Introduction

Nanotechnology has the potential to create many new materials and devices with a vast range of applications, it opens new prospects for biosensor design and construction [1–5]. The application of nanomaterials in biosensor design has grown rapidly within past few years. It increases selectivity and detection limits of analyte [6, 7]. The application of nanomaterials in electrochemical sensors was reviewed in detail [2, 6, 8, 9]. Carbon based nanomaterials [3] and gold nanoparticles (Au-NPs) [2–4, 6, 10] are characterized by high electrochemically-active surface area, favourable electronic properties and even by significant electrocatalytic effect what is very attractive for electrochemical biosensor design. Analysts in this field are always enthusiastic about finding new materials with good biocompatibility to improve the behaviour of biosensors. Considerable attention has been devoted to immobilization of enzyme on electrode surface, which is one of the main factors that affect the performance of bioanalytical system [6, 11, 12]. Enzymes have been immobilized on electrode surface by passive physical adsorption [13, 14], covalent attachment [15, 16], entrapment within an ultra-thin polymeric film [11, 17, 18] with an ultra-high density, or encapsulation within polymeric nanoparticles [19, 20]. Bioconjugates based on GOx and gold-nanostructures showed extraordinary stability at high temperature [21]. Some biosensors with immobilized enzymes are highly selective, sensitive, fast, reversible, practically applicable and demonstrates excellent catalytic activity [10, 22–24]. In this point, another important challenge is the improvement of performance of enzyme-based biosensors. Transduction of analytical signal in amperometric biosensors can be based on the electroactivity of the enzyme substrate or reaction product; utilization of red-ox mediators, or on direct electron transfer (DET) between the red-ox active biomolecule and the working electrode surface [11, 24–26]. However, some problems and/or limitations arise in the design of electrochemical biosensors and these are usually related to impossible DET between majority of red-ox enzymes and working electrode [26] because red-ox centre of enzyme is deeply buried in its electrochemically “insulated” peptide backbone, resulting in the inaccessibility of the red-ox centre. In this case electron-transfer mediators are needed for registration of amperometric signals [23, 25, 27]. In some designs of biosensors a low molecular weigh electron-transfer mediators are used to shuttle electrons between the enzyme and the electrode surface [27, 28]. Moreover, enzyme after immobilization on the electrode surface could change its conformation and lose biological and electrochemical activity.

Some nanoscale materials including Au-NPs are applied in the optimization of enzyme immobilization [17, 23, 24, 29, 30]. Au-NPs are useful for the design of electrochemical and optical biosensors [5, 30–32]. The Au-NPs have also been used as electron-transfer mediators and as ‘electric wires’ for enhancing the electron-transfer rate between the active centre of enzyme and electrode [33]. It allows electrochemical sensing without the application of other electron-transfer mediators [25, 29], for this reason various aspects of Au-NPs interaction with red-ox enzymes are highly important. Usually Au-NPs are prepared by chemical reduction of the corresponding transition metal salts in the presence of a stabilizer, which binds to their surface to impart high stability of colloid solution, to provide the surface charge and to increase their solubility [2, 32]. The size and surface morphology of Au-NPs can be controlled during a preparation procedure [22]. It was reported that amperometric biosensors modified by 10 nm Au-NPs are more sensitive [17] and can be used for the achievement of direct electron transfer between enzymes and electrode surface.

The key idea of this paper is to investigate the efficiency of colloidal solution of Au-NPs in mediated electron transfer. In this work the GOx was used for the design of amperometric glucose biosensor model.

Experimental

Chemicals

Glucose oxidase (GOx) (EC 1.1.3.4, type VII, from Aspergillus niger, 215.266 units mg−1 protein) and N-methylphenazonium methyl sulphate (PMS) were purchased from Fluka and Sigma-Aldrich (Buchs, Switzerland), respectively. D-(+)-Glucose (Glu), tetrachloroauric acid (HAuCl4·3 H2O) and tannic acid were obtained from Carl Roth GmbH&Co (Karlsruhe, Germany), sodium citrate – from Penta (Praha, Czech Republic), hydrochloric acid 37 % – from Acta Medica (Hradec Kralove, Czech Republic), sodium carbonate – from Lachema (Neratovice, Czech Republic). Before investigations glucose solutions were allowed to mutarotate for 12 h. All other chemicals used in the present study were either analytical grade or of highest quality. All solutions were prepared using deionised water purified with water purification system Millipore S.A. (Molsheim, France). The solution of sodium acetate buffer (0.05 mol L−1) with 0.1 mol L−1 KCl was prepared by mixing sodium acetate trihydrate and potassium chloride, which were obtained from Reanal (Budapest, Hungary) and Lachema (Neratovice, Czech Republic). Carbon (3 mm diameter, 99.999%, low density) was purchased from Sigma-Aldrich (St. Louis, USA), alfa alumina powder (grain size 0.3 μm, Type N) – from Electron Microscopy Sciences (Hatfield, USA) and 25% glutaraldehyde solution – from Fluka Chemie GmbH (Buchs, Switzerland).

