| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Abstracts of Oak Ridge Posters |
(Akubio Limited, Cambridge, UK;
aaddress correspondence to this author at : Akubio Limited, 181 Cambridge Science Park, Cambridge, CB4 0GJ, UK; fax 44-1223-225336, e-mail lhuang{at}akubio.com)
Escherichia coli O157:H7 (O157) is a rare serotype of E. coli that produces large quantities of a powerful toxin that can severely damage the lining of the intestine and induce extreme diarrhea, hemorrhagic colitis, and kidney damage. People are commonly infected by food or water; because bacteria can live asymptomatically in healthy ruminant mammals, meats can become contaminated during processing (1)(2). A rapid test to identify contaminated foodstuffs, water supplies, and infected persons would assist in reducing the risk of spreading this disease.
Many assays have been developed for detection of O157 (3)(4)(5)(6). These methods are time-consuming, however, if they require bacterial culture and may be too complex and costly for use in routine analysis. Rapid confirmation PCR assays for O157 have recently been developed (7)(8), but inhibition of PCR by humic compounds present in complex matrices such as food and surface waters can lead to difficulties such as false-negative results. To address these shortcomings, we developed a rapid, specific proof-of-principle assay for the detection of E. coli O157:H7 with Resonant Acoustic ProfilingTM (RAP) technology (9).
We obtained heat-inactivated E. coli O157:H7 cells from Kirkegaard and Perry Laboratory. Five clinical E. coli isolates provided by Dr Derek Brown (Addenbrookes, Cambridge, UK) and an American Type Culture Collection E. coli 25922 were used as negative controls. Five of these isolates were live bacteria, and heat-inactivated E. coli 170044 was used as a killed control. Affinity-purified goat anti-O157 polyclonal antibodies (pAbs) were obtained from Kirkegaard and Perry Laboratory. Anti-O157 pAb was isolated from a pooled serum from goats immunized with heat-killed whole cells of E. coli O157:H7. The lyophilized anti-O157 antibody was reconstituted with 0.3 mol/L sodium phosphate buffer (0.1 mol/L NaH2PO4, 0.2 mol/L Na2HPO4), pH 7.4, to a concentration of 10 g/L, and then further diluted to 1 g/L with phosphate-buffered saline (0.2 g/L KCl, 0.2 g/L KH2PO4, 8.0 g/L NaCl and 1.15 g/L Na2HPO4, PBS), and pH 7.4. Mouse IgG was purchased from Jackson ImmunoResearch Laboratory.
We conducted RAP experiments with an automated 4-channel instrument (RAP-id 4, Akubio Ltd) that applied the principles of quartz crystal microbalance, whereby high-frequency voltage applied to a piezoelectric crystal induced the crystal to resonate, and the resonance frequency was monitored in real time (10)(11). The RAP-id 4 apparatus featured automated sample handling and 2 pairs of oscillating crystal sensors mounted in parallel microfluidic flow cells, allowing sample to flow across 4 surfaces simultaneously. Different flow paths interchanged via electronically operated valves. Sensors were standard gold-coated quartz wafers, with a carboxylic acid–terminated chemical linker layer (used to provide a suitable surface for protein immobilization) mounted in an acrylic cassette. As the sample flowed across the control and sample sensors, binding to the sample sensors was measured as a change in the resonant frequency, with the control sensor acting as a subtractive sample reference. The instrument was fitted with a thermally stable sensor-mounting block that provided temperature control, and microfluidic and electrical connections to the 2 pairs of piezoelectric sensors. Buffer flow was maintained with syringe pumps (Tecan UK Ltd) under software (RAP Workbench) control. Microfluidics formed separate flow-paths to individual flow cells, combined with a common flow path split to address the 4 flow cells simultaneously.
Before the sensors were exposed to bacteria, we prepared the sensor surfaces by using conventional amine coupling chemistry to immobilize anti-O157 pAbs onto the sample sensor surface and mouse IgG onto the control sensor surface. Sensor surfaces were activated with a 1:1 mixture of 400 mmol/L 1-ethy 1–3-[3-dimethylaminopropyl] carbodiimide hydrochloride and 100 mmol/L N-hydroxysuccinimide, prepared in 0.22 µm filtered deionized water, and mixed immediately before use (final concentrations; 200 mmol/L 1-ethy 1–3-[3-dimethylaminopropyl] carbodiimide hydrochloride and 50 mmol/L N-hydroxysuccinimide). This mixture was injected simultaneously across 4 sensor surfaces for 3 min at a flow rate of 25 µL/min. We prepared antibodies for immobilization at 50 mg/L in immobilization buffer comprising 10 mmol/L sodium acetate, pH 4.5 and injected them simultaneously across separate sensor surfaces for 3 min at a flow rate of 25 µL/min. Nonreacted N-hydroxysuccinimide esters were then capped with 1 mol/L ethanolamine prepared in 0.22 µm filtered deionized water, pH 8.5. Running buffer between sample injections was HEPES-buffered saline, pH 7.4, containing 10 mmol/L HEPES, 150 mmol/L NaCl, 3.4 mmol/L EDTA, and 0.005% Tween 20, pH 7.4, at a flow rate of 25 µL/min.
