Bacteriophage-Based Rapid and Sensitive Detection of Escherichia coli O157:H7 Isolates from Ground Beef

Abstract

We investigated the efficacy of bacteriophage-based detection technology to detect Escherichia coli O157:H7 from ground beef. The assay involved a short enrichment period of 8 h followed by capture of the pathogen on O157-specific immunomagnetic beads. The captured cells were treated with O157-specific lytic bacteriophage, CSLO157. Upon phage-induced lysis, the enzyme adenylate kinase, which was released from the lysed cells, was measured in terms of relative light units using luciferin–luciferase assay. The plaque forming efficiency (e.g., phage susceptibility) and ability to capture cells with immunomagnetic beads were examined using an array of 74 E. coli O157:H7 isolates obtained from various clinical and foodborne samples. Immunmagnetic beads successfully captured all 74 isolates; however, only 53 isolates showed susceptibility toward the bacteriophage. Susceptible isolates were further classified into two broad groups, moderately sensitive isolates, which generated phage titer ∼10⁷ pfu/mL (group I, n = 15), and highly susceptible isolates, which generated high phage titer ∼10⁹ pfu/mL (group II, n = 38). We selected 15 isolates (9 from group I and 6 from group II) and individually spiked beef samples (ca. 3–8 cells/25 g beef ) to evaluate the bacteriophage-based detection system. Eight out of nine isolates from group I and all six isolates from group II were successfully detected. Pathogenic E. coli strains belonging to other serogroups (12 serogroups, 67 isolates) as well as nontarget microorganisms (n = 18) were not lysed by the bacteriophage and hence were not detected. The method is high-throughput compatible, is rapid, and can provide live culture the following day by streaking an aliquot before phage lysis on conventional selective agar media.

Introduction

F oods of bovine origin, such as ground beef, frequently serve as a vehicle for outbreaks involving Escherichia coli O157:H7, which causes bloody diarrhea, hemorrhagic colitis, and, in certain cases, hemolytic uremic syndrome (Kaper et al., 2004). E. coli O157:H7 has a low infectious dose, and the presence of this pathogen in beef is considered an adulterant. As such, there is zero tolerance for its presence in beef. To meet the highest sensitivity of detection, several polymerase chain reaction (PCR)-based methods are available (Fratamico et al., 1995; Bhagwat, 2003; Feng, 2007). Although these methods offer greatest sensitivities, the isolation of a live pathogen culture is still very essential. Relying solely on PCR results would likely impede epidemiological aspects of investigations of foodborne outbreaks and also would create uncertainty regarding the possibility of false-positive PCR results (Rangel et al., 2005). PCR and DNA isolation methods both are adversely affected by the presence of inhibitory compounds present in complex food matrices (Heller et al., 2003; Wolffs et al., 2004; Bhagwat et al., 2008; Chua and Bhagwat, 2009).

Detection of pathogens using phages may allow the distinction of living bacteria much more efficiently than standard molecular techniques (Favrin et al., 2001). However phage-based methods may not always provide the sensitivity of a PCR approach (Goodridge et al., 1999; Favrin et al., 2003). On the other hand, measurement of ATP using a bioluminescent assay offers sensitive and rapid method in which theoretical detection limits have been predicted in the range of 10³ bacterial cells (Stanley, 1989). However, when tested in complex biological matrices such as enrichments of vegetable and meat extracts, 10⁴–10⁵ cells were required (Thouand et al., 2008). In living cells, the enzyme adenylate kinase (AK, E. C.2.7.4.3) catalyzes the equilibrium reaction of generating two molecules of ADP from ATP and AMP in the presence of Mg⁺⁺ ions. Squirrell and Murphy (1997) increased the detection sensitivity of the bioluminescent assay by focusing on AK rather than ATP as a cell marker by adding exogenous ADP to the assay mix. The released intracellular AK amplified ATP by converting exogenous ADP to ATP and AMP. The resulting bioluminescent signal was proportional to AK activity and, ultimately, to the number of cells in the sample. The lack of specificity, a major drawback in the application of AK-based bioluminescent assay, was resolved by using specific bacteriophages to lyse target cells instead of lysis buffer (Blasco et al., 1998; Wu et al., 2001). These studies were performed with laboratory pure cultures of Salmonella and E. coli O157:H7, and efficacy of AK-based bioluminescence in phage lysed cells using food matrices has not been thoroughly examined (Sun et al., 2002). In this study we used O157 immunomagnetic beads to concentrate E. coli O157:H7 cells after a short 8 h selective enrichment and utilized bacteriophage CSLO157, which is specific to E. coli O157 to lysethe pathogen. The released AK activity was measured using the luciferin–luciferase assay (Blasco et al., 1998). We reasoned that the combination of two principles, specificity (provided by immunomagnetic beads and bacteriophages) and magnification of the signal (e.g., luciferin–luciferase reaction), would enable detection of E. coli O157:H7comparable to DNA-based methods. We also took upon the task to detect E. coli O157:H7 at low contamination levels (at ∼3–8 cells/25 g) on ground beef, which was subsequently stored at 4°C for 2 days to induce cold stress. Data from the experiments involving 74 E. coli O157:H7 isolates and several other nontarget organisms (toxigenic and commensal isolates) are presented.

