Nissui Glucose Fermentative Gram-Negative Rod Identification System EB-20 Gives a Unique Profile for Typical Non-Sorbitol-Fermenting Escherichia coli O157:H7

Abstract

The 98 non-sorbitol-fermenting (NSF) Escherichia coli O157:H7 strains identified on a Nissui glucose fermentative gram-negative rod identification system (EB-20) gave a unique biochemical profile number that was not detected in 85 pathogenic and 13 nonpathogenic E. coli strains. Thus, EB-20 is useful for the identification of NSF E. coli O157:H7 and provides a simple, cost-effective, and reliable tool for clinical laboratories.

Shiga toxin-producing Escherichia coli (STEC) is one of five currently recognized groups of E. coli that cause several distinct clinical syndromes associated with infectious bacterial gastroenteritis, and E. coli O157 is the most commonly identified diarrheagenic E. coli isolate in North America and Europe (7). Since STEC O157:H7 emerged as a public health threat following its initial identification as a pathogen in a 1982 outbreak of illness associated with the consumption of undercooked ground beef (26), it has been reported as a food-borne pathogen throughout the world. In Japan, the first outbreak of STEC O157:H7 was reported in a kindergarten in Saitama in 1990 (17). Since the report of an outbreak of STEC O157:H7 infection among schoolchildren in Japan in 1996 (32), the number of STEC serotypes of non-O157:H7 E. coli such as E. coli O26 (H11, nonmotile [NM]; and H not tested) and serotype O111 (NM, H not tested, and H untypeable [UT]) have also increased in Japan (20). However, STEC O157:H7 has remained the predominant STEC serotype in Japan since 1996. The use of Shiga toxin detection assays to screen stool samples for STEC has previously been reported (25); however, these assays are not dependent on a particular bacterial serotype. Therefore, it is necessary to isolate the bacterium from the sample after detection of the toxin and to perform biochemical and serological identification for reporting. During an outbreak of food-borne disease, large numbers of samples must be tested, and a simple identification method is needed. Currently, the most rapid and accurate Shiga toxin detection assays such as the PCR (5, 8) and rapid immunoassay (24) can be carried out in large clinical laboratories. However, these sophisticated assays are not suitable for detection of a particular bacterial serotype and are often not readily applicable for use in small clinical microbiology laboratories where simple and low-cost methods are needed. At a typical Japanese local health center, there is almost no automated equipment for optical scanning and interpreting and analyzing results for the identification of gram-negative bacilli. The detection of E. coli O157:H7 is based on its recovery from samples and the presence of its virulence-associated factors (verocytotoxins) or the detection of its O157 antigens (8, 28, 30). Selective media for isolating E. coli O157:H7 strains rely on the fact that most of these strains display characteristic biochemical reactions (30): no β-glucuronidase activity and no sorbitol (SOR) fermentation within 24 h at 37°C, except for some German or U.S. strains (11, 12). However, these last two biochemical features are not commonly used in routine clinical and food laboratories because of the simplicity of well-established commercial identification systems. The use of these systems on presumptive E. coli strains offers not only the advantage of confirming strains at the genus and/or species level but also the possibility of detecting the presence of STEC O157:H7 by identifying profiles that are unique to these strains. It has been reported that a few profiles generated by the API 20E identification system strongly suggest E. coli O157 (10) and that MicroScan generates one of two profiles with more than 90% of the strains of E. coli O157 (1). Results obtained from the ID 32E system showed atypical biochemical reactions but accurate identification at the species level and no unique biochemical profile numbers for E. coli O157, although these numbers were distinct from those of other serotypes (18). In this study, we evaluated the Nissui glucose fermentative gram-negative rod identification system (EB-20; Nissui Pharmaceutical Co. Ltd., Tokyo, Japan) for STEC, enterotoxigenic E. coli (ETEC), enteropathogenic E. coli (EPEC), and nonpathogenic E. coli. We also tested separately for the presence of β-glucuronidase in these strains. The presence of β-glucuronidase in E. coli is about 96 (9) to 97% (13). However, E. coli serotype O157:H7, which causes hemorrhagic colitis, has been reported to be β-glucuronidase negative (16). Serotypes O111 and O1 have also been reported to be β-glucuronidase negative (16). During the last decade, we have used the EB-20 system to study many intestinal bacteria isolated from clinical specimens and food samples. EB-20 is able to differentiate E. coli O157:H7 from pathogenic and nonpathogenic E. coli.

(Part of this study of E. coli O157:H7 with the EB-20 identification system and the β-glucuronidase test has been presented previously [H. Kodaka, Y. Uesaka, S. Mizuochi, and K. Horigome, Abstr. 91st Gen. Meet. Am. Soc. Microbiol., abstr. C-290, 1991].)

