Quartz crystal microbalance (QCM) as biosensor for the detecting of Escherichia coli O157:H7

Affinity purified antibodies E. coli O157:H7 were purchased from Abcam Company, UK. Protein A-soluble, 16-mercaptohexadecanoic acid (MHDA), 1-ethyl-3- (3-dimethylaminopropyl) carbodimide (EDC), ester N-hydroxysuccinimide (NHS), photphate PBS pH 7,4, bovine serum albumin (BSA) were supplied from Sigma Aldrich (USA). And ethanol, NaOH, acetone HCl, H₂SO₄ (98%), H₂O₂ were purchased and were used without treatment from Merck Company (Germany). All the reagents used were AR grade.

We applied a QCM system designed and fabricated by ICDREC-VNUHCM (figure 1) for this study, controlled by a laptop under Windows environment and connected with 5 MHz QCM devices provided by Stanford Research Systems Company. In addition, FE-SEM-MX-51(OLYMPUS Company, Japan) and atomic force microscopy (AFM, Model 5500 AFM system, Agilent Company, USA) were employed to analyze the surface of QCM biosensor.

Zoom In Zoom Out Reset image size

E. coli O157:H7 as target bacterium was supplied by the Pasteur Institute in Ho Chi Minh City, Vietnam. The bacterial concentration was determined by the conventional surface plating-count method. The culture was then heated in a 100 °C water bath for 15 m to kill all the bacteria, and diluted to the desired concentrations with PBS for further use.

2.3.1. Self-assembled monolayer (SAM) method-based immunosensor [14]

2.3.2. Electrochemical characterization of the gold surface on QCM with SAM

2.3.3. Protein a method-based immunosensor

The holder which had the antibodies-treated QCM sensor was a fitted 5 MHz QCM system. Then, the sensor was added with 1 ml PBS solution, while the frequency shift caused by the combination was collected until the curve reached a plateau. During the E. coli O157:H7 detection process, 1 ml of 10¹–10⁷ colony forming units CFU ml⁻¹ bacterial suspension was added into the detection cell for 1 h.

In this work we applied SAM method for the protein linkage interface and used MHDA, a long chain carboxylic acid terminating alkanethiol which was proved to be more stable than other shorter chains [21]. Besides, MHDA in the function of an oriented monolayer on gold surface of QCM sensor was shaped through the strong Au–thiolate bond. In additon, the co-addition of EDC and NHS will improve the stability of the linker compounds by activating the MHDA monolayer and will conjugate E. coli O157:H7 antibody by replacing the active NHS esters through amide bonds [22]. As shown by the Sauerbrey equation, the frequency shift will be deduced from the mass change on the surface. We can calculate the amounts of molecules of each layer on gold surface QCM sensor according to the mass as described by equation [20]

$$\begin{eqnarray}&&\Delta m=-\frac{A\;\Delta F}{2.26\times {{10}^{-6}}{{F}^{2}}}\cdot \end{eqnarray} \tag{ 2 }$$

Besides, the amounts of molecules N can be calculated by mass as shown by the following equation

$$\begin{eqnarray}&&N=\frac{\Delta m}{A{{M}_{w}}},\end{eqnarray} \tag{ 3 }$$

with molecular weight M_w and Avogadro's constant A (6.023 × 1023 mol⁻¹). Substituting expression (2) into equation (3) we obtain

$$\begin{eqnarray}&&N=-2.664\times {{10}^{29}}\frac{A\Delta F}{{{M}_{w}}{{F}^{2}}}.\end{eqnarray} \tag{ 4 }$$

The result derived from equation (4) and presented in table 1 showed that frequency shift of QCM crystal changes after each treated step. Based on the observed frequency shift the mass loaded onto the surface crystal was calculated. The frequency shift of MHDA-treated crystals is 322 Hz, equivalent to 3.853 × 10¹⁵ MHDA molecules attached onto electrode, density was 11.983 × 10¹³ MHDA molecules/mm², i.e. about 57 ng MHDA mm². The mass of MHDA attached onto gold electrode is 3.6 times as many as in the study of Wang's group [20]. It is explained that the active area of the Au electrode in two studies are different. Our study used a device with the basic frequency of 5 MHz, while Wang and Su [14, 20] used an 8 MHz device. Thus, after immersion in solution of MHDA, density of this molecule is disposed regularly and there were many attached molecules. This means that the immersion time (24 h) was suitable. This result was in agreement with the result recently reported in [14].

