Escherichia coli O104:H4 was associated with a severe foodborne disease outbreak originating in Germany in May 2011. More than 4000 illnesses and 50 deaths were reported. The outbreak strain was a typical enteroaggregative E. coli (EAEC) that acquired an antibiotic resistance plasmid and a Shiga-toxin 2 (Stx2)-encoding bacteriophage. Based on whole-genome phylogenies, the O104:H4 strain was most closely related to other EAEC strains; however, Stx2-bacteriophage are mobile, and do not necessarily share an evolutionary history with their bacterial host. In this study, we analyzed Stx2-bacteriophage from the E. coli O104:H4 outbreak isolates and compared them to all available Stx2-bacteriophage sequences. We also compared Stx2 production by an E. coli O104:H4 outbreak-associated isolate (ON-2011) to that of E. coli O157:H7 strains EDL933 and Sakai. Among the E. coli Stx2-phage sequences studied, that from O111:H- strain JB1-95 was most closely related phylogenetically to the Stx2-phage from the O104:H4 outbreak isolates. The phylogeny of most other Stx2-phage was largely concordant with their bacterial host genomes. Finally, O104:H4 strain ON-2011 produced less Stx2 than E. coli O157:H7 strains EDL933 and Sakai in culture; however, when mitomycin C was added, ON-2011 produced significantly more toxin than the E. coli O157:H7 strains. The Stx2-phage from the E. coli O104:H4 outbreak strain and the Stx2-phage from O111:H- strain JB1-95 likely share a common ancestor. Incongruence between the phylogenies of the Stx2-phage and their host genomes suggest the recent Stx2-phage acquisition by E. coli O104:H4. The increase in Stx2-production by ON-2011 following mitomycin C treatment may or may not be related to the high rates of hemolytic uremic syndrome associated with the German outbreak strain. Further studies are required to determine whether the elevated Stx2-production levels are due to bacteriophage or E. coli O104:H4 host related factors.
Citation: Laing CR, Zhang Y, Gilmour MW, Allen V, Johnson R, Thomas JE, et al. (2012) A Comparison of Shiga-Toxin 2 Bacteriophage from Classical Enterohemorrhagic Escherichia coli Serotypes and the German E. coli O104:H4 Outbreak Strain. PLoS ONE 7(5): e37362. https://doi.org/10.1371/journal.pone.0037362
Editor: A. Mark Ibekwe, U. S. Salinity Lab, United States of America
Received: January 10, 2012; Accepted: April 20, 2012; Published: May 23, 2012
Copyright: © 2012 Laing 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: This work was supported by the Public Health Agency of Canada and grants from the Natural Sciences and Engineering Research Council of Canada (www.nserc-crsng.gc.ca) and Alberta Innovates Technology Futures (www.albertatechfutures.ca). The funders 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.
A novel Escherichia coli O104:H4 strain was associated with a widespread and severe foodborne disease outbreak in Germany between early May and July, 2011 . Among the more than 4000 individuals from which the pathogen was isolated, approximately 800 people developed the life-threatening hemolytic uremic syndrome (HUS) and 50 people succumbed to the illness. Epidemiological studies eventually pointed to a single lot of imported Fenugreek seeds used to prepare sprouts for salads as the source of the organism .
The E. coli O104:H4 outbreak strain exhibited characteristics typical of other enteroaggregative E. coli (EAEC), including the presence of a paa-containing virulence plasmid, production of enteroaggregative adherence fimbriae (AAF) I, and formation of the stacked-brick adherence pattern on intestinal epithelial cells . The genome of the outbreak strain was also very similar to the EAEC O104:H4 reference strain 55989, which was isolated in Africa in 2002 . The 2011 outbreak isolates were shown to be clonal and part of multi-locus sequence type ST678, which is unique to E. coli O104:H4 strains, and part of the E. coli phylogenetic group B1, which contains a variety of other pathogenic serotypes . The outbreak strains also acquired an antibiotic resistance plasmid, conferring resistance to all penicillins, cephalosporins and co-trimoxazole , . Unusually, the outbreak-associated E. coli O104:H4 isolates also produced Shiga-toxin 2 (Stx2) and consequently were classified as Shiga-toxin producing E. coli (STEC). Stx production has only rarely been reported among EAEC strains. Prior to the outbreak, only seven sporadic cases of infections with E. coli O104:H4 strains that produced Stx2 had been reported worldwide in the preceding ten years .
Prior to this outbreak, HUS in Germany and elsewhere was most frequently associated with STEC infections in children and the elderly, and only 1.5%–10% of all reported STEC cases in Germany between 2006 and 2010 resulted in HUS in adults . By contrast, in the E. coli O104:H4 outbreak, illness occurred predominantly in otherwise healthy adults and approximately 20% of the cases resulted in HUS, and severe neurological complications were also observed in a large number of cases , . E. coli O104:H4 had previously been associated with only two cases of HUS; one in a woman from Tokyo and one from a child in Germany .
STEC can harbour one or more stx-genes which are encoded by inducible lambda phage integrated into their genomes, and the entire phage and specific regions within the phage can be gained or lost through horizontal gene transfer . The E. coli O104:H4 outbreak isolates contained two integrated Stx-like phages, one of which contained the Stx2 A and B subunit genes and another that did not .
Whole-genome phylogenies place the E. coli O104:H4 outbreak strain as most closely related to other E. coli O104:H4 and EAEC strains, whether they have acquired the Stx2-phage or not , . Because Stx-bacteriophage are mobile, their evolutionary history may differ from that of their bacterial host . This raises the question as to the possible origin of the Stx2-phage found in the E. coli O104:H4 outbreak strain.
In this study, we analyzed the Stx2-phage sequences from the E. coli O104:H4 outbreak strain and compared it among 51 other Stx2-phage sequences from Genbank and those obtained using de novo DNA sequence analysis. We demonstrate that the E. coli O104:H4 Stx2-phage is most closely related to the Stx2-phage from E. coli O111:H- strain JB1-95, and that a recent common ancestor likely gave rise to both it and the phage present in the outbreak strain. Additionally we show that Stx2 production by an E. coli O104:H4 outbreak-related isolate is significantly greater than that of two other outbreak-related E. coli O157:H7 strains following addition of mitomycin-C to cultures.
