White Wine Protein Instability: Origin, Preventive and Removal Strategies

Abstract

White wine protein instability depends on several factors, where Vitis vinifera pathogenesis-related proteins (PRPs), namely chitinases and thaumatin-like proteins, present an important role. These proteins can be gradually denatured and aggregate during wine storage, developing a light-dispersing haze. At present, the most efficient process for avoiding this wine instability is through the removal of these unstable proteins from the wine before bottling. To remove unstable white wines proteins, the sodium bentonite fining is the most used treatment, however, many alternative techniques such as ultrafiltration, the application of proteolytic enzymes, flash pasteurisation, other adsorbents (silica gel, hydroxyapatite and alumina), zirconium oxide, natural zeolites, chitin and chitosan, carrageenan and the application of mannoproteins have been studied. This chapter overviews the factors that influenced the white wine protein instability and explored alternative treatments to bentonite to remove white wine unstable proteins.

1. Introduction

Even wine contains a low level of proteins and glycoproteins, which usually ranged from 15 to 230 mg/L, proteins present an important role from a technological point of view [1]. Indeed, proteins greatly affect wine quality by contributing to its sensory and foam characteristics [2, 3, 4]. However, specific wine proteins, mainly grape wine unstable proteins, can be gradually denatured and aggregate/precipitate during wine storage, developing a light-dispersing haze, this being the main cause of post-bottling haze development in white wines [5]. Even the formation of protein haze is improbable to affect the olfactory or gustatory wine characteristics, turbid wines are usually rejected by consumers, causing significant economic losses for the wine industry and brand image [6, 7, 8]. Wine proteins are mainly derived from grapes, but they could also be formed by the metabolism of the several microorganisms (yeasts and lactic acid bacteria) present in the vinification process [9]. Most of these proteins are extinct after wine alcoholic fermentation and consequent fining processes. Nevertheless, the so-called pathogen-related (PR) proteins (β-glucanases, chitinases, thaumatin-related proteins) can persist in the final wine as they are resistant to proteolysis and low pH [10]. They are produced by the grapevine for defence against bacterial or fungal infections and in reaction to abiotic stress [11]. These proteins are resistant to the wine acid conditions, heat and proteolysis due to their compact structures [9]. Numerous research works have stated that wine total protein levels could not predict wine protein instability, as individual protein fractions synthesised in grape berries are responsible for haze development [12, 13]. Furthermore, the wine chemical composition such as metal ions, ionic strength, pH, alcohol level, polysaccharides and phenolic concentration could also play an essential role in protein haze development as these parameters might affect protein denaturation [8, 14]. Additionally, Pasquier et al. [15] mentioned that the climatic changes, with the rise in temperatures and the reduction in precipitation throughout the grape maturation phase, lead towards increase in the probability of wine protein instability. In the winemaking industry to avoid this instability, normally the unstable proteins are removed through bentonite fining before wine bottling. However, this fining agent is non-specific, removing other wine compounds besides unstable proteins, which may affect wine sensory qualities [16]. Therefore, in the previous years, several alternative solutions to bentonite fining for this aim have been searched [16, 17].

