Neelam Taneja, Department of Medical Microbiology, Post Graduate Institute of Medical Education and Research, Chandigarh 160012, India. E-mail: firstname.lastname@example.org
To study the prevalence and distribution of various variants in the stx gene of Shiga toxin–producing Escherichia coli (STEC) isolated from diverse environmental sources (animal stool, meat) and human illness, from a large geographic area in India, and to understand the association between variants, serotype distribution and human disease.
Methods and Results
A surveillance for STEC was conducted in the semi-urban and rural areas of Punjab, Himachal, Haryana and Chandigarh. Shiga toxin–producing Escherichia coli isolates (80 animal stool, 39 meat, 21 human stool from diarrhoea and HUS cases) were characterized for stx variants by PCR. Shiga-like toxin (Stx) was detected using Ridascreen-EIA assay. Variant stx2c was the most common (25·1%), followed by stx1d (13%), stx1c (10·7%) and stx2d (9·2%), whereas stx2e, stx2f and stx2g were absent. Only 8/21 (38%) human isolates harboured stx variants, of which stx2c and stx2d were found in 2 and 1 isolates, respectively. The low frequency of carriage of these potentially more pathogenic variants may explain the low severity of human illness seen in India. Shiga-like toxin was detected in only 42 of the isolates positive for the stx genes probably due to the low levels of toxins produced. Serogroup distribution was found to be diverse, suggesting the lack of any predominant circulating type.
The presence of stx variants 1c, 1d, 2c and stx2d in diverse environmental and human sources in India was demonstrated. The prevalence of the most common subtype stx2c found in this study in animal isolates may pose a threat to the public health. We report the subtyping of human STEC isolates and report the presence of stx1d subtype for the first time from India.
Significance and Impact of the Study
We demonstrated the presence of potentially pathogenic subtypes in the environmental specimens which may act as a reservoir for human infections. Serogroups new to India were also reported.
Shiga toxin–producing Escherichia coli (STEC) are important foodborne zoonotic pathogens of great clinical and public health concern that cause various clinical syndromes ranging from diarrhoea, haemorrhagic colitis (HC) to life-threatening sequelae like haemolytic uraemic syndrome (HUS). More than 200 STEC serotypes have been described. Of these, O157:H7 is the most predominant serotype reported world-wide from human illness, whereas O26:H11, O111, O113:H2 and O145 are the most common non-O157 serotypes causing human illness (Boerlin et al. 1999). Minced meat is the most common food implicated in the transmission of STEC, and cattle, sheep and goat are considered the most important reservoirs (Rangel et al. 2005). The cardinal pathogenetic factor of STEC is the production of Shiga-like toxins (Stx), which are classified into two broad immunologically different types, Stx1 and Stx2. Both the toxins have a common receptor for their entry into the cell and have the same intracellular mechanism of action. The stx1 and stx2 genes that encode Shiga toxins (Stx1 and Stx2) have 56% similarity at the amino acid sequence level (Jackson et al. 1987). The sequence of the stx1 gene is highly conserved, and only a few variants such as stx1c and stx1d have been reported (Brett et al. 2003; Burk et al. 2003). The Stx2 group is a closely related family of 20 variants with 84–99% similarity to stx2 gene (Johannes and Romer 2009). Of these, toxins Stx2c, Stx2d, Stx2e, Stx2f and Stx2g are most commonly reported (Schmitt et al. 1991; Melton-Celsa et al. 1996; Pierard et al. 1998; Schmidt et al. 2000; Jelacic et al. 2003). Formation of toxin Stx2e is associated with oedema disease in pigs and is rare in human isolates (Caprioli et al. 1993). Toxin Stx2f is found in feral pigeons as natural reservoirs. Toxin Stx2g has been described as a new and rare Stx2 variant occurring in bovine STEC (Beutin et al. 2007).
