Estimation of the sensitivity of a novel technique for the detection of E. coli O157 in non-heat-treated milk

Estimation of the sensitivity of a novel technique for the detection of E. coli O157 in non-heat-treated milk

Asmaa Elsayed Mohammed^1*

¹ Department of Bacteriology, Animal Health Research Institute, Agriculture Research Center (ARC), Sohag, Egypt.

ABSTRACT

Escherichia coli O157: H7 is an important pathogen that causes fatal infections. It has been the cause of several outbreaks in different areas world-wide and is associated with high mortality and morbidity. Multiple procedures were adopted to detect this serotype, but conventional culture techniques represent the gold standard. In this study we estimated peptide nucleic acid fluorescence in situ hybridization (PNA-FISH) technique for the rapid detection of E. coli O157:H7. PNA-FISH technique was performed for 50 strains; 20 were E. coli O157 H7, 20 were E. coli O157 non-H7 strains, and 10 strains belonging to other Enterobacteriaceae. Sensitivity, specificity, and accuracy of the FISH technique were 100%, 86.67%, and 92% respectively with Confidence Interval of 95% (82.5% to 98.4%). Both PNA-FISH and conventional culture technique were used for examination of 100 non-heat-treated milk samples which were artificially contaminated with E. coli O157:H7. PNA-FISH sensitivity, specificity, and accuracy were 97%, 100%, and 98.5% respectively with a Confidence interval of 95% (95.7% to 99.7%). The PNA-FISH technique had approximate results to the conventional culture methods, in addition, it reduced the diagnosis time to one day.

Keywords: E. coli  O157:H7,  PNA-FISH,  conventional culture method.

INTRODUCTION

1. coli is a Gram-negative, facultatively anaerobic, rod-shaped, and highly motile bacteria. It belongs to the Enterobacteriaceae family and represents a normal inhabitant of the gastrointestinal tract of both humans and animals. Few strains of E. coli can cause diarrhea and some extra-intestinal diseases (Kubitschek 1990).
2. coli can be classified into 6 pathotypic groups according to the mechanism of produced disease; Enteropathogenic E. coli (EPEC), Enterotoxigenic E. coli (ETEC), Enteroinvasive E. coli (EIEC), Diffusely adherent E. coli (DAEC) Enterohaemorrhagic E. coli (EHEC) and Enteroaggregative E. coli (EAEC) (Tassew 2015).

Within the EHEC group, E. coli serotype O157:H7 is the main factor of E. coli food poisoning outbreaks (Karmali et al. 2010 and Segni  et al. 2018).The major reservoir of E. coli O157:H7 is cattle, but the main vehicle for transmission of the organism to human is food (e.g meat, dairy products, and vegetables). Whereas, dairy products have been associated with many outbreaks of E. coli O157:H7 foodborne illness (Trevena 1999).  Clinical presentation of E. coli serotype O157:H7 infection includes colicky abdominal pain and bloody diarrhea, as well as the fatal complication; the hemolytic uremic syndrome (Griffin and Tauxe 1991). Early diagnosis of foodborne pathogens is essential to avoid outbreaks, reduce costs, and achieve ideal food safety management (Weng et al. 2021). The infectious dose of the organism is very low, probably 1 to 10 ingested cells. Therefore, the detection methods must be very sensitive (Jordan and Maher 2006). E.coli is identified by conventional culture procedures based on its inability to ferment sorbitol (Liesegang et al. 2000). The gold standard method for the detection of  E.coli O157:H7  is the conventional bacterial culture, a time-consuming and laborious procedure (Kim et al. 2005). The conventional culture process includes bacterial cultivation, isolation, recovery, and identification procedures which require 3–7 days, which has the disadvantages of depending on personal experience (Sugiarti and Nurhayati 2021 and You et al. 2020). The development of a rapid, sensitive, accurate, and cost-effective technique for detecting foodborne pathogens has therefore been an essential achievement in food safety (Nurmasytha et al. 2021). To attain the sensitivity needed to detect pathogens at low concentrations, signal amplification procedures such as enzyme-linked cascade amplification and nucleic acid amplification were developed and performed for pathogen detection to meet the rising measures for food safety (Zhao et al. 2015).

