The Evolution of Phenolic Compounds in Vitis vinifera L. Red Berries during Ripening: Analysis and Role on Wine Sensory—A Review

Abstract

The study of phenolic maturity in Vitis vinifera L. requires a multidisciplinary approach to understand how the evolution of berry flavonoids and cell wall material influence the colour and the textures of red wine. This is a challenging issue which involves researchers of viticulture and enology, and the results of their work are of particular interest for the producers of high-quality red wines. This review reports the current knowledge regarding phenolic maturity, describing the sensorial traits of the different compounds, the evolution of berry flavonoids and the methodologies used to evaluate their characteristics. Finally, the role of cell wall material in influencing the extractability of anthocyanins and proanthocyanidins was shown. By means of a critical review of the results, it can be hypothesised that prolonged ripening improved colour characteristics and mouthfeel properties, thanks to the higher amounts of extractable skin flavonoids associated with lower amounts of seed proanthocyanidins, and to the increased affinity of the cell wall material for the proanthocyanidins most involved in the perception of unpleasant astringency.

1. Introduction

It is well known that cultivar, meso and micro-climate, nutritional status, water availability and cultural practices have a significant effect on ripening and on the final chemical composition of Vitis vinifera L. berries. Moreover, maturity range can be affected by the desired wine style. The evolution of soluble solid concentration, pH and titratable acidity are measured during ripening to predict the main technological parameters of wine, such as alcohol content, pH and total acidity, and represent a basic grape ripening assessment; however, information regarding skin and seed phenolic compounds are key determinants for the production of high-quality red wines.

Considering the role that phenolic maturity of red-berry grapevine cultivars plays on grape quality, the aim of the current work was to review the current knowledge regarding this issue and to describe the changes that flavonoids undergo during ripening, and their effects on wine colour and textures.

1.1. Classification of Berry Flavonoids

The most representative classes of berry flavonoids are constituted by anthocyanins and flavanols. Anthocyanins accumulate in the skin of red berries starting from veraison; they are present as glucoside derivatives of delphinidin, cyanidin, petunidin, peonidin and malvidin [1]. Catechin, epicatechin, epicatechin-gallate and epigallocatechin are flavanols present from fruit set in berry skins and seeds as free monomers and polymeric forms, which are called proanthocyanidins (PAs) or tannins [2]. Skin PAs have a higher mean degree of polymerisation (mDP) and a lower proportion of galloylated forms as compared to those of seeds, while epigallocatechin is not found in seed flavanols. Thus, the epigallocatechin and epicatechin-gallate found in wine may indicate the berry tissue, skin or seed from which the PAs are extracted [3].

1.2. Role of Anthocyanins and Flavanols in Determining Wine Quality

The importance of each class of berry flavonoids to wine quality has been known for a long time. Anthocyanins are pigments responsible for red wine colour, and their contribution could be direct, due to their spectral properties, or indirect after copigmentation reaction with other phenolic compounds, such as flavonols which are present in minor quantities with respect to other flavonoids [4]. During wine ageing, the formation of pigmented polymers, resulting from the reaction between anthocyanins and PAs, modulates wine colour which shifts to a red-brown tint [5]. Colour has been widely recognised as a good predictor of red wine quality; Jackson and co-workers [6] and Parpinello and co-workers [7] have shown that higher overall quality, evaluated after a complete sensory assessment, was associated with higher colour intensity, for Beaujolais and Novello wines, respectively; similar results were found in a study conducted on young red wines from South Australia [8].

Anthocyanins, falavan-3-ol monomers and PAs contribute to the perception of important mouthfeel attributes (i.e., astringency and bitterness) and to the overall quality of the wine [9,10]. In this context, Mercurio and co-workers [11] reported that Australian Cabernet Sauvignon and Shiraz wines with a higher concentration of total phenolic compounds were allocated to grades with higher market value and, in the companion paper, the increase in projected bottle price in relation to the increase in colour density, tannin concentration and tannin mDP was observed [12]. Moreover, a study conducted on premium red wines from different Spanish Denominations of Origin, described a positive relationship between the overall quality, and the concentrations of the malvidin-catechin dimer and of the PAs linked to polysaccharides [13].