The capacity of STEC strains to cause disease in humans depends on the toxin genotype (Friedrich et al. 2002). A number of studies have implicated subtypes Stx2 and Stx2c more often in HUS than the other Stx2 subtypes (Caprioli et al. 1995; Eklund et al. 2002; Ethelberg et al. 2004). STEC strains harbouring stx2d and stx2e sequences have also been isolated from the patients with HUS (Ito et al. 1990; Lindgren et al. 1993). The variations in the Stx2 amino acid sequences have a direct impact on the properties of Shiga toxins. Therefore, subtyping of Stx 1 and Stx2 is an important tool in identifying strains with different degrees of virulence and predicting the potential risk of STEC to public health. PCR techniques have been developed to differentiate stx variants (Franke et al. 1995; Pierard et al. 1998; Schmidt et al. 2000).
In India, surveillance data regarding this organism are very sparse. Although demonstrated to be present in the food chain by a number of studies, STEC have not been implicated as a major cause of diarrhoea (Khan et al. 2002; Kumar et al. 2004; Bandyopadhyay et al. 2011). The purposes of this study were to determine the prevalence and distribution of various stx variants in STEC isolated from diverse environmental sources (animal stool, meat) and human illness from a large geographic area in India and to understand any association between variants and human disease. We also investigated whether there was any correlation of subtypes with serotypes and whether these toxin variants could be detected by a commercial immunoassay. This is a first study identifying subtypes of stx variants of STEC causing human illness in India.
Materials and methods
Surveillance for STEC was conducted in the semi-urban and rural areas of three states (Punjab, Himachal and Haryana) surrounding Chandigarh in north India. Animal rearing is commonly practised in these areas. Animal stool samples (550 from cows and buffalos and 100 from sheep and goats) were collected from 51 small and big diaries. A total of 450 meat samples (chicken, mutton and pork) were collected from local abattoirs and slaughterhouses. Human stool samples (n = 600) were collected from every fifth consecutive stool sample of patients presenting with diarrhoea and all cases of bloody diarrhoea and HUS presenting at our centre that were submitted to the Enteric Laboratory, Department of Medical Microbiology, PGIMER, Chandigarh, India. Ten human isolates were obtained from the National Institute of Cholera and Enteric Diseases (NICED), Kolkata, for geographic comparison.
Processing of samples
Two to three loopfuls of human or animal stool samples were directly inoculated for enrichment into 3 ml of EC (enrichment culture) medium (Difco, Sparks, MD, USA) and incubated overnight at 37°C under shaking conditions (200–300 rev min−1). For meat samples, 50 ml of EC medium was transferred into a sterile polythene bag and mixed well with 5–10 g of finely minced raw meat sample. After 2-h incubation at 37°C, the broth was transferred into a sterile conical flask and incubated as described above.
Screening of enrichment broth by polymerase chain reaction
One ml of overnight-incubated EC broth culture was centrifuged at 5000 g for 10 min, and the pellet was resuspended in 200 μl of TE buffer (20 mmol Tris–HCl, 2 mmol EDTA, pH 8·0). The suspension was boiled for 10 min in a water bath to achieve cell lysis. The lysate was centrifuged at 5000 g for 5 min, and the supernatant was removed and used as template in PCR for the detection of stx1 and stx2 genes using the primer sequences listed in Table 1. Escherichia coli strain EDL933 (obtained from Pasteur Institute, Paris, France) was used as a positive control for stx1 and stx2. PCR was performed in a 25-μl reaction volume containing 20 mmol l−1 Tris–HCl (pH 8·3), 50 mmol l−1 KCl, 1·5 mmol l−1 MgCl2, 0·001% gelatin, 1 U Taq polymerase, 25 mmol l−1 dNTPs, 1·5 μl (equivalent to 25 ng) of template DNA, 100 