Isolation and Characteristics of Shiga Toxin 2f-Producing Escherichia coli among Pigeons in Kyushu, Japan

An increasing number of Shiga toxin 2f-producing Escherichia coli (STEC2f) infections in humans are being reported in Europe, and pigeons have been suggested as a reservoir for the pathogen. In Japan, there is very little information regarding carriage of STEC2f by pigeons, prompting the need for further investigation. We collected 549 samples of pigeon droppings from 14 locations in Kyushu, Japan, to isolate STEC2f and to investigate characteristics of the isolates. Shiga toxin stx 2f gene fragments were detected by PCR in 16 (2.9%) of the 549 dropping samples across four of the 14 locations. We obtained 23 STEC2f-isolates from seven of the original samples and from three pigeon dropping samples collected in an additional sampling experiment (from a total of seven locations across both sampling periods). Genotypic and phenotypic characteristics were then examined for selected isolates from each of 10 samples with pulsed-field gel electrophoresis profiles. Eight of the stx 2f gene fragments sequenced in this study were homologous to others that were identified in Europe. Some isolates also contained virulence-related genes, including lpfA O26, irp 2, and fyuA, and all of the 10 selected isolates maintained the eae, astA, and cdt genes. Moreover, five of the 10 selected isolates contained sfpA, a gene that is restricted to Shiga toxin-producing E. coli O165:H2 and sorbitol-fermenting Shiga toxin-producing E. coli O157:NM. We document serotypes O152:HNM, O128:HNM, and O145:H34 as STEC2f, which agrees with previous studies on pigeons and humans. Interestingly, O119:H21 was newly described as STEC2f. O145:H34, with sequence type 722, was described in a German study in humans and was also isolated in the current study. These results revealed that Japanese zoonotic STEC2f strains harboring several virulence-related factors may be of the same clonal complexes as some European strains. These findings provide useful information for public health-related disease management strategies in Japan.


Introduction
Recent reports have shown that cases of human infection by Shiga toxin 2f-producing Escherichia coli (STEC2f) are becoming more frequent, with pigeons being implicated as a possible source of infection [1,2]. For example, van Duynhoven et al. [3] screened 211 human stool specimens suspected to contain Shiga toxinproducing E. coli (STEC), and reported that three of seven STEC strains were positive for STEC2f. Prager et al. [1] also reported that STEC2f was present in 2.5% of STEC strains identified in their study. In Japan, STEC2f strain O128:HNM was isolated from the feces of a 1-year-old boy with diarrhea and abdominal pain in 2002 [4]. In addition, Sonntag et al. [5] showed that a STEC2f O128:HNM strain from human diarrhea was identical to those of pigeon isolates belonging to the serotype O128:H2, based on the restriction fragment length polymorphism pattern identified in Germany. These studies have shown that the occurrence of STEC2f in humans is clearly higher than previously thought, and that zoonotic transfer from pigeons is a potentially serious threat.
In Europe, STEC2f strains among pigeons and humans were investigated using several methods, including isolation of the pathogen from city pigeons, analysis of the isolates' characteristics, and relationships between isolates. Isolation of stx 2f (or the toxin, Stx2f) from pigeon droppings was reported in Europe [6]. STEC2f serotypes are variable, and so far include the O15, O18ab:HNM, O25:H7, O45, O45:HNM, O66:HNM, O75:HNM, O128:H2, O132, O135:HNM, O152:HNM, and O-untypeable serotypes in Japanese [7] and European pigeons [5,8,9]. The stx 2f sequences were analyzed [2] and multi-locus sequence typing (MLST) of the pathogen in humans was conducted [1]. Given that the capacity of STEC strains to cause human disease is strongly associated with other virulence-related genes, in addition to stx [10], Sonntag et al. [5] evaluated the potential virulence of STEC2f strains containing such virulence-related genes, including lpfA O26 (major subunit of long polar fimbriae of STEC O26), sfpA (major fimbrial subunit of Sfp fimbriae), irp 2 (iron-repressible protein 2), and fyuA (ferric yersiniabactin uptake receptor). Other virulence-related genes, such as eae (intimin), EHEC-hlyA (enterohemorrhagic E. coli hemolysin), and cdt (cytolethal distending toxin), were also detected in STEC2f strains in Europe [1]; however, there is only one study of this emerging pathogen in Japan, which sampled only 67 pigeons [7]. The purpose of the present study was to investigate the prevalence of STEC2f-harboring pigeons in Kyushu, Japan, to determine the phenotypic and genotypic characteristics of any resulting isolates, and to compare the genotypes, including virulence-related genes, of Japanese isolates with those reported in Europe. Accordingly, we successfully isolated STEC2f strains from the pigeon droppings and identified some overlap with virulence gene clones identified in Europe, as well as documenting serotype O119:H21 as STEC2f for the first time. This study represents a critical appraisal of the occurrence of pigeons infected with the STEC2f zoonotic pathogen in Japan.

