Introduction

Wine is an ancient alcoholic beverage with relevant health benefits [1], also representing a lucrative and growing industry with a global market value of 28.9 billion euros and an annual production of 267 million of hectoliters in 2016 [2]. Wine is produced by liquid fermentation and microbial transformation of grape juice through a complex ecological and biochemical process involving the interaction of many microorganisms, including yeasts, fungi, acetic acid bacteria, lactic acid bacteria, and sometimes other microbial species of Streptomyces, Bacillus, and Clostridium [3].

The yeast Saccharomyces cerevisiae (Sacch) is by far the predominant species in wine production due to its ability to achieve complete fermentation of sugars to alcohol, carbon dioxide, and secondary end-products [4,5,6]. However, Sacch should be properly eliminated after fermentation to avoid uncontrolled growth at the re-fermentation stage, causing alteration of the chemical composition of wine, and accordingly, detracting from its organoleptic properties of appearance, aroma, and flavor. In fact, Sacch is a critical spoilage yeast, mainly for sweet wines, due to its unusual resistance and growing capacity at high ethanol concentrations [7]. For this reason, the development of reliable analytical approaches for the sensitive and fast detection of Sacch constitutes a priority for wine industry.

In general, traditional microbiology testing methods remain the core basis and standard practice for the detection and quantification of Sacch in wine [8, 9]. Despite the advantages associated with these microbial culturing methods, such as simplicity, specificity, and relatively low cost, these approaches are time consuming and are often inappropriate for fast decision making during wine production and processing [8, 9]. PCR-based molecular biology methods have also been proposed for microbial analysis of wine, but they require expensive instruments and reagents, as well as specialized personnel [7, 9,10,11]. For these reasons, it is imperative to provide wine industry with new analytical procedures for the fast and sensitive detection of Sacch.

Electrochemical immunosensors are promising tools for in situ determination of pathogenic and non-pathogenic microorganisms in food [12,13,14]. These devices offer relevant advantages such as high sensitivity and specificity, reduction or even elimination of sample treatment steps, and relatively low cost. Electroanalytical instruments can also be miniaturized to portable and user-friendly devices that can be easily employed by non-specialized personnel to perform reliable and fast analysis at the place of the sample collection. Electrochemical immunosensors can also be empowered by using advanced nanomaterials and nanohybrids as transduction elements and supports for the stable immobilization of the biorecognition elements. Furthermore, the use of disposable screen-printed electrodes as sensing interface allows to reduce price and to increase miniaturization and portability to the analysis [15,16,17].

In this work, we describe the construction of a novel and disposable affinity-based amperometric immunosensor for Sacch determination in wine. This electroanalytical device was assembled by covalent immobilization of a polyclonal anti-Sacch antibody on the surface of carbon screen-printed electrodes modified with propionic acid-functionalized graphene oxide (PA-GO/SPE). The sensing approach relies on a sandwich-type scheme, based on the initial capture of the yeast cells by the antibody molecules at the sensing interface, and further biorecognition of the glycoproteins at the Sacch cell surface with a concanavalin A-peroxidase (Con A-HRP) conjugate. Amperometric detection of the captured yeasts was finally performed by addition of H2O2/hydroquinone to the electrochemical cell.

Experimental

Materials

Carbon screen-printed electrodes modified with propionic acid-functionalized graphene oxide OHT-075 and OEH-000 electrode holder were donated by Orion High Technologies S.L. (Spain). Polyclonal anti-Sacch antibody (Ab) was purchased from Abcam (UK); Con A-HRP peroxidase conjugate (ConA-HRP) was acquired from Sigma-Aldrich Co. (USA). All other chemicals were of analytical grade.

Yeast strains used in this study were obtained from the University of Castilla La Mancha culture collection. Strains were growth on YPD agar (1% yeast extract, 2% bacteriological peptone, 2% glucose, and 2% agar–agar, w/v), incubated at 30 °C for 2–4 days and kept in slants of the same medium under refrigeration and refreshed weekly. To adjust the cell concentration in all cases, a Thoma hemocytometer cell count chamber was used.

