Generation of Single-Copy Transposon Insertions in Clostridium perfringens by Electroporation of Phage Mu DNA Transposition Complexes

Abstract

Transposon mutagenesis is a tool that is widely used for the identification of genes involved in the virulence of bacteria. Until now, transposon mutagenesis in Clostridium perfringens has been restricted to the use of Tn916-based methods with laboratory reference strains. This system yields primarily multiple transposon insertions in a single genome, thus compromising its use for the identification of virulence genes. The current study describes a new protocol for transposon mutagenesis in C. perfringens, which is based on the bacteriophage Mu transposition system. The protocol was successfully used to generate a single-insertion mutant library both for a laboratory strain and for a field isolate. Thus, it can be used as a tool in large-scale screening to identify virulence genes of C. perfringens.

Clostridium perfringens is a gram-positive, anaerobic bacterium that forms heat-resistant spores. It is widespread in the soil and commonly found in the gastrointestinal tract of mammals. It has been implicated in several medical conditions in humans, ranging from mild food poisoning to necrotic enteritis and gas gangrene. C. perfringens strains also cause a variety of important diseases in domestic animals, including several enteric syndromes, such as enterotoxemia in cattle, sheep, and pigs, necrotic enteritis in poultry, and typhocolitis in equines (17, 40).

Understanding the pathogenesis of these infections is of crucial importance for the development of new tools for the prevention and control of C. perfringens-related diseases. Genetic modification is a valuable approach to identify new virulence factors and to study their role in the pathogenesis of C. perfringens.

Since the 1980s, several tools for manipulation of C. perfringens at the molecular level have been developed (1, 5, 28, 35, 38). Among these tools, transposon mutagenesis is a method that is widely used for identification of virulence genes. Until now, the only reproducible method for transposon mutagenesis in C. perfringens was based on Tn916, a tetracycline resistance-encoding conjugative transposon originally isolated from Enterococcus faecalis (10, 11, 13). Tn916 has been used extensively for transposon mutagenesis due to its broad host range and has been proven to be valuable for the identification of genes in C. perfringens (3, 7, 22). Nevertheless, this method has major disadvantages; multiple Tn916 insertion events occur with an incidence of 65% to 75%, severely complicating identification of genes responsible for phenotype changes (3, 7, 19). Furthermore, Tn916 is still active after insertion, resulting in unstable mutants (6, 39, 42). To our knowledge, generation of Tn916-derived transposon mutants in C. perfringens field strains has never been described.

Although a variety of transposon mutagenesis methods are available for gram-positive bacteria (4, 37, 41, 43), the inherent species nonspecificity, as well as the lack of mobility of the integrated transposon, makes the bacteriophage Mu-based transposon delivery system a system of choice for a variety of species (16, 26, 46). The Mu transposition approach includes in vitro assembly of a complex between the transposon DNA and the transposase enzyme, the transpososome, followed by delivery of the transpososome into the recipient cells. Once inside a cell, the Mu transpososome becomes activated in the presence of divalent cations, resulting in genomic integration of the delivered transposon. The bacteriophage Mu transposition system is also functional in vitro (15, 32, 33), in contrast to the Tn916 mutagenesis strategy, which is restricted to transposon mobilization in vivo following conjugation or electroporation. Under the optimal in vitro conditions, the Mu transposition reaction requires only the MuA transposase, a mini-Mu transposon, and target DNA as macromolecular components (15).

In this study, a novel protocol is described for transposon mutagenesis in C. perfringens that exploits the bacteriophage Mu transposition system. To our knowledge, this report is the first report describing a mutagenesis method generating single-insertion transposon mutants in laboratory and field isolates of C. perfringens. This method is important for the identification of C. perfringens virulence factors involved in the numerous diseases caused by this bacterium.

MATERIALS AND METHODS

Bacterial strains and culture conditions.

C. perfringens strain JIR325 is a rifampin- and nalidixic acid-resistant derivative of strain 13, a toxinotype A strain originally isolated from the soil (22). C. perfringens strain 56 was isolated from the gut of a broiler chicken having necrotic lesions in the intestine (14). Escherichia coli laboratory strain DH5α (45) was used for routine plasmid DNA isolation.