Sampling of pigeon droppings
Samples of droppings from city pigeons were collected for both experiment 1 (single dropping collection) and experiment 2 (pooled dropping collection) (see below) in Kyushu (Fig. 1), Japan, between December 2008 and March 2009. Kyushu is Japan's third largest island. It is located southwest of the main island of Honshu. For both experiments, droppings were collected from the ground while the pigeons were eating feed that we supplied, except at location K, where droppings were collected from plastic boards that had been sterilized with ethyl alcohol (80%). Therefore, not every sample came from a separate bird. Sampling locations were selected randomly from parks, station squares, and religious sites. In experiment 1, 549 individual pigeon droppings were collected from 14 locations in Kyushu (A-E, G-I, K-P in Fig. 1). Each dropping was collected in a 15-mL conical tube. In experiment 2, which was carried out for additional collection of STEC2f-isolates, 19 samples consisting of pooled pigeon droppings were collected from 13 locations (A-L, P in Fig. 1) in Kyushu.
Detection of stx 2f gene and isolation of STEC2f from pigeon droppings Each dropping sample for experiment 1 was mixed with 10 mL of buffered peptone water (Oxoid, Hampshire, UK), while each pooled sample in experiment 2 was mixed with 225 mL of buffered peptone water. The collected samples in the buffered peptone water were incubated overnight at 42uC to reflect the body temperature of pigeons. Aliquots (0.1 mL in experiment 1 and 1 mL in experiment 2) of the cultures were centrifuged at 3,2206 g for 20 min, and pellets were re-suspended in 0.5 mL of Tris-EDTA buffer. Cell suspensions were then boiled for 10 min and centrifuged at 3,2206 g for 20 min for experiment 1 and at 20,8146 g for 10 min for experiment 2. A 1-mL aliquot of each supernatant was used as a template for PCR-based detection of stx 2f . PCR was performed as described previously by Nakao et al. [11]. Primers G3-F (59-TTT ACT GTG GAT TTC TCT TCG C-39) and G3-R (59-TCA GTA AGA TCC TGA GGC TTG-39), at a final concentration of 0.5 mM, and GoTaq Green Master Mix (Promega, Madison, WI, USA) were used in the PCR reaction mixture.
The stx 2f PCR-positive mixed cultures were streaked onto deoxycholate-hydrogen sulfide-lactose agar (Eiken Chemical Co., Tokyo, Japan) and the plates were incubated at 37uC overnight. Presumptive E. coli colonies (25-96/sample) were isolated on nutrient agar plates (Eiken Chemical Co.) and incubated at 37uC overnight. The isolates were then re-tested for stx 2f using the PCR method described above. The biochemical characteristics of the stx 2f -positive isolates were examined using triple sugar iron agar (Eiken Chemical Co.), lysine decarboxylase broth (Eiken Chemical Co.), sulfide indole motility medium agar (Eiken Chemical Co.), methyl red-Voges-Proskauer medium (Nissui Pharmaceutical Co., Tokyo, Japan), Simmons citrate agar (Eiken Chemical Co.), and cytochrome-oxidase test strips (Nissui Pharmaceutical Co.) to determine whether the isolates were E. coli. Serological characteristics of the isolates were investigated using O-and H-antisera (Denka Seiken Co., Tokyo, Japan). The isolates were also tested using a previously described PCR-based method for discriminating between E. coli and Escherichia albertii [12,13]. Genotyping of flagella antigen (H-typing), from isolates that were identified as having a type other than H-in the phenotyping assay, was also carried out by restriction fragment length polymorphism analysis, as previously described [14].

