Abstract
The effect of chitin [poly(N-acetyl-1,4-ß-d-glucopyranosamine)], an abundant, low-cost natural polymer, on white wine stabilization on a laboratory scale was studied in comparison with bentonite fining. Treatments of an unfined wine with increasing doses of chitin allowed a reduction of up to 80% of the haze induced by the heat test, which corresponded to a reduction in wine protein content of less than 29%. In contrast, bentonite fining, although allowing a complete stabilization, resulted in the removal of almost all the proteins from wine. These results suggest that chitin can remove from wine protein components involved in haze formation more specifically than bentonite. SDS-PAGE analysis of both the proteins remaining in wine and those adsorbed by chitin confirmed this specificity. Chitinolytic activity detection after SDS-PAGE separation demonstrated that a main protein component removed by chitin corresponded to the class IV chitinase of grape origin involved in white wine instability. Because class IV chitinases are characterized by bearing a chitin-binding domain, a specific interaction of these wine proteins with chitin can be suggested. Preliminary trials with chitin immobilized in a column system indicated the possibility to regenerate this matrix and to use it continuously for white wine stabilization. However, the effects on both the organoleptic quality and the long-term stability of white wines treated with chitin need to be determined in the actual winemaking conditions.
The presence of haze in bottled white wines results in a serious quality defect because turbidity makes the wine undesirable for consumers. Wine proteins, which have the tendency to insolubilize during wine storage (Bayly and Berg 1967, Hsu and Heatherbell 1987, Waters et al. 1992), are the main cause for this defect. However, the hazing potential of a given wine does not seem to correlate with its total protein concentration (Bayly and Berg 1967), suggesting a different contribution of individual wine proteins to the phenomenon of haze formation (Hsu and Heatherbell 1987, Waters et al. 1992).
The major proteins present in white wines were first identified by Waters et al. (1996), as pathogenesis-related (PR) proteins, the most abundant of the grape berry (Waters et al. 1998). These proteins, which include different chitinases [poly(1,4-N-acetyl-ß-d-glycosaminide) glycanohydrolase, EC 3.2.1.14] and thaumatin-like proteins, are particularly stable in winemaking conditions (low pH, proteolysis), and therefore pass into the wine, where they can precipitate, causing haze or sediments (Waters et al. 1996, 1998).
The traditional method to stabilize white wines against haze formation is based on bentonite fining. The negatively charged bentonite particles interact electrostatically with the wine proteins, allowing their removal (Ferreira et al. 2002). However, this treatment has some drawbacks because bentonite adsorption is rather aspecific and, in addition to proteins, also removes different molecules or aggregates, including aroma and flavor compounds, resulting in organoleptic changes in the wine (Miller et al. 1985, Voilley et al. 1990). Moreover, a certain quantity of wine is lost as bentonite lees (Lagace and Bisson 1990). For these reasons, alternative procedures for protein removal from white wines have been developed, such as fining with silica sol/gelatine (Millies 1975), use of immobilized tannic acid (Weetall et al. 1984) and proanthocyanidins (Powers et al. 1988), use of exchange resins (Sarmento et al. 2001), adsorption on the surface of metal oxides (Pashova et al. 2002, 2004), and ultrafiltration (Hsu et al. 1987). All these procedures allow some reduction of the protein content, but most have some drawbacks, mainly related to alteration of the organoleptic characteristics of the wine (Feuillat et al. 1987, Miller et al. 1985, Millies 1975).
Chitin [poly(N-acetyl-1,4-ß-d-glucopyranosamine)] is a low-cost, abundant, natural insoluble polymer present in the shells of crustaceans and in other invertebrates. It has several potential industrial uses, including food applications (Shahidi et al. 1999). Moreover, chitin is the substrate for the action of chitinases, some of which (class I and IV) possess a chitin-binding domain, different from the catalytic site, which is able to interact specifically with the polymer (Graham and Sticklen 1994). By exploiting this property, and by considering the major role of wine chitinases in haze formation (Waters et al. 1996), we have hypothesized the use of chitin to remove these enzymes from white wines in order to improve their protein stability. Here we report the preliminary results of such an approach.
Materials and Methods
Materials.
