Figures
Abstract
Since the first isolation of type E botulinum toxin-producing Clostridium butyricum from two infant botulism cases in Italy in 1984, this peculiar microorganism has been implicated in different forms of botulism worldwide. By applying particular pulsed-field gel electrophoresis run conditions, we were able to show for the first time that ten neurotoxigenic C. butyricum type E strains originated from Italy and China have linear megaplasmids in their genomes. At least four different megaplasmid sizes were identified among the ten neurotoxigenic C. butyricum type E strains. Each isolate displayed a single sized megaplasmid that was shown to possess a linear structure by ATP-dependent exonuclease digestion. Some of the neurotoxigenic C. butyricum type E strains possessed additional smaller circular plasmids. In order to investigate the genetic content of the newly identified megaplasmids, selected gene probes were designed and used in Southern hybridization experiments. Our results revealed that the type E botulinum neurotoxin gene was chromosome-located in all neurotoxigenic C. butyricum type E strains. Similar results were obtained with the 16S rRNA, the tetracycline tet(P) and the lincomycin resistance protein lmrB gene probes. A specific mobA gene probe only hybridized to the smaller plasmids of the Italian C. butyricum type E strains. Of note, a ß-lactamase gene probe hybridized to the megaplasmids of eight neurotoxigenic C. butyricum type E strains, of which seven from clinical sources and the remaining one from a food implicated in foodborne botulism, whereas this ß-lactam antibiotic resistance gene was absent form the megaplasmids of the two soil strains examined. The widespread occurrence among C. butyricum type E strains associated to human disease of linear megaplasmids harboring an antibiotic resistance gene strongly suggests that the megaplasmids could have played an important role in the emergence of C. butyricum type E as a human pathogen.
Citation: Franciosa G, Scalfaro C, Di Bonito P, Vitale M, Aureli P (2011) Identification of Novel Linear Megaplasmids Carrying a ß-Lactamase Gene in Neurotoxigenic Clostridium butyricum Type E Strains. PLoS ONE 6(6): e21706. https://doi.org/10.1371/journal.pone.0021706
Editor: Holger Bruggemann, Max Planck Institute for Infection Biology, Germany
Received: February 11, 2011; Accepted: June 6, 2011; Published: June 28, 2011
Copyright: © 2011 Franciosa et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: Funding was provided by the Istituto Superiore di Sanità, Rome, Italy (www.iss.it). The funder had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Clostridium butyricum is a butyric acid-producing anaerobic and spore-forming bacterium widely distributed in the environment and commonly present in the gut microflora of healthy humans and animals. It is generally considered a harmless saprophyte and it may be used for bioproduction of butyric acid, a widely applied food preservation and flavour-enhancing additive [1]. In addition, in certain countries, C. butyricum strains with proven absence of toxicological effects are allowed for use as probiotics in humans and animals, based on the beneficial effects that butyric acid and the bacteriocin(s) produced by those strains are believed to exert on the host intestine [2], [3].
On the other hand, an in vitro cytotoxic effect of butyric acid produced by C. butyricum has been demonstrated in various cells. This effect has been hypothesized to be responsible for the initiation of the intestinal lesions that lead to the mucosal necrosis of the ileum and colon observed in neonatal necrotizing enterocolitis, although conclusive evidence is still lacking [4], [5].
Most notably, some strains of C. butyricum have been described that are capable of synthesizing the type E botulinum neurotoxin (BoNT/E), which is one of the seven (A to G) antigenically distinct protein toxins that cause the flaccid paralysis of botulism. Neurotoxigenic C. butyricum type E strains were first isolated in Italy in 1984 from two distinct cases of infant botulism [6], an intestinal toxaemia by botulinum toxin-producing clostridia that affects babies under 1 year of age. Except for the ability to produce botulinum toxin, which is conventionally considered a distinctive characteristic of members of the C. botulinum species, these particular strains had all the phenotypic characteristics of the C. butyricum species [7]: their taxonomic identity was later genetically confirmed by DNA hybridization experiments and 16S rRNA gene sequencing [8]–[10]. Since then, neurotoxigenic C. butyricum type E has been implicated in further cases of infant botulism, as well as in several episodes of intestinal toxaemia in adults and of foodborne botulism (namely the classic intoxication caused by the ingestion of preformed BoNT in improperly preserved foods) [11]–[17]. Moreover, C. butyricum type E has been isolated from numerous lake sediments in China [18], where an outbreak of foodborne botulism had previously occurred [11]. Although neurotoxigenic C. butyricum type E has been rarely isolated up to now, it seems to be spread throughout the world, and in Italy it has been implicated in botulism cases more frequently than C. botulinum type E [19].
It is assumed that neurotoxigenic C. butyricum type E strains acquired the bont/E gene from a progenitor strain through mobile genetic elements [20]. Recently, the bont/A,/B and/F genes of certain C. botulinum strains were localized within circular plasmids [20]–[24], whereas the bont/G gene of C. argentinense has long been known to be plasmid-encoded [25], although whether the bont/G-encoding plasmid has a circular or linear structure remains unknown. Remarkably, some of the bont-encoding plasmids have been shown to be conjugative [26].
