A colony blot immune assay to identify enteroinvasive Escherichia coli and Shigella in stool samples

Diagnostic Microbiology and Infectious Disease 45 (2003) 165–171

www.elsevier.com/locate/diagmicrobio

A colony blot immune assay to identify enteroinvasive Escherichia coli and Shigella in stool samples Do´ra Szaka´la, Gyo¨rgy Schneidera, Tibor Pa´la,b,c,* a

Department of Medical Microbiology and Immunology, University of Pe´cs, Hungary b Department of Medical Microbiology, University of Kuwait, Kuwait c Department of Medical Microbiology and Immunology, United Arab Emirates University, Al Ain, United Arab Emirates Received 2 July 2002; accepted 10 October 2002

Abstract Using an IpaC protein-specific monoclonal antibody a colony blot immune assay was developed for the identification of enteroinvasive Escherichia coli (EIEC) and Shigella in fecal specimens, and was evaluated in a field study. By screening the entire culture plates the colony blot assay was significantly more sensitive than the investigation of 16 randomly selected colonies from artificially contaminated fecal specimens. Among the 165 stool samples from 121 patients with diarrhea the immune assay detected IpaC expressing colonies in 16 out of the 17 specimens positive with a Shigella-, and EIEC-specific polymerase chain reaction targeting the ipaH gene. Guided by the colony blots, Shigella was isolated from 12, while EIEC from four of the samples. The IpaC-specific colony blot immune assay is a simple screening method to detect EIEC in stool samples for laboratories not equipped with molecular techniques. © 2003 Elsevier Science Inc. All rights reserved.

1. Introduction Bacillary dysentery is an acute ulcerative infection of the human large intestine. The infective organisms multiply within, and destroy the epithelial cells of the colon leading to a vigorous local inflammatory response and to the development of superficial ulcers (DuPont, 1990). The disease is endemic in most developing countries, but it also occurs as sporadic cases, or occasional outbreaks of varying magnitude in the industrialized parts of the world (DuPont, 1990; Echeverria et al., 1991). The majority of infections is caused by the members of the Shigella genus, i.e., by S. dysenteriae, S. flexneri, S. boydii, and S. sonnei, respectively. Several studies have demonstrated that, depending on the geographical area and the diagnostic method used, for up to 10% of dysentery cases a group of Escherichia coli strains, called enteroinvasive E. coli (EIEC), are responsible (Abuxapgui et al., 1999; Beutin et al., 1997; Echeverria et al., 1989; Echeverria et al., 1992; Ke´tyi, 1989; Tamura et al., 1996; Wanger et al., 1988). Both Shigella and EIEC strains carry a large invasion plasmid, and express a similar set of * Corresponding author. Tel.: ⫹971-3-7039-480; fax: ⫹971-3-7671966. E-mail address: [email protected] (T. Pa´l).

proteins, some of them with already determined functions in the infective process (Parsot, 1994). The isolation and identification of Shigella strains is a straightforward, well-established procedure (Echeverria et al., 1991). On the other hand, the recognition of EIEC is hindered by the same difficulties associated with the identification of other groups of diarrheagenic E. coli (Echeverria et al., 1991; Nataro and Kaper, 1998). The lack of easily detectable markers differentiating pathogenic strains (e.g., EIEC) from their non-pathogenic counterparts present in the fecal flora, and the complexity of virulence assays using animal or tissue culture models prevent several laboratories from searching for these pathogens. In the past decades, the description of genes coding for various virulence factors has opened new pathways for the identification of pathogenic E. coli. As the most frequently followed diagnostic approach, a limited number of randomly selected E. coli colonies (usually three to 10) are subjected to DNA hybridization or to PCR targeting pathogen-specific sequences (Boileau et al., 1984; Echeverria et al., 1991; Nataro and Kaper, 1998; Taylor et al., 1986). The success of this approach depends on the ratio of the pathogen in the culture, and is limited by practical considerations, i.e., by the number of colonies that can be processed within technical and budgetary limits (Echeverria et al., 1991;

0732-8893/03/$ – see front matter © 2003 Elsevier Science Inc. All rights reserved. PII: S 0 7 3 2 - 8 8 9 3 ( 0 2 ) 0 0 5 1 2 - 6