Materials and Methods

Bacterial strains and culture conditions

A list of E. coli O157:H7 isolates and other nontarget strains used in this study are listed in Tables 1 –4. Reference stocks of E. coli O157:H7 and other nontarget strains were stored at − 70°C in 15% (vol/vol) glycerol in LB broth. A stock culture was kept at 4°C on tryptic soy agar (Oxoid) plates and was refreshed every week. Working cultures were prepared by suspending a single colony in 10 mL tryptone soy broth (TSB; Oxoid) and then were incubated overnight (∼18 h) with shaking (200 rpm) at 37°C. Alaska diagnostics bacteriophage specific to E. coli O157 (CSLO157, obtained from The Food and Environment Research Agency, Sand Hutton, York, United Kingdom) was used.

Table 1.

Escherichia coli O157:H7 Isolates Used for Spiking Beef Samples

USDA ID	Original laboratory ID	Phage susceptibility group ^a	Notes	Reference/source
106	2886–75	II	First US case of disease, 1975, California	Bhagwat et al. (2005)
105	93–111	II	1993 Washington outbreak reference strain	Bhagwat et al. (2005)
52	ATCC 43895	II	EK 279, Na^R,Rf^R wild type	Bhagwat et al. (2005)
A64	96.0428	II	Cow isolate	USDA, Beltsville, MD
A68	CC 0078	II	Fairfax Inova Hospital, VA	USDA, Beltsville, MD
A91	TW07961	II	Human isolate (F, 10 years), Ohio, 1998	NFST^b
A117	TW14313	I	Human isolate, (Michigan), 2006	NFST
A114	TW11516	I	Human isolate (M, 23 years, Michigan), 2005	NFST
A100	TW09098	I	Human isolate (F, 11 years, Michigan), 2003	NFST
A118	TW14359	I	Human isolate (F, 55 years, Michigan), 2006	NFST
A99	TW09087	I	Human isolate (F, 16 years, Michigan), 2003	NFST
A93	TW08080	I	Human isolate (F, 33 years, Montana), 1999	NFST
A65	96.0932	I	Human isolate	USDA, Beltsville, MD
A66	96.094	I	Fairfax Inova Hospital, VA	USDA, Beltsville, MD
A4	ATCC 43888	I	ATCC reference strain	ATCC, Manassas, VA

Phage susceptibility grouping was based on plaque forming efficiency using bacteriophage CSLO157 (see Fig. 1 for details).

National Food Safety and Toxicology Center (Michigan State University).

Table 2.