Pulsed-field gel electrophoresis
All of the STEC2f isolates (n = 23) were tested with pulsed-field gel electrophoresis (PFGE) to select representative isolates from each sample for further characterization. PFGE was carried out as previously described [15] for 11 isolates, while the other 12 isolates that showed Tris-dependent degradation were subsequently analyzed by PFGE with 50 mM thiourea described byRömling and Tümmler [16].

Full-length sequencing of stx 2f
Full-length sequence analysis of the stx 2f genes from the 10 selected isolates was carried out as previously described [17]. The sequences were assembled using the SeqManII program in the Lasergene software package (DNASTAR, Madison, WI, USA), and were compared with stx 2f reference sequences (AB232172, M29153, AJ010730, AJ270998, and AB472687 from the DNA Data Bank of Japan, or DDBJ). Cluster analysis was performed using the unweighted pair-group method with arithmetic averages using MEGA 4 software [18].

Toxin assay
Filtrates of cultures of each of the isolates, with mitomycin C (MMC) (0.2 mg/L), were tested for cytotoxic activity in Vero cells [23], and in a reversed passive latex agglutination assay (RPLA) using a VTEC-RPLA kit (Denka Seiken) to detect Shiga toxin, as previously described [17]. E. coli O157:H7 (ATCC 43894), harboring stx 1 and stx 2 , was also examined as a reference strain in both tests.

Multi-locus sequence typing (MLST)
Housekeeping genes adk (adenylate kinase), fumC (fumarate hydratase), gyrB (DNA gyrase), icd (isocitrate/isopropylmalate dehydrogenase), mdh (malate dehydrogenase), purA (adenylosuccinate dehydrogenase), and recA (ATP/GTP binding motif) were used for MLST analysis. Cell treatment and sequencing were performed as previously described [24]. Primers for the sequencing reactions are listed in the MLST Database at the Environmental Research Institute, University College Cork, Ireland (http://mlst.ucc.ie/, available April 1, 2012). Sequences were submitted to the database website and were assigned existing or novel allele type numbers and sequence type (ST) numbers defined by the database. Composite STs were assigned based on the set of allele types derived from each of the seven loci. Allele sequences for each strain were then concatenated in the order adk-fumC-gyrB-icd-mdh-purA-recA for a final composite length of 3,423 bp. The concatenated sequences were aligned using ClustalW software [25], and a phylogenetic tree was constructed using the neighborjoining method to compare with other published STs (Center for Information Biology and DDBJ). Phylogenetic analyses were performed using NJplot (http://pbil.univ-lyon1.fr/software/ njplot.html). A minimum spanning tree generated from the data for ST 20, ST 2684, ST 2685, ST 28, ST 722, ST 40, and ST 11 was also described using Prim's algorithm from the PubMLST site (http://pubmlst.org/, accessed May 2, 2013).

Ethics statement
No specific permits were required for the described field studies, locations, and activities. The sampled locations are not privatelyowned or protected in any way, and the field studies did not involve endangered or protected species. At no point did the researchers come into physical contact with the pigeons. This study was carried out in strict accordance with the guidelines of the Regulations for the Ethical and Humane Use of Experimental Animals at Fukuoka Institute of Health and Environmental Sciences, which is based on domestic standards, and approved by the Animal Ethics Committee of Fukuoka Institute of Health and Environmental Sciences under permit number 250802.