Previous studies have investigated the bont/E gene chromosomal or plasmid location in the genomes of neurotoxigenic C. butyricum type E. Although earlier experiments yielded discordant results, more recent studies indicated a chromosomal location for the bont/E gene in the neurotoxigenic C. butyricum type E isolates that had been analyzed [20], [27]–[29].
Here, we present evidence for the existence of previously unrecognized extremely large linear plasmids in ten neurotoxigenic C. butyricum type E strains isolated from clinical, food, and environmental sources in two widely separated geographical areas, Italy and China (Table 1). Given these unexpected findings, and because the genes encoding other BoNT types have been demonstrated to be chromosomally or plasmid located, depending on the clostridia strains being examined, the location of the bont/E gene in the genomes of the ten rare neurotoxigenic C. butyricum type E strains was re-evaluated. The genomic location of additional selected genes, including three different antibiotic resistance genes, was also determined in order to investigate the genetic content of the newly identified megaplasmids.
Results
Detection of mega-sized linear extrachromosomal bands in neurotoxigenic Clostridium butyricum type E isolates
Pulsed-field gel electrophoresis (PFGE) analysis of the undigested bacterial genomes has proven to be a valuable tool for demonstrating the existence of multiple replicons in bacteria [30], [31]. Our PFGE experiments revealed that the undigested genomic DNA preparations of each of ten neurotoxigenic C. butyricum type E strains contained a mega-sized DNA band of >600 kb in addition to the main chromosomal band (Figure 1). This result was apparently in contrast with the result previously obtained by Wang et al. [29], who detected by PFGE a single DNA chromosomal band in several neurotoxigenic C. butyricum type E strains, some of which were also analyzed in the present study (Table 1). However, the electrophoretic mobility of the DNA molecules is known to depend upon pulse times [32], and discrepancies between the above results are likely due to the different pulse time values used in the PFGE experiments: Wang et al used pulse times increasing from 3 to 20 s [29], whereas we used longer pulse times in different experiments (increasing from 5 to 60 s, or from 50 to 90 s) that allowed us to resolve the mega-sized extrachromosomal elements.
An additional extrachromosomal smaller band is evident in the DNA preparations from neurotoxigenic type E strains ISS-145/1, ISS-20, ISS-21, ISS-109, ISS-86, ISS-190 (lanes 2–7), and KZ-1890 (lane 9). Lane 1, undigested genomic DNA from non-neurotoxigenic C. butyricum strain ATCC 19398. S.e., XbaI-digested genomic DNA fragments of Salmonella enterica serotype Braenderup strain H9812 (ref. 35). S.c., Saccharomyces cerevisiae chromosomal DNA size marker (Biorad). PFGE conditions: 5–60 s pulse at 6 V/cm for 18 h (1a), and 50–90 s pulse at 6 V/cm for 20 h (1b). At the different PFGE conditions applied, the apparent sizes of the extrachromosomal mega-sized DNA bands (lanes 2–11) did not change in relation to the linear bands of the molecular size standards, consistent with the behaviour expected for linear DNA molecules. The mega-band sizes were estimated as follows (Figure 1b): >610 kb for strains ISS-20, ISS-21, ISS-109, ISS-145/1 (lanes 2–5); ∼825kb for strains ISS-86, ISS-190, LCL-063 and LCL-155 (lanes 6, 7, 10 and 11); and between 750 and 785 kb for strains KZ-1886 and KZ-1890 (lanes 8, 9). Migration of the extrachromosomal smaller DNA bands relative to the linear DNA markers was inconsistent (lanes 2–7 and lane 9 of Figures 1a and 1b), consistent with the behavior expected for circular super-coiled DNA molecules: as a consequence, the actual size of the extrachromosomal smaller bands of strains ISS-145/1, ISS-20, ISS-21, ISS-109, ISS-86, ISS-190 and KZ-1890 could not be determined.
The high intensity of the mega-sized extrachromosomal bands after ethidium bromide staining indicated that they were present in multiple copies or that they might have a linear structure. The latter assumption is based on the different PFGE mobility of linear DNA molecules compared with that of large circular DNA molecules. Specifically, linear DNA molecules are totally released from the wells of PFGE gels and freely migrate through the gel according to their molecular size, whereas circular closed DNA molecules are in part retained within the wells of PFGE gels or they can partially migrate out of the wells in a supercoiled form, the mobility of which depends on changes in the applied pulse time values rather than on the molecular size, resulting in variation in the apparent sizes of the DNA molecules [33]. Our results supported the linear structure of the mega-sized extrachromosomal elements observed in the neurotoxigenic C. butyricum type E strains. First, the apparent sizes of the mega-bands did not change in relation to the linear bands of the molecular size standards when different pulse time conditions were used, consistent with the behavior expected for linear DNA (Figure 1). Second, the migration rate of the mega-sized extrachromosomal bands was not affected by treatment with S1 nuclease, an enzyme used to linearize circular DNA (data not shown). Finally, the mega-sized extrachromosomal bands were degraded by treatment with ATP-dependent exonuclease, an enzyme that selectively hydrolyzes linear DNA without affecting supercoiled and circular DNA (Figure 2), confirming that the extrachromosomal mega-bands had a linear structure.