Additional Escherichia coli O157:H7 Isolates Used in This Study

USDA ID	Original laboratory ID	Notes	Reference/source
Group I: Intermediate plaque former (∼107 pfu/mL)
A95	TW08609	Human isolate (M, 1.4 years, Washington), 1999	NFST (MSU)^a
A89	TW07941	Human isolate (F, 48 years, Massachusetts), 1998	”
A107	TW11032	Human isolate (Michigan), 2002	”
A111	TW11137	Human isolate (Michigan), 2001	”
A116	TW14301	Human isolate (Michigan), 2006	”
A102	TW09178	Human isolate (F, 36 years, Michigan), 2003	”
Group II: High plaque former (∼109 pfu/mL)
A115	TW14270	Human isolate (F, 18 years, Michigan), 2005	NFST (MSU)^a
A88	TW07945	Human isolate (F, 53 years), 1998	”
A104	TW10022	Human isolate (Japan), 1996	”
A108	TW11039	Human isolate (Michigan), 2002	”
A112	TW11308	Human isolate (Australia), 2005	”
A84	TW05356	Human isolate (Washington), 1995	”
A83	TW04863	Human isolate (Washington), 1993	”
A110	TW11107	Human isolate (Michigan), 2004	”
A106	TW10245	Human isolate (M, 67 years, Michigan), 2002	”
A81	TW02303	Human isolate (Oregon)	”
A79	TW01663	Human isolate (Canada), 1987	”
A85	TW06591	Human isolate (Germany), 1990	”
A103	TW10012	Human isolate (Pennsylvania), 1998	”
A87	TW07763	Human isolate (Germany), 1991	”
A96	TW08612	Human isolate (F, 2 years, Washington), 2001	”
A86	TW07520	Human isolate (Australia)	”
A80	TW02300	Human isolate (Canada)	”
A94	TW08264	Human isolate (Japan), 1996	”
A3	ATCC 35150	–	ATCC, Manassas, VA
A5	ATCC 43889	–	”
A6	ATCC 43890	–	”
A7	ATCC 43894	–	”
A8	ATCC 43895	–	”
A1	96–014	–	U of Idaho
A63	96.0427	Cow isolate	USDA, Beltsville, MD
A67	96.008	Human isolate	”
A69	CC0079	Fairfax Inova Hospital, VA	”
A70	CC0080	Fairfax Inova Hospital, VA	”
253	DEC3A	Washington, 1985	Bhagwat et al. (2005)
255	DEC3D	Michigan, 1988	”
252	OK-1	1996, Okayama, Japan outbreak	”
250	OK-1	1996, Okayama, Japan outbreak	”
Group III: Phage-resistant strains
A82	TW02883	Human isolate	NFST (MSU)^a
A90	TW07957	Human isolate (F, 1.4 years, D.C.), 1999	”
A92	TW08030	Human isolate (F, 22 years, Montana), 2000	”
A97	TW08613	Human isolate (Washington), 2002	”
A98	TW08623	Human isolate (Washington), 2002	”
A109	TW11102	Human isolate (Michigan), 2004	”
A113	TW11346	Human isolate (Michigan), 2001	”
A105	TW10045	Human isolate (Connecticut), 1996	”
A2	8624	–	U of Idaho
A71	CC 0081	Fair Fax Inova Hospital, VA	USDA, Beltsville, MD
A72	CC 0324	Milk isolate	”
A73	2203	Composite manure	”
A74	2204	Cow feces	”
A75	2205	Composite manure	”
A76	2206	Flies from calf area	”
A77	2207	Composite manure	”
A78	2248	Exiting calf	”
42	H25349	Wisconsin outbreak strain	Bhagwat et al. (2005)
39	F15110	Wisconsin outbreak strain	”
98	CDC B 6914-MS1	–	”
256	DEC3E	Canada, 1988	”

National Food Safety and Toxicology Center (Michigan State University).

Table 3.

Toxigenic Escherichia coli Strains Other Than O157:H7 Serotype Used in This Study

USDA ID	Serotype/Notes	Reference/source
113	O5:HN bovine isolate, diarrhea	Bhagwat et al. (2005)
210	O26:H– strain from diarrhea	”
206	O26:H– strain from diarrhea	”
103	Nontoxigenic O26:H−, 1974	”
209	O26:HNM, Rhode Island, 1994	”
208	O26:HN, New Hampshire, 1979	”
207	O26:HN	”
102	O26:H11 strain from infant, 1977	”
201	O26:H11, Mexico, 1986	”
200	O26:H11 strain from HC	”
202	O26:H11, United kingdom	”
203	O26:H11, France, 1952	”
204	O26:H11, Wisconsin, 1961	”
205	O26:H11, Delaware, 1977	”
211	O44:H18	”
214	O55:H6	”
213	O55:H6, Dutch Guiana, 1958	”
215	O55:H6, Congo, 1962	”
217	O55:H7	”
108	Nontoxigenic O55:H7	”
115	stx2 + O55:H7	”
216	Nontoxigenic O55:H7	”
219	O55:H7, New York, 1950	”
220	O55:H7, Iran, 1963	”
222	O103:HN	”
112	O103:H6, Diarrhea, fecal	”
110	O104:H21 1994 bloody diarrhea	”
236	O111, Carries invasion insert	”
101	O111:H− Strain from HUS, 1985	”
235	O111:NM	”
234	O111:HN, Kenya, 1986	”
224	O111:H2 DEC12a, diarrhea SSI	”
225	O111:H2	”
227	O111:H8, Texas, 1999	”
100	O111:H8 strain from HC	”
99	O111:H8 strain from HUS	”
228	O111:H9 Diarrhea outbreak	”
229	O111:H11, Cuba, 1953	”
231	O111:H12, Guatemala, 1967	”
230	O111:H12, Germany, 1954	”
232	O111:H12, Italy, 1951	”
233	O111:H21, New Jersey, 1966	”
237	O113:H21 HUS	”
238	O119:H6, Ohio, 1984	”
240	O121:NM, Massachusetts, 1998	”
239	O121:H19, Montana, 1998	”
241	O125:HNM, North Carolina, 1986	”
114	O127:H6 Model EPEC organism	”
246	O128:NM, Virginia, 1979	”
263	O128a:H2, Missouri, 1971	”
242	O128:H7, Tanzania, 1974	”
245	O128:H21, Bangladesh, 1977	”
244	O128:H21, India, 1971	”
109	O133:H21 Diarrhea, agglutinin−	”
247	O142:H6, Brazil, 1983	”
248	O145:HNM, South Dakota, 1982	”
249	O156:H21, Bovine isolate, stx1 ⁺	”

Table 4.