Detection of Shiga toxin 2f-producing E. coli
In experiment 1, stx 2f was detected in 16 (2.9%) of 549 dropping cultures from four of the 14 locations by PCR. The incidence varied from 0-17.0% (Table 1). In experiment 2, stx 2f was detected in four of 19 pooled dropping samples at four (A, H, I and K) of the 13 locations. stx 2f was detected in one out of the two samplings and one out of five samplings at locations I and K, respectively. The combined results of both the experiments showed stx 2f was detected at seven of the 16 locations (Fig. 1). No common characteristics or features were found in the seven locations where stx 2f was detected.
Phenotyping and genotyping of selected isolates chosen by pulsed-field gel electrophoresis We obtained 23 STEC2f isolates from 10 samples at seven locations in total: 14 isolates from seven samples from four locations in experiment 1, and nine isolates from three samples from three locations in experiment 2. These isolates were identified as STEC2f-positive by PCR and as E. coli by biochemical tests. None of the isolates were identified as E. albertii by PCR [12,13].
In the PFGE analyses, no remarkable differences in fragment patterns (defined as more than three fragments [26]) were observed between isolates from the same samples, suggesting that these isolates were the same or closely related strains (Fig. 2). Isolates from sample location G (G5 in Fig. 2) and H (H1, H2, and H3 in Fig. 2) showed the same PFGE profiles. Therefore, we selected one isolate from each of the 10 unique samples covering all STEC2f-positive locations (n = 7) across the two experiments. The 10 selected STEC2f-isolates showed several serotypes and characteristics ( Table 2). The results of genotyping and phenotyping H-type assays were identical in each of six isolates. The 10 selected isolates showed a variety of O serogroups that were specific to location, and no O serogroup was shared between the different locations, except for untypeable. Selected isolates from each sample showed the following characteristics: supernatants from the isolate cultures containing MMC had titers of 16-256 in the RPLA test for Stx2 only, while ATCC 43894, the control strain, had titers of 16 and 8,192 for Stx1 and Stx2, respectively, in the MMC treatment assay. For the Vero cell assays using culture supernatants, the titers were between 4,096 and 262,144 with MMC, while ATCC 43894 with MMC treatment showed a titer of 65,536 (Table 2). Normal Vero cells were observed in a control plate containing only medium and MMC.
The isolates contained several virulence factors, as well as two different stx 2f STs (Table 2). Eight out of 10 isolates had stx 2f STs that were previously identified in European isolates [2], whereas two other isolates had sequences that showed identity to a human isolate from Nagasaki, Kyushu [27] (Fig. 3). The major characteristics of the virulence factors are summarized in Table 2. The gene that was detected by PCR with the primers for sfpA was confirmed as sfpA by DNA sequencing. The sequences from five isolates were completely identical to each other (525/ 525 bp, 100%), and only differed at 10 bases (10/525) from the sfpA sequence reported for E. coil O157:H-(AF401292). All of the isolates harbored cdt (we did not perform sub-typing) and astA, as well as eae. The dominant intimin type was type b in the 39 terminal region, and b1 type in the 59 terminal region. The MLST results revealed six STs in the selected isolates (Table 2, Fig. 4). No correlation was observed among stx 2f -sequence patterns, eae-types, and STs, although presence-absence patterns of virulence related genes might be correlated with intra-isolate STs ( Table 2).

Database accession numbers
The DDBJ accession numbers for the stx2f sequences of the isolates are AB499801-AB499803 and AB499807-AB499813. The sfpA sequences are submitted under numbers AB826561-AB826565. The sequences generated by multi-locus sequence typing (MLST) are available from the MLST database (http:// mlst.ucc.ie/).

