The NSS and SS from the 10 volunteers (five women and five men) between 25 and 67 years of age were collected in the morning between 9 and 11 a.m. and for 5 min (Figure 1). The NSS was naturally collected in 15 ml centrifuge tubes (VWR, PA, USA), while for the SS, the volunteers chewed a piece of Parafilm™ during the collection (30). After this, the saliva samples were centrifuged at 15,000 g and 4°C during 15 min (31). The supernatants from the NSS and SS were pooled into two different tubes. The pH was determined in both the saliva mixtures SS (pH = 7.7) and NSS (pH = 7.6). Then, aliquots of 1.5 ml of saliva were prepared in Eppendorf tubes (VWR, PA, USA) and were frozen (−80°C) until the experiments. This storage technique has been previously proven not to affect the saliva EA (32). Figure 1 shows a schematic summary of the experimental procedure.

The volunteers did not report oral diseases. The donors were not allowed to eat or drink anything 1 h before the saliva collection. The participants were informed about the purpose of the study, and they gave their written consent to participate. These experiments were authorized by the Bioethical Committee of the Spanish National Council of Research (CSIC), Spain.

The saliva EA was measured following the colorimetric method previously described in the study of Pérez-Jiménez et al. (32) by using six different p-nitrophenyl (p-NP)-linked esters substrates provided by Sigma-Aldrich (Germany) and TCI Europe (Belgium), as detailed in the following section. The reaction mixture was composed of 25 μl of Tris 50 mM (pH = 8), 74 μl of saliva, and 25 μl of McIlvaine buffer (pH = 5) with the corresponding p-nitrophenyl (p-NP) ester substrate at 1.4 mM. The buffers Tris 50 mM (pH = 8)and McIlvaine (pH = 5) were prepared according to the procedure described by Pérez-Jiménez et al. (32). Tris was provided by GE Healthcare Uppsala, Sweden, and the reagents for McIlvaine buffer were provided by Panreac Química S.A., Spain (citric acid), and Merck. Darmstadt, Germany (Na₂HPO₄·12H₂O). The pH of the reaction mixture was 6.7.

The EA determinations were performed in m96 microplates (VWR, PA, USA). To do so, the microplates were incubated for 35 min at 36°C, and the absorbance (λ = 410 nm) was measured every 5 min (eight times in total) by using a microplate spectrophotometer (Multiskan™ FC Microplate Photometer, Thermo Fischer Scientific, MA, USA). The EA was calculated as the units of EA (UI) per min (14). One unit of EA was determined as the absorbance change of 0.01 optical density. The saliva EA was measured three times.

To determine the specificity of the saliva EA toward different carboxylic esters, six p-NP-linked esters were employed as substrates: p-NP-acetate (NPA), p-NP-butyrate (NPB), p-NP-hexanoate (NPH), p-NP-octanoate (NPO), p-NP-decanoate (NPD), and p-NP-dodecanoate (NPDD) (Figure 1). They differed among their hydrocarbon chain length (from C2 to C12, respectively). Six different 70 mM stock solutions (one per p-NP-ester) were prepared in 5 ml of methanol (Lab-Scan, Poland). Then, 1 ml of these 70 mM stocks were made up to 10 ml with McIlvaine buffer pH = 5, to obtain a 1.4 mM final concentration of the p-NP-esters in the microplate well.

To check the inhibitory effect of wine flavonoids toward saliva EA in conditions more similar to those of wine consumption, a new experiment was carried out. For this, NPB (four carbons) and catechin were selected among the six ester substrates and the five single wine polyphenols (Figure 1). The inhibitory effect of catechin was evaluated in the presence of ethanol [11.3% (v/v)] and under more acidic pH conditions (pH = 5). These assays were carried out in both the SS and NSS samples.

To evaluate the effect of ethanol, 14 μl of absolute ethanol (Panreac Química S.A., Barcelona, Spain) was added to the microplate well. To maintain the final volume of the reaction mixture (124 μl), the volumes of the reagents were recalculated: 21.3 μl Tris (50 mM, pH = 8), 67.4 μl saliva, 21.3 μl McIlvaine buffer (pH = 5), and 14 μl ethanol.

To evaluate the wine tasting conditions in the terms of acidic pH on EA, a preliminary assay was performed to know the salivary pH after the wine tasting. The results are shown in Supplementary Figure 1. It was found that the pH of saliva immediately after wine rinses decreased to 5, thus, pH = 5 was selected to simulate the wine consumption conditions.