Fluorescence In Situ Hybridization (FISH) is a molecular procedure extensively utilized for the identification and localization of bacteria in samples (Amann and Fuchs 2008). This technique is based on the specific binding of nucleic acid probes to target RNAs, due to their higher numbers of copies inside the cells. The FISH technique was applied for the detection of Salmonella (Sousa et al. 2024b), E.coli (Sousa et al. 2024a), Helicobacter pylori (Sousa et al. 2023), Campylobacter spp. (Oliveira et al. 2024), and Coliform bacteria (Kuo et al. 2020).

A synthetic DNA analog, named peptide nucleic acid (PNA), capable of hybridizing to complementary nucleic acid targets, was developed and  made FISH technique simple and effective (Wilks and Keevil 2006). PNA-FISH technique was successfully used to detect multiple pathogenic microorganisms in clinical specimens and the food industry (Søgaard et al. 2005).

This work aimed to apply a PNA-FISH technique for the detection of E. coli O157:H7 in non-heat-treated milk samples and to compare its performance with that of the conventional culture technique.

MATERIALS AND METHODS

This work was conducted between March 2023 to November 2023 in the Bacteriology Department, Animal Health Research Institute, Sohag branch and Centre of Biotechnology at Alexandria University.

Bacterial growth and culture media:

This study included 20 E. coli O157:H7 strains, 20 E. coli non-O157:H7 strains, and 10 other strains from the Enterobacteriaceae family. They were obtained from the Centre of Biotechnology at Alexandria University. All strains were maintained on tryptic soy slant agar (Table 1).

PNA probe design:

To identify potentially useful oligonucleotides to use as probes, 16S and 23S rRNA gene sequences available at the Centre for Biotechnology Information were chosen. The suspected regions were chosen by sequence alignment via the ClustalW program, offered by the European Bioinformatics Institute. A conserved region in the 23S rRNA of all E. coli O157:H7 isolates was identified.

The following probe sequence was chosen: 5=-CAA CAC ACA GTG TC-3=. This sequence hybridizes between positions 1169 and 1183 of E. coli O157:H7 strain TW14359 (accession number CP_001368); hence, it was named EcoPNA1169.

Theoretical evaluation of the PNA probe performance:

 After designing the probe, its performance was assessed to define the theoretical values for sensitivity and specificity. Probe Check software, offered by the ARB Silva database was used to evaluate these parameters. For this theoretical assessment, only the high-quality sequences with at least 1,900 bp were considered along with E. coli strain sequences with the selected serotype. The probe was examined against the large-subunit ([LSU]; 23/28S) database and the small-subunit ([SSU]; 16/18S) database.

Hybridization procedure:

The hybridization temperature ranged between 53 and 61°C; the fixation step using ethanol 80%, and multiple hybridization times (30, 45, 60, and 90 min) were examined. Smears of each strain were prepared by standard technique and dipped in 4% (wt/vol) paraformaldehyde, followed by 50% (vol/vol) ethanol for 10 min each, and left in air to dry.

The smears were then covered with 20 µl of hybridization solution containing 10% (wt/vol) dextran sulfate, 10 mM NaCl, 30% (vol/vol) formamide, 0.1% (wt/vol) sodium pyrophosphate, 0.2% (wt/ vol) polyvinylpyrrolidone, 0.2% (wt/vol) Ficoll, 5 mM disodium EDTA, 0.1% (vol/vol) Triton X-100, 50 mM Tris-HCl (pH 7.5;), and 200nMEcoPNA1169 probe.

Specimens were protected by coverslips, put into humid containers, and incubated for 45 min at 59°C. Afterward, the coverslips were taken off and the slides were placed in a pre-heated (59°C) washing solution containing 15 mM NaCl, 1% (vol/vol) Triton X-100, and 5 mM Tris base (pH 10;). Washing was performed at 59°C for 30 min, and the slides were left to air dry. The slides were kept in darkness for 24 hours before being examined under a microscope. The experimental specificity and sensitivity of the probe were assessed (Almeida et al. 2013).

