Introduction

Wine plays an important role worldwide since it is one of the most popular alcoholic beverages. The wine industry is highly relevant in Europe since Spain together with Italy and France are the major wine producers (147.8 million Hl in 2018) and the main wine exporters (54.9 million Hl in 2018). Furthermore, Spain is the country with the largest area devoted to vineyards (969 thousands Ha in 2018), representing 13% of the world vineyard (OIV International Organization of Vine and Wine 2019).

The determination of sugars in wines is of a great economic interest because of its importance as a quality parameter of the wine. The hexoses glucose and fructose are the major sugars in grapes, and their content is mainly dependent on factors such as variety (Eyduran et al. 2015; Monteiro Coelho et al. 2018; Sabir et al. 2010), maturation stage and harvest time (Bindon et al. 2013; Sabir et al. 2010), climate (Monteiro Coelho et al. 2018; Schelezki et al. 2018) and agriculture practices (Parpinello et al. 2015; Yang et al. 2020a). Their levels in musts range between 170 and 260 g L−1, whereas in wines, both sugars are present at a low concentration due to the fermentation process, by which both sugars are converted into ethanol by yeasts. Therefore, the amount of both sugars fermented determines the final ethanol content, which in the case of a whole fermentation process leads to an alcoholic degree of 10–15% (Boto Fidalgo and Boto Ordóñez 2016). Residual non-fermented sugar affects the wine characteristics, and the Commission Regulation (EC) (2009)of 14 July 2009 classifies the wines into four types according to the total sugar content, expressed in terms of glucose and fructose: dry (≤ 4 g L−1), semi-dry (> 4–12 g L−1), semi-sweet (> 12–45 g L−1) and sweet (> 45 g L−1).

Sugars are involved in very important characteristics such as astringency and viscosity (Li et al. 2017), colour stability and taste (Harbertson et al. 2013; Yang et al. 2020b) and flavour (Shen et al. 2011). The analysis of sugars is crucial in all steps of the winemaking process in order to follow the grape ripening, determine the optimum harvest time, control the fermentation process and decide the bottling date. The determination of total sugar is therefore of high relevance in the wineries and in the wine industry; thus simple, fast and accurate methods, which can be established as routine quality control procedures, are needed.

Due to the great importance of sugars in wines, numerous procedures have been described for their determination. Most methods involve colorimetric reactions based on the properties of reducing sugars employing copper(II) and iron(III). The earliest methods developed were titrimetric procedures, which although present many disadvantages, such as wide use of reagents and careful control of reaction conditions, are in current use due to their simplicity and low cost (Parpinello et al. 2015). Some of these methods, such as the Lane-Eynon method, are still included as a reference method in many rules and regulations (AOAC International 2016).

UV–Vis spectrophotometry is the most widely used technique for the determination of sugars, as it allows the employment of many analytical procedures and accessible equipment. The major reported methods are based on the reaction of reducing sugars with copper(II) (Başkan et al. 2016; Leotério et al. 2015) or with hexacyanoferrate(III) (Aimo et al. 2016; Brasil and Reis 2017). However, these procedures are not suitable for the determination of different sugars in a mixture. Colorimetric reactions based on the use of enzymes increase the specificity. The main advantages of the enzymatic determination rely on the inherent selectivity of the enzymes and the short analysis time. A simple enzymatic method employed for the measurement of glucose in foods is based on the amount of hydrogen peroxide produced through the glucose oxidation by glucose oxidase. The hydrogen peroxide is then quantified, in the presence of peroxidase, by reaction with the reduced form of o-dianisidine (Biscay et al. 2012) or p-hydroxybenzoic acid and 4-aminoantipyrine (Xu et al. 2012). However, the hydrogen peroxide produced is a strong oxidant and can react with other reducing compounds such as polyphenols leading to a decreased glucose content (Biscay et al. 2011; Xu et al. 2012). Therefore, the International Organisation of Vine and Wine (OIV) recommends the hexokinase method to measure the glucose concentration in musts and wines (OIV International Organization of Vine and Wine 2009). This assay is based on the phosphorylation of D-glucose, catalysed by hexokinase, and the subsequent oxidation of the glucose-6-phosphate with nicotinamide adenine dinucleotide phosphate (NADP) in the presence of glucose-6-phosphate dehydrogenase. Finally, the absorbance of the reduced form of the NADP (NADPH) is measured at 340 nm. Furthermore, D-fructose also can be determined by using phosphoglucose isomerase in order to convert fructose-6-phosphate into glucose-6-phosphate. Electrochemical detection based on amperometric enzyme electrodes has also been successfully applied for the determination of glucose and fructose in wines (Biscay et al. 2012; Shkotova et al. 2016).