Nontarget Microorganisms Used in This Study

USDA ID	Original laboratory ID	Notes	Reference/source
S1	# 30	Shigella sonnei	Walter Reed Army Institute, MD
S3	Uban 12	Shigella dysenteriae 1(12uban 11c + 1)	”
S7	2457 T	Shigella flexneri2	”
SL1344	SL1344	S. enterica serovar Typhimurium, wildtype, his Sm ^r	Hoiseth and Stocker (1981)
68	UMD 221	Klebsiella pneumonia	USDA, Beltsville, MD
69	UMD219	Proteus vulgaris	”
70	UMD 225	Serratia marcescens	”
71	UMD 223	Citrobacter freundi	”
74	UMD 228	Staphylococcus aureus
204	Ribotype DUP-14591	Enterobacter aerogenes	Kim and Bhunia (2008)
205	Ribotype DUP-15301	Enterobacter cloaceae HK8	”
206	Ribotype DUP-18066	Hafnia alvei	”
75	ATCC 13061	Bacillus cereus	ATCC, Manassas, VA
78	ATCC 27270	Enterococcus faecium	”
79	ATCC 51299	Enterococcus faecalis	”
125	ATCC 13932	Listeria monocytogenes Serotype 4B	”
126	ATCC 35897	Listeria welshimerri	”
128	ATCC 33090	Listeria innocua	”

Experimental contamination of ground beef samples

Ground beef (20% fat) was purchased from a local grocery and was divided aseptically into 25-g portions in pulsifier bags (Microbiology International). Beef samples were artificially contaminated with ∼3–8 E. coli O157:H7 cells by adding 200 μL of appropriate dilutions. Viable cell counts were confirmed by retrospective spread-plating 10-fold serial dilution onto tryptic soy agar plates. Artificially contaminated beef samples were stored at 4°C for 48 h before subjecting to enrichment. All beef samples were tested for intrinsic E. coli O157:H7 by the conventional, as well as bacteriophage-based method, and were found to be negative. Each E. coli O157:H7 isolate was tested in 5 to 9 independent tests and each test included three replicate samples.

Selective enrichment protocol

To each 25-g ground beef sample, 225 mL of brain heart infusion broth (BHI; Oxoid) was added. The broth was amended with the following selective agents (μg · mL⁻¹ final concentration): novobiocin, 11.0; cefsulodin, 11.0; vancomycin, 8.8; maltol, 4.4; and acriflavine, 10.0. The samples were pummeled in a pulsifier (Microbiology International) for 2 min and were incubated at 41.5°C for 8 h without shaking.

E. coli O157 Dynabeads capture

The protocol for the E. coli O157 Dynabeads capture (Dynal) was performed as described by the manufacturer except that 11 mL of the enrichment sample was used after it was filtered through a Alaska food filter (Alaska Diagnostics) and 20 μL of antibody-coated magnetic beads was incubated at room temperature in a Magnetic Sample Rotator (Alaska Diagnostics) for 10 min. A magnetic field was applied for 5 min, and the supernatant was carefully aspirated and discarded. Samples were washed two times with phosphate-buffered saline containing 0.1% Tween-80 (PBST). The supernatant was discarded, and the particle–bacteria complex was resuspended in 500 μL of PBST. Twenty-microliter beads were streaked on sorbitol MacConkey agar supplemented with cefixime tellurite selective supplement (CT-SMAC; Oxoid), and the remainder of the magnetic beads were subjected to E. coli O157:H7–specific phage treatment as described below.