C. perfringens strains were grown anaerobically at 37°C in brain heart infusion broth (Oxoid, Basingstoke, United Kingdom) or TGY broth (30 g tryptone [Oxoid], 20 g Bacto yeast extract [BD Biosciences, San Jose, CA], 1 g glucose [Sigma, St. Louis, MO], and 1 g l-cysteine [Sigma]). The solid medium used for C. perfringens consisted of 3.9% Colombia agar base (Oxoid) supplemented with 5% defibrinated sheep blood (International Medical, Brussels, Belgium) or 2% egg yolk in plates (referred to below as egg yolk agar plates).

E. coli DH5α was grown aerobically at 37°C in Luria broth (Sigma).

When appropriate, plasmid maintenance and genomic transposon insertions were selected by addition of erythromycin (Sigma) to the growth media at the following concentrations: for E. coli, 200 μg/ml; and for C. perfringens, 10 μg/ml.

Plasmids.

Plasmid pLEB620 is a pUC19-derived carrier plasmid for the Em-Mu minitransposon (26). Em-Mu includes the ermB gene from Lactobacillus reuteri, which encodes resistance to macrolides, lincosamides, and streptogramin B antibiotics. In the transposon termini as an inverted repeat, the resistance cassette is flanked by a pair of 50-bp segments from the Mu right end, including critical MuA transposase binding sites (26, 33). Plasmid pTCATT is a derivative of the E. coli-C. perfringens shuttle vector pJIR410 (8). Besides the ermB gene, pTCATT carries the open reading frame of catP and can therefore be used as a reporter system. In this work, pTCATT was used as a replicative control plasmid to determine the competency statuses of the strains.

Electrocompetent cells.

Electrocompetent E. coli cells were prepared essentially as described by Sambrook and Russell (30). Electrocompetent C. perfringens cells were prepared using the method of Scott and Rood (35), with minor modifications.

C. perfringens cells were cultured overnight in liquid brain heart infusion medium and diluted 1:30 in 12 ml of TGY broth. The cells were grown to an optical density of 0.2 to 0.3 and harvested by centrifugation at 4,300 rpm (∼650 × g) for 20 min at 21°C. The cells were then rinsed twice with 1.2 ml of electroporation buffer (272 mM sucrose, 7 mM sodium phosphate buffer [pH 7.4]). Between the rinse steps, the cells were collected by centrifugation at 13,000 rpm (∼1,250 × g) for 10 min at 21°C. After the second rinse, 1.2 ml of electroporation buffer containing lysostaphin (Sigma) was added, and the cells were incubated for 1 h at 37°C. The two rinse steps with electroporation buffer were repeated. Finally, the cells were resuspended in 1.2 ml of electroporation buffer and divided into 400-μl aliquots.

Isolation of the Em-Mu transposon and transpososome assembly.

For transposon mutagenesis in C. perfringens, the Em-Mu minitransposon was used (26). Isolation of the Em-Mu transposon and transpososome assembly were performed as described by Pajunen et al. (26). Briefly, plasmid pLEB620 containing the Em-Mu transposon was propagated in E. coli DH5α and isolated using a plasmid midi kit (Qiagen, Hilden, Germany).

The Em-Mu transposon was released from pLEB620 by BglII (Sigma) digestion. The digested DNA was extracted sequentially using phenol and chloroform and concentrated by ethanol precipitation. The transposon fragment was purified chromatographically using an anion-exchange column (MonoQ HR 5/5; Pharmacia Amersham Biosciences, Piscataway, NJ). Fractions containing Em-Mu were pooled and concentrated by ethanol precipitation.

The in vitro transpososome assembly reaction mixture (80 μl) consisted of 4.4 pmol Em-Mu, 19.6 pmol (1,600 ng) MuA transposase (Finnzymes, Espoo, Finland), 50% (vol/vol) glycerol, 150 mM Tris (pH 6), 0.025% Triton X-100, 150 mM NaCl, and 0.1 mM EDTA. The reaction was carried out at 30°C for 4 h. Eight transpososome assembly reaction mixtures were pooled and concentrated by polyethylene glycol precipitation essentially as described previously (31). The pellet was resuspended in 50 μl TGD buffer (10 mM Tris-HCl [pH 6], 0.5% glycerol, 0.1 M dithiothreitol). Transposition complexes were stored at −80°C unless otherwise indicated. Successful complex assembly was monitored on 2% agarose (Nusieve 3:1; Cambrex) gels containing 87 μg/ml of heparin and 87 μg/ml of bovine serum albumin as described previously (21).

Transposon mutagenesis.