Chromatographic methods, mainly high-performance liquid chromatography with refractive index detection (HPLC-RI), have been employed for the simultaneous determination of several sugars (Eyduran et al. 2015; Monteiro Coelho et al. 2018; Sabir et al. 2010). However, they present the disadvantages of requiring expensive equipment and using toxic solvents and waste generation. For the determination of sugars in musts and wines, the OIV recommends an HPLC method (OIV International Organization of Vine and Wine 2016).

Since the automation of instrumental analysis is always desirable to limit human errors, decrease analysis time and improve precision, in 2018, OIV added the determination of D-glucose and D-fructose in wines by the automated enzymatic method to the Compendium of International Methods of Analysis of Wines and Musts, (OIV International Organization of Vine and Wine 2018). As the European Legislation classifies wines according to the total sugar content, expressed in terms of glucose and fructose, the main aim of this work was the validation of an automatic sequential analyser for the determination of the sum of glucose and fructose in wines by automated enzymatic analysis. The method was successfully applied to the analysis of wines from Galicia (NW Spain), where the wine industry is very relevant, with exports around the world. The obtained results were compared to those found by employing the OIV reference chromatographic method.

Materials and Methods

Samples

A total of 18 commercial samples of wine from different vine-growing areas of Galicia (North-western Spain) were analysed. All the wine bottle samples were purchased directly from the cellars and stored protected from light exposure until analysis. In order to check precision and trueness, reference materials were employed, obtained by participating in an interlaboratory comparison test organised by the Interprofessional Bureau of Analytical Studies (BIPEA, Paris, France) and the Ministry of Agriculture, Fish and Food (Madrid, Spain). For enzymatic analysis, samples were injected directly, whereas for chromatographic analysis, samples were filtered and half-diluted in acetonitrile:water (40:60).

Chemicals

Fructose and glucose analytical standards were supplied by Sigma-Aldrich (Inc.115 St. Louis, MO, USA). A commercial calibrator of glucose/fructose constituted by standards at five levels of concentration (0.90, 1.80, 3.60, 5.40, 7.20 g L−1) was obtained from Biosystems (Barcelona, Spain). Acetonitrile (HPLC grade) was from Panreac (Barcelona, Spain), and Milli-Q water was obtained from a purification system from Millipore (Billerica, MA, USA). Hexokinase/phosphoglucose isomerase kit for D-glucose/D-fructose analysis was purchased from Biosystems (Barcelona, Spain).

Enzymatic Method

The analysis of total sugar expressed as the sum of glucose and fructose was performed according to the enzymatic reference method of the OIV (method OIV-MA-AS311-02, 2009). The enzymatic analysis was carried out in an automatic sequential analyser from Biosystems (model Y15) equipped with an UV detector and software to control operation and processing data.

The enzymatic method is based on the measurement of the absorbance of NADPH produced from sugars through several enzymatic reactions. Firstly, the enzyme hexokinase catalyses the phosphorylation of D-glucose and D-fructose by adenosine triphosphate (ATP) generating the corresponding 6-phosphate sugars. Then, fructose-6-phosphate is converted into glucose-6-phosphate by phosphoglucose isomerase. Finally, glucose-6-phosphate is oxidised into 6-phosphogluconate by glucose-6-phosphate dehydrogenase in the presence of NADP, generating an equivalent amount of the reduced form (NADPH), whose absorbance is measured at 340 nm. For each analysis, 3 μL of sample and 300 μL of reagent were employed. The temperature was kept at 37 ºC, and the reaction time was 10 min. Figure 1 shows the enzymatic procedure.

Fig. 1
figure 1

Scheme of enzymatic reactions for the determination of D-glucose and D-fructose

Full size image

Chromatographic Procedure

Chromatographic analyses were carried out by high-performance liquid chromatography-refractive index detection (HPLC-RID) using a HPLC system from Waters (Milford, MA, USA). The system consisted of a quaternary solvent pump (2600 model), an autosampler (717 Plus model) and refraction index detector (2414 model). Data were processed using Empower 2 Pro software (Waters).

