencing analysis as described above. Six of the nine nonmotile mutants revealed transposon insertions in genes known or expected to be associated with motility (Table 2). The two remaining mutants had insertions in genes of unknown function not previously associated with motility. These results confirmed the usefulness of the Tn552kan-Campy in vitro transposition system for the isolation of random mutants.

TABLE 2.

Locations of Tn552kan-Campy insertions in C. jejuni that resulted in a motility defect

Nonmotile mutant	Location	Locus	Function
NM1	824,961	flaD	Putative flagellin
NM2	262,953	cheA	Chemotaxis histidine kinase
NM3	1,262,736	Cj1318	Unknown
NM4	357,157	Cj0390	Putative transmembrane protein
NM5	768,579	fliP	Flagellar biosynthesis protein
NM6	74,236	fliY	Putative flagellar motor switch protein
NM7	1,225,181	Cj1293 (flaA1)	Possible sugar nucleotide epimerase homologous to flaA1 from Helicobacter pylori and Caulobacter crescentus
NM8	624,961	rpoN	RNA polymerase sigma 54
NM9	1,639,362	flgE	Central flagellar-hook protein

Analysis of Tn552kan-Campy insertions.

In the system described here, interpretation of the phenotypes of transposon-promoted mutations depends on the assumptions that the recovered insertions do not contain gross rearrangements of the target DNA (such as large deletions) and that transposon-disrupted genes are incorporated by homologous recombination, replacing the wild-type chromosomal locus. Our expectation of homologous recombination was confirmed by PCR using primers complementary to chromosomal sequences bracketing the insertion site. In all six cases tested, the band corresponding to the preinsertion wild-type locus was lost and was replaced by a new band that was larger by ∼1.3 kb, which is the size of the transposon (data not shown).

DNA sequence analysis of the transposon-target junctions of 19 insertions (Fig. 4) showed that all were the result of simple concerted insertions at a single target site; none were accompanied by unexpected rearrangements such as deletions and inversions. However, target junctions were unusual and indicated that C. jejuni processes the gapped products of in vitro transposition very differently from E. coli. Only 3 of the 19 insertions were flanked by a simple target duplication of the type seen previously with many hundreds of in vitro Tn552 insertions recovered in E. coli (6). In most of the C. jejuni insertions, vestiges of a target duplication were observed, but in addition one junction contained from 1 to 4 bases derived not from the target sequence but from the 5′-GATC that extended from the 5′ end of the transposon substrate DNA. In three cases a portion (3 or 4 bases) of this GATC sequence had completely substituted for the target duplication.

FIG. 4.

Target site sequences of Tn552kan-Campy in C. jejuni 81–176 genomic DNA. Junction sequences were determined using outward primers complementary to the aphA-3-resistant gene located within Tn552kan-campy (5′ end, 5′-CCAATTCTCGTTTTCATACC-3′; 3′ end, 5′-GGATCAAGCCTGATTGGGAG-3′). Capital letters indicate the Tn552kan-Campy transposon, lowercase letters indicate the C. jejuni genomic DNA sequence, outlined letters indicate added nucleotides, and boldface letters indicate target site repeats. ‡, a single C is missing at the site of insertion.

In vitro, the Tn552 transposase joins the CA_OH-3′ transposon end to target DNA (attacking phosphodiester bonds on each DNA strand that are separated by a 6- to 9-base 5′ stagger), giving a product with 6- to 9-base gaps at each transposon target junction. Preexisting single-strand extensions on the 5′ ends of the transposon substrate (5′-GATC for the substrates used here) are unaffected by transposase in vitro and are efficiently removed in E. coli (probably by the 5′ “flap” exonuclease of DNA polymerase I) during the gap repair and ligation process. The most straightforward explanation for the retention of the 5′ flank in C. jejuni (and probable loss of some of the target duplication) is that it uses microhomologies to pair within the single-strand gap region and become ligated to the 3′ end of the duplex as repair synthesis proceeds.

While these anomalous target junctions have no affect on the utility of the transposon mutagenesis, their formation may be prevented by processing the gapped intermediate with DNA polymerase I (with deoxynucleoside triphosphates) in vitro before electroporation of C. jejuni.

Conclusions.

We have described an in vitro transposition system for the generation of random insertion mutants of C. jejuni. Unlike previously described systems, the Tn552-based system is very efficient, generating up to ∼8,000 mutants in a single reaction using 100 ng of target DNA. The efficiency of this system combined with the simplicity of the design of the element will allow the easy construction of derivatives suitable for signature-tagging mutagenesis, the isolation of in vivo-expressed genes, and random PhoA or LacZ protein fusions. In addition, the development of an in vitro transposition system combined with the availability of the entire nucleotide sequence of the C. jejuni genome will allow the application of specialized mutagenesis protocols to identify virulence as well as essential genes (1).