Phenotypic and Molecular Analysis of Tellurite Resistance among Enterohemorrhagic Escherichia coli O157:H7 and Sorbitol-Fermenting O157:NM Clinical Isolates

Abstract

A total of 66 (98.5%) of 67 enterohemorrhagic Escherichia coli (EHEC) O157:H7 strains had increased potassium tellurite (Te) MICs (32 to 1,024 μg/ml), grew on Te-containing media, and possessed Te resistance (ter) genes, whereas 83 (96.5%) of 86 sorbitol-fermenting (SF) EHEC O157:NM strains had Te MICs of ≤4 μg/ml, did not grow on Te-containing media, and lacked ter genes. Optimal detection of SF EHEC O157:NM strains requires Te-independent strategies.

Tellurite (Te)-resistant (Te^r) non-sorbitol-fermenting enterohemorrhagic Escherichia coli (EHEC) O157:H7 strains cause diarrhea and hemolytic-uremic syndrome (HUS) worldwide (17), but sorbitol-fermenting (SF) EHEC O157:NM (nonmotile) strains have emerged as pathogens only in Europe (1, 6, 8, 12) and Australia (3) so far. SF EHEC O157:NM strains are not distinguishable from commensal E. coli strains on sorbitol MacConkey agar (SMAC), and do not grow (11) on cefixime-Te (CT)-SMAC (22), which is frequently used for selective isolation of EHEC O157:H7 strains from feces, foods, and the environment (2, 4, 5, 9, 13, 20, 21). Te^r in EHEC O157:H7 is associated with the ter (terZABCDEF) gene cluster (19), duplicated in strain EDL933 in O islands OI 43 and OI 48 (14). One of these islands was originally identified in strain 86-24 (16) as the Te^r and adherence-conferring island (16). Te^r in SF EHEC O157:NM strains has not been investigated. Because Te susceptibility (Te^s) could thwart the detection of such strains on media containing Te, we investigated Te^r and the presence of ter genes in a large collection of SF EHEC O157:NM clinical isolates. We compared these characteristics with those of EHEC O157:H7.

Isolation and characterization of strains.

A total of 67 EHEC O157:H7 and 86 SF EHEC O157:NM strains were isolated between 1987 and 2003 from patients with HUS (n = 118) or bloody (n = 11) or watery (n = 19) diarrhea and from asymptomatic carriers (n = 5). To avoid biases from strains selected by their Te^r, only EHEC O157 strains isolated on Te-free media by methods described previously (6, 10, 11) were included in this study, and they comprise a subset of the 572 E. coli O157 strains recovered during this interval. The 67 EHEC O157:H7 strains belonged to Shiga toxin (Stx) genotypes stx₁ (2 strains), stx₂ (28 strains), stx₁ + stx₂ (6 strains), stx_2c (8 strains), stx₁ + stx_2c (4 strains), and stx₂ + stx_2c (19 strains). All 86 SF EHEC O157:NM strains contained stx₂ only. Te MICs (the lowest Te concentrations which inhibited growth) were determined using a microdilution broth method (15). Each strain was tested in duplicate and in two independent experiments using 4 × 10⁴ to 5 × 10⁴ CFU/well and serial twofold concentrations (1 to 1,024 μg/ml) of potassium Te (Sigma, Taufkirchen, Germany) in 100 μl of Luria-Bertani (LB) broth. The ability of EHEC O157 to grow on solid media containing the Te concentration routinely used for E. coli O157:H7 selective isolation was tested by inoculating 10⁵ CFU from overnight LB broth cultures on CT-SMAC (potassium Te 2.5, μg/ml; cefixime, 0.05 μg/ml [Oxoid, Basingstoke, United Kingdom]) and LB agar plus 2.5 μg of potassium Te per ml (LB-Te agar). The presence of ter genes was determined by PCRs using primer pairs TerZ1 plus TerZ2 (terZ), TerA1 plus TerA2 (terA), TerB1 plus TerB2 (terB), TerC1 plus TerC2 (terC), TerD1 plus TerD2 (terD), TerE1 plus TerE2 (terE), and TerF1 plus TerF2 (terF) (19), with strains EDL933 and C600 as positive and negative controls, respectively. Genomic DNA was digested (BamHI and PstI; New England Biolabs, Frankfurt, Germany), separated in 0.6% agarose, and probed under stringent conditions with digoxigenin-labeled terC (19) (DIG DNA labeling and detection kit; Roche Molecular Biochemicals, Mannheim, Germany).

Te^r and ter presence among EHEC O157.

Of 67 EHEC O157:H7 strains, 50 (74.6%) and 16 (23.9%) had high (256 to 1,024 μg/ml) or intermediate (32 to 128 μg/ml) Te MICs, respectively (Table 1). All 66 Te^r strains grew well on CT-SMAC and LB-Te agar and contained terZ, terA, terB, terC, terD, terE, and terF (Table 1). One EHEC O157:H7 strain (5288/91) had a Te MIC of <1 μg/ml, failed to grow on CT-SMAC and LB-Te agar, and lacked all ter genes (Table 1). In contrast, 83 (96.5%) of 86 SF EHEC O157:NM strains were susceptible to Te (Te MICs of ≤4 μg/ml) (Table 1). Of these 83 strains, 70 (84.3%) failed to grow on CT-SMAC and LB-Te agar and 13 strains (15.7%) were strongly inhibited on both media (<10 colonies grew after overnight incubation). All 83 Te^s SF EHEC O157:NM strains lacked ter genes (Table 1). Two of three Te^r SF EHEC O157:NM isolates (3226/98 and 3323/98) had Te MICs of 128 μg/ml, grew well on CT-SMAC and LB-Te agar, and contained all ter genes; the remaining Te^r (MIC, 256 μg/ml) SF EHEC O157:NM strain (4180/97) had no ter genes (Table 1).

