Construction of portable electrochemical immunosensors based on graphene hydrogel@polydopamine for microcystin-LR detection using multi-mesoporous carbon sphere-enzyme labels

Cuifen Gana, Zihong Suna, Li Lingb, Zuyu Hea, Hongtao Leib and Yingju Liu*a
aDepartment of Applied Chemistry, College of Materials & Energy, South China Agricultural University, Guangzhou 510642, China. E-mail: liuyingju@hotmail.com; Fax: +86-20-85282366; Tel: +86-20-85280319
bThe Guangdong Provincial Key Laboratory of Food Quality and Safety, College of Food Science, South China Agricultural University, Guangzhou 510642, China

Received 26th March 2016 , Accepted 12th May 2016

First published on 12th May 2016


Abstract

As one of the most common toxins of the toxigenic Cyanobacteria, microcystin-LR (MC-LR) has raised global concerns in water monitoring, environmental detection, toxicology research and epidemiology research. In this work, a portable electrochemical MC-LR immunosensor was fabricated for the detection of MC-LR. Firstly, graphene hydrogel@polydopamine (GH@PDA) was prepared by self-polymerization of dopamine on graphene oxide, followed by a hydrothermal reaction. The modification of polydopamine (PDA) not only acts as an important modifier of the inner structure of the hydrogel, but can also prevent graphene aggregation during further treatments. The transformation to GH@PDA by the hydrothermal reaction was a green synthesis, and this process can be also used to capture biomolecules based on nucleophilic reactions or Schiff base reactions between the catechol groups of PDA and the amino groups of antigens. Secondly, mesoporous carbon spheres, which were loaded with thionine as the electron mediator followed by captured Au nanoparticles, were used to immobilize horseradish peroxidase and secondary antibodies, providing a strong electrochemical response. Using this dual-amplification strategy, the immunosensor can detect MC-LR as a competitive method in the range from 0.01 to 10 μg L−1, with a detection limit of 0.0097 μg L−1. This method also showed good accuracy, acceptable precision and reproducibility. The proposed strategy provides biocompatible immobilization and sensitive recognition for the detection of chemical compounds, pollutants and food contaminants and for clinical diagnosis.


1. Introduction

With the rapid development of industrialization, the eutrophication of lakes, ponds and rivers, causing outbreaks of blue-green algae and serious pollution, is increasingly common. In the family of blue-green algae, microcystins (MCs) are potential hepatotoxins that are produced by Cyanobacteria. MCs exposure exerts great harm on the reproductive system of fish, deteriorates the quality of eggs and sperm and has further effects on the early developmental stage of fish.1 Especially, MCs can accumulate in aquatic organisms at high levels, representing a health hazard to animals and humans, because they have strong affinity to serine/threonine protein phosphatases, thereby acting as an inhibitor of enzymes.2 Microcystin-LR (MC-LR), containing leucine and arginine in the main variant positions, is the most frequent and toxic variant among the identified 80 MCs to date.3 Due to its harmful impact on human health, in 1997, the World Health Organization (WHO) set a maximum permitted level of 1 μg L−1 of MC-LR in drinking water and a daily intake value of 0.04 μg kg−1 MC-LR/body weight.4

To date, several traditional analytical methods, including protein phosphatase inhibition assays (PPIA),5 high performance liquid chromatography (HPLC)6 and enzyme-linked immunosorbent assay (ELISA),7 have been reported for the detection of MC-LR. For instance, Ma et al. developed C-18-functionalized magnetic silica nanoparticle (Fe3O4@SiO2@C-18 MNPs) based magnetic solid phase extraction followed by HPLC-diode array detection for the detection of MC-LR in reservoir water samples in the range of 0.1 to 10.0 mg L−1 with a detection limit of 0.056 mg L−1,8 while Liu et al. established a highly sensitive ELISA and an immunochromatographic assay for the semi-quantification of MCs.7 Although PPIA allows the sensitive detection of MC-LR in water through radioisotopic, colorimetric or electrochemical responses, false positives can easily occur if the enzyme is inhibited by other compounds in the sample matrix.9 HPLC, although it is more precise and allows the identification of individual variants, nonetheless requires expensive equipment, complex procedures and trained personnel. Especially, various structurally similar compounds of MCs increased the detection difficulty of HPLC. Conventional ELISA is usually based on the adsorption change of tetramethylbenzidine (TMB) in the presence of H2O2; however, its sensitivity can be influenced by the characteristics of ultraviolet-visible detection. Thus, it is still greatly important to establish a sensitive and specific detection method for MC-LR.