Selection of the milk samples for inclusion in the study:

Non-heat-treated milk samples were subjected to conventional culture methods to select 100 negative samples for E. coli O157:H7 according to ISO 16654:2001 (Tozzoli et al. 2019).

Milk sample inoculation with E. coli O157:H7:

1. A loopful of coli O157:H7 CECT 4267 was added to 20 ml of tryptic soy broth (TSB) and incubated for 18 h at 37°C and 120 rpm in an orbital incubator. Bacterial suspensions were diluted in a phosphate-buffered saline (PBS) solution and adjusted to a concentration of 0.5 McFarland Standard corresponding to approximately 1 ×10⁸ cells/ml. These were further diluted in PBS to obtain 1 ×10² CFU/ml. Cell concentrations were confirmed by plating on tryptic soy agar.
2. Twenty-five ml from each milk sample were mixed with 225 ml of modified tryptic soy broth supplemented with novobiocin.
3. The diluted milk samples were then artificially contaminated with coli O157:H7 at concentrations of 10 CFU/25 ml milk and 1 CFU/25 ml milk, then incubated for18 h at 37°C and 120 rpm in an orbital incubator.

Detection of E. coli O157:H7 in milk samples by conventional culture technique:

• The detection of coli O157:H7 by culture-based methods was done according to ISO 16654:2001 (Tozzoli et al. 2019).

A loopful from the artificially contaminated milk sample was inoculated into sorbitol MacConkey's agar and Eosin Methylene Blue (EMB) then incubated at 40^oc for 24 hrs according to Collee et al. (1996).

Typical colonies of E. coli O157:H7 produce large circular blue-black colonies with a green metallic sheen EMB agar and are colorless colonies on  Sorbitol MacConkey agar (Harrigan and McCance 2014).

• Microscopic examination: the direct film was prepared from pure culture and stained by Gram stain to observe the morphological characteristics of coli O157:H7 which is gram-negative short rods bacilli, that occur singly or in pairs (Cheesbrough 2006).
• Biochemical Confirmation of coli O157:H7: pure cultures of a single colony from sorbitol MacConkey agars were streaked onto a nutrient agar plate then incubated at 37^oc for 24 hrs. Oxidase, Catalase, Indole, Methyl red, Voges-Proskauer (VP), and Citrate tests were done to confirm the presence of E. coli O157:H7 in the test samples (Cerqueira et al. 2008).
• Serological identification of coli O157:H7: Isolates of presumptive E. coli O157:H7 for all biochemical tests were subjected to latex agglutination test (Kok et al. 1996).

Detection of E. coli O157:H7 in milk samples by PNA-FISH:

• 15 µl of each artificially contaminated milk sample were added to 15µl of a Triton X-100 solution (1%) directly on the microscope slides. All samples were dried (about 5 min at 59°C), and then hybridization was performed (Almeida et al. 2013).
• The smears were mounted with one drop of nonfluorescent immersion oil (Merck) and analyzed using an Olympus BX51 (Olympus Portugal SA, Porto, Portugal) epifluorescence microscope with one filter sensitive to the Alexa Fluor 594 molecule attached to the EcoPNA1169 probe (excitation, 530 to 550 nm; barrier, 570 nm; emission long-pass filter, 591 nm). Other filters in the microscope that were unable to detect the EcoPNA1169 probe fluorescent signal were used to verify that cells did not auto-fluoresce.
• For every sample, a negative control was performed simultaneously for which all steps as described were followed, but no probe was included during the hybridization procedure.
• The Olympus CellB software (Olympus Portugal) was used to capture all images at a magnification of ×1,000.

Statistical analysis:

The correlation between FISH and conventional culture method (Correlation coefficient r), Receiver operating characteristic (ROC) curve analysis, test agreement, sensitivity, specificity, 95% confidence intervals, Youden index, and the accuracy were calculated using a statistical software program (MedCalc for Windows, version 22.0.18, Med- Calc Software, Mariakerke, Belgium, https://www.medcalc.org).

Statistical significance was assumed at P<0.05. Inter-rater agreement was quantified by Weighted Kappa (K), interpreted as follows: < 0.20 poor; 0.21–0.40 fair; 0.41–0.60 moderate; 0.61–0.80 good; 0.81–1.00 very good (Altman 1990).