S.e., XbaI-digested genomic DNA fragments of Salmonella enterica serotype Braenderup strain H9812 (ref. 35). PFGE conditions: 5–60 s pulse at 6 V/cm for 18 h. − and + indicate absence and presence of treatment with ATP-dependent exonuclease, respectively. The megaplasmid bands disappeared after treatment with the ATP-dependent nuclease, an enzyme that selectively digests linear DNA molecules, indicating that all megaplasmids had a linear structure (lanes 2, 4, 6, 8, 10, 12). The smaller plasmids of strains ISS-109 (lanes 1 and 2) and ISS-190 (lanes 3 and 4) were not susceptible to the ATP-dependent nuclease treatment, indicating that they had a circular structure. The ∼167 kb band of strain KZ-1890 disappeared after the ATP-dependent nuclease treatment (lanes 7 and 8), indicating that it was a linear DNA molecule; however, the same plasmid of strain KZ-1890 showed inconsistent PFGE migration typical of supercoiled DNA (lane 9 of Figures 1a and 1b). Hence, the smaller plasmid of strain KZ-1890 can exist in both linear and supercoiled forms, consistent with the behavior expected for circular DNA.
At least four differently sized linear extrachromosomal bands ranging from ∼610 kb to ∼825 kb, as estimated by comparison with a molecular size standard (S. cerevisiae chromosomal DNA, Biorad), were detected in the ten neurotoxigenic C. butyricum type E strains examined in this study (Figure 1b). Specifically, a band greater than the 610-kb band of the molecular standard was observed in four Italian strains (ISS-20, ISS-21, ISS-109, ISS-145/1). A band approximately in line with the 825-kb band of the molecular size standard was observed in four strains, two from Italy (ISS-86 and ISS-190) and two from China (LCL-063 and LCL-155). The two remaining C. butyricum type E strains from China (KZ-1886 and KZ-1890) displayed slightly different bands ranging from 750 to 785 kb.
In addition to the mega-sized bands, the six neurotoxigenic C. butyricum type E isolates from Italy showed a smaller extrachromosomal band of an apparent size of ∼55 kb, whereas one of the strains from China (strain KZ-1890) displayed an additional extrachromosomal band, the apparent size of which was close to the 167.1-kb band of the molecular size standard (XbaI-digested genomic DNA fragments of Salmonella enterica serovar Braenderup strain H9812) [34] (Figure 1a). However, these smaller extrachromosomal bands disappeared from the gel after S1 nuclease treatment, indicating that they were susceptible to the enzyme activity (data not shown). In addition, the smaller plasmids of the Italian strains were not susceptible to the ATP-dependent nuclease treatment (Figure 2), indicating that they had a circular structure. The ∼ 167 kb band of strain KZ-1890 disappeared after the ATP-dependent nuclease treatment (Figure 2), indicating that it was a linear DNA molecule; however, since the same plasmid of strain KZ-1890 was susceptible to the S1 nuclease treatment and it showed an inconsistent PFGE migration (Figures 1a and 1b) typical of a supercoiled DNA molecule, it was concluded that this plasmid could exist in both linear and supercoiled forms, consistent with the behavior expected for circular DNA.
PFGE of the undigested genomic DNA obtained from the non-neurotoxigenic C. butyricum strains and from the C. botulinum type E strains included in this study (Table 1) either showed no extrachromosomal bands or extrachromosomal bands much smaller than the mega-sized ones observed in the neurotoxigenic C. butyricum type E isolates (Figure 3).
Strains: UC-9035 (lane 1); UC-9041 (lane 2); GP1 (lane 3); ATCC 19398 (lane 4); CDC-5234 (lane 5); CDC-4581 (lane 6); CDC-5380 (lane 7). S.e., XbaI-digested genomic DNA fragments of Salmonella enterica serotype Braenderup strain H9812 (ref. 35). PFGE conditions: 5–60 s pulse at 6 V/cm for 18 h. Extrachromosomal bands close to the 54.7 kb band of the molecular standard (S.e.) are evident in lane 3 (non-neurotoxigenic C. butyricum strain GP1) and lane 6 (C. botulinum type E strain CDC-4581).
Analysis of the XhoI and SmaI PFGE macrorestriction profiles of the genomic DNA from the neurotoxigenic C. butyricum type E strains revealed that the four Chinese strains clustered into four different PFGE groups with both endonucleases, and the six Italian strains clustered into two major PFGE groups with an intra-linkage homology level of ≥90% with both XhoI and SmaI endonucleases (Figure 4). Each of the PFGE groups was consistent with a different size of mega-sized extrachromosomal bands observed among the strains.
The similarity of PFGE profiles was evaluated using the Bionumerics software (version 4.0) (Applied Maths, Sint-Martens-Latem, Belgium); a similarity level of ≥ 90% was used to define PFGE groups. The 10 neurotoxigenic C. butyricum type E strains were divided in 6 PFGE groups with both endonucleases.
Bont/E gene carriage and copy number
Southern-blot analysis following PFGE of the undigested DNA of the neurotoxigenic C. butyricum type E strains with a specific bont/E gene probe revealed an hybridization signal corresponding to the chromosomal bands of all strains (Figure 5b).