Phage treatment and luciferin–luciferase (Fastr_AK™) assay to detect E. coli O157:H7

The phage treatment and luciferin–luciferase assay was performed essentially as described earlier for the Fastr_AK™ assay for Salmonella strains except that E. coli O157–specific phage CLSO157 was used (Patel et al., 2009; Hammack and Chen, 2010). Briefly, the suspended Dynabeads magnetic particles (from the previous step, see above) were divided into two 240-μL aliquots and placed into 2 wells of a 96-well microtiter filter plate (with 3.0 μm glass fiber/0.2 μm, Bio-Inert membrane; Europa House). The magnetic particle–bacteria complexes were washed three times with 200 μL PBST by placing the filter plate on a 96-well vacuum manifold. For each sample, one aliquot was treated with bacteriophage CSLO157 diluted to deliver 5 × 10⁸ pfu in 100 μL, while the other aliquots received 100 μL phage diluent buffer (ATSB solution, which is PBS with 5 mM glucose and 10 mM MgSO₄). The bacteriophage–magnetic bead–bacteria complex and control aliquots were incubated at 37°C for 2 h. Lysed bacteria or ATSB buffer eluate from individual wells were collected in a new plate using 96-well vacuum manifold. The plate was immediately placed in a Fastr_AK plate luminometer equipped with two injector-coupled tubes, one containing 10 mM ADP solution and the other with a proprietary luciferin–luciferase mixture. For each measurement, the following controls were included: bacteriophage suspension solution, adenylene kinase, unlysed magnetic bead–bacteria complex, and enrichment medium. Emitted light from wells was measured as relative light units. The relative light units from control wells (T₁) were subtracted from sample wells that received bacteriophages to get the background subtracted light unit readings (T₂). Each sample was assayed in triplicate and each treatment was repeated for at least three times. The data were plotted as signal-to-noise ratio (T₂/T₁) of relative light units.

For all statistical analyses, SigmaStat 3.0 software was used. Data were analyzed by one-way analysis of variance test or Student–Newman–Keuls method to determine statistical differences between means of treatments.

Results and Discussion

The aim of this study was to evaluate specificity and sensitivity of a bacteriophage-based assay for detection of E. coli O157:H7 in ground beef. We also optimized this assay while keeping the enrichment time at 8 h, artificial contamination level to ∼3–8 cells per 25 g ground beef, and stressing the cells at 4°C for 2 days before enrichment. Initially, we evaluated several enrichment media in addition to BHI such as buffered peptone water and modified buffered peptone water, basic listeria enrichment broth (Heller et al., 2003; Kim and Bhunia, 2008; Jasson et al., 2009), and TSB with yeast extract containing varying amounts of selective agents (novobiocin, cefsulodin, vancomycin, maltol, and acriflavine) (Jasson et al., 2009). These selective media successfully enriched E. coli O157:H7 isolates 52, 106, and 105 to a detectable threshold after 18 h incubation (data not shown). However, with a shorter enrichment period (e.g., 8 h), these media were successful only if beef samples were enriched immediately after spiking (e.g., without subjecting to cold stress). Additionally, enrichment in basic listeria enrichment broth and TSB with yeast extract media resulted in high background readings of the enzyme AK (data not shown). Among all different enrichment media examined, only BHI broth containing novobiocin, cefsulodin, vancomycin, maltol, and acirflavine generated a detectable number of E. coli O157:H7 and low background levels of adenylate kinase in the control untreated wells. The observations obtained with three E. coli O157:H7 isolates were further extended to analyze an additional 71 E. coli O157:H7 isolates (Tables 1 and 2).

Phage sensitivity of E. coli O157:H7 isolates

The ability to lyse E. coli O157:H7 is the key for this detection method; therefore, we determined phage susceptibility of all 74 isolates. Phage titer of CSLO157 initially determined on ATCC 43895 reference strain (9.5 ± 4.4 × 10⁹) was compared with all of the other isolates to determine the sensitivity to phage-induced lysis (Fig. 1). Out of 74 E. coli O157:H7 isolates, 21 isolates were not lysed by CSO157 and produced no plaques. The remaining 53 isolates were grouped as moderately susceptible (group I, n = 15) and highly susceptible (group II, n = 38). The mean phage titer (pfu · mL⁻¹) of group I isolates was 5.5 ± 8.5 × 10⁷ compared with 4.2 ± 1.7 × 10⁹ (p ≤ 0.05). All phage-resistant as well as phage-susceptible isolates (Tables 1 and 2) were determined to be of O157:H7 serogroup by latex agglutination test (Oxoid) as well as by their binding to Dynal O157 immunomagnetic beads by estimating percentage of cells recovered on CT-SMAC plates (data not shown).

FIG. 1.

Box plot of grouping of Escherichia coli O157:H7 isolates based on phage susceptibility. Phage titer (pfu · mL⁻¹) was determined by 10-fold serial dilution of the stock phage CSLO157. Out of 74 E. coli O157:H7 isolates, 21 isolates did not form plaques. Remaining 53 isolates were grouped based on their median plaque forming efficiency (group I, 0.85 ± 8.5 × 10⁷ and group II, 4.2 ± 1.7 × 10⁹; Kruskal–Wallis one-way analysis of variance on Ranks, p < 0.05). The boundary of the box closest to zero indicates the 25th percentile, a line within the box marks the median, and the boundary of the box farthest from zero indicates the 75th percentile. Whiskers (error bars) above and below the box indicate the 95th and 5th percentiles.