Instruments, electrodes and solutions

Electrochemical experiments were carried out with propionic acid-functionalized graphene oxide-modified screen-printed electrode OHT-075, consisting of a carbon ink working electrode (4 mm diameter) modified with the nanomaterial, a carbon ink counter electrode, and a Ag/AgCl ink pseudo-reference electrode. Amperometric measurements were carried out with an EmStat3 Blue potentiostat (PalmSens, The Netherlands). A 0.1 M sodium phosphate buffer at pH 7.2 solution was employed as working buffer.

Preparation of the antibody-functionalized electrode

The working electrode in PA-GO/SPE was first coated with 10 μL of 1 mg/mL solution of N-hydroxisuccinimide (NHS) and 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDAC) in 50 mM sodium phosphate buffer, pH 6.0. After 30 min of incubation at 25 °C, the electrode surface was washed with distilled water and dried under N2. Ten microliters of a 50 μg/mL Ab in working buffer was then dropped on the electrode surface and incubated at 25 °C during 30 min. The antibody-functionalized electrodes were further three times washed with cold working buffer and dried under N2. The working electrode surface was further incubated with 10 μL of 1.5% bovine serum albumin (BSA) in working buffer during 30 min at 4 °C. The resulting electrode (Ab-PA-GO/SPE) was then washed with cold working buffer, dried under N2, and kept at 4 °C until use.

Procedure for electrochemical biosensing

To detect Sacch content in wine and aqueous solution, samples were first properly diluted in working buffer. Ten microliters of the resulting mixture was dropped on the working electrode surface and incubated during 45 min at 25 °C. The electrode was then washed with working buffer and dried under N2.

Commercial Con A-HRP solution was 1000-fold diluted in working buffer, and then 10 μL was added to the surface of the working electrode. After 45 min incubation at 25 °C, the electrode was washed with working buffer, dried under N2 and then 45 μL of a 1 mM hydroquinone solution, freshly prepared in working buffer, were dropped on the electrochemical cell. Changes in the cathodic current were measured at −200 mV vs Ag/AgCl pseudo reference electrode after addition of 5 μL of a freshly prepared 5 mM H2O2 solution. Each measurement was carried out at least in triplicates, by using different Ab-PA-GO/SPE electrodes, unless stated otherwise.

Results and discussion

Figure 1 displays the different steps involved in the construction of the electrochemical immunosensor, and its use to detect Sacch by amperometric measurements. Disposable screen-printed electrodes coated with a propionic acid-funcionalized graphene oxide nanohybrid were selected as sensing interfaces due to the useful electrochemical properties of graphene oxide [18,19,20,21,22,23]. In addition, the introduction of propionic acid moieties increases the density of carboxylic acid groups on the nanomaterial surface, allowing high load of the specific antibody molecules and easy covalent immobilization through conventional carbodiimide-coupling chemistry.

Fig. 1
figure 1

Schematic display of the steps involved in the preparation and performance of the immunosensor for Sacch

Full size image

In this work, a Con A-HRP conjugate was used as signaling element. This selection was based on the combined ability of Con A to recognize the carbohydrate moieties in the mannoproteins at the surface of the yeast cells [24, 25] and the capacity of peroxidase to catalyze the oxidation of hydroquinone in the presence of H2O2, yielding p-benzoquinone which can be easily detected by amperometric methods [26,27,28].

The assembly of the immunosensor device was optimized by determining the influence of different experimental variables on the amperometric response of the immunosensor toward 104 CFU/mL Sacch cells. Accordingly, the amount of anti-Sacch polyclonal antibody incubated on the electrode surface during the preparation of the immunosensor was first optimized. As is shown in Fig. 2A, a clear increase in the analytical response of the electrode was observed when the amount of antibody increased, reaching a maximum value for 0.5 μg antibody. Higher amount of antibody caused a noticeable decrease in the amperometric response. This fact could be ascribed to a lower biorecognition capacity of the sensing surface toward Sacch cells due to the steric hindrance caused by the bulky protein layer, formed by the non-oriented immobilized antibody molecules.