S.e., XbaI-digested genomic DNA fragments of Salmonella enterica serotype Braenderup strain H9812 (ref. 34). PFGE conditions: 5–60 s pulse at 6 V/cm for 18 h (5a). Southern hybridization with a bont/E gene probe showing that the gene probe hybridized to the chromosome bands of strains (5b). Southern hybridization with a β-lactamase gene probe showing that the gene probe hybridized to the megaplasmids of strains ISS-145/1 (lane 1); ISS-20 (lane 2); ISS-21 (lane 3); ISS-109 (lane 4); ISS-86 (lane 5); ISS-190 (lane 6); LCL-063 (lane 9); LCL-155 (lane 10) (5c).
The XhoI- and SmaI-digested DNA samples of the neurotoxigenic C. butyricum type E strains were also hybridized with the bont/E gene probe to determine the copy number of the bont/E gene. The results showed a single hybridization band in all analyzed neurotoxigenic C. butyricum type E strains, indicating that they all contained a single copy of the bont/E gene (Figure 6).
S.e., XbaI-digested genomic DNA fragments of Salmonella enterica serotype Braenderup strain H9812 (ref. 34). PFGE conditions: 4–40 s pulse at 6 V/cm for 18 h (6a). Southern hybridization with a bont/E gene probe showing that the gene probe hybridized to single restriction bands, as indicated by the black arrows (6b).
Carriage of 16S rRNA and mobA genes
The genomic location of the 16S rRNA genes was investigated to determine whether the mega-sized extrachromosomal bands were plasmids or second chromosomes; by convention, the presence of ribosomal RNA genes is indicative of a chromosome because such essential genes of the translational machinery are generally not carried within plasmids in bacteria [35]. The specific 16S rDNA gene probe that we used only hybridized to the chromosomal bands of the ten neurotoxigenic C. butyricum type E strains (data not shown), indicating that the extrachromosomal bands were indeed mega-plasmids rather than second chromosomes.
BLAST comparison between the draft whole-genome sequences of two neurotoxigenic C. butyricum type E strains (accession numbers NZ_ACOM00000000 for strain BL5262 that corresponds to strain ISS-20 of this study, and NZ_ABDT00000000 for strain 5521 that corresponds to strain ISS-21 of this study) and a ∼8-kb bacteriocin-encoding plasmid from a non-neurotoxigenic C. butyricum strain (pCBM588, accession number NZ_ABDT00000000; the only available complete DNA sequence from a non-neurotoxigenic C. butyricum strain) [36], showed a high percentage of nucleotide identity (93%) in a 2-kb DNA region. This region included a gene for a mobilization protein of the MobA/MobL family that is essential for specific plasmid transfer [37]. A specific mobA gene probe that we designed only hybridized to the smaller extrachromosomal bands observed in the Italian C. butyricum type E strains (Figure 7).
S.e., XbaI-digested genomic DNA fragments of Salmonella enterica serotype Braenderup strain H9812 (ref. 35). PFGE conditions: 5–60 s pulse at 6 V/cm for 18 h. Southern hybridization with a specific mobA gene probe showing that the gene probe hybridized to the smaller plasmids observed in lanes 2–7 (strains ISS-145/1, ISS-20, ISS-21, ISS-109, ISS-86, ISS-190) (7b).
Carriage of antibiotic resistance genes
Several antibiotic resistance genes were present in both draft genome sequences of the Italian neurotoxigenic C. butyricum type E strains. As antibiotic resistance genes are often found within bacterial plasmids, we investigated the location of representative antibiotic resistance genes (i.e., the tetracycline tet(P) gene, the lincomycin resistance protein lmrB gene, and a ß-lactamase gene) in the genomes of the ten neurotoxigenic C. butyricum type E strains. The tet(P) gene probe that we used hybridized to the chromosomes of all ten strains that were analyzed. The lmrB gene probe hybridized to the chromosomes of the six strains isolated in Italy, whereas no hybridization signal was observed for the four strains from China (data not shown), indicating that only the Italian strains contained the lmrB gene. Finally, the ß-lactamase gene probe hybridized to the megaplasmid bands of eight of the ten neurotoxigenic C. butyricum type E strains (Figure 5c), of which six from Italy and two from China (LCL-063 and LCL-155); the two remaining Chinese strains (KZ-1886 and KZ-1890) did not show any hybridization signal with the ß-lactamase gene probe that we used, indicating that they did not harbor that ß-lactamase gene.
Discussion
In this study, we show the presence of a linear megaplasmid in the genomes of ten neurotoxigenic C. butyricum type E strains with different clinical and geographical origins. The linear megaplasmids displayed size heterogeneity, with at least four different sizes being recognized. The differently sized megaplasmids were distributed among the C. butyricum type E strains according to the distinct PFGE groups they belonged to, indicating that the megaplasmids likely contributed to the clonal diversity. Besides, some of the analyzed strains were shown to possess additional smaller circular plasmids. These unprecedented findings indicate that neurotoxigenic C. butyricum type E strains share a complex genomic organization, with multiple linear and circular replicons, thus providing new insights into the genetics of these microorganisms.