E. coli O157:H7–specific bacteriophages have been isolated and evaluated to control the target pathogen population in sheep and cattle feedlots (Raya et al., 2006; Callaway et al., 2008) and under vitro conditions (Niu et al., 2009). In the current study a single phage, CSLO157, specific to O157 was used. To improve the efficacy and robustness of this assay, a cocktail of phages containing diverse but O157:H7-specific phages are needed, which could be the subject for future investigations.

We tested Shiga-toxin-producing non–O157 E. coli strains isolated from clinical and foodborne outbreak studies (Bhagwat et al., 2005) for their susceptibility toward bacteriophage CSLO157. None of the serotype strains tested such as O26 (n = 13), O55 (n = 9), O111 (n = 15), or O128 (n = 5) generated any titer or visible plaques. Other serotypes tested and found to be resistant to bacteriophage CSLO157 were O44, O103, O104, O113, O119, O121, O125, O127, O133, O142, O145, and O156 (Table 3). Altogether, 67 toxigenic E. coli isolates belonging to 12 serotypes were tested and found to be resistant to the bacteriophage CSLO157. We also examined other enteropathogens such as Shigella sonnei, Shigella dysenteriae, Shigella flexneri, and Salmonella enteric serovar Typhimurium strains. Other nonpathogenic organisms tested for the phage susceptibility were Klebsiella pneumonia, Citrobacater freundi, and other commensal bacteria (n = 22) (Table 4).

Detection of group I and II isolates in artificially inoculated beef

Next we selected several group I and II isolates (Table 1) and individually spiked them on beef samples. The spiked beef samples were stored at 4°C for 2 days before enrichment. Each E. coli O157:H7 isolate was spiked in 5 to 9 independent tests and each test included three replicate samples. After 8 h enrichment at 41.5°C without shaking, samples were processed for phage lysis as outlined in the Materials and Methods section. For group I isolates, the ratios of relative light units released upon phage lysis to unlysed/untreated cells were >1.8 except for isolate A66 (Fig. 2A). Signal-to-noise ratio for isolate A66 was 0.92 ± 0.15 and was statistically similar to ratio from blank or control wells, which was 0.82 ± 0.82 (p < 0.05). Rest of the isolates showed higher signal-to-noise ratios with significant statistical difference with reference to blank or isolate A66. Detection efficacy was much better for group II isolates (Fig. 2B), where five out of six isolates had signal-to-noise ratio >10 (isolate A106 had median signal-to-noise ratio of 4.4 ± 1.4). From all samples, a portion of magnetic beads (20 μL out of 500 μL suspension after capture and wash step) were spread on CT-SMAC plates and after overnight incubation yielded typical O157:H7 type colonies.

FIG. 2.

Box plot of detection efficiency (signal-to-noise ratio, T₂/T₁) of E. coli O157:H7 isolates from group I (A) and group II (B) using Fastr_AK phage assay. The boundary of the box closest to zero indicates the 25th percentile, a line within the box marks the median, and the boundary of the box farthest from zero indicates the 75th percentile. Whiskers (error bars) above and below the box indicate the 95th and 5th percentiles (not shown where indistinguishable from the box). Ratios of relative light units (signal-to-noise ratios) obtained from lysed (T₂) and control/unlysed (T₁) cells for individual isolates are plotted, and blank ratio was calculated by measuring light units from wells treated with BHI broth (T₁) and phage CSL157 without host cells (T₂). Each E. coli O157:H7 isolate was tested in 5 to 9 independent tests, and each test included three replicate samples. *p < 0.05 by pairwise comparison using Student–Newman–Keuls method.

Although isolate A66 was not successfully detected by the phage lysis method, it was probably not due to insufficient number of cells after 8 h enrichment. In all 7 tests using isolate A66, we were able to isolate O157:H7 colonies on CT-SMAC plates (data not shown). It may be noted that isolate A66 was a poor host for the phage CSLO157 and generated lowest phage titer of 1.9 ± 0.89 × 10³ pfu · mL⁻¹ among all group I strains. It is interesting to note that although isolates A99 and A114 generated titers of 3.6 ± 1.7 × 10⁴ and 5.1 ± 2.2 × 10⁴ pfu · mL⁻¹ (respectively), their signal-to-noise ratio was significantly above the background (Fig. 2A). Thus, it appears that there was no effective lysis of A66 cells upon treatment with phage CSLO157 to release enough AK.