166

D. Szaka´ l et al. / Diagnostic Microbiology and Infectious Disease 45 (2003) 165–171

Nataro and Kaper, 1998). Alternatively, stool extracts, or that of mixed fecal cultures can be submitted directly to DNA hybridization, or to PCR (Echeverria et al., 1991; Frankel et al., 1990; Nataro and Kaper, 1998). In these cases, however, pathogen-containing specimens, rather than the pathogenic strain itself, are identified. Unless followed by the time consuming and expensive re-testing of individual colonies from positive samples, this latter approach does not lead to the actual isolation of the pathogen. Although highly sensitive and specific, the expenses and the methodological sophistication required by these techniques have been limiting their widespread use in countries facing the majority of enteric infections. Consequently, our understanding of the epidemiology and worldwide distribution of EIEC infections is still limited. Due to their technical simplicity methods based on the immune detection of virulence-specific antigens could provide and alternative to DNA based techniques for these laboratories. Indeed, several groups of diarrheagenic E. coli frequently secrete toxins and express antigens that could be targeted with specific antibodies (for a recent review see Nataro and Kaper, 1998). Recently, we have developed and successfully used in a field trial an ELISA detecting the IpaC antigen, a plasmid-coded invasion protein, to identify EIEC and Shigella strains (Pa´ l et al., 1997). However, the sensitivity of this diagnostic approach is also limited by the number of colonies chosen from individual specimens for the assay. Colony immunoblots combine the simplicity of immune assays with the increased sensitivity of testing large number of colonies for pathogenic E. coli. Indeed, the technique was successfully used, and compared to DNA hybridization, to detect Shiga toxin producing E. coli in stool samples (Hull et al., 1993). More recently, EIEC and Shigella were shown to express the IpaC antigen in amount sufficient to allow their immune detection with this technique when filtered from contaminated water samples (Szaka´ l et al., 2001). These previous data suggested that the development of a colony blot immune assay to screen for IpaC-expressing colonies on fecal culture plates should be attainable. Such a method was expected to facilitate and simplify the identification of pathogens causing dysentery, particularly that of EIEC. Here we report on the development and optimization of an IpaC-specific colony blot immune assay for fecal specimens, and also on its successful application in a field trial identifying EIEC and Shigella strains.

2. Materials and methods 2.1. Strains and culture conditions Eighty virulent EIEC strains of serogroups O28, O29, O112, O121, O124, O136, O143, O152, O164, O167, O171, respectively and 59 virulent Shigella strains (33 Shigella flexneri, 18 Shigella sonnei, 3 Shigella boydii and 5

Shigella dysenteriae isolates) were received from M. B. Martinez (Brasil), C. Sasakawa (Japan), S. Matsushita (Japan) and P. Echeverria (Thailand) and were partly from our strain collection. A further 100 strains of Salmonella, Klebsiella, Enterobacter, enterotoxigenic-, enterohaemorrhagic-, enteropathogenic E. coli, enterococcus, Staphylococcus, Proteus, Pseudomonas and the E. coli K-12 strain J53 were from our strain collection. In order to ensure their invasive character, in all experiments, EIEC and Shigella colonies pigmented on Congo Red agar (Maurelli et al., 1984) were used, only. For experiments to optimize the colony blot assay, and for artificial contamination of fecal samples strains were grown in Tryptic Soy Broth (TSB, Scharlau Chemie) overnight at 37°C on a rotary shaker (120 rpm). To test the specificity of the immune assay blots were prepared from plates inoculated with cultures of non-enteroinvasive strains, or with invasive strains mixed with E. coli J53 at different ratios. In order to prevent the overgrowth of colonies by swarming Proteus strains present in fecal specimens, all blots were prepared from Tryptic Soy Agar plates (TSA) made from TSB with 3% agar-agar added (Szaka´ l et al., 2001). 2.2. Colony blot assay Three different blotting protocols were compared. In the replica-plating technique TSA plates were flood-inoculated with 500 ␮L aliquots of serially diluted samples and incubated overnight at 37°C. Next day, a 82 mm nitrocellulose membrane (0.45 ␮m, Sartorius) laid over a fresh TSA plate was replica-inoculated with a sterile velvet-disk from a master plate containing approximately 500 colonies, and incubated at 37°C, overnight (Hull et al., 1993). Alternatively, plates with similar colony counts were overlaid with the filters and the colonies were blotted at 37°C for two hours. Finally, plates inoculated with a pre-determined dilution of the samples were covered with the filters and incubated overnight at 37°C allowing the colonies to develop under the membranes. Orientation of plates and filters were marked for the subsequent identification of colonies positive with the immune assay. Beyond flood-inoculation, when testing fecal samples, a set of TSA plates were also loop-inoculated with undiluted samples. After covering the plates with the filters they were incubated overnight at 37°C. This procedure regularly resulted in approximately 200 –300 isolated colonies. Filters carrying the blots of colonies were incubated in 10 mL chloroform for 20 min to kill bacteria. This was followed by the extensive washing of the membranes in phosphate buffered saline (PBS) on an orbital shaker with 100 rpm (4 times 5 min in 20 mL aliquots) to remove the solvent and the bacterial debris. To eliminate the endogenous alkaline phosphatase (AP) activity regularly present in fecal samples (Horrigan and Danovitch, 1974) and interfering with the detection of the immune reaction (Hull et al., 1993) filters were soaked in 0.5% Tween 20 for 10 min