pmol l−1 of forward and reverse primers (Sigma Aldrich, St Louis, MO) for stx1 and stx2 genes on thermocycler. The amplification programme included an initial denaturation step of 94°C for 4 min and 30 cycles each of denaturation (94°C for 1 min), annealing (53°C for 1 min), extension (72°C for 1 min) and final extension of 72°C for 10 min. Agarose gel (1·5%) electrophoresis (4 V cm−1, 50 min and 100 mA) in 0·5X TBE buffer (pH 8·0) using horizontal gel assembly was performed to detect the amplification products. Gels were stained in 1·0 ug ml−1 ethidium bromide solution and visualized under UV light. The images were digitized using a gel documentation system (Alpha Imager™ 3400; Protein Simple, Santa Clara, CA). Broth cultures that were positive for stx1 and/or stx2 were inoculated into Sorbitol MaConkey's agar (SMAC) (Difco) for the isolation of STEC that appeared after an overnight incubation at 37°C. Isolated presumptive E. coli colonies including nonsorbitol-fermenting colonies were picked up randomly from dilution plates and subjected to biochemical analysis for confirmation. The following tests were used: gram stain, oxidase, Huge Leifson's oxidative fermentation test, motility, triple sugar iron (TSI) medium, indole, urease, PPA (phenylamine pyruvic acid), LAO (lysine, ornithine and arginine decarboxylase) tests. Biochemically confirmed E. coli colonies were spot-inoculated on Luria agar master plates into rows and columns, and all colonies from each row and each column were pooled into a fresh EC broth. After an overnight incubation at 37°C, DNA was extracted and PCR performed for stx1 and stx2 as described above. A stx-positive colony was identified as the colony shared by row and column. STEC-positive isolates were inoculated into a brain heart infusion broth containing 15% glycerol and stored at -80°C until further use.
Table 1. Primer sequences for stx1, stx2 and their variants utilized for polymerase chain reaction (PCR) amplification in this study
The E. coli isolates possessing at least one of the above-mentioned genes were serogrouped by the National Salmonella and Escherichia Centre (Central Research Institute, Kasauli, HP, India).
Detection of Shiga toxins by ELISA
The E. coli isolates harbouring stx1 and/or stx2 were screened for the formation of Shiga toxins by using Ridascreen-EIA kit (R-Biopharm AG, Darmstadt, Germany) and following the manufacturer's instructions (http://www.r-biopharm.de/). Noninoculated growth medium was taken as negative control.
Subtyping of stx1 and stx2
STEC isolates were characterized into stx1 and stx2 subtypes. PCR assays were carried out by using primer sequences and published protocols for stx1c, stx1d, stx2c, stx2d, stx2e, stx2f and stx2g genes referenced in Table 1.
The data on prevalence of stx variants from different sources were analysed by chi-square test. A P value < 0·05 was considered to be significant.
A total of 130 (7·1%) samples were positive by STEC PCR with the following break-down: 80 (12·3%) animal stool samples (34 cow, 41 buffalo and 5 sheep samples), 39 (8·4%) meat (only mutton was positive) and 11 (1·8%) human stool samples including one from HUS case (Table 2).
Table 2. Number of samples collected and STEC isolated from different sources
No. samples collected
Positive for STEC (%)
Animal stool (cow, buffalo, sheep/goat)
Meat (chicken, mutton, pork)
Prevalence of stx1 and stx2
PCR assays showed that the stx1 gene was most common among the recovered STEC. It was present in 64 animal stool isolates (80% distributed among 27 cows, 34 buffalos, 2 sheep/goats), 32 meat (82%) and 21 (100%) human isolates including the Kolkata strains. The stx2 gene was present in 59 (73%) animal stool (23 cows, 31 buffalos, 5 sheep/goat), 25 (64%) meat and 10 (47·6%) human stool isolates. Overall, stx1 was found to be significantly more prevalent than stx2 (P < 0·05). Among the human isolates too, stx1 was significantly more prevalent than stx2 (P < 0·05).