Electrochemical immunosensors have been widely developed to determine various trace amounts of targets, including pesticides,10 environmental pollutants11 and some cancer markers12 based on the highly specific properties of immunoreaction. These immunosensors also boast promising advantages, including feasible miniaturization, sensitivity and cost-effectiveness. Recently, nanomaterials have been widely used in the preparation of immunosensors, including substrate modification and label magnification.13 Graphene, a two-dimensional carbon material consisting of a single-layer of sp2 carbon atoms, has attracted great scientific and technological interest due to its unique properties, such as high surface area, excellent electrical conductivity and strong mechanics.14 Usually, the preparation of graphene starts with chemical exfoliation of graphite into graphene oxide (GO), followed by effective reduction of GO using many reducing agents, including hydrazine,15 NaBH4 (ref. 16) and hydroquinone.17 Recently, a hydrothermal process for the controlled synthesis and structural adjustment of nitrogen-doped graphene hydrogels has been reported using organic amines, including urea18 and ethylenediamine.19 However, some reagents are corrosive, toxic or explosive, thus raising serious safety and environmental concerns. Currently, due to its good adhesion, hydrophilicity and biocompatibility, polydopamine (PDA), which can be easily prepared from the self-polymerization of dopamine, has been used in sensor modification.13,20 Herein, PDA was selected to modify GO, not only acting as an important modifier for its inner structure, but also preventing graphene aggregation during further treatments. The modified GO was then transformed into graphene hydrogel@polydopamine (GH@PDA) by a hydrothermal reaction; this results in a homogenous solution with high dispersion that can also be easily used to capture biomolecules based on nucleophilic or Schiff base reactions between the catechol groups of PDA and the amino groups of proteins.13

Additionally, the signal detection of the label for immunosensors is important. Various nanomaterials, including Au nanoparticles,21 Au nanorods,22 carbon nanotubes,23 carbon nanospheres24 and SiO2 nanospheres,11 have been used to prepare multi-enzyme nanocomposites. Recently, mesoporous carbon spheres (MCSs) have received extensive attention due to their mesoporous shell-hollow inner space shape, high chemical and thermal stability, low density and large surface area. In addition, MCSs exhibit more advantages in mass diffusion and ion transportation; especially, when electron mediators are introduced, they can be used in capacitors, biosensors or electrocatalysis.25,26 The template-based method has been widely employed to produce hollow mesoporous carbon spheres with controlled structures and inherited porosity. Due to the tedious, multistep procedures and even non-environmentally friendly conditions, the hard templates (e.g. SAB-15, CMK-3, MCM-48) which were generally used to introduce mesoporous structures into carbon frameworks27 have now been replaced by soft templates using cross-linked polymeric materials, including resorcinol-formaldehyde and phloroglucinol-formaldehyde as carbon-yielding components and amphiphilic block copolymers, including F127 and P123, as pore forming components.28 Herein, MCSs were prepared by a soft template and then loaded with thionine, which is not only an electron mediator due to its excellent electroactivity,29 but also contains thiol groups that can capture Au nanoparticles by Au–S bonds.30 As is known, Au nanoparticles are very popular in the immobilization of proteins;13,22 it is convenient to use these nanocomposites to prepare multi-enzyme labels, which can greatly increase the detection signals.

In this work, nitrogen-doped graphene hydrogel, GH@PDA, was prepared by an easy polymerization of dopamine on graphene and immediate hydrothermal treatment. After characterization with infrared spectroscopy (IR), transmission electron microscopy (TEM), and Raman spectroscopy, GH@PDA, due to its abundant functional groups, high surface area and electron conductivity, was used as the sensor substrate to immobilize antigens. After that, an electrochemical immunosensor was fabricated to sensitively and selectively detect MC-LR using MCSs-loaded multi-enzymes as the sensing label.

2. Experimental

2.1 Materials and apparatus

Graphite was purchased from Alfa Chemicals Co. Ltd, while dopamine and horseradish peroxidase (HRP) were bought from Aladdin Chemistry Co. Ltd and Guangzhou Qiyun Biotechnology Co. Ltd, respectively. MC-LR-BSA (Ag, primary concentration 4 mg mL−1) and MC-LR-antibody (Ab, primary concentration 4 mg mL−1) were a gift from the Guangdong Provincial Key Laboratory of Food Quality and Safety, while HRP-labeled secondary goat anti-rabbit antibody (Ab2, 400 μg mL−1) was obtained from Santa Cruz. F127, MC-LR and its related interferences, including MC-RR, MC-YR and nodularin, were all purchased from J&K Chemical Company. All other materials were of the highest available grade.

The supporting electrolyte buffer in the electrochemical detection experiments, 1/15 M phosphate-buffered solution (PBS) with various pH values, was prepared by mixing 1/15 M stock solutions of KH2PO4 and Na2HPO4 at different ratios. The washing buffer solution in the immunoreaction was prepared by dissolving 0.05% Tween-20 in 0.01 M PBS (PBST, pH 7.4), while the blocking buffer solution was composed of 0.5% skim milk in 0.01 M PBS (pH 7.4). The MC-LR stock solution was prepared by dissolving MC-LR in methanol, and a series of MC-LR solutions with different concentrations were made by diluting the stock solution with PBST.

The morphologies of the samples were characterized by transmission electron microscope (TEM, Tecnai 12, FEI, Holland) at the Instrumental Analysis & Research Center at South China Agricultural University. The N2 adsorption–desorption analysis was performed using a Micromeritics Gemini instrument (Gemini VII 2390, America), and the infrared spectra were acquired by Fourier transform infrared spectroscopy (FTIR, Bruker VERTEX 70). X-ray refraction diffusion (XRD) patterns were acquired on a Rigaku Ultima IV X-ray diffractometer. Electrochemical experiments were performed in a three-electrode system that was connected to a CHI660D electrochemical workstation (Chenhua Instruments Co. Ltd., Shanghai, China). The fabricated screen-printed carbon electrode (SPCE, Φ = 2 mm) was used as the working electrode, while a carbon electrode and an Ag/AgCl electrode served as the counter and reference electrodes, respectively.