A white wine (obtained from Vitis vinifera cv. Incrocio Manzoni 6.0.13 grapes) taken in the winery before bentonite fining was used in all the experiments. Two different types of chitin [poly(N-acetyl-1,4-ß-d-glucopyranosamine)] from crab shells (chitin A and B, purchased from Fluka, Milano, Italy, and Sigma, Milano, Italy, respectively) and bentonite (Volclay KWK, Vason, Verona, Italy) were used. Before use in wine, chitin samples were exhaustively washed with 2 M NaCl and 10% (vol/vol) methanol in succession.
Wine treatments with chitin and bentonite.
Aliquots of 10, 50, 100, and 200 mg of chitin were washed in a solution containing 10% (vol/vol) ethanol and 1.25 g/L tartaric acid and brought to pH 3.5 with KOH (washing solution). The suspension was centrifuged, and the washing was repeated again on the pellet. After recentrifugation, 10 mL of wine was added to the pellet and the suspension was gently stirred for 2 hr at room temperature. Samples were centrifuged (3500 g, 5 min) and the supernatants were filtered on a Whatman no. 4 filter paper.
After hydration for 16 hr at 5% (wt/vol) in water, 1.0, 5.0, 10, and 20 mg (dry wt) of bentonite was added to 10 mL of wine and left in contact as for chitin. Four replications were run for each treatment. Each replicate was heat-tested and analyzed for polyphenol and protein concentrations.
Heat test.
Aliquots of wine taken before and after treatments with chitin and bentonite were heated for 2 hr at 80°C and then cooled at 4°C for a further 2 hr (Pashova et al. 2002). After allowing the samples to reach the room temperature, their absorbance was read at 560 nm with an ATI-Unicam (Cambridge, UK) spectrophotometer. The blank was the wine before heating. Each measurement was the average of three replicates.
Measurement of total polyphenol content.
Total phenolics were measured with the procedure described by Waterhouse (2001) for small volumes of wine. Results were expressed as gallic acid equivalents. Each measurement was the average of eight replicates.
Measurement of total protein content.
Protein was precipitated from 1.0 mL of wine (three replicates for each sample) by the KDS method (Zoccatelli et al. 2003). Precipitated proteins were resolubilized in 1.0 mL of distilled water and quantified by the bicinchoninic acid method (Smith et al. 1985) using the BCA-200 protein assay kit (Pierce, Rockford, IL) with bovine serum albumin (Sigma) as the standard (Zoccatelli et al. 2003). Each measurement was the average of eight replicates.
Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE).
SDS-PAGE in gels containing a total polyacrylamide concentration of 14% (wt/vol) was performed according to Laemmli (1970) in a Mini-Protean III apparatus (Bio-Rad, Segrate, Milan, Italy). Samples were prepared by precipitating proteins from 0.2 mL of wine by the KDS method (Zoccatelli et al. 2003). Precipitated proteins were resolubilized in 30 μL of 0.5 M Tris HCl, pH 6.8, 15% (wt/vol) glycerol and 1.5% (wt/vol) sodium dodecyl sulfate (SDS) (sample buffer). Electrophoresis was carried out at 18 mA constant current until the tracking dye Bromophenol Blue ran off the gel. Gels were stained for total proteins by Coomassie Brilliant Blue R-250 and destained with 7.5% (vol/vol) acetic acid. Molecular weight standard proteins (Bio-Rad) were myosin (200.0 kDa), ß-galactosidase (116.3 kDa), phosphorylase B (97.4 kDa), bovine serum albumin (66.2 kDa), ovalbumin (45.0 kDa), carbonic anhydrase (31.0 kDa), soy trypsin inhibitor (21.5 kDa), lysozyme (14.4 kDa), and aprotinin (6.5 kDa).
Chitinolytic activity detection on SDS-PAGE gels.
Chitinolytic activity detection was performed with the procedure of Trudel and Asselin (1989), as described by Vincenzi and Curioni (2005). Briefly, the samples, solubilized in sample buffer, were loaded on SDS-PAGE gels containing 0.1% (wt/vol) of glycol chitin. At the end of the electrophoretic separation, the enzymes were renaturated by incubating the gels for 16 hr in 50 mM Na acetate, pH 5.5, containing 1% (wt/vol) Triton X 100 (Sigma). The gels were incubated for 20 min in a 0.5 M Tris HCl buffer, pH 8.9, containing 0.01% (wt/vol) Calcofluor white MR2 (Sigma), and then in water for at least 1 hr. Protein bands with chitinolytic activity were detected under UV light.
Protein elution from chitin.