Materials and Methods
Identification and Isolation of Stx2-bacteriophage Sequences
The Stx2 subunit A nucleotide sequence from E. coli O157:H7 strain EDL933 (NC_002655.2 1352290..1353249) was used to identify Stx2-bacteriophage containing bacteria within Genbank, both the closed and whole-genome shotgun sequence databases. These sequences were downloaded and the Stx2-bacteriophage sequence was identified using PHAST . Only continuous phage sequences of ≥20 kb from a single contig or closed-genome were used in subsequent analyses; Stx2-positive draft genomes that did not meet these criteria were excluded. The Stx1-like phage from Shigella dysenteriae was included as an outgroup for the analyses. The names, sources and accession numbers of all 52 bacteriophage sequences used in this study are presented in Table 1.
Phylogenetic Relationships Among Stx2-bacteriophage
A non-redundant Stx2-bacteriophage “pan-genome” of the 52 phage-genomes of this study was created, and fragments ≥500 bp present in at least three of the Stx2-phage genomes were aligned using the standalone-version of the Panseq program (http://lfz.corefacility.ca/panseq/) with the following settings: ‘minimumNovelRegionSize’ = >‘500’, fragmentationSize’ = >‘500’, ‘nucB’ = >‘200’, ‘nucC’ = >‘50’, ‘nucD’ = >‘0.12’, ‘nucG’ = >‘100’, ‘nucL’ = >‘11’, ‘percentIdentityCutoff’ = >‘85’, ‘coreGenomeThreshold’ = >‘3’ . The aligned fragments were visualized in SplitsTree 4 (http://www.splitstree.org/), using the uncorrected P distance and the neighbor-net algorithm . We have previously shown these fragmentation and sequence identity threshold parameters to be appropriate choices for E. coli . Figure S1 depicts the neighbor-nets of the same data fragmented from 50 bp –1000 bp at both an 85% and 95% sequence identity cutoff, all of which show similar relationships among strain groups.
A maximum-likelihood tree using the aligned Stx2-bacteriophage “pan-genome” described above was created using PhyML 3.0 with the following settings: PhyML_3.0_linux64 -p -m HKY85 -s BEST –rand_start –n_rand_starts 5–r_seed 5 . The resulting tree file with approximate likelihood ratio test (aLRT) branch support values was visualized in Dendroscope .
The Stx2-bacteriophage that were closest to the E. coli O104:H4 outbreak isolates in Figures 1 and 2 were aligned using MAFFT v6.857b with the command: mafft –maxiterate 1000–clustalout –thread 22 . Unlike the non-syntenic Panseq alignment, the MAFFT alignment was based on uninterrupted, whole bacteriophage genomes. The MAFFT-aligned genomes were subjected to maximum likelihood phylogenetic analysis as previously described and visualized in Dendroscope .
The “pan-genome” consists of 500 bp sequence fragments present in at least 3 bacteriophage sequences at an 85% sequence identity threshold. Labels represent the host-source of the bacteriophage as serotype followed by strain name. Red shading: Stx2-phage from O104:H4 outbreak isolates (light); O111:H- strain JB1-95 (dark). Purple shading: Stx2-phage from O157:H7 lineage I strains; related Stx2-bacteriophge; O145:H2 strain 4.0967. Green shading: Stx2-phage from O157:H7 lineage I/II strains; the sub-group of lineage I/II Stx2-phage most closely related to the O104:H4 isolate phage. Yellow shading: Stx2c-phage from O157:H7 lineage I/II strains and bacteriophage 2851 (light); O157:H7 lineage II strains (dark).
Labels represent the host-source of the bacteriophage as serotype followed by strain name. Red shading: Stx2-phage from O104:H4 outbreak isolates (light); O111:H- strain JB1-95 (dark). Purple shading: Stx2-phage from O157:H7 lineage I strains; related Stx2-bacteriophge; O145:H2 strain 4.0967. Green shading: Stx2-phage from O157:H7 lineage I/II strains; the sub-group of lineage I/II Stx2-phage most closely related to the O104:H4 isolate phage. Yellow shading: Stx2c-phage from O157:H7 lineage I/II strains and bacteriophage 2851 (light); O157:H7 lineage II strains (dark).
A non-redundant “pan-genome” of 42 STEC and Shigella dysenteriae whole-genomes was created, and fragments ≥500 bp present in at least three of the genomes were aligned using the standalone-version of the Panseq program (http://lfz.corefacility.ca/panseq/) with the following settings: ‘minimumNovelRegionSize’ = >‘500’, fragmentationSize’ = >‘500’, ‘nucB’ = >‘200’, ‘nucC’ = >‘50’, ‘nucD’ = >‘0.12’, ‘nucG’ = >‘100’, ‘nucL’ = >‘20’, ‘percentIdentityCutoff’ = >‘85’, ‘coreGenomeThreshold’ = >‘3’ . A maximum-likelihood tree with aLRT branch support values was created using PhyML as described above .
Overnight cultures of each strain from single colonies were inoculated in 1 ml of BHI medium and incubated at 37°C with shaking at 150 rpm in 2 ml microfuge tubes for 17 hours. The overnight culture was diluted 1∶100 in fresh BHI medium and incubated at 37°C with shaking at 150 rpm for 3 hours after which optical density at 600 nm was measured. Cultures were diluted in two serial dilutions by factors of two to reach final optical densities ∼0.6. Antibiotic concentrations of 0.015625 µg/ml ciprofloxacin and 0.125 µg/ml norfloxacin were used for the induction of EDL 933; 0.0625 µg/ml ciprofloxacin and 0.125 µg/ml norfloxacin were used for the induction of Sakai; 0.0625 µg/ml ciprofloxacin and 0.25 µg/ml norfloxacin were used for the induction of ON-2011. A concentration of 0.5 µg/ml mitomycin was used for all three strains. Cultures were incubated at 37°C with shaking at 150 rpm for 18 hours, which Vareille et al. (2007) found to be critical for maximum Stx2 expression . Optical density at 600 nm was measured with ten-fold dilutions of culture.
Quantification of Stx2 Production
Three Stx2-producing strains were used for the quantification of Stx2 production; two were E. coli O157:H7 strains associated with large outbreaks (strains Sakai  and EDL933 ) and were obtained from the American Type Culture Collection, and one a travel-associated isolate from the German 2011 E. coli O104:H4 outbreak (strain ON-2011 ). The amount of Stx2 produced by each strain was quantified as previously described using an enzyme-linked immunosorbent assay , with the following modification: cells were lysed by incubation with 0.5 mg/ml polymyxin B (Sigma) at 37°C for 60 min, rather than 1.5 mg/ml for 5 min. The values of three independent experiments, each with two replicates, were used to determine the average toxin production for each strain.