A total of 183 strains of diarrheagenic E. coli and 13 strains of nonpathogenic E. coli were tested with conventional biochemical tests and the EB-20 system at 37°C. These strains, from the American Type Culture Collection (ATCC; Rockville, Md.), the Osaka Prefectural Institute of Public Health (Osaka, Japan), Nagasaki University attached to the Institute of Tropical Medicine (Nagasaki, Japan), Toho University attached to Omori Hospital (Tokyo, Japan), the Tokyo Metropolitan Research Laboratory of Public Health (Tokyo, Japan), and the National Institute of Infectious Diseases, were assembled at Nissui Pharmaceutical Co. Ltd. The strains were as follows: (i) 6 ATCC E. coli O157:H7 strains, of which ATCC 43888 does not produce either Shiga-like toxin I or II and does not possess the genes for these toxins; (ii) 13 nonpathogenic E. coli strains from ATCC; (iii) 92 Japanese STEC O157:H7 strains from the Tokyo Metropolitan Research Laboratory of Public Health (6 strains), Toho University (2 strains), and the National Institute of Infectious Diseases (84 strains); (iv) 3 STEC O26:H11 strains and 3 STEC O111:NM strains from the Tokyo Metropolitan Research Laboratory of Public Health; (v) 8 EPEC strains from Nagasaki University; and (vi) 71 ETEC strains (24 heat-labile enterotoxin [LT]-producing strains comprising 12 strains from Nagasaki University, 10 strains from the Osaka Prefectural Institute of Public Health, and 2 strains from the National Institute of Infectious Diseases; 26 heat-stable enterotoxin [ST]-producing strains comprising 16 strains from Nagasaki University, 9 strains from the Osaka Prefectural Institute of Public Health, and 1 strain from the National Institute of Infectious Diseases; and 21 [LT-ST]-producing strains comprising 9 strains from Nagasaki University, 10 strains from the Osaka Prefectural Institute of Public Health, and 2 strains from the National Institute of Infectious Diseases).

Conventional methods.

All strains were tested by standard methods (4, 19). The β-glucuronidase assay using 5-bromo-4-chloro-3-indolyl-β-d-glucuronide cyclohexylammonium salt and 4-methylumbellifery-β-d-glucuronide (29) and the β-galactosidase assay using 5-bromo-3-indolyl-β-d-galactoside and 4-methylumbellifery-β-d-galactoside were carried out. The concentration of each chromogenic substrate was 0.01% (15).

EB-20.

The EB-20 test (Fig. 1) was inoculated, incubated at 37°C for about 18 to 20 h, and interpreted as recommended by the manufacturer. The EB-20 test required the oxidase and motility test results for organism identification. The EB-20 test menus of esculin (ESC), phenylalanine deaminase (PPA), malonate (MALO), adonitol (ADO), and raffinose (RAFF) differ from the menus of tryptophane deaminase (TDA), gelatinase, glucose, melibiose (MEL), and amygdalin for the API 20E test. The EB-20 biochemical menu of PPA and mannose differs also from the menus of glucose, MEL, TDA acetamide, tartaric acid, and nitrate for the MicroScan gram-negative rod panel.

FIG. 1.

Nissui identification test EB-20. Test wells include the following: hydrogen sulfide (H₂S), ESC, PPA, indole (IND), Voges-Proskauer (VP), citrate (CIT), lysine decarboxylase (LDC), arginine dihydrolase (ADH), ornithine decarboxylase (ODC), ONPG, urease (URE), MALO, ADO, inositol (INO), RAFF, RHA, SOR, SUC, and arabinose (ARA). The oxidase and motility tests were performed separately.

Bacterial toxin detection.

The ETEC strains were tested for the production of an ST by an enzyme-linked immunoassay (ELISA) (27) and for production of an LT by the bead ELISA (21) or the sandwich ELISA (2). Shiga toxins or verocytotoxins were tested at other places.

E. coli serotyping.

The serotyping for ETEC and EPEC were based on a positive agglutination reaction using 43 kinds of O (somatic) and 22 kinds of H (flagellum) antigen antisera for E. coli (Denka Seiken Co. Ltd., Tokyo, Japan). The O and H antigens of all strains were examined with the antisera by slide agglutination tests and tube agglutination tests, respectively. The presence of flagellar antigens was determined by flagella staining (14) before the H antigen testing.