D. Szaka´ l et al. / Diagnostic Microbiology and Infectious Disease 45 (2003) 165–171

167

followed by 10 min at 100°C in a hot oven. After three rinses in PBS membranes were incubated in 2% skimmed milk for two hours at room temperature to block the free binding sites. A cell-free tissue culture supernatant diluted to 1:100 in 0.5% skimmed milk containing the monoclonal antibody MAIC-1 (Floderus et al., 1995) was added to the filters for 1 h, at room temperature. This antibody specifically reacts with the invasion plasmid coded protein IpaC secreted by virulent Shigella and EIEC strains (Floderus et al., 1995). After three washing steps in PBS, anti-mouse Ig-alkaline phosphatase conjugate (DAKO) diluted 1:1000 was added, also for 1 h at room temperature. The reaction was developed with Fast Red TR Salt-Naphtol AS MX Phosphate substrate (SIGMA-Aldrich GmbH) as previously described (Pa´ l et al., 1989).

iron agar slopes (Mast Laboratories) and into urea broth. Cultures showing reactions suggestive of Shigella or Salmonella were serogrouped with antisera (Murex Diagnostics). The species was confirmed in the Vitek automatic system using the Vitek GNI card (bioMerieux). The presence of enteropathogenic E. coli was investigated by the agglutination of ten colonies from the MacConkey plates with the diagnostic pooled sera 2, 3, and 4 for pathogenic E. coli (Murex Diagnostics). Campylobacter strains were isolated from Karmali’s Campylobacter agar (Oxoid) incubated at 42°C. The O antigen of EIEC strains identified by the immunoblot technique was determined by slide agglutination using immune sera prepared by injecting rabbits i.v. with boiled cells of EIEC strains of known serogroups.

2.3. IpaC ELISA

2.6. Guinea pig keratoconjunctivitis test

The IpaC specific monoclonal antibody, MAIC-1 (Floderus et al., 1995), was used to test clones selected and inoculated into wells of flat bottom ELISA plates (Costar) containing 200 ␮L of TSB. After overnight incubation at 37°C the assay was performed as described (Floderus et al., 1995; Pa´ l et al., 1997). When investigating artificially contaminated fecal samples, 16 randomly selected colonies were tested from each specimen.

To asses the virulence of EIEC isolated from clinical specimens strains were inoculated into the conjunctival sac of adult, out-bread guinea pigs, as described (Sereny, 1955). 2.7. Statistical analysis The sensitivity of the different methods to detect EIEC in artificially contaminated fecal samples was compared by McNemar’s non-parametric test.

2.4. ipaH PCR To detect DNA sequences in fecal samples indicating the presence of Shigella or EIEC a PCR system specific to the ipaH gene of these pathogens was used (Sethabutr et al., 1993; Venkatesan et al., 1989). Samples were inoculated onto TSA plates. Next day, approximately one quarter of the culture was suspended into 1 mL of distilled water and extracted with 125 ␮L of DNA Extraction Reagent, (Perkin Elmer) by boiling the specimens for 30 min, as described (De Lamballerie et al., 1992, Szaka´ l et al., 2001). After centrifugation with 13,000 g for two min, the supernatants were used as samples in the PCR reaction with primers specific to the ipaH gene (Sethabutr et al., 1993). 2.5. Bacteriologic examination of fecal specimens One hundred sixty-five fecal samples, including 44 repetitive specimens, from a total of 121 patients, received by the Microbiology Laboratory, Mubarak Al-Kabeer Hospital, Kuwait with the diagnosis of “diarrhea”, “enteritis” or “dysentery” were processed according to the guidelines for hospital laboratories in Kuwait for common bacterial enteric pathogens, i.e., Salmonella, Shigella, Campylobacter and enteropathogenic E. coli (Johny et al., 1994). Briefly, MacConkey (Oxoid) and salmonella-shigella agar plates (Difco Laboratories) were inoculated before and after Selenite-F broth enrichment (Mast Laboratories). Colonies suspected for Shigella or Salmonella were inoculated into Kligler’s