Pre-treatment of working electrode

The working surface area of the carbon rod (CR) electrodes was 0.071 cm2. Carbon rod of spectroscopic carbon were cut and polished by fine emery paper and then polished by slurry of alfa alumina powder containing 0.3 μm grains of Al2O3. The electrode surface was rinsed with deionised water and dried at room temperature 20 ± 2 °C. Then electrodes were modified with GOx (GOx/CR) and sealed into silicone tube to prevent contact of the electrode side surface with the solution.

Working electrode modification by glucose oxidase

During preparation of GOx/CR, 3 μL solution containing 40 mg mL−1 of GOx were deposited on the electrode and water was evaporated at room temperature by intensive ventilation. After water evaporation electrodes were stored for 15 min over 25% solution of glutaraldehyde at room temperature in a closed vessel. The development and optimization of this immobilization procedure was described in details previously [34]. Prior to all electrochemical measurements, working electrodes were thoroughly washed with deionised water to remove non cross-linked enzyme. Working electrode was stored in a closed vessel over the solution of sodium acetate buffer at +4 °C until used in the experiment.

Electrochemical measurements

All electrochemical measurements were performed using a computerized potentiostat PGSTAT 302N/Autolab (EcoChemie, Netherlands) with GPES 4.9 software in amperometry mode at working electrode potential of +0.3 V vs. Ag/AgCl. A conventional three-electrode system comprising a working carbon rod electrode (modified as it was described above), 2 cm2 platinum as an auxiliary electrode and Ag/AgCl with 3 mol L−1 KCl Metrhom (Herisau, Switzerland) as a reference electrode was employed for all electrochemical experiments. All experiments were performed at room temperature in stirred 0.05 mol L−1 sodium acetate buffer, pH 6.0, with 0.1 mol L−1 KCl. Electrochemical detection of the analytical signal was performed at different concentrations of glucose in the presence of Au-NPs without and with 2 mmol L−1 of PMS. The activity of GOx was estimated by measuring the oxidation current of PMSH2 and/or H2O2 at +0.3 V potential vs. Ag/AgCl.

Synthesis of gold nanoparticles

The Au-NPs of 6 nm and 13 nm diameters were synthesized by reducing HAuCl4·3H2O by sodium citrate in the presence of tannic acid as described in the literature [35, 36]. An aqueous solution of tetrachloroauric acid (81 mL of 0.0125% [w/v] HAuCl4·3H2O) was brought into heated Erlenmeyer flask with magnetic stirring. Solution of sodium citrate (4 mL of 1% [w/v]) in deionised water, 0.500 mL and 0.025 mL of 1% solution of tannic acid for synthesis of 6 nm and 13 nm Au-NPs, respectively, were added to the flask while heated until 60 °C and rapidly stirred. The mixture of two solutions was heated until 98 °C and kept at this temperature for 3 min to yield colloidal solution of Au-NPs of 6 nm and 13 nm diameters (a wine-red colour of solution). Before use the colloidal gold suspensions were stored in a dark glass vessel at +4 °C until needed. The concentration of 6 nm and 13 nm Au-NPs was 23 nmol L−1 and 3.6 nmol L−1, respectively; it was calculated considering the amount of starting material, density of gold, the approximate size of the nanoparticles and assuming that the reaction yield was 100% [37]. Initial concentration of gold (according to mass %) in all used solutions was the same – 0.0058 %.

Calculations

The kinetic parameters: the maximal speed of an enzymatic reaction (V max) and the apparent Michaelis-Menten constant (K M(apparent)) – are correspondingly a and b parameters of hyperbolic function y = ax/(b + x) and were used for approximation of results. The kinetic parameters of the enzyme-catalyzed reaction were calculated using SigmaPlot software.

