O-serogroups, virulence genes, antimicrobial susceptibility, and MLST genotypes of Shiga toxin-producing Escherichia coli from swine and cattle in Central China

Background Shiga toxin-producing Escherichia coli (STEC) is a leading cause of worldwide food-borne and waterborne infections. Despite an increase in the number of STEC outbreaks, there is a lack of data on prevalence of STEC at the farm level, distribution of serogroups, and virulence factors. Results In the present study, a total of 91 (6.16%) STEC strains were isolated from 1477 samples including pig intestines, pig feces, cattle feces, milk, and water from dairy farms. The isolation rates of STEC strains from pig intestines, pig feces, and cattle feces were 7.41% (32/432), 4.38% (21/480), and 9.57% (38/397), respectively. No STEC was isolated from the fresh milk and water samples. By O-serotyping methods, a total of 30 types of O-antigens were determined, and the main types were O100, O97, O91, O149, O26, O92, O102, O157, and O34. Detection of selected virulence genes (stx1, stx2, eae, ehxA, saa) revealed that over 94.51% (86/91) of the isolates carried more than two types of virulence associated genes, and approximately 71.43% (65/91) of the isolates carried both stx1 and stx2, simultaneously. Antimicrobial susceptibility tests showed that most of the STEC isolates were susceptible to ofloxacin and norfloxacin, but showed resistance to tetracycline, kanamycin, trimethoprim-sulfamethoxazole, streptomycin, amoxicillin, and ampicillin. MLST determined 13 categories of sequence types (STs), and ST297 (31.87%; 29/91) was the most dominant clone. This clone displayed a close relationship to virulent strains STEC ST678 (O104: H4). The prevalence of ST297 clones should receive more attentions. Conclusions Our preliminary data revealed that a heterogeneous group of STEC is present, but the non-O157 serogroups and some ST clones such as ST297 should receive more attentions.


Isolation of STEC
A total of 1477 samples, including 432 samples of intestinal contents from pigs with diarrhea, 480 fecal samples from pigs with diarrhea, 397 fecal samples from cows with diarrhea, 99 samples of fresh milk and 69 water samples from dairy farms, were collected from four provinces of Central China (Hubei, Anhui, Hunan, Henan) for PCR detection of Shiga toxin encoding genes (stx) and STEC isolation. Of the 1477 samples detected, 119 (8.06%) samples were positive for stx 1 and/or stx 2 . STEC strains were isolated from 91 (76.47%) of the 119 stx-positive samples. The isolation rates of STEC strains from pig intestines, pig feces, and cattle feces were 7.41% (32/432), 4.38% (21/480), and 9.57% (38/397), respectively ( Fig. 1a, Table 1). However, there were no STEC strains being isolated from the fresh milk and water samples collected (Fig. 1a, Table 1). Biochemical tests showed that all isolates were capable of fermenting glucose, maltose, lactose, and xylose, raffinose, lysine, and ornithine, but were unable to use gluconate, phenylalanine, and citrate.

Cytotoxicity
Cytotoxicity tests showed that all isolates positive to stx 1 and/or stx 2 were capable of making Vero cells rounding and exfoliation. However, no cytopathic effect was observed in the cells inoculated with the isolates negative to both stx 1 and stx 2 .