Prevalence of stx1 and stx2 variants
Of 117 and 94 isolates positive for stx1 and stx2, variants were present in 34 (29%) and 48 (51%), respectively. Variant stx2c gene was the most frequent at 35 (25·3%) of the isolates followed by stx1d in 19 (13·5%) isolates, stx1c in 15 (10·7%) isolates and stx2d in 13 (9·2%) isolates. Variants stx2e, stx2f and stx2g were not detected in any of the isolates. Table 3 shows the distribution of these variants in different samples. The stx1c and stx2c subtypes were absent in sheep/goat stools, while these subtypes were found in five and three of the 39 mutton meat samples, respectively. The variant stx1d was present in five (34%) cow, seven (17%) buffalo and three (60%) sheep/goat stool isolates and absent in meat isolates. The stx2c gene was found to be significantly more associated with animal stool isolates (P < 0·05). Variant stx2d was present in cattle and sheep/goat stools and mutton samples (Table 3).
Table 3. Prevalence of stx1, stx2 and their variants in STEC isolates isolated from diverse sources
stx1 and stx2 variants (no. (%) of isolates possessing indicated stx variants)
In human stool isolates, subtypes stx1c, stx1d, stx2c and stx2d were present in 3 (15%), 4 (19%), 2 (10%) and 1 (4·7%) isolates, respectively (Table 3). The only isolate from a HUS case was positive for stx1c and stx2 and negative for others. Figure 1 summarizes the different combinations of stx and variants detected. The most common combination of subtypes was stx1, stx2c present in 30 isolates followed by stx2, stx2c (29 isolates); stx1, stx2, stx2c (24 isolates); and combination of stx2, stx1d and stx1, stx1c (14 isolates) each (Fig. 1).
Of the 140 STEC strains isolated in this study, 41 strains were O antigen untypable with somatic (O) antiserum and 17 strains were rough. The most common serogroups were O60, O69, O43 and O85 present in seven isolates each followed by O2, O22, O103, O168 in four isolates; O55, O30, O102, O95 in three isolates; O5, O110, O173, O41, O28 serotypes in two isolates each. Other serotypes were present in one isolate each. Serogroup distribution was found to be diverse, and no association of stx variants with the particular serogroup was observed (Table 4). Serogroup O64 was present only in human isolates, and serogroup O28 was common to human and meat isolates (Table 5).
Table 4. Distribution of serogroups according to stx variants (No. of isolates displaying same serogroup)
Detection of stx1 and stx2 variants by commercial immunoassay
We assessed the formation of toxins by isolates harbouring stx variants using the Ridascreen immunoassay. Of the 140 STEC isolates positive by PCR, only 42 produced toxins. The detection rates were 30·7% for stx1, 31·9% for stx2 and 32·2% for both stx1 and stx2. Among the variants, stx1c-encoded toxin was detected in five (33·3%) isolates, and stx1d-encoded toxin in five isolates (26·3%), stx2c encoded toxin in 15 isolates (42·8%) and stx2d encoded toxin in four (30·7%) isolates (Fig. 2). These differences in detection rates were not found to be significant (P > 0·05).
We performed stx subtyping on 21 human STEC isolates and report the occurrence of variant stx1d for the first time in India. Ten human strains from Kolkata were also included for comparisons. It has been important to study stx1 and stx2 subtypes harboured by STEC strains because they vary in their capacity to cause illness in humans because of the type and amount of toxins produced (Law 2000; Beutin et al. 2007; Scheutz et al. 2009). Toxin Stx2 is 1000 times more cytotoxic than toxin Stx1 towards human renal endothelial cells (Lindgren et al. 1993; Law 2000). Thus, STEC strains producing toxin Stx2 are more commonly associated with HUS than isolates producing Stx1 (Ostroff et al. 1989). Two variants of toxin Stx2, that is, Stx2c and Stx2d (an elastase mucus activatable toxin), are also reported to be associated with high virulence of STEC and cause of HC and HUS (Beutin et al. 2007); other Shiga toxins like Stx1c have been reported to evoke asymptomatic or mild disease (Zhang et al. 2002). Escherichia coli carrying stx1d and stx2g genes have not yet been reported to occur in humans, though, despite having animal reservoirs (Beutin et al. 2007).