1.3. Relationship between Flavonoid Composition and Mouthfeel Attributes of Wine

Although the relationship between the concentration of flavonoids and wine quality has been proved, it is important to note that the composition of these compounds could also alter the perception of the mouthfeel attributes. Astringency, in particular, can define the quality, complexity and persistence of a red wine better than other sensorial traits [14], and it is well known that a balanced level of astringency is a key component of high-quality red wines [15]. Gawel and co-workers [16] have described this very complex oral sensation with thirty-three different sub-qualities which were grouped into two ‘pleasant’ categories (complex and surface smoothness) and five ‘unpleasant’ categories (drying, harsh, unripe, dynamic, particulate). Although the majority of the astringency sensations can be considered undesirable, a study conducted on different red Italian wines has indicated that ‘unpleasant’ astringency, if well balanced, could also enhance wine quality, as found in premium Nebbiolo and Sangiovese wines [17].

Several studies have been conducted to describe the oral sensations of different flavonoids by tasting phenolic fractions or analysing the sensory properties of wines with different phenolic compositions. The results of these studies are herein presented.

1.3.1. Sensory Analyses of Flavonoid Fractions

The sensory analysis of flavan-3-ol monomers, dimers and trimers dissolved in aqueous ethanol (1% v/v) indicated that astringency increased with the degree of polymerisation while bitterness decreased [18]; however, more recent research regarding PA fractions isolated from Italian and Spanish red wines has showed that monomers and dimers contribute slightly to astringency [19] and to bitterness [10]. Vidal and co-workers [20] also found that the polymer length of the PA fractions was positively correlated with astringency and with some ‘unpleasant’ sub-qualities; however, the mDP did not influence bitterness. Moreover, the same author reported that a higher percentage of galloylation elicited rougher astringent attributes while epigallocatechin units slightly decreased those sensations. The PA fractions of Cabernet franc skin and seed were equally astringent [21], probably due to the equivalent contributions of the high mDP of the skin PAs and the high percentage of galloylation of the seed PAs. Conversely, seed PA extracts of Aglianico grapes were more astringent than those of the skin [22].

Some studies have been conducted involving the addition of flavonoids to acidified water or to commercial wines, and it was noted that epicatechin was more astringent and bitter than catechin [23]. Moreover, it should be noted that the concentration of PAs may prevail in the perception of astringency rather than the structural composition of the above-mentioned compounds [24].

In the literature, there is consensus regarding the role of anthocyanins in influencing mouthfeel attributes. These pigments are per se less astringent than PA; however, when added to wine, they can increase ‘pleasant’ astringency and the perception of fullness [10,21,25].

1.3.2. Sensory Characteristics of Flavonoids in Wine

The results of the studies based on wine sensory analysis appear to be in accord with the many findings observed by the sensory evaluations of phenolic fractions. As reported by Preys and co-workers [26] in a study conducted on 61 French and 60 German red wines, the higher the concentration of PAs, the galloylation rate and the mDP, the higher the astringency. On the contrary, perceived astringency was lowered by a higher concentration of epigallocatechin [26,27]. Considering these results, the higher amount of epigallocatechin found in Carménère wines, as compared to that of Cabernet Sauvignon, may explain the lower astringency of the former, although it was characterised by the higher content of PAs of a higher mDP [28]. In Shiraz wines, the ‘unpleasant’ pucker astringency was positively correlated with the concentration of PAs and pigmented polymers, but negatively correlated with the concentration of anthocyanins [29]. Thus, the concentration of seed-derived PAs and that of total phenolic compounds in Cabernet Sauvignon wines were positively associated with astringency and bitterness [30]. Finally, some studies conducted on Tannat wines, which are usually characterised by high levels of phenolic compounds, have described that wines with similar overall astringency differed significantly in ‘unpleasant’ sub-qualities [31], and that PA concentration was positively correlated to aggressive astringency and negatively to soft textures [32].