Discussion
STEC is a leading cause of foodborne and waterborne infections worldwide, food-producing animals such as cattle, and pigs are major reservoirs for STEC [23]. Among different kinds of food producing animals, cattle and other ruminants are considered to be the major reservoirs for STEC [23,27,28]. STEC strains are more frequently isolated from cattle and other ruminants than from other animals such as pigs, cats, and dogs [23,29]. In agreement with these suggestions, the rate of STEC isolation from cattle (9.57%) was higher than that from pigs (5.81%). However, the isolate rate of STEC from cattle feces (9.57%) in present study is different from the reports from the other countries [2,29]. These differences might be explained by differences in feed, seasonal peak, age, or detecting methods [2].
A total of 30 types of O-antigen were determined for the 91 STEC isolates by O-serotyping methods, with the exception of 4 isolates which were not typable (Fig. 2). This might be because there are only 50 types of O antisera are available, and these four strains do not react with the available antisera. Among these 30 categories of O-serogroups, the most frequently occurring serogroups were O100, O97, O91, O149, O26, O92, O102, O157, and O34 (Fig. 2). These serogroups have been isolated from pigs, cattle, sheep, and water in both China and the other countries [30][31][32][33][34][35][36][37]. It has been known that O157 is the most common serogroup that causes human illness in most parts of the world [4]. It has been also reported that cattle are the most common reservoir of E. coli O157 [38]. Corresponding to this suggestion, the O157 serogroup was only determined within the isolates from cattle in the present study (Fig. 2). Although there was no STEC O157 being isolated from pigs in Central China in the present study, a recent study has revealed that the overall prevalence of E. coli O157:H7 in pig  farms around Hubei, a province located in Central China, is approximately 12.8% (16/125) [13]. These findings suggest that the prevalence of E. coli O157 in this region is still a problem. In addition to O157, O26 also displayed a high proportion of identification (Fig. 2). It is worthy of note that this type of O-antigen has been declared by the U.S. Department of Agriculture (USDA) as one of the "Big 6" (O26, O45, O103, O111, O121, and O145) non-O157 serogroups that are most commonly associated with severe illness in humans [4]. STEC O26 has been detected and isolated from diarrheal patient in China [11]. It should be noted that another member of the "Big 6", the O145, was also identified in the present study (Fig. 2). In addition, STEC O149 has been also detected and isolated from diarrheal patient in China [11]. The determination of these non-O157 serogroups represents a great risk on public health and should also receive more attentions. Virulence genotyping based on the detection of six virulence genes (stx 1 , stx 2 , ehxA, eae, and saa) showed that the detection rates of stx 1 (82.42%), stx 2 (89.01%), and ehxA (86.81%) were higher than those of the other virulence genes (Fig. 3a); most of the STEC isolates possess stx 1 , stx 2 , and ehxA simultaneously (Fig. 3b). It is known that both stx 1 and stx 2 are responsible for encoding the Shiga toxin, which is the most important and common virulence factors of STEC [15]. In particular, the detection rate of stx 2 (89.01%) was higher than stx 1 (82.42%), and a small proportion of isolates (5.49%, 5/ 91) only carried stx 2 ( Fig. 3a and b). It has been reported that stx 2 is more often associated with severe disease [39]. Therefore, those strains might be more harmful. Both Stx 1 and Stx 2 have several subtypes, and some subtypes are more frequently associated with human disease [4]. It has been widely documented that STEC isolates from pigs normally harbor Stx 2e subtype [40][41][42], and in agreement with these studies [40][41][42], approximately 91.20% of the porcine isolates positive to stx 2 determined in the present study harbored this subtype (Stx 2e ).
STEC producing Stx2e is known to be closely associated with edema disease in pigs [43], the high proportion of stx 2e detection in STEC isolates from pig intestines and/ or feces in this study suggest a big threat to the pig health. Although Stx 2e -producing STEC strains are still not proposed as pathogens for humans [43], active actions are still required to control and decrease the prevalence of such strains in pigs in a One Health perspective. In the present study, we also identified several other Stxsubtypes such as Stx 1a , Stx 1c , Stx 1d , Stx 2b , Stx 2d , Stx 2a , Stx 2c , and Stx 2g . Among these subtypes, stx 2a and stx 2c are proposed to be associated with high virulence and the ability to cause hemolytic-uremic syndrome (HUS), Fig. 5 Distribution of the sequence types (STs) among the STEC isolates. Columns in pink displays the percentage of STECs isolated from the feces and intestinal contents of