One ml of overnight culture was transferred to RNase-free 1.5 ml microfuge tubes and spun for 5 minutes at 13,000 rpm in a Legend Micro 21 R (Sorvall). Following centrifugation, 990 µl supernatant was removed, filtered using 0.20 um filters (Corning), and stored at −20°C for use in ELISA. The bacterial pellet was dissolved in 1.0 ml of RNAprotect Bacteria Reagent (Qiagen) and incubated for 5 min at room temperature (RT) and subsequently centrifuged for 10 minutes at 5000 g to remove supernatant. The pellet was re-suspended in 100 µl TE buffer (10 mM Tris:1 mM EDTA pH 8.0) with 1 mg/ml lysozyme, and incubated for 5 minutes at RT, with vortexing in 2 minute intervals. The RNeasy Mini Kit (Qiagen) was used to complete the RNA extraction and eluted RNA was stored at −80°C.
Reverse transcription of isolated RNA was performed using 1 µg RNA. The genomic DNA elimination reaction consisting of 18 µl template RNA in DNase, RNase-free water (Qiagen) and 3 µl of 7× gDNA Wipeout Buffer from the QuantiTect Reverse Transcription Kit (Qiagen) was incubated for 2 minutes at 42°C in a Mastercycler pro (eppendorf). The reverse transcription reaction was conducted in a reaction mixture containing the 21 µl gDNA elimination reaction, 6 µl of 5× Quantiscript RT Buffer (Qiagen), 1.5 µl RT Primer Mix (Qiagen), and 1.5 µl Quantiscript Reverse Transcriptase. The mixture was incubated in a Mastercycler pro (Eppendorf) using a 2-stage program: 15 minutes of incubation at 42°C followed by 3 minutes at 95°C. The cDNA was stored at −20°C until needed.
Detection of gnd and stx2 Gene Expression by RT-PCR
RT-PCR reactions were performed using a Rotor Gene Q (Qiagen) in reaction volumes of 25 µl consisting of: 9.25 µl DNase, RNase-free water (Sigma), 12.5 µl PerfeCTa FastMix II (Quanta Biosciences), 1 µl of standard or cDNA, 7.5 pmol of each primer and probe. The probes were conjugated at the 5′ end with fluorescent reporter dyes 6-carboxyfluorescein (FAM) and proprietary fluorophore VIC for stx2 and gnd probes respectively. Both probes were also conjugated at the 3′end with the Black Hole Quencher BHQ-1 (Alpha DNA). The gnd housekeeping gene of E. coli was used to normalize values between RNA samples. Standard curves were generated using three concentrations (0.1 ng/µl, 0.01 ng/µl, and 0.001 ng/µl) of plasmids pStx2R-1 and pAY01 containing stx2 and gnd genes respectively. The RT-PCR cycling conditions used were: a hold for 2 minutes at 55°C preceding a hold for 2 minutes at 95°C and 45 cycles of 15 seconds at 95°C followed by 45 seconds at 60°C. Gain optimization was set to the first tube for both Green (FAM) and Yellow (VIC) Channels and acquisition occurred during the 60°C step in cycling. The values of three independent replicates were used to determine average mRNA copy numbers.
The neighbor-net in Figure 1 depicts the Stx-phage distribution among STEC strains and Shigella dysenteriae. The phage in S. dysenteriae was most distantly related to those in the rest of the strains. The E. coli O157:H7 Stx2-bacteriophage were clustered based on both genetic lineage and Stx2/Stx2c-toxin sub-type. All Stx2c-bacteriophage from E. coli O157:H7 lineage I/II strains were from hosts that also contained a second integrated Stx2-bacteriophage. The E. coli O104:H4 outbreak Stx2-phage were situated as a separate cluster, nearest to the Stx2-phage from E. coli O111:H- strain JB1-95.
The maximum-likelihood tree of the same data used to construct Figure 1 depicts the relationship among the Stx2-phage in tree form (Figure 2). Most Stx2-phage grouped together with other Stx2-phage isolated from bacteria of the same serotype, with the exception of E. coli serotype O111 strains. Again, the O157:H7 Stx2-bacteriophage were clustered based on both genetic lineage and Stx2/Stx2c-toxin sub-type. The E. coli O104:H4 outbreak isolate Stx2-phage are most closely situated on the tree to the Stx2-phage from E. coli O111:H- strain JB1-95, and more distantly to Stx2-phage from other O111, O103 and O91 serotypes.
As Stx2-phage are mobile, we compared the relationships among the host bacterial genomes to determine the concordance of the Stx2-phage with that of their host-genome (Figure 3). The groupings of Stx2-phage and host Stx2-genomes were surprisingly similar, with the notable exception of Stx2-phage and bacterial genomes from serogroup O111, which formed a discrete cluster in the whole-genome tree, but were more widely distributed in the Stx2-phage tree. All serotypes in the whole-genome tree of Figure 3 formed discrete clusters. The tree was divided into two separate branches, with O157:H7 strains comprising one branch, and all other STEC strains the other. Within the large O157:H7 branch of Figure 3, three sub-groups that corresponded to genetic lineages I, I/II and II were observed. O145:H2 strain 4.0967 and O103:H2 strain 12009 formed a distinct group in contrast to the Stx2-bacteriophage tree (Figure 2), where the O145:H2 strain 4.0967 Stx2-phage was part of the O157:H7 lineage I Stx2 group and O103:H2 strain 12009 Stx2-phage was most closely grouped to O111:H- strain 11128.
Red shading: O104:H4 outbreak isolates (light); O111:H- strain JB1-95 (dark). Purple shading: O157:H7 lineage I strains; O145:H2 strain 4.0967. Green shading: O157:H7 lineage I/II strains; the sub-group of lineage I/II strains with Stx2-phage most closely related to the O104:H4 isolate phage. Yellow shading: O157:H7 lineage II strains.