Considering the results described in this section, it would appear that high-quality red wines are marked by high colour intensity and balanced astringency, which should be mainly characterised by ‘pleasant’ sub-qualities. According to the previously reported studies, these properties may be conferred by anthocyanins and skin Pas, while seed PAs may be considered undesirable from a winemaking perspective [33]. Accordingly, Ristic and co-workers [34] described that the index Anthocyanins x Skin PAs/Seed PAs, calculated with the relative concentrations in grapes, was highly correlated with the “overall quality” of the resulting wines.

The interpretation of these results allowed the authors to speculate that the appropriate berry phenolic maturity for the production of a high-quality red wine might be that condition in which anthocyanins and PAs could easily be extracted from the skin while low quantities of flavanols were released by seeds. This condition is affected by many factors which are subject to changes during ripening, such as the concentration, composition and extractability of flavonoids, and the activity of other berry compounds (i.e., skin and flesh cell wall material) which interact with flavonoids.

In the following sections, the methodologies applied to estimating the phenolic maturity and the results of the studies aimed at evaluating the properties of flavonoids and cell wall material are reported.

2. Methods Used to Determine Phenolic Maturity

Determination of the phenolic maturity implies the extraction of phenolic compounds from berries and the measurement of their concentrations. Phenolic extracts can be obtained from grapes using different solvents and, after their extraction, methods of analysis could include evaluation of the total content spectrophotometrically or separation using high performance liquid chromatography (HPLC). The main methods used for analysing the phenolic maturity are described below.

2.1. Colorimetric Assays

The phenolic content of berries has been determined colourimetrically using Folin–Ciocalteau reagent for a long time [35]. The extraction of polyphenolic compounds is performed on berry homogenate or on ground tissues, using methanol [36] or acidified methanol [37]. The following quantification is based on a standard curve of gallic acid [38]. Colourimetric analysis determines the total amounts of polyphenols giving any information regarding colour characteristics and the different sub-qualities of astringency.

Another colourimetric assay, aimed at quantifying flavanol concentration in berries and wines, was proposed by Deshpande and co-workers [39]: grape extracts were mixed with vanillin 1% methanol and incubated with sulfuric acid (25% v/v methanol) before spectrophotometric measurements. This protocol was modified by Sun and co-workers [40] in order to ameliorate its reproducibility. Similarly, flavanol can be quantified after the reaction with p-dimethylaminocinnamaldehyde (DMCA), as described by McMurrough and McDowell [41]. This procedure was successively optimised for grape extracts and wine by Nagel and Glories [42], and it showed higher specificity than a vanillin assay [43].

2.2. Methods Involving the Precipitation of Flavanols

Specific protocols for the assessment of flavanols are based on the ability of different proteins to precipitate these compounds. Considering that astringency is caused by a reduction in oral lubrification due to the interaction between flavanols and salivary proteins, these methods may give more specific information regarding phenolic maturity.

Hagerman and Butler [44] determined plant-derived PAs after their precipitation with bovine serum albumin (BSA) and reaction with ferric chloride. This procedure was adapted to grapevine berry analysis to evaluate skin and seed contribution separately after the extraction of PAs using a solution containing acetone [45] or aqueous ethanol [46].

Gelatine, similar to human saliva, contains proline-rich proteins and is profitably used to determine the concentration of PAs in wine and its astringency [47,48,49] but is rarely used for skin and seed extracts [50]. Whole human saliva showed interesting results in the characterisation of the astringency of skin and seed extracts [22,51]; however, the heterogeneity of this precipitating agent has limited the use of this technique.