The chromatographic procedure was based on the reference method of OIV (method OIV-MA-AS311-03, 2016). A 10 mL of sample was diluted in 20 mL of acetonitrile:water (40:60), filtered through a 0.45-μm polytetrafluoroethylene (PTFE) filters (Teknocroma, Barcelona, Spain), and a volume of 10 μL was injected. The HILIC column was a Nucleodur®, 100–5 NH2-RP (250 × 4 mm, particle size 5.0 μm) purchased from Macherey Nagel (Düren, Germany) protected by a Purospher® STAR NH2 (5 μm) pre-column from Merck (Darmstadt, Germany). The temperature of the column compartment and the RID cell was kept at 35 ºC. The elution was isocratic employing acetonitrile:water (80:20) as mobile phase at a flow rate of 1.2 mL min−1.

Results and Discussion

Method Validation

Sample preparation of wines is simple, usually limited to filtration with several materials. In some cases, solid-phase extraction has been used for removing interfering compounds, especially phenolic compounds (Başkan et al. 2016; Sabir et al. 2010). However, it has been reported that the direct injection decreases the time and cost of analysis with sufficient precision and accuracy (Eyéghé-Bickong et al. 2012).

By applying the enzymatic method, filtration was unnecessary as wine samples appeared clear and had no turbidity. The procedure establishes that if samples are very coloured, they must be decoloured with polyvinylpolypyrrolidone (PVPP) by shaking for 1 min, followed by centrifugation or filtration. In order to check the need of discoloration, two reference materials, a red wine (1.42 g L−1) and a white wine (2.7 g L−1), were analysed in triplicate without using PVPP. Since the obtained values (1.45 and 2.63 g L−1, respectively) were in good agreement with the reference values, the use of discoloration was discarded in order to eliminate the sample preparation step.

The method was validated in terms of linearity, limits of detection and quantification, precision and accuracy, according to the Eurachem Guide: The Fitness for Purpose of Analytical Methods – A Laboratory Guide to Method Validation and Related Topics (Magnusson and Ornemark 2014).

The linearity of the calibration curves was calculated using eight concentration levels (0.08, 0.2, 0.4, 0.90, 1.80, 3.60, 5.40 and 7.20 g L−1) by triplicate analysis. The five highest levels were injected directly from the commercial calibrator, whereas the three lowest levels were prepared from the 0.90 g L−1 standard by dilution of the necessary volume in water. An excellent linearity was obtained with a coefficient of determination (R2) equal to 0.9998.

The limit of detection (LOD) and the limit of quantification (LOQ), calculated as 3 and 10 times Sa/b (“Sa” being the intercept and “b” the slope of the calibration curve), were 0.04 and 0.13 g L−1, respectively. These values proved that the method was adequate for this analysis since the LOQ was similar to the one reported when using the automatic enzymatic method of OIV (0.10 g L−1) and lower than the one obtained with the official method recommended by the OIV (0.4 and 0.6 g L−1 for fructose and glucose, respectively).

The precision and accuracy of the method were evaluated employing six wine reference materials, three red wines and three white wines, with concentrations ranging between 0.5 and 4.9 g L−1. The precision was calculated measuring ten replicates of each reference wine on the same day (repeatability) and on different days during a week (intra-reproducibility). The results, expressed as the relative standard deviation (RSD), are collected in Table 1; the values were lower than 3.0% and 3.5% for repeatability and intra-reproducibility, respectively. It is worth noting that the obtained relative standard deviations are comparable to that provided for repeatability by the automatic enzymatic method of OIV (3%) and lower than the repeatability and reproducibility limit for the sum of glucose and fructose reported by the official method recommended by the OIV (10%). Furthermore, the Horwitz equation (RSD (%) = 21–0.5logC) was applied. This equation establishes the maximum value for coefficient of variation taking into account the analyte concentration. These values were obtained by the equations RSD (%) = 1/2 (21–0.5logC) for repeatability and RSD (%) = 2/3 (21–0.5logC) for intra-reproducibility, where C is the mass fraction expressed as exponent of 10. As observed, results were highly satisfactory because in all reference materials, the obtained values of RSD (%) for repeatability and intra-reproducibility were lower than the corresponding values calculated with the Horwitz equation.

