Forty wine samples were analyzed: fino (n = 10), manzanilla (n = 10), amontillado (n = 10) and oloroso (n = 10). Due to the particular ageing system, the bottled wine was a mixture of wines from different years. The samples were purchased from supermarkets and small specialized wine shops and analyzed during 2003. The samples were stored in their original bottles at 4-5 °C until analysis was carried out. The day before the analysis, the samples were taken out the fridge in order to temperate them.

Ochratoxin A (99.6%) was purchased from Sigma (St. Louis, MD, USA). All of the reagents were pro-analysis grade. Sodium acetate and phosphoric acid were purchased from Panreac (Barcelona, Spain); acetonitrile and methanol LC grade were purchased from Riedel-de Haën (Seelze, Germany). Phosphate buffered saline solution (PBS) was prepared diluting 9.55 g of PBS DULBECCO from Biochrom AG (Berlin, Germany) in one liter of water. Ochratest Immunoaffinity Columns (IAC) were obtained from Vicam Inc. (Watertown MA, USA). Millipore type I water was used to prepare all of the aqueous solutions and was obtained daily from a Milli-Q water purifying system.

A stock standard solution of 100 mg/L was prepared by dissolving 1 mg of OTA in 10 mL of methanol. It was then stored at −20 °C. It has been reported that OTA solutions in methanol stored at -20 °C are stable over a period of several years [19]. The exact concentration was determined espectrophotometrically at 333 nm (MW: 403.8; ε: 5500 M⁻¹ cm⁻¹) [20]. Working standard solutions and calibration samples were prepared by dilution of stock solution with methanol. Prior to the LC analysis, 200 μL of the calibration samples were evaporated at 40 °C under nitrogen stream and redissolved in 200 μL of mobile phase, in the same way in which the methanol extracts from wine samples were prepared. Quality control of analysis has been carried out by analyzing standard solutions among the wine samples. Moreover, during the years 2002-2005 the laboratory, using the described method, participated in four rounds (1720, 1728, 1736, 1745) for OTA analysis in wine of Food Analysis Performance Assessment Scheme (FAPAS), obtaining satisfactory z scores (|z| ° 2) in each case.

The method was based on that which has been described by Otteneder and Majerus (2000) [11], with some modifications. A 5 mL aliquot of wine was diluted with 45 mL of PBS. After shaking, all of the mixture was passed through an IAC column which was preconditioned with 10 mL of PBS. Afterwards, the IAC was washed with 10 mL of PBS and 10 mL of water. Finally, it was dried by passing air with a syringe. OTA was eluted with methanol, 4 times with 1 mL each at a flow rate of 20-30 drops/min. The eluate was evaporated to dryness in a water bath at 40 °C under nitrogen stream, and the residue was redissolved in 200 μL of mobile phase before LC analysis.

The instrument used was an Agilent Technologies 1,100 liquid chromatographic system equipped with a fluorescence detector (model G1321A), controlled by a Chemstation 3D software. The chromatographic conditions were based on the method of Jimenez et al. [21]. Briefly, OTA was analyzed on a 5 μm (25 cm × 0.4 cm) Tracer Extrasil ODS2 column with a Tracer Extrasil ODS-2 precolumn, both from Teknokroma (Barcelona, Spain). The injection volume was 100 μL and the flow rate was 1.5 mL/min, with a mobile phase of 29:29:42 (v/v/v) methanol-acetonitrile-sodium acetate (5 mM acidified to pH 2.2 with phosphoric acid). The aqueous phase was filtered through a 0.45 μm membrane filter (Millipore, Ibérica S.A. Spain). Chromatography was performed at 40 °C and the fluorescence conditions were: Ex = 225 nm, Em = 461 nm. in order to achieve a low limit of detection (LOD) [22]. In these chromatographic conditions, OTA has a retention time of 5.5 min.

The selectivity of the method was assured by the use of immunoaffinity purification techniques and a very selective fluorescence detector. Furthermore, some experiments were carried out in order to assess selectivity. When OTA was added to positive samples, an increase of OTA peak area was observed. Moreover, when the proportion of the aqueous component of the mobile phase was increased, the retention time in the OTA peak was delayed, and no other peak, differing from that of OTA, appeared.

Two calibration curves (seven and eight data points, three replicates for each point) were generated by plotting peak areas of OTA versus concentration of calibration samples for two ranges, 0.4 to 1 ng/mL and 1 to 10 ng/mL (equivalent to 0.02-0.04 and 0.04-0.4 ng/mL in wine). Both calibration curves showed a good linear relationship between peak areas and OTA concentration in both ranges.