Aliquots of chitin (200 mg, dry wt) treated with washing solution were incubated with 10 mL of wine for 2 hr. After centrifugation, the precipitate was washed twice with washing solution and then treated alternatively with 10 mL of one of the following solutions: triethylamine 0.1 M pH 11, 2 M NaCl, or N-acetyl-d-glucosamine (40 mg/mL of water) for 2 hr. After centrifugation, the supernatants were dialyzed (3.5 kDa cut-off) against water, freeze-dried and dissolved in 0.1 mL of SDS-PAGE sample buffer, and 40 μL of the resulting solution was loaded on SDS-PAGE gels.
Wine treatments in chitin column.
A glass chromatographic column (20 x 1.8 cm) equipped with a postcolumn UV detector was filled with 2.0 g (dry wt) of chitin (not washed) suspended in water. During the experiments, the absorbance at 280 nm of the eluate was continuously monitored. The column was washed by flushing a 2 M NaCl solution until the absorbance was stable. The same was done with 10% (vol/vol) methanol. Finally 10 bed volumes of the washing solution (10% ethanol [vol/vol] and 1.25 g/L tartaric acid brought to pH 3.5 with KOH) were flushed and the column was left to dry. By using a peristaltic pump, 100 mL of unfined wine was recirculated through the column at constant volumetric flow of 1.5 mL/min until the absorbance was stable. The wine was recovered, and the whole procedure was repeated again with another aliquot of unfined wine, starting from the washing with NaCl. The two samples of wine obtained in this way were heat-tested for protein stability determination.
Results
Effect on wine stability.
A sample of Incrocio Manzoni 6.0.13 wine, taken in the winery before bentonite fining, was treated with increasing doses of two different commercial types of chitin. The reduction of the haze induced by the heat test was determined and compared to that obtainable by bentonite fining (Figure 1⇓). In the range of the absorbance values measured in these experiments, the relationship between particle concentration and absorbance at 560 nm was linear (not shown). Addition of bentonite to wine at doses as low as 0.5 g/L resulted in a complete disappearance of the haze induced by the heat test, confirming the effectiveness of this material in removing the factors responsible for white wine instability. In contrast, the same result was never obtainable with chitin, even at the highest dose tested. In fact, in the wine treated with 20 g/L of chitin, the greatest reduction of the haze induced by the heat test was ~70 to 80% with respect to the untreated wine. However, the slope of the dose/response curve changed at 1.0 g/L for both the chitin samples, indicating that the efficiency of chitin in reducing the heat-induced instability was dose dependent. For both the samples, at 1.0 g/L of chitin the haze caused by the heat test was approximately half of that of the untreated wine. These results indicated the possibility to exploit chitin to reduce the instability of white wines.
Effect on wine polyphenols.
Because wine polyphenols can participate in the phenomenon of haze formation (Ferreira et al. 2002), the wines treated with increasing doses of the two chitin samples were assayed for total polyphenol concentration (Figure 2⇓). Large differences in polyphenol concentration were noted among the samples deriving from the different replications of the same treatment with chitin, indicating that its effect on polyphenols would be highly variable. However, the shape of the dose/response curve for the reduction of the polyphenol content did not correspond to that of the reduction of haze (Figure 1⇑). In comparison to the control wine, chitin treatments led to a progressive reduction of the polyphenol content from average values of ~0.2 g/L to ~0.15 g/L of gallic acid equivalents, corresponding to a reduction of 25% at the highest dose of chitin (20 g/L) used (Figure 2⇓). This result confirmed a relatively low capability of this material to reduce the phenolic content in wine, as already reported by Spagna et al. (1996). For comparison, the wines treated with bentonite were also tested and, at the highest dose tested, a reduction of ~20% of the initial polyphenol content was obtained (Figure 2⇓).
Effect on wine protein concentration.