We wished to more thoroughly examine the relationships of the entire, intact Stx2-phage sequences from those that most closely resembled the Stx2-phage from the E. coli O104:H4 outbreak isolates. Figure 4 shows the Mauve aligned complete Stx2-phage sequences of a selection of Stx2-bacteriophage that were closely related in Figures 1 and 2. There were common conserved regions among the strains, and Stx2-phage such as O103:H2 strain 12009 and O157:H7 71074 had what appeared to be a very similar phage structure to the O104:H4 outbreak phage, despite not being as evolutionarily conserved at the nucleotide level, as demonstrated by the phylogenetic analyses of this study. Conversely, the O111:H- strain JB1-95 Stx2-phage was grouped most-closely to the O104:H4 Stx2-phage cluster in Figures 1, 2 and 5, but showed the absence of a phage region beginning at the 45 kb region of its phage that was conserved in all the Stx2-phage except O111:H- strain 11128. The O111:H- strain 11128 Stx2-phage was the least similar to the O104:H4 outbreak Stx-2 phage among those examined in Figure 4.
Bounded boxes indicate similar sequence composition among sequences. The black box indicates the location of the stx2ab gene cluster in each sequence.
Figure 5 shows the maximum likelihood phylogram of the sequences depicted in Figure 4 after MAFFT alignment; the alignment file is available as Dataset S1. This ML tree showed the E. coli O111:H- strain JB1-95 Stx2-bacteriophage to be most closely situated to the Stx2-bacteriophage from the German E. coli O104:H4 outbreak isolates, and the O111:H- strain 11128 to be most distantly related among the strains in Figure 5. The Stx2-phage from O111:H- strain JB1-95 was located on the tree between the O104:H4 outbreak cluster and a cluster of Stx2-phage from E. coli O157:H7 lineage I/II strain 71074, O103:H2 strain 12009, and O111:NM strain OK1180.
Lastly, we compared the Stx2-production by E. coli O104:H4 outbreak-related strain ON-2011 with that of two other outbreak–associated strains of E. coli O157:H7, EDL933 and Sakai. As shown in Figure 6, Stx2-production of ON-2011 was significantly lower than that of the two O157:H7 strains (Ρ <0.01). However, after the addition of mitomycin C, Stx2-production by ON-2011 was significantly greater than for both of the E. coli O157:H7 strains (P<0.01).
Error bars represent standard deviations from three independent replicates.
Stx2-mRNA levels were also compared between O104:H4 strain ON-2011 and O157:H7 strain EDL933 after induction by ciprofloxacin, norfloxacin and mitomycin C (Figure 7). All three treatments increased Stx2-mRNA production in the strains, with the mitomycin C treatment causing the largest increase in Stx2-mRNA production by both strains. The amount of Stx2-mRNA produced by ON-2011 after mitomycin C treatment was significantly greater than that produced by EDL933 (P<0.05).
The E. coli O104:H4 outbreak strain is thought to be derived from an E. coli O104:H4 progenitor that recently acquired the Stx2-phage . Seven isolates of this outbreak strain were used in this study, but a detailed description of differences among these and other isolates can be found in the study by Grad et al. (2012) . Stx-phage are temperate lambdoid phage, which exist as an ordered set of interchangeable modules that have a high propensity to recombine into novel mosaic configurations in the laboratory . As Stx2-phage can be readily horizontally transferred, the evolutionary history of the phage and its bacterial host would not be expected to be highly concordant. We wished to identify the Stx2-phage most closely related to the E. coli O104:H4 outbreak Stx2-phage, by considering the bacteriophage sequence independently of that of the host bacteria.
Given the potential for high levels of recombination among Stx2-bacteriophage, we initially examined the phylogenetic relationship of Stx2-phage from 51 E. coli strains and Shigella dysenteriae using a network, where competing phylogenetic signals would be shown. We also chose to limit the Stx2-phage “pan-genome” to loci present in at least three strains, as loci present in two or fewer are typically not informative phylogenetically .
Brzuszkiewicz et al.  had previously found that the VT2phi_272 O157:H7 bacteriophage from E. coli strain 71074 was closely related to the Stx2-phage from the E. coli O104:H4 outbreak strain. This was particularly interesting to us, as strain 71074 is known to be from a clade of E. coli O157:H7 lineage I/II strains that has been termed “hyper-virulent” due to its more frequent association with severe human infections compared to other E. coli O157:H7 genotypes , . However, it was unclear how many Stx2-phage sequences had been examined in the Brzuszkiewicz et al.  study, so we repeated the analysis using 51 Stx2-phage sequences and the outgroup Shigella dysenteriae phage sequence.
Both the neighbour-net and ML-tree examination of the pan-genomic data grouped the Stx2-phage from the O104:H4 outbreak isolates as a cluster, with the common closest phage being from O111:H- strain JB1-95. While the phage from O157:H7 lineage I/II strains bear a structural similarity to the phage from the O104:H4 outbreak strain, our study suggests the possibility of a lateral gene transfer event from an STEC host more closely related to JB1-95 than any of the O157:H7 strains.
Evolutionary Concordance of Stx2-bacteriophage and Host-genomes
Despite the expectation that Stx2-phage would be highly heterogeneous, most were highly concordant with the phylogenetic trees of their bacterial hosts, suggesting that the host and phage had stably co-evolved for significant periods of time. This suggests that the lysogen offers an advantage to its host in the natural environment of the bacteria that is greater than the evolutionary cost of the bacterium continually copying the DNA of the foreign bacteriophage genome. The natural reservoir of most Stx2-producing bacteria is the ruminant intestine ; therefore, the evolutionary forces acting to select for particular traits are those that confer an advantage in the ruminant host, or the “source” habitat . Humans are largely incidental hosts, with infection and clearance lasting only a few weeks and represent a “sink” habitat. In this source/sink evolutionary model, the source habitat selects for long-term evolutionary change and the sink habitat allows proliferation over a limited time-scale, with eventual clearance.
In the case of Stx, it may aid in the colonization of the gut  and has also been suggested to defend the population against grazing protozoa that inhabit the bovine intestinal tract . This resistance to predation may also facilitate increased environmental persistence and re-uptake by ruminants, allowing a single-clone to dominate in a herd, as has been shown in E. coli O157:H7 , . The increased environmental persistence if associated with toxin production would increase the risk of human infection by increasing the probability of exposure. Once infected, the amount of toxin produced may also help the pathogen survive the human immune response, as Stx production has been shown to allow proliferation of E. coli O157:H7 within human macrophages . Additionally, immunity to predatory bacteriophages in the rumen gut could also select for the persistence of lysogenic Stx-bacteriophage in E. coli hosts.