2.2 Preparation of PDA/GH

Graphene oxide (GO) was prepared from natural graphite powder using a modified Hummers method.31 Then, 30 mg GO was dispersed in 60 mL water and ultrasonicated for 60 min. After that, 15 mg hydrochloride dopamine and 73 mg Tris were successively added, and the mixture was ultrasonicated for 10 min. Then, the solution was transferred to a 60 °C water bath, stirred magnetically for another 24 h, and adjusted to pH 6.0 using HCl. After this lengthy heat treatment, a solid pad (partly reduced rGO@PDA) was obtained by centrifugation at 10[thin space (1/6-em)]000 rpm and washed until the supernatant was colorless. For further reduction of rGO, rGO@PDA was transferred into an autoclave with a volume of 30 mL water. Hydrothermal treatment of the mixture was performed at 180 °C for 24 h. Then, the autoclave was naturally cooled to room temperature and the product, GH@PDA, was removed. Meanwhile, the common rGO was prepared by the reduction of GO with hydrazine.

2.3 Preparation of MCSs

The MCSs were prepared using triblock polymer (F127) as the soft template and water soluble phenol-formaldehyde resin as the carbon precursor as follows. Firstly, the following solutions were prepared. Solution A: 0.768 g of F127 was dispersed in 15 mL water for 24 h. Solution B: 0.6 g phenolic acid was dissolved in 15 mL 1 mol L−1 NaOH and heated at 70 °C for 0.5 h. Secondly, solution A was slowly added to solution B and heated at 66 °C for 2 h. After 50 mL water was added, the mixture was heated for 16 h; thus, resol–F127 composite was prepared. Thirdly, a mixture containing 17.7 mL of resol–F127 composite and 56 mL of water was hydrothermally treated at 130 °C for 24 h. After centrifugation, washing and drying, the solid was crushed into powder, heated to 700 °C at 1 °C min−1, and calcinated at 700 °C for 3 h. Finally, it was cooled to room temperature; thus, the MCSs were obtained.

2.4 Preparation of multi-HRP-(MCSs/Thi@AuNPs)-Ab2

Gold nanoparticles (AuNPs) were prepared as follows. All glassware was thoroughly cleaned with freshly prepared aqua regia (HNO3[thin space (1/6-em)]:[thin space (1/6-em)]HCl = 1[thin space (1/6-em)]:[thin space (1/6-em)]3, by volume) and rinsed extensively with doubly distilled water. A twenty milliliter sample of aqueous 0.01% HAuCl4 was placed in a 50 mL flask. After the solution was brought to a boil under stirring and N2 flow, 0.4 mL of 1% sodium citrate solution was added; the colour of the solution changed from pale yellow to wine-red. After boiling for another 10 min, the solution was stirred for another 15 min at room temperature. Finally, the product was stored at 4 °C for further use.

To improve the dispersion of the MCSs in water, the MCSs were treated with H2SO4/HNO3 (1[thin space (1/6-em)]:[thin space (1/6-em)]1, by volume) for 3 h to obtain sufficient carboxylic groups. Then, 10 mg MCSs and 10 mg thionine were added to 10 mL water, ultrasonicated for 30 min and stirred for 12 h. After centrifugation at 10[thin space (1/6-em)]000 rpm, washing with water, and redispersing in 10 mL water, MCSs/Thi was obtained. Successively, 1 mL MCSs/Thi and 2 mL fresh AuNPs were mixed, stirred at 37 °C for 3 h and centrifuged at 8000 rpm; then, the solids were collected as MCSs/Thi@AuNPs. Finally, multi-HRP-(MCSs/Thi@AuNPs)-Ab2 was prepared by adding MCSs/Thi@AuNPs to 1 mL PBS 7.4 solution containing 5 μg mL−1 Ab2 and 3 mg mL−1 HRP, shocking sufficiently to obtain a homogenous solution, and stirring for another 3 h at room temperature. After centrifugation, the product was redispersed in 1 mL PBST 7.4 and stored at 4 °C before use. Multi-HRP-MCSs-Ab2 was prepared by the same method, using MCSs instead of MCSs/Thi@AuNPs.

2.5 Fabrication of the immunosensor

The SPCE was washed thoroughly with water and then activated by oxidation at 1.7 V for 4 min in 1/15 M PBS (pH 7.4). As shown in Scheme 1, 3 μL of 0.5 mg mL−1 dopamine solution (Tris) was deposited on the electrode surface, followed by the addition of 5 μL 0.5 mg mL−1 GH−1@PDA. To increase the immobilization of biomolecules, AuNPs were further deposited on GH@PDA/SPCE by continually scanning in 0.5 mM HAuCl4 + 0.5 M H2SO4 in the range from −0.3 to 0.3 V for 5 cycles at a scan rate of 50 mV s−1. Then, the AuNPs/GH@PDA/SPCE surface was rinsed with water and dried with N2. Subsequently, 5 μL Ag was decorated onto the activated AuNPs/GH@PDA/SPCE for 1 h at 37 °C. Finally, 5 μL blocking buffer was incubated to block possible remaining active sites for 1 h at 37 °C. After each step, the modified electrode was thoroughly rinsed with PBST. The obtained immunosensor was stored at 4 °C before use.
image file: c6ra07881h-s1.tif
Scheme 1 Schematic of the MC-LR immunosensor.