Then, the pH of the reaction mixture in the well was modified by replacing the Tris (50 mM, pH = 8) buffer with the McIlvaine (pH = 4.56) buffer. The effect of changing the buffer was previously checked, showing that the EA measurements were not affected.

To evaluate the effect of both factors (ethanol 11.3% and pH = 5), the volumes of the reaction mixture were recalculated as explained before and the Tris (50 mM, pH = 8) buffer with the McIlvaine (pH = 4.56) buffer. Then, the EA was measured following the same procedure as explained in section Saliva esterase activity measurement.

One-way ANOVAs and Tukey's tests for mean comparison were applied to check for the significant differences (p < 0.05) in the EA: between the SS and NSS, to check the inhibitory effect of the single and mixtures of wine polyphenols and to determine the influence of ethanol (11.3%) and acidic pH (pH = 5) in the inhibitory effect of catechin. To do so, version 2020.3.1 of the XLSTAT (Addinsoft, Paris, France) software was used.

To determine the specificity of esterases from both SS and NSS for carboxylic esters, the EA was measured by using six p-NP-ester substrates (NPA, NPB, NPH, NPO, NPD, and NPDD) that differed in carbon chain length (from C2 to C12). The mean values of EA (UI/min) and SD are shown in Figure 2.

As it can be seen (Figure 2A), the EA was higher in the SS than in the NSS for all the p-NP-esters, and these differences in the EA between the two saliva types increased from 11 to 50% as the length of the ester chain increased. In Figure 2B, the EA from SS was considered as 100%, and from here, the percentage of EA from NSS was calculated. As it can be seen, the EA from SS measured with smaller esters (NPA, NPB, and NPH) was between 11 and 37% higher than EA from NSS with the same esters (Figure 2B). In the case of longer esters (NPO, NPD, and NPDD) the differences in the EA between SS and NS were even higher (42–50% higher in SS than in NSS). The main difference between the two saliva types is that SS is mostly produced by the parotid gland, while NSS is produced by sub-mandibular glands (41). Therefore, these results indicated that the saliva EA mainly proceeds from parotid saliva, which is in agreement with the previous findings (42). Apart from parotid saliva, epithelial cells and oral microorganisms are proposed as the other important sources of saliva esterase (43, 44).

On the other hand, Figure 2A shows that the higher the carbon chain length of the ester, the lower the EA, independent of the type of saliva (SS and NSS). This trend was observed for all the p-NP-ester substrates, except for the shortest ester NPA (two carbons), in which the EA from both the saliva types was between 60 and 70% lower than in the NPB (four carbons). Greater hydrolysis toward butyrate than acetate esters by carboxylesterases from saliva (9, 15) and liver (45) is seen before. Apart from NPA, as the size of the ester increased, the EA decreased. Therefore, the NPB substrate showed the highest EA in the two saliva types, which was 50–55% greater than the activity using NPH (six carbons), and up to 97% higher than in the case of longer chain esters, such as NPD and NPDD (10 carbons or more). A lower EA toward the longer esters instead of smaller esters is shown by other authors in the case of human liver carboxylesterases (46) and carbonic anhydrases (47, 48). In the case of human liver carboxylesterases, an increase in the affinity constant between the ester and the enzyme as the hydrophobicity (log P) of the ester increased is shown (45, 46). A reduction in the maximum rate of reaction (Vmax) is produced as the size of the product (carboxylic acid) increased, since more hydrophobic products are located in the active site gorge, reducing the activity of the enzyme (45, 46). Therefore, on one hand, it seems that there is a greater affinity of esterases for the more hydrophobic esters, which could lead to the larger carboxylic acid products, which in turn, decreases the reaction rate, and thus, the activity of the enzyme. But, on the other hand, the activity of esterases increases for the smaller esters, which, although have a lower affinity for the enzyme, lead to small carboxylic acid hydrolysis products, which do not affect the activity rate of the esterase enzymes, contrarily to what happens with the long esters. Thus, the size of the degradation products seems more important for the enzymatic reaction rate than the affinity between the enzyme and the ester. Therefore, for small esters (such as NPA and NPB), the higher the log P of the ester, the higher the EA, while as the size of the carboxylic acid product increases (six carbons or more), the EA decreases. Although this premise was not previously proven for saliva esterases, it could be a plausible hypothesis to explain these results. Additionally, this hypothesis might explain the higher persistence of long than short chain esters in the breath after wine consumption (49).