We observed in the whole-genome tree of Figure 3 that serotypes invariably clustered together, suggesting that the discordance between the Stx2-phage and E. coli host whole genome trees was caused by recent acquisition (horizontal transfer) of the Stx2-phage. The evolution of pathogenicity in E. coli is thought to be driven by events such as phage-mediated horizontal gene transfer and our results suggest that once a phage has been acquired, it can be stably propagated along with the bacterial lineage . This is most evident in the case of the O157:H7 strains, where the whole-genomic sequences in Figure 3 are clearly distributed among the three lineages, and the Stx2-phage also group according to lineage (Figures 1 and 2), even in the case where a single strain has two integrated Stx2-phage (Stx2 and Stx2c). The grouping of non-O157:H7 Stx2-phage with O157:H7 Stx2-phage on the same branch of the tree, as was the case for the O145:H2 strain and O157:H7 lineage I Stx2-phage, likely indicates horizontal transfer of the phage between organisms. In the case of the O111 Stx2-bacteriophage, their wider distribution among STEC strains also suggests at least one horizontal transfer event.
Stx2-bacteriophage Sequence Alignment
E. coli O111:H- strain JB1-95 was found to be more closely related to the E. coli O104:H4 outbreak isolates than the previously reported Stx2-phage from O157:H7 lineage I/II strain 71074 (VT2phi_272) and among the other Stx2-phage sequences examined.
A plausible Stx2-phage evolutionary scenario involves an STEC O111:H- like common ancestor to O111:H- JB1-95 and the E. coli O104:H4 outbreak isolates undergoing a lateral transfer event to the O104:H4 outbreak progenitor strain. This would account for the similar evolutionary histories of the Stx2-phage, as well as account for the region missing from the JB1-95 Stx2-phage sequence but present in all other closely related strains, which seems to indicate a loss of genetic information in the O111:H- branch. However, it should be noted that some Stx2-phage sequences were incompletely recovered from the draft whole-genome sequence as a result of the Stx2-phage sequence being distributed among multiple contigs or the absence of Stx2-phage sequence from the draft-genome. Assembled genome sequences for all of these strains would be helpful in more precisely defining these relationships. The proposed evolutionary history of the E. coli O104:H4 bacterial ancestor giving rise to the 2011 E. coli O104:H4 Stx2-containing strain and EAEC strain 55989 has recently been presented by Mellmann et al. .
Stx2-phage are known to have an integration site preference in host bacteria . Because this integration is dependent on the host genome, the site occupied by an integrated Stx2-phage can be used to discriminate among members of the same species . The fact that the Stx2-bacteriophage in the O104:H4 outbreak isolates, the E. coli O157:H7 lineage I/II strains and the O111:H- strains are all inserted in the argW locus offers additional evidence that these bacteriophage are closely related. Other E. coli strains within serotypes O111:H- and O157:H7 possess Stx2-phage that are significantly different from those of the E. coli O104:H4 outbreak isolates, so we conclude that the Stx2-phage has been introduced multiple times within STEC through horizontal transfer.
The parallel evolution of this and other STEC virulence factors has been previously described . For example, evidence suggests the locus of enterocyte effacement (LEE) has been obtained through horizontal transfer multiple times in E. coli, giving rise to two distinct groupings of Stx-negative enteropathogenic E. coli (EPEC) and Stx-positive enterohemorrhagic E. coli (EHEC) strains that share a LEE insertion site . EPEC typically cause diarrhea in humans, while EHEC cause hemorrhagic colitis and HUS. These findings and the results of this study support the idea that multiple genomic backgrounds in E. coli are able to persist in the human gastrointestinal tract, and in some cases cause illness through a variety of distinct mechanisms; however, it is the Shiga toxin, and in particular Stx2 that is responsible for severe forms of human disease such as HUS. It is now clear that Stx2-phage are capable of integrating into diverse E. coli genomic backgrounds and of qualitatively transforming health risks associated with E. coli strains from other pathogroups such as EPEC and EAEC.
E. coli O157:H7 clade 8 lineage I/II strains are more frequently associated with HUS than other E. coli O157 genotypes and the incidence of human infection with clade 8 strains is thought to be increasing . It is interesting to note that the Stx2-phage from these “hyper-virulent” E. coli O157 strains bear a striking structural similarity to the Stx2-phage from the E. coli O104:H4 outbreak isolates. The E. coli O104:H4 outbreak isolates are also associated with high rates of HUS, suggesting that these phage share attributes that contribute to increased virulence through mechanisms such as increased expression of Stx2. We have previously shown that lineage I strains of E. coli O157:H7 produce significantly different amounts of Stx2 in vitro than lineage II strains, and that within lineage I, strains from humans produce significantly more Stx2 than strains from cattle . In the current study, we found that an E. coli O104:H4 outbreak-related isolate (ON-2011) produced significantly more Stx2 and Stx2-mRNA than outbreak-related lineage I E. coli O157:H7 strains after mitomycin C induction. However, without mitomycin C added to the culture, O104:H4 strain ON-2011 produced very little toxin. Stx2 is an extra-cellular toxin and has been shown to be released from STEC via two specific mechanisms . Therefore, it is unlikely that the high level of Stx2 produced by the mitomycin C-treated E. coli O104 strain was the result of differential toxin release from cell membranes by polymyxin B; however, this possibility cannot be excluded without further study.
A recent study examining the effect of antimicrobials on isolates from the German O104:H4 outbreak found that ciprofloxacin significantly increased both Stx2 and Stx2-mRNA production , which corroborates our current findings and suggests that the Stx2-phage from the E. coli O104:H4 outbreak strains is induced and undergoes lytic conversion in vivo in a significant proportion of the bacterial population in the presence of both ciprofloxacin and mitomycin C.
While it can be argued that conditions in the gastrointestinal tract could induce a similar response to that obtained with mitomycin C, further study is required to determine if Stx2 production by E. coli O104:H4 is greater or lesser than for E. coli O157 strains in vivo, particularly since other work has shown down-regulation of toxin-production by E. coli O157:H7 strains due to microbiota secreted factors present in the human gut .