The detection of MC-LR was based on the competitive combination between immobilized Ag and MC-LR in a solution with a definite amount of Ab. Therefore, MC-LR solutions with different concentrations were mixed with Ab; then, 5 μL mixture was placed onto the electrode surface at 37 °C for 1 h. After washing with PBST, the electrode was incubated with 5 μL multi-HRP-(MCSs/Thi@AuNPs)-Ab2 at 37 °C for 1 h. The electrochemical experiment was performed by recording cyclic voltammograms (CV) between −0.6 and 0.6 V in 1/15 M PBS containing 1 mM hydroquinone (HQ). The signal was recorded by the cathodic current change before and after the addition of 1.25 mM H2O2.

3. Results and discussion

3.1 Characterization of GH@PDA/SPCE

Here, GH@PDA was prepared by a two-step process, including the formation of PDA on GO and hydrothermal reduction at high temperature. Fig. 1A shows the TEM image of graphene oxide, which has a flat paper-like structure; the GO remained smooth when it was modified with PDA, suggesting that the PDA layer was thin and even, as shown in Fig. 1B. After it was hydrothermally reacted at high temperature, the surface of GH@PDA wrinkled slightly, as shown in Fig. 1C, which can provide a high surface area for protein immobilization. Fig. 1D shows the XRD patterns of the samples before and after hydrothermal reduction. GO has a sharp peak at 10.7°, corresponding to an interlay distance of 0.82 nm. However, for GH@PDA, this peak completely disappears, and a relatively broader peak centered at around 24.7° is observed. Additionally, a diffraction peak at 2θ around 43°, which corresponds to the (100) reflection of graphene sheets, can be found in GO (curve a); this peak is smaller in rGO (curve b) and disappears in GH@PDA (curve c), indicating a poor degree of graphitization.32 Comparing the FT-IR spectra (Fig. 1E) of rGO (curve b) with GO (curve a), the strong band at 1710 cm−1 due to the C[double bond, length as m-dash]O stretching mode nearly disappears; however, this band can be found in GH@PDA (curve c), suggesting the existence of PDA and also the incomplete reduction of GO. Meanwhile, the wide peak at 1120 cm−1 due to the heterocyclic N–H in plane deformation breathing and the peak at 1570 cm−1 due to the C–N stretching vibration can be found in GH@PDA; thus, these residual amino groups can be used for the subsequent deposition of Au nanoparticles and the immobilization of proteins.33 Finally, Raman spectra were employed to determine the significant structural changes from GO to GH@PDA. In Fig. 1F, both samples display two bands at around 1355 and 1597 cm−1, which correspond to the D and G bands, respectively. The integrated intensity ratio (ID/IG) of these two Raman bands is usually used to reflect the ordered and disordered crystal structures of graphene. As expected, the ID/IG ratio of GH@PDA was 0.87 (curve b), slightly higher than that of GO (0.82, curve a), suggesting that the modification of PDA and the hydrothermal treatment alter the structure of GO and introduce some structural defects.34
image file: c6ra07881h-f1.tif
Fig. 1 TEM images of GO(A), rGO@PDA (B) and GH@PDA (C); XRD spectra (D) of GO(a), rGO@PDA (b) and GH@PDA (c); IR spectra (E) of GO (a), rGO (b) and GH@PDA (c); and Raman spectra (F) of GO (a) and GH@PDA (b).

To further compare the electron transfer, GO/SPCE, rGO@PDA/SPCE and GH@PDA/SPCE were prepared by depositing these substances on SPCE, and CVs in HQ solution were recorded (Fig. 2A). The current of GH@PDA/SPCE (curve c) is much higher than that of GO/SPCE (curve a) and rGO@PDA/SPCE (curve b), while the peak-to-peak potential is much lower than that of GO/SPCE and rGO@PDA/SPCE, suggesting that the electron transfer is slightly faster on GH@PDA/SPCE. Then, the influence of GH@PDA concentration was investigated (Fig. 2B). With increasing concentration, the reduction peak potential shifts to the positive direction, suggesting that the oxidation state of HQ can be readily reduced on the electrode surface. However, if the concentration of GH@PDA is much higher than 0.5 mg mL−1, the peak potential becomes stable. In addition, the cathodic peak current also gradually increased with increasing concentration of GH@PDA and stabilized slightly when the concentration was higher than 0.5 mg mL−1. Thus, 0.5 mg mL−1 GH@PDA was selected to modify the electrode surface. Finally, the immobilization ability of protein on the different electrodes was compared by immersing them in 1 mg mL−1 HRP for 1 h, after which the electrochemical response to 1 mM H2O2 was recorded (Fig. 2C). The results show that after the immobilization of HRP, AuNPs/SPCE provides a current response of 19.3 μA, while the currents for AuNPs/GO/SPCE and AuNPs/rGO@PDA/SPCE are 16.3 and 19.8 μA, respectively. However, if the electrode is modified with AuNPs/GH@PDA/SPCE, the current response is as high as 26.3 μA, suggesting that the introduction of GH can increase the immobilization ability of protein due to the relatively high surface area.


image file: c6ra07881h-f2.tif
Fig. 2 Cyclic voltammograms (A) of GO/SPCE (a), rGO@PDA/SPCE (b) and GH@PDA/SPCE (c) in 1.0 mM HQ; (B) the influence of the GH@PDA concentration on SPCE on the reduction peak potential (blue) and the reduction peak current (pink) in 1.0 mM HQ; (C) the cathodic peak current changes on AuNPs/SPCE (a), AuNPs/GO/SPCE (b), AuNPs/rGO@PDA/SPCE (c) and AuNPs/GH@PDA/SPCE (d) after the immobilization of 1 mg mL−1 HRP in 1.0 mM HQ + 1 mM H2O2.