The preferred method for stx2 subtyping is PCR-restriction length polymorphism (RFLP). However, this method is vulnerable to single nucleotide changes, and interpretation can be difficult if small or similar size fragments are generated by RFLP. Therefore, for subtyping, we carried out PCR amplification by using allele-specific primers (Zhang et al. 2002; Beutin et al. 2007). In the present study, the stx1 gene was found to be significantly more prevalent than stx2 (83 vs 67·5%) (P < 0·05). Two variants of stx1 (stx1c and stx1d) and five of stx2 (stx2c, stx2d, stx2e, stx2f and stx2g) were detected. The variant stx2c was the most common variant present in 35 (25·1%) isolates, followed by stx1d in 19 (13%) isolates, stx1c in 15 isolates (10·7%) and stx2d present in 13 isolates (9·2%). In our study, variant stx2c was significantly more associated with cattle stool isolates as compared to meat and human STEC (P < 0·05), thus identifying cattle as a reservoir of this potentially pathogenic subtype. Bosilevac and Koohmaraie (2011) found variant stx2c in 77 (22·7%) of 338 isolates from ground beef samples in the USA. Beef is not eaten in this geographic region of India owing to religious beliefs.
Variant stx1c has been described to be commonly found among STEC isolates from sheep stool but seldom among those from cattle stool (Brett et al. 2003). It has also been isolated from asymptomatic or mildly ill humans (Ochoa and Cleary 2003). In our isolates, stx1c was absent in sheep/goat stool isolates, and of three human stool isolates positive for stx1c, one developed HUS. Variant stx1c was also detected in a small number (5/160, 3·1%) of mutton samples. This may be due to unhygienic practices at the site of slaughter leading to contamination from environment and the practice of rearing sheep and goat in the same premises as of cattle. The incidence of isolation of STEC isolates carrying the stx2d gene in the present study was high as compared to the findings of earlier studies from Australia and Brazil where prevalence was 1·4% and 0·9% in cattle stool isolates, respectively (Brett et al. 2003; Cergole-Novella et al. 2006). The mucus activatable toxin Stx2d has been associated with severe human disease and has been described to have limited occurrence in cattle reservoir (Tasara et al. 2008). In the present study, stx2d was present in nine (11·2%) cattle stool isolates and three mutton samples plus one human stool isolate from a case of severe diarrhoea (data not shown).
In this study, only 8/21 (38%) human STEC isolates harboured stx variants, of which stx2c and stx2d were found in two isolates and one isolate, respectively. The low frequency of carriage of these potentially more pathogenic variants may be one of the reasons for low severity of human illness in India, although hitherto unstudied host factors may also play a role. Human STEC isolated in this study included five each from watery and bloody diarrhoea and one from a HUS case following bloody diarrhoea. Up to 83% of the cases of diarrhoea-positive (D+) HUS in children occur after an infection with STEC (Gerber et al. 2002). In our study, although all cases were D+ HUS, STEC were isolated from only one of 10 cases. This low prevalence may be due to late presentation of cases. Shiga toxin–forming E. coli may have disappeared from stool by the time patients presented to our hospital. HUS can also occur as a complication of S. dysenteriae serotype 1 infection in our region (Taneja et al. 2005) but was also not isolated from these cases. In human stool isolates, the stx1c subtype was present in 3 (14·2%), stx1d in 4 (19%), stx2c in 2 (9·5%) and stx2d in 1 (4·7%) STEC isolate (P value = 0·7). Zhang et al. (2002) reported the presence of stx1c gene in 36 non-O157 human STEC strains with one being recovered from a patient with HUS. In the current study, other variants such as stx2e, stx2f and stx2g were not found. The reservoirs of these include feral pigeons for subtype stx2f (Schmidt et al. 2000) and healthy pigs (Caprioli et al. 1993) and piglets with oedema disease (Matise et al. 2003) for stx2e.