The occurrence of plasmids has already been reported in several Clostridium spp, including C. perfringens, C. difficile, C. tetani, C. argentinense and C. botulinum [38]. However, to our knowledge, the largest plasmid described so far in the clostridia genus is a 270-kb circular plasmid (pCLJ, Accession Number NC_012654) that simultaneously harbors two bont genes, the bont/A and/B genes [22]. This plasmid is considerably smaller than the megaplasmids of 600 kb to 850 kb that are described here. DNA molecules of these huge sizes have largely remained undetected due to the considerable technical challenges that their separation from the chromosomes involves. Considering that the sizes of two neurotoxigenic C. butyricum type E genomes have been estimated to be ∼4.5 Mb and 4.7 Mb [20], the megaplasmids would constitute 13–19% of the whole bacterial genome. Such genome sizes are larger than all C. botulinum genomes available to date in the GenBank database, which have reported sizes of 2.3 Mb to 4.2 Mb, in line with the presence of megaplasmids in the genomes of neurotoxigenic C. butyricum type E strains. However, comparison with the genome sizes of non-neurotoxigenic C. butyricum is not possible because genome sequencing data are lacking for that organism.
Furthermore, our results indicate that the megaplasmids identified in the neurotoxigenic C. butyricum type E strains have a linear structure. Bacterial linear plasmids cannot readily be recognized by genome sequencing owing to the inability to generate a closed circular DNA sequence; however, they can be identified by their migratory behaviour under different PFGE conditions and by the use of specific nucleases that selectively cut linear or circular DNA molecules. Although linearization of circular plasmids may occur due to cell death and/or during DNA extraction, the concomitant presence in some of the tested C. butyricum type E strains of additional smaller plasmids that were not linearized under the same experimental conditions provided an internal control for the absence of DNA breakage, indicating that the megaplasmids were native linear DNA molecules and not simple artefacts.
Linear plasmids have been established in both Gram-positive and Gram-negative bacteria, including Streptomyces spp, Mycobacterium spp, Bacillus spp, Borrelia spp, Lactobacillus spp, Rhodococcus spp, Micrococcus spp, Escherichia coli, Salmonella enterica, Klebsiella oxytoca, Yersinia enterocolitica, as well as in some eukaryotic microorganisms [39]. However, no linear plasmids have yet been reported in Clostridium spp.
In general, comparative genomic studies have indicated that bacteria with large composite genomic structures might be more ecologically successful in certain environments [40]. In particular, both linear and circular plasmids are usually assumed to confer a selective advantage to the host, by providing an additional repertoire of genes that significantly extend the metabolic versatility, fitness and stress resistance of the microorganism, even though cryptic linear and circular plasmids also exist. Our attempts to localize within the newly recognized linear megaplasmids some of the genes important for the biology of neurotoxigenic C. butyricum type E, including the bont/E-encoding gene, a plasmid mobilization protein gene, and two antibiotic resistance genes (tet(P) and lmrB genes) were unsuccessful. However, a major finding of this study was that the megaplasmids of eight of the ten analyzed neurotoxigenic C. butyricum type E strains carried a gene coding for a ß-lactamase, an enzyme that hydrolyzes the β-lactam class of antibiotics such as penicillins and cephalosporines. Interestingly, most of the C. butyricum type E strains that contained the megaplasmid-encoded ß-lactamase gene had originally been isolated from the clinical specimens of people affected with intestinal toxaemia botulism (five strains) or foodborne botulism (two strains). Thus, it is likely that the neurotoxigenic C. butyricum type E strains harbouring the megaplasmid-encoded β-lactamase gene might have a selective advantage in a clinical setting. Although antibiotics are not recommended for treating intestinal toxaemia botulism and foodborne botulism, their use may be necessary to control secondary infections [41]; β-lactam antibiotics are among the most commonly prescribed antimicrobial drugs, thus inducing selective pressure for resistance genes. Both neurotoxigenic C. butyricum type E strains whose megaplasmids did not hybridize to the ß-lactamase gene probe were from environmental sources (Chinese soil samples): whether the megaplasmids of those strains carry antibiotic resistance genes other than the β-lactamase gene remains to be determined.
The ß-lactamase probe that hybridized to the megaplasmids of eight C. butyricum type E strains was located at 279338–279745 within contig 1 (NZ_ACOM01000001.1) of the draft genome sequence of strain BL5262 (ISS-20 of the present study). Hence, the megaplasmid of strain ISS-20, whose size was estimated >610 kb by comparison with a molecular marker in the present study, is likely (part of) this contig (757,653 bp). Interestingly, another ß-lactamase-encoding gene, a gene coding for a metallo-ß-lactamase family protein, and a ß-lactamase domain protein gene are also present within contig 1. Besides, this contig contains other important putative genes, including the α-subunit of a DNA polymerase III (Gram-positive type) gene, the nitrogenase operon that is absent from the other sequenced contigs, genes coding for the phosphotransferase (PTS) systems for the import of sugars, genes reminiscent of phage proteins such as a phage membrane protein and a recombinase of the phage integrase family, and the genes encoding the CRISPR-associated Cas proteins involved in the acquired immunity against viruses and plasmids (a system evocative of the eukaryotic RNA interference system), several membrane sensor proteins and ATP-binding transporter (ABC transporter) of various substrates.