TABLE 1.

Te^r and the presence of ter genes among EHEC O157:H7 and SF EHEC O157:NM strains

Serotype	Total no. of strains	SFb	No. (%) of strains with a characteristic
			Te MIC (μg/ml) in range of:				Robust growthc on:		No or strongly reduced growthd on:		ter genese
			256-1024	32-128	2-4	<1	CT-SMAC	LB-Te agar	CT-SMAC	LB-Te agar	Present	Absent
O157:H7	67	0	50 (74.6)f	16 (23.9)f	0 (0)	1 (1.5)	66 (98.5)	66 (98.5)	1 (1.5)	1 (1.5)	66 (98.5)g	1 (1.5)
O157:NMa	86	86	1 (1.2)f	2 (2.3)f	40 (46.5)	43 (50.0)	3 (3.5)	3 (3.5)	83 (96.5)	83 (96.5)	2h (2.3)g	84i (97.7)

^aNM, nonmotile.

^bSF, sorbitol fermentation as determined after overnight incubation on SMAC agar.

^cRobust growth, >1,000 colonies of normal size per plate.

^dNo growth, no colonies grown; strongly reduced growth, <10 colonies reduced in size present after overnight incubation.

^eAs determined by PCRs targeting terZ, terA, terB, terC, terD, terE, and terF (19); present, all genes investigated were present; absent, all genes investigated were absent.

^fTellurite resistance (Te^r): EHEC O157:H7 versus SF EHEC O157:NM, P < 0.000001 by the χ² test.

^gPresence of ter genes: EHEC O157:H7 versus SF EHEC O157:NM, P < 0.000001 by the χ² test.

^hTwo of three Te^r strains; the third was negative for ter genes.

ⁱ83 Te^s strains and 1 Te^r strain.

Southern hybridization.

terC was carried on a ca. 9-kb DNA fragment in ter⁺ SF EHEC O157:NM strains 3226/98 and 3323/98 (Fig. 1, lanes 6 and 7), on a ca. 6.3-kb DNA fragment in strain EDL933 (lane 1), and on no DNA fragments in strains 5288/91, 4180/97, and ter-negative SF EHEC O157:NM strain 493/89 (lanes 3 to 5).

FIG. 1.

Hybridization of BamHI-PstI-digested genomic DNA from EHEC O157 strains and controls with the terC probe. M, molecular weight marker (1-kb DNA ladder; Gibco-BRL). In lanes 1 to 7, the following strains are displayed (serotype, Te-MIC in micrograms per milliliter, and the presence of ter genes as detected by PCR are given in parentheses): lane 1, EDL933 (O157:H7, 128, ter⁺); lane 2, C600 (not available; <1; ter negative) (negative control); lane 3, 5288/91 (O157:H7; <1; ter negative); lane 4, 4180/97 (SF O157:NM; 256; ter negative); lane 5, 493/89 (SF O157:NM; <1; ter negative); lane 6, 3226/98 (SF O157:NM; 128; ter⁺); lane 7, 3323/98 (SF O157:NM; 128; ter⁺). Two terC copies detected in EDL933 after DNA separation by pulsed-field gel electrophoresis (19) were not distinguishable by conventional gel electrophoresis.

Effect of Te^r on detection of EHEC strains.

Our study provides for the first time a basis for the inability of SF EHEC O157:NM to grow on CT-SMAC (11), which was until now only speculated to be caused by their Te^s (12, 16). In contrast to EHEC O157:H7, almost all SF EHEC O157:NM strains lack ter genes and are Te^s. Low Te MICs for SF EHEC O157 strains and the comparable growth inhibition of these isolates on CT-SMAC and LB-Te agar suggest that Te, and not cefixime, is the growth-inhibiting component in CT-SMAC. The Te^r and Te^s correlated with the presence and absence, respectively, of ter genes in all but one of the 153 EHEC O157 strains investigated. The single Te^r, ter-negative SF EHEC O157:NM strain (4180/97) is currently being investigated for other presently known mechanisms of Te^r (18). Interestingly, Southern hybridization suggests that the genomic positions of terC differ in the two ter⁺ SF EHEC O157:NM isolates and EHEC O157:H7 strain EDL933 (Fig. 1). Studies are under way to determine if the ter genes in these SF EHEC O157:NM strains are clustered, similar to those in EDL933 (14), and to determine the genomic location of the ter cluster as well as its copy number. Also, further studies should clarify the reasons for the substantially lower frequency of Te^s found among central European EHEC O157:H7 isolates (1.5%) than among North American E. coli O157:H7 (20%) (19). Taken together, our data demonstrate a significant difference between EHEC O157:H7 and SF EHEC O157:NM in the frequency of Te^r and ter genes, demonstrate a diversity among SF EHEC O157:NM strains as far as the presence of ter genes is concerned, and suggest that other mechanisms of Te^r exist in a minority of such strains. However, most importantly, our data clearly indicate that, because of their Te^s, most SF EHEC O157:NM strains are missed by strategies currently used for the isolation of EHEC O157:H7 strains in many clinical laboratories. A similar observation has been reported for a subset of other Te^s EHEC strains (5). The selectivity offered by incorporating Te into agar media, while appropriate for isolating E. coli O157:H7 (the most important EHEC serotype worldwide), hinders the assessment of the geographic distribution, medical significance, and epidemiology of SF EHEC O157:NM strains, which in Germany are the second most common cause of HUS (6). Detection opportunities that do not select against SF EHEC O157:NM and, optimally, specifically target these organisms (7) are needed to answer the question about the relative significance of both EHEC O157 pathogens in human diseases.