Rapid genetic typing of diarrheagenic Escherichia coli using a two-tube modified molecular beacon based multiplex real-time PCR assay and its clinical application

Background Diarrheagenic Escherichia coli (DEC), including Enterotoxigenic E.coli (ETEC), Enteroaggregative E.coli (EAEC), Enteropathogenic E.coli (EPEC), Enterohemolysin E.coli (EHEC) and Enteroinvasive E.coli (EIEC) causes diarrhea or hemolytic uremic syndromes among infants and travelers around the world. A rapid, reliable and repeatable method is urgent for identifying DEC so as to provide the reference for responding to diarrheal disease outbreak and the treatment of the diarrheal patients associated with DEC. Methods In this study, specific primers and modified molecular beacon probes of nine specific virulence genes, whose 5′end were added with homo tail sequence, were designed; and a two-tube modified molecular beacon based multiplex real–time PCR (rtPCR) assay for the identification of five Escherichia coli pathotypes, including ETEC, EAEC, EPEC, EHEC and EIEC was developed and optimized. Totally 102 bacterial strains, including 52 reference bacterial strains and 50 clinical strains were detected to confirm whether the target genes selected were specific. Then detection limits of the assay were tested. Lastly, the assay was applied to the detection of 11860 clinical samples to evaluate the specificity and sensitivity of the developed assay compared with the conventional PCR. Results The target genes were 100% specific as assessed on 102 bacterial strains since no cross-reactions were observed. The detection limits ranged from 88 CFU/mL (EHEC) to 880 CFU/mL (EPEC). Compared with the conventional PCR, the specificity and sensitivity of the multiplex rtPCR was 100% and over 99%, respectively. The coefficient of variation (CV) for each target gene ranged from 0.45% to 1.53%. 171 positive clinical samples were mostly identified as ETEC (n = 111, 64.9%) and EPEC (n = 38, 22.2%), which were the dominating pathotypes of DEC strains. Conclusion The developed multiplex rtPCR assay for the identification of DEC was high sensitive and specific and could be applied to the rapid identification of DEC in clinical and public health laboratories.


Background
Diarrheal disease has been considered as one of the most important public health problem, as well as the significant cause leading to children's death [1] and brings high disease burden to society [2]. Diarrheagenic Escherichia coli (DEC) is the main pathogen associated with diarrheal disease.
Based on the epidemiological, clinical features, specific virulence determinants and other molecular characteristics, DEC has been classified into five main pathotypes as follows: Enterotoxigenic E.coli (ETEC), Enteroaggregative E.coli (EAEC), Enteropathogenic E.coli (EPEC), Enterohemolysin E.coli (EHEC) and Enteroinvasive E.coli (EIEC) [2]. Each pathotype possesses specific virulence genes associated with the disease symptoms. For example, shiga-like toxin I (stx1), shiga-like toxin II (stx2), eaeA and escV genes are virulence genes for EHEC. EaeA encodes intimin that involved in attachment of bacteria to enterocytes, and escV is the main component of ring structure on inner membrane. EPEC also contains eaeA and escV, but without shiga-like toxin [3]. For ETEC, it carries the plasmid encoded heat-stable (ST) and/or heat-labile (LT) toxins. ST can be divided into two main groups, stp (also referred as STIa) and sth (also referred as STIb) [4]. Gene ipaH, which is related to invasion [5] expressed in EIEC and shigella spp.. The defining feature of EAEC is the distinctive aggregative-adherence pattern, which is mediated by aggR gene [6].
DEC will mostly result in persistent watery diarrhea, abdominal cramping and diarrhea. EHEC, such as O157: H7 will cause hemolytic uremic syndrome (HUS) and is life-threatening. DEC is also associated with foodborne disease outbreak or infantile diarrheal disease outbreak. So identification of different pathotypes of DEC will provide the critical information for determining appropriate therapies for patients with suspected E.coli infections and controlling of the outbreak. A variety of assays have been published for identification of DEC. Serotyping and biochemistry has been widely applied in the diagnosis of gastrointestinal pathogens, but they cannot be used for conclusive identification of DEC groups [7]. Molecular assays [8][9][10][11][12][13][14][15][16][17][18], like DNA hybridization [8], loop-mediated isothermal amplification (LAMP) [9], multiplex PCR [10,11,15,17,18], and real-time PCR (rtPCR)-SYBR green melting curve [12][13][14] have been described for such identification. Among these techniques, multiplex PCR is one of the most useful diagnostic methods, however, this method needs post gel electrophoresis, which is labor-consuming and can easily result in carry-over contamination. Thus it is not suitable for analyzing large numbers of clinical samples in clinical laboratory. Guion et al. [14] creatively used rtPCR SYBR green meltingcurve to detect eight virulence gene of DEC, and the results could be obtained from the peak of melting temperature(T m ). However, SYBR green melting-curve is not a stable method, since it's prudent to standardize the assay using control strains in each laboratory.
We have developed a multiplex rtPCR assay based multicolor combinational probe coding (MCPC) for the identification of eight foodborne pathogens [19] and applied to the rapid diagnosis of foodborne disease outbreak successfully. In order to quickly screen DEC from large number of stool samples in laboratory-based surveillance network of diarrheal disease and respond to diarrheal disease outbreak associated with DEC, in this paper, based on the principle of MCPC, we attempted to design modified molecular beacons and primers with homo tail sequences and also add internal amplification control to establish the multiplex real-time PCR assay for the identification of DEC. Considering the different modes of the real-time PCR instruments used in the different hospitals, we established a two-tube multiplex rtPCR assay for detecting nine genes simultaneously with the same thermocycling condition and DEC can be typed within 90 min.