3.2 Characterization of MCSs/Thi@AuNPs

The morphologies of the synthesized MCSs were characterized by TEM. Fig. 3A is the TEM of resol–F127 composite, where homogeneous spheres are found, suggesting that F127 and the phenolic resin have been assembled in an orderly fashion. After the carbonation of resol–F127 composite at 700 °C, the MCSs are still uniform spheres with diameters of about 80 nm (Fig. 3B), suggesting that the carbonation did not change the skeleton of the nanospheres. Provided by the N2 adsorption–desorption data in Fig. 3D, the specific surface area of the MCSs is 634.1 m2 g−1, much higher than that of resol–F127 composite (36.1 m2 g−1); meanwhile, the total pore volume of MCSs is 0.390 cm3 g−1, much higher than that of resol–F127 (0.007 cm3 g−1), suggesting that F127 is gasified during the carbonation, and thus many mesoporous pores are produced and the specific surface area increases. In addition, the pore size distributions of the MCSs were calculated by a density functional theory (DFT) model from the adsorption part of the isotherm. As shown in Fig. 3E, the pore diameters are mainly distributed from 3 to 6 nm, which can introduce thionine into the pores of MCSs by immersing them in thionine solution. Finally, AuNPs were adsorbed on the surface of MCSs by the thiol groups of thionine. From the TEM image of MCSs/Thi@AuNPs shown in Fig. 3C, AuNPs with a diameter of 15 nm are well-distributed on the MCSs. During the preparation of MCSs/Thi@AuNPs, thionine is not only an excellent electron mediator,29 but its thiol groups can also be used to capture AuNPs by Au–S bonds.30 In addition, AuNPs are very popular in the immobilization of proteins;13,22 thus, it is convenient to use such nanocomposites to prepare multi-enzyme labels, which can greatly increase the detection signals.
image file: c6ra07881h-f3.tif
Fig. 3 TEM images of resol–F127 (A), MCSs (B) and MCSs/Thi@AuNP (C); N2 adsorption–desorption isotherm (D) and pore size distribution from the adsorption branch through the DFT method (E) of MCSs.

3.3 Optimization of the detection conditions

In a traditional immunosensor, the concentrations of antigen and antibody, the incubation time and the pH of the detection system play important roles in the performance of the detection process; thus, all these influence factors were optimized. Fig. 4A shows the influence of the dilution ratio of the antibodies, while curves a and b give the response for the immunosensor without or with the addition of 1.0 μg L−1 MC-LR, respectively. For both curves, the current response (ΔI) gradually increases when the dilution ratio of the antibodies decreases from 800-fold to 100-fold. It is common that a lower dilution ratio can result in a higher concentration of antibodies in the solution and on the electrode surface, followed by a larger current response. However, for competitive immunoassays, the inhibition efficiency is influenced by the amount of antibodies. That is to say, if there are too many antibodies in the solution, the combination of the antibodies with the analyte cannot greatly influence its combination with the immobilized antigen. By comparing these two curves, the best inhibition effect was 74% at the dilution ratio of 1[thin space (1/6-em)]:[thin space (1/6-em)]300. Consequently, the optimum antibody dilution ratio was 300-fold. In addition, the dilution ratio of Ab2 was also investigated. As shown in Fig. S1, the 100-fold dilution ratio was selected as the optimum of Ab2.
image file: c6ra07881h-f4.tif
Fig. 4 The influence of the dilution ratio of Ab (A), the dilution ratio of Ag (B), the incubation time (C) and the pH of the detection solution (D).

Furthermore, the influence of Ag concentration was investigated. As shown in Fig. 4B, the current responses gradually increase when the dilution ratio decreases from 800-fold to 100-fold, while the current responses are very similar at 50-fold and 20-fold dilution ratios because this is close to the saturation of Ag. For a competitive immunoreaction, a saturated concentration of Ag cannot be selected even if it gives a high response signal, which is due to the fact that if Ag is oversaturated, the combination of the immobilized Ag with Ab cannot greatly influence the combination of the analyte with Ab, resulting in less current change. Therefore, the 50-fold dilution ratio was selected as the optimum of Ag.

Also, the incubation time plays an important role in the detection of MC-LR. As shown in Fig. 4C, with increasing incubation time, the current response rapidly increases up to 60 min. When the incubation time exceeds 60 min, the current response reaches a maximum and remains almost unchanged, indicating that Ab is thoroughly captured on the electrode. Thus, the incubation time of 60 min was selected.

Finally, the influence of the buffer pH is also essential to the sensitivity of the immunosensor, since pH affects not only the bioactivity of the antigen and antibody, but also the electrochemical reaction at the electrode surface. As shown in Fig. 4D, the current response of the immunosensor increases initially and then decreases, with the highest peak at a pH value of 7.38. From the electrode reaction, a much higher pH value will accelerate the electron transfer because it is relative to the proton, which can be seen in the reaction mechanism given below. However, if the pH value is too high, it will damage the bioactivity of the protein. Hence, the optimum pH value of the detection system was set at 7.38.