For electroporation, electrocompetent cells (400 μl) were mixed with the transpososome preparation (1 μl) on ice and transferred to a prechilled 0.2-cm electrode spacing cuvette (Bio-Rad, Hercules, CA). The cells were incubated on ice for 10 min. Electroporation was then performed with a Gene Pulser Xcell electroporation system (Bio-Rad) using the following settings: 400 Ω, 1.25 kV, and 25 μF. Following the pulse, the cells were incubated on ice for 10 min and then transferred to 1 ml of TGY broth containing 10 mM MgCl₂. JIR325 cells were incubated for 2 h at 37°C following the pulse. Depending on the protocol being tested, cells of field isolate 56 were incubated for 3 h at 37°C or for 4 h at 20°C or 30°C following the pulse. In the optimized protocol for field isolate 56, the cells were first incubated at 37°C, and then 1 μg/ml of erythromycin was added and the cells were incubated for another 1 h at 37°C in the presence of erythromycin. Following a total incubation time of 2, 3, or 4 h, the transformed cells were plated on egg yolk agar plates containing 10 μg/ml of erythromycin.

Southern blotting.

Genomic DNA was isolated from C. perfringens using the cetyltrimethylammonium bromide method (44) and was subsequently digested with PsiI (Fermentas, Burlington, Canada). The resulting fragments were separated by agarose gel electrophoresis and subsequently transferred onto a nylon membrane (Roche Diagnostics, Basel, Switzerland) with 1× SSC (0.15 M NaCl plus 0.015 M sodium citrate; Roche Diagnostics). Fixation was done by heating the membrane for 2 h at 80°C. A digoxigenin (DIG)-labeled Em-Mu probe was synthesized with a PCR DIG probe synthesis kit (Roche Diagnostics) using the following primers: Em-Mu probe fw (ACTGAATACTCGTGTCAC) and Em-Mu probe rev (GTCAGATAGATGTCAGACGC). For hybridization and immunodetection, a DIG Easy Hyb wash and block buffer set and CDP-Star (Roche Diagnostics) were used according to the manufacturer's guidelines.

Identification of transposon-flanking genome sequences.

To amplify the transposon-flanking genome sequences, a modification of the method of Kwon and Ricke (20) was used. Genomic DNA was prepared and digested with XapI (Fermentas). The Y-linkers were prepared as described previously (20) by annealing the following two oligonucleotides: Linker 1 (TTTCTGCTCGAATTCAAGCTTCTAACGATGTACGGGGACA) and Linker 2 (AATTTGTCCCCGTACATCGTTAGAACTACTCGTACCATCCACAT). The DNA fragments were ligated to Y-linkers (20) using T4 DNA ligase (Invitrogen, Merelbeke, Belgium). For amplification of the genome regions 5′ and 3′ of the inserted transposon, the primer pair Y-linker primer (CTGCTCGAATTCAAGCTTCT) and Em-Mu seq rev (ATCAGCGGCCGCGATC) and the primer pair Y-linker primer and Em-Mu seq fw (TCTGCAGACGCGTCGACGTCA), respectively, were used. Nucleotide sequences were analyzed using a BigDye Terminator v3.1 cycle sequencing kit (Applied Biosystems, Foster City, CA), a DyeEx 2.0 spin kit (Qiagen), and an ABI Prism 3100 genetic analyzer (Applied Biosystems).

Genomic transposon insertion sites were identified by comparing the sequences to the publicly available genomic sequences of C. perfringens strains 13, ATCC 13124, and SM101, using BLAST on the European Bioinformatics Institute server.

RESULTS AND DISCUSSION

Optimization of the electroporation process.

Several protocols have been described for electroporation of C. perfringens laboratory strain 13 (1, 2, 7, 18, 27, 35). However, field isolates often behave differently than laboratory strains, and the currently available electroporation protocols do not guarantee successful electroporation of field isolates.