Figure 2 indicated that after the MHDA-NHS/EDC treatment of the crystal for 2 h, the difference of its surface in comparison with the gold surface of QCM sensor was almost negligible. But the frequency shift of the MHDA-NHS/EDC treated crystal was very low (97/322 ∼ 30.12%). It means that NHS/EDC quantity was not more than that of attached MHDA. For example, with the frequency shift of 97 Hz, equivalent to 2.889 × 10¹⁵ NHS molecules attached onto electrode, density was 8.987 × 10¹³ NHS molecules/mm², about 17.2 ng NHS mm⁻², ratio NHS/MHDA of 75.5% shown that the activating efficiency of NHS is about 75.5%, namely almost 3 MHDA molecules could be activated by 2 NHS molecules (in table 1). Besides, the frequency shift of the immobilized antibodies crystal was about 205 Hz, equivalent to 4.636 × 10¹² antibodies anchored. It means that 800 MHDA molecules could be immobilized by 1 antibody. Density of antibody was 1.442 × 10¹¹ antibodies/mm², and hence about 36.3 ng mm⁻². Antibody attachment is lower than MHDA and NHS. N_Antibodies/MHDA is 0.12%, as good as the ratio in the study of Wang (0.14%). The result of this experiment in QCM sytem designed and fabricated by ICDREC is as good as the result of Wang [20].

Zoom In Zoom Out Reset image size

Analyzing the low immobilization efficiency, there are two possible reasons to be considered. First, the volume of anti-E. coli O157:H7 antibody is more gigantic than that of MHDA (about 520 times). Second, many active NHS esters would be hydrolyzed by H⁺ ions in the reagent during the course of immobilization reaction. It is important to adjust the pH value of the antibody solution to the alkalescent level so the OH^- ions could neutralize the H⁺ ions and inhibit the hydrolization of active NHS ester [20].

Figure 3 shows the CV of bare gold QCM electrode and SAMs coated gold QCM electrode in 5 mM potassium ferrocyanide at a potential scan rate of 50 mV s⁻¹. It can be seen from the figure that the bare gold electrode shows a typical CV for the redox couple where the electron transfer reaction is under controlled diffusion. In contrast, the monolayer of the SAM layer modified gold electrode does show a weak peak in the voltammogram since the redox reaction is significantly blocked by the monolayer.

Zoom In Zoom Out Reset image size

Figure 4 shows the impedance plots (Nyquist plots) of the monolayer of SAMs on gold QCM surface in potassium ferrocyanide. The EIS was carried out at a formal potential of [Fe(CN)₆]^3−/4− redox couple. It can be seen from the figure that the bare gold electrode shows a very small semicircle at high frequency region, and a straight line at low frequency region indicates that the electron transfer process of the redox couple is under controlled diffusion. On the other hand, the SAM modified electrode shows the formation of semicircle in the entire range of frequency used for the study, implying a good blocking behavior and complete charge transfer control for the electron transfer process. A very large semicircle obtained in the case of SAM of antibody on gold QCM surface compared to other SAM indicates a high charge transfer resistance and hence an excellent electrochemical blocking ability of the SAM.

Zoom In Zoom Out Reset image size

Atomic force microscopy (AFM) allows characterizing the organic thin films at a molecular resolution. The surface of the waveguide was then examined by AFM. The topography image on an area of 3 μm, and an example of a roughness profile, are shown in figure 5. A set of several lines drawn in different places of the waveguide and the profile curves show that the surface has an average surface roughness (R_ms) of 2.12 nm. The topography image of the sensor surface with a monolayer after binding of MHDA is also presented in figure 5. Examination of the curves of profile at different locations shows that the R_ms is of the same order of magnitude as that of the initial surface, about 5 nm. AFM confirms that the most uniform layer is obtained for a binding time of about 24 h in determined experimental conditions. Thus, for periods of the binding of 24 h, the R_ms is 1.79 nm. This result suggests that the molecules of MHDA bind uniformly on the roughened surface, as well as in the recesses on the bumps, and there was little aggregate.

