HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via CrossRef Citing Articles via Google Scholar Google Scholar Articles by Iijima, Y. Articles by Hayashi, K. Search for Related Content PubMed PubMed Citation Articles by Iijima, Y. Articles by Hayashi, K. Agricola Articles by Iijima, Y. Articles by Hayashi, K.

Improvement in the detection rate of diarrhoeagenic bacteria in human stool specimens by a rapid real-time PCR assay

¹Department of Microbiology, Kobe Institute of Health, Minatojima-nakamachi, Chuo-ku, Kobe 650-0046, Japan ²Department of Microbiology, Tenri Hospital, Mishima-cho, Tenri 632-8552, Japan

Correspondence Yoshio Iijima iijima{at}oak.ocn.ne.jp

Received January 21, 2004
Accepted March 26, 2004

A rapid laboratory system has been developed and evaluated that can simultaneously identify major diarrhoeagenic bacteria, including Salmonella enterica, Vibrio parahaemolyticus, Campylobacter jejuni and Shiga toxin-producing Escherichia coli, in stool specimens by real-time PCR. Specific identification was achieved by using selective TaqMan probes, detecting two targets in each pathogen. A positive result was scored only when both targets of a pathogen were amplified and the difference between threshold cycles for detection was less than five. Diagnosis of enteric bacterial infections using this highly sensitive method, including DNA extraction and real-time PCR, requires only 3 h. Forty stool specimens related to suspected food poisoning outbreaks were analysed: 16 (40 %) of these samples were found to be positive for diarrhoeagenic bacteria using a conventional culture method; 28 (70 %) were positive using the real-time PCR assay. Of the 12 PCR-positive but culture-negative cases, 11 patients had consumed pathogen-contaminated or high-risk food. Analysis of faecal samples from 105 outpatients who complained of diarrhoea and/or abdominal pain identified 19 (18 %) patients as being positive for diarrhoeagenic bacteria using the culture method. An additional six (6 %) patients were found to be positive by PCR analysis.

	INTRODUCTION

TOP INTRODUCTION METHODS RESULTS AND DISCUSSION ACKNOWLEDGEMENTS REFERENCES

 
Bacterial culture has been the ‘gold standard’ for the identification of the causative agent(s) of diarrhoea from stool specimens. However, this type of analysis usually requires 3–5 days. Recently, more rapid approaches for the direct identification of diarrhoeagenic bacteria in stool specimens have been developed using PCR-based methods (Cohen et al., 1996; Collins et al., 2001; Gentry-Weeks et al., 2002; Lawson et al., 1999; Lin & Tsen, 1999; Linton et al., 1997; Logan et al., 2001; Ramotar et al., 1995). Unfortunately, most of these PCR-based methods can only identify one pathogen, and its close relatives, at a time. Detection of multiple pathogens simultaneously is limited by the differences in optimal PCR conditions for each pathogen-specific primer set. To solve this problem, we have developed a rapid real-time PCR method that can identify several diarrhoeagenic bacteria in stool specimens using the same PCR conditions. Since the most prevalent cause of diarrhoea in Japan has been Salmonella enterica, followed by Vibrio parahaemolyticus, Campylobacter jejuni and Shiga toxin-producing Escherichia coli (STEC; Infectious Disease Surveillance Center, 2004), we have designed at least two sets of primers and corresponding fluorogenic TaqMan probes to detect each pathogen. These sets of primers and probes were evaluated using bacterial strains and human stool specimens.

	METHODS

TOP INTRODUCTION METHODS RESULTS AND DISCUSSION ACKNOWLEDGEMENTS REFERENCES

 
Bacterial strains.

The strains used in this study are listed in Table 1. Control strains of S. enterica and V. parahaemolyticus were obtained from National Institutes of Health, USA and Research Institute for Microbial Diseases (RIMD), Osaka University, Japan, respectively. Penner serotype strains and a strain obtained from the Japan Collection of Microorganisms were used as C. jejuni control strains. Control strains of STEC were obtained from the American Type Culture Collection and RIMD. Clinical strains of V. parahaemolyticus, C. jejuni and STEC were isolated from stool specimens at the Kobe Institute of Health. Clinical strains of S. enterica were isolated from blood and faecal specimens at the Kenya Medical Research Institute, Kenya. We also used bacterial strains from the Kobe Institute of Health that were isolated from liquid egg, retailed chicken, seafood, beef and food remnants associated with food poisoning outbreaks.