2. White wine protein instability

2.1 Proteins responsible for white wine protein instability

Several studies were performed concerning wine protein instability. Koch and Sajak [18] verified using electrophoresis that the heat-formed deposits enclosed two types of protein fractions with diverse heat sensitivities. Moretti and Berg [19] after fractionation and analysis of wine proteins concluded that, among grape and wine proteins, that the protein fractions with low isoelectric points and low molecular weights were more sensitive to heat treatment and responsible for wine protein instability. The main proteins related to the white wine protein instability have a low molecular weight (12.6–30 kDa) and isoelectric point (4.1–5.8) and contain glycoproteins [20]. Waters et al. [21] did the separation and the fractionation of wine proteins using a combination of salting out with ammonium sulphate and ultrafiltration, showing that the protein fractions with those characteristics (24 and 32 kDa) were more sensitive to high temperatures, contributing more to the white wine protein instability. The lower-molecular-weight protein fractions appear to be the major responsible for white wine haze, where the protein with 24 kDa produced nearly 50% more haze with identical concentration than protein fractions with 32 kDa [13]. Many works showed that pathogenesis-related proteins (PR) are the principal responsible for wine protein instability [13, 22]. Pathogenesis-related proteins are very important for plant protection and are associated with its disease resistance, growth and adaptation to stressful environments [23]. Vitis vinifera is the most used for winemaking; however, it is very sensitive to pathogens, particularly fungi and oomycetes, such as Botrytis cinerea and Plasmopara viticola, respectively [24]. Pathogenesis-related proteins (PR) are produced by the plant in response to infection by pathogens [25], to control the harm made to the grapevine [26]. In V. vinifera grape varieties, the thaumatin-like (PR-5 type, 24 kDa protein fraction) [13] and chitinases (PR-3 type, 28 kDa protein fraction) [13] are the two main pathogenesis-related (PR) proteins separated from wine, presented a globular structure and at wine pH a positive charge [27]. Other examples of PR proteins existing in lesser quantity in wine are osmotins, β-1,3-glucanases, invertases, lipid transfer proteins [28]; however, diverse isoforms of thaumatin-like proteins and chitinases have been recognised in grape musts of several V. vinifera grape varieties, with a molecular weight between 20 and 30 kDa, and an isoelectric point between 3.0 and 5.0 [7, 29]. These are the principal soluble proteins from V. vinifera [22, 30] and are responsible for haze development in bottled white wine during storage and transportation [27, 31]. These proteins are synthesised during grape development in function of the grape variety [32], region and year [24, 28] presenting higher levels in the ripening; that means that riper grapes are the more susceptive to protein instability [30]. Chitinases and thaumatin-like proteins present a significant amount of disulphide bonds that contribute to their chemical stable structures and some resistance during the vinification process (some resistance at the low grape juice pH (3.0–3.8)) and resistance to proteolysis [5, 33]. However, the low-molecular-weight proteins (chitinases) are sensitive to temperature changes [22] and wine pH [34]. Thaumatin-like proteins are more thermostable and insensible to wine pH variations, showing no significant structural variations or aggregation in different wine pH [34]. The different sensibility of chitinases and thaumatin-like proteins appears to be associated with the differences in the secondary structure of both proteins, elliptical for chitinases and globular for thaumatin-like proteins [34, 35]. The pathogenesis-related proteins are present in different concentrations in the grape juice of the diverse grape varieties of Sultana, Sauvignon Blanc, Pinot Noir, Muscat of Alexandria and Shiraz with chitinases/thaumatin-like proteins of 118/119, 76/119, 44/23, 21/35 and 9/18, respectively [30]. These researchers likewise verified that grape berry destruction throughout mechanical grape picking, related to long-distance transportation, could encourage the production of pathogenesis-related proteins by the grape defence mechanism before grape pressing [36]. In fact, chitinases and thaumatin-like proteins and its variances in heat stability give the impression that protein composition may influence haze development in wines [9, 37]. Thaumatin-like protein (24 kDa fraction) is the principal responsible for haze formation relative to chitinase (32 kDa fraction) [21, 38]. However, chitinases are very sensitive to precipitation, where a high correlation was verified between wine chitinases levels and wine haze obtained [37]. The thaumatin-like proteins shows a melting temperature of 62°C, with a determined denaturation half-life of 300 years at 25°C, and chitinases present a denaturation half-life of 6 minutes at 55°C, consequently extrapolating down to a denaturation half-life of 3 days at 35°C or 2 years at 25°C. It was observed that vacuolar invertase (GIN1), from the grapes, and β-(1–3)-glucanases (32 kDa fraction) can also influence haze formation. In fact, the existence of V. vinifera thaumatin-like protein bands, β-(1,3)-glucanase and maturation-related protein-like (27.4 kDa) Grip22 precursor have been associated with the natural protein haze of white wines [13, 39, 40]. In fact, there is not a correlation between the total amount of protein and wine protein instability [39].