Detection limits and reproducibility of multiplex rtPCR
The detection limits of this method for bacterial pure culture ranged from 88 CFU/mL (EHEC) to 880 CFU/mL (EPEC). Amplification efficiencies for bacterial culture ranged from 92.2% (EIEC) to 102.8% (EHEC) ( Figure 1 and Table 1).
The C T values from three replicate experiments for each ten-fold dilution (10 6 cfu/ml -10 3 cfu/ml) differed within one cycle and the coefficient of variation (CV) for each target gene ranged from 0.45% to 1.53%, which indicated that the assay was reproducible ( Table 2).

Specificity of target genes for multiplex rtPCR
All 102 bacterial strains were correctly identified and no cross-reaction or false positives were observed, demonstrating that the selected target genes were specific for their respective pathotypes (Additional file 1: Table S1).

Application of multiplex rtPCR for the identification of suspected DEC samples
Of the 11860 stool samples, 5100 samples were putatively E.coli positive on MacConkey agar and biochemical test. Among 5100 suspected DEC samples, 171 were DEC positive, including 38 ETEC, 111 EPEC, four EHEC, 10 EAEC and eight EIEC identified by multiplex rtPCR assay, whereas 150 were DEC positive including 29 ETEC, 99 EPEC, four EHEC, 10 EAEC and eight EIEC were identified by the conventional PCR.