3.4 Signal amplification strategy of the immunosensor

In the electrochemical catalytic detection of HRP-related immunosensors, the following mechanisms are usually used:13,22
 
HRP (Fe3+) + H2O2 → compound (I) + H2O (1)
 
Compound (I) + HQ → compound (II) + Q (2)
 
Compound (II) + HQ → HRP (Fe3+) + Q + H2O (3)
 
Q + H+ + 2e → HQ (electrode reaction) (4)
where HQ and Q are hydroquinone and the oxidation state of hydroquinone, respectively. With the aid of H2O2, HRP can be oxidized to compound (II), followed by the oxidation of HQ to Q. The final equation is the electrochemical reaction at the electrode surface, where the reduction peak is dramatically increased by the increase in Q from the HRP-catalytic reaction.

In the preparation of the immunosensor, the signal amplification strategy was based on the modification of the sensor surface and the preparation of the multi-enzyme label. To demonstrate the function of GH in the platform, AuNPs/PDA/SPCE was constructed by directly coating PDA and AuNPs on the surface of SPCE. Comparing Fig. 5A and B, the starting currents before the addition of H2O2 are much higher on AuNPs/GH@PDA/SPCE, suggesting that this substrate provides a much higher conductivity. In addition, after the addition of H2O2, the cathodic peak current response at around −0.2 V of the immunosensor in Fig. 5A is 4.51 μA, while it increases to 6.19 μA when the substrate is modified with AuNPs/GH@PDA/SPCE (Fig. 5B). This can be explained as follows: (1) the introduction of GH can increase the specific surface area to provide many more deposition sites for AuNPs, which can increase the binding sites for Ag; (2) GH has a much higher conductivity, which can increase the electron transfer at the sensor surface.34


image file: c6ra07881h-f5.tif
Fig. 5 Cyclic voltammograms of the immunosensors with different constructions: (A) AuNPs/PDA/SPCE and HRP-Ab2, (B) AuNPs/GH@PDA/SPCE and HRP-Ab2, (C) AuNPs/GH@PDA/SPCE and multi-HRP-MCSs-Ab2, (D) AuNPs/GH@PDA/SPCE and multi-HRP-(MCSs/Thi@AuNPs)-Ab2 in 1 mM HQ before (black, solid) and after (red, dashed) the addition of 1.25 mM H2O2.

To further achieve the signal amplification, multi-HRP-MCSs-Ab2 and multi-HRP-(MCSs/Thi@AuNPs)-Ab2 were prepared as the signal label. Comparing Fig. 5C and D, the starting currents before the addition of H2O2 were much higher when using the multi-HRP-(MCSs/Thi@AuNPs)-Ab2 label, suggesting that this label provides a much higher electron transfer rate. A new pair of peaks corresponding to thionine can be found. In addition, the cathodic peak current response at around −0.2 V increased to 8.66 μA (Fig. 5C) and 23.74 μA (Fig. 5D), respectively. Firstly, MCSs can provide a much higher specific area to immobilize HRP for the electrochemical response. Secondly, the introduction of thionine in MCSs can increase the catalytic activity of the multi-enzyme label due to the electron mediator function of thionine. Thirdly, the loading of AuNPs on MCSs/Thi can not only preserve thionine from elapsing out of the MCSs, but can also provide a high binding ability for HRP and Ab2 due to the reaction of the AuNPs with protein.

3.5 Performance of the immunosensor

To evaluate the performance of the developed immunosensor, different concentrations of MC-LR standard solutions were incubated with a certain amount of Ab, and then the mixture was incubated on the sensor surface. As shown in Fig. 6A, the current response gradually decreased with increasing concentration if the immunosensor was incubated with conventional Ab2. The inset of Fig. 6A exhibits that ΔI is proportional to the logarithm of C in the range of 0.025 to 5 μg L−1, with a regression equation of ΔI = (2.11 ± 0.04) − (2.35 ± 0.05) × log[thin space (1/6-em)]C (R = 0.996), where ΔI represents the current change of the peak before and after the addition of H2O2 at every concentration and C is the concentration of MC-LR. The limit of detection, corresponding to three times the standard deviation of the blank, is estimated as 0.02 μg L−1. If the immunosensor was incubated with multi-HRP-(MCSs/Thi@AuNPs)-Ab2, as in Fig. 6B, the current response also gradually decreased with increasing MC-LR concentration, suggesting that a higher concentration causes a lower current change. However, it should be noted that the current response is much higher than that of the immunosensor with conventional Ab2 at the same concentration. In the inset of Fig. 6B, the immunosensor shows a linear response in the range of 0.01 and 10 μg L−1 as ΔI = (8.21 ± 0.18) − (7.69 ± 0.18) × log[thin space (1/6-em)]C (R = 0.995), with a detection limit of 9.7 ng L−1 (ESI). Therefore, the use of the multi-HRP-(MCSs/Thi@AuNPs)-Ab2 label can improve the behavior of the immunosensor. The characteristic properties, including the linear range and the detection limit, are also compared with earlier published work in Table S1 (ESI), showing that the immunosensor showed comparable or even better performance.
image file: c6ra07881h-f6.tif
Fig. 6 The current response of the immunosensor with MC-LR concentrations by using the conventional HRP-Ab2 (A) and multi-HRP-(MCSs/Thi@AuNPs)-Ab2 (B). Inset: the calibration curve of the immunosensor. (C) The specificity of the proposed immunosensor of 1 μg L−1 MC-LR with 1 μg L−1 interferences.