Results and discussion

The principle of amperometric biosensors is based on monitoring of current associated with oxidation or reduction of an electroactive species involved in the process of biological recognition [11]. The aim of this research was to investigate red-ox mediating properties of dissolved Au-NPs in electrochemical system where GOx is immobilized on the surface of working electrode. Previously published study by Li et al. shows that GOx conjugate with Au-NPs posses high thermostability and can retain high electrocatalytic activity within broad pH range: the highest sensitivity was reached at pH 6.0 [24], although some researchers have reported that maximal activity of layered GOx and Au-NPs system was between pH 4.5 and 7.0 [17, 30, 38, 39]. Thereby 0.05 mol L−1 sodium acetate buffer, pH 6.0, was chosen for measurements described in this paper. Recently, the results of the kinetic parameters by use of 0.05 mol L−1 sodium acetate and sodium phosphate buffers, pH 6.0, were investigated and compared. The best sensitivity of glucose could be reached in sodium acetate medium [40]. It was illustrated that for GOx/CR electrode in 0.05 mol L−1 sodium acetate buffer the K M was 1.8 times higher, when compared with results registered in 0.05 mol L−1 sodium phosphate buffers. On another hand, the pH 6.0 of buffer was chosen as a compromise to retain good activity of enzyme. The probe-electrode was based on carbon rod modified by cross-linked GOx (GOx/CR). The schematic diagram for the reactions, which occur on GOx/CR, is presented in Fig. 1. The GOx in the presence of glucose and oxygen dissolved in water generate hydrogen peroxide and gluconolactone, which is hydrolysed to gluconic acid (Fig. 1). Electrons from red-ox centre of enzyme are transferred towards positively (+0.3 V vs. Ag/AgCl) charged electrode via dissolved PMS and registered steady-state current is proportional to the concentration of glucose in the sample. The choice of the working potential of +0.3 V for presented investigations was determined by the fact, that this value is the formal red-ox potential of the mediator – PMS. Amperometric measurements show that catalytic activity of GOx remained after immobilization procedure and after continuous measurements, what is in line with the investigations demonstrated in other studies based on application of other red-ox enzymes [19, 34].

Fig. 1
figure 1

Principle scheme of electron-transfer process during oxidation of glucose on GOx/CR electrode in the presence of Au-NPs and PMS as red-ox mediators

Full size image

Principle scheme of electron-transfer process during oxidation of glucose on GOx/CR electrode in the presence of Au-NPs and PMS is presented in Fig. 1. Application of metal nanoparticles in biosensor design represents a rapidly advancing field. Recently, research efforts on Au-NPs have flourished because of their good biological compatibility, excellent conducting capability and high surface-to-volume ratio. For this reason in some cases metallic nanoparticles could increase the electron transfer rate between enzyme and electrode surface [40, 41] and/or the red-ox centre of enzyme can be plugged to electrode surface via Au-NPs [2, 3, 23, 24]. Nanoparticles are chemically and electrochemically more active than usual electrode materials. Au-NPs are characterized by high surface energy and electrochemically active surface [6], what enhance the electron transfer towards the electrode surface and accelerate electrocatalytic reactions [24, 27, 31]. Willner’s group [42] has demonstrated an approach that is suitable for establishment of ‘electrical’ contact of red-ox enzymes (GOx) by the reconstitution of an apo-enzyme with a flavin adenine dinucleotide (FAD) covalently connected to Au-NPs. In such systems the Au-NPs act as an electrical relay element between the FAD red-ox site and the electrode. Moreover in previously performed spectrophotometric study showed that Au-NPs in the solution could act as red-ox mediator for the GOx [8].

Direct electron-transfer between GOx and electrode surface and the combination of the catalytic properties of nanoparticles [8] could give the possibility for the creation of reagentless biosensors [2, 7, 25]. From the point of this view the electron transfer from GOx towards carbon electrode in the presence of Au-NPs was tested in present study. Using GOx/CR electrodes hyperbolic dependences of amperometric signal on glucose concentration in the range of 0.1–100 mmol L−1 (Fig. 2) in sodium acetate buffer solution, pH 6.0, with PMS in presence and absence of Au-NPs were observed.

Fig. 2
figure 2

Calibration plots of carbon rod electrodes based on cross-linked GOx in the presence of 13 nm Au-NPs of different concentration: (1) 0.60 nmol L−1, (2) 1.5 nmol L−1 and (3) absence of Au-NPs. Changes in anodic current are presented as a function of glucose concentrations in 0.05 mol L−1 sodium acetate buffer, pH 6.0, containing 2 mmol L−1 PMS at +0.3 V vs. Ag/AgCl

Full size image

All presented hyperbolic dependences were in agreement with Michaelis-Menten kinetics and are presented in Table 1. The kinetic parameters: the maximal reaction speed (V max) of an enzymatic reaction and the apparent Michaelis-Menten constant (K M(apparent)) are correspondingly a and b parameters of hyperbolic function y=ax/(b+x). Coefficient of determination (R 2) shows how well our experimental results fit with predicted statistical model.