Wine proteins are components mainly involved in determining instability in white wines (Bayly and Berg 1967, Hsu and Heatherbell 1987, Waters et al. 1992); accordingly the reduction of wine protein content caused by the addition of both chitin and bentonite was measured (Figure 3⇓). Treatments with increasing doses of bentonite resulted in a steep reduction of the protein concentration in wine, which decreased from ~420 mg/L of the original wine sample to ~50 mg/L (88% reduction), confirming that bentonite has a very high capability to adsorb proteins in wine (Hsu and Heatherbell 1987). On the other hand, the wine samples treated with chitin showed only a low reduction of the protein content (from ~420 to ~300 mg/L, less than 29%), which was obtained with both the chitin samples used. Also in this case the curve was steeper with doses up to 1.0 g/L of chitin (14% of the original protein removed), whereas the curve was much flatter when higher doses of chitin (from 5.0 to 20 g/L) were used (Figure 3⇓), resembling what could be observed for the reduction of the heat-induced haze (Figure 1⇑). However, in the case of chitin treatments, the heat instability did not seem proportional to the quantity of protein remaining in the wine. A 14% reduction of the protein content, obtainable with 1.0 g/L of chitin, corresponded to a reduction of the haze induced by the heat test as large as 50% (comparing the results shown in Figures 1⇑ and 3⇓). By calculating the ratio between the percent reduction in haze and the percent quantity of protein removed from wine, a maximum value of 3 was obtained for chitin (at doses higher than 5.0 g/L); whereas when bentonite was considered, the same ratio reached a maximum of only about 1.6 for doses of 0.1 g/L and decreased to about 1 for larger doses (Figure 4⇓). This would indicate that chitin had a more specific effect compared to bentonite in removing from wine protein components more heat-instable than others.
Effect on wine protein pattern.
SDS-PAGE analysis of the protein remaining in wine after treatments with both bentonite and chitin was performed in order to detect qualitative differences in protein bands (Figure 5⇓). Protein samples were not reduced before the electrophoretic separation, because in this condition a higher number of bands could be detected in wine compared to the SDS-PAGE of reduced samples (not shown). In nonreducing conditions, the relative mobility of some protein bands was different from that observed after reduction of the samples (not shown) (Vincenzi and Curioni 2005). For example, the main 24 kDa component detectable in reducing conditions which was shown to correspond to a thaumatin-like protein (Pocock et al. 2000, Tattersall et al. 1997, Waters et al. 1996) migrated to a position corresponding to ≈20 kDa when the wine proteins were not reduced (Figure 5⇓).
The treatments with chitin, even when used at the highest dose, caused only very slight modifications of the wine protein pattern. In particular, the low molecular weight band of ≈9 kDa completely disappeared even at the lowest dose of chitin, whereas a progressive reduction of the 31 kDa band was noted. This band became almost undetectable when chitin was used at the highest dose (20 g/L). Unlike chitin, bentonite treatments resulted in a nonspecific removal of nearly all the protein components, which completely disappeared in samples treated with doses of 1.0 g/L or higher (Figure 5⇑).
Characterization of wine proteins adsorbed by chitin and bentonite.
In order to study the proteins removed by chitin, a chitin sample was incubated in wine and treated with different solutions that were shown to allow the release of the proteins bound to that polymer. Elution with triethylamine at pH 11 gave a protein sample with a SDS-PAGE pattern that was difficult to interpret because of a dark background appearing after Coomassie staining (not shown). The precise reason for this effect was not investigated, although some chitin degradation may have occurred at pH 11, releasing compounds of nonprotein nature stained by Coomassie. In contrast, both 2 M NaCl and N-acetyl-d-glucosamine, which is the structural monomeric unit of chitin, gave similar interpretable patterns (not shown). Therefore, 2 M NaCl was chosen to recover the proteins adsorbed by the chitin incubated in wine. The SDS-PAGE analysis of the fractions eluted by the salt solution showed the presence of only some bands of those present in the original wine (Figure 6⇓). In particular at the highest dose of chitin used (20 g/L), the salt solution released three main bands showing relative molecular masses (Mr) of ≈31, 20, and 9 kDa. These bands were certainly of wine origin, because the chitin sample was exhaustively washed with NaCl before its contact with the wine. At doses of 1.0 and 5.0 g/L of chitin, the band of ≈20 kDa was not detectable, indicating a low affinity of this protein for chitin. The other proteins of the original (untreated) wine sample were not or only barely detectable in the solution eluted from chitin, confirming that this polysaccharide interacted specifically with some protein components but not with others. For comparison, the same procedure was applied to the bentonite samples used in different doses for treating the wine (Figure 6⇓). In this case, the salt solution released protein bands that were different from those bound by chitin, with the exception of those of ≈9 and 20 kDa, confirming that chitin and bentonite had different effects in removing specific protein components from wine. The results obtained for chitin are in good accordance with those obtained when the proteins remaining in wine were analyzed (compare Figures 5⇑ and 6⇓). In the case of bentonite adsorption, which seemed to remove completely the proteins from wine at doses as low as 1.0 g/L (Figure 5⇑), the slight discrepancy between the results (Figures 5⇑ and 6⇓: see for example the behavior of the ≈31 kDa band, which is absent in the bentonite-treated wine and also in the sample eluted from bentonite with NaCl) could be due to the fact that the 2 M NaCl solution removed only some of the adsorbed protein components from bentonite.