Prior to the inhibitory assay of EA by the wine phenolic compounds, the use of B4NP as a positive control for the inhibition of saliva EA was checked, as B4NP is widely known to be a specific inhibitor of carboxylesterases (37, 50, 51). For this, B4NP was added to SS and NSS at different concentrations in the microplate wells (from 0.05 to 0.8 mM), and the EA was then measured following the same procedure as previously described.

The results showed that the inhibitory effect of B4NP was not dose-dependent as seen previously (37, 51, 52). The maximum inhibitory effect of B4NP toward EA was found at 0.05, 0.1, and 0.2 mM concentrations of B4NP (Supplementary Figure 2). At these concentrations, B4NP inhibited ~65–70% of EA in both the saliva types. These results confirmed that, although there are more enzymes involved in salivary EA (15, 18), carboxylesterases are the main enzymes that contribute to this activity in saliva (9, 17).

Therefore, 0.1 mM concentration of B4NP was selected for the following assays, which is within the range that was previously used by other authors (37, 50, 53, 54). This concentration provided the highest EA inhibition (65 and 69% in SS and NSS, respectively).

Finally, to evaluate how EA could be modified during the wine tasting conditions and if phenolic compounds could also inhibit this enzymatic activity, a new assay was set up. For this, the EA was first measured in the presence of ethanol and acidic conditions. The results (Supplementary Figure 1) showed, as expected, that salivary pH decreased (average values from 6.7 to 5.6,) immediately after wine expectoration (0 s). Interestingly, 30 s after wine rinsing, the salivary pH was practically the same (pH = 6.3) as the baseline pH (pH = 6.7), which indicated that a short recovery time is needed by saliva to return to its baseline pH value, as observed in the previous works (57).

The EA was then, measured in the saliva samples at acidic pH (pH = 5) or standard salivary pH (pH = 7), in the presence or absence of polyphenol (catechin), and also in the presence and absence of ethanol (11.3% v/v). The results, expressed as a percentage of EA, are shown in Figure 5.

As it can be observed (Figure 5), the presence of ethanol has a significant effect on reducing the saliva EA. A decrease of about 29% in SS (Figure 5A) and 15% in NSS (Figure 5B) compared with control saliva (without ethanol) was observed. Although the presence of ethanol on saliva EA is scarcely investigated, the effect of acute ethanol consumption has been shown to significantly decrease the activity of other saliva enzymes, such as peroxidases and isozymes (58). The presence of ethanol or its metabolites could have inactivated the saliva esterases. The reduction in the EA from saliva could explain the higher persistence of carboxylic esters after the rinses with the wines supplemented with ethanol (5 and 10%) compared with the wines without ethanol (0.5%) observed in the previous works (22). In addition to this, the EA of the two saliva types also reduced about 12% for SS and 11% for NSS in the presence of both catechin and ethanol compared with the saliva without catechin and ethanol (Figure 5). These results showed that the decrease in saliva EA was higher when both the factors (catechin and ethanol) were evaluated at the same time than when they were evaluated individually.

Regarding salivary pH, the EA of both saliva types (SS and NSS) was significantly reduced by more than 90% when the pH of the well-decreased to 5 and in all the assayed conditions, with or without catechin and ethanol (Figure 5). These results indicated that under the acidic pH conditions (pH = 5), the ability of esterase enzymes to hydrolyze esters was compromised, which is in agreement with the previous findings, in which a pH between 7 and 8 has been established as the optimal pH for saliva esterases (16, 20, 59). Despite this, in a recent work, the authors determined the EA in saliva collected after mouth rinsing with an acidic (pH 3.5) water solution (32). These contradictory findings could be explained by the short time required for saliva to return to normal pH values, as previously observed in Supplementary Figure 1. Although the esterases barely showed enzymatic activity at acidic pH, saliva is constantly being generated and replenished, and as the pH of the oral environment return to the normal values within few seconds (Supplementary Figure 1), the esterases from saliva could also keep their activity. This suggests that the saliva EA might be reduced for a short time when saliva is in contact with the wine, but the EA activity can quickly be recovered when the pH of the saliva returns to more neutral values. This fact could be of great importance for the hydrolysis of esters remaining in the oral cavity during the wine tasting and thus, for the long-lasting wine aroma perception. Further studies focusing on the dynamic changes in the saliva composition and EA during a consumption episode will be achieved in future works to confirm this premise.