Phage susceptibility and signal-to-noise ratio

In general, group II isolates had much better signal-to-noise ratio (median ratio 14.58 ± 6.2) compared to group I isolates (5.17 ± 3.6) (p < 0.05) (Fig. 3). If we assume that higher phage titer is equivalent to efficient lysis of host cells, better detection efficacy in terms of higher signal-to-noise ratio is expected for group II isolates. There was no difference in binding efficacy of Dynal O157 immunomagnetic beads to group I and II isolates (data not shown). Secondly, both groups of isolates did generate typical E. coli O157:H7 serotype colonies on CT-SMAC plates from 20 μL aliquot of the immunomagnetic beads, and no correlation was observed in number of colonies recovered versus signal-to-noise ratio after phage lysis. This observation indicates that all isolates had reached threshold enrichment to be captured on immunomagnetic beads, and degree to which they were susceptible to bacteriophage CSO157 was reflected in the signal strength they generated upon lysis (Fig. 3).

FIG. 3.

Relationship of phage titer and signal-to-noise ratio for detection of E. coli O157:H7 isolates using Fastr_AK phage assay. For the two-way plot, ratios from individual isolates from two groups were plotted against their phage titer values.

Conclusions

The potential of a new AK-phage assay for the detection of low numbers of E. coli O157:H7 strains in ground beef after 8 h of enrichment has been clearly demonstrated in this study. The assay successfully detected 14 out of 15 O157:H7 isolates. However, further work is needed to generate a cocktail of phages that will specifically lyse diverse range of O157:H7 isolates, as 21 out of 74 isolates were found to be resistant to phage CSLO157. A more robust assay outcome is expected using phage cocktails as opposed to single phage utilized in this study.

Disclosure Statement

No competing financial interests exist.

References

Bhagwat

. Simultaneous detection of Escherichia coli O157:H7, Listeria monocytogenes and Salmonella strains by real-time PCR. Int J Food Microbiol, 2003; 84:217–224.

Bhagwat

, Chan

, Han

, Tan

, Kothary

, Jean-Gilles

, Tall

. Characterization of enterohemorrhagic Escherichia coli strains based on acid resistance phenotypes. Infect Immun, 2005; 73:4993–5003.

Bhagwat

, Patel

, Chua

, Chan

, Cruz

, Aguilar

. Detection of Salmonella species in foodstuffs. Methods in Molecular Biology: Signaling Nucleic Acid Probes, Methods and Protocols, 429 Marx

, Seitz

. Totowa, NJ: Humana Press Inc., 2008; 33–43.

Blasco

, Murphy

, Sanders

, Squirrell

. Specific assays for bacteria using phage mediated release of adenylate kinase. J Appl Microbiol, 1998; 84:661–666.

Callaway

, Edrington

, Brabban

, Anderson

, Rossman

, Engler

, Carr

, Genovese

, Keen

, Looper

, Kutter

, Nisbet

. Bacteriophage isolated from feedlot cattle can reduce Escherichia coli O157:H7 populations in ruminant gastrointestinal tracts. Foodborne Pathog Dis, 2008; 5:183–191.

Chua

, Bhagwat

. A rapid and simple DNA extraction procedure to detect Salmonella spp. and Listeria monocytogenes from fresh produce using real-time PCR. Food Anal Methods, 2009; 2:96–101.

Favrin

, Jassim

, Griffiths

. Application of a novel immunomagnetic separation-bcteriophage assay for the detection of Salmonella enteritidis and Escherichia coli O157:H7 in food. Int J Food Microbiol, 2003; 85:63–71.

Favrin

, Jassim

, Griffiths

. Development and optimization of a novel immunomagnetic separation- bacteriophage assay for detection of Salmonella enterica Serovar Enteritidis in broth. Appl Environ Microbiol, 2001; 67:217–224.

Feng

. Rapid methods for the detection of foodborne pathogens: current and next-generation technologies. Food Microbiology: Fundamentals and Frontiers, 3rd Doyle

, Beuchat

. Washington, DC: ASM Press, 2007; 911–34.

10.

Fratamico

, Sackitey

, Wiedmann

, Deng

. Detection of Escherichia coli O157:H7 by multiplex PCR. J Clin Microbiol, 1995; 33:2188–2191.

11.

Goodridge

, Chen

, Griffiths

. The use of a fluorescent bacteriophage assay for detection of Escherichia coli O157:H7 in inoculated ground beef and raw milk. Int J Food Microbiol, 1999; 47:43–50.

12.