3. Results 3.1. Optimization of the colony blot assay To establish an effective and simple protocol to transfer antigens secreted by EIEC and Shigella strains to nitrocellulose filters different blotting procedures were compared. All three protocols tested, i.e., growing colonies on the surface of replica-inoculated nitrocellulose membranes, incubating filters placed over colonies for 2 h at 37°C, and growing colonies beneath the filters, respectively, resulted in colony prints clearly visible after the immune assay. The intensity of the color reaction on blots prepared by the 2 hour-long antigen-transfer procedure considerably varied when the same strains were tested repeatedly. On the other hand, no similar inter-assay variation was observed with the other two blotting techniques (data not shown). Therefore, after the initial experiments, the use of the 2 hour-long blotting protocol was discontinued. The colonies, and consequently their blots were smaller when bacteria were grown covered by the filters (Fig. 1, B) compared to prints obtained when growing colonies on the membranes (Fig. 1, A). However, this smaller size did not prevent the clear identification of positive colonies (Fig. 1, B), as confirmed by subsequent IpaC–specific ELISA of 10 positive, and 10 negative colonies taken from plates containing mixtures of E. coli K12 and 10 EIEC or 10 Shigella strains,

168

D. Szaka´ l et al. / Diagnostic Microbiology and Infectious Disease 45 (2003) 165–171

Fig. 1. IpaC specific colony blot assay of mixed cultures. (A) The membrane was replica-inoculated with a mixed culture (1:1) of EIEC 88 (O164) and E. coli J53, and incubated overnight at 37oC. (B) The membrane was placed over a freshly inoculated TSA plate and incubated overnight at 37°C. EIEC 88 and E. coli J53 (1:20). (C) A blot prepared as membrane B from a plate inoculated with a 1:1 mixture of Salmonella Enteritidis and E. coli J53. All plates contained approximately 500 colonies.

respectively (data not shown). Moreover, this procedure did not require the extra day needed to prepare the master plates for the replica-inoculation of the filters. Blots of the 80 EIEC and 59 Shigella strains were clearly marked by the immune reaction following this antigen transfer technique. The intensity of the color reaction was slightly lower with most of the EIEC strains than that observed with the Shigella isolates. Nevertheless, it was clearly distinguishable from the negative reactions seen with all the 100 nonenteroinvasive strains tested (Fig. 1). Therefore, during the subsequent experiments blots of cultures were prepared by allowing the colonies to develop while covered by the nitrocellulose filters. 3.2. Sensitivity of the colony blot using artificially infected samples Twenty fecal specimens diluted to contain approximately 109 coliform CFU/ml, were contaminated with increasing

germ counts of a virulent EIEC strain of serogroup O164. The samples were tested with the ipaH-specific PCR, with the IpaC-specific ELISA investigating 16 randomly selected colonies, and with the colony immunoblot assay. For the latter assay TSA plates were either flood-inoculated with samples further diluted to yield approximately 500 colonies per plates, or were loop-inoculated with the contaminated samples without further dilution. Without artificial contamination none of the diagnostic methods used gave positive reaction with any of the samples tested. PCR proved to be the most sensitive technique identifying all pathogen-containing specimens even at 1:5000 ratio of contamination (Table 1). On the other hand, ELISA of randomly selected colonies was the least sensitive method. Although all specimens containing EIEC as approximately 10% of their total coliform flora were identified by this method, the performance of this diagnostic approach considerably decreased at lower rates of pathogen content (Table 1).

Table 1 Sensitivity of different methods to identify EIEC in artificially contaminated fecal samples (n ⫽ 20) Approximate ratio of the pathogena

1:10 1:100 1:250 1:500 1:1000 1:5000 0 a

Number of positive samples identified PCR

20 20 20 20 20 20 0

ELISAb

20 8 1 0 1 0 0

Colony immunoblot Flood-inoculationc

Loop-inoculationd

20 20 18 12 8 1 0

20 20 16 7 3 1 0

As compared to the CFU of coliforms in the specimens. Testing 16 randomly selected colonies. c Plates were inoculated with titrated dilutions of sample to provide approximately 500 colonies. d Plates were loop-inoculated with undiluted samples yielding approximately 200 –300 colonies. b