It will be interesting to analyze and compare the gene contents of the other C. butyricum type E megaplasmids identified in the present study, specially those that are greater than the one of ISS-20 strain, in order to understand which genes are conserved within the megaplasmids and which ones have more recently been acquired.
In conclusion, the widespread occurrence of linear megaplasmids in the neurotoxigenic C. butyricum type E isolates, and their absence from non-neurotoxigenic C. butyricum and C. botulinum type E isolates, would suggest that they are essential for the survival and growth of the former microorganisms. The fact that the megaplasmids of neurotoxigenic C. butyricum type E strains associated with human botulism carry a β-lactam resistance gene could have contributed to the emergence of neurotoxigenic C. butyricum type E as a human pathogen. Research on the role of the megaplasmids in the spread of antibiotic resistance among clostridia and/or other microorganisms is warranted. We believe that findings reported in the present study can give new impulse for completing the genome sequences of the neurotoxigenic C. butyricum type E strains.
Materials and Methods
Clostridia strains
The clostridia strains used in this study are listed in Table 1. Ten strains were neurotoxigenic C. butyricum type E. Of these, six had been isolated in Italy from as many patients suffering either from intestinal toxaemia botulism (ISS-20, ISS-21, ISS-86, ISS-109, ISS-190) or food-borne botulism (ISS-145/1) [6], [13]–[15]: they were obtained from the ISS culture collection. The other four C. butyricum type E strains were from food (LCL-155) and human (LCL-063) specimens isolated from two different food-borne botulism outbreaks in China, and from soil samples (KZ-1886 and KZ-1890) collected from the areas where the outbreaks had occurred [11], [18]: they were a kind gift from Professor Nakamura, Kanazawa University, Japan.
All strains had genetically been confirmed as C. butyricum [8], [9], [42]; however, they all produced BoNT/E and carriage of the corresponding bont/E gene had previously been shown by PCR [13], [15], [29], [43]. In addition, four non-neurotoxigenic C. butyricum strains and three Clostridium botulinum type E strains were included in the experiments (Table 1).
All clostridia strains were stored at −80°C in microbank cryogenic vials (Prolab Diagnostics, Austin, TX). Clostridia stock cultures were grown for 48 h at 37°C on egg yolk agar (EYA) plates (Oxoid, Milan, Italy) under anaerobiosis (GasPack jars, Oxoid). For growth of broth cultures, TPGY broth (5% Trypticase, 0.5% peptone, 0.4% glucose, 2% yeast exctract, 1% L-cysteine hydrocloride monohydrate) was used.
Pulsed-field gel electrophoresis (PFGE)
Overnight TPGY cultures of the clostridia strains were used for DNA preparation. Cells were embedded in 1.5% low-melting-point agarose (Invitrogen, Carlsbad, CA, USA) plugs and genomic DNA was extracted as described previously [23]. Some DNA plugs were treated with S1 nuclease (MBI Fermentas, Lithuania) as detailed elsewhere [23]. Other DNA plugs were digested with ATP-dependent exonuclease (Epicentre Technologies, Madison, WI) according to the manufacturer's instructions. Restriction of the DNA plugs with XhoI or SmaI (New England BioLabs, Ipswich, MA) was performed under conditions recommended by the manufacturers.
Undigested and enzymatically treated genomic DNA samples were separated in a contour-clamped homogeneous electric field system (CHEF Mapper apparatus, BioRad Laboratories, Hercules, CA) through 0.8% Seakem Gold agarose gel (Cambrex, East Rutherford, NJ) in 0.5 X Tris-borate-EDTA buffer. Electrophoresis was performed at 6 V/cm and 14°C. The pulse time increased from 5 to 60 s (linear ramping factor) over 18 h. The DNA isolated from Salmonella enterica serovar Braenderup strain H9812 and restricted with XbaI (New England BioLabs, Ipswich, MA) was used as the molecular marker [34]. In some experiments, the undigested DNA samples were separated by increasing the pulse time from 50 to 90 s over a 20 h run, to more accurately estimate the sizes of the DNA molecules; when these parameters were applied, Saccharomyces cerevisiae chromosomal DNA (Biorad) was used as molecular marker.
Gels were stained with ethidium bromide and visualized using a GelDoc 2000 apparatus (Bio-Rad). The PFGE restriction profiles were compared using Bionumerics software (version 4.0) (Applied Maths, Sint-Martens-Latem, Belgium). Clustering was performed by applying the Dice coefficient and the unweighted pair-group method using arithmetic averages (UPGMA), with 1% optimization and 1% position tolerance.
Southern-blot analyses
Digoxigenin (DIG)-labeled probes were generated by a PCR DIG Probe Synthesis kit (Roche Diagnostics GmbH, Mannheim, Germany). Primers for the preparation of probes from the bont/E, 16SrRNA, mobilization protein (mobA), tetracycline tet(P), lmrB efflux, and ß-lactamase genes are described in Table 2 and were purchased from Eurofins MWG (Ebersberg, Germany). Southern hybridizations of pulsed-field gels with the gene probes were performed as previously described [23]. DIG detection reagents were obtained from Roche Diagnostics. The CSPD substrate (Roche Diagnostics) was used for detection of hybridized probes according to the manufacturer's instructions.