2.2 Factors that can affect white wine protein stability

The denaturation of some white wine proteins could result in aggregation and flocculation and sometimes in the development of deposits [25], other wine non-proteinaceous compounds can also be related to the wine protein haze development. Curiously, wines with an identical protein fraction can present different haze tendencies [41], and the wine ethanol level did not influence wine protein instability [42, 43]. The protein-polyphenol interaction is the major studied mechanism associated with white wine protein instability [44, 45]. The existence of procyanidins is necessary to develop wine turbidity, only the presence of wine proteins did not develop wine turbidity [46]. It was shown that the interaction between haze-active polyphenol and haze-active protein and the amount of haze formed is highly dependent on protein and polyphenol concentration and their ratio [47]. The turbidity of a protein-polyphenol complex increased with a pH rise from 2.5 to 3.7 (model wine solution with 10% ethanol) [48]. Some authors consider that protein haze formation is an isoelectric precipitation mechanism [49]. Some authors think that turbidity formed in white wines is related to hydrophobic interactions among proteins and tannins happening on the hydrophobic tannin-binding sites of proteins that can be exposed depending on heating and reduction [37]. Many phenolic compounds were detected in protein haze, such as tyrosol, trans-p-coumaric, vanillic, trans-caffeic, protocatechuic, gallic, syringic, ferulic, shikimic acids, (+)-catechin and ethyl coumaric acid ester; quercetin and cyanidin, after acid hydrolysis, the existence of procyanidins was also shown [39]. Phenolic compounds can increase haze formation by cross-linking denatured proteins provoking aggregate development [9], in fact, the removal of phenolic compounds from wines resulted in reducing haze development [38]. The X factors are factors essential for protein turbidity and are wine conditions such as pH, ionic strength, organic acid concentration [49], polysaccharides [50], metal ions [51] polyphenols/phenolic compounds [25] and sulphate anions. As mentioned before, wine pH is an important factor in protein haze development, with model wines at pH 4.0 inducing higher protein aggregation and turbidity development after heating than model wines of lower pH (pH 3.0) [52]. The application of sulphate anions or sodium cations that increase the wine electrical conductivity and ionic strength increases the tendency of haze formation after heating, by the decrease of the electrostatic repulsion of proteins [37]. In model wines, it was shown that other ions including tartrate, chloride, Fe^2+/3+ and Cu^+/2+, do not influence the turbidity formation [30]. Higher electrical conductivity (0.134 and 0.163 S/m) and protein levels (9 and 25 mg/L) provoke greater perceptible turbidity; however, the white wine with low iron levels (0.3 and 0.9 mg/L) and protein stability appears to increase so there is a negative correlation between wine turbidity and the iron levels [53]. There is evidence that polysaccharides could potentially decrease wine protein instability by forming a protective layer around unfolded proteins [54]. Organic acids could present interactions with phenolic acids, free amino acids, tannins, pectic compounds and sulphate ions, avoiding in this manner, their interaction with proteins [55]. The same authors verified that organic acids could influence wine protein instability by the electrostatic interactions that depend on the organic acid pKa and protein isoelectric point values and the medium pH. The sulphate ions could be a non-proteinaceous factor for protein instability, as they promote protein-protein hydrophobic interactions Pocock et al. [38], in addition to the suppression of the electrostatic repulsion between proteins by the increase of the ionic strength of the medium [37]. It was demonstrated that potassium hydrogen sulphate can influence haze formation [7]. Some authors suggest a three-stage process in the protein haze formation that included protein unfolding, protein self-aggregation and aggregate cross-linking, highlighting the role of sulphate ions in all stages [9]. Chagas et al. [56] verified the influence of sulphur dioxide, existing in wines in the irreversible denaturation and aggregation phenomena of thaumatin-like proteins and their influence on wine protein instability or turbidity development. The presence of ion bisulphite (HSO₃⁻) results in cleavage of the disulphide bonds of the thaumatin-like proteins, with the formation of S-thiosulfanates and free thiol-groups that contribute to the temperature-induced protein unfolding. The hydrophobic surfaces and the presence of free thiol-groups result in protein aggregation by formation of inter-protein disulphide bonds in thaumatin-like proteins, following a nucleation-growth kinetic mode.