Fig. 2
figure 2

Influence of the amount of antibody (A), time of antibody immobilization (B), and concentration of BSA (C) on the amperometric response of the electrode. Eapp = −200 mV; Sacch concentration, 104 CFU/mL; measuring solution, 0.9 mM hydroquinone and 0.5 mM H2O2 in working buffer

Full size image

Figure 2B shows the influence of the immobilization time for the antibody on the analytical response of the immunosensor. The functionalized electrode displayed maximum amperometric signals when the immobilization times were 90 min or higher. Then, further immunosensors were assembled by using 90 min as optimum immobilization time.

Finally, BSA was employed as coating protein to minimize non-specific adsorptions of other sample components on the electrode surface. In fact, Sacch secreted a large variety of glycoproteins [29] that can be recognized by the Con A-HRP conjugate upon non-specific adsorption on the carbon electrode surface, thus affecting the analytical response of the electrochemical sensor. The specific/non-specific immunosensor response, tested toward Sacch samples, exhibited an almost bell-shaped behavior with the highest response at 1.5% BSA, and this concentration was used for further experiments.

The working conditions for the immunosensor device were also optimized. Figure 3A shows the effect of the time of incubation with the yeast samples on the amperometric response of the electrode. Maximum electroanalytical signal was reached when the functionalized electrode was incubated with the yeasts during 45 min. It was also demonstrated that high analytical signal were recorded when this affinity-based recognition process was performed at 25 °C (Fig. 3B). Consequently, further experiments were performed by using these optimized conditions. In addition, amperometric measurements were performed by using a working potential of − 200 mV and HRP substrate concentrations of 0.9 mM hydroquinone and 0.5 mM H2O2 in 0.1 M sodium phosphate buffer at pH 7.2 according to our previous studies [26, 30].

Fig. 3
figure 3

Effect of the time (A) and temperature (B) of incubation with the yeast samples on the amperometric response of the electrode. Eapp = −200 mV; Sacch concentration, 104 CFU/mL; measuring solution, 0.9 mM hydroquinone and 0.5 mM H2O2 in working buffer

Full size image

The analytical performance of the immunosensor on buffered and undiluted commercial white wine samples (Don Simón, Spain) was then evaluated. To avoid the potential presence of autochthonous yeasts in the commercial wine, samples were filtered through a 0.45-μm membrane before use.

The average calibration curve for ten different experiments obtained with the immunosensor for Sacch in buffered solutions is shown in Fig. 4. The cathodic current values exhibited a linear behavior (r = 0.998) with the logarithm of Sacch concentration in the broad range of 10–107 CFU/mL, according to the following equation:

$$ {i}_{\mathrm{C}}\kern0.15em \left(\mathrm{nA}\right)=5.7\times \log \kern0.30em \left(\mathrm{Sacch},\mathrm{CFU}/\mathrm{mL}\right)-2.4 $$
Fig. 4
figure 4

Calibration curves obtained with the amperometric immunosensors toward Sacch in working buffer solution (empty circles) and commercial wine samples (filled circles). Eapp = −200 mV; measuring solution, 0.9 mM hydroquinone and 0.5 mM H2O2 in working buffer

Full size image

In addition, a low detection limit of 4 CFU/mL was estimated for this immunosensor. This parameter was calculated according to the 3Sb/m criterion, where m is the slope of the calibration curve and Sb was estimated as the standard deviation of ten different amperometric signals recorded for the lowest yeast concentration measured in the calibration graphs (10 CFU/mL).

The immunosensor was also tested toward Sacch in commercial and undiluted white wine samples, by using the standard addition method (Fig. 4). The immunosensor also exhibited a linear behavior (r = 0.995, n = 10) with the logarithm of yeast concentration in the range of 10–107 CFU/mL, according to the following equation:

$$ {i}_{\mathrm{C}}\kern0.15em \left(\mathrm{nA}\right)=7.1\times \log \kern0.15em \left(\mathrm{Sacch},\mathrm{CFU}/\mathrm{mL}\right)+13.2 $$

The slope of this calibration curve showed a significant difference at 95% confidence with respect to those obtained in buffered solutions, demonstrating a matrix effect for this wine sample. On the other hand, the immunosensor exhibited low detection limit values of 6 CFU/mL for Sacch in white wine samples.