Acknowledgments
We would like to thank Professor Shinichi Nakamura, Kanazawa University, Kanazawa, Japan, for the generous gift of the neurotoxigenic Clostridium butyricum type E strains isolated in China. We also thank Anna Maria Dionisi and Brunella Appicciafuoco, Istituto Superiore di Sanità, Rome, Italy, for their technical support.
Author Contributions
Conceived and designed the experiments: GF PDB. Performed the experiments: GF CS MV. Analyzed the data: GF CS PDB. Wrote the paper: GF PDB PA.
References
- 1. Zhang C, Yang H, Yang F, Ma Y (2009) Current progress on butyric acid production by fermentation. Curr Microbiol 59: 656–663.
- 2. Seki H, Shiohara M, Matsumura T, Miyagawa N, Tanaka M, et al. (2003) Prevention of antibiotic-associated diarrhea in children by Clostridium butyricum MIYAIRI. Pediatr Int 45: 86–90.
- 3. Van Immerseel F, Ducatelle R, De Vos M, Boon N, Van De Wiele T, et al. (2010) Butyric acid-producing anaerobic bacteria as a novel probiotic treatment approach for inflammatory bowel disease. J Med Microbiol 59 (Pt 2): 141–143.
- 4. Cashore WJ, Peter G, Lauermann M, Stonestreet BS, Oh W (1981) Clostridia colonization and clostridial toxin in neonatal necrotizing enterocolitis. J Pediatr 98: 308–311.
- 5. Obladen M (2009) Necrotizing enterocolitis. 150 years of fruitless search for the cause. Neonatology 96: 203–210.
- 6. Aureli P, Fenicia L, Pasolini B, Gianfranceschi M, McCroskey LM, et al. (1986) Two cases of type E infant botulism caused by neurotoxigenic Clostridium butyricum in Italy. J Infect Dis 154: 207–211.
- 7. McCroskey LM, Hatheway CL, Fenicia L, Pasolini B, Aureli P (1986) Characterization of an organism that produces type E botulinal toxin but which resembles Clostridium butyricum from the feces of an infant with type E botulism. J Clin Microbiol 23: 201–202.
- 8. Suen JC, Hatheway CL, Steigerwalt AG, Brenner DJ (1988) Genetic confirmation of identities of neurotoxigenic Clostridium baratii and Clostridium butyricum implicated as agents of infant botulism. J Clin Microbiol 26: 2191–2192.
- 9. Hutson RA, Thompson DE, Collins MD (1993) Genetic interrelationships of saccharolytic Clostridium botulinum types B, E, and F and related clostridia as revealed by small subunit rRNA gene sequences. FEMS Microbiol Lett 108: 103–110.
- 10. Hill KK, Smith TJ, Helma CH, Ticknor LO, Foley BT, et al. (2007) Genetic diversity among botulinum neurotoxin-producing clostridial strains. J Bacteriol 189: 818–832.
- 11. Meng X, Karasawa T, Zou K, Kuang X, Wang X, et al. (1997) Characterization of a neurotoxigenic Clostridium butyricum strain isolated from the food implicated in an outbreak of food-borne type E botulism. J Clin Microbiol 35: 2160–2162.
- 12. Chaudry R, Dhawan B, Kumar D, Bhatia R, Gandhi JC, et al. (1998) Outbreak of suspected Clostridium butyricum botulism in India. Emerg Infect Dis 4: 506–507.
- 13. Fenicia L, Franciosa G, Pourshaban M, Aureli P (1999) Intestinal toxaemia botulism in two young people, caused by Clostridium butyricum type E. Clin Infect Dis 29: 1381–1387.
- 14. Fenicia L, Da Dalt L, Anniballi F, Franciosa G, Zanconato S, et al. (2002) A case of infant botulism due to Clostridium butyricum type E associated with Clostridium difficile colitis. Eur J Clin Microbiol Infect Dis 21: 736–738.
- 15. Anniballi F, Fenicia L, Franciosa G, Aureli P (2002) Influence of pH and temperature on the growth and toxin production by neurotoxigenic strains of Clostridium butyricum type E. J Food Prot 65: 1267–1270.
- 16.
Dykes JK, Lùquez C, Raphael BH, McCroskey LM, Maslanka SE (2008) Laboratory investigation of the first case of Clostridium butyricum type E infant botulism in the United states. Abstracts of the 45th Interagency Research Coordinating Committee Meeting, Philadelphia, PA, USA.
- 17. Abe Y, Negasawa T, Monma C, Oka A (2008) Infantile botulism caused by Clostridium butyricum type E toxin. Pediatr Neurol 38: 55–57.
- 18. Meng X, Yamakawa K, Zou K, Wang X, Kuang X, et al. (1999) Isolation and characterisation of neurotoxigenic Clostridium butyricum from soil in China. J Med Microbiol 48: 133–137.
- 19.
Fenicia L, Anniballi F, Aureli P (2008) Human botulism in Italy, 1984-2007. Abstracts of the Clostridium botulinum - Epidemiology, diagnosis, genetics, control and prevention Meeting. Helsinki, Finland.