Table 1 Parameters of Michaelis-Menten kinetics
Full size table

Results presented in Fig. 2 and in Table 1 illustrate that in the presence of PMS the sensitivity of GOx/CR electrode depends on the concentration of AuNPs. Thereby using 1.5 nmol L−1 and 0.60 nmol L−1 of Au-NPs in the solution of sodium acetate buffer, pH 6.0, the maximal speed of an enzymatic reaction (V max) slightly increases (up to 11%) in the comparison with system not-containing Au-NPs. No changes of analytical signal were observed in the absence of electron mediator – PMS – and values of analytical signal in this case were similar. These results could be explained by relatively low mobility of 13 nm Au-NPs and the rate, what establishes relatively low electron transfer from GOx to carbon rod electrode. Here presented results demonstrate that in all cases the PMS increases the rate of electron transfer from GOx to the electrode.

We compared the analytical signal of sensors with different concentrations of 13 nm Au-NPs in buffer solution in the absence and in the presence of PMS (2 mmol L−1) (Table 1). The maximal speed of an enzymatic reaction (V max) increased by 6.425 times using 1.5 nmol L−1 concentration of 13 nm Au-NPs and PMS. Similarly V max increased by 7.175 times using 0.60 nmol L−1 concentration of 13 nm Au-NPs and PMS. It indicates that direct electron transfer from red-ox centre of GOx to electrode using only Au-NPs (without PMS) was less efficient. Analytical signal in the system with 2 mmol L−1 PMS depends on the concentration of Au-NPs and at 0.60 nmol L−1 concentration of 13 nm Au-NPs was higher than the signal at 1.5 nmol L−1 concentration of 13 nm Au-NPs.

The values of K M(apparent) were determined by analysis of corresponding slope and intercept for the plot in the reciprocal-coordinates of the steady-state current versus glucose concentration [40]. Michaelis-Menten constant of GOx/CR electrode in sodium acetate buffer solution, pH 6.0, with 2 mmol L−1 PMS and in the presence of 1.5 nmol L−1 or 0.60 nmol L−1 13 nm Au-NPs (Table 1) was 1.2 and 1.3 times lower, respectively, if compared with the same parameter for the system without Au-NPs. If PMS was not used in these systems, K M(apparent) was 1.1 and 1.9 times lower, respectively, in comparison with the system without Au-NPs. Although Au-NPs together with PMS increased electron transfer from red-ox centre to electrode and it was reflected as higher V max value. However we don’t observe any positive effect on the extension of the linear detection ranges of glucose. Also during presented investigations no clear electrochemical response was observed in the absence of soluble red-ox mediator – PMS. The K M(apparent) in the presence of gold nanoparticles in the solution was calculated to be 3.20–5.72 mmol L−1 without and 12.3–13.6 mmol L−1 with mediator (Table 1). It well correlates with K M(apparent) values reported by some other authors in the different systems without mediator: 4.58 mmol L−1 for the glucose biosensor based on GOx immobilized by cross-linking in the matrix of bovine serum albumin on a Pt electrode [39]; 3.5 mmol L−1 for the glucose biosensor fabricated by the deposition of chitosan-gold nanoparticles on glassy carbon electrode [43]. However, in some systems without mediator the calculated K M(apparent) values are slightly higher: 12.1 mmol L−1 for the glucose biosensor based on immobilized glucose oxidase at the gold nanoparticles electrodeposited on indium tin oxide electrode surface [44], and 10.73 mmol L−1 for the amperometric biosensor developed through the adsorption of glucose oxidase on the gold and platinum nanoparticles modified carbon nanotubes [45]. The calculated K M(apparent) values with PMS and Au-NPs (12.3–13.6 mmol L−1) is slightly lower if compared with the previous results: 17.6 mmol L−1 for the glucose biosensor based on immobilized GOx on carbon rod modified with gold nanoparticles [40]; 16.0 mmol L−1 for GOx directly immobilized on gold electrode modified with self-assembled monolayer and gold nanoparticles [46].

By next investigation step influence of Au-NPs concentration on amperometric response was evaluated. The 6 nm and 13 nm Au-NPs dissolved in sodium acetate buffer solution, pH 6.0 were used to facilitate electron transfer between immobilized GOx and electrode surface. Whereas the analytical signals without PMS were very low.