3. Preventive treatments and strategies to mitigate white wine unstable proteins

3.1 Effect of growing and harvest conditions on wine protein composition

By using principal component analysis and clustering techniques, Sarmento et al. [32] pointed out that the most important factor affecting wine protein profile was the grape variety, and the growing region, whereas vinification practice (industrial and laboratory scale) on the same varietal wine did not show a major effect. In grapevines, the synthesis of the PR proteins is regulated in a developmental and tissue-specific manner and occurs predominantly in the skins of the grapes [36, 57]. In V. vinifera cv. Muscat Gordo Blanco, both the concentration of the corresponding main thaumatin-like proteins and the berry-specific expression of the VvTL1 gene improved intensely after véraison and continued during grape maturation [58]. Identical developmental patterns were also found in the expression of genes encoding chitinases, some identical to those involved in wine protein haze [59, 60, 61]. Immunological research of V. labruscana cv. Concord likewise demonstrated that thaumatin-like proteins and chitinases accumulate during berry maturation [62]. PR proteins also exist in several other fruits such as banana [63], cherry [64] and kiwi fruit [58]. In all V. vinifera cultivars studied, thaumatin-like proteins and chitinases are the main soluble protein of grape berries [36, 58]. The prevalence of these PR proteins was evident at all phases of the grape berry growth next véraison [30]. Significantly, as the levels of extractable proteins in the grape berries continually rise during maturation, it can be supposed that the haze-forming potential growth as maturation continues [30, 58]. Pocock et al. [30] also showed that the increase of thaumatin-like proteins and chitinases initiated at berry softening for Muscat of Alexandria, Sultana, Shiraz grape varieties, Sauvignon Blanc and Pinot Noir grape varieties. As in healthy grape berries, PR protein synthesis seems to be caused by véraison, this does not signify that the traditional PR protein inducers, wounding, stress and pathogenic attack, cannot additionally modulate the grape berries’ PR proteins concentration. These grape proteins in vitro display antifungal activity to Botrytis cinerea, Uncinula necator, Phomopsis viticola, Elsinoe ampelina and Trichoderma harzianum general fungal pathogens of grapevines [35, 62, 65, 66, 67]. The antifungal activity shown in vitro replicates the major function of the PR proteins in vivo, their expression in grapes afterward veraison represents a defence mechanism for grapes. Jayasankar et al. [66] give additional credibility to this hypothesis by indicating that after in vitro selection, grapevines regenerate with E. ampelina culture filtrates presented high constitutive expression of PR proteins, comprising VvTL1 and higher disease resistance. Works in which the PR proteins synthesis is changed by gene technology would permit us to explore this hypothesis more. Currently, there are slight chances that the wine turbidity problems could be resolved by decreasing the PR protein expression in grape berries as this could lead to the grapevine disease. In leaves and grape berries from infected grapevines with pathogens, improved expression of some PR genes and higher levels of some PR proteins have been shown [68, 69, 70]. In greenhouse experimentations, Monteiro et al. [67] showed in infected grape berries with U. necator augmented concentration of thaumatin-like proteins than in uninfected grape. Jacobs et al. [68] observed that in response to powdery mildew infection β-1,3-glucanase activity and chitinases augmented in leaves and grape berries, and that genes expression (VvGlub, VvChi3 and VvTL2), for coding PR proteins, was powerfully induced. Only VvTL2 of the three putative gene products has been found as a soluble protein in grape must and wines [13]. In Chardonnay V. vinifera cv. grape bunches, Girbau et al. [71] showed that occasioned powdery mildew infection augmented the concentration of a grape berry lesser thaumatin-like proteins, VvTL2, in wine. In infections with higher intensities (>30% of infected bunches), the wine turbidity values measured after a heat test were significantly higher. Marchal et al. [72] showed that grape must from infected grape berries by B. cinerea presented lower protein concentration, in opposing to expectations that fungal diseases would lead to higher concentration of PR proteins in grape, and suggested that proteolytic enzymes from B. cinerea were responsible for this. In culture media and on fruits such as apple, secretion of proteases by B. cinerea has been observed [73] and in tomato [74]. Girbau et al. [71] also studied the influence of B. cinerea infected grapes on the vineyard and observed that infection resulted in noticeable reductions in the concentration of PR proteins in the grape berries. Similar although fewer tendencies of decreases in protein concentration were observed in laboratory experimentations in which otherwise healthy grape berries were inoculated with B. cinerea [71]. In this work even though these grapes were not vinified, the variance in protein concentration was predictable that between uninfected and infected grapes would also be shown in wines produced from uninfected and infected grapes. In the grape juice from Botrytis-infected grape berries, the decrease in protein concentration did not appear to be an artefact of reduced extraction into juice due to desiccation or shrivelling of the fruits, nevertheless could be due to proteolytic degradation of grape PR proteins by enzymes of B. cinerea as suggested by Marchal et al. [72]. In grape must, protein concentrations were also decreased when in this medium B. cinerea was grown [71]. If these effects are due to the activity of proteolytic enzymes from B. cinerea, these enzymes have the capacity to substitute bentonite fining for protein stabilisation in oenology, an objective of many research efforts worldwide. The consequences of mechanical grape picking, a harvesting operation that could cause wounding, on the grape berries PR proteins concentration, are therefore of attention. Paetzold et al. [12] showed that grapes picked up by hand, originated grape must with lower protein concentration compared with that of mechanically harvested grapes. The absence of stalk throughout crushing led to lesser polyphenolic concentration in the grape juice compared with the grape juice from grapes picked up by hand, therefore fewer proteins were lost in complexes with phenolic compounds from grape juice from the fruit picked up mechanically. Dubourdieu and Canal-Llaubères [75] showed that wine produced with destalked grapes with maceration during 18 hours presented higher protein concentration than wine produced immediately by pressing of whole bunches. It was not elucidated, if this rise in protein concentration was due to the wounding of grapes that occurs during destalking or maceration or from the elimination of the grape stalks. Pocock and colleagues [36, 76] observed the influence of mechanical harvesting on the PR proteins in grapes and wine. Mechanical harvesting together with long transport of the grape berries leads to greater PR protein concentration in the grape juice and wine. Indeed, white grapes harvested mechanically, following transport was found to double the concentration of bentonite necessary for the avoidance of protein haze when compared with grape berries harvested by hand and transported from the same vineyard [76]. This does not seem to be a consequence of an increase in protein synthesis, as evaluations among hand picked up grapes, mechanical picked up intact grapes, and the major form of mechanical picked up grape berries—a combination of damaged grape berries and grape must—showed that few if any protein was formed as a consequence of stress provoked by mechanical grapes picked up. Protein concentration increase in grape must from mechanical picked up grape berries consequently look to be due to protein extraction from grape skins rather than a physiological wounding answer by the grapes. The influence of water stress established under some viticultural management practices has been studied, on the PR proteins expression in grape berries by determination of the PR protein concentration of V. vinifera cv. Shiraz grape berries in irrigation essays [30]. The absence of irrigation, did not lead to higher PR proteins concentration in the grape, however it provided a clear physiological marks of grapevine water stress. On a fixed quantity of protein per grape, it was observed that in the grape must from water stressed grape the protein content was greater than that from irrigated grape since grapes from irrigated grapevines were greater and thus grapes solutes were fewer concentrated. The water stress influence on the grape dimension is an overall phenomenon [77] and it is probable that reports related to the wine turbidity problems, are higher in drought years and they are due many to a variation in the grape dimensions in these years instead of a direct physiological answer of the grapes to water stress in the formation of PR protein.