Bacteria grown on Luria–Bertani (LB) agar or Columbia blood agar with Campylobacter selective supplement were suspended at concentrations of approximately 1 x 10⁹ ml^–1 in Milli-Q water and boiled at 100 °C for 5 min. After centrifugation at 10 000 g for 5 min, the supernatants were diluted 10-fold with Milli-Q water and applied to the real-time PCR assay.

Stool specimens.

Stools were obtained from 16 healthy people and 40 patients who were linked to a total of 17 incidents of suspected food poisoning reported to the Kobe City Office from May to October 2002. We also obtained stool specimens from 105 outpatients receiving treatment at Tenri Hospital from June 2002 to April 2003 for diarrhoea and/or abdominal pain. All stool specimens were cultured for enteric pathogens by a standard method (Murray et al., 1999). Enrichment broth was used for S. enterica. Surpluses of stool specimens were frozen at –20 °C prior to DNA extraction. We also used frozen faecal specimens obtained from two patients infected with E. coli O157 in 1996. DNA was extracted from approximately 0.2 g of the stool specimens using the QIAamp DNA stool mini kit (Qiagen) according to the manufacturer's instructions, selecting the option to incubate the specimens in the lysis buffer at 70 °C. DNA extracts were kept at –30 °C until use.

Real-time PCR.

Primers and TaqMan probes were designed for two regions of the invA gene of S. enterica, the toxR and tdh genes of V. parahaemolyticus, the yphC and gyrA genes of C. jejuni and the stx1, stx2 and eaeA genes of STEC (Table 2). All probes were labelled with a reporter dye, 6-carboxyfluorescein (FAM), and a quencher dye, 6-carboxytetramethylrhodamine (TAMRA). Aliquots of 4.0 µl DNA extract were mixed with 12.5 µl 2x TaqMan universal PCR master mix (Applied Biosystems) and 8.5 µl of a set of primers and probe, resulting in final concentrations of 300 nM primers, 250 nM fluorogenic probe and 0.1 µg BSA µl^–1 in each well of a 96-well optical reaction plate.

After uracil N-glycosylase treatment at 50 °C for 2 min to prevent amplification of carry-over PCR products, followed by the activation of polymerase at 95 °C for 10 min, DNA was amplified for 45 cycles at 95 °C for 15 s and 60 °C for 1 min using an ABI PRISM 7900HT cycler. Amplification was detected by scanning the fluorogenic reporter dye signal resulting from the cleavage of the TaqMan probes by the 5'–3' exonuclease activity of Taq DNA polymerase. Based on this signal, the threshold cycle (C_t) value was determined. The stool specimen was considered to be positive when both pathogen-specific target sequences were amplified and the difference in C_t values was less than five. In each PCR run, Milli-Q water and bacterial DNA obtained from each of the pathogens served as negative and positive controls, respectively.

PCR inhibitors.

To confirm presence or absence of inhibitors in DNA extracts obtained from stool specimens, a positive amplification control was attempted for all DNA extracts. Approximately 1 x 10⁵ genome equivalents of DNA from S. enterica serovar Enteritidis strain KEMRI-013 was added to the reaction mixture containing 4.0 µl of each DNA extract and amplified using the same conditions as described above, with the invA-1 primer set and probe. When amplification was not seen or the C_t value was more than 40, the DNA extract was considered to contain PCR inhibitors.

Detection limit.

S. enterica strain KEMRI-013 grown on LB agar was suspended in Milli-Q and serially diluted up to 100 000-fold. Aliquots of 20 µl of diluted suspensions were plated on LB agar to count the number of bacteria and mixed with 215 ± 15 mg of stool of a healthy adult. DNA was extracted from these stool samples containing various numbers of the intact cells by the same procedure, and analysed by the same PCR conditions as described above, using the invA-1 primers and probe.