As the electroporation efficiency is the key factor affecting bacteriophage Mu transpososome delivery, an optimal electroporation protocol needed to be generated for C. perfringens. To prevent premature activation of the transpososome, contact with divalent cations outside the bacterium should be avoided, and therefore, electroporation buffers should be prepared without Mg²⁺ or Ca²⁺ ions. Several electroporation protocols (1, 7, 18, 35) were tested for field isolate 56. The highest electroporation yield, obtained using control plasmid pTCATT containing an erythromycin resistance gene, was generated with the modified protocol of Scott and Rood (35). Due to the altered ion content of the Mg²⁺- and Ca²⁺-free buffer (272 mM sucrose, 7 mM sodium phosphate buffer [pH 7.4]), new electroporation parameters needed to be established. The optimal settings were determined to be 400 Ω, 1.25 kV, and 25 μF. Electrocompetent cells of both laboratory strain JIR325 and field isolate 56 were prepared using the same protocol, except for the concentration of lysostaphin used. As field isolate 56 appeared to be more sensitive to lysostaphin, a lower concentration of this compound (2 μg/ml) was used for this isolate than for strain JIR325 (10 μg/ml). Electrocompetent cells of field isolate 56 could not be stored at −80°C but needed to be freshly prepared prior to electroporation.

Transposon mutagenesis.

In each electroporation experiment, 1 μl of a transpososome preparation was delivered into strain JIR325 or isolate 56 cells. Following the pulse, the transformants were allowed to recover in TGY broth before antibiotic selection was applied on egg yolk agar plates containing 10 μg/ml erythromycin. For strain JIR325, incubation for 2 h in nonselective TGY broth was sufficient to allow plating on selective agar plates. However, transformants of field isolate 56 required an incubation time of 3 h in nonselective medium. Sequencing data (not shown) revealed that 3 h of incubation in nonselective medium resulted in the presence of multiple identical mutants in each batch. Thus, several strategies to prevent mutant amplification in nonselective TGY broth were tested. A decrease in the incubation temperature to 20°C or 30°C combined with a longer incubation time appeared to be unsuccessful. Finally, erythromycin was added to the incubation medium at a concentration below the MIC for both strains (i.e., 1 μg/ml) after 1 h of incubation; the transformants could be plated on selective agar plates 2 h after the pulse, and multiplication of the mutants was avoided.

The yield of transposon mutants fluctuated depending on the quality of the electrocompetent cells used. As a control, plasmid pTCATT was electroporated with each batch of electrocompetent cells. An absence or a low yield of transposon mutants was consistently reflected in the absence or low number of positive transformants obtained with pTCATT. Following electroporation of plasmid pTCATT, the numbers of erythromycin-resistant bacteria obtained for strain JIR325 and isolate 56 were 32,700 and 23,600 CFU/μg plasmid DNA, respectively. Reflecting this difference in the competency statuses of these strains, the yield of transposon mutants was higher for laboratory strain JIR325 (239 CFU/μg transposon) than for field isolate 56 (134 CFU/μg transposon). A total of 3,200 mutants of field isolate 56 were obtained.

The level of efficiency of the Em-Mu transposon insertion into C. perfringens is workable and is intermediate compared to the levels for other gram-positive bacteria. The yield is similar to the yields obtained with Streptococcus suis (100 CFU/μg DNA) and Lactococcus lactis (110 CFU/μg DNA), lower than the yields obtained with Staphylococcus aureus (20,000 or 12,000 CFU/μg transposon DNA depending on the strain), and higher than the yield obtained with Streptococcus pyogenes (10 CFU/μg transposon) (26, 46). Compared to the efficiency in gram-negative bacteria, the efficiency in C. perfringens is 2 to 3 orders of magnitude lower (21).

Genomic integration.

A PCR was performed to confirm the presence of the Em-Mu transposon in the genomic DNA of 50 mutants. Genomic DNA from 30 erythromycin-resistant isolate 56 mutants was obtained, digested with PsiI, which does not cut transposon DNA, and analyzed by Southern hybridization with a DIG-labeled Em-Mu transposon probe. All of the mutants analyzed contained a single copy of the Em-Mu transposon. The results of Southern blotting for seven mutants are shown in Fig. 1.

FIG. 1.

Southern blot detection of genomic fragments harboring the Em-Mu transposon insertion. Genomic DNAs from mutants M1 to M7 (lanes 1 to 7) and wild-type parental strain 56 (lane 8) were digested with PsiI and hybridized with a DIG-labeled probe for the Em-Mu transposon. Lane 9 contained purified Em-Mu transposon (1.4 kb) as a positive control. Lane M contained a DIG-labeled DNA molecular weight marker (Roche Diagnostics); the fragment sizes are (from top to bottom) 23.1, 9.4, 6.5, and 4.3 kb.

Identification of the Em-Mu insertion regions.