It is also unknown whether the high-prevalence of serious human disease caused by the O104:H4 outbreak isolates and other Stx2-producing bacteria is due simply to human exposure to high-numbers of the organism, or whether the production of high-levels of Stx2 are responsible for the high-levels of serious human disease. Credence is given to the first hypothesis by the “supershedder” phenomenon in cattle, where it has been reported that phage type (PT) 21/28 of E. coli O157:H7 is shed in significantly higher numbers by cattle than other PTs, and PT 21/28 is also the most commonly associated with serious human disease, suggesting that higher levels of human exposure to this PT account for its increased association with human disease . Conversely, it may be that certain strains are more virulent in humans, simply by virtue of the amount of toxin or other virulence factors they produce. In theory, it would only take one such organism to establish and proliferate within the human intestine to cause severe human illness. Factors allowing the bacteria to withstand the hardships of the human gut and immune system, such as acid resistance and toxin-production could significantly influence disease outcomes. We have shown that in cultures containing mitomycin C, an O104:H4 strain produced significantly higher levels of Stx2 than E. coli O157:H7 strains. It was also recently shown that O104:H4 outbreak isolates also had a higher number of cells survive pH 2.5 conditions than E. coli O157:H7 strain EDL933 .
It has also been suggested that the increased virulence of the E. coli O104:H4 outbreak strain could in part be related to the EAEC bacterial host genome, which has evolved adaptations for attachment and survival in the human intestine and that these adaptations have facilitated the systemic absorption of Stx2, which in turn increased the risk of developing HUS , . If true, it is likely that the acquisition of any one of the many Stx2-phage would have led to an increase in virulence of the E. coli O104:H4 host. However, the relative importance of Stx2-phage heterogeneity and specific Stx2-phage features compared to those of the bacterial host in the virulence of the E. coli O104:H4 outbreak strain awaits further study.
It is clear from the German E. coli O104:H4 outbreak that future novel combinations of bacteriophage and bacterial host are likely and that their impact on human health could be devastating. According to the source-sink model, attributes such as Stx production by STEC from ruminants is thought to provide a selective advantage in the animal reservoir, not in the “dead end” human host. STEC virulence in humans is therefore not selected for but is an unintended consequence of accidental infection. As EAEC strains are thought to be human-restricted, the acquisition of a Stx2-phage from an O111:H- like strain had to occur in an environment where both EAEC and STEC or their phage were concurrently present. This phenotypic jump could have been in the intestine of an animal or human harbouring both E. coli pathotypes, or an environment where both human and ruminant feces were present. In the case of the E. coli O104:H4 STEC strain, we do not know if the acquisition of this Stx2-phage has provided a selective advantage for the bacterium in humans. While the infection was clearly foodborne, it would appear that human to human transmission has not been sustained and no new cases have been reported. The findings of this study suggest that efforts to control all E. coli pathogroups associated with enteric disease in both developed and developing countries as well as Stx2-phage from both human and animal sources are warranted to prevent the re-occurrence of similar devastating outbreaks.
The neighbour-net visualizations of the Stx2-phage “pan-genome” using fragmentation sizes from 50 bp to 1000 bp at both an 85% and a 95% sequence identity threshold. The number to isolate name correspondence is as follows: 1 = O147 UMNF18, 2 = O104:H4 ON211, 3 = O104:H4 CS70, 4 = O157:H7 EC4076, 5 = O157:H7 EC970520, 6 = O103:H2 12009, 7 = O157:H7 EC4206, 8 = Bacteriophage stx2_I, 9 = O157:H7 71074, 10 = O104:H4 C227-11, 11 = O104:H4 GOS2, 12 = O157:H7 TW14588_2, 13 = O121:H19 5.0959, 14 = O139 S1191 stx2e, 15 = O157:H7 EC869, 16 = O157:H7 EC4196, 17 = O157:H7 EC4115, 18 = O104:H4 GOS1, 19 = O111:H- 11128, 20 = O157:H7 EC1212, 21 = Shigella dysenteriae Sd197 stx1, 22 = O157:H7 EC4115, 23 = O157:H7 EC4486, 24 = O153 3.3884, 25 = O157:H7 LRH6, 26 = O104:H4 H112180540 CS110, 27 = O157:H7 EC4206, 28 = O157:H7 1044, 29 = O157:H7 EC4113, 30 = O147:H- 2.3916 stx2e, 31 = Bacteriophage 2851, 32 = Bacteriophage stx2_II, 33 = O157:H7 Sakai, 34 = O111:H- JB1-95, 35 = O157:H7 EC4024, 36 = O157:H7 EDL933, 37 = O157:H7 TW14588_1, 38 = Ont:H12 EH250 stx2d, 39 = O157:H7 TW14359, 40 = O104:H4 TY-2482, 41 = O111:NM OK1180, 42 = Bacteriophage 933 W, 43 = O157:H7 EC4042, 44 = O157:H7 TW14359, 45 = O145:H2 4.0967, 46 = O157:H7 EC508, 47 = O91:H21 B2F1, 48 = O73:H16 C165-02 stx2d, 49 = O157:H7 1125 ECF, 50 = O157:H7 EC4042, 51 = O157:H7 EC4401, 52 = O128:H2 9.0111.
We would like to thank the Canadian Food Inspection Agency for allowing this research to be conducted at the Animal Diseases Research Institute.
Conceived and designed the experiments: CRL JET VPJG. Performed the experiments: CRL YZ. Analyzed the data: CRL YZ VPJG. Contributed reagents/materials/analysis tools: MWG VA RJ. Wrote the paper: CRL JET VPJG.
- 1. Blaser MJ (2011) Deconstructing a lethal foodborne epidemic. N Engl J Med 365: 1835–1836.
- 2. Buchholz U, Bernard H, Werber D, Böhmer MM, Remschmidt C, et al. (2011) German outbreak of Escherichia coli O104:H4 associated with sprouts. N Engl J Med 365: 1763–1770.
- 3. Bielaszewska M, Mellmann A, Zhang W, Köck R, Fruth A, et al. (2011) Characterisation of the Escherichia coli strain associated with an outbreak of haemolytic uraemic syndrome in Germany, 2011: a microbiological study. Lancet Infect Dis 1(9): 671–676.
- 4. Mossoro C, Glaziou P, Yassibanda S, Lan NTP, Bekondi C, et al. (2002) Chronic diarrhea, hemorrhagic colitis, and hemolytic-uremic syndrome associated with HEp-2 adherent Escherichia coli in adults infected with human immunodeficiency virus in Bangui, Central African Republic. J Clin Microbiol 40: 3086–3088.