Binding interactions between E. coli O157:H7 and antibodies were monitored in real time. Both direct binding (bacteria only) and sandwich assays (bacteria followed by antibody) were constructed for detection of E. coli O157:H7 with anti-O157 antibodies covalently attached to activated sensors. The binding signal was quantified as the frequency change at the end of the contact time with the sample. Nonspecific binding was assayed by passing different E. coli controls over the anti-O157 antibody-immobilized sensor. We used prereconstituted E. coli O157:H7 and E. coli 170044 control samples to investigate the assay sensitivity. Because these reconstituted samples had been stored at –80 °C for >1 year, a centrifugation step (10 s at 2400g) was performed to remove large particulate aggregates, and supernatants were reserved for RAP assay experiments. A sample of each supernatant was gram stained and then counted with a hemocytometer to give a stock concentration (bacteria/mL).
E. coli O157:H7 from the supernatant, prepared as described above, was diluted in HEPES-buffered saline and then passed over anti-O157 antibody-coated sensors for 10 min at a flow rate of 20 µL/min. The dissociation of anti-O157 bound to bacteria O157 was monitored by passing HEBES-buffered saline for 3 min at a flow rate of 20 µL/min. The bound complex dissociates over time when no analyte in the solution phase is present in the buffer wash. Subsequent exposure to further anti-O157 antibodies enhanced the specific signal from the direct binding of E. coli O157:H7. A concentration of 50 mg/L of anti-O157 antibody diluted in HEBES-buffered saline buffer was injected onto sensors for 3 min at a flow rate of 25 µL/min. After a dissociation period of 3 min, sensor surfaces were regenerated to remove the secondary anti-O157 pAbs layer with a 1 min injection of 100 mmol/L HCl followed by a 30 second injection of 20 mmol/L NaOH.
Six dilutions of E. coli O157:H7 supernatant (20, 60, and then log-fold to 600 000 bacteria/mL) were injected onto anti-O157 immobilized sensors. In analyzed samples, the signals generated by direct binding or sandwich assay decreased with increasing dilution of E. coli O157:H7. The minimum measurable resonant change was 6000 bacteria/mL for direct O157 binding and 600 bacteria/mL for the sandwich assay (Table 1
, Fig. 1A
). Nonspecific binding for both assay formats was determined by passing E. coli controls (170044 at 600 000 bacteria/mL and other live bacteria at 107 bacteria/mL) over the anti-O157 immobilized sensor (Table 1
, Fig. 1A
). The E. coli O157:H7 supernatant sample at 600 bacteria/mL produced a mean frequency change in the sandwich assay that was >3-fold greater than that observed for the control 170044 isolate at a 1000-fold greater concentration. At higher concentrations of O157, specificity for selection of O157 over the control 170044 was observed for both the direct binding and sandwich assay formats. At 600 000 bacteria/mL (supernatant sample), we observed a 25-Hz signal for the control 170044 vs a 373-Hz change for the specific signal in the direct binding assay (a specific-to-nonspecific signal of 15:1). In the sandwich assay at the same concentration, nonspecific binding was 0.4% of the specific signal. The ratio of the specific to nonspecific signal was 258; 17-fold greater than that observed for the direct-binding assay.
|
|
Because we used heat-killed E. coli O157:H7, the thermal inactivation and subsequent lyophilization, storage, and shipping may have caused breakdown of the intact bacteria. To determine if any free antigen was present in the analyzed sample, we filtered the E. coli O157:H7 with a 0.1 µm membrane and used the direct binding and sandwich assays to analyze the filtrate at the same dilution as the 6000 bacteria/mL sample. Specific signal was found in the filtrate by the direct binding assay (21% of unfiltered sample signal) and sandwich assay (27% of unfiltered sample signal), suggesting that nonparticulate antigen was present in the sample and was captured by specific antibody in the assay. This result clearly demonstrated that the assay was sensitive to antigen in free solution.
To further explore the clinical utility of the assay, we prepared freshly reconstituted E. coli O157:H7 and freshly thawed E. coli 170044 whole-cell suspensions by gram staining and hemocytometer counting and added the E. coli, at a concentration 107 bacteria/mL, to undiluted horse serum, human urine, and HEBES-buffer saline buffer. We analyzed the suspensions by direct binding and sandwich assay as described above. This preliminary study showed that E. coli O157:H7 was clearly detectable in undiluted matrix whereas the signal for the E. coli 170044 control was negligible. Compared with the result in buffer, the signal for O157 in serum was decreased by 
15% for the direct binding assay (Table 1
) and by 80% for the sandwich assay (Table 1
, Fig. 1B
) but was unaffected by urine (Fig. 1B
). In the latter sample matrix, the signal intensity was within 10% of that obtained in buffer. The amounts of nonpecific binding with the 170044 control in the enriched serum and urine were similar to those found in buffer, suggesting that the undiluted complex matrices did not contribute significantly to the amount of nonspecific binding.
Acknowledgments
This work was funded by National Institute of Allergy and Infectious Diseases (grant number: U01 AI061243-02). The authors thank Dr. Derek Brown, Head of Clinical Microbiology, Addenbrookes Hospital, Cambridge, UK, for providing the control clinical E. coli isolates.
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
) and sandwich assay (
) vs signals for control E. coli 170044 supernatant in direct binding assay (
) and sandwich assay (
), n = 3. (B), representative binding curves (1 example from 5 repeats) for specific binding signals for E. coli O157:H7 in buffer (