For identification of the transposon-flanking genome regions, a modification of the method designed by Kwon and Ricke (20) was used. After each successful mutagenesis experiment, mutants were randomly picked for sequencing. A total of 200 mutants were sequenced. A 5-bp target duplication is present in all the clones (Table 1), which is characteristic of Mu transposition in vivo and eliminates other types of DNA-restructuring reactions as the cause of genomic integration. The sequencing data revealed a relatively even distribution of integrations, although rRNA gene regions appeared to be favored. In the transposon insertion sites of the 200 mutants sequenced, protein-encoding genes comprised 44.5% of the integration sites. A total of 43% of the mutants carried a transposon inserted into one of the rRNA genes, and 12.5% had a transposon inserted into an intergenic sequence; 2% carried a mutation in a pCW9-like plasmid. Preferential insertion of the Em-Mu transposon into the rRNA gene clusters has also been described for Saccharomyces cerevisiae, in which a positive correlation was found between GC richness and MuA integration frequency (25). In C. perfringens strains, the number of rRNA genes varies between 23 and 29, accounting for about 1.5% of the organism's coding capacity (23, 36). Furthermore, it was reported previously that rRNA genes have a significantly higher G+C content than the rest of the genome (12, 36). Both the high copy number and the G+C content of the rRNA genes could be an explanation for the higher prevalence of rRNA mutants.

TABLE 1.

Integration sites of transposons in the mutants derived from field isolate 56^a

Mutant	Integration site	Genomic location	Description	Transposon orientation
M1	AAAGCTATTGTACTT-Em-Mu-TACCTAAAGATTTAT	2540298-2540302	Heat-inducible transcriptional repressor HrcA	+
M2	ACATTAAAATGTAAA-Em-Mu-GTAAAGACTGTGGAG	1431407-143411	Conserved hypothetical protein	−
M3	TGTTCAGCTGACCGA-Em-Mu-ACCGATACTAATAGA	77125-77129	23S rRNA	+
M4	CTTCAGTTTAACTGA-Em-Mu-ACTGAAAGTTCTTTG	3215472-3215476	Intergenic	−
M5	TTCTGCCTCTTCTGA-Em-Mu-TCTGATAGTATTAC	2963631-2963635	Methionine aminopeptidase, type I	−
M6	TTACCTTTGTAGTTA-Em-Mu-AGTTACAACATCTCT	332463-332467	Transcriptional regulator, AbrB family	−
M7	GTAAAGGTGTTTCAG-Em-Mu-TTCAGAAGAAAAAAT	530994-530998	BFD-like iron-sulfur cluster binding protein	+

^aTarget site duplications are indicated by boldface T, A, C, and G. All genomic locations were determined by comparison with the complete sequence of strain ATCC 13124 (GenBank accession number NC_008261) (18). All genes described were transcribed in the sense direction. Transcription of the transposon was compared to the local transcription within the genomic location.

Compared to the use of Em-Mu, the use of Tn916 has several limitations. First, Tn916 shows a preference for regions with sequence similarity to its transposon ends (i.e., a 5- to 7-bp run of adenines followed by a similar number of thymines) (34). These AT-rich regions are usually found in intergenic regions in low-G+C-content genomes (34). Preferential insertion of Tn916 in AT-rich regions has been described for other low-G+C-content bacteria, like Mycoplasma gallisepticum, Haemophilus influenzae, and Streptococcus mutans (9, 24, 29). The exact proportion of intergenic Tn916 insertion mutants for C. perfringens is not known, as the insertion spectrum of the transposon has always been analyzed by Southern hybridization and not by sequencing (7, 19). Second, only 25 to 35% of the Tn916-derived C. perfringens mutants carry a single copy of the transposon, while no multiple insertions were detected in the Em-Mu-derived mutants (3, 7, 19). Taking both the multiple insertion events and the preference of Tn916 for AT-rich intergenic regions into account, the proportion of single insertions into protein-encoding genes must be less than 25% to 35%. Third, gene regions can be removed when the Tn916 insertion is followed by a deletion event (3). No such deletion events have been described for Em-Mu.

In conclusion, a new protocol for transposon mutagenesis in C. perfringens that is based on the bacteriophage Mu DNA transposition system was developed. The method described is better than the protocols based on Tn916 used previously since mutants containing only a single copy of the transposon are generated and the proportion of hits in protein-encoding genes is higher. Furthermore, the method is also applicable to field isolates of C. perfringens, if they can be made electrocompetent.