Sarnekis and co-workers [52] proposed a protocol for the quantification of grape PAs by precipitation using the methyl cellulose precipitable (MCP) assay after berry homogenisation and extraction in aqueous ethanol. The results obtained with this method are highly correlated with wine astringency such as those using proteins as precipitation agent and, thanks to its simplicity, it also appears suitable for wineries [53].

2.3. The Glories Method and the Following Modifications

A specific method for assessing berry phenolic maturity is based on the determination of the total amount of flavonoids and of the portion which can be extracted during the winemaking process [54,55]. Phenolic compounds are extracted from berry homogenates in acidified ethanol (pH 1.0) to quantify their total amount, and in a solution at pH 3.2 to evaluate the extractable portion. The concentrations of anthocyanins and flavanols are then usually determined using spectrophotometric analysis [56,57]. This method, easy to apply in the laboratory, enhances knowledge of the characteristics of berry flavonoids thanks to the characterisation of their extractability (i.e., the ratio between the extractable portion and the total amount), and makes it easier to establish the most suitable time for harvest. Nevertheless, since the analysis of flavanols is conducted on berry homogenates, it is not possible to determine the contribution of skin and seed flavanols and, as a consequence, some information regarding the prediction of astringency, due to the origin of the phenolic compounds, may be lost. Moreover, the disruption of skin and seed tissues following homogenisation might mislead the evaluation of anthocyanin extractability which is normally affected by the slow degradation of the skin cell wall.

As reported by Nadal [58], protocols were developed by modifying the Glories method; the composition and the pH of the solutions were changed by Peyron [59] and Mateos [60], respectively, while Lamadon [61] proposed using only one extracting solution containing 15% v/v of ethanol.

2.4. Methods for the Determination of Flavonoid Concentration and Composition

The characterisation of the total amount of berry flavanols requires exhaustive extraction of skin and seed powders in a solution using 70% acetone [2,62] or in a solution using 50% ethanol, even if the latter showed lower extraction efficiency than acetone [63]. High performance liquid chromatography carried out after the acid-catalysed cleavage of the PAs then allows quantification of the constitutive units of the flavanols and calculation of the mDP of the polymeric forms, increasing the accuracy of the information regarding the properties of these compounds. The cleavage of PAs may be carried out in the presence of excess phloroglucinol [64] or with thiolysis reactions [65,66]. It is worth noting that the results obtained with these methods, as with those based on protein precipitation, were highly correlated with the perceived astringency, while the results of colorimetric analysis, being nonspecific assays, showed a very low correlation coefficient [67]. Furthermore, in a comparison between the protocols proposed by Iland et al. [62] and Glories [54], the former showed the best correlation with wine data [68].

Other methods based on gel permeation chromatography (GPC) and size-exclusion chromatography (SEC) may be employed to evaluate the molecular mass of grape and wine flavanols [69,70]; however, they do not give information regarding their composition [71].

Determination of the total amount of anthocyanins can also be carried out using HPLC, and the exhaustive extraction of skin, or skin homogenates, can be carried out in methanol [72,73], acidified methanol [74,75,76], aqueous methanol [77], acidified ethanol [78] or aqueous acidified ethanol [62].

The same HPLC analysis can be carried out on the extractable portion of the flavonoids in order to obtain information regarding the quantity of anthocyanins and flavanols which can be effectively released during vinification. In this case, the skins and the seeds are soaked separately in model hydroalcoholic solutions characterised by an ethanol concentration and pH similar to that of wine for a duration varying from 5 h to 21 days [1,79,80,81]. In addition, some methods have been designed to mimic the increase in ethanol due to the fermentation and the temperature at which this biological process proceeds [82,83,84].

The use of wine-like extractant solutions has been proven to be effective in predicting the concentration of a wine’s flavanols [85]; this approach can also be aimed at evaluating flavonoid extractability during ripening, a characteristic which behaves differently depending on the berry tissue (i.e., skin and seeds).