Discussion
The current study produced three important findings in regard to the isolation and characteristics of STEC2f in pigeons in Japan. First, STEC2f was isolated from pigeon excrement in Kyushu, Japan. The percentage of stx 2f -positive droppings varied from 0-17% across the 14 locations examined. Second, some of the stx 2fgene fragment sequences identified in the current study were previously identified in Europe [2], and STEC2f O145:H34, ST 722, which has been isolated in Europe [1], was also obtained from pigeon droppings at location I in the present study. Last, the isolates varied in their serotypes, phenotypes, and genotypes. In particular, some isolates contained sfpA, which is reportedly restricted to some STEC strains [28], suggesting horizontal transfer of this virulence gene.
The isolation of STEC2f from pigeon droppings in Kyushu adds to the global surveillance data on STEC2f from pigeons. STEC2f from pigeons has previously been reported in other countries, including Finland [8], Germany [6], Italy [9], and India [29]. Kobayashi et al [7] reported that the frequency of STEC2f carriage in pigeons in Tokyo Bay, Japan, was 3.0%, while 3.5% of 29 studied pigeons in Finland carried STEC2f [8] Grobmann et al. [6] reported that 30.0% and 37.0% of 366 German pigeon Table 1. stx 2f PCR-positive droppings from pigeons from 12 locations in Kyushu, Japan (experiment 1).

Sampling location
No     Escherichia coli isolates that were isolated in this study.
dropping samples contained STEC2f, as analyzed by PCR and DNA-DNA hybridization, respectively. However, as not all of our samples came from separate pigeons, comparing our results with previous studies is difficult. Further investigation with direct sampling from pigeon cloaca is required to measure carriage rates of STEC2f in pigeons in Kyushu, Japan. The present STEC2fpositive isolates showed various lineages and harbored several virulence factors. All of the selected isolates contained eae, astA, and cdt, while some of these isolates also had ''the high virulence pathogenicity island'' characterized by the genes irp 2 and fyuA, in addition to genes for the major fimbrial subunit, LPFO26 (lpfA O26 ) [5]. Moreover, one O119:H21 and four O-untypeable:H16 isolates harbored sfpA (Table 2), which encodes the Sfp protein, the major pilin subunit of STEC O165:H2 and sorbitolfermenting STEC O157:NM strains [28]. This indicates the horizontal transfer of this virulence gene to other serotypes, which is a novel and an important finding. The presence of eae, the locus coding for enterocyte effacement, in the STEC2f-positive strains is similar to previous findings [6,9]. In particular, the confirmation of the b eae subtype in some of the present isolates is consistent with reports by Morabito et al. [9]. In other bacteria, some E. albertii strains examined in a report by Ooka et al. [30] encoded both eae (subtype N1.1 or N2) and stx 2f . In the present study, tested strains were confirmed to be E. coli, not E. albertii, using specific PCR and MLST analyses (Fig. 4)  Serotyping and MLST revealed that there may be a relationship between Japanese and European STEC2f strains. This is the first study to document serotype O119:H21 as being STEC2f, while O128:HNM, O145:H34, and O152:HNM were previously reported as human or pigeon STEC2f isolates [1,[3][4][5]27,29,32,33]. In particular, the O128 group, including  Table 2. doi:10.1371/journal.pone.0086076.g003 O128:HNM, have been isolated by many laboratories in Japan [4] and Europe [1,5,32]. In addition, the 1,230-bp stx 2f sequence of some of the present isolates was the same as that identified for European isolates. STEC2f O145:H34, which is ST 722, was detected in both German human samples [1] and in Japanese pigeon excrement. These results suggest that STEC2f strains in Kyushu, Japan, and those in Europe may be in the same clonal complexes or lineage. It is important to identify the source of transmission of this pathogen, harboring several virulence-related genes, between Japan and Europe. Our findings suggest that STEC2f strains in pigeons represent a potential hazard in Japan. The findings contribute to public health in Kyushu and other parts of Japan, where STEC2f, and its relation to pigeons has not been reported. Because commercial stx PCR primers seldom detect the stx 2f gene [17], we recommend testing E. coli isolates with stx 2fspecific PCR primers. This knowledge is particularly useful for public health in Japan.