Table 1 Data of precision of the automated enzymatic method evaluated by using reference wine samples (n = 10)
Full size table

Regarding accuracy, Table 2 presents the mean value obtained for the six reference materials with the proposed method (n = 10) and the assigned value, estimated through proficiency testing programmes, together with the corresponding tolerance values. As can be seen, in all cases, the mean values measured in the automatic analyser are in good agreement with the values of the reference materials. Furthermore, when comparing the values of the reference materials with the corresponding values obtained by the automatic analyser, a lineal relationship was established, showing that R2 was higher than 0.999 (Fig. 2). The 95% confidence interval calculated for the slope (0.9345–1.0090) and the intercept (− 0.0484–0.1302) did not differ significantly from the values of 1 and 0, respectively. Thus, the automated enzymatic method provided excellent results, in agreement with the declared values.

Table 2 Accuracy of the automatic enzymatic method evaluated by using reference wine samples (n = 10)
Full size table
Fig. 2
figure 2

Comparison between obtained and reference values for the content of glucose and fructose in wine reference materials

Full size image

Analysis of Samples

The sugar content, expressed as the sum of glucose and fructose, was determined in 5 red wines and 13 white wines, in duplicate analysis. For red wines, this content ranged from 0.20 to 0.50 g L−1, whereas this content was higher in white wines, ranging from 0.50 to 4.30 g L−1. According to the Commission Regulation (EC) (2009), all samples corresponded to the dry kind (values in the range 0–4 g L−1), except for one white wine (4.30 g L−1) which was classified as semi-dry. Although these values are consistent with those found in the literature, one can observe that the levels of residual sugar obtained in the analysed samples were lower than those reported by Tourtoglou et al. (2014) for white wines from Greece (1.23–12.29 g L−1). Regarding red wines, the obtained values were also lower than those reported by Bindon et al. (2013) in wines from Australia (0.40-–0.90 g L−1) and by Monteiro Coelho in wines from Brazil (0.20–2.05 g L−1). The lower levels observed in this study can be attributed to the fact that Galicia is a region located on the Atlantic coast of Spain characterised by a temperate climate and high rainfall amounts.

The Laboratorio Agrario e Fitopatolóxico de Galicia is an accredited laboratory (ISO 17025, 2017) for the determination of several physicochemical parameters in foods, drinks and soils. One of these accredited assays is the determination of sugars in several types of drinks by the reference chromatographic method. Therefore, in order to assess the quality of the obtained results with the automated enzymatic method, the 18 commercial samples were also analysed by the reference chromatographic method. For quantification purposes, individual stock standard solutions of 10 g L−1 of each sugar were prepared in acetonitrile:water (40:60). Then six working solutions of both sugars (0.1, 0.2, 0.5, 1.0, 2.5, 5.0 g L−1) were prepared by appropriate dilution of the 10 g L−1 standard solutions.

Figure 3 shows the relationship between the values obtained with the chromatographic method and the values measured by the automated enzymatic method. A good correspondence was achieved between the results obtained by the two methods, with a high regression coefficient (R2 = 0.9943). The 95% confidence interval calculated for the slope (0.9322–1.0039) and the intercept (− 0.0541–0.1119) included the values of 1 and 0, respectively. A Bland–Altman plot was also constructed to evaluate the agreement between the two methods (Fig. 4). As can be seen, the mean of the differences did not differ significantly from 0 (95% confidence interval from − 0.2569 to 0.1971). Therefore, there were no statistically significant differences between the results provided for both methods.

Fig. 3
figure 3

Correlation of the automated enzymatic method to the HPLC reference method for the determination of total sugar in wines

Full size image
Fig. 4
figure 4

Bland–Altman dispersion plot for the comparison of results obtained for the determination of total sugar in wines by the automated enzymatic method and the HPLC reference method

Full size image

Conclusions

In this work, an automatic analyser was validated to determine the total sugar content, expressed in terms of glucose and fructose, in wines. The procedure is based on enzymatic reactions followed by photometric determination of the generated NADPH. The method was validated in terms of linearity, limits of detection and quantification, precision and accuracy, achieving satisfactory results. As an application, 18 samples of wines from Galicia were analysed, verifying wine compliance requirements. The automated enzymatic method provided results statistically comparable to those obtained by the HPLC reference method of OIV.

The main advantages of this methodology when compared to other methods used to determine sugars in wines include rapidity of analysis and employment of a small volume of samples and reagents which, together with a minimum maintenance, involves a considerable reduction in laboratory costs. Furthermore, it can be considered an environmentally friendly methodology because of the avoidance of organic solvents and reduced waste generation.