	RESULTS AND DISCUSSION

TOP INTRODUCTION METHODS RESULTS AND DISCUSSION ACKNOWLEDGEMENTS REFERENCES

 
The sensitivity of the PCR assay using the primers and probes designed in this study was 100 % for S. enterica and V. parahaemolyticus strains and 96 % for C. jejuni strains (Table 2). Although the sensitivity for stx1 and eaeA of STEC was 100 %, that for stx2 was only 89 %, which may be related to the genetic diversity of the stx2 gene (Asakura et al., 2001).

Conventional analysis of a single stool specimen from each of 40 symptomatic patients associated with suspected food poisoning outbreaks at Kobe showed that 16 (40 %) specimens were positive for S. enterica, V. parahaemolyticus or C. jejuni (Table 3). No STEC was isolated, although an enteropathogenic E. coli was identified in a single stool specimen by a serological method. Clostridium difficile was isolated from one sample. Enteropathogenic E. coli- and C. difficile-positive samples were considered as negative in this study because the assay utilized did not target these pathogens. Among the 24 culture-negative specimens, 12 were PCR-positive for both regions of invA of S. enterica, toxR and tdh of V. parahaemolyticus or yphC and gyrA of C. jejuni. Among these additional 12 positive patients, five patients were directly linked to food poisoning outbreaks (incidents K-1, K-5 and K-12) and six patients had consumed raw or undercooked chicken meat (incidents K-2, K-3, K-9, K-10 and K-11). Given that S. enterica and C. jejuni were detected in 9.5 and 96 %, respectively, of chicken meat samples retailed in Japan (Doi et al., 2003; Ono et al., 2003), it is reasonable to infer that the six patients who had consumed raw or undercooked chicken were infected with S. enterica or C. jejuni and harboured these bacteria in their stools, as was evidenced by the real-time PCR result. The most probable number of S. enterica in chicken retailed in Japan was less than 10 g^–1 (Doi et al., 2003) and that of C. jejuni in 94 % of chicken was less than 10 g^–1 (Ono et al., 2003). Taken together with the detection limit of the real-time PCR assay, which was 10⁴ c.f.u. (g stool specimen)^–1 (discussed later), it was unlikely that these PCR signals could come from individuals who had consumed S. enterica or C. jejuni cells killed by cooking. All stool specimens from the 16 healthy adults were negative by both the culture and PCR methods. These observations support the conclusion that PCR-positive culture-negative samples were truly pathogen-positive.

We also analysed faecal specimens from 105 outpatients with diarrhoea and/or abdominal pain. By conventional culture, 19 (18 %) specimens were found to be positive for S. enterica, V. parahaemolyticus or C. jejuni (Table 4). Enteropathogenic E. coli was identified in 12 stool specimens and Clostridium difficile was isolated from one stool specimen. As described above, these samples were considered to be negative. Of the 86 culture-negative specimens, six were found to be PCR-positive for diarrhoeagenic bacteria (two S. enterica, two C. jejuni and two V. parahaemolyticus). In contrast, three specimens were C. jejuni-positive by culture but C. jejuni-negative by PCR. In summary, the sensitivity of the PCR assay for all stool specimens tested in this study was as follows: S. enterica, 100 %; V. parahaemolyticus, 100 %; C. jejuni, 84 %. Stool specimens obtained from two patients infected with STEC were recognized by the stx1, stx2 and eaeA primers and probes.

To eliminate false-positive results due to DNA contamination, DNA was extracted again from all of the 18 PCR-positive but culture-negative stool specimens and re-examined by the PCR method. Although C_t values were slightly different from the original assays, identical qualitative results were obtained for all specimens on retesting.