Discussion
In present study, two-tube multiplex rtPCR was successfully established for identifying five pathotypes of DEC using homo tail primers and modified molecular beacons simultaneously. The detection limits ranged from 88 CFU/ mL to 880 CFU/mL. Compared with the conventional PCR assay, detection data of 11860 clinical samples showed that two-tube multiplex rtPCR possessed 100% sensitivity and above 99% specificity. After optimizing cycling condition of the assay, five categorized DEC were detected at the same time, leading to a great reduction of cost and time. And the time required from DNA extraction to final result was less than 90 min. Therefore the assay was a high sensitive, specific and rapid method.
Traditionally, high throughput rtPCR was restricted by the wavelength of fluorescence channel and primer dimer formation, which may result in simultaneous detection of two or three genes in one reaction. The follows were used as means to overcome these limitations: First, part of the stem structure of probes was designed to participate in hybridization of template to increase the binding affinity and reduce the background of fluorescence signal greatly. Second, to reduce dimer formation, homo-tag assisted non-dimer (HAND) method [20,21] was put into use to ensure the stability and sensitivity of multiplex rtPCR. Also, the additional tag sequence increased the T m value of primer, making the optimization of the same cycling temperature for twotube assay easier. The above methods are able to enlarge  the signal-to-noise ratio, and enhance the efficiency of amplification. Comparing to gel electrophoresis and SYBR green melting curve, the use of fluorescence probes would avoid reopening the reaction tubes, thus greatly reduce the risk for carry-over contamination and false-positive and will also avoid errors from band discrimination when observed by naked eyes.
Comparing to the conventional PCR for DEC identification according to diagnostic criteria for infectious diarrhea (WS 271-2007, China) and other literature data, our assay covers more genes, such as escV, stp and sth. Since eaeA had a lot of single nucleotide polymorphisms, escV was also included for the identification of EPEC. The twelve extra positive samples were identified when escV was added. In addition, only using st gene without differentiation of stp and sth will lead to nine ETEC missing. Statistical analysis revealed that there was no difference among positive result of EAEC, EHEC and EIEC using multiplex rtPCR and the conventional PCR. All these suggested that more genes were preferred to improve the sensitivity of DEC identification method.
In order to indicate false-negative of multiplex rtPCR, an internal amplification control (IAC) was added to the multiplex rtPCR assay. We used the sequence of the human adenovirus (accession No. AY601634) as an internal amplification control to guarantee that proper reagents, such as magnesium, Taq polymerase and 1 × PCR buffer were suitable for PCR amplification.
Considering that clinical DEC strains should be isolated for the further antibiotics susceptibility test, we adapted the multiplex rtPCR to the suspected colonies for the identification of DEC, not directly to the fresh stool samples. Based on the research data of Barletta et al. [12], we used five-colony pool analysis from one clinical sample for typing DEC.
Epidemiological information of out-patients was collected, including patients'age and symptoms (data not shown). Among 171 positive cases, adults were the main infected population instead of children. This may be due to the population structure of Shenzhen, in which young workers constitute a big percentage of population. For DEC cases, ETEC was the predominant type, accounting for 64.9% of all cases. This was corresponded to a new  systematic review [22]. During summer, diarrhea cases caused by DEC increased rapidly (Figure 2), suggesting a high risk of infection in summer and an essential measure to take to promote health care.
In conclusion, a rapid and effective two-tube multiplex rtPCR for the identification of all DEC pathotypes was established. This method was time (90 min) and resources saving (including biochemical reagents, disposable medical supplies) and could be used in clinical and public health laboratories.

Reference bacterial strains and clinical strains
A total of 111 bacterial strains, comprising 24 species, including 61 reference bacterial strains from American Type Culture Collection (ATCC) or National Center for Medical Culture Collections of China (CMCC) and 50 clinical isolates, were collected as specific or nonspecific targets (Additional file 1:  Table S2).

Bacterial culture and DNA extraction
All positive reference bacterial strains were cultured on appropriate agar plates to isolate bacterial colonies. After overnight incubation at 37°C, single colonies of each culture were carefully picked and added into 200 μl sterile purified water. After shaking thoroughly and heating for five minute, slurry was centrifuged at 12000 rpm for one minute, and 5 μl supernatant was collected for PCR reaction.
Clinical stool samples were streaked on MacConkey agar plates. After overnight incubation at 37°C, a pool of five suspicious colonies was picked for DNA extraction as follows. A pool of five suspicious colonies was added to 200 μl sterile purified water and shaked thoroughly following heating lyses for five minute. After centrifuging at 12000 rpm for one minute, 5 μl supernatant was assimilated for PCR reaction.

Primers and probes
Specific virulence genes were chosen to detect specific pathogens: ipaH for EIEC, stp/sth/lt for ETEC, eaeA/ escV for EPEC, stx1/stx2 for EHEC, aggR for EAEC. The sequence of the human adenovirus (accession No. AY601634) was selected as an internal amplification control(IAC) for the multiplex rtPCR assay [23]. All the sequences of genes published in the Genbank were downloaded and aligned by Clustal X 2.0. Conserved regions were picked up for designing primers and probes after merging into GenBank. Based on the articles published online [24], molecular beacon probes were designed with the help of DNA folding form website (http://mfold.rit.albany.edu/?q=mfold/DNA-Folding-Form). All the 5′end of primers were added with a homo tail sequence to alleviate the forming of primer dimer. Molecular beacons were modified and designed. All primers and probes were listed in Table 4.