D. Szaka´ l et al. / Diagnostic Microbiology and Infectious Disease 45 (2003) 165–171

On the other hand, with the colony blot assay IpaCsecreting colonies were detected in all samples containing at least 1% EIEC. Even at 1:250 and 1:500 rates of contamination, respectively, this latter technique significantly outperformed the method based on random colony selection (p ⬍ 0.01) (Table 1). For the colony blot assay, seeding the plates by flood-inoculation was more effective than the loop-inoculation technique (Table 1). However, the difference between these two methods was not statistically significant at any ratio of pathogen content (p ⬎ 0.05) (Table 1). This fact, and the extra time and material requirements associated with flood inoculation (i.e., the need for the preliminary determination of the CFU content of the samples (Hull et al., 1993) prompted us to use the loop-inoculation method during the subsequent field studies. 3.3. Detection of EIEC and Shigella in clinical samples Hundred and sixty five fecal samples received by the Mubarak Al-Kabeer Hospital were investigated for enteric bacterial pathogens by the standard laboratory methods, by the ipaH-specific PCR to detect the presence of Shigella-, or EIEC-related DNA sequences, and were also subjected to the IpaC specific colony blot assay. Non-typhoid Salmonella, and Campylobacter jejuni were isolated from 26 and five samples respectively. E. coli strains expressing O antigens associated with infantile diarrhea were detected in five samples (2 strains of O111 and 3 of O86 serogroups). Shigella strains (five S. flexneri and seven S. sonnei) were recovered from 12 samples, all submitted from different patients. From none of the samples was more than one pathogen isolated. The ipaH-specific PCR detected sequences specific to enteroinvasive strains in 17 samples, including all the 12 yielding Shigella by culture. The colony blot assay revealed the presence of IpaC expressing colonies in 16 samples, all being positive also by PCR. The 16 samples found to contain IpaC positive colonies included those from which Shigella had been simultaneously cultured. No PCR negative specimens were positive with the immune assay. By locating colonies on the master plates with the aid of their respective blots, in 13 out of these 16 specimens colonies expressing IpaC could directly be identified and isolated. In the remaining three samples the colony blot assay showed clear color reaction at spots where the density of colonies was too high for the clear identification of the positive ones. However, by re-suspending cells in PBS from the respective area of the culture, and plating this suspension followed by a repeated colony blot assay we could isolate the pathogenic strains in these cases, too. From the 12 samples shown to contain Shigella by culture, indeed, strains belonging to the same genus and species were recovered by isolating colonies marked by the immune assay. Beyond the 12 Shigella-containing samples the colony blot assay also identified four specimens which

169

were also positive by PCR, but from which no enteric pathogens had been cultured. The isolation and identification of IpaC positive colonies revealed that these samples contained EIEC. These specimens derived from two patients, from samples submitted one and two days apart, respectively. From one of the patients an EIEC strain expressing the O28 cell wall antigen, while from the other that of the O164 serogroup was isolated. All EIEC isolated were invasive, as shown in the guinea pig keratoconjunctivitis test. Both patients excreting EIEC had diarrhea without microscopic blood or pus cells detected in the samples. One sample, repeatedly positive by PCR, did not yield any colonies reactive with the IpaC-specific antibody, even after testing blots of six loop-inoculated, and six floodinoculated plates (i.e., testing over 4000 colonies). According to these results all together 2.4% of the samples, received from 1.6% of the patients, were positive for EIEC, as shown by the immunoblot technique.

4. Discussion Currently, no diagnostic methods combining easy and cheap technology with high sensitivity and specificity are available for the identification of diarrheagenic E. coli, and in particular for that of EIEC (Echeverria et al., 1991; Nataro and Kaper, 1998). The absence of lactose fermentation, the lack of lysine and ornithin decarboxylase and mucinase, or the non-motile character are frequently associated with the invasive phenotype in E. coli. However, rather the combination of these features than any of them alone is suggestive for EIEC (Abuxapgui et al., 1999; Silva et al., 1980). Consequently, due to the need of testing large number of colonies in multiple biochemical reactions, biotyping has a limited value in screening for this pathogen. In the past years the number of E. coli serogroups with reported invasive character has considerably increased making serotyping inconvenient as a routine procedure, as well (Pa´ l et al., 1998). Furthermore, since neither biochemical markers nor any of the O antigens are specifically linked to the invasive character of the isolates, the virulence of the strain thus identified still needs to be confirmed. The molecular methods recently developed for the identification of diarrheagenic E. coli are highly specific and sensitive, and are suitable to test individual colonies or, directly, the extracts of the specimens (Boileau et al., 1984; Echeverria et al., 1991; Frankel et al., 1990; Nataro and Kaper, 1998). However, particularly in diarrhea-endemic areas of the world, these techniques are available for a rather limited number of laboratories, only, leaving the etiology of infections frequently undetermined. Although widely used to identify bacterial colonies expressing cloned antigens, the colony blot immune assay has not been used extensively in the diagnostic work, so far. Nevertheless, the number of reports on its successful application to detect a variety of pathogens has increased in the