Hyperbolic decrease of amperometric signal was detected by increasing the concentration of Au-NPs from 0.01 nmol L−1 to 1.5 nmol L−1 in the buffer solution without PMS (Fig. 3). Electron transfer from enzyme to the Au-NPs present in the solution could explain these results. The ratio of Au-NPs(red)/Au-NPs(ox) decreases by significant increase of Au-NPs(ox) concentration, because only limited amount of Au-NPs(ox) can be utilized by immobilized GOx (in this case it is over 0.01 nmol L−1). It means that by increase of concentration of Au-NPs(ox) significant part of electrons is transferred to Au-NPs(red) that remains in colloid solution and are not re-oxidized on the electrode.

Fig. 3
figure 3

The dependence of the maximal speed of an enzymatic reaction on the concentration of 13 nm (1) and 6 nm (2) Au-NPs in 0.05 mol L−1 sodium acetate buffer, pH 6.0, without PMS at +0.3 V vs. Ag/AgCl, concentration of glucose – 100 mmol L−1

Full size image

As it is seen from Fig. 3 and Table 2, for all analytical systems based on colloidal solution of 6 nm and 13 nm Au-NPs, V max values were similar for maximal concentration of Au-NPs in the solution (1.5 nmol L−1) and about 1.8 times higher for minimal concentration of Au-NPs in the solution (0.01 nmol L−1), if compared with the same parameter calculated for the GOx/CR electrode in absence of Au-NPs. However, higher concentrations of Au-NPs in colloidal solution make the electron transfer in the same system less efficient if compared with some lower Au-NPs concentrations. As it was shown in the Fig. 1, Au-NPs are also intending to accumulate some electrons generated by GOx and to transfer toward solution instead of being transferred to the electrode.

Table 2 Parameters of Michaelis-Menten kinetics for electrochemical sensors using different concentration of Au-NPs without PMS
Full size table

As it seen from Table 2, in the systems containing 6 nm and 13 nm Au-NPs without PMS the higher sensitivity (V max are 0.013 and 0.014 μA, respectively) was found in sodium acetate buffer solution, pH 6.0, with 0.01 nmol L−1 concentration of Au-NPs (the lower concentration of Au-NPs tested in our investigations). For future experiments to achieve the higher analytical signals it is recommended to use not only small concentration of Au-NPs in the solution of buffer, but also to use red-ox mediator (e.g. 2 mmol L−1 PMS).

In the system with 0.60 nmol L−1 of Au-NPs and 2 mmol L−1 of PMS the linear response range of GOx/CR electrode to the concentration of glucose was 10 mmol L−1, the limit of detection for this sensor was determined as 0.05 mmol L−1, at a signal to noise ratio of 3. The limit of detection for created biosensor is 7.4 times lower if compared with the biosensing system (Nafion-GOx/Au-NPs/glassy carbon electrode) reported by other authors [30]. The reproducibility of analytical responses for created electrodes close to detection limit was 27%, because noise at lowest detectable glucose concentration is close to value of detection limit. The storage stability of GOx/CR electrode was presented in previous investigations [40]: the τ 1/2 of working electrode was 49.3 days when the electrode was stored at +4 °C. The stability of GOx/CR electrode was tested during a 66-day period and 43% of the initial current response was retained.

The major advantages of newly created biosensing system (GOx/CR electrode in the system with 0.60 nmol L−1 of Au-NPs and 2 mmol L−1 of PMS) are: (i) simple and low-cost detection of biologically active analytes, (ii) wide linear range of analytical signal vs. glucose concentration and (iii) low detection limit of analyte. The K M(apparent) in the presence of Au-NPs in the solution were 3.20–5.72 mmol L−1 without and 12.3–13.6 mmol L−1 with mediator. It well correlates with K M(apparent) values reported by some authors in the different systems without [39] and with [40, 46] mediator.

Conclusions

The Au-NPs promote electron transfer from GOx. This effect was amperometrically investigated at concentrations of Au-NPs in the range from 0.01 to 1.50 nmol L−1. The GOx-based enzymatic electrode with 0.60 nmol L−1 of Au-NPs and 2 mmol L−1 of PMS in the solution demonstrates linear analytical signal dependence within 0.1–10 mmol L−1 of glucose with lowest detection limit of 0.05 mmol L−1. This biosensing strategy could have been extended to the application of other red-ox enzymes. However by this study we determined that higher concentrations of Au-NPs in the solution make the electron transfer in the same system less efficient if compared with lower concentrations of Au-NPs. The Au-NPs are also intending to accumulate some electrons generated by GOx and transfer toward solution instead of transferring them to the electrode.