2.5. Other Methods

Some methods based on texture analysis have showed that the mechanical properties of berries, in particular the skin break force, are highly correlated with phenolic compound concentration and extractability, proving to be powerful tools in the study of phenolic maturity [86,87,88,89].

Moreover, non-destructive analyses have been proposed to estimate the concentration of soluble solids and phenolic compounds, using a fluorescence method [90] or near infrared hyperspectral images [91]. After additional investigation and cultivar-based calibration curves, these procedures could be adopted for the fast and inexpensive monitoring of berry characteristics, which would be useful in establishing the harvest period and the quality of different lots of grapes with the objective of a precision viticulture approach.

Finally, considering that the change in colour of the seed coat was a good indicator of berry flavonoid properties [92], since a darker colour corresponded to higher anthocyanin extractability and lower percentages of seed flavanols [93], advanced methods were developed to assess phenolic maturity by the elaboration of seed images [94,95].

3. Evolution of Phenolic Compounds during Berry Ripening

3.1. Characteristics of the Total Amount of Skin and Seed Flavonoids

Considering the key role that the characteristics of anthocyanins and flavanols play on wine quality, many studies have been conducted to determine the evolution of these compounds during berry ripening in order to predict the optimal harvest time for the desired oenological aim. Evidently, the evolution of the characteristics of these compounds is affected by genotype and several environmental factors; however, it is possible to outline the main features.

Anthocyanins in grape berries are synthesised via the flavonoid pathway. All the structural genes involved in anthocyanin biosynthesis are located on the endoplasmic reticulum membranes or in the cytoplasm in which the anthocyanins are directly produced. Almost all the anthocyanins are then stored in the vacuoles. The accumulation of skin anthocyanins in red grape varieties begins with veraison and reaches its maximum in the latest phases of berry ripening when the synthesis decreases or stops [96]. In general, skin anthocyanins accumulate linearly from veraison to harvest; however, in the case of excessively high temperatures, a decline could occur after an initial increase [78,97]. The trend of accumulation, in which derivatives of cyanidin were more promptly synthesised at the very early stage of veraison, is relatively common in grapevines, even for cultivars whose anthocyanin profile is largely composed of malvidin and tri-substituted anthocyanins at harvest [98]. Glycosylation, methylation and acylation are necessary for anthocyanin stabilisation and transport into the vacuole. Glycosylation is an important modification for increasing the hydrophilicity and stability of anthocyanins which occurs after veraison in grapevines. Cyanidin 3-glucoside and delphinidin 3-glucoside may also be methylated to be converted to peonidin 3-glucoside and petunidin, or malvidin 3-glucoside, respectively. The acylation of anthocyanins leads to the production of 3-O-acetyl-, 3-O-coumaroyl- and 3-O-caffeoyl-monoglucosides by attaching acyl groups to the C6″ position of the glucose moiety [99]. The acylation of anthocyanins occurs from veraison to harvest; however, some cultivars, such as Pinot Noir, do not produce acylated anthocyanins [100].

Flavanols are synthesised in skin and seeds from berry-set to veraison [101]. As a consequence, the concentration of these compounds is highest at the onset of ripening and may subsequently decrease due to berry expansion and the oxidative crosslinking of the polymers [102]. These reactions determine the change in colour of the outer integument of the seed, which turns from yellow into dark-brown [92,103].

The reduction in flavanol concentration after veraison may initially be rapid and can then be followed by a plateau in the last weeks before harvest [84,104] or may not proceed with a specific trend [2,45,105,106,107].

A steady decrease in flavan-3-ol monomers throughout ripening was observed in the seeds of Cabernet Sauvignon, Syrah (also known as Shiraz), Pinot Noir and Sangiovese [84,102,106,108]. Moreover, a reduction in the galloylated forms of seed flavanols until the time of harvest has been reported [109,110,111] and, in some cases, also of their mDP [102,106,108,109]. Considering skin flavanols, Kennedy and co-workers [112,113] observed an increase in epigallocatechin percentage during ripening, as well as an increase in their mDP.