It is well known that some types of chitinases can bind to chitin (Graham and Sticklen 1994). Since the Mr of at least some of the protein bands eluted by NaCl from the chitin sample incubated in wine was similar to that of the chitinases that were previously shown to maintain their catalytic activity in wine (Vincenzi and Curioni 2005), chitinolytic activity detection on SDS-PAGE gels containing 0.01% glycol chitin (Trudel and Asselin 1989) was performed. As recently reported, the presence of this substrate in SDS-PAGE gels affects the mobility of some chitinolytic enzymes of grape origin (Vincenzi and Curioni 2005); therefore, the zymographic results had to be compared to those obtained by Coomassie staining of a SDS-PAGE gel containing the same quantity of glycol chitin, in which the ≈31 kDa wine band shifted to a position corresponding to a Mr of ≈35 kDa. After staining for chitinolytic activity detection, the solution eluted from chitin by NaCl showed the presence of the main chitinolytic band present in the original wine (Figure 7⇓), confirming that chitinases maintain their activity in wine (Manteau et al. 2003, Vincenzi and Curioni 2005) and demonstrating that the main band removed by chitin from wine is actually a chitinolytic enzyme. This enzyme had the same Mr of ≈31 kDa of the main chitinase isoform CHV 5 isolated from grape berries by Derckel et al. (1998), which corresponds to the class IV chitinase isoform cloned and sequenced by Robinson et al. (1997) and purified by Waters et al. (1996). Moreover, other minor chitinolytic bands of the original wine were also detectable in the samples eluted from chitin by NaCl, one with a Mr of ≈100 kDa and others with Mr ranging from 22 to 30 kDa (Figure 7⇓). Although the two bands of ≈20 and 9 kDa present in the Coomassie-stained pattern of the sample eluted from chitin (Figure 6⇑) did not show any chitinolytic activity (Figure 7⇓), the overall reported results suggest that chitin is able to bind rather specifically the chitinases of wine. In contrast, for the sample eluted by the NaCl solution from bentonite, a comparative examination of the patterns stained for total protein (Figure 6⇑) and for chitinase activity (Figure 7⇓) did not indicate any specific effect of bentonite in removing these enzymes from wine.
Development of a continuous chitin treatment.
The relatively rigid physical state of chitin made it possible to pack a column with one of the chitin samples (chitin A) used in this study. After washing (described in Materials and Methods), the column was left to dry, and the wine, in the same ratio with chitin as that giving the maximum stabilizing effect, was passed through and recirculated until its absorbance was stable (Figure 8a⇓). The column was regenerated by washing as before. During washing, the NaCl solution eluted a major peak (Figure 8a⇓, peak 3), which contained the proteins retained by the chitin during its contact with the wine, whereas a minor peak (Figure 8a⇓, peak 4) was released by methanol. This minor peak probably corresponded to the small amount of polyphenols retained by chitin. A second volume of wine was then recirculated and the washing procedure repeated (Figure 8b⇓). The absorbance profiles of the two successive trials were identical, indicating that the chitin column could be used more than once with the same performance. The two samples of wine successively passed through the chitin column were then heat-tested for protein stability. For the first and second samples, the haze induced by the heat test was 42% and 45% of that showed by the original wine, respectively.