Performance of multiplex rtPCR
The homo tail was added to 5′ end of primer sequences so that proper annealing temperature will fall into broad range in each individual PCR reaction. Five classes of DEC were divided into two tubes. For tube-one, stp/sth/lt for ETEC, aggR for EAEC and IAC were included, while ipaH for EIEC, eaeA/escV for EPEC, stx1/stx2 for EHEC and IAC were included in tube-two. Multiplex rtPCR reactions were performed on Bio-Rad CFX 96 Real-Time PCR System (Bio-Rad, Hercules, CA). Firstly, individual real-time PCR cycling condition and proper reagent concentration were optimized. Secondly, two-tube system reagents, including primers, probes, magnesium and Taq polymerase are mixed and optimized separately. The optimal assay was as follows: tube-one, 1 × PCR buffer, 3 mM MgCl 2 , 2 μl of deoxynucleoside triphosphate (2.5 mM), and 1 unit Taq polymerase, 0.12 μM to 0.3 μM specific primers and probes (Table 4) and 5 μl of DNA template; tube-two, 1 × PCR buffer, 2.5 mM MgCl 2 , 2 μl of deoxynucleoside triphosphate (2.5 mM), and 1 unit Taq polymerase, 0.2 μM to 0.4 μM specific primers and probes (Table 4)

Detection limits and reproducibility of multiplex rtPCR
Five ATCC reference strains (ETEC ATCC 35401, EAEC ATCC 33780, EIEC ATCC 43893, EHEC ATCC 35150 and EPEC ATCC 43887) were used for the analysis of detection limits. The target strains were cultured in Luria broth and the concentration of the bacteria was adjusted to 10 6 or 10 7 colony forming units (CFU)/mL using a turbidity meter, then 1 mL enrichment broth was used as the dilution, meanwhile an aliquot of each dilution was quantified via plate counting. The DNA template of serial 1:10 dilutions was extracted by heating lyses method mentioned above and was used for the detection limits analysis and four tenfold serial dilutions (10 6 -10 3 cfu/μL) were prepared for the analysis of the reproducibility in three replicate experiments.

Specificity of multiplex rtPCR assay
To confirm whether the target genes selected were specific for each pathotype of DEC, 102 well-characterized strains were tested by multiplex rtPCR.

Application of multiplex rtPCR for the identification of suspected clinical samples
The multiplex rtPCR was applied to detecting clinical samples to determine assay sensitivity and specificity with direct comparison with conventional PCR. All clinical samples were collected based on the inclusion criteria of World Health Organization (WHO) case definition for diarrhea. Diarrheal disease was described as the follows: having three or more loose or liquid stools per day, or as having more stools than normal person. The stool samples of totally 11860 out-patients were collected for DEC detection from 11 sentinel hospitals during 2011 to 2012 in Shenzhen, Guangdong province, China.
All specimens were streaked onto MacConkey agar and a minimum of five typical suspected E.coli colonies were picked and identified by biochemical test and the conventional PCR according to the diagnostic criteria for Table 4 The primers and probes used for multiplex real-time PCR assay in this study a F/R/P: forward and reverse primer, b concentration infectious diarrhea (WS 271-2007, China). Then DNA of a pool of five typical suspected colonies was extracted and performed for both two-tube multiplex rtPCR and conventional PCR simultaneously. For conventional PCR, the reaction volume and amplification condition were as the follows: 5 μl of DNA template was applied to PCR reaction at the final volume of 50 μl; the PCR reaction consisted of 10 mM Tris-HCl (pH 8.3), 50 mM KCl, 0.1% Triton X-100, 1.5 mM MgCl 2 , 2.5 U Taq, 0.2mMdNTP, and 0.125-0.5 μM primer pairs. The cycling condition of PCR contained 95°C for 10 min; and 35 cycles of 94°C for 1 min, 52°C for 1 min, 72°C for 1 min; then 72°C for 10 min. PCR products were analyzed with 1.5% agarose gel electrophoresis. The primer sequences were attached in Additional file 1: Table S2.

Statistical analysis
The results of the multiplex rtPCR assay were compared with conventional PCR using a 2 × 2 table to estimate indices of sensitivity and specificity. Corresponding 95% confidence intervals (CIs) were calculated by using approximate normal distribution and exact interval methods and Kappa value was calculated.