3.2 Preventive winemaking practices to avoid white wine protein instability

Regarding the mechanisms of wine protein turbidity development, there are numerous potential approaches for avoiding wine turbidity that would either decrease or remove the requirement for bentonite application. These comprise reducing the concentration of wine phenolic compounds; reducing the wine ionic strength; disrupting hydrophobic protein-protein interactions; stabilising wine proteins against thermal unfolding; degrading wine proteins enzymatically after heat treatment; application of alternative adsorbents or ultrafiltration to eliminate proteins [9].

Enzymes application to degrading haze-forming proteins in wine is a specially an attractive substitute to bentonite since it diminishes aroma removal and wine losses. Preferably, active enzymes would be applied to grape must without the requirement for future removals, as in the case of glucanases and pectinases [78]. The products of grape proteins degradation may also be used by yeast as nitrogen sources, theoretically decreasing the common necessity for nitrogen application and enhancing wine aroma quality [79, 80]. For wine protein degradation, there are two important kinds of enzymatic activities: the decrease of disulphide bonds by protein disulphide reductases and the hydrolysis of peptide bonds by proteases [81]. The difficulty in using proteases for specifically degrading haze-forming proteins in wine is related to the stability of the proteins in wine-like conditions. Protein disulphide reductases could, hypothetically, destabilise and precipitate haze-forming proteins throughout vinification via reduction of disulphide bonds [22]. Nevertheless, under wine conditions, there have been no published cases of active protein disulphide reductases.

In a Champenois Chardonnay wine, it has been shown that a 24/25 kDa protein was an N-glycosylated protein and underwent no modification throughout fermentation [82], whereas degradation or variation of the sugar moieties of the glycoproteins (12–30 kDa) was found to happen during winemaking for a hybrid grape variety (Muscat Bailey A) [83]. The hydrolysis of the sugar chains of grape derived glycoproteins by glycosidase treatment was found to rise turbidity with seed phenols in a model wine [84]. Instead, yeast-derived mannoproteins (420 and 31.8 kDa) could contribute to a stabilisation effect on wine proteins, decreasing haze development [85, 86]. Yeast derived mannoproteins (10–30 kDa) possessing both compositions of the hydrophobic and hydrophilic protein domains and mannose moiety also improved the foaming properties in sparkling wines [87, 88].

Another strategy to decrease the level of proteins in white wines is pre-fermentative skin maceration, for example, in the Albariño grape variety, pre-fermentative skin maceration augmented the concentration of polysaccharides and phenolic compounds extracted, however, reduce the quantity of protein extracted, mainly of the pathogenesis-related proteins, specifically the V. vinifera chitinases and thaumatin-like proteins. While the PRPs and total protein of the Albariño wine produced by pre-fermentative skin maceration were lesser, the wine presented higher protein instability in the heat test, perhaps the presence of higher level of polyphenols compounds [17].

4. White wine unstable proteins removal

4.1 Physical treatments

The removal of unstable white wine proteins could be performed by the use of ultrafiltration [5, 20, 87, 88, 89, 90], flash pasteurisation [91, 92, 93], high hydrostatic pressure [10] and ultrasound [94].