Discussion
As demonstrated previously, a protein with an apparent molecular weight of ≈32 kDa was shown to dominate in the wine fractions most susceptible to heat-induced haze (Waters et al. 1991, 1992). The same authors later demonstrated that this wine protein corresponded to a chitinase of grape origin (Waters et al. 1996), which is now identified as one of the main ones responsible for protein haze formation in white wines. This component is one of the most abundant proteins found in wine, representing about 40% of the total protein content of the grape berry at maturity (Derckel et al. 1998). The same protein, which was also labeled CHV 5 (Derckel et al. 1996), has been identified as being a class IV chitinase (Derckel et al. 1998, Robinson et al. 1997). Moreover, close sequence similarities to this main chitinase isoform have been found for other chitinases of the free-run juice (Waters et al. 1998). Plant chitinases belonging to class IV are characterized by having an N-terminal cysteine-rich domain, different from the catalytic site, which is involved in the binding of the protein to chitin (Graham and Sticklen 1994). Therefore, we have foreseen the possibility to exploit the chitin-binding capability of the main chitinase isoform to specifically remove it from wine in order to reduce its protein instability. To obtain basic information on chitin action in wine, this exploratory work was conduced on a laboratory scale in controlled conditions, which do not necessarily correspond to those of standard winemaking practices. By treating a white wine sample with chitin we obtained a good reduction of the haze caused by the heat test. Although further experiments are needed to assess whether this partial stabilization is sufficient to ensure stability of white wines during real storage conditions, from the preliminary laboratory results reported here it seems that the use of chitin should be considered a potential alternative to bentonite fining for white wines stabilization.
The mechanism involved in the stabilizing effect of chitin is not precisely known. Although also removing a certain quantity of polyphenols, which are claimed to participate in haze formation in white wines (Ferreira et al. 2002), the results reported here suggest that the partial removal of specific protein components is the main reason for the diminished heat-induced haze of the chitin-treated wine. These components should be among those most responsible for haze formation, because the efficiency of the chitin treatment (as calculated from the ratio between the percent haze reduction and the percent reduction in protein content) is about double that obtainable with bentonite, which seems to remove proteins from wine rather aspecifically. In other words, by using chitin, the reduction in heat-induced haze seems to depend on the removal of specific protein components. Although some physical protein adsorption on the chitin surface cannot be excluded, the most likely mechanism of interaction between chitin and wine proteins is a specific binding of these latter through their chitin-binding domain. This idea is supported by the fact that treatment with N-acetylglucosamine, the structural unit of chitin, of a chitin sample previously incubated in wine was able to release the protein adsorbed on the polymer as could be done by washing with a salt solution. Moreover, the analysis of this protein fraction by SDS-PAGE followed by chitinolytic activity detection demonstrated that it was mainly constituted of chitinolytic enzymes and, in particular, by the isoform with a SDS-PAGE mobility similar, if not identical, to that of the chitinase which has been shown to play a key role in the phenomena of haze formation in white wines (Waters et al. 1996). Taking into account that this chitinase isoform should be a class IV chitinase characterized by having a chitin-binding domain (Derckel et al. 1998), a specific interaction with chitin can be suggested. This conclusion is further supported by the fact that this chitinase isoform is retarded in SDS-PAGE gels containing glycol chitin, indicating a specific interaction with the substrate during the electrophoretic migration (Vincenzi and Curioni 2005).
The other chitinase isoform described by Waters et al. (1996) as having a Mr of ≈28 kDa was absent in the protein fraction that could be eluted from chitin. As already reported (Robinson et al. 1997, Waters et al 1998), the ≈28 kDa isoform may be derived from a proteolytic cleavage of the main 32 kDa chitinase, the predicted site of this cleavage being between the chitin-binding domain and the catalytic region. Moreover a lack of chitin-binding activity of a ≈29 kDa chitinase was also shown by Salzman et al. (1998). The absence of the chitin-binding domain in the ≈28 kDa isoform would eliminate its ability to bind chitin and therefore could explain the absence of this isoform in the fraction eluted from the polymer previously incubated in wine. However, the activity band corresponding to this protein is only barely detectable in the untreated wine (Figure 7⇑), indicating that, at least in the case of the wine sample used here, this chitinase isoform is present in very low amounts. This is confirmed by analyzing the gels stained with Coomassie Brilliant Blue, in which a ≈28 kDa band cannot be detected (Figure 6⇑). The degree of processing of the grape berry chitinases, including the proteolytic generation of the ≈28 kDa fragment, may be affected by several factors, such as the plant metabolism in the field, the crushing of the berries, and the fermentation process (Waters et al. 1998).