- 20. Hill KK, Xie G, Foley BT, Smith TJ, Munk AC, et al. (2009) Recombination and insertion events involving the botulinum neurotoxin complex genes in Clostridium botulinum types A, B, E and F and Clostridium butyricum type E strains. BMC Biol 7: 66.
- 21. Marshall KM, Bradshaw M, Pellett S, Johnson EA (2007) Plasmid-encoded neurotoxin genes in Clostridium botulinum serotype A subtypes. Biochem Biophys Res Commun 361: 49–54.
- 22. Smith TJ, Hill KK, Foley BT, Detter JC, Munk AC, et al. (2007) Analysis of the neurotoxin complex genes in Clostridium botulinum A1-A4 and B1 strains: BoNT/A3,/Ba4 and/B1 clusters are located within plasmids. PloS One 2: e1271.
- 23. Franciosa G, Maugliani A, Scalfaro C, Aureli P (2009) Evidence that plasmid-borne botulinum neurotoxin type B genes are widespread among Clostridium botulinum serotype B strains. PLoS One 4: e4829.
- 24. Umeda K, Seto Y, Kohda T, Mukamoto M, Kozaki S (2009) Genetic characterization of Clostridium botulinum associated with type B infant botulism in Japan. J Clin Microbiol 47: 2720–2728.
- 25. Zhou Y, Sugiyama H, Nakano H, Johnson EA (1995) The genes for the Clostridium botulinum type G toxin complex are on a plasmid. Infect Immun 63: 2987–2091.
- 26. Marshall KM, Bradshaw M, Johnson EA (2010) Conjugative botulinum neurotoxin-encoding plasmids in Clostridium botulinum. PloS One 5: e11087.
- 27. Hauser D, Gilbert M, Boquet P, Popoff MR (1992) Plasmid localization of a type E botulinal neurotoxin gene homologue in toxigenic Clostridium butyricum strains, and absence of this gene in non-toxigenic C. butyricum strains. FEMS Microbiol Lett 99: 251–256.
- 28. Zhou Y, Sugiyama H, Johnson EA (1993) Transfer of neurotoxigenicity from Clostridium butyricum to a nontoxigenic Clostridium botulinum type E-like strain. Appl Environ Microbiol 59: 3825–3831.
- 29. Wang X, Maegawa T, Karasawa T, Kozaki S, Tsukamoto K, et al. (2000) Genetic analysis of type E botulinum toxin-producing Clostridium butyricum strains. Appl Environ Microbiol 66: 4992–4997.
- 30. Li Y, Canchaya C, Fang F, Raftis E, Ryan KA, et al. (2007) Distribution of megaplasmids in Lactobacillus salivarius and other lactobacilli. J Bacteriol 189: 6128–6139.
- 31. Sayeed S, Li J, McClane B (2010) Characterization of virulence plasmid diversity among Clostridium perfringens type B isolates. Infect Immun 78: 495–504.
- 32. Herschleb J, Ananiev G, Schwartz DC (2007) Pulsed-field gel electrophoresis. Nat Protoc 2: 677–684.
- 33. Levene SD, Zimm BH (1987) Separation of open circular DNA using pulsed field electrophoresis. Proc Natl Acad Sci USA 84: 4054–4057.
- 34. Hunter SB, Vauterin P, Lambert-Fair MA, Van Duyne MS, Kubota K, et al. (2005) Establishment of a universal size standard strain for use with the PulseNet standardized pulsed-field gel electrophoresis protocols: converting the national databases to new size standard. J Clin Microbiol 43: 1045–1050.
- 35. Krawiec S, Riley M (1990) Organization of the bacterial chromosome. Microbiol Rev 54: 502–539.
- 36. Nakanishi S, Tanaka M (2010) Sequence analysis of a bacteriocinogenic plasmid of Clostridium butyricum and expression of the bacteriocin gene in Escherichia coli. Anaerobe 16: 253–257.
- 37. Meyer R (2009) Replication and conjugative mobilization of broad host-range IncQ plasmids. Plasmid 62: 57–70.
- 38. Bruggemann H (2005) Genomics of clostridial pathogens: implication of extrachromosomal elements in pathogenicity. Curr Opin Microbiol 8: 601–605.
- 39.
Meinhardt F, Klassen R (2007) Microbial linear plasmids. Microbiology monographs. Heidelberg: Springer. 252 p.
- 40. Konstantinidis KT, Tiedje JM (2004) Trends between gene content and genome size in prokaryotic species with larger genomes. Proc Natl Acad Sci USA 101: 3160–3165.
- 41. Arnon SS, Schechter R, Inglesby TV, Henderson DA, Bartlett JG, et al. (2001) Botulinum toxin as a biological weapon. JAMA 285: 1059–1070.
- 42. Pourshaban M, Franciosa G, Fenicia L, Aureli P (2002) Taxonomic identity of type E botulinum toxin-producing Clostridium butyricum strains by sequencing of a short 16S rDNA region. FEMS Microbiol Lett 214: 119–125.
- 43. Franciosa G, Ferreira JL, Hatheway CL (1994) Detection of type A, B, and E botulism neurotoxin genes in Clostridium botulinum and other Clostridium species by PCR: evidence of unexpressed type B toxin genes in type A toxigenic organisms. J Clin Microbiol 8: 1911–1917.