- 5. Brzuszkiewicz E, Thürmer A, Schuldes J, Leimbach A, Liesegang H, et al. (2011) Genome sequence analyses of two isolates from the recent Escherichia coli outbreak in Germany reveal the emergence of a new pathotype: Entero-Aggregative-Haemorrhagic Escherichia coli (EAHEC). Arch Microbiol 193(12): 883–891.
- 6. Denamur E (2011) The 2011 Shiga toxin-producing Escherichia coli O104:H4 German outbreak: a lesson in genomic plasticity. Clin Microbiol Infect 17: 1124–1125.
- 7. Scheutz F, Moller Nielsen E, Frimodt-Moller J, Boisen N, Morabito S, et al. (2011) Characteristics of the enteroaggregative Shiga toxin/verotoxin-producing Escherichia coli O104:H4 strain causing the outbreak of haemolytic uraemic syndrome in Germany, May to June 2011. Euro Surveill 16(24): 19889.
- 8. Askar M, Faber MS, Frank C, Bernard H, Gilsdorf A, et al. (2011) Update on the ongoing outbreak of haemolytic uraemic syndrome due to Shiga toxin-producing Escherichia coli (STEC) serotype O104, Germany, May 2011. Euro Surveill 6(22): 19883.
- 9. Frank C, Werber D, Cramer JP, Askar M, Faber M, et al. (2011) Epidemic profile of Shiga-toxin-producing Escherichia coli O104:H4 outbreak in Germany. N Engl J Med 365: 1771–1780.
- 10. Shaikh N, Tarr PI (2003) Escherichia coli O157:H7 Shiga toxin-encoding bacteriophages: integrations, excisions, truncations, and evolutionary implications. J Bacteriol 185: 3596–3605.
- 11. Mellmann A, Harmsen D, Cummings CA, Zentz EB, Leopold SR, et al. (2011) Prospective Genomic Characterization of the German Enterohemorrhagic Escherichia coli O104:H4 Outbreak by Rapid Next Generation Sequencing Technology. PLoS ONE 6: e22751.
- 12. García-Aljaro C, Muniesa M, Jofre J, Blanch AR (2009) Genotypic and phenotypic diversity among induced, stx2-carrying bacteriophages from environmental Escherichia coli strains. Appl Environ Microbiol 75: 329–336.
- 13. Zhou Y, Liang Y, Lynch KH, Dennis JJ, Wishart DS (2011) PHAST: A Fast Phage Search Tool. Nucleic Acids Res 39: W347–352.
- 14. Laing C, Buchanan C, Taboada EN, Zhang Y, Kropinski A, et al. (2010) Pan-genome sequence analysis using Panseq: an online tool for the rapid analysis of core and accessory genomic regions. BMC Bioinformatics 11: 461.
- 15. Huson DH, Bryant D (2006) Application of phylogenetic networks in evolutionary studies. Mol Biol Evol 23: 254–267.
- 16. Guindon S, Dufayard J-F, Lefort V, Anisimova M, Hordijk W, et al. (2010) New algorithms and methods to estimate maximum-likelihood phylogenies: assessing the performance of PhyML 3.0. Syst Biol 59: 307–321.
- 17. Huson DH, Richter DC, Rausch C, Dezulian T, Franz M, et al. (2007) Dendroscope: An interactive viewer for large phylogenetic trees. BMC Bioinformatics 8: 460.
- 18. Katoh K, Kuma K, Toh H, Miyata T (2005) MAFFT version 5: improvement in accuracy of multiple sequence alignment. Nucleic Acids Res 33: 511–518.
- 19. Hayashi T, Makino K, Ohnishi M, Kurokawa K, Ishii K, et al. (2001) Complete Genome Sequence of Enterohemorrhagic Eschelichia coli O157:H7 and Genomic Comparison with a Laboratory Strain K-12. DNA Res 8: 11–22.
- 20. Perna NT, Plunkett G, Burland V, Mau B, Glasner JD, et al. (2001) Genome sequence of enterohaemorrhagic Escherichia coli O157:H7. Nature 409: 529–533.
- 21. Alexander DC, Hao W, Gilmour MW, Zittermann S, Sarabia A, et al. (2012) Escherichia coli O104:H4 infections and international travel. Emerging Infect Dis 18: 473–476.
- 22. Ziebell K, Steele M, Zhang Y, Benson A, Taboada EN, et al. (2008) Genotypic characterization and prevalence of virulence factors among Canadian Escherichia coli O157:H7 strains. Appl Environ Microbiol 74: 4314–4323.
- 23. Vareille M, de Sablet T, Hindré T, Martin C, Gobert AP (2007) Nitric oxide inhibits Shiga-toxin synthesis by enterohemorrhagic Escherichia coli. Proc Natl Acad Sci U S A 104: 10199–10204.
- 24. Grad YH, Lipsitch M, Feldgarden M, Arachchi HM, Cerqueira GC, et al. (2012) Genomic epidemiology of the Escherichia coli O104:H4 outbreaks in Europe, 2011. Proc Natl Acad Sci U S A 109: 3065–3070.
- 25. Botstein D (1980) A theory of modular evolution for bacteriophages. Ann N Y Acad Sci 354: 484–490.
- 26. Wiens JJ, Morrill MC (2011) Missing Data in Phylogenetic Analysis: Reconciling Results from Simulations and Empirical Data. Syst Biol 60(5): 719–731.
- 27. Laing C, Buchanan C, Taboada E, Zhang Y, Karmali M, et al. (2009) In silico genomic analyses reveal three distinct lineages of Escherichia coli O157:H7, one of which is associated with hyper-virulence. BMC Genomics 10: 287.
- 28. Manning SD, Motiwala AS, Springman AC, Qi W, Lacher DW, et al. (2008) Variation in virulence among clades of Escherichia coli O157:H7 associated with disease outbreaks. Proc Natl Acad Sci U S A 105: 4868–4873.
- 29. Boerlin P (1999) Evolution of virulence factors in Shiga-toxin-producing Escherichia coli. Cell Mol Life Sci 56: 735–741.
- 30. Sokurenko EV, Gomulkiewicz R, Dykhuizen DE (2006) Source-sink dynamics of virulence evolution. Nat Rev Microbiol 4: 548–555.
- 31. Robinson CM, Sinclair JF, Smith MJ, O’Brien AD (2006) Shiga toxin of enterohemorrhagic Escherichia coli type O157:H7 promotes intestinal colonization. Proc Natl Acad Sci U S A 103: 9667–9672.