RESULTS

To assess the probe's experimental specificity and sensitivity, the PNA-FISH technique was carried out on 50 strains (Table 1); 20 strains were E. coli O157 H7, other 20 strains were E. coli O157 non-H7 strains and 10 strains were belonging to other Enterobacteriaceae. Results showed that the hybridization only occurs with E. coli O157:H7. On the other hand, two E. coli O55 strains (E. coli CECT 730, E. coli CECT 731) and two O157 non-H7 strains (E. coli CCC-18-12, E. coli CCC-26-12) (both tested negative for H7 antigen) were detected. All examined Enterobacteriaceae strains other than E coli were not detected by the FISH method (Table 1); therefore, sensitivity, specificity, and accuracy of the FISH method were 100%, 86.67%, and 92% respectively with Confidence Interval of 95% (82.5% to 98.4%) (Table 2) (Figure 1).

PNA-FISH method probe was tested in one hundred non-heat-treated milk samples artificially contaminated with E. coli O157:H7 CECT 4267 with concentration of 10 CFU/25 ml milk and 1 CFU/25 ml milk then examined on epifluorescence microscope (Figure 2)

PNA-FISH results were compared with conventional culture method as a gold standard to determine the sensitivity, specificity, and accuracy values.

The positive samples by the conventional culture method were 100 (100%) while FISH detected 97 (97%) samples and failed to detect three samples (Table 3).

Based on these results, PNA-FISH sensitivity, specificity, and accuracy were 97%, 100%, and 98.5% respectively with a Confidence interval of 95% (95.7% to 99.7%) (Table 4) (Figure 3).

Weighted Kappa 0.97 indicated that there was a very good agreement between the conventional culture method and the FISH technique.   

Table 1: Fluorescence In Situ Hybridization test on examined bacterial strains and species.

Table 2: Statistical analysis of FISH on examined bacterial strains and species.

Table 3: Comparison between FISH test and conventional culture method for detection of E coli O157:H7 in artificially contaminated milk samples.

Table 4: Statistical analysis of FISH in comparison to conventional culture method.

Figure 1: Sensitivity, specificity, and AUC of FISH on examined bacterial strains and species.

Figure 2: PN A-FISH result for non-heat-treated milk samples artificially contaminated with 10 CFU/25 g of E. coli O157:H7 CECT 4267.

Visualization of the E. coli O157:H7 microscopic field in the red channel, appeared as red fluorescent on black background by Olympus BX51 epifluorescence microscope with filter sensitive to the Alexa Fluor 594 molecule attached to the EcoPNA1169 probe.

Figure 3: Sensitivity, specificity, and AUC of FISH in comparison to conventional culture method.

DISCUSSION

PNA-FISH is known as a reliable microbial identification and detection technique (Cerqueira et al. 2008). PNA-FISH targets rRNA sequences, which are universal phylogenetic marks, EcoPNA1169 is combined with a probe targeting a specific sequence of the EHEC group. It can assess the desired detection limit of 1 CFU (Spano et al. 2005).

PNA-FISH has high sensitivity, high stability, cell visualization capacity, safety, short detection time, and multiple color labeling ability (Wagner and Haider 2012).

Oliveira et al. (2024) reported that PNA-FISH is a promising alternative for detecting Campylobacter spp. in food samples. The PNA-FISH technique presented sensitivity and specificity values of 92.0% and 96.9%, respectively Compared to the conventional culture method.

reported that FISH had a short detection time and a high accuracy in the identification of coliform bacteria in simulated water and domestic wastewater samples (Kuo et al. 2020).

Almeida et al. (2010) used the FISH technique a peptide nucleic acid (PNA) probe for the rapid detection of Salmonella spp. using. The probe's theoretical sensitivity and specificity were 100%. They recommended that this procedure be used as an alternative to culture-based techniques.

In the current study, FISH represented a rapid, accurate technique for the detection E. coli O157:H7. The sensitivity, specificity, and accuracy of the FISH method were 100%, 86.67%, and 92% respectively with a Confidence interval of 95% (82.5% to 98.4%) when compared to reference strains.