As can be observed in Table 1, the analytical properties of this amperometric immunosensor are similar or even better than those previously reported for other immunosensor devices, in terms of limit of detection and range of Sacch concentrations.

Table 1 Analytical properties of immunosensors employed for Sacch determination
Full size table

To evaluate the reproducibility of this immunosensor device, ten different electrodes prepared in the same manner were tested toward 104 CFU/mL Sacch in wine samples. This affinity-based sensor showed good reproducibility, with a relative standard deviation of 16.3%. It should be notified that the commercial propionic acid-modified screen-printed electrodes used showed a relative standard deviation of 5.7% toward the quinone/hydroquinone redox system. Accordingly, reproducibility of this immunosensor should be mainly affected by the complex and multi-step preparation protocol employed.

As can be observed in Fig. 5, the immunosensor showed excellent storage stability, retaining high sensing capacity after 3 weeks of storage at 4 °C in dry conditions. This fact could be ascribed to the high hydrophilic environment around the antibody molecules provided by the carboxylic acid enriched nanohybrid. This negative charged nanomaterial can also confer structural stability to the antibody molecules through the formation of multipoint electrostatic interactions at the protein surface. It should also highlighted the potential stabilizing effect of BSA molecules for the 3D structure of the antibody, as has been previously reported for other proteins.

Fig. 5
figure 5

Effect of time of storage at 4 °C in dry conditions on the normalized amperometric response of the immunosensor toward 104 CFU/mL Sacch. Eapp = −200 mV; measuring solution, 0.9 mM hydroquinone and 0.5 mM H2O2 in working buffer

Full size image

The selectivity of the immunosensor was also evaluated by comparing its amperometric response in commercial white wine samples toward 104 CFU/mL Sacch and the same concentration of other relevant yeasts (Zygosaccharomyces fermentati, Hanseniaspora uvarum, and Torulaspora delbrueckii). As can be observed in Fig. 6, the immunosensor did not provide a significant analytical response toward Z. fermentati, H. uvarum, and T. delbrueckii, which could be ascribed to the high specificity of the polyclonal antibody for Sacch.

Fig. 6
figure 6

Normalized electrochemical response of the immunosensor toward 104 CFU/mL Sacch (1), Z. fermentati (2), H. uvarum (3), T. delbrueckii (4), and a mixture of Sacch and the rest of yeast (5) in commercial and undiluted white wine samples. Eapp = −200 mV; measuring solution, 0.9 mM hydroquinone and 0.5 mM H2O2 in working buffer

Full size image

In addition, it was further demonstrated that the immunosensor showed similar amperometric response toward 104 CFU/mL Sacch in the absence and presence of a mixture of Z. fermentati, H. uvarum, and T. delbrueckii at the same concentration. Accordingly, we can conclude that these yeasts did not cause a significant interference to the determination of the Sacch in white wine samples.

The immunosensor was finally validated in white wine samples by comparing the electroanalytical results with those obtained by the conventional cell-counting method. All measurements were performed in triplicate. Commercial white wine samples were spiked with a theoretical concentration of 8.4 × 104 CFU/mL of Sacch, and total cells were counted on a Thoma hemocytometer cell count chamber after coloration with methylene blue [35].

The average Sacch content in the spiked wine samples was estimated as (85,200 ± 300) CFU/mL and (81,400 ± 500) CFU/mL by conventional cell counting and amperometric detection, respectively, representing 95.5% of recovery.

Conclusions

A novel amperometric immunosensor for Sacch analysis was constructed by immobilizing a specific polyclonal antibody on the surface of disposable carbon screen printed electroded modified with propionic acid functionalized graphene oxide. This analytical device showed excellent performance in term of linear range of response, reproducibility, stability, selectivity and limit of detection. The immunosensor was also succesfully employed to quantify Sacch in commercial white wine samples. According to our results, we can suggest that this nanostructured immunosensor could be a valuable tool for the determination of Sacch in wine industry.