We tested 183 E. coli strains that caused diarrhea and 13 strains of nonpathogenic E. coli with the EB-20 system according to the manufacturer's instructions. Of the 98 E. coli O157:H7 strains tested, 98 (100%) generated a biochemical profile number (0151453) that was not observed among any of the other non-O157:H7 strains tested (Table 1). This profile consisted of positive reactions for indole, lysine decarboxylase, ornithine decarboxylase, o-nitrophenyl-β-d-galactopyranoside (ONPG), RAFF, rhamnose (RHA), sucrose (SUC), mannose, and arabinose and negative reactions for hydrogen sulfide, ESC, PPA, Voges-Proskauer, citrate, arginine dehydrolase, urease, MALO, ADO, inositol, and SOR. All E. coli O157:H7 strains were β-glucuronidase negative. Two EPEC strains, 12 LT-producing strains, 3 ST-producing strains, and 13 LT-ST-producing strains were also β-glucuronidase negative. All E. coli O157:H7 strains were negative for SOR. One EPEC and one LT-ST-producing strain were also negative for SOR. All LT-producing and ST-producing strains were SOR positive. Two strains other than E. coli O157:H7 were both SOR and β-glucuronidase negative (one EPEC and one LT-ST-producing strain). The serotyping and EB-20 profile numbers generated were OUT:NM and 0111453, respectively, for the EPEC strain and O25:NM and 0151003, respectively, for the LT-ST-producing strains. Three STEC O26:H11 strains generated the profile 0151463, and three STEC O111:NM strains generated profile numbers 0101473 and 0141433. The 13 ATCC nonpathogenic E. coli strains were not identified as 0151453. During this study for pathogenic and nonpathogenic E. coli strains, we got 37 different profile numbers with the EB-20 system. When the penultimate digit in the EB-20 profile is 5, this identifies the strain as RHA positive, SOR negative, and SUC positive; the profiles of one EPEC strain and 98 E. coli O157:H7 strains were recognized in this way.

TABLE 1.

Frequency of EB-20 profile numbers of STEC, EPEC, ETEC, and nonpathogenic E. coli

Profile no.a	No. (%) of profile frequencies
	STEC				EPEC (n = 8)e	ETEC			Nonpathogenic E. coli (n = 13)b
	O157:H7 (n = 6)b	O157:H7 (n = 92)c	O26:H11 (n = 3)d	O111:NM (n = 3)d		LT (n = 24)f	ST (n = 26)g	LT-ST (n = 21)h
0011013	0	0	0	0	0	0	1 (4)	0	0
0101033	0	0	0	0	1 (13)	6 (25)	0	0	0
0101433	0	0	0	0	0	2 (8)	0	0	0
0101473	0	0	0	1 (33)	0	0	0	0	0
0111023	0	0	0	0	0	0	0	1 (5)	0
0111033	0	0	0	0	5 (63)	0	2 (8)	1 (5)	1 (8)
0111133	0	0	0	0	0	1 (4)	0	0	0
0111413	0	0	0	0	0	0	1 (4)	0	0
0111423	0	0	0	0	0	0	0	1 (5)	0
0111433	0	0	0	0	0	1 (4)	0	1 (5)	0
0111453	0	0	0	0	1 (13)	0	0	0	0
0111473	0	0	0	0	1 (13)	0	0	0	0
0111523	0	0	0	0	0	0	0	1 (5)	0
0141423	0	0	0	0	0	0	1 (4)	0	0
0141433	0	0	0	2 (67)	0	0	0	0	0
0141473	0	0	0	0	0	0	1 (4)	0	0
0150013	0	0	0	0	0	0	0	0	1 (8)
0151033	0	0	0	0	0	2 (8)	0	1 (5)	5 (38)
0151413	0	0	0	0	0	2 (8)	0	0	0
0151433	0	0	0	0	0	1 (4)	1 (4)	6 (29)	3 (23)
0151453	6 (100)	9 (100)	0	0	0	0	0	0	0
0151463	0	0	3 (100)	0	0	0	0	0	0
0151473	0	0	0	0	0	3 (13)	10 (39)	1 (5)	0
0151533	0	0	0	0	0	0	1 (4)	0	0
0153033	0	0	0	0	0	0	0	0	1 (8)
2101033	0	0	0	0	0	1 (4)	0	1 (5)	0
2101123	0	0	0	0	0	0	0	1 (5)	0
2101133	0	0	0	0	0	0	0	1 (5)	0
2111032	0	0	0	0	0	0	0	1 (5)	0
2111033	0	0	0	0	0	1 (4)	1 (4)	1 (5)	0
2111073	0	0	0	0	0	0	1 (4)	0	0
2111123	0	0	0	0	0	0	1 (4)	2 (10)	0
2111133	0	0	0	0	0	2 (8)	1 (4)	0	0
2151033	0	0	0	0	0	1 (4)	0	0	1 (8)
2151433	0	0	0	0	0	0	1 (4)	1 (5)	0
2151471	0	0	0	0	0	0	0	0	1 (8)
2151473	0	0	0	0	0	1 (4)	3 (12)	0	0

^aThese numbers represent a tabulation of positive biochemical results based on EB-20 analysis.