Other proteins were bound by chitin in wine, some showing chitinolytic activity and others not in the conditions used for its detection (Figures 6⇑ and 7⇑). Several chitinase isoforms have been shown to be constitutively expressed or induced by stress in grape berries (Derckel et al. 1996), most showing sequence similarities (Waters et al. 1998) and similar molecular weights as determined by both SDS-PAGE analysis and mass spectrometry (Pocock et al. 2000). Given these similarities, it is then likely that some of them pass into the wine, where they maintain their activity, as occurs for the main chitinase isoform (Vincenzi and Curioni 2005). The results reported here confirm the stability of the chitinases under the conditions of winemaking (low pH, protease activity, presence of alcohol and sulfite) not only as protein molecules (Waters et al. 1992) but also as active enzymes (Manteau et al. 2003, Vincenzi and Curioni 2005). However, the SDS-PAGE mobility of the minor chitinolytic bands, which could be eluted from chitin at least at the highest doses used (Figure 7⇑), does not correspond to that found for the chitinase isoforms previously described. Whereas the bands with SDS-PAGE mobility higher than that of the main isoform may be products of a limited proteolytic processing of the main isoforms (Manteau et al. 2003, Waters et al. 1998), the chitinolytic band appearing at the top of the gel (Figure 7⇑) should correspond to an isoenzyme with a different origin.
At least other two wine proteins not showing chitinolytic activity were present in the fraction eluted from the chitin sample used for treating the wine (Figures 6⇑ and 7⇑), including a low molecular weight band of ≈9 kDa and one with a mobility corresponding to that of the major band found in the original wine (Mr of ≈20 kDa). One possibility is that the enzymatic activity was lost or not detectable for these components. Another is that these proteins, even though bearing a chitin-binding domain, do not possess any chitinolytic activity in native conditions. Indeed several chitin-binding proteins not showing chitinolytic activity, such as some types of lectins (Beitema 1994), have been described in plants. Of these components, the ≈9 kDa band showed a high affinity for chitin, as well as for bentonite, because, unlike all the other wine proteins, it became undetectable in the wine treated with the lowest doses of both the adsorbents (Figure 5⇑). In this case more data are necessary to identify this low molecular weight protein and to speculate on the nature of its interaction with chitin.
An aspecific mechanism of interaction would instead explain the presence in the fraction eluted from chitin of the relatively low quantity of the ≈20 kDa protein (Figure 6⇑). This band showed a shift of mobility to ≈24 kDa when both the sample eluted from chitin and the total protein of the original wine were reduced before the electrophoretic separation (Vincenzi and Curioni 2005) (not shown). Therefore the characteristics of this wine protein (Mr, relative quantity in wine) correspond to those of the major thaumatin-like protein (Pocock et al. 2000), which is also involved in haze formation (Waters at al. 1996). Since there is no evidence that thaumatin-like proteins have chitin-binding properties and because of the very high relative quantity of this band in the original wine, a non-specific adsorption on chitin should be hypothesized to explain the presence of this protein in the fraction eluted from chitin. That only a minimal part of this component is removed by chitin might explain, at least in part, the residual heat-induced haze detectable in the chitin-treated wine samples.
Compared to bentonite, which at doses as low as 0.5 g/L could completely eliminate the heat-induced haze, the quantities of chitin needed to reach wine (partial) stabilization were much higher (Figure 1⇑). However, this problem is certainly moderated by the possibility of regenerating the chitin sample by simply washing it in succession with salt and methanol solutions as demonstrated here, which indicates the possibility of reusing the same adsorbent material for various cycles of wine treatment with low waste output. Moreover, as shown here on the laboratory scale, the possibility of using chitin for wine treatment in a column system would open the way to developing a continuous technique for white wine stabilization, thus allowing winemakers to greatly reduce waste products and to avoid the wine losses typical of bentonite treatments.
Conclusion
The specific effects of chitin at doses that reduce the tendency to form heat-induced haze suggests that this natural, low-cost polymer has the potential of stabilizing white wines without greatly modifying the wine protein pattern and content, as occurs in the case of bentonite fining. Because proteins can contribute, directly and indirectly, to the organoleptic properties of wines, the use of chitin for the stabilization of white wines can help in maintaining their sensorial quality. The effects on organoleptic properties and on long-term stability of white wines treated with chitin need to be established in actual wine-making conditions.
Footnotes
Acknowledgments: This work is part of the research activity of the Dottorato in Viticoltura, Enologia e Marketing delle Imprese vitivinicole, supported by Provincia di Treviso. The research was supported by a grant from Università di Padova (Progetto di Ateneo CPDA023747). The authors are grateful to Marzio Pol for providing the wine samples.
- Received October 2004.
- Revision received January 2005.
- Revision received March 2005.
- Copyright © 2005 by the American Society for Enology and Viticulture