Zoom In Zoom Out Reset image size

The results shown by FE-SEM (figure 6), AFM images, electrochemical characterization and frequency shift after each treated step indicated that SAM and anchorage of anti-E. coli O157:H7 antibodies have been successfully performed on a QCM device.

Zoom In Zoom Out Reset image size

We measure the frequency shift of QCM immersed in PBS solution for 2 h. At first, there is oscillation because of action on the electrode of PBS. After the establishment of stable surface, the shift barely changes with time. Hence, the baseline in PBS was determined. Results of measurement with PBS show that there was no frequency shift after the establishment of stability (<1 Hz) as antibodies do not capture any antigen. There was no increase of mass on the surface of QCM.

Figure 7 shows the result of the test of E. coli O124 suspension in 10⁶ CFU ml⁻¹ concentrations, when there was strong frequency change during the first 1000 s. After the time moment when measured system and suspension became stable, the frequency shift reached the baseline (influence of PBS solution), and the frequency did not change. This phenomena should be explained as specific antibodies have not captured any antigen in the suspensions, the mass increase on the SAM layer surface did not take place and there was not any frequency shift. For two remaining control bacteria (Salmonella typhimurium and Bacillus subtilize), results are as good as for E. coli O124. It can be concluded that antibody E. coli O157:H7 is specific.

Zoom In Zoom Out Reset image size

Figure 8 shows that in the entire working range of 10¹–10⁷ CFU ml⁻¹ of E. coli O157:H7, the higher the concentration, the greater the sensor responses. However, the cell concentration was 10¹ CFU ml⁻¹, the temporal response curves could not distinguish from the baseline of negative control (about 4.67 Hz). Because this concentration is very low, bacteria numbers captured by specific antibodies are not enough to make a remarkable frequency change.

Zoom In Zoom Out Reset image size

Running a sample solution of 10² and 10³ CFU ml⁻¹ of E. coli O157:H7, frequency shift was about 15 Hz and 34.67 Hz, respectively. This result allows QCM system (figure 1) to be used for qualitative and quantitative analysis of cell concentration in solution. Besides, testing a suspension of 10⁴–10⁷ CFU ml⁻¹ of E. coli O157:H7 for 2 h resulted in frequency decreases which change, respectively, with each other concentration. The higher the cell concentration, the greater the frequency shifts. As more E. coli O157:H7 bacteria have been captured by specific antibodies, the more the mass covered the surface of the Au electrode, and then the more frequency shift of measurement decreased. It takes 50 min to determine frequency shift. This result was as good as the result recently reported [20, 23, 24].

Frequency shift when running a sample solution from 10² to 10⁷ CFU ml⁻¹ of E. coli O157:H7 can be distinguished. It means that QCM system, made in ICDREC-VNUHCM, acts well with a detection range of 10²–10⁷ CFU ml⁻¹. This also proves that sensitivity of the QCM system is similar to few other systems [14, 25].

A protein A based immunosensor has been developed for detection of E.coli O157:H7 [19]. The SAM-based immunosensor was compared with the protein A in QCM system which was fabricated in ICDREC. As shown in table 2, when testing a sample solution of E.coli O157:H7 in three different concentrations, the frequency shift signals observed by MHDA method are more obvious than that by protein A. It would mean that the peptide binds between specific antibodies and NHS/EDC-MHDA SAM layer are better than those between specific antibodies and protein A.

Figure 9 showed that frequency shift values at any concentration also have remarkable difference. Frequency shift obtained from MHDA-SAM layer method is 2, 1.6 and 2.1 times (n = 3), respectively better than that from protein A method at 10², 10⁴ and 10⁶ CFU ml⁻¹. This problem demonstrated that MHDA-SAM layer is more sensitive than protein A method.

Zoom In Zoom Out Reset image size

Moreover, there is a greater number of antibodies which anchored onto NHS/EDC-MHDA SAM layer than onto protein A. The frequency shift is 322 Hz with MHDA–SAM layer and from 54 to 83 Hz with protein A when the antibodies were imobilized. Therefore, MHDA-SAM method attached antibodies 4.7 times higher than the protein A method.