pigs; Columns in sky blue displays the percentage of STECs isolated from the feces of cows; Columns in gray displays the percentage of the total STECs isolated herein (STECs from pigs plus STECs from cows) Fig. 6 Phylogenetic analysis on different STEC sequence types clones. The tree was constructed based on the MLST data by MEGAX [26], using neighbor-joining algorithm with 1000 bootstrapping while stx 2d , stx 2e , stx 1a , and stx 1c occurred in milder or asymptomatic infections [43,44]. The detection of those subtypes in STEC strains from food producing animals such as pigs and cows we detected in this study represents a high risk on public health. It is worthy of note that stx 1d , stx 2b , stx 2g have been also detected in STEC strains from patients in Demark, however, HUS does not develop in these patients [45][46][47].
In addition to stx 1 and stx 2 , the prevalence of ehxA (86.81%) was also very high, showing a good agreement with previously studies [48][49][50][51]. It is worthy of note that ehxA is generally used as a diagnostic indicator because the presence of ehxA is frequently correlated with the Shiga toxin [49,51]. In agreement with this conclusion, ehxA displayed a high detection rate from the stx-positive STEC strains in the present study (Fig. 3b). In contrast to these genes which have high rates of detection, the detection rates of eae, and saa were relatively low. These results are similar to previously studies [50,51], suggesting that these virulence genes are not common. However, their presence in particularly the detection of eae should be given a concern. It has been reported that the combination of eae and stx 2 has an especial association with the development of HUS and bloody diarrhea [46,47,52]. In the present study, all eae-positive STEC strains isolated in Central China were detected to be positive for stx 2 ( Table 1). The determination of such strains represents a high risk on public health in this region.
The antimicrobial resistance (AMR) of STEC is also a serious problem that the world is now facing. It has been reported that STEC isolates from both humans and foodproducing animals displayed resistance most often to tetracycline, kanamycin, trimethoprim-sulfamethoxazole, streptomycin, amoxicillin, and ampicillin [36,[53][54][55][56]. In agreement with these studies, a low proportion of STEC isolates from the present study was susceptible to those types of antimicrobials (Fig. 4). These findings suggest a serious profile of AMR in STEC in food-producing animals. While there is a number of articles reporting the colistin resistance prevalence in E. coli [57][58][59], it is worthy of note that all STEC isolates were sensitive to colistin in the present study.
MLST is also a strategy commonly used for STEC surveillance [36,60]. In this study, 13 types of STs were determined for the 91 STEC isolates. In particularly, many isolates belonging to different STs possessed the same serogroups (Table 1). These findings are consistent with the findings of other publications [61,62], suggesting that STEC isolates with the same serogroups might have genotypical diversity. Among the determined STs, ST297 possesses the highest rate of isolation (31.87%) compared to the remaining identified STs (Fig. 5). Interestingly, ST297 is rarely reported in STEC. A previous study determined five ST297 from 75 STEC food strains, with a detection rate of 6.67% [63]. In another study, the detection rate of ST297 among STEC isolates from cattle in Korea was only 4.69% (3/64) [64]. Our results are quite different from these studies, suggesting that the prevalence of the ST in different regions of the world might be different. The ST297 isolates harbored many types of O-antigens, including O26 and O145, the important members of the "Big 6" declared by the USDA [4]. In particular, all O26 isolates recovered in the present study are ST297 clone (Table 1). It has been reported that the STs of STEC O26 associated with a broad spectrum of diseases in Europe are ST29 and/or ST21 [65][66][67][68][69][70]. In addition, the ST297 clones isolated in this study displayed a close relationship to STEC ST678 (O104: H4) (Fig. 6). It should be noted that the STEC ST678 (O104: H4) isolates have caused the outbreak of human gastroenteritis and human hemolytic-uremic syndrome in Europe [8,71]. In the next step, we intend to do follow up study to determine the genetic and phenotypical characteristics of these ST297 clones. In addition, the sequence types of the two STEC O157 were determined as ST542 (Table 1). Although the sequence type of STEC O157 is normally determined as ST11 [61,72], O157 isolates determined as non-ST11 have been also documented elsewhere. For instance, four O157 isolates from the US and/or UK are determined as ST1804 [62]. These findings suggest there might be other STs for STEC O157. In the next step, we will do follow up study to determine the genetic and phenotypical characteristics of these two isolates.