Analysing total skin and seed flavonoids throughout ripening showed them to be good predictors of wine colour characteristics, but they were poorly correlated with wine flavanols and their sensory traits, probably because the evolution of the extractability of these compounds differed between skin and seed [114,115]. As a consequence, as regards the concentration and composition of total skin and seed flavanols, there may be a lack of information useful in assessing the phenolic maturity of grapes [116]. However, the use of hydroalcoholic solutions for the extraction of flavonoids enables the acquiring of more realistic information about the concentration and composition of the anthocyanins and flavanols which could be extracted during the maceration of black grapes.

3.2. Characteristics of the Extractable Portion of Skin and Seed Flavonoids

Studies based on the extraction of intact skins in wine-like solutions have showed that the concentration of extractable anthocyanins and skin PAs increased during ripening, even if the total amount of these compounds remained constant or even decreased [1,82,84,117,118,119]. Contrasting behaviour was observed with extractions conducted on ground skins of Cabernet Sauvignon and Carménère [120]. Moreover, homogenised Monastrell skins released lower amounts of extractable anthocyanins with the progression of ripening; however, the wine produced with riper grapes showed the highest colour intensity [121]. Considering these contrasting results, it appears that analysing the evolution of the extractable flavonoids on ground or homogenised skins may lead to incorrect results, as the mechanical disruption of skin tissues may overcome the effect of the cell wall degradation which occurs during ripening [122] and which modulates the effective release of phenolic compounds [86].

Conversely, the seed flavanols extracted under wine conditions are reported to decrease during ripening [82,84,123], following the progressive hardening of the seed coat due to lignification [92,103].

The composition of the extractable flavonoids did not show substantial changes during ripening, except for the monomeric epicatechin-gallate of Monastrell seed which decreased after veraison and was no longer detected beginning a month before the harvest and onwards [123]. Finally, the mean degree of polymerisation of skin and seed PAs did not change with the progression of ripening [82].

The astringency of skin and seed extracts, assessed using ovoalbumin as a precipitation agent, was lower at harvest [117], even when the extractable flavanols were higher than those at the previous sampling date [82]. The latter finding led to the hypothesis that other berry constituents might markedly affect the perception of astringency and, consequently, also the progression of phenolic maturity.

4. Role of the Cell Wall Material of Flesh and Skin in Phenolic Maturity

4.1. Characterisation and Evolution of the Cell Wall Material Constituents

In the berry, the structure of flesh (mesocarp) and skin (exocarp) tissues is maintained by the cell wall, which is composed of cellulosic microfibrils, hemicelluloses, pectins, lignin and structural proteins. Polysaccharides may account for 90% of the cell wall material (CWM) while the remaining 10% is mainly represented by proteins [124]; however, it is important to note that the composition of CWM varies broadly according to the cultivar [125]. Flesh contains higher concentrations of pectic polysaccharides and proteins than skin [126]; however, it has been reported that up to 75% of the whole berry CWM might be recovered in the skin [127].

During ripening, changes occur to polysaccharides, such as the solubilisation of pectins which determines berry softening [128,129] and simultaneously, the turnover of proteins determines the drastic increase in hydroxyproline-rich proteins, such as extensins, which may contribute to maintaining the structure of the cell wall, counteracting the ongoing loosening of the berry tissues [122,126].

4.2. Interactions between the CWM and PAs

The modifications which occur to the skin and flesh CWM from veraison to harvest have important technological implications as the CWM interacts with phenolic compounds, in particular PAs, modulating their extractability and their precipitation during vinification.

It is well known that PAs may be bound by the CWM via hydrogen bonds and hydrophobic interactions [130]; between the CWM constituents, proteins showed the highest binding capacity [126,131]. Considering the polysaccharides, pectins showed higher affinity for PAs than hemicellulose [132,133] while weak interactions were found with cellulose [134]. Comparing the binding capacity of the CWM of different varieties, that of Syrah, which is characterised by the lowest amount of pectins, retained lower quantities of PAs than Cabernet Sauvignon and Monastrell [135].