Hammack

, Chen

. Methods committe on microbiology. J AOAC Int, 2010; 93:11B–22B.

13.

Heller

, Davis

, Peak

, Wingfield

, Cannons

, Amuso

, Cattani

. Comparison of methods for DNA isolation from food samples for detection of Shiga toxin-producing Escherichia coli by real-time PCR. Appl Environ Microbiol, 2003; 69:1844–1846.

14.

Hoiseth

, Stocker

. Aromatic-dependent Salmonella typhimurium are non-virulent and effective as live vaccines. Nature, 1981; 291:238–239.

15.

Jasson

, Rajkovic

, Beart

, Debevere

, Uyttendaele

. Comparison of enrichment conditions for rapid detection of low numbers of sublethally injured Escherichia coli O157 in food. J Food Prot, 2009; 72:1862–1868.

16.

Kaper

, Nataro

, Mobley

HLT

. Pathogenic Escherichia coli. Nat Rev Microbiol, 2004; 2:123–140.

17.

Kim

, Bhunia

. SEL, a selective enrichment broth for simultaneous growth of Salmonella enterica, Escherichia coli O157:H7, and Listeria monocytogenes. Appl Environ Microbiol, 2008; 74:4853–4866.

18.

Niu

, Johnson

, Xu

, McAllister

, Sharma

, Louie

, Stanford

. Host range and lytic capability of four bacteriophages against bovine and clinical human isolates of Shiga toxin-producing Escherichia coli O157:H7. J Appl Microbiol, 2009; 107:646–656.

19.

Patel

, Uddin

, Ensell

, Crabtree

. Alaska fastrAK™ HTS PT1000 system for Salmonella analysis in processed seafood and processed meat groups. AOAC-RI Performance-tested methodsTM 110801, 2009. www.aoac.org/testkits/testedmethods_2.htm 2009 2010 July 07(Online.)

20.

Rangel

, Sparling

, Crowe

, Griffin

, Swerdlow

. Epidemiology of Escherichia coli O157:H7 outbreaks, United States, 1982–2002. Emerg Infect Dis, 2005; 11:603–609.

21.

Raya

, Varey

, Oot

, Dyen

, Callaway

, Edrington

, Kutter

, Brabban

. Isolation and characterization of a new T-even bacteriophage, CEV1, and determination of its potential to reduce Escherichia coli O157:H7 levels in sheep. Appl Environ Microbiol, 2006; 72:6405–6410.

22.

Squirrell

, Murphy

. Rapid detection of very low numbers of micro-organisms using adenylate kinase as a cell marker. A Practical Guide to Industrial Uses of ATP-Luminescence in Rapid Microbiology. Stanley

, Smither

, Simpson

. Surrey, UK: Cara Technology Limited, 1997; 107–113.

23.

Stanley

. A review of bioluminescent ATP techniques in rapid microbiology. J Biolumin Chemilumin, 1989; 4:375–380.

24.

Sun

, Khosravi

, Albrechtsen

, Brovko

, Griffiths

. Comparison of ATP and in vivo bioluminescence for assessing the efficiency of immunomagnetic sorbents for live Escherichia coli O157:H7 cells. J Appl Microbiol, 2002; 92:1021–1027.

25.

Thouand

, Vachon

, Liu

, Dayre

, Griffiths

. Optimization and validation of a simple method using P22::luxAB bacteriophage for rapid detection of Salmonell enterica Serotypes A, B, and D in poultry samples. J Food Prot, 2008; 71:380–385.

26.

Wolffs

, Grage

, Hagberg

, Radstrom

. Impact of DNA polymerases and their buffer systems on quantitative real-time PCR. J Clin Microbiol, 2004; 42:408–411.

27.

, Brovko

, Griffiths

. Influence of phage population on the phage-mediated bioluminescent adenylate kinase (AK) assay for detection of bacteria. Lett Appl Microbiol, 2001; 33:311–315.

Bacteriophage-Based Rapid and Sensitive Detection of Escherichia coli O157:H7 Isolates from Ground Beef

Abstract

Introduction

Materials and Methods

Bacterial strains and culture conditions

Experimental contamination of ground beef samples

Selective enrichment protocol

E. coli O157 Dynabeads capture

Phage treatment and luciferin–luciferase (FastrAK™) assay to detect E. coli O157:H7

Results and Discussion

Phage sensitivity of E. coli O157:H7 isolates

Detection of group I and II isolates in artificially inoculated beef

Phage susceptibility and signal-to-noise ratio

Conclusions

Disclosure Statement

References

Phage treatment and luciferin–luciferase (Fastr_AK™) assay to detect E. coli O157:H7