To assess the specificity of the proposed immunosensor for MC-LR, the immunosensors were incubated with samples containing some potential interferents with similar structures, such as MC-RR, MC-YR and nodularin. In Fig. 6C, the percentages of the response signal were between 96.6% and 106.6% for 1 μg L−1 MC-LR if 1 μg L−1 different interferents were added. Furthermore, when all the correlative interferents (MC-RR, MC-YR and nodularin) were mixed with 1 μg L−1 MC-LR, the percentage of the response signal was 95.1. Additionally, when the interferences of cations, anions and other compounds in natural water samples, such as SO42−, CO32−, Ca2+, Mg2+ and chlorpyrifos (a type of organophosphorus pesticide), were assessed by the same method, the percentages of their response signals were 101.2%, 99.1%, 98.5%, 99.6% and 97.4%, respectively, suggesting that the specificity of the proposed immunosensor was acceptable.

The intra-assay of the immunosensor was evaluated by analyzing three MC-LR concentration levels with one immunosensor four times, where the coefficients of variation were 1.07%, 4.90% and 4.82% at 0.05, 0.5 and 1.0 μg L−1 of MC-LR, respectively. The inter-assay of the immunosensor was evaluated by analyzing three MC-LR concentration levels with four immunosensors, where the coefficients of variation were 6.41%, 5.06% and 5.37% at 0.05, 0.5 and 1.0 μg L−1 of MC-LR, respectively. Thus, the reproducibility of the proposed immunosensor was acceptable.

The stability of the proposed immunosensor was also investigated by monitoring the substrate and the label. After Ag/AuNPs/GH@PDA/SPCE was blocked with blocking buffer and stored at 4 °C for 14 days, the signal response retained 93.5% of the initial response. Additionally, the response of the immunosensor using a 7 day-stored multi-HRP-(MCSs/Thi@AuNPs)-Ab2 label retained 96.6% of the initial signal. Thus, the proposed immunosensor process demonstrated satisfactory stability.

3.6 Preliminary application of the immunosensor

To monitor the feasibility of the application of the proposed immunosensor, a series of samples prepared by the standard addition of MC-LR to water samples from Poyang Lake in South China Agricultural University were analyzed by the proposed method. The experimental results, shown in Table 1, showed good recoveries varying from 95.0% to 110.0%, indicating that the proposed immunosensor provided a sensitive and alternative method for the detection of MC-LR in real water samples.
Table 1 Detection of MC-LR in spiked Poyang Lake water
Samples Added (μg L−1) Founda (μg L−1) RSD (%) Recovery (%)
a Each value is the average of three measurements.
1 5.0 4.90 3.45 98.0
2 1.0 1.07 4.94 107.0
3 0.5 0.49 2.89 98.0
4 0.10 0.095 1.76 95.0
5 0.01 0.011 4.30 110.0


4. Conclusions

In summary, GH@PDA nanocomposites and MCSs/Thi@AuNPs were prepared, characterized and proved to be excellent sensor surface materials and label materials with high analytical performance for MC-LR detection. The relatively high surface area of GH@PDA enhanced the immobilization capability of Ag, while the electron mediator and abundant Au nanoparticles on MCSs/Thi@AuNPs provided high amounts of HRP and Ab2 to increase the detection signal. With a competitive immunoassay format, the designed immunosensor for MC-LR showed a linear response of the current change with the logarithm of concentration in the range of 0.01 to 10 μg L−1, with a detection limit of 0.0097 μg L−1. The portable electrochemical immunosensor exhibited high sensitivity, selectivity and stability. This strategy could be a promising candidate for probing other electrochemical immunoassays for chemical analysis, pollutant analysis, food analysis and clinical diagnosis.

Acknowledgements

This work was supported by the National Scientific Foundation of China (21475047, U130214), the Science and Technology Planning Project of Guangdong Province (2013B02080007), and the Foundation for High-level Talents in Higher Education of Guangdong Province.