^bStrains from ATCC.

^cJapanese strains: 84 strains from the National Institute of Infectious Diseases, 6 strains from the Tokyo Metropolitan Research Laboratory of Public Health, and 2 strains from Toho University.

^dStrains from Tokyo Metropolitan Research Laboratory of Public Health.

^eStrains from Nagasaki University.

^fTwelve strains from Nagasaki University, 10 strains from the Osaka Prefectural Institute of Public Health, and 2 strains from the National Institute of Infectious Diseases.

^gSixteen strains from Nagasaki University, nine strains from the Osaka Prefectural Institute of Public Health, and one strain from the National Institute of Infectious Diseases.

^hNine strains from Nagasaki University, 10 strains from the Osaka Prefectural Institute of Public Health, and 2 strains from the National Institute of Infectious Diseases.

The 0151453 profile generated by EB-20 was not observed in the identification code of the 619 Enterobacteriaceae strains including 344 E. coli strains isolated from 3,562 urine samples during a previous study (15). A result with the combination of SOR negative, RAFF positive, and β-glucuronidase negative was reported by Krishnan et al. to be O157:H7 (16), but we found one EPEC (OUT:NM) with the same combination. The 98 non-sorbitol-fermenting (NSF) E. coli O157:H7 isolates were β-glucuronidase negative, SUC positive, dulcitol (DUL) positive, SOR negative, and RAFF positive. However, it has been reported that some strains of E. coli O157:H7 do not ferment DUL (16). The EB-20 system does not contain DUL. Therefore, if the strain does not ferment DUL, the profile number would remain the same. Of 91 strains isolated from sporadic or intrafamilial cases in Japan, 82 strains (90.1%) were β-glucronidase negative, SOR negative, DUL positive, SUC positive, and RHA positive; 8 strains (8.8%) were β-glucronidase negative, SOR negative, DUL positive, SUC negative and RHA positive; and 1 strain (1.1%) was β-glucronidase negative, SOR negative, DUL negative, SUC positive, and RHA positive (17). E. coli O157:H7 strains which do not ferment DUL were not obtained and therefore not tested in this study.

The identification of NSF E. coli O157:H7 by the EB-20 test was E. coli inactive (23), which refers to biochemically inactive strains of an E. coli group formerly known as Alkalescens-Dispar (22). This identification is inconsistent with E. coli O157:H7 because E. coli inactive is not usually motile. The β-glucuronidase test in not included in the EB-20 system, and so it must be done separately, but the test is not always useful because positive β-glucuronidase strains of STEC O157:H7 have been reported (11). Of 188 serotype O157 strains examined, 166 isolates were β-glucuronidase negative and verocytotoxin positive, while the remaining 22 isolates were all β-glucuronidase positive and verocytotoxin negative (29). According to the report, most (91.9%) of the E. coli O157:H7 strains had the same biotype according to the API 20E test, and a few failed to ferment either SUC (2.7%) or RHA (5.4%) (10). The EB-20 test menu of ESC, PPA, MALO, ADO, and RAFF differs from the API 20E test menu of TDA, gelatinase, glucose, MEL, and amygdalin. Although all STEC O157:H7 strains had the same profile on the EB-20 test in this study, E. coli O157:H7 strains that fail to ferment SUC or RHA would have a different profile. Therefore, if the strains that failed to ferment SOR are identified as E. coli, the slide agglutination test for E. coli O157 and the β-glucuronidase test should be carried out. Based on 657 strains of E. coli O157:H7 identified between 1995 and 1999, a total of 34 strains (5.2%) were aberrant strains (31); however, the aberrant strains are also clinically important and not negligible. We agree with Ware et al. (31) that direct detection of Shiga toxin-producing strains may be the future method of choice if it is financially feasible. Even though SOR-fermenting STEC O157:NM strains have been reported in sporadic cases (6) and outbreaks (3), the SOR-negative and β-glucronidase-negative E. coli O157:H7 strain is still the most predominant organism.

The incorporation of this unique biochemical profile could inform users of the presumptive identification of the typical NSF E. coli O157:H7 strain and should suggest additional confirmation tests to be performed, such as the slide agglutination test for E. coli O157 and the β-glucuronidase test. If NSF E. coli O157:H7 strains with aberrant biochemical properties that would lead to other profiles are identified as E. coli, the slide agglutination test for E. coli O157 and the β-glucuronidase test should also be carried out. The strain with aberrant biochemical properties should be sent to a reference laboratory for the important confirmation of the production of Shiga toxins. It is nonetheless important that there be a relatively simple and inexpensive technique for discriminating E. coli strains, which would be especially useful for laboratories which do not have an automated system that can take readings hourly by optical scanning and interpret and analyze results for the identification of the bacteria.