Hsu et al. [20] ultrafiltered a white Gewürztraminer and Riesling wine with Romicon and Millipore systems, worked with membranes of nominal molecular weight cut-offs (MWCO) of 10–100 kDa. According to these authors, protein stability could be achieved with MWCO of 10 and 30 kDa; nevertheless, if the protein stability was not achieved, bentonite required was reduced from 80 to 95%. However, according to Miller et al. [95] and Flores et al. [87, 88, 96], ultrafiltration could also lead to the depletion of wine aroma compounds responsible for the floral, fruity and honey/caramel descriptors (Table 1), changing, in this manner, the wine aromatic profile [97, 98]. Additionally, wines treated by ultrafiltration also showed a significant reduction in yellow colour (420 nm) and total phenols [87, 88], as well as a decrease in the ‘body’ and ‘mouthfeel’ related to the removal of colloids [99]. Furthermore, the high operation and equipment cost associated with the aroma decrease, making this procedure unattractive to the wine industry for eliminating unstable proteins.

	White Riesling			White Gewürztraminer
Descriptors	Control	UF1	UF2	Control	UF
Overall intensity	5.97	5.33	5.23
Fruity	4.83	3.93	4.10	5.22	4.39
Fresh fruit citrus	3.47	2.87	2.43	4.61	3.39
Floral	3.63	1.83	2.00
Vegetative	1.93	2.63	3.13
Cooked vegetative				1.97	2.22
Honey/caramel	2.43	1.63	1.57
Chemical	4.07	3.97	4.33	2.35	3.56

Table 1.

Mean scores* of the significant aroma descriptor ratings for white Riesling and white Gewürztraminer wine (adapted from [97]).

^*

Scored on a nine-point intensity scale (1 = none, to 9 = extreme); UF, ultrafiltration.

Wines heat treatments at medium temperature (45°C, several hours) and high temperature (90°C, 1 minute), with and without the application of proteolytic enzymes, lead to a decrease of the wine protein level and up to 70% of the bentonite needed for heat stability [91]. However, after sensory assessment of the wines submitted to the different treatments the panel members in some wines submitted to heat treatment without enzyme application and to heat treatment with enzyme application (Trenolin blank, 10 mL/L), observed slight effects on wine aroma descriptors [91].

The results obtained by Tabilo-Munizaga et al. [10] established that high-pressure treatments changed the β-sheet and α-helical structures of wine proteins. During 60 days’ storage period, the α-helix structure in high-pressure treatment samples was reduced. Structural modifications by high-pressure treatments (450 MPa for 3 and 5 minutes) increase wine proteins thermal stability and consequently delay the wine haze formation throughout wine storage.

Recently, Celotti et al. [94] developed a research work focused on the application of ultrasound for white wine protein stabilisation. The results showed that higher amplitude (90%) and treatment time (10 minutes) induced an increase in white wine protein stability. This effect is related to the protein charge neutralisation and surface electrical charges, intending positive conformational modifications in the wine proteins. This technique could be considered as a way to prevent wine protein precipitation and to decrease the amount of bentonite fining agents used in wineries.