- 32. Steinberg KM, Levin BR (2007) Grazing protozoa and the evolution of the Escherichia coli O157:H7 Shiga toxin-encoding prophage. Proc Biol Sci 274: 1921–1929.
- 33. Chase-Topping M, Gally D, Low C, Matthews L, Woolhouse M (2008) Super-shedding and the link between human infection and livestock carriage of Escherichia coli O157. Nat Rev Microbiol 6: 904–912.
- 34. Laing C, Pegg C, Yawney D, Ziebell K, Steele M, et al. (2008) Rapid Determination of Escherichia coli O157:H7 Lineage Types and Molecular Subtypes by Using Comparative Genomic Fingerprinting. Appl Environ Microbiol 74: 6606–6615.
- 35. Poirier K, Faucher SP, Béland M, Brousseau R, Gannon V, et al. (2008) Escherichia coli O157:H7 survives within human macrophages: global gene expression profile and involvement of the Shiga toxins. Infect Immun 76: 4814–4822.
- 36. Donnenberg MS, Whittam TS (2001) Pathogenesis and evolution of virulence in enteropathogenic and enterohemorrhagic Escherichia coli. J Clin Invest 107: 539–548.
- 37. Serra-Moreno R, Jofre J, Muniesa M (2007) Insertion site occupancy by stx2 bacteriophages depends on the locus availability of the host strain chromosome. J Bacteriol 189: 6645–6654.
- 38. Besser TE, Shaikh N, Holt NJ, Tarr PI, Konkel ME, et al. (2007) Greater Diversity of Shiga Toxin-Encoding Bacteriophage Insertion Sites among Escherichia coli O157:H7 Isolates from Cattle than in Those from Humans. Appl Environ Microbiol 73: 671–679.
- 39. Reid SD, Herbelin CJ, Bumbaugh AC, Selander RK, Whittam TS (2000) Parallel evolution of virulence in pathogenic Escherichia coli. Nature 406: 64–67.
- 40. Wieler LH, McDaniel TK, Whittam TS, Kaper JB (1997) Insertion site of the locus of enterocyte effacement in enteropathogenic and enterohemorrhagic Escherichia coli differs in relation to the clonal phylogeny of the strains. FEMS Microbiol Lett 156: 49–53.
- 41. Zhang Y, Laing C, Zhang Z, Hallewell J, You C, et al. (2010) Lineage and host source are both correlated with levels of Shiga toxin 2 production by Escherichia coli O157:H7 strains. Appl Environ Microbiol 76: 474–482.
- 42. Shimizu T, Ohta Y, Noda M (2009) Shiga toxin 2 is specifically released from bacterial cells by two different mechanisms. Infect Immun 77: 2813–2823.
- 43. Bielaszewska M, Idelevich EA, Zhang W, Bauwens A, Schaumburg F, et al. (2012) Epidemic Escherichia coli O104:H4: Effects of antibiotics on Shiga toxin 2 production and bacteriophage induction. Antimicrob Agents Chemother. In press.
- 44. de Sablet T, Chassard C, Bernalier-Donadille A, Vareille M, Gobert AP, et al. (2009) Human microbiota-secreted factors inhibit shiga toxin synthesis by enterohemorrhagic Escherichia coli O157:H7. Infect Immun 77: 783–790.
- 45. Rasko DA, Webster DR, Sahl JW, Bashir A, Boisen N, et al. (2011) Origins of the E. coli Strain Causing an Outbreak of Hemolytic-Uremic Syndrome in Germany. N Engl J Med 365: 709–717.
- 46. Darling AE, Mau B, Perna NT (2010) progressiveMauve: multiple genome alignment with gene gain, loss and rearrangement. PLoS ONE 5: e11147.
- 47. Ogura Y, Ooka T, Iguchi A, Toh H, Asadulghani M, et al. (2009) Comparative genomics reveal the mechanism of the parallel evolution of O157 and non-O157 enterohemorrhagic Escherichia coli. Proc Natl Acad Sci U S A 106: 17939–17944.
- 48. Plunkett G, Rose DJ, Durfee TJ, Blattner FR (1999) Sequence of Shiga toxin 2 phage 933W from Escherichia coli O157:H7: Shiga toxin as a phage late-gene product. J Bacteriol 181: 1767–1778.
- 49. Rohde H, Qin J, Cui Y, Li D, Loman NJ, et al. (2011) Open-Source Genomic Analysis of Shiga-Toxin-Producing E. coli O104:H4. N Engl J Med 365: 718–24.
- 50. Sato T, Shimizu T, Watarai M, Kobayashi M, Kano S, et al. (2003) Genome analysis of a novel Shiga toxin 1 (Stx1)-converting phage which is closely related to Stx2-converting phages but not to other Stx1-converting phages. J Bacteriol 185: 3966–3971.
- 51. Eppinger M, Mammel MK, Leclerc JE, Ravel J, Cebula TA (2011) Genomic anatomy of Escherichia coli O157:H7 outbreaks. Proc Natl Acad Sci U S A 108: 20142–20147.
- 52. Kulasekara BR, Jacobs M, Zhou Y, Wu Z, Sims E, et al. (2009) Analysis of the genome of the Escherichia coli O157:H7 2006 spinach-associated outbreak isolate indicates candidate genes that may enhance virulence. Infect Immun 77: 3713–3721.
- 53. Shepard SM, Danzeisen JL, Isaacson RE, Seemann T, Achtman M, et al. (2012) Genome sequences and phylogenetic analysis of K88- and F18-positive porcine enterotoxigenic Escherichia coli. J Bacteriol 194: 395–405.
- 54. Strauch E, Hammerl JA, Konietzny A, Schneiker-Bekel S, Arnold W, et al. (2008) Bacteriophage 2851 is a prototype phage for dissemination of the Shiga toxin variant gene 2c in Escherichia coli O157:H7. Infect Immun 76: 5466–5477.
- 55. Sato T, Shimizu T, Watarai M, Kobayashi M, Kano S, et al. (2003) Distinctiveness of the genomic sequence of Shiga toxin 2-converting phage isolated from Escherichia coli O157:H7 Okayama strain as compared to other Shiga toxin 2-converting phages. Gene 309: 35–48.
- 56. Yang F, Yang J, Zhang X, Chen L, Jiang Y, et al. (2005) Genome dynamics and diversity of Shigella species, the etiologic agents of bacillary dysentery. Nucleic Acids Res 33: 6445–6458.