170

D. Szaka´ l et al. / Diagnostic Microbiology and Infectious Disease 45 (2003) 165–171

past years, and includes the identification of colonies of Clostridium botulinum (Goodnough et al., 1993), TSST-1producing Staphylococcus aureus (See et al., 1990), Pediococcus acidilactici (Bhunia and Johnson, 1992), Rhodococcus equi (Takai et al., 1993) and Shiga-toxin producing E. coli (Hull et al., 1993). In the current study we have shown that the technique can also be applied for the specific and sensitive identification of EIEC and Shigella strains in fecal samples. We have used a monoclonal antibody specific to IpaC, a plasmid-coded protein of these pathogens (Floderus et al., 1995). This antibody proved to be highly specific for invasive strains, i.e., no strains other than EIEC or Shigella gave false positive reaction in the colony blot assay. In order to keep the protocol simple, various methods of plate-inoculation and colony blotting were compared. Flood-inoculation of the plates with properly diluted samples yielded more colonies for blotting than loop-inoculation (approximately 500 vs. 200 –300). However, loop-inoculation required only one plate per specimen, while for the alternative method several plates were needed to find the proper sample dilution. Inevitably, as was shown with artificially contaminated samples, loop inoculation resulted in a slight decrease in sensitivity of the assay (Table 1). Replicaplating, and subsequent growing of colonies on filters resulted in larger colony-print sizes than allowing colonies to grow on plates covered by the membranes. However, the latter protocol did not require the day to prepare the master plates for replica-inoculation while still yielded clearly visible colony blots (Fig. 1). It should be noted that with our protocol we did not reach the sensitivity of the immunoblot technique previously reported (Hull et al., 1993). These authors could detect Shiga toxin producing E. coli in blots of mixed cultures at as low as 1:1000 to 1:5000 pathogen to non-pathogen ratios. This could be due to the fact that they used the more laborious replicaplating method allowing more colonies to be screened. Testing 165 clinical fecal samples proved that the compromises made concerning sensitivity to keep the protocol cheap, fast and straightforward did not affect adversely the performance of the immunoblot assay under field settings. With the immune assay, among the 165 specimens tested, we found 16 containing IpaC positive colonies, just one specimen shorter than the 17 identified by the Shigella and EIEC specific PCR. Obviously, the outcome of the colony blot assay depends on the ratio of the pathogen among the colonies present on the plate investigated. Apparently, in at least 16 out of the 17 PCR positive clinical samples it was sufficiently high to find IpaC expressing colonies with the immunoblot method. It should be noted that by screening a few hundred colonies the immunoblot provides a significantly more sensitive method than those based on testing randomly selected colonies with molecular, virulence-, or immune assays. In these cases only a few, rather than a few hundred colonies are considered to represent the sample, considerably reducing sensitivity. This was clearly shown in this study even

though as many as 16 randomly selected colonies, instead of the 3–5 recommended (Nataro and Kaper, 1998), was tested by ELISA (Table 1). Nevertheless, our results do not exclude the possibility that some patient could shed the pathogen below the detection level of the immunoblot assay. Theoretically, this could be the case with the only PCR positive, but immunoblot, and culture negative sample encountered in this study. Alternatively, non-viable, disintegrating cells of the pathogen could also contaminate the sample with DNA fragments detected by molecular methods. Furthermore, as we have shown before, mutations resulting in the loss of IpaC expression could lead to a negative immune assay, while molecular methods still could demonstrate the physical presence of virulence specific sequences (Pa´ l et al., 1985). This could also be the case with the PCR system and the immune assay used in this study. Allels of the the ipaH gene, targeted by the PCR used, is present both on the chromosome and on the invasion plasmid, while the gene coding for the IpaC protein is located extra-chromosomally, only (Parsot, 1994; Venkatesan et al., 1989). Although, based on our results, we surmise that extreme low pathogen content may not be a frequent case in bacillary dysentery, its actual frequency needs to be determined in large scale studies. Furthermore, the sensitivity of the immunoblot assay could be further increased by blotting multiple, floodinoculated plates from selected samples, should clinical or epidemiologic data make the etiology likely. Although the purpose of the current study was not to investigate the frequency of EIEC in Kuwait, our findings (i.e., EIEC was isolated from 1.6% of diarrhea cases) is consistent with other reports on the incidence of this pathogen from different parts of the world (Abuxapgui et al., 1999; Echeverria et al., 1989; Echeverria et al., 1992, Ke´ tyi, 1989; Pa´ l et al., 1997; Tamura et al., 1996). In summary, the IpaC-specific colony blot immune assay is a simple, straightforward method to screen for EIEC and Shigella in fecal cultures. Its time requirement does not exceed that of manual ELISA tests extensively used in several laboratories in the developing countries. Although it does not reach the sensitivity of PCR performed on extracts of mixed stool cultures, it is clearly superior to testing limited number of randomly selected colonies, i.e., the most frequently used diagnostic approach in laboratories equipped to carry out DNA hybridization or virulence assays. The assay efficiently screens a few hundred colonies, a task practically unattainable with bio-, or serotyping. A further, significant advantage of the method is, over molecular assays carried out on extracts of specimens, that it allows the immediate isolation of identified colonies for further investigations, like antibiotic sensitivity testing or epidemiologic studies. This could be of particular importance since, as was shown in this study, too, EIEC is a pathogen present, but likely to be often unrecognized, in areas where it has only been occasionally reported before (Pa´ l et al., 1997). The possibility of combining antibodies