Within- and between-day precision and recovery were established by making 27 determinations in spiked wine samples, covering the range of the method (three concentrations of 0.02, 0.1 and 0.4 μg/L/three replicates each one/three different days) and achieving a recovery value of 95.4% ± 5.5%. The limit of detection (LOD) (signal/noise ratio of 3) and of quantification (LOQ) (the lowest concentration assayed that gave good precision and accuracy) were 0.009 and 0.02 μg/L, respectively. Moreover, the good results obtained from FAPAS tests demonstrated the suitability of the analytical methodology for OTA determination in wine. With this method, the laboratory was accredited during the years 2002-2005 for OTA determination in wine by the Spanish Accreditation Board (ENAC) under the Standard ISO 17025.

Figure 2 shows chromatograms obtained after analysis of samples of each type of wine.

The results obtained from the 40 samples analyzed for OTA content are shown in Table 1.

These results are summarized in Table 2. All of the data has been corrected by recovery (95.4%) and all of them have been taken into account in the calculation, even those values below LOD, in this case, a value of LOD/2 have been assigned.

All of the samples presented OTA concentration levels below the maximum tolerate level (2 µg/L). Thirty-two (80%) samples contained detectable OTA amounts, and the overall OTA mean concentration was 0.138 μg/L. Taking into account the mean value obtained in this study (0.138 µg/L) and the mean obtained from oloroso wine (0.319 µg/L), and assuming that a consumer weighs 60 kg with a 50 mL daily intake of these aperitif wines (consumption of this type of wine is occasional and rarely exceeds 50 mL/day [1]), the estimated weekly intakes of OTA are 0.8 and 1.9 ng/kg body weight (bw), respectively, far from the provisional weekly tolerable intake (PTWI) recommended by the Joint FAO/WHO Expert Committee on Food Additives (JECFA) for OTA (100 ng/kg bw) [2].

These types of wines are elaborated from white grapes and they are considered to be white wines. The results obtained in this study can be compared with those reported from different studies carried out on white wines in different countries, in which an OTA level range between <LOD-1.72 µg/L was obtained. These are summarized in Table 3.

Considering all of the wines studied and the fact that they are white wines, they presented OTA levels similar to those obtained in other studies carried out in Spain [5,25], Greece [8,10] or Italy [13].

From all of the wines studied in this work, manzanilla and fino wines are considered to be aperitif wines. In the reference literature, Burdaspal and Legarda [24] have studied 47 aperitif wines from six different European countries, obtaining an OTA mean concentration of 0.040 μg/L. Among them, Spanish wines (n = 27) presented a slightly higher OTA mean value, 0.054 μg/L. Zimmerli and Dick [12] studied two Sherry aperitif wines (type not mentioned) and obtained 0.042 μg/L as the OTA mean value, similar to the Burdaspal and Legarda data. These reported results coincide with the levels obtained in the fino and manzanilla wines (OTA mean levels 0.044 μg/L and 0.042 μg/L, respectively).

Comparison of the OTA levels in the four different wine groups was carried out using the median non-parametric test. There are significant differences between the groups fino-oloroso (p < 0.001) and manzanilla-oloroso (p < 0.001), while significant differences do not appear between fino and manzanilla wines (p = 0.656), fino and amontillado (p = 0.179), manzanilla and amontillado (p = 0.179) or amontillado and oloroso wines (p = 0.179).

The wines that suffer biological ageing have OTA levels which are significantly lower than oloroso wine, which after suffering an oxidative ageing, have the greatest OTA concentration; and although no significant differences have been observed, it appears that as the biological ageing diminishes, the OTA levels that are obtained tend to be greater. Also, the incidence of OTA in samples is lesser in those wines that suffer a biological ageing (70% in manzanilla wines and 50% in fino wines) in contrast with the 100% of the samples with detectable OTA levels in amontillado and oloroso wines, as can be observed in Figure 3.

Taking into account these considerations, it could be assumed that there is some type of relationship between the presence of OTA in wine and the presence or absence of the blooming veil, because the difference among these wine groups is the ageing process. Different authors have described that yeast can potentially be used for removing OTA from wine [29]. Several hypotheses could be made: yeast could use OTA as a source of carbon for obtaining energy, thereby reducing the OTA levels in these types of wines [17]. Also, yeast or yeast cell walls could adsorb OTA [30], acting like a sponge [18], by different mechanisms such as hydrogen bonding and/or ionic or hydrophobic interactions enhanced by low pH values (near 3.00) [31]; and the wines under study have an average pH of 2.9. Thirdly, it is known that there are certain enzymes with the ability to degrade Ochratoxin A, such as lipase A [32] or carboxypeptidase A [33,34]; these enzymes might be present in yeast, thereby decreasing the OTA level in the medium.