Notes and references

  1. W. Liu, Q. Qiao, Y. Chen, K. Wu and X. Zhang, Aquat. Toxicol., 2014, 155, 360 CrossRef CAS PubMed.
  2. J. Jia, W. Luo, Y. Lu and J. P. Giesy, Sci. Total Environ., 2014, 487, 224 CrossRef CAS PubMed.
  3. Z. Lin, H. Huang, Y. Xu, X. Gao, B. Qiu, X. Chen and G. Chen, Talanta, 2013, 103, 371 CrossRef CAS PubMed.
  4. W. Li, J. Duan, C. Niu, N. Qiang and D. Mulcahy, J. Chromatogr. Sci., 2011, 49, 665 CAS.
  5. C. Rivasseau, P. Racaud, A. Deguin and M. C. Hennion, Anal. Chim. Acta, 1999, 394, 243 CrossRef CAS.
  6. X. Guo, P. Xie, J. Chen, X. Tuo, X. Deng, S. Li, D. Yu and C. Zheng, J. Chromatogr. B: Anal. Technol. Biomed. Life Sci., 2014, 963, 54 CrossRef CAS PubMed.
  7. L. Liu, C. Xing, H. Yan, H. Kuang and C. Xu, Sensors, 2014, 14, 14672 CrossRef PubMed.
  8. J. Ma, F. Yan, F. Chen, L. Jiang, J. Li and L. Chen, J. Liq. Chromatogr. Relat. Technol., 2015, 38, 665 Search PubMed.
  9. J. Papala, K. Erkomaa, J. Kukkonen and K. Sivonen, Anal. Chim. Acta, 2002, 466, 213 CrossRef.
  10. Z. He, S. Zang, Y. Liu, Y. He and H. Lei, Biosens. Bioelectron., 2015, 73, 85 CrossRef CAS PubMed.
  11. C. Wang, M. Lin, Y. Liu and H. Lei, Electrochim. Acta, 2011, 56, 1988 CrossRef CAS.
  12. H. Li, Q. Wei, J. He, T. Li, Y. Zhao, Y. Cai, B. Du, Z. Qian and M. Yang, Biosens. Bioelectron., 2011, 26, 3590 CrossRef CAS PubMed.
  13. M. Lin, Y. Liu, X. Chen, S. Fei, C. Ni, Y. Fang, C. Liu and Q. Cai, Biosens. Bioelectron., 2013, 45, 82 CrossRef CAS PubMed.
  14. H. Huang, Y. Liu, Q. Gao, W. Ruan, X. Lin and X. Li, ACS Appl. Mater. Interfaces, 2014, 6, 10258 CAS.
  15. L. Zhang, T. Wu, X. Xu, F. Xia, H. Na, Y. Liu, H. Qiu, W. Wang and J. Gao, J. Alloys Compd., 2015, 626, 364 CrossRef.
  16. R. Ramachandran, M. Saranya, V. Velmurugan, B. P. C. Raghupathy, S. K. Jeong and A. N. Grace, Appl. Energy, 2015, 153, 22 CrossRef CAS.
  17. K. Han, P. Miao, H. Tong, T. Liu, W. Cheng, X. Zhu and Y. Tang, Appl. Phys. Lett., 2014, 104, 053101 CrossRef.
  18. G. Wu, R. Li, Z. Li, J. Liu, Z. Gu and G. Wang, Electrochim. Acta, 2015, 171, 156 CrossRef CAS.
  19. P. Chen, J. Yang, S. Li, Z. Wang, T. Xiao, Y. Qiao and S. Yu, Nano Energy, 2014, 2, 249 CrossRef.
  20. Z. Sun, Z. Luo, C. Gan, S. Fei, Y. Liu and H. Lei, Biosens. Bioelectron., 2014, 59, 99 CrossRef CAS PubMed.
  21. P. Duangkaew, S. Tapaneeyakorn, C. Apiwat, T. Dharakul, S. Laiwejpithaya, P. Kanatharana and R. Laocharoensuk, Biosens. Bioelectron., 2015, 74, 673 CrossRef CAS PubMed.
  22. S. Zang, Y. Liu, M. Lin, J. Kang, Y. Sun and H. Lei, Electrochim. Acta, 2013, 90, 246 CrossRef CAS.
  23. L. Jiang, J. Han, F. Li, J. Gao, Y. Li, Y. Dong and Q. Wei, Electrochim. Acta, 2015, 160, 7 CrossRef CAS.
  24. Y. Luo, A. M. Asiri, X. Zhang, G. Yang, D. Du and Y. H. Lin, RSC Adv., 2014, 4, 54066 RSC.
  25. B. Zielinska, B. Michakiewicz, E. Mijowska and R. J. Kalenczuk, Nanoscale Res. Lett., 2015, 10, 430 CrossRef PubMed.
  26. J. Tang, J. Liu, C. Li, Y. Li, M. O. Tade, S. Dai and Y. Yamauchi, Angew. Chem., Int. Ed., 2015, 54, 588 CAS.
  27. Q. Zhang, L. Li, Y. Wang, Y. Chen, F. He, S. Gai and P. Yang, Electrochim. Acta, 2015, 176, 542 CrossRef CAS.
  28. S. Wang, W. Li, G. Hao, Y. Hao, Q. Sun, X. Zhang and A. Lu, J. Am. Chem. Soc., 2011, 133, 15304 CrossRef CAS PubMed.
  29. A. Noorbakhsh and A. Salimi, Electrochim. Acta, 2009, 54, 6312 CrossRef CAS.
  30. Y. Liu, Z. Zhang, L. Nie and S. Yao, Electrochim. Acta, 2003, 48, 2823 CrossRef CAS.
  31. W. S. Hummers and R. E. Offeman, J. Am. Chem. Soc., 1958, 80, 1339 CrossRef CAS.
  32. Y. Yan, Y. Xiao, G. Ning, T. Weo and Z. Fan, RSC Adv., 2013, 3, 2566 RSC.
  33. X. Yu, H. Fan, Y. Liu, Z. Shi and Z. Jin, Langmuir, 2014, 30, 5497 CrossRef CAS PubMed.
  34. H. Gao, Y. Sun, J. Zhou, R. Xu and H. Duan, ACS Appl. Mater. Interfaces, 2013, 5, 425 CAS.

Footnote

Electronic supplementary information (ESI) available: Optimization of Ab2, estimation of detection limit, and comparison with other methods. See DOI: 10.1039/c6ra07881h

This journal is © The Royal Society of Chemistry 2016
Click here to see how this site uses Cookies. View our privacy policy here.