The ability of the CWM to bind PAs increases with the molecular weight and the percentage of galloylation [134] as larger polymers and the galloylated forms present more reactive sites which allow the associations [136,137,138].

Studies in which the same amount of flesh and skin CWM were combined with PAs of different varieties showed that flesh CWM has the ability to bind a higher quantity of PAs as compared to that of skin [139,140], and also that flesh CWM interacts selectively with high molecular mass PAs while skin CWM showed lower affinity for those compounds [141,142].

The effect of ripening on the properties of the CWM is different between flesh and skin; flesh CWM did not display any changes in its binding capacity during ripening while skin CWM showed the highest affinity for PAs at harvest, in particular for those of higher molecular mass and for the galloylated forms [143,144]. The variation of skin CWM binding capacity may be linked to the increase in porosity of the skin cell wall which allows the incorporation of larger molecules [143], and also with the increase or the turnover of protein which determines the accumulation of hydroxyproline in the skin cell wall [145].

Therefore, with the progression of ripening, skin CWM seems to be responsible for limiting the extraction of those compounds which, more than others, elicit the negative sub-qualities of astringency [144].

Finally, it appeared that the CWM interacted with PAs in two ways during vinification; skin CWM, binding these phenolic compounds, reduced their extraction from the berry while flesh CWM, which is easily released in the fermenting wine, may remove PAs of higher molecular mass by precipitation [146].

4.3. Interactions between CWM and Anthocyanins

The extraction of anthocyanins was mainly limited by ionic interactions with pectins of CWM [147] while proteins showed a lower binding capacity [148]. Moreover, the affinity of the CWM for acylated anthocyanins was slightly higher than that for non-acylated anthocyanins [147]. These results may explain the differences in the anthocyanin extractability found between varieties characterised by different CWM compositions [149] and also the difficulty in extracting anthocyanin from grapes characterised by both the elevated concentration of anthocyanins and high quantities of skin CWM [150,151]. Finally, Bautista-Ortín and co-workers [152] reported that anthocyanins may lessen the CWM adsorption site to PAs, increasing their extractability.

5. Conclusions

The phenolic maturity of Vitis vinifera L. has been studied for a long time so as to understand how the changes occurring to the berries during ripening may affect wine colour and textures. In order to achieve this goal, researchers in many areas of the wine sector have participated in this effort; in fact, as seen in this review, contributions have come from experts in berry ripening physiology, winemaking procedures, wine chemistry and sensory analysis.

Thanks to their work, some key features of phenolic maturity can be outlined:

• anthocyanins contribute to wine colour and may have beneficial effects on astringency;
• skin flavanols mainly contribute to ‘pleasant’ astringency while seed flavanols contribute to ‘unpleasant’ astringency;
• extractable skin flavonoids usually increase during ripening (irrespective of the evolution of their total amount) while that of seeds decreases;
• cell wall material affects the presence of anthocyanins and proanthocyanidins in wine, by limiting their extraction and enhancing their precipitation into the wine;
• the affinity of skin CWM for high molecular mass proanthocyanidins and for the galloylated forms increases during ripening;
• the use of hydroalcoholic solutions gives the most reliable results for a comprehensive evaluation of anthocyanin and flavanol extractability; however, these methodologies cannot be routinely applied in wineries;
• protein-precipitation methods are easy to carry out and give good results in the prediction of astringency.

As a concluding remark, it came to light that the optimal phenolic maturity for the production of high-quality red wines could be achieved by prolonged ripening as the amounts of extractable anthocyanins and proanthocyanidins of skin would be higher, and the concentration of seed flavanols would be lower. Moreover, the CWM would selectively reduce the presence of those proanthocyanidins more involved in the perception of ‘unpleasant’ astringency.