D. Szaka´ l et al. / Diagnostic Microbiology and Infectious Disease 45 (2003) 165–171

specific to different virulence factors (e.g., IpaC and Shiga toxins) in a single colony blot assay to screen for various groups of diarrheagenic E. coli is currently under investigation. Acknowledgments This work was supported by the grants T030201 of the National Scientific Research foundation (OTKA), Hungary, MI O92 of Kuwait University, Kuwait, and NP/02/25 of the UAE University. The constant support and interest of Professor T.D. Chugh during this work, as well as the skillful technical assistance of Mrs. Csilla Weber and Mr. Akbar P. Kalandath is highly appreciated

References Abuxapgui, F. J. J., Hoil, S. G. J., Navarrete, H. M. R., Franco, P. M. A., & Rosel, V. M. L. (1999). Four biochemical reaction for identification of probable enteroinvasive Escherichia coli strains. Rev Latinoam Microbiol, 41, 259 –261. Beutin, L., Gleier, K., Kontny, I., Echeverria, P., & Scheutz, F. (1997). Origin and characteristics of enteroinvasive strains of Escherichia coli (EIEC) isolated in Germany. Epidemiol Infect, 118, 199 –205. Boileau, C. R., d’Hauteville, H. M., & Sansonetti, P. J. (1984). DNA hybridization technique to detect Shigella species and enteroinvasive Escherichia coli. J Clin Microbiol, 20, 959 –961. Bhunia, A. K., & Johnson, M. G. (1992). Monoclonal antibody-colony immunoblot method specific for isolation of Pediococcus acidilactici from foods and correlation with pediocin (bacteriocin) production. Appl Environ Microbiol, 58, 2315–2320. De Lamballerie, X., Zandotti, C., Vignoli, C., Bollett, C., & de Micco, P. (1992). A one-step microbial DNA extraction method using “Chelex 100” suitable for gene amplification. Res Microbiol, 143, 785–790. DuPont, H. L. (1990). Shigella species/bacillary dysentery. In: G. L. Mandell, R. G. Douglas Jr., & J. E. Bennett (Eds), Principles and Practice of Infectious Disease. Churchill Livingstone, pp. 1716 –1722. Echeverria, P., Sethabutr, O., & Pitarangsi, C. (1991). Microbiology and diagnosis of infections with Shigella and enteroinvasive Escherichia coli. Rev Infect Dis, 13 (Suppl 4), S220 –S225. Echeverria, P., Sethabutr, O., Serichantalergs, O., Lexomboon, U., & Tamura, K. (1992). Shigella and enteroinvasive Escherichia coli infections in households of children with dysentery in Bangkok. J Infect Dis, 165, 144 –147. Echeverria, P., Taylor, D. N., Lexsomboon, U., Bhaibulaya, M., Blacklow, N. R., Tamura, K., & Sakazaki, R. (1989). Case-control study of endemic diarrhoeal diseases in Thai children. J Infect Dis, 159, 543–548. Floderus, E., Pa´ l, T., Karlsson, K., & Lindberg, A. A. (1995). Identification of Shigella and entetroinvasive Escherichia coli strains by a virulencespecific, monoclonal antibody-based enzyme immunoassay. Eur J Clin Microbiol Infect Dis, 14, 111–117. Frankel, G., Riley, L., Giron, J. A., Valmassoi, J., Friedmann, A., Strockbine, N., Falkow, S., & Schoolnik, G. K. (1990). Detection of Shigella in feces using DNA amplification. J Infect Dis, 161, 1252–1256. Goodnough, M. C., Hammer, B., Sugiyama, H., & Johnson, E A. (1993). Colony immunoblot assay of botulinal toxin. Appl Environ Microbiol, 59, 2339 –2342. Horrigan, F. D., & Danovitch, S. H. (1974). The origin of human fecal alkaline phosphatase. Am J Digest Dis, 19, 603– 608. Hull, A. E., Acheson, D. W., Echeverria, P., Donohue-Rolfe, A., & Keusch, G. T. (1993). Mitomycin immunoblot colony assay for detec-