Conclusions
In conclusion, the present study performed an isolation and a characteristic analysis of STEC from pigs and cattle. Our preliminary data revealed that a heterogeneous group of STEC is present, but the non-O157 serogroups and some ST clones such as ST297 should receive more attentions. In the next step, we intend to do a follow up study to correlate the pathogenicity of these STEC with the stx-subtypes as well as the ST clones.

Sample collection and bacterial isolation
A total of 1477 samples were tested in this study. These samples included intestinal contents from pigs with diarrhea (432 samples), fecal samples from pigs (480 samples) and cows (397 samples) with diarrhea, fresh milk (99 samples), and water samples from dairy farms (69 samples) (Fig. 1a). The 912 pig-associated samples (feces and intestinal contents) were collected from 323 pig farms in Central China (Hubei, Anhui, Hunan, Henan) between 2016 and 2017, while the 565 cow-associated samples (feces, milk, and water samples from dairy farms) were from three dairy farms in different regions Hubei Province in 2017 ( Fig. 1a and b). Bacterial isolation was performed following a previously described protocol with some modifications [2]. In brief, each of the samples were mixed in sterilized 0.9% normal saline by vortexing. After a centrifugation at 500×g for 1 min, 500 μL of the supernatant of the mixture was inoculated into 5 mL modified E. coli broth (Nissui, Tokyo, Japan) and incubated at 37°C for 18~24 h.
After that, genomic DNA was extracted from the cultures by boiling 100 μL aliquot of each incubated broth directly, as described previously [2]. The extracted DNA was evaluated by electrophoresis on a 1% agarose gel and/or using a Nanodrop2000 (Thermo Scientific, Waltham, USA). Presence of the Stx encoding gene stx 1 and/ or stx 2 was determined by PCR assays using the genomic DNA extracted herein as the template and the primers listed in Table 2. PCR reaction was performed in a 25 μL mixture containing 2 μL of the template DNA, 2.5 μL of 10× PCR Buffer (TAKARA, Japan), 2 μL of dNTP (TAKARA, Japan), 0.5 μL of rTaq (TAKARA, Japan), each of the forward and reverse primers 0.5 μL, and 17.0 μL of nucleotide-free water (TAKARA, Japan). Thermocycler conditions used for PCR were 95°C for 5 min, followed by 30 cycles of denaturation at 94°C for 30 s, annealing at different temperatures listed in Table  2 for 40 s, and extension at 72°C for 1 min, with a final extension at 72°C for 10 min before storage at 4°C. DNA from STEC O157:H7 strain EDL933 and nucleotide-free water were included as positive and blank controls, respectively. The PCR product was visualized using 1% agarose gel electrophoresis under ultraviolet light.
In the next step, bacterial cultures positive to at least one of stx 1 and stx 2 were streak-plated onto sorbitol MacConkey agar (Hangzhou Microbial Reagent CO., LTD, Hangzhou, China), and incubated at 37°C for 18~24 h. After this stage, the isolates were purified and cultured following the standard methods used for bacterial identification [73]. Presumptive isolates of E. coli were finally confirmed via Galanz staining, biochemical testing, and 16S rRNA amplification and sequencing.

Serotyping and virulence genotyping
O-polysaccharide antigens serogroups of STEC isolates were determined by Slide agglutination test based on the reaction of the bacterial strains against the 50 kinds of O antisera purchased from China Institute of Veterinary Drug Control (Beijing, China). STEC O157:H7 strain EDL933 was used as positive control.
Virulence genotyping was performed by PCR assays amplifying another three virulence associated genes eae, ehxA, and saa of with primers listed in Table 2. The PCR volume and procedure were the same as that used for determining the Stx encoding genes. Positive and blank control samples were included in each set of reactions. The PCR product was visualized using 1% agarose gel electrophoresis under ultraviolet light. The stx subtypes (stx 1a , stx 1c , stx 1d , stx 2a , stx 2b , stx 2c , stx 2d , stx 2e , stx 2f , stx 2g ) were also determined by PCR assays with primers and reaction procedures described previously [16].

Cytotoxicity
Vero cells (purchased from ATCC) were used to test the cytotoxicity of the STEC strains isolated herein. In brief, isolates were inoculated in Luria-Bertani (LB) broth (Sigma-Aldrich, MO) and shaken at 37°C for 18~24 h. Bacterial culture were then centrifuged at 20000×g for 40 min, followed by a filtration through a 0.22 μm membrane. Filtrate was inoculated into Vero cells and the cells were incubated at 37°C for 18~24 h to observe the morphology. Filtrates collected from STEC O157:H7 strain EDL933, E. coli DH5α, cell medium were included as controls.

Multilocus sequence typing
Multilocus sequence typing (MLST) was performed using the previously described protocols [60]. Nucleotide sequences of seven housekeeping genes (adk, fumC, gyrB, icd, mdh, purA, and recA) were amplified and sequenced using the primers listed in Table 2. PCR reaction was performed in a 50-μl reaction mixture containing 2 μl of the template DNA, 2 μL of dNTP PurA-R2 CATACGGTAAGCCACGCAGA mixture (TAKARA, Japan), 5 μL of 10 × PCR buffer (TAKARA, Japan), 0.5 μL of rTaq polymerase (TAKARA, Japan), each of the forward and reverse primer 1 μL, and 38.5 μL of nuclease-free water. The reaction was performed under the following standard cycling procedure: an initial denaturation at 95°C for 5 min, followed by 30 cycles of denaturation at 94°C for 30 s, annealing at 54-60°C for 45 s (see Table 2), extension at 72°C for 1 min, and a final extension at 72°C for 10 min. The PCR products were initially analyzed by electrophoresis on a 1% agarose gel. Products with the correct size were sequenced at Sangon (Shanghai, China). Nucleotide sequences of the housekeeping genes were submitted to the Escherichia coli MLST Database (http://mlst.warwick.ac.uk/mlst/dbs/Ecoli) to determine the sequence types automatically. Phylogenetic tree was generated based on the MLST data by MEGAX [26], using neighborjoining algorithm with 1000 bootstrapping.