4.2 Enzymatic treatments

Proteases hydrolyse the peptide linkages between the amino acid units of proteins. Protease activity exists in grape berries [100, 101] and yeast [101, 102, 103, 104, 105, 106, 107, 108] as described by several authors. One important aspect is their potential role in wine protein haze reduction [90, 109]; however, proteases have low activity concerning haze-forming proteins, which consequently persist during the vinification process. It is essential, that proteases have to be active under specific wine conditions, namely acid pH, the existence of ethanol, sulphites, phenolics and if possible act at low temperatures. One more challenge is the resistance of PR proteins against proteolysis due to their molecular features such as disulphide bonds and glycosylations. However, proteases from plants (papain from papaya, bromelain from pineapple) have been tested with some promising results concerning their effectiveness in the degradation of heat-unstable proteins from white wine [110, 111, 112, 113]. However, the search for fungal enzymes that could degrade wine proteins has so far remained ineffective [114]. As unfolded proteins are more easily cleaved by enzymes, the subsequent phase was the evaluation of the mutual effects of protease addition and heat treatment. Heat treatment joint with the application of proteolytic enzyme can decrease the formation of white wine protein instability; however, the low specificity of commercially disposable proteases for the haze-forming proteins seems to decrease significantly the possibilities of offering this strategy as shown by Pocock et al. [91]. A fungal acid protease resulting from Aspergillus sp. rich in aspergillopepsin I (EC 3.4.23.18) and aspergillopepsin II (or aspergilloglutamic peptidase, EC 3.4.23.19) in association with flash pasteurisation (75°C) of the grape juice was confirmed to eliminate haze-forming proteins and consequently stabilises the wines [92, 93]. The application of aspergillopepsin I to eliminate haze-forming proteins in grape must and wine is already authorised by the International Organisation of Vine and Wine [115] Resolution OIV-OENO 541A-2021 and Resolution OIV-OENO 541B-2021. Aspergillopepsin is active at juice and wine pH and at a temperature greater than the melt temperature of haze-forming proteins (chitinases and TLPs, 56 and 62°C, respectively). Therefore, after application of aspergillopepsin I, one short-term heating (60 and 75°C; 1 minute) must be performed as it contributes to the unfolding of haze-forming proteins and facilitates their enzymatic degradation by proteases, as well as leads to denaturation of the protease itself [115], Resolution OIV-OENO 541A-2021; Resolution OIV-OENO 541B-2021. In this context, a protease of Botrytis cinerea BcAp8 has been described to hydrolyse grape chitinases at moderate temperatures [116]. Also, evaluation of the effects of the joint use of heat treatment (75°C, 2 minutes) and application of proteases on the protein stability was recently studied by Comuzzo et al. [117]. These authors also evaluated the effect of the heat treatment with application of protease on the wine volatile composition and observed that the wines submitted to this treatment presented a lower content of esters produced during alcoholic fermentation and a higher concentration of esters that are characteristic of ageing such as ethyl lactate [117]. The potential of ultrafiltration (UF), in association with heat and proteolytic enzymes, to eliminate haze-forming proteins and stabilise white wine was evaluated by Sui et al. [90]. Since the treatment with enzymes (proteases) to eliminate wine haze-forming proteins needs a previous thermal treatment to denaturant them, recently the application of ultra-high-pressure homogenisation (UHPH) was suggested as a possible alternative to the heat treatment. In this way, the application of UHPH could be in the future a new technological solution for using enzymes in the wine protein stabilisation process and probably with a lower impact on the wine volatile composition [118].

4.3 Fining and adsorption treatments

These practices include the use of adsorbents [117], such as zirconium dioxide (ZrO₂) also known as zirconia [37, 119, 120, 121] carrageenan [6, 92, 122], silica gel, hydroxyapatite and alumina [42], magnetic nanoparticles [123] zeolites [124, 125] and dicarboxymethyl cellulose [126]. However, all of them are at the moment under investigation and therefore not allowed by the International Organisation of Vine and Wine (OIV) or by the European Union (EU) legislation for application in wine.

Mannoproteins [127] are already allowed to be used by the OIV [115]. Chitin and chitosan [127, 128] have been authorised by the European Union (EU) for removal of contaminant and heavy metals, avoidance of turbidity and decrease of unwanted Brettanomyces spp. population (EU) 53/2011), but only chitin (Oeno 367-2009 Chitin-Glucan [115] and chitosan (Oeno 368-2009 Chitosan [115] from the cell walls of Aspergillus niger or Agaricus bisporus are allowed to be applied in wine.

In recent times, some researchers also studied the application of nanomaterials to remove unstable wine proteins [129]. Magnetic steel nanoparticles coated with acrylic acid have been experimented for the selective removal of pathogenesis-related proteins from wines by cation exchange mechanism due to the existence of carboxylic acid groups in the modified surface, and the results showed that they are highly efficient in decreasing haze-forming proteins [122, 130, 131]. Although these nanoparticles have been found to be effective in removing proteins in protein-unstable wines, their efficiency in wines seems to be affected by the low pH of wines that affects the cation exchange capacity of the nanoparticles due to the protonation of the carboxylic acid groups. Also, mesoporous nanomaterials proved to have high efficiency in decreasing haze-forming proteins with lesser wine aroma decrease compared with bentonite fining [132].

Wine-unstable proteins could also be adsorbed by zirconium dioxide [4, 119, 120, 133], a metal oxide usually known as zirconia, and consequently stabilise the wine by removing, especially, wine protein fractions between 20 and 30 kDa. Also, zirconium oxide pellets enclosed into metallic cage submerged in wine at 25 g/L for 72 hours stabilised white wines by removing unstable proteins with the advantage to be regenerated [37].

Results show that the water-insoluble dicarboxymethyl cellulose successfully reduced the wine protein content and turbidity, producing heat-stable wines with concentrations higher than 0.25 g/L [126].