171

tion of Shiga-like toxin-producing Escherichia coli in fecal samples: comparison with DNA probes. J Clin Microbiol, 31, 1167–1172. Johny, M., Hassan, A. A., Ghafoor, M. A., Elhag, K. M., & Dhar, R. (1994). Microbiology procedure manual. Kuwait: Ministry of Health, pp. 40 –54. Ke´ tyi, I. (1989). Epidemiology of the enteroinvasive Escherichia coli, observations in Hungary. J Hyg Epidemiol Microbiol Immunol, 33, 261–267. Maurelli, A. T., Blackmon, B., & Curtiss, R. III (1984). Loss of pigmentation in Shigella flexneri 2a is correlated with loss of virulence and virulence-associated plasmid. Infect Immun, 43, 397– 401. Nataro, J. P., & Kaper, J. B. (1998). Diarrheagenic Escherichia coli. Clin Microbiol Rev, 11, 142–201. Pa´ l, T., Echeverria, P., Taylor, D. N., Sethabutr, O., & Hanchalay, S. (1985). Identification of enteroinvasive Escherichia coli by indirect ELISA and DNA hybridization. Lancet, ii, 785. Pa´ l, T., Formal, S. B., & Hale, T. L. (1989). Characterization of Virulence Marker Antigen of Shigella spp. and enteroinvasive Escherichia coli. J Clin Microbiol, 27, 561–563. Pa´ l, T., Al-Sweih, N. A., Herpay, M., & Chugh, T. D. (1997). Identification of enteroinvasive Escherichia coli and Shigella strains in pediatric patients using an IpaC specific ELISA. J Clin Microbiol, 35, 1757–1760. Pa´ l, T., Serichantalerg, O., Echeverria, P., Scheutz, F., & Cam, P. D. (1998). The Use of an IpaC-specific ELISA to identify enteroinvasive Escherichia coli strains of unusual serogroups. Diagn MIcrobiol Infect Dis, 32, 255–258. Parsot, C. (1994). Shigella flexneri: genetics of entry and intercellular dissemination in epithelial cells. In: J. L. Dangl (Ed.), Bacterial Pathogenesis of Plants and Animals. (pp. 217–241). Curr Top Microbiol Immunol, 192. See, R. H., Adilman, S., Bartlett, K. H., & Chow, A. W. (1990). Colony immunoblot assay for the detection of staphylococcal toxic shock syndrome toxin 1 (TSST-1) with anti-TSST-1 F(ab⬘)2 fragments. J Clin Microbiol, 28, 1488 –1490. Sere´ ny, B. (1955). Experimental Shigella keratoconjunctivitis. Acta Microbiol Acad Sci Hung, 2, 293–296. Sethabutr, O., Venkatesan, O., Murphy, G. S., Eampokalap, B., Hoge, C. W., & Echeverria, P. (1993). Detection of Shigellae and enteroinvasive Escherichia coli by amplification of the invasion plasmid antigen H DNA sequence in patients with dysentery. J Infect Dis, 167, 458 – 461. Silva, R. M., Toledo, M. R. F., & Trabulsi, L. R. (1980). Biochemical and cultural characteristics of invasive Escherichia coli. J Clin Microbiol, 11, 441– 444. Szaka´ l, D., Gado´ , I., & Pa´ l, T. (2001). A colony blot immonoassay to detect enteroinvasive Escherichia coli and Shigella in water samples. J Appl Microbiol, 90, 229 –236. Takai, S., Iie, M., Kobayashi, C., Morishita, T., Nishio, T., Ishida, T., Fujimura, T., Sasaki, Y., & Tsubaki, S. (1993). Monoclonal antibody specific to virulence-associated 15- to 17-kilodalton antigens of Rhodococcus equi. J Clin Microbiol, 31, 2780 –2782. Tamura, K., Sakazaki, R., Murase, M., & Kosako, Y. (1996). Serotyping and categorisation of Escherichia coli strains isolated between 1958 and 1992 from diarrhoeal diseases in Asia. J Med Microbiol, 45, 353–358. Taylor, D. N., Echeverria, P., Pa´ l, T., Sethabutr, O., Saiborisuth, S., Sricharmorn, S., Rowe, B., & Cross, J. (1986). The role of the Shigella spp., enteroinvasive Escherichia coli and other enteropathogens as causes of childhood dysentery in Thailand. J Infect Dis, 153, 1132–1138. Venkatesan, M. M., Buyssee, J. M., & Kopecko, D. J. (1989). Use of Shigella flexneri ipaC and ipaH gene sequences for the general identification of Shigella spp. and enteroinvasive Escherichia coli. J Clin Microbiol, 27, 2687–2691. Wanger, R. A., Murray, B. E., Mathewson, J. J., & DuPont, H. L. (1988). Enteroinvasive Escherichia coli in travellers with diarrhoea. J Infect Dis, 158, 640 – 642.