Data from India are limited. Kumar et al. (2001) reported stx1c from beef samples for the first time from Mangalore. Wani et al. (2007) from Srinagar reported stx1c in 13 (31·70%), stx2c in 8 (24·2%) and stx2d in 2 (6·06%) STEC isolates, isolated from diarrhoegenic (n = 134) and healthy (n = 53) calves. In another study, Wani et al. (2009) reported stx1c variant in 25 (73·5%) isolates and stx2c in 2 (6·8%) and stx2d in 6 (20·6%) STEC isolated from diarrhoegenic and healthy lambs. Bandyopadhyay et al. (2011) also reported the presence of stx1c (77%), stx2c (32·1), stx2e (10·7%) and stx2f (10·7%) variants among E. coli isolates from yak milk and yak cheese.
Six serotypes such as O26:H11 or NM; O45:H2 or NM; O111:H8 or NM; O103:H2, H11, H25; O121:H19 or H7; and O145:NM have been described by CDC as being to be responsible for 71% of non-O157 STEC disease (Brooks et al. 2005). In this study, serogroup distribution was found to be diverse without any predominant circulating type. The most commonly isolated serogroups were O60, O43, O85 and O69 (Table 4). Four isolates (two cattle, one each from human and mutton) belonged to the O103 serogroup, which is the third most common non-O157 serogroup associated with human illnesses (Bosilevac and Koohmaraie 2011). In the present study, new STEC serogroups reported for the first time in India include O55, O33, O173, O165 and O136. Only one isolate of the serogroup O104 was isolated from a mutton sample. Serogroup 104 was recently described in the largest outbreak of HUS that occurred in Germany (World Health Organization 2011).
Culture supernatants of all STEC isolates were tested for their reactivity in commercially available Stx EIA assay in order to establish toxin production. Of the 140 PCR-positive stx1 and stx2 STEC isolates, only 42 (30%) were positive. This finding is similar to the study by Pulz et al. (2003). As here, they found that a PCR technique that can cover all the variants of stx1 and stx2 is more sensitive than the kit for the detection of STEC. This may be related to the amount of toxin released in the culture, which may be below the detection threshold of the Ridascreen-EIA kit. Other possible explanations for the low detection rate could be that the mutation in stx genes leads to antigenic differences in the toxin molecule, or that synthesized toxins were not released from the bacterial cells because of defects in toxin export mechanisms. The above-mentioned finding has practical implications: detection of toxin genes by PCR indicates a potential risk and may be important for surveillance and public health point of view, whereas toxin detection by EIA may indicate clinical, ‘real-life’ risk. Data from CDC support the high sensitivity and specificity of EIA for the detection of both O157 and non-O157 STEC serotypes from patient stool samples (Marcon 2011). The main limitations of this study were low number of human isolates and the exclusion of asymptomatic populations. No outbreaks of STEC diseases have been reported from India to date. Such may be due to limitations in laboratory diagnosis that requires the use of molecular methods.
In conclusion, this study demonstrates the presence of potentially pathogenic subtypes in the environments that may serve as reservoirs for human infections. The relatively high prevalence of STEC subtype stx2c in animal isolates in this study may pose threat to public health. To the best of our knowledge, this is the first attempt at subtyping human STEC isolates as well as the report of the presence of stx1d in India.
We gratefully acknowledge the assistance of the Director, National Salmonella and Escherichia Centre, Central Research Institute, Kasauli 173204 (India), in serogrouping of the isolates. We would like to acknowledge Dr G.B. Nair and Dr T. Ramamurthy from National Institute of Cholera and Enteric Diseases for providing isolates of STEC from Kolkata. We are grateful to Dr Deepti Gupta, PU, Chandigarh, for contributing to language corrections in the manuscript.
The financial, administrative and technical support of Indian Council of Medical Research (ICMR), New Delhi, India (grant no. 5/8-1(213)-D/2006 ECD-II), is acknowledged.