Polysaccharides extracted from seaweeds were also studied by several researchers due to their negative charge at low pH, can electrostatically flocculate and precipitate positively charged proteins and remove wine unstable proteins [6, 122, 134]. Carrageenan uses at different winemaking stages were considered, and the application stage showed to be very important for its effectiveness [6, 92] More recently, Arenas et al. [17] showed that k-carrageenan reduced the content of pathogen-related proteins and consequently the wines protein instability, being even more efficient than sodium and calcium bentonites (Figure 1). On the other hand, these authors also showed that chitosan from fungal origin was unable to heat stabilise the wines, and it was also observed that after the application of this oenological product, the levels of pathogen-related proteins remained unchanged. Additionally, the application of the fungal chitosan decreased the concentration of wine polysaccharides by 60%, as also observed after the application of sodium and calcium bentonite (16–59%). However, the application of k-carrageenan did not change the concentration of wine polysaccharides.

Figure 1.

Reversed-phase HPLC results and percentage reduction of turbidity (NTU), total protein (mg/L), Vitis vinifera thaumatins (VVL, mg/L) and chitinase (mg/L) for Albariño white wine produced without pre-fermentative skin maceration and the impact of the different products applied for its protein stabilisation. Control wine without any additive; after addition of k-carrageenan (100 g/hL); after addition of sodium bentonite (120 g/hL); after addition of calcium bentonite (120 g/hL). All chromatograms were acquired by analysis of a 5 mg/mL solution of the high molecular weight after elimination of the low-molecular-weight material by application of 6 M urea and repeated ultrafiltration through a 10 kDa cut-off membrane (adapted from Arenas et al. [17]).

Chitin [135] and chitosan [128], polysaccharides mainly from Aspergillus niger, also have the capacity to decrease wine haze-forming proteins. It was observed that wine haze induced by the heat test is reduced by 50% after the addition of 1 g/L of chitin, while the addition of 20 g/L of chitin decreased the haze by 80%. The haze decrease perceived was related to the removal of the class IV grape chitinases [136]. Colangelo et al. [128] also showed that wines fined with 1 g/L of fungal chitosan-glucan enhanced heat stability at 55−62°C, and this was also due to the reduction of chitinases.

Mannoproteins existing in yeast cell walls have also been reported to have a protective effect on wine protein haze development [137, 138]. Waters et al. [54] showed that mannoproteins protect unstable wine proteins, avoiding wine turbidity when wine is exposed to high temperatures; these authors indicated that this action does not avoid the protein precipitation. Instead, they detected a reduction in particle size, justifying, in this way, the wine stabilisation observed when determined by turbidimetry. However, their effectiveness for protein stabilisation is highly dependent on the mannoprotein structural characteristic, according to Ribeiro et al. [137], the effectiveness of commercial mannoproteins was related to their chemical composition, namely their high mannose-to-glucose ratio.

5. Final remarks

White wine protein instability has still been an important problem in the wine industry by the frequency of haze formation on the white and rose bottled wine. The grape variety and its grape sanitary conditions, ‘terroir’, the climate conditions during the grape maturation, the mechanical harvest and some winemaking operations could influence significantly the levels of unstable proteins in the wines. The principal proteins responsible for the protein haze are chitinases and thaumatin-like proteins, considered pathogenesis-related proteins (PR) with different thermostability and sizes. Many factors could affect their wine stability, such as wine exposition to high temperatures, wine pH variation, organic acids levels, metals composition, sulphur dioxide levels and the presence of phenolic composition and its degree of polymerisation. Some factors are yet unknown (X factors) but they influence protein precipitation. Even after many works that have been done in the last years, sodium bentonite has still been the most effective treatment to eliminate unstable proteins from white and rose wines. In fact, many products and treatments had been tested to remove these unstable proteins, such as proteases, different polysaccharides (chitin, chitosan, CMC, carrageenan), yeast mannoproteins, some of them show an interesting efficiency, such as carrageenan in a recent work. Finally, white wine proteins stabilisation has still been a problem for the wine industry, and it is necessary to continue developing new approaches to remove or mitigate this important problem. It is necessary to get new solutions to decrease the amount of bentonite used in the wine industry per year by those negative sensory impacts after wine treatment and by environmental concerns.

Acknowledgments

We appreciate the financial support provided to the Research Unit in Vila Real CQ-VR Chemistry Research Centre—Vila Real (UIDB/00616/2020 and UIDP/00616/2020) by Fundação para a Ciência Tecnologia (FCT-Portugal) and COMPETE and by the European Regional Development Fund through NORTE 2020 (Programa Operacional Regional do Norte 2014/2020) to the project AgriFood XXI (NORTE-01-0145-FEDER-000041).

Conflict of interest

The authors declare no conflict of interest.