This study found that FISH is highly sensitive for the detection of E. coli O157: H7. On the other hand, two E. coli O55 strains and two O157 non-H7 strains (negative for H7 antigen) were detected. The match with two O157 non-H7 strains indicates that the technique is specific for O157 despite the H7 and toxin presence. This is an advantage for the identification of other E. coli (EHEC) O157 non-H7 strains (Alpers et al. 2009).

The match of two E. coli O55:H7 isolates may be a weak point. Similarly, PCR and enzyme-linked immunosorbent assay (ELISA) methods designed to identify O157 strains have also shown cross-hybridization with O55 strains (Arthur et al. 2005).  This may occur because O55:H7 is the serotype most closely related to O157:H7 (Zhou et al. 2010). Both serotypes show the locus of enterocyte effacement (LEE) island that causes diarrhea through an attachment-effacement mechanism (Garmendia et al. 2005). E. coli O157:H7 is thought to be originated from E. coli O55:H7 that, during evolution, acquired bacteriophage encoding Stx2 and/or Stx1 toxins (Nataro and Kaper 1998).

In this study, the PNA-FISH method probe was examined in one hundred non-heat-treated milk samples artificially contaminated with E. coli O157:H7 CECT 4267 in comparison to the conventional culture method with a concentration of 10 CFU/25 ml milk and 1 CFU/25 ml milk.

Conventional culture technique detected E. coli O157:H7 in milk samples with a percentage of 100% while FISH detected 97 (97%) samples and failed to detect three samples.

PNA-FISH sensitivity, specificity, and accuracy were 97%, 100%, and 98.5% respectively with a Confidence interval of 95% (95.7% to 99.7%). Also, there was a very good agreement between the conventional culture method and the FISH technique.

Although conventional culture technique is more precise for the detection of E. coli O157:H7, it is a time-consuming method, which often extends to 3–7 days, requiring professional staff and a complicated operation process to detect the pathogen.

FISH technique can detect E. coli O157:H7 in one day, so it is recommended to be used as an alternative to culture-based techniques.

The pre-enrichment step is also known as a critical step in different microbial detection techniques, mainly due to the low count of the desired bacteria, a high level of other competing microorganisms, and the limitations of the detection technique (Vimont et al. 2006). It is crucial to carefully optimize this step to attain high sensitivity values.

1. coli O157:H7 usually present in low contamination levels in the food samples (Garmendia et al. 2005), an enrichment step is recommended. This enrichment step can be done using many types of culture media, from complex rich media (such as TSB or BPW) to selective media, such as Gram-negative (GN) broth, or E. coli (EC) broth (Vimont et al. 2006).

TSB is known as the most used enrichment broth. Additionally, antibiotics like novobiocin (the most commonly used), cefixime, cefsulodin, and vancomycin, as well as other selective materials e.g., bile salts to inhibit the non-Enterobacteriaceae strains (Vimont et al. 2006).

The pre-enrichment step allows FISH to detect E. coli O157:H7 with the detection limit of 1 CFU/25 ml milk (Oliveira et al. 2024).

Regarding the enrichment temperature, it appears that the incubation temperature is not related to the type of serogroup (Vimont et al. 2006). However, some authors demonstrated that O157:H7 strains typically have an ideal temperature of approximately 40°C. This implies that adjusting temperature can control the existing microorganisms and support the growth of E. coli O157. Actually, the ISO recommended a pre-enrichment step in mTSBN at 41.5°C for O157 detection in food samples (ISO 16654:2001) (Tozzoli et al. 2019).

CONCLUSION

EcoPNA1169 is highly sensitive and specific to E. coli O157:H7 strains; however, some cross-hybridization can occur with the closely related O55:H7 serotype. PNA-FISH method can detect pathogen with concentration of 1 CFU/25 ml of non- heat-treated milk samples. Utilizing the PNA-FISH technique can reduce detection time by a minimum of 3 days in detection of E. coli O157:H7 compared to the conventional method. Finally, a comparison of results to those of the conventional culture method has shown high and sensitivity, specificity and accuracy 97%, 100%, and 98.5% respectively with a Confidence interval of 95% (95.7% to 99.7%).