When DNA extracts from samples containing varied species of organisms are amplified, non-specific amplification is unavoidable (Gentry-Weeks et al., 2002). Although we employed highly specific probes, some non-specific signals were detected during analysis of the 163 stool specimens (40 suspected food poisoning, 16 health control, 105 outpatients and 2 STEC-infected patients) by each set of primers and probe at the following frequencies: invA-1, 0.7 % (1/146); invA-2, 1.4 % (2/146); yphC, 2.2 % (3/137); gyrA 2.9 % (4/137); toxR, 0 % (0/153); tdh, 4.6 % (7/153); stx1, 0.6 % (1/161); stx2, 36 % (58/161); eaeA, 17 % (28/161: some of the 28 might be truly eaeA positive). However, false positives were eliminated as the sample was considered positive only when both targets of a pathogen were amplified and the difference in C_t values was less than five.

PCR-positive but culture-negative results may be the result of a non-culturable state of the organism in selective media, including dead and damaged bacterial cells. Our results, which reveal a significant proportion of PCR-positive but culture-negative specimens, were consistent with many other studies (Gentry-Weeks et al., 2002; Kulkarni et al., 2002; Lawson et al., 1999; Logan et al., 2001) and, taken together with the epidemiological data presented in this study, support the view that PCR-based testing is more sensitive than conventional culture for the detection of enteric bacterial pathogens in stool samples obtained from symptomatic patients. The PCR-negative result obtained with three specimens that were culture-positive for C. jejuni may be due to sequence diversity among C. jejuni isolates.

We did not obtain faecal specimens from patients infected with STEC; instead, we analysed previously stored stool samples obtained from two patients infected with STEC. Although the three sets of STEC primers and probes recognized the correct target sequences, further improvement of primers and probes targeting stx2 and eaeA is necessary because of their high non-specific signal detection rates of 36 and 17 %, respectively, associated with analysis of stool specimens.

PCR is sometimes hindered by inhibitors in stool specimens, including bile acids, bilirubins, haem and complex carbohydrates (Wilson, 1997). Of the 163 stool specimens tested, added S. enterica DNA template was not amplified in DNA extracted using the commercially available kit from five (3 %) specimens. Subsequent experiments confirmed that the inhibitors could be removed by phenol/chloroform extraction followed by ethanol precipitation. Retesting of the five inhibitor-free samples led to identification of an additional stool specimen containing C. jejuni.

The detection limit of the real-time PCR method was found to be 10⁴ c.f.u. (g stool)^–1 based on results obtained by mixing S. enterica with a culture-negative stool specimen at different concentrations. If it was assumed that the DNA recovery rate in the extraction process was 100 %, it can be estimated that approximately 40 copies of the target gene(s) per reaction well were detectable in the real-time assay described.

There are other rapid methods to identify pathogens in stool specimens, such as the ELISA test for Shiga toxin (Kehl et al., 1997), but the number of detectable pathogens is limited. Takeshi et al. (1997) have reported PCR detection of four genera of enteropathogens in 24 bloody stool specimens. However, sensitivity and specificity was not studied and their protocol required almost 6 h. Recently, Fukushima et al. (2003) reported PCR assays using melting point analysis of PCR amplicons for the detection of 17 distinct food- or waterborne pathogens. However, only enteropathogenic and enteroaggregative E. coli were detected from human stool specimens in that study.

Our method requires approximately 3 h, including 1 h for DNA extraction from the stool specimens and 2 h for the real-time PCR assay. It detected pathogens in culture- negative stool specimens derived from patients who had been linked to a food poisoning outbreak. Hence, we propose that this PCR-based method contributes to improved rapid diagnosis of enteric bacterial infections, whilst yielding higher detection rates of causative agents. Importantly, the number of identifiable pathogens detectable through such an assay could easily be increased by designing additional primers and probes, as long as the primers and probes have melting temperatures around 59 and 69 °C, respectively.

	ACKNOWLEDGEMENTS

TOP INTRODUCTION METHODS RESULTS AND DISCUSSION ACKNOWLEDGEMENTS REFERENCES

 
We thank the laboratory staff in both facilities for their assistance in culturing samples. This work was supported by a grant for International Health Cooperation Research (12A-1) from the Ministry of Health, Labour and Welfare, Japan.

	REFERENCES

TOP INTRODUCTION METHODS RESULTS AND DISCUSSION ACKNOWLEDGEMENTS REFERENCES